Annals ^ 
of the 
Missouri 

. Botanical 


Garden 
— v 


olume 93 


umber 


Annals of the 
Missouri Botanical Garden 


Volume 93, Number 1 
May 2006 


The Annals, published quarterly, contains papers, primarily in systematic botany, 
contributed from the Missouri Botanical Garden, St. Louis. Papers originating outside 
the Garden will also be accepted. All manuscripts are peer-reviewed by qualified, 
independent reviewers. Authors should write the Managing Editor for information 
concerning arrangements for publishing in the Annars. Instructions to Authors are 
printed in the back of the last issue of each volume and are also available online at 


www.mbgpress.org. 


Editorial Committee 
Victoria C. Hollowell 
Scientific Editor, 

Missouri Botanical Garden 
Beth Parada 

Managing Editor, 
Missouri Botanical Garden 
Diana Gunter 

Associate Editor, 

Missouri Botanical Garden 
Barbara Mack 

Editorial Assistant, 
Missouri Botanical Garden 
Adele Niblack 

MBG Press Assistant, 


Missouri Botanical Corden 


Ihsan A. Al-Shehbaz 
Missouri Botanical Garden 
Gerrit Davidse 

Missouri Botanical Garden 
Roy E. Gereau 

Missouri Botanical Garden 
Peter Goldblatt 

Missouri Botanical Garden 
Gordon McPherson 
Missouri Botanical Garden 
P. Mick Richardson 
Missouri Botanical Garden 
Charlotte Taylor 

Missouri Botanical Garden 
Henk van der Werff 


Missouri Botanical Garden 


For subscription information contact ANNALS 
or THE MISSOURI GARDEN, Y Allen Marketing 
& Management, P.O. Box 1897, Lawrence, KS 
66044-8897. Subscription price for 2006 is $160 
per volume U.S., $170 Canada & Mexico, $195 
all other countries. Four issues per volume. The 
journal Novon is included in the subscription 
price of the ANNALS. 


annals@mobot.org (editorial queries) 
http://www. mbgpress.org 


Ec ] jM J 


THE ANNALS OF THE MissoURI BOTANICAL GARDEN 
(ISSN 0026-6493) is published quarterly by the 
Missouri Botanical Garden, 2345 Tower Grove 
Avenue, St. Louis, MO 63110. Periodicals post- 
age paid at St. Louis, MO and additional mail- 
ing offices. POSTMASTER: Send address changes to 
ANNALS OF THE Missourt BOTANICAL GARDEN, Do 
Allen Marketing & Management, P.O. Box 1897, 
Lawrence, KS 66044-8897. 


indexed in AGRICOLA (through 1994), APT Online, BIOSISO, CAB Ab- 


strac/Global Health databases, ingenta, ISI databases, JSTOR, Research Alert®, and Sci Search®. 
Typesetting by Allen Press, Inc. using Advent3B2 automatic paging system. 


© Missouri Botanical Garden Press 2006 


The mission of the Missouri Botanical Garden is to discover and share knowledge about plants and 
their environment, in order to preserve and enrich life. 
This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 


Volume 93 
Number 1 
2()0soun BOTANICAL 


JUN 0 6 2006 
GARDEN LIBRARY 


Missouri 
Botanical 


Garden 


SPECIES RECONSIDERED: 
CONSEQUENCES FOR 
BIODIVERSITY AND EVOLUTION. 
INTRODUCTION” * 


P. Mick Richardson? 


The Annual Systematics Symposium that took place 
at the Missouri Botanical Garden in 2003 marked 
a major DK in the history of scientific discourse 
in the U.S.A. The Systematics Symposium is remark- 

able for its e run of : 
Also noteworthy 


50 years at the same 


institution. is the fact that the 
National Science Foundation has supported 48 of 
the symposia. What is the explanation for such a long 
run of successful symposia, and why does each 
symposium continue to fill the seats of a large 
auditorium? Certainly, some of the audience comes 
for an annual visit to the massive herbarium or the 
comprehensive library, to visit a large city for 
shopping, or for the chance of an exotic meal, but 
most of the audience come to hear the symposium 
speakers. There 
symposium because every year the list of speakers is 


chosen by a new set of organizers. This year was no 


is a distinct. freshness to eac 


exception, and the organizers comprised a zoologist 
and a botanist. Jonathan Losos and Peter Stevens 
invited a diverse group of speakers to shed some light 


on the perennial subject of *the species question." 


In order to bring attention to a display of Linnaean 
material in the Monsanto Research Building, the new 
location of the library, a Special Linnaean Celebration 
Talk was arranged on the Friday evening, an atypical 
time for one of our symposium talks. This was 
delivered by Mary P. (“Polly”) Winsor. Although it 
is published here with the succinct title “Linnaeus’s 
biology was not essentialist,” it was advertised and 
delivered with the intriguing title *Would you like 
your species concept rare, sir, or well-done? (A look at 
how the history of science is cooked, how you can tell, 
and whether it is any good for you anyway).” The room 
ed to overflowing, and especially notable in 


— 


was fil 
the audience were the local history of science pro- 
fessionals, a group not always seen at the symposium. 

Saturday, as usual, was the main day of the 
symposium. Jonathan Losos moderated the morning 
session, and Peter Stevens was the afternoon moder- 
ator. Peter Raven introduced the evening speakers, 
a husband and wife team of scientists. All but one of 
the speakers has submitted à manuscript in time for 


publication here. 


This and the seven articles that follow it are the proceedings of the 50th Annual Systematics Symposium of the Missouri 
Botanical Garden, Species Reconsidered: Consequences, for Biodiversity and Evolution. The symposium was held 10-11 October, 


E Missouri Botanical Garden in St. 


2003, at 


Losos and pen ens for selec 


their ru e beue an 1 1 rias assistance, M 
Freire-Fierro for her expertise in producing the sym 
en website in a timely fashion. Last but not least, I wish to thank the speakers for making the symposium so wonderfully 


process, and Alina 
Gard 
interestir Ing. 


ouis, Missouri, U. 
ae was Posee in part by the National Science Foundation under grant DEB-9981642. I thank Jonathon 
ig an interesting diversity of speakers. I also 


thank Sandy Lopez and Mary McNamara for 
ary Merello and Amy Pool for their help with the registration 
osium booklet and getting the information onto the 


2 The editors of the Annals thank Sophia Balcomb for her editorial contribution to these papers. 
S.A. 


3 Missouri Botanical Garden, P.O. Box 


299, St. Louis Missouri 63166, U.S 


ANN. Missouni Bor. Garp. 93: 1. PUBLISHED ON 31 May 2006. 


LINNAEUS'S BIOLOGY WAS 
NOT ESSENTIALIST' 


Mary P. 


Winsor? 


ABSTRACT 


method of definition taught by medieval followers of 


and “species” ruined the meaning they had in logic. and * 


he current pieture of the history of taxonomy incorporates x J. 
ini Aristotle. 
Contrary to some published statements, there is no evidence a Linnaeus ever studied logic. His 


Cain's claim that Linnaeus strove to apply the logical 
‘ads argument does not stand u p toe rilical examination. 
"genu is” 


The 


use of the were 


“essential” meant to him merely “taxonomically usefu 


essentialism story, a narrative that has most pre Darwinian biologists steeped in the 17 view of Plato and aen is ill- 


founded and improbable. 


Key words: A, J. Cain. 


essentialism. 


history of systematics. 


history of taxonomy. Linnaeus, 


Mayr. 


The Missouri Botanical Garden has celebrated the 


250th anniversary of the publication of Linnaeus’s 


Species Plantarum (Linnaeus. 1753). but there are 
doubtless some biologists who are not sure if they are 
proud of that landmark. After all, we have now 


departed so far from the beliefs of that eccentric old 
Swede that it is rather embarrassing to count him 
among our intellectual ancestors. Didn't he think that 
every species was directly created by God. in other 
words, wasn’t he a creationist and thus an enemy of 
evolution? More profoundly, didn’t Linnaeus believe 
that every species has its own essence, that is, type in 
the Platonic sense? Well, to be blunt, no. and no. His 
views on the fixity of species changed in his lifetime, 
and the business about essentialism is the scholarly 
equivalent of an urban myth, that is, a story everyone 
repeats but for which there is scant basis in fact. 
That the mature. Linnaeus abandoned his vouthful 
insistence on the fixity of species was pointed out by 
Edward Greene (1909) and fully documented by John 
Ramsbottom (1938). Linnaeus was perfectly comfort- 
that God His 


f creativity, per 


able imagining may have made, in 


burst 


- 


original only one species 


genus, while natural processes later caused the 


emergence of the others. well 
that 


fixity to 


Although it is very 


known Linnaeus underwent this shift from 


transmutation (Larson, 
Müller-Wille, 1999), 
it so clearly contradicts the urban 
The 


Platonic types. we have been told time and again. is 


limited 
1999: 


because 


absolute 
1971; 


vou of il 


Koerner, | remind 


myth about essentialism. whole point about 


that there are unbridgeable gaps separating them. Did 


the mature Linnaeus think that some species were 
created with an immutable essence and others with 


mutable essences? Ernst Mayr deserves credit. for 


admitting the problem (1982: 259), but he leaves us to 
conclude that Linnaeus was hopelessly confused and 
inconsistent. | suggest that it is we who are confused. 
for we attribute to Linnaeus a philosophical notion he 
never held. 

Actually. I suspect that many of us are allowing the 
spectre of modern creationism to bias our understand- 
ing of a pre-Darwinian creationist like Linnaeus. To 
a modern biologist who does not believe in God, 
Linnaeus’s explicit piety has the effect of making 
the essentialism story seem plausible, but the link 
connecting these realms was the insight of a later 


In 1857 


rising interest in Men 


generation. Louis Agassiz, disturbed by the 


«T 


argued in his “Essay on 


Classification" that the Creator had conceived each 
species in his divine intellect before giving it material 
1991). 


after Darwin’s revolutionary book appeared in 1859, 


existence (Winsor. He repeated these ideas 
and Agassiz remained adamant that the patterns 
recorded by taxonomists were direct evidence of God's 
thoughts. But all this was a century after Linnaeus. In 
Linnaeus’s day the “requirements of Christian faith.” 
(1982: 259), 
the possibility that the “kinds” mentioned in the Book 
rank of 


Those modern biologists 


contrary to Mayr's claim did not forbid 
of Genesis could have been at the taxonomists’ 
genus rather than species. 
who do believe in God will have more sympathy with 
Linnaeus than with Agassiz, because they accept that 


the Creator willed into existence our actual world. 


‘My research was supported by a grant from the Social Sciences and Humanities Research Council of Canada. for which | 


am very grateful. For their 


Jennifer Hertz. 


assis ance 


Eie Paradis, and Dennis Yeung. 
to the Missouri Botanical Garden's 


a ple asant v 


isi 
“Institute for he History and Philosophy of Science and Af 


Ontario M58 IK? 


93: 2-7. 


Park Crescent East, Toronto, Canada. polly. 


ANN. Missouri Bor. GARD. 


and encouragement I also thank Stephan Müller-Wille and my students Abby * 
| am particularly 


Slinger. 
indebted to Peter Stevens and Peter Raven for inv iting me 


50th Annual Systematics Symposium, and I thank Mick Richardson for arranging such 


Victoria College, University of Toronto. 73 Queen's 


1 
tLOPONtO.Cd 


PUBLISHED ON 31 May 2006. 


Volume 93, Number 1 
006 


Winsor 3 
Linnaeus's Biology 


including life with all its evolutionary complexity and 


history. Those biologists realize that belief in God 
does not logically entail belief in the fixity of species. 


The mature Linnaeus's limited transmutation was, 
Mayr (1982: 259) admits, 


everything he had said and believed before but was in 


“not only inconsistent with 
fact irreconcilable with essentialism.” It is time, then. 
to look more carefully at this supposed. doctrine or 
dogma of essentialism. Without a doubt. there does 
exist today a story about essentialist beliefs in past 
centuries. It is this story, that is, this historical claim 
about the world-view of past taxonomists, that I am 
challenging. (Some philosophers are now refurbishing 
the word essentialism for variously modernized 
concepts of natural kinds, but the biological and 
historical literature to which I refer was innocent of 


Whether 


kinds corresponding to such 21st century usage were 


such sophistication. concepts of natural 


held in the past is an entirely separate question nol 
relevant here [Boyd, 1999: 152; Ellis, 2002].) I begin 


by laying before you only one telling of the 


essentialism story, but it stands for other repetitions 
un- 


that number in the thousands, if we include 


— 


dergraduate lectures. The latest edition of one of the 
most widely used textbooks, Evolutionary Biology by 


Futuyma (1998: 448), puts it this way: 


Linnaeus and other early taxonomists held what E rnst 
ayr (1942, 63) has called a “TYPOLOGICAL” or 
"ESSENTIALIST" NOTION OF SPEC 

^ a given species if the y sufficiently conformed to that 


S. Individuals were members 


certain characters that were 


A 


a concept descended from 


dae | properties 


ixec 
Plato’s 1 (see chapter 1). 


Here are the relevant passages in his chapter 


(Futuyma 1998: 6): 
evolutionary wherein natural 


In positing an process 


selection sorts among hereditary variations, 
identified variation as a centrally 
biological systems. In doing so, he broke with 
year-old tradition that had "dum dated Western thought. 
The tradition stemmed from Plato, whose philoso] hy was 
built on the concept of the “eidos,” the “form” or “idea,” 
a transcendent ideal form impe erfe ctly im 
earthly representations. In hi: 
cave in The Republic, P late likened earthly 
as the triangles or horses we are familiar 115 
shadows cast on the wall of a cave by objects that pass by 
the entrance. Like people within the cave, bound so that 
they face the wall, we see only the shadows, the imperfect 
representations, of reality. Likewise, the reality—the 
ESSENCE—of the 
fectly captured by the 
of which are imperfect, and vary from the true, acier 


true equilateral triangle is only 1 "ai r- 


e draw or construct, 


triangles w 


Daru win 


? Mayr does set out the essentialism story on pp. 4-6 and 
16-17 s his 1963 book tua only the term "typological") 


but not in his 1942 bo 


And 


each has an 


triangle. so it is with horses, or any other species: 


immutable essence 
individual has imperfectio ions. In this 
ESSENTIALISM, variation is accidental eee ae tion; only 


essences matte 

Plato’s philosophy of essentialism became incorporated 
into Western philosophy. Its central tenet was that 
1owever much the objects in a class might accidentally 


— 


vary, the class still had a defining essence that could not 


horse, zebra. or ass, for 


Thus each species 


instance—has an essence, and one cannot be changed 
into another any more than a triangle can vary enough to 


become a rectangle 


The chief source of this historical narrative is Ernst 
Mayr (1959, 1963, 1968, 1976, 1982 


carried such authority that his claims have been 


). whose words 
have 
rarely questioned. Yet in its broad sweep across the 
history of systematics, this story is not merely 
inaccurate in ds it is wrong and harmful i 

ils basic message. According to the essentialism story, 
Platonic idealism dominated Western thought until 


act, William of 


century and other nominalists 


Darwin broke its spell, whereas in 
Ockham in the 14th 


dealt it erushing blows from which it never recovered. 


— 


Systematic biology evolved, largely independent of 


philosophy, from the 16th century onward, through 
the actions of an army of herbalists, encyclopedists, 
makers and cataloguers of collections, and other 
naturalists, whose joint efforts built up the mountain 
of data that Linnaeus confronted. None of Linnaeus's 
numerous enemies noticed any remnant of Plato’s 
ideal forms in the series of catalogues the Swedish 
professor and his followers kept churning out. Quite 
the contrary, one philosopher later noted with surprise 


and respect that naturalists’ success in their massive 


project of inventory seemed to involve an active 
neglect of the classic rules of logical definition 


(Whewell, 1847; Winsor, 2003). By suppressing this 
rich history, the essentialism story distorts the 
historical and logical foundation of Darwinism. 

To fully understand the creation and impact of the 
essentialism story would require us to take it apart 
and examine it piece by piece, but at present I will 
limit my attention to the claim that Linnaeus himself 
was in thrall to this philosophy. Mayr's sources for this 
claim are impressive, including such scholars as 
James Larson (1971) and Frans Stafleu (1971). Yet 
a close examination of all Mayr's sources reveals that 
instead of a literature of accumulating evidence, all 
derive from a single source, an article by Arthur J. 
Cain in the Proceedings of the Linnean Society of 
London (1958). There the Oxford zoologist maintained 
that Aristotelian 


essences; the 


Linnaeus thought in terms of 


backbone of his claim is that “the 


— 


method [Linnaeus] adopted was to classify by the rules 


of Logical Division, which involve the determination 


Annals of the 
Missouri Botanical Garden 


1958: 


Cain’s conclusions are now woven into the 


of the essence of each entity” (Cain, 


102). 


established knowledge, repeated. countless times by 


people who have never read him. Scholars who 


subsequently contributed fresh research on Linnaeus, 


ironically including Cain himself (Broberg, 1985: 
Stevens & Cullen, 1990; Cain. 1992, 1993. 1994, 


1995; Miiller-Wille, 1999) are perceived, if they are 
read at all, as supplementing Cain's 1958 conclusions. 
even though they paint a very different picture. 

A few years ago | undertook a careful examination 
of Cain’s 1958 article as part of a project concerned 
with scientists turn their attention to 


why some 


questions about the past (Winsor, 2001). My intention 
was lo investigate the phenomenon of scientists 
turning to history, and both Cain and Mayr were in 
the prime of their scientific careers when they began 
to write about history in the 1950s. My interest was 
but I 


weaknesses 1 


Cain’s motivation, not. Linnaeus's reputation, 


— 


found myself amazed by the many 


1 
Cain’s argument. It would seem that Cain's conclu- 
sions took hold because they meshed so well with two 


other 


— 


semi-independent and simultaneous lines o 
thought. One was Mayr's association of the morpho- 
logical concept of species with typology, which 
included tracing typology back to Plato’s cave (Mayr 
el al, 1953: 15; Mayr, 1976: 256-257). Another 
strand flowed from David Hull's 1965 article 
effect 


“The 
of essentialism on taxonomy—two thousand 
years of stasis.” The essentialism story known today 
consists of elements from each strand, which were 
twisted tightly together in 1968 when Mayr decided 
that the concept of species he called 
thinking 


essentialism he found in Hull. 


“typological 
' could be oed with the concept of 
intend to trace out 
elsewhere the distinct history of those three strands. 


Cain believed he had made a breakthrough in 
understanding Linnaeus when he learned that Aristo- 
telian logic mandated that definition should proceed 
by stating the kind (genus in Latin) to which a species 
belongs (man is an animal) and then stating the 
differentia, 


other members of that kind (man is a rational animal). 


the features that distinguish it from the 
Linnaeus’s rule that every species name must begin 
with the name of the genus to which it belongs was 
what had first made Cain suspicious, and then the 
differentia, 
seemed to confirm it. Indeed the case seemed beyond 
doubt 


telltale words definitio, and essentialis 


when Cain found a series of incriminating 


pronouncements in Linnaeus’s Philosophia Botanica, 
which he rendered : 
that 


peculiar to it, if there is one such, which will instantly 


s “The Essential Character. « 


a genus is which gives some characteristic 


serve to distinguish it from all others in the same 


natural order [Phil. Bot., 187] .... The true specific 


fabric of 


name is a Differentia essentialis (Phil. Bot., 


distinguishing that species from al 


251) 


others in the same 


genus.... The Character essentialis |of a species] 
a Differentia. (Phil. BO, 258)" (Cain 1958: 148). 
These words of Linnaeus equipped Cain to argue that 
however messy and complicated his system became in 
practice, fundamentally Linnaeus must have been 
engaged in this sterile scholastic game. 

The game was indeed scholastic, for the procedure 
of formulating definitions per genus et differentiam was 
familiar 


an exercise lo generations of scholars i 


medieval universities, based upon the 6th century 


scholar Boethius's commentaries on and Latin trans- 


lation of the 3rd century Greek scholar Porphyry's 


introduction to Aristotle (Kretzmann et al., 1982; 
Spade, 1994). Porphyry’s notorious “five words” are 
rendered by Boethius genus, species, differentia, 


propria, and accidens. What Cain did not realize was 


the extent to which teachers in Linnaeus’s day, and 
indeed many of the medieval schoolmen, regarded the 
whole business as a taxonomy of words. 
(Arnauld & Nicole, 91 5 (What 


would have thought of it [Balme. 


not things 
Aristotle himself 
1980, 1987: 73: 
Gotthelf, 1985] is quite nu matter.) If the task is 
to define the word 
the 


“horse,” a schoolboy learned not to 


mention features peculiar to his own mare 


Rosalind, such as her location and color, and not to 


— 


compose a tedious list of all those features, such as 


being four-legged and having a mane and tail, that she 
shares with asses and zebras; the frugal and proper 


approach is simply to state the group name (“equine”) 


plus the features. distinguishing horses from others 
that this 
elementary exercise arouses the thorny question of 


of this group. It was widely understood 


whether the abstract “horse” (the universal). the idea 


Rosalind 


belongs, really enjoys existence in some eternal place, 


or type of horsiness to which somehow 


as Plato said, or has a more limited existence as the 
active power within each individual horse. as Aristotle 
sald. or exists only nominally, as a product of our 
own em as William of Ockham said. Philosophers 
debated the question, but the relevant point here is 
that their debate did not touch the rules of proper 
definition. Even for a nominalist, it is efficient to 
describe an object that resembles other objects by 
first stating the set of similar things, and then pointing 
to the features by which this one differs. 


The supposedly essentialist views Cain quoted from 


Annaeus could have been multiplied, for Linnaeus 


also declared “The concept of a species consists of an 


essential feature, by which alone it is distinguished 


from all others in the same genus.... specific 


definition contains features in which the species 
differs from those in the same genus. But the specific 


name contains the essential features of the defini- 


Volume 93, Number 1 
2006 


Winsor 
Linnaeus's Biology 


tion.... Therefore the specific name is the essential 
definition” (Linnaeus, 1751: 219-220). (Note that for 


Linnaeus, the “name” of the species was not ils two- 


— 


word shorthand form, but the genus name modified by 
up to a dozen adjectives.) However, it would be a sin 
against a prime principle of historiography (under- 
standing actions and words in the context of their own 
time) if we were to assume that Linnaeus was using 
the word essentialis in the same sense as medieval 
as well as his own 


philosophers had. The context, 


definition, shows that the word only meant “taxonom- 
ically useful” and nothing more. What he called the 
"character" of a genus was the list or suite of features 
found to be dependable, and these he categorized with 
three adjectives: factitius, essentialis, and naturales. 
The first was a single feature used in some botanist's 
artificial system; the second was a single feature, or as 


few as possible, peculiar enough that it serves to 


distinguish this genus from the other genera in its 


natural order. The third was the full list of features. 


e character essentialis was desirable because it 
enabled the production of a succinct catalog. Nowhere 
did Linnaeus suggest that the "essential" features 
were any closer to the inner nature of a plant than its 
that 


Linnaeus 


other features. On the contrary, he insisted 


botanists pursue the characteres naturales. 


(1751: 


14.3) wrote, 


If the essential characters of all genera had been 
discovered, the recognition of plants would turn out to 
be very easy, and many 911 undervalue the natural 


characters, to their own loss. But they must understand 


that, without regard for the natural character, no one will 


turn out to be a sound botanist; for when new genera are 


2 


discovered, the botanist will always be in doubt if [he] 
Anyone who thinks that he 


E: 


neglects the natural character. 


understands botany from the essential character and 


disregards the natural one is therefore deceiving and 


deceived; for the essential character cannot fail to be 


deceptive in quite a num of cases. The natural 
character is dis 1 ona of is genera of plants. and no 
one has ever made a proper judgement shout a genus 
without its help; and so it is and always will be the 


absolute foundation of the understanding of plants. 
In erecting these terms, Linnaeus may have had in 
mind the famous exchange between John Ray and 
Joseph Pitton de Tournefort (Sloan, 1972: McMahon. 
2003), but if 


his nose al 


— 


So, he was siding with Ray and thumbing 


Tournefort, by choosing to define 
essentialis solely by its taxonomic convenience rather 
than by reference to a plant's essential nature. 

The telltale words that had first caught Cain's 


de actually testify most 
that 


division. As 


attention, "genus" and "species." 


damagingly against his claim Linnaeus was 


following the rules of logical every 


medieval schoolboy knew, Boethius’s genus and 


species were relative terms. One may correctly say 


that “bird” is a genus containing the species “swan,” 


but it is equally true that “bird” is a species in the 


genus “animal.” A person devoted to scholastic 
principles, who decided to set up a system running 
from Kingdoms containing Classes to Classes contain- 
ing Orders, would be obliged to avoid the logical terms 
genus and species for the next two ranks, for each Class 
is a species with respect to its Kingdom, but a genus 
with respect to its Orders. The fact that Linnaeus 
appropriated these old words and spoiled them by 


attaching. them absolute ranks in his hierarchy 


proves his utter disregard for the “Aristotelian” rules 
f logic. 
| am myself no expert on Linnaeus, so | was 


relieved to find that the skepticism to which I was led 
of Cain had already been 
Staffan Miiller-Wille 


(1999) through his close reading of Linnaeus. His 


by a close reading 


inde pendently adopted by 
work, which is not yet as well known as it deserves to 
be, makes me confident that the tide of opinion on this 
issue must turn. Yet the process will not be easy, for 
the essentialism story can bias our reading without our 
noticing. Consider one significant nm from Mayr's 
Growth of Biological Thought (1982: 


and he was evidently 


173): “In school 
Linnaeus had excelled in logic, 
deeply impressed by the precision of this method.” 
Here we have an instance of history by deduction, for 
Cain’s argument certainly does require us to imagine 


that Linnaeus had mastered scholastic logic. Yet Cain 


had no evidence to that effect, so he could only 
beguile his readers by saying that "every well- 


educated man in and before Linnaeus's time might 
have been given some instruction in the principles of 
Aristotle. 
” (Cain, 


he ambiguity of the pronoun “he,” 


classification in general, as laid down by 
This he would receive in the study of Logic... 
1958: 145). Notice 
which could easily be misread as “Linnaeus” though it 


woperly refers only to some “well-educated man.” 


T 


Leaving aside the quibble that Linnaeus was not 
particularly well-educated by 18th century standards, 
Cain’s claim is true only because of his carefully 
qualified verb *might have been given." An educated 
man might have studied Aristotelian logic, but then 
again, he might not. 

Actually there is no evidence that Linnaeus studied 
logic at all, much less excelled at it, according to 
Müller-Wille 


assertion of excellence in school is peculiarly at odds 


(pers. comm.). Indeed the positive 
with the slight evidence we do have. Among Mayr's 
sources, the one that reports on Linnaeus's schooling. 
Blunt (1971: 18), says this: 

In 1724 he passed, though with no great distinction, into 
Ed curriculum was designed t to fit 
esthood; t 


aa 


the Gymnasium. Her 


[m 


the needs of boys intended for the e 


emphasis was upon Greek, Hebrew meta- 


Annals of the 
Missouri Botanical Garden 


physies and oratory—subjects in which he was little 


gifted and even less interested. Apparently he shone : 


physics and mathematics. . . . 

This is an accurate summary of Linnaeus's own 
which claims excellence in science 
(Fries. 1923). The 


erroneous tidbit about his not only studving logic but 


autobiography, 
and makes no mention of logic 
hard 


being deeply impressed by it may now be 


erase, however, because of being stated so clearly in 
Mayr's authoritative book. 

The essentialism story that engulfed and incorpo- 
1958 


Absent from Cain’s article 


rated Cain's argument also transformed it. 


is the very word "essen- 
tialism,” which had been used by Karl Popper (1944: 
94) to expose a fallacious and 


thought. The fallacy is one into which reasonable 
people are prone to fall, namely, to believe that we can 
find out the true nature of something by concocting 
that (Actually 


Oth century scientist 


a definition of the word names il. 


Popper assumed that no 2 really 
believed this, so he aimed his criticism at “method- 
ological essentialism,” which means acting as though 
one believes it.) Popper's 1945 book The Open Societ 
and Its Enemies further developed the concept, but 
Mayr took 


which 


there is no evidence that either Cain or 
Hulls 1965 


Popper's word and concept, was almost certainly how 


notice. David paper. featured 


Mayr learned of them; Hull mailed him a copy of his 


paper. which was the start of Mayr’s mentorship of 
Hull. (Mayr very graciously allowed me to examine 
their correspondence. along with his other papers in 


Archives.) Hull 
that there were some problems with Cain's argument 
(Hull, 1965: 316n, 1967, 1985), but he did not quarrel 


with 


the Harvard University recognized 


Cain’s basic message, that Linnaeus was in the 


business of defining species like a logician rather 


than describing them like a natural historian. The 


connection between Poppers concept and what 


bothered Mavr—"tvpological thinking" about spe- 
: MI 8 8 | 
(Mayr Plato 


rather than Aristotle the enemy), but in general, Cain’s 


cies—was somewhat loose considered 
assessment of Linnaeus seemed to fit into both Hull’s 
and Mayr's views of the history of taxonomy. 

When we 


steeped in the literature and traditions. of early 


consider how deeply Linnaeus was 


modern botany and zoology, is it even likely that he 
would have fallen into the fallacy of essentialism? We 
need to realize that Popper was by no means the first 
thinker to warn people against it. The rhetoric of the 
Scientific Revolution, including the writings of Bacon, 
Descartes, and Locke, was consistently anti-scholas- 
tic. It was clear to every student of Linnaeus that the 
business of the true botanist was to interrogate nature: 


after all. he had emphasized that his own classes and 


dangerous habit of 


orders were artificial devices, invented to ease 


identification. 
What about Mayr's concern that individual differ- 
ences, so erucial for Darwinian evolution, would have 


been discounted as mere accidents by any follower of 


classic logic? Cain later showed (1996) that the 
logicians concept of “accident” had considerably 


faded by John Ray’s time, but Mayr was surely correct 


that taxonomists wanted to find constant characters 


and did their best to weed out variable characters. 


This was obviously a practical issue, but whether the 
concepts of scholastic logic played any role at al 
18th century taxonomy, or even in the two previous 
centuries, entirely remains to be demonstrated. 
Unfortunately, the essentialism story has tended to 
dampen interest in the rich and complex story of 
before and after 


axonomists, both 


exactly how 


Darwin, coped with the real-world challenges of 


comparing and identifying organisms. To loosen its 


grip. we could begin by recognizing that maligning 
Linnaeus distorts our understanding of the entire 


history of systematics. 


Literature Cited 


Arnauld, A. & P. Pe 1662 (1996). Logic or the Art of 
Thinking. ] uroker (translator). Cambridge Univ 


Balme, D. M. 1980 o biology was not essentialist. 
Arch. Gesch. Philos : 1-12. 

— 1981. vetas use ye division and differentiae. 
Pp. 69-89 in A. tthelf J. G. Lennox (editors). 


Ar ee s Biology. Cambridge 

. The € omple at Linnaeus: A Life of Linnaeus. 

iking p New York. 

Boyd, R. 1999. 
Pp. 141-185 in Species: New Interdisciplinary Essays, R. 
V. Wilson 1 MIT Press, Cambridge, 

Broberg, G. 198 


history discoveries. Pp. 1 


Homeostasis, species, and higher taxa. 
Massachusetts. 
. Linné's systematics and the new iod. 
53-181 in J. Weinstock (edito 
Contemporary Perspectives on Linnaeus. Univ. 
Maryland. 

Cain, A. J. 1958. pos and memory in Linnaeus's system of 

iudi 5 08 inn. Soc. London 169: 144—103. 


e Methodus of Arch. Nat. Hist. 


P ress E 


America, Lanham, 


"Linnaeus. 


; 1755 Linnaeus's Ordines Naturales. Arch. Nat. 
Hist. 20: 405—415. 
1994. Numerus, figura. proportio. situs; Linnaeus's 
definitory attributes. Arch. Nat. Hist. 21: 17-36. 
. 1995. Linnaeus's natural and antic al arrangments 


of lanis. Bot. J. Linn. Soc. 117: 131: 
e 96. John Ray on “accidents. An h. Nat. Hist. 23: 

8. 
E n " n 2002. The Philosophy of Nature: A Guide to the 
; Montreal. 
. Linnaeus (afterwards Carl von Linné 


ew Esse 1 MeGill-Queer's Univ. Press, 
M. 19 


fisse 


The Stony of vs | afe, B. D. Jackson (editor and translator 
H. F. & G. Witherby, 

Futuyma, p. J. 1998 
Sunderland. 


London. 
». Evolutionary Biology, 3rd ed. Sinauer, 
Massachusetts. 


Volume 93, Number 1 
2006 


Winsor 
Linnaeus's Biology 


Gotthelf, A. 1985. 


Notes towards a study of s and 
ssence in Aristotle’s Parts of Animals ii—iv. Pp. 25-54 


= 


le Gotthelf (editor), Aristotle on Nature and Living Things. 
Mathesis Publications, Pittsburg. 

Greene, E. L. 1909. Linnaeus as an evolutionist. Proc. 
Acad. Sci. 11: 17-26. 
Hull, D. 1965. The effect 

thousand years of stasis. Brit. J. 
16: 1-8. 
1967. Certainty wy circularity in evolutionary 
aes 21: 15 
Linné as an as otelian. Pp. 
(editor). C 1 ary 


Wash. 


of essentialism on laxonomy—two 
Philos. Sci. 15: 314—326 


= 


FROND 
. 19 37-54 in John 
Weinstoc K i 


. Univ. Press of America, 


Perspectives on Lin- 
Maryland. 


re and Nation. 


Lanham, 


— 


Koer Ther, 99. Linnaeus: Natu Harvard 


]niv. Press, © Cambridge, Massachusetts. 
Kretzmann, N., A. 1982. The 


Rd ud of Later Medie al Philosophy from the 


Kenny & J. Pinborg (editors). 


Scholas- 


— 
wn 
c 


discovery of Aristotle to the Disintegration of S 
ticism "| 1005 l 19 1 uA Univ. Press, 
. 1971. Reason and Experience. Univ. of California 


Cambridge. 


Larson, 


4 


ele 
I (2003). o Botanica. S. Freer 
Oxford Univ. E ord. 

1 Ist. ed. 


(transl: ator). 
— Species Impensis 
Laurentii Salvii, Stock 
Mayr. E. 1942. 
Columbia Univ. Press 
1959 Darwin. 1 the 


olm. 

Systematics and the Origin of Species. 
. evolutionary theory in 
010 Pp. 1-10 in Betty J. 1 (editor), Evolution 
and Anthro] Es i Anthropolog- 
963. Animal Spa ies Band Evolution. Be ‘Iknap Press 


of He me niv. Press, Cambridge, Massachusetts. 
06 


— 1 sen of i, classification. Nature 

220: 545— 2 

. 19 Ad Evolution and the Diversity of Life: Selected 

3e anap Press of Harvard Univ. Press, Cambridge., 
Massachusetts 

1982. T m 


Press of Harvard Univ. 


Essays. 


Growth of Biological Thought. Belknap 
dee 


Press, Cambridge, Massachusetts. 


1953. Methods and 
McGraw-Hill, New 


G. Linsley & R. 
Princ vil s of Systematic 
Yo 


„ singer. 


1 


Me Mahon. S. 2003. John Ray, Joseph Tournefort and 
essences in the seventeenth atto 


Paper de ‘liver red al 
July meeting of the International Society for the 
Philosophy, e Social Studies 

Müller-Wille, S. 1999. Botanik und We ltweiter Handel: Zur 
EH eines Natürli hen Systems der Pflanzen 

^ lar (1 70178). Verlag für Wissenschaft 

Popper. K. 1944 l. The p poverty of historicism J. Economica 11: 

3. 
1945. The 


T Enemies. G. 
Routledge, 


Open Society and Its 
London. 
Ramsbottom, = 1938. Linnaeus and the species concept. 
4 i eir 150: 193-219. 
John Locke, 
problem of 1 dole: system. J. Hist. 


. (editor). 1994. Five 


Problem of Universals: Porphyry, Boethius. Abel: ard, Duns 


m Ray. and the 
53. 


spade, p. Texts on e Medieval 


Scotus, Ockham. Hackett Publishing. Indianapo 
Stafleu, F. A. 1971. Linnaeus and the Linnaeans. ode 


Stevens, P. 


F. & S. P. Cullen. 1990. Linnaeus, the cortex- 
medulla theory, and the key to his understanding of plant 
orm and natural relationship. J. Arnold Arbor. 71: 


179-221 x 
W he well, 1847 (1967). The Philosophy of the Induc- 
Founded upon Their History, 2nd ed. 


e ie nees 1S 
Johnson Reprint Corporation, New 


Fac 8 reprint. 
‘or 


Winsor, M. P. 1991. Reading the 
Comparative Zoology at the Agassiz Museum. 


Nature: 
Univ. of 


Shape of 


— 


icago Press, Chicago. 
The scientist-historian as 


Biol. & Biomed. Sci. 


2001. Cain on Linnaeus: 
unanalysed entity. Stud. Hist. Phil. 
392. $ 3 9c 
2003. Non-essentialist methods in pre-Darwinian 


Bio. & Philos. 18: 387—400. 


taxonomy. 


PROBLEMS WITH SPECIES: David B. Wake? ^ 
PATTERNS AND PROCESSES OF 

SPECIES FORMATION 

IN SALAMANDERS' 


ABSTRACT 


How many species one recognizes within a given taxon remains a difficult question, especially when morphology is 


relatively stable or when clinal variation is present, thus complicating diagnosis. | accept the general lineage concept of 
species, and my goal is to recognize historically distinct evolutionary lineages that are likely to remain distinct. Here | analyze 
this task with respect to patterns of species formation in two genera of plethodontid salamanders in California. Ensatina is 
a ring species complex surrounding the Central Valley of California. At present it is a single species with seven subspecies 
that are linked by apparent elina varialion in intergrade zones, but there are also some narrow hybrid zones where 
morphologically and ecologically differentiated forms interact. In contrast, Batrachoseps, which has much the same 
distribution, has about 20 species in California, most occurring in sympatry with Ensatina. Divergence in the two taxa is base 
on two fundamentally different phenomena, and vet there are some common themes. Adaptive divergence in coloration is the 
dominant theme in Ensatina, whereas differentiation is largely perceived at the molecular level in Batrachoseps. Yet both have 
evolved in the same region and have been affected by many of the same climatic and earth historical phenomena. Within the 
Ensatina complex, different adaptations related to predator avoidance have evolved. Coloration has diverged in different 
directions in coastal and NN populations, even though genetic interactions continue to take place. Where coastal 
populations meet other coastal populations, ecologically and d duet 'ally similar 1 1 ions merge genetically, even if 
well differentiated in molecular traits. In contrast, where the ring is crossed and wher re Bu ally and d rua cally 


differentiated populations meet, they hybridize px or are sympatric and behave as if they are species. Within the ring- 
like distribution, elinal patterns of variation occur. The current polytypic taxonomy is retained, even thot T it is ium lematic, 
because alternatives are even less appropriale. In contrast, where genetically differentiated populations of Batrachoseps meet 


they typically do not merge. Instead, they replace one another spatially, in part because ie are so similar ecologically. 
Apparently the periods of isolation were sufficiently long that even in the absence of adaptive divergence there has been 
divergence of Polane mechanisms. Analysis of patterns of genetic differentiation in allozymes 9905 mtDNA in relation to the 
ecological history of California is used to generate biogeographic scenarios to help e xplain the contrast between Batrachoseps 


Key words: Batrachoseps, California paleogeography, Ensatina, historical biogeography, phylogeography, salamanders, 
ormation. 


species 


Every generation of biologists is destined to deal salamanders is based on such criteria. When 
with species problems, sometimes repeatedly, or, in molecular methods became accessible to systematists, 
my case, continuously throughout my career. |. roughly 35 years ago, many surprises resulted. Kay 
described my first new species in 1962, based on  Yanev. the first to use such methods in my lab, found 
a single specimen from Colombia that had been that allozymes uncovered different patterns than did 
collected in the nineteenth century and assigned to morphology in the Slender Salamanders (Batrachoseps 
a species that was thought, mistakenly, to range Bonaparte, 1839) of California (Yanev. 1978). Several 
throughout Middle America and into South Ámerica nominal species were species complexes, and a num- 
(Brame € Wake, 1962). That species has so far stood ber of undescribed species were identified (Yanev, 
the test of time, although even today it is known from 1978). James Hanken, also in my lab, made similar 
fewer than 10 specimens. The decision to describe the observations on tropical salamanders of the genus 


species was based solely on anatomy and color Thorius Cope. 1869 (Hanken, 1980). Richard Hig 


ion 
pattern, The vast. majority of all known species of was among the first systematists to use molecular 


Many 5 have. c ae d with me in the research summarized here, and I thank especially C. P M. Frelow, 
R. ee . E. Jockusch. S. Kuchta, C. Moritz, T. Papenfuss, D. Parks, C. Schneider, N. Staub, M. Wake, T. Wake. E 
Yanev. B rown permitted me to use many of his excellent photographs of Ensatina. The figures are 5 Karen. Klitz. 
appreciate the helpful comments of Shawn Kuchta and two anonymous reviewers, Financial support for my research from Ses ; 
the Gompertz Professorship of my university, and the Museum of Vertebrate Telos is grate fully acknowledged. 

“De vu nt of Integrative Biology and Museum of Vertebrate Zoology, University of California. Berkeley. California 
947 20-3160, U.S.A. wakelab@uc link, berkeley 

| dedicate this paper to R. C. Stebbins, ie remains an enthusiast for Ensatina, on the occasion of his 91st birthday, and to 

the memory of Arden H. Brame Jr. H (1934-2004), who first introduced me to Batrachoseps and its diversity. 


ANN. Missouri Bor. GARD. 93: 8-23. PUBLISHED ON 31 May 2006. 


Volume 93, Number 1 


Species Formation in Salamanders 


methods in comprehensive geographic sampling: soon 
he levels of differentiation. that were 
»reviouslv unimaginable (e.g. Highton, 1989). Before 
I ) 8 8 8 


the main impact of the new molecular systematics 


uncovered 


— 


— 
— 


revolution, the first modern complete catalog of 
amphibians reported 356 salamander species (Frost, 
985). The pow of salamanders has increased 
since 5696 (the 

AmphibiaW i 2006). These are surprising increases 
terrestrial 


most recent count is 550, 


of relatively well-known 


— 
A 


or a group 
vertebrates, especially given that many of the new 
reflect new discoveries. However, the new 


descriptions also highlight a modern species problem 


species 


that is not unique to salamanders, but common to most 


taxa—how to translate molecular findings into 
taxonomy. 

At the same time as new laboratory methodologies 
were drawing attention to species problems, new 


phylogenetic methods were having a major impact. In 


particular, methods ranging from mainly phenetic 
numerical taxonomy to cladistics were causing 


systematists to rethink species concepts and criteria 
(e.g., Cracraft, 1989, 1997; Mayden, 1997). 


Me ue became ever more sophisticated, renewed 


As genetic 
result (e.g.. Templeton, 
2001). 


publication on 


focus on species was one 
1989; Coyne & Orr, 2004; Hey. 


Given the sheer volume of recent 


species concepts and on the species problem ii 
general, it is a bit surprising that de Queiroz (1998, 
2000, 2005) concluded that no fundamental change in 
our species concept has taken place and that such 


a change is unnecessary. | agree with this perspective. 


here are indeed many ongoing debates about 
species, but these are fundamentally about criteria 


and the debate is mainly joined by proponents of 
different kinds of data or more explicit methodologies 
than used in the past. 

Here I examine patterns of species formation in 
display 


Californian salamanders that 


contrasting 
patterns. My goal is to highlight the connection of 


pattern and process in the formation of species. Lam 


especially interested in comparisons of tree-based and 


character-based methods of delimiting species and 
what it is that we have delimited using these methods. 


In addition, I explore whether species recognition and 
delimitation is an exercise in recovering history or in 
predicting the likely course of further evolution. 

I accept the general lineage concept of species that 
has been developed by Mayden (1997) and de Queiroz 
(1998, 2005). f 


metapopulations, or, more precisely, they are seg- 


© 


Species are historical lineages « 


ments of metapopulation lineages; all other considera- 


tions are secondary to this primary principle. In any 


given instance, one might rely on any of a number of 


secondary species concepts as criteria to delineate 


what species are recognized. The critically important 
point with respect to the general concept is that there 
are no necessary attributes of species; rather, there are 
different lines of evidence that can be used concern- 
ing separation of lineages and these are all contingent. 
The pragmatic matter of delimiting species remains, 
and 


however, many methods have been proposed to 


deal with the issue. All are problematic (reviewed by 
Sites & Marshall, 2004). Furthermore, even adopting 
the concept framed by de Queiroz does not mean that 
different taxonomists will reach the same conclusions 
because criteria can be interpreted in different ways. I 
suspect that de Queiroz (based on de Queiroz, 2005) 
would recognize more species than I would in most 
instances, because of his emphasis on detection of 


initial splits. 
PATTERN AND PROCESS IN SPECIES FORMATION 


Classifications of modes of species formation may 
be pattern-based or process-based. By pattern I refer 
in particular to geographic units inferred to be 
spe- 
cies. Criteria vary greatly, from some relative measure 
of degree of divergence to tree-based approaches. The 


simplest way to explain such patterns is vicariance 


independently evolving lineage segments, 1.e., 


plus something else, such as differential selection or 
even haphazard changes in gene frequencies. By 
process | refer specifically to adaptive processes that 


the establishment of 


are critically important to 
independent lineages. Although an enormous amount 
of literature deals with species formation, I focus here 


on adaptive divergence and nonadaptive radiation. 
These are alternative ways that lineages diversify. The 
main problem that has been identified with adaptive 
divergence is the establishment of genetic isolation 
between the adaptively diverging populations when 


are not allopatric. The main problem with 


hey 
nonadaptive radiation is the historic one of de- 
termining the status of geographically isolated but 
phylogenetically related parts of lineages, the classi- 
cal problem with allopatry. 
ADAPTIVE DIVERGENCE 

Papers by Bush (e.g.. Bush, 1969) and a book by 
Endler (1977) are the starting points for my own 


— 


interest in adaptive divergence (as opposed to the 
historical allopatry arguments) as a factor in species 
formation. Bush emphasized adaptive divergence in 
sympatry. Endler was more interested in geographic 
variation, in particular clines in characters or gene 
frequencies, and appearance of steps in clines, which 
might be preludes to species formation; his emphasis 


was parapatry or alloparapatry. 


10 Annals of the 
Missouri Botanical Garden 


My studies of adaptive divergence have focused on. picta Wood, 1940, E. e. platensis. E. e. xanthoptica 
the plethodontid salamander Ensatina Gray, 1850, Stebbins. 1949: Fig. I). He envisioned an ancestor 
which displays geographical differentiation interpreted having the traits of picta, restricted now to the ancient 
as clinal differentiation and adaptive divergence on the forests of southwestern Oregon and extreme north- 
way to species formation (Stebbins, 1949); it remains western California. This form has a somewhat varie- 
problematic whether subsets of the lineage have gated color pattern, and in it Stebbins envisioned the 
achieved independence and how many species warrant potential of all of the patterns found elsewhere. Each 
recognition. Ensatina is a ring species complex, subspecies was diagnosed on the basis of its color 
a taxonomically challenging group of populations pattern and some subtle morphological differences. 
spread from British Columbia to Baja California. although oregonensis was something of a default and 
generally west of the Cascade-Sierra Nevada mountain had highly variable color patterns. From a picta-like 
system. These are terrestrial salamanders that never ancestor Stebbins postulated. southward migrations 
enter water, and they never gather in assemblages. accompanied by divergent adaptation. In the Sierra 
They display direct development (lecitotrophy) in Nevada the forms became restricted to montane areas, 


which embryos form inside the egg capsules and hatch usually with closed canopy forests. and these were 


as miniatures of the adult. Ensatina is relatively long- thought to show progressive stages in the development 
lived (ca. 8 years), sedentary, philopat tric, and shows of disruptive coloration, from intergrades in the far 
little dispersal (Stebbins, 1954: Staub et al.. 1995). north to spotted red and brown platensis. to more vivid 
Prior to Stebbins (1949) four es were recognized. — yellow and black croceater in the Tehachapi Moun- 


and although some originally had been described as tains, finally culminating in the boldly blotched 


members of the genus Plethodon, Ensatina long has — klauberi in the far south (Fig. I). Along the coast 


been recognized as monophyletic (e.g, Dunn, 1926). — the generalized oregonensis is variable within and 
Three of the species, Ensatina croceater (Cope. 1868). — among populations in coloration. displaying generally 
E. klaubert Dunn, 1929, and E. platensis (Jiménez de — cryptic coloration, but not disruptive in pattern. As 
la Espada, 1875), were boldly marked (spotted or populations became increasingly associated with more 


blotched) salamanders assoc dated with forested regions — open habitats, a mimicry relationship developed with 
from inner montane areas from Mount Lassen, the dangerously poisonous newts (Taricha Gray, 
California, southward to San Diego County and 1850). Mimicry is most highly developed in xanthop- 


northern Baja California. A more widespread Species. tica, mainly found in the inner coast ranges east of San 


E. eschscholtzii Gray, 1850, was more uni ormly Francisco Bay (Kuchta, 2005). These aposematically 


colored, lacking spots or blotches. and was distributed colored animals display vivid. vellow and orange 


along the coast of California and inland as well, to the — coloration that contrasts sharply with the disruptive 
Cascade Mountains in Oregon, Washington, and British patterns of the blotched inland forms. Adaptive 
Columbia. This form is associated with woodlands in divergence was accelerated as a result of these 
the north, but in central and southern California it alternative adaptations, each different from the 


frequently occurs in more open habitats such as coastal ancestral condition. When the contrasting eschscholtzii 


sage scrub and chaparral. and Alaubert contact each other at the southern extent 
The conclusions of Stebbins's (1949) groundbreak- of their ranges, they look and act like different species 


ing study were revolutionary. He postulated that the (Fig. 1). although they often hybridize. Critical to 
four, then-recognized, species were part of a single — Stebbins’s interpretation was a “transvalley leak.” 
lineage that had originated in the north and had a relatively recent invasion of the coastal xanthoptica 
expanded its range southward, adaptively diverging into the foothills of the Sierra Nevada. where it came 


in different. directions in the coastal and inland into contact with platensis and the two hybridized. The 


mountains. Stebbins recognized seven subspecies ring was formed by the intergrading populations of 
(Ensatina o zii eschscholtzii, E. e. croceater, picta, oregonensis, and platensis at the northern end of 
E. e. klauberi, E. e. oregonensis (Girard. 1856), E. e. the Sacramento Valley. 


Figure l. Ensatina that are found in California (all except the first mentioned loci in that state). Diverse color 
kw | E 


patterns of the subspecies of Ensatina. Clockwise from upper left: Ensatina eschscholtzii oregone nsis from Kittilas Co., 
: „ (photo a D. B. M n 7. "a oregonensis from Sonoma Co. (C. W. Brown): inte 1 Pus ‘tween H. e. oregonensis 9 
atensis fron e Co. . Brown): oru) E. e. platensis from Calaveras Co. (C. W. bow. southern £ 
pes from T e Go. (D. B. N ke) : E. e. croceater from Kern Co. (C. W. Brown): E. e. me eri s: San Diego Co. (C. 
Brown): H. e. csg 5 zii from San Die go C o. (C. W. Brown): at bottom, a presumptive F hybrid n E. e. klauberi and E 
eolit E San Diego Co. de W. Brown): a hybrid bac 10 ss between E. e. xantho 7 m d E. eur from 
Calaveras Co. . Brown); K. e. a from Sonoma Co. (C. W. Brown); Z. e. le ae C PNE ras Co. (C. W. 


Brown): Z. e. ns fram Del Norte c o. (C. W. Brown). 


Volume 93, Number 1 Wake 11 
2006 Species Formation in Salamanders 


oregonensis- platensis 
"inte grade" 


| platensis Nor 


a EA 
ne TN ines N rau. M 
-a a Ts * AI 
X e^ ed: 3 
, de. ^ 4 
E Ši 4 
d “ 
d ^ r, 
EI uw a i= 2» 
2 * o 
A oe y * 
qu -- 
/ 
N 3$ fv 


noo 


Ensatina — | 
in California 


| klauberi x eschscholtzii 
ybri 


Annals of the 
Missouri Botanical Garden 


Subsequent research lent some support to the 
hypothesis of Stebbins, but added complexity. Al- 
though some hybridization occurs in southern Cali- 
fornia, there is much more in the Sierra Nevada 
(Brown, 1974). At the southern-most point of contact 
in San Diego County, no current or past hybridization 
detected (Wake et al., 1986) Many 


evolutionists picked up on this example as one 


has been 


py 


illustrating stages in a gradual process of species 


formation driven by adaptive divergence. Ensatina 
(1958) 


extended the hypothesis by envisioning gene flow 


became a textbook example. Dobzhansky 
throughout the ring retarding species formation and 
leading to the persistence hybridization in the 
1. Stebbins (1949) 


center of the ring and in the sout 


had not discussed gene flow: 


a continuous process but one in which there was 


fragmentation, isolation; and divergence, and then 

recontact. In other words, 

i vicariance and adaptive divergence. It was his 
| 

iphical 


geog scenario that led him to consider it 
a ring species (Stebbins, pers. comm.). 


When genetic methods became available I initiated 
studies of Ensatina, expecting it to be a complex of 
several to many species. Indeed, we found extensive 


genetic differentiation in allozymes (Wake & Yanev 


1986). but thought our sampling density too coarse to 


detect species borders (for a contrary view, see 


1998). 


south end of the distribution than in the region across 


Highton, Differentiation was greater at the 


the upper end of Sacramento Valley, and xanthoptica 
and platensis were less differentiated from each other 


than were eschscholtzii and klauberi. as would be 


predicted by the Stebbins model. Many of the genetic 
distances between adjacent samples were relatively 
high. and we suspected that finer. sampling was 
necessary to find either species borders or the gradual 
genetic transitions as expected under an isolation by 
distance scenario. We could readily reject Dobzhan- 
sky's hypothesis of ongoing gene flow throughout the 
complex. In general, the coastal forms clustered. with 
blotehed and 


the widespread oregonensis, and the 


= 


spotted forms in the south clustered together i 
phenetic analyses of the allozyme data. However, the 
blotched platensis was highly differentiated. with 
northern populations clustering with the unblotched 
coastal forms and the single southern population 
analyzed clustering with croceater and klauberi. 

In order to provide further resolution we expanded 
research to many populations extending from picta in 
northwestern California to Alaubert in the south (Jack- 
man & Wake, 1994), 


three clusters, 


The populations resolved into 


general within which patterns of 


isolation by distance were detected (Jackman & Wake. 


1994). A northern. cluster included what Stebbins 


he was thinking not of 


his view had elements of 


(1949) had labeled picta. picta/oregonensis intergrades, 
oregonensis, and oregonensis/platensis intergrades. A 
cluster 


second included only platensis from the 


northern and central Sierra Nevada, and the final 
cluster included platensis from the southern Sierra 
Nevada as well as croceater and klauberi. Northern 


platensis was closer to populations in the northern 
cluster than to southern platensis, but a population from 
Wagner i National Park, was 


intermediate between the two groups. 


Ridge, west of Yosemite 
We investigated these patterns in more detail using 
(Moritz et al., 


examining distributions of 


the mitochondrial cytochrome b gene 
1992). 


haplotype clades, sharp borders between groups of 


As expected, by 


populations were found, notably between northern and 


southern platensis. We also found a PD break 


between northern platensis and oregonensis/platensis 


intergrades. The scale still seemed too coarse. and 


subsequently we expanded the sampling to saturate 
» o 


E 


the California range of the genus, including nearly 
400 samples (Kuchta et al., in prep.). Based on several 
different analyses we found it convenient to recognize 
LI clusters of haplotypes (Fig. 2) based on patterns of 
phylogenetic relationships and geographic: distribu- 
tion. Most of these are well supported statistically, but 
some (e.g.. 


(e.g.. E) 


B) appear to be paraphyletic and others 
include members of two currently recognized 
subspecies. The northern (postulated ancestral forms) 
picta and oregonensis comprise six of these groups. 
Again, northern and southern platensis are. differen- 
tated, with southern populations clustering tightly 
with croceater (Fig. 2); there are three clades in this 
cluster, each with strong support. 

More detailed study of the Sierra Nevada contact 
zone focused on the hybrid zone between platensis and 
1989). 
fixed or 


xanthoptica (Wake et al., This transition is 


sharp, involving 8 nearly fixed allozymic 
markers that change in only a few hundred meters. 
The hybrid zones have been stable for about 40 years 
(first Brown, 1974). 
narrowed over a 20-year period (Alexandrino et al.. 
2005). 


lowland xanthoptica and the cryptic, upland platensis 


recorded by and may have 


The distinction between the brightly colored, 


occurs in an ecotone between lower elevation open 
oak-pine forest and chaparral and higher elevation 
closed canopy mixed conifer forest. Few F; hybrids 
are found, but there are many backcrosses. Because 
parental types are syntopic, they would be considered 
distinct species by most taxonomists were it not for the 
pattern in the remainder of the complex. The 
distribution of xanthoptica extends for about 100 km 
in the foothills of the Sierra Nevada. These popula- 
lions are little differentiated from coastal populations 
of xanthoptica in allozymes and mtDNA, as well 


coloration. Hybridization has only been studied in 


Volume 93, Number 1 
2006 


Wake 13 
Species Formation in Salamanders 


200 km 
ee sure a ds in California. Left. 
gram of Bayesian analysis of 
postear probability of su Poor at levels from 95%-—10( 
clades, 
labeled and keyed to the map. 


2005). but 


hybrid zones are also known to occur between southern 


detail in the north (Alexandrino et al.. 


platensis and xanthoptica. In contrast, where northern 
and southern platensis interact there is no evident 
difference in coloration. or habitat preference. The 
genetic distance (Nei. 1972) rei “pure” northern 
50.30, with 
nm loci. The 


transition as determined from allozvmes takes place 


and southern platensis is 0.2 major 


a 


frequency differences in genetic 
over about 300 km, but a major transition in haplotypes 
takes place near the southern end of this region, about 
75 km south of the major zone of allozymic transition 
(Wake & Schneider, 


color pattern is found in Yosemite Valley. This is where 


1998). Extraordinary variation in 


This is a preliminary alii 


glaciation in the region of present-day Mt. Lassen, i 


xanthoptica | 


xanthoptica H 


eschscholizii J 


oregonensis A 

picta B 

picta B 

oregonensis D 

platensis Q «NORTH 

oregonensis F 

platensis, E Sour» 

klauberi K 

oregonensis C 
— 


Outgroups 


Distribution of e clades identified using the mitochondrial DNA ¿ 


uences for ru 40 m of Ensatina from 


but they consist of relatively closely allied 91 5 Jes that are or la MICH The m groups are 


the allozyme transition first is detected, and I believe 
that the variation may be a consequence of the merger 
of two genetically differentiated populations. 

Jackman and Wake (1994) presented a possible 
scenario to account for the historical biogeography of 
Ensatina in the Sierra Nevada. They envisioned an 
early southward migration and differentiation. of 
spotted and blotched salamanders, then geographic 
isolation as a result of a gap in the distribution that 
developed midway along the Sierra Nevada. Sub- 
sequently the northern Sierra Nevada was occupied by 
precursors of northern platensis, which in turn became 


isolated from oregonensis by repeated volcanism and 


1 


Annals of the 
Missouri Botanical Garden 


northeastern California. Northern and southern pla- 


lensis. remained separated, probably by glaciation 
concentrated in the deep valleys of the central Sierra 


Nevada associated with the present-day Tuolomne, 


Merced, and San Joaquin river canyons, until late 
Pleistocene, when the two groups met and merged 
genetically. A selective sweep is responsible for rapid 
northward movement of the strongly adaptive blotched 
color pattern of southern platensis (Wake & Schnei- 
1998) 
the northern end of the range of the northern platensts 


the Mt. 


populations in 


der, . This pattern is postulated to have spread to 


haplotype group. in Lassen area. Stebbins 
(1949) that 


California were intergrades between oregonensis and 


thought northeastern 
northern platensis. An alternative interpretation is that 
strongly marked (but not blotched) salamanders in this 
that differ from northern platensis in both 
allozymes and mtDNA 


southern platensis color alleles have only recently 


area 


might be an indication that 


where 


reached this area (e.g. Fig. I). they are 
introgressing into oregonensis. 
The combination of allozymic and haplotype in- 


formation led Highton (1998) to conclude that there is 
a broad hybrid zone separating northern and southern 
which he considered to be separate (but 
He 


However, in my 


platensis, 


unnamed) species. was silent with respect to 


coloration. view, the zone of in- 


eraction is far too broad (at least 75 km) to be 


considered a hybrid zone. | consider a hybrid zone to 
be a site either where two different parental forms co- 
occur and form hybrid individuals, or where two 


different parental forms are separated by a distance 
equivalent to a few. on the order of tens. of home 
ranges, with the intervening area oceupied by hybrids 
and perhaps several generational backerosses. | have 
1997) that any of several 


problematic, 


argued elsewhere (Wake. 
taxonomic change are 


taxonomy of Stebbins (1949) 


suggestions for 
and have retained the 1 
until compelling evidence of species borders around 


(as contrasted with across) the ring-like distribution of 


Ensatina is found. When morphologically and eco- 


logically similar forms meet they merge genetically, 


across haplotype clade borders, which remain as 


markers of past distributional limits. In contrast, when 


morphologically and ecologically differentiated forms 


meel they hybridize, in ecotones in which both 


parental forms are present. No indication of post- 


mating isolation is found. The hybrids and back- 
crosses are hypothesized to be at an adaptive 


disadvantage t 


either parental class, and we have 
measured extraordinary levels of against 
40%-75%) (Alexandrino et al., 2005). The 


classic. explanation by Stebbins (1949) that intergra- 


selection 


them (ca. 


dation occurs in the north, hybridization in the region 


of the transvalley leak in the middle, and sympatry in 


the south 1s correct, but the intergradation in the north 
may be secondary rather than primary, and sympatry 
i one of four 


with no hybridization is found in only 


hybrid zones in the south. While sympatry with no 
hybridization is found at one site in the extreme south. 
hybridization is documented for three other Alauberi 


1986). 


The situation on the coast is far from simple. There 


isolates (Wake et al., 


are two separate haplotype groups associated with 
distinctive allozymic-based clusters of populations 
within xanthoptica, one on the southern San Francisco 
Peninsula and the other widely distributed to the north 
San Bay. the 


are complex interactions 


including ii 


[e] 


and east of Francisco 


Sierran foothills. There 
where oregonensis (with two haplotype groups in the 
San 


F urthermore, 


region) meets xanthoptica north and south of 
Wake, 1997). 


is more differentiated (with respect to 


Francisco Bay (Fig. 2; 
eschscholtzii 
haplotypes) than would have been anticipated from 
the early allozyme study, with distinctive northern and 
southern geographic segments that form only a weakly 
supported possible clade. 

These considerations led me lo propose an 
historical biogeographic hypothesis for the complex, 


approximately 5 million vears before present, which 


postulates a widely distributed oregonensis/platensis 


precursor in the north, an isolated precursor o 
northern. platensis in. the northern Sierran region, 
and a precursor of southern platensis/croceater/klau- 


The 


Central Valley of California was at this time an inland 


beri in the southern Sierran region (Wake, 1997). 


extension of the Pacific Ocean, and the precursors of 
xanthopticaleschscholtzii may have originated on an 


archipelago (ef. Batrachoseps scenario. below). How 


they got to the archipelago is uncertain, but because 
the 
in the south, Parks (2000) postulated. that an 


movement of oregonensis-like salamanders southward 


and associated with the Salinian Block originated 


early 


gave rise both to northern platensis and to the 
precursors. of xanthopticaleschscholtzit. Subsequent 


orogenic as well as plate movements led to the 
assembly of populations of Ensatina into the current 
form of a ring. Whereas klauberí has usually been 
BON as well-nested within the complex, this 


the 


alternative view (shown in Fig. 2, although 
branching near the base of the complex is not well 
supported) is that xanthoptica + eschscholtzii, which 
form a distinctive phylogeographic unit, might form 
a clade that is sister to everything else, even including 
the diverse populational groups identified as orego- 
nensis or picta. Some of Stebbins’s subspecies are 
historical units that can be diagnosed by morphology 


(klaubert, 


nonmonophyletic 


and molecular characters xanthoptica, 


eschscholtzii); others are either 


amalgams (platensis) or incompletely differentiated 


Volume 93, Number 1 
2006 


Wake 15 
Species Formation in Salamanders 


but adaptively diverged groups (southern platensis- 
croceater, picta-oregonensis). Finally, oregonensis is an 
undiagnosable grouping of differentiated lineages that 
branched independently (Fig. 2). 

illustrates 


Ensatina a complicated relationship 


between adaptive divergence (as exemplified by the 


E. 


evolution of aposematic coloration in association with 


mimicry in the coastal populations, and cryptic 


and 


— 


coloration in the inner montane populations 


vicariant events (for example. the differentiation of 


northern and southern platensis). While our under- 
standing of the complex has changed greatly since the 
work of Stebbins (1949), 


a case in which adaptive divergence has been 


the basic point that this is 


dominant still holds. This understanding does not 
help us to establish a better taxonomy, and with the 
present data, no alternative seems more appropriate 
than maintaining what we now have. a polvtypic 
species. Establishing taxonomic species within the 
complex creates new problems. For example. platensis 
is readily diagnosable on morphological grounds, but 
not with either DNA (diphyletic) or proteins, and 


croceater, also diagnosable on morphological grounds 


is nested within southern platensis in the DNA tree. 


conclude that this is a rare instance in which 


subspecies are helpful. They reflect what is in essence 
the shallowest (most recent) time depth. related to 
current adaptive antipredator strategies. M an in- 
termediate time level are the genetic interchanges 
associated with population-level mergers around the 
ring, revealed by allozymes. Finally, the deepest level 
distributions of the 


is reflected in the exclusive 


haplotype groups recognized in Figure 2, all of which 


have discrete geographic limits and may reflect. i 
part, allopatric episodes at different times in the past. 
NONADAPTIVE RADIATION 


The Slender 


have a geographic range in California that is simi 


Salamanders, genus Batrachoseps, 


ar 
to that of Ensatina (Fig. 3), but in other respects they 
contrast to Ensatina. 


are a sharp 


Batrachoseps occur in syntopy with Ensatina at most 


sites of range overlap, including a vast array of 
habitats ranging from temperate rain forest in 


northwestern California, to relatively high elevations 


a, tc 


(ca. 2000 m) in the forests of the Sierra Nevac 
chaparral and coastal sage scrub along the southern 
Like Ensatina, 


salamanders that lay direct-developing eggs on land. 


coast. these are strictly terrestrial 


They are sedentary, with home ranges that are even 
smaller than those of Ensatina; there is little evidence 
of dispersal. 

A detailed morphological analysis of Batrachoseps 


(Hendrickson, 


1954) recorded extensive geographic 


Species of 


variation and concluded that only a single species, B. 
attenuatus (Eschscholtz, 1833), was represented 

California. A related northern species, B. wrightorum 
1937), was 


Mountains of Oregon. Hendrickson's conclusions were 


found in the northern Cascade 


(Bishop, 
controversial because his two subspecies of 
attenuatus occurred in complete sympatry on Santa 


[22] 


— 


Cruz Island. off the coast of southern California. This 


unusual taxonomy resulted from his envisioning 
a complex biogeographic scenario in which the island 
was invaded by two somewhat differentiated. popula- 
tions, one from the north and the other from the south, 
establishing sympatry that formed from a ring-like 
pattern of differentiation (he was strongly influenced 


work of his 


Ensatina). 


y the adviser, R. C. Stebbins. on 


Subsequent research revealed many species bor- 
and currently 20 species of Batrachoseps are 
Only one, B. 


Analysis of allozymic and mtDNA data sets 


ders, 
recognized. wrightorum, is not found in 
California. 
identifies six major clades, all found in California. 
Species belonging to different clades often are so 
similar morphologically that they cannot be discrim- 


The three members 


inated without molecular an 


f the subgenus Plethopsis Bishop, 1937, the sister 


taxon of subgenus Batrachoseps, differ from the latter 
in some osteological traits, and while most species of 
subgenus Batrachoseps are more slender than those of 
Plethopsis, one (B. stebbinsi Brame & Murray, 1968) 
closely resembles species of Plethopsis. At the 


molecular level, however, B. stebbinst is sharply 


diverged from Plethopsis, differing both in allozymes 
(more than 10 fixed differences in conservatively 
evolving proteins) and in mtDNA sequences (Wake et 
al., 2002). 

Along the coast of RUE clades replace one 
The attenuatus 


another geographically (Figs. 3. 4). 


= 


clade in the north is replaced by the pacificus clade in 
the central coastal region, and it in turn is replaced by 
the nigriventris clade to the south, which finally is 
replaced by another member of the pacificus clade in 
the far south. Ecological transitions between the 
borders of the group are almost imperceptible, and the 
species represented are similar in morphology. in- 


cluding coloration, and ecology. In the central coastal 


region, the pacificus clade is represented by four 
morphologically similar species that once again 


replace one another from north to south. Furthermore, 
n southern California a geographical replacement 


pattern. is found within the pacificus clade, starting 
with one species on the northern Channel Islands and 


a second on the southern Channel Islands and 


adjacent mainland. The most southerly member of 


— 


the pacificus clade, B. major Camp, 1915, has severa 
clearly distinguished haplotype clades within it, and 


Annals of the 
Missouri Botanical Garden 


Batrachoseps complex 
in California 


Figure Distribution of the species 


comprising le central coastal cluster of the pacificus clade 1 by M. Carcia-Paris 
text: B. campi Marlow, Brode & Wak 
1998, B. regius Jockusch, Wake & unen 


where not provided in the 


Jockusch, Wake & Yanev, 


again they replace one another from north to south 
(Wake € Jockusch, 2000), this pattern apparently 
continuing into Baja California (unpublished). 

These patterns raise questions concerning species 
delimitation and circumscription. Because Batracho- 
seps attenuatus (attenuatus clade) occurs in sympatry 
with B. gavilanensis Jockusch, Yanev & Wake, 2001 


( pacificus clade) without evidence of interbreeding and 


with great genetic divergence between them (D as 
defined in Nei, 1972, 


even though they cannot be distinguished in the field. 


ca. 1.6). two species are present, 


Other instances of sympatry led Brame and Murray 
(1968) to undertake a progressive revisionary study that 
Brame and 


was a prelude to subsequent research. 


Murray recognized two species in southern California 


— 
w 


(B. major, on the mainland and southern Pacific 


Islands, and B. pacificus Cope, 1865, on the northern 
which separated from 6. 


Channel Islands), they 


attenuatus on morphological grounds. They also de- 
scribed three species from the southern Sierra Nevada 
B. stebbinsi, B. simatus Brame & Murray, 1968, and B. 


1968). The curious distribu- 


relictus Brame & Murray, 


Subgenus Plethopsis 
E B. campi 
SB. robustus 
Subgenus Batrachoseps 


pires bus relictus Group 
Wi B. atten E 


nigriventris Group " B. edu 
B. diabolicus 


6. gregarius pacificus ae 
E stebbinsi B B luc 

simati 
ll Upper Kern taxa — E B. incognitus 


is B gavilanensis 
B. pacificus 
3. major 

B. andus 


gabrieli Group 
E B. gabrieli 


of Batrachoseps in California. The photographs are of spec imens of the four species 
& 


'ake). Authorities for species 


Yanev & Hansen. 2002, B. kawia 


. 1979, i RS Wake. 


tion of B. relictus. (Sierra Nevada, central Coastal 


California, Santa Cruz Island, and mountains of Baja 
California) stimulated Yanev (1978, 1980) to conduet 


an extensive study of allozymes. She found unexpect- 


— 


large divergences and high levels of variation 
found that B. 


edly 
within the taxa she recognized. 
restricted to the region 


Yanev 
attenuatus was north of 
In the central coastal region, popula- 
» D. 
attenuatus were assigned to B. pacificus (an unnamed 


now B. 


parts of 


Monterey Bay. 
tions that Brame and Murray considered to 


semispecies, gavilanensis). Populations from 


more southerly the central. coastal. region 
identified as B. relictus by Brame and Murray were 


considered by Yanev to constitute another unnamed 


semispecies of B. pacificus (now three species: B. luciae 


Jockusch, Yanev & Wake, 2001. B. incognitus 
Jockusch, Yanev & Wake, 2001. and B. minor 
Jockusch, Yanev & Wake, 2001). Brame and Murray's 
D. major and B. pacificus were also considered 


semispecies by Yanev and treated taxonomically as 


B. pacificus. Yanev had narrower 


subspecies of 
perspective on B. relictus (restricted by Yanev to the 


Volume 93, Number 1 
2006 


Wake 


Species Formation in Salamanders 


30 mya 


mid-Oligocene 


12 Future 

Q San Andreas 
` , Fault 
So 


à 3 North 
N American 
2v ate 


other 

clades 

Pacific 
Plate 


Precursor of 
pacificus clade 


mya 
g late Miocene 
N 


` 
N gavilanensis 


AN 
SN 0 luciae 
^ vd 7 
s S ifi 
oF pack —— 
NN Transverse / 
iu Range 
A 


faults 


pacificus 


my 
early Pleistocene 
gavilanensis 
luciae 
pacificus 
so. major 
incognitus 


minor 


SS 


18 mya 
‘uture . g 
San Francisco mid-Miocene 
Bay 


mM 
: 


\ 


x o 
ub w- 
CMS ON 
"MT N Future 
À ~ V Sur Nacimiento 
` p ult 


,Fa 


o NE 
— i 


other 
clades 


gavilanensis 


' m 

` early Pliocene : 

' y avilanensis 

aN luciae 

SN . 

1 On, future . 

usc Range 5 ; 
. o major 


Y. incognitus 


major 


12 mya 


mid-Miocene 


ix. 


luciae 5 


luciae 


2 mya l . 
i . gavilanensis 
mud-Pliocene 


V luciae 


2" pt s 


N major 
y 
` N o incognitus 
— minor | 
`A | minor. 
A " 


incognitus v major oy ` 
; a 
southern a p e 
d yas Present Day 
S Distribution of 
se aci 
nass ; Batrachoseps pacificus 
cluster clade 
gavilanensis NL 
incognitus N 
minor. b 
N 
0 \ 
southern 
cluster 


— 


eei 


Historical biogeography of the Batrachoseps pacificus clade in southern and central California. This scenario is 


sed on 5 al reconstructions by Hall (2002) and the phylogenetic hypotheses of Jock usch et al. (2002) and Jockusch 
2 02). 


and Wake (2( 


gavilanensis 


Annals of the 
Missouri Botanical Garden 


southern Sierra Nevada), considered to be a semispecies 
within the pacificus complex, and she reduced it, too, to 
subspecific status. The populations from the mountains 
ol Baja California were considered to represent a final 
unnamed semispecies of B. pacificus. Populations on 
Santa Cruz Island, assigned to B. relictus by Brame and 
Murray, were assigned by Yanev to a resurrected 
species, B. nigriventris Cope, 1869. In many ways 
Yanev's recognition of B. nigriventris was her most 
important contribution. This widespread species oc- 
curred in sympatry with other species of Batrachoseps 
in central and southern coastal California, on Santa 
Cruz Island, and in the southern Sierra Nevada. It 
turned out to be the key in unraveling the complex 


history of the genus, which is more complicated than 


even Brame and Murray had conceived. Of special 


interest was the complex pattern of range overlap with 
no sign of hybridization. or 


genetic leakage of 


morphologically cryptic species in the Inner Coast 


Range in central California, where B. attenuatus and 


the current B. gavilanensis were narrowly sympatric in 
the north, and B. gavilanensis and B. nigriventris were 


narrowly sympatric in the south. In coastal central 


En 


California the geographic range of B. attenuatus 
overlaps that of B. gavilanensis, whose range abuts 
that of B. luciae. Parapatry, but as yet with no range 
overlap, occurs further south between B. luciae and B. 
incognitus, and B. incognitus and B. minor (Figs. 3, 4). 
However the range of B. nigriventris overlaps that of 
both B. incognitus and B. minor, with local sympatry, 
and the northwestern-most range limit of B. nigrirentris 
is within a few km of the southern-most range of B. 
luciae. 

The addition of mitochondrial gene sequences and 
the reinterpretation of unpublished data gathered by 
Yanev (1978), 
field, led to further taxonomic revision (e.g., Marlow et 
al., 1979; Wake, 1996; Jockusch et al., 1998, 2001; 
Wake et al., 2003), summarized by Jockusch and 
Wake (2002). The distribution of species and the 


inferred. pattern of species formation is what Gitten- 


as well as recent discoveries in the 


berger (1991) termed nonadaptive radiation, phyloge- 
netic diversification not accompanied by adaptation 
into distinctively different ecological niches (see also 
Wiens, 2004, who refers to ecological niche conser- 
The 


parapatric species, with the parapatric forms isolated 


vatism). result is a group of allopatric or 


geographically because of inferred competitive inter- 
actions at borders of species ranges. The levels of 
genetic divergence are sufficient to preclude in- 
terbreeding or hybridization, and the ecological 
similarities are sufficient to maintain largely exclusive 
distributions. 

A general feature of the nonadaptive radiation in 


Batrachoseps is the relatively limited sympatry. The 


only sympatry between two members of any one of the 
six major clades is between B. nigriventris and B. 
stebbinsi, which co-occur in the Tehachapi Mountains 
at the southern end of the Central Valley. These 
species display morphological and ecological di- 
vergence, unusual among close relatives in this genus. 
The smaller and more slender B. nigriventris is 
a habitat generalist, found in woodlands under small 
to large cover objects and using retreats in the soil. 
The substantially larger and more robust B. stebbinsi 
is most frequently found in rocky soil and small talus 
slopes. These two species, which are not sister taxa, 
may have undergone character displacement, and if so 
it is unique in the genus. 

There are some important exceptions to the general 
rule in Batrachoseps that species either replace one 
another without hybridizing or are sympatric. In 
southern California, where population density is low, 


e 


populations of “northern” and “southern” B. major 
that are morphologically identical have mtDNA (cyt b) 
genes that are about 9% diverged and are not sister 
(Wake & Jockusch, 2000). In fact, mtDNA of southern 


B. major is more closely related to B. pacificus, a more 


robust species that is morphologically distinct, than t 
northern B. major (Fig. 4). However, northern and 
southern populations are only slightly differentiated in 


roteins, and Wake and Jockusch argue that when 


they meet secondarily and interact genetically, alleles 


rom northern populations spread over southern 


populations, but the interactions have been insuffi- 


IE 


cient as yet to dislodge the maternally inherited 
This 


reminiscent of Ensatina e. platensis in t 


southern mitochondrial situation is 


genes. 


1e central 
and northern Sierra Nevada (see above). 

The apparently nonadaptive radiation of Batracho- 
seps results from the complicated. geological history 
of California (Yanev, 1980; Jockusch et al., 2001). 


The first split, an ancient one, was between mem- 


— 


bers of subgenus Plethopsis and subgenus Batracho- 
Plethopsis is 


seps. Today located peripheral to 
Batrachoseps, to the north and east of the main range. 
Within subgenus Batrachoseps the first split segregat- 
ed B. attenuatus from everything else. The attenuatus 


clade, with a single but differentiated species, is 


= 


distributed mainly northeast of the San Andreas Fault 
zone and north of the historically significant embay- 
ment of the Central Valley in the vicinity of present- 
day Monterey Bay. The relictus clade, perhaps the 
next to branch phylogenetically, is restricted to the 
Sierra Nevada and western adjacent lowlands. H 
displays a distinet pattern of regional displacement, 
with substantially more divergence than within 
the attenuatus clade. Four species are recognized, 
each occupying an exclusive geographic range. The 


northern species, B. diabolicus Jockusch, Wake & 


Volume 93, Number 1 
2006 


Species Formation in Salamanders 


Yanev, 1998, split from the remaining species 


n 
the vicinity of the mid-Sierran glacial region, a region 
in which there are north-south splits in diverse taxa. 
The nigriventris, and 
pacificus, are largely southern, with gabrieli isolated 


in the southern transverse ranges, 


remaining clades, gabrieli, 
nigriventris con- 
centrated in the southern Sierra Nevada, 
Mountains, 


Tehachapi 
and some more southerly and westerly 
extensions, and pacificus located in coastal regions, 
extending from northern Baja California to just north 
of Monterey Bay, 


Fault zone. 


south and west of the San Andreas 

The only clade that crosses the San Andreas Fault 
zone to any degree of significance is nigriventris, and 
in particular Batrachoseps nigriventris, a species that 


has violate d the range exec lusi Iveness 50 ( haracte Vis tic 


of the genus to establish populations sympatric with 
members of the pacificus clade along the central and 
southern California coast and on the Channel Islands. 
This taxon. displays greater ecological breadth than 
any other member of the genus and likely represents 
an extensive and relatively recent. range. expansion 
from its origin, inferred to be near the southern end of 


the range of its sister taxon, B. gregarius Jockusch, 


Wake & Yanev, 1998, in the vicinity of the Kern 
River. where so much of the evolution of the 


nigriventris and relictus clades has been centered. 
The pacificus group has been more affected by the 
massive land movements that have taken place west 
and south of the San Andreas Fault zone than any of 
the other Batrachoseps. The scenario. developed by 
Jockusch et al. (2001). mtDNA 


phylogeny, envisioned an origin of the lineage 


based on their 
in 
southern California. A new geological reconstruction 
of coastal California (Hall, 2002) has enabled a further 
4). The first. split 
separated the pacificus clade from a nigriventris or 


development of the scenario (Fig. 


gabrieli ancestral sister taxon, with the pacificus clade 
associated with the Pacific geological plate in the 
paleoterrane known to geologists as Salinia. This may 
have occurred as long ago as about 30 million years. 
Since Oligocene, extensive land movements have 
transported pieces of the continental crust associated 
The 


first split within the pacificus clade was associated 


with Salinia more than 160 km to the northwest. 


with early stages of the northwestward movement and 
fragmentation of Salinia. the 
member of the pacificus clade. B. gavilanensis. 


the only species to have penetrated territory north of 


Today northernmost 


is 


the ancient Central Valley embayment, in the vicinity 
of present-day Monterey Bay. Its penetration. north- 
ward, probably a relatively recent event, has estab- 
B. attenuatus. As one 
old species, B. 
gavilanensis shows the greatest degree of protein 


lished a zone of sympatry with 


would expect for a relatively 


and mtDNA diversity within the northern cluster of 
extant species in the pacificus clade. 

The next species to split (Fig. 4) was present-day 
Batrachoseps luciae, which now occurs on a plate 
fragment (Jockusch et al., 2001). After the first two 
splits our phylogeny is less robust. What is clear is 
that each of the central coastal species is sister not to 
ils nearest geographic neighbor but to some complex 
of populations to the south. Thus, B. pacificus appears 
to have been the next to move north, and out to sea, on 
the precursor terrane that gave rise to the northern 
Channel Islands, becoming isolated from populations 
on the mainland. Progressively, B. incognitus and B. 
minor were transported northward, leaving a compli- 
cated B. major in the south that is made up of several 
distinctive phylogeographic units, essentially left over 
from episodes of isolation. The northern phylogeo- 
graphic unit within B. major either has or is in the 
process of extending its range south and east. where it 
appears to be merging genetically with populations 
characterized by non-sister mtDNA haplotype clades, 
morphology, or both. For example the problematic B. 
aridus Brame, 1970, variously recognized as a sub- 
species of B. major or as a full species, is closely 
related to a southern phylogeographic unit of B. major. 

The pacificus clade displays at a fine scale what | 
consider to be a long-sustained pattern of nonadaptive 
radiation in Batrachoseps. An ancient lineage frag- 
mented again and again, with a resulting patchwork- 
quilt pattern of lineage segments, 1.€.. that 


close parapatry with 


species, 
generally show “tight stitching.” 


no hybridization. Some restricted sympatry occurs 
between members of different major clades, which in 
the cases of B. nigriventris and B. gregarius is more 
widespread and involves more than one other taxon. 
Why the radiation in Batrachoseps 
resulted in a patchwork-quilt pattern of distribution 


nonadaptive 


requires closer examination of the history of the 
central coastal region. The four species of the 


pacificus clade in this region occupy closely abutting 


geographic ranges. and they similar in 


morphology and ecology (Figs. 3. 4). Two of these. 


B. gavilanensis and B. luciae, occupy relatively large 


are very 


areas and have sufficient samples to disclose 


significant geographic and genetic variation. These 


two species show extensive geographic variation in 
both allozymes and mtDNA haplotypes. In fact, the 


differences within each the level of 


between them (Jockusch et al., 2001). 


in the border zone where the ranges of the two 


approach 
difference 
Yet, 


species abut, no syntopy or genetic interchange is 


detected. Genetic distances across this border remain 


high. with some fixed allozvmic differences and 


complete segregation of mtDNA haplotypes. but no 
ecological differentiation is observed. 


Annals of the 
Missouri Botanical Garden 


The patchwork pattern, which is general in the 
genus, likely arises from several factors. First, these 
salamanders are extremely sedentary, with limited 
dispersal. Second, today's species represent lineage 
fragments of former species, which themselves were 
geographically differentiated. Thus species are “born” 
with geographic variation. Third, species, especially 
close relatives within a clade, replace one another 
geographically because they are so similar ecologi- 
cally that there is what may be called preemptive 
occupancy of space. That is, the space-holders have 
relatively large and well-established populations that 
resist invasion by others that are virtually identical 
ecologically. Wiens (2004) has discussed such niche 
conservatism and considers it to be a major factor in 
the formation of similar species under allopatric 
conditions. Areas where populations merge with 
others having non-sister haplotypes typically are 
ecologically marginal, with populations scattered 
and small. Under such circumstances, populations 
may have very low densities, and individual organisms 
that meet may have few mating opportunities other 
than with heterospecific individuals. 

Sympatry between species is found in two contrast- 
ing situations. The first is where there is some 
ecological divergence, often related to differentiation 
in the body sizes of the two species. Examples include 
Batrachoseps major with B. nigriventris, B. nigriventris 
with B. stebbinsi, and B. nigriventris with B. gabrieli 
Wake. 1996. The second is in areas where the ranges 
of species become discontinuous and where popula- 
tions are highly localized. This is the situation in the 
generally inhospitable Inner Coast Range, a largely 
treeless area with exceedingly hot and dry summers 
and litle winter rain. Here broad regional ranges 
overlap, but little or no local syntopy is found, as in 
the cases of B. attenuatus and B. gavilanensis and B. 
gavilanensis and B. nigriventris. This is the same 
ecological situation in which genetic mergers mas 
lake place. One can only speculate that the interactors 
that fail to merge have been separated. sufficiently 
long (as estimated from allozymie genetic distances 
and degree of haplotype divergence) that genetic 
isolating mechanisms have arisen. incidental to di- 


' 


verse genetic processes during the long period « 


geographic isolation. 
The main feature of a nonadaptive radiation, then. 
is the 


general impression of allopatry or. parapatry 


with a lack of ecological or any other kind. « 


divergence other than in molecular traits. In situations 


like this, where there is little or no hybridization, i 
may be fruitful to investigate the evolution of isolating 
mechanisms. which might have arisen incidentally in 
allopatry but serve now to cause individuals from 


adjacent populations not to recognize each other as 


potential mates. In the case of salamanders, this is 
Houck & 


Arnold, 2003). Postulated premating isolating mech- 
I o T 


PN 


most likely to involve mating pheromones ( 
anisms might be incidental, a byproduct of the 
different genetic histories of lineages since time of 
divergence. If isolating mechanisms did arise at the 
borders of two incipient species, it is very difficult to 
imagine a scenario. in which they would spread 
“backwards” throughout the far-flung populations of 
each species, which are so genetically heterogeneous. 
In such situations the evolution of isolating mecha- 
nisms seems unlikely to be related directly to the 


formation of species. 
GENERAL DISCUSSION 


The contrast between Batrachoseps and Ensatina is 
great. Where morphologically and ecologically similar 
populations of Ensatina meet they exchange genes, 
whereas in Batrachoseps they do not. Why does this 
stark contrast exist? The two genera belong to the 
same major clade (Plethodontidae). have similar life 
histories and ecologies, and species of the two are 
often sympatric. Both are fully terrestrial. However. 
Ensatina is a larger salamander, with better locomo- 
home 
Stebbins, 1954, and Staub et al., 1995, with Cunning- 
ham, 1960, and Hendrickson, 1954). Both genera are 


phylogenetically isolated, having no close relatives, 


lory ability and larger ranges (compare 


and likely represent very old lineages (perhaps on the 
order of 60 million years or more, based on estimates 
from rates of albumin evolution and degree of 
divergence in mtDNA and other data, e.g., Larson el 
al. 1981, 2003: Chippindale et al.. 2004: Mueller. 
2005: Mueller et al., 2004). Batrachoseps is the more 


^ 


internally diverged and. differentiated, and is likely 
the older. Its lower vagility, fidelity to pieces of the 
planetary crust. and inferred great age may be 
responsible for the fragmentation of the primordial 
lineage and what has been mainly a nonadaptive 
radiation. The processes involved in differentiation of 
Batrachoseps appear to be largely related to factors 
associated with earth history, especially crustal 


movements, as first proposed by Yanev (1980). In 


contrast, Ensatina appears to be responsive to 
challenges from its biological community and has 
evolved several different antipredator mechanisms, 
which have concomitant. behavioral and ecological 
The i 


incompletely fragmented and very widespread, meta- 


consequences. result is a differentiated, but 
population structure. 

Both examples have strong geographic components 
to their histories. Most species formation in terrestrial 
vertebrates appears to be geographic in nature, as first 
Mayr (1942). 


generalized by Geographie variation 
O | c 


Volume 93, Number 1 


Wake 21 
Species Formation in Salamanders 


within living terrestrial vertebrate species (as well 


many other laxa) natural populations is nearly 


„ Avise, 2000). In the taxa I study. there 
that 


ubiquitous (e.g 


so much within 


n` 


dim ne species only 
adaptively significant alleles or allele complexes are 
likely to spread broadly and rapidly. Accordingly. 
Ensatina has experienced spread of adaptively signif- 
icant trails, and this has kept the lineage from 


permanent fragmentation. However, in Batrachoseps, 
where no clearly adaptive traits beyond those general to 
most species have been identified, divergence and 
lineage fragmentation has proceeded. To focus atten- 
tion on the evolution of isolating mechanisms in 
situations like this is problematic. When long-separat- 
ed populations that have diverged significantly come 
into secondary contact they may be sufficiently distinct 
that they 
mates (e.g. Kozak, 2003). I suspect that this has 


no longer recognize each other as potential 


happened in Batrachoseps, where one never finds 


hybrids, but instances of haplotype paraphyly and 
polyphyly are found, sus eeesting that sec ‘ondary contact 


Funk & Omland. 


produced in 


was followed by genetic merger (cf 
2003). 


contacts may be adaptively inferior. 


Alternatively, offspring such 
This is the case 
in the adaptively different Ensatina in the central 
Sierran and southern California hybrid zones, which are 
One might 


only a few home-range diameters wide. 


study isolating mechanisms in these instances, but 
genetic isolation likely evolved incidentally during the 
allopatric divergences, not upon recontact. Perhaps 
hybrid zone or alloparapatric interactions (Alexandrino 
2005) are 
reinforcement of isolating mechanisms 
studied (Servedio & Noor, 2003). 


common theme, whether species formation has pro- 


et al., settings in which the potential 
might be 
Geography is the 
eressed adaptively or has been incidental to events 
related to isolating mechanisms. 

At what level of lineage segmentation do we recognize 
2005) pointed 


out that while the process of evolutionary divergence leads 


TD 


and name species? Recently de Queiroz 
to the acquisition of different. properties by diverging 
differ 


particular property they emphasize. He argues that the 


with which 


ineages, taxonomists respect to 
adoption of different thresholds is the main cause of the 
“species problem.” I have attempted to use the same 
thresholds for the two examples presented here, but have 
reached very different conclusions. De Queiroz argues that 
“all separately evolving metapopulation lineages would be 
species” regardless of particular attributes, but in the case 
of Ensatina, for example, the issues of separateness and 
metapopulation limits remain unclear. The likelihood that 
once separate lineage segments are now genetically 
merged, or have in the recent past undergone secondary 
genetic merger, is high. In contrast, in Batrachoseps the 


same issues are more generally resolved. 


Biologists will continue to have problems with 
species. We may all share the same conceptual 
but it 


framework. is safe to predict that arguments 


concerning thresholds and attributes will continue to 


roduce controversy. However, regardless of taxo- 
| ) 8 
careful analysis of patterns. and 


will 


nomic treatment, 


processes associated with species formation 


inform meaningful discussion and debate. 


Literature Cited 


Alexandrino, J.. S. J. E. Baird, L. Lawson, J. R. Macey, C. 
Moritz & D. B. 2005. 


hybrids at a hybrid zone in the 


Strong selection = 


Ensatina ring specie 

comple x and its evolutionary implications. Evolution 59: 
1334-1347. 

ino abs E 
servation [web application]. 2006. Berkeley, California: 
AmphibiaWeb. Available: <http://amphibiaweb.org/>. 

Avise, J. C. 2000. i 
Formation of Species. Harvard Univ. 


Information on amphibian biology and con- 


Phylogeography: The History and 
Press, Cambridge, 
H 


Brame, H. Jr. & . Wake. 1962. A new species 
salama e from a and the status of Geotriton 
d Posada Arango. Contr. Sci., Nat. Hist. Mus. Los 

ngeles Co. 49: 1-8. 
— & K. F. Murray. 1968. Three new slender 


discussion of relation- 
Mus. Nat. 


moms s (Batrachoseps) with a 
shi D and id iation within the genus. Sci. Bull., 
Hist., Los Angeles Co. 4: 1-35 

ios. C. W. 197. 24. "Hybridization among the subspecies of 
= ple e Si Exe Ensatina eschscholizii. Univ. 

Calif. Publ. Zool. 98: 1-58, Pl. 1—4. 

Bush, G. L. 1969. 

speciation 


ormation and 


— 


pani host race 


frugivorous flies of the genus Rhagoletis. 


Evolution 23: 237-251. 
Chippindale, P. T., R. M. Bonet, A. Baldwin & J. J. 
Wiens. 2004. Phylogenetic evidence no ajor reversal 


of life 1 evolut tion in 9 3 

Evolution 58: 2809-28. 

Coyne, J. & H. A Orr. 
land, Massachusetts 

Cracraft, J. 1 


4022. 
2004. Speciation. Sinauer, Sunder- 
989. Spee lation and its ontology: The empirical 


s concepts for under- 
S > 


Consequences. Sinauer, Sunderland, Massachus 
. 1997 


Spe cies conce pts in syslemalies o conser- 


vation M An aa al viewpoint. Pp. 325-339 
) M 


in M.F. C laridge. H. A. Dawah & Wilson 1 


iugis The Units of Biodiversity. Chapman & Hall, 
Lone 

. J. D. 1960. Aspects of the ecology of the 
Pacific Slender Salamander. 7 1 1 in 


Southern California. Ecology 41: 88-9 


de Queiroz, K. 1998. The general lineage concept of species 

species criteria, and the process of speciation. Pp. 57-75 in 

D. J. Howard & S. H. Berlocher (editors), Endless Forms: 
k 


Species Pm Speciation. Oxford Univ. Press, New Yor 


The general lineage concept of species and 
the d iin properties of the spec ies category. Pp. 49-89 
in R. A. Wilson (editor), Species: New Interdisciplinary 
T MIT Press, Cambridge, Massachusetts. 

2005. Ernst 


Proc. Natl. 


Mayr and the modern concept « 


species. Acad. Sci. 102: 6600-6607. 


Annals of the 
Missouri Botanical Garden 


Mr usd T. 1958. Species after Darwin. Pp. 19-55 in S. 
Barnett (editor), A Century of Darwin. Heinemann, 

22 15 n. 
Dunn, E. R. 1926. Salamanders of the family Plethodontidae. 


Massachusetts. 
Variation, Speciation, and 


Smith College, Northampton, 
A J. 1977. 
: Princeton Univ. Press, Princeton, New Jersey. 

Frost, p. h. 15 Amphibian Species of the World. 


Allen Press and Association of Systematics Collections. 


ern 


(editor). 


Funk, D. J. 
and 11 1 


„ Omland. 2003. Species-level „ 


Frequency, causes, and consequend 


with insights { from animal e DNA. Annual 
Rev. Ecol. Syst. 34: 397—423. 

1991. What 0 non-adaptive 
Biol. A Ann. Soc. 43: 263-272. 


Hall. C. 


increasing zoogeographie provine ‘iality, molluscan extin 


Gittenberger, E. radiation? 


s 2002. Nearshore marine paleoclimatic regions, 


California, Late Oligocene (2 
M. 


iocene (2.5 MA). Geol. Soc. Amer. ea 


1 
] 

| 
85 


ap« + appendices and maps. 
Hanken, 1 1990. Mi ral and Genetic Investigations 
nus Thorius). Ph.D. 
University of California. Berkeley. 


of Miniaturization in Salamanders (Gen 
Dissertation in Zoology. 


Hendrickson, R. 1954. Ecology and = systematics of 
o rs of the genus 1 Univ. Calif. Publ. 
Zool. A6. 

Hey, 3 200 l. Genes. Categories. and Species: The 


E volutionary and Cognitive Causes of the Species 


Problem. Oxford Univ. Press; New York 
Highton, R. 1989. Biochemical evolution in the slimy 
salamanders of the Plethodon glutinosus complex in the 


United States. Chapter 1. Geographic 
92 Biol. x ogr. 57: 1-78. 
Pe atina eschscholtzii a 

278 


eastern protein 
variation. 
ring-species? 

He pe B. 'a 54: 254-278. 
Houck, L. D. & S. I. Arnold. 2003. Courtship and mating 
383424 in D. (editor). Re- 


productive Biology and Phylogeny of Urodela. 


ds Pp. d 
Science 
Publishers, Enfield. New pra 

B. Wa 1994. 


ariation 


Jackman. T. R. & D. Evolutionary. and 


historical. analysi protein in the blotched 


P 


forms ol oue m of the Ensatina complex (Amphibia: 
Ple M . Evolution 48: md 


Jockusch. E. L. & D. B. Wake. 2002. 


Falling apart. and 


merging: 1 rsification of slender salamanders (Pletho 
dontidae) in the American West. Biol. J. Linn. Soe. 76: 
361— 


— ——, p. B. Wake & K. P. Yanev. 1998. New 
slender salamanders, Batrac aa (Amphibia: Ple Cus 
tidae). from the Sierra Nevada of En fornia, 

Nat. Hist. Mus.. Los Angeles Co. 472: 1-17 

. K. P. Yanev & D. 2001. Molecular 

Salamanders, 


phylogenetic analysis genus 


Batrachoseps (Amphibia: Pleuiadohtidao) from central 


coastal. California. with. descriptions of four new species. 


Herpe 9 al Monogr. 15: 54-99 


Kozak, K. I 


in n. of the Eurycea bislineata species complex. 


2003. Sexual isolation and courtship behavior 


with comments on the evolution of the mental gland and 

pheromone delivery behavior in the Plethodontidae. S.F. 
ituralist 2: 281—292. 

Kuchta, S. R. 


coloration in 


2005. Experimental support for aposematic 


salamander ring species Ensatina 


eschscholtzii xanthoptica: desa for mimicry of 


pacific newts. Copeia 2005 271. 


species of 


D. B. Wake, 


molecular phylogenetic 


L. R. Maxson & R. Highton. 1981. A 


origins of 


Larson, A.. 
perspective on the 


morphological novelties in the salamanders of the tribe 


Evolution 35: 


Plethodontini (Amphibia, Plethodontidae) 
102—422. 

. D. W. Weisrock & K. 2003. 
systematics of salamanders (Amphibia: Urodela). a 
Pp. 31-108 in D. M. Sever (editor), Reproductive Biology 
Enfield, 


I. Kozak. Phylogenetic 


review. 

and Phylogeny of Urodela. Science Publishers, 
New Hampshire. 

Marlow, W. R.. J. M. Brode & D. B. Wake. 1979. A new 

salamander, genus 1 ui the Inyo Mountains 
i h 


of California, with a discussion of relationships in the 


genus. Contr. Sci. Mus. Nat. Hist., Los Angeles Co. 308: 
l-17. 
Mavden, R. L. 1997. A de of species concepts: 


The denouement in x saga the species problem. 
Pp. 381—424. in M. F. a H. A. Dawah & 
R. Wilson Mo ae The Units of Biodiversity. 
Oxford Uni Ne York 

Mayr, E. 155 and the Origin of Species. 
C eae Univ. Press. New pes 

Moritz. C., E. J. Schneider & D. B. Wake. 1992. Evolutionary 
re 11 within the Ensatina eschscholtzii a x 

Syst. Zool. 


Systematics 


eM the ring species interpretation. 
21 


3-29] 


Mueller, R. L. 2005 
Homoplasy. and the Phylogeny of Plethodontid Salaman- 
ly 


Genome Evolution, Morphological 


Thesis in Integrative Biology, University of 
California. Berkeley, 
R. Macey, M. Jackel, D. B. Wake & J. L. 


J. R. 0⁰ 
2004. Morphological TUN 115 history E and 


ders. Ph.D. 


historical biogeography of plethodontid salamanders 
inf I ee genomes, Proc. 


320— 


Ex 


Naturalist 106: 283—292. 
2000. 
tion, Wen and Species Boundaries in th 


Phylogeography. Historical Distribu- 

Salamander 
Mitoc hondrial 
Biology, 


Ensatina eschscholtzii as Measured with 


Thesis in Integrative 
University of California, Berkley. 
Servedio. M. R. & Noor, 2003. The role of 
re d tement in speciation: Theory and data. Annual Rev 
34: 339 


INA Sequences, Ph.D. 


Ecol. ua : 339-36 


Sites. J. a Jr. &. Mars hall. 2004. Ope br criteria 
for de owe ting species; Annual Rev. Ecol, . Sysl. 35: 
a 22 

Staub, . Brown & D. B. Wake. 1995. Patterns of 


1 1 ug movements in a population of Ensatina 
esc 'ehscholtz i e ( Caudata: M Ue in the 
rra Nevada, California. J. Herpetology 29: 593-590, 


Be h. €. 1949, Speciation in la of the 


plethodontid genus Ensatina. Univ. Calif. Publ. Zool. 48: 
311—520. 
—————. 1954. Natural history of the 


salamanders of the 

ple eden genus Ensatina. Univ. Calif. Publ. Zool. 5 
12 

R. 1989. The 

\ genetic perspective. Pp. 3-27 in D. Otte & 


Te 1 ton. X. meaning of species and 
spec lation: 
J. V. Endler (editors). Spe ciation and lts Conse quences. 
ADS r. a rland. Massachusetts. 
Wake. | . 1996. 
Ple 1 ) from the San Gabriel Mountains. uk 
California. Contr. Sei. Mus. Nat. Hist.. Los | 

| 12 


73: 1-12 


| new species of Batrachoseps ( ee 


\ngeles 


Volume 93, Number 1 
2006 


Wake 23 
Species Formation in Salamanders 


. 1997. Incipient species formation in salamande rs of 


ihe pM 
7761-7767 


J 


complex. Proc. Natl. Acad. 


Ea Jockusch. 2000. Detecting species bor- 
ders in diverse data sets: Plethodontid salamanders 
in California. Pp. 95-119 in R. C. Bruce, L. Houck & 
R (editors), The Biology of the  Pletho- 
dontidae. Kluwer Academic/Plenum Publishers, New 
Yo 

& C. J. Schneider. 1998. Taxonomy of the 
up salamander genus Ensatina. Herpetologica 

29 


1980. 
” the plethodontid salamander 
holtzii of we Sie North America. Évolution 


. P. Yanev. Geographic variation in 
allozymes in a UE species 
Ensatina eschse 


40: 702-715. 

1986. 
" the plethodontid salamander 
in southern. California. 


Brown. Intraspecific 


& C 
sympatry in a “ring species, 
i 1 c 


40: 86€ 


Evolution 


M. M. 


, —— 4 Frelow. 1989. Sympatry and 
iybridization in “ring species": The 


= 


ple thodontid 
salamander Ensatina esc e holizii Pp. 134-157 in D. 
Otte & J. A. Endler (editors), Speciation aud Its 
Consequences. Sinne r, Sunderland, Massachusetts. 

— v. W. Hansen. New species of slender 
salamander genus Batrach sep, from . southern Sierra 
evada of C abdo: Copeia, 2002: 1016-1028. 

2004. Speciation and iue revisited: 

‘Phyloge netic niche conservation and the origin df: species. 
Evolution pe 193-197. 

Yanev, K. P. 1978. Evolutionary uc of the Plethodontid 

Salamander ‘Ge nus 


University of 
98 


Batrac hoseps. | . Thesis in Zoology, 
California, Berkeley 

1980. Sein hy and: distribution of three 
— Ü salamander species in coastal and borc 


California. Pp. 531—550 in D. M. 


Proceedings 


[om 


erland 
ower (editor), The 
: of a Multidisciplinary 
Symposium. Santa Barbara Museum of Natural History, 
Santa Barbara, California. 


ADAPTATION, SPECIATION, 
AND CONVERGENCE: A 
HIERARCHICAL ANALYSIS OF 
ADAPTIVE RADIATION IN 
CARIBBEAN ANOLIS LIZARDS' 


Jason J. 


Jonathan B. Losos, Richard E. Glor. 
Kolbe, and Kirsten Nicholson? 


ABSTRACT 


Caribbean Anolts lizards are a classic « 
On each island 


ecological 5 ecomorphs—on each island. 
% 8 d 1 M y 


part, but not entirely. 
achieved by : 


Key words: adaptive radiation, Anolis, Caribbean. 


l, very similar patterns of evolutionary divergene 


studying multiple hierare hic al evolutionary levels f 


ec n s 


case of adaptive radiation, repeated four times across islands of the ( Greate r Antilles. 


1e sel of 


the most complete understanding of evolutionary diverse ‘ation and radiation is 


om clades to populatio 


lizard, speciation. 


Adaptive radiation is 


members of a single phylogenetic line into a variety of 


different adaptive forms” (Futuyma, 1998). Models of 


1953: Schluter, 2000) 


begin with a species in an environment in which 


adaptive radiation (Simpson, 


resources are plentiful, but few, if any, other species 
are present to use them. Such a situation could occur 


in a number of Ways, sue h as colonization of c new 


area, evolution of a trait that allows the species to use 
resources previously inaccessible to all species, or 
surviving a mass extinction. Through time, two things 


happen: species proliferate and resources. become 
limiting. In some scenarios, speciation occurs first and 
only later do resources become limiting—this seems 
to be the 1953) had in 


Alternatively, in models, 


model Simpson mind. 


other resource limitation 
occurs first, thus driving speciation (e.g., Orr & Smith, 
1998: Dieckmann & Doebeli, 1999). 


however, the clear: 


In either case, 


main point is 


subpopulations of one initial species) alter their 


behavior and habitat use to partition resources, 


minimizing competitive interactions. Over time, the 


species (or subpopulations) evolve features to increase 


their adaptation to their new niches (ie. character 
displacement occurs). In this scenario, multiple, 
possibly simultaneous, instances. of character dis- 


placement ensue. The end-result is that a single 


species has given rise to a number of species each 


“evolutionary divergence of 


species (or 


adapted to a different part of the environment—i.e., 
an adaptive radiation. 

Both the definition of adaptive radiation and these 
models say nothing about the species richness of an 
adaptively-radiating clade. Indeed, such clades may 
not be rich in species, despite their ecological and 
phenotypic disparity. Examples of adaptive radiations 
with unexceptional species richness include Darwin's 
finches (Grant, 1986 
Shine, 1994; 
lake fish (Schluter, 
adaptive radiations Lake cichlids (Ver- 
2003). (Lovette & 


Bermingham, 1999), Hawaiian silverswords (Baldwin. 


), pygopodid lizards (Webb & 
Jennings et al., 2003), and postglacial 
1996). Nonetheless. 
e.g., Rifi 


passerine 


many other 


birds 


— 


heyen el al., 


1997) —exhibit not only great ecological and pheno- 
typie disparity, but also exceptional species richness. 
not obtained. by 
different 


but rather by the existence of sets of 


This species richness is generally 
adaptation to an extraordinary variety of 
environments, 
species that are. ecologically similar, with each set 


adapted to a different part of the environment. 


— 


Understanding the genesis and maintenance of this 


diversity obviously requires a more complicated 
model than the simple model of speciation and 
character displacement just discussed. 

The goal of this paper is to examine one species- 


rich adaptive radiation, Caribbean lizards in the genus 


Anolis Daudin, 1802. By doing so, we hope, first, to 


! We thank K. T 


The M report 


e Queiroz, T. Jackman, and . 


article. d here 
Soc iely, 
Jepartment of Biol Campus Box 1137, 
le edi u 


ANN. Missouri Bor. GARD. 93: 


was supported E grants from the 


Washington 


Larson. who have collaborated on much of the work summarized in thi: 


National Science Foundation, the National Gene oe 


and the David and Lucile Packard Fondat on: among others. 


University, Saint Louis, Missouri 63130, U.S.A. losos@ 


24—33. PUBLISHED ON 31 May 2006. 


Volume 93, Number 1 
2006 


Losos et al. 25 
Caribbean Lizards 


explore whether the classic model of adaptive 
radiation can adequately explain the ecological 


diversification. of Caribbean anoles and, second, 


examine the determinants of the great species 


richness. of this radiation. and investigate 


determinants. are related to 


adaptive differentiation. 
THE EVOLUTIONARY RADIATION OF CARIBBEAN ANOLES 


The lizard genus Anolis is one of the largest genera 
of vertebrates (and the largest of amniotes), with well 
over 300 described species, 154 of which occur on 
2005). Of 


particular interest are the anole faunas of the islands 


islands in the Caribbean (Nicholson et al.. 
of the Greater Antilles (Cuba, Hispaniola, Jamaica. 
and Puerto Rico). On each of these islands. typical 
communities contain a number of species, each of 
which is morphologically distinct, behaves differently. 
and uses a different part of the structural habitat (e. g. 
on broad surfaces near the ground, on twigs, in grassy 
areas). Detailed functional and behavioral analyses 
support the hypothesis that the morphological differ- 


ences between the species—in trails such as hindlimb 


length and toepad size adaptations to the 
functional demands imposed by living in different 
parts of the environment (e.g., Irschick et al., 1996; 
1996; Elstrott & Irschick, 2004). 


The most striking aspect of this diversity. however. 


Larson & Losos, 
is its repeated occurrence. across all four Greater 
Antillean islands. For the most part, the same set of 


abitat specialists—termed “ecomorphs” and named 


pu 
© 


| 
for the part of the habitat in which they are usually 
found (Williams. 1972. 1983)—recurs on each island. 
In all, there are six ecomorphs: trunk, trunk-crown, 
trunk-ground, grass-bush, twig, and crown-giant. Four 


are found on all four islands, whereas grass-bush 
anoles are absent from Jamaica and trunk anoles are 
only found on Cuba and Hispaniola. Phylogenetic 
analysis of morphological and molecular data confirms 
Williams’s pre-cladistic conclusion (1972, 1983) that 
the ecomorphs are independently derived on each 
1998; Poe, 2004). 


Two aspects of the ecomorph story have received 


island (Losos et al., 


relatively little attention in the years since Williams 
(1983) First. 


represented on each island, not necessarily by a single 


classic review. the ecomorphs are 
species, but sometimes by many. At the extreme, there 
are 14 trunk-ground and 15 grass-bush species on 
Cuba. 


than one species on at least one island. and at least 


All ecomorph types are represented by more 


two trunk-ground and trunk-crown species are present 
on all islands (Table 1). 


Second, in addition to the ecomorphs that have 


FS 


evolved repeatedly on different islands, a number o 


whether 


the patterns of 


Table 1. 


type on each island in the Greater Antilles. 


Numbers of Anolis species of each ecomorph 


Cuba Hispaniola Jamaica Puerto Rico 
Crown-giant 6 5 1 l 
Grass-bush 15 7 0 3 
Trunk l 6 0 0 
Trunk- 5 | 2 2 
crown 
Prunk- 14 9 2 3 
ground 
Pwig 5 | 1 l 
Other/ 12 8 1 0 
unique 
Total 58 11 T 10 


habitat specialists have evolved on only one island. 


These “unique? types include the semi-aquatic 


streamside Anolis vermiculatus Duméril & Bibron, 
1837 on Cuba, the rock wall specialist A. bartschi 
Cochran, 1928 on Cuba, and the leaf-litter specialist 
1919 on 


Hispaniola, among others. In some cases radiation 


1. (Chamaelinorops) barbouri Schmidt, 


has occurred subsequent to the evolution of such 


specialists; for example, four large species on Cuba 
(formerly comprising the genus Chamaeleolis Cocteau, 
1838) are particularly slow-moving species special- 
ized for narrow perches high in the canopy. and may 
prefer to feed on molluscan prey (Leal & Losos, 2000). 
All told, 21 such unique specialist species exist, 12 on 
Cuba, 8 on Hispaniola, 1 on Jamaica, and none on 
Puerto Rico (Table 1). 


MOLECULAR PHYLOGENETIC PERSPECTIVE 


Recent molecular studies have provided a phylog- 
Caribbean 
radiation (Jackman et al., 1999, 2002; Creer et al.. 
2001: 2001: Glor et al., 2003, 2005: 
Harmon et al., 2003; Brandley & de Queiroz, 2004: 
Poe. 2004). The most recent phylogeny (Nicholson et 


eny for most of the species in the entire 


Schneider et al., 
al., 2005), based on sequence from a 1.8 kb region of 


DNA for 132 of the 154 


species, provides the ability to thoroughly examine the 


mitochondrial Caribbean 


Caribbean anole radiation, including both ancient 


events deep in the tree and many of the more recent 
divergences. This examination reveals three patterns 


of interest that characterize. anole evolutionary. di- 


versification in the Greater Antilles (Fig 
( Ecomorphs have generally evolved only once on each 
island. riven that each ecomorph has evolved on 


multiple islands and that multiple representatives of 
each ecomorph type are present on most islands. one 
might predict that ecomorph types have evolved more 
within islands. 


than once The phylogeny, however. 


reveals this not to be the case. Most ecomorphs have 


26 Annals of the 
Missouri Botanical Garden 


Cuban Ecomorphs Hispaniolan Ecomorphs [a runi 
| Crovn-Giant 


Trunk-Crown 


— Trunk 
| Tri 


LL 
=== J Gra Banh 

———— — Ini 

y Ironton 


1—— Gris Bush 
Trunk Ground 


1 


— | bra T 
Tig 
— | miim 


El Croma 


Figure l. Island-by-island examination of the evolution of habitat specialization in the Greater Antilles. Anole de ak 


from Nicholson et al. (2005): branch lengths made ultrametric "E penalized likelihood (Sanderson, 2002). Penalized 
likelihood was implemented using the program r8s (Sanderson, 2003). The smoothing parameter was 0.90, and brane lengths 
were scaled to relative time by assigning the root node an arbitrary age = 100. Central and South American branches were 


subsequently pruned out using the program Tree Edit Mamba iil (x Charleston, 2001). Terminal taxon names were removed for 
conciseness, but may be viewed, along with the tree file, online at ER I. wustl.edu/-lososlab/anolis. mb RE 2005/7 
Ecomorph designation is based on previous is or, for some newly described species, represents new interpretations 
based on the species’ description. In a few cases (Losos & de Queiroz, 1997: Glor et al., "i Ma e species occur on 


nearby islands (e.g.. the trunk-crown carolinensis group in Cuba also. contaii 


ally similar species in the 
Bahamas, Little Cayman, Navassa, and Florida). In the figure, these species are inc and d sat thei ‘ir Gre ater Antillean relatives 
(.e.. all carolinensis group species are indicated as Cuban trunk-crown anoles 


Volume 93, Number 1 
2006 


Losos eta 


Caribbean Lizards 


Puerto Rican Écomor 


[i 
—ĩ . 


Pr 


(^ 


= 


aunt ant 
UUW UTAH 


— 
— 


—£— 
— 

= 
> 
— 
— 
— 


= 


Vini paid 
cE | nique-Cuta 


Trunk-Ground 


== | Cra Bu 
d Trunk i 


TA 


8 
Tyi 
— 8 

— 7 ˙ — 

o 

A 

t 

Figure l. Continued. 

evolved only a single time on an island, though 
evidence exists for two instances of evolution of twig 


and grass-bush anoles on both Cuba and Hispaniola 


(however, a phylogeny in which the two twig lineages 


in Cuba are 
almost as well-supported as the phylogeny in Fig. 


sister taxa, implying a single origin, is 


1). 


V 


— 
— 
=> 


= 
=> 
= — E 


d 


I z 


n 


In addition, either trunk-ground or grass-bush anoles 
have evolved twice on Puerto Rico (the nested position 
of the inolis gundlachi Peters, 1876 


within a clade of grass-bush anoles makes distinguish- 


trunk-ground 


possibilities difficult; see 


= 
= 


ing between 
Brandley & de Queiroz, 2004) and trunk-crown anoles 


Annals 
MESS 9 Garden 


may have evolved twice on Jamaica (alternatively the 
€ 


crown giant garmani Stejneger, 16 i have 


evolved from an ancestral trunk-crown anole 


— 


Ecomorphs are generally not phylogenetic alo nested 


within other ecomorphs. With few exceptions (e.g.. the 


erass-bush anole, Anolis ophiolepis Cope, 1862, which 


arose from within a clade of Cuban trunk-ground 


anoles |Fig. 1] and the two other cases mentioned 


above), most ecomorph clades are monophyletic and 
do not give rise lo other ecomorphs. The “unique” 
habitat specialists also rarely, if ever, arise from within 
a clade of one of the ecomorphs (Fig. I: one possible 


exception is ambiguous: two Hispaniolan unique 


anoles, eugenegrahami Schwartz, 1978 and 
christophei Williams. 1960. form a clade with the e Jade 


sister taxon Lo 


` His paniolan crown-giants: in turn, the 


this elade is the Puerto Rican crown-giant A. cuvieri 
Merrem, 1820). 


Ecomorph and 


w 


“unique” habitat specialist clades are 


old. Examination of Figure | indicates that almost all 
ecomorph clades arose in the first half of anole 
evolutionary history on most islands (the exception is 
Jamaica, the short branches of which 18 0 sugges- 
tions that Jamaica was underwater until well into the 
Miocene: but see Hedges, 2001). 
make them in absolute 


But i old does that 
terms? Very 
First, 
like evolution of the mtDNA 


old. ace cording lo 


several lines of evidence. if one accepts elo 


region examined. the 
maximum pairwise difference among species exceeds 
multiple substitutions), 


(after correction. for 


suggests evolutionary divergence more than 
Second, microc e nt fixation, an 


indepe de nb molecular clock estimation based on 


divergence albumin 


T dn dig bee a very 
similar estimate for the earliest. divergence within 
Mee 3540 million „ears (Shochat & Dessauer. 
1981). Third. three fossil amber anoles from the 


Dominican Republic date to the Miocene or possibly 
older (Rieppel. 1980: de Queiroz et al., 
2002). Two of 
animals (the third, described by Polcyn et al. 
little than the 
morphometrically — indistinguishable from 


998: Poleyn 


et. al., these specimens are 


Q omprise 5 more he ad) Dr are 


trunk-crown anoles. n that al 
ecomorph clade was pr > 20 million years ago 
trunk-crown 


(de Queiroz et al., 1998) oe the 


anoles from Hispaniola, the chlorocyanus 
comprise an ancient clade (Fig. 
specimens cannol be osteologically distinguished from 


extant members of this group. 


THe ROLE « 
ANOLE DIVERSIFICATION 


E ECOLOGICAL INTERACTIONS IN. DRIVING 


The classic model of adaptive radiation posits that 
ecological interactions, primarily interspecific com- 
petition, drive ecological divergence and adaptation to 
different parts of the environment (Simpson, 1953; 
Schluter, 2000). A 


however. would be 


corollary of this hypothesis. 
the prediction thal once one 
specializes for a particular aspect of the 
difficulty 


species 


environment, other lineages should have 


occupying this niche. This prediction could explain 


the paradox of ecomorph evolution: repeated evolution 
across islands, but few instances of multiple evolution 
within an island. Thus, it would appear that once an 
ecomorph evolves on an island, it achieves ecological 
incumbeney and prevents other lineages from entering 
that inability of introduced 


niche. The species lo 


become invasive. when ecologically similar species 
already occur. in an area, in contrast to their 
invasiveness in areas lacking such species, further 


supports this prediction (Losos et al., 1993) 

The adaptive radiation. scenario. hinges on the 
that interact strongly. Indeed, 
1994; 
p more recent examples include Leal et 
1999: Campbell, 2000) 


interspecific interactions 


assumption anoles 


a wealth of data (reviewed in Losos, Rough- 


Losos & Spiller, 


attests to the strength of 


among anole species. These studies come in a variety 


of forms. including experimental manipulations. 


natural experiments, and comparative ecology. Stud- 
ies document effects in terms of differences in growth 


rates, habitat use, reproductive rates, and density. 


Despite the great variety of study systems and 


approaches taken, one result is readily apparent: 
sympatric anole species tend to interact: strongly. 
Review of these studies leads to several conclusions 


1994): 


ecologically, 


Losos. first the more similar 


species are 
the stronger are the interactions between 
them: second, anole species alter their habitat use in 
the presence. of. ecologically similar species; and, 
third, 


shifts in habitat use. 


anoles evolve phenotypically in response t 


EVOLUTION WITHIN ECOMORPHS 


The evolutionary divergence and stasis of ecomorph 
types can reasonably be interpreted as the result of 
interspecific interactions. These interactions. howev- 
er, also have likely played a role in within-ecomorph 
differentiation. 

When examining diversity within ecomorph clades, 
several patterns are apparent. First; many species are 
allopatrically or parapatrically distributed. At least 
some of these allopatric forms are located in different 
mountain ranges. The most parsimonious explanation 
of these distributions is that they reflect. the early 
stages of allopatric speciation: populations that have 
become isolated and diverged to the point that they 


are likely to be reproductively isolated. In this 


context, the use by anoles of their dewlap (an 


extensible flap of skin located on the throat; Fig. 2) 


in communication is fortuitous. Svmpatric anole 


species invariably differ in some aspect of dewlap 


“design” (color, pattern, size): these differences 


used as recognition signals 


2004). As 


appear lo be species 


(reviewed in Fleishman, 2000; Losos, 


Volume 93, Number 1 


Losos et al. 29 
Caribbean Lizards 


ewlap diversity in uus 


Å. „Cuba 


Figure 2. 
(Garman, tos pu ae 
allopatrically 


a result, it is possible to identify 


distributed taxa that are likely to be reproductively 
isolated in a non-arbit 


Many of 


rary way. 


— 


these allopatrically and parapatrically 


Ph 


distributed taxa (which correspond to the concept of 
1963) appear 


“allospecies” or “semi-species”: Mayr, 


1. Anolis allogus (Barbour & Ramsden, 1919), Cuba. —b. A. grahami 
to be ecologically very similar; that is, they are 


ecologically equivalent, but occur in different places. 
Of course, because the ranges of these species are 
often small and in out-of-the-way places, the ecology 
of many of these species is poorly known, so this 


statement must be considered tentative. 


Annals of the 
Missouri Botanical Garden 


Not all of 


however. 


these species are ecologically similar, 
In some cases closely related species have 
conditions. 


differentiated to adapt to local 


example, members of the Anolis cybotes Cope, 1862 


group in Hispaniola are adapted to use a variety of 
different habitat types, such as pine forest, rocky 
terrains, and  semi-deserts (Glor et al. 2003). 


Moreover, these ecological differences are so great 
as lo permit sympatry in some cases, 
which is the trunk-ground anoles of Soroa, in western 
Cuba, where four members of the sagrei group coexist. 


All four of 


ecologically 


these species can be distinguished 
by a combination of their thermal and 
structural microhabitats (one species occurs in hot, 

open habitats at the forest edge, a second in re 1 
open areas within the forest, and two in deep forest; 
the latter two, in turn, differ in their use of boulders 
within the forest, with one species, A. mestre Barbour 
& Ramsden, 
or rock walls: 


961). 


More generally, a pattern exists in the ecological 


1916, always found on or near boulders 


2003; 


Losos et a see also Ruibal, 


means by which closely-related members of the same 
ecomorph type are able to coexist. In the ecomorphs 
that occur on or near the ground, the trunk-ground and 
erass-bush ecomorphs, sympatric species usually 
differ in thermal microhabitats, as just discussed for 
Soroa (sympatry of grass-bush anoles in eastern Cuba 
requires further study; at this point, ecological 
differences are not known, but few data are available). 
By contrast, sympatry in the more arboreal ecomorph 
lypes o -crown and crown-giant) is correlated with 
differences in body size, which correlates with prey 


1994). Thus, 


grass-bush tend to 


size 1 in Losos, sympatric 


O 


trunk-ground and anoles 


approximately the same size, but occur in different 
microclimates, whereas sympatric trunk-crown, twig, 
and crown-giant anoles tend to be more similar in 
microhabitat, but differ in body size. 

These patterns of ecological divergence in members 
of an ecomorph class lead to the following scenario: 
first, speciation occurs in allopatry. The mechanisms 
contributing to allopatric speciation within islands are 
poorly understood, but one recent study suggests that 
1 seu levels and tectonic events have 
contributed to population fragmentation and allopatric 
divergence in Cuban trunk-crown anoles (Glor et al., 
2004). Most allopatric populations remain ecological- 


Webb 


This ecological similarity precludes 


ly similar (the concept of “niche conservatism’ 
et al., 2002) 
sympatry when allopatric species come into contact, 
leading in some cases to parapatry. Second, in some 
instances, allospecies diverge to adapt to their local 


environments. Third, such divergence sometimes 


permits sympatry, leading to the coexistence of 


For 


the extreme of 


members of the same ecomorph class, which partition 
resources along axes other than the structural habitat 
axis (which is the axis that is partitioned between 
ecomorphs). An alternative possibility, of course, is 
that the differences permitting coexistence arise after 
sympatry (the classic model of character displace- 
Although detailed 


ment). examination of particular 


cases Is required to distinguish these two scenarios, 


he existence of ecological divergence in allospecies 
that 


divergence. 


a 


indicates sympatry is not required to drive 


WITHIN-SPECIES DIVERGENCE 


Recent studies have added an unexpected new twist 
to the scenario detailed above. 
the mid-1990s on 


species revealed high levels of inter-populational 


Studies beginning in 


several Lesser Antillean anole 
genetic 
hotra & 


Schneider, 


differentiation in mitochondrial DNA (Mal- 
1994; Thorpe & Malhotra, 1996: 
Although a number of subspecies 


Thorpe. 
1996). 


had previously been described for some of 


— 


these 


species (as many as 12 for Anolis marmoratus Duméril 


& Bibron, 1837 on Guadeloupe: Lazell, 1972), levels 


of genetic differentiation exceeding 9% uncorrected 


sequence divergence between geographic haplotype 


clades were unexpected, particularly for densely 


populated, mobile organisms such as Caribbean 


anoles. 
Further research, however. reveals that the same 
pattern is found in Greater Antillean (Jackman et al., 
2002; Glor et al., 2003, 2004; Kolbe et al.. 2004) and 
(Glor et al., 2001). Ongoing 


that high 


Amazonian anoles 
levels of 


DNA 


research indicates extremely 


geographic differentiation in mitochondrial 

may be the norm (R. Glor, unpublished). 
These results are exciting for several reasons. First, 

and most indicate that, 


40 years of 


ec ‘ology, 


generally, they despite 
8 ) \ } 


intensive work on anole evolutionary 


one aspect of anole diversity has been 


overlooked. Second, the results suggest the existence 
of unrecognized evolutionary units, perhaps worthy of 
the 


be much greater 


recognition as species. As a result, species 


diversity of anoles may, possibly, 


than previously recognized. Of course, robust di- 


— 


agnoses of morphologically cryptic species wil 


require more than mitochondrial DNA data (Moritz. 
1994; Sites & Crandall, 1997). 


al a nuclear locus in one Cuban group supports the 


Analyses of variation 
major intraspecific differentiation 
mtDNA (Glor et al.. 


i al least one Lesser Antillean species where the 


identified by 


2004). but the same is not true 


pattern of gene flow inferred from microsatellite data 
disagrees with that of mtDNA (Stenson et al., 2002). 


Further work examining other loci is necessary to test 


Volume 93, Number 1 
2006 


Losos et a 31 


I. 
Caribbean Lizards 


the hypothesis that these forms are genetically 
distinct. 

If these forms are genetically distinct, then what are 
currently recognized as single, island-wide species in 
the Greater Antilles may, in fact, be complexes of 
Such 


would suggest that the forms are ecologically identical 


parapatrically-distributed species. parapatry 


and thus unable to coexist in sympatry. Certainly, 


ecological differences had not previously been 
identified for most of these populations, so the 


hypothesis is reasonable, though requiring further, 
detailed investigation. Nonetheless, to the extent that 
the hypothesis is correct, then these recent findings on 
intraspecific differentiation extend the scenario for 
differentiation. proposed above for semispecies: the 
first step in differentiation may be the evolution of 
ecologically similar forms that occur in allo- or 
parapatry; evidence of these forms comes not only 
from allopatrically distributed populations recognized 
the existence of 


as different but also by 


genetically differentiated populations that are para- 


species, 
patrically-distributed and may be reproductively 


isolated. 


CONCLUSIONS 


Evolutionary diversification of Anolis lizards in the 
Greater Antilles conforms to our ideas about adaptive 
radiation. On each island, species have diversified, 
producing a set of species adapted to different 
ecological niches. Studies on extant species indicate 
that sympatric species routinely experience strong 
the strength of which is 


ecological interactions, 


a function of how similar ecologically two species 
are. Moreover, species alter their resource use in the 
presence of ecologically similar congeners and. over 


reir altered 


— 


evolutionary time, evolve adaptations to t 
the evidence 


regime of resource use. In sum. Or 


ecological interactions as the driving force in adaptive 
radiation is probably as strong for Caribbean anoles as 
it is for any other group. 

whole for the 
Caribbean Many 


species are ecologically similar, but not sympatric. 


However. this is not the story 


anole radiation. closely-related 


Ecological divergence appears to play little role in the 


— 


speciation process of these species (at least ecologica 
divergence that relates to resource partitioning among 
sympatric species; see Losos, 2004). Moreover, some 
of these semispecies have diverged ecologically to 
adapt to the particular environment in which they 


occur. Thus, not all speciation is related to ecological 


divergence, and not all ecological divergence is 
related to resource partitioning among sympatric 


species. 


Nonetheless, when these closely related species— 
which occupy the same structural habitat niche and 
are members of the same ecomorph class—are found 


sympatrically, they do partition resources. Whether 


among these species arise prior to 


ae differences 
sympatry and are a necessary prerequisite for co- 
existence, or whether they arise after sympatry in the 
character. displacement process pre- 


same sort of 


sumably responsible for ecological divergence be- 
tween the ecomorphs, is not yet clear 

Regardless, these closely related sympatric species 
indicate that adaptive radiation has occurred within, 
as well as between, ecomorph classes. At least some, 
but certainly not all, of the diversity within ecomorphs 
is the result of the same sort of evolutionary ecological 
processes responsible, at a deeper phylogenetic level, 
for evolution of the ecomorphs themselves. 

A second message to be taken from the anole story 

that adaptive radiations often must be studied at 
multiple levels: at the level of clades. species, and 
populations. Only by studying all of these levels can 
we get the most complete picture of the patterns and 
processes responsible for evolutionary diversification 


in Caribbean anoles. 


Literature. Cited 
Baldwin, B. G. 1997. Adaptive radiation of the Hawaiian 
silv versword alliance: Congruence and conflict of phyloge- 
idence from id T non- -molecular inves- 
Pp. 103-128 in ish & Sytsma 
Molecular de e pee Radiation: 
Cambridge A ress, Cambric dge. 
Brandley K. de Queiroz. 2004. 
ecomory holas EN is and historical biogeography 
of us Anolis cristatellus series. Herpe Aologic al Monogr. 18: 
90— 


^ 


ne 
tigations. 
(editors), 
Univ. 

C. Phylogeny, 


= 


Eo T. 2000. Analysis of the Effects of an Exotic 
Lizard (Anolis sagrei) on a Native Lizard (Anolis 
carolinensis) in. Florida, Using Islands as Experimental 


Dissertation, University of Tennessee, 


` R. Jackman, J. B. Losos & A. 


Creer, A., K. de Queiroz, * 
51 5 2001. Systematics a the d roquet series of 
the Southern 7 ar Antilles. J. Her 5: 428-441. 

de Queiroz, K.. | . Chu & J. B. p second 


Anolis lizard | in 10 ‘an amber and the systematics and 

ecological morphology of Dominican amber anoles. Amer. 

Mus. Novit. 3249: 1-23 

Dieckmann, U. & M. Doe lie li. 1999. On the origin of species 

by sympatrie speciation. Nature 40 54-357. 

Elstrott, J. & D. J. Irschick. 2004. Evolutionary correlations 
among morphology, habitat use and clinging B 
in oe san Anolis lizards. Biol. J. Linn. Soc. 83: 
389-398 

Fe an 
evolution üf anoline lizard dewlap c 

dsen & G. Rosenqvist 1 Animal 

mmuni— 


J. 2000. Signal function, signal efficiency and the 
olor. Pp. 209—236 in Y. 

Espmark, T. Amun 

Signals: Signalling and Signal Design in Animal C 

: Press, Trondheim, Norway. 


3rd. Ed. Sinauer. 


cation. Tapir Acaden 
Futuyma, D. J. 1998. loa Biology. 


Sunderland, Massachusetts. 


Annals of the 
Missouri Botanical Garden 


Glor; R. E., L. J. & A. Larson. 2001. A molecular 
phylogenetic n: of dive jo ation in Amazonian 
Inolis lizards. 193 "C. L col. s 12008. 

. J. J. Kolbe, R. Powell, jn K. . Losos. 


2003. 


ical. diversification. in 


Phylogenetic nicus of 1 7 al and 1 1 1 
Hispaniolan amero anoles 

(Anolis cybotes 1 Evolution 57: 2 —2397 

—, M. E. Gifford, 


Schettino, A. R. a 


7 
ez 
2 


Rodrígue 2 
2004. 
Partial island submergence and s : 
artial island submergence and speciation in an adaptive 


izo Lara & T. R. Jac bone 


analysis : the Cuban 
London B 2 


3. Losos & A. Larson. 


radiation: A multilocus green 

: 2257-2265. 
e Out of Cuba: 
overwater dispersal and noo among vm in the 
1,14 24.32. 


anoles. Proc. Roy. Soc. 


Anolis carolinensis subgroup. Molec. Ecol. 14: 2 


Grant, P. R. 1986. Ecology and Enn i 3 
Finches. Princeton Univ. Press, Princeton. 
\ 


Schule IL A. Larson & J. B. 


2003. Terie d mode of ip 


Harmon. Losos. 


radiation. in 


iguanian lizards. Science 301: 961—906 
Hedges, S. B. 2001. Noman of the West Indies: An 
overview. Pp. 15-338 in C. A. Woods & F. E. Sergile 


(e dice. Biogeography of the West Indies: Patterns and 
Perspectiy 0 RC Press, 
Fisher, J. B. 


E comparative analysis of 


. Petren, 


=a 
J- 


Losos & 0. 
clinging ability among pad-bearing lizards. Biol. J. Linn. 
Soc. 59: 21-35 


Jackman, T. R., X. Larson, K. de Queiroz & J. Losos. 
1999, Phylogenetic re n and tempo of early 
diversific p in 1 05 lizards. Syst. Biol. 48: 254-285 

. D. J. Irschick. K. de TRO J. B. Losos & A 
Larson. 1 Molec 15 pios genelic perspective on 
ee of lizards of Bh Anolis grahami series. J. Exp 

ol. Dey. E vol.) 294: 1-16. 

a. W. B., E. R. D & S. Donnellan. 2003. 

Systemati ( lizard family Pygopodidae with 


implications for the diversification of Australian temper- 
ate biolas. ya Biol. 52: 757—780 

Kolbe, J. J. 5 lor. Ds 
Lara, A. eel & J. B. 
increases during biological invasion by 
Nature 431: 177—181. 


> 


Rodríguez Schettino. A. Chamizo 


Losos. 2004. Genetic variation 


Cuban lizard. 


Larson, B. Losos. 1996. Phylogenetic syste 9 s of 
adaptation. Pp. 187-220 in M. R. Rose & G. 
(editors), Adaptation. eene 


Lazell. J. D. Jr. 1972. Th 


auder 
Press, San 11 


> anoles (Sauria: 1 :J of the 


cers 1 Bull. Mus. € s Zool. : 1-115 
Leal, M., J. A. a ¿Robles & J. B. 2 1908, in 
expe S udy of inl erspecifie interactions between 


two Puerto i an Anolis lizards. Oecologia 117: 273-278. 


Losos, 994. Integrative approaches to evolutionary 


50 lizards Annual. Rev. 


25: 467-493 


ecology: 
Ecol. Syst. 
2004. a and 

335 34: 


as model systems. 


speciation in Greater 


Antillean anoles. Pp. 335—343 in U. Dieckmann, M. 
Doebeli, J. . Metz & D. Tautz (editors). Adaptive 
Speciation. Cambridge Univ. Press, Cambridge. 

& D. A. Spiller. 1999. Differential colonization 


success and asymme ps Sut interactions between two lizard 


AER i 58. 
. Marks UN Schoener, 1993, 


and ee al interactions of an introduced and a native 


spec ie S. 


Habitat use 


species of Anolis 17 wd on Grand Cayman, with a review of 
95: 


— 


the oulcomes of anole introductions. Oecologia 


529-5 


. R. Jackman. A. Larson. de Queiroz & 
Redes z-Schettino. 199% 
in replicated adaptive radiations of island lizards. Science 
279: 2115-21 5 
. M. Leal 

odias 7 Se ne tino. 
arson. 2003. a he 

a Caribbean lizard communil 
Lovette, I. J. & E. Be 

in the new 

London B 266: 
Macey. J. R. J. 


Rastegar-Pouy 


3. Contingency and determinism 


P lor. 


zu o Lara. 


de Queiroz, P. E. Hertz, L. 
Jackman & 


— 


lability in the evolution of 
Nature 423: 542—545 
999. Explosive speciation 


rmingham. 
world Dendroica warblers. Proc. Roy. Soc. 
eee 

A. b a N. B. Ananjeva, A. Lars 
d im & T. J. 


1998. Phy 92 melic re o among neamd lizards of 


on, N. 
Pape nfuss. 
the Laudakia caucasia species group: Testing hypotheses 
(hi i cladogram 


Evol. 10: 


fragmentation and an arca 
Plateau. Molec. Phylogenet. 
Malhotra, A. &. 

land 


1994. 


lizards suggests selection. on 


Parallels between is- 
DNA 


London B 257: 37- 


R. S. Thorpe. 

mitochondrial 
and morphology. Proc. Roy. Soc. 
42. 

Mayr, E. 1903. 
Press, Cambridge, 

Moritz. C. 1994. De 
for conservation. Trends Ecol. Evol. 

Nicholson, K. E.. R. E. Glor, J. ae Y Larson, S. B. 
Hedges & J. B. 2005. Bee colonization. by 
island lizards, J. Biogeogr. 32: 

Orr, M. R. & B. Smith. e 

Prends E 2. ol. 13: 5 

E. 5. 20 


Animal Species and Evolution. Belknap 
Massachusetts. 


fining ce PU units” 


Losos. 


rad and speciation. 
2 500. 


1 of Anolis. Herpetological Monogr. 


. Rogers II. Y. 


Combiited ci d of an Anolis 


Kobayashi & L. L. Jacobs. 
p in 
Dominican amber: Systematic, taphonomic, biogeographic. 
tions. Pala eontological Electroni l 


and s tionary 1 5 
Vo <http://palaeo-electronica.org/2002_1/amber/ 


ca, 
main.hti 
te Ta o. (^ 30. Green anole in Dominican amber. Nature 
150—157 

18 Caribbean: 
Oxford. Univ. 


Lizards 
Plate 


Anolis 


l. 
J- 1995 
Ecology, Evolution, and Tectonics. 
Press. 
Ruibal, R. 1961. Thermal relations of five species of tropical 
lizards. E one 15: 98-111. 
Schluter, D. 1996, 
Ta Trans. Roy. Soc. 
2000. The 
l de Press. Oxford. 
Schneider, C. J. 
secondary intergradation among morphologically differen- 


Ecological speciation in postglacial 
| 


fishes. London 351: 807-8 


Ecology of Adaptive Radiation. Oxford 
1996. Distinguishing between primary and 
lale a T of Anolis marmoratus. Molec. Ecol. 
239— 


& K. de 2001. 
the Anolis ee group from the 
Antilles. J. Herp. 35: 1-12, 

. € 1981. 


“albumins of 


Losos Queiroz. Evolutionary 


O ips of 
orthern Lesser 

T hat. N 
tive immunological stu 


& H. C. 5 1 Compara- 
Anolis lizards of 
Physiol. 68A: 


dy of 


the Caribbean islands. Comp. Biochem. 


67-73. 


Simpson. G. G. 1953. The Major Features Evolution. 


) 


Columbia Univ. Press. New York. 
Sites, J. W. & Crandall. 1997. Testing iu 
bound: 


aries in Duis srsity studies; Conservation Biol. 
97 


289-12 


Volume 93, Number 1 
2006 


Losos et al. 33 
Caribbean Lizards 


Stenson, A. G., A. Malhotra & R. S. Thorpe. 2002. Population 
differentiation and nuclear gene i in the Dominican 
88. 


anole ugs oculatus). RE ol. 11: 1679-16 
Thorpe. R. & A 1 E Molecular and 
phil a evolution within small islands. Philo 


H 


Trans. Ros . London 351: 815-822. 
E. 


Verheyen, Salzburger. J. Snoeks & A, Meyer. 2003. 
)rigin of te superflock of cic e fishes from Lake 
Victoria, East Africa. Science 300: : 329, 

Webb, C. 0., D. D. Ackerly, M. a Er ek & M. J. 


Donoghue. 2002. Phyloge ‘nies and community ecology. 
5 


Annual Rev. Ecol. Syst. 33: 475-5 


Webb. J. K. & 


and reproductive 
of 


Shine. 1004. 


DN of Aust 


Feeding habits 
ralian | pygopodid 


lizards the genus Aprasia. Copeia 1994: 390- 
398. 

Williams, E. E. 1972. The origin of faunas. Evolution of 
lizard congeners E a complex island fauna: A trial 


>: 47-89. 


.E NIA oe aunas, 


a 1 85 Biol. 
and diverse 
326-370 in 


. Schoener (editors). 


island size. 
end jin: in island radiations of Anolis. Pp. 
1. B. Hue i 


E. R. Pianka & 


Lizard Ecology. Studies of a Model Organism. Harvard 


Univ. Press. Cambridge. 


HYBRID SPECIATION IN Loren H. Rieseberg? 
WILD SUNFLOWERS' 


ABSTRACT 


lybrid speciation refers to the establishment of novel hybrid genotypes that are reproductively isolated from their parental 
species and genetically stabilize d. Most fre que ntly, reproductive isolation is achieved via an increase in ploidy. However, in 


some instances new hybrid species arise and become reproductively isolated 18 a change in chromosome number, 
a process known as diploid or “homoploid” hybrid speciation. The annual sunflowers of the genus Helianthus provide a well- 
studied example of this latter mode of speciation. Here, | review this work, placing individual studies in their proper context. 
These include (1) computer simulations that describe the evolutionary conditions under which hybrid speciation is most likely: 
(2) molecular phylogenetic studies that document the origins of three hybrid sunflower species: (3) comparative genetic 
mapping studies that describe the karyotypic changes associated with hybrid speciation; (4) experimental re-creations of 
homoploid hybrid species that allow genotypic and phenotypic ed s between synthetic and ancient hybrid lineages: 
(5) quantitative trait locus (QTL) studies that describe the genetic basis of 5 pic differences between the parental species 
and the mode of gene action underlying the generation of extreme 1 555 types in hybrids: (6) phylogeographic studies that 
estimate the ages and number of origins of each hybrid species: (7) selection studies that measure the strength of selection on 


individual traits and QTLs in synthetic hybrids transplanted into hybrid habitats: aun (8) candidate gene studies that search for 


correlations between candidate genes for ecological divergence and traits and QTLs shown to be under selection in the habitats 
of the hybrid species. Ongoing work includes searches for the molecular signature of selection during hybrid speciation, 


surveys of gene expression shifts associated with hybrid speciation, and experiments that evaluate the role of new hybrid gene 
combinations versus reproductive isolation in the ecological divergence of hybrid lineages. 
ey w 


ords: ecological divergence, genetic mapping, Helianthus. homoploid hybrid speciation, QTLs. 


eproductive 
isolation. 


Species often come into contact before reproductive from hybrid zones, which allows fit hybrid gene 


barriers are fully formed, resulting in the formation of combinations to become established (Buerkle et al.. 


hybrid swarms or hybrid zones. The consequences of 2000). Although seemingly rare, these stabilized 
hybridization are most frequently discussed in relation hybrid lineages are much easier to detect and study. 
to the reproductive isolation of the hybridizing As a consequence, much of what we know about the 
lineages. That is, the hybrid zones may be stable. role of hybridization in adaptive evolution comes from 


reflecting a balance between dispersal and selection the study of these hybrid races or species (Rieseberg 
(Barton & Hewitt, 1985). Alternatively, they may be et al., 2003). 
ephemeral, due either to the strengthening of re- Finally, a significant component of the literature 
productive barriers and cessation of hybridization on hybridization and its consequences has focused 
(Noor. 1999) or to the weakening of reproductive on the means by which these new hybrid lineages 
barriers and fusion of the hybridizing lineages (Wolf become reproductively isolated from their parental 
et al. 2001). All of these consequences are well species. Most commonly, isolation arises as a by- 
documented empirically, although their frequencies product of hybrid genome duplication or allopolyploi- 
remain poorly understood. dy (Stebbins, 1950). However, reproductive isolation 
‘he consequences of hybridization may also be sometimes develops without a change in chromosome 
discussed with respect to adaptive evolution (Ander- number, a mode of speciation called diploid or 
son, 1949: Arnold, 1997). In stable hybrid zones, for homoploid hybrid speciation (Grant. 1981). 
example. favorable alleles will move rapidly across In this paper. | review a series of studies conducted 
hybrid zones, and thereby contribute to adaptation in by my lab over the past 15 years that describe the 
the recipient species. This "adaptive trait introgres- origins of three diploid hybrid sunflower species. the 
sion” may be common, but it is difficult to detect means by which they became reproductively isolated 
because the introgressing alleles will spread rapidly from their parental species, and the theory underlying 
to fixation and are unlikely to be “caught in the act” the process. This narrative serves several purposes. 


(Barton, 2001). Less frequently, hybrids may escape — First, our theoretical and empirical work provide 


1 


he author's research on homoploid hybrid speciation has been supported by the U.S. National Science Foundation an 
the National Institutes of Health. 
Department of Biology. Indiana University, Bloomington, Indiana 47405, U.S.A. 


ANN. Missouri Bor. Garp. 93: 34-48. PuBLISHED ON 31 May 2006. 


Volume 93, Number 1 
2006 


Rieseberg 35 
Hybrid Speciation 


convincing evidence that hybridization can contribute 
to adaptive evolution and suggest that it may 
sometimes serve as a mechanism for large and rapid 
evolutionary transitions. Second, the work illustrates 
the kinds of experiments required to rigorously assess 
hybridization’s role in evolution. Finally, the sunflow- 
er hybrid speciation story is currently scattered across 
the literature, appearing in more than 25 papers. This 
narrative ties these pieces together, placing individual 
studies in their proper context. 
HYBRID SPECIATION: THEORY 

Hybrid speciation refers to the process in which 
novel genotypes generated by hybridization become 
isolated from their parental species and other hybrid 
he- 


productive isolation is straightforward for hybrids with 


genotypes and become genetically stabilized. 


increased ploidy because hybrids between polyploids 
and their diploid parents typically have an odd 
therefore be largely 


number of genomes and may 


— 


sterile (Grant, 1981). In contrast, homoploid hybrid 


speciation represents a challenge to evolutionary 


theory, because homoploid hybrid lineages must 
become established despite the possibility of back- 
crossing with their parental species (Turelli et al.. 
2001 

Three mechanisms have been proposed by which 
become reproductively 
First, the 


a homoploid hybrid may 
isolated from its parental species. new 
hybrid lineage may diverge karyotypically from its 
parental species through the sorting of chromosomal 
rearrangements that differentiate the parental species 
(Stebbins. 1957: Grant. 1981). 


lishment of new chromosomal rearrangements induced 


and/or by the estab- 
by recombination (Rieseberg et al.. 1995b). Second, it 
has been speculated that hybrid dere events may 
facilitate hybrid speciation by providing initial spatial 
isolation for the new hybrid lineage (Charlesworth. 
1995). Third. 
pitat or niche and thus become ecologically isolated 
1981). Hybrid 


species often occupy habitats that are very different 


a hybrid lineage may colonize a new 


— 


ha 


-. 


from its parental species (Grant, 
from those occupied by their parental species (Riese- 
2001: Gross & Rieseberg, 


This is probably because a novel hybrid 


7; Wang et al., 
2005). 
lineage is more likely to survive if it experiences little 
competition with its parental species and if di- 
vergence in habitat preference results in some spatial 
isolation with the parental species (Templeton, 1981: 
Charlesworth, 1995). 

Three simulation studies have examined the pro- 
cess of homoploid hybrid speciation (McCarthy et al.. 
1995; Buerkle et al., 2000, 2003). 


was a spatially explicit individual-based simulation, 


The first of these 


which monitored the consequence of hybridization 
between species that differed in karyotype (McCarthy 
et al., 1995). There were two possible outcomes—the 
generation of a stable hybrid zone or the origin of 
a novel hybrid genotype that displaced the parental 
species. The latter outcome was considered hybrid 
speciation, but it probably is more correctly viewed as 
the merger of species through hybridization. In most 
documented hybrid species, the parental species 
continue to coexist with the derived hybrid (Riese- 
berg, 1997). ' 


simulation model to investigate the conditions for this 


Phus, our group developed a computer 


more conventional kind of hybrid speciation (Buerkle 
Our 


model was similar to that of McCarthy et al. (1995) in 


et al., 2000), which is the focus of this paper. 


its inclusion of chromosomal rearrangements, but 
differed by providing a novel habitat in which hybrids 


were favored, as well as the possibility of spatial 


separation between the novel hybrid habitat and that 
of the parental species. 


Three 


model: 


possible outcomes were possible in our 


hybrid zone stasis, hybrid speciation, and 


adaptive trait introgression, in which one of the 
parental species colonizes the novel habitat by 


acquiring advantageous alleles from the other parent. 
This latter outcome was by far the most common. 
Hybrid speciation was less frequent, but it did occur 
at a significant rate when the sterility barrier isolating 
the parental species was weak and hybrid genotypes 
favored in novel. habitat (i.e., 


were strongly open 


strong ecological selection). In contrast. hybrid zone 
o O / J 


stasis was rare, apparently because of the presence 


of novel habitat in which hybrid genotypes had 
a fitness advantage. 
The conditions required. for the independent 


evolution of the new hybrid species following its 


origin are surprisingly different from those favoring 
2000). 


though a weak sterility barrier between the paren- 


establishment in the first place (Buerkle et a 
A 


tal species favors speciation, it also translates into an 


— 


even weaker barrier between the new hybrid lineage 


and its parents. As a consequence, the genetic 
integrity of the new hybrid lineage cannot be 


maintained in parapatry with parental populations. 
Thus, for hybrid lineages to evolve independently 
from parapatric parental populations, the initial ste- 
rility barrier between the parental species must be 
strong. a requirement that greatly reduces the likeli- 
hood of mode of speciation. Alternatively, 
isolation of the hybrid neospecies can be maintained 
by strong ecological selection and/or spatial isolation, 
that 
predictions that hybrid speciation is most likely 


a result accords well with earlier verbal 


following hybrid founder events, in which a few 


hybrids found a new population that is ecologically 


Annals 
Missouri Botanical Garden 


of th 


isolated from the parental species 


1995; Rieseberg, 1997) 


and spatially 
(Charlesworth, 


The third simulation study (Buerkle et al., 2003) 
was an extension of Buerkle et al. (2000), except that 


the novel habitat for hybrids was eliminated, the sizes 
the hybridizing parental populations were allowed 
and the 
This new model allowed us to assess the 


of 
to vary, extent of spatial isolation was 
reduced. 
relative frequency of the four most important outcomes 
of hybridization: hybrid zone stasis, hybrid speciation, 


one ol 


— 


adaptive trait introgression, and extinction o 
the parental species (typically the rare one). Hybrid 
zone stasis was most common, occurring in 84% of 
simulations. Extinction of the rarer species occurred 
only when the sterility barrier was weak and was 
essentially always the result of adaptive 
Hybrid speciation occurred in only 2.1% 


trait in- 
trogression. Ha 
of the simulations and was favored by a weak sterility 


barrier between the parental species and moderate 
strong ecological selection, However, hybrid specia- 
tion was almost always accompanied by extinction of 
one of the parental species. These results indicate that 


the presence of novel habitat greatly facilitates hybrid 


speciation, as has been suggested by numerous 
students of hybrid speciation over the past century 
(e.g.. Kerner, 1894-1895; Grant, 1981: Templeton, 
1981: Arnold, 1997: Riesebere. 1997). Also. without 
an open niche, one of the parental species is replaced 
by the hybrid, and total species diversity remains 
unchanged. 

All three simulations studies have one remarkable 
result in common. When hybrid speciation does 


happen, it occurs quickly, often in fewer than fifty 
Met (1995) referred to 
homoploid hybrid speciation as a punctuated mode 
which long periods of hybrid zone 


generations, Thus. varthy et al. 
of speciation, in 
stasis are followed by abrupt transitions 


new fit hybrid genotype becomes established. 


which a 


HYBRIDIZATION AND THE EVOLUTIONARY HISTORY 


OF SUNFLOWERS 


The role the diversification of 
North American (Helianthus L.) 
initially explored by Charles Heiser and his students 
(Heiser, 1947, 1949, 1951a, b; 1969). 
Heiser documented the occurrence of natural hybrids 
the H. 
and several of its congeners, including H. 
argophyllus & A. Gi 1951a), H. 
bolanderi A. Gray (Heiser, 1949), debilis Nutt. 
(Heiser, 19510). H. petiolaris Nutt. (Heiser. 
1947). This hybridization. allowed H. 


annuus to acquire favorabl 


of hybridization in 


sunflowers was 


Heiser et al., 


between widespread, common sunflower. 


annuus L. 
Torr. ay (Heiser, 
and 
Heiser argued, 
e alleles from the 
increasing its ecological 


adapted species, thereby 


locally 


amplitude and facilitating range expansion. In some 
instances, hybridization was believed to have contrib- 
uted to the formation of introgressive races (H. annuus 


1951b) and 


California (Heiser, 


subsp. texanus Heiser in Texas (Heiser, 
H. bolanderi subsp. bolanderi 
1949)) or new species (H. neglectus Heiser (Heiser el 
al., 1969)). 

My laboratory re-examined these hypotheses using 


a combination of molecular marker surveys and 
molecular phylogenetic analyses (Dorado et al., 
1992; Rieseberg et al., 1988, 1990a, b, 1991a, b: 


1991). We verified much of Heiser's work. 


Rieseberg, 
hybridization 


including the occurrence of natural 


between Helianthus annuus and each of the species 
1992: d 


^ 


listed above (Dorado et al., Rieseberg « 
1988. 1990a. 19098). 
origin of H. annuus subsp. 
1990b). On the other hand. 
failed to support a hybrid origin for H. neglectus or H. 
bolanderi subsp. bolanderi (Rieseberg et al., 1988). 
Thus. while the morphological criteria employed by 
hybrid 


b. as well as the introgressive 
texanus (Rieseberg et al.. 


the molecular evidence 


contemporary 
the d 


Heiser accurately identified 


swarms or zones, their utility in election of 


ancient, stabilized hybrid lineages was limited. This 
was nol unexpected, as it has long been recognized 
that morphological intermediacy may sometimes 


result Hom es other than hybridization (Gott- 
lieb. 1972 

In Po 
examples of hybrid lineage formation, 


to verifying or refuting putative 
molecular 
phylogenetic analyses detected unsuspected hybrid 
origins of three species (Fig. 1): Helianthus anomalus 
Blake. H. deserticola Heiser, and H. paradoxus Heiser 
(Rieseberg et al., 1990b, 100 la: 1991). 
The three hybrids appear to be derived from the same 
parental H. H. petiolaris. 


However, there are subtle differences in the parental 


Rieseberg, 


species, annuus and 


chloroplast and nuclear ribosomal DNA haplotypes 


found in each hybrid, suggesting that each was 
independently derived (Rieseberg, 1991; Rieseberg 
et al., 1991a). Over the past decade we have exploited 
this replicated. natural hybrid speciation experiment, 
in combination with theoretical studies. to derive 


a more predictive theory of hybrid speciation. This 
body of work is described below. 


NATURAL HISTORY 


The three hybrid species and their parents are self- 
-pollinated annuals, with the 
same chromosome number (n = 17). All five species 
are native to the continental United States. The 
parental species have widespread and broadly over- 
lapping distributions that are centered in the U.S. 
Great Plains (Fig. 2). They differ in soil preferences, 


incompatible, insect 


lwo 


Volume 93, Number 1 
2006 


Rieseberg 37 
Hybrid Speciation 


100 


100 


H. petiolaris 3 


perennial Helianthus 

H. niveus subsp. niveus 

H. praecox 

H. debilis 

H. niveus subsp. tephrodes 
H. niveus subsp. canescens 


H. neglectus 


NE 
YON " „ anomalus (sand dune) 
x 
8. cH deserticola (desert floor) 
* 
DES ^H. paradoxus (salt marsh) 
CL. 


5 H. annuus * 


H. argophyllus 


85 H. bolanderi 


] 


71 H. exilis 


Figure 1. Phylogenetic tree for Helianthus se D. He o based on combined chloroplast DNA and nuclear ribosomal 
fic 


DNA data (redrawn from Riesebe 199], fig. The 


below each branch. 


however, with Helianthus annuus largely restricted to 
heavy. clay soils, and A. petiolaris to dry, sandy soils 
(Fig. 3). Nonetheless, 
found in close proximity in the central and western 


— 


these two habitats are often 
United States, resulting in the production of numerous 
hybrid swarms and hybrid zones. The hybrid zones are 
narrow, often less than 30 m, and little evidence of 
introgression is found outside of the hybrid zones. 
apparently due to the synergistic action of several 
reproductive barriers (Rieseberg et al., 1995a, 1999: 
Schwarzbach et al., 2001 

The three hybrid species are much more limited in 
geographic distribution than their parents (Fig. 2). 
with restriction of Helianthus deserticola to the Great 
Basin Desert in Nevada, Utah, and northern Arizona. 
H. anomalus to a handful of sand dune habitats in 


Utah and northern Arizona, and H. paradoxus to saline 


wetlands in western Texas and New Mexico (Fig. 3). 
Despite proximity to parental populations, no natural 
hybrids have been reported between the three ancient 
hybrid species and their parents. In contrast, the three 


hybrid species are almost completely allopatric t 
each other, but I have observed hybrids between H. 


anomalus and H. deserticola at the only site where 


Dashed lines indic ‘ale parentage of homoploi 


imber of mutations are given above and bootstrap percentages 


hybrid speci ies 


Little Sahara Recreation Area in 


hey co-occur, 
Central Utah. 

Wild sunflower species produce an indehiscent 
one-seeded fruit called an achene (or cypsela). 
Although the achenes are eaten by small mammals 


and birds, dispersal by way of the animal feces seems 


— 


unlikely because most seeds are likely to be digested. 
It seems more likely that sunflowers are dispersed by 
large mammals such as bison. Wild sunflower achenes 
are covered with small hairs that can bind to fur, and 
sunflower achenes have been reported in buffalo fur 
(Asch, 


of an association between bison and sunflowers. For 


1993). There also are reports from early settlers 


example, in 1839 journalist Matthew Field described 


how “among the sunflower beds the huge back of 
a buffalo here and there was seen, as the ponderous 
brute broke down the stalks before him while pressing 


toward a fresher pasture ground” (cited in Asch, 1993: 
] 3 b 


— 


2). 

In addition to their role in dispersal, the distur- 
bance generated by bison may have created opportu- 
nities for hybridization. Hybridization has long been 
known to be associated with natural and anthropo- 
Anderson, 1948), and, as shown by 


genic disturbance 


38 Annals of the 
Missouri Botanical Garden 


A D 
SS 


SS 


i 


Helianthus annuus 
Bll Helianthus anomalus 

N Helianthus deserticola 
E Helianthus paradoxus 
Helianthus petiolaris 


Figure 2. Geographie distributions of the two parental species and their three hybrid derivative species. 


the following passage from Barsness (1985: 1), the To assess the role of genomic restructuring. in 
impact of bison on the Great Plains was profound: hybrid speciation, we have generated detailed genetic 
“The [bison] herds marked their territory with trails, linkage maps for the three hybrid species. Helianthus 
thousands of them, some shallow traces, some eight- anomalus, H. deserticola, and H. paradoxus. as well as 


to-ten-inch trenches, others so deep that the animal’s their. putative parental species, H. annuus and H. 


sides would rub the embankments.” Reflecting either petiolaris (Rieseberg et al., 1993, 1995b; | ngerer e 
cause or coincidence, microsatellite divergence al., 1998; Burke et al., 2002: Rieseberg et al., 2003). 
places the origin of the three hybrid Helianthus These maps were developed using a combination of 
species between 63,000 and 210,000 generations AFLP, RAPD, and SSR (microsatellite) markers, with 


ago (Schwarzbach & Rieseberg. 2002: Welch & more than 650 markers placed on each map. 


Rieseberg, 2002b: Gross et al., 2003). or just after Alignment. of orthologous markers across the five 

the colonization of North America by bison ca. maps allowed us to reconstruct. the chromosomal 

200,000 years ago (Dary, 1974). rearrangements that have accompanied hybrid speci- 
alion. 

KARYOTYPIC EVOLUTION All three hybrids have undergone massive karyo- 

typie re-organization. Not only does each of the hybrid 

As discussed earlier, theory indicates that the Species possess a unique combination of parental 


development of reproductive isolation between the chromosomal rearrangements, but also a minimum of 
new hybrid lineages and their parental species may be three, two, and five chromosomal breakages (and an 
facilitated by rapid karyotypic evolution (Grant, 1981; equal number of fusions) are required to achieve the 
McCarthy et al., 1995: Buerkle et al., 2000, 2003). Helianthus anomalus. H. deserticola. and H. paradoxus 


Most verbal and simulation models have assumed that genomes, respectively (Rieseberg et al.. 1995b. 2003: 


the karyotypic change occurs through the sorting of Burke et al., 2004: Lai et al. 2005). We have 
preexisting chromosomal rearrangements that differ- previously shown that approximately 50% of the 
entiate the parental species (Stebbins, 1957: Grant, — reproductive barrier between Helianthus species is 
1958). However, Templeton (1981) noted that re- caused by chromosomal rearrangements (Rieseberg et 


combination in hybrids might lead to additional al.. 1999b). Furthermore. crossability and interfertility 


chromosomal breakage. studies (Rieseberg, 2000. unpublished) indicate the 


Volume 93, Number 1 
2006 


Rieseberg 39 
Hybrid Speciation 


anomalus (hybrid) 


Figure 3. Photographs of the two yarental species 
n 


Photographs by Jason Ki 


three hybrid species are strongly isolated from each 


other and from their parental species (Fig. 4). Thus, as 
predicted by theory, rapid karyotypic evolution in the 
hybrid sunflower species appears to have facilitated 


the development of reproductive isolation with their 


Buerkle et al., 


parental species (Templeton, 1981: 


— = paradoxus- 
deserticola ( 0 0 (hybrid) 


and their three hybrid derivative species in typical habitats. 


EXPERIMENTAL SYNTHESES OF HoMOPLOID HYBRID SPECIES 


There is a rich literature describing the recovery of 
fertility in synthetic hybrid lineages and the degree of 
their isolation from their parental species (Stebbins, 
1957; Grant, 1966a, b, 1981). 
were important they 


These early studies 


because demonstrated that 


40 


Annals of the 
Missouri Botanical Garden 


synthetic 
lineages 


deserticola 


4.90 


annuus 


paradoxus 


anomalus 


Figure 4. 


derivative species. Percentages indicate pollen viability 


reproductive compatibility . 


homoploid hybrid speciation was feasible. However. 


0 


they did not attempt to replicate naturally occurring 


hybrid. species or compare the newly synthesized 


hybrid lineages with those already existing in nature. 


The Helianthus work fills this important gap in the 


literature, 
We attempted to replicate the origins of the hybrid 


species by synthesizing hybrids between the two 


parental species, Helianthus annuus and H. petiolaris. 


and subjecting the synthetic hybrids to selection for 


jen over several generations (Rieseberg et al., 
Because sunflowers are self-incompatible and 
generation. hybrids are semi sterile, hybrid 
speciation in this group likely involved both back- 
crosses and 


both 


Therefore, 
bul 


filial crosses. we employed 
the 


backcross and filial generations in three independent 


kinds of crosses. varied order 
experiments. 


The Alter 


generations of selection. pollen fertility 


results were gratifying. only four 
in the three 
5.6 to 91.8%, 


f fertility 


hybrid lineages had increased. from 


indicating that the recovery o not only was 


of 


Crossability and interfertility among the two parent il species, synthetic hybrid lineages, and three hybrid 


hybrids. Line thicknes 


irs gene ration 


Is proportional 10 


feasible, but that it potentially could occur very 
the 


hybrid lineages were strongly isolated from 


rapidly (Ungerer et al., 1998). In addition. 


synthetic 
their parental species, 
12% 


with mean pollen fertility 
ranging from 29 to for crosses with Helianthus 
annuus and 10 to 13% for crosses with H. petiolaris 
(Fig. 4; 2000). the 


independently generated hybrid lineages were highly 


Riesebere, Remarkably, three 


interfertile when crossed with each other: mean pollen 


39% 2000). 


fertility ranging from 83 to 8 
This result implies that the three hybrid lineages have 


(Rieseberg, 


converged onto a fairly similar set of parental genes 


or chromosomal segments. Indeed, meiotic analyses 
indicated. the synthetic lineages were similar in 


karyotype, varying in only one or two reciprocal 


2000). 


To explore the genomic similarity of the synthetic 


translocations (Rieseberg, 


— 


hybrid lineages, we screened ca. 60 individuals from 
lo 


Li 


1996). Genomic congruence, as measured by the q 


each lineage with 197 molecular markers specific 


one or the other parental species (Rieseberg et : 


coefficient (a standard measure of association that can 


Volume 93, Number 1 
2006 


Rieseberg 41 
Hybrid Speciation 


vary from — I to l; Sokal & Rohlf, 1995). ranged from 
0.65 to 0.75 (P < 0.0001) for comparisons among the 


three hybrid lineages, a result that accords well with 


the high fertility reported for crosses between these 
lineages (Rieseberg, 2000). 


Given the congruence in genomic composition 
among the synthetic hybrid lineages, it seemed sensi- 
ble to determine if the three natural hybrid species 
had found a similar solution to the problem of hybrid 
This was a more difficult problem that 
both He- 
determine the 
the 
| 905b. 


and 325 


sterility. 
required surveying natural „ ol 
lianthus annuus and H. petiolaris 1 
likely parental origin of each de mapped 
three hybrid species (Rieseberg et al.. 1993, 
2003; Ungerer et al., 1998). In all, 427, 290, 

of the markers mapped in H. anomalus, H. deserticola, 


and H. paradoxus, respectively, could be assigned to 
one or the other parental species and thus were 


with respect to genomic composition 
. 2003). Genomic 
between the three 
hybrid lineages and either H. anomalus or 
deserticola: ꝙ ranged from 0.44 to 0.50 (P < 0.0001) 
for comparisons with A. anomalus (Rieseberg. 2000) 
and 0.53 to 0.59 (P < 0.0001) for comparisons with H. 
itle 
congruence was observed between the three synthetic 
to 0.34 (P < 0.01). 


that levels of genomic congruence accord 


informative 


(Rieseberg et a congruence was 


high for comparisons synthetic 


— 
e 


deserticola (Rieseberg, unpublished). In contrast, 


lineages and H. paradoxus: 0.28 
Also, note 
crossing relationships (Fig. 4); the 


closely with 


synthetic hybrids are most compatible and most 


interfertile with H. 


predicted by genomic congruence, but least compat- 


anomalus and H. deserticola as 


ible and least interfertile with A. paradoxus. From 


a broader perspective, these results indicate that 


fertility selection plays a major role in determining 


hybrid genomic composition and that the number of 


solutions to the problem of hybrid sterility may be 
Clearly, H. 


a different solution than the other taxa. Possibly. there 


fairly limited. paradoxus has found 


— 


were different proportions of the parental species in the 
hybrid zone from which H. paradoxus was founded and/ 


or the genomic composition of H. paradoxus was 


strongly influenced by ecological selection (Lexer « 
al., 2003a, b: 2003). 


Rieseberg et al., 


Tempo or HYBRID SPECIATION 

Both theoretical and experimental studies of homo- 
ploid hybrid speciation indicate that it is likely to occur 
>rant. 1966a. b: 
1995: Buerkle et al.. 


methods 


very rapidly (Stebbins, 1957: ( Riese- 
1996; McCarthy et al.. 


until 


berg etal., 
2000), but 


to test this prediction. We 


recently no have been 


available have devised 


a method for estimating the tempo of hybrid speciation 


in nature (Ungerer et al.. 1998). The rationale for our 
method is that in a newly forming hybrid species the 
sizes of parental chromosomal segments are expected to 
become progressively smaller over time due to re- 
1953; Baird, 1995). 


continued reduction in segment size will be countered 


combination. (Fisher. However, 
by stabilization of the hybrid species” genome: sub- 
sequent recombination will be among segments derived 
1998). 
This analysis represents an empirical application of R. 
1953), 


tracks parental species segments by monitoring re- 


from the same parental species (Ungerer et al., 


V. Fishers junctions approach (Fisher, which 


combination breakpoints or "junctions? between het- 
erogeneous regions rather than all points on a genome. 
In a neutral case, we can imagine a “junctions clock. 


which is analogous to the molecular clock, but slows 
over time as heterozygosity decreases due to drift. By 
comparing the frequency spectrum of observed parental 
species" chromosomal segments with predictions based 
on computer simulations (Ungerer et al., 1998), 1 
possible to estimate the number of generations required 
to stabilize the genome of a hybrid species. Note that 
we are estimating the speed of hybrid speciation, 
not the age of the hybrid species. 

To date, we have only estimated the tempo of 
speciation for one of the hybrid species. Helianthus 
1998). 


argest number of mapped species-specific markers. 


anomalus (Ungerer et al.. This species has the 


thereby allowing the most precise empirical estimates 


chromosomal lengths (Ungerer et al., 


1998). 


arise in hybrid founder populations that are spatially 


of segment 


lt seems likely that the hybrid species will 


— 


and/or ecologically isolated from the parental species. 


Therefore. we modeled this scenario by simulating 


| hybrid swarm, starting with equal numbers of H. 
annuus and H. uM individuals, and allowing 


recombination to take its course in small. closed 


populations of en sizes and under different 


selection regimes. The speed of hybrid speciation was 


estimated by comparing the sizes of chromosomal 


segments in the simulation studies with those of H. 


anomalus. Chromosomal segment sizes in the H. 
anomalus genome were large, suggesting that its 


genome became stabilized in 10 to 60 generations. 


This result accords well with earlier simulation 


studies of hybrid speciation in which new hybrid 
established in ca. 50 
1995; Buerkle et al., 


2000) and argues that the tempo of homoploid hybrid 


species typically became 


generations (McCarthy et al.. 
speciation, like allopolyploidy, is very rapid. 
DIVERGENCE 


PHENOTYPIC 


The three hybrid species inhabit what are arguably 


the most extreme habits of any sunflower (dune, desert 


42 Annals of the 
Missouri Botanical Garden 


Table 1. Hypothetical example of transgressive segregation due to the complementary action of genes with additive 
effects (from Rieseberg et al., 2003) 


Phenotypic values 


QTLs Species A (AA genotype) Species B (BB genotype) Transgressive Fə Transgressive F» 
| +] =] +1(AA) — KBB) 
2 +] l +1(AA) — I (BB) 
3 +] =] +1(AA) = 1(BB) 
| =i +] +1(BB) — (AA) 
5 — +] +1(BB) — l(AX) 
Total +1 =] +5 =) 


floor. and salt marsh habitats; Fig. 3) and have conducted in the natural habitats of each of the 


diverged significantly from the parental species for parental species (Lexer et al.. 2003a; Gross et al.. 


many traits associated with these habitats (Schwarz- 2004: Ludwig et al. 2004) Thus. transgressive 
bach et al., 2001; Rosenthal et al., 2002) An Segregation, which is commonly observed in segre- 


— 


important question is whether hybridization could gating hybrid populations (Rieseberg et al., 19994), 
provide the phenotypic variation necessary for provides a simple mechanism for the ecological 
colonization of these habitats. divergence of hybrid lineages. 

We addressed this question by comparing pheno- 
typic variation in the three ancient hybrid species, Generics or TRANSCRESSIVE SEGREGATION 
synthetic hybrids (a BCs population of Helianthus 
annuus X H. petiolaris), and both parental species So how are extreme phenotypes generated in hybrid 


grown in a Common ereenhouse environme nt (R 


- populations? Numerous possibilities have been sug- 
senthal et al., 2002; Rieseberg et al., 2003). A total of gested (deVicente & Tanksley, 1993; Rieseberg et al., 


40 morphological, life history. and ecophysiological — 19992), including an elevated mutation rate, reduced 
trails were measured. developmental stability, non-additive effects of alleles 

As expected, given their divergent habitat prefer- at the same romana r different. (epistasis) 
ences, the ancient hybrid species differed significant- loci, unmasking of rare recessive alleles, and the 
ly from the parental species for many of the measured combination of alleles with effects in the same 


traits: 20 for Helianthus anomalus (8 intermediate, 12 direction from both parental species 1 


z 


extreme). 14 for H. deserticola (4 intermediate, 10 gene action: Table 1). 


extreme), and 24 for H. paradoxus (8 intermediate, 16 We employed quantitative trait locus (QTL) 
extreme). Intermediate traits are easily accounted for methods to determine the mode of gene action 
by hybridization because the entire phenotypic underlying transgressive segregation in Helianthus 
continuum between the parental species is found in (Rieseberg et al., 2003). Briefly, the same BCs 
segregating hybrids (Rieseberg et al., 2003). population employed for phenotyping in the previous 


Hybridization could account for most of the extreme section was genotyped for 96 molecular markers, and 

traits as well. Indeed, for 28 of the 40 traits, BC 185 QTLs were detected for the 40 traits. Remarkably, 
— X ~ 

plants had phenotypic values that significantly 39% of the QTLs had effects in the opposite « 


irection 
exceeded those of the parental species, a phenomenon of species differences and 34 of 40 traits had at least 
known as transgressive segregation (Grant, 1975). one opposing QTL. That is, for these QTLs, the V. 
Overall, 10 of 12 (83.3%). 10 of 10 (100%), and 11 of annuus allele produced a more H. petiolaris-like 
16 (68.8%) extreme traits in Helianthus anomalus, H. phenotype, whereas the H. petiolaris allele produce 


deserticola, and H. paradoxus, respectively, were a phenotype that was in the direction of H. annuus. 


within the range of the BC population. The handful This is exactly the kind of genetic architecture that is 
of trail values exceeding those in the BC» population conducive for transgressive segregation under the 
might have been transgressive in a reciprocal back- complementary gene action model. Hybrid individuals 
cross population or have arisen through mutational that combine all of the “plus” or all of the “minus” 


divergence rather than hybridization. In either case. 


it QTLs from both parents for a given trait will have 
is clear that the majority of extreme traits observed in transgressive phenotypes 


— 


Table 1). Although com- 
the ancient hybrid species could have been generated plementary gene action explained most of the trans- 
through ni N These results have now been gressive phenotypes, epistatic interactions were 


replicated. by similar common garden experiments detected for 18 traits as well. These results are 


Volume 93, Number 1 
2006 


Rieseberg 43 
Hybrid Speciation 


similar to those obtained for cultivated plants: 
complementary gene action is most frequent, but 
jistasis and overdominance do sometimes contribute 


(deVicente & 1993; Rieseberg et al.. 
19992). 


e 


Tanksley, 


GENETIC CORRELATIONS 


Transgressive segregation via complementary gene 
action provides a means by which extreme phenotypes 
may be generated for individual traits. For a hybrid 


lineage to successfully colonize a new habitat, 
however. all of the trait differences must be combined 
into a single individual which may be 
difficult 
QTLs underlying key traits. Thus, genetic correlations 


or genotype, 
because of linkage and/or pleiotropy for 
may limit ecological divergence through hybridization. 

To determine the role of genetic correlations in the 
ecological divergence of the hybrid sunflower species, 
we asked whether closely linked or pleiotropic QTLs 
in the BCs population described above have effects 
that are in the same direction with respect to the 
2003). 


These positive correlations would greatly facilitate 
| 8 ) 


hybrid species phenotype (Rieseberg et al., 
j I l Yl 8 


ecological divergence and hybrid speciation. 
We found linked QTLs were 


positively correlated in direction of effect (Rieseberg 


closely indeed 


et al., 2003: Fig. 1). The pe coefficient of association 
ranged from 0.32 (N = 256; P < 0.001) for 
Helianthus paradoxus, to 0.47 (N = 256: P < 


0.001) for H. anomalus, to 0.53 (N = 256: P 
0.001) for H 


deserticola. Thus genetic correlations 


likely facilitated rather than impeded the origins of 


the three species, particularly for H. anomalus and H. 


deserticola. However. our results also imply that 
genetic correlations may limit the number of ecolog- 
ically relevant multi-trait phenotypes, perhaps ex- 
plaining why only three new hybrid species have 


originated from this cross. 


HYBRID SPECIES GENOMIC COMPOSITION AND 
PREDICTED PHENOTYPES 


Although studies of synthetic hybrids demonstrate 
that most of the phenotypic differences associated 
with each of the hybrid species could have arisen 
t 
what actually happened. It could be, for example, that 


— 


rough hybridization, they fail to prove that this is 


the differences arose as a consequence of mutational 
divergence and that hybridization was incidental to 
phenotypic evolution. To distinguish between these 
two hypotheses, we compared the genomic composi- 
tion of the three ancient hybrid species with 
predictions from the QTL analyses of the BC» 


population of Helianthus annuus X H. petiolaris 


2003). That is. we asked whether 
of QTL 


alleles for producing their phenotypes. If hybridiza- 


(Rieseberg et al., 


the hybrid species had the predicted. se 


tion played a key role in phenotypic divergence. then 


significant correlation should be found between 


predicted and actual genomic composition. 

Genomic composition of the ancient hybrid species 
did accord closely with predictions from the QTL 
2003: Fig. 2). The $ 
coefficient of association ranges from 0.56 (N — 150: 
P < 0.001) for Helianthus paradoxus, to 0.58 (N = 
193; P < 0.001) for H. anomalus, to 0.65 (N = 94; 
P < 0.001) for H. deserticola (Rieseberg et al., 2003). 
hybridization does appear to be largely re- 


analyses. (Rieseberg et al.. 


Thus. 
sponsible for the phenotypic divergence of the three 
hybrid species. 

Earlier in this paper, I reviewed evidence suggest- 
ing that fertility selection plays a major role in 
shaping hybrid genomic composition (Rieseberg et al., 
1996), whereas the QTL comparisons described above 
(presumably | ecological) 


suggest that phenotypic 


selection must be important as well (Rieseberg el 


al., 2003). What is the relative importance of these 
two modes of selection? A preliminary comparison of 
the two data sets indicates that in many instances 
there is concordance between the predicted genomic 
OTL 


analysis of phenotypic differences, perhaps implying 


composition from fertility selection and the 
overlap between hybrid incompatibility genes and 
those responsible for phenotypic differences. Also. the 
combined data set more completely accounts for 


hybrid genomic composition than either individual 


data set does alone (Rieseberg. unpublished). 


PHYLOGEOGRAPHY 


Comparisons of the genomic composition of the 
ancient hybrid species with synthetic hybrid lineages 
and predictions from QTL studies (above) suggest that 
homoploid hybrid speciation may be repeatable. Thus, 
we asked whether the ancient hybrid species had 
This was accom- 
the 


arisen. multiple times in nature. 


lished by surveying natural populations of 
| \ ying po] 


parental species and their three hybrid derivatives 
for variation at chloroplast DNA and nuclear micro- 
satellite loci (Schwarzbach & Rieseberg, 2002: Welch 
& Rieseberg, 2002b; Gross et al., 2003). We also 
assessed crossability and interfertility among popula- 
tions of Helianthus anomalus and H. deserticola. In H. 
anomalus, the crossing studies were complemented by 
meiotic analyses of inter-populational hybrids to 


determine the chromosomal basis of variation in 
fertility. 
For a single origin, all populations of a given hybrid 


species are expected to have the same chloroplast 


Annals of the 
Missouri Botanical Garden 


— 


DNA haplotype, to form a monophyletic lineage based 


on nuclear microsatellite loci; and to be highly 
interfertile (ite have a single karyotype) when 
crossed. For multiple origins. different populations 


of a species may vary in chloroplast DNA haplotype 
and the species is unlikely to be monophyletic with 
respect to nuclear microsatellite loci. Also, significant 


variation fertility due to karyotypic variability is 
Note that 
multiple chloroplast DNA haplotypes and polyphyly 


for 


predicted. for inter-populational crosses. 


nuclear microsatellite loci can also result from 
a single origin followed by introgression with one or 
both parental species. However, substantial variation 
in inter-populational crossability and interfertility or 
karyotype is unlikely in this latter scenario (Schwarz- 
bach & Rieseberg. 2002). 

Molecular 
suggest al least three different origins for Helianthus 
2002). The 


species has three main chloroplast DNA haplotypes. 


and crossability/interfertility evidence 


anomalus (Schwarzbach & Rieseberg. 
one apparently derived from M. annuus and the other 
two from H. petiolaris. Likewise, the species has three 
fertility groups that are completely correlated with the 
distribution. of chloroplast DNA haplot vpes. Meiotic 
analyses indicate the fertility groups differ by one or 
two reciprocal translocations, as would he predicted 
for multiple origins. However, there was no apparent 
micro- 


signature of multiple origins in the nuclear 


satellite data, perhaps due to gene flow among the 
independently derived lineages subsequent to hybrid 
speciation. 


Helianthus 


deserticola were also most consistent with multiple 


Patterns of molecular variation in 
origins (Gross el al., 
different 


2003). The species displayed four 
chloroplast DNA haplotypes (one from M. 
annuus and three from H. petiolaris), polyphyly at 
microsatellite loci, and considerable inter-population- 
little 


however. 


al crossability and interfertility. There was 


congruence between the three data sets. 
making it more difficult to rule out the possibility of 
a single hybrid speciation event followed by in- 
trogression with populations of the parental species. 
Helianthus and H. 
H. paradoxus is derived from a single 
hybrid speciation event (Welch & Rieseberg. 20020). 


All surveyed populations of the species share a single 


In contrast to anomalus 


deserticola, 


chloroplast DNA haplotype (derived from H. annuus) 


and form a single clade with 99.8% bootstrap support 


ased on 17 microsatellite loci. 

The apparent numbers of origins of the three hybrid 
the the 
morphological and QTL analyses. 12 extreme 


species correlates well with results from 


Ten of 
traits in Helianthus anomalus and all 10 extreme traits 
H. deserticola could be re-created by hybridizing 


populations of the parental species. However, only 11 


X 16 traits in VV. 


replicated. in synthetic 


extreme 


paradoxus could be 
hybrids, implying that the 
creation of the H. paradoxus phenotype may be a rare 
event. & similar pattern was observed for genetic 
correlations. Tightly linked QTLs mostly had effects 
in the same direction with respect to the phenotypes of 
the apparently multiply derived H. anomalus ( 

0.47) and H. deserticola (9 = 0.53). In 
genetic correlations make it difficult to re-create the 
multi-trait phenotype of H. paradoxus ($ = 0.32). 


contras 


which appears to have arisen only once. 
ECOLOGICAL SELECTION 


Theoretical studies indicate that strong ecological 
selection is a successful 


1995; Buerkle 


Also. our studies of phenotypic differentiation 


prerequisite for hybrid 


speciation (McCarthy et al., al., 


No 


have assumed that ecological selection is the un- 


derlying force driving divergence. However, it may be 
that some of the phenotypic differences result from 


neutral processes or are a byproduct of selection on 


correlated traits. Correlated selection is likely to be 


particularly important in hybrid speciation, because 
the unit of selection is the chromosomal segment 
rather than individual gene or mutation. 

We have employed both comparative and experi- 
traits likely 


been under selection during hybrid speciation. The 


mental approaches to identify have 


comparalive approach focuses on traits in the hybrid 


species that are common to other taxa sharing its 
habitat. For example, the sand dune endemic. 


Helianthus anomalus, has many traits associated with 
dune habitats such as very large, cylindrical seeds, 
and succulent leaves 
Rosenthal et al.. 2002). 
Large seeds are thought to represent an adaptation for 


rapid early root growth. 


(Schwarzbach et al.. 2001: 


burial avoidance, whereas a round 


^ 


cylindrical 
shape may prevent seeds from being blown away from 
the dune habitat. Large seeds also likely contribute to 
rapid early root growth, which enables seedlings to tap 
into water reserves deeper in the sand dunes. Finally, 
succulent leaves may serve to reduce water loss and 
(Danin. 199]: 


Helianthus deserticola has many of the classic 


abrasion. by 
1996). 


features of a desert annual, including rapid flowering, 


blowing sand Bowers, 


reduced height at maturity. and small, 
al., 2002). 


rapid. reproduction following heavy seasonal rain, 


narrow leaves 
(Rosenthal The first two traits ensure 
whereas small narrow leaves decrease water loss and 
1987: ( 


3). Helianthus paradoxus shares several traits with 


avoid fatal overheating (Chapin et al.. 
190€ 


other 


ribson, 


sall-loving or halophytic species. including 
increased leaf sueculence and an efficient mechanism 


for reducing mineral ion uptake (Welch & Rieseberg. 


Volume 93, Number 1 


Rieseberg 


2006 Hybrid Speciation 
2002a; Lexer et al., 2003b). Both reduce the toxic observation that QTLs underlying both mineral ion 


effects of sodium and other mineral ions. 

Our experimental studies involved the transplanta- 
tion of reciprocal BC synthetic hybrids into the 
habitats of the three ancient hybrid species. We then 
measured the strength of selection on individual traits 


for Helianthus paradoxus) the QTLs underlying 


— 


and 
them. For H. anomalus, we focused on leaf ecophysi- 


ological traits, phenology, and vegetative. biomass. 
with fitness estimated as reproductive biomass 


Although we had predicted 


(Ludwig et al.. 2004). 
that selection in mile synthetic hybrids would mostly be 
in the direction of the H. anomalus phenotype. this 
was not necessarily the case. Only for leaf succulence 
in the BCoann hybrids (backerosses toward H. annuus) 
use efficiency in the BCopet. hybrids 
(backerosses toward H. 


toward the H. anomalus phenotype. For most traits. 


and water 


petiolaris) was selection 


the direction of selection was dependent on the 
genetic background, the 
multiple adaptive peaks for Helianthus in the dune 


O 
— 


suggesting existence « 


habitat. Although these results seem inconsistent with 
the multiple origins of the species (Schwarzbach & 
Rieseberg, 2002). it 
other 


may be that fertility selection or 


selection on correlated traits drive early 


generation hybrids toward just one of these peaks. 


Results from the selection analyses in the 
Helianthus deserticola environment were equally 
difficult to interpret (Gross et al., 2004). For two of 


the key traits, leaf area and flowering date. selection 
differentials and. gradients differed in sign in both of 
the BCs populations. Thus, while it is clearly feasible 
for directional selection to produce the H. deserticola 
phenotype for these traits, it is not possible to predict 
the phenotypic outcome of such selection from the 
key 


synthetic 


the third trail, stem 


the 
but in the opposite direction of the H. 


present experiment. For 


diameter, selection in hybrids was 
consistent, 
deserticola phenotype. This result is puzzling and may 
possibly relate to the experimental design, in which 
seedlings were propagated at Indiana University and 
then transported to Utah for transplantation. In such 
weak and etiolated seedlings, it may be that large 
stems would be favored, whereas narrow stems might 
be favored seedlings propagated under natural. 
high light conditions. 

Only in Helianthus paradoxus habitat was the 
direction of selection in the experimental hybrids 
the 


phenotype (Lexer et al., 


direction of the H. paradoxus 
2003b). Positive directional 


and Ca 


consistently in 
selection was detected for leaf succulence 
uptake, two traits that are known to mitigate salt stress 
in plants. In contrast, the uptake of Na and other toxic 
mineral ions was strongly negatively selected in the 
Of significance was the 


salt marsh habitat. greater 


uptake traits and survivorship in the salt marsh had 
effects in opposing directions (Lexer et al., 2003a). As 
architecture 


discussed earlier, this kind of genetic 


underlies transgressive segregation. and provides 
a simple explanation for how the salt marsh habitat 
of H. paradoxus might have been originally colonized 
by hybrids, 

The QTL data also allowed us to test a fundamental 
requirement for homoploid hybrid speciation, Homo- 
ploid hybrid speciation represents a kind of sympatric 
or parapatric speciation, and models of speciation 
with gene flow indicate the strength of selection at 
individual loci must be greater than the migration rate 
Maynard Smith, 1966). calculated 
selection coefficients for mineral ion QTLs in the salt 
marsh habitat (Lexer et al., 2003a). The selection 
(—0.08 to 40.13) than any 
conceivable migration rate for sunflowers, 


it would have been feasible for Helianthus paradoxus 


Therefore, we 


coefficients were larger 
indicating 
to arise in sympatry or parapalry with its parental 


species. 
CANDIDATE GENES FOR ECOLOGICAL DIVERGENCE 


Finally, we have assayed sequence polymorphisms 
for salt tolerance candidate genes in these same BC» 
populations to determine whether any of the candidate 
eenes map to QTLs of interest, as well as to identify 
additional genomic regions associated with sall 
2004). 
candidates were identified from an expressed se- 
quence tag (EST) library for Helianthus paradoxus 


based on homology to genes with known function. One 


tolerance (Lexer et al. The salt tolerance 


of the 11 Ca-dependent protein. kinase 


(CDPK), maps coincident with a previously identified 


genes, a 
QTL for mineral ion uptake. Two additional genes (an 
ER-type calcium ATPase and a transcriptional regu- 
lator) also exhibit a significant fitness effect in the 


wild. Of course, these studies are correlational only 


and function will have be verified by transgenic 


( omple nlé ntation or RN, At. Nonetheless, they indicate 


we soon may be able to examine the role of individual 


n ecological divergence and 


genes spec iation in 


sunflowers. 
CONCLUSIONS AND. FUTURE DIRECTIONS 
Over the past 15 vears, my lab has studied how new 
hybrid species may arise without a change in chromo- 
Molecular doc umentation of the hybrid origin of three 


the same pair of 


1990b: Rieseberg, 


some number. Our main contributions include: 
(1) 

distinct from 
parental species (Rieseberg et al., 
19 


991). 


46 Annals of the 
Missouri Botanical Garden 
(2) Development of the first genetic map for a homoploid tion when the species arose in nature more than 
hybrid species (Rieseberg et al., 1993). 63.000 years ago. Preliminary evidence suggests we 
2 Demonstration that homoploid hybrid speciation in E : : . 
P | can. Microsatellite markers flanking the three most 
wild sunflowers has been accompanied by rapid m . . 
POTU evolution m predic ied hy nica 0 5 „ strongly selected QTLs in the Helianthus paradoxus 
(Rieseberg et al.. 1005: Lai et al., 2005). habitat had significantly reduced variability as 
(4) 0 re-creation of homoploid hybrid spe- compared. with microsatellite loci from unselected 
cles in nflower and demonstration of genomic regions, vet there was no difference in variability 
congruence of synthetic and ancient hybrid species | lechel tl oa anu al 
evels between lese groups OCI IN lo parenta 
(Rieseberg et al. 1996). This demonstrated the j . p : ps . 
repeatability of homoploid hybrid. speciation and Species (C. Edelist et al., unpublished). 
the importance of fertility selection in shaping hybrid Second, my lab has developed a 3000-gene abiotic 
- genomic composition. stress microarray for sunflower (Lai et al.. in press). 
(5) De Mein urs ol the ra id tempo of hybrid speciation : . " 
i l ! | We wish to ask whether gene expression differences 
Das ed on the larg SIZOS ol wental e hromosomal bloc ks i AS . 
in a hybrid Ma en ae (Ungerer et al., 1991 are induced by hybridization and whether these 
(6) Computer simulation of homoploid hybrid speciation, induced. differences contribute to ecological diver- 
which verified verbal theory and demonstrated the gence and hybrid speciation. 
evolutionary conditions favorable to hybrid origin 


from those 
evolution of the hybrid species (Buerkle et al., 2000). 
(7) Re- 


spectes b. experimental hybridization of conte mpo- 


differ 


creation of phenotypes. of homoploid hybrid 


rary parental species api and transgressive 


)03b: Riese de re et i 
5 nthal et al.. 2005 
(8) De monstration thal coral mentary gene action is qe 
cause of transgressive segregation in exper- 


imental sunflower hybrids (Lexer et al., 20030; 
Rieseberg et al.. 2003). 
(9) Discovery that genetic correlations facilitate. rather 


than impede the origin of. the hybrid sunflower 
species (Riesebere et al.. 2003). 

De monstration that the genomic ee of the 
alysis of 1 55 nolypic traits. implying 
an important nie for ecological selection in shaping 


(11) 


— 
= 


arlier studies of the re pe sath jd hybrid specia- 
f 
Wek 


Lion Falsa & Rieseberg, 2002: Welch 
20 2003 


Rieseber 02b: Gross et al.. 2003). 


Demonstration that many of the transgressive traits 


— 


discovered in experimental erosses are under selec- 
tion in natural hybrid habitats 5 ex 2003a: 
. 2004: Gros 
Demonstration that selection coe illis ients for QTLs 
Helianthus 


er et al., 

Ludwig eta ss el al., 2004). 

(13) 
underlying ecological divergence in 
agin are large e nougli to account for the origin 

of the species in parapatry with its parental Species 
(Lexer et al.. 2003b) 

(14) Identification of seve SN Pu A candidate genes 

dive rgence 


2004). 


there is much left to do. 


for ec 010 elc al 1 He lianthus paradoxus 


P 


Of course, Previously. we 


attempted to replicate the early stages of hybrid 


speciation by planting segregating hybrids into the 


habitat of the ancient hybrid species and identifying 


traits and QTLs that were under selection. We are 
currently extending these experiments by asking 


whether we can detect the molecular 


selection during hybrid speciation and if the same 


chromosomal segments that were under selection i 


the experimental populations were also under selec- 


signature of 


Third. we wish to use the hybrid species system to 
address the fundamental question of whether adaptive 
divergence within a species is limited by mutation or by 


each 


eene flow from the center of species range 
(Kirkpatrick & Barton, 1997). That is; why did the 


parental species not colonize the extreme habitats of 
the ancient hybrid species? Put another way, was il new 
hybrid gene combinations or reproductive isolation that 
allowed the hybrid species to colonize new habitats? 
Finally. we will continue our work to identify and 
characterize the genes underlying ecological divergence 
in this system. This will involve a combination of EST 
(> 70,000 ESTs 


searches for genes with the signature o 


— 


sequencing already sequenced 


xy 


positive selec- 
tion (> 30 identified), candidate gene mapping (close to 
300 mapped). 


testing using transgenic complementation and RNAi. 
c o c 


microarray analyses. and functional 
Although this review has focused almost exclusive- 


ly on hybrid. sunflower species. there are. broader 
issues that need to be addressed as well. In particular. 
we need know how frequently. homoploid hybrid 
speciation occurs in nature, which will require better 
tools for detecting it in phylogenetic trees. Fortunate- 
Ne "WV 
reconstruction. are under development (reviewed i 
Linder & 2004). and 


sequence data from numerous nuclear genes should 


methods of network 


lv. help is on the way. 


Rieseberg. the availability E 


provide far more power for detecting reticulate 


histories than was possible in the past. This power 
could be further enhanced if clusters of linked genes 
were sequenced. Under the assumption of hybrid 
specialion, separate phylogenetic reconstructions of 
individual genes that are tightly linked should have 
the topology of only one side of the hybridization, 
These reconstructions would be topologically incon- 
upon clusters of 


eruent with reconstructions based 


genes from the other side of the hybridization. 


We also need to know whether many of the 
discoveries made in the sunflower system can be 
generalized to other homoploid hybrid species. How 


Volume 93, Number 1 
2006 


Rieseberg 47 
Hybrid Speciation 


are most new hybrid species reproductively isolated 
from their parental species? Is ecological divergence 
the rule rather than the exception? Are the habitats of 
the new hybrid species typically extreme relative to 
their parental species, or are they more accurately 
described as intermediate or recombinant? What is 
the 
hybrids? How repeatable is hybrid speciation in other 


genetic basis of the ecological divergence in 


lineages? Are homoploid hybrid species often multi- 
ply derived? What are the characteristics of taxa that 
make them prone to homoploid hybrid speciation? 
Hopefully, our sunflower work will provide an 
experimental and conceptual guide for studies of 


homoploid hybrid species in other organismal groups. 


Literature Cited 


Anderson, E. 1948. Hybridization of the habitat. Evolution 
2: 1-9. 
1949. Introgressive Hybridization. John Wiley. 


New York. 
Arnold, M. L. 

Oxford Univ. 
Asch, D. L. 1993, Common s 

The pathway toward domestication, Pp. 1— 
i Annual Meeting of the 
Archaeology. Society for 


1997. Natural Hybridization and Evolution. 
Press, Oxforc 
unflower (Heltantl nnuus l): 


15 in Proceed- 
Society 
American 


58th 
for American 
Vre hae 1 St. 

Baird, S. J. E. 1995. A simulation study of multilocus clines. 
E cie iid pelt )15. 


Louis. 


Barsness, 1985. Head, Hides and Horns: The C ue al 
Buffalo 11 Te xas Christian Univ. Press, Fort Worth. 
Barton, N. H. The role of hybridisation in evolution. 

Molec. E D es 551-56 
& C. M. Hewitt. 1985. Analysis of hybrid zones. 
Ann. Re E e. Syst. 16: 113-148. 
SONOS. 1 E 996. Seedling emergence on Sonoran desert 


m Environ. 33: 5 7 
Buela. C. AR J. M 
2000. The cea of 
Heredity 84: 441—45 
Wolf & L. H. CN 2 
extinc © of species through en Pp. 117-1 
in C. A. Brigham & M. W. Schwartz (editors), | 1 
1 in Plants. Springer Verlag, Berlin. 
Burke, J. M., S. Tang, S. J. Knapp & 1. H. 
j domestication. 


OTTIS nussen & H. 


Riesebere, capis oid b 
8 l 


speciation. 
D . The origin Jn 


2002. 
Genetics 


Rieseberg. 

Genetic analysis of sunflower 
161: E 

— Lai, M. Salmaso, T. Nakazato, S. Tang, A. 
— € S. J. Knapp & L. H. 2004. 


Comparative mapping and rapid UL evolution in 


Riesebere. 


the genus i. lianthus. D Ws 167: 44€ m 57. 
Chapin, F. S., A. J. Bloom, C. B. Field & R. H. Waring. 1987. 
Plant responses to mu in 5 factors. Bio- 
37: 49-51. 
Charlesworth, De 1995. Evolution under the microscope. 
urr. Biol. 5: 835-836. 


Danin, A. 1991. Plant X ONT in desert dunes. J. Arid 
Environm. 21: 193 r 

deVicente, M. C. & 8 Tanksley. 1993. QTL analysis of 
transgressive segregation in an interspecific tomato cross. 
Genetics 134: 585-5 


Dary, D. 1974. The ultus Book: The Full Saga of the 
American Animal. Sage Books, Chicago. 


. Rieseberg & D. 


southern 


Arias. 1992. Chloroplast 


California 


Dorado, 0. 
D! x introgression an sunflowers. 


Evolution 46: 566— 


Fisher, R. A. 1953. A le sr theory of junctions in inbreeding. 
Heredity 8: 187-1 
Gibson, A. C. 1998. Pares lic organs of desert plants. 


o 


Gottlieb, L. D. e suni of confidence in the analysis of 


boe. de Ed Ann. Missouri Bot. Gard. 59: 
135440. 
Grant, V. 1958. 


The regulation of recombination i in Bn 
23: 337 


Gold s Harbor Symp. Quant. Biol. 23: 337-306: 
966 f 


— — Selection. for and eniin in the 


vigor 


progeny of a highly sterile species hybrid. in Gilt 
Genetics 53: 757-775. 
1966b. The origin of a new species of Gilia in 
a hybridization expe riment. Genetics 54: 1189-1199. 
. 1975. Genetics of Flowering Plants. Columbia Univ. 
Press. New York. 
1981. Plant Speciation. Columbia Univ. Press, New 
York. 
Gross, p. L. & L. H. Rieseberg. 2005. The ecological 


> 
——, A. Schwarzbach & L. Rieseberg. 2003. 
Origin(s) of the diploid hybrid species Helianthus 

deserticola (Asterace Amer. J. Bot. 90: 1708-1710 
C. Kane, Eten . Ludwig, D. M. Rosenthal, 


onovan & E 2004. Reconstructing 
iita origin(s) of Helianthus deserticola: Survivorship and 


floor. Amer. Naturalist 164: 


E Tg. 

selection on the desert 
145-156. 

Heiser, €. B. 
species Helianthus annuus and H. petiolaris. | 
249—262. 

—— ———. 10949. Study in the 

pecies ee peus annuus and H. bolanderi. 

p ubl. ot. 23: 196. 


" bridization in the 


Hybridization between the sunflower 

Evolution I: 

evolution of the sunflower 

Univ. Calif. 

— due annual sunflowers: 

dh lianthus annuus X H. argophyllus. Amer. Naturalist 85: 
— (i 


6 2. 


1951b. Hybridization in the annual sunflowers: 
Melanii annuus X H. debilis. var. cucumerifolius. 
Evolution 5: 42-51. 

———, D. M. Smith, S. Clevenger € W. Martin. 1969. 
The North eran sunt Un Mem. Torrey 
ot. Club -218. 

Kerner, A. ae The Natural History of Plants. 
Vols. 1, 2. Blackie, London. 

Kirkpatrick, M. & N. H. Barton. vic Evolution of 

range. P Naturalist 150: 1-23 


er. : 
Salmaso, J. M. 13 irke, S. Tang. S. J. 
II. 19 berg. 2005. Extensive c bus 


PS and the evolution of re barriers in 
hybrid sunflower species. Genetics 171: 291-303. 

3. Gross, i; 1 & L. H. Rieseberg. In 
press. eris lecting dil fle rential gene expression in homoploid 
hybrid sunflower species using microarrays. Molec. Eco 

Lexer, C., M. Welch, J. | eR ¿€ L. H. het a 2003a 
Natural selection for sali tolerance QTL 1 


hybrids: Implications for the origin of Heliantus para- 
12: 


doxus, a homoploid hybrid species. Molec. Ecol. 
1225-1235a. 

; . Raymond & L. H. Rieseberg. 2003b. 
The origins 2 ec Mrd al me rgence in Helianthus 
paradoxus (Asteraceae): Selection on transgressive 


acters in a novel hybrid habitat. E EIE 51: 1989- 2000. 


Annals of the 
Missouri Botanical Garden 


Lar & L. II. Rieseberg. 2004. Candidate gene 
polymorphisms associated with salt tolerance in wik 
sunflower hybrids: Implications for the origin of He- 


lianthus paradoxus, a diploid hybrid species. New Phytol. 
161: 225-233. 

Linder, C. R. & L. H. 2004. 
evolutionary network of plant species: The need to account 
for population geneties. Amer. J. Bot. 91: 1700-1708 

Rose uu J. A. QU N. 


Rieseberg. Reconstructing the 


Ludwig. r D. 


Gross, È exer, Dudley, L. H. Rieseberg & 
Ds 01. 1 eee traits in 


a hybrid 1 0 5 species and early generation hybrids 
in a desert ode habitat. Evolution 58: 2682-2092, 
Maynard Smi J. 1900. Sympatrie speciation. Amer. 
PF i ay 
MeCarthy, A. Asmussen & W. Anderson. 1995, 


V theoretic 11 assessment of rec 5 speciation. 

Hori 7 7 E 502—509. 

. 1999, Reinforcement and other consequences 

of —— He 1 5 83: 503-508. 
: 19 


Rieseberg, L. Hoepli reticulate evolution in 


pe lianthus ( ste raceae): o ence from ribosomal genes. 
Amer. J. Bar 7 : 1218-12 
199 105 origins E plant species. Annual Rev. 
Ecol. d 28: 359—389. 


2000. Crossing relationships among ancient and 
Evolution. 54: 


experime ental sunflower hybrid. lineages. 
859-8 
— ———, D. E. Soltis & J. D. Palmer. 1988. A molecular 
amination of introgre e Mead 


Helianthus 
Evolution 42: 238. 
N K. 1000 


annuus and H. bolanderi. 
s. “Di : 
He 1 annuus Lexanus has chloroplast DNA aid 


Beckstrom-Sternberg 


^ 


nuclear ribosomal RNA genes of He ee debilis ssp. 
cue PUES, Proc. Natl, ad. Sci. U.S.A. 87: 593-597, 
— ——, A. Ai Å Arias. 199 la. P hylogenetic 


enema. inferences from chloroplast. DNA and 


variation in Helianthus sect. Helianthus Syst. Bot. 
76 


oad 
isozyme 
16: 50-76 

Carter & S. Zona. 1990b. 


: Molecular tests of the 
listiothi sized hybrid origin of "ge 


9 7 = lianthus 
5 Asteraceae). Evolution 44: )8— 

. Choi & D. Ham. 1991b. a ni i c dum ismi 
versus nuclear gene flow in Helianthus. J. He 
189—493 


— 


spectes 
redily 82 


Chan & C. 


Spore. 1993, Genomic map 
of a E mh species, He 293. 
A 5 


redity 70: 285-2 


esrochers & S. J. 1995a. Interspeci 
pollen metn as a reproductive: barrier 


between 
` Helianthus (Asteraceae). An J 


sympatric species 
Sot. 82: 515-519, 
1005. Hybrid 


reorganization in 


Van Fossen & A. Desrochers. 
speciation accompanied by genomic 


Nature 375: 313-3106, 


wild sunflo 


WETS, 
—— sinervo, C. R. Linder, M. Ungerer & D. M. Arias. 
1090. Role of gene interactic in hybrid. speciation: 


evidence from ancient and experimental hybrids. Science 


272: 741—745. 


Desrochers. 1998. Patterns of mat- 


— ———, S. Baird & A. 


ing in wild. sunflower. hybrid zones. Evolution 52: 
mm 

— M Archer & R. K. Wayne. 19993, Transgressive 
cd o adaptation, and speciation. Heredity 8: 
303—372 

—————, J. Whitton & K. Gardner. 1999b. Hybrid zones 
and the genetic architecture of a barrier to gene flow 


between Iwọ wild sunflower species. Genetics 152: 
eer 


Rosenthal. Z. Lai. 


Durphy. A. E. Schwarz- 


O. Raymond, D. 
mon T. Nakazato, J. L. 
bach, . Donovan & C. Lexer. 2003. Major ecological 
transitions in ae calles facilitated by hybridiza- 
1211- 1216. 

.. A. Donovan. O. 
T differen- 
their 


Schwarzbach, 

2002. 
ancie foal hybrid 

parental s JPL Sci. 162: 

x dH; Rieseberg & L. A. o 2005. Recreating 

complex multi-trait phenotypes 


~~ 
J: 

ines 

p 


tiation taxa and 
3 


3871 98. 


y 


ancient hybrid. species 
from 


carly generation (inm hybrids: 1 
sae a 26-4 
2. Likely multiple 


wee examples 
using wild sunflowers. Amer. Nat l. 

& Rieseberg. 2( 
origins de a diploid hybrid sunflower species. Molec. Ecol. 


L1: 1703-1717. 
ni 


Schwarzbach, A. E. EH. 


— ———, |. A. Donovan & L. H. Rieseberg. 2001. Trans- 
gressive character expression in a hybrid sunflower 
species. Amer. J. Hot 88: 270-277. 

Sokal i à . 1995. Biometry, Ard ed, W. H 


. Variation. and evolution in plants. 
Columbia U niv. Press. New Yor 


The hybrid origin of 


——. 105 microspecies in the 
ymus one complex. Cytologia Suppl. 36: 330—340. 
Te ie ton, X. R. 1981. Mechanisms of speciation——A 
a gene lic approac th. Annual Re >V. E col. Sysl 
12 Pu 
Ture Ji: ; . Barton N b : Coyne, 2001. Theory and 


ni E 16: 330-: 
Baird: J. 
Rapid isi speciation in 


Acad. Ser U.S.A. E 


ein 
1998 
Proc. Natl. 


E 8 Le I. 
wild sunflowers. 
1762. 


Ungerer, M. S. Rieseberg. 


1751-1 


Wang. N.-It, A. E. Szmidt & O. Savolainen. 2001. Genetic 
composition and 1595 hybrid 9 of a high 
mountain pine, Pinus e native to the Tibetan 
plateau. Ge “netics 159: 340. 

Welch, M. E. & L. H. Ri ee . Habitat divergence 


between a homestand hy brid Ee t species, Helianthus 


m vus (Asteraceae), and its progenitors. Amer. J. Bot. 
X): 


178. 


——— P . H; Rieseberg. 2002b. Patterns of genetic 
variation suggest a single, ancient origin for the diploid 
hybrid species Helianthus paradoxus. Evolution 50: 
212 62137. 
Wolf. D. E., N. Take bavashi & L. H. Rieseberg. 2001. 


Predicting the i da extinetion through hybridization. 
1053. 


: 1039— 


Conservalion Biol 


ADAPTIVE RADIATION AND 
EVOLUTION OF BREEDING 
SYSTEMS IN SCHIEDEA 
(CARYOPHYLLACEAE), AN 
ENDEMIC HAWAIIAN GENUS! 


Ann K. Pio Stephen G. Weller,? Warren L. 
Holly Nepokroeff,' and Theresa M. 


Wagner 


Culley’ 


ABSTRACT 


multi-disciplinary approach, 

f; that have pron 

Se hiedea is the fifth largest lineage in the native Hawaiian flora 
The 


elucidate the selective factors ted spec 


82 185 Most inter-island colonizations were from ol 


sufficient isolation to result in 


lation and shifts 


enus is monophyletic and shares a common ancestor 15 a clade 
der to younger islands, 
formation of new species 0 are single-i sl 


including phylogenetic analysis, population biology, and quantitative genetics, has helped to 
h 


in breeding systems in Schiedea (Caryophyllaceae). 


and the most diverse dus 'age with re iur ct to breeding systems. 
consisting of two arc r boreal-north te mperale 
and most movement between islands led to 


ind endemics rather than species with multi-island 


distributions. Closely related species pairs occurring on older islands tend to diff fer in habitat and are isolated ecologically on 


the same island, while s pecies pairs on younger islands tend to be in similar habitat on different islands. Speciation within this 
| 


lineage has bee 
(obligate aedes 
Dimorphic breeding systems appear 


The « 


obligate autogamy have both evolved ho 


Sex toc alon patte ms, and the evolution p yr in this 
E 


assoc iated with shifts in habitat, pollination system, and breeding system 
facultative uae mixed mating systems, and 1 (gynodioecy, 


. including evolution of selfing 


subdioecy, and dioecy). 


nn been derived inde spe nder A twice in Schiedea, and facultative M en and 


S 
da 
& 
un 
=< 
ds 
B 
e 
3 
Z 


and high selfing rates. Many morphologic T traits associated with allocation to male and female function 


are highly heut able. and genetic correlations in general do not appear to constrain the evolution of dimorphism in Schiedea. 


ords: al aplive radialic aulogamy, 


Schiedea. selfing. sexual d morphism 


breeding systems, 


Hawaii, inbreeding depression, resource allocation, 


Flowering plants reproduce with a remarkable 


diversity of breeding systems, even among closely 


related species, and the selective forces responsible 


for this diversity continue to be of great interest. 
Adaptive radiations in island systems have been 


particularly useful because the isolation of remote 


archipelagos has made it easier to identify selective 


actors and to discern evolutionary patterns (e.g., 


Wagner € Funk, 1995; Givnish «€ Sytsma, 1997 
Grant, 1998). The Hawaiian Islands have been of 


particular interest in the study of plant mating systems 
because of the high incidence of dioecy in the flora. 
Worldwide, the 
angiosperm floras is 


Ricklefs, 1995), 


incidence of 
4-609 


but the Hawaiian 


average dioecy in 


(Renner & 


Islands have the 


about 


highest incidence of dioecy of 
studied (14.796). followed closely by the New Zealand 
flora (12% 13%: S 19954). 


kai et al., 
Phylogenetic studies have elucidated the relation- 


any angiosperm flora 


axa in 


ships among some of the larger Hawaiian 
lineages (e.g.. Baldwin et al., 1991; Baldwin, 2003; 
Givnish et al., 1995; Wagner & Funk, 1995; Weller et 
al, 1995; Sakai et al, 1997b; Kim et al., 1998; 
Ballard & Sytsma, 2000; Baldwin, 2003; Nepokroeff 
et al., 2003, 2005: Lindqvist et al., 2003; Carlquist et 
al., 2003; Price & Wagner, 2004). Most 
Hawaiian lack a complete phylogeny, but 
because of the great isolation and small size of the 


1020 


2005a), it has been possible 


summary In 


groups 


native Hawaiian angiosperm flora (about 


species; Wagner et al., 


! We 


with the figures, al nd the me any unde rgraduate researc hers NS 


their contributions to studies 


thank the N: ational Tropical Botanical Garden for logistical support. This work was funded in part by the National Geographic 
76 


2 

00 

5 

co 
í 


Science Foundation 


“De partment of Ecology and Evolutionary Biology 
* Department of Botany, MRC-166, Smithsonian Institution 


, University 


, P.O. Box 37012 


thank Amy Dunbar-Wallis and Allen Andres for their invaluable help in the A. a greenhouse, Denise Mix for help 


n the Sakai-Weller lab at University of California-Irvine for 
of selfing rates, inbreeding Xu ssion, resource allocation, and quantitative genetics. We also 


1 
Smithsonian 


89-18360, 


the 


DEB 92-07724, 


Undergraduate 


DEB 98-15878), 


Research Opportunities Program/ 


Vagner's contribution was completed while he 


10 
Chair for Hawaiian Plant Studies at the National Tro opical Beal Garden. 


of California-Irvine. Irvi Pe 92697, 


D.C. 20013-7012 


ne, 
,W O 


Biology Department, University of South Dakota, Ve lion: South Dakon 5 706€ .l 


Department of Biological Sciences, MLO06, University of Cincinnati, Cincinnati, uon 4 


ANN. Missouni Bor. 


5221-0006. U.S.A. 


GARD. 93: 49-63. PuBLISHED ON 31 May 2006. 


Annals of the 
Missouri Botanical Garden 


to identify plant lineages descended from a common 


colonist and examine evolutionary patterns in the 
Hawaiian Islands with respect to speciation (Price & 
Wagner, 2004), 
et al., 2002), 

1995a, b). The ability 


flora examination of 


rarity and conservation status (Sakai 
and breeding systems (Sakai et al., 
to identify lineages in the extant 


allows traits correlated with 
breeding systems that are less confounded by 


than analyses at the 
(Sakai et al.. 


these 


phylogenetic relationships 


generic level 


Steiner, 1988). Based on studies of 


lineages, dioecious species in the Hawaiian Islands 


are descended from dimorphic colonists or from 


hermaphroditic colonists with in situ evolution of 


dioeey in the Hawaiian Islands. Ten percent of 


successful colonists had dimorphic breeding systems 
lineages with dimorphic species. 
both 


and gave rise 


These dimorphic lineages (with dimorphic 
colonists and descendents) account for over half of 
Bakers law that 


difficulties in obtaining mates will limit colonization 


the dimorphic species. suggests 


by dioecious species more than hermaphroditic 


species, but the success of dimorphic colonists as 
well as biogeographic patterns within the Hawaiian 
Islands suggest that obtaining mates did not severely 
limit colonization by dimorphic species (Sakai et al. 
1995a). The incidence of dimorphism is also high in 
the Hawaiian Islands because in at least 11 lineages 
dimorphism evolved. from hermaphroditic colonists, 
accounting for about one third of the current. di- 
morphic species. These lineages, with autochthonous 
have higher 


of dimorphism, tended to 


diversification rates 


more species per colonist) than 
either hermaphroditic lineages or dimorphic lineages 
(Sakai et al, 19954). 
indicates that breeding systems have evolved 
that promote 
outcrossing (e.g. dioecy), but ways that 


1995; Price & Wagner. 


\nalysis of lineages also 


in the 


Hawaiian Islands not only in ways 
also in 
promote selfing (Weller et al.. 
2004. Weller et al., 2005). 


While flora-level analyses of factors 


associated with 


breeding system evolution reveal. general patterns, 


selective factors promoting speciation and the evolu- 


tion of dimorphism may be most. evident. through 


studies of hermaphroditic lineages evolved 


dimorphism after colonization. of the Hawaiian 


Islands. We have focused on Schiedea Chamisso & 


Schlechtendal (Caryophyllaceae), a monophyletic lin- 


eage with a hermaphroditic ancestor in which 10 of 


the 34 species evolved dimorphism in the Hawaiian 


Schiedea is the fifth largest lineage in the 


Islands. 
native flora and the most diverse Hawaiian lineage 
with respect to breeding systems (Weller et al., 1995; 
005b). We have 


used a multi-disciplinary approach, including phylo- 


Sakai et al., 1997b, Wagner et al., 2 


genetic analysis, population biology, and quantitative 


genetics, to understand both the genetic potential for 
the evolution of breeding systems in these species and 
the selective factors that have promoted. speciation 
and shifts in these breeding systems. 
SPECIATION AND. PHYLOGENETIC ANALYSES OF SCHIEDEA 
The genus Schiedea, with 34 species and 35 taxa, 
includes deciduous perennials in coastal habitats. 
sprawling subshrubs in mesic forests, woody shrubs in 
dry, mesic, and wet forest, rainforest vines, and small 
subalpine subshrubs. Schiedea exhibits the greatest 
diversity in breeding systems of any native Hawaiian 
angiosperm genus (Weller et al., 1990; Weller & 
Sakai, 1990; Weller el al., 1995; Sakai et al., 1997b; 
Wagner et al., 2005b), including hermaphroditic 
species that are obligately autogamous, facultatively 


autogamous, and partially selfing with mixed mating 


systems, as well as dimorphic species that are 
evnodioecious (females and hermaphrodites in popu- 
lations). subdioecious (females, males, and a few 
hermaphrodites in populations). or dioecious (females 
and males in populations). In this section we review 
the changes in biogeography, habitats; and breeding 
systems that have occurred. during diversification of 
the lineage. Shifts in breeding systems are associated 
with shifts in habitat and changes in pollination 
biology in Schiedea, and we have used phylogenetic 
analysis and the comparative method to help infer 
causal factors in the evolution of breeding systems 
(Weller & Sakai, 1999). 

Karlier phylogenies of Schiedea based on morpho- 
logical characters, chloroplast DNA, and ribosomal 
DNA restriction-site analysis (Wagner 1995: 
Weller et al., 1995; Soltis et al., 1996; ls 


1997b) have been refined more recently using se- 


el al., 


Sakai el ¢ 


£5 


quence analysis of ITS and ETS genes and a revised 
and expanded morphological data (Nepokroeff et al., 
2005. Only 


cluded in the study so that combined analyses could 


unpublished). exlant species were in- 


be conducted. Maximum parsimony analyses were 
performed using unweighted parsimony for the 


morphology and nuclear (ITS and ETS regions) data, 
both separately and combined. Maximum likelihood 
searches on the molecular data utilized an iterative 
approach to evaluate models and optimize model 
parameters for an initial set of trees resulting from 
parsimony analysis, and analyses were then performed 
under the fully defined model parameters (Nepokroeff 
et al., 2005). PAUP* y. 4.0b10 (Swofford, 2001) was 
used for maximum parsimony and maximum likeli- 
hood analyses. Bayesian. analyses of morphological 
and molecular data were also conducted (Nepokroeff 


et al., 2005). The combined molecular and morpho- 


Volume 93, Number 1 
2006 


Sakai et al. 51 
Breeding Systems in Schiedea 


logical data sets resulted in 222 characters in- 
formative in parsimony analysis. Using equal weight- 
ing of characters, 24 most parsimonious trees resulted, 
with a length of 613. The ITS and ETS sequences 
yielded a single likelihood tree, and a single Bayesian 


consensus tree was also produced. These trees were 


— 


highly congruent in topology with each other and with 
the parsimony trees and are not shown here. Bio- 
geography, habitat, and breeding system diversity 
were optimized using MacClade 4.0 (Maddison & 
Maddison, 2000) onto one of the 24 most parsimonious 
trees showing the fewest unresolved polytomies, 
although several weakly supported nodes indicate 
that conclusions about the number of evolutionary 
transitions are tentative. 

Sequence analysis of ITS, ETS, matK, and trnL-F 
has provided a better understanding of the extant 
sister group to Schiedea. Based on these analyses. the 
extant sister group to Schiedea is strongly supported 
and consists of a lineage comprised of circumboreal- 
north temperate Honckenya peploides (L.) Ehrh. and 
Wilhelmsia physodes (Fischer ex Seringe) McNeill. 
a species with an Alaskan to northeastern Asian 
distribution (Nepokroeff et al., 2005). Analyses of the 
strict consensus tree of the 24 most parsimonious trees 
(Fig. 1) indicate that Schiedea is a strongly supported 
monophyletic lineage that arose from a single common 
ancestor of cireumboreal or Alaskan origin. Although 
much of the phylogeny is only weakly supported, 
several areas with moderate to strong support are 
present, including a number of well-supported species 
pairs. In the strict consensus tree, with the exception 
of gynodioecious Schiedea apokremnos St. John, all 
other dimorphic species occur in a single moderately 
Schiedea) that 


includes three hermaphroditic species. Within section 


well-supported clade (sect. also 
Schiedea, there is moderate support for a terminal 
subclade of several dimorphic species (S. salicaria 
Hillebrand, S. ligustrina Chamisso & Schlechtendal, 
S. adamantis St. John, S. kealiae Caum & Hosaka, and 
S. spergulina A. Gray). Unfortunately, this section, 
which is of greatest interest in the evolution of 
dimorphism, is also one of the most poorly resolved 
parts of the tree, perhaps because it has evolved more 


recently than other clades. 


BIOGEOGRAPHY AND PATTERNS OF SPECIATION 


About 10% of angiosperm colonists in the Hawai- 
ian flora have temperate affinities. including the 
ancestor of the lineage comprised of Schiedea (Fig. 2: 
Nepokroeff et 
islands have been forming over the Hawaiian hot spot 
for the past 85 Ma (Clague, 1996 


al.. 2005. unpublished). Although 


„ most colonization 


of the current high islands has occurred. within the 


past 5 Ma, because dispersal between islands was 
previously limited by an extended period when only 


small, low, widely spaced islands were formed (Price 


— 


& Clague, 2002). The current islands with species of 
Schiedea range from the older islands of Nihoa and 
Kaua'i to the youngest island of Hawaii (Fig. 3). In 
this analysis we consider Moloka'i. Lana'i. Maui, and 
Kaho'olawe as a single island (Maui Nui). because 
they were interconnected for more than 75% of their 
existence (Price & Elliott-Fisk, 2004). Species in the 
most basal clade of Schiedea (S. membranacea St. John 
and S. helleri Sherff) occur on the older island of 
Kaua‘i, and the sister relationship of these two species 
is strongly supported. Schiedea appears to have 
colonized the older current major high islands, and 
the basal clades of Schiedea also exhibit the greatest 
1995). The 


ereater morphological diversity in these older clades 


morphological diversity (Wagner et ¢ 


may result from diversification into the greater range 
of environments that develop as islands age and from 
adaptation to more diverse modes of pollination. The 
higher probability of extinction of intermediate forms 
in the older clades may also emphasize differences 
among species. 

Most inter-island colonizations appear to be from 
older to younger islands (Wagner et al.. 1995), but 
a few back colonizations also may have occurred (e.g.. 
from Kaua'i to Nihoa (Schiedea verticillata V. Brown in 
Christopherson & Caum). from Maui Nui to O'ahu 
(several dimorphic species). and from O'ahu to Kauai 
(S. spergulina)). In most cases, movement between 
islands led to sufficient isolation in space and time to 


result in formation of new species that were single- 


[om 


island endemics, rather than species with multi-islanc 
distributions. Only three species have multi-island 
distributions. Schiedea globosa H. Mann is a small 
subdioecious subshrub that grows only on coastal 
cliffs of Oʻahu, Maui Nui, 


disperse relatively easily between islands by rafting 


and Hawaii and may 
(Wagner et al., 1995): floating mats of 5. globosa have 
been observed floating offshore of Moloka'i (Wagner 
et al., 2005b). Two other species (S. hookeri A. Gray 
and S. nuttallii Hooker) are both found on O'ahu and 
Maui Nui in diverse mesic forest. 

Several Hawaiian lineages show inter-island colo- 
followed by habitat 


nization to islands 


diversification (Wagner & Funk, 1995). Species pairs 


younger 


that are well supported in phylogenetic analyses give 
some indication of the importance and differences in 
inter-island versus intra-island speciation in Schiedea. 
Most species pairs occurring on older islands are on 
the same island but differ in habitat, are isolated 
ecologically, and, in some cases, show pronounced 
morphological differentiation. Species pairs on youn- 


ger islands tend to be in similar habitat on different 


52 Annals of the 
Missouri Botanical Garden 


. spergulina 
S. kealiae 
70 5. adamantis 
>. ligustrina 
5. salicaria 


. lydgatel 


_ | Schiedea 
>. haleakalensis 


S. mannil 


S. globosa 


3. sarmentosa 


S. menziesii 


S. hookeri 
S. pubescens 
>. pentandra 


S. nuttallii 


S. diffusa 
S. jacobii Mononeura 


». hawaiiensis 


». laui 


c 
— 


S. stellarioides 


un 

Ke 
N 
O 


5. perlmanii 


S. kauaiensis 


S. apokremnos | Anestioschiedea 


5. VISCOSA : . 
Nothoschiedea 
S. lychnoides 
96 
. obovata A 

] | Alsinidendron 
. trinervis 


61 a 
— S. verticillata | Polyneura 
80 S. attenuata | Leucocalyx 
AZ 
98 S. membranacea 


S 
S 
S 
S 
S 
S 
S 
S 
S 
$ 
S 
S 
S 
S 
— 
— 8. kalac 
— SY 
— 5 
S 
S 
S 
S 
S 
S 
5 
S 
S 
S 
S 
S 
S 
8 Alphaschiedea 


S. helleri 


—.— Honckenya peploides major 
66 Honckenya peploides diffusa 
3 Wilhelmsia physodes 


Minuartia mochringioides 


100 pe Minuartia rossii 
LL Scleranthus biflorus 


Geocarpon minimum 


gure l. riel consensus tree of 24 MP trees. Two major partitions of data were analyzed i nee data from 
n (E v ETS regions and morphology. y -one morphological characters were scored, nella * acters 
published in previous analyses (Weller et al., : Wagner et al., 1995; characters listed in Wagner et al., 2005b). The 


sections of the genus are shown to the right of cac 185 lade. Bootstrap values greater than 50% are shown above the lines. Two 
extinct species, Sehiedea amplexicaulis H. Mann and S. implexa (Hillebrand) Sherff, are omitted from the analyses (Wagner el 
al., 20050). Be for y 10 authorities are not given in the text include S. haleakalensis Mattf.. S. mannii St. John. S 
sarmentosa Degener & 1 J. kaalae Wawra . S. stellarioides H. Mann, S. perlmanti W. V. Wagner & Weller. S. kauaiensis 


2 


ohn, S. attenuata W. E agner, Weller & Sakai, Honckenya „ subsp. major (Hooker) Hultén (Oregon), Honckenya 
ido subsp. diffusa Nome mann) Hultén ex V. V. Petrovsky (Baffin), Minuartia rossii (R. Brown ex Richardson) Graebner, 
Winuartia d n DC.) Mattf.. Scleranthus biflorus Hodie f.. Geocarpon minimum Mackenzie. Figure adapted from 


Wagner et al., 50. 


Volume 93, Number 1 Sakai et al. 
2006 Breeding Systems in Schiedea 


S. adamantis 
S. ligustrina 
S. spergulina 
S. kealiae 


S. salicaria 


S. lydgatei 


S. haleakalensis 


mannil 
elobosa 


S. sarmentosa 


». menziesii 


S 
S. hookeri 
S 


. apokremnos | 
— — > S. pubescens | 
B S. pentandra | 
S. hawailensis 
| L E S. diffusa diffusa 
mE S. diffusa macraei 
7 S. nuttallii 
-- S 


i . kaalae | 
1 
S. jacobii | 
S. laui 
S. stellarioides 
S. perlmanii 
S. kauiensis 


S. viscosa 


S. lychnoides 
S. obovata 


S. trinervis 


S. verticillata 


uem S. attenuata 


om membranacea 
LL 8. helleri 


— Wilhelmsia physodes 


BEEN n Minuartia rossii 
LL Scleranthus biflorus 
Geocarpon minimum 


= OOL 2 Oa A z = OF © 


A 


K 


— | e Honckenya peploides major 


Honckenya peploides diffusa 


Minuartia moehringioides 


Figure 2. Biogeographic hypothesis for Schiedea using one of 24 most parsimonious trees reconstruc ted with MacClade 


0 (Maddison & Maddison, 2000; figure adapted from Wagner et al.. 2005b). Island abbreviations and ages: 


Nihoa 


(N, 


T: zd Kaua'i (K, 4.7 Ma), Oahu (O, 3.0-2.6 Ma). Maui Nui (M, 2.0-1.2 Ma). Hawaii (H, 0.6-ongoing Ma), Extra-Hawatian 


1 


Zquivocal regions are indicated in gray. 


For species with multi-island distributions, coding indicates the island where it is most likely that the species evolvec 


Annals of the 
Missouri Botanical Garden 


>) Kaua i (4.7) 


Y Niihau (5.2) 


Figure. 3. 


islands shown. 


Map of the Hawaiian Islands, with island age 


I 


islands and. show less morphological. differentiation. 
On the of En 


a perennial herb, occurs in the mesic forest of several 


older island Kauai, membranacea, 
valleys, while closely related S. helleri is a vine found 
at higher elevations in wet montane forests. In the four 
species of the former genus Alsinidendron H. Mann. 
inter-island dispersal differentiation i 
On D: Mann and 
lychnoides Hillebrand are separated by elevation 


preceded 


habitat. Kauai. viscosa H. 


Uu 


and habitat, although there is some geographic 


overlap. Schiedea viscosa occurs at slightly lower 


elevations in diverse mesic forest in limited areas. 
while S. /ychnoides occurs throughout higher elevation 
wel s on Kauai. On O'ahu, S. obovata (Sherff) 
W. I. Wagner & Weller occurs in forest a 
in elevations throughout the Waranae Mountains. 
while S. trinervis (H. 


mesic 


Hoffmann is 
the 


species 


Mann) Pax & K. 


found only in higher elevation wet forest in 


northern Waianae Mountains. In the five 


pairs on younger islands, both species occur in similar 
habitats, but geographical isolation has occurred 


through dispersal to different islands or mountain 


anges on the same island (or former island in the case 
of Maui Nui). Two of these pairs occur in wet habitats 
(S. jacobii W. L. Wagner, Weller & Medeiros and S 
laui W. L. Wagner & Weller on Maui Nui (Maui and 
Molokaʻi, the two subspecies of S. 
diffusa A. S. diffusa subsp. diffusa and S. 
diffusa subsp. macraei (Sherf) W. Wagner & 


respectively); 


Gray, 


X O'ahu (2.6 - 3) 
EN 


M 


millions of years. Nihoa (7.3 Ma) occurs to the northwest of the 


Weller on Maui Nui and the island of Hawari). One 


pair occurs in mesie habitat. (&. pentandra W. L. 
Wagner & E. Harris and S. pubescens Hillebrand on 
Oahu and Maui Nui, respectively), and one pair 


occurs in dry habitats (S. spergulina and S. kealiae on 
Although both S. 


adamanlis occur on O'ahu 


Kauai and O'ahu. respectively). 
ligustrina and dry 
habitats, there is clear geographical separation by 
mountain ranges. Schiedea ligustrina occurs through- 
out the Waianae Mountains (formed 3.2-2.5 Ma) in 
western O'ahu in dry forest to diverse mesic forest, 
often on cliffs. Schiedea adamantis occurs in a single 
population on steep dry slopes of open shrubland on 
Diamond Head crater in southeastern O'ahu (formed 
about 0.5 Ma: Juvik & Juvik, 1998). More definitive 
biogeographical patterns of other Schiedea species 


awail better resolution of their phylogenetic relation- 


ships. 
Several species may be undergoing directional 
selection for greater fitness in dry environments 


colonized relatively recently. For example, Schiedea 
hawaiiensis Hillebrand may have only recently shifted 
from mesic to dry habitat. It occurs in dry. high light 
environments, but exhibits morphological traits (large 
leaves, open pendent inflorescences, vining habit) and 
physiological traits (maximum levels of photosynthe- 
sis, water use efficiency) more characteristic of mesic 


site species (Mishio et al.. 


unpublished). On younge 


islands, ecological divergence of species may also be 


Volume 93, Number 1 
2006 


Sakai et al. 55 


Breeding Systems in Schiedea 


limited by the later development of geological features 
creating drier habitats. Wet habitats are present on 
vounger islands as soon as they become high enough 
to intercept moisture-laden tradewinds. but the drier 
leeward slopes and cliffs preferred by many Schiedea 
species develop only after erosion and/or subsidence 
age (Walker, 1990a, b). 


as the islands 


HABITAT SHIFTS AND CHANGES IN BREEDING SYSTEM 
The spectacular adaptive radiation in breeding 
systems in Schiedea apparently evolved. within the 


past 5-7 million years. The association of changes in 


breeding with habitat shifts has made 


difficult to discern cause and effect. 1.€.. 


syslem 
to determine 
if changes in habitat have driven shifts in breeding 
whether shifts in breeding system have 
Based 


character optimization, the ancestor giving rise to 


system, or 


allowed movement into new habitats. on 


Schiedea appears to have been hermaphroditic, and 
likely 


is di- 


the ancestral habitat of Schiedea was most 


mesic forest. Although Honckenya peploides 
oecious, Wilhelmsia Rchb. and other close outgroups 
are hermaphroditic, and it appears that dioecy was 
derived independently in Honckenya Ehrh. and also 


within Schiedea (Fig. 1; Nepokroeff et al., 

Dimorphic breeding systems appear to have been 
derived independently twice in Schie D 
apokremnos and once in section Schiedea. Within 


section Schiedea, it appears that S. lydgatei Hille- 
brand is hermaphroditic by reversal, possibly through 
Molokai. 


Facultative autogamy has evolved three times—once 


loss of females during colonization of 
in the former genus Alsinidendron, in the species pairs 
S. jacobii-S. laui, and in S. hawaiiensis-5. diffusa. 
Each of these groups with facultative autogamy has 
also evolved obligate autogamy (S. diffusa subsp. 
and 8. 


systems is 


Evolution of 
linked to 


Diversification into wet habitats 


macraei, S. laut. trinervis). 
autogamous breeding also 
changes in habitat. 
has occurred independently five times, and in several 
of these cases, a shift to facultative or obligate 
autogamy has occurred. The repeated. independent 


evolution of autogamy with wet habitats may result 
from historically small population sizes in these 


with selection for closed pendent. flowers 


species, 

that are less likely to be adversely affected by 
extremely wet conditions (Weller et al., 1998; Wagner 
el al.. 2005b). Some exceptions occur to these 


patterns: S. helleri occurs in wet montane forest, but 
shows no indication of autogamy, and S. hawatiensis 
occurs in dry forest, but is autogamous rather than 
sexually dimorphic. 

Evolution of dimorphic breeding systems is also 
linked to changes in habitat (Fig. 4). In both cases, the 


independent evolution of dimorphism is associated 


with a shift to dry habitat. All ten dimorphic species 


occur in dry and windy habitat and most hermaphro- 


ditic occur in mesic to wet habitat. Four 


hermaphroditic species also occur in dry habitats, 
although Schiedea lydgatei may be hermaphroditic by 


reversal. This pattern, where all dimorphic species as 


spec les 


well as a few hermaphroditic species occur in dry 
habitats (e.g.. S. menziesii Hooker, 5. hawaiiensis, and 
S. verticillata), suggests that the shift to dry wind 
habitats precedes the shift in breeding sys 
dimorphism, but that the evolution of dimorphism is 


not inevitable after a shift to dry habitats. Once the 


Lo 


em 


shift in habitat has occurred, selection may simulta- 


neously favor the evolution of dimorphism as 


a mechanism to promote outerossing as long as pollen 
1998). The 


evolution of adaptations promoting wind pollination 


is not limiting for females (Weller et al.. 
becomes critical with the loss of biotic pollination 
limiting pollen dispersal in these habitats (Rankin et 
al.. 2002). 
evolve because of low pollen production, as in 5. 


Alternatively, wind-facilitated selfing may 
hawaiiensis. a hermaphroditic dry-site species derived 
from a lineage of mesic-wet forest species with low 
pollen production. 

EVOLUTION OF DIMORPHISM 


SELECTIVE FACTORS IN THE 


AND SELFING 


Inbreeding depression, selfing rates. and resource 
allocation are all factors that can lead to changes in 


either selfing or. 


breeding systems by favoring 


alternatively, the separation of the sexes and the 
evolution of dimorphism in these species (reviewed in 
Sakai € Westneat, 2001). In 
inheritance of male sterility, as in Schiedea (Weller & 
Sakai, 1991), 


selection for 


cases with nuclear 
theoretical models suggest that strong 
may occur as a mechanism to 
1975; Charlesworth & 


dioecy 


promote outcrossing (Lloyd, 


Charlesworth, 1978; reviewed in Geber et al., 1999; 
Weller & Sakai, 2005). If inbreeding depression (9) is 


high, selfed progeny of hermaphrodites will have 
lower fitness than progeny derived from outcrossing. 
Unisexual individuals, usually females, will be 
favored because they produce outcrossed progeny 
with relatively higher fitness than the selfed progeny 
of hermaphrodites. Under the simplest. conditions, 
females will be favored if k +50 > 0.5, where k is the 
seed production of females relative to hermaphrodites 
k = 0 if females and hermaphrodites produce equal 


numbers of seeds) and s is the selfing rate. Because 


pra 


dioecious species produce only unisexual individuals. 
inbreeding (with selfing as the most extreme form of 
inbreeding) is much less likely in dioecious species 


than in hermaphroditic species. Separate sexes may 


56 Annals of the 


Missouri Botanical Garden 


vo 


| 


adamantis 
ligustrina 
spergulina 
kealiae 
salicaria 
lydgatei 


haleakalensis 


| 
| 
| 
| 
| 
mannii 
globosa | 
sarmentosa | 
menziesii 
hookeri | 
apokremnos | 
pubescens 
pentandra | 
hawaiiensis 
diffusa diffusa 
diffusa macraei | Oa 
nuttallii 
h 
fa 


| 
| 
laui | oa 


kaalae 


jacobii 


stellarioides 
perlmanii 

kauiensis 

viscosa 

lychnoides fa 
obovata 
trinervis | oa 
verticillata 


attenuata 


. membranacea 


helleri 


S 2 U Zz OZ 


VES = 


= 


Honckenya peploides major 


Honckenya peploides diffusa 


Wilhelmsia physodes 


Minuartia moehringioides 


— Minuartia rossii 


dg c 


Scleranthus biflorus 


Mac € ade 


the right. E a “al regions 


an 


lor breeding systems are 


herma rini a= 1 aulogamous: oa = 
" o 


Geocarpon minimum 


; eding system evolution hypothesis for Schiedea using one of 24 most parsimonious trees reconstructed with 
LO (Maddison & Maddison, 2000: modified from Wagner et al.. 


2005b). He 1 0 ts of species are 7 in the bar to 
Habitats: D = 


nodioeci IOUS: d 


indicated in gray. Iry pis s and c aut M = Mesic forest 


Bre aaa sd bea h 


eding Systems: g = 


1 D. autogamous. 


Volume 93, Number 1 
2006 


Sakai et al. by 
8 5 Systems in Schiedea 


also be favored in the absence of high inbreeding 
depression and high selfing rates if females compen- 
sate for the lack of male function (pollen production) 
by producing more than twice as many seeds as 
hermaphrodites ( = 1; Charlesworth & Charlesworth, 
1978; Lloyd, 1975). A 


greater seed production by females and higher levels 


combination of somewhat 
of selfing and inbreeding depression may also promote 
separale sexes. 

Resource allocation to primary sex traits such as 
change 


biomass of pollen and seeds will obviously 


with the introduction of unisexual individuals in 
a population of hermaphrodites (i.e.. the evolution of 
eynodioecy and dioecy from hermaphroditism). The 
more challenging problem has been to discover if 


he 


these changes in resource allocation patterns are 
driving force in the evolution of separate sexes, with 
subsequent changes in inbreeding depression and 


selfing rates, or alternatively. if changes in selfing 


rates and inbreeding depression occur first. The order 


of these changes is difficult to deduce in dioecious 


species, where changes both in resource allocation 


and in levels of inbreeding and inbreeding depression 


already have occurred. Gy nodioecious species that are 


under continuing selection for changes in the 


frequency of females may offer more insights than 


n the 


dioecious species into the factors important 
evolution of both gynodioecy and dioecy. 
Comparisons of six species of Schiedea species give 
some clues about the relative order and importance of 
changes in inbreeding depression, selfing rates, and 


seed production with the evolution of breeding 


systems in this lineage (Table l; Norman et al.. 
1995, 1997; Sakai et al., 1997a; Culley et al., 1999; 


2002: Weller € Sakai, 2005: Weller et 
Measurements. of seed production were 
Selfing 
rates were measured using starch gel electrophoresis 


field- 


Rankin et al., 
al.. 2005 


taken from 


field-collected inflorescences. 


grown in the greenhouse from 


remote steep cliffs, levels of inbreeding depression 


were measured in the greenhouse on progeny resulting 


from controlled crosses involving self and outeross 


o calculate inbreeding 


— 


pollinations. Traits used 
depression included the number of seeds per capsule 
following the controlled pollinations, percent germi- 
nation, percent survival until flowering, and either the 
number of flowers per inflorescence or inflorescence 
biomass. The relative fitness of selfed progeny as 
a proportion of values for outcrossed progeny was 
calculated for each trait for each family and used in 
a multiplicative fitness function. Inbreeding depres- 
calculated as l-relative. fitness. Our in- 


sion was 


breeding depression studies of these perennial plants 


in the greenhouse probably underestimate the levels 


of inbreeding depression expressed in the field over 
more than one flowering season in presumably harsher 
field conditions. 

Schiedea membranacea. is a member of the basal 
clade of the lineage and has a hermaphroditic 
breeding system. It is a perennial herb endemic to 
Kauai that grows in mesic environments typical of 
hermaphroditic species. Schiedea membranacea has 
rate) that 
prevents expression of the high inbreeding depression 
detected. fol 
house (Culley et al., 


Pe, 


a high level of outcrossing (low selfing 


owing controlled crosses in the green- 
1999 


diversity (number of alleles per locus, percent of loci 


Measures of genetic 


polymorphic, and percent heterozygosity) also are 
consistent. with other outerossing species (Weller el 
al. 1996). 


biotically pollinated, with a relatively 


Schiedea membranacea is apparently 


low pollen : 


ovule ratio and an open inflorescence (Table 1). 
Because of the low selfing rate, expression of 


inbreeding depression is low, and the hermaphroditic 
breeding system in this species is apparently stable. 
In contrast to Schiedea membranacea, S. viscosa has 


stable hermaphroditic breeding system because it 


^ 


has a high selfing rate, but much of the inbreeding 
depression characteristic of outcrossing species has 
been eliminated from these populations (Weller et al., 
2005). 


selfing, the pollen 


Typical of populations that are historically 
: ovule ratio is extremely low, and 
S. viscosa measures of 


also has very low genetic 


diversity (Table 1). Schiedea viscosa (formerly in the 
genus Alsinidendron) is a rare hermaphroditic vine 
that occurs in mesic habitats of Kauai. The flowers of 
S. viscosa are pendent and remain relatively closed, 
traits that may be adaptive in rainy conditions and 
that may have promoted the evolution of selfing. 
Schiedea viscosa also produces nectar (0.64 ul/flower/ 
24 h in the greenhouse), suggesting an earlier history 
of biotic pollination (Weller et al.. 1998) 

Two gynodioecious species illustrate the changes 
that may occur with an increase in the frequency of 
females in populations. Schiedea salicaria, a gynodioe- 
cious species with a low frequency of females (13%), 
habitats of the West Maui 


Mountains. Schiedea adamantis is also gynodioecious 


s found in dry windy 


with a higher frequency of females (39%). and is 
found on the more recent dry slopes of Diamond Head 
Oahu. In 


differentiation between females and hermaphrodites 


Crater on these two species, grealer 


sex traits occurs in 
2005). In $. 


females and hermaphrodites show only slight differ- 


in primary and secondary 


adamantis (Golonka et al., salicaria, 
entiation in floral and inflorescence traits and nearly 
identical seed production (Weller & Sakai, 2005). In 
S. adamantis, females produce 2.3 times more seeds 


than hermaphrodites in the field. This occurs because 


— 


Table 
1998: Culley et al.. 


populations are « 


1999: Rankin et al.. 
of flowers per inflorescence and inflorescence c pie ane were not measured in S. viscosa. 


30 individuals: | > 100 nara Estimates of genetic diversity (mean number o 


2002: 


‘eller et a 


large populations are 


Habitat, breeding system, se 10 rate, inbree ing depression. and differences in resource allocation f 


or six species of Schiedea. For methods see Norman et al.. 


1997: Weller et al.. 


)05. Resource allocation differences between females and he rmaphrodites cannot be calculated in herm: iphroditic species. Number 


heterozygosity) were based on 24 to 39 plants per population (Weller et al.. 1996). 


fl 


fl 


E 
INHOTFesce 


Inflorescence condensation is the number of flower: 
f alleles per locus, percentage of loci 1 and mean 


h 
/ 


1IIIIOTESCEnC t 


e length in em. Small 


Habitat 


forest herb 


Breeding system 


Pollinator 


S. membranacea mesic 


S. viscosa mesic 


forest 


vine 


S. salicaria dry 


shrubland shrub 


S. adamantis dry 


shrubland shrub 


gvnodioecious 
3% 


evnodioecious 


menziesii dry 


shrubland shrub 


S. lydgatei dry 


shrubland shrub 


hermaphroditic 


Ovule (N) 
Pollen (N) 
P:O ratio 
Pollen size (p) 
N fl/inflo 
Inflorescence 
(em) 
Inflorescence 
condensation 
Nectar 


Selling rate 
Inbreeding depression 
Mlocation 


Population size 


Mean N alleles/locus 


% of loci polymorphie 


Mean heterozygosity 


Assessment 


hermaphroditic hermaphroditic e F) 9% F) hermaphroditic (reversal) 
biotic biotic wind wind not adapted to wind biotic (moth) 
29 122 35 17 46 32 
8.037 3,122 14.756 22.177 13.935 11.211 
282 25 119 1308 304 351 
32:5 38.0 33.5 28.7 34.8 32.5 
94.3 E 30.1 54.5 26.6 24.5 
length 42.3 — 6.9 5.2 4.4 10.6 
2:2 E 4.5 10.5 6.0 23 
(uL/24 h) 0.030 0.643 0.236 0.155 0.385 0.325 
low high high high high low 
(0.3) (presumed 1.0) (0.7) (0.7) (0.8) (0.2) 
high OW high high high high 
(0.7) (0.0) (0.8) (0.( (0.8) (0.6) 
— no difference between F produce 2.3X more — — 
al seeds than H 
large small large large large large 
3.0 1.0 2.2 1.6 2.7 2.7 
66.7 0 11.8 1 77.8 88.9 
0.566 0 0.305 0. 0.340 0.322 
stable: outerossing stable: facultative unstable: selection for unstable: se sale tion unstable: selection stable: outerossing 
hermaphrodites selfing female for females for females hermaphrodites 


USPIBO) jealuejog unossi|A 


ay} jo sjeuuy 


Volume 93, Number 1 
2006 


Sakai et al. 59 
ca Systems in Schiedea 


females and hermaphrodites produce similar numbers 
of inflorescences and flowers per inflorescence in the 
field, but 
inflorescence 
1997a). 


adaptation toward wind pollination, with a higher 


capsules per 
(Sakai et al., 


1 greater 


females produce far more 


than hermaphrodites 


— 


Schiedea adamantis also shows muc 


pollen : ovule ratio, smaller pollen grains, and more 


condensed inflorescence than S. salicaria. Schiedea 


salicaria, although wind pollinated, has pollen: 
ovule ratios, pollen grain size, and inflorescence 


more similar to biotically pollinated 


1998: Golonka et al., 2005 


Schiedea salicaria and S. adamantis also differ n 


architecture 


species (Weller et al., 


levels of selfing and inbreeding depression. Schiedea 
salicaria has a mixed mating system, with a high 


rate (0.7: Weller € Sakai, 2005). 


extremely high inbreeding depression (0.8: 


selfing but with 

Sakai el 
al.. 1989). The low frequency of females in 5. salicaria 
is consistent with these estimates, although if selfing 
and inbreeding depression have been underestimated 
in these greenhouse studies, selection for increased 
representation of females may be strong. In contrast, 
rates and high 
1997a). In 


greater seed production of 


adamantis has both high selfing 


inbreeding depression (Sakai et al., 
combination with the 
these rates of selfing and inbreeding de- 
in 5. 
adamantis. The predicted frequency of females (based 
on Lloyd, 1975) is 4296, remarkably close to the 
observed frequency of 39% (Sakai et al., 1997a). The 


~ 


in &. 


females, 


pression. suggest strong selection for females 


low estimates of genelic diversity adamantis 
suggest a severe population bottleneck in the past 


(Weller et al., 1996). 


population fluctuated 


More recently, the size of this 


has widely over the past 


20 years, from a high of about 400 flowering plants 
t 
drought and competition from alien species. 


a current low of 2 plants because of extended 


Schiedea menziesii, one of the few hermaphroditic 
Schiedea species growing in dry habitats, may be 
under selection for evolution of sexual dimorphism. 


Unlike other dry site species, 5. menziesii does not 


— 


appear to be well adapted for wind pollination (e.g.. 


low pollen f 
wind pollination in a wind tunnel (Weller et al.. 1998). 


ralio) and shows no evidence « 


o 


: ovule 


This species has both high inbreeding depression and 
conditions that favor females in 
2002). Females 


field-collected seeds grown in 


high selfing rates, 


populations (Table Rankin et al., 
occasionally occur in 
the greenhouse, but have not been detected in the 
field. 
are a result of a mutation for male-sterility (Weller & 
Sakai, Without 
females may not be able to establish and increase in 


Greenhouse crosses suggest that these females 


unpublished). biotic pollinators, 
the population without adaptations to wind pollina- 


tion. The lack of adaptations for intrafloral selfing 


suggests that the evolution of autogamy is unlikely, 
even with high selfing rates. 

Schiedea lydgatei is another hermaphroditic species 
found in dry habitats on Moloka‘i, and based on 
phylogenetie evidence, appears to be hermaphroditic 
by reversal from a sexually dimorphic breeding 


U nlike S. 


selfing rate, which limits expression of the high levels 


system. menziesii, S. lydgatei has a low 


of inbreeding depression found in this species. As 
a consequence, the breeding system is stable despite 
Schiedea 


the high inbreeding depression levels. 


— 


lydgatei is apparently moth-pollinated (Norman e 
al. 1997), and the occurrence of biotic pollination 
may be critical in the maintenance of high outcrossing 


rates and hermaphroditism. 


— 


The patterns of inbreeding depression, selfing rates. 
and allocation shifts in these six hermaphroditic and 
that the 


dimorphism is associated with a shift to dry, windy 


dimorphic species suggest evolution of 


— 


habitats. and the evolution of wind pollination i 
Schiedea. The 


appears to occur before changes in sex allocation. For 


colonization of windy, dry habitats 


most hermaphroditic species, the shift to dry. windy 


abitats apparently is associated with a loss of insect 
pollinators, resulting in an increase in the selfing rate. 
The 


expression of the high inbreeding depression rates 


increased selfing rate may result in greater 
found in these historically outcrossing populations. 


Under these conditions, females could be favored 
because they are obligately outerossing, but only if 
pollination is not limiting populations (Weller et al., 
1998 
pollination may be followed by the rapid appearance 
Once 


troduced into the population, there is strong selection 


In these environments, the evolution of wind 


f females in populations. females are in- 


for different allocation. patterns favoring hermaphro- 
dites with greater male function. In these dry habitats, 


selection may act on traits associated with wind 


pollination and/or simultaneously on traits associated 


with resource allocation to increasingly male or 
female function. 
GENETIC POTENTIAL FOR THE EVOLUTION OF DIMORPHISM 


Theoretical models suggest that the evolution of 
dioecy may be favored because inbreeding depression 
levels and selfing rates are both high or because shifts 
in allocation of resources to a single sexual function 
result in accelerating fitness returns for individuals 
(Charlesworth € Charlesworth, 1978; Charnov. 1982: 
Lloyd, 1984). 


result 


These accelerating fitness gain curves 


may from a variety of ecological factors, 


including specific dispersal agents and shifts. in 
pollinators (reviewed in Bawa. 1980: Lloyd. 1964 
Sakai & Weller, 1999). In 


Thomson & Brunet, 1990; : 


Annals of the 
Missouri Botanical Garden 


Schiedea, shifts to wind pollination, as well as high 
inbreeding depression and high selfing rate 
to be strong selective pressures favoring the evolution 
The ability to respond to 
f 


breeding systems will depend upon the genetic 
: à | | : 


of gynodioecy and dioeey. 


these selection pressures with further evolution 


potential for changes in allocation to male and female 
function. 


— 


Ji 
n patterns of allocation 
male and female function. 


of these models assume the existence 


heritable genetic variation i 


lo 


tradeoffs in 
allocation between the two sexual functions are also 
assumed (Stanton & Galloway, 1990: 

1991: Geber 


genetic 


Genetic 


Charlesworth & 
el 1999), 


male 


Morgan. 
Negative 


reviewed in al.. 


correlations between and 


female traits, such as stamen biomass and ovary 


hermaphrodites, may enhance the evolu- 
dioecy, 


biomass 


tion of while positive genetic correlations 
between these traits would impede evolution of dioees 
and instead imply selection for larger or smaller 
flower size. 

We know little about patterns of genetic variation 
and covariation in characters related to sex allocation 
for natural populations. Only a few recent studies have 
explored the quantitative. genetics of sex allocation 
traits in hermaphroditie plant species ( 


& de Jong. 


van Noordwijk 
980: Stanton & Galloway, 1990: Houle, 


1991: O'Neil & Schmitt, 1993; Agren & Schemske. 
1995; Delesalle & Mazer. 1995: Mazer & Delesalle. 
1996: Campbell. 1997: Mazer et al.. 1999). These 


studies used controlled crosses to create families to 
estimate genetic variation and covariation of morpho- 
logical characters, but they generally were not placed 
in Other have 
examined the quantitative genetics of sex allocation 


trails in gynodioecious (Ashman, 1999, 2003; Delph 


a phylogenetic context. studies 


el al., 2004, 2005). andromonoecious (Elle, 1998). 
and dioecious species (Meagher, 1992, 1994). Even 


fewer studies have examined the quantitative genetics 
of physiological traits in gynodioecious species (Poot 
al., 1996; Caruso et al., 2003) 

Closely related species of Schiedea with different 
proportions of females and with 


el 


variation in male 
function in hermaphrodites should be especially 
informative the un- 
derlying these sex allocation differences. In Schiedea. 
male sterility is controlled by 


in studies of 


genetic changes 
a single nuclear gene 
(Weller & Sakai, 1991). but quantitative variation in 
other traits related to sex allocation females and 
hermaphrodites is likely to be controlled by a number 
of genes. Some gynodioecious species may have 
breeding systems that are in transition, with allocation 
patterns that are under directional selection. Given 
this range of breeding systems and allocation patterns 


in such closely related species, we expect significant 


S. appear 


heritability of biomass allocated to male and female 
function in gynodioecious species. Significant herita- 
bility of biomass would indicate the potential for 


evolutionary change. assuming that genetic correla- 


We 


genelics approach to 


tions do not otherwise constrain 


response. 
have used a quantitative 
examine the genetic potential to evolve new breeding 
systems and traits related to allocation to male and 
female function in gynodioecious S. salicaria, a spe- 
cies with 139 females, and in S. adamantis. a species 
with 39% females. 

We used heterozvgous hermaphrodites and females 
as parents to. generate families that would produce 
both hermaphrodites and female progeny. In Schiedea 
salicaria, thirty-five families (each consisting of one 
heterozygous hermaphrodite and one related female) 


were used pollinations to produce the plants for 


a single generation experiment to estimate the 
heritabilities and genetic covariances for the focal 


traits. We employed a partial diallel design similar to 
that of Meagher ( 


or greater statistical power, 


1992). but with larger sample sizes 


Each hermaphrodite was 
used as a pollen donor and crossed with three females 


randomly chosen from different families. and each 
female was crossed with three unrelated 


hermaphrodites. In most cases, five females and five 


different 


hermaphrodites from each sibship were measured for 
more than 30 traits. Additive variances were 


estimated separately for females and hermaphrodites 


genetic 


by examining the component of variation among 
paternal half-sib families using restricted maximum 
likelihood methods (see Kearsey. 1965; Shaw, 1987 
Meagher, 1902). Narrow sense heritabilities were 


calculated for each trait in each sex as the additive 


genetic variance (four times the paternal variance 


component for this design) divided by the total 
variance (paternal variance + maternal variance + 
error: Falconer & Mackay, 1996). Genetic covariances 


and correlations were estimated separately for females 
and hermaphrodites, 


predictors (BLU Ps) 


using best linear unbiased 


of sire values obtained 
from our Proc Mixed analysis (SAS, 2001: 
., 2003). Pearson correlation coefficients between 
BLUPs were calculated f 


hermaphroditic data sets were also combined and 


breeding 


Conner el 


or each sex. The female and 


Pearson correlation coefficients were calculated 


between homologous trait 


Heritabilities, 


s of the two sexes. 

and genetic correlations of females 
of Schiedea Most 
floral traits show significant heritabilities in females 


and hermaphrodites salicaria. 


and hermaphrodites. In general, the floral biomass of 


stamens (measured only in hermaphrodites) and 
carpels both sexes shows high heritability. Fruit 
biomass is not heritable in terminal flowers. but is 


heritable in the more numerous lateral flowers in both 


Volume 93, Number 1 
2006 


Sakai et al. 61 
Breeding Systems in Schiedea 


sexes (Sakai et al., unpublished). At the inflorescence 


level, both females and hermaphrodites exhibit high 


heritability in the number of flowers and fruits per 


inflorescence (Weller et al., in press). These results 
suggest that significant genetic variation is present in 
this species and that a response to selection could 
result in further differentiation of females and 
hermaphrodites, depending upon genetic correlations 
of traits. 
Genetic correlations within hermaphrodites are 
unlikely to promote or hinder the evolution of dioecy 
in this species. Within hermaphrodites, female bio- 


mass (ovary, fruit, or seed) and male (stamen) biomass 


are independent of each other. Female biomass in 


females and male biomass in hermaphrodites also 


showed no genetic correlation, suggesting that spe- 


cialization toward greater female function in females 
and greater male function in hermaphrodites would 
not be constrained. Other genetic correlations may 
impede selection for sexual dimorphism, with positive 
genetic correlations between similar traits in females 
and hermaphrodites. For example, the female biomass 
of females and of hermaphrodites is significantly 
genetically correlated, making specialization more 
difficult (Sakai et al., 


Schiedea salicaria has likely undergone a 


unpublished). 

recent 
transition from biotic to abiotic wind pollination, and 
significant narrow-sense heritabilities were detected 
for inflorescence condensation and other traits related 
to wind pollination in Schiedea species (Weller et al., 
Heritabilities generally higher in 


in press). were 


hermaphrodites than in females. The presence of 
significant narrow-sense heritabilities for traits. asso- 
ciated with wind pollination suggests that selection for 
more effective wind 


pollination in the windy, 


pollinator-limited environments where S. salicaria 


grows could lead to the evolution of the highly 
condensed inflorescences characteristic of other wind- 
pollinated species of Schiedea. A recent study of eco- 
adamantis 


physiological traits of S. salicaria and 5. 


showed few differences between females and her- 
maphrodites, but significant narrow sense heritabil- 
ities for stomatal conductance and specific leaf area, 
suggesting that these traits could also respond 

2006). 


trails related to floral and inflorescence architecture, 


selection (Culley et al., Further analyses of 
wind pollination, and physiology should clarify how 


other genetic Moa may affect the evolution of 


dioecy in this species. 


CONCLUSIONS 


A multi-disciplinary approach, including phyloge- 
netic analysis, population biology, and quantitative 


genetics, has been instrumental in understanding both 


the genetic potential for different breeding systems to 
evolve in Schiedea and the selective factors that have 
promoted speciation and shifts in breeding systems. 
Autogamy and dimorphism evolved independently 
several times in Schiedea, and diversification has been 
associated with changes in breeding system, pollina- 
tion system, and habitat. In general, dimorphic 
species have evolved in dry, windy habitats and are 
wind pollinated, while autogamous species occur in 
wetter habitats. Dimorphic breeding systems may have 
evolved as a mechanism to promote outcrossing 
because of the expression of high levels of inbreeding 
depression resulting from high selfing rates that occur 
with the loss of pollinators related to a shift to drier 
habitats. Quantitative genetics studies suggest that 
most morphological traits associated with wind 
pollination and resource allocation are heritable and 
Strong 


habitats and for 


could respond to selection. selection for 


outcrossing in dry selfing in wet 
habitats suggests that a further response to selection 


in breeding systems and other traits is possible. 


Literature Cited 
1995. 
semiovata. 


Sex allocation in the 
Evolution 49: 


\gren, J. & D. W. Schemske. 


monoecious Que Begonia 
121-130. 
Ashman, T.-L. 


a gy nodioeci :10US 


1999, Quantitative genetics of floral traits in 


wild strawberry Fragaria virginiana: 


Implications for the independent evolution of female 
Heredity 83: 


and hermaphroditic floral 


7133-74 " 


phe nolypes. 
. Constraints on the evolution of males and 
sexual its ism: Field estimates of genetic architecture 
of reproductive trails in three populations of S 10US 
Fragaria virginiana. E We 57: 2012-202 

Baldwin, B. G. 2003. 


V phyloge netic E on ES 
origin and evolution of 93-228 


Madiinae. Pp. 193-228 in 
Carlquist, B. G. Baldwin & G. D. Carr (editors). 1 
Missouri 


and Silverswords: Evolution of the Madiinae. 


Botanical Garde n Press, St. Louis. 

D. Kyhos, J. Dvo rak € G. D. Carr. 1991. 
ropas DNA evidence for a North American origin 
of the Hawaiian silversword alliance (Asteraceae) 
Proc. Natl . Sci. U.S.A. 88: 1840-1843. 

Ballard, H. E. Jr. € K. J. Sytsma. 2000, Evolution and 
biogeography of the woody Hawaiian violets (Viola. 


Violaceae): Arctic origins, he b 'eous ancestry, and bird 
dispersal. a 54: 1521- 

Bawa, K. S. 1980. Evolution of iie in flowering plants. 
Ann o . Ecol. Syst. 11: 15-40. 

Campbell, D. R. 1997. 
allocation to male and fe Ps Py. lions in a natural plant 

79. 
pM E G. p. 
ju me p ords: 


Genetic correlation between biomass 


5 | redi 


— 
= 


rr (editors). 2003. 
E of 


Carlquis 
the Madiinae 
5 1 1 Garden Press. St. Louis 
Caruso, C. M., H. Maherali & R. 2003. 9 
Specific de and e implications for the 
maintenance of females in gynodioecious Lobelia siphili- 


. Oecologia 135: 524—431. 


ackson. 


Annals of the 
Missouri Botanical Garden 


Charlesworth, B. & D. Charlesworth. 1978. A model for the 


evolution of dioecy and gynodioecy. Amer. Naturalist 112: 
975—991. 

Charlesworth, D. & M. T. Morgan. 1991. Allocation of 
resources lo sex functions in flowering plants. Philos. 
l'rans., Ser. B 332: 91-102. 

Charnov, E. L. 1982. The reory of sex allocation. 
(Monographs i in Aa Biology: 18). Princeton Univ. 


, Princeton, i w Jers 

a The ar and subsidence of the 
Hanalan: Emperor ae anie chain, Pp. 35-50 in A. Keast 
& S. E. Miller (editors), The Origin and Evolution of 
Pacific Is T Biotas, New Guinea to Eastern Polynesia: 
Patterns and Processes. SPB Academic Publishing. 
Amsterdam 95 Netherlands. 

Conner, J.. R. inks & C. Stewart. 2003. Expression of 


MIO genetic variances and covariances for wild radish 


floral traits: comparison between field and greenhouse 
environments. E pn e mes 
Culley, T. M., . Weller, Sakai & A, 


. Rankin. 
1999, 1 5 1 br selfing rates in a self- 
compatible hermaphroditic species. Schiedea membrana- 
cea (Caryophyllaceae). Amer. ae Bot. 86: 980-987. 


Dunbar-Wallis, K. Sakai, S. G. Weller & D. 


R. Campbell. 2006. Minera 'artation of m “al 
traits in [wo gynodioecious s species of ; 1 (Carvo- 
phyllaceae). New Pod 169: 589-601. 

Delesalle, V. A. & S. J. Mazer. 1995. The structure of 


phenotypic variation in gender and floral traits within and 


among populations of pee marina (Caryophylla- 
10. 


ceae). Amer. J. Bot. 82: 798-8 
Delph. L. F.. J. L. Gehring, F. M. Frey, M. Arntz & M. 
Levri. 2004. Genetic constraints ral evolution in 


on 

xually dimorphic plant revealed by d ial selection. 

nnd 58: —1946, 
Arntz, 

1 1 95 with lle 


M. Levri & F. M. Frey. 2005. 

al display lead to sexual 
5 in the cost of reproduction. Naturalist 
A 

Elle, 


Genetic 


Amer. 


1990 The quantitative genetics of sex allocation in 


e po nnn cious perennial, Solanum carolinense. 
Here dity 80: 181—168 

Falconer, D. S. & T. F. C. Mackay. 1996. Introduction to 
DE veneties. Prentice 1 ew Yo 

Geber, M. T. E. Dawson & L. F. De Iph am 1999, 
Sexual aud Gender 1 in Flowering Plants. 


n e ag, Heidelbe 


a 
rg. 


Givnish, J. & K. J. Sytsma (editors). 1997. Molecular 
E "in d Adaptive. Radiation. Cambridge Univ. 
Press. Cambridge. 

.J. F. Smith & W. J. Hahn. 1995. Molecular 


evolution ^ adaptive radiation, and geographic 
in Cyan a (Campanulac eae, Lobelioideae) 
in W & V. 


in “agner & A. Funk (editors), 


Hawaiian 
n i 


Bioge e ‘volution on a Hot 
i adu eda M pps 
A. 


Golonka, A. M.. A. K. Sakai & S 


Archipelago. 


— 


por 2005. Wind 

ao sexual dimorphism, changes in floral traits 

of Schiedea Amer. J. Bot. 92 
l 192- 502. 

Grant, P. R. (editors). 
Univ. Press, Oxford. 

Houle, D. 1991. Genetic 
What genetic correlations are made of and why it matters. 
Evolution 45: 6: Pa 

Juvik. S. P. & J. ( 
edition. 


(Caryophyllaceae). 


1998 Oxford 


Evolution on Islands. 
fitness correlates: 


covariance of 


018. 
Juvik. 1998 
1 i Press, 


Atlas o 


Honolulu. 


© Hawaii. third 


Univ. 


Kearsey, M. J. 1965. Biome trical analysis of a random mating 


A comparison of five 
Heredity w 205-255, 

EO: ». Keeley, P. S. Vroom & R. K. Jansen. 1998. 
Molec "e o: nce d an African origin of the Hawaiian 


. Natl. Acad 


experimental. designs. 


ende "mie de rdc ge : n raceae). Proc 
Ser. U.S.A. 95: 15440-15 

Lint rs * » Motlev, J. " Jeffrey & V. A. Albert. 

ladoge nesis and re lic 9 in P endemic 
mints (Lamiaceae T Cladistics 19: 480—495 

Lloyd, D. G. 1975. The maintenance es 
androdioecy in angiosperms. 


2002. 
Hawaiian 


evnodioeey and 
39 


Genetica 325— 
984. Gender as allons im outcrossing cosexual 


plants. Pp. 277-300 1 . Dirzo & J. 


Perspectives in Plan onmes 


Sarukhan (editors). 

Ecology. Sinauer, 
Sunderland. Massac lies ls. 

Maddison, W. P. & D. R. Maddison. 2000. MacClade 4: 
Analysis of phylogeny and character evolution. Version 
1.0 Sinaue i 1 rland, Massachusetts. 

. Delesalle. 1996, 


in Spergularia marina (Caryophyllaceae): 


Covariation among 


"alls 
Geographic and temporal one dn phenotypic and 


among-family corre p J. Evol. . 9: 993-1015 


„ & P. R. Neal. ines p MONA of feral 
traits to selection. on primary sexual investment in 
Spergularia marina: The battle between the xes. 
Evolution 53: 717-7: 

Meagher, T. R. 1992 The quantitative genetics of sexual 
dimorphism in Silene latifolia (Caryophyllaceae). 1 
Genetic variation. Evolution 46: 445—157 

1994. The quantitative genetics of sexual di- 


foglie in Silene latifolia (Caryophyllaceae). II. Re- 
sponse to sex-specific selection. Evolution 48: 939-951. 
Nepokroeff, M., K. J. Sytsma, W. I. Wagner & E. A. 


2003. Reconstructing ancestral 


Zimmer. 
illerns. of colonization 


pe 
understory tree 


and P "sal in the Hawaiian genus 


ria (Rubiaceae): d of parsimony and 

Helen approaches. Syst. 1 Siol. 52: 820-838 
I. Wagner, P. S. Soltis. S. 

Soltis, i K. Sakai & E. A. Zimmer. 2005. 
dde Pp. 13-20 in W. L. Wagner. S. G. Weller & 

Sakai (editors), Monograph of Schiedea ph 

V Syst. Bot. Monogr. 72: 

Norman, J. K.. K. Sakai, 5. G. Weller & 

K 


95. h ibreeding sion in 


Weller, D. E. 


logeny of 


T. E. 
morphological and 
physiological traits of Schiedea lydgatei ee eae) 


Dawson. 
depre 


in two environments. Evolution 49; 297: 

G. Weller & Sakai. 1997. Polli- 
nation TEH a outcrossing rates in ba ps 
Schiedea lydgatet (Caryophyllaceae). Amer. J. Bot 
0641-0648. 

O'Neil, P. & J. Schmitt. 1993. Genetic constraints on the 


indepe de nb evolution of male and female reproductive 

characters in the tristylous plant Lythrum salicaria. 
Evolution 47: 1457-1471. 

Poot, P.. J. 8 T. 1990. 


Pilon a Pons. 
characteristics of leaves of male-sterile and hermaphrodite 
of l 


Photosynthetic 


antag 


sex ly pes o lanceolata grown under conditions of 


contrasting uisum and light availabilities. Physiol. PL 
98: 780—190. 

Price. J. P. & D. A. Clague. 2002. How old is the Hawaiian 
biota? Geology and phylogeny suggest recent divergence 
Proc. Royal. Soc. London. B 269: 2429-2435, 

K D. Elliott-Fisk. 2004. preda history of the 
Maui Nui complex, Hawai'i, y its implications for 


biogeography. Pacific Sei. 58: 27 


Volume 93, Number 1 
2006 


Sakai et al. 63 
Breeding Systems in Schiedea 


2004. 


consequence, 


& W. L. Wagner. 
5 lineages: Cause. 
volution 58: 2185-2200. 

1 7 00 A. E., S. G. Weller & A, K. Sakai. 2002. Mating 


system instability in Schiedea menziesii (Caryophyllaceae). 


Speciation in Hawaiian 
and mode. 


Evolution 50: Va v 

Renner S. & R. E. Ricklefs. 1995. Dioecy and its 
correlates. in the 1 plants. Amer. J. Bot. 82: 
596-00 

Sakai. A. K. & S. G. Weller. 1999. Gender and sexual 


dimorphism in flowering plants: A review of terminology. 


biogeographic patterns, ecologic al correlates, and phylo- 
genetic approac hes. Pp. 1-31 in M. Geber. T. E. 
Dawson € L. F. Delph (editors), Sexual and Gender 


Dimorphism in Flowering Plants. Springer-Verlag. 
Baa 

v D. Mating systems. Pp. 193- 
206 in C. Fo f & D.J. 
E clan Ecology: ; onc 7115 and Case Studies. Oxford 
U Oxford. 

"has: & S. GW eller. 1989. 

depresión in Schiedea globosa and S. salicaria (Caryo- 


. Westne P zu 
D ff 


Fairbairn (editors). 


niv. 
Inbreeding 
phyllaceae). subdioecious and gynodioecious Hawaiian 
Amer. J. Bot. 76: 437444. 


species. 


. W. L. Wagner. D. M. Ferguson & D. Herbst. 
19953. Dies of dui in the native Hawaiian. flora. 
Ecology 76: 2517-25 

V ———. 1995b. Biogeographical 


and e olor i corre lates of dioecy in the Hawaiian flora. 
543 


Ecology 16: 


Chen, S.-Y. Chou & 


G 'eller. . 
and mainte- 


E 9 ee of gynodioecy 


Missa 1 
females: The 


= 


nance of of inbreeding depression, 


oulerossing rales, and resource allocation in Schiedea 
adamantis (C aryophyllaceae). Evolution 51: 724-730. 
. L. Wagner, P. S. Soltis & D. E. Soltis. 


1997b. 


Pied perspectives on the evolution of 
dioecy: Adaptive radiation. in the endemic Hawaiian 
genera X 9 and Alsinidendron (Carvophyllaceae: 

\lsinoideae). Pp. 455 Givnish & K. Sytsma 


(editors). Molecular lato - Adaptive Radiation. 
. Press. Cambric 
Mehrhoff. e 


flora. Syst. Bio 


Cambridge Univ 
V 


Wagner & L. 
l 


endangerment in the Hawaiian 
76— 
SAS Ho 2001. SAS, Version 8.2. SAS Institute, Cary, 
NC. 
1987. 


genetics. of natural populations. 


Shaw. R. G. Maximum -likelihood approaches to 
or 
41: 812-820. 

Solus. P. S.. D. E. Soltis. 
Wagner. 1996, 
Hawaiian endemics Schiedea and 

phyllaceae), Syst. Bot. 21: 365-371 


Evolution 


A. R. Sakai & W. 


analysis of the 


S. G. Weller. 
Molecular phylogenetic 
Alsinidendron (Carvo- 


Stanton. M. L. Galloway. 1990. Natural sclection and 
allocation to re i. tion in flowering plants. Pp. 1—50 in 
M angel (editor), Allocation and Sex Change: 


Experiments Me Models. American Mathematical Society. 
Prov ide nce, js ode Island. 
88. Dioecism and its correlates in the Cape 
flora of f South Africa. Amer. J. Bot. 75: 1742-1754. 
Swofford, D. L. 2001. PAUP*: 


a 5 und di: methods). version 4.0b6. Sinauer. 


ri 


Phy les netic 


Sunderland. Massachusetts. 


Patterns of 
31 5l: 


Analysis Using 


> 


Brunet. 1990. 


Thomson, . 
evolühon of lotes in seed plants. Trends Ecol. 
11-16. 


Hypotheses for ine 


uv vol. 


e Jong. 1986. Ac quisition and 
c ir influence on variation in life 
Naturalist 128: 137—142. 

Funk (editors). 1995. 
puc on a Hot Spot Arc hipelago. 


van Noordwijk, A. J. & G. 
allocation of resources: 
a T. 

Wagner, W. A. 

Bioge y 


histories tac g 
Hawaiian 


x Institution E ss. Washington, D.( 
G. Weller « ; Sakai. 1995. 

e ote n Alsinidendron (C aryophyl- 
W. agner & V. Fun 


Bis uma ny and 
i 
md Pp. 25] —258 in i 
. Hawaiian Biogeography: Evolution on a Hot Spot 
1 Smithsonian Institution Press, Washington. 
2005a. Flora of 


http://ravenel.si.edu/botany/ 


. D. R. Herbst € D. H. Lorence. 

the Hawaiian Islands website. 

pacific ru dr [accessed 12/9/05 |. 

— ——, S. G. Weller A. Sakai. 2005b. ee 
of Schiedea (C bd Mondo ded Syst. Bo 

Monogr. 72: 1-169. 

Geology. Pp. 21 
R. Herbst & S. H. 
flowering plants of Hawaiʻi. Un 
Museum Special Publication 97 á 

g 90b. 
~ ac E Sci. 44: 315— 

Weller, > K. Sakai. 1990. Bei of dicliny in he 

a Do. Caryophyllacaeae. Pl. Spec. Biol. 


a Wagner, D. 


Mania of the 


—35 in W. 
Sohmer 1 
. Hawaii Press (Bishop 


n and volcanology of the Hawaiian 
315 


83— 

& —— —. 1991. 
in Schiedea (C o pu an 
Heredity 67: 265-27: 

& 19 


999, seen shylogenetic approaches 
o is 


The genetic basis of male sterility 
endemic Hawaiian 


genus. 


for the malls of plant breeding system evolution. 
Annual Rev. Ecol. Syst. 30: 167-199. 

& —— 9005. Se ing and resource allocation in 
Schiedea salicaria (Curyopl lace ie). a gynodioecious 
species. J. Evol. 3 18: 301 — 

roses "i D R. Herbst. 1990. 
E solution of ho | Schiedea (C Eri eae: Alsi- 


noideae) in the 3 Isl ee Biogeographical anc 
ecologic al factors. Syst. Bot 266—210. 


„Wagner & A. K. S 1 


mu of Schiedea and . Alsinidendron: (C 1 eae: 


phyloge “netic 


Alsinoideae): Implications for th he evolution of breeding 
systems. Syst. Bot. 20: 3k 5-337 
S "Straub, 1996. 


Allozyme dive RI 


and ic Alsinidendron 


(Caryophyllaceae: Alsinoideae) in Hawaii. Evolution 50: 
23-34 
2—————, A. E. Rankin. A. Golonka, B. Kutcher & K. 

E. Ashby. 1998. dem and the evolution of wind 

pollination in Schiedea and Alsinidendron (C ae 

ceae: Alsinoideae) in the Hawaiian Islands. Amer. J. Bot 
85: 1377-1388. 

————, — 0D. A. Thai, J. Tom * A. E. Rankin. 2005. 
N . RE heterosis in populations of 
Schiedea viscosa, a highly selfing spec 1 1 Evol. Biol. 18: 
14 


34-1444. 


: D. R. Campbell, T. M. Culley & A. Dunbar- 
lis. | ress. Predic 5 dhe pathway lo wind 
pollination: heritabilities and genetic 1 of 
inflorescence trails associated with wind ar in 


J. Evol. 


Schiedea salicaria (C a eae). 


CONTRASTING PATTERNS AND Bruce G. Baldwin? 
PROCESSES OF EVOLUTIONARY 

CHANGE IN THE TARWEED- 

SILVERSWORD LINEAGE: 

REVISITING CLAUSEN, KECK, 

AND HIESEY'S FINDINGS 


ABSTRACT 


Jens €. Clausen, David D. Keck. and William M. Hiesey's biosystemalie research on continental tarweeds (Madiinae: 
[MENS provided. diverse Ea s of evolutionary change for Clausen’s synthesis, Stages in the Evolution of Plant 
Species. Subsequent anatomical work by Sherwin Carlquist demonstrated that the tarweed lineage also includes a spectacular 


example of adaptive radiation, the Hawaiian silversword alliance. Molecular phylogenetic data and evidence from genetic and 
dw cd studies have allowed additional perspectives on E lausen et als and Carlquist's hypotheses of. tarweed— 
silversword evolution. In Californian Lavia, Clausen et al.’s evidence for 0 loteo nn rsification for then = 7 taxa 
accords with patterns of molecular divergence and decay o ] inte " tility across lineages inferred from a rate-constant rDNA 
tree. In contrast, recent evidence on patterns and timing of diversification in an n = 8 Layia clade indicates multi ple examples 
of accelerated phenotypic evolution, unresolved by Clausen et al.. that evide ide eflect rapid "budding off" of morphologically 
distinct lineages in ecologically novel settings. In rDNA trees of € alifornian Halo Pa E a 
eryptic biological species. documented by Clausen, appear to predate the origin of a mor| ear all and ecologically 


distinctive taxon (M. macradenia (DC.) Greene) that retains inte de silty with relatives of ance dine phenotype: at fine-scale 
levels of divergence, a disconnect is evident between evolution of intrins „ post-mating reproductive barriers and phe notypic 
evolution in Holocarpha. Clausen’s evidence for strong interste "m 10 rs between the mostly annual, continental species of 
the “Madia” lineage contrasts with Gerald D. Carr and Donald W. Kyhos's subsequent finding of partial to full interfe ruility 
between the phenotypically disparate, insular species of the 1 silversword alliance, a monophyletic group that 
descended from continental ancestors in the “Madia” lineage. Molecular phylogenetic ms indicating major ecological 
uh 
silversword alliance n Carlquist's pus thesis of adaptive radiation of the group and help explain the lack of substantial. 
internal barriers to gene flow across lineages 


changes associated with diversification. a brief timeframe for diversification. and a shift to woodiness in the ancestry o 


verein. Results of recent investigations have shown that highly dynamic 
evolutionary change in Madiinae. both in phenotypic characters and in modes and patterns of diversification, extends to even 
finer-scale evolutionary levels than indicated by Clausen et als ele want studies. general, current evidence on 
dive rsification in Madiinae appears to be consistent ne Clausen et al^s views concerning the importance of ecological factors 

incipient evolutionary divergence. Phylogeny of Madiinae is no longer the intractable problem perceived by Clausen: 
relatively little is known about the biologic al basis for the extreme curi Bondi propensities of tarweeds, 

Key words: Adaptive radiation, cryptic diversity, diversification, edaphie endemism. Holocarpha, Layia. Madia. 
Madiinae, peripheral isolates. reproductive isolation. speciatio 


"Original research presented here was supported by the National Science Foundation (NSF-DEB 9458237). the Lawrence 
R. Heckard Endowment Fund of the Je p on Herbarium, and Roderic B. Park and other generous Contributors to the Jepson 
Herbarium. Figures J. 2. 6-8. 10, 11 nd 15 were illustrated by the late Jeanne Janish (lettering added). from Abrams. 
Leroy and Ferris; Roxana Stinchfie ld. TRE Flora of the Pacific States; Volume IV. Copyright © 1960 by the Board of 
Trustees of the Leland Stanford Jr. University. All rights reserved. Used with the permission of Stanford University Press. 
www.sup.org. Figures 19-24 were illustrated by Lesley Randall (lettering added). and Figure 25 was illustrated b Kare n Klitz. 
I thank Peter H. cen for inviting me to participate in the Missouri Botanical Garden's 50th Annual Syste S d 
where a version of this paper was | presented, and Peter R. Grant, B. Rosemary Grant. Jonathan B. Losos. Lore 
Rieseberg, Ann K. Sakai, Christopher " Schneider, David B. Wake, and Mary P. Winsor for sharing their rid Rer he Y 
insights on evolutionary diversification at the symposium. Falso thank 90 L. Strother for reviewing the manuscript, Bridget 


L. Wessa for lab assistance, Susan J. Bainbridge for field assistance, Rexford E. Palmer for providing living seed collections of 
Holocarpha, Gerald D. Carr for adapting Figure 3. and all Madiinae researchers, past and present, for their extensis 
contributions to ut m. ling of this fascinating group of plants. 

* Jepson He 1 and 9 0 5 nt of Integrative Biology. 1001 Valley Life Sciences Building 442465. University of 
California, Berkeley, CA 91720-2465, U.S.A. bbaldwin@berkeley.edu 


ANN. Missouri Bor. Garp. 93: 64-93. PUBLISHED ON BL May 2000. 


Volume 93, Number 1 


Baldwin 
Clausen, Keck, and Hiesey Revisited 


Contrasting modes of diversification. (speciation) 
have been suggested to explain evolutionary change 
within different lineages of subtribe Madiinae (Com- 
positae), a monophyletic plant group of 119 species 
(see Baldwin, 20034). 


belies the great morphological 


commonly known as tarweeds 
The name “tarweed” 
and ecological breadth of plant life encompassed by 
Madiinae, especially by the Hawaiian silversword 
alliance, a famous example of adaptive radiation (see 
2003b, c). The Californian or continental 


tarweeds, in general, are less diverse ecologically than 


Baldwin. 


the silversword alliance but contain about three-fold 


more recognized species and have been under 


evolutionary investigation far longer than have the 
A recent book, Tarweeds & Silver- 
of the 


includes extensive reviews on natural history, 


Hawaiian taxa. 


swords: Evolution Madiinae (Asteraceae), 


mor- 


phological and taxonomic diversity, chromosome 


— 


evolution, hybridization, glandular structures, leaf 
and wood anatomy, secondary chemistry, molecular 


evolution, conservation. genetics, evolutionary. rela- 


tionships, and processes of diversification in Califor- 
2003). 


first evolutionary studies of tarweeds began at 


nian and Hawaiian Madiinae (Carlquist et al., 
The 
the Carnegie Institution of Washington at Stanford. 


— 


California, with pioneering biosystematic investiga- 
tions by Harvey Monroe Hall and E. B. Babcock on 
DC. (Babcock & Hall, 1924) and by 
Hall and his two junior colleagues David D. Keck 
and William M. Hiesey 
Madiinae (Hall, 1932). Hall recruited. Danish. plant 
evolutionist Jens Clausen to join the research team at 
the € 
position. of 
following Hall's untimely death in 1932, and pursued 


Hemizonta 


on various lineages of 


Carnegie Institution in 1931; Clausen assumed the 


principal researcher. four months later, 


decades of tarweed research there as leader of the 
renowned Clausen, Keck, and Hiesey team (see 
French, 1989). 


Clausen, Keck, and Hiesey are best known for their 
reciprocal transplant studies along Hall's (1932) west- 
east ecological transect of California (e.g., Clausen et 
al., 1940, 1948), which helped establish the impor- 
tance of local adaptation in explaining variation in 
widespread plant species. Less widely appreciated are 
Clausen, Keck, 
understanding how diversity arises in plants and their 


and Hiesey's broader interests in 
extensive experimental studies on Californian. tar- 
weeds. In his synthesis, Stages in the Evolution of 
Plant Clausen (1951) showed that plant 
eroups recognized by taxonomists as species represent 
diversity of that 
in relative levels of morphological, 


Species, 


a wide evolutionary entities can 


differ 


ecological, 


greatly 


and genetic divergence and can arise 


by distinct processes. Californian tarweeds con- 


stitute about half of the evolutionary examples 


presented by Clausen in his 1951 book, which has 
been a major inspiration for generations of Madiinae 
researchers. 

Here, I revisit some of Clausen, Keck, and Hiesey's 
hypotheses of evolutionary change in Madiinae and 
evolutionary hypotheses of subsequent tarweed re- 
searchers who, at least in part, based their studies on 
previous l 
Carnegie team. Addition of a molecular phylogenetic 


[e 


work by one or more members of 
framework has proven useful for re-evaluating pre- 
viously suggested modes of diversification in various 
eroups of Madiinae, bearing in mind the limitations of 
phylogenetic data for inferring such evolutionary 
processes (see Losos & Glor, 2003). In light of data 
from those recent and ongoing studies, I suggest that 
the contrasts in processes of tarweed evolution so well 


Keck, 


comparisons across genera are far more extensive 


documented by Clausen, and Hiesey in 


than earlier appreciated. In addition, new molecular 
perspectives demonstrate that phylogenetic structure 


in Maditnae extends to finer-=scale levels than 


in part probably reflecting 
that 


Clausen et 


previously suggested, 


divergence of ecologically distinct lineages 


conform, al least in some measure, 


al.’s (1939, 1940, 1948) ecotype concept. 


EVOLUTIONARY CONTRASTS BETWEEN MAJOR LINEAGES 
OF LAYIA 


Gradual allopatric divergence in the n = 7 


Layia lineage. A widely studied hypothesis of 
diversification in the California tarweeds considered 
by Clausen, Keck, and Hiesey concerns the group of 
taxa in Layia Hook. & Arn. ex DC. with seven pairs of 
chromosomes. Members of Layia are spring ephem- 
erals that sometimes have white-tipped ray corollas 
and are commonly called “tidy-tips” (Figs. 1, 2; se 
Baldwin, 2003b, c). Clausen et al. (1941) ee 
the evolution of 


examining crossability and interfertility among all 


Layia extensively, in part by 
members of the genus. Their crossing diagram for 
Layia (Fig. 3) illustrates the intrinsic. potential for 
gene flow estimated from seed set and extent of 
chromosome pairing in F, hybrids. Clausen et al. 
(1941) regarded 
on the left side of the figure, as constituting 


axa with seven pairs of chromo- 


somes, 
one of two “major blocks of species” in Layia, which 
com- 


in turn was subdivided into three, informal * 


plexes,” each united by morphological characteristics 
and moderate levels of interfertility: (1) L. chrysanthe- 
moides (DC.) A. Gray and £. fremontii (Torr. & A. 
Gray ex A. Gray) A. Gray, (2) L. jonesii A. Gray, L. 
leucopappa D. D. Keck, and L. munzii D. D. Keck, and 
(3) L. platyglossa (Fisch. & Mey.) A. Gray. 
Stebbins (1949: 232) regarded the interfertile 


cC 
— 


— 


laxa in 


Annals of the 
Missouri Botanical Garden 


Figures | and 2. Widely sympatric, + intersterile members of wie with seven pairs of chromosomes. Figure | (left). 
Layia e ee —A. Base of plant. —B. Capitulescence. —C. dug ral bract (abaxial view). —D. Disk floret. 
—E. Recept 1c ae bract. Figure 2 (right). Lavia Maii pud —A. Habit. —B. Ray floret and enfolding 1 ral bract 
(adaxial view). . Involucre al bract (abaxial view). Disk floret. —E. Pappus elements (two forms). 


each of the first two complexes as “species in the state 


.. 


of becoming. 


Stebbins showcased Clausen. and 


(1966) Keck. 
Hiesey's (1941) biosystematic data on the members of 
Layia with n = 
divergence (see also Stebbins, 1982). Stebbins (1966) 
taxa of Layia with moderate interfertility 


l as exemplary of gradual, allopatric 


noted that 


are invariably allopatric: in contrast, taxa that are 


sympatric are moderately to highly intersterile, with 
reduced chromosome pairing at meiotic metaphase | 


in Fy hybrids (Fig. 4). 


interfertility 


Under the assumption that 
of common 
1985). 


consistent with the 


degree of reflects recency 


ancestry (a risky premise; see Donoghue, the 


patterns shown Figure 4 are 


hypothesis that sympatric taxa underwent earlier 


from a common ancestor compared de 


- 


divergence 
allopatric taxa or populations in each of the three 
subgroups. and that sympatry is secondary, following 
a period of lineage divergence in geographic isolation. 
Warwick Gottlieb (1985 


hypothesis of gradual. allopatric divergence among 


and re-examined the 


EN 


the n = 7 members of Layia from the perspective of 


allozymes. Their results mirrored the findings of 


Clausen, Keck, and Hiesey (1941) and the scenario 


Stebbins (1966) by 
reduced genetic similarities in pairwise comparisons 


ihe 


outlined by showing highly 


among three subgroups. relative to genetic 
identities within each group. 
A molecular phylogenetic perspective on diversifi- 


the n = 7 members of Layia based on 
(rDNA) internal transeribed 
2003a. 
unpublished) reinforces the evolutionary conclusions 
of Clausen et al. (1941). Stebbins (1966), 
Warwick and Gottlieb (1985). Monophyly of the n = 


7 group is well supported, as are each of the three 


ion of 


ca 
nuclear ribosomal DNA 
Baldwin, 


spacer (ITS) sequences (Fig. 5: 


and 
subgroups (“complexes”) proposed by Clausen et al. 


(1941). 


spond well to Clausen et als (1941) model of gradual. 


The tree topology and. branch-lengths corre- 


allopatric diversification, with sympatric taxa being 


deeply divergent relative to allopatric members of 


each subgroup and in comparison lo lineages i 


the (n. = 8) sister-group to the n = 7 lineage. The 


results provide no evidence of accelerated loss of 
interfertility within any part of the n = 7 group. 


wherein internal barriers to gene flow may well reflect 
byproducts of gradual evolutionary divergence. Not 
the 


shown in Figure 5 is finding (Baldwin, un- 


Volume 93, Number 1 
2006 


aldwin 
Clausen, Keck, and Hiesey Revisited 


jonesii 
2, 


2 ) 
N. m 
Va 


leucopappa//A E we 


E 


xm 


chrysanthemoides y 


[2277 D 
Y 


18 


2085 
heterotricha 


Crossing diagram of Layia, adapted from Clausen, Keck, and Hiesey (1941). Circles represent different species; 


ure 3. 
size of circles roughly illustrates relative abundance of each species. Width of shaded connections between species indicates 


estimated intrinsic potential for gene flow based on seed set in Fy hybrids. Width of lines connecting species indicates degree 


of chromosome pairing at meiotic metaphase | 


in F, hybrids; number(s) of meiotic chromosome pairs in Fy hybrids or 


vegelative condition of non-flowering hybrids are presented along lines. Dotted lines between species or groups of species 


indicate major discontinuities in morphological variation. Used 


of Washington. 


published) that rDNA sequences from different, often 
widely separated populations of each species were 
resolved within a common, monophyletic lineage, as 


expected under Clausen et als (1941) evolutionary 


mode 
The n — 7 lineage of Layia evidently conforms to 
Clausen's view of the “most normal pattern of 


„554 . 
speciation,” with “a more or less simultaneous and 
gradual separation in morphologic, ecologic, genetic, 
and cytologic characteristics” (Clausen, 1951: 90). 


Groups that exhibit approximately rate-constant di- 


by permission of the publisher, Carnegie Institution 


vergence of characteristics across lineages present 
minimal problems for systematists; even application of 
phenetie criteria should provide accurate reconstruc- 


^ 


tions of relationships under such conditions (se 
Felsenstein, 2004). These attributes are putatively 
reflected by congruence between Clausen et al/s 
(1941) phylogenetic hypotheses for the n = 7 Layia 


group (based on essentially phenetic considerations of 


morphological similarity, chromosome-number simi- 
larity, and levels of meiotie chromosome pairing and 


interfertility) and clades resolved from molecular 


68 Annals of the 
Missouri Botanical Garden 


Distribution and interfertility of six species of Layia 
Data from J. Clausen, 1951 


—— — o — 


t e 


L 5 e 
L. fremo 


Lt . jonesii, L. munzii 
L. leucopappa 
[ B platyglossa 


4) Meiotic pairing, F, hybrid = 711 Fertility 25-30% 
> Meiotic pairing, Fi hybrid = 5-71] Fertility 5-20% 
4> Meiotic pairing, Fi hybrid = 2-6ll Fertility 0.5-2% 


Y 


Figure J. Distributions and interfertilities of species of Layia with seven paisg Occ. Letters alongside species 
dist 1 e 1 to the first letter(s) in ips ies epithet of those taxa (C „a Chrysanthemoides; F = L. fremontii: | = 
jon esii: I. cxx iS 5 -— L. munz TR P | = +p alyglossa). Ni ole mode ‘arate lo high inte rte ruility of most allopatric s spe 8 16 S 


and low interfertility of widely sympatric species. d printed from Stebbins (1966), by permission of Pearson Education, U Ipper 


Saddle River, New Jerse 


phylogenetic analyses. As noted above, the three n = Accelerated phenotypic divergence in n = 8 
T “species complexes” proposed by Clausen et al. Layia lineages. In contrast to ene ſor gradual. 
(1941) correspond precisely to clades 1-3 in Figure 5: allopatric divergence in the n = 7 Lavia lineage. 


in addition, Clausen (1951: 129-130) indicated that experimental and molecular Jata from the congeneric 


“Layta platyglossa is almost a species complex inn = 8 sister-lineage have revealed multiple examples 
itself... It 


is genetically sharply separated from of taxa that likely arose rapidly in peripheral or 
oo. and Fremontii and...is more close- otherwise isolated, ecologically distinct settings. 
ly related.. . leucopappa, Munzii, and Jonesii^. in Phylogenetic and phenotypic patterns and timing of 
complete Mrd with the rDNA tree topology diversification resolved in two n = 8 sublineages are 
(Figure 5). similar. at least in part. to expected outcomes of 


Volume 93, Number 1 Baldwin 69 
2006 Clausen, Keck, and Hiesey Revisited 


L. platyglossa 
L. platyglossa 
7 M 
L. platyglossa 
3A platyg 
| 6-7 Il L. platyglossa 
711 L. jonesii 
| . munzii 


n=7 2-6 Il 2 ^ 
Layia —» p 


group A/S . leucopappa 


JE 
L 
L. chrysanthemoides 
7 M mL L. chrysanthemoides 
ba | 
L 


. fremontii 


. fremontii 


p L. glandulosa 
E. e ió 8 Il L. glandulosa 
iu P L. discoidea 
L. pentachaeta 
L. pentachaeta 
_ ae A 


. gaillardioides 


L. gaillardioides & relatives  gaillatdioltce 
n= 


L 
L 
L. hieracioides 
L. hieracioides 


A = allopatric 


( L. carnosa 

P = parapatric 
S = sympatric L. carnosa 
L. carnosa 


—— <80% bootstrap 
=m >80-90% bootstrap 
ö >90% bootstrap 


L. septentrionalis 


L. septentrionalis 


L. heterotricha 
y 


| L. heterotricha 
— — 0.005 substitutions/site 


Figure 5. Phylogenetic hypothesis for Layia based on nuclear 185-265 rDNA . transcribed spacer sequences 


(Baldwin, 2003a, unpublished); one of two maximally o trees ( = clade unresolved in strict consensus tree). 
Branch lengths correspond to relative time since divergence, under maximum-likelihood 1 iion: rate constancy of ITS 
evolution across lineages could not be rejected using Felsenstein’s (1988) likelihood-ratio test. In then = 7 group, clades 1A, 
2A, and 3A represent Clausen et al.’s (1941) three “complexes” of allopatric species, disc Based in the text. Numbers of meiotic 
chromosome pairs in Er hybrids between taxa in a clade are indicated above branches (Lh = hybrid with diploid £. 
hieracioides; Le n oe with L. carnosa; NH = no hybrids reported), based on Clausen et als (1941) crossing data for Laya 
(see Fig. 3). The tree was rooted. using Árnica mollis Hook., Hulsea algida A. Gray, Raillardella pringlei Greene, and 
Ide nisi eile alas (Brandegee) D. D. Keck as the outgroup. Terminal taxa of the same name represent geographically 


disjunct populations. See Appendix | for authorship of species names not mentioned in text. 


70 Annals of the 
Missouri Botanical Garden 


*— 8 
a- = 
— 


als 
2 O 
— 
€ 
>. 
= 
N 


se = 
Shak 


2 
23 


UJ 


Figures 6 and 7. Ecologically distinct. interfertile members of Lavia with eight pairs of chromosomes. Figure 6 (left). 


La yia discoidea. A. Habit. —B. Fruit with pappus. —C. Pappus element (: dr view). —D. Pappus element (lateral view). 
Figure 7 (right). Layia 1 —A. Habit (bisected). —B. Involucral bract . Fruit with pappus. —D. Pappus element 


(lateral view). 


g 


peripatrie speciation (Mayr, 198: 


2), quantum evolution examples of 
(Simpson, 1944). or quantum speciation (Grant, 1963 


1971: Lewis, 1966), as discussed below. 


species pairs in plants. that represent 


putative examples of lineages wherein one species 

descended from ancestors in another: he discussed 
Evolutionary divergence on a local geographic evidence for budding-off of three annual diploid 

scale in general has been widely suggested to be of species of angiosperms. including an n = 8 member 

major importance in plant evolution (e.g.. Ehrlich & of Lavia, I. discoidea D. D. Keck 

Raven, 19609: Levin, 1993), although convincing The best-studied hypothesis of rapid evolution 

phylogenetic evidence for a putatively common among the n 


O Layia taxa concerns the origin of 
pattern of local diversification. i.e.. “budding off of I discoidea (Fig. 6). which is morphologically so 
atively unusual that it was not. initially 
uniform set of related lineages, is usually lacking in a tarweed 


distinctive, new lineages from within a re 


— 


recognized to be 


and was about to be described in 
studies of suspected “progenitor-derivative” species 
pairs (see Gottlieb, 2004). Expected. paraphyly of Roxana S. Ferris and Ira L. Wiggins)—until Clausen 
a widespread “progenitor” species relative to a “de- ct al. (1941 


monotypic genus (“Roxira.” for co-discoverers 


i 


) observed some similarities with tarweeds 
rivative isolate in some models of local diversifica- anch subsequently found that the plant was completely 
tion may fail to be resolved in phylogenetic analyses — interfertile with £. glandulosa (Hook.) Hook. & Arn. 
because of insufficiently high rates of evolution in (Fig. 7). Despite being fully interfertile. L. glandulosa 
the characters under study, extinction or inadequate and J. discoidea differ 
sampling of populations (see Neigel & Avise, 1986), 


or gene flow among lineages within the paraphyletic 


greatly im morphology and 
ecology. For example, like most tarweeds. L. glandu- 


losa has ray florets and involucral bracts: L. discoidea 


progenitor taxon (e.g. for genes under strong 
selection: see Rieseberg & Burke. 2001). Based on marginal receptacular bracts constituting a 


various theoretical considerations. Rieseberg & voluere. The two taxa 


acks both ray florets and true involueral bracts. with 


alse in- 
also differ substantially i 
Brouillet (1994) suggested that paraphyly of plant features of the pappus. 


1 
Layia glandulosa occurs 


species may be common. Gottlieb (2004) listed 31 mostly in coarse, sandy soils. including sand dunes, 


Volume 93, Number 1 
2006 


Baldwin Al 
Clausen, Keck, and Hiesey Revisited 


in semi-arid and montane regions of California and the 


western deserts; I. glandulosa has the widest 


distribution of any obligate outcrosser in Madiinae 
(see Baldwin, 2003b, 


serpentine barrens in a small region of the Inner South 


c). Layia discoidea is confined to 


Coast Ranges of California (San Benito and western 
Fresno counties). Serpentine erodes to clay (structur- 


has 


ally unlike the sandy habitats of L. glandulosa 
a high magnesium/calcium ratio unfavorable for most 
plants, is low in some essential plant nutrients, and 


contains concentrations of heavy metals that can be 


toxic to angiosperms (see Kruckeberg. 2002). Ap- 
proximately 9% of minimal-rank vascular plant 


taxa (species, subspecies, and varieties) endemic. to 
Province are confined to 


1992) 


interpreted either as products of recent, in 


the California Floristic 
serpentines (Kruckeberg. and have been 
silu 
evolution (neoendemics) or as evolutionary relicts 
(paleoendemics) that became secondarily restricted to 
serpentine habitats, possibly by competition or in- 
terference with other taxa, in association with late 
Cenozoic climatic changes (see Raven & Axelrod, 
1978). 

On the basis of morphological and ecological 


Clausen (1951: 82) that 


Layia discoidea was “probably an edaphic race (of L. 


considerations, concluded 


glandulosa) adapted to the serpentine soil, and 


possibly an ancient relict.” Clausen et al. (1947: 
121 


the two taxa: 


were unable to resolve the relationship between 


— 


"There is no way of determining which 
race was first, the rare, inconspicuous one from the 
serpentine (L. discoidea), or the common, showy one 
from sandy habitats (£. glandulosa).^ 

In the 1980s, Gottlieb and colleagues revisited the 


origin and age of Layla discoidea. Gottlieb et al. 
(1985) found that genetic similarity between 7. 
discoidea and L. glandulosa, based on allozymes, 


was almost as high as values typical of conspecific 
plant populations. They concluded that the serpentine 
endemic L. 
a peripheral isolate that diverged relatively rapidly 
and recently in an ecologically marginal setting from 
an ancestor referable to the geographically widespread 
(1989, 1990) 


corroborated Clausen et als (1947) conclusion that 


L. glandulosa. Ford and Gottlieb 


only two major genes are responsible for presence or 


abse nce of ray florets and associated involucral bracts 


L. glandulosa and L. discoidea and showed that 


morphological differences between the two taxa are 

controlled by “a complex admixture of genes with 

laree and small, qualitative and quantitative effects” 
B 1 | 


(Ford & Gottlieb, 1990: 44). 


In the ITS trees for Layia (Fig. 5), resolution of 


a clade comprising I. discoidea and L. glandulosa 


corroborates Clausen et al.’s (1947) hypothesis that 


discoidea represents an example of 


the two phenotypically dissimilar taxa constitute 


a natural group. Strong support for the nested 
placement of L. discoidea within L. glandulosa 


corroborates Gottlieb et als 


(1985) hypothesis that 
L. discoidea descended from a lineage of L. glandu- 
losa. Gottlieb et al.’s (1985) additional conclusion that 
L. discoidea represents an example of relatively recent 
evolution is also evident from the ITS tree: the branch 
separating L. discoidea from the most closely related 
lineage of L. glandulosa in the rate-constant ITS tree 
length to branches of 


is comparable or shorter in 


conspecifie sequences in other members of Layta, 
e.g., I. pentachaeta A. Gray, the sister-species to L. 


glandulosa and L. discoidea. 


— 


A more extensive phylogenetic analysis of relation- 
ships within the Layia glandulosalL. discoidea lineage 
based on expanded sampling of populations and 
sequences, with combined ITS and ETS (external 
transcribed spacer) data, yielded a nearly noise-free 
tree that reinforces support for recent descent of L. 
discoidea from a sublineage of L. glandulosa (Bz ildwi in, 
2005). [. 


discoidea sequences of [. 


Sequences of the five populations of 
sampled are identical: 
glandulosa are heterogeneous and are resolved. as 
a grade of lineages that are geographically distinct 
and, to some extent, morphologically divergent. For 
example, the sister-group to all other lineages in the £. 
glandulosa/L. 


and ecologically discrete group endemic to coastal 


discoidea clade is a morphologically 
sand dunes. Phenotypic disparity among the lineages 
of L. 


morphological and ecological differences between L. 


glandulosa is minimal, however, compared to 


discoidea and relatives in L. glandulosa. Vine-scale 
patterns resolved in the ETS+ITS tree are consistent 
Gottheb’s (1990) that 


yellow-rayed populations of L. glandulosa, sometimes 


with Ford and conclusion 
treated as L. glandulosa subsp. lutea D. D. Keck, are 


the closest relatives of L. discoidea, which lacks ray 
florets but has one or more modifier gene(s) for vellow 
ray-corolla color; most populations of L. glandulosa 
have white ray corollas and all but one of the sampled 
white-rayed populations are resolved, based on the 
molecular data, as more distantly related to L. 
discoidea than are the sampled yellow-rayed popula- 
tions. 

As noted by Gottlieb (2004), the evolutionary origin 
of Layia discoidea did not involve the major genotypic 
effects predicted under quantum speciation (Grant, 
1981); complete interfertility with L. glandulosa and 
the highly contrasting edaphic settings of the two taxa 


implicate ecological selection without radical genetic 


— 


reorganization in divergence of L. discoidea, ii 
keeping with some recent, general views on acceler- 
ated evolution in peripheral populations (e.g.. Barton 


& Charlesworth, 1984; Coyne, 1994). Mayr's (1982: 5) 


72 


D 


B 

M. 
in) ae X 

> 


A 


Figure 8. Layia gaillardioides (n = 8). —A. 
bract (abaxial view). —D. Involucral bract (lateral vie w). — 


(lateral view). 


refinement of his earlier ideas on rapid divergence of 
peripheral isolates founders (“peripatrie specia- 
tion”) allowed for the possibility of “rapid change... 


increased selection 


due more to greatly pres- 
sures... than to the genetic consequences of in- 
breeding," in closer accord with the proposed 


evolutionary scenario for discoidea. Proximity of 


habitats of L. 


even leaves open the possibility that divergence of L. 


potential glandulosa and L. discoidea 


discoidea may have occurred in the presence of 


potential for gene flow with I. glandulosa. although 
adjacency of populations of the two taxa has not been 
documented and no evidence of hybridization is 
apparent from the rDNA sequences, 

A less widely studied example of rapid evolutionary 
change among the n = 8 Layia 


gaillardiovides (Hook. & Arn.) DC. 
relatives. Clausen (1951: 


^^ 


laxa concerns L. 
(Fig. 8) and close 
16) regarded the extensive 
interpopulational variation in ray corolla color, 
vegetative morphology, and flowering time within J. 
In part based on 
thal 


gaillardivides. 


gaillardioides as “very spectacular.” 


common-garden studies, he concluded some 


variation among populations of 7. 


e.g., In leaf margins (Fig. 9), is heritable and follows 


a west-east geographic pattern “correlated with 
ecological differences between the outer and inner 
Coast Range” (Clausen, 1951: 14-15). Phylogenetic 


analysis of rDNA sequences of Layia corroborates 


FA 


Stem segment, 


. Ray floret (adaxial view). —F. 


Annals of the 
Missouri Botanical Garden 


—B. Capitulescence, with peduncle det: Tm . Involucral 


Disk floret. — dE ue 


Clausen’s (1951) suggestion that west-east interpop- 


ulational differences in L gaillardioides | reflect 
genetic divergence, though at a deeper evolutionary 
level than he proposed. Four major. allopatric 


evolutionary lineages within L. gaillardioides are well 
supported by rDNA data and three of the four lineages 
replace one another from west to east through the 
Coast Ranges (Baldwin, unpublished). 

The four semi-cryptic, allopatric lineages of Layia 
gaillardioides do not constitute a clade in rDNA trees: 
the morphologically and ecologically distinct „. 
(Nutt.) Torr. & A. Gray, L. 
(DC.) Hook. & Arn., and L. 


Keck are nested among lineages of L. 
[e] fe 


carnosa hieracioides 
septentrionalis D. 

gaillardioides 
and represent examples of accelerated 


gaillardioides-like 


Fig. 5 


appear to 


phenotypic divergence from £. 


ancestors (Baldwin, unpublished: includes 
representalives of three of the four major lineages of 
L. gaillardioides). 10). a 


endemic of outer coastal foredunes. occupies a mar- 


Layia carnosa (Fig. rare 
ginal ecological setting where L. gaillardivides does 
that J. 
the 


not occur. \ hypothesis carnosa. arose as 


peripheral isolate al coastal-most limits of 


é 
a widespread progenitor is consistent. with distribu- 
lions of the two closest relatives of “beach lavia” 
(I) a lineage of L. 
windward slopes of the North Coast Ranges and San 


11). 


gaillardioide 5 B occurs along 


Francisco Bay Area and (2) I. hieracioides (Fig. 


Volume 93, Number 1 
2006 


Baldwin 73 


VIII 
di 


un 


SARATOGA SUMMIT 


MUIR BEACH 


Figure 9. Bas 
individual grown unden common-garden 
sented at left 


are repre three populations from the 


equates to al to- south distribution of populations). Note deeper leaf- 
Evolution of Plant Species, 


the state. Reprinted from Stages in the 


Used by permission of 


— 


the publisher, Cornell Univ. Press. 


which oceurs widely in the South Coast Ranges, in 
part on sandy soils and even coastal backdunes. Self- 
and £L. hieracioides 


compatibility in I., carnosa 


(Clausen, 1951)—otherwise unknown in Layia—is 
associated with some conspicuous | morphological 
differences from other members of the L. gaillar- 
reduced sizes of ray corollas and 
shift 


dioides clade (e.g.. 
The 


inbreeding mating system in the self-compatible taxa 


heads). putative toward a more highly 


can be reasonably suspected to have promoted 


— 


evolutionary change, as suggested by Gottlieb (1973 
the self- compatible Stephanomeria 
Gottlieb 


a well-studied example of local. 


for origin of 


malheurensis (Cichorieae; Compositae), 


rapid phenotypic 
(see 


divergence from self-incompatible ancestors 


Gottlieb, 2004). Layia septentrionalis, the other taxon 


he 


— 


nested among lineages of L. gaillardioides in 


sal-leaf variation within and among six populations of Layia gaillardioides. 
conditions at Stanford.“ 


KNOXVILLE 


Clausen, Keck, and Hiesey Revisited 
| ISABEL CREEK 
| LEWIS CREEK 
Each leaf represents a different 


Three populations from the outer Coast Ranges of California 


inner Coast Ranges of California are represented at right (top-to-bottom 


— 


obing in populations from the drier, hotter interior of 
by Jens Clausen. Copyright 9 1951 by Cornell University. 


outcrossing, 


rDNA (Fig. 5), 


narrowly distributed plant of poor, sandy or serpentine 


trees is an obligately 
soils that occurs in interior regions of the North Coast 
Ranges, where populations of one major lineage of L. 
gaillardioides occur sporadically; the two taxa have 
not been reported to occur together. 

Parallels between patterns of evolutionary change 
in the L. the L. 
glandulosa/L. discussed above, do 


not extend to patterns of interfertility. In contrast 1 


gaillardioides assemblage and 


discoidea lineage, 


) 
complete interfertility between members of L. glan- 
members of L. gaillardioides 


dulosa and L. discoidea, 


are largely to completely intersterile with members of 


L. carnosa and L. septentrionalis and are largely 
interfertile with diploid members of L. hieracioides 
(Fig. 3; Clausen, 1951: Baldwin. unpublished). 


Consideration of the observed interfertility patterns 


74 Annals of the 
Missouri Botanical Garden 


NDS 


Figures 10 and . Self-compatible taxa in Layia with eight pairs of ee pal Figure LO (left). Lavia carnosa. 
. Habit. —B. Involucral bract (abaxial view). —C. Ray floret (adaxial view). —D. Disk floret. —E. Pappus element. 
Figure 11 (right). Layía hieracioides. —A. chc m, leaves, and roots. —B. E —C. Ray floret and 
enfolding involucral bract (adaxial view). —D. me bract (abaxial view). —E. Disk floret. —F. Pappus element. 
in a phylogenetic context leads me to conclude that J The n = 8 Layia clade encompassing L. glandu- 
carnosa and L. septentrionalis underwent accelerated — losa/L. discoidea and the paraphyletic L gaillar- 
evolution of intrinsic, post-mating reproductive bar-  dioides and relatives. stands out in Layia as an 
riers, as expected with quantum speciation (Grant, exceptional example of how levels of interfertility or 
1971). although no evidence of chromosome repat- morphological divergence can be misleading about 
terning has been detected in Layia (Clausen, 1951), phylogeny, unlike in the n = 7 Layia clade, where an 
unlike in such well-studied, plant examples of quan- evidently different mode of evolutionary divergence 


tum speciation as Clarkia lingulata Harlan Lewis € — led to congruent patterns of interfertility, morpholog- 


M. Lewis (Lewis & Roberts. 1956: see Gottlieb. 2004) ical similarity, and re latigaship; Whether such stark 


and Stephanomeria malheurensis (Gottlieb, 1973, — contrasts between the n = 7 and n = 8 sister lineages 
2004). Clausen (1951: 130) concluded that I carnosa in patterns of evolutionary change are attributable to 
and L. septentrionalis each represents a “monotypic intrinsic or extrinsic factors in the history of the two 


species complex,” largely based on intersterility or groups remains unknown. The principal conclusion 
lack of crossability with other layias. In addition, that has emerged from molecular phylogenetic ri 


(1941) regarded L. carnosa as highly search previewed here is that evolutionary processes 
l B BH yd 


Clausen et al. 
divergent morphologically from other species or in Layía have been more dynamic than was evident 


groups of species in Layia (Fig. 3). The major based on Clausen et als (1941) biosystematic data 
morphological discontinuity between I, carnosa and alone. Evidence for multiple instances of accelerated 
other members of Layia perceived by Clausen, Keck, — phenotypic divergence in Layia underscores the 


and Hiesey now appears to reflect accelerated potential importance of peripherally isolated popula- 
morphological evolution in L. carnosa, as documented tions in plant evolution (see Rieseberg & Brouillet, 
in the origin of L. discoidea (Ford & Gottlieb, 1990), — 1994; Gottlieb, 2004). For systematists, the possibility 
rather than a distant evolutionary relationship with that diversification is often marked by major changes 


other members of the same genus. in rates of phenotypic evolution and rates of decay of 


Volume 93, Number 1 
2006 


Baldwin 75 
Clausen, Keck, and Hiesey Revisited 


interfertility across lineages, as inferred for the 
principal n = 8 Layia clade, makes risky a reliance 
on phenetie considerations for estimating relation- 
ships or circumscribing taxa, as noted earlier (e.g.. 


Donoghue, 1985). 


CRYPTIC BIOLOGICAL SPECIES AND RAPID PHENOTYPIC 
EVOLUTION IN HOLOCARPHA 
As noted by Clausen (1951: 94): “There is one 


genus of the tarweeds of California which differs 


= 


strikingly from the pattern of speciation in Layia of 


the same subtribe of the Compositae in that it has very 
strong barriers of sterility even between neighboring 
populations of one species. These populations are so 
similar in external appearance that they cannot be 
distinguished. This genus is Holocarpha.” Holocarpha 
Greene also differs greatly from Layia in ecology. For 
example. members of L, escape summer drought 
by flowering and fruiting during the wet spring 
months: most members of Holocarpha tolerate summer 
drought and flower in summer and fall (see Baldwin. 
2003b, c) 

Clausen (1951) emphasized that reproductive iso- 
lation in plants can be achieved by different means. 


He provided examples of “groups having predomi- 


nantly ecological barriers” (e.g.. physiological/habitat 


“predominantly morphological differen- 
floral 
“predominantly genetic 
90—107) in 


wherein 


differences). 
tiation” (e.g.. in features associated with 
pollination), and barriers” 


1931s 


typical species” 


(Clausen, opposition to "more 
“the pattern of develop- 
ment... from ecological races is through small steps 
and chromosome systems and 


1951: 107). 


Clausen's (1951) main example of diversification 


involving the genetic 


the morphological characters? (Clausen. 


dominated by the evolution of genetic barriers to 
gene flow across lineages was Holocarpha. a tarweed 
genus with four taxonomic species (recognized on the 
basis of morphological characteristics and chromo- 
some numbers) and an indefinite number of un- 
recognized, cryptic biological species. 

Clausen, Keck. and Hiesey's experimental studies 
in Holocarpha revealed that interpopulational crosses 
within each of two 5 taxonomic species. H. 
Keck and H. 
Gray) D. D. Keck, were either unsuccessful or yielded 
12: Clausen, 1951). 


Chromosomal studies revealed karyotypic differences 


heermannii (Greene) D. D. virgata (A. 


inviable or sterile hybrids (Fig. 


between intersterile, often morphologically indistin- 
guishable plants from different populations (Fig. 13: 
1951). Clausen (1951: 99) concluded that 


chromosome evolution via structural rearrangements 


Clausen, 


had resulted in each population of H. heermannii and 


H. virgata becoming “a breeding unit by itself anc 
e o o y 


genetically sharply separated from its neighbors." 
Hybrids between members of H. virgata (Fig. 14) and 
a rare species, H. macradenia (DC.) Greene (Fig. 15). 
exhibited modest fertility, in contrast to intersterility 
between members of different “conspecific” popula- 
tions of H. Palmer (1982) conducted addi- 


tional cytogenetic and morphological 


virgata. 


studies it 


Holocarpha that extended Clausen’s (1951) charac- 
terization of patterns of interfertility and chromosomal 
evolution in the genus. 


An ongoing molecular phylogenetic study of 


Holocarpha (Baldwin, unpublished) has revealed 


evidence of a disconnection between evolutionary 


divergence in conspicuous morphological and ecolog- 


ical characteristics and the origin of intrinsic, post- 
mating reproductive barriers in the group. Well- 
supported rDNA trees based on ITS and ETS 
sequences have confirmed that each of the three 


widespread species recognized by Clausen, Keck, and 
Hiesey. i.e.. H. heermannii, H. obconica (J. C. Clausen 
& D. D. Keck) D. D. Keck, and H. virgata, descended 
rare H. 


descended from an ancestor within (paraphyletic) A. 


from distinct ancestors; the macradenia 


virgata (Fig. 16). Detailed comparisons of the relative 
timing of evolutionary changes in Holocarpha are now 
possible because of minimal noise in the molecular 
dataset, phylogenetic structure both among and within 
recognized and inability to 


species, reject rale 


constancy of molecular evolution across lineages 
using Felsenstein's (1988) likelihood-ratio test. 

Based on branch-length comparisons across the 
e-constant rDNA trees, 


distinctive Holocarpha macradenia from an ancestor 


of the highly 


ra divergence 
within H. virgata was more recent than evolutionary 


divergence of some intersterile lineages 


within H. 


origin of H. 


cryptic, 
16: Baldwin. 


macradenia is consistent. with 


virgata (Fig. unpublished). 
Recent 


modest interfertility between H. macradenia and some 


populations of H. virgata, in contrast to strong 
intersterility barriers between most studied popula- 
1951: Palmer. 1982). In 
short. extensive morphological and ecological evolu- 


tion of H. 


development of 


tions of H. virgata (Clausen, 


macradenia was not accompanied by 


intersterility with closely-related 


members of virgata; conversely, evolution of 


intersterility between “conspecific” lineages within 
H. virgata (and H. heermannii) was accompanied by 
relatively minor or insubstantial phenotypic. diver- 


gence, despite longer evolutionary timeframes for 
such divergence to have occurred compared to the 
relatively rapid rise of H. macradenia. 

Based on available data, Holocarpha macradenia 
appears lo represent yet another example among 
continental tarweeds of a peripheral isolate thal 


underwent rapid morphological and ecological di- 


Annals of the 
Missouri Botanical Garden 


Greenhorn $ 


7! 


ener Crossin failed 
us imits o Species 


Figure. 12. Crossing diagram ol Holocarpha. Note 
populations in M. UIN and H. 
hybrids 


G copyright © 


vergence in a novel habitat. 


Holocarpha, which are found extensively in interior or 


southern Californian habitats with extremely dry. hot 
summer climates, H. macradenia occurs on coastal 
terraces in the San Francisco Bay and Monterey Bay 
areas, where summer drought and temperatures are 
greatly ameliorated by onshore winds and fog. Clausen 
(1951: 10 


virgata 


5) suggested that H. macradenia and H. 


“probably were interconnected as maritime 


and inland races of one species before the Coast 


Ranges arose.” He further posited that the rise of the 


d Ad osos Pass 
ts. 
White River: : 


strong. 
virgata, high interfertility among populations ol 
> between A. macradenia and H. virgata, Reprinted from Stages in the Evolution of Plant e by Jens Clausen. 
1951 by Cornell University. Used by permission of the publisher. Cornell Univ. Pre 


Unlike other members of 


A 


o only; no germination 


postmating reproductive barriers. between “conspecific” 


H. obconica, and partial fertility of 


Coast Ranges 7... provided the dry interior habitats 
interior 
Palmer (1982. 1987) 


macradenia 


now so completely exploited by the three 
1951: 105). 
instead that H. 


members of Hf. virgata that invaded the coastal region 


species” (Clausen, 


suggested arose from 
during an unusually warm period within the last 
10.000 vears. He based his hypothesis in part on 
patterns of. chromosome evolution and interfertility 
within and between the two taxa and on morphological 
considerations, 


including a suggestion by Carlquist 


(1959a: 305) that distinctive distal-leaf morphology of 


Volume 93, Number 1 
2006 


Baldwin 
Clausen, Keck, and Hiesey Revisited 


HEERMAN NII 


VV li 
VV Ll 
MN UVA 


VIRGATA 


VV UOI 
MN UI 


OBCONICA 


Figure 13. III 


and spe C le S of Holoc arpha (one c hromosome of eac h pair Is sho own). Ni ol 


of H. heermannii and H. virgata. Reprinted from MA | 


nthe Evolution bee nl Species. by Jer 


ustrations of somatic chromosome sets from root-tip pre parations for representatives of different populations 


e extensive chromosomal variation among populations 
s Clausen. Copyright © 1951 by 


Cornell University. Used by permission of the ableton Cornell Ui 


H. macradenia may represent merely “a juvenile 


form” of the distal-leaf condition in other taxa o 


Holocarpha. Molecular phylogenetic data are most 
consistent with Palmer's (1987) evolutionary scenario: 
the actual time of divergence of H. macradenia is still 
under investigation, 


As in the origin of Layia discoidea. 


evolutionary 


ivergence of Holocarpha macradenia was not accom- 
panied by acquisition of strong internal barriers to 
Keck, 
fertile Fə 
generation from crosses between H. macradenia and 


gene flow with close relatives: Clausen, and 


Hiesey produced a vigorous, partially 
H. virgata (Clausen, 1951). Given the great potential 
for chromosome evolution in Holocarpha, the lack of 


major chromosomal repatterning associated with 
evolution of the distinctive H. macradenia is consis- 
tent with the conclusion from studies of L. discoidea 
that genomic reorganization may be less important in 
rapid evolutionary divergence of peripherally isolated 


populations than has been widely suggested (see 


previous section). In contrast, the potential evolution- 
ary importance of ecological selection in the di- 
vergence of hypothesized peripheral isolates in. both 


Layia and Holocarpha is difficult to ignore: climatic 


and/or edaphic differences between habitats of 


paraphyletic “progenitor” taxa and nested "deriva- 


tive” lineages are extreme. In Holocarpha, ecological 
selection also may have played a significant role ii 


maintaining long-term phenotypic stability among 


morphologically cryptic, intersterile lineages within 


H. heermannii and H. virgata; genetic cohesion by 
g | 
gene flow cannot explain morphological stasis in each 


of the two taxa. 


ADAPTIVE RADIATION AND THE “Manía PATTERN” 


Clausen, Keck, and Hiesey's biosystematic inves- 
tigations in Madiinae led to recognition of extensive 


intersterility between species then treated as members 
of Madia Molina (Fig. 17). Clausen (1951: 134) stated: 


78 


Annals of the 
Missouri Botanical Garden 


ld and 15. 


Figures 


Figure 14 (left). Holoc s virgata. pitulescence. - 


ue (adaxial view). Figure 


E. Disk floret. —F. 


. Involucral bract (abaxial view). 


“The species of Madia are cytologically and geneti- 
cally much more isolated from each other than are the 


Layias....Most of the species of Madia can be linked 


together through hybrids. but these are sterile, and, 


moreover, many hybrids have unpaired, nonhomolo- 
gous chromosomes. There is, therefore, usually only 


In 


designating five general categories of fertility patterns 


one species lo a species complex in Madia.” 
in vascular plants, Grant (1971: 101) aptly considered 


the Madia 


with... 


pattern” exemplary of “annual herbs 


related species usually separated by incom- 
patibility barriers and by chromosomal and genic 
sterility barriers.” 
and Grant (1971). s 
plant 


As recognized by Clausen (1951) 
terility barriers in Madia and other 
do nol 


groups exhibiting the “Madia pattern” 


Clausen. 
Keck, and Hiesey (1945) demonstrated that y 
largely Madia 


o produce fertile, 


necessarily preclude reticulate evolution: 
1ZOrous. 
in had considerable 


sterile hybrids 


stable, allopolyploid 
“M. 


the wild species, M. citrigracilis D. D. Keck. 


potential t 


lineages, such as their synthetic nultrammii ^ and 


phylogenetic studies 


that 


Subsequent molecular 0 


Madiinae and relatives revealed Madia. as 


treated in all previous senses. was nol monophyletic: 


the smallest clade encompassing all species of Madia 


ecologically se ie d. partially inte ES rtile members of Holoc 9 0 with four B ole 1 s. 
-C. 


A. 
15 (right). itis arpha aria — 


. Involucral bract (a floret. Ras 


Habit. —B. 


hay fruit (adaxial view). 


axial view). Di 
Ste 


om se Sime ( Mei 


Ray iu E ri al view). 


sensu Clausen (1951: Fig. 18) also included plants 
that never had been placed in the genus: two 
perennials from montane California —Anisocarpus 


(5. = 
scabrida (Eastw.) Rydb.] 
(A, Gray) B. G. 


muirii (A. Gray) Rydb.] (Fig. 


Baldwin [= Raillardiopsis 
(Fig. 
Baldwin [= Kaillardiopsis 
20) 
Hawaiian silversword alliance, a monophyletic group 


30 


Ireyroxiphium DC., 


scabridus (Eastw.) B. 
19) and Carlquistia 
mulrti 


and the endemic 


woody and semi- woody 


comprising spec les i 


and Wilkesia 


Dubautia Gaudich.. 


V. Gray (Baldwin et al., 1991; Baldwin, 1992, 1996, 
2003a; Barrier et al.. 1999). The two Californian 
perennials (A. scabridus and C. muirii) and the 


silversword alliance were not shown convincingly to 
be members of Madiinae until publication of Carl- 
quist's (1959b) anatomical investigations, after Clau- 
sen, Keck, and Hiesey concluded their experimental 
Carlquist (1965, 1974. 


also championed the hypothesis that the Hawaiian 


studies of the subtribe. 


axa represent a major example of insular adaptive 
The 


shrubs, mat-plants, cushion plants, woody vines, and 


radiation. silversword alliance includes trees. 


rosette plants. spans most of the wide spectrum of 
habitats found in the Hawaiian Islands from dry serub 


< 400 rainfall/yr) to rainforests and bogs 


mm 0 


Volume 93, Number 1 
2006 


Baldwin 79 
Clausen, Keck, and Hiesey Revisited 


98. 


K 


macradenia 
98 


virgata 


0.001 substitutions/site 


Figure Phylogene tic hypothesis (one of two maxi- 


mally parsimonious trees) for Holoc B ds macradenia and H. 
virgata based on ani 1 5 of nuclear 185-265 rl external 
and internal transe ribed spacer 10 8 5 nd ITS) sequences 
from representatives of 13 
(Baldwin, unpublished): the clade shown is from a larger tree 


widely-separated populations 


including representatives of H. heermannii and H. obconica 


and E taxa in pu [D. e (D. D. Kec 
3. C : 


maximum-likelihood optimization: \ 
evolution across lineages could not be rejected using 
Felsenstein’s (1988) likelihood-ratio test. 

branches are parsimony bootstrap values (only values > 859 o 


Numbe TS along 


are shown). 


(> 12,300 mm/yr), and extends along an elevational 
gradient from near sea level (75 m) to high alpine 
(3750 m) (Figs. 21-24; Carr, 1985; Carr et al.. 1989; 
Robichaux et al., 1990; Baldwin € Robichaux, 1995: 
1997, 2003a, b, c). 

Clausen, Keck, and Hiesey were aware of Gray's 
(1852) 
Argyroxiphium 
(1936: 8) conducted : 


Baldwin. 


early Hawaiian genera 


Keck 


Argyrox- 


inclusion of the 
and Wilkesia in Madiinae: 


| systematic study of 


iphium with the expectation of treating it in 


Madi(i ie? 


the Hawaiian 


and astutely suggested that members of 


silversword alliance, as currently 


recognized, 


. would appear to constitute an insular 


eroup unto themselves” (Keck, 1936: 10), although his 


inclusion of. Wilkesia in Argyroxiphium has not been 


upheld (Carlquist, 1957: Carr & Kyhos, 1986; 
Baldwin et al.. 1990: Baldwin € Robichaux. 1995: 
Baldwin, 1997). At the same time, Keck rejected 


Gray's (1852) placement of Argyroxiphium and 


Wilkesia in Madiinae and surmised that the silver- 
sword alliance was .. probably without close 
relatives” (Keck, 1936: 10). Keck (1936: 11) 


concluded: “By thus divorcing Argyroxiphium from 


the American genera to which it has been thought 
related, the most persistently proposed connection 
between the ancient element in the Hawaiian flora 
and the New World has been shattered.” 

Molecular phylogenetic data on historical ecology 
and rate of diversification of the 


and on timing 


Hawaiian silversword alliance uphold Carlquist's 
(1965) hypothesis that the Hawaiian taxa represent 
recent adaptive radiation from a tarweed ancestor 
rather than an ancient lineage without close relatives, 
as suggested by Keck (1936). Based on rDNA ITS 
trees, Osborn’s (1902) original, ecological criterion for 
adaptive radiation is met by the silversword alliance; 
major ecological shifts between wet and dry habitats 
were evidently associated with diversification on each 
of the four major island groups occupied by the 
lineage, with the possible exception of the youngest 
island, Hawaii (Fig. 25; Baldwin & Robichaux, 1995 
Robichaux et al., 1990). (1944) 


rapid-diversification criterion. for adaptive radiation 


see also Simpson’s 


is also met: inability to reject rate constancy of 


molecular evolution across clades in the ITS trees and 


time-calibration of a node outside the Hawaiian 


lineage allowed estimation of a maximum age for the 


most recent common ancestor of the silversword 


— 


alliance at 5.2 + 0.8 million years and a minimum 


diversification rate for the group of 0.56 + 
species per million years (Fig. 26: Baldwin & 
Sanderson. 1998). The estimated maximum age of 


the silversword alliance is roughly comparable to the 
Kauai (ca. 


and 


age of the oldest modern high-island. 
5.1 million years; Clague & Dalrymple, 1987) 


much younger than the 10 million-year-old. esti- 


— 


mated ages of some other prominent examples of 


adaptive radiation in the Hawaiian Islands, such as 
Hawaiian drosophilids (Thomas & Hunt, 1991) and 


Hawaiian lobelioids (Givnish et al, 1996). The 
minimum diversification rale estimated for the silver- 
sword alliance falls within the upper range of 


estimated rates for various continental radiations of 
plants and animals and was regarded by Nee (2001: 
661) as 


An expanded perspective on Madia that encom- 


“remarkably high.” 


passes all of the “Madia” lineage (Baldwin, 1996; 
Fig. 18), Le, the clade including all of the taxa in 


Annals of the 
Missouri Botanical Garden 


Bolanderi 


glomera 


! 46 
exigua (6) 
ta 
14 


e hilensig 6 : 


zd DO 
ca pitata ^ 


2 E 
DD 272277777727 7 Le 


sativa 


gracili ZN 
II SUIDA, 


- subspicata’ © 


nÂ | 
eitrioderg es 


nó 


madioides 
n=7 


— 
— 


<= d 
[e] 
3 
3 


\ 11 
7 / 


bu ^ maj. 
— / [11 Nosemitana 
X CMS 1 [fu 

A 


— ey 


elegans 


XWheeleri 


 Citrigracilis 


«eus Species Complex 

Fully tertile hybrid 
Partially sterile hybrid 
—— — Sterile hybri 
---— — Crossing failed 


Figure . Crossing diagram of Madia sensu Clausen (1951). Fully interfertile taxa shown in the diagram are now treated 
as taxonomically indistinct (Madia capitata Nutt. and M. chilensis (Nutt) Reiche are now treated as synonyms of M. sativa 
t 


Molina: Madia wheeleri (A. Gray) D. D. K 


k is now treated ¿ 


isa synonym of M. elegans D. Don ex Lindl.). Some species are 


now treated in other genera, ie., Inisocarpus Nutt. | A. madioides Nutt.]. Harmonia B. G. Baldwin H. hallii (D. D. Keck) B. G. 
Baldwin, Y. nutans (Greene) B. G. Baldwin]. Hemizonella (A. Gray) A. Gray |M minima (A. Gray) A. Gray]. Jensia B. G. 
Baldwin [J. rammii (Greene) B. 6. Baldwin. J. yosemitana (Parry ex A. Gray) B. G. Baldwin]. and Ayhosta B. G. Baldwin K. 
bolanderi (A. Gray) B. G. Baldwin]. All taxa shown are annuals except for Anisocarpus madioides and Kyhosia bolanderi, which 
are perennial herbs. The “Madia” lineage (Baldwin. 1996) also includes Anisocarpus scabridus. Carlquistia muirii. and the 
Hawaiian silversword alliance (30 species in Argyroxiphium. Dubautia, and Wilkesia). Reprinted from Stages in the Evolution 
of Plant Species. by Jens Clausen. Copyright © 1951 by Cornell University. Used by permission of the publisher, Cornell 
University Press. 


Madia sensu Clausen, Keck. and Hiesey (Fig. 17) plus 
Inisocarpus scabridus, Carlquistia muirii; and the 
Hawaiian silversword alliance, reveals strongly con- 
trasting patterns of phenotypic divergence. Life-form 


evolution and habitat shifts among continental 


members of the “Madia” lineage (and among conti- 


nental tarweeds in general) are dwarfed by adaptive 
radiation of the Hawaiiian silversword alliance: the 
continental taxa of the group are annual and perennial 
herbs that occur along a relatively narrow precipila- 
lion gradient compared to the wide range of dry, wet. 


and boggy situations occupied by the semi-woody and 


Baldwin 81 


Volume 93, Number 1 
Clausen, Keck, and Hiesey Revisited 


Arnica mollis — <70% bootstrap 


Hulsea algida — >/0-90% bootstrap 
m >90% bootstrap 


Raillardella 


Adenothamnus validus 


* Hemizonia congesta . : . 
= o "Hemizonia" lineage 
Blepharizonia laxa 
Carlquistia muirii 
Madia elegans 
AA 9 
Jensia rammii 
Jensia yosemitana 


* 


Argyroxiphium sandwicense 
zu -- Wilkesia gymnoxiphium 
E Dubautia laevigata 


Anisocarpus scabridus 


Anisocarpus madioides 
* Harmonia stebbinsii 
— ~ 
Harmonia hallii 
| J Kyhosia bolanderi 
Hemizonella minima 


Holozonia filipes 


oDeeui| ,eIpelN, 


Layia chrysanthemoides 


" * Layia heterotricha 
— Lagophylla minor 
Achyrachaena mollis 
Blepharipappus scaber 
Centromadia perennis 
fe Deinandra greeneana 
Holocarpha virgata "Calycadenia" 
lineage 
Osmadenia tenella 
Calycadenia fremontii 


50 changes 


oDeoul| ,e1Áe7, 


: Madiinae based on simultaneous analysis of 


Figure 18. Phylogenetic hypothesis (the most JR ed tree) f 
chloroplast DNA (rri K intron) and 185—265 nuclear ribosomal DNA external and internal transcribed spacer (ETS and ITS) 
sequences (Baldwin, 2003a, unpublished). The dark box nested within the “Madia” lineage encloses the clade corre sponding 
to the Hawaiian silversword alliance (Argyroxiphium, Dad ts Wilke 8 Asterisks (+) indicate that branch support rises 
to the next level indicated by branch widths with removal of sequences for the three long-branch, Ri genera: 


Achyrachaena Schauer, E Hook., and Holozonia Greene. Raillarde lla (A. Gray) Benth. is represented by a irnk 
intron sequence of R. argentea (A. Gray) A. Gray and ETS and ITS sequences of R. pringlei. Allo 1 of the Hawaiian 
silversword alliance (Barrier et al., 1999) is not re ae d by chloroplast DNA or nuclear rDNA sequences. See Appendix I for 


d of species names not mentioned in text. 


Annals of the 
Missouri Botanical Garden 


QW ON 
M | | 
V á | | 
i| 
IN 


N 


and 20. 


Figures 19 
): lth both diploid species w 
fere genome in the 

. Figure 19 (left). 


closely related to a diff 


ge eee el al., 


Californian montane pere nnials of the 
© previously treated in Raillardiopsis Rydb.. 


Inisocarpus scabridus. — 


— 


E AN 

AOS 

—— 
SO es 


— 


l 


Madia” lineage, not included in Madia sensu Clausen 


the genome of each is evidently more 


(allotetraploid) Hawaiian us sword alliance than S the genome in the other 


. Habit. —B. Leaf. . Head. . Ray floret. 


— isk floret. "o d ie half-enveloping MET ral braci Sen ‘ly adaxial view). Figure 20 17 0 Carlquistia 
mulrii. inii ab re 3. Leaf. cad. —D. Floret. —E. Fruit with pappus. Both species have iae N 
crosses with one a and 1 members of the 11 an silversword alliance (Baldwin et al.. 1991: Carr et al.. 199€ 
Barrier et al.. 1999), 


woody Hawaiian species (see Baldwin, 2003a, b, c). 


Phylogenetic evidence for origin of the silversword 


alliance after considerable diversification of the 
“Madia” lineage (Figs. 18. 26: see also Barrier et 


al.. 1999) the 


Hawaiian lineage represents a major acceleration. of 


indicates that ecological diversity in 


evolutionary change associated with colonization. of 


the Hawaiian Islands. As suggested for diverse 


oceanic-island lineages in general, adaptive radiation 


of the silversword alliance probably has been in 


arge part a consequence of ecological opportunity 


(Simpson, 1953; see Schluter, 2000), with availa- 
bility of a wide array of empty niches in the 
Hawaiian Islands and lack of interference with 


other organisms, possibly aided by intrinsic. factors 


promoting evolutionary change, such as the allo- 
polyploid constitution of the group (Barrier et al., 
1999). 


Extension of Clausen's (1951) crossing results to 


encompass perennial members of the “Madia” lineage 


has indicated. strikingly different fertility patterns 


between the continental and Hawaiian taxa. Crosses 


between the continental perennials and between the 
continental and Hawaiian perennials of the “Madia” 
lineage, not attempted by Clausen, Keck, and Hiesey 


(see Fig. 17). either have failed or yielded vigorous E. 


hybrids of extremely low pollen stainability (an 


estimate of fertility), with most stainable grains being 


abnormally large, tetraporate, and putatively diploid 
1990: 


Barrier et al 


(Kyhos et al., Baldwin et al.. l; Carr et al., 
1990: 1999). Observed. lack of 
interfertility between the continental perennials and 
between the continental and Hawaiian perennials 
Keck. 


the lineage 


conforms to the results obtained by Clausen. 


and Hiesey for the annual members of 
(Clausen, 1951) and to the * 
(1971). In 


members of the Hawaiian silversword alliance have 


Madia pattern” described 


y Grant contrast, crosses. between 


yielded vigorous, partially to fully fertile hybrids in all 


interspecific and intergeneric combinations al- 


tempted, despite major morphological and ecological 
taxa (Fig. 27: Carr & Kyhos, 198 


2003a). Reduced fertility in some 


5 of the 
1986: Tr. 1985, 


o hybrid 


combinations can be attributed 


Volume 93, Number 1 


Baldwin 83 


Clausen, Keck, and Hiesey Revisited 


S 
SN 
y 
E 


SS 
==, F 


2 

(f 

a 
FES 

SS 


in ati) 

4 2 

ib d^. z 
i. . 6 


3 
n 
— 2 AS 
pp ie 


UJ icm 
77 

A 
LZ 


j N 
— a O Y 


E 
SYN NG, 
SY NK IZ. 
= SN NUS ILL AS 
—— Cp 2 > 


Figures 21—2 


Haleakala. East 


Members of the Hawaiian silversword alliance. 
subsp. macrocephalum (A. Gray) Meyrat, a monocarpic, thick-leaved rosette plant of dry, alpine, cinder slopes and flats on 
I 


Figure 21 (upper left). Argyroxiphium sandwicense DC. 


Maui. —A. Habit. —B. Head. Figure 22 (upper right). Dubautia latifolia (A. Gray) D. D. Keck, a liana (woody 


climber) of mesic to wet forests on Kaua'i. Figure 23 (lower left). Dubautia reticulata (Sherff) D. D. Keck, a tree or large shrub 
of wet forests on East Maui. Figure 24 (lower right). Wilkesia gymnoxiphium A. Gray, a monocarpic, fibrous-leaved rosette 
plant of dry slopes and ridges on Kauai. —A. Habit. —B. Segment of capitulescence. —C. Head and distal peduncle. 


mostly or solely to one to three whole-arm reciprocal 
translocations that differentiate genomes of the group 
(Carr & Kyhos. 1986: Carr, 2003b). Even hybrids of 
low fertility, i.e., between Argyroxiphium sandwicense 
DC. subsp. macrocephalum (A. Gray) Meyrat and 


Dubautia menziesii (A. Gray) D. D. Keck, have been 


shown to serve as effective pollen parents in back- 
crosses, with recovery of full fertility in a subset of By 
rogeny (Carr, 1995, 2003a). Ecological (post-dis- 
progen) | 


e 


Y 


persal) selection against hybrids appears to | 
important in limiting gene flow in the Hawaiian group 


(Carr, 1995), although such selection has not pre- 


84 Annals of the 
Missouri Botanical Garden 


c ¥ 
N 
} i 
\ q 
1” EN 
— cid XY > 
SR 
S NN 
E P» 7, { x 
gy ON " 
7 y t 
7 1 
* N = 
j : a 
Y) y 
y NEL 
YY ` — si 


400 km 


| Maui WY- California 
/ Hawai v^ N 


o 
0) 
c 
o 
3 
(e 
x 
X 


(Dubautia/Wilkesia) Ñ 
> | Pa 

A. sandwicense m. 

A. grayanum WM 

A. caliginis 

A. scabridus 

K. bolanderi 


C. muirii 


D wiB | wie | B| W/B 
D or W|RPorS D, W, or B | RP Maui Nui 


i » D W 
Kaua'i or O or W H. BP or S e CAFP 
Maui Nui 


; TUNE D or W | H 
CA FP, Kaua'i, or Maui Nui CAFP 


11287). Ecological shifts and major dispersal events during diversification of each of four principal 
1 


Figure 25 (pp. 84- 


ae s of the Hawaiian silversword alliance. based on the semi-striet consensus of eight maximally parsimonious nuclear 
ITS trees (reprinted or adapted an Bal lwin, 1997: data from Baldwin & Robichaux. 1995). —A. M Sra in 
a voung-island group (Maui and Hawaii), shown in context of "Dubautia/M 1 lineage (detailed in Figs. 25B. C. and D) 
and western North American outgroup laxa ntane, perennial tarweeds of the “Madia” lineage (in Anisocarpus, d S 
and Ayhosia). —B. Wilkesia sister group ipn of d restricted to e i Habitat abbreviations and symbols: B = 
bog: D or sun symbol = dry wel: cloud symbol = wet or p. iin symbols: H = herb: L = liana o climber): M = 
mi nu LUE RP = rosette 1 75 (monocarpic or polye arpic); S wür = tree, Geographic ps CA FP = California Floristic 
ince: Maui Nui = prehistoric island uniting Lanai. 111 i. and e (i i Kaho'olawe, where no me we rs of Madiinae 
are 1 to occur). Placement of geographic area, habitat, and habit along branches is where ancestral state is une quivocal 
based on parsimony criterion (Baldwin & Robic lo aux, 1995). Island occurrences and life form information for the silve rsword 
alliance conform with Carr (1985, 1999), Taxonomic abbreviations: Los or a mo = subsp. macrocephalun m, 5. = 
1 eigene; D. ciliolata (DC) D. p. Keck— “e.” = subsp. ciliolata, * © = subsp. glutinosa G. D. Carr; D. knudsenii 
Hille = sul 8b. filiformis G. D. 6 eT = sip ne “n” = Ss sp. nagatae (H. St. John) G. D. Carr: D. laxa 
Hook. 4 Aat; — “h.” = subsp. hirsuta (Hill C. D. ( uL uds laxa: D. linearis 1 : D. D. K os 
subs wo p Mann) C. D. Carr, /.“ = subs Lo p Hagin Gaudich.— “BH” (Blue iv = e 
magnifolia (Sherff.) G. D. Carr, AD — subsp. humilis 6. . Carr, /.“ = subsp. Pom D scabra (DC. i Keck— 
ULU = subsp. ph (A. Gray) G. D. Carr, “s.” = 1 5 on See Appendix 1 for authorship of species names not 
mentioned in tex 
cluded establishment of hybrids: 38 naturally occur- e.g. for D. scabra (DC) D. D. Keck (Baldwin et al.. 
ring hybrid combinations between different pairs of 1990: Baldwin. 1997) 
species (including intergeneric hybrids) have been The stark contrast between the Hawaiian pattern of 
documented by Carr (20034). and evidence for both interfertility across phenotypically disparate. lineages 
recent and ancient introgression has been obtained and the pattern of intersterilily seen across the less 
from studies of hybrid zones (Crins et al, 1988: distinctive continental species of the “Madia” lineage 
Caraway et al., 2001) and comparison of phylogenetic has evolutionary implications. Lack of strong. intrinsic 


data from cytogenetic, nuclear rDNA, and chloroplast. reproductive barriers between species and genera of 
DNA studies (Baldwin et al.. 1990; Baldwin, 1997, the Hawaiian silversword alliance leaves open the 


2003c). Hybrid speciation, as well, may explain some potential for homoploid hybrid speciation or intro- 


conflicts between different lines of phylogenetic data, — gression involving plants. that differ greatly in 


Volume 93, Number 1 
2006 


W. gymnoxiphium 


UJ 


U 


RP 


aldwin 
Clausen, Keck, and Hiesey Revisited 


D. paleata 


D. raillardioides 


S 
2 
5 
C 


WIiLorS 


Kaua'i D or W | HP or S 


Figure 25. Continued. 


life-form or ecological setting, particularly under 
environmental conditions favorable to recombinant 
Carr, 1995, 2003a). Intersterility barriers 


among continental species of the “Madia” lineage may 


phenotypes 


limit the evolutionary potential of hybridization to 


a 


formation of allopolyploids, as demonstratec experi- 


mentally by Clausen et al. (1945) for some annual 
species and as meiotic behavior and diploid pollen 
formation indicate for hybrids between the continental 
1996; 1999). 


Such allopolyploidy has contributed only modestly to 


perennials (Carr et al., Barrier et al., 


diversity in the continental lineages (e.g. Madia 
citrigracilis), but evidently gave rise to the lineage 
Islands 
and became the silversword alliance (Barrier et al., 


1999). 


Phylogenetic 


that successfully colonized the Hawaiian 


data from ITS sequences on time 


since divergence of perennials in the “Madia” lineage 


1998) 


to be placed in an 


(Baldwin & Sanderson, allows patterns of 


interfertility understandable 
perspective (Fig. 26). Lack of interfertility between 
continental perennials and between continental and 
Hawaiian perennials may be explained simply by the 
longer timeframe for evolutionary divergence between 
those lineages compared to the time since divergence 
of the 


with gradual breakdown in interfertility through time 


Hawaiian lineages from a common ancestor, 


as a by-product of divergence. Interfertility between 


members of the silversword alliance is consistent with 


fertility patterns seen in other (mostly perennial 
lineages of Hawaiian angiosperms (see Carr, 1998 
Baldwin, 


common fertility pattern seen in young woody 


o 
— 
— 


1998) and conforms more closely t 
groups in general, ie. the “Ceanothus 
(Grant, 1971), 


widely in annuals (Grant, 1971). Although the relative 


paus rn” 


than to the “Madia pattern,” seen 


Annals of the 
Missouri Botanical Garden 


D. knudsenii xk ——s»- 


D. pauciflorula 
D. knudsenii f. 
D. knudsenii n. 


D. laxal 
D. laxa h. 


- 
if 


i 83 — 
o d c 8 E E 
a hy QU Qj o > 
E D D ® 
Q = £ 8 & y 2 

D D D s 

S 8S8 Y S S S$ P 
5 c c € S > > 
= mS S g E 8 o 
E a a a = y S 
Q Q Q Q Q Q Q 


Figure 25. Continued. — 
pers. comm. . native (but not all ds mic) Lo 1 b 
(not on. Kauai). For additional details, see p. 84. 


timing of divergence of annuals and perennials in the 


“Madia lineage" has not vet been resolved, the 
generally higher rates of rDNA evolution in the 


Baldwin & 
1998) accords with findings. for 
Sidalcea 
& Baldwin, 2001) and may reflect more general rates 


annuals compared to the perennials 


Sanderson, some 


other angiosperms (e.g.. V. Gray; Andreasen 


of molecular evolution, which, in turn. could bear on 
a putatively higher rate of acquisition of post-zygotic 
reproductive barriers between annuals than. between 


perennials in general. Similar results have been 


obtained in comparisons between phylogenetic and 
crossing data for continental annual and insular 
perennial lineages of the tarweed genus Deinandra 
Greene (Baldwin, unpublished). These findings re- 
inforce Grants. (1971) observations on associations 
among life history, patterns of interfertility. and the 
evolutionary potential of hybridization in plants. and 
illustrate further contrasts in patterns and processes of 
evolutionary change among closely related lineages of 


Madiinae. 


Members of Dubautia with genomic arrangements 


Kaua'i W | S 


DGI and DG3 (Carr & Kyhos, 1986. 


—D. Dubautia sect. Railliardia, restricted to one or more young island(s) 


CONCLUSIONS 


Evolutionary investigations of Madiinae in the wake 

E Pa] 
of Clausen, Keck. and Hiesey's pioneering research 
have led to a greatly expanded cireumseription of the 
subtribe and a more detailed understanding of 
processes and patterns of diversification in the group. 
Progress in resolving evolutionary questions in 


Madiinae lines of 
Keck, 


and Hiesey and reaffirms the value of systematic data 


through integration of diverse 


evidence follows in the tradition of Clausen. 


for tackling process-oriented evolutionary questions. 
Keck, 


evolutionary 


as appreciated by the Carnegie team. Clausen, 
and Tliesey’s evidence for a 
Madiinae, 


morphological. 


dynamic 
history of both in relative rates of 


ecological, and genetic change and 


in modes of diversification, has been upheld and 
extended to finer-scale levels than in the comparisons 


(1951). 


processes of 


among genera described by Clausen Even 


within individual genera of. Madiinae. 


evolutionary change and the biological properties of 


Volume 93, Number 1 
2006 


O'ahu 


1 45 é 
77771 0 ^ 


(ONE c 
2n 


D. herbstobatae 
D. sherffiana 
D. menziesii 


O 
= 
O 
O 


O'ahu 


aldwin 87 
Clausen, Keck, and Hiesey Revisited 


I a 
6755 
DAP 


Kaua'i, O'ahu 


: MF 2: | 
Be E =» . « © = AR. XE 
a 3 o T T S © — — 
> 3 [e] 3 E Q Ks] D G 
[s x Q 2 2 00 © G 
a o G 3 S 8 8 ES = 
G a a G Ci a C 2 Ci 
WIT DIT WIM W D D 
\ Dh ~ - 
Maui Nui | Hawai'i 
Hawaii Maui Nui or Hawai'i 


„ DorW|S 
or Maui Nui. p 


Figure 25. Continued. 


ma] 


lineages may differ greatly. These results underscore 
the value of phylogenetic data in process-oriented 
evolutionary studies and the importance of phyloge- 
netic considerations in classification at all levels of 


evolutionary divergence. 


= 


A recurring finding in the re-examination of 


tarweed evolution using modern systematic ap- 
o J 
proaches has been the discovery of eryptic or semi- 


eryplic diversity, even in groups that were studied 
Keck, 


experimental biosystematic perspective (e.g.. Laya). 


intensively by Clausen, and Hiesey from an 
In part, such diversity appears to reflect ecological ls 
differentiated and evolutionarily divergent—though 
often interfertile—lineages, as in L. gaillardioides. 
Similar examples have emerged recently from molec- 
ular phylogenetic and ecological studies of other plant 
taxa that were a focus of previous experimental work, 
such as Lasthenia Cass. (Ornduff, 1966; Chan et al. 
2002; 2003). 


importance of ecologically distinct, cryptic groups 
o J 4 [e] 


Rajakaruna et al., The potential 


was well appreciated by Clausen (1951: 29-30): "[t is 
now established as a general biological law that 
species that occupy many kinds of environment are 
able do so because they have evolved series of 
physiologically distinct races, each of which survives 
within its native zone but is less able to compete in 
neighboring zones and usually is unable to survive in 
the extreme ranges of the species. This fitness is 
primarily physiological: it is determined by genes, and 
it may or may not be expressed in the external 
appearance of the plant. Ecological races or ecotypes 
have therefore generally been overlooked by taxono- 
mists, geneticists, and ecologists, all of whom looked 
that 


however, the ecological race is a far more 


could classify. 


or visible characters they 
Actually. 
important biological entity than the morphological 


Although 


‘ecological races’ 


subspecies.” infeasible during Clausen’s 


life, resolving * ' that correspond to 
natural groups worthy of taxonomic recognition is now 


possible and desirable, especially in light of the 


88 Annals of the 
Missouri Botanical Garden 


K. bolanderi 509 

C. muirii 618 

A. scabridus 676 
A. madioides 488 


5] 


spaamie | 


W. gymnoxiphium 76.022 


D. latifolia B675 


. paleata 1375 
. raillardioides 670 
. menziesii 522 


. reticulata 664 


. arborea 527 


. Scabra l. 778 


. ciliolata c. 529 


. herbstobatae 1244 


. Sherffiana 515 
. laevigata 671 


. imbricata 667 


. plantaginea p. 1180 


. plantaginea h. 1183 


douelſſe PIOMSI3AJIS UHREN 


D 
D 
D 
D 
D 
D 
D 
64 L D. ciliolata g. 659 
D 
D 
D 
D 
D 
D 
D 


. plantaginea" BH" 776 


e 
O 
O 
O 


. microcephala 1044 
. laxa |. 662 
. laxa h. 833 


. knudsenii k. 1047 
. knudsenii n. 1322 
. caliginis 680 


D 
D 
D 
D. pauciflorula 668 
D 
D 
A 
A 


. sandwicense s. 657 


72 
r 
LI oT PF LE TT TTT TT T] 
15 0 
Ma Ma 


Figure 26. A phylogenetic hypothesis (one of four maxim: ally parsimonious trees) for perennial members of the “Madii x 
lineage, with time-calibrated branches. based on nuclear e N. rs sequences (adapted from D; dodi & Sanderson, 1998 
Continental tarweeds (Anisocarpus, Cader uistia B. G. Baldwin. and Kyhosia) and the Hay ir silversword alliar 
(Argyroxiphium, Dubautia. and Wilkesia) were included in the an: ilysis. The outgroup (Ade S 1 and Ratllarde Ila 


A een is not shown. jud ‘constancy of mole c ul; ar evolution across line ages could nol be rejected using Fe ‘else lein’s > (1 M 


uM on the assumption pu "iversifio alion above ne "es would. nol M Hc e Wer due AM sere 
(15 million years ago). when summer precipitation in western North America began to (nearly all continental 
tarweeds are restricted to areas of summer- dry climate: see Baldwin & Sanderson. 1998), S © Figure 25 caption for 
ibus wiations of f subspe cies in Dubautia. 


Volume 93, Number 1 
2006 


14 


DUBAUTIA N 


$ 
22 
EN 


MAT-FORMING 
SUBSHRUB 


Figure 27. 


silversword alliance based on data of Carr (1985) and Carr and Kyhos (1981. 1986). Crosses 


b 

SAN 
L 

cm r 

* ^ 


Crossing diagram showing synthetic and natural hybrids between species of different life-form in the Hawaiian 
) 


Baldwin 
Clausen, Keck, and Hiesey Revisited 


MONOCARPIC 
ROSETTE 


= 


Lysa NN 


PA 
Y > 


in all attempted combinations 


vere successful and yielded hybrids of at least partial fertility. Reprinted from Carr et al. (1989) with permission of the 


y 
publisher. Oxford University Press. 


potential evolutionary importance of such lineages in 
; An 


challenge in the coming years for systematists working 


the face of rapid global change. important 


in well-studied, threatened floras, such as the 
California flora. will be to resolve ecologically 


significant, overlooked diversity through application 
of molecular phylogenetic and experimental ap- 
proaches, much in the spirit of Clausen. Keck, and 
Hiesey's classic investigations (see Baldwin, 2000). 
The potential of modern plant systematics to 
contribute to understanding of evolutionary processes 
and fine-scale diversity is just beginning to be 
realized. with exciting results (e.g.. Rieseberg, 2006 
Sakai 2006 


Spectacular progress by systematists in resolving 


this volume: el al. this volume). 


higher-level plant relationships through application 


of 


an invaluable framework for integrative systematic 


molecular phylogenetic methods has created 
studies of young lineages. Recent methodological 
advances in systematics that allow rigorous esli- 
mation of divergence times and diversification 
rates and the explosive growth in comparative 
methods have much potential to aid systematists in 
taking phylogenetic hypotheses beyond the limitations 
of strictly pattern-based considerations. and toward 
better informed conclusions about how and why 
evolutionary changes have occurred. The promise of 
collaborative studies of plant diversification that 
bring together the combined strengths of system- 
atists, ecologists, and geneticists, so well demonstrat- 
ed by Clausen, Keck, and Hiesey, has never been 


greater. 


Literature. Cited 


Andreasen, K. & B. C. Baldwin. 2001. Unequal evolutionary 
rales between 10 0 and perennial lineages of checke 


mallows (Sidalcea, Malvaceae): Evider 


= 


Biol. ae Rp 2 I4. 
Babcock, E. B. & . Hall. 

genetic, 1 taxonomic study of. the 

larweeds. n Calif. Publ. Bot. 13: 15-100. 
Baldwin, B. G. 1992, 


transeribe 1 


1924. Hemizonia congesta: 
c hay-field 
utility of the internal 
acers of nuclear ribosomal DNA in plants: 
e sample from the Compositae. Molec. 


Phylogenetic 
Phylogen. Evol. 


6. 
. 1996. Phylogeneties of the California tarweeds and 
the 0 silversword allianc m Madiinae; Heliantheae 
37 1-30 li 100 " J. N. J Hind & K H. 


rosita Sys stematics 


sensu lato). P} Beentje 
(editors), Proceedings of the 
Kew. 1994, Vol. I. 


International Compositae Conference, 


Gardens, Ke 


A 


Royal Botanic ( 
— 1007. Adaptive 1 alion of the Hawaiian silver- 


sword alliance: Congruence and conflict of phylogenetic 
evidence from molec ular and non-molecular investiga- 
10 


tions. Pp. 103-128 in 


Givnish & K. J. Sytsma 
(editors), Molecular Evolution and Adaptive Radiation. 


Cambridge Univ. Press, Cambridge. 


——. 1998. Evolution in the endemic Hawaiian C ;ompos- 
itae. Pp. 49-73 in T. F. Stuessy & M. Ono (editors). 
Evolution and Speciation of I Island Plants. Cambridge 
. Press, Cambridge 
modern plant systematics in 
\ conservation of fine-scale biodiversity. 
Madroño 47: 219-229, 

20 \ phylogenetic perspective on the origin and 
Madina Pp. 193- a in S. Carle 


). Carr (editors). ' 


evolution i ol 


s ( wes 'Taceae). Missouri Botan- 

en Press, St. Lor 

— — — 2003b. il l Tene of the continental tarweeds 

Hawaiian (Asteraceac— 

. Baldwin & G. 
> Evolution of 

Botanical 


1 


eraceae). Missouri Garden 

Characteristics and dive rsity of Madiinae. 

in S. Carlquist. B. C. Baldwin & G. D. Carr 

(editors), Tarweeds & Silverswords: Evolution of the 

e (Asteraceae). Missouri Botanical Garden Press. 
. Louis. 


x 17—52 


— — 2005. Origin of the se rpentine-endemic herb Lavia 
discoidea from the widespread L. glandulosa (C. -ompositae). 
Evolution 59: 2473-2470 

N 


& K. II. Si ane. 1995. Historical O 


and ecology of ihe Hawaiian  silversword allia 

(Asteraceae): New molecular ogee perspectives. 
Pp. 259-287 in W. vedior « Funk (editors). 
Hawaiian Biogeograp le Evolution on a Hot Spot 


Archipelago. Siillisonion Institution Press. Washington, 
Di. 

& M. Sanderson. 1998. Age and rate of 
diversification of the 9 
0402— 


dna e 
(Compositae). Proc. Nath Acad. Sei; U.S.A. 
9406 
W. Kvhos & J. Dvorak. 1990. Chloroplast DNA 
elu dox a p" iplive radiation in the Hawaiian silve 
word alliance \steraceae-Madiinae). Ann. Missouri Bot. 
(and. 96-109. 


fan 185-2065 


A internal and external sella spacers. Molec. 


Annals of the 
Missouri Botanical Garden 


D. Carr. 1991. 0 


DNA evidence for a North American origin. of th 
Hawaiian silversword alliance (Asteraceae: 
Proc. Natl. ad. Sci. U.S.A. 88: 1840-1843. 
oo Baldwi R. H: 


oi, 1999. Interspecifi 


Madiinae). 


> 


a 
= 


Robichaux & 

e hybrid ancestry of a plan 
adaptive radiation: bid. of the Hawaiian silve 
sword alliance (Asteraceae) inferred from floral homeotic 


Evol. 16: 1105-1113. 


1984. Genetic revolutions. 


gene duplications. Molec. Biol. 
Barton, N. H. & 


founder effects, and speciation. 
33-104 


B. Charlesworth. 
Annual Rev. Ecol. Syst. 


15: 133-1 

Caraway. V. G. D. Carr & E W. Morden, 2001. Assessment 
of hybridization and introgression in lava-colonizing 
Hawaiian. Dubautia 1 Madiinae) using RAPD 
care Amer. J. Bot. 88: 1688-1694 

Carlquist, S. 1957. Leaf anatomy and ontogeny in Argyro- 


iium a = Wilkesia. Amer. J. Bot. 44: 696—705 
9a. Glandular structures of Holocarpha " their 
m re J. Bot. 46: 300-308. 


uc ny. 


959b. Studies on Madinae: Anatomy, cytology. and 
On relationships. Aliso 4: 171-2306. 
—.. |965. Island Life. Natural History Press; New York. 


— ——— 1974. Island Biology. Columbia Univ. Press, New 
Vorl 

— ——, 1980. Hawaii, a Natural History: Geology, Climate, 
Native Flora and Fauna above a Shore line. 2nd ed. 


Pacific s ‘al Botanical Garden. Law: 
. Baldwin & G. D. Carr ( 
^ ral Evolution - of the 


2003. Tar- 
1 Madiinae 
the Hawaiian Madiinae 
(Asteraceae): Dubautia. and Wilkesia. 


VE. yroxiphium, 
Allertonia 4: 29 


. 1995 A fully fertile intergenerie hybrid derivative 
from Ten iphium sandwicense ssp. se cle din 


steraceae) and its relevance to 


Amer. J. Bot. 


Dubaulia menziesii 
plant evolution in the Hawaiian Islands. 
82: 1574-1581. 
1998. Chromosome evolution and e in 
Hawaiian flowering plants. Pp. Stuessy & M. 
Ono (editors). E MEN and T ciation of Island Plants. 
Cambridge Univ. 


~ 


5—47 in T. F. 


Press, Cam 

22 igs O in D iin Pp. 79-1 
Ros ae . Baldwin & G. D. Carr (editors). 
Evolution of n M: idiinae 
Mis Botanical Garden Press. St. Loui 


M in S. 
Tarweeds 
Silve 9 a (Asteraceae). 


030. 5 e sation in tiene Pp. 53- 


78 in a Carlquist. B. C. Baldwin & C. D. Carr (editors). 
Tarweeds ; Silverswords: Do of the Madiinae 


(Asteraceae). Missouri Botanical Garden Press. St. Louis. 
D» . Kyhos. 1981. Adaptive 


NA e allian 


radiation in the 


(C ompositae -Madiinae). 
of nde hybrids. : 


Cy loger IK 


543- 


Evolution 35: 


— —. 1986. Adaptive radiation in the Hawaiian 
silve i alliance (C ompositae-Madiinae). H. 


ruificial and 


loge E 


elics of natural Evolution 40: 
a 

. B. C. Baldwin & D. W. Kyhos. 1996. € vlogenetic 
mplications of artificial bu. between the 


silversword 


hybrids. 


lawaiian 


alliance and North American been" 


(Asteraceae: Heliantheae—Madiinae). Amer. J. Bot. 83: 
053-000 

. R. H. Robichaux. M. ter & D. W. byhos. 
1989. Adaptive radiation of Mahon silversword 


alliance (Compositae-Madiinae): A comparison with 


Volume 93, Number 1 
2006 


79-97 in L. V. 


Hawaiian pic ture-winged Drosophila. Pp: 
Y & W. Anderson (editors) 


Giddings, K. Y. Kaneshiro , 

Ger netics, Speciation and the F bate r Principle. Oxford 
55. y New OK. 

. B. C. Baldwin & J. L. Strother. 2003, Accepted 

names A synonyms for spec 15 and m cific taxa in 

229-243 in S. Carlquist, B. G. Baldwin & 

& 1 Evolution 


Madiinae. 
C Carr (editors), 
of the Madiinae (Asteraceae). Missouri Botanical Garden 
Press, St. Louis. 

Chan, R.. B. G. Baldwin & R. 


molecular phylogenetic 


Pp. 2 
Tarweeds 


Ornduff. 2002. Cryptic 
re-investigation of 
and close relatives 


Amer. J. Bot. 89: 


californica sensu lato 


(Compositae: Heliantheae sensu lato). 
1103-1112. 

Clague, D. A. & G. B. Dalrymple. e = Hawaiian- 
Emperor volcanic chain. Pp. 1254 i . Decker, T. 
W right & P. H. Stauffer (editors), V "e 'anism in Hawaii. U. 

Geological Su 350. B 
Gove us Printing Office. Washingk 
951. Stages in the ia " plant s 


rvey Professional Pap: rd 


Clausen, )ecies 
Cornell 11115 Press, Ithaca (reprinted by Hafner Publish- 
ing C rx New var in 1962). 

Ke ec ko & 


species ae on experime i e J. Bot. 26: 


. Hiesey 


— 


. 1939, The concept of 
103—100. 
). Experimental Studies on 


i ct s Varied E 


the | Nature of e l. invironments on 


Western North American Plants. Carnegie Institution of 
W 1 Publication 520. 
& . 1941. Experime ntal taxonomy. 


Book no. 40: 


Camegie Institution o E ashington Year 
160—17 
; & ———. 1945. Experimental Studies on 
Nature of II. Plant throug 
Amphiploidy and Autoploidy with Exa the 
Madiinae. Carnegie Institution of W he Publication 
504 


Species Evolution 


mples from 


— . ——— & ———— 4. Heredity of geographically 
and ec ~ ally isolate races. American Naturalist 81: 
|14- 


> 


c 


; n 8. Experimental Studies on 
the Nature of Species. | T nvironmental Responses of 
Climatic 
Washington a, 58 

A. 199 EN Mayr and the origin of species. 


Races of s villea. Carnegie Institution of 
Coyne, J. 
Evolution 48: D 
Crins, W. J.. D. pod & G. D. Carr. 
indicators odi 5 in a mixed population of lava- 
Heliantheae- 


1988. Flavonoids as 


colonizing Hawaiian tarv veeds (Asteraceae: 
Madiinae). zu 2: 19:3 


Donoghue, M. J A critique " the biological species 


567-57 


concept and recomme encase for a phylogenetic alterna- 
live. The Bryologist 88: 172-181. 

Ehrlich. P. R P. H. Raven. uu Differentiation. of 
populations: Gene flow seems to be less 


Er 


important in 

speciation = the neo- 5 thought. Science 165 
1228-123 

Felsenstein, a 1988. Phylogenies from molecular sequences: 


nference and reliability. Annual Rev. Genel. 22: 

521-5 
2004. 

se ai 


Ford. & I. D. 


in b. (€ e 


Inferring Phylogenies. Sinauer Associates, 


9 05 b. 1989. Morphological evolution 
Character recombination in hybrids 
glandulosa. Syst. ot. 14 


between L. iur and L. 


284-206. 


Baldwin 91 
Clausen, Keck, and Hiesey Revisited 


. 1990. Genetic studies of floral evolution 
11 64: 29-44. 
S 9. Jens Christian Clausen: March 11, 1891- 
rad. Sci. U.S.A. 


in Layla. 
French, C. 5 
Nove ubet? 22, eed Biogr. Mem. Natl. Ac 
58: 75-95. 
Givnish, T. J., E. Knox. T. B. D. Palmer & K.. 


Sytsma. 1996. The Hawaiian lobelioids are monophyletic 


Patterson, J. 


and underwent a rapid initial radiation roughly 15 
million years ago. Amer. J. Bot. (supplement). 83: 159 


Gottlieb. L. D. 1973. Genetic differentiation. sympalric 
speciation and the origin of a iploid species of 
Stephanomeria. Amer. J. Bot. 60: 545 553 

2004. Rethinking classic examples of recent 


alias in plants. New D 161: 71-82. 
S. I. W 1985. Morphological 


and elec 0 tic divergence. between d discoidea 


arwick € V. S. Ford. 


and L. o Syst. Bot. 10: 48441 
Grant, V. 1963. The Origin of o 0 olumbia Univ. 
Press, New pira 
1971. Plant Speciation. Columbia Univ. Press. New 
York 


. 1981. Plant Speciation, 2nd ed. Columbia Univ. 
Press, New York. 

Gray, A. 1852. Quarterly meeting, August 8. 1949. 
Amer. Acad. Arts 2: 159-160. 

Hall. H. M. 1932. Heredity and environment—as illustrated 
by transplant studies. Sei. Monthly 35(October): 289-302. 


Proc. 


ck, D. D. 1936. The Hawaiian silverswords: Systematics, 
affinities, "ed phytogeographic problems of the genus 
Argyroxiphium. Occas "ap. Bernice Pauahi Bishop Mus. 
11: 1-38. 


Kruckeberg, A. R. 1992. Plant life of western North 
American ultramafics. Pp. 31-73 in B. A. Roberts & J. 
Proctor (editors), The Ecology of Areas with Serpentinized 
Rocks: A World Kluwer i 
Dordrec ne 

2002 

eee and Rock 

ass, Seattle and London. 

Kyhos, D. W. G. D. Carr & B. G. 1990. 
Biodiversity and cytogenetics of the tarweeds (Asteraceae: 
He cl sae-Madiinae). Ann Bot. 
84-95 

Levin, p. A. 1993. 105 al speciation in plants: The rule not 

. Bot. 18: 197-2 


66. "ue iation in flowering lasts Science 152: 


View. Academic Publishers. 


Life: 


Univ. 


The Effects 
of Washing- 


Plant 
on Plants. 


and 


Geology 
Types 
ton Pre 
Baldwin. 


Missouri Gard. 77 


= 


xce plion. S 
ein H. 19 


167-172 


& M. R. Roberts. 19: 7 5 The origin of Clarkia 
lingulata. dd 10: 126-1: 

Losos, J & R. E. Glor. 2003. A comparative 
me A pe Um MET of speciation, Trends Ecol. 
Evol. 18: 220-227. 

Mayr. ^ pes . Processes of speciation in animals. Pp. 1-19 

. Barigozzi (editor). Ms of Speciation. Alan 
" n New York 

Nee, S. 2001. Inferring us iation rates from phylogenies 
Evolution 55: 66 1-60 

Neigel. J. E. & J. C. iue 1986. Phylogenetic relationships 


of mitochondrial DNA under various is demographic models 
Karlin € E. Nevo 
Academic 


of ape ciation. Pp. 5155: 


„ E 1 Processes and Theory. 


survey of the eoldfield 
Helenieae). Univ. Calif. 


Ornduff, R. 1966. A biosystematic 
genus Lasthenia (Compositae: 
92. 


Publ. Bot. 40: 1— 


Osborn, 902 The 
TN P vs 
R. E. 198 


Holoc arpha. “e D. dissertation, University of California. 


law of adaptive radiation. Amer 


Palmer, E ological and Evolutionary Patterns in 


avis 

— 1987. Evolutionary pad of e 
macradenia. PP. 295-304 in T. Elias (editor 
Conservation and Management "3 jm an 2 
lodangered Plants. California Native Plant Society. 
Sacramento. 

Rajakaruna, N., B. G. Baldwin. R. Chan, A. M. Desrochers. 
D. . Bohm & J. Whitton. 2003. Edaphie races and 


phylogenetic taxa in the Lasthenia californica (Asteraceae) 
complex: An hypothesis of parallel evolution. oe Ecol. 
12: 1675-1679 


haven, P. H. & D \xelrod. 1978, Origin and relationships 


of the California fora. Univ. Calif. Publ. Bot. 72: 1- 
134. 
Rieseberg, L. H. 2000. du uo cialion in wild sunflowers. 
(e Bot. Gard. 93: 34-48. 


Brouillet. 
paraphyletic? Taxon 43: 21-32. 
& J. M. Burke. 2001. The 


Gen ne flow, selection, 


— & 


Are many plant species 
) 


biological reality o 
and. collective evolution. 
. Pearcy. 


Robichaux, R. H.. G. D. Carr. M. Liebman & R. 
( 


Adaptive radiation of the Hawaiian vd 


alliance (Compositae: Madiinae): Ecological. morphologi- 


Annals of the 
Missouri Botanical Garden 


cal. and OE 'al diversity. Ann. Missouri Bot. Gard. 


TT: 04-72 
Sakai, A. K., S. C. Weller, W. L. Wagner, M. Nepokroeff & T. 
M. D 2 Adaptive radiation and evolution. o 


breeding Systems in Sehiedea (C EP MS an 


endemic Hawaiian genus. Ann. Missouri Bot. Gard. 
19-63. 
Schluter, D. 2000. The Ecology of 
Oxford Univ. Press. Oxford. 
y 


Simpson. G. G. 1944. 


Adaptive Radiation. 


Tempo and Mode in Evolution. 
Columbia Univ. Press. New Yor 
1953. The 17 05 Features of. Evolution. Columbia 


Univ. Press; New York. 
Stebbins, C. L. 1949. Rates of evolution i in plants. Pp. 229— 
242 in G. L. Jepsen, E. Mayr & G. G. Simpson (editors). 


Genetics, Paleontology, and chia Princeton Univ. 
Press, de inceton, 
». Processes of organic evolution. Prentice-Hall, 
Engle en Cliffs. 
-. 1982. Plant speciation, Pp. 21-39 in C. Barigozzi 


\lan R. Liss, New 


(editor). Mechanisms o Spec lation, 


. Hunt. 199]. 
of the 1 de 5 nase locus and E 11 15 ny of 


or 
"damus II. The molecular evolution 
Hawaiian Drosophila. Molec. Biol. Evol. 8: 087—702 
Warwick, S. I. & L. D. Gottlieb. 1085. Ge 
and geographic speciation in Layía (C pipe ). Evolu- 
tion, 39: 1236-1241 


ietie divergence 


Volume 93, Number 1 
2006 


aldwin 
Clausen, Keck, and Hiesey Revisited 


APPENDIX. 1l. 
when otherwise nol provided in the t 


listing of accepted names and synonyms for 


infraspecific taxa in Madiniiae, see Carr et al., 2 


Achyrachaena mollis Se auer 


= 
x 
~ 
= 


Argyroxiphium grayanum (Hillebr.) 0. De: 5 
Argyroxiphium kauense (Rock & M. Neal) € 
Deg. 

Blepharipappus scaber Hook. 
Mepharizonia lax 
Cah cadenia a A. Gray 


Greene 


Ceniromadia perennis Greene 
Deinandra greeneana ie B. 6. o 
eck 


Gray 3 D. 


Dubautia herbstobatae 8 


Dubautia arborea (A. 


Dubautia imbricata H. S 
Dubautia laevigata A. 5 ray 
Dubautia microcephala Skottsb. 


1 5 8 G. D. Carr 


Authorities for spec ies pres sented in the figures 
For a complete 


specific and 


). Deg. 


2003. 


& I. 


Dubautia paleata A. Gray 


Dubautia nd ur H. Sı. dol 5 G. D. Carr 


Dubautia sherffiana Fosbet 
Harmonia ede (T. W. Nelson & J. 
Baldw 


Hemizonia congesta D( 

ilipes (Hook. 8 Arn.) Greene 
Lagophylla minor (D. D. Keck) D. D. Keck 
Layia heterotricha (DC. ) Hook. & A 


yok. 
Madia gracilis (Sm.) D. D. Keck & J. C. 
App i egal 
Madia a 1 Kellogg 


Madia ica p. D. Keck 


Osmadenia tenella N 


utt. 
Wilkesia hobdyi H. St. John 


“lause n ex 


P. Nelson) B. G. 


SPECIES BEFORE SPECIATION Peter R. Grant? and B. Rosemary Grant? 
IS COMPLETE! 


ABSTRACT 


The adaptive radiation of Darwin’s finches in the Galapagos archipelago stands as a model of species multiplication. The 
radiation began two to three million years ago, and resulted in 14 species being derived from the original colonizing species. 
This system is highly suitable for investigating the causes of speciation because closely related species occur sympatrica 


= 


several combinations and in environments with relatively little anthropogenic disturbance. The role of natural selection and 


adaptation to feeding nic hesi in the allopatrie phase of spe ciation has be en demonstrated repeatedly. In the sympatric phase of 


speciation, differences in song and Ue dd actas a premating barrier to gene exc e This form of reproductive isolation 


evolves al least partiy as a passive consequence or byproduct of adaptive divergence in beak morphology. Song characteristics 


diverge in allopatry, largely independent of be ak omg and for a variety of reasons, not all of which are well understood. 
The barrier to gene exchange in sympatry is nol completely effective, however: species hybridize rarely, and under some 


circumstances the hybrids are surprisingly fit. These re sults ch allenge some current notions of species. For example, the 
ground finch species Geospiza scandens Gould and G. fortis Gould on the island of 1 1 5 Major have lost morphological 
diagnosability, as a result of introgressive hybridization, while retaining vocal diagnosability. Speciation is a process of 
divergence, and therefore these two populations are currently despeciating. With a change in e matte conditions they are 
expected to respeciate. Such merge-and-diverge dynamics may occur frequently in hybrid zones and in relatively young 
radiations in habitats mulie cl to strong environmental fluctuations 

Key words: allopatric divergence, conservation, evolutionary wo nal; hybrid fitness, imprinting, introgression, song. 
species barrier, species concept. 


Two inter-related questions in evolutionary biology radiation. One key component is adaptive divergence 
have been discussed intensively in the last decade. in resource (food) exploiting traits. Three major lines 


These are: what are species, and what role does of evidence demonstrate the role of natural selection 


hybridization play in speciation? The debate about in divergence (P. R. Grant & B. R. Grant, 1997a). 
species is about concepts and criteria (Coyne & Orr. First. populations differ ecologically on different 
2004; de Queiroz, 1998, 2000: Harrison. 1998: Hey, islands that have different arrays of food resources 
2001: Mayr. 2000, 2001; Orr, 2001; Shaw, 2001; Wu. (Abbott et al.. 1977; Smith et al.. 1978). Second. the 
2001a, b). The debate about hybridization centers on particular species on an island and their beak sizes 


whether it has a creative role in speciation through are predictable to a large extent from the food supply 


enhancement of multilocus genetic variation, or (Schluter & Grant. 1984). Third. when the food 


whether it has no effect or even retards or reverses environment changes, evolutionary change occurs as 


the process (Mayr, 1942; Stebbins, 1950; Svürdson, a result of natural selection on heritable variation in 
1970: Mecham, 1975: Raven. 1976: Arnold. 1997: beak size and shape as well as body size (Boag & 
Hercus & Hoffmann, 1999: Barrier et al., 2001 Grant, 1981; B. R. Grant € P. R. Grant, 1989; P. R. 
Barton, 2001: Martinsen et al., 2001; Ortíz-Barrientos — Grant & B. R. Grant, 1995, 2002a). 

et al., 2002: Rieseberg et al., 2003). The two subjects A second key component is divergence in courtship 
are inter-related because when populations hybridize, Signals and responses leading to reproductive iso- 
and therefore are not completely isolated reproductively lation of two populations derived from one. This is the 


from each other, the question arises as to whether they focus of our article. We address four questions: whal 
should be considered as one species or two. constitutes a barrier to gene exchange between closely 


e address the two issues of species and — related species, how does it originate. how might it be 


= 


hybridization by discussing recent findings from field breached, and what are the consequences of occa- 
studies of Darwin's finches on the Galapagos islands. sional hybridization? Answers to these questions 
The studies are designed to investigate the causes of constitute the background to a concluding discussion 


speciation within the broader context of an adaptive of species as taxonomic: units, 


Our research has been carried out under the auspices of the Servicio Parque Nacional Galápagos. Charles eae 
DE and Charles Darwin Research Station. It has been supported by research grants from NSERC (Canada) and N 
(L. S. K.). We thank the numerous assistants who have helped with fieldwork, listed in previous publications. 

x 11 155 nt of Ecology and Evolutionary Biology, Princeton Unive rsily, Princeton, New Jersey 08544-1003, U.S.A. 


ANN. Missouri Bor. Garp. 93: 94-102. PuguisHED oN 31 May 20006. 


Volume 93, Number 1 
2006 


Grant & Grant 95 
Speciation 


THe BARRIER TO GENE EXCHANGE 

Darwin’s finch species are similar in the type of 
as far as we can tell. in their 
1947: Ratcliffe & Grant. 
1983a). Species in the same genera are similar in 


P. R. 


two. traits 


nest they build and. 
courtship behavior (Lack. 


P 


plumage but differ in beak morphology and song 
Grant. 1999). 


would appear to be the best candidates for barriers to 


Differences in these latter 


gene exchange and, indeed, field experiments with 
sympatric species in the ground finch genus Geospiza 
Gould have confirmed their potential role (Ratcliffe & 
1983a, 1985). 


The two sets of traits are interestingly different. Beak 


Grant, 


size and body size are polvgenic traits that display high 
heritabilities (P. R. Grant & B. R. Grant, 2000: 
et al., 
culturally transmitted from father to son (B. R. 
P. R. Grant, 19964). 1983) d 


lab experiments that songs of Darwin’s finches are 


Keller 
2001). whereas songs, sung only by males. are 
Grant & 


ted with 


Bowman ( 


learned, in an imprinting-like manner, between 
approximately day 10 and day 30 after hatching. These 
20 days cover the period from the last few days in the 
nest to the end of the fledglings’ dependence on their 
parents. The father sings throughout this time. Our field 


studies support Bowman's laboratory findings (B. R. 


Grant & P. R. Grant, 1989: Gibbs, 1990). They show 
that a strong resemblance exists between song 


characteristics of sons and fathers and even paternal 
erandfathers but not maternal grandfathers (B. R. Grant 
1996a). 


Often exceptions are more revealing than the rule. 


P. R. Grant. 


Imprinted traits, being learned, are subject to mis- 
imprinting if young birds are exposed to the song of 


a different species rather than, or more than, the 


fathers song, during the short sensitive period of 
learning. This has been observed in a study of Geospiza 
fortis Gould (medium ground finch) and G. scandens 
Gould (cactus finch) on the small island of Daphne 
Major (B. R. Grant & P. R. Grant, 1996a, 1998; P. R. 
Grant & B. R. Grant, 1997b). Misimprinting happens 
rarely, in three types of circumstances. It occurs when 
(1) a father dies and the offspring learn the song of the 
neighbor, which can be the 


nearest natal 


another species: (2) nests of two species are close 
together and one male sings louder, more persistently. 
and chases off the other male; and (3) a nest with one 
of the original eggs in it is taken over by a pair be- 
longing to another species, the egg hatches, and the 


nestling is fed by and learns the song of its foster father. 


INTROGRESSIVE HYBRIDIZATION 


Following the fates of 16 misimprinted finches has 


revealed that they mate mainly according to song type 


bird of 


and not morphology. and consequently hybridize and 
produce hybrid offspring. If they are males, they sing 
the same single song as sung by their fathers an mate 
with a female whose father sang that song type (P. R. 
Grant & B. R. Grant, 1997c). If they are MA > 


choose a mate that sings the same song as their fathers’. 


Even though Fy hybrids and backcrosses are morpho- 
logically intermediate between the parental species, 
they breed strictly according to song type (B. R. Grant 
& P. R. 1998). ' 


misimprinting, introgression of genes occurs 


Grant, Thus. through occasional 
both 


directions between Geospiza fortis and G. scandens. 


HYBRID FITNESS 


Hybrids have been formed rarely (~1% of breeding 
pairs are interspecific), but approximately continuous- 
ly. throughout our 30-year study on Daphne Major. 


However, in the first 10 years none survived to breed. 


At this time we considered two possible reasons for 
their failure: genetic incompatibility and ecological 
insufficiency. Few birds, hybrids and non-hybrids, 
survived the generally poor feeding conditions in the 


beak 


sizes. were unable to crack the large and hard seeds 


early years. Hybrids, with their intermediate 


of Tribulus cistoides (L.) that the more robust-billed 
took 


significantly longer to crack open the seeds of Opuntia 
g y 8 | l 


Geospiza fortis were feeding on, and they 
echios Howell, which is the main dry-season food of G. 
scandens (B. R. Grant & P. R. Grant, 1996b). After the 
unprecedentedly high rainfall in the El Niño year of 
1983. the ecological conditions on the island changed 
dramatically (P. R. Grant et al., 2000b). 


common large and hard seeds of T. cistoides and O. 


The previously 


echios diminished in the seed bank, these plants having 
been smothered by vines and other small-seed pro- 
ducing plants. Under these altered. conditions of an 
and soft seeds, hybrids and 


abundance of small 


backerosses survived to reproduce. Therefore. their 
previous failure is attributable to ecological insuffi- 
ciency and not to genetic incompatibility. 

After 1983, hybrids survived as well as, if not better 
than, their parental species, and reproduced as well as 
them, there. being no difference in number of eggs. 
nestlings, or fledglings produced (P. R. Grant & B. R. 
1992). We detected no loss of fitness in the Fy 


and backcross generations in terms of relatively poor 


Grant, 


survival, mating success, or reproduction (P. R. Grant 
& B. R. Grant, 1992; B. R. Grant & P. R. Grant, 1998; 
P. R. Grant et al., 2003; but see Barton, 2001). 


IMPLICATIONS OF INTROGRESSION 


implications of the 
First, 


There are two important 


documented. introgressive hybridization. post- 


Annals of the 
Missouri Botanical Garden 


mating isolation appears not to have evolved between 


these and other pairs of ground finch species (B. R. 


Grant & P. R. Grant. 1989, 1998: P. R. Grant et al.. 
2005). Even more distantly related species of 
Darwin’s finches have been known, or suspected, to 


me 


of fitness (P. 


This is not surprising. 


hybridize without a recorded loss 
1999). 


comparative evidence from other bird species it takes 


Grant, According t 


more than two million years for post-zygotic in- 


compatibilities to evolve. (Price & Bouvier, 2002). 


All. or almost all, of the adaptive 


Darwin's finches took place in less than this time 


(P. R. Grant, 1999; Sato et al. 2001: Burns et al., 
2002). Second, hybridization and backerossing has 


enhanced the evolutionary potential of these two 


species by increasing additive genetic variances 
while reducing constraints from genetic correlations 


(P. R. Grant & B. R. 1994). 


strength of correlations. between beak size 


Grant, The reduction in 
genetic 
trails. is due to the interspecific. differences in 
allometries. In. principle, a third. implication is that 
hybrids have the potential to invade new habitats 


Birch. 


species in another environment (see Rieseberg, 200€ 


(Lewontin & 1906) and even form a new 


- 


this issue). We conjecture that introgressive hybrid- 
izalion may have contributed to speciation in the past. 
although we lack evidence that this happened in any 


particular case. 
THe ORIGIN or A BARRIER TO GENE EXCHANGE 


Based on these findings, reproductive isolation of 


Darwin's finches is explained by a theory of 
discrimination by song in association with morphology 
that is learned in a sexual imprinting-like process 
early in life (B. R. Grant & P. R. Grant, 2002: P. R. 
Grant & B. R. Grant, 2002b). 

We now apply contemporary findings from Daphne 
to the question of how reproductive isolation evolves. 
using as a framework the allopatric model of 
speciation. The model invokes ecological adaptation 
in allopatry as a requirement for the establishment of 
sympatric coexistence of two species derived from 


one. For this there is plenty of evidence (see 


introduction to this article). A total or near-total lack 


of interbreeding is another requirement. The mode 
invokes an allopatric origin of a pre-mating barrier to 
gene exchange. Experiments show that some degree of 
differentiation in the cues used in mate choice must 
arise allopatrically. 


Experiments with museum specimens from different 


— 


islands were conducted to simulate the second ary 
contact phase of the speciation cycle (Ratcliffe & 
1983b). In a 


ground finch species, we consistently found weak or 


Grant, variety of tests with different 


radiation of 


no discrimination between an immigrant and a (resi- 
dent) member of the responding bird’s own population 
when beak differences between them were small, and 
strong discrimination when the differences were large. 
cues were controlled for in these 


Acoustic experi- 


ments by eliminating them; dead museum specimens 


do not sing! 

Playback. experiments without museum specimens 
as visual cues showed that ground finches discrimi- 
nale acoustically between residents and immigrants in 


a similar way to their visual diserimination: they 


discriminate when differences are large and fail to do 


so when they are small (Ratcliffe & Grant. 1985) 


The morphological barrier arises as a result o 
adaptation. As such it conforms to an old idea dating 
back al (1937) that 


reproductive isolation originates as a byproduct of 


least 65 years to Dobzhansky 


adaptive differentiation. We know less about how the 
2002; 


2b). One possibility Is 


song barrier arises (B. R. Grant & P. R. Grant. 
& D. 
that song characteristics covary with beak size for 
the 
range of frequencies of individual notes (Podos, 1997) 


P. R. Grant R. Grant, 200 


biomechanical reasons, affecting. trill rate and 
and, therefore, when beak size changes adaptively, 
songs change as a passive consequence. There is 


2001): 


less, correlated. effects of beak change appear to be 


some supporting evidence (Podos. neverthe- 
PI : 


insufficient to set up a barrier to gene exchange 
because coexisting sister species of Geospiza do nol 
differ discretely in these two song features (B. R. 


Grant & P. R. 2002). In 


interspecific similarity, populations of 


— 


Grant, contrast to this 


the same 
species may differ substantially in songs independent 


po Wa). 


ability of 


of beak size differences (P. R. Grant et al., 
the 


sympatric species with different beak sizes to sing 


The hypothesis is undermined by 
each other's songs with scarcely altered. character- 
1983: B. R. Grant & P. R. Grant 
1998: P. R. Grant & B. 1997b. c) 
effects of may 


contribute to a 


istics (Bowman, 
1989, 


Correlated 


h. Grant, 


body size changes also 
influence on 
Bowman, 1983). 


Since beak size and bodv size generally covary. a third, 


barrier. through an 


frequency characteristics of songs 


— 


and more realistic, hypothesis is that they are jointly 
involved in determining characteristies of songs and 
hence the barrier to interbreeding. Another possibility 


that a new song type may originate as a cultural 


mutation through miscopying of father’s song and 


increases in frequency by chance or selectively. A 


selective advantage may arise if the songs transmit 


better in the new environment and. as a result. more 


effectively repel intruding males or attract. females. 


The origin of reproductive isolation is still not fully 


understood, and these possibilities need further 


Investigation. 


Volume 93, Number 1 Grant & Grant 97 
2006 Speciation 

ADAPTIVE DIVERSIFICATION THROUGH REPEATED SPECIATION 2002: P. R. Grant & B. R. Grant, 20026). On the 

island of Genovesa, occupied by Certhidea fusca 

The foregoing explanation of how two species are Sclater, we tested birds with their own songs, songs 


a 


formed from one serves as a basis for explaining the 


Grant & B. 


Grant, 2002c). Fourteen species evolved from a single 


o) 


adaptive radiation as a whole (P. R. 


ancestral species by a simple repetition of the process 
of division, with the species produced at each step 
differing according to the particular ecological 
circumstances that guided each pathway. This was 


(P. R. 1981, 1999). 


Furthermore, we believed most niche space would 


our starting point Grant, 


be occupied in the early history of the radiation. with 


ater speciation resulting in gap-filling. as illustrated 


recently by Jonathan Losos and colleagues with 
Caribbean Anolis Daudin lizards (Harmon et al. 


2003: Losos et al., 2006 this issue). For experimental 


and observational work we concentrated on the ground 


finch twigs of the phylogenetic tree, because these 
species are most similar to each other and have been 
genealogically separated for a short time. To see if the 
same processes of divergence in morphology and song 
occurred early in the history of the radiation, we have 


gone to the base of the tree. There, five years ago. we 


encountered a surprising fact. The initial divergence 
is frozen at the allopatric stage. Reconstructions that 
have used both mitochondrial and nuclear genetic 
markers (Freeland & 1999a, b), including 
microsatellite DNA (Petren et al., 1999), show a basal 


split between two groups of warbler finches that are 


Boag, 


now allopatric on different sets of islands (Petren et 
al., 1999; 


Why have they remained allopatric? What has 


Tonnis et al., 2005 


prevented them from establishing sympatry? We do 
not have complete answers to these questions. Part of 
any answer requires greater knowledge than we have 
of ecological opportunity for joint occupancy of an 
island. Another part requires knowledge of potential 
reproductive isolation, and this we do have. 
Morphologically the two groups of warbler finches 


are similar; correspondingly, they are similar ecolog- 


— 


ically as well. They also share the peculiar feature of 
reversed sexual dimorphism in beak length. in which 
trait they differ from all other Darwin's finch species 
and from all continental relatives (P. R. Grant & B. R. 


2003). 


extent in plumage color and beak size, 


Grant, They differ from each other to some 
but scarcely 


enough to prevent them from interbreeding to judge 


— 


from differences in pairs of sympatric species o 


Darwin's finches. However, their songs do differ (B. R. 
Grant & P. R. 2002). With a 


question of the origin of reproductive isolation, we 


Grant, focus on the 
performed a set of playback experiments. similar to 
earlier ones in which we simulated immigration of 
birds from another island (B. R. Grant & P. R. Grant, 


from another population of C. fusca, and songs from 
two populations of C. oliracea Gould. We performed 
similar experiments with the same playback tapes on 
olivacea. We 


an island, Santa Cruz, occupied by C. 


expected to find strong discrimination, manifested in 


— 


the pattern of contrasts shown in Figure 1 (above). 


Instead we found little evidence of discrimination of 
island 


heterotypic (heterospecific) song on either 


Fig. I. below). This is fascinating because it contrasts 


with clear discrimination between sympatric and 


vounger species. One would have thought warbler 
finches of the two groups have surely had enough time 
to diverge in both beak morphology and song to enable 
them to coexist without interbreeding. Yet the 
implication of the playback results is that the oldest 
(allopatric) lineages would interbreed, whereas the 
youngesl (sympatric) ones do not, or do so very rarely. 
Moreover, the genetic evidence tells us that despite 
their apparent potential to exchange genes, they have 
not done so, at least not to an appreciable extent, for 
a long time (Petren et al., 2005). 

There are three other 19 5 ations of this striking 


contrast (B. R. Grant & P. R. Grant, 2002). Th 


that rates of song and morphological divergence vary 


e first is 


among the species; hence, the rate of evolution of 
sexual isolation varies considerably and is not a simple 
function of time. The second is that either divergence 
in song is accelerated in sympatry. perhaps by 
selective reinforcement of initial differences if off- 
spring of mixed matings are at a selective disadvan- 
tage, or establishment of sympatry is made possible 
only by the acquisition of pronounced song differences 
in allopatry. The third implication is that a knowledge 
of neutral genetic differences among taxa in young 
radiations like Darwin’s finches is not enough for 
inferring their potential to interbreed and exchange 
genes. Geographical, ecological, and behavioral in- 


2005; 


Ferguson (2002) has argued 


formation is needed as well (Petren et al. 
2005). 


similar points more generally. 


Tonnis et al., 


SPECIES RECONSIDERED 


(2002). the 


question of systematic biology is. what is a 


According to Cracraft first greal 


species? 
Templeton (1989) has suggested this question must be 
answered before the process of species formation can 
be investigated. Like other evolutionary biologists 


attempting to understand speciation, we have not 


© 


heeded this advice. Instead of defining the species we 
study, we have found it sufficient to simply invoke the 


biological species concept with its emphasis on the 


Annals of the 
Missouri Botanical Garden 


Expected Response to Playback 
1k SC G 


Observed Response to Playback 


| SC G 
I 
P 
CF 
0 


C. olivacea C. fusca 


Santa Cruz Genovesa 
Figure |. E 


sponses of warbler finches on Santa Cruz ( 


xpected (above) and observed. (below) 
Certhidea olivae a 
and Genovesa (C. fus 


experiments were conducted. Controls were Cassin’s finch 


(CF) songs recorded in North s rica. Populations on Santa 
Cruz and Isabela belong to the olivacea e a and 
populations on Genovesa and cd belong to the l 
Figure based on B. R. 
2002 by the 


C. fusca 
2002: 


lineage. Grant & P. R. Grant 


fig. 3 (* University of Chicago). 


reproductive isolation and not on 
classification (Harrison, 1998: Hudson & 
2002). De Queiroz (1998: 63) has 


paleontologist G. G. Simpson (1961) in arguing 


evolution of 
Coyne. 
followed the 
“there 
is really only one general species concept in modern 


and 


SYS tematic evolutionary biology species are 
segments of population level evolutionary lineages.” 
He then offers an important distinction: “A species 
concept is an idea about the kind of entity represented 
that is, about the kind of 


by the species category, 


entity designated by the term species. A species 
criterion is a standard for judging whether a particular 
entity qualifies as a member of the species category, 
that is, for judging whether a particular entity is or is 
(de Queiroz, 1998: 65). 


bypasses the plethora of 


not a species" This neatly 
tangled arguments about 
species concepts and allows debate to be focused on 
how species should be recognized. 

Ernst Mayr (1942) defined species as groups of 
actually or potentially. interbreeding natural popula- 
lions, which are reproductively isolated from other 
such populations. The criterion here is clearly one of 
this definition to 


interbreeding. Application of 


individuals in order to refer them to species has 


repeatedly encountered practical difficulties with 
allopatric populations and with populations connected 
occasionally or rarely by interbreeding (Zink & 
McKitrick, 1995; see & Wollenberg, 1997, 


and Wake, 2000 this issue). These practical difficul- 


Avise 


lies have been faced head-on by ornithological 


systematists. For example. the Taxonomic Subcom- 
mittee of the British Ornithologists’ Union confronted 
assist in the 


British Bird 


They started with a position statement: 


them it guidelines to 


drawing up 
assessment of species rank for the 


Species List. 


— 


“If we define species as population lineages main- 


taining their integrity with respect. to other such 


lineages through time and space, this means that 
species are diagnosably different (otherwise we could 
not recognize them), reproductively isolated. (other- 
wise they would not maintain their integrity upon 
contact) and members of each (sexual) species share 


a common mate recognition and fertilization system 


(otherwise they would be unable to reproduce 
(Helbig et al., 2002: 519). 


two recommendations. First. 


Then they followed with 


"Diagnosable taxa will be 
ranked as species if they are broadly sympatric. i.e.. 
over areas beyond the average natal dispersal distance 


of the species involved.. .. and do not hybridize... « 


hybridize only rarely, so that gene flow between them 
either does not occur (because hybrids are sterile or 
reasons) or occurs al such 
(often difficult to detect) that it is 
(Helbig 


populations 


do not backcross for other 

low frequency 
unlikely their gene pools will ever merge...” 
et al., 2002: 522). 


Second, allopatric 


Volume 93, Number 1 
2006 


Grant & Grant 99 
Speciation 


should be considered different species "if they have 
diverged to the extent that merging of their gene pools 
(Helbig et al., 2002: 519), 
"predictions. about possible reproductive 
that differ 


slightly (e.g., in size or darkness of plumage) are very 


in the future is unlikely? 
whereas, 
isolation between allopatric taxa only 
uncertain. Such taxa are best treated as subspecies” 
(Helbig et al., 2002: 519). 


and recognizing our ignorance of post-mating com- 


On the basis of the latter. 


patibility, the warbler finches are best treated 
subspecies. We consider them to be well-differenti- 
ated genetic lineages. 


These prescriptions for sympatric and allopatric 


species fit very well with our past 30 years « 


thinking and practice based on the biological species 


concept and notwithstanding the fuzziness (Harrison, 


1998; Hey, 2001: Lee, 2003) of species boundaries 
caused by rare hybridization. The prescriptions 


clearly apply to sexually reproducing species gener- 
ally and not just to birds. Nevertheless Darwin's finch 
studies expose two hidden problems with the di- 
agnosis of sympatric species. First, species may lose 
diagnosability. Second, if they do it is an arbitrary 
matter to decide when they lose their species status. 

At the beginning of our study on Daphne Major 
diagnosably 


scandens were 


different on morphological criteria (P. 


1993). With hybridization backerossing and natural 


Geospiza fortis and G. 
R. Grant, 


selection occurring, the populations have converged 
in morphology and are now more similar than they 
were at the beginning (P. R. Grant & B. h. 
2002a; P. R. 2004). They 
10096 diagnosably different on the same morpholog- 


Grant. 
Grant et al., are no longer 
ical criteria. Have they lost their species status in 


a couple of decades? They are merging, but ver 
slowly. Members of the two populations can still b« 
recognized by us on the basis of the song they sing 
(males) or the song of their mates (females). as well as 
by beak morphology in almost all instances. The 
problem that this poses for the delineation of species 
is that song is a learned, culturally inherited, trait, 
without a genetic basis in many species of birds, and 
yet 7... characters. used in diagnosis must be the 
dli E evolution: they must be genetically based and 
'aused by environmental factors such as 
(Helbig et al., 2002: 520). 


the basis of 


not merely « 
nutrition” A way out of this 
dilemma is to broaden 
that 


important as genetic inheritance in trans-generational 


diagnosis by 


recognizing cultural inheritance may be as 


maintenance of biologically important traits used 

reproduction. We recommend this because song is the 
cultural equivalent of the Y chromosome in these 
birds: paternally inherited, non-recombining, and 
subject to change only by mutation, cultural mutation 
variation has been found to be 


in this case. Song 


useful in systematic studies of other groups of birds 
(Ahlstróm & Ranft, 2005). 

The second problem has no simple solution. “We do 
not know of a single documented case of breakdown of 
to full. re- 


productive compatibility between species that were 


reproductive isolation (i.e. a reversal 
incompatible before) in any class of organism? (Helbig 
al.. 2002: 521). Kat (1985) described 


a possible case of breakdown of reproductive isolation 


Howe Ver, 


= 


with two species of Anodonta L. freshwater mussels. 


According to fossil evidence, they coexisted without 
interbreeding about 200,000 years ago, and now at the 
The 


irst 


same locality they hybridize relatively commonly. 


— 


two finch species on Daphne may become the 
exception to be witnessed in progress. At present they 
are on a reversed course of speciation (P. R. Grant et 


al.. 2004), in fact they are despeciating, and the point 


at which they should no longer be considered two 
species, but only one, is arbitrary. Morphological and 
genetic convergence could easily be reversed if there 
is a change in climate and food composition, in which 
case they will be respeciating. Many other pairs of 
taxa, not just Darwin’s finches, might go through 
some 


2000). 


possibly also in young adaptive radiations such as the 


similar merge-and-diverge oscillations, as in 


shifting hybrid zones (e.g., Carney et al., 
silversword alliance in Hawaii (Barrier et al.. 2001: 
Baldwin, this volume) and cichlid fish in the African 


Great Lakes (Salzburger et al., 2002; Smith & 
Kornfield, 2002). 
Such fluidity can be bewildering to those who 


recognize species by inferred ancestry rather than by 
interbreeding and its consequences. For instance Zink 
(2002), 
differentiation in mitochondrial DNA among species 
(Freeland & Boag, 1999a, b; 1999) that are 


upheld by other criteria, has argued there are far fewer 


basically restricting attention to the lack of 
Sato et al., 


species of Darwin’s finches than are currently 


recognized, perhaps only half. Aside from an un- 
fortunate reliance on a single, potentially misleading. 
molecule (Ballard et al., 2002; Hudson & Coyne, 
2002; Ferguson, 2002; Machado & Hey, 2003: Rokas 
et al., 2003), a view such as this overlooks the fact 
that as many as 10 species coexist on the same island. 
ecologically differentiated, morphologically recogniz- 
able, and reproductively isolated by and 
morphology (P. R. Grant, 1999). 


These problems illustrate the well known fact that 


song 


systematics deals with discrete categories, whereas 
evolutionary change is a continuous and gradual 


Harrison, 1998; Hey, 2001; Hey et al., 2005). 


valuable to evolutionary biolo- 


process 
Darwin's finches are 


gists precisely because they so well exemplify the 


graded nature of evolutionary transitions. They are 


challenging to systematists for the same reason. 


100 


Annals of the 
Missouri Botanical Garden 


CONSERVATION OF BIODIVERSITY 


Darwin's finches are valuable to conservation 


biologists for another reason. They deliver a message 
about preserving biodiversity: since species, 


environments, are continually changing, both must be 


conserved to allow adaptation to further change. For 


any given focal species, or evolutionarily significant 
(Hey et al., 2003), 


species in the community are part of the environment, 


population of the species other 


be they food, competitors, predators, or parasites. This 


is especially clear in the case of interbreeding 


species, because their ecological and evolutionary 
future depends to some extent upon the genes they 
exchange with each other (Cade, 1983: B. R. Grant & 
1989; Rhymer & Simberloff, 1996). 


clear with competitor species that persist only when 


P. R. Grant, It is 


predators prevent one species from competing the 
2002). It is clear from the 
Grant € P. R. Grant 
1983) on one o 


seasons of 


others to extinction (Paine. 
dependence of finches (B. R. 
1989) and primates (Terborgh, 


specific 


r [wo 
food acute food 
And 
effects that are manifested in the trophic networks of 
Wootton, 1993; 


than 


sources al 
limitation. is clear from the myriad. indirect 
complex communities (e.g.. 
2003). 


species, are the appropriate units in 


Dayton, 


Thus ecosystems, rather individual 
programs of 


long-term conservation of biodiversity, 


Literature Cited 


Abbott, I., L. K. 


ecology ol 


Abbott & P. R. 


Galapagos ground 


1977. 
Evaluation of the 


Grant. s 
finches: 
importance of floristic dive Sil and interspecific compe- 
Ecol. Monogr. 47: 151-184. 
JWR, 2003. The 
and the HOUR of bird sound archives. 
rnit - Club 123A: 114-135. 
a M. L. 19 Vide F and 


tution. 


Ranft. use of sounds in avian 


evolution. 


Oxford Univ. e ress, New ] 
Vise, J. C. . Wollenberg. -1997.] Phylogenetics and bp: 
origin T species. Proc. Natl. Acad. Sei. U.S.A. 94A: 


7748-7755. 
Baldwin. B. G. 


evolutionary 


2006. Contrasting patterns and processes of 


change in the tarweed=silversword lineage: 


Revisiting Clausen, Keck, and Hiesey's findings. Ann. 
Missouri Bot. Gard. 93: 64—93 
Ballard, J. W. O. U. Chernoff & James. 2002. 


Divergence of mitochondrial DNA is not corroborated by 


nuclear DNA, morphology, or behavior in 


Drosophila 
simulans. s 56: 527-545. 
Barrier, M., R. I 

Accelerated 


Robichaux & M. 2001. 


an adaptive 


D. Purugganan. 


11 5 gene evolution in 


radiation. Proc. Nath Acad. Sei. U.S.A. 08: 10208— 
10213 

Barton, N. H. 2001. The role of hybridization in evolution. 
Molec. Ecol. 10: 551-508. 


Boag. P. T. & P. R. Grant. 198 


a population. of Mere finches (Geospizinae) 


. Intense natural selection in 


n 
Galápagos. Science 214: 82-85. 


like their 


Bowman, R. J. 1983. The Jarwin's 


evolution. of song 1 
finches. Pp. 237-537 in R. J. 


Leviton e Patterns of Evolution in Galápagos 
Organisms. American Association for the Advancement 
of Science, Pacific Division. San Francisco, California. 
Burns, K. J., S. J. Hackett & N. N. Klein. 2002. Phylogenetic 
relationships and morphological diversity in Darwin's 
fine 3 . Evolution 56: 1240-1252 


and ES ir relatis 
19 


Cade, T. J. .H ybridization and gene exchange among 
in re a lo conservalion. 5 288-3009 in C. 95 
Schonewald-Cox, . Chambers, 
rene E s and C onservation, 
Menlo Park. 
Carney, S. E., K. A. dide & L. H. 
over the fifty- 
a shes population of sunflowers (He 1 1 Evolution 
: 462—474. 
; & A. R. Orr. 
9 Massac husells. 
Cracraft, J. 2002. 
biology: 


Thomas (editors). ( Be 5 1 
G ummings. 
2000. 


history of 


Rieseberg. 


Evolutionary changes 


mee 2004. Speciation, Sinauer, 


The seven great questions of systematic 
An essential foundation for conservation and the 
sustainable use of biodiversity. Ann. Missouri Bot. Gard. 
89: 127-144. 
Dayton, " K. 2003 
lo conservation. 
1998. 
species criteria, and the process of spe 
in D. J. Howard & S. H 


Forms: Species and Speciation. Oxford Univ. 


. The importance of Aa natural sciences 
Naturalist 162: 1-13. 


The general um 'age concept of spec cles, 


ere 
de Queiroz. K. 
lation. Pp. 
ie (e 8 11 . 
Press; New 
( 

2000. The general lineage concept of species and 

the ds 1 7 5 properties of the species category. Pp. 49-89 
i R. A. Wilse 


Essas s. M M Press, ( 


n (editor), Species: New Interdisciplinary 


vambridge, Massachusetts 


Dobzhansky, T. 1937. Genetics and the Origin a Species. 
Columbia Univ. Ney r 

Ferguson, J. W. H. 2 P x use of genetic divergence 
for identifying species. > Biol. J. Linn. Soe. 75: 509—510. 

Freeland. J. R. & P. T Boag. 1999a. Phylogenetics of 
Darwin's finches: Paraphyly in the tree finches, and two 
divergent lineages in the warbler finch. Auk 116: 
571—588. 

& ———. 1999b. lear 


The mitochondrial and nuc 
genetic homogeneity of the phenotypically diverse Dar- 
wins iua finches. Pus 53: 1553-1563 

Gibbs, L. 1990. Cultural evolution of male song types in 


RA medium ground finches, Geospiza fortis. Anim. 
Behav. ap 759-771. 
Grant, B. R. & P. R. Grant. 1989. Evolutionary Dynamics of 


a 1 5 The Large 


7 


Cactus Finch of the 
oe d Univ. Chicago Press, Chicago, Illinois. 


——— XK ———. 19064. Cultural inheritance of song and 
its role in the e Usus of Darwin's finches. E duos 50: 
2471-2487. 

V ———. |996b. High survival of Darwin's finch 
hybrids: effects ul beak morphology and diets. Ecology 77: 
500—500 

K 1998 Hybridization and speciation in 
Darwin’s finches. sexual 
a culturally transmitted trait. Pp. 
Howard & S. H. Berlocher E ie s 


Species Ban Speciation. Oxford Univ. Press, New York 


yee ae on 
104-422 in D. J. 
(editors), Forms: 


A & — ——. 2002. Lack of premating isolation at v 
\mer. Naturalist 160: 1— 
Speciation and the w 1 i 


Sci. 69: 653— 


base E a phyloge netic tree. 
Grant, P. R: 


e s ^u 55 S. Amer. 


Volume 93, Number 1 
2006 


Grant & Grant 101 


Speciation 


ä 1993. Hybridization of Darwin's finches on Isla 
d Major, Galápagos. Philos. Trans., Ser. 340: 


om Ecology and Evolution of Darwin's Finches. 
Princeton Univ. Press, Princeton, New Jersey 

& B. R. Grant 1992. Hybridization of bird species. 
d 256: 193-197. 

. 1994. * anche genetic effects of 


ERE UTE in Darwin E e thes. Evolu 
( 


1995. Pre die ‘ting microevolutionary 


responses to directional se lection on heritable variation. 


Evolution 49: 241-251. 
= 1997a. Genetics and the origin of bird 
species. Proc. Natl. Acad. Sei > 7768-7775 
1997b. dao diis imprinting 
and male choice. Amer. Naturalist 149: 1— 
. 19970. Mate patterns of 11 s finch 
hybrids Biol. J. Linn. Soc. 60: 317-343. 
— ———, 2000. Ola genetic variation in 
"E E of Da wwin's finches. Pp. 340 in T. 
Mousseau, B. Endler a \daptive 


& 
Genetic lona in ume W ild. Oxford Univ. Press. New 


Yor 


E ——— 2002a. Unpredictable d in a 30- 
year study of Darwin’s finches. Science 296: 70 7—1 711. 
— & ——— 2002b. Simulating — contact in 
allopatric speciation: An empirical test of. premating 

isolation. Biol. J. Bus Soc. 76: 545—5 e. 
& ———. 2002c. Adaptive radiation of Darwin's 
finches- Amer. ea 90: 130-139. 
. 2003. Re verse p se dd dimorphism in the 
Ibis 145: 341-3 


beak of a finch. 
& p nn e ii allopatric phase of 


meet The be raked ground finch (Geospiza 


difficilis) on the Galápagos islands. Biol. J. Linn. Soc. 
69: 287-317. 
. Keller & K. Petren. 2000b. Effects 


of E : Nino events on Darwin's finch productivity. Ecology 
81: 2442-2457. 

— —g̃— — arkert & K. Petren. 2003. 
Inbreeding and inte ds e 11 in Darwin’s finches. Evolu- 
tion 57: 2911-2916. 

& K. 


caused by 


— Markert. L. F. Kelle Petren. 
2004. Conve ent Tu NR n of Darwin's finches 
introgressive rybridization and selection. Evolution 58: 
1588-1599, 

& K. Petren. 2005. Hybridization in the 
Naturalist 106: 50-07 
II. X. J. a & J. B. 


radiation in 


recent past. Amer. 
Eom L. J., J. A. Se Losos. 
2003. Tempo and d 

: BOL: 961—9€ 


iguanian ian Scier 


Harrison, R. 1998. L king evolutionary 9 5 rn and process: 
The relevance of species concepts for the study of 


speciation. Pp. 19-31 in D. J. Howard & S. H. Berlocher 


(editors), Endless Forms: Species and Speciation. Oxford 
Y A 


Iniv. Press. New 
Helbig, A. J., A. G. 19 10 D. T. Parkin. G. Sangster & M 
Collinson. 2002. Guidelines for assigning species E 
Ibis l | |: 518-525 
Hercus & . Hoffmann. 1999, Does interspecific 


1 influence evolutionary rates? An e 
tal study of laboratory adaptation in hybrids between 
Drosophila serrata and Drosophila birchii. Proc. Roy. Soc. 
London. Ser. B 266: 2195-2200. 

Hey. J. 2001. Genes, Categories and Species. Oxford Univ. 


New Y York 


xper imen- 


Press, 


. Waples, M. L. Arnold, R. K. Butlin & R. G. 
pu oe and confronting species 
Trends Ecol. 


————, Hh. 
Harrison. 


uncertainty in biology and conservation. 


Evol. 18: 597—003. 
Hudson, R. R. & J. A. Coyne. 2002. Mathematical 
consequences of the genealogical species concept. 
Evolution 56: 1557-1505. 


Kat, P. 1985. Historical evidence for 5 in levels 
of o 1 e 116 

Keller, L. F., P. R. € . B. R. Grant A K. Petren. 2001. 
1 of 1 traits in Darwin's finches: 
Misidentified paternity and maternal effects. Heredity 87: 
325-336 

Lack, D. 194 S Finches. Cambridge Univ. Press. 
Cambridge, U 

Le S. Y. 200 03. Species concepts and 1 ies reality: 
Salvaging a là innaean rank. J. Evol. Biol. 79-188. 

Lewontin, R. & L. C. Birch. 1900. 1 üs 


a source of variation. for adaptation to new environments. 
E iro 20: 315 
Losos. J. B., R. E. Glor. J. J. Kolbe & K. 
\c s speciation, and convergence: A hic 
analysis of pu radiation i in Caribbean . Anolis lizards. 
93 33. 


2006. 


rarchical 


Nicholson. 


inn. Missouri Dot. rd. 93: 24— 


Machado, C. A. & J. 19 2003. The causes of phylogenetic 


conflict in a classic Drosophila species group. Proc. Roy. 
Soc. London. Ser 3 210: —1202. 
Ma s .G. D. T Whitham. R. J. Turek € P. Keim. 


)01. lieb pains selectively filter gene introgres- 
33: 


u 
N 
yY 


sion . tween species. Evolution 55: 
94 


Mavr, 1942. eae and the origin of species. 
C VM Univ. Press, New York 

2000. The biological species concept. Pp. 17-29 in 

Q. D. EY heeler & R. Meier (editors), Species Concepts and 


Theory. A Debate. Columbia Univ. Press, 


Phylogenetic 
New York 
. 2001. Wu's genic view of speciation. J. Evol. Biol. 


Mecham, J. S. 1975. Incidence and significance of in- 
trogressive hybridization in anuran populations. Amer. 


Zool. 15: 83 

Orr, H. A. 2001. Some doubts about (yet another) view of 
species Evol. Biol. 14: 1 

Ortiz- did ntos, D., J. Reiland, J. y & M. A. F. 

2002. Rec 1 and the fs rgence of hybridizing 

11 Genetica 116: 167—178. 

Paine, R. T. 2002. “Trophic « control of production in a rocky 

tertidal community. Science 296: 730-7: 
"i K., B. R. Grant & P. R. Grant. 1999, A 9 8 ny of 


Noor. 


Darwin's finches bass d on microsatellite DNA le ngth 
variation. Proc. Roy. . London. B 266: 321—329. 
. P. R. Grant, B. n Grant & L. F. Keller. 2005. 


C — landscape genetics and the adaptive radia- 
tion of "api bus d Ms role of peripheral isolation. 

Molec. Ecol. 14: 2043-2 

. 1997 i pe nites constraint on the evolution of 
voc mia ons in a songbird (Passeri- 

a E o srizidae). E 1 al 
. Correlated evolution " morphological and 


Nature 409: 


subfamily 
531—551. 


voc al ue structure in Darwin's finches. 
€ 


85 88 
Price, T. D. & M. M. Bouvier. 2002. The evolution of Fy 
postzygolie incompatibilities 15 pe Evolution 56: 


2083-2089. 
Ratcliffe. L. M. & P. R. Grant. 198: 
1 finches (Geospiza, 


a. Species recognition in 
Gould). 


J. Discrimination by 
Behav. 31: 1139-1153 


morphological cues. Anim. 


102 


Annals of t 
Missouri 9 Garden 


. 1983b. Species recognition in Darwin's 


finc le s (Geospiza, Gould). H. Geographic variation in mate 
gra] 
preference. . Behav. 31: 1154-1165. 
& ——— i 985. cies recognition Darwin's 


Gould). "m 
2 different song types. dialects and heterospecific songs. 
cha. 33: 290-307. 

97 


us matics and plant population biology. 


nches NEN Male responses to playback 


Anim. 


— 
5 
2 
P 
EN 


Syst. Bot. 1: 284 
Rhymer, J. M. & P EU" rloff. 
hybridization and introgression. 

T: 8: 3H JO 
Rieseberg, L. H. 20006. Hybrid speciation in wild sunflowers. 
n. Missouri Bot. 


1990, 
\nnual 


E xtine E S 


. Ecol. 


Gard. 93: 34-48 


————. O. Raymond, D. M. Rosenthal, Z. Lai, 
Livingstone, T. Nakazato, J. L. Durphy, A. E. Schwarz- 
bach. L. A. Donovan & C. Lexer. 2003. Ma jor ecological 


transitions in wild i TS fac ilitated by hybridization. 
Science ae I211— 


Rokas. A.. B. N N. King & S. B. Carroll. 2003. 
Genome-se i approaches to resolvi w incongruences in 
molecular phylogenies. Nature 425: 708-804. 

Salzburger. W.. S. Baric & C. AG 2002. S 


pecialion 


via SRT a in East African cichlids? 

Molec. 2 I: —02 

i V. oa. Figueroa, ^ R. Grant. B. R. Grant. 
. Tic & J. Klein. 9 n. “ny s Darwin's finches 


as reve by mitoc hondrial i V sequences. Proc. Natl. 
M US: > 96: 5101— 
E Hugin, ^ 
175 the origin of Darwins finches. 
. Evol. 18: 5 311. 
N R. Grant. 1904. 
Uis in 


. Tichy, Grant, B. R. Grant & J. 


Molec. 


Determinants of 
morphological 5 Darwin's 


\mer. 
2001. 


communities ol 
175-196 


finches. | Natur: alist 123: 175— 


The genealogical view of speciation. J. 


. Biol. 14: 880-882. 


) 
iT 

Simpson, G. G. 19 00 1 of Animal Taxonomy. 
rk 


ress, New 


Columbia Univ. 


Smith, J. N Ie . B. R. Grant. I. Abbott & L. K. 
Abbott. i da oe variation in feeding habits of 


137-1150. 
. Phyloge 'ography of Lake 


Darwin's anad finches. Ecology 59: | 
Smith, Kornfield. 2002 
18 cichlids of the genus Pse 5 


. 


Significance 

of allopatric colour variation. Proc. Re . London, Ser. 
B 269: 2495-2502. 

Stebbins, G. L., Jr. 
C -olumbia Univ. 


1950. Variation and Evolution in P 
Press, New Y rk 

. 1970. Significance of pose ssion in coregonid 
33-59 in C. C. 
Biology of Coregonid. Fishes. 


ants. 


Svárdson. ( 
Lindsey & C. S. Woods 
Manitoba 


1 9 Pp. 


(editors). Univ. 


ress, Winnipeg. 

l'empleton, H 1989. The meaning of species and 
Irs V genetic perspective. Pp. 3-27 in D. Otte & 
J. A. Endler (e ditor) Speciation and Hs Consequences, 


Sunde rland, Massachusetts. 
1983. Five New World Primates. 


Press. Princeton. xe Jersev. 
P. 


ba 


Terborgh. Princeton 


Univ. 


Tonnis. B.. R. Grant, B. R. Grant & K. betten. 2005. 
Habitat se elec ‘Lion and ec uo al speciation in e 
warbler finches (Cert Hora a olivacea s C. fusca). Proc. 
toy. Soc. London, Ser. B 272: 819— 

Wake. D. B. 2006. Problems with species: Patterns and 


processes of species formation in salamanders. Ann. 
93: 8-23. 


Indirect effects and habitat use 


Missouri Bot. Gard. 
Wootton, J. T. 1993. 


intertidal community: 


man 
Interac ne on and interaction 
Naturalist : 71-89 


he genic view of s process of speciation. 


dees sation Amer 

51-805. 

Genes and speciation. J. Evol. Biol. 14: 

Zink. R. M. 2002. 

history of s irwin's finches. Auk 119: 861—897 
& M. Mekitriek. 1995. The debate over species 

idu E its implications for ornithology. Auk 113: 
—719, 


A new perspective on the evolutionary 


POLYPHYLY IN GUETTARDA L. Frédéric Achille; Timothy J. Motley." Porter P. 
(RU BIACEAE, GUETTARDEAE) Lowry Il. and Joël Jérémie? 

BASED ON nrDNA ITS 

SEQUENCE DATA"? 


ABSTRACT 


The genus nda (Guettardeae—Rubiaceae) e m E Lu Da ly 150 55 cles, ranging from eastern Africa through 
the islands of the Indian and Pacific Oceans to the Neotropics. Sequence data from the nuclear ribosomal DNA internal 
1 spacer (ITS) region were used to test the i of Guet parda and its relationships to closely related genera 


results indicate that Guettarda and two smaller genera, Antirhea and Ste nostomum, are ele tic. 


within Guettardeae. The 
"spread Indo-Pacific 


Most Guetta irda species fall into two distinct groups: a Ne 5 'al lineage ha also includes the 
s G. speciosa (the type of the genus), and a New Caledonian lineage that, along M eru and Timonius. 
comprises a dioecious Paleotropical clade. The Hawaiian endemic Bobea, traditionally conside red close 10 Timonius anc 


may have evolved twice within the tribe. The use of traditional evnoecium characters 
TS phylogeny; other features, such as inflorescence arc ite ture, se cur. system, and 1 0 
appear to correlate more closely with the molecular phylogeny. 

Key oe Guettardeae, islands, ITS, Neotropics, Paleotropics, Rubiaceae. 


systemalics. 


biogeography, dioecy, Guettarda, 


— 


2002: Andersson & 


Rubiaceae, one of the largest families of flowering — 1996: Rova, 1999; Rova et a 
plants with ca. 13,000 species and 650 genera Antonelli, 2005). 
(Delprete, 2004), comprise 44 tribes (Robbrecht, As part of a systematic study of the New Caledonian 
1988) currently assigned to three subfamilies (Bremer species of Guettarda, a molecular phylogenetic 
& Jansen. 1991: Bremer et al., 1995; Bremer. 1997; analysis was conducted to evaluate the monophyly of 
Bremer et al., 1999). The genus Guettarda L. and its this widespread tropical genus, whose definition has 
varied considerably. Originally described by Linnaeus 


— 


relatives have been regarded for most of the last two 
as monotypic, Guettarda has progressively 


— 


centuries as a fairly clear-cut group, generally (4753 
recognized as tribe Guettardeae, and sometimes even grown with the description of additional species and 
as a distinct. subfamily under the invalid name the inclusion of at least five other genera: Matthiola L. 
Guettardoideae |= Antirheoideae| (Verdcourt, 1958; and Laugieria Jacq. (Lamarck, 1792); Antirhea 
Bremekamp. 1966). While assigned to subfamily — (^Antirrhoea," Mueller, 1875); and, more recently, 
). a series of recent Tournefortiopsis Rusby and Dicrobotryum Willd. ex 
molecular phylogenetic studies have shown that Roem. & Schull. (Standley, 1931). The broadest 
Guettardeae are best placed in Cinchonoideae definition of Guettarda by far was that of Baillon 


(Bremer & Struwe, 1992; Bremer et al., 1995; Bremer. (1879, 1880), who included all Guettardeae species 


Antirheoideae by Robbrecht (1988 


! We thank K. Wurdack and P. Delprete for helpful advice; D. Lobreau-Callen for her invaluable assistance and for advising 
first author during his pollen study; two anonymous reviewers for pro sea useful comments on an earlier version of the 


the 

manuscript; N. Damico for d: Ip with pre poore poe n A s; and T. Jaffré. . Dagostini, and F. Rigault for their hospitality 
in New Caledonia and for access to facilities at IRD (Nouméa). We m ihe aea des Ressources Naturelles of the 
Province Sud and the Service Forêts, Bois et 5 of the Province Nord in New Caledonia for permission to collect 
specimens; the curators of GB, NOU, NY, and P for providing access to their hr tions and/or for allowing the moval of 
material for mole e le analysis: the curators of CANB, L, MEL, and Z for loans of specimens; and G. Areces and y n for 
supplying leaf material preserved in silica gel This research was conducted with hindins from the as 


Pluriformations—Biodiverstté Terrestre en Nouvelle-Calédonie of the Muséum National d'Histoire Naturelle (Paris), a gift left 
by H. S. MacKee, one of the most important specialists on the flora of New Caledonia, and the Lewis B. & Dorothy Cullman 
F oundation for molecular studies. 

2 The editors of the Annals thank Sopa Balcomb for her editorial contribution to this article. 
* Muséum E d'Histoire Naturelle, Département Systématique el Evolution, USM 602, 57, rue Cuvier, CP 39, 75231 
Paris CEDEX 05, France. lema. 
The Lewis B. and Dorothy Cullman Program for Molecular Systematics Studies, The New York Botanical Garden, Bronx, 
New York 10458-5126, U.S.A. tmotley@nybg.org. 
5 Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A. pete.lowry@mobot.org. 


Ann. Missouri Bor. GARD. 93: 103-121. PUBLISHED ON 31 May 2006. 


Annals of the 
Missouri Botanical Garden 


(189]) 


Baillon’s ideas and restricted the circumscription. of 


known at the time. Schumann rejected 


E 


Guettarda, excluding taxa usually placed in Antirhea 
and Laugieria. While Schumann's concept has been 
followed by most authors since (e.g.. Standley, 1934: 
Bremekamp, 1966: Airy Shaw, 1073; Steyermark, 
1972, 1974: Robbrecht, 1988). 


slve review of the genus is lacking. 


a modern comprehen- 


Guettarda is currently regarded as a large, nearly 


and shrubs. H 


7 


pan-tropical genus of trees 


characterized by flowers with usually imbricate 


corolla lobes, 2- to 9(to 20)-locular fruits with a stony 


endocarp and one seed per locule, and a non- 


persistent calyx. In the Neotropics, Guettarda is 
thought to comprise anywhere from nearly 80 (Taylor 


et al., 2004) to more than 130 species (Andersson, 
1992), and is morphologically quite heterogeneous, 


described or placed in 


including taxa originally 


regarded as distinct, e.g.. 
Matthiola. anc 


Antirhea (or 


several genera once 


a 


Dicrobotryum, Laugieria, Tournefor- 


tiopsis. Furthermore, the inclusion of 
Stenostomum, to which Neotropical Antirhea species 
are currently assigned) can also be inferred based on 
the recent placement of the South American 97 
A. surinamensis (described by Bremekamp. 1959) i 
under Guettarda acreana K. Krause 
Taylor et al. (200 


In the Paleotropics, a 
| 


Sy nonymy 


single, common, coastal 


species (Guettarda speciosa. V), which ranges from 
Africa 
of Guettarda's 


World 


sively 


most 
Old 


exclu- 


o Polynesia, accounts for 


eastern 


broad distribution. The other 
all) are almost 


New 


diverse 


species (ca. 25 in 


narrow endemics on Caledonia and 


a morphologically These 


Guettarda by Baillon (1879) and 


comprise group. 


were assigned 


Guillaumin (1948) in accordance with their broad 
concept of the genus. although they differ from its 


other members by several important. characters. in- 
cluding a dioecious sexual system. and persistent 


calyx. As currently circumscribed. the genus Guel- 


tarda thus has two main centers of diversity, the 


Neotropics and New Caledonia, showing an unusual 
trans-Pacific distribution shared with only a few other 
genera of Rubiaceae, such as Augusta Pohl sensu 
Kirkbride (1997) (including Lindenia Benth.) 
brecht, 1988). 


Recent molecular phylogenetic studies have pro- 


(Rob- 


vided little evidence regarding the monophyly of 


Guettarda. Molecular sequence data for the genus are 
scarce, and, apart from the inclusion of a few species 
family-level phylogenies, the only taxa studied to 


date are those used by Rova (1999) and Rova et al. 


m 


(2002). who examined rps16 and trnl.-F sequences of 


a limited number of exemplar taxa. However, due 1 


insufficient levels of sequence variation, trees based 


on these markers place the six Guettarda species they 
sampled in a polvtomy along with several allied 
genera. viz. Chomelia, Malanea, Neolaugeria. Stenos- 


tomum (referred to as Antirhea) and Timonius. 


Considering the wide range of interpretations (often 
contradictory) of relationships based on the morpho- 
logical. biogeographical, and molecular data, it seems 
opportune to examine more closely the phylogenetic 


status of Guettarda, as currently defined. and its 


affinities with its apparent close allies. 
Guettarda and 1 other genera currently included 


in Guettardeae (Table l) are generally grouped 


together based on a set of common characters 


including: a woody habit. entire interpetiolar stipules 
in Neoblakea), 


(divided axillary inflorescences. (ter- 


minal in Dichilanthe, Machaonia. and Neoblakea). 
a single, apically attached, pendulous ovule in each 
cell of the ovary, fleshy fruits with 2 to many 


unilocular plurilocular endocarp or 


"s 


pyrenes or 


putamen (dry capsular fruits in Machaonia). and 
elongated seeds with poorly-developed endosperm 
(Robbrecht, 1988). the tribe 


contains perhaps 12 to 15 genera and more than 500 


Using this definition, 


species, ranging in distribution from coastal East 


Africa 


unpublished data). 


o Polynesia and the Neotropics (Achille. 
The 


(2002) using sequence data from the chloroplast H 


analysis of Rova et al. 


intron and traL-F spacer suggested a much broader 


circumscription of Guettardeae (Table 1) that includes 


several genera formerly placed in two other tribes. 


Isertae and Rondeletiae. However, the expanded 


notion of Guettardeae proposed by Rova et al. 


because the additional 


2002) seems problematic 


taxa they included. have morphologies strikingly 


different from those of Guettardeae s. str.. as reflected 
that 
include 


by the fact these elements (whose distinctive 


features strictly terminal inflorescences. 


mostly pluriovulate locules and capsular fruits. or 
sometimes sub-capsular fruits in Gonzalagunia) were 
placed by Robbrecht (1988) in subfamily Cinchonoi- 


inti- 


deae, whereas Guettardeae were assigned to 
rheoideae. 
Within 


elements 


Robbrecht), several 


related to 


Guettardeae (sensu 


appear to be most closely 


Guettarda (as currently circumscribed). together 


comprising core Guettardeae (as defined in Table 
viz. Antirhea (including Guettardella), Bobea, Chome- 


lia, Guettarda, Malanea, Neolaugeria, Ottoschmidtia. 


Pittoniotis, Stenostomum, and Timonius. These 10 
genera form a morphologically homogeneous group 
whose recognition is consistent with available molec- 
ular data. In the phylogenies from Rova (1999) and 
Rova et al. (2002). the core Guettardeae they sampled 
(except Bobea) comprise a strongly supported mono- 


phyletic group. while the remainder of the tribe (sensu 


Volume 93, Number 1 
2006 


Table enera included in Guettardeae by Robbrecht (1988) and in the * 


Achille et al. 
Polyphyly in Guettarda 


105 


—— 


Gueltardeae” s.l 


— 


. clade in the phylogenies o 


Rova me bon Rova et al. (2002); those retained in core Guettardeae (as defined here) are indicated in the right column. 


excl. = excluded: n.s. = genus not studied. 
Guettardeae sensu Guettardeae clade of Rova Genera in core Guettardeae 
Genus Robbrecht (1988) (1999) & Rova et al. (2002) las defined here) 
Allenanthus Standl. excl, Allenanthus 
Antirhea Comm. ex Juss.' Antirhea n.s. Antirhea 
Arachnothryx Planch. excl. Arachnothryx 
Bobea Gaudich Bobea excl. Bobea 
Chomelia Jacq. Chomelia Chomelia? Chomelia 
Cuatrecasastodendron Standl. & excl. Cuatrecasasiodendron 
Steyerm. 
Dichilanthe Thwaites Dichilanthe n.s. 
Gonzalagunia Ruiz & Pav. excl. Gonzalagunia 
Guettarda . Guettarda Guettarda” Guettarda 
Guettardella Champ. ex Benth. Guettardella n.s. within Antirhea 
Javorkaea Borhidi & Komlodi excl. Javorkaea 
Machaonia Bonpl. Machaonia Machaonia 
Malanea Aubl Malanea Malanea? Malanea 
Veoblakea Standl. Veoblakea Veoblakea 
Veolaugeria Nicolson Veolaugeria n.s. Veolaugeria 
Ottoschmidtia Urb. Ottoschmidtia n.s. Ottoschmidtia(n.s.) 
Rogiera Planch. excl, Rogtera 
Pittoniotis Griseb. P/ Hh n.s. Pittoniotis 


Stenostomum C. F. Gaertn. within Antirhea 


Timonius DC Timonius 


! Sensu Chaw & Darwin (1992) 


E Ge nera assigned to the core Guettardeae clade | 
Ex 


authors actually sample o in one species. 
(Borhidi & Fernandez, 1993-19 


Robbrecht, 1988). 


sampled), are scattered among a paraphyletic group. 


except Dichilanthe (which was not 


Furthermore. core Guettardeae, as circumscribed 


here, which correspond to Baillon’s (1879, 1880) 


broadly defined Guettarda, display none of the 
distinctive features (mentioned above) that character- 
ize Dichilanthe, Neoblakea, 
newly included in the tribe by Rova (1999) and Rova 
et al. (2002). 
Although 


Guettardeae, Rova (1999) suggests that it would have 


Machaonia, or the genera 


Bobea has always been included in 
to be excluded in order to render the circumscription 
of the tribe (s.l.) monophyletic. Morphology, however, 
provides very strong support for the inclusion of Bobea 
within the core group of Guettardeae. Bobea is hardly 
s placed in 
and Baillon (1880) 


Guettarda s.l. More 


distinguishable from Timonius, where it wa 
Candolle (1930). 


within 


synonymy by 
both 


Darwin and Chaw (1990) regarded Bobea as 


included genera 
recently, 
distinct, but closely related to Timonius on the basis of 
morphology. In the field, individuals of Bobea are 
immediately recognizable as belonging to Guettardeae 
on the basis of fruit structure and overall morphology 
Notwithstanding Rova’s (1999) sugges- 


(pers. obs.). 


Antirhea?” Stenostomum 


Timonius” Timonius 


w Rova (1999) and/or van et al. (2002 
Intirhea acutata (DC.) Urb.. 


which is cute placed in Stenostomum 


tion, we have thus provisionally chosen to regard 


Bobea as a member of core Guettardeae. 

Circumscriptions of the genera comprising core 
Guettardeae have varied widely from one treatment to 
another. Two genera, Antirhea and Timonius, in 
addition to Guettarda itself, play a central role in 
the group, as most species have at one time or another 


The 


Neotropical genus Chomelia (including Anisomeris C. 


— 


seen associated with at least one of them. 


Presl) comprises ca. 50 species and may be associated 
2004). 


difficult to distinguish from one another on the basis 


with Guettarda (Taylor et al., as the two are 
of unequivocal characters. For example, Schumann 
1891), Standley (1934), (1974 


separated Chomelia from Guettarda on the basis of 


and Steyermark 


— 


corolla aestivation (valvate vs. imbricate, respective- 
ly), but Burger and. Taylor (1993) recognized both 
irnL-F 
sequence data, although limited, suggest that the two 
J., 2002), 
Chomelia sampled (C. tenuifolia Benth.) clusters with 


aestivation types Chomelia. Furthermore, 


genera are related (Rova et à as the only 


one of the five Guettarda species (G. crispiflora Vahl). 
Guettardella), 


Piitoniotis 


Antirhea (including Neolaugeria, 


Stenostomum, and appear to comprise 


106 


a complex distributed in the Neotropics and the Indo- 
lo 


2 to 12 united pyrenes, ai 


— 


1 


Pacific region, characterized by mostly 
imbricate corolla lobes, 
a persistent calyx. All the Paleotropical species in this 
group appear to be dioecious, a fact that was not 


1992). The 


Neotropical genus Pittoniotis, described 


discovered until recently (Chaw & Darwin, 
monolypic, 
by Grisebach (1858), is readily distinguished by its 
long-exserted stamens, perfect flowers, and paniculate 
Hemsley (1881) transferred 


inflorescences. However, 


Pittoniotis to Antirhea, a decision that has been 


followed by some authors (Schumann, 1891: Dwyer. 
1980). Hooker (1873) merged Stenostomum, a Carib- 
and Guettardella, 


genus, into Antirhea. Most subsequent authors have 


bean genus, a small South Asian 


regarded Stenostomum and Antirhea as congeneric 
(e.g.. Schumann, 1891: Standley, 1934; Bremekamp. 
1966; Airy Shaw, 1973; Steyermark, 1974; Robbrecht, 
1988). According to this broad generic circumscrip- 


Antirhea, 


— 


tion. like Guettarda, exhibits an unusual 
trans-Pacific distribution. 

Recently, Jansen (198 
species from the Mascarene Islands and Madagascar, 
other Old World 


delimitations were based 


) restricted Antirhea to two 


and assigned all the 
Guettardella. 


largely on the number of flowers per inflorescence, 


species L 


His generic 


number of ovary cells, geographie distribution, and 
also an erroneous interpretation. of floral sexuality. 
By contrast, Chaw and Darwin (1992) retained all the 
Old World species 
which they distinguished three subgenera, A. 
A. subg. Guettardella (Champ. ex Benth.) 
Chaw, and A. subg. Mesocarpa Chaw, while tentatively 
that 


in Antirhea (36 species). within 
suby. 
Antirhea, 
Slenostomum distinct 


suggesting represents a 


genus characterized by hermaphroditic, (A to)5- 
merous flowers, ebracteolate inflorescences. and 


a Neotropical distribution. Following their suggestion, 
Borhidi and Fernandez (1993-1094) formally trans- 
ferred the 39 of Antirhea to 


Stenostomum. and 


American species 


reduced Neolaugeria (4 to € 


= 


Caribbean species, distinguished by the united 
stipules and the 4- to 6-locular ovary) to a section of 
Stenostomum. However, Moynihan and Watson (2001) 
rejected the inclusion of Neolaugeria in Stenostomum 
based on phylogenetic evidence from ITS sequence 


variation. of three 


Stenostomum. 
The large genus Timonius is traditionally charac- 


terized by having fruits with separate pyrenes, 


Bobea. 


) species (Darwin, 


although this feature is also encountered 
Timonius comprises more than 180 
1994. 


and from the Philippines to northern Australia. It is 


— 


distributed. from the Seychelles Polynesia 


morphologically heterogeneous, but is classically 


interpreted to include all dioecious Guettardeae with 


species of Neolaugeria and one of 


Annals of the 
Missouri Botanical Garden 


valvate corolla lobes and separate pyrenes. Bobea 
comprises four species endemic to Hawaii (Darwin & 
Chaw, 1990), which differ mostly from Timonius by 
the presence of imbricate corolla lobes. 


Finally, two Neotropical genera, Ottoschmidtia and 


Malanea, stand out because their cire umscription: 


have rarely been questioned. Ottoschmidtia (3 spe- 
cles) is easily separated from other Guettardeae by its 
2-locular ovary and zygomorphic corollas. Malanea, 
with ca. 35 species (Taylor et al., 2004), is recognized 


habit, to. spiciform in- 


by its twining paniculate 


florescence, exserted stamens. and stipules with 


a distinctive morphology, although Taylor et al. 
(2004) regard the distinction between Malanea and 


Chomelia as unclear. 


In an effort to evaluate the phylogenetic status of 
Guettarda, as currently defined, we have endeavored 
to include a representative group of species belonging 
to the genera of core Guellardeae. Our study employs 
DNA sequences from the internal transcribed spacer 
region (ITS, spacer 1 & 2, and 5.85 gene between the 
DNA 785-265 
of Guettarda (as 


nuclear ribosomal genes). Using 


a broad sampling of 21 species 
currently circumscribed) and 21 species belonging to 
e 


we thus (1) evaluate the monophyly of Guettarda: 


10 other genera of Guettardeae (sensu Rova. 


provide some insights regarding the 1 


n relation to Guettarda. 


status of other core genera 


especially those whose cireumse riptions have varied 


widely; and (3) attempt to identify morphological 


characters that might prove valuable for further 
taxonomic and phylogenetic studies within Guettar- 
geographic 
light of 


the results of our phylogenetic analysis, as these two 


deae. We have also examined the 


distributions of Guettarda and Antirhea in 


genera are among the only Rubiaceae to exhibit 


a disjunct, nearly pantropical distribution. 


MATERIALS AND METHODS 


PAXON SAMPLING 


^ 


» ensure the best possible representation of 
infrageneric groups within Guettarda, for which no 
comprehensive infrageneric treatment is available, we 
delimited informal morphological groups on the basis 


P and NY, 


of examination of herbarium material 


field observations made by the first author in New 
Caledonia, and information gathered from "e TM 
ture (in. particular Howard, 1989: Standley. 1934; 


Standley & Williams. 1975). Furthermore, we 
attempted to include species from various phytogeo- 
graphical subdivisions of the Neotropical regions, as 
defined by Andersson (1992). We have also sought to 


sample. as exhaustively as possible, at the generic 


Volume 93, Number 1 
2006 


Achille et al. 107 


Polyphyly in Guettarda 


(and sometimes infrageneric) level within the other 
core genera of Guettardeae. To delimit infrageneric 
we consulted the treatments of Antirhea by 
Chaw and Darwin (1992) and Stenostomum by Borhidi 
and Fernandez (1993-1994). and the partial revisions 
of Timonius by Darwin (1993, 1994). supplemented by 


informal morphological groups that were identified in 


groups, 


— 


the same manner as Guettarda (Valeton, 1909: Wong. 
988). 
Special attention was given to including the type 


species of each genus, and, when possible, a minimum 


of 2 species were sequenced per genus, subgenus, « 


informal group. The 42 taxa used in the analysis 


represent 9 of the 10 core genera of Guettardeae 


(Table 2), including the type species of 8 currently 


recognized genera (Guettarda speciosa, Antirhea 
borbonica. Bobea elatior, Chomelia spinosa, Neolau- 


geria resinosa, Pittoniotis trichantha. Stenostomum 
lucidum. and Timonius timon) and 3 genera currently 
placed in synonymy, namely, Dicrobotryum. Guettar- 
della, and Matthiola (typified by Guettarda divaricata. 
Intirhea chinensis, and Guettarda scabra, respective- 
y). Because of a lack of material, Ottoschmidtia could 
and a single species each was 


not be included, 


sequenced for Chomelia and Malanea. For Antirhea 


and Timonius, we included 4 and 5 species. re- 
spectively, which were selected to represent as much 
morphological diversity as possible. Three previously 
published sequences were used for Neolaugeria 
densiflora, Stenostomum myrtifolium, and Timonius 
nitidus (Moynihan & Watson, 2001). 39 new 


Guettardeae sequences were generated. Two outgroup 


while 


taxa (Cuatrecasasiodendron spectabile and Gonzalagu- 


nia rosea) were selected from Guettardeae sensu 
Rova, since they are both placed near to the genera of 


core Guettardeae in his phylogenies (Rova, 1999). 


DNA EXTRACTION 


Leaf samples were collected either in silica gel or 
from herbarium sheets. Genomic DNA was extracted 


from approximately 1 em? of dried leaf tissue using 


a modified CTAB methodology. Leaf material was 
ground in a lysing matrix “A” tube (Qbiogene, 


Carlsbad, CA) and pulverized for 15 sec. in a Fastprep 


machine FP-120 1 bead mill at speed 


NAA 500 ul of Carlson Lysis Buffer (2 g 
TAB, 8.18 g NaCl, 0.7: 2 g EDTA, 10 ml M Tris/ 


HCI pH 7.0, nanopure water to 100 ml, verified to 
pH 9.5, autoclaved, then 1 g PEG 400 
and 75 ul of B-mercaptoethanol were added to 
each tube and incubated 74°C 
shaking for 60-90 min. Following incubation, 575 ul 
of SEVAG (24 : 


added to each tube: the tubes were placed on a tipping 


JO added when 
cool) 


with occasional 


| chloroform : isoamyl alcohol) were 


board for 30 min. at room temperature, and then 
centrifuged at 14,000 rpm for 1 min., and ~350 ul of 
supernatant was removed and added to new tubes 
containing 1050 ul of Nal solution, 20 ul Glassmilk, 


4 ul TBE modifier (Qbiogene). The tubes were 


and 
placed on a tipping board for 30 min. at room 
temperature. Afterward, the tubes were centrifuged 


at 14.000 rpm for J min. and all of the supernatant 
Next. each Glassmilk pellet was 
washed three times with 800 ul and once with 
150 ul of ice cold New Wash solution (Qbiogene). 
After the all of Wash 
aspirated from the Glassmilk pellet. and 50 ul. of 
10 mM Tris-Cl (pH 8.5) elution buffer were added to 
resuspend the DNA. 
—55 € for —10 min. and then centrifuged for | min. 


at 14.000 rpm. The supernatant containing the DNA 


was discarded. 


final wash, the New was 


tubes were incubated 


The 


was removed and transferred to new tubes and stored 
—20"C. 


a 


DNA AMPLIFICATION 


PCR was performed in 
5 ul 10X buffer with 
Po City, CA), 9.3 ul 
autoclaved nanopure water, 2.5 ul BSA (bovine serum 
albumin), 2.5 ul dNTP, 1 a each of two 20 uM 
primers. 5 Ul betaine, 0.2 ul Taq polymerase (Qiagen, 
Valencia, CA), and 1 ul of genomic DNA. All PCR 
and cycle sequencing reactions were run on a Gene 
\mp PCR system 9600 (Applied Biosystems, Foster 
City. CA). The ITS region was amplified using primers 
NY183 . (5'-CCTTATCATTAAGAGGAAGGAG-3!) 
and NY43 (5'-TATGCTTAAAYTCAGCGGGT-3’), 
and when necessary two additional internal e rs 
were employed, ITS2 (5'-GCTACGTTCTTC: 
GATGC-3') and ITS3b (5'-GCATCGA' ee 
TAGC-3') (White et al., 1990; Baldwin, 1992; 


For DNA amplification, 
a 25 ul mixture consisting of 2 


MeCl, (Perkin Elmer, 


Baum 


et al., 1998). The PCR conditions i ap of 
the ITS region were: | cycle of 97 € for 50 sec.: 30 
cycles of 97 € for 50 sec., 53 € a 50 sec.; aad l 


cycle of 72°C for J min. 50 sec., hold 72°C for 7 min., 


hold at 4 C. 


DNA SEQUENCING 


To detect successfully amplified products, as well 
as the possible presence of multiple ITS paralogues of 
varying length (as discussed by Alvarez € Wendel, 
2003) and possible contamination of negative con- 
trols, PCR products were examined on 1% agarose 
ethidium bromide and visualized 
under light. 
purified with spin columns from the QI Aquick. PCR 


urification kit (Qiagen) following protocols provided 
| : 31 I 


gels stained with 


ultraviolet Amplified products were 


108 Annals of the 


Missouri Botanical Garden 


Table 2. Sp 


information, with collections italicized, is given for new sequences: literature citation is provided for previously published 


ecies, origin, and voucher information for the 42 ITS sequences used in the present. study. Voucher 


sequences of three taxa. Herbarium acronyms follow Index Herbariorum (Holmgren et al.. 1990). Partial sequences are 


indicated in parentheses following GenBank accession number. 


Voucher information / 


Species 


Origin 


literature citation 


GenBank accession 


Antirhea borbonica J. F. Gmel Madagascar Rakotomalaza 1248 (P) DQ003000 

Intirhea chinensis (Champ. ex Benth.) China Tam s.n., Hong Kong (P) DQ063702 
F. B. Forbes & Hemsl. 

Intirhea rhamnoides (Baill) Chaw New Caledonia Achille 688 (P) 10063086 


Intirhea smithii (Fosberg) Merr. 
L. M. Perry 
Bobea elatior Gaudich 
Bobea sandwicensis (A. Gray) Hillebr. 


Chomelia spinosa Jacq. 


Cuatrecasasiodendron spectabile Steyerm. 

Gonzalagunia rosea Standl. 

Guettarda acreana K. os 

Guettarda combsit Urb. 

Guettarda crispiflora Vahl 

n divaricata (Humb. & Bonpl.) 
ndl. 

e elliptica Sw. 

Guettarda heterosepala Guillaumin 


Guettarda hirsuta (Ruiz & Pav.) Pers. 


Fiji 


Hawaii 
Hawaii 


Costa Rica 


Colombia 
Costa Rica 
French Guiana 
Belize 
Colombia 


Guyana 


Puerto Rico 
New Caledonia 
Ecuador 


. C. Smith 7085 (NY 


Motley 2536 (NY) 
Takeuchi 3302 (NY) 
Seigler 12306 (NY) 


Rova et al. 2095 (GB) 
Achille san., Cult. NYI 
Mori 18842 (NY) 
Lentz et al. 2653 (NY) 
Andersson et al. 2088 
Jansen 4908 (NY) 


„ 3000 n ) 
lehille € 
Eres s m ) 


) 


3G (P) 


(GB) 


DQ063667(I T8 1) 


110063608 
110063069 
5000370301781!) 
DQ063704(1T52) 
100063670 
50003071 
110063672 
DQ063073 
50003074 
50003075 


50003070 
5000307 
DQ003078(1 TS 1) 


Guettarda humboldtensis Guillaumin New Caledonia Achille 672 (P) DQ063070 
Guettarda krugii Urb. Puerto Rico Struwe 1102 (NY) 50003080 
Guettarda macrosperma Donn. Sm. Costa Rica Hammel 19883 (NY) 1006308 

Guettarda nannocarpa Urb. & Ekman West Indies Liogier 16631 (NY) 10063082 
Guettarda noumeana Baill. New Caledonia lehille 662 (P DQ063083 


Guettarda pohliana Mill. Arg. 


Guettarda pungens Urb. 


Bolivia 


Puerto Rico 


Vee 10556 (NY) 
Grimes 3261 (NY) 


950003084 


Guettarda scabra (L.) Lam. Puerto Rico Acevedo Rodriguez 2840 (NY) 5003087 
Guettarda sp. “Cuba” Cuba lreces 0538 ) 100603084 
Guettarda speciosa L. New Caledonia lehille 661 (P) 110063689 
Guettarda splendens Baill. New Caledonta lehille 1010 (P) DQ063690 


Guettarda stenophylla Urb. 


Guettarda trimera Guillaumin 
Guettarda uruquensis Cham. & Schltdl. 
Malanea macrophylla Baril. ex Griseb. 


Veolaugeria densiflora (Griseb.) Nicolson 


Neolaugeria resinosa (Vahl) Nicolson 
Pittontous trichantha Griseb. 
Stenostomum acutatum DC. 
Stenostomum lucidum (Sw.) C. F. Gaertn. 
Stenostomum myrtifolium Griseb. 
Timonius densiflorus Valeton 
Timonius flavescens Baker 


Timonius nitidus Vern.-Vill. 


Timonius polygamus (Forst. f.) Robinson 
pot v; 


Timonius timon (Spreng.) Merr. 


West Indies 


New Caledonia 
Brazil 
Guyana 


Bahamas 


Puerto Rico 
Panama 


Puerto Rico 


Dominican Republic 


Bahamas 


Papua New Guinea 
Malaysia 


Guam 


French Polynesia 


Papua New Guinea 


Liogier 10950 (NY) 


lehille 870 (P) 
12227 (NY) 


Jansen= Jacobs 5396 (NY) 
Moynihan & Watson (2001) 


Taylor & Gereau 1050 
Croat 9311 (NY 
Ixelrod 3288 (NY) 


Acevedo Rodriguez 8408 


1 (NY) 


(NY) 


Moynihan & Watson (2001) 


Takeuchi 8089 (NY) 


Moynihan & Watson (2001) 


Motley 2044 (NY) 
Motley 2309 (NY) 


DQO063705( TS 1) 

DQO63 7060 T82) 
DQ06369 | 
DQ063692 
100063693 
VE323055(IT8S 1) 

A F323056(ITS2) 
110003694. 
DQO63695(IT81) 
DQ063696 
— 50003097 
AF323059( ITS 1) 

\F323060(ITS2) 
110063608 
50003099 
VE323003(0T8 IH) 

AF323064(T82) 
DQO063700 
DQ063701 


Volume 93, Number 1 
2006 


Achille et al. 109 


Polyphyly in Guettarda 


by the manufacturer. Purified products were cycle 


sequenced with dye terminator ABI Prism? Ready 
reaction mix (Applied Biosystems) using dRhodamine 
596 dimethyl 
sulfoxide. Primers for cycle sequencing were the same 
used in the PCRs. 
conditions were: | cycle of 95 C 
96 C for 10 sec., 50°C for 5 sec 
hold at 4 C. 
hydrated Sephadex 6-50. DNA Grade F columns 
NJ) and 
dehydrated : a speed vac. The DNA was resuspended 


in 2.2 ul « 


or Big Dye v1.0 (1/8 reaction) and 
Cycle sequencing 
for 1 min.: 32 cycles 
. 60 C for 


Products were purified using 


as those 


3 min:: 


and 
Piscataway. 


(Amersham Pharmacia Biotech, 


` formamide (83.5%) and EDTA blue- 


dextran Mun dye (16.5%), heated at 95°C for 
2 min., and immediately placed on ice. Sequences 


were run on a polyacrylamide gel on an ABI Prism 
377 DNA sequencer (Applied Biosystems). Gels 
consisted of 18 g urea ultrapure grade (Amresco. 
Solon, OH). 0.5 g Amberlite mixed bed resin MB-150 


(Sigma-Aldrich, Mi WI). 27 ml 


water, and 5.6 ml Acryl/Bis (Amresco). 


waukee, nanopure 


PHYLOGENETIC ANALYSIS 


Sequences were edited and aligned in Sequencher 
Ann Arbor. MI) 


followed by manual optimization. Gaps were inserted 


version 3.1.2 (Gene Codes Corp., 
to keep the number of informative characters (indels 
and substitutions) to a minimum. Indels were treated 
as unordered, multistate characters; those of. equal 
length that occurred in more than one sequence were 
considered homologous and scored as separate binary 
characters added to the data matrix. Insertions that 
had slightly different sequences in two or more taxa 
1994), and the 


variable bases within these sequences were then re- 


were treated as homologous (Barriel, 


coded as new characters (and treated as missing for 
the taxa in which the insertions were absent). The data 
matrix is available from the first author. 


A heuristic parsimony search was performed using 


PAUP* version 4.01b (Swofford, 1998), with 100 
random addition replicates, TBR branch swapping. 
and MULPARS option in effect. Relative levels 


sel were 
(CI), the 


scaled consistency 


homoplasy and synapomorphy in the data 


calculated using the consistency index 


and the 


retention index (RI). 
index (RC). 
branch support was estimated with bootstrap analysis 
TBR 


branch swapping and simple taxon addition, Pairwise 


as implemented in PAUP*. Internal 


(Felsenstein, 1985) using 1000 replicates with 
sequence divergence between species of core Guet- 
calculated with PAUP* (“p 


excluding ambiguously aligned sites 


lardeae was ” distance, 


(Bl) was 


dis uned with the program MrBaves 3.0b4 (Huelsen- 


Bayesian inference of phylogeny 


beck € Ronquist, 2001) using the GTR+I+T model of 


nucleotide substitution, a general time reversible 
model that allows (1) a proportion of invariable sites 


and (2) 


a gamma distribution. 


among-site rate heterogeneity following 
[e] o y o 


This model was selected as 
the best fit model for our data from a comparison of 
24 models using the Akaike information criterion 
(Akaike, 1974) as implemented in MrModeltest 2.2 
(Nylander, 2004), a modified version of Modeltest 3.6 
1998). 
data set (excluding indels) was run using 2 X 10° 
Markov 


and 


(Posada & Crandall, The analysis of the ITS 


generations, which were performed on four 


chains (the temperature of the chains other 
parameters were left at default values) with trees 
sampled every 100 generations. Markov chain con- 
vergence occurred rapidly (before 2 X 10 genera- 


and 2000 trees (2 X 10? 
s the 


generations) were 
The 


then used to construct a 50% 


Lions). 


"burn-in" of the 


£5 


discarded ¢ process. 
remaining trees were 
majority rule consensus tree, which provides estima- 
tions of each clade’s posterior probability. As a control 
(Huelsenbeck et al., 2002), four independent searches 
All saved 
trees from the independent runs (“burn-ins” excluded) 
Unlike 


posterior probabilities are 


were run, which yielded similar results. 
were pooled to obtain a consensus tree. 
bootstrap support. values, 
interpreted as true probabilities for each clade under 
the assumed model (Rannala & Yang. 1996). 
RESULTS 


Direct sequencing of the purified PCR products 
yielded single bands on agarose gels and unambigu- 
ous ITS sequences (without obvious multiple bases), 
showing no evidence of paralogs of varying length. 
The length of the sequences (including ITS] + 5.85 + 
ITS2) ranged from 


stenophylla) to 609 bp (in Gonzalagunia rosea). The 


590 base pairs (bp) (in Guettarda 


aligned matrix comprises 080 positions (655 of which 
represent the aligned sequences and 25 the coding of 
A total of 84 


sites were excluded from the analysis because their 


the indels, including autapomorphies). 


alignment was uncertain (most of the indels present 
among these 84 positions were, however, retained in 
their recoded form as indicated above). Of these 596 
analysis, 421 


uninformative, 


characters included in the 


constant, 76 variable but 

potentially informative. For two species (Guettarda 

hirsuta and Pittoniotis trichantha). ITS? sequences 
1 


could not be obtained. Chomelia spinosa and G. 


stenophylla could not be sequenced for a region in the 
5.85 cistron (17 and 19 bp long, respectively), and the 
end of ITS2 could not be sequenced for Timonius 
flavescens (46 missing bp). In these cases, the bases 


were treated as missing data. 


110 


Annals of the 
Missouri Botanical Garden 


The 5.85 cistron is 165 bp in length for all taxa. 
except in the sequence of Stenostomum myrtifolium 
provided by Moynihan and Watson (2001), which is 
2 bp shorter. As often observed (e.g. by Andreasen el 
al.. 1999: Moynihan & Watson, 2001), few mutations 
The ITS I and 2 


the 40 species of core Guettardeae included in our 


occur within this region. regions in 
sample, range in length from 213 to 223 and 207 1 

224 bases, respectively. These lengths fall within 5 
range reported for other angiosperms (Baldwin et al. 


1995) and for other groups of Rubiaceae (McDowell & 


Bremer, 1998: Andreasen et al., 1999; Persson, 2000: 
buco qus qs ina & Bremer. 2002). Percent. pair- 
wise distances between core Guettardeae species 
range from 0 to 14.2 for ITS 1 (mean = 7.6 + 2.4), 
0 to 13.0 for ITS 2 (mean = 5.3 + 2.1), and O to 9.8 
(mean = 5.0 — 1.5) for the whole ITS region. This 
amount of [ITS variation is comparable to levels 


usually reported for a single genus (Baldwin et al.. 


1995). such as Viburnum L. (Donoghue & Baldwin. 
1993; reviewed in Baldwin et al., 1995) or infra- 
generic taxa such as Gilia sect. Giliandra A. Gray 


reviewed in 


(Porter, 19903; 


Compared lo 


Baldwin et al.. 1995). 
other groups of Rubiaceae, variation 
within core Guettardeae (ca. 500 species) is similar to 
(or in the case of ITS1, lower than) that reported by 
McDowell and (1998) for Exostema ers.) 
Rich. ex Humb. & Bonpl. (25 species). and similar to 
the level observed by Nepokroeff et al. (1999) within 


the main lineages of tribe Psychotriae, but much lower 


Bremer 


than for the tribe as a whole (1000 to 1650 species). 

The MP analysis yielded 1293 most parsimonious 
CI = 0.577, R 
= 0.401). One of the shortest trees is given 
2) is well 


although numerous clades are not signifi- 


trees, each with a length of 445 = 
0.695, RE 


— 


Figure The strict consensus. (Fig. 
resolved, 
cantly supported in the bootstrap (BS) analysis (BS < 
50%). 


99%) and the topology at the base of the ingroup is 


The ingroup node has strong support (BS = 
fully resolved but poorly supported. in contrast to the 
Eight | 


divaricata—G. elliptica, G. rd stenophylla, 


terminal branches. airs of species (Guettarda 


G. hrugti-G. sp. “Cuba.” Stenostomum acutatum—S. 
myrtifolium. Bobea elatior-B. sandwicensis, G. crispi- 


lord. 


Neolaugerta 
¿ 


hirsuta, Timonius nitidus-T. polygamus. 


densiflora—N. 


bootstrap support, ranging from 94 to 1006 


high 
. Pitto- 


the 1 0 poe the 


resinosa) e 


niotis is sister to the rest of 


remainder of the genera comprise two main clades. 


The first of these clades contains most species of 


1e Neotropical genera, and 
Bobea. Within 


clade, two Guettarda species appear isolated as sisters 


Guettarda, the majority of 


the Hawaiian endemic this 


genus 


of Bobea (comprising the “Bobea clade”). The second 


major clade comprises all the Paleotropical genera. 


some Guettarda species (mostly from New Caledonia), 
and the West Indian taxa Neolaugeria and Stenosto- 
mum sect. Stenostomum. Within this second main 
clade, a weakly-supported Paleotropical dioecious 
subclade (BS = 63%) unites all the Paleotropical 
species included in the analysis except one (Guettarda 
speciosa), while the Neotropical subclades (the first 
second with Stenos- 


comprising Neolaugeria and the 


lomum secl. Stenostomum + Guettarda acreana) are 
paraphyletic at the base of the clade (no BS support). 

Guettarda appears to be polyphyletic in our trees. A 
majority of the Neotropical taxa sampled are associ- 
ated with the Paleotropical type species, G. speciosa. 
and thus comprise the “Guettarda s. str.” clade (BS = 
57%). The three remaining species that occur in the 
New World are 
Stenostomum lucidum. while the other two are part 
All the other Guettarda 
Old World and are 


included within a Paleotropical Dioecious Clade. The 


isolated (one is associated with 
of the Bobea Clade) (Fig. 


species sampled occur in the 


complexes of genera usually associated with the three 
principal members of core Guettardeae (Guettarda. 
Intirhea. and Timonius) are not evident in our trees. 
Our 


Bobea are monophyletic and, although often consid- 


limited sampling suggests. that Timonius and 


ered as close relatives, are placed in very distant 


clades in our trees. Antirhea + Stenostomum (= 
Antirhea s.l.) is polyphyletic, with Antirhea species 


forming a para hyle Le group with respect to Timonius 
| | 
New 


Paleotropical Dioecious Clade, while the members of 


and the Caledonian Guettarda species in the 


Stenostomum are represented in two additional clades, 


suggesting it, too, may be polyphyletic. 


The BI tree is given in the right half of Figure 2. Its 
overall topology is quite similar to that of the MP tree 
and the two analyses essentially recover the same 
major clades. More clades are well supported in the BI 
analysis (with posterior probabilities exceeding 0.95) 
than in the MP analysis (some of which have BS 
50%), 
supported in the MP analysis are likewise in the Bl 


New 


species is 


values below but those clades that are well 


tree. For example. the clade comprising 


Caledonian Guettarda and 3 Antirhea 


strongly analysis 
0.99). but unsupported in the MP 


50%), whereas the clade comprising 


supported. in the B 
P = 


analysis (BS < 


(posterior 


probability. 


G. divaricata and three other Neotropical Guettarda 
species has strong support in both MP and BI analyses 
(BS = 97%, P = 1.00). 


resolved than the MP consensus tree. 


The BI tree is slightly less 
Veolaugeria 
appears sister to the rest of the ingroup. which forms 

polytomy of three elades. The first of these clades 
encompasses most of the Neotropical genera, in- 
cluding Guettarda s. str., Pittoniotis and Malanea (P 


= 0.57); the second comprises the Bobea Clade (P = 


Volume 93, Number 1 Achille et al. 111 
2006 Polyphyly in Guettarda 


Gonzalagunia 
— — Pittoniotis trichantha | 
Malanea macrophylla 
Guettarda macrosperma 
Guettarda divaricata 
Ó ri allintira 
; O 
Guettarda pohliana c 
D 
Guettarda uruquensis S 
Guettarda nannocarpa 8 
58 rs + A fannnhyvilla [02] 
e D 
b 2 
E 
Guettarda pungens o 
HER phus. 
— 99 Guettarda scabra a 
AAA 3 
Guettarda speciosa | 
Guettarda krugii 
Guettarda sp. Cuba 
94 Stenostomum acutatum 
52 Stenostomum myrtifolium 
| (^h. I f 
| p 
100 | Bobea elatior D 
: > 
Bobea sandwicensis D 
100 e Harda n ispiflora O 
, Ry 
L— Guettarda hirsuta Q 
Antirhea rhamnoides 
- Antirhea smithii 
EN Antirhea chinensis "CU 
w 
—— Guettarda splendens o 
Sa | —— Guettarda heterosepala 3 
Guettarda noumeana 2 
e 
63 B Guettarda humboldtensis — 
66 Oo 
Guettarda trimera Q 
Til us densifloru 5 
63 Timonius flavescens 5 
Timonius nitidus 2 
Timonius polygamus - od 
— | Timonius timon ] 
———À Antirhea borbonica | 
52 Guettarda acreana 
ma OE Stenostomum lucidum | 
Al J ; ry, ifl 
100 | J 
— Neolaugeria resinosa | 
5 changes 
Figure l. Phylogram of one of the 1293 most parsimonious trees (length = 445; CI = 0.577; RI = 0.695) resulting from 


the MP mE of ITS data. Numbers on internodes indicate bootstrap support (in 96). Guettarda species are in bold. Type 
species of currently recognized genera are placed in a box; type species of genera currently placed in synonymy are underlined 
(see text for details). 


112 Annals of the 
Missouri Botanical Garden 
Cuatrecasasiodendron 
Gonzalagunia 
Pittoniotis trichantha — 085 
Malanea macrophylla unl 
8 Guettarda macrosperma 
2 75 100 [ Guettarda divaricata 
2 i in Guettarda elliptica 
E T Guettarda pohliana 
- Guettarda uruquensis 
D 
S ES 100 > Guettarda nannocarpa 
D a Guettarda stenophylla 
9 Guettarda combsii 
73 
m Guettarda pungens 
57 Guettarda scabra 
99 - Guettarda speciosa 
ma Guettarda krugii 
Guettarda sp. Cuba 
Lm Stenostomum acutatum assem d 
52 Stenostomum myrtifolium — 
DL m Chomelia spinosa 
Bobea — 
Clade Bobea elatior 
= Bobea sandwicensis ; 
100 — Guettarda crispiflora — i 0.63 
Guettarda hirsuta 
—— BÉ Antirhea chinensis 
(0) Antirhea smithii 
3 T | 
5 | - Antirhea rhamnoides = 
— 3 Guettarda heterosepala Duo 
8 Guettarda splendens === 1 
o | [ES] — 
8 As if Guettarda humboldtensis ———— 
3 63 53 | a Guettarda trimera EN 
-= | | | 0.84 
a — Guettarda noumeana po 
o6 i us + ii ri. ¡fl ES 
E: | | 
E E Es Timonius flavescens 
m Timonius nitidus — : UL 
: Timonius timon pao 
Antirhea borbonica 
se[ . . Guettarda acreana 0.89 
Stenostomum lucidum = 
100 — Neolaugeria densiflora * 
iin Neolaugeria resinosa ml 
MP tree Bl tree 
Figur Comparison of the consensus trees from the MP and BI analyses of ITS data. The tree on the left is the strict 


B 
consens 


" 
us of 1293 most parsimonious trees (numbers on internodes indicate bootstrap values over 50%). The tree on the right 


is the 50% majority rule Bayesian consensus tree under the GTRH+P model of substitution from four 2 X 10° generations 


runs (numbers on internodes indicate posterior probabilities). Guettarda species are in bold. Clades discussed in the text are 


indicated by double lines. Geographic distribution of terminal taxa is indicated as follows: gray = paleotropical: dark gray = 


Hawaii endemic; light gray = endemic to New Caledonia: all others are Neotropical. 


Volume 93, Number 1 
2006 


Achille et al. 113 


Polyphyly in Guettarda 


0.92); and the third clade (P = 0.80), which is also 
recovered in the MP analysis, includes the type of 


as 


Stenostomum (S. lucidum), Guettarda acreana, and the 
well-supported Paleotropical Dioecious Clade (P 

0.99). Most of the smaller clades recovered in the BI 
tree are also found in the MP tree, but the position of 


Pittoniotis is inconsistent between the two analyses. 


Discussion 


The poor support of the larger clades identified in 
the MP analysis (most with BS < 50% and none 


exceeding 63%) is unusual as IT'S data are generally 


— 
=) 


most informative at suprageneric levels (Baldwin et 
al.. 1995). 
(Lantz et al.. 
(19 genera, ca. 600 species) that was examined using 


For example, in a study of Vanguerieae 


2002). a similar-sized rubiaceous tribe 


a comparable data set of 41 ingroup taxa, the major 
clades are both well-resolved and strongly supported. 
The situation found in Guettardeae, with low support 
and/or resolution among the internal nodes, has also 
been observed in several other groups of Rubiaceae. 
1. ex DC. (Persson. 2000) and 
the Catesbaeeae-Chiococceae complex (Motley et al.. 
2005). as well as among groups of 
(Baldwin & Sanderson, 1998; 
2001). Fabaceae (Allan & Porter, 2000), and Apiales 
(Plunkett € Lowry, 2001; Chandler & Plunkett, 2004: 
Plunkett et al., 2004). This 


consistent with a hypothesized rapid radiation early 


such as Alibertia A. Ric 


Asteraceae 


^ 


Garcia-Jacas et al.. 


pattern would be 
in a group's evolution, and could explain the lack of 
support at the base of our trees, which stands 

contrast to the morphological diversity that has served 


to segregate the core genera within the tribe. 


POLYPHYLY OF GUETTARDA 
ITS 


circumscribed, is 


Based on our analysis of sequence data, 


Guettarda, as currently clearly 


with distributed among four 
(Fig. I). Most of 
species belong to the first of these. the Guettarda s. 
str. clade (BS = 58%, P = 0.91), which includes the 


type species, G. speciosa, from the Indo-Pacific region. 


polyphyletic. species 


distinct clades the Neotropical 


— 


The topology within this clade is well resolved (Fig. 2 
even though the species are morphologically quite 
similar to one another. Members of this group are 
characterized by usually globose fruits with more than 
2 locules, dichasial cymose inflorescences (sometimes 
reduced to a solitary flower), and a caducous calyx 
with a circular abscission scar (Fig. 3). These last two 
characters, although highly distinctive of Guettarda s. 
str., also occur (although not in combination) in three 
New Caledonian Guettarda species, where they may 


have evolved independently from the core Neotropical 


lineage: G. noumeana and an unnamed close ally (not 
sampled) have a dichasial cymose inflorescence 
structure, and a recently discovered undescribed 


species (not sampled) has a dehiscent calyx with 
circular abscission. 

Within the Guettarda s. str. clade, there are three 
main subclades. The first comprises species from 
Continental America (BS = 75%, P = 0.94), a few of 
which, such as G. elliptica, extend slightly into the 
Caribbean Islands. Guettarda divaricata, the m 
species of Dicrobotryum Willd. ex Roem. & Schult., 
Members of 


characterized by small- to medium-sized, membrana- 


part of this clade. this group are 


ceous leaves, and possibly also by their pollen (Fig. 2). 


— 


which is similar to that found in Antirhea s. str 


(oblate-spheroidal 3-colporate pollen with reticulate 
(Achille. The 


includes Caribbean species (BS = 


exine) unpublished data). second 


subclade two 


100%, P = 1.00), G. nannocarpa and G. stenophylla, 


with very small membranaceous leaves, solitary 


The third subclade, while 
weakly supported (BS = 57%, P = 0.78), unites 


flowers, and small fruits. 


8 


group of morphologically distinct species mainly 
from the West Indies along with the Paleotropical G. 
speciosa, the type species. Guettarda scabra, the type 
of Matthiola (widely regarded as a generic synonym 
since Lamarck, 1792), also belongs to this group and 


closely resembles G. speciosa. The fruits of taxa in this 


third subclade are rather large, as are the leaves. 
which are often coriaceous or bullate and have 


scalariform venation. Guettarda pungens stands out 


spiny leaves, and 
third 


wilhin this group in having small. 


small fruits. However, all members of the 
subclade share a distinct pollen type, characterized 
by an oblate-spheroidal shape. 3 to 4 pororate 
apertures, a perforated to reticulate tectum, and short 
columellae (Achille, unpublished data). 

The inclusion of the widely distributed Indo-Pacific 
species Guettarda speciosa in an otherwise strictly 
Neotropical clade is perhaps best explained. by 
relatively recent long-distance dispersal. Guettarda 
speciosa occurs in strand vegetation and coastal forests 
on coral limestone, and its fruits have air-filled 
cavities, making them well adapted to transport by 
ocean currents. This species is not found in Hawaii or 
the eastern. Pacific, although this may simply reflect 
the random nature of dispersal. It is perhaps more 
surprising to note that G. speciosa appears to be most 


closely related to species from the West Indies (such 


[om 


as the widespread coastal species G. scabra mentione 
above) rather than from the Pacific coast of the 
Americas, as one might expect. 

The second clade containing members of Guettarda 
includes two Neotropical species (G. crispiflora and C. 
hirsuta) that are sister in our analysis to Bobea (BS < 


114 Annals of the 
Missouri Botanical Garden 


Ñ Ss » ds ve 
Ors 


se SV 
DL Cr| Cp 7 H Cuatrecasasiodendron E 
Pi i Cr Cp [Pn| H.. Gonzalagunia E 
: Du Cp |Pn| H Pittoniotis trichantha 
Du Cp [Pn] H  Malanea macrophylla 
Du H Guettarda macrosperma 
Du H Guettarda divaricata 
Du H Guettarda elliptica 
Du H Guettarda pohliana 9 
2 

Du H Guettarda uruquensis T 
Du H Guettarda nannocarpa 8 
Du H Guettarda stenophylla 1 
Du H Guettarda combsii 2 
Du H Guettarda pungens o 
Du H Guettarda scabra a 
Du H Guettarda speciosa ? 
Du H Guettarda krugii 
Du H Guettarda sp. Cuba 
Du Cp Ci H  Stenostomum acutatum 

¿Pi; Du Cp C1 H  Stenostomum myrtifolium 

A Du Cp Ci H Chomelia spinosa 


i Bo Cp C1 Bobea elatior D 
p Bo Cp C1 Bobea sandwicensis E 
"T xi EJ Du Cp C! H Guettarda crispiflora 9 
bu Cp C! H Guettarda hirsuta 2 
An Du Cp C1 Antirhea smithii 
An Du Cp Ct Antirhea rhamnoides 
An Du Cp Ci! Antirhea chinensis v 
ja —— ^n Du Cp C: Guettarda splendens 3 
An Du Cp Ca Guettarda heterosepala E 
— — An Du Cp Ci H Guettarda humboldtensis — 8. 
An Du Cp C1 Guettarda trimera = 
pas —— —— An Du Cp Guettarda noumeana 9 
|| An Eua Cp C! Timonius densiflorus > 
An NN Cp Ci Timonius flavescens 8 
0.99 1 y 745 2 
, An pig Cp C: Timonius nitidus 8 
An [ui Cp C: Timonius polygamus ® 
0,60 An UU Cp C: Timonius timon 
An Du Cp C1 Antirhea borbonica 
Age An Du ? C1 H Guettarda acreana 
An Du Cp Ci H Stenostomum lucidum 
— An Du Cp Ci H  Neolaugeria densiflora 


1 
na p Du Cp Ci H Neol 


Figure 3. Important morphologic al characters mapped on the 50% majority rule Bayesian consensus tree. Numbers on 
inte da indicate posterior probabilities. Guettarda species are in bold. The key urs ate the state for each category of 
characters (Pollen type: Pi = Pittoniotis: An = e Gu = Guettarda; Bo = Bobea; ? = Le Fruit type: Cr = 

capsular; Du = se ans with united. pyrenes; Df o with ud pyrenes. Calyx type: C caducous B 
cna "ssion; Cp = persistent on fruit; ? = variable. hojas scence type: Pn = paniculate; C2 Ee MO dichasial cyme: 
- fundame sli monochasial cyme. Sexual system: H = hermaphroditism; D = in cy) 


Volume 93, Number 1 


Achille us 


2006 ROT in O 
50%. P = 0.92). These taxa exhibit several as trimerous flowers, corollas with valvate reduplicate 


distinctive characters that separate them from other 


with a few 


Along 


closely related species from Central and 


members of Guettarda. 
apparently 
South America that were not sampled, they form 
a well-defined group characterized by large stipules 
with membranaceous margins, scalariform venation, 
bifid monochasial inflorescences (which differ from 


the dichasial cymes found in other Neotropical 
Guettarda), small 4-ribbed ovoid fruits with a minute 
and persistent calyx (see also species descriptions in, 
1934; Steyermark, 1972; 


“Bobea type” pollen (see below 


e.g., Standley, Taylor et al.. 


2004), 


under Bobea). 


as well : 
Although additional molecular data will 
be required to confirm whether these taxa and Bobea 
are closely related, their distinctive morphological 
characters suggest that recognition at the generic rank 
might be in order in which case the name 
Tournefortiopsis would be available. as it is typified 
by T. reticulata Rusby. (= Guettarda tournefortiopsis 
(Rusby) Standl.), a close relative of G. crispiflora. 
The Guettarda species from New Caledonia consti- 
tute a third clade (Fig. 2) that is very well supported in 
the BI analysis (P = 0.99) although unsupported in 
the MP analysis (BS < 50%). 


three species of Antirhea (A. chinensis, A 


This clade includes 


. rhamnoides. 


and A. smithii) that would either belong to the 
segregate genus Guettardella according to Jansen 


(1984) or to Antirhea subg. Guettardella and Antirhea 
subg. Mesocarpa according to Chaw and Darwin 
(1992). This 
Paleotropical Dioecious Clade (see below). and is 
The New 


species of Guettarda are morphologically very distinct 


clade is nested within a 


the sister group of. Timonius. Caledonian 
from the Neotropical members of the genus and from 
fact, 


dioecism, a calyx that is almost always persistent 


G. speciosa. In several characters. me luding 


in 
fruit, and monochasial eymes (Fig. 3). are reminiscent 
of Antirhea s.l. The placement of these species within 
the Paleotropical Dioecious Clade is thus consistent 
and their inclusion in a broadly 
Baillon (1879). Guillaumin 
(1948). and more recently Jaffré et al. (2004) merely 


reflects an out-of-date concept of the genus that has 


with morphology. 


defined Guettarda by 


been abandoned in most other parts of the world. H 
therefore would seem appropriate to exclude the New 
Caledonian species from Guettarda and transfer them 
to another genus (see discussion below under 
Antirhea). 

It is worth noting that the resolution within the New 
Caledonian Guettarda clade is very weak despite the 
fact that its members are morphologically quite 


distinct from one another. Many New Caledonian 
species of Guettarda show one or more characters that 


are rare or unknown elsewhere in Guettardeae. such 


other 


broader 


aestivation, and a woody endocarp with up to 20 
The lack of 
reflects a high level of similarity in the IT 
of the 

cies ae New ( 
(Achille & Motley, unpublished). 
sequences are identical or differed by only a few 
that 


a closely related group within which rapid morpho- 


locules. resolution within this clade 


S sequences 
species sampled. In another study of ca. 20 
Caledonia and surrounding areas 
we found that the 
suggesting this clade comprise 


bases, may 


logical diversification has taken place. perhaps in 
a manner analogous to that observed within Guettar- 
deae as a whole. The pattern of diversification 
apparent among the New Caledonian species differs 
that of 


where resolution is better and 


substantially from the mostly Neotropical 


Guettarda s. str. clade, 
branches are longer (Fig. 1), even though there is 
comparatively little morphological variability among 
its members. In order to evaluate a hypothesized 
radiation centered on New Caledonia, it would first be 
necessary to sample additional species of Antirhea 
subg. Guettardella and subg. Mesocarpa (sensu Chaw 
& Darwin, 
New 


These 


1992). the apparent closest relatives of 
Caledonia Guettarda, according to our results. 
subgenera are most diversified. in Southeast 
Asia. but we were able to include only one species 
from South China (Antirhea chinensis). one from Fiji 
(A. smithii) and one from New Caledonia (A. 
rhamnoides). 

The fourth group of Guettarda identified in our 
study comprises a single species, C. acreana, which 
appears as sister to Stenostomum lucidum. This well 
known Guettarda, widely distributed in northern South 
America, was sampled in our analysis because of its 
distinctive morphology, particularly its calyx, which is 
mostly caducous but not shed by abscission. lts 
affinities are discussed below under Stenostomum. 

Our finding of polyphyly in Guettarda is consistent 
with morphological data, as several unique characters 
are associated with each of the clades revealed in our 
analysis (Fig. 3). Guettarda s. str. can be recognized 
by its caducous calyx. which appears to be a synapo- 
morphy for the lineage. However. there has been some 
confusion between what we regard as a truly caducous 
calyx (shed by abscission) and a calyx that is minute 
or lost as it disintegrates (without leaving an evident 
abscission scar) as seen in G. crispiflora, which has 
been misinterpreted by several authors (e.g.. Dwyer, 
1980: 1931. 1934; Standley € Williams. 
1979; 1974). Species of Chomelia have 
also been reported to possess both a deciduous and 


(Burger & 1993). but 


additional developmental studies will be required to 


Standley. 
Steyermark, 
persistent calyx Taylor, 
evaluate whether the nature of the calvx in this genus 


is homologous with that seen in Guettarda s. str. Our 


116 


Annals of the 
Missouri Botanical Garden 


molecular study does 


a close relationship between the lype species of 


Chomelia and Guettarda, as currently circumscribed, 


even though, according to Taylor et al. (2004), the 
morphological distinction between them remains 


unsatisfactory. With our more thorough sampling, i 


also appears that the association of Chomelia with G. 
l. (2002) may not 


reflect a close relationship between the two genera in 


crispiflora suggested by Rova et a 


their stricter sense and thus does not support the 


inclusion of Chomelia within Guettarda s. str. 


las 


defined here 
Further analyses will be necessary before extensive 


nomenclatural changes can be made within Guettarda 


22 


as most of the larger clades identified in our study 
are insufficiently supported in the MP bootstrap and 
Bl analyses. and broader sampling is needed for some 


of them. It would. however. seem appropriate to 


exclude the New Caledonia species from Guettarda, 
which clearly constitute a group that is distinct from 


all other members of the genus, as currently 


circumscribed, and i from the 


| particular type 
species, G. speciosa. 


OTHER GENERA OF CORE GUETTARDEAE 
Relationships within core Guettardeae. as within 
Guettarda, appear well correlated with geography. All 
the Paleotropical species sampled, except the Hawai- 
ian taxa and the widespread coastal species C. 
speciosa, belong to a Paleotropical Dioecious Clade 
).99: 


genera 


that is well supported in the BI analysis (P = ( 
Fig. 2). Since the closest allies of the 
comprising core Guettardeae are Neotropical (Bremer, 
1996: Bremer & Thulin. 1998: Rova et al., 2002). this 
New World origin for the group, with 
presumably two separate events involving dispersal 


suggests i 
suger sts £ 


[ex] 


(or vicariance) to account for the presence in the Old 
World of the Paleotropical Dioecious Clade and of C. 
speciosa. Dioecy characterizes all members of the 
Anti- 
rhea and the New Caledonian Guettarda species, in 


Paleotropical Dioecious Clade. viz. Timonius, 


contrast. to other core Gueltardeae. whose sexual 


system is almost always hermaphroditism, the only 


exception being the dioecious Hawaiian endemic 


genus Bobea (cited as polygamodioecious by Darwin 
& Chaw, 1990 

Timonius. Despite the presence of considerable 
morphological heterogeneity within Timonius (Wong. 
1988). 


easily distinguished 


genus is among 
Paleotropical Guettardeae by its free pyrenes. The 
five species included in our sample form a don that 
is well supported in the BI analysis (P = 1.00), but 
only weakly so in the MP analysis (BS = 60%). The 
notion that Timonius is related to the endemic 


not provide any evidence of 


Hawaiian genus Bobea is not confirmed by our 


results, despite the fact that they are both dioecious. 
a character found throughout the Paleotropical clade, 


and that they share free pyrenes, which are not known 


anywhere else in the tribe. Timonius is further 
distinctive among Rubiaceae because of its large 


number of ovary locules, which can exceed 500 


some species (Darwin. 1994), whereas most members 


of the family have bilocular ovaries. However, several 
other groups also show increases (albeit much more 


modest) in locule number. including some Neotrop- 


ical dedos Bobea, and most members of the 


Paleotropical Dioecious Clade. Thus. 
the 


most 


although char- 


acters of gynoecium are usually considered 


among the important in infrafamilial classi- 


fication of Rubiaceae. it seems very likely that an 


increase in the number of locules has occurred several 
times independently within Guettardeae. although 
the evolution of this character within the tribe may 
not be strictly homologous in the lineages in which it 
occurs., 
Antirhea. Although numerous genera and species 
have been associated with or included in Antirhea, our 
data suggest thal a restricted generic circumscription 


may be appropriate. The monotypic Neotropical genus 


Pittoniotis has sometimes been included in Antirhea 
(ee. bv Dwyer, 1980), but neither of our trees 
provides evidence for a close relationship with either 
Antirhea or the previously included Neotropical 
segregate genus  Stenostomum (Fig. 2). The 
morphological basis traditionally invoked for this 
putative relationship. is rather weak, essentially 
limited to the presence in both Pittoniotis and 
Antirhea of calyx-crowned drupes with few united 
pyrenes, characters that are in fact fairly common 


within core Guettardeae. Pittoniotis differs, however. 
in several important ways, sueh as having paniculate 
inflorescences, exserted stamens, and distinctive 
prolate 


latter 


3-colporate pollen with perforate exine (the 


feature occuring also in Stenostomum secl. 


Resinanthus Borhidi, represented in our analysis by S. 
acutatum and S. myrtifolium), whereas Antirhea has 


monochasial cymes, included stamens. and oblate- 
spheroidal 3-colporate pollen with reticulate exine 


( Achille, data). All 


features of Pittoniotis mentioned above are 


unpublished the distinctive 
shared 
with Malanea (Fig. 3), but occur nowhere else within 
in this 


core Guettardeae (the outgroup taxa used in 


study, however, exhibit some of 


obs. ). 


and 


. pers. 
This suggests an affinity. between Pittoniotis 
the 
sister relationship indicated by the BI analysis (P = 


Malanea that would be congruent with 


— 


9.85). 
The ITS data clearly indicate that Antirhea s.l. (i. C., 
the 


including Stenostomum) is polyphyletic. Even 


Volume 93, Number 1 
2006 


Achille et al. 117 


Polyphyly in Guettarda 


more restricted. definition of the genus proposed by 
Chaw and Darwin (1992), 


Paleotropical species but excluded Stenostomum, 


who included only the 


appears to be paraphyletic with respect to Paleotrop- 


ical members of Guettarda and Timonius. Our 
sampling, while modest, suggests that in order to 
render Antirhea sensu Chaw and Darwin (1992) 


monophyletic, it would be necessary to include both 


Timonius and the New Caledonia species of Guet- 
tarda. M such a broad definition were used, Antirhea 


would constitute a very large genus (comprising about 


— 


230 species) that corresponds more or less 
Schlechters (1906) 


Adopting such a circumscription would, however. 


broad concept of Timonius. 


make it very difficult to define Antirhea based on 
morphological characters and thus does not seem 
advisable. 

It would therefore seem best to limit Antirhea to 
the 


borbonica (the type species) and two close re ‘latives 


species from Indian Ocean, comprising A. 


(A. bifurcata (Desf.) Hook. and A. madagascariensis 
Chaw), which together correspond to Antirhea sensu 
Jansen (1984). or Antirhea subg. Antirhea sensu Chaw 
(1992). 
(represented by A. borbonica only) is sister to all the 
= 63%, P = 


Antirhea would be 


and Darwin In our analysis, this group 


other Paleotropical Guettardeae (BS 
0.99). 


easily characterized by its monomorphic male and 


A more narrowly defined 


female inflorescences (i.e. which have similar 
branching patterns). ovaries with only 2 or 3(4) 
locules, and the presence of erystals in the leaf 


epidermis. 
If this 
adopted, a new generic placement would be needed 


narrower interpretation of Antirhea is 
for the species belonging to Antirhea subg. Guettar- 
the 


type species of Guettardella, and by A. rhamnoides) 


della (represented in our study by chinensis, 


and Antirhea subg. Mesocarpa (represented by 
smithii). Chaw and Darwin (1992) recognized both 


these subgenera, which together formed a clade in 


their analysis. The first subgenus comprises e 


mostly from Malesia, Southeast Asia, and the 


southwest Pacific with small, 3- to 11-locular fruits. 


and the second includes Melanesian species with 
large, 8- to 16-locular fruits. Taken together, this 
group corresponds to Guettardella sensu Jansen 
(1984), which could be resurrected to accommodate 
New 
Guettarda may belong to the same broad clade as 
the Guettardella group (BS < 50%, P = 0.99), 


could perhaps also be included in an even more 


these taxa. As the Caledonian 


they 


broadly defined Guettardella. If this relationship is 


confirmed by future studies and a substantially 


enlarged Guettardella is recognized, it would, howev- 


er, constitute a heterogeneous group, since the New 


species ( : 


Caledonian species are morphologically quite variable 


and different from those currently included in 
Guettardella. 
Stenostomum. Our results suggest that Steno- 


stomum may also be polyphyletic, with S. lucidum in 
a weakly => sister relationship to Guettarda 


acreana (BS = = 0.89) and two other species 


acutatum B 5. 1 1 forming a separate 


clade (BS = 94%, P = 1.00) that is sister to Chomelia 


in the MP analysis (BS — 52%). As mentioned above, 
Borhidi and Fernandez (1993-1994) placed S. 


lucidum in Stenostomum sect. Stenostomum, whereas 
the other two species in our sample belong to their 
section Resinanthus. Our study supports the idea that 
these two groups are distinct, as suggested by 
morphology (Borhidi € Fernandez, 1993-1994), but 
their placement in different parts of our trees would 
not be consistent with maintaining them as a single 
genus. The possible sister relationship of G. acreana 
to the type species of Stenostomum, although weakly 
supported in our trees, would on the other hand be 
consistent. with morphology, as several characters 
shared between these taxa (e.g., similar inflorescence 
structure, small ovoid fruits, and a calyx that is not 
found in members of 


shed by abscission) are not 


Guettarda s. str. For example, Bremekamp (1959), 
who was familiar with Stenostomum (under the name 
Central 


Antirhea) America and the West Indies, 


while studying an undetermined specimen of G. 


acreana collected in Surinam, recognized it as an 


Antirhea. He consequently did not suspect an 
relation of his specimen to Guettarda and described 
it as a new species, A. surinamensis Bremek. It might 
thus prove appropriate to include the South American 
G. acreana within Stenostomum, a relationship that 
: 1972) perhaps 
Stenostomum 


the West 
Borhidi & 


earlier authors (e.g... Steyermark, 


overlooked simply because occurs 


from Indies to 


'arther to the north, 


America Fernandez, 


northern Central 
1993-1909 
Despite being well distinguished by having stipules 
united into a sheath, Neolaugeria was merged into 
Stenostomum (as a third section) by Borhidi and 
Fernandez (1993-1904). We 


this interpretation, as the two species of Neolaugeria 


are unable to confirm 


in our sample (which comprise a well supported pair. 
35 = 100%, 
with the S. lucidum clade in the MP tree, whereas they 


— 


P = 1.00) form a paraphyletic group 


appear as sister to the rest of the ingroup in the Bl 
tree, ee both topologies are unsupported (BS < 
50%. P = 0.63). Moynihan and Watson (2001) also 
cee the inclusion of Neolaugeria in Stenostomum 
using ITS sequence data. Their results were, however, 
based on an analysis that included only one species of 
a member of section 


Stenostomum, S. myrtifolium, 


118 


Annals of the 
Missouri Botanical Garden 


—_— ͤ—- —ꝛꝛꝛꝛꝛꝛꝛññ̃ ́R Qꝗ ³ “!?!! 


Resinanthus, which in our study was resolved. in 


a clade distant from the Neolaugeria group. 


Stenostomum has been considered a synonym of 


Antirhea by some authors (e.g., Schumann, 1891: 
Standley, 1934; Bremekamp, 1966: Airy Shaw, 1973; 
Steyermark, 1974; Robbrecht, 1988) because these 


two genera are morphologically similar in several 


aspects (e.g, monochasial cymes, a stony endocarp 
often with few locules, and a persistent calyx). Our 
results suggest that the type species of Stenostomum 


is indeed closely related to Antirhea, and in particular 


0 borbonica (the type of that genus), although 


placed sister to the Paleotropical clade. Our data are 


thus consistent with Chaw and Darwins (1992) 
suggestion thal Stenostomum is best. treated as 
a distinct genus. 

Bobea. The exclusion of the Hawaiian genus 
Bobea from Guettardeae was proposed by Rova 


(1999), but both our analyses suggest that Bobea is 
nested well within the tribe (BS = 99% and P = 
1.00). Furthermore, the placement of Bobea within 
core Guettardeae has long been recognized based on 
morphological features. Bobea has numerous 
characters in common with other genera of core 
Guettardeae, including typically secund, monochasial 
inflorescences that are once-dichotomous at the base: 
fleshy fruits with numerous similarly-structured 


pyrenes; elongate cylindrical seeds with smal 
cotyledons and topped by an obturator; and overall 
vegetative and floral morphology which closely 
resembles some species of Antirhea, Guettarda. and 
Timonius. The position of Bobea in the tree presented 
by Rova (1999) thus seems at odds with both ITS data 
and morphology, suggesting that Rova's placement 
may have resulted from an error. in sampling or 
analysis. 

Within core Guettardeae, Bobea does not appear to 
be closely related to Timonius, despite the fact that 
these two genera were once considered as synonyms 
(Candolle, 1830) or, more recently, as close allies 
(Darwin & Chaw, 1990). On the contrary, as indicated 
earlier, Bobea may instead be sister to G. erispiflora + 
G. hirsuta, thus forming the Bobea Clade, although 


the BI 


analysis (P = 0.92) and unsupported in the MP tree 


this relationship is weakly supported in 


(Fig. 2). It is worth noting, however, that all of these 
species are well-characterized (Fig. 3) by the presence 
of “Bobea type” pollen (spheroidal porate pollen with 
very thick, reticulate exine; Achille, unpublished 
data), which appears to be unique in the tribe, and 
also by their inflorescence structure (bifid) mono- 


chasial cymes), distinctive from Guettarda s. s 


r. 
New World origin of Bobea may thus be plausible. 
since G. crispiflora and its relatives are distributed in 


the Neotropies and in particular on the Pacific coast of 


Central and South America and the Cocos Islands. 
rather than a western Pacific origin, as implied by 
a traditionally assumed relationship to the Paleotrop- 
ical Timonius. If Bobea is indeed derived from 
a Neotropical ancestor, its dioecy may have evolved 
independently from the Paleotropical | Dioecious 


Clade, perhaps as an adaptation to an insular 
environment, a pattern that is seen in many other 
Hawaiian groups (Sakai et al., 1995). Hawaii has one 
of the highest incidences of dioeey (14.7%) of any 
llora in the world, and many dioecious lineages are 
derived from dimorphic ancestors, which was thought 
to be the case for Bobea (Sakai et al.. 1995). Bobea 
may thus be a 13th Hawaiian lineage that is believed 
to have evolved dioecy from hermaphroditic colonists 
(Sakai et al., 1995). 


CONCLUSIONS 


Our study. using a broad sampling that represents 
J : | 8 


most of the morphological and geographical diversity 


within Guettarda, indicates that the genus is poly- 
phyletic. Guettarda should probably be restricted to 
a group of species characterized by dichasial cymes 
(sometimes reduced to solitary flowers) and a caducous 
calyx with a circular abscission zone, which is well- 
diversified in the Neotropics, but also comprises 
a single widespread species in the Paleotropics. The 
numerous endemic New Caledonian species currently 
assigned to Guettarda clearly belong to a distinct 
Intirhea-Guettardella complex, whose internal. rela- 
tionships are in need of further study. Our results also 
that 


Neotropical species traditionally included in Guet- 


indicate several morphologically distinctive 
tarda are placed outside Guettarda s. str., although 
their exact affinities remain unclear. 

Our study also provides some interesting insights 
concerning the delimitations of other genera in core 
Guettardeae and the distribution of characters within 
the tribe. Our data confirm that the Neotropical genus 
Stenostomum is distinct from Antirhea, which should 
the Old 
World. Furthermore, Antirhea and Stenostomum both 


thus be considered as strictly limited te 


appear to be polyphyletic and may thus require the 


application of narrowed generic cireumseriptions. 
Features of the gynoecium (such as the number of 
locules and fusion of the pyrenes), which have often 
been considered among the most important characters 
for recognizing suprageneric taxa within Rubiaceae. 
do not appear to be particularly well correlated with 
the groups identified in our analysis. Several in- 
teresting correlations have, however, been found 
between our phylogeny based on ITS data and the 
distribution of certain previously-ignored or under- 


utilized characters. For example, a large Paleotropical 


Volume 93, Number 1 
2006 


Achille e 119 


E in 1 


Dioecious Clade is characterized by its distinctive 
sexual system. The recognition of several pollen types 
among the members of core Guettardeae, and the 
homogeneity in pollen type seen within several clades 
identified in our phylogeny, also suggest that 
palynological characters may be valuable for further 
studies at the generic and/or specific levels within the 
tribe. Finally, with regard to biogeography, although 
Guettarda s.l. and Antirhea s.l. have generally been 
distributions with 
in the Old 


study 


considered to show trans-Pacific 


two disjunct areas of diversification, one 
World 


suggests that this interpretation was erroneously based 


n the Neotropics, our 


— 


and another 


on non-monophyletic generic concepts. 


Literature. Cited 


Airy Shaw, H. K. 1973. A Dictionary of K "lowering Pla ints and 
sth edition. € 
1974. A new look al pni 


IEEE 


— 


Ferns, € Cambridge Un 
Akaike, H. 8 

ide ntific ation. rans. FM Control 
716-72 


Allan. G. J. 


phyle 


M. Porter. 2000. Tribal de limitation and 


relationships s of 


ogene tie I oleae and oroni 


(Faboideae: Fabaceae) with special reference e Lola tus: 


Evidence > from nuclear “‘bosomal ITS sequences. Amer. 
J. Bot. 87: 1871-1881. 

Alvarez. ] $ J. F. Wendel. 2003. Ribosomal ITS sequences 
and plant p "o netic s Phylogenet. 
Evol. 29: 417— 


b. A na Checklist of Neotropical 
: 1-199 


inference. Molec 


Andersson. l. 

Rubiaceae. Seripta Bot. Belg. 1 l 

& A. 2005. tribe 
(Rubiaceae). its position in Cinchonoideae, 


Antonelli. Phylogeny of the 
Cinchoneae 
and description of a new genus, Ciliosemina. Taxon 54: 
l 7-28 

Andreasen, K.. B. Baldwin & B. 

utility of the N 
the subfamily Ixoroideae (Rubiaceae): Cotaparicons with 
cpDNA a sequence data. PL Syst. Evol. 217: 119-15 

Baillon, E. 1879. Sur les limites du genre 

: 316-318. 

Histoire ps 

Pp. 257-503. Hachette, Paris. 

ol netic 


1990, 
ITS region in 


Bremer. 


Phylogenetic nuclear rDNA 


Gue pcd 
uci 12 


Plantes, Vol. 7. Rubiaceae. 


QQ 
OO 


Baldwin, B. G. 1992. utility of the internal 
transcribed spacers of nuclear ribosomal DNA in plants: 
An example from the C ou ius Molec. Phylogenet. 


Evol. 1: 3-16. 

& M. 

diversification of the Hawaiian 

( Evolution 95: 9402 
. J. M. Porter, M 

Guna ll & M. J. 


nuclear ribosomal DNA: 


Sanderson. 


5 alliance 


— 


ompositae). 


— 


zu ojciec 
1995. The 
valuable source 


SH 
pe evidence 
Gard. 82: 


on angiosperm phylogeny. pu" Missouri. Dot. 


241—211 Te 
Barriel, V. 1994. Phylogénies moléculaires et. insertions- 
délétions de nuc p ompt. Rend. “Hebd. Séances 
Acad. Sci. 317: € 01. 


Baum. D. A. R. Small & J. F. Wendel. 1998. 
Biogeography and floral e evolution of Baobabs (Adansonia, 
Bombac aseo as inferred from multiple data sets. Syst. 
Biol. 47: 181—207. 


1998. Age and rate of 


ski, 
region an 


1993-1994. [1996]. The 

the 

resurrec A of the New World Antirhea species. Acta Bot. 

Hung 55 

M C. 
(Rubiaceae) from Suriname. 


Fernandez. 


Borhidi, A. & M. Z. 


Stenostomum C. F. Gaertn. (Rubiaceae) o 


genus 


E. B. 1959. A species of Antirhea 
Acta Bot. Neerl. 8 179-481. 
emarks on the position, the de ia and 

I5: 


Acta Bot. Neerl. 


the subdivision of the Rubiaceae. 


Bremer, ie 1996. Phylogenetie studies within Rubiaceae and 
relationships to other families based on molecular data. 
Opera Bot. Belg. 7: 33-50. 

997 (as 1996). ( 

—— and molecular data in Rubiaceae. Cladis- 

tics 12: 21-40. 

& R. K. Jansen. 


Combined and separate analyses of 


199]. 
1 of chloroplast DNA implies new phylogenetic 


Comparalive re striction site 


5 ships within Rubiaceae. mer. J. ot. 70 
19 ues 

— —— & L. Struwe. 
the Loganiaceae: Congruence or conflict between 1 5 
logical and molecular data. Amer. J. Bot. 79: 1 17 71— 
N 


M. Thulin. 1999. 


establishment of Mussaendeae, and a new genus of 


1992. Phylogeny of f the Rubiaceae and 


Collapse of 


Sabiceeae (Rubiaceae); phylogenetic re cone based 
on eee data. PL . Evol. 211: 71-92 

Andreasen & D. Olsson. 1995. Subfamilial and 
Rubiaceae based on rbcL 
. Missouri Bot. Gard. 82: 383—397. 
K. Jansen, b. Oxelman, M. Backlund, H. 


K.-]. Kim. 1999, More characters or more taxa for a robust 


abel re rations in 
sequence data.“ 
¢ Lantz & 


— 


phylogeny in a from the Coffee family 
Syst. Biol. 
C. M. 1 9 1 Flora 
Family 202, 1 Fieldiana Bot. 33. 
V. P. de. . Prodromus Systematis Naturalis. 
Regni Vege M o 
Chandler. G. T. & G. M. Plunkett 2004. 
ales; Nuclear and chloroplast markers together in 
I: ] 


(Rubiaceae). 


burger, W. 


Costaricensis. 


Candolle, 

Evolution in 

(almost) pe lec : ES Bot. J. Linn. Soe. 23-147. 

Chaw, S. M. € S. P. Darwin. 1992. A systematic study of the 
Old World species of Antirhea (Rubiaceae: Guettardeae). 
Tulane ES Zool. Bot. 28: 25-118. 


Darwin, < . 1993. A revision of Timonius ic nus 
Timonius 1 iceae: Gue tarde sae). Allertonia 7: 1-30 
1994, Systematics of ee 


` Timonius subgenus 


Jd 


(Rubiac eae: ign prs Sy 
& S. 


st. Bot. M 

Bobea Gaud. (Rubiaceae). 
). Herbst & S. Sohmer 

Hawaii, 


E^ oun 

(editors), Manual of the Flowering | 

Vol. 2. Univ. Hawaii Press & Bishop Museum Press. 
De e te, P. G. 2004. ay 331 
Heald, A. 


Flowering 


N. 
Henderson, S. A. Mori & D. W. 


Plant Families of the 


American Tropics. New York Botanical Garden Press, 
New York/Princeton Univ. Press, Princeton. 
Donoghue. M. J. & B. G. Baldwin. 1993. Phylogenetic 


analysis of Viburnum based on ribosomal DNA sequences 
from internal transcribed spacer regions. Amer. J. Bot. 80 
(supplement E Abstract]: 145. 
Dwyer, J. D. 1980. Flora of Panama, Part. IX. Family 
79. 1 Part L Ann. Missouri Bot. Gard. 67: 
255. 


Confidence 


1985. 


An approach using the bootstrap. Evolution 39: 
791 


“elsenstein, J. limits on phylogenies: 
783 
10 


120 


Garcia-Jacas, N., A. Susanna, T. Garanatje & R. Vilatersana. 


: JOI. Generic de tao and phylogeny of the subtribe 

Sentaureinae (Asteraceae): combined nuclear and 
e DNA analysis. Ann. Bot. 87: 503-515. 

Grisebach, l. 1858. Novitiae Florae panamensis. 


Bondisn d 6: 6. 
Guillaumin, A. 1948. Flore analytique et synoptique de la 
ouvelle-Calédonie; | phanérogames. fice de la re- 
cherche scientifique coloniale, Paris. 

Biologia Centrali- Americana Botany 
12. F. D. Godman and O, Sal 

P. K., N. H. 

Index Herbariorum. I 


ed. 


He ig y. W. B. 1881. 
Vol. 2: Lone 

Holmgren & Barnett. 1990, 

Part 1. The B of the World, 8th 

20 


vin. 
1 17 55 TN 


> 
75 
N 


. Rubiaceae. Pp. 7-151 i 
J De Me 11 5 1 Genera 1 en 
Howard, R. A. 1989. Flora of a 
and am Islands. Vol. 
vard University, Massac 9 10 
Huelsenbeck, J. P. & E. 


Bayesian 
755 


e ntham & 


Tas ard 


Har- 


lesser 1 


Arnold Arboretum. 
200 
Bioinformatics 


honquist. 


l. MRBAYES: 

inference of phylogeny. 17: 
754-755. 

——, B. Le R. E. Miller V. Ronquist. 2002. 
Points] applications and pitf alls 7 Bayesian inference 
of  phyloge ny. Syst. Biol. 51: 073-688. 

Jaffré, T., P. Morat. J.-M. Veillon, F. Rigault & G. Dagostini. 
2004. Composition and Characteristics of the native flora 
of New Caledonia, 2nd ed. IRD 

Jansen, M. E. 1984. A synopsis of Guettardella Benth. and 

Old World species of Antirhea A. bs 
(Rubiaceae: Guettardeae). Bluse a 29: 566- 
Kirkbride, J. H. Jr. 1997. 

Brittonia 49: 354—379 
amarck, J. B. de. 1792. CUR encyclopédique et 
méthodique pa trois règnes de 


Panckoucke, 


irgel. 


Nouméa. 

de Jussieu 
20 

38 


Manipulus an 'earum- VI. 


la nature. Botanique: 2. 


Lantz, IL, K, 17 usen & B. Bremer. 2002. Nuclear 
rDNA ITS sequence data used to co 7 t the first 
phylogeny of Vanguerieae (Rubiaceae). Pl. Syst. Evol. 


230: 173-187. 


Linnaeus, E 1753. Species plantarum. Stockholm. 


„ i D 1907. The Plant Book. 2nd ed. Cambridge 
niv. Press, Cambri 95 5 


a l. & B. 


distribution in 


. 1998. 


(Rubiaceae): 


Phylogeny, diversity and 
Exostema implic ations of 
Syst. | 


m al and molecular analysis. PL vol. 
212: 215-246 
Motley, T. Wurdack & P. G. 


Delprete. 2005. 


Molecular systematics of the Catesbaeeae—Chiococceae 
complex (Rubiaceae): Flower and fruit evolution and 
biogeographic implications. Amer. J. Bot. 92: 316- 


329. 


Moynihan, J. & L. E. Watson. 2001. Phylogeography. generic 
allies, and nomenclature of Caribbean endemic genus 
Mage ides Buby iceae). based on ITS sequences. Int. 
. - 1 2 39: 3—101. 


. Fragmenta. Phytographiae Australiae: 9, 

Nepokroeff, M., B. Bremer & 
organization of the genus 
b infe “rre ed f 
Syst. Bot. 

Nylander, ^ 


distributed by tł 


1999, Re- 


Psychotriae 


. J. Sylsma. 
Psychotria and tribe 
from ITS and eL 


sequence data. 


. 2004. 
1e author. 
University. 

|> 


MrModelte v2. 
Biology Centre. 
<http://vww.ebe.uu.se/sy stzo0/st all / 


Program 
E Ku 
Uppsala 


nylander.htn 


Annals of the 
Missouri Botanical Garden 


Persson, C. 2000. Phylogeny 


group (Rubiaceae). with 


of the Neotropical Alibertia 
emphasis on the genus Alibertia, 


ee. som - and 58 ribosomal DNA sequences. 

. J. Bot. 87: 1018-1028 
1 G. " T P. P. Lowry IL 2001. Relationships 
among "anc ient araliads” and their significance or the 
Apiales. Molec. Phylogenet. Evol. 19: 


systematics of 
276 


owry II. 2004. Infrafamilial 


classifications: and Insights 


from the phylogenetic analysis of nuclear (ITS) and 
plastid (trnL-4rnF). sequence data. PL Syst. Evol. 245: 
1-39. 

Porter, J. 1993. Phylogeny of Gilia sect. Giliandra 
(Pole ——] Ph.D. Dissertation. University of Ari- 
zona. Tucson. 

Posada. D. & K. A. Crandall. 1998. 


Modeltest: Testing the 
model B DNA substitution. Bioinformatics 14: 81 
& 1996. 


7-818 


os 7. Yang. Probability distribution of 


lec i ar edition. wy trees: A new method of phyloge- 
netic inference, J. ae Evol. 43: 30431 1. 
Razafimandimbison. 5. & B. Bremer. 2002. Phylogeny and 


aceae) inferred from 


; and 1 m morphological data. 


mer. E ig : E 027-1041. 
Robbrecht, 195 1 ‘al woody Rubiaceae. Character- 
istie features and progressions, 


b r de to a new 
lg. 1: | 


subfamilial c lassifi ic EM Opera Bot. Belg. -271. 


Rova. J. 1999. The Rondeletie 1 55 condaminecaec— 
PL ae Complex. (Rubiaceae). Ph.D. Dissertation. 
Göte 5 University, Sweder 
. Delprete, L. / T & V. A. Albert. 2002. A 
Irnl- 1) NA "LE study of the Condamineeae— 
H 


nple x with implications on the 


phylogeny of Ms Rubiaceae. Amer. J. Bot. 89: 145-159. 
Rusby. | 1907. An enumeration of the plates collected 
in ee by 15 55 Bang. Part 4. Bull. New York Bot. 
"ed s 
Sakai, / a a Wagner. D. M. Ferguson & D. R. Herbst. 
Pu A the Hawaiian flora. Ecology 


Origin e dioecy 

76: 2517-2529, 

Schlechter. F. R. 1906, Beiträge zur Ke nntnis der Flora von 
dene Bot. Jahrb. Syst. : ds 25 

Schumann, K. 1891. Rubiaceae. Pp. LA. Engler & K. 
Prantl (editors). d naturlichen P 1 15 Rae n, IV. 4. 
W ilheli m E nee ur I ¿e pig. 

Standley, P. 


51-258. 


The Rubi: aceae of Venezuela. Publ. 
Field. ÓN 15 1 Bot. Ser. 7(4): 340-455. 
1934. Rubiaceae. North Amer. Fl. 32: 214—277 
K L. O. Williams. "a 75. Flora of Guatemala— 
Rubiaceae. Fie wee Bot. 24: 1-275 
Steyermark 1972. 9 ina Pp. 237-832 in 


Maguire & ll alors (editors), Botany of the ( 


Guayana 
Highlands, Part. Bot. ( 23. 


Mem. New York yard. 


Rubiaceae. 


. 1974. Guettardeae. Pp. 772-867 in T. 

Lasser & J. A. Steve pad fe editors}, Flora de Venezuela 9. 
Swofford, D. L. 1998. PAUP*: Phylogenetic Analysis Using 
Parsimony ( and other me sthods ) ver. 4. Obl. Sinauer, 


Sunderland, Massac hu: 
Taylor, C. M., J. A. 


sells. 


Steyermark, P. G. Delprete. A. 


Vincentini, R. Cortés, D. Zappi. C. Persson, C. Bestetti 
osta & E. Araujo ds Anunciação. 2004. Rubiaceae 
Pp. 497-847 in Steyermark, P. Berry. K. Yatskievych 


ni. A. 
Hols (editors) 


& 15 Flora of the Venezue telan Guayana, 
Vol. 8. Missouri Borana Garden Pres Louis. 
Valeto E 1909. Beiträge zur Kenntniss der Gattung 
Timonius Bull. Dept. Agric. Ind. Néerl. 26: 1-61. 


Volume 93, Number 1 
2006 


Achille et al. 


Polyphyly in Guettarda 


Verdcourt. B. 1958. Remarks on the 
Rubiaceae. Bull. Jard. Bot. État 28: 209— 
White, TJ. T. 


Bruns, 


& J. Lom 1990. 
Amplification and direct sequencing of fungal 5 


(A. genes for phylogenetics. Pp. 315-322 in M. Linn 


17 8 wae of the 


D. CRUSH J. 


s K. M. 1988. 


Malay Pe 1 


Sninsky & T. White (editors). PCR 


A Guide to Methods and Applications. Aca- 


Diego. 


The ph es of the 


Kew Bull. 


T —————— ——————— Ru 
A SYSTEMATIC REVISION OF Cristina Inocencio,* Diego Rivera,” 
CA PPARIS SECTION CAPPARIS M" Concepción Obón, PRISE Alcaraz. 

. E ia and Jose-Antonio Barreña 
(CAPPARACEAE)! 
qn TOA AA 


ABSTRACT 


systematic revision of Capparis sect. Capparis, from western and Central Asia. North Africa. and E urope. is presented 
here. The taxonomy of this section has been approached combining morphologic “al, biogeographical and molecular data when 
available. Ten species are recognized, including two new species, Capparis atlantica and C. zoharyi. ln addition, four new 
subspecies are presented: Capparis ovata subsp. myrtifolia, C. Tu or sphaerocarpa, C. sicula subsp. n me sopotamica, 
di 


and C. sicula aun: sindiana. Lectotypes are da | for C. aegyptia, C. hereroensis, C. mucronifolia. C. elliptica, C. 
M Boiss subsp. rosanoviana, C. rupestris, C. ovata, C. pm C. spinosa var. canescens. C. sicula subsp 


verbacea, and C. sic a subsp. leucophylla. A full taxonomic treatment, keys, and distribution m: aps of hes recognized specie 
are provided. The Iwo new species are illustrated. 
Key 


aac 


A ords: aapparae eae, G apparis, taxonomy. 


—˙.¼ LLL 


Capparis (KAnTAPIG) is a name coined by Theo- transferred. to section Monostichocalyx Radlk. and 
phrastus (4th century BC) and endorsed by Dioscor- — section Busbeckea (Endl.) Benth. & Hook. Subsection 
ides (Ist century AD). It seems to have come into wide — Seriales DC. Racemose inflorescences or in series on 
use after the spread of the Arab culture in the Middle the stems. Representative species: C. zeylanica L., C. 


Ages. The Genus Capparis was created by Linnaeus acuminata Willd.. C. quadriflora DC.. C. rotundifolia 


(1753, 1754) with the description of Capparis spinosa Roul., C. brevispina DC. Subsection Corymbosae DC. 
L. and other Capparis species (Jarvis et al.. 1993), Corymbose inflorescences. Representative species: C. 
Capparis comprises around 250 species distributed in sepiaria L., C. umbellata Brown ex DC., C. incanescens 
tropical and subtropical zones of southern America, DC. subsection Octandrae DC. Flowers with 8 


Europe, Africa, Madagascar, Asia, Australia, and the stamens. Representative species: C. racemulosa DC., 
Pacific Islands (Willis, 1988). C. oleoides Burch.. C. coriacea Burch. 

New World taxa are included in five sections. 
TAXONOMIC HISTORY Section 2. Capparidastrum DC. Stems generally 


thorny, oval sepals without gland, and short gyno- 


Candolle (1824), who proposed a sectional division phore. Representative species: C. brasiliana DC., C. 
of the genus, provided the first comprehensive macrophylla HBK, C. cuneata DC. Section 3. 
systematic approach to Capparis. Old World taxa Cynophalla DC. Stems generally thorny, oval sepals 


belong to Section 1. Capparis (= Eucapparis DC.). possessing basal gland or foveola, and large gyno- 
Representative species: C. spinosa L. with four phore. Representative species: C. cynophallophora L., 
subsections. Subsection Capparis (= Pedicellares C. guayaquilensis HBK, C. amplissima Lam. Section 
DC.). Flowers always solitary at leaf axils. In addition A. Calanthea DC. Rounded fruit section, thin sepals, 
to C. spinosa and related species, other representa- linear, and sharp. Representative species: C. nemo- 
tives include species such as C. cartilaginea Decne. rosa Jacq.. C. pulcherrima Jacq. Section 5. Breynias- 


Some species with geminate flowers or in bune hes of trum. DC. Triangular sepals, sharp: always very short 


three, rarely alone, such as C. horrida L. 


. E gynophore. Representative species: C. ferruginea 
pubiflora DC., and C. canescens DC., were later C. incana HBK. C. indica (I.) Rave. & Rendle. 


This work was part of the Ph.D. Dissertation presented by C. Inocencio to the Faculty of Biology. i la, Spain. P3 thank 
d Wes herbaria and 11 0 ‘ir ves 20 1 loans, access to collections, and assistance: BISH, BM. BR. C. E. G. HUB. 
HUJ, K, JE. JEPS, LIV, MA, MARSSJ, MUB. OXF. P. RNG, RSA, UMH, and US. We thank 5 Perales for Mene 
Figure P 59 grati d is due to a hen Müller (JE), Vladimir Dorofeev (LE) gone Marner (OXF), Donna Young (LIV). 
and Aljos Farjon (K) for their help in finding type material. This researc d was supported by a grant from the Spanish Ministry 
of Education and Culture and by the project AGF96-1040 of DGICY 
“The editors of the Annals thank 9 Balcomb for her editorial US to this article. 
"Departamento de Biología Vegetal, Facultad de Biología, Universidad de Murcia. E-30100 Espinardo, Murcia, Spain. 
drivera@um.es. 
‘Departamento de Biologia Aplicada, Escuela Politécnica Superior de Orihuela, Universidad Miguel Hernández, E-03312 
Orihuela (Alicante) ). Span 


ANN. Missouri Bor. Garp. 93: 122-149. PUBLISHED on 31 May 2000. 


Volume 93, Number 1 
2006 


Inocencio et al. 123 


Revision of Capparis Sect. Capparis 


fruit section. 


DC. 


C. jamaicensis Jacq C. 


Section 6. Quadrella Angular 
Representative species: 
intermedia HBK, C. 
C. decidua (Forssk.) Edgew. with deciduous leaves, up 
lo 20 X 3 mm in size, is included by Candolle (1824) 
in the genus Sodada Forssk. 

This was followed in part by Bentham & Hooker 
(1862), although they 
section Sodada (Forssk.) Benth. & Hook. including C. 
decidua: section Busbeckea (Endl.) Benth. € Hook. 
including Australian species; and section Beautemp- 
sia (Gaud.) Benth. & Hook. including one American 


crotonoides HBK. The species 


created three new sections: 


species. 

Zohary (1960) proposed new systematics based on 
a partial geographical review of this genus restricted 
West He 


recognized two biogeographical groups: the tropical, 


to the Mediterranean region and Asia. 


including Capparis decidua, C. cartilaginea, C. 


mucronifolia Boiss., and the Mediterranean, including 


species that have lost their links with the tropical 


African stock (C. spinosa, C. sicula Veill., C. 
leucophylla DC.). All species belong to section 


Capparis in the sense of Candolle (1824). 

Jacobs (1965) attended to the Capparis species from 
Pacific, and organized them into four 
with C. 
with C. 


a new 


India to the 


sections: l. section Capparis, monotypic 


spinosa, s.l; 2. section Sodada, monotypic 
3. section Monostichocalyx Radlk., in 


the 


decidua; 


circumscription containing most of species 
formerly included in section Capparis (= Eucapparis 
DC.), with about 65 species in the area under revision; 
4. section Busbeckea, with 14 species in all. 

Higton and. Akeroyd (1991) reviewed the diversity 


of Capparis in the Mediterranean region, especially in 


connection with C. spinosa, examining six species in 
one section, which were finally reduced to a single 
species with two subspecies. 

Hall et al. (2002) analyzed sequence variation for 
a large sampling of Brassicaceae and Capparaceae, 
using two chloroplast regions, irnL-trnk and nd. 
The 


strongly supported the monophyly of Brassicaceae 


results of parsimony and likelihood analyses 


plus Capparaceae and recognized three clades: 


Capparoideae, subfamily 
Habit 


three 


Capparaceae subfamily 
fruit 


All 


Capparoideae are woody, which is the plesiomorphic 


Cleomoideae, and Brassicaceae. and 


characteristics demarcate these clades. 


condition for the Brassicaceae and Capparaceae 
clades. The herbaceous habit is generally found in 
Cleomoideae and Brassicaceae. Indehiscent, fleshy 


fruits are plesiomorphic in Capparaceae and Brassi- 
caceae and are the dominant fruit type of Cappa- 
roideae. Brassicaceae and Cleomoideae both have 
capsules with a replum, a synapomorphy 


The 


dehiscent 


shared by these two clades. 


dehiscent fruits of 


Capparoideae are almost always fleshy in nature in 
contrast to the often dry fruits of Cleomoideae and 
Brassicaceae. Floral symmetry, stamen number, leaf 
type, and fruit type all show homoplasy. Clades within 
Capparoideae show a biogeographical pattern based 
on this sampling with a New World clade (all New 
World representatives of Capparis, Belencita Karst.. 
Morisonia L.. Boscia Lam., and Cadaba Forssk.) and 
Old World (Maerua Forssk., Tylachium 
Ritchiea R. Br., The Old World 
species Capparis tomentosa Lam. is nested within an 
otherwise New World clade (based only on trnL-trnF). 
the 
Capparis, is nested within an Australasian clade, with 
Apophyllum anomalum F. Muell. (western New South 


Wales) and Capparis callophylla Blume (Java) (based 


an clade 


and Cadaba). 


Our., 


Old World Capparis spinosa, lype of genus 


only on nE). 


DNA sequencing of the chloroplast rebL gene 
nested New World Capparis hastata Jacq. and Old 
World C. sandwichiana DC. in the same clade, but 


sampling was restricted to these two Capparis species 


and comparison was made with Crataeva L. species 
(Rodman et al., 1998; Cummings et al.. 2003) 


Taxonomic confusion of Capparis sect. Capparis 18 
reflected in the number of combinations and changes 
of rank, with frequent placement under C. spinosa 
(Zohary, 1960: 1965: Higton & Akeroyd, 
1991). Capparis has become a blanket 


identification level of 


Jacobs, 
spinosa 
used to cover the scarce 
the taxa in Capparis sect. 
1984). This has both taxonomic and 


economic implications, as Capparis flower buds are 


definition of Capparis 


(Greuter et al., 
the erude material used as commercial capers and 
variability in this economic product is determined 
mostly by taxonomic 2001: 
2002). Molecular studies of Capparis 


differences (Inocencio, 
Inocencio et al., 
sect. Capparis have been scarce or restricted to small 
sampling (Fici, 2001). 

A genetic fingerprinting technique (AFLP) was 
the 


used by Inocencio et al. (2005) to examine 


relationships among Capparis species. Genetic dis- 


tances, based on AFLP data, were estimated for 45 
accessions of Capparis species from Spain, Morocco, 
and Syria. The results of this analysis support the 
differentiation of four (C. orientalis Veill.. C. sicula, C. 
aegyptia Lam., and C. ovata Desf.) of the five taxa 


sampled. The fifth, excluded, species was C. 


The group of plants recognized as C. spinosa on the 


spinosa. 


basis of morphological characters such as shrub 
procumbent or somewhat erect, stipules usually weak 
or vestigial, rarely strong, very long and thin (0.3— 
0.6 em long). indument on leaves always very lax, and 
300—500 um long), is found 


in cultivation. Several cultivars of 


— 


trichomes thick and long 
almost exclusively 1 
C. spinosa appear in an intermediate position between 


124 


Y 


C. orientalis. and C. sicula and overlap with C. 


orientalis. The other two species, C. aegyptia and C. 
ovata, are morphologically distinct, characterized by 
their habit, stipules, and leaves (C. 


aegyptia: shrubs 


somewhal erect; twigs gray-green or blue-green with 


waxy cover; stipules curved, retrorse: leaves obovate 


to ovale, bases and apices rounded, 2-3 X 1.83 em: 
C. ovata: shrub pendulous; adult leaves ovate, 
texture subcoriaceous; stipules curved, mostly an- 


trorse). 

The objective of the present revision is to provide 
a general taxonomy, based fundamentally on morpho- 
logical data for Capparis sect. Capparis. 


TAXONOMIC CONTEXT OF C IPPARIS SECT. CAPPARIS 


Capparis belongs to the subfamily Capparoideae 
(Capparaceae), that also includes Cadaba. Crataeva. 


Morisonia, Boscia, and other New World and Old 
World genera (Hall et al., 2002). The different genera 


overlap in molecular studies (Hall et al., 2002). 


although a marked biogeographical distinction is 
found between New World and Old World groups. 


recognized at the level of section or subsection 


(Candolle, 1824: Bentham € Hooker, 1862: Hall 
el al. 2002). A tentative infragenerie division is 


Table l in order to 
There 


taxonomic and nomenclatural i issues for infrastructure 


presented in contextualize 
Capparts sect. Capparis. are many unresolved 


| Capparis. Therefore taxa in Table l are merely 


presented as a synthesis of major groups. 
GEOGRAPHIC DISTRIBUTION AND. ENDEMISM 


Capparis is mostly a Pantropical genus, but section 
Capparis is almost strictly Holarctie with six exclusive 
(C. aegyptia, C. 
Obón & Alcaraz, 


ovata, C. 


species atlantica Inocencio. D. 


Rivera, sp. nov., C. orientalis, 


spinosa, C. zoharyi Inocencio, D. Rivera, 
Obón & Alcaraz, sp. nov.), one paleotropical species 
(C. hereroensis Schinz). and three holaretic and 
paleotropical areas (C. 


Boiss., C. 


calegories 


mucronifolia, C. 


parviflora 
sicula). Here we name the floristic 


(kingdoms and regions) 


Takhtajan (1986). 


according to 


Capparis sect. Capparis has its 


maximum diversity in the Mediterranean Region 
(Holarctic) with seven species (C. aegyptia, C. 


J 


atlantica, C. orientalis, C. ovata, C. spinosa, C. 


zoharyi, C. sicula). Wt is followed next by the Saharo- 
Arabian Region (also Holarctic) with six species in 
North Africa and the 


C. ovata, C. 


sicula). 


Arabian Peninsula (C. aegyptia, 
zoharyi, C. mucronifolia, C. parviflora, C. 
The Irano-Turanian Region (Holarctic) is 
inhabited by four species e 


West and Central Asia (C. 


xtending along most of 


aegyptia, C. mucronifolia, 


Annals of the 
Missouri Botanical Garden 


The 


is inhabited by three species 


C. parviflora, C. sicula). Sudano-Zambezian 
Region (Paleotropical) 
in tropical East Africa, littoral Arabia, western India. 
and southern Pakistan (C. mucronifolia, C. parviflora, 
C. sicula). Karoo-Namib 


regions are inhabited by one single species each (C. 


The African and Indian 


hereroensis and C. sicula. respectively). Capparis 


sicula is the most widespread species, stretching from 
the Western Mediterranean to the Himalayan Moun- 


ains and the Rajasthan of India (Indian Region. 
Paleotropical). Capparis hereroensis, in western Na- 
mibia, is the only species confined to the African 
1980). 


demic taxa are infrequent in this section, although 


Karoo-Namib Region (sensu Takhtajan, En- 
C. 
atlantica is endemic to the Atlas range in Morocco, C. 
mucronifolia Boiss. subsp. rosanoviana (B. Fedtsch.) 
Inocencio, D. Rivera, Obón & Alcaraz is endemic to 
central Tajikistan, and C. parviflora Boiss. subsp. 
sphaerocarpa Inocencio, D. Rivera, Obón & Alcaraz is 
endemic 


to the western. provinces (Nimruz, Farah, 


Herat) in Afghanistan. 
ECONOMIC IMPORTANCE 

Species of Capparis sect. Capparis are widely used 
in the Old World. 


spinosa is almost exclusively 


as food and medicine Capparis 


known in cultivation. 
The supposed wild individuals of C. spinosa are often 
remnants of 


— 


ancient caper fields or escaped from 


cultivation. Local caper cultivars and. ethnovarieties 
are recognized throughout the Western and central 
Mediterranean region (Spain, France, Italy, continen- 


tal and insular). 


EN 


These principally belong to 
spinosa, but also to C. 


1999, 


orientalis (Rivera et al.. 


2003b). Flower buds, consumed as brined product, are 
a rich source of the antioxidant phenolic compound 


rutin (Inocencio et al.. 2000). 
species by humans has been traced to the Prehistory 
and early historic times of Western Europe (1st-2nd 
AND 


region (900t 


The use of Capparis 


cent. Tongeren, Belgium). the Mediterranean 
)- 7400 B.C.. 
West Asia (9th-8th mill. 


as evidenced. by the 


Franchti Cave, Greece). and 
B.C., Tell Mureybit, Syria). 
presence of these species in 
archaeological sites (Rivera et al.. 2002). 


MATERIAL AND. METHODS 


The herbaria and se ol the following institu- 


tions were consulted: BM. E. JEPS. K. MA. MUB. 
OXF, P, RSA, and UMH. niina) on loan were 
received from BISH, BR C, G, HUB. HUJ. MARSSJ. 


RNG, and US. Special collection trips were conducted 
in North Africa. 
Spain to obtain fi 


western Asia, and Mediterranean 


esh material for molecular studies. 


for the in vivo study of floral characters, and for 


Table J. 


Tentative infrageneric division of Capparis presented in order to contextualize Capparis section Capparis. 


Infrageneric rank 


Characters 


sree species 


Distribution 


Sect. Breyniastrum (Plum.) DC. 


Sect. Calanthea DC. 


Sect. Capparidastrum DC. 


Sect. Capparis 


Sect. Corymbosae (DC.) Span. (— Subsect. 


Corymbosae DC., 


Radlk, p.p.) 


Sect. 


Sect. Cynophalla DC. 


Wonostichocalyx 


Branches unarmed. Leaves large (more than 0.3 cm wide) 


and persistent. Flowers in simple or compound 
inflorescences (racemes, umbels, corymbs), although 
solitary flowers at the leaf axil may appear in some 


individuals. Sepals triangular, sharp; gynophore always 


very short. Flowers with numerous stamens 


(> 50). Fruit rounded in cross section. 


Leaves large (more than 0.3 em wide) and persistent. 


Flowers in simple or compound inflorescences 
x). although solitary flowers 


(racemes, umbels, corym 
at the leaf axil may appear in some individuals. Sepals 
thin, linear, and pointed. Flowers with numerous 


stamens (> 50). Fruit 1 in cross section. 


Stems generally thorny. Flowers in simple or compound 


inflorescences (racemes, umbels, corymbs), although 
solitary flowers at the leaf axil may appear in some 
individuals. Leaves large (more than 0.3 em wide) and 
persistent. Sepals oval without gland. Flowers with 
numerous stamens (> 50). Gynophore short. Fruit 


rounded in cross section 


Flowers always solitary at leaf axils. Plants with flowers 


slightly zygomorphic, abaxial sepal not galeate or 
slightly galeate. Flowers with numerous stamens 
(> 50). 

Branches unarmed. Leaves large (more than 0.3 em wide) 
and persistent. Flowers in corymbose inflorescences. 
Sepals rounded; gynophore very short or large. Flowers 
with numerous stamens (> 50). Fruit rounded in cross 
section. 

Stems generally thorny. Flowers in simple or compound 
inflorescences (racemes, umbels, corymbs), although 
solitary flowers at the leaf axil may appear in some 
individuals. Leaves large (more than 0.3 cm jus and 


persistent, Sepals ovate possessing basal gland « 


— 


foveole. Flowers with numerous stamens (> 50 


Gynophore large. Fruit rounded in cross section. 


C. M rM L., C. incana HBK, 
C. indica (Ie) 1 & Rendle. 


C. nemorosa Jacq.. C. pulcherrima Jacq. 


brasiliana DC., C. macrophylla HBK, 


C. cuneata DC 


C. spinosa L. 


C. sepiaria L., C. umbellata Brown ex DC.. 


C. incanescens DC. 


C. cynophallophora L., C. 


C. amplissima Lam. 


guayaquilensis HBK, 


New World 


New World 


New World 


Old World 


Old World 


New World 


900c 


| JequinN SG euinjoA 


sueddeo oes sueddey jo uoisI^eH 


‘ye ja OlOUBDOU| 


Sol 


Table Il. Continued. 


Infrageneric rank 


Characters 


Representative species Distribution 


Sect. Galeatae Inocencio et al. 


Sect. Quadrella DC. 


Sect. Seriales 1 55 ) Spa 


0 5 bec alyx Radlk. 


Sect. Sodada (Forssk.) Benth. & Hook. 


Subsect. Octandrae DC. 


Subsect. Pedicellares DC. (= Sect. 


Vonostichocalyx EUN b. p., Sect. Busbeckia 


(Endl.) Benth. & Hook. p.p.) 


. (= Subsect. uias 


Flowers always solitary at leaf axils. Plants with flowers 


strongly zvgomor] hic. abaxial se pal strongly galeate. 


Flowers with numerous stamens (> 50). 


weaves large (more than 0.3 em wide) and persistent. 
Flowers in simple or compound inflorescences 
(racemes. umbels. corymbs), although solitary flowers 
at the leaf axil may appear in some 0 8 

F 


in cross section 


owers with numerous stamens (> 50). Fruit angular 

Branches unarmed. Leaves large (more than 0.3 em wide) 
and persistent. Inflorescences racemose or in series on 
the stems. Sepals rounded. Flowers with numerous 
stamens (> 50). Gynophore very short or large. Fruit 
rounded in cross Meng 

0.1—0.3 em. deciduous. Flowers in 


simple or ME inflorescences (racemes. umbels. 


Leaves thin, 0.3—2 


corymbs), although solitary flowers at the leaf axil may 
appear in some individuals. Flowers with numerous 
stamens (8-18). 

Branches unarmed. Leaves large (more than 0.3 em wide) 
and persistent, Flowers in simple or compound 
inflorescences (racemes, umbels. corymbs). although 
solitary flowers at the leaf axil may appear in some 
individuals. Sepals rounded. Flowers with 8 stamens. 
Gynophore very short or large. Fruit rounded in cross 
section. 

Branches unarmed. Flowers geminate or in bunches of 
three, rarely alone. Leaves large (more than 0.3 em 
wide) and persistent. Sepals rounded. Flowers with 


numerous stamens (> 50). in kee very short or 


arge. Fruit rounded in cross sectio 


Old World 


C. cartilaginea Decne. 


New World 


G. e Jacq., i: intermedia HBK, 


7. crotonoides H 


C. zeylanica L, C. acuminata Willd., Old World 
C. quadriflora DC., C. rotundifolia Roul.. 


C. brevispina DC. 


C. decidua (Forssk.) Edgew. Old World 


Old World 


C. racemulosa DC., C. oleoides Burch.. 


— 
=) 


C. coriacea. Bure 


C. varada L. Old World 


2. CANESCENS ; DL 


Es = DC., 


UepJer) jeoiuejog unossi|A 


9c. 


au Jo s¡euuy 


Volume 93, Number 1 Inocencio et al. 127 
2006 Revision of Capparis Sect. Capparis 


Table 2. List of morphological characters considered in the study. 


Character States 
. Plant habit Erect / procumbent / pendulous 

; Height Maximum length of the stems in meters 
3. Twig shape Tortuose / straigh 
4. Twig color Green / yellowish / reddish / waxy 
5. Internodes Length in millimeters 
6. Stipule shape Curved / somewhat curved / straight / setaceous 
7. Stipule orientation Spreading / retrorse / antrorse 
8. Stipule base Decurrent / not decurrent 
9, Stipule color Orange-vellow / orange / golden-yellow 

10. Leaf shape Rounded / ovate / lanceolate / oblong / elliptical / obeordate /obovate 

l. Base of the leaf Obtuse/ tapering / acute/ cordate 

12. Leaf apex Acute / rounded / obcordate / obtuse / truncate 

13. Mucro presence and length Long (1—1.5 mm) / small (0.5-1 mm) / very small (0.1—0.5 mm) / lacking 
14. Muero shape Straight / curved 

15. Leaf texture Herbaceous / fleshy 

16. Leaf veins prominence Prominent / not prominent 

7. Petiole length Very short (less than 0.5 em) / short (0.5-1.5 cm) / long (exceeding 1.5 em) 
18. Leaf abaxial indument Very dense / dense / dense to lax / lax / very lax 

19. Trichome thickness Thick (25-50 um) / thin (15-25 um) 
20. Trichome length Long (250-900 um) / short (50-250 Um) 

21. Fertile floral pedicel length Long (greater than 4 em) / short (1.54 em) 
E Fertile floral pedicel thickness Thick (over 1 mm) / slender (less than or equal to 1 mm) 

. Flower bud apex Veute / rounded 

2 Abaxial sepal Galeate / slightly galeate 
25. Flower symmetry Zygomorphic / somewhat zvgomorphic 
26. Number of stamens Numerous (100—150) / not so numerous (40—80) 
27. Anther length Very small a mm) / small (2-3 mm) / large (3 mm) 
28. Anther apex Rounded / a 
29, Fruit shape Rounded / 1 0 / obovate / oblong 
30. Pulp color Red / vellow 
31. Seed color Brown/ dark brown 
32. Seed size Length X width X depth in millimeters 
preparing exsiccatae. Voucher specimens were de- Distribution of species comprising several subspecies 
posited in MUB and UMH. is represented in Figures 2, 4, and 5. Habitat is 


Here we base species on morphological and described using the available data on herbarium 


— 


biogeographical features. The species represented labels and those reported in the protologue of each 
are more or less distinct, heterogeneous, and variable taxa. Phenology data are restricted to the flowering- 


morpho-physiological entities, the origin of which is — fruiting period according to the herbarium specimen 


associated with a particular environment and area in labels and the protologue. 

agreement with Vavilov (1931). Taxa are defined 1 

such a way that it is relatively easy to determine the TAxoxowic CHARACTERS IN. CAPPARIS SECT. CAPPARIS 
ascription of each specimen to either one or another. 

Therefore, extremely large and variable species were We have shown that molecular, phytochemical, and 
avoided. Discontinuities, both geographical and in vivo data are useful for understanding patterns of 
morphological (see vestiture, stipules, leaves, inflo- variation and, as such, were considered for the 
rescences, flowers, and fruits sections in this paper), populations present in Spain, Morocco, Syria, and 
are good markers, but hybridization and hybrid Lebanon (Inocencio, 2001; Inocencio et al., 2000, 


swarms have obscured the definition of species and 2002, 2005): however, these methods have not vet 


subspecies. been applied to the entire section. Therefore. we 
New taxa are represented in Figures 3, 6-9, selected characters generally available in herbarium 
Geographic distributions have been plotted exclu- specimens (Table 2), some of which have been used 


sively using the information from herbarium sheets. for the first time in this study. For example, anther tip 


Á "m 
Y. Pad, 
% S y 

7 0 E B 


— — 
- oa ſ— 
N T teet c km E 


e 


1 r >| 23 . me 
3 o az Ros 
— —-— Tel 
Figure l. Plant habit: —A. Erect, e.g.. Capparis zoharyi. Drawn from photo taken in Llano del Beal, Murcia, Spain. —B. Procumbent. e.g.. C. spinosa. Drawn from photo taken in Llubí. 
Mallorca. Spain. —C. Pendulous, e.g.. C. orientalis. Drawn from photo taken in Palma de Mallorca. Spain. 


80 


əy} Jo s¡euuy 


dope jeoiuejog unossiJA 


Volume 93, Number 1 
2006 


Inocencio et al. 
Revision of Capparis Sect. Capparis 


shape was recognized as a useful character for 


identifying living, herbarium, and processed material 
2002): 


(Inocencio et al., once characterized in fresh 


material, it is relatively easy to determine in 
herbarium specimens by re-hydrating the anthers 
and using a stereomicroscope. All the selected 


n malure specimens with 


characters were studied 
open flowers and ripe fruit. The quantitative char- 
acters are expressed using SÍ metric units. 


PLANT HABIT IN CAPPARIS SECT. CAPPARIS 


Capparis comprises small trees, shrubs, or geo- 
phytes. All species of section Capparis are erect, 


Y 


or pendulous shrubs (Fig. 1). Some (C. 


procumbent, 
sicula) behave as a true geophyte, with underground. 
branched perennial stems and decaying annual aerial 
sensu Bocquet & | Aeschiman, 


parts (géothamne, 


1981). 


VESTITURE IN CAPPARIS SECT. CAPPARIS 
Plants glabrous or with simple trichomes as the 


indument. Trichome types and the density of 
pubescence on various parts of the plant are useful 
characters for identifying some Capparis species. 
However, high infraspecific variation is common in 
collections identified as C. sicula, for example. More 
consistent vestiture is found on the abaxial part of the 
thickness, and length of the 


leaves, where density, 


indument (Table 1) may be analyzed for each taxon. 


STIPULES AND LEAVES IN CAPPARIS SECT. CAPPARIS 


Spiny stipules developing at the base of the petioles 
are absent in some species (e.g., Capparis orientalis) 
ovata, C. 


or are early shedding (e.g.. C. hereroensis, C. 


spinosa). Stipule shape. color. direction of curvature 
and decurrence at the base are distinctive characters 
(Table 1). 

Leaves are simple, 


petioles well differentiated. 


with 


a 


not divided, alternate. 
Leaf morphology fur- 
texture, base 
Table 2). 


The size and, much less, the shape of the leaves are 


nishes distinctive characters (shape, 


type, apex type. presence and type of mucro) 


variable within many species. Leaf size especially 
depends on water availability and exposure to winds 
and sun during the growing season. Petiole length. 
although variable within species, can be sorted into 
three groups (Table 2). 


INFLORESCENCES AND FLOWERS IN CAPPARIS SECT. CAPPARIS 


The flowers are solitary in the axil of the leaves. 


This is a distinctive character for the section. In other 


sections (except sect. Galeatae) the flowers form 


buds 


Flower 


Flowers 


corymbose or racemose inflorescences. 


may have acute or rounded apices. are 


bisexual and more or less zygomorphic. Four green 
sepals are always present, free, concave, having a more 
less galeate (helmet-shaped) abaxial sepal. Four 
white or pink petals are always present, free. oval, 
frequently unequal. From 50 to numerous stamens. 
Anthers are small to large (Table 2) with a rounded or 
sharp apex: androphore absent. Nectary situated 
the floral disk. between the insertion of the petals and 
sepals, triangular in form, apex directed toward the 
flower's interior. Nectary morphology can be a highly 
valuable taxonomic resource, but only available 
fresh or well-preserved material (Inocencio et al., 
2002). A gynophore is present, usually exceeding the 
stamens in length. The ovary is ellipsoid. situated at 
the end of the gynophore, unilocular, with (2 to)4(to 


10) placentas. 


FRUITS IN CAPPARIS SECT. CAPPARIS 


The fruit is an oblong, ovoid, ellipsoid, or globose 


berry, that is green in color with well-defined 
longitudinal nerves, along which dehiscence later 


Seeds 


generally brown in color when mature, 


occurs. are from one to numerous, and are 


immersed in 


a reddish or yellow pulp. Seed shape, color, and 
dimensions have been noted to be of limited 
taxonomic value (Rivera et al., 2002). 


TAXONOMIC TREATMENT 


Capparis L., Sp. pl: 503. 1753. TYPE: Capparis 


spinosa L. 

Voyage Bonite Bot. Atlas: tab. 5 
TYPE: Beautempsia cenit 
Bonite Bot. Atlas: tab. 56. 12 


e Gaudich., 

2 [1844—46]. 

ae h.. Voyage 

[1844—46]. 

Busbeckea Endl., Prodr. Fl. Ins. Norf. 64. 1833. TYPE: 

Jusbeckea nobilis Endl., Prodr. Fl. Ins. Norf. 64. 1833. 

Flora 22(1) (Beibl.): 25. 1839. TYPE: 
Flora 22 (1) (Beibl.): 


Colicodendron Mart., 
¿olicodendron yco Mart., 
183 39. 


Destrugesia Gaudich.. Voyage Bonite Allas: tab. 57. 


1842 [1844-1846]. TYPE sudan scabrida Gau- 
dich.. Voyage Bonite Bot. Atlas: tab. 57. 1842 [1844 
840]. 

Hombak Adans., Fam. 2: 402, 408. 1763. TYPE: Genus 


described referring to Lippi MS. Acar list of 
one collected in Egypt by I There are no 


species associated with the genus in the protologue. 


However, the inflorescence type, lack of leaves, and 
origin (Lippi, hence Egypt) likely 150 to Sodada 


Capparis aphylla Roth) 
109. s 


decidua Forssk. (= 
Oligloron Raf.. S 
Oligloron zeylanica Raf. 
Olofuton Raf., Sylva Telluriana: 108. 1838. TYPE: Olofuton 


racemosum Raf. 


ylva Telluriana: TYPE: 


130 


Annals of the 
Missouri Botanical Garden 


Pseudocroton Müll. Arg. Flora 55: 24. 1872. 

Pseudo M tinctorius Mii 

Quadiella (DC.) J. S. Presl, 

"nirozenosti: Rostlin 2: 260. 

MER m JS; >. P resl. 

nth. Nov. Gen. Sp. 5: 95. 1 

11 9 I 'orssk., \egypt.- res 8 
decidua Vor e 


TYPE: 


1825. TYPE: 7 
(= Capparis crotonoides 
321). 

. 1775. TYPE: Sodada 
Nov. Hort. 


Uterveria frondosa Bertol. 


Uterveria Beten PL Bonon. 2: 8. 1839. TYPE: 


Capparis section Capparis TYPE: Capparis spinosa L. 


A total of 10 species and 12 subspecies of economic 


relevance are recognized, distributed in the tropical, 
subtropical, and Mediterranean zones of both hemi- 
spheres. Widely represented in Asia. and reaching 
southern Europe, eastern, northern, and southwestern 


Africa. A key for the 


nothospecies of 


species and recognized 
Capparis sect. Capparis is presented 


here. 


KEY TO THE SPECIES OF CAPPARIS SECT. 
AND AFRICA 


CAPPARISAN ASIA, EUROPE. 


la. Plants. unarmed, or with slipules vestigial or 


North 


Croatia, Greece, lta 


Europe and 


Algeria, 
‘urk 


Malta, Spain, Turkey] .. ... 5. C. orientalis 


Ib. Plants always spiny oo 2 
2a. 9 retrorse to horizontally oriented . . . . . . . . 3 
Za. Leaf texture somewhat fleshy . . . . . . . .. . . . . . . . | 


à Y Or gray-green, nol elaucous. without 
bloom; 
Middle East. [ Afghanistan, 
Tajikistan, United Arab Emirates] 


stipules golden yellow, not de- 


current. Iran, Oman, 

. C. mucronifolia 
Ib. Twigs green to rede purple. elaucous due to 
a Waxy COV ering: stipules orange, decurrent „ A 
Ha. Shrubs erect: stipules strongly decurrent. Medi- 
le rranean Europe, North Africa, Middle I 
1 „Algeria. Israel, Jordan. 


b Tac, Morocco, Spain, Syria, Turke 


sast into 
Egypt. Greece, 
C. zoharyi 


5b. Shrub procumbent, or somewhat erect or pendu- 


lous: stipules decurrent to not decurrent 


6a. Shrub pendulous; leaves lance 5 lo. ovale- 
lanceolate. North Africa [ Algeria, Chad, Libya. 
Morocco, Tunisia] . . . . . . . . . . . . . . .. 6. C. ovata 
Ob. Shrub procumbent or somewhat erect: leaves 


rounded or obovate to ovate 


la. Shrub procumbent: twigs purple-red; without 


waxy bloom: leaves rounded, tip acute. base 
rounded to cordate. North Africa [Morocco 9 
CCC 2. C. 1 
7b. Shrubs somewhat erect: twigs gray-green or 
glaucous, with waxy bloom: ie saves obovate to 
ovale, bases and s rounded. North Aidan, 
Middle las! into India [Egypt India. Israel. 
Jordan, Saudi Arabia] . . . . . . . . . .. l. C. aegyptia 
3b. Leaf texture herbaceous . . . . . . . . . . . . . . . . . .. 8 
8a. Flowers F abaxial sepal helmet- 
shaped, 1.7-2.5 cm long. 0.7-1.2 em deep . .. . .. 9 


Ha. Shrub procumbent: stipules usually stout: pubes- 


cence or leaves from lax lo very dense (rarely 


in e & J. resl., 


very lax). Me s Europe. North Africa, 
Middle East into. India Ansa Albania, 
Algeria. Azerbaijan, Cyprus, Georgia, Greece, 


India, Iran, Iraq, Israel, Haly, ea, Kazakhstan, 


Mongolia, Morocco, Pakistan, Saudi Arabia. 
pain, Syr Purkey, Turkmenistan. Ukraine. 
zbekistan, Vemen] .. . . . . . . . . . . . . 8. C. sicula 


Ob. Shrub procumbent or somewhat erect: stipules 
i; indument 
‘Mediter rranean Euroj 

Middle Fast into Turkey [France. Greece. Italy. 


Turkey | 


spain; Turkey), 4.4 6a haw eR 9. C. spinosa 
Ob. Flowers slightly zygomorphie; abaxial | sepal 

slightly galeate only to 1.4 em long, to 0.6 cm 

deep. Middle East [Afghanistan, Iran, Iraq. 


Arabia. 


Pakistan, Saudi Turkmenistan 


2b. Stipules mostly antrorse in orientation 


1 Oa. 


mature leaves ovale: 
North 


[ Algeria. Chad, Libya, Morocco, Tunisia] 


Shrub pendulous: texture 


Africa 


subcoriaceous: stipules e urved, 


6. C. ovata 


10b. Shrub erect: mature leaves oblong to obovate: 


A 


se "a eous. 
3. C. hereroensis 


t coriaceous; stipules 
South Africa [Namibia] 


Method. Bot. 
605. 1783. Capparis spinosa var. aegyptia 1 
Boiss., Fl. 120. 1867. TYPE: [Egypt] 


"Lipi a observé ce Caprier en Egypte (v.s. in herb 


l. Capparis aegyptia Lam.. Encycl. 


orient. l: 


Su.)“ (lectotype. designated here, P-JU!: speci- 
men on sheet 11.248 right). 

in Duhamel, Traité Arbr. Arbust. Ed. 
. 1801. TYPE: [Egypt] “Habite. L'Arabie 


Mont Sinai et. les montagnes 


Capparis sinaica h ill., 
2. Vol. I: 14 
pétrée. sur y" qui 

environment, auprès du village du Pharagou. et sur 

les coe: entre le Mont Sinai et le chateau ou la ville 


eae awa et Shaw’ ye t. 112 in Shaw. 
Pl. s Ali 7 381, designated by Rivera et al. 

€ 3a: a 8.300. 
Capparis deserti (Zohary) e & Boulos, Publ. Cairo 
Univer. Herb., 5: 14. 1972 [197 oe m Ls 
var. zd died Eee "di Res. Coun. Israel 8D: 56. 
1960. TYPE: [Israel] Wadi a entrance lo ehe 


1 15 be rheba). 27 Mar. 
Feinbrun 5076-7 (holotype. 


1944, P. H. Davis & N. 
J!) 


— 
posl 


Shrub somewhat erect, glabrous: twigs straight, up 
to 3 m long. gray or blue green due to waxy cover that 
appears over the entire plant; internodes 1.5—4 em: 
stipules curved, retrorse, slightly decurrent, orange. 
0.3-0.4 cm long. 0.1-0.2 em wide at the base. Leaves 
3 * 
leaf veins not prominent; bases and apices rounded: 
0.1—0.5 


Flower 


obovate to ovate. 2— 1.8-3 cm, somewhat fleshy: 


mucro very small, curved: 
floral 


2.5— em: flowers slightly zvgomor- 


mm. straight or 


petioles 0.5-1 em. buds rounded: 


pedicels stout. 


phic: abaxial (odd) sepal slightly galeate at apex, I. 4— 
1.6 cm long. 0.6-0.8 em deep; stamens 30 to 80, 
anthers 1.2-1.5 mm. with rounded apices. Fruit 


Volume 93, Number 1 
6 


Inocencio et al. 131 


Revision of Capparis Sect. Capparis 


259. 


— 
5° 20° 
Figure 2. 


Capparis aegyptia also occurs in India, beyond map range. 


oblong, pulp color unknown; ripe seeds dark brown. 


* 2.8-3 X 1.8-2 mm. 
Illustrations. Plate 31(3) in Delile, 1812. Zohary 
(1960: 52. fig. I). Tückholm. (1974: 163, pl. 48c). 


Migahid (1989: 
Phenology. 

March to August (December). 

Saharo-Arabian, extend- 


48, pl. 26). 


Flowering and fruiting from 


— 


January 


Distribution and habitat. 


into the Irano-Turanian and Mediterranean 


ing 
Regions. North Africa, Middle East into India 
(Egypt. India, Israel, Jordan, Saudi Arabia]. Rocky 


places, steep slopes, at elevations from 0 to 2000 m, 


often in the vicinity of human dwellings. It is the 
common caper in Egypt and is often associated. with 
the Hyparrhenia hirta (L.) Stapf community. Figure 2. 

The type specimen of Capparis aegyptia has one 
flower, no fruits. It is part of the herbarium A. T. 
which is cited by Lamarck (1783): 
herb(ario) Isn(ard).” H 
in the Jussieu herbarium (P-JU). 


Danty d'Isnard, 
“v(idit) 


included in 1857 


s(iccam) in 
The specimen was collected “en face de Minia" in 
Egypt by one botanist who gave it to Isnard: "Doni per 
tu” [Gift for you]. It was, presumably, D. Lippi (1078— 
1704) himself. 
plants—now at P-JU—collected by Lippi in Egypt 
P 1872). 


The basionym of Capparis deserti (Zohary) Tückh. & 


during his last trip in 1704 (Pritzel, 


Boulos, Capparis spinosa L. var. deserti Zohary, was 


validly published by Zohary. However he states “This 


is a ‘weak’ variety because small-leaved forms occur 


T 


almost in all groups. 


Distribution map for Capparis ovata (3&) (both subspecies): Capparis spinosa (0); and Capparis aegyptia ($ 


Isnard wrote the manuscript list of 


— 


Selected specimens examined. EGYPT. Cairo, Schwein- 
furth 995 (K); Gebel Ez Zebir, Sinai. Tadmor S-417 (K): 
Minia. Lippi ? (P); Tadmor & Shmida 5-120 = ): Pie: Bové 
273 (K). INDIA. Bom Bomb: s.n. (E). 
. Stait 266 (RNG 5 y adi n It. pe 4888 


A C 
D 
As 
N 


Mujib, Madaba, Boulos 5856 q. SAUDI ARABIA. Wadi 
Lakus, Jebel, Collenette 7228 


2. Capparis atlantica Inocencio, D. Rivera, Obón & 
Alcaraz, sp. nov. TYPE: [Morocco] “Safi, 20-6 
1999. Inocencio 60026" (holotype, UMH': 

MO!, K!, E!, MA!). Figure 3. 


130- 
pe 8. 


Suffrutex decumbens, usque ad 60 em, caulibus purpure- 
l. 


foliis pans apice aculis, base rotunc is, 1.5-3 cm 
—2.5 cn ! 


longis .2-2.5 c latis, a C. stipulis 
te muor bus. al ichu foliorum acutis Dua und nec 


cordatis a C. 1 et C. aegyptia diffe 


Shrub procumbent, up to 60 cm high, glabrous: 
2m 
sometimes green; internodes 1-3 cm; stipules curved, 
0.3-0.6 em long, 
1.5-3 


not 


twigs straight, up to long, reddish purple, 
retrorse, slightly decurrent, 
0.2-0.3 cm wide at the base. 
X 1.2-2.5 cm, 
bases rounded to cordate, 
mucro 0.1-0.5 mm, 
short, 0.3-1 cm. Flower buds rounded; floral pedicels 


orange, 
Leaves rounded, 


somewhat fleshy; leaf veins 


prominent; apices acute: 


very small, straight; petioles 


slender, short, 2-3 cm; flowers slightly zygomorphic: 


abaxial (odd) sepal slightly galeate, 1.4-1.6 cm long. 


— 


0.5-0.6 cm deep: stamens 30 to 80, anthers very 


small, 1.8-2 mm, apices rounded. Fruit oblong. pulp 


132 Annals of the 
Missouri Botanical Garden 


— x i 

10 cm 

Figure 3. Details of the new species Capparis atlantica Inocencio, D. Rivera, Obón & F. Alcaraz. —A. Stem and flower. 
—B. Detail of leaf. (X. B drawn by J.-A. Barreña from /nocencio 70100, UMH.) 


— 
Perm 


color unknown; ripe seeds brown, 2.1—2.6 X 2.2-2.4 — elevations from 0 to 2000 m. often in the vicinity of 
human dwellings. In the surroundings of Moulay 


X 1.6-1.8 mm. 
Brahim (Morocco), it grows on clayey slopes along 


Phenology. Flowering and fruiting from May to with Ephedra L. sp.. Teucrium fruticans L., Lavandula 
August. L. sp.. Chamaerops humilis (J. C. Archibald. 4586, E). 
Distribution and habitat. Mediterranean Region. Figure 4. 


North Africa [Morocco]. Ilt. is the common caper Capparis atlantica is a procumbent shrub with 


species in the High Atlas Mountains of Morocco. 


purple-red twigs; C. aegyptia is somewhat erect with 
gray-green or glaucous twigs; and C. zoharyi is erect. 


LG 


Found in rocky places, slopes, on calcareous substrata 


or marls, occasional on metamorphic substrata; at The stipules are not so strongly decurrent i 


Volume 93, Number 1 
2006 


Inocencio et al. 
Revision of Capparis Sect. Capparis 


133 


Figure 4. 


atlantica as in C. zoharyi; the leaf tips of C. atlantica 
are acute, while those of C. aegyptia and C. zoharyi 


are rounded. 


MOROCCO. Asni., Inocencio 60005 (UMH): 
falls of Ouzoud, Jury 8764 (RNG): na 
(RNG): Moulay 117 Archibald 1586 (E): 

Inocencio 60003 (UN Siksoua. Greater e Joker s.n. 
(K): Tafraoute. s.n. d (RNG): Tizi n'est. pnus 60034 


(UMH). 


Paratypes. 


3. Capparis hereroensis Schinz, Bull. Herb. Bois- 
sier 1(3): 396. 1895. TYPE: [Southwest Africa = 
Namibia] “Südwestafrika: Zwischen Wortel und 


Walfischbai, October 1886, Schinz 10067 (lec- 
totype, designated here, Z). 

Shrub erect, almost glabrous; twigs d up to 

3 m long, yellowish green: internodes 1.3-2.5 cm: 

stipules antrorse, mostly setaceous, not T 


sometimes falling or weak, yellow, 0.2-0.3 cm long. 


0.05-0.1 em wide at the base. Leaves oblong-obovate. 


234.2 X 1-1.6 em, somewhat fleshy, yellowish 
green; leaf veins prominent: bases tapering lo 


rounded, apices rounded to somewhat truncate; muero 
small, 0.5-1.0 mm, straight: petioles 0.5-1 cm. od 
er buds rounded: floral pedicels thick and short, 2.5— 
4.5 em: flowers slightly zygomorphic; abaxial «e 
sepal slightly galeate, 1.6-1.8 em long, 0.6-0.8 cm 
deep: stamens 30 to 80, anthers 2.5-3 mm, with acute 
apices. Fruit ellipsoidal, pulp color unknown: ripe 
3.84 X 3.6-3.8 X 2.7-3 mm. 


[om 


seeds dark brown, 


Phenology. Flowering from January to April. 


Distribution map for Capparis atlantica (Y Capparis orientalis (, and Capparis zoharyi (4). 


Distribution and habitat. Karoo-Namib Region. 
South Africa [Namibia]. Dune-forming sandy 
substrate, in coastal zones: at elevations from 0 to 


100 m. 
The Schinz herbarium is at Z. At Z there are two 
material of Capparis hereroensis 


sheets with 


because holotype is not designated in the protologue. 


type 


The specimen accompanied by a single handwriting 
following text, “Capparis 
Walfischbai/18860/2 Bo- 


is here designated as the lectotype: 


sheet with the 

hereroensis Schinz/Wortel, 
gen/N 1006." 
the specimen on the other sheet, with identical Schinz 


on the 


collection number, is a paralectotype. 


Selected. specimens examined. NAMIBIA. Walvis Bay 
Town. Salworks, Ward 9250 (K); Hereroland, Walfisc 
Schinz 1000 (Z); Conception Hut, Ward & Ward 158 (K). 


:hbai, 


Diagn. Pl. Ori- 
3. 1843. Capparis spinosa L. 


4. Capparis mucronifolia Boiss., 
ent., Ser. I, Vol. I: 
var. mucronifolia (Boiss.) Hedge & Lamond, Fl. 


Ir. Hoch. Umr. Geb.: 7. 1970. TYPE: [Iran or 
Oman] “Hab. in Persia Australi et Regno 
Mascatensi, Aucher 4189..." (lectotype. desig- 


nated here. G). 


KEY TO THE SUBSPECIES OF CAPPARIS MUCRONIFOLIA IN THE MIDDLE 
EAST AND CENTRAL. ÁSIA 
stipules curved, retrorse: 


la. Twigs slightly tortuose; 


leaves ovale, pube scence very lax: flower buds 


e 
Ah Apen a. C. mucronifolia subsp. mini 


Twigs straight: stipules almost straight or some- 
[el o o 
what curved, retrorse or often spreading: leaves 


134 


Annals of the 
Missouri Botanical Garden 


ii 20° 


45" 


Figure 5. Distribution map for Capparis parviflora (3%) (three subspecies): Capparts sicula (@) (all five subspecies): and 


Capparis mucrontfolia (4) (both subspecies). 


oblong-lanceolate. pube scence lax: flower buds 


rounded . . . . . b. C. ada subsp. rosanoriana 


Distribution. Figure 5. 
da. Capparis mucronifolia Boiss. subsp. muero- 


nifolia 


Hausskn. & 

maskatensis Hausskn. & Bornm. ex oe Mitt. Thür. 

Bot. er. N.F. VE 49. 1894, TYPE: [Oman] "foliis 

rupibus ad 11 5 - Bornmüller| 
JE 


type, "de ssignated here, 


Capparis E Bornm. ex Bornm. var. 


latioribus ovatis: i 
ex. 46” (lect 


Shrub somewhat erect, heavily branched, irregu- 


arly and widely spreading. up to ] m high: twigs 
slightly tortuose. approximately 2 m long, yellowish or 
grayish green; internodes 0.5-3 em; stipules curved, 
retrorse, not decurrent, golden yellow, apex orange. 
basally, 0.2-0.6 cm 
2-4 


sometimes pubescent, al least 


long, 0.1—0.2 em wide at the base. Leaves ovate, 
X 0.5-1.5 cm, somewhat fleshy: indument very lax. 


trichomes thick and short to long. (20)30—10 Xx 200- 


— 


400 um: leaf veins not prominent; bases usually 
rounded. sometimes tapering. apices acute: mucro 


small, 0.5—1.0 mm. straight; petioles very short, 0.2— 
0.4 em. Flower buds acute: floral. pedicels slender. 
short, 2.5—3.5 em: flowers slightly | zvgomorphic: 
abaxial (odd) sepal slightly galeate, L.4—1.6 em long. 

10.9 em deep: stamens 30 to 80. anthers 1.3— 
1.5 mm, with round apices. Fruit oblong, pulp color 
unknown; ripe seeds brown, 2.1—2.8 X 1.6-2 X 1.6- 


1.8 mm. 


Phenology. 


Distribution and habitat. 


Flowering from March to September. 
Sudano-Zambezian and 


Saharo-Arabian, extending to the Trano-Turanian 


Regions. Middle East Afghanistan, Iran. Oman. 


United Arab Emirates]. ravines and 


slony plains of deserts. wadies 


Rocky slopes. 
| Acacia seyal Del. 
hammada. at elevations from 0 to 1000 m. 

There is no designation of holotype for Capparis 
mucronifolia by Boissier: the type material cited 
comprises different specimens collected by P. M. R. 
Aucher Eloy in southern Iran and Oman (former 
kingdom of Muscat). with collection numbers 4189, 
1190. 1192. 


Therefore lectotypification is needed. The 11 5 with 


and during his travels 835-1838. 


the lectotype contains the following labels: | “Cap- 


paris. mucronifolia / Boissier”. II " Aucher-Eloy- 
Herbier d'Orient N. 4189.” Paratypes are Aucher 
1190 and 4192, G! Isolectoypes are in K! 


bs elliptica Hausskn. & Bornm. ex Bornm. 


pl. exs. 


Thür. Bot. Ver. N.F. VI: 40. 1804. is a later 
enl " Capparis elliptica Span. ex F. Muell. 
Fragmenta Phytographiae Australiae 9: 172. 1875: it 


is within the range of Capparis mucronifolia subsp. 
mucrontfolia. 

There is no designation of holotype for Capparis 
elliptica var maskatensis Hausskn. & Bornm.: the type 
malerial cited comprises different specimens collect- 
his “Her 
Haussknecht's 
1904. as 
curator in JE, but sold his herbarium to B. Most of the 


ed. under number 46, by J. Bornmiiller in 
1892-1893.” 


Bornmiiller 


Persico—turcicum 
barium is at JE: worked. from 
Bornmúller Capparidaceae material at B 
stroyed during the Second World War. One wa 
number 46 is at JE (J. 


Therefore mota Is 


specimen with a 


Müller, 


needed and EE 


pers. comm. 
The sheet with the lectotype 
contains the following labels: I “Isotypus/ Capparis 
Hausskn. &  Bornm/ var. 
& Bornm.” II “J. 


elliptica maskatensis 


Hausskn. Bornmüller: Iter Persico- 


Volume 93, Number 1 


Inocencio et a 135 


Revision of l Sect. Capparis 


turcicum/ 1982-93/ No 46/ Capparis spinosa L. v./ 
Maskatensis Hsk. & Bornm./ determ.:/ Arabia aus- 
in saxosis ad Maskat/ 1893. 25-5 legit: J. 


Bornmüller." 


tralis: 


elected specimens examined. AFGHANISTAN, Griffith, 
Leman 3 11 (K). IRAN. Bangar Langeh, Davis & Bokhari, D. 
56176 ; Chahbahar, Baluc histan, Runemar 22417 (E) 
Mo & Hewer 263 (K); Hormuz, J. 
K). OMAN. Istal, Wadi Bani Kharus, 
Radcliffe- Smith 4045 (K); Ruwi near Muscat, Miller 6003 
(E). UNITED ARAB EMIRATES. Hatta, Dubai, Western 267 
(E); Fujairah, Lumley 38 (K) 


4b. Capparis mucronifolia Boiss. subsp. rosanovi- 
ana (B. Fedtsch.) Inocencio, D. Rivera, Obón & 
Al 


rosanoviana B. Fedtsch., 


araz, comb. et stat. nov. Basionym: Capparis 
Beih. Bot. Centralbl. 
20: 297. 1906. Capparis rosanoviana B. Fedtsch., 
Consp. Fl. Turk. Vol. II: 98. 1909. TYPE: 
[Tajikistan] “Ost-Buchara: am östlichen Ab- 
Berge Aryktau, hóher als Goranty, 
auf Felsen, 1900-2000”, am 3. [= 15.] April 
1883 (A. Regel!).” (lectotype, designated here. 
LE). 


^ 


hange der 


Shrub somewhat erect, well branched from base, 
irregularly and widely spreading, up to 1 m high: 
twigs straight, erect or decumbent, up to 2 m long, 
1-2 em; 
stipules almost straight or somewhat curved, retrorse, 
0.2— 
0.1-0.2 cm wide at the base. Leaves 
1.5-2 0.4—2 cm, somewhat 
fleshy; indument very lax, trichomes thick and 
long, 20-30 X 200—400 um; leaf veins not prominent; 


acute; 


yellowish or grayish green; internodes 


often spreading, not decurrent, yellow-golden, 


0.4 em long, 


Rem 


oblong-lanc eolate, 


mucro long, l— 
0.1—0.3 cm. 


Flower buds rounded; usually with indument at least 


base usually rounded, apices 


1.5 mm, straight; petioles very short, 
at the base; floral pedicels slender, short, 2-3 em; 
flower shape unknown. Fruit unknown; ripe seeds 


unknown. 


air Bobrov (1939: 7, tab. 1, fig. 2: 1970: 
8, tab. 1, fig. 2) 
noo. Flowering and fruiting from July to 


August (Bobroy 1939: 3; 1970: : 
Distribution and habitat. 2 N Region. 
Middle East [Tajikistan]. At the eastern slope of the 


mountains Aryktau, above Goranty and on the left 


— 


bank of the River Vakhsh, between Kurgan-Tyube anc 
Lechman. Limestone, sunny rocky slopes, and sandy 
substrata; from 600 to 1500 m. Also in the Kafimigan, 
Pani, and Amudarya river valleys, in southwestern 
Tajikistan (Anonymous, 2005). 

LE), the fresh 


flowers display a corolla with yellowish tints, and the 


According to Hegel (in sched., 


staminal filaments are reddish. The dried fruits are 


oblong-lanceolate, with marked longitudinal nerves 
(Fedtschenko € Fedtschenko, 1906). 

There is no designation of holotype by B. 
Fedtschenko; the type materials cited are different 
specimens collected by E. A. von Regel in south- 
western Tajikistan (former Bukhara region) during his 
travels in 1883. Therefore lectotypification is needed. 
Two sheets have been found by Vladimir Dorofeev in 
The sheet with the 
*A. Regel, 


montium 


LE containing type material. 
lectotype contains the following label: 1 
Orient. 
Aryktau [am östlichen Abhange der Berge Aryktau|/ 
supra Horanty [höher als Goranty] 1200—20007/ 3—15/ 
VII 1883." The sheet with the paralectotype 
contains the labels: 1 1883/ 14/ Sand- 
strecken zwischen Kurgantüfe/ und Lechman, 1200— 
13007 (links Wachschufer).” II “A. 
Turkestanicum/ Sandstrecken zwischen Kur-/ gantüfe 


1200-13007 (links Wachschufer)” 


Iter Turkestanicum/ In decliv. 


"August 
Regel, Iter 


und Lechman 
(LE!). 
TAJIKISTAN. 


between Kurgan Tyube and Lechman, Regel s.n. 


Selected. specimens examined. Kadajian, 


s.n. 277 (E) 


(LE): eastern slopes of Aryktau mountain, Regel s.n. (LE). 
5. Capparis orientalis Veill., in Duhamel, Traité 
Arbr. Arbust. ed. 2, 1: 142. 1801. Capparis 


spinosa L. subsp. orientalis (Veill.) Jafri, Flora of 
Vol. XII: 3. 1977. TYPE: “Capparis non 
spinosa fructu majore c Bauh. Pin. 480.. 
(lectotype, designated by Rivera et al. (2006), de 
image in J. Bauhin, J. Cherler & D. Chabrey, 
Hist. Plant. Vol. II: 63. 1651!; epitype, desig- 
nated by Rivera et al. (2006), [Greece] Sokastro, 
Dodecanese, Th. Raus 8382, E!). 


Libya, 


Prodr. Vol. I: . 1809. 
Em) m 
18. C. spinosa LE Warn 


Capparis owe Sm., Fl. Graec. I 
C 


subsp. 


a 


aris spinosa L. rupestris 
J. Eur.: 


E: [Greece, Crete] “In Creta et 
dos aro insulis, ad rupes" (lectotype, designated here, 
OXFI). 


Shrub pendulous or decumbent, sometimes reach- 
ing great dimensions in shaded sites, glabrous; twigs 
straight, dark green, sometimes reaching more than 
3m long; internodes 2-7 cm; plants unarmed, 
occasionally vestigial stipules are early falling. Leaves 


38.5 X 2.38 cm, 


base 


rounded, or somewhat ovate, 


somewhat fleshy; leaf veins not prominent; 
rounded, sometimes cordate, apices obtuse, some- 
times obcordate, rarely acute; mucro lacking or very 
small, 0.1—0.5 mm; petioles long, 1-2.5 cm. Flower 
buds rounded, rarely acute; floral pedicels thick and 
long, 5-8 em; flowers slightly zygomorphic; abaxial 


(odd) sepal slightly galeate, 1—1.5 cm long, 0.6- 


136 


Annals of the 
Missouri Botanical Garden 


0.8 em deep; stamens 30 to 80, anthers 2-3 mm, with 


round apices. Fruit ellipsoidal, with apices nipple- 


shaped, Pn red; ripe seeds dark brown. 3.2-3.6 X 
2.8-3 X 2.8-3 mi 

Illustrations. Figure | Ali & Jafri (1977: 5). 
Guerau & Torres (1981: 31). 

Phenology. Flowering from May to October. 


Distribution and habitat. Mediterranean. Region. 
Mediterranean Europe and North Africa [Albania 
Algeria, Croatia, Greece, Italy, Libya, Malta, Spain. 
Turkey]. Rocks, cliffs, old 


elevations from 0 to 600 m, often surrounding human 


walls of buildings, at 


dwellings. Figure 
Capparis orientalis was described by Veillard 
(1801: 142). 


based on the pre-Linnaean literature. He 


Duhamel apparently fundamentally 
gives three 
synonyms, listing the authors and references where 


these have appeared previously. Also, further material 
is cited in pages 142-143, after the discussion of the 
species. The references to the habitat and distribution 


are restricted to: “Les rochers de l'isle de Créte et des 
isles de l'Archipel. particulièrement celle d'Anti- 


aros: en Syrie et dans la Palestine.” Both localities 
| à 


and bibliographical references—including illustra- 
tions—furnish fundamental elements for determining 
the original material of Veillard. The author himself 
places different values on the material used. Part of 
the material is only indirectly cited in the discussion 
several additional 


alter the protologue. in which 
authors, localities. and icon are cited. The new 
species is not illustrated in the 


Veillard’s 
apparently part of the missing Duhamel herbarium 
(Stafleu & 1976). 
materials available for typification are the two images 
Veillard. The 


unillustrated. Lonicer's 


original paper. 
herbarium is unknown because it was 


Cowan, Thus the only original 
references are 
1679: 1006) 
may be discarded, as Veillard himself noted the poor 


cited. by remaining 


Lonicer, 


icon 


quality of this drawing. However, it clearly repre- 


sents an unarmed caper bush with rounded leaves. 


(165 


by Veillard in the pre-Linnaean synonymy of 


The illustration in Bauhin et al. 


1: 63) is cited 
C. 
orientalis and was selected as lectotype by Rivera etal. 
(2000). 


\ccording to 


— 


Veillard in Duhamel (Ii 


Palestine 


301). this 
and Syria by 
We 


found any plant material of this species from the 


species was collected in 


travelers such as Pockocke and Shaw. have not 


Levant, neither from these collectors nor others. 


However, it is likely that the species occurred in the 
area. 


There is no designation of holotype for Capparis 


rupestris Sm. Serena Marner (pers. comm.) did find 
a single specimen. (one blossoming branch and 


sn. (RNG). 


a branch with a young fruit) of Capparis rupestris in 
one sheet of the Sibthorpian Herbarium at OXF 


Prodr. Vol. | 


which is also associated with 


corresponding to the Fl. ( 355. 
809, number | 190, 
Fl. Graec. T. 48 


a very typical one from the Sibthorpian Herbarium. 


Graec 
The sheet of Capparis rupestris is 
The cited illustration [Tab. 487] was later published 


(omit. 1825): 


"Folia quam in praecedente crassiora et 


however, it is also original material. 
succo pleniora. Olivier" is certainly Olivier (1801— 
Olivier was a 
Turkey, 


and other Middle Eastern countries between 1792 and 


1507). Guillaume Antoine French 


naturalist who traveled in Greece. the Levant, 
1798. Therefore, lectotypification is needed. We have 
seen the sheet (OXF) with the lectotype. 


contains 


which 


neither references. to localities 


lectors. However it is clearly labelled "C 
Sib.” [written in the hand of J. E. 
Sibthorp, M.D.” [None of the 
written directly on the sheets in 
Sibthorpian Herbarium at OXF.] The 
Ivpical of i 


material 
localities 
specimen is 
specimen collected on the first voyage 
1787 (Serena Marner. 
in agreement with the description in the protologue. 
The 


with reference to the 


in 1786 or 


pers. comm.). [t is 


sheet contains further stamp and annotation 


Prodromus Flora Graeca [added 


> 


by M. Lawson, Sherardian Professor of Botany at 
Oxford (Lack, 1997)]. the 


reasonably original material and is designated here as 


Therefore, specimen is 


a lectotype. 


Selected A examined, ALBANIA. Berirk Saran- 
da, Krendl s.n. : Sarandë, Alston 2248 (K). Al pe *. 
ape Carbon, 19255 Davis 529601 (E). CROATIA. Cavtat, 


Lapad, Dubrovnik. Larsen s.n. (C) 


(RNG). GREECE. Ar golis. Tol m. Bowen 
0507 (RNG) Korgadhos: Islet Sokastro. Davis 8382 (E); 
Sokastro. Dodecanese. Th. Raus 8382 (Eh Kiklades 
Andros Jury 271 (RNG): Corfu, Markos. Dai 545. 
E): Gerolimenas, 28379 (C); Island Zakynthos. Bomam 
728 (C); Port Kheli, Lewis 642 (K); Rhodes, Ahaussen s.n 
(C). ITALY. t 1 


Florence, s.n. (C); Palermo, R. Coll 3 (E): 
ani 429 


Perugia, Ransone 410 (E); Pisa. Savi 429 (K); Rome 
Larsen 4874 (C); Sicily, Taormina, Island Bella Ostenfeld 
«n. (C); Messina, S. P. Brookes 5754 (RNG): Favignana 
Cape Calarossa. J. R. Akeroyd 545 (RNG): Island Leranzo. 
Davis 40174 (RNG). LIBYA. Coefía. N of Benghazi, Davis 
20177 (E): Gebel Ne ‘foussa, een 19670 (E): Kouf. J. H. H. / 


26 (K): Shahat. Ali 624 (E). MALTA. Addaloute 
Lanfranco 5967 11 s Wright s.n. (K). SPAIN. 
Alicante: Gram 2039 (C); Santa Barbara Castle. Inocencio & 
10 araz 00048 10 B). Barcelona: Barcelona. Sennen 1587 
Ibiza: Santa Eulalia, 3250 (RNG). Mal- 
Mallorca. Inocencio & Alcaraz 48696 

€ 

€ 


Inocencio & Alcaraz 


Cemetery, 


zy 


Cannon 
Palma de 
\leudia. 
Menorea: Punta Nati, 
Benifayet, 
G70I (MUB). Valencia: Lliria, 
(MUB). TURKEY. Mug 
Prance 116 (E). 


Valldemosa. Jacobsen s.n. (C). 


‘ar ragona: 


a Province. Marmaris District 


Volume 93, Number 1 
2006 


Inocencio et al. 137 


Revision of Capparis Sect. Capparis 


6. Capparis ovata Desf., Fl. Atlant., Vol. I: 404. 
1798. Capparis spinosa L. var. ovata (Desf.) B 
in Batt. € Trabut., Fl. Algérie: 82. 1688. 15 5 
[Algeria] Habitat in fissuris rupium prope Oran” 
s.c. (lectotype, designated here, P 948). 


KEY TO THE SUBSPECIES OF CAPPARIS OVATA IN THE. MEDITERRANEAN 


REGION, THE MIDDLE East AND NORTH. AFRICA 


la. Young leaves usually ovate to DES lanceolate, 
X i 


mature leaves ovate, 2.5—5 m; stipules 


anttorSe 4 nage 9 e REESE C. ovata subsp. ovata 
lb. Young leaves ar lanecolae mature leaves 
ovate-lanceolate, 2.5 ).9-1.9 em: stipules 
retrorse e b. C. ovata subsp. myrtifolia 
Distribution. Figure 2. 


6a. Capparis ovata Desf. subsp. ovata 
Capparis pd Veill. var. kruegeriana Pamp., L Agric soltura 

Col. 22: 459, 1926. Capparis spinosa L. var. 
forma 5 Pamp.) Pamp., Prodr. F 
234. 1931. Capparis spinosa L. subsp. orientalis ( (Veill.) 
Flora of Libya. 


Jafri. var. Y a (Pamp.) Jafri., 


Vol X 


Shrub pendulous, sometimes reaching great dimen- 
sions in shaded sites; twigs straight, up to 3 m long, 
dark green or with a reddish tint in younger twigs, 
adult ones becoming woody, acquiring grayish or 
brown color: internodes 1.5-2.5 cm: stipules curved, 
mostly antrorse, sometimes spreading, not decurrent, 
sometimes very small or early falling, yellow, 0.15— 
0.4 em long, up to 0.1 em wide at the base. Leaves 
usually ovate, when young ovate to ovate-lanceolate. 
2.5-5.5 X 2-4.5 em, subcoriaceous; indument lax, 
trichomes thick and short, 30-40 X 200-250 um; leaf 


rounded or 


veins not prominent; base usually 
somewhat tapering, apices acute; mucro small, 0.5— 
| mm, straight; petioles short, 0.5-1 cm. Flower buds 
acute; floral pedicels thick, long, 4.5-6.5 cm; flowers 
slightly 


em long, 0.7-0.9 cm deep: stamens 


on 


slightly 11 abaxial (odd) sepal 
galeate, 1.5-1. 
30 to 80, 1 2-3 mm, with round apices. Fruit 


obovate, pulp yellow; ripe seeds dark brown, 2-2.2 X 


2.2-2.4 X 1.6-1.8 mm 
Illustrations. P. Ozenda (1991: 246, fig. 68): 
Lewalle & Montfort. (1997: 25); A. Benchelah et al. 


(2000: 147); Charco (2001: 283). 


Phenology. Flowering from April to December. 
Distribution and habitat. Saharo-Arabian and 
Mediterranean Regions. North Africa [Algeria, 


Libya, Morocco, Tunisia]. On rocks or walls of old 
buildings, at elevations from O to 2000 m, often in the 


vicinity of human dwellings. 


Selected specimens examined. ALGERIA. Ahaggar, Tez- 
zeit, 1750 m, Maire 180 (P); Bejaia, Reverchon s.n. (E): 
Djanet, Tamli of Tafalelet, Lhote 107 (P); Oranais, Faure 


LIBYA. Benghazi, Davis 50477 (RNG) 
oussa, Giado, Davis 49678 (RNG); Tripolitania, 
, Keith M (K). MOROCCO. Driouch, 
60029 (M UB); Gorge du Ziz, Er-Rachidia, Jury 1781 
ING); Fez, us 6653/95 (RNG): Imzoúrene, Jury 15602 
UN Msemrir, Dadés, Jury 17776 (RNG): Nador, 
Homo Pet (RNG); Ouarzazate, Brooks E.5349 
RNG); Di. Sarhoro, Ouarzazate. ua 53476 (E) River 
`~ Inocencio 60021. (MU B); , 
Taroudannt, Jury 14453 me Taza, Jury 6602 
G); Sidi Belkassen, Pu i s.n. (RNG); 
Ratherford BV1281/93 (RNG); Tazeka, Jury 16777 (RNG); 
Zaio, Inocencio 60000 (MUB). TUNISIA. Kabylia, Letourneux 
s.n. (P); Fedj el Khe 


Taura s.n. (P ). 


HM 


s.n. (E). 


~ 
- 
= 
— 


Inocencio 
9 


bl 
2 
e 
= 


Khargued, Letourneux s.n. (P): 


The sheet 948 at P contains two branches with 
leaves and without flowers or fruits; the specimen to 
the right is selected as a lectotype. It is labelled as 
“Herbier de la Flore Atlantique/ doné au Muséum par 
label 


" 


Desfontaines/ N/ Capparis ovata.” Another 


states “Habitat in fissuris rupium prope Oran.” In the 
protologue, Desfontaines also cites seven pre-Linnae- 
an references associated with images of a Capparis 
species known since at least the 1st century AD from 
the coasts of Marmarica (northern Libya and Egypt), 


== 


which is without doubt Capparis ovata s. str. However, 
other authors interpreted these as belonging to C. 
spinosa or C. sicula. 

Candolle (1824) raised the question of homonymy 
with Capparis ovata M. Bieb. On account of this 
supposed homonym he proposed naming the species 
from Algeria with the type material by e 
. 1 DC. Prodr. Vol. I: 245 
TYPE: “in fissuris rupium Mauritaniae prope Or- 
an... V.S. Desf.” 
adopt this view of Candolle because C. ovata M. Bieb. 


e fl. in h. There is no reason to 


is a later homonym (Bieberstein, 1808). 


6b. Capparis ovata Desf. subsp. myrtifolia Ino- 
cencio, D. Rivera, Obón & Alcaraz, subsp. nov. 
TYPE: | “Habitat in fissuris rupium Enneri 
Gousa, 3000, 15 Mar. 1953, Guichard 
KG/TIB/40” (holotype, the specimen with the 
“KG/TIB/40”, P!). 


Tibesti, 


label 


Fon angustis lanceolatis, stipulis valde retrorsis a typo 


Shrub pendulous, sometimes reaching great dimen- 


— 


sions in shaded sites; twigs straight, dark green or with 
a reddish tint in younger twigs, adult ones becoming 
woody, acquiring grayish or brown color; internodes 
1.5-2.5 em; stipules curved, retrorse, not decurrent, 
0.2-0.4 em long, up to 0.1 em wide at the base. 
ovate-lanceolate, the young leaves 
.9-1.9 cm, 
ceous; indument lax, trichomes thick and short, 30—40 
x 200-250 um; leaf veins prominent; iue dios 


Leaves usually 


usually lanceolate, 2.54 X 0 subcoria- 


tapering, apices acute; mucro small, 0.5-1 mm, 


138 


straight; petioles short, 0.5-1 em. Flower buds acute: 

floral pedicels thick and long, 4.5-6.5 em: flowers 

slightly zygomorphic: abaxial (odd) sepal slightly 

galeate, 1.2-1.5 em long, 0.5-0.7 em deep: stamens 

Fruit 

obovate, pulp yellow: ripe seeds dark brown, 2-2.2 X 
1.6-1.8 mm. 


30 to 80, anthers 2-3 mm, with round apices. 


Phenology. Flowering from (September)January to 
March. 

Distribution and habitat. Saharo—Arabian Region. 
North Africa [Algeria, Chad]. On rock crevices, from 
1000 to 2000 m. 

The Capparis ovata populations of the Central 
Sahara mountains show distinct retrorse stipules (as 
most species in section Capparis) in contrast to C. 
ovata populations of North Algeria and Morocco which 
have antrorse, sometimes 


mostly spreading, nol 


decurrent, sometimes very small or early-falling 


stipules. The type of stipules is a character very 
distinct and constant. The leaves are narrower and the 


floral pedicels are longer in the Central Sahara 


specimens. This led us to distinguish this new 
subspecies, that is subordinated to C. ovata. Other- 
wise it has some likeness to C. cartilaginea Decne., 
which is also present in the area, but the strongly 
zygomorphic flowers of the latter species (which we 
included in another section) are not found in C. ovata 


subsp. myrtifolia. 


Paratypes. ALGERIA. 
ae 28 (K); Tamanrasset, 
. Dalloni s.n. (P); Ennerdi, 
(omini Ennerdi, St. Serole 57 (P); 


Brookt 52 (K): 
(P) 


Darmouilly, W of Tamanrasset, 

Chipp 28 (P). CHAD. Aouzi. 
Dalloni s.n. (P); 
Gorges. L'Oudingueur, 
1000 m. 


Tibesti. 


Enneri Gousa. Tibesti. 


Guichard s.n. 

7. Capparis parviflora Boiss.. Orient., 
Ser. J. Vol. I: 4. 1843. 
parviflora (Boiss.) Boiss., Fl. Orient. Vol. E: 420. 
1867. Capparis leucophylla DC. var. parviflora 
(Boiss.) Zohary, Bull. Council Israel 8 D: 
59. 1960. TYPE: [Iran] “Hab. in Persia australi. 
Aucher pl. exs. N 


Diagn. Pl. 
Capparis spinosa var. 
Res. 


119] et 4191 
18503). 


(lectotype. 


designated here, GI. ( 


KEY TO THE SUBSPECIES OF CAPPARIS PARVIFLORA IN THE MIDDLE 
LAST AND CENTRAL ASIA 
la. Fruit oblong or 


elliptical; twigs yellow gree 
Xx ie 


n 
mature leaves ovate-rounded, 0.5—2.5 0.5- 
2 cm, pubescence from dense to almost E oes 2 


Fruit oblong: mature leaves ovate-rounded. 


0.5-2 em; pubescence from dense to 


ax 
n~ parviflora subsp. parviflora 
2b. Fruit e FM mature le 'aves ovate-lanceolate, 
2.5 X 0.7 


— 


rarely obovate 2 em: pubescence 


almost glabrous . . . . . b. C. parviflora subsp. kurdica 


Annals of the 
Missouri Botanical Garden 


b. Fruit rounded: twigs yellow green to gray-white: 

mature leaves obovate to ovate-rounded, 0.7-1 X 
0.6-1 em: pubescence dense . . . . . . . . . . . . . 
c. C. parviflora subsp. sphaerocarpa 


Distribution. Figure 5. This Capparis species 
to) 


shows a geographical pattern of variation. 
7a. Capparis parviflora Boiss. subsp. parviflora 


Bomba 


idia] “On veins of trap rock in the 1 i 


Capparis murrayana Graham, Cat. PL 
TYPE: [h 


ravines al Mahableshwur [Mahabaleshwar], al 
Loghur (Sir C. 
Gibson)" 


Vacolm.) - about Hurryc hunder jee (Dr. 
(not seen). 


Shrub procumbent, up to 75 em high: twigs straight, 


up to 2m long, light green or yellowish | green: 
internodes 0.5-1.5 em: stipules somewhat curved, 


retrorse, not decurrent, golden yellow, contrasting 
with the twigs, 0.3-0.5 em long, 0.1-0.2 em wide al 
-2 X 0.5-2 cm, 
indument variable from dense to 
20-25 X 300-900 um: 


acute to tapering, 


the base. Leaves ovate-rounded. 


— 


herbaceous: 


trichomes thin and long, 


leaf. veins not prominent: base 
apices rounded or acute; mucro very 
0.3-0.5 em. 


floral 


0.5 mm, straight; petioles very short, 


Flower buds rounded or slightly acute; 


pedicels slender, short, 1.5-2.5 em: flowers slightly 
zygomorphic: abaxial (odd) sepal slightly galeate, 1.2— 
LA em long. 0.5-0.6 em deep: stamens 30 to 80, 
anthers 1.3-1.5 mm, with round P Fruit oblong, 
pulp red; ripe seeds brown, 2-2.2 8-2 X 1.6- 
1.9 mm. 
Phenology. Flowering and fruiting from (April) May 
lo September. 
habitat. — lrano-Turanian and 


Distribution and 


Saharo-Arabian Regions, extending to the Sudano- 


Zambezian Region. Middle East and Central Asia 
[Afghanistan, Iran, Pakistan. Saudi Arabia. 


Turkmenistan]. Cliffs, stony places, and slopes, in 
1850 m. 

A holotype was not designated for 
Graham (1839). The 


barium has not been found. 


semi-deserts; from 650 to 
Capparis 
Graham Her- 


murrayana John 


Selected AFGHANISTAN, Her: 

Hedge & Lamond s.n. (K); Kandahar Pirzada. Ede berg 205 » 
Hedge & Lamond 20524 (€): 
(E) Kajakay, Hedge > Lamond 
2 km to the N of Lashe 

Shin Dand, 10 mi. S se s 
1958, Chapman 26096 
Irchibald 2940 (E); 
Kotschy 309 (V); 


rs B Jafri 


specimens examined. 


(C) Kandahar Pirz ada, 


1 Petersen 


; Province of C hakhansur, 
Jowayn. red 1928 (E); 
(K): Zint eds Mam 
(K). IRAN. Fars, Kuh-i-Bamus, Se hiras, 
Kuh-e-Hari, A. J. Lee 83 (K): 
Hohenacker 309 (C). 


district, 8 June 


Schiras, 


2106 (E). SAUDI ARABIA. Aflja Well, J. D. den 
(RNG): Turayf, Collenette 4495 (K). Tl cR 
Ashjabat, s.n. (E). 


Volume 93, Number 1 
2006 


Inocencio et 139 


Revision of uM Sect. Capparis 


Tb. Capparis parviflora Boiss. subsp. kurdica 
(Zohary) Inocencio, D. Rivera, Obón € Alcaraz. 
et comb. nov. Basionym: Capparis ovata 
Desf. var. kurdica Zohary, Bull. Res. Council 
Israel. Vol. 8D, 56. 1960. TYPE: [Iraq] “Rupes 
Mt. Singarae [Djabal Sindjar]. Mai 1867 C. 
(holotype, K!). 


stat. 


Haussknecht S. n.“ 
Shrub procumbent, almost glabrous: twigs straight. 
up to 2 m long, yellowish green; internodes 1.5—3 em: 


stipules somewhat curved, retrorse, somewhat to very 


widely spreading, not decurrent, golden yellow. 
contrasting with the light twigs. 0.2-0.4 em long. 


0.1—0.2 em wide at the base. Leaves ovate-lanceolate, 
1-2.5 X 0.7-2 cm, 


leaf veins not prominent; base rounded or somewhat 


rarely obovate, herbaceous: 


tapering. apices acule lo rounded: mucro very 


small, 0.1-0.5 mm, straight; petioles very short, 0.3— 


0.5 cm. Flower buds rounded; floral pedicels 
slender, short, 2-3.5 em: flowers slightly zygomor- 


phic: abaxial (odd) sepal slightly galeate, 0.9-1.2 cm 
long. 0.4-0.6 em deep: stamens 30 to 80, anthers 1.5— 
1.8 mm, with rounded apices. Fruit ellipsoidal. pulp 
red: ripe seeds dark brown, 3.2-3.8 X 2.2-2.0 X 2.1— 
2.2 mm. 
Phenology. Flowering from June to September. 
Distribution and habitat. Irano Turanian Region. 
Middle East [ Afghanistan, Iran, Iraq]. Cliffs, ravines, 


in stony places, from 200 to 700 m. 


Selected specimens examined. AFGHANISTAN. Cha- 
khansur. Zaranj. Breckle 4905 (E). IRAN. 9 Se 1 
Haussknecht s.n... (K). IRAQ. Bilas, Raui 294 K); 
Darvidikhan. 1536 (E): Zint gorge. Mam district, a 


26096 (K) 
7e. Capparis parviflora Boiss. subsp. sphaero- 
arpa Inocencio, D. Rivera, Oben & Alcaraz. 
subsp. TYPE: “Herat: 105 
Km S Herat versus Shindand, K. H. Rechinger. 
Iter Orientale 1967/37579” ) 


Figure 6. 


nov. | Afghanistan] 


(holotype. E”. 


Foliis minoribus. usque ad | em longis dem fruetibus 
brevioribus globosis non oblongis a typo differ 


Shrub procumbent, up to 75 em high: twigs str aight, 
up to 2 m long, yellowish green or erayish-whitish: 
young twigs very thin with very short internodes (0.1 
0.3 em) and very spiny: longer adult twig internodes 
0.3-1 em: stipules somewhat curved, retrorse, usually 
very open, slightly spreading to widely spreading. 

g yellow 
0.30.5 cm long. 0.1— 


ovate- 


slightly decurrent to not. decurrent, golden 
contrasting with whitish twigs. 
Leaves obovate to 
0.7-1 X 0.6-l cm, 
herbaceous; indument dense, trichomes thin and long, 


20-25 X 350-800 Lun: 


0.2 em wide at the base. 


rounded, the young rounded, 


leaf veins not prominent; 


bases tapering to rounded, apices rounded or weakly 


acute: mucro very small, 0.1-0.5 mm. straight; 


Flower buds rounded: 
1.5-2 em: 


abaxial (odd) sepal slightly 


petioles very short, 0.1-0.3 cm. 

floral pedicels slender, short, flowers 

slightly zygomorphic: 

galeate, 0.8—1.1 cm long, 0.3-0.6 cm deep: stamens 

30 to 80, anthers 1.2-1.5 mm, with round apices. 

Fruit rounded, pulp red; ripe seeds dark brown, 2.6- 
8 X 1.6-1.8 X 2.3-2.8 mm. 


Phenology. Flowering and fruiting from (April) 
May to August(September). 


Distribution and habitat. \rano-Turanian Region. 
Central Asia [Afehanistan, Turkmenistan]. Cliffs and 
stony slopes, from 600 to 2000 m. 

The sheet with the holotype (at E) is labelled as 
follows: | “K.-H. 1967/ 
37579/ Capparis spinosa L./Var. parviflora (Boiss.) 


Herat:/105 Km S Herat Ver- 


Rechinger, Iter Orientale 


Boiss./Sw- Afghanistan, 


sus Shindand, 1300 M." II “Det. Hedge & Lamond, 
1968/ T. VIII.“ 

Paratypes. AFGHANISTAN. — Chakhansur. Zaranj. 
Breckle 4928 (E); Fariah, Hedge. M _ & Ekberg W 


und Dand. 


Vikitin & 


7674 (E); Herat. Hedge & Lamond 57579 (K): 5 
5488 (K). TURKMENISTAN. ce] 


Ivanov s.n. (E). 


Furse 


8. Capparis sicula Veill.. in Duhamel, Traité Arb. 
Vrbust., Ed. 2, 1: 159. 1801. Capparis ovata var. 
sicula Neill.) Zohary, Bull. 
Israel. 8D: 55. 1960. Capparis spinosa subsp. 
sicula (Veill.) Holmboe. Borg. Mus. Skr. Ny 
Raekke. Bird L 2: 1-344. 1914. TYPE: [Italy] 
“Capparis Sicula duplicata spina, folio acuto. 
Boccone, 1666 [1674]. Icon. et descript. Rarior. 
Plantar. Sicil. P. 79. tab. 42. f. 
Cathol. P. 36” (lectotype, designated by Rivera 
20006. 
descriptiones rariorum plantarum Siciliae. Meli- 


tab. 42. f. 3. 16744). 


Res. Council of 


3. Cupani Hort. 


et al. Boccone’s image. [cones el 


tae. Juliae et Italicae: 79. 


KEY TO THE SUBSPECIES OF CAPPARIS SICULA IN THE MEDITERRANEAN 
REGION. NORTH AFRICA, THE MIDDLE East. AND CENTRAL Asta 


la. Plants with stipules straight. spre 'ading 


pad CD 8b. C. sicula subsp. herbacea 
lb. Plants with stipules curved, retrorse . ...... .... 2 
2a. Leaf pubescence lax, very lax, or almost glabrous . 3 


Ja. Twigs dark green: leaf pubescence lax, trichome 


thick and long. 20—50 X 400-800 um; floral 

pedicel thick and long, 5.5-7 m.. 

CERTI ee + eee ee ee 8a. C. sicula subsp. sicula 
3b. Twigs light green. white, or yellow: leaves 


glabrous or with very lax pubescence, trichomes 
thin 15-20 x 100-250 um: floral 
pedicel slender and short, 1—1.5 em 


and short. 


4a. Leaf pubescence very lax: stipules hess 
trorse, yellow-orange: anthe rs small, 2-2.5 mm. 
AAA 3d. C. sicula ls mesopotamica 


140 


Annals of the 
Missouri Botanical Garden 


tb. Leaf almost glabrous: stipules somewhat curved, 
8 I 
retrorse or spreading, golden yellow contrasting 


with the light color of twigs: athe rs very small, 
€ 


1.6-2 mm | sicula subsp. sindiana 
2b. Leaves appe aring white mus d dense pub )escence 
e sicula subsp. leucophylla 


Distribution. figure 5. 


9a. Capparis sicula subsp. sicula 


Capparis spinosa var. canescens Coss., Anal. Sci. Nat., 11: 28. 
19. Cap 


18 aris spinosa subsp. canescens (Coss.) A. & 
O. Bol Misc. Fontsere: 88. 1961. vic lol ovata 
Desf. var. canescens oss.) Heywood, Feddes rl 


ome n 
69: 1964. TYPE: [Spain] “In rupestribus aca 
hes prope Ps rez (E. Bourgeau. pl. Esp. 1849. 


n. (lectotype, designated here. P!). 

D ovata var. palaestina 1 Bull. Res. Council 
Israel 8D: 55. 1960. TYPE: [Israel] Galilee. 
Wadi Hindaj, 25 June 1 M. Zohary 110 ERAN 
HUJ!) 


Shrub procumbent; twigs straight, sometimes reach- 
ing up to 3 m long, dark green; internodes 1.5-5 em: 
slipules curved, retrorse, not decurrent, occasionally 
0.3-0.0 cm 


Leaves ovale. 


slightly decurrent, yellow-orange, long. 


0.2-0.3 em wide at the base. oblong or 
elliptic, 3-5 X indument lax, 
trichomes thick and long, 20-50 X 400-800 um: leaf 


veins not somewhat 


2—4.5 em. herbaceous: 


prominent; base rounded or 


tapering, apices acute; muero long, 1-1.5 mm, usually 


curved: petioles short, 1-1.5 em. Flower buds acute: 


floral pedicels thick and long. 5.5-7 em: flowers 


zygomorphic: abaxial (odd) sepal galeate, 1.7-2.5 em 
long. 0.7-1.2 em deep: stamens 100 to 150. anthers 


2.5-3 mm, with acute apices. Fruit oblong. pulp red: 


ripe seeds dark brown, 2.6-3 X 2.4-2.8 X 2-2.2 mm. 

Illustrations. | Boccone (1674: 79, tab. 42. f. 3): 
Zohary (1966: 359); Valdés et al. (1987: 374): 
Castroviejo et al. (1993: 520, pl. 142); Plitmann et 


al. (1983: 81): O. Fragman et al. (2001: 312-3. 


pl. 133). 
Phenology. Flowering May to October. 
Distribution and habitat. Mediterranean and 


Irano-Puranian Regions: locally introduced in the 
Saharo-Arabian Region. Mediterranean Europe. North 
Africa. Middle East into Turkey [Albania. 


Cyprus, Greece, Haly, Morocco, Spain, Syria, Turkey |. 
Stony 


Algeria. 


marls or clayish soils; Scrubs. at 


to 600 m: 


places. 


elevations from 0 in the vicinity of human 


dwellings. 


lected specimens examined. ALBANIA. Below Levani 
"e Frascula LE 15 Barat. Baldaci s (K). ALGERIA. 
PRUS. Cape Greco. Gold 29 (RNG). 
Island E 


Orán. Faure s.n. (K 


GREECE. Baths E 5 80 iboea. Stamatiadhou 


17290 (C): Crete, Kissamos. Bickerich 15091 (K); lrakliou. 
Bowen 8913 (RNG): Fokis, Delfi, Pinset 107 (RNG): Island 
Poros, Strid 29680 (€); Mitilini. Lesbos, Hansen 4918 (( 5); 


Hanson 67—1-1. (RNG). ITALY. 
Basilicata. Potenza, Akeroyd 3328 (RNG): Gigerti. Agrigento, 
1 s.n. (C). MOROC CO. Fez. Inocencio 60016 (MUB): 
: Inocencio 00014 (MUB): Meknès, Jury 15502 
m Karcen, 60017 (MUB). SPAIN. 

Reservoir of Mearaz did (MUB): 
» Inocencio &. Alcaraz 180 


Valimiú, Peloponessos. 


Inocencio 
\madoiro, 


Alcaraz 48710 (MUB): ( „ Or. 

Mearaz Hu 14 (MUB); La as C las de D. Pepe. V 
Inocencio & Alcaraz 8682 e B) 
Vearaz 18723 (MUB) 
& Alcaraz 48713 
18728 (MUB): Serón, 


Sierra de Gador. | 


Vcaraz 
oven (MU B): 
i es Inocencio 
2524/08 (RNG). 
Cannon 692 (RNG): Mala, 
1 0 13 337 (RNG): Lanjarón. Inocencio & Alcaraz 18718 


1 & Ale 


~ 
a 
2 


(MUB): Reservoir of Negratín, Inocencio & Alcaraz 18703 
(MI 9 Guadix. Inocencio & Alcaraz 48712 (MUB): Loja. 


Inocencio & \learaz 18081 (MU B): Baza. Inocencio & Mearaz 

18080 (MUB): Salobreña. Inocencio & Alcaraz 18717 (MUB). 

VIDES. Inocencio & Alcaraz 48688 

Inocencio & 

1 & 

Mallorca: from Pim de Mallorca to Andr ailx. Inocencio & 
la 


Alearaz 18726 (ML 1 0 Murcia: San 


Lange s.n. (C); Murcia, Lange s.n. (C); Ve del Olivo. 
Inocencio & e EL (MLB): Barinas. /nocencio & 
Vearaz 46704 (MUB); Albudeite. Inocencio & Alcaraz 48709 


te, 

(MUB); Lorca, Inocencio & Alcaraz 48722 (MUB): 
Inocencio & Alcaraz 48711 al 
Inocencio & . Nemaz 16721 (MUB); 
Menor. os & Alcaraz 48727 (MUB); 

e MUB). Sevilla: Osuna. /nocencio & Alearaz 
8679 (MUB): S of Spain, J. W. Carr 20706.6 (RNG). 
sh RIA. Djebel Casioun. Damascus. Pr NA 1518 (K). 


lo Guadiana 
Jódar, Inocencio & 


URKEY. Ankara. Davis & Coode 37217 (E): Dardanelles. 
sn. (E) Denizli, A Hormia 588 (RNG : E mm 
Baytop 14306 D Hakkari, Zab river, arce ny gd 
K d d nli. Mersin, He nnipman 1073 (RNG); iei 


kale. Denizli. 191 55 10251 (1): Sarayc ra Osmane N Tobey 
9) 


250 Ñ o). 


8b. herbacea (M "e 
Obón & Alcaraz. stat. 
comb. nov. Basionym: Capparis herbacea N 155 
Enum. Pl.: 560. 1800. 
herbacea (Willd.) Zohary. 
Israel 8D: 56. 1960. TYPE: 
Yzerbaijan] “Capparis herbacea. Marschall a Bie- 
Habitat ad 
here. B! 
BA. 


Capparis sicula subsp. 
Inocencio, D. Rivera. 
Capparis ovata var. 
Bull. Res. 


[Russia, Georgia or 


Council 


berstein. Caucasum" (lectotype, 


designated unlabelled blossoming 
specimen, folder 10034, right of sheet 2). 
ete he ie Willd. var. microphylla Ledeb., Fl. Ross.: 
2. Capparis le d DC. 

Mere Tiekh. 5 tuden 
* : [Kazakhstan or Turkmenistan] * 


var. microphylla 
Flora of Egypt. Ed. 2 
"n 


litt. Orie ntm. Caspii! Karelin” (holotype. LE not seen). 
Shrub procumbent; twigs straight. sometimes reach- 
3m light 


internodes 1—3 em: stipules straight, spreading, not 


ing up lo long, green, herbaceous; 


Volume 93, Number 1 
2006 


Inocencio et al. 
Revision of Capparis Sect. Capparis 


JAC 


10 cm 


Figure. 6. 
—A. Stems and flowers (drawn Eb 
from Nikitin & Ivanov s.n., E). — 


A. Barreña from Fu 

. Fruit (drawn by I-A 
decurrent, golden 0.3-0.0 em 0.1- 
0.2 cm wide at the base. Leaves elliptic or oblong, 
3-6 indument lax, tri- 
chomes thin and short to long, 15-20 X 200—500 um; 


leaf veins prominent; bases rounded, apices obtuse to 


yellow, long, 


1.5-4 em, herbaceous; 


— 


acute; mucro small, 0.5-1 mm, curved; petioles short, 
11.5 cm. floral 


pedicels thick and long, 4—6 em: flowers zygomorphic: 


— 


Flower buds acute or rounded: 
abaxial (odd) sepal helmet-shaped, 1.8-2.4 em long, 
0.6-1.2 cm deep; stamens 100 to 150, anthers 2.3— 
3 mm, with acute apices. Fruit obovate to oblong, pulp 


3—3.6 X 2.4-2.8 X 2- 


red; ripe seeds dark brown, 


2.4 mm. 
Note. Mongolian populations display stouter 
stipules and smaller leaves. 
Illustrations. Takhtajan. (1966: 58, tab. 22). 
Phenology. Flowering from May to September. 


Details of the new subspecies Capparis 1 ol ec Inocencio, D. Rivera, Obón, & Alcaraz. 
5488, 3. De 


ES 


— tail of Mur. rr M y 


J.-A. Barreña 


"Barrel Ta m edes & Lamond 37579, E). 


Distribution and habitat. — lrano-Turanian Region 


extending to the Euro-Siberian. Middle East, Central 
Asia, and Caucasus [Afghanistan, Azerbaijan, 
Georgia, Iran, Kazakhstan, Mongolia, Turkey, 


Turkmenistan, Ukraine, Uzbekistan]. Stony places in 
calcareous soils and rocky grounds, slopes of low hills 
and walls of abandoned buildings, at elevations from 
O to 2000 m. 


Selected specimens examined. AFGHANISTAN. Bamian, 
Darrah Siakar, Hedge 3423 (E); Den Hundie, Edelberg 1889 
(C); Kundut, Khanabad. Carter 390 (K): Obe h, Herat, Hedge 
W7777 (E) Samangan, Hewer 1128 (E); Takhar, Mughul. 
Podlech 11377 (E); Yawarzan. Badkshan, Hedge W9471 (E). 
AZERBAIJAN. Morghak, from Bileh 
Lamond 3097 (E). GEORGIA. S.L, Frierk s.n. (E); Caucasus. 
Hohenanker s.n. (E); Tbilisi, o 55 (K). IRAN. Amol, 
Andersen 244 (E); Khvoy, 
Elburz, 7 
Moraweh Tappeh. 


icons to Savar, 


Cowan 1560 (K): ** 
Furse & Synge 491 (K); Mianeh, Bowles 2423 (K 
Hewer H3804 (E): Sanganeh E Dagh. 


142 


Annals of the 
Missouri Botanical Garden 


Ghorashi-Al-Hosseni 4956 (RGN). KAZAKHSTAN, 
gorozi loshnee sel. 
GOLIA. Mongolia. Potanin s.n. (K); Eastern 
Mongolia, Gobi, Prezwalski s.n. (K). TURKEY. Artvin, 
Coruh, Davis & Hedge D32427 (E); Hakkari, Kalolans, Davis 
23870 (Ey Kagizman. Watson 375 (K): Karabük. Baytop 
11383/65902 (E); Kars, Davis 46680 (©) (K): Konya. 
\kyokus, Dural 576 (E); Osmanive, Balls 1199 (K): Sinop, 
Kargi, Tobey 2809 (E); Tokat-Niks: Davis 24882 (N): 
0 Karabuk, Davis & a D3 39050 (E). TURK- 
MENISTAN. Aschabad, 1 260b (E); Syr-Darja, Golike 
F. /I. 19 UKRAINE. Kryi : Toige 1896 (RNG). 
UZBEKISTAN. Altyn Tepe, 190 m. 


(K): Chauvast. Samarkand. "Nan 277 (C). 


Tehavto- 
Tamerlanovski, Priajin s.an. (K). MON- 
Tian-Shan, 


Vasak 


There is a clear reference in the protologue to the 


(53 


raucasus, which indirectly points to the specimens 
from this area in the herbarium Willdenow (surpris- 
ingly the protologue does not mention the Mussin- 
Pushkin expedition, vid. infra). The reference in the 
Marschall 
a Bieberstein” is presumably pointing do Capparis 
Bieb 1808. 1819). In the 
herbarium Willdenow (B) are two sheets within folder 
nr. 1003 
1972). 


Capparis herbacea foliiis/ subrotunde ellipticis ovalis/ 


protologue to a “Capparis herbacea. 


ovata M. (Bieberstein. 


pertaining to Capparis herbacea (Hiepks, 


The folder is labelled: “Polyandria Monogynia/ 
Pedunculis/ unifloris/ Habitat Cauca- 
Sheets | and 2 


axillis spinosis. 
are numbered and annotated both 
and “W.” The 
Pushkin? W.” 
Count Apollo Apollosovich Mussin-Pushkin (1760— 
1805). 


collector. He 


sus.” 


as follows: "C. herbacea” ` folder has two 


other labels: “Mussin This refers t 


who was a Russian explorer and plant 
led a botanical expedition to the 


1800-1602. label 


illegible. It is not clear whether the specimen was 


Caucasus in The other is almost 


originally collected. in the Caucasus by Mussin- 
Pushkin or. presumably, cultivated. in the Royal 


Botanic Garden of Berlin from seeds gathered by this 


collector (the species is published in a catalog of 
plants actually grown at Berlin). This species was. in 
fact. 


cultivated in another botanic garden in Saint 


Petersburg: “Cultam in tepidario saepe fruticosam 


fieri et C. spinosae assimilari nunciat Fischer in 
litteris” [Friedrich Ernst Ludwig von Fischer was 
director of the Imperial Botanic Garden of Saint 


Petersburg (Pritzel, 1872)| (Bieberstein. 1819). 


9c. Capparis sicula Veill. subsp. leucophylla (DC.) 
Obón & Alcaraz, 
comb. nov. Basionym: m : ee Dus 

Prodr. Vol. I: 246. 1824. 2 [Iraq] “Inter. 
Bagdad et Alep. Oliv. et 5 155 in Herb. Mus 


Par.)“ (lectotype, designated here, P!). 


Inocencio, D. Rivera, stal. el 


Capparis spinon L. var. 0 ens Zohary, Bull. Res. Council 
Israel 8D: 56. 1960, TYPE: [Egypt] 9 10 rocks, 
1100 ft.. 1944, P. E 1 3062 (holotype, K! 


Shrub 


sometimes reaching up to 3m long, glaucous: 


procumbent: twigs straight. semi-erect, 


internodes 1.5—5 em: stipules curved, retrorse, not 


decurrent or somewhat decurrent, golden yellow, 0.3— 


0.6 em long, 0.1-0.3 em wide at the base. Leaves 
elliptic to rounded, sometimes ovate, 2.5—1.5 X 2- 


whitish, 
trichomes thin and long, 20-25 X 200-500 um; leaf 


De 


3.5 em, herbaceous: indument very dense, 
veins not prominent; bases rounded, apices obtuse or 


acute: mucro long, 1-1.5 mm, straight or somewhat 
curved: petioles short, O. SI em. Flower buds round- 
ed or acute; floral pedicels thick and Short. 2.5— 
(odd) 
galeate, 1.5-2.2 em long, 0.7—1.1 em deep: stamens 
100 to 150 
Fruit oblong, pulp red: ripe seeds dark brown, 2.7-3 


X 2.6-2.6 X 1.8-2 mm. 


352 em: flowers zygomorphiez abaxial sepal 


anthers 3.5—1 mm, with acute apices. 


Hlustrations. Mandaville (1990: pl. 69-70). 


Phenology. Flowering from May to September. 


Distribution and habitat. \vano-Turanian and 
Saharo-Arabian Regions. North Africa, Middle East 
into Pakistan Afghanistan, Iran, Iraq. Israel. 
Pakistan. Saudi Arabia, Yemen]; also in Egypt 
(Zohary, 19600). Oasis in semi-deserts. flooding 
plains. sometimes in somewhat saline soils: at 


elevations from 0 to 1000 m. 

The sheet with the lectotype of Capparis leucophylla 
is labelled: | “Capparis leucophylla DC/ (De Candole 
Script.)“ H “de Bagad à Alep./ Olivier de Bruguiere” 
| The 


coincidence. between labels and references to type 


| “Herb. Mus. Paris./ Capparis leucophylla DC.” 


material in the protologue led us to suppose Candolle 
was implicitly designating a holotype. provided the 


author used only one element. 


Se 1 8 specimens examined. 


AFGHANISTAN, Baghlan, 


SW of Doshi. Hewer 1153 (K): KE un Furse 7735 (K). 
IRAN. E Leonard ys (K); Kerman, Parris 75405 (E): 


Rudak, [= Dehbarek]. Davis & y D. 56505 (E). IRAQ. 
Karin & Noort 39981 Tell Kotehek-Senonal. 
J. B. Gillet 108 o 8 Desert. S. of AI 
Salman, Ravi. Agnew & Haines 1656 (E); Shaikhiva. S 


Hamah, 
Mosul-Liwa. 


Salma- 


Samawa, Al-Shehbaz s.n. (RNG). ISRAEL. Gilboa mountain, 
seen 1 (E). JORDAN. Jordan V. L. Y.. S/1113 (K). 
PAKI Chitral. Tirich Stainton 2780 (E). 

AR UN Rumah. White 73 (K). MIN. Huth, Miller 
3150 (E): Sanaa, Miller 3401 (E). 


8d. Capparis sicula Veill. subsp. mesopotamiea 
Inocencio, D. Rivera, Obón & Alcaraz, subsp. 
TYPE: 9 Sep. 1918, W. 


Edgar Evans, (holotype. designated 


nov. [Iraq] “Amara, 


M/100 (VY 


here, EI. specimen E 65908). Figure 7. 


Indumentum — trichomatibus | brevioribus, 100250 um 


longis. el tenuioribus; 15-20 um latis; internodiis. aculeis 


ad basim te Er XA 0.1—0.2 em. pedicelis brevioribus. 3— 


1.5 


em. a typo diffe 


Volume 93, Number 1 Inocencio et a 143 
2006 Revision of et Sect. Capparis 


A 
L 4 " 4. -— 
10 cm 
Figure 7. Details of the new subspecies Capparis sicula subsp. me. Ld e Inocencio, D. Rivera, Obón, & Alcaraz. 
—A. Ste m and flower. —B. Detail of leaf and stipules. (A, B drawn by J.-A. Barreña from Furse 9021, K.) 
Shrub procumbent; twigs straight, up to 2 m long. deep; stamens 100 to 150, anthers 2-2.5 mm, with 


yellowish green; internodes 0.5-3.5 em; stipules acute apices. M i pulp red; ripe seeds brown, 
curved, retrorse, not decurrent, yellow-orange, 0.2— — 2.8—3 X 2.4-2. 1.7 mm. 


0.4 em long, 0.1-0.2 em wide at the base. Leaves m ' : 
ong ne EN M Illustrations. Townsend & Guest (1980: pl. in 
obovate or oblong, 2.2—3.5 X 1.3-2.5 em, herba- E : . 
8 , N front of the title page). 

ceous; indument very lax, trichomes thin and short, 


Phenology. Flowering and fruiting from July to 
100-250 x 15-20 um; leaf veins prominent; base 


September. 
acute, apices acute or rounded; mucro long, l- Distribution and habitat. — lrano-Turanian. Region. 
1.5 mm, straight; petioles short, 0.5-0.7 em. Flower Extending somewhat into the Mediterranean and 
buds rounded or slightly acute; floral pedicels slender — Saharo-Arabian Regions. Middle East [Iran, Iraq, 
and short, 3—4.5 em; flowers zygomorphic; abaxial Israel, Syria]. Sandstone, often near orchards and 


(odd) sepal galeate, 1.2-1.8 em long, 0.6-0.8 em groves, at elevations from 0 to 2300 m. 


144 


Annals 
con an Garden 


10cm 


\ 


Details of the new me cies Capparis sicula 
.—C 


] 


Figure 8. 
. Stem and flower. —B. Detail of leaf and stipules 


Paratypes. IRAN. Gulestan, Furse 9021 (K); 


Masjed-c- 


Saleyman, Lee 68 (K); Mazanderan, Tehran, s.n. 10530 (©). 
IRAQ. Amara, Edgar Erans (A (E): Saádiva, Al-Kaisi 
12894 (K): Sirk. Garmah. R. Wheeler 322 (©). ISRAEL. 
Mizde Dragot, Dead Sea. Danin : Knees 342 (RNG). SYRIA. 


Ain Dara, Rivera & Oban s.n. (MUB). 

8e. Capparis sicula Veill. subsp. sindiana Inocen- 
D. Rivera, Oben & Alcaraz, subsp. nov. 
[Afghanistan] Barak, 20 Aug. 1986, Mr. 


KY). Figure 8. 


cio, 

TYPE: 
Lulman, 92 (holotype, 
Mts. Vol. 


sh, India] “Hab. 


Kinawur]? 


iris pe Rs Ilinois Bot. Himal. 
18 de 


2: [Himachal Pra 


1 Kunawur [Kkinnaur. 


l s 


1 (type not 
seen) 


. Fruit. 


Veill. subsp. síndiana Inocencio. D. Rivera, Obón & Alcaraz. 
drawn by J.- K.) 


(A-C 


. Barreña from Lulman 92. 


Caulibus viridescentibu luteis, trichomatibus laxis, 


vel 
pedunculis floralibus racionis brevioribus, a 
differt. 


mesopotamica differt. 


typo 


Stipulis aureis, antheris minutis a subspecies 


Shrub procumbent, almost glabrous; twigs straight, 
3 m long. light green or 


1-3 em: 


retrorse, 


sometimes reaching up to 


vellowish: internodes stipules curved or 


somewhat curved. occasionally spreading. 


not decurrent to somewhat decurrent, golden yellow 
contrasting with the twigs, 0.1-0.4 em long, 0.1— 


0.2 em wide at the base. Leaves obovate or elliptical, 
1-3.5 


base 


1-3 em, herbaceous; leaf veins prominent: 


rounded, apices acute: mucro very small, 


Volume 93, Number 1 
2006 


Inocencio et al. 145 


Revision of Capparis Sect. Capparis 


slightly curved; 


short, 0.3-1 em. Flower buds acute; floral pedicels 


0.1-0.5 mm, straight or petioles 


slender, short, 1.54 em; flowers  zygomorphic: 
abaxial (odd) sepal galeate, 0.8—1.2 em long. 0.4— 
0.6 cm deep: stamens 100 to 150, anthers 1.6-2 mm, 
with acute apices. Fruit ellipsoidal to oblong, 15 
red; ripe seeds dark brown, 2.8-2.6 X 2.4-2.0 X 1.8— 


2 mm. 


me Royle (1839: 73); Narvi & Ali 
(1973: 8, fig. Za): Polunin & Stainton (1984: pl. 16, 

161). 

Phenology. Flowering from July to September. 


Sudano-Zambezian, Indian 
Middle East into India 


Distribution and habitat. 
and lrano-Turanian Regions. 
[ Afghanistan, India, Pakistan]. Walls, meadows, and 
1000 to 3000 m. 


In the lower left-hand. part of the sheet with the 


human dwellings: from 


holotype of Capparis sicula Veil. subsp. sindiana at K 
"FLORA OF 
. 92 / Name. Capparis spinosa L.. / 


is a label with the following script: 
AFGHANISTAN N 
Native Name: Barak: dry riverbed. shale / Locality & 
Altitude: / sunny widespread growth. straggling, / 
Notes: procumbent stems. Fruit when ripe / dehisces 
along 3 lines to open out as / mass of bright red sticky 
fluid seeds / embedded in latter / alt. 1430 m. 7 
Collector: Mr Lulman / Date: 26.8.1968.” 

Herbarium is a 


JV; 


The John Forbes Royle 
however. Donna Young (pers. comm.) has checked in 
that 


specimen for Capparis ovata is there. Other herbaria 


the Roylean herbarium and verified no type 


where possible original material may be (DD and K) 
were unsuccessfully contacted. Therefore a neotype is 


neccessary. 


Lulman 92 (K). 
Punjab, 


Paratypes. AFGHANISTAN. Barak, 
INDIA. Wangtu to Sholtu, Cholto bridge, Bashahr, 
Lace 178 (E): Li. Bushahr, Simla, Eastern Punjab, Parma- 


nand 725 (E). PAKISTAN. Balti, Iskalkoo, Winterbottom s.n. 
(K); Gaud nullah, Gupis, Gilgit, 3 Omer 266 (E): 
Gilgit-Karimabad. 1500 m, Qaiser, Omer & Husain 8444 
(RNG); Kharipur, Jafri 2421 (E); Shardu, Baltistan, Shah 


249 (E); above the Indus river, Shardu, Kashmir, Baltistan, 
2000 m, Webster & Nasir 5771 (K) 
9. Capparis spinosa L., Sp. pl.: 503. 1753. TYPE: 


[France?| “Habitat in Europae australis arenosis, 
(lectotype, designated by Burt & 
Bull. 299. 1949, BM, Herb. 
Capparis No. 912.348-50, vali- 
1965). 


ruderatis” 
Lewis in Kew 
Clifford: 2031, 
dated by Jacobs, Blumea 12/3: 417. 


Capparis peduncularis Presl., Delic. Prag.: 20-21. 1822. 


Shrub procumbent; twigs straight, sometimes reach- 


ing up to: dark green; internodes 1.5-3.5 cm: 


stipules ie retrorse, not decurrent, slender, weak 


or vestigial, rarely strong, usually very long and thin, 


dark yellow, 0.3—0.6 cm long, 0.1-0.2 em wide at the 
4—5 
herbaceous; indument very lax, trichomes thick and 
long, 25—40 X 300—500 Um, early falling: leaf veins 


not prominent; base rounded or somewhat tapering, 


base. Leaves ovate or obovate, —3.5 em. 


apices acute; mucro very small, 0.1—0.5 mm, straight: 


petioles short, 0.7-1 em. Flower buds acute; floral 
pedicels thick and long, 5-6.5 cm; flowers zygomor- 
phic: abaxial (odd) sepal galeate, 1.8-2.4 cm long. 
0.6-1.1 cm deep; stamens 100 to 150, anthers 2.5— 
Fruit ee pulp red; 


2.6-2.8 X 2 


2.8 mm, with acute apices. 
ripe seeds brown, 3-3.2 X 2 2-2.5 mm. 
Illustrations. Coste (1900: 142); Woodville (1794: 
plate 228, drawn and engraved by James Sowerby). 
Phenology. Flowering from May to October. 

Distribution and habitat. Mediterranean Region. 
Europe, Middle East into Turkey 
Italy, Spain, Turkey]. Cultivated, 


sometimes found in secondary habitats as a feral or 


Mediterranean 
|France, Greece, 
living among the parental species, at elevations from 
O to 250 m. 


Rivera et al. 


Figure 2. 
(2006 


typification of Capparis spinosa. 


reviewed the status and 


— 


In summary, mor- 


shological, reproductive, and molecular data suggest. 


— — 


ut do not prove, that this is a hybrid species kept in 
cultivation. The origins of the different populations 
the 


cultivated individuals i 


are related to coincidence of wild and/or 


| proximity, belonging to C. 
sicula and C. pa especially in the Western 
Mediterranean [Sicily, Mallorca], but also in Greece. 


y within this erop shows a pattern of 


The diversi 
variation interme e between the putative parentage 
(Rivera et al., 

Selected Dum examined, FRANC E. 
moa 243 (E); sl. Agardh. s.n. (C). GR 
: Periol Toae 1 5 1 
Lasfeu S. u. (C). IT ALY. Sicily, Agrigento, Davis 40229 
(RNG): Stromboli, B. Larsen s.n. (C). SP je 
er lo, Inocencio & Alcaraz s.n. (MUB); Paraiso Beach, 


Villajoyosa. Hewat H 1040 Ses ü Barcelona: Barcelona, 


5 lier. 
Rhe 


des, 


Sennen 1587 (RNG). Cordo | Templete, C. López 
CLI971/86 (RNG). a Too Inocencio & 
Mearaz s.n. (MUB); Campanet Caves, Inocencio & Alcaraz 
s.n. (MUB): from Santa Maria to Inca, d & Alcaraz 


C). TURKEY. 
Ehim 373 


Christensen 1372 ( 
Eskisehir-Súndihen, 


s.n. (MUB); Valldemosa, 
Hänel 00.437 (E); 


Denizli, 
(E). 


10. Capparis zoharyi Inocencio, D. Rivera, Obón 
& Alcaraz, sp. nov. TYPE: [Spain] “El Llano del 
Beal, Murcia, Spain, 1999, Inocencio 
42669" (holotype, designated here, UMH!; para- 
types, ede & Alcaraz 70102, 70103, 48689, 
sent to K, E, and MO!). Figure 9. 


July 


146 Annals of the 
Missouri Botanical Garden 


1mm 


F 


Figure 9. Details of the new species Capparis zoharyt Inocencio, D. Rivera, Oben & Alcaraz. —A. Habit. —B. Detail of 
leaf variability. —C. Stems. —D. Detail of the stipules. —E. Detail of nectary. —F. Detail of anther. 
by P. Perales & J.-A. Barreña from Inocencio & Alcaraz 18689.) 


G. Fruit. (A-G drawn 


Volume 93, Number 1 
2006 


Inocencio et al. 
Revision of Capparis Sect. Capparis 


147 


Frutices erecti; usque ad 2 m; stipulis. decurrentibus, 
similibus rosaris: foliis rotundis-obcordatis, rarius ovalis, (2— 
2-4 cm latis). 
m HUE aegyptia differt. In memoriam Michaelis Zohary 


dic 


Lem longis. apice emarginatis vel obtusis 


Shrub erect, glabrous: twigs straight, + erect, up to 
2 m long, green to reddish purple, older twigs bluish 
due to a waxy covering; internodes 1-5 cm; stipules 
curved, retrorse, strongly decurrent, rose type, orange. 
0.3-0.6 cm long, 0.3-0.4 em wide at the base. Leaves 
rounded to obcordate, 
leaf 


sometimes 


rarely ovate, 2-4 X 2-4 cm, 


somewhat fleshy; veins not prominent; base 


rounded or 
0. 1— 


buds 


rounded, cordate, apices 
slightly obcordate; muero absent or very small. 
short, 0.7-1 cm. Flower 


0.5 mm: petioles 


rounded: floral pedicels thick and long. 3-5 cm: 


flowers slightly zygomorphic; abaxial (odd) sepal 
slightly galeate, 1.5—1.7 
stamens 30 to 80, 


apices. Fruit oblong, pulp yellow: ripe seeds brown, 


em long, 0.6-0.9 em deep: 


anthers 1.3-1.5 mm. with round 


3.4-3.8 X 3-32 X 2-2.2 mm. 
Illustrations. Figure 14 in Zohary (1966: 358 
Phenology. Flowering and fruiting from 105 lo 
October. 


Distribution and habitat. Mediterranean Region. 
Mediterranean Europe, North Africa, Middle East into 
Turkey | Algeria, Egypt, Greece, Israel, Jordan, Lebanon. 


Syria, Turkey]. Walls, 


pronounced slopes, at elevations from 0 to 


rocks. 
200 m. 


often in the vicinity of human dwellings. Figure 4. 


Morocco, Spain. 


Paratypes. ALGERIA. Chiffa. Blida, Davis 59527 
(E). EGYPT. S.l. N. Tadmor € 4. Shanda S-420 (E); Sinai, 
Bove 
jl 


5733 (MUB). ISRAEL. E Du Js 9/99/71 
(C); Kfar Gile D y Ud Curle 65 e Manara, C. M. Curle 
143 (E): Mount Gilboa, Davis 4667 (E); Wadi Qelt. Davis 
Yarmuk. Davis 4604 (E). JORDAN. South 
of Rum Rest ous Jallad et al. 7636 (Ex Wadi 
ji hi 5856 (K). LEBANON. Beirut, 
- DUM 1999, Inocencio 00025 

a & a 60049 (MUB): 

a & Obón 1 D : Maaraba, Rivera 
MU & Obon 60051 
B) TURKEY 


s.n. (E 


Adana, E K E^ ils 11 
Antalya, Smith 4 (K). 


Tin Haytop 9701 (E): 


DISCUSSION 


This paper introduces changes in the Capparis sect. 
Capparis taxonomy adopted by authors such as Zohary 
(1960). who organized the diversity within this section 
around two core species, viz. Capparis spinosa and C. 
ovata. Most of the endemic taxa that previously were 


subordinated to one or the other of the above species 


at the rank of variety or subspecies have been here 


recognized as species or subspecies. One of the 


reasons for the synthetic approach of Jacobs (1965) or 


Zohary (1960) is the relative frequeney of intermedi- 


— 


ate individuals in herbarium specimens that obscure 
the clear distinction among species. We refer to most 
of these as hybrid individuals, as revealed by our 
fieldwork in the Iberian Peninsula and North Africa 


2001) 


Hybrids have been reported from different areas 


(Inocencio, 


in which presumably hybrid swarms occur between 


two different Capparis species growing together. 


Hybrids are frequent in Iraq and neighboring 
countries of the Near East (Blakelock € Townsend, 
1980), 


The most 


shadowing the distinction between species. 
relevant 
is C. 


of the genus. It occurs spontaneously ir 


interspecific hybrid, for its 


economic uses, spinosa, which is also the type 


“species” 


populations of C. orientalis growing close to those 


of C. sicula in the western Mediterranean (Inocencio, 
2001). 


into cultivation. There have been no reported in- 


Only this nothotaxon has been widely taken 
tersectional hybrids within Capparis subgenus Cap- 


paris, although Capparis ovata subsp. myrtifolia 
seems to be an intermediate between C. ovata subsp. 
and C. 


their hybridogen. Capparis ovata is here restricted to 


ovata inermis and, therefore, presumably 


the Algerian type and those populations closely 
related morphology that extend from Morocco to 
Chad. 

The 


Turanian 


arge complex of Mediterranean and Irano- 


taxa formerly subordinated to Capparis 


spinosa is here combined under C. sicula because 
this is the name available according to the principle of 


priority. 


Literature Cited 
Ali, S. & S. 
8. Ali 


University, 


Jafri. 1977. 


(editors). 


Jafri & 
Faateh 


Capparaceae. Pp. 1-20 in 5. 
Flora of Libya, Vol. XII. Al 
Tripoli. 

2005. CBD Strategy and Action Plan—Tajiki- 
stan, Part IV. [English version]. <http://www.biodiv.org/ 
doc EA nbsap-01-p04-en.pdf>. 


Anonymous. 


Baul J.. J. Cherler 8 D. Chabrey. 1651. Historia 
a ntarum universalis, Vol. Il. Yverdon 
Benchelah, . (ng RE M. Maka & C. Ouahes. 2000. 


Fleu ra 1 5 Ed. Ibis Press, Paris 
11 9 G. & J. 1 1862 

‘ol. I. Reeve, 
von Bieberstein, 


Vol. IL 


Coen plantarum, 
London. 
A 


. 1808. 
Typis Academicis, Charkouiae. 
1819. 
s Ac a micis, Charkouiae. 
Hake 180 k, .& C C. 

0 15 C. C. Tow 


Flora Taurico-Caucasica. 


Flora Taurico-Caucasica, Vol. HI. Suppl. 


Townsend. 1980. A 


nsend & E. Guest (editors). Flora 
" uma ‘ol. IV. Pai l. Ministry of Agriculture & Vier 
Reform, Baghdad. 


148 


Annals of the 
Missouri Botanical Garden 


Bobrov, E. G. 1939. —14 


Capparidaceae. 


in V. I. 
Izdate’stvo 


Komarov (editor). Flora S.S. S. HR. Vol. E Il. 
Akademii Nauk SSSR, Leningrad- Moscow. 
1970. 5 Pp. 3-13 in V. L. Komarov 


(editor), Flora of the l 


. Vol. VIEL Ter ael Program for 
Scientific 1 1 e 
Boccone, P. 1674. leon. et deseript. Rarior. Plantar. Sicil. 


p et Italiae. Theatro Sheldoniano, Oxford. 
Une modification de la 
Ellenberg & 


2271 8. 


bene C. & Aeschiman. 1981. 


classification formes Pie ge 
27 


Mueller- Dou Candollea : 
\ | 


x 


Candolle, A. P. de. 1824 


= 


1 1 Pp. 245-2 
PI | 


Prodromus systematis naturalis regni. vegetabilis. 2 
Treuttel et Würtz. Paris. 


Castroviejo, S., M. Laínz, G. López González, P. Monserrat. 


F. Munoz rus J. Paiva & E 1993. Flora 
Ibérica. Vol. III. CSIC. Madrid. 

Charco, x 200 a ues de Arboles \rbustos del Norte 
Africa. E A ones Cultura Hispanic: a, Madrid. 

Coste, H.! 900. Fl. deser. France, Vol. I. Paul Klincksieck. 
Paris. 

Cummings, M. P., J. M. Nugent, R. - Olmstead & J. D. 
Palmer. 2003. ge netic anc m reveals five in- 


dmg ist gene rbcL to the 


Genet. 43: 


5 8 IN transfers of t 
mitochondrial genome in angiosperms. Curr. 
131—138. 

De ne A. 1812 


2. Imprime rie pus riale, 


. Deseription de l'Egypte, Histoire Naturelle. 


Paris. 


Desc: . L. 1801. Traité des Arbres et Arbustes. Ed. 2. 
ol. I. 1 
Fedtschenko, B. & O. Fe "e he $m px E mius n Flora 


Beih. Bot. 'albl. 


Turke stanica. 


Fici, S. 20( l. Intraspe cific variation 7 e san bre E 
in Cami spinosa L. (Capparaceae). PI. . Evol. 2 
123-1 f 

Fragman, O., R. & P. Christodoulou. 2001. Flowers of 
the cH rn 115 e ranean. Ruppell, A.P.G., Gertne 


Graham, J. 1839, . PL Bombay. The Government Press, 
ombay 
Greuter, W Burdet & G. Long. 1984. in Med-Checklist. 
Vol, L Conservatoire et Jardin Botaniques de la Ville de 
Geneve, Genes 


Guerau, C. & N. pM 1981. Nova Aportació al coneixe- 


ment de les plantes d'Eivissa i Formentera. Institut 
d pc Eivissences, Ibiza. 
s K. lsma & II. II. Hus. 2002. Phylogeny of 


Capparaceac hase : on chloroplast 
842. 


sequence data. Amer. 


Herbarium Willdenow. 


Hiepks, P. 1972. TER tical Index 
Museum Botanicum Berolinense. IDC, Ag Zug. 
Higton, = N. X J. royd 1991 Variation i in Capparis 


spinosa L. Bot. J. 
) 


Inocencio, 2001. ( aracterización Capparis subgénero 


( 
Unive sidad de Murcia, 


Capparis (Capparaceae). Murcia, 
(Ph.D. e sis.) 
. Rivera, F. Alcaraz & F. Tomás-Barberán. 2000. 
F woe content of commercial capers (Capparis 6 5 
C. sicula and C. orientalis). Eur. Food Res. Technol. 212 
70-74. 
b. Alcaraz, F. Calderón, €. Obón & D. Rivera. 2002. 


The use of 


determine 2 botanical d ge E ‘al origin of capers. 
339 


floral characters in Capparis sect. Ca apparis to 


Eur. Foo p ol. 3 35— : 

—— . p. River ra, - Ale araz, F. Fay & M. 200 
AFLP finge 1 in ari: 1. subes nus Gaya 
(Capparaceae) related to the commercial sources of capers 
Genet. Resources & C Op Evol. : 137-144 


Jacobs. M. 1965. The Genus Capparis from the Indus to the 
Pacif - Blume 'a 12(3): 385-541. 

Jarvis, E „H. Barrie; D. M. Allan & J. L. Reveal. 1993. A 
List of hi innaean Generic Names and Their Types. Koeltz, 
Königstein. 

Lack, W. 1997. 
Guidelines n its use. Taxon 46: 253-206: 

Lamarck, J. B. 1783. Method. Bot. 
oucke. 


Lew alle, 
dle 


The Sibthorpian Herbarium. at Oxford— 
3, 


Encycl. Vol. J. Panck- 
A 
J. & N. Montfort. 


] 
Verlag 


1997. Fleurs Sauvages du Maroc. 
gessellschaft, Rossdorf. 
Linnaeus, C. 17523. Species Plantarum. Stockholm. 
. Genera Plantarum. Stockholm. 
Lonicer, A. 1679, Kreuterbuch. Ulm. 
1 J. 1990. Fl. E 
-ommission for Wildlife Conservation. 
Migahid, “ M. 1988. 
S. Unive reity, Riyadh. 
Narvi. M. & Kli. 1973. FL West Pakistan. 
University d Kari hi. Rawalpindl, 
Olivier. G. A. 1801-1807 
l'Égypte et la 


Arabia. 
Riyadh. 


3rd Ed. | 


astern Saudi National 


Flora of Saudi Arabia. 


Vol. XXXIV. 


Voyage dans l'Empire Othoman, 


Perse, fait par. ordre du gouvernement 


pendant les six premieres. années de la république. 


3 Vols.. H. Agasse, Paris 
Ozenda, P. 1991. Fl. Sahara. Éditions du CNRS, Paris. 
Plitmann, U., C. Heyn, A. Danin A. Shmida. 1983. 


Massada, Tel / 
. Stainton. 1984. 
Oxford Univ. PIED. Delhi. 
Pritzel, G. A. 372. 
Broc 71 9 Leipzig. 


Pictorial F lora Israel. 


Polunin. O. F es n 111 the Himalaya. 


Thesaurus Literaturae Botanicae. 
. Obón & I 
. 1 Panouri 
Gap in the western Mediterranean. Pp. 451— 


mus ra. D. Alcaraz. C. Inocencio, Carreño. 


Y 
2 
O 


n axonomic study of sect. 
155 in 
Alexander (editors). 

Third 
. Kew. 


Carreño. A. 


Andrey . €. Leslie & C. 
Plants: 
Gardens 


— 7 


um " Culpvated International 


ii rn Royal Botanic 


. Inocencio, €. Obón, E. Reales & F. 


1 55 2002. Archaeobotany of capers (Capparis) 
(Capparaceae). M Hist. Archacobot. E 295-313. 
Friis, C. Inocencio, C. Obón, F. Alcaraz & A. 


Reales, 90 The 
Forssk.. C. 
(Capparaceae). 


typification. of 5 inermis 
Ve p oe 
7 0 : 307 
1 & AS 
Medicinal 


sinaica cartilaginea Decne 
2003b. Review 
subgenus 
3 


—. C FN 
; Food and 
. Bot. 57: 5 5: 
F. nen 2000. 


application 


S 

= 

— 
S 
a 
= 


Capparis 1 1 L. (Capparaceae N 
d of the 


sicula Neill. and C. 


of the name iue s Ivpifi- 


cation of C. orientalis Veill. Taxon. (in 


Soltis; D. E. 


S: Soltis. 
Parallel 


ss). 
Rodman. J. E., P. 
1998. 


synthesis 


K. J. Sytsma & K. G. 


Karol evolution of glucosinolate bio- 


inferred from congruent nuclear and plastid 


gene phylogenies. Amer. J. Bot. 85: Ere 

Royle, J. F. 1839. III. bot. Himal. Mts.. Vol. I. Allen, 
London 

Smith, J. E. 1825. Flora Graeca. Vol. Richard Taylor, 


Stafleu, F R. S. Cowan. "ia oe Literature, 
Vol. | LE We g. 94: 36. 
& —— — . 1979. m D ne Vol. II: He 
Le. Regnum Veg. d 991. 
— Y ————. 1981. Taxonomie Literature, Vol. III 
Lli-O. Regnum Veg. 105: 1—9% 


Volume 93, Number 1 
2006 


Inocencio et a 


149 
Revision of TEN Sect. Capparis 


a : ;L 


— — —., 1983. Taxonomic Literature, Vol. IV 
Sak. Regnum Veg. 110: 1-1214 
N 1985. omo Literature, Vol. V: Sal- 
Ste. Regnum Veg. 112: 1-1066. 
X 1988. Taxonomic Literature. Vol. VII: 
Lis de BT V. eg. 116: 1-653. 


Y 
Fickholm, V. 1974. 
Beirut. 

Takhtajan, A. 1966. Flora Armenii, Vol. 
d SSR, Erev 


"Shade mts” FI. Egypt. Cairo University, 


V. Akademii Nauk 


F o 
aon bre ss. Berkeley 
Townsend, C. C. € E. 


Re gions of the World. 


Univ. 
Guest. 1980. FI. Iraq. Vol. IV(I). 


Ministry y Agriculture and Agrarian Reform. Baghdad. 


Valdés, B., S. 0 ra & E. 
Occide e Vol. 


Fernández. 1987. Fl. Andalucía 
^ . Linnaeus species as a 
E net. Seleke. 
Breed.) 26(3): 109-13 
Willis. J. €. 1988. A o of the Flowering Plants and 
Kerns. Eighth Edition. Cambridge Univ 
M W. 1794. Medical 


Londor 


sstem. Trudy 


as 
(Bull. appl. Bu Gen. Pl. 


Cambridge. 
James Phillips, 


. Press, ( 


Botany. 


Zohary, M. 1960. The species of Capparis in the. Mediter- 
ranean and the Near Eastern Countries. Bull. Res. Council 
Israel. 8D: 29-64. 

1966. Fl. Palaestina. Vol. I. Plates. The 


Israeli 
Academy of Sciences and Humanities, Jerusalem 


THE CARDUEAE (COMPOSITAE) Alfonso Susanna,” Núria ne dubio qd Oriane 
REVISITED: INSIGHTS FROM Paid Roser Vilatersana,* and Teresa 
ITS, trnL-trnF, AND matK i 

NUCLEAR AND CHLOROPLAST 

DNA ANALYSIS"? 


oo 
ÅBSTRACT 


The new outline of relationships in basal branches of the family Compositae Giseke confirms that the sister group to the 
Tare nA 
Kostel. This as implie s that the monophyly of the Cardueae must be reassessed on a molecular basis. Moreover. 
collections in recent years allow us to extend our sampling to 70 of the 74 genera of the tribe. We performed a new 1 
study of the tribe using one nuclear region (ITS) and two c hloroplastic markers H and matk) in addition to a more 
appropriate outgroup. Our results confirm that the Cardueae is a natural group but indicate some changes in subtribal 
delineation: the subtribe Cardopatiinae Less. is recognized and some genera are moved to other subtribes ( 57 Boiss., 
Nikitinia iin, Syreitschikovia Pavlov, and the We ina r group). A recapitulation of a number of intere sting questions 
that remain unresolve d in the classification of some large genera is presented. 


tribe Cardueae Cass. are not Mutisieae Cass.. but rather a group of African genera now classified as the tribe 1 


= 


ey words: vardueac, Compositae, genus delineation, ITS. mark, phylogeny, tribal delimitation, trnL.-trnF. 


RESUMEN 


El nuevo esquema de las relaciones en las ramas basales de la familia C «omposttae Giseke ha confirmado que el grupo 


r 


hermano de la tribu Cardueae Cass. no es la tribu Mutisieae Cass., sino más bien un grupo de géneros africanos que ahora se 
clasifican como tribu 1 b 


Pare een ae Kostel, Este cambio implica que la monofilia de las Cardueae establecida sobre bases 
moleculares debe confirmarse. Además, nuestras nuevas recolecciones en los últimos años hacen que nuestro muestreo se 
extienda a 70 de los 74 géneros de 


1 
a tribu. Hemos llevado a cabo un nuevo estudio molecular de la tribu usando una región 
nuclear (ITS) y dos c loroplásticas, además de seleccionar un grupo externo más adecuado. Nue ie resultados confirman que 
las Cardueae son un grupo natural, pero aconsejan algunos cambios en la delimitación subtribal: el reconocimiento de 
subtribu Cardopatiinae Less. y el cambi 


a 
de algunos géneros a otras subtribus (Myopordon Boiss., Nikitinia Ijin, 
Syreitschikovia Pavloy y el grupo Xeranthemum V). Presentamos también una . ion sobre ciertas cuestiones 


interesantes que quedan sin re 1 r en la clasificación de algunos grandes genere 
— ꝛꝛñꝛ I 


INTRODUCTION a classification of two subfamilies and 16 tribes that 


OVERVIEW OF COMPOSITAE SYSTEMATICS ganen panera n P e us 
two latest revisions of the family, toward the end of the 

The systematics of Compositae is marked by three 20th century (Dittrich, 1977; Bremer, 1994), followed 
milestones, each one involving deep changes in the Hoffmann's classification. The third set of large-scale 
classification of the family. Since the history of this changes was produced by the introduction of methods 
classification has been revised in depth in Funk etal. based on DNA analysis, First came the pioneering 
(2005), here we give only a short summary. The first study by Jansen & Palmer (1987) using. cpDNA 
attempt to classify Compositae was made by Cassini restriction site polymorphisms, which led to the 
(1819), who defined 20 tribes. A more synthetic description of a third. subfamily, Barnadesioideae 
system was proposed by Bentham (1873) and, soon (Benth. & Hook. f.) K. Bremer & R. K. Jansen: this 


after, Hoffmann (1894). Both authors proposed — proposal was reflected in Bremer (1994). Second, the 


! Financial support [rom the Dirección General de Enseñanza Superior, Spain (Project PB 97/1134), Ministerio de Ciencia y 
Tecnologia, Spain (Projects PB. BOS2001-3041-CO2-02 and PB. BOS2002-11856-E ) and Generalitat de Catalunya (“Ajuts 
a grups de recerca consolidats” 19995GR00332 and 20015GR00125) is grate fully acknowledged. We thank Christine Hidalgo 
for her help in sequencing the trnl.-trnF region and Miquel Veny for keeping the collections of living plants. We also thank 

Santiago Ortiz, University of Santiago, Spain. who provided mate rial and sequences of Br aie Oldenburgia, and 
lios The collaboration of the botanical gardens listed in Table 2 is also acknowle dged. We thank Vicki Funk and 
one anonymous reviewer for their valuable suggestions for improving the manuse ript. 

“The editors of the Annals thank Sophia Balcomb for her Editorial contribution to this paper. 

t Botanic Institute of Barcelona (CSIC- Ajuntament. de Barcelona). Pg. del Migdia s. n... E-08038 Barcelona. Spain. 
asusanna@ibb.c 


4 


ANN. Missouri Bor. Garp. 93: 150-171. PuBLISHED ON 31 May 2000. 


Volume 93, Number 1 
2006 


Susanna et al. 151 


Cardueae (Compositae) 


Table 


Outline of two different proposals of classification of Compositae (only basal branches). 


BREMER (1994) 

BARNADESIOIDEAE (Benth. & Hook. f.) 

K. Bremer & R. K. 

e BARNADESIEAE D. Don 
CICHORIOIDE. un Chevall. 

* MUTISIEAE 

e CICHORIE v 

* CARDUEAE 

* VERNONIE 15 

* LIABEAE 9 

„ ARCTOTEAE Ca 
ASTEROIDEAE (Cass.) Lindl. 


17 tribes) 


Lam. & DC. 


(3 subfamilies. 


BARNADESIOIDEAE (Benth. & Hook. f.) 
isen 

e BARNADESIEAE 
Stifflioideae clade (provisional) 

e Stifftieae clade (provisional) (S. AMERICA, AFRICA) 
MUTISIOIDEAE 

e MUTISIEAE A) 
GOCHNATIOIDE 5 (Benth. & Hook. f.) Panero & V. 
HECASTOCLEIDOIDEAE 

e HECASTOCLEIDEAE 
CARDUOIDEAE 

e DICOM 

. 1 i i AE 


Jansen Jat 


PANERO & FUNK (2002) 


K. Bremer & R. K. 


D. Don 


w ) Lindl. 

‘ass. (S. AMERIC 

A. Funk 
Panero & V. A. Fun 

Panero & V. A. Funk 

ex Sweel 

SAE Panero & V. A. Funk 

Kostel. 


Cass. 


* CARDUE 


latest and more revolutionary study by Panero & Funk 
(2002) analyzed sequences of nine chloroplast regions 
across the entire family and proposed a new classi- 
fication with 11 subfamilies and 35 tribes: this was in 
some ways closer to Cassini's analytical views than to 
synthetic approaches. The dramatic differences. be- 
tween Bremer's (1994) and Panero and Funk’s (2002) 
classification are illustrated in Table 1 (only the basal 
groups are shown). The high statistical support for the 
latter and its sound correlation with morphology leads 
us to believe that the new classification of Compositae 
is near to being definitive. 


THE TRIBE CARDUEAE 


Cardueae Cass. is one of the largest tribes of 
Compositae, with ca. 2500 species. Previous studies 


DNA both 


(Susanna et al., 


based on sequence analyses, nuclear 


1995) 


nuclear (Garcia-Jacas et al., 


and combined chloroplast and 
2002), confirm Cardueae 
as monophyletic. However, the new classification 
shows that our previous outgroup choice was not the 
best choice. In the classic system of Compositae (e.g., 
1977: Bremer. 1994), Cardueae were 


classified in subfamily Cichorioideae Chevall., close 


Heywood et al.. 
to tribes Cichorieae Lam. € DC. and Mutisieae Cass. 
(Table 1). Therefore, in our first nuclear-DNA-based 
1995), 
f one Cichorieae (Tragopogon L.) 
(Ainsliaea DC.. 
& Coss.). 


Ainsliaea, 


phylogeny (Susanna et al., the outgroup was 


and 


^ 


composed 


Mutisieae Gerbera Ie, and 


Benth. 


we replaced 


three 
Warionia In Garcia-Jacas et al. 
(2002). 


Warionia because of the increasing difficulties i 


Tragopogon, and 


~ 


1 
aligning the ITS region, and we used two Mutisieae as 
Mutisia V. f. 


according to the new classification by Panero & Funk 


outgroups, Gerbera and However. 


(11 ON 36 tribes) 


(2002). tribe Cichorieae (Tragopogon) is derived 
relation to Cardueae; Ainsliaea and Warionia do not 
belong to Mutisieae but to Pertyeae Panero & V. A. 
Funk and Gundelieae H. Rob. & Brettell, respective- 
ly. both tribes also derived with regard to Cardueae: 
and Gerbera and Mutisia are placed in Mutisieae 
sensu stricto, phylogenetically far from Cardueae 
(Table 1). 
will always be monophyletic, and monophyly of the 
The 


outgroup should be chosen from the clade formed by 


With these outgroup species, Cardueae 


tribe has always been a controversial issue. 


Tarchonantheae plus the genus Oldenburgia L., 
that 
Cardueae with bootstrap support values of 100% in 
Panero and Funk (2002). In fact, ( 
Panero & V. X. Funk, Tarchonantheae (plus 


compose a monophyletic 


tribe 


a clade appears as the true sister group to 


‘ardueae, Dicomeae 
and 
subfamily, 


Oldenburgta 


Carduoideae Cass. ex Sweet, also with the highest 
statistical support (Panero € Funk, 2002). Morpho- 
logical connections between Cardueae and the rest of 


subfamily Carduoideae are, however, unknown to date. 


TRIBAL LIMITS OF CARDUEAE 


1819). 


tribes: 


In the earliest classification (Cassini, 


present Cardueae were divided in three 
Echinopeae, Carlineae, and Cardueae, the latter with 
two subtribes: Carduinae and Centaureinae. Bentham 


(1873) and Hoffmann ( 


three tribes in a single tribe 


894) proposed grouping the 
Cardueae that held four 
subtribes: Echinopinae (Cass.) Dumort., Carlininae 
(Cass.) 


(Cass.) Dumort. This was a conservative approach that 


Dumort., Carduinae Cass., and Centaureinae 
was generally accepted for a very long time. However, 
discussion on the status of Echinopinae restarted 


when Wagenitz (1976) proposed the segregation of the 


Annals of the 
Missouri Botanical Garden 


— . . ů ö —-— . ̃ —ꝛñ—— —ꝑ.ñ . —— 


subtribe as a separate tribe, Echinopeae. Dittrich 
(1977) returned to Cassini's early views and proposed 
the restoration of Echinopeae and Carlineae. Finally 
Bremer. (1994) 


proach with only one tribe, Cardueae, which, accord- 


reintroduced the conservative ap- 


ing to our molecular studies, is a better solution 


(Susanna et al., 1995: Garcia-Jacas et al., 2002). 


SUBTRIBAL CLASSIFICATION 


Within Cardueae, there is general agreement in 
accepting four groups, regardless of the rank (tribe or 
subtribe) adopted. Three subtribes are natural (Carli- 
ninae, Echinopinae, and Centaureinae) and the fourth 
(Carduinae) is a paraphyletic assemblage (Garcia- 
Jacas et al., 2002) 

Subtribe Carlininae is sister to the rest of the tribe. 


A striking and probably plesiomorphic character is 


the presence of true ray florets in at least one genus o 
Carlininae, Atractylis L., while remaining subtribes 
have only disk florets. Capitula are usually subtended 
by pectinate-pinnatisect leaf-like bracts; corolla lobes 
are very short, only 1-3 mm long: and the pappus has 
long, plumose bristles, often connate at the base 


forming broader, robust scales (Susanna € Garcia- 
eo 


Jacas, in press). 

Subtribe Echinopinae is easily characterized by its 
second-order inflorescences (uniflowered | capitula 
Our 


Echinopinae 


clustered in a large synflorescence). latest 
that 


should also include the genera of the Xeranthemum 


molecular phylogeny indicates 
group and we previously proposed that the small 
heads of the genus Xeranthemum and allies could be 


: n 
SVIIIIO Ie 


interpreted as reduced scences (Garcia-Jacas 
et al., 2002). 
Subtribe Carduinae is a paraphyletic complex of 


genera with some well defined groups (Arctium L. 


group, Onopordum L. group, Saussurea DC. group, or 
the thistles) together with genera of problematic 
ascription like Berardia Vill. or Staehelina L. All 
the genera of Carduinae have basal or basal-abaxial 
insertion areole of achenes and, usually, a simple 
pappus, and are often spiny, 

Finally, subtribe Centaureinae is the most derived 
group and is characterized by achenes with lateral- 
adaxial insertion areole, a double pappus. and, with 
few exceptions, unarmed leaves. However. examining 
the 


represents a challenge, because differences lie ii 


limits between Carduinae and Centaureinae 


microcharacters of the achene and pappus that are 


difficult’ to observe in incomplete or immature 
herbarium materials. The examples of Nikitinia and 
Syreitschikovia illustrate these difficulties (Susanna et 
al., 2002) and the ascription of these and other genera 


should be checked against a molecular phylogeny. 


(2002) 


maybe a fifth subtribe could be recognized, Cardopa- 


n Garcia-Jacas et al. we suggested that 
tiinae Less., with two genera: Cardopatium Juss. and 
Cousiniopsis Nevski. Cardopatiinae were placed in an 
intermediate position between Carlininae and the rest 
of the tribe (Garcia-Jacas et al.. 2002). However. we 
have postponed the restoration of this subtribe until 


more unambiguous evidence has been collected. 


GENERIC LIMITS IN TRIBE CARDUEAE 


Other points of interest are genus affinities and 


limits i 


1 Cardueae, a tribe with some of the largest 
genera of the family. Regarding genus affinities. on 
the basis of morphology and partial molecular studies, 
the two largest subtribes (Carduinae and Centaurei- 
nae) were subdivided into informal groups (Susanna & 
Garcia-Jacas. in press), which should be checked 
against a more comprehensive molecular phylogeny. 
As to genus limits, in our latest revision of Cardueae 
(Susanna & Garcia-Jacas, in press) we adopted a broad 


generic concept for Cousinia Cass. (600 species), 


x, 


Jurinea Cass. (200). and Saussurea (400) because o 
the lack of recent systematic revisions for all three. 
Recently, on the basis of a partial study of DNA 
sequences and achene morphology, Raab-Staube 
(2003) proposed the restoration of two small genera, 
Frolovia (DC.) Lipsch. and Lipschitziella Kamelin, and 
described a new genus. Himalaiella Raab-Staube. all 


of these within the Saussurea group. 


SCOPE AND AIMS 
With 


sampling covers 


the addition of new materials. our DNA 
70 of the 74 accepted genera of 
Cardueae: only Aneathia DC. (Carduinae. Central 
Asia), Centawrodendron Johow (Centaureinae, Juan 
Fernández archipelago), Goniocaulon Cass. (Centaur- 
einae, India and East Tropical Africa). and Takei- 
kadzuchia Kitag. € Kitam. (Carduinae. Mongolia) are 
absent. However, the position of these within the tribe 
and their subtribal ascription has never been 


challenged on a morphological basis (Susanna & 
Garcia-Jacas, in press). To test our broad generic 
concept. we included the genera Frolovia, Lipschit- 
ziella, and Modestia Kharadze & Tamamsch.. which 
we submerge 


? 


in Jurinea, and Anura (Kult) Tscher- 


neva and Tiarocarpus Rech. f. which we had 
previously considered in Cousinia (Susanna & 
Garcia-Jacas, in press). We also included the 


published sequence of the recently described genus 
Himalaiella. Vor this wide representation of Cardueae, 
we completed the ITS and matK regions and. in view 
X the 


low resolution of basal 


^ 


groups in previous 


analyses, we added a new marker. Low resolution in 


Volume 93, Number 1 
2006 


Susanna et al. 153 


Cardueae (Compositae) 


many molecular phylogenies may be solved by adding 
more data to DNA sequence matrices as discussed in 
Panero & Funk (2002). We used a chloroplast marker, 
the nn which is widely 
utilized in Compositae (Bayer & Starr, 1998: Liu et 


al., 2002: Oberprieler, 2002; Panero & Funk, 2002). 


intergenic region, Is 


Our gvals were to: 


; from the 


(a) verify S of Cardueae using species 
sister clade Tarchonantheae and Oldenburgia as an 
»utgroup: 

(b) c clari subtribal classification and define the position 
of Cardopatiinae, which might constitute a fifth sub- 
tribe: 

(c) examine whether the informal species groups defined 
in subtribes Carduinae and Centaureinae are natural 
and check the 1 position within these groups 
of genera not included ii previous studies: and 

(d) verify the suitability of a broad generic concept in 


certain large genera of Cardueae by analyzing species 
from genera that we had previously rejected on the 


basis of morphological characters. 


MATERIAL AND METHODS 
PLANT MATERIAL 


Sampling was defined on the basis of Garcia-Jacas 
et al. (2001), Garcia-Jacas et al. (2002), Susanna et al. 
(2003). & Garcia-Jacas (in press). in 
order to of tribe 

Cardueae. Thirteen accepted genera (Amphoricarpos 
Vis., Karvandarina Rech. f 
& Tamamsch., Lamyropsis Khatradze) Dittrich, Myo- 
pordon Boiss., Nikitinia. Olgaea Iljin, Plagiobasis 
Schrenk. Polytaxis Bunge, C. Winkl., 
Syreitschikovia, Siebera J. Gay. Tricholepis DC., and 


Tugarinovia Ijin) are sequenced here for the first 


and Susanna 


represent most the genera of 


.L amyropapplttis Knor r ing 


Russowia 


time. Six other genera that were not accepted in our 


latest revision of the tribe (Susanna & Garcia-Jacas, in 


press) are Aegopordon Boiss.. Anura, Frolovia, 
Lipschitziella, Modestia, and Tiarocarpus. Two more 
Dolomiaea DC. and  Himalaiella, were 


genera, 
obtained from sequences published elsewhere. Three 
outgroup species were chosen, two among Tarcho- 
nantheae and another from the genus Oldenburgia, 
because the tribe and the genus form the sister clade 
to Cardueae (Panero & Funk, 2002). Many of our ITS1 
and ITS2 sequences from previous studies (Garcia- 
2001. 2002) have been completed with 
the sequence of the 5.8 S gene, and some of our old 
1995) were re- 


Jacas et al. 


manual ITS sequences (Susanna et al., 
sequenced confirmed by automatic sequencing. 
Both previously published (Garcia-Jacas et al., 2002) 
and new sequences of the matK gene were used in this 
analysis. All the rnit sequences analyzed are 
with the exception of Dolomiaea (from Liu, 


of Saussurea (from 


new, 
unpublished) and some species 
Raab-Staube, 2003). The number of new sequences is 


283 for a total of 466. The origin of the samples and 
GenBank 


Table 2. 


their acc ession numbe IIS are give n in 


DNA EXTRACTION, AMPLIFICATION, AND SEQUENCING 


DNA was extracted following the 


as 


Total genomic 
miniprep procedure of Doyle € Doyle (1987 
modified by Soltis et al. (1991) and Cullings (1992), 
from silica gel-dried leaves collected in the field or 
from fresh leaves of plants cultivated in the Botanic 


of 


material was used. 


Institute Barcelona. In some cases, herbarium 


cpDNA tral tral REGION STRATEGIES 


The plastid irnL-trnF region includes the tral 
intron. the 3^ tral (UAA) exon, 
spacer between trnL (UAA) and trnF (GAA), that were 


Universal primers 


and the intergenic 


amplified and sequenced together. 


trnL-c. forward, and trnL-f, reverse (Taberlet et al., 
1991), were used for amplifying the nE region. 
In some cases, irnL-d, reverse, and trnL-e, forward, 
were used. Polymerase chain reaction (PCR) was 
conducted in a thermocycler (MJ Research PTC 100). 
The 
minute 35 seconds, followed by 80°C during which 
Ecogen S.R.L., Barcelona, 


de- 


CR procedure included a warm start at 95 € for 


the polymerase (Ecotaq. 
Spain) was added, with 34 cycles of ] 
naturation at 93 C, ! 
extension at 72°C. and a final 10 min. extension at 


min. 
min. annealing at 58 C. I min. 


C. 

PCR products were cleaned with a QlAquick PCR 
Purification Kit (Qiagen Inc., Valencia, CA) and 
sequenced with the trnL-c and trnL-f primers. Direct 
the amplified DNA segments was 
performed BigDye 
Sequencing v3.1 (PE Biosystems, Foster City, CA), 


following the protocol recommended by the manufac- 


sequencing of 


using the Terminator Cycle 


turer. Nucleotide sequencing was carried out at the 
Científico-Tecnies of the University of 
Barcelona on an ABI PRISM 3700 DNA analyzer 
(PE Biosystems, Foster City, CA). 


— 


Servels 


cpDNA MaK GENE STRATEGIES 


We have sequenced the first 1000 base pairs at the 
the 


5. because this includes most of 
variability in the matK plastid gene (Hilu & Liang. 
1997). Partial matK was amplified by PCR with the 
(Johnson & Soltis, 1995) and 
l., 2002). The PCR pro- 
cedure included a warm start at 94 C. for 
20 sec., followed by 80°C during which the poly- 


merase (Ecotaq. Ecogen S.R.L., Barcelona, Spain) was 


end, part 


primers. trnK-710 F 


AST-1R (Garcia-Jacas et ¿ 
min. 


Annals of the 
Missouri Botanical Garden 


Table 2. Origin of the 


sequences are boldfaced ). 


„ herbaria where the vouchers are 


deposited, and GenBank accession numbers (new 


Species 


Acantholepis orientalis Less. 


Acroptilon repens ( DC. 


Vfredia acantholepis Kar. & Kir. 


Alfredia cernua (IU) Cass. 
Vfredia nivea Kar. & Kir. 


Amberboa turanica Iljin 


Imphoricarpos autariatus 


Blecic & Mayer 


Amphoricarpos exsul. O. Schwarz 


Iretium lappa L. 
Irctium minus Bernh. 


Atractylis cancellata. 1.. 


Atractylis carduus (Vorssk.) Christ. 


VMractylis humilis U 


Atractylodes japonica Koidz. 


ex Kitam. 


Berardia subacaulis Vill. 


Brachylaena discolor DC. 


les i nitens (M. Bieb. 


Will d.) tey. 


B Pe (I.) Pers. 


Carduncellus duvauxti Batt. & Trab. 


Carduncellus mareoticus (Delile) 


lanelt 
Carduus carlinoides Gouan 
Carduus defloratus | 


Carduus pycnocephalus U 


Carlina acanthifolia All, 
Carlina falcata Svent. 


Carlina gummifera (L.) Less. 


Carlina lanata L. 


Carlina macrophylla (Desf) DC. 


Carlina vulgaris |. 


Carthamus creticus L.. 


Carthamus leucocaulos Sibth. & Sm. 


Carthamus oxyacantha M. Bieb. 


Carthamus tinctorius L. 


Carthamus turkestanicus Popov 


Centaurea alba |. 


Centaurea amadanensis Sch. Bip. 


Voucher 
Garcia-Jacas et al.. 2002 
Garcia-Jacas et al.. 2001 


Kazakhstan, Susanna 


Garcia-Jacas et al., 2002 

Kazakhstan, Susanna 
2090 et al. (BC) 

Garcia-Jacas et al., 2001 


Serbia & Montenegro. 
Stevanovic et al. 
1.10.91 (BC) 

uike , Susanna 
2250 et al. (BC) 

Garcia-Jacas et al., 2002 

Garcia-Jacas et al., 2002 

Garcia-Jacas et al., 2002 

Garcia-Jacas et al., 2002 

Garcia-Jacas et al.. 2002 


Garcia-Jacas et al., 2002 


Garcia-Jacas et al.. 2002 

South Africa, Ortiz 20.3.02 
(NBG) 

Garcia-Jacas et al., 2001 


Garcia-Jacas et al., 2002 

Vilatersana et al.. 2000; 
Garcia-Jacas el 5 2001 

Vilatersana et al., 2000 


Garcia-Jacas et al., 2002 
Spain, Garnatje 18 (BC) 
Garcia-Jacas et al., 2002 


Garcia-Jacas et al.. 2002 
Garcia-Jacas et al., 2002 
Garcia-Jacas et al., 2002 
Garcia-Jacas et al., 2002 
Garcia-Jacas et al.. 2002 


Garcia-Jacas et al., 2002 
Vilatersana et al.. 2000 
Vilatersana et al.. 2000 


Vilatersana et al., 2000: 
Garcia-Jacas et al.. 2001 
Vilatersana et al.. 2000 


Vilatersana et al.. 2000 
Garcia-Jacas et al., 2000. 
2001 


Garcia-Jacas et al.. 2000 


ITS accession 


Irnl.-trnE 


accession 


math 


accession 


AY826222 
AY826223 
41326224 


AY826225 
41326220 


AY012275, 
AYO12311 
AY826227 


AY826228 


AY826229 


AY826232 

AF319052. 
\F319106 

AY826233 


AY826234 
AY826230 


AY820237 


41326238 
AY826239 


11140480, 
114048 
AY326240 
AY826241 
AF319111, 
11013528 
AY826242 
AY826243 
AY826244 
AY826245 
VE319062. 
AF319116 
AY826240 
AY826247 
VE 140400. 
VET40461 
AY826248 


11140458. 
111404509 
AY826249 
AY829446 


114058830, 
110588641 


AY772269 
AY772270 


AY772278 
AY772280 


AY77228 


— 


AY772282 
AY772283 


AY772284 


AY772285 
41772286 
AY772287 
AY772288 
AY772289 
AY772290 


AY772291 


AY772292 


AY785086 
AY013489 


AY013519 
AY785087 


AY785088 


AY785089 


AY013520 
AY013521 
AY013522 
AY013523 


AY013524 


AY013525 
AY785090 


AY013492 


\Y013526 
11013493 


AY013527 
AY 70500] 
AY013528 


AY013529 
AY013530 
AYOI353] 
AY013532 


AY013533 
AY785092 


AY013494 


AY785093 
AY013495 


Volume 93, Number 1 
6 


Susanna et al. 


Cardueae quam ome 


155 


Table 2. Continued. 
trnL-trnF math 
Species Voucher ITS accession accession accession 
Centaurea behen L. Garcia-Jacas et al., 2000, AY826250 AY772294 AY013496 
2001 
Centaurea benedicta L. France, Nancy Bot. Gard. AF058850, — AY013508 
(BC) AF058875 
Centaurea bruguierana (DC.) Garcia-Jacas et al., 2000, AY826251 AY772295 AY013497 
Hand.-Mazz 2001 
Centaurea calcitrapa L. Egypt, Susanna 1866 & AY826252 -— — 
Vilatersana (BC) 
Centaurea carolipauana Fern. Garcia-Jacas et al., 2001 AY826253 AY772296 AY013498 
asas & Susanna 
Centaurea cyanus L. Spain, Garcia- Jacas & AY826254 — — 
Susanna 2076 (BC) 
Centaurea depressa M. Bieb. Garcia-Jacas et al., 2001 AY826255 AY772297 AY013499 
Centaurea involucrata Desf. Susanna et al., 1995; AY826256 AY772298 AY013503 
Garcia-Jacas et al., 2001 
Centaurea kotschyi (Boiss. & Heldr.) Garcia-Jacas et al., 2000 AY829441 — -— 
Jaye 
Centaurea lagascana Graells Garcia-Jacas et al., 2001 AY826257 AY772299 AY013504 
Centaurea lingulata Lag. Garcia-Jacas et al., 2000, AY826258 AY772300 AY013505 
2001 
Centaurea linifolia L. Garcia-Jacas et al., 2000 AF058832, — — 
AF058857 
Centaurea macrocephala Muss.- Susanna et al., 1995 L35873 — — 
Puschk. 
Centaurea montana L. Susanna et al., 1995 135887 — — 
Centaurea polyacantha Willd. Susanna et al., 1995 L35878 — — 
Centaurea rhizantha C. A. Me Garcia-Jacas et al., 2000 AF058842, — — 
AF058807 
Centaurothamnus maximus Garcia-Jacas et al., 2001 AY826259 AY772301 AY013506 
Vagenitz & Dittrich 
Chardinia orientalis (L.) O. Kuntze Garcia-Jacas et al., 2002 AY826260 AY772302 AY013534 
Cheirolophus benoistii (Humbert) Susanna et al., AF045415, 
olub AF079942 
Cheirolophus mauritanicus Susanna et al., 1999 AY826261 AY772303 AY013507 
(Font Quer) Susanna 
Cheirolophus sempervirens (L.) Pomel Susanna et al., 1999 AF021150, == € 
AF021173 
iier e teydis (C. Smith) Susanna et al., 1995 AY826262 AY772304 A Y 785094 
X we 
"en m arvense (L.) Scop. Susanna et al., 1995 L35807 — — 
Cirsium echinus (M. Bieb.) Garcia-Jacas et al., 2002 AY826263 AY772305 AY013535 
Hand.-Mazz. 
Cirsium ochrolepidium Juz. Uzbekistan, Susanna AY826204 AY772300 AY785095 
2048 et al. (BC) 
Cirsium palustre (L.) Scop. Garcia-Jacas et al., 2002 AY826205 AY772307 AY013536 
Cousinia alberti Regel Schmalh. Susanna et al., 2003 AY373721, — 
AY373088 
Cousinia arachnoidea Fisch. & C. A. Susanna et al., 2003 AY373722, = — 
Mey. AY373089 
Cousinia astracanica (Spreng.) Susanna et al., 2003 AY826260 AY772308 AY373070 
Tamamsch. 
Cousinia caespitosa C. Winkl. Susanna et al., 2003 AY373724, —_ — 
AY373091 
Cousinia canescens DC. Garcia-Jacas et al., 2002 AF je iade = — 


AF319122 


156 


Annals of the 
Missouri Botanical Garden 


Table 2. Continued. 


Voucher 


ITS accession 


ral tral 


accession 


matK 


accession 


Species 
Cousinia chrysantha Kult. Susanna et al., 2003 41373725, — — 
19373092 
Cousinta congesta Bunge Susanna et al., 2003 1Y373720, = — 
AY373093 
Cousinia coronata Franch. Susanna et al., 2003 AY820207 AY772309 AY373002 
Cousinia dissecta Kar. & Kir. Susanna el al., 2003 AY373728 


Cousinta esfandiarii Rech. f. & 
Aellen 
Cousinia grandifolia Kult. 
Cousinia karatavica Regel & 
Schmalh. 
Cousinia lappacea Bunge 
Cousinia microcarpa Boiss. 
Cousinia neubaueri Rech. f. 
[= Tiarocarpus neubaueri 
.) Rech. f.] 


Cousinia onopordioides Ledeb. 


Cousinia pallidivirens Kult. 
[=Anura pallidivirens (ult) 
Tschern. | 


Cousinia platylepis Schrenk 
Cousinia polycephala Rupr. 


A. Mey. 


Cousinia syrdariensis Kult. 


Cousinta purpurea C. 


Cousinia ttanshanica Kult. 


Cousinia triflora Schrenk 
Cousinia umbrosa Bunge 
o atractyloides 
C. Winkl.) N 
5o lium creticium 
(Boiss. & Heldr.) N. Garcia & 
Susanna 
Crocod ylium pumilum (L.) 


N. Garcia & Susanna 
Crocodylium syriacum Cass. 


Crupina crupinastrum (Moris.) Vis. 

Crupina vulgaris Cass. 

Cynara cornigera Lind. 

Cynara humilis |. 

Dolomiaea unp. sp. (D. tibetica in 
GenBank) 

Echinops niveus Wall. 

Echinops persicus Stev. & Fisch. 

Echinops ritro L. 


Echinops spinosissimus Turra 


Echinops tschimganicus B. Fedtsch. 


Echinops viscosus D( 


Garcia-Jacas et al.. 2002 


Susanna et al.. 


Susanna et al., 


Susanna et al., 


Susanna et al., 


2003 
2003 


2003 
2003 


Afghanistan, Dieterle 


HO (W) 


Garcia-Jacas et al., 


2002 


Uzbekistan, Botschantzer 


20.6. ra (L 


Susanna el al., 
Susanna et al. 
Susanna et al.. 
Susanna et al.. 
Susanna et al., 
Susanna et al., 
Susanna et al., 


Garcia-Jacas et al.. 


Garcia-Jacas et al., 


E) 


2003 
2003 
2003 
2003 


2003 


2003 
2003 


2002 


2001 [as 


/ lags dea AE 


Boiss. & H«c 


Garcia-Jacas el 7 


> 


2001 [as 


Aegialophila pumilio (L.) 


Boiss. 


Garcia-Jacas et a 


Centaurea E a L.] 


Susanna et al.. 


K Heldr.] 


. 2001 [as 


Garcia-Jacas el i 2 Sip 


Garcia-Jacas el 


Portugal, Blanqué 8 (BC 


Liu (unpubl.) 


Garnatje et al., 


Garcia-Jacas et al., 


t al., 2002 


2005 


Garnatje et al., 2005 


2002 


Garcia-Jacas et al., 2002 


Garnatje et al., 


Garcia-Jacas et EN 


2005 
2002 


AF319123 
„ 
132. 

AY e 3099 
AY826269 
AY826270 
AY826271 


AF319070, 
AF319124 
AY826272 


AY373131, 
AY373704 
AY826273 
AY826274 
\Y373741, 
a 7 at 08 


AY 826 278 


AYO12272, 
AY012308 


AY826279 


135884 
41326230 
AY826281 
AY826282 
AY360334. 


AY538034 

\¥ 538639 

AF319074, 
AF319128 

AF319075, 
AF319129 

A Y538633 

AY820283 


AY772311 
AY772312 
AY772370 


AY772313 
AY772314 
AY772315 
41772310 


41772317 


AY772318 


AY772319 


AY772320 
AY772321 
AY772322 
AY330342 


\Y013537 


1Y373079 


AY373077 
11373007 
AY785096 


AY373008 


AY373675 
\Y373676 
AY 785097 


AY013490 


AY785098 


\Y013538 
AY785099 


AY785100 
AY785101 
AY013539 


AY785102 
\Y013540 


Volume 93, Number 1 


Susanna et 


Cardueae € 


157 


Table 2. 


Continued. 


Species 


Voucher 


ITS accession 


irnb-trnE 


accession 


matK 


accession 


Femeniasia balearica (]. J. Rodr.) 


Susanna 


Galactites tomentosa Moench 


Hypacanthium echinopifolium 
] pt 


(Bornm.) Juz. 


Jurinea albicaulis Bunge 


Jurinea berardioides (Boiss.) 


O. Hoffm 


Jurinea carduiformis Boiss. 


Jurinea humilis (Desf.) DC. 


Jurinea lanipes Rupr. 


Jurinea macrocephala DC. 


lvalochaete modesta (Boiss.) 


go modesta Boiss. 2002 
[= 


Ne h & Rec 


sE] 


Jurinea moschus (Habl ) Bobrov 


[=Jurinella moschus (Habl.) 


Bobrov 


Jurinea robusta Schrenk 


Jurinea sp. ined. 2 |= Modestia 
Winkl.) 


Kharadze & Tamamsch.] 


darwasica (C. 


Jurinea sp. ined. 


Jurinea stoechadifolia (Bieb.) DC. 


Jurinea suffruticosa Regel 


Karvandarina 


Rech. f., 


Klasea serratuloides (DC. 


Greuter & 


a iun 
Vellen & Esfand. 


Wagenitz 


* schakaptaricus (B. 
&T 


1 0 sis cynaroides (Lam. 
yro[ : 


“edtsch.) 


Dittrich 


Leuzea conifera 


morr. 


L.) DC 


amamsch. 


Mantisalca 1 (L 
| 


vill. 


Myopordon aucheri Bot 


.) 


Briq. & 


een hyrcanum m m ) 


Wager 


1 persicum Boiss. 
Nikitinia leptoclada (Bornm. & 


Sint.) Ijin 


Notobasis syria 


a (L.) 


Cass. 


Oldenburgia inte 1 Bond. 


Olgaea baldschuanica 


Iljin 


(C. Winkl.) 


Garcia-Jacas et al., 


Garcia-Jacas et al., 
2003 


Susanna et al.. 


Susanna et al., 


Pakistan, Rechinger 


(W) 


2001 


2002 


2003 


28450 


[as Aegopordon 


berardioides Boiss.] 


Garcia-Jacas et al., 


2002 [as 


Outreya carduiformis 


Jaub. & Spach| 
1995 
. 2003 


Susanna et a 


Susanna et : 


Garcia-Jacas et al., 


Garcia-Jacas et al.. 


2002 


2002 


Garcia-Jacas et al.. 2002 


Susanna et al., 20 


(LE) 


Garcia-Jacas et al., 200 


Ukraine, 
(BC) 


03 
Tadjikistan, Bubanov 21.9.01 


— 


Romo 10321 et al. 


Kazakhstan, Susanna 2166 


et al. (BC) 
Iran, Soják 6379 


Garcia-Jacas et al.. 


Kyrgyzistan, Poljakov 29.8.5 


(LE) 
Turkey, 


Garcia-Jacas et al., 


Garcia-Jacas et al., 


Iran, Carls s.n. (W) 


( 


W) 


2001 


Susanna 2207 et al. 


2001 
2001 


Iran, Koelz 16395 (W) 


Iran, Remandieri s.n. (W) 


e nistan, 


4 (LE) 


me la- a as el al., 


Markova 


2002 


South Africa, Ortiz 3.4.02 


(NBG) 


— Kamelin et al. 


6 (LE) 


3 


AY826284 
AY826285 
AY826286 
AY826287 
AY826288 


AY826289 


1.35808 
AY373748, 
AY373715 
AF319081, 
AF319135 
AF319080 
AF319134 


AF319083, 
AF319137 


AY826291 
AY820290 


AF319082, 
AF319136 

41826292 
AY826293 
AY826294 
AY826295 
AY826296 
AY826297 


AY826298 
AY012292, 


AY012328 


AY820299 
41826300 


AY826301 
AY829442 


AY826302 
AY826303 


AY826304 


AY772327 
AYT 


AY772330 
AY772331 


AY772332 


AY772333 


AY772334 
AY772335 
AY772330 
AY772337 


AY772338 
AY772339 


AY772340 
AY772341 


AY772342 


AY013509 


AYO13541 


AY 37308 


AY 785103 


AY013543 


AY013514 


AY785104 


AY785105 


AY013545 
AY785106 


AY785107 


158 


Annals of the 
Missouri Botanical Garden 


Table 2. Continued. 


Species 


Voucher 


ITS accession 


tral tral 


accession 


math 


accession 


Olgaea pectinata Mjin 


Oligochaeta divaricata (Fisch. & 


Mey.) K. Koch 


Oligochaeta. minima (Boiss.) Briq. 


Onopordum leptolepis DC. 
Onopordum nervosum Boiss. 


Onopordum tauricum Willd. 


Phonus arborescens (I.) G. López 


Phonus riphaeus (Font Quer & Pau) 


G. López 


~ 
n 
un 


Pienomon acarna (L.) 


Plagiobasis centauroides Sehrenk 


Polytaxis lehmanii Bunge 
Polytaxis winkleri Iljin 


Psephellus dealbatus (Willd.) 
K. Koch 
Psephellus gilanicus (Bornm.) 


Wagenitz 


Psephellus incanescens (DC.) Boiss. 
Psephellus persicus (DC.) Wagenitz 


Psephellus pulcherrimus (Willd.) 


Wagenitz 


Psephellus xantocephalus (DC.) 


Mey. 


Ptilostemon afer (Jacq.) Greuter 


Ptilostemon diacantha (Labill) 


Greuter 
Ptilostemon echinocephalus 


Ild.) Greuter 


= 


Ptilostemon hispanicus (Lam.) 


Greuter 


ji [p ana (Lam.) 
M 


pi ab. neuter 


Rhaponticoides haan 
M. V. Agab. & Greuter 
Rhaponticum acaule DC. 
Rhapontic um australe 
(Gaud.) Soskov 
Russowia sogdiana (Bunge) 
B. Fedtsch. 


Saussurea alpina (L.) DC. 


Saussurea asbukinii Ijin | =Frolovia 


asbukinú (Iljin) Lipsch. 


Tzvel.) 


Kazakhstan, Susanna 2187 
al. (BC) 
Garcia-Jacas et al.. 2001 


Moers o ashkent Bot. 
Gard. 
oo et al., 2002 


France, Dijon Bot. Gard. 
(BC) 
Germany, Berlin Bot. Gard. 
3C) 
Vilatersana et al., 2000 [as 
Carthamus arborescens | 
Vilatersana et al., 2000 [as 
Carthamus pg 
Garcia-Jacas et al., 2002 
Kazakhstan, Susanna a 2130 
et al. (BC 
s Kamelin et al. 
3.4.86 (LE) 
tiki, Botschantzev 
et al. 27.4.82 (LE) 
1995 [as 
Centaurea dealbata] 
lL, 2001 las 


Centaurea gilanica 
8 


Susanna el nia 


Garcia-Jacas et a 


~ 


narcia-Jacas et al., 2001 [as 
Centaurea incanescens | 
Garcia-Jacas et al., 2001 las 


Centaurea gaubae| 


n 


rarcia-Jacas et al., 2001 [as 
Aetheopappus 
pulcherrimus] 
Garcia-Jacas et al., 2001 [as 
Centaurea xantocephata | 
Garcia-Jacas et al., 2002 
Turkey, Susanna 2313 et al. 
(BC) 
Ukraine, Romo 10365 et al. 
(BC) 
Spain, Mateos & Garcia- 
Garcia, 18.9, (BC ). 


Susanna el al., 
Garcia-Jacas et al., 2001 


Susanna et al., 1995 


Australia, Funk s.n. (BC) 


Tadjikistan, Botschantzev 
9 (LE ) 
Garcia-Jacas et al., 2002 


Tadjikistan, Kamelin 24.6.70 


(LE) 


AY820305 
AY826306 
AY826307 
AF319086 
AF319140 
41826308 
141326309 
AF140444, 
11140445 


AY820310 


AY8263II 
AY826312 


41320313 


AY820314 


AY820315 
\Y012283, 
1Y012319 
AY826316 


AY820317 


AY829445 


AY826318 
41320319 


AY829442 


AY829444 


L35803 


41320235 


AY820334 
AY820335 


AY826320 


AF319091, 
AF319145 


AY826321 


AY772343 


AY772344 


Y772348 


AY772349 


AY772350 


AY772279 


AY772309 
AY772370 


AY785108 


AYOI3547 


AY785109 


AY785110 


AY013512 


AY013549 


AYO13501 


AY013500 


\YO1349] 


AY785111 
AY785112 


1Y013502 


1Y0I3515 
AY785120 


Volume 93, Number 1 
2006 


Susanna et al. 
Cardueae (Compositae) 


139 


Table 2. Continued. 


VIE 


math 


Fisch. & C. A. Mey. 


l. (BC) 


Species Voucher ITS accession accession accession 
Saussurea carduicephala (Ijin) Hjin Tadjikistan, Smakov & AY826322 AY772357 — 
o carduicephala Dengubianko 6.8.80 (LE 
(Iljin) Kamelin] 
Saussurea ceratocarpa Decne. Raab-Straube. 2 AJ606170. AJ606138 — 
|= Lipschüziella ceratocarpa AJ606210 
(Decne.) Kame lin] 
Saussurea deltoidea (DC.) Sch. Bip. Raab-Straube, 2 AJ606 169, AJ606137 — 
nalaiella p ltoidea (DC.) AJ606209 
Raab-St raube | 
Saussurea discolor (Willd.) DC. Garcia-Jacas et al., z AF319092, — — 
AF319146 
Saussurea elegans Ledeb. Susanna et al., 2 AY820323 AY772358 — 
Saussurea frolowii Ledeb. [5 Frolovia | Waab-Straube, 2 AJ606171. — — 
frolowii (Ledeb.) Raab-Straube | AJ606211 
Saussurea maximowiczii Herder Susanna et al., 20 AY820324 AY772359 — 
Schischkinia albispina (Bunge) Iljin Garcia-Jacas et al., 2 AY826325 AY772360 AY785113 
Schmalhausenia nidulans (Regel) Susanna et al., 2 AY820326 4177236 AY373681 
Petr. 
Serratula coronata L. Garcia-Jacas et al., 2 AY826327 AY772362 AY785114 
Siebera pungens (Lam.) DC. Turkey, Susanna 2316 et al. 41326328 41772363 41735115 
» 
Silybum marianum (L.) Gaertner Garcia-Jacas et al., 2 AY826329 41772364 AYO13551 
Staehelina baetica DC. Garcia-Jacas et al., 2 AF319095, 
AF319149 
Staehelina dubia L. France, Garnatje 25 & Luque AY826330 AY772365 AY785116 
(BC 
Staehelina fruticosa L. Greece, Kriti, Garnatje 147 & AY826331 AY772360 AY785117 
Luque (BC) 
Staehelina lobelii DC. Turkey, Susanna et al. 2272 AY820332 41772367 41735118 
(BC 
Staehelina uniflosculosa Sibth. & Greece, Raus & Hay led AY826333 AY772308 AY785119 
Sm. Berlin Bot. Gard. 
Stizolophus balsamita (Lam.) Garcia-Jacas et al.. 2 41826336 AY772371 AY785121 
Cass. ex Takht. 
Stizolophus coronopifolius Cass. Garcia-Jacas et al.. 2 41326337 AY772372 AY013516 
Synurus palmatopinnatifidus Garcia-Jacas et al., 2 Y826338 AY772373 AY013552 
(Makino) Kitam. 
Syreitschikovia spinulosa (Franch.) Kazakhstan, Susanna AY820339 AY772374 AY785122 
Pavlov 2200 et al. (BC) 
Tarchonanthus camphoratus L. South Africa, Ortiz 17.3.02 AY820340 AY772375 AY785123 
(NBG) 
Tricholepis tibetica Hook. f. & Pakistan. Niisser 1055 (B) AY826341 — = 
Thomson 
Tugarinovia mongolica Iljin Mongolia, Trubov et al. AY826342 AY772377 AY785124 
(LE) 
Tyrimnus leucographus (L.) Cass. Garcia-Jacas et al., 2 AY826343 AY772378 AY013554 
Volutaria crupinoides (Desf.) Garcia-Jacas et al.. 2 AY820344 41772379 AY785125 
Maire 
Xeranthemum annuum L. Turkey, Susanna 2362 et al. AY820345 AY772380 AY785126 
(BC) 
Xeranthemum cylindraceum Sm. Denmark, Copenhagen Bot. AY820346 — — 
Gard. (BC) 
Xeranthemum inapertum (L.) Miller Garcia-Jacas et al., 2 AY826347 AY772381 AY013555 
Xeranthemum longepapposum Kazakhstan, Susanna 2162 AY826348 AY772382 AY785127 


160 


Annals of the 
Missouri Botanical Garden 


Table 2. Continued. 


Species Voucher 


tral-irnk math 


ITS accession accession accession 


Zoegea baldschuanica C. Winkl. 
(BC) 
Zoegea leptaurea l. 


las Z. mianensis | 


Uzbekistan, Ahassanov s.n. 


Garcia-Jacas et al., 2000 


AY012305. — E 


AY012341 
AY820349 AY772383 AY013517 


added, and 40 cycles of 45 sec. denaturation at 94 C, 
l min. annealing at 58 C. 2 min. extension at 72 C. 
and a final 10 min. extension at 72°C. PCR products 
were cleaned with QlAquick PCR Purification Kit 
(Qiagen Ine., Valencia, CA) and sequenced with srn K- 
710 F and AST-IR primers. Direct sequencing of the 
amplified DNA segments was performed as for the 


trnL-trnF region. 


arDNA TTS REGION STRATEGIES 


The three nuclear ITS] spacer, 5.8 S gene, and 
ITS2 spacer (the ITS region) were amplified and 
sequenced together. The ITS region was amplified by 
PCR with 1406 F (Nickrent et al.. 1994) and ITS] 
(White et al., 1990) as forward primers, and. ITS4 
(White et al., 1990) as reverse primer, referring to the 
protocol described in Soltis and Kuzoff (1993). PCR 
products were purified using the QlAquick PCR 
Purification Kit (Qiagen Ine., Valencia, CA). Se- 


— 


quencing primers 1406 F and ITS4 were used. Direct 
sequencing of the amplified DNA segments was 


performed as for the trnb-trnF. region. 


PHYLOGENETIC ANALYSIS 


— 


Nucleotide sequences were edited. with Chromas 
1.56 (Technelysium, Tewantin, Australia). The trnl- 
irk and matK sequences were aligned visually by 
sequential pairwise comparison (Swofford & Olsen, 
1990 


putative proteins with GeneJockey (Biosoft, Cam- 


. The matK sequences were translated for their 


bridge, U.K.) to verify the absence of internal stop 
codons among those for amino acid codons. Due to the 
high level of variability of the ITS sequences. our 
alignment was checked with the ITS alignment for the 
whole Compositae by Goertzen et al. (2003) and 
adjusted manually. In order to conserve the phyloge- 
that 
the trab-trnk 


netic information of insertions and deletions 


constituted most of the variation of 
region, and at the same time avoid an overestimation 
of lengthy indels, they were coded as presence- 
absence characters and added to the end of matrices 
in the combined analyses. The aligned data matrices 
are available on request. from the corresponding 


author. 


The ITS matrix was analyzed by Bayesian in- 


ference, because heuristic parsimony search was 


~ 


impossible due to the size of the data matrix (190 
species; the search for most-parsimonious trees was 
loo time-consuming and soon became unpractical). 
Bayesian. inference (BI) estimation was calculated 
using MrBayes 3.01 (Ronquist & Huelsenbeck. 2003). 
The 


best-available model of molecular evolution, 
required for Bayesian estimations of phylogeny. was 
selected using hierarchical likelihood ratio tests 
(hLRT) and Akaike criteria (AIC) as 
implemented in the MrModeltest 1.1b 
(Nylander, 2002), which considers only nucleotide 
substitution models that are currently implemented in 
PAUP and MrBayes 3.01 (Huelsenbeck & Ronquist, 
2001; Ronquist & Huelsenbeck, 2003). The best-fit 


model of nucleotide substitution for the ITS dataset 


information 


software 


was the same in both methods: the symmetrical model. 
with some sites assumed to be invariable and variable 
sites assumed to follow a discrete gamma distribution 
(GTR+I+G; Yang, 1996). Bayesian inference analyses 
were Initiated with random starting trees and were run 
for | X 10° generations. Four Metropolis-coupled 
Markov chain Monte Carlo (MCMC) chains were 
sampled every 100 generations, which resulted in 
10.000 sample trees. A critical aspect of the Bayesian 
i that the Markov 


reached stationarity. All sample points prior to 


analysis is tọ ensure chain has 
stationarity are essentially random and are discarded 
as “burn-in” 1,000 samples trees, because they do not 
contain useful parameter. estimates. Internodes with 
95% 


statistically significant. A majoritv-rule consensus 


posterior probabilities were considered 
tree was calculated with PAUP version 4.0b4a (Swof- 
1999 


estimated to be significant for nodes with PP > 0.95. 


ford, . Posterior. probability support (PP) was 

For the combined data sets, parsimony analysis 
involved. heuristic searches conducted with PAUP 
version 4.0b10 (Swofford, 1999) using Tree Bisection 
Recognition (TBR) branch swapping with character 
states specified as unordered and unweighted. All 
most parsimonious trees (MPT) were saved. To locate 
islands of most parsimonious trees (Maddison, 1991), 
we performed 100 replicates. with random taxon 
TBR 


lengths, consistency index (CD. and retention index 


addition, and with branch swapping. Tree 


Volume 93, Number 1 
2006 


Susanna et al. 


161 


Cardueae (Compositae) 


Table 3. 


homoplasy indices are calculated by excluding uninformative 


| 


Comparison of results from the ITS, ITS + trnL-trnF, and ITS + trnL-trnF + matK data sets. The consistency and 


laracters. 


Data sel ITS ITS nE ITS + trnL-trnF + math 
Number of taxa 190 2] 111 
Total characters 519 1699 2708 
Informative characters 390 493 677 
Number of MPTs — 298 3122 
Number of steps — 2970 3288 
Islands — 8 5 
Consisteney index (Cl) Ea 0.2914 0.3259 
Retention index (RI) — 0.6800 0.0783 
Homoplasy index (HT) — 0.7086 0.6741 
Range of divergence, ingroup (%) 0-5 0.13-2.43 0.12-1.8 


excluding uninformative 


(RI) 


characters. Two combined analyses were performed. 


are always given, 
with different data sets: the ITS + trnL-trnF sequence 
data and the ITS + trnL-trnF + matK data. 

Bootstrap (BS) and Bremer support (Bremer, 1988: 
1992) 


carried out to obtain support estimates for the nodes in 


Donoghue et al., or decay index (DI) were 
the consensus trees. Bootstrap analysis was performed 
(Felsenstein, 1985) using 1000 replicates and heuris- 
tic search with the default options. In the nrDNA ITS 
data matrix, we used the approach by Lidén et al. 
(1997) 


with 20 replicates, and no branch swapping. For the 


using 1000 replicates, random taxon addition 


two combined matrices, DI was calculated for each 
node by successive analyses using the clade con- 


straint approach, as discussed in Morgan (1997), with 


— 


10 replicates. ACCTRAN (accelerated transformation 
character-state optimization was used for all illustrat- 


ed trees. 
RESULTS AND. DISCUSSION 


Since we were unable to obtain DNA sequences for 
all three genie regions for every taxon sampled, we 
performed three distinct analyses: (1) ITS alone, to 
examine the position of some genera not included in 
previous analyses and for which we were not able to 
amplify any chloroplast region; (2) ITS and trnL-trnk 
regions combined, to study the generic limits in the 
Saussurea group; and (3) the three regions (ITS, IrnL- 
trnF, and matK) combined. to elucidate subtribal 
limits and to confirm the naturalness of the informal 
eroups in Carduinae. The numeric results of the three 
analyses are summarized in Table 3. The resulting 
n Figure 1A and IB (Bayesian 


trees are. shown 
majority rule consensus for the ITS dataset alone), 
Figure 2 (parsimony strict consensus of the combined 
ITS and unn sequences), and Figure 3 (parsi- 
mony strict consensus of the combined ITS, ITE 


and mak sequences). The Bayesian majority rule 


consensus trees of the two combined data sets are 


argely coincident with the parsimony consensus trees, 
and therefore we have added the Bayesian support 
(PP) to those branches that have PP > 0.95 but are 
nol, or are only weakly, supported by parsimony 
(Figs. 2, 3). 


Only the Carlininae branch in the 
combined Bayesian analysis of the three regions is 
illustrated (Fig. 4), because it confirms the position of 


Tugarinovia within Carlininae. 


DELINEATION OF CARDUEAE 


The monophyly of Cardueae was confirmed with the 
new outgroup in all the analyses with high statistical 
support: PP 1.00 (Fig. IA). BS 100%, 100% 
(Figs. 2, 3), and DI = 11 (Figs. 2, 3). Thus, the most 


appropriate status for Echinopinae and Carlininae is 


subtribal. Indeed, Cardueae could be divided into five 
tribes, but we consider it unpractical to fragment 
a natural group that can be so easily recognized on the 


basis of macromorphology. 
SUBTRIBAL CLASSIFICATION 
The four subtribes recognized by the latest report 


.. 2002). Carlininae. 


Echinopinae, Carduinae, and Centaureinae, were 


on the tribe (Garcia-Jacas et a 


confirmed. Subtribe Cardopatiinae must be restored 
and some changes made to correlate molecular 
phylogeny and subtribal delineation. However, in 
view of the moderate support for these basal branches 
(they collapse in a polytomy in all the Bayesian 
analyses, cf. Fig. 1A), subtribes Carlininae, Echino- 
pinae, and Cardopatiinae should be considered 
a currently unresolved polytomy basal to Carduinae— 


Centaureinae. 


CARLININAE AND TUGARINOVIA 


Our results do not modify the circumscription of 


Carlininae in our latest surveys of Cardueae (Garcia- 


162 Annals of the 
Missouri Botanical Garden 


ECHINOPINAE | [CENTAUREINAE 


Cirsium ochrolepidium 


0,85 
- Cirsium arvense 


rduus deflor 
(^ arduus p spn HOC “ph alus 


Notobasis syriaca 
Pic nomo y M arna Carduus 
Cyr group 


1.00 O nara hun mi 
Gali tites tomentosa 
yn 


ic 
1 hinoce, halu 
ine erardioides | Ae, onda 
Jurinea carduiformi | Outreya] 
Jurinea lanip 


z 
T 


tic 
1.00 Jurinea sp. 2 | Mid stia] 
Juri nea macro roce 79 

. Jurinea albicault 

094 1.00 T 1.00 Jurinea sp. 1 Saussurea 
— . Jurinea humilis 
1.00 Jurinea stoechadifolia OUP 

u WOSC. 


Hyalocha MdL 
1.00 10% 8 surea carduicepha | Lene iella] 

, Saussurea ceratocarpa f Lipschitziella] 
(.94 Saussurea deltoidea [~Himalaiella] 

1.00) 3 Dolomiaea tibetica 

Saussurea asbukinii [ Frolovia) 

Poh adage olowii | Frolovia] 

5 da 

Polvtaxis win 

Saussur rea à EE A 

Saussurea discolo 19 

Saussurea elegan 

Saussurea MAIMON dco 

Cousinia pa allidivirens [7 [ Amra] 

e QUITE 10 aher cea 


0.76 


(1,98 


1.00 


Cousinia "ran lifolia 
Hypacanthium echino, de id 
Se malhaus enia HIA ans 


09 


CARDUINAE 


1.00 n hr Saniha 2 
coi Md Cousinia 
inia syrdar ha 
ic group 


C ousinia 


Cousin esceL 
C ousinia i omopordioides 
Cousinia purpu 
Cousinia esfan idi ari 
e Oe neubauert [> Tiarocarpus] 
Cousinia coronata 
Cousinia microcarpa 
reda acantholepis 
redia nivea 
Aure H oerni ua 3 
am appus schakaptaricus O 
E 725 dschuanica ` 
0 gaca a pec tinata 9 
Sara $ palmatopinnatifidus Q 
Syreitschikovia spinel 1 o 
1.00 [ —— Onopordum leptolep £ 
—.— Onopordum ne ire ed o 
»ordum tauricum 
00 Berardia subacaulis 
" SU 7 Staehelina 1 
1.00 Staehelina 1 — Stachelina ¢ 
071 Staehelina prm ulosa 
: Staehelina AUT osa 
a taehelina lobelii 
Amphoricarpos autariatus 
A mph oricarpos exsul 
Chardinia orientalis 
Siebera pu ngens 
No ranthemum annuum 
Xerant e inapertum 
Xeranthemum longepapposum 


0.98 X 7 num cylindraceum 
Atracty 18 c an ce Ha 7177] 
tract ui 
nilis 


1.00 


Xeranthemum 1.00 
rou 


0.80 


1 
yl a 
Atractylodes japonica 
100 Tug ee ¿ ngolica 
Nt dopatium COrymbosutn 
1400“ —— 660 Mov a atrüctyloides CARDOPAT 
E [  — Brachylaena discolor 
— Tarchonanthus camphoratus 
Oldenburgia intermedia 


Fig (pp. 162-163).—A (p. 102). Bayesian NE rule consensus tree of the ITS sequence data matrix (basal part 
of me tree vti Numbe rs above branches are Bavesian Posterior Probabilities 155 ). CARDOPAT = Cardopatiinae. —B (p. 
163). 


branches are Bayesian Ne rior Probabilities (PP) 


Jayesian majority rule consensus tree of he TS sequence data matrix (upper part of the tree only). Numbers above 


Volume 93, Number 1 
2006 


Susanna et al. 163 


Cardueae (Compositae) 


Acanthole opis orientalis 


(99 


1.00 


Echinops niveus 
. ECHINOPINAE 


ius on repi 
Leuz ea coni 
Myopordon a 
Myopordon h hy roan ium 
Oligochaeta divaricata 


0.87 


Olig vochaeta minima 


1 p 
Carthamus c 
Carthamus 555 od cates 
C arthami US turke "stanic us 


CENTAUREINAE 


nauritanicus 
Ci heiroloph us sem, N 
Cheirolophus ey. 
Cru, 


ina cru, inas a um 


Em L— — Crupina vulgaris 
Mantisalca salmantica 


Stizoloph us balsa 


722 hus coronopifolius 
Schis ^hkinia albis, 


BASAL BRANCHES 


Continued. 


Figure 


Jacas et al., 2002; 5 


SISTER GROUP 
OF CENTAUREINAE 


Susanna & Garcia-Jacas, in press). 


genera that were classified by other authors (Dittrich, 


The subtribe is monophyletic without Bayesian 1977, 1996b; Bremer, 1994) in Carlininae belong 
support in the ITS analysis (PP 0.89, Fig. e either to Carduinae (Staehelina and the Xeranthemum 
and with high support in the combined analyses (BS group) or Cardopatiinae (Cardopatium and Cousiniop- 
95%, DI = 7, Fig. 2; BS = 100%, DI = 9, Fig. 3. sis), based on molecular phylogenies as suggested by 


and includes Atractylodes DC., 
and Thevenotia DC. 


but confirmed in our previous work). 


Atractylis, Carlina L., 
(not included in present analyses, 


Phe remaining 


the sequence data. The classic definition of Carlininae 


was based mainly on achene characters (Dittrich, 


1977, 1996b): parenchymatic pericarp usually hir- 


Annals of the 
Missouri Botanical Garden 


164 


Acantholep is orientalis 
C inops hiveu 
Echinop. 
c 6205 


ECH 


15 de 


iç amnus maximus 
cum 
m 


SCOPE uvauxii 
meniasia alearica 
Us riphaeys X 
s çreficus i 
amus turkesta 
4% 


cus 
Aeg be (ean a 


CENTAUREINAE 


Ps a 
0 ep a . 
uritanicus 
heirolop p mauri 
hapon icoides „ 
ru, upina \ vu igari 
atula co 
jasoa RIIE RS 
5 d. es 


E és 


E 2dop ordon| 
Jurinea b adu formis | ya] 
Jurinea r 1 
urinea albicaulis 
aussurea ceratocarpa | =Lip 1 la] 
aussurea c Wee EDT pire 
imala 


ubat f » Frolovia] 
rolovil [=Frolovia 


em Arctioid 
Hypacanthium eci ec inopifolium 
alhaus senia, ardu ans group 
ousinia la, 
ousinia tri 978 
e nia a umbrosa 
15 a inioi 
651003 Cousinioid 
1 a group 
ousia ba er 2 Tiarocarpus| 
lviaxis enm 


ar 
E ET ssurea ele ans A 
10 


Arctium group 


Saussurea group 


irsium ochrolepidium 
arduus us C flora 

arduu genes 
Las deucographus 
irsium echinu 


CARDUINAE 


Ptil mon 

tilostemon diacantha 

am ra gornidera naroides 
nara pus 


41 8 Ale Pda An filv ea. 
Lamyropapr i 1 
1.0 100 Bea 
0 ] u 91113 8 ec tn: m 
iiid eto = t Synurus palmatopinnatifidus 
6 Sy reitschikov via 1 Spin nulos 
m 


or m 


faeh eli ina q^ osculosa 


rui 19953 


aeh gna 


horic 
66 os autarjatus 
oricarpos exsu 
PUE ol 17015 
1 Be Tannu 
eran emum ina ertum 
erant. Dum lo. 97605 ponien 


170 0 208 ls 2 6 5 loi CP 


5 Xeranthemum group 


100 | 
23 L7 


arlina falcata 
arlina lanata . 

ari ag mifera 
arlina 


u. 
u 15 : 
et lodés Um 


hbnamhus. s camphoratus 
5105 2998470 usean "phi 


Strict consensus tree of the most parsimonious trees resulting from the ITS and trnl-rnF combined data matrix. 


Figure 


2. 
Bolded numbers are Bayesian Posterior Probabilities (PP) which are detailed only for branches unsupported by parsimony 


bootstrap. Numbers above branches or before a slash are bootstrap percentages (BS); be low branc he s or after a slash, decay 
indices (DI). A dash indicates bootstrap support less than 50%. Abbreviations of subtribe Carlininae; CP 


Cardopatiinae; ECH = Echinopinae. 


Volume 93, Number 1 Susanna et al. 165 
2006 Cardueae (Compositae) 


Ac anii holepis orientalis 
chinops niveus 

Echinops persicus E C H 
Eeninops ps techimganicus 


100 
- 30 |100 1 
8 
1 Tugarino lame olaa 
84 Acroptilon repens 
„ australe 
Leuzea conire 11 
Rha aponticu aule 
555 ; divaricata 
eu 88953 minima 
cepha Ane 
Conta 1455 eres "maximus 
elfolo us sp eu 


_ 100 — 

15 == Cheirolophus 
atula coronata 

Re serratuloides 


CENTAUREINAE 


Centaurea pragnie rana 
Centaurea involucrata 


Crupina vuigaris. 8 
Zoegea lepta 
tzolophus balsamita 
phus coronopirolus 


1 Arc um 15 
Md Hune Arenum lappa. 
3 99 Cousinia i grandifoli 
5 malhausenia nidulans 
ace 


Cousinia a heubaueri Tiarocarpus 
87 —— Cou C pusl 


28 5 1 . ceda Outreya] 
or i utre 
1 — . ——— Sea . " 


Carduus defloratus 
Tyrimnus loucographus 


0.99 


54/1 
Es EE f ilo h 
10 r a Urt sara 
orn 


100 
11 


79/3 Alfredia niv T 
— — — —— |Lamyropa, us schakaptaricus 

1.0 109 — — Qlga doa ba um blanca” 
Onopordum group 97 Sy Murus pa e linatopinnatifidus 
S yreitschikovi via spinu uosa 


CARDUINAE 


Berardia and Staehelina - rardia subacaulis 
i 100 — i 


f, Sta 1 1 
5 a 


- Xeranthemum group 100 Amphoricarpos exsul 
NES m 28 alis 


6 rdopatium corym m 
— Cardopatiun cory. bosur CP 


„5 ecd Atractylis cancel 2 
100 Arachis carduus 
13 100 Carina UT id 
arlina falcata 
100 100, 10 190 3 Carlina lanata | CL 
ari 
9 18 —————— — |Carlina 
ode aponica 
92 Brachylaena a discolor 
5 Tarchonanthus ; camphoratus 
Oldenburgia intermedia 


Figur Strict consensus tree of the most parsimonious trees resulting from the ITS, L- „n F, and matK combined data 
matrix. EL numbers are Bavesian Posterior Probabilities (PP) detailed only for branches unsupporte d by parsimony 
Wir Numbers above brane hes or before a slash are bootstrap percentages (BS): below branches or after a slash, decay 
indices (DD. A dash indicates bootstrap support less than 50%. Abbreviations of subtribes: CL = Carlininae: CP 


Coe ECH = Echinopinae. 


166 


Annals of the 
Missouri Botanical Garden 


BASAL 
POLYTOMY 


Figure 4 Det 


Atractylis 1 Enid 
Atractylis carduu. 
Carliria acanthifolia 
Carlina falca 


ail of the branch of the subtribe Carlininae from the Bayesian majority rule consensus of the ITS, trn -trnE. 
| 


and mat K potins d data matrix. Numbers above branches are Bayesian Posterior Probabilities (PP) 


sule, and pappus setae very long with plumose 


directly attached to the pericarp. However, these 


characters must be interpreted as plesiomorphic, 


because they appear across all basal subtribes 


(Carlininae, Cardopatiinae, and Echinopinae) and 


even in Carduinae. If we y only on achene 
characters for classification, the resulting definition 
1977, 1996b) differs greatly 


from the delineation on the basis of DNA sequence 


of Carlininae (Dittrich, 


analyses and macromorphology (Susanna & Garcia- 
Jacas, in press). 

Our molecular analyses confirm that Tugarinovia, 
a puzzling monotypic genus of dioecious plants from 
Mongolia that was placed in Carlininae by Dittrich et 
al. (1987). 


subtribal placement is supported only by the Bayesian 


belongs to Cardueae. Curiously, this 
combined analysis of the 


Fig. 4). 


placed on the basis of morphological affinities (leaves, 


three regions (PP = 1.0, 


Indeed. the only subtribe where it can be 


— bracts, and pappus) is Carlininae (Dittrich 
et al., 1987; Susanna & Garcia-Jacas, in press). Our 
analyses show no connection of Tugarinovia with the 
other Carlininae, 


only Fast Asian representative of 


Atractylodes, or with any other genus of the subtribe, 


reinforcing its isolated position. 


CARDOPATIUNAL 


This subtribe had moderate support in our 
gen analyses (BS = 65%, 81%: DI = 4. 6; 
Figs. 2, 3), pe rhaps due to the different evolutionary 


rates of 11 5 Cousiniopsis and perennial Cardopa- 
tium. The Bayesian pros for 11 8 il however, 
(PP — 1.00, Figs. 1A, 


Cardopatiinae, as first E "a" included 


is very high ( 
Subtribe ( 


only the east Mediterranean genus Cardopatium. 


Later, Nevski (1937) described a monotypic genus 
from central Asia. Cousiniopsis. closely related to 


Cardopatium (it was first described as Cardopatium 
atractyloides C. Winkler). Classic monographers of 
Compositae (Bentham, 187: 1894: Dit- 
1977; 1994) consistently placed both 


genera among Carlininae, but the only characters that 


: Hoffmann. 


trich, Bremer, 


connect these two groups are those of achenes, which 


could equally relate Cardopatium and Cousiniopsis to 


Echinopinae. It is tempting to interpret the corymbose 
inflorescence of Cardopatium, formed by very small, 
few-flowered capitula. as a first step towards synceph- 
aly. On this basis. Petit (1997) considered Cardopa- 
tum sister to Echinops L. and placed Cardopatium 
and Cousiniopsis in Echinopinae. On the basis of our 
results, we prefer to interpret these similarities as 
convergence, because syncephalies at various states 
of development involving small. few-flowered heads 
n all the subtribes across Cardueae (Garcia- 
Jacas et al., 2002). 


occur 


ECHINOPINAE 


Our results demonstrate, contrary to our previous 


studies (Garcia-Jacas et al.. 2002), that Echinopinae 
include only Echinops s.l. (Echinops and Acantholepis 
with strong support (PP. = 1.00, Fig. 1B: BS = 
100%, DI = 26, 30, Figs. 2, 3). In fact. a 


molecular study indicates that Acantholepis is a re- 


Less.) 


recent 


duced, unarmed species of Echinops (Garnatje et al., 
2005). as originally described (Echinops acantholepis 
Jaub. & Spach). Our combined analyses reveal that 
the Xeranthemum group does not belong to Echino- 
pinae, but rather to Carduinae as sister to the rest of 
IA, 2, 3) 


origin of the 


this subtribe (Figs. 


The compound inflorescence « 


Echinops cannot be tracked. on molecular grounds, 


because the subtribe does not show supported affinity 


to any other group in Cardueae. Cardopatiinae and 


Carlininae are the best candidates for sister groups to 
the subtribe Echinopinae (because the structure of the 
Dittrich, 1977) 


achenes is very similar, cf. 


CARDUINAE 


If monophyletic Centaureinae are recognized as 


a distinct subtribe, the Carduinae constitute a para- 


phyletic assemblage (Figs. IA. 2. 3) However. 


allernate solutions are not practical. Either a subtribe 
level is ascribed to all the monophyletic groups 


recognized in present Carduinae and a fragmented 


classification results. a Single large subtribe 


Carduinae, including Centaureinae, is maintained. 


Volume 93, Number 1 
2006 


Susanna et al. 167 


Cardueae (Compositae) 


which thereby encompasses almost ninety per cent of 


the species of the tribe (Garcia-Jacas et al., 2002). 
Even in this disparate assemblage, some well- 


defined groups emerge, together with genera without 


known affinities like Berardia or Staehelina. 


BERARDIA AND STAEHELINA 


Our molecular analyses show that these two genera 
present no obvious affinities. They cluster in an 
isolated position within Carduinae, in both combined 
analyses, without statistical support (Figs. 2, 3). 
Berardia was ranked among Mutisieae on the basis of 


achene characters (Dittrich, 1977) and we agree in 
(a) 


that the pericarp cells, with thickened u-shaped 
walls, are very similar to the type found in 
Gochnatiinae Benth. € Hook. f, a subtribe of 


Mutiseae (Dittrich, 1996a). Further support for the, 


albeit weak, relationship between Berardia and 
Staehelina is that the pericarp of Staehelina is also 
(Dittrich, 1996a). However, we 


state whether this similarity represents convergence 


“gochnatioid” cannot 
or is a very old character conserved in these two 
anomalous genera 

Staehelina was previously placed among Carlininae 
(Bentham. 1873; Hoffmann, 1894; Dittrich, 1977: 
Bremer, 1994), but Petit (1997) proposed moving it to 
Carduinae. For Dittrich (1996b), the two species of 
Staehelina with hirsute pericarp (S. fruticosa L. and 5. 
lobelii DC.) should be classified in a distinct genus, 
Hirtellina Cass. Al our analyses grouped the included 


species of Staehelina (five out of eight) in a robust 


clade with very high support (PP = 1.00, Fig. 1A; BS 
= 100%, DI — 34, 47, Figs. 2, 3). However, both 


combined analyses divided the genus into two well- 
supported clades that coincide with Staehelina s.s. 
and Hirtellina (Figs. 2, 3), which is compatible with 
the division of the genus. Nevertheless, morphological 


differences other than presence or absence of achene 


pilosity are virtually non-existent, and we prefer to 


keep a single genus with Staehelina and Hirtellina 


recognized at sectional rank within Staehelina 


(Susanna & Garcia-Jacas, in press). 


THE XERANTHEMUM GROUP 


In a previous study, the Xeranthemum group 


(genera Amphoricarpos, Chardinia Desf., Siebera, 
and Xeranthemum) was placed among subtribe 


Echinopinae. This unexpected result led us to propose 
that the very small and peculiar heads of the genera of 
the group, with very large receptacular bracts, could 
200: 


Our new analyses show that this view was erroneous 


Nw 
— 


constitute a syncephaly (Garcia-Jacas et al., 


and, in fact, Harris (1995) had already demonstrated 


inflorescence of Xeranthemum was not 
the 


in the combined analyses, 


that the 


a syncephaly. Indeed, Xeranthemum group 


appears, as part of the 
Carduinae, sister to the rest of the subtribe, with low 
parsimony bootstrap support but very high Bayesian 
support (PP = 1.00, Figs. 1A, 2, 
position. The monophyly of the group also has very 
high support (PP= 1.00, Fig. 1A; BS = 100%. DI = 
23, 28, Figs. 2, 3). Traditional classification (Dittrich, 
1977; Bremer, 1994) placed it in Carlininae, and Petit 


3). in an isolated 


(1997) was the first to suggest Carduinae. Species of 
three of the genera of the group (Chardinia, Siebera, 
and Xeranthemum) are annual colonizers of arid and 
waste-land thoroughout the Mediterranean region. In 
contrast, species of the dwarf shrubby genus Ampho- 
ricarpos are narrow mountain endemics, sister to the 

rest of the genera of the Xeranthemum group in all the 


analyses (Figs. 1A, 2, 3) 


THE ONOPORDUM GROUP 


The usual definition of this group is based on the 
A pitted, naked 
receptacle is otherwise rare in the tribe. However, 
(Alfredia 


ass.) show an epaleate receptacle. In addition to this 


absence of receptacular bracts. 


not all the species of at least one genus 


— 


character, achenes are also peculiar. with pericarp 


diversely pitted, wrinkled, or rugulose (Susanna & 


=~ 


Jarcia-Jacas, in press), but seldom smooth (Olgaea 


and Syreitschikovia). The group has considerable 
negative importance, because species of Onopordum 
include some highly noxious weeds widespread in the 
Mediterranean region and the American west like O. 


and O. 


that can reach up to 3 m high. 


— 


acanthium L. nervosum Boiss., giant thistles 


Three genera not included in previous studies, 
Lamyropappus, Olgaea, and Syreitschikovia, are 
classified in the Onopordum clade in all three 


analyses, which confirm the group as a natural one 
with significant support (PP = 1.00, Fig. LA; BS = 
91%, 97%, DI = 6, 4, Figs. 2, 3). Syreitschikovia had 
been previously placed by Dittrich (1977) and Bremer 
(1994) in Centaureinae. Its classification in Carduinae 
and its relationship to the Onopordum group was 
tal. (2002) on the basis of 


reported by Susanna el 
morphology. 

Generic definitions in the group are unclear, with 
Alfredia f 


a polytomy with Lamyropappus, Olgaea, Synurus Ijin, 


forms 


the only exception of Onopordum. 


and Syreitschikovia (Figs. 1A, 2, 3). The inclusion of 
more species of Olgaea, which comprises some 15 
taxa from the Tien Shan mountains of Kyrgyzstan and 
Tajikistan, and its strange relative Takeikadzuchia 


Kitag. & Kitam. from Mongolia, may contribute to 


a better definition of the genera in the group. 


Annals of the 
Missouri Botanical Garden 


> 


THE CARDUUS GROUP 


This represents a large complex of very spiny plants 
which are usually called thistles. 
large-sized heads, spiny leaves, and a long pappus 
detachable as a single piece. Qur results indicate that 
a large number of the genera (Carduus L., Cirsium 
Mill., Silybum 
Adans., and Tyrimnus Cass.) form a natural group with 
significant parsimony support (BS = 9196, 100%. DI 
= 10, 14, Figs. 2, 3). 


Galactites Moench, Lamyropsis (Kharadze) Dittrich. 


Notobasis Cass., Picnomon Adans.. 


Remaining genera, Cynara l. 
and Ptilostemon Cass., are also placed in this group in 
this is 


(PP 


all analyses, but supported. by the 


1.00 in all the cases. 


only 
Bayesian analyses 
LAs 2.3): 

As pointed out by Hiiffner and Hellwig (1999) and 


Garcia-Jacas et al. (2002), phylogenetic relationships 


Figs. 


and generic boundaries within the clade are obscure 
(Fig. 


co-existence of 


LA, 2, 3). One of the reasons for this is that the 


annual or biennial species (most of 


Carduus, Galactites, Picnomon. Silybum, or Tyrimnus) 


Cirsium. 


together with perennials (many Cynara, 
Lamyropsis, and Ptilostemon) hinders the assessment 
of the two aspects from a molecular standpoint. 


Differences i 


mutation rates between annuals and 
1997; 


2001) make comparison of 


perennials (Gaut et al.. Laroche et al.. 1907: 


Andreasen & Baldwin. 


DNA 


expected results, 


reliable tool. In fact. 
like the 


annual genus Galactites, could be a result of these 


sequences a un- 


strange position of the 
differences: Galactites is placed close to the base of 
the combined analyses 
Ptilostemon (Figs. 2, 3), 


the thistles in . grouped. with 
thereby contradicting mor- 


phological evidence (Galactites is morphologically 


similar to Carduus or Cirsium). Lamyropsis, the only 
genus of the thistles missing in our previous studies 
and sequenced here for the first time, appears related 
to Ptilostemon in the combined analyses, without 
(Figs. 2, 3). 


dentate-spiny 


support Species of Lamyropsis have 


leaves with very prominent veins 
beneath, similar to many species of Ptilostemon. The 
affinities between the two genera were pointed out by 
Dittrich (1971). 

‘aking into account our low sampling for such an 
(ca. 500 n total). 


remark on the thistles would be pre- 


enormous group species any 


concluding 
mature. The Carduus group, together with the two 


following ones, requires a more comprehensive 


molecular analysis. 


THE ARCTIUM GROUP. 


This group has been the subject of a recent 


preliminary molecular survey, using ITS and matk 


All share medium or 


The results herein. 


Susanna et al., 2003). 


including the un 


sequences 
region (Figs. 2, 3). do not 
change our previous main conclusions that the limits 
Our study 


of Arctium L. and Cousinia are unclear. 


Susanna et al. 


03) demonstrated two principal 
the Arctioid 


(supported wi d the two combined analyses with 


clades in the Arctium group: clade 


BS = 85%, 100%, DI — 5, 7, Figs. 2, 3) and the 
C pue c e ue ted by all three analyses with 
PP = „Fig. IA.; BS = 92%, 94%, 


Figs. 2. 3, The 


molecular, chromosome. and pollen characters. but 


Iwo groups can be segregated by 


this grouping is not consistent. with eme two 


| I] Winkl. 


genera of the group. Schmalhausenia 
pude on the 


and 
Hypacanthium Juz., are affined with 


basis of pollen, chromosomes, and DNA sequences 


(Figs. IA. 2), but in other respects are morphologi- 
cally much closer to Cousinia. In addition to an 


Arctioid species group within Cousinia, there is also 
a Cousinioid group seen in Arctium. More sampling of 
the obscure Cousinia subgenus  Hypacanthodes 
Tscherneva from Central Asia is required, but it is 
| 
highly probable that all four genera will have to be 
8y 1 t 
grouped in Arctium. 
Finally, our ITS analysis (Fig. 1A) confirms that the 
E P O 
purported genera Anura and Tiarocarpus, as pre- 


viously proposed by Susanna and Garcia-Jacas (in 


press), cannot be segregated from Cousinia, to which 
they are united with good support (PP = 1.00 in both 


cases, Fig. LA). 


HE SAUSSUREA GROUP 

The only genera placed in the Saussurea group by 
Susanna and Garcia-Jacas (in press) that were not 
included in our previous studies are Dolomiaea and 
Polytaxis. The ITS and the combined ITS + tral. -trnF 
analyses place Polytaxis basal to Saussurea with high 
Bayesian (PP = 0.99, Fig. LA; PP = 1.00, lig. 2) and 
DI = 6, Fig. 2). 


Because species of Polytaxis are the only annual taxa 


good parsimony support (BS = 87%, 
in this clade, its basal position could originate in the 


faster evolution of annuals relative to perennials (Gaut 


el al. 1997; Laroche et al, 1997; Andreasen & 
Baldwin, 2001). in the same way that annual 
Acantholepis always appears basal to perennial 


Echinops (Vig. 1B, 2, 3). By 
grouped with high support (PP = 1.00, Fig. LA; 
0456. DI — 6. Fig. 2) in the ITS and the combined ITS 
and irnk with Frolovia (Saussurea 
5. frolovit Ledeb.). H 


adequate do consider Frolovia, a genus restored. by 


its side, Dolomiaea is 


we) 
ya 


trnl.- analyses 


asbukinit Iljin and seems 


Raab-Staube (2003), a synonym of Dolomiaea. 
Another taxon that was not included in our previous 
study is the (Jurinea 


purported genus Aegopordon 


Volume 93, Number 1 
2006 


Susanna et al. 169 


Cardueae (Compositae) 


berardioides (Boiss.) O. Hoffm. in Figs. 1A, 2), which, 
according to Susanna and Garcia-Jacas (in press). 
should be considered a synonym of Jurinea. The 
combined ITS + trnL-trnF analysis (Fig. 2) place it in 
a robust clade (BS = 91%, DI — 4) with 
carduiformis (Jaub. & Spach) Boiss., formerly 
considered a distinct genus (Outreya Jaub. & Spach) 
2002). 


Our results confirm that the limits between Jurinea 


Jurinea 


also 


hat we merged in Jurinea (Garcia-Jacas et al.. 


and Saussurea are not well established (as pointed out 
2004), because some species 
the 


recently by Kita et al., 
in Saussurea are grouped it 


formerly included 
genus Jurinea xd LA, 2). Saussurea carduicephala 
(Hjin) Hjin and S. deltoidea (DC.) Sch. 
considered by R 3 (2003) 
genus Himalaiella. Saussurea ceratocarpa Decne. was 
for Raab-Staube (2003) a restored genus Lipschitziella. 
Both purported genera form a robust clade in the ITS 
and the ITS + trnL-trnF analysis (PP = 1.00, Fig. IA; 
BS — 98%, DI = 6, Fig. 2). 


we grant the genus level to this clade, Himalaiella 


Bip. were 


as the distinct 


According to this result, if 


should be considered a synonym of Lipschitziella. 


However, we prefer to consider it a synonym of 


Jurinea, because both purported Himalaiella and 
Lipschitziella form a monophyletic clade with Jurinea 
s. str., with very high support (PP = 1.00, Fig. 1A: BS 
= 100%, DI = 15, Fig. 2). 

No 


entanglement of 


final conclusions can be drawn from this 


genera, because our sampling of 
Jurinea was very limited. However, a redefinition of 
the boundaries between Jurinea and Saussurea is 
clearly required. The clarification of these limits, and 
indeed the description of new genera in a complex in 
which no less than 15 have been already described 
(Susanna & Garcia-Jacas, in press), calls 


more comprehensive sampling than any performed to 


for a much 
date 


CENTAUREINAE 
Our results confirm the general outline of Centau- 
reinae proposed by Garcia-Jacas et al., 2001, this time 


on the basis of three regions of the genome (Figs. UB. 
2. 3). Here we describe only the most important 
results, namely the inclusion of two genera formerly 
classified in Carduinae, Myopordon and Nikitinia, in 
the subtribe. 

Myopordon was considered related to Onopordum 
(hence the name) and placed in subtribe Carduinae 
because of the absence of receptacular setae (Wa- 
genitz, 1958; Dittrich, 1977). In the ITS 
analysis relates Myopordon to the genus Oligochaeta 
K. Koch of subtribe Centaureinae with very high 
support (PP = 1.00, Fig. 1B), and the combined ITS + 


irnL-trnF analysis places Myopordon deeply nested in 


contrast, 


the Rhaponticum Vaill. group with high support (BS = 
85%, DI = 5, Fig. 2). Difficulties in interpretation of, 
even apparently unambiguous, characters consistently 
occur in tribe Cardueae: as we have seen above, the 
naked receptacle is a supported character of the 
Onopordum group (Susanna & Garcia-Jacas, in press), 
but there are many exceptions. Epaleate genera are 
present in almost every subtribe: Tugarinovia in 
Carlininae, Dolomiaea and part of the Onopordum 
group in Carduinae, and Myopordon and Russowia C. 


Winkl. in 


Myopordon within Centaureinae on morphological 


Centaureinae. To verify the position of 


grounds, the characters of the achenes are critical; 
however. we were unable to find herbarium material 
Mouterde (1983) described the 


as oblique, a character 


with mature fruits. 
insertion areole of the achenes 
of Centaureinae. This observation contrasts with that 
of Wagenitz (1958), 
straight, which therefore points towards Carduinae. 


who reported the insertion as 


Vikitinia was described in Carduinae. and in recent 


reviews of the tribe was maintained in that subtribe 


(Dittrich, 1977; Bremer, 1994). However, achene 
characters are undoubtedly centauroid (especially the 
double pappus, illustrated in Susanna et al., 2002) 


and relate it to the genus Klasea Cass. as confirmed by 


molecular analyses (PP = 1.00, Fig. 1B). 


CONCLUDING REMARKS 


suitable and with the 


With a 


addition of the trnL-trnF region, 


more outgroup 
the systematics of 
Cardueae now appears to be more fully resolved. 
However, there 
unclear and their clarification requires better sam- 


are taxonomic issues that remain 
pling and more morphological and molecular data. In 
addition to only moderate support for basal branches 
in the combined analysis of the three regions, doubts 


remain regarding problems typical of delimitation of 


arge genera such as are frequently found ii 


Very 


Compositae (classic examples are Aster L., Erigeron 


|... or Senecio L.). In Cardueae, generic boundaries are 


difficult to establish for Carduus, Cirsium, Cousinia, 


Jurinea. and Saussurea. In the case of Carduus and 
Cirsium, extensive sampling in Africa and North 
America is needed. For Cousinia, Jurinea, and 


Saussurea, which are the easternmost representatives 
of the tribe in Eurasia, intensive collections are called 


for in Central and East Asia. 


Literature Cited 

Andreasen, K. & B. G. 
rales between annual and 8 5 nnial lineages of checker 
mallows (Sidalcea, Malvaceae): Evidence from 18 8-26 S 
rDNA internal and external mita spacers. Molec. 


Biol. 18: 936—944. 


3aldwin. 2001. Unequal evolutionary 


2 


Evol. 


170 


Bayer, R.. 


Asteraceae 


— 


K J. R. Starr. 1998. 
based on two 


Tribal P ded A the 
non—coding chloro se- 
, intron s in intergenic: spacer. 
. Gard. 85: 242-256. 

—5233 in G. Bentham 
Lovell Reeve & Co., 


373. 1 5 P 


). Hooker, Genera Plantarum. 


ce 


Bremer, K. The limits of amino acid se quence data in 


angiospe rm n logenetic reconstruction. Evolution 42: 
795-803. 


1994. 


Timber Press, 


Asteraceae. Cladisties & | Classification. 


Portlar 


Cassini, H. 1819 Vario articles in: Dictionnaire de 
K Di 


sciences Naturelles, Paris. Cited by King, R. & Dawson. 


H. W. 1975. 


5 assini on Composilae. Oriole, New York. 


Cullings, K. W. 1992. Design and testing of a plant-specific 
PCR primer da ecological and evolutionary studies. 
Molec. Ecol. J: 233-240. 


Dittrich, M. 1971. Dittrich—Zur 
Frucht—und Blütenmorphologie einer kritise 


Lamyropsis (Charadze) 
ien. Gruppe 
aus der reu IN erwandtsehaft. Candollea 26: 97-102. 
s matic review. zu 999-1015 
. Harborne & 
eae and Chemistry pe the C aeta 


7. Cynareae 


4 1 
1 J. Turner, The 


Academic 


Press, London-New York-San Francise 
19964. Bemerkungen zur Tribus zugehorigkeit von 
Berardia. Ann. Naturhist. Mus. Wien 99(B, suppl.). 
329-342. 
— ——. 1996b. Die Bedeutung morphologischer und 
anatomischer Achünen— Merkmale für die Systematik 


der Tribus Ke N Cass. und Carlineae Cass. 
Boissiera 51: 9-102. 
. Kastner & H. Meusel, 1087. 


Carlineen=Sippe? Bot. 


Tugarinovia—eine 

aa he Jahrb. Syst. 108: 
167-18 

"REDE M J. R. C. Olmstead, J. F. Smith & J. D. Palmer. 
1992. Phylogenetic. re lationships of Dipsac 5 5 Ss base d on 
Tg sequences. Ann. Missouri H Gard. 333—: 

Doyle, J. J. & J. U. Doyle. 1987. i D isolation 
procedure for eis rM a fresh leaf tissi 
Phytochem. Bull. 19: 11- 

Felsenstein, J. 1985. 1 0 mee limits on a An 
approach using the o Evolution 39: 783-701. 

. Keeley, R. Chan, L. Watson, B. 

. Panero, B. G. Baldwin, N. 

N . Jansen. 2005 


Funk, V. A., R. 


Gemeinholzer, Ji. S 


ny 
— 


Garcia-Jacas, A. Sus 


Everywhere hal Antarctica: Using a supertree do un- 
derstand the diversity and distribution of the C :omposilae. 
Biol. Skr. 55: 343-373 

Varslan. 


sanna, V. Mozalfarian & R. 


Ornelas, | 
The natural de ‘limitation of 9 


2000. rurea (Asteraceae: 
Cardueae): ITS sequence analysis 5 Centaurea jacea 
group. Pl. Syst. Evol. 223. 185 o 


2001. 


subtribe 


R. Vilatersana. 


y — T. Garnatje & 
Generic delimitation and phylogeny of the 


~ 


nuclear and 
503—515. 


S im Sama. 


;entaureinae (Asteraceae): A combines 
chloro i 11 A d Ann. Bot. 
na & 2002. 
Tribal d 1 d M and phylogeny of the 


ral rnatſe 

Cardueae (Asteraceae): A combined nue ^ ‘ar ne e chloro- 

plast DNA a Molec 

A. Susanna, N. Garcia-Jacas, : : 
Valles. 2005. A first approach to the molec de 

1 7 ny of the genus Echinops L. (Asteraceae): Sectional 

delimitation and re lationships s A genus da 

107 


nalysis. Phylogenet. Evol. 


\ Giant & 


Less. Folia Geobotanica 40: 4 


Annals of the 
Missouri Botanical Garden 


Gaut, B. S., L. G. Clark. Wendel & S. V. Muse. 1997. 


Comparisons of the o c M evolutionary process at rbe, 
and ndhF in A e grass family (Poaceae). Molec. Biol. Eval. 
14: 769-777. 

a n, L. R.. J. J. Cannone, R. R. Gutell & R. K. Jansen. 


003. ITS secondary structure de rived from a 
1 Implications lor sequence alignment. and poy- 


loge ny of the Asteraceae. Molec. Phylogenet. Evol. : 
216—234. 
Haffner, E. & F. H. Hellwig. 1999, Phylogeny of the tribe 


Cardueae (C 1 7 with emphasis on the subtribe 
Carduinae: An analysis based on 


Willdenowia 29: 27-39, 
i LOM 


TS sequence data. 


Inflorescence and floral ontogeny 


Harris, E. M. 1995. in 
Asteraceae: A synthesis of historical and current concepts. 
Bot. a v. 61: 93-278. 

Heywood, V. II., J. B. Harborne & B. L. Turner. 1977. An 
s L 


ee Pp. 1-20 in V. H. Heywood, 
3. 


Harborne & B. L. Turner (editors), The Biology and 


1 5 of the Compositae. Academic Press, London- 


E, 


ew York—San Francisco. 
Hilu, K. W. & H. 1997. 
variation and applis ation 
Bot. 19 30-839. 
m ann, 
OK 


Liang. The math gene: Sequence 
t | 


plant systematics. Amer. J. 

. 1894. Compositae. Pp. 324-333 in A. Engler 
se Die natürlichen Pflanzenfamilien. Wilhelm 
s nge 51 Leipzig. 


Huelsenbeck, J. P. & F. R. Ronquist. 2001. MRBAYES: 
Baye sian inference of phylogenetic. trees. Bioinformatics 

: 754—755. 
Jansen, R. K. & J. D. Palmer. 1987. Chloroplast DNA from 
lettuce and Barnadesia (Asteraceae): Structure, gene 


and characterization of a large inversion. 

urr. Genet. I: 355—964. 

& D. E. Soltis. 
in Saxifragaceae sensu stricto and Gilia (Polemoniaceae 
using math Missouri. Bot. Gare 

9-175 

E Dod K. PBüflbawa: M. Ho, H. Ohba & M. Kato. 2004. 

Molec ilar phylogenetic wid s and systematics of the 


localization, 


Johnson, L. A. 1995, Phylogenetic inference 


sequences. Ann. 


genus Saussurea and related genera (Asteraceae, Car- 
dueae). Taxon 53: 679-690 
Laroche, J.. P. Li, L. Maggia & J. Bousquet. 1997, 


Molecular evolution of angiosperm mitochondrial introns 
and exons. Proc. Nat. Acad. Sei. U.S.A. 94: 5722 
5121. 

Lidén, M., T. Fukuhara. J. Rylander & B. 


Phylogeny and classification of 


1997. 


with em- 


Oxelman. 

Fumariaceae, 

phasis on o 85 J. hend on the plastid gene rpsló 
intron. Pl. Syst. Evol. 206: 420 

Liu. J. -N. Lo Gao, Z.-D. 155 en & A.-M. Lu. 2002. 
Molecular phylogeny and biogeography of the Qinghai— 


E 


Asteraceae). Molec. 


Tibet Plateau endemic Nannoglottis 
Phylogenet. Evol. 23: 307—325. 
Maddison, D. R. 1991. The 

multiple islands of most parsimonious trees. Syst. Zool. 40: 
315-328 


discovery and importance of 


Morgan. D. R. 1997. 

dn as data. Taxon 46: 509—517. 
Mouterde, P. 1983. Nouvelle Flore du Liban et de la Syrie. 
zu dee & M. Dittrich (editors). E onservaloire et Jardin 


Decay analysis of large sets of 


tiie s de la Ville de Genève. 


Nevski. S 1937. Beiträge zur Flora n 's Kuhil: dq baa 
seiner birge. Trudy Bot. Inst. Akad. Nauk S er. 
l, 4: 199—340. 


Volume 93, Number 1 
2006 


Susanna et al. 171 
Cardueae (Compositae) 


Schuette & M. Starr. 1994. A 
bee (V 1 based 


Nickrent, 


molec e 


K. P. 
)h als ny of 
internal transcribed. spacer 
sequences. wer . Bot. 81: 1149-1160 
A. Mr. Wodeltes v. I.I b. 
uted by the e Department of Systematic Zoology. 
S ~ w.ebc.uu.se/ 
2005 


on nuclear ribosomal DI 
P rogram distrib- 
Uppsala University, ppsala. 
systzoo/stal bi ae imi 0 Dec. 
Oberprieler. 2002. 


melum s 


phylogenetic ui sis of Chamae- 


(C 1 related 
genera based upon \ 


IGS — nee variation. Bol. J. Linn. Soc. 55-273. 
V. A 


Pane J. & Funk. 2002. Toward a ARR IA netiec 
ubfamilial c in -ation for the C ene (Aster: 
Proc. Biol. Washington 115: 909-922, 

Petit. D. P. Tm ;eneric 0 e of the Cardueae 


(C ee : A cladistic — of morphological data. 
PI. Syst. Evol. 207: 173-2 
Raab- mu E. von. "s los netic re iens in 


Saussurec 


a 


(Compositae, ‘ardues 1€) sensu lato. inf red 


from morphologic al; ITS and trnL-trnk sequence data, with 
. Lipschüziella d 


a synopsis of „„ gen. nov. 


Frolovia. Willdenowia 33: 379—402. 

Ronquist. F. & J. P. Hue e eee k. 2003. MRBAYES 3: 
Bayesian phylogenetic infer rence under mixed models. 
Bioinformatics 19: 1572-15 

Soltis. D. E. & R K. Eno 1993. ITS sequence 


homogeneity within and among populations of Lomatium 

laevigatum (Umbelliferae). Molec. Phylo- 
uy ). 

E T. C. Collier & M. 199]. 


des agaceae): 911 5 nce v c 10655 


grayt and a 

genet. E je 
— P. . Edgerton. 
The 9 9 1 group ( 
plast transfer and paraphyly. Amer. J. Bot. 78: LO 2 
A. & N. Garcia-Jacas. In press. The tribe C june "ae. 

| Kadereit & K. Kubitzki (editors), Compositae. The 
Fani s and Genera of Vascular Plants. Springer-Verlag. 


idelb 


Susanna, 


a e, dq E Soltis & D. S Soltis, 1995. 
Phylogenetic re juam in tribe Cardueae (Asteraceac) 
Amer. J. Bot. 82: 1056-1068. 


based on ITS sequences. 


Pp. 411—501 in D. Hillis & C. 


Garcia-Jacas. 


„ T. Garnatje & N. 


vid Molecular 
phylogeny of Cheirolophus (Asteraceae: C 


rdueae-Cen- 
laureinae) based on 9 75 sequences of nuc 5 ribosomal 
DNA. Pl. Syst. Evol. 214: 147—160. 

—— & R. Vilatersana. 2002. On the 
correct subtri bil placement of the genera Syreitschikovia 
and VVV in (Asteraceae, Cardueae): Carduinae or 
3 Linn. Soc. 140: 313— 

R. Vilatersana & T. Ge 


C enlaureinae? Bot. J. 


N. Garcia-Jacas, irnalye. 


2003. Ge neric boundaries and evolution of characters in 


nuclear and chloroplast. DNA 
e e celona) 26: 101-118. 
Analysis Using 


2 
B] 
un 
Y 
~ 
D 
^ 
> 
— 
a 
— 


. Phylogenetic 
Version 4. Sinauer, 


Swofford, LI 
1 (and Other Methods) 
Sunderland. 

& G. J. Olsen. 1990. Phylogeny reconstruction. 

Moritz (editors), Molecular 
Systematics. Sinauer, Sunde rland. 

Taberlet. P.. L. Gielly, Pautou & J. Bouvet. 1991. 
Unive isa prime rs for sen of three non-coding 

NA. 7: 


Molec. Biol. 


c 


chloroplast I 


Vilatersana, 
2000. Generic de limitation and phylogeny of the Cardun- 
ITS 


„Susanna. N. Garcia-Jacas & T. Garnatje. 


9 9 ex ( n raceae) based on 
. Evol. 221: 05. 


19 0 Gattung 1 Boiss (Compo- 


cellus-Carthamus 
sequences. Pl. Sy 

Wagenitz, G. 1958. 
sitae-Cynareae). Ber. Deutsch. Bot. Ges. 71: 271-277. 

1976. Systematies and phylogeny of the Compositae 
(Asteraceae). Pl.! Syst. Evol. 125: 20-16. 

White. T. J.. T. Bruns. S. Lee & J. 1990. 


id anu and died sequencing of fungal oe 


Taylor. 
V genes for phylogenetics. Pp. 315-322 in: M. Innis. 
¿ | | 


J. Sninsky & T. White ia PCR 
zui Lo A and Applications. 


— 
X 
N 
= 


Academic Press, San Diego. 
Yang, Z. 1996. 


on phylogenetic analyses. 
372 


Among-sile variation and its impact 
à . 367 


Trends Ecol. Evol. 11: 367— 


Volume 93, Number J. pp. 12172 of the Assas or Tig. Missouri BOTANICAL GARDEN 
was published on May 31, 2006. 


ARA 


| www.mbgpress.org 


CONTENTS 


Species Reconsidered: Consequences for Biodiversity and Evolution, the 50th Annual 
Systematics Symposium of the Missouri Botanical Garden 
Species Reconsidered: Consequences for Biodiversity and Evolution. Introduction 
— ich Richardson l 
Linnaeus's Bul Was NotEssentialist Mary P. Winsor 2 
Problems with Species: Patterns and Processes of Species Formation in Salamanders 


Adaptation, Speciation, and Convergence: A Hierarchical "NETS of Adaptive Radia- 


tion in Caribbean Anolis Lizards 
CNN Jonathan B. Losos, Richard E. Glor, s E Kolbe & Kirsten Aichalton 
Hybrid Speciation in Wild Sunflowers. Loren H. Rieseberg 
Adaptive Radiation and Evolution of Breeding d in Schuss (Caryophyllaceae). 
an Endemic Hawaiian Genus __ Ann K. Sakai, Stephen G. Weller. 
ae Warren L. Wagner, Molly Nepokroeff & Theresa M. Culley 49 
Contrasting Patterns and Processes of Evolutionary Change in the Tarweed—Silversword 
Lineage: Revisiting Clausen, Keck, and Hiesey’s Findings Bruce G. Baldwin 64 
Species Before Speciation Is Complete Peter R. Grant & B. Rosemary Grant 94 
Polyphyly in Guettarda L. (Rubiaceae, Guettardeae) Based on nrDNA ITS Sequence Data 
Frédéric Achille, Timothy J. Motley, Porter P. Lowry II & Joel Jérémie 103 
A Systematic Revision of Capparis Section Capparis (Capparaceae) —— 
Cristina Inocencio, Diego Rivera, 1 8 Obón, 
"rancisco Alcaraz & Jose-Antonio Barreña 122 
The Cardueae (Compositae) Revisited: Insights from ITS, trnL-trnF, and matK Nuclear and 
Chloroplast DNA Analysis Alfonso Susanna, Núria Garcia-Jacas, 
Oriane Hidalgo, Roser Vilatersana & Teresa Garnatje 150 


N 


4 
4 


uu 


Cover illustration. Capparis sicula subsp. mesopotamica Inocencio, D. Rivera, Obón « Alcaraz, 
drawn by Jose-Antonio Barreña. 


Annals 
-of the 
Missouri 
Dotanical 


Ga va 


Volume 93 


umber 


Annals of the 
Missouri Botanical Garden 


Volume 93, Number 2 
August 2006 


The Annals, published quarterly, contains papers, primarily in systematic botany, 
contributed from the Missouri Botanical Garden, St. Louis. Papers originating outside 
the Garden will also be accepted. All manuscripts are peer-reviewed by qualified, 
independent reviewers. Authors should write the Managing Editor for information 
concerning arrangements for publishing in the Annars. Instructions to Authors are 


printed in the back of the last issue of each volume and are also available online at 


www.mbgpress.org. 


Editorial Committee 
Victoria C. Hollowell 
Scientific Editor, 

Missouri Botanical Garden 
Beth Parada 

Managing Editor, 
Missouri Botanical Garden 
Diana Gunter 

Associate Editor, 

Missouri Botanical Garden 
Barbara Mack 

Editorial Assistant, 
Missouri Botanical Garden 
Monica Anderson 

MBG Press Assistant, 


Missouri Botanical Garden 


Ihsan A. Al-Shehbaz 
Missouri Botanical Garden 
Gerrit Davidse 

Missouri Botanical Garden 
Roy E. Gereau 

Missouri Botanical Garden 
Peter Goldblatt 

Missouri Botanical Garden 
Gordon McPherson 
Missouri Botanical Garden 
P. Mick Richardson 
Missouri Botanical Garden 
Charlotte Taylor 

Missouri Botanical Garden 
Henk van der Werff 


Missouri Botanical Garden 


For subscription information contact ANNALS 
or THE Missourt BOTANICAL GARDEN, Y Allen Mar- 
keting & Management, P.O. Box 1897, Lawrence, 
KS 66044-8897. Subscription price for 2006 is 
$160 per volume U.S., $170 Canada & Mexico, 
$195 all other countries. Four issues per volume. 
The journal Novon is included in the subscription 
price of the ANNALS. 


annals@mobot.org (editorial queries) 
http://www.mbgpress.org 


The ANNALS OF THE Missouri BOTANICAL GARDEN 
(ISSN 0026-6493) is published quarterly by the 
Missouri Botanical Garden. 2345 Tower Grove 
Avenue, St. Louis, MO 63110. Periodicals post- 
age paid at St. Louis, MO and additional mail- 
ing offices. Postmaster: Send address changes to 
ANNALS OF THE Missouri BOTANICAL (GARDEN, Yo 
Allen Marketing € Management, P.O. Box 1897, 
Lawrence, KS 66044-8897. 


The Annal bst Vor indexed in AGRICOLA (through 1994), APT Online, BIOSISO, CAB Ab- 
stract/Global Health ee ingenta, ISI databases, JSTOR, Research Alert®, and Sci Search®. 


Typesetting by Allen Press, Inc. using Advent3B2 automatic paging system. 


© Missouri Botanical Garden Press 2006 


The mission of the Missouri Botanical Garden is to discover and share knowledge about plants and 
their environment, in order to preserve and enrich life. 
This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 


Volume 93 Annals 
umber 2 of the 
006 Missouri 


Botanical 


haan 


MISSOURI DOTANICAL 
LATIN AMERICAN Alan Graham? 
BIOGEOGRAPHY—CAUSES AND AUG 3 1 | 2006 
EFFECTS. INTRODUCTION” ? 


GARDEN LIBRARY 


The 51st. Annual Systematics Symposium held at sperm biogeography and past continental move- 
the Missouri Botanical Garden, October 8-10, 2004, ments,” in the Annals of the Missouri Botanical 
had as its theme Latin American Biogeography) Garden. Vhat paper did much to revitalize interest in 

gra] pal 


Causes and Effects. The papers in this volume were biogeography and to modernize prevailing views about 
either presented at the symposium or are invited the physical context within which these patterns and 
papers by specialists on related topics. The fall of affinities originated. It still resonates well today 
2004 was an appropriate time to consider patterns of because it focused attention on two components 
plant and animal distributions within Latin America essential to understanding the arrangement of organ- 
and their affinities with other lands because it was isms and communities over the landscape. One is the 
30 years ago, in the fall of 1974, that Raven and need for a taxonomy that reflects distributions that are 


Axelrod published their landmark paper. “Angio- biologically meaningful rather than ones that are 


"This and the 10 articles that follow it are the proceedings of the 51st Annual Systematics Symposium of the Missouri 
Botanical Garden, Latin . American 1 5 270i — Causes and Effects. The symposium was held 8-10 October, 2004 at the 
La l 


Missouri Botanic ? Garden in St. Louis. Missouri. U.S 

This was the 49th Missouri Botanic alt jarden camel Systematics Symposium to be supported by a grant from the National 
cience Found 1555 (grant DEB-9981( ie \mong the many persons contributing to the planning and operation of the meeting 
and to the publication of the papers are Olga Martha Montiel (( r for Conservation and Sustainable Developing nt): Mic 
tichardson (Research. Graduate Studies) ina Freire-Fierro. 1 pun opez; Victoria Hollowell (Director, Missouri 
Botanical Garden Press). Carmen Ulloa Ulloa. Amy e e and Beth Parada: Donna Miller and Zubin Chandran 
ani s Technology): William Guy (Facility Support): and Mary MeNamara (oameni) This symposium, together with 
34 previous ones, again benefited conside 9 rom the d. support, and encouragement of Peter Raven. 

"This symposium volume is dedicated to Barbara W. Leyden, our esteeme 4. 'olleague and an invited speaker to the 
e who died in February 2006. She e MUN d herself as an Ps contributor to Quaternary vegetation history 


al "x inte rglac ial boundaries i 


and deed through papers on rapid. climate e and periods of aridity 
Guatemala and in the Lake Valencia Basin, Venezuela. Barbara gained further inte mod reputation and s ct for her 
later work on e environments of the Lowland Maya region of the Yucatán Peninsula and adjacent Central America. We join 
with our editor Victoria Hollowell in celebrating her critical and broad-based expertise and her vibrant mq s to the 
literature. 
'The editors of the Annals thank Sophia Balcomb for her significant editorial contribution to these papers. 
"Missouri Botanical Garden, P.O. Box 299, St. Louis. Missouri 63166, U.S.A. alan.graham(Omobot.org. 


ANN. Missouri Bor. GARD. 93: 173-177. PUBLISHED ON 23 Aucusr 2006. 


174 


Annals 
1 Ed Garden 


artifacts of nomenclature. The other is the need for 


sound models of landscape evolution and environ- 


mental 1 as a context for reconstructing 
migrations through time 
An outgrowth of this revitalization in biogeography 


has been the application of new information and 


analysis from molecular biology (phylogeography) to 
augment and test long-standing knowledge derived 
from traditional approaches. There are now more 
temporally exact models of plate 
1994; Pindell & Barrett, 1990; 


1999). and there is a better appreciation of its 


accurate and 
tectonics (Pindell, 
Mann. 
consequences, 


including the direction and rates o 


crustal movement, the reconstruction of land bridges. 
the uplift of mountain systems and d and the 


C hange 8 1 drainage 


— 


partitioning of. populations by 


systems. There is also an expanding 1 
basis for tracing the effects of these events on the 
distribution of lineages and communities and an 


increasing awareness of the 
techniques from independent disciplines now avail- 
able for reconstructing environmental change and the 
biotic responses (see below). 


Also important, but still in the formative stage, is 


development of a database of fossil plants that 


provides a more complete and reliable checklist of 


taxa and their age. Identifications of megafossils in the 
older literature are notoriously unreliable, as are many 
of the 


palvnological publications for Cretaceous and Ceno- 


biological affinities cited in stratigraphic 
¿ote plants. His incomprehensible in this modern era 
| | 


that key 


inaccessible. 


information for fossil plant remains is so 
These data are badly needed for dating 
phylogenies, calibrating molecular clocks, and con- 
ducting studies on vieariance and dispersal. They are 
also valuable as historical background for formulating 
speciation models and in making paleoenvironmental 
reconstructions as analogues for the effects of current 
The online electronic frame- 


environmental trends. 


work for such databases already exists, but a co- 
ordinated effort by paleobotanists will be required to 
filter into the system more complete records assessed 
for accuracy. 

The new analytical tools available. and the novel 
concepts being generated, are part of the reason for 
renewed interest in biogeography. There have been 
three major symposia devoted to the subject within the 
past three vears— Tropical. Intercontinental Disjunc- 
tions, organized by Susanne Renner and Thomas 
Givnish (2004): Long-Distance Dispersal, edited | 
Michael Cain, Ran Nathan, and Simon Levin (2003): 


and the present volume. 


— 


The goals herein are, first, to illustrate and update 


present and past distributions for several major groups 


of organisms, with particular reference to Latin 


impressive arsenal of 


America and other lands of the Southern Hemisphere. 
Second, to document some of the forcing mechanisms, 
especially plate tectonics, land bridges, and climatic 
change, that help explain former distributions and 
their residual patterns in the modern biota. Third, to 
show the variety of methodologies now available for 


documenting these forcing mechanisms and their 


biological consequences. The most enlightening of 


these methodologies often come from outside the 


biological sciences and include satellite imagery « 
almospheric events; recent innovations in quie lime. 
] | 


pale Bia pre 


detailed, | multi-factored computer 


animations (Scotese, 2004): geoc 5 mistry; geophysics; 


marine sedimentology: and the myriad of factors 
determining weather and longer-term climate. Rela- 
tionships between seemingly disparate observations 
are a rich source of new insights into biogeographic 
problems unresolved by traditional approaches. To 
this end, a fourth goal of the symposium was to bring 
together speakers not. traditionally. part of a single 
gathering to discuss results from diverse disciplines 
relevant to biogeography. 
Consistent with these goals, the discussions begin 


with a summary of the biogeographic history of 


Gondwanan vertebrates by David Krause and others 


and a historical perspective of South American land 
mammals by Rosendo Pascual. The next papers focus 


on development of the land bridges that connect North 


and South America and the history of migrations 
across those bridges. Evolutionary biologist Blair 


thinking about the 


West Indian vertebrate fauna within 


Hedges summarizes current 
evolution of the 
that physical context, and S. David Webb reviews the 
Great American Biotic Interchange, bringing a lifetime 
of experience to interpreting the complex events that 
led to the formation of the Panamanian land bridge 
and to documenting the biological consequences of 
the union of South America with North America after 
55 million years of separation. 

More recent events and the biotic responses that 


modernized the landscape, environments, and com- 


munities of Latin America in the Quaternary are 
discussed in three papers. Sarah Metcalfe reviews 
information from spores, pollen, diatoms, rodent 


middens, dendrochronology, and deep sea cores to 
climate for the northern 
deserts and the central Transvolcanic Belt of Mexico. 
Late 


events, documented initially for the high latitudes, are 


reconstruct. fluctuations 


Although the record. is complex, Quaternary 


shown to have affected northern and central Mexico. 


including a dry mid-Pleistocene, a more moist 
Holocene, and El Niño-Southern Oscillation (ENSO) 
cycles. In contrast to older views that conditions and 
communities became more stable and were essentially 


unchanged toward the tropics, a wealth of evidence 


Volume 93, Number 2 Graham 175 
2006 Introduction 
now documents that the climates and biotas at the pollen. record that the genus first appears at the 


lower latitudes had a dynamic and complex Quater- 
nary history. 

Human influence becomes a factor in shaping the 
America by 20,000 yr. BP, 


and that history is recounted by Dolores Piperno using 


biotas of lowland Central 


plant microfossils, charcoal, and, innovatively, a group 


of plant erystals called phytoliths and starch grains of 


agriculturally important plants preserved on grinding 
tools. 

One long-standing barrier to the detailed recon- 
struction of Quaternary vegetation history from high 
has been 


altitudes in the 


x 


tropic S the lack of 


continuous or near-continuous records. Cores from 


which paleobiological information on terrestrial biotas 
are retrieved mostly extend back only 


60.000 yr. BP 


previous interglacial at 


lo. approxi- 


mately less, and rarely into the 


100,000 yr. BP.“ 


that records must be spliced together from different 


his means 


sites and studied through a variety of techniques. The 
results have been less-than-consistent accounts of the 
climatic, biotic, and uplift history for several regions 
of the Neotropics (e.g., the Atacama Desert and the 
high elevations of Chile and Bolivia; Grosjean et al.. 
2003: Baker et al., 


for one 


2001). This limitation was removed 
site in northern South America when cores 
357 m and 586 m long, dating to 1.6 million and 3 


drilled 


sediments of the interandean basins around Bogotá, 


million years, respectively, were through 
Colombia. These cores were obtained by a group from 
the 


colleagues led by Henry Hooghiemstra. 


University of Amsterdam and their Colombian 
The history of 
montane vegetation is revealed in 


páramo and 


considerable detail from Funza, and these and other 
studies in the northern Andes are revealing abrupt 
Little 


and Dansgaard- 


short-term fluctuations in climate such as the 
lce Age, Younger Dryas, Heinrich, 
Oeschger events long recognized in sediments from 
the Northern Hemisphere. 

The fullest understanding of tropical ecosystems 
will require a seamless integration of studies on their 
history with those on the existing temporal versions. 
This synthesis has yet to be fully achieved, but there 


is some information on the Quaternary Atlantic Forest 


of Brazil (Ledru et al., 2005). This will complement 
nicely the studies of Rodolphe Spichiger and others 


(e. g., Perret et al., this volume) on the biogeographic 
history of the Gesneriaceae in the Atlantic. Forest via 
dispersal-vicariance analysis. 


My own contributions on the geol record of the 


gi 


^ 


Rhizophoraceae and on the origin of African-South 
American biotic affinities make two points not widely 
emphasized in the current biogeographic literature. 
With reference to the history of Rhizophora in the 


Caribbean region, it is clear from the extensive fossil 


middle to late Eocene transition (ca. 45 million years 
ago) and has been more or less lb present 
One the 


molecular evidence is that, rather than distributional 


ever since. interpretation. of emerging 


and taxonomic stasis, the genus disappeared for an 


— 
= 


estimated 3 million years between ca. and 11 
ago, that the black 


mangroves are derived from progenitors that arrived 


million years and modern 
ca. 11 million years ago. It is possible that both 
results are valid because the stenopalynous pollen 
establishes the presence of Rhizophora, as a genus, 
but 


does not preclude the extinction and re-introduction 


continuously in the region for 45 million vears, 


of the molecular-defined extant lineages throughout 
this interval. A duality of stasis versus more dynamic 
change will likely become commonplace as paleobio- 
geographic patterns and evolutionary histories from 
the fossil record and from molecular data are 
compared. 


My ol 


observations from different disciplines regarding the 


ier. contribution involves a synthesis of 


increasing number of elements that are recognized as 


— 


having migrated between Africa and the Caribbean 


South American region. This interchange has been 


poa e in a variety of ways, including vicari- 


ance in Cretaceous and Paleocene times, expressed 


primarily at the higher taxonomic categories; direct 
overland migration across the North Atlantic land 
bridge until about the middle Eocene; and long- 

8 8 


distance dispersal occurring throughout this time but 
becoming essential after about the middle Eocene. 

The relative importance of these overlapping modes 
depends on the dispersal potential of the propagules. 
the ecological requirements of the migrants, the time 
during. which the lineage under consideration. has 
been in existence, and the physiographic and climatic 
conditions prevailing in the region being transversed 


Observations 


+ 


at the time the migrations occurred. 
relevant to long-distance dispersal between Africa and 
South America are: (1) the enormous quantities of dust 
and organic debris being blown from the Sahara and 
Sahel into the Caribbean region, especially during 
periods of major hurricanes; (2) the increase in 
hurricane intensity, and possibly hurricane frequency, 
with global warming; and (3) the existence in the 
geologic past of intervals when global temperatures 


were considerably warmer than at present (e.g., during 


the early Eocene climatic optimum and in the middle 


Pliocene; Schmitz et al., 2000; Cronin & Dowsett. 
1991). Furthermore, the Eocene was the time of 
emergence of the Greater Antilles, and the distance 


between Africa and South America was about one-half 
to one-third less than at present. Thus, the opportunity 


for wind transport for some organisms and/or their 


176 


Annals of the 
Missouri Botanical Garden 


propagules may have been greater during these peaks 


f warmth than would seem plausible under modern 
conditions. There are extant. plant genera in which 
molecular evidence suggests disjunctions between 
Old World and New World 


(viz.. in the 


laxa originated ca. 34 


million years ago middle Pliocene; 


Graham. this volume). 
Collectively, the papers of the symposium illustrate 
some of the characteristics. and 


past and present 


affinities of the Latin American biota and the causes 


o its modernization. Interwoven throughout 


leading 
are examples of the various techniques being utilized 
This 


increasingly recognized as important to understanding 


o reveal their history. record is becoming 


the origin, development, and functioning of the 


present faunal and = floristic communities and in 


anticipating future environmental and biotic change. 
To emphasize the considerable array of techniques 
available for investigating 


ecosystems, the following list augments the ap- 


proaches noted in the papers of this volume (see also 


Graham, 1999; . 4—Methods, Principles, 

Strengths, and Limitations, 

(1) Iti OW Ds lo de ‘termine the Pw pus of 
COs and other greenhouse gasses in the Earth's 
nee are ‘the past 100 million vears, correlate 
the ió with temperature trends based on 
oxygen isotope ratios, date the events through 
paleomagnetic and e studies, an 
relate the reconstructed. elim to the history of 
lineages and communities. 1 5 is accomplished. by 
such diverse techniques as analyzing gas in air 
bubbles trapped in ice and by calculating the 
changing mass of the Earths carbonate sediments 
through time. It is also possible to calibrate the 
stomatal index of leaves from extant plants growing in 
environments of known COs concentration, as well as 
from herbarium sheets, and to ‘ppi the indices to 
fossil cuticles of different ages. In this way, past 
concentrations can be estimated, and the fluctuations 
can be compare ad with de mperalure treni from 
oxygen isotope studies in the marine realm 55 with 
vegetational histories. from the terrestrial. environ- 
ment. 

(2) Changes in albedo (reflectivity), precipitation, carbon 

“ag inks). and carbon release (pumps) can be 


specific places and times as the Fa 
a itc 4 cycles). 


(2 cal / enn / 
i. ugh 


position relative 
It is also aa that the ilm constal 


min) has varied) over time, and these 


fluctuations cannot be precisely quantified and do 


in concert with extinctions. 


no moto. vary 


the historical aspects of 


evolutionary peaks, migralional histories, they 
constitute a de ampe ning or amplif ving ove lay lo the 


Milankoviteh 


mechanisms. 


variations and other. climate-forcing 


( 1) The he ighi of mountains al various points in lime can 
be estimated by caleulatii the mean annual 
temperature of time- iube en ti | fossi floras from sea 
evel and from high elevations using the modern 
analogue method and/or leaf margin analysis (Gre- 
gorv-Wodzi 2000: Graham et al. 200 By 
applying the average global lapse rate ) m, 
an approximation of the KA of the highland flora 
can be made. These estimates, together with other 
geologie information, have revealed that about hall 
the elevation of the Central Andes has been attained 
in relatively recent times (viz... since. the late 
Miocene). That has implications for such topics as 
the time of origin of páramo environments and 
vegetation, speciation by vicariance the region, 
and the migration of temperate and high-altitude 
organisms along the axis of the high Andes. 

(5) The growth of stalagtites, and the midges and miles 
they often contain, is being used to estimate 
precipitation history in the deserts of the southeastern 
United States, in adjacent 1 0 rn. Mexico, and 1 
Brazil (e.2., Polyak et al., 1). 

(6) Amino acid racemization " n radiocarbon-dated. emu 
eggshe I] fragments is a function of temperature, and il 

s being used to reconstruct climates in 1 of the 
Southe vn Hemisphere (Johnson et al., 9), 

(7) \lkenones are long-chained organic in cules from 
algae that preserve in ocean sediments as bio- 
chemical fossils. The degree of unsaturatic Is 
temperature se nsitive (incre asing unsaturation pa 
cales cooling water) and this affords a method of 
detecting. El Niños during times besomd historical 
records. The modern eyele of El Niños along the 
Peruvian coast appears to have begun ca. 5800 vr. 
BP., and it corresponds to a time of intense monument 
building, while the onset. of even stronger events 
corresponds to their abandonment. 

(8) The solubility of noble gases is dependent on the 
temperature of the water, and "C-dated les s from 
eroundwater in Brazil indicate a cooling of 5 € at the 
Last Glacial Maximum ca. 18.000 P. This is one 
example from a wealth of new porem uon that 
indicates lowland quads | experienced significant 
cooling during the Quaternary. Whether it also 
experienced drying and, if so, when, where, and t 
what extent, is still unknown di ‘spite proc N to 
the contrary in the much-debated issue of refugia. 

(9) Titanium is a terrestrial mineral that is found in off- 
hore basins of Ven Decreases in the amounts 
along sections of mari cores reflect periods of 
re int ed runoff and river flow. and it is one indication 
of aridity associated with cooling in Amazonia during 
intervals of the Quaternary 

(10) Geographical information stems (GIS) tools. re- 


motely sensed data. and nic ie modeling techniques 
are being integrated with botanical data to study 1 je 


ecology and e volution of pla ni geographic ranges an 


spatial (Trisha Consiglio. 


Cons servation and Sustain- 
able De ye 19 nt, Missouri Botanical Garden: pers. 
comm... 

These and 1 5 . demonstrate how far we 


have come since the early great synthetic minds 


Volume 93, Number 2 Graham 177 
2006 Introduction 

biology laid the foundations. of biogeography. Lin- Gregory-Wodzicki, K. M. 2000. Uplift history of the central 
naeus, through the 186 dissertations of the Amoeni- 05 northern Andes: A review. Bull. Geol. Soc. Amer, 112 
tates Academicae, noted that the vegetation of distant 91-1105 


regions showed floristic similarities (e.g., Kamchatka 
and eastern North America in the 1 
that is, there were patterns in the 
Von 


from his observations 


Jonas P. Halenius): 
arrangement. of organisms over the landscape. 


Humboldt established in 1807. 


in the Ecuadorian Andes, that vertical changes i 
communities along mountain slopes were comparable 
to horizontal changes with latitude. He concluded that 
climate is an important factor in determining these 
patterns, and that did much to place the study of 
biogeography on a scientific basis. More recently. the 
time and the occurrence. of 


consideration. of pasl 


events has become 


increasingly recognized as a 
essential component of biogeography. as it was when 
geologist Lyell implanted the concept that lineages 
and communities extend back through time in the 
thinking of Darwin and Wallace on evolution. The 
papers presented here, and in other recent symposia, 
statement of the being made in 


are a progress 


understanding the causes and the implications of 
plant and animal distributions. The likes of Linnaeus. 
Von Humboldt, Darwin, and Lyell would undoubtedly 


be amazed and immensely pleased. 


Literature Cited 
Baker. P. A. et al. (+ 8 authors). 2001. 
changes at millennial and seba timescales on the 
Bolivian pm Nature 109: 698-701 
1. L., R. Nathan & S. A. 


nce s sal. Ecology 84: 1943-2020 


Tropical climate 


Levin ‘Glos: 2002. Long- 


dist 
Clapperton, C. 1993, Quaternary Geology and 5 
8 


ogy of South America. Elsevier Science Publ 


T. M. & H. J. 


climates. Quatern. Sci. Rev. 10: 115 


Cronin, ae ee 1991. Pliocene 


Graham, A. History of 


1999, Late Cretaceous a CM 
| Oxford 


orth American Vegetation (North of Mexico). 
Univ. Press, Oxford. 
Gregory-Wodzicki & K. L. 


nding | in Seaton ti al paleobotany. 


Wright. 2001. 

Mio-Pliocene 
DE from the Eastern C 9 ra, Bolivia: Implica- 
lions for the ame history of the Central Andes. Amer. J. 
Bot. 88: 1515-15 


750 dissertation of 


Mein M.. L Cartajena, M. A. Geyh & L. 


From prox 


2003. 
y data to Am e interpretation: The mid- 


Nunez. 


Holocene paradox of the Atacama Dese E Hohe rn C Bue: 


Palaeogeogr. PRIME 'oclimatol. Palaeoecol. 194: 247—2 


150. Plantae Rariores 7 hatcenses. 
Thesi dad: a of Uppsala, Upsala. 


Johnson, n. i Fogel, J. W. Magee, M. 


Halenius, J. 


Miller, M. L. 


Gagan & da i hivas. 1999, 65.000 vears of vegetation 
change in 1 n and the Australian summer 
monsoon. me lence 284: 1152. 


. D.-D. B 8 15 W. Cruz, Jr., C. 
. & L. Martin. 2005. 
during the last 100,000 vr from a record in the Brazilian 


ticcomini, 
Renin Pale climate changes 
Atlantic rainforest region and interhemispheric compari- 
son. Saas les. 44—450. 
s Caribbean Basins. 
erdam. 


Else; ier Science 
A .E E of the ( 
5 E. - 3—40 

‘ar ‘bbe an Ge. qus 


zulf of i xico and the 
\ K 


onovan & | Jackson 


An sedis l aan 


of the est Indies Publishers” Association (UWIPA). 
Mona, Kingston. 
— ——— & S. F. Barrett. 


1990. Geological evolution of the 
Caribbean re . A plate-tectonic pe rspective. Pp. 405— 
132 in G. Dengo & J. E. 
North America, Vol. 
Society of America. Bou 
Polyak, V. J., J. 
\smerom. 2001 
in the southwestern U States from mites preserved in 
stalagmites. Geology 29: 613—640. 
Raven, P. H. € D. J. sg 1974. 
ecography and a continental movements. Ann. Missouri 
Bot. ¢ Gard. 61: —6 7: 
. Giv vidi 
‘dis antions. 


tase (editors). The ( Geology of 


H. The Caribbean LS Geological 


„Colorado 


2. A. Norton & 


; FPEM her, 
. Wetter and cooler late Holocene e ‘climate 
li 


~ 
£ 


\ngiospe rm bio- 


Renner, (organizers). 2004. Tropica! al 
I 1 


intercontine en "lant Sci. 165 
SI-S138 

Schmit B. D. 
2000. 
dynamics. GFF 
1-189. 


Sundquist € F. P. 
ee ne 


Andreasson (editors 
warm Cia S an 
Society of 


Early mosphere re 
122: 


(Geological Sweden) 

Scotese, C. R. 
Changing terresti ial biogeographic pathways 
M. V. R. He 


904. Cenozoic and Mesozoic paleoge el 
. Pp. 9-26 in 
Frontiers of 


molino & L. aney (editors), 

Bioge pee New Directions in 110 Geography of Nature. 

ciates, Sunderland, Massachusetts. 

Von 1 A. 1807 
I 


a sur la Géographie des 


tant Levrault et Schoell, Paris. [See also, Ideen zu 
einer Geographie der Pflanzen, Ed eprint 1962, 


Wissenschaftlichen Buchgesellschaft, Darmstadt. | 


LATE CRETACEOUS David W. Krause,” Patrick M. O'Connor, 
TERRESTRIAL VERTEBRATES Kristina Curry Rogers,! Scott D. Sampson,” 
FROM MADAGASCAR: Gregory A. Buckley, and Raymond R. Rogers‘ 
IMPLICATIONS FOR LATIN 

AMERICAN BIOGEOGRAPHY!' 


ABSTRACT 


The iW Basin Project, initiated in 1993 and centered in Upper Cretaceous strata of northwestern Madagascar, has 
resulted i discovery of some of the most complete, well-preserved, and significant specimens of Late Cretaceous 
vertebrale a rom the Southern Hemisphere and indeed the world. Among the most important finds are various 
specimens of crocodyliforms, non-avian dinosaurs, and mammals; these finds have the potential to provide key insights into 
the biogeographic and paleogeographie history of Gondwana. Malagasa ar has been physically isolated from Africa for over 
160 million years and from all other major landmasses for more than 85 million years. The closest known relatives of many of 
the Late Cretaceous Malagasy taxa are peneconte piace ous forms from South America (primarily Argentina) and India, thus 
documenting a previously unrecognized high level of cosmopolitanism among Gondwanan vertebrates near the end of the 
Cretaceous. The family-level taxa that are shared among Madagascar, South America, and the Indian subcontinent are not 
known from penecontemporaneous horizons in mainland Africa, but it cannot yet be confidently determined if this is due to 


differential extinction, poor sampling. true absence (iter, the taxa were never present on Africa), or some combination thereof. 
Nonetheless, currently available geologic and paleontologie data are most consistent with the Africa-first model, suggesting 


] 
that Africa was the first of the major Gondwanan landmasses to be fulls isolated prior to the Albian/Cenomanian boundary, and 
that its terrestrial vertebrale faunas became progressively more provincial during the Cretaceous, while those on other 
g aes landmasses remained relatively cosmopolitan until the later stag wes of the Late Cretaceous 

ards: biogeography, Cretaceous, Madagascar, South America, vertebrates 


RESUMEN 


El proyecto de la cuenca de Mahajanga, iniciado en 1993 y centrado en los estratos del Cretácico supe rior en el noroeste de 


Madagascar, ha resultado en el descubrimiento de algunos de los especímenes más completos, mejor preservados y más 
8 | 

significativos de animales vertebrados del Cretácico tardío del hemisferio sur y de hecho de | ndr E ntre los hallazgos más 
importantes son los varios especímenes de crocodiliformes, dinosaurios no-aviares y mamíferos; estos hallazgos tienen el 
potencial de ade ur en "eru claves en la historia ELE da a y paleogeográfica de ( HU SA Madagascar ha estado 
fisicamente aislada de Africa por más de 160 millones de años y del resto de masas importantes de tierra por más de 85 

| i | I 

millones de anos. Los re cercanos conocidos de 1 de los taxones malgaches del Cretácico tardío son formas 
penecontemporáneas de América del Sur (sobre todo de Argentina) v la India, docume nando así un alto nivel cosmopolita 
previamente desconocido entre los verte . de Gondwana cerca del final del Cretácico. Los taxones al ne de familia 
compartidos entre Madagascar. América del Sur y el subcontinente Indio no se conocen i 1 ntes pet de 


Mrica continental, pero todavía no se puede " ‘le rminar con certeza si esto se debe a extinción dife renci uda: escaso muestreo, 

ausencia verdadera (1.e.. los taxones nunca estuvieron presentes en Africa), o una cierta combinación de esos. No obstante, los 

datos dd y paleontológicos actualmente 1 s son los más consistentes con el modelo de Africa-primero, que 

sugiere que Africa fue la primera de las masas importantes de tierra de Gondwana que estuvo alslada completamente antes del 

limite del Albiano/Cenomaniano. y que su fauna de vertebrados terrestres se volvió progresivamet 
los 


"más provincial durante el 
Cretácico mientras que aquellos en las masas terrestres de Gondwana permanecieron re a nte cosmopolitas hasta las 
ases finales del Cretácico tardío. 


'We thank the organize rs of the symposium and the editors of this volume for the invitation to participate: our collaborators 
at the University of Antananarivo and the Institute for the Conservation of Tropical Environments and the people of Berivotra 
for logistical support: M. Hallett and R. Ridgely for drawing Figures 7B and 5B, respectively, and L. Betti-Nash for executing 
and compiling all of the remaining figures: M. Carrano, C. Forster. : Georgi, J. Groenke, N. Kley, M. Loewen, T. Pinegar, A. 
Turner, and L. Witmer for advice on fossil restorations; L. Jacobs. aus y, and C. Scotese for discussion: P. Sereno and J. 
Wible for insightful review of the manuscript; and the National Science Foundation and the National Geographic Society iu 
financial support. 

“Department of Anatomical Sciences, Stony Brook University, Stony Brook. New York. 11794-8081. U.S.A. 
dkrause@noltes.cc vsb.edi 


Department of Biome dical Sciences, 22% 


Irvine Hall, io University, Athens, Ohio 457 155 
E: r 


aleontology Department, Science Mise un of Minnesota, est Ke Hoge Boulevard, St. Paul, once sola 55102, U.S.A. 
Utah Museum of Natural History and De ~ nt of Geology and Geophysics, University of Utah, 1390 East Presidents 
Cire 125 salt E City, Utah 84112-0050, U.S.A. 
^ Evelyn T. Stone University College, Bun sevell University. Chicago, Hlinois 00605, U.S 
“De iia ni of Ge ology, Macalester College, 1600 Grand Avenue. St. Paul, Minnesota 55105, U.S.A. 


— 
— 
2 


ANN. Missouri Bor. GARD. 93: 178— 


pond 
— 


PuBLISHED ON 23 Aucusr 2006. 


Volume 93, Number 2 
2006 


Krause et al. 
Late Cretaceous Terrestrial Vertebrates 


179 


The first question that logically might occur to 
“Why is Madagascar 


Latin 


readers of this contribution is 
relevant to a consideration of American bio- 


eeography?" The answer is not intuitively obvious 


because over 7500 km separate the two landmasses 

today and Madagascar's extant biota is highly endemic 
and imbalanced. However, in the context of a mobilist 
paradigm of Earth history, the question has different 
degrees of relevance depending upon the geologic 
time interval under consideration. We will assert in 
this paper that near the end of the Late Cretaceous the 
island-continent of Madagascar hosted a terrestrial 
that 
taxonomic composition to that of South America. This 


vertebrate fauna was strikingly similar in 
high degree of similarity is unexpected in the context 
of most recent paleogeographic reconstructions, which 
depict the southern. supercontinent Gondwana as 
been fragmented into its constituent 


having long 


landmasses by this time. Specifically, Madagascar, 
although still connected to the Indian subcontinent, is 
usually illustrated as having been physically isolated 
from all other Gondwanan landmasses since approx- 
imately 120 million years ago (Ma). This presents 

o 'ographie conundrum, as noted by Hay et al. 
(100 

i. configuration of Gondwana changed dramati- 
cally during the Late Jurassic and Cretaceous as it 


broke apart into isolated landmasses. The dispersion 


of these landmasses undoubtedly had profound 
consequences for the geographic distribution and 


subsequent evolutionary trajectories of the resident 
terrestrial. vertebrate faunas. Reconstructions of the 
Mesozoic fragmentation of Gondwana, however. are 
based almost exclusively on geophysical and strali- 
graphic evidence and remain poorly tested paleonto- 
logically. Recent discoveries of fossil vertebrates on 
from the Late 
, 1999: 


rom South 


southern landmasses, | particularly 


Cretaceous of Madagascar (cf. Krause et a 


2 


Krause, 2003), complement those known 
America and elsewhere and have profound implica- 


tions for testing hypotheses concerning the timing and 


sequence of Gondwanan breakup. Purportedly iso- 
lated in the Indian Ocean for over 85 million years. 
Madagascar is of unique biogeographic importance: il 
occupied a central geographic position within Gond- 


wana and was among the first (western margin) and 


— 


last (eastern margin) major landmasses to be involved 
in fragmentation of the supercontinent. 

The objective here is to provide an overview of the 
recovered 


terrestrial vertebrate assemblage recently 


from the latest Cretaceous stage (Maastrichtian) of 
Madagascar and to compare it with penecontempora- 
neous and earlier assemblages from South America 
and other Gondwanan landmasses in an attempt to 


elucidate biogeographic patterns. Terrestrial verte- 


brates are the focus of this report because they are 
tied to land (some presumably more than others) and 
are thus the most appropriate vertebrate animals (i.e. 
relative to, for example, fishes or birds) for examining 
biogeographie hypotheses involving subaerial land- 
masses. We specifically focus on erocodyliforms, non- 
avian dinosaurs, and mammals for the simple reason 
that their representation in the Gondwanan terrestrial 
fossil record is better than for most other terrestrial 
vertebrate clades and, as a result, their phylogenetic 
relationships are more highly resolved. This report is 
not a comprehensive analysis of Gondwanan bio- 
geography. Rather, it is an attempt to compare and 
evaluate the currently available data provided by 
southern 
South 

ig 
phi 


terrestrial vertebrate fossils from major 


andmasses, particularly Madagascar and 


America. in the context of recent paleogeogra 
We 


identify and draw attention to sampling problems that 


reconstructions of Gondwana. also strive to 
limit our ability to address particular biogeographic 
questions. 

LATEST CRETACEOUS 


TERRESTRIAL VERTEBRATES FROM THE 


OF MADAGASCAR 


The sample of latest Cretaceous (Maastrichtian) 
vertebrates from Madagascar is the most diverse and 
arguably the most significant in terms of completeness 


and preservation of specimens among Gondwanan 


assemblages outside of South America. This is 
particularly striking considering the small size of 


Africa, 


the Gondwanan landmasses, is over 50 times larger in 


Madagascar; indeed, mainland the largest of 


vertebrate 


area and yet its. penecontemporaneous 
fauna is much more poorly known. 

1e Late 
Cretaceous of Madagascar have been discovered a 
Mahajanga Basin Project (MBP), 
conducted jointly by Stony Brook University and the 
The MBP was initiated in 


1993 and has included eight expeditions, the most 


The vast majority of specimens known from th 
a result of the 


University of Antananarivo. 


— 


recent in 2005. It is focused on the vertebrate 
paleontology and geology of Upper Cretaceous strata 
in the Mahajanga Basin of northwestern Madagascar 
(Fig. 1). Most of the fossil vertebrate specimens have 
been recovered from a small study area (measuring 
approximately 20 sq. km) near the village of Berivotra 
and from one thin stratigraphic interval (measuring 
some 12 m thick). Existing stratigraphic data indicate 
the Anembalemba Member of the 
latest 


cannot discount the 


that this interval. 


Maevarano Formation, is of Maastrichtian 


Cretaceous) age: however, we 
possibility that lower reaches of the member might be 


2000). 


Member is of fluvial origin and accumulated in a semi- 


Campanian (Rogers et al.. The Anembalemba 


180 


Annals of the 
Missouri Botanical Garden 


| 
46° E 
ER M limestone 
n) 


E Berivotra Formation 

(Maastrichtian) 
Maevarano Formation 
(?Campanian-Maastrichtian) 

E Marovoay beds 

— (?Santonian-?Campanian) 
Ankazomihaboka beds 
(Coniacian-?Santonian) 


5 o S 


^ 


DEA D | 
RR e^ 


asalts 
(Coniacian) 


Berivotra 
Study Area 


e 
Ambato Boeni 


Berivotra | 1 
Study Area {f 


Morondava ` 
Basin 3 


— 
0 40 Km 
| | 


Figure l. 


The Berivotra Study Area in the Mahajanga Basin of northwestern Madagascar and the outerop area of Upper 
E PI 


Cretaceous and Paleocene strata in the basin. Que sion marks indicate unce rlainly conce rning age estimatio 


arid setting characterized by repetitive flood events 


that triggered fine-grained debris flows. These debris 
flows in turn served to entomb and preserve massive 
quantities of vertebrate skeletal material (Rogers et 
al.. 2000: 2005). 


Our ongoing efforts have dramatically increased the 


> ES 
Rogers, 


previously known (pre-1993) species diversity of Late 


Cretaceous vertebrates from the island (ef. Krause e 
al.. 1997a. 1999: 2003b). The 


known to include fishes, frogs, turtles, lizards, snakes, 


Krause, fauna is now 


crocodyliforms, non-avian dinosaurs. birds. and 


mammals. Many new genera and species have been 
taxa represented 


discovered and many of the higher 


are the first documented occurrences for the pre=Late 


Pleistocene of Madagascar (e.g.. frogs (Asher & 
Krause. 1998), lizards (Krause et al.. 20030). birds 


1996). mammals (Krause et al.. 1994,)) 


laxa represent the 


(Forster et al.. 


Furthermore, some of the higher 


fossil or 
1997b) 


among mammals), and 


only known occurrences. from Madagascar, 
gondwanatheres (Krause et al.. 


2001) 


and marsupials (Krause, 2 


Recent (e. g. 
some represent the first Late Cretaceous records from 
large portions. of Gondwana (e.g... lizards. birds, 
marsupials). 

Study of these vertebrate fossils and the sediments 
that encase them provides important information on 
the anatomy, paleobiology, ecology. and phylogenetic 
relationships of several vertebrate higher taxa and has 


resulted in important insights into, among other 
topics, the origin and early evolution of birds (Forster 
1996, 1998. 2003: Forster & O'Connor, 2000: 
Chinsamy & Elzanowski. 2001): the Gondwanan 
anch distribution of erocodyliforms 
1997, 2000: 
Rasmusson. 2002: 


2004a. b). 


et al.. 


diversification 
(Buckley et al.. Buckley & Brochu, 
19909: 
2001: 


Rasmusson & Buckley. 


Turner, saurischian dinosaurs 


Volume 93, Number 2 
2006 


Krause et al. 181 
Late Cretaceous Terrestrial Vertebrates 


Table 1. Crocodyliform genera from the Anembalemba 


Member of the 


(Maastrichtian). Mahajanga Basin, Madagascar. 


Maevarano Formation. Upper Cretaceous 


Mesoeucrocodylia 
Metasuchia 
Notosuchidae 
Simosuchus Buckley, Brochu, Krause & Pol 
Peirosauridae 
Genus indet. 
Trematochampsidae 
Trematochampsa Bulletaut € Taquet 
Metasuchia incertae sedis 
Iraripesuchus Price 
VI. Buckley & Brochu 
Genus indet. 


Genus 1 


(Sampson et al., 1998, 2001: Curry Rogers & Forster. 
2001, 2004; Curry Rogers, 2002, 2005: Carrano et al. 
2002. 2004: O'Connor & Claessens. 2005; O'Connor, 
in press). and mammals (Krause et al., 1997b: Krause. 
2001): the highly 
endemic 
et al. 1997a, 


biogeographic origins of the 
vertebrate fauna (Krause 


2003b) and the 


Malagasy 
1999: 


extant 
Krause. 
stratigraphy, 
Upper Cretaceous rocks in the Mahajanga Basin 
(Krause & Hartman, 1996; Rogers et al., 2000, 2001: 
2003; 2005). 


significantly, these discoveries have profound impli- 


Casey et al., Rogers, Perhaps most 
cations for testing biogeographic hypotheses that. in 
turn, address broader questions concerning the timing 
and sequence of Gondwanan fragmentation (Krause el 
al., 19072. b, 1999: 
2003b). 


The record of Late Cretaceous terrestrial verte- 


Sampson et al., 1998; Krause. 


brates from Gondwana is spotty at best and therefore 
difficult to analyze and interpret in a biogeographic 


context. Among the vertebrate groups represented in 


the Late Cretaceous of Madagascar, the best known in 


terms of their completeness and preservation. and 


therefore the most precisely identified, are the 


crocodyliforms, non-avian dinosaurs, and mammals. 


These taxa are therefore the most relevant for 


81 81 a e L: ons 1 Tha 
consideration of biogeographic relationship The 


following provides a brief overview of the taxonomic 


and anatomical diversity of each of these taxa. 


CROCODY LIFORMS 


At least crocodyliforms, all 
metasuchians, are present in the Maevarano Forma- 
tion (Table 


therefore not. surprising that a range of adaptations 


seven species of 


1). This diversity is extraordinary, and it is 


suggestive of habitat specialization is evident. In 


addition to extreme differences in body size and form, 


sedimentology, and geochronology of 


ranging from small and gracile to large and ponderous. 
there is a broad range in skull shape. from robust and 
broad to long and slender to short and blunt. 
Mahajangasuchus insignis (Fig. 2). originally de- 
nearly complete skeleton 


seribed on the basis of a 


acking only the skull, is the first crocodyliform genus 
and species from the Late Cretaceous of Madagascar 
to be named on the basis of MBP discoveries (Buckley 
& Brochu, 1999). Since publication on this specimen, 
two nearly complete skulls have been discovered 
(Buckley & 2001). 


a large carnivore, measuring almost 4 m in length, and 


Brochu, Mahajangasuchus was 
is distinctive among crocodyliforms in exhibiting an 
extremely broad and flat, hippopotamus-like snout 
and deep lower jaw. 

Simosuchus clarki (Vig. 3) is a new, bizarre, pug- 
nosed species represented by a complete skull and 
articulated anterior half of a posteranial skeleton that 
were preliminarily described by Buckley et al. (2000). 
The 


oventrally positioned occipital condyle, underslung 


lunt. shovel-like snout, deep cranium. poster- 


lower jaw, and areas for extensive neck musculature 
are suggestive of burrowing adaptations, whereas the 
anteriorly positioned jaw joint and clove-shaped teeth 
may reflect adaptations for herbivory (Fig. 3). 

A new species, Araripesuchus sp. indet., is being 
described by A. Turner (in press). It is known from 
skulls and skeletons of at least five individuals, all 
recovered from a single locality, that reveal a small, 
eracile-limbed form (Fig. 4). Araripesuchus tsangal- 
sangana represents the geologically youngest occur- 
rence, yel most primitive known member of the genus 
2004a, b; in press). 


Trematochampsa oblita Buffetaut & Taquet, named 


(Turner. 


on the basis of three fragmentary dentaries, is the only 


— 


erocodyliform species from the Late Cretaceous of 
Madagascar that was described prior to inception of 
the MBP 1979). A 


preserved. dentary, as well as skull and 


(Buffetaut & Taquet, better- 
several 
posteranial elements, were recovered by MBP expedi- 
tions (Rasmusson & Buckley, 2001). The new material 


of T. oblita. provides the opportunity for a better 


understanding of phylogenetic relationships of trema- 
(2002) 
trematochampsids are monophyletic, but 
genus Trematochampsa is paraphyletic. Although still 
clear that T. oblita, like 


Mahajangasuchus, was a massive animal and likely 


tochampsids; Rasmusson determined that 


poorly known, ib is 


one of the top predators of its time. 


named erocodyliform taxa, 


n addition to these 
there are at least three as yet unnamed species 
recovered from the Maevarano Formation. One is 
represented by several skull elements preliminarily 
identified as peirosaurid. The second is a medium- 


sized (i.e.. smaller than Mahajangasuchus insignis and 


182 Annals of the 


Missouri Botanical Garden 


Mahajangasuchus insignis 


10 cm 


H 


50 cm 


Figure 2. Mahajangasuchus insignis Buckley & Brochu, a large. broad-mouthed metasuchian erocod 


vliform from the Late 
Cretaceous of Madagascar. Skull in left lateral view and restoration of the skull and skeleton in left le 


aleral view 


Trematochampsa oblita, but larger than Araripesuchus 


laquet, 1979; Sues, 1980; Fig. 5). One of tl 


le most 
spectacular fossil discoveries of the MBP to date is an 
exquisitely preserved and virtually complete skull and 
al. lower jaws of M. gp. discovered in 1996 
2003). The third is poorly known, represented only by (Sampson et al., 1998). 
isolated elements, primarily partial skull and jaw bears fused 
material. 


sp. indet.), slender-snouted metasuchian, represented 
by a partial skull. a complete lower jaw, and 


considerable posteranial material (Buckley et 


The short and deep skull 
nasal bones with a large interior 
pneumatic chamber, and a midline projection. or 
“horn” positioned above the eyes. with a notable 
NON-AVIAN DINOSAURS parietal eminence capping the skull roof posteriorly, 
Complementing this specimen are several other, more 
Dinosaur material is abundant in the Maevarano 
Formation, but to date is limited to the Saurischia. At 


fragmentary, and less well-preserved skulls discov- 


ered recently, in addition to three partial skeletons, 
least four non-avian species have been discovered: one of them representing a juvenile individual; thus, 
two theropods and two sauropods (Table 2). virtually all bones of the skeleton are now represented 
The best represented of the two theropods is the — for this animal 


the primary exceptions being parts of 
the pelvis and the distal portions of the forelimbs). 
o be These specimens confirm referral of Majungatholus to 
a pachycephalosaurid (“dome-headed” ornithischian 


mid-sized (approximately 6.2 m in length) abelisaurid 
Majungatholus  atopus, previously thought t 
the Abelisauridae and are described in detail in 


dinosaur) based on fragmentary skull material (Sues & a monograph-length study (Sampson & Krause. ii 


Volume 93, Number 2 
2006 


Krause et a 183 


| 
Late Cretaceous Terrestrial Vertebrates 


Simosuchus clarki 


10 cm 


Simosuchus clarki Buckley, Brochu, Krause 


Figure 3. 


& Pol, 


small- to mid-sized metasuchian crocodyliform from the 


Late Cretaceous of Madagascar. Skull in left lateral view and restoration of the skull and skeleton in left lateral viev 


press). Numerous bones in the Maevarano assemblage 
exhibit tooth marks (some of which even preserve 
denticle drag marks) that can be attributed only to 
thus carnivorous 


Majungatholus, attesting to ils 


habits. Interestingly, many of the bones are those of 
Majungatholus itself, thereby providing the first well- 
documented evidence for cannibalism among dino- 
saurs (Rogers et al., 2003). 

Masiakasaurus knopfleri, measuring approximately 
1.8 m long, was a much smaller and more. gracile 
predator than Majungatholus atopus (Fig. 6). It is 
represented by isolated bones of both the skull and 


postcranial skeleton and is thus less completely 


known. Masiakasaurus is remarkable. however, in 
having an anterior dentition that consists of pro- 


cumbent, spearing teeth, a unique condilion among 


Dinosauria. Masiakasaurus was described by Sampson 


et al. (2001) and Carrano et al. (2002), who concluded 
that this Malagasy taxon should be included within 
the small-bodied abelisauroid clade Noasauridae, 
previously known only from Argentina. The placement 
of Masiakasaurus—together with the penecontem- 
poraneous Indian form Laevisuchus—into Noasauridae 


therefore 


(previously known only from Argentina) 
ereatly expanded the geographic range of this clade to 
encompass much of Gondwana. This finding demon- 
strates that, at least in a number of ecosystems, 
noasaurid abelisauroids were small-bodied counter- 
parts to their larger-bodied cousins, the abelisaurids, 


it 


> 


a manner parallel to small-bodied maniraptoran 


coelurosaurs (e.g.. troodontids, oviraptorosaurs, dro- 
and large-boded tyrannosaurids in many 
Recent field- 


has produced additional 


maeosaurids) 


— 


ale Cretaceous Laurasian ecosystems. 


work in 2003 and 2005 


184 Annals of the 
Missouri Botanical Garden 


Araripesuchus sp. indet. 


9 cm 
Figure d. Araripesuchus sp. indet.. a small. gracile-limbed metasuchian erocodyliform from the Late Cretaceous of 
Madagascar. Skull in left lateral view and restoration of the skull and skeleton in left lateral view. 
skeletal remains of Masiakasaurus, including several Titanosaurian sauropods were among the first 
key. previously unrepresented cranial and posteranial — fossils described from Madagascar and are abundant 


elements that will significantly increase knowledge of components of the Late Cretaceous fauna. At least two 


species of lithostrotian titanosaurs are present in the 


noasaurid anatomy (Carrano et al., 2004, in prep. 
fauna, and at least one possessed osteoderms ( 


Table 2. Non-avian dinosaur genera from the Anem- 5 3 : 
bones”) that range in size from small ossicles to bony 


balemba Member of the Maevarano Formation. l pper | | 10 in di Dené 189 
: Doha : : dales greater than em in diameter (Veperet. 1890: 
Cretaceous (Maastrichtian), Mahajanga Basin, Madagascar. pr o B (Jer 


Dodson et al.. 1998). 


Theropoda Rapetosaurus krausei is known from approximately 
Ceratosauria 90% of its skeleton, including several well-preserved 


\belisauroidea associated Specimens representing a range of ontoge- 
Abelisauridae nelic stages (Curry Rogers & Forster, 2001, 2004: 
Vajung s Sues & Taquel : " l i 
lajungatholus Sues & Taqui Curry Rogers, 2005; Fig. Rapetosaurus is partic- 
Noasauridae " Pp . . 
ularly significant. because it preserves. cranial and 


Masiakasaurus Sampson, Carrano & Forster Ñ i . . 

posteranial data key to elucidating phylogenetic 

Sauropoda . / è A 
relationships among one of the most temporally and 


"Manosauria 
Phases geographically widespread of dinosaurian groups. 
Saltasauridac Rapetosaurus has already provided the first cranial 


Genus indet, (Malagasy Taxon B of Curry Rogers, data in support of Tit if s monophyly and 
2002) has helped to increase resolution of lower-level 
Lithostrotia incertae sedis Uitanosaurian relationships (C urry Rogers & Forster, 
J > " * x a - 2 y A ^ 
Rapetosaurus Curry Rogers & Forstei 2001: urn y Rogers, 2005). 


Volume 93, Number 2 
2006 


Krause et al. 
Late Cretaceous Terrestrial Vertebrates 


185 


Majungatholus atopus 


1m 


a 
mein in defi lateral view 


“Malagasy Taxon B” (Table 2) is the second 
titanosaur species from Madagascar. Although repre- 


sented by only several isolated caudal vertebrae, 
a series of articulated distal caudals, and an isolated 
coracoid (Curry Rogers, 2002, 2005), Taxon B can be 
easily distinguished from Rapetosaurus. The caudal 
centra of Malagasy Taxon B are distinctively dorso- 
ventrally compressed and subrectangular in transverse 


section. In contrast, Rapetosaurus caudal centra have 


subequal height and width dimensions resulting in 


a round transverse section. Coracoids are also di- 


Majungatholus atopus Sues & Taquet, a mid-sized theropod « dinosaur from the 
and restoration of the skull and skeleton in left lateral v 


ate Cretaceous of Madagascar. 


agnostic: Malagasy Taxon B exhibits a square coracoid 
with broad scapular facet, whereas Rapetosaurus has 


a rounded coracoid with narrow scapular articulation. 
Curry Rogers (2005) recently conducted a phyloge- 


netic analysis and concluded that Malagasy Taxon B 


[om 


and Rapetosaurus are distant relatives within Titano- 
sauriformes (Curry Rogers, 2005). Even more signif- 
icantly, the compressed caudal vertebrae and low 
neural spines of Malagasy Taxon B indicate that it is 
a member of 


2003) 


Saltasaurinae (sensu Wilson & Up- 


church, with close affinities to the South 


Annals of the 


Missouri Botanical Garden 


Masiakasaurus knopfleri 


Figure G. Masiakasaurus knopfleri Sampson, Carrano & 
Madagascar. Skull in left lateral view and restoration of the 


American saltasaurine genera Saltasaurus, Neuquen- 


saurus, and Rocasaurus. 


MAMMALS 
The Mada- 


gascar is still poorly understood. lt is represented by 


Late Cretaceous mammalian. fauna of 
five, or perhaps six, taxa, each known from nothing 
more than a fragmentary tooth or two (Krause et al.. 
1994, 1997b; Krause & Grine, 1990: 2001, 
2002: Table 3). One still-undescribed taxon. however. 
is represented by a nearly complete, well-preserved, 
03a). 


ly, none of the known taxa can be considered to be 


Krause, 


and articulated skeleton (Krause, 2X Interesting- 


potential ancestors of the island’s highly endemic 
extant mammalian fauna, all of which are placentals. 
Two fragmentary mammalian teeth were assigned to 


a new genus and species, Lavanify miolaka Krause. 


Forster, a small theropod dinosaur from the Late Cretaceous of 
skull and skeleton in left lateral vi 


2A 
z 


Sahni & Grine, a sudamer- 
1997b). 


mammals i 


Prasad, von Koenigswald, S 
icid gondwanathere (Krause et al.. Sudamer- 


icids are unique among Mesozoic 
possessing hypsodont cheek-teeth separated from 
eliriform incisors by a pronounced diastema. Just as 
for noasaurid theropods (see above), the discove ry of 
sudamericids in Madagascar and the Indian sub- 
continent considerably extended the geographic range 
of the clade (previously known only from Argentina) 
and provided evidence for a previously unknown high 
degree of cosmopolitanism that encompassed both 
western and eastern Gondwana (see below). 

M least one other tooth may also be assigned 
Gondwanatheria (listed as Mammalia incertae sedis 
A in Table 3 
and considerably lower erowned than those assigned 
to Lavanify (Krause, 2000). Still less high-crowned is 


another isolated tooth that exhibits a complex occlusal 


Genus indet. ). although it is much larger 


Volume 93, Number 2 
2006 


Krause et 
Late Cretaceous Terrestrial Vertebrates 


al. 187 


Rapetosaurus krausel 


Figure 7. 


Rapetosaurus krausei Curry Rogers & Forster, a tilanosaurid sauropod dinosaur from the 


Late Cretaceous of 


Madagascar. Subadult skull in left lateral view and restoration of the skull and skeleton in left lateral view. 


morphology (W-shaped ridge with two infundibula 
separating each of the three limbs of the W) on an 
could 


obliquely-oriented surface. This specimen 
conceivably also pertain to Gondwanatheria but, 
Table 3. Mammalian genera from the Anembalemba 


Member of the Maevarano Formation, Upper Cretaceous 


Maastrichtian), Mahajanga Basin, Madagascar. 


Mammalia 
Multituberculata 
Genus indet. 
Marsupialia 
Genus indet. 
Gondwanatheria 
Sudamericidae 
Lavanify Krause, Prasad, von Koenigswald, ' 
Grine 
Mammalia incertae sedis 
;enus indet. À 


Genus indet. B 


Sahni & 


based on current information, it could equally well 
belong to a previously unrecognized higher taxon of 
Mesozoic mammals. 

Another isolated tooth, a tribosphenic lower molar, 
as that of a marsupial 


— 


was regarded by Krause (2001 
and therefore as potentially the earliest evidence of 
marsupials in the Southern Hemisphere. This identi- 
fication was confirmed by Case and Krause (2002), 
but has since been disputed by Averianov et al. 
(2003 


detailed analysis that supports the marsupial affinities 


; Case (in prep.) is nearing completion of a more 


— 


of the taxon represented by this specimen. 
Another molar 
exhibiting cusp-in-line morphology characteristic. of 


multituberculates, a clade that is well represented and 


isolated specimen is a fragment 


extremely diverse in penecontemporaneous, as well as 
earlier and later. horizons in Laurasia (Krause & Grine, 
1996). Multituberculates are poorly known from the 


Mesozoic of the Southern Hemisphere, previously found 


188 


Annals of the 
Missouri Botanical Garden 


only in the Early. Cretaceous of Morocco (Sigogneau- 


Russell, 1991: Hahn & Hahn, 2003) and possibly the 


Late Cretaceous (Campanian or early Maastrichtian) of 


Argentina (Kielan-Jaworowska et al., 2004). 

taxon (listed 
Bin 3) 1s 
represented by a virtually complete, articulated, sub- 
adult skeleton (Krause, 2003a). Thi 


Finally, one recently discovered 


Mammalia incertae sedis Genus indet. ‘able 
This skeleton, still in 
the process of being prepared and studied, represents 
the largest and most complete specimen of a mammal 
yet known from the Mesozoic of Gondwana. It exhibits 
a striking mosaic of primitive (e.g.. septomaxilla with 
prominent septomaxillary canal, epipubic bone) and 
derived. (e.g, specialized dentition with prominent 
diastema, well-developed humeral trochlea, reduced 
fibular-calcaneal contact) features. There is no doubt 
that this specimen will introduce a substantial amount 
of character conflict 


mammalian phylogeny. It is also safe to conclude, 


even al this preliminary stage, that this animal cannot 
be “shoehorned” into any currently recognized higher 


taxon of mammals; it 


represents a major new 


nontherian clade. The relative completeness of this 


extraordinary specimen promises to elucidate the 


anatomy. functional morphology, and phylogenetic 


position of the clade it represents. 


TERRESTRIAL VERTEBRATES FROM THE 
ANKAZOMIHABORA SANDSTONES 
Terrestrial vertebrate fossils have also been re- 
covered from the Ankazomihaboka sandstones. a unit 
that is purportedly interbedded with, and definite ly 
basalts of Coniacian age (Besairie, 1972; 
Storey et al., 1995, 1997). 


(Curry, 1997) indicates the presence of at least three 


overlies, 
V preliminary report 
erocodyliform taxa and two to three taxa of nonavian 


dinosaurs, including sauropods and theropods; no 


mammals were found. Subsequent collections have 


been made and a more extensive analysis « 1e 


assemblage from this unit is under way (Curry Rogers 
et al. in prep). The taxa derived from the 
Ankazomihaboka sandstones are not considered here 
because of the tentative nature of the current 
identifications and because the age of the rock unit 
has not been ascertained, although. conservative sly, 

can be constrained to have been as early as 1 1 


and as late as Maastrichtian. 


CAMPANIAN/MAASTRICHTIAN CROCODYLIFORMS. 
Non-AVIAN DINOSAURS. AND MAMMALS FROM 
OTHER GONDWANAN LANDMASSES 


The primary purpose of this section is to compare 


the latest Cretaceous. terrestrial vertebrate assem- 


into previous. topologies of 


blages of Madagascar and South America. However. 
e 7 


similarities between the assemblages 
t the 


Cretaceous make it clear that physical and biotic 


the taxonomic 


from these two. landmasses : end of the 


connections with other landmasses were also involved 
and must be considered. An overview of penecontem- 
poraneous terrestrial faunas from potentially con- 
Africa. 


warranted. 


nected landmasses—namely Antarctica. and 


the Indian subcontinent—is therefore 
Australia and southern Europe also contained Gond- 
wanan-aspect faunas during the Late Cretaceous, but 
that 


either was directly connected to Madagascar or South 


there is no independent evidence indicating 


America during this interval; as such, and owing to 
space considerations, their faunas are nol considered 
here. 


The faunal lists in Tables 4-6 provide an overview 


of the generic diversity of erocodyliforms. non-avian 


2 


dinosaurs, and mammals known from the Campanian 


Africa, 


Campa- 


and Maastrichtian stages of South America. 
Antarctica, and the Indian subcontinent. The 
nian and Maastrichtian, the last two stages of the Late 
Cretaceous, represent a broader time slice (approxi- 
mately 18.5 million years) than is likely represented 
Member of the 


Formation in Madagascar. 


by the Anembalemba Maevarano 


Nonetheless, the uneven 
comparison is necessitated, at least in part, because of 
the poor sampling of penecontemporaneous Gondwa- 
nan horizons and because of the imprecise dating of 
it is better 


productive strata: given these constraints. 


be too inclusive rather than too exclusive. The 


faunal lists have been extracted. from numerous 
sources for crocodyliforms (e.g... Gasparini et al., 
1996; Bertini & Carvalho, 1999; Wilson et al.. 2001: 
Prasad & de Lapparent de Broin, 2002), but primarily 
from Weishampel et al. (2004) for non-avian dinosaurs 
(2004) for mammals, 


with other sources noted in table headings. 


and Kielan-Jaworowska et al. 


SOUTH AMERICA 


Sou 


h America has the largest and most diverse 
samples of Gondwanan Late Cretaceous. vertebrates, 
many from Campanian and Maastrichtian horizons. 


Indeed, there are over 40 family-level taxa of 


vertebrates known from the Campanian 


and Maastrichtian of South America. than 


more 


double the number from any other Gondwanan 


landmass. This is in large part the result of intense 
sampling over the past several decades, much of i 
owing to the efforts of José Bonaparte of the 117 
Argentino de Ciencias Naturales Paleontologia (Bue- 
nos Aires) and his students. 

lt is not possible to determine which of the family- 


level taxa currently unknown from. elsewhere were 


Volume 93, Number 2 
2006 


Krause et al. 
Late Cretaceous Terrestrial Vertebrates 


Table 4. Crocodyliform genera from Campanian and 
Maastrichtian horizons in South America, Africa, 
Antarctica, and the Indian subcontinent. List for Soutl 


America compiled primarily from Gasparini 
Bertini and Carvalho (1999), for Africa from Buffetaut (1982) 
and Brochu (1997), and for the Indian subcontinent from 
Wilson et al. (2001) 
(2002). 


and Prasad and de Lapparent de Broin 


SOUTH. AMERICA 
Mesoeucrocodylia 
suchia 
Notosuchidae 
Uruguaysuchus Rusconi 
Baurusuchidae 
Cynodontosuchus Woodward 
Baurusuchus Price 
Stratiotosuchus Campos, Suarez, Riff & Kellner 
Peirosauric 
MUN Price 
Lomasuchus Gasparini, Chiappe & Fernandez 
Uberabasuchus Carvalho, Ribeiro & dos Santos Avilla 
Trematochampsidac 
ltasuchus Price 
Metasuchia incertae sedis 
Sphagesaurus Price 
Neosuchia 
Dyrosauridae 
Suleusuchus Gasparini & Spalletti 
Hyposaurus Owen (questionably from Maastrichtian) 
Eusuchia 
Dolichochampsidae 
Dolichochampsa Gasparini & Buffetaut 
AFRICA 
Mesoeucrocodylia 
Neosuchia 
Dyrosauridae 
Sokotosuchus Halstead 
ANTARCTICA 
No record 
INDIAN SUBCONTINENT 
Mesoeucrocodylia 
Metasuchia 
Baurusuchidae 
Pabwehshi Wilson, Malkani & Gingerich 


Metasuchia incertae sedis 


~ 


enus indet. 
Neosuchia 


Dyrosauridae 


~ 


senus indet. 


indeed restricted to South America and which were 
more broadly distributed in the Southern Hemisphere, 
simply because of the relative paucity of discoveries 
on other Gondwanan landmasses. However, it is 
significant to note that of the seven family-level taxa 
of crocodyliforms (Notosuchidae, Peirosauridae, Tre- 
matochampsidae). non-avian dinosaurs (Abelisauri- 


dae, Noasauridae, Saltasauridae (including Saltasaur- 


inae)), and mammals (Sudamericidae) that have been 


identified as occurring in the Maastrichtian of 


Madagascar, all are known from the Campanian/ 
Maastrichtian of South America. This is suggestive of 


close biogeographic ties and, indeed, a degree of 
cosmopolitanism that was unexpected before the MBP 


discoveries and that, moreover, is difficult to coni 


in the context of mosl recent pale og ographic 
reconstructions. 
Crocodyliforms known from Campanian and Maas- 


South 


comprising at least 12 named genera, most of them 


trichtian horizons. in America are diverse, 
metasuchians. In addition to notosuchids, peirosaurids, 
and trematochampsids, the metasuchian family Baur- 
usuchidae is represented. Neosuchians and eusuchians, 
for which definitive evidence has vet to be found in the 
Late Cretaceous of Madagascar, are also present in 
South America. The non-avian dinosaur fauna from the 
South 


dominated by saurischians. Among Theropoda, abeli- 


American Campanian and Maastrichtian is 


sauroid ceratosaurians (including both abelisaurids and 
noasaurids) are the most diverse and, among the 
Sauropoda, lithostrotians, including saltasaurines, dom- 
inate. Ornithischians are represented by spotty occur- 
rences of ankylosaurs, euornithopods, and hadrosaurs. 
It is intriguing that, in contrast to the dominance of 
ornithischian herbivores in most Campanian terrestrial 
ecosystems on Laurasian-derived landmasses, or- 
nithischians appear to be only minor components of 
Campanian ecosystems on most Gondwanan land- 
1996). 


Campanian and Maastrichtian mammals from South 


masses, | present at all (Currie, 


America include a div ersity of are pens nontribosphenic 


including Austrot 


forms, . Bondesiidae, 


Brandoniidae, ela o Mesun- 
eulatidae, Reigitheridae, and Sudamericidae. Of these, 
the Sudamericidae is known to also occur in 
1997b). In addition to this 


diversity of nontribosphenic taxa, there are two Spes 


only 
Madagascar (Krause et al., 
of eutherians and questionable occurrences of “pedio- 


myid” and peradectid marsupials. 
ANTARCTICA 


Not unexpectedly, considering its ice-cover and 


harsh climate today, Antarctica has grudgingly 


yielded fossils of Late Cretaceous terrestrial verte- 
brates. They have been recovered from the Antarctic 


Peninsula 


— 


Vega and James Ross Islands) and include 
only specimens of dinosaurs (Hooker et al.. 199] 
Gasparini et al., 1996; Rich et al., 1999; Case et al., 
2000, 2003). This appears to be largely the result of 
float” 


carcasses are washed oul to sea and skeletal elements 


“bloat and taphonomic scenarios, where 


are buried in, and recovered from, marine sediments. 


190 


Annals of 
Missouri Botanical Garden 


Table Non-avian dinosaur genera from Campanian 
and Maastrichtian horizons in South America, Africa, 
Antarctica, and the Indian subcontinent. Lists compiled 


et al. (2004), Een additions from Novas 
Agnolin (2004), Novas et al. (2004) for South America, 
Wilson and Upehureh (2003) for 11 and Suberbiola et al. 
(2004) Africa. 


placement. 


for Question marks indicate tentative 


SOUTH AMERICA 
Theropoda 
Ceratosauria 
\belisauroidea 
\belisauridae 
Carnotaurus Bonaparte 
\belisauridae indet. 
Noasauridae 
Voasaurus Bonaparte & Powell 
Tetanurae 
vetheropoda 
Quilmesaurus Coria 
Maniraptora 
Unquillosaurus Powell 
?Oviraptosauria indet. 
Coelurosauria indet. 
Theropoda indet. 
Sauropoda 
Lithostrotia 
Saltisauridae 
Saltasaurus Bonaparte & Powell 
Lithostrotia incertae sedis 
leolosaurus Powell 
Intarctosaurus Huene 
Laplatasaurus Huene 
Neuquenosaurus Powell 
Pelliginisaurus Salgado 
Rocasaurus Salgado & Azpilicueta 


Lithostrotia ine 


Titanosauria indet. 
Sauropoda indet. 
Thyreophora 
Ankylosauria 
?Nodosauridae indet. 
Ankylosauridae indet. 
Ornithopoda 
Euornithopoda 
leuanodontia 
Talenkauen Novas, Cambiaso & Ambrosio 
Hadrosauridae 
Hadrosaurinae indet. 


Lambeosaurinae indet 


Euornithopoda inde 
AFRICA 
Theropoda 
Ceratosauria 
g lisauroidea 
Abelisauridae 
Genus indet. 
\vetheropoda incertae sedis 


Bahariasaurus Stromer 


Table 5. Continued. 


Sauropoda 
Titanosauriformes indet. 
ANTARCTICA 
Theropoda indet. 
Thyreophora 
ikvlosauria 
Nodosauridae indet. 
Ornithopoda 
Euornithopoda 
adrosauridae indet. 


INDIAN SUBCONTINEN 


Theropoda 
Ceratosauria 
Abelisauroidea 
Abelisauridae 
Indosaurus Huene & Matley 
Indosuchus Huene & Matley 
Rajasaurus Wilson, Sereno, Srivastiva, Bhatt. 
Khosla & Sahni 
Genus indet. / 
Genus indet. B 
Noasauridae 
Laevisuchus Huene & Matley 
Ceratosauria indet. 
?Carnosauria indet. 
Compsosuchus Huene & Matley 
Ornithomimidae inde 
Theropoda indet. 
Sauropoda 
Lithostrotia 
Isisaurus Wilson & Upchurch 
Jainosaurus Hunt, Lockley, Lucas € Meyer 
Antarctosaurus 
Titanosauria indet. 
Sauropoda indet. 
Thyreophora 
\nkylosauria 
Ankylosauridae indet. 
Stegosauria 


Stegosauridae indet. 


This also may account for the lack of discovery of 
small vertebrate taxa (e.g., lizards, turtles, crocodyli- 
forms, mammals), although climate may have also 


played an important role. Interestingly, the dinosaur 


fossils that have been recovered are of taxa (e.g.. 
Euornithopoda, Hadrosauridae, Nodosauridae) that 


are not represented in the Campanian/Maastrichtian 
of 


Instead they 


rica, Madagascar, or the Indian subcontinent. 
are shared with penecontemporaneous 


horizons in South America. 


AFRICA 
Africa is, by far, the largest Gondwanan landmass, 


roughly 70% larger than the next largest, South 


Volume 93, Number 2 
200 


Krause et al. 191 
Late Cretaceous Terrestrial Vertebrates 


Campanian and 
Africa, 


Antarctica, and the Indian subcontinent. Lists compiled 
. (200 


ble 6. Mammalian genera from 
S 


a 
Maastrichtian horizons in South America, 
from Kielan-Jaworowska et ¢ 004), with a from 
Rana and Wilson (2003) and Khosla et al. (2004) for the 
Indian subcontinent. Question marks indicate questionable 


occurrences and quotation marks indicate paraphyletic taxa. 


SOUTH AMERICA 
?Docodonta 
Reigitheriidae 
Reigitherium Bonaparte 
Eutricondonta 
Austrotriconodontidae 
strotriconodon Bonaparte 
ie re ulata incertae sedis 
Ger 
Archaic Piles P E 


Bondesiidae 


Bondesius Bonaparte 
Stem Cladotheria 
Dryolestidae 
Groebertherium Bonaparte 
Leonardus Bonaparte 
;enus indet 
Mesungulatidae 
Mesungulatum Bonaparte & Soria 
Brandoniidae 
Brandonia Bonaparte 
?Casamiguelia Bonaparte 
Marsupialia 
?Peradectidae 
Genus indet. 
?"Pediomyidae" 
Genus indet. 
Eutheria 
Family incertae sedis 
Perutherium Grambast, Martinez, Mattauer & Thaler 
Genus indet. 
Gondwanatheria 
Sudamericidae 
Gondwanatherium Bonaparte 
Ferugliotheriidae 
Ferugliotherium Bonaparte 
AFRICA? 
No record 
ANTARCTICA 
No record 
INDIAN SUBCONTINENT 
Euther 
TN 


Genus indet 


y incertae sedis 
es Prasad & Sahni 
Sahnitherium Rana & Wilson 


Fami 


eccanolest 


Infraclass incertae sedis 
Gondwanatheria 
Sudamericidae 


Genus indet. 


America. Despite its vast superiority in size among the 
five Gondwanan landmasses considered here, Africa's 
diversity of Campanian/Maastrichtian terrestrial ver- 
tebrates is the second poorest, little better than that 


from Antarctica. This is presumably due in large part 


to the virtual absence of suitable sedimentary 
depocenters of the right age and environment, but 


further exacerbated by limited exploration and a lack 
of definitive age control independent of the vertebrate 
fossils themselves. 

Of the seven identified families of crocodyliforms, 
non-avian dinosaurs, and mammals represented i 
Madagascar, only one, Abelisauridae, is represented, 
albeit questionably, in the Campanian/Maastrichtian 
of A 


crocodyliform clade that dominates in. penecontem- 


Metasuchia, the 


— 


rica. Among crocodyliforms, 


oraneous horizons South America, Madagascar, 


— 


and the Indian subcontinent, is absent. Instead, only 


the neosuchian dyrosaurid Sokotosuchus is known. 
Among non-avian dinosaurs, there is only poorly 


preserved. material, none of which is precisely 


identified; significantly, however, no ornithischians 
are currently known. Campanian/Maastrichtian mam- 
mals have yet to be discovered from mainland Africa. 
although the possible gondwanatherian recently de- 
scribed by Krause et al. (2003b) could conceivably be 


from this horizon. 


THE INDIAN SUBCONTINENT 


The record of terrestrial vertebrates from the latest 
Cretaceous of the Indian subcontinent has increased 
dramatically over the last three decades, largely the 
result of efforts by Ashok Sahni of Panjab University 
(Chandigarh) and his students. The assemblage. from 
below and interbedded within the Deccan Traps (the 
infra- and intertrappean beds) was recently summa- 
rized by Khosla and Sahni (2003). 


considered to be of late 


The infra- and 
intertrappean beds are 
Maastrichtian age. The diversity of terrestrial verte- 
brates from the Indian subcontinent is low, and 


surprisingly low in one clade that is diverse 


elsewhere, the Crocodyliformes. Two metasuchians 
(Baurusuchidae and Metasuchia incertae sedis) and 
one neosuchian (Dyrosauridae) are present, but none 
of the metasuchian families identified in Madagascar 
Trematochampsidae) 


(Notosuchidae, Peirosauridae, 


have been definitively identified on the Indian 
«— 

A possible . mammal was recently 
described by Krause et al. (2003b), but the age of the 


orizon from which it was 1 is of uncertain. age 


within the Cretaceous period. 


192 


Annals of the 
Missouri Botanical Garden 


subcontinent. Non-avian dinosaurs were discovered 
on the Indian subcontinent over 175 years ago. Many 
specimens have been discovered since and demon- 
strate that a diverse fauna existed at this time. 
Unfortunately, however, few articulated specimens 
have been recovered and, as a result, the alpha 
taxonomy remains ambiguous. Nonetheless, it is clear 
that the most common dinosaurs present at this time in 
Madagascar (abelisaurid and noasaurid theropods and 
lithostrotian sauropods) were also present on the 
Indian subcontinent. The mammalian fauna is repre- 
sented by fragmentary isolated teeth that have been 
assigned to at least three eutherian taxa (Deccano- 
lestes, Sahnitherium, and a possible otlestid) and 
& Sahni, 1988: 
1994; 


tana & 


a sudamericid gondwanathere (Prasac 
Godinot & Prasad, 1994: Prasad & Godinot. 
Prasad et al.. 1994: Krause et 1997b: 
Wilson, 2003: Khosla et al., 2004). 


^ 


SUMMARY 


It is clear. based on currently available samples. 


that the greatest similarity in taxonomic composition 


7 


aumpanian/Maastrichtian terrestrial verte- 
South 


of known 


brate faunas to that of Madagascar occurs ii 


America and India. 


PRE-CAMPANIAN CRETACEOUS DISTRIBUTION OF 
CROCODYLIFORMS, NON-AVIAN DINOSAURS. 
AND MAMMALS ON GONDWANAN LANDMASSES 


In addition to comparing the taxonomic composition 
of crocodyliforms, non-avian dinosaurs, and mammals 
from Campanian/Maastrichtian horizons of Madagas- 
car, South America, Africa, Antarctica, and the Indian 
is. relevant to record the 


subcontinent, it is pre- 


— 


ampanjan Cretaceous distributions of these same 


order to 


clades on these same landmasses ii 


potentially reveal deeper histories. As such, Ta- 
bles 7-9 provide an overview of the generic diversity 
of crocodyliforms, non-avian dinosaurs, and mammals 
from the pre-Campanian Late Cretaceous. whereas 
Tables 10-12 provide the same information for the 
Early Cretaceous. The pre-Campanian Late Creta- 
ceous and Early Cretaceous distributions are listed 
because there is general agreement 
that 


South America and Africa separated near the Early 


individually 


| ] 
parcogecosgrapuers 


among and paleontologists 
o D 


Late Cretaceous boundary (see below). Each of the 
clades will be considered in turn. 
CROCODYLIFORMS (TABLES 7. 10) 


Notosuchids, but not trematochampsids. are known 


from the pre-Campanian Late Cretaceous of South 


Table 7. 


Cretaceous 


Crocodyliform genera from pre-Campanian 


South America. Africa, 


horizons ii 


ate 


Antarctica, and the Indian subcontinent. List. for 


America compiled primarily from Bertini and Carvalho 
(1999), and for Africa from Buffetaut (1982) and Larsson and 
Gado (2000). 


SOUTH AMERICA 
Mesoeucrocodylia 
Metasuchia 
Notosuchidae 
Votosuchus Woodward 
Mariliasuchus Carvalho & Bertini 
Comahuesuchus Bonaparte 
AFRICA 
Mesoeucrocodylia 
Metasuchia 
Libycosuchidae 
Libycosuchus Stromer 
Trematochampsidae 
Hamadasuchus Buffetaut 
Trematochampsa 
Eusuchia 
Stomatosuchidae 
Stomatosuchus Stromer 
leevptosuchus Stromer 
ANTARCTICA 
No record 
INDIAN SUBCONTINENI 
No record 


America, whereas trematochampsids, but not notosu- 
chids. are known from the pre-Campanian Late 


Cretaceous of Africa. Both notosuchids and tremato- 


champsids, as well as the unplaced metasuchian 
Araripesuchus, are recorded from Early Cretaceous 
both South 


reveals a deeper history on those landmasses. pre- 


horizons in America and Africa. This 
sumably prior to separation of these landmasses near 
the Early/Late Cretaceous boundary. Peirosaurids are 
Africa but not 


known from the Early Cretaceous o 
South America and from neither landmass during the 
pre-Campanian Late Cretaceous. Unfortunately, the 
record of identifiable pre-Campanian Cretaceous 
crocodyliforms is nonexistent for Antarctica and the 


Indian subcontinent, 
NON-AVIAN DINOSAURS (TABLES 8, 11) 
Pre-Campanian non-avian dinosaur faunas from 


Africa South 


America, are generally much better characterized 


Gondwana, particularly from and 


than their Campanian/Maastrichtian counterparts, i 
part owing to the higher incidence of basin formation 
(with concurrent sedimentation) during these times. 


Abelisaurids and lithostrotians are known from Africa 


and South America during both pre-Campanian Late 


Volume 93, Number 2 
2006 


Krause et al. 


Late Cretaceous Terrestrial Vertebrates 


193 


Table 8. Non-avian dinosaur genera from pre- 
Campanian Late Cretaceous horizons in South America, 
Africa, Antarctica, and the Indian subcontinent. Lists 
compiled from Weishampel et al. (2004), with additions 
rom González Riga (2003) and Apesteguía (2004) for South 
America and Sereno et al. (2004) for Africa. Question marks 
indicate tentative placement. 


SOUTH AMERICA 
Theropoda 
Ceratosauria 
Xenotarsosaurus Martinez, Gimenez, Rodriguez & 
Bochatey 
Velocisaurus Bonaparte 
Abelisauroidea 
Abelisauridae 
Abelisaurus Bonaparte & Novas 
Aucasaurus Coria, Chiappe & Dingus 
Ilokelesia Coria & Salgado 
Abelisauridae indet. 
Tetanurae 
?Coelurosauria 
Aniksosaurus Martinez & Novas 
Avetheropoda 
Allosauroidea 
Giganotosaurus Coria & Salgado 
Dromaeosauridae 
enlagia Novas 
Megaraptor Novas 
Dromaeosauridae indet. 
Troodontidae indet. 
?Ornithomimidae indet 
Tetanurae indet 
Theropoda indet. 
Sauropoda 
Diplodocoidea 
Amazonsaurus Carvalho, Avilla & Salgado 
Rayososaurus Bonaparte 
Titanosauria 
Andesaurus Calvo & Bonaparte 


Argentinosaurus Bonaparte & Coria 


— 


rgyrosaurus Lydekker 
Bonitasaura Apesteguia 
Epacthosaurus Powell 
Mendozasaurus González Riga 
Titanosauria indet. 
Lithostrotia 
Antarctosaurus 
Laplatasaurus 
Rinconsaurus Calvo & González Riga 
Lithostrotia indet 
Saltasauridae 
Neuquensaurus Powell 
Saltasaurus 
Sauropoda indet. 
Ornithopoda 
Euornithopoda 
Notohypsilophodon Martinez 
Iguanodontia 
Inabisetia Coria & Calvo 


Gasparinisaura Coria & Salgado 


Table 8. Continued. 


Hadrosauridae 


Secernosaurus Brett-Surman 


?lguanodontia indet. 


Ornithopoda indet. 


AFRICA 
Theropoda 


Ceratosauria 


> 


;elisauroidea 


Abelisauridae 
Rugops Sereno, Wilson € Conrad 


Abelisauridae indet. 


Noasauridae 


Jeltadromeus Sereno, Dutheil, 


Lyon, Magwene, 


Tetanurae 


Sigilmassasaurus Russe 


Spinosauroidea 


Spinosauridae 


l 


Spinosaurus Stromer 


Avetheropoda 


Carcharodontosauridae 


Carch arodontosaurus Stromer 


Avetheropoda incertae sedis 


Bahariasaurus 


Dromaeosauridae indet. 


Theropoda indet. 


Sauropoda 


Diplodocoidea 


Dicraeosauridae 


cf. Dicraeosaurus Janensch 


Titanosauria incertae sedis 


Aegyptosaurus Stromer 


Lithostrotia 


P um Smith, 


Smith, 


le, Giegengack, Attia 


Dicraeosauridae indet. 


L ithostrotia indet. 


Sauropoda indet. 


Ornithopoda 
Euornithopoda 
Iguanodontia 


Hac 


[om 


rosauridae 


. Ouranosaurus Taquet* 


TN M indet.* 


Iguanodontia indet.* 


Ornithopoda indet. 


ANTARCTICA 
Theropoda indet. 
INDIAN SUBCONTINENT 
Theropoda 
Theropoda indet. 


Sauropoda 


larochene. Larsson, 
Sidor, Varricchio & Wilson 


Lamanna, Lacovara, Dodson, 


Sauropoda indet. (= Bruhathkayosaurus Yadagiri & 
Ay 


'yasami) 
Stegosauria 


Stegosauridae 


Dravidosaurus Y ac 


A 


agiri & Ayyasami 


* [ndicates taxa listed from only Marsa 


(? Albian—Cenomanian) 


Matruh, 


Egypt 


194 


Annals of the 
Missouri Botanical Garden 


Table 9. 


Cretaceous horizons in South America, Afric: 


Mammalian genera from pre- 1 ate 


Antarctica, 


and the Indian subcontinent. ed FEMA Kielan- 


Jaworowska et al. (2004). 


Lists compl 


SOUTH ÁMERICA 
Mammalia indet. 
AFRICA 

Mammalia indet. 
ANTARCTICA 

No record 

INDIAN SUBCONTINEN 


No record 


Table | 


horizons in South 


).  Crocodyliform genera from Early Cretaceous 
Arica, 
South 
primarily from Bertini and Carvalho (1999), and for Africa 
Buffetaut (1982) and (2000). 


Question mark 11 9155 tentative placement. 


America, Antarctica, and the 


Indian subcontinent. List. for America compiled 


[rom Larsson and Gado 


SOUTH AMERICA 
Mesoeucrocodylia 
Metasuchia 
Notosuchidae 
Candidodon Carvalho & Campos 
T Pal N 


s 
Amargasuchus Chiappe 
?Trematochampsidae 

Caririsuchus Kellner 
Metasuchia incertae sedis 
Araripesuchus 
Neosuchia 
Pholidosauridae 
Sarcosuchus Marsh 
Meridiosaurus Mones 
AFRICA 
Mesoeucrocodylia 
Metasuchia 
Notosuchidae 
Malawisuchus Jacobs. Winkler. 


Anatosuchus Sereno, Sidor, 


Downs «€ Gomani 
Larsson & Gado 
Peirosauridae 
Stolokrosuchus Larsson & Gado 
Trematochampsidae 
Hamadasuchus 
Libycosuchidae 
Libycosuchus 
Metasuchia incertae sedis 
Araripesuchus 
Neosuchia 
Pholidosauridae 
Sarcosuchus 
Neosuchia incertae sedis 


Brunet & Hell 


Brillanceausuchus Michard, de Broin. 


INDIAN SUBCONTINENT 
No record 


Cretaceous and Early Cretaceous intervals, thereby 


indicating the initial diversification of these clades 
prior to the breakup of Gondwana. Noasaurids are also 
same intervals, bul 


known from Africa during these 


their occurrence in South America is limited to the 


pre-Campanian Late Cretaceous. The phylogenetic 


ambiguity resulting from the relatively fragmentary 


skeletal material of many of these taxa, particularly 
noasaurids and lithostrotians, however, necessarily 
limits biogeographic inferences derived from them. 
Similar to the situation described above for 
Campanian/Maastrichtian strata of non-Madagascan 
Gondwanan landmasses, pre-Campanian Cretaceous 


deposits, particularly of South America and Africa, 


also preserve diverse non-avian dinosaur aunas 


characterized by numerous clades of tetanuran 
theropods, non-lithostrotian sauropods, ornithopods, 
ankylosaurians, and stegosaurians. Particularly prob- 
lematic for Gondwanan-wide biogeographic recon- 
structions is the virtual absence of Early Cretaceous 
dinosaur discoveries. from Antarctica. India. and 


Madagascar. 


MAMMALS (TABLES 9, 12) 


None of the lower-level mammalian taxa recovered 
from the Maastrichtian of Madagascar are shared with 
pre-Campanian Cretaceous horizons from other Gond- 
wanan landmasses, but the biogeographic relevance of 
this information is limited by 


the extremely poor 


knowledge of the mammalian fossil record for the 
entire Cretaceous of Gondwana. With the possible 
exceptions of a caudal vertebra from Libya (Nessov et 
al., 1998) and a dentary fragment from Brazil (Bertini 
et al., 1993), both of which were recovered from poorly 


age-constrained horizons — (Santonian-Campanian), 


there are no known pre-Campanian Late Cretaceous 


mammals known from Gondwanan landmasses. In 


addition to an important assemblage from Australia, 
Early Cretaceous Gondwanan mammals are known 
only from Africa and South America. A Barremian site 


1 Cameroon has yielded evidence of at least three 


nontribosphenic therians, only one of which. the 


"eupantotherian"  Abelodon. has been 


1990). By 


Berriasian 


peramurid 


named (Brunet et al.. contrast. a diverse 


mammalian fauna of age has been re- 
covered from Morocco and includes eutriconodontans. 
archaic "eupantotherians;" and 
1998). A 
late Hauterivian or early Barremian site in Argentina 
skeletal 


Hopson & Rougier. 


“symmetrodontans.” 
“tribotherians” (Sigogneau-Russell et al., 
remains of the 
1993). 


but this is the only South American mammal site of 


has yielded numerous 


zatherian Vincelestes (e.g.. 


definitive Early Cretaceous age. 


Volume 93, Number 2 
2006 


Krause et al. 
Late Cretaceous Terrestrial Vertebrates 


195 


genera from Early 


Antarctica, 


Table 11. Non-avian dinosaur 
Cretaceous horizons in South America, Africa, 


and the Indian subcontinent. Lists compiled from 
Weishampel et al. (2004) with additions from Leanza et a 


(2004) for South America and Sereno et al. (2004) for Africa. 


SOUTH AMERICA 
Theropoda 
Ceratosauria 
Ligabueino Bonaparte 


Ceratosauria indet. 


Abelisauridae indet. 
Tetanurae 
Spinosauroidea 
Spinosauridae 
Irritator Martill, Cruickshank, Frey, Small & 
larke 
Ingaturama Kellner & Campos 
Avetheropoda 
Allosauroidea 
Carcharodontosauridae indet. 


o 
85 


Tyrannosauroi 
Santanaraptor Kellner 
Compsognathidae indet 
Oviraptorosauria indelt. 
Theropoda indet. 
Sauropoda 
Diplodocoidea 
Imargasaurus Salgado & Bonaparte 
Rayososaurus 
Rebbachisauridae indet 
Titanosauria 
Agustinia Bonaparte 
Chubutisaurus del Corro 
Titanosauria indet. 
Thyreophora 
Stegosauridae indet. 
Ornithopoda 


Euornithopoda 


guanodontia indet. 
Ornithischia indet. 
AFRICA 
Theropoda 
Ceratosauria 
on Sereno, Wilson & Conrad 
\belisauro 
5 rui indet. 
Abe 
Tetanurae 
Afrovenator Sereno, Wilson, Larsson, Dutheil & 


isauridae indet. 


Sues 
Spinosauroidea 
Spinosauridae 
Spinosaurus 
Suchomimus Sereno, Beck, Dutheil, Gado. I 
Lyon, Marcot, Rauhut, Sadlier, Sidor, 
Varricchio, Wilson & Wilson 
Spinosauridae indet. 
Avetheropoda 
Carcharodontosauridae 


Table 11. Continued. 


arsson, 


Carcharodontosaurus Stromer 
Avetheropoda incertae sedis 


ariasurus Stromer 


e 
= 
2 


Coelurosauria incertae sedis 
Nqwebasaurus de Klerk, Forster, Sampson, 
Chinsamy & Ross 

Tetanurae indet. 

Theropoda indet. 


Sauropoda 


Diplodocoidea 
Rebbachisauridae 

Nigersaurus Sereno, Beck, Dutheil, Larsson, Lyon. 

Sidor, Varricchio, Wilson & 


Moussa. Sadlier, 
Wilson 
Rebbachisaurus Lavocal 
Diplodocidae indet. 
Macronaria 
Jobaria Sereno, Beck, Dutheil, Larsson, Lyon, 
Moussa. 


Varricchio; Wilson & Wilson 


Sadlier, Sidor, 


E 


Titanosauriformes 
3rachiosaurus Riggs 
Lithostrotia 
Talawisaurus 
Lithostrotia indet. 


Sauropoda indet. 


Thyreophora 


\nkylosauria 
Nodosauridae indelt. 
Stegosauria 
Paranthodon Nopcsa 
Thyreophora indet. 
Ornithopoda 
Euornithopoda 
Iguanodontia 
Lurdusaurus Taquet & Russell 
Valdosaurus Galton 
Hadrosauridae 
Ouranosaurus 
lguanodontia indet. 
Ornithischia indet 
ANTARCTICA 
No record 
INDIAN SUBCONTINENT 


No record 


SUMMARY 


Current evidence suggests a number of close 
biogeographic ties linking pre-Campanian Cretaceous 
aunas from Africa and South America. However, it 


must be noted that data to evaluate faunal links with 


other Gondwanan landmasses, including Madagascar, 
are minimal. The undescribed faunal assemblage from 
the Ankazomihaboka sandstones (Curry, 1997; Curry 
el al., in prep) may have a significant bearing in this 


regard. 


196 


Annals of the 
Missouri Botanical Garden 


Table 12. Mammalian genera. from Early Cretaceous 


horizons in South America, Africa, Antarctica, and the 
Indian subcontinent. List compiled from Kielan-Jaworowska 
(2004). with additions from Hahn and Hahn (2003). 


Question mark indicates tentative placement and quotation 


el al. 


marks indicate paraphyletie taxa. 


SOUTH AMERICA 
“E upantothe ria“ 
Zatheria 
Vincelestida 
Vincelestes Bonaparte 
VERICA 
Mammalia 


Family incertae sedis 


a 


enus indet. A 
Genus indet. B 
Eutriconodonta 
Amphilestidae” 
Genus indelt. 
Family indet. 
Dyskritodon Sigogneau-Russell 
Ichthyoconodon Sigogneau-Russell 
Multituberculata 
Hahnodontida 
Denisodon Hahn & Hahn 
Hahnodon Sigogneau-Russell 
Genus indet. 
Archaic 


“amily indet. 


“symmetrodontans” 


Atlasodon Sigogneau-Russell 
licroderson Sigogneau-Russell 
Thereuodontidae 


Thereuodon Sigogneau-Russell 


Stem Cladotheria (eupantotherians’ 


) 


Family indet. 


ae 


senus mdet. 


Afriquiamus Sigogneau-Russell 
nimus Sigogneau-Russell 
Donodontidac 
Donodon Sigogneau-Russell 
Peramuridae 
Abelodon Brunet, Coppens, Dejax, Flynn, Heintz, Hell. 
Jacobs, Jehenne, Mouchelin, Pilbeam & Sudre 
PN i Owen 
Stem Boreosphenida 
Aegialodontidae 
ypomylos Sigogneau-Russell 
Family indet. 
Tribotherium Sigogneau-Russell 
ANTARCTICA 
No record 
INDIAN SUBCONTINENT 
No record 


PHYLOGENY AND BIOGEOGRAPHY OF LATE CRETACEOUS 
CROCODYLIFORMS, NON-AVIAN DINOSAURS. 


AND MAMMALS FROM MADAGASCAR 


— 
— 
— 


the seven family-level taxa of. erocodyliforms, 


non-avian dinosaurs, and mammals known from the 


Maastrichtian of shared with 
C . /M 1 


Madagascar, all are 


htian faunas of South America. 
This is suggestive of close biogeographic ties and, 


indeed, a degree of cosmopolitanism that is difficult to 


explain in the context of most recent paleogeographic 
reconstructions depicting separation of most Gondwa- 
nan landmasses by great distances at this time (e.g.. 
1992; Smith et al., 1994; 
Tikku. 2001: 


Lawver et al.. Reeves & de 
Wit, 2000; Marks € 
2001; Scotese, 2001; O'Neill et al., 
2003: Bernard et al., 2005). 


reconstructions, Madagascar had long been isolated in 


Rotstein eb a 
2003: 
According to 
the Indian more 
South 


America, by whatever route, had been severed some 


Ocean by the Campanian and, 


specifically, any terrestrial. continuity with 


50 million years earlier. However, the record of Late 


Cretaceous. terrestrial. vertebrates from Gondwana is 
spotty at best and therefore difficult to interpret in 


biogeographie context. The best possibilities for 


— 


obtaining a biogeographic signal come from erocodyli- 


forms. non-avian dinosaurs, and mammals, in part 


because of their relatively good preservation. and 


therefore relatively precise identification, and in parl 
the 


recent. publication of two monumental works, by 


because they are relatively well-studied. Indeed, 
Weishampel et al. (2004) on dinosaurs and. Kielan- 
Jaworowska et al. (2004) on mammals, facilitates the 
compilation of occurrence data that are used herein to 
reveal distributional patterns for these taxa. Here we 
document the available taxonomic and phylogenetic 
information that might have a more immediate bearing 


on our understanding of Gondwanan biogeography. 


CROCODYLIFORMS 


Phylogenetic analyses have been presented for only 
three of the seven species of crocodyliforms known from 
the Late Cretaceous of Madagascar. The phylogenetic 


relationship of Mahajangasuchus insignis to other 


metasuchians was examined by Buckley and Brochu 
(1999) and Buckley (2000). 


indicated a clade that consisted of Trematochampsa + 


| al. 


These analyses 


Mahajangasuchus + Peirosauridae, with Araripesuchus 
basal to this clade. This result tends to support the 
contention of Buffetaut (1988, 1989), who argued that 
Peirosauridae should be considered a junior synonym of 
Trematochampsidae. Without more conclusive evidence 
and a more thorough understanding of trematochampsid 


laxa, Mahajangasuchus was classified as Metasuchia 


Volume 93, Number 2 
2006 


Krause et al. 197 


Late Cretaceous Terrestrial Vertebrates 


incertae. sedis. Tykoski et al. (2002) and Turner and 
Calvo (2005) obtained identical results to Buckley and 
Brochu (1999). Carvalho et al. (2004) did not inclu 


any traditional trematochampsids in their analysis, but 


— 


e 


found Mahajangasuchus to be embedded within the 
Peirosauridae, with Uberabasuchus Carvalho, Ribeiro & 


os Santos Avilla from the Late Cretaceous of Brazil as 


— 
a 


its most closely related sister taxon. 
Buckley et al. (2000) determined that Simosuchus 
and its closest sister taxon, Uruguaysuchus, from the 
Late Cretaceous of Uruguay, formed a clade with 
Malawisuchus, and that these three taxa were question- 
able members of the Notosuchidae (which included 
This 
relationship was supported by several 1 
(2002) recovered 


with Simosuchus and eee. 


Notosuchus + Libycosuchus in their analysis). 


analyses. Tykoski et al. | nearly 


identical tree, 
linked as closely related sister taxa within a notosu- 
chian clade. Sereno et al. (2003) placed Simosuchus 
within the Notosuchia (including a Comahuesuchus + 
Anatosuchus clade and a Simosuchus + Araripesuchus + 
traditional sebecosuchians + Malawisuchus + Notosu- 
chus clade). Tree topology regarding this clade is nearly 
identical with the earlier study by Buckley et al. (2000), 


with the exception of the exclusion of Araripesuchus 


from the Notosuchia. Sereno et al. (2003) additionally 


recognized an expanded concept of Notosuchia, 
roughly equivalent to the Ziphosuchia proposed by 
(2000). 


Anatosuchus, and Baurusuchus + Sebecus Simpson. Pol 


Ortega et al. by including Comahuesuchus, 
(2003) recovered a tree similar to those of Buckley el 
al. (2000) and Sereno et al. (2003), with Simosuchus 
The constituency of 


firmly nested within Notosuchia. 


his notosuchian clade strongly reflects that proposed by 
Ortega et al. (2000) and Sereno et al. (2003), with the 
only key difference being the exclusion of Araripesu- 
chus, as was also proposed by Buckley et al. (2000). It is 
worth noting that Ortega et al. (2000) and Pol (2003) 
not included in the other two 


include several taxa 


analyses, resulting in a more geographically wide- 
spread notosuchian (or ziphosuchian) clade. These taxa 
include Chimaerasuchus Wu, Sues & Sun from the 
Early Cretaceous of China and /berosuchus Antunes 
X05) 


essentially mirrored the earlier results of Buckley et al. 


from the Paleogene of Europe. Turner and Calvo (2t 


(2000) in producing a Malawisuchus + Uruguaysuchus 
+ Simosuchus clade within the Notosuchia. 


` 


The most contrary hypothesis regarding the r 
lationship of Simosuchus to other mesoeucrocodylians 
was put forth by Carvalho et al. (2004). They placed 
the genus, along with Chimaerasuchus, within the 
Chimaerasuchidae, a clade basal to their Notosuchi- 
morpha. Other differences also appear in their results, 
including the exclusion of Uruguaysuchus from the 


Notosuchia and the placement of Malawisuchus within 


the Peirosauroidea. It is likely that these discrepan- 
cies are due to the authors’ reweighting of characters 
based upon rescaled values, a practice not followed by 
other studies. 

A recent biogeographic study by Turner (2004b) is 
highly 


croc sodyliform distribution. Turner (2004b), employing 


relevant to the question of Cretaceous 
a time-slicing protocol adapted from Upchurch et al. 
(2002), conducted a cladistie biogeographic analysis 
of a diverse sample of Cretaceous crocodyliform taxa, 


mostly from Gondwana. Included in his sample were 


three taxa known from the Late Cretaceous. of 
Madagascar: Simosuchus clarki, Mahajangasuchus 


insignis, and Araripesuchus sp. indet. The sister taxon 
Late 


American form Uruguaysuchus, 


of Simosuchus in Turners analysis is the 
South 


thus supporting an earlier assessment by Buckley el 


Cretaceous 


al. (2000). Turner resolved a sister taxon relationship 
between Mahajangasuchus and the South American 
Peirosauridae, which also supports the earlier analy- 
ses of Buckley and Brochu (1999) and Buckley et al. 
(2000). Araripesuchus sp. indet. occurs at the base of 
a clade of other Araripesuchus species from both 
Africa and South America. In his analysis, Turner also 
included Pabwehshi, a ?Maastrichtian mesoeucroco- 
the Indian subcontinent recently de- 
Wilson et al. (2001). Turner (2004b) 


confirmed inclusion of Pabwehshi in the Baurusuchi- 


dylian from 


e 


scribed by 


dae, otherwise only known from the Late Cretaceous of 
3razil and Argentina. Late Cretaceous crocodyliforms 
from the Indian subcontinent are poorly known, based 
(Prasad € de 


However, 


on isolated teeth 
2002). 


available information, Malagasy and Indian crocodyli- 


almost exclusively 


Lapparent de Broin, based on 


forms share closest affinities with roughly contempo- 
raneous taxa from South America. 


Turner (2004b) refined the methods of Upchurch et 


al. (2002) and revealed two continent-level vicariant 


events: (1) separation of Africa, South America, and 


Indo-Madagascar from other non-Gondwanan land- 
masses earlier in the Cretaceous, and (2) separation of 


\ 


in the Cretaceous. Turner (2004b: 


rica from South America and Indo-Madagascar later 
2007) states that 
“this later event depicts a rather non-traditional 
biogeographic relationship and, in that respect, this 
results are similar to 
conclusions of Sampson et al. (1998), : 
3rochu (1999), Krause et al. (1999), Pap : 2000); 


and the geological data of Hay et al. (1999 


study's 


NON-AVIAN DINOSAL RS 


It could be argued that non-avian dinosaurs are 


perhaps better suited than crocodyliforms for re- 


— 


vealing biogeographic pattern as it relates to subaeria 


198 


Annals of the 
Missouri Botanical Garden 


landmasses, because they were likely more closely 


tied to the terrestrial realm. Moreover, most taxa were 
very large. making it less likely that they would have 
been able to raft across great. distances on floating 
has been documented. for 


mats of vegetation, as 


smaller-bodied extant 


1998). As for 


avian dinosaur | 


vertebrates (Censky et al., 

crocodyliforms, the majority of non- 
taxa known from the Late Cretaceous 
of Madagascar share closest affinities with penecon- 


axa from South America and India. 


lemporaneous 
Preliminary phylogenetic analyses have been pre- 
sented for all four of the saurischian taxa known from 
2001; Curry 
hogers, 2002. 


Madagascar (Sampson et al., 1998 
2001, 2004; 
2005; Carrano et al., 2002). 


Majungatholus, an abelisaurid, has been included 


Rogers & Forster, Curry 


in several recent phylogenetic analyses. Sampson et al. 
(1998) placed this Malagasy abelisaurid in a polytomy 
with /ndosaurus and Indosuchus from the Maastrichtian 
of India, Abelisaurus from the Santonian of Argentina, 
and Carnotaurus from the Campanian-Maastrichtian of 
Argentina. Relationships within this polytomy were 
somewhat better resolved by Sampson et al. (2001) and 
Carrano et al. (2002), who posited that Majungatholus 


and Carnotaurus were sister taxa and that they. 
together, were the sister taxon of Abelisaurus (In- 


dosaurus and Indosuchus were not included in the 


analyses). These results were in essence supported in 


a 


Rowe 
(2004). Coria et al. (2002) grouped Majungatholus with 


Abelisaurus and Hokelesia in a polytomy that com- 


a more comprehensive analysis by Tykoski anc 


prised the sister taxon of Aucasaurus + Carnotaurus. 
Wilson et al. (2003) described 


Rajasaurus, from 


new abelisaurid. 


the Maastrichtian of India and 


reported results of another cladistic analysis « 
abelisaurids (later presented by Sereno et al., 2004), 
which concluded that Abelisaurus, Rajasaurus, and 
Majungatholus + Carnotaurus occupied. successive 


nodes on a tree. The later study by Sereno et al. (2004 


also included a newly named Cenomanian abelisaurid 
from Africa, Rugops (referred to as "Niger taxon 27 
2003). 


shuffling positions among abelisaurid genera. all 


Wilson et a Most significantly, despite (s 
analyses have concluded that the closest. known 
relatives of Majungatholus lie in Campanian-Maas- 
trichtian horizons of Argentina and India. 

considered to be 
2004: 


An analysis by 


Masiakasaurus is currently 


a noasaurid abelisauroid (Carrano et al.. Sereno 
el al., 2004; Tykoski & Rowe, 2004). 
Sereno et al. (2004) concluded that 


lies in an unresolved polytomy with Noasaurus from 


Masiakasaurus 


the “late Campanian-Maastrichtian of Argentina. an 
unnamed genus from the Aptian-Albian of Niger, and 
Deltadromeus (originally regarded as a basal coelur- 


osaurian by Sereno et al., 1996) from the Cenomanian 


— 


of Morocco. When originally described by Sampson et 
al. (2001) and Carrano et al. (2002). the relationships 
of Masiakasaurus were also unresolved, although it 
was recognized that its affinities lay most closely with 
Genusaurus Accarie, Beau- 
Michard & Taquet, and /lokelesia, 


and secondarily with the abelisaurids Majungatholus, 


Noasaurus, Laevisuchus, 


doin, Dejax, Fries, 
Carnotaurus, Xenotarsosaurus, and Abelisaurus. Addi- 
tional remains of Masiakasaurus recovered during the 
2003 study by M. 
Carrano, S. Sampson, and M. Loewen. In a preliminary 
(2004: 44A) the 


Noasauridae to be abelisauroids “from the Cretaceous 


field season are currently under 


report, Carrano et al. regard 


of Argentina (Noasaurus, Velocisaurus), India (Laevi- 
suchus). Madagasca ar (Mastakasaurus), and possibly 
Europe and Africa. 

Rapelosaurus has been included in several phylo- 
genetic analyses, the most comprehensive by Curry 
Rogers (2005), but also by Curry Rogers and Forster 
(2001) and Wilson (2002). 
their 


All three of these analy Ses 


agree in resolution of a close relationship 


between Nemegtosaurus Nowinski from the Campa- 
nian-Maastrichtian of Mongolia and Rapetosaurus. but 
postulate different successive oulgroups and close 
relatives. Curry Rogers and Forster (2001) and Curry 
Rogers (2005) identify a clade that includes Malawi- 
Y Malawi and several taxa 


saurus from the Aptian 


from India and/or South America (e.g.. Antarctosaurus. 
leustinia). Wilson (200 


saurus as a more basal lithostrotian, with Rapetosaurus 


2) instead identifies Malawi- 


derived 
Jam & Ban- 
f 


taxa from North 


as the sister taxon to a clade of 


^ 


more 


litanosaurs including /sisaurus colberti 


dyopadhyay (from the Campanian-Maastrichtian 
India) and Saltasauridae (including 


America, Mongolia, and South America). 


r 


Malagasy Taxon B has recently been included 

a more comprehensive analysis of lilanosaur phylog- 
eny (including over 29 purported litanosaurians: 
2001. 2005): 
clearly resolves it as nested within the 
Wilson & 2003). Saltasaurines 
traditionally South 


genera (Saltasaurus, Neuquensaurus, and Roc SOUS) 


Curry. Rogers, the strict consensus tree 
Saltasaurinae 
(sensu Upchurch, 


include only three American 


and are uniquely characterized by the presence of 
strongly. procoelous, dorsoventrally compressed distal 


caudal vertebrae. 
MAMMALS 


Most of the mammalian taxa from the Late 


Cretaceous of Madagascar provide little biogeographic 
information, either. because the specimens are too 
fragmentary to identification to a 


permil ower 


— 


taxonomic level, or because the taxa represented are 


unknown from any other landmass. Such is clearly the 


Volume 93, Number 2 
2006 


Krause et al 199 


Late Cretaceous Terrestrial Vertebrates 


case for one isolated tooth of a large taxon and for 
a nearly complete skeleton (Krause, 20032); both taxa, 
if they are different, cannot be identified as yet 
beyond Mammalia incertae sedis. 

The 


important. The two teeth questionably identified as 


presence of gondwanatheres, however, is 
gondwanatherian are relatively uninformative, but the 
two specimens that have been assigned to Lavanify 
miolaka, a sudamericid, provide important. biogeo- 
graphic data. Sudamericids are elsewhere known from 
the Late Cretaceous and Paleocene of Argentina, the 
Late Cretaceous. (Maastrichtian) of India, and the 
Eocene of Antarctica. Lavanify appears to be most 
closely related to the unnamed sudamericid from 
India (Krause et al., 1997b). 

PALEOGEOGRAPHIC RECONSTRUCTIONS: WERE SOUTH 
AMERICA AND MADAGASCAR CONNECTED IN THE 

LATE CRETACEOUS 


the 


— 


Despite Madagascar's current position ii 
Eastern Hemisphere and South America's location in 
the Western Hemisphere, and despite differing details 
in depictions of Gondwanan paleogeography. it is 
clear that the two landmasses were much closer to one 
another in the Mesozoic than they are today. Prior to 
the break-up of Pangea, a non-obstructed (by oceanic 
waters) overland route across Africa spanning some 
3000 km separated the two areas of interest: today the 
straight-line distance between South America and 
Madagascar is almost 8000 km. With the fragmenta- 
tion of Gondwana, which is generally agreed to have 
commenced in earnest in the Late Triassic to Early 
Jurassic (Lawver et al., 1992; Torsvik et al., 2001: de 
Wit, 2003; Wells, 2003), Madagascar, as part of “East 
Gondwana” (also including the Indian subcontinent. 
Antarctica, and Australia), began to separate from 
“West Gondwana” (South America and Africa). 
Initial rifting between the Indo-Madagascar block 
and Africa began during the Permo-Triassic, and 
seafloor spreading between the conjugate-rifted mar- 
gins of southern Somalia, Kenya, and Tanzania 
(Western Somali Basin) 


commenced by the late Middle Jurassic (Lawver et al., 
1992; Wells, 2003). By the Late Jurassic. (approxi- 
mately 160 Ma), a narrow seaway separated the east 
coast of Africa from Madagascar and the rest of the 


East Gondwana block. During the Late Jurassic, Indo- 


Madagascar shifted southward along the Davie 
Fracture Zone, ultimately coming to rest some 


400 km off the east coast of Mozambique in the Early 
Cretaceous (130-120 Ma). 

At about this same time (mid Early Cretaceous). 
seafloor spreading commenced between the Indo- 


Madagascar block and Antaretica-Australia (Lawver 


and northern Madagascar 


et al., 1992). Most workers (e.g., Lawver et al., 1992; 
Müller et al., 1993: Roeser et al., 1996; Marks & 
Tikku, 2001; Coffin et al., 2002; Kent et al., 2002; 


O'Neill et al., 2003) posit that a through-going seaway 


— 


intervened between Indo-Madagascar and Antarctica- 
Australia by the mid to late Early Cretaceous (130— 
110 Ma) and that subsequent spreading between these 
landmasses proceeded rapidly. Several recent paleo- 
geographic reconstructions, for instance, illustrate 
a separation between Antarctica and Indo-Madagascar 
of approximately 950-1100 km (employing 111 km/ 
degree of latitude conversion) by 96 Ma and 1700— 
1850 km by 83 Ma (Rotstein et al., 2001: fig. 9; 
Bernard et al., 2005: figs. 4, 5; Schettino & Scotese, 
2005: figs. 4, 33). In part to reconcile the developing 
record of vertebrate fossils, Hay et al. (1999) boldly 
proposed an "alternative global Cretaceous. paleoge- 
ography” in which intermittent land connections 
persisted. between Indo-Madagascar and Antarctica, 
across the Kerguelen Plateau, well into the Late 
Cretaceous (perhaps as late as 80 Ma). Case (2002) 


also proposed a persistent. land bridge, although 


arther west, across the Gunnerus Ridge and Kainan 
Maru Seamount. 

Finally, in the mid Late Cretaceous, another major 
rifting event linked to the Marion hotspot led to the 
separation of Madagascar and India. This episode of 
seafloor spreading was accompanied by an outpouring 
of flood basalts on both landmasses. Dates from these 
basalts cluster in age from approximately 91 to 84 Ma 
(Storey et al., 1995, 1997; Torsvik et al., 1998, 2001). 
With this final episode of Late Cretaceous rifting, the 
Indian subcontinent moved. rapidly northeastward 
toward Eurasia (Randrianasolo et al., 1981; Storetvedt 
et al, 1992; 1995, 1997), 


Madagascar, situated several hundred kilometers from 
e 


Storey et al., and 
mainland Africa across the Mozambique Channel, has 
remained isolated in the Indian Ocean ever since. 

“West South 
Africa Early 


Cretaceous, and by approximately 120 Ma an arm of 


Rifting in Gondwana? between 


America and commenced in the 
the South Atlantic extended well northward between 
the two landmasses (Lawver et al., 1992; Müller et al., 
1993; Smith et al., 1994; Scotese, 1998; Hay et al., 
1999). Geophysical data indicate that a through-going 
seaway intervened between South America and Africa 
by the beginning of the Late Cretaceous (Nürnberg & 
Müller, 1991; Lawver et al., 1992, Müller et al., 1993; 
Pletsch et al., 2001). and normal marine communica- 
tions, as evidenced by the distribution of Cretaceous 
echinoids and fishes (Maisey, 2000; Néraudeau & 
Mathey, 2000), were apparently established between 
the Western Tethys and the Southern Ocean domain 
by approximately 100 Ma. The Antarctic. Peninsula 


and South Orkney group remained contiguous with the 


200 


Annals of the 
Missouri Botanical Garden 


southern tip of South America throughout the Late 
‘retaceous and well into the Tertiary (Lawver et al., 
1992: Woodburne € Case, 1996; Hay et al., 1999; 
2002: Lawver & Gahagan, 2003). 


There remains strong indication that Africa was the 


Reguero et al., 


first major landmass to become isolated by circum- 
continental seaways during the breakup of Gondwana, 
with both geophysical and biogeographic data sug- 
gestive of isolation. before the end of the Early 
Cretaceous (approximately 100 Ma). Whether or not 
Indo-Madagascar maintained intermittent connections 
with Antarctica via the Kerguelan Plateau (Hay et al., 
1999) and/or 02) into 


the later remains 


the Gunnerus Ridge (Case, 20 


stages of the Late Cretaceous 


debatable. Faunal data described in this report are, 


however. consistent with a scenario of faunal in- 


terchange among "East Gondwana” landmasses (Ma- 
dagascar-Indian subcontinent-Antarctica) and South 


America until late in the Late Cretaceous (see below). 


GONDWANAN FRAGMENTATION AND. LATE 
CRETACEOUS BIOGEOGRAPHY 


Fragmentation of the Gondwanan supercontinent 
and dispersion of its constituent landmasses during 
the latter half of 


profound effects on resident faunas of terrestrial 


of the Mesozoic Era undoubtedly had 
vertebrates and their subsequent evolutionary histo- 
ries. Combining geophysical and stratigraphic data for 
Gondwanan breakup with observations of faunal 
distributions provides an unprecedented opportunity 
to reveal large-scale biogeographic patterns. However, 
invoking ceteris parabis, there is no reason to assume 
hold priority 


that geologie data, by their very nature, 


over paleontologie data or vice versa—yet all things 


are rarely equal. Unfortunately, for the Cretaceous of 


Gondwana, seldom are both the geologie and 


paleontologie data sufficient for specified temporal 


slices and geographic areas. For some times and 


places, the geologic data supporting paleogeographic 
reconstructions are sound, plentiful, and derived from 
and resulting 


independent sources, interpretations 


have been made with confidence. For other times and 
places, the fossils are numerous, well preserved, and 
have been incorporated. into. rigorous phylogenetic 
frameworks. In this regard, the developing Cretaceous 
record of terrestrial vertebrates from Gondwanan 
landmasses is beginning to provide an opportunity to 
further clarify the timing and sequence of Gondwanan 
fragmentation. 

Tables 4-12 reveal a number of notable patterns in 
the distribution of Cretaceous. Gondwanan terrestrial 
vertebrates. The same family-level taxa of crocodyli- 


forms. non-avian dinosaurs, and mammals thal were 


shared among Madagascar. the Indian. subcontinent, 


and South America during Campanian/Maastrichtian 


times are nob. known from penecontemporaneous 
horizons in mainland Africa or. for that matter. 
Antarctica. Nonetheless, despite the intensity of 


recent exploration and collecting, the fossil record 


not fully up to the task of testing whether or not this 


real, a point made repeatedly by our 
10070. 
Rogers et al.. Carrano el 
2002: 2003b: 


The geologic stages of the 


pallern is 
working group and others (e.g., Krause et al., 
19099; 1999; 2000: 
2002; Lamanna et al., Krause, 
2000). 


Cretaceous for which the records of terrestrial. fossil 


Forster, 


O'Connor et al., 


vertebrates from Madagascar, the Indian subconti- 
nent, and South America are reasonably well sampled 
are, for the most part, the very stages for which the 
Mrican record is poor, and vice versa. As emphasized 
(1999: 6). 7 


virtual 


by Krause et al. one of the key stumbling 


blocks . . . is the lack of terrestrial and 
freshwater vertebrates from the 
Late Africa? The 


terrestrial vertebrates from mainland 


post-Cenomantan 


Cretaceous. of fossil record of 


Africa is much 


better for the Early Cretaceous and pre-Campanian 


Late Cretaceous than it is for the later stages of the 


Late Cretaceous, whereas the reverse is true for the 


other three landmasses, particularly Madagascar and 


the Indian subcontinent. To further complicate 


malters, the record terrestrial vertebrate fossils 


from the Cretaceous-of Antarctica remains all but 
nonexistent. 

This uneven temporal sampling has inhibited. and 
sull inhibits. definitive testing of two competing 
biogeographic hypotheses (Fig. 8). The first. recently 
hypoth- 


formulated and labeled the “pan-Gondwana” 


esis by Sereno et al. (2004), stipulates that various 
clades of terrestrial MENU were broadly distrib- 


uted throughout Gondwana during the Early. Creta- 
ceous and that the much more limited distributions in 
the post-Cenomanian Late Cretaceous (including in 
Africa) are the 


differential extinction. 


result of poor sampling and/or 


A corollary of this hypothesis 
Africa and 


states that land connections between (1) 


South America, (2) South America and Antarctica. 
and (3) Antarctica and Indo-Madagascar were all 


interval at the 
JO-90 Myr 


Implicit 


severed “during a relatively. brief 
the Late Cretaceous (ca. l 
(Sereno et al.. 2004: 1328 


hypothesis, therefore, is the prediction that terrestrial 


beginning. of 


n this 


ago)” 


vertebrate faunas on each of these major landmasses 
would become increasingly endemic during the Late 
Cretaceous, beginning at approximately 90 Ma. The 
second hypothesis. formulated by our working group 
(e.g. Krause et al., 1997b, 1999: 1 al; 
1998, 2001) and recently dubbed the "Africa-first^ 


hypothesis by Sereno et al. (2004), posits cosmopol- 


Sampson 


itanism of Late Cretaceous terrestrial Gondwanan 


Volume 93, Number 2 Krause et al. 201 
Late Cretaceous Terrestrial Vertebrates 


PAN-GONDWANA HYPOTHESIS AFRICA-FIRST HYPOTHESIS 
120 Ma 120 Ma 


[er m] RH 
Africa Africa 
South South 
America America 
Indo- 
Madagascar 


Antarctica 


Antarctica 
100 Ma | 100 Ma 


Africa Africa 
South South 
America America 
Indo- 


Antarctica Antarctica 
80 Ma 80 Ma 
R Se 
Madagascar Madagascar 
Africa Africa 
South South 
America America 
Antarctica Antarctica 
60 Ma 60 Ma 
Madagascar Madagascar 
Africa Africa 
South South 
America America 


2 


Antarctica | Antarctica 


Figure 8. Schematic paleogeographic maps of major Gondwanan landmasses exclusive of Austr: alia and southern Europe 
at 120, 100, 80, and 60 Ma de Pic ting the major differences in purported land connections between the pan- -Gondwana 


hypothesis (left column) and the Africa-first hypothesis (right column), The pan-Gondwana hypothesis posits se paration ol 
0 REA rica from Africa, South 2n erica from Antarctica, and Antarctica from Indo-Madagascar in a narrow lime interva 
100-90 Ma. The Africa-first hypothesis posits separation of South America from Africa before the beginning of the Late 


1 South America from Antarctica in the Eocene, and Antarctica from Indo-Madagascar late in the Late Cretaceous. 


Annals of the 
Missouri Botanical Garden 


faunas exclusive of Africa. the 
atter model, generally consistent with the paleogeo- 
graphic reconstruction of Hay et al. (1999), invoked 


Antarctica (in combination with two key land bridges) 


More specifically, 


as a dispersal route between Indo-Madagascar (the 
South 


circum- 


Indian subcontinent plus Madagascar) and 


America following isolation of Africa by a 


African seaway prior to the beginning of the Late 
Cretaceous, According to this model, terrestrial 
vertebrate faunas on Africa are predicted have 
become progressively more provincial during the Late 


Cretaceous (after separation between South America 
and Africa prior to the end of the Early Cretaceous), 
while t 


— 


1ose on other Gondwanan landmasses re- 
mained relatively cosmopolitan until considerably 
later in the period. 

Sereno et al. (2004: see also Mahler, 2005) recently 
reported. the important discovery of a number of 
definitive abelisauroid theropods (abelisaurids and 
noasaurids) in Aptian-Albian strata of Niger (approx- 


imately 110 Ma) and Cenomanian deposits of both 


Niger and Morocco (approximately 95 Ma), thus 
providing the first conclusive. evidence for the 


presence of this clade in Africa. These finds support 
the pre-Late Cretaceous origin of Abelisauroidea on 
Gondwana, as previously indicated by several earlier 
Argentina (Coria. & Salgado, 1998 
Lamanna et al., 2002: Rauhut et al., 2003). Thus, one 
alternative hypothesis of Sampson. e (1998: 
1050)—that sometime 


the Early Cretaceous after the tectonic 


discoveries. in 


al. 


“abelisaurids originated in 
isolation of 
Africa“ is clearly refuted. Based primarily on the 
African (2004) 


concluded that the evidence therefore supports the 


new 


abelisauroids, Sereno e 


« 


pan-Gondwana hypothesis. We disagree with the latter 
contention, and regard the Africa-first hypothesis to 
be the 


admittedly limited, evidence (see below). 


more consistent with available, although 


Before assessing this issue, it is necessary to 


(2004). 
. these authors claimed that the original formu- 


address several statements by Sereno et al. 
First 
lation of the Africa-first hypothesis by Sampson et al. 
(1998) included the stipulation that Africa and South 
America separated by 140—120 Ma. 
^h: 1328) stated that 
the Africa-first model, as portrayed by Sampson et al. 


were fully 


Specifically, Sereno et al. (20t 
"shows a cireum- African seaway in the 
140-120 Myr 


other 


Early 
Cretaceous (ca. ago) that isolated the 
continent from Gondwanan — landmasses.” 


Whereas Sampson et al. (1998: 1050) did state that 


“South America separated from Africa before 
100 Ma” and, in the caption for Figure 4, specified 
“circa 120 Ma,” claims for separation as early as 


140 Ma were not made in thal paper, or in any other 


paper published by our working group. Indeed, all of 


] [s 
the pateogeograptic 


reconstructions presented by us 
Africa and South America still 
joined at 120 Ma (Krause et al., 1997b: fig. 1, 1999: 
fig. 6: 5 1998: fig. 4; 2003a: fig. 


2.17; we specifically did not use the 120 million year 


consistently show 


Sampson et al., Krause, 
reconstruction 1n Hay et al. (1999: fies. 12, 15), which 
illustrates full separation between South. America and 
Africa at | 


— 


us time, but instead employed a recon- 
struction generated from the website operated by the 
Hay et al. working group (<http://www.odsn.de/odsn/ 
/pal ip/paleomap.html >), 


these landmasses still connected by a 


Service which | shows 


narrow sub- 
aerial passage at 120 Ma). In any case, the portrayal 
(2004) of our Africa-first hypothesis 
as requiring isolation of Africa as early as 140 Ma is 
Incorrect. 

Second, Sereno et al. (2004: 1328-1329) concluded 
that “a 


hy Sereno et al. 


permanent equatorial seaway of significant 
depth between South America and Africa was in place 
no earlier than the end of the Albian,” and that “trans- 
Atlantic interchange may have been operative as late 
as 95 Myr ago.” Elsewhere in the same paper, Sereno 
el (2004: 1328) stated that “well-constrained 
geological evidence (Reyment € Dingle, 1987: Pitman 
et al, 1993; 2000 final 
separation of South America and Africa in the latest 
Albian (ca. 100 Myr ago), significantly 
proposed | "Africa-first" 
120 Myr To clarify 
Sereno et al. do not make zd definitive assessments. 
Dingle (1987: 99) stated that “final 
continental separation was probably completed in Late 
[not latest} 


al. 


Maisey, the 


— 


pinpoints 


than 
| 40— 


. the three papers cited by 


later 


v the model 


(ca. 
ago). 
Reyment and 
Albian time” (emphasis and bracketed 
words added). Pitman et al. (1993: 23) stated that “the 
in time of final separation must be between 
106 Ma.” (2000: 285) 
concluded that a permanent equatorial seaway joining 
the South 


and and Maisey 
Atlantic and the western Tethys Ocean had 
the 


approximately 112 


developed by late Aptian (which ends at 


Ma; ( )04). That 
should be pointed out that there is other well- 


radstein et al.. 2 
said, it 
constrained geological evidence (Nürnberg € Müller. 
199]: Pletsch et al.. 2001) for a 


connection between the central and southern parts of 


permanent marine 


the Atlantic Ocean, separating South America from 


11 by the late Aptian-early Albian. approximately 


18-106 Ma (Gradstein et al., 2004). Again, we have 
never argued for separation of Africa and South 


\merica as early as 140 Ma, but it must be noted that 
current geophysical and paleogeographie evidence is 
suggestive of separation before the end of the Albian. 
As such, the primary distinction concerning this point 
is that Sereno’s formulation of the pan-Gondwana 
hypothesis views separation between South America 


and Africa as having occurred. afier the end of the 


Volume 93, Number 2 
2006 


Krause et al. 203 


Late Cretaceous Terrestrial Vertebrates 


Early Cretaceous, whereas the Africa-first hypothesis 


supports separation before the end of the Early 
Cretaceous. 

Third, Sereno et al. (2004: 1328) made the claim 
that “other faunal evidence (notosuchian crocodylo- 
morphs (Buckley et al., 2000) and gondwanatherian 
mammals (Sampson et al., 1998)) no longer supports 
ae "Africa-first" 


evidence is far from strong or highly resolve 


model." While we agree that this 


Sereno 


— 
5 


el al/s conclusion simply does not follow from 
their explanatory statement that “African crocodylo- 
morphs of mid-Cretaceous age (Aptian-Albian) are 
most closely related to taxa of comparable age on 
South America (Buffetaut & Taquet 1977, 1979; 
2003), and the absence of gondwa- 
reflects only the 


Sereno et al., 
natheres non-existent record o 
mammals on Africa during most of the Cretaceous." 
Close relationships among African and South Amer- 
ican crocodylomorphs of Aptian-Albian age, which we 
do not dispute, cannot serve as evidence for no longer 


supporting the hypothesis that Africa was the first 


major Gondwanan landmass to become isolated. 
Furthermore, the recent cladistic biogeographic 


analysis by Turner (2004b), reviewed above, demon- 
strated that crocodyliform evidence is indeed consis- 
Finally, the 


gondwanatheres in the 


tent with the Africa-first hypothesis. 
presence of sudamericid 
Campanian/Maastrichtian of South America, Mada- 
gascar, and India was primarily used to support the 
hypothesis that this enigmatic clade of mammals was 


much more cosmopolitan in its distribution in the 


— 


latest Cretaceous than previously realized (Krause el 


al.. 1997b). The 


natheres from the Cretaceous. of mainland 


absence of sudamericid gondwa- 
Africa 
(which, in fact, may not be the case—see Krause el 
al.. 2003b) is no less supportive of an Africa-first 
hypothesis than it was when the paper by Krause et al. 
(1997b) was published. 

In light of the discussion above, let us reexamine 
the current evidence and assess the claims and 
implications of the competing biogeographic hypoth- 
eses by focusing on established geologic and biologic 
constraints. First, the pan-Gondwana hypothesis, as 
defined by Sereno et al. (2004), posits the existence of 
three narrow, intermittent land bridges. all severed in 
the early Late Cretaceous, approximately 100-90 Ma: 
Africa 


between South America and Antarctica, and a third 


one between and South America, another 


between Antarctica and Indo-Madagascar. Geologic 
evidence in support of this model was based on the 
of Scotese (2001). 
a revised Africa-first model, incorporating recent data 
Albian and 
of South 


America and Africa prior to the Early/Late Cretaceous 


paleocoastline maps In contrast, 


on African abelisauroids from the 


Cenomanian, postulates the separation 


—ͤ— 


boundary, but the persistence of land bridges that 
permitted faunal exchange through Antarctica be- 
tween South America to the west and Indo-Madagas- 
car to the east until well into the Late Cretaceous. 
According to this view, early stocks of abelisaurids 
(and other vertebrates) were present on at least South 
America and Africa (and perhaps other Gondwanan 
landmasses) by the late Early Cretaceous. Rifting of 
South America and Africa toward the end of the Early 
Cretaceous isolated at least two stocks of abelisaurids, 
one on each continent. Currently, there is no reason to 
assume that they had spread into Gondwanan land- 
masses outside of Africa and South America at that 
early stage, although this scenario is not inconsistent 
with available evidence. Similarly, in the absence of 
fossil evidence, we cannot yet know what happened to 
the isolated stock of abelisaurids on Africa during the 
Late Cretaceous. However, on the rest of Gondwana, 
a basal stock of Cenomanian abelisaurids (of which 
there is evidence in Argentina; Lamanna et al., 2002) 
diversified into closely related forms observed later in 
and the 


the Cretaceous in Argentina, Madagascar, 


Indian subcontinent. Lacking fossil evidence, this 
hypothesis does not stipulate when during the Late 
Cretaceous those basal stocks (or their descendants) 
first arrived on Madagascar and the Indian sub- 
continent; that is, they may have been present at the 
time South America and Africa separated or they may 
have dispersed much later from South America. 
However, this view is consistent with a// phylogenetic 
evidence presented to date (Sampson et al., 8; 
Carrano et al., 2002; Coria et al., 2002; Wilson et al., 
2003; Sereno et al., 2004) and posits that the derived 
abelisaurids (including several horned forms) present 
in the Maastrichtian of India and Madagascar on 
t 
South 


common ancestor than either did with abelisaurids 


ie one hand, and the Campanian-Maastrichtian of 
America on the other, shared a more recent 
from Africa. 

Concerning paleogeographic reconstructions, none, 


including Seotese (2001), 


depicted a situation ii 


— 


which all three land severed at 


bridges were 
approximately the same time in the early Late 
Cretaceous, approximately 100-90 Ma. As reviewed 
above, current evidence suggests that the South 


America/Africa land bridge ceased to exist prior to 
the Albian/Cenomanian boundary (i.e. prior to 
100 Ma). Indeed, even Scotese (2001) depicted the 
two continents to be already well separated by 94 Ma 
(Schettino & Scotese (2000) 


100 Ma). Current geologic evidence is most consistent 


show separation al 
with the separation of Africa and South America (and 
thus the isolation of Africa) by 100 Ma (Lawver et al.. 
1992; Smith et al., 1994; Marks & Tikku, 2001: 
Scotese, 2001; Kent et al., 2002). By contrast, a land 


= 


204 


Annals of the 
Missouri Botanical Garden 


South America and Antarctica is 


have been present throughout the Late 


bridge between 
thought to 


Cretaceous and until at least the early Eocene (Lawver 


et al., 1992; Woodburne & Case, 1996; Hay et al.. 
1999; Reguero et al., 2002; Lawver & Gahagan, 
2003) there is no evidence for a termination. of 


as detailed above. 


100-90 Ma. 


the timing of separation of 


connection Finally, 


Antarctica from Indo- 


Most 


that Indo-Madagascar became isolated from 


Madagascar is controversial. reconstructions 
indicate 
all other Gondwanan landmasses about 120 Ma. some 
50 million vears prior to the time that the Maevarano 
as well as those from India) 
was deposited (e.g.. Lawver et al., 1992; Smith et al.. 
1994; Marks & 2001; Rotstein et al., 2001: 
Scotese, 2001; Kent et al., 2002: Bernard et al., 2005). 


Notably, and in stark contrast, the tectonic 


vertebrate 


assemblage 
Tikku. 


however, 
modeling of Hay et al. (1999) provides support for the 
scenario posited by the Africa-first hypothesis in that 


a land bridge between Antarctica and Indo-Madagas- 


car was maintained until approximately 80 Ma (cf. 
2002). 


indicate synchroneity 


Case. In neither case is there evidence to 


with the separation of South 
Antarctica at 100— 
(2004) pan- 


America from either Africa or 


90 Ma. as postulated by Sereno et alis 
Gondwana model. 

Concerning faunal evidence, the crocodyliforms. 
nonavian dinosaurs. and mammals recovered from the 
Maevarano Formation of northwestern Madagascar are 
known from 


similar to those 


axonomically most 
Campanian/Maastrichtian horizons of South America 
and the Indian subcontinent (Krause et al.. 997b: 
Buckley & Brochu. 1999: Buckley et al.. 2000: 
2001: Wilson et al., 2001: Rogers. 
2002; Prasad & de Lapparent de Broin, 


Curry 
2002; Turner, 


2004b). Because strength of biogeographic signal 


Krause, 


proportional to the number of phylogenetically in- 


dependent groups possessing the same congruent 
I 8 8 
pattern of area cladograms, it is significant to note that 


the patterns of similarity are repeated among a growing 


number of phylogenetically independent groups ( 


o 
Que 
notosuchids. peirosaurids, trematochampsids. and the 
unclassified metasuchian genus Araripesuchus among 


and salt- 


crocodyliforms; abelisaurids. noasaurids. 
asaurines among nonavian dinosaurs; and sudamer- 
icids among mammals). This provides evidence for 
cosmopolitanism among latest 
\frica that 


was nol fully appreciated prior to recovery of the 


a high degree of 


Cretaceous Gondwanan faunas outside of 


Maevarano assemblage and is unpredicte d based on 


most recent paleogeographic reconstructions of. the 
southern supercontinent (e.e.. Lawver et al.. 1992: 
Smith et al., 1994: Scotese, 1998, 2001: Marks & 
Tikku, 2001: Rotstein et al.. 2001: Kent et al.. 2002: 


Bernard et al., 2005: Schettino & Seotese. 


2005) or 


the pan-Gondwanan hypothesis of Sereno et. « 


2 


the of abelisaurids and 


noasaurids on Africa 25 or more million years prior lo 


Finally. mere presence 
their occurrence in Madagascar and India does not 


constitute evidence refuting the Africa-first hypothe- 


sis. Although recent discoveries. demonstrate the 
presence of abelisauroids in the Early Cretaceous 
and earliest Late Cretaceous of mainland Africa. 


certainly an interesting and significant finding. 


phylogenetic analysis does not include placement « 
the recovered forms among the more derived members 
al., 2004). the 


abelisaurid Rugops is postulated. by Sereno et al. 
80] | y 


Indeed, 


of this clade (Sereno e 


(2004) as the basalmost member of the group. Thus, 


although it appears that Africa did not separate from 
South 


boundary, the closest relatives of Majungatholus and 


America until near the Early/Late Cretaceous 
Masiakasaurus from the Maastrichtian of Madagascar 
are still found in Campanian/Maastrichtian horizons of 
South 
faunal evidence to refute the hypothesis that Africa 


America and India. As such, 
was the first among major Gondwanan landmasses to 
be fully isolated and, perhaps most importantly, thal 
Gondwanan faunas 


the terrestrial vertebrates of 


outside of Africa were shared until late in the Late 
Cretaceous. 

In summary, we regard the growing weight of latest 
Cretaceous biogeographic data as evidence in support 
\frica-first The 


terrestrial vertebrates that lived on 


l a modified hypothesis. close 


relationships of 
the Indian subcontinent and Madagascar near the end 
of the Cretaceous to penecontemporaneous taxa in 
South America appear incompatible with any pro- 
posed. lengthy separation of Indo-Madagascar and 
South 


that duration was approximately 50 million years. as 


America prior to the Maastrichtian. whether 


indicated by most recent paleogeographic reconstruc- 
tions of Cretaceous Gondwana, or approximately 25 


million years. as posited by the pan-Gondwana 


hypothesis. 


Literature Cited 


\pestiguia, S. 2004. Bonitasaura salgadoi gen. et sp. 
nov.: beaked sauropod from the a ee of 


Patagonia. Naturwissenschaften 91; 493— 
Asher, R. J. & D. W. Krause. 1998. The ed as 
(Cretaceous) record of Anura from N 1 ] 


x 696—699, 

Archibald & T. 
alleged marsupial 
Veta Palaeontol. 


Vertebrate 
Averianov, A. O. J. D. 
Placental 
Cretaceous 
149-151 


Bernard. 


Paleontol. lé 

) Martin. 2002. 
nature of the from the 
of Madagascar. Polonica 18: 
. M. Munsehy. Y. Rotstein & D. Sauter. 2005. 
Refined spreading history ab the Southwest Indian Ridge 


for the last 96 Ma. with the aid of satellite gravity data. 
Geophys. J. Int. 162: 765-778. 


Volume 93, Number 2 


2006 


Krause et al. 205 
Late Cretaceous Terrestrial Vertebrates 


Bertini, 


& L 


Carvalho. 1999. Distribuição 


Seen 'a em crocodilomorfos notossüquios e ocorrén- 


clas nas Bacias Cretacic 
Simpósio sobre o Cretáce 
L. G. Marshall, 
Vertebrate faunas from 
formations (upper Bauru 


as Brasileiras. Bole tim do 5 
o do Brasil, : 

M. Gayet à P. Brito. 1993. 
the Adamantina and Marilia 
Group, Late Cretaceous, Brazil) 


in their d s and AO context. 
71-101. 


Neues Jahrb. Geol. Palio 
Besairie. H. 1972. Géologic 


sédimentaires. Annales G 
K. 19 


ntol., Abh. 188: 
> de Madagascar. I. Les terrains 
éol. Madagaso ar 35: 1463. 


Brochu, C. A. 1997. Morphology, fossils, divergence timing. 


and the phylogenetic relationships of Gavialis. Syst. Biol. 
522 


Brune M. et al. 
1 0 res du Crétacé 
de l'Oue ste. Compt. Rene 


10 authors). 1990. Nouveaux 


inférieur du Cameroun, Afrique 
J. Acad. Sci. 310: 1139-1146. 


Buckley, G. A. & C. A. Brochu. 1999, An enigmatic new 
crocodile from the Upper Cretaceous of Madagascar. 


Pp. 149-175 in D. Unwin (editor), Cretaceous Fossil 


Vertebrates: Special Pape 


Palaeontological Associat 
& ————. 2001. 


Tf 


rs in Palaeontology No. 60. The 
ion, London. 


A skull of „ 


insignis (( y 
Moran: J. 1 
& UM 


the i iC ¢ 


) from the Uppe! Cretaceous of 
» Paleontol. 21: 30A. 

. 1997. e ee and 
rigins of the Late Cretaceous 


croc odylitontis E E ar. J. Vertebrate Paleontol. 17: 
35A. 


croc odio dora the | 
Nature : 941—944 
& J. A. 


shouted crocodyliform fre 


). Pol. 2000. A pug-nosed 
ate Cretaceous of Madagascar. 


Georgi. 2003. A new slender- 
m the Upper Cretaceous Mac- 


varano Formation, Madagascar. J. Vertebrate Paleontol. 
23: 37A. 


Buffe tauta. E. 1982. Radiat 
biogé 1 d 5 crocod 
Géol. Fr : 60: 1-88. 
1988. The QAUM 
Messel: A reassessment. 
enberg 107: 211-22 
8 


ion évolutive, paléoécologie et 
iliens mésosuchiens. Mém. Soc. 


mt mesosuchian crocodile from 
Courier Forschungsinst. Senek- 


19089. A new colado mesosuchian crocodile 


from the Eocene of Alge 
1-10. 


& P. Taquet. 1977. 


suchus in 


. 1979. 
suchien dans le Campai 
Te oblita, n. sp. 
3-188. 


aid M. T., S. D. Samp: 


ec 


ria. Palaeontographica A. 208 


The giant erocodilian Sarco- 


the Early Cretaceous of Brazil and Niger. 
Palaeontology 20: 203-208. 
& 


Un nouveau Crocodilien méso- 


ven de Madagascar: Tremato- 
ull. Soc. Géol. France 21: 


son & C. A. Fors 2002. 


Phe 
osteology of 4 ea knopfleri, a al is lisauroid 


ao Theropoda) from the Late Cretaceous o 
Madagascar. J. li: rtebrate Paleontol. 22: 510-534 
M. Loewen. 2004. New discoveries. of 


i pes 
Noasauridae (Dinosauria: 
leontol. 24: 44; 

Carvalho, |. de 80152 Za: Lo 
Avilla. 2004. Uberabasuc 
Mein Mus from tl 
ceous), Brazil. Gondwana 

Case, J. X. a i new biog 


Late Cretaceous vertebrate 


Vertebrate Paleontol. 22 


and the morphology of the 
Theropoda). J. Vertebrate Pa- 


B. Ribeiro & L. dos Santos 
hus terrificu a a new 
1e a Basin 
Res. 7: 975-1002. 


os model for dispersal of 


(Uppe r Creta- 


s into Madagascar and India. J. 


N 
=> 


& D. W. Krause. 2002. A tribosphenic lower molar 
from the Maastrichtian of Madagascar: Phyletic affinities, 
biogeography and a new dispersal model. International 
Palaeontological Congress, Sydney. Australia (abstract). 

et al. (+ 6 authors). 2000. The first duck-billed 
dinosaur (Family Hadrosauridae) from Antarctica. J. 
Vertebr: ate Paleontol. 20: 612-614. 

. J. E. Martin, D. S. Chaney € M. Reguero. 2003. 
Late Ca taceous dinosaurs from the "na tic Pe 8 


— 


remnant or immigrant fauna? J. Vertebrate Paleontol. 2 
30; 

Casey. M., R. Rogers, M. Jackson & G. Buckley. 2003. 
Magnetic stratigraphy of the Upper Cretaceous Maevarano 
Formation (C ampanian?— Maastric pees northwestern Ma- 
dagascar. J. Vertebrate Paleontol. 

Censky. E. J.. K. Hodge € J. Dud 155 1008. Over-water 

) 


dispersal of lizards due to hurricanes. Nature 395 


Chinsamy. A. & A. Elzanowski. 2001. Evolution of growth 
pattern in birds. Nature 412: 402-403. 

Coffin, M. F. et al. (+ 6 authors). 2002. Kerguelen 51 
magma output since 130 Ma. J. Petrology 43: 1121-113 


oria, R. A. & L. Salgado. 1998. A basal Abelisauria o 


1992 sage aie C ee from the Cretaceous of 
Patagonia, Arge n . Perez-Moreno, T. R. Holtz 
Jr. J. L. Sanz " 3 Mc dli (editors), Aspects of 


Theropoda Paleobiology. Gala 15: 89-102. 
L. 


= 
= 


uappe & L. Dingus. 2002. A new close 
re lative af Giors sastrel eer 1985 (Thero- 
poda: Abelisauridae) from the Late Cretaceous of 
Patagonia. 5 Vertebrate Paleontol. 22: 460-465. 
urrie, P. J. 1996. Out of Africa: ore eating 5 that 
challenge Tyrannosaurus rex. Science 272: € 

Curry, K. A. 1997. Vertebrate fossils ES ^ Upper 
Cretaceous a nA Sandstones, m Ba- 

\ 


sin. Madagascar. J. Vertebrate Paleontol. 17: 40A. 


Curry Rogers. K. 2001. A new sauropod from eee 


Implications for titanosaur lower-level phylogeny. 
Verte brate Paleontol. 21: 43A 

2002. Titanosaur alls tell two tales: A second 
Malagasy titanosaur with TON affinities. J. Verte- 
brate Paleontol. 22: 47A-48, 

2005 nadas p phylogenetic overvie 
Pp. 50-1( 3i in K. Curry Rogers & J. A. Wi 


The Sauropods: Evolution and Paleobiology. Univ. Cali- 


" 
son tors 


fornia Pre ‘ss. Berkeley. 
& C. A. Forster. 2001. The last of the dinosaur titans: 
^ new sauropod f rom Madagascar. Nature 412: 530-534. 
4. 


2004. The skull of dor male krausei 
(Sauropoda; Titanosauria) from the Late Creta s of 
Madagascar. J. Vertebrate Paleontol. 24: 121-1: ] E 


Depéret, € 


. 1896. Note sur les Dinosauriens Sauropodes el 
Thé 1 5 s du Crétacé ou de Madagascar. Bull. 
—19: 


Soc. Géol. France 21: 17 1. 
de Wit, M. J. 2003. Madagascar: Heads it's a contine nt, tails 
its an island. Annual Rev. Earth Planet. 31: 213-248 


Dodson, P.. D. W. Krause, C. A. Forster, 5. DS Sampson & F. 
Ravoavy. 1998. Titanosaurid P oste ode rms from 


the Late Cretaceous of Madagascar. J. tebrate 

Paleontol. 18: 563—568. 
Forster, C. “ 909. Gondwanan dinosaur evolution and 
bioge ES analysis. J. African Earth Sei. 28: 169-185. 
& P. M. O'Connor. 2000. The Athan of the Upper 


Cretaceous Maevarano Formation, Madagascar. J. Verte- 
brate Paleontol. 20: 41 A-42A. 

, L. M. Chiappe, D. W. Krause & S. D. Sampson. 
1996. The first Cretaceous bird from Madagascar. Nature 
382 34 


D 


Annals of the 
Missouri Botanical Garden 


. Krause. 


New evidence from 


—— ———, 8. D. Sampson. L. M. Chiappe & D. 
1008. The theropod ancestry of birds: 

the Late Cretaceous of 
1915-1919. 


D 


Madagascar. Science 27 


— Chiappe, S. D. < 
2003. us berivotrensis, a primitive bird from the 
268-280 ir 
& L. Witmer (editors), Mesozoic Birds, 19 1 the Heads of 
Berkeley 

Biogeographic oe of the 
I 59-184 in 
Contributions of Southern South 


Sampson & D. W. Krause. 


Cretaceous of Madagascar. Pp. 
aurs. Univ. California Press, 
1996. 


nerican crocodilians. Pp. 


Dinos 
South 
. Arratia rani 
America to Vertebrate 


Abhandlun- 


Gasparini, Z. 
An 


Paleontology. München geowissenschaftliche 


gen, Reihe A, Geologie und Paláontologie 3 
. Pereda-Suberbiola & R. Molnar. 1996. New data 


on the ankylosaurian dinosaur from the Late Cretaceous of 
In F. E 


Novas & R. E. Molnar 

Gondwanan Dinosaur Sym- 

. Mem. Que I n "s Mus. 39: 583-594. 

Codinot, M. & €. V. 1994. 
Cretaceous arboreal e a rians. Nalurwissenschaften 81: 
4 79-81. 

González Riga, B. J. 2003 


Sauropoda) from the L 


ie Peninsula. n F. E. 


P M 


\ new titanosaur e 


oper Cretaceous lendoz: 
Argentina. Ameghiniana 40: 155-172 
Gradstein, F. M., J. G. Ogg & „ Smith. 2004. 
Time Seale 2004. unidas Univ. Press, Cambridge. 
Hahn, G. & R. Hahn. 2003. New multituberculate teeth from 
the Early Acla 


Province, 


\ Geologic 


j Cretaceous of Morocco. Palaeontol. 
Polonica 48: 349-350. 


Hay, W. + 10 authors). 1999, An alte native global 
. Pp. le 


Cretaceous pss p. 


„etal. ( 


8 in E. Berrera & C. 


ed (editors), ee A of Cretaceous Ocean/Climate 

stems. Geological Society of America Special Paper 

Hooker. J. J.. Milner & S. E. 
ornithopod lim from the Late Cretaceous o 
Antari Antarctic Res. 3: 331—332 

Hopson, J. A. & G. W 


1991. An 
f West 


K. Sequeira. 


“Llica. A 

\ Rougic r. 1993, Braincase structure in 
the oldest known skull of a therian mammal: Implications 
for mammalian systematics and cranial evolution. J 
Sci. 293: 268—299 


Kent, R. W., M. 


Amer. J. 


. Pringle, D. Müller. A. D. Saunders & 


N. C. Chase, 2002. ""Ar/?Ar geochronology of the 
Rajmahal Basalts, India, and their re 1 to the 
Kergue 2 n Plateau. J. Petrology 43: 1141-11 

Khosla, A. & A. Sahni. 2003. Biodiversity M i Deccan 


volcanic Earth Ser 21: 
895- 6s 


MUR S 


pen episode. J. Asian 


rasad, O. Verma, A. K. Jain & A. Sahni. 


2004. SEA of a micromammal-vielding Deccan 

intertrappean site near Kisalpuri. Dindori District 

Madhya Pradesh. Curr. Sci. 87: 380-368: 
Kielan-Jaworowska, Z.. R. Cifelli & ZX. Luo. 2004. 


Mammals from the Age of Dinosaurs: Origins, Evolution, 


New York. 


New mammalian specimens from the 
y 


and ee ure. Columbia Univ. Press, 
Krause, D. 2000. 
ES relaceous of Madagascar. J. Vertebrate Paleontol. 
20: 357 
2001. Fossil molar from a Madagascan marsupial. 
atun 412: 497-498 
2002. Late Cretaceous (Maastrichtian) mammals 
from Madagascar: Implications for the evolutionary and 
biogeographic history of Gondwanan mammals. Sth In- 
Terrestrial 


ternational Symposium on Mesozoic Ecosys- 


tems, Programme and Abstracts. p. 51 


ale 


. M. Chiappe 


Discovery of 


— 2003 
lian specimen from the Late Cretaceous of Madagascar. J. 
Vertebrate Paleontol. 23: 69A 

— — —, 2003b. Late Cretaceous vertebrates from Madagas- 
car: À window into Gondwanan biogeography at the end of 

| 407 in S. M. Goodman & J. P. 

Natural Elision of Madagascar. 


. Discovery of a relatively complete mamma- 


the Age of Dinosaurs. Pp 
Benstead (editors), The 


niv. Chicago Press, Chicago 
— ——— & F. E. Grine. 1996. The | 


first multituberculates from 

Madagascar: Implications for Cretaccous biogeography. J. 

Vertebrate Paleontol. 16: 46/ 
K J. II. 1990. la 


: Cretaceous fossils from 


Madagascar and thei 1 155 ations for bioge i 
re relationships with the Indian subcontinent. Pp. 135-15 
Sahni (edito Crel Stratigraphy 11 


Palaeoenvironments. a Geological Society of India 
37 


. (+ 8 authors). 1994. p an Panas mammals 
Nature 308: 
N. 


[from Mu ar]. | 
H artman & 


E Y 


Cretaceous verte 


— 


Wells. 


M: ulagaso “an 


1997a. Late 

‘brates from _ Implications 

10 ed c p ange in deep time. Pp. 343 . D. Goodman 

. D. Patterson (editors). 19 vin and Human 
LEY 


19 i Madagascar. Smithsonian Institution Press, 
W m) D.C. 
. V. R. Prasad, W. A. Sahni & F. 
T . 1997b. C osmopolitanism an among Late Cretaceous 
donde anan mammals. Nature 39 : 04—507. 
eta 5 authors). 1996 The Latc 


vertebrate a of Madagascar: 1 ations for Gondwa- 


von Koenigswald, 


Cretaceous 


nan pale obioge ography. GSA Today 9: 
— p & K.-Q. Gao. 2002a. 
rec en of 55 "SOZOLC bapa from Madagascar. J. 
2 


n irst definitive 
Vertebrate 


Evans 


Paleontol. 23: 842-8506. 
.M 


b. Gottfried. P. M. O'Connor & E. 


Roberts. 


2003b. Cretaceous 9 from ‘Tanzania. Acta 
Palaeontol. Polonica 48: 321-330 

Lamanna, M. C.. R. D. Martínez € J. B. Smith. 2002. A 
definitive abelisaurid theropod dinosaur from the un 


Late : retaceous of Patagonia. J. Vertebrate Paleontol. 2 
58-04 

& B. Gado. 2000. A new Early 
Neues Jahrb. € 


Larsson, IL. B Cretaceous 


croc bli from Niger. „col. Paláontol.. 
Abh. 217: 131-142 
A. & L. 


seaways in the 


cack 1 M. ee 2003. Evolution of Cenozoic 
cireum-Antarelie region. 
Palaeoecol. 198: 11—37. 
M. F. Coffin. 1992, The de ve 17 05 nt of 
Salon oseaways around Antarctica. Pp. 7-30 „ Kennett 
& D. ALN 


Antarctic (Cod 


Palacogeogr. 
Palaeoclimatol. 


arnke (editors), The 
i Antarctic. Re- 
American Geophysical Union, Wash- 
Dru p. 2: 
. S. Apesteguia, F. as & M. S. de la 
2004. from the 
Neuquen Basin (Argentina) and their tetrapod assem- 


O 
Fuente. Gretaceous c beds 


blages. Cretaceous hes. 25: 01-87. 

Mahler, L. 2005. Record of Abelisauridae (Dinosauria: 
Theropoda) from dpi m of Morocco. J. Verte- 
brate Paleontol. 25: 236-239 


Maisey, J. G. 


2000. Contine ee bre ak up and the distribution 
of fishes C 


in western a during the Early Creta- 

ceous. retaceous Res. 21: 281-31 

Marks, M Tikku. a 

Lions S East "RA Alric a and Madagascar. 
49: 


Planet. Sci. Lett. 186: 479— 


G relaceous reconstruc- 
Earth 


Volume 93, Number 2 
2006 


Krause et al. 
Late Cretaceous Terrestrial Vertebrates 


Müller, R. D., J.-Y. Royer & L. A. 


plate motions relative to the 


. 1993. 


hotspots dium combined 


Lawver Revised 


Atlantic and Indian Ocean hotspot tracks. Geology 21: 
215-21 
Néraudea E Mathey. 2000. Biogeography and 


& 
diversity v of South Adaniie Cretaceous echinoids: Implica- 
tions for circulation patterns. Palaeogeogr. Palaeoclimatol. 


Palaeoecol. od / 1—88. 
Nessov, L. A.. V. I. Zhegallo & A. O. Averianov. 1998. A new 


locality of Late Cretaceous snakes, mammals and other 


vertebrates in Africa (western Libya). Ann. Paléontol. 84 
265-27 


Novas, F. E. & F. L. Agnolin. 2004. Unquillosaurus ceibali 
Powell, a giant maniraptoran (Dinosauria, Therope n n. 


the Late Cretaceous of Argentina. Revista Mus. qe gent. Ci. 
Nal., new ser., 6: 61-66. 

Cambiaso & A. Ambrosio. 2004. new 
basal olaaa (Dinosauria, Ornithse hia). from 
the w er Cretaceous of Patagonia. Ameghiniana 4l: 


Nüraber TE. xin & R. D. Müller. 1991. The tectonic evolution of 

the South. Atlantie Ocean p Late Jurassic to present. 
ectonophysics 191: 27—5 

O'Connor, P. M. In e 


press. pneumalicity: an 


evaluation. of solis tissue influences on the posteranial 
skeleton and the reconstruction of pulmonary anatomy in 


10saurs. J. p 5 


AA 


arc 
& L. Claessens. 2005. Basic avian pulmonary 
design and 1 5 ea ventilation in non-avian theropod 
Nature 436: 253-256 
al. (+ 6 authors). 2006 


'elaceous Red $ 


pecie 
A new vertebrate fauna 
Sandstone Group, Rukwa Rifi 


from 5 
3 can Earth Sei. 44: 


Mu ui ane ida, Tanzania. J. Afri 
278. 


O Nei "n C.. D. Müller & B. Steinberger. 


2003. Geodynamic 
implications of moving Indiar h 


Ocean hotspots. Eart 
Planet. Sci. Lett. 215: 151-16 
1 


E F.. Z. Gasparini, Buse alioni & J. 
2C 


Pu "e eous ‘of Paleo. 
T Arge e 1 Vertebrate Pale ontol. 20: 57-76. 
Pitman. W. C. III. S. Cande, J. LaBreque & a 1 11. 1993. 
ascia ntation of Gondwana: The d ce of Africa from 

uth America. Pp. 15-34 in Goldblatt 
Biological Africa 


aven. 


(editor). 
Relationships between and South 
America. Yale Univ. Press, New 
. T. et al. (+ 7 authors). 2001. Cretaceous separation 
of Africa and South America: The view from the West 
Mica n margin (ODP Leg 159). J. 5. Amer. Earth Sci. 14: 
7-174. 
Pal. D. 2003. New 
(c ly] pl Mesoeucrocodylia) 


Brazil. J. 


remains of Sphagesaurus | huenei 
from the Late 


Paleontol. 23: 


Cretaceous of Vertebrate 
7-83 1. 
Prasad, G. \ 
from the 
892—902 


R. & M. 


Late 


Godinot. 1994. Eutherian tarsals 
| 


of India. J. 


co 


Cretaceous Paleontol. 6€ 
& A. Sahni. 1988. First Cretaceous mammal from 
Nature 332: 638-040. 

& F. de Lapparent de Broin. 2002. 


India. 


ate Cretaceous 
crocodile remains from Naskal (Ir 
ffinities. Ann. 
J. Jaeger, A. Sahn 
1994. Euthe e al 
nee taceous (Maastrichtian) Intertrappean bec 
ndhra Pradesh, India. J. 
e 


adia): pd and 
Palé 'ontol. 88: 19-71. 

Gh Meus & C.K 
Upper 
Is of Naskal, 
Paleontol. 14: 


biogeographic a 


————., J. 
Khajuria. from the 


lan 


Vertebrate 


Rana, S. & ( . Wilson. 2003. New Late Cretaceous 
1 from a Intertrappean beds of Rangapur, India 
and paleobiogeographic framework. Acta Palacontol. 
Polonica 48: 331-348. 

dos Acs J. 


imba. 1981. 


iras ntation marine au nord-est de M 


Zimmerman, R. Rambeloson & G. 


-L. 
Précisions sur lage de la 


premiere 
T en liason 


avec la dislocation du bloc Seyche elle s-Indes-Madagascar. 
Compt. Rend. Séances Acad. u^ DEK: 3, Sci. Vie 292 
1039-1042 

Rasmusson, E L. 2002. Interrelationships of tremato- 


champsid crocodyliforms. J. Vertebrate Paleontol. 22: 
Buckley. 2001. New > 
e croc codyliform 7 ja hampsa oblita, from the 
Vertebrate 


material of 


». A. 
ile b retaceous of Madagascar. J. Paleontol. 
2]: LA. 

aa O. W. 

)03. . 
ni Chubut 
187—497 

Reeves. C. & M. de 


Gondwana: 


>. Cladera, P. Vickers-Rich & T. 


remains from the 


Rich. 


Lower Cretaceous of 
> 


Group, Argentina. Cretaceous Res. 24: 


Wit. 2000. Making 


Retracing the transforms of the 


ends meet in 
Indian Ocean 
and reconnecting continental shear zones. Terra Nova 12: 
272-280. 

Marenssi & S. 2002. 
South (Patagonia) 
terrestrial faunas and environments: 


Reguero, M. . Santillana. 
Antarctic 


Paleogene 


Peninsula and America 
Biogeo- 
Palaeogeogr. Palaeoclimatol. Pa- 
0 


tevment, R. A. € R. V. Dingle. 1987. 
Africa during the Cretaceous period. Palacoge 
climatol. Palacoecol. 59: 975 16 

Rich, T. H., P. Vickers-Rich, 

) 


probable 


Paleogeography of 


gr. Palaeo- 


. Fernandez & S. Santillana. 


rre from Seymour jp 


© Peninsula. Pp. 219-222 in Y. Tomida, 
P. Vi 


'kers- Rich: (editors). Proceedings e 115 
Second Gondwanan Dinosaur Symposium. National Sci- 
ence Monographs No. 15. 

Roeser. H. X., J. Fritsc i & K. Hinz. 1996. The development 
of the crust off Dronning Maud E and, East 
Pp. 243-264 in B. C. Storey, E. C. King & R. A. Li 
(editors), Weddell Sea Tectonics ain Minis Break-u 
Geological Society Special Publication No 


Marctica. 
ivermore 


Rogers, R. R. 2005. Fine-grained debris ire and 
extraordinary vertebrate burials in the Late Cretaceous 

of Madagascar. Geo ogy 3: i 7-300. 
J. H. Hartman & D. W. Krause. 2000. Stratigraphic 


analysis of Upper Cretaceous rocks in the Mahajanga 
Basin, northwestern Madagascar: oleo for ancient 
and modern faunas. J. Ge E 108: 215-301. 

— — — & —————. 2001. 


Gone r Cretaceous roc ke in the Wan Basin, north- 


Stratigraphic analysis 


western Madagascar: Implications for ancient and modern 
faunas: Reply. J. Geol. 109: 674-676. 
): 


— 
o 


D. W. Krause & K. Curry Rogers. 
C annibalism i in the Msg om an dinosaur Majungatholus 
atopus. Natur 422: 515-518. 


lotsteln, 15 M. Munschy & A. 


Kerguelen Province 


2001. The 


Additional constraints on 


Bernard. 
revisited: 
the early development of the Southeast Indian Ocean. Mar 
Geophys. Res. 22: 81-100. 
Sampson, S. D. € D. W. Krause (editor In. pre 
i lisauridae ) 
rtebrate 


m 
Vries ee le crenatissimus (Theropoda: : 
Cr 


from the Late Cretaceous of Madagascar. J. V 


Paleontol. Memoirs. 


208 


Annals of the 
Missouri Botanical Garden 


Sampson, S. D. et al. (+ 6 authors). 


remains from Madagascar: 


1998. Predatory dino- 
saur Implications. for 
C Vp. biogeography of 


10481 


Gondwana. Science 2 


E Carrano & C. A. Forster. 2001. A bizarre 
pre deta dinosaur from the | Late Cretaceous of 1 1 0 


car. Nature 409: 504—500. 
Schettino, A. € C. R. Seotese. 2000. A Synthetic APWP for 
Africa (Jurassic-Present) and Global Plate ‘Tectonic 


Reconstructions. Eos Trans. 
———— K ——— 2005. Apparent polar wander paths for 
the major continents 200 O Ma to the present day): 

palacomagnetic reference frame for E i plate tectonic 

reconstructions. „ IJ. Int. 103 
Scotese. C. R. . Continental drift (0-750 million years), 
PALEOMAP 
Arlington, Arlington, ' 
\tlas of Earth History, PALEOM 
Texas. 


: 727-159 


a quie ektime e animalion: 


P roject. 
University of ' 
2001 
Arlington, 
Sereno, P. C. et al. (+ 8 authors). 1990. Predatory dinosaurs 
from the Sahara anch! 
Se it 


exas al 


texas. 


AP proje cl, 


Late Cretaceous faunal differentia- 
272: 986-99]. 
race HH. C. E. 
new notosuchian from the | 
Vertebr: P Paleontol. T. 177 
J ilson & J. 
1 in 
\ . London B 271: 
r . 1991. 


Lion. 


Larsson & B. Gado. 2003. A 
2 Cretaceous. of Niger 2. 
1 182. 

2004. 
mid-Cretaceous. 
325-1330. 


First evidence of Multitubereu- 


. Conrad. New dinosaurs 
Proc 


link io 


Roy. Soc 


lata (Mammalia) i the Mesozoic of Africa. Neues Jahrb 
Geol decas . Monatsch. 119-125. 

“vans. J. F. Levine & D. A. Russell. 1998. The 
Fado a re à iceous microvertebrate locality of Anoual. 


Moroc CO: 


of Af 


V glimpse at the small verte huis assemblages 
rica. Pp. 177-181 in S. G. Lucas, J. land & J. 
(editors), Lower and 1 p: relaceous 
Te cen Ecosystems. New Mexico Mus. Nat. Hist. 
Bull. 
Smith. . 
5 
Press, Cambridge, 
m M., J. 6 
& 6 1992 


2 


Estep 


). G. 


and 


Smith & B. M. 

Cenozoic 

Eng land. 
Mitchell, M. C. Abranches, S. Maaloe 
The coast- euro dolerite dykes of 


Maderas age of 


1994. Atlas of 


Cambridge 


Funnell. 


Coastlines Univ. 


i Robin. 


east remagne lization and 


tectonic aspects. J. African Earth Ser. 15: 237-249. 
Storey. M.. J. J. Mahoney & . D. Saunders. 1997. 
Cretaceous basalts in Madagascar and the transition 
between ae and continental lithosphere mantle 
sources. Pp. 95-122 in Large leneous Provinces: Conti- 
nental, Oceanic, and Planetary Flood Volcanism. Geo- 


physical | i in 100, American Geophysical Union. 
— et al. (+5 1995. Timing of hot spot-related 
volcanism and i clap al Madagascar and India. 
rence 207: 052-055 
1980. A paolo phalosaurid dinosaur from the 
Upper Cretaceous of Madagascar and its paleobiogeogra- 
phical implications. J. Paleontol. 54: 954—902 


uthors). 


P. Taquet. 1979, A pachycephalosaurid dinosaur 
from Madagascar and a Laurasia-Gondwanaland connec- 


tion in the Cretaceous. Nature 270: 633— 


Suberbiola, X. P.. N. Bardet, larochene, B. Bouya & M. 
pus 2004. The first ie of a 1 dinosaur 
from the Late Cretaceous. phosphates of Morocco, 
African ae Sci. 40: 81-88. 

Torsvik, el 


t al. (+ 5 authors). 1998. 
1 in Madagascar: Palacomagnetic evidence for 
a stationary Marion hotspot. Earth. Planet. Ser. Lett. 16 
2]-232 


Late Cretaceous 


— ———— et al. (+6 authors). 2001. Late Cretaceous India- 
Madagascar fit and timing of break-up related magmatism. 
Terra Nova 12: : 

Turner, A. H. 2004a. A 
5 . from the I 

Verte 

—— 2004b. 


new species of Araripesuchus 
on Cretaceous of Madagas- 


: 123A 


1 5 8 during the 


rate Paleontol. 


Cretaceous: Evidence of Gondwanan  vicariance pon 
biogeographic analysis. Proc. Roy. Soc. London B 2 
200: JOO 

Purner, : press. Osteology mal huge ny of a new 
species of Araripesuchus (C y Mesoeucroco- 


dylia) from the Late Cretaceous of Madagascar. Hist. Biol. 
———— & J. 0. Calvo. 2005. 
erocodyliform from the I 


new  sebecosuchian 


ate Cretaceous of Patagonia. 
rate Paleontol. 25: 87-98. 

Tykoski. R. S. & T. " 
B. Weishampel, 


The Dinosauria. 


o Pp. 47-70 
in D. smólska (editors). 
Univ. 
RA. Ketchan & M. 
Calsoyasuchus valliceps, a from the 
y. sic Kayenta F ormation of Arizona. J. Vertebrate 
Paleontol. ET 593-6 | 
Upchureh, P., C eine D. B. Norman. 2002. An analysis 
of 1 en ography: Evidence for the existence of 
vicariance and 


new erocodyliform 


dispersal patterns caused geological 
events. 


W V MEE D. SQ 


Dinosauria. 2. ed. 
IE A. 


. Osmóls an 


Univ. Caliber Press, Berkeley. 


The 


Wells, N. 2003. Some hypotheses on the Mesozoic an 
Cenozoic pu oenvironmental 
Pp. 16-34 in S. M. Goodman & J. 
The Natural Bist of Madagascar. 
Chicago. 

Wilson. J. A. 2002. Sauropod pp. 5 ny: Critique 
and Era analysis. Z in 136 

x 


history of Madagascar. 
P. Benstead (editors). 


Univ. 


= 
= 


Chicago Press, 


Zool. J. 
2003. 


: 217-270. 


> 


P. Upchurch, revision of Titanosaurus 


35 poda). the first dinosaur 
genus with a “Gondwanan” distribution. J. Svs Pal: aeon- 
tol. 1: 125-100. 

. M. Sadiq Malkani & P. D. Gingerich. 2001. New 
crocod ion m (Reptilia, Mesoeucrocodylia) from the 


Upper Cretaceous Pab Formation of Vitakri, Balochistan 


aki tan). Contr. Mus. Paleontol. Univ. Michigan 30: 
el Lui 5 authors). 2003. A new abelisaurid 
(Dinosauria, Theropoda) from the Lameta Formation 


(Cre : aceous, Maastric "ud of India. Contr. Mus. 


Paleon- 


tol., . Mic wai 31: 1-42. 

m M. O. y Case. 1996. Dispersal, vicartance, 
and the Late eee to Early Tertiary land mammal 
biogeography from South America to Australia. J. 
Mammalian Evol. 3: 121-162 


EVOLUTION AND GEOGRAPHY: Rosendo Pascual? 
THE BIOGEOGRAPHIC HISTORY 

OF SOUTH AMERICAN 

LAND MAMMALS' 


ABSTRACT 


Based on the unequaled southern Late Cretaceous-Cenozoie record, the history of South American land mammals i 
divisible into two major episodes we term the Gondwanan Episode and the South American Episode. The former ale is 
y th i 


distinguished b: northern and central Patagonian Argentinian record, while the latter is mostly based on the extra 
Patagonian Argentinian record. The Gondwanan Episode is so termed because it is exclusively represe ented by endemic 
mammals of Gondwanan origin, i.e., Mesozoic lineages. In contrast, the South American Episode is almost exclusively 


distinguishe d by ende mic The ‘rian mammals whose ancestors e migrate d f from the Laurasian North ih rie an C ontine nt. 


These two major, successive, and for some time superimposed episodes are the result of the geologic history of the South 
| [ | 8 


American Plate. (1) This plate was part of the Gondwana Supercontinent until about 120 millions of years before Present 


(mybP). when it began to separate and drifted westward (but always south and relatively near the North American Plate). Until 


about 30 mybP it was connected to Western Antarctica, and through it to most of the Eastern Gondwanan continents. (2) By 
Don ly 125 mybP, Donnelly \f i j 
ectonic features around the Caribbean began, i.e., subduction and island ares or continental margin magmatism. Thes 


s “Flood Basalt” was initiated. After its cessation by 85 mybP a variety ol romps 


us probably permitted the first known inter- American exchange of tetrapods, dinosaurs such as Hadrosa s n the 
rising North American continent and Titanosauridae Saltasaurinae from the South American continent. Probably by the latest 
Cretaceous these geologic features also permitted the immigration of the first Therian, which gave rise to the native South 
American land mammals. (3) The geologic definition of southern Central America is the last and most important phenomenon 
related to the final connection of both Americas. By 12 mybP the submarine connection of Central America with South 
America besan. and subsequent volcanic island ares permitted the beginning of The Great a rican Biotic Interchange. 


According to the Patagonian record, the absence of tcibesphe nic mammals and the total extinction of the endemic non- and 
pre-tribosphenic mammals are the most outstanding events characterizing the Ge mdwanan Episode. Up to the beginning of the 
Soulh American E ee (e ab Paleocene), few Gondwanan lineages of mammals (a native Gondwanatheria and an endemic 
Dryole sstoid) survived in the South American continent: another related Gonder taxon survived up to the Late Eocene 


— 


estern id tica. These Gondwanan survivors lived together with the first, but advanced, immigrant Therian that 
initiated the South American Episode. Although not conclusively 1 there are some suggestions that this 
superposition began earlier, probably by the pre-Campanian Cretaceous. The history of endemic Therian mammals 
characterizing the South American Episode began to be known in 1948 thanks to the efforts of the Ameghinos and G. G. 
i iese represented the only known higher taxa until 1985 when Bonaparte recorded the first non-therian Gondwanan 


E 
7 


E 1 Late Cretaceous (C ampanian) Patagonian beds. Successive authors contributed to this history, ratifying Simpson’s 
rr ment that the South American Cenozoic mammal evolution was episodic. Thus, the most characteristic geobiotic features 


‘the two major episodes are based almost exclusively on the most eloquent Campanian-Quaternary Argentinian record. 
Ke ywords: South America, land mammals, evolution, geography, Gondwanan Episode, South American Episode. 
RESUMEN 


De acuerdo a registros sin iguales del Cretácico-Cenozoico tardío meridional. la historia de los mamíferos terrestre 
suramericanos se puede dividir en dos episodios importantes que llamamos el Episodio de Gondwana y el Epis odio 
Suramericano. El primer episodio se distingue por el registro argentino patagónico del norte y centro, mientras que el segundo 
S 


e basa sobre todo en el registro extra argentino patagónico. El Episodio de Gondwana es así llamado porque está represe EN 


! extend special thanks to Peter Raven for his invitation to be one of the speakers of this symposium. To be selected among 
such prominent scientists is a greal honor, especially by so distinguished an investigator as the Director of the well-known 
Missouri Botanical Garden. | would like to address a very special ackno 1 to Alan Graham, as well as to Mick 
Richardson and Victoria Hollowell, whose assistance was not only uncommon but also decisive in making our stay a real 
pleasure. Without exception. we are also indebted to all the organizers of this pleasant and successful symposium. Special 
thanks to Olga M. Montiel. Most of the figures were masterfully designed by Agustin Viñas; others by Laura Zampatti, Carlos 


Vildoso Morales, and Marcela Tome As 0 ways, Laura snp contributed to the composition of the paper, suggesting 
invaluable syntactic ideas. Constru a cl nía suggestions were made by the referees, Richard Cifelli and Guillermo 
Rougier, as well by Victoria C. Hollowe ll Sei le entific Editor) i Sophia R. Balcomb (editor); all their suggestions and 


corrections, without any si stion, improved this paper. lo is obvious that English is not my native language. Hence, I left quite 
a Job for my friend R. Cifelli and the editors!! To all of them, I am very grateful. 
Departamento eg Vertebrados: Museo de La Plata, Paseo del Bosque s/n, 1900 La Plata, Argentina. 


5 


ANN. Missouni Bor. GARD. 93: 209-230. PUBLISHED ON 23 AuGusT 2006. 


210 Annals of the 
Missouri Botanical Garden 


exclusivamente por los mamíferos endémicos de origen gondwánico, i.e. linajes mesozoicos. En contraste, el Episodio 


Suramericano se distingue casi exclusivamente por mamíferos terios endémicos cuyos antepasados emigraron del continente 


laurásico norteamericano. 


Estos dos episodios importantes, sucesivos y por cierto tiempo sobrepuestos son el resultado de la historia geológica de la 
placa suramericana. (1) Esta placa formaba parte del supercontinente de Gondwana hasta cerca de 120 maap, cuando 
comenzó a separarse y alee. hacia el oeste D ro siempre al sur y relativamente cerca de la FM ica norteamericana). Hasta 
hace cerca de 30 maap estuvo conectada a Antártica occidental, y ravés de esta a la mayoría de los continentes 
gondwanicos orientales. (2) Alrededor de 125 maap, empezó la ' a ion basáltica" de Donnelly. Después de su cesación 
hace 85 maap, comenzó una variedad de carac 5 rístic as tectónicas eompresivas alrededor del Caribe, i.e., arcos de islas y 


subduceión o magmatismo marginal continental. Estas características probablemente permitieron el primer intercambio inter- 
americano conocido de te E dinosaurios 85 como Hadrosauridae del continente norteamericano en levantamiento y 
Titanosauridae Saltasaurinae del continente suramericano. Probablemente hacia fines del Cretácico estas características 


geológicas también permitieron la migración del primer terio, que dio origen a los mamíferos terrestres suramericanos. (3) La 
definición geológica del sur de América Central es el último y más i ral fenómeno relacionado con la conexión final de 
ambas Américas. Hace 2. maap comenzó la conexión submarina de América Central y del Sur, y subsecuentes arcos de isla 


volcánicos permitieron el inicio del Gran Intercambio Biotico Americano. 


—. 


Según el registro patagónico, la ausencia de mamíferos tribosfénicos y la extinción total de los mamiferos endémicos no- 
ud nicos son los acontecimientos más exce pe ionales que caracterizan el Episodio de Gondwana. Hasta prine ipio del 
Episodio Suramericano (Paleocene temprano), pocos linajes gondwánicos de mamiferos (un Gondwanaterio nativo y un 

W D 


Dryolestoide endémico) sobrevivieron en el continente suramericano; otro taxón Gondwanaterio relacionado sobrevivió hasta 
el Eoceno tardío en Antartica occidental. Estos sobrevivientes de Gondwana vivieron juntos con el primer, pero avanzado. 


inmigrante terio que inició el Episodio Suramericano. Aunque no se ha de mostrado en forma cone vene hay algunas 


sugerencias de que esta superposición comenzó ra 5; 5 mente por el Cretácico pre-Campaniano. La historia de los 
mamíferos te Ta “micos que caracterizan el Ej ricano se empezó a conocer en 1948 gracias a los esfuerzos de 
los Ameghinos y G. G. Simpson. Estos represe 11 js únicos taxones avanzados conocidos dud 1985 cuando Ap eid 


registró los ee linajes no terios de Gondwana en depósitos ds del Cretácico tardío (Campaniano). Autor 


posteriores contribuyeron a esta historia. ratificando las afirmaciones de Simpson que la evolución de los is 


7 
= 

7 
in 
7 
E 
= 
F 


suramericanos CChnozolcos fue episódica. de los dos episodios más 


importante S se basan casi exe Jusivame ante ene | mas loc tuente registro arge nino de | C ;ampaniano- Cuale rnario. 


INTRODUCTION AND. METHODS The history of South American mammals is 
me. Within the 


The purpose of this paper is to analyze the first framework of plate tectonics, varying paleogeographic 


something like two histories in 


order geologic-geographic phenomena, and the States divided the history of land mammals recorded 


n the southern South American continent into two 


related climatic-environmental changes, that gov- 
erned the evolutionary history of land mammals distinct episodes: the Gondwanan Episode and the 
P: 


recorded to date in the present Patagonian region of South American Episode. In fact. the tribosphenic 


E: 


the South American continent. This is not a mere mammals recorded in Late Jurassic beds of 
analysis of the systematic changes that successively present South American continent (Fig. 2). as well 


occurred in. the mammal communities during the as an advanced pretribosphenic mammal recorded in 


short time that South America was an island conti- Early Cretaceous beds (Fig. 3), both in Patagonia, 
nent (a short time because, according to Zachos et al. correspond to two somewhat distinct geographical 


(2001). the island continent began to be totally situations when South America was an emerging 
isolated from ca. 30 mybP (Fig. I). not from the continent still not well separated from the African 
beginning of the Cenozoic as Simpson (e.g., 1980) and continent. Thus, the Gondwanan and South American 
followers thought). While the history of South Episodes are the only two presently well-established 
American land mammals has been extensively major episodes in the history of South American 
treated, virtually all such treatments were published mammals. With that in mind, I will not detail the 
before recognition of the recently discovered Gond- trends and features of the successive mammal 


wanan mammals (Campanian and the last earliest communities as, for example, was done following the 


Paleocene representatives). Thus, this paper empha- previously recognized faunistic cycles (Pascual et al., 
sizes the distinct history of the Gondwanan mammal 1996). For example, within the South American 


communities with respect to the relatively well-known Episode | will only point out the main events that 
Cenozoic and Recent South American. mammal characterize what Stehli and Webb (1985) distin- 


communities. guished as The Great American Biotic Interchange, 


Volume 93, Number 2 
2006 


Pascual 
South American Land Mammals 


211 


Figure 
continents 
Land masses 30 mybP are outlined in black. 


which has been aptly developed by Webb (2006 this 
issue) and, independently, but complementarily, by 
Coates (2003). 

As Simpson (1950, 1980) correctly indicated, the 
history of South American mammals was episodic and 
divisible into significant intervals of compositional 
changes, and thus provides the biochronological 
framework in which to understand their evolution and 
the evolution of their environments. Because of this, 
most North and South American paleomammalogists 
use the Land Mammal Age (LMA)—in our case, the 
South American Land Mammal Age (SALMA) 


basic biochronological unit (but see Cione & Tonni, 


as the 


1995). "Mammal Ages were intended to represent 
divisions of the Cenozoic, based on characteristic 
eroups of fossil mammals whose temporal relationships 
and overall stage of evolution were thought to be 
indicative of a particular interval of geologic time... 
attention was given to first and last occurrences, index 
fossils’ were noted, and characterizing assemblages 
were listed” (Woodburne, 1987: 1). 


|. The geographic situation of South America and Antarctica 30 mybP. According to Zachos et al. 
were totally separated by this time, thus beginning the transcendental isolation of the South American continent. 


2001), these 


hierarchical 


The of the SALMAs 


reveals clusters that represent evolutionary episodes 


arrangement 


of mammal communities, not only taxonomically, but, 
more importantly, according to ecological modifica- 
tions. Ortiz-Jaureguizar (1986) performed a multivari- 
ate analysis of similarity among SALMAs, a method 
that was subsequently revisited and updated by 
Pascual & Ortiz-Jaureguizar (1990) and Pascual et 
(1996). As Ortiz-Jaureguizar (1986) did, these 
latter authors used SALMAs as operational taxonomic 


al. 


units (OTUs) and families of the Late Cretaceous and 
Cenozoic Land-Mammal Ages as “characters.” They 
were able t 
SALMA 


Units; from lower to higher these were Faunistic 


| recognize four hierarchical ranks of 
organization, distinguished as Faunistic 
Subeycles, Faunistic Cycles, Faunistic Supercyles, 
and Faunistic Megacycles (Table 1). Major changes 
observed among mammal communities (Pascual & 
Ortiz-Jaureguizar, 1990; Pascual et al., 1996; Pascual, 
2001) appear to be concomitantly related to major 


environmental changes, apparently correlated with 


212 Annals of the 
Missouri Botanical Garden 


D P | : | 


Cafiadón Asfalto Fn 
(Asfaltomylus patagonicus) 
ncia Laguna Manantiales 

rae pie iu patagonicus) 


— ae p umm 

6. still 
This apparently explains the persistence. of 
the Patagonian Late Cretaceous 
2002 is a Late Jurassic 


The paleogeographic situation of the South American contine : x the end of the Jurassic, 155 mybP, i 


Figure. 2. 
nnected to the remaining Gondwanan continents (fide Smith et a 
E ule rian mammals in what was going to be the Patagonian territory, ind thei ‘ir ee nce during t 
Gondwanan Episode. Asfaltomylus e Rauhut, Martin, Ortiz-Jaureguizar & Puerta, 
tribosphe nic 1 find in Patagonia (Chubut). and Ameghinichnus patagonicus Casamiquela, 1961 is a Late Jurassic 
ichnita of a supposed Pantotheria, also from Patagonia (Santa Cruz). Grey shading ne nts land masses 155 mybP. 
(Modifie d from Smith et al., 1994; reprinted with the permission of Cambridge University Press 
worldwide diastrophie phenomena. We use diastroph- nological classification to Late Cretaceous mammal 
ost Comprehensive sense communities from Patagonia known to date (Scillato- 
Yané & Pascual, 1984, 1985; Bonaparte et al., 1993), 
earliest mammal- 


= 


ie phenomena in its 
including the fragmentation and drifting of continents. 
We recently (Pascual & Ortiz- 
1990; Pascual & 


including the recently discovered. 


bearing Paleocene formation that documents the 


sea level changes, ete. 
Jaureguizar, 1992; Pascual et al., asl 


Ortiz-Jaureguizar, in press) also applied this biochro- known Gondwanan survivors. To avoid repetition, and 


arga Fn 
(Vincelestes 5 


Figure 3. The primordial South American and Mrican continents had somewhat drifted apa but were still tenuously 
connected by 135 mybP. This the Early Cretaceous Vinee euque northwestern 
marga Form: on, is not part of the Gondwanan Episode. 
994: reprinted with the permission of Cambridge University Pre 


Grey shading re Dor nts land masses 


8.) 


explains why 


Patagonian bed of La 
135 mybP. (Modified Com Smith et al.. 1 


Table l. Land-mammal Ages: d 


Compiled by 
PALMER and 
GEISSHAN 


d 2 2 kee 
Bl 1:11 


Yuu 2 
¡CLIC v 


LE 


eres 
HEIRS EIN 


(199 


^ 


I 


ates. faunistic units, and the main climatic and environmental indicators. 


EP AGES FAUNISTIC UNITS CLIMATIC AND ENVIRONMENTAL INDICATORS 
HAQ etal. | BERGGREN] North AMERICA | South AMERICA | FAUNISTIC | FAUNISTIC | FAUNISTIC FAUNISTIC | EUSTATIC CURVES | TERTIARY | MARINE TEMPERATURE | MAIN SOUTH AMERICAN BIOTIC 
(1987) | eta. [BERGGREN et a. modifie SUBCYCLES| CYCLES |SUPERCYCLES| MEGACYCLES | (HAQ etal., 1987) |  ANDINE PC) (SAVIN et 21,1875) | AND CLIMATIC EVENTS 
5 1985) MARSHALL DIASTROPHISM 
o) (1985) | (1985) 
be 9 Ww 29 3 
“RECENT | RECENT | RECENT | RECENT '| POSTPAMPIAN Iu up m) 
: PLEISTOCENE | PLEISTOCENE É es NADAN] PAMPIA 1 ic iiia Extinction 
TE |PLIOCENE |PLIOCENE| BLANCAN | a = F DIAGUITA—] t record “legionarios” of North 
x Aiba es origin 
ARAUCANIAN á 5 Climax of high-crowned “ungulates” cursorial 
HEMPHILLIAN | HUAYQUERIA 8 gigantic rodents and large piter edentates 
a AN 2 6 Major expansion of southern plai 
E MAYOAN id 2 : Beginning of Patagonian aridity 
w CLARENTONIAN| CHASICOAN $ o Last record of Primates in high latitudes 
u doc ersten sued z T o 
E z | LAVENTAN Jmorwuceus| £ 3 S 8 
8 di o BARSTOVIAN | FRIAS! 8 2 E QUECHUA — 5 
= A 2 a ha i 9 First emigration (caviomorph rodonts and 
o SANTACRUCIAN ST c O — tardigrade edentates) to West Indies 
e PANSARTACRUCIAN 8 us 2s 2 10 e 5 and grasslands 
* u = bo 2 ( nahs) 
? z = SE a 7 Kash 5 40 "caviomorphs" and Platyrrhini 
- z ne 2 uj | 12 First record of euhypsodont “ungulates” 
8 oa o 
8 ES 8 
ARI KAREECEAN v] LX [4 
w & ge 3 
z DESEADAN | DESEADAN a € o a 
9 PI E 
8 E z zt} og 
m a L pru — a 
8 8 8 R " PEHUENCHE 9 13 Record of fossil mammals in Antarctic 
1 m Peninsula indicates to warm climates 
8 5 eee 14 Climax of Pleslapid- lik [ 
PIT Polydolopimorphia 
CHADRONIAN 
L DIVISADERAN | IVISADERAN 15 The grassland environments begin to be 
Js — 4 ——— prominent 
16 First record of protohypsodont "ungulates" 
z 17 N differences 8 lower 
2 (forested) and higher fepe latitude habitats 
ul DUCHESNEAN o 8 Increase of the diversity of xenarthran 
= M E o 
8 < e 
g E x g g 
5 z — 9 2 21 Unique association of “northerners” with 
2 5 UINTAN z a a 8 Gondwanan mammals 
— o a < 2 22 Record of the first non · Aus trallan monotreme 
z o ow o (ornithorhynchid) and last occurrence of 
E x z g iondwanan pre-nontribosphenic mammals 
= ~ z with endemic South American marsupials 
BRIDGERIAN CASAMAYORAN 3 m: —d.,noe . O A ES 
2 
T RIOCHICAN & 19 Warm to temperate climates and 
Ze wees predominance of [oreste uae 
B 2 LARKFORK! TABORAI 20 Highest diversity of marsu; 
3 o 37. A 
à — o TIFFANIAN PELIGRAN 
4 úl ú 3 PANRIOCHICAN | EOPATAGONIAN See ziand 22 
a E x 
o 
a 
1 „55 MEIST 
ES PUERCAN TIUPAMPAN | COCHABAMBAN | PROTOCENOZOIC | a 
o 
CIA ? ? ? ? E 
z 
- tri i 
ALAMITAN | SOMUNCURAN |CUADRADAN | VERDIAN $: O aa Sondwanán 
JUDITHIAN š: 
2: 
COE MINE ? ? ? ? $ : 24 No record of mammals 
o. — 


9002 


z JequinN 'e6 SUNJOA 


s[euue|N pueg ueoueuly ynos 


¡enosed 


Elz 


Annals of the 
Missouri Botanical Garden 


to provide a synopsis of essential features, I herein 
summarize the main evolutionary events that mark the 
history of South American mammals according to the 
higher Faunistic Cycles (Table I), although in some 
cases transcendental events punctuate these inclusive 


Faunistic Cycles and Subeyeles. 


ANTECEDENTS 


three of the most outstanding 
South 
extant land mammals, Carlos and Florentino Ameghi- 
1948: 19-26) and G. G. Simpson 


(1980), were aware of the recently incorporated South 


None of the 


connaisseurs of the American extinct and 


nos (see Simpson, 
American Mesozoic mammals. However, F. Ameghino 
and Simpson, based on distinct, supposedly positive, 
knew 


dinosaur-like 


them. For 
teeth 
associated with mammals of Casamayoran age (early 
1906). F. 


teeth as 


evidence, did consider that they 


example, C. Ameghinos found 


Eocene) in central Patagonia (/ Ameghino, 


Ameghino interpreted these zyphodont 


pertaining to carnivorous dinosaur (Ameghino, 
1906). Simpson (1932, 1933, and particularly, 


1937a, 1937b), using relatively complete, and differ- 
ent, remains, demonstrated that these teeth pertained 
to a curious and unique terrestrial crocodilian species, 
which he named Sebecus icaeorhinus Simpson and 
included in a new peculiar infraorder he named 
S 1937a). In 1946, E. H. 


request to finish his study 


— 


vecosuchia (Simpson, 
Colbert accepted Simpsons 
and published a monographie study on the best 
specimens collected by Simpson and his staff of the 
f New York 
Colbert corroborated both the eroc- 
that it 


straligraphically associated with mammals of Eocene 


American Museum of Natural History o 
(Colbert, 1946). 


odilian nature o 


this specimen and was 


age (Casamayoran SALMA). However, even at the 


time he studied those erocodilian remains. he 


considered the relationships of this new erocodile to 


be questionable (Colbert. 1946; see also the conclu- 


— 


sions of Gasparini, 1984). 
Of course Simpson’s papers, and Colbert’s conclu- 
Ameghino, who died 


sive paper, were unknown to F. 


in 1911 thinking that supposedly carnivorous dino- 
saurs lived in South America during the Cretaceous 
and were associated with more advanced mammals 


than were known from anywhere else in the world. 


^ 


Consequently, he regarded these mammal-bearing 
beds with supposedly carnivorous dinosaur teeth as 
Cretaceous in age, whereas Simpson later considered 
be Eocene in age nnen 1937a). 
«asamayoran SALMA (Table 1). 


The transcending point is that, on this basis, F. 


these beds 


representing the ( 


Ameghino built a chronological scale that was too old 


and which he used as standard not only for South 


American land mammal-bearing beds but also for all 
1906). 


This was his péché originale, which so negatively 


related. lithostratigraphical units (Ameghino, 


— 


influenced. his otherwise extraordinary work. This 


transcendent error produced an almost universal 
reaction against him, especially when he suggestively 
named monkey (Ceboidea) remains, found in what he 
took as early Eocene beds of Patagonia, Homunculus 


1891). 


.e€s un mono de caracteres muy 


patagonicus Ameghino (Ameghino. About this 


species he wrote 7 
elevados, y lo considero como formando parte de la 
linea que conduce al hombre y a los antropomorfos” 


1891: 29 


Simpson (1980), in turn, accepted the Cretaceous 


Ameghino, 


age of the mammal-bearing Umayo Formation discov- 
ered by the French paleontologist B. Sigé (Grambast 
1967; Sigé, 1972) in 


Andes. These French paleontologists correlated. this 


et al., the southern Peruvian 
formation. with the marine Vilquechico Formation, 


exposed further tọ the north and undoubtedly 


Cretaceous in age, and wrongly used the same name. 
The mammals collected in the Umayo Formation were 
early Paleocene, nol 
1979). What Simpson 


(1980: 39) expressed about these Peruvian mammals 


later demonstrated to be 


Cretaceous, in age (Crochet, 


Is quite instructive: 7... it would have been expected 
that like 


Cretaceous of South America." 


mammals these would occur in the 
Simpson did not admit 
that non-tribosphenic and pre-tribosphenic mammals 
were inhabitants, at least, during the Late Cretaceous 
of the South American continent, and by extension— 
considering his non-acceptance of the Plate Tectonic 
rationale—that what we presently accept as the 
Gondwana supercontinent had no such pre-Cenozoic 
mammals as inhabitants. 

be remarked that what was known by 
1980) as the “ 


American. mammals was based only on the 


lt has 
Simpson (Simpson, curious history” of 
South 
endemic Cenozoic mammals, the majority of which 
descended from Laurasian mammal immigrants from 
the primordial North American continent. No Gond- 
1980s, 
Bonaparte and staff (Bonaparte et al., 1984; 
& Soria, 1985: Bonaparte, 1986a, 1986b, 1988, 1990, 
1996) found the first unquestionable Late Cretaceous 


wanan mammal was known until. the when 


Jonaparte 


mammals, and simultaneously Pascual and staff (e. g. 
Scillato-Yané & Pascual, 1984, 1985) found the last 
relictual representatives, also in Patagonia but in beds 
well 


of early Paleocene age. None of the relatively 


known Cenozoic and living mammals show features 


indicating a Gondwanan ancestry. 
The Gondwanan mammals completely disappeared 


by the early Paleocene, at least in Southern South 


America, in contrast to what happened, for example. 


with the Monotremata in Australia. Probable Gond- 


Volume 93, Number 2 
2006 


Pascual 215 


South American Land Mammals 


os Alamitos region 
Cono patagonicum) 


a Colonia region 
(Reigitherium bunodontum) 
Figure Bet 


and by its end the fir 


ween 85 and 63 mybP a seaway bisected the 


st Therian immigrants from North Ame rica inaugurated the South American Episode. Los 


American continent into 


South two 
provinces 1 Broin & de la Fuente, 1993). It is likely that most of this interval was the climax of the Dou iude. Es de. 
Alamitos region 


and La Colonia region are the two coetaneous Campanian Patagonian localities where the two indicated Dondwanan laxa 


were found. 


wanan survival candidates are the Xenarthra, despite 
the fact that in South 
American Late Cretaceous or earliest Paleocene beds. 
1950: 363) 


“episodic history” 


none have yet been found 


Consequently, Simpson's (Simpson, 
interpretation of what he termed * 
of South 


applicable to the Cenozoic mammals o 


American mammals was correct, but only 


— 


Which he was 
knowledgeable. 
DISCUSSION 


South 


exclusively 


The best current Mesozoic record of 


American land mammals is almost 
Patagonian and Late Cretaceous (Campanian) in age 
(Fig. 4). Compared to the Cenozoic record from the 
entire continent, the Late Cretacaeous evidence 


demonstrates that the entire South American land 


mammal history is radically divided into two major 
and unrelated episodes occurring under two distinct 
geographic sceneries. These two distinctive episodes 
were separated by a relatively long Maastrichtian— 
early Paleocene hiatus; there is only one dubious early 
Paleocene Patagonian specimen, a tribosphenic molar 
2005). This molar 


may well represent one of the oldest therians to have 


found quite recently (Goin et al., 


emigrated from Laurasia, embodying the beginning of 


The South American Episode. Additionally, an 
assemblage from Bolivia, of middle Maastrichtian 


age based on interbedded marine taxa, includes an 


undoubtedly eutherian mammal, based on a molarized 
premolar (non vidi, fide Gayet et al., 2001). This may 
well represent the Ancient Immigrants from North 
America that inaugurated the South American 
Episode. We originally recognized those two major 
episodes as stages, plus a third, the Late Pleistocene 
“Neotropical Stage” (Pascual, 1996, 1998; Vizcaino et 
al.. 1998: 202). However, 


major and well distinguishable episodes : 


herein I rec ognize i edi two 


S 


a 


Gondwanan Episode and The South American Epi- 


sode. 


THE GONDWANAN EPISODE 


To date, the Gondwanan Episode is best represent- 
ed by mammalian remains from two middle Patago- 
both Late Cretaceous in age (Los 
Chubut; Fig. 4). Both 


-atagonian Late Cretaceous mammal assemblages 


nian. localities, 


Alamitos and La Colonia, 


(apparently Campanian) are comprised only of non- 
tribosphenic and pre-tribosphenic mammals. 

Based on the Patagonian records, and the Bolivian 
one (Patcha Pata) mentioned above (if the supposed 
origin), The 


eutherian is really of Gondwanan 


Gondwanan Episode extended from the latest Creta- 


— 


ceous (Campanian) to earliest Paleocene (Danian) 
interval, and thus far is represented only in these two 
central Patagonian localities, and probably Bolivia. 
and land mammal 


Similar stratigraphic sequences 


Annals of the 
Missouri Botanical Garden 


Figure 5.1. A comparison of right dentaries in (a) labial. 
Gondwanatheria Sudamerica ameghinoi Scillato Yané 
Dinomyidae Fetrastylus Ameghino, 1886 sp. Scale bars 


Gondwanatheriar 


« Pascua 
lem. 
(A) Ferugliotherium windhausent Bonaparte, ( 


(b) lingual, and (e) occlusal views of (A) the early Paleocene 
xt with (B) the Late 


V comparison of the 


Miocene ¢ 


NE 


O 
Figure 5. 


(B) OS um patagonicum Bonaparte, and (C) 


Sudamerica 9 with the late Miocene caviomorph rodent Dinomyidae (D) Pentastilomys Kraglievich, 1926. Scale bars: 


| mm. 


genera present at both Patagonian localities indicate 


synchronicity, representing continental facies of 


a wide-ranging epeirogenic marine Campanian-Da- 


nian transgression (Bonaparte, 1987a, 1987b. 1996; 
Pascual et al. 2000a: Fig. 4). This transgression 
D o 


extended along the length of the eastern region of the 


Andes, as far as the Bolivian Lake Titicaca. During 


the Campanian a coetaneous “Caribbean” transgres- 


sion connected to the former by the Titicaca region. 


which, as a seaway, divided the South American 


territory into two Late Cretaceous—early Paleocene 


lerre sstrial | id ographic al pas inc es Cs prov inc [e e nord- 
d 
d 


and “province suc 
o verified by various 
abiotic and biotic evidences (Wilson € Arens. 2001: 
Fig. 4). 

The 


ormative and eloquently representative base of the 


eondwanienne" 


DULCI ATIC THe. 


sensu Broin & de la Fuente, 


Patagonian Late Cretaceous record (the most 


— 


in 


Gondwanan Episode) shows the total absence of 
| 


tribosphenic mammals. This absence and the ad- 


vanced morphology and endemism of non-tribosphe- 
nic and pre-tribosphenic mammals illustrate the most 
mammalian evolutionary that 


important processes 


occurred in. Patagonia during the Gondwanan Epi- 
sode. For example. among the non-tribosphenic taxa 


were the quite derived Gondwanatheria. together with 


dubious Laurasian Triconodonta. Symmetrodonta, and 
Docodonta. plus an array of quite advanced Dryoles- 


loidea (Figs. 5. 6). The record of the Bolivian 


Maastrichtian Eutherian (see above) may indicate 


that this is not a valid generalization for the whole 


continent, and that Patagonia—as insinuated by some 
e P 


other fossil vertebrates (see below)—may have had 
a distinct biogeographic history. at least from the 


second half of Mesozoic to the Present. In any 


case, this does not refute the existence of the two 
distinct major episodes that characterized the South 
American land mammal history. 
Bonaparte (1988) discovered an Early Cretaceous 
mammal, Vincelestes neuquenianus Bonaparte, from 
another north western Patagonian locality (Fig. 3). 
which he considered to be Eupantotherian. To 


(1987). Butler (1990). 


Rougier el al. (1992), the level of organization of its 


Bonaparte and Rougier and 


molars appeared to be“... between the Late Jurassic 
Peramus Owen and the Early Cretaceous Aegialodon 
& Musset” (Butler, 1990: 542). 


States of V. neuquenianus 


Kermack, Lees 


Therefore, the character 


placed it in an intermediate state between “non- 
therian” mammals and Therian taxa. Thus. Rougier 


(1993) grouped this advanced taxon with therians in 


a clade he named Prototribosphenida. Two additional 


Volume 93, Number 2 
006 


Pascual 
South American Land Mammals 


Figure 


ha "Bonde STUS DELE (AY ingual. (B) 
poe henna 5 
sepulvedat oe (Triconodonta?): (C) labiz 

views of a rig wer molar. 6.3. I 
view: (B) Leonards Pb e Bonaparte: (a) 


(a) labial. (b) occlusal, and (c) li 


) labial. (b) 


mesial views Ns Casc miquelia rionegrina Bonaparte: (a) 


(A) labial and (B) po views ae a 
land (D) lingual views of : 
55 e Upper molars of 
lingual, and (c) 
igual views: (D) Paraungulatum rectangularis ie 


mes ial 


occ bn P ower el ar. > 
"lustro 
right uppe °) lingual 
(A) Westm DI Bonaparte & So 
occlusal views; (C) Brandonia „ 1 
(a) distal, (b) « 


occlusal views. P 


clusal; and ( 


Doc don? 


(b) distal. and (c) 


Reigitherium i Bonaparte. Fragment of a left dentary: (A) labiz a 10 lingual. and (D) ocelusal views of premolar 4— 


molar 2, (C) posterior view behind molar 2 


mammals, from two more distant Late Jurassic 
localities, also in Patagonia (Casamiquela, 1901: 
Rauhut et al. 2002; Fig. 2). as well as two Late 


mammal-bearing Brazilian 


2001: 


Triassic-Early Jurassic 


localities (Ribeiro et al., Bonaparte et al., 


2003), suggest that there were at least two more 
Mesozoic minor “episodes,” though they are not well 


known to date. Supporting those Early Cretaceous and 


Late Jurassic minor “episodes,” the paleogeographic 


evidence shows that these mammals evolved in 
different geographic situations. During the Late 


Triassic, what was going to be the South American 
continent was connected, on the one hand, to what was 
going to be the North American continent, and, on the 


other hand. to what was going to be the Gondwanan 


. and (E) distal view of molar 2 


1994, fig. 29). By 


be the South 


Supercontinent (fide Smith et al., 
the Late Jurassic what was going 


American continent was already separated from what 


was going lo be the North American continent, but 
still connected to Africa (Fig. 2) as part of the 


Gondwanan Supercontinent. By the Early Cretaceous, 
South 


separating, with the northern sector somewhat drifted 


southern America and Africa were already 


but still tenuously connected (Fig. 3: see also Storey, 
1995, fig. 2c; compare these two figures with Fig. 1). 
The extinction of most of the Gondwanan mammals, as 
well as the 
Paleocene laxa, must have occurred while the South 


the extinction. of relictual earliest 


American continent lay in the geographical position 


depicted Figure 4, i.e., south of and near to the 


218 Annals of the 
Missouri Botanical Garden 
North American continent, but still connected to the with some archipelagie volcanic islands between each 


Australian continent. 

The seaway mentioned above existed between ca. 80 
and 64 mybP and bisected the pre-South American 
continent into. the two biogeographic provinces, as 


11-214) 


"province 


recognized by Broin and de la Fuente (1993: 2 
on the basis of distinct chelonian faunae: 
nord-gondwanienne” dominated by the Pelomesudae 
“province sud-gondwanienne” 


Notably, 


represented in the Atlantie border of equatorial Africa 


and dominated by the 


Chelidae. the former were, and are. well 


while the latter were, and are, common inhabitants of 


the Australian. continent. The dipnoan fishes offer 
another, similar testimony: Lepidosirenidae (Lepidosi- 
ren Fitzinger in the province nord-gondwanienne vs. 


Protopterus Owen in intertropical Africa) and Cerato- 


— 


dontidae in the province sud-gondwanienne (plus 
Ceratodus Agassiz in the Late Cretaceous of Patagonia 
vs. Neoceratodus de Castelnau presently in Australia) 
1973: Pascual & Bondesio, 1976: 
1977). 


The Late Cretaceous mammal-bearing sediments 


(see Fernandez et al.. 


Bondesio & Pascual. 


from Patagonia, representing the Gondwanan Episode. 
are continental facies of that southern marine seaway. 
The faunal biota of the province nord-gondwanienne 
closely resembles other Gondwanan equalorial faunal 
biotas, such as that from northern Africa. which. i 

turn, are distinct from the southern South American 
faunal biota that more closely resemble other austral 
biotas, such as those from 


Gondwanan faunal 


Antarctica and Australia (Zinsmeister, 1987; Pascual 
el al., 1992b, 1996). For example, the Lepidosirenidae 
lung-fishes during the Cretaceous were. and still are, 
nord-gondwanienne and 


endemic to the province 


northern Africa. This is also the ease for the Chelidae 
endemic. to the sud- 
Africa. 


Ceratodontidae lung-fishes and the Chelidae turtles 


turtles which are province 
gondwanienne and southern Similarly, the 


were endemic to Patagonia during the Cretaceous and 
(1991) 


discuss many other examples of living Insecta, Fungii, 


are presently endemic to Australia. Crisci et al. 
Gymnosperma, and. Angiosperma, which have similar 
vicariant distributions in southern South America and 
Gondwana. 
South 


American Late Cretaceous—Paleocene province nord- 


the continents that were part of Eastern ( 


The hypothesis about the existence of a 
gondwanienne and province sud-gondwanienne was 
tested and confirmed by a quantitative comparison of 
Late Cretaceous and Paleocene palynofloras, distinct 
biotic elements, from austral and equatorial South 
American representatives (Wilson € Arens, 2001) 

By the time of The 
South 
nected to Antarctica and had drifted to just south of, 
rising North 


— 


Gondwanan Episode, the future 


American continent was still somewhat con- 


and nearer to. the American continent, 


continent (Fig. 4). As a Noah’s Ark (sensu McKenna. 
1972). 


relictual Gondwanan mammals. This is supported by 


the South American Plate had been carrying 


remains from the Patagonian Los Alamitos Formation 
(Campanian—Maastrichtian; Bonaparte 1996), as well 
as those in the correlated La Colonia Formation 
2000b, 2002: 


impossible to know when the massive extinction. of 


(Pascual et al., Figs. 5. 6). It is 


Gondwanan mammals began. The separated South 


American continent. situated south of the 


North 


the Campanian, 


always 


American Plate, drifted north-northwest. unti 
when its mutual connection by the 
Panamanian “Bridge” began (ef. Figs. 2-4). As 
pointed out above, the complete extinction definitely 
occurred while the continent was already situated 
near to and south of the North American continent. nol 
very far from its present position, around 80 mybP. At 
this time, the defined, 
alt 
first interamerican faunal exchange (Fig. 12): hadro- 


North 


continent and titanosaurid saltasaurine dinosaurs 


Caribbean sea Was nol vel 


ough intermediate volcanic islands enabled the 


saurid dinosaurs from the rising American 


FS 


rom 
the South American continent. 

Closely related Gondwanatherians found in Late 
Cretaceous beds of Madagascar and peninsular India 
(Krause et a 


Gondwanan 


1997) also appear to be relictual 
mammals; they likewise survived on 
Noah's Ark-like continents (Fig. 7) that resulted from 


f 


agmentation of the Gondwanan Supercontinent. 


THE SOUTH AMERICAN EPISODE 


The last steps of The First Great Turnover (sensu 
2001) 


a hiatus which spanned the Maastrichtian—earliest 


Pascual et al., must have occurred during 


Paleocene. This involved the extinction of all the 


Gondwanan mammals and their “replacement” by 


Laurasian therians, which emigrated from the North 
American sector. Apparently this marked extinction 
effect of the 


Extinction. 


was another massive Crelaceous— 
Cenozoic This event marks the end of 
The Gondwanan Episode just as the subsequent 


Laurasian therian immigration marks the beginning 
South 


occurred some time during the 


of the American Episode. Both episodes 


latest Cretaceous— 


earliest Paleocene span, as evidenced by records in 


the middle Patagonian Hansen Member—usually 
known as Banco Negro Inferior—of the marine 


Salamanca Formation (early Paleocene), exposed 
Gulf. ca. 20 km north 
of Comodoro Rivadavia (Chubut) (Scillato-Yané & 
Pascual, 1984, 1985: Pascual et al., 1992a, 1992), 
1999: Bonaparte et al., 1993; Gelfo & Pascual, 2001; 
See map in Pascual et al., 10920). 


the coastal region of San Jorge 


Volume 93, Number 2 
2006 


Pascual 
South American Land Mammals 


Figure 7. 
Gondwanatherians, similar to the South 
masses 80 mybP. (Modified from Smith et al., 


early P 


American 


An older Paleocene isolated, tribosphenic molar 
recently found by Goin et al. (Goin et al., 2005) in the 
Danian Lefipan Formation from central-western 
Patagonia appears to be the oldest known Laurasian 
immigrant to South America, even older than the 
Laurasian immigrants found in northern Bolivian beds 
of the Tiupampan SALMA (early Paleocene, although 
not earliest: 1991). Examples of Laur- 
asian immigrants that inaugurated the subsequent 
include the Marsupialia, 


see Muizon, 


South American Episode 
Borhyaenoidea, Borhyaenidae (Mayulestes de Mui- 
zon), Peradectidae (Peradectes Matthew & Granger). 
Caroloameghiniidae (Robertohoffstetteria Marshall, de 
Muizon & Sige), Didelphidae (Andinodelphis Marshall 
& de Muizon, Incadelphys Marshall & de Muizon, 


Mizquedelphy Marshall & de Muizon, Pucadelphis 
Marshall & de Muizon, Carolopaulacoutia (Mckenna 
& Bell)(=Sternbergia Paula Couto). Jaskhadelphys 


Marshall € de Muizon, Chulpasia Crochet & Sige, 
Khasia Marshall & de Muizon, Sillustania Crochet 
and Sige) as well as the Placentalia Leptictida 
(Palaeoryctidae?, ef. Cimolestes Marsh), Condylarthra, 
Muizon & Marshall), 
& 


Pucanodus 


Hyopsodontidae (Andinodus de 
Mioclaenidae (Kollpania Marshall 
Molinodus de Muizon & Marshall, 
Muizon & Marshall, Simoclaenus de Muizon & Cifelli, 
Tiuclaenus de Muizon & Marshall). 


Valen), and Pantodonta (Alcidedor- 


de Muizon, 


de 


Periptichidae 
(Mimamuta Van 
bignya inopinata. de Muizon € Marshall) (but see 
1993, and Gayet et al., 2001). 


The tribosphenic molar from the Lefipan Formation 


Bertini et al., 


appears to be older than the Australian and Laurasian 


aleocene Sudamerica. ame, 
1994; reprinted with the permission of Cambridge University Press 


By aproximately 80 mybP Madagascar and India, as Noah’s Ark-like continents, kept living Late Cretaceous 


inoi. Grey shading 1 d land 


— 


immigrants recorded in the Patagonian Peligran Land 
Mammal Age beds (see Pascual et al., 1992a, 1992b: 
1993; Forasiepi € Martinelli, 2003; 


submitted to press). 


Bonaparte et al., 
Pascual & Ortiz-Jaureguizar, 
These immigrants include the Platypoda, Ornithor- 
Ortiz 


hynchidae (Monotrematum Pascual, Archer, 


Jaureguizar, Prado, Godthelp & Hand; Fig. 8); the 
Marsupialia, Didelphimorphia, Didephidae (Dero- 


rhynchus Paula Couto, Didelphopsis Paula Couto), 
Bona- 


p 


the Polydolopimorpha, | Bonapartheriidae 
partherium Pascual), Polydolopidae gen. nov.; as well 
as the Placentalia, Condylarthra, Mioclaenidae (Escri- 
bania Bonaparte, Van Valen € Kramartz, Raulvaccia 
Bonaparte, Van Valen &. Notopterna 


Notonychopidae (Resquisia Bonaparte and Morales). 


Kramartz), 


In the mammal from the Paleocene 


Patagonian Lefipan Formation plus the taxa from the 
Bolivian Tiupampan and the Patagonian Peligran 


summary, 


appear to be among the oldest Laurasian immigrants, 
i.e., Simpson’s (1950) Ancient Immigrants. These taxa 
evolved on a proto-South American substratum that 
included at least the Antarctic Peninsula (Fig. 9) up 
to about 30 mybP (Fig. 9). At the beginning of the 
South American Episode (early Paleocene) two 
Gondwanan relictual taxa, a gondwanatherian and 
a dryolestoid, as well as the first and unique 
immigrant of an Australian ornithorhynchid mono- 
treme, persisted in central Patagonia (Fig. 8). Another 
relictual Gondwanan mammal was an Eocene gonwa- 
edd) found in the Antarctic Peninsula (Reguero 
. The Patagonian mammals of the Peligran 


P: the 


et al., 


SU pons aleocene) suggest that basic 


220 Annals of the 
Missouri Botanical Garden 


e Didelphimorphs 


didelphimorphs 


5j 


sudameric ids? 


$ . F 1 8 
R p microbiotheriids 


, 
Sudamerica ameghinoi 


or ines i 
d. 


; e e 0 
Microbiothertids Ornithorhynchids tn 


Figure 8. Inferred geographic relationships of the primordial South American continent between 85 and 63 mybP. The 
South American Episode began during the Late Cretaceous when the South American continent was still connected to 
ntarelica, 1. 


] „ not vet completely isolated. The first known immigrant tetr "e rom North America was the Campanian 
. Hadrosauridae Arytosaurus Brown. 1910 sp. Probably by the Late Cretaceous. and certainly by the early Paleocene, it 

vas followed by Metatherian and Eutherian mammals, ancestors of the taxa depicted here. However, their immigration ma 
155 begun by the pre-Campanian Late Cretaceous. By the Late Cretaceous, marsupials emigrated to Australia an 
the same route, by the early Paleocene, South America received the first Australian immigrant, the monotreme la hid 
No. sudamericanum Pascual, Archer, Ortiz Jaureguizar, Prado. Godthelp & 11 ind, 1992b. Dark grey bullets and 
arrow indicate the first immigrants from North America to South America. Among them the microbiotheriids were the first 
marsupials to emigrate from South America to Australia, through the Antarctic Sonne Thylacini 


— 
= 


y ds are one of the end 
products differentiated in Australia from South American marsupials (Sparassodonta?). Sudamerica ameghinoi Scillato Yané 
& Pascual, 1984 is the first and last Gondwanatheria found in Late Cretaceous and carly Paleocene | 


eds from Patagonia. 
Ornithorhynchids are the first monotreme found outside (Late Cretaceous—earl y | 


Paleocene of Patagonia) of. Australasia. 


diversification of the immigrant therians was well on primates and rodents (apparently from Africa: 
its way. IH those therian mammals recorded by Bertini big. 10). This interval corresponds to what we have 
et al. (1993) and Gayet et al. (2001) really were of distinguished as the “most autochthonous part of the 
Late Cretaceous age and immigrants from Laurasia history“ (Pascual et al.. 1985: 230-231: Figs. 9, 10). 
(North America), then they are the oldest known. during which the basic diversification of the South 
immigrants, Le. Simpson's (1950) Ancient Immi- — American. mammals occurred. This event also com- 


grants. prised the Protocenozoic and Infracenozoie Super- 


In accordance with the main turnovers withstood by cycles, which probably extended back to include the 
the Cenozoic land-mammal communities, otherwise last part of the long Patagonian sterile Maastrichtian— 
related to first order diastrophie and climatic earliest Danian span. 
phenomena, it is clear that four main stages of the (2) The end of the Infracenozoie Superevele led to 
South American episode can be recognized: the beginning of a new evolutionary cycle, marked by 

(1) The span between the extinction of the last the late Eocene-early Oligocene Turnover (also 
Gondwanan land mammals and the very first immi- known as La Grande Coupure or the Terminal Eocene 
gration of the Laurasian Marsupialia and Placentalia Event). The archaic lineages of ungulate mammals 
(in part contemporaneous) as well as the late Eocene— (brachydonts) that had gradually changed from 


early Oligocene immigration to South America of — protohypsodont (Eopatagonian Cycle) to almost eu- 


Volume 93, Number 2 
2006 


Pascual 
South American Land Mammals 


Utaetini 


e, 
Notostylopids — 


Figure 9. 
Therians had 111 55 ntiated into the most advanced and v 


xenarthran dasy 


aried lineages of mammals, 


Leontiniids 


Nf o 


VIÑAS 4 


zy the late Eocene (ca. 30 mybP) the South American continent was Paige d. During the Eocene, the immigrant 


„the marsupial Groebertid, the 


odids, Utaetini, and the notoungulates, Ne and Leontiniids. We recognized this interval as the 


most autochthonous part of the South American mammalian history. 


hypsodont types (sensu Mones, 1979, 1982) (Pre- 
patagonian Cycle), was replaced by a quite modern 
stamp of native ungulates (euhypsodont), which 


characterized most. of the native ungulates of the 
Patagonian Cycle and all subsequent cycles up to the 
end of the Pleistocene (Panaraucanian and Pampam- 
pean Cycles; Table 1). We qualified the beginning of 
the Patagonian Cycle) as “The First 
(Pascual et 


this process (1.€e., 
Major Change Toward a Modernization” 
al., 1985: 232-234). At this point, the history of South 
American mammals passed from “Early Experimen- 
tation to Modern Standardization” (fide Gould, 1983: 
21). This turnover appears to be regionally connected 


to climatic and/or environmental changes, related to 


— 


of geotec tonic i- 


a series phenomena, e.g. the c 
astrophic Inea Phase, and quite possibly also the 
Pehuenche Phase (see Tertiary Andine Diastrophism 
in Table 1), the completion of the Drake Passage, and 
America from 
with all 


the consequent separation of South 
Antarctica as an island continent (Fig. 10), 


the climatic consequences derived from the formation 


— 


the encircling Antarctic Current. 


Thus. the evolutionary processes leading to the 


differentiation of the unique South American land 


mammals, until ca. 30 mybP, also had the Antarctic 
Peninsula, and probably some other portion of the 
Antarctic continent, as substratum. Simpson's (1950) 


Ancient Immigrants developed on this scenery. 


Simpsons two other episodes (Faunal Strata in 
a figurative sense), which he named Old lsland 


Hoppers and Late (Island Hoppers and) Immigrants 
(Simpson, 1950: 364). respectively, developed while 
the continent was isolated, except toward the end 
when the inter-American terrestrial connection began 
with the Bridge (ca. 
2.5 mybP), and through it the terrestrial transit (i. e. 
walking) of what Stehlin and Webb (1985) termed The 
Biotic During the 


preceding this the 


newborn Panamanian Land 


Great American Interchange. 
isolation period interchange 
mammals increased, 


particularly the 
hypsodontism of the herbivore grazers among the 


modernization of the native 


manifested in augmentation of 
native placental mammals, as well as in some peculiar 
marsupials, such as the Argyrolagidae (Proargyrola- 
gus Wolff) and the 


Two main subeycles characterize this 


> Patagoniidae (Patagonia Pascual 
and Carlini). 
interval (Pascual et al., 1985): the Deseadan Subeycle 


and the Pan-Santacrucian Subeycle. 


222 


Annals of the 
Missouri Botanical Garden 


INFRACENOZOIC SUPERCYCLE 
OLD ISLAND HOPPERS 


Figure 10. 


aviomorph ancestors 


| Platyrrhini ancestors 


EF 


VIGAS A. 


Approximately coincident with the separation of the South American and Antarctic continents (ca. 30 mybP), 


rodents inm ericetids) and primates immigrated from Africa by sea. 


During the Deseadan Subeyele 
earliest Miocene), according to the Patagonian record, 
over the southern extreme of the continent the most 
primitive Paleogene mammals vanished or disap- 
peared and some very important cladogenetic pro- 
cesses came lo an end, especially within very diverse 
lineages. On the other hand, from small contempora- 
neous mammals related to the latter, new cladogenetic 
processes began. For example, certain notoungulates 
evinced the precocious acquisition of high-crowned 
teeth, and caviomorph rodents were recorded for the 
first time at austral latitudes (see Patterson & Pascual, 
1972: Pascual et al., 1985, and literature therein). 
According to Zachos et al. (2001), 


between two glaciations, 


this cycle occurred 
known as Glaciation Oi-l 
Antarctica 


and Glaciation Mi-1, which developed 


and the southernmost part of the South American 
continent. Furthermore, also according to Zachos et 


al. (2001), completion of the Drake I 


in between these two glaciations (Table I). i.e.. 


Passage occurred 


around 30 mybP. 


(late Oligocene— 


The following Pan-Santacrucian Subeycle repre- 
the South 


related to the end of 


sents the southernmost record of any of 
American SALMAs. 
the Mi-! 


conditions occurred. 


Apparently 


Glaciation, a change to relatively mild 
These mild conditions culminat- 
ed about 15 mybP, with the so-called Mid-Miocene 
Climatic Optimum, and correspond to the time that 
preceded the Quechua Phase of the Tertiary Andean 


Webb 


medium to 


Diastrophism. Considering the entire fauna, 
1978) remarked that 


arge-sized mammals strongly infers a broad environ- 


its rich diversity of 


mental impact, Le. an oplimum balance between 


grasslands and woodlands provided by savanna. An 
intense drop in sea level ca. 30 mybP occurred due to 
glaciation, followed by intermittent drops in sea levels 
E with the Plio—Pleistocene  glaciations 


(Table | 


native 


Around 18 mybP the first emigration of 
En 


although not to North America but to the West Indies. 


American land mammals occurred, 


\round that time, caviomorph rodents and mega- 


lonychid xenarthrans (Fig. 11) 
E e L 


and probably also the 


Volume 93, Number 2 
2006 


scual 223 
South American Land Mammals 


NORTHERN 
CENTRAL 


AL AMERICA 
i A 


m^ 


Figure II. 


Heptaxodontid ancestors 


Megalonychids 


y 18 mybP the first dispersal event of South American mammals occurred: Xenarthran Megalonychids and 


Dima ime emigrated (as island-hoppers) to the West Indies 


platyrrhine primates—populated the Antilles, most 
likely as 


a notable evolutionary radiation. 


island-hoppers. There they underwent 
The megalonychid 
xenarthrans, the heptaxodontid rodents, and the 
platyrrhine primates became extinet quite recently, 
within historical times, while the capromyid rodents 
These are regional descendants of those 
South American immigrants, or probable vicariant 
taxa according to MacPhee & Iturralde-Vinent (1994). 
Curiously, both the megalonychid xenarthran and the 
heptaxodontid rodents differentiated into some pecu- 
liar ecological types, including tree-sloth-like forms 
(Matthew & Paula Couto, 1959; Paula Couto, 1967: 
1990) and  cursorial, elasmodont 


heptaxodontid rodents very similar to the Miocene 


Pascual et al., 


South American phoberomyine neoepiblemids and 


eumegamyine dinomyids (Pascual et al. 1990) 
Among the West Indies rodents, all those with 


elasmodont molariform dentition have been lumpec 


into a single family, Heptaxodontidae, although 
Ray, 1964). Neverthe- 
suggestive that the peculiar Jamaican 
Clidomys Anthony, 1920 (see MacPhee et al.. 1983; 


MacPhee, 1984) is very similar to the South American 


sometimes with hesitation (e.g 


less it is 


Neoepiblemidae in dental structure. On the other 


hand, the remaining Heptaxodontidae are more 


similar to the highly diversified Dinomyidae. This 


difference may indicate that there was more than one 


Fig. 11). 


immigratory event of rodents 


— 


The following Pan-Araucanian Cycle. as we called 


(Pascual et al, 1985), was with the diastrophic 


Quechua Phase, known by most geologists as the 
orographic phase (e.g., Vicente, 1972). The elevation 
of the Andes displaced the most propitious environ- 
ments for land mammals to both sides of the cordillera 
1993). On the eastern sector 


extensive plains arose, 


(Ortiz-Jaureguizar et al., 
Andes, 


comprised of pyroclastic mammal-bearing sediments, 


of the rising first 
and then clastic sediments; the middle pampas region 


also had an important component of pyroclastic 
sediments. 

Mammals on opposite sides of the cordillera had 
long been considered as hd ue two distinct and 
& Salinas, 1990). On 


the basis of a multivariate fate Ortiz-Jaureguizar 


successive ages (e.g., Mars 


et al. (1993) suggested that they are correlative and 
proposed that both be maintained under the original 
name of the Friasian SALMA. The Frias Formation, 
both in Patagonia and in extra-Patagonian regions of 
Argentina, was deposited under the influence of the 
withdrawal of an extensive middle Miocene marine 
1996: 291) made the 


interpretation that the distinction of the Friasian land 


transgression. We (Pascual et al., 


mammals with respect to the preceding subtropical 


Santacrucian one was related to the climatic and 


environmental effects of the first subphase of the 
complex Quechua orographie phase, as well as to the 


global climatic changes, as evidenced by the 


Annals of the 
Missouri Botanical Garden 


The rising Panamanian land bridge estab- 
ll P land. bridge tal 
North and 


mybP the interamerican 


Figure 12. 


| definitive connection between South 
—A. By 85 


terrestrial tetrapods began: hadrosaur dinosaurs from North 


lished ë 


\merica. “exchange 


America and titanosaur sallasaurines from South America. 
Apparently this occurred through some kind of land bridge 
(archipelagic) related to Donnelly’s (1973, 1985) Flood 
Basalt Episode. M — Mexican block: Y — Yucatán block: 
Cl Chortis block. (A modifie d from e lly. 1985 
permission of Springer Science and Business Media.) —B. 


5. with 


of 


relatively sudden divergence between high and low 
latitude marine temperatures. According to Savin and 
Stehli (1975), 
dramatically, but low-latitude temperatures remained 


high-latitude temperatures dropped 


constant or even perhaps increased. In a period 


probably not exceeding 2 to 3 million years, minimum 


temperatures appear to have dropped 3 C, or at a rate 
of about ! €. per My. Based on fossil land mammals 
and other tetrapods, the wide-ranging plains that built 


up on the eastern. slopes of the Andes became 


successively more and more arid to the west, with 


more humid portions to the northeast influenced. by 


the humid Atlantic winds. The successive sedimen- 


tologic mammal-bearing units deposited toward the 


8 


northeast part of Patagonia (Chasicoan, Huayquerian, 
Montehermosan, and Chapadmalalan SALMAS) pro- 
in North 


American immigrants corresponding with a change 


gressively contain. a notable increase 


toward finest granulometric sediments (silly—loessoid), 
as well as with the pace that the Panamanian land 
bridge was being built up. 

Contrary to the increasing trend of eupsydontism in 
the native ungulates, the immigrant. ungulates were 
brachydont, or at the most mesodont. The withdrawal of 
the Miocene marine transgression was succeeded. by 
a period of similarly widespread and varied plains that 
we distinguished as the “edad de las planicies australes” 
(Pascual € Bondesio, 1982: 29). In such a way the late 
Miocene—Pleistocene pampas of present times initiated 
with 


their differentiation. Mammal-bearing sediments. 


good ecological indicators, clearly show the north- 


northeast environmental displacement. At the beginning. 
i.e.. during the middle Miocene, the last appearances of 
Pascual 


some Santacrucian mammals are recorded 


al., 1965: 177). In general, and in accordance with the 


Pn 
About 15-16 mybP. the present 1 anian Land Bridge 
was occupied by an abyssal sea more than 2000 meters deep. 


€. Some time before 7 mybP the ( 1 1 region began 


to be built up and. the “Island Hoppers” interchange 
1950) began. —D. 


terrestrial inter- American connection was almost completed 


(Simpson. Between 2.5 and 3 mybP the 


and, consequently, the formation of the present American 


Atlantic coast was also almost complete. This. and the 
final connection, began the massive interchange 
of terrestrial mammals, i.e.. Simpsons “Late (Island Hoppers 
id Dotted 
ge ographic val ene open triangles: 


and) lines indicate the presen 


calc-alkaline volcanism: 


closed triangles: subduction and island-are or continental- 


margin 1 around the Caribbean. B-D. Grey 


un 
YT 


shading: continental blocks: dotte 


environments: grids: deep marine » 2 


anes with arrows: seaw: iw.) (B. 
Paseo Pantera: Una Wd de P 
Centroamerica by Anthony G. 
Smithsonian Books). 


Used by permission of the loan 
Institution. Copyright 2003.) 


Volume 93, Number 2 
2006 


Pascual 
South American Land Mammals 


THE LAST STEPS LEADING TO THE PRESENT 


LUJANIAN 


Figi About 8000-10,000 ybP the 
modern 1 ‘al mammal asse „mble age 
native 


he abundance of ungulate 


— 


fossil record, 


herbivores was markedly reduced, accompanied by 
a greal increase in the numbers of such specialized 
herbivores as huge edentates and caviomorph rodents. It 
is important to emphasize that there was a successive 
eradual increase in immigrants of North American origin 
toward the end of the Miocene, from the Late Island- 

through the rising 
Panamanian Bridge. 1.e.. (1950) Late (Island 
Hoppers and) Immigrants (cf. Fig. 12 B-D). Within the 


temperate regions of the continent, forest progressively 


hoppers to the early walkers. 


Simpson's 


ands. As 


— 


declined and was replaced by extensive grass 


a consequence, cursorial and grazing native communi- 


“Megafaunal Extinction” 


in the 


and subsequent newcomers began to feature 


ties of mammals spread, setting the stage for the 
forerunners of the pampean's (late Pliocene—Pleistocene) 


peculiar inhabitants. 
(3) The third stage is marked by the beginning of 


the interamerican exchange of land mammals, which 


prima facie is related to the beginning of the land 
connection between the Americas, in part due to the 
sustained low sea level that followed the Quechua 
Phase (Table l. Eustatic Curve). 
North 


occurred; the supposedly strong influence of these 


As a consequence, 


a nolable increase of American. immigrants 


immigrants on the extinction of native mammals has 


Patterson & Pascual, 


been strongly disputed (e.g.. 


1 tage ie, | A ES 
Toxodontids ~^ 


Vd Canids 
\ \ ( As 
r \ \ Camelids 
\ \ / / 
Agoutids X . N V. (Arc tothe teo Foo "rium m group) 
ù 25 , Tayassuids 


en IR 4 


Mammals that became extinct 
by ca. 10.000 years 


p (The Megafaunal Extinction) 


— | Murids 


Atelids Callitrichids iiic 


Figure 14. Between 2.5 and 3 mybP the principal interchange of the Great American Biotic Intere hange (Stehli € Webb. 1985). whose acme was about 1.5 mybP. began. “The predominance 
70 1 


eo 
of savanna-adapted mammals indicates a continuous corridor of such vegetational formations as thorn scrub or seasonally arid d whereas they are now separated by at least 1700 km o 
hund tropical Tore st... The final link between American mesic equatorial biotas evidently was established in the late Pleistocene” (Webb, 1985: 381). The fates of mammals that participated in 
the Great American Bini Interchange. up to the present Neotropical Region, are indicated on each of the most representative se ted taxa depicted here. 


dope jeoiuejog unossi|A 


9cc 


au) Jo s¡euuy 


Volume 93, Number 2 
2006 


Pascual 227 
South American Land Mammals 


1972; Cifelli, 1985). During the 5th International 
Theriological Congress in Rome (1989), Ortiz-Jaur- 
eguizar (1989: 277-278), Goin (1989: 271—272), and 
Tonni et al. (1989: 285-286), among others, again 
questioned the competitive model for extinction of the 
South American latest Pleistocene mammals. 

There is hard evidence that some South American 
megalonychid and mylodontid xenarthrans (Webb's 
(1985: 358-359) South"). 


apparently some North American carnivores such as 


"heralds of the and 
the procyonids, were able to cross the marine barriers 
rom about 8 to about 9 mybP, i. e., 
before the completion of the Panamanian land bridge, 
estimated at 2.5-3 mybP (Fig. 12). 
among Simpson's Late Island Hoppers. Thus, the final 


more than 5 Ma 


lang] 


These taxa were 
terrestrial connection of the Americas by the newly 


built Panamanian bridge initiated the massive inter- 


american exchange, whose acme occurred by the 
Ensenadan SALMA, between 1.8 mybP and ca. 
60,000 ybP. (Cione & Tonni, 1995, 2001). This 


occurred concomitantly with a progressive extinction 
of most of the native mammals, culminating in what 
by 


has been called "The Megafaunal Extinction” 


Martin (1984: 355; see figures encircled on Fig. 14). 


which was an extinction phenomenon that, with some 


temporal differences, occurred almost all over the 
world. 
(4) The last and fourth stage that marks the history 


of South American mammals is the so-called “Mega- 


Fig. 13). This is an extraordinary 


EM 


faunal Extinction” 


event that took place about 10,000 years ago. 
Together with the preceding arrival of man about 
13,000 years ago, this event shaped the basic 


mammalian fauna of the present Neotropical Region 
bullets Fig. 14). 


Le. in post- 


(see representative taxa with in 


somewhat later, 
Pleistocene times, the last North 
grants (quite probably inhabitants of Central America) 


Simultaneously or 


American immi- 


arrived. These included rodents such as Heteromyi- 


? Exhaustive details related to the integrated gronin 
I 


phenomena that led to The Great American Biotic In- 
terchange can be fo not only in this Symposium, 
particularly the paper by S. D. Webb, but also in the 


1985 book The Great . American Biotic Intere hange, Stehli and 


5" International Theriological Congress in Rome, R Pascual 
and 5. D. Bund 5^ ITC, Rome, 1989, Abstract of 


pp. 260-291; Chapters 1, 2, 4, 
Un Ina. Historia de la 
(editor), 
;. and London: 
Tropical 

s 


Papers and Posters, Vol. . 
and 5 of the 2003 bool Paseo Pantera: 
Coates 


— 
NEL 
go 
— — 
= 
= 


Smithsonian Institution Pr 
the 1996 ok Evolution 
America, J. B. ( n Jackson, A. F. Budd, and A. G. 
(editors). The University of C ved Press: and en 1999 book 
A Paleobiotic Survey of Caribbean Faunas ibd. e 1 8 
S. Collins and / . Coates 


ot. 


a Environment i 
A | 


the Isthmus of Panama, L. 


of 
(editors), Bull. Amer. Paleontol. 35 


dae, Geomyidae, and Sciuridae; Lagomorpha such as 
Also 
included were some genera of families already living 
in South 
Rafinesque (see the lower part of Fig. 13). 


Leporidae; and Insectivora such as Soricidae. 


America, such as the cervid Odocoileus 


Literature Cited 


Nuevos restos de mamíferos fósiles 


Ameghino en el Eoceno inferior de la 


Ameghino, F. 1891. 


descubiertos por a 


Patagonia austral. »ecies nuevas, adiciones y correc iones. 
Revista Argent. His st. Nat. : 289-: 
es os sé entes du Crétacé 


ire de Patagonie avec un 5 le 


— 


supérieur et du Tertia 


entre leurs faunes p i a s et celles de l'ancien 


continent. Anales Mus. Nac. Hist. Nat. Buenos Aires 
1568. 

Bertini; R. J., L. G. Marshall, M. Gayet & P. Brito. 1993. 
Vertebrate faunas from the Adamantina and Marília 
formations (Upper Baurú ae late Cretaceous, Brazil) 
in their stratigraphic and paleobiogeographic context. 
Neues Jahrb. Geol. Palüontol.. “Abh. 188: 71-101. 


1 


Bonaparte, J. F. 1986a. History of the terrestrial Cretaceous 
Vertebrates of b e IV Congr. Argent. Paleontol. 
eee Actas 2: 63-95. 

1986b. 1 7 e houssayl y nuevos 
mia ros Arge a IV Congr. 
Vr y PA dem. : 48-61. 

Ta. The Late Cretaceous 1 i Los Alamitos, 

ps Part VIII. The mammals. Revista 

s Rivadavia" Inst. Nac. 


cretácicos de I 


iade 


Patagonia, 
V rgent. E Nat., 
i: „ Paleontol. : 
(edil une 1987b. The 
Alamitos, Patagonia, Argentina. ji lus 
Nat., “Bern davia" Inst. Nac. 1 
Paleontol. 3: 102 
p Speci Cretaceous 

mE (rom Patagonia. /n: D. E. Russell. J.-P. Santoro & 
D. Sigogneau- 9 (editors), Teeth Revisited: Proceed- 


Mus. / 


3-] 


Late Cretaceous Fauna of Los 
Argent. Ci. 
Ci. Nat., 


ene Riva 
3-178. 


ialized dentitions in two 


ings of the International Symposium on Dental 
Morphology. Es 19 Mém. Mus. Natl. Hist. Nat., 
Paris, (Série C) 53 201-215. 
. 1990. New Late Cretaceous mammals from the Los 
Alamitos Formation, northern Patagonia. Natl. Geogr. Res 
: 63-93. 
1996. Cretaceous tetrapods of Argentina. /n 


1 (editor), Contributions of Southern South America 


to Vertebrate Paleontology. Münchner Geow Abh. 
Reiche 190 Geol. Paliiontol. m 73-103. 

——— . W. Rougier. The hadrosaurs. /n J. F. 
m— (editor), The od ca Fauna of Los 
Alamitos, Patagonia, Argentina. Revista Mus. Argent. Ci. 
Nat «Be pera Rivadavia” Inst. Nac. Invest. Ci. Nat 


55-16 
s 1985. RE sobre el primer mamífero 
1 Maastrichtiano 
17 3 


Pa 15 ontol. ; 


del Cret tacico argentino, 
(Condylarthra). D 21: 
„ M. R. 


Franchi, J. E. Powell 8 E. 


— el G Sepúlveda. 
1984. La Formación Los pu (Campaniano-Maas- 
trichtiano) del sudeste de Río Negro, con descripción de 

Significado 


Kritosaurus australis n.sp. (Hadrosauridae). 
paleogeográfico de los vertebrados. Revista Asoc. Argent. 
Geol. 39: 289-299. 

. Van Valen & A. Kramartz. 1993. La fauna local 
de Punta Peligro. Paleoceno inferior de la provincia de 
Chubut, Patagonia, Argentina. Evol. Monogr. 14: 1-6 


Annals of the 
Missouri Botanical Garden 


Martinelli, C. L. Schultz & R. Rubert. 2003. 
Mes sister group of 11 1 small cynodonts from the 


Late Triassic of Southern Brazil. Revista Brasil. Paleontol. 


: 0—ZÍ. 

Bondesio, P. & R. Pascual. 1977. Restos de Lepidosirenidae 
(Osteichthyes, 
de Colombia. Sus de 5 8 iones paleoambientales. 
Asoc. Argent. Geol. 32: 34-43, 

Broin, F. de & M. S. de > Puente. 1993. Les tortues fossiles 

Paléontol. 79: 169-232. 


trends in the 
65: 5 


Dipnoi) del com Honda (Mioceno tardío) 
Revista 


D 


d'Argentine: synthèse. Ann. 
Butler, 
tribos 


pit la, I 


P. M. 1990. 
phenic p Biol. Rev. 65: 529—552. 
. 1961. Sobre la presencia de un mamífero 
en el primer de neo (ic 1 de vertebrados del 
1 isico i Patagonia. Physis 22: 225-233. 
felli, R. L. 1985. South 1 an 1 ungulate e and 
249-266 in F. G. Stehli & . Webb 
aie The Great American Biotic a Bs 
Press, New Mun and London. 
Blanes n . P. Tonni. 1995. 
"Land Mammal Ages" i 


extinction. Pp. 


Chronostratigraphy and 
in the Cenozoic of Southern South 
America: principles, He Te ot es and the “Uquian” problem. 


. Paleontol. 69: 135— 


200 15 
Holoc ‘ene southern South American and E ;uropean verte- 
Se Paleontol. Ital. 40: 


Correlation of Plioc lo 


brate-bearing units. Boll. Soc. 


107-174, 


y ón 
Institution Press. Washington DC and London. 
Colbert, E. 1946 


suborder of E Groc odiliad 
70. 


represen ntalive of a peculiar 


from Patagonia. Bull. Amer. 
Mus. Nat. Hist. 87: 217-27 
Crisci, J. V.. M. M. ( 
1991. is a E ography of southern South America. 
Syst. Zool. -]l. 
P^ "9. 
paléogéographique des Dide 0 we EMail Y. Compt. 
ie Sel. Paris; Sér. D 288: 1457-1460. 
Donnelly, T. W. 1973. E HN 
9 111 an: a possible flood basalt province of vast size 
1 . Union 54: 1004 
1985. Mesozoic and Cenozoic plate evolution of the 
1-121 in F. C. Stehli & 
American Biotic 


agliano, J. J. Morrone & S. Roie-Juñent. 


Crochet, Données nouvelles sur l'histoire 


aceous ! asalts from the 


Prans. p 
9 


= 


Caribbean Bea Pp. 80 

W he Great Intere fiai ange. 
on. 

sio & R. Pascual. 1973. 


“Lepirosiden nive (Osteichthyes, Dipnoi) de la forma- 


Restos de 


¿Eoceno?) de Jujuy. Considera- 
ciones estratigráficas, paleo as y ANDE 
ficas. Ameghiniana 10: 152-172 

A. V K 2003. 


10 Femur of 
Monotremata) from the Early 


Forasie spi, A. M. & A. C. wtinelli. 
a monotreme (Mammalia. 
Paleocene Salamanca ‘a Formation of Patagonia. Argentina. 
ime 10: 625-030. 

IBA. 


Nes ww Tertiary Sebecosuchia (Crocodylia: 


Gasparini, Z. 19 


Me "SsOsUC hia) from Argentina. J. Vertebrate Paleontol. 
IS 


Middle 


reptiles, 


Gavel. M. et al. 
Ma istrichtian 


: al. 2001. 

vertebrates (fishes, amphibians, 

95 (Bolivia). 

id paleogeographic op e Palaeo- 
geogr. Palaeoclimatol. Palaoecol. 169; 39-6 

Gelfo, J. N. & R. Pascual. 2001. 


(Mammalia. „ 


. e 


mammals) from | pon Biostratigaphic, 


paleoecologic, al 


Pe oe tropicalis 


from the early Paleocene of 


Patagonia, a survi from a EUM Gondwanan 
37 


i 
radiation. Geodiversitas 23: 360— 


evolution. of 
2 


von, F. J. 1989. 


vate C enozoic South American marsupial 


and placental carnivores: changes in predator-prey 
evolution. Fifth Int. Ther Congr., Abstract of Papers 
and Posters J. 271-27 

et al. (+ 14 ahi 2005. El mamife ro Cenozolco 


América del Sur 
I] Congr. Latino- Amer. 
aneiro, RJ, Brasil, 10-12 


Resumos: 125-1206. 


más antiguo i Pon 


temprano, Chubu Argen ntina 


Paleontol. Verte 1 5 (Rio de. 
agosto 2005). Boletim de 
. Museu Naciona 
Gould. S. J. 1983. Nature's 


Hio de 
great era of experiments. Nat. 


Grambast, L., M. Martínes, M. Mattauer & L. Thaler. 1907. 
Perutherium altiplanense, nov.gen, nov.sp., premier mam- 
mifère mésozoique d'Amérique du Sud. Compte. Rend. 
Hebd. Séances Acad. Sci. 264: 707-710 

Krause, D. W., G. V. R. Prasad, W. 

l E. Grine. 1997. SE among p 

Nature 390: 504—507 

RC A J. > Z. Kielan- doomed & 
(edito . 1976 

E 1 History, I-X 


von Koe nigswald, 


Cretaceous mammals. 
J. Clemens 
. Mesozoic Mammals. The First 1 
University of California Press. 
erke = Los Angeles, London. 

15 Phee, D. E. 1984. Quaternary mammal localities and 


Novit. 


Amer. Mus. 


2 pe rodents of Jamaica. 
2803: l- 


& M. A. Iturralde-Vinent. 


1994. First Tertiary land 
mammal from Greater Antilles: an Early Miocene sloth 
from Cuba. Amer. Mus. 


(Xenarthra, Megalonychidae) 
Novit., 3094: 1-13 

SCA. Wo ods 8 G. 5. Morgan. 198 
rodent hc major and the 1 biogeography 


. The Pleistocene 


of a. Palaeontology 26: 831-837. 
Marshall. K P. Salinas. 1990. Stratigraphy of the Río 
Frías natin (Miocene), along the Alto Rio Cisnes. 
Aisen, Chile. Revista Geol. Chile 17: 57— 
Martin. E S. 1984. overkill: 
po 354—403 in P. lartin & R. G. Klein 


P vehitur the Et il model. 
y 1 
Quaternary Bde V dE Revolution. Uni 
Arizona Press, w ‘son, Ariz 
Matthew, W. D. & C. de I A Couto. 
} Ae Mus. Nat. 


2, Was Europe connected directly to North 


1959. 
st. 117: 1-56. 


The Cuban 


America n le U middle Eocene? 


Dobzhansky, Hecht & W. C. Steere (editors). | OM 
tionary eda 5 6. Appleton-Century-Crofls; New York. 
Mones. A. 1979. Los dientes de los vertebrados. Una 


introducción a su estudio. Direc. Gral. Extensión Univ.. 


Div. Publ. Edic.. 


Montevideo, Urugu 


1982. An equivocal nomene dae what means 
hypsodonty? Palüontol Z. 56: 107— 

Muizon, C. de 1901. La fauna de e de ws ra 

(Paleoceno inferior, Formación Santa Lucía). Bolivia. 


Pp. 575-624 in R. Suárez-Soruco (editor). Fósiles y 45 acies 
de Bolivia, Vol. : 
Ortiz-Jaureguizar. 


I. Vertebrados. Revista Téen. YPFB. 

E. 1986. Evolución de las comunidades de 
mamíferos Cenozoicos Sudamericanos: un estudio basado 
en técnicas de crea multivariado. J Congr. Argent. 
91-207. 

1989. Análisis of the . al changes of the 


Paleontol. Bioestratigr. 


Miocene- 


Fifth Int. 


South American mammal fauna during the 


Pliocene (Panaraucanian Faunistic ycle). 


Theriol. Congr. Abstract of Papers and Posters J. 
211—218 

etal. (+ 7 authors). 1993. Un análisis de similitud 
entre. | “Edad-mamifero  Friasense 


faunas de la 
Ameghiniana 30: 5 


ee 


51-352 


Volume 93, Number 2 
2006 


Pascual 
South American Land Mammals 


229 


1996. 
An approach to South 
Mastozool. Neotrop. 3: 133-15 

1998 The > history of South 
'elaceous-Paleocene transition. Pp. 9-18 i 


Pascual, 

American geobiotic 
52 

American land n 

the seminal € 


S. Casadío (e FEN Paleógeno de América de | 155 y 5 jh 


Península Antártica. Asoc. Paleontol. Argent., Publ. 
5. Buenos Aires. 


2001. 
denn (editor), 


South America, an island continent. /n 


No 
Bot. 


Towards Gondwana Aliv 
Biodiversity and Stemming the Sixth Extinction. Natl. 
Inst. 5 South Africa. 

. Bondesio. 1976. Notas sobre los vertebrados de 
la ms ra Cretácico-Terciaria. III. Ceratodontidae (Peces. 


Osteichthyes, Dipnoi) de la Fo ormación Coli Toro y de otras 
Sur de 


Mendoza. Sus d anclas pales 'obiogeográficas. Actas VI 


unidades del Cretácico tardío de Patagonia y 
Geol. Argent. I. 565-577. 

& EO Un 
(Hydrochoeridae) de la Edad Huayqueriense 


Congr. 
Cardiatheriinae 
(Mioceno 


ya. Sumario de los ambientes terrestres 


roedor 


tardío) de La Pam 
en l: 
19-35. 

& A. 1987. A 
extensive radiation of South American Paleogene marsu- 
pials. Fieldiana. Zool. 39: 99-110. 

1990. Evolving climates and 
South 


Argentina durante el Mioceno. Ameghiniana 10: 


Carlini. 


new superfamily in 


& E. Ortiz- Jaureguizar. 


e aunas in Cenozoic America. J. Human 
Evol. 23-00. 
& 1992. Evolutionary pattern of | 
mammal faunas dadas the Late Cretaceous and Paleocene 
America: a comparison with the 
pattern: In Forstén, M. Fortelius 
| dne Mammals and Evolution. yee 


Zoll. Fennici 98: 2 
& 


3E 


American 
Episodes: two major and unrelated episodes in the history 
of the South American mammals. J. Human Evol. (Sub- 
mitted to press.) 


p" and South 


1965. 


a Argentina, con 


, E. J. Ortega Hinojosa, D. Gondar & E. Tonni. 


Las edades del Cenozoico mamalífero de 


especial atención a d c. del territorio bonaerense. 
Anales come. pe x mos Aires. VI: 1605-193 

ies 6. J. Scillato-Yané & M. Bond. 
1085. is trus of mana diversification. in 
South America. Pp. 219-247 Stehli & S. D. Webb 
(editors). The ien American 5 Interchange. Topics 


New York 


= 


in Geobiology, vol. 4. Plenum Press. and 
London. 
1990, 


edentates 


xXxtinet and recent South 
and rodents: 
Atti Conv. 


Biogeogr. 


American and Caribbean 
standing examples of isolation. 
Naz. Lincei (Int. 


87: 621—040 


Symp. Aspects of Insularity) 


et al. (+ owen 1992a. The first discovery of 
monotremes in 520 America. Nature 356: 704-705 
et al. (+ 5 authors). 1992b. 11 75 first 1 


monotremes: an early Paleocene South American pale 


(Monotreme. Ornithorhynehidae). Pp. 1-14 in: L. 
Augee (editor), Platypus and Echydnas. Roy. Zool. Soc. 
New Se uth la ales, Sydney. 
liz-Jaureguizar & J. L. Prado. 

South 
Vrratia (editor), Contributions of 
Vertebrate 
205-319 


1906. 
American Geo- 


Land- 


mammals: ae for Cenozoic 


Evolution. J. G. 
5 America * 
Münchner Geowiss. Abh.. 30(A): 


Paleontology. 


Late Cretaceous-Recent land-mammals. 
: evolution. 


. W. Krause, E. Ortiz-Jaureguizar & A. 
A. € Caplin: 1999, i first gnathic remains of Sudamerica: 


implications. for 1 n c relationships. J. Verte- 


brate Paleontol. 12: 373-382. 

; Es 3alarino & D. Udrizar Sauthier. 2000a. 
New dental remains of Monotrematum sudamericanum. 
Their bearings on monotreme relationships an 
Gondwanan u e radiatio dom Argent. 


(La Plat: 
92-93 

González, A. Ardolino & P. F. Puerta. 
bello derived dur odont from the Patagonian 


Mastozool —10 Novio uis 2000). Libro de 


Resúmenes: 


2000b. 

Late Cretaceous: evolutionary implications for Gondwanan 

mammals. Geodiversitas 22: 3954.14. 

. M. L. Balarino & D. Sauthier. 2001. The 

K/T and Te rliary-Pleistocene ERE American mammalian 
Sy 


Idrizar $ 
turnovers. Similar phenomena? VII Int. Mesozoic 
Terrestr. Ecosyst. (Buenos Aires, September 26-0 ctober 
. 1999.) Asoc Argent., Publ. 
nee 


Paleontol. 


Especial 


al. (+ authors) 2002. Un mamitero 


5 gran 
MUN Drvolestida) del Cretácico superior de Patago- 


nia. Argentina. VIII Congr. Argent. Pale ontol. DONNE 
Corrientes, 7—10 octubre on Resúmenes: 
Patterson, B. € R. Pascual . The fossil d Fauna 


A. Keast, F. C. Erk € 
Southern 


Continents. S v. New York Press, Albany. 


Paula Couto, C. de 1967. P lei pens edentates of the West 
Indies. Amer 1158 Novit. 2304: 1—55. 

Rauhut, O. W. T. Martin. E. Ortiz-Jaureguizar & P. 
Puerta. 2002. 2 la mammal from South America. 
Nature 416: 165-168. 

Ray. C. E. 1964. The taxonomic status of „ e 
ontogeny it 1 Elasmosdontomys and - m rhiz entia. 
Cavin Bull. Mus. Comp. Zool. 109 2. 

Reguero, M. S. A. Marenssi & 5 m Santillana, 2002. 


Antarctic South 


faunas and environments: biogeo- 


1 and America | (Patagonia) 


Paleogene terrestrial 


graphic QE 8 Palaeoclimatol.. Pa- 
O: 5210 


aeoecol. 17 


Ribeiro, A. M. 5 ou & F. Abdala. ue Primeiro 
ane (M ha es | | 

g ( Pricon Mesozóico 

da América do Sul. XVII Congr. Brasil. RAN (Rio 


Branco-Acre, 5-9 agosto 2001), Boletim de Resumos: 169. 
Rougier, G. W. 1993 (Unpublished). Vincelestes neuquenia- 
nus Bonaparte (Mammalia, Theria) un Primitivo Mamífero 
del Cretácico Inferior de la Cuenca Neuquina. Ph.I 

Aires. 

&. A. Hopson. 1992. Reconstruction of 


eda in the 


dissertation. 100 Nac., Buenos 
J. R. Wil 


the cranial Early Cretaceous. mammal 


Vincele "stes n e ations for the evolution of 


the mammalian cranial vascular system. J. Vertebrate 
Paleont l. 12: 188-216. 

Savin, S. M. & F. C. Stehli. 1975. Tertiary marine 
alentó A ratures. Geol. Sor. Amer. Bull. 86: 1499-1510 


Scillato-Ya G. J. Pascual. 1984. Un pec ilias 
Paratheria, E dentata aama del Paleoceno me de 
Patagonia. Primeras Jornadas Argent. Paleontol. Verteb- 
rados (L a Plata. 28-29 mayo 1984), Resú 

985. Un Xe silva del 

Paleoceno medio de Patagonia (Argentina). Su importancia 

ie mática de los 


peculiar 


Paratheria. Ameghiniana 21: 


B. 1972. La 
supérieur de | 
Mus.! Natl. 


u Crétacé 
Bull. 
2 5—40 JO. 


faunule de mammifères 
aguna - mayo reges s Es ruviennes). 
99, 3 


Hist. Nat., msn . Terre 19 


Annals of the 
Missouri Botanical Garden 


Simpson, G. G. 1932. The supposed association of dinosaurs 


with P » Tertiary type in Patagonia. Amer. Mus. 
ovil. 

“19 d ne new 1 0 an 1 from 2 SUME beds 

of Patagonia. Amer. Mus. Novit. 623: 
——— |93 Ta. New pie bom H p of South 

Ameria, Amer. Mus. Novit. 927: 1-3. 
3 Tb. An ancient eusue es crocodile from 

"mes Amer. Mus. Novit : 1-2 
— — — 1948. The beginning 15 e of mii in South 


Mu "Nat. Hist. 91: 1-232. 


Amer. 


America. Part 1. Bull. Amer. 

———— 195 50. er of a Fauna of Latin 1 
Sel 361-3 389. 

1 Splendid. Isolation. 

Yale Univ. 


The Curious History of 


South American Mammals. Press, New Haven 


Smith. A. G., D. C. Smith & B. M. Funnell. 1994. Atlas of 
Mesozoic and Cenozoic Coastlines. Cambridge Univ. 
Press, New York. 

stehli, F. G. & S. D. Webb (editors). 1985. The Great 
American Biotic Interchange. Plenum Press. New Yor 

Storey, B. C. 1995. The role of mantle plumes in contine inia 
breakup: case 1 S from Gondwanaland. Nature 377 
301—308 

Tonni, E. P., J. L. Prado, M. S. Bargo, A. N. Menegaz & W. 


D. Berman TA Pleistocene-Recent E 0 com- 
munities. Pp. 285-286 in: Fifth Int. riol, Congr. 
(Rome, 22-29 August 1989), Said 89 ale 
Cenozoic res dispersal between Americas” (R. 
Pascual € . Webb, conveners), Abstracts of Papers 


and 1 1 
Vicente, J. C. 1972. 
thiliennes centraux au an illele de 
Sect. 24—430. 


\pergu sur l'organization des Andes 
l'Aconca- 
Montreal, 


Argentine-( 


gua. 2dth Int. Geol. Congr., 


Canada 


Vizcaíno. S. F., Pascual, . Reguero & F. J. Goin. 
1998. Antarc ‘lic a as bac bs for mammalian evolution. 


Pp. 199-209 in: S. Casadío (editor), Paleógeno de América 
del Sur s de là Península Antártica. Asoc. Paleontol. 
Argent., Publ. Especial 5. 

Webb. S. D. 1978. A history of savanna vertebrates in the 
New World. Part Il: South Pa and the Great 
Interchange. Annual Rev. Ecol. Syst. 9: 393—126. 

985. Main pathways of a dive e 7700 
in North America. Pp. 201-217 in: F. Stehli & ebb 


(editors), The Great American Biotic i ats os 
in ay Vol. 4. Ple 
200€ Greal 


patterns and processes. Ann 
. EE 957 


num Press. 
American Biotic 
Missouri Bot. 


06. The Interchange: 
Gard. 93 
245-257. 

n G. P. & l 


Arens. 2001. The evolutionary impact 


veirie seaway on Late Cretaceous and Paleocene 
\merica. VII Int. Symp. Mesozoic 
s. September 26-October 


f: 
muon s of South 


Terrestr. Ecosyst. (Buenos Aires 


l, 1906 ). | 'aleontol. Argent... Publ. Especial 7: 
185—189 
Wolff, R. 6 1984 new early Oligocene argyrolagid 
(Mamm: "e oe from Salla, Bolivia. J. Vertebrate 
Paleontol. -113 


Woodburne, 1987, Mammal Ages, Stages, and Zones. 
Pp. 18-23 in: M. Woodburne (editor). 
Mammals of North America. Geochronology and Bioestra- 

Ur 


Cenozoic 


tigraphy. Univ. Calif. Press, Berkeley. Los Angeles. 
London 

Zachos, J., M. Pagani. L. Sloat Thomas & K. Billups. 
s Trends. boo and she rations in global climate 


Science 292: 686-693. 
Cretaceous paleogeography of 
Palaeoecol. 59: 


> Ma to Present. 
rus Aster, W. J. 1987. 
Antarctica. Palaeogeogr.. Palaeoclimatol., 


191—200. 


PALEOGEOGRAPHY OF THE S. Blair Hedges“ 
ANTILLES AND ORIGIN OF WEST 

INDIAN TERRESTRIAL 

VERTEBRATES’ 


ABSTRACT 


The fauna of the West Indies includes more than 1300 native terrestrial vertebrate species and is characterized by high 
levels of endemism. Several theories have been proposed to explain how these animals arrived to the islands, inc luding 
dispersal, vicariance, and land bridges. The dispersal theory proposes that most of the West Indian terrestrial biota arrived by 
flying or by flotsam. The vicariance theory suggests that there was a pr roto-Antillean land mass, or masses, connecting North 


J oo 
and South America in the late Cretaceous that traveled eastward as the Caribbean geologic plate developed and carried an 
ancient biota with it. One widely discussed land bridge theory proposes that much of the Antillean biota a de by 
dispersal over an unbroken dry land connection, the Aves Ridge. between South America and the Greater Antilles 35-33 Ma. 


Geologic evidence cannot unambiguously support or refute any of these mode ls, despite claims to the contrary. Othe +r evidence 


bearing on these three theories, or mechanisms, comes from the taxonomic composition of the fauna, their phylogenetic 
re sa fossil record, "n oge ography, ecology, climate. water currents, and divergence time estimates from molec ular 
c | This evidence supports an origin by overwater dispersal for most or all of the West Indian terr estrial vertebrate Es 
The strongest support comes from a reduced higher-level taxonomic composition of f the fauna (now and in the past). the 
‘land 


presence of unusually large adaptive radiations, and the finding of f divergence time estimates between island and main 


eroups that are not astral l around a particular time. In addition, the mé gods of terrestrial (non-flying) groups have closest 
relatives in South America, which is consistent with the direction of water currents and most hurricane tracks. Some of the 
same evidence in support of dispersal argues against a mid-Cenozoic land bridge from South America. Several ancient, 


relictual groups (e.g., xantusiid lizards and en nodontoid shrews) may have arisen through proto-Antillean vicariance, and 
molecular clock analyse ses—here revisited— provide support for this, but an origin by dispersal can also be argued. No model 
can be completely disc ounted. Although the general pattern (dispersal) has emerge «d, many details remain to be determined 
concerning the origin of the fauna 

Keywords: ls bioge s WE Caribbean, dispersal, systematics, vertebrates, vicariance, West Indies. 


RESUMEN 


a fauna de las Antillas incluye más de 1300 especies nativas de vertebrados terrestres y se caracteriza por altos niveles de 
Eo mismo. Se han propuesto varias teorías para explicar cómo estos animales llegaron a las islas, incluyendo dispersión, 
vicarianza, v corredores terrestres. La teoría de la dispersión propone que la mayoría de la biota terrestre antillana llegó 
volando o en y restos flotantes. La teoría vic ariante sugiere que hubo una masa (o masas) de tierra proto-antillana, que conec taba 

y América del Norte y del Sur en el Cretácico tardío y que se desplazaba hacia el este mientras la placa geológica del Caribe 
se desarrollaba y llevaba consigo una biota antigua. Una teoría ampliamente dise ulida acerca del corredor terrestre propone 
a mayoría de la biota antillana se originó por dispersión sobre una conexión seca intacta de tierra, el Promontorio de 


que 
\ ves 


intre América del Sur y las Antillas Mayores 125 'e 35-33 Ma. La dns ncia geológica no puede apoyar o refutar en 
forma inequívoca eae de estos ae los, a pesar de que se opine lo contrario. Otra evidencia concerniente a estas tres 
rías, o mecanismos, proviene de la composición taxonómic a de la fauna, de sus relaciones filogenéticas, del registro fósil, 
ans paleogeografia, la ecología. el clima, las cor s de agua, y esti vA de tiempo veras mies de los relojes 


4 
— 


de 

moleculares. Esta evidencia apoya un origen por dis spersión para la mayoría de la fauna vertebrada terrestre antillana. E 
apoyo más fuerte proviene de una reducida composición taxonómica de alto nivel de la fauna (en el presente y el pasado). la 
presencia de radiaciones adaptativas inusualmente grandes y el | 1allazgo de divergencia en los estimados de tiempo entre 
erupos de las islas y del continente que no se agrupan alrededor de un tie mpo particular. Además, la mayoría de grupos 


terrestres (que no vuelan) tienen parientes más cercanos en América del Sur, que es consistente con la dirección de las 
corrientes de agua y de la mayoría de huracanes. Algunas de las mismas e videncias de apoyo a la dispersión argumentan en 


contra de un corredor terrestre desde América de | Sur en el Cenozoico medio. Varios grupos antiguos, relictos (.. lagartos 
xantusiidos y musarañas solenodontoides) pudieron haber surgido a través de vicarianza proto- -antillana, 3 . anälisis de 


relojes moleculares—aquí revistos—dan apoyo a esto, pero también se pue «de argüir un origen por dispersión. Ningún modelo 
se puede descartar totalmente. Aunque el patrón general (dispersión) A sobresalido, ndana quedan por de ‘terminar muchos 


detalles referentes al origen de la fauna 


l thank Alan Graham for the invitation to paro ipate in the E e Ann Budd, Robert Henderson, Robert Powell, and 
Tom Zanoni for comments on the manuscript; Pervaze Sheikh for the photograph of flotsam; George R. Proctor for reminding 
me of Lefty the hurricane; and Jaime Blair for assistance with analyses. This research was supported by grants from the 
National Science re ation. 
"pe oe nt of Biology, Pennsylvania State University, 208 Mueller Laboratory, University Park, Pennsylvania 16802- 
5301, Us ot )su.ec 


ANN. Missourt Bor. Garp. 93: 231-244. PUBLISHED ON 23 Aucusr 2006. 


Table 1. 


Annals of the 
Missouri Botanical Garden 


Taxonomic diversity of native West Indian terrestrial vertebrates. 


Genera Species 


Group Orders Families? Total 
Freshwater fishes 6 9 14 
Amphibians l 4 6 
Reptiles 3 19 50 
Birds 15 19 204. 
Mammals: 

Bals l 7 32 
Other! | 9 36 
TOTALS 30 97 342 


Endemic % Endemie Total Endemic £6 Endemic 
6 13 74 71 96 
l 17 173 111 99 
9 18 499 178 96 
38 19 125 150 35 
8 25 58 29 50 
33 92 90 90 100 
05 28 1319 089 75 


“Includes one endemic family of birds (Todidae) and four of mammals (Capromyidae, Heptaxodontidae, Nesophontidae, an 


Solenodontidae). 
" Edentates, insectivores, primates, and rodents. 


The West Indies has captured the attention of 
18 


reasons for this include a large but 


biogeographers for over a century (Wallace. 
the 


taxonomically peculiar fauna, high levels of endemism, 


Some of 


and an intriguing physiography. For example. there are 
four large islands, the Greater Antilles, now known to 
be quite old (ca. 100 million years): a large group of 
smaller, flat 
Bahamas Bank; and a textbook-like are of volcanic 


islands on a limestone. plateau. the 
islands, the Lesser Antilles. When plate tectonies and 
past. sea-level fluctuations are added, this physical 
selling. presents a complex challenge for the bio- 
geographer. The primary task is to understand how and 
when the biota arrived. A comprehensive review of 
Caribbean biogeography is not possible in this limited 
and thus for additional insights see recent 
reviews by Hedges (1996a, c, 2001), Iturralde-Vinen 
& MacPhee (1999), Graham (2003a, b), and Santiago- 
Valentin and Olmstead (2004), some of which will be 


discussed and contrasted herein. Although models of 
historical biogeography apply to beth plants and 
animals, examples will be drawn from the latter, and 
primarily land vertebrates of the West Indies. 

Most groups of West Indian invertebrates are poorly 
known. Examples of some groups that have been given 
axonomic attention are ants (Wilson. 1985, 1988: 
Morrison, 1998), termites (Scheffrahn et al.. 2003), 
butterflies (Smith et a 
1996), land snails (Gould & Woodruff, 1986: Good- 
friend & Gould, 1996), scorpions (deArmas, 1982). 
mites (Niedbala, 2004), and selected ol 
1988, 1992). 


amber have added significantly to the knowledge of 


, 1994), earthworms (James, 


er groups 
(Liebherr, Fossils. from Dominican 
West Indian invertebrate diversity. In some cases the 
amount of information is so detailed and material of 
that 


behavioral interactions can be inferred. (Poinar & 


such high quality complex ecological and 


Poinar, 1999). However, studying the historical bio- 


geography of a group is difficult when a majority of the 
living species has vet to be discovered. West Indian 
vertebrates, on the other hand, are better known 
taxonomically, and therefore most zoogeographic 
analyses have focused on these animals. 

Currently 1319 species of living vertebrates are 
known from the West Indies (Table 1). Relatively few 
of 


encountered, indicating that knowledge of the species 


new species mammals and birds are being 


diversity of those two groups is nearly complete. 
However, amphibians, reptiles, and to a lesser degree 
freshwater fishes; show more steeply rising discovery 
curves, suggesting that undescribed species still exist 
(Hedges, 1996a. c). 


reptiles (499 species). and birds (425 species) are 


Amphibians (173 species). 
the most species-rich groups, although levels of 
endemicity differ considerably. Only about one-third 
of bird species (35%) are endemic to the West Indies, 
whereas nearly all species of amphibians (9966). 
reptiles (9696). and freshwater fishes (96%) are 
endemic. That the ability to fly is the reason for this 
difference is seen clearly in mammals; only 50% of 
bals are endemic, whereas all non-flying mammals are 
endemic (Table 1). 

Not unexpectedly, most West Indian vertebrate 
fossils are from the Quaternary, <1.8 million years 
ago (Ma), and most often are from regurgitated 
remains of owl meals and vietims of natural pitfalls 
(Pregill & Olson, 1981: Pregill, 1986; Pregill et al., 
1992). Approximately 15 to 20 vertebrate taxa are 
known from the Paleogene (65-1.8 Ma). including 
those in amber (MacPhee & Iturralde-Vinent. 1994. 
1995; MacPhee & Grimaldi, 1996; Poinar & Poinar, 
1999). The amber fossils are mostly from the same 
formation, but the date (30-15 Ma) has been disputed 
(Grimaldi, 1995; Hedges. 1996a: Iturralde-Vinent & 
MacPhee, 1996). With 


Paleogene fossils, a clear picture of the Paleogene 


such a small number of 


Volume 93, Number 2 
2006 


Hedges 
Paleogeography of the Antilles 


233 


V 


f 


ertebrate fauna is not available. However, current 
ossils suggest that the fauna was similar to the 


Quaternary fauna in higher-level taxonomic composi- 
tion. For example. fossils of frogs (Eleutherodactylus 


Dumeril 
Sphaerodactylus Wagler), a snake (Typhlops Oppe 
and a mammal (Solenodontidae) 
vertebrate 


(MacPhee 


and 
). 


are representatives of 


& Bibron). lizards (Anolis Daudin 


groups present today on the islands 


& 1996; 


Giant megalonychid ground sloths (Order 


Grimaldi, Poinar Poinar, 
1999). 
Edentata) 
the Antilles, as evidenced from Quaternary fossils, 
the Early 


underwent a major adaptive radiation in 


and their fossils are known starting from 


33 Ma (MacPhee & Iturralde-Vinent, 


Oligocene, ca 


1995). Primates are also known from the Paleogene 
and Quaternary of the West Indies (MacPhee € 


Iturralde-Vinent, 1995; MacPhee & Horovitz, 2004). 
No sloths or native primates occur in the West Indies 
A rhinoceros relative (Hyrachyus Leidy) i 
50 Ma) of Jamaica 
ungulates are 


un 


today. 
Eocene (ca. 
(Domning et al., 1997). 
known from the West Indies subsequent to that record. 
In fact. Hyrachyus may represent a survivor of the 
proto-Antilles or an early disperser from Central 
1997; 1999; 


known [from the 


but no other 


al., Pregill. 


ph 


America (Domning « 


Hedges, 2001). 

Although many naturalists have commented on the 
West Indian fauna over the centuries (e.g.. Sloane, 
1725: Gosse, 1851; Darwin, 1859). Wallace (1981) 
was one of the first to discuss the historical 
zoogeography of the islands. He noticed that the 

taxonomic composition of the West 


higher-level 
Indian fauna was reduced when compared to that of 
a fact also noted by subsequent 


mainland areas, 
zoogeographers (Matthew, 1918; Simpson, 1956; 
Darlington, 1957; Williams, 1989). Such patterns 


are characteristic of oceanic islands where the fauna 
arrived by dispersal, rather than of islands 
previously connected to mainland. areas (Wil 
1989; Paulay, 1994). the. debate 
during the first half of the 20" Century concerned the 
possibility of land bridges (Scharff, 1912: Barbour, 
1916: Schuchert, 1935). Since about 1960, knowledge 
of seafloor topography and the acceptance of plate 


tectonics have greatly changed the physical backdrop 


has 
liams, 


However, much of 


for zoogeographic studies, and rendered the sugges- 
tions of those land bridges obsolete. For this reason, I 
briefly review the current knowledge of geology, water 
currents, and climate. before discussing contemporary 


models of Caribbean zoogeography. 


GEoLocic HISTORY AND PALEOGEOGRAPH) 


Caribbean region was formed when the 


The 


supercontinent Pangaea broke apart in the Jurassic 


76 Ma 


Proto-Antilles 


Costa Rica- 
th 
Panama Arc 3 
America 
ASS 
1000 km 
l. Location of the Proto-Antilles in the late 


Figure 
Cretaceous (ca. 76 Ma), redrawn and modified with permis- 
sion from Pinde 11 (1994). Present coastline of southern North 
ines without teeth 
are fault zones: teeth 
subduction zones (teeth indicating direction of subduction). 
that may or may 


ica is rem as narrow line; 


heavy lines with are 


Shading denotes continental and arc areas 
not have been above sea level. This remains speculative for 
the Proto- Antilles, which may have been eme 
islands (as in the current Lesser Antilles). or 
submerged. The Costa Rica-Panama Arc is shown to the west. 


"gent, a chain of 
completely 


(200—150 Ma). The initial ocean floor that was created 
by a spreading center has since been recycled through 
subduction. The current Caribbean Plate formed 

the mid-Cretaceous and has 


American Plate 


the eastern Pacific 


moved east-northeast relative to the 


(Dengo & Case, 1990; Donovan & Jackson, 1994: 
Pindell, 1994; Pindell & Kennan, 2002). although 
models that do not require high rates of plate 


movement have been proposed (Meschede & Frish, 
1998). The present-day Greater Antilles were formed 


mostly from subduction, and resulting island-arc 
volcanism, of the North American Plate under the 


lighter Caribbean Plate during this eastward move- 
ment (Fig. 1). Antillean 
with the Bahamas Platform i 


However, collision of this 


island arc | the early 
Paleogene caused subduction to stop and a fault zone 
and short spreading center (Cayman Trough) to form 
south of Cuba, Cuba eventually 
northern Hispaniola and Puerto Rico to the American 
Plate. Subduction continued along the leading (east) 
Plate during the Cenozoic, 


thus fixing and 


— 


edge of the Caribbean 


always feeding an island arc on the overriding 
(Caribbean) plate. Initially the are was the Aves 


Ridge (probably then a chain of islands), but later in 


the Cenozoic the expression of are materials (andesitic 
volcanism) transferred slightly eastward to its present 
position as the Lesser Antilles. Among the Lesser 


Antilles, Barbados is an exception, being emergent 
Atlantic sea floor caused by the buckling of the 


Atlantic oceanic plate. The northern boundary of the 


234 


plate has remained largely a transform fault and minor 


spreading center. 
The Bahamas Platform has been a stable carbonate 
the North Plate the 


and reef development has kept the 


block on American since 
Cretaceous, 


platform near sea level apparently for much of that 


time. Although Jamaica and the southern portion of 


Hispaniola were initiated. during the development of 


the proto-Antilles, their histories have been quite 
different from that of Cuba, northern Hispaniola, and 
Puerto Rico. They have been attached to the northern 


boundary of the Caribbean Plate and have moved 
eastward relative to Cuba and northern Hispaniola 
during the Paleogene. Jamaica has remained isolated 
during this time, but southern Hispaniola collided 
with northern Hispaniola in the Miocene (Huebeck & 


Mann, 


Cul de Sac- 


1985), with the zone of attachment being the 
Valle de Neiba of present-day Hispaniola. 
The similar origin and proximity of Cuba, northern 
Hispaniola, and Puerto Rico raises the possibility that 
two, or all three, of these islands were attached 
some point during the Paleogene, although that has 
yet to be firmly established (Pindell, 1994). The close 
similarity of the fauna of the Virgin Islands and Puerto 
Rico reflects the fact that they reside on the same 
bank. A more detailed account of Caribbean geologic 
history can be found elsewhere (Pindell, 1994: Pindell 
& Kennan, 2002 

Also of 


paleogeography of the West Indies. 


the 


This is a compli- 


importance for biogeography is 
cated problem because several poorly known variables 
must be considered simultaneously, such as sea level 


fluctuations, geologic uplift, and erosion, Some high 


mountain ranges (e.g. Blue Mountains of Jamaica) 
were uplifted over a period of only a few million years 
(late Miocene and Pliocene) and even higher ranges 
were eroded in less than a million years, as evidenced 
by the Eocene Wagwater Trough formation of Jamaica 
(Comer, 1974). Therefore, the current topography is 
a relatively poor guide of paleogeography during most 
of the Paleogene. Clearly, at some point between the 
inception of subduction in the mid-Cretaceous and 
today, the Antilles became emergent. 
islands first rose above the 
if at all, 
they were connected is not a simple task, and is often 


Determining when the 
sea (and remained above sea level) and how. 
based on the distribution and 


type of marine 


sediments that presumably came from erosion of 


emergent areas. This type of information has 
suggested that the proto-Antillean island are of the 
late Cretaceous probably was not a continuous (un- 
America but 
islands (Donnelly, 1989: Perfit & 
1990, 1992; Pindell € 


although the details are unknown. 


broken) land area resembling Central 
rather a chain of 
Williams, 1989; 
Kennan, 2002), 


Donnelly, 


Annals of the 
Missouri Botanical Garden 


Some workers have claimed that no land areas in the 
Antilles were constantly 
before 45 Ma (MacPhee & 
MacPhee & 
MacPhee, 


cannot be backed by existing evidence. 


level 
1994: 


lturralde-Vinent & 


Greater above sea 
lturralde-Vinent, 
1990: 


1999). although such absolute. statements 


Grimaldi, 


To support 
this claim would require having an uninterrupted 
geologie record from most areas in the Caribbean 
region, and this is not yet available (Dengo & Case, 
1990). In addition, small areas low in elevation would 
not be expected to yield large quantities of terrestrial 
sediments. 

The claim by Iturralde-Vinent & MacPhee (1999) 
the Antilles 
5 Ma, 


sea level before 45 


that no. land areas in Greater were 
constantly above and their 


advocacy of a mid-Cenozoie dry land connection 
(Aves Ridge land bridge) with South America for the 
purpose of emplacing land mammals to the Antilles 
(see below). demonstrate that their models are not 


“paleogeographv” but rather biogeography. The dis- 
} BCograpin Beograptn 


linction is important because it provides the in- 
dependence needed to falsify or test biogeographic 
models. For example, if a biogeographic model claims 
that animals must have walked across land to reach 
another area, and paleogeographic evidence says that 
no such land existed, then the biogeographie model 
can be rejected. 

Finally, an important detail of Earth history of 
the the 
Paleogene bolide impact of 65 Ma. This was one of 
the 


relevance to Caribbean is Cretaceous- 


argest impacts in the inner solar system since the 
Precambrian, and is believed to have exterminated the 
dinosaurs and other groups (Hildebrand & Boynton, 
1990). The 


under sediments of the Yucatan Peninsula (Kring & 


impact site is in the Caribbean region 


Boynton, 1992). At the time of impact, the islands of 


the Greater Antilles were only 1-3 crater diameters 


away (Pindell, 1994) and undoubtedly experiel iced 


massive waves hundreds to thousands of meters in 


height (Maurrassee, 1991). In addition to the elobal 
climatic effeets of the impact, these local effects must 
have resulted in much extinction of the proto- 


Antillean fauna present on the islands at the time 


(Hedges et al., 
WATER CURRENTS 


The direction and speed of water currents is of 
importance. to the dispersal of terrestrial organisms 


unable to fly. Even those non-piscine vertebrates 


capable of swimming are unlikely to swim against the 
current for long distances. For nearly all West Indian 


vertebrates, this mode of transport would involve 


rafting on floating vegetation (flotsam) washed. into 


ocean currents from the mouths of rivers following 


Volume 93, Number 2 
2006 


Hedges 235 
Paleogeography of the Antilles 


Figure 


rivers. The upper photo shows the entire island: the lower photo is a close-up of the left side. 


A floating island of vegetation (ca. 600 m^) near Santarém, Brazil, at the confluence of the Amazon and Tapajos 


with a cattle egret (Bubulcus ibis). 


approximately 50 em in height. Photograph taken by Pervaze Sheikh in February, 1997 (used with permission). 


storms (Fig. 2). This method of dispersal is well 
known (Guppy. 1917; King, 1962; Heatwole & Levins, 
1972), and has been documented for West Indian 
vertebrates (Henderson et al., 1995; Censky et al.. 
1998; Knapp. 2000). 

In the West Indies. 
s predominantly from the southeast 


the present-day surface current 


o the 


flow 
northwest, and would have been similar in the past 
he 


(Fig. 3). This is because the Caribbean is i 
southwestern portion of the North Atlantic Gyre. the 
clockwise flow of water in the Atlantic. The clockwise 
flow is the result of the differential rotational velocity 
of the Earth 
from the west coast of Africa across to northeastern 


Coriolis Force). Water currents come 


— 


South America and then up and across the Caribbean 
to the southern tip of Florida where they form the Gulf 
Stream. One of the best documented cases of flotsam 
dispersal involves a West Indian vertebrate, green 
iguanas (/euana iguana Linnaeus), and is consistent 
with this pattern. The distribution of this species and 
hurricane paths suggest that the iguanas originated in 
Guadeloupe and were transported to Anguilla, about 
250 km to the northwest (Censky et al., 1998) 
The closest relatives of most West Indian terrestrial 
and freshwater 


vertebrates (excluding birds, bats, 


fishes) are in South America, a fact that is compatible 


with the hypothesis that current flow is responsible for 
most overwater dispersal. The Amazon and many other 
major rivers in northeastern South America (Guiana 
Shield) drain into the Guiana Current, which flows 


he Caribbean, and presumably carried the 


— 


into 
ancestors of much of the West Indian biota (Fig. 4). 
The closest relatives of a few West Indian vertebrates 
1917) are in 


Africa. In those cases, the dispersal event would have 


(Hedges, 1996c) and plants (Guppy. 
taken at least several months based on the present 
rale of current flow (Guppy. 1917). 

The West Indies has been north of the equator since 
the origin of the Caribbean Plate. and therefore the 
general east-to-west current flow would have been the 
same in the past, even prior to the emergence of the 
Isthmus of Panama (Fig. 3). An 
made elsewhere (Iturralde-Vinent & MacPhee, 1999) 


that a dry land bridge between South America and the 


argument has been 


[om 


— 


Greater Antilles in the mid-Cenozoic would have 


altered that general. current flow. In addition, those 
authors have argued that, based on past ocean current 
flow, most animals carried on flotsam during that time 


would not have ended up on islands in the West 


Indies. However. these suggestions are speculative 
(see below). Even in the presence of such a land 
bridge, the North Atlantic Gyre still would have 


Gulf Early Oligocene 
\ Stream 36-30 Ma 
X 


Pacific 
Ocean 


Figure 3. Water current Due ros in the West Indies in 
the (A) Early Oligocene (36-30 Ma) and (B) Pliocene/ 
Quaternary (4-0 Ma). (Copyright 2001. from 
8 5 ography of the West Indies: Patterns and Perspectives by 

Voods & F. E. Sergile (editors). Reproduced by 
| Routledge/Taylor Francis, LLC.). 
Guiana Shield (Guiana Curre nl) in the past is 


alter | ele 5 


permission. of N current 
flow along the ( 
inferred. given its lento ab that time. Carbonate 
platforms (horizontal bars) may oe affected current flow in 
the Caribbean (Droxler et al.. 8). —A. 
the Aves Ridge (shaded crescent) was submerged or was 
ret al., 1998) s current. flow 
(dashed arrows) would have passed on to the Pacific 
the Aves Ridge 

suiana Current would have 
been deflected to the s st along the Antillean land- 
Gulf S 
South pos rica into the 
prevailing current would have provided a source of flotsam 
for the Antilles. —B. Pli 


current continues to flow 


Early Oligocene. If 
a chain of islands (Drox some 
Ocean, 
as il did during the Miocene. 
land bridge (dotted lines) the 


were a dry 


masses and up to the . Under both scenarios, 


rivers in not "the Paslern draining 


ocene/Quaternary. The Guiana 


along the coast of South America 


and into the Caribbean, bringing flotsam to the Antilles. 
brought currents from northeastern South America to 
the Antilles (Fig. 3; Hedges, 2001). 


where flotsam from particular rivers in South America 


Also. the focus on 


would go, rather than on where flotsam reac hing the 
| logic. If 


lizards riding on flotsam from a river in UR d MA 


Greater. Antilles originated, was an error 
South America were carried out to the Pacific in the 
(turralde-Vinent & MacPhee, 1999), 


that was their unfortunate luck. To understand the 


mid-Cenozoic 


origin of the Antillean fauna, the hits are important: 
the misses do not count. If the Antillean hits are still 


mostly from South America, as opposed to Central and 


North America, those arguments concerning misses 
are of no consequence. Even if the implication is that 
fewer hits would occur, that also is of little 


consequence when the odds being discussed already 


are as low as one in 65 million vears. 


Annals of the 
Missouri Botanical Garden 


Distribution of : 


species of snake (Corallus 
pera "us; 1 in South 

s (C. cooki Gray and C. grenadensis Barbour: 
dashed lines) in Lesser Antilles (Henderson & He dges, 1995: 
Henderson, 1997), 


Figure 
hortulanus America and its 


closest relativ 


Jased on their distributions. 
direction of current flow, the island endemics likely arose by 
dispersal from South America of a snake (or snakes) floating 


on flotsam in the Guiana current (arrows) 


HURRICANES 


Hurricanes are important for Caribbean biogeogra- 
phy because flotsam usually is the result of storm 
(King. 1962) which affects air 


currents. For geographic and climatic reasons, the 


aclion and water 
West Indies represents a hot spot for hurricanes, and 
the 


hurricanes and 


most storm tracks are similar in direction to 


surface current flow (Fig. 5). Thus. 


current flow work ae lo give a 


the West 


For example. it would be difficult for flotsam 


rectionality for. over-water. dispersal in 
Indies. 
(or organisms) to travel from Cuba to the Lesser 
Antilles: 
Although generate al least some 
flotsam in the West Indies (Censky et al., 1998), 


thunderstorms probably are responsible for nearly all 


the reverse would be much more likely. 


hurricanes may 


flotsam emanating from rivers in South America. 
Despite the directionality of dispersal mechanisms 
1 the West 


example. in 


Indies, exist. For 
1999, 
Lenny began near the Cayman Islands and ended near 


Antilles (Fig. 5). 
B 


| 


some exceptions 


the rare west-to-east Hurricane 


Nevis in the Lesser This unusual 


direction gave it the nickname “Lefty.” Although 
west-Lo-east tracks are rare (only two others during the 
last five years), this indicates that some unusual 


distribution patterns might be explained by atypical 


hurricanes. Another possible factor is the counter- 


Volume 93, Number 2 
2006 


Hedges 
Paleogeography of the Antilles 


T 


90 


78 


— 1 


Figure 5. 


Hurricane tracks in the West Indies (1995-1999). The hurricane track labeled “L” 


is Lenny (also called 


Lefty“) of 1999, Source of hurricane track data: National Oceanic and Atmospheric Administration. 


clockwise airflow of the hurricane vortex itself. 
Although the hurricane track may be moving in 


a westerly direction, strong surface winds to the south 
of the storm will be moving in the opposite (easterly) 


direction. This would not explain dispersal over long 


distances but might be a factor for short (tens of 


kilometers) distances. such as between islands 


(Powell, 1999). 
CONTEMPORARY MODELS OF HISTORICAL ZOOGEOGRAPHY 


In the several decades since plate tectonics became 
accepted, zoogeographers have had to consider the 
complexities introduced by past movements of islands. 
Early on, a new biogeographic theory was proposed 
the West 
originated on a proto-Antillean land area connecting 
North and South America in the late 
(Rosen, 1975, 1985). 


model, the present-day Antillean fauna did not arrive 


— 


suggesting. that much o Indian biota 
Cretaceous 


According to this vicariance 


by dispersal but instead was carried on the islands as 
they drifted, tectonically, to their current positions. 
Subsequent geologic models for the Caribbean differ 
considerably in details but not in the general notion of 
a proto-Antillean are. However, as discussed above, 
the paleogeographic details concerning which land 
areas were emergent (and when), critical for this 
vicariance model, are still poorly known. 
model have used 


Advocates of the vicariance 


ationships of organisms, almost ex- 
1980: 


phylogenetic re 


clusively, as evidence (Guyer & Savage, 


Crother & Guyer, 1996). This has been done with 
the assumption that the historical pattern of land area 
relationships should be reflected in the phylogeny of 
organisms that underwent. vicariance (Rosen, 1975, 


985). A similar application of this approach was 


— 


attempted for West Indian mammals (Dávalos, 2004). 
However. dispersal patterns often have directionality, 
as in the Caribbean, and can yield concordant 
phylogenies of groups (Hedges et al.. 1994; Hedges, 
1996b. c, d) Also, f land 


relationships in the West Indies is not yet possible 


reconstruction 


^ 


area 


jecause of uncertainty in details of the geologic 
history and paleogeography. 
Ot 


dispersal 


— 


1er approaches to clarifying the vicariance vs. 
problem integrate multiple sources of 
information, including times of divergence of species 
and groups. The fossil record has been used in an 
indirect manner to show that times of divergence (1.e., 
the origin of West Indian groups) are too recent to 


1981). 


support an origin by 
However, fossil-based divergence times are always 


vicariance | (Pregill, 
minimum estimates. Time estimates from molecular 
data pertain to the actual splitting event. because 
genetic differences begin to develop immediately 
following speciation. The most extensive molecular 
data available for testing these competing models of 
Caribbean biogeography have been immunological 
estimates of protein sequence divergence in serum 
albumin of West Indian vertebrates (Fig. 6). Gener- 
ally, time estimates from these data have been too 


recent to support proto-Antillean vicariance (Hass, 


238 Annals of the 
Missouri Botanical Garden 


Vicariance Dispersal 
= —— 4 
» — — 3 
x 5 
—.—|; 
— e 
9 
————$— —— 
11 
„ 13 
— — — 
— —— 15 
P» 17 
Oo — — 
S SS 19 
E , 
o o 21 
2 23 
© 
a 25 
„ 
29 
— 31 
— — 
33 
— 35 
e 
37 
Cretaceous | Paleocene Eocene Oligocene Miocene IP la 
| | 
80 60 40 20 
Million years ago 
Figure 6. me ies of origin for the 37 inde pende nt lineages of endemic West Indian amphibians and reptiles, mostly from 
imi al estimates of sequence divergence in the protein serum albumin (redrawn and modified with permission from 
Hedges, 19960). ia eans and error estimates (in some cases, ranges) are shown. The lineages our cated by numbers at right) 
(1) ne 5 Vschudi group. (2) Colostethus chalcopis Kaiser, Coloma & Gray, (3) Osteopilus IE (4) 
Ele ee uméril & Bibron, (5) Leptodactylus albilabris (Gunther) Boulenger, (6) Le 1 d fallax Müller. (7) 
‘rocodylus s Cuvier, (9) . ee ee Linnaeus, (9) € 1 Gray. (10) N Wiegmann, (11) Gy 1 
pleei Bocourl. (42) his Daudin, (13) Cyclura Harlan, (14 ana delicatissima Laurenti; (15) Leiocephalus Gray. (16) 


Aristelliger Cope. (17) Ph sllodactylus an her Gray. (18) 1 Vagler, (19) Tarentola Gray. (20) Antillean Ameiva 


Meyer, (21) C emir vanzol Baskin & Williams, (22) Cricosaura typica Gundlach & Peters, (23) Epicrates Wagler, (24) 
Chironius pe 3oulenger, (25) Mastigodryas bruesi (Barbour) Schwartz & Henderson, UA Clelia 7 
Underwood, (27) Lio, iis cursor Schwartz & Henderson group. (28) Ba iky andresensis Schw : Henderson. (29) 
nora variabili Duméril, Bibron & Duméril. (30) alsophine sna (31) lom pu Klauber, (32) 
Tropidophis Bibron in de la Sagra, (33) 20 Oppel. (34) Pas M iem (35) Trachemys Agassiz, (36) pelomedusid 
turtles, and (37) Geochelone Fitzinger. Note: P = Pleistocene uaternary. 


1991: Hedges et al., 1992: Hass et al.. 1993; Hedges — 1996a, c). Among the exceptions are organisms that 
et al., 1994; Hedges, 1996c: Hass et al., 2001). Also, — actively disperse, such as most freshwater fishes, bats, 
those divergence time estimates are not strongly and birds. In those cases, geographic distance rather 
clustered at one particular point in time but tend to be than current patterns appear to be more important 
more spread out, as one would expect from the action since such organisms show closer affinities with 
of a random mechanism such as over-water dispersal Central and North American faunas. 

(Hedges, 1996c). One possible exception is the frog The reduced higher-level taxonomic diversity 
genus Eleutherodactylus, which may have arrived in remains a major piece of evidence in the comparison 
the West Indies in the late Cretaceous or y | 
Paleogene (Hass & Hedges, 1991: Hedges, 1996c). It predicts that the Antillean fauna should be a cross 


=m 
= 


early of contemporary models. The vicariance moe 


is widespread, abundant, and diverse (i.e... not section of the continental fauna. Extinction is nol 
relictual), and the divergence times estimated be- excluded under the vicariance model, but no evidence 
tween mainland and Antillean species are in the late exists in the fossil record that many major continental 


Cretaceous or early Paleogene (Hass & Hedges, 1991; groups ever were present. For example, no primary 


Hedges et al., 1992). freshwater fishes (intolerant of salt water), salaman- 
The phylogenetic relationships of West Indian ders, caecilians, marsupials, carnivores, or lago- 


vertebrates reveal that most have their closest morphs are known to have occurred in the islands. 


relatives in South America (Hass. 1991; Hedges. Most families of frogs, turtles, and snakes are also 


Volume 93, Number 2 
2006 


Hedges 
Paleogeography of the Antilles 


missing. Ungulates with their large bones fossilize 
well, yet no record of any ungulate exists aside from 
the early ca. 50 Ma rhinocerotoid from Jamaica, as 
noted above. 

Other that 
composition extended into the past comes from the 


evidence this peculiar taxonomic 


unusually large radiations of some groups, such as 


— 


ground sloths, capromyid rodents, eleutherodactyline 
frogs, and anoline and sphaerodactyline lizards. This 
is to be expected when niches are left vacant and is 
characteristic of oceanic islands never connected 

continents (Williams, 1989; Woods, 1990; 
1996a). Probably as a result of these open niches, 


some of the largest and smallest species of various 


Hedges, 


groups occur on islands, including the West Indies 
(Estrada & Hedges, 1996; Hedges & Thomas, 2001). 
dition to dispersal and vicariance, organisms 
Greater. Antilles by 
a land bridge (Rosen, 1975; Holcombe & 
1990: Woods, 1990; MacPhee & Iturralde-Vinent, 
1994, 1995; Tturralde-Vinent & MacPhee, 1999). 
Unlike earlier land bridge models, proposed before 
plate tectonics, this land bridge is based on modern 


geology. It proposes that the. Aves Ridge, now below 


[om 


In ac 


may have reached the way of 


Edgar. 


sea level, formed a chain of islands or dry land 


America and the 


connection between northern South 
i the 


Antilles in the mid-Cenozoic. However. 


(unbroken) land bridge 


Greater 
difference between a stable 
and a chain of islands is important for biogeography. 
Even narrow land bridge corridors, such as the current 
major interchange. of 
Williams, 1989). A 


hand, acts as a filter 


— 


Isthmus of Panama, allow a 
faunas (Stehli & Webb, 1985; 
chain of islands, on the other 
selecting. for species with more optimal dispersal 
abilities. 
been at least a chain of islands, much like the Lesser 


The Aves Ridge has been assumed to have 


Antilles, that enhanced the over-water dispersal 

organisms to the Greater Antilles (Rosen, 1975; Perfit 
& Williams, 1989). 
Aves Ridge was a unbroken land 
(Iturralde-Vinent & MacPhee, 1999) is of interest 


This model proposes only a narrow window of 


Thus, the recent assertion that the 


dry and bridge 
here. 
time (35-33 Ma) when this terrestrial 
(Iturralde-Vinent & MacPhee, 1999) 

In the process of providing evidence for their dry 
land bridge & MacPhee 


(1999) eliminated the possibility of proto-Antillean 


corridor existed 


model, [turralde-Vinent 


arguments against over-waler 


the Aves Ridge land 


vicariance and 


gave 


dispersal. They suggested that 
bridge plaved a major role in the origin of West Indian 
terrestrial vertebrates. However, the same evidence 
that supports over-water peral for much of the 
West Indian fauna and argues against proto-Antillean 
vicariance also argues against the Aves Ridge land 
This includes a reduced higher-level 


bridge model. 


taxonomie diversity (now and in the past), unusually 


large adaptive radiations of groups present, and 
divergence time estimates that are not strongly 


clustered at one point in time. The existence of a land 
bridge would instead predict a more continental-like 


auna (at least in the Paleogene fossil record) and 


divergence time estimates clustered strongly around 
35—33 Ma. Nevertheless, because the Aves Ridge was 
probably a chain of islands, it should be expected to 
dispersal at different times in the 
much like the Lesser 


have enhanced 
Cenozolc, mid-Cenozoic, 
Antilles in the Late Cenozoic. In that respect it may 


have facilitated dispersal of mammals and other 


x. LY 
E. g., 


groups from South America to the Greater. Antilles, 
as suggested by previous researchers (Holcombe & 
Edgar. 1990; Woods, 1990). However, this is different 
from the Iturralde- 
Vinent & 

yermitted a 
American biota to enter the 


dry land connection claimed by 
MacPhee (1999), 

diverse section 
Antilles. 


cross section is not seen in the fauna. 


have 
of South 
Such a diverse 


which would 


cross the 


The dry land bridge initially was proposed as a way 


to explain the origin of Antillean land mammals 


~ 


(MacPhee &  Hurralde-Vinent, 1994, 1995). For 
example, Iturralde-Vinent € MacPhee (1999: 58) 
stated. that they “doubt that the sloths swam the 


distance.” Yet, as they acknowledge, some species of 
ground sloths were highly adapted to marine life 


(Muizon & McDonald, 1995), this 


proposed biological need for a dry 


thus removing 


land bridge. 
Unfortunately, the creation of a dry land bridge also 
has implications for other aspects of the paleogeog- 


raphy of Iturralde-Vinent & MacPhee (1999): as 
previously mentioned it is used to suggest the 


direction of water current flow in the past and 


influence on dispersal 
Iturralde-Vinent & MacPhee (1999) also claim that 


no permanent landmasses were present in the 


he middle Eocene, eliminating 


— 


Caribbean prior to 
the possibility of proto-Antillean vicariance. Although 
the absence of direct evidence for an uninterrupted 
land occurrence is used as support of that claim, there 
is also an of against 
uninterrupted land occurrence. In fact, such evidence 
would be difficult to obtain, as it would require nearly 


abse nce direct evidence 


continuous sedimentation. present from all areas and 
sampled at all time intervals. Because some land 
areas were at least close to the ocean surface (sea 
level) during part of that time, as acknowledged by 
Iturralde-Vinent & MacPhee (1999), the possibility of 
uninterrupted land areas cannot be completely ruled 
out. Moreover, organisms are surprisingly resilient to 
changing landscapes and are known to survive as 
years in Island 


continuous lineages for millions of 


chains such as the Hawaiian Islands and Galapagos 


240 


Annals of the 
Missouri Botanical Garden 


— 


Islands these 


(Sequeira et al., 2000). In cases, 
temporal continuity in the long term is likely 


accomplished by dispersal among ephemeral islands 


in the short term. 


Thus, the biota of an archipelago 
can be maintained over time even while individual 


islands rise and fall. 
Tur PROBLEM OF RELICTUAL GROUPS 


Two endemic groups of West Indian animals, the 
giant shrews (Solenodon Brandt) and Cuban night 
lizards (Cricosaura Gundlach and Peters). have been 
suggested to have arisen through proto- Antillean 
Hedges et al., 1991: 
1992: Roca et al., 2004). However, the 


both 


vicariance (MacFadden, 1980 


Hedges et al., 


distributions o groups, including mainland 


relatives, have receded through time (i.e. are 
relictual), which presents a problem for interpreting 
their biogeographic history. Before discussing these 


biogeographic implications, some clarification is 


needed concerning the estimated time of divergence 
for the Cuban night lizards from their closest relatives 
in the family Xantusiidae. 

The lizard family Xantusiidae occurs in North and 
West 


species is known from the West Indies, Cricosaura 


Central America and the Indies. Only one 


typica Gundlach and Peters, restricted to a small 
M 


region of eastern. Cuba. In. the original studies 
reporting the xantusiid lizard sequence data (Hedges 
et al, 1991; Hedges & 1993), 


suggested that Cricosaura might represent a remnant 


Bezy, the authors 
of the proto-Antillean fauna, based on molecular 
phylogeny and fossil data. Although a molecular clock 
analysis was not explicitly used, they demonstrated 
phylogenetically that the West Indian lineage (Crico- 
most ancient in the family, and its 


saura) was the 


divergence from other xantusiids must have logically 
` Palaeoxantusia fera. Hecht (ca. 
1963; 


pre dated fossils of 
60 Ma) assigned to the other lineage 
Schatzinger, 1980 


Savage, 
1991) 


reanalyzed 


; Hedges et al., 


Those sequence data were later. by 
Vicario et al. (2003), who recovered the same topology 
and performed a molecular clock analysis, arriving at 
a different conclusion concerning the time of origin of 
the Cricosaura lineage. Although much emphasis was 
eiven to this difference. it was the sole result of those 
60 Ma fossils of Palaeox- 


antusta fera. from consideration, 


authors removing the ca. 
without explaining 
why they disagreed with earlier authorities concerning 
15 Ma 


a mainland. divergence. (Xantusia 


that fossil. Using the next earliest fossils (ca. 


for calibration of 
Baird/Lepidophyma | Duméril), their molecular clock 
time interval for the Cricosaura divergence from other 
64.9-43 Ma. They stated that 


consistent. with the Hedges et 


xantusiids was this 


estimate “is nol 


(1992) proposal that the Cricosaura line survived the 
the K/T 


bolide impact : (Cretaceous/Paleogene) 
t 65 MYBP, 
estimate is nearly 100,000 vears after that momentous 
003: 257). However, there are 
(1) they 


to molecular 


boundary : as the earliest. limit of our 


event” (Vicario et al., 2 


several problems with this interpretation: 


ascribed considerably more accuracy 
the difference between 
64.9 and 65.0 Ma is only 0.15%), especially given the 
modest sequence data set, (2) they 
off by 17 
nullifies the difference of 100,000 years, and (3 


interpreted the 


clocks than is justified. (e.g.. 


conceded that the 
which 
) they 


time 


calibration could be million years, 


divergence date as a fixed 
estimate, when it is more correctly interpreted. ¿ 


2004). 


data were reanalyzed more recently by 


S 
a minimum (Hedges & Kumar, 


The same 


Roca et al. (2004), who disagreed with Vicario et al. 
(2003) and supported the 


vergence for 


earlier. (Cretaceous). di- 
16 Ma. 
a different method of time estimation (Thorne et al., 
1998) with the same 13 Ma for 


YantustalLepidophyma) as minimum and added Pa- 


Cricosaura, al They used 


a 


calibration point 


laeoxantusta fera (00 Ma) as maximum. To explore the 
effect. of. calibration point on this time estimate, | 
reanalyzed the data using the same methods but used 
instead = Cricosaura/non-Caribbean divergence for 
calibration (a Bayesian estimate can be obtained even 
for calibrated nodes). Regardless of whether that node 
s treated as a fixed calibration or minimum 


43 Ma or 60 Ma, and 


the ingroup root prior (50-100 Ma), the confidence 


calibration, al regardless. of 


interval in all cases extends into the Cretaceous. If 
Palaeoxantusia fera is correctly assigned to the 
FP 


— 


yma lineage, then a fixed calibra- 
tion of 60 Ma for the Cricosaura/non-Caribbean split 


is the preferred method, yielding a time of 75 Ma (00— 


116 Ma) using an ingroup prior of 75 Ma. similar te 
the result of Roca et al. (2004). 


be interpreted as a minimum time estimate and nol 


This date should then 


a mean time estimate in drawing conclusions. 
The time of divergence of the West Indian shrews 
relatives on the 

16 Ma Ma) 

using a large sequence data set of 16 nuclear and 

2004). A direct 

divergence for the 

that the West 


Indian representatives originated by vicariance on the 


(Solenodon) from their closest 


mainland was estimated to be (72-8 


three mitochondrial genes (Roca et al., 


interpretation of this ancient 


shrews and for xantusiid lizards is 


proto-Antilles in the late Cretaceous, whether it was 
dry land or an island chain, and survived the bolide 
impact in that region at 65 Ma. However, both of these 
that 


alternative hypotheses should be considered. Nantu- 


groups appear to be relictual, which means 


siid lizards are pate hy 1 In distribution and have Cc le arly 


receded from a wider distribution in the past (Hedges 


Volume 93, Number 2 
2006 


Hedges 
Paleogeography of the Antilles 


2004). 


there is uncertainty regarding the identi- 


etal., 1991: Sinclair et al., 
al. (2004), 


fication of the closest relatives of West Indian shrews. 


As noted by Roca et 


Nonetheless, some North American soricomorphs 
(e. g., Apternodus Matthew), all extinct, have been 
proposed, and the general conclusion is that soleno- 
dontoids and their relatives represent a relictual group 
of animals (MacFadden, 1980). 

The primary alternative hypothesis raised by these 
relictual groups, such as West Indian shrews and 
xantusiid lizards, is that they arose by dispersal from 
Paleogene and not by 
199064). 


According to this hypothesis, the phylogenetic. di- 


the mainland in the early 


vicariance in the late Cretaceous (Hedges, 


vergence still occurred in the Cretaceous, as estimat- 
ed by molecular clocks, but it resulted in descendant 


lineages that initially occupied only mainland areas. 


dispersal 
h 


the mainland. representatives of those “West Indian” 


At some later time, in the Paleogene, 


occurred from the mainland to the West Indies, wi 


lineages later becoming extinct. Advantages of this 
hypothesis are that it does not require the presence of 
emergent land in the Antilles since the Cretaceous 
and survival of the local effects of the bolide impact 


(65 Ma) by the Antillean group. Nonetheless, a better 


fossil record of mainland relatives would help to 
distinguish between this hypothesis and one of 
vicariance. Currently, neither hypothesis can be 


strongly favored. 


DISPERSAL FROM ISLANDS TO MAINLAND 


In discussions of island biogeography, dispersal 
from the 
mainland Moore, 


2005). However, the Greater Antilles are large and old 


usually considered to be in one direction: 


to the island islands (Cox & 


areas for 


land masses capable of being source 
dispersal. At least three Antillean groups appear to 


have colonized mainland areas. At some time during 
the mid-Paleogene, estimated as 37 Ma with a molec- 


clock, 


dispersed. from Cuba 


ular one or more frogs (Eleutherodactylus) 


Central America (Hedges, 1989; Hass & Hedges, 


1991: Hedges et al., 1992). The founders evolved into 
a clade of Eleutherodactylus species recognized today 


1989). 
American pond turtles of the genus 
1996) and a large 
2005) are 


also considered to have originated by dispersal from 


as the subgenus Syrrhophus Cope (Hedges. 
Some Central 
Trachemys Agassiz (Seidel, 199€ 
clade of anoline lizards (Nicholson et al., 


the Antilles to the mainland. The West Indies was 


probably the source for other mainland groups as well 


and this should be considered in any discussions of 


Caribbean biogeography. 


to the adjacent mainland of 


CONCLUSIONS 


This brief overview of West Indian biogeography 


focuses on the major competing models. Anyone 


a 


his fiel 


familiar with the literature will admit that 
generates contrasting opinions. In part, this is because 
of the complexity of the problem and evidence that is 


often limited. All agree that over-water dispersal can 


+ 


and has occurred, but some dispute the degree of the 
contribution of this mechanism. The paleogeographic 
basis for the other two proposed mechanisms, proto- 
Antillean vicariance and a mid-Cenozoie land bridge, 
has not yet been confirmed. In the author's opinion, 
both of those mechanisms remain as possibilities, 
however remote, and the author disagrees with the 
assertion that proto-Antillean vicariance must be ruled 
out (Iturralde-Vinent & MacPhee, 1999). 


| | e 
JALCOLCOLTAPH¢ 
| pepe 


At the same 


evidence does not currently 


time, 
support the Aves Ridge land bridge model, in contrast 
to claims by its strongest advocates [turralde-Vinent 
and MacPhee (1999), In their advocacy of that model, 
with the corresponding need to construct a mid- 


Cenozoic walkway for land mammals, they have blurred 


— 


the distinction between paleogeography (past land- 


scapes) and historical biogeography (past distributions 


— 


of organisms). They refer to their model as “paleoge- 


ography” but in reality it is a biogeographic model 


— 
n 


biased by their desire to create a corridor for mamma 

when there is no physical evidence to support such an 

Biologists and biogeogra- 
aM 


unbroken corridor of d. 
p 


to evaluate their evidenc e from organisms 


iers need unbiased paleogeogray reconstructions 

As noted here, and in more detail Nod (Hedges, 
1996a, 
origin by over-water dispersal for most of the West 
) 


the reduced higher-level taxonomic composition of the 


c, 2001), the weight of the evidence supports an 


— 


Indian vertebrate fauna. This conclusion stems from ( 
fauna (now and in the past) (2) the presence of 
unusually large adaptive radiations, (3) the finding that 
closest relatives of most Antillean groups are from South 
North 


America (active dispersers), and (4) the finding o 


America (passive dispersers) Central and 


— 


divergence time estimates that are nol strongly 
clustered. The first point was noted over a century ago 
and has been dubbed more recently as the central 
problem (Williams, 1989). The second point is related to 
the first, and infers that existing groups have taken over 
niches of the missing groups (thus suggesting that the 
missing groups have been missing for a long time and 
not just recently extinct). The third point is consistent 
with the directionality of water current flowing through 
T 


the prediction of over-water dispersal. 


ie fourth point follows 


the region (now and in the past). 


Despite this general (major) pattern, several 


Antillean groups may have originated on the proto- 


242 


Annals of the 
Missouri Botanical Garden 


Antilles in the Late Cretaceous or dispersed there in 
the early Paleogene. 
(fish), 


Cricosaura, and the insectivore mammals. 


These include the Cuban gars 
eleutherodactyline frogs, the xantusiid lizard 
Additional 
data are needed for these and other so-called ancient 


groups. In the future, divergence time estimates will 
be based on sequence data from multiple genes, rather 


than immunological estimates from a single protein 


, 


and therefore will increase in 


precision, Such time 
estimates from many Antillean groups will permit 


a more accurate test of these competing models. 


Literature Cited 


Barbour, T. 1916. Some remarks - p iu ws “climate 


and e Coe — New | Acad. : 1-15. 
Censky, Hodge & J. Dudley. po 8. Ove T- Pid 
dispe a a paris due to hurricanes. Nature 395: 


Comer, J. B. 197 

09: 1251— . 
C. B. & P. 
Ecological and Evolutionary 
chusetts. 


Genesis of Jamaican 1 Econ. e e 


Cox, 


= 


Moore. 2005. Biogeography: An 


1 
Approach. Malden, Massa- 


1996, 


persal-vicariance debat 


Crother, B. I. & C. Guyer. Caribbean historical 
biogeography: was the dis 
natec by 
140—416 

Darlington, P. J. 1957. Zoogeography: 
Distribution of Animals. Wiley 

Darwin, €. 1859. The 
London. 

Dávalos. L. M. 2004 
Caribbean mammals. 

deArmas, I. F. 198 
aste: oon UE antillana. Poeyana 238: 

Dengo, C. & J. E. ( 


imi- 
an 1 bolide? Herpe lass 52: 


The Geographical 
. New York. 
Origin of Species. John Murray. 


ea and biogeography of 
Biol. J. Linn. Soc. 81: 373 

982. Algunos aspectos zoogeográficos de la 
1-17 

ase (editors). 1990. The Ge 188 of North 


America. Volume H. The Caribbean Region. The Geo- 
logical ae iely of sia ‘a, Boulde T. 

Demme: D. P... J. Em . Portell, . Donovan & K. 
S. Schi writ 1997. i st West bs land mammal: 


rhinoceratoid ungulate from the Eocene of Jamaica. J. 
Vertebral ile e Paleonto „ 7 6 
Donnelly, T. W. 1980, History 
terrestrial 


of marine barrier ars and 
connections: i 
ference from p nt analysis. Pp. 103-117 7 
A. Woods (e ditor) igh of the West Indies: Past, 
Present, hill €i 
Caribbean biogeography: 
the 


In 


and Future. Sand rane Press, Gainesville. 
Xx 


( 


= 


geological consid- 
erations on i 
Pp. 595-609 
Aspects of 


jr of vieariance vs dispersal. 
A. (editor). 


1 


atolli Bioge O al 


Insularity. Nazionale dei Lin 
Rome. 
1992. 


Mesoamerica. Pp. 


Geological setting and tectonic 
1-13 Quintero & 


Panama and 


history of 
Niello 
Oxford 


in 
(editors). Insects. of Mesoamerica. 


Univ. Press, Oxford. 
Donovan, S. K. & T. A. Jackson acne 1994. Caribbean 
Geology: An Introduction. Unive sity of the West Indies 


\ssociation, alca. 
W., K. C. PASS A. E T ‘cunningham, ES C. Hine 


. Hallock & 


on 


. Robinson. 
Garibbean 
Mlantic i 


constraints 1 between 


and Pacific Oceans over the past 40 Million 


years. Pp. 169-191 in T. J. Crowley 


Tectonic 


& K. C. 


Conditions 


\ Burke 

(editors). Boundary for Climate 
* ‘tions. rur Univ. | New York. 

Estrada. A. R. & . Hedges. 1996. At the lower size limit 
in tetrapods: diminutive frog from Cuba (Lepto- 
dactylidae: lester Copeia 1996: 852-856 

Goodfriend, A. K 1990. Paleontology 
a of two e ede transitions by hy 


in the Bahamian land snail Cerion. 
L 1851 


ress, 


Gould. od 
b ridization 


Science 274: 18 394-1897. 


— 


osse, P. H. 5 Naturalist’s 


Sojurn in Jamaica. 
Longman, Brown, Green, and Longmans, London. 
s. 


Gould, S. J. & D. vodruff. 1986. Evolution and 
systematies of Cerion (Mollusca, Pulmonata) on New 
Ad Island—a Len revision. Bull. Amer. Mus. 
Nat. Hist. 182: 391—19 

Graham. *. 20022. E RT models and Cenozoic 
a mts of the Caribbean region. Syst. Bot. 

378 Er -386. 
. 2003b. Historical phylogeogi iphy of the Greater 
Antilles. Brittonia 55: 357— 
Grimaldi, a A. 1995. The sol os amber. Pp. 203— 
217 . B. Anderson & J. C. Crelling (editors). Amber. 


Resinite. ni F ossil Resins. 


American Chemical Society. 
W gr 


— 
— 


Guppy. | 1917. Plants, Seeds, and Currents in the West 
Indies n Azores. Williams and po on. 
Guyer, C. & J. M. 96 


Londo 


I 


Savage. Cladistic re eli tionship 


among PX auria: Ieuanidae). Syst. Zool. 35: 509-53 | 

Hass, C. 199]. Evolution and n of West 

Indian share (Sauria: Gekkonidae): A molecular 
appro: J. Zool. 225: 525-561. 

& S. B. Hedges. 1991. Albumin evolution in West 

Indian frogs of the genus Eleutherodactylus: Caribbean 


biogeography and a oe 15 = 
logical clock. E i 413 gc 


.. 1 


albumin immuno- 


1255 Molecular insights 
1 and biogeogr: p of West Indian 
anoline lizards, Biochem. Syst. Ecol. 21: 97-114. 

L. R. Maxson & S. B. Hedges. dd Relationships 
and en ‘rgence limes 


into the 


Vest Indian oes and 
reptiles: insights vp albumin immunology. Pp. 157-174 
in C. A. Woods & F. E. Sergile (editors), Biogeog sah of 
the West Indies: 1 and e MR RC 
Boca Raton, F 


3 


Florida. 


Heatwole. II. & R. Levins. 1972. Biogeography of the Puerto 
ican Bank: flotsam transport of terrestrial animals. 
Ecology 53: 12-117 


Hedges, S. B. 1989 Paluun and biogeography of West 

Indian frogs of the genus £ „ slow-evolving 
loci and the major dn 305-371 0 in C. A. Woods 
(editor), Piogeugruphy dq of 


and Future. Sandhill € 


| Present. 


19964. TUM ioe al Wes Indian 
vertebrates. Annual Rev. Ecol. : 163-19 

1996b. More on € 1 zooge RM is elo- 
logica 52: 622-024. 

1996c. 


The origin of West Indian E os and 
repliles. Pp. 95-127 in R. Powell & R. W. Henderson 
Contributions to est. Indian  Herpeto- 
> A Tribute to Albert Schwartz. Society for the Study 
of Amphibians and Reptiles, Ithaca. 
10904. 


biogeography. 
200 


—— cartance and dispersal in Caribbean 


Herp 1 52: 466—473 


1 1 


a MH. Caribbean biogeography: an liga: Pp. 
F. E. Se 


204 . Woods & E. Sergile (editors). Bioge raphy 
of the Y ES sl Indies: Patterns and Perspectives. CRC Pr 


Boca Raton, Florida. 


Volume 93, Number 2 
2006 


Hedges 
Paleogeography of the Antilles 


2 


i. L. Bezy. 
concern for data and analysis. Molec. Phylogen. Evol. 

74 
& S. Kumar. 2004. Precision of molecular 
Genet. 20: 242-247. 
2001. 
F a new diminutive lizard from the West Indies. 
Caribbean J. Sci. 37: 168-173. 

R. L. Bezy & L. R. 


re lationships an 1d bioge ography of 


ip Trends 


& R. Thomas. At the lower size limit in 


Maxson. 1991. Phylogenetic 

lizards 

inferred from n mitoc hondrial DNA sequences. Molec. Biol. 
161—180 


Evol. 8 


ntusiid 


Hass & L. R. Maxson. 1992. Caribbean 
Y Molecular evidence for 5 in i 


Acad. S 


— n 


Indian terrestrial vertebrates. Proc. Natl. 
89: 1909-191: 


& 


— ee . 1994. 
ge a of he cari san. Cladistics 10: < 
Henderson. R. W. 


hortulanus 1 of Neotropical tree boas. Caribbean J. 
Sci. 33: 198-2 


Maxson. Towards a bio- 


A taxonomic review a ike Corals 


. Hedges. 1995. Origin of West Indian 
populations of the geographically widespread boa Corallus 
enydris inferred. from mitochondrial. DNA sequences. 
Molec. 4: 88-92. 


> 


Phylozen. Evol. 
. W. P. Mi 


G. Puorto Bourgeois. 


— icucci. W. 
1995. E MEM al correlates and patterns in the pcd 


tion of Ned tropical boines 50 nles: pA a pre- 


ist. oe 
on 
impact deposits in the 
: 843-847. 

1990. 
pem » Caribe zan ridges and rises with 


liminary assessment. Herpe tol. Nal. 


Hildebrand. A. 8 on Proximal 


Cretaceous- bun 11 
Caribbean. Science 105 


Holcombe, T. L. & N. T. Edgar. Lale Cretaceous and 

Cenozoic 

special reference to paleogeography. Pp. 611-626 in ^. 

Azzaroli (editor), Biogeographic ‘al Aspects of. Insularity. 

Accadem. Nazionale dei Lincei, Rome. 
C. & P 


Huebeck, Mann. 1985. 


Cenozoic tectonic history of the 


Structural geology am 
southeastern termination 
Cordillera Central, Dominican Republic. 
Pa seol. Soc 
hara Vent u. A. “D. E. 
| 


origin of 


ol e Special 


Mac Phe e. 1996. 


Dominican amber. 


Age and 


Science 
13% 1850-1852. 


. 1999. Paleogeography of the Caribbean 
: B 


region: implications for Cenozoic biogeography. Bull. 
Amer. Mus. Nat. Hist. 238: 1-95 
James. S. O. ine new species of Dichogaster 


(Olizoc haeta, Megase E x ) a Guadeloupe (French 


West Inc E Us cripla 25: 21—34. 

King. F. W. occurrence of rafts for dispersal of 
land m into 15 West Indies. Quart. J. Florida Acad. 
Sci. 2 pos 52. 

pl 


2000. Alsophis vudú s Brown Racer). 
31: 244 
1902 


= 


)verwaler RS Herpetol. 
& W. V 


Kring, D. X. é oynton. Petroger mesis of an 
igite hei caring melt rock in the C 


a and its 


re lationship to K/T impact sphe mh in Haiti. Nature 358: 
41-144. 

rud J. 1988. „ of 
nsects. Cornell Univ. Press, Ithac: 


1992. Phylogeny 


K. (editor). Caribbean 


and revision of the Platynus 


degallieri species group (Coleoptera. Carabidae. Platini). 
zull. Amer. Mus. Nat. Hist. 214: 1-1 14. 
MacFadden, B. 1980. Rafting mammals or drifting islands? 
Biogeography of the Greater Antillean 


inseclivores 
—22. 


Nesophontes and Solenodon. J. Biogeogr. 7: 


1993. Phylogeny of xantusiid a 


time 


Davis & W. B 


MacPhee, R. D. E. & D. 


in Dominig an amber. 
. Ho 


A. Grimaldi. vdd Mammal bones 
1 0 380: 489— 

. New dd remains of 
the Quale rnary s an 1 Xenothrix 


rovitz. 
megregori 
Xenotrichini, Callicebinae, Pitheciidae), with a reconsid- 
eration of the Aotus hypothesis. Mus. Novit. 3434 


Amer. 
al. 

LN: M mp 
the Greater early Miocene sloth 
1 irthra. Megalonc 5 uba. Amer. Mus. Novit. 

: 1-13. 
& 


1994. First Tertiary land 


mammal f rom ntilles: an 


Antillean 
ym Cuba and 


——— & —— oe Origin of the Greater 
land mammal funds . New Tertiary Dr n 

E a Novit. 314.1: 

Affinities and origin q the Antillean 

Amer. 29: 657—660. 

Maurrassee, F. J.-M. R. 1991. Impacts, tsunamis and the 

aitian Cretaceous-Tertiary boundary layer. 5 

1690-1693. 

Meschede. M. & W. Frish. 


the Mesozoie and Early C 5 history of the Caribbean 


cience 252: 
1998. A plate-tectonic model for 


. Tectonophysics 296: 209-291. 
L. W. 19€ 


plat 
1 t 7 1998. A review of Bahamian ant 
5 5 1 Formicidae) biogeography. J. Biogeogr. 
50 


"o C. ^ & H. G. MeDonald. 1995. An aquatic sloth 
from the Pliocene " Peru. Nature 375: Wi i. Te 

Nicholson, K. E.. R. E. Glor, J. J. Kolbe, arson, S. B. 
He K J. B. Losos. 2005. 1 5 by 


island lizards. J. Biogeogr. 32: 929—938. 


Niedbala, W. 2004. Ptyctimous mites E Acari. Oribatida) of 
the neotropical region. Ann. Zool. 54: 1-288. 
Paulay, G. 1994. Biodiversity on oceanic islands: its origin 


9 Zool. 34-144. 
.W ama. 


11 1 in the 


and e A Lion. 
ih " Geological constraints 
volution of the 


Caribbean Sea and its islands. Pp. T in 
Woods (editor). Biogeography of the t Indies: Past. 
Present, Future. Sandhill Crane E ress. Gaine: sville. 


Pindell. J. L. 1994. E a a the Gulf of Mexico and the 


1 an. s Pp. 13—39 in . Donovan & T. A. Jackson 
(editors). Caribbe an Ge e og An 1 1 The 
University of th re West Indies as Publishers” Association, 


Kingston, Jan 
&l 


om ennan. 2002. 


Plate model for the Caribbean: 


httpz/www.ig.utexas.edu/C aribPlate orum/pindell/pindell_ 
5) 


kennan. htm (accessed 17 May 200 
Poinar, G. i. Poinar. 1999, The Abe Forest. Princeton 
niv. Press. Princeton, New Jersev. 
Powell. R. 1999. ud tology of Navassa Island, West 
Caribbean J. Sei. 35: 1-13 


J: 
e 


1981. 
hypothesis of Caribbean biogeography and its application 
to West Syst. Zool. 
147—155 


1986. 1 size 


pu 1 of the vicariance 


Indian terrestrial. vertebrates. 


of insular lizards 
. Evolution 40: 997 
1999, Eocene lizard from Jasisie a. Herpetologica 


: a pattern. of 
Puig dwarfisi )08. 
999 


— 8. L Olson. 1981. Zooge cd of West Indian 
vertebrates in relation to Pleistocene climatic cycles. 
Ecol. Syst. 12: 7 5-98. 


Annual Rev. 
R D. W. Steadman. L. K. Gordon. F. V 
Hilgartner. 1992. Living and late Holocene 


2 Cockpit 


. L Crombie. 


rales, and the 


Atoll Res. 


fossil verteb e 


Country, Jamaica. Bull. 353: 


244 


Annals of the 
Missouri Botanical Garden 


Roca, A. n > K. 
Maria, 
Mesozoic origin for West Indian insectivores. 
649-51. 


Bar-Gal, iz irik. Helgen, R. 
Springer, S. J. 5 & ^ A Me 2004. 
Nature 429: 


hosen, D. E. 1975. vicariance model of Caribbean 
biogeography. Syst. Zool. 24: 431-464. 
1985. Geological a rarchies and id 


I 

congruence in the Caribbean. Ann. Missouri Bot. Gard. 
12: 036-059 

Santiago-Valentin, E. & 2004. 


R. Historical 
biogeography of Caribbean 5 995 introduction to current 


. Olmstead. 
knowledge and possibilities from a phylogenetic perspec- 
tive. Taxon 53: 299-319 

is J. M. 19 

ing genera. Contrib, Sci. Los Angeles County Mus. 71: 


63. Bon the lizard family Xantusidae. 
s 

Scharff. R. F. 1912. Distribution and 
America. Macmillan, New York 

Sehatzinger, R. A. 1980. New 
TO 


origin of life i 


species of Paleoxantusia 
Sauria) from Uintan (Eocene) i San Diego Co 
60-17 

Tones, J. renee J. Chase. J. 
2003. 
(Isoptera : 


talifornia. J. Paleontol. 54: 
Se he ffrahn, R. H., S. C. 
Mangold & N. Y. Su. 


noles on the 


~ 


Taxonomy, a and 


* lermiles Kalotermitidae, Rhino- 
Termitidae) of Puerto Rico and the 5 
Amer. 96: 


Historical 


termitidae, IS Virgin 


Islands. Ann. Entomol. Soc. 

Schuchert, C 1935 

190 0 an n gion. 

Seidel, 88. Revision of a "iei an emydid turtles 
(Te sili " jus r. Mus. Novit. III. 


Current status of E. 1 of the West 


Geology of the 


y \ntillean- 
John Wiley and Sons, New York. 


ee 05 in the genus mydidae). 
Pp. 169-174 in R. Powell & 


Contributions to West Indian a 1 A Tribute to 


S hemys (E 


. Henderson (editors), 


Albert Schwartz. Society for the Study of Amphibians and 
Reptiles, Itha 
Sequeira, X. S. a. A. V. Seataglini, V. 
K B. I 2000. Are 
Galapaganus weevils older than the Galapagos Islands 
they inhabit? Heredity 85: 20-29 


Lanteri, M. 


Farrell. flightless 


Simpson, G. G. 1956. Hoogeography of West Indian land 
T 


mammals; Amer. Mus. Nox : 1-28. 
B 


Sinclai , E. ORO amarillo, 


history: Genus is (Squamata : Xantusiidae). Amer. 
Naturalist 164: 4.14. 
Sloane, H. 172! 


Nieves, S. C " iristophers, 


> voyage to the islands Madera, Barbados, 
with the 
four-footed beasts, 


and d Jamaic a. natural 


history of the herbs and trees, fishes, 
birds, insects, reptiles, Qe, of the last of those islands. H. 
Sloane, London. 
Smith. D. S.. L. D. Miller & J. Y. Miller. 
Butterflies of the West Indies and South Florida. 
Univ. Press, Oxford. 
Stehl F. C. & D. S. Webb (editors). 1985. The 
American e Interchange. a lenum Press. 
Thorne. J. L.. H. Kishino € J. S. Painter. 19€ 
the rate ol cai of 1185 ix 
Molec. Biol. 
Vicario, S., 2003. 
“night” lizards: a puzzling phylogenetic problem revisited 
mtDNA 


1994. The 
Oxford 


Greal 


Yo 
. Estimating 
of molec salar evolution. 
Evol. 15: 1647-1657. 
s & 


files Xantusiid 


using likelihood-based Bayesian methods on 
Phylogen. n 26: 243-201. 


. Island Li 


1050. Old 15 ms and new dr aad sin 


Harper, New York 


Wesl indian MR Pp. I-4 oods 


(editor), Biogeography of the : Indies: Past. Preso nt. 
and bows Sandhill Crane Press, Gainesville. 


Wilson, E. O. 1985. 


wed ant fauna: 


| West 


amber. 


Invasion and extinction in the 
evidence. from Dominican 
Science 229: 265-201. 

1988. The O of the West Indian ants 
(Hymenoptera: Formicidae). Pp. 214—230 in J. K. 
(editor taribbean Insects. Cornell Univ. 
Press, is a. 

. 1990. The 
d 


Liebherr 


. Loogeography of € 


fossil and recent mammals of the 


an analysis of the origin, evolution, ane 


extinction of an insular fauna. P p. 042-080 in A. Azzaroli 


(editor). Biog e ral Aspects of Insularity. Aecadem. 
Nazionale del Li . Rom 


THE GREAT AMERICAN BIOTIC S. David Webb? 
INTERCHANGE: PATTERNS > 
AND PROCESSES! 


ABSTRACT 


When the Panamanian land bridge was e pas ed about 2.7 Ma, it triggered the Great American Biotic ra (GABI), 
a major mingling of land mammal faunas between North and South America. Four families of northern immigrants 
(Procyonidae, Felidae, r ssuidae, and Camelidae) diversified at moderate rates, while four others, Canidae, Mustelidae, 
Cervidae, and especially Muridae, evolved explosiv ely. As a consequence, half of living South American genera are 


descendants of ae ‘rn immigrants. The other major consequence of the interchange was the conquest of tropical North 
America by immigrants from Amazonia, an episode that justifies the term Neotropic ‘al Realm. 
Key words: biostratigraphy, evolutionary rates, isthmian land bridge, mammals, Neotropical Realm. 


RESUMEN 


Cuando el corredor terrestre ews se ubicó hace cerca de 2.7 Ma, AS el Gran Intere ambio Biótico Americano 


GABI), una mezcla importante de las faunas de mamiferos terrestres en » América del Norte y del Sur. Cuatro familias de 
e s del norte (Procyonidae, Felidae, Tayassuidae, and C ame fiu ) se diversificaron a un ritmo 1 tado, mientras 
que otras cuatro, Canidae, Mustelidae, Cervidae y especialmente Muridae, evo sjlucionaron en forma explosiva. Por 
consiguiente, la mitad de géneros suramericanos actuales son E de inmigrantes del norte. La otra principal 


consecuencia del intercambio fue la conquista de Norteamérica tropical por inmigrantes amazónicos, un episodio que justifica 


el término de Rei "uno Neotropic cal. 


The roots of the Latin American biota can be traced living faunas, along with their fossil predecessors, 
down through the epochs of earth history to several record both the rapidly changing distributions and the 
iversifications that illuminate 


[om 


formative chapters. During most of the Mesozoic Era. dynamic evolutionary ¢ 
the South American biota was an integral part of the the origins of the Latin American biota. 

Gondwanan biota. Through most of the Cenozoic Era, My own introduction to the nature of the GABI 
the terrestrial and freshwater biotas of South America began in the 1960s when the Florida Museum « 


were isolated from those of North America and all Natural History was actively filling chronological gaps 


— 


other continents by deep ocean troughs. The most in that state's. rich late Cenozoic record ol land 


recent chapter begins with the Great American Biotic vertebrates. One fossil site in particular opened 
Interchange (GABD. This occurred approximately a critical window into the late Pliocene, which is 
three million years ago during the late Pliocene. known in the system of North American land mammal 


when the powerful forces of plate tectonics raised the ages as the Blancan. Unlike any preceding stage. that 
isthmian land bridge in a final phase of uplifting the interval introduced to North America a large cohort of 
northern and southern cordilleras to connect them as land mammals previously restricted to South America. 
the backbone of the Americas. Such dramatic physical Our site, Inglis 1A (Fig. 1), yielded the richest 
changes inevitably triggered major new biological sample of late Pliocene vertebrates in eastern North 
interactions between the two previously separate America. Our team excavated and screenwashed 
American continental biotas, and that is the focus of about 100 tons of fossiliferous sediment, which 
this paper. vielded about 120 species of land and freshwater 

In the following pages, | consider the major animals. Early in our studies, | realized that ten 
evolutionary features of the GABI as registered by genera of new immigrant taxa from South Ámerica 
and mammals. The mammals offer the richest, most accounted. for about 20% of the entire rich fauna 
both (Webb, 1976). There were three families of ground 


nearly coherent. record of faunal changes ii 
American continents during the late Cenozoic. The sloths; three families of shelled edentates, ranging 


thank Peter Raven pa Alan Graham for inviting me to participate in the 51st Annual Systematics Symposium at the 
Missouri Botanical Garde n. Lam also grateful to many colle agues who have expanded my perspectives in both biological and 


geological sciences de R. H. Tedford. M. C. McKenna, B. J. MacFadden, the late F. Eisenberg, and the 
partic ipants in our symposium. This paper is Contribution No. 575 from the Florida Museum of Natur il History. 
? Florida Museum of Natural History. University of Florida, Gainesville, Florida 32611, U.S.A. david@millereeklodging.com 


ANN. Missourt Bor. GARD. 93: 245-257. PUBLISHED ON 23 AUGUST 2006. 


Annals of the 
Missouri Botanical Garden 


Figure 1. 


barge canal, west-central Florida. Photo by author. 


from tiny Dasypus to gigantic TGlyptotherium^; as well 
as two families of rodents, namely Erethizontidae and 
Hydrochoeridae. The truly surprising forms were 
a vampire bat, Desmodus Wied, and a giant pre- 
daceous bird named tTitanis Brodkorb. Additionally, 


nearly 40% of the fauna at Inglis 1A represented 


mm 


North American genera, representatives of which had 


dispersed to South America during roughly the same 


lime in the late Pliocene. There were relatives of such 
living Neotropical taxa as llamas, peccaries, tapirs, 
Jaguars, raccoons, foxes, and spectacled bears (cf. 
Table 1 for taxa). Thus, a majority of the land mammal 
fauna recovered from Inglis 1A, and living in Florida 
nearly three million years ago, were participants, in 
one direction or the other, of the GABI. I was also 
realize how similar the 
Inglis 1A 


contemporaneous interchange faunas through the 


astonished to mingled 


elements of the local fauna were to 
\merican tropics and all the way into Argentina and 
the “southern cone” of the continent. Thus, in Florida 
we had stumbled fortuitously into the acme of the 
GABI (Webb, 1976). 

The 


represents a dramatic experiment in rapid cladogen- 


interamerican evolution of land mammals 


Ká——————4+ OOO] 


"y represents extinct taxa. 


Paleontological excavations by the Florida Museum of Natural History at Inglis LA (late Pliocene) at cross-state 


esis on an intercontinental scale. We may ask how the 


modern South American fauna differ from the pre- 


ceding fauna living in isolation. Likewise, we may ask 


— 


how the present Central American fauna differ from 


1 the southern 


the preceding fauna that had lived i 
portion of North America. Clearly, the land bridge 
calalvzed a rapid remodeling of the Latin American 
Such 


underlying cause that 


fauna. geological changes represented the 
triggered major biological 


effects. 


We begin by indicaling some preliminary land 
mammal dispersals that preceded the GABI. I refer to 
these as herald taxa. Thereafter, we will consider the 


much larger cohorts that burst across a fully acces- 


sible Panamanian land bridge, which | refer to as 
legions taxa. We also briefly review current geo- 
chronological evidence regarding the timing of the 
GABI. Finally, we will consider evolutionary and 
ecological factors that influenced the origin of the 
South and Central American land mammal faunas. 


HERALD TAXA 


A very small number of interamerican land 


mammal dispersals have been recorded earlier than 


the latest Pliocene in North and South America. These 


Volume 93, Number 2 
2006 


Webb 
Great American Biotic Interchange 


Families of land mammals in the GABI. 


Table 1. 


Family Common name 


Northern legion taxa (17 families) 


Soricidae shrews 
Leporidae rabbits 
Heteromyidae pocket mice 


pocket gophers 
squirrels 

field mice 

cats 

skunks and otters 
foxes and wolves 
Procyonidae raccoons 
Ursidae bears 
Gomphotheriidae elephantoids 
Tapiridae i 
Equidae 
Tayassuidae 
Camelidae 
Cervidae 
Southern legion taxa (21 families) 
Didelphidae opossums 
Dasypodidae armadillos 
Chlamytheriidae giant armadillos 
ae tank-like edentates 


Glyptodontic 
Megalonychidae bear-sized ground sloths 
Mylodontidae middle-sized ground sloths 
Megatheriidae elephantine ground sloths 
Bradypodidae three-toed tree sloths 
Myrmecophagidae 
Desmodontidae 
Callithricidae 
New World monkeys 


capybaras 


Cebidae 
Hydrochoeridae 
Erethizontidae porcupines 


Caviidae guinea plgs 


Agoutidae pacas 
Dasyproctidae agoutis 


Echimyidae spiny rats 
Toxodontidae rhinoceros-like ungulates 
Trichechidae sea cows 
Phororhachidae giant predaceous birds 


herald taxa appear during the middle Miocene some 
six or seven million years earlier than the legion taxa. 
Both Mexico and Central America have increasingly 
strong records of late Cenozoic land mammals, yet in 
those areas no South American immigrants other than 
the herald taxa have been recognized before the latest 
Pliocene (Webb, 1997; Ferrusquia-Villfranca, 2003). 
North (e.g.. 


slingshot-horned Protoceratidae) that had become 


Some native American groups the 
increasingly rare at temperate latitudes evidently 
thrived in the more tropical landscapes of Mesoamer- 
ica (Webb et al., 2003). As stated by Ferrusquia- 


Villfranca (2003: 321). all known mammal faunas in 


— 


the Miocene of Mesoamerica "show strict North 


American affinities.” 

According to present records, the herald taxa 
consist of only two distinct genera of ground sloths 
in North America and one genus of large raccoon in 
South America. As indicated by Tedford et al. (2004), 
+Thinobadistes Hay, a mylodontid sloth, and +Plio- 
metanastes Hirshfeld & Webb, a megalonychid sloth, 
arrived in the early Hemphilllian about 9 million 
years ago (Ma). One record of tPliometanastes from 
California yields the best age estimate for these early 
immigrant sloth genera. There it occurs in sediments 4 
meters below the Mehrten Tuff, which is radiometri- 
cally dated at 8.2 Ma (Hirschfeld & Webb, 1968). 

tArctonasua Baskin is a large procyonid, well 
known in North America during the Clarendonian and 
Hemphillian (Baskin, 2003). +Cyonasua Ameghino, 
a closely related genus, appears in the Chiquimil 
fauna of northwestern Argentina in the Huayquerian 
mammal age, middle to late Miocene (Marshall, 
1985: Flynn € Swisher, 1995). By Chapadmalalan 
time. TCyonasua had been replaced by its very large 
descendant, +Chapalmalania Ameghino. 

Flynn et al. (2005) suggested that two additional 
genera of South American. ground sloths separately 
invaded North America. This suggestion is based on 
late Hemphillian collections from near Guanajuato in 
central Mexico. It be more 


may parsimonious. 


however, to view the two genera, identified as 


— 


tMegalonyx Leidy and +Glossotherium Harlan, as 
North American descendants of the closely related 
sloth taxa already. present in the early Hemphillian. 
etailed 


— 


The question may be resolved by more « 


analysis of the Hemphillian sloth material from 
Guanajuato. 

For three reasons, the appearances of these herald 
taxa have been interpreted as waif dispersals, not as 
crossings of a continuous land bridge. Firstly, the 
hiatus of some six million years, during which no other 
crossings have been documented, suggests a fortuitous 
mode of island hopping. Secondly, living sloths and 
raccoons are particularly adept at floating and 
swimming, so that they are among the most likely 
groups of land mammals to make water crossings 
(Webb, 1985). Thirdly, the very limited number of 
other concurrent arrivals in either direction provides 
strong negative evidence that no broad highway had 
opened between 9 and 3 Ma. The strength of this 
interpretation depends on one’s optimism about the 
completeness of relevant parts of the fossil record. 
Suffice it to state that the late Cenozoic record of land 
mammals in North and South America is exceedingly 
rich and continues to improve each year. 

Occasionally, one reads that the murid rodents also 


jumped the isthmian gap and entered South America 


Annals of the 
Missouri Botanical Garden 


One 


Marshall (1979) encouraged the idea of some mid- 


before the latest Pliocene. early paper by 


murid immigration, even though the word 
the title 
discussion was hypothetical. In their discussion of the 


timing of the GABI at about 2.7 Ma. 


Swisher (1995) reiterate the possibility that two murid 


Tertiary 


"model" in made it quite clear that his 


Woodburne and 


genera had reached South America during the 
Montehermosan stage, generally correlated with the 


4 Ma). 


The supposed records that they cite (by "UN isis in 


latest Miocene and earliest Pliocene (ca. 7 


Marshall, 1985), however, were not explicit and have 
not been verified by subsequent faunal and strati- 
graphic studies. Indeed, in his review of fossil murid 
records in the Pampean region of Argentina, Pardinas 
(1995: 


Reig’s two type specimens to be “dubious.” 


229) finds the stratigraphic 


An additional strong inference comes from paleon- 
tological studies of murids in Miocene-Pliocene sites 
in North America. There the presence of appropriate 
plesiomorphic sister groups such as +Prosigmodon 
Jacobs & Lindsay and tBensonomys Gazin, just prior 
strengthens the view 


to the time of the interchange, 


that these lineages did not jump the gap prematurely. 


acobs and Lindsay (1984), for example, make a strong 
case for Miocene-Pliocene diversification of murids in 
North 


branches during, 


America followed by dispersal of multiple 
Blancan GABL 


but not before. the 


LEGIONS TAXA 
Table | records the families of land mammals that 


moved each direction across the land bridge. In 


addition to including land and freshwater mammals, 
have included the +Phororhachidae, a family of large. 
predaceous birds that followed some of their favorite 
mammalian prey northward across the land bridge. 
tTitanis from Inglis LA and 
1 Florida. 
a balanced md faunal interchange, as noted by 
Webb (1976). Webb Marshall (1982), 
Marshall et al. (1982b). The body of paleontological 


that the mingled 


The record is based on 


other Blancan sites These data indicate 


and and 


evidence also indicates fauna on 
each American continent supported a greater diversity 
GABL 
diversity lasted about a half a million years before it 
decreased to its prior level (Webb, 1976: Marshall et 
al., 1982b). Wilson (1992: 120 


geographic pattern as follows: "When biologists see 


than before the This increase in apparent 


— 


explains this bio- 


a number go up following a disturbance and then fall 
back to the original level; whether body temperature. 


density of bacteria in a flask, or biological diversity on 


— 


a continent, they suspect an equilibrium.” The 


interchange numbers neatly fit a large-scale example 


of the species equilibrium hypothesis. After the initial 


provenance of 


mutual enrichment of both continental faunas, they 


encounter limits to diversity, a kind of carrying 


that 


mammalian groups that one continent can support. 


limits the number similar 


capacily f broadly 


CHRONOLOGY OF THE INTERCHANGE 


The reciprocal arrival of a cohort. of terrestrial 
mammal immigrants deep within each of the two 
American continents marks the Eee of the 
faunal in- 


GABI as 


precisely as possible, because it represents the datum 


interamerican land bridge and a major 


lerchange. [t is important to ilis the 
by which ensuing evolutionary and biogeographic 


changes can be calibrated. 


The best method to date emplacement. of the 


'anamanian land bridge comes, perhaps surprisingly. 


not from any geologic data in Panama but from 
excellent. stratigraphic sequences with rich fossil 
faunas in temperate latitudes of North and South 


America. Only a few examples of this paleontological 
record need be cited here to establish the quality of 
the record and the basic chronological framework. 
The Anza-Borrego Desert lies in southern Califor- 
Salton Sea Basin and within the 
that 


accumulated very rapidly during the Pliocene and 


nia, just north of the $ 


San Andreas Fault zone. Sediments ii 


region 


Pleistocene, first entombing marine fossils within the 
Imperial Formation, an arm of the Gulf of California, 
non-marine vertebrates and their trackways 
Vallecito Creek 


provides a wealth of land 


and later, 
within the Formation. The latter 
formation mammals, 
including early appearances of some South American 
TErethizon. stirtont (White) Fra- 


a vertical sequence of some 4000 meters. 


immigrants, notably 
zier. in 
Furthermore, these fossil beds include a radiometric 
date of 2.3 Ma and are 


a long, detailed paleomagnetic profile (Opdyke et al., 


directly intercalated) with 
1977). Figure 2 exemplifies the outstanding exposures 


of long fossiliferous sections in the Anza-Borrego 
Desert 
An equivalent 


North 


11! Ranch 


fossiliferous section in 


a 


known as 
Tusker 


associated both with an ash dated 


America, in Arizona. 
Fauna. which is 


2.33 Ma and 


Sequence, produces the 


an excellent magnetic profile (Galusha et al.. 1984) 
Similarly, in southwestern New Mexico, local faunas 
from the lowermost strata. in the Mesilla Basin 
(Mesilla A in the Camp Rice Formation) and Pearson 
Mesa clearly tie South American immigrants into 
magnelostraligraphically controlled sections (Morgan 


2003). 


In each of these southwestern examples, the new 


& Lucas. 


immigrants occur within an interval of normally 


oriented magnetic samples. constrained above by the 


Volume 93, Number 2 
2006 


Webb 249 


Great American Biotic Interchange 


HA gS 


bx e 


Tum 


Figure. 2. 
W Wi indicating collecting areas of Los 


radiometric date and the Gauss/Matuyama magnetic 
reversal boundary (fixed at 2.58 Ma) and below by the 
Kaena Subchron (at 3.04 Ma). On this basis. mam- 
malian stratigraphers often cite the time of first 
appearance for Neotropical immigrants into North 
America as 2.6 or 2.7 Ma 
1995; Bell et al., 2004). 


The foregoing faunas from the southwestern United 


(Woodburne & Swisher. 


States that most concisely date the first appearances 
fall within the late Blancan land mammal stage. This 
same biostratigraphic rubric includes several Florida 
faunas, most notably the Inglis IA site, which records 
the grealest wealth of Neotropic 'al immigrants from 
radio- 


any one locality. The unfortunate absence of 


metric control on Florida faunas weakens our ability 


to specify how concurrently the entire cohort of 


Blancan immigrants from South America arrived 
temperate North America. It is reasonable to suggest 
that they had all come within about a two-million-year 
period, for that is the approximate span of the Blancan 
interval (Bell et al., 2004). 

Some evidence of Blancan immigrants comes from 
lower latitudes in Mesoamerica. One unique genus is 
+Meizonyx Webb & Perrigo from El Salvador. 


megalonychid ground sloth is larger than any Blancan 


This 


Vallecito Creek Formation Pliocene-Pleistocene exposures in Anza Borrego State Park, California. 
Angeles County Museum. 


Dr. John A. 
Photo by author. 


species within ils temperate sister genus, Y Megalonyx. 
yet it possesses triangular (plesiomorphic) caniniform 
teeth (Webb & Perrigo, 1985) 

Intensive geochronological and biostratigraphic 
studies of sedimentary sequences near Guanajuato 
in central Mexico record several immigrant genera 
from South America earlier in the Blancan than the 
well-dated southwestern faunas cited above (Miller & 
2001; Flynn et al, 2005: 
Montellano-Ballesteros & Jimenez-Hidalgo, 2006). 


+Plaina, a small sister genus of THolmesina Simpson, 


Carranza-Castaneda. 


is recorded at about 4.6 or 4.7 Ma, in the very early 
Blancan. TGlyptotherium and Y Neochoerus appear in 
the interval between 3.9 and 3.1 Ma (Flynn et al., 
2005). This may suggest, as proposed by Flynn et al. 
(2005), 


mammalian province than southwestern United States. 


that central Mexico represented a different 


It is possible that some Florida Blancan faunas, 
15A and Santa Fe River 
these and other immigrant land mammals from South 


notably Haile l, producing 


— 


merica, may also pertain to the early Blancan. These 
faunas are not well enough constrained by geo- 
chronology. and it is notoriously difficult to distin- 
guish early from late Blancan biostratigraphically 
(Bell et al., 2004). 


The new discoveries of earlier 


Annals of the 
Missouri Botanical Garden 


Figure 3. 


Blancan immigrants in Guanajuato indicate that the 
North 


America on a more haphazard schedule than pre- 


legion taxa may have reached temperate 

viously suggested. 
Critical evidence from southern, temperate parts of 
South America. produces results broadly similar to 
United Long 


sequences of fairly continuous fossiliferous sediments 


those from the southwestern States. 
provide an essential framework for integrating bio- 
stratigraphic evidence with both radiometric dates and 
paleomagnetic reversal chronology (Marshall et al., 
1982a; Flynn & Swisher, 1995). 

A central part of Barranca de Los Lobos near Mar 
del Plata 


— 


Argentina (Fig. 3), along with other 
typical Uquian sites in the province of Buenos Aires, 


North 


American taxa, including equids, mustelids, tayas- 


records a number of first appearances of 


suids, camelids, and murid rodents (Pascual et al.. 
1985; Alberdi et al, 1995). Paleomagnetic work, 


integrated with land-mammal biostratigraphy, places 
these northern immigrants just below the Gauss/ 
Matuyama boundary, justifying a first appearance 
datum of about 2.5 Ma (Marshall et al., 19822), later 
revised to about 2.7 Ma (Woodburne & Swisher. 
1995). 

A curious claim of a much earlier immigration of 
northern land mammals into South America has been 
advanced by Campbell et al., (2000: 33A) in a brief 


Barranca de Los Lobos, Buenos Aires Province. 


Argentina, Pliocene-Pleistocene exposures. Photo by author. 


abstract. Based on presumed correlations to sediments 
underlying an ash date of 9.01 Ma in the Madre de 


Dios Formation in eastern Peru. these authors 


recognize. “numerous Amazonian paleofaunas” that 


include “at least two species each of proboscideans, 
peccaries, tapirs and camelids.” The discrepancy 
between their evidence from unspecified Amazonian 
faunas versus the wealth of previously documented, 
late Miocene land mammal faunas in South America is 
profound. The classical sequences at Chiquimil and 
Corral Quemado in Catamarca Province, northwestern 
Argentina, offer a wealth of excellent fossil mammals 
lypifying the Huayquerian land mammal stage. 
Additionally, rich faunas from Micana in the eastern 
altiplano of Bolivia also span relevant, late Miocene 
intervals (see recent references in Flynn & Swisher, 


1995). 


element in any of 


Yet the only hint of any North American 


these well-studied faunas is 
TCyonasua, discussed earlier from the Chiquimil 
1985; Flynn & 


Swisher, 1995). Even the next younger Chapadmalalan 


— 


auna from Argentina (Marshall. 


and/or Montehermosan mammal ages give no hint of 
the mysterious immigration proposed by Campbell et 
al. (2000). In apparent anticipation of such skepticism, 
that 
undoubtedly hindered the movement of the earliest 
orests” (2000: 33A). 


This ad hoc defense of such a novel hypothesis 


these authors claim "ecological | restrictions 


— 


dispersers out of lowland tropical 


Volume 93, Number 2 
2006 


ebb 
Great American Biotic Interchange 


underlines its improbability. It would be quite 


astonishing if eight or more species of large, mostly 

herding herbivore taxa had remained sequestered in 

southeastern. Peru for six or more million years. 
Recently, the assertion of Campbell et al. (2000) 


received minor encouragement from Ferrusquia-Vil- 


— 


afranca (2003) on the grounds that three of the four 
family groups in question (TGomphotheriidae, Tayas- 


suidae. and Camelidae) are known in the middle 


ha 


— 


Miocene of Mesoamerica. This might suggest 
these three family groups could have been available 


for their supposed early dispersal into one corner of 


the Amazon Basin. Such a charitable view does no 


relieve those who would hypothesize a much earlier 
dispersal by a novel, geographic route from the need 


to provide more substantial evidence. 
LATER FEATURES OF THE INTERCHANGE 


After the early events of the GABI, with balanced 


dispersals and augmented familial diversity in both 


continental faunas, the evolutionary pathways of 


immigrant land mammals in the two American 
continents diverge quite dramatically. Evidently, the 
GABI conformed to the balanced mutual enrichment, 
predicted by the species equilibrium hypothesis, 
during the first million years, within an ecological 
time frame. Thereafter, however, within a more 
extended evolutionary. time frame, the northern and 
southern legions embarked on radically different 
evolutionary pathways. 

In South America, the most remarkable result of the 
GABL is the rapid diversification of many of the 
immigrants from the north. In each successive stage. 
the fossil record throughout the continent. reflects 
a continual increase in the numbers of taxa derived 
from the first wave of immigrants. Webb and Marshall 
(1982: 50) quantified this trend by a census of the 
They 


modern mammalian fauna at the generic leve 


— 


found that more than half (about 53%) of the present 


South American land mammal fauna consisted. of 
taxa that had entered the continent by crossing 


— 


the isthmian land bridge less than three million years 
ago. 

'This result becomes even more notable when one 
turns to the North American record and follows the 
very different evolutionary fate there of immigrants 
from South America. Little or no diversification takes 
place. Some genera may be considered quite success- 
ful by some measures. For example, fMegalonyx, 
a bear-sized ground sloth, evolved fairly rapidly and 
extended its range very widely, even reaching Alaska. 
However, none of the South American immigrants 
diversified above the species level. In the recent fauna 


f temperate North America, only four genera of 


^ 
© 


southern immigrants survive, namely Didelphis L., 
Dasypus L., Erethizon Cuvier, and Trichechus L. 


Wuy Dip NORTHERN TAXA SUCCEED IN SOUTH AMERICA? 


This striking asymmetry in the results of the GABI 
requires explanation. In particular, the immense 
suecess of northern immigrants in South Ámerica 
represents one of the greatest natural experiments in 
rapid evolution on a large scale. The most cogent 
general hypothesis may be that the northern. groups 
that spread into South America had a long, wide- 
ranging history not only in North America, but also 
before that in Eurasia. This cosmopolitan background 


strongly contrasts with the long Cenozoic isolation 


— 


experienced by the pre-GABI South American fauna. 

V particularly intriguing part of this inequity is the 
presence among the northern legions of efficient 
mammalian carnivores. In South America, by contrast, 
the only mammalian carnivores were primitively 
modified members of the marsupial family, Borhyae- 
nidae. The group had declined markedly before the 
interchange. It seems evident that the few remaining 
taxa could not compete with progressive carnivorans 
that 
Furthermore, the large and small herbivores of South 


rapidly spread into their native continent. 


American descent, lacking any previous experience 
with advanced mammalian carnivores, were over- 
whelmed by the rapid onslaught of wolves, jaguars, 
sabercats, and many smaller kinds of carnivores in the 
families Felidae, Canidae, and Mustelidae. 

To place the GABI in a broader context, let us turn 
briefly to the earlier and more diffuse intercontinental 


Asia North 


lists the times and generic numbers 


interchange between and America. 
Table 2 of 


immigrant cohorts that entered North America from 


^ 


Asia. These data summarize recent tabulations for 
Miocene land mammals by Tedford et al. (2004). Of 
particular interest in this context are the several sets 
of immigrants that reached the New World in the 
middle to late Miocene. The greatest of these episodes 
occurred about 19 Ma (middle Miocene), when 17 
Eurasian genera are recorded as new arrivals in this 
continent. In Woodburne and Swisher's (1995) similar 
codification of Cenozoic immigration events in North 
America, this is Interchange Event #6. Two million 
years later, another large wave of immigrations 
appeared in Interchange Event #7 (Woodburne & 
Swisher, 1995). A particularly important member of 
that cohort was +Copemys Wood, a probable de- 
scendant of a Eurasian murid such as TDemocriceto- 
don Fahlbusch. The future impact of this North 
American immigration event with respect to the GABI 
surely could not have been anticipated. Nonetheless, 


one epoch later, when a few inconspicuous descen- 


252 


Annals of the 
Missouri Botanical Garden 


Miocene immigrants to North America from 


Table 2. 


Asia.“ 

Age (Ma) No. of genera Selected genera** 

6 T Lutra (Mustelidae). 
* Eocotleus (Cervidae). 
tMegantereon (Felidae), 
T Trigonictis (Mustelidae) 

7 8 Felis (Felidae) 

8 9 TPlionarctos (Ursidae) 

9 6 

12 2 

E 9 *Gomphotherium 
(Gomphotheriidae) 

16 2 

17 11 TGopemys (Muridae) 

19 17 

20 | 

23 0 


ih igures compiled from Tedford et al. (2004). 
* See text. 


" Includes both fossil and living genera. 


dants of these murid rodents colonized South America. 


they produced. by far the greatest monofamilial 
diversity of land mammal genera in that continent. 
The next middle Miocene immigration episode 
introduces a taxon that is extremely far removed from 
murid rodents, namely, the first proboscidean in the 


New World. Out of 


elephantoid genus tGomphotherium Burmeister. New 


Africa by way of Asia came the 
World descendants of this basal member of the family 
Gomphotheriidae diversified greatly in Central Amer- 
ica and, following the GABL, 

Following a mid-Miocene lull, the late Miocene i 
North 


immigration episodes from 


South America. 


another series of 
Asta. New 

genera included the first New World deer, 
Webb; 


TTrigonictis Hibbard, a large extinct mustelid more or 


\merica witnessed strong 
immigranl 
Focoileus 


otters of the modern genus Lutra Blünnich: 


less closely related to the living tropical genus Lira 


(Hamilton) Smith (the tayra); and +Plionarctos Frick. 


the oldest known tremaretine (spectacled) bear. 
presumably an early sister-genus of the living 


Neotropical Tremarctos Gervais. Finally, two immi- 
gration events occurred involving the large cats: Felis 


K Jobert. 


TSmilodon 


L. and tMegantereon Croizel which is the 


probable antecedent. of Lund, the great 
sabercat of the 
these Miocene immigrants contributed key elements 
to the North American cohort during the GABL 
Evidently, these Miocene immigrants extended 
their ranges through the polar latitudes of the Bering 
Strait before climatic deterioration produced extreme- 
ly cold, dry conditions. Virtually all of the omnivores, 


frugivores, and herbivores involved, based on reason- 


Pliocene-Pleistocene. All eight of 


able interpretations of their dental adaptations and on 
the habits of their living relatives, required relatively 
mesic habitats with year-round provisions of nutritious 
browse, fruit, or seeds. 

By the beginning of the Pliocene, the ecological 


o shift 


valences of immigrants from Asia. began 


toward forms adapted to cold steppes. Such grazers as 
tBison H. Smith, 


microtine rodents spread widely through most of North 


TMammuthus Brookes, and diverse 
America but were not subsequently involved in the 
GABI. 


themselves of an 


Evidently, they were too late to avail 


open-country route through the 


isthmian region. There is good evidence that t Mam- 


muthus, during the early Pleistocene, and tBison, i 


the latest Pleistocene, extended their ranges south- 


ward along the relatively dry Pacific terrain into 


Honduras and Nicaragua. These areas today support 
extensive pine savannas and thorn serub, especially 
along the seasonally drier Pacific slopes. Beyond that. 


however, these grazers did not penetrate, presumably 


because they were blockaded by the prevalent 
rainforest (Webb & Perrigo, 1984; Webb, 1997) 


— 


The purpose of this digression into Holarctic faunal 
movements is to show that the Bering Strait had an 
important effect on the GABI Clearly, many groups of 
land mammals that joined the legion taxa from North 
America had previous. evolutionary. records in the 
World Continent. 
may help explain their success in South 
More 


rodents, cervid ruminants, and other northern families 


We suggest that this background 
America. 


than two million years after the GABI. murid 


with similar backgrounds burgeoned into more than 


half of the genera of modern South American 
mammals (Webb & Marshall, 1982). Wilson (1992: 


130) offered the following eloquent explanation: “The 
World Continent 


lines, built tougher competitors, and perfected more 
8 | | 


has tested more evolutionary 


defenses against predators and disease. This advan- 


c 


age has allowed its species to win by confrontation. 
They have also won by insinuation: many were able to 
penetrate sparsely occupied niches more decisively, 
quickly. With both 
the World Continent 


radiating and filling them 
confrontation. and insinu: iion. 


mammals gained the edge.” 


EVOLUTIONARY PATTERNS OF NORTHERN INTERCHANGE TAXA 


To develop a full understanding of how the legions 


M land mammals from the north realized their full 


evolutionary potential in South America remains 


a great challenge. More than a century of dedicated 


paleontological spadework on both American con- 
tinents has provided a reliable framework in two 


respects. Firstly, it has identified the set of 17 families 


A land mammals (Table I). Secondly, it has de- 


Volume 93, Number 2 
2006 


Webb 
Great American Biotic Interchange 


termined a fairly well-dated beginning point for 
reciprocal immigrations of legion taxa between about 
3.1 and 2.7 Ma (Webb, 1985; Flynn et al., 2005 

Yet, from an evolutionary perspective, this is just 
the beginning of the story. The other end of the story 
has also been well wrought by several generations of 
mammalogists working out the modern systematics, 
ecology, and distribution of Neotropical mammal 
eroups. Taken together, these basic facts imply some 
the 


evolution on any continent within the entire 


mammalian 
Age of 


most of the real phylogenetic 


examples of 


— 


of most dynamic 


Mammals. However, 
history of these groups still remains to be told. 
Presumably, the opening of a new continent. after 
long isolation, accompanied by wholesale introduction 
of new kinds of mammals, provides a perfect theater 
for rapid, innovative evolution. The diverse nature of 
South 


topographic and ecologic opportunities i 
America was already evident in the mid Tertiary. 
especially after the uplift of the Andes (Flynn & 
1995; 1999). This 


erand panorama of opportunities may have expanded 


Swisher. Burnham & Graham, 
even more broadly after the GABI as a consequence of 
increased climatic and sea level fluctuations during 
the Pleistocene. 

The 


Pliocene 


the late 


generalization 


basie fossil mammal data from 
also permit an ecological 


about the availability of an open country route. A 
the time of the GABI, the isthmian region was broadly 
accessed by herds of grazers and browsers (Webb, 
1991). 


populated by such large herbivores as horses, 


The legions of the north were primarily 
lamas, 
and deer. An especially significant role may be 
attributed to gomphotheres; presumably they modified 
the landscape by clearing trees, especially during dry 
seasons, much in the manner of African elephants 
today. 

Likewise, a majority of the southern immigrant taxa 
reflect the same set of open landscapes. The analogy 


the key 


relatives. 


because most of 


lack 


Nonetheless, one can reasonably impute to toxodonts 


is weakened somewhat 


families of large herbivores living 


the role of rhinocerotids and confirm it with carbon 
isotopic data. Three or four distinct families, broadly 
brigaded as ground sloths, included. grazers, mixed 
feeders, and browsers. The tank-like glyptodonts are 
generally perceived as herding grazers of savanna 


settings. In the southern set, there were also some 


aquatic herbivores, namely the Hydrochoeridae (ca- 


pybaras) and the Trichechidae (seacows); both groups 


were grazers, the former arrayed along lowland 


waterways and the latter surely moving along the 


coast. In. general, as emphasized by Webb (1978; 
1997), the GABI data imply a broad highway of 


savanna and open woodland that extended from the 


Great Plains of North America to the Pampas of 
Argentina during the late Pliocene. 

The most controversial part of this supposed 
savanna highway is the isthmian portion. In their 
comprehensive review of the complex late Cenozoic 
history of Neotropical vegetation, Burnham and 
Graham (1999: 560) grant that “fossil plant data are 
generally consistent with estimates of habitats (sa- 
vannas) and climate (temperate) based on mammalian 
fossil faunas, but with some differences.” It may be 
necessary to argue for the GABI savanna by calling for 
relatively short-lived rain shadows during the late 
GABI data do not conflict with younger 
belts of 


Pliocene. 
evidence indicating that broad rainforest 
probably dominated the region during most, if not all, 
of the Pleistocene (Colinvaux, 1997 

In South 


northern immigrants, 


America, the remarkable radiations of 


briefly outlined below, moved 
along similar ecogeographic lines. Marquet and Cofre 
(1999) conducted detailed analyses of South Ameri- 
can mammal diversity partitioned into ten. biomes. 
Their results find northern immigrant taxa dominant 


in the following three biomes: temperate grassland, 


cold winter desert, and mountain systems. 


Continuing systematic studies of each of the South 


American phylogenetic lines promise to fill in the 


detailed patterns of their deployment across the 


continent during the Pleistocene. Molecular studies 


provide a particularly important set of independent 


evidence. Based on the degree of its generic di- 


versification, each immigrant family may be assigned 
a bradytelic, horotelic, or tachytelic rate of phyloge- 
netic evolution (Table 3). The nine bradytelic families 


are Soricidae, Sciuridae, Heteromyidae, Geomyidae, 
Leporidae, Ursidae, Gomphotheriidae, Tapiridae, and 


Equidae. Five of these families still consist only of the 


same genus that entered the continent. The four 
families exemplifying intermediate (horotelic) rates 


are Procyonidae, Felidae, Tayassuidae, and Camelidae. 
With nine or ten genera each, Canidae, Mustelidae, and 
Cervidae exemplify rapid (tachytelic) taxonomic evo- 
lution following the GABI. The fourth tachytelic family, 
Muridae, 


producing dozens of genera and several 


the exhibits truly explosive taxonomic 
evolution, 
tribes uniquely differentiated in South America. 

The dazzling diversity of Neotropical Muridae and 
their interrelationships continue to challenge sys- 
tematists. These field mice experienced multiple 
diversifications in southern deserts, in tropical low- 
lands, and in Andean uplands. The sense of their 
succcess everywhere is captured nicely by Pearson 


(1982: 280) in 


altiplano and desert regions of Peru, wherein he cites 


his work on small mammals of the 


the Muridae as “a tribute to the evolutionary ferment 


in South America.” 


Annals of the 
Missouri Botanical Garden 


Pable 3. Generic. diversification among northern land 


mammals in South Ámerica. 


Family N 


. of genera 


Brad ytelic 


Soricidae | 
Sciuridae 3 
Heteromyidae l 
Geomyidae | 
Leporidae l 
Ursidae 2 
TGomphotheriidae 3 
Tapiridae l 
Equidae 3 


Horotelie 


Procyonidae 6 
"elidae 4 
Tayassuidae 1. 
Camelidae 5 
Tachytelic 

Canidae 10 
Mustelidae 9 
Cervidae 9 
Muridae 16—00 


F Includes both fossil and living genera. 


An important consideration. that will increasingly 


concern students of phylogenetic differentiation. of 


northern mammals in South America is the question of 


how much diversification had already been accom- 
plished in tropical North America before the GABL 
As noted above, the paleontological record documents 
that some Murid diversity (e.g. TProsigmodon and 
TBensonomys) was staged before the GABI in low 
latitudes of North America during the Hemphillian 
1984). 


that molecular clock estimates of divergence times 


and Blancan (Jacobs & Lindsay, This means 
among murid rodents may date events that preceded 
the actual entry of those branches into South America. 

Baskin (2003 


example of pre-GABI differentiation: the new genus and 


— 


recently described a surprising new 


species +Parapotos tedfordi Baskin representing the 
procyonid tribe Potosinae in the middle Miocene 
This 


member and the first recorded 


pon 


Barstovian) of Texas. taxon is the only extinct 
fossil of the Potosinae. 
Without this ancient kinkajou record, it would seem 
reasonable to assume that modern forms developed their 
arboreal, nocturnal habits and their marvelous. pre- 
hensile tails as post-GABI evolutionary products of their 
present range in multistratal rainforests of the Amazon 
1989) 


A recent phylogenetic analysis of New World deer 


and lowland Central America (Eisenberg, 
(Webb, 2000) documents the geographic complexity of 
another radiation that had been thought to be purely 
South 1982). The 


American (Hershkovitz, Andean 


endemics, Hipppocamelus Leuckart and Pudu Gray, 


— 


are shown to be members of the tribe Rangiferini 
defined by Webb, 2000) with an extinct sister taxon, 
Kurtén, 


temperate North America. This tribe of mountain deer 


as 


+ Navahoceros in the southwestern parts of 


had already diverged in North America from Mazama 


Rafinesque, Ozotoceros Ameghino, and Blastocerus 


— 


sray, members of the tribe Odocoileini (as defined by 
Webb. 


persals. It seems likely that there were three or more 


2000) before their separate southward dis- 


separate entries of what are presently South American 
deer across the Panamanian land bridge. Thus, the 


issue of Central American diversification and staging 


of northern groups must be considered in phylogenetic 


analyses of South American immigrant groups. 


A SOUTHERN SUCCESS STORY WITHIN THE AMERICAN TROPICS 


The great success of the northern immigrants in 


South America has been well documented by 


paleontologists working in Andean localities, Argen- 


tina, and other parts of temperate South America. 


Those efforts, however, largely miss the contrary 


success story in which Central America was con- 
quered by the tropical fauna of South America. Á few 
sites in tropical parts of South and Central America 
must be more intensely developed before the full 
paleontological perspective can elaborate. on this 
tropical counterdispersal. 

I like to call this episode The Central American 
Paradox. During the Miocene, Central America had 
been a southern adjunct of North America, widely and 
deeply separated from northern South America by the 
Vertebrate fossil sites in Mexico, 


Bolivar Trough. 


lem 


Honduras (Gracias Formation), El Salvador, ane 
Panama (Cucaracha Formation) corroborate the pa- 
leogeographic data by producing no evidence of any 
land mammals from South America (Webb & Perrigo, 


1984: 2003). Yet 


Central America and southern Mexico are faunisti- 


Ferrusquia- Villafranca, today, 
cally and floristically brigaded within the Neotropical 


Realm. This was obvious to A. R. Wallace in his great 
07 
876 


work on biogeography (Wallace, „). In that classic 


synthesis, he extended the Neotropical Realm from its 
Amazonia u 


major center in » to about the Tropic of 


> 


Cancer. 


The predominant groups of mammals in 
Central America today had their origins in Amazonia 
before the establishment of the land bridge. These 
were stalwart members of the ancient isolated fauna, 


New World 


such as armadillos, anteaters. 


including a diversity of monkeys; 


edentates, and sloths; 


and hystricognath rodents. 


— 


Unfortunately, the critical, transitional phases of 


this Neotropical success story are nearly opaque to 


paleontologists. In essence, one can study only the 


Volume 93, Number 2 
2006 


Webb 255 
Great American Biotic Interchange 


Figure 4. 


Miocene, which is too early, or the modern fauna, 
which, in some sense, is too late. It is a problem 
familiar to any paleontologist who has attempted to 
overcome the refractory nature of the fossil record in 
tropical settings. The pressing need in this case is to 
redouble our efforts in primary field work in Pliocene- 
Pleistocene exposures throughout tropical America. 
Ultimately, this challenge will be answered and this 
mystery will yield to such ongoing endeavors as those 
of Laurito et al. (1993) in Costa Rica and Raney in the 
bb & Raney, 1996). Figure 4 


illustrates one of the valuable late Cenozoic localities 


— 


western Amazon (We 


on the Rio do Acre, Brazil accessible during the dry 
season. 


CONCLUSIONS 


The causes of the GABI can be traced readily and 
powerfully to the forces of plate tectonics in lower 
Central America. Hundreds of cubic kilometers of 


ignimbrites and other massive volcanic productions 


— 


1ad already greatly augmented the volume of nuclear 
Central America during the middle Miocene. The 
eradual shallowing of seaways in the isthmian region, 


driven by persistent. subduction of oceanic crust by 


— 


Pacific plates, had major consequences well before the 
final emplacement of a continuous land bridge (Cronin 


& Dowsett. 1996). Jones and Hasson (1985) reviewed 


Paleontological excavations by Universidade Federal do Acre party, led by Dr. Alceu Raney, on Rio do Acre in 
late Cenozoie sediments. Photo by author. 


the many studies that demonstrate the separation and 


divergence of marine faunas on both sides of the 


isthmus. In general, deepwater species undergo 
geminate speciation earlier than shallow-water forms, 
and such patterns can be demonstrated in many 
organisms both across a broad range of taxonomic 
levels and also by rapidly diverging molecular systems, 
notably by mitochndrial DNA (Collins, 1996). Rapid 
divergence between Atlantic and Pacific descendants 
of planktonic foraminifera, such as the globigerinid 
+Globorotalia multicamerata, indicates that the seaway 
was breaking up between about 3.6 to 3.1 Ma. The 
importance of this process in reorganizing oceano- 
graphic currents and intensifying the Gulf Stream was 
indicated by Keigwin (1982) and Cronin and Dowsett 
(1996), who also noted its probable role in generating 
northern hemisphere ice sheets beginning about 
3.2 Ma and again around 2 Ma. 

The final emplacement of a continuous isthmian 
land bridge is faithfully recorded by the appearance of 
legions of land vertebrates reciprocally dispersing 
North South 


evidence from long sections in temperate latitudes of 


between and America. Substantial 
both continents places the GABI events between 
about 2.7 and 3.1 Ma (Webb, 1985; Flynn et al 
2005). 


During the first million years or so of the GABI, the 


.. 


contribution of each continent to the faunal in- 


256 


Annals of the 
Missouri Botanical Garden 


terchange was roughly equal at the family level, and 
each continental fauna was taxonomically enriched. 
Subsequently, by early Pleistocene time, the in- 


lerchange, as recorded by new immigrants, was 


essentially completed and family-group diversities 
had dropped back to previous levels, in accord with 
predictions of species equilibrium theory. 

The northern legion taxa experienced astonishing 
evolutionary success as measured in the modern South 
half of the 
When that broad record is 
that 


with 


American fauna, capturing more than 


present generic. diversity. 


partitioned, it is seen nine families. simply 


entered the continent little or no taxonomic 


diversification above the species level (ef. Table 3. 


bradytelic families). Four other families o 15 5 


and mammals evolved. taxonomically at moderate 


(horotelic) rates: these include Procyonidae, Felidae. 


Tayassuidae, and Camelidae. The four families that 


really seized the new opportunities in South Ámerica, 


evolving at tachytelic rates, were Canidae, Musteli- 


dae, Cervidae, and especially Muridae. It is worthy of 


note that their success focused ecologically on the 
very biomes that most nearly resembled temperate 
had come. According to 
(1999) detailed analyses of 


mammalian diversity in ten biomes, 


biomes from whence. they 
Marquet and Cofre’s 
northern immi- 
grant taxa are dominant in temperate grassland, cold 
winter desert, and mountain systems. 

A countervailing conquest of Central America by 
the Amazonian biota, while nearly opaque to paleon- 
tologists, has been fully appreciated by most biologists 
since Alfred Russell Wallace (1876) proposed the 
Neotropical Realm. There are hints the relatively 
weak tropical portions of the vertebrate fossil record 
that this phase of the GABI came after the Pliocene 
and that during the Pleistocene the Panamanian land 
bridge was more fully occupied by rainforest (Webb. 
1991; Webb & Rancy. 


makes a 


1996). Paleobotanical evidence 


case for this (Colinvaux, 1997; 


1999), 


and mammals, the prominent protagonists 


strong 


Burnham & Graham, 


Among 
in this second phase, into Central America, were old 
southern taxa including marsupials, primates, xenar- 
thrans (both cingulates and pilosans), and hystricog- 
nath rodents. It is not surprising that these same old 
southern taxa outnumber the northern newcomers in 
the two. South American 
Marquet and Cofre (1999) 


and tropical humid forest. 


biomes designated by 


as temperate rain forest 


Literature Cited 


M. T. Alberdi, G. Leone & E. 1995. 
Evolución i adi Y € 1 de la Región Pampeana 
on s de Anos. 


Tonni (editors). 


durante los Ultimos Cinco 


Mus. Nac. Ci. Nat., Madrid | 


Monogr. 


Baskin, J. 2003. 


and Barstovian of the 


New Procyonines from the Hemingfordian 
inc s 
the first 


Flynn ertebrate Fossils and Their 5 xt: 
Contributions in Honor of Richard H. Tedford. Bull. Amer. 
Mus. Nat. Hist. 

Sell, €. J. et al. (+ 8 authors). 2004. The Blancan. 


desde ae Rancholabrean Mammal Ages. Pp. 232- 


314 in Woodburne (editor), Late Cretaceous and 
Cenozoic 1 of North America: ol and 
Columbia Univ. Press, New York. 
N 1999, uM 
developments and status. 
D l6— 589. 


5 authors). 2000. 


Geochronology. 


Burnham, R. J. é Graham. history of 


a al vegetation: New 
. Missouri Bot. Gard. 86: 
Miocene 


Campbell. K. et Late 


dynamics of the Great American Interchange: Waifs are 
out. J. Vertebrate Paleontol. (abstract supplement) 20: 
33A. 

ee P. 1997. The history of forests on the isthmus 


B to the Present. Pp. 123-136 in 
0 oales (editor). C central \merica: A Natural and € alal 
History. 
Collins, T. 
species pairs: Rates and patterns of evolution. Pp. 303- 
33 J Jackson, A. F. Budd. & A. G. 


(editors), Evolution and Environment in Tropical America. 


996, Molecular comparisons of transisthmian 


4 in J. B. C. Coates 


Univ. Chicago Press, Chicago. 
Cronin, T. M. & H. J. Dowsett. 1990. 


Biotie and 


oceanographic response to the Pliocene closing of the 
|J. 


Y. C. Jackson, 


Central American isthmus. T T0-l04in J. | 
Budd. & A Evolution. and 


= 


F. Budd. € . G. Coates (editors). 


D MM nb in Tropical \merica, Univ. Chicago Press 
Chicago. 

1989. 
Northern Neotropies, vol. 1. Univ. Chicago Press, Chicago. 


io mE Middle Miocene 
Pp. 321-347 in L. 


Eisenberg. J. F. oe of the Neotropics: The 


Ferrusquia-Villafranea, J. 


mammalian assemblages: An overview. 


Flynn (editor), Vertebrate Fossils and Their Context: 
Contributions in Honor of Richard H. Tedford. Bull. Amer. 
us. Nat. Hist 
Flynn, J. J. & C. C. Swisher III. 1995. Cenozoic South 
American land mammal ages: Correlation. to global 


— 


geochronologies. Pp. 317-333 in W. 


Geochronology. 


. Berggren (editor). 

Time Scales and Global Stratigraphic 

Correlation. SEPM Special ee No. 5 

— et al. ( 6 authors). 2005. 
He mM Blancan aged strata, 


Geochronology of 
Guanajuato, Mexico, 
and implications for timing of the Great American Biotie 
Intere 5385 „col. 11368): 287-307. 
[.. . Johnson, E. H. Lindsay. 
ie 1984. 
ik late es rocks, = 
Soc. Amer. Bull. 14— 
Hershkovitz, P. 
Pudus. dc Gray. Fie dn. Zool. Ul: 
Hirschfeld, E. & S. D. Webb. P TM Pleistocene 
Me die D sloths ol Nut America. Bull. Florida State 
Mus.. Biol. dee 8 
JL. " E. H. 


Neogene 


xd N. D. Opdyke & 
Biostratigraphy and magnetostrati- 
Ranch. Arizona. Geol. 


171 5 mn a deer (Ce pridas); Part . 


Jacobs, | Lindsay. 408. . Holaretie radiation of 


muroid rodents and the South 
American Cricetids. J. 

Jones, D. S. & P. 
of the marine invertebrate faunas separated by the Ce mitral 

Stehli & 


American Biotic 


F. Hasson. 1985. History and development 
American isthmus. Pp. 325-355 in F. G. 
Webb (editors). The ( 


Topics in Geobiology, vol. 


Great Interchange, 


New York. 


J. Plenum Press, 


Volume 93, Number 2 


Webb 257 


Great American Biotic Interchange 


Keigwin, E. 1982. Carib- 


Isotopic paleoceanography of the 
bean and east Pacific: : 


Role of Panama uplift in Neogene 
time. Science 217: 350-353. 

C. A. W. Valerio & E. Vega. 1993 
Península de 


Laurito, Nuevos 
hallazgos paleovertebradológicos en la 
pipe ancias paleoambientales y culturales de 


Amer. Centr. 16 


Nicoya: 
la fauna de Revista Geol. 
3-115. 

P. A. € H. Cofre. 1999. 


spatial scales in the structure of 1 


Nac aome. 


Marquet, arge temporal and 
mammalian assemblages 
in South America: A macroecological approach. Oikos 85: 
1979. A model for paleobi | 
rodents. 


o pa I Md 
outh American cricetine Paleobiology 5: 


. 1985. Geochronology and land-mamm: r biochronol- 


ogy ol the Qua in faunal interchan 49-85 in 


F. C. Stehli & S. D. Webb (editors), The C ;real erica 


Ae Intere hanes: do. s in Geobiology, vol. 4. Plenum 
Press, New York. 
—— —, R. Butler. R. E. Drake & R. E. Curtis. 1982a. 


Geochronology of type Uquian (Late Cenozoic) land 
i. reentina. Science 216: 986-989. 
‘ebb. J. J. Sepkoski & D. M. Raup. 1 


American 


d age, 
1982b. 
Inter- 


ee ps and the Greal 
a 


f = ience 215: 1351-1357. 


Miller. & O. Carranza-Castaneda. 2001. Late Cenozoic 
mz m. pas the 7 of central Mexico. Boll. 
Paleontol. Hal. 40: 235-242. 


ante lane Ballesteros, jo & E. 
Mexican fossil mammals: Who, when. where? Pp. 


Jimenez-Hidalgo. (c ditors). 
)-213 


in F. Vega et al. (editors), Studies on Mexican im ontol- 


2003. Mammalian bioc hronology 


— 

Q 

as s 
D 

ps 

Jf. 

— 


La aucas. 


ipa faunas from New Mexico. Pp. 209-32 


(editor). Vertebrate Fossils and Their Context: Contribu- 


tions in Honor of Richard H. Tedford. Bull. Amer. Mus 
Nat. Hist 

Opdyke. D.. E. H. Lindsay. N. M. Johnson & T. Downs. 
1977. The paleomagnetism and magnetic pelar stralig 
raphy of the mammal-bearing sectior Anza- boues 
State Park, California. Quatern. Res. a 316-329 


Pardinas. U. ^ J. 1995. Los roedores Cricetidos. Pp. 229— 


253 in M. Alberdi, G. Leone & E. P. 


Evolución redd y Climática de la Región Pampeana 


Tonni (editors), 


durante los Últimos Cinco Millones de Años. Monogr. 
Mus. Nac. : Nat, Madrid 12. 

R.. M. G. Vucetich, G. J. Scillato-Yane & M. Bond. 
85. Main ln of mammalian diversification. in 
Stehli & S. D. 


Interchange, 


America. 


] 
\ e . (e 5 
To 


Pp. 219-247 in F. G. 
The Great 


American Biotic 
seobiology, vol. 4. Plenum 


Press, 
a. 
Pearson, O. 1982. Distribución pequeños EE ros en e l 
altiplano y los desiertos Perú. Pp. 263-204 
Salinas (editor), Actas del A 7 Congreso Latinoamericano 


Mérida. 


e Zoología. 


Tedford, 


( "s 
ne of the Ar 


— 


2004. 


through 


authors). Mammalian 


Ikareean Hemphillian 
Interval (Late Oligocene through Early I "liocene epochs). 
Pp. 169-231 in M. O. Woodburne (editor Late 
and Cenozoic Mam America: 
Biostratigraphy and Geochronology. Duet Univ. Press, 


Cretaceous d mals of North 


. R. 1876. 
Animals, vols. | and 2. 
Webb, S. D. 


American Interchange. Pale 
978. 


The Geographical Distribution of 

Harper & Brothers, New Yor 

1976. Mind da faunal dynamics of the € 
2: 216-234 


Greal 
)bi IC ology 
| listory xa savanna vertebrates in the New 
W E art " ju pus rica and the ( 
Ann. Rev. Ecol. Syst. 9: 426. 


real Interchange. 


1985. 11 eee ami dispe rsals SpA 
the maño as. Pp. 357-3 tehli & S. ! ebb 
(editors), The Great American Bit 1 

Vork. 


O in F. 
s 


— 
= 
£ 


in Geobiology, vol. 4. Pler ress, or 
99]. 


A 0 Ecoge 7 Mg and the Great American 
Interchange. Paleobiology 17: 280 
. 1997. The Great 
97-122 in A. G. 
: Natural and C EN au. 
Haven. 
. 2000. Evolutionary History of New World Cervidae. 
S. Vil Schaller. (editors). 


1 rican Faunal Interchange. 
es (editor), Central Ameri- 


Yale Uni 


T 


. Press, New 


Pp. 38-64 in E. ' rba & G. B. 
Antelopes, Deer and Relatives: Fossil Record. Behavioral 
Ecology. -~ Yale Univ 


New Hav 
& 


stemalics and Conservation. . Press 
rshall. 1982. 
recent EE n rican | land mammals. | 
Mares & li 
5 otropical 1 ol. 
cology, Univ. Pittsburgh 
& S. C. Perrigo. 1984. 
from 7 ii and El Salvador. n Vertebrate Paleontol. 
37—2 


Historic gl biogeography o 
. 39—54 in M. 7 
Evolution of 


noways (ee 11 


» Pymatuning Laboratory of 


ate Cenozoic verte avi 


==, 1985. New d hid sloths from El 
Pp. J E 3—120 in G. 


Evolution 1 Ecology of 


. Montgomery (editor). The 
Sloths and Ve rmi- 


Salvador. | 

s idillos. 

linguas. Smithsonian Institution Press, Washington, D.C. 
& A. Rancy. 1996. Late Cenozoic evolution A m 
3 


. Coates (editors), E 1 


Neotropical mammal an Pp. 335-358 
Jackson. A. F. Budd & 
and Environment in T sut 


America. Univ. Chicago 


Pre 
cos y & G. Poinar Jr. 2003. New evidence of 
Mioce ene Protoce ub including a new species from 
Chiapas, Mexico. Pp. 348-367 in L. Flynn (editor). 
Vertebrate Fossils and Their Context: Contributions in 
Honor of ^» s H. Tedford. Bull. Amer. Mus. Nat. Hist. 
Wilson. E. O. 1992. The Diversity of Life. Harvard Univ. 
Pre: Cant. 


Swisher 111 1995. Land-mammal 
intercontinental overland 


335-304 


\\ 1 M. O. & C. C. 
high-resolution geochronology. 
Tus rsals, sea level, climate and vicartance. Pp. : 


in . Berggren (editor). Geochronology, Time Scales 
al G Juba PIE UAE Correlation. SEPM Special 


1d 
Publication No. 


LATE QUATERNARY Sarah E. Metcalfe 
ENVIRONMENTS OF THE 

NORTHERN DESERTS AND 

CENTRAL TRANSVOLCANIC 

BELT OF MEXICO' 


ABSTRACT 


The i tailed nature of climatic change over the late Quaternary remains poorly understood for northern and central Mexico. 
scarcity of records in the former and great complexity in the latter have hindered a thorough reconstruction of changing 


environments. Pre viously published research by Mete alfe et al. highlighted questions relating to conditions at the last glacial 


maximum (LGM), the nature of the transition from glacial to inter gl: acial conditions, and change over the e ne, including 
the role of phenomena such as El Niño-Southern Ose illation (ENSO). Here, data from the Sonoran Desert. the Chihuahuan 


Desert, and the Trans-Mexican Voleanic Belt (PMVB) (and saline cent oceans) are reviewed. In the rene regions, the mid- 
Pleistocene may have been drier than the late Pleistocene. which was signifie antly cooler than present and saw more winter 
precipitation derived from midlatitude frontal systems. There was a significant expansion southward of woodland taxa. 
the 
modern desert, Conditions wetter than present persisted into the Holocene, but the modern summer rainfall regime may nol 
have become established al a 9 9000 uncalibrated radiocarbon-dated years before present (UC vr. BP). Fully modern 
conditions started about 4000 ! . BP. In the TMVB, sparse lake sedime nt records indic ate that the mid-Pleistocene may 
have been wetter than the late 1 0 cene. Further data are “til required to confirm whether the proposed pattern of a wel 


although many fossil vegetation assemblages apparently have no a ‘rn analogues. Extensive paleolakes existed i 


west and a dry east around the LGM holds true. Most lake sediment records show major anthropogenic influence from the mid- 
Holocene on, although there is evidence for increasing climatic variability in the late Holocene. New deep sea core records 
indicate the glacial meltwater was re-routed into the Gulf of Mexico after the Younger Dryas cool event, helping to explain the 
delayed onset of the modern summer rainfall pattern in re 


— 


ation to general warming. High-resolution records are still confined 
to deep sea cores and tree rings, but highlight the region's vulnerability to climatic change. 
Keywords: Chihuahua. climate change, late Quaternary, Mexico, Sonora, Trans-Mexican Volcanic Belt. 


RESUMEN 


La naturaleza detallada del cambio climatic o durante el Cuaternario tardío en el norte y centro de México es aún poco 
entendida. La escasez de registros en el primero y la complejidad del segundo, han dific nn una reconstrucción detallada de 


ambientes es La ea 5 ies publicada por Metcalfe et al. destaca preguntas relacionadas a 


condiciones durante la máxima 


as 
de último glacial (LGM), la naturaleza de la transición de glacial a las condiciones 
interglaciales, y de cambio durante el Holoceno, incluyendo el papel de fenómenos tales como El Niño-oscilación meridional 
(ENSO). Aquí se revisan los datos del desierto de Sonora, el desierto de Chihuahua y el cinturón volcánico trans-mexicano 
(IMVB) (y los océanos adyacentes). En las regiones desérticas, el Pleistoceno medio pudo haber sido más seco que el 
Pleistoceno tardío, que fue significativamente más frío que el presente y tuvo más precipitación invernal derivada de sistemas 
frontales de latitudes 5 medias. Hubo una expansión significativa de taxones leñosos hacia el sur, aunque muchos grupos fósiles 
de vegetación al parecer no tienen ningún análogo moderno. Paleolagos extensos existieron en el desierto moderno. 
Condiciones más húmedas que las actuales persistieron en el Holoceno, pero el régimen de lluvias de verano PCR 'rno pudo no 
haberse establecido hasta despues 9000 años de padlina arbono no RS antes del presente (C yr. BP). Condiciones 
000 "C B MV 


completamente modernas comenzaron cerca de 400 > yr. BP. En el” B, escasos registros de sedimentos lacustres 
indican que el Pleistoceno me ‘dio pudo haber sido más hümedo que i 1 tardío. 
adicionales para confirmar si el propuesto » patrón de un oeste húmedo y un este seco alrededor del LGM es verdadero. La 
mayoría de registros lacustres demuestran una. importante. influencia antropogénica desde mediados del Holoceno hacia 
ade 


ante, aunque hay evidencia de un aumento en la variabilidad climática en e : Holoc 'eno Prio. d Los registros profundos del 
fondo del mar indican que el aguanieve glacial fue reencaminada hacia el golfo de Méx s del evento del enfriamiento 


del joven Dryas, lo que ayuda a explicar el inicio retrasado del patrón moderno sa la prec 18 ion de verano en relación al 
calentamiento general. Registros de alta resolución todavía se confinan a muestras del fondo del mar y a anillos de árboles, 
pero destacan la vulnerabilidad de la región a los cambios climáticos 


thank a number of my collaborators and colleagues who have he ped me in various ways with the material for this paper: 
boi Stahle (University of Arizona, U. S. X.). Socorro Lozano, Victor Magaña. and Margarita Caballer o (UNAM, Mexico) 
Isabel Israde and i Chacón (UMSNH, Mexico), Melanie Leng (NIGL, U.K i V J 
O Newton (University of Edinburgh. U.K.). Vera Markgraf (INSTAAR, U 
tahle and Socorro Lozano also provided helpful reviews of the original manuscript. A Leverhulme Study Abroad l 'ellowship 
E AF/2004/ 044: 3) gave me the time to write this pape 
School of Geography, University of M. T NG7 2RD, United Kingdom. sarah.metcalfe& nottingham. 
k 


ce 


ac.u 


ANN. Missouni Bor. GARD. 93: 258-273. PUBLISHED ON 23 Aucusr 2006. 


Volume 93, Number 2 
2006 


Metcalfe 
Late Quaternary Environments in Mexico 


259 


The late Quaternary saw the transition from a glacial 
climate (with the global glacial maximum at 18,000 
uncalibrated radiocarbon-dated years before present 
("C yr. BP)) into the modern interglacial climate at 
about 10,000 “C vr. BP. Although the broad patterns 
of climatic 


change over this period in tropical and 


sub-tropical areas have been identified (e.g.. Gasse, 


ias 


2000). the detailed nature of change remains poorh 
understood. This is particularly true for northern and 


central Mexico. Records from northern Mexico, 


especially terrestrial. records, are scarce, and in 
central Mexico the complexity of the environment 
has hindered making climatic reconstructions. In 


practice, most records are also confined to the past 
70.000— 


equivalent to marine 


70,000 vears or so (the mid-Pleistocene ca. 
45.000 before present (BP), 
isotope stage 3) 

Global climatic change during the Pleistocene and 
Holocene brought about major changes in precipita- 
tion regimes (amount and seasonality), as well as 
changes in temperature and sea level. In general, the 
glacial world was drier than the interglacial world, but 
there are regional variations to this pattern. The area 


to the southwest of the Laurentide ice sheet, including 


present day northern Mexico, was one area of 
increased moisture availability around the LGM 
(COHMAP Members, 1988; Thompson et al., 1993). 


In tropical and subtropical areas, moisture availability 
is a key aspect of climatic change, bringing about 
major shifts in drainage systems and vegetation 
distributions. 

The nature of climatic and environmental change 
can be reconstructed using a range of sources, or 
Most of these proxies 


proxies. provide records of 


relatively low temporal resolution (e.g., decadal. 
centennial) depending on the rate at which material 
accumulates. Key sources of low-resolution data are 
lake 


middens. and paleosols. Some records, however, have 


cores, ocean cores, glacial records, packrat 
annual (or even seasonal) resolution; these include 
some ocean cores and lake cores, ice cores, tree rings, 
and historical and instrumental records. This paper 
reviews climate records from central and northern 
Mexico and the adjacent oceans based on a wide 


range of proxies. The focus here is on records that 


= 


have been published since the reviews of Metcalfe et 
al. (2000) and Fritz et al. (2001). The chosen study 
area includes the modern Sonoran and Chihuahuan 
deserts and the highlands of the Trans-Mexican 


Voleanic Belt (TMVB) (Figure 1). 
THE CLIMATE OF MEXICO 


The modern climate of Mexico is dominated by 
seasonal shifts in the position of the Inter-Tropical 


Convergence Zone (ITCZ), the position and intensity 
of subtropical high-pressure cells (e.g., the Bermuda- 
High), 


midlatitude westerly depressions. In winter, with the 


Azores and the frequency and extent of 


ITCZ in a southerly position (Fig. 2a), most of Mexico 
is dominated by high pressure, resulting in dry 


conditions. Cyclonic systems, originating in the 
Pacific Ocean, affect northwestern and central Mexico 


rainfall. A 


northwest Mexico (primarily the west coast of Baja 


and bring frontal very small area of 

California) receives most of its precipitation in winter. 
On the east side of Mexico, some precipitation is 
brought by nortes, cold winds originating over North 
America that pick up moisture over the Gulf of Mexico 
(Mosiño 1974). 
ITCZ moist, 


established over the country (Fig. 2b). 


Aleman & Garcia, In summer, the 


moves north, and easterly flow is 
This has been 
identified as a monsoonal-type circulation. The period 
June to September is the main wet season over most of 
Mexico. The principal source of moisture is the Gulf of 
Mexico, with flows extending up into the Great Plains 
of the United States and Canada. On the west side of 
called the Mexican Mon- 


Mexico, a low-level flow, 


soon, becomes established, drawing moisture north- 
ward along the Pacific margin, through the Gulf of 
California (or Sea of Cortez), and into the south western 
United. States 1993; 
1998). Convective storms and tropical storms (origi- 
both the Pacific Gulf of 
Mexico/Atlantic Ocean) make a significant contribu- 


(Douglas et al., Higgins et al., 


naling over Ocean and 


tion to summer Englehart & 


Douglas, 2001) 


precipitation | (e.g., 


The position and strength of the key features 
influencing. pressure. distributions and, hence, 


cipitation are known to be affected by climatic 
phenomena such as El Niño-Southern Oscillation 


(ENSO), the North Atlantic Oscillation (NAO), and the 
Decadal Oscillation (PDO). Although the 


effects of ENSO are complex in this region, in broad 


Pacific 


terms El Niño years see increased winter precipitation 
Mexico and drier summers overa 
2003), while La Niña years result in 


more summer precipitation. 


1 northwest 


(Magaña et al., 


Across the whole of Mexico, both precipitation and 
temperature are strongly affected by altitude and 
The highlands of the TMVB, which 
cross Mexico at about 19%N (Fig. 


ably temperate climate because much of the land lies 


aspect (Fig. 3a). 
1), have a remark- 
above 2000 m, with the highest peaks reaching more 
than 5000 m and supporting permanent ice cover. 


Pine and oak forests are the dominant natural 
vegelation over most of the TMVB. In northern 


Mexico, the highlands of the Sierra Madre Oriental 
and Occidental are similar to the TMVB (Fig. 3b). The 


plateau area between these ranges (the Mesa Central) 


260 Annals of the 
Missouri Botanical Garden 


/ 


Baja) 
California g 


Ae 
Gulf of X 
Mexico 25 N— 


Pacific 


^f BELIZE 


| 
Highlands S 
— 12 m 200 "C GUATEMALA 75 5°N — 
L Ip : ^ 1 
< 
as fe 

E. 
Tow 100° W 90° W =~ 
— | | | j 
Figure l. Main topographical features of Mexico showing the locations of the study areas. —a. Baja/Sonora. —b. 

C hihuahua —c. Trans-Mexican Volcanie Belt. 


declines in elevation. northward toward the modern boojum tree (Fouquieria columnaris (Kellogg) Kellogg 
Mexico-U.S.A. border. Lower altitudes, the increasing ex Curran). Rzedowski (1973) has highlighted. the 
predominance of subtropical high pressure, and diversity of vegetation in Mexican dryland areas and 
distance from moisture sources give rise. to the the distinct differences between the Chihuahuan and 


Chihuahuan Desert. Annual precipitation in the Sonoran floras. 


border area is about 200 mm, and there are areas of 

active dunes. The vegetation of the modern Chihua- BHSON ORA 

huan lowlands is dominated by creosote bush (Larrea 

tridentata (Sessé & Mocino ex DC.) Coville). with Today. this area of northwest. Mexico (Fig. 1) is 
a range of other shrubs: yuecas, especially soaptree extremely dry (see above), and the range of terrestrial 
yucca (Yucca elata (Engelm.) Engelm.): agaves; cacti: Sites that have preserved paleoclimatic records is very 
and grasses. In northwestern Mexico, dry conditions limited. Sources of paleoclimatic data include tree 
are exacerbated by the occurrence off-shore of the rings, packrat middens, lake sediments, and deep sea 
cold California current. Areas of Baja California and cores (Fig. 4). Although tree ring records from this 


Sonora are the driest in Mexico, with less than area extend back into the 15th century, there are no 


— 


100 mm of rain annually (Fig. 3a). This broad area other proxy records for comparison, and no attempt 
lies within the Sonoran Desert. Here a summer rainfall has been made to relate. the tree ring records to 
maximum is the norm, except for parts of Baja archival records. As a result, the tree ring data will not 


California and coastal Sonora that are exposed to be discussed here. 


winter frontal systems (see above). Although creosote Most of the paleoclimate data for this region comes 
bush occurs in the Mexican part of the Sonoran from packrat (Neotoma Say & Ord) middens. Packrat 
Desert, it is not as abundant as it is farther north. middens preserve plant macrofossil remains. which 
Lowland vegetation includes many forms of cacti, allow reconstruction of local vegetation to species 


yuecas, agaves, and forms of ocotillo, including the level. They may also contain pollen, which allows 


Volume 93, Number 2 
2006 


Metcalfe 
Late Quaternary Environments in Mexico 


261 


I— 459N 
—30°N 
5N 
0° ( ; | pe A 3 


— 40°N 


m 


ge 


————A. 


Narth American/ 
Mexican Monsoon 
anticyclone 


NAM 


—A—A_ cold front 
U cyclonic flaw main moisture 
SOUrCES 
— — flow at 500 mb 
A anticyclonic flow 
Figure 2. Main features of the atmospheric circulation across Mexico. —a. In winter. —b. In summer. 


8 O 2. 
Intertropical Convergence Zone. Redrawn after Metcalfe et al. (2000). 


ITCZ = 


Annals of the 
Missouri Botanical Garden 


Pacific 
Ocean 


Gulf of 
Mexico 


25 N= 


E ia ( 
[aeuze 


PAR 


o A 
. GUATEMALA ,” 15 N— 


2 


PA i 
90° WS 
| 


Pacific 
ean 


b 


dm e 


25 N— 


Y GUATEMALA / 15 NN — 


" 


Figure 3. —a. Total annual precipitation, —b. 


ajor vegetation types across Mexico. (a. based on data for 1941-2002 


from the Servicio Mete Or ee Mi ‘ional, Mexico: E pus d on data from the Instituto Nacional de Estadística, Geografia e 
4). 


Informatica, and Rzedowski (19 


Volume 93, Number 2 
2006 


Metcalfe 263 


Late Quaternary Environments in Mexico 


Tinajas Mountains 
es 


— 
e — 
Hornaday Mountains 
m 3 


(o) ME 
© — — 
E 
© L. Seca de S 
San 1: 
Sierra Catavina 
| 
prs 4 | 28°N 
3 


Sierra San Francisco 


Guaymas 
Basin 


© Tree rings 


Sources of paleoclimatic data for the nm 
'erred to in the text are name 


Figure 4. 


Sonora re gion. Sites ref 


reconstruction of both regional and local vegetation 


1995), 


(Anderson € Van Devender, but only to 


family level. Species-level identification 


generic or 
of paleovege tation is an asset to reconstructing 


climate, because it can be used to identify the season 
of precipitation; this is a key concern in both Baja/ 
The longest midden records 
(Sierra Catavifia San 
30,000 years (Van De- 
1997; Van 


record. from the Tinajas 


Sonora and Chihuahua. 
from Baja 
Fernando) extend back ca. 
1990a; Peñalba & Van Devender, 
Devender, 1997), while 
Mountains (just over the border in Arizona) reaches 
back to 43,000 BP (Van Devender, 1990a). Although 


the amount of data is limited, it appears that in this 


California and 


vender, 


area the mid-Pleistocene may have been drier than the 
1990a). The mid- 
Pleistocene middens from Tinajas include Joshua tree 
(Yucca brevifolia Engelmann), a Mojave desert species 
that is not found in the late Pleistocene middens. In 


contrast. single leaf pinvon (Pinus monophylla Vorrey 
8 pin pr à 


ate Pleistocene (Van Devender, 


& Fremont), although present, is more common in the 
late Pleistocene. The late Pleistocene (full. glacial, 
equivalent to marine isotope stage 2) seems to have 


been significantly cooler than present (by 5 to 6 C). 
with more winter precipitation. Pinyon (e.g., Parry 


pinyon (Pinus quadrifolia Parlatore ex Sibworth)). 


juniper (e.g.. western juniper (Juniperus occidentalis 
Hooker)), and chaparral species were present more 


than 400 km south of their present day distributions 


in southern California and northern Baja California. A 
record from the Sierra San Francisco (Rhode, 2002), 
300 km south of Sierra Cataviña (Fig. 4), which is 
today vegetated with scrub and succulents, confirms 
the southward expansion of juniper and chaparral 
vegelation even at elevations (< 
800 m). Based on the modern distribution of Cali- 
fornia juniper (Juniperus californica Carrière) and 
other taxa, this area of central Baja California may 


relatively low 


have experienced a mild, Mediterranean climate 11 


the late Pleistocene and early Holocene, with at least 


twice as much winter precipitation as it receives 
today. 

It is interesting to note that many of the late 
Pleistocene vegetation communities recorded in the 
middens have no complete modern analogues. This 
may indicate that although the late Pleistocene 
climate here was similar to that of present day 
southern California, there were some differences. 


Woodland plants apparently persisted in this area into 


the early Holocene (at least at elevations above 
250 m). Cooler summers and greater winter pre- 


cipitation seem lo have continued. Middens from the 
Hornoday Mountains in northwest Sonora, close to the 
Gran Desierto, however, do not show the woodland 
plants found in other Sonoran Desert sites 
From about 9000 "C yr. BP, the winter rainfall regime 
have collapsed and the modern. summer 
estab- 


at this time. 


seems 
rainfall-dominated climate regime became 


red. Pinyon-juniper-oak woodland. with chaparral 


lis 
species, were replaced by juniper-oak chaparral and, 
finally, by Sonoran desert scrub (Van Devender, 
1997). 

The middle 
wetter than present (more summer rain). Middens from 


and chaparral 


Holocene was warmer and generally 


ea 


Cataviña, however, show that woodlanc 
elements had died out and been replaced by mesquite 
(Prosopis sp.) and then cactus (Van Devender, 1997), 
An increasing abundance of C4 grasses is often taken 
as being indicative of more summer rainfall, although 
the fossil. record has been 
disputed (Van Devender et al., 1990; Holmgren et al.. 
2003). McAuliffe and Van Devender (1998) report 


that frost-intolerant taxa were present in the northern 


their ls in 


js 


Sonoran Desert and suggest that temperatures in the 
early to mid-Holocene were 2 C warmer than present. 
They also point out that more frequent tropical 
cyclones may have contributed to increased overall 
precipitation levels. These mid-Holocene conditions 


are consistent with insolation changes driven. by 


Milankoviteh forcing. 
Sonoran desert scrub and cactus seems to have been 
established about 4000 BP. 


records 


The present day vegetation of 


sediment from Baja / Sonora are 


— 


Ake 


extremely scarce. There are paleolake basins, but 


264 Annals of the 
Missouri Botanical Garden 
poor microfossil preservation and difficulties with terrestrial records in the late Pleistocene. The authors 


obtaining a reliable chronology have severely limited 
the usefulness of those that have been studied. The 
Laguna Seca San 
1999; Lozano 
García et al., 2002). with a core covering the period 
from about 70,000 to 4000 "C yr. BP. 


indicates a relatively dry mid-Pleistocene followed by 


best dated record comes from the 


Felipe in Baja California (Ortega et al., 
The sequence 
a wetter late Pleistocene. The pollen record indicates 


the presence of open pine and juniper woodland in the 


mid-Pleistocene, with an expansion of juniper wood- 


lands in the late Pleistocene. Strong summer cooling 


and increased winter precipitation are proposed as the 


most likely explanation for the observed changes in 


vegetation. Planktonic, saline water diatoms are 
present for the period from 34.000 to 28.000 BP, 


and lacustrine conditions seem to have persisted until 
about 12.000. BP. There is some suggestion that the 
Younger Dryas (11.000 to 10,000 "C yr. BP) was dry 
but that conditions wetter than present marked the 
early Holocene through to 7000 BP. No pollen record 
was preserved in these sediments. Aeolian sediments 
mark drying in the mid-Holocene, and about the last 
4000 vears of 


Although the record from Laguna Seca has a number 


the record seem to have been lost. 


of problems. it does seem to be remarkably consistent 


with that derived from the much more abundant 
midden data. 
Unlike the terrestrial environment, conditions in 


the Gulf of California (Fig. 1) are conducive to well- 


preserved and high-resolution paleoclimate records. A 


arge number of deep sea cores have been retrieved 
from this area, particularly from the Guaymas Basin. 


High productivity and anoxic conditions at depth have 


resulted in the deposition of laminated sediments 
giving the potential for seasonal resolution (Pike & 
1997). 


(massive) 


Kemp, There are periods of non-laminated 


sedimentation, such as during the full 


glacial and Younger Dryas periods. Despite these 
promising conditions, the interpretation of records 
Gulf of California is far from straightforward. 
JP56 


4) that suggested weak 


from the 
Sancetta (1995) reported results from core 
(western Guaymas Basin, Fig. 
westerly winds in the late glacial and Younger Dryas 
periods, but more El Nino-like conditions and an 
increase in westerly winds into the early Holocene. 
This pattern of change is quite inconsistent with that 
derived from terrestrial records (in. Baja/Sonora and 
U.S.A.) and 


the mid-Holocene. 


the nearby southwest with modelling 


results. By however, the record 


from JP56 is consistent with terrestrial records as 
modern conditions are indicated. 

A more recent record (Barron et al., 2004) from the 
DSDP 840) has 


of matching marine and 


eastern Basin (core 


highlighted the 


Guaymas 
difficulties 


use percent weight CaCO; as an indicator of tropical 


ocean influence. High percentages in the late glacial 
and Younger Dryas periods seem to indicate warm 
conditions, and the Younger Dryas sediments also 
have a higher number of tropical diatom species. The 
straightforward. interpretation of the Younger Dryas 
reflect the 


tropical ocean influence (today a summer and autumn 


sediments is that they persistence of 


phenomenon). However, the authors note that this 
apparent warmth is inconsistent with both the midden 
data (see above) and the results from deep sea cores 
Santa Barbara Basin and 
The DSDP 


shows an abrupt change in the early Holocene, with an 


taken farther north in the S 


elsewhere off the California coast. 840 core 


increase. in productivity. Cooler and more saline 


conditions in the early to mid-Holocene are explained 
as being due to intensified northwest winds; however, 
difficult to 


this is reconcile with the terrestrial 


records. A spike in productivity at 8200 BP (ca. 
7400 “C vr. BP) is attributed to a cold event in the 
North Atlantic brought about by the catastrophic 


drainage of Lake Agassiz. This cold event (also seen 
in the Greenland GISP2 ice core) apparently resulted 
in stronger westerlies and more upwelling, driving an 
increase in productivity. Around approximately 6200 
BP (5400 “C vr. BP), the core record shows the onset 
ENSO conditions, 
with increasing frequency over the last 3000 years. 
ENSO 
s for comparison, but the onset of the modern 
ENSO regime about 5000 BP has been reported 
1999: Tudhope et al., 2001; 


of modern which have occurred 


Unfortunately, there are no long terrestrial 


CO 


elsewhere (Rodbell et al., 
Moy et al., 2002). 

Over ENSO records are pre- 
1993). 


responsive to 


shorter. timescales. 


In 


— 


served in tree rings (Stahle & lic oi 
this area, tree growth appears to be 
increased winter pree iain in El Niño events 
: PDO shows ENSO- 


with 


(although summers are dry 
type features over pg time scales, warm 
(positive) phases being El Nino-like (eastern Pacific 
warm) and cool (negative) phases being La Nina-like 
2003). Evidence 
(40-vear and 


(eastern Pacific cool) (Higgins et al., 
\ DO 


for low frequency variability 80-year 
eycles) consistent with the PDO has been found by 
combining tree ring and reconstructed streamflow data 
from the Gulf of California watershed (Brito Castillo et 


al., 2003). 


CHIHUAHUA 


This area is taken to include the Mexican part of 


the Chihuahuan Desert, covering the modern states of 


— 


Chihuahua, Coahuila, Durango. northern Zacatecas. 


and parts of Nuevo Leon and Tamaulipas (Fig. 1). 


Volume 93, Number 2 
2006 


Metcalfe 265 


Late Quaternary Environments in Mexico 


Pla - —— € —À m 
| Valey er S Hueco Mountains 
RE [Lake 


Sierr 
M 


— 22^ N 


© Tree rings 31°N — 
O Packrat middens 


Lake sediments 


Puerto de La 
Ee s 
- J A CN 9 
| 


EN 


Figure 5. 


described above, summer rainfall predominates here, 
with strong altitudinal and latitudinal gradients. Ás in 
Baja/Sonora, key paleoclimatic data for the Chihua- 


huan area (Fig. 5) have come from packrat middens, 


although there are few records for the Mexican part of 


the Chihuahuan Desert compared with the part within 


the U.S.A. Middens from the Hueco Mountains (west 


Texas. Fig. 5) provide a record covering about 
42.000 years (Van Devender, 1990b). These record 


the presence of woodlands with sandpaper bush and 


big sagebrush (Artemisia tridentata Nuttall type) in the 


mid-Pleistocene. Sandpaper bush and big sagebrush 


are generally considered to be Great Basin plants; 
their presence indicates that their range extended 


southward in this period. They are only rarely present 


in late Pleistocene and Holocene middens. There are 


three records from the Bolson de Mapimi (Coahuila) 


that cover about the last 13,000 years (Van Devender 


Sources of paleoclimatic data for the Chihuahua region. Sites referred to in the text are named. 


1985). 


period, the vegetation of the area was a woodland of 


& Burgess, These show that in the late glacial 
juniper and papershell pinyon (Pinus remota (Little) 
y & Hawksworth), but that 
replaced by Chihuahuan desert scrub and sueculents 
between 12,000 and 9000 "C yr. BP 


scrubland 


Bailey these were 


This change 


from woodland to vegelation occurred 


earlier here than in Baja/Sonora. 
By including midden data from the whole of the 
Chihuahuan Desert (e.g.. Betancourt et al., 2001; 


Holmgren et al., 2003), a longer and more complete 


record of climatic change can be obtained. In the 


Chihuahuan Desert, it appears that the mid-Pleisto- 


cene was drier than the late Pleistocene only at the 
lowest elevations (cf. Baja/Sonora). The late Pleisto- 


cene was marked by cooler summers and mild. wet 
winters, but it seems that there was still some summer 


rainfall (shown by C4 grasses and summer annuals). 


Annals of the 
Missouri Botanical Garden 


These wetter conditions resulted in an expansion of 


pinyon-juniper woodland, including papershell pinyon 


and Colorado pinyon (P. edulis Engelmann). as far 
pin 8 


south as 20 N. As in Baja / Sonora, there are no modern 
analogues for some of these late Pleistocene commu- 
nities, particularly those in modern New Mexico. 

In the early Holocene, winters still seem to have 
been wetter than present, although the limit of winter 
rainfall was apparently tracking northward by 9000 
BP. The modern climatic regime was established 
between 9000 and 8000 BP as desert shrubs replaced 
the woodland taxa (except at high elevations). It has 
been noted that some vegetation types responded very 
rapidly to increasing summer temperatures and pre- 
cipitation. Conditions wetter than present apparently 
persisted through the early Holocene, driven by 
summer (monsoonal) rain, although midden records 
are quite scarce. Holmgren et al. (2003) explain the 
lack of mid-Holocene (8000 to 4000 BP) middens in 
the Playas Valley (southwestern New Mexico, in the 


U.S.A.—Mexico border area) as being due to persistent 
that 


winter drought led to a decline in woody 
perennials. As in Baja/Sonora, modern conditions 
seem to have set in about 4000 BP with the arrival of 


desert scrub species including creosote bush. 


As in the Sonoran Desert region, lake sediment 
records in Chihuahua are few and microfossil 


preservation is often poor. There does, however, seem 
to be much more potential for records from the 


Chihuahuan area because it is the southward 


extension of the Basin and Range province and 
because there are extensive paleolake basins. The 
earliest work in the area was a pollen record from the 
Bolson de Mapimi published by Meyer (1973). This 
apparently showed little change in vegetation (and 
hence climate), although the midden data from the 
same area contradict this. More recent records have 
come from farther north and confirm that the area has 
experienced significant climatic change over the late 
Quaternary. The Babicora Basin, at 2200 m in the 
foothills of the Sierra Madre Occidental (Fig. 5). has 
been the focus of most work (Metcalfe et al., 2002: 
1900; 


2002). Cores from the basin floor and sections from 


Ortega Ramirez et al. Palacios-Fest et al., 


the basin margins have provided a record covering 
about the last 65,000 years. Diatom, pollen, and 
ake 


occupied the Babicora Basin throughout the mid- to 


geochemical data show that an extensive 


late Pleistocene and into the early Holocene. The 


diatom record shows strong fluctuations in conditions 


1 the mid-Pleistocene (with a period of enhanced 
evaporation) that are consistent with the midden data. 
Cool conditions around the LGM are indicated by the 
presence of forests of fir (Picea A. Dietrich) and pine. 


lt is clear from the Babicora record and from lake 


level records elsewhere in the Great Basin that the dry 
conditions of the late Holocene are very unusual. The 
increased moisture availability across this area in the 
middle and late Pleistocene is thought to be the result 
of increased winter precipitation (Bradbury et al., 
2000). The Holocene record from Babicora is rather 
patehy, particularly from the cores. A change in 
dominant diatom taxa in the early Holocene is thought 
to indicate a shift to summer precipitation (Metcalfe et 
al., 2002). A permanent lake seems to have persisted 
into the mid-Holocene, but with strong variability in 
1998). Dry 


conditions are indicated around 5000 BP. An analysis 


water level (Ortega Ramirez et al., 


of Mg / Ca ratios in ostracods (Palacios-Fest et al.. 
2002) from the late Holocene (after 4000 BP) has 
been interpreted as indicating colder winters with 
more effective moisture, possibly more winter pre- 
cipitation. This would be consistent. with the south- 
ward shift of the ITCZ from the mid-Holocene (see 
below), making summer precipitation less dominant. 

A more detailed Holocene record has been 
published for lakes El Fresnal and Santa Maria a little 
farther north of Babicora (Castiglia & Fawcett, 2001, 
2006). 
paleolake Palomas (Fig. 5), but Castiglia and Fawcett 
(2001, 2006) report the occurrence of large lakes in 
The 
argest Holocene lake seems to have been present 
around 8000 BP and covered more than 5600 kn. 


The mid-Holocene showed high variability, with 


These basins were part of Pleistocene 


the Holocene. including the Little Ice 


a reasonably extensive lake in 6000 and 4000 BP, 
but with dry conditions at 5000 BP. The authors 
suggest that millennial-scale high stands may be 
linked to Bond cycles in the North Atlantic and that 
the Holocene high stands may reflect increased winter 
precipitation associated with increasing El Niño 
events (see above). 

Consideration of all the published terrestrial long- 
term climate records from northern Mexico and 
adjacent areas yields a very consistent. picture of 
change since the mid-Pleistocene. The Pleistocene 
was cooler than present (up to 5° to 6 C around the 
LGM) and generally wetter, with the mid-Pleistocene 
drier than the late Pleistocene. Pinyon and juniper 
woodland extended over substantial areas of what are 
the and Sonoran deserts, and 


now Chihuahuan 


chaparral vegetation. extended into southern Baja 


California. Some of the vegetation communites 
recorded in packrat middens have no modern 


analogues. Extensive lakes and wetlands occupied 
basins in present day Chihuahua and Coahuila. The 
data seem to support previous explanations that wetter 
conditions were driven. by increased winter pre- 
cipitation originating over the Pacific Ocean and 


driven. southward by the mass of the Laurentide ice 


Volume 93, Number 2 
2006 


Metcalfe 267 


| 
Late Quaternary Environments in Mexico 


— 20" N 


| | 
© Tree rings 
Q Glacial records 
Lake sediments 


Figure 6. 


sheet (COHMAP Members, 1988; Thompson et al. 
1993; Benson et al, 2003). There are signs 


the LGM. Conditions 


wetter than present persisted into the early Holocene, 


fluctuating conditions around 


even as temperatures increased. Earlier debates about 
the source of early Holocene moisture remain largely 
unresolved, although it appears that summer rainfall 
became established earlier in the Chihuahuan Desert 
than in the Sonoran Desert. Greater effective moisture 
in the early to mid-Holocene is attributed to increased 
summer precipitation driven by insolation forcing. By 
about 5000 BP, conditions seem to have become drier, 
fully modern environments being established 
from about 4000 BP. There is some 
wetter conditions around 3000 to 2000 BP. 


Over the last 1000 years, tree rings provide very 


with 
evidence for 


high-resolution paleoclimatic data, and the number of 
records is increasing rapidly. Most of the records from 
northern Mexico come from Douglas fir (Pseudotsuga 


menziesii (Mirbel) Franco), with the longest sequence 


coming from Durango (AD 1376 to 1993). The 
Durango record from Cerro Baraja and El Salto 


— 
, 


(Fig. 5) (Acuna-Soto et al., 2002; Cleaveland et a 
2003) shows the importance of winter precipitation for 
soil moisture and subsequent tree growth, even in an 
area where summer rainfall dominates. It shows that 
persistent/recurrent La Niña conditions, with dry 
winters, lead to drought (for the trees). The most 
severe droughts in the record occurred between 1540 
1579 (the 


associated with epidemics of hemorrhagic fever), in 


and “Megadrought,” which was also 


Sources of paleoclimatie data for the Trans-Mexican Volcanic Belt. Sites referred to in the text are named. 


the 1860s, and in the 1950s (this major drought also 


affected the southern and southwestern U.S.A.). 
Relating the tree ring records to instrumental climate 
records and historical data has shown that tree rings 
respond to ENSO and to the onset and strength of the 
Therrell et al., 


2002). Unfortunately, none of the other proxy records 


summer monsoon (Diaz et al., 2002: 
cover this time period or have this resolution. 


TRANs-MEXICAN VOLCANIC BELT 


This area is quite different from northern Mexico. It 
is geologically recent, is tectonically and volcanically 
has long been a focus for human 
The 


network has created a large number of lake basins, 
still 


active. and 


fluvial 


settlement. tectonic disruption of ¿ 


most of which contain water (except where 


drained deliberately or affected by groundwater 


abstraction). As described above, a number of the 
highest peaks are still glaciated and a larger number 
were glaciated in the past. These offer an additional 


source of information (Fig. 6). 


Lake sediments provide the greatest number o 
records from the TMVB (Fig. 
the Holocene. To 


records have been low resolution, partly because of 


6); these are mainly for 


and date, the 


ate Pleistocene 


sampling and dating. but also because there are few 
lakes that deposit laminated sediments. The interpre- 
tation of lake sediment records in terms of paleocli- 
mate is affected by tephra deposition and the impact 
Some pollen 


of long-term anthropogenic activity. 


268 


Annals of the 
Missouri Botanical Garden 


records have been interpreted in largely anthropo- 
gente & 1998). 


published records reveal very complex patterns « 


terms (Goman Byrne, Previously 


change (Metcalfe et al., 2000; Caballero et al., 2002). 
In the late Pleistocene, there may be a contrast 


between the western part of the region (wetter than 


present) and the eastern. part (drier than present 


— 


(Bradbury, 1997). The terminal Pleistocene seems to 
have been very dry in many areas, with wetter 
conditions being established the early Holocene. 


There are indications of dry conditions around 5000 
BP followed by more overall variability in climate in 
the second half of the Holocene. Many records show 
evidence for human impact on the environment over 
the last 3500 years, but there is some evidence 
another dry episode around 1000 BP. 

As in northern Mexico, there are few long climate 
records, 


f rr 


but a 27 m core from Lake Cuitzeo (Fig. 6) 
may extend back more than 120,000 years (Israde et 
al., 2002). The present lake is highly alkaline. and 
saline and has a maximum depth of less than 2 m. The 
age of the base of the core is only an estimate because 
the oldest "C date, 42.400 BP There are 
a number of substantial volcanic deposits in the core, 
including a 0-em thick ash layer dating to around 
25,000 BP. The base of the sequence (ca. 120 to 
90.000 BP, equivalent to MIS 5) indicates a freshwater 
lake 
being indicated at the end of the period prior to the 
10 m volcanic 


deposits. The mid-Pleistocene saw a brief deepening, 


is from 9 m. 


f moderate depth, with a lake transgression 


deposition of of silts, clays, and 


followed by shallowing and increasing | salinity. 
Around 35,000 BP. the lake was very low. After 
deposition of the 70 em of ash, the lake became 


deeper again (around the LGM), but then shallowed. 
Pollen data (Velazquez Duran et a Ol) reflect 
conditions wetter than present between 35,000. and 
22.000 BP. (Alnus Miller). 
hornbeam (Carpinus L.), hazel (Corylus L.). and willow 
(Salix L.). Unfortunately, there is a hiatus between ca. 
17,650 and 8000 BP. This may reflect a dry period 
similar to that recorded in the northern part. of. the 
Basin of Mexico (Caballero et al.. 1999), but recent 


isotope data for part of this period from another core 


with pollen of alder 


from Cuitzeo indicate wet conditions. Very low lake 
level, with a possible hiatus, is indicated around 5000 
BP, and the lake has been shallow and increasingly 
saline over the last 3000 years. These conditions seem 


have been particularly pronounced over the last 
1000 years. 


similarities 


The record from Cuitzeo shows strong 
that 


conditions 


Lo from Babicora 
the key period from the 
LGM into the early Holocene remain unresolved. 


(see above). 


although for 
ike. Pátzcuaro is one of the best studied basins in 


L: 
the TMVB. Watts and Bradbury (1982) and Bradbury 


(2000) provide details of a record from the southern 
part of the lake covering the last 44,000 years. This 
indicated wet conditions in the basin in the full glacial 
and latest Pleistocene. 
the 


Drying apparently only set in 
mid-Holocene, the 


confused by human impact. Cores from the north and 


from although record was 


southwest of the basin (Metcalfe et al.. 


in prep.) 
confirm the persistence of a relatively deep lake in the 
basin through the LGM and into the early Holocene. 
As in Babicora, a change in the dominant diatom taxa 
at the start of the Ho 


ocene seems to indicate a change 
in the seasonality of precipitation from winter-domi- 
nated to summer-dominated. The later Holocene record 
shows increased inputs of soil from the catchment, but 
analysis of òO on authigenic carbonates and the 
occurrence of multiple ostracod layers show a number 
of episodes of drying. One of these episodes occurred 
shortly after AD 1342 to 1396 (calibrated “C date), 
corresponding to evidence for low lake level from an 
archaeological site in the southwestern part of the basin 
(Fisher et al., 2003) and to early documentary evidence 
for drought from the Basin of Mexico. From this basin, 
it is clear that lake sediment records have the potential 
to record climatie change of the recent past, but they 
must be interpreted with care. 

Some of the highest peaks of the ‘I 


records 


MVB preserve 
whether the 
glaciers in central Mexico responded to temperature 


of glacial advance, although 


r precipitation (or some combination of both) is still 


not clear. Early work on the area's glacial history was 


carried out in the 1950s and 1960s (e.g.. White, 
1962). with further studies by Heine (1988). Un- 
fortunately, there was disagreement over both the 


identification of moraines and the timing of events. 


More recent work has used 


cosmogenic exposure 
dating (in situ °C), K-Ar, and dated tephras to 
improve the chronology (e.g.. Vázquez Selem. 1998 


It now ap 


M 


ears that the maximum extent of glaciation 
of Iztaccthuatl occurred between about 150.000. and 
126,000 BP ( 2004). 


coincides with the high stand at Lake Cuitzeo. 


Vázquez Selem & Heine, This 
There 
were also advances (on Iztaecihuatl and the Nevado 
de Toluca, Fig. 6) in the late Pleistocene (including 
the LGM), when the equilibrium line altitude was 
about 1000 m below present. The Nevado de Toluca 
records the (when 
conditions seem to have been relatively dry) and 

the late 


readvance in Younger Dryas 


Holocene. Global evidence from tropical 
glaciers suggests that cold conditions alone are not 


sufficient for glacial expansion (L. Thompson. pers. 


comm.): they must have some source of moisture. 
Given the location of. Mexico's highest mountains, 
however, it is possible that at different times they 
could have received moisture from either tropical or 
midlatitude sources. 


Volume 93, Number 2 
2006 


Metc 269 


alfe 
Late Quaternary Environments in Mexico 


In many parts of the world, pollen records have 
played a key part in reconstructing climate change 
Hooghiemstra, 2006). Unfortunately, pollen records 
from the TMVB have been difficult to interpret and 
are not always consistent with other proxies (Lozano 
García & Xelhuantzi López, 1997). They are also 
strongly affected by human activity in catehments. Á 
pollen record from a very high altitude site (3860 m 


— 


avoids human impact and provides evidence of tree 
line fluctuations since the late Pleistocene (Lozano 
García & Vázquez Selem, 2005), reflecting the 
interaction of precipitation and temperature. The 
record indicates that the tree line was 700 to 500 m 
below present in the late Pleistocene/early Holocene. 
The forest expanded in the early mid-Holocene (ca. 
6500 to 6000 BP) under warmer 
tions, but was then replaced by grassland as the 
climate dried. through to about 5000 BP. Wetter 
conditions are indicated about 3000 BP (by the 


mg 


and wetter condi- 


c 
E 


presence of Pinus hartwegii Lindley), corresponc 
to the period of wetter conditions in northern Mexico 
and ice readvance on some on the peaks of the TMVB. 
Cooler and moister conditions are also indicated 
around 2000 BP. T | 
boundary between forest and alpine grassland makes 


it climatically sensitive. A similar picture of lowered 


1e location of this site close to the 


— 


treeline and an expansion of alpine grassland in the 
late Pleistocene has also been recorded in the Upper 
Lerma basin (2570 m) (Lozano García et al., 2005 
Sites at lower altitudes in the TMVB appear not to be 
as climatically sensitive and are more likely to be 
affected by human activity through the Holocene. 

The late Holocene climatic variability indicate 
by tree ring 


— 


— 


by 
lake and bog records is also confirmed 
records emerging from this part of Mexico. The TMVB 
and states immediately to the north (e.g.. San Luis 
Potosi) have recently become a focus for dendrocli- 
matology. Records from this area seem to be most 
sensitive to summer precipitation (unlike those from 
Durango, see above). A climatie record from Douglas 
fir tree rings from Cuauhtemoc La Fragua. Puebla 
(Stahle et al., 2003), covering the period AD 1474 to 
2001 has been correlated with maize (Zea mays L.) 
vields. Seven periods of low yield coincide with 
drought. famine, and social unrest. Tree ring records 
clearly have considerable potential, especially when 
combined with historical records (Therrell et al., 
2004). to extend our knowledge of climate change 


back from the rather short-term instrumental records 


— 


and forward from the more traditional geological 
archives. Even tree ring records, however, are nol 
immune from human impact. A Montezuma bald 
cypress (Taxodium mucronatum Tenore) record from 
Chapultepec Park in the center of Mexico City 


(Villanueva-Diaz et al., 2003) appears to lose its 


climate/ring width relationship over the last 80 years 
as a result of effects of groundwater abstraction. 
There is still little evidence from the TMVB for 
conditions in the last interglacial period, but there is 
-Pleistocene 


some indication that it was wet. The mic 
may also have been wetter than the late Pleistocene: 
this contrasts with the situation in northern Mexico. 
ll remains 
unclear because more records are needed from west of 
101°W to confirm that it was indeed wetter, while the 


east was drier. Records from the Basin of Mexico 


The pattern of conditions at the LGM sti 


indicate very dry conditions in the late glacial and 
early Holocene, but again there are few other records 
for comparison. Pátzcuaro was clearly not dry over this 


period. The patterns of change over the Holocene 


indicated by earlier studies have generally been 
confirmed with drier conditions around 5000 BP and 


again about 1000 to 900 BP. 


SUMMARY 


Pleistocene conditions in Mexico were clearly very 


ferent from present conditions. Northern Mexico 


— 


di 
was much more forested (pinyon-juniper woodland. 
chaparral), with extensive lakes and wetland areas. In 
some areas, this woodland persisted for at least 
30.000 vears before the establishment of the modern 
desert scrub vegetation in the early to mid-Holocene 
(Van Devender, 1990b). The balance of evidence 


indicates that increased moisture was brought by 


westerly winds bringing more winter precipitation 
farther south than today. In central Mexico, glaciers 
and alpine grasslands expanded, and there 
extensive forests of pine, oak, spruce, and 


farther south in Mexico, the Caribbean. 


Evidence from 
and northern South America clearly indicates that the 
summer rainfall regime had largely broken down; 
therefore, it seems likely that some moisture reached 
this area from midlatitudes. Whether this was 
sufficient to drive higher lake levels in the west of 
Mexico and 
Mexico's highest peaks remains to be established. A 
transition toward the modern climate regime occurred 
about 9000 BP as the north of Mexico started to dry 


out. Pinyon-juniper woodlands retreated north, and 


explains the expansion of glaciers on 


spruce (Picea) died out. In the central highlands, very 
dry conditions are recorded in the Basin of Mexico, 
je 


— 


but there is little clear evidence from other sites. T 
early to mid-Holocene was warmer and wetter than 
present in both the TMVB and northern Mexico, 
presumably due to an enhanced summer monsoon in 
response to the insolation maximum. From about 4000 
BP, modern desert conditions (desert scrub and 
succulents) became established in the north, and 


subtropical elements entered the TMVB (again, as 


270 


Annals of the 
Missouri Botanical Garden 


[—45^N 


r—30?N 


—15°N 


— 0° 


ed 


Figure 7. Locations of deep sea cores in the Gulf of Me 


present), Strong variability seems to mark the | 


Holocene, but many records from the north do not 


cover this period (no deposition or lack of preserva- 
lion): furthermore, in the central highlands, records 


are confounded by human disturbance. 


CAN WE IDENTIFY THE CAUSES OF CLIMATE CHANGE? 
Changes in insolation (Milankovitch forcing), sea- 
surface temperature, the extent and height of the 


Laurentide ice sheet, and CO» concentrations have all 
climatic and 
There 


IC timing and magnitude of 


been identified as possible drivers of 


vegelalion change over the late Quaternary. 


7 


— 


also increasing interest in 1 


ENSO, 


ENSO-type. events and their impact on 


Holocene climate change. The inevitably patchy 
terrestrial records of change provide some indication 
of the role of these forci mgs. Evidence. for the 


influence of insolation and the Laurentide ice sheet 


is clear from the literature. and CO) concentrations 
least 
2001). Some further evidence 


can be derived from deep sea cores from the Gulf of 


have played some role in vegetation change, at 


regionally (Huang et al., 


Mexico/tropical Atlantic (Fig. 7). Results from the 
Orca basin and Louisiana slope (Broecker et al., 1989; 
Poore et al., 2003) have shown pulses of meltwater 


from the Laurentide ice sheet entering the Gulf of 
Mexico periodically from as early as 16,000 years ago. 
The occurrence of three major pulses in the late 


elacial period has been well established (Kennett et 


ater 


NI 


0 and Caribbean basin referred to in the text. 


al.. 1985), but Aharon's (2003) study shows renewed 


inputs of meltwater between about 9900 and 
6900 years ago. These inputs may help to explain 
why the summer rainfall regime in the circum- 


Caribbean region did not become fully established 


lagging the insolation 
of the 


sheet has to be considered both through its rearrange- 


Holocene, 
the 


until the early 


forcing. Hence, influence Laurentide 


ment of upper air circulation and the impacts of its 


meltwater. While meltwater entering the Gulf of 
Mexico might seem the most likely to have affected 
the climate of Mexico, an increasing number. of 


records (including those from the Gulf of California) 


indicate that meltwater pulses into the North Atlantic 
(Heinrich events) also had an effect on the Mexican 
climate. 

Some of the most detailed ocean rec a have been 
the ( 
(Fig. 7 


(). 


sediments. 


obtained. from Cariaco Basin off. the coast of 


Venezuela vhere pud results in 


(2001) 


shifts int 


aminated Haug et al. report 


evidence from the Cariaco Basin for 


location of the ITCZ since the last glacial period. 
in the 


a proxy for runoff from the adjacent land, which itself 


Concentrations of Ti in sediments are used as 


responds to the changing location of the ITCZ driven 


by insolation and ENSO. The long-term record shows 


runoff increasing in the early Holocene as the ITCZ 
moved north in response to insolation, but then 
decreasing from the mid-Holocene as the ITCZ 


returned southward. At this time, drving occurred 


Volume 93, Number 2 
2006 


Metcalfe 271 
Late Quaternary Environments in Mexico 


the Caribbean, central and northern Mexico, and 
tropical north Africa. Dry events are recorded in the 
Younger Dryas and at 8200 BP in response to cooling 
in the North Atlantic. 
shows the importance of the equatorial Pacific through 
the onset of stronger ENSOs from about 5000 BP (also 
Basin). The impact of ENSO is 


seen in the position of the ITCZ, which lies farther to 


At higher resolution, the record 


seen in the Guymas 


the north in La Niña years and farther to the south in 
El Niño At the highest 
variations in insolation occur on time scales of the 
, 11, 22 years). 
have also been ene in lake sediment records from 


2001). 


years. resolution, annual 


solar cycles (e. These solar cycles 


the Yucatan peninsula (Hodell et al., 


CONCLUSIONS 


— 


A number of important questions relating to the 
nature and timing of late Quaternary climatic change 
in northern and central Mexico remain unresolved. In 


some cases, it may simply be a matter of working on 


— 


more sites to try to fill some of the gaps in our 
knowledge. Late glacial and early Holocene condi- 
tions were very strongly influenced by the Laurentide 
ice sheet and its history. Reorganizations of the major 
features of the atmospheric circulation in the late 
Pleistocene led to major changes in seasonality and 
amount of precipitation. These changes in precipita- 
tion were reflected by significant changes in vegeta- 
tion distribution, bringing together assemblages of 
plants not found together today. Evidence for early 
warming seems to be recorded in some deep sea core 
records, but is not apparent in the terrestrial records. 
The modern summer rainfall regime only became fully 
established after 9000 BP, 


was probably wetter 


and the early Holocene 


over wide areas than today. 
Changes through the Holocene are consistent. with 
insolation forcing, the resulting position of the ITCZ, 
and enhanced monsoon precipitation (Harrison et al., 


2003; Ruter et al., 


records show the onset of the modern ENSO regime 


2004). So far, only deep sea core 


from the mid-Holocene, although many different types 
of proxies show increased climatic variability after 
Only the 
(annually laminated deep sea cores and tree rings) 
show ENSO and solar cycles. 


It is clear that we still need better understanding of 


this time. highest resolution records 


what drives change in many of the systems from which 
we derive our paleoclimatic records (e.g., vegetation, 
glaciers, lake levels). Better monitoring in the present 
day would help, but it does seem that there may be no 
analogues for conditions at certain periods in the past. 
The better integration of terrestrial and marine records 
understanding, but as the 


should improve our 


examples from the Gulf of California indicate, this is 


— B. Urn F. Valdez 


not always a straightforward task. Obtaining contin- 
uous, long, high-resolution terrestrial records would 


facilitate this comparison. For the more recent past, 


the integration of proxy records with historical and 
instrumental data could shed more light on both the 
nature of climatic variability and the ways in which 
human societies have coped (or not) with the vagaries 


of the climate system. 


Literature Cited 


Acuna Soto, R., D. W. Stahle, M. K. Cleaveland & M. 
Therrell. 2002 Mex negada ah in ~ 
century M 11 Emerg. Infect. 50 

Aharon. P. 200: . Meltwater 11 a events 1n id Gulf of 


Mexico iae Implications for rapid climate changes 
during the last de glaciation. 
no. 1079. 

Anderson, R. S. & 1. R. Va 
listory and -o of the 

Mexic 

Arid Environm, ine 

Barton. J. A., D ry Mo L. eae 2004. High resolution 


paleoe Na of the Guaymas Basin, Gulf of Cali- 


„ 1804): art. 
nú? 


doi: 10.1 n UUZI 


an Devender. 1995. Vegetation 


coastal lowlands of 


Sonora, Er n rec 'ords from pack-rat middens. J. 


fornia, during the past 15,000 years. Mar. Micropaleontol. 

50: 185-207. 
Benson, L., S. Lund, R. Negrini, B. Linsley € M. Zie. 2003. 
Basin lakes to 
Dansgaard- Oeschger oscillations. Quatern. Sci. Rev. 22: 
2239-2251. 


Response of North American Great 


M. C. Peñalba & J. L. 
Late d vegetation history of 

ral New Mexico, U.S.A. P. 
1 ol. 165: 71— 


Betancourt, J. L., K. Aasen Rylander, 
MeVikar. 2001. 
tough Canyon, south-ce 


Palaeoclimatol. 
P. 1997 


alaeo- 

geogr. 

Bradbury, , l Sources of glacial moisture in 
43/44: 97-110 


history of Lago 


Mesoamerica. Guster Int. 
2000. 
Michoacán, Mexico for the past 48,000 years: 


iu 


Limnologic Pátzcuaro, 
Impacts of 
c xd 5 man. Palaeogeogr. Palaeoclimatol. Palaeoecol. 


163: 
E rosjean, S. Stine & F. Sylvestre. 2000. Full and 


— p^ 
late alaca lake records along the PEPI transect: Their 


role in Ari interhemispheric paleoclimatic inter- 


actions. Pp. 265-291 in V. Markgraf Pa tor). Interhemi- 


spheric C ina L inkages. Academic Press, San Diego. 
Brito Castillo, L., S. Díaz Castro, C. Salines 7 Zavala « A. 
9 1 2003. Reconstruction ong-term winter 
streamflow in the Gulf of California watershed. J. Hydrol. 
278: 39-50. 


1989. Routing of 
Laurentide lee Sheet during the 


Nature 341: 318-321. 


authors). 


Broecker, W. S. et al. (+ 6 
meltwater from the 


Younger Tas cold episode. 


Caballero, Lozano, B. Ortega, J. Urrutia & J.-L 
acia 1000. Environmental characteristics of Lake 


northern Ar in 20 for the last 


50,000 years. J. Pale olimnol. 
eu d JE A Macias et al. 
22-ka lake level record and 


human 


2002. Sta. 


climatic 


Cruz Atizapan: 


implications for the late Holocene 


Mexico. 


occupation in the Upper Lerma Basin, Central 


Palar onr em Palaeoclimatol. Palaeoecol. 186: 217—235. 
Castiglia, P. & P. J. Fawcett. 2001. Shoreline and lacustrine 

records of late Quaternary climate change in 

Chihuahuan desert, Mexico. Eos Trans. 82: 7 


272 


Annals of the 
Missouri Botanical Garden 


— & — 20006. Lar 
in Tw C iren de sert, Geology 

Cleaveland, M. e, D. W. Stahle, M. D. 

Villanueva-Diaz & B. T. 2003. 


structed winter pro c i tation rand Iropic e teleconneetions 


» Holocene lakes 5 climate 
: 6. 


* 
c 


change 


ae 


in Durango, Mex Change 59: 369-388. 
COHMAP Members. 
18.000 years: Observations and model simulations. Sci- 
e 31052. 
. Therrell, D. W. Stahle & M. 
(Mexico) winter-spring precipitation 
tree-rings, 1647-1902. Clim. Res 


e hanges of the last 


ence 24 
Diaz, 5. C. 
2002. 


reconstructed 
2: MA 


K. Cleaveland. 

C trud 

from 
37-2 

Doris M. W.. R. A. Maddox. K. Ru ard & 5. Reyes. 1993, 


The Mexican monsoon. J. Climate 6: 1665-1677. 


15 8 P. & X. R. Douglas. 20 MI. The role of eastern 
north. Pacific tropical storms in ilis rainfall climatology of 
western Mexico. Int. J. Climatol. 21: 1357-1370, 


Fisher. C. T... H. P. 


Garduño-Monroy & S. K. 


Pollard, l. Israde-Aleantara, V. H. 
Banerjee. 2003. 
of human-induc ie environmental change within the La 


V reexamination 


Pátzc "uaro Basin, Michoacan, Me . Proc. Natl. Acad. 
J.S.A. 100: 19 74962. 
Fritz, S. C., S. E. Metcalfe & W. Dean. 2001. 


Holocene 
climate patterns in 9 5 Americas 1 from. paleolim- 
. 241-263 in V. Markeraf (editor). 
Interhemispherie C 1 Linkages. » “ademie Press, San 
Dies 


go. 
Gasse, F. 


nologic al rec cords. 


2000. Hydrologic al changes in the African tropics 


since the Last Glacial Maximum. Quatern. Sci. Rev. 19: 
189-211. 
Goman, M. & R. Byrne. 1908. A 5000-year record of 


agriculture and tropical forest clearance in the Tuxtlas. 
Veracruz. Mexico. Holocene 8: 83-89. 
Harrison. S. P. et al. (+ 7 2003. Mid-Holocene 

Americas: A dynai 


climates of the 
Clim. 


authors). 
nical response to 
changes in ans : 
Haug, G. H., . Hughe 5 
U. Rohl. 795 E migration of the Intertropie “al 
Zone 293: 


Peterson € 


Convergence through the Holocene. Science 
130 
Heine, " 8 Late 
Mexican voleanoes. Geowisse schaften 7: 197-205. 
Higgins, R. W.. K. C. Mo & Y. Yao. 1998 


abla of the US. 


Quale nary glac ial chronology of the 
2 


Interannual 
with 


summer precipitation. regime 
Climate I1: 


emphasis on the southwestern monsoon. J. 
sears 


(+ 16 authors). 2003. Progress in Pan American 

CLIV A researc I: 
Atmosfera m 294 
N Drent VJ. II. Curtis & T. 


2001. Solar 1 of mie frequency in the 
zl 


The North American monsoon system. 


Guilderson. 
Mava 
lowlands. Science 292: |— 
Holmeren, C. M. C. Pe der K. A. Rylander & I. 
Betancourt. 2003. A 16.000 C-14 vr BP packrat ode n 
series from the U.S.A.-Mexico borderlands. Quatern. Res. 
60: 319-329. 
Da H. 2006. 
the extremes: The High Andes and 
Missouri Bot. Gard. 93: 297-32 
Huang, Y. et al. (+ 2001. 


dominant control on. elacial-interglac E variations in Ca 
C4 pl l. 


Quaternary vegetational history al 
Amazonia. Ann. 


5 authors). Climate change as the 


y plant abundance, Science 293: 1647-105 


Israde Alcantara, 
Murillo. 2002. 


tardío el centro del 


Gar dans ams & R. Ortega 


cuaternario 


Hidrobiol. 


Pale Wr inte lacustre del 


1 8 


480 de 


1-78 


Kennett. J. . Elmstrom & N. L. Penrose. 1985. The last 
sulf of Mexico: High 
asalto planktonic foraminifera changes. Palaeogeogr. 
159—210. 


des nd a ie Orca basin. Gulf 
aeg ide 50: 
& M. 


usu ms in Pet 


no 
1997. Some 
cords of central 


Lozano García, Xelhuantz López. 
ate ENS pollen re 

xico: Basin of Mexico and Zacapu. Quatern. Int. 43/44: 
117 —12; 


Vázquez Selem. 2005. A high elevation 

Holoce 'ene bolle n record from id acc aie itl volcano, central 
ico. Holocene 329-338 

. B. Ortega Guerrero & S. Sosa Nájera. 2002. Mid- to 

te-Wisconsin 8 n record. of 2d Felipe Basin, Baja 


a 


d Quatern. Res. 58: 84-92 
. 5. Sosa Nájera, Y. Sugiera & M. Caballero. 2005. 
.000 vr of vegetation history of the Upper Lerma. 


a a ‘al high-altitude basin in Central Mexico. Quatern, 
n 70-82. 

Magaña, V. O. J. L. J. L. Perez & J. B. Perez. 2003. 
1 of [E | Nino on precipitation in Mexico. Geofis. Int. 


Vazquez, 


12: 313-330. 
Me 1 Y: K T. R. Van Devender. 1998, A 22,000-vear 
record of vegetation change in the north-central Sonoran 


Desert. Palaeogeogr. Palaeoclimatol. Palaeoecol. 141: 
253-275. 

n S., A. Say, S. Black, R. McCulloch & S. O'Hara. 
2002. Wet FEE a during the last glaciation in the 
qi dud desert, Alta Babicora Basin, Mexico, Quatern. 
Res. i 91-101. 

Metcalfe, S. E., S. L. O'Hara, M. Caballero & S. J. Davies. 
2000. 1 of Late Pleist ree eee climatic 
change in Mexico—A review. Quate Sci. Rev. 19 
099-721. 

Meyer, E. R. 1973. Late-Quaternary paleoecology of the 
Cuatro Cienegas Basin, Coahuila, Mexico. Ecology 54: 
082-995 


Mosino Ale man, P. no García. 1974. The climate o 


A. & 
Mexico. Pp dine in R. À. 


H. Bryson & F. K. Hare 

(editors). Climates North America, Vol. Elsevier, 
: or 

Moy, C. M., G. O. Seltzer, Rodbell & D. M. Anderson. 


n 1 of El UE rn Oscillation activity 


ab millennial timescales during the Holocene epoch. 
Nature 420: P 
Ortega Guerrero. B.. Caballero Miranda. S. Lozano García 


& M. De la O 11 1909, Palaeue nvironmental 


date of the last 70,000 yr in San Felipe Basin, lee 
desert, Mexico: Pre ‘liminary results. Geofis. Int. 38: 1-11. 
Irlega - zd Valiente Banuet, J. Un 155 Fucu- 
gauchi, C. Mortera Gutierrez & G. Alvarado Valdez. 1998 


ee changes during the Late Pleistocene- 


Holocene in Laguna Babicora. 
S 


Ww 
yı 


t. Mexico. Canad. J. Earth Sei 
Palacios-Fest, VI., A. I. Carreno, J. Ortega- 1 1 E G: 


2002. A palacoenvironmental recon- 


struction of laguna Babicora, Chihuahua, Mexico, based 
on ostrocod palae pecology and trac element shell 
chemistry, 1 Paleolimnol. 27: 185-206. 

Peñalba. M. C. & T. R. Van 1 1997. Pollen a 


of p " Era and Holocene packrat (Neotoma) 
middens from San Fernando and Catavina, B E a 
Botanical Symposium, Symposium Abstracts: Peinado to 
Zirahue n. San Diego Natural History Museum. 


Pike, J. & x. E. S K 


variability 


San Diego. 


emp. . Early Holocene decadal- 


scale ocean recorded. in bed California 
8 


laminated sediments. Paleoceanog. 12: 227— 


Volume 93, Number 2 
2006 


Metcalfe 273 
Late Quaternary Environments in Mexico 


Poore, R. Z., H. J. Dowsett & S. Verardo. 2003. Millennial- to 
century-scale variability in Gulf of Mexico Holocene 
climate records. Paleoceanog. 1802): art. no. 1048. 
doi: 15 1029/2002 PK000868. 2003. 
Rhode. D. 2002. juniper 
ru taxa in the C p 0 Calif 
aM rn. Res. —10 
Rodbe ll al. (+ 5 ji 9 5 
ord d a Niño-driven 
Science 283: 516—520. 
„S. Vavrus, R. A. Bryson & J. E. 


and environment 


Early Holocene woodland and 


— 


ifornia peninsula. 


= 


s Án — 
alluviation in 


15.000-vear 
southwestern 
Ecuador. 
Ruter, A. y 
2004. e of the dir s and 
tropical Americas (NH) in the mid-Holoce 
= obse Pes with oe 
: 663-6 


> 


Kutzbach. 


e: Comparison 

model ide Quatern. 
. Rev. 23: 663-076 

1 J. 1973. NE ata the flora of 

Mexican dry regions. Pp. ;raham (editor), 


Vegetation and O History 1 di Latin 
i Yo 


61—72 in 


America. Elsevier, New 
994 Mexico DF. 


California— 


. 1994. Vegetación de 8 Limusa, 

Saneetla, C. 1995, 
Seasonal flux patterns and the se dime nt record for the last 
15,000 years. Paleoceanog. 10: 

Stahle, D. W. ES M. K. Cle 
extremes reconstructed from tree rings of the 5 
1 11 0 Great 
12 


Diatoms in the Gulf of 


1993. le rn oscillation 
erra Madre 
Plains. J. Cina te 6: 


eaveland. 


and southern 

. Therrell, J. Villanueva-Diaz € E. H. Cornejo- 

re Pos El bosque centenario de Abeto pd 

(Pseudotsuga menziesii) cerca de Cuauhtemoc La Fragua, 

Puebla, Mexico. woes d manuse oript. 
M. W 


Therrell, . Stahle Cleaveland & J. 
Villanueva- Diaz. a Warm season tree growth and 
precipitation over Mexico. J. Geophys. Res. 107(D14): art. 


no. 4205. doi: 10.1029/2001 JD000851.2002 
————, ——— Acuna Soto. 2004. Aztec drought and 
the “Curse of One Rabbit. Bull. Meteorol. Soc. 


> 


Amer. 


Bartlein, S. P. Harrison 


93. C hee ES as in the western 


Thompson, R. S.. Whitlock, PME 


& W.G. Sn 19 


United States since 18,000 yr A . 318-356 in H. E. 
Wright, J. E. Kutzbach & T. e TT (editors), Global 
Climates Since the Last 11 9 1 Univ. Minne- 


Press, Minneapoli 

(+8 NN 2001. 
E Niño-Southe Tn Osc d um a glacial-interglacial 
cycle. Science 291: 11-15 


W. et al. Variability in the 


Van Devender, T. R. 1990a. Late Quaternary vegetation and 
United States and Mexico. 


L. Betancourt, T 


climate of the Sonoran Desert, 


p. 134-165 in J. 


— 


Van Devender & P. 


S. Martin (editors), Packrat Midde ms: The Last 
40,000 vears of Biotie Change. Univ. Arizona Press, 
Tucson 


1990b. Late Quaternary vegetation and climate of 
the ire Desert, United States and 
P E d —]33 in J. L. Betancourt, T. R. 


Mexico. 
Van Devender & 


7 ns (editors), Packrat Middens: The Last 
0. of Biotic Change. Univ. Arizona Press, 
Tucson. 

1997. 21.000 years of vegetation. change in the 


northern Vizcaino Norte, Baja California. Botanical 
dd Abstracts, San Diego Natural History Museum, 
E 
& Zur . 1985. 


55 i Mapimi: 


Late Pleistocene woodlands 


A m Me the Chihuahuan 
Res. 2 
5 


vegetation of the Hornaday 


in th 
ee rt biot: a? Quatern. 

. L. Burgess, 1990. 
A. of the 
San Diego Soc. Nat. Hist. 


A . Turner. 
Holocene 


[m 


northwest Sonora, Mexico. Proc. 
2: 1-19 


Vázquez Selem, L. 1998. Glacial chronology of Iztaccihuatl 


volcano, central Mexico, based on cosmogenic “Cl 


5 ages and tephrochronology [abstract]. American 


Nu 


Puerto Vallarta, Mexico. 
. Heine. 2004. Late Quaternary 

233-242 in ]. 
p dn 0 mary Glaciations—Extent and Chronology, 
: uth Americ: Antare- 


Association, 
& glaciation in 


Es Pp. Ehlers & Gibbard 


Asia, Africa, Australia, 


Elsevier, Amste 1 RM 

felina Duran. R.. I. Israde Alcantare & S. Lozano García. 
2001. 
itados en el lago de 
últimos 35,000 anos. Ci. 


Significado ecológico de los palinomorfos depos- 
Cuitzeo. Michoacán. durante los 


Nicolaita 29: 45—61. 


Villanueva Diaz, J. et al. (+ 8 authors). 2003. Registros 
climáticos de los Ahuehuetes de Chapultepec en los 
ultimos 450 años. Bol. Arch. Hist. Agua (January-April): 
34. 

Watts, W. A. & J. P. Bradbury. 1982. Paleoecological studies 
at Lake Pátzcuaro on the west-central Mexican Plateau 
95 at Chalco in the Basin of Mexico. Quatern. Res. 17: 

—(0. 


W 11 5 S. E. 1962. 
west side of Iztaccihuatl, 


: 935-95 


Late Pleistocene glacial sequence for the 


Amer. Bul 


Mexico. Geol. Soc. 


QUATERNARY ENVIRONMENTAL Dolores R. Piperno! 
HISTORY AND AGRICULTURAL 

IMPACT ON VEGETATION IN 

CENTRAL AMERICA 


ABSTRACT 


The corpus of historical data from lake sediments relating to the climate, vegetation, and human land use of the lowland 
Central American = al forest between ca. 20,000 BP and the time of European contact is reviewed. Pollen. phytolith, and 
charcoal records identify the distribution and composition of tropical vegetation and fire patterns during the late Pleistocene, 
when they were ay ae altered from today’s, and earliest. Holocene, when plant communities reassembled and 
interglacial representatives began to coalesce the landscape. The significance of the environmental perturbations that 
occurred during the transition from the Pleistocene to the Holocene for human occ tin of the lowland is 'al forest and the 
geography and chronology of agricultural origins is discussed. Fire was employed by hu nters sand gatherers and farmers alike 
during the past 11,000 years as a primary pe of forest modification. The AE effects of an ancient pre-Columbian 
development of plant food production and, subsequently, slash and burn agriculture between ca. 10,000 BP e 1000 BP can 
be seen on lowland forests from Mexico to the Amazon Basi 

Key words: Climate, human land use, prehistoric ls Quaternary history. seasonal tropical forests, slash and burn 
cultivation, vegetation. 


RESUMEN 


Se revisa la recopilación de datos históricos de sedimentos lacustres que conciernen el clima, la vegetación y la utilización 
000 A. 


= 
— 


) ALP. y el momento de 


humos del suelo de los 1 E or “ales de tierras bajas de América Central entre ca. 2 
olen, fitolitos y carbón identifican la distribuci ión y composición de la vegetación y los 


mtacto europeo. Registros de 


patrones de fuego durante el Ple en no tardío, cuando fueron marcadamente alterados en relación al presente y ali inicio del 
Holoceno, cuando las comunidades de plantas se reasociaron y los representantes interglaciales comenzaron a unirse 


P i 
paisaje. Se discute el significado que mi perturbaciones ambientales durante la transición del Pleistoceno al Holaa 
tuvieron en la ocupación humana del bosque tropical de tierras bajas, v la geografía y cronología de los orígenes agrícolas. 


fuego fue e mplea ado por cazadores y recolectores y granjeros por igual durante los últimos 11.000 años como una herramienta 
primaria de la modificación del e Los efectos profundos de un antiguo desarrollo pre-colombino de la producción de 
plantas alimenticias y, posteriormente, de la agricultura de tala y qus ma entre ca. 10,000 AP y 4000 AP pue a verse en los 
bosques de tierras bajas desde México hasta la cuenca amazonic 


This paper presents a summary of Late Pleistocene aged records for climate change now available from 
and Holocene environmental conditions in the Central lower elevations of Central America. The quality and 
American lowland tropical forest. Natural and human- quantity of information dating to the past 10,000 years 
caused changes in | vegetation and climate that are rapidly increasing, and it is beyond the scope of 
occurred during the past 20,000 years are discussed this paper to cover it (see Grimm et al., 2001, for an 
with reference to the accumulated evidence acquired excellent review, including the burgeoning literature 
from. botanical and minerological studies of lake on the timing and effects of past ENSO (El Niño/ 


sediment cores. Insights that historical data may Southern Oscillation) cycles). E do attempt to provide 


provide about how to best conserve and restore a more thorough examination of the evidence for past 
tropical forests are also outlined. I focus primarily human influences on the lowland tropical forest— 


on paleoecological records that date to the Late including a discussion of a few of the most important 


Pleistocene and early Holocene periods. Debates records from South America—from the initial stages 
surrounding vegetation and climate in the tropical of colonization, through the earliest periods of 
forest during that time have been fervent and long- agriculture, until the arrival of Europeans. The reader 


standing. but now appear to be converging on should nonetheless bear in mind that it is possible to 


c 


consensus with regard to some important issues. | cite but a small fraction of the associated archaeo- 


do not provide a systematic overview of the Holocene- logical literature generated during the past 20 years. 


Smithsonian Tropical Research Institute. Panama, and Department of Anthropology. National Museum of Natural History. 
\\ 1 DC. pipernod@si.edu 


ANN. Missouri Bor. GARD. 93: 274—296. PUBLISHED ON 23 Aucust 2006. 


Volume 93, Number 2 
2006 


Piperno 
Quaternary Environmental History 


Paleoecological data attesting to the considerable 
changes in climate and vegetation that impacted the 
lowland Neotropical forest during ice ages have 
accumulated primarily during the past 25 to 30 years. 
Before t 


very limited, theories of environmental stasis domi- 


— 


vat, because direct empirical information was 


nated questions concerning tropical forest ecology and 


environmental history, including explanations of 


1960: Slobodkin & 


These theories were elegant, and at 


species diversity (e.g.. Fisher. 


1969). 


the time they made a good deal of sense because it 


Sanders, 


was logical to assume that the presence of a constantly 
warm and wet climate over millions of years would 
promote a continuous accumulation of species while 
creating little stress on the existing biota and causing 
only low rates of extinction. 

At about the same time that theories of long-term 
tropical environmental stability held influence. what 
would become a famous archaeological debate was 
developing over the importance of ancient human 
occupations in the tropical forest. The dry highland 
regions of Mesoamerica and South America, where the 
(then) best understood and most famous high civiliza- 
New World, 


some prominent prehistorians as the major sources of 


tions had arisen in the were viewed by 
most pre-Columbian cultural innovations and ad- 
vances, including crop plant origins, pottery manu- 
facture, and large and permanent village settlements 
(e.g. Mangelsdorf, 1953; Mangelsdorf et al., 1964: 
Meggers, 1954, 1971). Skepticism abounded in these 
circles as to whether the tropical forest was a fit 
environment for human habitation and innovation. 
including the development of effective agricultural 
systems (Mangelsdorf, 1953: Meggers, 1954, 1971: 
1950, 1952, and Lathrap, 1970, 1973a. for 


the alternative viewpoint that tropical forest cultures 


see Sauer, 


had been very important). Some scholars would argue 


— 


into the 1980s that the first people to colonize Centra 


[om 


and South America largely avoided tropical forest, anc 
that they and later tropical hunters and gatherers 
could not have survived for long without access to 
a cultivated food supply because wild food resources 
were scarce (e.g. Hart & Hart, 1986: Bailey et al.. 
1989: Sponsel, 1989). 


Consequently, the degree to which cultural forest 


— 
— 


modification had taken place during the pre-Colum- 


— 


bian era was in serious question. The “Ecologically 


Noble Savage” concept that tropical forest peoples 
8 I peo] 


lived in harmonie. balance with their natural re- 


disrupting the landscapes on 
settled, was in full. 
1993, and Dods, 2002 


discussion of the Noble Savage concept in anthropol- 


sources, minimally 


which they politically-correct 
swing (see Alvard, . for a good 
ogy and its historical antecedents). Descriptions of the 


extant tropical forest as “pristine” floral communities 


practically untouched by the human hand were 


common both the anthropological and ecological 
literature. 

During the 1960s and 1970s, pioneering studies by 
Van der Hammen and Gonzalez (1960), 
(1975), and Martin (1964) provided the first empirical 
data demonstrating that highland regions occupying 


— 


Livingstone 


tropical latitudes in Africa and America experienced 


dramatic climatic and vegetational changes at the 


same time periods during which glaciations were 


strongest in higher latitudes of the Northern Hemi- 
sphere. These investigations, along with a now-classic 
application of environmental reconstruction through 
palynology in the Classic Maya area (Deevey et al., 
1979), 
tropics. 
phytoliths that dated to the past 20.000 years, and 


thus that could provide evidence on the most extreme 


motivated similar research in the lowland 


Lakes with sediments rich in pollen and 


climatic oscillations of the last transglacial cycle. 


occurred more commonly in lowland tropical forest 


than was originally thought (e.g., Leyden, 1984, 1985: 
Bush € Colinvaux, 1988, 1990; Bush et al., 1989, 
1992; Piperno et al., 1990). The furious constructions 


of modern reference collections of pollen and 
phytoliths that accompanied these initial paleolimno- 
logical efforts would lead to the identification of many 
of the lacustrine plant microfossils at more precise 
taxonomic levels than had been thought possible. 

that 


environments through the same types of paleolimnol- 


became clear robust reconstructions of past 


ogy that had long been pursued in Europe and North 


America were achievable in the lowland tropical 
forest 
Also beginning in the 1970s and 1980s. 


ologists began to intensify their explorations in the 


archae- 


tropical forest by investigating new areas and in- 


creasing their coverage of others. These efforts 


included carrying out intensive foot surveys using 
statistically significant sampling designs that not only 
identified more early human occupations, but also 
robustly documented. regional trends in site number 
and size and population density through time (e.g.. 
Cooke & Ranere. 1984; Lathrap et al, 1975). In 
addition, new and rigorous methods were applied to 
recover and study plant and other organic cultural 
1978; Piperno, 1985a, b, 1988). 


such as the analysis of visible 


remains. (Pearsall. 
Traditional. methods, 
remains of seeds. fruits, and nuts, had not worked in 
studying early plant use and domestication and, 


indeed, were providing highly biased pictures of 


— 


man plant exploitation because organic materials 
the inimical warm and 
the Our 


and processes 


do not survive for long under 


humid conditions found in tropics. un- 


derstanding of the natural human 


affecting the lowland Neotropical forest during the 


276 


Annals 
7 Sl Garden 


Lake Chichancanab 


Gatun Basin 
El Valle 


0 1000 km 
l 


La Yeguada 


Monte Oscuro 


lowered Pleistocene sea levels at 18 
1500 m asl. Modified from Piperno and Pearsall. 
Lake Chic ja m (Hodell et al.. 

Yeguada (Piperno et al., . 199laz Bush et al., 
Basin (Bartlett & A 5 Piperno et al., 


(Piperno, this paper. and Piperno et al., 200: 


as follows: F 
1992): 


witha 


Podoc ‘Arpu s L'Hér. 


often 
Vnus Mill.. 


sufficient uu existed. to support a forest. ( 


x Pers., Quercus, Ilex) were 


elevation forest elements occurred, especially in moister areas of the zone. 


areas with sandy soils may have contained savanna 


undifferentiated thorn woodland, 


2). River- and stream-side locations supported a forest 


past 20,000 years and more has been dramatically 


altered as a result of the above-mentioned work. 
VEGET VITON AND CLIMATE 


LATE PLEISTOCENE 


Most existing lake records of Late Pleistocene age 
from the Americas, including the tropics. date to nol 
more than 20,000 to 25,000 years ago (all ages in the 
are in The 


of an age 


paper uncalibrated) radiocarbon years). 
bulk 
appropriate for evaluating the problems addressed 


10,000 years o 


transglacial cycle (between about 20.000 and 


the available data are. however. 


the last 
0.000 


BP (before present) has long been a period of intense 


here because the final 


interest to paleoecologists and archaeologists alike. 


covers the Last Glacial Maximum (18.000 BP) and 


d 2.4. with permission from : 2 vier. 

Lake Que xil 
El Valle (Bush & ( 1 5 1990 Pi ibe rno et al., 100 la); 
1992); A Oscuro (Piperno & Jones, 
1). The vegetational reconstruction is as follows: (1) Largely male ‘n moist forest. 
mixture of presently high-elevation and lowland forest elements. In $ 
consplc uous. 


2) Forest containing drier elen 


low scrub. and wooded savanna vege 


1 Guat 
elements (e.g.. Juniperus). Areas receiving greater than 2000 mm of rainfall today may still have am da drier forest. 


Location of paleoecological sites with records dating to the Pleistocene and the reconstructed vegetation of 
No ipis val Middle and Central America between 20,000 BP and ca. 
.000 BP. the 1 50 tod of maximum elacial advance. 


10.500 BP Gray areas represent land exposed by 
Black areas indicate clevations above 
Sites and their investigators are 
(Leyden et al.. | Lake Salpeten (Leyden, 1987): La 

Gatun 
2003): Lakes Ixtaevola and Ixtapa 
some areas, montane forest elements (e.. 
Anta] a was lower than today. but 


than characteristic today. High- 


Areas near na 2000 mm precipitation s : and 
woodland. The vegetation may have been patchy. (3) sty 
gelation, Some regions (e.g emala) had empere 


1 


subsequent termination of the Pleistocene. and 


accommodates all chronologies for 
human colonization of the Americas (Dillehay, 1997. 
2000: Meltzer. 1997: Meltzer et al.. 1994). Further- 


more, although paleoecological records from the now- 


now-accepted 


humid lowlands are still few in number, they originate 
from a representative spectrum of the major ecological 
zones found in the lowland Neotropics from ever-wet. 
aseasonal habitats to those characterized by à marked 
seasonality of rainfall. 


Figure | shows the locations of the major lakes and 


arge swamps Central America that have yielded 
records dating to the Pleistocene period. Vegetational 
and other reconstructions derived from these paleoen- 
vironmental sequences were often based on pollen, 


phytolith, diatom, and minerological data analyzed in 


Volume 93, Number 2 
2006 


Piperno 
Quaternary Environmental History 


tandem. Figure 1 also contains a postulated. recon- 
the 
late-glacial period. 


struction of vegetation for Central America 


lowlands during the For areas 
where sequences of Pleistocene age do not yet exist, | 


the 


assumed the paleovegetation be similar te 
paleovegetation in areas which are presently similar 
ecologically. 

The 
indicated from the records have been described in 
detail elsewhere (e.g., Bush et al., 1992; Leyden et al., 
1993: Piperno & Pearsall, 1998) and are summarized 


and 


major environmental trends and patterns 


herein. The lowland Neotropics between 20.000 
10.000 BP appear to have been considerably cooler 
and drier than they are today. Reconstructions of tree 
line descent indicate that temperatures over land 
4°C to 750 the 


present lime: these estimates are largely concordant 


surfaces were from lower than al 
with those derived from newer studies of Pleistocene 
sea surface temperatures derived from coral oxygen 
(Bush et al., 1992; Colinvaux et al.. 1996: 
1994: 1993: Peltier & 


The degree to which the Pleistocene 


isotopes 
Guilderson et al.. Leyden et al., 


2004). 


Was 


Solheim. 


precipitation reduced is under considerable 


discussion. Many of the lakes shown in Figure | were 


lower or completely dry during the Late Pleistocene 
and, as will be described shortly, many of their 
watersheds appear to have contained few trees. These 


factors have led many investigators, the author 


included, to believe that intertropical annual rainfall 


reduced on the order of about 30% to 
50% (Leyden, 1984: 1993; Piperno & 
Jones, 2003; Van & Hooghiemstra, 
2000). Others would disagree, believing that rainfall 
reduction was instead only about 10% to 15% below 


1996, 2000). 


There is no doubt that glacial-age atmospheres 
t S | 


was probably 
Leyden et al., 


der Hammen 


— 


today's values (e.g.. Colinvaux et al., 
contained much less CO» than they would after the 
1987; 


CO» concentrations were 


Cowling & 
33% lower 
12,500 BP, 
when they were still 25% lower than the PIV, not 
rising close to PIV 9000 BP 
(Barnola et al., 1987). The lower CO» concentrations 


probably exerted significant influences on the vege- 


Pleistocene ended (Barnola et al.. 
Sikes, 1999). 
than. preindustrial values (PIV) until ca. 


values until about 


tation in ways we do not understand very well 
present (Cowling & Sikes. 1999: Mayle & Beerling, 
2004). One possible major effect was that Pleistocene 
forest Canopies were more open than today’s due to 
lower light use efficiency during photosynthesis (Sage, 
1995) 

In many regions of the tropics these identified ice 
age conditions—much cooler, probably substantially 

much atmospheric CO» 
2.000 to shortly after 10.500 years 


ago, such that modern vegetation communities did not 


drier, and with lower 


persisted until ] 


begin coalescing until that time. For example, at Lake 
Quexil, located in the Peten, Guatemala, seminal work 
by Barbara Leyden and her colleagues (Leyden, 1984; 
1993) showed 


f the modern semi- evergreen forest in 


Leyden et al., that trees now charac- 


teristic of the 
region are not recorded in pollen records until a short 
after ca. 10,300 BP, t took : 


another hundred 9 1 8 a 


and that i east 


time 


several years Md 
veloped tropical forest was established. During the 


contrast, 


Late Pleistocene, 1 the region was a cool 
and very dry place in which few trees grew. 

Existing paleoecological data allow us to identify the 
following major, Late Pleistocene vegetational patterns. 
First. where the modern potential vegetation is de- 
forms of semi-evergreen forest. 


ciduous or drier 


a 5 of forest is evidenced by open vegetation 


similar to but probably not exactly equivalent with 
today’s 1 woodlands, thorn serublands, and savan- 
na. This is empirically demonstrated in the lowlands of 
southwest Mexico. Guatemala, Haiti, and Pacific-side 
Panama. at Lakes Ixtacyola, Ixtapa, Quexil, Salpeten. 
and Monte Oscuro. 


rainfall is between about 2 


Second, where today the annual 
G and 4 m and the actual or 
potential vege talion is evergreen and semi-eve rereen 
forest, the Pleistocene vegetation was largely forested. 
This is empirically demonstrated in Caribbean-side 
the 


watershed sites in Panama at elevations of between 500 


Panama al Gatun Basin, as well as at Pacific 
and 700 m above sea level (asl). such as La Yeguada 
and El Valle. 
throughout the Caribbean watershed of Central Panama 
\t the middle 
the 


Valle, La 


This situation presumably prevailed 
and other low lying wetter areas as well. 


500—1000 m 


Pleistocene vegetation. was forest (e.g.. El 


elevation sites (ca. asl) where 


Yeguada), it can be seen particularly well how the 


— 


arboreal component of the vegetation responded to the 


cooler temperatures; trees that today generally grow 
only in mountainous areas above 1500 m were forced 
about 800 to 1200 m downslope and co-existed with 
those lowland arboreal elements that were better at 


tolerating lower temperatures. 


— 


It is not surprising in view of all of this that few to 


none of the vegetational signatures from the Pleisto- 
There were 


cene appear to have modern analogues. 


different forest facies likely involving the reduction 
and partial replacement of lowland evergreen forest by 
arboreal elements now primarily confined to drier 
types of forest. However, because species-specific tree 
identifications usually cannot be made from pollen 
and phytolith records, details of forest composition are 
difficult to reconstruct with any certainty. This is an 


area of research where increasing our efforts to 


retrieve and identify seeds and other macro-fossil 


plant remains from lake sediments may prove to be 


rewarding. 


Annals of the 
Missouri Botanical Garden 


Figure 2. The crater and surrounding rim of Monte Oscuro, Panama. 


Poaceae and other Herbs —B—QÀ € Arboreal 


d o o 
S 
S & a Rg E uS 
Pe IS 
S Se OOO > E 
AS SS CS ds ES 
AN A oÑ A oS 


A 


20 20 40 20 20 40 60 20 20 40 60 80 20 40 60 80 


2 
A PIN : n 
Lake Ashand Ephemeral Soil Subsoil Volcanic 
Sediment — Gravels ter Ash 
Figure 3. The percentage of each major phytolith taxon found at Monte Oscuro. Panama, The Cl bal: ae/Othei 
Vrboreal category contains predominantly Chrysobalanceae phytoliths (usually > 859-90%). with UE URS also from 
nd Flacourtiaceae. Percentages for 


loraceac, 5 n nonaceae, Connaraceae. a 
Only short cell phytoliths from grasses (egi bilo 
‘aya attributions are as follows: C husquea Kunth. (Poaceae). Re oid jd Pon m Pipe rno and 


Protium. other Burseraceae. 
l ates, saddle-shapes) 


Jones (2003), with permission from Elsevier. 
annual rainfall presently is 1.8 m (Piperno & Jones. 


The paleobotanical records from a few of these sites 
2003: Figs. 2—4). This sequence is the first from 


are now discussed in more detail to better illustrate 
the interpreted findings The author and John Jones 
recently studied phytolith and pollen. records from 


Monte Oscuro, Panama, a large lake located in an 


Panama dating to the Pleistocene from the seasonally 
dry Pacific coastal plain. where the potential 
vegetation is a deciduous tropical forest and the 
l 


extinct crater 75 km west of Panama City where forests, therefore, would be particularly prone 


Volume 93, Number 2 
2006 


Piperno 
Quaternary Environmental History 


279 


Monte Oscuro Pollen 


Herbs 7 — Arboreal — —— 
J " 
i Le Se D S 
SCS « @ AÀ 
oS & SS Sows Ser 
* ORI qo 
560 7 
50 
100 
150 
2290 - = 200 
o 
ES 250 
4750 - 300 
350 
400 ` 
7500 450 ] i : E 4 —— "rns 
20 20 40 60 20 40 20 20 20 40 60 80 20 40 60 
Figure 4. the percentage of each major pollen taxon found at Monte Oscuro, Panama. Cheno-Amar refers 
Chenopodiac . Taxa 1 are as follows: Bombacopsis (Bombac aceae) Pittier. Reprinted from 


Piperno n lon s (2003), with permission from Elsevie 


fragmentation under conditions of significant climatic 


drying. The sedimentary sequence from Monte Oscuro 


ake bed 


was persistently dry, such that a paleosol formed and 


shows that during the Late Pleistocene the | 


supported an associated vegetation. Pollen nol 


is 


preserved in these sediments because they were 
subjected to oxidation, but phytoliths are common, 
The phytolith data from the paleosol indicate that 
during the Late Pleistocene, the Monte Oscuro region 
grasses, 
The 


archetypal indicator of savanna in the Neotropics, the 


supported a vegetation containing many 
sedges, and other herbaceous plants (Fig. 3). 
shrub or small tree Curatella americana L., is also 
present during this period. Trees that are character- 
istic of the modern deciduous forests of the area 
appear to have grown in low numbers. Grass phytolith 
that 


grasses) and Panicoid (tall grasses) sub-families are 


morphology | indicates the Chloridoid (short 
common in the phytolith assemblages. These grasses 
predominantly use the C4 photosynthetic pathway 
(only a small proportion of modern Panicoid grasses 
are Cs). further adding to the picture of a dry and open 
arcoal (not shown) and burned grass 
that the 


landscape was periodically subjected to fires. During 


landscape. Ch 


phytoliths (Fig. 3) also occur, indicating 


this interval, arboreal phytolith taxa are represented 


balanaceae and Arecaceae. 


mainly by the Chryso 


There are a few phytoliths from Protium Basm. f.. 


the Annonaceae, and the Moraceae, indicating these 


— 


trees were also growing nearby. 
At an estimated date of ca. 10,500 BP (at a depth of 
7m), a substantial water table rise and initial ponding 


occurred, leading to the establishment of a permanent 


The 
for the filling of the Monte Oscuro crater is consistent 
lake 
America that were dry or much lower during the Late 
1999: Ledru et al., 1998 

Both the phytolith (Fig. 3) and pollen (Fig. 4) records 


waler body that lasted into the modern era. date 


with the inundation of other beds tropical 


Pleistocene (Curtis et al., 


show that during the early Holocene a diverse tropical 
deciduous forest expanded into the Monte Oscuro 


By 


sedge phytoliths had declined greatly 


watershed. 8300 BP, percentages of grass and 


from their 


Pleistocene occurrence frequencies and Curatella 
By shortly after 7500 BP. 


when sediments first contain high amounts of pollen, 


is no longer recorded at all. 


a number of trees occur in the pollen assemblages 
whose modern ecological affinity is seasonal tropical 
L.. Protium. 
of 


axa 


forest (e.g... Zanthoxylum L., Spondias 
La; Celirs dua 


Moraceae; arboreal 


Bursera Jack., Erythrina species 
: | 


Anacardiaceae and other 


such as Trema L., Machaerium-type, Cavanillesia 


Ruiz & Pav.. 
(Fig. 4). 


During the Pleistocene, 


and Calophyllum, which are not shown) 


the Monte Oscuro region 


supported open kinds of vegetation consisting of 


e 


a large number and variety of herbaceous taxa, 


including grasses and sedges, and small shrubs and 
trees, including Curatella and palms, adapted to dry- 
land habitats. Trees that form important components 
of seasonal tropical forests in the region today were 
not entirely eliminated from the landscape but appear 

have grown in low numbers. Some of them were 
probably growing along the margins of water courses. 
amount of modern 


1800 mm, 


In light of these factors and the 
annual precipitation the area receives, TET 


Annals of the 
Missouri Botanical Garden 


likely that annual rainfall in the Monte Oscuro region 


was reduced during the Late Pleistocene by at least 


35%. It is also reasonable to extrapolate from the 


Monte Oscuro record to predict that a significant part 


of the Pacific coastal plain of the Panamanian land 
bridge was a dry and open habitat. It might be unwise. 


however, to use the word savanna to describe the Late 


Pleistocene vegetation. Thorny serub associations 
(those with many legumes, cacti, and also some 


euphorbs) produce few phytoliths that can be 


identified even to the family level, and are thus 


difficult to document in phytolith records. It is likely 
thal 


landscape, 


these were the 


types of plants present 


perhaps in considerable number, and. 
along with the plant taxa that can be documented 
through their microfossils, formed species combina- 
lions that were well adapted to Pleistocene dry 
climatic phases but that do not at the present time 


grow together. 


How heterogeneous and probably patchy the 
Pleistocene vegetation was across small distance 
scales in regions such as Panama, where one can 


move from tropical deciduous to tropical evergreen 


ride, is illustrated 


forest in a leisurely 90 minute ca 


by the pollen. and phytolith sequences from La 


Yeguada and El Valle studied by Mark Bush, Paul 


Colinvaux, and the author (Figs. 5, 6). These sites are 


located in the Pac 5 walershed uplands between 500 


and 650 m asl (Fig They receive between 3 and 


Em of rainfall ma today. twice the amount as on 


the Pacific coastal plain where Monte Oscuro is 


located. The climate is still highly seasonal at these 


two sites: over 90% of the rain falls during the wet 


season and there is a long and marked dry season of 


five months’ duration. In each case. the Pleistocene 


vegetation was a forest containing trees now found 
mainly at elevations of 1500 m and above, such as 
Quercus L., Hex le, and Magnolia L. (Fig. 5). 


However, presence of the clay type illite—a mineral 
that typically forms when the soils and rocks of a lake's 
catchment are little weathered from rainfall—in the 
Pleistocene but not in the Holocene deposits at Lake 
1992) is a good 


Pleistocene 


a Y eguada (described in Bush et al.. 
that Late 


prec ipitation 


he 


indication during the 


region's was considerably less than 
today’s. At El Valle, a surge in Chenopodium L. and 
Alternanthera Forssk. 19.000 and 
12,000 BP indicates the lake contracted severely at 
thal 


drying (Fig. 6: Bush & Colinvaux. 


pollen between 


lime, almost certainly a response lo climatic 


1990). 
The most recently generated data dating to the 
Pleistocene from the Central American lowlands. which 
are still in a preliminary state and cannot be discussed 

detail here, come from the Central Balsas River 


Valley located in tropical southwestern Mexico, where 


no information on the paleoenvironment was pre iously 


available. The author began a project there five years 


ago to reconstruct. the environmental and cultural 


history of the region. This is where the extensive 


molecular work carried out by John Doebley and 


colleagues during the past 15 years places the origin o 


maize, whose ancestor they have determined to be 
a species of teosinte, Zea mays L. 
Iltis & Doebley, native to the pice Balsas region 
1990: Matsuoka et al., 2002: Fig. 7). As 


‘the Pacific watershed of C ius \merica, 


subsp. parviglumis 


(e. g., Doebley, 
over much of 
the potential vegetation of the region is a tropical 
information. from two 


deciduous forest. Preliminary 


akes, Ixtacyola and Ixtapa, located in and near the 
lguala Valley on the eastern side of the Central Balsas 
region ab an elevation of 700 m asl (Fig. I) suggests 
that there, also, the Late Pleistocene climate was much 
cooler and drier than today’s and the glacial vegetation 
trees than did early Holocene 


contained fewer 


associations (Piperno et al., 2004). 


Wir Have We 
PLEISTOCENE 


SUMMARY: LEARNED ABOUT 


ENVIRONMENTS? 


The Pleistocene climate in the lowland Neotropics 
is now understood to have been a peculiar combina- 
tion of low temperature, precipitation, and atmospher- 

The cumulative effect of these conditions on 
tropical plant communities and whether all were 
effecting community structure and floristic composi- 
lion to degree, are under 


an equally significant 


considerable discussion. It has been suggested that 
low CO» concentrations alone might explain much of 
the displacement of forest by herbaceous vegetation 
seen in the Pleistocene lake records from Central and 
Mayle & 


lo more 1 study this 


South America (e.g., Colinvaux et al., 2000; 
Beerling, 2004). 
issue, Huang et al. (2001) analyzed leaf wax carbon 
Mexico (Lake 


Babicora). and Peten, Guatemala (Lake Quexil). It is 


In order 


isotopes in lakes from Chihuahua. 


well understood that the Chihuahua and Guatemala 


regions were, respectively, fairly wet and very dry 


during the Late Pleistocene due to being under the 


influence of different sources of moisture. Today. the 


wo regions still experience contrasting precipitation 
f what they were 
(2001) were 


Babicora there 


patterns. which are the opposite of 


during late glacial times. Huang et al. 


able to show conclusively that at Lake 
was much more C} vegetation during the late-glacial 
period than during the Holocene. The opposite was 


true at Lake Quexil, where, in accordance with the 


palynological data indicating the presence of open. 
dry-land environments, the Late Pleistocene vegeta- 
C4 plants than 
that 


tion was shown to contain many more 


did the Holocene vegetation. The wet climate 


Volume 93, Number 2 Piperno 
Quaternary Environmental History 


006 


aa o 
e 
sod x e 
x o 
os Qu "a e e 
= uv 
rc: 8 G. V e D QU g^ Se 
a SOS ó 2S os S > P 
Q SV S EEN 9 X ug 
D 9 Meer nS SIS XS $ > 
$ à VPS g ¿ES A e «e 
Ex La | | 
| ! 10 20 30 40 50 60 70 80 10 20 30 40 50 60 I 10 2 0 40 50 60 70 80 90 100 110 120 130 140 150 160 170 


10 20 | 10 20 30 40 50 60 70 80 90 


E 


| N 
| T" 
| P. 
| "d 
| Jn 
| P d 
» d 
A d 
1700 10 10 20 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 | 10 20 30 40 50 60 70 80 | 10 20 30 40 50 60 | 70 
PN S S 
Ge 4$ $ 
| g y N 
— 8 J KS 9 
= o? v f 
2 E o go S SH 
O Ss e o > Ò 
=e o? S 4 9 e 
m. l æ d d $ e 
S xx 
S 8 e Se N e S 
“eer eee A peered ter AAA F TTF ice as ee E 
0 W 1 * saco vanos". pa» SS % » n» aucun m m m» s» €- m min im tM imu 
] : | . 
1004 AAA 
500 * 
200 4 x 
1000 
300 4 A : 
400 y 
2000 
500 4 
3000 => 
600 y 
3500 
700 y Pz 
5000 
6000 800 + 
7000 
7500 900 - 5 
8000 1000 | 
90001109 | 
10,000 


1200 y 
11,000 
1 
12,000 1400 | 
1500 — 
13,000 
00 
ä ß , MNT OEC EE 
F IAI i 1-5 LLL LLL LLL LLLLLLCLLLLLLLLLL I ILLI LS 


La Yeguada, Panama. Taxa 
Perebea Aubl. 
99] b). 


A e EE influx (number of phytoliths per 20 em? per yr. : diagram from Lake 
(Magnoliaceae), Carex sp. ined. aff. pe olysticha Boeck (Cyperaceae). 


(Hyme oe eae). From Piperno et al. 


‘igure 5a. 
attributions are as follows 
(Moraceae), Prottum 1 f. 


: Magnolia 
(Burseraceae), Trichomanes L. 


282 


Annals of the 
Missouri Botanical Garden 


b d o ao 2 8 E: 
99 y 8 8 88 2 2 „ $ 
o 8 g 2 $5 E S g 
$ E pu c 
M 3 3 S 8 $2 — S 2 E 
— S2 2 GS neg o S — 2 
2000 4 
4000 
| J 
8000 - I 
10,000 
r 
12.000 * a 
14,000 | 
LL L 1 1 4 L aa n 1 
5 10 5 5 
Pollen influx grains cm^ yr x10° 


Figure. 50. 
allributions as follows: 


thus had far more 


Lake 
the 


atmospheric CO» concentrations did. 


prevailed al Babicora 
influence on regional vegetation. than the low 


Therefore, vegetational shifts from C4-dominated 


forests to Ci-dominated grasslands in the tropics 
significant 
CO, 


concentrations acting largely separately from climatic 


cannot be driven without concurrent 


reductions. in precipitation. Furthermore, low 


influences probably cannot explain why many pres- 
ently deep lakes from Mexico to southern Brazil were 
dry lake beds or shallow water bodies during the Late 
g. Ledru et al., 1998: Leyden. 1984 
Leyden et al., 1993; Piperno & Jones, 2003; Piperno 
2004) and why 


substantially reduced 


Pleistocene (e. 
el al., incontrovertible 
precipitation, such as the 


evaporitie mineral gypsum and the clay illite, formed 


in some of them (Bush et a Leyden et al., 


1993). While low CO» may certainly have exacerbated 


drought stress and further promoted the expansion of 


Cy and CAM open-land plants at the expense of some 


A pollen influx (number of pollen grains per en? y! 
Typha L. (Typhaceae). Limnocharis Bonpl. m recae). From Piperno et al. (1991b). 


signals of 


X107) diagram from La Yeguada, Panama. Taxa 
b 


C3 forest species, the tropical ice age climate appears 
to have been substantially cooler and drier. 

Formerly thought to be incompatible physical 
forces of the Pleistocene climate, cooling and drying 
can now be understood to have been mechanistically 
linked. The f 


evaporation that took place over cooler tropical ocean 


amount 


and causally negligible 


surfaces, for example, meant that much less water 
vapor was available to become rainfall over adjacent 
land surfaces (Webb et al., 1997). The persistence of 
forest in higher precipitation areas such as at La 
Yeguada and El Valle, 


and others think was probably a 30% to 40% decline 


even under what the author 


of precipitation, is not surprising because sufficient 


moisture would still have been available to support 


species-diverse forests. Furthermore. in these Cx- 


dominated forests, cooler. Pleistocene temperatures 


probably would have counter-balanced the effects of 
lower precipitation by improving carbon and water 
| | À | ¿ 


Ul eradients 


balance and reducing € | l | 


EL VALLE PANAMA 


— Arboreal — Herbs —— — Aquatics — Spores —> 
Es 
E š o o o 
a a o 6 b: 
— en E a a T c 2 8 g E: D c o 
> One 3 p o Es D = o 8 E 
E £ £ 8282 226.92 E23 B 2 285% s $ 8 i 8 2 8 
a 2 S 283 3 F28e8 ge 5 E S 8 6 S S S $ 8 9 p 8 8 3 
7 $99.98 8 8 58 8 989 9 E E & 5 S 5 £ 2 E E: 2 £ 3 +8 8 
S 2355 S S 8 ® EEE S s 225 £ E ES £2 9 2 £ SE 3 
8350 + 150 S S FES >55IO=>= Fas O = S S SS E zi [e Our mou 8 28 
riL 14,800 +250 3-9 — + - J v5 
110 19,400 + 450 P r P 2 " 
$ 4 F T . 
1800 « 1800 M | | | | | : b ? us 
: 535,000 b ` i bo US "opu an EN i 
«+ 30,700 + 800 15- 4 
È ) , i > > ! | : i 
> 2- > IE. | WM SG d Ap m Rh | 
X b 
8 | EV3 
x 25 — Y L I 
s 2 y ‘ i 
N 1834 > 8 b i | 
i 30- d | 
A i | !| y 
235560 35 117 } 
id » > d E 1 | 2 4 t 
40-f ) 38! ) > 1 =e 
& | b | > Ev 2 
“| b II ) 
s0- \ 
55— 1 1 L 1 1 1 L L 1 1 1 L L 1 ü k uk L L A L 1 t L 
10% 10% 10% 10% 10 20 10% 10% 10% 10% 10 20 10 20 30 40% 10% 10% 10 20 30% 100 200 300 
Gyttja Cay [III Laminated gyttja E Sandy clay Sandy gravels a concentration grains cm? X10 
+=< 1% 


Figure 6. The percentage of each major pollen taxon found at El Valle, Panama. Total pollen concentrations are shown at the right. Taxa attributions are as follows: Hura L. (Euphorbiaceae), 
Myrica L. (Myricaceae), Byrsonima Rich. ex Kunth (Malpighiaceae), Alternanthera Forssk. (Amaranthaceae). Reproduced with permission from Bush and Colinvaux (1990). 


9005 


c JequinN ‘£6 SunjoA 


ouJadid 


KOSIH [Je3ueuiuoJiAu3 Meuran 


88 


Annals of the 
Missouri Botanical Garden 


Figure 7. The wild : 
and photorespiratory carbon loss (Cowling & Sykes, 
1999). In contrast to forest persistence in high rainfall 
zones, forests were often displaced by savanna/thorny 
scrub kinds of vegetation during the Pleistocene, where 
the modern potential vegetation is a highly seasonal 


forest with annual rainfall of 1.2-2 m a year. Here, the 


incestor of maize, Zea mays subsp. parviglumis, growing in the Central Balsas River Valley. Mexico 


dry climate, aided by the effects of low COs, resulted in 
landscapes where forest growth was minimal. 

The Central and South American lowland tropical 
forests are thought to house more plant species than the 
tropics of Asia and Africa combined (Richardson et al., 


2001). There are increasing indications that a good deal 


OR e 


520 + 60 - 


3920 + 90 - 


4790 * 90 - 


Depth 
d 


5630 = 100» 


7290 + 1401 
35 1 
20 20 40 60 80 20 40 60 
450 
Figure 8 p 
Swamp, Belize. E ultigens are maize, ton (Gossypium | 


ions are as follows: Cladium E 
99 


rom Jones (1 Lj: 


Se 
S80" 80 p Ec 


| 


ollen profile (perce E abundance of gs nat depth in em below surface: age in 


rowne (Cyperaceae), Borreria G. 


Tr 


20 40 60 80 


DOT! ME EN — — — . — — 

20 40 20 20 40 60 20 20 40 
~ BP) from ( ;iobwe b 
taxa 


(Malvaceae)). and chile pepper 1 (Sol. anaceae)), 


Mey (Rubiaceae). Re 1 with permission 


Volume 93, Number 2 
2006 


Piperno 285 


Quaternary Environmental History 


Sg 
s SS 


es 


S 
ch 95 
See a 


PPP eoccce co o o o coco 099 0000699 


20 20 40 60 80 


425 | 


pollen profile (percentage abundance of pollen at depth in cm below surface: 


EM — p 


80 


20 40 60 


20 


. BP) from Kob 


age in 


Figure 9. A 
Swamp. Belize. Cultigens are manioc (Manihot esculenta Crantz (Euphorbiaceae)) and cotton. cant e d Sh permission 


from Jones (1994). 


of this diversity is pre-Quaternary, perhaps dating to 
the Eocene (Knapp & Mallet, 2003: Wilf et al.. 2003), 
although recent investigations also show that some tree 
species of Central and South American seasonally dry 
forest along with speciose genera of evergreen and 


orests in South America probably did speciate 


ur 


other 
during the past few million years (Richardson et al., 
2001: 2004). Whether 
specialions occurred in islands of 
lowland moist forest surrounded by savannas (Haffer. 
1969; Whitmore & 1987), been 


completely resolved. However, interpretations of Late 


Pennington et al., those 


refugia, intact, 


Prance, has not 
Pleistocene vegetation that rely on refugial theory 
appear increasingly less plausible as more paleoeco- 
logical data accumulate. indicating that due to drier 
conditions postulated refugial areas were nearly devoid 
of trees, or that these areas contained many colder- 
adapted taxa presently confined largely to 
forests (Colinvaux et al., 2000; Knapp & Mallet, 2003; 
Leyden, 1984, 1985). 

It is relevant with respect to refugial theory that 
Panama, 


montane 


phytolith evidence from Monte Oscuro, 
discussed above, located in a highly seasonal area 
that was sparsely forested during the Pleistocene, does 


point to the presence of a number of lowland forest 


— 


as Protium and other Burseraceae, 


Annonaceae, Chrysobal 
the lake during the driest times before 10,000 BP 


trees such 


ae, and Moraceae near 


There were almost certainly other tree species present 


there and at other similar locations that for various 


reasons (low pollen productivity or an absence of 


identifiable phytoliths) are not visible in the paleo- 


botanical records. Some arboreal species may, 
therefore, have grown in lower densities than is usual 
today in more mesic pockets on the drier Pleistocene 
landscapes; for example, they may have been 
harbored alongside water courses. Studies of modern 


Kellman (1994) and 


these linear of 


riparian forests by Meave and 
that 
much greater diversity of tree 


others have shown stretches 


vegelalion contain a 
species than was previously thought possible. Sir 
Ghillean Prance's (1987) 


were important refuges for low 


idea that riparian forests 


trees in 


— 


and forest 
Amazonia thus has considerable merit, and his view 
applies equally well to Central America. These 


findings and ideas on the historical significance of 
riverine forests should also be remembered by modern 
planners when they are considering what the sizes, 
shapes, and species composition of modern reforesta- 
tion programs should be 

We've learned that the questions we first asked of 
the historical records—was the Pleistocene vegetation 
a forest or was it a grassland; were the lowland tropics 
cooler, or were they drier during an ice age: where did 


trees go during periods when savannas and scrub 


lands appear to have dominated landscapes now 
underneath forest? —were far too simplistic. We have 
begun to consider more realistic, if increasingly 


286 


Annals of the 
Missouri Botanical Garden 


complicated, Pleistocene scenarios involving the 


likelihood of different forest facies: the existence of 
species combinations not presently found together: 
heterogeneous, even patchy, vegetation across small 
distance scales: and stretches of forest alongside water 
courses in regions where forests were significantly 
reduced that may have held and preserved at least 
some of the diversity we see today during the past few 
million years. On average, glacial periods last four to 
five times longer than interglacials. Hence, during the 
past two million years, tropical forest. plants and 
animals spent over 80% of their time in, and must 
have been well accustomed to, a world of lower COs. 


Our 


canopied forests appear to be little more than short- 


temperature, and precipitation. modern, closed 
term, interglacial floristie associations that coalesced 
or expanded onto landscapes the last time beginning 
11,000 to 10,000 years ago. | 
key to the present, then what we have learned about 
ook to 


among other 


—ͤ— 


the past is indeed the 


— 


tropical forest history should cause us to 
ancient records for advice concerning. 
things, modeling plant succession following major 
ecosystem disruptions and predicting and patterning 
forest regeneration when planning regrowth programs 
see also Birks & Birks, 2004, and Wardle et al., 
2004). 


PRE-COLUMBIAN HUMAN IMPACTS ON THE 
NEOTROPICAL FOREST 


During the final 1000 years or so of the Pleisto- 


cene, the climate warmed markedly and became 


substantially wetter and more highly seasonal. and 
atmospheric concentrations of CO» increased, creating 
conditions in which tropical plants could really 
Humans had already arrived in Central and 
South 3000 years 
demonstrated by Dillehay's (1997) excavations at 
Monte Verde, Chile. 
Clovis First paradigm for human colonization of the 
New World, is 
a 13,000 South 
America by a strong consensus of American archae- 
1997). At the close of the Pleisto- 


human occupants. of Central and South 


flourish. 


America by at least earlier, as 


This work, which overturned the 


acceptec 
BP human presence in southern 
ologists (Meltzer, 
cene the 
America, having spent the late glacial period in cool 
and dry landscapes where the amount of forest cover 


was far less than it is today and large. slow-moving. 


now-extinet game animals abounded, found them- 
selves in surprising and dramatically changing 


ecological circumstances. Tropical forests. with their 
far fewer and much smaller animals and more highly 
dispersed and toxic plant species, were expanding 
rapidly on landscapes, and new strategies for living in 


these habitats would become necessary in fairly short 


as convincing proof of 


One cultural adjustment to the changing 


ecological circumstances that turned out to be a highly 


order. 


pus 


successful adaptation was plant food production. 

The transition from a hunting and gathering way of 
life to one which human societies became de- 
pendent on a food base underwritten by plants they 


domesticated was accomplished independently in at 


least eight regions of the Old and New World, 
including both hemispheres of tropical America, nol 
long after the Pleistocene ended (Diamond, 2002: 


1998; Kennett & Winterhalder, 
20006). It marked the emergence of humans as owners 


Piperno & Pearsall, 


and extraordinary manipulators of their landscapes. In 
southern Central and northern South America, the 
best studied regions to date, microfossil (phytolith, 
starch grain, and pollen) research indicates that the 
domestication and spread of important native crops 
like maize. manioc (Manihot esculenta Crantz). two 
species of squash (Cucurbita moschata Duchesne and 
C. Cutler & Whitaker), arrowroot 
(Maranta arundinacea L.). yams (Dioscorea trifida 
. .). (Calathea (Aubl.) Lindl.) 
occurred between 10,000 and 5000 years ago (Bray, 
2000; Mora et a 2001: 


C. ecuadorensis H. 


and líren allouia 


991; Pope et al., Piperno, 


2006a; Piperno & Pearsall, 1998; Piperno & Stothert, 
2003: Piperno et al., 2000a, b; Pearsall et al., 2003, 
2004). Furthermore, together with evidence from 


this tells us that the 


both Central and 


molecular biology and botany, 
most important lowland crops 
South 


tion and domesticated in the seasonal, 


America were originally brought under cultiva- 
o » D 

nol the ever- 

wel tropical forest, and that the interior of the Amazon 


Basin had little at all to do with the origins of major 


subsistence crops. (e.g. Matsuoka et al., 2002; 
Olsen & Schaal, 1999; Piperno & Pearsall, 1998: 


Piperno, 2006a; Sanjur et al., 2002: Westengen et al., 
2005). 

Paleoecological data from lakes and large swamps 
often located near important archaeological sites make 
it clear that the development and spread of agriculture 
in the American tropics exerted profound influences on 
the structure and composition of the vegetation. The 
numbers of sequences showing human agricultural and 
other alterations of the lowland tropical forest in 
Central America during the early and middle Holocene 
are rapidly increasing, and not all of them can be 
considered here. It is briefly noted that in addition to 
the records discussed below, good data also exist from 
Costa Rica, El Salvador, and coastal Guatemala. as well 
as from Puerto Rico (see, for example, Clement & Horn, 
2001: Dull, 2004: Horn & Sanford. 1992: Arford & 
Horn, 2004). Virtually everywhere the records show 
that fire was an important instrument of vegetational 
modification for hunters and gatherers and people 


practicing agriculture alike. 


Volume 93, Number 2 
2006 


Piperno 
Quaternary Environmental History 


287 


At Lake La Yeguada in Panama, a disturbance 
horizon indicating frequent firing of the forest and 
small-scale vegetation clearance begins at 11.000 BP 
after being absent for the first 3000 years of the lake’s 
history, when the climate was much drier (Piperno et 
al., 1990; Bush et al., 1992; Fig. 5a). The disturbance 
is represented by a large (several orders of magnitude) 


and rapid increase of charcoal and plants of forest 


[om 


gaps such as Heliconia and grasses, and it is sustainec 
throughout the early Holocene. The initial appearance 
of the the 
settlement of and 
gatherers. These and later populations who occupied 
the La 11,000 
10,000 BP left numerous characteristic stone artifacts, 


including a projectile point on the shore of the lake 


vegetation disturbance coincides with 


the region by Clovis hunters 


Yeguada watershed between and 


itself, that have been found by archaeologists during 
systematic surveys (Ranere & Cooke, 2003) 

More than 70% of the disturbance taxa phytoliths 
11.000 to 9000 BP, 


providing direct indications that they had been fired 


are continually charred from 
repeatedly (burnt phytoliths can be easily recognized 
because they acquire a black coating on their surface: 
see Piperno, 2006b). Such a scenario would be highly 
unlikely under burning caused by natural processes 
because fire return rates in tropical forests documen- 
at La 
Yeguada, the sampling resolution was often about 
100 years). Furthermore, ENSO became a part of the 
post-Pleistocene climate system only after 6000 to 
5000 BP (Moy et al., 2002; Sandweiss et al., 1996), 
and thus persistent. fire events during the early 


E" 


ted for the modern era are never that frequent 


Holocene cannot reasonably be attributed to El Niño 
dry periods. Hence it appears that early hunters and 
gatherers repeatedly subjected the forest around La 
Yeguada to fire and small-scale clearing. 

The disturbance at La Yeguada continued unabated 
and intensified. Then at 7000 BP a marked decline in 
arboreal and associated species characteristic of older 
forest growth (which the phytolith record is more 
adept than pollen at identifying) and continued high 
frequencies of charcoal indicate the development of 
slash and burn cultivation. Pollen data, which are very 
sensitive to a set of woody secondary forest taxa 
generally not visible in phytolith profiles, 


Pilea Lindl., 


indicate a concurrent substantial increase « 


such as 
Trema, Ficus L., and Cecropia Loefl., 

i , 
f such 


— 


trees (Fig. 5b), confirming the decline in older forest 
growth recorded by the phytoliths. At 4000 BP, even 
these pollen types decline greatly, as does charcoal, 
indicating that an intensification of forest clearance 
taxa were left in the 
7000 BP, 


when the earliest clear indications of slash and burn 


occurred and not many woody 


watershed (Fig. 5a, b). Also starting at ca. 


cultivation begin, nearby archaeological sites contain 


EI Venancio, 
„ An Bin Ch 


Central Balsas Valley 
Zea mays ssp. parviglumis 
i pu Guila Naquitz 


watershe 


Figure 10. Map of the Central Balsas 
Mexico showing the Iguala Valley (grey shading) and coring 

Juntas de Río Chico. 
archaeological 


Venancio and Las 


Valley and Guila Us are 


locations El 
Tehuacan 

settlements located in the dry 
research has documented early bal 11 tion, 


nds where previous 


the first phytoliths and starch grains from maize and 
that 
arrowroot and Cucurbita moschata were also present 
by that date (Piperno, 1988, 2006a; Piperno et a 
2000b: Piperno & Pearsall, 1998). 

At about 350 BP, when Europeans are known to 


have first entered the region, a forest resurgence is 


manioc. These records show crops such as 


evident at La Yeguada. It is almost certainly a result 
of the terrible decimation of indigenous populations 
that occurred soon after the Spanish contact on 
account of disease, warfare, and slavery. Hence, the 
forests around La Yeguada were resilient to long-term 
prehistoric agriculture. Even so, some of the trees first 
recorded during the early Holocene do not appear 
again when indigenous population pressure decreases 
(e. g. Protium), while other taxa, especially forest floor 


herbs such as the Marantaceae, appear to be present 


1 higher frequencies than they were before the 
vegetation was fired and cleared. 

Similar early manifestations of human firing and 
clearing of tropical forest have been demonstrated 
from Mesoamerica to the Amazon Basin. For example. 
two pollen and charcoal records studied by John Jones 
from swamps called Kob and Cobweb, located about 
55 km apart in northeastern Belize, show an early 
phase of intensive deforestation resulting from slash 
and burn agriculture (Jones, 1994; Pohl et al., 1996; 
Figs. 8, 9). The sites are near major pre-Classic 
archaeological occupations, and both 
pollen greatly declines and maize is present by 4000 
BP At 


increases of charcoal fragments (not shown) and early 


— 


cases tree 


this time, there are also sudden, major 


successional herbaceous taxa. Among arboreal taxa, 
even pollen from Brosimum Sw. nearly disappears, 


indicating that the fruit of this tree, called rámon, 


probably was not a primary food source for Classic 


EL VENANCIO 
POLLEN DIAGRAM 


(Bulk) 4090240 
(Phyt) 42304290 


Ta 


(Fabace > Ale hor 


vA CUA 


Th 


"C yrs BP 


e] 
nec 


1 


erce 
Sw. 


Pinus 


Cav. (Malpighiaceae). 


Arboreal 


ES 

. S 
S . 
Lo . PES DEI Q 00 ou N 
NF Sy ES N S 
E S NUS SF 4 Pe o, D 9 oN 
PP ue. OS Lye AO NAS: 
PFO PEE CS ox g IESO 
"PEBEEESS ELENA 


i io 10 10 19 10 10 


il Fine compact 
clay 


ntage 


and Croton L. 
(Pinaceae); Piper L. 


r3 
lo 0 


ffffff 


. 


r 


| 


| 
| 


10 20 30 40 50 eo . 


4 


ar" Edi 


Hu 


f E 
10 1020 


— 


10 20 30 40 BO BO 70 BO BO 100 


Io 


| 
| | 
— — | 
$0090 100000 150000 
Pollen groins/cc 


| 
I 


5% exaggeration. 


Percentages based on Gymno/Angiosperm pollen sum. 


(Euphorbiaceae): 


| 
Rh! 
Mo! 10 0 10 10 10 10 PO YO 40 50 80 TO 10 10 10 vw 10 10 
Note: Hatched areas = 
Oxidized Weathered 
cloy bedrock 


abundance of pollen of each taxon found in 


(Piperaceae): 


Polygonum L. 


El Venancio, 


(Polygonaceae): 
eee} 


Mexico. 


Sida | 


Taxa attributions are 
: Cuphea P. Browne (Lythraceae): Cyperus L. (Cyperaceae); F 


(Malvaceae); 


orsteronia P 
Smilax 


as follows: 


Acacia L., Pithecellobium Mart.. and Pterocarpus Jacq. 
Mey. (Apocynaceae): Hyptis Jacq. (Lamiaceae); Ipomoea L. 
(Siler 'aceae): Spathiphy ilum Schott (Araceae); Tetrapteris 


uapled [eoiuejog unossi|A 


880 


ay) Jo sjeuuy 


Volume 93, Number 2 
06 


Piperno 289 
Quaternary Environmental History 


Maya peoples, as was thought before Jones's pollen 


work and other studies showed otherwise (Jones, 


1994). As in central Pacific Panama, the fertile soils 


of northern Belize were able to sustain intensive 
cultivation. for thousands of years. Agricultural 


activity stops at about the time of European Contact 
at the Belize sites, but in those cases the forest did not 
recover. The reasons why are presently unclear. 

T 


drainage of 


— 


ae seasonal tropical forest of the Rio Balsas 


Mexico appears, based on molecular 


studies of modern crops and their nearest wild 


relatives, to have been an important center of crop 
plant origins, where maize, the silver-seeded squash 
(Cucurbita argyrosperma Huber), and other plants 
2002; Sanjur el 
litle archaeological o 


were domesticated (Matsuoka et al., 
al., 2002) 
paleoecological research has been undertaken there, 
that 


assumed that crop plant evolution largely occurred 


However, very 


owing largely to the fact investigators long 


in the dry mountainous zones of the 
Therefore, virtually nothing is known of the earlier 
segments of pre-Columbian human history (ca. 12,000 


BP to 3000 BP) in the Balsas watershed. During the 


past five years, the author has directed reconnais- 


country. 


sance and coring of lakes and swamps in the Central 
Balsas region, where the modern populations of Zea 
mays subsp. parviglumis are genetically closest to 
2002; Fig. 10). The 


Valley and its environs, located in Guerrero on the 


maize (Matsuoka et al., Iguala 
eastern side of the Central Balsas drainage. appears to 
have been a smaller-scale version of the lake district 
in the Petén, Guatamala, as there are large lakes and 
numerous other basins that held permanent water 
until recently (Piperno et al., 2004). The 
date to 


sequences 


he Late 


from two of these sites, which 
Pleistocene, are still being investigated in the author's 
laboratory and cannot yet be discussed in detail. 

In the Central Balsas watershed, about 120 km west 
of Iguala near the town of Ciudad Altamirano, still in 
Guerrero but close to the border of Michoacán, one 
large swamp and the remnants of an oxbow lake were 
located and cored. These sites are called, respective- 
ly, El Venancio Las Río Chico 
(Fig. 10). The elevation of the area is 300 m asl, 
annual 1400 mm. The 


potential vegetation is tropical deciduous forest, but 


and Juntas d 


and precipitation is about 


the region is apparently too low-lying and warm t 
support the growth of teosinte, which has never been 
collected there. A 1.6 meter-long sequence retrieved 
from El Venancio has two concordant dates of 4090 + 
40 BP and 4230 + 290 BP obtained, respectively, 
from sediments and phytoliths retrieved from the same 
depth of between 90 and 100 cm 


beneath the surface (b.s.) (Fig. 11). There and up the 


sediments at 


core to about 30 em b.s., sediments are dark, organic- 


rich, silty clays containing abundant pollen grains and 
phytoliths. Beneath 100 cm, sediments are organic- 
poor clays largely devoid of any plant material. 

The pollen analysis, carried out by Enrique Moreno 
of the Smithsonian Tropical Research Institute, shows 
that by 4000 BP the vegetation contained few trees 
and was instead dominated by early successional 
plants indicative of landscape clearing, such as the 


(Fig. 11). 


time (and is 


— 


and Asteraceae Maize pollen 
that 


throughout the later portions of the sequence), along 


Poaceae 
occurs al present continually 
with abundant fragments of charcoal (not shown). This 


picture is repeated at nearby Las Juntas de Río Chico 


(Fig. 12). Here, a 1.5 meter-long sedimentary se- 
shows that the same taxa indicative of 


quence 
landscape clearing present at El Venancio by 4000 
BP dominate the pollen assemblages found at a depth 
of 110 to 120 em b.s., dated to 3050 + 40 BP Maize 
pollen and abundant charcoal particles occur here. 
Sediments beneath 120 em b.s. contain relatively few 
pollen grains, making it difficult to interpret. the 
vegetation with the degree of confidence possible in 
stratigraphically higher deposits. It can be said that 
the basal-most pollen, which on the basis of estimates 
derived from sedimentation rates probably dates to 


about 4000 BP, is mostly from grasses, and there is no 


arge spike of fern spores that would indicate 
differential degradation of pollen grains (the pine 
pollen is almost certainly a result of long distance 
transport). Importantly, there are also abundant 
charcoal particles in the deepest sediments. 


We 


Central Balsas region, thought now to be a cradle of 


thus have our first indications that in the 


agriculture in Mesoamerica and the hearth of maize. 
agricultural systems using maize that resulted in 
intensive forest clearing have considerable antiquity. 
The few trees and lianas remaining on the landscape 
t 4000 BP (Cedrela P. Browne, Spondias, Bombaca- 
ceae, Bauhinia L., palms such as Oenocarpus Mart.) 
provide but a small hint of the species-rich forest that 
existed before humans burned and cleared the 
vegetation. 

A discussion of the South American paleoecological 
records that speak to ancient human modification of 
the environment for agricultural and other practices is 
beyond the scope of this paper. However, one 
important pollen record studied by the prominent 
and her 


1991). 


The sequence comes from the wet evergreen forest of 


Colombian paleoecologist, Luisa Herrera, 


colleagues should be mentioned (Mora et al.. 


the now-remote middle stretches of the Rio Caquetá in 
the center of the Colombian Amazon (annual pre- 
cipitation > 3000 mm). It shows clearly that by 4700 
BP a slash and burn system of agriculture using 


maize, manioc, and other crops was well established 


LAS JUNTAS DE RIO CHICO 
POLLEN DIAGRAM 


— ers e 


A ————————————— — Lianas > 


/- Arboreal 7 = 


S 
x 
o 
ES 
n. DS 
m e 
o 
e iS 
> 0 2 > 
o e 3 © 
8 Y pou 
| | 
| 
| | 
| | 
| 
serena] | 
LR LLL) A AZ EL. il. „ a md LL 
1020 10 10 30 10 10 “Yo eo 30 40 50 80 7o 90 90 o 30 10 10. 10 20 30 40 89 80 10 10 "0 10 10 10 10 102030 10203040 5060708090 10 Yo zo 3040 10 8000 10000 1800020000 
Pollen grains/cc 
Sd Note: Hatched areas = 5% exaggeration. Percentages based on Gymno/Angiosperms pollen sum 
Soil Fine compact Oxidized 
clay clay 


The percentage abundance of oes n of each taxon found in Las Juntas de Río Chico, Mexico. Taxa attributions are as follows: Aechmea Ruiz & Pavón (Bromeliaceae); 


Figure 12. 
a (Fabace 


Evolvulus L. (Convolvulaceae): Mimosa 


O6 


əy} Jo speuuy 


uepier) jeoiuejog OSN 


900 


MAIN DIAGRAM 
ANNONACEAE 
Annona sp. 
Manihot 
Pouteria cainito 
Capsicum 
chinensis 
Theobroma 
bicolor 
Psidium sp. 
Zea mays 

c JequinN ‘E6 SUNJOA 


utilissima 


ZONATION 


4330 + 45 B.P. 
4695 + 40 B.P. 


29 Oo O [v] [s] 
» v — K 
% E NC S DT RES 
Md Wow Now VO ON Nw M ON OW UON 
acu aW Wu W uw. y 9M MN 
* x ON ON NON TT US 
e eo ow Q 32888988922 
«22228 8 92 8 £8 2 g g SZ 8 8 8282888335 
oars an oa MO en u A = 0 v A C) O — UO A O M Ur 
— 
— 
— a — 2 


E Š 314 
0% 25% 50% 75% 100% 0% 15% 30% 45% 
FAA Al OA? EDIT 


T 


Fine silty sand TROPICAL FOREST ELEMENTS 2 HUMIDITY INDICATORS 
ES Silt with fine sand LJ Slity sand E ANDEAN FOREST ELEMENTS OTHER FOREST 
ELEMENTS 


N N 
iiil Silty sand with charcoal Sand SAVANNA VEGETATION 


Figure 13. Pollen (the percentage of pollen of each taxon) and charcoal profiles 
" : 


os jejueuiuoJiIAu3 ÁleuJajeno 


found in the Río Caquetá region of the Colombian Amazon. Taxa attributions are as follows: Annona L. 
(Annonaceae), Manihot utilissima Pohl (Euphorbiaceae), Persea americana Mill. (Lauraceae), Pouteria cainito (Ruiz & Pavón) Radlk. (Sapotaceae), Capsicum chinense Jacq. (Solanaceae). 
Theobroma bicolor Bonpl. (Sterculiaceae), Psidium L. (Myrtaceae). Reproduced with permission from Mora et al. (1991). 


L6c 


292 


Annals of the 
Missouri Botanical Garden 


(Fig. 


had supplanted the species-rich forest. 


13). By this time, grasses and other field weeds 


Central American sequences, presence of abundant 
4700 BP testify that 


played an important role in field preparation. At about 


fragments of charcoal at fire 


775 years ago, when associated archaeological data 
indicate the site was abandoned by its human 


occupants, higher frequencies of trees returned. and 
This the 
Amazon can be contrasted to those from three distinel 
“Hill of Six Lakes” 


forest in 


burning ceased. record. from Colombian 


watersheds in the region, located 
in wel where 
annual precipitation is greater than 3000 mm (Colin- 
vaux el al., 1996; Bush et al., 2004). These sequences 


demonstrate that few forest fires 


evergreen northern Brazil 


and no recognizable 
human vegetational modification occurred in the area 
during a very long time period stretching from more 
than 40.000 years ago to the modern era. 

Sauer (1958) argued that humankind's successful 
adaptation the 


seasonally dry one, was aided greatly by the use of 


to the tropical. forest, especially 


fire. He presciently predicted that because Neotrop- 
ical trees are thin-barked and thus poorly insulated, 
having had apparently a 


short evolutionary history 
with fire, early hunters and gatherers and farmers 
alike could use fire as a highly effective tool for 


exploiting the forest. This made it possible to kill trees 


easily and create substantial openings or enlarge 
existing open areas without having to depend on 


a sophisticated stone tool technology (Sauer, 1958). In 
Panama, the Colombian Amazon, and other areas not 
discussed here where paleoecological records testify 
to the appearance of slash and burn cultivation 
several millennia and more ago, associated archaeo- 
that axes were nol 
commonly used before 3000 BP (Piperno & Pearsall, 


1998) 


logical data indicate slone 


SUMMARY: Wuar Have We LEARNED ABOUT ANCIENT 


HUMAN ĪMPACTS ON NEOTROPICAL VEGETATION? 
Human cultures had profound effects the 
the lowland tropical 


forest long before Europeans arrived. Human-induced 


on 
structure and composition of 
perturbations were often greater and more widespread 
than those brought about by Pleistocene climatic and 
Many 


herbaceous taxa were in large or smaller part removed 


other physical factors. arboreal and even 


from landscapes that were underneath indigenous 
agricultural systems. Other plants experienced range 
extensions as they were widely dispersed together with 
erop plant species, or otherwise 


were exchanged 
between human cultures, and 


grown outside of their native habitats. Existing records 


experimented with, 


also suggest that different species and plant commu- 


As in the 


nities had contrasting regenerative capacities follow- 
Without 


paleoecological data it might be difficult to know how 


ing severe disruption, as they do today. 
much of any extant forest was truly very old or not an 
artificial product of human intervention. and sub- 
sequent abandonment. 

Although we now understand that early peoples 
substantially altered their landscapes, we should not 
assume that every tract of lowland tropical forest was 
heavily settled and disturbed during the prehistoric 
era. With little doubt, 


resulting from the inherent fertility of soils, 


variability in land use patterns 
type of 
agriculture practiced (root/tree- vs. seed-based), and 
associated human population densities resulted in 
differential impacts on vegetation. The contrasting 
records from the Hill of Six Lakes in Brazil and the 
Rio Caquetá region of the Colombian Amazon 
discussed above are but two examples (see Piperno 


& Becker, 1996, 2000, 


others from the Brazilian Amazon). 


and Bush et al., for three 

The timing and duration of landscape modification 
would also have varied depending on whether a region 
was a cradle of erop plant evolution or otherwise 
witnessed early developments of agricultural systems 
through participation in the initial dispersals of 
domesticated plants, as occurred in Central Pacific 
Panama, for example. One might suspect, then, that 
highly seasonal tropical forests, which on present 
evidence were the geographic cradles of origin for 
many staple seed and root crops of the Neotropics, 
were more greatly affected over more protracted time 
scales than were forests in wet, aseasonal, and poorly 
soiled regions. While these are reasonable suggestions 
that accord well with available information, data from 
many regions are still few and there is much to be 
learned. 

Unlike the thinking that prevailed in human studies 
little more than 20 years ago, many anthropologists 
now decline to argue that modern land use practices 
in the tropics by small-scale, politically autonomous 
indigenous societies preserve or increase plant di- 
versity, or that there is an inherent conservation ethic 
that simpler human societies still adhere to 
Alcorn, 1996, Redford € Padoch. 1992. Alvard, 
1995, for good discussions of these issues). I doubt 


that these ideas constitute accurate assessments of 


(see 


and 


human relationships with, and impacts on. natural 


resources al any moment in time. None of the now 


substantial paleoecological records indicate they were 
true in the past. If Europeans had not arrived little 
more than 500 years ago and tragically all but ended 
indigenous life in many areas of Central and South 
America, it is very much an open question as to how 
sustainable intensifying or even maintaining existing 


agricultural practices would have been under ever- 


Volume 93, Number 2 
2006 


Piperno 
Quaternary Environmental History 


increasing human population densities, even with the 
use of relatively simple technologies (stone tools and 
fire) as instruments of forest clearing. Archaeological 
records from the tropics and elsewhere also tell stories 
of over-exploitation of marine and other resources 
during the prehistoric era (e.g., Jones et al., 2004; 
Mannino & Thomas, 2002; Porcasi et al., 2000; Stahl, 
1996; Steadman, 1995). Moreover, rigorous stuc 


modern indigenous resource use in tropical South 


— 


les of 


America make it clear that short-term self-interest and 
a desire to capture game in an energetically efficient 


manner structure hunting and other subsistence 


pursuits much more so than does a long-term interest 


in conserving resources (e.g., Alvard, 1993, 1995; 
Winterhalder & Lu, 1997). 
In order to better preserve and protect the 


remaining forests, we have to straightforwardly assess 
past and present human behavior, its proclivities, and 
its goals. At the very least, we should more carefully 
define what we mean by the word conservation when 
we describe modern indigenous resource use (is it 
a designed attempt to preserve an existing biomass 
and diversity, or an epiphenomenal by-product of 
factors such as low human population and high prey 
densities?), no matter how sustainable the resource 
usage appears to be (see Alvard, 1993, and Smith & 


Wishnie, 2000, for excellent discussions of this issue). 


Finally, there is little doubt now that the views 
originally taken by Donald Lathrap and Carl Sauer 
1970, 1973b; 1936, 1947, 1950) 


concerning the importance of the lowland Neotropical 


(Lathrap, Sauer, 
forest in the emergence and spread of domesticated 
species, the development of effective agricultural 
systems, and the flourishing of cultural life in general, 
were correct in many aspects. In fact, as Lathrap had 
predicted, the origins of New World pottery occurred 
in the heart of Amazonia 7000 years ago (Roosevelt et 
al., 1991), the first 
villages were created by people who lived in the 


and permanent, agricultural 
seasonal forests of southwest Ecuador 5000 years ago 


1975; 2004). Newer 


explorations in the TM Amazon have documen- 


(Lathrap et al., Pearsall et al., 
ted that beginning a few centuries before the time of 
Christ, 


settlements existed in the middle to lower stretches of 


very large, dense, and permanent human 


the main Amazon river channel and Upper Xingu 


region, which are now sparsely bros and under 


forest cover (Heckenberger et al., ). 2003). As 
with the forces associated with 1 today. 
these prehistoric advances probably came with 


negative consequences for the native flora and fauna. 


Profound human alteration of the tropical landscape 


— 


with substantial loss of biodiversity is hardly new, but 


we are the first societies with the wherewithal to do 


something about it. 


Literature Cited 
Alcorn, J. B. 1996. ls 7 0 0 conserved by indigenous 
peoples? Pp. 234-238 in S. K. Jain (editor). Ethnobiology 
in ee y elfare. Deep Publications, New Delhi. 
Alvard, 993. Testing the “ecologically noble savage" 
bu 1 prey choice by Piro hunters of 
Human Ecol. 21: 355-387 


— 


Amazonian Peru. 
1995 Amazonian 

789-818. 

. Pollen evidence of the 

Latin Amer. 


hunters. -~ 
Arford, M. R. & S. P. 
earliest maize agric riculture in Costa Rica. J. 
Geogr. 3: 10 
Bailey, R. C. et al (+ 
gathering in tropical 
„ 91: 59-82. 
Barnola, J. M., D. Raynaud, Y. S. Korotkevich & C. Lorius. 
1987. ee ice 160,000-vear 
108414. 
1973. Phytogeographic 


1989. Hunting and 
Is it possible? Amer 


5 authors). 
rain. forest: 


core provides record of 
atmospheric COs. Nature 329: : 

Bartlett, A. S. & E. Barghoorn. 
15 5 of the Isthmus of 
12,000 years. Pp. 203-299 
Vegetation. and 5 


Panama during the past 
] raham (editor), 


History of Northern. Latin 


o 
‘a. Elsevier, Amster dai 


ET The rise and fall of 


Nature 408: 


Bray, W. 2000. Ancient food. te thought. 
45-146 
Bush. M. B. 8 P. A. Colinvaux. 1988. > 7000 year-old 4i 
Vegetatio 76: 141-15 


A pollen record of a mia 
Ve l: 


record from the Amazon lowlands. 

———— & — . 1900. 
glacial cycle from lowland Panama. J. 

105-118. 

D. R. Piperno & P. A. Colinvaux. 1989. A 6000 year 


Nalure 340: 


y 
history of Amazonian maize cultivation. 
305. 
et al. (+ 6 authors). 1992. A 14,300-yr paleoeco- 
logical prof tile of a lowland tropical lake in Panama. Ecol. 
Mono. 62: Tas 
Bush, M. B., M C. Miller, P. E. De & P. 
Colinvaux. 2000. Two Melones of environmental change 
and human disturbance in eastern lowland Amazonia. The 
da 10: 543—55: 
.De Oliveira, P. A. Colinvaux, M. C. Miller = 
Amazonian paleoecological historie 
Palaeogeogr. Palaeoc E 


Oliveira 


hill, three watershe E 
347-350 

Up. 1 2001. Pre-Columbian land use 
Rica: A 3000-y« cord of forest 


e 
Palaeoecol. 2 
Clement, R. M. QS 8 


in Costa 


m 
arance, agriculture, and fores from 5 Zoncho. The 
Holocene 11: 419-426. 
Colinvaux, P. A., P. E. De Oliveira, P. Moreno, M. 
Miller & M. p. Bush. 1996. A long ae record 19 5 


lowland Amazonia: Forest and cooling in glacial time 
Science 274: 85— 
Amazonian 


= > 


„ Bush. 2000. 
Neotropic al plant communities on glacial time-scales: The 
failure of the ar 7955 
Rev. 19 va 16 

Cooke, R. 


aria’ 


and 
and refuge hypotheses. Quatern. Sci. 


Ranere. The “Proyecto Santa 


A. i 1984. 
E N S jlinary analysis of prehistorie ada] vta- 
tions to a tropical watershed in Panama. Pp. 3-30 in F. 
Lange (editor), Recent Developments in Isthmian Archae- 
ology. British Archaeological Reports International Series 
212, Oxford 
Cowling, S. A. € M. T. Sykes. 1999, 
significance of low D € O, for plant-climate 
-242 


Physiological 


interactions. Quatern. Res. 


294 


Annals of the 
Missouri Botanical Garden 


Curtis, J. H., M. Brenner & D. 
change in the Lake 


Hodell. 
Valencia Basin, Venezuela, 
The Holocene 9: 609-619, 

(+ 5 authors). 
a tropical karst environment. 


12600 vr 
BP to Present. 
1979. Mayan urbanism: 
Science 200: 


S. et al. 


Diamond, J. 2002. 
plant g animal domesticatie lature « 

Dillehay, T. D. 1997. Me m 
ce ber in Chile, Vol. 
and Interpretation. 5 Institution Press; Wash- 
ington, 


100—707. 
Vende A bf Pleistocene 


Archaeological Context 


2000. The 


e hae 


New 


Americas. A 

Basic Books, New 

1. R. 2002 The Smokey Bear: The 
ecodisaster myt th and forest management prac 555 es in 
prehistoric s America. World Archaeol. 33: : 

Doebley, J. 1990. Molecular evidence 
maize. Econ. 2 14(3 supplement): 6-27. 

Dull, R. A. 2004. An 8000 
and human disturbance 


Settlement: of the 
rk 


15 S, death of 


and the 


-year record of vegetation, climate. 

from the Sierra de Apaneca. El 

Salvador. Quatern. Res. 61: 159-167. 

Fischer, A. 1960. Latitudinal variations in organic diversity. 
Evolution 14: 64-81. 

Grimm, E. C., S. Le 
2001. 


Americas. 


ozano-Garcia,. 


H. Behling & V. Markgraf. 


Holocene vegetation and climate variability in the 
325-370 in V. Markgraf (editor), In- 


Academic Press. San 


Pp. 325- 
terhemispherie Climate Linkages. 
Diego. 


Guilderson, T. m. V C. Fairbanks & J. L. Rubenston. 1994. 
Tropical temperature variations. since 20,000 years ago: 
Mode ling | inte M un ric climate change. Science 263: 
60300 

Haffer, | 000. Speciation in Amazonian forest birds. 


Science 165: 131-137 
Hart, T. & J. Hart. 1986 The 


gatherer subsistence in Mi an rain lore ests: The Mbuti of 


ecological basis of hunter- 


Eastern Zaire. Human : col. 14: 29-55. 
Heckenberger, M. J., J. B. Peterson & E. G. Neves. 1999, 


Two archae- 
Antiquity 10: 


lage size and permanence in Amazonia: 


ological examples from Brazil. Latin Amer. 

353-376. 

el al. 2003. 
forest or la parkland? Science 301: 

Hodell, . et al. (+ 1991. 

change over the 
Nature 352: 790-792. 

Horn. S. P. & R. I. Sanford Jr. 1992. 
Costa Rica. 7 e 24: 354-3061. 
Huang, Y. et al. 2001. 

dominant 1 1 on glacial- a al variations in Cs 
1647-1651. 
ehistoric Forest 


— Amazonia 1492: Pristine 
1710-1714. 

Reconstruction of 
past 10,500 


© authors). 


+ © authors). 
years. 


Holocene fires in 


> authors). Climate change as the 


and Cy plant abundance. Science 
Jones, J. G. 199 
Modific ation and Maya Cultivation in Belize. Unpublished «d 


Pollen. Evidence ears 


Ph.D. Dissertation. De des nt of Anthropology. Tex 
A&M E niversity, College Station. 
994. Pollen e ide nce for early se P "me ent and 
agric d in northern. Belize. Palynology 18: 205-21 I. 
Jones, T. 1 . Hilde ‘brandt. D. J. Kennett ut un F. 


Porcasi. 2004. 
northeastern de review of some new e 
alif. Great Basin Anthropol. 24: 69-80. 
Kenneth, D. & B. W tenidas (editors). 2006. 
Ecology and the Transition to Agriculture. Univ. 
Press, Be rke ley 


Knapp. S. & J. Mallet. 2003. Refuting refugia? Science 300: 
11-72 


Prehistorie marine mammal overkill in the 


vidence. 


Behavioral 
California 


1999. Climate 


Evolution, 9 iud ae future of 


ee of 


1970. The Upper Amazon. Praeger Publish- 


— D. W. 
. New Yor 
1973a. Gifts of the « 


2 Some thoughts on the 
subsistence basis of Chavin. Pp. 91-105 in D. W. I 


cayman: 
athrap 
& J. Douglas (editors). Variation in inia oss Essays 
McGregor. Archaeological 


n Honor of John Ilinois 


Survey, Urbana. 
973b. The 


distance trade 


antiquity and importance of long 
relationships in the moist mene of pre- 


E paa South 9 rica. Wor 


Collie Ch 1 1975. Ancient. Ecuador: 
"mus Clay, pe 1 3000-300 B.C. Field 
Museum of Natural History, Chicago 
Ledru. M.-P.. J. dd cd V. Sifeddine & K. Suguio. 1998. 
Absence al La rlacial secre records in. lowland 
tropic j^ forests. 19 Res. 40: 233-237 
Leyden, B. 1984. Guate 0 forest synthesis altar Pleisto- 
cm Proc. Natl. Acad. Sei. U.S.A. 81: 4856-4859. 
1985. ale 1 aridity and Holocene 
moisture fluctuations in the Lake Valencia Basin, 
Venezuela. Ecology 66: 1279-1295 


1987. Man and climate in the lowlands. 
bsp" Res. 28: 407—414. 
V. Hodell & J. H 


dim in the 
„ p. K 


Maya 
— 1993. Late 
lowlands. 
McKenzie 


Continental 


. Brenner, . Curtis. 


Central American 
a 178 in P. K. Swart, K 
4 js (editors). 


Ple EY ene 

C. Lohmann. J. 

Climate Change in 
wit. 


ic Records. Geophysical Monogr: aph 78. American 


— 
eS 


Geophysical Union, Washington, 
1975. Late Ande rary e año change ii 
Mrica. de Ke v. Ecol. : 249-280. 
Mangelsdorf, s 1953. de "view a 
Ty Carl O. Sauer. 


Livingstone, D. A. 


“Agricultural Origins 
and Dispe E Amer. Antiquity 19: 
86-90. 

. R. S. MacNeish & G. R. Willey. 
agriculture in Middle America. Pp. 427-445 in R. C. West 
(editor). Handbook of Middle American e poe 
Environment and Cultures. Vol. Univ. Texas 


1964. ges of 


Early 


Press. Austin. 


2 


Mannino, D. Thomas. 2002. Depletion of 


a resource? The impact of prehistoric human foraging on 
significance. for 


World Ar- 


I 
intertidal ll communities and its 


human se p me ient mobility, and dispersal. 
chaeol. 33: ATA, 
: s. 9 Paleoclimatology and a tropical pollen 


310-323 in Report of the Vith International 


Congress on Quaternary (Warsaw 1961), Vol HE: Palaeo- 
Climatological Section. Lodz 
Matsuoka, Y. et al (+ 5 authors). 2002. A single 


domestication for maize Sour : muliiloc "us microsatel- 
lite genotyping. Proc. Natl. Acad. Sci. U.S.A. 99: 
60€ 0— 0084. 

Mayle, F. E & D. J. 2004. Late 


changes in Amazonian ecosystems and their implications 


Beerling. Quaternary 


ie cycling. Palaeoclimatol. 
Palaeoecol. SL IS25 
& " 1994. Maintenance 


diversity in riparian forests of tropic 


for global « Palacogeogr. 


Meave. J. ps men. of rain forest 


cal savannas: Implica- 
lions for species conservation during P leisten drought. 
TE 533. 


1954. Environmental 


. Biogcogr. 21: 
Me weers, B. 


development ot culture. Amer. 
19 


limitations on the 
)1 -824. 


— 
oo! 


\nthropol. 5€ 

. 1971. Amazonia: Man e Culture 
Paradise. Aldine, Chicago 

Meltzer, D. 1997. Monte 


peopling of the Americas. 


in a Counterfeil 


Verde and the 


Pleistocene 
Science 276: 154-155. 


Volume 93, Number 2 
2006 


Piperno 
Quaternary Environmental History 


M. Adovasio & Dillehay. 1994. On 


a Ple stocene human € al t Pe dra Furada, Brazil. 
1. 


Antiquity 68: 605 5-1 
Mora, 5. C. ` Herrera, I. Cavelier & C. Rodríguez. 1991. 
Cultivars, Anthropic Soils s Stability. Univ. Pittsburgh 


Latin Amer. Archaeol. Rep. No. 2. Univ. Pittsburgh. Dept. 
Noth.  Piteburh 
Moy, C. M., . Seltzer, D. T. Rodbell & D. M. Anderson. 


2002. Variability of El T ign Oscillation activity 

escales during the Holocene epoch. 
Na 

Olsen, K. M. & B. 4. 5 Schaal. 1999, Evidence on the origin of 
cassava: M Ei of Manihot esculenta. Proc. Natl. 
Acad. Sei. U.S.A. 96: 5586-5591 

Pearsall. D. M. 197 8. Phytolith analysis of archaeological 
soils: E rs for maize cultivation in formative Ecuador. 

> 199: 177-17 

K. Ch: 1 Eze I] & A. Chandler-Ezell. 2003. 


Ide gti ing maize in Neotr s als sediments and soils using 


Science 


Archaeol. Sei. 30: 611-6: 
J. X. 2x e 2004. Maize in ancient 


cob phytoliths. J. 
———— & 
: c s Results of residue 1 70 sis of stone tools from the 


teal Alto site. J. Archaeol. . 31: 423-442. 
11 9 W. R. & L. P. Sell im. 2004 The climate of the 
earth at the Last Glac 


ial Maximum: Statistical equilibrium 
state and a mode of internal variability. Quatern. Sci. Rev. 
335-357. 


Historical 
ry 
forest plants show patterns of both Tertiary and Quaternary 


Pennington, R. T. et 2004. 
limate change and speciation: Neotropical seasonally ¢ 


authors). 


¡om 


diversification. Philos. Trans. Ser. B. 359: 515-536. 
| 985a. Phytolith analysis and tropical paleo- 
of sili- 


Produc ‘Lion and | signific ance 


y laxonomic 
ceous forms in New World plant domesticates and wild 
les . Palvnol. 45: 185-228 
taphonomy ne d in 
archaeological sediments from Panama. J. Archaeol. 
12: 247-267. 
1988. Phytolith Analysis: An 
Geological Perspective. Ac ademic Press, San Diego 
2006a. A behavioral ecological approach to the 
origins of plant 


\rchaeological and 


and domestication in the 
New World. /n D. Kennett 
Behavioral Ecology and the 


cultivation 
seasonal tropical forests of the 
& B. Winterhalder (editors), 
Univ. California Press. 

A Comprehensive Guide for 
AltaMira Press, 


Transition to Agriculture. 
2006b. Phytoliths: 
Archaeologists and Paleoecologists. 


an- 
1 
. Becker. 1996. Vegetational history 
Amazon Basin derived from phytolith po 
4 


Res. 


ham. 
of a site in 
the jue 
Quatern. 


charcoal. records from natural. soils. 


202-200 


J. Jones. 2003. Paleoecological and archaeological 


implications of a late Pleistocene/early | Holocene record of 
vegelalion 
Panama. Quatern, Res. 59: 79-01. 
& D. M. Pearsall. 1998. 
in the Lowland Neotropics. 
& K. E. Stothert. 2003. Phytolith evidence for early 
Holocene ee domestication in Southwest Ecuador. 
"ci 2 5 
e us d & P. 
me iiis 2 is 1 
Res. 33: 108-116 
———— & — . 1991a. 
ectives on human adaptation in Central Panama. I. The 
10 6. 


The Origins of Agriculture 


Academic Press, San Diego. 


cience 2 
M. 1990. Paleoenviron- 


in Late-Glacial 


. Colinvaux. 
Panama. 
Qualern. 
Paleoecological per- 


p i sistocene. Geoarchaeology 6: 210-22 


and climate from the Pac ific coastal plain of 


——— & 1991b. Paleoecological per- 
spec tives on oos adaptation in Central Panama. H. The 
Geoarchaeology 6: 227—250. 

— (— — & . 1992, Patterns of articulation of 
c lie 1 the plant world i in prehistoric Panama: 11,500 
BP-3000 BP. Pp. 109-127 in O. R. Ortiz-Troncoso & T. 
Van der Hammen (e dins. Archaeology and Environment 

Amster 


Holocene. 


in Latin America. Universiteit van Amsterdam, 


dam. 


Stare 


Panama- 


. Ranere, I. Holst & P. Hansell. uM 

veal early root crop horticulture in the 
894—897 

— & K. E. Siothe srt. 2000b. 
Phy toliths i in Ciia and other Neotropical Cucurbita- 

early 1 3 sites É 


grains 
nian 1 'al forest. Nature 407: 
lolst, C. Andres 
ceae and their occurrence in 
from the lowland American tropics. J. Archaeol. 
193-208. 
al. (+ 


History 


6 authors). 2004. Environmental and 
in the Central age atershed, 


Paper a al 


Agricultural 


+ 


Mexico: Results o 
the Annual Meeting of the 
Archaeology, Montre al 

Pohl. M. D. et al. (+ 10 authors). 

the May a lowlands. Latin 

0) 


Pre ‘liminary Research. 
Society for American 

Canada. 

1996. Early agriculture in 


Amer. Antiquity 7: 355-372. 


Pope, (+ 6 authors). 2001. Origin and 
1 selling of ancient agriculture in 
lowlands of 5 rica, Science 292: 1370-137: 

Porcasi, Jones & L. M. Raab. 2000. Trans- 


Holoc 'ene marine 1 „ on San Clemente 


Island, California: A tragedy of the commons revisited. J. 
Anthropol. 19: 2 00-220 
Prance, G. T. 1987. Pp. 2t 945i in T. C. Whitmore 
G 


T. Prance (editors), Biogeography and Duman 
I 


Archaeol. 
Vegetation. 


& G. 
History in Tropic val America, ds 115 rendon Press, Oxford. 
Ranere. A. J. & R. G. Cooke. 
Holocene occupation of Centre Al 55 rican tropic i orests. 
Pp. 210-248 in J. Under the 


The Archacology of pe ‘al Rain Forests. Rutgers Univ. 


vate glacial and early 


Mercader (editor), Canopy: 
Press, New Brunsy 

Redford, K. H. m 1992. 
Neotropical F orests. c 'olumbia Univ. Press. New York. 

F. D. Pennington & P. 


Padoch. Conservation. of 


r 


Richardson, J. E., R. T. Pennington, 
M. H eo 2001. Rapid diversification of a species- 
rich genus of Neotropical rain forest trees. Science 293: 
2242-2245. 
Roosevelt, A. 
Maranca & | 
from a prehistoric shell midden in the Brazilian Amazon. 
Science 254: 1621-1624. 
Sage, R. F. 1995. Was low atmospheric CO» during the 
Pleistocene a limiting E tor for the origin of agriculture? 
Biol. 33-106. 
Sandweiss, D. H.. J. E Ric hardson III. 
Rollins & K. A. V ) 
i from Peru for a 


Housley, M. Da Silveira, S. 
Eighth millennium pottery 


(., R 
i 


A. 
Johnson. 1991. 


Global Change 
E. J. Reitz, H. B. 


laasch. 19906. Geoarchaeologic a 


evidence 5000 years B.P. onset of 
Niño. Science 273: 1531-1533 

Sanjur, O., D. R. Piperno, T. C. Andres & L. Wessell-Beaver. 
2002 


nships among domesticated and 
wild species of Cuci iita (Cucurbilaceae) inferred. from 
a d enun gene: Implications for crop Eus 2 
: roc. Natl. Acad. Sci. U.S 540. 

American agricultural 
and agriculture. Pp. 279-298 in 


Sauer, C. 0. 193 origins: 
consideration. of nature 
R. H. Lowie du ditor), Essays in 


. Unn 


Ear re Jations of man to plants. Geogr. 


e in Honor of 
California Pies Berkeley. 
Rev. 


1947. 


(New York) 37: 1-25 


Annals of the 
Missouri Botanical Garden 


1950. 
America. Pp. 4 


Cultivated E of South and Central 
in J. Steward (editor), Handbook 
of South American P Bus au of American Ethnology 
Bulletin. 143, Vol. U Printing Office 
Washington, 
——— 1952. 


37-543 


~ 


yovernment 


— 


cde ultural Origins and. Dispersals. Amer- 
ican Gi 'ographical Society, New York. 
— . 195 


58. Man in the i» ogy of tropical America. 


: rocee Pr of the Ninth Pacific Science Congress, 
3angkok, 1 20: 104—110. 
stola i A K H. L. Sanders. 1969. On the e e 


0 5 predictability to species div 
dap qe n Symposium in Biology 22: 82-95. 
V 


Smith, E. A. & 


Wishnie. 2000. Conservation and 
son nce in small-scale societies. Annual Rev. Anthro- 
pol. 2 is s 

Spons 5 1989. Farming and foraging: A necessary 


Pp. 37-45 in S. Kent 

l'armers as Hunte rs: The Implications of Sedent- 
| 

bridge Univ. Pre 


c ibi a mentarity s Amazonia? 
(editor), 
ism. Cam 


Stahl, P. W. 


Cambridge, Massachusetts 


1996. zie ene pne silly: An archaeological 


perspective from the Americas. Annual Rev. \nthropol. 
25: 105-20. 

Steadman, D. 1995. Prehistorie extinctions of Pacific Island 
birds: Biodiversity meets zooarchae ology. Science 267: 


1123-1131. 


^ 


Van der Hamı . & 


Pleistocene 


Gonzalez. 1960. Upper 
te and vegetation of the 
South 


and ‘Holocene clim 
s de as (Colombia, 
ol. e 126-315. 
& H. 1 8 nstra. 2000. Neogene and Quaternary 
Te of vegetation, 
Amazonia. Quatern. Sci. . 19: 725-142. 
Wardle, D. A. L. R. Walker K R. D. Bardgett. 2004. 


Ecosystem properties and forest decline in contrasting 


America). Leidse 


Mee and plant diversity in 
( 


pel 


Webb. R. S. R. D. Rind. S. J. Lehman. R. J. Healy & D. 
Sigman. 1997, s 'e E ocean heat transport on the 
climate of the Last Glacial Maximum. Nature 385: 

o. 

Westengen, Huaman & M. Heun. 2005. Genetic 

diversity ond Ee pattern in early South American 
cotton domestication. Theor. Appl. Genet. 110: 392-402, 
Whitmore, T. C. & G. T. (editors). 1987. 


5 1 Tropical. America. 


Prance 

nd Quaternary History 
C adond P ress, Oxford. 

Wilf, P. et al. (+ 5 authors). 2003. High plant diversity 
m 'ene | America: Evidence from Patagonia. Science 
300: 

Winterha j m & F. Lu. 1997 forager-resource 
population ec voles model and implic ations for indigenous 
conservation. Conservation Biol. JJ: 1354-1364 


= 


THE PALEOBOTANICAL RECORD Henry Hooghiemstra,? Vincent M. Wijninga.? 
OF COLOMBIA: IMPLICATIONS and Antoine M. Cleef” 

FOR BIOGEOGRAPH Y 

AND BIODIVERSITY" 


ABSTRACT 


Plant microfossil and macrofossil associations pis d from six dated sections from the area of the basin of Bogotá (2550 m 


Eastern. Cordillera, Colombia) show the evolutior the late Neogene Andean montane forest, triggered by the Andean 
orogeny. Progressive adaptation of warm s 'al taxa to i montane c renditions, evolution of new ne 8 ‘al montane ta 
and immigration of temperate Laurasian, Holarctic, and Austral-Antarctic elements gave shape to the present-day montane 
forest. Vegetational characteristics inferred from fossil s associations reveal the altitude at 985 time of deposition. Neoger 
forests are floristically compared with contemporary forests at comparable altitudinal intervals in the surroundings of la 
Bogotá basin: however, the absence of taxa that had not yet arrived, or evolved, is most salient ie shows that non-analogue 
plant communities are common 

he main 11 55 s of montane 1916 st development are: (1) pre-uplift phase of the late Miocene with abundant lowland taxa with 


tropic di or t neotropic 'al affi pies Lire d &unth, An nanoa Aubl., Ceiba Miller, and representatives of Humiriaceae); montane 
forest rich in Podocar tentially including Nageia Gaertn., Podocarpus L Hér., Prumnopitys Phil., and Retrophyllum C. 


[em 


. Page) cov eicd ae n We uplifted areas in 1 85 region; (2) toward the early Pliocene the area was uplifted to ca. 1000 n 
the relative proportion of temperate taxa of North American and southern South American stock increased and occupied the 
slopes of the low mountains; and (3) by the ups Pliocene uplift had proceeded to ca. 2000-2200 m and tropical lowland 
taxa, which are now restricted to altitudes below 1000 m, are no longer recorded in the fossil plant associations: thei increase in 
the number of newly appearing montane taxa 17 0 L., Turpinia Vent., Gunnera L., Bocconia L., Gaiadendron G. Don f., 
is C. Martius, and Monnina Ruiz & Pav.) suggests a significant increase of diversity 

itil the late Pliocene there is little to no evidence for the existence of the páramo; taxa such as Xyris L.. Hypericum L., 


Carex L., Montia L., and Ranunculus L. might have formed swamp or bog vegetation only. It is be a that these taxa 
colonized mountaintops with half-open vegetation; these areas extended when the mountains reached above upper forest line 
elevati tions. 

The distribution areas of the endemic genera of the Espeletiinae largely coincide with the youngest parts (< 5 Ma) of the 
northern Andes. Changing climatic conditions forced individual plant species to migrate vertically. Composition of plant 
communities Pe d contir me and vegetation belts frequently were altitudinally squeezed or “offered possibilities for 
expansi This long process stimulated speciation and provoked sequential non-analogue vegetation types. Thus, the 
bi odiversity hotspot al Inc ie Andes has a dynamic 71 

y words: Andean uplift, biome evolution, C olombia. Espe dedii, montane forest, Neogene, páramo, phytogeography. 


RESUMEN 
Los microfósiles de plantas y asociaciones ds macrofósiles obtenidos de seis secciones fechadas del área de la cuenca de 
bia) J 


Bogotá (2550 m, Cordillera Oriental, ( stran la evolución del bosque andino del Neógeno tardío, accionado por 
la orogenia andina. La adaptación progresiva i: taxones tropicales calidos hacia las condiciones frías montanas, la evolución 


de nuevos taxones neotropicales montanos y la migración de elementos templados laurásicos, holárticos y austral-antárticos 
dieron forma al bosque montano actual. Las características de la vegetación deducidas de las asociaciones de plantas fósiles 


revelan la altitud al momento de los depósitos. Los bosques neógenos se comparan floristicamente con bosques 


cont aa os a intervalos altitudinales comparables en los 5 dedores de la cuenca de Bogotá; sin embargo. la ausencia de 


= 


taxones que todavía no habían llegado, o desarrollado, es sobresaliente y demuestra que las Nö de plantas no 
análogas son comune 

as fases princ ud s del desarrollo del bosque montano son: (1) fase pre-levantamiento del Mioceno tardío con Wer m a 
de taxones de las tierras bajas con afinidades tropicales o ne Bio. dies (Mauritia Kunth, Amanoa Aubl., Ceiba Miller y 
representantes de Humiriaceae); un bosque montano rico en Podocarpaceae (potencialmente incluyendo Nageia etd 

] 3 itys Phil. y Retrophyllum C. N. Page) cubría otras áreas previamente levantadas en la región; (2) 

hacia el Plioceno temprano el área se elevó ca. 1000 m: la proporción relativa de taxones templados de origen norteamericano 
suramericano meridional aumentó y ocupó las vertientes de las montañas bajas: y (3) hacia el Pine 'eno medio, el 


! We thank the Netherlands Foundation for the Advance: 'ement of Tropical Research (WOTRO, grant W 75-314) and the 
Hugo de Vries Foundation for finance a support to VMW. The Instituto de Ciencias Naturales (ICN) of the Universidad 
Nacion al de Colombia, Inderena (C. Barbosa), and Ingeominas ce 4) are thanked for their cooperation. The following 
persons assisted in the field: L. Beccerra, J. Numpaque, W. Diaz, and C. Sarmiento. H.H. thanks the organizers of the St. Louis 
EE Alan Graham in particular, for the opportunity to participate; this was a direct stimulus to prepare this ag We 
thank Beryl Sees Michael Wille, and Orlando Rangel for constructive comments on an earlier draft of this pape 

TIMOR for Biodiversity and Ecosystem Dynamics (IBED), Palaeoec be and panos scape j cology: Faculty of Se ience, 


University of Amsterdam, Kruislaan 318, 1098 SM ee The Netherlands. hooghiemstra@science.uva. nl. 


ANN. Missouni Bor. GARD. 93: 297-324. PUBLISHED ON 23 Aucusr 2006. 


298 


Annals of the 


Missouri Botanical Garden 


levantamiento había d hasta ca. 2000-2200 m y taxones ps vales de las tierras bajas, i actualmente se restringen 


a las altitudes ya no se registran más 


ceno tardío, hay 


2 

— 

— 

m jah 
= = 
Š 5 
~. U 
z 

3 

C3 


Cree que c stos taxones colon nn 


es montanos nuevos (Myrica L., Turpinia Vent., Ce L., Bocconia L., 


oco a ninguna evidencia de 
ex L., M ontia L. y Ranunculus L. pudieron haber v formado sólo vegetación de 
zaron las cimas de las montañas c 


gistros fósiles de asociaciones de plantas: el aumento del 


Gaiade ndron G. Don f., Daphnopsis 


giere un aumento significativo de la diversidad. 


xistencia del páramo; taxones como por ejemplo Xyris 


p E 


los pantanales o las ti ibas. $ 


1 vegelac IÓN S se mi- abie rta; estas áre as se exte andie ron c uando 


las montanas alcanzaron e i. vaciones sobre el límite superior del bosque. 


as áreas de distribución 
(< 5 Ma) de los Andes del norte. 
verli 9 1 N . La 


Condicion 
composición de | 
aa altitudinalmente o se ofrecieron posibilidades 
tipos de vegetación secuenciales no análogos. 


árida. 


es climáticas eee intes 


de expansión. Este 


de los géneros endémicos de Espeletiinae coinciden en gran parte con las regiones más jóvenes 


forzaron a especies individuales de plantas a migrar 


las comunidades de plan las cambia continuamente y los cinturones de ve ge tación fue eron 


argo proceso e 'slimuló espec lac ion y provoc 0 


Así el punto caliente de biodi ve rsidad de los Andes del norte tiene una historia 


Increasing knowledge of the present-day flora and 
vegetation in the northern Andes promoted a simulta- 
neous interest in its history. Paleoecological, primarily 
palynological, studies of Neogene sediments have 
provided a general understanding of the composition of 
the flora in the past and its changes over time. Van der 
Hammen et al. (1973) demonstrated the influence of 
tectonics and climatic change on the development of the 
Andean forest. Since then, understanding of how and 
when the present-day Andean forest acquired its floristic 
composition has increased by integrating paleodata 
with uoloniiation about the present-day Andean forest 
(Raven Axelrod, 1974; van der Hammen. 1974; 
Simpson, 1975; Gentry, 1982; Cleef et al., 1983; van der 
Hammen & Cleef, 1986; van der Hammen, 1989; 
Simpson & Todzia, 1990; Hooghiemstra & Cleef, 1995; 
2004 


This paper focuses on results obtained from plant 


Hooghiemstra & van der Hammen. 


macrofossils and fossil pollen from sediments located 
in the basin and surrounding areas of Bogotá, Eastern 


Cordillera of Colombia. Sites are sections from 


outcrops or deep bore holes and are all located 


ca. 2550 m (Fig. 1). Fossil plant assemblages from the 


late Miocene (sites Salto de Tequenedama I/II: 
Wijninga, 19964), early Pliocene (site Río Frío: 
Wijninga. 1996c), middle Pliocene (site Subacho- 


que-39; Wijninga & Kuhry, 1990; site Facatativá-13; 
1990b). 
Wijninga & Kuhry, 1993) present snapshots of the 


Wijninga. and late Pliocene (site Guasca; 
long-term paleoecological development of the north- 
| | 8 I 
em Andes (Table Paleoecological data are com- 
pared with present-day northern Andean flora in core 
Fuquene-2 (van Geel & van der Hammen, 1973) to 
shed light on speciation, migration, plant diversity, 
and phytogeographic composition (Cleef, 1979: Cua- 
1979; & Cleef, 1986; 


Luteyn, 1999). In addition, a synthesis of these data is 


trecasas, van der Hammen 


presented to support our understanding of how 


evolution of the Neogene sedimentary environment 
and vegetation in northwestern Amazonia relates to 


the uplift of the northern Andes. 


The objective of the present paper is to compare the 
paleobotanical evidence from the northern Andes with 
paleobotanical studies from Amazonian lowland sedi- 
ments in order to reconstruct the evolution of the 
northern Andean biota. The phytogeographic compo- 
sition of the fossil plant assemblages shows snapshots 
of a long process of gradual change. Uplift created 
large areas of habitat that had previously not existed 
regionally, offering new establishment possibilities 5 iu 
floral elements from different 


source areas. ‘he 


biostratigraphic position of the recorded plant taxa 
is evaluated. 

ALTITUDINAL ZONATION OF THE PRESENT-DAY NORTH 
ANDEAN VEGETATION AS a Basis FOR PALEO- 
RECONSTRUCTIONS 


The altitudinal zonation of the neotropical montane 
vegetation is primarily due to the effect of decreasing 
temperature with increasing elevation. Many studies 
have been dedicated to the composition and structure 
of the neotropical montane forest (e.g., Cuatrecasas, 
1934; 


González, 


— 


963; van der 


1963: Vareschi, 


Grubb et al., Hammen & 


1980: Cleef et al., 


Cleef & Hooghiemstra, 1984; Huber, 1980: Ulloa 
Ulloa & Jørgensen, 1993; Rangel, 1995; Rangel et al.. 
2003; Cleef et al., 2003). Cuatrecasas (1958) 


published an altitudinal zonation for the Colombian 


vegetation, later depicted by van der Hammen (1974) 


(Fig. 2). Their scheme of altitudinally arranged 


vegelation belts is used in this study, but new 
information from Wille et al. (2001) and Torres 


2006) has been incorporated. This altitudinal zona- 


tion corresponds principally to the global vegetation 
1974, 


1977) and is the basis for interpreting paleorecords in 


zonation in high tropical mountains (Grubb, 


terms of uplift, climate change. and changing floristic 
compositions. 

The following account of the present-day altitudinal 
vegetation distribution in the Eastern Cordillera of 


Colombia is needed to support the interpretation of 


Volume 93, Number 2 
2006 


Hooghiemstra et al. 299 


Paleobotanical RECON of Colombia 


Laguna de Fuquene «> 


5030’ 


Sorg 


2 


Rio Bo 


22 
onp’ : 
2-00 E Funza Il & 


8 —— 


xv 


4 3 
7 
study area Q 


] 


AX 


E 3500-4500 m. alt 
3000-3500 m. alt 


5 e | 
15 Mies JUL 


Salto de Tequendama I & lo 5 1 2x 
n. moo A NGA 


PORUM O 3000 
[73 high 9255 a Bogotá 
2550 m. alt 


j d ES lake 


20km 


74030’ 
Figure 1. 


s outcrops and deep bore holes) with 
(1984), reprinted with permission from E 


fossil assemblages in paleoaltitudes. Here we use the ` 


family level because a significant proportion of 


pollen- and macrofossil taxa were identified to that 
level. Moreover, given our objectives, Keizer's (2000) 


vegetation analysis of the Colombian Eastern Cordil- 


lera indicates that analysis at the family level is 
sufficient. During the Miocene and Pliocene, in- 


sediments of an age from 
. Schweizerbart’sche Verlagsbuchhane 


N 
74900' 


Map of the high plain of Bogotá, Colombia, and surroundings. Black dots indicate locations of sites (sections 


late Miocene to late Pliocene. Modified after Hooghiemstra 


llung. 
dividual taxa may have had slightly different 


altitudinal ranges compared to today (Torres, 2006: 
Torres & 


number of taxa used to infer paleoaltitudes and the 


Hooghiemstra, in prep.) but the large 


relatively large envelope of uncertainty associated 
with the paleoaltitude estimates compensate for such 


potential taxa-specific errors. 


Table J. Site-specific data of sections and cores in and around the basin of Bogotá, Colombia, used in this study (see Fig. 1). 
Salto de Salto de 
Tequendama | Tequendama Il Río Frío 17 Subachoque 39 Facatativá 13 Guasca 103 Fuquene 2 Funza | 
Site (section) (section) (section) (section) (section) (section) (core) (core) Funza 2 (core) 
Coordinates 4^50'N, SON, 4'N, 5 06'N. 457 N, 513'N, 5°27'N, 4°50'N, 4°59'N, 
74°20'W 714 20'W 14^29'W 14729'W 14^33'W 74°15'W 73 46'W 74°12'W 74°22'W 
Modern altitude (m) 2450 2475 3165 2820 2750 50 2580 2550 2550 
Lithostratigraphy Lower Tilata Lower Tilata Lower Tilat Lower Tilata Lower Tilata Upper Tilata — E 
Formation/ Formation/ DURUM Formation/ Formation/ Formation/ 
Tequendama Tequendama Tibagota l'ibagota Tibagota uasca 
Member mber Member Member Member Member 
Depositional stream deposit. floodplain. stream deposit, floodplain peat. swamp, swamp, lake lake swamp, lake, river sw lake, 
environment river channel swamp deposit lake deposits ver deposits 
Sample interval (m) 34.70-34.7 1:20—1:25 0.60— " 65 0.35-0.40 1.90-1.95 26.75-26.80 12-0 3574 m 


Sample material peaty silt peaty clay silt peaty silt peat peat (peaty) clay. clay, sandy clay, 
peat peat pe 
Main macroremains wood. cuticles cuticles seeds wood wood, cuticles cuticles — — — 
Age (Ma) 17-11 1— us mundi + 54 7 * 04 mes! ca. 320 ka ca. 1.6 Ma-30 ka .2 Ma- 
30 ka 
No. of taxa included in 03 03 18 52 59 32 — — — 
percentage 
calculations 
Biozone I I II II IH VII V, VI, VII IV, V. VI. VII 
Inferred paleoaltitude 700 (+ 500) < 700 (+ 500) ca. 1000 1000 (+ 500) 2000 (+ 500) 2200 (+ 500) — — — 
(m) (maximally 
[s 500) 
References Wijninga, 1996a: Wijninga 1996a; Wijninga. Wijninga & Wijninga, Wijninga van Geel & Hooghiemstra, 1984, Andriessen et 
1996d 1996d 1996c: Kuhry, 1990; 1996b: & Kuhry, 1993; van der 1 ssen al.. 1993 
1996d Wijninga. 1996d Wijninga. Hammen, et al., 1993; Torres et al., 
1996d 1996d 1973 Torres, pico 2005: Torres, 
2006 
Relevance l ] 1 1 1 l 2 3 


35. n 


clay, sandy clay, 


l. Paleoflora from which paleoaltitude has been inferred. 


2. Modern flora used for comparison with paleofloras. 


3. Pollen spectra reflect Quate 


ernary setting after the upheaval of the Eastern Cordillera was completed 
4. Pollen spectra reflect late Pliocene and Quaternary setting after the upheaval of the Eastern C ondilke 


era was completed. 


00€ 


aul Jo sjeuuy 


uopJer) [eoiuejog unossi|A 


Volume 93, Number 2 Hooghiemstra et al. 301 
2006 Paleobotanical Record of Colombia 
Present 20,000 - 14,000 yr BP 
(Interglacial conditions) (Glacial conditions 
9990 Perennial snow 
4599-7] Superparamo l 
40007 ] Grassparamo / 
3500 - paramo — A 
30007 Andean forest 
2500 
— 20007 
E Subandean forest 
— 15004 
o Z 
Tv 100074 Z 
2 500 d A XAT, 17 ie 
— avannas > Savannas 
= est m — 
* High plain of Bogotá Z Xerophytic vegetation —- Upper forest line 
eure ‘matic cross section through the Eastern Cordillera at the latitude of Bogotá. The altitudinal distribution of 


"me vegetation be lts for the ps sent and the Lasl Glacia 


occurs in the interandean valleys. Its presence is shown be 


Maximum are shown. Xerophyt 


Cause 


c vegelalion is mainly azonal and 


xerophytic elements so present i associations of 


e are alse 
paleofloras of this study. Modifie J after van der setae in (1974). re printe d with permission from Blac iie iL Publishing. 


The present-day tropical lowland belt extends from 
sea level to 1000-1200 m, and annual temperatures 
I Cat 


The vegetation is princi- 


average 30 C at the lower altitudinal limit to 2 


the upper altitudinal limit. 


pally characterized by arboreal representatives of the 


Anacardiaceae, Annonaceae, Apocynaceae, Areca- 


ceae, Bombacaceae, Burseraceae, Caryocaraceae, 


Cecropiaceae, Clusiaceae, Euphorbiaceae, Humiria- 


Lauraceae, 


ceae, Lecythidaceae, Leguminosae, Mela- 
stomataceae, Meliaceae, Moraceae, Myristicaceae, 
Rubiaceae, Sapotaceae, Sterculiaceae, Tiliaceae, and 


Jochysiaceae. Savanna vegetation is dominated bi 
Vochy 5 getal | ted by 
Poaceae and characterized by woody genera including 
The 


most species-rich families include Fabaceae, Myrta- 


Dilleniaceae, Malpighiaceae, and Rubiaceae. 


ceae, Orchidaceae, Piperaceae, Rubiaceae, and 
Sapindaceae. 
The present-day subandean forest belt (= lower 


montane forest belt) extends from 800-1200 to 2300— 
2500 m. and average annual temperatures are from 
24°C at the lower altitudinal limit to 16 C at the upper 
altitudinal limit. Common arboreal taxa include 
representatives of the Araliaceae, Arecaceae, Brunel- 


liaceae, Cecropiaceae, Clusiaceae, Cyatheaceae, Eu- 


phorbiaceae, Fabaceae, Fagaceae, Hippocastanea- 
ceae, Lauraceae, Malpighiaceae, Melastomataceae. 


Meliaceae. Proteaceae, Rubiaceae, Rutaceae. and 


Solanaceae. The most species-rich families include 


Leguminosae, Orchidaceae, Polypodiaceae. Rubia- 
ceae, and Solanaceae. 

The present-day Andean forest belt (= upper 
montane forest belt) extends from 2300-2500 to 
3200-3300 m. and annual temperatures average 


16 C at the lower altitudinal limit to 9 C at the upper 


altitudinal limit. The upper altitudinal limit reflects 


the upper forest line. Palynological studies from deep 
cores from the basin of Bogotá show that during the 
Pliocene the upper forest line ene eee to a higher 
temperature than today apa mstra, 1984: Torres, 
2000: | prep.). 
the Mioce na 'ene temperature range 
its 


Araliaceae, 


Torres & Hooghie amstra, 
may 


been ca. l6 C to ca. Common families 


eri med Betulaceae, Caprifolia- 
Cu- 


noniaceae, Cyatheaceae, Elaeocarpaceae, Ericaceae, 


ceae, Chloranthaceae, Clethraceae. Clusiaceae. 


— 
d 


“uphorbiaceae, Fagaceae, Juglandaceae, Lorantha- 


ceae, Melastomataceae, Myricaceae, Myrsinaceae, 


Sola- 


Thymelaeaceae, 


Myrtaceae, Arecaceae, Proteaceae, Rosaceae, 


naceae, Symplocaceae, Theaceae, 


and Winteraceae. The most species-rich families 


include Asteraceae, Ericaceae, and Polypodiaceae. 
The present-day subpáramo zone is immediately 
above the upper forest line and extends from 3200- 
3300 to ca. 3000 m, and annual temperatures average 
9"C 


subpáramo belt is dominated by shrub and dwarf 


7 C at its upper limit. The lower part of the 


forest, whereas the upper part is characterized. by 
The 
subpáramo belt developed further 
2006: 


Hooghiemstra, in prep.). During the late Miocene and 


dwarf shrub vegetation. 
the 


during the late Pliocene (Torres, 


pollen record of Funza-2 
shows that 


Torres & 


early Pliocene, the status of the subpáramo belt i 


bl 


uncertain and perhaps it had not developed into 


a discrete belt as it did during the Quaternary. 


Characteristic subpáramo taxa include Asteraceae, 


Clusiaceae, Ericaceae, Escalloniaceae, Melastomata- 


ce? Myricaceae, Rosaceae, Rubiaceae, and Sero- 


phulariaceae. 
The present-day grasspáramo belt extends from ca. 


3500 to ca. 4200 m, and annual temperatures average 


Annals of the 
Missouri Botanical Garden 


L 1 
Late Miocene 
Holocene 


=> sediment flux 


Atlantic Ocean 


Pacific 
Ocean 


~ Guyana shield 


n "d 


Amazon basin 


Amazon ubi 


L 
late Middle Miocene 


l 

=> sediment flux 

marine incursions 
limit marine incursions 

developing 


Magdalena S] 


drainage 


system 


Atlantic Ocean 


Pacific 
Ocean 


Llanos 


|i 
, basin 
Xx — 
x 
"oi E 
* 


yO 


— 
NS EN 
sal fresh water Oe, E 
1 fluvio-lac string system alte rations — \ 
estuarine SRM "i wes | 
— Em ertet 
pi Teo Amazon ae ^ ru e. 
River Pd mo s 
— A i 


Late Oligocene 
early Middle Miocene 


L (i 
mountains in development 
-»- sediment flux 


Pacific 
Ocean S basin . , 

GSE f ae Guyana shielc 
Ñ i ) 


MON uer EN UN 
VAS Amazon basin f i Sur 


Solimoes basin 


24 
3 
I bot 
uvio-lacustrine system 
estuarine character P . 
salUfresh water 
alternations >) 
cene to Present. (mide P ) late Middle Miocene, and (bottom) early 
; un à ps yon nt of northern 


` Amazonian rainforc 
own. During periods of 
basin, mangrove vegetation developed under brackis 


ndes, the s separation of Chocó and Lower 
. and the a lle Miocene fluyio-lacustrine system in the 
[high s sea level stands, at various places in the present Amazon 

vater conditions (the extension of an inland sea is indicated by a dotted 
line). Between the late Miocene and the Quaternary, the modern Orinoco River system developed from the paleo-Orinoco River 
and the Amazon River developed as a transcontinental drainage system toward the Atlantic. Reprinted from Hooghiemstra & van 
der Hammen (1998 


), with permission from Elsevier. The Hooghiemstra & van der Hammen (1998) figure was created based on 
Hoorn et al. (1995). 


Volume 93, Number 2 
2006 


Hooghiemstra et al. 303 


Paleobotanical Record of Colombia 


TC at the lower altitudinal limit to 3°C at the upper 


altitudinal limit. The vegetation is dominated by 
graminoid communities associated with stem rosettes 
of mainly Espeletia Mutis ex Bonpl. The most species- 
rich families in the subpáramo and grasspáramo belts 
are Asteraceae, Ericaceae, Poaceae, and Polypodiaceae. 

The present-day superpáramo belt occurs above the 


4200 m up to the 


snowline at ca. 4800-5000 m. The annual tempera- 


grasspáramo and extends from ca. 


ture averages between 3°C and 0 C. The vegetation 
cover is scarce and patchy due to frequent frosts and 
unstable soil caused by frost heaving. The vegetation 
of this belt does not play a relevant role in estimating 


paleoaltitudes from paleofloras. 


CHANGES OF THE PALEOENVIRONMENTAL SETTING OF 
NORTHERN SOUTH AMERICA SINCE THE MIOCENE 


The geologic and paleogeographic history of the 
northern Andes is principally related to the relative 
motion of continental and oceanic plates. During the 
South 


Farallón Plate in the Nazca and 


late Oligocene of northwest America, the 


breaking up of the 
Cocos plates enhanced tectonic activity along the 


Pacific side of the continent and induced the uplift of 


the Andes 
Smith, 
the closing of the Isthmus of Panama during the 
Pliocene (Keigwin, 1978, 1982; Duque Caro, 1990). 


The uplift history of the three Cordilleras in 


Wortel & Cloetingh, 1981; Wortel, 1984: 


1985). Plate motions were also responsible for 


— 


Colombia stretches across the Neogene, but in many 
places uplift was most important during the period 
between the Oligocene and the Pliocene, with highest 
rates during the Pliocene (van der Hammen, 
1962; van Houten € Travis, 1968; 
van der Hamen et al., 1973; Fabre, 1983; Kroonen- 
l., 1990; van der Wiel, 1991; 
1995). Uplift in some parts of the Western Cordillera 


uplift 
1961; 


Harrington, 


bere et a Hoorn et al.. 
e 

occurred during the Miocene. Parts of the Colombian 

22 and 

parts of the 


Central. Cordillera rose between 18 million 
(Ma), 
Cordillera were uplifted between 10 and 4 Ma. The 
Massif of the 
uplifted between 16 and 12 Ma (Kroonenberg et al., 


vears ago whereas Eastern 


Santander Eastern Cordillera was 


1990). In the Upper Magdalena Valley, the Garzón 
Massif was uplifted around 12 Ma, which resulted in 
a fundamental change in the Magdalena Rivers 
direction of flow (van Houten & Travis, 1968: van 
der Wiel, 1991; Guerrero, 1993; Flynn et al., 1994). 


SHORT RECAPITULATION OF PALEOBOTANICAL EVIDENCE FOR 
UPLIFT FROM SITES IN LOWLAND AMAZONIA 
Miocene and Pliocene sediments from exposures in 


river valleys in Colombian Amazonia showed pollen 


records that could be dated based on biostratigraphy 
(Hoorn, 1993; Hoorn et al., 1995). Pollen records from 
these landlocked sites showed alternations between 
vegetation characteristic of inland wet forest, open 


vegetation. of coastal plains (with abundant. grasses 


and palms), and mangroves. These environmental 
changes strongly suggest that during those time 


intervals marine incursions were able to reach the 
area that is covered at present by the rainforests of 


northwest Amazonia. In particular, during periods 
with high sea-level stands, marine waters penetrated 

freshwater 
lores!) 


(mangroves) 


far into the Amazon basin, changing 


ecosystems (swamps and inundated into 


brackish and saltwater 


(Fig. 3). 


such changes would have been at extremes, during the 


ecosystems 
Because ecophysiologic conditions during 
Neogene a significant part of Amazonia had to have 
been repeatedly fully repopulated with a “new” 
ecosystem. 

Uplift of the northern Andes is substantiated by 
a change in the drainage pattern: during the early 
Mi 


source of sediment and the paleo-Orinoco River 


dle Miocene the Guayana Shield was the main 
drained northward into Lake Maracaibo (mainly east- 


to-west flow direction in Fig. 3). During the late 


Miocene and Pliocene, the Andes gradually became 
the main source of sediment that substantiated the 
process of uplift (mainly a west-to-east flow direction). 
Closure of the Andean chain by the uplift of the 
Eastern. Cordillera forced the paleo-Orinoco River to 


migrate eastward to its present-day position. 


PALEOBOTANICAL EVIDENCE FOR UPLIFT FROM SITES IN THE 
HIGH ANDES 
macrofossils were 


Fossil pollen and botanical 


studied sediments from sections collected at six 
2550 m altitude in the surroundings 
ag. l; 


The fossil taxa were arranged according to 


exposures al ca. 


the present-day high plain of Bogotá ( 
Table 1). 


their modern altitudinal distribution. 


M any taxa have 


altitudinal ranges in common and such ecological 


fact 


groups constitute vegetation belts, which are ii 
abstractions. Modern altitudinal ranges are based on 
studies of the existing vegetation. The assignment of 
a fossil taxon to a particular ecological group is based 
on the principal altitudinal range of its present-day 
shows the relative pro- 


raphic source areas (Fig. 4; 


relatives. Every time slice 


portion of the phytogeog 
Appendix 1). The phytogeographic units are after van 
der Hammen and Cleef (1986) and Gentry (1986). 


Additional information on the phytoge ographic distri- 


— 


bution of the identified fossil taxa was obtained from 
Kappelle et al. (1992) and Mabberley (1993). The 


biostratigraphic range of selected pollen and macro- 


304 Annals of the 
Missouri Botanical Garden 


| all fossil plant taxa l taxa of regional vegetation only 
| - 3 — B UM 
LL] — 
Z, — 
1 > 40 n=73 07 n=38 
35 
O o 30 m 
e 8 2 25 
; 20 
N 15 15 
— O 10 10 
O = 5 Füquene 2 
aB ea 1 2 3 4 5 6 7 1 2 3 4 5 67 8 9 
— 
— 
— 2 n=68 5 n=40 
o 30 30 
o [e 25 25 
a 20 20 
— = E 15 
S 10 10 
E 5 Guasca 103 5 II 
ea l 2 1 4. 56789 1 2 3 4 5 67 8 9 
LH 
40 n=97 40 n=64 
35 351 
2 30 30 
— 25 25 
"d 20 201 
S 15 15 
E 10 F TEE H 
neata y 5 
5 acatativa 13 5 = — 
23 4 5 6 7 8 9 


PLIOCENE 
| 


40 n=68 40 7 n=57 


35 


30 

25 251 

20 20- 

15 15 

10 i E 10 1 

III Subachoque 39 3; i 
l 2 3 4 5 6 7 8 9 


Biozone II 


„ 
— 
t2 
o 
D 
o 


Ey 
3 19 n=29 e n-22 
30 30 
\ 25 25 
15 15 
\ 10 10 
" 4 Río Frío 17 ‘i NE NB 
9 1 2 3 4 5 6 7 8 9 


c DOE xoxo 5 4 
MV 
3 8 2 
El] o 
pud 40 1 =98 a S = 
LL — € EN ral n=98 E 3 n=70 
Ol3|| 2 2 xol E: 
— N 5 251 = 
Os O 2 20 3 
— 2 — 215 & 
S| [emi ^s = ü 
¡Arma [L1 L1.——J a — . = í Salto de - Cow = 
l 2 3 5 6 8 E l 2 3 $ 9 
Tequendama I & II 
1 unit 5 ae unit 
a wide tropical 4: tropical amphi-pacific 7: holarc 
2: Africa-tropical America 5: wide temperate 8: 12 55 T 
3: neotropical 6: austral-antarctic 9: cosmopolitan 


Figure . Phytogeographic composition of neogene fossil plant assemblages and paleofloras (only selected taxa are shown) 
from six sections located at the border of the basin of Bogotá, Colombia. Sections are arranged from the late Miocene (bottom) 
to the late Pliocene (top) and the present-day composition is shown above the dashed line for comparison. Percentage 
calculations are based on all identified fossil plant taxa (left). or on a selected group that represents the regional vegetation 
only. Plant assemblages representative for the Holocene are taken from core Fuquene-2 (van Geel & van der Hammen, 1973). 


Volume 93, Number 2 
2006 


Hooghiemstra et al. 
Paleobotanical Record of Colombia 


fossil taxa obtained from these Neogene sediment 


sequences is presented in Figure 5. 


STATISTICAL ANALYSIS OF FOSSIL ASSEMBLAGES 


Floristic similarity between fossil assemblages was 


statistically analyzed (cf. Graham. 1992). Data per 
ecological group and per section were not normally 
distributed using 95% confidence intervals (CI). 


Therefore, the nonparametric Kruskal-Wallis 1-sam- 
ple test (Sokal & Rohlf, 


the degree of dissimilarity between the fossil pollen 


1981) was chosen to evaluate 


spectra of sections Salto de Tequendama I and ll. 
Subachoque 39, Guasca 103. 


was expected because 85 in- 


Facatativá 13. and 
Pseudosignificance 
dividual comparisons were carried out. Pseudosig- 
nificance means that if for all tests a 95% CI is used, 


5% 


decisions based on conventional levels of significance 


X random data lie outside the 95% CI and 


may be incorrect. Therefore, a more conservative 


approach was used that lowers the critical probability 
P for 


probability of an erroneously rejected null hypothesis 


each individual comparison so that the 
in the entire series of tests does not exceed the general 


critical alpha. The level of confidence for the 


individual comparisons was calculated as follows: 


(L represents the number of comparisons) 


In this case, the level of confidence is 0.0006. 

The results of the Kruskal-Wallis tests are shown in 
Table 2. 
differed significantly in all 


The tropical lowland taxa assemblages 


sections. For the sub- 
andean-tropical lowland and Andean-subandean as- 
Tequendama II differed 
Additionally, 
“acalalivá 13 


semblages, site Salto de 


sienificantly from all other sites. for the 
o J 


Andean-subandean assemblage, site 


differed significantly from site Guasca 103. For the 
Andean assemblage, sites Salto de Tequendama I and 


each other as well a 


de 


Il differed. significantly from 


from all the other sections. For the subpáramo 
assemblages, both sites Salto de Tequendama I and 
Subachoque 39 differed significantly from Facatativá 
13 and Guasca 103. 


affinity between the studied associations is consistent 


The general decrease in floristic 


with the decreasing age of these associations. This 


trend shows that each flora is slightly different in 


floristic composition from its neighboring floras. 


PALEO-PHYTOGEOGRAPHA 


The fossil floras from sites in the Bogotá area show 
a gradual change in plant N from the late 


Miocene to late Pliocene. The late Miocene warm 


tropical lowland taxa are lr replaced by 


taxa with pre-montane and ultimately montane 


affinities (Figs. 4 and 5). During the late Miocene to 
late Pliocene, the following trends in the phytogeo- 
graphic composition of the floras can be observed 


(Fig. 4): (1) 


decreased slightly, (2) neotropical elements decreased 


the percentage of wide tropical taxa 


markedly. (3) the percentage of cosmopolitan taxa 


increased; (4) Andean neotropical taxa were present 


from the middle Pliocene onward. and (5) wide 


temperate taxa increased significantly. During the 


late Miocene, Amazonian-centered families made up 


the majority of the local forest taxa (Fig. 6). Toward 
the late Pliocene, the proportion. of this group 


Laurasian elements showed the 
der Hammen «€ Cleef, 1983). 


Andean-centered 


eradually decreased. 
opposite pattern (van 
laxa 


whereas the proportion of 


remained approximately the same. The calculated 
similarity values reveal the degree of floristic affinity 
(Table 2). The 


values of 


associations general 
that 


affinity decline with an increasing difference in age 


between the 


observalion is calculated floristic 


between the floras. 
Van der Hammen et al. (1973) used the successive 
such as 


arboreal taxa, 


— 


immigration o 
Alnus Mill., 
Myrica L., 


lemperate 
Hedyosmum Sw. (not strictly temperate), 
and Quercus L., to construct a biostrati- 
biozones 
for the sediments in the study 3) (ef. 
Kuhry & Helmens, 1990: van der Hammen & 
Hooghiemstra, 1995; Wijninga, 1996d: 


Torres & Hooghiemstra, 


graphie framework. He established seven 


area (Table : 
van der Ham- 
men & Hooghiemstra, 1997; 
in prep.). 
UPLIFT or THE NORTHERN. ANDES EVIDENCED BY 
PALEOBOTANICAL DOCUMENTS FROM LOW AND 
LIE ALTITUDES 


In the Bogota area, middle Miocene sediments show 
dominant warm lowland taxa, whereas late Miocene 


and Pliocene sediments show increasing proportions 


of laxa associated with high-elevation vegetation 
(Table 4: Figs. 4 and 5). This trend in vegetation 


change is primarily explained in terms of a gradual 


Phytogeographic re 7 E x of the identified taxa are according to van der Hammen & Cleef 
The neotropical and Andean neotropical groups have no taxa in common. (cf. genus name = sp. ined., 
alf. genus name). Modified after Wijninga (1996d). 


Kap] elle et al. (19 
e 


genus nante; 1 genus name = sp. ine d.. 


f (1986). Mabberly ( ia and 


Annals of the 
Missouri Botanical Garden 


Miocene 


| Middle? Lates Early 
hs 


Tequendama 
Member 


r Tilatá Formal 


Tibagota Member 


Pliocene | 
Chronostratigraphy 
| Middle | Late 
ES 
tion J 


Lithostratigraphy 


= Biozones 
[ss] 


Veguen ndama — Frio 
& Il 17 


Salto de Rio Suba- 
choque 
E 


Facatativa Sections 
13 


Guasca 103 


cf. Chrysobalanaceae (others) 
Juncaceae 3 rs) 
heedia 


el. Connarus 
Catostemma 
Sapotaceae (others) 
cf. Thoracocarpus 


Tiiaceae - == == == 4 
Conve aulicis 


Parinari 
Humiriastrum 
Sapindaceae 


cf. Chrysobalanaceae (others) 
ncaceae (others) 


heedia 
clinusa 
Asplundia 
Connan 
atostemm 


Thorac arpus 


Byttneria 
Cordia 
yanraadshooyvenii 
Ceiba 
Pseudobombax 
cf. Psychotria 
Anthurium 
Annonaceae 
Passiflora 
Pityrogramma 
Vochysia 
Tapirira 
Tilrac 
0 DAVAN ulag eae 


croton — — — — 4 roton 
Iriartea Iriartea 
Macrolobium B ^" Macrolobium 
Protiu ^ Protium 
Selaginella 4 Selaginella 
Schefflera 4 Schefflera 
Warszewiczia 4 Warszewiczia 
Eugenia 4 Fugenia 
ocynacea: 4 Apocynaceae (others) 
Leguminosae (others) — — — — 4 E Leguminosae (others) 
Meliaceae (others) — — — — 4 4 Meliaceae (others) 
Euphorbiaceae (others) — — — — 4 4 Euphorblag eae (others) 
Cedrela = — — — 4 4 Cedrela 
Pouteria F- "i Pouteria 
Faramea — Faramea 
Podocarpac NER (others) — — — — 4 — E Podocarpaceae (others) 
nihaceae Pp — — — +4 p+ 4 Acanthaceae 
Pha pec ee e c 91 3 Ficus 
Ericaceae [o — — ĩ +t ——— ricaceae 
Salvia — — — J — — — 4 alvia 
Malpighiaceae — —— ——— — — Malpighiaceae 
Clusia PT . alee pui ec Mira Clusia 
Alchorne. Alchorne 
Urticales (others) Urucale (others) 
Cecropia — — — —| Cecropia 


cf. Evodianthus 
Palmae (others) 


Ev odianthus 


Palmae (others) 


aff. Thuidiaceae 
aff. Hookeriaceae 
Anacardiaceae (others) 


Proteaceae — SS Se — Proteaceae 
Commouedub j — — — Compositac-tub. 
Umbelliferae E cm — — — OO Umbelliferae 
a Freziera 

olla oos SO. S ae o Azolla 
Malyorum Hedyosmum 
Byrsonima — Byrsonima 
Vismia r Vismia 
Bauhinia — — — Bauhinia 
ursera — — Bursera 
Thelypteris E — Thelypteris 
Muehlenbeckia E - Muehlenbeckia 
Xylosma — — Xylosma 


aff. Thuidiaceae 
aff. Hookeriaceae 


Anacardiaceae (others) 


= Biozones 


— 


Figure 5. 
39, es apu vá- 13, and Guasca- 103. 
Pe lines = 10 


E 0.5 Ma. Percentages (thic 


increase in the elevation of the sedimentary environ- 
ment and thus reflects a decrease in temperature. 
However, a temperature decrease is also expected at 
the transition from the late Pliocene to early 
Pleistocene, when temperatures decreased on a global 
scale and the period of Pleistocene ice ages started 


& Carlson, 1981: Kennett, 1995; 
Shackleton, 1995). Despite this cooling trend, paleo- 


(e.g... Lohman 


climatic studies show that low latitude sea surface 


temperature showed much less variation, generally in 


Marker taxa for biozones 
k %: thin lines 
representation in the pollen or macrofossil diagrams of the respective paleofloras (Wijninga, 1996 


see ible 3 3 


Biostratigraphic range of fossil plant taxa in sections S 79 P Te MAN Land HL Rio F 1 5 7. . 1 
| Mi 


) are under . Time control: a; 2 


= oo ri | lines = 1%) are 1 on average 
l) 


the range of 3°C 
Shackleton. 1982; 


therefore assume 


1975; 
Tiedemann et al. 
thal 


montane equatorial South America were maximally 


Savin, 1977; 
1994). We 


temperatures T 


(Savin et al., 


Neogene 


| higher than at present. A maximum difference of 
: s corresponds to a 500 m maximum shift of the 
altitudinal vegetation belts when the modern temper- 
ature gradient of 6 C/1000 m is applied (Meyer, 1992; 
Witte, 1994). This potential temperature shift does not 


explain the difference between present-day altitude 


Volume 93, Number 2 
2006 


Hooghiemstra et al. 307 


Paleobotanical Record of Colombia 


Miocene Pliocene 
Chronostratigraphy 
Middle? EN Early [ Middle Late 
Lower Tilatá Formation Upper Tilatá k. 
| Lithostratigraphy 
Tequendama Tibagota Member Guasca Member 
Member 
„ 8 
E Le 
3 sd Eos Paca Guasca 103 Sections 
1 17 2 
Araliaceae (others) Araliaceae (others) 
Valeriana Valeriana 
Rhus pe ae, Rhus 
Tristerix ilius! Tristerix 
Sapium — — — Sapium 
Guarea [- os Gua: 

Weinmaania Weinmannia 
Myrsine — Myrsine 
Hypericum — — Hypericum 
Solanaceae — — — — — — - Solanaceae 
Viburnum H — — —- Viburnum 
Gunnera - =- — p— Cun 
Typha - = — E — — — ~| Typha 
Billia pem ek: aks eee DIN 
Salix [tec cime Salix 

-lig. 
Daphnopsis Daphnopsis 
Celtis Celtis 
Piperaceae Piperace 
lea Vallea 
Macrocarpaea Macrocarpaea 
Aonnina Monnina 
asia Artemisia 
Bocconia Paccani 
Bomarea 
Bombacopsis ops is 
Bignoniaceae (others) [riens (others) 
Bromeliaceae eae 
Calceolaria 
Althernanthera 
Dicksonia 
Cystopteris 
Gaiadendron 
Lathyrus 
Mabea 
Malvaceae 
Rumex 
Qu ibea 
Relbunium 
Myriophyllum 
Tournefortia 
Sericotheca 
Lythraceae 
Xyris 
Elatine 
Callitriche [moe cem 
Carex pee ec m 
Acalypha — — — 
Antidaphne — — — 
i 5 
Montia EE 
Borreria fece dmm 
Myrica 
Eleocharis 
Draba 
Lepidium 
Hesperomeles 
Ranunculus 
Rubus 
Turpinia 
Sphagnum 
28 
aff. Prionolejeunia 
ydista 
pecia 
Fuchsia uchsia 
Decu 758 €——— Decussocarpus 
Hippocratea/Pristimera ÉL——-—- Hippocratea Pristimers 
Lust X. a a Cou 
Ipomoea | [[[ TE — — — Ipomoea 
aff. Lejeuna | | | | TT — — — 3/Y) aff. Lejeunia 
aff. Breutelia |. |. [( T — — — aff. Breutelia 
Alismataceae — p Alismataceae 
Cyperaceae other — Cyperaceae (others) 
Gri ineae ramineae 
Podocarpus Podocarpus 
Cyatheaceae Cyatheaceae 
Hex Hex 
Symplocc —— — — — Symplocos 
Melastomataceae elastoma ie 
Myrtaceae (others) eee € Myrtaceae (others) 
Hyeronima Hyeronima 
Rubiaceae (others) A Uer, Tum. pui Rubiaceae (others) 
CITA que rm mmm om exec U M Clethra 
Lycopodium — — — — 4 — — —— - Lycopodi 
Nats pe te cms aay [prie cep qum CaS 
Polygonum — — — — - 4—-—-L---- Polygonum 
Bombacaceae (others) PS = 3 ot = — — — — -| Bombacaceae (others) 
Loranthaceae ud Lora N (others) 
Ludwigia Ludw wigi 
1 Urticaceae? 
— = Biozones 
> c 
NE 
Figure 5. Continued. 
and inferred paleoaltitude of sections Salto de span, the montane forest evolved. The rate of 
"p T "p y 5 1 7 . E | . . 
l'equendama-l, Salto de Tequendama-ll, Río Frío- development in the study area may have been 


17, As a the 


change 


and Subachoque-39. consequence, 


n vegetation and the inferred temperature 


decrease must have been primarily caused by tectonic 
uplift. When uplift was quantified and plotted against 
time, the Pliocene was identified as the period. with 


main uplift (Fig. 7). 


From the late Miocene onward, a period « 


approximately 2 to 3 million vears elapsed before 
the area of sedimentation was uplifted to altitudes 
equivalent to the present-day lowermost limit of the 
Andean forest belt (2300-2500 m). During this time 


accelerated. by the presence of montane forest. in 
other areas. Several middle to late Miocene pollen 
records from northwestern South America show pollen 
erains of Podocarpaceae. Podocarpaceae pollen grains 
may potentially belong to Nageia Gaertn., Podocarpus 
Prumnopitys Phil., and/or Retro- 
phyllum C. N. Page (Lorente, 1986: Hoorn, 1993, 
1994a, 1994b; Behling, 1996; Wijninga, 1996a). In 


the northern Andes, representatives of the Podocar- 


lL'Hér. ex Pers., 


paceae occur mainly as single trees or as groups of 


trees in the Andean upper montane rain forest; a 


308 


Annals of the 
Missouri Botanical Garden 


Table 2. 
Bogotá: 


Miocene to late Pliocene, respectively. 


Significant scores are in bold: — — no data available. 


Individual significance. level, 


Results of the Kruskal-Wallis I-sample test applied to the floral assemblages of five sections from the area of 


Salto de Tequendama I and H. Subachoque 39, Facatativá 13, and ao 103: sections are dated from middle 


P = 0.0000: 


global significance level, % = 0.05. 


Results show floristic end belween five sections for five different 


ecological assemblages. ranging from tropical lowland to subpáramo. 


5. de Tequend. H 


Subachoque 39 


'acatativá 13 Guasca 103 


Tropical lowland assemblages 
5. de Tequendama | 0.0000 

S. de Tequendama H 

Subachoque 39 


‘acatativa 13 


Subandean-tropical lowland assemblages 
S. de Tequendama | 0.0000 


S. de Tequendama I 


— 


Subachoque 3 


acatativá 13 
\ndean-subandean assemblages 
5. de Tequendama | 0.0000 


5. de Tequendama H 


— 


Subachoque 3e 


“acatativá 13 
Andean assemblages 

5. de Tequendama | 0.0000 
5. de Tequendama II 


Subachoque : 


"acalalivá 13 
Subpáramo assemblages 

S. de Tequendama | — 
S. de Tequendama H 


Subachoque 39 


“acatativá 13 


0.0000 0.0000 0.0000 
0.0000 0.0000 0.0000 
0.0000 0.0000 
0.0000 
0.0104 0.0051 0.8114 
0.0002 0.0000 0.0000 
0008 0.3717 
0.1524 
0.5207 0.0010 0.0002 
0.0000 0.0000 0.0000 
0.2072 0.0148 
0.0000 
0.0000 0.0000 0.0000 
0.0000 0.0000 0.0000 
0.5002 0.5100 
0.1646 
0.5940 0.0000 0.0000 
0.0002 0.0003 
0.3103 


present Podocapaceae have been removed in most 


forests to produce timber. Stands of Podocarpus- 
dominated. forest have been reported from, among 
Mérida 2500 m 


Serranía de Perija between 


others. the Andes ca. (Vareschi, 
1980), the 3300 and 


3500 m (J. O. Rangel, personal communication). and 


— 


near Loja, Ecuador, in the Podocarpus National Park. 
The lowermost part of the Funza-2 pollen record 
(Torres. 2006: Torres & Hooghiemstra, in prep.) shows 
that before 2 Ma Podocarpaceae were more abundant 


in equatorial montane forests. Patches of modern 


Podocarpaceae-dominated forest may represent the 


ast relicts of the original montane forest. Podocarpa- 
ceae-dominated forest was more common before the 
immigration of arboreal elements from northern and 
southern latitudes and before warm neotropical taxa 
adapted to cool and cold conditions. The late Miocene 
evidence of Podocarpaceae pollen grains points to the 
existence of montane forest, although it has to be 
stressed that only a few modern specles of. Podocar- 


paceae occur as rare trees scattered over the lowlands 


of Amazonia and in the Chocó biogeographic region 
2004). 


Podocarpus 


along the Pacific coast of Colombia (Rangel, 
According to MW. TROPICOS (2006). 
Buchholz & N. E. Gray 


from Panama to the Bolivian Amazon and in Colombia 


magnifolius J. is recorded 
and Venezuela up to ca. 2000 m in the lower montane 
rain forest belt. 

Earlier uplifted areas in the environs of the basin of 
Garzon. Santander 


Bogotá are the Quentame. and 


massifs, and the Central Cordillera. All of these may 


have been populated by some type of early Andean forest 


and may have been the sources for a newly developing 


flora and vegetation ie newly uplifted mountain 
areas. Taxa already Pos d to cool climatic conditions 
may have accelerated. the formation of monodominant 
upper montane forest in the area under study. 

Forty taxa of the modern páramo belt have been 
recorded in the fossil assemblages. Among these taxa 


are Xyris L., Miconia Ruiz & Pav.. Eleocharis R. Br., 
which are shared with lowland 


flora (Cleef et al., 1993). 


and Scirpus L.. savanna 


Vyris probably originated 


Volume 93, Number 2 Hooghiemstra et al. 309 
2006 Paleobotanical Record of Colombia 


45 % 


40 % 4 


35 % 5 


T 
SER 
A +, 
5 


+ 


A 


+ 
es 


2 


+, 

A 

* 
t 


y + +, 
30% 4 EM 22 5 
Woy? VS . ~ 
ue AH Bl Amazonian-centered Gondwana 
25 om 
25 K um 5 
25 Y 98500 Andean 1G l 


ttt 
1 
ODO 


E] Laurasian 


Others 


+++ 
n 
m 

CS 
Se 


C» 


9 


> 


12 5 


“xd 
<> 
e 


e 
2^ 


percentage of recorded families 
KDE 
S 


10 * 


ms 
ere: 
XS 


950.6. 
<> 


Y 

¡nee 
IS 
eS 
— 


Guasca 103 Facatativá 13 Subachoque 39 Río Frío 17 Saltc 
Tequendama I & II 


ate Miocene 


— 


Late Pliocene 


\ndean-centered Gondwanan. and Laurasian 


Figure 6. Relative contribution of Amazonian-centered Gondwanan. 
families to the composition of the fossil plant assemblages of sections Salto de Tequendama-l and H (Middle Miocene). Rio 
Frío (late Miocene to early Pliocene), Subachoque-39 (early Pliocene), Facatativá-13 (middle Pliocene), and Guasca-103 (late 
Pliocene). Modified after Wijninga (1996d). 
from the open warm tropical lowlands and underwent In addition, other characteristic taxa of modern 
subpáramo and grasspáramo, such as Acaena Mutis ex 
L.. Polylepis Ruiz & Pav., Aragoa H.B.K., Arcyto- 
phyllum Willd. ex Schult.. Bartsia L., Caryophylla- 
macroremains, of Xyris ceae, Clinopodium L. (= Satureja), Eryngium 
1996b), indicate swamps are present in the different Plantago L., and Puya Molina, were absent during the 
Thi former Pliocene. Van der Hammen and Cleef (1986) 


gradual adaptation to cool climatic conditions during 
the uplift of the Eastern Cordillera. Finally, Xyris 
became part of the páramo vegetation. However, plant 
ij in particular (Wijninga, 


— 


forest belts. This does not exclude their 
occurrence in exposed mountain savannas as sug- postulated the presence of a hypothetical “pre- 
páramo" (i.e., an open type of vegetation occurring 
locally on exposed hilltops before the final uplift o 
the Eastern. Cordillera) and also a “proto-páramo” 


(i.e. early páramo vegetation with characteristic 


gested by Cleef et al. (1993). Miconia is mainly an 


important component of the understory of (succes- 
sional) lowland and montane rain forest. Wijninga 
(1996b) identified remains of Pliocene seeds to the 
taxonomic levels of the Miconieae and Miconia. It is päramo taxa, but still poor in species). The suggested 
difficult to assess whether species of Eleocharis and local presence of late Miocene open vegetation on 
hilltops may represent pre-páramo vegetation (Bio- 


in the northern Andes are from tropical or 
zone l; Grubb, 1971: Wijninga, 1996a). The fossil 


Scirpus 
temperate origin. 

Table 3. Main characteristics of northern Andean Biozones I to VII (after van der Hammen et al., 1973). Biozones IV to 
VII also follow from cluster analysis of the pollen assemblages of the Funza-2 record (Torres, 2006; Torres & Hooghiemstra, 


in prep.). 


Biozone Age First appearance date/Taxon dominance 


Biozone I middle to late Miocene abundant Mauritia. Iriartea, Bombacaceae 
Biozone lI early Pliocene start of the record of Hedyosmum 
start of the record of Myrica 


Biozone IH ate Miocene 
abundant Borreria 


Biozone IV middle Pliocene 
Biozone \ late Pliocene abundant Juglans and Plantago 
Biozone VI early Pleistocene start of the record of Alnus 


Biozone VII middle Pleistocene start of the record of Quercus 


Recovered pollen taxa in six sections located at the border of the basin of Bogotá, arranged from the late Miocene (left) to the 


able 4. 
modern ecological and altitudinal affinities. Data from Wijninga (19960). 


ate Pliocene (right). 


Taxa are arranged based on 


Salto de Tequendama 1 Salto de Tequendama II 


Río Frío 17 


Subachoque 39 


Facatativá 13 


Guasca 103 


Grasspáramo taxa 
(3500-4200 m) 

Subpáramo ta 
(3200— 3n. m)/ 
montane open 
vegetation taxa 

Andean forest 
belt taxa 
(2300-3200 m) 


Andean forest belt- 


ubandean 


— © 


orest belt taxa 


(1000-3300 m) 


Poaceae! Poaceae’ 
Asteraceae-tubuliflorae 
Asteraceae-liguliflorae 

Ericaceae 

Clethra Hedyosmum 
Hedyosmum 


Symplocos 


Clusia Melastomataceae 
Clusiacae (others) Myrtaceae (others) 
Melastomataceac Podocarpus 
Myrtaceae (others) Proteacae 


Podocarpus 


Proteaceae 


Poaceae’ 


Ericaceae? 


Hedyosmum 


Symplocos 


Me 


pl 


astomataceae 


Myrtaceae (others) 


Podocarpus 
Proteacae 


Poaceae! 


Asteraceae-tub.! 


Hypericum 


Clethra-type 
Eugenia-type 


Hedyosmum 


Weinmannia 


Xylosma 


Araliaceae 

Billia 
Clusia-type 
Clusiaceae 
Melastomataceae 
Myrtaceae 
Podocarpus 
Proteaceae 

Salix 


Solanaceae 


Poaceae’ 


Asteraceae-tub. | 
Hypericum 


Bocconia’ 
Clethra” 
Gaiadendron 

lum 
Ilex 


Macroc arpea 


Symplocos 

Va 

Weinmannia 
Araliaceae? 

Celtis 

Clusia-type 
Daphnopsis 
Euphorbiaceae Men) 
Melastomatac "eae! 
Monni 

Myrtaceae (incl. Eugenia) 


Solanaceae 


Poaceae! 


Asteraceae-tub. | 
Hypericum 


Clethra-type 
Eugenia 
Hedyosmum 
Hesperomeles-type 
Ilex 

Myrica 

Myrsine 
a x 
Symploco 

Viburnum 


Weinmania 


Billia 

Clusia-type 
Melastomataceae! 
Myrtaceae (others) 
Podocarpus 
Proteaceae 


Solanaceae 


USPIEL) jeoiuejog OSS 


018 


S JO sjeuuy 


Table 4. Continued. 


Salto de Tequendama I 


Salto de Tequendama H 


Subachoque 39 


Facatativá 13 


Guasca 103 


Subandean forest 


belt-tropical 
v 


(0-2300 m) 


Alchornea 
Apocynaceae (others) 


Arecaceae (others) 


Burseraceae/Sapotaceae 


Cecropia 


cf. Euphorbiaceae 
Fabaceae (others) 


Rubiaceae (others) 
Schefflera 
Urticales (others) 
Warszewiczia 


Alchornea 
Apocynaceae (others) 
Arecaceae 

Cordia? 

Faramea 
Malpighiaceae 

cf. Rubiaceae 
Schefflera? 


Urticales (others) 


Alchornea 


Apocynaceae (others) 


Bauhinia 

ropia 
Cedrela-type 
Euphorbiaceae (others) 


Guarea-type 


Paraprotium-type 
Rhus 


Rubiaceae (others) 


Urticales (others) 
Vismia 
Warszewiczia 


Zanthoxylum? 


Acalypha 
Alchornea 
Apocynaceae 
Arecaceae 
Cecropia' 
Cedrela 
Faramea 
“icus? 
Guarea 
Hyeronima 
Fabaceae (others) 
Mabea 
Malpighiaceae” 
Meliaceae 
Rhus 
Rubiaceae-type 
ubiaceae 
Sapium 
Tournefortia 
Urticales (others) 


Warszewiczia 


Cecropia” 

Hyeronima 

Malpighiaceae 
pig 


Meliaceae 


Salix 


Urticales (others) 


900 


z JequinN ‘£6 euinjoA 


BIQUIOJOD JO p1ooeH [eoiuejoqoo[ed 


e 19 eusuieiuBooH 


LIE 


Table 4. — Continued. 


Salto de Tequendama I 


Salto de Tequendama H 


Río Frío 17 


Subachoque 39 


Facatativá 13 


Guasca 103 


T 


ropical low land Acacia 
taxa (0-1000 m) Amanoa 

innonaceae 
Istrocaryum 
Bernoullia 
Bombacaceae (others) 
Brownea 
Byttneria 
Catostemma 
Ceiba 
cf. Chrysobalanaceae 
cf. Connarus 


Dalech ampia 


Ficus 
Humiriastrum 
Ilex 

Iriartea 

cf. Machaerium 
Macrolobium 
Mauritia 
Pouteria 
Protium 
Pseudobombax 
Rheedia 
Sapotaceae (others) 
Spathiphyllum 
Tapirira 

Virola 


Vochysia? 


Amanoa 


Annonaceae” 


Astrocaryum” 


Bombacaceae (others) 
Byttneria 
Catostemma 
eh 

Clusia? 
Eugenia? 

Ficus 
Humiriastrum 
Hex 

Iriartea 

cf. Machaerium 
cf. Macoubea 


Mac rolobi um 


cf. Psychotria 


Warszewiczia” 


Amanoa 


Astrocaryum 


Bombacaceae 
Ilex 

lriartea 

cf. Macoubea 
Mac vbi 
Maur 
9 


Amanoa 
Anacardiaceae? 
Bombacaceae 
Bursera” 
Byrsonima 
Humiriastrum 
Iriartea 

cf. Macoubea 
Macrolobium 
Mauritia 

cf. Sacoglottis 


Bignoniaceae 
Bombacopsis 
lriartea 
Macrolobium^ 
Pouteria? 
Protium? 


Quararibea 


Trichantera/Bravaisa? 


Bombacaceae 


' Poaceae may reflect mainly local ve gelation (and are excluded from the percentage calculations anc 


alc 


?Taxa with [um frequency (omitted from the statistical analvsis elsewhere in this paper). 


l statistical analysis elsewhere in this paper). 


USPIEO) jeoiuejog unossi|A 


cr€ 


eui Jo sjeuuy 


Volume 93, Number 2 Hooghiemstra et al. 319 
2006 Paleobotanical Record of Colombia 


inferred 
paleo 
Biozones altitude 
(m) 
IV-VII 2500 3 Ma (present elevation 2550 m) € 
1 
1 
i 
L 
1 1 
IIB 2000 h 
Oe l'acatativá 13 
— | D 
1 
Ex 
p 
1500 — 
HA 1000 4 
A Salto de Tequendama | & II 
EN 
500 — 
6 em 
| | | | | | 7 
1 2 3 4 5 6 16 Age (Ma) 


QUATERNARY PLIOCENE MIOCENE 


Figure 7. Main chronological. uu eri paleoenvironmental, and paleoaltitudinal properties of the five sections 
he n of Bogotá. Colombia. representing the middle Miocene to late Pliocene. 


from exposures in the outer valleys o 
Elevation of past depositional environments was estimate al by c ue. paleofloras with present-day equivalents. Se dim. rents 
were dated by fission track dating of intercalated volcanic ashes. ions make a diagonal in this age vs. paleo-altitude 
diagram, indic ‘ating uplift of the Zastern Cordillera duum the late Mi ocene and Pliocene. Vertical arrows correspond to an 
estimated uncertainty of ca. 3 C. The basin of Bogotá itself started to accumulate sediments when the Ande əs had reached its 
present-day elevation, which X reflected by the position of Funza-2 pollen spectra. For Biozones see Table 3. Modified 
after Wijninga (1996d). 


[om 


assemblages discussed here (Biozone H to Biozone HI) — record, and consequently first appearance dates from 


do not show evidence for the existence of proto- van't Veer and Hooghiemstra (2000) have been 

páramo vegetation. The first possible record of proto- corrected (Torres, 2006; Torres € Hooghiemstra, in 

páramo or páramo-like vegetation is from the late prep.: Torres et al., in prep.). 

Pliocene (Biozone IV; van der Hammen et al., 1973: Toward the late Pliocene, the fossil floras show at 

Helmens & van der Hammen, 1994; Torres & — the family level an enrichment of taxa with a Laurasian 

Hooghiemstra, in prep.). affinity, and taxa with an Amazonian-centered affinity 
Despite the fact that the uplift of the Eastern became less important. These trends are in accor- 


Cordillera ceased by the end of the Pliocene, the dance with the change from tropical lowland 9 8 
floristic composition of the Andean forest was subject ments to Andean forest conditions. The slight de- 
to change up to the late Quaternary. Throughout this crease in the percentage of Andean-centered elements 
period, forest types without modern analogues de- could be caused by the fact that in the montane forest 
per these families represent predominantly epiphytic and 


— 


veloped, existed for a while, and became furt 
modified by new migrants, such as Alnus and Quercus understory shrubs (Gentry, 1982), which are generally 
(van der Hammen & Gonzalez, 1964: van der underrepresented in the fossil record. 

Hammen et al, 1973: van der Hammen, 1974; 

Hooghiemstra, 1984; Hooghiemstra & Cleef, 1995; — Tye Dynamic PÁRAMO ARCHIPELAGO AND SPECIATION 
vant Veer & Hooghiemstra, 2000). The latest 

chronology of the sediments of the basin of Bogota During 


he Quaternary the upper forest line 


comes from astronomical tuning of the arboreal pollen migrated between ca. 2000 m (during the coldest 


314 Annals of the 
Missouri Botanical Garden 


0? 7 78? z 
— E fae - : — 
1000 m| 8000 [| € 
S QNO | 
i Gr 
Q 
* 


O [9] 
20 29 
< 
a Y us 
99199 Et | OO] 
© a O d / 
| 4j) 
| v E 
| = — as 25 2 
Páramos q iJ Páramos 
Subandean and — — Subandeon and 
— El Andean forest s = | is E Andean forest 0 
8 0 — 
0 50 100 150 E 0 100 150 : 
—* | Carramboa ECUADOR | 12. 7 "|  Coespeletia 


mE <l 1000m|3000m — 
pre] 
© — 


i pe 1000m LE. 
e — = | 
o = =N 909m — 


Paramos 


= —3 Subandean and : au a — Subandean and 
0° e | 8 E] Andean fores! ee 5 =Á — EJ Andean forest 5 
ri E ESE A AA 
o 3 o 
La 0 50 100 150 : be | 0 50 100 150 
ù — us Espeletia ECUAQOR — Hu 
EN -— 1 
79° 76° n? 722 70 78? 76° 705 72° 70° 


Figure 8. Maps showing the distribution of eight genera belonging to the Espeletiinae (Asteraceae). All genera are 
8 I Bh g ging I 8 
endemic to the northern Andes, and many distribution areas coincide with the geologically youngest parts of the Andes. 


Volume 93, Number 2 Hooghiemstra et 315 


al. 
2006 Paleobotanical Record of Colombia 


mal" 


a 
o 


Er 

— 462 
o9 o9 
zo 2° 
Tu . 
uy — wy — 

Qo| _O O9 | 4° 
og 


Paramos 
E de and 


dean forest 
0 


0 50 100 150 7 
| Paramiflos 


i 


ii 
2) 
P ; 


O 
20 
＋ i 
oo á 
0 0 
eee pu 4 
z o 


5 
À AY 
N. 
[9] 


Páramos Páramos 
Subandean and Subandean and 
EJ Andean forest Andean forest 0 


o 


Y vel Ruilopezia Tamania 
km 


78? 75 2 728 


Figure 8. Continued. 


3URITACA, SIERRA NEVADA CORD. ORIENTAL COLOMBIANA | 
DESANTA MARTA & VAUPÉS (MITU) 


SERRANIA S. LUIS p^ 


| temperate — 1 
> compone J 
PARAMO | E j 


number woody genera 


ANDEAN 


FOREST | 
100% 
SUBANDEAN 1 
| FOREST | | 
| h c m ar p n 


EQUATORIAL 
WETFOREST 


EQUATORIAL 
DECIDUOUS 
FOREST 


ll 


A 


Figure 9. Representation of geographic elements of woody genera for six different vegetation belts in northern Andean and adjacent lowland vegetation showing the d of 
neotropical lowland taxa to cool uw matic conditions. Data are specified (from left to right) for the Serrania San Luis (Venezuela) (subandean forest (1200-1500 m) and dec e fores 


CHOCÓ: BAJO CALIMA 


| 


1200 m) after Steyermark (1975)): the Buritaca area in the Sierra Nevada de Santa Marta (páramo (> 3300 m). Andean forest (2500-3300 m). and subandean forest (1300— 


Rangel & Jaramillo (1984)): 1 (dry deciduous lowland forest after Dugant (1941)) 


Amazonia near Trojita-La Brea, after Cuatrecasas (1958)). The bar to the left of each graph shows the number of woody genera and the proportion o 


elements: n = neotropical: p = pantropical: af = 


aupés (lowland rainforest of Amazonia near Mitú-C 
Cordillera of Colombia (Andean forests after van der Hammen & Jaramillo ( DR ished data). and páramos after Cleef (1979) with corrections): and Chocó (Bajo Calima) idani o st of 


African and American: m = southeast Asian and American: e = „ aa = Austral- 


Numbers in parentheses show representation (percentages) of floral x ‘ments based on a low total number of genera. After Cleef. unpublished data. 


aruru, after Cuatrecasas (1 Fs 


‘tropical and temperate taxa. Phytogeographic 


Antarctic: h = Holarctic: w = wide temperate. 


dope jeoiuejog LINOSSIN 


91€ 


ay) Jo s¡euuy 


Volume 93, Number 2 
2006 


Hooghiemstra et al. 
Paleobotanical Record of Colombia 


episodes of glacials) and ca. 3200-3400 m (during the 
warmest episodes in interglacials) (e.g.. van der 
1974; Hooghiemstra, 1984; van't Veer & 
2000: Wille et al., 2001). 
quence, the area covered with páramo vegetation has 
l 


Hammen, 


Hooghiemstra, As a conse- 


changed accordingly, from many small areas : 


mountaintops during interglacial periods (high posi- 
tion of the upper forest line) to large united areas 
during glacial periods (low position of the upper forest 
Hammen & 
the 


line) (van der Hammen, 1974; van der 
Cleef, 1986). 


Quaternary 


The series of ice ages during 


caused these altitudinal migrations to 
repeal many times, potentially leading to a relevant 
mechanism of speciation. 

Cuatrecasas (1979) studied species diversity and 
distribution areas of the Espeletiinae (Asteraceae). He 
identified the northern part of the Eastern Cordillera of 
Colombia (i.e., the Cocuy area and the area north of 
Chicamocha Valley up to the Mérida Andes) as the area 
with the highest number of species. The Espeletiinae 
include the genera Carramboa Cuatrec., Coespeletia 
Cuatrec.. Mutis ex Bonpl., 
Cuatrec., 
Ruilopezia Cuatrec., and Tamania Cuatrec. 
Espeletia 


Colombian and Venezuelan parts of this area, while 


Espeletia Espeletiopsis 
Libanothamnus Ernst, Paramiflos Cuatrec.. 
The genera 
and  Espeletiopsis are exclusive to the 
the genus Coespeletia is exclusive to the Venezuelan 
part of this area. These genera comprise morphologi- 
cally and ecologically diverse plants and occur in the 
altitudinal range between 2500 and 4500 m (Rauscher, 
2002). Figure 8 shows distribution maps of the genera 
of the Espeletiinae and shows this area as the center of 
origin of this taxonomic group (cf. van der Hammen $ 
Cleef, 1986). Notably, 


among Espeletiinae is also the place where the 


the area of highest diversity 
northern Andes is youngest, geologically speaking. 
According to Luteyn (1999), 


species of Espeletia and ca. 16 species of Espeletiopsis. 


HM 


Colombia has ca. 
Venezuela has ca. 15 species of Espeletia, ca. 5 species 
the 


Coespeletia. Many of these taxa show a very limited 


of Espeletiopsis, and ca. 6 species of genus 


distribution and sometimes are restricted to 


a single mountaintop (Luteyn, 1999). On the Quater- 


nary time scale, the Holocene is among the periods 


area 


during which distribution areas have a minimal surface. 
Placing the recent anthropogenic pressure on the 


páramo ecosystem in this context, many species are 
relatively close to extinction. 
The 


development and distribution of the present-day 


Andean uplift played a major role in the 


northern Andean flora. Numerous neotropical lowland 


taxa adapted to the cool climatic conditions at high 


elevation. Additionally, the Andes served as a route 
for temperate taxa to migrate from temperate latitudes 


to cool tropical montane areas. The result of these two 


phenomena can be seen in the present-day composi- 
tion of the flora of the Colombian montane forest and 
páramo. Figure 9 supports this view as it shows the 


representation of geographic elements of woody 


genera for five different main vegetation belts in 


northern Andean and adjacent lowland vegetation. At 


— 


high elevations, temperate taxa are abundant while at 
low elevations neotropical and pantropical taxa are 
abundant. A minority of Holarctic elements appear 


below the Andean forest belt. 


DISCUSSION AND SYNTHESIS 


During the late Miocene (Biozone I). the area of the 
present-day Bogotá basin was a lowland riverine 
environment (ca. 700 m elevation) dominated by 
tropical lowland taxa which, at present, are found in 
the modern várzea forest of the Amazon basin. This 


“pre-uplift” vegetation. principally consisted of taxa 
with a wide tropical or neotropical origin (e.g.. 


Bombacaceae, Cyclanthaceae, Humiriaceae, /riartea 
Ruiz € Pav.. Mauritia L. f., Melastomataceae, and 
Parinari Aubl.). Amazonian and Andean-centered 


Trees of Podocarpaceae, 
Clethra L., and 


families were predominant. 
accompanied by Symplocos Jacq.. 
probably Weinmannia L., apparently were constitu- 
ents of the Neogene montane forests found on elevated 
areas near the study area. 

By the late Miocene to early Pliocene (Biozone ILA), 
the proportion of tropical lowland taxa had decreased 
from 75% (Biozone I) to ca. 60%. The share of taxa 
from present-day Andean and subandean vegetation 
belts increased from 5% (Biozone I) to 20%. During 
HA. the 
increased from 10% to ca. 23%. 


Biozone proportion of Laurasian families 
These changes in the 
floral composition show a transition from tropical 
to subandean forest and suggest the local 

the 1000 m 


Some such as 


lowland 


uplift of Eastern Cordillera to ca. 


elevation. tropical lowland taxa. 
persisted. in 


1000 m 


A number of present-day montane taxa, 


Humiriaceae, Mauritia, and Parinari, 


areas, such as valleys, that were still below 
elevation. 
including those of neotropical origin (e.g., Billia Peyr., 
ex Willd., 


Forst.), appeared for the first time in the study area, 


Freziera Sw. Myrsine L., and Xylosma G. 
indicating a more diverse subandean and apparently 
Andean forest. 

During the middle Pliocene (Biozone HB). the floral 
composition was still characterized by a dominance of 
taxa with a tropical affinity, whereas only 25% of the 
taxa had either tropical amphipacific, wide temperate, 
Austral-Antarctic, Holarctic, or Andean neotropical 
affinities. Based on the affinity of their taxa, the floras 
of Biozones HA and IB are very similar. Alternative- 


ly, a comparison between floral composition has been 


Annals of the 
Missouri Botanical Garden 


made on the basis of plant affinity to present-day 
Around 50%-60% of the 


Pliocene taxa also belong to present-day subandean 


altitudinal vegetation belts. 


and Andean vegetation belts. In contrast to a compar- 


based on phytog hi 


ison composition, the 
difference in floral coms of between Biozones 
IA and HB is now significant. 
the vegetation suggest that the area had reached an 
altitude of ca. 2000 m. Prominent taxa in the montane 
forest included trees of Podocarpaceae, as well 

Ilex L., 
L. Taxa that first appeared in the studied sequence are 
Asteraceae (Liguliflorae), Boc- 


conia Plum ex L., Borreria G. May, Calceolaria L., 


Hedyosmum Sw., Myrsine L., and Weinmannia 


Alternanthera Forssk., 


Callitriche L., Daphnopsis Mart., Gaiadendron G. 
Don., 


Aubl., Sericotheca Rafin., 


Elatine L., Montia L., Mye La e 
Vallea Multis ex L. f., 
Xyris L. The majority of these taxa are 1 n 
the present-day Andean forest and páramo vegetation 
(van der Hammen & Cleef, 1986; Cleef et al., 1993). 
During the late Pliocene (Biozone III), continued 
uplift increased elevation to ca. 2600 m. About 40% 
of the fossil flora taxa have a tropical affinity and 35% 
have a temperate affinity. The Amazonian-centered 


Andean-centered and Laurasian 
elements contribute 18% and 60%, 


The 


Hedyosmum Sw., 


families and the 
respectively, to 
the  paleoflora. main forest taxa, including 
Podocar- 
Myrica 
are recorded for the 
of the late 
Andean forest differs from that of the late Quaternary 
Notable is the absence of Alnus Mill., 


Juglans L., and Quercus L., which immigrated into the 


Alchornea Sw., Myrica L., 
paceae; Weinmannia L., Hesperomeles Lindl., 
L., Rubus L., and Turpinia Vent., 
first time. The composition Pliocene 


Andean forest. 
during the Pleistocene. The contribution of 


Myrica, the 


Andean vegetation cover was apparently higher in 


area 


and 


© 


Hedyosmum, Podocarpaceae t 
the late Pliocene than in the Quaternary (Wijninga & 
Kuhry, 1993). It is possible that the late Pliocene 
forest had a more open character locally (in the area of 
the high plain of Bogotá), perhaps related to a dynamic 
river valley environment (Hooghiemstra, 1984; Torres 

al., 2005). 
characteristics of this early Andean forest are already 


Biozone IIB) 


To some extent, the above-mentioned 


apparent during the middle Pliocene 


— 


and prevail until the early Pleistocene. 

During the Quaternary, repeated. climate change 
forced main vegetation belts to migrate altitudinally 
along mountain slopes. Combining the palynological 
information from ecotone forests (e.g., van der Ham- 
men, 1974; vant Veer & Hooghiemstra, 2000) with 
2001), 


the subandean forest belt 


the dynamics at lower elevations (Wille et al., 
that 
maximally squeezed, during glacial periods in partic- 


1300 m 


we concluded was 


ular, from the present-day ca. vertical 


The characteristics of 


e Although in some parts of the | 


extension (ca. 1000-2300 m) to 600 m vertical 
Last Glacial Maximum (LGM) 
(ca. 800-1400 m). Other vegetation belts lost less of 
expansion during the LGM: tropical 
lowland forest lost 200 m (0-800 m LGM expansion) 
and Andean forest lost 300 m (1400-2000 m LGM 
expansion). The páramo belt kept the same altitudinal 
extension (2000-3000 m) during the LGM as today, 


but this belt in particular gained much surface when it 


extension during the 


their vertical 


— 


moved to lower elevations. In consequence, Quater- 
nary dynamies were highest in the páramo belt. This 


reconstruction is consistent with the observation that 


— 


the high mountain flora, in particular, reached a high 
level of diversity. In the Espeletiinae group, highest 
diversity occurs in the area where the uplift history is 
youngest (Cuatrecasas, 1979), suggesting that in the 
northern Andes the dynamic Quaternary environment 
has contributed positively to the present-day high 
level of plant diversity. 


CONCLUSIONS 
* Differences in the composition of fossil plant associations 
recovered from sediments in the area of the Bogotá basin 
form a series from middle Miocene to late Pliocene and 
reflect changing climatic conditions that result gend 
from the Pliocene tectonic uplift of the Easter 
* Global temperature change during the Neogene, estimat- 
ed at 3 C, may explain at most one third of the difference 
in inferred paleoaltitude, and hence the difference in 
floristics between the oldest and youngest fossil plant 
assemblages. 
e The fossil plant assemblages show that the late Miocene 
vegetation was dominated by taxa with a wide tropical or 
neotropical distribution. Early montane forest seems to 
aave been relatively rich in Podoc jdn eae. Dur 


Pliocene, the diversification of 


2 
NE 

o 

2 


increased. Taxa 


^ 
o 
B 
A 
© 
i 
e 
< 
= 
D 


IS val a afic, c 
the div 
span, e 1 r of 


of the montane forest. P the same time 
1 


azonian-centered families de- 
creased in favor of families of Laurasian origin 
ipi A reconstructed forest types are non- ales to 
ain extent an be compar d to their modern 
13 


aking int 


the absence of ta 


elena 
al immigrated into the 1 area at a later time. 
Eastern C ordille: ra uplift 
in the middle the area of 
Bogotá remained below 700 m altitude until the 


Miocene. 


had already started Miocene, 


ale 


middle 


Holarctic and 


Pliocene onward, the number of 
Asia-derived (e.g., N For- 
man, Symplocos Jacq.) taxa increased.“ ndicates the 
route via the 


Migration 


importance of the northern iua 


Panamanian isthmus for plant taxa. from 


southern South America possibly re before the 
ral-Antarctic 
axa arrived in the northern Andean montane rain forest 


late caer ne. A substantial number of Aus 


biome before immigration from the Northern Hemisphere 
took place, but in the present-day flora there are fewer 


immigrants from the south than from the north. 


Volume 93, Number 2 
2006 


Hooghiemstra et 
Paleobotanical 1 1 Of Colombia 


[09] 
= 
e 


e Based on the studied fossil plant assemblages, the first 


records of páramo-like vegetation are from the late 


liocene, confirming earlier conclusions by van der 
Ta 


Hammen et al. (1973). 


— 


Mainly during the Quaternary; climate change forced 
s lt rir 


lencia extension aud 1 . again. 
causing populations to be jeatedly isolated and 
remerged. In the northern pons s, high diversity 
correlates strongly with a dynamic environment in which 
migration nn clon played key ro 


Literature Cited 


Andriessen, P. A. M., K. F. Helmens, H. Hooghiemstra, P. A. 
Riezebos & T. van der Hammen. 1993. Absolute 
chronology of the Pliocene- Qu aternary sediment sequence 


ev. 12 


mu 


Bogotá area, Colombia. Quatern. Sci. 
483-501. 


Behling, H. 1996. First report on new evidence for the 


"E 


occurrence of Podocarpus and possible human presence at 
the mouth of the Amazon during the Late-glacial. Veg. 
i : 241-246 


Cleef, A. M. 1979. The TE posilion of the 
neotropical vasct ‘a paramo flora with special reference to 
the Colombian Cordillera Oriental. Pp. 175-184 in 
Larsen & L. B. Holm-Nelsen (editors), Tropical Botany. 
Academic zh ‘ss, Lone 

& H. Hohen 984. Present vegetation of 108 
ea of the hig ain of Bogotá. Pp. 42-66 in 
Hooghiemstra (edito " Vegetational and ee oe 

Bot. 


of the High Plain of Bogota. Diss. 

). Rangel & S. Salamanca. 1983. Reconoe 'imiento 
de la vegetación de la parte alta del transecto Parque 
Los Nevados. Pp. 1 in T. van 2 1 
A. Perez Preciado & O. Tinto Escobar (editors), La 
Cordillera Central Colombiana Transecto Parque Los 
Nevados. Studies on Tropical Andean Ecosystems 
Cramer, Vaduz. 

. T. van der Hammen & H. Hooghiemstra. 1993. The 
savanna relationship in the Andean páramo flora. Opera 
Bot. 121: 285-290 

x 0. "angel & S. Salamanca. 2003. The Andean 
rain forests of the Páramo Los Nevados Transect, 
Cordillera Central Colombia. Pp. 7 : 
Hammen & A. G. Dos Santos (editors), La Cordillera 
Central Colombiana Transecto Parque los Nevados 


| 
— 
aS 
— 
S 
— 
e 
3 
a 


Studies of Tropical o Ecosytems 5. Cramer- 
Borntraeger, are Stut 
Cuatrecasas, 


— 


Oh iones geobotánicas en 
Colombia. Trab. 19 1 Nac. Ci. Nat., Ser. Bot. 27: 1-144. 
——. 1058. Aspectos de la vegetación natural de 

ea Revista Acad. Colomb. Ci. Exact. 10(40): 
221-264 


A n: 9. Comparación fitogeográfica de páramos entre 
varias Cordilleras. Pp. 89—99 in M. L. Salgado-Labouriau 
D El Medio Ambiente Páramo. Ediciones Centro de 

tudios Avanzados, Mérida 

e) A. 1941. Estadios. geobotánicos Colombianos: 
Descripción de una sinecia típica en la sub xerofitía 
del litoral Caribe. Rev. Acad. Colomb. Ci. Exact. 
135-141. 

Duque Caro, H. 1990. Neogene stratigraphy, paleo-ocean- 
ography and paleobiogeography in northwest South 
America and the evolution of the Panama Seaway. 
Palaeogeogr. Palaeoclimatol. Palaeoecol. 77: 203-234. 


Fabre, A. 1983. La subsidencia de la Cuenca de Cocuy 
(Cordillera Oriental de Colombia) durante el Cretáceo y el 
Tertiario. Segunda parte: Esquema de evolución tectónica. 
Geol. Norandina 8: 21-27. 

Flynn, J. J., J. Guerreo & C. C. Swisher. 1994. Geo- 
edm. of the Honda Group (Middle Miocene), 

Colombia. In R. F. Kay, R. Madden, R. Cifelli & J. Flynn 
(editors), A Histo a 1 1 vical Fauna: Vertebrate 
Paleobiology of the Miocene of Tropical South America. 
Smithsonian Institution Press, Washington, DC 

Gentry, A. H. 1982. Neotropical floristic diversity: Phyt 
geographical connections between Central and South 
America, Pleistocene climatic fluctuations, or an acc m 
of the Andean orogeny? Ann. Missouri Bot. Gard. 

5 


. 1986. Species richness and floristic composition of 
Chocó region plant communities. Caldasia 40: 71-91. 

Graham, A. 1992. Utilization of the isthmian land bridge 
during the Cenozoic 


paleobotanical evidence for timing. 
and the selective inf duis of altitudes and climate. Rev 
Palaeobot. Palynol. 72: 119-128. 
Grubb, P. J. 1971. 3 of the Mae e 
effect on tropical mountains. Nature 229: / 5. 
. 1974 


E 


Factors controlling the ebia of forest 
types on tropical mountains: New facts and a new 
perspective. Pp. 13—45 in J. Flenley (editor). Altitudinal 
Zonation in Malesia. Transactions Third. Aberdeen-Hull 

Symposium on T ln n Ecology, Dept. of Geography. 
Univ. a Til. Hull. 

. 1977. Control of forest growth and distribution on 
wet tropic ical mountains: With s speci rw reference to mineral 
nutrition. Annual a Ecol. Syst. 8 33-107 

oyd & T. D. anite a )3. A comparison 
of montane and rer rain forest in Ecuador. I. T 
forest structure, physiognomy, and floristics. J. Ecc A 5l: 
567—001. 

Guerrero, J. 1993. Magnetostratigarphy of the Upper Part of 
the Honda Group and Neiva Formation. Miocene Uplift of 
the Colombian Andes. Ph.D. Dissertation, Duke Univer- 
sity, Durham, North Carolina. 


= 


nia: H. J. 1962. Pa 5 development of 
uth America. Bull. Amer. Assoc. Petrol. Geol. 46: 

eds 14 
———— & T. van der Hammen. 1994. The Pliocene- 
Quaternary of the high plain of Bogotá (Colombia): 
history of tectonic aptis basin development and climatic 
Quatern. Int. 


2 


change. 
Howie mstra, H. 1984. Vespa a climatic history of 
high aa of Bogotá. Diss. Be 
19 


d and Upper : lioc ‘ene glaciations 


and Tu dn elopment in the tropical Andes: Evidence 


rom a Oe iigh- 5 pollen record from the 

sedimentary basin o gola, mae Palaeogeogr. 
Palaeoclimatol. Palae oec B 725 

& A. M. Cleef. 1995 be climatic change 

and environmental and generic dynamics in the north 


van ph T unm 1998. Neogene and 
Quaternary ju d of the neotropical rain forest: 
The forest 0 ypothesis, and a literature overview. 
-Sci. Rev. 4 A | 47— 
— 2004. Quatemar ice-age dynamics in 
Colombian 19 Developir E 990 understanding of 
our legacy. Philos. Trans., Ser. 5 3 17 


Annals of the 
Missouri Botanical Garden 


Hoorn, €. 199; 
Andean tectonies on the Miocene depositional history of 
northwestern Amazonia: Results of a palynostratigraphic 
study. Palaeogeogr. Palaeoclimatol. Palaeoecol. 105: 
T 
. Fluvial y i ts in tl 
Bas in (Early 
ed Mia). 


1e intracra- 
Middle 


Palaeoclimatol. Pa- 


tonic ieee Mince 'ene-early 


Miocene, Palacogeogr. 
laeoecol. 109: 1-55. 

9 reconstruction. of the 
Tiddle to Late Miocene, NW 


Palaeoecol. 112: 


env ironmental 


— 


a 
paleo-Amazon River system (V 
Amazonia ). Palaeogeogr. Palaeoclimatol. 
l 

A. Sarmiento & M. X. 


se for changing drainage 


JUerero, Lorente. 


1 8 M. ade 
1995 een tectonics as a caus 

te rns in Miocene northern South America. Geology 23: 
237-240 

p 0. dia 1986. La Selva Nublada 
Grande, Parque Nacional Henri Pittier. 


de Rancho 
Fondo Editorial 

\cta Científica Vene ed Caracas. 
Kappelle, M., A. M. Cleef & 1992. 


raphy of Talamancan montane Quercus forest. Costa Rica 
J. T 4 315 


. Chaverri. Phytogeog- 
Bioge ogr. m TE 
Keigwin, L. D. Plioce 

Panama. base n on a evidence from nearby 


ene closing of the Isthmus of 


Pacific Ocean and Caribbean Sea cores. Geology 6: 
630-634. 

. 1982. Neogene planktonic foraminifera from Deep- 
Sea Ds Project sites 502 and 503. Pp. 269-288 in W. 
L. Prell € J. V. 
Deep-Sea Drilling Project 68. U.S. Gov. 
Washington. D.C. 
J. 2000. Vascular Plant Family C a. of the 
Colombia: Divers Patterns, 
Ph.D. ou Fm Univer- 


Gardner (editors), Initial Report of the 


Print. Office. 


Cordillera of 
co Me and Ordination. 
sity of Amsterdam. 

Kennett, J. P. 1995. 
during the Neogene, based on the marine sediment record, 
Pp. 49-64 in E. S. Vrba, C. Denton. L. H. Bui 
Partridge 


V review of polar climatic evolution 


ekle & T. C. 


(editors). Palaeoclimate and Evolution with 
Emphasis on Human Origins. Yale Univ. Press, N 
Haven. Connecticut 


Kroonenberg, S. B... . Bakker & A. M. 


1990. Late ON MS and paleogeographs 


van der Wiel. 
of the 
Colombian Andes: 5 on the development of 
high-Andean biota. Geol. Mijnb. 69: 270-200, 
K. 1990. 
biostratigraphy ane ee oenvironments. 
Helmens (editor), Neogene-Quaternary 880195 of the 
High Plain of ae 


Paleoenvironments and 
163 


Ruhr. P. & F. Helmens. Neogene-Quaternary 
K. 


. 89-132 in 


častern Cordillera, Colombia: 


Landscape Evolu- 


1981. 


la ate Miocene calcareous nanno- 


Lohman, G. P. & J. Carlson. Oceanographic 
Pacifi 


plankton. Mar. Vien 6: 553-579. 
te, M. A. 1986. 


significance of 


Loren 1 "alvnology and Palvnofacies of the 
Upper Tertiary in 1 Ph.D. Dissertation, Univer- 
sity of Amsterdam, Diss. Bot. 9 

Luteyn, J. L. 1999, Páramos: A checklist of plant diversity. 


vu ‘al distribution. and botanical literature. Mem. 
New York Bot. Gard. 84. 
Mabberley. D. J. 1993. The 


Rien of the Higher Plants. 


Plant-Book: A 
Univ. 


Portable 
Cambridge Press, 
ened ge. 
m . W. 
to estimating paleoaltitudes from ge floras 
Palaeoecol. 99: 71— 


1992. Lapse rates and other variables applied 
;. Palae 08e ogr. 
Palaeoclimatol. 


Marine incursions and the influence of 


J. 0. 


Andino de Colombia. Pp. 
N 


Rangel, 1995. La diversidad Er en el P 
187-205 in S. P Churchill, 
Balslev & J. L. Luteyn (editors dee e 
Conservation of Neotropical Montane Forests. Mem. New 
York Bot. Gard. 78. 

(editor). 2004. 


Chocó Biogeográfico/Costa Pacifica. Universidad Nacional 


Colombia Diversidad Biotica IV. El 
de ( D Conservation International, Bogotá. 
& R. Jaramillo. 1984. 


he rhorizado en el Transecto Buritaca-La Cum mbr e. Pp. 155- 
1 


Lista comentada del material 


176 in T. van der Hammen & P. 


Nevada de 


Buriotaca—=La 


1 75 (C ie ra 
Cumbre. Studies o ypical 
a 2. Crame r Bormiraeger, b 
Cleef & S. Salamane ‘a. 2003. 
inte 1 and sub: indea a 
Nevados I 
204 in T. van 


a age a Central C olombiana TERUR? cto. Parque los 
Ti 


Sierra Santa 


The « quatorial 


der Hammen é Jos Santos (editors). 


Nevados. Studies of Tropical Andean Ecosytems 5. 


Berli sere 
Molecular 


Espe letia complex ( 3 evidence from nrDNA 


Crame 5 T. 
J. ) 


on her. T. 2002 phylogenetics of the 


ITS sequences on the closest id ves of an Andean 
adaptive radiation. Amer. J. Bot. a 1-108 
haven, P. H. & D. Axelrod is l. 81 rm bio- 


UC and past continental movements. Ann. Missouri 
Bot. Gard. 61: 539-67: 

Savin, S. M. 
temperature during the last 100 million years. Annual 
Rev. s anet. Sci. 5: 319-3 

. Douglas & F. > Stehli. 1975 ddr marine 
ie ratures. Geol. Soc. Amer. Bull. 86: 1494—1510. 

Shackleton, N. J. 1982. The den sea di nb record. « 
climate variability, Oceanog. 11: 218 

1995. New data on the 
E variability. va 242-248 
L. H. Bure 


and Evolution with eue on Human Origins. Yale 


1977. The history of the Eartl's surface 


Progr. 
1 5 on . 
Vrba. G. Denton. 
cle & T. C. Part erm le a. Palaeoclimate 
Univ. 995 cw Haven. Connecticut. 
Simpson. D. B. 1975. Pleistocene changes in the flora of the 


high b Andes. Paleobiology 1: 273-294. 


& C. A. Todzia. 1990. Patterns and processes in the 


development of the high Andean flora. Amer. J. Bot. 77: 
1419 : 132 

Smith. D. L. 198 5. c an ne relative motions 
18 x F. C. Stehli S. D We (editors The ( 
American Biotic Plenum, New NE 


Sokal. R. R. & 


Ney Vork 


) ¿real 
1 


. J. Rohlf. 198 


. Biometry, 2nd ed. Freeman. 


Steyermark. J. A. 1975. Flora de la Sierra de San Luís 


(Eastado Falcón. Venezue i y sus afinidades fitogeográ- 


licas. Acta Bot. Ven. 10: 131-218. 
Tiedemann. R.. M. Exin & N. J. Shackleton. 1994. 
Astronomic timescale for the Pliocene Atlantic C70 and 


dust influx — of ren Drilling Program Site 659, 
Paleoceanography 9: 619-6; 


Torres. MV. 2006 


F Lora Vegetation. and 


3 Evolution 
C limate: A 
Se dus 5 cal Study of a 580 m core from the Bogota 


Basin. Ph.D. 


sterdam. 


Palvnological and 


Colombia. Dissertation, University of 
andenberghe & H. Hooghiemstra. 2005. An 


environme le reconstruction of the se dime nb infill of the 


Bogotá basin (Colombia) during the last 3 million years 


from abiotic and _ proxies. Palaeogeogr. Palaeocli- 
| 


matol. Palaeoecol. 226: 127-148 


Volume 93, Number 2 
2006 


Hooghiemstra et al. 
Paleobotanical Record of Colombia 


321 


Ulloa Ulloa, C. & . Jorgensen. 1993. Arboles y arbustos 
de los Andes de d ee AAU Rep. 30: 1-264. 

van der Ha 1901. 
stralig pi and lec Bis nesis of the Colombian Andes. 
Geol. Me 40: 181- 


The Pleistoc 


mmen, T. Late Cre e eous and Tertiary 


ene changes of vege lation and 
2€ 


us in in top South America. J. Biogeog 26. 
. 1989. History of 10 montane forest of A SE eA 
Pl. Sy Evol. 162: 109-114. 


1 Z. 


P 


Historia de clima and 
vegetación de | Pleistoceno o Superior v del Holoceno de la 


bs 11: 189-266. 


Sabana de Bogotá. Bol. 
& 1 a E vollen 


diagram from the 
Quaternary of die Sabana de Bogotá (Colombia) and its 
significance for the geology of the Northern Andes. Geol. 


Mijnb. 43: 113-117. 

& A. Cleef. 1983. Trigonobalanus and the 
tropical amplis pacific element in the Northern Andean 
orest. J. Biogeogr. 4374 

& 1986. De idus nt of the high un "an 


Vuilleu- 

Monasterio (editors), High Altitude Tropical 

Biogeography. Oxford Univ. 
& 


páramo flora and vegetation. Pp. 153-201 in F. 
mier & M. 
ress, New York 


FA 


Hooghiemstra. 1995. Cronoestratigrafia 


— 


correlación del Plioceno y Cuaternario de Colombia. Ann. 
Geogr. Bogota) 24: 51—67. 

. 1997. Chronostratigraphy and correlation 
of dh m and Quaternary of Colombia. Quatern. Int. 


40: 81-6 


Werner & H. van Dommelen. 1973. 
Palynologic al rec E of the upheaval of the Northern 
Andes: A study of the Pliocene and Lower Quaternary of 


5 Colombian K: astern C ion ra and the earl 


v evolution 
of its high Andean biota. . Palaeobot. Palynol. 16: 


1-122. 
van der Wiel. A. M. 1991, Uplift and Volcanism of the SI 
Colombian Andes in Relation to Neogene Sedimentation in 


the Upper Magdalena Valley. Ph.D. Thesis; University of 


van der Hammen. 1973. Upper Quaternary 


vegetational and climatic sequence of the Fuquene area 


S ¿astern C as ra. : olombia). Palaeogeogr. Palaeoclima- 
-02. 


. Palaeoecol. 


van Houten, F. B. & R. B. Travis. 1968. Cenozoic deposits, 
Magdalena AAPG Bull. 52: 


upper Valley, Colombia. 


75702. 


vant Veer, R. & H. Hooghiemstra. 2000. Montane 

duris the last 650,000 yr in Color 

multivariate ee hi based on pollen record Funza- P 

Quatern . Sci. 15: 329—346 

W^ TROPIC OS. ver. U. 5 
W3T/Search/vast. 1 55 

Wijninga. V. M. 19962 


Tequendama site " 75 m alt., 


forest 
evolution A 


— 
Y 


)06. E 1 mobot.org/ 
accessed 18 May 2006. 

. Neogene ous ^ the Salto de 
Cordillera Oriental. 
The paleobotanical 


lowland forests. Rev. 


Colombia): record of apes and 

Palaeobot. Pt )2: 97-156. 

6b. A Pliocene Podocarpus forest mire from the 
H 


llera Oriental. C 11 9 5 


high plain of N ies li ev 
Palaeobot. Palynol. 92: 157—205 
—— ——,. 1996c. B. and plot: of the early 


dg ra Orie nak, 


Frío 17 (( 
Biostratigraphical and 
9 92: 329- 350. 
m and Palynology of Neogene 
Bogota (Colombia): 
Andean Flora from a Paleoecological 


9 section Río 


implications. Rev. Palaeobot. 
1996d. 
Sediments from the High Plain of 
Evolution of the 
Perspec tive. Ph.D. Thesis, University of Amsterdam. 
P. Kuhry. 1990. A Pliocene flora 
Subac ous Valley (Cordillera Oriental, Colombia). Rev 
62: 249—290 
& . 1993. Late Pliocene paleoecology of the 
Guasca Valley Oriental. Colombia). Rev. 
Pu Punt 78: 69-121. 
.H. 


Wi à 


from the 
Palaeobot. Palynol. 
(Cordillera 


Hooghiemstra, H. Behling. K. van der Borg & 


: Ne grel. 2001. Environmental change in "tlie 
C na subande an forest belt from 8 pollen records: 
Archaeobot. 10: 61-77 

Past Vegetation and 


Northern 
A Quantative Approach. 


(Cordillera Central, 
Ph.D. 


Climate in ac Andes 
l Dissertation, 
Universily of Amsterdam. 


Wortel. M. J. R. 1984. Spatial and temporal variations in the 
Andean subduction zone. J. Geol. Soc. 141: 783—191. 
& S. Cloetingh. 1981. On the origin of the Cocos- 


Nazca spreading center. Geology 9: 425—430. 


322 Annals of the 
Missouri Botanical Garden 


Appendix I. cklist of identified polle n laxa, their phytogeographic relationships, and their absence/presence in the 
studied pollen "i macrofossil floras. Sites: S. de T. III = Salto de 9 . MII (middle Late Miocene); Río F 17 = Río 


Frio- 17 (early Pliocene): Subach. 39 = Subachoque-39 (early Pliocene): Faca. 13 = Facatativá-13 (middle Pliocene); Guasca 
103 (late Pliocene). Modern phytogeographic units: | = wide tropical: » = tropical Africa-America; 3 = neotropical; 4 = 


tropical amphipacific; 5 = wide temperate; 6 = austral-antarcetie; 7 = holaretic: 8 = andean neotropical; f 


cosmopolitan. 


cf. genus name = sp. ined., cf. genus name). Appendix modified after Wijninga (1996d). 


— 


Taxon Phyt. unit S.de T. I/II Rio F 17 Subach. 39 Faca. 13 Guasca 103 

Acacia Miller l X 
Acalypha L. | X X 
Acanthaceae l X * 
Alchornea Sw. l x X x x x 
Alismataceae 9 x x x 
Amanoa Aubl. 3 x x x 
Anacardiaceae l X 
Anacardium L. cf. 3 X 
Annonaceae l x 

Anthurium Schott. 3 x 
Antidaphne Poeppig & Endl. 3 x x 
Apiaceae 9 X X x 
Apocynaceae l x x x 
Araliaceae l X 

Arecaceae (Palmae) l x X x x 
Arte LL. 7 x 

Asplundia Harling cf. 3 x 

Asteraceae - liguliflorae 9 x x 
Asteraceae - tubuliflorae 9 x x x x x 
Astrocaryum G. Meyer 3 x x 

Azolla Lam. 9 * * 
5 Jacq. Ex Scop. 3 x 

Bauhinia L. l X 

Bernouillia Oliver 3 X 
Bignoniaceae l X 
Billia Peyi 3 X x 
Bocconia L 4 X 
Bomarea Mirbel 8 x 
Bombacaceae l X x X x 
Bombacopsis Pittier 3 x 

orreria G. Meyer 6 x 
Bromeliaceae 3 x 
Brownea Jac : 3 x 

ursera Jacq. ex L. 3 * 
Byrsonima Ric h. ex Kunth 3 x 
Byttneria Loefl. l x 

Calceolaria L. l * 
Callitriche L. 5 x x 
Carex L. 5 x x 
Catopsis Griseb. 3 x 
Catostemma Benth. 3 x 
Cecropia Loefl. 3 x x X x x 
Cedrela P. Browne 3 x X x 
Ceiba Miller l x 

veltis | 7 x 
Chrysobalanaceae cf. | x 
Clethra | 4 x x x 
Clusia L. 3 x x x x 
Clusiaceae s.l. (Guttiferae) 9 x x 
Convolvulaceae 9 x 
Connarus L. ¢ l x 
Cordia L. | X 
Coussapoa Aubl. 1 x 
Crassoretitriletes l x 
Croton L. l x x * 


Volume 93, Number 2 
2006 


Hooghiemstra et al. 


Paleobotanical Record of Colombia 


323 


Appendix l. Continued. 

Taxon Phyt. unit S.de T. MII Rio F 17 Subach. 39 Faca. 13 Guasca 103 
Cyatheaceae 1 x x x x x 
Cydista Miers 3 x 
Cyperaceae 9 x x x x x 
1 Bernh. 7 x 
Dalec al. ] x 
E C. Martius 3 x 

ecussoc 1 1 aubenf. 1 x 
Dicksonia UH 4 x 
Duroia L.f 3 x 
Ecclinusa C. Martius 3 x 
Eleocharis 9 x 
Ericaceae 9 X * 
Eugenia l x x * 
Euphorbiaceae cf. 9 x x x 
Evodianthus Oersted. cf. 3 x x x 
Fabaceae 9 * x * 
Faramea Aubl. à x x 
"icus L. ] x * 
Freziera Willd. 8 x x X 
Fuchsia L. 6 x 
Gaiade d . Don f. 8 x 
Guarea Allam ex L. 2 * * 
Gunn 6 x x 
Hedyosmu 4 * x * x x 
Hesperomeles Lindl 8 x 
Hippocratea L. /Pristimera Miers l x 
Humiriastrum (Urban) Cuatrec. 3 x x 
Iveronima Allemáo 3 x x * * 
Hypericum L. 5 * x X 
Ilex L. 9 x x x x * 
Ipomoea L. 1 x 
Triartea Ruiz & Pavón 3 x x x X 
Iridaceae 9 * 
Juncaceae s.l 9 X 
Juncus L 5 x x X x 
Lathyrus I 7 x 
Lepidium L 9 x 
Loranthaceae 9 x x x 
Ludwigia 9 x x x x 
W L 9 X x x x 
Lythraceae 9 x x 
Mabea 3 x 
Machaerium Pers. cf 3 x 
Macoubea Aubl. cf 2 X * * 
Macrocarpea (Griseb.) Gilg 8 x 
Macrolobium Schreber 3 * * * * 
Malpighiaceae 1 x x x x 
Malvaceae 9 * 
Mauritia Kunth 3 x x * 
Melastomataceae l x x X x x 
eliaceae l x * * 
1 Ruiz & Pavón 3 x 
Montia 5 x x 
Muchlenbechi Meissner 6 * 
Myr 7 X x 
My nm L. 9 x 
Myrsine L 1 x x x 
Ayr e ] * x X x x 
Oreopanax Decne. & Planchon 3 x x 
Paraprotium-type Cuatrec. 3 x 


324 Annals of the 
Missouri Botanical Garden 


Appendix 1. Continued. 


Taxon Phyt. unit S.de T. IH Rio F 17 Subach. 39 "aca. 13 Guasca 103 

Parinari Aubl. 1 x x 
Passiflora L. 3 x 

Piperaceae l x 
Pityrogramma Link 2 x 

Poaceae 9 x X X x x 
Podocarpaceae 6 x x x 
Podocarpus L'Hér. ex Pers. 6 x x N N x 
Polygonum L. 5 x x x 
Pouteria Aubl. l X x 
Protium Burm.f. l x x 

Pseudobombax pS 2 x 
Psychotria L. l X 
Quararibea e 3 x 
Ranunculus 5 X 
theedia L. 2 * 
Rhus L. 7 x X 
Rubiaceae 9 x X X x 
Ru 5 N 
Rumex L. 5 X 

Sacoglottis C. Martius 9 x x x 

Sagittaria L. 9 x 
Salix L. 7 X X 
Salvia L. n X x 

Salvinia Séeuier l x 

Sapindaceae l N x 
Sapium P. Browne. 3 x x 
Sapolaceae l x 

sti Forster & Forster f. j x x x x 
Scirpus L. cf. 9 X 
Selaginella Pal. 9 x x x 
Sericotheca Rafin. 3 X 
Solanaceae 9 x x x 
Spathiphyllum Schou 3 x x 
Sphagnum l. 9 X 
Symplocos Jacq | x x x X X 
Tapirira Aubl ) x 
Thelypteris Sehmidel 5 X 

Thoracoc une Harling cf. 3 x 

Tilia 9 x 

Tournefortia L. l X 
Trichantera Kunth / Bravaisia D.C. 3 X 
Tristerix Mart. 3 X X 
Turpinia Vent. I x 
Typha L. 9 x x 
Urticaceae 9 x x 
Urticales 9 x x x x x 
Valeriana L. 5 x x 

Vallea 11 055 ex L. 8 X 
Vantanea Aubl. 3 N x 
Viburnum U. 7 x X 
Virola Aubl. 3 x 
Vismia Vand. 3 x 
Vochysia Aubl 3 X 

Varszewiczia Klotzsch 3 X X x 
Weinmannia L 6 X x x 
Xylosma Forster f. | x 
Xyris L. | x x 
Zanthoxylum L. l x x 


PALEOBOTANICAL EVIDENCE Alan Graham? 
AND MOLECULAR DATA IN 
RECONSTRUCTING THE 

HISTORICAL PHYTOGEOGRAPH Y 

OF RHIZOPHORACEAE’ 


ABSTRACT 


The fossil record of the Rhizophoraceae includes numerous questionable identifications from strata of unconfirmed age. 
together with others representing clear evidence of the 1 family in deposits whose age is well constrained by independent or 
multiple lines of evidence. Bruguiera or the Bruguiera lineage. and Ceriops are known from the early Eocene, Kande lia from 

l 


the middle Eocene, and We are from the late Eocene. Combretocarpus is likely present by the middle Miocene, bu 


belongs to a family (A pl ) earlier placed in the Rhizophoraceae that is now considered unre slated to that family. 
Rhizophora, as a genus, has ‘existed essentially uninterruptedly in the Caribbean region since the late Eocene, but cooling 
events at ca. 15-14 Ma, ca. 3.4 Ma, and ca. 1.6 Ma, and the subsequent 18 to 20 glacial intervals provide a mec hanism for 
development of the e zn T (e 0 tele rant populations toward the northern part of its range i 1 the New World. In reconstructing the 


biogeographic history of Rhizophora in the Caribbean region. it is worthwhile noting Ds molecular-defined lineages (with little 
morphological or palynological variation) may have been introduced and gone extinct on multiple occasions. This allows for 
variation in molecular patterns and apparent stasis in the fossil record—a pos ssibility not widely considered in the literature. 
The most recent molecular-defined common ancestor to eats populations is estimated to have arrived in the New World ca. 
11 Ma. using an origin in the eae ocene (ca. 60 Ma) as the basis for calculation 
| IL | 
| phy, Rhizoy : 


Key words: fossils, historic al | 


RESUMEN 


El registro fósil de Rhizophoraceae inc luye numerosas identificaciones cue sstionables de estratos sin edad confirmada, junto 
| j 
familia se encuentra en depósitos cuya edad está forzada por líneas de 


con otros que representan clara e videncia de que la fi 
emprano, Aandelia 


evidencias independientes o múltiples. Bruguiera o el linaje Bruguiera, y Cer riops se conocen del Eoceno 1 
del Eoceno medio y Rhizophora del Boc “eno DEAD Combretoc ‘carpus está presente probablemente a partir del Mioceno me «dio, 


pero pertenece a una familia (A ) ubicada anteriormente en R 1 que ahora se considera sin 
relación a esa familia. Rhizophora, como género, ha e a esencialme : in n ión en la región del Caribe desde el 
Eoceno tardío. pero eventos de enfriamiento hace ca. 15-1 a. 3.4 Ma v ca. 1.6 Ma, y los subsecuentes 18 a 20 


intervalos glaciales proporcionan un mecanismo para el de 5 ES AUR iones e rantes al frío hacia la p norte de s 
distribución en el Nuevo Mundo. Al reconstruir la historia biogeográfica de Rhizophora en la región del Caribe. 
(con poca variación morfológica o palinológica) pudieron haber aparec ido y ex vni en múltiples 
sibilidad no 


> 8 
Z 


que linajes moleculares 
a variación en patrones moleculares y aparente estasis en el registro fósil — una 
antepas a común de definición molecular más reciente de la 80 130 iones 

Ma. si se usa un origen en el Paleoceno (ca. 60 Ma) como la base 


ocasiones. Esto permite 
n nte considerada en la literatura. El 


existentes se estima que llegó al Nuevo Mundo hace ca. 
lo 


para n e 
RO 


Investigations of fossil Rhizophora L. in the Carib- morphology, and fossils will allow continuing revision in 


phylogeny concepts, and the increasing integration of the historical 


bean, in association with studies on 
(Schwarzbach € Ricklefs, 2000), have led to this review record. with biogeographic and phylogenetic patterns 
of reports of the family Rhizophoraceae worldwide. One derived from modern systematic approaches. 

purpose of the review is to discern from the numerous 


reports of Rhizophoraceae, from strata purportedly MODERN DISTRIBUTION AND GLOBAL Fossil RECORDS 


v EA — > 
YE RHIZOPHORACEAE 


ranging in age from the Cretaceous to the Recent, 


identifications and ages that afford a reliable basis for 
reconstructing the evolution and historical biogeogra- The cireumscription of extant Rhizophoraceae and 
he family. From that base, the gradual Anisophylleaceae (included within the Rhizophora- 
ceae by some authors; e.g., Melchior, 1964) is listed in 


phy of 
accumulation of new evidence from genes, molecules. 


! Information on molecul 
on the Rhizophoraceae, and on the fossil records by Aaron M. E llison, Jeri Kvace K. Macphail, and Peter Saeng 
manuscript w | by Robyn J. Burnham, Aaron M. Ellison, Shirley A. Graham. 11 puris sa Schwarzbach. 

Missouri Bolts al Garden, P.O. Box 299, St. Louis, MO 63166-0299, U.S.A. alan.graham(Omobot.org 


ar systematics was provided by Andrea Schwarzbach from on-going studies by her and collaborators 
k, M er. 


ANN. Missouri Bor. Garp. 93: 325-334. PuBLISHED ON 23 Aucusr 2006. 


326 


Annals of the 
Missouri Botanical Garden 


Table Circumscription of Rhizophoraceae and Aniso- 
phyllaceae by various authors as given in Schwarzbach and 
Ricklefs (2000). * 


record. 


indicates genera reported in the fossil 


Rhizophoraceae 
Macarisieae 
Anopyxis Pierre 


18 
Blepharisenma \ W a ex Benth. 


Dactylopetalum Benth. 
Macarisia Thou. 
Sterigmapetalum Kuhlm. 
Gynotrocheae 
Carallia Roxb. 
Crossostylis J. R. Forst. & G. Forst. 
Gynotroches Blume 
Pellacalyx Korth. 
Rhizophoreae 
*Bruguiera Lam. 
Arn. 
*Kandelia (DC.) Wight & Arn. 
*Rhizophora L. 


* Ceriops 


Anisophylleaceae 
Anisophyllea R. Bi 
„ Hook. | 

Pie 


"ri Ducke 


Poga 


Table lt 


recognized as a 


that the 
eparate and unrelated family by 
Sehwarzbach & Ricklefs (2000), and of the 15 genera 


included in the Rhizophoraceae, only four (tribe 


is noted Anisophylleaceae is 


Rhizophoreae) are exclusively mangroves. 

Only five of these genera have been reported in the 
fossil record. For some, the age of the strata has nol 
the identifications have nol 


been confirmed or have been questioned, and broad or 


been firmly established, 


tentative identifications by the original author(s) have 
achieved greater certainty with repetition in the later 
literature. For purposes of biogeography, paleoenvi- 
ronmental | reconstructions, and 


tracing lineages, 


calibrating molecular data for taxa in which there is 
a relatively extensive fossil record, it is useful to have 
one set of paleobotanical evidence limited to reports 
in which the taxonomic level of the identifications is 
unambiguous, the age of the strata is well established, 
and the implications of the record are conservatively 
interpreted. For the Rhizophoraceae this record 
begins in the early Eocene (ca. 50 Ma; million years, 
millions of years ago), and for Rhizophora it begins in 
45 Ma; Table 2). The 
incorporation of more tentative data allows develop- 
ment of 


the middle to late Eocene (ca. 


urther theories and models of interest 


provided the assumptions, speculations, and accep- 


ance of provisional identifications are clearly desig- 
nated. This more inclusive record for the Rhizophor- 
aceae and Rhizophora extends the range into the 
1. OO Ma; e.g., Gruas-Cavagnetto, 1976; 
1987). Where both kinds of evidence 
are part of the overall record, it is worthwhile to have 


Paleocene (c: 


l'hanikaimoni, 


two calculations that bracket the potential evolution- 
ary history of the lineage. 

Pollen of the Combretocarpus-type (Anisophylla- 
ceae; Muller, 1981) was recovered from late Miocene 
deposits in northwestern Borneo (Anderson & Muller. 
1975). 
Berakas coal, and the pollen became frequent in the 
Holocene Marudi peat. Morley (197 


middle Miocene of the same area. 


A single grain was found in the Miocene 


7) found it in the 
Thus, if the family 
has a fossil record, and this has not been established. 
it consists of pollen of the Combretocarpus-type. 
possibly representing the Anisophylleaceae, from the 
middle Miocene to Holocene of Borneo. If the pollen 
s Combretocarpus, that the original 
author (Muller, 1964) did not report it from other 


material studied from Borneo, the Maracaibo region of 


and 


knowing 


northern South America, or from extensive projects by 
Bataafse 
(Shell Oil 


(Germeraad et al., | 


Internationale Petroleum Maatschappij 


Company) in and elsewhere 
908). 
that the middle Miocene wed the early appearance 


of Combretocarpus of the Anisophylleaceae. This 


Nigeria 


tis reasonable to assume 


n` 


younger than the minimum age for the Rhizophor- 
aceae (see below), and it will be interesting to see if 
this temporal sequence of the two families, and 
phylogenetic true with further 
phylogenetic and paleobotanical study. 

In the Rhizophoraceae, fruits of Kandelia Wight & 
Arn. are reported from the middle Eocene of Alaska 


(Wolfe, 1972, 1977). TR 


Certops cantiensis is reported from the | 


implication, remain 


A hypocotyl of = fossil) 
ondon Clay 
and (early Eocene, Ypresian; 52-49 Ma 
(Collinson, 1983)), and pollen is reported from the late 
Miocene of the Marshall Islands (l eeopold 1969; listed 
as pending in Muller, 1981) and possibly from the late 
973); 
The latter is not a definite report of Ceriops, but the 


flora of Eng 


locene of southwestern Australia (Churchill, | 
author regarded it as possibly present among the 
abundant and morphologically similar Rhizophora 
pollen. Hypocotyls resembling Bruguiera are known 
from the London Clay and are identified as - 
1983) 


those of 


brugutera (Collinson, 


— 


Starch grain casts and 


molds similar to modern Ceriops and 
Bruguiera have also been found in the London Clay 


Wilkinson, 


generic 


1983). Reports of fossils with uncertain 


identifications include pollen similar to 
Rhizophora or Kandelia from the earliest Eocene of 
1988) and Bru- 


in the Paris Basin 


al., 


gutera-type pollen of the same age 


France (Gruas-Cavagnetto el 


Volume 93, Number 2 Graham 327 
2006 Rhizophoraceae 


Table 2. Reports of Rhizophora in the fossil record of the Caribbean region. Fm(s) =formation(s). BP = before present. 
Age / formation Locality References 
vocene/El Bosque, La Trinidad Fms Chiapas, Mexico Tomasini-Ortiz & Martínez-Hernández, 
1984 
Middle to Late Eocene/Gatuncillo Fm Panama Graham, 1985 
Late Eocene to Present Maracaibo Basin, Venezuela Germeraad et al., 1968; Lorente, 1986 
Oligocene Colombia Lorente, 1986 
Middle Oligocene/San Sebastian Fm Puerto Rico Graham & Jarzen, 1969 
Middle Oligocene/San Sebastian Fm Puerto Rico Hollick, 
Oligocene to Present Brazi Lorente, 1986 
Oligocene to Miocene Colombia Lorente, 1986 
Early Miocene/Culebra, Cucaracha, Panama Graham, 1988a, b, 1989 
La Boca Fms 
Early Miocene/La Quinta Fm Chiapas, Mexico Langenheim et al., 1967; Graham, 
1999c 
Early Miocene/Ixtapa Fm Chiapas, Mexico Martinez-Hernandez, 1992 
Early to Middle Miocene Venezuela Lorente, 1986 
Early to Middle Miocene Venezuela Berry, 1922 
Karly to Late Miocene Brazil/Colombia Hoorn, 1994a, b, e 
Late Miocene/Gatun Fm Panama Graham, 1991a, b 
Pliocene/Herrería Fm Guatemala Graham, 1998 
Middle Pliocene/Paraje Solo Fm Veracruz, Mexico Graham, 1976 
Middle Pliocene/Uscari Fm (age fide Costa Rica Graham, 1987 
Collins to Burnham, pers. comm.) 
Quaternary 
ate glacial-Holocene Yucatan Peninsula Leyden et al., 1994 
Holocene Middle American Trench, Habib et al., 1970 
Tehuantepec, Mexico 
Holocene Isla Juventud, Cuba Moncada Ferrera et al., 1990-1991 
35,000 yrs BP Gatun Lake, Panama Bartlett € Barghoorn, 1973 
12,000 yrs BP Gatun Lake, Panama Bartlett & Barghoorn, 1973 
7000 yrs BP Playa Medina, NE Venezuela Vegas & Rull (in ms.) 
5100 yrs BP 5 Peninsula Pará, Brazil Behling et al., 2001 
Holocene ridad Berry, 1925 
Holocene 1 [Other reports of Behling, 1996 
Rhizophoraceae/Rhizophora 
from Latin America]! 
Paleocene/Cerro Boraró Fm Chubut, Argentina Petriella, 1972? 
Late Eocene to Early Oligocene/Potoco Fm Altiplano, Bolivia Horton et al., 2001^ 


rious re pu of pollen from the Paleocene to Eocene of northern South America by van der Hammen & Wijmstra (1964) 
and Wijmstra (1971) are now assigned by the authors to the late Eocene or younger. Another South American record is 
TZonocostites ramonae (pollen) from various sites on 185 continental platform off eastern Brazil, Late Cretaceous to Oligo- 
liocene, in a technical report by Pares Regali et al. (19 74a). 


2 


E 


Tentative reference of generalized wood to m. eae 
* TZonocostites ramonae, biologic cal affinities are not cited by the author. 


and in southern France (Gruas-Cavagnetto, 1991; bers of the Rhizophoraceae, those with the most 
Gruas-Cavagnetto et al., 1980). Pollen of Bruguiera- convincing fossil record are Kandelia from the late 
type (see B. gymnorhiza) is reported from the late middle Eocene of Alaska (Wolfe, 1972, 1977), and 
Miocene of the Marshall Islands (Leopold, 1969; Ceriops and Bruguiera both from the early Eocene of 
considered pending by Muller, 1981), and from the England (Collinson, 1983; Wilkinson, 1983) and the 
Oligocene of England (Bruguiera parviflora (Roxb.) ^ Miocene of the Marshall Islands (Leopold. 1969). 

Wight & Arn.-type; Machin-Pallot, 1971). With The earliest fossil record of Rhizophora is difficult 
regards to the latter, Muller (1981: 63) notes that to establish because of the uncertainty of some 
“Careful comparison failed to confirm this identifica- identifications and the unsettled age of some strata. 
tion. Points of difference are the more prolate shape, Several geologic range charts have been published 
less pronounced costae along the colpi, the indistinct that include Rhizophora (Germeraad et al., 1968: 


endoaperture and thin, psilate exine.” Among mem- Kico-Cray, 1993; Saenger, 1998; Ellison et al., 1999; 


328 


Annals of the 
Missouri Botanical Garden 


Morley, 2000). and summaries are available of the 
global record (Muller, 1981; Plazait et al.. 2001) and 
of more specific areas: Mexico (Langenheim et al., 
1967). the Caribbean Basin region (Graham. 1995). 
Maracaibo Basin (Rull, 1998), the southeastern 
Pacific (Hekel, 1972; Churchill, 1973: Anderson & 
Muller, 1975: Mepham, 1983: Ellison, 1991), and 
India (Prakash, 1960: Thanikaimoni, 1987). 

Muller (1981: 62) was of the opinion that, “The 
oldest. records [of Rhizophora] are still the upper 
Focene occurrences of Zonocostites ramonae Germer- 
aad, Hopping & Muller reported by Germeraad, 
Hopping, and Muller (1968) from the Caribbean area 
and confirmed by Pares Regali, Uesugui, and Santos 
(1974a, b) in Brazil." 
rejected by Muller: the middle Eocene of India 
(Venkatachala & Rawat, 1972, 1973), because the 


basis for the age assignment is unclear, “and since the 


Several older records were 


subzone is the uppermost one, an upper Eocene age 
appears possible” (Muller, 1981: 63): the Paleocene 
and lower Eocene of western Europe (Gruas-Cavag- 
netto, 1970; Krutzsch, 1969, 1970; Krutzsch & 
Vanhoorne, 1977), because the authors allied. T Tri- 
colporopollenites mansfeldensis with Rhizophora on the 
basis of similarities in morphology that were ques- 
tioned; and the Cretaceous (Maestrichtian) of Canada 
(Rhizophora-type: Jarzen, 1978), because the fossils 


are characterized by much wider endoapertures. lack 


of polar costae, lack of colpi, and have a uniform 
granular structure. & record in the llerdien or 
Y presian (earliest Eocene) of “Rhizophoraceae?” from 
the Pyrenees region (Gruas-Cavagnetto et al., 1988 
221) is uncertain because distinct colpi transversalis 
are nol evident in the photograph: it should be noted 
that the specimens are identified in the text as 
“Rhizophoraceae?” while on the range chart the 
question mark does not appear (Gruas-Cavagnetto et 
al.. 1988: 221 and 222, respectively). In the TOP 
(International Organization of Palaeobotany) Plant 
Fossil Record Data Base (www. biodiversity.org.uk). 
the Herdien/Y presian. record is listed under Rhizo- 


phoraceae as "first occurrence, 


There are reports of Rhizophoraceae in even older 


deposits in the early literature, but none of these have 


een confirmed: NHR ies bombacaceus Bayer 
(leaf, affinites cited as with R hizophoraceae) from the 
mal (early late Cretaceous) of Bohemia 
(Bayer, 1914) and +Rhizophorocarpus dekapetalus 
Velenovsky € Viniklár (fruit, Rhizophoraceae) from 
the Cretaceous of Bohemia (Velenovsky & Viniklár, 
1926) as listed in the Andrews (1970) catalog. No pre- 
late Eocene records for the family or the genus 
Khizophora are listed in the LaMotte (1952) catalog. 


The Bohemia fossils would be older by 50 Ma or more 


than any confirmed Rhizophoraceae. The specimen of 


R. dekapetalus is in the National Museum in Prague (J. 
Kvacek, pers. comm.); the location of R. bombacaceus 
is unknown. Neither genus is regarded as belonging to 
the Rhizophoraceae by Kvacek, who is working on the 
Bohemian Cretaceous floras, and he notes (pers. 
comm.) that in almost every case the generic 
determinations by Bayer and particularly Velenovsky 
are proving incorrect. 

The same concerns raised by Muller (1981) about 
the original identifications, subsequent. confirmation 
or acceptance of the identifications, and/or age of the 
strata also apply to other pre-late Eocene reports 
listed in the range charts previously mentioned. These 
include Rhizophora from the Paleocene (Thanikai- 
moni, 1987) and early Eocene (Prakash, 1960; 
Lakhanpal, 1970) of India. The authors efforts to 
communicate about the status of the India specimens 
were nol successful. 

Rhizophora was reported. from Borneo in beds 
initially assigned to the middle Eocene (Muller. 
1964), but the age was later emended to the late 
Eocene by Muller (1981). In the Australasian region 
in general, and in Australia in particular where the 
Cenozoic plant microfossil record is extensive, “the 


earliest Rhizophoraceae (Zonocolpites ramonae) are 


middle late Eocene from the offshore Browse Basin in 
NW Australia” (MacPhail, pers. comm.: MacPhail. 
1999), This is in contrast to older records (late 


Paleocene (60 Ma). and early Eocene (50 Ma)) from 


Australia and India cited in the summary by Saenger 


(pers. comm.: Saenger, 1998). 
Considering the critical importance of the earliest 
o 
records of a lineage in deciphering its phylogeny and 
biogeography, a conservative reading of the evidence 


is warranted. Rhizophora, as a modern genus reflected 


by the morphology of its pollen and its mangrove 
habitat, is first known in the late Eocene and appears 
1 the Old and New World 


tropics. This implies an earlier origin for the genus. 


almost. simultaneously i 


possibly in the southeast Asia-eastern Tethys Sea 
region (Raven & Axelrod, 1974; Ellison, 1991). and 
the change from its absence in the middle Eocene to 
its dominance in many tropical coastal habitats in the 


late Eocene suggests rapid migration, and an origin in, 


( 


r a rapid adjustment. to, its present specialized 
habitat. 

Among genera of Rhizophoraceae, Rhizophora 
appears somewhat later (late Eocene) than Kandelia 
(late middle Eocene) and Ceriops/Bruguiera (early 
Eocene), according to the data presented. However. 
the sequences of appearance are geologically close in 
lime, and additional paleobotanical information and 
phylogenetic studies are needed to establish the time, 
place of origin, diversification, and radiation of the 


family. 


Volume 93, Number 2 Graham 329 
2006 Rhizophoraceae 
Tue CARIBBEAN Fossi RECORD or Riizornora (Tante 2) hot house interval near the early Eocene/middle 


Confidence that advancement has been made 
understanding complex biological systems is en- 
hanced when results from two independent lines of 
inquiry are congruent. Opportunity for achieving a new 
paradigm is presented when incongruous or partially 


incongruous data sets are reconciled, when one 


data set limits the choices or supports a single 


interpretation from among several generated by 


another approach. Investigations into the historical 
biogeography of Rhizophora in the Caribbean region 
through paleobotanical techniques and by molecular 
approaches provide such an opportunity for discover- 
ing consistencies in interpretation, identifying appar- 
and formulating hypotheses to 


ent inconsistencies, 


explain the differences. There is clearly much left to 
do, as is evident, for example, from the view expressed 
by Ellison et al. (1999: 95) that 


diversity [of mangroves] resulted from in situ di- 


regional species 


> 


" and by Plaziat et 
that 


versification after continental drift 
al. (2001: 1961) that “It is 
continental drift had a limited role in the dispersal 


therefore clear 
and development of modern mangrove floras.” The 
Caribbean region is a suitable place for comparing the 
fossil record with speciation models derived from 
of Rhi- 


zophora there is one of the best documented in the 


molecular systematics, because the history 


world. 
A scenario for the origin and spread of the 
mangrove lineage (4 of 15 genera) within the 


retation of presently 


Rhizophoraceae, based on inter] 
available molecular data, and augmented by in- 
formation from the fossil record, is presented by 


Schwarzbach € Ricklefs (2000, 2001). The essentials 


of that scenario. as it relates to the Caribbean Basin, 


are as follows: 


(1) Rhizophora reached the New World about 40 Ma (based 
on evidence from the fossil record). 

(2) The n underwent range contraction and extinction 
in the late Tertiary associated with cooling climates 
(evidence for cooling climates is based on the fossil 
record and other geological e vidence). 

(3) The genus may have disappeare «d bon the Atlantic and 

Caribbean regions (a possible interpretation of the 
molecular data). 

(4) : so. Rhizophora was prob: ably re-introduced about 


1 Ma (a possible interpretation of the molecular data 


using a date of 60 Ma for the origin of the family). 


To better constrain the time during which Rhi- 
zophora most likely would have disappeared from the 
probably asso- 
1999a: 87), 


the following summary is presented of the relevant 


Caribbean region (“the late Tertiary, 


ciated with cooling climates” (Graham, 
global paleoclimatic trends as a context for assessing 


the Caribbean record. Beginning with the end of the 


Eocene boundary (ca. 52 Ma): 
Early Eocene—earliest mountain glaciers first appeared on 

Antarctica (Birkenmajer, 1990 
Late middle Eocene and early Oligocene 


Antar 


limited ice sheets 
appeared along the margin of etica (Leg 119 
Shipboard Scientific 
Farly/late Oligocene boundary (ca. 30 Ma) 
ice developed on Antarctica (Miller et al., 
Matthews. 1984), as reflecte 
> 1 


more extensive 
1987: Poore & 
ed ina m drop in global 
19994: 95), 


Middle Miocene Ma)—permanent ice formed on 
Antarctica and glac al climates developed in ils Arctic 
Middle Pliocene (ca. 3.4 Ma) extensive northern hemi- 


1971). 


began; present in- 


sphere glaciation began (Shac "a Aon. & Opdyke, 
1.6 Ma)— 


11 Kyr (thousand years, 


ic e age 


c 


Qu: ternary 
terglacial hee gan ca. thousands 


of years ago). 


If cooling climate was a factor in the possible 


disappearance of Rhizophora from the Caribbean 
region in the late Tertiary, 15-14 Ma would be 
a likely time as shown by the initiation of permanent 
ice on Antarctica, glacial climates in the Arctic 


region, and a number of terrestrial vegetational 
responses on a global scale (e.g., the trend toward 
replacement of forest by grasslands in mid-continent 
North America (Graham, 1999a) and the disappear- 
ance of warm elements from the flora of New Zealand 


(Lee, 


changes also occurred in the vegetation of northern 


pers. comm.). Beginning about this time, 


atin America that would be consistent with. and 


reciprocal to, a restriction in the range of Rhizophora 
due to cooling climates. The principal change was the 


introduction of an increasing number of northern 


temperate elements progressively farther south into 


the lower latitudes (Abies Mill., Picea A. Dietr., Pinus 
L., Alnus L., Celtis L., Fagus L., Juglans L., 
Liquidambar L., Myrica L., Populus L., Ulmus L.: 


Graham, 1999b). These events identify the most likely 
beginning for any disappearance of Rhiz puo from 
the Caribbean 14-15 Ma. The 


period between 14-15 Ma and Rhizophora’s estimated 


region at 


return at ca. 11 Ma corresponds to the Serravallian 
Stage of the middle Miocene (10.4—14.2 Ma: 
1989). and is part of the “ice house” interval 


Harland 
et al., 
when global climates were becoming cooler. In the 
combined analyses by Sehwarzbach € Ricklefs (2000, 
2001; Schwarzbach, pers. comm.). a maximum age of 
60 Ma is used for the earliest Rhizophora from which 
a New World/Old World split in Rhizophora, and its 
reintroduction into Vorld at 11 Ma is 
calculated. When the 45 Ma is 
used (Schwarzbach, in prep., pers. comm.), bracketing 
the possibilities afforded by the fossil record, that split 
New World 


the initial appearance of 


the New 


younger age of ca 


and any reintroduction into the has 


a younger dating. Thus, 


Annals of the 
Missouri Botanical Garden 


thizophora in the Caribbean and its subsequent 
record is of interest. 

Beginning in the Quaternary, it is probable that the 
range of Rhizophora in the Caribbean region fluctu- 
salinity, and sea 
1996). However, if Rhizophora 


disappeared as a consequence of the colder-than-al- 


ated with changes in temperature, 


level (see Blasco et al., 


present intervals of the past 2 Ma, then accommoda- 


tion must be made for the likelihood of 18 to 20 of 


these presence/absence cycles, because that is the 


number of such cold intervals documented. for the 
Quaternary Period in the northern latitudes (Milanko- 
vitch eccentricity cycles of 100 Kyr; see papers cited 
1999a: 37—41); the effect of these cold 


intervals on biotas is now known to have extended into 


in Graham. 


the low latitudes, as shown by strontium/calcium ratio 
1994) 
and North Atlantic Heinrich events reflected in pollen 
sequences from south-central Florida (Grimm et al.. 
1993; see papers cited in Graham, 1999a: 283). There 


are no plant fossil-bearing sections that extend back 


data from Barbados corals (Guilderson et al., 


into one or more of the glacial maxima in the 


Caribbean region to provide direct evidence of 

history during the cold periods (e.g, at 18 Kyr). 
However, the paleobotanical evidence that is avail- 
able suggests that Rhizophora did not Pru and 
Ma in that 


For example, sediment cores 


return 18 to 20 times during the past 2 


region (Table 2). 
Lake, 


Rhizophora 


Ls 


rom 
that at ca. 35 Kyr 
14% of the total 
palynomorph assemblage, and it has been present in 
12 Kyr (Bartlett & 


Barghoorn, 1973). At any rate, brief, relatively recent, 


Gatun Panama show 


pollen comprised 


varying amounts for the past 
rapid, and partial range restrictions are not the kind 
needed to explain that aspect of the molecular data 
3—4 Ma 
absence from the Caribbean region between 15-14 to 
11 Ma. The 5—14 Ma, 3.4 Ma, 


1.6 Ma, and the fast-paced a fluc- 


open to the possible interpretation of a 


cooling events al 


tuations of the Quaternary Period do, however, provide 
a mechanism for the presence of cold(er) tolerant 
southern Florida, 


populations of Rhizophora in 


amaica, St. Croix, and Texas mentioned by Sehwarz- 
bach & Ricklefs (2001). 


For the Tertiary, 


the question is whether it is 
a viable option, in interpreting the molecular data, to 
assume that Rhizophora was absent from the Carib- 


bean region for a significant span of time after its 


initial appearance in the late Eocene. The Caribbean 
fossil record of Rhizophora is based on fossil pollen, 
known in the stratigraphic literature as Zonocostites 
ramonae. Numerous well cores have been drilled and 
the 


aribbean Basin: Rhizophora pollen is 


examined extensively for plant microfossils in 
petroleum-rich ( 


distinctive and easily recognizable at the generic level 


(Langenheim et al., 1967; Muller & Caratini, 1977) 
and is an indicator of depositional environments—an 
important aspect of petroleum geology. For these 


reasons, the stratigraphic range of Rhizophora pollen 


n the Caribbean Basin is well known and can be 


confidently fixed as beginning in the late Eocene. The 


oldest occurrences are in the late Eocene Gatuncillo 
Formation of Panama, and the early to middle 
l 


Germeraad et al., 


part of 


1e late Eocene of Venezuela (Maracaibo Basin: 
1968; Lorente, 1986), corresponding 
to a time period around 45-40 Ma 

Regarding the subsequent history of Rhizophora, 
Lorente (1986) studied a section from the late Tertiary 


of Venezuela that includes the interval from the ecrly 


ae, 


to middle Miocene (Burdigalian to Langhian), and 
Zonocostites ramonae is present throughout. It has also 
been reported from the Oligocene to Miocene of 
Colombia, from the Oligocene to the Recent of Brazil, 
and from the late Eocene to the Recent generally 
throughout northern South America (Germeraad et al., 
1968: 1986); in early (Graham, 1988a, b. 
1989) and late (Graham, 199la, b) Miocene deposits 


Panama; and 


Lorente, 


Pliocene sediments from Mexico 
(Graham, 1976). In both published and unpublished 
studies, Rhizophora is almost universally present from 
the late Eocene to the Recent in pollen/spore-bearing 
sediments from coastal habitats throughout the 
Caribbean Basin, and there is no evidence for a 3— 
4 Ma gap between 15-14 and 11 Ma, oral any time in 
the post late Eocene record. 

A second point based on paleontological evidence 
concerns a cause for a possible disappearance of 
Rhizophora, and the timing of its subsequent return. Lf 
cooling temperatures were a reason for ils disappear- 
15-14 Ma, 
paleotemperature curve shows that it was even colder 
at 11 Ma (Graham, 19992: 89). 


lo explain its re-introduction under conditions more 


ance al it must be noted that the global 


This makes it difficult 


severe than those purported to explain ils disappear- 
ance. There are no abrupt, short-lived, extensive 
catastrophic events known for the Tertiary interval 
involved, and fossil records from other groups do not 
show a response to such an event. Habitat reduction is 
unlikely during the time under consideration, because 
coastal environments were expanding with the fall in 
sea levels as glaciers began to develop more 
extensively, 

It is clear that phylogenetic information alone does 
not discriminate among all possible patterns. and 
information from the fossil record can help to limit the 


Also, 


between a 


different options. the fossil record. « 


discriminate continuous presence 
periodic absences as a result of catastrophic events. 


here should be 


In the case of a continuous presence 


a relatively uninterrupted fossil record, while cata- 


Volume 93, Number 2 
2006 


Graham 331 


Rhizophoraceae 


strophic events should produce gaps in that record, 
likely accompained by similar evidence from adjacent 
regions, and this evidence is not present. 

However, there are several limitations that must be 
acknowledged in reconstructing the historical bio- 
geography of Rhizophora based on evidence from 
fossil pollen. Its history (Graham, 1995) is derived 
that 


intermittent for several reasons. 


from records for any one region may be 


One is that many 
samples are derived from roadside, streamside, and 
surface mine exposures that are widely separated both 
geographically and stratigraphically. Another is that 
even in cores resulting from drilling operations, there 
are zones missing or barren of terrestrial plant 
microfossils because there are changes in depositional 
environments (e.g., inundation, marine regression and 
erosion), the complete section was not studied, or 
results were considered and 


selected proprietary 


unpublished. A third reason is that tropical pollen 
and spore-bearing lignites, clays, and siltstones may 
be lain down under different depositional environ- 
ments. Those deposited at the coastal interface 
between terrestrial and marine habitats. frequently 
contain. Rhizophora pollen (e.g.. the middle Pliocene 
Paraje Solo Formation of Mexico; the early Miocene 


Culebra, Cucaracha. and La Boca formations of 
Panama; the late Miocene (fide Collins, pers. comm. 


Table 2). 


Similar sediments, but deposited in inland swamps, o 


to Burnham) Gatun Formation of Panama: 


upland bogs and marshes, removed from the influence 
of marine waters, do not contain Rhizophora pollen 
(e.g. the late Miocene Padre Miguel Group of 
For 


continuous lines (e.g., Germeraad et al., 


charts with 
1968, fig. 15) 


imply the continual presence of a laxon over millions 


Guatemala). these reasons, range 


of years, but this can be misleading because sections 


the 


for Tertiary are not sufficiently continuous 1 
or tens of 


the 


c 


detect absences of several thousands 


thousands of years. For Rhizophora, however, 
combined records reduce the possibility of an absence 
even for this duration. 

Another limitation to the paleopalynological ap- 
proach for the present study is the taxonomic level to 
which Rhizophora pollen can be recognized. lt is 
important to that 


molecular-defined lineages; in terms of morphology, 


note lineages here refer to 


including pollen morphology, the genus is not readily 


subdivided into clearly recognizable taxa. Langen- 
heim et al. (1967) and others have studied the 


morphology of Rhizophora pollen. and based on the 


complete geographic and morphologic range of the 


genus they have concluded that no ecological or 
taxonomic categories can be consistently recognized 


on the basis of pollen characters. This leaves open the 


possibility that different lineages. defined by molec- 


ular and/or non-pollen morphological characters, may 
have migrated in and out of the region, leaving no 
evidence of the fluctuations in the plant microfossil 
record. This possibility has not been widely empha- 
sized in the literature tracing the geologic record of 
important as 


but it will be increasingly 


— 


ineages, 
histories based on paleobotanical evidence are in- 
tegrated with those derived from molecular studies. In 
the case of Rhizophora, the stenopalynous condition of 
the pollen allows for the possibility of repeated 
extinctions of molecular- and 


introductions and 


otherwise-defined lineages. However, the coarse 


resolution of the fossil record, sea-level and paleo- 
temperatures trending in the wrong direction. and 
undetected or selective catastrophies cannot be 
invoked to support a prolonged absence of Rhi- 
zophora, as a genus, from the Caribbean region after 


t 


— 


ie late Eocene. 
According to the fossil record, Rhizophora has been 
widespread in the New World 


appearance in about the late Eocene. No major gaps 


since its initial 


are apparent in its distribution since that time, and 


there is no evidence for region-wide catastrophic 


events, The model around which consensus is de- 
veloping is one that involves frequent extinction and 
reintroduction of molecular-defined lineages, and 
pollen that documents the presence of Rhizophora, 
in the Caribbean region since the late 


as a genus, 


Eocene. According to recent calibration 
Schwarzbach & Ricklefs, unpublished data) the most 
recent common ancestor or colonizer that spread 


throughout the New World arrived about 11 Ma and 


diversified in response to environmental change. 


Consistent with the model are the documented 
climatic fluctuations and changes in the physical 


environment (Graham, 2003a, b, c) now known to have 
the 
changes in sea level and ocean salinity concentrations 


affected Caribbean region and the associated 


resulting from rapid glacial and interglacial transi- 


tions. Results from molecular-based studies may 
assume (l) intuitively a Paleocene origin (ca. 


60 Ma) for the family because of its distribution and 
an Eocene origin (ca. 


— 


diversity in the Eocene, and (2 
45 Ma) based on a conservative reading of the existing 
fossil record. These bracket the plausible options, and 
define eritical time intervals where additional, verifi- 


able records should be sought. 


Literature Cited 

Anderson, J. A. R. & J. Muller. 
a Holocene peat and a Pus "s 5 1 from NW. 

Rev. Paleobot. Palynol. 19: 291—351. 

Andrews, H. N. Jr. 
fossil plants, 182( 


1-354. 


1975. Palynological study of 


Borneo. 


1970. ni. er generic name 


0-1965. Bull. U.S. Geol. Surv. 1300; 


Annals of the 
Missouri Botanical Garden 


Bartlett. A. S. & Barghoorn. 1973. Phytogeographic 


ustory of th Jeune of Panama during the past 
12,000 years (a history a vegetation, cli 
Pp. 203-2 in A 


[e of 
Amste 1 


level change). 
Vegetation and Vege o 0 
America, E ae cience HE HA ee 
Bayer, E. 1914 ke príspé 


českých a. vrstev d 


cy ku Boni 
Arch. Prir. Vvzk 


Cech 25: 1-66. 
Behling, | 

occurrence of Podocarpus and possible human presence at 
the Late-glacial. Veg. 


1996. First report on new evidence for the 


An during 
st. & Archaeobot. 241-246 
— U 15 G Cohen n & R. Jo 2001. 


Holocene mangrove ecosystem dynamies of the Bragança 


a Pu of the 


Lara. Studies on 


Peninsula in north-eastern Tu razil. Palacogeogr. 


2 


Palaec 00 en es 'oecol. 167: 225-242. 
W. I9 


v 
, 


rtiary pe plants from Venezuela. 

9 E Mus. : 59: 553-579 

Pleistocene flora 
Nini Boos. U. S. Natl. . 66: 

Birkenmajer, J. 1990. G 
phy of Tertiary glacial a 


eoc in and c ‘limatostratigra- 
George Island, South Sh 


+ Workshop on Antarctic geochronol- 
Se id "izerbart sche: Verlagsbuchhandlung, Stutt- 


1996. 


Ic 67-178 


. Saenger & E. 
indicators of pin change. Catena 27 
Churchill, D. M 
mangroves in the early 
\ustralia Special Pap. No. 4: 


Janodet. Mangroves as 
78. 


113. The ecological significance of tropical 


Tertiary floras of southern 


a 
— 
= L 
co 
a 


E. 1983. Fossil plants of the London Clay. 


Fossils No. . The Paleontological 


J. Farnsworth € R. E. Merkt. 1999, Origins 


of mangrove ecosystems and the mangrove biodiversily 
15 


n 


5- D. 


Global Ecol. Biogeogr. 
1991. The Pacific 


(Rhizophoraceae). Bot. J. Lir 


Ellison, J. C. 
oe pc 1 L. 
. d )5 -284. 
1. A C. A. Hopping & J. Muller. 1968. 
Palynology of Tertiary sedime nts from tropical areas. Rev. 
189—348 


Wt. 


jermera: d. 


~ 


Paleobot. Palynol. 6: 8. 

76. Studies in neotropical paleobotany. IL The 

ae 'ene communities of Veracruz, Mexico. Ann. Missouri 
me 


Bot. Gard. 63: 787—842. 


Graham, A. 197 


1985. Studies in neotropical paleobotany. IV. The 
Eocene communities of Panama. Ann. Missouri Bot. Gard. 
72: 504-534. 


1987. Mie 


me us at southern Costa Rica. 
15 


ene Communities and paleoe nviron- 


Amer. J. Bot. 74: 1501— 


paleobotany. V. 
communities of | Panama—the 
Missouri Bot. Gard. 75: 1440 


588 . Studies in neotropical 


The lower Miocene 
Culebra Formation. Ann. 
1466 

1958 


Miocene communities of Panama— 


b. Studies in neotropical paleobotany. VI. The 


lower the Cucaracha 


Formation. . Missouri Bot. Gard. 75: 1467-1479 
989, ee in neotropical 1 VII. The 
lower. Miocene communities of Panama the La Boca 
DUE Ann. Missouri Bot. Gard. 76: 50-66, 
—. 1900 Ja. Studies in ne oe “al 15 obotany. IX. The 
Pliocene communities of Panam (dicots). 


—angiosperms 
3 


Ann. Missouri Bot. Gard. 78: 9 


from the island of 
0. 


palacogeography of 
B ] 


199] b. 


: Studies in neotropical paleobotany. X. The 
Pliocene communities of Panama 


composition, numeri- 


cal. representations, and o dn nviron- 


mental reconstructions. Ann. Missouri Bot. zard. 78 
465—475. 
1995. Diversification of Gulf/Caribbean mangrove 
communities 1 5 Cenozoic time. Biotropica 27: 20-27 
1998. Studies in neotropical paleobotany. XI. Lat 
Tertiar wry e anch environments heastern 


E soul 
: palynofloras from the Mio-Pliocene Padre 
Migue ; Croup and the Pliocene Herrería Formation. Amer. 
J. Bot. 85: 1409-1425. 
19993. Late 


North American vegetation (north of Mexico). Oxford U niv. 


Cretaceous and Cenozoic history of 

p ress, v ww Yo 

999». The 

temperate element in the northern Latin American biota. 
38 


Tertiary history of the northern 
Amer. J. Bot. 86: 32— 
1999c. Studies in neotropical paleobotany. XIII. 


Oligo-Miocene  palynoflora from Simojovel (Chiapas. 
Mexico). Amer. J. Bot. 86: 17-31. 
— ——-, 2003a. ( 


re ohistory models and Cenozoic paleoenvi- 
. Bo 378-380 


deprime “i and 
401 — 


ronme 5 of the Caribbean region. Sys 1. 28: 


003b. The concepts of 


on 


lean hot- -spols. : Syst. Bot. 

. 2003c. Historical 2 deu) of ee Greater 
re S. Brittonia 55: 357-38. 

D. 1960. 

pale 1 l. The .ligocene 


Studies in neotropical 


— 


Jarzen. 
communities of Puerto 
Rico. Ann. Missouri Bot. Gard. 50: 308-357. 


„X. X. Watts. B. C. S. Hansen 


Grimm, E. W^ ». L. 
& 


K. Maasch. 1993 50.000- year record of climate 
Osc n from Florida and its te mper al correlation with 
the Heinrich events. Science 261: 198-200 


Td Ard C. 1976. Etude papuo qie du Paléo- 
M : 1-49 


gene du sud de l'Angleterre. Cah. ropaléontol. 1: — 


99]. Pollens de dA o ées de PLlerdien moyen 
(Kocene inférieur) de Fordones (Corbières. France). Cah. 
Mieropaléonte S. 0: 51-60. 

aurain & R. Meyer. 19 


dans les lignites du Soissonnais (Y présien) à 


80. Un sol de mangrove 


= 


fossilizé 
Geobios 16: 795-80 
Villatte 
Thanétien et 


0 


Verzenay (Marne ). 
Y. 


bareau & Données 


an 
MS vécologiques nouvelles sur le 
de l'avant-pays pyrénéen et de la Montagne Noire. Tras 
sect. Set. Techn. Inst. Franc; Pondichéry 25: 219-2: 
T. P., R. G. . Rubenstone. 1994, 
Tropical temperature variations since 20,000 years ago: 


— 
T: 


Guilderson, 


modulating interhemispheric climatic change. Science 

263: 603-605 

Habib, D.. D. Uds . D. Ross & J. 1970. 
Holocene ma Ivnology of the Middle American Trench near 
1 . Mexi Mem. Geol. Soc. Amer. 126: 

26 1 

Es. 2 „K. L. Armstrong. A. V. Cox, L. E. Craig. 
G. e * 5 G. Smith. 1989, A ge al lime se 2 198 
Cambridge Univ. Press, Cambridge. 

lekel, H. 1972. 


Donahue. 


Pollen and spore assemblages from 
Queensland Tertiary se a its. Publ. Geol. Surv. O 
Pap. 30): 1-33. 


Paleobotany of Porto 


Jueens- 
land 355 (Palaeontol. 

Hollick. A. 1923. Rico. New 
\cademy of Sciences Scientific Survey of Porto Rico and 
the Virgin Islands 7: 177—393 

Hoorn, C. 1994a. Miocene Palynostratigraphy and Paleoen- 


vironments of Northwestern Amazonia, Evidence for 


Marine Incursions and the Influence of Andean ‘Tectonics. 
| i 


hD. Dissertation. Univ. Amsterdam. 


Volume 93, Number 2 Graham 333 
2006 Rhizophoraceae 
O4b. An environmental reconstruction of the Moncada Ferrera, Hernández Fuentes & M. 


palaeo-Amazon River system (middle-late Miocene, NW 
Amazonia). Palaeogeogr. Palaeoclimatol. Palaeoecol. 112: 
187-2: 

. Fluvial paleoenvironments in the intracra- 
iocene-early middle 


Palaeoclimatol. Pa- 


tonic bo. Basin (early 


Miocene, ae ibia) la ee 
laeoecol. 1 54. 


Horton, B. K., " A. Hampton & G. L. 


synorogenic 


Waanders. 2001. 


Paleogene sedimentation in the Altiplano 


Plateau and 9 0 0 ations for initial mountain building in 
l 


the Central Andes. Geol. Soc. Amer. Bull. 113: 1387- 
1400. 

Jarzen, D. M. 1978. Some Maestrichtian palynomorphs and 
their phsiogeographie 'al and paleoecological implications. 
Palynology 2: 29-38 

Krutzsch, W. 1969. Über einige stratigraphisch wichtige 


neue longaxoner- is aus dem mittleuropáischen Alt- 
9-31 


tertiár. Ge es xe 18: ). 
1970. Die 1 :h verwertbaren Sporen- und 
Pollenformen des DISCUS hen Alttertiárs. Jahrbü- 
cher für Geologie 3: 309-374 
& R. N e 197 bi 


e Pollenflora von Epinois 

und oe in Belgien. Buts hica " 163 
1-11 

p R. 1970. 


»earing on the historical geology of the region. 
25-604 


floras of India and their 
Taxon 19: 


Tertiary 


* 


LaMotte, R. S. . Catalogue of the Cenozoic plants of 


North America Pes 1950. Mem. Geol. Soc. Amer. 51: 
1-381. 
Langenheim, J. H., B. Hackner € A. Bartlett. 1967. 


Mangrove pollen at the depositional site of the Oligo/ 
Miocene amber of Chiapas, Mexico. Bot. Mus. Leafl. 21: 
289-324. 


Leg 119 a Scie aa d 1988. Early glaciation of 
Antarctica. Nature 333: 304. 

Leopold, E. B. 1969. 1 pollen and T flora of 
Eniwetok Atoll. Marshall Is lands Profess. Pap. U.S. Geol. 
Surv. 260-I1: 1133-1 85. 

. Brenner, 
994. Orbital dnd 

Yucatan Peninsula for the 


A. Hodell & J. A. Curtis. 


1 forcing of climate on the 
ca. Palaeogeogr. 
Palae 'oclimatol. Palaeoecol. 


Lorente, M. A. 198€ 


. Pa — 1980 palynofacies of the 
9. J. 


i Tertiary in Venezuela. . Bot. 9 Cramer, 
erlin. 
115 hin- Pallot, J. 1971. Plant microfossils from ee 'riary 
deposits " the Isle of Wight. New tol. P: 851 
Macphail, K. 1999, 5 . "id 
Basin, o southeastern Australia. E. 23: 
197—240. 


Martínez-Hernández, E. 1992. Caracterización ambiental del 


Terciario de la region de Ixtapa, estado de € 1 un 

enfoque palinoe E o. Revista Univ. Nac. 
México 10: 54-04. 

Melchior, H. 1964. 
Melchior (editor), Engler's 
12" ed., vol. 2. 

Mepham, R. l 


continents, part 1: des ge 9 lical origin of tide: Pacific 


d ae. Pp. 
Syllabus der Pfl. anzenfamilie n, 


mangrove genera and the deve 258 and 19 nt status 
of the Australian mangroves. S. African J. 8. 
Miller. K. C., R. G. Fairbanks & C. 9. D 1987. 


y oxygen Isotope aaike sis, sea le vel history, and 
2: 1-19. 


contine ntal margin erosion. 


1990-1991. Anális 
se sedime antos marinos del occidente de la Isla de 
a wa 5 Hung. 36: 145-161. 
olo 77. Te 
a nts in 1 "asl Asia. Proceeding of the Indonesian 
2" ¢ 


Cabrera Cee non polínico de 


la Juventud 


dro of Tertiary and Quaternary 


Pe Helen "um Ássocic nnual Convention: 255-276. 


Evolution of Tropical Rain 
Forests. 75 Wiley & Sons, New Yor 

Muller, J. 1964. A palynological contribution to the history of 

3342 in L. M. 


Hawaii 


the mangrove vegetation in Borneo. Pp. 
Cranwell, Ancient Pacific Univ. 
Honolulu. 


Floras. Press, 

———. 1981. Pu pollen records of extant angiosperms. 
Bot. Rew. 4T: 1-142. 

& 


8 


d pum 1977. Pollen of Rhizophora 


(Rhizophorac eae) as a guide fossil. Pollen € Spores 19: 

Uesugui & A. S. Santos. 1974a. 

ee re se 1 meso-Cenozoicas do Brasil (1). 

len. dii (Rio i 5 17 

)74b. Pilisglosta dos sedi- 

mentos meso- pde oicas 5 Brasil (II). Bol. 
Petrobras (Rio de Janeiro) 17: 263—301. 

Petriella, B. 1972. E 
Tertiario inferior del area 
Bororo). Revista Mus. La Plata 6: 

Plaziat, , C. Cavagnetto, J.-C. 

2001. pom He biogeography of the mangrove ecos 
of 


M Regali, 
: 17 4 T 

Téen. 
Estudio de maderas petrificadas del 


central de Chubut 
159-254. 


(Cerro 


Koeniguer & F. Blatzer. 
tem, based on a critical reassessment of the 
i d. Wetlands Ecol. 
[Special issue: Life on the Edge: 
Mangroves, N. C. D . Schwarzbach, editors. | 
Poore, R. Z. & R. K. Matthews. 1984. 


ranking of late Eocene and Oligocene planktonic forami- 


leon 
Managem. 9: 161-179. 
Past and 1 of 
Oxygen isotope 
© Oligocene sea-surlace tempera- 


nifera: ii ps ations for 

tures and global ice-volume. Mar. Micropaleontol. 9: 

1111: 

rs U. 960. A survey of the Deccan Intertrappean flora 
ia. 12 jl. 34: 1027-1040 

ns n, " if 


ge e past continental movements. Ann. Missouri 
3€ 


Axelrod. 1974. Angiosperm bio- 


Bot. Gard. 61: 539-673. 


hico-Gray, V. 


1993. Origen y rutas de dispersión de los 


mangles: una revisión con énfasis en las especies de 
uS d Acta Bot. Mex. 25: 1-13. 


Rull, 


c m 's in the 
287-29 


998. Middle Eocene mangroves and vegetation 
Maracaico Basin, Venezuela. Palaios 13: 
a 

P. 1998. 


Austral. 


Saenger, 
perspe ctive. 


Mangrove vegetation: 


J. Mar. 


an evolutionary 
Freshwater Res. 49: 277 


& R. E. Ricklefs. 2000. 


affinities of Rhizophoraceae and Anisophylleaceae, and 


Sbado A. E. Systematic 
intergeneric relationships within Rhizophoraceae, based 

chloroplast DNA, nuclear ribosomal DNA. and 
Amer. J. Bot. 87: 547—564 


on 
morphology. 
& ———. 2001. The 


use of molecular data in 


mangrove plant research. Wetlands Ecol. Managem. 9: 
195-201. [Special issue: Life on the Edge: Past and Futurc 
X Mangroves, N. C. Duke & A. E. Schwarzbach. editors. : 


1977. Oxy 
dd northern Ba 


Sio kleton, N. J. & N. D. Opdyke. 
paleomagnetic evidence a 
glaciation. Nature 270: 21€ 

Thanikaimoni, G. 1987 
Sci. Techn. Inst. Rane: Pondichéry 24: 1-100. 


gen Isotope a 


1 palynology. Trav. Sect. 


Annals of the 


Missouri Botanical Garden 


'omasini-Ortiz, A. C. & E. Martínez-Hernández. 198: 
Palinologia ds | Eocene-Oligoceno de Simojovel, C lapas: 
Paleontol. Mex. 50: 1-49 

van der Hammen, T. & T. A. Wijmstra. 1964. A 

palynological study on the run and Upper Creta- 

ceous of British Guiana. Leidse Geol. Meded. 30 
183-241. 


Velenovsky, 


K I. Viniklár. 10926. Flora Cretacea 
Bohemiae. C zechoslovakia Statniho Geologického Ustavu 
Rozpravy pt. 2: 1-51, 8-17. 


E he d B. 5. & M. 5. Rawat. 1972. ve of the 


Tertiary sediments in the Cauvery Basin. oes 'Ocene- 
Eocene palynoflora Í rom the subsurface. n 12-335 
A. E esip Seminar in Paleopalynology Indian 


Stratigraphy. A. K. Ghosh, Calcutta. 


& ————. 1973. Palynological 


Tertiary. sub-crop ie nce in the 


Caus 


zonation of the 


ery 


Basin. south 


India. Proceedings of the I International Palynology 
Conference, Palynology of the Cenophyte: 73-78. 
Wijmstra. T. A. 1971. The palynology of the Guiana coastal 


basin. Drukkerij ds Kempenaco, Oegstgeest. 
1 : 
Wilkinson, II. P. 1983. Starch grain casts and moulds 


Eocene (Tertiary) fossil mangrove hypocotyls. 


5]: 39-45. 


Inn. Bot. 


Wolfe, J. A. 1972, An interpretation of Alaskan Tertiary 


floras. Pp. 201—233 


Paleofloristics of Asi 


977. Paleogene 


region. Profess. Pap. 


a and 
Elsevier Science Publishers. 


Graham. 


Eastern 


Surv. 


Amsterdam. 
floras from the 


US. Geol. 997: 


Floristics and 


North 


America. 


Alaska 


MODERN PROCESSES AND Alan Graham 
HISTORICAL FACTORS IN THE 

ORIGIN OF THE AFRICAN 

ELEMENT IN LATIN AMERICA 


ABSTRACT 


The combination of factors that account for present-day distributions of organisms is unique to individual lineages and 
varies over time. An observation relevant to some lineages al some point in their history is that hurricane frequeney and 
intensity appears to be increasing with global warming. If so. then directional winds from Africa to the Caribbean region (the 
trade winds), from South America to Africa (the westerlies), the ocean currents they induce, and the transport of the floating 
islands they carry. were likely more intense during the generally warmer-than-present Tertiary Period and especially at peaks 
of exce ibid 1 00 These peaks occurred in the Pale 'ocene/Eocene (05—45 Ma), in the early to middle Miocene (broadly 
between ~ 2 Ma). in the middle Pliocene (3—4 Ma), and probably extended arl ater in the lower latitudes. The 
Paleocene 9255 ne interval includes the time when the distance between Africa and South America was one-half to Iwo- thitds 
that of the present. and when the Greater Antilles island arc was first becoming emergent as increasing target areas for 
propagules. The second and third intervals of warming include the times when 170 evidence suggests divergence 
between several African and New World lineages. Thus, wind and ocean transport of organisms and propagules throughout the 
Tertiary, and especially at peaks of warmth, was likely a more important means of dispersal than would seem plausible under 
present conditions. 

Key words: Africa, Caribbean. dispersal. dust transport 


RESUMEN 


ac acon ion de factores que explican la distribución actual de los organismos es ünica a los linajes individuales y varía 

y través del tiempo. Una observación relevante para algunos linajes en un cierto punto en su historia es que la frecuencia y la 
intensidad de ur acanes parecen aumentar con el calentamiento global. De ser así, entonces los vientos direccionales desde 
Af los 4 vientos alisios), desde América del Sur hacia Africa (los vientos del oeste), las corrientes 


— 


rica haci ‘ld la region de >] Caribe 


oceánicas que inducen y el transporte de las islas flotantes que llevan, fueron posible mente más intensos durante el período 
Terciario, generalmente más caliente que el presente, y especia TB durante los picos excepcionales de calor. Estos picos 
ocurrieron en el Paleoceno/Eoceno (65-45 Ma). en el Mioceno temprano a me 1 d sliamente entre —23-12 Ma). en el 


lemente se extendieron levemente más adelante a las latitudes más bajas. El intervalo de 


Plioceno medio (3—4 Ma) y proba 
Paleoceno/Eoceno inc luye el tiempo en que la distancia entre Africa y América d 4 Sur era la mitad a dos tercios de la actual, 
y ace el arco de las Antillas Mayores recién empezaba a emerger como área blanco para los propágulos. El segundo y el 
tercer intervalos de calentamiento incluyen los períodos en que la evidencia molecular sugiere divergencia entre varios linajes 
africanos y del Nuevo Mundo. Así, el transporte de los organismos y propágulos por viento y océano durante el Terciario, 

espec ialmente durante los picos de calor, fue probablemente un medio más importante de dispersión de que parecería 


plausible en condici “ones acluale 


estimated 114 species showing relationships between 


— 


rica 


An increasing number of plants with A 
affinities is being recognized in the Gulf/Caribbean — Africa/Madagascar and the Neotropics of which 27 are 


and adjacent region. For example, the American common to both areas and 87 are species pairs. 
Lythraceae genera Crenea Aubl. and Ginoria Jacq. Migration seems to have been primarily from South 
have African origins (S. Graham, 2002). The Domin- America to Africa, but, as the authors note (Moran & 


ican amber has long been regarded as a product of the Smith, 2001; Smith, pers. comm., 2004), the direction 
legume Hymenaea L., but Hueber and Langenheim of movement is not always clear, and extinctions may 
(1986) have shown that rather than being produced by have obscured the pattern. In a study of 123 species of 
the widespread New World species H. courbaril L., its Elaphoglossum Schott ex J. Sm. (Elaphoglossaceae). 
spectral pattern is most similar to the African H. Rouhan et al. (2004) concluded on the basis of 


verrucosa Gaertn. In the pteridophytes there are an molecular evidence that at least 21 long-distance 


The author gratefully acknowledges information on extant African-South American pteridophytes provided by Alan R. 
Smith, on various Compositae and other angiosperm groups by Karin Tremetsberger, and on dust particles in the Caribbean 
region by 1 Arimoto, Virginia H. Garrison, and Joseph M. Prospero. The manuscript was read by Shirley A. Graham, Alan 
R. Smith. is Carmen Ulloa Ulloa. 

Missouri Botanical Garden, P.O. Box 299 St. Louis, MO 63166-0299, U.S.A. alan.graham@mobot.org. 


ANN. Missouni Bor. GARD. 93: 335-339. PUBLISHED ON 23 AucusT 2006. 


336 


Annals 
1 E Garden 


dispersal events occurred between the African-Indian 
henner (2004 


further lists 110 angiosperm genera with species on 


Ocean region and the Neotropics. 


both sides of the tropical Atlantic Ocean. 

The principal explanation for the origin of biotic 
affinities between Africa and the Americas at the level 
of family or + distantly related genera is vicariance 
dating to the separation of the New and Old World 
continents in Jurassic and Cretaceous times (Nelson & 
1981: Wiley, 1988). For more recently evolved 
laxa, one option is direct overland migration available 
After that 


Rosen, 


until about the middle Eocene (see below). . 
time, long-distance dispersal (LLD: Cain et al., 2003: 
henner & Givnish, 2004) becomes progressively more 
important for those capable of crossing significant 
marine barriers; for example, for the 14 polypodiac- 
eous families (sensu lato) listed by Moran and Smith 
(2001) as appearing primarily in the Tertiary. 

In the earlier days of biogeography, long-distance 
transport essentially constituted a non-theory because 
it was invoked to explain how any organism could get 
anywhere given enough time. This was countered by 
the equally fatuous view that vicariance was the 
means of dispersion, and that dispersal biogeography 
was “a 


science of the improbable, the rare. the 


mysterious, and the miraculous” (Nelson, 1978). As 
noted elsewhere, vicariance, as thus proclaimed, has 
certain mystic qualities of its own (Graham, 1999: 


318). More balanced hypotheses tailored to the age. 
morphological features, and systematic relationships 
of individual lineages have been proposed (Givnish & 


Renner, 2004) within a context of temporally better- 


defined tectonic events and = climatic conditions 


(Graham, 2003a, b; 


present symposium). 


It is also increasingly recognized. that patterns of 
Oo. e 


biotic affinities are immensely complicated, and that 


their origin is not likely resolvable by single-factor 


explanations. The most satisfactory models will be 


based on a systems approach, wherein a broad 


spectrum of factors varying in kind and degree. and 


in different combinations over time, form the basis for 


models accounting for the presence of. identical 
closely related organisms in distant geographic 
requisite information includes: (1) 


that 
and the level of relationship: (2) a knowledge of the 
I., 2003); (3) the 


principal transporting vectors (directional atmospher- 


regions. The 


accurately reflects real affinities 


taxonomy 


dispersal processes (Higgins et : 


ic and ocean currents, animal migration routes): (4) an 


adequate fossil record broadly establishing the 


probable time, place, and migration history of the 


lineage; (5) a knowledge of the relevant geologic 
events and landscape configurations during the 


existance of the lineage; and (6) a knowledge of the 


climates and climatic change. at the relevant times. 


The North Atlantic 


across the area being transgressed. 


land bridge became progressively discontinuous from 
the south (the Thulean route) to the north (the 
Svalbard route) after about the middle Eocene 


1999: 61-64). and this was 
at about the same time that global climates at the high 


(Tifíney, 1985; Graham, 


latitudes were trending from the generally tropical 
hothouse interval of the Paleogene toward the 
temperate icehouse conditions of the Neogene. The 
Panama land bridge connecting North and South 
America was discontinuous until ~3.5 Ma (Ma = 


million years, millions of years ago), continuous but 
with lowland habitats suitable for tropical elements 
1997), 
scattered upland habitats available for more temperate 
1992: Burnham & 


Graham, 1999), Thus, the relevancy of these routes to 


until ~2.5 Ma (Coates, and continuous with 


elements after that time (Graham, 


a particular lineage depends on when the migration 
11 place and the ecological requirements of the 
migrants. It is obvious that much of the 1 
necessary for documenting migration routes and the 
origin of disjunct distributions is unknown for most 
organisms, for most places, for most of geologic time. 
Consequently, information that increases the range of 
feasible dispersal opportunities is useful in building 
a database commensurate with the complexity of the 
problem. 

An observation particularly apparent in the satellite 
bacteria, diatoms, 


era is that large amounts of dust, 


phytoliths, fungi, spores, and probably lichen. frag- 


ments are blown into the Caribbean region each year 


via the northeast and southeast trade winds from the 
Sahara Desert and the Sahel, the drought-plagued 
region south of the Sahara in Africa (BioBriefs, 1991; 

1000: Cadée, 1998; 2001; 
Harrison et al., 2001: Kohfeld & Harrison, 2001: 
Prospero. 2001: Prospero & Lamb. 2003: Ryan, 2001: 
2002: Garrison et al., 2003: 2003). 
amounts are impressive—13 million tons/yr. 


L. 2002), 
phosphate/yr./acre 


Reid et al., Arimolo, 


Griffin et al.. Toon, 
The 
(Griffin et a depositing one pound of 


into the Amazon Basin. The 
phenomenon has been recognized since the time of 
that it 
everything on board [the Beagle] and even hurting 
1998: 16). 


atmospheric transport of dust are primarily concerned 


Darwin, who complained was “dirtying 


people's eyes” (Cadée. Studies on the 
with the climatic effects, the bleaching of corals, the 
health. but the 


propagules is 


spread of pathogens, and human 


potential for disseminating small 
obvious, if remote under present conditions. 

However, conditions change, and another observa- 
(Trenberth, 2005) 


r the 


tion is that possibly the frequency 


intensity of hurricanes i 


and certainly the 


Atlantic Ocean and elsewhere has increased in recent 
2002: Otvos. 2002: 


decades (Moreno et al.. Emanuel, 


Volume 93, Number 2 
2006 


Graham 337 


Origin of African Element 


2005). There has been an estimated two-fold increase 
in the number of hurricanes over the past six. years 
compared to the previous 24 years, and the number of 
major hurricanes has increased 2.5 times (Bengtsson, 
2001: Goldenberg et al., 2001: Hoyos et al., 2006: 
Webster et al., 2005). The suspected cause is global 


warming (Knutson et al., 1998; Goldenberg et al., 


2001), 


and the change has been induced by an 
estimated global rise of only 0.6 € for the past century 
and 0.2 C-0.3 C in the last 40 years. An association 


not widely noted in the current literature is that this 


brings the observation on modern hurricanes into the 
realm of paleontological interest, because throughout 
the Tertiary temperatures were generally warmer than 
at present (see global paleotemperature curve; e.g., 
Graham, 1999: 89), and there have been times in the 
geologic past when they were much warmer than at 
present. 
After. the 


Cretaceous Period 65 Ma, climates began to 


asteroid impact at the end of the 
warm 
and reached a threshold between the early Paleocene 
(65 Ma) and the early Eocene (45 Ma) termed the 
LPTM (late Paleocene thermal maximum) and EECL 
(early Eocene climatic optimum), when temperatures 
were as high or higher than at anytime in the 
Phanerozoic Era. (Pomerol & Premoli-Silva, 1986; 
Prothero, 1994; Huber et al., 2000; Schmitz et al., 


2000). There was a rise in the early Miocene and 


extending into the middle Miocene (23-12 Ma with 
a peak at 18-14 Ma), and in the Middle Pliocene (3— 
4 Ma; Wrenn et al., 1999). The temperature values 
were ~12°C-15°C, 4°C-8 
than at present, respectively, in the high latitudes 


2. and 3 CAC warmer 


— 


Graham, 1999: 89), and these periods of unusua 


— 


warmth probably extended somewhat later in the mid— 
latitudes. If global warming is part of the cause for the 
modern rise in global temperatures, then warmer times 
of the past, and particularly these exceptionally warm 
intervals, should have witnessed an enhanced flow of 
directional winds and ocean currents between Africa 
and the Caribbean region, depending on the altered 
continental configurations (mostly in the Eocene) and 
the distribution of high and low pressure systems. 
Two other observations are relevant to the kinds 
and amount of biological debris arriving from Africa 
during past warmer times, and especially during peaks 
of warmth. One is that at the time of the LPTM and 
EECL, the distance between Africa and the South 
America/Caribbean region was one-half to two-thirds 
that of the present. Intervening land may have been 
available along the mid-Atlantic Ridge, or above hot 
spots as at present on the Tristan da Cunha Group 
(oldest of the present islands is Nightingale Island at 
~18 Ma fide www.btinternet.com) in the path of the 


opposite-flowing westerlies from South America to 


Africa. The other is that it was in the Eocene that the 
Greater Antilles began their principal period of 


emergence (Graham, 2003a, b), providing increased 


arget areas for new arrivals. 
There is considerable information available from 
ojects such as PRIDE (Puerto Rico Dust Experi- 


— 


ment) on the chemical composition, source, and 
configuration of dust particles being transported under 
present conditions, and these particles average 30U— 
401 in size (Colarco et al., 2003; Maring et al., 2003; 
5. Reid et al., 2003; J. Reid et al., 2003; R. Reid et 
al., 1996; Prospero et al, 2005). However, the 


maximum size and density of particles recovered is 


not tabulated, although it is probably about LOO (or 
less fide Prospero, pers. comm., 2004). Silica minerals 
ereater than 754 have been found in the atmosphere 
and in water columns transported from eastern Asia 
across the Pacific Ocean (Betzer et al., 1988). These 
minerals and mineral aggregates are far heavier than 
the less dense and often more aerodynamically suited 
biological material, and it may be speculated that 
biological debris to a maximum size of 3004-5004 is 


occasionally transported under present conditions 


— 


and rarely even larger material; see Renner, 2004: 
$31); probably larger material arrives under hurricane 
conditions (perhaps as much as 5004-7004); and still 
larger material was likely transported more frequently 
during times of maximum warmth and greater wind 
velocities in the LPTM and EECL (65-45 Ma). the 
early to middle Miocene (23-12 Ma), and the middle 
Pliocene (3—4 Ma; ?700u-1200u 


needed on the 


— 


. Obviously, addi- 


nalure and 


— 


tional information is 


maximum size and density of material presently being 
transported, as well as comparable data on relevant 
especially during times of 


modern  dissemules, 


maximum wind velocity. 


The point is that, collectively, these factors in the 
geologic past afforded greater possibilities for the 
introduction of a broader array of plants from Africa 
into the Caribbean region, possibly in the opposite 
direction at lower latitudes, across the Pacific Ocean, 
between Patagonia and Antarctica and beyond 
(Iriondo, 2000; Muñoz et al., 2004), and perhaps in 
other parts of the world, via wind, water, and drift than 
would seem plausible under modern conditions. As 
taxonomic relationships become better defined, the 
number of organisms recognized with African-New 
World affinities will likely increase, and the above 
observations of enhanced hurricane intensity, shorter 


distances, and increasing target areas should be 


considered in developing multi-faceted models te 
explain the introduction. of some of them. This is 
particularly true for those with means of long-distance 
dispersal (Givnish & Renner, 2004), such as pieces of 


lichen, the spores of bryophytes and ferns, vegetative 


338 


Annals of the 
Missouri Botanical Garden 


reproductive fragments, and flowering. plants with 
minute, plumose, and/or floating seeds such as in the 
2004). Ginorea (Lythra- 
ceae; S. Graham, 2002), some Myrtaceae (Sytsma el 


Bromeliaceae (Givnish et al., 


al. 2004), Orchidaceae, Piperaceae, and others. 
Fossil and/or molecular evidence also suggests 


correspondence between times of introduction. and 
the exceptional warm intervals of the Tertiary. Period 
23-12 Ma, 3-4 Ma. 


In addition to some Melastomataceae and. Lemnaceae 


centered around 65—45 Ma. and 
(Wolffiella Megelm.) estimated to have dispersed 
between Africa and South America 10-11 Ma (Ren- 
ner, 2004, table 2), 
2004) provides the following examples of plants with 
Old World-New World relationships, and estimates of 


Tremetsberger (pers. comm.. 


the time of separation corresponding to ~3—4 Ma: 


Evidence for long distance dispersal across the Atlantic 


(Asteraceae: two inferred dispersals from the Ol the 
New World; Coleman et al., 2003). The Mediterranean 
and southern African floras were not distinguishable as 


sources of the main New World lineage, estimated to have 
become established during the middle Pliocene (34 Maz 
3 


Coleman et al., 2003). 


Küss and Wink (1997) hypothesized. long distance 
dispersal of Old. World. (Mediterranean and African) 
lupines (Fabaceae) to the eastern parts of South America 
(“Atlantic region”) and, independently, to North America 
and western ri 
the Miocene 


‘gions of ae \merica cus the end of 


(—9 Ma y r dispe "sal to eastern parls o 
South America) and TU ene (~3—4 Ma for dispersal to 
North America and western regions of South America). 

Hypochaeris |. (Asteraceae, Chichorieae) has about 15 
species in the Old World, including Morocco, and more 
than 40 species in South America ds metsberger et al., 

2004). DNA sequencing suggests 
the South American group with a distance 
Atlantic 
evidence, the trans-Atlantic dispersal 
South 


— j or less; 


\frican origin for 
dispersal 
clock 


from Morocco. to 


across the Ocean. Based on molecular 


America can be estimated to have taken place 


1.6. ee Pliocene or Pleistocene 
limes ( Aine aes et al., 2005). 

Among other examples, Davis et al. (2004) suggest 
that in the Malpighiaceae, six amphi-Atlantie. dis- 
Africa 
early 
Miocene). 
Divergence of neotropical and African. Symphonia 
LE 30 + 1.53 Ma (Dick et al., 2003). 


As dh list increases it is worthwhile to monitor the 


junctions occurred. from South America to 


00-31 Ma 


Oligocene) and at 


between (middle Paleocene to 


about 17 Ma (early 


is placed at 17 


examples for estimated times of dispersal, correspon- 
dence with warm intervals of the geologic past, and 
the distribution potential of the propagules, lo 
determine if significantly increased directional winds 
and currents might be a factor in those patterns 
difficult to explain by vicariance or gradual overland 


migration alone. 


Literature. Cited 
Arimoto, R. 2001. Eolian dust ini climate: relationships to 
sources, troposphe ric chemistry, transport and deposition. 
Earth-Scienee Reviews 54: 2 


Bengtsson, L. 2001. Hona A Science 293: 
140—441. 

ds R. P. etal. (+11 authors). 1988. Long-range 1 
of giant 1 e particles; Nature 336: 568-357 1. 

BioBrief 1991. Amazon rain forest sucks up Si haran soil. 


BioScience 41: 439. 
]. & A. 


neotropical vegelation: 


Burnham, R. Graham. 1999. The 
New developments and status. 
nn. Missouri Bot. Gard. 86: 546-589. 

. 1998. Darwin on dust at sea. PAGES: News Int. 


n community O: 


history 


Cadée. G. 6 


Paleosci. 


Cain, M. L., R. Nathan &! . Levin (coordinating/euest 
editors). 2003. Long- its e dispersal. Ecology 84: 
943—2020 
Coates, A. G. (editor). 1997. C entral America, a Natural and 
Cultural History. Yale Univ. Press; New Haven. 


R. et al. (* 
transport to the € 


2003. polit dust 
aribbean during PRIDE: 


Colarco, P. at Bib 

Transport, 
vertical profiles, and deposition in 1 of in situ 

a remote sensing observations. J. Geophys. Res. 108: 6- 

1-6-16. 

J. W. Kadereit & R. J. Abbott. 2003. 


Repeat inter rcontinental dispersal and Pleistocene speci 


Coleman, . Liston, 


ation in sjunc pee “an and desert Senecio 
(Asteraceae >. Amer. J. Bot. 90: 1446-1454. 

Davis, C. C, P. W. » Bell & S. Mathews. 
High- latitude migrations of an 
ea clade: groper ie Malpighiaceae. Int. J. PL 

10 Suppl.) 5107— 

\bdul- i ES E. 


analys 


2004. 


Tertiary exclusively 


2003. 
Tertiary 


TA C. W. K. 
Molecular 


diversification. of ; 


Bermingham. 

ro s reveals cryptic 

wide spa tropical rain forest. tree. 
Amer. P iuralisl 162: 091-703 

Emanuel, 2005. Increasing destructiveness of tropical 
cy 9 5 over the past 30 years. Nature 436: 680— “688. 

Garrison, V. H. et al. (+ 9 authors). 2003. African and Asian 
dust: from desert soils BioScience 53: 
169—190. 

Givnish, P. J. & S 


to coral reefs. 
S. Renner. 2004. Tropical intercontinental 
disjunctions: Conde ana breakup, immigration from the 
boreotropies, and transoceanie dispersal. Int. J. PL Sci. 
165: 51-56 

al. (+ 6 authors). 2004. 


recent long-distance dispersal? Inferences about phylog- 


Ancient vicarianee or 


eny and South American—African disjunclions 


Rapa- 
leaceae 75 Brome liaceae based on ndhk 
Int. J. Pl 35-554. 

Goldenberg. 3 Landsea, A. M 


sequence dala. 
. 105: 8. 


. Mesta-Nunez & W. 


M. Gray. 2001. The recent increase in Atlantic herria 
aclivily: causes p in . ience 292: 474—479 
Graham. A. land bridge 


. Utilization. of the isthmian 
during the ES enozole l 


Sed evide 
and the selective qus nee z a and climate. Rey 
Palynol. 72: 119 

1990, Late 1 5 1 o Cenozoic 
North ‘rican vegetation (north of Mexico). Oxford U niv. 


Palaeobot. 


iistory of 


P ress, New or 
. 2003a. Wu den phytogeography of the Greater 
s. Brittonia 55: 372-398 

2005 


— — — 2003b. Geohistory models and Cenozoic folate. i- 
vironments of the Caribbean region. Syst. Bot. 2 
378-380 


Volume 93, Number 2 
2006 


Graham 
Origin of African Element 


Graham, S. A. 


2002. Phylogenetic relationships and bio- 
seoeraphv of the endemic Caribbean genera Crenea. 
o o | 2 E 


Ginoria, and Haitia (Lythraceae). Caribbean J. Sci. 38: 


195-204. 


Griffin, D. W., . Kellogg. V. H. Garrison & E. A. Shinn. 
2002. The BU transport of dust. Amer. Sci. 90 
228-235. 

Harrison, $ . Kohfeld, C. Roelandt € T. Claquin. 
2001. The ili 0 i Um inc ps changes ae at the last 


Farth-Science Reviews 


e ial maximum and in the future. 
30. 


11 7 S. I. R. Nathan & M. L. Cain. 2003. Are 
e sal events in plants usually 


long- 
distance caused by 
nonstandard means of dispersal? Ec 75 ie 1945-1056. 


Veudelo, P. J. Webs & J. A. Curry 


1 of the fac 


Hoyos, 
2006. 
increase in global 


‘tors pcre ea to the 
hurricane intensity. Science 312 
Huber T, . MacLeod & S. L. Wing 

Warm 5 1 1 ss in Earth History, ris ae niv. 


2000. 


Press, 


editors). 


Cambridge 
M. & J. 


African 


. Dominican amber 
8-10. 
. 2000. Patagonian dust in Antarctica. Quatern. 


Langenheim. 1986 

ancestors. Geotimes 31: 
Int. 6 ) d 86. 

Kiss, E. & M. Wink. 1997. 


phylogeography of Lupinus (Leguminosae) in 


Molecular phyloge ny and 

cis from 

sequences 5 the rbcL gene and ITS 2 

regions of 1 1 P l; . Evol. 2208: 139-16 

Knutson, T. R.. ae va a Y. Kurihara. 1998. Salted 
increase of 1 1 „ane intensities in a C Os-warmed climate 
Science 279: 1018 

Kohfeld, K. E. & 


record of dust. 


nucleotide 


2001. DIRTMAP: The 


Harrison. 
t Earth-Science 


leviews 54: 


Maring T L. Savoie, M. A. Izaguirre & L. Custals. 2003. 

line a dust aerosol size distribution change during 

1 transport. J. Geophys. Res. 108: 8-1-8-6. 
A 


Moran, R. C. & Smith. 2001. 
relationships between neotropical and African-Madagas- 


Phytogeographic 


can pteridophytes. Brittonia 53: 304-351. 

Moreno, A. et al. (+ 6 authors). 2002. 

and high-latitude glacial climatic variability: The 
Res. 58: 318-32 


Saharan dust transport 
Alboran 
Sea rec E 1 8 rn. 


Muñoz, J.. . Felicísimo, F. Cabezas, A. R. Burgaz & 
l. ud E Wind as a long-distance 


pagers il 


vehicle in the Southern Hemisphere. Science 304 


1144-11 


Nelson, G. 19 . From Candolle to C vus comments on the 
history of T geography . Hist. Biol. 209-305. 
D. osen (e Ex 19 9 poe Bio- 


ge T A € ritique. Columbia Univ. Press, New York. 
G. 2002 


Oty E. discussion of “Prehistoric landfall 
fre “quencies of cat astrophio. QNM S... (Liu and Fearn. 
2000). Par ern. Res. 57: 425-428. 

Pomerol, & J. Premoli-Silva ue 1986. Terminal 


b sers Developments in Palaeontology and 


Stratigraphy 9. ESA Amsterdam. 
Prospero, J. M. 2001. o dust in 
(November, 2001), 2 
& P. 100 


America. Geotimes 


2003. African droughts and dust 
Caribbean: climate change implications. 


024—1027. 


transport to the 


Science 302: 


E. Blades. G. Mathison & R. 


=== Naidu. 2005. 
Inte als mispherie transport. o 


iable fungi and bacter 


— 


Africa to the Caribbean with soil dust. 9 

19. 

Prothero, D. R. 10994. The 

paradise lost. Columbia Univ. Press, N 
al. 2003. 


African dust transported to Puerto Rico by individual 


from 
21: 1- 


Eocene-Oligocene transition, 


— 
z 


Reid, E. et al. (+ 7 authors). 5 01 


partic F and size segregated bulk analysis. J. Geophys. 
Res 08: 7-1-7-22. 
Reid. E S. et al. (+ 21 authors). 2003. Analysis of mea- 


surements of Saharan dust by airborne and ground-based 


remote sensing methods during the Puerto Rico Dust 
Experiment (PRIDE). J. 5 . Res. 108: 2-12-27. 
Reid 8 Cary & . Ross. 1996. Late bun rnary 


sedimentation in the epe Antilles island arc. Geol. Soc. 
108: 78-100. 
2004. 


by wind and sea currents. Int. Sci. 165: 


Amer. Bull. 
Renner. S. S. 
Atlantic 
823-833. 


0 mr 


Plant dispersal across the tropical 
J. Pl. 


Givnish (organizers). 2004. Tropical 


intercontine mia disjunctions. Int. J. Pl. Sci. (suppl.) 165: 

SI-S138 

Rouhan, G. 2004. Molecular phylogeny o 
WE (Elaphoglossaceae) and : 


al. (+ 5 authors) 


— 


rds hips be- 


tween species from the n eee and the African-Indian 
20 


Ocean area. Abstract, Botany 2004 1 (Snowbird, 
UT, July ae . 2004): 2. 
Ryan, J. €. 2001. The € 15 san gels dusted. BioScience 51: 
334-338. 
S6 buit B.. B. Sundquist & F. P. Andreasson (editors). 
2000. Early Paleogene Warm Climates and Biosphere 


Dynamics. GFF (Geologiska Föreningen), Vol. 122. Geo- 
logical Society of Swe :den, Stockholm. Sweden. 
Sytsma, J. J. et al. (+ 7 authors). 2004. 


Historical and 


Clades, clocks, and 


biogeographical analys 
Myrtaceae, Vochysiac ee x relatives in the Southe m 
Hemisphere. Int. J. PI. 65: 885 5105. 

Tiffney, B. H. 1985. North. Atlantic land bridge: 
Its importance in Tertiary and modern pario ee of 
the 1 eee sphere. J. Arnold Arbor. 66: 213 

Toon. O. B. . African dust in Florida c NEN 
623-62: 1. 

Tremetsberger, K. « 


continents: 


The E ocene 


Nature 424: 


NW African 


haeris (As 


l. (+ 10 authors). 2004. 

origin of Pa eee an Hypoc steraceae, 

Chichorieae) and their subsequent rapid radiation eluci- 
dated by DNA sequencing and fingerprinting (AFLP). 
Abstract, Botany 1 Meetings (Snowbird. UT, July 31— 
August 5, 2004): 

————— eta 6 "m 2005. 

and karyotypes indicate a NW 


Nuclear ribosomal DNA 
African origin of South 
American 5 (Asteraceae, Cichorieae). Molec. 
Phylogen. e J. 35: 102-116 
905 1 in hurricanes and global 

warming. p ience 308: 1753-1754. 

Webster, P. J., G. J. Holland, J. A. Curry & H.-R. Chang. 

2005. Changes in 1 tropical cyclone numbe sr, duration, and 
Science )9: 


Trenberth, 


intensity in a warming environment. 
1844-1846. 

Wiley. E. O. 1988. Vicariance biogeography. 
Ecol. Syst. 19: 513—542. 

. J.-P. Suc & S. A. C. Leroy (editors). 1999, The 


ino Time of change 


Annual Rev. 


— 


American Association o 


Stratigraphic Palynologists Foundation. Dallas. 


DISPERSAL-VICARIANCE Mathieu Perret,” Alain Chautems,? and Rodolphe 
ANALYSES IN THE TRIBE Spichiger” 

SINNINGIEAE (GESNERIACEA E): 

A CLUE TO UNDERSTANDING 

BIOGEOGRAPHICAL HISTORY OF 

THE BRAZILIAN 

ATLANTIC FOREST! 


ABSTRACT 


The historical bioge ography of the tribe Sinningieae (Gesneriaceae) was analyzed based on distribution data a molecular 
species-level phylogeny of 76 species. This plant group is distributed from Mexico to northern Argentina, but by far the highest 
diversity occurs within the Brazilian in us forest. The dispersal-vicariance analysis method and a c ladistic approach were 
used to infer ancestral areas as well as patterns of dispersal and vicariance. Results indicate that the Sinningieae probably 
arose in the coastal rain forest or in the neighboring tropical area delimited by the Sáo Francisco river in Brazil. The majority 
of the 5 vicariance episodes were reconstructed between the Brazilian Atlantic rain forests and their neighboring 
inland areas (i.e., Paraná and São Francisco regions). In contrast, few dispersal-vicariance events were reconstruc ted between 
the 99 al and A al areas of the Atlantic rain forest and between the Paraná and São Francisco regions. These results, 
together with ancestral areas inferred at the root of the main lineages, indicate an early north-south disjunction in Sinningieae. 
Occurrence of Sinningieae species in other areas, such as the andes; the cerrado, Amazonia, northern South America, and 
Central America, are mainly explained by independent range expansions of single species from either the Paraná or the São 
Francisco regions. According to the dispe rsal-vicariance optimization, 20% of the speciation events are dicens "quent to inter- 
area dispersal and range expansion, 23% are associated with inter-area vicariance, and 57% occurred at a lower geographical 
scale within a single biogeographic area. These results pinpoint the need for a phylogenetic framework to correctly understand 
area relationships and the relative contribution of dispersal and vicariance events in present-day distribution patterns. 

ey words: Atlantic forest, Brazil, dispersal, historical biogeography, Mata Atlántica, semi-deciduous forest, speciation, 
tropical rain forest, vicariance, 


RESUMEN 


5 


analizó la biogeografia histórica de la tribu Sinningieae (Gesneriaceae) sobre la base de datos de la distribución y 
filogenia molecular al nivel de especie de 76 especies. Este grupo de plantas se distribuye desde México hasta el norte de 
Argentina, pero la diversidad más alta se encuentra en el bosque atlántico brasileño. Se usaron el método de análisis de 
dispe rsion-vicarianza y un acercamiento cladístico para inferir las áreas ancestrales así como los patrones de dispersión y 


vicarianza. Los resultados indican que las Sinningicae surgieron ane mente en el bosque hümedo tropical costero o en e 


área tropical vecina delimitada por el río São Francisco en Br reconstruveron la mayoría de los episodios de dispersión- 


vicarianza entre is bosque húmedo atlántico bre sellos ño y sus áreas interiores vecinas (Le., regiones de Paraná y de Sao 
Francisco). En contraste, se reconstruyeron pocos eventos de dispersión-vicarianza entre las áreas tropicales y subtropicales 


giones de Paraná y de Sao Francisco. Estos resultados. junto con las áreas 


i 
eo 
Nn 
= 


del bosque osa allántico y entr 


ancestrales inferidas en la raíz de los linajes o indican una temprana disyunción norte-sur en Sinningieae. La 
presencia de especies de Sinningieae en otras áreas tales como los Andes, los cerrados. la Amazonia. el norte de América del 


Sur y América Central se explica prine ipi almente por e xpansiones indepe eli ntes de la distribución de especies 1 iles 


sea desde la región de Paraná o de Sao Francisco. Según la optimización de la dispersión-vicarianza, el 20% de los 
acontecimientos de espec lación es subsecuente a la dispersión inter-área y a la extensión de la distribución, el 2: a se asocia 
con vicarianza inter-área, mientras que la mayoría de los eventos de especiación (el 57%) ocurrieron en una escala geográfica 
más baja denim de una sola área biogeográfica. Estos resultados indican la necesidad de un marco filogenético para entender 
correctamente las relaciones del área y la contribución relativa de los eventos de dispersión y vicarianza en los actuales 
patrones de distribución. 


Clark, and M. Price for their critical reading of the manuscript and their he mu comments or its improvement. This y bs was 
Su 


"The authors thank C. Chatelain and N. Wyler for their assistance in preparing the distribution maps and L. Schulman. , 
supported by the Swiss National Science Foundation (grant number 31-52378.97), the Swiss Ae aden my of Natural Sciences, the 


» 


Georges and Antoine Claraz Fund, the Augustin Lombard Grant, and the Academic Soci ¡ely of Genev: 
‘Conservatoire & Jardin botaniques de la Ville de Geneve, Impératrice I. C.P. 60, CH- 1292 E E, reneva, 


Swilze ree mathieu.perret&ville-ge.ch 


ANN. Missouri Bor. GARD. 93: 340-358. PUBLISHED ON 23 AucustT 2006. 


Volume 93, Number 2 
2006 


Perret et al. 341 
Dispersal-Vicariance Analyses in Sinningieae 


The Brazilian Atlantic forest is known for its high 


level of species diversity and endemism, as wel 
being one of the most threatened tropical forests in the 


world (Mori et al., 1981; Terborgh, 1992; Morellato € 


Haddad, 2000; Myers et al., 2000). According to 
a broad definition, the Atlantic forest biome is 


composed of two types of vegetation. One is the 
coastal rain forest on the eastern slope of the mountain 
chain that runs along the coastline from southern to 
Brazil. The 
semideciduous forest that extends farther inland in 
Brazil, 
Argentina (Fernandes & Bezerra, 1990; Spichiger et 
al., 1995; Morellato & Haddad, 2000). Despite recent 


efforts to characterize this biome (e.g.. Oliveira-Filho 


northeastern other is a neighboring 


with small portions entering Paraguay and 


& Fontes, 2000). little is known about its bio- 
geographic history. Affinities among the different 


phytogeographic units that compose the Atlantic forest 
and their relationships with other areas in South 
America were generally based on floristice data. 
distribution patterns, or paleoenvironmental evidence 
(Smith, 1962: Harley, 1988; Prado & Gibbs, 1993; 
Oliveira-Filho & Ratter, 1995: Safford, 1999: Oli- 
veira-Filho & Fontes, 2000; 2000: 
Spichiger et al., 2004; 2004). 
Floristic studies indicate a strong link between the 
coastal Atlantic 
semideciduous forests 


2000). On a larger 


in Brazil and Paraguay were shown to be related with 


Pennington et al., 


feo 


Paylor Zappi. 
forests and the 
(Oliveira-Filho & 


scale, 


rain adjacent 
Fontes, 


the semideciduous forests 


other fragments of tropical seasonal forest scattered 


— 


along a large peri-Amazonian are that passes through 
northeastern Brazil, the Paraguay-Paraná river basin. 
the Andes, and northern South America (Prado & 
Gibbs, 1993: Pennington et al., 2000). Other evidence 
based on distributions of closely related species 
indicates biotic exchanges between the Atlantic forest 
and either the Andes or the 
(Rambo, 1951: Smith. 
1999; 2000). 


the direction and f frequency P. range movements are 


Guayana Highlands 
1962; Granville, 1982: Safford. 
Prado, However, the place of origin or 
unknown for most taxonomic groups that occur in the 
Atlantic forest (cf. Berry et al., 2004: Sánchez- 
Baracaldo, 2004; Taylor € Zappi, 2004). It is also 
unclear how vicariance and dispersal have interacted 
to determine current distribution patterns. In this 
tribe of 


distribution data anc 


study, we investigate these issues in a 


a 


Gesneriaceae using detailed 
a nearly complete species-level phylogeny inferred 
from a simultaneous analysis of seven DNA regions 
(Perret et al., 2003). 

The tribe Sinningieae 
Vell. ex 
Vanhouttea Lem., and a total of 81 species (Clayberg, 


1968; Wiehler, 1983, 1984; 1990, 1991, 


includes three genera, 


Paliavana Vand., Sinningia Nees, and 


Chautems, 


as for 


1995, 2002; Wiehler & Chautems, 1995; Chautems et 
al.. 2000). Their habit varies from herbaceous plants 
with a perennial tuber in most Sinningia species to 
woody shrubs without tubers in Paliavana and 
saxicolous, but 
(Chautems & Weber, 


1999). Most of the Sinningieae taxa are found along 


Vanhouttea. They are mainly some 


species are epiphytic or terrestria 


mountain ranges (up to 2200 m) within the Brazilian 
Atlantic rain forest or in the adjacent areas occupied by 
caatingas s.l. of 
2000). the 


and the Misiones 


seasonal vegetation types, i.e., the 


— 


eastern. and northeastern Brazil (Prado. 
cerrado biome in central Brazil, 
nucleus in the Paraguay-Paraná river basin (Prado & 
Gibbs, 1993). A restricted number of species (5) are also 
northern 


found farther along the Andes, in Amazonia, 


— 


Colombia, Venezuela, the Guayana Highlands, Central 


America, and southern Mexico. The Sinningieae is 
therefore an appropriate group for investigating general 


Brazilian 


patterns in historical biogeography of the 
Atlantic forest and its relationships with other areas in 
South Ámerica. 

The biogeographic history of the Sinningieae was 
reconstructed using a phylogenetic framework and 
methods, such as the dispersal-vicariance analysis 
that considers. dispersal and extinction explicitly 
1997). 


assumes no particular speciation mechanism (Bremer, 


(Ronquist, and a cladistic approach that 
: (J) to reconstruct the 

tribe; (2) to 
directions and frequency of dispersal events, espe- 
Atlantic 
between this region and other biogeographic areas in 
South 


dispersal and vicariance processes have interacted 


1992). Our goals are several 


ancestral distribution of the infer 


cially within the domain of the forest. and 


America; and (3) to evaluate how inter-area 


during Sinningieae diversification. 


MATERIAL AND METHODS 


SAMPLING AND PHYLOGENETIC ANALYSIS 


Sixty-two species of Sinningia, six species of 


Paliavana, and eight species of Vanhouttea were 
included in the phylogenetic analysis. This sampling 
represents all currently recognized species in the tribe 
ep cu with the exception of Sinningia helleri 


Nees, S. 


tems, ue S. 


schomburgkiana (Kunth & Bouché) Chau- 


sulcata (Rusby) Wiehler, for which no 


material was available. The two selected outgroups, 


Smithiantha laui Wiehler and Nematanthus villosus 


(Hanst.) Wiehler, belong to the tribes Gloxinieae and 
and are closely related to 


Episcieae, respectively, 


Sinningieae based on a molecular d e 
analysis of the entire family (Smith et al., 1997 
ull list of 


is provided in Appendix I. 


eh 


taxa with author attributions and cee 


342 


Annals of the 
Missouri Botanical Garden 


The plastid DNA spacers trn T-trnl, trnlb-trnF, trns- 
trnG, atpB-rbcL, introns trib and rp! 16, together with a 
portion of the nuclear encoded ncpGS gene were se- 
quenced following the procedure described in Perret 
et al. (2003). All sequences have been deposited 
EMBL/GenBank (accessions AJ438352-AJ438434, 
AJ439249—A 439331. AJ439745—A 439829, AJ439900— 
AJ439984, | AJA87702-AJ487786, AJ459006- 
AJ459691). Because these found. com- 
patible (Per rret et al., 2003), we performed a combined 
PAUP*4.0b8 


50% majority-rule 


regions were 
analysis 5812 characters using 
(Swofford, 1999). We 


bootstrap consensus tree from 1000 bootstrap repli- 


made a 
cales (Felsenstein, 1985) using maximum parsimony. 
tree bisection-reconnection swapping, simple addition 
of taxa (SIMPLE), and a limit of 500 trees retained at 
each replicate (MAXTREES = 


used as a constraint tree in a maximum likelihood 


500). This tree was then 


analysis using the HKY85 model, incorporating rate 
variation across sites and proportion of invariable sites 


(Yang. 1996). 


BIOGEOGRAPHIC DATA AND DELIMITATION OF AREAS 


Distribution data were obtained for all 76 species 


of Sinningieae used in the phylogenetic analy- 


sis. Species distributions were based on over 3000 


accessions from 67 international and Brazilian 
herbaria (i.e. ALCB, ASE, B. BH, BHCB, BOTU. 
CAY. CGE. CEN, CEPEC, CESJ. CONN, E, EAN, 


G. GB, CFP. GUA, HAS, HBR, 
HEPH, HUEFS, IAC, IAN, IBGE, ICN, INPA, IPA, K, 
MBM, MBML. MG, MO, NA, NY, P, PACA, PEUFR, 
PKDC, PY, R, RB, RBR, RUSU, S, SEL, SP, SPF, 
SPSE, UB, UC, UEC, UFG, UFMT, UFP, UPCB, UPS, 
US, VEN, W. WIS. WU, Z). Part of this chorological 


information is available in 


ESA, F, FGO, FUEL, 


various published floras 
and taxomomic treatments for Argentina (Toursarkis- 
1969), Paraguay (Chautems, 1993), and Brazil 
(Chautems, 1990, 1991, 1995, 2002, 2003a, 2003b; 
Woodgyer, 1995; Chautems et al., 2000; Araujo et al., 
2005). Representative specimens for all 
included in this 
Distribution maps were produced using the software 
ArcView GIS 3.2 (ESRI. 2000). The total : 


Sinningieae in South and Central 


sian, 


species 
study are listed in Appendix 2. 
range of 
America was 
partitioned into 12 biogeographic areas (Fig. 1). These 
are: (1) the northern portion of the Brazilian Atlantic 
rain forests that lies north of the Tropic of Capricorn 
(An); (2) the southern portion of the Brazilian Atlantic 
rain forests that lies south of the Tropic of Capricorn 


(As) (3) the São Francisco region (F) 


province s.l. according to Prado, 2000). including the 


(caatingas 


semideciduous forests and the campos rupestre 


distributed east of the Sáo Francisco river in Brazil 


and the caatingas of northeastern Brazil; (4) the 
Paraná region (P) (sensu Cabrera & Willink, 1973), 
including the semideciduous forest of the Paraguay- 
Paraná river system in eastern Paraguay, northeastern 
Mato 
Araucaria forest in 
(6) the Chaco 
(O); (7) the Pampas in southern Brazil and Uruguay 
(U); (8) the southern Andes (Da); (9) the 
Andes (Dx) (10) (11) 
Colombia-Venezuela Guayana 

V): and (l 


(M). We are aware that some areas are composed of 


Argentina, and southwestern Grosso do Sul 


(Misiones nucleus) and the 


southern Brazil: (5) the cerrado (C): 
northern 
Amazonia (Z); northern 
and the Highlands 


2) Central America and southern Mexico 


distinct units that are usually 


phy loge Un! aphic 
considered separately. 


ANCESTRAL AREA RECONSTRUCTION 


Ancestral distributions were reconstructed. on the 
phylogenetic tree using dispersal-vicariance analysis 
(DIVA: Ronquist, 1997), as implemented in the 
program DIVA 1.1 (Ronquist, 1996). This method is 


and 


event-based allows for the inference. of the 


ancestral distribution in a given phylogeny without 


— 


forcing vicariance events to conform to a hierarchical 


pattern. DIVA is thus appropriate to infer biogeo- 
graphic patterns even when area relationships are 


2001). The 


reconstruction of 


reticulated (Sanmartín et al., method 


searches for optimal ancestral 
distribution by assuming a vicariance process bul 
incorporates the potential contribution of dispersal 
and extinction in shaping the current distributional 
pattern. The optimal solutions are those that minimize 
the number of dispersal and extinction events under 
a parsimony criterion. In our analysis, we limited the 
number of areas allowed for an ancestral distribution 


to a maximum of three, using the “maxareas” option. 
DIVA reconstructions were used to determine the 
direction of movement as well as to quantify. the 
biogeographic processes such as dispersal. vicariance, 
and speciation within a single area. 


DIVA, 


Sinningieae and of its major 


In addition to ancestral areas of the 


lineages were also 
determined using the method proposed by Bremer 


(1992). 


which each area is treated as a single character that 


This approach is a eladistic procedure i 


can be optimized on a cladogram. For a particular 


e 


clade. one can calculate the necessary number 


gains and losses of each area by using forward 

reverse Camin-Sokal parsimony, respectively. The 
ratio between gains and losses provides an estimate of 
the likelihood that one particular area is part of the 
ancestral region. The area with the highest gain:loss 
ratio is considered to have 100% probability of being 


part of the ancestral region. 


Volume 93, Number 2 
2006 


Perret et al. 343 
Dispersal-Vicariance Analyses in Sinningieae 


Tropic of Capricorn. 


Tropic of Cancer 


Figure 1. 


Bioge 'ographie areas use «d to delimit S A 'inningle: ac 


l 
the northern and southern portions of the e Atlantic rain forais that lie respective ‘ly north e Stt of the Trop 
= the 


Capricorn; F = the São Francisco region; P = 
southern Brazil and Uruguay; Ds 
Venezuela, a 


5 re 
oes dud s 

nd the Guayana Highlands; 
species 5 in each a 


species distributions in South and Central America. Ax, 
pie of 
gion; C = the cerrado region; O = t he 1 9 in 
northern Andes: Z ree mia; E: — northern Colombia, 


M = Central feda and southern Mexico. Numbers indicate the total number of 
umbers in parentheses indicate the number of species endemic to the area. Delimitations of 


areas were modified after Gentry (1982), IBGE (1993), and Pennington et al. (2000) 


RESULTS 


PHYLOGENETIC ANALYSES 

The phylogenetic tree resulting from the maximum 
likelihood analyses and the bootstrap supports (BS) 
based on maximum parsimony are shown in Figure 2. 
Tribe Sinningieae is monophyletic, but both genera 
Vanhouttea 


and Paliavana are polyphyletic and 


embedded in Taxa are distributed into 
main clades (BS 100%), 
Corytholoma (BS 73%), and Sinningia (BS 6496), as 
well as into two smaller early-diverging clades named 
Vanhouttea (BS 100%) and Thamnoligeria (BS 73%). 
The tree used here is in full 


Sinningia. 


three named  Dircaea 


with our 
previous complementary and comprehensive phyloge- 


2003) 


agreemenl 


netic analyses (Perret et al., 


DISTRIBUTION PATTERNS 


The tribe Sinningieae shows a continental-wide 
distribution between southern. Mexico and northern 
Argentina, but the highest species richness occurs 
within southeastern Brazil (Fig. 1). Out of 76 species 
analyzed, 33 (4396) are restricted to the Brazilian 
Atlantic rain. forest. Within this area, species are 
located either in the tropical zone (22 species) or 
south of the Tropic of Capricorn (10 species), whereas 
only one species is distributed across both zones 
(Fig. 3). Species distributed farther inland are either 
restricted to the areas of Sáo Francisco, the cerrado, 
and Paraná (totaling 23 species, 30%), or occur across 
the tropical portion of the Atlantic rain forest and the 
Sáo Francisco region (10 species, 13%; Fig. 3). The 


remaining 10 species are widespread within Brazil or 


344 Annals of the 
Missouri Botanical Garden 


ue nivalis 
ingia sp. indet. 7 
Sinningia leopoldii 


Sinningia insularis 
Sinningia striata 
„ piresiana 
nningia rupicola 
Sinnincis douglasii 
Sinningia calcaria 
Sinningia mauroana 
Sinningia canescens 
Sinningia macropoda 


Dircaea 


ingi a laterit 
india pl ae 
Sinningia bulbosa 
Sinningia hatschbachi 


A 
: As 
AsP 52 100 AsnP Sinningia cooperi 
As, ASN AN Sinningia 1 
zn 81 An Sinaia Tar ae 


As Sin 1 5 micans. 
Cy, 


pl 


p f : 
È Sinningia lineata 
P Sinningia macrostach ya 
225 pde 
Sinningia E t 
8 PFCUO Sinningia allagophylla 
ee ua 


Sinningia curtifl 
P Sin nningia a amambayensis 
AsNPFC Sinningia aggregat 
PO Sinningia tubiflora 
F Sinningia carangolensis 
AN Sinningia valsuganensis 
F Sinningia araneosa 
ANF Sinningia brasiliensis 
NF Sinningia sceptrum 
VZM  Sinningia incarnata 
TAsNPFCD: Sinningia elatior Hl 
C DSI. Sinningia sp. indet. 1 


Corytholoma 


D 
3 
E 
3 
2 
& 
> 


\N Sinningia barbata 

N Sinningia aghensis 
N inningia pusilla 

N Sinningia concinna 
4 Sinningia richit 


N Vanhouttea ps 
N Vanhouttea lanata 


ISD 
Sinningia hirsuta 
Sinningia gigantifolia 


+ 


Vanhouttea fruticulosa 

Sinningia lindleyi 

Edd 1 
nningia spec 

Sinningia 7 

Sinningia villosa 


AN, ANF,AsF, 
ASN 60 


Sinningia 


We HE We UL NL WI uS AL HE NE 


Sinningia tuberosa 

Paliavana prasinata 
ANF Paliavana tenuiflora 
F po uen sericiflora 


> 
z 


AsNF 
100 


F Vanhouttea hilariana 
Ans Sinningia schiffneri 
ES O sp. in et 4 
ianthus ieee 
mientra laui 


—— 
Vanhouttea 
Thamnoligeria 


areas. The tree is based on 
a simultaneous analysis of plastid and nuclear DNA sequences (regions J . . alpB-rbel, rpl V6, and 
ncpGs). Numbers below the branches are bootstrap proportions (only if = 50%). Ancestral areas at each internal node were 
reconstructed using DIVA (Ronquist, 1997). Abbreviations for cin dde areas are defined in Figure 1. The asterisk 


Figure 2. Phylogenetic hypothesis for the tribe Sinningieae used to reconstruct ancestra 


indicates ambiguous reconstruction of the ancestral area. Names of the five major clades are indicated, and te three 
subelades discussed in this study are identified with Roman numerals. 


Volume 93, Number 2 
2006 


Perret et al. 345 
Dispersal-Vicariance Analyses in Sinningieae 


2 3 areas 
7 


AstAn (1) 


An (22 


— 


F+An (10) 


uA 


Figu ie chart summarizing the number of species 


De 
Né ae to thet distributions: endemic to a single area 


across two areas, or in a combination of at least three areas. 
5 areas are. delineated and abbreviations. are 
defined in Figure 1. 


occur outside in the Chaco (2). the 


northern 


of the country 
Andes (3), 
South America (2), and Central America and southern 
Mexico (2, Fig. 1). Most of 
distributed in a combination of more than t 
(Fig. : 


P: 


Pampas (1). the Amazonia (2). 


the latter species are 


two areas 


atterns of geographical distribution differ among 
the main lineages within Sinningieae (Fig. 4). The 
range of the clade Corytholoma (23 species) covers 
the full do M extent of the tribe Sinningieae 
(Fig. 4A). In Brazil, 


is the highest north a the Tropie of Capricorn within 


species richness of Corytholoma 
the northern Atlantic rain forest and the Sao Francisco 
region. Occurrences outside of Brazil are due to a few 
species distributed in Paraguay and southern. Argen- 


N 


tina (Sinningia aggregata (Ker Gawl.) Wiehler, 5. 


allagophylla (Mart) Wiehler, S. elatior (Kunth) 
thautems, S. tubiflora (Hook.) Fritsch, and 5. 


warmingii (Hiern) Chautems). in the southern Andes 
(S. elatior, S. (Mart.) Wiehler, 
warmingii), in the northern Andes (S. elatior and S. 
Denham), 


America (S. elatior and S. incarnate). and in Central 


sellowii and S. 


incarnata (Aubl.) D. L. in northern South 
America up to southern Mexico (S. incarnata and &. 
All of these species are widespread 
and III 


with the i d of the first 


richii Clayberg). 


and clustered in the subclades H within 


Corytholoma (Fig. 2 


diverging species of the clade, S. richii, which is 


a narrow endemic in the state of Veracruz in Mexico 
(Clayberg, 1968). 

The clade Sinningia (19 species) is distributed from 
eastern to northeastern Brazil (Fig. 4B). The majority 


of these species (12) are restricted to the northern 


Atlantic rain forest, whereas the remaining seven are 


— 


ound in the adjacent Sào Francisco region along the 
Mantiqueira and Espinhago mountain ranges. 

The clade Dircaea (27 species) is distributed from 
LC). Within this 


clade. 16 species have a range limited to the Atlantic 


southern to southeastern Brazil (Fig. 


ain forest, 


whereas the remaining 11 species are 
restricted to either the Paraná (8) or to the southern 
part of the Sáo Francisco region (3). The species that 
Atlantic 
Sinningia macropoda (Sprague) H. E. 


extend the farthest from the coast are 
Moore, which 
1993). and 


magnifica (Otto & A. Dietr.) Wiehler, which is found 


reaches eastern Paraguay (Chautems, 
“Cadeia do Espinhaço” in Minas Gerais 
2005). 


early-diverging 


along the 
(Araujo et al., 


distributed in 


The two clades Vanhouttea 


species) and Thamnoligeria (3) are 
eastern Brazil (Fig. 4D). The species of the clade 
Vanhouttea occur at elevations usually between 1200 
and 2200 m in a vegetation type locally called campos 
de altitude in the southern part of the São Francisco 


region along the mountain ranges of Manliqueira, 


Espinhaço, and Caparaó. Distribution of the clade 


Thamnoligeria is more disjunct, including elements 


from the Atlantic rain forest (Sinningia schiffneri 
Fritsch and an undescribed species of Sinningia, S. 
sp. indet. 4) together with a mountainous species 
(Paliavana plumerioides Chautems), which occur in 
the central part of the “Cadeia do Espinhaço” 


(Chautems, 2002). 


ANCESTRAL AREAS AND DISPERSAL PATTERNS IN BRAZIL 


the DIVA 


Figure 2 and Bremer’s method (Table 1), the ancestral 


According t optimization shown in 


area reconstructed at the root of the tribe Sinningieae 
likely the Atlantic l 


toge ather with the São Francisco region. 


most comprises rain forests 


The ancestral 
area of the clade Dircaea restricted to the 
Atlantic 


Atlantic region was probably part of the ancestral 


was 


southern rain forest, whereas the northern 


distribution of the clade grouping Corytholoma and 


Sinningia (Fig. 2 and Table 1). Subsequent to this 
early north-south disjunction, these two lineages 


followed different biogeographic pathways. In the 
clade Dircaea, dispersal-vicariance events occurred 
predominantly between the southern Atlantic rain 


forest and the Paraná region, whereas in the 
Corytholoma and Sinningia clades, dispersal-vicari- 
ance events were more fre "quent between the northern 
Atlantic rain and the Sáo 
(Table 2). Overall, 22 dispersal events were recon- 


structed between the coastal rain forest and the more 


forest Francisco region 


seasonal vegelalion of neighboring inland areas (1.e., 


Paraná and Sáo Francisco: Fig. 5). Movements inland 


346 


Annals of the 
Missouri Botanical Garden 


Uu cr 
i 


ait T 
; 
» 
e B 
j t e 
mm 
E 
F * & n 
* eh t hes > 
as , ; 
ll >3000 m N 
[I] 2000-3000 m T 
EJ] 1000-2000 m > 
E] 500-1000 m » 
[ ] 0-500 m 
Figure 4. Distribution of i main clades within the Sinningieae as defined in Figure 2. —A, Corytholoma (23 species). 


—B. Sinningia (19 species). —C. Dircaea (27 species). —D. 


from the coast were more frequent than in the opposite 


direction (Fig. 5). In contrast, following a pattern 


parallel to the coast, only seven dispersal events were 


reconstructed between the northern and southern 


Atlantic 


São Francisco regions (Fig. 


rain forest and between the Paraná and the 
5). 


DISPERSAL PATTERNS AT A CONTINENTAL SCALE 


Species found in the Chaco, the Pampas, the Andes, 
South 


America are widespread and clustered in subelades 


Amazonia, northern America, and Central 


H and HI within the Corytholoma clade (Fig. 2, see 
below for an exception). Ancestral areas reconstructed 


at the root of subelades. IH and HI indicate that 


Vanhouttea and Thamnoligeria (7 species). 
expansions of these taxa outside of Brazil occurred 
from two distinct biogeographic units: the Paraná and 
Sao Francisco + cerrado, respectively (Figs. 2 and 5). 
Such dispersal events were optimized on terminal or 
nearly terminal branches, indicating that they most 
likely are recent and concern mainly individual taxa 
(Fig. 2). 


Andean piedmont in the southern Andes is indicated 


The tie between the Paraná region and the 
by the distribution of Sinningia sellowii (Fig. OB) and 
Fig. 6A). The link between the 5 


Francisco region and the Andes through the cerrado is 


Sao 


S. warmingii 


indicated by the range of S. elatior (Fig. 6C), which 
also spreads into Amazonia and northern Colombia 


and Venezuela. In addition to these Brazilian— 


southern Andean connections, a tie between eastern 


Volume 93, Number 2 
2006 


Perret et al. 347 
Dispersal-Vicariance Analyses in Sinningieae 


Table J. Ancestral area 


Bremers (1992) 1 for the Sinningieae and ils main 


reconstruction base on 


clades defined in Figure 2. Ancestral area scores (AA) are 
probabilities that a given area is part of the ancestral area for 


the clade. G = number of necessary gains under forward 


— 


Camin-Sokal parsimony; L = number of necessary losses 
under reverse Camin-Sokal parsimony; AA = G/L quotients 


rescaled to a maximum value of | by dividing with the largest 


able 2. Number of « 


among areas vs 


and dispersal events 
Atlantic 


Sinningieae main clades (inferred from the 


cariance 
the Braun 
structed for the 


DIVA analysis in Figure 2). Values in parentheses indicate 


the number of - dispersal « evenls reconstructed on interna 


branches only. Abbreviations for biogeographic regions are 


defined in Figure 1. 


G/L value. Abbreviations for biogeographic regions are Clades 
defined in Figure !. 
Events Dircaeae  Corytholoma Sinningia 
Areas G L G/L AA Vicarance s between: 
Sinningieae NT 0 2 3 
An 21 25 084 | As 7 i i M 
F 20 31 066 0.77 d M " M 
As 13 25 0.52 0.62 Ay E e " 
P 11 26 042 0.50 lg Ag Q 9 9 
C 7 18 039 0.46 F-P 0 0 
Dy + V+Z+M 3 Ap ex g Pepas bermeen 
Ds 3 18 017 020 yr 110) $49) 9 0) 
Dircaea Ag ee T 5 (4) 3 (1) ) 
5 "M i W | (0) 2 (0) 0 
ds 4 8 05 05 Med | (1) B el 
P 6 12 05 0.5 ÁN © As 2 (1) 0 0 
F 4 11 036 036 3 i a) 2 
Corytholoma + Sinningia 
An 15 14 107 1 
: p 1 9 Mod Out of the 15 dispersals reconstructed on internal 
p 5 12 042 039 branches, 12 (16% of the speciation events) corre- 
Dy + V+Z+M 3 8 037 0.35 spond to range extensions of rain forest lineages into 
As 4 15 027 0.25 more seasonal areas of the Paraná or São Francisco 
D. 3 15 02 0.19 regions. All of these events were immediately followed 
by vicariance that further split the combined area into 
"" ils constituents. These dispersal-vicariance episodes 
Brazil and the Guayana Shield region through 


Amazonia can be also hypothesized based on the 
current. distribution of the sister species 5. sceptrum 
(Mart. 


to the ancestral area reconstruction, the range of 5. 


— 


Wiehler and S. incarnata (Vig. 6D). According 


incarnata expanded in a northwestern direction from 
the Sáo Francisco region (Figs. 2 and 5). Finally, we 
cannot exclude an ancient colonization event through 
Central America according to the basal position of 5. 
richii within Corytholoma and its occurrence in the 
state of Veracruz in Mexico (Figs. 2 and 5). 
RELATIVE IMPORTANCE OF DISPERSAL AND VICARIANCE 

In Sinningieae, DIVA analysis identified a total of 
49 inter-area dispersal events, among which 15 are 
reconstructed on internal branches. Because there are 
75 internal nodes in the phylogenetic tree of the 
Sinningieae, this means that 2096 of the speciation 
events were preceded by range expansion. The total 
number of vicariance events is 17 (2396), whereas the 
remaining 43 speciation events (5796) occurred within 


a single area. 


were reconstructed on deep branches within the main 
lineages as well as on branches closer 


(Fig. 2). 


to the tips 


DISCUSSION 


ANCESTRAL AREA AND DISPERSAL PATTERNS IN BRAZIL 


Dispersal-vicariance reconstruction and results of 
Bremers method indicate that the Brazilian coastal 
Atlantic rain forest and the adjacent Sáo Franciso 
the likely 


Sinningieae. Other plant groups that are reported to 


region are ancestral areas of the tribe 


have diversified in the Brazilian Atlantic. regions 
include Nematanthus Schrad. (Gesneriaceae; Chau- 
1988), Fuchsia L. Quelusia Vand. 
(Onagraceae; 1989; Berry et al., 2004), 
Cattleya Lindl. alliance (Orchidaceae; van den Berg 
& Martins, 1998; van den Berg et al., 2000). 
Ruiz & Pav. Faria et al., 
herbaceous Bambusoideae (Soderstrom & Calderon, 
974), and the Cactaceae 
(Taylor & Zappi, 2004). We hope that the increasing 


tems, sect. 


Berry, 


Aechmea 


2004), 


(Bromeliaceae; 


several genera within 


Annals of the 
Missouri Botanical Garden 


body of monographie works and species-level phylo- 
genetic analyses will soon provide opportunities to 
evaluate the biogeographic significance of our results 
taxonomic 


in relation to other groups. 


Within the three major clades, Dircaea, Coryth- 
loma, and Sinningia, several independent dispersals 
between the Brazilian Atlantic rain forest and 


adjacent inland areas of the Paraná and São Francisco 
J 


regions were reconstructed. These floristic exchanges 


are consistent with the floristic similarities. found 


and semideciduous forests 


Atlantic 


between rain forests 
enclosed in the Brazilian 
Filho & 2000). 


migrations were more frequent from the 


region (Oliveira- 
that 


rain forest 


Fontes, t is worth noting 
to neighboring areas than in the opposite direction 
(Fig. 5). the 
constitute the predominant source of Sinningieae 
the regions. 
According to Oliveira-Filho and Fontes (2000: 808). 
“the tree flora of semideci 


Therefore, coastal rain forest could 


evolution in more seasonal adjacent 


[om 


uous forests is a fraction of 


the much richer rain forest flora, and probably is 


composed of species able to cope with relatively 


seasons. 


longer dry Similarly, in Sinningieae, the 
repeated range extensions into the semideciduous 
forests most likely are correlated with independent 
evolution of lineages that are more tolerant to rainfall 
seasonality. 

Dispersal patterns between coastal and adjacent 
inland areas differ among lineages. In the sister clades 
Corytholoma and Sinningia, dispersal events occurred 
mainly between the northern portion of the Atlantic 
rain forest and the Sao Francisco region (Table 2). 
These dispersal events, as well as the fact that a large 
number of laxa are distributed in both areas, conform 
and climatic 


to the relative gradual floristic eradient 


that exists from the Atlantic coast to the interior above 


the Tropic of 
2000). | 


reconstructed. between the southern portion of the 


Capricorn (Oliveira-Filho & Fontes, 
| Dircaca, 


dispersal events were mainly 


coastal rain forest and the Paraná region. Dispersals 
along this route were also hypothesized for elements 


within the plant families Myrtaceae and Palmaceae, as 


well as for various groups of vertebrates (Mueller, 
1973; Cracraft, 1985; Spichiger et al., 1995). These 


taxa were considered to be expansive rain forest 
elements because they are well represented in the 
Atlantic 


meridional Planalto region and eastern. Paraguay. In 


rain forest and also penetrate into the 
these cases, the coastal mountain range Serra do Mar 
does not act as a hermetic barrier, despite its role in 
creating a sharp transition between subtropical and 
typical seasonal conditions (Oliveira-Filho & Fontes. 
2000). 

The ancestral areas reconstructed at the root of the 


main lineages Dircaea and Corytholoma + Sinningia 


support an early north-south disjunction between the 
subtropical and the tropical portions of the Atlantic 
forests (Fig. 2). The paucity of subsequent exchanges 


be tween northern and southern coastal regions and 


between the Sao Francisco and Paraná regions 
indicates that isolation of these lineages largely 
persisted during their diversification. Only two 
noticeable exceptions were reconstructed on internal 


One 


branches. is the northward shift of subclade 


along the coast (cf. Fig. 2): the other involves 
colonization of the Paraná region from the Sao 


Francisco 
(Fig. 2). 


along a 


region by the widespread. subelade H 
The relatively low frequency of dispersals 
northeast-southwest axis conforms to the 
strong floristic differentiation found between northern 
and southern blocks of Atlantic forests (Oliveira-Filho 
& Ratter, 1995: Oliveira-Filho & Fontes. 2000) and to 
the latitudinal divisions of the coastal rain forest into 
f 1973: 


different centers of endemism (Mueller. Brown, 


1987: Prance, 1987). Similarly, at the populations 
level, recent studies on the genetic differentiation 
of the Brazilian Cherry (Eugenia. uniflora along 


the Atlantic rain forest support the hypothesis that 
a past barrier obstructed gene flow between southern 
Brazil and the rest of the country (Salgueiro et al., 
2004). 


180 


Factors proposed to explain the relative 


— 


ation of tropical and subtropical areas of Brazil 
include mean temperature gradient along the coast 
(Oliveira-Filho & Fontes, 2000), contrasting vegeta- 


tion history on each side of the Tropic of Capricorn 


during the Quaternary (Brown, 1987; Behling & 
Negrelle, 2001). and the barrier effect of the dry 


climate and unforested landscape that locally per- 
sisted in the region of Cabo Frio in the state of Rio de 
1973). In Sinningieae, 


contributed to 


Janeiro (Mueller, these factors 


may have prevent. Sinningia and 


Corytholoma from moving south or Dircaea fron 


moving north. 


DISPERSAL PATTERNS AT A CONTINENTAL SCALE 

Occurrence of Sinningeae outside of Brazil in the 
Andes, Amazonia, northern South America, or bevond 
into Central America and southern Mexico most likely 
is explained by several independent range expansions 
of single species embedded within the Corytholoma 
clade. The analvses of these biogeographic events and 
present-day species distribution provide an opportu- 
nity to infer the pathways along which these areas 
were colonized. 

The four Sinningieae species extending their range 
in the Andes are also found in the Paraná region, the 
Sáo Francisco both Fig. 6). 
Without widespread species are 


clustered in subclades H and HI within Corvtholoma 


region, or in 


regions 


exception, these 


Volume 93, Number 2 
2006 


Perret et al. 349 
Dispersal-Vicariance Analyses in Sinningieae 


P 8 SE S 
5 — Tropic of Cancer 
N D 
`~ e 
BN 
e > = 


Tropic of Capricorn —u— 3 — M È 


E 


Figure 5. Direction and number of dispersal events between biogeographic areas of South and Central America in the 


Sinningieae (inferred from the DIVA analysis on Figure 2). The t 


dispersal events. Number of dispersal events is indic 


ated when greater than | 


rickness of the arrows is proportional to the frequency of the 
. Dotted arrows. indicale range expansion in 


sev p biogeographic areas reconstructed on a single branch of the phylogenetic tree (in subclade III. Fig. 2). Abbreviations 
1] 


for bioge 'ographie areas are def fine «d in Fig Zu re 


(Fig. 2). 


occupied by tropical seasonal forest, 


Distribution of these species in areas 
such as the 
nucleus, the 


foothills 


Misiones Andean piedmont, and the 


eastern and intermontane valleys of the 
Andes, fit with the Pleistocene are distribution pattern 


described by Prado and Gibbs (1993). 


recent. distribution maps available for the Cactaceae 


These data and 


indicate that this pattern may concern a larger number 
of taxa than had been initially thought (Taylor & 
Zappi. 2004). Direction and frequeney of dispersal 
events between the Brazilian Atlantic regions and the 
Andes have rarely been tested using molecular 
phylogenies. However, the few existing studies have 
already indicated that the direction of such dispersal 
events may differ according to the taxonomic group 
(Voelker, 1999; Berry et al. 2004; 
Sánchez-Baracaldo, 2004: Taylor € Zappi, 2004). In 
Sinningieae, dispersal-vicariance analyses support 


three independent migrations into the southern Ándes 


considered 


along two distinct biogeographic pathways (Fig. 


\ 5). In 
subclade IL colonization of the Paraná basin predated 
the independent dispersions of Sinningia warmingii 


The 


distribution of these species suggests the existence of 


and S. sellowii into the southern Andes (Fig. 2). 


a narrow bridge between the Paraná basin and the 
sub-Andean piedmont through the northern extremity 
of the Pe i 
(Fig. OA, B). 


identified as « 


Paraguay and southeastern Bolivia 


This strip of land has already been 


crucial connecting area for several 
woody elements that are found both in the Parana and 
t 


xeromorphic 


— 


1e Andean piedmont areas but that are missing in the 
vegetation (Prado, 2000; 
Spichiger et al., 2004). A second dispersion route is 
indicated by S. 


Chaquean 


elatior, linking the Sáo Francisco 
region with the eastern foothills of the Andes through 
the cerrado, the Pantanal. and Bolivia (Fig. 6C). 


species 


This 


occurs mostly in grasslands on marshy 


grounds at altitudes not exceeding 1200 m. Ancestral 


350 


Annals of the 
Missouri Botanical Garden 


e $ ^ 
n 


3000 m 

[I] 2000-3000 m 
EJ] 1000-2000 m 
L] 500-1000 m 


[C] 0-500 m 


Figure 6. 


1 85 rn South America. Central America. and southern Mex 


—D. S. incarnata (solid circle) and S. sceptrum (solid triangle). 
ö 


area reconstructions within subelade HI indicate that 
expansion of 5. elatior (or its close ancestors) into non- 
Brazilian areas occurred. from the cerrado and/or the 
Sáo Francisco region (Figs. 2 and 5). The westward 
extensions of 5. elatior follow the northern migration 
track described by Prado (2000) for elements of the 
seasonal tropical dry forest that cross the cerrado 


following a net of gallery forest along a northeast- 


southwest are (Oliveira-Filho & — Ratter, 1995). 
Occurrence of S. elatior further north in South 
America most likely is explained by northward 


dispersal along the Andes, although the few speci- 


mens collected in the Brazilian states of Pará and 


sellowii. —€. S. elatior 


Distributions of 1 sad species showing a connection between Brazil, the Andes ss, and/or other regions in 
C. S. elatior 


—A. Sinningia w arming. —B. 5. 


Amazonas may also support a migration route through 
the Amazon basin. 

Dispersal patterns of these three widespread 
species are consistent with the hypothesis about 
historical plant migration routes involving dry sea- 
2000: 


Hence, dispersion of these 


sonal forest formations (Pennington et al., 
2004). 


Sinningieae species along the Andes from either the 


= 


Spichiger el 


Sáo Francisco regions could have been 


Paraná or the S 
promoted by the extension of these formations, which 
and cooler 
2000). 


careful test of this hypothesis would 


may have taken place during the drier 
periods of the Pleistocene (Pennington et al.. 


| lowever, a 


Volume 93, Number 2 
2006 


Perret et al. 351 
Dispersal-Vicariance Analyses in Sinningieae 


tree t 


require a calibration of the phylogenetic ) 
estimate the absolute timing of such dispersal events. 


So far. 


because of the lack of dated fossils or pollen records 


such an attempt led to ambiguous results 
for the Sinningieae. 

In addition to migrations along the Pleistocene arc, 
the sister relationship between Sinningia incarnata 
and S. sceptrum provides evidence for a link between 
the Shield Brazil 
(Fig. 6D). The few known ocurrences of 5. incarnata 


Guayana region and eastern 


within Amazonia in the Brazilian states of Maranhão. 
ará, Amazonas, and Roraima may indicate the route 


followed by this species during its northwestern 


expansion into northern South America and beyond. 
Similarly, geographic disjunction between eastern 


Brazil and the Guayana Shield region was observed 


of closely related 


1993; Taylor 


in different species or groups 
species (Harley, 1988; Prado & Gibbs, 
& Zappi. 2004). According to Prado (2000), migra- 
tions through Amazonia for species of tropical 
seasonal forests could have been promoted by the 
existence of a former dry vegetation belt connecting 


the Atlantic rain 


— 
=) 


the Guayana Shield region wit 
forest in Brazil (van der Hammen & Hooghiemstra, 
2000). Finally, the basal position of S. richii within the 
clade Corytholoma and its endemism to the state of 
Veracruz in Mexico (Clayberg, 1968) imply a puzzling 
this 


distribution and phylogenetic position indicate that 


biogeographic scenario. Indeed. geographic 


a large geographic disjunction between the Brazilian 


and Central America 


— 


Atlantic forest might have 
predated diversification in the clade Corytholoma 


(Fig. 4). This that S. 


constitute a remnant of an early northward migration 


resull suggests richit could 


into Central Ámerica. 


RELATIVE IMPORTANCE OF DISPERSAL AND VICARIANCE 
Diversification in Sinningieae is associated with 
inter-area dispersal (20%). (2396). or 


within-area speciation (5796). The similarity between 


vicariance 


the frequency of dispersal and vicariance events can 
be explained by the alternating occurrence of these 
two processes during the diversification of the group. 
In southeastern Brazil, the majority of these recon- 
structed dispersal-vicariance episodes occurred be- 
tween the coastal rain forest and adjacent inland areas 
characterized by a more seasonal climate (i.e., Parana 
and Sáo Francisco regions). These episodes initially 
consist of range expansions of rain forest lineages 
toward the interior followed by vicariance events that 
subsequently divide the source and the invaded areas. 
According to ancestral area optimization, these series 
of events occurred repeatedly at different heights in 


the phylogenetic tree and, therefore, probably oc- 


curred at different times. Cyclical range expansion- 
vicariance events may indicate the transitory nature of 
the barriers existing between coastal rain forest and 
neighboring inland areas. Climatic changes, especial- 
ly in rainfall seasonality, may have been responsible 
for a succession of periods that either constrained or 
allowed expansion by producing more or less stringent 
ecological conditions toward the interior of Brazil. We 
hypothesize that this alternation between dispersal 
and vicariance along a changing climatic gradient has 
driven diversification of Sinningieae in the Brazilian 
Atlantic forests. Such a mechanism can be related to 
the species-pump hypothesis proposed by Stebbins 
(1974) in order to explain the high species richness of 
tropical forests bordered by tropical plant communi- 
ties adapted to drier conditions. 

In addition to inter-area biogeographic events, our 
results indicate that more than half of speciation 
events (57%) occurred within a single biogeographic 


area. For example, diversification of the Sinningia 


clade took place, in a large proportion, within the 
northern Atlantic rain forest (Fig. 2). To investigate 
the biogeographic circumstances of these speciation 
events, further biogeographic analyses at a finer scale 
are needed. A possible approach is lo compare 
geographic distribution of sister taxa across all nodes 


in the phylogenetic tree (Barraclough & Vogler, 


2000). Preliminary results in this direction indicate 


hat, to a vast majority, diversification within a 


pF 


particular area might be the product of dispersal 
and allopatric isolation operating at a narrow geo- 
graphic scale, in the context of the highly heteroge- 
neous environment that characterizes the Brazilian 
Atlantic forests (Perret et al., unpublished data). We 
are convinced that such phylogenetic approaches at 


— 


ferent geographic scales are necessary to assessing 
South 


di 
hypotheses about biological diversification in 
Whether 


gieae represent a general pattern of diversification can 


America. the results obtained for Sinnin- 
only be tested by evaluating their congruence with 
other plant or animal groups with high diversity in the 


Brazilian Atlantic forest hotspot. 


Literature Cited 


\raujo, A. 0. Chautems V. Souza. 2005 

: sesneriaceae da cadeia 7 : pinhago e em Minas Gerais, 
Rev. EN Bot. * -13 

e lough, T. C. & A ae 2000. 


geographic 4 n 5 speciation from species-level 
134. 


rasil. 
Detecting the 


phylog renies. Nat. 155: 41 
B. s lle. 

Atlantic lowland, s 
Quatern. Res. 56: 
revision n Fuchsia sect. 
Missouri Bot. Gard. 76: 


2001. pic ‘al rain forest and 


Behling, H. & R 


climate dn s of the uthe rn. Brazil, 


I 


during late Quaternary. 89 
Berry, P. E. 1989. A systematic 
Quelusta (Onagraceae). Ann 


584 


„W. J. Hahn, K. J. Sytsma, J. C. Hall & A. Mast. 
2004. Phylogenetic relationships and biogeography of 
Fuchsia (Onagraceae) based on nonc 7 0 nuc "i ar and 


sia 
chloroplast DNA data. Amer. J. Bot. 91: 601—610: 
1992. 


~ 


Bremer, K. Ancestral areas: A 1 1 


tion of the center of Syst. Biol. 41: 
145 
K. S. 19 


ical butte flies. p p. 


origin concept. 


Brown. x rn a evolution of neolrop- 
66-104. in . Whitmore & T. 
Prance (editors), Biogeography and m rnary ions in 


S 


Tropical America. Clarendon Press, Oxford. 
Cabrera, A. L. & A. Willink. 1973. iain ia de América 
Latina. Secretariat General OEA, W ashington, DC 
08 


Révision Loto el 


AZ 
e 


Chautems. A. possibilités 


Phybridations de Nematanthus Schrader (Gesneriaceae), 


genre endémique de la forêt côtière brésilienne. Diss. Bot. 
112: 1-120. 
— ———. 1990. ‘Taxonomic revision of Sinningia Nees: 


Nome nc ‘latural changes and new synonymies. Candollea 
386. 
-— 400 l. 5 revision of Sinningia Nees (Gesner- 
Candollea 46: 411—425. 
Pp. 140 in R. Spichiger & L. 
Ramella (editors). Flora del Paraguay. vol. 
loire et Jardin botaniques, Ville de ( 
1995 


laceae) I: Ne 'w species from Brazil. ( 


<= 

= 
= 
D 


————. ] Gesneriaceae. 
22. Conserva- 
Genève, Gene 

p 


new 


—. 5. Taxonomic revision of ungia 


Int 
(Gesneriaceae) III: New igo from Da] and 


combinations. Gesneriana 1: J. 
2002. New ve 


Candollea 56: 261— 
20( 


— 


jesneriaceae ien Minas Gerais, Brazil. 


——. 2003a. Gesneriaceae. Pp. 75-103 in M. G. 
W Wande rley, G. Shepherd, A. M. Giulietti & T. S. 
Melhem (editors). Flora Fane rogámic a do Estado de Sáo 


RiMa Editora, $ 


Gesneriaceae 


Sao Paulo 


Paulo, vol. 3. : 
Pp. 185-19 


20030. 
Cavalcanti B A. E. 


7 in T. B. 
tamos 4 101 rs). Flora do Distrito 
Federal, vol. 
Brasí (lia 
& A. 


E mbrapa Recursos Genéticos e Biotecno- 
logia, 
ES eber. 1999, Shoot 
architecture in the ne pe val genus Sinningia (Gesner- 
. 305-322 in M. H. Kurmann & A. R. 
The Plant 
— Ks ns, Kew. 
———— . Baracho & J. 
species 5 Sinningia 
- Brittonia 52: 
). 1968. 


Baileya 16: 5- 


and inflorescence 


laceae). 


I 


Hemsley 


(editors), Evolution 2 Architecture. Royal 


A. Siqueira Filho. 2000. A new 


ne riaceae) from northeastern 


No 


Blas hans. n D. Sinh. richii, a new species 
(Gesneriaceae). 5 
Cracraft, J. 1985. Historical biogeography and patterns of 
differentiation within the South American avifauna: 
Amer. Ornithol. Union Monogr: 30: 4e 
ArcView GIS version 3.2. Environmental 


Research Redlands, CA, http://www. 
2006. 


Areas 
of endemism. 

ESRI. 2000. 
Systems 


IZA. 


ia. 
esr.com, ace cessed 10 May 

Faria, A. P. 
re us of Aechmea (Brome 1 eae, pee lioideae) 

Ann. Missouri Bot. Gard. 91: 303-317 

. Confidence 1 5 on a phy loge nies ii 
approach using D bootstrap. Evolution 39: 783- 791. 

Fernandes, A. & I 1990. Estudo Uis 'ográfico do 
Brasil. Fortaleza. 


Vendt € 6. K. Brown. 2004. Cladistic 
and allied genera. 


Felsenstein, J. 19€ 


. Bezerra. 
Stylus Com: 


Gentry, A. H. 1982. Neotropical floristic diversity: Phyto- 
) | ] 
geographical connections. between Central and South 


America, Pleistocene climatic 1 or an accident 
of the \ Missouri Bot. Gard. 69: 


creo 


Andean orogeny? Ann, 
Wl 503 


Safford. H. D. 


Salgueiro, F., D. Felix. J. F. Caldas 


Annals of the 
Missouri Botanical Garden 


Granville, J. J. 1982. Rain E "sl flor: a xeric flora refuges 


in French Guiana. Pp. 15916 . Prance (editor), 
Biological A in the 1 Columbia 
University Press, New York. 

Harley, R. M. 1988. Evolution and distribution of 
Ertope (Labiatae), and its relatives, in Brazil. Pp. 71— 
20 in P. E. 1 & "n lever. (editors), 
Proceedings of a Workshop on RUNE, Distribution 
Patterns. Academia Brasileira de Ciencias. Rio de 
Janeiro. 

IBGE. 1993. Mapa de vegetação do Brasil. Instituto 


Brasileiro Ri Pe e estatística, Rio de Janeiro 

Morellato, I K C. F. B. Haddad. 2000. n The 
Brazilian wn 15 5 32: 786-7‘ 

Mori, S. A.. B. M. K ( Prance. 198 
pallerns b conservation of eastern E coastal 
forest tree species. Brittonia 33: 233-245 

Mueller, P. 1973. The Dispersal 7 5 


300m & C. T. m 


Terrestrial 


Vertebrates in the Neotropical Realm: A E in the 
Evolution of the 5 ‘al Biota and lts Native Land- 
scapes. Junk Publishers, The Hague. 

Myers. N., R. Mittermeier, C. G. Mittermeier. G. 
Fon A J. Kent. 2000. Biodive sy hotspots for 


conservation prioritas: Nature 403: 853-858 
Oliveira-Fil 1995. A study of the 
origin of central Brazilian forests by the analysis of P D 


10. A T. & J. A. Ratter. 


species. distribution patterns, Edinburgh. J. Bot 
141—194. 

—— —— K M. A. Fontes. 2000. Patterns of floristic 
differentiation among Atlantic Forests in southeastern 


bun and the influence of climate. Biotropica : 
93— 


810. 


Pennington, 


D. E. Prado 2000. 
Neotropical se dena dry forests n Mid vege- 
1-273. 


e ndry. 


tation changes. J. Biogeogr. n 27. 
erret, M., A. 


Savolainen. 


& V. 


tribe 


Chautems Spichiger, G. Kite 


2003. 


(Gesneriaceae): H 


N. 
Systematics and evolution of 


Sinningieae vidence from phylogenetic 
analyses of six d pes D na regions and nuclear nepG5. 
PS J. Bot. 145461 

Prado. D. E. 2000. Se 1 dry forests of tropical South 
i rica: From forgotten ecosystems to a new phytogeo- 
eral mg unit. Edinburgh. J. Bot. 57: 437— 

K P. E. Gibbs. 1993 
m de dry 
Missouri Bot. 


— 
a 


. Patterns of species distribution 


seasonal forests. of South America. Ann. 
—927 


Gard. 80: 902 


Prance, G. T. m Biogeography of Neotropical plants. 
Pp. 46-65 in T. Whitmore € G. T. Prance (editors). 


Biogeography and Dieii History in Tropical Amer- 


ica. Clarendon Press, Oxford. 
Rambo, B. 2 51. O elelmento andino no ae riogrande "se. 
R 


Anais Bot. Herb. “Barbosa Roc 


rigues” 3: 7 
Ronquist, a 1996. DIVA, ver. I. J. Users Mess. UD 
www.ebe.uu.se 10 aaa ssearch/diva/diva.html >. access- 
ed 10 May 
icon, oe analysis: A new ap- 


proach to the quantific ation of historical. biogeography. 
Syst. 1 16: 195-2 
1999, Bri azilian Paramo L An introduction to 
the 1 al environment and vegetation of the campos de 
altitude. J. Biogeogr. 26: 693-712. 

. M. Margis-Pinheiro & R. 
2001. Even population differentiation for maternal 
biparental gene marke 


Margis. 


anc 


ers in Kugenia uniflora, a widely 
1 species from the a an coastal Atlantic rain 
forest. Diversity Distrib. 10: 201—210. 


Volume 93, Number 2 
2006 


Perret et a 
Dispersal-Vicariance Analyses in Sinningieae 


Sánchez- Baracaldo, . 2004. Phylogeneties and. biogeogra- 
the ne 9 9 1 cat Jan nesonia and Eriosorus 
(Pteridaceae). Amer. J. Bot. 74— 
Sanmartín, I., H. Enghoff & | ane 2001 
inimal dispersal, vicariance e and diverellicalim in 
5-390. 
wn, C. L. Carroll & D. S 


iaon ip in the pee eae: 


poy of 


. Patterns of 
the 


E 
B 
= 
z 
Es 


Denton. 1997. Tribal 
Evidence from DNA sequences of the chloroplast gene 
€ 


ndhV. Ann. Massa Bot. Gard. 84: 50-66 
Smith, L. B. 196 1 at the flora of — Brazil. 
Contr. U.S. N Natl. Herb. 35: 215-249. 


Primitive forest 


Do 
— 


Soderstrom, T. R. & a Calder on. 1974. 
grasses and evolution of the Bambusoideae. Biotropica 6: 
141-153. 

Spichiger, R., R. Palese, A. Chautems & L. 1995. 

Origin. affinities and diversity hot Ds of the Paraguayan 


Ramella. 


de a loras. Candollea 50: 515-53 
1. Calenge & B. Bise. 2004. TEN zonalion 
in "a Neotropies of tree species s "a 1e 
'araguay- -Paraná aw J. Biogeogr. 31: 1489-15 
Stebbins, G. L. 1974. Flowering Plants, Evolution 1 115 the 
Species | eve m Be ee Press of Harvard University Press. 


. 1999, PAUP*. Phylogenetic Analysis Using 


Methods). vers. 4. 


Sinauer 


setts. 
ylor, N. & D. Zappi. 2004. Cacti of Eastern Brazil. Royal 
Botanic 5 , Kew. 
Terborgh, 902. M Wo nance o of the diversity in tropical 


forests. 91 a 24: 2 


1969, Gesneriáceas Argentines. 
E 
van den Berg, C. & P. S. Martins. 1998. 


Brazilian mb Ca. distribution, mor; mua 


Las 


Toursarkissian, M. 
Darwiniana 15; 
Biogeography 


cal variability, evolutionary and taxonomic conse quences. 
Pp. 315-320 in C. E. B. Pereira (editor), Proceedings of 
the 15th World Orchid Conference (Rio de Janeiro, 1996). 
dapes. d Turriers 

. W. E. Higgins, R. L. Dressler, W. M. Whitten, M. A. 
& M. W. Chase. 2000. A 


(Orchidaceae) base d 


"n renas. Culham 


phylogenetic analysis of Laeliinae 
on sequence data from an transcribed Pn ers (ITS) 
of nuclear ribosomal DNA. Lindleyana 15 : 

van der Hammen, T. & H. Hooghiemstra. 2000. (ia and 

Quaternary history of vege ipo climate, and plant 
diversity in m . Qui . Rev. 19: 725-7 

Voelker, C. ). Dispe c vicariance, and clocks: Historical 
E and speciation in a RUE passerine 

Evolution 53: 1536-155: 


E 


genus (Anthus: Motacillidae). 
Wiehler, H. 1983. A synopsis of the 5 Gesner- 


laceae. Se "es 6: 1-219. 
. 1984. Miscellaneous new species in the Gesner- 
laceae. ae ana 7: 328-347. 
& A. Chautems. 1225. A reduction of Lietzia to 
e Ge ssneriana l: 5-7. 
Woodgye sneriaceae. Pp. 327-328 in B. L. 


St 9 (e editor: Flora of the Pico das Almas. C ae 
Bahia, Brazil. Royal Botanic Gardens, Kew 
Among site rate 1 and ils impact on 


. Trends Ecol. Evol. 11: 367— 


Diamantina, 
Yang, Z. 1996 


phylogenetic analyses 


354 Annals of the 
Missouri Botanical Garden 
APPENDIX l. Species analyzed in this study and vouchers of 6 (G): S. tuberosa Mart.) . Moore. Chautems & Perret 
l 


the specimens sequenced. Acronyms in parentheses indicate 


the herbarium where the vouchers are de ‘posited 


Sinningieae: Paliavana gracilis (Mart.) Chautems, Leoni 
oe (GFJP); P. E Chautems, Chautems 460 (G); 
- Chautems & Perret 00-013 
545 (G): 


> 


^. prasinata (Ker Gawl.) Be 


(G); P. sericiflora Benth., 1 P. tenuiflora 


Mansf., Salviani 1541 (Instituto Plantarum); P. werdermannii 
Mansf.. Carvalho et al. 6552 (CEPEC): Sinningta aggregata 
(Ker Gawl.) Wiehler, Chautems & Perret 97-001 (G); S 


25. allagophylla 
amambayensis 


aghensis : hautems, Pereira et al. 522 (VIES); 
Wiehler, Chautems 325 (O): 
, Salviani 1118 (Instituto 1 S. araneosa 
Chaute ms, Chautems & Perret 00-016 (G): S. 1 Nees 

& Mart.) C. Nicholson, Chautems & UR 99-056 (G); S 
brasiliensis (Regel & Schmidt) Wiehler & eon 
Chautems & Perret 97-002 (G); S. bulbosa (Ker Gawl.) 
Wiehler, Chautems & Perret 97-003 (G); S. calcaria (Malme) 
Cervi et al. AC466 (UPCB): (Mart.) 
Wiehler, Chautems & Perret 97-004 (G); S. Mino. 
Chautems, Chautems 376 (G); S. 1 0 (Lehm.) H. E. 
Moore, Chautems & Perret 97-015 (G); S. cochlearis (Hook.) 


— 


Chautems, 5. canescens 


Chautems, Carvalho et al. AC535 (CEPEC); S. concinna 
(Hook. f.) Hanst., Salviani 1540 (Instituto N S. 
conspicua (Seem.) G. Nicholson, Chautems & Perret 00-008 
(G): S. cooperi i (Paxton) Wiehler, Chautems & Perret 01-032 
C); S. curtiflora (Malme) Chautems, Cervi et al. AC477 
(UPCB); S. defoliata (Malme) Chautems, Chautems & Perret 
01-031 (G); S. douglasii (Lindl.) Chautems, Chautems &. 
Perret 001 (G); latior (Kunth) Chautems, Salviani 
1889 (Instituto Plantarum): S. eumorpha H. E. Moore, 


97-006 (G); S. „ Chautems, 
z j) 


glazioviana (Fritsch 
Chautems, Chinita "ns & Perret 97- s G): S. guttata Lindl., 
Chautems & Perret 01-030 (G); S. e Wiehler & 
Chautems, Carvalho & Chautems 3235 (G): S. hatschbachii 
Chautems, Chautems & Perret 99-065 (G) 0 hirsuta (Lindl.) 
G. Nicholson, Chautems & Perret (G) larae 
ae et al. AC54] (CEPEC): S. incarnata 

. Denham, Chautems & Perret 98-004. (G); S 


1 7 
(Aubl.) 
ps TM line) Chaute 
A : S. lateritia (Lindl.) 
Chautems, Chautems & Barcia 321 (( A ad (Planch.) 
b € Perret 99-054 (G); S. leucotric 18 
. Chautems & Perret 97-008 (G); ; 
T, Chauiems & Perret 97-016 (G); S. lineata 


Imq.) Chautems, Chautems & Perret 97-007 (G); S 
macrophy. Nees & Mart.) Benth. & Hook ex Fritsch 
hum et a 001 (CEPEC); S. macropoda (Sprague) H. E. 
ool Chautems & Perret 97- (G); &. macrostachya 


— 
yl 
ES 
Kho 
T 


P lantarum) S. micans (Fritse E C "haute Ins. ia el S 4C LO: 
(UPC B) 


a 


& ‘Siqueira Filho 6065 04 (Gs. piresiana rm hne) 
Chautems viani 640. (Instituto. Plantarum): S. pusilla 
(Mart) Baill., Chautems & Perret 99-049 (G); S. reitzii 
(Hoehne) L. E. Skog, Chautems & des 98-001 (G); S. richii 


Clayberg. Chautems € Perret 01-029 (6); S. rupicola (Mart.) 
Wiehler, unknown; S. sceptrum Ne Wiehler, Carvalho et 
al. AC528 (CEPEC) S. schiffneri Fritsch, Chautems & Perret 
97-010 (G); S. sellowii (Mart.) Wiehler, Chautems & Perret 
01-028 (G); S. speciosa (Lodd.) Hiern, Chautems & Perret 98- 
003 (G); S. striata (Fritsch) Chautems, Chautems & Peixoto 


306 

00- 17 „): S. tubiflora D ‘ F ritsch, Chautems & Perret 
97- 1): S. valsuganensis Chautems, Chautems & Perret 
99- 0 ) S. villosa Lindl., Chautems & Perret 99-059 (G); 
S. w armingit | (Hie 15 Chautems, Chautems & Perret 97-012 
(G); 5. sp. indet. 1, Romero et al. 1709 (G); S. sp. indet. 4. 
Chautems & a 99-051 (G); S. sp. indet. 7, Cervi et dl. 
AC484 (UPCB): lic qs brueggerí Chautems, 

Carvalho et al. 


Salimena 
SJ); V. calcarata Lem., 
V. fruticulosa 0 
Cen. et T 338 (€) gardneri (Mook.) 
Chautems & Perret 01-024 (C: V. hilariana Chautems. 
Salimena Pires et al. oe V. 
Carvalho et al. AC527 (CEPEC): V. Icon Chautems 
1202 (GFJP): V. pendula Chautems. Leoni 4197 (GEJP ). 
Gloxinieae: Smithiantha laui Wiehler, 
Episcieae: 5 villosus (Hanst.) Wiehler, 
lems e Perret 99-041 ( 


rlaziou ex Hoehne) Chautems, 
Fritsch, 
lanata Fritsch, 

. Leoni 


unknown. 
Chau- 


APPENDIX. 2. 


List of representative specimens for the 76 
species of | | 


Sinningieae USCC lo produce the 


~ 


Figures 3 and Specimens are 
cally by species name. Acron 


the herbarium where the specimens are deposited. 


yms in parentheses indicate 


BR: | Minas Gerais: Anderson 


Paliavana gracilis: 


6875 (F, K. MO. NY, R. UI S); Zrwin et al. 28622a (NY): 
Leoni 2849 (GFJP); Leoni on (GEJP). 
Paliavana plumerioides: BRAZIL. Minas Gerais: Arbo et 


al. 4384 (G, SPF); Davis et al. 2285 (E. UEC). 


Paliavana prasinata: BRAZIL. Espírito Santo: Boone 
6668 (CEPEC. MBL M); Boudet Fernandes et al. 2809 
(MBLM): E & A. Cervi 51354 (CEPEC, MBM); 
Hatschbach et al. 69163 (G. MBM): ue ima 14056 (E, UEC). 


Minas Gerais: Belém 1619 (CEPEC, UB): Forzza et al. 2173 
Hio de e iro: Braga & M. G. 
c (CEPEC. K, MG, NY); 

s RB): Mello Silva & J. R. 
. : Me nm (a a et al. 2642 (G): Pabst 8106 
(HBR 31664): 17 & E. L. Costa 313 (RB); Smith 6473 
(IAN, NY, R, SP, US); Strang 490 (GUA) 

1 wana A BRAZIL. Minas Gerais: pa orson 

5809 (F, MO, NY, UB, US); Anderson 84 ua a MO, 

E. US): 55 et al. A (SPF) ps “a 
115 (G); Gibbs et al. . (MBM, 

Paliavana imis Bn AZIL. Vals de ( 
(054 (CEPEC) Alt 5 (CEPEC pend et al. 2 
(ALCB, CEPE C. K, NU Esteves & > P. Ly) 1 e 
(SPF): Giuletti et al. D 3309 (ALCB. CEPEC. 
22320 (CEPEC, | MO, US): 10 0 et al. 1 (C EPEC, 
US): Santos 2717 (C E PEC, SEI Minas Gerais: Leoni 
5570 (GEJP). Paraíba: Felix s (EAN). Pernambuco: 
Andrade el tal. 68 (G); Felix & G. Dornelas 1120 (EAN 4211); 
Felix & G. V. Do 17 5 1316 (K); Pietrobom- Silv a et E 4607 
(MBM): Sigueira Filho & J. A. Vicente 965 (G. UFI 


a 


¿aro de 


Paltavana i AE BRAZIL. Bahia: Lo el al. 
54363 (HUE oque et al. CFCR 11971 (ES PF) 
„ et al d ur (SPF). Minas Gerais: 1 
CFCR € CEPEC. SPF, US); Pirani i al 2271 (NY, SPF); 
a et al. 2 15147 (BHCB, ESA, SPF 


BRAZIL. 5 Federal: Kirk- 
bride | (US). Irwin et al. 11808 (K, US): 
al. 051 (F. NV. UB). Mato Grosso: rud 
. Mato da osso uio Sul: Hatschbach et al. . 

M. MBM, MO. UEC, US linas Gerais: . 
da 037 (G): É ouem & M. Peixoto 401 ) Gouvea et al. 
(UEC): Melo & L. N. Asuncao 1251 (BHCB): Mota 172 


inn ungla 488 gala: 
»Olás: 


Volume 93, Number 2 
2006 


Perret et a 355 
Dispersal-Vicariance Analyses in Sinningieae 


(G, BHCB 53328); 1 et al. 3846 rie NY, UEC). 
Parana: Cervi et 77 5 G); Cervi et al. 4 (G, NY. 
UPCB); Dusén 157% ) isotype, is holotype): 
Hatschbach 15900 ( MS Hatschbach 26350 (MBM, MO, 
US) Hatschbach 3170 (HBR, MBM, PACA, UPCB); 
Hatschbach 5061 (HBR. MBM, PACA, Jes p B); Medri 
et al. 885 (UEC). Rio de Janeiro: Andrec 542 
(RB, RUSU); Black & Fróes 11324 (IAN); ped 3 1196 
. & Dionisio 456 (G, RB). Santa E a 
Hatschbach et al. 78061 (G); Mattos 12036 (HAS, 
& R. M. Klein 4001 (HBR, 10 R. UC, US); Rei 
Klein 5349 (B, F, NY, R, US); Reitz 2031 (ABR, 

Smith & R. M. Klein 14030 B, BH, F, MO, NY, R, US WIS). 
Sáo Paulo: Ferreira 3177 (GUA); Mattos & Moura 14960 
(SP); Meira Neto 21332 (UEC, VIC); a 676 (SP. SPF, cult. 
11.1952). PARAGUAY. EN ssler. E. 11218 (G). 
E. & R. Degen E aa Q, G, US). 


Paraguarí: V A. et E 1226 65 ( 


= 


Cordillera: Zardini, 


Sinningia ed NSIS a Espirito I Boone 980 
(CEPEC, MBN oled. el a s.n. (UEC 

acia alcoi ARGENTINA. Cha yaco: Meyer, T. 
17211 (LIL). Corrie pes Ibarolla 3808 (LIL, MO). Entre 
Rios: Burkart Pere 2 (SI). Formosa: Morel 5314 (K, LIL) 
A. 6. 0377 (NY). Montes. J. E. 171 
.. Distrito Federal: /rivin et al. 26114 
i MC, MO, NY , R. SEL, UB). Goiás: Chautems 402 (G, 
UFG); Hatschbach & J. M. Silva 59963 (G, UFP); Hatschbach 
42002 (MBM); Melo & França 521 (SPF); Rizzo 8845 (UFG). 
Mato Grosso do Sul: Krapovickas & Cristóbal 34521 (MO). 
Minas Gerais: Atkins et al. CFCR 13793 (SPF); Cardoso 
5691 (R); Davidse et al. 10656A (MO, SP): Hatsc 10 28650 
(MBM); Leoni 1436 (G, GFJP); Macedo 2894 ( (MO, P); Mello 
Silva et al. 1624 (SP, SPF); Williams & V. " 
US) Paraná: Dusén 13456 (C. F, MO. Ny, S. US) 
1 1899 (MBM, US): Hatschbach 8577 (B. F. 
MBM. UPCB); Krapovickas & e 39714 (MBM): 
Aummrow & Graham 2644 (MBM, US) Kummrow 966 
MU io de a Bello, 1 342 (R): Mello Filho, 
: Martinelli, G. 8888 (CEPEC 


Misiones: 


— 


paa el ee s.n. (HAS); Palacios-Cuezzo 1158 (MO): Pivetta 
817 (PACA): Rambo 29391 (ICN): 5 el b 5165 (N 
MO): m 2 HBR. INPA, MBM, PACA); Silveira & 
D. Faria F. 3 (I E Sobral & Esposito 3657 (V. MBM). 
Santa eee A Chautems & R. Reitz 351 (G, 
Klein 3539 1 EA s.n. (HAS 
14019 (B, HBR, US, WIS) iva 4910 (HBR, R, US): 
Smith & R. 7 Klein 10399 (GH, R, S, US); Smith E R. M. 
Klein 11362 (B. NY, R, US); Smith & M. Klein 14060 
(US); Smith & R. M. Klein 7438 (HBR, R, US): Smith & R. 
Reitz 14276 (HBR, NY, US); Smith & - e 8621 (HBR, R, 
110 Smith et al. 9263 (HBR, NY, R, US); Ule 1533 (P. 
CA). São Paulo: Buzato & M. c uma 28017 (G, SPF); 
Canalo & J. R. Carvalho 11601 (UEC); Cuatrecasas 26566 
geilen & L. T. Eiten 2432 (BH, NY, US); Gemtchignicov 
y: Gibbs et al. 1707 (UEC); Grotta s.n. (SPF 
15097); Mimura (NY, SP, US). PARAGUAY. Alto 


Paraná: Schinini, A. 8107 (G). Amambay: Schinini, A. & 
M. Dematteis 33459 (G). Caaguazú: Schinini, A. 5772 (G) 
Caazapá: Soria, N. 4032 (MO). Canindeyú: Hassler, E 


IC, W). Central: Schinini, A. 


. Fiebrig 4626 (G, K). Cordillera: 
Bordas E. & G. Sc hme de 1120 (FCO. US). Guairá: 
Jörgensen, P. 3729 (V, . NY, S). Misiones: Schinint, A. 


& R. Vanni 26078 (G). Eo aguarí: Zardini, H. & T. Tilleria 


30279 (G). Pre sidente ae Hassler, E. & T. o ue 5 7 
edro: Pedersen, T. M. 9 
NY ). Rosengurtt B 5242 


Montevideg: Arechavaleta, J. 53 (G 


A Maldonado: 0 A 
Sinningia 5 PARA( Av. Amambay: 
monis, J. H., J. Hahn & R. Duré Rodas 223 ( 
, PY, RB, US = 
z nningia araneosa: BRAZIL. São Paulo: Marcondes 
Ferreira et al. 783 (SP). 
e barbari BRAZIL. eee Aeg ipd et al. 7002 
(CEPEC): » 1541 (CEPEC, HUEFS 5 990 


(CEPEC); 5 1 & Grupo Pedra 5 1 108 (CEPEC 


CB): Silva 4486 (HUEFS); Thomas = al. 11956 (C : PE C. 
NY): Thomas et al. 9104 (CEPEC( . Espírito Santo: 
Boone 219 (G, MBML); Cola 27 727 O 

Sinningia brasiliensis: BRAZIL. Bahia: Franca 1726 


(HUEFS); Carvalho & G. Lewis 911 (Cl PEC, 
Chautems et al. 226 (CEPEC); França et al. 
Harley et al. 19418 (CEPEC, K); 5 et al. 
(CEPEC); Santos 99 8 (CEPEC, R). Espírito Santo: Ar 
; Kollmann et al. 2722 (G, 


= 
= 
is 
Y 
= 
== 
— 


Y); Giuletti et al. CFC 
SP F): Hatschbach 47816 (MBM. US): Leoni 402 (GFJP); Mota 
1350 (G). Rio de Janeiro: Sucre, D. & Braga 2655 (RB). 
1 Se 11 BRAZIL. Rio de Janeiro: Carauta, P 
RB): Farney, C. et al. 2438 (RB): Martinelli, G. 
et al. ry Smith, " B. 6394 (R, UD); Widgren R 870 
UPS). 

Sinningia 1 BRAZIL. Paraná: Hatse hbach 32. us 
(BH, MBM, . WAG, Z); Tiepolo & Svolenski 682 
(MBM). Sáo 19 Fiaschi & A. C. NASA 221 (G): 

Sinningia canescens: BRAZIL. Paraná: Coates et. al. 
359 (FUEL, G, US): Hatschbach & O. Guimaraes 25477 
(MBM. S. WAG): Hatschbach 1036 (MBM. US): Hash 
15657 (CEPEC, MBM). São Paulo: Barreto et al. . 


zZ 


> 


Sinningia B M BRAZIL. Minas Gerais: aa 
586 (GFJP, US). 

Sinningia coals BRAZIL. Rio de Janeiro: Mark- 
Mur F. 64 (RB). 

Sinningia Lolli BRAZIL. Rio de Janeiro: Mark- 
graf. F. 1 (RB). 


Sinningia conspicua: BRAZIL. Paraná: Dusén 16545 (5). 
Santa Catarina: Smith & R. M. Klein 14168 (R, US). 

Sinningia cooperi: BRAZIL. Espírito Santo: 5 & 
E. Bausen 5670 (MBML). Minas Gerais: Leoni 4950 (GFJP). 
Parana: Hatschbach 12400 (MBM, WAG); Hatschbach 
8578 (F. MBM): Santos et al. 7 
Martinelli, G. & Maas, P. 32 
F, GH, S); Reitz & R. M. 
US). São Paulo: Buzato & M. Sazima 26666 
Sinningia 5 RAZIL. Paraná: 2 
Cervi 332 (G, UPCB, US); Gentry & E. Zardini 49756 ( MBM. 
RBR, US). us p ande 5 Sul: Silveira 10619 (H: s 
Santa Catarina: Reitz & R. M. 3 790 4105 (G, NY, UC, 

1 & R. M. Klein 8270 x NY, S); Reitz C1427 " 1 
V R). Sao Paulo: Miyagi et E 139 (ESA, HRCB, SP, 
: o 

Sinningia a BRAZIL. Goiás: Irwin & Soderstrom 
7371 (BH). > Grosso: Dorrien Smith 308 (K). Minas 
1 0 3882 ( SP. US). 

Sinningia douglasii: ARGENTINA. Misiones: Renvoize, 
S. A. 3245 (K). BRAZIL. Minas Gerais: Leoni 3495 (GFJP); 
Meireles & M. M. F. A. Pereira 660 (UEC); Tamashiro et al. 
1283 (UEC). Paraná: Chagas e Silva & Soares 908 (FUEL. 
MBM); Dusén 7045 (MO, S); Hatschbach 1080 (LIL, MBM, 
US); Hatschbach 22564 (MBM : Hatschbach 32866 (MBM); 


— 


Dusén s.n. ( 


Gerais: 


356 Annals of the 
Missouri Botanical Garden 
Hatschbach 3390 (MBM): Hatschbach 03045 (B. S): Rambo 9892 (PACA): Silveira 3847 (HAS): 


10179 (MBM, NY. 
UEC): Koczicki 267 (MBM); Lindeman & J. H. Haas | 
(NY). Rio de Janeiro: Carvalho. M. & Martinelli, G. 
(CEPEC, G, RB): Martinelli, G. et al. 17779 (RB): Pessoa. 
V. A. 109 (RID. Rio Grande do Sul: Camargo 2606 (B, 
p en d he sn. PACA 30678 (B. PACA): Silveira 
: Waechter 1793 (HAS. 
eur . eed 7070 (HBR): 
Lo 22742 (HAS): Reitz & R. M. 
6 (HBR, NY, | . US); Reitz & x 5 Klein 13532 
(HBR): “Reitz & R. M. Klein 7345 (HBR): Reitz & R. M. Klein 
(0505 (HBR): Reitz & R. M. Klein 9286 (HBR): Reitz 5909 
(HBR); Reitz C762 (HBR, RB); Smith & = M. Klein 13303 
. M. Klein 7338 (HBR. S. US); Smith & R. 
Reitz 8853 (HBR, US). São Paulo: 1 & J. 
Ribeiro 2034 (SP): Carmelo et al. 77 (BOTU, G); Gibbs el al. 
3251 (MBM, NY, I y Leitão Filho 32912 (UEC): Leitão 
Filho et al. 3167 (MBM, NY, UEC): 1 A. 
Custodio Filho 247 (SP): yea et al. 51 (SP). 
elatior: ARGENT Corrientes: | Arapo 
A. et al. 29369 (G, MBM, 11 Entre Ríos: Br 


> 


Neto & 


Sinnin gia 
vickas, 


& 5. Crespo s.n. (Her. Burkart. SD. Misiones: don 
718 (BAFC, LIL, SI. US). BOLIVIA. Beni: Beck. : 13143 
d La Paz: Santa 


d J. C. 9339 (MO. NY. 15 S: 
: Steinbach, F. / (F, G, MO, NY, 
Bang. M. 543 a MO, NY, US). 


US). Yungas: 
. Amazonas: 
; Janssen 209 (INPA. M, 


11 1 19757 (CEPEC, G, K) 
MG, SP); Traill 521 (K). Bahia: Jun 1267 (HUEFS): 
UEM et al. 19757 (CEPEC, K); Harley et al. 22323 


(CEC, K). Distrito Federal: /rivin et al. 8954 (F. K. NY. 
UB. US). Espírito Santo: Folli 2403 (RB). Goiás: Anderson 
6276 (NY): Ferreira et al. 2658 (UFG); Hunt & Pun 5/25 
(SP): Irwin et al. 13506 (UB): Irwin et al. 18839 (F, NY, UB, 
W): froin et al. 24168 (BH, F, TAN NY. US): s et al. 
p (F, NY, UB): Maguire et al. 50981 (NY, UB); Rizzo & 
Barbosa 3354 (UFG); Thomas et al. 5786 (SPF). Mato 
TES Ekman da Silva et al. R 2110 (R 7110): Hatsehbach 
17 (Z); Hatschbach 34143 (MBM): Hoehne 569/Rondon 
3064 (R); Hunt & J. F. Ramos 5725 (K. NY. UC. UB): 
P : & A. Ferreira 5684 (K); Philcox & J. F. Pin 1395 
(K. NY, UB): Pires € M. R. Santos 16587 (MG): Saddi et al./ 
UFMT 213 (UFMT). Mato Grosso do Sul: 
Gerra s.n. (G, RB 102323); Garcia s.n. (UEC 
J. F. Ramos 5959 (K): Rodrigue "s et al, 322 (RB): Sucre 10554 
(RB): Tsugaru & H. A. G 
Gerais: A: 
(R); Cavalcanti et al. CFSC 9108 en : Cordeiro & Simonis 
"M Y. RB 


UB): Irwin et tal. 2 
et al. 26052 (F, TAN 
(G ESJ); Macedo 2213 (MO, US); Pirani et al. 5867 (G. SP 
3 al. 1696 (G); Romero et al. 5147 (l 
et al. HHU 15087 (SPF): Salimena Pires et al. AC-492 
(CESJ. G): Shepherd 193 (UE C): Warming s.n. (US). Pará: 


EC): Sakurai 


Pires et al. 6120 (VAN Paraná: do ach 16227 
(MBM, UPCB); 2 22664 (MBM. NY): Hatschbach 
24125 (MBM, MO, UPCI o 26221 (M. MBM 


MO): Hatschbach 26411 11 Z); Hatschbach 5466 (MBM): 


ri 5060 (MBM, PACA): Hatschbach 615 (MBM, 
PA : Aummrov & J. D. Stutts 1791 (MBM): Reitz & 
Ale 15 5s V. US): Santos et al. 650 (G); Yunigoshi & R. 
Kummrov 4597 (CEPEC., MBM, US). Rio Grande do 
Bueno 4751 (HAS); Jürgens 193 (B); Mattos & Mattos 23178 
(HAS); Mattos et al. 22012 (HAS): Rambo & Schultz 36505 
(ICN): Rambo [37449/347492] (MO): od 19 05 (PACA) 
Rambo 26465 (PACA): Rambo 34749 (PACA, S); Rambo 
38074 (PACA): Rambo 56662 (B. HBR. PAC s Rambo 


Silveira ud (HAS): W aechter 1977 (ICN). Santa Catarina: 
Reitz & R. M. Klein 14405 (R); 18 1618 (HBR. UC. US): 
Reitz C 1406 in H. HBR. NY. UC. US); Senhem A (HBR): 
Smith & R. M. Klein 10577 (HBR, io R, RB, UC, US); 
Smith & R. 1975 6619 (HBR. R. US). m 
1209 (BOTU, G) Filho 2150 (SP, SPSE); 
Paischten & Kummrow 37122 (COL, vod MU): Lutz & 
Lutz 1931 (R); Mimura 271 (SP. US): Muller 32172 (UEC); 
Ratter et al, 4980 (NY, UEC); Souza et al. 7250 ( (HRC B. SP, 
SP F. UEC); Sugivama & Mantovani 128 (SP). 
Gardner 3872 (K). COLOMI 
3/080 (NY). Guaviare: Carde "nas, 
MO). Meta: García. M. C. et al. 524 (US). 
Amambay: Solomon, J. E et al. 7128 (MO, PY) 
Krapovickas, et al. 12559 (UC). € 
2709 (MO). Canindeyú: Schinini, 
(C). Central: Pedersen. T. M. 9496 (K). Er ed 
Pedersen, T. M. 1 Cordillera: Zardini. E. ) 
: 5 . A. 2091 
(US Para aguari: Zardini, E. 
Hassler, 1 1501 (G, GH, K. N P, UC, W). PERU. 
Amazonas: Díaz, C. i d 5 3037 (MO). 
Weberbauer, A. 6152 (F. US 
17278 (US). | e 
Dios: Vuúnez, ‘ 


H 1 1 i ? 31 1 0 


Custodio 


JA. Casanare: Uribe 
D. et al 6490 (C OAH, 
PARAGUAY 


Caaguazú: : 


(G). Misiones: ale : 
6665 (US). San Pe dro: 
Y. 


Cajamarca: 


ES 


: Vargas Calderón. J. 
N. yo (MO). Madre de 
69460 (MO). S 
U RI GUAY 


— 
= 
- 
— 


in: Smith 
1 E 


, VEI J. Apure: 
Williams, 
131788 


z A. 15739 (MO). 
Ste yermark 


Avacun: 
et a 
] Renee. C. 
„ . 9 US 
). Mon: agas: 
Wurdack, J. & Waste. J. 

UE 85 VVV; BRAZIL. 


— 
- 
= 
= 
* 
> 
= 
= 
z 
= 


Maguire. n et aL 3562 
39537 (NY). 
Sáo 1 1 Costa s.n. (R 
03451): San Martin-Gajardo s.n. (UEC 35824). 

nn a gigantifolia: BRAZIL. i. Guillau- 
Leoni 4569 (GFJP) Yamamoto (UEC). Rio 
de Janeiro: Marii nelli. G. et al. 13409 (RB); Martinelli, G. &. 
Farney, C. 8088 (RB); Silva Neto. S. J. et Ro 1185 pes 
Snow. D. 20 (K): Sucre, D. & Braga 2318 (R. RB). Sao 
: Arzolla & C. C. R. de Paula 411 (SPS F) 


Sao Paulo: 


as Gerais: 
mon s.n. (SPFS): 


Sinningia glazioviana: BR: 1 la 2560 


(G): Shirasuna et al. (SP, UEC 
un AZIL. Kuhlmann 


Sinningia guttata: s de Janeiro: 


6056 (RB). 


Sinningia harleyi: BRAZIL. Bahia: Carvalho & A. 
10115 3235 (CEPEC. holotype); Harley et al. 51552 
(HUEFS 

Sinningia hatschbachii: BRAZIL. Paraná: Hatschbach 
2: 50% (MB! Y. UC. WAG). São Paulo: Chautems s.n. 

C-1479 (C). 


n gta hirsuta: BRAZIL. Rio de Janeiro: 
(RB): Vidal H-5544 (R). 


m AZIL. Mendonça 13 (R): 


Farney, C. 
. Gomes 4157 
1 Iarae: Sáo Paulo: 
Silveira 9861 (HAS 

Sinningia inc s bh d 
(IAC. IAN, UB); 
V bred 5 Si 
et al. 14 
13 (G). 1 1 


5 ds Fróes, R. I. 25620 
"s ME C. Cavalcante 52651 (NY), 
1. (C. T 37015); 
7 M. 181 
end . & Eit 
610 (RB). Pe e Pickel 3060 


oraima: Pi C Magalhães 1 A 
(IAN, UB): Ducke As s C. G. RB): Milliken. K 
Bowles 397 (E K. P te NO E. L. 503 We 


Tocantins: E et al. m 3s 6 (INP A. MG, NY. SEL. US). 


Volume 93, Number 2 


Perret et al. 
Dispersal-Vicariance Analyses in Sinningieae 


COLOMBIA. Cauca: Cuatrecassas, 
US). Cundinamare a: Pennell, F. W. 2753 (F, 
o W. 553 (US). 


J. 19538 (F, MO, NY. 
MO. NY. US). 
Meta: Grant, M. L. 9220 

0 (US). COSTA RICA. 


ES 
m 
` 
A 
y 


ys J. ley, K. 2626 (MO). Guanacaste: 

. & Hd re. M. 22177 (F). EL SALVADOR. 
Ahuachapán: Croat, T. 42107 (MO). La Libe pid DUMMIES 
J. D. 427 (MO). Sonsonate: Croat, T. 42259 (MO). 
FRENCH GUIANA. Bassin de l'Oyapock: Prevost, M. F. 
& P. Grenand 2050 (CAY, K). : as du Ha: hi e. J. el 
al. 9780 (P). GUATEMALA. 

E. & J. D. Dwyer 3048 1 0 0 Chiquimula: 
Molina 26774 (V); ss 1467 (F). I 
Goiás: Steyermark, J. ). Jutiapa: Sleyermark. J. 
30366 (F). Quiché: i " 1 1433 (F). Santa Rosa: 
Harmon, W. E. &. Dwyer 3233 (MO). Zacapa: 
Steyermark, J. 29533 (F). GUYANA. Upper Takutu-Upper 
Gillespie. L. J. 1993 (CAY). 1108 55 RAS. 
Peaster C. el al. 12727 (F, MO - Copan: 
: Molina R. A. el ( 


: Harmon, W. 
Molina . & 
Iuehuetenan 


Essequibo: 


Croat, T. 
NY). Mood: 


Ocotepeque: 


aes R. & Molina 31080 ( F, 


Bárbara: Molina R. A. 2 (F). Valle: Molina i 4. d 
Molina 22791 (F, NY). ME 15 ^ C Le Breedlove, * E. 
27103 (MO, NY). Oaxaca: King, M. R. 1958 (NY). 
Veracruz: Moore, E ; ae (BH). NICAR To A. 

Stevens al. 20552 (MO). Esteli: 


Chontales: 
2 are 2 " (MO). 1 Moreno. P. V. 
0). Masaya: Stevens 37 (MO). Mata- 
7710 (MO). ids eh govia: Stevens, 
& Krukoff 1 0 (MO). Rio San Juan: Seymour, F. 
e a. S. B. 6191 (MO). PA NAMÁ. € anal Arcas 
Dressler. R. I. 5037 (F. MO). Coclé: RUN J.L. I : 
MO). Panama: Hammel. B. 3755 (MO). Sl nus 
i: Granville, J et al. 13738 (CAY, US): Old- 
1860 (NY). VENEZUELA. Amazonas: Davidse, 
2/300 (VEN). Aragua: Bunting. G. S. 4518 
Steyermark, J. 131256 (MO). Distrito 
253 (US). Mérida: Steyermark, J. 
Aristeguieta, L. 3206 (NY). 
5360 (MO). 


Sao Paulo: Rossi et al. 435 


= 
~ 


Miranda: 
Tachira: Werff. H. van der & González, A. 
Sinningia insularis: BRAZIL. 
(SP). 
Sinningia kautskyi: BRAZIL. i ph Santo: Chautems 
& M. Peixoto 269 (Herbarium UFES, Vitória, ho otype 
Sinningia lateritia: BRAZIL. Rio de Janeiro: 7 
A. & Barcia 321 (G, R): Martinelli, G. et al. pos (RB). 
Jo leopoldii: BRAZIL. Santa Catarina: Cervi et al. 
AC-483 (G, UPCB); Reitz 4549 ( o US). 
A 5 "ucotricha; BRAZI "araná: 
(Sb 56347). Santa Catarina: Reitz 5 R. 
(HBR). 


Motosima s.n. 


M. Klein 13603 


Sinningia A BRAZIL. Rio de Janeiro: Araujio & 
Vianna 926 (G 
Sinningia 128 BRAZIL. Rio Grande do Sul: 


Hatschbach et al. 78338 (G): Silveira & C. J. Manson 6071 


(HAS). Santa Catarina: Chautems & R. Reitz 348 (G, HBR, 
US). 

Sinningia macrophylla: BRAZIL. Bahia: Thomas et al. 
11936 (CEPEC, NY). 

Sinningia macropoda: BRAZIL. Paraná: Carneiro 317 


(MBM); Chautems et al. 360 (FUEL, G): Hatschbach 1035 

(MBM, US): Hatschbach 45655 (CEPEC. MBM. MO, Z); 

Vie ira et al. 131 (FUEL). Sáo Paulo: SanMartin- O s.n. 

" r 8 PARAGUAY. 
,US 


Paraguarí: Hahn, W. J. 2067 


Sinningia macrostachya: BRAZIL. Rio Grande do Sul: 
Bueno p (F, HAS, US); Chontems & J. Waechter 330 (G. 

CN, US): Costa Sacco 1226 (F, H: Fx HBR, IAN, R); Meyer 
et al. 63 (HAS); Rambo a (B, S): Senhem 3630 (B, 
MBM): Silveira 175 (HA ): Silveira aum (H AS); Sobral & C. 
* 3936 (ICN); 9 et al. 4232 (1C 
al. 1308 (UEC). Santa Catarina: Paci et al. 61239 
(G ): Rei & R. M. Klein 9241 (HBR, R). 


— 


N); Ste lunann et 


nningia magnifica: BRAZIL. Espírito Santo: Hatsch- 
952 j 17696 (CEPEC, MBM, SPF, US); Kollmann & C. Fraga 


3210 (MBML). Minas Gerais: Chautems & M. Peixoto 364 
(G): Cosenza 23462 (CESJ); Dulith 31207 (UEC); Furlan et 
al. CFCR 3061 (CEPEC, SPF f US): m 1651 (FCAB. G 
Irwin et al. d (F, S): Mayo et al. CFSC 
7092 (CEPEC, SPF, US): Oliveira & F. R. Salimena CESJ 
24790 (CE A s. himann et al. 2454 (G). Rio de Janeiro: 
Forzza et al. ra & G. Gottsberger 117- 
16471 (G); Puck 1286 (R). Sáo Paulo: Handro s.n. SP 
un (SP 39421); s 7 847 (HRCB, SP, UEC). 
inningia mauroana: ; 10 

e (MBM. NY. S). s Pa iue: Barros 2250 (G, 
SP): Chautems & M. Peixoto 283 (CEPEC m 6. s 
S): Kirizawa 2787 (SP); Silveira 9862 bach 

Sinningia micans: BRAZIL. São Paulo: Dedecca Kug & 
d s.n. TAC 63. ii (IAC 8338); Wettstein 2 Schiffner s.n. 


WU sintipo a micans Fritsc 
Z 


cp 


= 
— 


Paraná: 


isolypes 


Sinningia nivalis: BRAZIL. Rio Grande do Sul: Bueno 
3062 (HAS) 9 1 56197 (B. HBR). Santa Catarina: 
Hatschbach et al. 78119 (G), Pereira s.n. (ABR, NY, R. US). 

Sinningia nordestina: BRAZIL. Alagoas: Cervi et m 6017 
(C). Bahia: Santos 5 (ALCB). Ceará: Eugenio Leite 1075 
(RB). Par 1 9 Agra et al. 1461 (G, JPB). Pernambuco: 
5 075 [ME 20 (IPA); Pietrobom-Silva et al. 4602 (MBM, 

P): Siqueira Filho & G. Baracho 663/704 (G isotype. 
a y a itp Viana 1026 (ASE): 
SE) 


Sinningia pire. siana: 


Sergipe: Viana 052 


Macedo 275 


BRAZIL. 
56345 Helga de 
). 


Sao Paulo: 
Rechsteineria 


M n 


(ES 
piresiana, Salino 130. ES 
1 gu BR AZIL. Rio de pei ess J. M. 


ires s.n. 


715 (RUSU); Mello Filho, L. E. & Emmerich, M. (R). 
Sinningia reitzii: BRAZIL. Santa Catarina: a R. M. 
Klein 4068 (B. C. HBR holotype, MBM, NY, PACA, R, SP, 


; Silveira 9716 (H = 1 
p oa richii: ME 


T 


. Veracruz: Clayberg 26 (BH). 
Sinningia rupicola: s "e Minas Gerais: Braga el al. 

sn. (G); Irwin et al. 19805 (F, K, MO. NY, UB, US); 

Stehmann & C. E. S. Ferreira 2404 (G). 

7 Bahia: Lewis 737 (CEPEC, 

K). Espírito Santo: Arbo et al. 7720 (CEPEC): Kollmann 64 


(MBML, UEC); Sucre 5585 (CEPEC, RB). a ras Gerais: 
Assis et al. 959 (G); ed 0 18 (GFJP); Leoni 3243 (GFJP): 
Mexia 4343 (F. MO, NY „ US): Moreira & S. H. Borges 


24583 (CESJ). Rio de ~ : Duarte, A. 6298A x RB): 
Laclette, P. 921 (R); E K. & Viera, A. O. S. 26212 
UEC). 


Sinningia schiffneri: BRAZIL. Rio de Janeiro: Silveira. 
N. 9855 (HA 2 Sao Paulo: Cordeiro 2377 (NY, SP) 
Cordeiro et al. 525 (G); Gomes da Silva et al. 277 (RB, SPF); 
Leitão Filho et i 34616 (SP); Wettstein & Schiffner s.n. (WU 
holotype). 

ARGENTINA. Corrientes: Utter & 
Eskuche 


Sinningia 7 
Lonsdale 60-3 1 (K). 

À. 5 a: Beck, 5. G. 935 
Silva, R. et b 2079 qu BRAZIL. 
Hatschbac 74337 (G). 
Hatsc bach pen (MBM, UPCB); 


Misiones: 


Mato Grosso do Sul: 
Parana: Dusén 8944 (S); 
Mans hbach 23136 (F, NY, 


= 


Annals of the 
Missouri Botanical Garden 


MO): pium 31052 (MBM, US, Z); Lindeman & J. 1 
Haas 3271 (K , US): Poliquesi & J. Cordeiro 268 (G). i. 
Grande do 7 ees s.n. (HAS): Bueno 3266 (HAS); 
Bueno 5187 (HAS); Chautems & J. Waechter 373 (G); Eggli 
et al. 2488 (ZSS); Lindman A1297 (S, UPS); Mattos 
Bassan 28573 (MAS); Mattos & N. Mattos 24080 (H; » 
Mattos & N. Mattos 31062 (HAS): Mattos & N. Silve 
23305 (ICN); Mattos & N. Silveira 26437a (ICN); Mattos 4 
N. Silveira 26968 (HAS); Mattos & R. Frosi 23743 (HAS); 
Meyer et al. 65 (HAS); Senhem 3631 (ALCB, B, INPA 
MBM); Silveira & J. Mansan 6065 (MAS); Waechter 1840 
(ICN). Santa Catarina: Chautems & R. Reitz 347 (G. HBR. 
US): Reitz 3785 (ABR, US): Smith & R. M. Klein 13838 (NY, 
R. US). PARAGUAY. Amambay: Brunner, D. R, et al. 956 
(PY). Caninde i s rnández Casas, J. & J. Molero 4195 (G). 
Concepción: J. D. 2704 (S). 57 5 1. 

1800 (MO). Par aguart: Bernardi. L. 18146 (F, G, MO, NY). 
URUGUAY. 

a speciosa: BRAZIL. Bahia: p 1071 i (CEPEC, 
Boone 967 (CEPEC, MBML 5 

Mattos & bat oes 57 (H AS, 
(CEPEC FJI 


"AT 
i 
— 
ae 


^ 
= 
8 
— 
= 
zx 
— 
cw 
= 
a 
— 
mel 
- 


Y) Espírito Santo: 
Martinelli et al. 2206 (US); 
SP). Minas Gerais: Leoni s.n. 


Janeiro: Araujo, D. & Souza, S. R. ps (C Hn Braga 
(photo); Mello Filho, L. E. 2011 (R); Segadas-Viana. F. 1469 


(R); Sucre, D. 9731 (BH, CEPEC). 

Sinningia striata: BRAZIL. Minas Gerais: Krapovickas & 
Cristobal 35450 (MO, NY, US). 

Sinningia tuberosa: PS AZIL. iud Nis rais: 
Renato 927 (BHCB 15943): Irwin et al. 29414 (E 
US): Mello Silva et al. 1555 (RB, SP, 22 ). Rio i Des iro: 
Duarte, A. 1278 (RB). 

Sinningia tubiflora: ARGENTINA, ^ a 0: 8 'hulz 850 (K. 
Stuchert, T. 195 (6). rrientes: 

LIL). Entre E Burkart 2. (SI). 
Maldonado, R. 1731 


pues x 


AL). Córdoba 
Schwarz 8818 (NY, 
Formosa: Morel 1743 (LIL). Santa Fé: 


(K). Santiago del Estero: m s 975 (LP); Maldonado 
Bruzone 1568 (LP). PARAGUAY. Central: Hassler, E. 1 

(C. k. NY, P). Concepción: 1 

Cristóbal 15088 (G). Neembueu: Meyer, T. 1. wi U IL " 
Presidente Hayes: Brunner, D. R. 1334 (PY, ^ San 


5023 (G, NY, S, UC). 


Sinningia valsuganensis: BRAZIL. Espírito Santo: Boone 


1104 (CEPEC, MBMI, MO ) Rossini, A. Chautems & M. 
Peixoto 555 (MBML). 
Sinningia villosa: BRAZIL. Bahia: Jardim 3945 (NY 


Espírito Santo: ae et al. 1683 (G, MBML). 
Andrade & M. A. nie 676 (CEPEC, G). 
ARGENT Castil- 
5). Chaco: PA M 
ee T. M. 
„ LIL). Jujuy: 
al. 1900 (C: Si ever: L. J. s.n. (MO 3099831). 
Tucumán: Olar, D. 53 (NY). BOLIVIA. Santa Cruz: 
Herzog 1170 (G, S. IL W. Z). Tarija: Fiebrig 2109 (B not 
. Minas Gerais: Anderson et al. 36287 (F. 
K, MO, NY. UB, l IS); Melo & J. A. Lombardi 396 (G 
Hatschbach 22893 (F. HBR. K. MBM. MO, NY). 


HUEFS). I 
Minas Gerais: 
Sinningia warmingii: Catamarca: 


(C orrientes: 


"ar aná: 


Rio de Janeiro: Araujio, D. & Maciel, N. C. 5 (GUA). 
Rio Grande do Sul: edet 17 (HAS): us dd 31177 
(MBM, Z): Mattos et al. 2 (HAS); Meyer et al. 198 
(HAS): eon 52123 ae e 5 A, Sy; Dr 6569 (HAS); 
Silveira et al. 1156 (HAS). Santa C alarina: Hatschbach et 
al. 79222 (G); Lourteig 2120 (HBR, P, S, US): 7 9 10 2324 
(HBR, US); Reitz & R. M. Klein 7990 c HBR, NY, UC, US); 
Reitz C411 (HBR, R, RB); Ule 1058 (US). São ae 
Edwall s.n. CGG 5906 (SP). ECU ADOR. Loja: Harling G. & 
Inderson. L. 22594 (GB, MO, US). PARA A Alto 
Paraguay: Mereles, F. & L. Ramella 2886 (G ! Amambay: 


= 
= 


Hassler, E. 8491 (G, NY). Central: Zardini. E. 02 (MO). 
Concepción: Fie rA ig, K. 4608 P GI LO 111150 M). 
Cordillera: Fiebrig, K. 825 (K. E; IBG, P) 


Neembucú: Bernardi. L. 20474 cH Preside nte "pin 
id E. et al. 1 75 (US). San Pedro: Hassler, 
GH, NY, P, S, UC tL. Amazonas: 10 F. H. 
1 F. H. Dobson 816 7 "ONN). Apurimac: Alfaro, R. 
(MO). Cajamarca, Sánchez Vega, Í. et al. 5831 de MO, 
50964 m Diehl, 
Weberbauer. A. 6446 (F). 1 


Pa 
< 


, US). Cuzco: Croat, T. 
Huanea uve x tar 
204 (U 
que: Lis d S 


La Libertad: Vásques, i 6310 
& Cruz, H. 1191 
. Puno: Vargas C. 
Sinningia sp. indet. I: (Minas Gerais: Nakajima 
& , André 2481 (UEC); Simão Picadas CS. Bianchini 1202 
SP). 
Sinningia o indet. 4: BRAZIL. Paraná: Hatschbach 
1, Z). Sáo Paulo: Barbosa et al. 884 (G). 
. indet. 7: BRAZIL. Rio Grande do Sul: 
HAS): Wa a 2282 (ICN). Santa 
7241 (MBM holotype, BH 


Silveira 8824 (G, 
a: Hatschbach et al. 
UC 


Vanhoutte "1 bruce: BRAZIL. Minas Gerais: Brugger & 
24699a): Forzza et al. 1803 (G). 
Vanhouttea ae BRAZIL. t Santo: Brade 
19529 (R, RB): Kollmann et al. 2336 (G. MBM, MBML). 
Gerais: Hatschbach 47674 MS US) Krieger & 
Roth 1044 (R, SP). Rio de Janeiro: Anderson 11719 (NY, 
US): Boudet Fernandes 784 (GUA): Costa et al. 493 n EPEC, 
G, RB, SP): Martinelli & C. Farney 8690 (RB, US 
Vanhouttea fruticulosa: BRAZIL. Rio de 
Chautems et al. 338 (G, RB, US 
co gardneri: BRAZ 


— 


7. SOUZA S.N. 


Minas 


— 


a Rio de Janeiro: Anderson 


11722 (N IS) Oliveira 0 10 UA): Smith 6687 (NY, R. 
SP, US). 

Vanhouttea hilariana: BRAZIL. Minas Gerais: Andrade 
921 (CESJ, RB): Farney & S. A. Geromel 1065 (G); Lombardi 


et al. 4079 (G). 
Vanhouttea lanata: BRAZIL. Rio de Janeiro: 
. 2619 (G); 9 el je 788 (GUA). 
leoni: ZIL. Minas Gerais: 
E M Silva 55446 (G, MBM). 
Vanhouttea pendula: BRAZIL. 
al. 23735 (ESA). Minas Gerais 
(VIC): bb. 
MBM. NY, 


Mello Silva 
Catafa 5 
Espírito Santo: Souza sl 


: Caiafa & M. L. Batista 17: 
M. Hatschbach & J. M. Silva 55448 6 


A PHYTOGEOGRAPHIC 
ANALYSIS OF ARACEAE OF 
CABO CORRIENTES (CHOCÓ. 
COLOMBIA) AND COMPARABLE 
LOWLAND TROPICAL 
AMERICAN FLORAS' 


Marcela Mora,’ Rodrigo Bernal,” Thomas Croat,” 


and Jorge Jácome? 


ABSTRACT 

We studied the Araceae of the 
those of La Selva (Costa 
(Peru). We found 114 native species 
Philodendron (36). v 
< hemiepiphytic. Most of the 


1 14 genera at Cabo ( 


epiphytic or 


Syngonium, Monstera, and wa the latter three containing exclusively epi phytic or hemiepiphytic specie 
vas most similar to that of La Selva, Barro C olorado Island, and Bajo Calima: KENN 


flora of Araceae of Cabo Corrientes 
with Río Palenque was low. Our find; 
Nicaragua-Costa Rica boder and is divi led 


Key words: 


Araceae. aroids, 


RESUMEN 


Estudiamos la flora de Araceae de la región de Cabo Corrientes en la 
Barro Colorado (Panamá), 
14 especies nativas distribuidas en 14 géneros: los géneros más 


flora de aráceas de La Selva (Costa Rica), La Isla de 


or) e Iquitos (Peru). En Cabo Corrientes 


(Ecuac encont 


‘orrien 
which gus 1 for 65% af the species. Sever 
epiphytic species belonged to the genera Anthurium, Philodendron. Rhodospatha. 


ings support Lellingers view that the 
into a northern and a southern flora. 
Chocó biogeographic region. € colombia. species richness, tropic ‘al rainforest. 


ramos | 


Cabo Corrientes region on the Pacific Coast of Colombia and compared its aroid flora with 
Rica), Barro C :olorado Island (Panama), Bajo Calima (Colombia). Río Palenque (Ec uador), and Iquitos 


tes; the largest genera were Anthurium (38 species), and 


ity-four. percent of the species were exclusivelv 


s. The 


:hocó biogeographic region extends to the 


Costa Pacífica de Colombia y la comparamos con la 


Bajo Calima (Colombia). Río Palenque 


grandes fueron Anthurium (38 de ies) y Philodendron (36), los cuales comprendieron 65% del total de especies. Setenta y 


cuatro pore iento de las especies ron exclus 


a los géneros e 5 „ Gerl 


ı flora « 


éner epifitos o hemiept 
E 


géneros 
Isl. de Ban Colorado y Bajo 


raceae 


ivamente el ta 80 he mie 1 La mayoría de espec les e pifitas pe rlenecieron 


Monsterc siendo los últimos tres 


e Cabo 99 5 ‘ntes fue más similar con la de La S 


Stenospermation, 
elva. La 


Calima: la oe con Rio Pale inque fue baja. Nuestros resultados coinciden con los de 


Lellinger que señalan que la región del Chocó biogeografico se extiende hasta la frontera de Costa Rica con Nicaragua y está 


dividida en una flora norteña y una flora sureñ 


Colombia is probably the world's richest, but most 
poorly known, country for Araceae in South Ámerica 
(Croat, 1992). 
mented for other groups of organisms (e.g., McNeely et 
al.. 1990: 1995). 


complex $ the 


Colombia’s biodiversity, well docu- 


Henderson et al., is due to the 


mountain system of country and its 
location on the crossroads between North and South 
America. Croat (1992) has suggested that the richest 
area for Araceae in the country is the Pacific slope of 
the Andes and the adjacent wet lowlands. This area 
comprises most of the Chocó biogeographic region, 
which. as currently defined, stretches from the western 
slopes of the Andes of Colombia and Ecuador to the 


Pacific Ocean and from eastern. Panama to central 
coastal Ecuador (Gentry, 1982). The Andean mountains 
have effectively isolated this region for millions of years, 
resulting in a high degree of endemism for plants and 
(Gentry, 1993). This fact, coupled with a rainfall 


1nimals 


that exceeds five meters per year in many areas (Gentry, 
1986) and the absence of a long dry season, cause the 
Pacific lowlands to be an amazing expression of aroid 
diversity. 

The aroid flora has been extensively documented in 
only one locality in the Pacific lowlands of Colombia 
(Bajo Calima) (Bay, 1996), although a preliminary 


survey of another lowland locality (Bahia Solano) has 


!We thank Fundación Inguedé and the Embassy of the Netherlands in Bogotá for funding field work, 
al de Colombia for supporting M. Mora's and J. 


Sciences of Universidad Nacion 


the Faculty of 


Jácome's visit to MO, Mike Grayum for helping 


with identifications. Dorothy Bay for allowing us to use the Araceae checklist of Bajo Calima, and Gloria Galeano and Julio 


Betancur for eritically reading the manuscript 


? Instituto de Ciencias Naturales, Unive srsidad Nacional de Colombia, Apartado 7495 Bogotá, Colombia. marcelamora un(e 


hotmail.com 


? Missouri Botanical Garden, Box 299, St. Louis, 


P.O. 


Missouri 63166 U.S 


A. 


ANN. Missourt Bor. GARD. 93: 359-366. PUBLISHED ON 23 Aucusr 2006. 


Annals of the 
Missouri Botanical Garden 


been published (Croat, 1992). This paper presents the 


results of a detailed study of Araceae at a second 


lowland locality, the Cabo Corrientes area. located on 
the Pacific coast in the Department of Chocó. We 
compare the aroid flora of Cabo Corrientes with those 
America and 


of other localities in southern Central 


northern South America and discuss the affinities of 


the biogeographic Chocó flora. 


MATERIALS AND METHODS 
STUDY SITE 

The study was centered at El Amargal Biological 
Station, Pacific 


ll km north of Cabo Corrientes. 


located on the coast of Colombia 
The Cabo Corrientes 
area is a small peninsula in the Department of Chocó, 
at the southern end of the coastal Serranía del Baudó. 
The peninsula is 15 km long and ca. 6 km wide and 
ranges in elevation from sea level to about 120 m; its 
terrain is deeply dissected by numerous creeks and 
small rivers, resulting in an abrupt topography of hills 
and ravines covered by scarcely disturbed tropical 
rain forest. The total area of the peninsula is ca. 
8000 ha: our 
coastline to the east, 
4000 ha. 


sampling covered ca. 2.5 km from the 


thus including a total area of ca. 


COLLECTIONS AND COMPARISONS 


We collected Araceae at El Amargal from July 
September 1998, January to March 1999, 
2000. Specimens were collected near the visitors” 


house at the Biological Station, along the trail to the 


lo 


and June 


nearby village of Arusí, along the Arusicito river ca. 


2 km east of the Station, and at the southern end of 
the peninsula near Cabo Corrientes. Canopy plants 
were collected by using tree climbing equipment. 
COL and MO and all 
identifications confirmed by 
(COL) and/or Michael Grayum (MO). 


We compared the aroid flora of Cabo Corrientes 


Specimens were identified 


were Thomas Croat 


with available treatments of aroid floras for five 


La Selva Biological Station in Costa Rica 
Wilbur et al., 1994): 


in Panama (Croat, 1978): Bajo 


localities: 
(Croat & Grayum in Barro 
(BCI) 
Calima on the Pacific 
1996); 


Ecuador 


Colorado Island 


lowlands of Colombia (Bay. 


Río Palenque on the coastal lowlands e 


(Dodson & Gentry, 1978, complemented with 
1996): 


Iquitos Reserves in Amazonian Peru (\ asquez, 1997). 


more recent information, Grayum, and the 
Information for all localities was complemented. with 
recent data from the W^ TROPICOS database 
(2006) at the The 


complete data set for all localities is available upon 


more 


Missouri Botanical Garden. 


request from the first author. 


DATA ANALYSIS 


We compared the aroid floras of all pairs of 
localities using Sgrensen’s index (Mueller-Dombois € 
Ellenberg, 1974). For all localities. we excluded from 
the analysis those species that were introduced and we 
assumed that all unidentified species were not shared 
between localities. Thus, our similarity index for any 
two localities will slightly underestimate similarity if 


some of the excluded species prove to be shared. 
RESULTS 
SPECIES DIVERSITY 


At Cabo Corrientes we found 114 native species of 


Araceae belonging to 14 genera and two introduced 


The 


Anthurium (38 species) and Philodendron (36. spe- 


species (Appendix I). largest genera were 
cies). Together both genera comprise 65% of the local 
aroid flora. Nine of the species were new records for 
South 


species. 


America. Twenty-one species (19%). two sub- 


and one variety were new to science 
(Appendix 1). Of particular interest is the finding of 
a new subspecies of Anthurium eminens, a species so 
far known only from the Amazon basin. 

The pooled data for the six localities gave a total 
Only 


Philodendron inaequilaterum and Syngonium podophyl- 


number of 302 species. two of the species, 


lum, were shared by the six localities; 267 species 


of them at Cabo 


The species-richest area was Bajo Calima, 


occurred only at one locality, 41 


Corrientes. 
followed by Cabo Corrientes and La Selva (Table 1). 


The largest genera for the pooled floras were also 


Anthurium and Philodendron. In all cases Philodendron 
was predominant, except at Cabo Corrientes and Bajo 


Calima. where Anthurium surpassed Philodendron by 


hree and two species, The overall 


respectively. 
predominance of Philodendron over Anthurium agrees 
with Croats (1994) statement that Philodendron is much 
more abundant at lower elevations than Anthurium. 

MI similarity indices between localities were low 
(Table 2). 


Palenque and Iquitos with the other localities were the 


Similarities in species richness of Río 


lowest. The highest similarity between two localities 
was between La Selva and BCI and between La Selva 
and Cabo Corrientes, although these figures were nol 
much different from the similarity indices of Cabo 
Calima and BCI. 


similarity of Bajo Calima with any of these localities. 


Corrientes. with Bajo In. contrast, 


except with Cabo Corrientes, was low. 


GENERIC DIVERSITY 
Of 19 genera composing the pooled floras, most (11) 


were shared by all six localities. whereas onlv a few 


Volume 93, Number 2 
2006 


Mora et al. 


361 


Araceae of Cabo Corrientes, Colombia 


Table J. Generic and specific richness of native Araceae in six neotropical lowland localities. 

2 = E 8 = S Y = = = 8 = Jes E = = E S 

5 3 & B 8 $i P f: $ 3 B boS 2 P & 4 $ 

E E 5 = E E E 5 s * & $ 32 & 5 & A 3 

* — E S $ $ Bb ô E P$ E 

8 a = 3 = x 3 À E S ^ E 
E 5 a Š 

La Selva 25 l 0 7 10 0 30 0 2 0 5 2 7 4 97 
Barro 
Colorado 13 1 0 3 | 0 1 3 1 17 1 2 0 2 l 2 0 3 51 
Island 
Cabo 38 1 2 4 1 1 3 6 36 0 5 0 4 4 7 0 2 114 
Corrientes 
Bajo " F 4 " 
ar 49 2 l 3 ] 0 5 0 47 0 3 0 2 14 3 0 4 135 
Calima 
Rio 14 1 2 » -0 4 0 1 0 2 0 | 4 0 | 83 
Palenque 
Iquitos 20 1 0 8 1 4 2 6 1 28 l l 1 | l l l 2 80 
(5) occurred at only one or two localities (Table 1). witchcraft. Additionally, two common, well-known, 
The number of genera at Iquitos (17) was higher than introduced species, Xanthosoma sagittifolium (“oto”) 
that of all other localities (either 13 or 14 genera). and Colocasia esculenta (“achin”) are cultivated in the 


Chlorospatha was the only genus present at the two 


Colombian sites and at the one in Ecuador, but 
lacking at the other localities. 
GROWTH HABITS 

The aroids of Cabo Corrientes are terrestrial. 


epiphytic, or hemiepiphytic (Table 3). Of the total 
number of species, 33 (28%) were true epiphytes. 52 
(44%) were true hemiepiphytes, 29 (26%) were 


terrestrial, and two (296) were either hemiepiphytes 


— 


or terrestrial. Of the 14 native genera found at the 
study site, seven were exclusively terrestrial and four 
were totally epiphytic (true epiphytes or hemiepi- 
phytes). The only genus that showed all three different 


habit types was Anthurium. 


USES 


— 


the native species are used locally by 


African-Colombian people for medicine. construction, 


Five of 


Y Anthurium 


or magical-religious purposes. Leaves 


seca”), and Dracontium spruceanum are used for 


trilobum (“tres dedos”), Anthurium panamense (“hoja 


treating snake bite. The aerial roots of Heteropsis 
) l 


house beams and are reputed to be the best binding 


oblongifolia (“piquigua”) are used as cords for binding 
material in the region. As a consequence of pressure 
on this resource, populations of Heteropsis are now 
severely depleted. Live plants of Dieffenbachia long- 


ispatha (“hoja de chucha”) are planted at one of the 


n^ 


front corners of dwellings as a protection against 


nearby village of Arusí. Their stareh-rich tubers are 


used for human and animal food. 


DISCUSSION 


The number of aroid species in the Cabo Corrientes 
area is higher than the total number of species listed 
for this family by Forero and Gentry (1989) in the 
whole Chocó Department, even though their list 
includes the western slopes of the Cordillera Occi- 
dental. This fact stresses the poor knowledge of the 
aroid flora of the Chocó Department. Thus. for 
example. Grayum (1996) 


Philodendron subg. Pteromischum (Schott) Mayo in 


noted a marked absence of 


this department and suggested that there might be 
a discontinuity of distribution of the group in this area. 
However, at Cabo Corrientes we found ten species of 
this subgenus, which represents 45% of all the 
species of this group known in Colombia. Of these 


ten species, Philodendron croatii and P. rayanum are 


— 


new records for the South American flora. 

The poor knowledge of the Araceae of this region 
reflects both the incomplete exploration of the area 
and the bias of recent inventories, most of which have 
studied forest composition in terms of trees, ignoring 
the diversity represented in other growth forms, as 
pointed out by Galeano et al. (1998b). The latter 
authors found that at Cabo Corrientes and nearby 
areas the Araceae were by far the most species-rich 
family, accounting for 10.3% of all vascular plants. 


Not surprisingly, the biogeographic Chocó has been 


Annals of the 
Missouri Botanical Garden 


D 


Table 2. 


Serensen's index matrix for paired comparisons of aroid species richness at six neotropical localities. 


Locality La Selva BCI Cabo Corrientes Bajo Calima Río Palenque Iquitos 
La Selva | 0.392 0.391 0.216 0.160 0.124 
BCI 0.392 l 0.329 0.150 0.154 0.183 
Cabo Corrientes 0.391 0.329 l 0:355 0.205 0.114 
Bajo Calima 0.216 0.150 0.355 0.191 0.056 
Río Palenque 0.160 0.154. 0.205 0.191 0.105 
Iquitos 0.124. 0.183 0.114 0.056 0.105 l 


pointed to as the world’s richest region for Araceae 
(Croat, 1992) 

The richness of Araceae in this region is probably 
due to a combination of a high level of endemism and 
the great affinity of its flora to that of southern Central 
America. The high level of endemism in the Chocó, 
estimated for all plant groups to be near 20% (Gentry. 
1982), the 


number of new species found in 


is evidenced. for Araceae by the large 
the 
studies so far carried out in the lowlands (Bay, 1996; 
Croat & Mora, 2004; 2006): 37% of the 
205 species in the pooled floras of Bajo Calima and 
Cabo Corrientes were new to science. The f 


most of these species have not been found before in 


two detailed 
Croat et al., 
act that 


better-explored Costa Rica or Panama suggests that 
they may be endemic to the Chocó. 

The affinity of the Choco flora with that of southern 
Central America has been shown by authors 
1982; Croat, 1992; Galeano, 1992; 
Bernal & Galeano, 1993; Grayum. 1996; Galeano et 
19984). fact, the 


Corrientes are also in Central 


many 


(e.g. Gentry, 


al., In 73% of aroids of Cabo 


found America, and 
48% of all aroid species are known only from Central 
the 


published data). 


America and Chocó region (Mora et al, un- 


Serensen's similarity indices for the 


the Araceae of Cabo 


All species within each genus 


habit of 


Corrientes; Chocó, Colombia. 


Table 3. Growth 
have that dee cular growth habit unless otherwise noted with 
the numbe One of 


oL Sp 
Philodendron and one of Rhodospatha are either terrestrial 


ecies in parenthesis. species 


or hemiepiphitic. 


Terrestrial Hemiepiphytes Epiphytes 


Inthurium (8) IRURE (1) Anthurium (29) 


Caladium Stenospermation 
Philodendron (32) 
Rhodospatha (6) 


Syngonium 


Spathiphyllum 


Xanthosoma 


six aroid floras (Table 2) further stress this affinity: 
the similarity of Cabo Corrientes with La Selva and 
BCI is comparable to that between La Selva and BCI. 
Interestingly, Río Palenque and Bajo Calima seem to 
have a lower affinity with Central America. In fact, the 
similarity indices between La Selva and other sites to 
the south showed an abrupt reduction at Bajo Calima, 
despite its closeness to Cabo Corrientes. On the other 
hand, Río Palenque showed a low similarity with all of 
the above-mentioned localities, despite its purported 
affinity to the Chocó biogeographic region (Gentry, 
1982. 1986): the Río 


Palenque and other localities were comparable 


similarity indices between 
those of Iquitos. 
The the 


flora with that of southern Central America suggests 


strong affinity of Cabo Corrientes aroid 
that the northern limit of the Chocó biogeographic 
region does not lie near the Panama-Colombia border. 
1986, 1993) and 


currently recognized, but the region may extend north 


as suggested by Gentry (1982 


to the border between Nicaragua and Costa Rica, as 
initially stated by Lellinger and de La Sota (1978). On 
the other hand, the dissimilarity of Bajo Calima and 
of 


aroids, supports Lellinger's (1975) distinction. based 
PI 8 


lerms 


Río Palenque with Central America, 
on ferns, of a northern and a southern Chocó flora. 
According to Lellinger, the division between both fern 
floras would be near the southern boundary of the 
Chocó Department (Le., near Bajo Calima). If Bajo 
the 


Lellingers northern and southern Chocó floras, 


Calima indeed lies near contact area between 
this 
might perhaps explain its extreme endemism in 
54 of the 


(1996) were new lo science. 


aroids, indicated by the fact that species 
y Bay High 


endemism had already been noted by Gentry (1982) 


found 


for the Bajo Calima and neighboring areas in the Valle 


Department, as well as for Río Palenque. The low 


similarity between the aroid floras of the latter 
localities suggests that endemism in the southern 


Chocó flora is not uniformly distributed, and that al 
least two areas of endemism may exist. Unfortunately. 
the 


intervening Cauca and Nariño Departments of Co- 


the poor floristic knowledge of the lowlands in 


lombia currently prevents any further analysis. 


Volume 93, Number 2 
2006 


Mora et al. 
Araceae of Cabo Corrientes, Colombia 


363 


Literature Cited 


Bay, D. 1996. Araceae of the Bajo Calima region, Colombia. 
Unis Ph.D. Dissertation, Saint Louis University, 
. Louis, Missouri. 

R. G. & G. Galeano. 1993. Las palmas del andén 
‘ifico. 700 ik 0-231 in P. Leyva (editor), 
. Fondo FEN Colombia, Bogotá. 
Croat, T. Flora of = Colorado Island (Panama 
Canal Zone). Stanford Univ. 
1992. Species diversity of Araceae in C ~ A 

. Missouri Bot. Gard. —28. 
'axonomic status of 1 d ud 


Colombia 


X 
~l 


Press, Stanford. 


r 


pre Jean Survey. 
16 


Aroide sana 17: 33—00. 
. M. Mora. 2004. New taxa of Araceae from Cabo 


Co ‘orrientes in Chocó, Department of Colombia. Aroideana 


90-129. 
D. C. Bay € E. D. Yates. 2006. New taxa of 
Anthurium (Ara from the Bajo Calima region 
Valle, Chocó), Colombi and Ecuador. Novon 16: 25-50. 


9 C. H. & H. Gentry. 1978 
Palenque Science i enter, 
1-628 


. Flora of the Río 
Los Híos, Ecuador. Selbyana 4: 
1989. Lista 


: Anotada de las 
Plantas del De spe eater del Chocó, Colombia. Biblioteca 


Forero, E. Gentry. 


„Instituto de Ciencias Naturales, 
Galeano, G. 1992. Patrones de distribución ds las a de 
nue E Inst. Franc. Études Andines 21: 399-607 

Cediel & M. Pardo. 1998a. Structure and floristic 
composition of a one-hectare plot of wet forest at the 
Pacific coast of Chocó, Colombia. Pp. 
Dallmeier € J. A. Comiskey (editors), Forest Biodiversity 
in North, Central and South America, and the Caribbean: 
Research s UNI Man and Biosphere Series 21. 
UNESCO, Pa 


551-568 in F. 


— 


— S. Suárez & II. 8 1998b. Vascular plant 
species count in : t forest in the Chocó area on the 
Pacific Be & Conservation. 7: 


coast of C rcu 
TO: 


1563-15 


Gentry, A. 1982. EU patter rns as evidence for 
1 Chocó refuge. Pp. 112-135 in G. 


Prance (editor), 
Biological Diversification in ie Tropic s. Columbia Univ. 
P ress, New ‘or 
— 19 and floristic 

0 noh :ó region plant communities. 
2 


Species richness composition 
Caldasia 15: 71- 
í 
. 1993. Rique za de especies y composición florística. 
Pp. 200- 219 in 
Vol. Pea t E N du 
ES M. . Revision of T mds subgenus 
Pteromisc tum de ad for Pac “¡fe P Caribbean tropical 
America. Syst. Bot. Monogr. 47: 
/ Galeano & R. Be 
P Eins of the 


E Colombia Pacífico, 


n 1995. Field "an 


Americas. Princeton. Univ. 


175 
. p. p. 1975 


E a ographic analysis of Chocó 
105-114. 


Pte ridophyte S. Fer rn. Gaz. 11: 105-11 
— & E. de la Sota. 1978. »hytogeography of the 
Pteridophytes of the De o del Cl Colombia. 


Res. Rep. Natl. Geogr. Soc., V Wasting z DC i 
ic J. A., K. R. Miller . Reid, Mittermeier 
Werner. 1990. Conse B the e EC s Biological 
9 A The World Bank, 
IUCN, Conservation 1 NS 
Mueller- i. D & H. Aims and 
Methods of Vegetation Ec 1 5 John Wiley and Sons, New 


1 


orld Resources Instit tute, 
Washington, 
Ellenberg. 1974 


Ork. 
Vásquez Martínez, R. 1997. Araceae. Pp. 737-750 in 
1997. Flórula de las 


Perú: 


Reservas Biológicas de 1 1 
¿xplora- 


ra Bot. 


1 Mishana, Explornapo Can 
737-756 Monogr. Syst. B. 


ma | odge Pp. 7 
zard. 
" no i S ver. 1.5. 2006. «http: e zn org/ 
W3T/Seare ye uml, accessed 20 May 
Wilbur, R. L., Collaborators. 1994, 
interim checklist. Pp. 350-378 in L. A. MeDade, 1 
H. A. Hespenheide & G. S. Hartshorn (editors), La 
Selva: Ecology and Natural History of a Neotropical Rain 
Forest. Chicago Press, Chicage 


Bawa, 


Univ. 


364 Annals of the 
Missouri Botanical Garden 


Appendix l. Checklist of the Araceae of Cabo Corrientes, Colombia, and growth habits of the species. E, epiphytic: H. 
hemiepiphytic; T, terrestrial. Genera are in bold: taxa marked with solid circles are new species (Croat & Mora, 2004) and the 


two marked with an asterisk are introduced, 


Taxon Habit Voucher 
Anthurium 
acitangulium Engl. E Mora 10 (COL, MO), Mora & Croat 281 (COL. MO) 
aculibacca Croat & M. M. Mora? „ Mora = (COL. MO), Jácome 335 (COL), Mora & Croat 363 


amargalense Croat & M. M. Mora? 
arustense Croat & M. M. Mora” 


| Mora & Croat 385 (COL. MO) 
| Mora & Croat 352 (COL, MO), Jácome 414 (COL) 
brownti Mast. | Mora 32 (COL, MO), Mora & Croat 55 (COL. MO) 
cogolloanum Croat & M. M. Mora’ „ Mora 64 (COL). Mora & Croat 393 (COL, MO) 
colonense Croat l Mora & Croat 313 (COL, MO) 
l 


cucullispathum Croat Mora 76 (COL, MO), Mora & Croat 316 (COL. MO) 


cuspidatum Matuda T Mora & Croat (COL) 

debilis Croat & D. C. Bay? T | Mora 7 (COL) 

diwyeri Croat „„ Mora & Croat 357 (COL, MO) 

eminens Schott ssp. longispadix Croat & M. M. Mora? E Mora 33 (COL. MO). Mora & Croat 331 (COL. MO) 
Jormosum Schott T | Mora 62 (COL. MO), Mora & Croat 347 (COL, MO) 


friedrichsthalti Schott Mora 24 (COL, MO). Mora & Croat 318 (COL. MO) 
Mora & Croat 329 (COL, MO) 


| 
: | 
grandicataphyllum Croat & M. M. Mora? „ Mora 66 (COL. MO), Mora & Croat 376 (COL. MO) 
E 
| 
| 


galeanoae Croat & M. M. Mora? 


hacumense Eng Jacome 224 (COL) 
hodgei Croat, Ni. . Mora & Oberle? Mora & Croat 391 (COL, MO) 
kunthii Poepp. Jácome 419 (COL) 


ancifolium Schott Mora 4 (COL, MO), Mora & Croat 301 (COL) 


michelii Guillaumin | Mora 45 (COL. MO). Mora & Croat 375 (COL. MO) 
morae Croat | Mora 31 (COL. MO), Mora & Croat 343 (COL) 
obtusilobum Schott E Mora 41 (COL, MO). Mora & Croat 382 (COL, MO) 
obtusum (Engl.) Grayum Mora 30 (COL), Mora & Croat 349 (COL. MO) 
pallidicaudex Croat & M. M. Mora? T Mora 43 (COL, MO). Mora & Croat 344 (COL, MO) 
paludosum Engl. E Mora 23 (COL, MO). Mora & Croat 365 (COL. MO) 
panamense Croal I Mora 14 (COL, MO), Mora & Croat 276 
promininerve Croat & M. M. Mora? T Mora 28 (COL. MO), Mora & Croat 332 (COL. MO) 
propinquum Sodiro T Mora 48 (COL, MO), Mora & Croat 315 (COL. MO) 


ramonense Engl. ex K. Krause Mora & Croat 314 (COL) 
Mora 32 (COL). Mora & Croat 277 (COL. MO) 


| 
ravenii Croat & R. A. Baker | 
rotundistigmatum Croat „ Mora 55 (COL, MO), Mora & Croat 374 
E 
| 


salvinii Hemsl. Mora & Croat 356 (COL. MO) 


trilobum Hort. Linden ex André Mora 18 (COL, MO), Mora & Croat 312 (COL, MO) 


variilobum Croat & M. M. Mora? T Mora 11 (COL, MO). Mora & Croat 350 (COL. MO) 
warocgueanum Moore Mora 47 (COL, MO), Mora & Croat 295 (COL. MO) 
sp. indet. | [sect. Porphyrochitonium Schot] I Mora & Croat 363 (COL, MO) 
sp. indet. 2 [sect. Digittinervium Sodiro] E Mora & Croat 385 (COL. MO) 
Caladü 
2 (Aiton) Vent. T Mora 154 (COL) 
Chlorospatha 
kolbii Engl. TO Mora 50 (COL. MO). Mora & Croat 345 (COL. MO) 
mirabilis (Mast.) Madison TO Mora 50 (COL). Mora & Croat 304 (COL, MO) 
Colocasia 
esculenta (L.) Schott“ T Mora & Croat 392 (COL) 


Dieffenbachia 


davidset Croat & Grayum* To Jácome 350 (COL). Mora & Croat a (COL. MO) 

longispatha Engl. & K. Krause T Mora 72 (COL), Mora & Croat Hoi OL. MO) 

nitidipetiolata Croat & Grayum? T Mora 49 (COL). Mora & Croat 270 : 00. MO) 

tonduzii Croat & Grayum T | Mora 51 (COL), Mora & Croat 310 (COL. MO) 
Dracontium 

spruceanum (Schott) G. H Zhu T. Jácome 399 (COL), Mora & Croat 335 (COL. MO) 


Volume 93, Number 2 
2006 


Mora et al. 365 


Araceae of Cabo Corrientes, Colombia 


Appendix l. Continued. 


Taxon 


Habit Voucher 


Heteropsis 
oblongifolia Kunth 
omalomena 
erythropus (Mart. ex Schott) Engl. ssp. allenii Croat* 
peltata Mast. 
wendlandii Schott 
Monstera 
adansonii Schott 
var. laniata (Schott) Madison 
amargalensis Croat & M. M. Mora* 
dubia (Kunth) Engl. & K. Krause 
minima Madison 
obliqua Miq. 
pittieri Engler 
pinnatipartita Schott 
spruceana (Schott) Engl. 
Philodendron 
alliodorum Croat & Grayum 
amargalense Croat & M. M. Mora* 
angustilobum Croat & Grayum 
bakeri Croat & Grayum 
croatii Grayum 
ensifolium Croat & Grayum 
fragrantissimum (Hook.) Kunth 
grandipes K. Krause 
grayumit Cron 
hebetatum Croal 
hederaceum (Jaeq.) Schott 
heleniae Croat 
ichthyoderma Croat & Grayum 
immixtum Croat 
inaequilaterum Liebm. 
jodavisianum G. S. Bunting 
laticiferum Croat & M. M. Mora* 
ligulatum Schott 
Mr APA ulatum Croat & M. M. Mora? 
opacum Croat & Grayum 
panamense ^ Krause 
platypetiolatum Madison 
rayanum Croat & Grayum 
rhodoaxis G. S. Bunting ssp. lewisii 125 roat & Grayum 
roseocataph yllum Croat & M. M. Mor 
sagittifolium Liebm. 
scalarinerve Croat & Grayum 
senatocarpium Madison 
5. Bunting 
subhastatum Engl. 
sulcatum K. Kraus 


sulcicaule Croal ^s — 


strictum G. 


tenue K. Koch & Augustin 
tripartitum (Jac a Schott 
sp. indet. | 


sp. indet. 2 


H Jácome 390 (COL). 401 (MO) 


T Mora 38 (COL), Mora & Croat 271 (COL) 


T Mora 77 (COL, MO), Mora & Croat 301 (COL, MO) 

T Mora 75 (COL. MO). Mora & Croat 336 (COL, MO) 

I Mora & Croat 324 (COL, MO), Jácome 423 (COL, MO) 
H Mora 65 (COL), Mora & Croat 324 (COL, MO) 

H Mora & Croat 324 (COL, MO), Jácome 423 (COL, MO) 
Mora 54 (COL, MO). Mora & Croat 340 (COL, MO) 

H Jácome 280 (COL) 

H Mora 53 (COL, MO) 

H Jácome 280 (COL) 

H Mora 42 (COL, MO), Mora & Croat 306 (COL, MO) 


H Mora 26 (COL), Mora & Croat 265 (COL, MO) 


H Jácome 226 (COL). Mora & Croat 285 (COL. 


H Mora & Croat 280 (COL. MO), Mora & Croat B o OL. MO) 
H Mora 60 (COL, MO) 

H Jácome 228 (COL) 

H Jácome 369 (COL) 


II. Jácome 337 (COL) 

H Mora 37 (COL) 

T Mora 27 (COL, MO): Mora & Croat 272 (COL) 

H Mora 16 (COL, MO), Mora & Croat 264 (COL, 
II. T Mora 67 (COL, MO), Mora & Croat 325 (COL, 

H Mora 76 (COL), Mora & Croat 319 (COL, MO) 


MO) 
MO) 


H Mora 36 (COL, MO), Mora & Croat 321 (COL, MO) 

T Mora 74 (COL). Mora & Croat (COL) 

H Mora & Croat 354A (COL, MO) 

T Jácome 263 (COL) 

H Mora 3 (COL, MO), Mora & Croat 339 (COL, MO) 

H Mora 9 (COL, MO) 

1 Mora 17 (COL. MO), Mora & Croat 266 (COL. MO) 

H Mora 61 (COL) 

H Mora 71 (COL, MO). ps is Croat 359 (COL. MO) 

H Mora & Croat 369 (COL, MO) 

H Mora 29 (COL, MO), 175 pS Croat 322 (COL, MO) 

H Jácome 322 (COL) 

H Jácome 243 (COL). Mora & Croat 292 (COL, MO) 

H Mora 95 (COL, MO). Mora & Croat 384 (COL. MO) 
COL). Mora & Croat 280 (COL, MO) 


a MO), Mora & Croat 277 (COL, MO) 
H Mora & Croat 326 (COL, MO) 

T Mora 59 (COL) 

H Mora & Croat 355 (COL. 
H Jácome 355 (COL) 

H Jácome 219 (COL) 

H Mora 57 (COL). Mora & Croat 348 (COL, MO) 
H Mora 127 (COL), Mora & Croat 316 (COL, MO) 
H Jácome 262 (COL) 

H Mora 68 (COL). Mora & Croat 327 (COL) 


MO) 


366 


Annals of the 


Missouri Botanical Garden 


Appendix 1. 


Continued. 


Taxon Habit Voucher 
Rhodospatha 
brachypoda G. 5. Bunting H = Mora & Croat 361 (COL) 
monsalveae Croat & D. C. Bay* H Mora 40 (COL) 
moritziana Schott H. T Mora | (COL, MO). Mora & Croat 282, 287 (COL. MO) 
pellucida Croat & Grayum H Jácome 426 (COL) 
wendlandii Schott H Jácome 248 (COL) 
sp. indet. ! H Mora & Croat 299 (COL) 
Spathiphyllum 
dressleri Croat & F. Cardona? T Mora & Croat 366 (COL) 
friedrichsthalid Schott T Mora 25 (COL), 91 (COL. MO). Mora & Croat 387 (COL. 
MO 
phryniifolium Schott T Mora 15 (COL), Mora & Croat 328 (COL, MO) 
laeve Engl. T Mora 5 (COL, MO), Mora € Croat 311 (COL, MO) 
Stenospermatio 
angustifolium 15 msl. E. Mora 20 (COL, MO), Mora & Croat (COL, MO) 
latifolium Engl. E Mora & Croat 263 (COL) 
multiovulatum (Engl.) N. E. Bi E Mora 13 (COL). Mora & Croat 279 (COL. MO) 
robustum Engl E Mora 34 (COL), Mora & Croat 275 (COL, MO) 
Syngoniu 
5 We y ex G. S. Bunting H Mora 69 (COL), Mora & Croat 290 (COL, MO) 
macrophyllum H = Mora 21 (COL), Mora & Croat 368 (COL) 
chocoanum E H Mora 7 (COL, MO), a 380 (COL) 
podophyllum Schott Mora & Croat 341 (C Q) 
triphyllum Birdsey ex Croat H Mora & Croat 334 91905 MO) 
Xanthosoma 
daguense Engl. T Mora 19 (COL, MO), Mora & Croat 268 (COL, MO) 
var. amargalense Croat & M. M. Mora* T Mora & Croat 300 (COL. MO), Jácome 400 (COL) 
sagittifolium (L.) Schott & Endl. T No voucher (pers. comm., Croat) 


SCIENCE WITH A SOCIAL 
CONSCIENCE: DIGGING FOR 
DINOSAURS AND HELPING 
CHILDREN IN THE LAND THAT 
TIME FORGOT 


David W. Krause! 


Madagascar has been good to us. In 1993, I led 

a reconnaissance expedition in search of Cretaceous 
dinosaurs, mammals, and other vertebrate animals to 
this, the fourth-largest island in the world. Not in my 
wildest dreams did I anticipate the fossil riches that we 
would find on that expedition, or on the seven field 
'ampaigns since. Many of our discoveries are docu- 
mented in an article in this volume (see pp. 178-208). 
What we also found, however, were some of the 
most abandoned. and destitute people on the planet. 
Many of Madagascar citizens, particularly in rural 
areas, do not know how to read or write and have 
never been seen by a doctor or dentist. Malnutrition is 
ubiquitous. Children, who are otherwise rail thin, have 
abdomens that are visibly distended, largely a result 
of protein deficiency but oftentimes also owing to 
endemic parasites that inhabit, multiply within, and 
thus enlarge spleens and livers. Malaria is rampant. 
Upper respiratory, diarrheal, and dental infections are 
common and, owing to the lack of access to basic 

hygiene and antibioties, all too frequently fatal. 

These same Malagasy villagers welcomed us into 
their communities, directed us to productive field 
areas, and helped us carry massive plaster jackets 
containing fossil specimens over kilometers of rough 
terrain to waiting vehicles. With a desire to repay 
them for their many kindnesses, we asked what we 
could do. They humbly requested a school for their 
children. It was a “no-brainer” when we learned we 
could get things started by hiring a teacher for only 
US$500/year. 

In 1998 
expanded our efforts by establishing the Madagascar 
Ankizy Fund (MAF). The primary mission of MAF is 
to provide education and health care to children 


and I formalized and 


— 


my colleagues 


(*ankizy" in the Malagasy language) living in remote 
areas of the island nation, which currently has 
a population of 18 million. To date, MAF can claim 


Department of Anatomical Sciences, Stony 
a @notes.cc.sunysb.edu 


dKraust 


Brook University, 


the following successes: (1) the building and in- 
auguration of two elementary schools, one of them in 
our field area in northwestern Madagascar (Fig. 1), the 
other in a rainforest in the east (two more are in the 
planning stages); (2) establishment of annual health 
care clinics in remote areas, staffed largely by Stony 
Brook University personnel; (3) drilling of several 
wells and initiation of a water disinfection program; 
(4) initiation of an arts and crafts industry in a village 
starving because the primary occupation, wood- 
cutting (charcoal production), is no longer feasible 
since the forest has been depleted; (5) renovation of an 
orphanage in the city of Mahajanga: (6) initiation of 
community-wide education programs in nutrition, 
hygiene, and sexual health; (7) establishment of 
sister-institution status between the Stony Brook 
University School of Dental Medicine and the only 
dental school in Madagascar; and (8) initation of 
a program through which used and surplus hospital 
equipment and supplies are shipped from hospitals in 
the U.S.A. to those in Madagascar. 

Funding for MAF projects is derived from several 
sources: grants from organizations such as Rotary 
International, various administrative units at Stony 
Brook University, private individuals, and American 


children who conduct fundraisers in their schools. The 


— 


arly gratifying because it has a re- 


alter is particu 
ciprocal effect: children in Madagascar benefit from 
the relative wealth of children in the U.S.A. while the 
atter learn a great deal about the country and people 
of Madagascar. Importantly, American schoolchildren 


have the additional benefit of discovering that they 


can make a meaningful difference in the lives of much 
less fortunate children. 

Why am I writing this article? Partly because the 
editors of this volume kindly invited me to do so but 
primarily for two reasons: (I) to seek assistance for 
MAF projects by directing readers to visit the MAF 
Brook, New York 11794-8081 U.S.A. 


Stony 


Ann. Missourt Bor. Garp. 93: 367-368. PUBLISHED ON 23 Aucusr 2006. 


Annals of the 
Missouri Botanical Garden 


AGG ͤ A 
MOSES NON A : AIN 


SEKOLY RÍAS 


THE 5 
INAUGUI 


Dr. David W. Krause (left 


Figure !. 


ONY 2 
REE LE 


MOOK SCHOOL 


KOOK 
15 JUILLET 2C 


with teachers and children at inauguration of the Berivotra Elementary School, built 


by the Madagascar Ankizy Fund. Photo credit: Ashutosh Kaushesh. 


website at www.ankizy.org. and (2) to encourage and 
hopefully inspire other researchers working in de- 
veloping countries to initiate similar efforts; As 
scientists, both professionals and students, we have 
collected and studied literally thousands of specimens 


from Madagascar, and our careers have been advanced 


immeasurably as a result. This is simply our way of 
giving back to a country that has given us so much. 
NAF is administered by the Stony Brook Founda- 
tion, a 5010003 not-for-profit. education corporation 
established by Stony Brook University to receive and 


manage philanthropic contributions. 


Volume 93, Number 2 Solomon & Magill 369 
2006 Statistical Summary 
STATISTICAL SUMMARY OF SOME OF THE ACTIVITIES IN THE MISSOURI BOTANICAL GARDEN HERBARIUM, 2005 
Vascular plants Bryophyte Totals 
Acquisitions of Specimens 
Staff Collections 23.186 1.820 25.000 
Purchase 12.550 0 12,550 
Exchange 19,142 3,285 22,427 
Gifts 15.910 069 16.579 
Total acquisitions 71.388 5,774 77,162 
Mountings 
Newly mounted at MO 96,089 16,033 112,122 
Specimens mounted when acquired 0 0 0 
Total mountings 96.089 16,033 112,122 
Repairs 
Specimens repaired 20.975 0 20.975 
Specimens stamped 1.359 0 1.359 
Total repairs 22,334 0 22.334 
Specimens sent 
On exchange 24,293 6,455 30.748 
As gifts 18.111 1.238 19.349 
Total 12,404 7.093 50.097 
Loans sent 
Total transactions 295 34 329 
Total specimens 25.891 1,424 27.318 
To US institutions 
Transactions 153 16 169 
Specimens 13.620 307 13,987 
To foreign institutions 
Transactions 142 8 160 
Specimens 12,274 1,057 13.331 
To student investigators 
Transactions 53 5 58 
Specimens 5.066 140 6,306 
To professional investigators 
Transactions 242 20 211 
Specimens 20.028 984 21.012 
Loans received 
Transactions 160 68 228 
Specimens 13,727 7,241 20.974 
From U.S.A. From abroad Total 
Visitors 237 109 346 


ANN. Missouri Bor. GARD. 93: 369-370. PUBLISHED ON 23 Aucusr 2006. 


370 Annals of the 
Missouri Botanical Garden 


The Garden’s herbarium is closely associated with its database management system, TROPICOS (see 
www.tropicos.org). The charts below summarize some of the statistics from TROPICOS both for the calendar year 
2005 and year-end totals. At the end of 2005 the Garden’s herbarium held 5.749.641 numbered sheets: 
9,283,152 vascular plants and 466,489 bryophytes. 


TROPICOS RECORD SUMMARY 


Calendar year 2005 Vascular Plants Bryophytes Totals 
Specimens 525,672 13,742 539,414 
Names 18,163 304 19,407 
Synonyms 0,137 2,965 9.102 
EN 12,965 251 13,216 

ypes 6.299 101 6.100 
bos aphy 3,511 2,800 6.311 

Year end 2005 Totals Vascular Plants Bryophytes Totals 
Specimens 2,720,537 206,922 2,927,459 
Names 879,290 108,407 987,697 
Synonyms 132,007 74,238 506,245 
Distributions 877,120 13,710 920.890 
Types 322,977 9,823 332,800 
Bibliography 10,074 31.005 107,079 
In TROPICOS, taxonomic synonymy is always linked to a reference in Bibliography and directly with at least 


two records in Names. i.e. the synonym and the correct. name according to the reference. Additional 
nomenclatural synonymy may be derived from these direct links. e . all other combinations of a basionym 


treated as a synonym of a give n name are also synonyms of that name. 


—James C. Solomon and Robert E. Magill 


Volume 93, Number 2, pp. 173-370 of the ANNALS OF THE Missouri BOTANICAL GARDEN 
was published on August 23. 2006. 


ica 


3 1753 00336 8260 


00 26 


www.mbgpress.org 


CONTENTS 


Latin American Biogeography— Causes and Effects, the 5lst Annual Systematics 
Symposium of the Missouri Botanical Garden 

Latin American Biogeography—Causes and Effects. Introduction Alan Graham 

Late Cretaceous Terrestrial Vertebrates from Madagascar: Implications for Latin 

American Biogeography — David W. Krause, Patrick M. O'Connor. 

Kristina Gia Rowen. Scott D. Sampson, Gregory A. Buckley & 

Raymond R. Rogers 


E "M an PIES "The Biogeographic MEA E Sen American Land 
Rosendo Pascual 


Mammals 
Paleogeography of the Antilles and Origin F West Indian Terrestrial Vertebrates 
AA ETE S. Blair Hedges 
The Great American Biotic Interchange: Patterns and Processes S. David Webb 


Late Quaternary Environments of the Northern Deserts and Central Transvoleanic Belt 
of Mexico Sarah E. Metcalfe 
(Quaternary Puvisoumentat re E pon TUNE Impact on Vegetation in Central 


Dolores R. Piperno 


The olaaa Record of Colombia: n for Biogeography and 
Biodiversity... Henry Hooghiemstra, Vincent M. Wijninga & Antoine M. Cleef 
Paleobotanical Evidence and Molecular Data in Reconstructing the Historical Phyto- 
geography of Rhizophoraceae Alan Graham 


Modern Processes and Historical Factors in the Origin of the African Element in Latin 
America Alan Graham 

Dispersal- rasan insu in e Tribe obleas 1 A Clue to Un- 
derstanding Biogeographical History of the Brazilian Atlantic Forest 


Mathieu Perret, Alain Chautems & Rodolphe Spichiger 


A Phytogeographic Añalyals of Araceae of Cabo Corrientes (Chocó, Colombia) 
and Comparable Lowland Tropical American Floras Marcela Mora, Rodrigo Bernal, 
fae _ Thomas Croat & Jorge Jácome 

Science with a Social nen ience: De ie Minos ad Helping Children in the Land 
that Time Forgot —. David W. Krause 


Pipe) 'al Summary of Some of ha TE in the Missouri Botanical Garden Herbarium, 


. James C. Solomon & Robert E. Magill 


Cover illustration. Capparis sicula subsp. mesopotamica Inocencio, D. Rivera, Obón & Alcaraz, 


drawn by Jose-Antonio Barrefia. 


Annals 
Ol the 

Missouri 
Dotanical 


Garden 
na VG 


olume 93 


umber 


Annals of the 
Missouri Botanical Garden 


Volume 93, Number 3 
October 2006 


The Annals, published quarterly, contains papers, primarily in systematic botany, 
contributed from the Missouri Botanical Garden, St. Louis. Papers originating outside 
the Garden will also be accepted. All manuscripts are peer-reviewed by qualified, 
independent reviewers. Authors should write the Managing Editor for information 
concerning arrangements for publishing in the Annals. Instructions to Authors are 
printed in the back of the last issue of each volume and are also available online at 


www.mbgpress.org. 


Editorial Committee 
Victoria C. Hollowell 
Scientific Editor, 

Missouri Botanical Garden 
Beth Parada 

Managing Editor, 
Missouri Botanical Garden 
Diana Gunter 

Associate Editor, 

Missouri Botanical Garden 
Barbara Mack 

Editorial Assistant, 
Missouri Botanical Garden 
Monica Anderson 

MBG Press Assistant, 


Missouri Botanical Garden 


Ihsan A. Al-Shehbaz 
Missouri Botanical Garden 
Gerrit Davidse 

Missouri Botanical Garden 
hoy E. Gereau 

Missouri Botanical Garden 
Peter Goldblatt 

Missouri Botanical Garden 
Gordon McPherson 
Missouri Botanical Garden 
P. Mick Richardson 
Missouri Botanical Garden 
Charlotte Taylor 

Missouri Botanical Garden 
Henk van der Werff 
Missouri Botanical Garden 


For subscription information contact ANNALS 
or THE Missourt BOTANICAL GARDEN, % Allen Mar- 
keting & Management, P.O. Box 1897, Lawrence, 
KS 66044-8897. Subscription price for 2006 is 
$160 per volume U.S., $170 Canada & Mexico, 
$195 all other countries. Four issues per volume. 
The journal Novon is included in the subscription 
price of the Annals. 


annals@mobot.org (editorial queries) 
http://www. mbgpress.org 


The ANNALS or THE Missouri BOTANICAL GARDEN 
(ISSN 0026-6493) is published quarterly by the 
Missouri Botanical Garden, 2345 Tower Grove 
Avenue, St. Louis, MO 63110. Periodicals post- 
age paid at St. Louis, MO and additional mail- 
ing offices. POSTMASTER: Send address changes to 
ANNALS OF THE Missouri BOTANICAL GARDEN, % 
Allen Marketing € Management, P.O. Box 1897, 
Lawrence, KS 66044-8897. 


The Annals are abstracted and/or indexed in AGRICOLA (through 1994), APT Online, BIOSISO, CAB Ab- 
stract/Global Health databases, ingenta, ISI databases, JSTOR, Research Alert, and Sci Search®. 


Typesetting by Allen Press, Inc. using Advent3B2 automatic paging system. 


© Missouri Botanical Garden Press 2006 


The mission of the Missouri Botanical Garden is to discover and share knowledge about plants and 
their environment, in order to preserve and enrich life. 
This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 


| 


Volume 93 
umber 3 


* 

3 18e BC, 

RAISSOURI AG Y] 
Wii) UIN be 


Annals 

of the 
ISSOUTI 
otanical 
arden 


A TAXONOMIC REVISION OF 
CARIBBEAN ADIANTOPSIS 
(PTERIDACEAE)' 


Michael S. Barker** and R. James Hickey? 


ABSTRACT 


Idiantopsis Fée (Pteridaceae) is a relatively unstudied tropical cheilanthoid fern genus. In the present work, we evaluat 


the taxonomy and relations 


x 
ups among Caribbean Adiantopsis by examining 136 characters from approximately 500 herbarium 
specimens. This study identified nine Caribbean Adiantopsis species, three of which are newly described (A 1 


parvisegmenta 


pentagona, and A. vincentit). Additionally, an intriguing pattern of morphological and reticulate evolution was revealed by the 
analyses. Adiantopsis consists of three different laminar morphologies: palmate, pedate, and pinnate. The two pedate taxa are 


hy pothesize «d lo be fi 


ertile allotetraploid derivatives of the palmate A. 
iod they parallel the origin of the South American A. Xaustralopedata Hickey, 
analyses, il appears that pedate laminar morphologies in 


radiata (L.) Fée We two different 5 taxa. In this 
M. S. Barker € Ponce 


Idiantopsts dependan 


Based on our 


originated tele limes via 


hybridization. This study provides testable hypotheses of morphological and reticulate adn in the genus and presents 


a novel view of Caribbean Adiantopsis. 
Key words: 


Adiantopsis, Caribbean flora, cheilanthoid. 


ferns. monilophytes, 


Pteridophytes. 


Adiantopsis Fée is a cheilanthoid fern genus found 
throughout tropical America (Tryon & Tryon, 1982 
Africa (Moran & Smith, 2001). A 


combination of echinate spores, distinct pseudoindu- 


— 


and possibly 


sia, adaxial carinae, and unique septate-capitate hairs 
on the axes and laminar surfaces distinguishes 
Adiantopsis from other. cheilanthoids. Historically, 


the genus was interpreted to include seven species, 


four of which occur in the Caribbean, predominantly 
in Cuba and Jamaica (Tryon & Tryon, 1982). In this 
regional revision, we recognize nine Caribbean 
species out of an estimated 20 Adiantopsis species 
worldwide. In the Caribbean, Adiantopsis is a genus 
primarily of calcareous forests and cliffs. Six of the 
nine Caribbean Adiantopsis species are limited to 


narrow ecological niches, such as serpentine soils, 


The authors thank the cited curators and institutions for providing es imen loans. M. Vincent and MU deserve — for 


handling loans and making space for our material. J. Budke, 


thank W. s for assistance with specin 
Hauk, V. 1 1 Jed Mickel, and an anonymous rey 


K. Heafner, and S 
assisted us collections and unique nomenclatural information at US. M. Supdtte was very hel 
n data entry and J. G 
iewer signific antly i improved the manuscript. 


Shaw provided support in the lab. D. Le ‘Hinge 2 


— 
a 

w 
= 
un 


;astony for reviewing our Latin MC Comments from W. 


E 
Ohio Academic Challenge DRM the Botany Department at Miami University, and the Willard She man Turrell Herbarium 


F und for financial suppor 


e editors of the, 2228 thank Sophia Balcomb for her editorial contribution to this article. 


¿Departa of Botany, Miami University, Oxford, Ohio 45056 U.S 


Present 


ANN. Missouri Bor. GARD. 


Address: Department of Biology. Indiana University, Jordan Hall 142, 1001 East Third Street, Bloomington 
icis 47 405- 7005 U.S.A. (e-mail address: msbarker@indiana.edu). 


93: 371—401. PUBLISHED ON 24 OcToBER 2006. 


372 


Annals of the 
Missouri Botanical Garden 


and are endemic to small geographic areas of single 
islands. 


Relative to other cheilanthoid genera, a paucity « 


information exists about Adiantopsis. It is one of many 


small, overlooked. tropical genera that have never 
been examined using formal taxonomic methods. 


500) recognized this problem nearly 
"New total 
eraphs...to reveal the kinds of groups which can be 
tackled by the 


for studying Adiantopsis were simply to describe and 


Holttum (1975: 


30 vears ago, and called for mono- 


methods of biosystematics.” Reasons 


enumerate the taxa present, and 
knowledge of this genus. Such knowledge is a pre- 
molecular, and 
Additionally, 


axonomic research on Adiantopsis provides data for 


requisite for future systematic, 


evolutionary research on Adiantopsis. 


the eventual reorganization of the polyphyletic 
Cheilanthes Sw. 


monophyletic) classification scheme, 


(Gastony, unpublished data) into 
a natural (i. e. 
something that is not possible without basic research 
on Adiantopsis and other understudied 
Cheilanthes. 


Adiantopsis also possesses a striking range of laminar 


morphology. The laminae may be palmate (radiate), 
pedate, or pinnate. 
in many other fern genera, especially ones as small as 
Adiantopsis. However, without a solid taxonomy of the 
species, understanding the evolution of this variation is 
not possible. Thus, another reason for studying Adian- 
topsis was to develop and evaluate hypotheses about the 
origin of these laminar morphologies. 

The results presented in this revision are a compo- 
nent of a complete monograph of Adiantopsis that is 
currently in progress. In this publication, a taxonomic 
revision of the Caribbean Adiantopsis taxa is pre- 


sented: the Central and South American taxa will be 
included in a future, more comprehensive monograph. 
Although this taxonomic treatment deals only with the 
Caribbean species, the discussion of Adiantopsis 


biology in this revision should apply to the entire 


genus. 
MATERIALS AND METHODS 


We examined approximately 500 specimens from 
the following herbaria for this revision: B. BM, DUKE, 
F, Fl, GH, K, L, MICH, MO, MU, NY, P, UC, and US. 
In all, 136 characters were scored for 


For the 


each species. 


most part, morphological characters were 


recorded without disturbing the specimen, and were 
collected by eye or with a dissecting microscope. An 
Olympus BH-2 compound light microscope was used 
methodolo- 


to examine microscopic characters. The 


ejes for describing characters requiring special 


preparation are described below. 


increase our 


elements of 


This range of variation does not occur 


For length and width measurements, scales of the 
and rachis were mounted in Hoyer's 
1954) on a 


Measurements were made on a compound light 


rhizome, stipe, 


solution (Anderson, microscope slide. 
microscope. Scales were collected only from speci- 
mens with numerous scales, so that specimens would 
sull possess scales for future research. 

Guard cell measurements were made with a com- 
pound light microscope using prepared laminar tissue. 
Ultimate divisions were collected from various parts of 
the frond to avoid any difference in guard cell length 
altributed to position. The collected laminar tissue 
30% weight/volume (w/v) NaOH 


solution for 20 minutes to remove epicuticular wax. 


was placed into a 


To decolorize the tissue, it taken through an 
ethanol series of 2096, 50%, 85%. and 100%, 


then backward 1 that series lo water. 


was 
and 
The tissue 
was kept at each stage of the series for 10 minutes. 
The decolorized tissue was stained using a modifica- 
tion of Fosters Tannie Acid — Iron Chloride method 
(Johansen, 1940): decolorized tissue was placed in 


Jo w/v aqueous solution of tannic acid for 


IO minutes, briefly rinsed water, and transferred 
to a 3% w/v aqueous solution of ferric chloride for 
approximately 30 minutes. The tannic acid and ferric 
chloride steps were repeated until the cell walls were 
i Once the 


the tissue was mounted 


dark enough to be readily seen. visibility of 


the cell walls was satisfactory, 


ides using Hoyer’s solution, and guard cell length 
light 


ons 
measurements were made on a compound 
microscope. 

Sporangia were collected from herbarium speci- 
mens using forceps and mounted on a slide with 
Hover's solution. Two measurements were made from 
collected sporangia: the number of indurated arcus 
Data 


a compound light microscope. 


cells. and their height. were collected on 


Spores were collected from herbarium specimens 
and examined using both Scanning Electron Micros- 
Spores for SEM 


were collected from open sporangia to ensure that the 


copy (SEM) and light microscopy. 


spores were mature and that there were no contam- 
inating spores from other specimens. For counting 
spore number per sporangium, complete intact 


sporangia were collected from each specimen. The 
closed sporangia were collected using fine forceps 
and transferred to an aluminum SEM stub that was 
prepared with a double-sided sticky tab and a drop of 
water. The water allowed the spores to be easily 
scattered over the surface of the stub and kept the 
spores on the stub when breaking intact 
SEM 
coated for 90 seconds at 45 milliamps. Spores were 
JEOL JSM-T200 SEM at 


working distances, voltages, and spot sizes. depending 


open 


sporangia. stubs were gold-palladium sputter 


examined on a various 


Volume 93, Number 3 
2006 


Barker & Hickey 
Caribbean Adiantopsis (Pteridaceae) 


373 


— 


on the stub. Pictures were taken using an attache 


a 


Land camera with Polaroid Polapan 55 black anc 
white film. For light microscopy, spores were collected 
with fine forceps from open sporangia. These spores 
were mounted on a slide with Hoyer's solution and 
measured on a compound light microscope. 

Data analyses were conducted using SAS (Version 
8: SAS, 2000) and OpenOffice.org Cale (Version 1.1; 


OpenOffice.org, 2003). Guard cell, arcus cell, and 


spore size dala were compared in SAS using one-way 
ANOVAs 


comparisons ( = 0.05) to identify significantly 


with Tukey-Kramer multiple pairwise 


different values. Ranges, means, modes, and standard 


deviations for all quantitative characters were calcu- 


ated using OpenOffice.org Cale. Box plots and 
histograms for select data sets were constructed using 
OpenOffice.org Cale. 


Character his revision generally 


ct 


terminology in 


— 


ollows Lellinger (2002). One deviation is the scoring 
of planar shapes; these were characterized using the 
“Symmetric Plane Figures” guide published by the 
1962). 


complex 


Systematics Association Committee Addition- 


peres 


ally, as Adiantopsis possesses laminar 


architectures, terms relating to the pinnae and 
pinnules have been slightly modified from Lellinger 


(2002). In 


compound, the term ultimate division, which is not 


taxa that are more than bipinnately 


in Lellinger (2002), is used to describe any portion of 
the lamina that is not further dissected or compound, 
regardless of its position. Thus, an ultimate division 
may be a pinna, a pinnule, or a pinnulet, so long as it 


is nol further divided. When using the term ultimate 


division to describe these various structures, no 


homology is implied. Thus, pinnae are still the 


primary division of the lamina, but when pinna 


morphology is described it applies only to pinnae 


— 


that are compound themselves (i.e., bipinnate and 
more compound species). This same logic applies to 
pinnules. Thus, all final divisions of the lamina are 
grouped under ultimate division; divisions that are 
themselves compound are described using standard 


positional terms (pinna, pinnule, etc.). 


Ecoroca 


Adiantopsis species are ecologically variable. 
Caribbean Adiantopsis frequently occur in moist, 


forested areas, or on moist rock walls, often in gullies 
or cave entrances. In South America, many species 
also prefer moist, forested areas, often occurring near 
streams or on stream banks, although some South 
American Adiantopsis, such as some elements of A. 
chlorophylla (Sw.) T. Moore, do prefer exposed, rocky 
alka 


a requisite. for 


— 


habitats. An ine substrate appears to be 


many Adiantopsis taxa. In the 


Adiantopsis taxa in Ponce & Morbelli, 1989). 


Caribbean, this includes limestone, serpentine, and 


— 


amphibolite substrates. Adiantopsis radiata (L.) Fée 
may be able to grow in a variety of pH substrates. 
However, at least one South American species, A. 
monticola (Gardner) T. Moore, prefers acidic sub- 
strates. Adiantopsis taxa occur from near sea level to 


as high as 2800 m (Tryon & Stolze, 1989). 
MonPHOLOGY 


No previous generic-level description of Adiantop- 
sis morphology exists. The following description is the 
first of its sort for the genus. However, it cannot be 
considered comprehensive as it only represents data 
and patterns observable from herbarium specimens 
and excludes possible Old World members. 

Adiantopsis is a genus of generally erect, small to 
large herbaceous ferns. Among the Caribbean species, 
the plants (as a function of leaf length) range from 
small (9.5 em in A. asplenioides Maxon) to medium 
sized (to 74.0 cm in A. parvisegmenta M. S. Barker € 
Hickey): most individuals are around 20—40 cm long. 
occur in South 
Xaustralopedata Hickey, M. 


that may exceed one and a half 


arger taxa America, such as 


"ES 


chlorophylla and A. 
Barker & Ponce, 
meters in length. 
stiffly 


Kunze) Fée, the plants are also erect, but the stipes 


Most Adiantopsis taxa are strict, 


erect. In A. asplenioides and A. 1 


are not as stiff as other taxa. Caribbean Adiantopsis 


axa are generally smaller than their South American 
congeners, and less habit variation is observed. 
The rhizomes of Adiantopsis taxa are all similar. 


Rhizomes range from erect, de- 


© 


to ascending, te 
cumbent. It is not clear if rhizomes are distinctly 
subterranean or marginally surficial, but given the 
amount of soil on most rhizomes they appear to be at 
least partially hypogeal. In this respect, no difference 
is observed among taxa. Rhizomes range to 5.9 cm 


2.3 cm 


scales densely cover the rhizomes (Figure 1A); these 


long and diameter. Persistent, bicolorous 
may play some role in preventing dessication and/or 
infection by fungi or bacteria. In all species, a dense 
tangle of fibrous roots emerges from all sides of the 
rhizome. Stipes also emerge from all sides of the 
rhizomes, suggesting a fundamentally radial symmetry 
(Figure 1A). No positional relationship was observed 


between the stipes and the roots. 


Frond axes in Adiantopsis are typically chei- 
lanthoid. They are generally persistent and atropur- 


pureous to atrocastaneous, becoming darker at 


senescence, Adiantopsis axes are also carinate 


adaxially (Figure 1B). These carinae are in the form 
of paired, golden ridges raised above the surface of the 
“sclerotic ribs” for putative 


axis (described as 


Tryon 


1gur 


= 


D= 


374 


Annals 


Missouri Botanical Garden 


— 


li l. 
surface, a E 1 5 altached stipes (C. Wright 904, 
. reesti stipe (E ‘kman 9987, UC). Sci 
retaining P ‘ir e j^ — 
Seale bars C, 2 


Se un bar — 0,25 cm. Adaxial carinae (indicate 
xr = 1.5 mm. —C. Persistent costae " Í paupercula (H 
D. Marcescent costae 4 . vincenti (C. 


V. 


. Longitudinal section of Idiantopsis ab ones rhizome with olorous scales densely covering rhizome 
US). : 
cale bar 
m. 


d by 10 on 
espenheide 1191, US 
Morton 10389. US) airline upon themse 5 8. 


Volume 93, Number 3 


Barker & Hickey 
Caribbean Adiantopsis (Pteridaceae) 


carinae distinguish 


(1982) used the 


Adiantopsis from Cheilanthes. 


and Tryon 
However, characteriza- 


tion of Cheilanthes as lacking carinae is not accurate. 


Some Cheilanthes, such as C. aemula Maxon. C. 
microphylla (Sw.) Sw.. and C. wrightii Hook. are 


adaxially bicarinate (Barker, pers. observ.). In the 


bicarinate Cheilanthes, the carinae are usually green 
w pale green, approaching the color of the laminar 
tissue, whereas Adiantopsis carinae. are distinctly 


golden, independent of laminar color. Furthermore 


carinae in Cheilanthes are typically restricted to the 


costae and costules, whereas in Adiantopsis they 


Additionally, Cheilanthes 


laminar 


= 


generally occur on all axes. 


carinae are. contiguous. with the tissue. In 


Adiantopsis, with the exception of lamina and pinnae 


apices, carinae. remain distinct. from the laminar 


tissue. However, this may be more an architectural 


phenomenon than a reflection of the absence o 


homology. Thus, carinae distinguish Adiantopsis from 


Cheilanthes by their form, not their presence. 
associated with the 


Of all axis characters, those 


stipes are the most taxonomically informative. In 
Adiantopsis, the stipes may grow to 47.0 em long and 
2.0 mm in diameter. Pneumatothodes were observed 
on stipes of all species. Taxonomic differences emerge 


when comparing the ratio of stipe length to overall 


rond length. In the pinnate Caribbean taxa, stipes are 
one third the overall frond length, whereas they are 
equal to or longer than the lamina in the pedate taxa. 
Stipes of A. radiata are consistently longer than the 


lamina, and in A. paupercula they approach equity 


with laminar length, being variously shorter or longer 


a 


than the lamina. There are also taxonomic differences 


in the extent of carinae. development. In A. pauper- 


cula, the carinae begin in the upper half of the stipe. 


frequently starting at the stipe mid-point. Carinae 
3 B | 


begin in the dl third and quarter of stipes in A. 


pentagona M. S. Barker € Hickey and A. rupicola 


axon. In A. 1 A. pedata (Hook.) T. 
Moore, A. reesti (Jenm.) C. Chr., and A. vincentii M. S. 


Barker & Hickey, carinae begin at the stipe bases or 
basal half of 
lacks 


present, they begin at the base of the stipe. Carinae 


in the stipes. Adiantopsis radiata 


frequently carinae on the stipes, but when 

are always absent on the stipes of A. asplenioides. 
Laminar axes in Adiantopsis are generally persis- 

lent. In most taxa, the original laminar axis position 


and form are retained even after the loss of ultimate 


divisions (Figure 1C). However, in senescent fronds of 


A. rupicola and A. vincentii the costae are marcescenl 
and curl upon themselves becoming almost circinale 
in appearance (Figure 1D). 

In his revision of Lindsaea Dryand. 


(1957: 123) noted that “In few fern genera such a great 


diversity of leaf-pattern is found as in Lindsaea.” If 


ex Sm., Kramer 


the above statement is true for Lindsaea. then 


Adiantopsis may have the largest diversity of laminar 
variation seen in any fern genus, and it is certainly the 
most diverse among the cheilanthoid genera. Three 


basic laminar architectures. occur in Adiantopsis: 
e. pinnate, and pedate (Figure 2). The palmate 


(Fio- 


ure 2A). In this species, the pinnae are born from 


palmat 


architecture is observed only in A. radiata 


and no noticeable 


single point at the stipe apex. 


rachis is present. A pinnate laminar architecture 


occurs in A. asplenioides, A. parvisegmenta, A. 


paupercula, A. reesti, A. rupicola, and A. vincenti 


(Figure 2B). Adiantopsis pedata, A. pentagona, and A. 
Xaustralopedata are pedate (Figure 20). As in 
pinnate species, the pinnae of pedate laminae are 
disposed along a central rachis. However, in the 


pedate taxa, the basal basiscopic pinnules of the first 


pinnae pair are considerably longer than the other 


pinnules. In the pinnate and pedate taxa, the lamina 


lies in the same plane as the stipe and rachis. 


However, in Ed A. radiata, the lamina is 


geniculate to the stipe. 


The overall 


shape of Adiantopsis laminae is 
a reflection of architecture. Palmate Adiantopsis 
radiata has an orbiculate lamina. Pinnate taxa are 


variously triangular to lanceolate. linear in 4. 


asplenioides. Both pedate taxa are pentagonal, as 


— 


reflected in the name of A. pentagona. Laminar apices 
in the pinnate and pedate taxa are difform, with the 
exception of A. paupercula, which is conform. 

The laminae in Adiantopsis range from small and 
simple to large and complex. Adiantopsis asplenioides 


has the smallest laminae, with a maximum length of 
14.5 em and a width of 1.1 em. It is also the simplest, 


Of the 


the longest laminae al 


with pinnate-pinnatifid to bipinnate laminae. 
Caribbean taxa, A. reest has 
60.0 cm, 


pentagona at 48.2 cm wide. Adiantopsis parvisegmenta 


whereas the widest laminae are in A. 


and A. paupercula may have quadripinnate laminae 


and are the most compound among Caribbean 


Idiantopsts taxa. However, in South America, laminae 


may be pentapinnate as in A. chlorophylla. With the 


— 


exception of A. asplenioides, most Caribbean Adian- 
topsis laminae are 20—40 cm long and 10-20 cm wide. 

Adiantopsis laminar tissues vary from papyraceous 
to spongiose. Papyraceous laminar tissues are found in 
Adiantopsis asplenioides and 


A. reesit and A. vincenti. 


A. pedata are chartaceous. Adiantopsis rupicola may 
have chartaceous to thin-spongiose tissue, whereas A. 
radiata and A. pentagona have thin-spongiose tissue. 
Finally, spongiose laminar tissue is found in 
parvisegmenta and A. paupercula. 

Randomly dispersed on the adaxial laminar surface 
are white epidermal cells. These cells are hard, and 


probably calcified. Hydathodes are also located on the 


Annals 


376 of the 
Missouri Botanical Garden 


wy 
wy 
= m" 
S p. EN 
XN , 4 
— +4 4 
* 
O 
C ! 
— 


A. 2255 
„ 
. if 


— E 2 * 
7. 0 | 


. its: 


We: 
Los MNT dod: 
Ke) tei 


PU 


. MU). —B. Pinnate lamina morphology 
Wright 3949, MO). Scale 


Figure 


—A, Radiate lamina morphology of ; jii ue radiata (J. I. Main s.n. 
of A. i uiid (W. Palmer 242 & J. H. Riley, US). 


Pedate lamina morpholog y of A. pentagona (C. 


bars = 5 em 


Volume 93, Number 3 
2006 


Barker & Hickey 
Caribbean Adiantopsis (Pteridaceae) 


377 


Table 1. Sample sizes, means, standard 


Taxa that do not share letters are significantly different from each other 


deviations, and ranges (min, max) for Caribbean Adiantopsis guard cell lengths. 


(P < 0.0001). 


Standard deviation (um) 


Min (Um) Max (Um) 


Taxon Sample size Mean (Um) 

1. asplenioides" 50 52.42 
A. parvisegmenta" 125 15.5 

A. paupercula” 125 61.22 
A. pedata' 125 65.80 
A. pentagona' 125 61.04. 
1. radiata? 125 49.71 
|. reesii 125 44.34 
A. rupicola” 125 49.75 
A. rincentit? 50 15.84 


3.78 45.64 57.05 
1.44 32.6 50.60 
5.59 48.9 08.46 
5.88 523.19 78.24 
8.85 32.6 83.13 
1.95 35.06 03.57 
1.49 32.6 53.79 
1.32 42.38 57.05 
3.53 40.75 53.79 


adaxial surface of the laminar tissue. In Adiantopsis. 
these hydathodes are sub-marginal on single vein tips, 
and are generally covered with a hard, amorphous, 
white mass. As Adiantopsis prefers calcareous sub- 
strates, this material appears to be a guttation by- 
product. This material, and the wall constituents of the 
while epidermal cells, is most likely calcium or some 


other mineral collected from the substrate. 


Guard cells are located only on the abaxial side of 


the laminar tissue and are anomocytic. They may be 


surrounded by two, three, four, or as many as five 
subsidiary cells, even on the same ultimate division. 
Length differences do exist among the guard cells of 


(Table 1). The 


A. parvisegmenta, A. 


77 
— 


different taxa guard cells of 
asplenioides. radiata, A. reesii, 
A. rupicola, and A. vincentii form a group of taxa with 


guard cell length means centering around 45 um. 
Adiantopsis paupercula has a mean guard cell length 
of 61.22 Um, whereas A. pedata and A. pentagona 


form a group of taxa with a combined mean guard cell 


— 


length of 66.45 Um. The guard cell lengths of all three 
groups are significantly different from each other (P. < 
0.0001) in Tukey-Kramer pairwise comparisons. 
Pinnae in Adiantopsis are generally triangular to 
fusiform. The triangular basal pinnae of the pedate 
taxa are strongly inequilateral, a function of the much 
extended basal basiscopie pinnules. All other pinnae 


of the pedate taxa are more or less equilateral. In A. 


parvisegmenta and A. paupercula, pinnae are also 
somewhat inequilateral with basiscopic pinnules 
slightly longer than the acroscopic pinnules (Fig- 


Pinnae of A. asplenioides reach 6.1 mm long 


ure 1C). 
and 6.0 mm wide. Those of other species range to 
to 19.9 cm 


alternate, with the exception of the subopposite basal 


24.1 cm long and wide. Pinnae are 


pinnae of the pedate taxa. In A. radiata, the pinnae 
radiate from the stipe apex. Pinnae in Adiantopsis are 
generally ascending, but a number of taxa also have 
patent pinnae. In A. parvisegmenta, A. paupercula, A. 
reesil, A. rupicola, and A, vincentii the proximal pinnae 


are frequently or occasionally arcuate, with only the 


distal pinnae straight. This is quite different from the 
curling observed in senescent fronds of A. rupicola 
(Figure ID). In 


incurved pinnae increase their curvature dramatically 


and A. vincenti these taxa, the 


becoming circinate in appearance. In all other taxa, 
the senescent pinna axes retain their shape, whether 
straight or somewhat incurved. 


Adiantopsis venation and architecture is anadro- 


mous. Pinnules are typically narrowly triangular t 


linear, and occasionally lanceolate. In the pinnate 
taxa, the basiscopic pinnules are slightly longer than 
the acroscopie pinnules and may be up to 7.5 cm long 
and 2.5 cm wide. The basal basiscopic pinnules of the 
pedate taxa are much longer and reach 15.3 cm in 
length. Pinnules may be patent to ascending. 


Adiantopsis ultimate divisions are an important 


source of taxonomic information. For example. the 


ultimate divisions of A. paupercula are strongly 


articulate, with a clear enlargement and differentia- 
tion of the axis at its juncture with the laminar tissue 
(Figure 3A). 
topsis taxa, while articulate with age, lack the swollen 
of A. 


The ultimate divisions of A. parviseg- 


The ultimate divisions of other Adian- 


ind distinct articulation point paupercula 
(Figure 3B). 
menta are by far the smallest of the Caribbean taxa, 
reaching a maximum size of 3.5 mm long by 2.5 mm 
wide. The ultimate divisions of other Adiantopsis taxa 
generally overlap in size, with a maximum length of 
13.5 mm in A. radiata and maximum width of 4.5 mm 
in A. paupercula. There is considerable overlap in 
ultimate division shapes among Caribbean Adiantop- 
sis, with shapes ranging from oblong, to oblanceolate 

Ultimate 


and are often 


lanceolate, elliptic, trullate, and rhombic. 


division bases are generally cuneate 


uniauriculate on the acroscopic side. The apices of 4. 
reesit and A. pedata ultimate divisions are acute and 


distinctly toothed, whereas in other taxa they are 


generally acute to round and entire. Adiantopsis 


asplenioides ultimate divisions are distinctly lobed, 
laxa are entire to 


whereas the margins of othei 


crenulate (Figure 3C, D, E). All ultimate divisions are 


378 Annals of the 
Missouri Botanical Garden 


Figure 3. . Swollen stalk apex at junction with lamina tissue (arrow) of Adiantopsis | cula (C. Wright 962, UC). Ultimate 
divisions are articulate and break cleanly at this swollen point. Scale bar = | mm. —B. Non- 10 n stalk apex al bem with 


7 


lamina tissue (arrow) of A. pentagona (J. G. Jack 7903, US). Ultimate divisions are articulate, but do not break cleanly from the st 
Scale bar = 0.3 mm. —€C. Toothed ultimate division apices of A, reesit (E. L. Ekman 10608. US). —D. Round. entire ieee division 
apices of A. radiata (J. A. Shafer 3777, F). —E. Lobed ultimate divisions of A. asplenioides (Bro. Alain 1226 & Acuna, US). — 

\baxial view of basal flabellate divisions of A. radiata (G. L. Webster 13125, MICH), indicated by arrow. The divisions are attached to 
the stipe apex between pinna pairs. Note similarity of shape between basal i 
basal flabellate divisions, which suggests they are hom 95 to ultimate divisions of the pinnae. Scale bars C-F 


iltimate and flabellate divisions. Sori are visible on these 
5 cem. 


Volume 93, Number 3 


Barker & Hickey 
Caribbean Adiantopsis (Pteridaceae) 


deciduous with persistent stalks ranging from 0.1— 
6.3 mm in length. 

flabellate 
As their 


name implies, these divisions are generally flabellate 


Adiantopsis radiata possesses basal 


divisions, an autapomorphy for this taxon. 


and occur on the 
(Figure 3F). 


to be homologous to ultimate divisions. The 


stipe apex between pinnae 


These basal flabellate divisions appear 
overall 
shape of basal pinnules and these flabellate divisions 
Additionally, 


late divisions become fertile 


is similar (Figure 3F). the basal flabel- 


with the other ultimate 
divisions. They are likely a by-product of axial 
compression in A. radiata, as they appear to be 
ultimate divisions developed within the stipe crown at 
a point below the stipe surface. These basal flabellate 
divisions are usually restricted to mature Á. radiata 
plants with larger than average fronds and more than 
five pinnae. 

Adiantopsis indument consists of hairs and scales. 
The hairs are septate-capitate, with well defined septa 
and bulbous Mn cells. On the 


axes, they are 


concentrated at axis junctions and distributed dif- 
fusely across ds abaxial surface. They are generally 
appressed and may be up to ten cells long, although 
most are much shorter. Septate-capitate hairs are also 
found on the abaxial laminar tissue. Typically, the 
laminar hairs are three celled, 


celled. 


with an elongate basal 


although rarely may 
four These septate-capitate 
cell, 
a relatively short middle cell, and a bulbous apical 
cell. In 
apical cell is very bulbous and overarches the middle 
cell. h 
still 


overlap the middle cell. 


they be two or 


hairs are clavate. 


some taxa, such as A. parvisegmenta, the 


1 A. pentagona and others, the apical cell, while 


bulbous, is generally elongate and does not 


The basal cells are usually 
clear, but the middle and apical cells are variable, 
being red, gold, white, or clear, but usually the two are 
not the same color. While some taxonomic trends in 


color have been observed, they are too variable 


These 


reliably differentiate taxa. septate-capitate 
hairs are present in some putative South American 
Adiantopsis taxa as well and apparently demonstrate 
utility (Ponce & Morbelli, 1989). 


Their ultimate taxonomic use within the genus can be 


some laxonomic 
re-evaluated once all Adiantopsis taxa have been fully 
examined. 

The type of hairs observed in Adiantopsis appears to 
be a synapomorphy for the genus. Tryon and Tryon 
(1982) stated that Adiantopsis is closely related to the 
Cheilanthes microphylla group of Cheilanthes. How- 
ever, a number of differences between the hairs of the 


two groups are observed. Hairs of the C. microphylla 


— 


group are much longer. more filamentous, and with 


indistinct or oblique cross walls, characteristics not 
In the C. 


associated with Adiantopsis. microphylla 


group, both adaxial and abaxial surfaces of the lamina 
and axes are covered with hairs, whereas in 


Adiantopsis the hairs are restricted to abaxial surfaces. 
Additionally, hairs in the Cheilanthes taxa are much 
denser than Adiantopsis hairs. The septate-capitate 
hairs of Adiantopsis have not yet been observed i 
Cheilanthes. 


Adiantopsis scales are found on the rhizome, stipe, 
and laminar axes. Rhizome scales are bicolorous 
(Figure 1A), with persistent, black to dark brown 


centers and deciduous, golden or, in A. asplenioides, 
white, translucent margins. Scales densely cover the 


rhizome surface. These scales are generally lanceate 
slightly 


repand margins, and acute, gland tipped apices. The 


to acicular, with truncate bases, entire to 


rhizome scales of all species are similar and generally 


ack taxonomic value. 


Stipe scales are uniformly concolorous and 
Overall, 


occasionally subulate. 


an. 
the scales range from lanceolate to acicular, 


or 


Adiantopsis stipe scales are 
frequently biauriculate and appear pseudopeltate, but 
some scale bases are not auriculate and are truncate. 
Scale margins range from entire to slightly repand. 
Apices are acute to attenuate and gland tipped. In all 
laxa, stipe scales are deciduous, as the remnants. of 
scale bases are often observed and the number of 
scales varies considerably. These scale characters 
possess relatively little taxonomic value. However, the 
stipe scales of 4. 


parvisegmenta, A. paupercula, A. 


reesil, A. rupicola, and A. vincentii are divergent from 
the stipe surface, whereas A. pedata, A. pentagona, 
and A. radiata stipe scales are appressed to the stipe 
surface, only occasionally divergent. Stipe scales are 


distributed at the base of 


stipes, typically to 


a maximum of 7.1 em above the rhizome. However, 


stipe scales on some A. reesti specimens are 


distributed along the entire length of the stipe to the 
rachis. It is possible that stipe scales are distributed 
higher on the stipes of other taxa and are simply not 
observed because of their deciduous nature. Stipe 
scales are conspicuously absent in A. asplenioides. 

more 


Distal to the stipe, scales are rarer, easily 


overlooked, and are usually no more than five cells 


wide and eight cells long. They are concentrated 


- 


laminar axis junctions and are diffusely scattered on 
abaxial axis surfaces. Scales are typically lanceate in 
overall shape and are gland tipped. Some scales are 
suggesting 


biseriate and closely resemble axis hairs, 


that hairs and scales in Adiantopsis form a develop- 
mental and evolutionary continuum. Laminar scales 
are not taxonomically informative. 

Venation in Adiantopsis is free. In most species, 
veins are pseudodichotomous and anadromous, al- 
though in A. paupercula veins are flabellate. Veins are 


typically obscure to occult and are generally im- 


380 


Annals of the 
Missouri Botanical Garden 


n the laminae. Veins always terminate 


mersed 
marginally or sub-marginally in adaxial hydathodes. 


Adiantopsis sori are marginal and consistently 
Sori both 


basiscopic division margins; in A. asplenioides they 


uninerved. occur On acroscopic and 
frequently occur on the interior sides of lobes. For the 
most part, a single, distinct pseudoindusium covers 
each sorus, although in A. asplenioides, pseudoindusia 
may be partially confluent across sinus bases. The 
generally distinct pseudoindusia distinguish Adian- 
topsis from Cheilanthes, in which a series of confluent 
entire margin of an 
1982). 


pseudoindusia are typically lunate to quadrangular, 


pseudoindusia comprise the 


ultimate division (Tryon & Tryon, Adiantopsis 


with entire lo erose margins. 


Sporangia in Adiantopsis show a mixed develop- 
ment and are generally subglobose and long stalked. 


The stalks are typically three cells long. However, in 


A. asplenioides, and an undescribed Bolivian and 
Paraguayan taxon, the sporangia are globose and 


essentially sessile, with apparently only a single cell 
The 


number of indurated arcus cells shows taxonomically 


layer between the sporangium and receptacle. 
informative trends in Adiantopsis. Indurated arcus cell 
numbers show narrow ranges of variation, with modes 
that are useful for diseriminating 
(Figure 4). 


vincentil 


among laxa 


For example, A. pedata, A. reesti, and A. 


have a mode of 14, A. parvisegmenta, A. 
paupercula, A. pentagona, and A. rupicola have a mode 
of 16, A. asplenioides has a mode of 11, and A. radiata 
mode of 20. 


contain 64. 


has a Adiantopsis sporangia generally 


spores; however A. asplenioides has 32 


spores per sporangium. 


— 


Indurated arcus. cell sizes also differentiate taxa 
into two groups (Table 2). One group consists of the 
pedate taxa, with a mean indurated arcus cell height 
All other 


in the second group. with a mean indurated arcus cell 


of approximately 48 Um. taxa are included 
PI j 


height of approximately 38 Um. These two groups 


were significantly different from each other in Tukey- 
Kramer pairwise Comparisons. 


Adiantopsis spores are tetrahedral-globose and 


generally echinate (Figure 5). The only species to 


deviate from this is A. asplentoides, which has cristate 


spores with anastomosing strands below the muri 


(Figure 5B). 


Among the echinate spored species, 


some notable differences are observed. The echinae 
bases of A. pedata, A. pentagona, and A. radiata are 


whereas the echinae bases in other taxa 


Also, 


all dissected, 


are complete. the echinae of A. paupercula 


spores are more compacl relative to the echinae of 


other species? spores. The laesurae are obscured by 


the orname ntation in all spores exce pl. 4. asplentoides: 


s laesurae are easily observed with < 


i 


dissecting 


microscope. Spore sizes Adiantopsis are divided 


into three distinct groups all statistically different 
Tukey-Kramer 


). The first group, with 


from each other, as determined by 
pairwise comparisons (Table 3 
spore lengths averaging around 37 um, includes A. 
radiata, A, reesii, A. 


parvisegmenta, A. paupercula, A. 


rupicola, and A. vincentii; Adiantopsis pedata and A. 
pentagona compose the second group with average 
40 um. Finally, A. 
asplenioides forms a third group, with a mean spore 


length of 75.83 Um. 


spore lengths of approximately 


The echinate spores of Adiantopsis help distinguish 
Cheilanthes. The 


with only a few 


ib. from spore ornamentation of 


Cheilanthes is variable, Australian 
species possessing echinate spores (Tryon & Tryon, 
1982). light of these echinate 
Cheilanthes, Tryon (1982) 


Idiantopsis based primarily on the echinate spore 


Even in Australian 


Tryon and recognized 


ornamentation, as the echinate Australian Cheilanthes 
are easily distinguished from Adiantopsis by vegeta- 
live characters. The only exception to echinate spores 
in Caribbean Adiantopsis is A. asplenioides. Given Ms 
other unique characters, A. asplenioides may belong in 
another, perhaps new, genus, although some putative 
South American Adiantopsis taxa that are currently 
placed in Cheilanthes also have cristate spores (Ponce 
& Morbelli, 1989). Adiantopsis asple nioides is retained 
f South 


asplenioides is 


here only until a full analysis o American 


Adiantopsis is completed. If n 
segregated from Adiantopsis, echinate spores would 
characterize the genus. 

Taxonomic Hisrory 
In 1852, 


various species 1n 


Fée erected the genus Adiantopsis for 
Cheilanthes, 
Hypolepis Bernh. with distinct pseudoindusia for each 


Adiantum 1... and 


sorus. Originally, Fée included four species in 


Adiantopsis: A. capensis (Sw.) Fée, A. chlorophylla, 


paupercula, and A. radiata. He designated A. 


paupercula as the type species for the genus. 
Christensen (1900) later lectotypified the genus with 
l. radiata; this was unnecessary because Fée (1852: 
145) clearly designated (with the term Diagnosis) A. 
paupercula by referencing Kunze's (1850) publication 
of the species. As a result, some authors (e.g., Mickel 
& Smith, 2004) recognize A. 


whereas others (e.g., Tryon & 


radiata as the type, 
1982; Proctor, 


A. paupercula. It seems clear to us 


Tryon, 
1985) recognize 
A. paupercula should be recognized as the type 
species for the genus, as Fée unambiguously desig- 


nated it as such. 


Fée (1857) also published two subgenera within 
Idiantopsis: Euadiantopsis Fée (= Adiantopsis) and 


Cheilanthastrum Fée. Although he did not provide 


characters for the subgenera, an evaluation of the list 


75 — 75 — — 75 — 
60 —— 60 60 E 
45 45 45 
30 30 — 4 
15 - — 15 — 
Q T TT LE rT TT TT EFT T T T Td T T1 T T TT 
9 10 11 12 13 14 15 16 17 18 19 20 2122 23 24 25 26 27 9 10 1112 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 9 10 1112 13 44 15 16 17 18 19 20 21 22 23 24 25 26 27 
75 — 75 75 
> 9 enc - — 
Q^ 
i 
H | 
l III. 3 
| y m 1 iT INS EST i | | d I E faa d | a od OP | 
awoke IC P ERN Us 9 10 1112 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 
75 -— — 75 
60 -1— 6 60 — 
45 45 as tp. 
30 — 80 
15 — $5 
"b | d bou 1 | | T if bt i T E T T 1 1 0 | d | ot d d | od Iu 
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 9 10 1112 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 9 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 
Figure 4. Frequency histograms d ads da arcus cell number. X-axis is arcus ce 95 number, Y-axis i frequency (absolute value). —A. A. asplenioides 
parvisegmenta. —C. A. paupercula. — . pedata. —E. A. pentagona. —F. A. radiata. —G. A. reesii. —M. A. rupicola. —1. A. vincenti. 


Ww 


900 


€ JequinN 'ee aunjoA 


(aes9ep11914) sisdojueipy ueeqqueo 


FOH Y jexeg 


18€ 


382 Annals of the 
Missouri Botanical Garden 


Table 2. Sample sizes, means, standard deviations, and ranges (min, max) for Caribbean Adiantopsis arcus cell height. 


laxa that do not share letters are significantly different from each other (P < 0.0001) 
Taxon Sample size Mean (Um) Standard deviation (Um) Min (Um) Max (Um) 

|. asplenioides* 50 34.38 2.47 30.97 39.12 
|. parvisegmenta" 125 10.78 5.19 32.6 8.9 
|. paupercula 125 10.59 13 35.80 17.27 
|. pe data 125 18.8 1:00 10.75 55.42 
|l. pentagona 125 18.06 1.46 10.75 30.00 
L radiata" 125 39.04 2.76 30.07 10.46 
|. reesit? 125 30.3 6.47 21.19 18.9 
i. d 125 37.16 5.33 32.6 12.38 
|l. vincenti" o0 38.01 131 32.6 52.16 


af 


la 
9. 


* 


Figure 5. Electron micrographs of Caribbean Adiantopsis spores. —A. Echinate A. radiata spore with dissected echinae 
S (LA o TITAN). —B. Cristale A. asplenioides spore with anastomosing str beneath nuri (Br. Alain A1220. 


. Echin: p: I. paupercula 


Fehinate A. parvisegmenta spore with complete echinae bases (J. G. Jack 66877. US). 
spore wi 1 a (Maxon 42539. US). | Echinate A. reesit spore with complete echinae ibi „ Ekman 9987 


E). —F. Echinate A. rupicola spore with complete bs ie bases (Britton et al. 3407, US). Echinate I vincentit spore 
oe echinae bases (Egger 5399, V), —H. Echinate A. pedata spore with dissected echinae bases (Harris 10878, US) 
Behin: |. pentagona spore with dissected echinae bases (C. Wright 3940, US). Scale bars = 20 Un 


Volume 93, Number 3 
06 


Barker & Hickey 
Caribbean Adiantopsis (Pteridaceae) 


Table 3. 


Sample sizes, means, standard deviations, and ranges (min, 100 of ( 


Caribbean Adiantopsis spore lengths. Taxa 


that do not share letters are significantly different from each other (P < 0.0001) 
Taxon Sample size Mean (um) Standard deviation (Um) Min (Um) Max (um) 

A. asplenioides* 50 75.83 6.74 58.08 88.02 
A. parvisegmenta* 125 40.59 2.41 34,23 47.27 
A. paupercula" 125 41.4 2.96 35.86 15.64 
A. pedata* 125 45.5 3.85 38.31 4.61 
A. pentagona* 125 45.11 3.4 40.75 58.9 
A. radiata” 125 39.94. 2.35 34.23 17.27 
A. reesii" 125 36.34 2.44 30.97 14.83 
A. rupicola” 125 37.75 1.68 34.23 40.75 
A. vincenti" 50 38.16 3.07 32.0 14.01 


of species he included in each subgenus identified 
Eua- 


diantopsis (Adiantopsis radiata, A. paupercula, and A. 


possible distinguishing characters. Subgenus 


pteridoides Moore |= Cheilanthes pteridoides Sw.|) is 
apparently characterized by more oblong ultimate 
divisions, whereas those of subgenus Cheilanthastrum 
chlor- 


ophylla|. and A. capensis |= Cheilanthes E 


(Adiantopsis chlorophylla, A. spectabilis |= A 


are more elliptic to oblanceolate. Additionally. two of 


the taxa in subgenus Fuadiantopsis, A. paupercula and 


A. radiata, have ultimate division shapes that are 


more like those found in Adiantum than in Chei- 
lanthes, whereas the species placed in subgenus 


Cheilanthastrum have ultimate division shapes more 
similar to Cheilanthes. Although these subgenera are 
not recognized here, they do reflect morphological 
trends that have caused taxonomic confusion in the 
past. 

Past authors have not always recognized Adiantop- 
sis. For example. Hooker and Baker (1883) placed 
Adiantopsis as a subgenus within Cheilanthes in their 
Synopsis Filicum. However, they did use Fée's 
characters to recognize subgenus Adiantopsis. The 
placement of Adiantopsis within Cheilanthes is in line 
with their overall generic concept, which recognized 
a small number of large fern genera with a substantial 
number of subgenera. Copeland (1947) also did not 
that it 


recognize Adiantopsis, suggesting was in- 
sufficiently distinct to warrant generic status: he 


placed it in synonymy under Cheilanthes. 

Other authors have recognized Adiantopsis, but with 
qualification. For example, Smith (1981) suggested 
that if 


seemannii (Hook.) Maxon are included in Adiantopsis, 


species such as A. chlorophylla and A. 
the distinction between it and Cheilanthes disappears. 
Specifically. he cited the morphological similarities of 
these species to C. aemula and C. microphylla. 

Many other researchers have recognized Adiantop- 
s as construed by Fée. Christensen (1906) recog- 
Filicum. Maxon (1908) also 


recognized Adiantopsis and published a useful synop- 


=. 


s 


nized it in his /ndex 


sis of the Cuban taxa. Most modern tropical fern floras 
Adiantopsis as distinct. from Cheilanthes 
1959, 1961, 1972; Tryon & Tryon, 1982; 
Proctor, 1985; Lellinger, 1989; Proctor, 1989; Tryon 
el al., 1990; Pacheco, 1995: Mickel € Smith, 2004). 


In general, these researchers, and Fée, utilized a more 


recognize 


(Sehnem, 


narrowly defined generic concept than either Hooker 
and Baker (1883) or Copeland (1947 


1947) articulated problems with the recognition of 


). Only Copeland 


Adiantopsis, although he did not refer to a number of 


distinctive characters, such as spore ornamentation, 
adaxial carinae, and hair morphology. 
Although 


fern floras and indices, little comprehensive research 


Adiantopsis has been included in many 
has been done at the species level. One of the most 
significant. contributors to species level taxonomy in 
Adiantopsis was Sehnem. He published a series (1959, 
1961. 1972) of keys 


various floras in southeastern Brazil and the tri-border 


and species descriptions for 


region of Argentina, Brazil. and Paraguay. In these 


treatments, he provided excellent kevs to the species 
of the region using characters that were overlooked by 


other Further, he 


pteridologists. provided some 
clarification to the A. chlorophylla species complex 


by recognizing lwo new species, A. perfasciculata and 
A 8 8 I per 

Unfortunately, most pteridologists, in- 
(1982), 


Sehnem’s Adiantopsis research. 


A. occulta. 


cluding Tryon and Tryon were unaware of 


Tryon and Tryon (1982) were the most recent 


axonomists lo review Adiantopsis. They regarded the 
genus as closely related to their Cheilanthes micro- 
phylla group (their Cheilanthes group 1). That group 
and Adiantopsis share dark axes and less hair than 
other cheilanthoid groups. The main character used 
by Tryon and Tryon (1982) to distinguish Adiantopsis 
from Cheilanthes was the echinate spores in Adian- 
Additionally, 


laminar axes, 


topsis. they noted adaxially bicarinate 


asymmetrical and articulate ultimate 


divisions, and distinct indusia covering single sori as 
distinguishing features of Adiantopsis (Tryon & Tryon, 


1982) 


Annals of the 
Missouri Botanical Garden 


Tryon and Tryon (1982) recognized seven Adian- 
topsis o in "us including Adiantopsis pauper- 
cula, A. pedata, radiata, and A. reesii from the 
ed In 175 ‘ir treatment, they sank A. rupicola 
into A, reesii, stating that it does not appear to be 
asplentoides to be 


With 


distinct. and considered A. 


regard lo 


a juvenile form of A. paupercula. 


South American taxa, Tryon and Tryon (1982) did 
recognize one of Sehnem’s pos A. minutula, 


but did not acknowledge his A. perfasciculata and 


. occulta. It is not clear why they did not 
recognize these two distinctive species. Although 
Tryon and Tryon's (1982) treatment of Adiantopsis 


ailed to mention some key species, they provided the 


most comprehensive overview of the genus ever 
published. 

Recent molecular work by Gastony (unpublished 
data) supports the generic status of Adiantopsis. A 
molecular analysis with rbcL DNA sequences supports 
a sister relationship between A. radiata and the South 


American A. chlorophylla with 23 synapomorphies 


and 100% bootstrap support (Gastony, unpublished 
data). The 
cheilanthoid genera from all geographic regions and 


analysis included representatives of 


provides support that Adiantopsis is a distinct genus. 


Further, it refutes Smith's (1981) contention that 


including A. chlorophylla Adiantopsis makes the 
genus indistinguishable from Cheilanthes because of 
morphological similarity to C. aemula. Cheilanthes 


aemula was included in Gastony’s (unpublished data) 


analysis and was distantly related to the Adiantopsis 


group. Thus, it appears that any morphological 
similarity observed by Smith (1981) between the two 
taxa is likely convergent. Additional molecular 


research with a broader sampling of Adiantopsis is 
needed to evaluate the boundaries of the genus, but it 
does appear to be phylogenetically distinct from other 
Cheilanthes. 


SPECIES CONCEPT 


In this revision 


M Adiantopsis, a morphological 


species concept was employed discriminate spe- 


cies. Essentially, if groups of collections were morph- 


ologically distinct they were recognized as distinct 


taxa; if not, they were maintained as a single species. 


Implicit in this usage of the morphological species 
concept is the assumplion that morphological dis- 


tinctness reflects underlying genetic distinctness 


produced as a result of unique evolutionary histories 


If future research using molecular 


for each species. 


data, such as population genetie markers that sample 


multiple loci across the genome, does not identify 


partitioned genetic variation, then species boundaries 


should be re-evaluated. 


EVOLUTIONARY RELATIONSHIPS 


Based on a comparison of morphological data 


obtained from Caribbean collections, a total of nine 


morphological species are recognized for this region. 
seven are putative diploids 


Of these nine species, 


and two are putative polyploids. In Adiantopsis, 
chromosome counts are ap only from A. radiata. 
The counts of n = 30 (Walker, 1973: Smith & 


Foster, 1984) represent the base erm for the 


cheilanthoid genera, and thus it is assumed that 4. 
radiata is a diploid species. As material was nol 
available to count the chromosomes of other Adian- 


axa, indirect estimates of ploidy level were 
(1986) thal 


guard cell and spore lengths in leptosporangiate 


Lopsis 


used. Barrington et al. demonstrated 


ferns frequently correspond to ploidy level. Indurated 


arcus cell size has also been correlated with 


Using these 


= 


ploidy level by Butters and Tryon (1948). 
cell sizes and spore length (Tables J. 2. 3), the ploidy 
levels of Adiantopsis taxa were estimated, with A. 
radiata serving as a reference for the diploid ploidy 
level. These data partition the nine species into two 
statistically significant groups. All pinnate taxa and A, 
radiata form a group of putative diploid species. 
whereas the two pedate taxa appear to be polyploid 
species, most likely tetraploids. The pedale taxa 
consistently possess larger cell sizes. The cell length 
differences between putative diploid and tetraploid 
comparable to differences 


Adiantopsis laxa are 


observed between diploid and tetraploid leptospor- 


— 


angiate ferns by Barrington et al. (1986) and Butters 
and Tryon (1948). 
and A. 


species, 


Based on these data, A. pedata 


pentagona are supported as tetraploid 


whereas all other Caribbean taxa appear to 
be diploid. 

Two exceptions to this pattern in cell size exist. 
Guard cells of Adiantopsis paupercula are somewhat 


but not 


arger than all other putative diploid taxa. 
quite as large as those of the putative tetraploid laxa 
(Table 1). Spore and indurated arcus cell size place 
this spec ies with the putative diploid laxa (Tables 2 25 
3). Hs not 


intermediacy. 


clear why its guard cells show this 


that ils 


arcus cell sizes are within the putative diploid range. 


Given spore and indurated 


it seems probable that this is a diploid taxon and that 


its guard cell lengths are influenced. by other, 
unknown factors. 
The other. exception Adiantopsis asplenioides, 


a species with a number of unique characteristics 
cell 


durated arcus cell sizes group this taxon with other 


among Caribbean Adiantopsis. Guard and in- 


putative diploids (Tables 1, 2). However, it has the 
largest spore size of all Adiantopsis taxa (Table 3). it is 


the only species with 32 spores per sporangium, and it 


Volume 93, Number 3 Barker & Hickey 385 
2006 Caribbean Adiantopsis (Pteridaceae) 
has indurated pseudoindusia that completely and and generally ascending ultimate divisions. Addi- 


the sori. 


explain these anomalous features. First, 


persistently cover Two hypotheses may 


the species 
may be apogamous, a hypothesis supported by the 


1964: Klekowski. 


number 


reduced spore number (Evans, 
1973). Alternatively, 


a function of a reduced number of spore mother cells 


the spore may be 


and a greater resource allocation fewer spores. 
Under this hypothesis, the large spores are probably 
not indicative of a higher ploidy level but rather are an 
adaptation for limiting spore dispersal. Barrington el 


al. (1986) observed 


pedatum ssp. calderi Cody, a fern that is limited to 


similar pattern. in. Adiantum 


serpentine rocks. He hypothesized that the large 


spores may limit spore dispersal and thus prevent 


hem from leaving the “serpentine islands” to which 


hey are adapted. Adiantopsis asplenioides vici to 
al/s (1986) model well. it ds 


a serpentine species with very large spores e based 


fit Barrington e 


on other features appears to be a diploid. Further. the 
persistent and indurated pseudoindusia of A. asple- 
nioides may represent an | additional modification to 
the 


number, large spores, and indurated pseudoindusia 


limit spore dispersal. Thus, reduced spore 
appear to represent a suite of characters adaptive in 
keeping propagules within its geographically restrict- 
ping 8 Beogra] à 
ed edaphic niche. 
Adiantopsis pedata and A. pentagona are hypothe- 
H 8 ) 
between A. 


sized to be allotetraploid derivatives 


radiata and two different pinnate taxa. A hybrid 
origin for the pedate Caribbean taxa is supported by 
the occurrence and morphology of the sterile hybrid 4. 
Xaustralopedata in South America. This taxon docu- 
ments that the pedate laminar morphology is of hybrid 
origin between A. radiata and a pinnate second parent 
(Hickey et al., 2003). Also, 


bases on spores of A. radiata and the pedate taxa 


the dissected echinae 


support A. radiata as one parent, as all other taxa have 
Additional 


evidenced by common laminar development in the 


complete echinae bases. support is 


two pedate taxa and A. radiata. The fronds of young 
plants, as estimated by rhizome size, are strongly 
ternate in both species; progressively older, larger 
plants of these species show increasingly complex 
frond development. In A. pedata the basal basiscopic 
pinnules become extended and the penultimate pinna 
pairs enlarge, becoming increasingly prominent. In A. 
radiata, additional pinnae pairs are added to the stipe 
This pattern of growth is only found in A. 
radiata, A. Xaustralopedata, and the pedate Carib- 
bean taxa. The pinnate Adiantopsis taxa lack such 
ontogenetic architectural changes and show a common 
architecture on both young and old plants. 

the 


parent of A. pedata on the basis of crenulate apices 


Adiantopsis reesii is supported as second 


tionally, the carinae of A. reesii and A. pedata start at 
stipe bases or in the basal half (A. pedata) or quarter 
(A. reesii) of stipes. Both species also share a mode of 
An A. reesii X A. 


pedata is further supported by the 


14 indurated arcus cells. radiata 
origin for A. 
distributions of the taxa, as collections of A. pedata 
are currently only known from areas where the three 
CO-OCCUL. 
Numerous morphological comparisons support 
Adiantopsis rupicola as the most likely second parent 


of A. pentagona. The overall ultimate division shape of 


A. pentagona compares most favorably with the 
generally trullate divisions observed Adiautopsis 


rupicola. Furthermore, the acute and generally entire 
ultimate. division apices of A. pentagona reflect that 
character state in A. rupicola. The generally patent 
nature of A, pentagona ultimate divisions compares 


favorably with the patent disposition of the ultimate 


= 


divisions in A. rupicola, and both A. rupicola and 
pentagona have a mode of 16 indurated arcus cells. 
Finally, the carinae in A. rupicola and A. pentagona 

start in the upper portions of stipes (upper Ys in 4. 
pentagona and % in A. rupicola), a character unique 
to these two species. 

Adiantopsis pentagona, however, is quite variable, 
and there are suggestions that it contain 


may yet 


another 


eryptie species. Evidence for this comes 
from some A. pentagona specimens that have 
relatively small ultimate divisions, larger and 
slightly more compound laminae, and somewhat 


revolute fertile divisions. It should be noted. 


however, that all of these character states tend to 
intergrade across the suite of specimens referable to 
this 


is not clear from 


A. pentagona, and 


morphological analysis whether the observed varia- 


lion is due to habitat and ecological differences 


among plants, simple genetic variation, or multiple 


origins with different maternal parentages (e.g.. 
rupicola the maternal parent for some specimens and 
A. radiata for others). Alternatively, A. pentagona 
may be a case of cryptic allotetraploids wherein two 
distinct species share a common A. radiata parent, 
but each has a different pinnate second parent. The 
variation in character morphology mentioned above 


could easily be accounted. for if A. parvisegmenta 


The 


geographic distribution of specimens does not pro- 


Was i second collections. 


parent of some 


vide any insight about the source of the morpholog- 
ical variation, and resolution may require molecular 


analyses. Future molecular studies of this group 


should encompass the morphological range o 


pentagona specimens, Á. rupicola, Á. parvisegmenta, 
and A. vincentii, because any one could be a genetic 


contributor. 


386 Annals of the 
Missouri Botanical Garden 
TAXONOMY fusiform, widely rhombic, oblong, or obovate. 


Adiantopsis Fée, Gen. Filie. [Mém. Foug. 5| 145. 
1852. TYPE 


|= Adiantopsis paupercula (Kunze) Fée]. 


Adiantum pauperculum Kunze 


Rhizomes erect, ascending, decumbent, or repent, 


stiff; scales dense, overlapping, persistent, bicolorous 


with a black. dark brown. or atrocastaneous center 


and ephemeral, golden or translucent margins. 


acicular, lanceate. lanceolate, linear, Or rarely 


subulate, basifixed or pseudopeltate, bases truncate 


or cordate, margins entire, slightly 


repand, apices acute, attenuale, 


cuneate, gland tipped. Fronds aggregated, strict or 


erecl. Stipes persistent, alropurpureous, alrocasta- 


neous, or castaneous, always, occasionally, rarely. e 


" 
never bicarinate adaxially, with scattered. pneuma- 
tothodes: carinae when present, golden: scales when 
present, deciduous, divergent or appressed, con- 
colorous, tan, or bicolorous with broad castaneous to 
tan center and golden margins, acicular, lanceate. 
subulate, pseudopeltate o 


lanceolate, or rarely 


basifixed, bases biauriculate, truncate, or cordate, 


margins entire, slightly repand, or repand, apices 
acute or attenuate, gland tipped: hairs diffuse, rare. 
septate-capitate or rarely eiliform. Laminae pinnate, 
pedate, or palmate, drying brown, green, dull green. 
black, 


apices difform or conform, attenuate, long attenuate. 


dark green, yellow-green, olive-green, or 


acuminate, acute, or cuneate; laminar tissue papyr- 
aceous, chartaceous, thin-spongiose, or sponglose. 


adaxially glabrous with diffuse, glossy-white epider- 


mal cells, hydathodes marginal or in sinuses 


surrounded by white to pale blue epidermal cells: 


hairs diffuse abaxially, septate-capilate, three 


celled: stoma abaxial, complanate or impressed. 


anomocytie. Laminar axes persistent or marcescent 
with costae curling acroscopically, atropurpureous. 
alrocastaneous, or castaneous, grading into lamina 
and taking its color and texture, or not grading into 
bicarinate; carinae golden or 


lamina, adaxially 


colored as axis, interrupted at axis junctions, 
grading into lamina or not grading into lamina: 
scales when present, concentrated at rachis-costa 
junctions, diffuse abaxtally, deciduous. appressed or 


patent, concolorous, tan, linear, lanceate, narrowly 


lanceolate, narrowly rhombic, narrowly triangular. or 


biseriate. basifised. bases truncate, acute. or 


cuneale, margins entire, apices acute or attenuate, 
gland tipped; hairs concentrated at axis junctions, 
diffuse abaxially, septate-capitate. Pinnae ascend- 
ing, patent, or spreading radially from stipe apex, 
incurved in 


alternate, sub-opposite, or opposite, 


basal half of lamina and straight distally, or straigh 


` 


narrowly triangular, linear, triangular, narrowly 


repand, or 


acuminale, or 


inequilateral, slightly inequilateral, or equilateral, 


apices attenuate, acute, or round. Pinnules anadro- 


mous, slightly ascending, patent, or ascending. 


linear, narrowly triangular, or narrowly lanceate, 
apices acule, cuneate, attenuate, or obtuse. Ultimate 
divisions patent, slightly ascending, or ascending. 
deciduous, articulate or strongly articulate, oblong, 
oblong. oblanceolate, obovate, narrowly 


flabellate, 


narrowly 


elliptic, elliptic, trullate. lanceolate, 


ovate, widely rhombiform, or rhombiform, bases 


cuneate, acute, or truncate, may be uniauriculate 


acroscopically, margins entire, crenulate, slightly 


erenulate, or deeply lobed, apices acute, round, or 


— 


acuminate; stalks persistent, may be swollen. a 
junction with laminar tissue. Veins free, pseudodi- 
chotomous and anadromous or flabellate, immersed 
or impressed in lamina, occult, obscure, or promi- 
nent. Sori marginal on both acroscopic and basi- 
scopic margins, uninerved. Pseudoindusia distinct or 
occasionally confluent, lunate, quadrangular, re- 
niform, or subreniform, green to white proximally 
and hyaline distally, black maculate, hyaline, or 
white to pale blue, margins slightly erose, entire, or 
erose. Sporangia subglobose to globose, long stalked 
) 


o sessile. Spores O or 32 per 


sporangium, 
tetrahedral globose, golden or white at maturity. 


white to pale yellow when immature, echinate o 


densely echinate with echinae bases complete ( 


dissected, or eristate with anastomosing strands 
below the muri, laesura obscured by ornamentation 
or nol. Chromosome number 2n = 60 (Walker, 1973: 


Smith & Foster, 1984). 


Generic distribution. The Caribbean and central 
Mexico to Central America, and through northern South 
America to northern Argentina (Tryon & Tryon, 1982). 
For many Adiantopsis species, collection data are not 
clear enough to determine their modern political 
districts. For collections that indicate an historical 
political district with insufficient data to determine its 
modern location, the historical district is included in 
the specimens examined lists. Collections lacking data 
to determine a political district at all are indicated as 
“unknown” in the specimens examined lists. 

Fée (18 


type 


UI 


Vomenclature. 2) designated Adiantopsis 


paupercula the species by unambiguously 
referencing Kunze’s 1850 publication of Adiantum 
(1906) later 


lectotypified the genus with A. radiata. Fee's (1852) 


pauperculum. However, Christensen 
original typification is recognized as correct here. For 
more details, see previous discussion in the Taxo- 
nomic History section. 
Distinguishing characters. Adiantopsis is distin- 


guished from Cheilanthes by the following combina- 


Volume 93, Number 3 Barker & Hickey 387 
2006 Caribbean Adiantopsis (Pteridaceae) 


tion of characters: echinate spore ornamentation distinct in color and position from laminar tissue; and 


(excluding A. asplenioides); separate and distinct finally septate-capitate hairs with distinct, transverse 
pseudoindusia, each covering a single sorus; an septae and bulbous apical cells, these distributed 
adaxial pair of golden carinae. that occur on all diffusely on abaxial surfaces of axes and laminar 


laminar axes and generally on stipes, and which are tissue, although often concentrated at axis junctions. 


KEY TO THE SPECIES OF CARIBBEAN ADIANTOPSIS 


ds eee d y a aa Agee eee? 6. A. radiata 
lb. Lamina pedate or pinnate. 


2a. Lamina pinnate. 
: Pseudoindusia e with whitened to pale blue cells; sporangia sessile; lamina once pinnate to 
bipinnate; ultimate divisions deeply lobed 2... 0.000.000.0000 Ii... . . .. A. asplenioides 
3b. Pseudoindusia not 1 d. hyaline to green: sporangia stalked; lamina bipinnate or more compound: 
ultimate divisions entire to crenulat 
4a. Laminar apex conform; ae divisions with stalk swollen at juncture with laminar 98 
h ⁵ ( ẽè̃ͥ᷑72ẽœ„ A. paupercula 
Ab Laminar apex difform; ultimate divisions without stalk swollen at juncture with TEN tissue. 
Lamina quadripinnate; ultimate divisions less than 3.5 mm long and 2.5 mm wide; fertile 
divisions strongly revolute 0⸗j;n;iſDMvꝛ:mim:ꝛeqnuwaõẽm ee 2. A. parvisegmenta (new species) 
5b gae tripinnate or less one ultimate divisions more than 3.5 mm long and 2.5 mm 
wit de: $ ferti tile divisions more or less flat. 
Ga. Carinae beginning in upper quarter of stipe; laminar tissue thin-spongiose to 
chartaceous; mode of 16 indurated arcus cells ooo... 8. A. rupicola 
0b. Carinae beginning at base of or in lower half of stipe: laminar tissue papyraceous; mode 
of 14 indurated arcus cells. 
Ta. Apices of ultimate divisions crenulate; rachis basal diameter 1.1-1.5 mm; 
ultimate divisions ascending: laminar axes persistent, the costae straight; Jamaica 
and. Hispaniola, ica ato hn ome eae ea ua a 1. A. reesit 
7b. Apices of ultimate divisions entire; rachis basal diameter 0.6-1.2 mm; ultimate 
divisions patent to slightly ascending; laminar axes marcescent, the costae curling 
acroscopically: (Cuba O aa ee se we eae der. 9. A. vincentil (new species) 
2b. Lamina pedate. 
8a. Apices of ultimate divisions crenulate: carinae beginning in basal half of stipe: mode of 14 annulus cells: 
ultimate division stalks 0.1-0.6 mm long: ultimate divisions slightly ascending: basal basiscopic 
pinnules 2.2—5.] times longer than basal acroscopic pinnule: Jamaica and Hispaniola . . . . . A. pedata 
8b. Apices of ultimate divisions acute to round; carinae beginning in apical half of stipe; mode of 16 e 
cells: ultimate division stalks 0.9-6.3 mm long: ultimate divisions more or less patent; basal basiscopic 
pinnules 1.8-13.1 times longer than basal acroscopic division; Cuba . . . . . 5. A. pentagona (new species) 
1. Adiantopsis asplenioides Maxon, Amer. Fern J. short, tan or clear, apical cell elongate-bulbous, clear, 
22: 14. 1932. TYPE: Cuba. Pinar del Río, Río white, or golden; stoma complanate, guard cells 
del Medio, at Las Pozas, very scarce, 10 Sep. 45.64—57.05 um long. Laminar axes persistent, atro- 
1923, E. L. Ekman 17456 (holotype. US!) — castaneous to castaneous, grading into lamina and 
Figure OA, B. taking its color and texture; rachis 0.2-0.5 mm basal 
diam.; carinae colored as axis or golden, grading into 
Rhizome erect, to 1.5 em long, 0.3-0.8 em diam.: lamina; scales absent. Pinnae patent to slightly 
scales bicolorous with an atrocastaneous center and ascending, 16-20 pairs, alternate, sub-opposite ba- 
translucent. margins, acicular to lanceate (subulate), Sally, widely rhombiform, obovate, to oblong, slightly 
basifixed. 1.60-5.20 mm long, 0.12-0.55 mm wide, — inequilateral to equilateral, to 2.7-6.3 mm long, 1.7— 
bases truncate, margins entire to slightly repand, CO mm wide, apices round to acute. Ultimate 
apices attenuate. Fronds erect, 9.5-18.5 cm long. divisions patent to slightly ascending, articulate, 
Stipes atrocastaneous to castaneous, never bicarinate widely rhombic, flabellate, obovate, to oblong. to 
adaxially, shorter than lamina, 1.5-4.0 em long, 0.2— 2.0-0.3 mm long. 1.5-6.0 mm wide, bases cuneate, 
0.6 mm basal diam.: carinae absent; scales absent; margins deeply 2-6 lobed, apices round to acute; 


hairs rare, septate-capitate or rarely ciliform. Laminae — stalks persistent, not swollen at junction with laminar 
pinnate-pinnatifid to bipinnate, linear, drying dull tissue, 0.1-0.4 mm long. Veins pseudodichotomous, 
green, 8.0-14.5 cm long, 0.5-1.1 em wide, apices anadromous, immersed in lamina, obscure to occult. 


yr confluent across sinuses, 


> 
^ 


difform, long attenuate; laminar tissue chartaceous, — Pseudoindusta distinc 
hydathodes marginal in sinuses, surrounded by  reniform to sub-reniform, white to pale blue, occa- 
prominent white to pale blue epidermal cells; hairs sionally hyaline, 0.50—0.78 mm long, 0.36-0.50 mm 


with basal cell elongate, clear or white, middle cell wide, margin entire to slightly erose. Sporangia 


Annals of the 
Missouri Botanical Garden 


388 


Figure 6. —A. Adiantopsis asplenioides (Bro. Alain 
divisions with 1 d pseudoindusia. Scale bar = 0.5 em. 
30 


bar = 3 em. —D. Small, revolute ultimate divisions of. A 


receptacle, arcus of 11-15 
indurated 30.97-39.12 um tall. 


32 per sporangium, tetrahedral-globose, white to pale 


globose, sessile on 
cells, these Spores 


yellow at maturity, white to pale yellow when 
immature, 58.08-88.02 um diam., cristate with anas- 
tomosing strands below the muri, laesura not obscured 
by ornamentation, Chromosome number unknown. 


Distribution and habitat. A rare endemic of Pinar 


del Río and an unknown site in eastern Cuba. 
Restricted to moist, serpentine soils or rocks. In 
Pinar del Río, Cuba, the species is found on 


serpentine mogotes. Although there are few collec- 
tions of this geographically and edaphically restricted 
taxon, it is likely still extant in Cuba and protected 
within the confines of Valle de Viñales and other 
national park 

m asplenioides is the only Carib- 
than 


Notes. 


"M Adiantopsis to possess cristate, rather 


echinate, spores. It shares this trait with some putative 


South American Adiantopsis, such as Cheilanthes 
dichotoma Sw. (Ponce & Morbelli, 1989). The ultimate 
generic position of these cristate taxa will be 


determined by ongoing research: investigating South 
American Adiantopsis and morphologically similar 
Cheilanthes. 

Adiantopsis asplenioides is morphologically similar 
and 


to an undescribed Bolivian Paraguayan taxon 
(Barker, pers. obs.). Both 


laxa have sessile, elobose 


1220 & 
C. Adic 


parvisegmenta. 


Lobed ultimate 
IS). Scale 


. —B, 
Morton 1166, 


Acuna. US). Scale bar = 3.0 cm 


p 


tantopsts 3 (C. V. 
Seale bar = 0.15 en 


sporangia, deeply lobed margins, white to pale yellow 
and prominent white or blue epidermal cells 
surrounding the 


distinctive. and their relationship with the rest. of 


spores, 
hydathodes. These two taxa are 
Adiantopsis should be tested with molecular data. 
Adiantopsis asplenioides is unique among Caribbean 
Adiantopsis in having 32 large spores per sporangium 
Although it 


this species is likely a diploid, as guard 


and indurated pseudoindusia. has the 
largest spores, 
and indurated arcus cell sizes place it among other 


putative diploid taxa. [tis possible that the large spore 
size and persistent, indurated) pseudoindusium are 


distribution to. the 


adaptations tọ restrict spore 
serpentine substrates on which it occurs. The species 
may also be apogamous, a common cause of reduced 
spore number. 


del Río: 
Bro. Alain 
pus & Alain 15787 


. Wright 681 (US). 


Pinar 


CUBA. 


Sierra del Rosario, 


Select 
Loma Zambumbia, 
1175 S); La Cajalbana, La Palma, 
. Eastern Cuba: Cuba Orientali, 


ed specimens examined. 
Rangel, 


n 


Barker € 
Villas 


Buenos 


2. Adiantopsis parvisegmenta M. 5. 
Hickey, sp. TYPE: Cuba. Las 
(formerly Santa Clara), Trinidad Mins., 
Aires, 900 m. C. 
Morton 4166 (holotype, US! isotypes, GH! UC!, 

S). Figure 6C, 


nov. 


rare fern on rocks, alt. 


M 


Volume 93, Number 3 
6 


Barker & Hicke 
Caribbean Adiantopsis (Pteridaceae) 


Laminae quadripinnatae; segme E ultima pusilla, 
2 / 


3.5 mm longa, 1.2-2.5 mm lata. Ab A. vincentii segmentis 


ultimis fertilis valde revolutis et textura spongiosa diffe 


Rhizome ascending to decumbent, to 2.5 em long, 
1.2-2 
dark brown center and golden margins, lanceate to 
4.50—6.00 mm 0.33— 
0.50 mm margins entire to 
slightly repand, 54.0— 
14.0 em long. Stipes atrocastaneous, always bicarinate 
15.5-28.3 cm long, 


carinae beginning at stipe 


2.3 em diam.; scales bicolorous with a black to 


acicular, basifixed. long, 


wide, bases truncate, 


apices acute. Fronds strict, 
adaxially, shorter than lamina, 
1.6—2.3 mm basal diam.; 
base or in the basal 4 of stipe and continuing into 
rachis; scales extending 1.3—3.4 em up the stipe, 


divergent, concolorous, tan, lanceolate to acicular 


(subulate), pseudopeltate to basifixed, 3.40—4.63 mm 


long. 0.30-0.48 mm | wide, bases biauriculate to 
truncale, margins entire slightly repand, apices 


Laminae quadri- 
dull 


26.3—45.7 cm long, 6.5-28.0 cm wide, apices difform, 


acute; hairs rare, septate-capitate. 
pinnate, triangular to lanceate, drying green, 


cuneate to acute; laminar tissue spongiose, hy- 


dathodes marginal; hairs with basal cell elongate, 
clear or white, middle cell short, tan or clear, apical 
cell bulbous. clear, white, or golden; stoma compla- 
nate, guard cells 32.60-58.68 um long. Laminar axes 
persistent, atrocastaneous, grading into lamina and 
taking its color and texture; rachis 1.7—2.3 mm basal 
diam.; carinae grading into lamina; scales 
0.65- 


bases truncate, 


golden, 
patent, narrowly triangular, 
1.10 mm long, 0.02-0.10 mm wide, 
apices acute. Pinnae ascending, 26—45 pairs, alter- 


appressed lo 


nate, frequently incurved along basal half of lamina, 
straight distally, triangular to narrowly triangular, 
inequilateral, to 8.5-16.5 em long, 3.3-12.5 cm wide, 
apices acute to cuneate. Pinnules patent to slightly 
asce 2e narrowly triangular, to Dal: 5 cm long, 


0.5— 
"iie acroscopic pinnules, apices attenuate to cuneate. 


5 em wide, basiscopic pinnules slightly larger 


Ultimate divisions slightly ascending, 


articulate, ovate, elliptic to oblong, 1.9-3.5 mm long, 


patent to 


1.2-2.5 mm wide, bases cuneate, frequently uniaur- 
iculate acroscopically, margins entire, apices round to 
acute; stalks persistent, not swollen at junction with 


laminar tissue, 0.2-0.5 mm long. Veins pseudodicho- 


tomous, anadromous, immersed in lamina, occult. 
Pseudoindusia distinct, lunate to quadrangular, green 
to white proximally, hyaline distally, occasionally 
black maculate, 0.50-0.72 mm long, 0.22-0.48 mm 
wide, margin erose to slightly erose. Sporangia 


subglobose, long stalked, arcus of (12-)14-25 in- 
durated cells, these 32.60-48.90 um tall. Spores 64 
per sporangium, tetrahedral-globose, golden at matu- 
i yellow when immature, 


rity, white to 


47.27 um diam., 


pale 
echinate, echinae bases complete, 


34.23- 


laesura obscured by ornamentation. Chromosome 


number unknown. 


Distribution and habitat. A rare endemic in the 
Escambray mountains of Cienfuegos, Sancti Spiritus, 
and Villa Clara Provinces, Cuba. Adiantopsis parvi- 
segmenta has been collected on rocks, most likely 


amphibolite at approximately 800 m altitude. Like 


other Adiantopsis taxa, A. parvisegmenta is known 
from very few collections and from one small 
geographic region. This species appears to be 
confined to the amphibolitie substrates of the 
Escambray Mountains, as it has not been found 


outside of the region. The Trinidad and Valle de los 
Ingenios National Park and World Heritage site occur 
in the range of A. parvisegmenta and should afford it 
some protection. 

Notes. 
menta is distinguished from most other pinnate taxa 


As its name implies, Adiantopsis parviseg- 


by its small ultimate divisions. However, it may be 
ultimate 
Three 


> [wo 


because the 


— 
S 
— 
^ 
ay 
= 
=~ 
EN 
= 


confused with 
divisions of the two taxa overlap in size. 
characters can be used to reliably distinguish the 
species. Adiantopsis parvisegmenta has a spongiose 
amina in contrast to the papyraceous lamina of A. 


The fertile ultimate divisions of A. parvi- 


vincenti. 
segmenta are strongly revolute, whereas those of A. 
vincentil are more or less flat. Finally, A. parviseg- 


menta is quadripinnate with a generally broader 


lamina than A. vincentii, which is tripinnate and 


generally narrower. Adiantopsis parvisegmenta is most 
likely a diploid species, based on indurated arcus cell, 
guard cell, and spore sizes, and may contribute to the 
parentage of some elements of the putative allotetra- 
ploid A. pentagona. 

It should be noted that specimens of Adiantopsis 
parvisegmenta were frequently determined as A. reesit. 
In folders at US, Proctor noted that these collections 
were not A. reesii, and he was correct. Thus, Cuban 
material determined as A. reesii may be A. parviseg- 
vincentii, as these species 


menta, A. rupicola, or A 


were frequently included under A. reesii by past 
researchers. 
Paratypes. Las Villas: Las Laguanas, Buenos 


ae we pe (NY), 6877 (US), 7033 (US [2]) 


Aires, 


3. Adiantopsis paupercula (Kunze) Fée, Gen. Filic. 
|Mém. Foug. 5]: 145. 1852. Adiantum paupercu- 
lum Kunze, Farrnkráuter 2(13): 65. t. 127. 1850. 
Fil. Hort. 


Cheilanthes paupercula (Kunze) Mett., 


Bot. Lips. 52. 1856. Hypolepis paupercula 
(Kunze) Hook., Sp. Fil. 2: 73. t. 88-C. 1856. 
TYPE: Cuba. Oriente. Mt. Líbano, 1844, J. 


Linden 1864 (holotype, LZ not seen; isotypes, 


390 


Annals of the 
Missouri Botanical Garden 


> 


a 5 e. Maxon 
= ( ic 


Figure 


pape re VM 5 uU )5 em. 


. pedata with tecla d apices. Se Md jn r = 0.75 em. 


BL BM!, 
Figure 7A, 


BM photo US!, K!, L!, P!, US). 


os 


5.9 em long, 0.3— 
with a black to dark 
golden margins, lanceate to 
3.61-13.93 mm 0.30— 


0.87 mm wide, bases truncate, margins entire, apices 


Rhizome decumbent to erect, to 
L.l em diam.: scales bicolorous 


center and 


acicular, — basifixed, long, 


attenuate, Fronds erect, 12.5-65.5 em long. Stipes 


atropurpureous to atrocastaneous, always bicarinate 


adaxially. longer than or equal (shorter) to lamina, 
8.5—35.5 em long. 0.4-2.3 mm basal diam.: carinae 
beginning near midpoint of stipe and continuing into 


rachis; scales extending 1.4-6.4 em up the stipe, 


divergent, concolorous, tan, lanceate. pseudopeltate lo 


basifixed, 3.2-9.5 mm long, 1.2—5.2 mm wide, bases 


biauriculate to truncate, margins entire, apices 


attenuate: hairs septate-capitate. Laminae tripinnate, 
bipinnate, to quadripinnate, triangular, drying green 
to brown, 9.5-30.0 cm long, 4.5-19.0 em wide, apices 
attenuate; laminar tissue 


conform, spongiose, hy- 


dathodes marginal; hairs with basal cell elongate. 


clear or white, middle cell short, clear. red, or golden. 


1239, US), Seale ba 
intopsis pedata (Wm. Harris 1087 6. G 110 Se sale ie = 


Ultimate divisions 


—D. Ultimate divisions 


apical cell bulbous. clear, red, or golden: stoma 


complanate, guard cells 48.90-08.46 Um long. Lam- 
inar axes persistent, atropurpureous to atrocastaneous, 
not grading into lamina at apices: rachis 0.5-1.4 mm 
basal diam.; carinae golden, not grading into lamina: 
21-33 


pairs, alternate, occasionally incurved in basal half of 


scales absent. Pinnae ascending to patent, 


amina, straight distally, linear to narrowly triangular, 
» 3.1-12 0.9—7.4 em wide, 
apices acute to cuneate. Pinnules slightly ascending to 
0.5- 


1.0 em wide. basiscopic pinnules slightly larger than 


inequil A em long. 


patent, linear. basal pinnules 1.1-6.0 


em long. 


acroscopic pinnules, apices cuneate to acute. Ultimate 


divisions slightly ascending to patent, strongly artic- 


ulate, flabellate; rhombiform. to oblong, 3.2-8.0 mm 
long, 2.5-4.5 mm wide, bases acute tọ truncate, 


stalls 


swollen at. junction. with laminar tissue, 2.4— 


margins entire, apices round to acute: persis- 


lent. 


18 mm long. Veins flabellate, immersed lamina, 


occult. to obscure. Pseudoindusia distinct. lunate, 
hyaline. occasionally black maculate. 0.8-0.9 mm 
long. 0.25-0.60 mm wide, margin erose. Sporangia 


subelobose. long stalked. arcus of 14-20 indurated 


Volume 93, Number 3 
2006 


Barker & Hickey 
Caribbean Adiantopsis (Pteridaceae) 


cells, these 35.86-47.27 um tall. 


sporangium, tetrahedral-globose, golden at maturity, 


white to pale yellow when immature, 35.86-45.64 um 


diam., echinate, echinae compact with complete 
bases, laesura obscured by ornamentation. Chromo- 


some number unknown. 


Distribution and habitat. 
Cuba, 


limestone or serpentine slopes and cliffs between 


Jamaica, and Puerto Rico on shaded, rocky 
400-1000 m altitude. Adiantopsis paupercula is the 
most widespread endemic Caribbean Adiantopsis. This 
is probably because the species can occur on both 
limestone and serpentine substrates, whereas other 
endemic Caribbean Adiantopsis appear to have greater 
substrate specificity. It should be noted that on two 
recent collecting trips to Puerto Rico A. paupercula 
was not found, and it may no longer be on the island, 
because it was previously known from only one 
collection (Proctor, 1989). However, it probably still 
occurs on Cuba and Jamaica. In Cuba, some of its 
habitat is protected by the Alejandro de Humboldt 
National Park and World Heritage site. 

Notes. Adiantopsis paupercula is one of the more 
. The strongly 


late, flabellate to rhombiform em dissi with 


distinctive Adiantopsis species articu- 


M 


a swelling at the junction. of the stalk and laminar 
tissue distinguish this species from other Adiantopsis. 
It is likely a diploid species, because its indurated 
arcus cell and spore sizes place it within a diploid 
group of taxa. However, its guard cell sizes are larger 
than other putative diploids, and it is not clear why. 


Selected specimens examined. CUBA. Guantanamo: Top 


of El Yunque, limestone mtn. near Baracoa, E. L. Ekman 
3977 (US): Monte Libanon, San Fernandez, E. L. Ekman 
10308 (B, US). Holguin: Mina Cayoguan, Bro. Clement 5982 
(US): Vedado-Habana, 10 of Cayoguan River N of Sierra 


de Moa. Fres Lein 20910, M. Victorin, icd (US): Sierra 
de Nipe. he adwaters i E ilots River, E. L. Ekman 2515 (B. 

BM, US): Camp La Gloria, 5 of Sierra vm J. A. Shafer 8174 
(K, US); Sierra Nipe, near Woodfred, Rocky Arroyo 1 


~ 

— — 

— 5 

m. UN 

mr 

b x 

- 

la] 

p 

t 
vw 
8 
— 
— 
^ 
A 
> 
"S 
© 
= 
WS 
Y 


d. Linden 1862 (K): 
caverns P The mola $ vic., W. R. Maxon 4259 (P, l 
20: mte Verde, C. Wright 962 (UC): Verde, € 
Uriah 964 (BM [2], x [3]. I ). US 3). 
CA 


Clarendon: 


I. 19 1 
P. US 


£z. 


Monte 


ortland: E slope of Johi 
elesdown, G. R. Pr octor - 5729 
i of Macon ie Savanna, . 

BM) ^ On limestone cliffs, A M. 
Underwood Pu (US, P); Six mi. WNW of Troy on Crown 
4. Hes 1191, E. B. i ar JE: 

„ R. E. 17 ( , UC, US): Crown Lands Road 
ean ca. 5 mi. NW e Ls G. R. on 31099 (DUKE): 


) 
Doug: 


- 
A 
- 
arm D 


Spores 64 per 


Found on the islands of 


4 mi. WNW of Troy on Crown Lands px H. A. a 
Pr Ee On limestone rocks, 
(US); Near Troy, W. Harris 8791 D 
PUERTO RICO. Maricao: Maricao State Forest, Barrie 
Tane headwaters stream of Rio Seco, below Road 120. 


km. 10.15, G. R. Proctor 41649 (US 


. Underwood 330: 


Adiantopsis pedata (Hook.) T. Moore, Ind. Fil. 

18. 1857. Hypolepis pedata Hook., Sp. Fil. 2: 73, 

t. 92-A. 1852. Cheilanthes pedata (Hook.) A. 

Braun, Index Seminum. 1857. TYPE: Jamaica. 

Purdie s.n. (holotype, K!, K photo B!, K photo 

MICH!; isotypes, BM!, BM photo B!, BM photo 
MICH!, P!). Figure 7C, D. 


0.6— 


scales bicolorous with a black to dark 


Rhizome erect to decumbent, to 3.0 em long, 
1.4 em diam.; 
brown center and golden margins, lanceate, basifixed, 
3.50—4. 0.43-0.45 mm 


truncate, margins entire, apices acute. Fronds strict, 


75mm long, wide, bases 


21.7-32.3 cm long. Stipes atropurpureous to atrocas- 


always bicarinate adaxially, longer than or 
11.3-19.0 cm 0.8-1.3 mm 
basal diam.; carinae beginning at stipe base or in the 


75 of stipe and continuing into rachis; scales 


aneous, 


amina, long, 


equal to 
basal 
extending 2.3-3.4 cm up the stipe. appressed, con- 
3.50— 


bases biauricu- 


colorous, tan, lanceolate, — pseudopeltate, 
0.53—0.58 mm wide. 


repand, apices acute; 


4.00 mm long, 


ate, margins entire to slightly 
hairs rare, septale-capitate. Laminae pedate, tripin- 
nate to rarely quadripinnate basally, abruptly bi- 
drying 

8.5- 


difform, 


broadly deltate-pentagonal, 
green, dark 
9.4-25.7 cm 


tissue 


pinnate above, 


o brown, 


vellow-green, 
19.4 cm 


altenuate: 


green 


long. wide. apices 


^ 
— 


laminar vartaceous, hydathodes 


marginal; hairs with basal cell elongate and clear, 


middle cell frequently short and red, golden red, 
eolden, or white, apical cell bulbous and clear, white, 
cells 53.79— 


18.24 um long. Laminar axes persistent, atropurpur- 


or golden; stoma complanate, guard 
eous to atrocastaneous, grading into lamina and taking 
its color and texture; rachis 0.4—1.0 mm basal diam.: 
carinae golden, grading into lamina: scales appressed 


0.45-0.88 mm long, 0.02-0.07 mm 


wide, bases truncate, apices acute. Pinnae ascending 


to patent, linear, 


to patent, 21-28 pairs, basal pinnae opposite to sub- 


opposite, straight, triangular, inequilateral, to 4.7— 
12.9 cm long, 1.7-6.0 cm wide, apices attenuate to 


acute: penultimale pinnae opposile to sub-opposite, 


narrowly fusiform linear, abruptly shorter, 2.5— 


7.1 cm long; distal pinnae alternate, quickly reduced 
in length. Pinnules patent to slightly ascending, 
linear, apices attenuate to cuneate; basal basiscopic 
narrowly triangular Lo 


1.0-5.6 em long, 2.2-5.1 


pinnule, basal 


pinnules of basal pinnae 


fusiform, much extended, 


times longer than basal acroscopic 


392 


Annals of the 
Missouri Botanical Garden 


acroscopie pinnule 0.5-1.3 em long, distal pinnules of 


first pinnae equivalent, abruptly shorter than basal 


basiscopie pinnule. Ultimate divisions ascending, 
articulate, oblanceolate, narrowly elliptic, to narrowly 
4.0—9.0 mm 


cuneate, uniauriculate acroscopically, margins entire 


oblong, long, 1.4—3.5 mm wide, bases 
proximally, crenulate distally, apices round to acute; 
stalks persistent, not swollen at junction with laminar 


tissue, 0.10-0.60 mm long. Veins pseudodichotomous, 


anadromous, immersed in lamina, occult to obscure. 
Pseudoindusia distinct, lunate to quadrangular, green 
to. white proximally, hyaline distally, occasionally 
0.60-0.84 mm long, 0.30-0.56 mm 


margin slightly Sporangia subglobose, 
long stalked, 


black maculate, 


wide, erose. 


arcus of 13—15(—20) indurated cells, 
these 40.75-55.42 um tall. Spores 64 per sporangium., 


tetrahedral-globose, golden at maturity, white to pale 


yellow when immature, 38.31-54.01 um diam., 
densely echinate, echinae bases dissected, laesura 
obscured by ornamentation. Chromosome | number 


unknown. 


Distribution and habitat. Adiantopssis pedata is 
a rare fern known only from the Dominican Republic 
and Jamaica. The species has been collected in humus 
on wooded hillsides scattered. with limestone rocks al 
800—1000 m altitude. 


have been made, it occurs in very rural areas that have 


Although no recent collections 


relatively little economic development, such as the 
Cockpit Country of Jamaica. Given that ecotourism is 
becoming an industry in this part of Jamaica. the 


habitat of A. pedata is likely to be preserved, and the 


species may still be found there if collections are 
sought. 
Notes. Adiantopsis pedata formerly included 


plants from Cuba. However, the Cuban material 


recognized here as A. pentagona is clearly distinct. 
Carinae beginning in the basal half of the stipes, 
generally ascending ultimate divisions with crenulate 
apices, and a mode of 14 indurated arcus cells 


distinguish A. pedata. Adiantopsis pentagona has 


carinae beginning in the apical half of the stipes. 
patent or slightly ascending ultimate divisions, entire 
or only slightly crenulate ultimate division apices, and 
Both 
be confused with the South 


a mode of 16 indurated arcus cells. Caribbean 


taxa may American 4. 
Xaustralopedata, which is a sterile hybrid that occurs 


Brazil, 


Paraguay. The aborted spores, crenulate margins, and 


in the tri-border region of Argentina, and 


generally much larger frond size distinguish A. 


australopedata from the 
al., 2003). 


„ pedata appears to be a tetraploid 


Caribbean taxa (Hickey et 


derivative of A. radiata and A. reesit. Indurated arcus 


cell, guard ig and spore sizes strongly suggest that il 


is tetraploid. Its pedate morphology suggests that it is 
a hybrid taxon, with A. radiata as one parent (Hickey 


et al., 2003). Also shared with A. radiata are spore 


echinae that are basally dissected, and a common 
pattern. of laminar development. With A. esl, A. 
pedata shares distinctly toothed ultimate division 
apices, a mode of 14 indurated arcus cells, and 


carinae beginning in the lower portions of the stipe. 


Based on all these data, A. pedata is supported 


a tetraploid derivative of A. radiata and reesil. 
Future cytological research needs to be conducted to 
confirm its ploidy level, and future molecular research 


should test the parentage of A. pedata. 


DOMINIC | RE PUBLIC. 


dus lected vide ue examined. 


onte Cristi: Cordillera Central. Monción. E. L. Ekman 
17 0 (B. BM t K, US [2]). 

AMAICA. Clarendon: Peckham Woodland. Harris 
10. 878 (B, GH, K, US [2]: Glenwood n along 


road between Balcarres & St mbr Iry, G. è Proctor 


R. 
35655 (US). Unknown Parish: Jamaica. Walker-Arnott s.n. 
N 


Y). 


5. Adiantopsis pentagona M. 5. Barker & Hickey, 
sp. nov. TYPE: Cuba. ae Clara, vic. of 
Sopapo. Buenos Aires, Trinidad Mtns., shaded 
limestone cliff, 4 Aug. 1936, L. B. Smith 3320, A. 
R. Hodgdon, & F. Gonzalez (holotype, GH! 
isolypes, NY!, US!). Figure 8A, B. 

Laminae pedataez pinnae supernae bipinnatae; pinnae 
basales tripinnatae, pinnulis valde basalibus elongatis 
magnopere. Ab A. pedata carinis in dimidio superno stipitis 


x 
3 


s, segmentis ultimis patentibus vel leviter ascen- 
dentibus prae ‘bentibus apices integros vel leviter erenulatos, 
arcu e cellulis induratis 16 (modus) composito, et ab 
Xaustralopedata sporis fertilibus (nonabortivis) differt. 
Rhizome erect to decumbent, to 3.8 em long, 0.5- 
1.8 em diam.; scales bicolorous with a black to dark 


brown and 


center golden margins, acicular to 
lanceate, basifixed, 2.13-7.58 mm long, 0.08- 
0.88 mm wide, bases truncate, margins repand 


entire, apices acute. Fronds strict, 33.2-72.1 cm long. 
Stipes atropurpureous to atrocastaneous, always bicar- 
inate adaxially, longer than or equal to lamina, 17.1— 
44.0 cm 0.8-2.1 mm 


beginning in the upper / (2) of stipe and continuing 


long. basal diam: carinae 


into rachis: scales extending 2.0—3.7 em up the stipe, 


appressed, concolorous, tan, acicular to lanceate 


(subulate). pseudopeltate, 1.05-6.38 mm long, 0.13— 


0.83 mm wide, bases biauriculate, margins entire to 


slightly repand, apices acute; hairs rare, septate— 
capitate. Laminae pedate, tripinnate to quadripinnate 
basally, abruptly bipinnate above, broadly deltate- 
pentagonal, drying brown, dark green, green to yellow- 
green, 14.7—28.1 em long, 12.8-48.2 em wide, apices 


difform, long attenuate: laminar tissue thin-spongiose, 


Volume 93, Number 3 
0 


Barker & Hick 
Caribbean piis (Pteridaceae) 


393 


Figure 8. ROOM du 6 i sd 3320 et al., US). Scale bar = 3 em. —B. Ultimate divisions of A. 
pentagona ice oded apices. Scale bar = E. anodes ne (C. R. 1 29256, F). Scale bar = 3 cm. 
I). Ultimate divisions of A. ag ean Scale Pu = 07 18 0 
hydathodes marginal: hairs with basal cell elongate divisions patent to slightly ascending. articulate, 

J o o o J [o] 


and clear, middle cell frequently short and white, 


golden, golden red, or red, apical cell bulbous and 
clear, white, or golden; stoma complanate, guard cells 


48.90-83.13 um long. Laminar axes persistent, atro- 


— 


purpureous to atrocastaneous, grading into lamina anc 
taking its color and texture; rachis 0.5-1.3 mm basal 


diam.; carinae golden, grading into lamina; scales 


appressed to patent, lanceate to linear, 0.77-1.45 mm 
0.08-0.17 mm 
attenuate. Pinnae patent to ascending, 28—40 pairs, 


long, wide, bases truncate, apices 


pinnae to sub-opposite, straight, 
inequilateral, to 6.4-24.1 cm long, 2.8- 
19.9 em wide, apices attenuate to acute; penultimate 
].4— 


14.2 cm long; distal pinnae alternate, quickly reduced 


basal opposite 


triangular, 
pinnae sub-opposite, linear, abruptly shorter, 
in length. Pinnules patent to slightly ascending, linear 


to narrowly lanceate, apices cuneate to attenuate; 
basal basiscopic pinnules of basal pinnae fusiform to 


1.8—15.3 cm 


1.8-13.1 times longer than basal acroscopic 


narrowly triangular, much extended, 


long, 
pinnule, basal acroscopic pinnule 0.4-5.8 em long, 
distal pinnules of first pinnae equivalent, abruptly 


shorter than basal basiscopic pinnule. Ultimate 


narrowly oblong, oblanceolate, lanceolate, to narrowly 
6.1-11.1 mm long, 


cuneate, frequently uniauriculate acroscopically, mar- 


elliptic, 2.5—4.2 mm wide, bases 
gins entire to slightly crenulate, apices round to acute; 
stalks persistent, not swollen at junction with laminar 
tissue, 0.9-6.3 mm long. Veins pseudodichotomous, 
anadromous, immersed in lamina, occult to obscure. 
Pseudoindusia distinct, lunate to quadrangular, hya- 
line with a brown or black area proximally where 
hydathode terminates, occasionally black maculate, 
0.35-1.10 mm 0.20-0.60 mm 


entire to slightly erose. Sporangia subglobose, 


long, wide, margin 
long 
stalked, arcus of 14-22 indurated cells, these 40.75- 
58.08 Um tall. 
globose, golden at maturity, white to pale yellow when 
immature, 40.75-58.90 um diam., 
laesura obscured by orna- 


Spores 04. per sporangium, tetrahedral- 


densely echinate, 
echinae bases dissected, 


mentation. Chromosome number unknown. 


Distribution and habitat. A Cuban endemic found 
in the province of Pinar del Río, and the Escambray 
Mountains of Cienfuegos, Sancti Spiritus, and Villa 


Clara. Adiantopsis pentagona has been collected on 


[om 


moist, shaded limestone or amphibolite soils, rocks, 


394 Annals of the 
Missouri Botanical Garden 
cliffs, and cave entrances. When A. pentagona has Buenos Aires, J. G. Jack 7903 (GH [2], NY, US): On the crest 


been collected in soil, it was generally found on slopes 
scattered with limestone or amphibolite rocks. 
Adiantopsis pentagona is most likely still found 
Cuba. In Pinar del Rio, it is probably protected within 
the boundaries of Valle de Vinales National Park, and 
by the Trinidad and Valle de los Ingenios National 
Park in the 
Notes. 
overall pentagonal outline of the lamina. 


Escambray Mountains. 
the 
Adiantopsis 


Adiantopsis pentagona is named for 


pedata and A. Xaustralopedata also have pentagonal 


laminar outlines and may be confused with A. 
pentagona. Carinae beginning in the basal half of 


the stipes, distinctly crenulate ultimate division 


apices, generally ascending ultimate divisions, and 


mode of 14 indurated arcus cells, distinguish A. 
pedata from A. pentagona, which has carinae 


beginning in the apical half of the stipe, entire or 
slightly crenulate ultimate division apices, palent to 
slightly ascending ultimate divisions, and a mode of 


16 indurated arcus. cells. Adiantopsis Xaustralope- 
data, a hybrid distributed in the tri-border region of 
Argentina, Brazil, and Paraguay, may be distinguished 
from the Caribbean species by its aborted Spores, 
crenulate margins, and generally larger laminae. 
Idiantopsis pentagona appears to be the allotetra- 
The 


large spores, guard cells, and indurated arcus cells 


ploid derivative of A. radiata and rupicola. 


relative to. putative diploid taxa, such as A. radiata 
and A. that A. 


tetraploid. Its pedate morphology suggests that it 


rupicola, suggest pentagona is 


hybrid taxon, with A. radiata as one parent 
(Hickey et al., 2003). 
echinae that 


a common pattern of laminar development. With A. 


Is a 
Also shared with A. radiata are 
spore are basally dissected, and 
rupicola, A, pentagona shares carinae beginning in the 


apical portion of stipes, a mode of 16 indurated arcus 


cells, and the overall ultimate division shape agrees 
most favorably with A. rupicola. The type of A. 
)entagona, Smith 3320 et al.. was selected in part 
/ ¿ | 


because its morphology supports an allotetraploid 
origin involving A. radiata and A. rupicola. However, 
pentagona, specifically those 
that 


parvisegmenta 


some collections of A. 


with relatively small ultimate divisions are 


revolute when fertile suggest that A. 


may be the parent of a second, cryptic species. This 


putative eryplie species has been retained in A. 
pentagona because the available data do not clearly 
differentiate it. Future molecular research should 


encompass the morphological range of A. pentagona 


and include A. rupicola, A. parvisegmenta, and A. 


vincentit, to evaluate if cryptic elements are included, 


a BA. Las Villas: Las Vegas de s agua, 


Paratypes. 


Trinidad Mti V. Morton 10547 (DUKE. US), C. 
Morton TET a DUKE. US [2]); Las Vegas de 111 


Vain 6715 (MICH. 
Morton 4179 (GM. 


Trinidad Mins.. 1 5 


Prinidad Mtns., C. V. 


of Pico Petrerillo. 
US): Buenos Aires, ! 


UC, US): Sierra del Bibisial, Loma de Ponciano. Sancti 
Spiritus Mtns., F. Leon 6712, F. Clement (NY); San Blas- 
Buenos Aires. Trinidad Mtns.. gulley behind Gavinas, L. A 


Howard 6505 (GH. MO): Trinidad Mtns.. Glen . 
(MICH): Deep limestone cave near San Jose, R. A. 
5154 (GH, MO, NY [2]: Buenos Aires, Roig 6124, 
(NY) Hanabanilla Falls 
1 85 : Lane 

A: Ho T 5387 (GH, MO, NY): 


NY): San Blas-Buenos Aires, Tanided 


Howard 146, W. Briggs. 
> aranjal, Trinidad 
„J. A. 


zo 
E 
E 
D 
=< 
e. t 
q 
Jr 
ise) 
S: 


Pitajones 


Mins., uide near Miss Carlota, A. Gonzales 513 (BM, 
GH, US) San Blas-Buenos Aires; Trinidad Mtns.. Gavinas, 
R. A. lbi 6539 (GH): Las Lagunas. Buenos Aires, J. C. 
Jack 6869 (US): Buenos Aires, Trinidad. Hills. J. G. Jack 
7773 E Pinar del Hío Sierra de las Animas, F. L. 
1 10510 (NY, US). Sancti Spiritus: Mogote Caburni. 
P. Acevedo-Rdgz.. 6101 n" Oviedo. M. Fernandez. N. Vera 
(Us). Unknown Province: C. Wright 3947 0 . Wright 
3949 (GH, K, NY |2]: 1867, e. l. bos. Desd. : FI) . 
Sow t deed (L.) Fée, Gen. Filic. [Mém. 


Foug. 5|: 145. 1852. Adiantum radiatum l., Sp. 
PI. 2: 1094. 1753. Cheilanthes radiata (I.) John. 


1841. Hypolepis 
72. 1852. TYPE: 
LINN 
York 


Smith. J. Bot. (Hooker) 4: 159, 
radiata . Hook., Sp. Fil. 2: 
Sloane "Adiant. G radiatum” 
1252.1, chosen by Lellinger, 
Bot. Gard. 23: 3. 1972 


(lectoty pe. 
Mem. New 
). Figure 8C, D. 

0.5-0.9 em diam. 


black to dark brown center 


Rhizome erect, to 3.3 em long, 
scales bicolorous with a 
and golden margins, lanceolate to acicular, pseudo- 


0.41— 


margins 


peltate to basifixed, 754.43 mm long, 


0.58 mm wide. bases cordate to truncate, 


repand, apices acute to acuminate. Fronds strict, 


9.02.3 cm long. Stipes atropurpureous lo atrocas- 
taneous, rarely bicarinate adaxially, longer than 
lamina, 10.5—47.0 cm long, 0.8-2.0 mm basal diam. 


carinae when present, beginning al stipe base and 


continuing into lamina; scales extending 0.2-3.5 em 
up the stipe. appressed, bicolorous with broad 


margins, or 
3.00— 


bases biauricu- 


castaneous to tan center and golden 


conconlorous. tan, lanceolate. pseudopeltate, 


5.50 mm long. 0.45-0.75 mm wide. 
late, 


septale-capitate. Laminae palmate, 


margins repand, apices acute: hairs rare, 
radially bipinnate, 
orbiculate, geniculate, drying olive-green, dark green, 
10.2-26.5 em 4.4— 


apices difform, acute; laminar tissue 


long, 


black, to yellow green, 
25.0 em wide, 
thin-spongiose, hydathodes marginal; hairs with basal 
cell elongate and clear, middle cell frequently short 
and eolden or red, apical cell bulbous and clear or 


white: stoma complanate, guard cells 35.86-03.57 um 


nd Laminar axes persistent, atropurpureous lo 


alrocastaneous, grading into lamina and taking its 


color and texture; carinae golden, grading into lamina: 


Volume 93, Number 3 
2006 


Barker & Hicke 


y 395 
Caribbean Adiantopsis (Pteridaceae) 


scales absent. Pinnae spreading radially from stipe 
apex, 3—9, straight, linear to narrowly fusiform, apices 
acule; central pinna to 9.2-15.3 em long, 1.4-2.7 cm 
to 2.6-9 1.0-2.2 em 


wide; basal flabellate de often fertile, attached 


wide; basal pinna 9.2 cm long, 


to the stipe apex between pinnae. Ultimate divisions 


patent to ascending, articulate, oblong to narrowly 


oblong, 6.0-13.5 mm long, 2.1-3.7 mm wide, bases 
cuneate to acute, margins entire lo crenulate, apices 


stalks swollen al 


junction with laminar tissue, 1.0-6.0 mm long. Veins 


round to acute; persistent, not 


— 


pseudodichotomous, anadromous, immersed ami- 


na, obscure to occult. Pseudoindusia distinct, lunate 
to quadrangular, hyaline, occasionally black macu- 


late, 0.4—0.84 mm long. 0.22-0.56 mm wide, 


entire to slightly erose. Sporangia subglobose. long 


margin 


stalked, arcus of 16-27 indurated cells, these 30.97— 
46.46 Um tall. Spores 64 per sporangium, tetrahedral- 
globose, golden at maturity, white to pale yellow when 
immature, 34.23—47.27 um diam., densely echinate. 
echinae bases dissected, laesura obscured by orna- 


mentation. Chromosome number n = 30. 


Distribution and habitat. Adiantopsis radiata is by 


far the most common and widely distributed Adian- 
topsis species. The species is widespread throughout 
tropical America and occurs from the Caribbean and 
It has most often been 


Mexico south to Argentina. 


collected on moist, calcareous or serpentine, rocky, 


wooded slopes and has also been collected on 
limestone cliffs and walls. 
Notes. 


Adiantopsis species, as it is the only palmate species. 


Adiantopsis radiata is the most distinctive 


As the name implies, the pinnae radiate outward from 
a single point at the stipe apex. Young A. radiata may 
be confused with voung A. pedata and A. pentagona 


because all three taxa share a similar ternate juvenile 


form. The most reliable character to distinguish these 
taxa is indurated arcus cell number. Adiantopsis 


indurated arcus cells, 
14 and A. 


entagona has a mode of 16. Also, the stipes of A. 
/ & l 


of 20 


whereas A. pedata possesses a mode of 


radiata has a mode 


radiata are typically lacking carinae, whereas both A. 
pedata and A. pentagona stipes are always adaxially 
bicarinate. Another character that is useful for 
distinguishing A. radiata and A. pedata is ultimate 
division shape and orientation. The ultimate divisions 
of A. radiata are usually oblong with rounded apices 
and are generally patent. Adiantopsis pedata ultimate 


divisions are more frequently oblanceolate with 
a distinctly toothed apex and are ascending. 
Adiantopsis radiata is implicated in an interesting 
pattern of reticulate and morphological evolution. The 
species apparently hybridizes with various pinnate 


taxa to produce pedate derivatives. In the Caribbean, 


these hybrid derived pedate taxa are A. pedata and A. 
pentagona, and in South America A. Xaustralopedata 


(Hickey et al., 2003). 


JA. Oriente: Decidu- 
a, J. A. Shafer 3777 (F); 
MO [2]. 


Selected. specimens examined. CUB 
ous woods near base o 
Monte Verde, C. Wright 963 (V. 
1889, Eggers 4813 (F). 


3 
E 
"OX: 
C5 CD 
jan] 
— 
[e 


Unknown 
province: 


JAMAICA. Clarendon: Peckham Woods. . Proctor 
68145 (MO): Peckham Woodland. W. Harris ~ (F): 
Summit area of Crofts Min.. G. R. Proctor 29256 (F). 
Portland: 1906, A. Moore (MICH). Trelawny: Vic. of Troy, 


OQ 


W. R. Maxon 2954 (MICH). 1 Parish: 188 
390 (MO); W. Harris s 

| 5 ta: Cliffs in woodlands on 
River ucl W 


— 


6. J. P. 


— 
= 
= 


E 10 0 of Morne Brule, 
5t. Paul, Morne Cola 
5 (MIC H, 310) Martinique: 


). Montserrat Cudjor Lerad, J. A. S} 


the North 
Portsmouth, 
Anglais, G. L. Webster 1342: 
1868. Sidier s.n. (F 
153 (F). 


Blanchau s.n. 


vafer 
Guadeloupe: Falaises de la rivière des Peres, 


(©). Unknown Island: 1881, J. L. Main s.n. 


— 
= 


TRINIDAD. 1877-1878, A. Fendler 7 herb. chas. H. Ford 
1354] (F, MICH, MO): Blanchissense * 
Vale on Banks Main Road, 


Road r Verdant 


> 
"TO 
— 


A. Hambersley s.n. 


7. Adiantopsis reesii (Jenm.) C. Chr., Ind. Fil. 22. 
1905. Cheilanthes reesii Jenm., J. Bot. Brit. For. 
24: 267. 1886. TYPE: Jamaica. St. Elizabeth 
Parish, Oxford, Rev. J. L. Rees s.n. (holotype, Kl. 
K photo US BMI. BM photo US). 
Figure 9A, B. 


isotype, 


Rhizome decumbent to ascending, to 2.8 em long. 
0.6—0.8 em diam.; scales bicolorous with a black to 
lanceate 
0.52- 
margins slightly 
17.1-60.0 cm 


always bicarinate adaxi- 


dark brown center and golden margins, 


to acicular, basifixed, 4.05-5.50 mm long. 
0.6l mm wide, bases truncate, 
repand, apices acute. Fronds strict, 


Stipes atrocastaneous, 


long. 
ally, much shorter than lamina, 2.2-22.0 em long, 
1.0-2.6 mm basal diam.; carinae beginning at stipe 
base or in the basal 1⁄4 of stipe and continuing into 
rachis: scales extending 1.9—4.2 cm up the stipe or to 
concolorous, tan, lanceate to 


1.25—5.75 mm 


margins 


the rachis, divergent, 
acicular (subulate), pseudopeltate. 


long, 0.20-0.74 mm wide, 


entire to slightly repand, apices acute; hairs rare, 


bases cordate, 


lanceolate to 
5.0-53.0 em 


long, 5.5-14.6 cm wide, apices difform, attenuate to 


seplale-capilate. Laminae tripinnate, 


narrowly triangular, drying dull green, 
acuminate; laminar tissue papyraceous, hydathodes 
marginal; hairs with basal cell elongate and clear, 
middle cell frequently short and clear, brown, or red, 
apical cell bulbous and frequently covering middle 
cell, 


cells 32.00-53.79 um long. Laminar axes persistent, 


white, clear, or golden; stoma impressed, guard 


alrocastaneous, grading into lamina and taking its 


396 Annals of the 
Missouri Botanical Garden 


Figure 9, —4A. Adiantopsis reesit (F. E Ekman 9987, V). Scale bar 3 em. —B. Ultimate divisions of j reestt with 
toothed apices. 9 a bar = 0.25 em. —C. Adiantopsis rupicola (Britton 7497 et al., US). Scale bar = 3 em. —D. Ultimate 
divisions of A. rupicola. Scale bar = = cnt. 
color and texture; rachis 1.1-1.5 mm basal diam.: lden at maturity, white to pale yellow when 


carinae golden, grading into lamina; scales appressed immature, 30.97-44.83 um diam., echinate. echinae 


lo patent, narrowly rhombic to lanceolate, 0.70— bases complete, laesura obscured by ornamentation, 
1.60 mm long. 0.05-0.19 mm wide, bases acute, Chromosome number unknown. 

apices acute. Pinnae ascending, 23-55 pairs. alter- 

nate, slightly incurved in basal half of lamina, straight Distribution and habitat. Adiantopsis reesii occurs 
distally, narrowly triangular to linear, lightly inequi- on the islands of Hispaniola and Jamaica. It has been 


lateral to equilateral, to 5.7-8.2 em long. 1.8-3.7 em collected on shaded limestone cliffs and slopes, at up 
wide, apices cuneate to acuminate. Pinnules ascend- t0 900 m altitude. Like many other Adiantopsis 


ing to slightly ascending, narrowly triangular to species, A. reesit is known from only a few collections 


is 


narrowly lanceate, to 1.2-2.3 em long, 0.7-1.10 em and appears to be rare. However, the species 
wide, basiscopie pinnules slightly larger than acro- probably still extant, because it occurs in rural areas 


scopic pinnules, apices obtuse to acute. Ultimate with little economic development. Additionally, recent 


divisions ascending, articulate, narrowly oblong, ecotourism business in the Cockpit Country of 


oblanceolate. to nearly flabellate. 3.4-9.1 mm long, Jamaica may protect ils habitat from economic 


1.2-3.8 mm wide, bases cuneate, margins entire tọ development. 


crenulate proximally, crenulate distally, apices round, Notes. Adiantopsis reesii is distinctive in its 


acuminate to acute: stalks persistent. not swollen al toothed ultimate division apices that distinguish i 


junction with laminar tissue, 0.3-0.0 mm long. Veins from other pinnate Adiantopsis. Further, the ultimate 


pseudodichotomous, anadromous, impressed to im- divisions of A. reesii are rather elongate compared to 
mersed in lamina, prominent. Pseudoindusia distinct, other pinnate taxa such as A. rupicola, A. parviseg- 
lunate to quadrangular, green to white proximally, menta, and A, vincenti. 

hyaline distally, occasionally black maculate, 0.44— Adiantopsis reesii is a putative diploid, as evidenced 


0.92 mm long, 0.30-0.44 mm wide, margin erose to by the sizes of its guard cells, spores, and indurated 


entire. Sporangia subglobose, long stalked, arcus of arcus. cells. This species appears to be one diploid 
9—17(—23) indurated cells; these 21.19-48.90 um parent, along with A. radiata, of the putative tetraploid 


tall. Spores 64 per sporangium, tetrahedral-globose. A. pedata. The distinctly toothed apices and ascending 


Volume 93, Number 3 Barker & Hickey 397 
2006 Caribbean Adiantopsis (Pteridaceae) 

ultimate divisions of A. reesii are evident in A. pedata. scopie pinnules slightly larger than acroscopic 

Future molecular research should examine this — pinnules, apices acute. Ultimate divisions slightly 

hypothesized relationship. ascending to patent, articulate, oblong to trullate, 3.8— 

9.5 mm long, 2.0—5.5 mm wide, bases acute to 

Selected specimens HAITI. Grand’ Anse: cuneate, frequently uniauriculate acroscopically, mar- 


examined. 

Morne Rockhelois, Meragoaire, at 
Trou-Gorie, E. I. Ekman 9987 (BM, F [2], NY, UC, US): 
Massif de la e tte, W group, Tiburon, Morne Sentier, E. J 
Ekman 10608 (K, US); Massif de la Hotte, gr. M. Rockhelois, 
Meragoaire, Ekman 7965 (US). 


Massif de 1 Hotte, gr. Mor 


JAMAICA. Trelawny: Cockpit Country ca. 5 
Quick Step, above Aberdeen P.O., G. R. Proctor 4090 (U 3 
Unknown Parish: R. V. Sherring s.n. (BM). 


8. Adiantopsis rupicola Maxon, Contr. U.S. Natl. 


Herb. 10: 485. 1908. TYPE: Cuba. Pinar del Rio, 


collected in mtns. near El Guama, crevices of 


partially shaded cliffs, limestone, rare, W. 
Palmer 242 & J. H. Riley (holotype, US 


isotypes, KI, MO!, US). Figures 9C, D. 


Rhizome ascending to decumbent, to 2.2 cm long. 
0.6-1.2 cm diam.; scales bicolorous with a black to 
dark brown center and golden margins, acicular to 
1.60-3.60 mm long, 0.30-0.36 mm 
wide, bases truncate, margins slightly repand, apices 
23.5-50.2 cm 


atropurpureous to atrocastaneous, occasionally bicar- 


linear, basifixed, 


acute. Fronds strict, long. Stipes 


inate, shorter than lamina, 4.0-17.5 cm long, 0.6- 
1.7 mm basal diam.; carinae when present, beginning 


in upper % of the stipe and continuing into the rachis; 
scales extending 1.9-7.1 cm up the stipe, divergent, 
concolorous, tan, acicular to lanceate, pseudopeltate 
to basifixed, 1.00-6.64. mm long, 0.06-0.70 mm wide, 
bases biauriculate to truncate, margins slightly repand 
to entire, apices attenuate; hairs septate-capitate. 
Laminae tripinnate, narrowly triangular to lanceate, 


drying dull green to brown, 11.0-39.5 cm long, 4.2— 


17.5 cm wide, apices difform, attenuate; laminar 
tissue thin-spongiose to chartaceous, hydathodes 
marginal; hairs with basal cell elongate, clear or 


white, middle cell short, tan or clear, apical cell 


bulbous, clear, white, or golden; stoma complanate, 


guard cells 42.38-57.05 um long. 


marcescent with costae curling acroscopically, atro- 


Laminar axes 
purpureous to atrocastaneous, grading into lamina and 
taking its color and texture; rachis 0.6-1.5 mm basal 
scales 
0.67— 
bases acute, 
21-24 pairs, 


alternate, frequently ineurved in basal half of lamina, 


diam.; carinae golden, grading into lamina; 


patent, narrowly lanceolate, 


0.08-0.16 mm 


Pinnae 


appressed to 
1.12 mm 


attenuate. 


long, wide, 


apices ascending, 
straight distally, narrowly triangular, slightly inequi- 
lateral to equilateral, to 3.2-16.0 em long, 0.9-5.5 cm 
wide, apices acute to attenuate. Pinnules patent to 
slightly ascending, narrowly triangular to linear, basal 
0.5-1.5 cm wide, 


pinnules 0.7-3.9 cm long, basi- 


gins entire, apices acute to round; stalks persistent, 
0.1— 


0.5 mm long. Veins pseudodichotomous, anadromous, 


not swollen at junction with laminar tissue, 
immersed in lamina, occult to obscure. Pseudoindusia 


distinct, lunate to quadrangular, green to white 
proximally, hyaline distally, occasionally black mac- 
0.42-0.70 mm long, 0.30-0.40 mm 
margin erose to slightly erose. Sporangia subglobose, 
long stalked, arcus of 12-21 indurated cells, 
32.60-42.38 um tall. 
tetrahedral-globose, golden at maturity, white to pale 


340.75 Um diam., 


ulate, wide, 
these 
Spores O4 per sporangium, 
yellow when immature, 34.2 echi- 
nale, echinae bases complete, laesura obscured by 


ornamentation. Chromosome number unknown. 


Distribution and habitat. Adiantopsis rupicola is 
endemic to the province of Pinar del Río, Cuba. The 
species has been collected on shaded, rocky limestone 
slopes or cliffs. Similar to other Adiantopsis taxa, A. 
rupicola is known from few collections. Although there 
are no recent collections of A. rupicola, it probably 
still occurs in Pinar del Río and is likely protected 
within the boundaries of Valle de Viñales National 
Park. 
Notes. 
A. vincenti. 


Adiantopsis rupicola may be confused with 
The carinae of A. rupicola begin in the 
apical quarter of the stipes, whereas the carinae of A. 
Also, A. 


rupicola has a mode of 16 indurated arcus cells, 


vincentii begin in the basal half of the stipes. 


whereas A. vincentii has a mode of 14. The generally 
obovate to oblong ultimate divisions and 


of A. 


distinguish it from A. rupicola, which has generally 


smaller, 


lanceate to lanceolate lamina vincenti. also 
larger ultimate divisions that are trullate to oblong and 
Additionally, 


A. vincentii appears to be hemi-dimorphic with smaller 


a narrowly triangular to lanceate lamina. 
sterile fronds, whereas A. rupicola shows no signs of 
dimorphism. 

cell, 


sizes place dum DE rupicola in a group of putative 


Guard spore, and indurated arcus cell 


diploid species. This species may be a parent, 


along with A. radiata, of some component of the 


putative tetraploid A. pentagona. The overall ultimate 
division shape of A. pentagona resembles the 
generally trullate divisions of A. rupicola more than 
those of any other extant, pinnate species. Addition- 
ally, having the carinae begin in the upper half of the 
stipes is unique to these two species. However, A. 
rupicola may not be the parent of all elements of A. 
Some specimens of A. have 


pentagona. pentagona 


398 


Annals of the 
Missouri Botanical Garden 


a tendeney toward relatively larger and slightly more 


compound laminae and smaller ultimate divisions. 


r these 


suggesting an A. parvisegmenta parentage fc 
collections. 

Idiantopsis rupicola is frequently spelled incor- 
(1982) 


some 


rectly as A. rupincola, Tryon and Tryon 


consistently use this incorrect spelling, and 


specimens are determined with However, Maxon 


1908) published the species as A. rupicola, and on 


subsequent determinations used that spelling. There- 
fore. rupicola does not appear to be an orthographic 
error and should be used. 


del Río: 


Guacamalla. P. 


ected specimens examined, CUBA, Pinar 

iil om N to San Juan De 

Wilson 9349 (NY [2]. US): Vedado, Habana, 5 Pozas 

` ` inii on Bro. Alain 2416 (US): Valle de | 

i . R. Proctor 16565 (GH m 1 
G. Britton, C. 


de San Vicente, Bro. Leon 14352 (NY "n 

9. Adiantopsis vincentii M. 5. Barker & Hickey, sp. 
nov. TYPE: Villas, Las Vegas de 
Mataguá, Trinidad Mtns.. Feb. 22. 1956. C. V. 
Morton 10819 (holotype, USH isotype, DUKE. 
Figure 10. 


Cuba. Las 


Laminae tripinnataez axes marcescentes. Ab A. parviseg- 


menta segmentis ultimis le rtilis planis, d xlura pi re ea, ab 


o 
lanceolatis; arcu e cellulis induratis 14 (modus) Pe 


differt. 


Rhizome erect to repent, to 2.7 em long. 0.5-1.5 cm 


(diam: scales bicolorous with a black to dark brown 
lanceate to 
0.20— 


entire to 


center and margins, narrowly 


2.0 


truncate, 


eolden 


)-5.60 mm long. 


=~ 


acicular, — basifixed, 
0. 
slightly repand, apices attenuate 


13.2-36. 


alrocastaneots, 


wide, bases margins 


— 


lo mm 
Fronds 


e to cuneate. 


strict, 5 em long. Stipes atropurpureous to 
always bicarinate adaxially, shorter 


2.9-13.8 cm 


carinae. beginning n 


than lamina, 2 long, 0.7—1.8 mm basal 


basal % of stipe and 
| 


diam.: 
continuing into rachis; scales extending 5412.13 
the 
lanceate to acicular. 


0.06—( 


Z. Jem up stipe, divergent, concolorous, tan, 


narrowly pseudopeltate, 0.70- 


2.80 mm long. ).46 mm wide, bases biauricu- 


late to truncate, margins slightly repand to. entire. 
apices attenuate; fairs septate-capitate. Laminae 


tripinnate, lanceate to lanceolate, drying green 


brown, 12.2-28.5 em long. 4.2-10.5 em wide. apices 
difform, attenuate; laminar tissue papyraceous, hy- 


dathodes marginal: hairs with basal cell elongate, 


clear or white, middle cell short, tan or clear, apical 


cell bulbous, clear, white. or golden: stoma compla- 


nate, guard cells 40.75-53.79 um long. Laminar axes 


marcescent with costae curling acroscopically, atro- 


purpureous to atrocastaneous, grading into lamina and 


taking its color and texture; rachis 0.6-1.2 mm basal 


diam.; carinae golden, grading into lamina: scales 


patent to appressed, linear, biseriate, 0.62-1.24 mm 
long, 0.02—0.04 


apices attenuate. Pinnae ascending to patent. 23-32 


mm wide, bases cuneate to truncate. 
pairs, alternate to sub-opposite, frequently incurved in 


half of 


triangular to linear, 


basal lamina, straight distally. narrowly 


2. 


slightly inequilateral, to 


wide, 


7.8 cm long. 0.7-1.8 em apices attenuate. 
Pinnules patent to slightly ascending, narrowly tri- 
angular to linear, basal pinnules 0.6—1.7 cm long. 


0.3-1.0 em wide, 
than acroscopie. pinnules, apices cuneate le 
Ultimate slightly 


articulate, 


basiscopic pinnules slightly larger 


acule. 


ascending, 
2.7 0.0 mm 


margins entire, 


divisions patent to 


obovate, oblong to elliptic. 


13.0 mm wide, bases cuneate, 


long. 
apices round to acute; stalks persistent, not swollen at 
junction with laminar tissue, 0.1—0.5 mm long. Veins 
pseudodichotomous, anadromous, immersed in lami- 
na, obscure to occult. Pseudoindusia, distinct, 
ereen to white proximally, 

0.46-0.70 mm 


erose [o 


to quadrangular, 
distally, occasionally black maculate. 
long. 0.40— 


entire. Sporangia subglobose, 


0.50 mm wide, margin slightly 


long stalked, arcus of 
32.00-52.10 um tall. 


Spores 64 per sporangium, tetrahedral-globose, golden 


12-16 indurated cells. these 


al maturity, white tọ pale yellow when immature, 
32.00—44.01 Um 


complete, laesura obscured by ornamentation. Chro- 


diam.. echinate. echinae bases 


MOSOTNE number unknow ll. 


Distribution and habitat. Adiantopsis vincentit is 


a rare endemic in the Escambray Mountains of 
Cienfuegos, Sancti Spiritus, and Villa Clara Prov- 
inces. Cuba. 

Notes. Adiantopsis vincentil is most likely to be 


confused with either A. rupicola or A. parvisegmenta. 


The generally smaller, obovate oblone ultimate 


= 


divisions and lanceate to lanceolate lamina « 


vincentit distinguish it from A. rupicola, which has 


generally larger ultimate divisions that are trullate to 


oblong and a narrowly triangular to lanceate lamina. 


Adiantopsis vincentii may be distinguished from A. 
parvisegmenta by three characters. The lamina of A. 
vincentil is papyraceous whereas the lamina of A. 
parvisegmenta is spongiose. Adiantopsis vincentit is 


tripinnate with a generally narrower lamina than the 


quadripinnate and broader. lamina of 


Also, 


parviseg- 


menta. the fertile divisions of A. vincentil are 


Volume 93, Number 3 Barker & Hickey 
2006 Caribbean Adiantopsis (Pteridaceae) 


399 


e 
UA 


Te 


Figure 10. —A. Adiantopsis vincentii (C. V. Morton 10389, US). Scale bar = 3 cm. —B. Ultimate divisions of A. vincenti. 
Seale bar = 0.5 em. 


400 


Annals of the 
Missouri Botanical Garden 


more or less flat, whereas the fertile divisions of A. 


parvisegmenta are strongly. revolute. 
and 


Based on the sizes of 


guard cells, spores. 


indurated arcus cells, Adiantopsis vincentii is placed 
with a putative diploid group of species. Adiantopsis 


vincenlii is also the only hemi-dimorphic Caribbean 


Adiantopsis. On herbarium specimens with large 


rhizomes, and thus presumed 
the fertility of the 


with 


be mature plants, 


fronds is strongly correlated 


— 


rond size. The smallest fronds are consistently 


sterile, with a gradual increase in the number of sori 
and sporangia on progressively larger fronds. This 


pattern is not evident on specimens of other 
Adiantopsis taxa, and may serve as a useful character 
for recognizing the species if it can be confirmed to be 
unique. 
Presently, Adiantopsis vincentit is not convincingly 
implicated in any patterns of reticulate evolution in 
that 


elements of A. pentagona may have an A. vincentit 


Adiantopsis. However, it is possible some 


parentage although no explicit hypothesis is pre- 


sented. Future molecular studies should include A. 


vincentii, because the range of ultimate division 


variation in A. includes the character 
that 


vincentii parent. 


pentagona 
states would likely be inherited from an A. 

Adiantopsis vincentii is named in honor of Michael 
V. Vincent, curator of the Willard Sherman Turrell 
Herbarium (MU) at Miami University, Oxford, Ohio. 
Dr. Vincent has continued to expand the collection at 


MU and has 


world of 


raised the Institutions standing in the 


botany through his activities as curator. 


Further, he has spent a significant amount of time 


dealing with research loans, and through these 
activities facilitates the research of many at Miami 


University. 


Paratypes. CUBA. Las Villas: Trinidad Mins., San Blas- 

Aires, R. A. Howard 6542 (GH, MO, NY [2]: 
Trinidad Mtns.. Buenos Aires, C. p Morton 10389 (US): Las 
Vegas de Matagua. Trinidad Mtns., C. V. Morton 10544 (US). 
Unknown Province: Cuba. Kegers s (F) 


55 


Buenos 


Literature Cited 
Anderson, L. E. 1954. Hover's solution as a rapid permanent 
242-244. 
K T. A. Ranker. 1986 

é . A. Ranker. 
stemalic inf rences 1 ee an stomate size in ile 


mounting me dium for bryophyté s. Brvologist 57: 


= 


ioe cla Paris 
5 rns. Amer. 
Butters. E K. 
a Woodsia ses 


Fern. J. 76: 149 
& R. " Ten 100 A fertile mutant of 
Amer. J. Dot. : 132 

. 1906. Index Filicum. H 


Y 


Christensen, Hagerup, Copenha- 
gen. 

^ land, E. B. 1947. Genera Filicum. Chronica Botanica. 

Waltham, Massie husetts. 


9 A. M. 1964. 


life cycle in the ferns. Science 143: 261-203, 


Ameiolic alternation a gene rations: a new 


C. A. L A. 


Vol. 5 [Genera Filicum. 


1852. Mémoires sur les Familles des Fougeres, 


= 


Exposition des genres. de 
famille des Polypodiacées |. 
—————, 1857. Mémoires sur les Familles des Fougères, 


7 |leonographie des espèces nouvelles]. 


0 
B. R. J., M. S. Barker € M. Ponce. 2003. An 


„ hybrid from northeastern Argentina and 
cinto . Fern J. 93: 42-44. 
s H. . 1975. Carl Christensen. Taxon 24: 499— 


& J. C. Baker. 1 
London. 
1940. Plant Microtechnique. MeGraw-THill. 


383. Synopsis Filicum. 2nd ed. 


11 1 0 .J.4 
W. H. s n, 
nos n, D. / 
lew York 
Kle kows E. J. 1973. 
1 PO DIIS a new hypothesis. 
Bot 60: 535-5 
Kramer, K. U. 1957. A revision - a genus Lindsaea in the 
"n ‘ta e Neerl. 7-290. 
ae . 


Sexual and subsexual systems in 


mer. J. 


new world. 


Kunze, G. 1850. Die J. U. Fleischer, 
Leipzig. 
Lellinger, D. B. gs Polypodiaceae sublam. Adiantoideae. 


Pp. 1-23 in Maguire (editor), The Botany of the 


Guayana i M IX. Mem. New York Bot. Gard. 
23: 1-832. 

—— ——, |989. The ferns and fern-allies of Costa Rica, 
Pusat un and the Chocó (part l: Psilotaceae through 
Dicksoniaceae). Pteridologia 2A: 1-364. 

. 2002. 4 modern multilingui il glossary for taxonomic 
pteridology. Pteridologia 3: 1—2( 

Maxon, W. R. 1908. The Cuban species of Adiantopsis. 
Contr. U.S. Natl. Herb. 10: 485-4806. 

Mickel, J. T. & A. R. Smith. 2004. Pteridophytes of Mexico. 
Mem. New York Bot. Gard. 88: 1-1055 
oran, R. C. & A. R. Smith. 2001. aea rela- 


E elween neotropical and African-Madagascan 
pterido ne Brittonia 53: 304—351. 
OpenOffice.org 2003. OpenOffice.org Cale. 
OpenOffice.org 

9 


Pacheco, L. ] 55. 


Version l.l. 
2. http://www.openolfice.org. 
Idiantopsis. 106 in Davidse, M. 
lapp (gen. editors) & R. C. Moran & R. 
Riba (pte atis editors), Flora Mesoamericana. Vol. 
Vutó- 


— 


> 


Sousa S. s 


Psilotaceae a Salviniaceae. Universidad Nacional 


noma de México, México D. E; Missouri Botanical 
Garden, Louis; The Natural History Museum, London. 
Ponce, M. M. & M. A. Morbelli. 1989. The Cheilanthes 
dichotoma group of South America. Amer. Fern J. 79 


127-135. 
Proctor. C. R. 
Vat ete 
. Ferns of e uerto Rico and the 

Mem. W ay Bot. Gard. 53: 1-389. 
SAS Institute. 2000. E System for * indows, 
SAS Institute, Cary, North ( 


Sehnem, A. coleção de 


Ferns of Jamaica. British Museum of 
5 

Virgin Islands. 
Version. 8. 
carolina. 

Ds E fitos do Rio 


l. Pesquisas 3: 495 


— 1% E na legio de 5 
Sul, \ Seat 5: 19-2 

— . 1972 Pleridac eas, Pp. 1-244. in P. R. Reitz (editor) 
Flora Parte L Herbario Barbosa 


Smith. A. R. ! a 
California Acad. Sci.. San Francisco. 
——— & M. 8. a 198 


ecological observations of ferns from EL’ 
321—329 


loja of Chiapas. Part 2: Pteridophytes. 


. Chromosome numbers and 


. Paraguay. 
Fern Gaz. 12: 


Volume 93, Number 3 
2006 


Barker & Hickey 
Caribbean Adiantopsis (Pteridaceae) 


Systematics e lation. Committee for Descriptive Termi- 
O of simple symmetrical plane 
45-156, 245-247. 

1982. Ferns and allied plants 
Springer- 


shapes ie wu Taxon 11: !“ 
N. M. & A. Tryon. 
with special 1 to tropical America. 
Verlag, New York. 
& R. G. Stolze. 1989. Pteridophyta of Peru: Part H: 
Fieldiana, Bot. 


13. Pteridaceae—15. Dennstaedtiaceae. 
28 


22: l- 


1990. 
Green (editors), 
Plants, Vol. 1 
Springer-Verlag, 


——., A. F. Tryon & K. Kramer. Adiantopsis. 
P. 212 in K. U. Kramer 
The Families and Genera of Vascular 


Pteridophytes and Gymnosperms. 


Walker, T. G. 1973. Evidence from cytology in 
classification of ferns. Pp. 91—110 in A. C. Jermy (editor), 
The Phylogeny and Classification of Ferns. Bot. J. Linn. 


Soc. 67, suppl. 1. 


SUPPORT FOR AN EXPANDED 
SOLMS-LAUBACHIA 
(BRASSICACEAE): EVIDENCE 
FROM SEQUENCES OF 
CHLOROPLAST AND NUCLEAR 
GENES'* 


Ihsan A. Al-Shehbaz.^ 


Ji-Pei Vue. Hang Sun," 
and Jian-Hua Li” 


ABSTRACT 


Sequences of the plastid maturase (math) and nuclear chalcone 


assess hy loge rie tie re lati lions 


synthase (Chs) were analyzed separately and in combination to 


ips of Solms- nio Muschl. (all nine known species and two undescribed ones) to the genera 


Baimashania Al She ar Tae a R. Br. Desideria Pamp.. Leiospora (C. A. Mey.) Dvořák. Christolea Cambess.. and 
Phaeonychium O. | aceae). B VHC is clustered with Aubrieta deltoidea (L.) DC. and Arabis ble pharopholla 
Hook. & me indus POM De dina, Leiospora. s a. and Phaeonychium are more closely related to Mattht 


than to CE l they forn ell-supported clade. 


Vithin this clade, Leiospora is sister to a subclade containing i sister 


eroups Christolea ud Solms- cda sd. The Solms- "es hia s.l. group c ontains Solms-laubachia, Desideria, and a species of 


Pas onychium. 


within the ¢ group. Therefore, our results suggest that 


Nevertheless. more bi cies of Desideria, Parrya. and Phaeonychium are needed to further test our findings. 
Chs 


Key words: Brassicaceae. Christole Desideria, 


laubachia. 


Solms- laubachia be ex pal mded 1 inch ide 


Letospora, 


either Solms-laubachia nor Desideria is monophyletic. whereas Plhaeonychium jafrii Al-Shelbaz is e uns dded 


Desideria and P. jafru. 


A 


math, Parrya. Phaeonychium. phylogeny. Solms- 


The Brassicaceae (Cruciferae) consist of approxi- 
340 genera and 3350 species (Al-Shehbaz. 
1984: Appel € Al-Shehbaz. 2003 


phyletic family characterized by a number of synapo- 


mately 
They are a mono- 
morphies, including flower structure (e.g.. four sepals. 


four petals, often six tetradynamous 11 and often 


a dehiscent capsule derived from a bicarpellate pisti 
with a septum dividing its ovary into two locules. Many 
plants of this family are common vegetables (e.2.. 


cabbages, broccoli, cauliflower, Brussels sprouts. 


radishes. 
(L.) Heynh., 


biology, 


turnips, watercress). Arabidopsis thaliana 
the model organism in plant experimental 
belongs to the Brassicaceae. 

Before the advent of molecular systematics, classi- 
fication the family 
Schulz. 1936: Janchen. 1942: 


based almost exclusively on morphology. 


systems of [8. p. 
\l-Shehbaz, 198 l) were 
Over the 
molecular studies have greatly contrib- 
better 


past decade, 


uted tọ a understanding of phylogenetic 


relationships within the Brassicaceae. Such studies 
been commonly 
the 


show that many characters that have 


used to define tribal and subtribal divisions ii 


family are homoplasious. As a result, many tribes as 
traditionally recognized (e.g Arabideae DC., 
dieae DC., Euclidieae DC., Sisymbrieae DC 


.) are 
highly artificial (Koch et al.. 


Lepi- 


2003: Al-Shehbaz et al.. 
2006). Generic limits within the family have also been 
controversy. (AM-Shehbaz. 
Appel & Al-Shehbaz, 2003; Al-Shehbaz et al.. 2006). 


However, molecular data have shed enormous light on 


a major source of 


the monophyly of various genera. For example. results 
from the nuclear ribosomal DNA internal transeribed 
spacer (nrDNA TTS) sequence data show that Arabis L. 
and Arabidopsis Heynh., as traditionally delimited, are 
polyphyletic because they included purportedly un- 
related species that had been assigned to many other 
1999; Al-Shehbaz et al., 1999: Al- 
& Al-Shehbaz. 2003: Al- 


genera (Koch et al., 


Shehbaz, 2003: O' Kane 


"This study was supported by grants to HE Sun from the Chinese Nal tural Science Foundation (CNSE, grant numbers 4033202 1. 
X2 


304201200419), the geane Project of the Chines 


a Mercer Fellowship to J. P. Yue from the Arnold Arboretum. 
“The editors of the Annals thank Sophia Balcomb for het 
‘Laboratory of Biodiversity and Biogeography. 


Kunming. Yunnan 650204, People’s Republic of China. The 


Veademy of Sciences (KSC 
ent Contributions in Science and Tec 1 to Prof. Wu Zhe nevi in 2001 (KID-WU- 


Kunming 19 110 ile 


00), the Yunnan Province Government 
2001-02). and 


editorial contribution to this article. 


of Botany. Chinese Academy of Sciences. Heilongtan, 


authors contributed e qually to the work. hsunOmail.kib.ac.cn. 


"Graduate School of the Chinese Academy of Sciences, Bei “jing, 100039, People’s Republic of China. 


| P.O. Box 299, St. 


^ Arnold Arboretum and Harvard University Herbaria, 


"Missouri Botanical Garden Louis. 


Missouri. 
22 Divinity 


U.S.A. 63160-0299, ihsan.al-she P . org. 


Venue, Massachusetts 02138 College of Lif 


Sciences, Zhejiang University, Hangzhou 310029, People's Republic of China. Author for corre 9 nee: : iios b. a e d: 


ANN. Missouni Bor. Ganp. 93: 402-411. 


PUBLISHED ON 


24 Ocroger 2006. 


Volume 93, Number 3 
2006 


Yue et al. 403 


Support for an Expanded Solms-laubachia 


Shehbaz. 2005). Warwick et al. (2002, 2004) have 
shown that generic limits of Sisymbrium L. and 
Smelowskia C. A. Mey. need reconsideration, 
Another taxonomically difficult group involves 
several Himalayan and Central Asian genera, in- 
cluding Solms-laubachia Muschl. (9 spp.). Desideria 
Pamp. (12 spp.), Leiospora (C. A. Mey.) Dvorak (6 
spp-), Christolea Cambess. (2 spp.). Parrya R. Br. (25 
spp.), and Phaeonychium O. E. Schulz (7 spp.) (Appel 


& Al-Shehbaz, 2003). These genera occupy similar 
habitats (alpine scree slopes and limestone crevices) 
and exhibit) overlapping morphological characters 
e.g., perennial habit, latiseptate or terete fruits. 

1 


well-defined 


basal rosettes, non-saccate sepals). 
Consequently, the same species have often been 
placed in different genera. For example. Desideria 
pumila (Kurz) Al-Shehbaz was originally described as 
Parrya (Kurz, 1872) and later transferred to Christolea 
(Jafri, 1955). Vvedenskeyella Botsch. (Botschantsev. 
1955). or Desideria ( M-Shehbaz. 2001). 

Parrya is most highly diversified in central Asia but 
has three species occurring in. Arctic and subarctic 
Asia and North America (Appel & Al-Shehbaz. 2005). 
Four of the nine known species of Solms-laubachia. 
three of the 12 of Desideria, and all six of Letospora 
were originally described as species of Parrya 
(Schulz. 1936). Recent morphological (Al-Shehbaz. 
2001; Al-Shehbaz € Yang, 2001) and cytological 
(Yue et al., 2003, 2004) studies have indicated that 
while Desideria, Solms-laubachia, Christolea, and 
Leiospora are closely related to one another. together 
they are distinct from Parrya. Nevertheless, none of 
these genera. have been included in phylogenetic 
analyses based on molecular data (Koch et al.. 2003). 

Baimashania Al-Shehbaz (2 spp.) is a recently 
described genus from China based on specimens that 
have been misidentified as species of Solms-laubachia 
or Leiospora (Al-Shehbaz, 2000b 


The objective of this study was to infer phylogenetic 


relationships of Solms-laubachia to the presumably 


related genera Baimashania, Christolea, Desideria. 


Leiospora, Parrya, and Phaeonychium based on 
sequences of the chloroplast gene malK (maturase) 
and the nuclear gene Chs (chalcone synthase). These 
two markers have been used successfully in resolving 
generic relationships within Brassicaceae (e.g.. Koch 
el al., 2000, 2001). 


MATERIALS AND METHODS 


PLANT MATERIAL 


— 


"fly samples were used in this study. including 11 


species of Solms-laubachia (nine in total plus two 


undescribed species), four of Desideria (of 12). two of 


Leiospora (of six). and one each of Baimashania (of 


two). Phaeonychium (of seven), Christolea (of two). 


— 


and Parrya (of 25). The remaining samples represent 


24 genera of Brassicaceae. This broad sampling of the 
these 


because none of genera 


amily IS necessary 


mentioned above have been included in previous 


phylogenetic analyses and it is unclear whether they 
form a clade and what taxa are their outgroups. Ás in 
Koch et al. (2001), we used Aethionema grandiflorum 
Boiss. & Hohen. as the outgroup because Aethionema 
R. Br. is sister to the rest of Brassicaceae (Zunk et al., 
1996: Galloway el 1998). In their recently 
proposed tribal classification of the family, Al- 
Shehbaz et al. (2006) placed Parrya with Matthiola 
R. Br. in the tribe Anchonieae DC.. Baimashania in 


al., 


the Arabideae, and Solms-laubachia. Desideria, and 


Leiospora in the Euclidieae. 


DNA EXTRACTION, PCR, AND SEQUENCING 


racted from. silica 


Total genomic DNAs were ex 
gel-dried leaves using a Qiagen DNeasy Plant Mini 
Kit following the manufacturers protocol (Santa 
California). Genomic DNAs of three acces- 
sions of Desideria and one each of Leiospora and 


Solms-laubachia were kindly provided by Jason R. 


~ 


rant (Université de Neuchâtel, Switzerland). 
Plastid math 


Double-stranded DNAs of matK were amplified using 
primers trnK-7 LOF* and trn K-2R* originally designed by 
Johnson and Soltis (1994) and slightly modified later by 
Koch et al. (2001). A 50-uL polymerase chain reaction 
(PCR) included 50-100 ng DNA, 5 uL of 10X Tag 
polymerase buffer. 4 uL of dNTP (2.5 mM). 4 uL of 
MgCl» (25 mM), 2 uL of each primer (10 uM). 0.5 uL of 
Taq polymerase (5 units/UL), and 29.5-31.5 uL of 
sterilized water, depending on the amount of DNA used. 
The PCR profile used to amplify the AK region 
included 34 cycles (94 € for 1 minute 30 seconds, 48— 
52 € for 2 minutes, 72°C for 3 minutes) and a final 
extension at 72°C for 15 min. PCR products were gel- 
purified usinga Qiagen Gel Purification Kit, following the 
manufacturer's. protocol. Both strands were sequenced 
using the ABI BigDye Florescent Terminator Chemistry 
(Applied Biosystems, Foster City, California) and the 
trnK-2R*, mat K-1F., 
matK-1089R, 


following primers: trnK-710E*, 
matK-3121, YF, mat K-A95R, 
and matK-1089R (Koch et al., 2001). 


Nuclear Chs 


PCR amplification of the entire gene, including the 
I : 8 


promoter region, was performed using the primers 


Annals 
1 Eni Garden 


Chs-Forl and Chs-Rev 5 (Koch et al., 1999). 
Components of the PCR reaction were the same as 


CR 
5-minute denaturation at 95°C 


those described. for n matK gene, and the 


program included : 
and 35 cycles, T by an additional 15-minute 
Each eyele consisted of | minute 
95 C, 


2-minute extension al 


extension al 72 C. 
| minute e annealing at 55 C. 
All of the PCR 


amplifications resulted in a PEN 1 band 


of denaturing at 
and a 
on an agarose gel. PCR products were gel-purified 
Kit, 
manufacturer's protocol, and then were cloned into the 
pGEM-T 


Wisconsin). 


using a Qiagen Gel Purification following the 


cloning vector (Promega, Madison, 
To detect possible allelic variation. two 
clones from two independent PCR reactions were 
sequenced respectively for each sample. Sequencing 
primers included vector primers T7 and Sp. and 
Koch et als (1999) primers (Chs-FOR2, Chs-REV2. 
Chs-REV3, Chs-FOR4, and Chs-REVA). 

Sequences were analyzed using an Automated ABI 
Genetic Analyzer 3100, and were then edited using 


Ann 


The edited sequences were then 


Sequencher (version 4.0, Gene Codes Corp., Ine., 
Arbor, Michigan). 
aligned manually. 


PHYLOGENETIC ANALYSES 


Phylogenetic analyses were performed for the data 
sets individually and in combination using maximum 
(Farris et al., 1970: Fitch, 
Bayesian methods (Rannala & Yang, 


parsimony 


heuristic searches for maximum parsimony analysis 
PAUP*4.0 (Swofford. 2002) 


conditions: sequence 


were conducted using 


with the random 
and 10 


each step, tree bisection-reconnection branch swap- 


following 
addition with 1000 replicates trees held 
ping, and MULPARS on and the steepest descent off. 
Character state changes were unordered and P 
weighted in the analysis, and a were treated : 
1000 1 cus 


were performed to estimate the support for individual 


missing data. Bootstrap analyses « 


clades using heuristic tree search options as in the 
maximum parsimony analyses. To further assess 
relative support for the NE clades, a decay 


1996) was also 
DNA 


The best 
substitutions for the 
Bayesian analysis was determined using Modeltest 
3.06 (Posada & Crandall, 1998). Bayesian 
1996: Mau et al., 1999) 
carried oul MrBayes 3.0b3 
(Huelsenbeck & 2001. available at 


—httpZ/mrbayes.esit.fsu.eduz7) with the model 


analysis (Sorenson, used. 


evolutionary model of 
version 
analyses (Rannala & Yang, 
were using version 
Ronquist. 
pa- 
rameters as estimated in Modeltest. Bayesian analysis 


was started from random trees and employed four 


Markov chain Monte Carlo runs, monitored over 2 X 


10° generations, sampling trees every 100 generations. 


The 


generalions 


Runs were repeated twice to confirm. results. 


resulting log likelihood and number o 
were plotted to determine the point after which the log 
likelihoods had stabilized. After discarding the trees 
saved prior to this point as burn-in, the remaining 
trees were imported into PAUP*, and a 50% majority- 
rule tree was produced to obtain posterior probabil- 
ities of the clades (Huelsenbeck & Ronquist, 2001). 
Internodes with posterior probabilities = 95% were 
considered statistically significant. 

length difference (ILD) test 
1994, as implemented in PAUP*) was 


The  incongruence 


(Farris et al., 


used to assess potential conflicts between the 
phylogenetic signals from the two different DNA 
fragments. For the test, 100 replicates were analyzed 
with a heuristic search, each with 10 random 


sequence addition replicates. In addition, we com- 
pared matK and Chs trees to examine whether there 
were well-supported (> 70% bootstrap support), bul 


conflicting clades. 
RESULTS 
SEQUENCE CHARACTERISTICS 


and 24 


newly gathered in this study. 


Chs 


An additional 26 matk 


Twenty-five math sequences were 
and Chs sequences were obtained from GenBank and 


(Table 1). 


The amplified segment also included 


included in this study 
a 300-bp 


fragment at the 3“ end of the irn region. However, 


the alignment of this region across Brassicaceae was 
ambiguous, and thus our analvses did not include 
sequences of the segment. The sequences of the math 
gene were aligned readily by hand, resulting in a data 
set of 1536 bp. of which 530 sites were variable and 
237 sies (15.43%) were potentially parsimony in- 
formative. The 
6.02%-8.27% 


and from 0.067%-—7.93% within the ingroup. 


sequence divergence ranged from 


between. the oulgroup and ingroup 


Limits of the introns and promoter regions of the 
Chs gene were determined by comparison with 
published) sequences (Koch et al.. 1999). These 


regions could not be aligned unambiguously across 
Brassicaceae and thus were removed from phyloge- 
nelic analyses. However, exon sequences were aligned 
readily by hand, and the alignment resulted in a data 
set of 1174 bp. of which 511 were variable sites and 
304 sites (81%) were potentially parsimony informa- 
19.3%. 


M between the outgroup and ingroup and. from 


live. The sequence divergence ranged from 
.3496—17.996 within the ingroup. 
Results of the ILD test showed that our math and 


Chs data sets were not significantly different (P = 


Volume 93, Number 3 
2006 


Yue et al. 
Support for an Expanded Solms-laubachia 


405 


Table 1. 


from GenBank. 


Vouchers and sources of species used 


in this study. 


An asterisk (*) indicates that sequences were obtained 


Taxon Voucher Source matK Chs 
Aethionema grandiflorum Boiss. & Hohen. AFI44354* — AFI12082* 
Alliaria petiolata (M. Beib.) Cavara & Grande AF144363* A 144537* 
Arabidopsis thaliana (L) Heynh AFI44348* — AF112086* 
Arabis blepharophylla Hook. & Arn. AFI44353* ñ AFII2087* 


Arabis divaricarpa A. Nelson 
Aubrieta deltoidea (L.) DC 
Baimashania pulvinata Ne Shehbaz 


c 
— 
m — 
E 


irbarea vulgaris R. Rr. 
pen rubella Reuter 
Cardamine amara L. 
Cardamine penzesii Anéev & Marhold 
Cardamine rivularis Schur 


Christolea crassifolia Cambess. 


Cochlearia danica | 


Canade himalaica (Edgew.) Al-Shehbaz, 


cane & R. A. Price 
Desideria baiogoinensis (K. C. 
Al-Shehbaz 


Desideria himalayensis (Cambess) Al-Shehbaz 


Desideria linearis (N. Busch) Al-Shehbaz 


(Anderson) Al-Shehbaz 


Erysimum handel-mazzettii Polatschek 


Desideria stewartii 


Fourraea alpina (L.) Greuter & Burdet 

Halimolobus perplexa (Henderson) 
Rollins var. perplexa 

Ionopsidium abulense (Pau.) Rothm. 


Leiospora exscapa (C. A. Mey.) Dvořák 


Leiospora pamirica (Botsch. & Vved.) 
Botsch. & Pachom. 

Lepidium campestre (I.) R. Br. 

Br. 


=> 


Matthiola incana 
Meyer 


1 (L. 
Microthlaspi 1 aes (L.) F. K. 
* e (Hook. f 


e 


-Shehbaz cane & R. A. 


5 pumila (Stephan) AL "Shehbaz, 


"Kane & R. A. Price 
Parrya nudicaulis (L.) Regel 
Phaeonychium jafrú Al-Shehbaz 
Rorippa amphibia (L.) Besser 
Sinapis alba L. 


Sisymbrium irio | 


Solms-laubachia eurycarpa (Maxim.) Botsch., 1 


Solms-laubachia eurycarpa (Maxim.) Botsch., 2 


Solms-laubachia lanata Botsch. 
s sd aca (W. W. Sm.) 
Schu 


5 minor Hand.-Mazz. 


Kuan & Z. X. An) 


L & es 
I 


Yue 0222 (KUN) 


B. Bartholomew 


et al. 9499 (MO 


Yue 0246 (KUN) 


McBenth 1486 (E) 
B. Bartholomew 

al. 9549 (M. 
McBeath 2105 (E 
Yue 0369 (KUN) 


D 
oe 


— 


Murray et al. 457 
(ALA) 
B. Bartholomew 


et al. 9790 (MO 


Yue 0244. (KUN) 
Yue 0233 (KUN) 


S; ise ji 15 shu Exped. 


78 (E) 
ie » 58 (KUN 


Yue 0254 (KUN) 
Yue 0157 (KUN) 


Yue 0379 (KUN) 


) 


)) 


) 


China, Yunnan, 


Deqing 


China, Xinjiang 


China, 


Tibet, 


Mozhugongka 


Nepal 


China, Xinjiang 
Jiang 


India 


China, Yunnan, 


Zhongdian 


Russia 


China, Xinjiang 


China, 


China, 


Tibet, Lhasa 


Tibet, Lhasa 


China, Qinghai, 


Yushu 


China, Yunnan, 


Deqing 


China, 


Tibet, Lhasa 


China, Yunnan, 
Jeqing 


( hina, Sic huan, 


'anyuan 


AF144351* 


AF 14435 
DQ409251 


AF144330* 
AF144334* 
AF144337* 
AF144364* 
AF144365* 
50400250 


AF174531* 


AF144356* 
DQ409252 


50409200 
50409254 


50400205 
DQ409262 


AF144335% 
AF144346* 
AF144308* 
DQ409263 
DQ409255 
AFI144359* 


AFI44361* 
AF144362* 


AF144358* 


AF144345* 
DQ409253 


DQ409261 
F 


DQ409264 
DQ409243 


50409240 
50409240 


50400257 


AF112090* 
AF112109* 
DQ409227 


AF112108* 
AF112106* 
AFII2085* 
AF144538* 
AF144539* 
DQ409232 


AF144532* 


AF144531* 
D(409228 


DQ409242 
DQ409230 


5040924] 
50409238 


AF112102* 
AF112094* 


l144542* 
50400230 


5040023] 


AF144534% 
AJ427536* 
AF144536* 


AFI44533* 


AF112092* 

DQ409229 

DQ409237 

: > 44530* 
437* 

i " 4454 1% 

DQ409240 


DQ409219 


DQ409222 
DQ409225 


DQ409233 


406 


Annals of the 
Missouri Botanical Garden 


Table |. Continued. 


Taxon 


Voucher 


Source math Chs 


Solms-laubachia platycarpa (Hook. f. & Thomson) 
Botsch. 


Solms-laubachia pulcherrima Muschl. 


Solms-laubachia retropilosa Botsch. 


Yue 02: 


Yue 01553 (KUN) 


Yue 0248 (KUN) 


35 (KUN) China, Tibet, 50409245 DQ409221 


Danexion 


China. Yunnan, 10409247 100409223 


Lijian 
China. Sichuan. 


110409248 10409224. 


Kangding 

Solms-laubachia sp. indet. D. E, iie China, Sichuan, 50409253 DQ409234 
el d 1727 Xiangcheng 
(HUI 

Solms-laubachia sp. indet. 2 D. E. P s China. Tibet, 100409260 DQ409236 
et al. 31975 Riwoge 
(HUH) 


Solms-laubachia xerophyta (W. W. Sm.) 
Comb 
Solms-laubachia xerophyta (W. W. Sm.) 
Comber, 2 
Solms-laubachia zhongdianensis J. 
M-Shehbaz & H. 


Thlaspi arvense L. 


. Yue, 


Sun 


Turritis glabra |. 


Yue 0250 (KUN) 
Yue 0251 


Yue 0156 (KUN) 


China, Sichuan, DQ409244 100409220 


Daocheng 


(KUN) China. Yunnan. 110409259 DQ409235 
Zhongdiar 

China, Yunnan, DQ409250 DQ409226 
Zhonedian 

AF144360* AF144535* 


AF144333*  AF112091* 


0.51). Furthermore, maximum parsimony and Bayes- 


ian analyses based on the two individual data sets 
resulted in several well-supported monophyletic 
groups, and there were no well-supported (bootstrap 
95%) but 
Therefore. 
the 


— 


support > 70% and posterior probability > 


incongruent clades (trees not shown). 


results of phylogenetic analyses based on 


combined data set are presented below. 
PHYLOGENETIC RELATIONSHIPS 

The combined data set of math and Chs had 50 taxa 
and 2710 sites, of which 1669 (61.6%) were invari- 
able, 440 (16.2%) were unique nucleotide changes. 
and 601 (22.2%) were parsimony informative. The 
sequence divergence 
the 


7.93% within the ingroup. 


ranged from 
between outgroup and ingroup and 0. 


— 
— 
=~ 


Parsimony analyses yielded six trees of 2889 steps: 
the strict consensus is shown in Figure 1 (consistency 
index (CI) = 0.50, retention index (RI) = 0.69, and 
rescaled C) 0.35). The 
relationships of the taxa whose matk Chs 
sequences he (cf. 


were consistent with those al. 


consistency index (R = 


and 


were obtained from | GenBank 
Fable 


(2001). 


Baimashania pulvinata 


of Koch et 


To these, we included our 

Al-Shehbaz and 
Polatschek. The 
clustered with Aubrieta deltoidea (L) DC. and Arabis 
blepharophylla Hook. & Arn. (Clade A) and was 
remotely related to both Solms-laubachia and Leios- 


pora, whereas the latter was positioned in a 


samples of 
Erysimum 
handel-mazzettit 


former spectes 


clade 


including Arabidopsis, Turrits L.. 
O'Kane & I 
Halimolobus 
Al-Shehbaz, O'Kane & 

Parrya nudicaulis (L.) Regel was strongly supported 
the clade (Clade B, Fig. 1) containing 
Vatthiola incana ] R. Br., Leiospora, Christolea. 
Solms-laubachia, Desideria, and Phaeonychium jafrii 
Al-Shehbaz. Matthiola was sister to the clade contain- 
C, Fig. 1). In Clade C, 
wo accessions of Leiospora formed a well-supported 
group. 


Olimarabidopsis Al- 
Price, Capsella Medik., 


Tausch, 


Shehbaz. t 


Arabis 


and Crucihimalaya 
Price. 


as sister to 


ing the latter five genera (Clade 


occupying a sister position to other remaining 
taxa that formed Clade D (Fig. 1). Species of Solms- 
laubachia, Desideria 


. and P. jafrii formed a clade 


(Fig. l. Clade E), with C. crassifolia Cambess. as 
sister group. Within Clade E, there were three 


sube e the first of which (Clade F) included D. 
(N. Busch) Al-Shehbaz. D. stewartii (Ander- 


son) Al-Shehbaz, and D. himalayensis (Cambess.) Al- 


linearis 


Shehbaz. The second had P. jafrii alone and was sister 
to the third clade (Clade G), which contained D. 
Kuan € Z. X. An) Al-Shehbaz 
and all species of Solms-laubachia. While interspe- 
cific relationships of Solms-laubachia were weakly or 
5. Botsch. I). 
batogoinensis Clustered together with strong support 
94%). Clade G trichotomy 


including D. baiogoinensis + 5. lanata, S. platycarpa 


baiogoinensis (K. C. 


moderately supported. lanata and 


(bootstrap = was a 


(Hook. f. & Thomson) Botsch.. and Clade H. Within 
this clade retropilosa Botsch. and S. linearifolia 
(W. W. Sm.) O. E. Schulz formed a moderately 


supported clade (bootstrap = 74%) that was sister to 


Volume 93, Number 3 Yue et al. 407 
2006 Support for an Expanded Solms-laubachia 


Aethionema grandiflorum 


E 98 . Desideria linearis 
5 98 Desideria stewartii 
5 Desideria himalayensis 


78 Solms-laubachia xerophyta | 
le fp Solms-laubachia minor 
- Solms-laubachia xerophyta 2 
BET Solms-laubachia „„ 
157 l Solms-laubachia pulchert 
FT Solms-laubachia sp. indet. T 
Solms-laubachia sp. indet. 2 


——— Solms-laubachia eurycarpa | 

25 Solms-laubachia eurycarpa 2 
99 74 Solms-laubachia retropilosa 

D 4 Solms-laubachia linearifolia 


Solms-laubachia platycarpa 
94 "e Solms-laubachia lanata 

4 Desideria baiogoinensis 
Phaeonychium jafrii 
Christolea crassifolia 

100 — Leiospora pamirica 

25 L— Leiospora exscapa 
Matthiola incana 

Parrya nudicaulis 


Thlaspi arvense 
Alliaria petiolata 
Microthlaspi perfoliatum 


Sisymbrium irio 
Sinapis alba 
Fourraea alpina 
Cardamine penzesii 
Cardamine rivularis 


Cardamine amara 
Barbarea vulgaris 
Rorippa amphibia 


Arabidopsis thaliana 
Capsella rubella 
Arabis divaricarpa 
Halimolobus perplexa 
Crucihimalaya himalaica 
Turritis glabra 
Olimarabidopsis cabulica 
Olimarabidopsis pumila 
Erysimum handel-mazzettii 
Lepidium campestre 

85 Aubrieta deltoidea 
100 A[ 55 Arabis blepharophylla 
e LLL Baimashania pulvinata 
100 —— Ionopsidium abulense 
55 Cochlearia danica 


uu 
2 


JT 


Figure Strict consensus of six parsimonious trees based on the combined data set of matK and Chs genes. Numbers 
above and I branches are bootstrap percentages and decay indices, respectively. Letters A-J refer to c Jade 's discussed in 
the text 


408 Annals of the 

Missouri Botanical Garden 
a polytomous branch (Clade I, bootstrap = 57%). Specimens of Baimashania pulvinata, a species 
Clade I contained S. eurycarpa (Maxim.) Botsch., two endemic to northwestern Yunnan. had been mis- 
undescribed species (Solms-laubachia sp. indet. land identified as Solms-laubachia ciliaris (Bureau & 
2), and Clade J, which consists of S. minor Hand. Franch.) Botsch. (Al-Shehbaz, 2000b). Our sequence 


-Mazz. + one accession of S. „ (W. W. Sm.) 


Comber (bootstrap = 78%), S. zhongdianensis J. P. 
Yue, Al-Shehbaz & H. Sun + 5. E emus Muschl. 
(bootstrap = 64%), and the second accession of 5. 


xerophyta (W. W. Sm.) Comber. 

The best-fit model of evolution for the combined 
matK and Chs data set was GTR + I+ G, as indicated 
by the Modeltest. The base m ncies were A = 
0.287, C = 0.195, G = 0.19, T 0.328, and the 


gamma distribution shape parameter was 0.6244. 


Bayesian analyses with the estimated parameters 


produced trees with identical topology Figure !. 


with a few discrepancies in the clade containing 
species of Solms-laubachia, Desideria baiogoinensis, 
and Phaeonychium jafrú (Vig. 2). Phaeonychium jafrii 


the 


formed a sister relationship with Clade G i 


maximum parsimony tree (Fig. I). while in the 
Bayesian tree S. platycarpa was sister to the clade 


consisting of D. baiogoinensis, P. jafrii, and the 


É 


remaining species of Solms-laubachia (Vig. 2). In the 


maximum parsimony tree, 5. eurycarpa and the two 
undescribed Solms-laubachia species formed a poly- 
However, 


tomy. in the Bayesian tree, one undescribed 


species (Solms-laubachia sp. indet. 1) clustered with 


the two accessions of 5. eurycarpa, while the other 


(Solms-laubachia sp. indet. 2) was sister to the clade 


containing S. eurycarpa, S. sp. indet. 1, S. xerophyta, 
S. minor, S. zhongdianensis, and S. pulcherrima. 


Nevertheless, the topological differences between 


the maximum parsimony and Bayesian. trees were 


weakly supported. 
DISCUSSION 


Schulz (1936) placed Solms-laubachia and Parrya 


(including Letospora) with Matthiola in the tribe 
Matthioleae O. E. Schulz. Christolea in the tribe 


Sisymbrieae, and Desideria (as a synonym of Ermania 
Cham.) and Phaeonychium in the tribe Arabideae. 
These tribes were defined largely on the basis of silique 
fruits with incumbent cotyledons and open calyx 
(Sisymbrieae), aceumbent cotyledons and closed calyx 
(Matthioleae), or aceumbent cotyledon and open calyx 
(Arabideae). However, in the math + Chs tree (Fig. 1). 
Matthiola is sister to the clade containing Christolea, 
Desideria, Solms-laubachia. J 

= 78% 


pora (bootstrap = 


Phaeonychium, and Leios- 
), suggesting that these genera 
are Closely related to one another and, together, to 
Matthiola. Our results provide further evidence that 
1930) are not NOUS 
2003: Al-Shehbaz et al.. 200€ 


most. tribes of. Schulz ( 
(Koch et al.. : 


data did not place Baimashania close to Solms- 


laubachia. close re- 


Instead, these data supports! 
lationship of Baimashania with Aubrieta Adans. and 
Arabis. Therefore, our results support the recognition 
of Baimashania as a separate genus from Solms- 


laubachia, and Al-Shehbaz et al. (2006) assigned 


these two genera to the tribes Arabideae and 
Kuclidieae, respectively. 
In both the maximum parsimony and Bayesian 


trees (Figs. I. 2), Parrya is sister to the clade (B) 


containing Matthiola, Leiospora, Christolea, Desideria, 
Solms-laubachia, and Phaeonychium, indicating that 
none of the latter five genera is closely related 

This is supported. by morphological and 
cytological evidence (Al-Shehbaz, 2001: Al-Shehbaz 
& Yang, 2001: 20023. 2004). 


laubachia differs from Parrya in that it has wingless 


Parrya. 


Yue el al. Solms- 


seeds, rounded replum concealed by the strongly 


angled valve margins, valves apically united. with 
obsolete style and replum, fruits readily detached 
basally from the pedicel, entire stigmas or with non- 


decurrent lobes, and eglandular papillae. By contrast, 


Parrya has winged seeds, strongly flattened and 
readily visible replum, flat and not angled valve 
margins, valves readily separated from the well- 


developed styles. fruits persistently united. basally 


with pedicels, stigmas with decurrent and connate 
lobes, and often glandular papillae (Al-Shehbaz & 
Yang, 2001). 


tolea are more similar to one another in chromosome 


Solms-laubachia, Desideria, and Chris- 
morphology than they are to Parrya (Yue et al., 2003 
2004). Nevertheless, an extensive sampling of Parrya 
is needed to further test the phylogenetic position of 
the genus. 

and Phaeonychium 


Solms-laubachia, Desideria. 


Jafri form a well-supported clade (E, Figs. J. 2). 


and their close relationship is supported by the 
(Al-Shehbaz. 2000a. 2001). 


Together, these three genera form a clade sister to 


smooth fruit valve 


Christolea. Several morphological synapomorphies 


offer support for the relationship, such as the pres- 
ence of cauline leaves, equal and non-saccate 
sepals, obsolete or distinet style, non-decurrent stigma 
lobes, and wingless seeds. By contrast, in Leiospora, 
Christolea, Solms- 


laubachia, Desideria, and Phaeonychium jafrii. cau- 


which is sister to the clade of 


line leaves are absent, sepals unequal and saccate. 


style absent, stigma lobes strongly decurrent, and 
seeds winged (Al-Shehbaz, 2000a, 2001). 
clade (F, Figs. J. 2 


except for D. batogoinensis, which is embedde d with- 


Species of Desideria form a 


Volume 93, Number 3 Yue et al. 409 
2006 Support for an Expanded Solms-laubachia 


** grandiflorum 
100 F [ — Desideria linearis 
100 175 stewartii 
Desideria himalayensis 
100 Solms-laubachia 1 l 


Solms-laubachia minor 
Solms-laubac i xer e 
Solms-laubachia zhongdianensis 


trj 


Solms-laubachia a herrima 


2 
S 


| 


Solms-laubachia s et. | 

Solms-laubachia euryearpa | 
Solms-laubachia eurycarpa 2 
Solms-laubachia s 


et. 2 


2 
Q 


c 
— Solms-laubachia retropilosa 
L— Solms-laubachia linearifolia 
100 — Desideria baiogoinensis 


BY 


Solms-laubac hig lanata 
100 Phaeonychium 


B Solms-laubac 155 platycarpa 
100 Christolea cr issifoli l 

100 . Leiospora pamirica 

100 — Leiospora exscapa 
Matthiola incana 


Parrya nudicaulis 
Sinapis alba 
Sisymbrium irio 
Fourraea alpina 


100 Thlaspi arvense 


Alliaria petiolata 


Microthlaspi perfoliatum 
100 Cardamine penzesii 
100 Cardamine rivularis 


100 Cardamine amara 


100 Barbarea vulgar 
Ro 


orippa ma 


Arabidopsis thaliana 
Capsella rubella 


© 


Arabis divaricarpa 
Halimolobos perplexa 
100 Crucihimalaya himalaica 
Turritis glabra 
Olimarabidopsis cabulica 
Olimarabidopsis pumila 
Erysimum handel-mazzettii 


Lepidium campestre 
A 100 1 deltoidea 
100 bis blepharophylla 
— iip nia pulvinata 


100 — lonopsidium abulense 


Cochlearia danica 


Figure 2. Majority consensus tree inferred from Bayesian analyses of the combined data set of matK and Chs genes. 
Numbers at branches are posterior probabilities. and letters A-J refer to clades discussed in the text. 


in Solms-laubachia, suggesting that Desideria may chia and D. baiogoinensis or is embedded within 
be polyphyletic and Solms-laubachia paraphyletic. — Solms-laubachia (G, Fig. 2). Therefore, our results 
Phaeonychium jafrii is either sister to a weakly support an expanded Solms-laubachia, i.e., Solms- 


supported clade (G, Fig. I) containing Solms-lauba- — laubachia s.l.. including species of Desideria and P. 


410 


Annals of the 
Missouri Botanical Garden 


Jafrii as well. However, Phaeonychium is a heteroge- 
central 
(Al- 


species 


neous genus of seven species distributed 
Asia and centered in southwestern China 
Shehbaz. 2000a; Zhou et al.. 2001). More 
are needed to further test the phylogenetic relation- 
ships of Phaeonychium before a proposal can be made 
and Desideria. 


with Solms-laubachia 


lo merge i 
Nevertheless, sequences of matK and Chs suggest 


the union of Solms-laubachia and Desideria. 


Within Solms-laubachia, S. retropilosa and S. 
linearifolia are closely related (bootstrap = 74%. 
Fig. I. and posterior probability = 100%, Fig. 2). 
Solms-laubachia minor, S. pulcherrima, S. zhongdia- 


nensis, and two accessions of S. xerophyta form a elade 
1 — 06%, Fig. 
1009 Fig. 2). 


eit ae do not cluster. together, 


|. and posterior probability = 
However, two accessions of S. 
with one (Yue 
0250) from Sichuan and the other (Yue 0251) from 
Yunnan. Further morphological studies are needed to 
whether they might different 


determine represent 


species. Given the low phylogenetic resolution from 


math and Chs data for interspecific relationships of 


Solms-laubachia s.l.. further studies with additional 


data from more rapidly evolving DNA regions are 


warranted, 


Literature Cited 


Al-Shehbaz. L A. 198 
ceae ) in s southeastern United States. J. 
37 


. The tribes of Cruciferae (Brassica- 
\rnold Arbor. 
65: 3-373. 

- 2000a. A revision of the 
pori Nordic J. Bot. 20: 157-163 
2000b. o 5 c, a new genus 
Novon LO: 


view hs gamose perl in the Brassicaceae 


genus Phaeonychium 
from C bina: 
— . 2001. 


a revision ol Desideria, 


an wilh a 


related. genera. . Missouri Bot. Gard. 87: 510-563 
— . 2003 
irabis 1 to Boes hera (Brassicaceae). Novon 13: 381-391. 
2005. Nomenclatural Arabis 


noles on 1 


(Brassicaceae). Novon 15: 519-524. 

> Yang. 2001. revision of Solms-laubachia 
"lisse aceae). Harvi d 1 05 Bot. 5: 37 381. 

—— —, — & K. . 1999, 
spes ies excluded luni DT 1 cee. Novon 
9: 296-307 


eilstein & E. A. Ke logg. 
and hse of the Brassicaceae. 
Evol. ): 89 aros : 

Appel, 0. à I. Al-Shehbaz. 2003. Cruciferae. Pp. 75-174 
in K. 1 (editor), 
Plants, Vol. 5. Springer-Verlag, Berlin, Heidelberg. 

Botschantsev, V. 1955. De 
later. Gerb. Inst. 
160-17 

Farris, J. S.. 
approach to 

191. 


. M. Källersjö, 


2006. Systematics 
an overiew. Pl. Syst. 


Families and Genera of Vascular 


Crueiferis notae. criticae. Bot. 
Komarova Akad. Nauk S.S.S.R. 17: 
G. Kluge & M. J. Eckardt. 


phylogenetic 


197 0. j numerical 

systematics, Syst. Zool. 19; 

. Kluge & C. Buil. 1994. 
315 


Testing 
ET ance of incongruence. Cladistics 10: 310, 


Critical evaluation of 


Transfe "of most North American a cles of 


e placement of 


Fitch, W. M. 1971. 
Minimal change for a specific tree topology. 
106—116. 

Clan G. Ls Rd 


Toward defining the course of o 
Syst. Zool. 2 


1. A. 


nuclear gene 


Price. 1998 
arginine de- 


Molec. Biol. 


Malmberg & 


Phylogenetic utility of the 


9 ase: An example from Brassicaceae. 


Evol. 15: 1312-1320 

Hayek, à 1911. Entwurf. eines Cruciferensystems auf 
phylogenetischer Grundlage. Beih. Bot. Centralbl. 27: 
127-335. 


PR 2001. MRBAYES: 


Bayesian info rence of phylogenetic trees. Bioinformatics. 


17: 75 


Huelsenbeck, Ronquist. 

155% 

. H. 1955. Christolea: With special reference to the 

sin N. W. Himalayas, 

Notes Roy. mus Gard. 

Janchen, E. 1942. Das 

Zei Tu 

Johnson. L. A. & D. E. Soltis. 1994. math DNA sequences 
and phylogenetic reconstruction. in 

ot. 19: 1 


. Pakistan and Afghanistan. 
9-59 


Edinburgh 22: 


System der Cruciferen. Oster. Bot. 


Saxifragaceae sensu 


stricto. Syst. 


Koch, M.. J. Bishop & T. Mitehell-Olds. 1999, Molecular 
systematics of eae and Arabis. Pl. Biol. E: 
— 


aubolc | a 


aie analys 


. Mitehell-Olds. 2000. C 1 
E chalcone s ui 
dehydrogenase Pei in ö 


genera (Brassicaceae). Molec. Biol. Evol. 17: 1483-1408. 
———, — ——— & — ———. 2001. Molecular systematics of 

the Brassicaceae: E Vis nce from coding ph SU math 

0 nuclear ^ sequences. Amer. J. Bot. 88: 534-544. 


2002. 
slemalies, evolution and popu iion. biology 
. Missouri Bot. 


. I. -Shehbaz & K. 1 
Molecular sy 


3 the utd family (Brassicaceae). An 


Diar neue Tibetische Pflanzen. Flora 55: 
Mau. B.. M. 
phylogenetic E erence 85 
methods. Biometrics 55: 1— 
O'Kane, S. L. Jr. ü a4 1 2003. 
position anch generic limits of „ (Brassicaceae 


INA. Ann. 


Newton & B. Larget. 1999. Bayesian 
Meus Chain Monte Carlo 


Y 


Phylogenetic 


ased on sequences of nue ribosomal 
Missouri Bot. Gard. 90: 603-6 12. 

N V. Crandall. 1998. 
DN 


the model of \ 
"T 8. 


MODELTEST: 


substitution. 


Posada. D. & K. Testing 


Bioinformatics 
817-818 
Rannala, B. & Z. H. Yang. 1996. Probability distribution of 
molecular evolutionary. tree new me thod of phyloge- 
311 


E a : 304— 


s 58 1 


ic inference lolec. 


. Engler & 


K. Prantl (e ditors), Die Natürlichen Pflanze e n, Vol. 


17B. Verlag von Wilhelm Engelmann. Leipzig. 
Sorenson, M. D. 1996. Treeltot. University of Michigan. Ann 
rDOT. 
Swofford, D. L. 2002. PAUP*: Phylogenetic analysis using 
parsimony (* and other me kein) Version 4.0. Sinauer, 


Sunderl: ind, Ma issachusel 
Warwick, LA. Al-She 1955 R. X. Price & C. ider. 


2002. legs ny e Sisymbrium (Brassicaceae) di | on 


ITS sequences of nuclear ribosomal DNA. Canad. J. Bot. 
90: 1002-1017 

Q0————, C. A. Sauder, D. F. Murray & K. 
Mummenhoff. 2004. Phylogeny of Smelowskia and related 


ger 


era (Brassicaceae) based on TTS sequences of nuclear 
ribosomal DNA and tnb a of chloroplast DNA. Ann. 
Missouri Bot. Gard, 91: 99-12 


Volume 93, Number 3 
2006 


Yue et al. 411 
Support for an Expanded Solms-laubachia 


Yue. J. P.. H. Sun, L A. Al-Shehbaz € Z. J. Gu. 
2003. 


Cytological studies of five Chinese 


o laubachia (Brassicaceae). Harvard Pap. Bot. 7: 
7-473. 
„J.. J. Gu, L A. Al-Shehbaz & H. Sun. 2004. 
Cytological studies on the Sino-Himalayan endemic 


Solms-laubachia and two related genera (Brassicaceae). 
Bot. J. Linn. Soc. 145: ) 


species of 


Zhou, T. Y., L. L. Lu, G. Yang € 1. A. Al-Shehbaz. 2001. 
Brassicaceae (Cruciferae). Pp. 1-193 in Z. Y. Wu & P. H. 


Raven (editors), Flora of China, Vol. 8. Science p Tess, 
Beijing; Missouri Botanical Garden Press, St. Lot 
Junk, K., K. Mummenhoff, M. Koch & H. ua. 1000. 


Phylogenetic relationships of Thlaspi s.l. (subtribe Thlaspidi- 
nae, Lepidieae) and allied genera based on chloroplast DNA 
92: 375-38 


restriclion-site variation. Theor. Appl. Genet. 


REVISIÓN DEL GÉNERO ACAENA Alicia Marticorena? 
(ROSACEAE) EN CHILE 


RESUMEN 


Se realizó una revisión taxonómica del género Acaena en Chile, utilizando caracteres morfológicos clásicos. Se reconocen 


vel "species para Chile continental y Archipiélago de Juan Fernández, las que se ubican en seis secciones. Los caracteres 
de mayor utilidad ps reconocer las especies son el tipo de inflorescencia, forma de los folíolos, asi como la forma N 
ornamentación de la cupela. Las siguientes especies son lectotipificadas: A. alpina Poepp. ex Walp., A. caespitosa Gillies 
Hook A t 


Arn., A. integerrima Gillies ex Hook. et Arn., A. macrocephala Poe ahy A. magellanica Dad) ahl. A. 0 


Ruiz et Pav., A. platyacantha Speg., A. e Hook. et Arn. y Alboff. Las siguientes especies fueron 
neotipificadas: A. lucida (Aiton ) Vahl, A. sericea J. Jac I. trifida Ruiz el UN var. glabrescens Regel et Korn. Los siguientes 
sinónimos fueron lec ay sane ados: A. argentea Ruiz el [a var. breviscapa Bitter, A. argentea pos et Pav. var. coriacea Bitter, 
. argentea Ruiz et Pav. var. interrupte-pinnata Bitter, A. argentea Ruiz et Pav. Deus Bitter, A. argentea Ruiz et Pav. 
var. subcalvescens Bitter, A. cadilla Hook. f., A. canescens Phil., A. cuneata Hook. « 1 e Phil. A. digitata Phil. var. 
latifoliolata Bitter, A. digitata Phil. var. subpinnata Bitter, A. euacantha Phil. „ A Dc Phil., A. glandulifera Bitter, A 
grandistipula Bitter, A. hirsuta Phil., A. ischnostemon Bitter, A. krausei Phil., A. krausei Phil. subvar. glabratula 8 A. 
longiaristata M. Ross [= A. magellanica (Cam.) Vahl], A. lucida var. villosula Bitter, A. macrostemon Hook. f. subsp. 
longiaristata (M. Ross) Bitter var. basipilosa Bitter, A. macrostemon Hook. f., A. „ (Lam.) Vahl subsp. pygmaea 
Bitter, A. magellanica (Lam.) Vahl var. glabrescens Bitter, A. microcephala Schltdl., A. multifida Hook. f. subsp. intercedens 
Bitter, A. neglecta Bitter, A. petiolulata Phil., A. pinnatifida Ruiz et Pav. subsp. hy 1 Bitter, A. pinnatifida Ruiz et Pa 

subsp. nudiscapa Bitter, A. pinnatifida Ruiz et Pav. var. M inn Bitter, A. quinquefida Phil., A. splendens Hook. et Arn. 
ar. eR te la Bitter, A. splendens Hook. et Arn. ve evisericea Bitter, A. tenutpila Bitter, A. trifida Ruiz et Pav. var. 
argentella Bitter, A. trifida Ruiz et Pav. var. acopio Bitter, A. trifida Ruiz et Pav. var. mollissima Bitter y A. venulosa 


Griseb. b. siguientes sinónimos fueron neotipificados: A. calcitrapa Phil., A. closiana Gay, A. hirta Citerne y A. macrostemon 
Hook. f. subsp. elosiana (Gay) Bitter 


ABSTRACT 
laxonomie revision of the genus Acaena in Chile was carried out, using morphological characters. Twenty species 


\ 
belonging to six sections are recognized for continental Chile and Juan Fernandez Archipelago. The most useful characters for 
recognizing species are int 1 type, the form of leaflets, as well as the form and ornamentation of the fruit. The 


following species are lectotypified: A. alpina Poepp. ex Walp., A. cr Gillies ex Hook. et Arn., A. integerrima Gillies ex 
Hook. et Arn., A. macroc 1 Poepp., A. 1 (Lam.) Vahl, A. nal Ruiz et Pav., A. A Speg., A 
splendens Hook. et Arn., and A. tenera Alboff. The following species were neotypified: A. lucida (Aiton ) Vahl, A. sericea J. 


acq., and A. trifida Ruiz et var. glabrescens Regel et Kirn. The following synonyms were lectotypified: A. argentea Ruiz 
et Pav. var. breviscapa Bitter, A. argentea Ruiz et Pav. var. coriacea Bitter, A. argentea. Ruiz et Pav. var. interrupte- d 
Bitter, A. argentea. Ruiz et Pav. var. lanigera Bitter, A. argentea Ruiz et Pav. var. subcalvescens Bitter, A. cadilla Hook. 
o © 5 


canescens Phil., A. cuneata Hook. et Arn., A. digitata Phil., A. digitata Phil. var. latifoliolata Bitter, A. digitata 11 var. 
subpinnata Bitter, A. euacantha Phil.. A. fuegina Phil., A. glandulifera Bitter, A. grandistipula Bitter, A. hirsuta Phil.. 

ischnostemon Bitter, A. krausei Phil. 2 krausei Phil. subvar. glabratula Die A. longiaristata M. Ross [= A. mage 1 
(Lam.) Vahl], A. lucida var. 1 . macrostemon Wook. f. subsp. longiaristata (H. Ross) Bitter var. basipilosa Bitter. 
|. macrostemon Hook. f., A. mage E 9 5 (Lam.) d subsp. pygmaea Bitter, A. magellanica (Lam.) Vahl var. glabrescens 


Bitter, A. microcephala Schltdl., A. multifida Hook. f. subsp. intercedens Bitter, A. e Bitter, A. petiolulata Phil., A. 
pinnatifida Ruiz et Pav. subsp. hypoleuca Bitter, A. pinnatifida Ruiz et Pa discapa Bitte \ pinnatifida Ruiz et 
av. var. uspallatensis Bitter, A. quinquefida Phil., A. trifida Ruiz et Pav. var. pred er (Phil.) Reic he. 1. splendens Hook. et 
Arn. var. brachyphylla Bitter, A. i en Hook. et Arn. var. brevisericea Bitter, A. tenuipila Bitter, A. trifida Ruiz et Pav. var. 


ue Bitter, A. trifida Ruiz et Pav. var „ Bitter, A. trifida Ruiz et Pav. var. mollissima Bitter. and A. venulosa 
Griseb. The following synonyms were ne 5 : A. calcitrapa Phil., A. elosiana Gay, A. hirta Citerne, and A. macrostemon 
Hook. f. subsp. clostana (Gay) Bitter. 

Key words: Acaena, Chile, Rosaceae, taxonomy. 


! Este estudio fue posible g gracias al financiamiento de la Mellon Foundation, vire eto 91.032.001 -4 y del Proyecto Flora de 
Chile. Deseo agradecer a Nelson Moya por la delicada confección i de las figi Agradezco a Mélica Munoz (SGO), a Edmundo 
Pisano (f) (HIP) y Jorge Victor Crisei (LP) P el prést; cura material de he 1 Además a Pa as las personas que me enviaron 
Imágenes y fotoco jas de mate ‘rial tipo; W. Gre ute er ( Ste inhof f (BREM), M. Dillon (F), F. Jaquemoud (6). E. Lucas y D. 
Zappi (K i 
y 


= 
— 
= 
7 
b 
[ev] 
mmn 
= 
Y 
r 
T 


risci. gui y D. Gutiérrez (LP), F. Se i rk (M), A. Changy, T. Deroin 
RSL). A Roy Gereau y T. Lammers por la com he ta revision del texto y valiosas 

) | I 
sugerencias. En especial agradezco a Victoria Hollowell, Carlos Ae 


= 
Y 
= 
rm 
PP 
* — 
ur 
pA 
z 
a 
2 
N 
E 
— 


„ Kanchi Gandhi y Neil Harriman por su siempre atenta 
colaboración, paciencia y acertados comentarios que mejoraron sustancialmente el manuscrito, principalmente con los 
problemas de tipificación. A mis compañeros de viaje, ya que sin su apoyo no podría haber realizado las salidas a terreno. 

Departamento de Botánica, Facultad de Ciencias Naturales y Oceanograficas, Universidad de Concepción, Casilla 160-C 
Concepción, Chile. amarticGudec.cl 


ANN. Missouri Bor. Garp. 93: 412-454. PUBLISUED ON 24 OctoBER 2006. 


Volume 93, Number 3 


Marticorena 413 


Acaena en Chile 


La familia Rosaceae Adans. posee cerca de 3000 
especies distribuidas en alrededor de 100 géneros 
1988; Mabberley, 1997) los que corres- 


ponden a plantas leñosas y herbáceas, principalmente 


(Kalkman, 


de regiones templadas y Sp duelo del hemisferio 
norte (Cronquist, 1981: 
1964-1967). Se 


alternas, raramente opuestas, simples, compuestas o 


Heywood, : Hutchinson, 


caracteriza por o hojas 


disectas, en general acompañadas de un par de 


Heywood, 1985). 


a menudo dispuestas en inflorescencias, en general 


Las flores son solitarias o 


ps 


estípulas 


son actinomorfas, hermafroditas, con un hipantio 
notorio (Cronquist, 1981; Mabberley, 1997). 
Respecto de la polinización, las flores de las 
Rosáceas son poco especializadas, produciendo. gran 
cantidad de polen, el que atrae a una amplia variedad 
de insectos grandes y pequeños (Jude et al., 1999). 
Mutis ex L. 


Sin embargo, géneros como Acaena 
Poterium L., poseen flores muy reducidas, en parte 
unisexuales, sin pétalos ni néctar, y polinizadas por el 


1985). En 


carácter está la gran diversidad de frutos, 


viento (Heywood, contraste con este 


tales como: 


aquenio, poliaquenio, drupa, polidrupa, cápsula. 


pomo, folículo, entre los más comunes (Spjut, 1994). 


los que entregan importantes características. para 


diferenciarla en subfamilias (Heywood, 1985 


En Chile. la familia está formada por 15 géneros y 
| 8 j 


51 especies (Marticorena, 1991). La mayoría de las 
especies del género Acaena (aproximadamente 43), se 
distribuyen en el hemisferio sur, encontrándose 
también algunos representantes en el hemisferio norte 
como por ejemplo en México, California y Hawaii 
1964-1967). El género pertenece a 


tribu Poterieae Dumort. (Sanguisorbeae DC.) de 


— 


Hutchinson, a 


— 


a 
subfamilia Rosoideae Arn., y se encuentra relacio- 
nado morfológicamente a Sanguisorba Poterium, 
ambos de zonas templadas del hemisferio norte, y 
Polylepis Ruiz et Pav. y 


hemisferio sur, 


a Margyricarpus Ruiz et Pav., 
Tetraglochin Kunze ex Poepp.. del 
todos ellos pertenecientes a la misma tribu, los 
comparten algunos caracteres tales como el tipo de 
fruto rodeado por el tubo del cáliz persistente, que 
puede presentar ornamentaciones y en las flores 
simples. dispuestas en cabezuelas o espigas (Hutch- 
1964-1967; 1984; Zardini, 1973). 


El género Acaena comprende subarbustos y herbá- 


inson. Grondona. 
ceas, generalmente leñosas en la base, decumbentes o 
extendidas, con las ramas floríferas a menudo erectas. 
Sus 


dentadas o aserradas, con estípulas envainadoras en 


— 


rojas son alternas, compuestas, imparipinnadas, 


— 


la base. Las flores son pequefias y están dispuestas en 
capítulos o en espigas, en el ápice del tallo: pueden 


ser verdes, rojas o blancas y estar acompañadas de 


brácteas: el tubo del cáliz es persistente y puede ser 


desnudo, tuberculado o armado con espinas y garfios; 


los pétalos están ausentes y los estambres son 1-10, 
sésiles, insertos en la garganta del cáliz; el ovario es 
perígino, dialicarpelar, el estilo es subterminal, corto, 
con estigma peltado, espatulado o dilatado y fimbriado 
o penicilado. El óvulo es solitario y péndulo, el fruto 
es un aquenio que va incluido en el tubo del cáliz 
endurecido o esponjoso, tuberculado o con cerdas, 
ion que recibe el nombre de cupela (Grondona, 
c 984; Hutchinson, 1964—1907). 


1994), el tipo de fruto de Acaena corresponde a un 


= 


Según Spujt 


diclesium. Este difiere de otros antocarpos secos por 
el pericarpio que está incluido dentro del exocarpo. el 
que usualmente deriva del perianto. En frutos 


derivados de un ovario ínfero, como es el caso de 


Acaena, la presencia de costillas o alas sobre el 
exocarpo es usualmente considerada derivada desde 
un perianto acrescente. Sin embargo. es un nombre 
poco utilizado, y para esta revisión se conserva el 
nombre de eupela, por ser de uso más frecuente en los 
lextos. 

En cuanto a las características de las flores. éstas 
son poco vistosas, con perianto reducido, general- 
mente de color verde, o café oscuro a rojo, o bien 
puede estar ausente o ser deciduo (Faegri & van der 
Pijl, 1971). 


captadores de polen se encuentra incrementada por la 


La efectividad de los estigmas como 
exposición hacia afuera, más allá de las brácteas o el 
perianto. También es importante la exposición de las 
anteras, para que el aire pueda llevar lejos el polen. 
Los estambres se caracterizan por presentar filamen- 
tos largos, dejando a las anteras fuera del perianto. 
Como una condición no morfológica para el éxito de la 
anemofília, está lo que se refiere al número de 
individuos componentes de una población, los cuales 
dependen de cierta masividad de polen incidente 
Pijl, 1971). 


rísticas mencionadas para plantas con polinización 


(Faegri & van der Todas las caracte- 
anemófila, se presentan en el género Acaena, y debido 
a la simplicidad de estas, es que los caracteres de 
mayor utilidad son los reproductivos, en especial los 
caracteres de la cupela, debido a una mayor 
diversidad y riqueza de formas y ornamentaciones. 
Además, son de utilidad la disposición de las flores y 
las cupelas y la forma de la inflorescencia y de la 
infrutescencia. 

Un carácter presente en algunas especies de 
Acaena es la formación de flores cleistógamas. Estas 
se caracterizan por una reducción en el tamaño y 
número de las piezas florales, en especial las anteras, 
y por modificaciones en el perianto (Frankel & Galun, 
1977; 1981). En las 


género, se observa un gradiente de tipos florales, 


Lord, especies chilenas del 


desde chasmógama a cleistógama, dentro de un mismo 


individuo; las flores consideradas cleistógamas no 


presentan una modificación notoria que las haga 


414 


Annals of the 
Missouri Botanical Garden 


claramente distintas de las chasmógamas. lo más 


evidente son la ubicación y protección por las vainas 


foltares. Esta transición de tipos florales indica que la 
cleistogamia es una condición derivada de la 
chasmogamia (Frankel & Galun, 1977). 

Se considera que son factores ambientales, tales 


como temperatura, sequedad. humedad, luminosidad, 


los responsables de la formación de flores cleistóga- 


mas (Uphof, 1938; Lord, 1981). Es probable que 
algunos de estos factores, solos o en conjunto, 


provoquen la formación de flores cleistogamas en 
algunas de las especies presentes en el país. Sin 
embargo, esto no explica la falta de ellas en algunas 


especies que habitan las mismas áreas que aquellas 


que sí poseen flores cleistogamas. Sin embargo, Lor 
(1981) afirma que en condiciones naturales, ambas 
formas, chasmógamas y cleistógamas, aparecen en un 
mismo momento durante el ciclo de vida de la mavoría 
de las plantas. 

En cuanto al período de floración y fructificación se 
observó que no existe una etapa distinta de la otra, 
ambas ocurren casi simultáneamente, por lo cual sólo 
se señala el período de fructificación ya que además 
es el más útil para la identificación. 

Dentro de los caracteres vegetativos, los más útiles 
fueron la forma y disposición de los folíolos. Acaena 
Arn. 


son muy parecidas, pero si no presentan frutos se 


= 


alpina Poepp. ex Walp. y A. splendens Hook. et 
pueden diferenciar debido a que los folfolos de la 
primera se disponen de forma digitada y en A. 


splendens se presentan de manera pinnada. Para otras 


especies en cambio es necesario observar tanto 
caracteres del fruto como de las hojas. 
Otros caracteres vegetalivos cuantitativos como 


longitud del pedúnculo floral, tamaño de las hojas y 


de los folíolos, tamaño de las brácteas basales entre 


los más notorios. son altamente variables en la 


mayoría de las especies, por tanto su utilidad es 


— 


minima. Esto ha provocado que tanto Bitter (1910) 
como Philippi (1862, 1863, 1872, 


un gran número de especies y variedades, generando 


893) describieran 


un gran número de nombres que han sido sinonimi- 


zados (Grondona, 1964: Walton, 1975) 
Ha sido un largo trabajo la revisión de las 
descripciones de los taxa creados por Bitter, y 


reconocer que correspondían a un número muc ho 


menor de especies. Como va se comentó, este autor 
describió cerca de 100 taxa para Chile, lo que refleja 
el estrecho concepto de especie que tenía. No fue 
posible acceder a todos los materiales tipos de los taxa 
descritos por Bitter, debido a que la mayoría de ellos 
el cual fue 
Mundial. Sin 


las evidentes similitudes morfológicas de 


se encontraban en el herbario de Berlín, 


destruído en la Segunda Guerra 


embargo, 


dichos taxa, reconocidas mediante las completas y 


exactas descripciones realizadas por Bitter, además de 


a localización de los lugares de colecta en áreas 


donde sólo crece determinada especie, permiten 
considerar que es altamente probable que dichos taxa 


principalmente infraespecíficos, se refieren a la 
especie bajo la cual se encuentran. 

Simpson (1979), 
género Polylepis realizada por Bitter (1911), 


refiriéndose a la revisión del 
señala 
que el concepto tipológico del autor y su tratamiento 
de individuos más que de poblaciones, dejó una 
fragmentación taxonómica que hace imposible identi- 
ficar nuevas colectas. Obviamente esta situación es la 
misma para el género Acaena. 

La especie más citada por la literatura es Acaena 
magellanica (Lam.) Vahl, ya que presenta un gran 


variabilidad morfológica. además de anormalidades 


como desarrollo de capítulos subsidiarios, flores 
pediceladas, capítulos incompletos, poca elongación 
del escapo y de los entrenudos lo que produce hojas 
1979 


Uno de los mayores problemas fue determinar a qué 


en roseta (Walton. 


— 


especie correspondían algunos individuos. debido a la 


combinación de caracteres que presentaban. Se 
observaron individuos que compartían caracteres de 
Acaena magellanica y A. ovalifolia Ruiz et Pav., de los 
cuales la mayoría proviene de la XII Región, donde 
ambas especies son muy abundantes. Á pesar de ésto, 
Grondona (1964) y Walton y Greene (1971) con- 
sideran que estas especies, aunque habitan juntas, 


aparentemente no hibridizan, como lo reafirma Walton 


— 
— 
— 


zn general, la variación morfológica de una 


Pn está ligada a la plasticidad fenotípica o a la 


iversidad genética (Briggs & Walters, 1984), luego 


se considera necesario realizar un estudio más 


profundo entre las especies A. magellanica y A. 
ovalifolia que habitan en Chile, analizando previa- 


mente el complejo A. magellanica, utilizando in- 


formación de datos moleculares para verificar la 


existencia o no de divisiones infraespecíficas reco- 
nocibles, o si sólo corresponde a una gran variación 
morfológica debida a plasticidad fenotípica y con esto 
causas de la variabilidad 


también establecer las 


presente en algunas especies. 


= 
= 


Marticorena (2000) reconoció patrones celulares 
en la epidermis de la cupela, y determinó tres tipos 


básicos, los que se corresponden con las secciones 


y subgéneros a las que pertenecen las especies, 
niveles taxonómicos que han sido circunscritos sobre 
la base de caracteres morfológicos tradicionales, \ 
que fueron denominados como acaena, ancistrum y 


axillares. 


Volume 93, Number 3 Marticorena 415 
2006 Acaena en Chile 
DISTRIBUCIÓN GEOGRÁFICA nümero de especies en la zona sur de Chile. 


La familia Rosaceae presenta una distribución 
1985), 


mente a climas templados. Se la considera originaria 


cosmopolita (Heywood, restringida esencial- 
de Laurasia y que luego llegó a Sudamérica, a Malasia, 


desde allí a la región Pacífico-Australiana, y por 
último a Africa, esto unido a su predominancia en 
Europa, Norteamérica y Asia continental (Kalkman. 


1988). 


hipótesis más aceptada es la que considera su 


Respecto del género Acaena, al parecer la 


evolución a partir de una Sanguisorbeae primitiva, 


Sudamérica | relativamente 


que llegó al sur de 
temprano, mediante dispersión a larga distancia, y 
que luego desde allí se dispersó ampliamente a la 


1974), 
América Central y California con 3 


región Antártica y al norte (Raven & Axelrod, 
hasta alcanzar 
especies (Moore, 1972). 

La dispersión de plantas insulares de Acaena es 
indudablemente debida a las aves marinas. mientras 
que la abundancia y amplia distribución de las 
especies continentales, en especial de Sudamérica, es 
debida a la adhesión en las plumas de aves, al pelo de 


los animales E tales como viscachas (Gron- 


dona. 1964). y también a pul doméstico (Johow. 
1896: ee 1954) y a la vestimenta (Ridley, 
1930). La teoría de que el género se desarrolló en el 


hemisferio sur, está fundamentada por el número de 
laxa que se presentan mayoritariamente en el sur de 
Sudamérica, con aproximadamente 22 
1964, 1984), 
cies, Macmillan, 1988), Australia (5 especies, Beadle 
et al., 1982). 
Central y México, y una especie en Hawaii (Acaena 
californica Bitter), 


y en islas subantárti- 


especies 
(Grondona, Nueva Zelandia (16 espe- 


Luego, un menor número en América 


exigua A. Gray), California (4. 


Sudáfrica (A. latebrosa Aiton), 
South 


Georgia, Macquaire v 


1964: 


cas, e.g. Kerguelen, 
Falkland (A. magellanica) (Greene, 
1968; Walton, 1979, Pisano, 1984). 


En Chile, el género se distribuye ampliamente. 


Moore. 


ocupando todo el territorio, con un mayor número de 


taxa presente desde la V a la XII Región. Se las puede 


encontrar en variados ambientes, que van desde 
lugares secos hasta húmedos, entre rocas, suelos 


arenosos o en vegas. En Chile insular, el Archipiélago 
de Juan Fernández posee tres especies, una endémica. 
Chile 


A. argentea Ruiz et Pav. y A. ovalifolia. 


Acaena masafuerana Bitter, y dos nativas de 
continental, 
Estas últimas se comportan como malezas muy 
agresivas, cubriendo terrenos de pastoreo, praderas 
en las costas, claros en el bosque, terrenos húmedos 
a orillas de los esteros (Johow, 1896: Matthei, 1995; 
1897; Skottsberg, 1954) 


En general las 


Reiche. 


especies siguen la distribución 


impuesta por la diagonal árida, presentando un mayor 


comenzando desde el sur-este de Argentina, er 


especial la zona patagónica, cruzando los Andes en 
sentido noroeste. 
HISTORIA DEL GÉNERO 


El género Acaena fue propuesto por Mutis en 1771 
200) 


material 


válidamente publicado por Linnaeus (1771: 


basado en la especie A. elongata L.. 


proveniente de México según Linnaeus, pero en 
realidad de Colombia 


1771). 


nuevas 


(introducción de Stearn, 
Ruiz y Pavón (1798-1802) descri- 


especies de las 


Linnaeus, 
bieron 6 cuales 4 se 


encuentran en Chile. El primero en subdividir el 
género fue de Candolle (1825), quien reconoció dos 
secciones que denominó Euacaena DC. y Ancistrum 
(J. R. Forst. et G. Forst.) DC., 


posición de las espinas en la cupela. Este último 


basándose en la 


nombre fue tomado del género Ancistrum publicado 


. R. Forster y G. Forster en 1775, el que Vahl 
(1804-1805 
(1847) incluyó 12 


por 


consideró sinónimo de Acaena. 


— 


Gay 
especies en la sección Acaena y 9 

Ancistrum, describiendo dos nuevas especies, una 
para cada sección. Sin embargo este trabajo entrega 
sólo descripciones, con un breve comentario sobre 


distribución, siendo algunas descripciones sólo copia 


de la original. También hace referencia a las 
similitudes entre las especies y comenta sobre las 
características que permiten reconocerlas de otras 


I looker 


sin considerar 


afines. (1844-1847), reconoció 10 especies 


secciones y describió 3 especies 
nuevas, dos de las cuales son citadas para Chile. 


Weddell (1855-1861) utilizó las secciones 


Candolle y a su vez las subdividió basándose 


de de 
en el 
tipo de inflorescencia, ya sea espiciforme o capituli- 
forme, 
Chile. 


descripciones, con algunas observaciones sobre las 


donde 11 de las 16 especies son citadas para 


En este trabajo sólo se entregan breves 


similitudes entre especies. Philippi entre los años 
85 


357 y 1893 describió 22 especies para Chile, de las 
cuales varias fueron basadas sobre material incom- 


pleto, ya sea sin flores ni frutos o con flores 


mismo consideraba 


inmaduras, caracteres que él 


importantes para el reconocimiento de las especies, 
lo que de este modo hace que las especies descritas 
Citerne (1897) hizo un 


que 


sean de un valor dudoso. 


estudio del género Acaena en el incluyó 


descripciones de los órganos vegetativos y reproduc- 
tivos, morfología y anatomía de los frutos y semillas. 


También se refirió a las afinidades con algunos 


géneros como Polylepis, Sanguisorba y Poterium, y 


presentó un esquema de clasificación donde dividió al 
basándose en el tipo de 


género en 7 secciones 


inflorescencia, la sea axilar o terminal de las 


Annals of the 
Missouri Botanical Garden 


ramas floríferas y la posición de las espinas en la 


cupela; estas secclones son: Acaena (como Pleurosta- 


chya Citerne), Lachnodia Citerne, Brachycephala 


Citerne, Pleurocephala Citerne, Acrostachya Citerne, 


Acrocephala Citerne, Anoplocephala Citerne. Reiche 


(1897) utilizó las secciones publicadas por de Candolle 


(1825), pero las elevó al rango de subgénero; de tal 


reconoció en el subgénero Acaena (como 
“Fuacaena”) 15 especies y en Ancistrum 10 especies, 
considerando como carácter principal el tipo de 


inflorescencia y como segundo carácter la posición 


de las espinas en la cupela, siendo el primero en 
confeccionar una clave para el reconocimiento de las 
24 especies reconocidas para el país. Bitter (1910) 
realizó una monografía del género aceptando algunas 
secciones creadas por Citerne y dividió al género en 
dos grandes grupos, Axillares y Terminales, los cuales 
el subdividió considerando numerosos otros subgrupos 
dentro de Acaena. No obstante, la mayoría presenta 
límites poco claros, donde especies que pertenecían 


a distintas subgrupos (secciones según Bitter) son 


sinónimos de una sola especie, e.g. Subantarcticae, 
Dolichanteroideae o Glaucophyllae, que en términos 
generales sólo comprende variantes de A. magellanica. 
En el presente trabajo sólo se consideró la división del 
género hasta sección. Del tratamiento que hace Bitter 
para el género, se reconoce que alrededor de 100 serían 
los taxa presentes en Chile, distribuídos en aproxima- 
damente 60 especies con subespecies y variedades. 
cuales fueron sinonimizadas 
rondona (1964) y Walton (1975). 


consideraron que Bitter utilizó caracteres extremada- 


muchas de las or 


~ 
— 


Estos autores 


mente variables e inadecuados para la diferenciación 
taxonómica de los taxa infraespecificos, tales como la 
pilosidad de los tallos, número y tamaño de los folíolos, 
longitud del pedúnculo, o que sólo utilizó poblaciones 
femeninas, las que correspondían a especies ya 
descritas, como es el caso de A. magellanica. 

Grondona (1964) hizo una revisión del género para 
donde l4 de las 20 


compartidas también con Chile. Fue el primero en 


Argentina, especies, son 


incluir cortes anatómicos de frutos, que desde el 
punto de vista taxonómico no fueron una herramienta 
útil para la identificación de especies, debido a que el 
número de aquenios dentro de las cupelas puede 
variar de uno a cuatro en una misma planta, esto 
unido a que las secciones transversales de las eupelas 
pueden presentar distintas formas, también dentro de 


una misma planta. Este trabajo se caracteriza por 
reducir drásticamente el nümero de especies recono- 
cidas por Bitter y algunas de Philippi y Gay. En 1984, 
Grondona, reconoce a las especies Acaena antarctica 


Hook. f. y A. Alboff 1964 


considerado sinónimos o no había podido identificar- 


tenera que en había 


las con precisión. Walton (1975 


— 


completó la va larga 


lista de sinónimos para las especies A. magellanica 
con 28 


especies, y A. ovalifolia, 


antarctica 9 
A. lucida (Aiton) Vahl y 


pumila Vahl, con una especie 


especies sinonimizadas, 4. 


cada una. El autor se 
basó en material tipo y en la descripción original, y 
concordó con el amplio concepto de especie adoptado 
por Grondona (1964). El catálogo de la flora vascular 
de Chile (Marticorena & Quezada, 1985) está basado 
en la opinión que se han formado los autores a través 
de la literatura y su experiencia. En él señalan que 
existen géneros que no han sido revisados desde hace 
mucho tiempo, algunos de los cuales presentan una 
nomenclatura caótica y que luego de una revisión 


pueden cambiar, dentro de los cuales se puede 
mencionar a Acaena. Luego en un trabajo estadístico 


1991). 


se entregan valores expresados en número de taxa, de 


sobre la flora vascular de Chile (Marticorena. 


géneros, de especies, y para Acaena se listan 22 taxa 


que corresponden alrededor del 40% de la familia en 
el país y que lo constituye en el género con mayor 
número de representantes dentro de ella. Los trabajos 


antes indicados han generado gran cantidad de 


información sobre el género: sin embargo fueron 


realizados, ya sea para áreas que presentan algunas 
también crecen en 


especies Chile (Grondona. 


1964, 


que 
1984) o son muy extensos, con gran cantidad de 
basados en 


han sido 


taxa que caracteres poco 

confiables por su variabilidad. 

TAXONOMIA 

Acaena Mutis ex L., Mant. pl. 2: 145, 200. 1771. 
TIPO: Acaena elongata L. 

M R. Forst. et G. Forst., Char, Gen. 3, tab. 2. 1775. 
TIC 12 9 9 R. 11 0 et G. 5 Posh 
= 1 775 anserinifolia (J. R. Forst. et G. Forst.) J. 
Armstr.] 


Hierbas perennes, generalmente leñosas en la base. 


Hojas alternas, imparipinnadas, con vainas foliares en 


a base, a veces acompañadas por apéndices estipu- 


lares de forma variable; folíolos enteros, incisos, 


aserrados o pinnatifidos, sésiles. Ramas floríferas 


terminales o axilares: di 'scencia capituliforme o 


espiciforme. Flores her algunas pistiladas, 


chasmógamas o 5 hipantio profundamente 
cóncavo, angosto en el ápice, con un fino reborde donde 
se disponen las piezas florales; sépalos (4) 5 (6). 
ovados a 


persistentes, lanceolados u oblongos 


l 
orbiculares; pétalos ausentes; estambres 2-6, anteras 
púrpuras o amarillentas: estilo con estigma oblongo o 
lanceolado u oe ‘ular, 


laciniado. Cupela ovoide 
oblongoide a obeoni | 


el abra a serícea, lisa o 


costillada, con o sin espinas largo variable, que 


pueden o no terminar en gloquidios; aquenio 1-2. 


Volume 93, Number 3 Marticorena 417 
2006 Acaena en Chile 


CLAVE PARA LAS SECCIONES DEL GENERO ACAENA EN CHILE 


la. Inflorescencia Hee ete globosa, multiflora; cups ‘las | con 2—4 espinas en el ápice. 
2a. Cupelas olx , pubescentes, no glandula spinas más largas que el ovario; epidermis de la e ipe sla con 
límite ue » van ME eff y ecc. Ancistrum 
2b. Cupelas ovoides, con tricomas EE: ss o glándulas sésiles en la superficie; espinas de e tamaño o 
más cortas que el ovario; epidermis de la cupela con límite celular hundido . . . . . . . . . Secc. Acrobyssinae 
lb. Inflorescencia espiciforme, iden 0 alargada, o también de capítulos paucifloros; cupelas con 3 a numerosas 
espinas, con espinas rudimentarias o sin espinas. 
Za. Cupelas sin espinas o con espinas Ue snlarias de aspecto giboso—-—— .... Secc. Pleurocephala 
3b. C Cupe ‘las con ye notorias, no gibosas. 
: ela con numerosas espinas cortas, de igual largo, con gloquidios gruesos; cara inferior de la hoja con 
células paj pilos NS ONO Secc. Subt La m 
4b. Cupe la con espinas de distinto largo, con y sin gloquidios; cara inferior de la hoja sin células papilos 
Ha. Cupela con 3-4 espinas apicales, sin gloquidios; células epidérmicas de la cupela con el límite 
celular acanalado, pare s antic D ondulada, pared periclinal convexa ... .... Secc. Patagonicae 
5b. Cupela con más de 4 espinas de largo variable irregularmente dispuestas en la supe: dici ‘ie, con y sin 
gloquidios; células epidérmicas de la cupela con el límite celular levantado, grueso, s 
anticlinal recta a curva, pared perielinal plana 2... — in iũmͥ .. Secc. Acaena 
l. Acaena secc. Ancistrum (J. R. Forst. et C. Forst.) Inflorescencia al principio  espiciforme y 
DC., Prodr. 2: 592. 1825. TIPO: Ancistrum reunida, alargada en estado fructífero. Cupela 
anserinifolium J. R. Forst. et G. Forst., Char. ovoide, glabra, totalmente cubierta de espinas cortas, 
Gen. Pl. 3, tab. 2. 1775. [= Acaena anserinifolia de igual tamaño, con gloquidios. Hojas glabras, sin 
(J. R. Forst. et G. Forst.) J. Armstr.] estipulas; cara inferior del folíolo con células 


: : yapilosas. 
Inflorescencia capituliforme, globosa, densa. Cu- Pie l l 
pela obeónica, pubescente, con 2—4 espinas largas Especies en Chile: A. pumila. 


con gloquidios en el ápice. Hojas con estipulas. 


Especies en Chile: A. argentea, A. magellanica, A. 5. Acaena secc. Patagonicae A. E. Martic., Novon 
ovalifolia. 9: 227. 1999, TIPO: Acaena patagonica A. E. 
Martic. 
* - VE. RES v . » . . : . . 
2. Acaena secc. Acrobyssinae Bitter, Biblioth. Bot. Inflorescencia espiciforme corta, alargada en es- 
FITA). C ( TIPO. 2 e ics ‘ ^ » ; 
17(74): 49. 1910. TIPO: Acaena acrobyssina — tado fructífero, con glomérulos dispersos a lo largo del 
Bitter. pedúnculo. Cupela ovoide, glabra a pilosa, con 34 
espinas apicales, sin gloquidios. Hojas pilosas, sin 
Acaena secc. d Citerne, Rev. Nat. Oues pues Bod Jas | E 
7(2): 39. 1897. TIPO: Acaena UE WA Schltdl. 3 estípulas. 
Acaena antarctica Hook. f.] Especies en Chile: A. patagonica. 


Inflorescencia | capituliforme, pequeña. Cupela 
ovoide, con tricomas glandulares o glándulas sésiles ERE E 
mE 1 2 j 6. Acaena secc. Ácaena, Mant. Pl. 2: 145. 1771. 
esféricas en la superficie, con 4 espinas cortas con Nm 
e E . y MPO: Acaena elongata L. 
gloquidios, en el ápice. Hojas con estípulas. 


Especies en Chile: A. antarctica, A. masafuerana, Inflorescencia generalmente espiciforme, elon- 
A. tenera. gada o subglobosa, raro de capítulos globosos. 
Cupela ovoide, obcónica o piriforme, general- 


; "e mente pubescente y con numerosas espinas en 
3. Acaena secc. Pleurocephala Citerne, Rev. Sci. l a p 


Nat. Ouest 7(2): 39. 1897. TIPO: Acaena lucida 
(Aiton) Vahl. 


toda la superficie, la mayoría de las veces dis- 


puestas en costillas notorias, con o sin gloquidios. 


Hojas en general sin estípulas o con estípulas 

Inflorescencia en capítulo paucifloro. Cupela rudimentarias. 
ovoide, con tricomas lanosos, sin espinas o algunas Especies en Chile: A. alpina, A. caespitosa Calles 
rudimentarias. Hojas sin estípulas. ex Hook. et Arn., A. integerrima Gillies ex Hook. et 
Especies en Chile: A. lucida. Arn., A. leptacantha Phil., A. macrocephala Poepp., A. 


pinnatifida Ruiz et Pav., A. platyacantha Speg.. A. 
4. Acaena secc. Subtuspapillosae Bitter, Biblioth. poeppigiana Gay, A. sericea J. Jacq., A. splendens, A. 


Bot. 17(74): 38. 1910. TIPO: Acaena pumila Vahl. trifida. 


Annals of the 
Missouri Botanical Garden 


CLAVE PARA Las ESPECIES DE ACAENA EN CHILE 
la. 


Inflorescencia e e esférica, eupela obconica, ovoide u obovoide, con 2-4 (raro más) espinas acieulares, 
de base angosta, apicales, eot 
2. 


Za. 


1 fuertes gloquidio 
Inflorescencias e infrute sscencia de más de 6.5 mm de diámetro, cupela obconica, pubescente, 
glandulares; espinas más largas que el ovario. 

C 


Ba. 


sin tricomas 


upela con tricomas tiesos, caedizos, de color amarillo pálido, con 2 (raro 3) espinas: cabezuelas rojizas: 
folíolos oblongo-elípticos a e npa 100S Lacs eu Gabe OPER PuCRESPCSS Qd Say Eg qa 11. A. ovalifolia 
stentes, blancos, con 4 (raro 3 6 5) espinas: 
amarillas; foltolos linear-lanceolados u obovados. 
da. 


— 


upela con tricomas blandos, persis cabezuelas nande 0 


Folrolos linear-lanceolados, cara superior glabra, verde oscuro, eara inferior sericea, plateada 


A. argente: 
lb. Folfolos obovados, cuneados 


w 


. caras superior e inferior glabras a pubescentes, verde claro 


ge ellaulé ‘a 


"mmu EM A. 
2b. Inflorescencias e infrutescencia menores o iguales a 6.5 mm de diámetro, cupela ovoide u obovoide. 
Iricomas los o glándulas sésiles esféricas; espinas casi del largo del ovario. 

Ca 


Sa. pela con glándulas sésiles esféricas en la 


superficie del ovario, con un mechón de tricomas 
al: an ares en la base de los folfolos, sin tricomas glandulares entre las cupelas . . . . . . . tenera 
5b. Cupela sin glándulas sésiles « vas en la superficie del ovario, sin un mechón de tricomas 1 
en ja base de los folíolos, con tricomas glandulares entre 


seri 


en la Jase de las cupe ‘las 


D 


Pedúneulo glabro o con pocos tricomas esparcidos; generalmente espinas más cortas que la mitad 
del largo de la cupela; habita sólo en Más Afuera . . . . . . . . . . . . . . . . . .. . A. masafuerana 
Ob. Pedünculo pubescente a glabro; generalmente espinas más largas que la mitad del | 
cupela: habita al sur del continente Sudamericano 2 
Ib. Inflorescencia espiciforme, corta o larga, cupel 


argo de la 
a ovoide, oblongoide o piriforme, con 3-4 espinas de Dur ancha, 
dispuestas en el tercio supe rior y otras de menor tamaño entre las mayores, o totalmente cubierta de espinas de 


2. A. antarctica 


tamaño similar, con o sin gloquidios, o sin espinas. 
Ta. Cupela sin espinas o con espinas rudimentarias m iiwmwm i²m¹:ꝛi ee 7. A. lucida 
Tb. Cupela con espinas notorias. 

O 

od. 


E sericeas, pecíolo más largo que la lami 
Oa.  Folfolos con 3 pares de dientes grandes por lado, base cuneada 
9b. 19 5 enteros o incisos, bi 
10a. Cupela de 7-15 mm de largo, con espinas rojizas, blandas: 
,, ⁰⁰ se eras 
10b. Cupela de 2.5-12 mm de 


— 


e 17. A. sericea 


se no cuneada. 

raquis casi nulo, ue de 

. A. alpina 
argo. con espinas café-claro, duras: raquis notorio, i is sin 


Plantas de no más de 13 em de altura, folíolos partidas, cupela glabresce nile, con 
numerosas costillas notorias, totalmente recorridas 


por espinas cortas con fuertes 
elo rr. ³ð ⁰⁰ . caespitosa 
Plantas de más de 13 em de altura, folíolos enteros o con algunos dientes pequeños, 


upela o sin costillas nol 


110. 


orias. 


12a olíolos de color verde e la aro, enteros o con hasta 22 dientes pequeños: cupela 
EE PSOE La T" "T" 18. A. splendens 
12b. Folíolos de color gris plateado, enteros o tridentados en el 


apio e; cupela 
jvordeselobOSHA -u ied acce y ESO 


A. integer rima 


Hojas glabras o pubescentes, pecíolo de igual largo o menor que la lámina. 
3a. Folíolos orbiculares, lobulados, glabros . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16. A. pumila 
3b. Folíolos aserrados o 5 en el m 

Ida. F olíolos pinnatise clos, con 5-7 pi de segmentos linear-lanceolados 


con pe efolo notorio, sin ap e ndice ss esti pulare 8: CUpe Jn as dispu slas en una 
é l: 


] 
DTI A. pinnatifida 
ósiles, con apéndices e stipu lares de pap uos o rudimentarios: cupelas 


5 stas en una cabezuela o espiga corta, con espinas flexibles. 
loa. Inflorescencia en espiga contraída, esférica, de 1: i45 m de diámetro, 
cupelas con espinas pubescentes, con tricomas largos, 5 elabras al 


apie €; sin gloquidios: ape “ndice 288 € ene) ires rudime ntarios 


E 
8. A. macroc niet 
lob. Inflorescencia en espiga corta st 


. no contraída, de 2—4 em de largo. que 
as con espinas glabras y sin gloquidios o espinas con 
tricomas retrorsos y gloquidios: apéndices estipulares desarroll: i" 


alarga al madurar: eupel 


A. leptacantha 
l4b. Folrolos 5 a tripartidos 


ra. Foltolos de hasta 3.5 5 mm de largo, segmentos de los folíolos oblongo-e 9 os, ápice 
1 0, margen no revolulo i mmꝓe(i i;mmvmemq p iv . A. poeppigiana 
Vib. 


Folíolos la mayoría mayores de 3.5 mm de largo. segmentos de los isis line ares 
a lanceolados, ápice agudo, margen revoluto. 


Za. Cupela con 1-2 espinas por costilla, como alas aplastadas lateralmente, 
triangulares, cortas; superficie rugosa entre las cada costilla . . . . . . . . . . 


Volume 93, Number 3 Marticoren 419 
2006 Acaena en Qm 


| mm 


Fig vena alpina. —A. Hábito. —B. Cupela (A. Marticorena et al. 233, CONC). Acaena antarctica. —C. Hábito. 
p ve lupe : E et al. 0855, CONC). 


18b. Cupela con SEM más o menos cónicas o subaciculares, sin superficie rugosa 
entre las espin 
19a. Plantas grác ee cupelas de 1.5-2.5 mm, dispuestas en una espiga corta 
oblongoide, de hasta 6.5 mm de diámetro; costillas 3—4, con una espina 
subacicular por costilla, dispuesta en el tercio superior, raro una espina 
rudimentaria bajo ella u hör ͤ ͥ ͥ[ [mMͥ mr 12. A. „ 
19b. Plantas robustas; cupelas de 2.5-6 mm, dispuestas en una espiga corta, 
redondeada, de más de 8 mm gi dm pedúnculo floral con 15 
dispersas a lo largo, costillas generalmente 5, con 2 espinas cónicas, de 
tamaño similar por costilla .. ̃mũ ſT ooo... 20. A. trifida 
l. Acaena alpina Poepp. ex Walp., Nov. Actorum VV a fa 70 5 862. 
, , , x C nvernada de Maule,” Volckmann s.n. 
Acad. Caes. Leop.- PUE Nat. Cur. 19, Suppl. 1: Üéctotibucspui deco SCO 40861): 
326. 1843. TIPO: “Chile boreal inter fragmentas Acaena digitata var. latifoliolata Bitter, Biblioth. Bot. 17(74): 
syenitica ad nives perpetuas. Estero del Peñón 89. 1910. TIPO: Chile. “Cordillera de Maule, 1855, Ph. 
rajado. Cordillera de la Rosa. Debr. 1827," Germain s.n." (lectotipo, ae . W-110760!; 
duplicado, G, foto 27423 en CONC). 


> m : e a a JI las 
PDpUg + >C sig , Wl, las 
Poeppig 517 (lectotipo, aquí designado, WI, las Acaena digitata var. subpinnata n ie blioth. Bot. 17(74): 


dos superiores). Figura LA, B. 90. 1910. TIPO: Chile. “Cordill. de Baucagna [sic]. 


= 


420 


Annals of the 
Missouri Botanical Garden 


1856-1857, e 
nado, W-112 


id s.n." (lectotipo, aquí desig- 


Planta de 8-30 em de alto; rizoma 4-5.5 mm de 
diámetro; tallos suberectos; entrenudos cortos, cubier- 
tos por las vainas foliares. Hojas 2-7 em de largo; 2 
a 3 pares de folíolos, 1” y 2 par de folfolos separados 


por menos de 2 mm, raquis casi nulo, dándole un 


aspecto palmado o digitado, obovado-lanceolados, 21— 


aserrado en el tercio 


26 mm de largo, el margen 


superior, seríceos; vainas foliares con el dorso seríceo 


y la cara interna glabra; apéndices estipulares 


ausentes. Rama florífera terminal; pedúnculo de hasta 


27 em de largo, piloso; inflorescencia espiciforme, 


abierta, laxa, 4.5—10 em de largo, con alrededor de 6— 


10 flores; 


margen irregular. Sépalos 5, ovado-oblongos, 4—5 mm 


brácteas basales lanceoladas a lineares, el 


de largo, la cara interna glabra, la cara externa 


pubescente con tricomas blancos, blandos, ascen- 


3—5, 


largo, anteras orbiculares, 


dentes; estambres filamentos de ca. 3 mm de 


1-1.5 mm de largo; estilo 


corto; estigma globoso, laciniado, ca. | mm de 


diámetro. Cupela ovoide-oblongoide, 7-15 mm de 


largo, sericea, totalmente cubierta por tricomas 


blancos, lanosos, que contrastan con las espinas 


blandas de color café rojizo, de tamaño similar, de 


2-7 mm, con tricomas retrorsos esparcidos; aquenio J. 


Icones. Grondona, 1964: fig. 52: 1984: 438. 
Fenología. | Fructifica desde noviembre (Grandjot 


(Zoellner 6476). 


Distribución y habitat. Especie nativa que tambi- 


s.n.) a abril 


habita 


én crece en Argentina, n terrenos duros y 
secos, en los faldeos de cerros, desde la provincia de 
los Andes (Laguna del Inca) a la provincia de Malleco 
(Lonquimay), 3150 m de altitud 
(Fig. 11A). 
Nombre comun. 


cadillo (Ja/ffuel s.n.). 


Caracteres distintivos. 


entre 350 y 
Cepacaballo (Hollermayer 712), 


Primer y segundo par de fo- 
líolos parten casi del mismo punto, dando el aspecto 
de hoja palmada. Cupela lanosa, oblongoide, con 
espinas café rojizas. 


Acaena alpina pertenece a la Sección Acaena. Es- 


a especie es muy semejante a A. splendens, 
especialmente. en la pubescencia y forma de los 
foliolos, por lo que frecuentemente se confunden en 
estado vegetativo. No obstante, posee hojas de aspecto 
palmada y sus cupelas son de mayor tamaño, con 
espinas de color marrón-rojizas. Se le encuentra entre 
los 350 3150 m de altitud, 


aunque su mayor 


abundancia va desde los 2000 a 2800 m. en zonas 
cordilleranas de Chile central. 


Bitter (1910) creó la subsección Splendentes, en la 


cual incluyó a Acaena alpina, A. splendens y A. 


integerrima, de acuerdo a que presentan pubescencia 
serícea, sus cupelas son ovoides a fusiformes, con dos 
a un aquenio, y los folíolos con el ápice inciso, con 
pocos dientes, o sin ellos. 

Respecto de la tipificación, se eligió como lectotipo 
ma carpeta depositada en W, la que presenta cinco 
plantas, de las cuales las P» superiores corresponden 
a A. alpina, y las demás a A. splendens. Posteriormente 
Bitter incluyó. etiquetas aclaratorias y señala “diese 
beiden kleinen Pflanzen sind A. digitata Phil. die 
grossen: A. splendens M. et A. 1909 Bitter,” 


refiere a A. alpina como A. digitata, un sinónimo. Se 


donde se 
eligió esta carpeta para lectotipificar, a pesar de la 
mezcla de material, debido a que corresponde con la 
y que además la etiqueta 
Philippi. (1862) 


digitata, y señala en el protólogo varios 


información del protólogo, 
está escrita a mano por Poeppig. 
deseribió A. 
lugares, los cuales corresponden con dos carpetas. 
una colectada por Volekmann en Invernada Maule, y 
la otra por Germain, de la Cordillera de Linares. Se 
eligió como lectotipo la carpeta 5GO-49801 2. 
que presenta una espiga más desarrollada. Para la 
variedad latifoliolata Bitter, se estudió material de W 
yv G, y se seleccionó como lectotipo la carpeta W- 
110760, que posee una planta bien desarrollada, con 
las cupelas casi completamente maduras. El protólogo 
Bilter, 


depositado en 


de la variedad subpinnata menciona un 
271 


presumiblemente destruido, y un material de Germain 


— 


material de Philippi y 
en W-112652, el cual fue elegido como lectotipo, que 
es una planta con las cupelas bien desarrolladas. El 
tipo de Acaena digitata Phil. var. pleodactyla Bitter, 
fue destruido en Berlín, y no existen isotipos. Sin 
tipo en CONC (N 


. alpina. 


embargo, he podido ver una foto del t 


17998), que corresponde con A 


Material representativo examinado. CHILE. Y Región: 
Los Andes, Laguna del Inca, Arroyo 81277 (( 


> 
I 
Pa 
N 
c 
X 
zi 
8 
= 
= 
2 
2 
= 


ee Cajón del Maipo, Baños 
N 


(CONC): Maipo. Cordillera de Aculeo, 15 Ene. 1955. 
Germain s.n. (SGO). VI 0 

Flaco, ene 6476 (CONC). VH. Región: 
de Teno, C. Marticorena et a 22 (G ONC): T 


Maule, M de la laguna, A. Marticorena et al. 233 

(CONC). VIH Región: Ñuble, Camino a Te armas de Chillán, 

15 Ene. 1950. Barros s.n. pr X Región: Malleco, 

Lonquimay, Hollermayer 7 712 2 (C ONC). 

2. Acaena antarctica Hook. f., Fl. Antarct. 1: 269. 
1846. TIPO: Chile: "Hermite Island, , cape Horn: 
on the mountains, rare, 1000 feet.” A. Menzies 


n. (holotipo, K!). Figura 1C, D. 


=~ 


Acaena e Schltdl., Linnaea 28: 403. 1857, TIPO 
Chile. "Alia specimina fructifera lecta sunt Cordillera 


Volume 93, Number 3 
2006 


Marticorena 
Acaena en Chile 


1854," Lechler 2951 (lecto- 
tipo, aquí de 1 G, foto 27425 en CONC! 

Chile. 84: 626. 1893. 
naea 33: 67 non Acaena 
JP . * ahl). 1804. TIPO: Chile. “In 


andibus de Kane 0 die ‘tis, et quidem in pau nr c. 
R 


de Ranco glareosis, Dic. 


Acaena peana Phil, / Wales Univ. 


ha 00 p. s. m. elevato invenit ornat. R. Pearce, 
o, SGO-49890!). 
Biblioth. Bot. 


de Villarrica. 


rce s.n. (holotip 
1910. 
Neger 


17(74): 52. 
1897," F. W. 


Acaena tenuifolia Bitter, 
TIPO: Chile. “Andes 
ds 1). 


Buys "a el var. glabrinervis Bitter, Biblioth. Bot. 1 707 4): 
TIPO: Chile. *Andes 1897," F. 
W. n s.n. (holotipo, M!). 


. Villarrica, 


P 


Hojas 


anta hasta l6 cm de alto; entrenudos cortos. 


3.3-10.5 em de 


obovado-lanceolada; 4— 


— 


argo: lämina en contorno 


5 pares de folíolos, obovados, 
ápice truncado, base cuneada, asimétrica, 3.5— 
15.5 mm de largo, el margen pinnatipartido, con 7— 


raquis y el ápice de los segmentos; vainas foliares con 


13 segmentos, cubiertos por tricomas blancos en 


tricomas largos, blancos en el dorso; apéndices 
estipulares cortos, como una pequeña elongación, de 
1-3 mm de largo. Rama florífera terminal; pedúnculo 
hasta 13 cm de largo, glabro a pubescente; inflo- 
rescencia capituliforme, ca. 4.5 mm de diámetro, en 
fruto de hasta 7.5 mm; brácteas basales obovadas, ca. 
2 mm de largo, anchas, con el ápice dividido 
irregularmente, con un mechón de tricomas tiesos en 


el ápice y el contorno. Sépalos 4, oblongo-lanceola- 


dos, ca. | mm de largo, con pequeñas manchitas rojas 


a lo largo, glabros; estambres 2-5, filamentos de ca. 

0.5 mm de largo, anteras globosas, ca. 0.3 mm de 
8 8 

largo; estilo corto, estigma globoso, aplastado, más 
o o o 

ancho que largo, laciniado, ca. 0.5 mm de diámetro. 


largo, con tricomas 


Cupela ovoide, 1.5-2 mm de 
glandulares articulados, de color café rojizo, bri- 


llantes, desde la base, 4 espinas de base angosta, de 


1-2 mm de largo, con fuertes gloquidios; aquenio 1. 
8 8 l 


Icones. Bitter, 1910: tab. 3 b, c. Grondona, 1964: 
fig. 1-3; 1984: fig. 445. 
Fenología. | Vructifica desde noviembre (Pisano & 


Henríquez 0882) a marzo (Sparre & Constance 10705). 


Distribución y hábitat. | Especie nativa, se la en- 
cuentra en bosques, en vegas o también en laderas de 
ceniza volcánica, y la zona austral en los sitios mejor 
drenados del turbal, entre cojines de Bolax gummifera 
(Lam.) Spreng. Se distribuye desde Osorno 
a XII Región (Isla Wollaston, 
1800 m (Fig. 
También crece en Argentina (Grondona, 1964, 
y las Islas Falkland (Moore, 1968). 


Caracteres distintivos. 


(Paso 


caleta 
LILA). 
1984) 


Puyehue) : 


Lientur), altitudes entre 10 y 


Capítulos de alrededor de 
4.5 mm de diámetro; pelos glandulares café rojizos 
cupela glabra, con 4 


entre las cupelas, espinas 


apicales pequeñas. 


Acaena antarctica pertenece a la Sección Acroby- 


ssinae. Es una especie cercana a A. masafuerana, de 
la cual es difícil de distinguir en estado vegetativo, ya 


que ambas son de pequeño tamaño, con tricomas 
glandulares de color dorado a rojizo entre las cupelas, 


folíolos con tricomas en las nervaduras, cupelas de 


amaño similar con pequeñas diferencias en el largo 
de las espinas, color de las hojas y pubescencia del 
pedúnculo floral. Walton y Greene (1971), quienes se 
basaron en material tipo, consideraron que ambas 
especies poseen leves diferencias, como la pilosidad 
del escapo y de los folíolos, y el color de estos, pero 
que sin embargo son consistentes, y que en conjunto 
permiten separar una de otra, criterio ace ptado en esta 
revisión. Á esto se puede añadir, que las espinas de la 
cupela son más largas que la mitad de ella, 
a diferencia de A. masafuerana cuyas espinas no 
alcanzan la mitad de su largo. 

Para Acaena antarctica se estudió el material tipo 


depositado en K, que presenta cinco plantas y tres 


— 


etiquetas, dos de las cuales onde con la 
información del protólogo, es decir Menzies y Hermite 
Island. Dentro de los sinónimos, para A. microcephala 
existen dos carpetas referidas al protólogo con los 
números 3029 y 2951, de donde esta última colectada 
por Lechler y depositada en G, se eligió 


lectotipo. Philippi (1864) describe A. pumila Phil., 


pero como ya existía A. pumila Vahl, 


como 


le cambió el 
1893). 


según el tipo depositado en SGO corresponde a A. 


nombre a A. pearcei Phil. (Philippi. la que 


antarctica. Grondona (1964) excluyó a A. antarctica 


de su revisión, y sólo menciona a A. ioni is con 


los sinónimos A. pumila Phil., A. pearcei Phil. y A. 
tenera Alboff. Posteriormente, Walton y 955 'ene 
(1971) demostraron que A. microcephala es sinónimo 


Para la variedad A. antarctica Hook. 


de A. 


f. var. argutidentata Bitter, no fue posible localizar 


antarctica. 
material tipo. No obstante, la fotografía en 
Bitter (tab. 3 b, e, 1910), 


corresponde a variaciones mínimas de la especie. 


según 


se considera que sólo 


Walton (1975) la incluye en un listado de sinónimos, y 
señala que reexaminó el material tipo de todas la 
Bitter. 
última fue destruido en Berlín, pero he podido ver una 
foto en CONC (N^ 3382) que claramente corresponde 
a A. antarctica. 
Schltdl. 


tipo, sin embargo Walton (1975) la señala que vió el 


especies excepto para A, valida El tipo de 


Para la especie A. microcephala 


var minuta Bitter, no fue posible localizar e 


tipo, y la sinónimiza bajo A. antarctica. 


ARGENTINA. 
(HIP). CHILE. 


Material e nds examinado. 


Tierra del Fuego: Rio Olivia, McSweeney 25 


X Regió Malle eco, Tolhuaca, Pennell s.n 
egién: Osorno, P. N. Puyehue, Antillanca, Gardner & 
Knees 3889 (CONC ole & Constance 


'án Osorno, Sparre 
Vic 


; Vo 
10705 (CONC); e P. N 


ente Pérez Rosales, 


422 


Annals of the 
Missouri Botanical Garden 


Figura 2. 


e 
ea 
= 
C 
— 
— 


Tesoro, Schwabe 70 (CONC). XII Región: Ultima Esperanza. 
res ET i 7017 


; Magallanes, Laguna El Parrillar, 
(HIP): Tierra del Fuego, Cordón Baquedano, sector río 


Acaena argentea. A. Habito. —B. Cupela (Johow s.n., CONC 122372). Acaena integerrima. —C. Hábito. 
). Cupela (C. Marticorena et al. 1254, CONC). 


Verde, Pisano & Henríquez 6882 (HIP); Antártica Chilena, 
Isla Wollaston, Caleta Lientur, Pisano 5120 (HIP). 


3. Acaena argentea Ruiz el Pav., Fl. Peruv. Chil. 1: 


O7. 1798, non Acaena argentea Bertero ex 


Volume 93, Number 3 
2006 


Marticorena 423 


Acaena en Chile 


Steudel, (Steudel) i. 9. 
1840. Ancistrum argenteum (Ruiz et Pav.) Kunth, 
in Humb., Bonpl. & Kunth, Nov. Gen. Sp. (quatro 
ed.), 6: 230. 1824. TIPO: Chile. 


habitat abunde locis uliginosis et arvis..., 


Nomencl. Bot, ed. 2 


"Concepción, 


Itatae et Puchacay Provincianum, ad Nipas, 
Yeguaraqui, Collico, Chequen, Loicaca, Pelcho- 
quin et Chaymavida praedia," Ruiz & Pavón s.n. 
(holotipo, MA [Herbarium Peruvianum, Ruiz et 
Pavón, 10/77]!, foto 29606 en CONC!; 
CONC-29701!). Figura 2A, B. 


isolipo, 


17(74): 

“In graminosis siccis a Con- 
150-6 6000," Poeppig 664 (lecto- 
W-359495!; duplicados, WII. 


Acaena 5 var. lanigera Bitter, Biblioth. Bot. 
217 . TIPO: Chi 
cepcion e Antuco, 
tipo, aquí designado 
WRSL!). 

Acaena argentea var. grandiceps Bitter, Biblioth. Bot. 
217. 1910. TIPO: Chile. 


1774): 


“Juan Fernandez (Challenger 


Exped., herb. Vindob.)." Moseley (holotipo, W- 
13511). 
Acaena a forma viridis Bitter, Biblioth. Bot. 17(74): 
10 TIPO: Chile. San 11 in pascuis, Dec. 
1 311 (holotipo, BREM) . 
mes argentea forma epargyrea 1 Biblioth. Bot. 
17(74): 217. 1910. TIPO: Chile. “Hacienda de San 


Juan (Valdivia), Nov. 
mezclada, dos plantas en la parte inferior] ( 
BREM!). 
Acaena argentea -o nigricans ED Biblioth. Bot. 17(74): 
217 de le. San Juan, Dec. 
olección mezclada, una planta en 
MI 


1853." Philippi 311 [colección 


(holotipo, 


Hacienda de 
1854,” putem 


la parte 1 f 111 1 BRE M!). 


Acaena argentea. var. interrupte-pinnata Bitter, . Bot. 
1707): 217. 1910. TIPO: Chile. 
Delessert. et Vindob.,” Ph. 


Arca urbem 


Conce ds ion, js: b. Germain 


lectotipo, aquí designado, W-110756!: duplicado, 


6. 10 visto 95 
Acaena argentea var. coriacea Bitter, Biblioth. Bot. 17(74): 
218. 1910. TIPO: Chile. *In monte ignivomo pr. 
Monac. Rom., Vratisl.," Lechler 
designado, Wi: duplicado, 


Villarrica, herb. 
316a 
WRSI 
argentea var. subcalvescens Biblioth. Bot. 
17 (74): 216. 1910. TIPO: Chile. herb. 


Monac., F. . (lectotipo, aquí designado, 
M!). 


(lectotipo, aquí 
AJ. 


Bitter, 
“Conce 


xd 
peron, 


Neger s.t 


A 


Acaena LL var. breviscapa Bitter, Biblioth. Bot. 17(74): 
218. 1910. TIPO: Chile. *San Cristo al, cria 
Lion, 2 Der 1896, herb. Stock. Dus $ 
(lectotipo, aquí designado, 5-8037! 


Planta hasta 35 cm de alto, rizoma de 3.5 mm de 


diámetro, tallo 2.54 mm de diámetro, cubierto por 
tricomas blancos, lanosos. Hojas en contorno obovadas, 
4-12 cm í 


lanceolados, 33-35 mm de largo, el margen aserrado 


de largo; 3—5 pares de folíolos, linear- 
a pinnatipartido, con ca. 10 segmentos, con un pequeño 
mucrón en el ápice, cara superior glabra, verde oscuro, 
inferior serícea, plateada; vainas foliares con igual 
pubescencia que las hojas, apéndices estipulares 
lineares, 10-18 mm de largo. Rama florífera terminal: 


— 


pedünculo de hasta 20 em de largo, velloso; inflor- 
escencia capituliforme, morada o amarilla, de hasta 


4—8 mm 


, lanceolados, 


14 mm de diámetro; brácteas basales lineares, 
de largo. Flor chasmógama; sépalos 3-4, 
espatulados, con un mechón de tricomas en el ápice, de 
1.5-2 mm de largo, cara interna brillante, cara externa 
cubierta por tricomas largos, transparentes, con un 
mechón en el ápice; estambres 2-5, filamentos hasta 
2.5 mm de largo, anteras globosas, de 0.3-0.5 mm de 
largo: estilo corto, ca. 0.6 mm de largo, estigma 
oblongoide, laciniado, morado, de 1.4-2 mm de largo. 
Cupela obcónica, tetrágona, de color café rojizo, 11.5— 
14 mm de largo, pilosa, con tricomas blandos persis- 
tentes, blancos, con 4 espinas de base angosta en el 
ápice, desiguales, de 1.5-8.5 mm, glabras, con fuertes 


gloquidios; aquenio 1. 
Grondona, 1964: fig. 8-10; 1984: fig. 
1798: tab. 103, fig. b. 


Fructifica de octubre (Lépez y 


Icones. 
442. 


Fenología. 


Ruiz y Pavón, 
Mar- 
quez 108) a marzo (Stuessy et al. 7255). 
Distribución y hábitat. Especie nativa, en Chile 
habita en suelos secos a húmedos, generalmente a la 
orillas de río. Crece desde 
desde 5 a 1000 m 


Juan 


sombra, también a 
Colchagua (La Rufina) a Palena, 
de altitud, encontrándose como maleza en 
Fernández, tanto en Más a Tierra como en Más Afuera 


(Fig. 11B). 


Ecuador 


También se le encuentra en Argentina. 
Perú. Johow (1896) denominó a Acaena 
argentea como “la maleza más común en la isla," y 
Skottsberg (1921) la señaló como “la maleza más 
ampliamente distribuida y nociva” en Más a Tierra. Su 


reproducción vegetativa mediante largos estolones, esto 


unido a las espinas de sus frutos que le permiten 
adherirse a la piel de animales y a la ropa, la distinguen 


como la herbácea invasora más seria. Greimler et al. 


— 


200 2a. 2002 b) señalan que A. argentea, cubre el 
9.21% 


monodominante, siendo desde hace más de 100 años la 


de Más a Tierra, en una comunidad casi 
invasora más común (Johow, 1896). 

Cadillo (Pfister 933), trun (Spa- 
rre 116), amores secos, zarzaparrilla, proquín (Baeza, 
1921). 


Caracteres distintivos. 


Nombres comunes. 


Capítulos de alrededor de 


lem de diámetro, cuando maduros de color café 


rojizo; folíolos lanceolados, cara inferior sericea, 
superior glabra; cupela con tricomas blancos, 


lanosos, con 4 espinas apicales. 

Acaena argentea pertenece a la Sección Ancis- 
trum. El material tipo corresponde a una colecta de 
Ruiz y Pavón depositada en el herbario de Madrid. La 
103, fig. b.) 


corresponde sin dudas a la especie. De los sinónimos 


lámina en Ruiz y Pavón (1798: tab. 


Annals 
3 E ed Garden 


— 


Figura 3. Acaena caespitosa. —A. Hábito. —B. Cupe 
Cupe da (tler 5558, CONC). 


es interesante señalar el material tipo para las formas 


epargyrea y nigricans de la variedad contracta, 


Philippi 311 depositado en BREM, el que pertenece 


a una carpeta con tres plantas, la superior 


corresponde a la forma MIFICAns y las dos inferiores 


a la forma epargyrea, con sus etiquetas. con 


erencias en las fechas, yv que se relacionan a las 


— 


(li 
inferior 


dos formas, aunque la planta en la parte 
a cual la relacionó Bitter. 


BREM de Philippi 


y que es 


derecha es difícil decir 
Además existe otro material 
311, que corresponde a la forma nigricans, \ 
idéntica a las demás. Para la variedad lanigera se 
eligió como lectotipo una carpeta de W que posee tres 
ejemplares, y además cuenta con etiquetas originales. 
Bitter (1910) describió dos variedades de A. argentea: 
var. subcalvescens y var. breviscapa, y en el protólogo 
de cada una menciona dos materiales distintos, 
depositados en diferentes herbarios. Para la variedad 
subcalvescens señala los herbarios M y G, de los cuales 


se pudo revisar el de M, que corresponde con la 


a (Landrum 843 


IN xil 
XU 


„ Hábito. —D. 


7, CONC). Acaena leptacantha. — 


especie y que fue elegido como lectotipo. Para la 
variedad breviscapa señala los herbarios S y M, de los 
cuales se pudo examinar el espécimen de S-8037, que 
fue elegido como lectotipo. 

Para algunos taxa no se pudo localizar material 
lipo, como por ejemplo el de Acaena argentea Ruiz el 
Pav. var. gracillima Bitter, que estaba depositado en 
B. y el que probablemente fue destruido. Otros taxa 
1 por Bitter, 4. Ruiz et Pav. var. 
brevifoliolata Bitter y 


corresponden a leves variaciones de A. argentea, sin 


argentea 


var. pluribracteata Bitter, sólo 


embargo no se pudo revisar material tipo. Para la 


€ 


última pude examinar la foto que aparece en 


revisión de Bitter, la que sin duda corresponde a 
argentea. 
Material A i ntativo examinado, CHILE. 
Valpara Juan Fernández, Más a Tierra. La 
Crawford e et al. 11663 (CONC): Más a Tierra. V. 
Sparre 116 (CONC); Más Afuera, Quebrada Casas 
Lammers 8443 (CONC). VI Región: Cole ism. 


V Región: 


a Rufina, 


Volume 93, Number 3 
2006 


Marticorena 
Acaena en Chile 


425 


l'undo Bellavista, 4 
Región: 


Ene. 1951, Ricardi s.n. 
56 


(CONC); 


971, Silva s.n. (CONC). 


o e 
iu camino de Los Puquios a La Pun 


atilla, / . Martic orena 
et al. 272 (CONC); Ñuble, Atacalco, Pfister ps CONC); 
Cone p. ión, 1, Ramunte ho, Ricardi 367 (CONC); Concepción, 
Coronel, Lépez y Márquez 108 (CONC); Arauco, Laraquete, 


„ (CONC); B iobio, Camino de Ralco a Lepoy, A. 

Marticorena et al 298 (C 'aylor et al. 10380 (CONC). 
Gunckel 38441 
aldivia, San Juan, Ene. 188 9, ONT 
Trumao, Hollermayer 62. CONC); 


IX Región: Cautín, Padre Las Casas, 
(CONC). X Región: V 

s.n. (SGO); 
Llanq tihi 


Palei 


72 55 (ts 


Osorno, 


— 


15 Ene. 19 


Pto. 


Cayutué, 
4 km SE 


73. Martínez s.n. 
Ramírez a Palena, 


CONC); 
Stuessy et al. 


4. Acaena caespitosa Gillies ex Hook. et Arn., Bot. 
Misc. 3: 307. 1833. TIPO: Argentina. “El 
Paramillo de las Cuevas, El Alto de la Laguna. 

Chili.“ Gillies 

K. foto 1227 en 

3688 en S609. 


Andes of Mendoza and s.n. 
(lectotipo, 


SGO!; 


Figura 3A, 


aquí designado, 


duplicado, GL, foto 


B. 
hasta 13 em 
1.5-2.5 cm 


Planta de alto, entrenudos cortos. 


Hojas e largo, lámina en contorno 


obovado-orbicular; 2 pares de folíolos separados por 


más de 2 mm, ovado-orbiculares, 4.5—7.5 mm de 
largo, generalmente doblados a lo largo del nervio 


medio, el margen entero, seríceos, de color verde 


claro; vainas foliares con tricomas largos, blancos, 


tiesos, en la parte central de la cara externa y en el 


margen. Rama florífera terminal; pedúnculo hasta 
9 em de largo, velloso; inflorescencia globosa, sub- 
globosa o espiciforme, de 7-10 mm de diámetro: 


brácteas basales oblongas, lanceoladas o lineares, de 


2.5—4.5 mm de largo, pilosas en la cara externa, 


glabras en la interna. Sépalos 5, oblongo-lanceolados, 


de color verde claro, de 1.5-2 mm de largo, 


pubescente como las brácteas; estambres 2—5, anteras 


o 
o 


obosas, ca. 0.4 mm de largo; estigma lanceoloide, 


aciniado, ca. 1.2 mm de largo. Cupela oblongoide- 


anceoloide, de 4—5 mm de largo, glabrescente, con 


alrededor de 10 costillas finas, recorridas por 8-11 


espinas ca. 0.4 mm de largo, duras, con fuertes 
gloquidios y tricomas retrorsos; aquenio !. 
Icones. Grondona, 1964: 25-28; 1984: 443. 
Fenología. Vructifica de diciembre (Landrum 
8437) a enero (Pfister s.n.). 


Distribución y hábitat. Planta nativa de la zona 
1964, 
1984). Crece en suelo arenoso y en el país sólo se 
a ha encontrado en Tierra del Fuego (XII Región) 
de 10 m (Fig. 11C). 


distintivos. 


austral de Chile y Argentina (Grondona, 


a una altitud 


Caracteres Cupela con costillas cu- 


(CONC). VH 
Talca, 
(SGO); Linares, 
hos Región: 


bierlas de espinas cortas con gloquidios gruesos; 


hojas seríceas, folíolos enteros, doblados por el 
nervio medio. 
Acaena caespitosa pertenece a la Sección Acae- 


na. Esta especie es afín a A. alpina, A. splendens y A. 
integerrima, principalmente por poseer pubescencia 
serícea, con folíolos generalmente enteros. Bitter (1910) 
agrupó a estas tres últimas especies en la subsección 
Splendentes, a diferencia de A. caespitosa que inc 
De 


características que comparte con las demás especies, 


uyó en 


la subsección Dispersiflorentes. acuerdo a las 


se encuentra más relacionada con las especies 


mencionadas, que con las demás de su subsección, 


por lo que sería más adecuado incluirla dentro de 


a 
subsección Splendentes. Sin embargo, en esta revisión 
las subsecciones no han sido consideradas, debido 
principalmente a lo poco claro de sus límites. La 
descripción original señala a la especie para la zona de 


Los Andes de Mendoza y Chile; sin embargo, nunca se 


a ha colectado en el país a esas latitudes. Es probable 
que esto se deba a que los límites territoriales en esa 
época, no eran los de hoy, luego el material tipo se 
colectó en una zona que se consideraba chilena. Esta 
especie es bastante escasa, y sólo se le encuentra en la 
Los materiales 
K, 


corresponden con el protólogo. Ambos materiales están 


zona austral del país. lipo están 


depositados en GL y con etiquetas que se 


bien conservados y completos, pero se eligió el de K 


debido a que la carpeta incluye además un dibujo de la 


cupela, las espinas y las hojas. 


Material representativo examinado. 
but: Carrenleofu, CUBE N. 
Lago Argentino, Dic. 1900, S/C (LP). 
Cabo Goodall 191 (LP). CHILE. XH 
unta 8437 
(C Tierra. del R Ba 1952 
UD s.n. (CONC). 


ARGENTINA. Chu- 


Domingo, 


Landrum 
23 


Delgada, 


nia Felipe, Ene. 


5. Acaena integerrima Gillies ex Hook. Arn., 
Bot. Misc. 3: 306. 1833. TIPO: Argentina. “Near 
El Arroyo de los Potreros, Andes of Mendoza; 
Gillies 113 (lectotipo, aquí designado, KI; 
duplicado, GL, foto en 5GO-68707!). Figura 2C, 
D 


Acaena integerrima var. i Bitter, Biblioth. Bot. 
19 il 


7(74): 85, fig TIPC “Patagonia 
australis-occidentalis: in satan aris, Eberhardt, herb. 
Stockh.,” O. Borge 157 (holotipo. S -03902!). 


Planta hasta 35 em de alto. Hojas en contorno 


obovadas, de 4.2-11 cm de largo; 2-4 pares de 
folíolos, obovado-lanceolados, de 4.5-13 mm de 


largo, tridentados en el ápice, seríceos, de color verde 


Annals of the 
Missouri Botanical Garden 


claro a gris plateado: vainas foliares seríceas, sin 
apéndices estipulares. Rama florifera terminal, hasta 
32 em de largo: pedunculos vellosos, de 20-31.5 cm 
de largo. Inflorescencia capituliforme, de 9.5-11 mm 
de diámetro, en fruto espiciforme o capituliforme, 


hasta 16 mm de diámetro; brácteas basales linear- 


anceoladas, pubescentes a seríceas, 3-5 mm de 
largo. Sépalos 5-6, lanceolados, de 2.5-3 mm de 


largo, glabros en el interior, pubescentes en el 


exterior: estambres Dco. anteras ovoides, ca. 0.7 mm 


de largo; estilo corto, raramente 2 estigmas, lanceo- 


loides, laciniados, de 0.7—1 mm de largo. Cupela 


ovado-orbicular, de 2.5-7.5 mm, pilosa a vellosa, con 


costillas poco notorias, recorridas totalmente por 


duras espinas de tamaño similar, con gloquidios y 
tricomas retrorsos; aquento l. 
1964: 


Fructifica de 


Icones. Grondona, 53-57. 


Fenología. noviembre (Gunckel 


hi 


Especie nativa, crece en 


A 


50679) a marzo (Sparre & Constance 10875 

Distribución y hábitat. 
cojines sobre faldeos secos, desde el Cajón del Maipo 
hasta Tierra del Fuego (Bahía Felipe), de 10 a 2000 m 
de altitud (Fig. 11B). 1 


Caracteres distintivos. 


también en Argentina. 
Folíolos seríceos. de color 
eris plateado, enteros o tridentados en el ápice; cupela 
ovoide cubierta de espinas cortas con gloquidios. 
Acaena integerrima pertenece a la Sección Acae- 
na. Es una especie similar a A. splendens, debido a la 
pubescencia serícea y a lo parecido de sus cupelas. 
Grondona (1964) consideró a ambas especies distin- 
1 1984 


splendens. De acuerdo a los espect- 


euibles, sin embargo e sinonimizó a Á. in- 


tegerrima bajo 
menes examinados, ambas especies son distinguibles 


debido a la forma, color y tamaño de los folíolos, y por 


— 


a forma de la cupela. Los materiales originales de A. 
integerrima (Gillies 1 13) se encuentran depositados en 
K y GL, de los cuales se eligió como lectotipo el 


espécimen de K, debido a que es una carpeta con 


cuatro plantas completas, con etiquetas originales 


manuscritas que corresponden con los datos del 


protólogo. De los sinónimos, para la variedad oliga- 


cantha Bitter, el autor cita dos materiales, uno de 


Philippi 274, de la Cordillera de Linares, depositado en 


B y probablemente destruido, y el otro de Germain, de 


la Cordillera de Maule en el herbario FI. Ninguno de 
los dos se pudo localizar, sin embargo es posible que 
sólo presenten leves diferencias en relación al número 
de espinas de la cupela (oligacantha) que correspon- 
dería sólo a variaciones dentro de la especie. 

ARGENTINA. Men- 
(LP). Neuquén: Cerro 


Material representativo examinado. 


doza: Cerro Guanacos, Carrette 215 


Chapelco, Schajovskoy 94 (LP . Rio Negr 0: C erro Anecón, 
p. S 1 


Ferruglio 49 (LP). Santa 


LE. Región Metropolitana: C 
Maipo, Las Vertientes, Gunckel 50679 (CONC). V H Región: 
Palea, Entre Paso Pehuenches y Laguna del Maule, Ricardi 
et al. 971 (CONC): Linares, Valle Botacura, Schlegel 3577 
(CONC). VIH Región: Biobío, Laguna de la Laja, Ricardi & 
C. Marticorena 5735 (CONC). IX Región: Malleco, ( 
imas Pino Hachado, km 13, Ricardi 
Marticorena 5058 (CONC); Lonquimay, > km SI 
Sparre & Constance 10873 (CONC); Hapa 


amino 
& C. 


5 del pueblo, 


—Paso 


Raíces. cam. Lonquimay-Malalcahuello. articorena el 
al. 1254 (CONC). XI Región: os 


Montero. 54 
mee Ibañez, 12 Feb. 1974. 
XII Región: Ultima Esperanza, Cerro 
70316 (CONC); 
Magallanes, Punta Dungeness. Pisano 4551 (HIP): Tierra del 


E. 


Donoso, 10 a as Chinas, Arroyo et al. 8 


Fuego, Punta Espora, Pisano 3187 (HIP). 


1804. 
FIPO: Chile. "Cum A. macrocephala a cl. Gay in 
Andibus Talcaregue lecta confusa jacebat,” Phi- 
lippi s.n. (holotipo, SGO-49805)). Figura 3C, D. 


6. Acaena leptacantha Phil., Linnaea 33: 66. 


Acaena capitata Phil., Anales Univ. Chile 84: 623. 1893. 

TIPO: Chile. “Habitat in 

Chillanenses, Febr. 1892,” 
560-9905). 


Acaena macrocephala Poepp. var. 
jd B . N 3 


Andibus ad Thermas 
(holotipo, 


Philippi s.n. 


Nuovo 


negeri E. Duse, 


Giorn. Bot. Ital. N.S. 12: 357. 1905. Acaena lepta- 
cantha var. negeri (E. Duse) Bitter, Biblioth. Bot. 
7(74): 148, tab. 15a. 1910. TIPO: Chile. “Andes, 
Villarrica, 1897, Herb Monac.” F. W. Neger s.n. 


(holotipo, M!). 


ai 1 1 var. p 3 T. 9 Bot. 17(74): 


. 1910. TIPO: Chile. "From limits of vegetation, 

a hib: Dessauer in her. Monae. “ Neger 4039 
(holotipo, M!). 

Acaena _leptacantha val longissima 5 Biblioth. Bot. 


7(74): 148, fig. 26. 1910. TIPO: Chile. "Patagonia 


media occidentalis: Vallis unde 1 pars supe- 


rior, 1000 m. 23 Feb. 1897," Dusén 592 (holotipo, S- 
03-9011). 

Acaena 5 subsp. connectens DIR agave Bot. 
17(74): ). 1910. TIPO: Chile. simis 


en G 7 ue nes el 1 Tal caregue, prope nives 
perpetuas," Gay 78 (holotipo, P. excepto la planta de la 
derecha, foto 34720 en CONC!). 

Planta hasta 20 em de alto: tallo cubierto por las 


hojas; entrenudos cortos. Hojas sésiles, de 11-37 mm 


de largo, lámina en contorno oblonga; 7-8 pares de 
folíolos, cuneados en la base, de 4.5-8 mm de largo, 
el margen pinnatisecto, 3 pares de segmentos, cara 
superior pilosa, envés pubescente a seríceo: vainas 
foliares pubescentes con tricomas largos. blancos: 
apéndices estipulares oblongo-lanceolados, 


1.8-2 mm de 


pedúnculo hasta 


pubes- 


centes, de argo. Rama florifera 


terminal; 17 cm de largo, piloso 


a seríceo; inflorescencia espiciforme corta, de 24. cm 


Volume 93, Number 3 Marticorena 427 
2006 Acaena en Chile 
de largo, abriéndose al madurar; brácteas basales brachyacantha Bitter y A. leptacantha Phil. var. 


ovado-oblongas a lineares, el margen irregular, de 3.5— 
2.5- 


3 mm de largo, cara externa pilosa, interna elabra. con 
o [e] 


5o mm de largo. Sépalos 5, ovado-oblongos, de 


un mechón de tricomas en el ápice: estambres 2-3, 
anteras oblongoides a globosas: estigma globoso. 
laciniado, ca. 1 mm de diámetro. Cupela ovoide, de 
1.5-1.7 em de largo, pilosa, con la epidermis levan- 
tada, cubierta por espinas largas, de tamaño similar, de 
color café rojizo, con gloquidios y tricomas retrorsos o 


sin gloquidios y glabras; aquenio I. 


Icones. Grondona. 1964: fig. 33-36; 1084: fig. 
147. 
Fenología. | Fructifica de enero (Stuessy & Baeza 


15560) a abril (Rodríguez y 
hábitat. 


Parra 185). 
Distribución y Planta nativa, crece en 
terreno. húmedo. rocoso, formando matas cespitosas 


laxas. Se distribuye desde el volcán San José (Región 


Metropolitana) hasta Antillanca (X Región). en a 


tudes de 750 a 2500 m (Fig. IIC). También e 


1 
Argentina. 
Nombre común. Pimpinela (Pfister 


s.n. ). 
Caracteres distintivos. — Folíolos pinnatisectos; espi- 
nas de la cupela con gloquidios y pelos retrorsos tiesos 
o sin gloquidios y glabras: inflorescencia espiciforme. 
Acaena leptacantha pertenece a la Sección Acae- 


na. 


= 


Es una especie muy similar a A. macrocephala, 
difícil de diferenciar en estado vegetativo, pero que en 
estado reproductivo se distingue por las espinas de las 
Iricomas 


con relrorsos í 


cupelas con gloquidios, 
totalmente glabras, y por la inflorescencia espiciforme 
corta, que en estado fructífero se abre y se hace más 
laxa. Bitter (1910) incluyó a ambas especies como los 
únicos taxa dentro de la Subsección Astenoglochinae. 
por presentar cupelas algo infladas, con la base de la 
espinas más o menos ensanchadas, las hojas apretadas 


un la 


^ 


base. con pelos algo amarillentos en la cara 


inferior. No obstante, se decidió no incluir divisiones 


menores que la sección, ya que los límites son poco 
claros. 
Philippi (1857) describió Acaena andina Phil., de 


un material recolectado por Germain s.n. (holotipo 


SGO-49888!), el que posee semejanzas con 


leptacantha, lo que le daría prioridad en el nombre. 


Sin embargo. la inflorescencia se encuentra poco 


desarrollada, lo que deja la posibilidad de que 


también corresponda con A. macrocephala en su 
primera etapa de desarrollo, por lo tanto se prefirió 


dejar este taxón como dudoso. 


En relación a los sinónimos, las variedades 
descritas por Bitter, Acaena leptacantha Phil. var. 
grosseaculeata Bitter, A. leptacantha Phil. var. 


dolichacantha Bitter, los materiales originales se 
encontraban depositados en Berlín. por lo cual 
presumiblemente fueron destruidos. Para los dos 


últimos taxa. se pudo examinar la foto que aparece 
1910). las que 


sin duda corresponde a A. leptacantha, es decir que el 


en la revisión de Bitter (tab. 15 d y e, 


autor estableció sus variedades sobre la base de 


caracteres poco estables. 
Material representativo examinado. ARGENTIN A. Neu 
San Martín de los Andes, 
Rio Negro: P. N 
per < Soli 784 (LP): 
gión T teeta 
pss 97 (SGO). 


Volcán ne o II Región: 


aso gara, C. Marticorena & Matthei 1019 (C 
Palea, Laguna Maule, faldeo de los cerros, Rodríguez & 
arra 185 (CA VIII Región: Ñuble, Termas de Chillán, 


Biobío. Laguna del Laja, 
IX Región: Malleco, 
. Marticorena et al. 1417 (CONC): ( 
1909 (CONC): 
ONG 
C ONC 
2 (CONC). 


puni & b 157 (CONC): 
oellner 9998 C) Pino 
ii 
Volcán 


X Región: 


; Osorno, 


Llaima. Montero 


A 


Pfister s.n. (6 
Montero 4985 
52. 


1 & Matthei 


Volcán blema a, 


~ 


7. Acaena lucida (Aiton) Vahl, Enum. PI. 1: 296. 
1804. non Acaena lucida (Lam.) Vahl. Enum. Pl. 
|: 296. 1804, nom. illeg. Ancistrum lucidum 
\iton, Hort. Kew. I. 15. 1789. TIPO: Falkland 
Island, Antarct. Expedition 1839-1843. J. D. 
Hooker 1827 (neotipo, aquí designado, G!). 
Figura 4A, B. 

MI iae a Lam., Tabl. Encycl. 1, Ilustr. 1: 77, tab. 

791. Nom. 2. g., non Ancistrum lucidum 
Ait T oe Kew. . 1789. 

leaena. lucida am.) um iom PL 1: 296. 1804. Nom. 
illeg. 

leaena 75 D Phil., Anales. Univ. Chile 84: 622. 1893; 


caena lucida var. parvifolia (Phil.) Reiche, Anales 
l Ps a ule 98: 161. 1897. TIPO: 


orientali 


Argentina. “In 
lecta." Ortega s.n. (holotipo, SGO- 
leaena lucida var. villosula Bitter. Biblioth. Bot. 17(74): 61. 
tab. s 1910. TIPO: Islas Falkland. “Insula Falkland.” 
J. D. Hooker s.n. (lectotipo, aquí designado, P-106291!: 


( E S-R8052!, CONC-29690!. P-106290!). 
Acaena lucida var. abbreviata Bitter, Biblioth. Bot. 17(74): 
61. 1910. TIPO: Argentina. “Fuegia orientalis: Rio 


Grande, pl. 
280501). 


in reg. Magell..” P. Husen 377 (holotipo, S- 


Planta de 3.2 


3.5 mm de diámetro: 


—6.5 em de alto: rizoma leñoso. ca. 


tallo subdecumbente o erecto. 


con entrenudos cortos. Hojas en contorno linear- 


lanceoladas, de 1.7-2.2 em de largo; 6-7 pares de 


folíolos, linear-lanceolados, profundamente bisecta- 


dos, con el segmento superior más largo que el 


inferior, de 1.5-2.2 mm de largo, glabros. de color 


Annals of the 
Missouri Botanical Garden 


Figura leaena lucida. —A. Hábito. —B. Cupela 
—D. Cupela (Ricardi y Marticorena 5038/1422. CONO). 


verde claro; vainas foliares glabrescentes; apéndices 


estipulares ausentes. Rama florifera axilar: pedüncu- 


los de 3.3-5 em de largo, pilosos: inflorescencia 
capituliforme, de 4.5—5.5 mm de diámetro, con menos 
de 10 flores. todas acompañadas por brácteas florales 
ovadas. de ápice agudo, cóncavas, pubescentes en la 
cara externa y glabras en la interna, de 2-2.5 mm de 
largo. Sépalos 3-5. ovado-orbiculares. de 0.8-1 mm 
de largo. pilosos, con un mechón de tricomas en el 
ápice; estambres 2-3, anteras globosas: estigma 
eloboso, laciniado, ca. 0.3 mm de diámetro. Cupela 
ovoide, de color morado en la parte superior y 
amarillo-verdoso en la inferior, de 2-3 mm de largo. 
con tricomas blaneos, blandos, escasos, dispuestos en 
el tercio superior de la cupela o totalmente pilosa, 
superficie rugosa: espinas rudimentarias en la parte 
superior o sin ellas; aquenio I. 

Icones. Grondona, 1964: fig. 22-24; 1984: fig. 451. 


Fructifica de diciembre (Elvebakk 538) 


Fenología. 


a marzo (Elvebakk 470). 
Distribución y hábitat. Plantita naliva, bastante 
escasa que crece en laderas rocosas. formando 


1 


(M. T. 


C. Hábito. 


K. Arroyo 92-393, CONC). Acaena macrocephala. — 


pequeñas champas, en sitios con buen drenaje de la 


la XI 


le altitud (Fig. 11D). También crece 


turbera. Se encuentra el 
l 1100 m 


en las Islas Falkland (Moore, 1968) y en Argentina. 


Región, desde los 


a 


Caracteres distintivos. Cupelas sin espinas o muy 
rudimentarias: sépalos persistentes: folíolos bisecta- 


dos, los segmentos de distinto largo. 


Acaena lucida | único representante de la 


jS € 
Sección Pleurocephala. Esta especie es fácilmente 
reconocible debido a que sus cupelas no poseen 
espinas. o sólo algunas prominencias hacia el ápice. 
Aiton (1789) y Lamarck en Lamarck y Poiret (1791) 
describen independiente Ancistrum lucidum basado en 


una colección. proveniente de las Islas Falklands. 


amarek (Lamarck v Poiret, 1791) no se refiere a Aiton, 
quizás porque su trabajo ya estaba en prensa cuando el 
vió el de Aiton (Gandhi com. pers.). Luego Vahl (1804) 


refiere tanto a Aiton como 


a combina a Acaena, y se 
Lamarck y Poiret (Lamarck y Poiret, 1791), en ese 


orden. Sin embargo, Ancistrum lucidum Lam. es un 


nombre inválido por ser un homónimo posterior, y l 


Volume 93, Number 3 
2006 


Marticorena 
Acaena en Chile 


combinación correcta es la de Aiton por ser e 
basiónimo más antiguo. Aiton depositó materiales en 
BM y G-DC, pero no se encontraron. Se estudió un 
material de Green Plant Herbarium (TRT), pertene- 
ciente al Scrapbook de Adam White, que tiene tres 
plantas, una etiqueta que dice A. lucida, Falkland, y se 
presume es de J. D. Hooker. Sin embargo, el material 
elegido como neotipo se encuentra depositado en G, y 


es una carpeta con ocho plantas, la que además fue 


usada para la ilustración que aparece en The Botany of 


the Antarctic Voyage (Hooker, 1844-1847, pl. 94). 
Respecto de los materiales citados para Acaena 

Bitter, 

examinar tres ejemplares, dos de P, y uno en CONC, 


lucida (Lam.) Vahl var. villosula se pudo 
Se eligió como lectotipo el espécimen P-106290, que 
tiene nueve plantas, cinco de las cuales se encuentran 
en estado reproductivo. El material P-106291 tiene 
tres plantas, dos en estado reproductivo, en tanto el 
ejemplar de CONC 


estado reproductivo. 


presenta tres plantas, una en 


Material imd ntativo examinado. C HILE. 


XII Región: 
J 


Elvebakk 538 (CONC); P. 
Elvebakk 470 (C ws D ps del 
RM a Bahía Felipe, 19 Mar. ! 

J. (SCO). ISLAS FALKL NUN 1 (560). 


Refugio Pehoé, 


8. Acaena macrocephala Poepp., Fragm. Syn. PI. 
25. 1833. TIPO: Chile: *Cr. in Chile austral, 


rupibus basalticis ad pedem montis ignivomi 
Antucensis.” Poeppig 45 (lectotipo, aquí desig- 


. foto en CONC?). Figura 4C, D 


nado, 


Planta hasta 24 cm de alto; tallo totalmente 


cubierto por las vainas foliares; entrenudos cortos. 
Hojas sésiles, de 1.44.5 em de largo; lámina en 
5-10 


ovados, que se pliegan a lo largo y se disponen uno 


contorno ovado-oblonga; pares de folfolos 


sobre otro, de 3-6 mm de largo, el margen pinnati- 

secto, 3 pares de segmentos, cubiertos por tricomas de 
color amarillo pálido; vainas foliares muy imbricadas, 
con tricomas largos por todo el margen y el dorso: 
apéndices estipulares rudimentarios, hasta 2.5 mm. 
Rama florífera terminal; pedúnculo hasta 11 em de 
largo, con tricomas largos 


pubescente, delgados; 


inflorescencia capituliforme o raramente subespici- 


forme, de 1.3—4.5 em de largo; brácteas basales 


pubescentes, con un mechón de tricomas en el ápice, 
de 1.8-3 mm de 


espatulados, de 2.5-3 mm de largo, pubescentes en el 


argo. Sépalos 4, ovado-lanceolados, 


dorso, con un mechón de tricomas blancos en el ápice: 
estambres 3-5, filamentos ca. 1.5 mm de largo, anteras 


globosas, moradas, de 0.5-0.6 mm de largo: estigma 


globoso, de 0.5-1 mm de largo. Cupela ovoide, de 1— 
1.2 em de 


velloso-tomentosa, cubierta por espinas blandas, de 


— 


argo, con la epidermis desprendiéndose, 


tamaño similar, pilosas con tricomas largos y finos casi 
hasta el ápice glabro y sin gloquidios, la base 


ensanchada o dilatada; aquenio I. 
Icones. Grondona, 1964: fig. 


Fructifica de 


32; 1984: fig. 448. 
(Ricardi & C. 


Marticorena et al. 107). 


Fenología. enero 
Marticorena 5709) a marzo (C. 

Distribución y hábitat. | Especie nativa, habita en 
desde el Descabezado del 
Maule (Talca, VII Región) a Valdivia, volcán El Mocho 
(X Región), 900 a 2500 m 
(Fig. 11D). 1 


Caracteres distintivos. 


terreno arenoso Y rocoso, 


entre altitudes de 
'ambién en Argentina. 
Inflorescencia capituliforme 
grande: espinas de la cupela pilosas a todo largo, con 
pelos largo y suaves: folíolos pinnatisectos; apéndices 
estipulares rudimentarios o ausentes. 

Acaena macrocephala pertenece a la Sección Acae- 


na. Es una especie muy cercana a A. leptacantha, de 


a cual se puede diferenciar sólo en estado fructífero 
por sus espinas con tricomas blancos y finos, con el 
ápice glabro, y por la forma de la inflorescencia o 


infrutescencia que es capituliforme. Bitter (1910) creó 


a subsección Asthenoglochinae donde incluyó a ambas 
especies. Presenta una distribución bastante restrin- 
gida, siendo además relativamente escasa. El material 
tipo de A. macrocephala son dos carpetas de Poeppig 45 
depositadas en M y B (fotos 19315 y 3380 respecti- 
vamente en CONC). El espécimen de B presumible- 
mente ha sido destruido, por lo tanto se eligió el de M 
como lectotipo. La carpeta posee dos plantas, una de 
las cuales presenta una cabezuela bien desarrollada, 
característica de la especie. Bitter (1910) describió la 
variedad caput-medusae utilizando los tipos de A. 
por lo tanto dicha variedad es un 
Bitter (1910) 


intermedia Bitter considerando que presenta un tamaño 


macrocephala, 


nombre inválido. creó la variedad 


intermedio entre la variedad caput-medusae y A. 
leptacantha. El protólogo señala el material tipo para 


B, pero se presume que ha sido destruido. Sin embargo, 


a foto que aparece en su revisión (tab. 15c), muestra 


claramente que las características de la variedad caen 


[om 


dentro de la variación natural de la especie. 


CHILE. VII Región: 


Material bd esentativo ex ón 
che 


‘alca, Pas 


^ehuene Marticorena et al. 107 (CONC); 
4 5 del n „A. . et al. 228 (CONC). VIII 
ón: Ñuble n rmas de Chillán; 3 Feb. 1972. Duek & 


Faldeos del Volcán 
Laguna de la Laja. Eoo & C. Marticorena 5709 (CONC). 
IX Región: Malleco, Termas da Manzanares-Lonquimay, km 


430 Annals of the 
Missouri Botanical Garden 


Figura 5. Acaena magellanica. —A. Wi a : upela (C. Marticorena et al. 429, CONC). Acaena ovalifolia. —C. 
Hábito. —D. Cupela (C. Marticorena et al. 423, CONC 

20. Ricardi & C. Marticorena 5014 (CONC); Pino a hado. 223.” Bridges 522 (lectotipo, aquí designado, K, foto 
10 Ene. 1948, Pfister s.n. (CONC). X Región: Valdivia. 1244 en SCO). 

Volcán El Mocho. 1 Feb. 1953, Pinto s.n. (CONC). de cadilla Hook. Fl. Antaret. 1: 269. 1846, TIPO: 

hile. “Hab. in near Valdivia.” punt 772 

8 ectotipo, aquí designado, N. m 1235 en D. OD. 

9. Acaena magellanica (Lam.) Vahl, Enum. PL l: Acaena closiana Gay, Hist. Chile, Bot. 2. 298. 1847. Acaena 

297. 1804. Ancistrum magellanicum Lam.. macrostemon subsp. closiana on 11 5 Biblioth. 

Illustr. 1. p. 76, tab, 22, fig. 2. 1791. TIPO: Bot. 17(74): 193. 1910. TIPO: Chile. . . en las 


ET ras de la provincia de Coquimbo, a lo largo 


Chile. "E. Magellania. Commers. Ex herbario i 
de los riachuelos. y a la altura de 6.000 a 10,000 


Jussiaei.” Catal. 14218 (lectotipo, aquí desig- i f ; 
. . . e : E Pies: say s.n. (neotipo, aquí designado. P- 
nado, P, foto 34722 en CONC!). Figura 5A, B. 063101). 
Acaena venulosa Grise b., Syst. Bemerk. 30. 1854. Acaena 
l 


Acaena macrostemon Hook. f., El. Antarct. B: 209. 1846. aevigata T. Aiton var. venulosa (Griseb.) Reiche. 
TIPO: Chile. “Hab. Cordillera of Chili. Herb. Cuming Anales Univ. E hile 98: 171. 1897. Acaena magellanica 


Volume 93, Number 3 
2006 


Marticorena 
Acaena en Chile 


subsp. venulosa (Grise e BI Biblioth. Bot. 17(74): 
168. 1910. Pr Chi 
collibus pr. Sandy 


enins. Brunswick: in 


. Octobri," Lechler 978c 


Point, 


sclolipo, aquí de signado. Wi I! 
Pl Aan: ama: 192. 
me oridionalis deserti 


1560. TIPO: 


Alacamensis 


Acaena canescens Phil.. Re 
Chile. In partis 
vallibus editioribus rara non est, e.g. Cachinal de la 
Sierra, 7000 p.s.m., Sandon 9000 p.s.m.," Philip, 
596a (lectotipo, aquí designado, SGO-39729!: d 
cados, SGO-49883). 
hil, Linnaea 33: 63. 
“Pariter de planitie 


1864. TIPO: Chile. 
bs ad radicem. Andinum 
C. Cox s.n. (holotipo, SGO- 


^ 
S 
es 
— 


1 ano 180: 
198 
Acaena 11 Phil., Anales Univ. 


Chile 41: 711. 1872 
TIPO: Chile. “Común en la provincia de Valdivia, 
señaladamente cerca de 1838,” 
s.n. (lectotipo, an designado, S6O-49934/). 
leaena hirsuta Phil., Anales Univ. Chile 41: 712. 1872. 


Corral, Ene. Philippi 


caena 1 var. B sue (Phil.) Reiche. Anales Univ. 
c hile 98: 170. 1897. TIPO: Chile. “De la UM la de 
alc Ne común en mi fundo de San Juan, Feb. 1858 
(lectotipo, aquí designado. SGO-49858! 
Chile 84: 624. 1893. 
Santiago loco 
> Philippi 559 ( u 
SGO- 397 2313 


n pl s.n. 


Acaena petiolulata Phil., Anale s Univ. 
TIPO: Chile. “In 


Valle del Yeso, Ene. 


=~ 
= 
T 


designado, 
19885!) 

leaena ipe Citerne. Rev. Sci. Nat. 
TIP Chile austral., 


duplicados, SG 


Quest 7(2): 43. 1897, 
' colector desconocido (neotipo, 
P, foto 34719 en CONC). 

a H. Ross, Oesterr. Bot. Z. 57: 449. 1907. 
| subsp. longiaristata (H. 


f 192. 1910 TIPO: 


( uf 1 


Acaena longiaristat 


anitz, herb. 1 TIC. Nr. 


aquí 


1263." €. Buchs ien s.n. 
BREM!: 


designado, duplicado, 


Acaena gluucela us Biblioth. 10 0 170674): 157. 1910. 
TI Punta Arenas, 16 Dic. 1895," Dusén 188 
(isolipo, pos no visto, sinonimia según y po 1975) 

Acaena macropoda Bitter Biblioth. LAE 59. 1910. 
TIPO: Chile. “Patagonia 1 Erro 11 5 Svensk 
exp. Patag. 18 Mar. 1899, sub nom. A. laevigata 
\iton.” O. Borge 214 (holotipo, 5-80531). 

a magellanic ca var. esca. ep Biblioth. Bot. 


): 168. 1910. 


SE pr. | 


TIPO: Chile. 


Point, m. 


Penins. Brunswick: in 

Sandy Octobri," Lechler 978€ 

(lectotipo, aquí designado, BREM!) 

rers Bot. 

20: Chile. “Cord. de V 
18 


gies ann. 


Acaena _magellanica wid eran Bitter, 
74): 169. 1910. ' 
dum 


la ule, sub 


56-1851. l. 

(lectotipo, aquí designado, W-3594144). 

Acaena su Bitter, Biblioth. Bot. 17(74): 182. 1910. 
TIPO: Chile. “Regio Mage llanica: Cape Gregory Bay, 
1 [sic] of Magellan.” Capt. King 16 (holotipo, G, 
foto 27426 en CONC!). 

Bitter. Biblioth. Bot. 
TIPO: Chile. “Magellansland, N.J. ( 
ae Vahl.” 

ado. 5-8058! 

Acaena breit Bitter, oe Bot. 17(74): 204. 1910. 
TIPO: Chile. “Cord. de Stockh. 
nom. e (holotipo, S- 
80491). 


Germain s.n. 


leaena ne d 


17(74): 183. 1910. 
(1852), sub nom. A. 
280 (lectotipo, 


Anderson aquí 


anares, in herb. sub 


canescens "uem Germain s.n. 


Acaena Sp Bitter, 


Acaena exaltata. Bitter. 


Acaena macrostemon Hook. f. var. basi pilosa Pitter, Biblioth. 
1 
O: 


Bot. 17(74): 1910. TIPO 


der chilenischen Hoche oiler (36 . Breite) 
Juncal. 2300 m: 7L. 1] 1 . 0. Buchtien 5464 
(lectotipo, aquí di PCM 

Acaena 19 un itter, ick Bot. 17(74): 205. 


1910. Chile. 


A. canescens ; Phil. 


“Cordillera de Colchagua. sub nom. 

Herb. Berol., ` Philippi. s.n. (isolipo, 

FI, no visto, sinonimia según Walton, 1975). 

leaena 0 88 rata Bitter, Biblioth. Bot. 17074): 206. 1910. 
TIPO: Chile. “Cordille Maule. uni 
A. ¿ d ni Bitt. sub nom. A. pomo. Phil." Ph. 
DS 1 66 8b (holotipo, FI, no visto. 

75). 


era de uná cum spec. 8l 
sinonimia según 
Walton, 19 
"S ~ Bitter, 
TIP hile. 
T s.n. (holot 
Acaena avandia Biner Silo te Bot. 17(74): 206. 1910. 
TIPO: Chi dillera de Maule, 
de dedos 0 e Ph. Germain s.n. 
v. foto 


Biblioth. Bot. 17(74): 206. 


1910. 
“Andes des W 


Santiago, herb. Monac..” V. 


sub nom. A. 
(lectotipo, aqui 
241 en SGOL duplicado, F 


designado, 
Visto). 
Biblioth. Bot. 


"Magallanes, 


17(74): 210. 


sub nom. 4. 


1910. 


leaena ise a Bitter. 
TIPO: Ch laevi- 


ile. “Hab: 

ie Alton," Philippi s.n. (lectotipo, aquí designado, 
W-110007!). 

une 17 ure Biblioth. Bot. 


agonia 


17(74): 211. 1910. 
australis: ( Cape Fairweather 
ji: le sert.“ Capt. King 10 (holo- 
tipo, G, foto 27424 en (C! 


Chile. “Pat 


1 i 


Acaena ee Bitter, ios m 17(74): 212: 1910. 
TIPO: Chile. ee ulio: 
ste Isl.) Mission du Cap Horn 1882-83, he ib. 
Paris > ades ae ‘holotipo, P-106312 
Acaena drausei | Phil. var. massonandra Bitter, 
74 10 TIPO: Chile. 
Webbianum in herb. 


Regio magellanica: 


Ds 
Biblioth. Bot. 


glabratula Toten 
1910. TIPO: Chile. "Chile: 


* Ochsenius s.n. 


Acaena krausei Phil. subvar. 
17(74): 223. 
Berol., Brem. 

. BREM! . 
doen gil Bitter, Biblioth. Bot. 17(74): 223. 1910. 
Chile. "Patagonia occidentalis: In valle 1 
ysén, sub nom. 4. 


Bod 
(lectotipo, aquí desig- 


adscendens Vahl, 14.1.1897 
Upsal.,” 
1. (lectotipo, aquí designado, S-R8044!). 
e sulci Bitter. Biblioth. Bot. 17(74): å 
TIP Chile. “Pal 
P 8 in reg. Mese il lectae sub n a 
Vahl, herb. Dusén 190 (holte. 
visto, sinonimia segün Walton, 1975). 
Biblioth. Bot. 17(74): 


Chile. “Patagonia australis: 


flore mem et fructiferam herb. Stockh.. Dusén 


australis: 


Upsal.." UPS, no 
225. |910. 
Arenas, Hio 
„ cum fructibus PORE und cum, 
. exped. suec. 1907-1909, 20.2.1908," 
Skottsberg 175 (holotipo, SGO-! 2 9 0 

. ra i Bitter. en Bot. 17(74): 227. 

Chile. “Punt enas. Te 


1 de la 


Punta 


1910. 


Tains secs pelouses 


= 


1 elo gera sub 4. 
rb. Mus. 


1 7 ens Vahl var. macroc a Franch, 
) (holotipo, P- 


Paris, 10 Feb. 18772 
1063091), 


avatier s.n. 


Bablioth, Bot. 17(74): 2 tab. 24. 
1910. TIPO: Chile. “Patagonia australis: 8 Arenas, 
Rio de las Minas, cum fructibus submaturi, uná cum 


A. 
1907-09, 20.2.1908," 


1 


rubescens Bitt., exped. suec. 


432 


Annals of the 
Missouri Botanical Garden 


Skottsberg s.n. (isotipo, UPS, no visto, sinonimia según 
Walton, 1975). 
Acaena e Bitter, Biblioth. Bot. 17(74): 230. 1910. 


TIPO: Chile. "Fretum magellanicum: Peninsula Bru 
105 Port Famine, una cum specimine e A. ovali ifolia R R. 
et P. sine nomine in herb. Deles > Capt. King 46 


(holotipo, G, foto 27428 en CONC! 0 
Acaena brachyglochin Bitter, Biblioth. Bot. 17(74): 291. 
1910. TIPO: Chile. “In cordillera at 3500 ped., pl. 
129," Philippi s.n. (holotipo, BREM, no visto, 
sinonimia según Walton, 1975). 


chilensis 


Planta hasta 25 em de alto; rizoma ca. 5.5 mm de 


diámetro; tallo glabro, ascendente; entrenudos cortos. 
Hojas en contorno oblongas a linear-lanceoladas, de 
22-125 em de 


base cuneada, 


folfolos, 


argo; 5-10 pares de 


de 6-33 mm de 


segmentos, 


obovados, argo, el 


margen pinnatipartido, 4-7 pares de 


elabros a pubescentes, con tricomas cortos, delgados, 


blancos a transparentes, más abundantes en el envés: 


vainas foliares presentes, apéndices estipulares lan- 


ceolados, de 3-5 mm de largo, el margen irregular, 


elabrescentes. Rama florífera terminal; pedúnculo 


hasta 20 em de largo, glabro a pubescente; inflo- 
rescencia capituliforme, morada o amarilla, hasta 
15 mm de diámetro; brácteas basales alargadas, de 
forma irregular, hasta 6 mm de largo: algunas flores 


encontrándose a lo largo del pedúnculo floral, 
Flor hermafrodita con sépalos 4— 
de 0.8-2 
erna glabra, brillante, cara externa pubescente con 


„ fila- 


1 


cercanas al ápice. 


5, ovado-elípticos, mm de largo, cara 


in 


tricomas transparentes, largos; estambres 2-5 


mentos de 2-3.5 mm de largo, anteras 


a lanceoloides, moradas, de 1.2-2 mm de largo; estilo 


corto, estigma obovoide a lanceoloide, laciniado, 


amarillo a morado, 1-2.5 mm de largo. Flor femenina 


con estigmas notorios, 3 mm de largo, estaminodios 


0.5-1 mm de largo. Cupela obeónica, tetrágona. de 
6 mm de largo, pilosa con tricomas blandos, persis- 
4 espinas apicales, de base 


10 mm. 


tentes, blancos, con 


angosta, raro 3 6 5, desiguales, hasta con 


fuertes gloquidios; aquentos 1. 
1984: fig. 446. 


Marticore- 


1964: fig. 11-13; 


Fructifica de diciembre (A. 


leones. Grondona, 
Fenologta. 
na et al. 98) a junio (Loyola & Morales 


Se observaron flores femeninas (Arancio de Squeo 


10296, Ricardi 2900, Squeo 88068), y también 
algunas con características masculinas (Zoellner 


0793, Rodríguez 3184, Rodríguez y Ruiz 3715). 


Distribución y hábitat. Especie naliva, crece en 


suelos arenosos húmedos. en vegas, en grietas de 
rocas, a orillas del agua. Se distribuye ampliamente 
desde la | Región (Arica, Puquios) hasta la XH Región 
(Caleta Misión, 


1200 m de altitud (Fig. 11E). Presenta una amplia 


. del nivel del mar hasta 


— 


Isla Host 


distribución la que comprende el sur de Sudamérica, 


Islas Marion, Islas South 
Georgia, Islas Macquarie e Islas Falkland (Greene, 
1964: Moore. 1968: Walton, 1979: Pisano. 1984). Sin 


embargo, ES interesante señalar la ausencia de esta 


y las Islas Kerguelen, 


especie en el Archipiélago de Juan Fernández, donde 
si crece abundantemente Acaena argentea, y en menor 
cantidad A. ovalifolia (Marticorena & Cavieres, 2000). 

Nombre común. Amor seco (Montero 12650), cadilla 
1921), trun (Aravena 29). 


Cupela — obcónica-tetrá- 


(Baeza, 


Caracteres distintivos. 
gona, con 4 espinas apicales glabras a algo pilosas; 
verdoso, de hasta 15mm de 


capítulo maduro 


diámetro. 
Sección 


magellanica pertenece a la 


Grondona (1964) y Greene (1964) señalan 


Acaena 
Ancistrum. 


ï 


adscendens Vahl (1804) como sinónimo 


E 
magellanica, sin embargo no se ha conseguido localizar 
o ver el tipo, por lo tanto no se ha podido confirmar la 
sinonimia. Para el sinónimo A. elostana Gay (1847), no 
fue posible localizar el material Upo, por lo tanto se 
eligió un neotipo de un material cuyos datos se 
corresponden con los originales, excepto por el año 


1858, pues la especie fue descrita en 1947. 


"ara Acaena macrostemon, estudié un fotografía del 


material tipo (K), y las características generales 


pertenecen a A. magellanica, cuya descripción 


original dice “...calycis aristis. 2 caeteris. duplo 
longioribus.... y establezeo que esto se refiere 


a que posee dos espinas del doble del largo de las 
otras, es decir son cuatro, lo que se corresponde con 
A. magellanica. 

Lechler 978c, 


proveniente de M, con el que están relacionados los 


Se ha estudiado material de uno 
A. laevigata var. venulosa, 
uno de BRIM 


glabrescens. El 


nombres Acaena venulosa, 


A. magellanica subsp. venulosa. y 
asociado a A. magellanica var. 
asignado como lipo del epíteto 


BREM 


malerial de M fue 


venulosa, y el material de como tipo 
a glabrescens. 

En SGO se encuentran dos materiales con los datos 
de localidad, el N 
inflorescencia inmadura, cuya eliquela dice Acaena 


49883 


39729 presenta una planta con la 


canescens y el nombre deserticola rayado, el N 
tiene una planta sin inflorescencia, con 
queta con el nombre A. deserticola Phil., y con datos 


esta no fue 


de localidad más completos, pero 
publicada, luego se designó como lectolipo N 


39729 por ser una planta más completa. Para el 
material SGO-49934 presenta 


tres etiquetas, de las cuales una corresponde con el 


nombre A. krausei, el 
protólogo de la especie, es una planta completa con 


frutos maduros. Para A. hirsuta se encontraron dos 
SGO, pero uno correspondía con la 
058. v fue 


estudiaron 


materiales en 


fecha de recolección feb. elegido con 


lectotipo. Dos materiales se para A. 


Volume 93, Number 3 
2006 


Marticorena 
Acaena en Chile 


petiolulata, ambos provenientes de SGO, uno estéril 
(SGO-39723) y el otro fértil (SGO-49885) que fue 
elegido como lectotipo. 

Se estudiaron dos materiales de Acaena longi- 
aristata, de BREM y WRSL, pero se eligió el de 
BREM por presentar dos plantas con inflorescencias, 
en tanto la planta de WRSL es infértil. Bitter (1910) 


menciona de modo general, materiales de A. macro- 


— 


stemon var. basipilosa depositados en M, de los cuales 
se estudiaron dos, pero no son duplicados, ya tienen 
números de colector distintos (5410 y 5464). „ se 


eligió el N 5464 por presentar dos plantas bien 


— 


desarrolladas y tres hojas extendidas. Para A. 
magellanica subsp. pygmaea, se citaron dos materia- 
les, ambos de W, pero uno colectado por Philippi en 
la Cordillera de Linares, y el otro de Germain en la 
Cordillera de Maule. Ambos fueron estudiados y se 
eligió el N 3594.14 de Germain, que presenta dos 
plantas fértiles. Bitter (1910) menciona dos materiales 
para A. neglecta, en UPS y S, pero sólo se pudo 
estudiar el de S-8058, que presenta dos plantas, una 
fértil, y fue elegida como tipo. Para A. ischnostemon 
menciona materiales para B y W, pero sólo se pudo 
estudiar dos materiales de W, y se eligió el que 
presenta tres inflorescencias, con el N^ 31361 en una 
reglilla, y como duplicado W-116067. Para A. krausei 
subvar. glabratula, Bitter (1910) menciona materiales 


para B, y BREM, sólo se pudo estudiar el material de 


— 


BREM, que corresponde a dos plantas bien desarro- 
lladas, y que fue elegido como tipo. Para A. hirta 
Citerne se designó un neotipo que corresponde a un 
material de Gay (N 71). depositado en P, debido a que 
no se encontró el material original. 

Entre los taxa descritos por Bitter (1910), muchos 
fueron basados en caracteres. variables dentro de la 
i 5 22 


subvariedades; por 


especie. Se pueden mencionar especies, 5 
subespecies, 5 variedades y 2 
ejemplo: folíolos blanquecinos (Acaena glaucophylla 
Bitter), brácteas notorias entre las flores (A. frondosi- 
bracteata Bitter), escaso número de dientes en los 


(A. Bitter), 


lobulados (A. obtusiloba), número de eloquidios 


inferiores 
A. 


triglochin Bitter), largo de los sépalos (A. longisepala 


folíolos oligodonta folíolos 


— 


Bitter), número de dientes del folíolo (A. noremdentata 
Bitter), v así la lista continua. 


materiales tipo que fueron depositados en 


Mundial. Dentro de estos he podido ver fotos del tipo 
de Acaena triglochin, A. longisepala. A. novemdentata., 
A. molliuscula, A. acutifida Bitter y A. stellaris Meyen, 
los que sin duda por sus características corresponden 
a A. magellanica. Sin embargo, para otros no se 
encontró material fotográfico, como A. purpureistigma 

itter, 4. 


Jrondosibracteata y A. 


= 


krauset Phil. subvar. pilosior Bitter, A. 


oligoglochin Bitter var. doli- 


choglochin Bitter. A pesar de esto, dichos taxa son 


reconocibles con A. 


magellanica, de acuerdo a la 
detallada descripción que hace Bitter (1910), pero no 
se pudo localizar material relacionado. Otra situación 
se presenta con taxa donde no se mencionó material 
tipo. Para algunos fue posible determinarlos como A. 
magellanica, debido a que Bitter (1910) en su revisión 
incluyó fotos de algunos taxa, como para A. 
basibullata Bitter (tab. 23). Kalela (1940), describió 
A. subnitens Kalela de cinco materiales provenientes 
de Magallanes, pero no señaló herbario. En el artículo 
realizó una completa descripción, que concuerda con 
A. magellanica, lo que se corrobora al examinar las 


dos fotos (tab. 14, 15) que en él aparecen, y con lo 


cual Grondona (1964) también concordó. Por último. 
para algunos taxa no fue posible localizar material 


tipo, principalmente porque no fue señalado, aunque 


las descripciones permitieron reconocerlas como A. 


— 


magellanica. Dentro de estas podemos mencionar / 
denudata Reiche, a la que Walton (1975) señala como 
su sinónimo. Acaena laevigata, según señala Gron- 
dona (1964), fue considerada por W. T. Aiton (1810— 
1813) lo que Lamarck (Lamarck y Poiret, 1791-1823) 


describieron como Ancistrum magellanicum var. B, la 


» 


que probablemente sólo corresponde a Acaena ma- 
gellanica. Un caso distinto es el de Acaena glandu- 
lifera 


tampoco se localizó material original, pero que 


Bitter subsp. nordenskjoeldii Bitter, donde 


además su descripción parece haber sido hecha desde 
un material algo inmaduro, debido al diámetro de la 
inflorescencia, y al tamaño de las espinas de la 
No 


bastante similar a A. magellanica. 


cupela. obstante, en términos generales es 


Respecto de su número cromosómico. se ha 
determinado la presencia de poblaciones diploides 
con un 2n = 42, y tetraploides con 2n = 84, las que 
se encuentran en islas subantárticas, y en el sur de 
Chile e Islas Falkland respectivamente (Moore $ 
Walton, 1970). 


relación con las glaciaciones del Pleistoceno. donde 


Estas diferencias parecen tener 
hubo dispersión luego que los hielos se retiraron 
colonizando zonas boscosas del sur (Moore, 1972; 
Roulet, 1981). Se sabe que esta especie sólo hibridiza 
con Acaena tenera, y que esta falta de hibridación con 
otras especies se deba a que la mayoría de sus 
poblaciones en Sudamérica son probablemente tetra- 
ploides (Walton, 1979). 


En relación a su sistema reproductivo, 


>s una 
especie ginodioica (Arroyo & Squeo, 1989), aunque es 
posible encontrar algunas plantas con caracteres 
predominantemente masculinos, con grandes anteras 
y estigmas pequeños, pero no se conoce si son o no 
funcionales (Walton, 1979 


variar considerablemente, y puede presentar algunas 


— 


Su inflorescencia puede 


anormalidades como cupelas pediceladas o más de 


434 


Annals of the 
Missouri Botanical Garden 


cuatro espinas (Walton, 1977, 1979). Esta variabili- 


dad provocó que se describieran muchos taxa. como 


el caso de Bitter (1910) y Philippi ( (1862. 1862, 
ve 2, 1893), algunos de los cuales fueron. posteri- 
ormente pasados a sinonímia (Grondona, 1904; 


Walton, 1975). Esta especie corresponde claramente 
a un complejo y es probable que existan divisiones 
infraespecíficas reconocibles. 


ARGENTINA. Sta. 
iscachas, J. B... (HIP); 
a eae! Est. Sofía, J. B.. 5142 
e 2005 (HIP): 
. T Región: Arica. 


, Región: El Loa. 


Material representativo examinado. 
z: Güer Aike, Est. E 2537 
Rio Gallegos, J. B. U.. 
(HIP) eer del Fuego: Est. Cullen, 
bs Thetis. Goodall 2283 (HIP). CHIL 
Puquios, ion da & Morales 4 (CONC). 


Toconao, 7 Ene. 1950. Pfister s.n... (CONC); Socaire, 
Rodríguez 3184 (CONC); Quebrada de Socaire. Rodríguez 
€ Ruiz 3715 (CONC): Antofagasta, Vega Aguada de 


Choscha. Arroyo & Villagrán 831286 (CONC): Quebrada 
del Chaco, Arancio & Squeo 10296 (C ONC). III Región: 
Maricunga, 

Internac. 


Chañaral. Camino de Potrerillos a Salar de 
cardi l (CONC): 
del 
ONC); \cerillos, 

qe s.n. (SGO); Río del Estrecho, 
C). IV Región: Elqui, Baños del Toro. Ricardi el 
(CONC): El Squeo 68068 (CONC): Limart. 
La Laguna. Ricardi et E 22 (CONC): Choapa. 
Loca. Jiles 4215 (CONC). hace Los Andes, 
. Ricardi. 2900 (CONC); Zoellner 6795 
(CONC Ew 5 Santiago, Potrero Grande, 
222 1474 (( Cordillera. Baños Morales. A. 
Marticorena et a p (CONC). VI Región: Cachapoal. 
Sewell, Pennell 12333 (SSO Colchagua. Termas del Flaco, 
ey SC 


9 59 Cam. 
Ricardi et al. 1662 


15 Nov. 


Arancio et al. 


Ricardi et 
' Cuesta Colorado. 
\tacama, 


Desierto de 


Indio. 


Es 0 illo. 


15 Feb. 1986. Nuñez & Labra s.n. (GO). VIE Región: 
Curicó. Paso Vergara. C. Marticorena & Matthei 1007 
(CONC); Talca, Entre Laguna del Maule y Paso Pehuenches. 


Ricardi et al. 966 * ONC): Linares, Cordillera de Linares. FV. 
VIII Reston: Nuble, 


: Arauco. " iordillera de 


Termas de Chillán, 
Nahuelbuta, 
Pino del Finado, e 228 (C io. Camino de 
Ralco a Lepoy, Ae el ON N 
Región: Malleco, Entre 1 P. N. Nahuelbuta, 
( ONC); C | Liucura-Pino Hachado. km 19, 
5065 (CONC); Cautín. Ref 
Matthei 5305 (CONC): 
avena 29 (CONC). Región: 


JOY 


Taylor et 


Panguipulli, cam. Pullinque—Conaripe Marticorena et 
J. 429 Md : San José. Montero 126; 50 (C ONC): Osorno. 
P. N. Puyehue, Refugio Antillanca, Gardner & Knees 5619 
(CONC) bailas de: tto Montt. Los Alerces. 3 Feb. 
1960, Saa s.n. (CONC): Chiloé. Isla Doña Sebastiana. C. 
Marticorena 1616 (C 5115 ). XI Región: Coihaique. Lago 
Seco, Schlegel 2318 (CONC); Aisén. ! tafae 


(SGO); General Carrera. Lagi 
2460 (HIP); Capitán Prat, Glaciar 
inliot 6140 (SGO). XII Región: Ultima Esperanza. Est. La 
Pisano & Cárdenas 4044. (HIP): 


Porfiada. Sierra Baguales, 


Cerro Donoso, Río de Las Chinas. Arroyo et al. 870202 
(CONC) Torres del Paine, Cerro Diente, Arroyo & Squeo 
860084 (CONC); P. N. Torres del Paine. Laguna Mellizas 
Elvebakk 35 (CONC): Sierra del Toro. Arroyo et al. 92165 
(CONC): Magallanes. rial de Pali Aike, Dollenz 207 
220, 241 (HIP); Pto. 1 Seno Skyring, Pisano & 
Henríquez 60782 (WIP) Península Muñoz Gamero. Pl 
Ramírez, Pisano & Dollenz 5765 (HIP): Punta Arenas, 


Chabunco, 8 Ene. 1952, Pfister & Ricardi s.n. (CONC): Tres 
Brazos. 15 Ene. 1960. Saa s.n. (CONC) Bahía Morris. Isla 
Capitán Pisano 3301 (CONC, HIP): Tierra del 


Moore 2369 (WIP) Cuarto Chorrillo, 
Pisano 3169 (HIP); . Pisano 
China C a pst Por fin, 
( Se Baquedano, 5 
6901 (HIP): 
li 


útil, Est. 


“spora, 
one ome 
2552 (HIP); 
Río Ma Pisano 3210 (HIP): 
1074 (SGO); Antártica Chilena, 
ape 3106 (CONC. HIP); I 
Navarino, Puerto Toro. Dollen m P); Caleta Awaiakirrh, 
Isla Hoste. ISO. Pisano 3 ers 21 (HIP): Isla Wollaston. 
Caleta Lientur, Pisano 4969 (HIP 1 aleta Misión. Isla Hoste. 


Dunas 
. Moore 


Fiordo Pane C anoa, Isla 


Bahía Orange, 15 Nov. 1975, Samsing s.n. (HIP). ISLAS 
FALKLAND: West F. Fox Island. Skottsberg 52 (SGO). 
SOUTH GEORGIA: King Edward Point. Headland 562 
(HIP) 
Acaena masafuerana Bitter. Biblioth. Bot. 
17(74): 45. tab. 2. 1910. TIPO: Chile. “Juan 
Fernandez: Mas a fuera, una cum A. ovalifolia R. 
et P. var. insulae-exterioris Bitt. 1100-1300 m 
(expedit. suec. 1907—09)." Skottsberg s.n. (holo- 
tipo, S-R8057!). Figura OA, B. 


Planta hasta 9.5 em de alto, leñosa en la base: tallo 


cubierto por las vainas foliares: entrenudos cortos. 


Hojas de 1-7.5 em de largo: lámina en contorno 


obovado-lanceolada: 3—4 pares de folíolos. obovados. 


asimétricos en la base, 3-7 mm de largo. el margen 


INCISO. con una separación entre par y par mayor que 


el largo de los folfolos, con tricomas cortos y 


brillantes. en las nervaduras en ambas caras: vainas 


foliares glabras: apéndices estipulares de base 


ensanchada. de 0.5-2 mm. Rama florifera terminal: 


pedúnculo hasta 9.4 em de largo. glabro a glabres- 
cente: inflorescencia capituliforme, de 3.5-6.5 mm de 


diámetro: brácteas basales obovadas. el margen 


pubescente, de 1-1.5 mm de largo. con tricomas 


articulados de color anaranjado oscuro. brillantes, 
entre las brácteas \ las cupelas. Sépalos 4. ovados, de 
0.9—1.3 mm de largo. glabros, con manchitas de color 
2. anteras globosas, 


pm 


café rojizo a lo largo; estambres 


ca. 0.3 mm de largo: estilo corto, estigma globoso. 
0.4-0.6 mm de 


de 1-2 mm de largo, con tricomas glandulares entre 


aciniado. de argo. Cupela ovoide, 


y en la base de las cupelas. 4 espinas cortas de 


base angosta en el ápice. con fuertes eloquidios: 
: ? | 


aquento l. 


Icon. | Skottsberg, 1921: 131. 
Fenología. Fructifica de enero (Stuessy & Lammers 


9211) a febrero (Landero & Ruiz 9592). 


Especie endémica de 


Distribución y habitat. a is- 
a Más Afuera del 


(Fig. 12A). Crece en el fondo de las quebradas, entre y 


Archipiélago de Juan Fernández 


sobre rocas. donde se reconoce por sus sépalos 


amarillo-verdosos y estambres de color violáceo. 


Volume 93, Number 3 Marticorena 435 
2006 Acaena en Chile 


N 


SARD 
(e 
Q 


NO 
E 
SODA 
WT 
ANS 
— 
Ef) 


JAR 


AA 


NY 
Dy 


7 


Figura 6. Acaena masafuerana. —A. Habito, —B. Cupela (Valde 


benito y Landero 9030, CONC). Acaena patagonica. . 
Hábito (Arroyo & Squeo 870189, CONC). —D. Cupela (Elrebakk 532, CONC). 


) 


)m de — observó que presentaban células papilosas, carácter 


= 


formando cojines, desde los 900 a mas de 130 
altitud. que también observó en las especies A. pumila y A. 
Caracteres distintivos. Capítulos de hasta 6.5 mm exigua. ambas pertenecientes a la sección Subtuspa- 
de diámetro: cupela con 4 espinas apicales más cortas pillosae. También menciona la presencia de estolones, 
que la mitad del largo de la cupela: folíolos obovados. „ advierte que las anteriores especies también los 
asimétricos en la base, con tricomas glandulares entre poseen. Debido a que sólo contaba con material estéril, 
las cupelas. y a la similitud anatómica de las tres especies incluyó 
Acaena masafuerana pertenece a la Sección Acroby- a A. masafuerana dentro de la sección Subtuspapillosae. 
ssinae. Esta especie es afín a A. antarctica en cuanto Sin embargo, señala que las células papilosas de A. 


a su pequeño tamaño, la forma y tamaño de las cupelas. pumila y A. exigua están mucho más desarrolladas que 


forma de los folíolos, y los tricomas glandulares as de A. masafuerana, y mantiene muchas dudas sobre 
J e P4 R 
dispuestos entre las cupelas. La especie fue descrita su inclusión en la sección. Posteriormente, Skottsberg 
vor Bitter (1910) basado en material incompleto sin (1921) estudió material fértil. y pudo ver que la especie 
} ) | Y] I | 
flores ni frutos. Al estudiar la cara inferior de las hojas compartía muchos caracteres con A. antarctica y por lo 
J | y | 


436 


Annals of the 
Missouri Botanical Garden 


tanto la ubicó en la sección Acrobyssinae. Se observó 


además que A, masafuerana presenta el mismo patrón 
celular de la cupela (tipo axillares) que A. antarctica, A. 
tenera, A. lucida y A. pumila (Marticorena, 2000). 

Mat 


Glens. Juan Fernandez 


CHILE. 
s Afuera, N Quebrada Casas, 
Stuessy & Lammers a (( 1 Quebrada Pastos, Stuessy 
& Lammers 9416 (CONC): Cordón La Cuchara. 
& Landero 9050 (CONC); Cordón Atravesado, Stuessy et al. 
9305 (CONC): Plano de Rodríguez, Vale Jr 9132 
(CONC); Quebrada El Cane lo, Valde benito & Landero 6432 
(CONC); Entre Que bra 
et al. 9323 (CX 
(CONC); A p "de, SY 
Stuessy & Ruiz 9448 (CONC 


rial representativo ora V Región: 


g aldebenito 


ebrada Guaton, Stuessy 


(CONC); Quebrada 


Larga, 


Ll. ‘ 'aena ovalifolia Ruiz et Pav., Fl. Peruv. Chil. 
67. 1798. TIPO: Perú. Pillao “Huassa-huassi 
humidis et umbrosis," Ruiz et Pavón s.n. 


(lectotipo, designado por Romoleroux (1996: 69). 


MA [Herbarium Peruvianum, Ruiz et Pavón, 10/ 


80|D. Figura 5C, D. 

Acaena 179 0 Gay, . Chile, Bot. 2: 297, 1047. Acaena 
ovalifolia var. ee (Gay) Reiche; Anales Univ. Chile 
08: r 69. 1897. Acaena reine E 5 0 
(Gay) Bitter, Biblioth. Bot. 1774): 242 . TIPO 
Chile. “La encontramos en i Cerros a de la 
hacienda de Talcaregüe, en departame dn Pa San 
Fernando (1891), Gay 80 ( ra 5 1063 

leaena. tenuipila Bitter, Biblioth. Bot. 17(74): i 1910 
TIPO: Chile. “Prov. Valdivia. pl Het nsis, herb. 
Stockh., Upsal.." Z 


584 “eatin, aquí desig- 
nado, S-R8072!, 15 15 UE CONC-29697 8 


Acaena micranthera Bitter, Biblioth. Bot. 3 . 1910. 
TIPO: Chile. "Loco speciali non To n. Paris, i 
Gay s.n. (holotipo, P, foto 34723 en CA 1). 


20 em de 


2 mm de diámetro: 


Planta. hasta 


alto: 


tallos de | 


rizomas largos. Ca. 


m 


—2.5 mm de diámetro, 
Hojas de 2.3— 


3—5 pares de 


seríceos; entrenudos de 0.5-1 mm. 


IO em; lámina en contorno obovada: 
foltolos, oblongo-elípticos a elípticos. de 7.5-32 mm 


de largo, base oblicua, margen aserrado, crenado, 


pinnatifido, con 710 segmentos, pilosos a seríceos en 
el envés, especialmente en los nervios: vainas foliares 
pilosas; apéndices estipulares lineares, agudos, de 3— 
8.5 mm de largo, el margen levemente aserrado. Rama 
florifera terminal; el pedúnculo hasta 10.5 em de 
largo. piloso: inflorescencia capituliforme, rojiza en 
fruto, basales 
Flor 
l. oblongos, de 1-2.5 mm 
5 8 


Z—0 


de diámetro; brácteas 


de 2.5 


chasmógama con sépalos 3— 


hasta 40 mm 


lineares, laciniadas, mm «de largo. 


de largo, cara interna sericea: estambres 


filamentos de 1.24 mm de largo, anteras globosas, 


de 0.2-0.3 mm de largo: estilo de 0.5—1 mm de 


argo: 
estigma oblongoide, laciniado, de 0.5-1 mm de largo. 


Cupela obcónica, de 17.5-19 mm de largo, cubierta 


por tricomas retrorsos de color amarillo pálido, tiesos, 


caedizos, con E largas espinas de base angosta en el 


ápice, raro 3, de 14-15 mm de largo, con fuertes 
gloquidios; aquenio l. Inflorescencia cleistógama 


hasta 8 mm de diámetro: flores solitarias o en grupos, 


vedunculadas o sésiles: 1-1.5 mm 


— 


sépalos de de 


argo, estigma ca. 1.5 mm de largo, ovario ca. 0.5 mm 


de diámetro. 

Icones. Grondona, 1964: fig. 
1798 103, 
Fructifica de agosto (Rodríguez 


a marzo (Gunckel 15348) y en 


4—7; 


fig. c. 


1984: 


= 


fig. 449, 
Ruiz y Pavon, tab. 
Fenología. 


64.3) 


algunos casos se 


e 
puede encontrar con flores desde junio (Gunckel 
15632). 

Distribución y hábitat. Planta nativa, crece en sue- 
lo seco, a orillas de río o de 
IV 


Navarino). 


también en laderas 
Jorge) a XII 
de 5 a 3000 m de altitud 
Se encuentra como maleza en las islas Más 


Más Afuera del 


camino, 


húmedas, desde la 


Región (Fray 
Región (Isla 
(Fig. 111). 
a Tierra y Archipiélago de Juan 
Fernández. También crece Islas Falkland 
(Moore. 19608), 1999), Ecuador 
(Romoleroux, 1999) y Perú (Gereau, 1993). 


en las 


en Argentina (Zardini, 
Nombre común. 1921). 


Inflorescencia capituliforme 


Cadillo (Junge s.n.: Baeza. 


Caracteres distintivos. 


madura de color rojizo; eupela con pelos tiesos 


retrorsos y dos espinas apicales largas con gloquidios. 


Acaena ovalifolia pertenece a la Sección Ancis- 


trum. Esta especie es cercana a A. magellanica y 


A. argentea, con las que comparte en términos 


generales, el tipo de hábito. la inflorescencia y le 


l 
forma de la cupela. A diferencia de ellas. A. 
ovalifolia presenta una distribución mucho más 
amplia en América, llegando hasta Ecuador 


(Romoleroux. 1999). 


En relación a sus sinónimos, Acae "na 55 
Bitter, se examinaron especímenes de 5 y CONC que 
corresponden a Lechler 564. y se eligió como lectotipo 


a la muestra de S-R8072, porque presenta una planta 


con una infrulescencia madura. 


en lanto que la de 
CONC-29697 está poco desarrollada. Para las varie- 
dades Bitter, Bitler y 
subglabrescens Bitter, no fue posible localizar materia- 


dolichacantha inciserrata 


les originales. y se presume que fueron destruidos ya 


— 


que estaban depositados en Berlín. No obstante. al 
estudiar las descripciones, el autor basó sus varie- 
dades en caracteres muy variables, 
d. \ 


í ] 
(SUDELUOrescens), 


COMO la pubes- 


cencia 


el grado de división de los 
dientes de los foltolos (inciserrata) y en el largo de las 
espinas (dolichacantha). Tampoco se encontró mate- 
rial tipo para la variedad serrata, el que estaba en 


UPS, 


anteriores, 


sin embargo sucede lo mismo que para las 


fue establecida de acuerdo a caracteres 


Volume 93, Number 3 Marticorena 437 
2006 Acaena en Chile 
variables, en este caso los dientes agudos del folíolos. ^ pares de folíolos, de 3—6.5 mm de largo, obovado 


El tipo de la especie A. subtussericacens Bitter, estaba 
en B, por lo que se supone destruida. De acuerdo a la 
descripción la cupela presenta tres espinas, lo que 
a veces es posible observar en A. ovalifolia, además se 
señala que la cupela tiene pelos tiesos numerosos, un 


carácter propio de la especie. 


xa ARGENTINA. Santa 
Cruz: Glaciar iste no, Anliot 6141 (SGO); Güer Aike, Es 
Rospentek, T.B.P.A. 3263 (HIP). CHILE. 
Limarí, Talinay, Jiles 1685 (CONC). Y Región: Valparaíso. 
Juan Fernández, Más a Tierra, Villagra a El Mirador, Stuess) 
et al. 11777 (CONC); Más Afuera, Quebrada Sándalo, Stuessy 
& Lammers 9278 (CONC); N Quebrada Casas, Stuessy et al. 
9064 (CONC); Cordón 1 0 Stuessy & ned 9214 
(CONC). VI Región: Agua de la Vida, Hda. 
Cauquenes, 15 Oct 5 s.n. (SGO); ( 
Talca 'aregue, F. yI 


Material ı 


Cac 55 oal, 
; Colchagua, 
(SGO). I Región: Curicó, 
/ à Marticorena & 
Zoellner 


Termas de Chillán, Pfister 
15 Oct. 1965, 
(CONC); Arauco, Cordillera de Nahuelbuta, 
Cóndores, Matthei & Quezada 51 (CONC); 
Laraquete, 1 Mar. 1936, Junge s.n. (CONC 


: Ñuble, 


Cone epción, Chiguayante, 


: Biobío, Camino 
Marticorena et al. 299 (CONC). IX 
N. Nahuelbuta. 1 et al. 2085 
Montero 9987 (( 
Marticorena et al. 33 


de Ralco a Lepoy, A. 
Region Malleco, P. 
a : Termas Rio Blanco, M 2): Cautín, 
5 (CONC). 
Valdivia, "à „ Reducción Curaco, C. 
Marticorena et al. 423. (CONC); Comal Chaihuin, Gunckel 
15632 (CONC); 1 1 0 Quinchilca, Gunckel 15348 (CONC); 
P. N. Puy Refugio Rodríguez 1050 
Llanquihue, P. N. Vicente Pérez Rosales. 
Marticorena et al. n (CONC): rien Pu- 


22 


is Paraguas, A. 
Región: V 


pu. Puye hue, Antillanca, 


ONC): 


llinque, € . Marticorena 1582 (CONC); Camino entre Castro 
( honalit: pos 12 & a 730 (CONC); 1 
Punta Lapa, C. Mart orena et al. 81 (CONC); Palena, 
Isla Taleán, C i 1653 INC). XI Región: 
én, Islas Guaitecas, Melinka, Marti 


1096 (CONC); 
1 (CONC 


corena Camino de Pto. 
isen, Parra 
P. 18 General Carrera, Laguna Fon- 


un Pio Ud 221 (HIP). XII Región: Ultima I 
I 


. Cerro Diente, Ar 


Coyhaique, Coihaique, 
S de 


Porres del Pain i 

ONC): Est. Dos el “nero, Valle de las Chinas 
Cárdenas 1838 (HIP); Magallanes, Isla Riesco, Est. Río de 
& Ricardi. s.n. (CONC); 
aguna El Parrillar (al N), Dollenz 991 
Harris, Rodríguez 843 (CONC); 
Ricardi & Matthei 219 (CONC); 
frente 


los Palos, 4 Ene. 1921, Pfister 
Península Brunswick, | 
(HIP); Is 
Tierra del Fuego, 
Antártica Chi 
Rodríguez 670 (C 


1 Dawson, Puerto 


Vicuña. 


E 


ena, dela Navarino, islote Snipe, 


12. Acaena patagonica A. E. Martic., Novon 9: 
227. 1999. TIPO: Chile. “Ultima Esperanza, P. 
N. Torres del Paine, Cerro Diente, 900 m, 15. 
Dic. 1985," Arroyo € Squeo 850829 (holotipo, 
CONC-100000!). Figura 6C, D. 


Planta hasta 20 em de alto, con entrenudos cortos. 


Hojas lineares; lámina de 1.3-2.5 cm de largo; 5-8 


a obovado-orbicular, 3-5 partidos, segmentos lineares 


linear-lanceolados, margen revoluto, cara superior 


elabrescente, con tricomas blancos esparcidos, cara 


— 


inferior sericea, base cuneada; vainas foliares con e 
margen con tricomas separados, apéndices estipulares 
ausentes. Rama florífera terminal, pedúnculos hasta 
16 cm de largo, pubescentes a seríceos. Inflorescen- 
cia espiciforme corta, de 6.5 mm de diámetro, con 
pequeños glomérulos de flores y/o frutos a lo largo del 
pedúnculo; brácteas basales lineares a triangulares, 
de 1-1.5 mm, con el márgen partido. Flor chasmó- 
gama con 4-5 sépalos, lanceolados, de 1.5 mm, cara 
superior glabra, cara inferior serícea, con un mechón 
estambres 4, anteras 


de tricomas en el ápice: 


orbiculares, de 0.3 mm; estigma orbicular, laciniado, 
de 0.5 


de 1.5-2 


mm de diámetro. Cupela ovada a casi redonda, 
.5 mm, pubescente a glabra, 3—4 costillas, en 
las cuales se dispone una espina de 0.5-2 mm de 
largo, con tricomas retrorsos desde el ápice, dispuesta 
en el tercio superior de la cupela, a veces puede 
presentar otra espina rudimentaria bajo la principal. 


Aquenio I. 


Icon. Marticorena, A., 1999, fig. 
Fenología. | Fructifica de diciembre (Elvebakk 532) 


a febrero (Arroyo € Squeo 670169). 


Distribución y hábitat. Planta nativa, crece entre 
rocas, en sitios secos y también en vegas de altura 
en la XII Región (en la zona del Paine y en Tierra 
del Fuego), desde el nivel del mar hasta los 900 m 
de altitud (Fig. 12A). También en la Patagonia 
Argentina. 

Caracteres distintivos. | Folíolos 3-5 partidos, seg- 
mentos revolutos; cupela ovoide a redonda con 3—4 
costillas poco notorias terminadas en un espina con 
tricomas retrorsos en el ápice. 

Acaena patagonica pertenece a la Sección Patago- 
nica. Especie relacionada con A. pinnatifida, de la 
que se diferencia por sus folíolos con menos segmen- 
tos y por sus cupelas con 3—4 espinas en el tercio su- 
perior. Sin embargo, el patrón celular de la cupela es 
distinto, y la asocia más a las especies de las Seccio- 
Subtuspapillosae 


nes Pleuroc rm Acrobyssinae 


(Marticorena, 200( 


ARGENTINA. Sta. 
/iscachas, cono Pan de Azüca ar, 
Cabo Buen Tiempo. T.B.P.A. 

camp, Goodall ps 


Material jas examinado. 


(LP). CHILE. XII Región: ii Esperanza, Sierra 
Baguales, Cerro Santa Lucía, Arroyo & Squeo 670169 
(CONC); Torres del Paine, Cerro Diente S 


3 (CONC, Typus); Cerro Corona. 
. Pisano 5631 (HIP); 4-5 km SE de Est. ! 
Elvebakk 532 (CONC); Lago Sofía, Pisano & Henríquez 0827 


Annals of the 
Missouri Botanical Garden 


Lig 3 


Figura 7. 


—D. Cupela anda 7932 


(HIP); Puerto Natales, Muñoz s. ONC): Tierra del Fuego, 


Punta Delgada. Buque Quem: 1150 ns 195. 197 (HIP). 


Nu 
— 


13. Acaena pinnatifida Ruiz el Fl. Peruv. 
Chil. I: 68. 1798. TIPO: Chile. 
apricis el Gonceptionis 
Puntilla, 


Pavón s.n. 


“Habit lab in 
collibus 
Gavilan et Hualpen tractus,” Ruiz ef 
(lectotipo, aquí designado, MA 
| Herbarium Peruvianum, Ruiz et Pavón. 10/83], 
foto 29605 en CONCI: CONC- 
29692!). Figura 7A, B. 


duplicado, 


aena ps are da Hábito. —B. € pela (Taylor y Gereau 10907, CONC). Acaena trifida. — 


Chile, per 


cm 


i, 
* 
zu 
" 


C. Habito. 


leaena multifida Hook. Fl. Antaret. J: 265. 1847, TIPO: 
Chile. “Hab. Strait d Magalhaes: Port Gre gory.” Capt. 


King s.n. (holotipo, K, foto 1225 en SGO!) 

Linnaea 33: 60. "M 
pinnatifida var. oligacantha ó hil.) Reiche. 
Univ. Chile 98: 165. 1897. Ace 
o (P hil. ) Bitter, 
1910. TIPO: Chile. 
Arañas, Ene. 1861,” 

D; 


leaena 
Anales 
na d T 
Biblioth. łot. 17(74): 13 
“In Andibus prov. 


leaena. oligacantha Phil., 


Santiago, Cor 


7 


Landbeck s.n. Cuota: 


leaena. calcitrapa Phil... Linnaea 33: 67. 1864. Acaena 


pinnatifida var. e e (Phil.) Reiche. Anales Univ. 
Chile 98: 166. 1897. TIPO: Chile. “In Andibus de 


Volume 93, Number 3 
2006 


Marticorena 
Acaena en Chile 


Linares et de Talcaregue dictis. crescit,” colector 
desconocido (neotipo, aquí designado, SGO-49855!). 
Acaena euacantha Phu nales Univ. Chile 84: 620. 1893. 
Acaena „ var. 9 (Phil.) Reiche, Anales 
Univ. Chile 98: 165. 1897. TIPO: Chile. “In Andibus 
provinciae Santiago lecta est, Mar. 1883,” Philippi s.n. 
(lec n aquí designado, SGO-49936!; duplicado, 
SG0-39724!). 
Icaena leptophylla Phil., Anales Univ. Chile 84: 622. 1893. 
Acaena o var. d Pw ) Reiche. 
Anales : 166. 18 


procul a thermis ^ Chillan ae invenit orn.. 


Univ. Chile € Chile. "Haud 


sub 


nom. 4. myr 5 Ene. 1883." A. Borchers s.n. 
(holotipo, SGO-49838!) 

Acaena longifolia Phil., mm Univ. Chile 84: 624. 1893. 
Acaena „ var. Wo (Phil. Reiche, 


Anales Univ. Chile 98: 166. 18 p na a 
subs longifolia mi Bitter, Aen Bot. 17(74): 
127. 1910. TIPO: Chile. “In Andibus provinciae 
Valdivia habitat, um Queni dicto; Ene. 1887.“ 


Philippi s.n. (holotipo, SGO-49840!). 
Acaena e i subsp. hypoleuca Bitter, Biblioth. Bot. 
17(7 121. 1910. TIPO: Chile. “In collibus ad San 
myriophylla Lindl., 
a. 


Jerol.. 


e 


sub nom. 4. herb. I 


Brem.. Stockh., Ene. 185 Philippi 

(lectotipo, aquí designado. BREM!: duplicado, Pl, 
R8062!). 

leaena pan var. da Alb 


17(74): 122. 1910. TIPO 


Bitter. Biblioth. Bot. 
Chile. “Prope urbem Valdivia 
A myrioj hylla | Lindl.-Schlecht., 
Lechler 294 (holotipo, P-106290!). 
leaena pinnatifida subsp. nace Bitter, Biblioth. Bot. 
1774): 122. 1910. TIPO: Chile. “Patagonia occidenta- 
lis: Vallis fluminis herb. ‘kh. et 
. Dusén s.n. (lectotipo, aquí 
; no visto). 
leaena. pinnatifida var. 55 Bitter, Biblioth. Bot. 
17(74): 125. 1910. TIPO: Chile. la der 
chilenischen 1 33 südl. i 
den Bergen, (Baenitz, herb. Americ. nr. 1125 
Brem. et Lauson. neque numerus in herb. Vratilsav.). 


sub nom. herb. Paris." 


Ap n. Berol., Stoc 


— 
[97] 
a 


29 Ene. 1903." O. Buchtien s.n. (lectotipo. aquí 
designado, BREM!: duplicado. M!). 

leaena O var. macrura Bitter, Biblioth. Bot. 17(74): 
129. TIPO: Chile. “Prov. Colchagua, San 
Fer il n. collibus, lieri "rs, Dec. 1830.“ Gay 
81 elende 51062955). 

Acaena rhe 9 3 bu r. Biblioth. Bot. 


: Chile. 
ise. 


17(74): 130. 1910. * s pelouses des 
hautes 1 ps 
basaltique, frue 5 ra, herb. Paris. Febr. 
(holotipo, P-1062 

Acaena MGE var. 


17(74): 133. 1910. 


commune, terrain 


1831," Gay 76 


Bitter, Biblioth. Bot. 
“Hacienda Cauquenes, 


vile rmedia 


TIPO: Chile. 


Cajon del n ypres, an den Quellen" sub nom. Acaena 
cfr. Poeppigiana Clos det. R. A. Philippi, herb. 
Monac." Philippi s.n. (holotipo, M!). 


leaena „ subsp. heterotricha Bitter. Biblioth. Bot. 
7(74): 133. 1910. TIPO: Chile. “Cordillera de 

ci ac Phil. et 
ibuit A. Philippi sub 
* Philippi s.n. loi W-116066 . 

3iblioth. Bot. 
“Habitat f 
Magell. ; 


Santiago, una cum A. 


me 


giana Gay distr 
oligacantha Phil. 

\caena 1 subsp. intercedens Bitter, 
17(74): 143, tab. 14b. 1910. TIPO: CUN 
Magellanici angustias: Punta Arenas. pl. 


arenosis maritimis sub nom. ‘A. pinnatifida R. et P. — 


Grisebach,’ herb. Paris, Ups. Lechler 1149 (lectotipo, 
jut designado, P-106293!, duplicado, P-106300)). 
1 tii t quinquefida Bitter, Biblioth. 
TRE (4): 143. 1 TIPO: Chile? “Regio Magellanica, 
9 (holotipo. BREM!). 
Biblioth. 


Hab. in apertis 


unta d e 
Acaena 1 var. utrinquepilosa Bitter, 
iden 1910. TH 
95 Famine (Détroit de 1 voyage 
press et de la Zélée, 1841), Paris, 
(holotipo, P- 1063021). 


Chile. 


verb. 


Hombron s.n. 


Planta hasta 73 cm de alto. entrenudos cortos. 
Hojas 4-18 em de largo: 8-10 pares de folíolos 


ovado-oblongos, de 6-21.5 mm, pinnatisectos, con 5— 


7 pares de segmentos linear-lanceolados, revolutos en 


las hojas adultas, cara inferior sericea, superior 
pilosa, las hojas jóvenes completamente seríceas; 
vainas foliares vellosas en el dorso: apéndices 


estipulares ausentes; pecíolo notorio. Rama florífera 


terminal; pedúnculo del tamaño de la planta, pilosos 


con tricomas blancos, delgados, blandos: inflorescen- 


cia espiciforme, en general laxa, de 2.3-11 cm de 
largo, hasta 45 cm e interrumpida en fruto: brácteas 
basales lineares, de 1.5-2.5 mm de largo, pubes- 
centes. Sépalos 4-6, oblongo-lanceolados, de 1.5- 
3 mm de largo, cara interna glabra, cara externa 


serícea, con un mechón de tricomas duros en el ápice: 
2-6, 


anteras globosas, de 0.5—1 


estambres filamentos ca. 2.5 mm de largo, 


mm de largo; estilo corto. 


mm de diámetro. 


estigma globoso, laciniado, ca. 


Cupela ovoide, de 4—10 mm de largo. vellosa 


a tomentosa, cubierta por espinas largas, finas. de 


amano similar, sin gloquidios, con tricomas retrorsos; 


aquenio l. Flores cleistogamas acompañadas por un 


bráctea con forma de hoja pequeña o folíolo, 


pinnatisecta; espinas desde reducidas a largas como 
en las flores chasmógamas; flores y frutos de la espiga 
pasan gradualmente a ser flores o frutos cleistógamos: 
cupelas en poca cantidad, generalmente una. 

4549; 1984: fig. 453. 


fig. c. 


Grondona, 1964: fig. 
1798: 104, 


Fructifica de octubre (Garaventa 6579) 


Icones. 
Ruiz y Pavón. tab. 

Fenología. 
a abril (Parra & Rodríguez 166). 

Distribución y hábitat. Especie nativa, crece en 
suelos arenosos, secos, entre piedras, sobre escoria 
volcánica, también en lugares sombríos y sitios 
húmedos. Habita desde la IV Región (Cordillera de 
Ovalle) a XII Región (Isla Navarino), desde los 5 a más 
de 3000 m de altitud (Fig. 


Cadillo (Grandjot s. I.). 


12B). También en Argentina. 
Nombres comunes. amor 
seco (Pfister s.n.), pimpinela (Pfister 936). pimpinela 
cimarrona (Schwabe 179; Baeza, 1921), cepacaballo 
(Jiles 1585). 
Caracteres distintivos.  Folíolos pinnatisectos con 


= 


5 a 7 pares de segmentos lineares; espiga larga, algo 


péndula. 


Annals of the 
Missouri Botanical Garden 


Acaena pinnatifida pertenece a la Sección Acae- 
na. Es una especie muy variable, y al igual que el 


tratamiento realizado por Bitter para A. magellanica, 


el autor describió numerosos taxa basados en 
caracteres variables. 
Se estudió dos materiales tipo para Acaena 


pinnatifida, de MA y CONC, el primero fue elegido 
como lectotipo (foto 29605 en CONC), y 


una planta completa bien desarrollada. 


comprende 


Para e 
sinónimo A, calcitrapa, no se encontró material tipo, 
sin embargo se seleccionó la carpeta de SGO-49855 
como neotipo, ya que aunque el año es 1896, posterior 
(1864), 


características de la planta, además la etiqueta está 


a su descripción corresponde con las 
determinada y escrita por Philippi como A. calcitrapa, 
y proveniente del mismo lugar. Para A. euacantha, se 
examinaron dos materiales depositados en SGO, 
ambas muestras estériles y con sólo una hoja, pero 
se eligió como lectotipo SGO-49936, debido a que es 
una hoja más completa. Para A. pinnatifida id 
se estudiaron los tipos de BREM. P y S. Se 


examinaron dos carpetas BREM, ambas de 11 


hypoleuca, s 


327, pero una con la fecha Ene. 1853 como se senala 
en a protologo, y la otra de Dic. 1854. Esta ültima 
parece ser dudosa, por lo que se descartó como 


material original. Se eligió como lectotipo la muestra 
de BREM que corresponde con la fecha, y además es 
una planta bien desarrollada con muchas hojas. Para 
l. pinnatifida subsp. nudiscapa, se eligió como 
lectotipo S-R8064 que corresponde a dos plantas 
completas y maduras. El material tipo de UPS no se 
vió. De la variedad uspallatensis, se estudió material 
original de M y BREM, del Buchtien, del Herbarium 


Americanum del Dr. Baenitz. con el número 1125. 


et 


se eligió como lectotipo la muestra de BREM, la que 
además de estar completa, tiene la etiqueta original de 
Bitter. 


canum y el mismo colector, 


Bajo el mismo número del Herbarium Ameri- 
se utilizó una muestra que 
Bitter describió como A. pinnatifida subsp. pliacantha 
Bitter y su variedad depauperata Bitter, el que estaba 
depositado en WRSL, pero que no fue localizado. No 
obstante, por las características en la descripción 
original y por corresponder a la misma muestra, son 
reconocibles como A. pinnatifida. Para A. multifida 
subsp. intercedens, se examinaron dos materiales de 
Lechler 1149 depositados ei 


lectotipo la carpeta P- 106293, que tiene una planta 


se designó como 


completa con bastantes hojas. 
Numerosos materiales tipo fueron depositados en 


Berlín, 


los que presumiblemente fueron destruidos 


durante la Segunda Guerra Mundial. Entre estos 
tenemos Acaena pinnatifida subsp. brevispica Bitter 
con las variedades longiuscula Bitter y pallidiflora 


Bitter, A. 


gracilescens Bitter, 
¿ 


innatifida subsp. hypoleuca Bitter var. 
/ l | M 


A. pinnatifida subsp. mansoénsis 


Bitter, 


pinnatifida subsp. mollis Bitter var. humilior Bitter. 


A. pinnatifida subsp. membranacea Bitter y A. 


Otros taxa de los cuales no se pudo localizar material 
tipo fueron A. montana Phil., material no localizado 
en SGO: A. oet subsp. mollis Bitter var. elatior 
Bitter en G, y 


el protólogo en Z. Para A. multifida subsp. multi- 


l. multifida subsp. abavia Bitter según 


glomerulans Bitter, con A. multifida subsp. alaticon- 
nata Bitter y A. multifida var. ramificans Bitter, no se 
localizó material tipo, el que según los protólogos se 
encontraba en Z, y en Z y HBG respectivamente. 
Grondona (1964) las reconoció como sinónimos de 4 
pinnatifida, criterio con el que se concuerda en esta 
revisión. 

Es interesante señalar que Bitter tenía un concepto 
estrecho, 


de especie muy lo que provocó que se 


deseribiera numerosos taxa sobre la base de carac- 
teres. variables. i.e. la consistencia de las hojas 
(membranacea), el color de los sépalos y las anteras 


del 


(gracilescens), por ser suave y alta (mollis var. elatior). 


illidiflora), a lo delgado y lareo escapo 
n | 8 b 8 | 


También se observó la presencia 
de inflorescencias ramificadas (CONC 143890) lo que 


hizo deseribir a Bitter (1910) la variedad ramificans. 


entre mucho otros. 


Material representativo examinado. ARGENTINA. Neu- 
quén: Lacar, cerro Chapelco, Cabrera 23010, 23023 (LP); 
Rio Negro, Pampa del Toro, Cabrera 20506 (CONC). Tierra 
del Fuego: Tennessee oil camp, Goodall 140 (LP). CHILE. 
IV Región: Cabreria, Jiles 1585 (CONC): Cordillera 
de Ovalle, Río Tascadero-El Polvo, Jiles 6370 (CON( 

San Felipe. J 


P 


Limarí 


Valparaíso, Quintero, 
(CONC): Quillota, 


x 


utc . Gu 
Garaventa 0379 (C ONC 
Región Metropolitana: "a Cuesta de Chacabuco, C. 
612 (C 


Cerro Cruz, 


ires «€ Weldt ONC): Cerro Abanico, Nov. 
932. Grandjot s.n. (C ONC ): K 1 1 ra, Baños Morales, A. 
Vs rei et al. 


179 (CONC): 


; Refugio Lo Valdés, Schwabe 
& Gereau 10907 
: Cachapoal. Termas de C 
ister s.n. (CONC): 
nr 74168 (CONC): Cardenal Caro, Pichilemu, Saa 
ONC). VH A Curicó. Camino de El 

icorena et al, 162 (CONC): T. 


camino al Melado, III Región 
Nuble, Termas de 1 Pfister 930 eum y Bi 


C amino de Ral CO d | po 


Nacime ento, Fundo EL 17 7 1 1 No 950, Pfister s.n 
(CO IX Región: Malleco, Entre Angol y P. 
1 itn Taylor et al. 10383 (CONC): Termas de 


Manzanares a Lonquimay, km 26, Ricardi & C. 
5007 (CONC): 
CONC) X Región: 
Philippi. s.n. (S 
E " 


Cautín. Volcán Llaima, Montero 


Valdivia, Huahum, Ene. 7 
Ranco. Queni. 


13.4 km SE de 


Osorno, Erud de 


; Pale Da. 


camino La Tapera, 
. Río Exploradores. Seki 


310 (CONC) General Carrera. S of R. Ibanez. Villa Cerro 


Volume 93, Number 3 Marticorena 441 
Acaena en Chile 
$ 


Figura 8. 


—D. Cupela (Jiles 4438, CONC). 


Castillo, Braynshaw 202 (HIP) XII Región: Ultima 
Esperanza, Cordillera de Paine, Von Bohlen & Cavieres 337 


(CONC) ecion de Cerro Toro, Pisano 4121. 1122 (HIP): 
Lago Sofía. Pisano & Henríquez 6843 (HIP): Magallanes. 


Pd oe 8446 (CONC): Gordonier. Pisano 
erra del Fuego, Río C 


Antártica Chilena. 
Navarino, Pus 264 (CONC). 


a e es 


rico, Pisano & Henríquez 


Williams. Isla 


Puerto 


Speg., Revista Fac. 
Nac. La Plata 3(30-31): 515 
Argentina. “Santa Cruz, Golfe 
1896. 


designado, 


14. Acaena  platyacantha 


Univ. 
TIPO: 


Jorge. 


Agron. 
1897. 
San 

(lectotipo, 


ra BA, B. 


de 


" 


C. Ameghino sı 


LP-126921). 


anno 


— 


aquí Figu- 


Acaena platyacantha. —A. Habito. —B. Cupela (Elvebakk : 


. CONC). Hábito. 


Acaena poeppigiana. —C 


Planta erecta, hasta 20 em de alto. raro más. tallo 


rojas; entrenudos cortos. Hojas hasta 
5-8 


argo, 


cubierto por las 


2 cm de largo: lámina en contorno oblonga: 


pares de folíolos, obovados, de 3-12 mm de 


pilosos a seríceos con tricomas blancos, tripartidos, e 
margen revoluto; segmentos lanceolados, el central de 


mayor tamaño que los laterales: vainas foliares con 


tricomas largos y blancos en el margen, el dorso 


piloso: apéndices estipulares ausentes. Rama florífera 


terminal: pedúnculo hasta 9.2 cm de largo, raro más, 


sericeo, con tricomas blancos. blandos, largos, 


delegados: inflorescencia capituliforme, hasta 19 mm 


de diámetro, con flores aisladas a lo largo del 


lineares a ovadas, de 


pedúnculo; brácteas basales 


442 


Annals of the 
Missouri Botanical Garden 


—3 mm de largo, pubescentes. Sépalos 5. oblongo- 
elípticos, de 1.5-2.5 mm de largo, glabros en la cara 
interna y pubescentes en la externa, con un mechón 
ápice; estambres 2-3, 


de tricomas blancos en el 


filamentos ca. 0.5 mm de largo, anteras globosas, ca. 


laciniado, ca. 


0.5 mm de largo; estigma globoso, 


0.7 mm de diámetro. Cupela ovoide-oblongoide, de 3— 
7.5 mm de 


largo, tri-tetragona con ángulos aliformes, 


cubierta de tricomas cortos, duros, con ! (raro 2) 
espina de base ancha en cada ángulo y dispuestas en 
el tercio superior de la cupela, con espinas pequeñas 
entre las espinas mayores y generalmente rugosa entre 
los ángulos con pequeñas áreas hundidas, de forma 


irregulz ar y y de color oscuro; aque nio l. 


lcones. Grondona, 1964: fig. 37—40; 1984: fig. 
450. 

Fenología. | Fructifica de diciembre (Landero 671) 
a febrero (Zoellner Eo 

Distribución y hábitat. Crece en lugares secos 


desde Malleco, Paso Pino Hachado (IX Región) 
Seno Skyring (XII Región), desde alrededor de 


o0 a 1290 m de altitud (Fig. También en 


Argentina. 


126): 


Caracteres distintivos. Cupela tri a tetrágona, con 


ángulos aliformes, con zonas hundidas de forma 
irregular entre los ángulos. 

Icaena. platyacantha pertenece a la Sección Acae- 
na. Spegazzini (1897) en su descripción original 


indica tres formas para la especie: typica, elata y 
villosa, pero no las tipifica. En el protólogo menciona 
dos lugares, para Río Santa Cruz, y Golfo San Jorge. 
depositados en LP. Estos materiales corresponden 
a dos clases de plantas algo diferentes. una con la cara 
superior de las hojas oscura y glabra, margen revoluto. 


cupela con alas O costillas con dos espinas super- 


puestas, subhorizontales, pubescencia corta. la que 


correspondería con la forma elata. (LP-11942). Los 
otros materiales (LP-12682, 12577). podrían pertene- 


cer a las otras dos formas. con pn verde claro, 


cara superior de las hojas pubescente, alas de la 
cupela elíptica a ovada, con una a dos espinas. He 
podido revisar materiales que presentan caracteres de 
ambas formas. los que hacen difícil definirlas como 
reconoció Spegazzini (1897), por lo tanto se considera- 


ron como una sola especie sin formas. Como material 


lipo se eligió el espécimen LP-12682, ya que parece ser 


a forma más frecuente. 


ARGENTINA. Santa 


ischs, cerro Las Viscachas. 


Material representativo e 'xaminado, 
Cruz: Güer Aike, Est. Las 
J.. . I. 2597 (HIP) Colonia 
110 Von Thüngen 102 (LP); $ 
Rio Gallegos, año 1882, Spegazzini s.n. (LP-11942, 12577 


Carlos Pellegrini Est. La 


Santa Cruz, Bahía Gregorio. 


CHILE. IX Región: Malleco. Paso Pino Hachado, A. 
Marticorena et al. 42] NC ONO). XI Región: Aisén, Cerro 
Castillo, Zoellner 7566 (CC C) General Carrera, Chile 
Chico, 12 Die. 1954, Pfister s.n. (CONC). XII Región: 
Ultima Esperanza, 4-5 km SE Est. Cerro Guido, Elvebakk 
537 (CONC): Sierra de los Baguales. Est. la Cumbre, Landero 


671 (CONC): M; 9 0 S. Seno Skyring. 8 Ene. 1952, Pfister 


& Ricardi s.n. (CC 


Chile 2: 284. 


“Forma céspedes en las 


15. Acaena poeppigiana Gay, Fl. 
TIPO: Chile. 
cordilleras de Sotaquí (provincia de Coquimbo) 
a una altura de 9 a 10,000 
(holotipo, P- 10603189). Figura 8C, D. 


l7 em de 


1847. 
Gay 527 


ples. 


alto; entrenudos 


lanta. erecta, hasta 
Hojas hasta 


5—0 


pares de folíolos, ovados, de 3-3.5 mm de largo. el 


cortos, cubiertos por la vainas foliares. 


T em de largo: lámina en contorno lanceolada: 


margen partido, con 3—5 segmentos oblongo-elípticos, 
glabrescentes a seríceos: vainas foliares con tricomas 
blancos y blandos en el margen: apéndices estipulares 


ausentes. Rama florifera terminal; pedúnculo hasta 


Sem de largo, piloso: inflorescencia espiciforme 


corta, hasta 3cm de largo en fruto, a veces 


presentando flores a lo largo del pedúnculo: brácteas 
53.5 mm de largo, 


basales ovadas a lineares. de 2 


pubescentes en los márgenes y el dorso. con Iricomas 


le color amarillo pálido, blandos. Sépalos 4. oblongo- 


elípticos, de 0.71 mm de largo: cara interna glabra. 
externa con tricomas blancos y un mechón en el ápice: 
filamentos de 0.3-2.3 mm de largo. 
0.3-0.5 mm de 


corto; estigma globoso, laciniado, de 0.3-0.5 mm de 


estambres 2-5: 


anteras globosas, de argo; estilo muy 


diámetro. Cupela ovoide-oblongoide, de 2-4 mm de 


largo, tritetrágona. cubierta por tricomas cortos, 


hirsulos, con 2-3 espinas grandes de base ancha en 
0.5-1. 


ángulos espinas pequeñas: aquenio I. 


cada angulo, de ( 5mm de largo y entre los 
Grondona, 1964: fig. 41-44: 1984. fig. 452. 
Fructifica de noviembre (Jiles 2110) 


a abril. (Pisano 2545). 


leones. 
Fenología. 
habitat. 


Distribución Crece al borde de vegas, 


— 


orilla de lagunas, en suelo arenoso, formando champas. 
Habita en la IV Región (Limarí). en la VI y XH Región, 
desde 80 a 3500 m de altitud (Fig. 12C). También en 
Argentina. 

los foltolos 


Caracteres distintivos. Segmentos de 


ovado-oblongos: cupela con 2 a 3 espinas en cada 
ángulo. 

Acaena poeppigiana pertenece a la Sección Acae- 
na. Vs una especie algo variable que suele confun- 
dirse con A. platyacantha cuando esta crece en 
condiciones poco favorables, sin embargo los folíolos 
presentando segmentos ovados 


son inconfundibles, 


Volume 93, Number 3 Marticorena 443 
2006 Acaena en Chile 


19, Acaena pumila, —A. Há 
Cupela (Marticorena & Matthei 599, Ct 


Figuri 


bito. —B. Cupela (Pisano et al. 8639, CONC). Acaena sericea. —C. Hábito. —D. 
)NC). 


Ene. 1886, Germain s.n. (SGO). XH Región: Ultima 

herbicen: Esperanza, P. N. Torres del Paine, Laguna Mellizas, Elvebakk 
34 (CONC): Tierra del Fuego, Est. Brazo Norte, Pisano 2545 

Material representativo examinado. ARGENTINA. Santa (CONC, HIP). 

Cruz: Colonia Carlos Pellegrini, Est. La Flora, Von Thüngen 

17 (LP). CHILE. IV Región: Limarí, Cordillera de Ovalle, 

Río Molles, Jiles 21 10 (CONC): Potrero Grande, Ramadilla, 

Jiles 4438 (CONC). VI Región: Cachapoal, Sewell, Pennell 

12341 (SGO). VII Región: Linares, Cordillera de Linares, 


característicos, además de un aspecto general más 


p 


"Le 


16. Acaena pumila Vahl, Enum. Pl. 1: 298. 1804. 
TIPO: Chile. “Habitat ad fretum Magellanicum. 


Annals of the 
Missouri Botanical Garden 


Ex herbario Jussiaei, Oct.? 1767.” Commerson 


n. (holotipo, P-106320!). Figura 9A, B. 


Acaena glaberrima Phil., Linnaea 28: 685. 18 Acaena 
mila var. ge n hil.) Bitter, 1 0 Bot. 
170674): 41. 1910. TIP hile. “In insulis Chonos legit 

orn., 18 Ene. fe 12 Fonk 105 (holotipo, SG 


Biblioth. Bot. 17(74): 


Acaena pumila var. acrocoma Bitter, 


40. 1910. TIPO: Chile. “Adhuc solum e regione 
Magellanica nota: Paroni australis: Puerto Bar- 
row...; Peninsula Brunswick; Port Famine, herb. 


Lambert (herb. Upsal),” Lambert s.n. (holotipo, UPS!). 


cubierto 
3.5- 


7.5 em de largo; 9-12 pares de folfolos, orbiculados, 


Planta hasta 30 em de alto; tallo erecto, 


por las vainas. Hojas en contorno lineares, de 


base asimétrica, de color verde grisáceo, de 3-7.5 mm 


de largo, el margen revoluto, lobulados, glabros; 


vainas foliares glabras; apéndices estipulares au- 


sentes. Rama florífera terminal; pedúnculo de 17.5— 


29 cm de largo, glabro a glabrescente; inflorescencia 


brácteas 
con 2-3 


4, ovado-orbiculares, de 


axa, ca. l.5 em de 


de ] 


lóbulos desiguales. Sépalos 


espiciforme, largo: 


basales glabras, 22 mm de largo, 


0.8-1.2 mm de largo, rojizos, glabros; estambres 2—4, 


filamentos ca. 0.2 mm de largo, anteras globosas, c: 


0.2 mm de largo; estilo ca. 0.3 mm; estigma globoso, 
laciniado, ca. 0.2 mm de diámetro. Cupela ovoide, de 
2-2.5 mm de largo, glabra, cubierta completamente 
por espinas cortas de tamaño similar, con gloquidios; 


aquento l. 


Icones. 


Grondona, 1964: fig. 50-51; 1984: fig. 454. 
Fructificación de diciembre (Pisano 
0558) a marzo (Henríquez & Palma 98). 

y hábitat. 


Fenología. 
Distribución Crece en cojines en zonas 
pantanosas, vegas y turbales, desde Valdivia (Cordillera 
Pelada de Chaihuín) a la XII Región (Isla Hornos), de 
los 5 a más de 1000 m de altitud (Fig. 12D). También 
crece en las Islas Falkland (Moore, 1968). 

Caracteres distintivos. | Folíolos orbiculares, lobu- 
lados, cara inferior con células papilosas; hojas con 
9-12 pares de folíolos; cupela cubierta de espinas 
cortas, gruesas, con gloquidios. 

Acaena pumila pertenece a la Sección Subtuspa- 
A. Gray 


anterior por la 


— 


pillosae, la que también incluye a A. exigua 


de Hawaii, que se diferencia de la 
ausencia de gloquidios en las espinas, y un menor 
número de folíolos. Bitter (1910) incluyó además a A. 
masafuerana dentro de la sección, pero posterior- 


mente se estudió material fértil, que la ubicó dentro 


de la sección Acrobyssinae. 

CHILE. X Región: 

stero, Gardner & Knees 4108 
Junge 328 (CONC). XI 


Material repre sentativo e a 
Valdivia. Mon. Nat. Alerce 
(CONC Chiloé. 1 


Región: Capitán Prat, Caleta Tortel, parte alta en una 
Teneb 4077 (CONC). MI Región: 
Ultima Esperanza, Isla Wellington, Fiordo Triple, Dollenz 


1106 (HII 


Pisano 6558 (CONC 


turbera, Rodr IU 


es 


?: Bahía Micaela, Isla Saumarez, Canal Grappler, 
. HIP): Magallanes, Fiordo Nevado, Isla 
Santa Inés, Bahía Meus real, Pisano & Henríquez 6633 (HIP): 
Tierra del Fuego, Lago Fagnano. orilla oeste, Skousberg 225 
Parry, P. N. Alberto. de 
8639 (CON > 

Contralmirante Martínez, braz , SE. 
98 (CONC); Puerto W illiams, 
Schlegel 8197 (CONC); Isla Navarino, 
405 (CONC); Isla Grevy, Rada 


Fiordo 
Pisano et al. 


Norte, Gretton, 


Pisano 5389 (HIP): Isla Hornos, Cabo de p Dollenz 

016 (HIP). 

17. Acaena sericea J. Jacq., Ecl. Pl. Rar. 1: 81, tab. 
55. 1816. TIPO: “Acaena sericea. HB 1812. 


Herbar. Jacquin fil." (neotipo, aquí designado, 


W!). Figura 9C, 


Acaena cuneata. Wook. et Arn., Bot. Misc. 3: 307. 1833. 

TIPO: Chi le. “Cordillera of Chili, Cu uming (N. 300), 
> Agua y Cordillera (1832).” Predio 1 
signado. GL, foto 3689 en SGO!; 
SGO). 


isotipo, K, foto 1228 en 


Acaena Tugi Phil., Anales Univ. Chile 84: 620. 1893. 

Argentina. “In Fuegia orientali musei 

1 0 e Pablo Ortega s.n. (lectotipo, aquí e 
560-49831*; duplic 1 SC Q-39704!), 

Planta hasta 35 em de alto; rizoma leñoso; 


entrenudos cortos. Hojas en contorno oblonga, de 4— 
10 em de largo: 3-5 pares de folíolos de color verde 
oscuro, obovados, de 7-13 mm de largo, base 
cuneada, ápice inciso con 3-5 segmentos, pubes- 
centes a sericeos; vainas foliares de margen piloso: 
apéndices estipulares ausentes. Rama florífera termi- 
nal, hasta 31 em de largo, pilosa; pedúnculo de 12— 
zem de largo: inflorescencia capituliforme, hasta 


15 mm de diámetro; brácteas basales ovado-lanceo- 


adas, de 1.5-4 mm de largo, cara externa pubes- 
cente, cara interna con tricomas largos, 


].4— 


brácteas 


glabra, 


amarillos. Sépalos 5-6, oblongo-elípticos, de 


3 mm de largo. pubescentes como las 


basales, la cara interna brillante; estambres 2-5; 


estilo ca. 0.3 mm de largo; estigma globoso, laciniado, 
de 0.5- 


piriforme, de 3.5-5 mm de largo, glabra a pubescente 


0.7 mm de diámetro. Cupela obovoide, 
con tricomas cortos, con 4-5 costillas y 4-5 espinas 
duras en cada una, de base ensanchada, más largas en 
la parte superior de la cupela, con gloquidios; aquenio 
en ramas floriferas basales 


|. Flores cleistógamas, 


dispuestas en vainas foliares en número de 1-5. 


Icones. Grondona, 1964: fig. 58-60; 1984: fig. 
[e] o 

451. 
Fenología. Fructifica de noviembre (B. Parra 


476) a enero (Marticorena & Matthei 599). 


Distribución y hábitat. Planta nativa, habita en la 


Volume 93, Number 3 
2006 


V Región (Los Andes) a la Región Metropolitana (Cajón 
del Maipo), y en la XII Región, desde 200 a 2600 m de 
altitud (Fig. 12D). Crece en suelo arenoso y en la zona 
sur en estepa de Mulinum (Pisano € Cárdenas 4818), 
entre pastos duros y arbustos, como rosetas aisladas. 
También en Argentina. 

Caracteres distintivos. Cupela obovoide, con 4-5 
costillas y 4-5 espinas en cada una, cortas, de base 
ensanchada. 

Acaena sericea pertenece a la Sección Acaena. Jac- 


(1816) 


cultivado en el Jardín Botánico de París, y señala que 


quin describe la especie desde material 


no conoce su origen, pero la consideró distinta de las 


demás. La descripción viene acompañada de una 
lámina que se reconoce como la especie. De Candolle 
(1825) la cita para “Nova-Hispania, ad portum De- 
sideratum,” la que se pensó era una localidad 
mexicana. Bitter (1910) señala que la especie debió 
Debido 


a que sólo se contaba con una lámina, en la búsqueda 


ser de Patagonia, Puerto Deseado, Argentina. 
de material para neotipificar se localizó una carpeta del 
herbario de Viena que presenta una muestra completa 
de la planta, donde la etiqueta señala A. sericea, con la 
fecha 1812 anterior a la descripción (1816), y el 
nombre impreso Herbar. Jacquin fil., por lo que se 
consideró que cumplía con las características del 
protólogo. 

Dentro de los sinónimos, para Acaena cuneata. se 
pudo examinar dos fotos de material tipo depositados 
en K y GL, donde ambos ejemplares se encuentran en 
buen estado, pero se eligió el material de K por tener 
más información original incluida en la carpeta. Para 
A. fuegina pude ver dos materiales tipo de Ortega. 
ambos depositados en SGO, de los cuales se eligió 
como lectotipo al SGO-49831 por presentar dos 
plantas completas y etiquetas originales. 

Philippi (1893) describió Acaena sericea Phil., 
Du- 


sén (1900) le da el nombre nuevo A. philippi Dusén. 


ero ya existía la especie de Jacquin (1816). 


— 


Según Bitter, dicha especie es sinónimo de A. sericea 
J. Jacq. 


.ARG ENTINA. Men- 
122 (LP); 
Trayecto desde Las Loicas hasta Paso Pehuenches, Villagrán 
et al. 8176 (CONC). Santa Cruz: Giier Aike, Est. Las 
, T.B.P.A. 2445 (HIP). ' 

a N side, Moore 2499 
Goodall 506 (L P). CHILE. V Región: Los Ande 
de Portillo, C. Marticorena & eo 599 
Recreo Alto, B. Parra 476 


Material representativo examinado. 


doza: Puente del Inca, Ruiz Leal Dawson 


Tierra del Fuego: Cabo 


P); Tennessee oil 


— 


es, 4 km antes 
CONC); 


Región 


Valparaíso, f 
ia a C 1 ra, Cajón del n l n abajo de 
. Landrum & Martínez 8227 ES GO). XII 
Sierra. Baguales, Dos de 
1818 (HIP): E Baguales, 


Embalse del Y 


Región: U Th E speranza 


Enero, Pisano & Cárdenas 4 


Marticorena 445 
Acaena en Chile 

Cerro Santa Lucía, Arroyo et al. 841072 (CONC, HIP): 
Magallanes, Escorial de Pali Aike, Cerro Diablo, Pisano 
5185, 5222 (HIP); Río Pescado, Islote Sin Nombre, Pisano 


2787 (HIP); Tierra del 
Dollenz 4515 (HIP). 


Fuego, Punta Espora, Pisano & 


18. Acaena splendens Hook. et Arn., Bot. Misc. 3: 
306. 1833. TIPO: “Cordillera of Chili, Herb. 
Cuming (N.299): who finds it also at Sierra Bella 
Vista, Aconcagua (1832), 
aquí designado, K!; duplicado, GL, 


SGO!). Figura LOA, B. 


” Bridges 3 (lectotipo, 
foto 3686 en 


17(74): 


Acaena 1 var. gracilis Bitter, Bil 1 5 Bot. 
8l. e los Andes, sub 


TIPO: Chile. “Cordill. 
ee 2 hololigo, 


nom. mn pe herb. Vindob., 
WI). 

Acaena splendens var. De Bitter, Biblioth. 
Bot. 17(74): 81. 1910. TIPO: Chile. “Prov. Aconca- 
gua, dr fragmenta syenitica ad nives perpe- 
tuas, Estero i Peñon rasgado, Cordill. de Sa. 
Rosa uná cum Á. digitata Phil. ambae sub nom. ‘A. 
alpina’ Poepp. | i chil. exs. 517." Poeppig 131 
(lectotipo, aquí puris W- -3594971: duplic a W- 
3594981). 


Acaena splendens var. macrophylla Bitter ar, Biblioth. 
Bot. 17(74) 82. 1910. TIPO: Chile. ordill. de 
Vindob.," Philippi. s.n. 1050 W- 


Santiago, herb. 


116061). 


\caena ees brevisericea Bitter, Biblioth. 
Bot. 17(74): 82. 1910. TIPO: Chile. “Cordill. de 
Maule, herb. Delessert in herb. Genev., Vindob 
(1855)." Ph. Germain s.n. (lectotipo, aquí designado, 
W-110770!) 

Planta hasta 44 em de alto; entrenudos cortos. 


Hojas de 6.7-13 cm 5 largo, de aspecto palmado; 2— 
4 pares de folíolos, 1^ y 2^ par de folíolos separados 
por más de 2 mm, a raquis notorio, de 8— 
26 mm de largo, seríceos, de color verde claro, el 
margen entero en la parte inferior del folíolo, parte 
apical entera o con 3-22 dientes; vainas foliares 


seríceas; apéndices estipulares ausentes. Rama flo- 


rífera terminal; pedúnculo hasta 43 cm de largo, 
seríceo: inflorescencia espiciforme, laxa; brácteas 


basales lineares, seríceas, de 3.5—4.5 mm de largo. 


de 2.54 mm de 


glabros en la interna; 


Sépalos 5, lanceolados, largo, 


seríceos en la cara externa, 
estambres 3-6, filamentos 3—4 mm de largo, anteras 
ovoides, de 0.5-0.7 mm de largo; estilos generalmente 
2. uno con cada ovario, estigma globoso, laciniado, de 
0.5-1 mm de diámetro. Cupela oblongoide-elipsoide, 
de 4.5-12 mm de 


brillantes, cubierta por espinas de 


largo, café claro, con tricomas 


largos, lanosos, 

tamaño similar, de 1-2.5 mm de largo, de color café 

rojizo, con gloquidios y tricomas retrorsos; aquenios 2. 

Presencia de flores cleistógamas protegidas por vainas 
o o 

foliares; cupelas cleistógamas del mismo tamaño y 


aspecto de las chasmógamas. 


Annals of the 
Missouri Botanical Garden 


Fig va splendens. —A, 
p. 0 lupe D (Se 1 55 6134, CONC). 


Icones. Grondona, 1964: fig. 61; 1984: fig. 456. 
Fenología. | Vructifica de diciembre (Baeza 279) 
a febrero (Ricardi 3249). 


Distribución y hábitat. Planta nativa; crece en 
lugares húmedos, desde la IV Región (Loica) a la VII 
Región (Los Queñes) desde 800 a 2400 m de altitud 
(Fig. 12E). 1 


Nombres comunes. 


"ambién en Argentina. 
choncli (Behn 


amores secos, cepacaballo (Baeza, 1921) 


Abrojo, cadillo, 
s.n.), 


Caracteres distintivos. Hojas seríceas, de color 


verde claro; folíolos enteros o con hasta 22 dientes 


pequeños en la mitad superior: cupela oblongoide, 
cubierta de espinas. 

Acaena splendens pertenece a la Sección Acae- 
na. Esta especie es similar a A. alpina, y suelen 
confundirse debido a la similitud de sus hojas, tanto 
en la forma de los folíolos como en su pubescencia. 


Sin embargo, se diferencia por la forma de la hoja que 


Hábito. —B. Cupela (Taylor y Gereau 10948, CONC). Acaena tenera. —C. 


C 


Hábito. 


es pinnada y las cupelas ovadas, con espinas de color 


marrón claro. 


Se pudo examinar dos materiales e de 


Acaena splendens. depositados en K y GL. Se eligió 
como lectotipo el de K, que comprende tres s 
bien desarrolladas, y además los datos de colección 
están escritos directamente sobre la carpeta. 


Para brachy- 


S 


| sinónimo Acaena splendens var. 


phylla. he 


visto dos materiales de Poeppig 131, 


= 


cuales se eligió 
lectotipo la carpeta W-359497, 


plantas con la 


depositados en de los como 
que presenta dos 
espiga desarrollada. Respecto de la 
variedad brevisericea, en el protólogo se mencionan 
materiales originales depositados en y G, y se 
designó como lectotipo la carpeta W-110770 que 
tiene dos plantas completas, con las espigas con flores 
sinónimos, no se pudo 


maduras. Dentro de los 


localizar material tipo para las variedades angustior 


Volume 93, Number 3 Marticorena 447 
2006 Acaena en Chile 


74° a 70» 68° 66° 74* 72% 70% 68% 68% 74* 72 70% 68% 66? 
1 | J 
i | A, | 
[as x | \ 
A [à BÉ t | C f 5 
\ ? \ 
20°- : 20° — 20% ME - 
\ n | ¡a | t | 
n PA | 
Ü ) 0 7 
2 
TE po 
25°- 190 — 2885 yo) - 
l l 
] ¿ N 
j j j 
f M 
F z 
( J ( 1 
30°- J j + 3094 — L 
( / ( 
$ 
35% , ^ — 35 — 
¡e 
40-4 | bo 40] — L 
( 
n a 
[ n 
i x ( E > 
„ QUE 9 
45 Re 5 451805 L 
bae g j 
ey 3 ay GS 3 
yf 2 J 2 Jj 
as | m j 5 | 
SW / D oer / 
aN, / Ny j 
500 ae ( Boo Ey: ( L 500 dt 
AS 05 EP RN 
T» Yi 
ON | 
55^. L^ 55) SOROR po 59. L a 
| 
T T T T T T T T T T T T T T ar p Sea ea | T T T T T T T T T T T T T T T Tr” 
76° 74° 72^ 70° 68° 66° 64° 76° 74° 72° 70° 68° 66° 64° 76° 74° 72° 70° 68° 66° 64* 76° 74° 72° 70° 68° 66° 64° 76° 74° 72° 70° 68° 66° 64° 76° 74° 72° 70° 68° 66° 64° 


Figura I. 
(A). —B. Acaena argentea (8): A. iger 
3 (0). —E. Acaena dua ED. —MV. Acaena ovalifolia. 


1 (A). —C. Acaena caespitosa (A): A. leptacantha (). -D. Acaena lucida (A); 


Mapas de distribución para las especies de Acaena presentes en Chile. —A. Acaena alpina (8): A. Pu 
e 


Bitter y subintegra Bitter, las que segün los protólogos subsp. epilis Bitt., exped. suec. 1907/09, 3.3. 1908," 
estaban depositadas en HBG y P, respectivamente. Shottsberg 208 (holotipo, 5GO-588501). 
Ambas fueron descritas sae la base de caracteres Planta 
variables de la especie, tales como el margen casi Hojas en contorno lineares, hasta 10 cm de largo: 5-8 
entero de los folíolos (subintegra) o por lo angosto de pares de folfolos, ovados, de 5-8 mm de largo, hacia 
la base subopuestos, pinnatipartidos a pinnatisectos, 


con 7-12 segmentos lobados, de 1.5-2 mm de largo, 


hasta 15 em de alto, entrenudos cortos. 


ellos (angustior). 


Material representativo examinado. CHILE. IV Región: 
Limarí, Loica, Jiles 4725 (CONC). V Redon 3 San Felipe. 
Jahuel, Barrientos 1761 (CONC) Los Andes. Portillo. El 


base asimétrica, con un mechón de tricomas pluri- 


celulares, glandulares: vainas foliares glabras: apén- 


Juncal. Simon 199 (8GO). Región T Santiago, dices estipulares lobados, de 1.5-2 mm de largo. 
Quebrada de San Ramón. 1 Nov. 1879 : es ippi s.n. > OTTO n 2 i ; 
E en Rama florífera terminal; pedúnculo hasta 14 cm de 
(SGO): Cordillera. o Becket et all (SGO): 

largo. glabro; inflorescencia capituliforme, de 5.5- 


Cordillera, along Embalse El Yeso, Taylor e uh 1948 
(CONC): Subida a Potrero Grande, 8 Dic. 1933, Behn s.n. 
(CONC). VI Región: Cachapoal, El 1 Zoellner 10142 


6.5 mm de diámetro: brácteas basales orbiculares, 


lineares o lanceoladas, con el ápice dentado, de 0.5— 


0 agué e 20 279 (CONC); ‘ tz : 

(( NC); Colchagua, pur as del Flaco, Baeza 279 ( 2 mm de largo. Sépalos 3, ovados, de 1-1.5 mm de 
San Fe mando d Jegi | Flaco, Ricardi 3249 (CONC). VH . . 

Región: Curicó, Camino oe Los Queñes a El Planchón. A. largo: estambres y anteras no vistos: estigma globoso. 
Marticorena et al. 135 (CONC). Cupela obovoide, tetrágona, de 1-1.5 mm de largo, 


sin las espinas, hasta 3 mm de largo con las espinas, 
, con la superficie cubierta de glándulas sésiles 
19. Acaena tenera Alboff, Revista Mus. La Plata 7: AN MM 7 E HM . 
us esféricas de color naranjo (manná), 4 espinas de base 
367. 1896. TIPO: Argentina. “In regione alpina l VET : Tan 
un i angosta en el ápice, ca. 1 mm de largo, con fuertes 
montis Pyramidis (supra torrentum Ushuala), 29 | T 10 ] 
. inl M. 3 eloquidios; aquenio J. 
Feb. 1896," Alboff 234 (lectotipo, aquí desig- 


nado, LP-28054!). Figura 10C, D. Icones. Grondona, 1984: fig. 455. Skottsberg, 

i 2s "m TERN 1905: tab. 1-2. 
Acaena antarctica var. laxiuscula Itter. 1bhoth. ol. 4 E : lam DE E 5 
; Fenología. Fructifica de diciembre (Pfister. & 


17(74): 54. tab. 3a. 1910. TIPO: Chile. “Fuegia: in 


alpinis prope Rio Azopardo, unà cum A. tenerá Alb. 


Ricardi s.n.) a febrero (Pisano et al. 8271-A). 


Annals 


118 Eos Garden 
74" 72? 70° 68° 66° 74° yd 70% 68° 66% 740 72° 70° 68° 66? 74° 72° 70° 68% 66° 74° tas 70° 68° 66° 74° 72' 70° 68% 66° 
1 1 | | 1 1 L L 1 1 1 | L EFT L L SE 
Jr J^ : X 
A C D (ti E F f 
20" ) 6 20°- ) t 20°- 1 20° ¿ 
| 5 | ) E | 5 A 
PA PA EE ER 
J i: J \ ] L | he 
0 t / 0 7 a ] 
25° 1 + 25% ) d 25°- ( 
l j 5 ) ) 
j $ | $ ) S 
p^ { f* 
| / ES Pur 
( ( ¿ Cd) 
30°- } + 304 ( / 3055 l- 
3 2 
e ) ] o e | 
35° J d L 3] Í 355 ( - 
/ 1 / / [ ) 
( PE. «4 
DA PA | 
| ) \ ) y 
) J ] 4 ) oy 
40° ( $ L Po 40° A 3 : E sod / 3 t 
| m gd 
"n n) M 
M1 M ; $ 
ey C D 8.5 
St SDO SD O 
45°. “Be 5 45°- A j 45°- A 
ay Ph: j Fas j 
E, E, y PO d 
Qu f % ' 
50° 4 de t + 50%- d ( 
“Y T 
A br 
RW). 
B 
SE f 
| MO 
- 550- 
TT TT T TT 2 T TT 7 wp "e gai T TT T VVV 
76% 74° 72% 70° 68% 66° 64° 76° 74° 72° 70° 68? 66° 64° 76° 74° 72° 70° 68° 66° 64° 76° 74° 72° 70° 68° 66° 64% 76° 74° 72° B " 8 64* 76° 74° 72° 70° 68? 66? 64° 
Figura 12. Mapas de distribución para las i pbi de Acaena presentes en Chile. —A. n masafuerana (): A. 
patagonica (A). —B. Acaena pinnatifida (@). —C. Acaena a antha (A): A. O (e). —D. Acaena pumila (A): A. 
sericea (0). —E. Acaena splendens (): A. tenera dir . Acaena trifida (8): A. trifida var. 9 (A). 
Distribución y habitat. Planta nativa, habita en la Heuhuepen, Moore 2 2 CHILE. e Región Ultima 
Poro, Arroyo et al. 46 (CONC); 


XII Región, desde 120 a 700 m sobre el mar, en sitios 
húmedos y entre cojines Bolax gumifera (Fig. 12E). 
Pambién crece en Argentina y en las Islas South 
1964). 
Caracteres distintivos. 
folíolos con un mechón de tricomas 


Georgia (Greene, 
Cupela con glándulas esfé- 


ricas. (manná): 

pluricelulares en la base, la que es asimétrica. 
Acaena tenera pertenece a la Sección Acrobyssi- 

Especie poco recolectada, parecida a A. antarc- 


nae, 
tica y a A, masafuerana, de las cuales se diferencia 
fácilmente por un mechón de tricomas glandulares en 
la axila de los folfolos y por las glándulas esféricas 
a cupela. Según Walton 
A. magellanica. 
laxiuscula, 


(manna) en la superficie de | 
(1979), esta especie hibridiza con 
Respecto del sinónimo A. antarctica var. 
Walton (1975) la consideró sinónimo de A. antarctica 
sin embargo las características de las hojas y cupela 
son claramente de A. tenera. 

Para la subespecie pilosella Bitter, no fue señalado 
el herbario donde fue depositado, y no fue posible 
localizar el tipo. No obstante, Bitter (1910) describió 
la subespecie basado en la pilosidad de los folíolos, 
carácter muy variable. El autor incluyó una foto que 
claramente concuerda con la especie. 


examinado. 


of 


Material 


de Fuego 


representativo 
DON side 


L ago F agnano, 


Tierra 


ARGENTINA. 
Monte 


Esperanza, Sierra del 
Magallanes, Punta Arenas, 31 Dic. 1951, 5 5 & uale: 


(CONC); Tierra del Fuego, Sector Río Cóndor, Pisano et 
= 8271-A (CONC): Antártica Chilena, Isla 1 Puerto 
Williams, Cerro Bandera, Schlegel 8134 (CONC). 


Chil. 
Regni 


Acaena trifida Ruiz et Pav., Fl. Peruv. 
67. 1798. TIPO: Chile. "Habitat in 
Chilensis pascuis, campis et collibus Rere, Itatae 
ad Huilquilemu, 


20. 


et Conceptionis. Provinciarum, 
Culenco, Penco, Talcahuano, Hualpen et Car- 
camo tractus,” Ruiz € Pavón s.n. (holotipo, MA 
| Herbarium Peruvianum, Ruiz et Pavón, 10/82]!) 


). 


Figura 7C, 
Planta hasta 36 em de alto; tallo cubierto por las 
hojas; entrenudos cortos. Hojas de 2-6 em de largo; 
4-8 pares de foltolos, obovados, de 6— 
13.5 mm de largo, pinnatisectos o partida, con 3-5 
apéndices estipulares 


pedúnculo piloso 


ovados a 


lineares, agudos, 


florifera 


segmentos 
ausentes. Rama axilar, 
a pubescente, con tricomas largos, blancos, blandos: 
inflorescencia capituliforme: brácteas basales lineares 
—3.5 mm de largo, con hojas 
pequeñas a lo largo del pedúnculo protegiendo 
algunas flores o frutos. Sépalos 4-6, ovado-oblongos 


o lanceoladas, de 


a oblongo-lanceolados, estambres 4-6, anteras globo- 
sas, estilo corto, estigma globoso, laciniado, ca. 1 mm 


Volume 93, Number 3 
2006 


Marticorena 
Acaena en Chile 


de diámetro. Cupela ovoide, generalmente con 4-5 
costillas con 2 espinas en cada una, base ensanchada, 
con tricomas retrorsos, sin gloquidios; aquenio J. 
Acaena trifida pertenece a la Sección Acaena. Esta 
trifida y la 


las cuales se 


especie tiene dos variedades, la var. 
variedad glabrescens Regel et Kórn., 
diferencian. principalmente. por la pilosidad de los 


tallos y de la cupela. 


CLAVE PARA LAS. VARIEDADES 


la. Plantas pubescentes, cupelas sericeas 


%% . trifida var. trifida 
Ib. Plantas glabras, cupelas glabras .......... 

3%ͥ;ö;%⅕d“ trifida var. glabrescens 
20a. Acaena trifida Ruiz et Pav. var. trifida 


Linnaea 28: 686. 
trifida var. quinquefida e Reiche, 
97. 


1858. 


Anales Univ 


Acaena quinquefida Phil., Acaena 


Chile 98: 167. 1897 Chile. “Provincia de 
Valparaiso, Cordilleras T CR Nov. 1854.7 
Germain s.n. (lectotipo, aquí designado, 5GO-49930!). 
Acaena 1 var. 5 Bitter, Biblioth. Bot. 1774): 


102. 10. TIPO: Chile. “Concepción, herb. Berol., 
Monac eon 1894,” a W. Neger 1466 (lectotipo, aquí 

de lena, Mh). 
a S ies var. 19 Bitter, Biblioth. Bot. 
). TIPO: Chile: 


pon ae [sic]. 


17(74): 
“sine loco spec ali indic.. sub 
nom. herb. 
Berol., Paris.” 
P-106322!). 
Acaena 0 var. Ve 0 ant s Biblioth. Bot. 17(74 
). TIPO: Chile: * Nuble, 
cum 7 E var. a Bin. 
Neger 5099 (holotipo, M!, las dos plantas inferiores). 
Acaena trifida var. nanodes Bitter ; Biblioth. Bot. 17(74): 103. 
1910. TIPO: Chile. “Prov, Nu ble, San Carlos, uná cum 
varie lale alter * Bitt. herb. Monac., 
F i us 5099 (holotipo, M!, las cuatro plantas 
superiores) 


Dombey mus. 


Dombey s.n. (lectotipo, aquí designado, 


San Carlos, uná 
Monac.," F. W. 


herb. 


— 
x 


Bot. 17(74): 
‘Chile australis: San Cristobal 
Monac., Stockh., (1894—95),” 
5100 (lectotipo, aquí ea M!; dupli- 


Acaena trifida var. brachyphylla Bitter, Biblioth. 
103. 1910. TIPO: Chile. * 
(Eisenbahnstation), herb. 


P. Dusén 
ad 


S. no visto). 
Biblioth. Bot. 17074 


Valparaiso: 1 he 


cado, ? 
Acaena 1 var. Pos Bitter, 
104. 1910. TIPO: Chile. * 

Vina del Mar, ui 1 hilensis in Baenitz Herb. Amer 

A. trifida, von F Dr. Dus 

tP m zeichnet; 


er. sub 


Später als 4. 
“Bushee 44.56 


— 
T 
Em 
doo 
AN dE 
a 
= 
=. 
"m 
= 
a 
Sm 
Em 
a 


(holotipo, M!). 


Hojas de 2-6 cm de largo; lámina en contorno 
7-8 


pinnatisectos, seríceos; 


linear-lanceolada: pares de folíolos, ovados 
vainas foliares seríceas en el 
dorso, cara interna glabra. Rama florífera hasta 29 cm 
de largo; inflorescencia de 11-14 mm de diámetro; 
brácteas basales de margen irregular, seríceas en la 
cara externa, glabras en la interna. Sépalos de 2.5— 
3.5 mm de largo, café rojizos, pubescentes como las 


brácteas basales, persistentes en el fruto; anteras de 


l mm de diámetro. Cupela de 3-6 mm de largo, 


serícea, con tricomas cortos, blancos. 


Icon. Ruiz y Pavón, 1798: tab. 104, fig. c. 
Fenología. | Fructifica de septiembre (Montero 12649) 


a febrero (Ledezma 626). 
Distribución y hábitat. Especie endémica, crece 


en suelo arenoso, laderas rocosas, en bosques 
esclerófilos abiertos, desde la provincia de Choapa 
(IV Región) al Volcán Villarrica (IX Región), desde 
aproximadamente los 15 a los 1800 m de altitud 
(Fig. 12F). 

Nombres comunes. Pimpinela (Pfister 277; Baeza, 
1921), amores secos (Behn s.n.). 

En el protólogo de Acaena quinquefida, dice que 
crece en Valparaíso y Cordillera de Santiago, lo que 
está basado en dos carpetas depositadas en SGO. Para 
tipificar, se eligió la carpeta de Valparaíso por ser más 
representativa de su distribución y ser reconocible 
como la A. trifida. 

En el protólogo de A. trifida var. argentella, se 
menciona materiales originales en B y M. Se presume 
que el material de B fue destruido, por lo tanto se 
eligió lectotipo la carpeta M (Neger 4466). Para la 
variedad mollissima, el autor menciona dos materia- 
les, uno de Chile con duplicados en 
P 


B y P, y otra 
muestra de Perú depositada en B. Lo materiales de B 
no se localizaron, presumiblemente destruidos, luego 
se eligió como lectotipo la carpeta de P- 106322. De la 
varie dad brachyphylla, se menc ionan ene | protologo 
los herbarios M y 5, de los que sólo se examinó el de 
M, el que fue 11 7 como lectotipo. El tipo de la 
variedad valdiviensis Bitter estaba depositado en B, y 
Se supone que fue destruido. 

muchos de los taxa 
Bitter, 


i.e., la pilosidad de 


Como ha sucedido con 


infraespecíficos descritos por estos fueron 


basados en caracteres variables, 


as hojas (argentella), por la suavidad y flexibilidad 
(mollissima), por las hojas cortas (brachyphylla). 
(1829) describió 


señaló 


Lindley Acaena 


Lindl., 


un homónimo posterior por existir la especie de 


pinnatifida 


pero no material tipo, además era 


Ruiz y Pavón. La lámina que acompaña al texto 
muestra que corresponde con las características de 


A. trifida. 


IV Región: 


Material representativo examinado. CHILE. 
I 
. V Región: 


Choapa, Pichidangui, ed ro 10803 (CONC) 
2 Philippi s.n. (SG 


O): Valpara- 
Laguna V ie camino M VHS, Lammers et al. 
| A 929, Beh 


Petorca, Catapileo, Sep. 
íso, 3 km N 


7761 (CONC); Lago Peñuelas, 2 go. | n s.n 

(CONC). VI Región: Cardenal Caro, Pichilemu, Montero 

12649 (CONC); Colchagua, La Rufina, Montero 12648 

a VII Región: Talca, Constituci ión, Dic. 1891, Reiche 
| à 


SGO). VIII 1 8 
Mie: 558 (CONC 
ONC); Concepción, Camino de Florida a Per 

d: Florida, Landrum 7932 (CONC); Cerro Carac il. Pier 277 


p 


Annals of the 
Missouri Botanical Garden 


(CONC): Biobío, Cordillera de Poleura, Ledezma 626 (CONC). 
IX Región: Malleco, Angol, cerros, Montero 5277 (CONC): 
Cautín, Volcán Villarrica, Ene. 1935, Pfister s.n. (CONC). 


20b. Acaena trifida Ruiz et Pav. var. glabrescens 


Regel et Kórn., Index Sem. Hort. Petrop. 1857: 8. 
TIPO: “Chili. Philippi .. F (neotipo, aquí 
1 


1858. 


designado, LE 


Hojas de 2-3.4 em de largo: lámina en contorno 
4—6 


subopuestos, de 8-12 


algo obovado; pares de folíolos, obovados, 


mm de largo, 3 (4-5) partida, 


cara superior glabra, cara inferior glabra a algo 


pubescente; base ancha 0 de aspecto angosto por el 


margen revoluto, vainas foliares glabras. Rama 


S 


florifera hasta 18 em de largo, inflorescencia de 8— 
10 mm de diámetro: brácteas basales glabras. Sépalos 
de 11.5 mm de largo, glabros; anteras de 0.5 mm de 
diametro. Cupela de (1.5)2.5-3.5 mm de largo, glabra 
a algo pilosa, espinas de ca. 2 mm de largo. 

Fenología. | Fructifica de noviembre (Schlegel 7926) 
a diciembre (Gunckel 21873) 

Distribución y hábitat. Variedad endémica, se le 
encuentra desde la provincia de Antofagasta, Cerro 
Moreno, 


Chañaral, 


sector de El Rincón, Paposo, también en 
| 


Aguada Grande y en la zona de Las 
Condes en la Región Metropolitana, desde los 200 
a 1800 m de (Fig. 1215). 


escasa que crece e laderas escarpadas, en arena y 


altitud Es una planta 


entre rocas, en lugares húmedos entre fisuras. 


variedad fue descrita desde 
Chile 


pero 


Sta ung planta 


proveniente de que crecía en el Jardin 


Petropolitano, Regel y Kornicke (1858) no 
indicaron el tipo. El carpeta designada como tipo se 
encuentra depositada en LE, cuya etiqueta dice “Ex 
Horto bot. Petropolitano.” por lo cual es muy probable 
Bitter (1910) 


observó en el herbario de Firenze (FI) un material 


que corresponda la material original. 


proveniente de S Santiago y menciona que la planta en 
general es glabra, pero presenta algunos ide con el 


ápice esférico, compuestos por muchas células. 


Material representativo examinado, CHILE. II Región: 
Antofagasta, Cerro Moreno, eie 5016 (CONC): Cerro El 
9 ada & 170 53 (CONC): 

Ruiz 373 (C ; Paposo, Quebrada Guanillos. 
n (CONC): Dus pd Agua Grande, entre Pan de 
; 5380 (CONC) 
Región Ps 5 0 Santiago, Cordillera de Las Condes, 
Gunckel 21873 (CONC 


Rincon, Quebrada Peralito, 
Quezada & 
men 


aleta Esmeralda. Pisano & Bravo 


Literatura Citada 
Viton, 1789. Hortus Kewensis: Or. 
P E C male d in the 
Vo Is. 1 . London. 


of the 
Garden at Kew 


a Catalogue 
Royal Botanic 


Viton. W. 1810-1813. Hortus Kewensis; Or, 

of the 1 15 Cultivated in the Royal Botanic Garden at 
the late William Aiton. The Second 
Townsend Aiton, ( 


a Catalogue 


Edition 
Gardener to his 
London. 

yo. . Squeo. 1989, Re ee between 
plant. bre iine systems and . En S. kawano 


(editor). Biological \pproaches and | 1 Trends 


in. Plants. Eo mic Press. London. 


Baeza. V. a 19 


silvestres de 


Los nombres vulgares de las plantas 
5 i 
científicos. Anales Univ. 
Beadle, N. C. W. 
M. D. Tindale. 1982. ewe P: 
the Sydney 3rd ed. 


su concordancia con los nombres 
Chile 147: aan 
Evans, R. 


— 


C. Carolin & 
1 en Flora of 
Region. Kyodo-Shing Loong, Singa- 
Bitter, G. 1910. Die Gattung Acaena. Vorstudien zu einer 
Biblioth. Bot. 174): i, 1-330. 

Gattung Polylepis. Bot. Jahrb. 


Monographie. 


1911. Hevision der 


Sy D: 904=030; 
Briggs, D S. M. Walters. 1984. Plant Variation and 
Evolution. 19 5 ed. Cambridge Univ. Press 


7. Du genre Acaena. Rev. Sci. Nat. Ouest 7( 
21-24, 2) 33- Id. 
Mos - 198 


. An Integrated Syste m of Classification of 
Press, New Yor 


Flowering 0 ants. Columbia Ui 


de Candolle . 1825. TASA Pp. 5 


3 V. P. de 
Candolle pee. parte por N. C. 
Seringe.) Treuttel & 

Dusén, P. 1900. Die 


nebst 


Prodromus 2. (Fi 

ürlz, Paris, London, Strasbourg. 

Gefüsspflanzen der Magellanslinder 

einen Beitrage zur Flora der Ostküste von 
Patagonien. Wiss. Ergebn. Schwed. Exped. Magellanslän- 
dern 3(5): 77—260. 

Faegri, K. & L. van der Pijl. 1971. The 


Principles 


of Pollination Ecology. 2nd ed. Pergamon Press 
Germany. 
Forster, J. R. & G. Forster. 1775. Characteres Generum 


Plantarum. London. 
N 


Frankel, R. E. Galun. 1977. Pollination Mechanisms, 
Reproduction and Plant Breeding. Springer-Verlag, Ger- 
many. 

Gay, C. 1847. Rosáceas. Pp. 266-315 en C. 2 Historia 


Física y Política de Chile. Botánica. Vol. Thunot, 
i Natural, Santiago. 
. Rosaceae. Pp. 1003-1010 en L. 


Laruc né 1 1 ). 


Brako & J. 
f Catalogue of the Flowering Plants 
and Gymnosperms of Peru/Catálogo de las Angiospermas y 
Gimnospermas del Perú. Monogr. Syst. Bot. Missouri Bot. 
Gard. 45. 


Greene, S. W. 1904. The vase ^a me of South Georgia. Sel. 
Rep. Brit. Antare. Surv. 45 
Greimler, J., T. F. Stuessy, U. Swenson, C. M. Baeza & 0. 


Matthei. 
ago. Biol. Invas. 4: 73-85. 

n Li 1 . F. Stuessy & T. Dirnbock. 2002. The 
vegetation of Robinson Crusoe Is SIMA (Isla Más a Tierra). 
Pacific Sei. 5603): 


200 02a. Plant invasions on an oceanic are bene l- 


la Fernández Archipelago, Chile. 
03-284. 
5 E. 1964. Las 


Acaena on eae). 


argentinas. del 
200—342. 


© Rosaceae, en Correa. Flora Patagónica, parte 


especies género 
Darwiniana 13: 


A-b: 48.88. 


leywood, V. H. 1985. Plants of the World. 
Prentice X Englewood Cli v Jerse 
o J. D. 1844-47. The Botany S 115 Had arctic Voyage. 


S Reeve Brothers. London. 


F 5 
[fs 


. Flora 


Volume 93, Number 3 
2006 


Marticorena 
Acaena en Chile 


Hutchinson, J. 1964-1907. 
(Angiospermae) Based d 
Bentham and J. 

. 1-2. Clarendon Press, m 
Jacquin J. a von 18 111-1844 
. Vindobo 
F. 1806. Es 
Fernández. 


Judd, W. S., C. 


The Genera of Flowering Plants 


on the Genera Plan- 
larum of G. Hooker. Dicotyledones. 


: ; Eclogae Plantarum Rariorum. 
iae, Wien. 

je pene mn la Flora de las Islas de Juan 

Im; e nia Cervantes, Santiago. 


Campbell, E. A. n & P. 


Slevens 1999, Plant Systematics iylogenetic 
Approach. Sinauer Associates, ad es hu- 
setts. 

Kalela, A. 4 10, Beitrage zur Kenntnis der Flora Ostpata- 
goniens Sektion ae der Gattung Viola; M. 
Caune "bun III. Gattung. Carex. Ann. Acad. Sci. 


"enn., Ser. A. 54(5): 1-70. 
Kalkman, C. 1988. The phylogeny of the Rosaceae. Bot. J. 
Linn. Soc. 98: 37-59, 
Lamarck, J. B. A. P. M. de & J. L. M. Poiret. 1791— 
Tableau Encyclopédique et Voie des Trois 
de la Nature. 1-6. Chez 


1823. 


Regnes 


Botanique. Vols Pancoucke, 


Icaena pinnatifida. Edwards's Bot. Reg. 


Altera. 

onis II. Holmiae, Vapenis 

I-VI. (Facsímile, J. LK 
961). 


wann. Hist. Nat. Class. 8. 196 
M. 


Linnaeus, €. 1771. Mantissa Plantarum Generum 


Editionis VI et Specierum Editi 
Salvii. 


Laurenti ;ramer 


Lord, 


E. M. 1981. Cleistogamy: A tool for the study of floral 
morphogel nesis, i and evolution. Bot. Rev. (Lan- 
caster) 47: 421—450. 

Mabberley, S pe The Plant-Book. 2nd ed. Cambridge 


Univ. e SS. 


e Cr B. H. 1988. Acaena. En C. J. Webb, W. R. Sykes 


sarnock-Jones (editores), Flora of New Zealand, 

i . Naturalised Pte "ibi s, Gymnosperms, Dicot- 
1 Boleny Division, D.S.LR.. Chris tchur ch. 

Marticorena, A. 1999, Una nueva espe de 80 


(Rosaceae) en Sudamérica Austral. 1 9. 227. 

2000. Patrones celulares de la epidermis des la 

cupela de Acaena L. (Rosac me chile nas; Gayana, Bot. 5 
)-156 


149— 


L. Cavieres. 2000. Acaena 5 (Lam.) 
Vahl (Rosac eae). Flora Biológica de Chile l. Gayana, 
Bot. 57: 107-113. 
E C. 1991. Contribución a la estadística de la 
flora vascular de Chile. Gayana, Bot. 47: 85-113. 
& M. Quezada. 9 Catálogo de la flora vascular 


de Chile. Gayana, Bot. 42: 1-157. 
M: > [ hei, O. 
hile. Alfabeta pa sores, Santiago. 
Tin D. M. 1968. The 
Islands. a Bos Brit. 
1972. Connections between cool temperate floras, 
with "eic e reference to southern South 
Pp. 115-138 en D. H. (editor), 
a and E SIRO Academic Press, London. 
Moore, D. M. & Walton. 1970. 
numbers of Acaena from South Georgia. Brit. 
Surv. Bull. 23: 101—103. 
Philippi, R. 


Centuria prima. Linnaea 28: 609-655. 
wkd 


995. Manual E las Malezas que Crecen en 
s ` flora of 3 Falkland 
Antarc. Surv. 60: 1— 


America. 


Valentine l'axonomy, 


C hr omosome 
Antarct. 


1857. ea novarum chilensium. 


52. Descripción de unas plantas nuevas recogidas 
en el verano pasado en la provincia del Maule i en Chillan 


por don Jerman Volckmann. Anales Univ. Chile 21: 


442-450 


and hybridization on South ( 
13. 


. 1863. Catálogo de las plantas rec ne En G. E. 
Cox, Viaje a las regiones septentrionales de la Patagonia. 
Anales Univ. Chile 448—474 

1864. 


inclusis quibusdam Mendocinis et Patagonicis. Linnaea 
33: 1-308. 


1 novarum chilensium. Centuria 


1872. Descripciones de las plantas nuevas incor- 
en ültimamente en el herbario chileno. Anales Univ. 
Chile 41: 663-746. 
1555 Tots 
mw as, M cup as i oe más familias del tomo H de 
Univ. Chile 84: 619-634. 


Pisano, E. 1984. 


nuevas chilenas de las familias 


Gay. Anales 

Pe ciones cireumantárticas de la flora 

vascular del are hipiélago del C abo de Hornos. Anales Inst. 

Patagonia, Ser. Ci. Nat. 15: 5-24. 

H. & D. I. Axe Dr 1974. Angiosperm bio- 
geography and E continental movements. Ann. Missouri 
Bot. ig Ol: 673. 

Regel, E. A. von yi f. A. Ko 


quae ned botanicus impe rialis pe tropolitanus pro mulua 


Im 


Raven, 


órnicke. 1858. Index Seminum 


commutatione 8 0 Accedunt animadversiones botanicae 
nonnulla St. Petersbu 
D E: pir iy crítica OS s sobre la flora de Chile. 
iv. 7 98: 

. The Dispe sal of Plants throughout the 
>, a 
yc 


Romoleroux, T 1 O. Rosaceae. 
. Andersson (editores), 
Göteborg; Riks Stockholm; I 

dad Católica del Ecuador, Quito. 


. 1999, Rosaceae. Pp. 851—855 en P. M. 


leon Yánez (editores). Catálogo de las 


p. 1-151 en G. Harling & 
Flora of Ec 


D 


wuador 56. Univ. 


museum, Pontificia U niversi- 
Jorgensen & 
llantas 


Vase o del 


[e] 
Ecuador. Monogr. Syst. Bot. Missouri 
Bot. 75 
B “1981. Contribution à l'étude du genre Acaena L. 


aaa dans les Andes argentines et chiliennes. Bull. 
Soc. Neuchatel. Sci. Nat. 104: 145-151. 

Ruiz, H. & Pavón. Flora Peruvianae, e 
Chilensis. Vols. 1-3 [Vol. 1. 1798: Vol. 2, 1799: Vol. 3, 
1802]. Sancha, Madrid. 


Simpson, B. B. 1979. A revision of genus Polylepis 
(Rosaceae: Sanguisorbac). solia Contr. Bot. 43: 
i-i, 1-62. 


Skottsberg, C. 1905. Die Gefüsspflanzen Südgeorgiens. Wiss. 


Ergebn. = 5 Südpolar-Exped. Ta 1903 4(3): 1-12. 
The VVVVVV—¼ e Juan Fernandez 
Is ms m 95— 240 en C. Skotts dud 0 0 b ve Natural 
History of o 5 z and Easter Island, Vol. 2. 
. Notas sobre la vegetación de d 15 de Juan 
F ae Peri Univ. (Santiago) 38: 195-207. 
Spegazzini, 897. Plantae Patagoniae er Revista 
Fac. Agron. Vete rin. La Plata 3(30-31): 485—589. 
Spjut, R. W. 1994. Systematic treatment of fruit types. Mem. 
New Yor 182. 


a Gard. 70: 
Th. 1€ Cleistogamous flowers. Bot. Rev. 


4: 214 

ahl, M. 1804-1805. 

9 9 Copenhagen. 

Walton, W. H. 1975. 

Pp of genus Acaena L. 
09. 


500—! 


Vols. 1, 2. 


dnd Plantarum. 


— 


‘axonomic notes on South American 


(Rosaceae). Darwiniana 19: 


577. Studies on Acaena (Rosaceae). Seed 


ee growth and establishment. Br. Antarct. Surv. 
Zull. 29-40, 

e 9. Studies on Acaena (Rosaceae). HI. Flowering 
,eorgla. Br. Antarct. Surv 


Bull. 


452 Annals of the 
Missouri Botanical Garden 
v S. W. Greene. . The Ww m species Garaventa 6379 (13); Gardner 3450 (2), 3849 (9), 3889 
of aa "na s their 99 0 hybrid. Br. Antaret. Surv. (2). 4168 (16): Gay 527 (15). 76 (13). 78 (6). 80 (11). 81 (13), 
Bull. —44. s.n. (11) (P), s.n. E (P-106316); Germain 88b (9), s.n. (1) 
Wedde IL H. PF 1855-1861. Chloris Andina. (W-112652), s.n. (1) (W- 10760), s.n. (1). s.n. (15). s.n. (18) 


. 1855-1857; Vol. 


173. Los géneros de Ros 


Chez | 
Bertr pos Apo Vol. 1 2, 1858-1861.] 
Zardini, E. M sáceas espontáneos 
en la [T a Argentina. Bol. Soc. Argent. Bot. 15: 
22 

99. Rosaceae. Pp. 981-992 en F. Zuloaga & O. 
Morrone noe Catálogo de las Plantas Vasculares de 

a República Argentina. Monogr. Syst. Bot. Missouri Bot. 

Gard. 74 
APÉNDICE l. NDICE DE COLECCIONES NUMERA 
DAS. CADA ESPÉCIMEN ES CITADO POR EL APELLIDO DEL PRIMER 
COLECTOR EN EL CASO EN QUE OTROS COLECCIONISTAS PARTICIPEN DE 
LA COLECCION. EL NUMERO ENTRE PARENTESIS CORRESPONDE A LA 
ESPECIE NUMERADA EN LA LISTA. 


LISTA DE ESPECIES I 


l. Acaena alpina Poepp. ex Walp. 

2. Acaena antarctica Hook. f. 

3. Acaena argentea Ruiz et Pav. 

| Acaena caespitosa Gillies ex Hook. et Arn. 
5. Acaena integerrima Gillies ex Hook. et Arn. 
6. Acaena leptacantha Phil. 

7. Acaena lucida (Aiton) Vahl 


8. Acaena macrocephala Poepp. 

9 lcaena magellanica (Lam.) Vahl 
10. Acaena masafuerana Bitte 

ll. Acaena ovalifolia Ruiz et Pav. 
12. Acaena patagonica A. E. Martic. 


13. Acaena pinnatifida Ruiz et Pav. 
l4. Acaena platyacantha Speg. 

15. Acaena poeppigiana Gay 

10. Acaena pumila Vahl 

I7. Acaena sericea J. Jacq. 

18. Acaena splendens Hook. el Arn. 
19. Acaena tenera. Albo 

20a. Acaena trifida Ruiz et Pav. 

20b. Acaena trifida Ruiz et Pav. var. glabrescens Regel el 
örn. 


Alboff 234 (19), : 5); Andersson 280 (9); Anliot 6140 
(9), 6141 (11); Arancio 1 (9). S 201 (9); Aravena aa (9); 
Arroyo 81274. , 81277 (J), 286 (9). 841072 (17). 
850829 (12). 8. 50820 10 2). 7 pe 860086 (11), 87 0179 
202 (9). 870316 (5), 92-146 (19), 


; Barrientos 1733 (13), 1761 (18); Barros 
s.n. 4422 (9). 4471 mud Behn s.n. (18), s.n. 
(20a); Borchers s.n. (13) B 19838 ge (5). 

(9); Braynshaw 202 (13), 221 (11), 246 105 Bridges 3 (18), 4 
(17), 522 (9), 772 (9), 773 e Buchtien 4456 (20a), 54.64: (9), 
s.n. (13) (BREM), s.n. (9) (BREM). 

Cabrera 19193 (6), 20506 (13), 23010, 23023 (13); Capt. 
King 10 (9), 16 (9), 46 (9), s.n. ( a (K); Carrette 215 (5); 
Catalán s.n. (11); Claro de la Maza s.n. (7); Commerson s.n. 
(16) (P-106320); Cox s.n. (9) (SGO- 498: 32), € 81 5 11863 


1 enz ME 220, 241, 201 (9), 405 


Dessauer s.n. H 1 


E Ivebakk 34 (15), 35 (9). 470 (7). 532 (12). 537 (14), 538 


"Fabris 2168 (6), 784 (6); Ferruglio 49 (5); Fonk 105 (16). 


(W-110770), s.n. en 1) (SGO-49930), s.n. (3). s.n. (6) (SCO 
49888), s.n. (9) (K), s.n. (9) (5-8049), s.n. (9) (W-359414); 
Gillies 113 (5), s.n. 8 Goodall Bard 191 555 x bre, 
2283 (9), 506 (17); Grandjot s.n. (1), ickel 
15348 (11), DEIN 21873 (20b), NE e 
38441 (3), 50079 (5), 70058 (3); Gutiérrez 97 (6) 

Headland 562 (9); Henríquez 98 (16); 


oO 


Hollermayer 625 


om 712 (1): Hombron s.n. (13) (P-106302); Hooker 1827 (7), 
1 (P- 107 11); Hyades 901 (9). 
Em in 1900 (4). 


* Jiles 1585 (13), 1685 (11), 2110 (15), 4215 
4725 (18), 6370 (13); Junge 328 (16), s.n. 


hl sn (1) 
(9), 4438 (15), 


PEN rt s.n. (16) (UPS); Lammers 7761 (20a); Landbeck 


s.n. (13) (8GO-49871); Landero 671 (14), 9592 (10); 
ium 7932 (20a), 8227 (1 D. 8437 (4). 7 16 (13); Lechler 
BE 13), 294 (13), 2951 (2), 318a (3), 584 (11), 978 (13), 


978c (9); Ledezma 620 (20a), 
Loyola 4. (9). 

O A. 105 (13), 
(1), 272 (3), 296 (13), 297 (9), 
121 (14), 98 (9); Marticorena, C. 
, 1616 (9), 


626 (20a): Lépez 108 (3): 
135 (18). 162 (13), 228 (8), 233 
298 (3), 3 335 (11), 
106 (13). 107 (8), 1254 (5). 
1633 (11), 1949 (11), 22 (J). 


Matthei 


(9); Moore 1074 (9), 2369 (9). 2499 (17), 2603 (9), 2828 (19 
Moseley s.n. (3); Muñoz 0099 (9), s.n. 1 

Navas s.n. (5); Neger 4466 (20a), 4639 (6), 5099 (20a), s.n. 
(2). s.n. (3) s.n. (6) (M), s.n. (9) (M); Nuñez s.n. (9). 

Ochsenius s.n. (9) (BREM); Ortega s.n. (17) (8GO- 
s.n. (7) (SGOA nu 

Parra 1 (11), 168 (13); 
Penne 1i | re (9), 1234] 


19831), 


Parra, B. 


ii s.n.( 


476 (17); Pearce s.n. (2) 
2); Pfister 277 (20a). 480 


— 


S. n. (6), s.n. Sin. 

(1 n. s.n. i s.n. Al 1). s.n. ie s.n. E Philippi, F. s.n. 
(9). s.n. (J). s 8): Philippi, 95 S. n. 
(3). 327 (1 3.5 
s.n. (13) p. O- 40936). s.n. (1 - n -] 16066), s. 
116061), s.n. (20a). Sd ys . (6) (8GO-4« iw s.n. 
(SGO- 49905). s.n. 0 (BRE M. s.n. (9) (FI), s.n. (9) (S 
49858), s.n. (9) (SGO- qs . . (9) E 110007 de 
M); Pinto s.n. (8); Pisano 2545 (15), 2552 (9), 27 
. 3169 (9), 3187 O) EAR D 0 (9). : 
4044 (9), 


S: 
w 
Si 
— 
— 
— 


33 (16), 6795 (13), 


78 
vs Y A (19), 8639 (16); 
1 (3 


T 


Poe 'ppig 131 (18), 45 (8), 517 ( 
Quezada e E (20b), 228 (9), : m a 
Reiche s.n. (20a), s.n. (3); 

(9), 722 ae 966 Or 971 (5), 


5232 (6), 5 (9). 5658 (5. $ 
Rodríguez E s 558 (20a), 843 (11), 870 (11). 1050 (11). 
3184 (9), 3715 (9), 4077 (16); Ruiz n 10/ 
10/82 1 0 10/83 (13); Ruiz Leal 122 (17 

n. (9). s.n. (13); Samsing s.n. (9); Savatier s.n. a (P- 
106; 095 Schajov oy 94 (5); Schlegel 2318 (9), 3577 


3577 (5). 
3583 (11). 7926 (20b), 8134 (19), 8197 (16); Schwabe 179 


m 


Volume 93, Number 3 
2006 


Marticorena 
Acaena en Chile 


453 


(13), 70 (2); Seki 310 (13); 5 1. (3); Simon 199 (18); 
E 24 (7), 32 (9), 175 o 208 (19), 225 (16), s.n. (9) 
(UPS), s.n. (10) (S-R8057); 9 A (2), 10873 (5), 116 
(3); B s.n. (LP-11942 577); Squeo 88068 (9); 
Stuessy 5 (3), 7 ; , 6443 (3), 9064 (11), 
9121 (10), 9211 (10), Ee 1), 9278 (11). 9323 (10), 9365 
(10), 9416 (10), 9448 (10), 11777 (11), 15560 (6). 

Taylor 1948 (18), 10380 (3), 10382 (9), 10383 (13), 10907 
(13); TBPA 50 (12), 223 (9), 2397 (14) 2445 (17), 2537 (9), 
5), C 3 (11); Tsuji 284 (13). 

Valdebenito 6432 (10), 1 uo 9132 (10); Villagrán 
6855 (2). 8176 0 D Villanueva s.n. (9); Vole a s.n. (1); 

337 (13); Von Thüngen 102 (14). 5 

Weldt 697 (1), 730 (11). 

Zoellner 1474 (9), 5016 (20b), 5558 (6), 6476 (1), 6793 
(9), 7566 (14), 10142 (18), 10379 (11). 


St — 


APÉNDICE 2. INDICE DE NOMBRES CIENTÍFICOS. LOS NOMBRES ACEP- 
TADOS ESTÁN EN NEGRITA Y LOS SINÓNIMOS EN CURSIVAS 
Acaena Mutis ex L. 
> caen 
Secc. Ac robyssinae Bitter 


Secc. Ancistrum (J. R. Forst. et G. Forst.) DC. 


Secc. 1 a " C jterne 


Secc. Pteracaena Bitter 
Secc. Rhopalachanta Bitter 
Secc. Subtus COS Bitter 
acrobyssina Bitte 
adscendens V ahl 


— 


var. macrochaeta Franc 
affinis Hook. f. 
alpina Poepp. ex Walp. 
anserinifolia (J. 7 Forst. et G. Forst.) J. Armstr. 
antarctica Hook. 

var. 1 s Bitter 


1 


argentea Ruiz et Pav. 
var. breviscapa Bitter 
var. contracta Bitter 
var. coriacea Bitter 
var. grandiceps Bitter 
var. 1 pinnata Bitter 
var. lanigera 
var. subcalvescens Bitter 


forma nigricans Bitter 
forma viridis Bitter 
cadilla Hook. f. 
caespitosa a s ex Hook. et Arn. 


y 
colchaguensis Bitter 
coxt Phil. 
cuneata Hook. et Arn. 
depauperata Bitter 


digitata Phil 
var. latifoliolata Bitter 
var. subpinnata Bitter 


fuegina Phil. 


i. alia Bitter 
^hi 


integerrima Gillies ex Hook. et Arn. 

var. ee Bitter 
ischnostemon Bitte 
krausei Phil. 

var. hirsuta (Phil.) Reiche 

var. massonandra Bitter 

var. metonandra Bitter 

subvar. eas Bitter 

laevigata W. T. 
le pd 'antha Phi 
ee Bitter 


— — 


E 


Mord connectens Bi itter 


var. conferta Bitter 

var. longissima Bitter 

var. negeri (E. Duse) Bitter 
leptophylla Phil. 


longiaristata H. Ross 
hil 


lucida (Lam.) Vahl 
var. abbreviata Bitter 
var. parvifolia (Phil.) Reiche 
var. villosula Bitter 
macrocephala Poepp. 
var. negeri E. Duse 
macrophyes Bitter 
macropoda Bitter 
macrostemon Hook. f. 
subsp. closiana (Gay) Bitter 


subsp. longiaristata (H. Ross) Bitter 


var. basipilosa Bitter 
magellanica (Lam.) Vahl 
subsp. pygmaea Bitter 
subsp. venulosa (Griseb.) Bitter 
var. glabrescens Bitter 
masafuerana Bitter 
Mele ra Bitte 
microcephala Se "m 
multifida Hook. f. 
subsp. intercedens Bitter 
subsp. „ Bitter 
var. utrinquepilosa Bitter 
ovii a 
eglecta Bitter 
obtusiloba Bitter 


ovalifolia Ruiz et Pav. 


n var. venulosa (Griseb.) Reiche 


454 Annals of the 
Missouri Botanical Garden 


subsp. elegans (Gay) Bitter rubescens Bitter 
var. elegans (Gay) Reiche ericea J., 
parvifolia Phil. 

patagonica A. E. Martic. 
pearcei Phil. 


"An e Bitter 
var. gracilis Bitter 


var. sericella Bitter 


var. glabrinervis Bitter aen ndens Hook. et Arn. 
petiolulata Phil. 5 1 
pinnatifida Ruiz et Pav. var. brevisericea 
subsp. herir Bitter var. gracilis Bitter 
subsp. hypoleuca Bitter var. ee Bitter 
subsp. liocarpa Bitte subflaccida Bitte 
subsp. longifolia (Phil) Bitter lene Alboff 
subsp. nudiscapa Bitter subsp. ae Bitter 
subsp. oligacantha Phil.) Bitter tenuifolia Bitte 
var. calcitrapa (Phil.) Reiche tenuipila Bitter 
var. eucantha (P i ) Reiche trifida Ruiz et | 


var. inl ia Bi var. argente Mla Bitte 
var. leptophylla N il. ) Reiche var. E Binter 
tongo n |.) Reiche . glabrescens Regel et Korn. 
Var. Macri iller 
var. la (Phil ) Reiche 
var. parvifrons Bitter 
var. uspallate nsis Bitter 
dlatyacantha Speg. 


var. trift 
poeppigiana Gay triglochin Bitter 
pumila Phil venulosa Griseb 


Seb. 
pumila Vahl Ancistrum J. R. Forst. et G. Fo 
var. acrocoma Bit anserinifolium J. R. Forst. et va Forst. 
var. glaberrima (P ^il ) Bitter idum. Aiton 
yurpureistigma Bitter 


r] am. 
quinquefida Phil. mage 1 Lam. 


SCALE-DEPENDENT 
CLASSIFICATION OF XERIC 
LIMESTONE PRAIRIES: ANNUAL 
OR PERENNIAL GRASSLANDS?! 


Patrick J. Lawless,” 
C. Baskin?" 


and Carol 


> Jerry M. Baskin,” 


ABSTRACT 


Xeric limestone prairies (XLPs) 


Missouri and Pennsylvania south to Arkansas and Georgia (l 
classification of XLPs due to their similarity with jocis cedar lade (CCS), an edaphic climax community t 
to unglaciated southeastern United States. Although C4 perennial grasses typica 
sites with extremely shallow soils (S 0.1 m) are typified by local dominance of Sporobolus vaginifloru 


are herbaceous plant communities that occur on shallow rocky calcareous soils from 


There has been considerable confusion regarding the 
ype restricted 
ally are the dominant taxa in XL Ps, portions of 


Torr. ex Gray) Alf. 


Wood, a Cy summer annual grass that is the characteristic dominant in LCGs. In this study, we assess 1 relative . 


of Cy perennial grasses versus 5. vaginiflorus in XLP community types identified in Kentucky over a range of scales. 
XLP vegetation and physical environmental conditions and to further compare IIb. 
01 


use these data to analyze variability in 
with LCGs. Sporobolus vaginiflorus had | 
identified in this study. 
all 12 ol 


extremely shallow-soiled regions in XLPs 


f the community types identified at the largest scale 


successional community types dominated by cryplograms and/or C 


total range of environmental conditions and the variety 
considered when classifying the vegetation of 


Key words: cedar glades, community classification. 


ugh frequencies in many of the fine-scale community types (0. 
However, dominance of C4 perennial grasses (partic ularly Schizachyrium nc (Michx.) Nash) in 


. XLPs also differ irom LCGs in that the 
and extent of plant community types present 


calcareous rock outcrop communities in the eastern United States 


rock oute rop. 


then 
and 0.1 m’) 
aall areal extent of 


100 m^) is indicative of the relatively 


— 


ormer generally lack Ce primary 


a annual forbs. Thus, the results of our study suggest that the 


in a site should be 


se “ale, xeric lime "stone prairie S. 


Herbaceous plant communities developed on shal- 
low rocky calcareous soils are broadly distributed in 
the Eastern. Deciduous Forest region (sensu Braun, 
1950) in the eastern United States. As is the case in 
many other types of rock outcrop vegetation, species 
composition and fine-scale structure in. calcareous 
rock outcrop communities are strongly dependent on 
local physical environmental conditions, particularly 
soil depth (e.g.. Quarterman, 1948; Skinner, 1979; 
Somers et al., 1986: Ver Hoef et al., 1993; Rollins, 
1997; Baskin & Baskin, 1999). Consequently, c 


fication and description of these plant communities 


assi- 
may be affected significantly by the scale at which 
community sampling is conducted and the heteroge- 
neity of physical environmental conditions within 
sites. In this study, we consider the effect(s) of grain 
(i. e, quadrat size) on the classification of species-rich 
xeric limestone 


1994; Baskin & 


calcareous grasslands known as 
prairies (XLPs) (sensu Baskin et al., 
Baskin, 2000; Lawless, 2005). 


distributed from Missouri (Nelson & 


XLPs are 
Ladd, 1983; Baskin € Baskin, 2000) and Pennsylva- 
nia (Laughlin & Uhl, 2003) south to Arkansas 


(Keeland, 1978: Baskin € Baskin, 2000) and Georgia 
(DeSelm, 1993). Baskin et al. (1994) and Baskin and 
Baskin (2000) described these herbaceous communi- 
lies as Cy perennial grasslands dominated by 
Schizachyrium scoparium (Michx.) Nash and/or Bou- 
teloua curtipendula Torr. However, transitions in 
dominance from C, perennial prairie grasses (S. 
scoparium, Sorghastrum nutans (L.) Nash, Andropogon 
gerardii iman, and B. curtipendula) to C4 annual 
spp.) 
response to decreased soil depths (Ver Hoef et al., 
1993; Baskin € Baskin, 2000; Rhoades et al., 2004). 
Dominance of Sporobolus vaginiflorus (Torr. ex Gray) 


Alf. Wood in the shallow-soil regions of some XLP 


grasses (Sporobolus R. Br. oflen occur in 


sites has resulted in confusion of this vegetation type 
with limestone cedar glades (LCGs) (sensu Baskin & 
Baskin, 1999, 2000, 2003, 2004), an edaphic climax 
vegetation type in which the C, annual grass 5. 
vaginiflorus is the characteristic dominant (Quarter- 
man, 1948; Somers et al., 1986; Rollins, 1997; Baskin 
& Baskin, 1999). 

In this study, we determined the effect of grain on 


the classification and description of XLP vegetation in 


! We thank Rick Gardner for f 


management and extraction. 


ield assistance at Crooked Creek Barrens and Joel 


Gramling for assistance with data 


? Department of Biology, University of Rentuc ky, Lexington, Kentucky 40506, U.S.A. patric ‘k.lawless@uky.edu 
* Department of Geography, University of Kentucky, Lexington, Kentucky 40506, U.S.“ 
Department of Plant and Soil Sciences, Unive rsity Val Kentuc nn Lexington, Kentuc ky 405 46, U.S.A. 


ANN. MissouRi Bor. 


GARD. 93: 455-464. PUBLISHED ON 24 OCTOBER 2006. 


Annals of the 
Missouri Botanical Garden 


the Interior Plateau. ecoregion in Kentucky (sensu 
Woods et al., 2002). We assess the relative impor- 


tance of C4 perennial grass species versus Sporobolus 
vaginiflorus in XLP community types identified in this 
study for comparison with XLP and LEG community 
types described in other studies. Additional biotic 
variables and physical environmental conditions also 
are considered as potential sources of variability 
within and among XLP sites and between XLPs and 
LCGs. Our research affords insight into the primary 
source(s) of variability within and among XLPs and 
identifies key compositional and structural differences 
between XLPs 


tion of these vegetation ly pes. 


and LCGs, thus facilitating classifica- 


in Kentucky, because 


(1) numerous XLP sites occur in the state, (2) many of 


This study focused on XLPs 


these XLPs are owned and/or managed by state 
agencies and/or conservation organizations, thus 


affording relatively easy access, and (3) quantitative 
vegetation data are lacking for the vast majority of 
sites in the state. It should also be noted that the 
of both XLPs and LCGs Kentucky has 


resulted in considerable confusion regarding the 


presence 


classification of herbaceous vegetation associated 
with calcareous rock outcrops in this area (Baskin et 


al., 1994). 
METHODS 


SAMPLING 


2003. the 
Plateau 


for sample site 


July through October of 2002 and 
vegetation of 18 XLPs in the 
Kentucky (see Table 


using a nested quadrat 


Interior 
ecoregion 
locations) was sampled 
sampling design adapted from Peet et al. (1998) 
(0.01-, 0.l-, l- 
of all nonwoody vascular plant species was recorded 
in the 0.01- to 10-07 
the largest quadrat (100-m* 


10-, and 100-m° quadrats). Presence 


quadrats. Each taxon present in 
sampling module) was 
assigned to one of 10 cover classes. Soil depth was 
recorded to the nearest centimeter in the four corners 
of all modules, and soil samples were collected from 
25 Soil 


25 cm in the same locations. 
composited and 


the top of soil 


samples from each module were 
submitted to the University of Kentucky Soil Testing 
Laboratory for the following analyses: pH, buffered 
pH, texture, water holding capacity, percent organic 
cation exchange capacity (total 
and Na), base 
K, Ca, Mg, Zn, 


Cu, Mn, Na, Fe, and Al. Three measurements of slope 


matter, total nitrogen, 


and individually for K, Ca, Mg. 
saturation, and concentrations of P, 
(96) and aspect of : slope (in degre es east of true north) 
were collected for each module using a compass and 


clinometer, respectively, and an average was calcu- 


e 18 


Ecoregions 


locations of 

| Kentucky. 

follow Level IV ecoregions of the ivo Plateau (Level HI 
2 


2002). 


and ec 'oregional 


Table 1. 


xeric limestone prairies sampled ir 


County 


ecoregion; sensu W oods el al., 


"ite County [ecoregion 


Hardin Knobs-Norman 


| Fort Knox Military 


Reservation | plan 
2 Fort Knox Military Hardin Knobs-Norman 
Reservation 2 Uplan 


Knobs-Norman 


Upland 


3 Cedar Creek Farms Hardin 


4 Scudder Glade Hardin Knobs-Norman 
Upland 

5 Hardin Co. Cedar Knobs-Norman 
Glade Upland 

Knobs-Norman 
Upland 


Hardin 


6  Muldraugh's Barren Hardin 


7 Mixed Grass Barrens Larue Knobs-Norman 
Upland 
9 Spalding Glade Larue Knobs-Norman 
Upland 
9 Med Creek Larue Knobs-Norman 
Glade Upland 
10 Pine Creek Barrens Bullitt Outer Bluegrass 
II Crooked Creek Lewis Outer Bluegrass 
Barrens | 
12 Crooked Creek Lewis Outer Bluegrass 
Barrens 2 
13 Grayson Co. Barren Grayson Crawford-Mammoth 


Cave Uplands 


]4 — Knights Barren Hardin Crawford-Mammoth 
Cave Uplands 

15 Lapland Barrens Meade Crawford-Mammoth 
Cave Uplands 

l6 Lapland Road Barrens Meade Crawford-Mammoth 


Cave Uplands 


17 Logan Co. Glade Logan Western Pennyroyal 
Karst Plain 
18 Logan Co. Barrens Logan Western Pennyroyal 


Lurst Ple 


lated for each parameter. The average aspect of slope 
(D) for 
a corresponding heat load index (HLD using the 


= (1 cos (S — 45)/2 


each module then was used to calculate 


q 


following equation: HLI = 


CLASSIFICATION AND ORDINATION 


The flexible beta linkage agglomerative clustering 
method in PC-ORD (MJM Software Design; McCune 
& Mefford, 1999) was used 
types over the range of spatial scales sampled. A beta 
value of —0.25 


all analyses. Indicator species analysis was 


— 


to identify community 
and Sorensen's distance measure were 
used 
used to determine the optimal number of groups at 
each scale via selection of the group number with the 


lowest average P value (Dufréne and Legendre, 1997). 


Volume 93, Number 3 
2006 


Lawless et al. 
Xeric Limestone Prairies 


Nonmetric multidimensional scaling (NMS) (Kruskal, 
1964; Mather, 1976) was used to determine variability 


n community structure at the 100-m? scale and to 
investigate potential sources of variability. A random 
starting configuration and Sorensen's distance mea- 
sure were used in the analysis of ground-layer cover 
class data, which included 40 runs with real data and 
50 with randomized data (PC-ORD: MJM 
Software Design; McCune & Mefford, 1999). Abiotic 
environmental data and cover 


runs 


class data from the 
primary matrix were entered into a second matrix for 
calculation of coefficients of determination (r^) with 
ordination axes. Only variables with an value greater 
than or equal to 0.35 were included in joint plots. 


RESULTS 
CLUSTER ANALYSIS 


. Um Community types 


[2m 


Of 174 0.01-m? quadrats sampled, 159 contained 
one or more individuals and were included in cluster 
analysis. Four community types were identified at the 
0.01-m? scale, and percent occurrence data resulted 
30), 


all of which are graminoids (see summary of charac- 


in only four characteristic taxa (£6 occurrence = 


teristic taxa in. Table 2) (full analysis of community 
types at each of five scales is available from the authors 
by request). Three taxa, Schizachyrium scoparium, 
1 8 5 vaginiflorus, and Carex crawei Dewey ex 

, were characteristic in two community types, and 
characteristic 


Pu wan nutans was a taxon it 


~ 


a single community type. The 5. scoparium—S. vagini- 
florus community type contained more quadrats (48) 
and occurred in more sites (15) than any other 
community type. The S. vaginiflorus-C. crawei com- 
munity type contained 46 quadrats and occurred in 14 
of 18 sites. The community type present in the smallest 


— 


number of sites (8) and quadrats (22) supported no 


characteristic taxa. 
0.1-m? Community types 

Of 174 0.l-m° quadrats, 173 contained one or more 
individuals and were included in cluster analysis. Ten 
community types were identified at the 0.1-mY scale, 
and percent occurrence data resulted in a total of 19 
characteristic taxa (£6 occurrence = 40) among these 


community types. Schizachyrium scoparium was 
a characteristic taxon in six of these community 
types, more than any other taxon, and was followed by 
Sporobolus vaginiflorus (4) and Carex crawei (3). Eight 
taxa had a percent occurrence greater than or equal to 


all 


including S. 


other taxa in one or more community types, 


scoparium (4 community types), 5. 


vaginiflorus (2), Andropogon gerardii (1), Sorghastrum 
nutans (l), C. Dewey (1), C. crawei (J). 
Ruellia humilis Nutt. (1), and Rosa carolina L. (1). 
The S. scoparium—S. vaginiflorus-Echinacea simulata 


McGregor—Hedyotis nigricans (Lam.) Fosberg commu- 


meadi 


— 
— 


nity type and the 5. vaginiflorus-Heliotropium tenel- 
lum Torr. Hypericum dolabriforme Vent. community 
type occurred in the largest number of sites (9). The C. 
crawei=5. scoparium-—S. vaginiflorus-h. humilis com- 
munity type was represented in the largest number of 


quadrats (40). 


I Community types 


All 174 of the 1-m? quadrats sampled contained one 
or more individuals and were included in cluster 
analysis. Ten community types were identified at the 
I-m? scale, and percent occurrence data resulted i 
a total of 39 characteristic taxa (£6 occurrence = 50) 
among these community types. Schizachyrium scopar- 
ium was a characteristic taxon in nine of 10 com- 
munity types and was followed by Sporobolus vagi- 


niflorus (6 community types), Echinacea simulata (5) 


— 
5 


Euphorbia corollata L. 
Benth. 


humilis (4). Schizachyrium scoparium had a percent 


(5) Physostegia virginiana 
(5) Sorghastrum nutans (4), and Ruellia 
O) 8 ) 


occurrence greater than or equal to all taxa in five 
Other 
rence greater than or equal to all taxa in two or more 
community types included E. simulata (4 community 
types) and S. vaginiflorus (2). > 


community types. taxa with a percent occur- 


The S. scoparium—E. 
simulata—S. vaginiflorus community type occurred in 
the most sites (8) and was represented in the largest 
number of quadrats (51 


— 


10-m?* Community types 


Twelve community types were identified at the 10- 
mn? scale, and percent occurrence data resulted i 


1 


— 


a total of 73 characteristic taxa (% occurrence = 60) 
among these community types. Schizachyrium scopar- 
ium was a characteristic taxon in all 12 community 
types. Only eight additional taxa were characteristic 

seven or more community types, including Sporobolus 
vaginiflorus (8), Echinacea 
(8), Euphorbia 
corollata (8), Ruellia humilis (7), Solidago nemoralis 
Aiton (7), and Juniperus virginiana L. (7). Schizachyr- 


ium scoparium had a percent occurrence greater than 


Sorghastrum nutans (8), 
simulata. (8), Physostegia virginiana 


or equal to all other taxa in 8 community types and was 
followed by S. vaginiflorus (4) and E. simulata (4). The 


S. scoparium—S. nutans-E. simulata-S. vaginiflorus 


community type occurred in five sites in the Knobs- 


Norman Upland ecoregion and was represented i: 


more quadrats (36) than any other community type. 


458 Annals of the 
Missouri Botanical Garden 


D 


Table 2. Summary of characteristic taxa identified at all five spatial scales (0.01-100 m^) in XLPs in Kentucky, U.S.A. 


Taxonomy is in accordance with the United States Department of Agriculture, Natural Resources ( 


conservation Service (2004). 


All voucher specimens were collected in. Kentucky and are deposited at KY. Site localities are in accordance with site 


numbers listed in Table 1 
Family Species Site locale Date Specimen ID 
Acantliaceae Ruellia humilis Nutt. 17 3 July 2002 Patrick Lawless s.n. 
Agavaceae Manfreda virginica Salish. 17 3 July 2002 Patrick Lawless s.n. 
Alliaceae Allium cernuum Roth l 25 Aug. 2003 Patrick Lawless s.n 
Anacardiaceae Rhus aromatica Aiton 10 14 June 2002 Patrick Lawless s.n 
Apocynaceae Apocynum cannabinum L.. | 10 July 2002 Patrick Lawless s.n 
Asclepiadaceae Asclepias viridiflora Ral. 8 11 Aug. 2003 Patrick Lawless s.n. 
Asteraceae Ambrosia artemisiifolia V. 13 20 Aug. 2002 Patrick Lawless s.n. 
Brickellia eupatorioides (L.) 17 3 July 2002 Patrick Lawless s.n. 
Shinners 
Coreopsis tripteris L. 17 10 July 2002 Patrick Lawless s.n. 
Echinacea simulata McGregor ni 5 July 2002 Patrick Lawless s.n. 
Eupatorium altissimum U. 17 3 July 2002 Patrick Lawless s.n. 
Helianthus hirsutus Ral. 4 10 July 2002 Patrick Lawless s.n. 
Liatris aspera Michx. 4 10 July 2002 Patrick Lawless s.n. 
Liatris cylindracea Michx. 11 25 July 2002 Patrick Lawless s.n. 
Liatris squarrosa Michx. 10 l4 June 2002 Patrick Lawless s.n. 
Liatris squarrulosa Michx. 10 20 Sep. 1991 Barry Dalton s.n. 
Oligoneuron rigidum var. glabratum | 10 July 2002 Patrick Lawless s.n. 
(E 3raun) G. L. Nesom 
Parthenium integrifolium L. 4 10 July 2002 Patrick Lawless s.n. 
Silphium pinnatifidum Elliot 9 7 Aug. 2003 Patrick Lawless s.n. 
Silphium terebinthinaceum Jacq. B 25 June 2002 Patrick Lawless s.n. 
Silphium trifoliatum L. 14 28 July 1978 Ray Cranfill s.n. 
Solidago nemoralis Aiton 5 20 Sep. 2003 Patrick Lawless s.n. 
id laeve (U.) A. Love 4 10 July 2002 Patrick Lawless s.n. 
N D. 
do sum urophyllum (DC. | 10 July 2002 Patrick Lawless s.n. 
G. L. Nesom 
Boraginaceae Heliotropium tenellum Torr. 17 3 July 2002 Patrick Lawless s.n. 
Lithospermuan canescens Lehm. 17 3 July 2002 Patrick Lawless s.n. 
Caesalpiniaceae Cercis canadensis L. 6 22 Sep. 2002 Patrick Lawless s.n. 
Campanulaceae Lobelia spicata Lam. 17 5 July 2002 Patrick Lawless s.n. 
Clusiaceae Hypericum Ae Vent. 17 3 July 2002 Patrick Lawless s.n. 
Cornaceae Cornus drummondii C. A. Mey. 12 26 July 2002 Patrick Lawless s.n. 
Cupressaceae Juniperus virginiana L. 6 22 Sep. 2002 Patrick Lawless s.n. 
Cyperaceae Carex crawet Dewey ex Torr. 6 19 May 2002 Patrick Lawless s.n. 
Carex meadii Dewey 17 3 July 2002 Patrick Lawless s.n. 
Garex umbellata Schkuhr 6 19 May 2002 Patrick Lawless s.n. 
Fimbristylis puberula (Michx.) Vahl 10 14 June 2002 Patrick Lawless s.n. 
Ebenaceae Diospyros virginiana L. 13 26 Aug. 2002 Patrick Lawless s.n. 
Euphorbiaceae Croton capitatus Michx. 9 13 June 2002 Patrick Lawless s.n. 
Croton PN MR Michx. 4 10 July 2002 Patrick Lawless s.n. 
Euphorbia corollata L. 10 14 June 2002 Patrick Lawless s.n. 
Fabaceae Desmodium glabellum DC. 11 25 July 2002 Patrick Lawless s.n. 
Galactia volubilis (U.) Britton 17 7 July 2002 Patrick Lawless s.n. 
Lespedeza virginica (L.) Britton 17 3 July 2002 Patrick Lawless s.n. 
Melilotus alba Medik. 11 3 July 2002 Patrick Lawless s.n. 
Gentianaceae Sabatia angularis (L.) Pursh | 10 July 2002 Patrick Lawless s.n. 
Iridaceae Sisyrinchium albidum Raf. | 10 July 2002 Patrick Lawless s.n. 
Lamiaceae Blephilia ciliata Rat. 10 14 June 2002 Patrick Lawless s.n. 
Isanthus brachiatus Britton, Sterns | 10 July 2002 Patrick Lawless s.n. 


& Poggenb. 


Monarda fistulosa L. 17 > July 2002 


Patrick Lawless s.n. 


Volume 93, Number 3 
2006 


Lawless et al. 
Xeric Limestone Prairies 


Table 2 Continued. 


Family Species Site locale Date Specimen ID 
Physostegia virginiana Benth. 7 6 Aug. 2003 Patrick Lawless s.n. 
Pycnanthemum tenuifolium Schrad. 17 5 July 2002 Patrick Lawless s.n. 
Linaceae Linum sulcatum Riddell 17 3 July 2002 Patrick Lawless s.n. 
Oleaceae Fraxinus americana L. 13 3 June 2002 Patrick Lawless s.n. 
Onagraceae Gaura filipes Spach 4 10 July 2002 Patrick Lawless s.n. 
Poaceae Andropogon gerardii Vitman 4 10 July 2002 Patrick Lawless s.n. 
Aristida purpurascens Poir. 10 — Julian Campbell s.n. 
Bouteloua curtipendula Torr. 17 3 July 2002 Patrick Lawless s.n. 
Danthonia . ata (L.) P. Beauv. ex | 18 June 2003 Rob Paratley s.n 
Schult. 
Dani acuminatum (Sw.) 9 13 June 2002 Patrick Lawless s.n. 
Gould & C. A. Clark 
1 sphaerocarpon (Ellis) 17 3 July 2002 Patrick Lawless s.n. 
Gould 
Panicum flexile (Gattinger) Scribn. | 10 July 2002 Patrick Lawless s.n. 
Schizachyrium scoparium (Michx.) I 10 July 2002 Patrick Lawless s.n. 
Nash 
Sorghastrum nutans (L.) Nash | 10 July 2002 Patrick Lawless s.n. 
Sporobolus compositus Merr. 17 7 July 2002 Patrick Lawless s.n. 
Sporobolus vaginiflorus (Torr. ex 1 10 July 2002 Patrick Lawless s.n. 
ray) Alf. Woo 
Polygalaceae Polygala verticillata L. 17 3 July 2002 Patrick Lawless s.n. 
Ranunculaceae Clematis pitcheri Torr. & A. Gray 17 5 July 2002 Patrick Lawless s.n. 
Rosaceae Potentilla simplex Michx. 17 5 July 2002 Patrick Lawless s.n. 
Rosa carolina | 14 26 June 1992 Julian Campbell s.n. 
Rubiaceae Galium circaezans Michx. 4 10 July 2002 Patrick Lawless s.n. 
Hedyotis nigricans (Lam.) Fosberg 4 10 July 2002 Patrick Lawless s.n. 
Houstonia canadensis Willd. ex 17 3 July 2002 Patrick Lawless s.n. 
Roem. & Schult. 
Santalaceae Comandra umbellata Nutt. 10 14 June 2002 Patrick Lawless s.n. 
Serophulariaceae Agalinis tenuifolia Raf. 10 8 Sep. 1992 Julian Campbell s.n. 
Castilleja coccinea Spreng. 12 25 July 2002 Patrick Lawless s.n. 
Smilacaceae Smilax bona-nox | 17 3 July 2002 Patrick Lawless s.n. 
Ulmaceae Celtis tenuifolia Nutt. 17 3 July 2002 Patrick Lawless s.n. 
Violaceae Viola pedata L. 17 3 July 2002 Patrick Lawless s.n. 


100-m° Community types 


Forty-one taxa had an average cover value of 3% or 
greater in one or more community types. Fifteen taxa 
were dominant (avg. cover = 10%) in one or more 
community type(s). Six of these 15 taxa were dominant 
Schi- 


zachyrium scoparium (10 community types). Sporobolus 


in three or more community types. including 


vaginiflorus (5), Sorghastrum nutans (5), Echinacea 
simulata (5), Andropogon gerardii (4). and Carex erawei 
(3). Schizachyrium scoparium had the highest average 
cover in 10 of 12 community types identified in this 
1 in 20 of 22 other studies of XLPs 


eastern United States for which vegetation data were 


study in the 


available (Table 3). Andropogon gerardii and Silphium 


terebinthinaceum Jacq. each had the highest cover i 
one I00-m? community type. Schizachyrium scoparium 


five of 12 


and 5. vaginiflorus were co-dominant in 


community types, and 5. scoparium and 5. nutans were 
co-dominant in four of 12 community types. Only seven 


three or more community 


taxa were subdominant it 


types. Andropogon gerardii was subdominant in seven 


community types, and six taxa, including Æ. simulata, 


C. meadii, Hedyotis nigricans, S. nutans, Juniperus 


virginiana, and Cercis canadensis L., each were 


subdominant in three community types. 


ORDINATION 


Monte Carlo test results were statistically signifi- 
cant (P 


analyzed. 


0.02) for each of the six ordination axes 
The final NMS ordination consisted of 400 
iterations of a three-dimensional solution and resulted 


in final stress equal to 15.33 and instability equal to 


2 


0.00829. Percent of variance (equal to cumulative r 


X 100) collectively explained by the three axes was 


460 Annals 
1 . Garden 


Table 3. Summary of dominant taxa in XLPs of the eastern United States. Data (quantitative and qualitative) are 


organized by Level HI ecoregions (sensu Omernik, 1987). 
Ecoregion Dominant taxon Reference Data 
Ozark Highlands Sporobolus neglectus* Hall. 1955 frequency 
Schizachyrium scoparium Kucera & Martin, 1957 avg. cover 
Schizachyrium scoparium Keeland, 1978 avg. cover 
Schizachyrium scoparium? Skinner, 1979 avg. frequency 
Sporobolus neglectus Skinner, 1979 avg. frequency 
Schizachyrium scoparium Hicks, 198 avg. imporlance value 
Schizachyrium scoparium Logan, 1992 semi-quantitative 
Schizachyrium scoparium Ver Hoef et al., 1993 avg. geometric mean 
of cover! 
Panicum virgatum* George, 1996 ave. cover 
Schizachyrium scoparium George, 1996 avg. cover 
Eastern Corn Belt Plains Schizachyrium scoparium Maxwell, 1987 qualitative 
Interior River Hills and Schizachyrium scoparium McClain & Ebinger. 2002 avg. cover 
Lowlands Bouteloua curtipendula McClain & Ebinger, 2002 avg. cover 
Hedyotis nigricans McClain & Ebinger, 2002 avg. cover 
Schizachyrium scoparium Heikens, 1991 avg. cover 
Interior Plateau Schizachyrium scoparium Heikens, 1991 avg. cover 
Silphium terebinthinaceum Heikens, 1991 avg. cover 
Schizachyrium scoparium** Kurz, 1981 avg. frequency 
Schizachyrium scoparium Braun, 1928 avg. frequency 
Andropogon gerardii Braun, 1928 avg. frequency 
Schizachyrium scoparium Baskin & Baskin, 1977 frequency 
Schizachyrium scopartum DeSelm, 1988 importance values 
Schizachyrium scopartum DeSelm, 1991 avg. cover 
Schizachyrium scopartum DeSelm € Webb, 1997 avg. cover 
Sporobolus clandestinus* DeSelm & Webb, 1997 avg. cover 
Schizachyrium scoparium this study avg. cover 
Indropogon gerardii this study avg. cover 
Silphium terebinthinaceum this study avg. cover 
Ridge and Valley Bouteloua curtipendula Bartgis, 1993 avg. cover 
Solidago arguta var. harrisii” Bartgis, 1993 qualitative 
Monarda fistulosa var. brevis* Bartgis, 1993 qualitative 
Paronychia virginica” Bartgis, 1993 qualitative 
Schizachyrium scoparium DeSelm, 1993 avg. cover 
Andropogon gerardii DeSelm. 1993 avg. cover 
Schizachyrium scoparium Ludwig, 1999 avg. cover 
Indropogon gerardii Ludwig, 1999 avg. cover 
Schizachyrium scoparium Allison & Stevens, 97 qualitative 
Bouteloua curtipendula Laughlin & Uhl, semi-quantilalive 
Southeastern. Plains Schizachryium scoparium Harper, 1920 qualitative 


* Authors and families of taxa that are in this table but do not appear in Table 2 are as follows: Sporobolus neglectus Nash 
1 8 . Panicum virgatum L. (Poaceae), Sporobolus clandestinus Hitchcock (Poaceae), Solidago arguta. Spreng. vi 
harrissii (E. S. Steele) A. W. Cusick (Asteraceae), Monarda fistulosa var. brevis Fosber & Artz (Lamiaceae), Pata 


y fend 
+ Based on frequency date: collected in 0.01 and O. He quadrats centered around the rare focal species Stenosiphon 
owe (Nutt. ex E. James) Heynh. (Onagraceae). 

t Based on frequency data collected in 0.01 and . Une quadrats centered around the rare focal species Penstemon cobaea 


purpureus Pennell (Se pp sacra ) and Centaurium texense (Griseb.) Fernald (Asteraceae). 
Twenty-three of 32 sample sites were located in the Interior Plateau and nine in the Ozark Highlands: however, data were 


nol stratified by site. 


in the eastern 


82.9%, and the increment r^ values of the three axes Only quadrats from the two sites i 
were as follows: | = 0.234, 2 = 0.226, and 3 = Outer Bluegrass ecoregion (Crooked Creek Barrens 
7 o 


0. Orthogonality between axis pairs ranged from Sites J and 2) had relatively large negative cigenvec- 
97.2% to 99.196. tors (~ —1) on axis three (Fig. 1). and as a result the 


Volume 93, Number 3 
2006 


Lawless et al. 
Xeric Limestone Prairies 


Axis 3 


18.1 
ECHSIM LANE T AC 
SPOVAG  .. A 


4 
PY 


3.9 jg 


HYPPRO 


Depth 


Figure J. Joint plot (axis 1 vs. 3) incl 
0.35. Each 100-nm? quadrat (i 


Axis 1 


uding abiotic environmental ariables and taxa with 7” values greater than or ec ual 
„Sample module) is repre 


al to 
d by a triangle labeled with site number (from Table 1) followed 
by module number within site. Taxon codes are formed by the first three letters of the genus and species epithet, respectivel 
pth” symbolizes soil depth. 


distances between these quadrats and all others were 
high. There was a strong negative correlation betwee 
soil depth and axis three, and cover of 11 taxa also 
was correlated (1? value = 0.35) with axis three. Cover 


of Silphium terebinthinaceum, Symphyotrichum laeve 
(L.) A. Löve & D. Löve, Hypericum prolificum L., 
Pycnanthemum tenuifolium Schrad., 
cea Michx., 


Liatris cylindra- 
A. Mey., Castilleja 


, Cornus drummondii C. 


Annals of the 
com Botanical Garden 


coccinea Spreng., Oligoneuron rigidum var. glabratum 


(E. L Braun) G. 


DC. was correlated negatively with axis three, 


„ Nesom, and Desmodium glabellum 


and 


cover of Sporobolus vaginiflorus and Echinacea 


simulata was correlated positively with this axis. Axis 
three eigenvectors of quadrats from Logan pe 
Glade and Logan County Barrens were 


= 


arge (= (). 
Only quadrats from Pine Creek Barrens had FACH 


— 


arge positive eigenvectors (= 0.6) on axis two. Cover 


of Carex crawet, Aristida purpurascens Poir., Euphor- 
bia corollata, Symphotrichum urophyllum (DC.) G. L. 


Nesom, and L. Michx. 


positively with axis two. 


squarrosa was correlated 


DISCUSSION 


Results of this study confirm that XLPs in Kentucky 
are C4 perennial grasslands as described by Baskin el 
al. (1994) and Baskin and Baskin (2000). 


Sporobolus vaginiflorus being a dominant taxon (avg. 


Despite 


cover = 10%) in five of 12 OO community types, 
collective average cover of Cy perennial grasses 
(primarily Schizachyrium scoparium, Sorghastrum nu- 


tans, and Andropogon gerardii) exceeded cover of &. 


vaginiflorus in all 12 100-m° community types. 
Schizachyrium scoparium had the highest average 


100-m*^ 


in this study and was dominant (Le.. 


cover in 10 of 12 XLP community types 


had 


the highest cover, frequeney, or importance value) in 


identified 


of 19 studies conducted in XLPs in other regions 


of the eastern United States. Thus, 5. scopartum should 
be considered the characteristic dominant taxon in 
XLPs of the eastern. United States. 

However, Sporobolus vaginiflorus was a characteris- 


‘taxon in the majority of community types identified 


in this study at scales less than 100 m”. In one of four 


0.01-m? community types and two of 10 0.I-m* 
community types, S. vaginiflorus was the only 
characteristic grass taxon, thus indicating local 


dominance of this taxon in a considerable number of 


sites. This finding is consistent. with those of Hall 
(1955). Kucera and Martin (1957), Skinner (1979), 


Hicks (1981), Nelson (1985). Logan (1992), and Ver 
Hoef et al. (1993) on XLPs in the Ozarks, which were 


summarized by Baskin and Baskin (2000). Therefore, 


when considered at fine spatial scales, portions of 
some XLP sites fit the physiognomic description of 
cedar glades provided by Baskin and Baskin (2003). 

Sporobolus vaginiflorus and Schizachyrium scopar- 
in the 0.01-m? and 0.1- 
that 


ium were both characteristic 


m^ community types identified in this study 


contained the largest number of quadrats. Therefore. 
S. vaginiflorus is not necessarily confined to portions 


of sites in which perennial grasses are absent, which 


typically have extremely shallow (= O. m) soll 


depths, and apparently is capable of competing with 


these grasses over a range of local environmental 
conditions. Frequencies of both S. vaginiflorus and Ca 


perennial grasses (Bouteloua curtipendula and/or 
Sorghastrum nutans) also were high in the shallow- 
soil community types (“rocky glade” and “shallow soil 


Ver 


(1993). In addition. high frequency of co- 


glade”) on Gasconade Dolomite described. by 
Hoef et a 
occurrence of S. vaginiflorus and Andropogon virgini- 
cus L./S. scoparium (uncertainty resulting from spring 
sampling) was observed in one of the 14 0.1-m^ LCG 
community types identified ie Rollins (1997). 

In the NLPs sampled 1 


Sporobolus vaginiflorus cover also were significant 


1 Kentucky. soil depth and 


sources of variability at the largest grain (100 m). 
However, 13 of the 17 taxa with r^ values greater than 


or equal to 0.35 were herbs. This phenomenon is due 


small number of characteristic grami- 


argely to the 
noid taxa and their high percent occurrence values in 
this vegetation type. Thus, the patchy distribution of 
herbaceous taxa is largely responsible for differences 
the 


vast majority 


in Species composition among sites, because 
dominant graminoids are present in the 
of sites. This patehy distribution pattern most likely is 
attributable to the rarity of this vegetation type at the 
landscape scale and small area of XLPs in comparison 
with the regionally dominant deciduous forests and 
. 2004). 


colonization, extinction; and factors that affect such 


Therefore, 


agricultural lands (Lawless et : 


processes (e.g.. sile area, distance from nearest site, 


regional species pools). strongly influence species 


composition in XLPs and are primary sources of 
variability in this vegetation type throughout its range 


in the eastern. United States. 


COMPARISON OF XL Ps ano LEGS AND THEIR CLASSIFICATION 


The 


important component of both XLP a 


C4 annual grass Sporobolus 1 is an 


LOG commu- 
nity types. In XLPs, S. raginiflorus is n ally dominant 
= 0.1 


areas of greater soil depth, 


in areas with extremely shallow soils ( m), and it 
is abundant in where 
Schizachyrium scoparium (or other Cy perennial 
Lack of dominance 

100 mn 


types identified in this study suggests that extremely 


grasses) is the dominant taxon. 


of S. vaginiflorus in any of the community 


shallow soils are not areally extensive in XLPs. 
Conversely, dominance of S. vaginiflorus in LCG 
communily types identified over a range of scales, 


1948: Somers et al.. 


1997). is indicative of the pervasive- 


from 0.01 to 100 m? (Quarterman, 
98 


— 


Rollins. 


ness of extremely shallow soils in the LCG vegetation 


soil depth and the range of 


LCGs 


type. Therefore, 


soil depths in XLPs 


different. 


average 


and are considerably 


Volume 93, Number 3 
2006 


Lawless et al. 
Xeric Limestone Prairies 


XLPs and LCGs differ not only with respect to their 


characteristic dominant taxa and soil depth charac- 
teristics but also in the different community types that 
occur in each kind of vegetation and in the species 
Although 
lichens, blue-green algae (Nostoc commune 


composition of these community types. 


mosses, 
and macrobiotic 


Vaucher), crusts were observed 


some sites sampled in this study, they were not 


observed high frequency. Conversely, early-stage 
primary successional community types dominated by 
cryptograms are common in LCGs (Quarterman, 1948: 
1986; Rollins, 1997; Baskin & Baskin. 


1999). as well as many other rock outerop vegetation 


Somers et al., 


types in the eastern. United States, including alvars 


(Catling & Brownell, 1999), granite rock outcrops 


(Oosting & Anderson, 1939; Burbanck & Platt, 1964: 
Shure & Ragsdale, 1977; Shure, 1999), sandstone 


rock outcrops (Perkins, 1981), rock ledges (Winter- 
& Vestal, 1956), high-elevation rock 
outcrops of various substrates (Wiser & White, 
1999). In addition, XLPs generally lack the C3 annual 


forb-dominated seral stages that typically occur in 


ringer and 


LCGs subsequent to soil genesis by cryptogram 
1948; Somers et al., 
Baskin & Baskin, 1999), 


forb-dominated community 


5 types (Quarterman, 
; Rollins, 1997; 


» peur 


— 


These 


~ 


types are 
restricted to areas of LCGs with very thin soils over 
bedrock, 


a considerable 


and they are the primary habitat for 


number of taxa (e.g. Leavenworthia 
Torr. spp.) endemic or near endemic to LCGs of the 
southeastern United States (sensu Baskin & Baskin. 
1999, 2003). Thus, absence or infrequency of these 
physical environmental conditions and corresponding 
XLPs 


may explain the low frequency of occurrence of LCG 


C5 annual forb-dominated community types 


endemics and/or near endemics in XLPs of the eastern 
United States (Lawless, 2005). 

Community types identified in this study represent 
compositional and structural variation within and 
among XLP plant communities. Some XLP community 
types identified in this study and in others (particularly 
those dominated by Sporobolus vaginiflorus) closely 
resemble those present in LCGs of the southeastern 
United States. 
apparent. between XLPs 


However, distinct differences are 
and LCGs when the total 
range of environmental conditions and the variety and 
extent of plant community types present are consid- 
ered. Prevalence of C4 perennial grass-dominated 
community types, infrequeney or absence of early stage 
primary successional seres composed of cryptograms 
and/or 


— 


C3 annual forbs, and extremely variable soi 
depths (< 0.01 to > 1 m) are all characteristics of 
XLPs 


should be used in any classification system involving 


that distinguish them from LCGs and, thus, 


these calcareous rock outerop vegetation types. 


———— & . Webb. 1997. 


Literature. Cited 


Allison, J. R. & T. E 2001 
Dolomite outcrops in Bibb County. 
154—205. 


flora of Ketona 
Alabama. Castanea 06: 


. Stevens. . Vascular 


Bartgis, R. L. 1993. The limestone glades and barrens of 
West Virginia. C Pim a 58: 69-89. 

Baskin, J. M. & C. askin. 1977. An undeseribed cedar 
glade community s middle Tennessee. Castanea 42: 
140—145 


. 1999. Cedar glades of the southeastern 
Pp. 206-219 in R. Anderson, 5. 
. Baskin (editors), : 
Plant 


Lp ress, 


N 
United Rue 


Savannas, Barrens, 2nd 
N 


Communities of North America. 
( 


Cambades Uni ambridge. 


= 


000. Vegetation of limestone and 
dolomite glades in 15 Ozarks and Midwest regions of the 
United States. Ann. Missouri Bot. Gard. 87: 

———— & — 2003. The vascular flora of 3 glades 
of the southeastern United States and it 
ical oo J. Torrey Bot. Soc. 130: 

—— — € •ͤmů— 2004. 
and 


X ograph- 
101-118 
‘cedar des 
rocky 


Tennessee. Bot. 


History of the use of * 
other desc riplive tern Jr vege lation on 
limestone Ph in the Central Basin of 
hev. 70: 403— 

s ^ Chester. 1994. The 


Region of Kentucky M: Tennessee: 


Big Barrens 
Further observations 
and considerations. Castanea 59 
Braun, E. L. 1928. The vegetation of the Mineral Springs 
Region of Adams County, Ohio. Bull. Ohio Biol. Surv. 15 
Ol) 9% a 75 
195 


226-254. 


. Deciduous Forests of Eastern North America. 
Blakiston, E Pw cw 

Burbanck, M. B. Platt. 1964. Granite outerop 
communities E i E. dmont Plateau in Georgia. Ecology 
45: 292-306 


2 ). 
Catling, P. M. & V. R. Brownell. 1999, Alvars of the Great 
Lakes Region. Pp. 375—405 in R. Anderson, J. S. 
Fralish & J. M. Baskin (editors), and 


Savannas, Barrens, 
North 


Rock Outcrop Plant Communities of North America. 


Cambridge ie niv. Press; Cambric lec. 
DeSelm, H. R. Western Highland 


3. The barrens of the 


Rim of Tennessee. Pp. 199-219 in D. H. Snyder (editor) 
Proceedings of the First Annual Symposium on the 
Natural History of Lower Tennessee and Cumberland 
River Valleys. The Center for Field Biology, Austin Peay 


Clarksville, 
99]. Barrens of the Cental Basin of 
-20 in D. H. Snyder (editor). 

Routh Annual Symposium on the Natural History of Lower 


State Univ., 
19 Tennessee. 
Proceedings of the 


Tennessee and Cumberland River Valleys. The Center for 
Field Biology, Austin Peay State Univ... Clarksvill 
Tennessee. 


993 - Bartons and f: of the southern. Ridge 


8 Walley ». 81-136 in S. W. Hamilton, E. W. Chester 
A. E 80 ann (editos). Proceedings of the Fifth 
nnual n on the Natural History of Lower 


The Center 
Clarksville, 


zumberland River Valleys. 


Tenne ssc ane 
i Austin Peay State 


for H Field pos Univ., 
Tennesse a¢ 

Vegetation studies of barrens 
in Rcs Stern Alabama. Pp. 204-213 in A. F. Scott. S 
W. Hamilton, E. W. Chester & D. S. White (editors), 
Proceedings of the Seventh Annual Symposium on the 
Natural History of Low 
River Valleys. 
State Univ.. 


and Cumberland 
The ae for Field Biology, 


Clarksville, Tennessee. 


Tennessee 


Austin Pea 


Annals 
11 E unm Garden 


Dufréne, M. € P. 


indicator species: The 


Legendre. 1997, Species assemblages and 
need for a a ible asymmetrical 
approach. 1 55 pi d 07: 345— 

A Vegetative nea s of Two Central 
M.S. Thesis, Central Missouri 


George, J. R. 
Missouri | Amestone Glades. 
Warrensburg. 


Hall, M. T. 1955. 


Comparison of Juniper populations on an 


Ozark glade and old fields. Ann. Missouri Bot. Gard. 42 
171-194 

Harper, R. M. 1920. The limestone. prairies. of Wilcox 
County, Al ban: Ecology 1: 198-203. 


Natural Forest 
Southern 


991. Classification of the ! 


Heikens, A. L. 
i Ph.D. 


in Southern Illinois. Thesis. 
E n ale. 

A Vegetative Analysis of Hercules Glade 
M S. Thesis, Southwest Missouri State Univ., 
Keeland, B. D. 1978. 
; Northwest 
Miro 
Kruskal, J. B. 1964. 

numeric "i method. Psychometrika: 29: 1-27. 
vucera, C. L. & S. C. Martin. 1957. 
relationshi M ps in the glade 
Ecole 38: 285-291. 


1 


Vegetation and Soils in Calcareous 


Arkansas. M.S. Thesis, Univ. of 


> 
* 
E 
os 


Nonmetric multidimensional scaling: A 


soil 


southwestern 


Ve 'getation. anc 


Miss souri Ozarks. 
1. 1981 


lora of elades in Ilinois. 
teese (editors). The 
l k Pro- 
Conference. 

Ohio Biol. Biol. 


. Columbus, Surv. 


limestone 


8 C. F. Uhl. 2003. The 


prairies i aa Castanea 68: 300— 


Laughlin, | xeric 
16 


Lawless, P. J. 2005. Neric Limestone Prairies of Eastern 
United States. Ph.D. Dissertation, Univ. of Kentucky, 
Lexington, 

— J. M. Baskin & C. C. Baskin. 2004. The floristic 


m^ of xeric pun prairies in Kentucky, and 


limestone cedar. glades and deep-soil 
) 


5. Sida 21: 1055-107 


Logan, J. M. 1992. The Glade "e the Bull ffalo National River, 
Arkansas. M.S. Thesis, lowa State mes. 


Ludwig. J. C. 1999, The flora of dns and limestone 
barrens in southwestern Virginia. Castanea 64: 200-230. 

Mather, P. M. 1976. Computational Methods of Multivariate 
| and Sons 


Analysis in Physical Geography. J. Wiley 


n 
Maxwell, R. H. 1987. Natural area remnants within the 


Indiana Army Ammunition vi Charlestown, as 
The little bluestem glades. Proc. Indiana Acad. 90: 
1134 ps 

McClain, W. E. & J. E. Ebinger. 2002. A comparison of the 


| glades in Calhoun County, 
Das 

. Multivariate 
MIM Software 


vegetation of pw 
Illinois. 5. x^ N: 
McCune, B. 
analysis of 
Design, Gleneden Beach, Oregon. 
Nelson, P. 1985. The 


Missouri. Missouri 


data, version 4. 


Terrestrial 


Natura 


Department of 
Jefferson Cit 


Natural Communities of 


Resources, 


2 


& D. Ladd. 1983. Preliminary 
identification, distribution, and classification of Missouri 
ucera (editor), Proceedings of 
South- 


report on the 


p. 5976 in C. L. Ki 
dale rican Prairie Conference. 
„Springfield. 


19 7. Ec coregions of the conterminous United 
States. Ánn ids mer. Geogr. 77: 11 
Oosting, H. J L. E. Ande son. 1939, 1 succession on 


eastern. North Carolina. Bot. Gaz. 100: 
— Js 
Peet, R. K., T. R. Wenworth & P. S. White. 1998. A flexible, 
multi-purpose method for p vegetation composi- 
tion and 0 ture. Castanea 63: 202-274, 
. 1981. Vege "ation of Bendstone prs rops on the 


M.S. Thesis, Univ. Te 


Perkins, B. 
Cumbe Plateau. ennessee, 
Knoxville. 

apo rman, E. L. 194 


in iddle 


l A en 


J. Plant Communities of Cedar Glades 


Duke Uni 


— 


Tennessee. Ph.D. Dissertation. 


^ Miller & M. M. Shea. 2004. Soil 


properties and soil nitrogen dynamics of prairie like forest 


.entucky's. Knobs 
l. 


Rhoades, C. C. $ 


openings and surrounding forests in 
Midl. Naturalist 152: 1-1 
Calcareous Glade Communities in the 
Effects of Scale on 
{ North 


region, Amer. 
Rollins, 5. 1997 
Basin. of Tennessee: The 


M.S. 


Central 

Community Classification. Thesis, Univ. o 
Carolina, Chapel Hill. 

Shure, D. 1999. Granite per of the 
United niue Pp. 99-118 in R. 
& J. 3askin (editors), Savannas, 


s. 
Outcrop 1 70 Communities of North America. 


southe aste rn 


Cambridge 

Univ. Press, Cambridge. 

&. L. 1977. Patterns of 
sion on granite outcrop surfaces. Ecology 58: 993-1000. 

Skinner, B. J. 


Ragsdale. primary succes- 
1979. Location and Site Requirements of Three 
Endangered/Phreatened Missouri Glade 
5 of Missouri, Columbia. 
Hammel & E. L. 


pli ant communities and 


Plant Species. 
Univ. 
R. Smith, P. B. 
analysis of 


Thesis. 
Somers, P., L. 
Preliminary 
s in cedar glades in middle 

S. B. Bull. 33: 178-192. 
United States Department of Agriculture, Natural Resources 
0 PLANTS Database, 
National P 


Bridges. 


Te mnnessee 


Conservation Service. 2004. The 
<http:/plants.usda.gov>). 


ant 

e Rouge 

Ver Boot ni M. Ro Raat & D. C. 
Glades of the 59 5 Scenic Riverways, Missouri. 
to the National Park Service. U.S 

Midwest Region, Omaha, 


1993. 
A report 


Glenn-Lewin. 


. De RM of the 
Nebrask 
Winterringer, G. S. & A. G. 1056. Rock-ledge 

vegetation in ewe rn Ilinois. Ecol. Monogr. 26: 105-130. 
. White. 1999, High-elevation outcrops 
ede i p southern 
Pp. 119-134 in R. C. 
Baskin (editors). 
Plant omine: of 


Interior. 
estal. 


uw Mountains. 
alish & 
‘Roe k Outerop 


Anderson, 


Savannas, Barrens, ae 
North America. Cambridge Uni 
Press, Cambridge. 


m / or 2 Omernik, W. H. Martin, G. x Pond, W. 


Andrews, S. M. Call, J. A. > 1 1 & D. D. Taylor. 
1 5 Ecoregions of Kentucky. United ue s Geological 
Survey, Reston, Virginia. 


POLLINATION BIOLOGY IN A Leandro Freitas’ and Marlies Sazima' 
TROPICAL HIGH-ALTITUDE 

GRASSLAND IN BRAZIL: 

INTERACTIONS AT THE 

COMMUNITY LEVEL'? 


ABSTRACT 


Surveys of local assemblages of plants and their pollinators are among the most useful ways lo evaluate specialization in 
pollination and to discuss the patterns of plant- 57 interactions among ecosystems. The high-altitude grasslands from 
southeastern. Brazil constitute diminutive island-like formations surrounded bs montane rainforests. We registered the floral 
traits of 124 species from the Serra da Bocaina grasslands (about 60% of the animal-pollinated species of this flora). and 
determined de pollinators of 106 of them. Asteraceae (40 species) and 1 'eae (10 species) were prominent, while 
most families were represented by few species. The predominant floral traits were: dish or short-tubular sh: ape; nectar as 
a reward; and greenish or violaceous colors. Pollinators were divided into eight func on groups (small bees, syrphids, other 
dipterans, ete.) and small bees, wasps, and large bees were the most important pollinators. Butterflies. beetles, and 
hummingbirds were poorly re presented, ae no bats, hawkmoths. or odor-collecti ing bees were detected. Plants were grouped 
in nine pollination systems, among which nectar-flowers ds ited b bees (28%), by wasps or wasps and flies (2196) 


or by 
several insect groups (19%) were the most representative. 


Vith regard to the degree of specialization, plant species were 


classified according to their number of pollinator groups. pes 3396 of the species were monophilous and 3096 were 


Mar (1.e.. pollinated by one or two functional groups, respectively). The remaining species were either polyphilous 
lo) or lentillas 19%), a highly generalist system in which at least three groups act as indistinct pollinators. The general 


s of the floral traits "ie jap vests interactions at the Bocaina grasslands resemble those of biogeographic- 
connected ecosystems, such as the Venezuelan arbustal, and the Brazilian oe rupestre and cerrado. However, in the 
Bocaina grasslands, the mean number alee types per plant was 2.09, one of the highest values obtained for worldwide 
floras. The origin of the high-altitude peer is linked to episodes of expans sion and retraction due to glacial events. Such 
a situation may have favored species able to quickly occupy new habitats, including those that do not depend on a few highly 
specialized. pollinators. The prevalence of Asteraceae may also be linked to more generalized pollination systems. 
Alternatively, some floral rails, such as spontaneous self- rod and long-lived flowers, may be advantageous for s 


spe C ie S 
with more 17 7 cialized systems in these grasslands with harsh climatic conditions and low rates of ` pollinator visitation. 
Key words: Asteraceae, p "es, 5 floral biology. 1 pollination ecology, specialization, syndrome. 


Wasps. 


This paper is part of the first author's doctoral dissertation deis to the Programa de Pós-Graduação em Bo. 


NU: Universidade Estadual de € Campina as. We would like t than Alves-dos-Santos, A. A. A. Barbosa, P. E. Berry, D. R. 
Campbell, P. Feinsinger, R. A. Figueiredo. J. Semir, and S ee | for em comments on the early versions of this manuscript: 


R. Goldenberg and two anonymous reviewers for their careful revision of the manuscript: A. Francois for the E nglish revision: 
LE p E. Gross, R. Ramos, A. L. Rave ‘tla, and I. San Martin-Gajardo for field and I: aboratory assistance; A. Amaral Jr., 
M. C. E. Amaral, N. M. Bacigalupo, V. Bittrich. M. Bovini, E. Cabral, L. Capellari . B. 

¿hukr, d Condena. . P. Correa, J. H. A. D ailh, € „ L. Esteves, R. L. Esteves, A. D. Faria, M. A. Farinaccio, A. Fillietaz, / 
Flores, M. ppo, A. Guglieri, P. J. pu e R. oe Harley, N. Hind, S. L. Jung-Mendagoli, M. L. Kawasaki, L. S. 


e 
=] 
> 
~ 
7 
> 2 


e e Longhi- agner, M. C. H. Mamede, M. C. M. Marques, A. B. Martins. V. F. Miranda, R. Monteiro, i 
Moraes. J. "e Nala: 1 Oliveira. E P 1 R. Pirani, M. F. Salimena, P. T. Sano. J. Semir. G. J. She phe "i 
Simao-Bianchini, A. O. Simoes, R. B. Singer, J. P. Souza, V. C. Souza, J. R. Stehmann. J. Tamashiro. A. M. G. Tozzi. A. . 
Vieira, F. A. Vitta, M. G. L. : ande rley, K. Yamamoto and C. S. Zickel for plant identifications; I. Alves dos Santos, S. 0 p. 
\marante, K. S. Brown, e M. F. de Camargo. S. A. € ar B. W. Coelho, A. V. L. Freitas, E. Giannotti, A. X. Linhares. U. R. 
Martins, G. A. R. Melo, A PE S. R. M. Pedro, I. Sazima, C. Schlindwein. O. T. Silveira, 3 D. Urban for animal 
identifications: J. B. n n rand R. C. Forzza for the nomenclatural verification of plant taxa; and IBAMA for allowing us to 


work in the PARNA Serra da Bocaina (through M. A. B. Rondon). CNPq, CAPES, and FAEP-U a AMP provided financial 


uc ^n 


ie editors of the Annals thank Sophia Balcomb for her editorial contribution to this article. 

‘Jardim Botánico do Rio de Janeiro. Rua Pacheco Leão 915, 22460-030. Rio de Janeiro-RJ. Brasil: le andro@ jbrj.gov.br 
partamento de Botánica, Caixa Postal 6109, Universidade Estadual de Campinas, 13083-970. Campinas- SP, Brasil: 
msazima@unicamp.br. 


t 


ANN. Missouri Bor. GARD. 93: 465-516. PUBLISHED ON 24 OcronkRn 2006. 


Annals of the 
Missouri Botanical Garden 


Plant reproductive processes are believed to help 
determine the composition and structure of commu- 
nities (Heithaus. 1974, 1979; Bawa, 1990; Oliveira & 
Gibbs, 2000). Among such processes, the plant- 
pollinator interactions form a dynamic, yet somewhat 
cohesive, ecological subunit of a community, which 
can be studied in terms of species diversity and 
distribution, resource utilization, and niche packing. 
for instance (Moldenke & 1979). In the 


Neotropies, pollination biology at the community level 


Lincoln. 


has been studied in forest areas (Bawa et al.. 1985: 
Beach, 1994), vegetation 
(Silberbauer-Gottsberger & Gotlsberger. ha- 
mirez, 1989, 2004: Barbosa. 1997; Oliveira & Gibbs. 
Brito, 1992) and sub- 
1982). In 


Brazilian Atlantic Forest Domain, there are several 


Kress & savanna-like 


2000: see also Ramírez & 


Andean vegetation (Arroyo et a the 
studies on pollination biology in forest areas on both 
individual species and flower assemblages (Sazima el 
al., 1995, 1996, 1999; 2000 


references therein). However, for the high-altitude 


Buzato et al., and 


grasslands or campos de altitude—a subtype of the 
Atlantic Forest Domain—information on plant-polli- 
nator interactions is limited to a few case studies 
(Martinelli, 1997: Freitas & Sazima, 2001. 20024. 
Freitas et al.. 2000). 


The high-altitude grasslands in southeastern Brazil 


) 


comprise an archipelago of mountaintop formations 


(ca. 1500 m elevation or higher) along the main 
mountain ranges, with a North/South orientation. 


Their origin is connected to glaciation episodes since 
al least the Late Pleistocene, when these grassland 
areas are supposed to have experienced periods of 
1998). The 


present high-altitude grasslands have strong bio- 


expansion and retraction (Ledru et al., 


geographic connections with other high-altitude 
ecosystems in South America, such as the Andean 
and Central- American (sub-) alpine habitats. as well 
as with the Central-Eastern Brazilian rocky and 
grassy habitats (campo rupestre) (Giulietti & Pirani, 
1988; Safford. 1999a and references therein). Thes 


are also connected lo the central Brazilian savannas 


(cerrado) (Silveira & Cure, 1993; see also Modenesi, 
1988 for cerrado components in palinofloras from the 


grasslands). These grasslands present—for such re- 


stricted areas of habitat—extraordinarily rich floras 
with many endemies. About one third of the ca. 400 


km”). for 


instance, appear to be endemic to the high-altitude 


species of the Itatiaia plateau (< 50 
grasslands (Martinelli, 1989). However, many species 
on the high-altitude grasslands have small populations 
with few individuals. Owing to such features as high 
species richness, high endemism, small populations, 
"island-like 


and harsh climatic conditions for a tropical place, the 


occurrence, biogeographic connections, 


Brazilian high-altitude grasslands are interesting 
communities in which to study plant-pollinator 
interactions. 

In pollination biology, the degree of specialization 
in plant-pollinator relationships has generated a de- 
the the 


syndrome concept (see Faegri & van der Pijl, 1979) 


bate in which traditional view based on 
has been counterbalanced by skepticism about the 
specialized nature of pollination systems (Waser el al. 
1996; Herrera. 1996; see 
2000 and Fenster et al.. 


views). Surveys of entire local assemblages of plants 


also Johnson & Steiner, 


2004 for reviews of both 


and their pollinators provide the best evidence to 
assess the degree of specialization in pollination 
Waser 


pollinator interactions may be examined on a broad 


et al. 1996). because particular plant 


ecological scale, reducing the sampling bias inherent 
to the comparison of data originating from small scale 
studies, such as those concerning particular plant 
taxonomic groups. This paper describes the commu- 
nity-level interactions between plants and pollinators 
of high-altitude grasslands at the Serra da Bocaina, 
southeastern Brazil. We studied the floral biology of 
plant species, recording their sexual systems, flower 
shapes. dimensions, resources, colors, and pollinators. 
Combining the data on floral features and functional 
groups of pollinators allowed us to characterize nine 
pollination systems in these grasslands and to group 
plant species in four categories that express the 
degree of specialization/generalization of plant-polli- 
nator interactions in this community. In addition, we 
monitored. the flowering phenology so as to assess 
flower resource availability, throughout the year. for 
the main pollinator groups. We provide a discussion 
about general patterns of plant-pollinator interactions 
at the Serra da Bocaina grasslands in relation to other 
ecosystems with strong biogeographic connections. as 


well as in relation to the current debate about the 


prevalence of generalization versus specialization in 


l 


e plant-pollinator interactions. 
AREA AND VEGETATION STUDIED 


The Parque Nacional da Serra da Bocaina (PNSB) 
ca. 100.000 hectares 
States of Rio de 
southeastern Brazil. Since it extends from sea level 


s located between the 


Janeiro and Sao Paulo. in 


to the highest peaks in the Serra da Bocaina (ca. 


2 


00 m). the PNSB shelters many vegetation types of 
the Brazilian Atlantic Forest Domain. The lower areas 
are covered mainly by lowland and submontane 
rainforest, followed by montane and high-altitude 
high-altitude grasslands 


1700-1800 m. 


nowadays, due to human disturbance of the forested 


forest. Originally, these 


probably occurred only above 


Volume 93, Number 3 
2006 


Freitas & Sazima 467 


Pollination in a High-Altitude Grassland 


1450 m. 


These grasslands are composed of a matrix of Poaceae 


areas, they can be found from as low as ca. 


— 


and Cyperaceae species mixed with shrubs and herbs 


from many families, mainly Asteraceae and Melasto- 


mataceae. Scattered small trees, mainly species 
belonging to the genus Eremanthus Less. (Astera- 


ceae), can also be found in more protected fields. 

In southeastern. Brazil, the montane areas encom- 
pass two climate regions (Kóppen, 1948). Grasslands 
from the highest summits (above ca. 2000 m) are Cwb. 
with = 12°C mean annual temperature, cool summers, 
moderately cold winters with frequent frost, and rare 
snow. Montane areas at lower elevations. can be 
classified as Cfb (Cwb and Cfb as in Köppen. 1948). 
12°C-20°C average annual 
and mild, wet sum- 
mers (Segadas-Vianna & Dau, 1965; Nimer, 1977: 
Safford, 1999a, b). Most grassland areas at Serra da 
Bocaina are Cfb. 
2100 mm. Rains are concentrated in summer, 


.e., mesothermic, with 


temperature, moderate winters, 


Annual precipitation is up to 


mainly 


from December to March. The dry season (Apr. & 
a ) is driest from June to August, with precipitation 

< 50 mm per month. The effects of seasonal drought 
on the vegetation are partially offset by thick banks of 
orographie fog that shroud most Bocaina grasslands. 
for much of the year. 


mainly in the early morning, 


a. 15€ at 150 


^ 


Annual mean temperature is 
1600 m, and minimum temperatures around 0 C are 
common during winter (L. Freitas, pers. obs.). During 
the study period (1998-1999), frosts — on more 
than 30 days per year (L. Freitas, pers. obs.). General 
information about the vegetation, climate, and topog- 
raphy of southeastern Brazil is available in Segadas- 
Vianna & Dau (1965). Eiten (1970, 1992), 
(1977), Nimer (1977), Moreira & Camelier (1977). and 
Safford (19992, b). 


Despite its long history in Bocaina highlands. 


Alonso 


human disturbance was low until the beginning of the 


20" century. The primary source of revenue of the 


inhabitants of this region comes from livestock. 
Regular fires and grazing are supposed 1o have 
negative impacts on the physical environment. the 
vegetation structure and dynamics, and the welfare of 
some of the small and endemic plant populations of 
these grasslands (Safford, 1999a, but see Safford. 
2001). Natural fires are apparently rare on these 


erasslands and are restricted to the wet season, when 
lightning occurrence is frequent. Anthropogenic fires. 
however, are frequently set during the dry season. 
because such an action favors the sprouting of grasses 
after the rains begin. Some plants may have been 
regular fires, in particular, 


eliminated. by these 


species restricted to mesic habitats. However, many 


of the typical high-altitude taxa, such as the 
geophytes, have a strong capacity to sprout, appar- 


ently as a pre- nee ae frost, that also seems 
to be efficient against 

This study was 1 out in three grassland sites 
that experienced different fire regimes over the years. 
Since there is no official register of fire occurrences, 
information for the last 45-50 years was obtained from 
both the PNSB staff and the inhabitants of the region. 
irst area (22°43'57"S, 44 37'00"W) is adjacent 


to a private farm (Fazenda Mariana) and used to burn 


The 


every two to three years. The last fire in that area had 
five months before the beginning of this 
study (Dec. 1997). The second area (22744'50"S, 
44 36'57"W) is close to Santo Izidro waterfall 
The last 


occurred ca. 


and 
used to burn at intervals of five to 10 years. 
fire in this area had occurred five years before the first 
year of this study, and the area burned again in 1998 
Since il presented many treelets and large shrubs, the 
third area (22 44 1278, 44 36'55"W) showed marked 
structural differences. There was no record of fire here 
when the project started, but the area burned in 
September 1999 

About 260 plant species belonging to 46 families 
were recorded during monthly collections in the three 
study areas and through non-systematic collections in 
other grassland areas located between 1450 m and 
2100 m. 


ed by a few species, some families, such as oe ‘eae 


Although most families were only represent- 


(ca. 70 spp.), Poaceae (ca. 35 spp.). Cyperaceae (ca. 
15 spp.). Melastomataceae (18 spp.). Orchidaceae, 
Iridaceae, Rubiaceae, and Solanaceae, were common. 
Most species were herbs or (sub)shrubs, with small, 
often lignified and hairy leaves. Many species had 
leaves with a cupressoid or rosette arrangement, and 
several perennial plants showed well-developed un- 
derground organs (e.g., Tibouchina minor. Microlicia 
isophylla, Esterhazya macrodonta, Escallonia farina- 
cea). Such vegetative features are typical of plants 
(Camerik & Werger, 
19994). 


high-altitude tropics 


1987; Safford, 


living in 
1981; Smith & Young, 
MATERIAL AND. METHODS 

Between December 1997 and February 2000, we 
made 25 field trips to Serra da Bocaina, for a total of 
211 
community 
| km long in each of the three study sites. We studied 


days of fieldwork. Pollination biology at the 


level was studied along a transect ca. 
all plant species located within 5 m of the transects, 
except (Table 1). The 


following floral attributes were noted in the field: 


Cyperaceae and Poaceae 


shape, symmetry, size, color, and rewards (Table 2). 
Following Faegri and van der Pijl (1979) and Endress 
(1994), the flowers were classified into eight different 
types: inconspicuous, dish (bowl), brush, bell (funnel), 
flag. and revolver. Tube types include 


gullet, tube, 


468 Annals of the 
Missouri Botanical Garden 


able J. Plant species, their habit, flowering time, and flowering peak at the Serra da Bocaina grasslands. All collections 


a 
y P authors are deposited at the UEC, and the collection number (L. Freitas) follows the taxon name in parentheses. 


Phenological observations occurred from January to December 1999, Species that were found only in the area that burnt in 
f 


September 1999 are indicated by the symbol (f). 


Family/Species Habit Flowering time Flowering peak 
Amaryllidaceae 
Hippeastrum glaucescens Mart. (713) herb Oct.—Nov. Nov. 
Apiaceae 
Eryngium canaliculatum Cham. & Schltdl. (792) herb Nov.-Dec. Dec. 
Eryngium horridum Malme (140) herb Jan.-Mar. Jan. 
Apocynaceac 
Gonioanthela hilariana (E. Fourn.) Malme (329) vine Mar.—July (f) Mar.—May 
Mandevilla erecta (Vell.) Woodson (86, 509) sub-shrub Jan.—Feb. Jan. 
Oxypetalum appendiculatum Mart. (123) vine Jan.—Aug. (f) Apr.—June 
Oxypetalum sublanatum Malme (13, 491) vine Jan., Mar.—May, Jan., Apr., Nov.—Dec. 
)ct.- Dec 
Tassadia subulata (Vell.) Fontella & E. A. Schwarz (686) vine May-Aug. June 


Aquifoliaceae 
Ilex amara (Vell.) Loes. (656, 797) sub-shrub Nov.—Dec. Nov.—Dec. 


Asterace: 


Achyroc 1 satureioides (Lam.) DC. (276, 330) shrub Jan.—Aug., Dec. Mar.—June 
Baccharis aphylla (Vell.) DC. (707) sub-shrub Sep.-Nov. Sep.-Oct. 
Baccharis curitybensis Heering & Dusen (498) sub-shrub Jan.-Mar., Dec. Jan. 

Baccharis dracunculifolia DC. (372) shrub Mar.—Oct. Apr.-May, Aug. 


Un 
z 
= 
= 
= 
| 
=> 
E 
oo 


Baccharis intermixta Gardner (418 July 


Baccharis leptocephala DC. (392, 443, 688) herb Apr.—Sep. May—Aug. 


Baccharis pentzitfolia Sch. Bip. ex Baker (589, 691, 728) shrub Jan.—Dec. Feb., May-Aug.. Oct.— 
Nov 
Baccharis platypoda DC. (334, 863) shrub Mar.—June May 
Baccharis tarchonanthoides Baker (733) sub-shrub Oct.— Oct.—Nov. 
Baccharis L. sp. indet. | (407) shrub PL ee July 
Baccharis sp. indet. 2 (441, 689) shrub May-July June-July 
Baccharis sp. indet. 3 (108) shrub Jan.—Mar. Jan. 
Barrosoa betonictiformis (DC-) R. M. King & H. Rob. herb Jan.—May Feb.—Mar., May 
(512, 591) 
Campuloclinium „ (Mart. ex Baker) R. M. herb Jan.—May Veb.-Apr. 
King & H. Rob. (101, 596) 
Chaptalia integerrima nt Il.) Burkart (744) herb Jan. Feb., Oct.—  Oct.—Nov. 
Jec. 
Chaptalia runcinata var. graminifolia (Dusen) Burkart herb May-Nov. July-Aug. 
(406, 697) 
Chromolaena ct. decumbens Gardner (100) herb Jan.—July Jan.—Mar. 
Chromolaena xylorrhiza (Sch. Bip. ex Baker) R. M. herb Jan.—Feb. Jan. 
King & H. Rob. (835) 
Eremanthus erythropappus (DC.) MacLeish (403) treelet Jul.-Aug. (f) Aug. 
Erigeron maximus (D. Don) Otto ex DC. (64) shrub Jan.—Mar., Oct.-.— Jan.—Feb., Dec. 
Dec. 
Eupatorium l. sp. indet. 1 (60, 61) shrub Jan., Nov.-Dec. Nov. 
Eupatorium sp. indet. 2 (99) herb Jan.—July, Dec. Jan.—Feb. 
Eupatorium sp. indet. 3 (326) herb Mar.—June Apr.-May 
Eupatorium sp. indet. 4 (374) herb Apr. Apr. 
Gochnatia paniculata (Less.) Cabrera (438, 705) sub-shrub Aug.—Nov. Sep.—Oct. 
Grazielia „ (DC-) R. M. King & H. Rob. shrub Jan.—June Jan.—Feb. 
(226, 59 
33 e Baker (481) herb Jan.—Feb.. Oct.—  Jan.—Feb., Nov.—Dec. 
Dec. 
Lucilia lycopodioides (Less.) S. E. Freire (398, 838) herb June-Sep. July-Aug. 
Mikania lundiana DC. (393) vine May-Jul. May 


Volume 93, Number 3 
2006 


Freitas & Sazima 
Pollination in a High-Altitude Grassland 


Table 1. | Continued. 

Family/Species Habit Flowering time Flowering peak 
Mikania nummularia DC. (440, 690) shrub May-Aug. June-Aug. 
Mikania sessilifolia DC. (318, 417, 871) shrub Mar.-May Apr 
Senecio oleosus Vell. (391, 409) shrub Jan., Mar.-Sep., Mar., July-Aug. 

Nov 

Stevia myriadenia Sch. Bip. ex Baker (88, 291) shrub Jan -May Feb.—Mar. 
Symphyopappus compressus (Gardner) B. L. Rob. (57) shrub Jan.—Feb. (f) an. 
Vernonia herbacea (Vell.) Rusby (741) herb Oct.—Dec. Nov.-Dec. 
Vernonia megapotamica Spreng. (36, 536) herb Jan.—Apr Jan.—Feb. 
Vernonia cf. rosea Mart. ex DC. (275, 592, 844) sub-shrub Jan.—Feb Feb. 


Vernonia tomentella Mart. ex DC. (67, 588) 


Vernonia tragiaefolia DC. (626) 

Vernonia westiniana Less. (92) 
Bromeliaceae 

Dyckia tuberosa (Vell.) Beer (495) 
Campanulaceae 

Lobelia camporum Pohl (41) 


Wahlenbergia brasiliensis Cham. (232, 280 
Clethraceae 
Clethra scabra Pers. var. scabra (332) 


Convolvulaceae 
Convolvulus crenatifolius Ruiz & Pav. (37, 89) 
Ipomoea procumbens Mart. (76) 
Jacquemontia grandiflora Meisn. (118) 


Cunoniaceae 

Weinmannia organensis Gardner (54, 125) 
Droseraceae 

Drosera montana A. St.-Hil. (127) 


Ericaceae 
Agarista hispidula (DC.) Hook. p ex . (23, 431) 
Gaylussacia chamissonis Meisn. 
Gaylussacia jordanensis Sleumer 2 432) 


Eriocaulaceae 
Paepalanthus paulensis Ruhland (706) 
Paepalanthus polyanthus (Bong.) Kunth (18) 


Erythroxylaceae 
Erythroxylum microphyllum A. St.-Hil. (373, 496) 


Escalloniaceae 
Escallonia farinacea A. St.-Hil. (507) 
Euphorbiaceae 
Croton dichrous Müll. Arg. (17, 619) 
Gentianaceae 

Calolisianthus pedunculatus (Cham. & Schltdl.) Gilg 

(430, 520) 

Calolisianthus pendulus (Mart.) Gilg (80, 264) 

Deianira nervosa Cham. & Schltdl. (348) 

Helia oblongifolia Mart. (1, 95) 

Zygostigma australe (Cham. & Schltdl.) Griseb. (622) 
Gesneriaceae 

Sinningia allagophylla (Mart.) Wiehler (9) 


sub-shrub 


sub-shrub 
shrub 


herb 
herb 
herb 
treelet 
vine 
vine 
vine 
treclet 


herb 


sub-shrub 


sub-shrub 


herb 


herb 


shrub 


Jan.-May, Oct.— 
Dec 


Jan.-Feb., Apr. 
Jan.-June 


Sep.-Nov. 
Jan.-June, Nov.— 
Dec. 
Jan.-Dec. 
Feb.—Aug. (f) 
Jan.—Feb., Dec. 
Jan.—Mar. 
Jan.-Aug., Nov.— 
Dec. 
Jan.—Aug. (f) 


Jan.—Feb., Nov.— 


ec 
June-Nov 
Apr.-Dec 


Jan.-May, Aug.— 
Dec 


Sep.-Dec. 
Jan.—June, Nov.— 


Jan.-May, Nov.— 
Dec. 

Nov.—Dec. 
Jan.-Dec. 
Jan.-Apr., Dec. 
Jan.-May, July 
Feb.—Apr., June 
Jan.-Mar 
Feb.—Apr. 


Jan.—Mar., Nov.— 
Dec 


Jan.—Feb., Nov.—Dec. 


Feb. 
Feb. 


Oct.-Nov. 


Jan.—Feb., June, 


Feb., June-Sep. 


Mar.-May 


Jan.. Dec. 
Feb. 
Jan.—Feb., 


Dec. 


Feb.-Mar., May 


Jan., Dec. 


Aug.—Sep. 
Oct.-Dec. 
Sep.- Oct. 


Sep.-Nov 
Jan.-Apr., Dec. 


Jan.—Feb., Dec. 


Nov. 
Mar.—Jul.. 
Jan.-Feb. 
Feb. 
Feb.—Mar. 
Jan.—Feb. 


l'eb.-Mar. 


Jan., Dec. 


Dec. 


May-July, 


Nov.—Dec. 


470 


Annals of the 


Missouri Botanical Garden 


Table . Continued. 


Family/Species Habit Flowering time Flowering peak 
Hypericacea 
Hypericum Pm A. St-Hil. (218. 508) shrub Jan.—May, Dec. Jan 
Hypericum ternum Choisy (71, 494) herb Jan.—Mar., May- Jan -Heb., Nov.-Dec. 
Aug.. Oct.-Dec 
Iridaceae 
Alophia geniculata Klatt (376) herb Mar.-Apr., Nov.— Mar., Nov. 
Jec 
Alophia Herb. sp. indet. herb Jan.—Feb Jan 
Calydorea campestris 1 10 Bake r (236, 368) herb Jan.—Mar.. Oct.— jue Nov Dee. 
Dec. 
Sisyrinchium micranthum Cav. (702) herb Jan.—Feb.. Nov.— — Nov.- Dec. 
Dec. 
Sisyrinchium vaginatum Spreng. (346, 434, 515) herb Jan., Mar.-Dec..— May-Oct. 
Lamiaceae 
Hyptis lippioides Pohl ex Benth. (399, 423) herb July-Oct. July 
Typtis plectranthoides Benth. (5, 606) herb Jan.—May, Oct— Jan., Dec. 
Dec. 
Hyptis umbrosa Salzm. ex Benth. (310) shrub Jan.-Apr. Feb.-Mar. 
Peltodon radicans Pohl (313, 625) herb Mar.-May Mar.—Apr. 


Leguminosae 


Chamaecrista Moench sp. indet. | (85) 


Crotalaria breviflora DC. (308, 818) 


Lupinus velutinus Benth. (427 


Lythraceae 
Cuphea glutinosa Cham. & Schltdl. (20) 


Malpighiaceae 
Byrsonima variabilis A. Juss. (112, 490) 
Malvaceae 
Pavonia kleinii Krapov. & Cristobal (105) 


Sida L. sp. indet. 1 (24, 456 


Leandra erostrata Cogn. (113) 
Leandra Raddi sp. indet. 1 (19) 
Leandra sp. indet. 2 (401) 


Leandra sp. indet. 3 (402) 
Microlicia isophylla DC. (103) 


Tibouchina frigidula (DC.) Cogn. (25. 


Tibouchina martialis Gogn. (309) 


19) 


Tibouchina minor Cogn. (32) 

Trembleya parviflora (D. Don) Cogn. (341, 420) 

Trembleya phlogiformis DC. (44, 74) 
Ochnaceae 


Engl. (425) 


Ouratea semiserrata (Mart. & Nees 
Orchidaceae 

Epidendrum secundum Jacq. (1 

Habenaria parviflora Lindl. o a 
(282. 305) 


Oncidium barbaceniae lind 


Oncidium blanchetii Re P f. (611) 


Oncidium Sw. sp. indet. 1 (501) 


sub-shrub 


sub-shrub 


sub-shrub 


sub-shrub 


sub-shrub 


vine 


herb 


herb 


herb 


shrub 
shrub 


sub-shrub 


shrub 
shrub 
herb 

shrub 


sub-shrub 


treelel 


Feb.—Mar.. Oct.— 
Dec. 

Jan.—Apr., Dec. 

July-Oct.. Dec. 


Jan.- Dec. 


Jan.—Feb.., May, 
Oct.-Dec. 


Jan.—Feb., Dec. 
Jan.—Mar., Oct.— 


| ec. 


Jan.—Feb.. Oct.— 
Dec. 

Jan.—Feb.., Nov.— 
Dec. 

July-Oct. 

July-Dec. 

Jan. ia July- 


Jan.—Mar. 
June—July 


Jan.—Aug. (f) 
Jan.—Mar.. Dec. 
Jan.-Mar., May- 


Dec. 


Jan.—Dec. 


Jan., Dec. 


Nov.—Dec. 


Jan.—Mar. 
Aug.-Sep. 


Jan., Apr.-June, Aug., 
Nov. Jec 

Nov.—Dec. 

Feb. 

Jan., Nov.—Dec. 

Jan.. Nov.—Dee 

Jan.. Nov.—Dec. 

Aug.-Sep. 

Oct.—Nov. 

Jan.—Feb. 

Jan.-Mar., Dec. 

June 

Jan.—Feb. 

June-July 

Jan.—Mar. 

June 

Jan., Apr.—May 

Jan.—Feb. 

Jan.. May. July. Sep.- 
Dec. 

July-Nov. 

Jan. 


Volume 93, Number 3 
2006 


Freitas & Sazim 


a 471 
Pollination in a High-Altitude Grassland 


Table J. | Continued. 
Family/Species Habit Flowering time Flowering peak 
Orobanchaceae 
Esterhazya macrodonta Cham. & Schltdl. (244) shrub Jan., June-Oct. July-Aug. 
Polygalaceae 
Polygala brasiliensis L. (309, 366, 683) herb Feb.-Sep., Nov.— Mar., June-Aug., Dec. 
Dec. 
Polygala cneorum A. St.-Hil. & Moq. (390) herb Apr.—Sep., Nov. May—Aug. 
Rubiaceae 
Borreria capitata (Ruiz & Pav.) DC. (322) herb Feb.—July Mar.—Apr. 
Borreria tenella Cham. & Schltdl. (192, 667) herb Mar.—July May-June 
Declieuxia cordigera var. angustifolia Müll. Arg. (6. herb Jan.-Dec. Jan.-Feb., June, 
727) aces: 
Galianthe angustifolia (Cham. & Schltdl.) I Cabral sub-shrub a e July, — Jan.-Feb., Dec. 
(50) De 
Galianthe brasiliensis Spreng. (618) shrub jun. a Oc t= Jan.—July, Nov. 
Dec. 
Galium hypocarpium (L.) Endl. ex Griseb. (371, 802) sub-shrub Jan.—July, Oct.— Jan., Apr.—May, Nov.— 
Dec. Dec. 
Solanaceae 
Solanum aculeatissimum Jacq. (483) herb Jan.—Feb., May, Jan.—Feb., Oct.-Dec. 
Det.—Dec. 
Solanum americanum Mill. (488, 675) herb Jan., Nov.-Dec. Nov. 
Solanum pseudocapsicum L. (768) sub-shrub Nov—  Jan.—Feb., Nov. 


Solanum swartzianum Roem. & Schult. (538, 659) 


Solanum viarum Dunal (559, 742) 


Verbenaceae 

Verbena hirta Spreng. (31, 243) 
Violaceae 

Viola cerasifolia A. St.-Hil. (124) 
Xyridaceae 


Xyris asperula Mart. (846) 
Xyris tortulla Mart. (2, 350) 


Jan.—Feb., 
Dec 


, Apr.—June, 


shrub Jan. Apr.—May 
Aug. 

herb Jan.—Mar., Aug.— Feb., Sep., Nov.-Dec. 
Dec 

sub-shrub Jan.-Dec. Jan., Sep.-Dec. 


herb Jan.—Feb., Dec. Jan., Dec. 
herb Feb.—Mar. Feb. 
herb Jan.—May, Dec. Jan.—Mar. 


both salverform and spurred flowers. Asteraceae 
florets were included in the tube type, and their head 
as a whole was classified as either dish or brush type. 
Floral measurements with calipers were performed on 
at least five fresh flowers, each from different plants. 
The main flower color was determined using a color 
guidebook (Kornerup & Wanscher, 1963). For our 
analysis at the community level, the flowers were 
grouped into six color sets: i. violet —showy combina- 
tions of blue and red, such as purple and magenta; 11. 


light colors based on blue or red, such as rose, 


pink 


lavender, and lilae; iii. red and red; iv. 


orange 


— 


yellow—bright and deep yellow: v. greenish—colors 


with sallow tonalities such as greenish, yellowish, ot 
brownish white; and vi. white—bright, apparently 
pure, white. For many species, pollen fertility was 


estimated by cytoplasmic stainability, using the acetic 


carmine test (Radford et al., 1974), and stigmatic 
receptivity was verified with the HzO catalase 


activity test (Zeisler, 1938). In some cases, flowers 


were tagged and bagged at bud stage, and nectar was 


— 


extracted on the following day with a graduated 
microliter syringe (Hamilton, U.S.A.). Nectar volume 
was registered immediately, and nectar sugar concen- 
tration was measured with a hand refractometer 
(Atago, Japan). 

Sexual systems were determined by the presence of 
both functional anthers and stigmas. Flowers were 
categorized as either hermaphroditic (monoclinous) or 
unisexual (diclinous). Plants with unisexual flowers 
further classified as 


were monoecious, dioecious, 


andromonoecious, or gymnomonoecious. Other in- 
formation about sexual systems—such as the presence 
of dichogamy and heteromorphy—was noted for some 
of the 


verified for some species by the report of fruit set in 


species. Spontaneous self-pollination was 


unmanipulated bagged flowers (Kearns & Inouye, 


993). Some pistils were fixed in formaldehyde- 


Table 2. 


Floral features of plant species and their pollination system at the Serra da Bocaina grasslands. In Asteraceae the shape of the blossom (head) and its width are given in parentheses 
in addition to the tubular florets measurements. Flower size for tube, gullet, b 


ell, and revolver type is effective corolla length vs. opening width: for dish, brush, and inconspicuous type it is the 
flag type it is keel (k) vs. flag (f) ind. due symmetry and size for monoecious (U-M) and dioecious (U-D) they 


flower width (diameter: d): and for ' plants correspond to female flower first and 


male flower second, and for gymnomonoecious (GymnM) plants they correspond to hermaphrodite flower first and female flower second. Gynoecium/Androecium = pistil height from the corolla 


base X stamen height from the corolla base (or spur base in spurred flowers). Flower measurements are the mode 


Sexual Floral and Flower symmetry, Gynoecium/Androecium Color 
Family/Species system reproductive traits Floral shape size (mm) (mm) Flower color! group' resource Pollination system 
Amaryllidaceae 
„ glaucescens* H tube Zy, 14 (82) X 7 114 x 96 dark red RE N hummingbird 
Apiac 
Deme ari canaliculatum H protandry inconsp. Ac, dl 1x1 vellowish green GR N.P | uncertain 
Eryngium horridum H protandry inconsp. Ac. dl 2x4 vellowish green GR N. P several insect groups 
Apo ceae 
1 hilariana H dish Ac, dl 1x] pale vellow GR N wasp/fly 
Wandevilla erecta H herkogamy tube Ae, 13 X3 5 * 8 pale vellow GR N undet. 
Oxypetalum appendiculatum H revolver Ac, 8 * 2 * 2 greenish lilac GR N wasp/fly 
Oxypetalum sublanatum I revolver Ac. 8 X 5 5 * 5 pale green Gh N bee nectar-flower 
Tassadia subulata H lish Ac. dl 1x1 dark ruby GR N syrphid nectar-flower 
Aquifoliaceae 
Ilex amara U-M temporal bell Ac, 4 X 2 2X vellowish white GR N bee nectar-flower 
dioecism Ac, 4 * 2 x2 N, P? 
Asteraceae 
Achyrocline satureioides GymnM protandry. 2a tube Ac, 4 X <1 (14) 5. light vellow GR N wasp/fly 
(dish) Ac, 6 X <] 10 x N 
Baccharis aphylla U-D tube Ac, 5 X <1 (2) 6 X greenish white GR N wasp/fly 
(dish) Ac, 4 X 1 (5) x6 N 
Baccharis curitybensis U-D tube Ac, 5 X <l (7) 5X pastel yellow GR N wasp/fly 
(dish) Ac, 5 X «I (6) x 6 N 
Baccharis dracunculifolia U-D tube Ac, 2 X <l (1) 3 X yellowish green GR N wasp/fly 
(dish) Ac, 2 X «] (4) x3 N, P 
Baccharis intermixta U-D tube Ac, 2 X <l (2) Lx pale yellow GR N wasp/fly 
(dish) Ac, 3 X <1 (4) x 4 N 
Baccharis leptocephala U-D tube Ac, 2 X <l (2) 3 * pale yellow GR N wasp/fly 
(dish) Ac. 3 X «1 (3) X4 N 
Baccharis pentziifolta U-D tube Ac, 3 X «1 (2) 5 * vellowish green GR N wasp/fly 
(dish) Ac, 3 X <1 (4) x5 N.P 
Baccharis platypoda U-D tube Ac, 3 X <1 (5) 7 * erayish yellow GR N wasp/fly 


dp N [eoiuejog unossi|A 


C4 


əy} Jo sjeuuy 


Table 2. Continued. 
Sexual Floral and Flower symmetry, Gy ium/And ium Color Floral 
Family/Species system reproductive traits Floral shape size (mm) (mm) Flower color' group! resource Pollination system 
(dish) Ac, 4 X 1 (6) x6 N,P 
Baccharis tarchonanthoides U-D tube Ac, 3 X 1 (7) 4 X whitish yellow GR N several insect groups 
(dish) Ac, 2 X 1 (5) x 3 N 
Baccharis sp. indet. | U-D tube Ac, 3 X «1 (2) 7X pale vellow GR N wasp/fly 
(dish) Ac. 4 X <1 (4) x7 N 
Baccharis sp. indet. 2 U-D tube Ac, 4 X «1 (2) 5 * greenish yellow GR N wasp/fly 
(dish) Ac. 4 X 1 (6) x5 N 
Baccharis sp. indet. 3 U-D tube Ac, 2 X <1 (1) | * pale vellow GR N wasp/fly 
(dish) Ac, 2 X «1 (2) x3 N 
Barrosoa betonicuformis H protandry, 2a tube Ac, 3 X <1 (10) 6X5 purple VI N several insect groups 
(brush) 
Campuloclinium H protandry, 2a tube Ac, 4 X «1 (16) 9x8 reddish purple VI N several insect groups 
megacephalum (brush) 
Chaptalia integerrima GymnM protandry, 2a tube Ac, 10 X 1 (22) 14 * 3 white WH N, P several insect groups 
Ligulate (brush) Zy,2 X «] 3 * N 
Chaptalia runcinata var. Gymn! protandry, 2a tube Ae, 3 X «1 (12) 4X4 vellow YE N. P several insect groups 
graminifolia Ligulate (dish) Zy,2 X <] 3 * N 
Chromolaena cf. decumbens H protandry, 2a tube Ac, 5 X <I (9) 7X6 reddish violet VI N bee nectar-flower 
(dish) 
Chromolaena xylorrhiza H protandry, 2a tube Ac, 4 X 1 (4) 6X4 purple VI N undet. 
(dish) 
Eremanthus erythropappus HI protandry, 2a ibe Ac, 5 X <1 (14) 10 x 8 purple VI N, P several insect groups 
(brush) 
Erigeron maximus GymnM protandry, 2a tube Ac, 4 X 1 (35) 8X 7 vivid yellow YE N. P several insect groups 
Ligulate (dish) Zy, 3 X <] TX N 
Eupatorium sp. indet. 1 H protandry, 2a tube Ac, 7 X <1 (10) 15 x 13 white WH N, P bee nectar-flower 
(brush) 
Eupatorium sp. indet. 2 H protandry, 2a tube Ac, 6 X «1 (8) 8 * 7 violet VI N several insect groups 
(dish) 
Eupatorium sp. indet. 3 H protandry, 2a tube Ac, 9 X 1 (12) 19 x 18 purple VI N bee nectar-flower 
(brush) 


900c 


€ JequinN ‘£6 euinjoA 


eullZeS Y Sepal 


pue|ssec) p- e ul uoneurog 


£Zv 


Table 2. Continued. 


Sexual Floral and Flower symmetry, Gynoecium/Androecium Color Floral 
Family/Species system reproductive traits Floral shape size (mm) (mm) Flower color! group! resource Pollination system 
Eupatorium sp. indet. 4 H protandry, 2a tube Ac, 6 X 1 (8) 9x8 white WH N, P_ several insect groups 
(brush) 
Gochnatia paniculata H protandry. 2a tube Ac, 3 X <1 (10) 8X1 greenish vellow GR N, P — wasp/fly 
(brush) 
Grazielia gaudichaudeana H protandry, 2a tube Ac, 3 X «I (5) 7X5 white WH N. P several insect groups 
(brush) 
Hypochaeris gardneri GymnM protandry, 2a tube Ac. 7 X «] (16) 10 x 9 vivid vellow YE N bee nectar-flower 
Ligulate (brush) Ly, 5 X <l 11 x 
Lucilia lycopodioides GymnM protandry. 2a tube Ac. 6 X 1(3) 7X6 vellowish white GR N wasp/fly 
Ligulate (dish) Zy,6 X] 7 * 
Mikania lundiana protandry, 2a tube Ac. 8 X <l (4) 11 x9 vellowish white GR N.P bee nectar-flower 
(brush) 
Mikania nummularia H protandry, 2a tube Ac. 4 X 1 (3) 5X4 white WH N wasp/fly 
(brush) 
Mikania sessilifolia H protandry, 2a tube Ac, 2 X 1 (5) 4x3 vellowish white GR N wasp/fly 
(dish) 
Senecio oleosus GymnM protandry, 2a tube Ac. 7 X | (40) 10 * 9 vivid yellow YE N. P svrphid/bee nectar- 
Ligulate (brush) Ly, I NI 7 * flower 
Stevia myriadenia H protandry, 2a tube Ac, 5 X «1 (10) 7X6 pink PK N. P several insect groups 
(brush) 
Symphyopappus compressus M protandry, 2a tube Ac, 6 X 1 (7) 12 x 11 white WH N. P' several insect groups 
(brush) 
Vernonia herbacea H protandry, 2a tube Ac. 5 X «1 (15) 11 X 10 purplish magenta VI N bee nectar-flower 
(brush) 
Vernonia megapotamica H protandry, 2a tube Ac, 2 X 1 (12) 5X4 purple VI N, P several insect groups 
(brush) 
Vernonia cf. rosea H protandry, 2a tube Ac. 10 X 1 (21) 15 X 14 purple VI N bee nectar-flower 
(brush) 
Vernonia tomentella H protandry, 2a tube Ac. 9 X «1 (20) 15 X 14 purple VI N.P bee nectar-flower 
(brush) 
Vernonia tragiaefolia H protandry, 2a tube Zy, 4 X «1 (9) 9x7 magenta VI N bee nectar-flower 


(brush) 


USPIBE) [eoiuejog unossi|A 


5 


əy} Jo sjeuuy 


Table 2. Continued. 


900c 


Sexual Floral and Flower symmetry, Gynoecium/Androecium Color Floral 
Family/Species system reproductive traits Floral shape size (mm) (mm) Flower color! group! resource — Pollination system 
Vernonia westiniana H protandry, 2a tube Ac, 6 X <1 (11) 10 x 7 reddish purple VI N.P bee nectar-flower 
(brush) 
Bromeliaceae 
Dyckia tuberosa H herkogamy tube Zy. 13 X 5 10 x 12 reddish orange RE N hummingbird 
Campanulaceae 
Lobelia camporum H protandry, 2a gullet Zy. 8 X 3 14 X 12 lilac PK N bee nectar-flower 
Wahlenbergia brasiliensis H protandry, 2a bell Ac, 3 X 3 5X4 lilac PK N. P syrphid/bee nectar- 
lower 
Clethraceae 
Clethra scabra H bell Ac,2 X2 2X2 yellowish white GR N wasp/fly 
Jonvolvulaceae 
Convolvulus crenatifolius H heterandry bell Ac, 12 X 14 7X 7,6 lavender PK NP syrphid/bee nectar- 
flower 
Ipomoea procumbens H herkogamy, bell Ac, 38 X Il 26 X 25,21 pink PK N,P bee nectar-flower 
heterandry 
Jacquemontia grandiflora H herkogamy bell Ac, 18 X 23 14 X 11 violet VI N, P? bee nectar-flower 
Cunoniaceae 
Weinmannia organensis H dish Ac, d3 2x2 pale vellow GR N several insect groups 
Droseraceae 
Drosera montana H dish Ac, d10 3x2 pink PK P syrphid/bee pollen- 
lower 
Ericaceae 
Agarista hispidula H 2a tube Ac, ll X 2 8X 7 red RE N hummingbird 
Gaylussacia chamissonis H 2a bell Ac, 7 * 3 7X5 white WH N bee nectar-flower 
Gaylussacia jordanensis H bell Ac,9 X 6 10 * 5 white WH N. P bee nectar-flower 
Eriocaulaceae 
Paepalanthus paulensis U-M inconsp. Ac, dl 2X vellowish white GR N several insect groups 
inconsp. Ac, dl x2 N,P 
Paepalanthus polyanthus U-M inconsp. Ac, dl 2 * greenish white GR N several insect groups 
inconsp. Ae, dl x 2 N, P 
Erythroxylaceae 
Erythroxylum microphyllum H heterostyly dish Ae, d7 5X2,2 X5 pastel yellow GR N wasp/fly 


€ JequinN 'e6 euinjoA 


BwIZeS Y Seal 


pue|sseJc) pH- e ur uoneuijog 


SLY 


Table 2. Continued. 


Sexual Floral and Flower symmetry, Gynoecium/Androecium Color Floral 
Family/Species system reproductive traits Floral shape size (mm) Flower color’ group! resource Pollination system 
Escalloniaceae 
Escallonia farinacea H protandry, bell Ac, 3 X 2 5X4 white WH N bee nectar-flower 
herkogamy 
Euphorbiaceae 
Croton dichrous U-M temporal dish Ac, d2 2 * yellowish white GR N wasp/fly 
dioecism dish Ac, d3 x2 . 
entianaceae 
Calolisianthus pedunculatus H protandry tube Ly, 32 X 12 30 x 29 deep red RE N? undet. 
Calolistanthus pendutus H protandry, tube Zy, 25 X 9 22 x 20 violet VI N? undet. 
protogyny 
Deianira nervosa H herkogamy dish Ac, dl5 3x2 reddish lilac PK P syrphid/bee pollen- 
protogyny ower 
Helia oblongifolia H protandry tube Zy, 20 X 3 16 X 15 vellowish green GR N 
ygostigma australe H protogyny dish Ac, d20 3x2 purple VI P? syrphid/bee pollen- 
flower 
Gesneriaceae 
Sinningia allagophylla H protandry, tube Zy, 14 X 3 11 X 14 reddish orange RE N hummingbird 
herkogamy 
Hypericaceae 
Hypericum brasiliense H apomixis dish Ac, d22 4x3 vivid vellow YE P syrphid/bee pollen- 
flower 
Hypericum ternum H apomixis dish Ac, d22 4x3 vivid yellow YE P? uncertain 
Iridaceae 
Alophia geniculata H herkogamy bell Ac, 20 X 14 23 X 12 violet VI P, N? bee pollen-flower 
Alophia sp. indet. 1 H herkogamy bell Ac, 20 X 12 10x 8 lilac PK P. N? uncertain 
Calydorea campestris H dish Ac, d28 5x4 purplish violet VI P syrphid/bee pollen- 
flower 
Sisyrinchium micranthum H tube Ac, 5 X2 3 * 3 purple VI P,O  syrphid/bee pollen- 
ower 
Sisyrinchium vaginatum H dish Ac, d26 6 * 5 vivid yellow YE P. O0 syrphid/bee pollen- 
ower 
Lamiaceae 
Hyptis lippioides H protandry gullet Zy,4 X 2 12 x 8 grayish violet VI N bee nectar-flower 
Hyptis plectranthoides H protandry gullet Ly, 2X 1 6X5 violet VI N bee nectar-flower 


dope [eoiuejog D 


94 


eui Jo sjeuuy 


Table 2. Continued. 
Sexual Floral and Flower symmetry, Gynoecium/Androecium Color Floral 
Family/Species system reproductive traits Floral shape size (mm) (mm Flower color! group! resource Pollination system 
yptis umbrosa H protandry gullet Zy.3X] 5X5 olet VI N bee nectar-flower 

Peltodon radicans H protandry gullet Ly, 5X] 8X6 bluish violet VI N bee nectar-flower 
Leguminosae 

Chamaecrista sp. indet. l H herkogamy, bell Zy, 12 X 7 12 X 12, 4 vivid yellow YE P bee pollen-flower 

eterandry 

Crotalaria breviflora H heterandry, 2a flag Zy, k12 X f12 11 x 10 vivid yellow YE N, P bee nectar-flower 

Lupinus velutinus H heterandry, 2a flag Zy: k13 X f16 12 x 12 violet VI P bee pollen-flower 
sythraceae 

Cuphea im H protandry tube Zy,8 X 2 9x9 purple VI N bee nectar-flower 
Malpighia 

Byrsonima pu H dish Zy, d13 X 10 3 * 3 vivid yellow YE O.P bee oil-flower 
Malvaceae 

Pavonia kleinii H herkogamy bell Ac, 9X 9 7X6 yellow YE N.P bee nectar-flower 

Sida sp. indet. 1 H herkogamy bell Ac, 18 X 25 17 x 9 pastel pink PK N bee nectar-flower 
Melastomataceae 

Leandra erostrata H apomixis? dish Ac, d7 6 X6 in PK P? uncertain 

Leandra sp. indet. 1 H apomixis? dish Ac, d6 5 * 5 white WH P? uncertain 

Leandra sp. indet. 2 H apomixis? dish Ac, d7 14 * 12 pink PK P? uncertain 

Leandra sp. indet. : H apomixis? dish Ac, d8 15 X 13 pink PK p? uncertain 

Microlicia isophylla H apomixis?, dish Zy, d20 3X4 magenta VI P ndet 

heterandry 

Tibouchina frigidula H heterandry dish Zy, dad 14 X 16 deep violet VI P bee pollen-flower 

Tibouchina martialis H heterandry dish Zy, d55 15 x 20 reddish violet VI P bee pollen-flower 

Tibouchina minor H heterandry dish Zy, d30 6x7 iolet VI P bee pollen-flower 

Trembleya parviflora H heterandry dish Zy, d18 0X5 ink PK P bee pollen-flower 

Trembleya phlogiformis H heterandry dish Zy, d22 4 X4 purple VI P bee pollen-flower 
Ochnaceae 

Ouratea semiserrata H protandry dish Ac, dl6 6X5 vivid yellow YE E bee pollen-flower 
Orchidac 

bur. secundum H tube Zy, 5 X <] 6 * 6 purple VI N undet. 

Habenaria parviflora H tube, spur Zy, 8 X <l 9x9 grayish green GR N undet. 

Oncidium barbaceniae H dish Zy, d22 x 19 3x2 vivid yellow YE 07 uncertain 

Oncidium blanchetii H dish Zy, d24 X 15 2x2 vivid yellow YE 0? undet. 

Oncidium sp. indet. ! H dish Zy, dl5 X 11 2 * 2 vivid yellow YE 0? undet. 


900c 


€ JequinN ‘g6 euinjoA 


pue|sseJc) epniniv-uBi e ur uoneurjog 


ewIzes y Seal 


¿Lv 


Table 2. Continued. 
Sexual Floral and Flower symmetry, ( /And ium Color Floral 
Family/Species system reproductive traits Floral shape size (mm) (mm) Flower color. group' resource Pollination system 
Orobanchaceae 
Esterhazya macrodonta H protandry, gullet Zy, 24 X 10 38 X 34 reddish orange RE N hummingbird 
herkogamy 
Polygalaceae 
Polygala brasiliensis H protandry? flag Zy, k2 X f3 2x2 reddish purple VI N. P? bee nectar-flower 
Polygala cneorum H protandry? flag Zy, k3 X f4 3x3 grayish magenta VI N. P? bee nectar-flower 
Rubiaceae 
Borreria capitata H herkogamy, tube Ac, 3 X] 4x 4 white WH N several insect groups 
protogyny? 
Borreria tenella H herkogamy, tube Ac, 2 X |] 3x3 purple VI N wasp/fly 
protogyny 
Declieuxia cordigera var. H heterostyly tube de, 7X] 8 * 6.6 X 8 purple VI N bee nectar-flower 
angustifolia 
Galianthe angustifolia H heterostyly tube Ac, 3 X] * 2,2 X4 — yellowish white GR N several insect groups 
Galianthe brasiliensis H heterostyly tube Ac, 3 X ] J * 3. 3 * 4 White WH N several insect groups 
Galium hypocarpium H dish Ac, dl 2x] greenish vellow GR N wasp/fly 
Solanaceae 
Solanum aculeatissimum AndroM dish Ac, d12 10 x 8 vivid vellow YE P bee pollen-flower 
Solanum americanum H dish Ac, d4 l l vivid yellow YE P bee pollen-flower 
Solanum pseudocapsicum H dish Ac. dll 1x3 vivid vellow YE P bee pollen-flower 
Solanum swartzianum H dish Ac, d17 5X4 vivid yellow YE P bee pollen-flower 
Solanum viarum AndroM dish Ac, dl6 10 x 8 yellowish green GR P bee pollen-flower 
Verbenaceae 
Verbena hirta H herkogamy, tube Ac, 6 X | 3 * 4,5 purplish violet VI N bee nectar-flower 
heterandry 
Violaceae 
Viola cerasifolia H herkogamy tube Zy,6 X3 5X4 violet VI P.N — bee pollen-flower 
Xyridaceae 
Xyris asperula H dish Ac, d13 4X4 vivid vellow YE P syrphid/bee pollen- 
flower 
Xyris tortulla H herkogamy dish Ac, d22 3x3 deep vellow YE P syrphid/bee pollen-flower 
Abbreviations: Ac = actinomorphic, ie = andromonoeci ious. D = dioecious. GymnM = gymnomonoecious. H = hermaphrodite, inconsp. = inconspicuous, Ligulate = ligulate floret, M 
= monoecious, N = nectar, O = oil, P = pollen, U = unisexual 15 ae n undet. = undetermine d. Zy = zygomorphic, 2a = secondary oe n Sl cg 


' Main color el ius except 


Sina ‘eae (anthe TR) m. Cunoniac 'eae Sepa 


n 


flowe 1 
Bl 


* Hummingbirds y 


] 
the large Pl 


PK = pink, RE = red, VI — 


in brackets) 7 visits. Nectar is concealed in a tube at the corolla base. 


violet. 


WH = vy 


YE 


= yellow. 
n 72 7 id 


WHOSE SIZE 15 ue 


lla length. 


dp jeoiuejog unossi|A 


84v 


əy} JO sjeuuy 


Volume 93, Number 3 


Freitas & Sazima 


2006 Pollination in a High-Altitude Grassland 
alcohol-acetic acid (FAA) at 12 to 60 hours after self respectively: others were designated as secondary 


or cross hand-pollination and analyzed under fluores- 
cence microscopy to observe pollen tube growth 
(Martin, 1959). Plants 


checked monthly from December 


the transects were 
1998 to February 


2000, and two phenology parameters were recorded: 


along 


flowering time—months in which each species was in 
and the activity blooming peak—months in 
50% of 


species presented flowers (Table | 


flower; 


which more than f the individuals of each 


The pollinators for each species were determined 
through direct observation of flowers for at least four 
hours (Tables 3, 4). Observation periods were one to 
two hours, but longer periods (three to eight hours) 
were necessary for some species with very low 
frequency of visits; these were usually observed for 


a total of 12 or Table 4). We 


concentrated our observations on the most probable 


more hours (see 


time of pollinator visitation for each species. Most 


observation sessions occurred in the middle of the 


day, because insects are generally more active in this 
period. The total observation times only include the 


observations under good climatic conditions and 


during the flowering peak of each species. To be 


considered pollinators, visitors had to have pollen 


erains attached to their bodies and contact the stigmas 
during their visits. Hummingbirds were identified in 
the field or through photographs. Insects visiting the 
with an 


collected, when possible. 


— 


lowers were 
entomological net. Insect specimens were identified 
ZUEC, 


MPEG. Non-captured insect visitors were grouped 


by specialists and deposited at MZSP, and 
into different hierarchical categories, at least to order 
and, in a few cases, to genus level (e.g.. Bombus Latr.). 
For the broad community study, we categorized 
captured and non-captured insect visitors into seven 


os (after Root, 1967, Cummins, 1973): 


other 


functional grou 


— 


large bees, small bees. syrphids, 


dipterans, butterflies, and beetles (Tables 4. 5). Large 


wasps. 


bees are those larger than 12 mm (because it is not 


indigenous to Brazil, Apis mellifera was not included 
in any bee group). 

Plant species were divided into nine pollination 
systems. For those species that were pollinated by one 
or two functional groups (monophilous and oligophi- 
lous; see below), pollination system was based on sets 
of floral traits common to plants pollinated by each 


functional group. A second set of species were 
pollinated by three or more functional 


which one or two were clearly identified as main 


pollinators (polyphilous), and pollination system was 


Main 


designated according to the main pollinator. 


pollinators were defined based on pollinator morphol- 
ogy, behavior, and frequency of visits: usually over 


60% or 80% of visits for one or two main groups, 


groups of 


pollinators. The final set of species was pollinated by 
three or more functional groups of apparently similar 
importance (holophilous) and thus did not have any 
main pollinator(s). These species were grouped in the 


pollination system named "several insect groups” 


(SIG), 


indistinct to distinguish 


and their pollinators were designated as 


them from the secondary 
and main pollinator categories. 

Plant species were classified into four categories 
representing the degree of specialization/generaliza- 
tion of their interaction with their pollinators: 
(Ta- 


categories 


monophily, oligophily, polyphily, and holophily 
ble 4). 
differs from that given by Faegri and van der Pijl 
1979), 


groups rather than pollinator 


The definition of our first three 


because our unity of reference was functional 


species. Thus, here, 
monophily refers to species exclusively pollinated by 
one functional group. Plant species pollinated by two 
functional groups (either one or both are their main 


pollinator[s]) are considered. oligophilous. Species 


pollinated by three or more functional groups are 
classified either as polyphilous, when one or two 


eroups act as their main pollinators, or holophilous. 
whe “th the re is no preponde rance of one or two groups, 


i.e., when groups act as indistinct pollinators. 


RESULTS 


We studied a total of 124 species from 35 families 


Asteraceae 


(Table 1). Forty species belonged to 
(32%). followed by Melastomataceae (10 spp.. 8%). 
On the other hand, 16 (46%) and seven (20%) 


families were represented by only one and two 


species, respectively. 


FLORAL BIOLOGY AND POLLINATOR AGENTS AT THE 


Community LEVEL 
V. FLORAL TRAITS 

The most common flower color groups were violet 
29.8% of the 124 species), greenish (28.2%), and 
vellow (17.0%). Pink (10.5%), white (9.7%), and red 
(4.8%) were the least represented groups (Table 2). 
Mosl (47.696) 


8.296). Since we considered the shape of the head of 


flowers were tubular or dish-shaped 


Asteraceae as a whole instead of each of its florets, 
dish-shaped the 
(43.6%). followed by the brush (16.9%), 
15.3%) types. Bell (11.3%), and gullet, inconspicu- 
2.8% altogether) were the 
Table 2). Most plants had 
or zygomorphic flowers 


4.8%) bore both 


blossoms were most frequent 


and tube 


ous, flag. and revolver (1 


least represented flowers ( 


— 


either actinomorphic (69.4% 


(25.8%), 


— 


though a small percentage 


480 Annals 
MES a Garden 


Table 3. Pollinator species and the plant species it was collected from or observed pollinating (in the case « 
hummingbirds) at the Serra da Bocaina grasslands. Pollinators that were seen but not collected are included in Table 4. List of 
the abbreviations of the plant families: AMA, Amaryllidaceae; API, Apiaceae; APO, Apocynaceae; AQU, Aquifoliaceae; AST, 
Asteraceae; BRO, Bromeliaceae; CAM, Campanulaceae; CLE, Clethraceae; CON, Convolvulaceae: CUN, Cunoniaceae; DRO, 
Droseraceae; ERIC, Ericaceae; ERIO, Eriocaulaceae; ERY, Erythroxylaceae; ESC, Escalloniaceae: EUP, Euphorbiaceae; 
GEN, Gentianaceae; GES, Gesneriaceae; HYP, Hypericaceae; IRI, Iridaceae; LAM, Lamiaceae; LEG, Leguminosae; LYT, 

Lythraceae; MALP, Malpighiaceae; MALV, Malvaceae; MEL, Melastomataceae; OCH, Ochnaceae; ORC, Orchidaceae; ORO, 
Orobanchaceae; POL, Polygalaceae; RUB, Rubiaceae; SOL, Solanaceae: VER, Verbenaceae; VIO, Violaceae; 
XYR, Xyridaceae. 


Pollinator family and species Plant family Plant species 


Trochiliformes 


Trochilidae (hummingbirds) 


Chlorostilbon aureoventris berlepschi ERIC: Agarista hispidula 
Pinto 
Clytolaema rubricauda (Boddaert) GES: Sinningia allagophylla 
Colibri serrirostris (Vieillot) BRO: Dyckia tuberosa 
GES: Sinningia allagophylla 
Leucochloris albicollis (Vieillot) AMA: Hippeastrum glaucescens 
ERIC: Agarista hispidula 
GES: Sinningia allagophylla 
ORO: Fsterhazya macrodonta 
Stephanoxis lalandi lalandi (Vieillot) GES: Sinningia allagophylla 


Coleoptera (beetles) 


Buprestidae (jewel beetles) 


Conognatha Cobos sp. indet. | ERIO: Paepalanthus paulensis 
Cantharidae (soldier beetles) 
Discodon tucumanum Pic RIO: Paepalanthus paulensis 
Cantharidae sp. indet. | API: Eryngium canaliculatum 
AST: Baccharis pentziifolia, Baccharis tarchonanthoides, Barrosoa 


betoniciiformis, „ megacephalum, 11 
integerrima, Erigeron maximus, Eupatorium sp. 
Grazielia pM PS 
ERIO: — Paepalanthus polyanthus 
Cantharidae sp. indet. 2 CLE: Clethra scabra 
CUN: Weinmannia organensis 
Cerambicidae (longhorn beetles) 
Rhinotragus festivus Perty CLE: Clethra scabra 
E 


— 


ateridae (click beetles) 
Cardiorhinus Eschscholtz sp. indet. 1 API: Eryngium canaliculatum 
Melyridae (softwinged flower 
beetles) 
Astylus sexmaculatus Perty AST: Eupatorium sp. indet. 4 
Diptera (flies) 


Bombyliidae (beeflies) 


Euprepina Hull sp. indet. | AST: Campuloclinium megacephalum 
Exoprosopa Macquart sp. indet. | RUB: Horreria capitata 
Paravilla Painter sp. indet. 1 AST: Campuloclinium megacephalum, Chaptalia runcinata, Stevia 
myriadenia, Vernonia megapotamica 
Culicidae (biting ag! 
Culicidae sp. indet. API: Eryngium horridum 
Curtonotidae (c n flies) 
Curtonotidae sp. indet. ! API: Eryngium horridum 
AST: Baccharis platypoda, Mikania sessilifolia 
Curtonotidae sp. indet. 2 RUB: Galianthe angustifolia 


Sarcophagidae (flesh flies) 
Sarcophagidae sp. indet. 1 RUB: Galianthe angustifolia 


Sarcophagidae sp. indet. 2 ERIO: Paepalanthus paulensis 


Volume 93, Number 3 
2006 


Freitas & Sazima 481 
Pollination in a High-Altitude Grassland 


Table 3. Continued. 


Pollinator family and species 


Plant family 


Plant species 


Sciaridae (sciarid flies 


— 


Sciaridae sp. indet. 1 


ma 


Syrphidae (hover-flies 


Allograpta exotica (Wiedemann) 


Palpada rufipedes Thompson 
Pseudodoros clavatus (Fabricius) 
Syrphus phaeostigma Wiedemann 


Toxomerus watsont (Curran 


Toxomerus Macquart sp. indet. 1 


Syrphidae sp. indet. 1 


Syrphidae sp. indet. 
Syrphidae sp. indet. : 
Syrphidae sp. 11 4 
Syrphidae sp. indet. 5 


Syrphidae sp. indet. 6 
Tachinidae (tachinid flies) 


Cylindromyia dorsalis (Wiedemann) 


Jurinella cf. corpulenta (Townsend) 


Tachinidae sp. indet. 1 


Tachinidae sp. indet. 2 
Tachinidae sp. indet. 3 
Tachinidae sp. indet. 4 


Tachinidae sp. indet. 5 
Tachinidae sp. indet. 6 
Tachinidae sp. indet. 7 


Tachinidae sp. indet. 8 
Tachinidae sp. inde L. d 


Tephritidae (fruit f 


Trupanea Sc puc sp. indet. 1 


AST: 
ERIO: 


RUB: 


Baccharis aphylla, Baccharis tarchonanthoides 
Paepalanthus paulensis 


Wahlenbergia brasiliensis 

Zygostigma australe 

Alophia sp. indet. 1, Sisyrinchium micranthum 

Galianthe an ala 

Tassadia subulat 

Borreria capitata, " Calianthe brasiliensis 

Baccharis leptocephala 

Tassadia subulata 

Achyrocline satureioides, Baccharis leptocephala, Baccharis 
pentziifolia, Baccharis tarchonanthoides, Chaptalia runcinata, 
Gochnatia 1 A Mikania sessilifolia, Senecio oleosus 

Wahlenbergia brasiliensi: 

Drosera montana 

Paepalanthus paulensis, Paepalanthus polyanthus 

Croton dichrous 

Alophia d Sisyrinchium vaginatum 

Cuphea glutir 

Borreria 1 Galianthe brasiliensis 

Xyris asperu da, . Xyris tortu 

Tassadia subulata 

Achyrocline satureioides, Chaptalia runcinata 

Paepalanthus paulensis 

Sisyrinchium vaginatum 

Borreria capitata 

Drosera montana 

Xyris asperula 

Alophia geniculata 

Baccharis pentziifolia 

Xyris tortulla 

Mikania sessilifolia, Stevia myriadenia 

Deianira nervosa 

Calydorea campestris 


Chaptalia runcinata 


Grazielia gaudichaudeana 
Hypericum brasiliense 


Ery ngium horri dur 


Baccharis dracunculifolia Baccharis platypoda, Mikania 
sessilifolic 
Eupatorium sp. indet. 2 
Galianthe brasiliensis 
Clethra scabra 
Paepalanthus a 
Baccharis sp. indet. 
Galianthe 1 8 
Eremanthus erythropappus 
'ernonia westiniana 
Galianthe angustifolia 
A line satureioides, Baccharis dracunculifolia, 
Symphyopappus compressus 


Eryngium horridum 


Galium hypocarpium 


Annals of the 
Missouri Botanical Garden 


Table 3. Continued. 


Pollinator family and species 


Plant family 


Plant species 


“amily indet. (flies) 
Diptera sp. indet. | 
Diptera sp. indet. 2 
Diptera sp. indet. 3 
Diptera sp. indet. 4 
Diptera sp. indet. 5 
Diptera sp. indet. 6 
Diptera sp. indet. 7 
Diptera sp. indet. 8 

Hymenopter: 

Andrenidae, Panurginae (dagger 
bees) 

Anthrenoides Ducke sp. indet. 

Apidae, Apini (honey bees) 


\pis mellifera Linnaeus 


Apidae, Bombini (bumblebees) 


Bombus atratus Franklin 


Bombus brasiliensis Lepeletier 
Bombus morto (Swederus) 


Apidae, Centridini (digger bees) 


Centris burgdorfi Friese 


ERIO: 


VIO: 


AQU: 
AST: 


Eryngium horridum 
Chaptalia runcinata 
Chaptalia runcinata 
Baccharis dracunculifolia 
Paepalanthus ae 
Erigeron maxin 

Clethra scabra 


Paepalanthus polyanthus 


Viola cerasifolia 


llex am 
5 0 dim unculifolia, „ pentztifolia, Baccharis 
platypoda, Baccharis sp. indet. 2. Barrosoa betonicitformis. 


:ampuloclinium megacephalum., p eu erythropappus. 
Eupatorium sp. indet. 2, Grazielia gaudichaudeana, 
Hypochaeris gardneri, Mikania lundiana, Stevia myriadenta. 
Symphyopappus compressus, Vernonia westiniana 

Wahlenbergia brasiliensis 

Clethra scabra 

Convolvulus crenatifolius 

saylussacia chamissonis, Gaylussacia jordanensis 

Croton dichrous 

Sisyrinchium vaginatum 

Hyptis lippioides, Hyptis plectranthoides, Hyptis umbrosa, Peltodon 
radicans 

Cuphea glutinosa 

Borreria capitata, Galianthe angustifolia, Galianthe brasiliensis 


Verbena hirta 


Oxypetalum sublanatum 

Campuloclinium megacephalum, Ghromolaena cf. decumbens, 
Stevia myriadenia, Symphyopappus compressus, Vernonia 
tomentella, Vernonia westiniana 

Lobelia camporum 

Ipomoea procumbens, Jacquemontia grandiflora 

Alophia geniculata 

Hyptis lippioides, Hyptis plectranthoides 

Crotalaria breviflora, Lupinus velutinus 

Cuphea glutinosa 

Tibouchina frigidula, Tibouchina minor, Trembleya phlogiformis 

Declieuxia cordigera, Galianthe angustifolia 

Verbena hirta 

Stevia myriadenia, Vernonia westiniana 

Crotalaria breviflora 

Cuphea glutinosa 

Verbena hirta 

Vernonia westiniana 

Cuphea glutinosa 

Galianthe angustifolia 

Verbena hirta 


Lobe lia camporum 


Declieuxia cordigera 


Volume 93, Number 3 
006 


Freitas & Sazima 


Pollination in a High-Altitude Grassland 


483 


Table 3. Continued. 


Pollinator family and species 


Plant family 


Plant species 


Centris discolor Smith 


Centris cf. insularis Smith 


Centris klugi Vriese 


Centris tarsata Smith 


Centris (Melacentris) Moure sp. inde 
Apidae, Ceratinini (dwarf carpenter 
ees) 


Ceratina cf. asuncionis Strand 


Ceratina Latreille sp. indet. 1 
Ceratina sp. indet. 2 
Apidae, Erierocidini 
Mesonychium caerulescens Lepeletier 
& Serv 


Apidae, Eucerini (long-horned 
bees) 
Gaesischia nigra Moure 


Melissoptila aureocincta Url 


E 


1 


Apidae, Meliponini (stingless bees 


Melipona bicolor bicolor Lepeletier 


Melipona quadrifasciata anthidioides 
Lepeletier 


Paratrigona subnuda Moure 


Plebeia droryana (Friese) 


Plebeia saiqui (Friese 


— 


Scaptotrigona bipunctata (Lepeletier) 
Schwarziana quadripunctata 


(Lepeletier) 


l 


MALP: 


AST: 
AST: 


Byrsonima variabilis 
Tibouchina frigidula 


Byrsonima variabilis 


Hypochaeris a Vernonia cf. rosea, Vernonia tomentella 


Chaemaecrista sp. indet. 
Cuphea glutinosa 
Byrsonima variabilis 
Tibouchina minor 
Byrsonima variabilis 
Verbena hirta 


Byrsonima variabilis 


Senecio oleosus, Vernonia herbacea, Vernonia tomentella 


Lobelia camporum 
Sida sp. indel 
Verbena hirta 
Cuphea glutinosa 


Convolvulus crenatifolius 
Vernonia tomentella 
Cuphea glutinosa 


Galianthe angustifolia 


Verbena hirta 


Vernonia tragiaefolia 


Symphyopappus COTUpressics 


Eremanthus erythropappus, Eupatorium sp. indet. 3, Grazielia 


gaudichaudeana, Mikania lundiana, Vernonia megapotamica, 


Vernonia tomentella, Vernonia westiniana 


Clethra scabra 
Gaylussacia chamissonis 
Alophia geniculata 
Hyptis lippioides 
Trembleya parviflora 

0 


uratea semiserrata 


Campuloc! linium megacephalum, Eremanthus erythropappus, 
ik 


~ 
= 
=. 


— 


UNGLANGa 


Achyrocline satureioides 


Eryngium horridum 
Ilex amara 

Vernonia westiniana 
Gaylussacia jordanensis 


Alophia geniculata, Calydorea campestris 
Hyptis lippioides 
Byrsonima variabilis 


Clethra scabra 


indet. 4, Grazielia gaudichaudeana, Mikania lundiana 
Clethra scabra 


Hypericum brasiliense 


indiana, Symphyopappus compressus 
Chaptalia NT. Eremanthus erythropappus, Mikania 
li 


Baccharis sp. indet. 2, Eremanthus erythropappus, Eupatorium sp. 


Annals of the 
Missouri Botanical Garden 


Table 3. Continued. 


Pollinator family and species 


Plant family 


Plant species 


Trigona spinipes (Fabricius) 
g ping 


Apidae, Tapinotaspidini (digger 

bees) 

Monoeca Lepeletier & Serville sp. 
indet. 

Paratetrapedia (Lophopedia) ct. 
pygmaea (Schrottky) 


Paratetrapedia (Xanthopedia) 
Michener & Moure sp. indet. 1 
Paratetrapedia (Trigonopedia) Moure 
sp. indet. 1 
Apidae, Xylocopini (giant carpenter 
bees) 


Xylocopa brasilianorum (Linnaeus) 


Braconidae (braconid wasps) 


Braconidae sp. indet. 


— 


Chrysididae (cuckoo w ea 
Chrysididae sp. indet. 
Colletidae, Colletini 1 

vees) 


Colletes Latreille sp. indet. 1 


Eurytomidae (eurytomid wasps) 
Eurytomidae sp. indet. | 
Gasteruptiidae (gasteruptid wasps) 
Casteruptiinae sp. indet. | 
Halictidae, Augochlorini (sweat 
bees) 
Augochlorodes turrifaciens Moure 


Augochloropsis cf. cognata Moure 


Augochloropsis cyanea (Schrottky) 


Augochloropsis iris (Schrottky) 


Augochloropsis Cockerell sp. indet. 1 


5 


AST: 
ERIO: 
LAM: 


MALP: 


AST: 


XYR: 
MALP: 


AST: 


AST: 


AST: 


AST: 


ESC: 


RUB: 


EUP: 


Baccharis sp. indet. 2 
Paepalanthus paulensis 


Hyptis lippioides 


Byrsonima variabilis 
Stevia myriadenia 


Xyris tortulla 
Byrsonima variabilis 


Stevia myriadenia 


Symphyopappus compressus 
Ipomoea procumbens 
Tibouchina frigidula, Tibouchina martialis 


Verbena hirta 
Baccharis platypoda 


Baccharis dracunculifolia 


Hypochaeris gardneri 


gd] farinacea 
Galium hypocarpium 


Croton dichrous 


Viola cerasifolia 

Ilex amara 

Campuloclinium megacephalum, Vernonia tomentella 

Clethra scabra 

Croton dichrous 

Calydorea campestris 

Hyptis lippioides 

Byrsonima varias 

Pavonia klei 

Borreria 5 Galianthe brasiliensis 

Barrosoa betoniciiformis, Campuloc linium megacephalum, 
Chaptalia runcinata, Eupatorium sp. indet. 4, Grazielia 
gaudichaudeana, Vernonia tomente lla 

Gaylussacia jordanensis 

Croton dichrous 

Cuphea. glutinosa 

Galianthe angustifolia, Galianthe brasiliensis 

Solanum aculeatissimum, Solanum pseudocapsicum, Solanum 
swartzianum, Solanum viarum 

Xyris asperula 


Campuloclinium megacephalum 


Alophia geniculata 


Eupatorium sp. indet. 2, Stevia myriadenia 


Volume 93, Number 3 
2006 


Freitas & Sazima 
Pollination in a High-Altitude Grassland 


485 


Table 3. Continued. 


Pollinator family and species 


Plant family 


Plant species 


Ceratalictus Moure sp. indet. 1 


Paroxystoglossa cf. jocasta 
(Schrottk y) 

Paroxystoglossa Moure sp. indet. 1 

Pseudaugochlora cf. graminea 
(Fabricius) 

Augochlorini sp. indet. 1 

Halictidae, Halictini (sweat bees) 
cod Guerin-Meneville sp. 
inde 


3L. 
TUNE T" sp. indet. 


Dialictus sp. indet. 2 

Pseudagapostemon cyaneus Moure & 
akagami 

Halictini sp. indet. 1 


Halictini sp. indet. 2 


Leucospidae EUM wasps) 
Leucospidae sp. 1 
Megachilidae, Hem (mason 
and carder bees 
Anthidium sertanicola Moure & 


Urban 


9 1 5 autumnale (Schrottky) 
Megachilida egachilini (leaf- 
cutter, mason and cuckoo bees) 
Coelioxys Latreille sp. indet. 1 
Coelioxys sp. indet. 2 
Megachile cf. anthidioides 
Radoskowski 


Megachile iheringi Schrottky 


Megachile laeta Smith 


Gaylussacia chamissonis, Gaylussacia jordanensis 

Hyptis umbrosa 

Cuphea. glutinosa 

Trembleya din 

Polygala cneorum 

Galianthe Dm 

Solanum pseudocapsicum 

Ilex amara 

Vernonia westiniana 

Gaylussacia jordanensis 

Erythroxylum microphyllum 

Zygostigma australe 

Alophia sp. ndet. 1 

Hyptis 15 Hyptis plectranthoides, Hyptis umbrosa 
Borreria capitata, Borreria tenella, Galianthe brasiliensis 


Alophia geniculata 


Chaptalia integerrima 


Verbena hirta 
Declieuxia cordigera 
Symphyopappus compressus 


Gaylussacia jordanensis 

Alophia geniculata, Calydorea campestris, Sisyrinchium 
micranthum 

Ouratea semiserrata 

Viola cerasifolia 

Chaptalia runcinata 


Mikania lundiana, Senecio oleosus 


Hypericum brasiliense 
Alophia geniculata 


Verbena hirta 


Baccharis intermixta 


Crotalaria breviflora 


Cuphea glutinosa 


Stevia myriadenia 


Erigeron maximus, Senecio oleosus 
Stevia myriadenia 
Symphyopappus compressus 


Cuphea glutinosa 
Eupatorium sp. indet. 1, Vernonia tomentella 
Gaylussacia ee 
Hypericum brasiliense 
Lupinus velutinu 
Lupinus n 
'erbena hirta 


Annals of the 
Missouri Botanical Garden 


Table 3. Continued. 


Pollinator family and species 


Plant family 


Plant species 


Megachile terrestris Schrottky 
Pompilidae (spider wasps) 
Pompilidae sp. indet. 
Pompilidae sp. indet. 2 
Pompilidae sp. indet. 3 


Pompilidae sp. indet. 


Scoliidae (digger wasps 
Scoliidae sp. indet. | 
Sphecidae (mud-dauber and sphecid 
wasps) 
Bicyrtes paranae Bohart 
Pryonyx thomaz (Fabricius) 
Sphex dorsalis Lepeletier 
Sphex opacus Dahlbom 

l 


Cercerini sp. indet. 2 


Cercermi sp. ine 


Larrinae sp. indet. 1 
Tenthredinidae (sawflies) 

Tenthredinidae sp. indet. 
Tiphiidae (tiphiid wasps) 

Tiphiidae sp. indet. 
Vespidae (paper wasps) 


Agelata vicina (Saussure) 


Brachygastra lecheguana (Latreille) 


Mischocyttarus drewseni Saussure 


Polistes billardieri Fabricius 


Polistes cinerascens Saussure 


Polybia fastidiosuscula Saussure 


Polybia minarum Ducke 
Polybia scutellaris (White) 
Polybia sericea (Olivier) 


Protonectarina sylveirae (Saussure) 

Protopolybia sedula (Saussure) 

Synoeca cyanea (Fabricius) 
Family indet. (wasps) 

Hymenoptera sp. indet. 


Hymenoptera sp. indet. 2 


AST: 


ST 


AST: 


Eupatortum sp. indet. 1. Vernonia tomentella 


Baccharts sp. indet. 2 
Baccharis platypoda 
Baccharis sp. indet. 2 


Baccharis sp. indet. 2 


Symphyopappus compressus, Vernonia westiniana 


Galianthe angustifolia 
Baccharis dracunculifolia 
Galianthe angustifolia 
Eupatorium sp. indet. 4 
Clethra scabra 

Mikania nummularia 
Clethra scabra 


Achyrocline satureioides 
Galianthe brasiliensis 
Baccharis sp. indet. 2 


Baccharis intermixta 
Paepalanthus „ 
Baccharis sp. indet. 

Ilex amara 


Achyrocline satureioides, Baccharis aphylla, Baccharis 


dracunculifolia, Baccharis pentziifolia, Baccharis sp. indet. 2, 


Chaptalia integerrima, Eupatorium sp. indet. 4, Gochnatia 
paniculata 

Clethra scabra 

Erythroxylum microphyllum 

Croton dichrous 

Galianthe brasiliensis 

Ilex amara 

Achyrocline satureioides, Baccharis sp. indet. 2, Baccharis sp. 
indel. 3, Gochnatia paniculata, Mikania 5 

UN ylum microphyllum 


Galianthe brasiliens 


2 


Gonioanthela hilariana 
accharis dracunculifolia, Baccharis platypoda 
Baccharis platypoda, Baccharis sp. indet. 2 
Clethra scabra 
Galianthe angustifolia 
Clethra scabra 
Weinmannia organensis 
Baccharis tarchonanthoides 
Croton dichrous 
Gontoanthela hilariana, Oxypetalum appendiculatum 
Borreria tenella 
Baccharis curitybensis 
Baccharis sp. indet. 2 


Clethra scabra 


Achyrocline satureioides, Mikania nummularia 


Lucilia lycopodioides, Mikania nummularia 


Volume 93, Number 3 
2006 


Freitas & Sazima 487 


Pollination in a High-Altitude Grassland 


Table 3. Continued. 
Pollinator family and species Plant family Plant species 
ERY: Erythroxylum microphyllum 
RUB: Borreria capitata, Borreria tenella 
Hymenoptera sp. indet. 3 AST: Mikania sessilifolia 
LYT: Cuphea glutinosa 
Hymenoptera sp. indet. 4 AST: Baccharis aphylla 
Hymenoptera sp. indet. 5 AST: Baccharis pentziifolia 
Hymenoptera sp. indet. 6 AST: Eupatorium sp. indelt. í 
Hymenoptera sp. indet. 7 AST: Achyrocline satureioides 
Hymenoptera sp. indet. 8 AST: Achyrocline satureioides 
Hymenoptera sp. indet. 9 AST: Achyrocline satureioides 
Hymenoptera sp. indet. 10 AST: Baccharis pentziifolia 
CLE Clethra scabra 
Hymenoptera sp. indet. 11 LYT Cuphea glutinosa 
Hymenoptera sp. indet. 12 RUB Galianthe brasiliensis 
Hymenoptera sp. indet. 13 AST: Grazielia gaudichaudeana 
Hymenoptera sp. indet. 14 AST: Baccharis dracunculifolia 
Hymenoptera sp. indet. 15 RUB: Galianthe angustifolia 


Lepidoptera (butterflies) 


Hesperiidae (skippers and darters) 


Sarbia cf. damippe (Mabille & 
Boullet) 


Sarbia cf. m (Latreille) AST: 
Thespieus Godman sp. indet. 1 AST: 
Urbanus Hübner sp. indet VER: 
Hesperiinae sp. indet AST: 
Pyrrhopyginae sp. indet. AST: 
Lycaenidae (coppers and n streaks) 
ecla Vabricius sp. indet. 1 AST: 
RUB 
Riodininae sp. indet. | API: 
Nymphalidae (danaids and browns) 
Agraulis vanillae maculosa (Stichel) AST: 
Vanessa myrinna (Doubleday) AST: 
Vanessa Fabricius sp. indet. ! AST: 
Yphthimoides ochracea (Butler) AST: 
CLE: 
Pieridae (whites and yellows) 
Eurema nise tenella (Boisduval) LYT: 
VER 
Hesperocharis erota (Lucas) AST: 


Eupatorium 5 inde 


Barrosoa betonictiformis, Campuloclinium megacephalum 


Vernonia westiniana 

Campuloclinium megacephalum 

Verbena hirta 

Grazielia gaudichaudeana 

Chaptalia integerrima, Vernonia westiniana 


Chaptalia runcinata, Mikania nummularia, Vernonia westiniana 
Galianthe brasiliensis 


Eryngium horridum 


Vernonia megapotamica 


Campuloclinium o Eupatorium sp. indet. 


4. Vernonia tomentella 


Eupatorium sp. indet. 3 
Symphyopappus compressus 
Clethra scabra 


Cuphea glutinosa 
Verbena hirta 


Vernonia westiniana 


types. Small flowers were predominant. Out of the 65 


species with tubular or gullet flowers, for instance, 


55.4% had corolla tubes = 5 mm long, whereas only 
13.8% had >10 mm Nectar 


exclusive or main resource of 71.0% of the species. 


tubes long. was the 
Although pollinators fed on or collected pollen from 
more than half of the species, it was the exclusive 
resource of only 17.7% of the species. Even though we 
only measured the nectar production of a few species, 


low nectar production (1.e., less than 2 ul) prevailed. 


The great majority of species presented hermaph- 
(82.396), and a few were either 
dioecious (8.996) (Table 2). 


atter group only comprised the 11 species of 


roditic flowers 
monoecious (3.296) or 


Baccharis Mill. (Asteraceae). Seven other species of 


Asteraceae (5.696) presented both hermaphroditic and 


unisexual (ligulate) flowers (gymnomonoecism). We 
detected spontaneous self-pollination in some species 
including plants that were also pollinated by animals 


(Table 4. 


Table 4. Plant species, time of observation, pollination system. 


5 N — nocturnal observation; undet. — = undetermined pollination system 


= large bees, BS = small bees, 


1 to be confirmed; uncoll — 


CO = beetles, DI = 


uncollected insect o (only observed). 


except syrphids, HB = bien LE = 


degree of 1 VV ae collected and observed pollinators at the Serra da Bocaina grasslands. 
ONO = monophily, OLIGO = oligophily, POLI = polyphily, HOLO = oe n/a = not applicable: 
pe, SSP = spontaneous self-pollination, SY yrphids, WA = wasps; ? = 


Time of observ. 


Degree of specialization/ 


Family/Species (hours) Pollination system generalization Pollinator species 
Amaryllidaceae 
Hippeastrum glaucescens 18.5 hummingbird MONO HB - Leucochloris albicollis 
Apiaceae 
Eryngium canaliculatum 4 uncertain CO - Cardiorhinus sp. indet. 1, Cantharidae sp. indet. 
Eryngium horridum 4 several insect groups HOLO BS - Plebeia saiqui; DB- 1 cf. corpulenta, Culicidae sp. 
indet. 1, pM" sp. indet. 1, Tachinidae sp. indet. 9, 
d sp. indet. I: LE - Rp sp. indet. 1; SY - uncoll 
Sp.): WA - 1 (1 sp.) 
Apocynaceae 
Gonioanthela hilariana 5.5 wasp/fly MONO WA - Polistes cinerascens, Polybia sericea 
Wandevilla erecta 18 N undet. n/a not visited 
Oxypetalum appendiculatum 1.5 wasp/fly MONO - Polybia sericea 
Oxypetalum sublanatum 11 bee nectar-flow MONO BL - Bombus atratus 
Tassadia subulata 5 syrphid nectar- flower MONO SY - Palpada rufipedes. Toxomerus watsoni, Toxomerus sp. indet. 
l. uncoll (1 sp.) 
Aquifoliaceae 
Ilex amara 1.5 bee nectar-flower OLIGO "e o BS - Augochloropsis cf. cognata, Ceratalictus sp. 
ebeia saiqui; WA - Mischocyttarus drewseni. Polistes 
vdd uncoll Vespidae (1 sp.). Pompilidae (1 s 
Asteraceae 
Achyrocline satureioides 6 wasp/fly POLI BS - Plebeia droryana; DI - Tachinidae sp. indet. 8: SY - 
dion watsont, Toxomerus sp. indet. 1, uncoll (2 spp.): 
schocyttarus drewseni, dae billardiert, Larrinae sp. 
indet E Hyn menoptera sp. indet. 1, Hymenoptera sp. indet. 7, 
Hymenoptera sp. indet. 8, EL. sp. indet. 9, uncoll (1 
sp. ) 
Baccharis aphylla n wasp/fly OLIGO DI - Sciaridae sp. indet. 1; WA - Mischocyttarus drewseni, 
Hymenoptera sp. indet. 4, uncoll (1 sp.) 
Baccharis curitybensis 4 wasp/fly MONO WA - Protonectarina sylveirae 


uapey [eoiuejog UNOSSIIN 


88r 


au} jo s¡euuy 


Table 4. Continued. 


Time of observ. 


Degree of specialization/ 


Family/Species Pollination system generalization Pollinator species 
Baccharis dracunculifolia 7.5 wasp/fly OLICO Apis buds DI - Jurinella cf. 17 5 nta, Tachinidae sp. indet. 
8, Diptera sp. indet. 4; WA - Mischocyttarus drewseni, Polistes 
cinerascens, Pryonyx DUIS Chrysididae sp. indet. 1, 
Hymenoptera sp. indet. 
Baccharis intermixta 4 wasp/fly MONO WA - Agelata vicina, Leucospidae sp. indet. 1, uncoll (1 sp.) 
Baccharis leptocephala 4 wasp/fly OLIGO SY - Syrphus phaeostigma, Toxomerus watsoni; WA - uncoll (2 
spp.) 
Baccharis pentziifolia 9.5 wasp/fly POLI Apis bp ds CO - Cantharidae sp. indet. 1; DI - e a Sp.): 
SY - Toxomerus watsoni, Syrphidae sp. indet. 3; WA 
5 5 ttarus drewseni, Hymen „ Sp. 1 5. 
Hymenoptera sp. indet. 10, uncoll (1 sp.) 
Baccharis platypoda 4 wasp/fly POLI Apis mellifera; BS - uncoll ip bir. a E DI - Jurinella cf. 
corpulenta, Curtonotidae sp. indet. 1; WA - Polistes 
cinerascens, Polybia fastidiosuscula, Braconidae sp. indet. 1, 
Pompilidae sp. indet. 2 
Baccharis tarchonanthoides 4 several insect groups HOLO BS - uncoll Augochlorini a sp.); CO - Cantharidae sp. indet. 1; DI 
Sciaridae sp. indet. 1, uncoll Tachinidae (2 spp.); SY - 
Toxomerus watsoni, uncoll (1 sp.); WA - Polybia scutellaris 
Baccharis sp. indet. 1 4 wasp/fly MONO 7 oll (1 sp.) 
Baccharis sp. indet. 2 7 wasp/fly M veli BS - Schwarziana quadripunctata, Trigona spinipes; 
achinidae sp. indet. 4; WA - Mischocyttarus drewseni, 
ne billardier Polybia fastidiosuscula, Protopolybia 
sedula, AR pd is indet. 1, Pompilidae sp. indet. 3, 
Pompilidae sp. indet. 4, Tiphiidae sp. indet. 1, uncoll 
doe (2 
Baccharis sp. indet. 3 4 wasp/fly MONO WA - 1 lecheguana, Polistes billardieri 
Barrosoa betoniciiformis 4 several insect groups HOLO ud meli 1 ochloropsis cyanea; CO - Cantharidae sp. 
95 arbia cf. damippe, uncoll (1 
Campuloclinium megacephalum 11.5 several insect groups HOLO n 1 T elena quadrifasciata, pue hloropsis 


BL - 
I - Euprepina 


cyanea, A. iris, A. cf. cognata, uncoll n shlorini (1 sp.): 


— 


Bombus atratus; CO - Cantharidae sp. i 
sp. indet. 1, Paravilla sp. indet. 1, unc ol a (2 sp.): 
L 
myrinna, uncoll Vanessa (1 sp.), Hesperiidae (1 sp.) uncoll (1 
sp.); WA - uncoll Pompilidae (1 sp.), Sphecidae (1 sp.) 


e] 


- Sarbia cf. damippe, Thespeius sp. indet. 1, Vanessa 


900 


€ JOQUNN 'e6 euin|oA 


pue|sseJc) op- UH e ul uoneurog 


eullZeg Y Seol 


68v 


Table 4. Continued. 


Family/Species 


Time of observ. 


(hours) 


Pollination system 


Degree of specialization/ 


generalization 


Pollinator species 


Chaptalia integerrima 


Chaptalia runcinata var. 


graminifolia 


Chromolaena cf. decumbens 
Chromolaena xylorrhiz 


~ 
a 


Eremanthus erythropappus 


Erigeron maximus 
Eupatorium sp. indet. 


Eupatorium sp. indet. 


[em 


Eupatorium sp. inc 


Eupatorium sp. indet. 


Gochnatia paniculata 


et. « 


— 


J 


On 


uw 2 


several insect groups 


several insect groups 


bee nectar-flower 
undet. 


several insect groups 


several insect groups 
bee nectar-flower 


several insect groups 


bee nectar-flower 


several insect groups 


wasp/fly 


HOLO 


HOLO 


MONO 
n/a 


HOLO 


HOLO 


OLIGO 


HOLO 


OLIGO 
HOLO 


OLIGO 


BS - Paratrigona subnuda, Paroxystoglossa sp. indet. 1, uncoll 
Halictidae (2 spp.): CO - Cantharidae sp. indet. 1; LE 
Pyrrhopyginae sp. indet. 1, uncoll Lycaenidae (1 sp. i WA - 
Mischocyttarus drewseni. uncoll Eunemidae (I sp.). Vespidae 
(1 sp.) 


BS - A cyaned, O sp. indet. 2 


; DI - Paravilla 
. Diptera sp. indet. 3, uncoll 
Bombyliidae (1 Sp.): LE - Thecla sp. indet. 1, uncoll 
Hesperiidae (1 Sp.): SY - Toxomerus watsoni, Toxomerus sp. 
indet. I. Syrphidae sp. indet. 6 


sp. indet. I. Diptera sp. indet. 


BL - Bombus atratus 

not visited 

Apis mellifera; BS - Melipona bicolor. M. 55 
Paratrigona subnuda, Schwarziana qun | BL 
uncoll Bombus (1 Sp.): DI - Tachinidae sp. indet L E - uncoll 


7 - Gael P Sp.): WA - 


Hesperiidae (1 sp.). Vanessa (2 spp.):? 


uncoll Pompilid: ie (I sp.). Sphecidae (1 sp.) 
BS - Coelioxys sp. indet. 1, uncoll 1 hlorini (1 Sp.): CO - 
Cantharidae sp. ide 1: DI -Diptera sp. indet. 6 
BL - 1 iheringi, M. terrestris; LE - Vanessa myrinna, 
| Lycae DE (J sp.) 


11 1 1 ra: - Augoc ee sp. indet. 1; BL - uncoll 
Bombus (1 sp.): CO - Cantharidae sp. indet. 1; DI - Tachinidae 
sp. indet. 1: LE - uncoll Hesperiidae (2 spp.); WA - 
Hymenoptera sp. indet. 6, uncoll (1 sp.) 

BS - Melipona bicolor: LE - Vanessa sp. indet 

BS - Schwarziana quadripunctata, Augochloropsis cyanea; CO - 
Astylus sexmaculatus; DI - uncoll Bombyliidae (1 sp.). 
Tachinidae (1 Sp.): LE - Vanessa myrinna: SY - uncoll (1 sp.): 
WA - Mischocyttarus drewsent, 85 opacus 

SY - Toxomerus watsoni, uncoll (2 Sp.): WA - Mischocyttarus 

drewsent. Polistes billardieri. ud Pompilidae (1 sp 

Vespidae (2 spp.) 


uapied [eoiuejog unosst|A 


au) JO sjeuuy 


06v 


Table 4. Continued. 


Family/Species 


Time of obse rV. 


(hours 


Pollination system 


Degree of specialization/ 


generalization 


Pollinator species 


Grazielia gaudichaudeana 


Hypochaeris gardneri 
Lucilia lycopodioides 


Mikania lundiana 


Mikania nummularia 


Mikania sessilifolia 


Senecio oleosus 


Stevia myriadenia 


Symphyopappus compressus 


Vernonia herbacea 


m 


) 


several insect groups 


bee nectar-flower 
wasp 


bee nectar-flower 


wasp/fly 


wasp/fly 


syrphid/bee nectar- 


flower 


several insect groups 


several insect groups 


bee nectar-flower 


HOLO 


OLIGO 
MONO 
POLI 


OLIGO 


POLI 


POLI 


HOLO 


HOLO 


OLIGO 


Apis mellifera; BS - Melipona bicolor. Schwarziana 


ea Augochloropsis cyanea: CO - Cantharidae sp. 
ndet. 1, uncoll (1 sp.); DI - . dorsalis; LE - 


cui ui sp. indet. 1; WA - Hymenoptera sp. indet. 13, 


uncoll (1 sp.) 


Apis e BS - Colletes =P. indet. 1; BL - Centris klugi 


ate ra sp. indet. 2, unc Vespidae (3 


Apis mellifera; B 5 - Melipona bicolor, M. quaa es 


didi e subnuda, Schwarziana quadripunctata 
Pseudagapostemon cyaneus; LE - uncoll Vanessa 4 sp.); WA - 
uncoll Pompilidae (1 sp.) 

LE - Thecla sp. indet. 1, uncoll Hesperiidae (1 sp.); WA - 
Cercerini sp. indet. 2, Hymenoptera sp. indet. 1, Hymenoptera 
sp. indet. 2 

DI - Meada cf. corpulenta, Curtonotidae sp. indet. 1, uncoll 
Curtonotidae (1 sp.). Tachinidae (3 spp.); SY - 1 
watsoni, Syrphidae sp. indet. 5; WA - Polistes billardiert, 
Hymenoptera sp. indet. 3, uncoll Vespidae (2 spp. 

BS - Ceratina cf. asuncionis, Pseuda pi d cyaneus, 
Coelioxys c n . uncoll Halictini (1 sp.); BL - uncoll 
Bombus (I sp.); S in oxomerus watsoni, ancoll (2 spp.) 

Apis mellifera; BS - b ( Lophopedia) cf. pygmaea, P. 
(Trigonopedia) sp. indet. 1, Augochloropsis sp. indet. 1 
Epanthidium autumnale, Coelioxys sp. indet. 2, uncoll 
Augochlorini (1 sp.); BL - Bombus atratus, B. brasiliensis, 
uncoll Centris (1 pi DI - Paravilla sp. indet. 1, uncoll 
Bombyliidae (1 sp.); SY - Syrphidae sp. indet. 5, uncoll (1 sp.) 

Apis mellifera; BS - Melipona * Melissoptila 


T 


aureocincta, Agapostemon sp. . 1, Megac chile e 
anthidioides; BL - Pd atratus, 1 1 DI 
- Tachinidae sp. indet E - Yphthimoides ochracea, uncoll 


[om 


Nymphalidae (2 sp.), sim (1 sp.): WA - Scoliidae sp. indet. 
l, uncoll (1 sp.) 


BS - Ceratina cf. asuncionis; BL - uncoll Bombus 


900 


€ JequinN ‘g6 euinjoA 


pue|sseJc) op- e ui uoneuliiog 


BWIZES Y Seya 


165 


Table 4. Continued. 


Time of observ. Degree of specialization/ 
Familv/Species (hours) Pollination system generalization Pollinator species 

Vernonia megapotamica 12 several insect groups HOLO BS - Melipona bicolor. uncoll pans (2 spp.), Halictidae (3 
spp.): DI - Paravilla sp. indet. 1; LE - Agraulis vanillae; WA - 
uncoll Pompilidae (1 sp.), vd cd (2 spp.) 

Vernonia cf. rosea 4 bee nectar-flower MONO BL - Centris klugt. uncoll Bombus 

Vernonia tomentella 8 bee nectar-flower POLI BS - Melipona bicolor, Mesonychium caerulescens, Ceratina cf. 
asuncionis, Augochloropsis cf. cognata, A. cyanea, uncoll 
Halictini (1 sp.): BL - Bombus atratus, Centris klugi, Megachile 
iheringi, M. terrestris, uncoll Bombus; LE - Vanessa myrinna 

Vernonia tragiaefolia 4 bee nectar-flower OLIGO BS - Gaesisc je: nigra; BL - uncoll Bom! hus 

Vernonia westiniana 8 bee nectar-flower POLI Apis vites - Melipona bicolor. Plebeia saiqui, Ceratalictus 
sp. indet. 1, uncoll Meliponini (1 sp.). Halictidae A spp.); BL - 
Bombus atratus, B. brasiliensis, B. morio, uncoll Centris (2 
spp): DI - Tachinidae sp. indet. 6, uncoll Tac rinda (2 spp.): 
LE - Sarbia cf. xanthippe. Pyrrhopyginae sp. indet. 1, Thecla 
sp. indet. 1, Hesperocharis erota, uncoll o (1 Sp.). 
Nymphalidae (5 spp.), Pieridae (1 sp.). Sphingidae (1 Sp.): WA 
- Scoliidae sp. indet. 1, uncoll Pompilidae (1 sp.). Vespidae (1 
sp.) 

Bromeliaceae 
Dyckia tuberosa 10 hummingbird MONO HB - Colibri serrirostris 
Campanulaceae 

Lobelia camporum 7 bee nectar-flower OLIGO BS - Ceratina cf. asuncionis; BL - Bombus atratus, Centris 
burgdorfi, uncoll Bombus 

Wahlenbergia brasiliensis 15 syrphid/bee nectar- OLIGO Apis mellifera: BS - uncoll Augoc e i spp.): SY - Allograpta 

flower exotica, Toxomerus watsoni, uncoll (2 spp.) 
ethracea 
Clethra scabra 1.5 wasp/fly POLI Apis mellifera: BS - Melipona bicolor, Scaptotrigona bipunctata, 


Schwarziana quadripunctata, Augochloropsis cf. cognata, 
uncoll Meliponini (2 PP „ Halictidae (3 spp.): CO - 


Maura sp. indet. 2, Rhinotragus festivus: DI - Tachinidae 
sp. indet. 2, Diptera sp. indet. 7. uncoll Sare ud (l Sp.). 
Tachinidae (1 Sp.): LE - Yphthimoides ochracea; 


Mischocyttarus pul Synoeca cyanea, Polybia 
ba d qe P. minarum, Cercerini sp. indet. 1, Cercerini 


sp. indet. 2, Hymenoptera sp. indet. 10, uncoll (1 sp. 


C6 


ay} Jo sjeuuy 


uepJer) e,, og unossi|N 


Table 4. Continued. 


Family/Species 


Time of observ. 


(hours 


Pollination system 


Degree of specialization/ 


generalization 


Pollinator species 


Convolvulaceae 
Convolvulus crenatifolius 


Ipomoea procumbens 

Jacquemontia grandiflora 
Cunoniaceae 

Weinmannia organensis 


Droseraceae 


Drosera montana 
Ericaceae 
Agarista hispidula 


Gaylussacia chamissonis 


Gaylussacia jordanensis 


Eriocaulaceae 


Paepalanthus paulensis 


Paepalanthus polyanthus 


Erythroxylaceae 


Erythroxylum microphyllum 


syrphid/bee nectar- 


ower 


— 


vee nectar-flower 


bee nectar-flower 


several insect groups 
syrphid/bee pollen- 


hummingbird, SSP 


bee nectar-flower 


bee nectar-flow er 


several insect groups 


several insect groups 


wasp/fly 


OLIGO 


MONO 
POLI 


HOLO 


MONO 


MONO 
OLIGO 


POLI 


HOLO 


HOLO 


OLIGO 


Apis mellifera; BS - Ceratina sp. indet. 2, uncoll Ceratina (1 sp.); 


SY - uncoll (1 sp.) 
BL - Bombus atratus, Xylocopa brasilianorum, uncoll Bombus 
ini (1 sp.), Halictidae 


- Bombus atratus, uncoll Bombus, Centris (1 sp.): 


jon 


BS - uncoll Meliponini (1 sp.), Tapinotaspic 
(3 spp.); BL 


WA - unco 


| Vespidae (1 sp.) 
CO - Cantharidae sp. indet. 2; LE - uncoll (1 sp.); WA - Polybia 


minarum 


SY - Toxomerus watsoni, Syrphidae sp. indet. 1, uncoll (1 sp.) 


HB - Chlorostilbon aureoventris, Leucochloris albicollis 
P 8 0 BS - Melipona bicolor, Augochloropsis sp. indet. 1, 
oll Me d (I sp.: WA - uncoll Polybia (2 spp.) 
m mellife ra; ebeia saiqui, Augochloropsis cyanea 
Vni sp. a t. I. Ceratalictus sp. indet. J. Dalei 
ndet. 1, uncoll Meliponini (1 sp.). Halictidae (1 sp.); BL - 
Megachile ee WA - uncoll Vespidae (1 sp.) 


BS - Trigona da CO - Conognatha sp. ES: 1, Discodon 
2, Sciaridae sp. 
; LE - uncoll 
Hespe sio (J Sp.): SY - Toxomerus watsoni, Pescar sp. 
indet. l, uncoll (3 spp.): WA - 
CO - a. sp. indet. 1; DI — wna sp. indet. 5, Diptera 
sp. indet. 8, uncoll Bombyliidae (I sp.): LE - uncoll (1 sp.): SY 


- Toxomerus watsoni; WA - Agelaia vicina 


tucumanum; DI - Sarcophagidae sp. ue 
indet. 1, Tac hinidae sp. indet. 3, uncoll ( 


‘oll Vespidae (1 sp.) 


BS - Ceratalictus sp. indet. 1, uncoll Meliponini (1 sp.). 
Augochlorini (2 spp.): WA - Mischocyttarus drewseni, Polistes 
2, uncoll Pompilidae (1 


billardieri, Hymenoptera sp. indet. : 


sp.). Vespidae (2 spp.) 


900 


€ JequinN ‘g6 euin|oA 


pue|sseJc) pH e ur uogeutljog 


ewIzes Y Seal 


£6v 


Table 4. Continued. 


Time of observ. 


Degree of specialization/ 


Family/Species hours) Pollination system generalization Pollinator species 
Escalloniaceae 
Escallonia farinacea 6 bee nectar-flower MONO BS - Colletes sp. indet. 1 
Euphorbiaceae 
Croton dichrous 6 wasp/fly POLI Apis mellifera; BS - We cf. cognata, A. cyanea: DI - 
oll Tachinidae (2 spp.): - Toxomerus watsoni: WA - 
Mischocyttarus drewseni, 1 scutellaris. Gasteruptiinae sp. 
indet. 1, uncoll Vespidae (2 spp.) 
Gentianaceae 
Calolisianthus pedunculatus 19.5 undet., SSP n/a not visited 
Calolisianthus pendulus 38.5 undet., SSP n/a not visited 
Deianira nervosa 6 deal /bee pollen- MONO SY - Syrphidae sp. indet. 5 
flower, SSP 
Helia oblongifolia 20 N undet.. E n/a not visited 
Zygostigma australe 8.5 syrphid/bee pollen- OLIGO BS - Ceratalictus sp. indet. 1: SY - Allograpta exotica, uncoll (1 
flower, SSP sp.) 
Gesneriaceae 
Sinningia allagophylla 21 hummingbird MONO HB - Clytolaema rubricauda. Colibri serrirostris, Leucochloris 
albicollis. Stephanoxis lalandi 
Hypericaceae 
Hypericum brasiliense 8 syrphid/bee pollen- POLI BS - Schwarziana quadripunctata, Halictini sp. indet. 1: BL 
flower, apomictic Megachile iheringi: DI - Cylindromyia dorsalis: SY - uncoll (1 
sp.) 
Hypericum ternum - uncertain, apomictic n/a n/a 
Iridaceae 
Alophia geniculata 11 bee pollen-flower POLI BS - Melipona bicolor. Plebeia saiqui. Augochloropsts tris, 
Paroxystoglossa cf. jocasta, Dialictus sp. indet. 1, Halictini sp. 
indet. 2, uncoll Tapinotaspidini (1 Sp.): BL - Bombus atratus: 
SY - Toxomerus watsoni, Syrphidae sp. indet. 2 
Alophia sp. indet. 1 7 uncertain n/a BS - Ceratalictus sp. indet. 1: SY - 1 exotica 
Calydorea campestris 8.5 svrphid/bee pollen- OLIGO BS - Plebeia saiqui. Augochloropsis of « cognata, Dialictus sp. 
flower indet. 1: SY - Syrphidae sp. indet. 5. uncoll (2 spp.) 
Sisyrinchium micranthum 5 OLIGO BS - Dialictus sp. idet 1; SY - poe exotica 


syrphid/bee pollen- 
flower 


dope jeoiuejog unossi|A 


6 


ou Jo sjeuuy 


Table 4. Continued. 


Family/Species 


Time of observ. 


(hours) 


Pollination system 


Degree of „ 


generalizatio 


Pollinator species 


Sisyrinchium vaginatum 


Lami 
b oe 


Hyptis plectranthoides 
Hyptis umbrosa 


Peltodon radicans 
Leguminosae 
Chamaecrista sp. indet. 
Crotalaria breviflora 
Lupinus velutinus 


Lythraceae 
Cuphea glutinosa 


Malpighiaceae 


Byrsonima variabilis 


Malvaceae 
Pavonia kleinii 
Sida sp. indet. 1 

Melastomataceae 


Leandra erostrata 


21.5 


syrphid/bee pollen- 


flower 


bee nectar-flower 


bee nectar-flower 

bee nectar-flower 

bee nectar-flower 

bee pollen-flower 

bee nectar-flower 

bee pollen-flower, 
SSP 


bee nectar-flower 


bee oil-flower 


bee nectar-flower 


bee nectar-flower 


uncertain, 


apomictic? 


OLIGO 


OLIGO 


OLIGO 


OLIGO 


MONO 


MONO 


OLIGO 
MONO 


POLI 


OLIGO 


MONO 
MONO 


n/a 


Apis mellifera; BS - uncol 


E 


Augochlorini (2 spp.): SY - Toxomerus 


watsoni, Toxomerus sp. indet. 1, uncoll (1 sp.) 


Apis mellifera; BS - Melipona bicolor, Plebeia saiqui, Trigona 


spinipes, Augochloropsis cf. cognata, Ceratalictus sp. indet. 1: 


ombus atratus 


Apis mellifera: BS - Ceratalictus sp. indet. 1; BL - Bombus atratus, 


uncoll Bombus 
Apis mellifera; BS - Augochloropsis sp. indet. 1, Ceratalictus sp. 


indet. 1: WA - uncoll Vespidae (2 spp. 


Apis mellifera 


BL - Centris klugi 

BS - Anthidium sertanicola; BL - Bombus atratus, B. brasiliensis 

BL - Megachile il . M. laeta, Bombus atratus, uncoll 
Bombus, 1 d spp.) 


Apis mellifera; BS - Ceratina sp. indet. 1, Mesonychium 


caerulescens, Augochloropsis cyanea, m b sp. inde 

Anthidium „ Megachile cf. f. anthidi ioides 
Bombus atratus, B. brasiliensis, B. » Centris Hugi: LE - 
Eurema nise tenella; ne - 5 watsoni 


Hymenoptera sp. indet. 3, Hymenoptera sp. a 11 


BS - Plebeia saiqui, Augochloropsis cf. cognata, Monoeca sp. 


indet. 1, Paratetrapedia ( Xanthopedia) sp. indet. 1, Centris ef. 


insularis; BL - Centris discolor. C. klugi. C. tarsata, Centris 


(Melacentris) sp. indet. 1, uncoll Centris (2 spp.) 


BS - Augochloropsis cf. cognata 


BS - Ceratina cf. asuncionis, uncoll Ceratina (1 sp.) 


not visited 


9005 


€ JequinN ‘£6 euinioA 


pue|ssec) epniniv-ufirg e ul uoneuillog 


eBWIZeS Y Seya 


965 


Table 4. Continued. 


Time of observ. 


— 


Degree of specialization/ 


Family/Species (hours) Pollination system generalization Pollinator species 
Leandra sp. indet. | 6 uncertain, n/a not visited 
apomictic? 
Leandra sp. indet. 2 11 uncertain n/a not visited 
apomictic? 
Leandra sp. indet. 3 6.5 uncertain n/a not visited 
apomictic? 
Microlicia isophylla 14.5 uncertain, n/a not visited 
apomictic? 
Tibouchina frigidula 24 bee pollen-flower MONO BL - Bombus atratus, X) 1 8 brasilianorum, Centris discolor. 
uncoll Bombus. Centr spp.) 
Tibouchina martialis 4 bee pollen-flower MONO BL - Xylocopa edi uncoll Centris (1 sp.) 
Tibouchina minor 10 bee pollen-flower OLIGO BS - uncoll Augochlorini (1 sp.): BL - Bombus atratus, Centris 
klugi, uncoll Centris (2 spp.) 
Trembleya parviflora 6 bee pollen-flower MONO BS - Melipona bicolor 
Trembleya phlogiformis 7 bee pollen-flower OLIGO BS - Augochloropsis sp. indet. 1: BL - Bombus atratus 
Ochnaceae 
Ouratea semiserrata 7 bee pollen-flower OLIGO BS - Melipona bicolor, Dialictus sp. indet. 1: BL - uncoll Centris (1 
Orchidaceae 
Epidendrum secundum 14 undet. n/a not visited 
Habenaria parviflora 5. 13 N undet. n/a not visited 
Oncidium barbaceniae 19.5 uncertain n/a uncoll Bombus 
Oneidium blanchetii 14 undet. n/a not visite 
Oncidium sp. indet. 1 14 undet. n/a not visited 
Orobanchaceae 
Esterhazya macrodonta 72.5 hummingbird MONO HB - Leucochloris albicollis 
Polygalace: 
Polygala ee 6.5 bee nectar-flower, MONO BS - uncoll Augochlorini (1 sp.) 
Polygala cneorum 10 bee nectar-flower. MONO BS - Augochloropsis sp. indet. 1 


uapseyd jeoiuejog unossilA 


96v 


ay} Jo s¡euuy 


Table 4. Continued. 


Family/Species 


Time of observ. 


(hours) 


Pollination system 


Degree of specialization/ 


gene sralization 


Pollinator species 


Rubiaceae 


Borreria capitata 


Borreria tenella 


Declieuxia cordigera var. 


angustifolia 


Galianthe angustifolia 


Galianthe brasiliensis 


Galium hypocarpium 
Solanaceae 

Solanum aculeatissimum 

Solanum americanum 

Solanum pseudocapsicum 

Solanum swartzianum 


Solanum viarum 


8.5 


several insect groups 


wasp/fly 


bee nectar-flower 


several insect groups 


several insect groups 


wasp/fly 


bee pollen-flower 
bee pollen-flower 
bee pollen-flower 
bee pollen-flower 


bee pollen-flower 


HOLO 


OLIGO 


POLI 


HOLO 


HOLO 


OLIGO 


MONO 
MONO 
MONO 
MONO 
MONO 


Apis mellifera; BS - Augochloropsis cf. cognata, Ceratalictus sp. 


indet. 1; DI - i cd m indet. 1, uncoll Sarcophagidae (1 
Sp.): SY - a is, Toxomerus watsoni, Toxomerus 
et. 2, 


sp. indet. 1, uncoll (1 sp.); WA - Hymenoptera sp. inc 
uncoll Vespidae (1 sp.) 

BS - Ceratalictus sp. indet. 1; WA - Polybia sericea, Hymenoptera 
sp. indet. 2, uncoll Vespidae (2 s 

BS - Augochlorini sp. indet. 1, uncoll Augochlorini (1 sp.): BL - 
Bombus atratus, Centris M uncoll 9 0 LE - uncoll 
Hesperiidae (1 sp.); WA - uncoll Vespidae sp.) 


Apis mellifera; BS - 1 caerulescens, p hloropsis 


cyanea, uncoll Augochlorini (2 spp.): BL - Bombus atratus, B. 
indet. 2, 


Sarcophagidae sp. indet. 1, Tachinidae sp. indet. 7; LE - 


morio, uncoll Centris (2 T „.): DI - Curtonotidae s] 


uncoll od (l Sp.): SY - Allograpta exotica, uncoll (2 
spp.) WA - Bicyrtes paranae, Sphex dorsalis, Polybia 
1 Hymenoptera sp. indet. 15, uncoll Vespidae (2 
spp.) 

Apis mellifera; BS - Augochloropsis e cognata, Augochloropsis 
cyanea, Augochloropsis sp. indet. 1, Ceratalictus : indet. 1; 
DI - Tachinidae sp. indet. 1, Tac 11 155 sp. indet. 4; L 
Thecla sp. indet. 1; SY - Pseudodoros clavatus, 1 
watsoni, uncoll (1 sp.): WA - Mischocyttarus drewseni, Polistes 
ed Tenthredinidae sp. indet. 1, Hymenoptera sp. indet. 

uncoll Pompilidae (1 sp.), Sphecidae (1 sp.), Vespidae (3 
55 


DI - Trupanea sp. indet. 1, WA - Eurytomidae sp. indet. ! 


BS - Augochloropsis cyanea 

BS - uncoll Augochlorini (1 sp.) 

BS - Augochloropsis cyanea, Augochloropsis sp. indet. 1 
BS - Augochloropsis cyanea 


BS - Augochloropsis cyanea 


9005 


€ JequinN 'ee asunjoA 


pue|sseJc) pH UH e ul uoneurpog 


PuIZeS 9 Sealy 


465 


Table J. Continued. 


Time of observ. Degree of specialization/ 
Family/Species (hours) Pollination system generalization Pollinator species 
Verbenaceae 
Verbena hirta 10 bee nectar-flower POLI Apis mellifera: BS - Mesonychium caerulescens, Ceratina cf. 
asuncionis, P sci lora cf. graminea, Halictini sp. indet. 
| ombus atratus. B. brasiliensis, B. morio. Xylocopa 
1 Bois tarsata, Megachile laeta; LE - Urbanus 
sp. indet. 1, Eurema nise, uncoll Vanessa (1 Sp.): SY - uncoll (2 
spp.): WA - 1 Vespidae (1 sp.) 
Violaceae 
Viola cerasifolia 21 bee pollen-flower MONO BS - Anthrenoides sp. a l. Augochlorodes turrifaciens. 
Dialictus sp. indet. 
Xvridaceae 
Xyris asperula | syrphid/bee pollen- OLIGO BS - Augochloropsis cyanea, uncoll Meliponini (1 sp. 
flower Augochlorini (1 sp.): SY - Toxomerus watsoni, 9 Sp. 
indet. 1. uncoll (2 spp.) 
Xyris tortulla 6 syrphid/bee pollen- OLIGO BS - Paratetrapedia cf. pygmaea: SY - Toxomerus watsoni, 
flower Svrphidae sp. indet. 4, uncoll (2 spp.) 


86r 


eui Jo SE 


uepJe*) jeoiuejog unossi|A 


Volume 93, Number 3 Freitas & Sazima 499 


2006 Pollination in a High-Altitude Grassland 


Table 5. Number of plant species exclusively, mainly, and secondarily pollinated by each pollinator group at the Serra da 


Bocaina grasslands. Percentages are given in parentheses (n = 106 plant species). Because many species are pollinated by 


two to six pollinator groups the sum of the totals exceeds 100%. 


Role in pollination 


Pollinator group Exclusive Main* Secondary or indistinct Total 
Small bees 8 12 (11.3%) 21 (19.8%) 38 (35.9%) 71 (67.0%) 
Wasps 7 (6.6%) 15 (14.29) 26 (24.5%) 18 (45.3%) 
Large bees 8 8 (7.6%) 18 (17.0%) 10 (9.466) 36 (34.0%) 
Syrphids 3 (2.8%) 12 (11.3%) 18 (17.0%) 33 (31.196) 
Other flies 0 ) (2.896) 26 (24.5%) 29 (27.3%) 
Butterflies 0 2 (1.966) 23 (21.796) 25 (23.0%) 
Beetles 0 0 13 (12.3%) 13 (12.3%) 
Hummingbirds 5 (4.7%) 0 0 5 (4.7%) 


* Includes plant species that have multiple pollinator groups but only one or two are the main pollinator group(s) (see 


methods section of text). 
8 Apis mellifera not included. 


B. POLLINATOR AGENTS 


A wide array of pollinator agents were recorded 
(Tables 3, 4). Hymenopterans, followed by dipterans. 
were the most important pollinators. and all insect 
eroups except large bees were found more often as 
secondary agents than as exclusive or main pollination 
agents (Table 5). Hummingbirds, which were the only 
vertebrate acted as exclusive 
(Table 5). 


Hymenoptera. 


group, pollinators 
Bees and/or wasps were among the 
pollinator agents of 98 of the 101 plant species 
pollinated by insects. In addition, hymenopterans 
were the pollinators of three-fourths of the plants 
pollinated by one group only (Table 5). We recorded 
at least 54 native species of bees (Table 3). The 
families Apidae (including “Anthophoridae”), Halic- 
tidae, and Megachilidae were represented by 28, 16, 
and 8 species, respectively. Andrenidae and Colleti- 
dae were only represented by one species each. 
Halictidae bees—owing mainly to  Augochlorini 
species—pollinate about half of the species in this 
community. Meliponini bees played a minor role in 
pollination, occasionally pollinating some generalist 
plant species, and acted more as pollen thieves in 
several species (e.g.. Baccharis and Melastomataceae 
species, Fig. 11). However, Melipona bicolor—which 
is markedly larger than other Meliponini bees in that 


was an important pollinator of several species 


area 

(Fig. 1B, Table 3). Small and large bees belonging to 

Megachilidae pollinated 12 species from six families. 
8 p | 


o the 


ex 


The large Apidae bees belonging genera 
Bombus, Xylocopa Latr., and Centris Fabr. also were 
important pollinators in the community. The three 
bumblebee species—mainly Bombus atratus—polli- 


nated at least 29 species. and they were either the 


most frequent or the exclusive pollinators of 12 of 


them. We recorded at least 46 species of wasps acting 
as pollinators, the most important of which were the 
social Vespidae, in particular Polistes Latr. species, 
followed by Pompilidae and Sphecidae. 

The introduced honeybee, Apis mellifera, acted as 
a pollen robber in Sinningia allagophylla but 
pollinated the other 31 species they visited (Table 3). 
These bees were frequent visitors to some species 
classified as having bee nectar-flower pollination 
systems (i.e., pollinated by bees in search of nectar: 
see “Pollination Systems” below), such as /lex amara, 
Gaylussacia chamissonis, G. jordanensis, and Cuphea 
glutinosa. Honeybees were the most important 
pollinators of the three Hyptis Jacq. species (Lamia- 
ceae) and the only flower visitor for the other 
Lamiaceae, Peltodon radicans. 

Diptera. Syrphids constituted a noticeable group 


of pollinators in the community, feeding on both 


nectar- and pollen-flowers. Species of the genus 
Toxomerus Macq. were particularly important and. 
for instance, T. watsoni acted as a pollinator of 21 
species (Table 3). Flies other than syrphids were 
represented mainly by Tachinidae and Bombyliidae 
(beeflies) and, usually, were secondary or indistinct 
pollinators of polyphilous or holophilous species 
(Fig. 2B, Tables 5, 6). 

Lepidoptera. Pollination by lepidopterans (Fig. 21) 
was poorly represented. Butterflies (“Rhopalocera”) 
and diurnal moths (“Heterocera”) generally played 
a secondary role for plants pollinated mainly by bees 
(Tables 4, 5, 6). They were among the most frequent 
pollinators of only a few Asteraceae species. Butter- 
flies belonging to the genus Vanessa Fabr. (Nympha- 
linae) were the most frequent flower visitors of this 
group. In spite of the scarcity of their visits to flowers. 
many butterflies were recorded during the summer 


and fall, and small butterflies belonging to Hesper- 


500 Annals of the 
Missouri Botanical Garden 


Figure l. Flowers and pollinators at the Serra da Bocaina grasslands. XC. Bee nectar-flowers. —A. Augochloropsis alf. 
cognata (Halictidae) visiting a flower of Paonia kleinii (Malvaceae e). —KB. A stingless bee Melipona bicolor (Apidae) entering 
a bell-shaped flower of 5 chamissonis (Ericaceae). —C. Bombus brasiliensis (Apidae) visiting a head of Vernonia 
ni ly f —D. Polistes mne (Vespidae) visiting a he id of Gochnatia paniculata 
(Asteraceae). which is xls wells ited bs y syed —E. Polybia sericea (Vespidae) visiting a flower of Gonioanthela hilariana 
(Apocynaceae). F-J. Bee pollen-flos 


m 
z 
= 
Ew 
= 
Ex 
— 
Y 
Y 
— 
= 
r 


\ flower of Tibouchina minor (Melastomataceae). Note the long poricidal 
anthers. —6. Augoc 17 sp. indol a ‘cling pollen by vibration from a single anther of a Trembleya phlogiformis 
(Melastomataceae) ) flowe s This bee e occasionally touches the stigma (arrow). —H. Be ed atratus collecting pollen by 
vibration in Trembleya phlogiformis. which is mainly pollinated by this bee. —I. Paratrigona subnuda (Apidae) collecting 
pollen from a flower of Trembleya parviflora. This bee removes pollen from the anther by inserting its proboscis and acts as 
pollen thief since it does not transfer pollen to the stigma of this species, which is pollinated by another stingless bee, 
Melipona bicolor. —J. Megachile iheringi (ogi mua collecting ee n un a ke el-shaped flower of ane velutinus 
(Leguminosae). Note the bee abdomen touching the keel petals (arrow). K, L. Syrphid/small bee pollen-flowei K and L. 

Syrphids species feeding directly on pollen from flowers of Xyris 12 me iceae) and Calydorea 1 (Iridaceae). 
respectively, See Table 2 for flower measurements, 


Volume 93, Number 3 Freitas & Sazima 501 
2006 Pollination in a High-Altitude Grassland 


Figure 2. Flowers and pollinators in the Serra da Bocaina grasslands. Flowers pollinated by several insect groups: —A-C. 
Species ‘of three FP groups visiting the heads of Eupatorium sp. indet. 4 (Asteraceae), respective El a wasp (Sphex 
opacus, Sphecidae). a efly (Bombyliidae), and a small bee (Augochloropsis cyanea, Halictidae). Oil-flower: —D. 
Paratetrapedia sp. indet. collec ting pollen from flowers of Byrsonima variabilis (Malpighiaceae). Note the flower oil-glands 
or elaiophores (arrow). —E. Centris cf. insularis (Apidae) collecting oil from flowers of Byrsonima variabilis (Malpighiaceae). 
Syrphid/small bee nectar-flowers: —F. Ceratina sp. indet. 2 (Apidae) visiting a flower of Convolvulus crematifolius 
(Convolvulaceae). —G. Pseudagapostemon cyaneus (Halic tidac) 1 a head of Senecio oleosus (Asteraceae). Minor 
pollinator a —H. A beetle (Cantharidae sp. indet. 2) visiting a flower of the mainly wasp-pollinated Clethra : -— 
(Clethraecae). A butterfly, Thecla sp. indet A | (Lycaenidae), visiting the generalist flowers of Galianthe brasiliensis 
(Rubiaceae). ^r ul pollination systems: —J. A female Monoeca sp. indet. 1 (Apidae) with a pollinarium of a species of 


Oncidium (Orchidaceae) attached to its clypeus. This digger bee was captured while collecting oil on flowers of roni 


variabilis, and its role in the pollination of Oncidium species is uncertain. — K. Flower of Helia A odi (Gentianace: 
Floral traits point to pollination by nocturnal moths, but no pollinators were observed. —L. Flower of 1 
pedunculatus (Gentianaceae). The attractive flowers of this species seem to be 55 ipled to pollination bn o da but 


they are nectarless and spontaneously self-pollinated. See Table 2 for flower measurements. 


Annals of the 
Missouri Botanical Garden 


Table 6. The 
solvphilous, and holophilous (most generalized): within 
polyphil 1 holophil (most g lized) will 


Bocaina grasslands. 


number (and percentage) of plant species that are 


monophilous (most specialized), oligophilous. 


each pollination system and pollinator group at the Serra da 


Monophily Oligophily Polyphily Holophily Total of species? 
All species 36 (34.066) 32 (30.2%) 18 (17.0%) 20 (18.8%) 106 
Pollination system 
Small and large bee 10 (33.3%) 12 (40.0%) 8 (20.7%) 0 30 
nectar-flowers* 
Wasp and wasp/fly 7 (30.4%) 8 (31.8%) 7 (31.8%) 0 22 
nectar-flowers 
Several insect 0 0 0 ) (100%) 20 
groups 
Small and large bee 11 (13.3%) 3 (20.0%) 1 (6.766) Ü I5 
pollen-flowers 
Syrphid and 2 (22.2%) 6 (66.1%) 1 (11.19) 0 9 
syrphid/small 
bee pollen- 
flowers 
Hummingbird 5 (100%) 0 0 0 5 
flowers 
Syrphid/bee nectar- 0 2 (66.1%) ] (33.366) 0 3 
flowers 
Small and large bee 0 | (10066) 0 0 | 
oil-flow« 1 
Svrphid nectar- 1 (100%) 0 0 0 | 
flowers 
Undetermined or — — — — 18 
uncertain 
renee groupss 
all bees $ 12 (16.966) 25 (35.2%) 16 (22.5%) 18 (25.4%) 71 
Mns 7 (14.696) 11 (22.9%) 14 (29,2%) 16 (33.396) 18 
Large bees E 8 (22.2%) 12 (33.3%) 10 (27.8%) 6 (16.7%) 30 
Syrphids 3 (9.1%) 10 (30.3%) 9 (27.3%) 11 (33.39%) 33 
Other flies 0 3 (10.3%) 9 (31.1%) 17 (58.6% 29 
Butterflies 0 3 (12.096) 6 (24.0%) 16 (04.0%) 25 
Beetles 0 0 2 (15.49) 11 (84.6%) 13 
Hummingbirds 5 (100%) 0 0 0 5 


Does not include species whet 


* Includes Gaylussacia 1 and the specic 


Since man 
t Apis 


y spe 
mellifera not included. 


inae (probably species of Corticea Evans) were very 
common in all seasons at the study sites. 

Coleoptera. Beetles pollinated a few species and 
always acted as secondary or indistinct. pollinators 
(Tables 5, 6). ( 
Cantharidae sp. indet. 
(Table 3 


importance in pollination at the community level. 


zantharidae (Fig. 2H), most particularly 
l. which pollinated 10 species 
) was the only beetle family with a notable 


Vertebrata. The interactions between humming- 
bird-pollinated plants and their agents were the most 
specialized at the community level. We recorded five 
Trochilinae study 


Table 3 


in the high- 


species in the areas (Fig. 3, 


). These species could be observed year-round 


altitude forest and grasslands, excepl 


e the (RA system was nol determined or was mn 


amiaceae pollinated exclusive Des or mainly by ipis mellifera. 
JO 


are pollinated by more ke one pice group, the sum al of plant species exceeds It 


Colibri serrirostris, which seems to lo the 


migrate 


highlands during the wet season. 


POLLINATION SYSTEMS 


= 


The nine pollination systems differed markedly 


RE 


with regard to their number of species. For instance, 
the three most important systems, respectively, small 
and large bee nectar-flowers, wasp and wasp/fly 
nectar-flowers, and several insect groups, encompass 
72 species (08% of the 106 species total: Table 6). On 
the other hand, two of the systems, respectively small 
and large bee oil-flowers and syrphid nectar-flowers, 


only comprise one species each (Tables 4. 6). 


Volume 93, Number 3 Freitas & Sazima 503 
2006 Pollination in a High-Altitude Grassland 
Figure 3. Hummingbirds and their flowers in the Serra da Bocaina grasslands. —A. Leucochloris albicollis visiting a very 


large flower of Hippeas drum glaucescens 0 
female Colibri serrirostris vi 
F lowe r of Esterhazya 1 nta (Orobanchaceae), which 
siting flowers of Sinningia allagophylla 
hl e: ricaceae); originally published by Plant 
ermission. : Table 2 for flower measurements 
A. NECTAR-FLOWERS POLLINATED MAINLY BY SMALL AND 
LARGE BEES 


Plants of this group had flowers pollinated mainly 
by large or small bees that were searching for nectar 
(Fig. LA-C 


floral traits, and we observed specific associations 


. This group shows great variability in 
with pollinators among its species. As a whole, this 
group is the most representative pollination system of 
the community (Table 6). 

five 


This includes 


plant group first species 
pollinated exclusively by small bees. Species of 


Malvaceae (Fig. 1A) and Polygalaceae were pollinat- 
ed by Ceratinini or Halictidae bees, and Escallonia 
farinacea was B s only by Colletes sp. indet. 

(C olle tidae) (Table 4 


spec ies 


An important sub-group com- 


saat pollinated by small 


bees, in which wasps either were rare—Mikania 


lundiana, Hyptis umbrosa, and Gaylussacia jordanen- 
sis—or acted as secondary agents—/lex amara and 


(Fig. 1B). The 


species have white, bell-shaped flowers. 


Gaylussacia chamissonis last three 


Large bees were the exclusive pollinators of 
the large flowers of Ipomoea procumbens and Oxype- 


talum sublanatum. and also, of two Asteraceae 


e). Note the Tu oot gripping 
ting flowers of pum oe (Bromeliaceae 
is pon ited by L albicollis. i. 
(Gesneriaceae). —E. A female Chlorostilbon aureoventris visiting 


Systematics has Evolution. 


the inferior tepal. —B. 
ote pollen on the bird's bill and forehead. — 
A female Stephanoxis Jalón 
flowers of Agarista 


2006. Springer-Verlag; 5 with 


species, Chromolaena cf. decumbens and Vernonia 


cf. rosea. Likewise, Lobelia camporum. and Hyptis 


lippioides were mainly pollinated by large bees. 
although small bees acted as their secondary 
pollinators. Despite their typically melittophilous 


flowers (sensu Faegri & van der Pijl, 1979), Cuphea 
glutinosa and Verbena hirta were each pollinated by 
five different agents, of which large bees were by far 
the most frequent ones: for C. glutinosa, the pollen 
load deposited on the stigmas in a single large bee 
visit was significantly higher than that deposited by 
small bees, wasps, or syrphids (L. Freitas, unpubl. 


data). Both C. 


uously (Table 1) and are among the most common 


glutinosa and V. hirta bloom contin- 
plants in the study area. 


Plant species pollinated by large bees produced 


larger quantities of nectar than the other insect- 
pollinated species, which typically secreted less than 


2 ul of nectar per flower per day. Oxypetalum 
sublanatum and Estela camporum. flowers, for in- 


8 ul (n = 10) and 4—7 ul (n = 


respectively. 


stance, secreted 5 
17) of nectar per day. 
Another sub-group in this system is composed 


of species in which both large and small bees 


seemed to be the main pollinators—Hypochaeris 


Annals of the 
Missouri Botanical Garden 


gardneri, Vernonia herbacea, V. tragiaefolia, Jacque- 
montia grandiflora, Hyptis plectranthoides, Crotalaria 
breviflora, and 


lion systems of some species of Asteraceae diverge 


Declieuxia cordigera. The pollina- 
slightly from this last sub-group, because, in addition 
to small or large bees as their main pollinators, 


butterflies were either important agents—Aupatorium 


sp. indet. 3 and Vernonia westiniana (Fig. 1C)—or 
secondary agents—Kupatorium sp. indet. | and V. 
tomentella. 

B. NECTAR-FLOWEIRS POLLINATED EITHER BY WASPS OR BY 


WASPS AND DIPTERA 


This group includes species from several families. 
such as Gonioanthela hilariana (Fig. 1E), Oxypetalum 
appendiculatum, Erythroxylum microphyllum, Croton 
dichrous, Clethra (Fig. 2H). 
tenella, for either 


scabra and Borreria 
the exclusive 


pollinator or the main pollinator, usually associated 


which wasps were 
with flies. Wasps were also either the main « 
of 


satureioides 


the 


Asteraceae. e. 


n 
exclusive pollinators many m 
Ich yrocline Gochnatia 
(Fig. 1D), Lucilia lycopodioides, 


mularia, 


paniculata 
and Mikania num- 
and of all species of the genus Baccharis 
tarchonanthoides and B. dracunculifolia. 


as well as l 


n Mikania sessilifolia am 


the spontaneously self-pollinated Galium hypocar- 
pium, pollination by flies and wasps was equally 
important. 

Flowers of this group typically have easily 
accessible nectar and are usually greenish. Pre- 
dominant flower shapes are short tube (e.g... Mikania 
Willd.) and dish (e.g... Gonioanthela Malme and 


Erythroxylum P. Browne), but flowers may be large 
in width, allowing the insect to enter and reach the 
neclar Wasps 


and some 


(e.g... Oxypetalum appendiculatum). 


carry few pollen grains on their bodies. 
species belonging to this pollination system may be 


additionally pollinated by wind. 


C. POLLEN-FLOWERS POLLINATED BY SMALL AND LARGE BEES 


Nectarless flowers with poricidal anthers are the 


most characteristic which includes 


species of Melastomataceae and Solanaceae, among 


of this system, 
others (Fig. l E-I). Plants of this group were pollinated 
by large and small bees, which collected pollen by 
vibration. Large Apidae bees—belonging to the genera 
Bombus, Xylocopa, and Centris—were the exclusive 
pollinators of Tibouchina frigidula, 


Chamaecrista sp. indet. 


T. martialis, and 
Large bees were also the 
pollinators of Tibouchina minor (Fig. IH). Trembleya 
phlogiformis (Fig. 1G—H). and Ouratea semiserrata, i 


addition to such small bees as Melipona bicolor and 


some Halictidae species. The remaining plant species 
with poricidal anthers—Trembleya parviflora (Vig. 11) 
and the five Solanum L. 


species—have relatively small 


lowers, which were only pollinated by small bees. Four 


species of Solanum were exclusively pollinated by the 
same small bee—Augochloropsis cyanea (Halictidae) 
This bee (Fig. 2C) is endemic to the Brazilian coastal 
range and is markedly larger (mean body length 10.1 + 
0.99 mm, n = 


study area. 


10) than other Halictidae species in the 


Bees searching for pollen also pollinated some 


flowers without poricidal anthers. Viola cerasifolta 


flowers have five anthers opening rupi bul 


their androecium is arranged so as to function. as 


es 


single poricidal anther. This eie was mainly 
pollinated by small Anthrenoides Ducke bees (Andre- 
nidae), which showed a singular pollen collecting 
behavior on these flowers, in addition to vibration ( 
Freitas & Sazima. 2003a for details). 


The nectarless flag-shaped flowers. « 


See 


Hi 
velutinus were only pollinated by large bees belonging 
to the genera Bombus and Megachile Latr. lJ). 
During the the the bee 
the 


(Fig. 


bumblebee visits. weight of 


flexed wing petals downward. and pollen was 
therefore pressed out in portions, at the tip of the keel 
“macaroni pump? 
1994). which contacted the abdomen of the 


bee. However, pollen collection by bumblebees was 


type of pollen presentation. sensu 
Endress. 


usually improved by vibration. Leafcutter bees seem 


to lack the necessary weight to flex the wings and 
press pollen out of the keel. 


a flower, 


During their visits to 
these bees used their head as a lever, and the 
wines were thus depressed 


e. IJ). 


7 


— 


y their front and middle 
The hind legs stroked the sides of the 
eak. 
and pollen was then packed on the ventral abdomen of 
the bee 19 


Bombus atratus, collecting pollen by vibration, was 


legs (Fi 


keel, forcing a stream of pollen out of the keel 


72 
140). 


(“milking action.” sensu Wainwright, 
the effective pollinator of Alophia geniculata flowers, 
although its visits were infrequent. Small bees and 
syrphids frequently fed on pollen of this species, but 
\ 
similar pollination system is expected for Alophia sp. 
indet. 


they only occasionally pollinated its large flowers. 


l, although we were unable to record any 
bees on its flowers (Table 4) 


arge 


D. POLLEN-FLOWERS POLLINATED EITHER BY SYRPHIDS OR BY 
SMALL BEES AND SYRPHIDS 

Syrphids were the exclusive pollinators of Drosera 
montana and Deianira nervosa and the co-pollinators. 
in association wilh small bees, of Zygostigma australe, 
Calydorea campestris, Sisyrinchium vaginatum. 
asperula, and Xyris tortulla (Vig. IK-L). Species in 


this group typically bear small, dish-shaped, actino- 


Xyris 


Volume 93, Number 3 
2006 


Freitas & Sazima 
Pollination in a High-Altitude Grassland 


morphic, and vividly colored pollen-flowers. Pollen 


grains are easily accessible pollinators of these 
species because their anthers open longitudinally. 
Flowers of Sisyrinchium micranthum bear oil-secret- 
ing trichomes at the base of the staminal column (see 
& Vogel, 2001, | 


failed to detect any oil-collecting bees 


Cocucci Truylio et al., 2002). 
However, we 
on these flowers, which were pollinated by syrphids 


and small pollen-collecting bees. 


E. NECTAR-FLOWERS POLLINATED BY SEVERAL INSECT 


GROUPS (SIG) 


This group contains species pollinated by three or 


more functional groups with similar importance 


(holophily). The SIG system includes some species 
flowers—Eryngium horri- 


with very small, greenish 


dum (and probably E. canaliculatum), Paepalanthus 
paulensis, P. polyanthus, and Weinmannia organen- 
sis—that fit the “diverse small insects” (d.s.i.) 
oo of Bawa et al. (1985). 

e SIG system also includes species from several 
1 genera (Baccharis tarchonanthoides, Barro- 
soa betonictiformis, Campuloclinium megacephalum. 
Chaptalia integerrima, C. runcinata, Eremanthus ery- 
thropappus, Erigeron maximus, Eupatorium sp. indet. 2 


ig. 2A-C]. Grazielia 2 
Symphyopappus com- 


Eupatorium sp. indet. 


chaudeana, Stevia myriadenia, 


pressus, and Vernonia megapotamica) and three Rubia- 
ceae species (Borreria capitata, Galianthe angustifolia. 
such as 


and G. brasiliensis Fig. 21]. Large insects, 


bumblebees and Pompilidae wasps, were among the 


pollination 
Asteraceae and Rubiaceae, which thus differed. from 
the species of Apiaceae, Eriocaulaceae, and Cunonia- 


ceae, which have very small flowers (Table 4). 


V. HUMMINGBIRD-POLLINATED FLOWERS 


Hummingbirds were 
five species from five families. These species bear 
that 


distal fmm flowers o 


tubular/urceolate flowers make them easily 


— 


other species in this 
community (Fig. 3). Except the nectarless, spontane- 
ously self-pollinated flowers of Calolisianthus pedun- 
culatus (Fig. 2L), all species with red-orange flowers 
were pollinated by hummingbirds. 

Each of these plant species was pollinated by one to 
four hummingbird species (Tables 3, 4). Leucochloris 
albicollis 


pollinating four 


was the main visitor in the study area, 


species and acting as the single 
pollinator of Esterhazya macrodonta and Hippeastrum 
glaucescens. This Hippeastrum Herb. species bears 
very large flowers (tepals ca. 140 cm long). and the 


birds needed to perch on one of the inferior tepals to 


agents of some holophilous species of 


the exclusive pollinators of 


^). Pollen 


was deposited mainly on their wings. On the other 


reach the nectar at the flower base (Fig. 3, 


hand, Esterhazya macrodonta has protandrous flowers 


(Fig. 3C) that last up to six days. Nectar removal 
the first days of anthesis strongly affects nectar 


production (see Freitas & Sazima, 2001 for details). 


Sinningia allagophylla bears a long hairy in- 


lorescence with more than 40 flowers, which last up 
to seven days and present partial herkogamy associ- 
ated with protandry. This gesneriad was mainly 
pollinated by Leucochloris albicollis and Stephanoxis 
lalandi (Fig. 3D). 


Clytolaema rubricauda. The flower disposition along 


but also by Colibri serrirostris and 
the inflorescence of Dyckia tuberosa resembles that of 

allagophylla (Vig. 3B. D), and we only observed 
Colibri serrirostris females pollinating this terrestrial 
bromeliad (Fig. 3B). Agarista hispidula bears dozens 
of flowers arranged in dense inflorescences (Fig. 3E). 
poricidal anthers, and 


‘lowers are urceolate with 


pollen is secondarily presented on hairs at the corolla 


arly in the morning, 


opening. It was pollinated very 
Chlorostilbon 
albicollis (see 


mainly by the small aureoventris 


(Fig. 3E), but 


Freitas et al., 


also by Leucochloris 
2006 for details). In 


philous species were rare in these grasslands. We 


general, ornitho- 


found only four, seven, and eight flowering individuals 
of, respectively, Agarista hispidula, Dyckia tuberosa, 
and Hippeastrum glaucescens in the study areas. It was 
common for there to be < 10 Esterhazya macrodonta 
plants along each | km transect; however, along one 
exceptional transect we found 55 flowering individu- 
als. whose nectar features were also studied (Freitas & 
2001). In 


common in many places in the Serra da Bocaina. 


Sazima, contrast, S. allagophylla was 
Individuals of this species were usually distributed in 
small clusters (three to eight plants) separated from 
each other by many meters. However, 5. allagophylla 
plants occurred in large clusters in some areas—with 
dozens of individuals per 100 m*—probably due to 
clonal reproduction. Esterhazya Mikan., Agarista D. 
Don, and Dyckia Schult. f. 


receiving, respectively, 0.05, ( 


n were rarely visited. 
8, and 0.20 visits per 
hour (see Table 4 for times of o Each plant 
of Hippeastrum glaucescens received ca. one visit 


every two hours. Although plants of Sinningia Nees 


erowing in small clusters received ca. only one visit 


every three hours, when in large clusters, humming- 


birds visited each patch between three and six times 
individuals were visited 


per hour so that most 


intervals of one to two hours. 
of the 


bird species basically acted as low-reward trapliners 


Because low flower availability, humming- 
n the studied grasslands (foraging strategies after 
& Colwell. 1979). 


behavior was recorded only in the large clusters of 


Feinsinger Territorial. foraging 


Annals of the 
Missouri Botanical Garden 


Sinningia allagophylla. Clytolaema rubricauda and 
Stephanoxis lalandi behaved territorially in these 
circumstances, and Leucochloris albicollis occasional- 
ly acted as a parasite in territories set by Clytolaema 
rubricauda. Thus, Leucochloris albicollis may alternate 
its foraging strategy from low-reward traplining to 


territory-parasitism in some cases. 


G. MISCELLANEA 


Many oil-collecting bees—mainly large bees be- 
longing to Centris—were recorded pollinating Byrso- 
nima variabilis (Fig. 2D-E). The mechanism for oil- 
collection in this species was similar to that recorded 
for other Malpighiaceae species (see Vogel, 1990). 
Byrsonima variabilis was also pollinated by small bees 
belonging to Apidae (Meliponinae) and Halictidae in 


search of pollen, but to a minor degree. Some oil- 


collecting bees also collected pollen on this Byrso- 
nima Rich. ex Juss. species. 
The 


exclusively pollinated by svrphids, which carried the 


nectar-flowers of Tassadia subulata were 
pollinia on their heads. The elaborate. small, ruby 
flowers of this species are typically myophilous (sensu 
Faegri & van der Pijl, 1979). Small bees were also 
associated with syrphids on the nectar secreting 
flowers of Senecio oleosus, Wahlenbergia brasiliensis, 
Convolvulus crenatifolius (Fig. 2F-G). The last 


two species bear bell-shaped, lilac/lavender flowers. 


xs 


Pollinators first fed on the nectar of these three 
species, although they also searched for pollen. on 
them. Syrphids and small bees seemed to be co- 
pollinators of these species (svrphid/bee nectar-flower 


system). 


H. SPECIES WHOSE POLLINATION SYSTEMS ARE DOUBTFUL 


The flowers of 18 species were either not visited 
during the observation sessions or the role of their 
visitor in pollination was not clear. The results and 
brief discussions of these cases are presented herein. 

Although many Oncidium Sw. species bear flowers 
that produce nonvolatile oils with a chemical compo- 
sition suitable for larval nurturing by " Anthophor- 
bees (Reis et al., 2000: A. D. 


comm.). only one single observation of a bee actively 


idae” Faria, pers. 
collecting oils on Oncidium flowers has been reported 
so far (Singer & Cocucci, 1999), We did not observe 
Oncidium 
Table 4). 


However, we collected an individual of Monoeca sp. 


oil-collecting bees visiting flowers of 


species (total 47 hr. of observation, 


indet. 1 (Tapinotaspidini) (Fig. 2J) and an individual 
of Megachile aureiventris Schrottky (Megachilidae) 
each carrying an Oncidium pollinarium attached to its 


head. During an observation. session on Lupinus 


velutinus flowers. a female individual of Bombus— 
after collecting the pollen of Lupinus L. flowers— 
quickly visited a flower of Oncidium barbaceniae. 
During this visit, the bee vibrated in a failed attempt 
to collect pollen. We did not collect this individual 
bumblebee but observed that the visited flower had no 
pollinia, indicating that this bee could have removed 
them. These observations indicate that bees in search 
these 


of pollen could be acting as pollinators in 


populations of Oncidium. Thus, in addition to the 
expected system involving oil-collecting bees, Oncid- 


ium flowers could be pollinated by deceit by pollen- 


collecting bees. The following aspects point. t 


a pollination system by deceit in the studied 


populations: (1) bees in search of pollen visited their 
flowers: (2) bees of the “Anthophoridae” group 
intensively collected oil on Byrsonima flowers but 
were not recorded on flowers of the Oncidium species 
growing in the same patch: (3) visits by bees from any 
group were scarce on Oncidium flowers. and fruit 
production was very low in the three species. In short, 


pollination of the three species at Bocaina is unclear. 


and more comprehensive studies are required. t 
understand the pollination biology of this genus. 
The tubular flowers of Calolisianthus pendulus do 
not produce floral nectar, except for some rare flowers 
with minute nectarioles (sensu Vogel, 1998) located 


on petals. No visits to their flowers were recorded, and 


delayed self-pollination seems to ensure the repro- 
duction of this species, as well as that of the other 
four studied gentians of Bocaina, which include 
the ornithophilous Calolistanthus pedunculatus with 
nectarless flowers (Fig 4L) (see Freitas. 2004. for 
details). 
Although the 


(Mandevilla erecta, Habenaria parviflora, and Helia 


floral features of three species 


— 


oblongifolia [Fig. 2K]) point to phalaenophily and 
sphingophily (after Faegri & van der Pijl, 1979), no 
nocturnal moths were recorded. visiting their flowers 
(Table 4). T 


wel season 


iese species flower in the middle of the 
(Table 1). The 


greenish or yellowish tubes (spurs in M. parviflora), 


flowers have narrow 
presenting both odor and nectar at night. The long- 
lived flowers of H. oblongifolia are spontaneously self- 
Freitas, 2004). The 


other two species were rare in the studied areas. 


pollinated at the end of anthesis 


Vandevilla erecta bears large inflorescences with 
dozens of flowers, but we only recorded one fruit 
produced during three subsequent blooming seasons 
1998 to 2000). Several Habenaria parviflora 


lants produced fruits in the study areas. Although 
| | ) 8 


(from 


this species is self-compatible, spontaneous self- 
2001). Hh is 


pollinated by crepuscular crane-flies and nocturnal 


pollination is improbable (Singer. 


moths in 


an area al sea level by the Sao Paulo 


Volume 93, Number 3 
2006 


Freitas & Sazima 
Pollination in a High-Altitude Grassland 


A sb ee 
* > n = 
50 + A Pad 
40 + ln | X ut 
| IIIIII | 
0 ee ee e 
LD J F M A „HJ IAs ON DIF 


LD JF M A H J 


number of species 


D IF KM A KNM J J AS 


DJFMAMJ IAS ONDIJI F 


D J 
1998 


FMAM J J OND J F 


2000 


A C 
a 
oh — 


months 
1999 


Figure 4. Number of species blooming (line) and i 
998 to Feb. 2000, in the 
Distribution at the 


| 
rom Dec. | 


— 


flowering peak (bars) 


da Bocaina pue —A. 
4). —B-E. 


of each a system. —B. 


Serra 
mn level (n 12 Distribution of plant 


specie ctar-flowers 
= 30) —( 


1 mainly by small and large. bees a — 


Pollen-flowers pollinated by small and large bees (n. — 15). 

—D. Nectar-flowers me pose either by wasps or by wasps 

and dinis rans (n 22). —E. Nectar-flowers pollinated bs 
veral insect groups (SIG) (n = 20). 


2001). 


seldom use floral resources from grassland plants. 


coastline (Singer, Results indicate that moths 


The two studied species of Hypericum L. have 
extended flowering times, with individual plants 


flowering in different months year-round. The poly- 


androus flowers of both species are similar (Table 2). 
The flowers of the fourteen individuals of Hypericum 
fernum in the study area did not produce pollen, but 


all flowers developed fruits with well-developed 


— 


seeds. Similar results were found for the ca. 40 
individuals of Hypericum brasiliense. However. two 
individuals of the latter were found to produce flowers 
with pollen grains (ca. 95% viability) December 
1998. One of these plants had been followed since 
January 1998 


without pollen until December 199€ 


. and, curiously, it produced only flowers 
3. This capacity for 
sexual and apomictic reproduction within the same 
indiy idual is named 


environne ntally influenced (Asker & Jerling. 


facultativism and may be 
1992). 


In the pollen-producing flowers of Hypericum brasi- 


liense, self- and cross-pollen tubes grew down the 
style, but only crossed ones seemed to penetrate the 
ovule. We observed these pollen-producing individu- 
als in December 1998 


that insects, mainly syrphids and small bees, searched 


and January 1999 and found 


for pollen and pollinated their flowers (Table 4). 


Empty anthers were also observed in herbarium 


specimens of Hypericum species from lowland forest 
areas (V. Bittrich, pers. comm.). It is expected that H. 
ternum also occasionally produces pollen. 

No floral visitors were registered during the 28 hr. 
f observation of the four Leandra Raddi species. 
There are records of apomixis and pollenless flowers 
in species of this genus (Goldenberg & Shepherd, 
1998; R. 


Goldenberg, pers. comm.); thus the Leandra 


T at Bocaina may also be apomictic. Another 
Melast . Microlicia isophylla, was not visited 


e Has v. This species is very common in grasslands and 
bears attractive flowers, with magenta petals and vivid 
vellow anthers. Their smooth pollen grains are highly 
Bees 


vibration are the expected 


viable and are produced in large quantity. 
collecting pollen by 
pollinators of this species. although it could also be 
apomictic. 

greenish flowers of Eryngium canalicu- 
Floral 


canaliculatum point to a SIG pollination 


The small. 
latum are very similar to those of E. horridum. 
traits of H. 
system, visits by two species of 


although only four 


beetles were recorded on these flowers. Further 
observations are necessary to ascertain their pollination 
system. We did not observe pollinators visiting the 
showy flowers of Chromolaena xylorrhiza and Epiden- 
drum secundum that fit in the melittophilous syndrome. 
DEGREE OF SPECIALIZATION OF POLLINATION SYSTEMS 
third of the 


pollinated by only one pollinator group (monophilous, 


Around one plant species were 


Table 6). For 50 out the 70 species pollinated by two 


Annals 
MEN 5 Garden 


Sinningia - — 
Esterhazya —— 

Agarista 

Dyckia 


Hippeastrum 


D| J|F |M| AIM| J| J| A SIOIN[ DJ |F 


— 


1998 


Figure 5. 
1998 to February 2000. Thin and thick lines 


are abbreviated ace cording to their genus 


December 
species 


to six pollinator groups. we were able to determine one 
or two insect. groups acting as main pollen vectors 
(oligophilous and polyphilous). The remaining 20 
species were holophilous. since they presented no 
Monophily 


and oligophily 


dominant pollinator group. dominates 


among pollen-bee flowers, among 
syrphid and bee pollen-flowers. Species belonging to 
the bee nectar-flowers and wasp/fly flowers systems 
are more equitably distributed among monophily. 
oligophily, and polyphily (Table 6). 

More than half of the species pollinated by small and 
large bees are monophilous or oligophilous (Table 6). 
In contrast, most species pollinated by flies, butterflies. 
and beetles are holophilous. The species pollinated by 
| distributed 


wasps and syrphids are better among 


oligophily, polyphily, and holophily. 
FLOWERING PHENOLOGY 


Flowering duration in grasslands was long (mean 
5.5 + 2.92 months, n = 124), and the mean flowering 
peak duration was 2.0 + 


1.56 months (n = 124 


(Table 1). Flowering pattern in this community was 
seasonal, with a peak during the middle of the wet 
season (Fig. 4A). Over 60% of the species flowered in 
January and February, and ea. one-third of the species 
were al their flowering peak during these two months. 
There was a drastic reduction in flowering during the 
end of the dry season/beginning 
(Aug.—0ct.). The fire 


September 1999 may have only slightly affected these 


of the wet season 
that occurred in one area in 
general patterns, because only eight species from this 
area did not occur in at least one of the other two 
studied areas (see Table 


Blooming patterns of the species with nectar- or 


pollen-flowers pollinated by bees followed the pattern 


indicate 


months 
1999 


2000 


Flowering patterns of the five Te pollinated species m the Serra da Bocaina grasslands from 


| respectively. flowering time and flowering peak. Plant 


Table 1 for 1 n 


observed in the community (Fig. 4A-C). However. 
there was a slight increase of species al peak 
flowering in June (Fig. 4B-G). In contrast. most 


species with nectar-flowers pollinated either by wasps 

by wasps and dipterans reached their flowering 
HD). Wasps are very 
important pollinators during that season. For example, 
May, 
1999 were exclusively or mainly pollinated by wasps 
(Fig. 4A, D). 


resembles the 


peak during the dry season (Fig. 
I] out of 23 species at their flowering peak 


The flowering pattern of SIG species 
general pattern of the community, but 
a few plants at their flowering peak were recorded 
LA, E). 


five hummingbird-pollinated species concentrated. at 


throughout the year (Fig. The flowering of the 


the beginning of the wet season (Fig. 5). No plants 
flowered in April and May, and no plants were at their 
flowering peaks during five months of the year. 
DISCUSSION 

Although Martinelli (1989) reported a total of 215 
flowering species in the Bocaina grasslands. we 
collected ca. 260 About 20% ol 260 


species belong to the families Poaceae. Cyperaceae, 


) species. f these 
and Juncaceae and are anemophilous. This study thus 
the ca. 210 


animal-pollinated species from the Bocaina grass- 


included around 60% (124 species) of 
lands, and pollinators were determined for about half 
of n zoophilous species (106 species). In addition, 


ca. 80% of the 47 


plant families found in the Bocaina 
Modi de were represented in this study, so that our 
sample was fairly representative of these high-altitude 
erasslands. 

We carried out intensive pollinator observations for 
some plant species and brief ones for many others, as 


expected in studies on plant-pollinator interactions al 


Volume 93, Number 3 
2006 


Freitas & Sazima 
Pollination in a High-Altitude Grassland 


the community level which encompass many species 
with low population density (see Momose et al.. 1998). 


Both the short observation times and the generally low 


— 


flower visitation rates explain why we may have failed 
to ascertain the typical pollination system of some 
species in the Bocaina grasslands. However. since we 
used broad categories to describe the pollination 
systems, we may oulline general interaction trends in 
this community and establish general comparisons. 
There are no comparative data on the floral and 
pollination biology of any other high-altitude grass- 
land areas in southeastern Brazil. However, commu- 
nity level pollination data from vegetational types with 
both abiotic similarities and strong. biogeographic 
connections are available. These include the high 
temperate Chilean Andes, especially in the sub- 
1982). 


Venezuelan Guiana high-altitude shrubland or arbus- 


Andean scrub zone (Arroyo et al. the 


tal (Ramírez. 1989), the Brazilian open savannas or 
cerrado (Barbosa, 1997: see also Silberbauer-Gotts- 
berger & Gottsberger, 1988 and Oliveira & Gibbs, 


) 


2000 for other cerrado physiognomies), and. t 
some extent, data from the Brazilian campo rupestre 
(Faria. 1994). We focus our discussion on. compar- 
isons between these formations and the Bocaina 


erasslands. 
FLORAL TRAITS 


Flowers with light colors—including greenish. 


are well represented 


pink, and white color groups 
in the grasslands, resembling the arbustal and cerrado 
communities. in which light-colored flowers predom- 


Silberbauer-Gottsberger & Gottsberger. 1988: 


— 


inate ( 
Ramírez. 1989: Barbosa. 1997). However, more than 
half of the species at Bocaina have showy flowers 
(violet, vellow, and red color groups). This dominance 


is due to the presence of Asteraceae and Melastoma- 


l 


aceae, with their showy violet-colored flowers. i 
these grasslands. Furthermore, greenish flowers were 
more continuously distributed throughout the year, in 
contrast to violet flowers that are concentrated during 
the wet months. Using canonical diseriminant analysis 


(L. Freitas. unpubl. res.). we only detected a weak 


relationship between floral color and the frequency of 


visits for each pollinator group in the community. Red 


flowers were strongly associated with hummingbird 
visitation and greenish flowers weakly associated with 


ions primarily by beetles and wasps. Although 


visita 
we used the human color spectrum, such results for 
insect responses to flower color resemble the general 
trends detected by Chittka et al. (1994) and Waser et 


al. (1996). who found no statistical differences among 


insect pollinator groups, at the order level, in terms of 


the colors of the visited flowers. 


The prevalence of nectar as the main floral resource 
at Bocaina was recorded in other grassy communities 
(Arroyo et al., 1982; Ramírez, 1989; Faria, 
Barbosa. 1997). In fact. nectar-flowers are the most 
important flower class among angiosperms in general, 
because most pollinator groups are nectar consumers 
(Endress. 1994). The percentage of pollen-flowers at 
Bocaina was similar to that registered in the campo 
rupestre (Varia, 1994) and the cerrado (Barbosa, 1991). 


Mal n | 


IVECO TAO a 


species are the most important 


pollen-flowers at Bocaina, as well as in the campo 
rupestre (Faria, 1994). Species of this family are also 
important pollen sources in the arbustal (Ramirez, 


1989). However, species of other families, such as 


eguminosae and Myrtaceae, constitute the chief 
sources of pollen in the cerrado (Barbosa. 1997). 
Malpighiaceae are represented by several species in 
the campo rupestre and the cerrado (Faria. 1994: 
Barbosa, 1997). which explains why oil-flowers are 
more frequent in those communities than in the 
Bocaina grasslands. 

Although small flowers predominate in the arbustal 
and the cerrado (Ramirez, 1989; Barbosa, 1997), the 
flowers of these two communities are larger than those 
at Bocaina. As for the floral color patterns. the relative 
in the 


patterns. 


Asteraceae in community 
the The 


prevalence of small tubular flowers secreting nectar 


predominance of 
strongly influences floral size 
at Bocaina is a direct reflection of the number of 
Asteraceae presenting these floral traits and appar- 
ently not the result of selection by the local pollinator 
agents. 

The prevalence of hermaphroditic flowers is typical 
of angiosperms, and, consequently, characteristic of 
1993), including the 


erassland flora at Bocaina. However, many of the 


most communities (Ramírez. 


hermaphroditic flowers exhibited spatial or temporal 
separation of the male and female functions, through 
dichogamy. herkogamy, or heterostyly. In fact. a high 
(protandry) is expected 


frequency of dichogamy 


the 


in a flora dominated by Asteraceae, owing t 
frequent presence of secondary pollen presentation on 
the style among the Asterales (see Yeo, 1995). In 
addition, the female ray-florets of many Asteraceae 
species impart a state of protogyny to the head as 
a whole. 

All the dioecious species of the studied community 
belong to the genus Baccharis, which was primarily 
associated with wasp pollination. Dioecy has often 
been associated with a generalist mode of pollination 
by small opportunistic insects (Bawa, 1994: but see 
Renner & Feil, 1993) or with wind pollination (Culley 
et al., 2002). Although this study did not focus on 
some evidence suggests the 


abiotic pollination, 


existence of ambophily among the studied taxa. in 


510 


Annals of the 
Missouri Botanical Garden 


which wind could be a pollen vector in addition to 
The 


category, 


insects (e.g. Croton dichrous). species of 


Baccharis may also fit in this as female 
heads have well-exposed stigmata and male flowers 
produce small pollen grains. Furthermore, we found 


Baccharis 


pollen grains of species adhered t 


microscope slides covered by elvcerol, during explor- 


atory experiments at the study sites, and most species 


of this genus flower during the winter. when insect 
activity is reduced. Pollinator limitation, dry weather. 
and the presence of recurrent wind in the winter 


could favor the evolution of anemophily and amboph- 


ily in the high-altitude grasslands (see Culley et al., 
2002). 
POLLINATION AGENTS 

Hymenopterans, markedly small bees. are the 


predominant pollinators in the Bocaina grasslands. 
as well as in the cerrado, arbustal, 
1982 


OZ, 


and Subandean 
1989: Barbosa. 


reported as 


zone (Arroyo et al., Ramírez, 


1997). Bees have been 


the main 
pollinators for most communities (e.g... Moldenke. 
1976; Kevan & Baker, 1983: Roubik. 1989: Bawa. 
1990; Momose et al., 1998; Nakano & Washitani. 


2003: Ramirez, 2004). and different patterns—for 
example the dominance dipterans—seem to be 


restricted to very singular habitats; such as Arctic and 
alpine areas or oceanic islands (e 
Primack & 


However, 


g., Kevan & Baker. 
1993: 


communities 


Inouye, Anderson el al.. 


certain with similar 


percentages of bee pollination may exhibit: marked 
differences among their general patterns of pollina- 
lion. Bees belonging to different groups show great 
variation in, for instance, their preferences, abilities. 
1989), 


lead to 


and behavior on flowers (see Roubik. and 


such variation is expected to different 
pollination patterns among communities with distinct 
bee faunas. 


The 


Bocaina grasslands are the large bumblebees and 
e o 


two main functional groups of bees in the 


small halictid bees. Agents of one or both of these 


groups were among the pollinators of 67 out of the 106 


species (03%). Bombus and Halictidae species 
pollinated, respectively, 78% of the large bee- 
pollinated species (28 out of 36). and Ap Jo of the 


small bee-pollinated species (54 out of 7 


. Both bee 


groups exhibited a generalist sa behavior. 


visiting flowers of different shapes, sizes. and colors, 


which, in many cases, were also pollinated by other 


groups. An illustrative example is that of an individual 


of Bombus atratus observed visiting the violaceous 


flowers of seven species. belonging to Asteraceae. 


amiaceae, and Rubiaceae, in a single foraging trip 


(for ca. three min). Such generalist behavior is typical 


of the Brazilian species of Bombus (Alves-dos-Santos, 
1999; 2000). Bumblebees pollinated 


most of the large bee-pollinated species at Bocaina 


Barbola et a 


and predominated in almost half of them. 


the 


including 


terms of 


richest ones n 


nectar, such as 
Oxypetalum sublanatum, Lobelia camporum, Cuphea 
glutinosa, and Verbena hirta. Moreover. these bees 
showed a trapline behavior, favoring pollen dispersal 
al great distances. Considered together, the variety of 
visited flowers (including nectar- and pollen-flowers), 
foraging behavior, dominance of the richest resources. 
and year-round activity make bumblebees the main 


pollinators of the plant group related to large bee 


pollination. Bumblebees are known to be one of the 
most important pollinators in cool temperate climates 
1993: Nakano & Washitani, 2003). and 
the harsh climatic conditions, for a tropical place, that 
the 


(Kevan et al., 


characterize Brazilian high-altitude grasslands 


may be related to the high importance of bumblebee 
pollination at Bocaina. 


P / 


A , 
oC 177051677097 
S 


(Halictidae) visited 
only flowers of Mikania lundiana and Senecio oleosus 
2G). Similarly, 


oligolectic 


Cyaneus 


(Asteraceae) al Bocaina (Fig. this bee 


species Was d associated to F. 
in a grassland area at Lapa, State of 
Brazil (Barbola et. al.. 2000). 


Pseudagapostemon CYaneus occurs from southernmost 


narrrow 
oleosus flowers, 
Paraná. southern 
Brazil to the State of Sao Paulo. In this last region, it is 
1000 m 1989). The 


particular interaction between Senecio oleosus and 


only reported above (Cure. 


Pseudagapostemon cyaneus at both Bocaina and 1 9 0 


s one 


of the most illustrative example S < bio- 
geographic connections between the mountain ranges 
of southeastern Brazil and the cool lowland areas in 
1997 


Pollination by wasps has been observed 


southernmost regions (see Behling, 
some 
Ophrys 


). as well as among generalist plants 


highly specialized and infrequent cases (e.£.. 


„ Ficus 
pollinated by several insects (Faegri € van der Pijl. 
1979). In 


reported as minor agents (e.g.. 


most communities, been 
1982: 
1985; Silberbauer-Gottsberger & Gotts- 
1998: Oliveira & Gibbs. 


wasps pollinated 45% 


wasps have 
Arroyo et al., 
Bawa et al., 
berger, 1988: 
2000). In 


species in the Bocaina grasslands and acted as the 


Momose et al.. 


contrast. 


f the 


exclusive or main pollinators of more than 20% of the 
animal-pollinated species in this community (Ta- 


ble 5). Although wasps have their importance in the 


pollination of cerrado and arbustal. they are the 
exclusive pollinators of i | few species in these 
communities (Ramirez, 198€ ee 1997). In fact 
as far as we know, 1 importance of wasp 


pollination as observed in the Bocaina grasslands 
has never been reported for any other ecosystem. 
I 3 à 


Whether the wasps are minor pollinators or their role 


Volume 93, Number 3 
2006 


Freitas & Sazima 511 


Pollination in a High-Altitude Grassland 


n pollination has been neglected in Neotropical 


communities is still an inadequately explored issue 
that deserves further study. However, at least in some 
subtropical habitats, the rates of floral visitation by 
wasps, as well as their probable importance as 
pollinators at the community level, seem to have been 
underestimated (P. Feinsinger, pers. comm.). 

At Bocaina, 


pollinators of species with a generalist pollination 


lepidopterans were generally co- 


system or secondary pollinators of species pollinated 
mainly by bees. Similar results were found in cerrado 
and arbustal areas (Ramírez, 1989; Barbosa, 1997; 
Oliveira & Gibbs, 2000). However, butterflies are very 
important pollinators in the open vegetation of the 
Chilean 


(Arroyo et al., 


Andes, especially in the subnival zone 
1982). In spite of the scarcity of their 
visits to flowers, butterflies (mainly Hesperiidae) may 
have an important role in the gene flow of many 
at Bocaina, 


species transferring pollen at long 


distances due to their traplining behavior. Despite 
some highly specialized cases, such as those related to 
long-tongued flies in South Africa (e.g., Goldblatt et 
al., 1995), fly pollination has been considered mostly 
unspecialized, since flies do not feed their young and 
may have other food sources than flowers (Kearns, 
1992; 1996). Most species pollinated 
by the three most important fly groups at Bocaina— 


Proctor et. al., 


Syrphidae, Tachinidae, and Bombyliidae—have small 


nectar-flowers that were pollinated by several agents. 
However, syrphids, associated with small bees, were 
the pollinators of many species, particularly those 


with pollen-flowers. These last species present 


— 


monophilous or oligophilous pollination systems. This 


result shows that pollination by flies can be 


specialized even in the absence of extreme proboscis 
elongation. If we consider the species for which 
syrphids and other dipterans are either the main or the 
percentage of species 


exclusive pollinator, the 


pollinated by flies at Bocaina is similar to that found 


in cerrado and arbustal areas (Ramírez, 1989; 
Barbosa. 1997). Nevertheless, in the Chilean Andes, 
myophily is more common (Arroyo et al., 1982). The 


ereat importance of flies and butterflies in the Chilean 


Andes, as compared to the other communities 


considered here, probably reflects the unfavorable 


climatie conditions for bee activity in the Andes 


(Arroyo et al., 1982; see also Kevan. 1975 and 
Kearns, 1992). In the Bocaina grasslands, climatic 
conditions are not harsh enough to limit the de- 


velopment of bees, which dominate the pollination in 
this community. 

At Bocaina, the percentage of hummingbird- 
pollinated species is similar to that observed in 
cerrado areas (Barbosa, 1997), but it is lower than that 


of other high-altitude communities, such as arbustal 


— 


12%) (Ramírez, 1989) and campo rupestre areas 
10%) (Faria, 1994). 


ornithophilous species constitute secondary nectar 


In the Bocaina grasslands, 


ieee 


sources for hummingbirds that find their main nectar 


sources i surrounding high-altitude forest 
(Sazima et al., 1996; Freitas € Sazima, 2001). Two 


First, in these 


erasslands, no ornithophilous species were flowering 


factors, at least, support this idea. 
in April or May or were in their blooming peak during 
February to June. Therefore, this community would be 
unable to sustain hummingbirds throughout the year. 
Secondly, the five ornithophilous species were seldom 
visited by hummingbirds, while hummingbirds in- 
tensively fed on ca. 30 species of the surrounding 
1996; Buzato et al., 2000; 
Bocaina, the 


forest areas (Sazima et al., 


L. Freitas, pers. obs.). Hence, at 


continued replacement of forest areas by grass- 
lands—mainly because of fires—is a risk factor for 


populations of the five hummingbird species observed 


— 


in the grasslands, in particular for Stephanoxis l. 
lalandi, which lives exclusively in some high-altitude 
areas of southeastern Brazil (Grantsau, 1989; Sazima 


1990). 


We did not observe any sola by either bats or 


et al., 
other mammals in the Bocaina grasslands. Moreover, 
no species were collected with floral features in 
accordance with the mammal pollination syndromes, 
suggesting the absence of such pollination systems in 
these grasslands. Reports of bat-pollinated species 
from mesic habitats in high-altitude grasslands are 
restricled to three Bromeliaceae species (Martinelli, 


1994, 1997). In 


represented in the high-altitude forest areas that 


contrast, bat pollination is well 
surround the grasslands (Sazima et al., 1999 

The Bocaina grasslands also lack perfume flowers, 
another highly specialized pollination system related 
1993). 


Euglossine bees have restricted altitudinal limits of 


to male euglossine bees (e.g., Sazima et al., 


distribution (Roubik, 1989), and aromatic compounds 
traps did not detect any such bee in the Bocaina high- 


altitude grasslands. 


— 


In contrast to bats, nocturnal pollination by moths 
is expected since the Bocaina grasslands present at 
least three species adapted to it. However, we did not 
observe these pollinator agents, and the low fruit set in 


Mandevilla erecta indicate low—perhaps unpredict- 


able—rates of flower visits by nocturnal moths 
these habitats. As suggested for the cerrado area 


(Oliveira & Gibbs, 2000), strong winds at the study 
areas may be harmful to the moth pollinating species. 
restricted time of the 


In addition, the flowering 


species that might be moth-pollinated in the Bocaina 


grasslands may be linked to migrating pollinator 
agents, which would increase the difficulty of 


observing pollinators. 


512 


Annals of the 
Missouri Botanical Garden 


FLOWERING IN RELATION TO POLLINATORS 


At Bocaina, the bees of all families reduced their 


activity during the dry season. Such reduction in the 


number of bee species and individuals in activity 


during the winter was observed in connected 


ecosystems in southern Brazil ( Mves-dos-Santos, 


1999; Barbola et. 
factors, such as the reproductive phenology of bee 


al., 2000) and is related to several 


species, climatic restriction for flying, and flower 


availability. The flowering patterns of species polli- 


nated mainly by bees followed the pattern of the 


community as a whole. Thus, flower resources [or 


bees, especially large bees, decrease during the dry 


season. In addition, such climatice conditions as low 


temperatures, low precipitation, and occasional frosts 
are apparently unfavorable for flowering during the 
winter months. Moreover, the flowering pattern of 
species with either nectar- or pollen-flowers over- 
lapped, and this result indicates reciprocity between 
nectar) 


the availability of resources for adults (ite. 


bees ab the 


and pollen) community 


arvae (1.6., 


o ascertain if these 


is difficult 


level, although i 
patterns are not casual. 


constant activity pattern throughout the year, 


more 


1 contrast to bees, social Wasps showed 
and 
pollination by this group was prominent during the dry 


season at Bocaina. Because. in this area. several 


species that are pollinated mainly by wasps are 


apparently also pollinated by wind, anemophily. 


rather than pollination by wasps. may be an ultimate 


actor driving flowering concentration during the dry 
season in this group of species. If we consider the 


flowering phenology of species belonging to the most 


representative pollination systems at Bocaina, only 
the few species that flower during the dry season (wel 
season for wasp-pollinated species) seem to be 


I^ UN 


fundamental to support populations of insects. 
example, the floral nectar of the nearly year-long- 
flowering species Cuphea glutinosa, Galianthe brasi- 
and the winter-flowering 


liensis, and Verbena hirta. 


Eremanthus erythropappus constitutes a keystone 
. 1993) for 


These four species 


plant resource. (sensu. van Schaik et a 
bee pollinators during the winter. 


are large shrubs or treelets (E. erythropappus) with 


many flowers and are particularly important for 


bumblebees during this period low flower avail- 


ability. 
common during the dry 


Anthropogenic fires are 


season in the Bocaina grasslands. A few months after 


fires, we observed mass-flowering of several species 


such as Sinningia allagophylla, 1 isoph vlla. 
T.. Galianthe angustifo- 


Tibouchina frigidula, T. minor. 


lia, and Xyris o Flowering controlled by fires 


is a well-known and widespread phenomenon among 


1990), but, ) 


Safford (2001). based on his observation in the Serra 


cerrado species (Coutinho, according | 


do Caparaó, it is not characteristic of the Brazilian 
high-altitude grasslands. A possible reason for the 


difference between Safford’s observation and ours is 
that the ecological similarities between cerrado and 
high-altitude grasslands may be more pronounced al 
Bocaina, due to intense fire regimes, than at other 
erasslands of southeastern Brazil (see 
1989 for 


Bocaina grass- 


high-altitude 
19904, 2001 


comparison). 


Safford. and also Martinelli. 


floristic In this sense. 
lands may be more pyrogenic than other grassland 
areas, such as those from Pico do Hatiaia, Serra dos 


Órgãos. and Serra do Caparaó. 


SPECIALIST VERSUS GENERALIST POLLINATION SYSTEMS 


based on the 
1979) 


traditional view in studies 


The 
syndrome concept (see Faegri & van der Pijl. 
supports the idea that pollination systems mostly tend 
toward increased specialization. Such a view has been 
questioned more recent times due to evidence of 
more widespread generalization in pollination systems 
(e.g.. McDade, 1992; Waser et al.. 1996; 

1996: Ollerton, 1996: Olesen & Jordano, 2002). These 


field observations of regional or local flora. contexts 


Herrera. 


show that many plant species are visited by many 
animal species, and, as a result, the mean number of 
species of animal visitors per plant species is high. 


functional 


\lternatively, we used the number. « 
groups of pollinators per plant species as a parameter 
of the degree of specialization of the pollination 
systems al Bocaina. Functional groups are more suited 
t 
reasons (see Fenster et al., 2 
l 


lo exert similar selection pressures on the floral traits 


» this purpose than pollinator species for several 
004). Different species of 


1e same functional group, for instance, are expected 


related to pollination mechanisms. Thus, considering 
entities some 


that 


such species as distinct generales 
Another 


pollinator species belonging to each functional group 


the number of 


artificiality. actor is 


varies drastically within a local community, as does 
the species richness of a given functional group among 
distinct communities or regions. Thus, plant species 
pollinated by more diverse functional groups would 
simply appear more generalized, despite the bi- 
ological proprieties of their pollination system. 
Ramírez (2002) used the functional group approach 
the 


temperate 


de gree of specialization of nine 


O Compare 


tropical and communities. The mean 


number of ne groups per plant species ranged 
from 1.0. in a beach dune vegetation of northern 
Brazil. to 1. 
Plain. Using the same functional groups as Ramírez 
2002). 


35. in the forest of the Venezuelan Central 


the Bocaina grasslands has a mean of 2.09 


P 


Volume 93, Number 3 
2006 


Freitas & Sazima 513 


Pollination in a High-Altitude Grassland 


pollinator types per plant. probably one of the highest 


= 


values for worldwide floras. The higher degree o 
generalization al Bocaina is also evident if we 
consider the distribution of species along the four 
specificity categories (i.e... monophily to holophily). 
Monophily, as used here, encompasses both the 
monophily and oligophily categories of Ramírez 
(2002). More than half of the species, at 
belonged to these two categories in six tropical 


54% 100%: 


2002). in contrast to the Bocaina grasslands where one 


least, 


communities (between and Ramírez. 


third of the species are monophilous. 

However, the functional group approach to de- 
termine the specialization degree of pollination 
systems may also induce some artificiality. One 
well-demonstrated theory is that different pollinators 
of a given plant species may vary in their pollinating 
abilities in such a way that this plant species may 
effectively specialize on its most efficient visitor 
(Stebbins, 1970; Schemske & Horvitz, 1984: Ollerton, 
1996; Fenster et al., 2004). Another point is that the 
definition of functional groups of pollinators (e.g., 


small based on 


vees, moths, bats) is typically 
systematic boundaries. Thus, two pollinators in 
distinct orders are in general placed in different 


functional groups, even if they probably exert similar 
selection pressures on floral traits. For example, at 
Bocaina, some of the pollen-flowers are pollinated by 
syrphids and small bees and some of the nectar- 
flowers are pollinated by small bees and wasps: we 
combine 


named this situation oligophily. [f we 


monophily and oligophily, we conclude that almost 
two thirds of the zoophilous species in this community 
present a specialization onto functional groups. Like- 


wise, a reanalysis of the dataset from the flora of 


Carlinville, Ilinois, indicated that about 75% of the 
species showed specialized pollination systems (Fen- 
ster et al., 2004) contrasting with a previous 
conclusion of generalization predominance in this 
community (Waser et al., 1996), which was based on 
the number of pollinator species per plant species. 
Even above "wide 


using the perspective of 


^ 


functional groups” (i.e. oligophily considered as 
a specialized system), a considerable degree of 
generalization (more than one-third of species) is 
found in the Bocaina grasslands. Broadly, generaliza- 
tion is predicted as long as temporal and spatial 
variance in pollinator quality is appreciable; different 
pollinator agents do not fluctuate in unison and are 
similar in their pollination effectiveness (Waser et al., 
1996). Thus. the climatic conditions of the high- 
altitude grasslands may favor the occurrence. of 
generalists at Bocaina. Generalization in pollination 
seems to 


systems be more frequent in naturally 


inclement or unstable areas, in the modern agricul- 


tural-urban mosaics or human surroundings, and 
among short-lived plants (Vogel & Westerkamp. 
1991: Johnson & Steiner, 2000, 2003). In this sense. 
frequent fires may be an additional factor that makes 
the environment harsher for specialists (both plants 
and animals) in the Bocaina grasslands. 

Generalist’ pollination systems have often. been 
characterized by small and white or greenish-colored 
flowers pollinated by small insects, the diverse small 
insect (d.s.i) syndrome of Bawa et al. (1985). However. 
some of the small greenish flowers at Bocaina present 
highly specialized pollination systems (e.g... Go- 
nioanthela hilariana), and the pollination of some 
holophilous species involves such large insects as large 
bees, wasps, and butterflies. Therefore, the degree of 
generalization should not be inferred from the sizes of 
flowers and pollinators, as previously postulated. by 
Ramirez (1989). Although represented mainly by the 
families Asteraceae and Rubiaceae, the holophilous 


species at Bocaina vary in both their floral traits and 


their flowering phenology. In addition, some gener: 
with holophilous species also have monophilous or 
oligophilous species: for example, wasp pollination in 
Baccharis and Borreria G. Mey. and bee and butterfly 


pollination in Vernonia Schreb. and in the Eupatorium 


a alliance. For such genera, the floral features that 
determine whether a few or several pollinator groups 
are attracted to the flowers of specialists or generalists, 


respectively, remain unclear. Moreover, since the 


degree of specialization/generalization of pollination 
systems may be influenced by factors such as plant life 
history, phylogenetic constraints, vegetation. strata, 
successional status, plant abundance, breeding system, 
and local fauna (Stebbins, 1970; Vogel & Westerkamp. 
1991; Waser et al., 1996; Ollerton, 1996: Johnson & 
Steiner, 2000, 2003; Ramirez, 2002, 2004; Nakano & 
Washitani, 2003: Fenster et al.. 2004), floral traits 


alone may be inadequate to address this issue properly. 


CONCLUDING REMARKS 


The correlation between certain taxa and particular 
ecological conditions may enhance the abundance of 
given plant groups. Thus, ecological attributes, un- 
related to sexual reproduction, may promote certain 


taxonomic groups with particular reproductive trails, 


and hence, may bias the frequency distribution ¢ 
these traits in the community (Vogel & Westerkamp. 
1991; Ramírez, 1993). Apart from the wind-pollinated 
species of Poaceae and Cyperaceae, the Bocaina 
erasslands are dominated by species of Asteraceae, 
which strongly influence the floral features (1.e.. short 
tubular flowers that offer nectar and have exposed 
reproductive parts) at the community level. In these 


erasslands, general trends related to the types of 


Annals of the 
Missouri Botanical Garden 


pollination systems and degree of generalization may 


be connected to the phylogenetic floral trends of 


species with certain abilities—not necessarily related 


to pollination—to occupy these habitats, such as short 


living-cyeles and adaptations to frost, low 


mean 
temperature, high daily temperature variation, acidic 
soils, and, more recently, some kind of adaptation to 
fire (see Safford, 19994 

Biogeographic and palvnologic studies indicate 
that high-altitude grasslands in southern and 


Brazil 


the Pleistocene cold and dry 


southeastern have to be understood as 


relicts of climates 
and of the early- and mid-Holocene warm and dry 
1997). 


these grasslands are unlikely 


climates (Behling. Because of their rela- 


tively recent origin, 
have produced highly coevolved plant commu- 


nities, although some relictual plant-pollinator 


interactions could have been retained in 


Cases 
where both the plant and its pollinator colonized 
the mountain tops. These grasslands are linked to 


episodes of expansion and retraction due to glacial 


[e] 
events during the Quaternary (Behling, 1997; 
Safford, 1999a). Such a situation may have favored 


species with the ability to occupy new habitats 


quickly. Dependence on a few specialized polli- 


nators may be less suitable than more generalist 


pollination systems in the occupation of a new 


habitat, because previous pollinator agents may be 


left behind. In fact, generalized pollination 


systems may have been an ecological advantage 


for plants colonizing post-glacial landscapes 
| 8 I : | 


(Johnson & Steiner, 2000, 2003). Alternatively, 
for those species with more specialized pollination 
systems, some floral traits that are frequent at 


Bocaina, such as spontaneous self-pollination 


mechanisms and long-lived flowers, may be 


both 


maintenance in these grasslands characterized by 


advantageous to plant establishment and 


unpredictable and scarce pollinators. 


Literature Cited 
Monso, M. T. A. 1977. 

Brasil. 
\lves-dos-Santos, J. 


Vegetação da região sudeste do 
Geogr. 34: 91-118 

1999. 
mala atlántica, restinga e dunas do litoral norte do estado 


Brasil. 


Revista Brasil. 

Abelhas e plantas melíferas da 

do ps x uus do Sul. Revista Brasil. Entomol. 
191- 

Anderson, G. J. G. Be 

2001. 


selected plants e E mie to Juan Fernández Islands. Amer. 


Stuessy & D. J. 


and pollination « 


nardello. T. F. 


Crawford. Bn ding system 


J. Bot. 88: 220-233. 
Arroyo, M. T. f © Primack & J. Armesto. 1982. 
Community studies in piani ecology in the high 


temperate Andes ol Central Chile. I Pulli Mon mee han- 
sms and 5 variation. Amer. J. . 09: 82 
Jerling. 1992, 


Boc “a 1 0 


e. S. E. a in Plants. CRC 


| ^n 88. 


Barbola. 


de recurs 


aroca & M. 


„8. . Almeida. 2000. 
IS florais 


ias silvestres (Hyme noplera. 


Utilização 
por 5 
\poidea) da Floresta Estadual BN T (Lapa, Paraná, 
Brasil). 


Entomol. -19 
Barbosa. A. A. A. 75 


Revista Brasil. 
A. A. 199 cue de uma 


Uberlandia/MG. Ph.D. 


Campinas, Campinas, Sao Paulo. 


1 
Comunidade de e 
thesis. Univ. 
K. 8. 


Bawa, 19 m Pla ant- a interac Bu in tropical 
2 


rain forests. Annual Rev. Ecol. Syst. 21: 4.22. 
. 1994. Pollinators of is ee 1 
a reassessment? No, D Amer. J. Bot. 81: 456-460. 
. Bullock, h - Pure R. E. Coville & M. H. 


c x Sade aoe biology of tropical lowland 


rain forest trees. II. Pollination systems. Amer. J. Bot. 72: 
3 7 
Behling, 1997. Late Quartenary vegetation, climate and 


fire et of the Araucaria forest and campos region from 
. Paraná State 
10 09-121. 
Buzato, H M. ar ima & l. 
dA js Tod at three 
32 


Serra Campos Gera (South Brazil). Rev. 


2000. Hummingbird- 


Sazima. 
lante Forest sites. Biotropica 


8 


lal erik. Pa 


n Leaf. character- 
istics of 


V, Werger. 1981. 

T D of the high plateau of Hatiaia, Brazil. 
5 39—48 

Chittka, L.. A. a Menzel. 1994. 


U ore as d aa nt of flower reflections, and the 


Vorobyev & R. 


colour a ME of Hymenoptera. Vision Res. 34A: 
| 189- 
Cocucci KS. Vogel. 2001. Oil- producing flowers of 


pollinators i 


ahi species il idaceae) and their 1 
96: 2 


~ 


southern S E. 97 rica. Flora 196: 26— 


in the ecology of the Brazilian 
1055 in J. G. 
ils Tropical p e a erlag, Berlin 
alley, T. M., G. Weller 
e of ue uta in angiosperms. 
301-369, 
9 T W. 
Annual Rev. 


Coutinho, L. M. . Fire 


vado. Fu Goldhammer (editor), Fire in 
2002. The 
TREE 17 


K. Sakai. 


~ 


1973. es relations of aquatic insects. 
: 183-200. 
Cure, J. R. 1989. Revisão de 1 Schrottky e 


Entomol. 


descrigáo de Oragapostemon, gens n. 
hietidae). Revista Brasil. 
Eiten, G. 1970. A vegetacao do E "wa de S 
Inst. Bot. (Sáo Paulo) 7 1147. 
92. Natural Brazilian vegel ation types and their 
Y 9 N 
K 


Endress. P. 


(Hymenoptera, Ha- 
Entomol, 33: 220-335, 


— 


sao Paulo. Bol. 


causes. Brasil. 64: 35-65 


. Diversity s uolumus biology o 


tropical lose TES 8 Univ. Press; Cambridge. 


Faegri, K. L. van der Pil. 1979. The peer of 
Ss ecology. Pergamon Press. New Yor 

Faria, G. M. 1994. A Flora e a Fauna Apícola de um 
Leossistema de Campo is stre, Serra do Cipó - MG, 
Brasil: Composição, Fenologia las ioe 
Ph.D. thesis. Univ. Est 0 Tou. "Rio Claro. Sáo 
Paulo. 


Feinsinger, P'. & R. K. Colwell. 1978. Community organiza- 
lion among Neotropical feeding birds. Amer. Zoologist 18: 
779-795 


ke 3 B 


M. S. Armbruster, P. Wilse . Dudash & 

J. | Thomson. 6 Pollin: 9 ded and floral 
spec latin Annual Rev. Ecol. Evol. Syst. 35: 37540: 

Freitas, L. 2004. Mecanismos de auto-pe olinizacà 'ào espontá- 


ntianaceae em i 


€ 


ea em cinco espécies de ( 
campo de altitude. Pp. 1-8 in 


Congresso da Sociedade Botânica de São Paulo. Ubatuba, 


Sao Paulo 


Volume 93, Number 3 
2006 


Freitas & Sazima 515 


Pollination in a High-Altitude Grassland 


Nectar features in Esterhazya 


= 


Sazima. 2001. 
macrodonta, eG pollinated Sc 1 eae 
in southeastern rn Brad jl. J. Pl. Res. 114: 18 9] 
2003a. poe biology em pollination 
meds end in two species—from nectar to pollen 
Bound Ann. Bot. (London) 9 311—317. 
———À 003b. Daily blooming pattern. and 
pollination by m in Sisyrinchium vaginatum (lrida- 
5 130: 


in southeastern. Brazil. J. Torrey Bot. Soc. 131 


55— 5 


zi 


EM 
8 


& M. 2000. 


hummingbirds and bees in eight syntopic species and 


Galetto Sazima. Pollination by 
ative hybrid of Ericaceae in southeastern Brazil. PL 
. Evol. d 49-61. 

E A. M. . R. Pirani. 1988. Patterns of geographic 
distribution Pa some plant species from the Espi 
Range, Minas s Ge rais and Bahia, Brazil. Pp. 39-69 in X 

P. F (editor 


a e 


md 


. Vanzolini 
of a shop on Neotropical Distribution Pattern 
Academia y Brasile ‘ira de Ciências, Rio de Janeiro. 
Goldblatt, P., J. C. 1995. 
Pollination biology of Lapeirousia subgenus Lapeirousia 
Africa; 


adaptation for 0 fly pollination. Ann. Missouri 
—5: 


Manning € P. Bernhardt. 


(Iridaceae) in southern floral divergence and 
Bot. Gard. 82: 5 


Shepherd. 


. J. 1998. aa on the 
hol of im rp eae in “cerrado 
. Evol. 211: 13-29. 
R. t 9 5 o 
Expressão e C ~ Rio de Janeiro. 
Heithaus. E. R. 197 
tions in dete rmining community structure. Ann 
Bot. Gard. 75—091. 
1970 


s T visitation records and resource overlap 


Grantsau, beija-flores do Brasil. 
. The role of plant-pollinator interac- 


Missou iri 


of bees = wasps in Northwest Costa Rica. Brenesia 10: 

9-52, 

Herrera. C. M. 
insect 1 a devils's advocate approach. 35—01 
in D. G. Lloyd € S. C. H. Barret (editors), | 


Studies on Floral E volution in Animal-pollinated Plants. 


1906. Floral traits and plant o ition to 

See ae 

Chapman & Hall, New 
.S. D. & 


. E. Steiner. 
spec cal ation in em pollinator systems. TR 
10-143 


Johnsor 2000. Generalization versus 
EE 15 


cs Specialize n Laie) systems in 


southern Africa. S. African J. Set. 345-348. 


Kearns, C. A. D Akoh ^ 1 across an 
ele Sion gradient. Amer. Midl. Naturalist 127: 172-182. 
¿ D. W. Inouye. 1993. 


Techniques for Pollination 
Niwot. 
rnaces in high arctic 


mi 
= 


Biologists. Univ. Colorado Press, 
Kevan, P. G. 1975. Sun tracking solar fur 
flowers: significance for pollination and insects. Scien 
189: 723-126. 
& H. G. Baker. 1983. 
¡ads Annual Rev. Entomol. 
, E & M. Usui. em 
Sha in ike pollination ecology of the boreal zone. 
Re S. 8: 247-46 M. 
Kóppen, W. 1948. Climatología: con un estudio de los climas 
de la tierra. F. C. E., 


Ins 11 5 as dp visitors and 
1—453. 


= 


í De 
Insects and 


Ecol. 


Tikhmenev 


Ciudad de Méxi 


Kornerup, A. & J. H. Wanscher. 16 "g l'aschenlexikon der 
Farben. Musterschmidt, Zürich. 
Tess. W. «€ J. H. Beach. 1994. Mure plant 


A. McDade, K. S. Bawa. 


re podig tion. Pp. 161-182 in L. 
I a Selva. 


J. A. Hespenheide & G. S. Hartshorn 2 


Univ. Chicago Press, Chicago. 


Ledru, M.. M. L. Salgado-Laboriau & M. Lorscheitter. 


1998. Vegetation dynamic 
oe the last 10,000 years BP. Rev. 
z 1B31- 


in south ut central. Brazil 


Palaeobot. Palynol. 
142 
"Nut F. N. 1959, Staining and observing pollen tubes 1 in 
the style by fluorescence. Stain Tech. 34: 
d 
Martinelli, 


de Janeiro. 
1994. Reproductive Biology of Bromeliaceae in the 
Pl 


means ol 


. 1989. Campos de Altitude. Editora Index, Rio 


kiano Rainforest of Southeastern Brazil. Ph.D. thesis. 


niv. Saint A Andrews, Saint / Andrew 


— 


Biologia reprodutiva 7 o na 
Reserva Ecológica de Macaé de Cima. Pp. 213-250 in H. 
C. Lima & R. R. Guedes-Bruni (editors). Serra 3 Macaé 
de Cima: Diversidade Florística e Conservação em Mata 
\tlantica. 
Janeiro. 


McDade, L. 


Rio de 


Jardim Botânico do Rio de Janeiro. 


. 1992. Pollinator re PUE biogeography, 
20. 


and 1 netics. BioScie 2: 21— 

Modenesi, M. C. 1988. Signifie en p de 19 correlativos 
Hi enel em Campos do Jordi - Sao Paulo: 
im] lica 'acóe leocli e] | lógi . Bol. Inst. 


y 1-155. 
9 5 m A. R. 1976 


California M ecology. and 
cgetation types. Phytologia 34: 305-361. 
& P.C 


». Lincoln. 1979. Pollination ecology in 

montane Colorado. Phytologia 42: 549—379 
Momose. K, T. Yumoto, T. Nagamitsu, M. Kato, H. 
Nagamasu, S. Sakai, R. D. Harrison, T. Itioka, A. X. 


1998. Pollination biology in a lowland 
Characteristics 


Hamid & T. Inoue. 


dipterocarp forest in Sarawak, Malaysia. I. 
of je Pas pollinator community in a lowland dipterocarp 
Be 


forest. Amer. X. 85: 1477-1501. 
Moreira, A. A. N. & C. Canale 1977. Relevo da região 
ds ste do Brasil. Rev. Brasil. Geogr. 34: 1-50. 


Nakano, C. & J. 1 2003. Variability and special- 
ization of plant- ed systems in a northern maritime 
grassland. Ecol. Res. 18: 221-246. 

Nimer, E. 1977. 


Rev. Brasil. € 


1 da regiao sudeste do Brasil. 
34: : pa 8. 

>, Jordano. 2002. Geographie patterns in 

Ecology 83: 


„Or. 

Olesen, J. M. & 

plant- 70 tor mutualistic networks. 
2416-2424. 

Oliveira, 4 E. & P. E. 
woody plants in a cerrado community of Central Brazil. 
Flora 195: 311—329 

1996. 

patterns: the 5 1 of plant 

. Ecol. 84: 767— 

19 50 1993, 


biogeographic comparison. 


Gibbs. 2000. Reproductive biology of 


Ollerton, J. Reconciling ecological with 


phy 


processes 


ogenetic 
ollinator systems. 
D. W. 


pollinator visitation rates: a 
257-26: 


Primack, R. B. & ioa affecting 
5 9 ) 
Yoo & A. Natural History 
of ination Timber Press. 
Radford. E., W. C. Dic 
1974. > asc bs Plant Systemalics 


, New Yor 


Proctor, Lack. 1996. The 
Portland. 
R. Mas & C. R. Bell. 


e $ R ow Publish- 


'kinson, J. 


1989. 


Ramírez, N. Biología de polinización en una 
comunidad arbustiva 1 de la Alta Guayana 
T 1 ana. iiid a 21: 319—330 
. Re produc tive e ina tropic cal shrubland 
of m ue 1 ;uayana. J. Veg. Sci. 4: 5-12. 
2002. Pollination specialization and time of 


solli tion on a d Venezuelan plain: variations in 
Bot. J. Linnean Soc. 145: 1-16 


time and space. 


516 


Annals of the 
Missouri Botanical Garden 


. Ecology of p in a tropical Venezue- 
i | 
: 171-189. 


200 
lan savanna. PI. Ecol. : 
K Y 5 biology a 


Drito. t 


swamp — in the Venezuelan Central P lane Bol. 


ilm 


J. Linnean . 110: 27 Qs 

Reis. M. G., \. pus p Faria, V. Bittrich. M. C. E. Amaral & A. J. 
Marsaioli. 2000. The e > ‘mistry of qe T re san Dni 
(( Irchidaceae). J. Brazil. € hem. Soc. 11: 600— 

Renner, S. S. € J. P. Feil. 10 alin e oe 


: 1100-11 

The niche exploit ition 5 rn of oH i lue- 

317-350. 

989. Ecology and Natural History of Tropical 
lg 


dioecious angiosperms. Amer. J. Bot. 

Root, R. B. 1967 
gnalc he r. Ecol. 
Roubik. DAN 
Bees. Candid Univ. Pre 
1999a. 
to the 9 al environment on vegetation of the campos 
de oe Je 712 
9 


». Brazilian ed IL 


Monogr. 37: 


Can nbric at 


Salford, Brazilian Páramos. n An introduction 
Biogeogr. 26: 09: 2. 


lacro- and mesocli- 


= 


mate ee se campos de altitude and affinities with high 
mountain climates of the tropical Andes and Costa Rica. J. 
Progeogr 26: 713-731. 

. 2001. B mazilian Páramos. III. Pa 
E regeneration in the eampos de altitude. Biotropica. 
236 —302. 

Pune p . Buzato & M. Sazima. 1995. 
hermit Ramphodon naevius 1 its flowers in southeastern 
Brazil. J. Ornithe l. 136: 195-206. 

1996. An 
— E pollinated flowers in ¿ 
Acta 25 


é 19 


The saw-billed 


——— assemblage of 
montane forest in 
149—100. 

Bat-pollinated flower 


southeastern Brazil. Bot. 
K 


= 


asse bla ages and bat visitors à two Atlantic forest sites in 


Brazil. . Bot. 83: 705-712. 


K C. Hausner. 1993. The 


ES sli A. 
perfume flowers of C vom (Solanaceae): pollination 


Cocue 


by euglossine bees, bell hanism. osmophores, and 
volatiles. Pl. Syst. d 87: 51-88 

Schemske, D. W. & 1 1984. Variation among 
floral. visitors 


in lito ability 

mutualism s ur exar Science 225; 5]9— 
Segadas-Vianna, F. & L. Dau. 1905. 
soutien Brazil. II. Cli 


a precondition for 
52] 


Ecology of the Itatiaia 

range, imales and altitudinal 

clima Arq. Mus. ps ional 53: 
N 


Silberbauer-Gottsberger, 


lic zonation. 31-53. 


^ 
— 
= 
e 
= 


A 
polinizag āo de N do cerrado. Revista Brasil. Biol. 
663 


18: 651— 


G. Got 1 1 l 


Herns and rates of 


Silveira, F. X. & J. 
southeastern Brazil: 


1993. High-altitude bee fauna of 
implications for. bi 8 pal- 
\poidea). Stud 


R. Cure. 


Neotro Fauna 
€) 


terns (Hymenoptera: 
Environm. 28: 47-55. 


R. B. 2001. 


parviflora ( 


Singer, The d biology of Habenaria 


(Orchidaceae: oe in southeastern 
Darwiniana 39: 201—2( 
& A. A. Cocucci. 1999, 
four sympatric ones rn Brazilian Epidendroideae orchids. 

agar 114: 47- 

Smith, X. P. & T. P. Young. 1987. wi dues UNS plant 
ecology. pM n Syst. 18: 137- 

Stebbins, G. L. 1970. Adaptive radiation E pM 
caia in x I: 
1 . Ecol. Syst. 1: 307-320 

Pruylio, ! P dunt Mt & W. 2002. Biologia 
floral « iiti: ào de ia. micranthum (lrida- 


Drazil. 


: ibus mechanisms in 


Ecol. 
po mechanisms. 
p mds 


ceae) na Região do Planalto das Araucárias do Rio Grande 

do Sul, We ec 10: 11-24. 

Van Schaik, C. P., J. W. Terborgh & S. J. Wright. 1993. The 

phenology i 1 5 forests: adaptive e and 

consequences for primary consumers. Annual Rev. Ecol. 
Sy St. ue 995—911. 

Vogel, 1990. History of the Fou ea eae in the light 8 
TA ecology. Mem. New York Bot. Gard. 
130-142. 

1998. 


a perspectives. II. 
x 


temarkable nectaries: structure, ecology, 
Nectarioles. Flora 193: 
— 8 C. 


lactor of 


Westerkamp. 1991. Pollination: an integrating 
biocenoses. Pp. 159-170 in 


Seitz & 
Loescheke (editors), Species ec aes a Population- 
biological Approach. Birkhäuser Verlag. Basel. 

Wainwright, C. M. 1978. The f 
ecology of two desert lupines. Bull. Torrey Bot. Club 105: 
24—38. 

Waser, N. M.. L. Chittka, 
Ollerton. ae sene ST in sm systems, = 
why it matters. Ecology 77: 1043- 

Yeo, P. F. 1993. Pollen 


Function and Evolution. Springer (Plant Syst. Evo 


floral biology Pie pollination 


N. M. Williams & , 


. Price, 


Sec VEN 2 resentation: Form. 


. Suppl. 
6.), Vienna. 

Zeisler, M. 1938. Über die 
Narbenflic a mit Hilfe von 
Gentralbl. 58: 308-318 


Abgrenzung der eigentlichen 


Reaktionen. Beih. Bot. 


EMBRYOLOGY, KARYOLOGY, Christoph Dobes,++% Marcus Koch,’ and Timothy 
AND MODES OF REPRODUCTION F. Sharbef 

IN THE NORTH AMERICAN 

GENUS BOECHERA 

(BRASSICACEAE): A 

COMPILATION OF SEVEN 

DECADES OF RESEARCH! 


ABSTRACT 


The North American-Greenlandie genus Boechera A. Live & D. Löve (Brassicac i is distinguished from other 
phylogenetic lineages recognized within the artificial taxon Arabis L. s.l., in which it has formerly been included, by its base 
chromosome number x = 7. Based on outgroup comparisons, we consider this boom number to be derived from an 
ancestral genome sce eight chromosomes. Chromosome counts are now available for about half of the species names 
listed in the most recent taxonomic treatments of the genus. An analysis of the karyological differentiation with respect to 
bers was performed, and these patterns were correlated with data on pollen viability and e genetic 


chromosome num 
groups: sexual, amphi- 


measures in order to de dis modes of reproduction. This approach allowed us to distinguis 1 three main; 
apomictic, and apomictic species, respectively. We focused further on the cytology of gametes, espe an on that of alk n, 
and we discuss its role in the formation of new genotypes and eytotypes. Evidence for ihe following e mn phenomena is 
currently available: (1) allopolyploidization, (2) homoploid hybrid formation, and (3) establishment of viable offspring by 
means of fusion of highly aneuploid gametes, which resulted from irregular meiotic divisions. In contrast, there is no strict 
evidence demonstrating the occurrence of autopolyploidization. Boec hera experts have just started to integrate eve lutionary 
concepts derived from molecular analyses into taxonomic research, and profound changes, with respect to species definition 
expect ted in the near future. The identified modes of reproduction and evolutionary phenomena 


and circumscription, are to 
axonomic classifications and species 


will contribute to the deve bue nt of a theoretical and practical basis on which future 


conce pls can be based. 
“wore agamospermy, Boechera, Brassicaceae, chromosome number, embryology, evolution, hybridization, kary- 


1 1 W genetics, speciation, taxonomy. 


The genus Boechera Á. Lóve & D. Lóve (Bóchers | Shehbaz, 1997; Koch et al., 1999), Fourraea Greut. et 
rock cress, Brassicaceae) has recently been recog- Burdet (A. pauciflora (Cima) Garcke: Koch et al. 
nized as a monophyletic isolated lineage (Koch et al., 1901, 2001), Pseudoturritis Al-Shehbaz (A. turrita L.; 
1999; Koch, 2003) of North American—Greenlandic Koch et al. 1999, 2001), and Turritis L. (A. glabra L.: 


distribution, based on DNA sequence variation. The Koch et al., 2001). Based on the occurrence of two 
genus, which according to current taxonomic treal- base chromosome numbers within Arabis s.l., Love 


ments (Rollins, 1993; Mulligan, 1996) comprises and Lóve (1975) proposed the splitting of this aar 
about 80 species, was formerly included within distributed genus into a New W orld taxon (x 7 
broadly defined genus Arabis L. s.l. The latter is Boechera) and a second group chiefly of Old World 


highly polyphyletic, containing species that are to be distribution (x = 8, Arabis s.s.). This classification is 
treated within the genera Streptanthus Nutt. (A. supported by several morphological characters (Al- 
petiolaris A. Gray; l. Al-Shehbaz, pers. comm.). Shehbaz. 2003). 

Pennellia Nieuwl. (e.g., A. microsperma Rollins; Al- Recently Koch et al. (2005) identified a basal 


Shehbaz, 2003), Arabidopsis Heynh. (O'Kane & Al- phylogenetic split within the Brassicaceae based on 


— 


! We are grateful to Ihsan Al-Shehbaz, Missouri Botanical Garden, and Eric Schranz, Max Planck Institute for Chemical 
E elos for critical reading of the manuscript and for helpful discussion. Additional thanks go to Michael Windham and 
second anonymous reviewer for extensive comments and substantial improvement of the manuscript. We also thank Victoria 
Hollowel for her editorial impact covering various taxonomic issues. 

Heidelberg Institute of Plant Science, De c a of Biodiversity and Plant Systematics, Heidelberg University, 


Heidelberg, 1 cdobes(hip.uni-heidelberg.de. 
* Apomixi rch Group, Department of Ce netics, Institut für Pflanzengenetik und Kulturpflanzenforschung (IPK), 
D- 06466 qd Germany. 


* Corresponding author. 


ANN. Missourr Bor. GARD. 93: 517—534. PUBLISHED ON 24 OCTOBER 2006. 


518 


Annals of the 
Missouri Botanical Garden 


the occurrence of a pseudogenic tandem repeat found 
in the trmmlLvoaptrnF ove intergenic spacer. The 
repeated region constitutes a synapomorphy support- 
cardaminoid and arabi- 
2002). with the 


an evolutionarily derived 


ing a clade comprising the 
dopsoid Brassicaceae (Heenan et al., 
genus Boechera placed 
The base chromosome numbers of sister taxa 
Al-Shehbaz. O Kane & R. A. Price 
Medik.) and of emerging 


deeper from ils evolutionary history are primarily x 


position. 
(Crucihimalaya 
and Capsella branches 


= 8. The phylogenetic pattern of cytological differ- 
entiation shown in Figure | provides strong evidence 
that 


axon, 


the ancient base chromosome number of the 


and 


which has given rise to the cardaminok 


arabidopsoid Brassicaceae was indeed x = 8. This 
ancestral base chromosome number is furthermore 
supported, at least in the arabidopsoid lineage, by the 
high degree of collinearity 
Capsella rubella Reut. and Arabidopsis lyrata subsp. 
petraea (L.) O'Kane & Al-Shehbaz (both 27 = 16) as 
has been demonstrated by Koch & Kiefer (2005) (cf. 
fig. 1). Consequently, the x — 7 condition in Boechera 
must be considered to be a derived condition resulting 
from chromosome reduction. Comparative map-based 


studies are now available that explain structural 


changes in plant genomes, as has been done in the 
case of the highly derived chromosome set (x = 5) of 


(Acarkan et al.. 


have nol yel been 


Hevnh. 


maps 


Arabidopsis. thaliana (L.) 
2000). 


constructed. for 


Such genetic 


Boechera, although several hundred 


microsatellite loci have been isolated and the 
construction of linkage maps is currently underway 
(T. Mitehell-Olds, 
vances in the mapping of 
(Kowalski et al., 1994; Lagercrantz, 1998; Kuittinen 
2004; Lysak et al., 2003, 2005), we can expect 


these data to provide resolution on the structur: al 


pers. comm.). From recent ad- 


Brassicaceae genomes 
et al., 
changes responsible in the origin of the distinct x = 7 
base chromosome set characterizing Boechera. 

The genus Boechera has long been known for its 
polymorphism, which is 


extensive morphological 


thought to have resulted from the combined evolu- 


tionary forces of hybridization, polyploidy, and 
apomixis (Rollins; 1966, 1983: Dobeš et al., 2004a. 
2004b, 2000). However, during the past seven 


decades research activities have focused mainly on 
taxa: B. divaricarpa (A. Nelson) Á. Lóve & D. 
B. holboellii (Hornem.) Á. Lóve & D. Lóve, and 
the only one that is a proven P UE species, B. 
stricta (Graham) Al-Shehbaz (fi 
Only 


evolutionary patterns and processes from the remain- 


three 


Love, 


Arabis drum- 


mondit A. Gray). a data on the 


ing species of this genus have been gathered. 
The aim of this article is to synthesize the available 
population. genetics, 


data on karyology, palynology, 


between the genomes of 


reproductive mode, and associated cytological path- 
ways on a genus-wide level, with a special focus on 
the resulting chromosomal composition of gametes 
and its evolutionary significance in the establishment 
of new genotypes and cytotypes, respectively. 


Because of the enormous. taxonomic. difficulties 


within this genus, one has to take into account that 


many past documentations refer to incorrectly identi- 
fied material. For this reason, we are introducing a web- 


based. gateway for ongoing Boechera and Brassi- 
e E eo D 


caceae research ( (<http://ephedra.hip.uni-heidelberg. 
lel >) 


aGae/Drassicacedae- 


Through this gateway, all re- 


examinations of vouchers. referring. to cytological 
references provided during the past seven decades 
will be accessible. This is also true for the summarized 
and discussed data of this review. In the future, we aim 
to open and link several other database resources on 
crucifer research (Ihsan Al-Shehbaz and co-workers, 
Michael Windham and Donovan Bailey). 

Although the cytological and embryological mech- 
anisms and phenomena addressed here may be invalid 
they are applicable to 


taxonomic context, 


in a 
evolutionary questions (i.e. they mirror evolutionary 
processes occurring in the genus Boechera). Conse- 
quently, they contribute to a theoretical and practical 
basis on which we can build evolutionary models and 
on which we Can develop taxonomic and species 
concepts in the near future. Therefore, we recommend 
reading of this review in an evolutionary context, even 


though we, for practical reasons, have to refer to the 


only currently available, but highly artificial, taxo- 
nomic Classification. 
SOMATIC CHROMOSOME NUMBERS 

Karyological studies in Boechera have focused 


mainly on the estimation of chromosome numbers, 
and karyotypic analyses of their small chromosomes 
(3-6 um) have recently been performed (Sharbel et 
al., 2004, 2005). 
structural peculiarities of the Boechera genome. For 
Mulligan (1964: 
North American 


species have one pair of somatic chromosomes with 


These studies have revealed several 


instance, 1517) stated that diploid 
native Arabis? (genus Boechera) 
satellites; while triploid plants of A. divaricarpa A. 
Nelson had three somatie chromosomes with satellites. 
Sharbel et al. (2004) described the diploid genome of 
B. holboellii to be composed of autosomes, which have 
median to subterminal centromere positions. 

Another conspicuous feature is the presence of 
chromosome fragments) 
holboellii 


la and Ib) and that can be differentiated from 


aneuploid chromosomes (or 


that were mainly observed in Boechera 


(Table 
the normal autosomes by their conspicuously smaller 
size (Bócher. 1947 1964: 1964: 


Mulligan, Packer. 


100 DIT lonopsidium prolongoi Batt 
lonopsidium abulense (Pau) Rothm. 
Cochlearia danica L. 
ochlearia pyrenaica DC. 
aL. 


‘cop. 
Aubrieta deltoidea (L.) DC. 
ieta deltoidea 


Arabis turrita L. 

Microthlaspi perfoliatum (L.) F.K. Mey. 

Thlaspi a el. 

Alliaria 0 (M. Sieb.) Cavara & Grande 
Sinapis alba L. 


xs sativus L. 
Sisymbrium 
Fourraea "iin e ) Greuter & Burdet 


68| Barbarea vulga 
Rorippa PA " * Besser 
100 n ? P 
F i 8 Cardaminoid Brassicaceae 


Rorip 


* 
I 


mine amara 
e penzesii M.E. Ancev & Marhold 
Cardamine rivularis Schu 
idium Ai ?" ) R. Br. Xx S 


Lep. 
= petraea (L.) O'K 


2 Al-Shehb 


^ 
| | Ara bidopsts Ita ssp. petra x=8 
L| dopsis lyrata ssp. lyrata (L 105 Kane & Al-Shehbaz 
92 (OC bites he alleri (L.) O'Kane & Al-Shehba: 
| | Arabidopsis thaliana ln ) Heyhn. 
| 
E Turritis iia aL. l - =6 u -— 
| Oli bid bul 55 Al-Sheht O'Kane & RA Pri x=8 


i Arabidopsoid Brassicaceae 


Capsella Abel Reut 
— Boechera lyallii (S. Watson) Dorn) 
- Boechera parishii (S. Watson) Al-Shehbaz 


— Boechera ~ 


Weber 
res Halimolobos ! (L. F. Hend.) Rollins X27 


O'Ka R Pri 
pis ae x =8(7, 9) 
h I Hook.f. & Thomson) Al-Shehbaz, O'Kane & R. A. Price 
Matthies incana (LOR 
Aethionema ee Boiss. & Hohen. 


0.05 substitutions/site 
Fig l. Phylogenetic relationships among cruciferous plants based on chs and matK sequence data (redrawn from Koch et al., 2001, reprinted with *** e The Botanical Society 
of America). The asterisk marks the initiation of a fru pseudogene tandem repeat at the basis of the cardaminoid and arabido psoid Brassicaceae (Heenan et al., 2; Koch et al., 2005). The 
ogenetic ds e ME aber among the cardaminoid and arabidopsoid Brassicaceae is taken as evidence that the base chromosome 1 af the progenitor of these 


phy 


lineages was most ike 


9005 


€ JequinwN '£6 euinioA 


(eeeoeoisseig) e/19y990g ueoueuly YHON 


e 1e SO 


618 


Table la and Ib (pp. 520-524). Evidence for sexual and apomictie reproduction taken from karyological investigations, analyses of pollen, estimation of seec 
reported for Boechera species and affined taxa. Species names are according to Rollins (1993): taxonomic ranks are acc -ording to Mulligan (1996). 
Boechera (cf. Al-Shehbaz, 2003. for synonomy with Arabis). four Arabis species are also included. Arabis petiolaris and A. serotina probably 
were included for clarity. The remaining two Arabis species | 


fertility, and population studies 


Although most species are treated within genus 


have to be transferred to other genera (see below) but 
iave not yet been formally transferred to Boechera, but are included based on the similarity in genus cytology. Superscript index 


numbers refer to the publications listed in the footnote. Gray fields signify features associated with apomictic reproduction and species showing evidence of apomixis. 
Table la 
Taxon Chromosome number? Meiosis* Type ot Fer Cr bait y Population 
Pollen genetic 
— 22 um) « y = i 
2n= 14 2n P 2 UN T others Sex? Apomeiosis" | Tetrads Monads | % uis pra studies 
+2 +2 Dyads viabilit et 
Arabis boivinii’ G.A. Mulligan 2 2M* 
A. Nelson & P.B. Kenn. 1* M* 
Arabis de 2n = 28? 
Arabis serotina” 19 
B. breweri (S. Watson) Al-Shehbaz 17 90 * 9?! 7] «29?! 
B. canadensis (L.) Al-Shehbaz 11° 
B. cobrensis (M.E. Jones) Dorn p p? 
B. constancei (Rollins) Al-Shehbaz 1 
B. crandallii (B. L. Robinson) W. A. Weber 336 Pp Iso - Sex 
B. demissa (Greene) W.A. Weber 16 poe 
B. divaricarpa 1 l, B: 1 9 4,B:1 2 See T able | b f or d eta ils 
73519 p p” 55 r 
B fenden V 18 : p | | | | 
B. glaucovalvula (M.E. Jones) Al-Shehbaz p p? | | | | | 
3 BER TICO e E > 0 
B. gunnisoniana : 22 E E.P 22 22 22 
- 1 D 58^ 38^ Iso - Apo 
B. holboellii 85 4, B: 5 28 4, B: 6 7 e Table lb for detai s 
B. inyoensis (Rollins) Al-Shehbaz 1 17 
B. laevigata (Muhlenberg ex Willdenow) Al- gm 
Shehbaz 
h 
B. lemmonii (S. Watson) W. A. Weber 3 " à 4 
1 2M abundant 
B. lignifera 3322 p? T? 91? 6” Iso - Apo?” 
B. lyallii ; 541021 2M* 2p? 34 + 282 65 +22 
3.10 3 poen p 
B. microphylla (Nuttall) Dorn 2 j3 12 E p^? p? 


06 


au} jo sjeuuy 


uapsJed jeoiuejog OSS 


Table la and Ib. | Continued. 


900c 


€ 1equinN ‘g6 euinjoA 


B. parishii B 1? p? 
B. pendulina (Greene) W.A. Weber 1? uniform? 
B. perennans (S. Watson) W.A. Weber 7 p? 
B. perstellata (E. Braun) Al-Shehbaz 96.15 Mé” 
B. platysperma (A. Gray) Al-Shehbaz 121 p?! 911)! | 77 «25? 
: 14 pem E un - M* 
B. puberula (Nuttall) Dorn T 9 Mi 
379 P 
B. pulchra (M.E. Jones ex S. Watson) W.A. A a 
Weber 2 
B. repanda (S. Watson) Al-Shehbaz 3 | pP 
B. schistacea (Rollins) Dorn 12? uniform? 
B. selbyi (Rydberg) W.A. Weber c 15 
B. shortii (A. dentata) 2 128 MP 
o p p P.M: 88-975 
B. sparsiflora (including Arabis columbiana) 31721 55 23 = 46 2p?, plé?9 M.D?7? | 62 4 46?! 734257! 
; 2 B. 17 o d 
2n= 
: 32!7* 
B. stricta : 20 2 1 l See Table lb for details 
Ti e x 2 — UMS 5 55 
B. subpinnatifida (S. Watson) Al-Shehbaz qu pe 95 +2” 1 E 182 
+18 
B. williamsii (Rollins) Dorn Mor-uniform? 
B. crandallii x B. holboellii 1° 
B. divaricarpa x B. holboellii 92829 p22? DA 
B. holboellii x B. stricta 15 LE 19 


Je 19 58900 


(eeeoeoisseig) ejaeuyoeog ueoueuly YON 


LoS 


Table la and 1b. Continued. 


0 


au} Jo sjeuuy 


Table 1b 
Somatic Meiosis‘ Type of Fertility eee Citati j 
Taxon l genetic i 
Material f e : 
Chromosome Pollen s[udies! 
denotation a 
number Sex! Apomeiosis* Tetrads Monads | % Pollen % Seed 
Dyads | viability Set” 
B. divaricarpa [15] ("n 2 15”) p^ 3 
16 P [?] 3 
14 M 4,5: 11 
15; 21; 28 2M 4 
14; 22 6 
13 + 2B; 14; 20 + 2B; 21; 10, 31, 33 
28 
1 22 2P 64+38 | 65+19 21 
B. holboellii Colorado 21 V E (0%--5%) E (95%-- 95 l 
100%),P 
including Greenland 14 E (11%--55%), | E (45%--89%), | Mixture of reduced and 33 l 
E P P unreduced pollen 
Arabis holboelli H / 
a e elt ornewahn] — Greenland ncc E (2%--17%) | E (83%--98%) 50 
A. collinsii Fernald 
collinsii Ferna 12 biotypes 14 P 92--99 2 
A. exilis A. Nel 
abd MR 2 biotypes 28:28 + If P 89--94 2 
A. pendulocarpa A. Nelson 14 P.M 3-13 36 
A. pinetorum Tidestr. 
[14] (‘n= 14”) 3 
A. retrofracta Graham E 
[21] (“n = 21") Pp 3 
14 M (mostly) ? M (few) 4 
C21 2M D 4 
ODOREM —— | mÈ 4, 29 
15 M 2M 4 
c. 22 : — | ?M 4 


USPIBE) [eoiuejog LNOSSIIN 


Table la and 1b. | Continued. 


14 Mor-uniform 5; 6, 10, 28, 
29, 30, 31, 32 


21 — 6.102.829 


9005 


2111 7 
14 (n 2 c.8 in a few cells) P WP 8 
13 + 2B 10 
20 ＋ 2B Doe 10 


14+ IB, 14 « If 10, 12. 30 


€ JequinN 'eg euinjoA 


14 (n =7; 9) P ?P 32218 7213 21 

5 46 22 

21+1f D (42)88-- | 38--83.3 | Mol Apo?” 22 
o | 95 


ie] 


28 


„„ D 37-97 | (1557- : 22 
. 17 


86--92 32--51 22 


Disko 1 
Disko 2 
S. Str. 10 


e 1e 59900 


S. Str. 3 
S.Str. 9 


Plant fertile, one 27, 28 
completely sterile shoot 
Eqaluit 
Alaska 
No 3090 


Normal 26, 27, 28 


No 2820 35 (of accidental origin) 28 
No 2820 2x vp : D 28 


(9oe9a9e91sselg) 2/9y990g ueoueuly YON 


£cS 


Table la and Ib. | Continued. 
No 6 214 1f lee | P oD 29 
No6 17 shrivelled 29 
No 12 14 M T 29 
No 16 21 + 1B or 20 + 2B 29 
No 20, A-Batch 14; 14 + 1B; 16; 17 T D 29 
No 20, B-Batch 21; 21 + 1B M T D 29 
22 39 
B. stricta 14 M,P 98 2,4 
(A. drummondii) 14 P 3,8,34 
21 „55 2M 4 
14 5; 6, 10, 31 
20 = 10 
28 P Reduced 29 
“Number of published ch lassified di ] th f additional ch . Somatic triploidy was taken as evidence for apomictic 


g ploidy 1 
reproduction. " B indicates that the additional chromosome(s) was a B chromosome. E and P denote female or male gametogenesis, respectively; "M" signifies that the gender of the gametophyte 


has not been stated by the authors.* The count 27 = 32 for B. da was based on a misidentified true Arabis (M. Windham, pers. comm.) and, therefore, has not been considered in further 


statistics provided in the manuscript. ^ Cell division listed in this col d cl istics indicati plete, regular pairing of chromosomes at meiosis 
(bivalents); regular meiotic di l ith red d ch ber. Crisscrossed fields underline that, so far, evidence for normal regular meiosis d trated i 
association with e Divisions listed in this colu howed characteristics indicative for ap d desis, apomeiosis, gametophyte with unreduced chromosome number. 
Aq di irregular, often i plete, pairing of chromosomes and irregular divisions, lini in FOE with irregular chromosome numbers. D, M, and T denote 
dyadic, monadic, and tetradic pollen. "Reduced" and "Unreduced" refer to chromosome number. * The p ble pollen grains is provided; estimat f fertility of pollen produced by 


sexuals are given in bold. " Values provided by Vorobik (1985) refer to fruit set, while those given in Roy dios) are the percentage of germinating seeds. ' ! Evidence for sexual ("Sex") and 


apomictic ("Apo") reproduction taken from population studies; “Iso” ref l g “Mor” are observations on the variability of morphotypes. Records of 
identical information content have been merged. " This species may belong to the genus Pennellia (Al-Shehbaz, pers. com.). * This species may bel th Streptanthus (Al-Shehbaz, pers. 
com.; Koch & Dobeš, unpubl. data). ” This species has formally not been transferred to Boechera. Therefore. it is 1 d as an Arabis. ^ Pollen was highly vanable among 


1 fi A E 


1) p while the majority of pollen grains were of irregular si d sh 2 fl f d dyad poll ly. Numbers in superscript 


uepier) jeoiuejog OSN 


ves 


əy} JO sjeuuy 


Volume 93, Number 3 Dobes et al. 525 
2006 North American Boechera (Brassicaceae) 


Bócher, 1969; Johnson, 1970; Rollins & Riidenberg, 


1971). Recent work has demonstrated these extra 


ob 

— 

E m EP. chromosomes to be B chromosomes (Camacho et al., 

S 5 doc 2000). as is evidenced by their interspecific origin, 

- 2 = o y a ee ; E 

a 3 a variable morphology, high content of heterochromatin, 
— — . TE » 

2 1 = 3 and independent evolution from the | autosomes 

E pa „ (Sharbel et al., 2004). Interestingly, the B chromo- 

ril i somes have a wide geographic distribution and can 
— [ex] ^ 4 . š . . 

PE e be found in both diploid and triploid Boechera 

e Z = E o 

S E g (Sharbel et al., 2005). Furthermore, DNA sequence 

— 0 e 

1 * e analyses of B chromosome alleles suggest a single 

= 3 origin for these elements (Sharbel et al., 

o [22] " 

— m However, chromosome number is cn most in- 


tensively studied karyological property of Boechera 
that has been studied (Böcher, 1938; Smith, 1938; 
Rollins, 1941; Böcher, 1947; Böcher & Larsen, 1950; 
Böcher, 1951, 1954; Rollins, 1960; Easterly, 1963; 
Mulligan, 1964; Packer, 1964; Raven et 1965; 
Rollins, 1966; Taylor & Broekman, 1966; Böcher, 
1969; Lóve, 1969: Johnson, 1970; Rollins & Rüden- 
berg, 1971; Love, 1975, 1976; Rollins & Rüdenberg, 
1977; Taylor & Taylor, 1977; Kovanda, 1978; Rollins 
& Rüdenberg, 1979; Lóve, 1979, 1982; Rollins, 1983; 
Ward, 1983; Vorobik, 1985; Wieboldt, 1987; Dal- 
gaard, 1988; Stace, 1989; Roy & Rieseberg, 1989; 
Roy, 1995; Mulligan, 1996; Stace, 1998, 1999; 
Naumova et al., 2001: Sharbel et al.. 2004; Tsakin 
et al., 2004), the data for which is now aon for 36 


species (34, if Arabis petiolaris and A. serotina E. S. 


6 
= (Packer, 1964); 13 
(Ward, 1983); 20 = (Wieboldt, 1987); 21 = (Vorobik, 1985); 22 = (Roy, 1995); 


(Bócher, 1951); 28 


(Lóve, 1982); 32 = (Lóve, 1979); 33 = (Lóve, 1975); 34 = (Lóve, 1976); 35 = (Love, 1969); 36 = (Dalgaard, 1988); 37 


— 


2 
(Bócher, 1954); 29 = (Bócher, 1969); 30 


(Stace, 1998); 38 


(Rollins, 1983); 


Steele are excluded from the genus), comprising about 


half of the Boechera species recognized in the 


(Mulligan, 1964); 11 = (Taylor & Brockman, 1966); 1 


(Taylor & Taylor, 1977); 18 2 (Kovanda, 1978); 19 


monography of Rollins (1993). Much of the current 


taxonomy in Boechera may be expected to be subject 


to far-reaching revision in the near future (Windham 


Johnson, 1970); 3 = (Rollins, 1941); 4 = (Mulligan, 1996); 5 


et al., 2004), but there is no doubt that x = 7 


( 


& Rüdenberg, 1977); 9 = (Rollins & Rüdenberg, 1979); 10 


1963); 15 = (Rollins, 1960); 16 = (Raven et al., 1965); 17 


constitutes the common and probably only base 


chromosome number of the genus. The only taxon 


deviating from this rule may be B. shortii (Fernald) Al- 
Shehbaz (A. dentata Torr. & Gray), for which a single 
count of n = 6 Il (bivalents) was published by Smith 
(1938). A total of 289 chromosome records (excluding 
counts reported repeatedly and for taxa probably 
belonging to other genera) are currently. available. 
This total is strictly limited to those taxa currently 
recognized as Boechera s.s. (cf. Al-Shehbaz, 2003) and 
to direct observation. of chromosome number rather 


than inference by flow cytometry or pollen size. In 


Continued. 


3 = (Roy & Rieseberg, 1989); 24 = (Bócher, 1947); 25 = (Bócher, 1938); 26 = (Bócher & Larsen, 1950); 27 


refer to the following publications: | = (Naumova et al., 2001); 2 


E order to minimize propagation of any error in species 

E S recognition, we do not provide here the original data, 

<a c E which were collected in a taxonomic context. Instead. 
E : = : in addition to Table la and lb, an annotated and 
3 E S 2 fully documented database summarizing published 
E EN aS ë data on karyology, pollen, and population genetics is 


available at <http://ephedra.hip.uni-heidelberg.de/ 


brassicaceae>. (In both data files we applied the 


526 Annals of the 

Missouri Botanical Garden 
taxonomy and nomenclature suggested in Rollins, duplication of the somatic chromosome number with- 
1993, Mulligan, 1996, and Al-Shehbaz, 2003, re- in an individual has nonetheless been reported from 


spectively.) These databases are continuously updated 
in regard to taxonomic identification. In addition, two 
that 


formation. will be 


databases might be useful for additional in- 


accessible in the future. A 


chromosome number database is compiling most 
references of chromosome counts in eruciferous plants 
(Warwick & Al-Shehbaz, 2006), and a 
database is listing many laxonomic 


taxonomy new 


treatments (Warwick et al., 2006). The most recent 


findings on Boechera are nol ib in these 
databases, but updates will follow continuously. 
From an evolutionary point of view, several obvious 
features with respect to chromosome numbers can be 
this overview. (1) Diploidy is very 
190 out of 


records. (2) Polyploidy is primarily expressed on the 


deduced from 


common, comprising 2890 chromosome 
triploid ploidy level (2n. = 3x), although higher ploidy 
levels are also found in a limited number of reports 
(N = 12). 


quently and are associated with both even and odd- 


3) Aneuploid chromosomes occur fre- 


ploids. (4) Prov ided that a sufficient number of counts 


were carried out, many of the individual taxa ex- 
hibited a variety of eytotypes, including diploids, 


polyploids, and aneuploids. However, it remains to be 
verified whether the latter feature represents a signif- 
icant cytological pattern of differentiation or whether 
if mere ly reflects an erroneous assignment of 
accessions and artificial taxonomic 
populations. 

The unevenness in sampling must be considered, as 
6196 of the data (195 out of 289 records) describe the 
karyological differentiation in accessions attributable 
(Boechera | divaricarpa, B. 


to three only 


holboellii, and B. stricta), while 6996 of the species 


species 


so far investigated (25 out of 36) are represented by 
one to three chromosome counts only. 

In addition, some of the karyological data require 
revision or reinterpretation, taking into consideration 
any ambiguities that may have resulted from varying 
degrees of synapsis and nonreduction in chromosome 
(Bócher, 1951; 


more detail 


numbers during meiosis Johnson, 


1970). As will be 


next section. (“Development of 


discussed in the 


— 


gametophytes”), the 
karyotypic irregularities common to Boechera embryo- 
genesis and pollen development are associated with 
agamospermy and hybridization and may have 
resulted in several unreduced gametophytic chromo- 
some numbers being published as haploid ones 
(Rollins, 1941), including B. divaricarpa, a count of 
n = 15: B. fendleri (S. Watson) W. A. Weber, 14 and 
21: and B. holboellii, 14 and 21 (cf. comments in 
Bócher, 1969; Mulligan, 1990). 


levels probably do not exist in Boechera. A 


Such high ploidy 


true 


crucifer 


classification. of 


the somatic tissues of P. sparsiflora (Nutt.) Dorn, 2n = 
46 (Bócher, 1969), the apparent 


outcome of an endomitotic event. 


ca. 23 and ca. 2n = 


Aneuploidy has been reported (in terms of species 
numbers) eight times. However, it can be inferred 
from the observed pattern of eytological differentiation 
within Boechera holboellii that autosomal aneuploidy 
may be the exception rather than the rule. Thus only 9 
139 records 


were aneuploidies not 


somatic chromosome numbers out of the 


ascribed to this species 


altributable B chromosomes. The infrequeney of 


aneuploids may reflect the need for proper chromo- 
some balance for pseudogamous endosperm formation 
in apomicts. In contrast, the flow cytometric data of 


Sharbel and Mitehell-Olds (2001) demonstrate. high 


requencies 


or aneuploid chromosomes of two size 
5 in Sharbel & Mitehell-Olds, 2001). 
of which likely represents the autosomal 


aneuploid chromosome. 


classes (cf. fig. 
the larger 
This larger aneuploid chro- 
restricted to a single (or very few) 
(Sharbel & | Mitchell-Olds, 


2001) and therefore may reach very high frequencies 


mosome Was 


chloroplast haplotype 
in specific geographic regions or genetic backgrounds. 
The smaller of the aneuploid elements is a B 
chromosome, which has a wide geographic distribu- 
2005) and can be 


divergent genetic backgrounds (Sharbel et al., 


found in 
2004). 
ut has what appears to be a single origin based on B 
2004). 


apomictic 


tion (Sharbel et al., 


allele sequencing (Sharbel et a This non- 


random distribution of Bs in individuals 


only, in conjunction with its wide distribution (1.e.. 


evolutionary maintenance), implies that it either has 
Camacho et al., 2000) 
with 


a meiol 
r ior 


reproduction. 


: c drive mechanism (cf. 


is somehow associated apomictic 


( 


— 


While the Boechera holboellii genome frequently 
contains additional chromosomes (both B and aneu- 
ploid chromosomes), natural populations are typically 
composed of both diploid and triploid eytotypes (Roy. 
1995: Sharbel et al., 2005). 


palynological, and molecular research has provided 


Recent cytological, 
us with the opportunity to describe and understand 
this pattern in more detail. Using pollen morphology 
Dobeš et al. 
2004a) described diploidy and triploidy among 7 
Sharbel and 
Sharbel et al. (2005) 


the occurrence B chromosomes in 


and microsatellite allelic variation, 


-—] 


poe 


accessions of this whereas 


Mitehell-Olds (2001) and 


demonstrated 


species, 


both of these cytotypes using flow-cytometry and 


karvology. These results are concordant with those 
collected. from light microscopical investigation of 


(2001) 


chromosome counts on 37 plants from 


chromosomes. For example, Naumova et a 


carried oul 


Volume 93, Number 3 Dobes et al. 527 
2006 North American Boechera (Brassicaceae) 
Colorado and Greenland representing eight separate — ell-Olds, 2001; Sharbel et al., 2005). Subsequent 


seed batches and one bulked seed collection. but 
found exclusively euploid diploids and euploid trip- 
loids. Because no tetraploid individuals have been 
identified in these mixed populations, it is hypothe- 
sized that environmental effects (e.g., cold shock) may 


trigger the production of unreduced gametes, which 


eventually contribute to triploid zygote formation 
(Sharbel & Mitchell-Olds, 2001; Sharbel et al., 


2005). 


triploid cytotypes within natural populations and 


The frequent and repeated generation. of 


different genetic backgrounds (Sharbel € Mitchell- 
Olds, 2001; 2005) provides 
evidence that gamete production is 
common in this genus (cf. Böcher, 1951). 
demonstrated to be of allopolyploid 
2004; Dobeš et al., 2004a). 
unclear genome 
differ 
dependently derived triploid lineages (Sharbel et al., 
2005). 
frequent sharing of haplotypes between diploids and 
(Sharbel & Mitchell-Olds, 2001). lt 


shown that certain diploid and triploid 


further 


Sharbel et a 
unreduced 
Triploids 


origin 


were 
(Shar 


this case it is 


and in 


— 


el et al., 
how and 


effects 


dosage 
subsequent phenotypic between in- 
Autopolyploidy was suggested based on the 
triploids wis 
further 
eytotypes share identical to genetically highly similar 
nuclear ribosomal ITS sequence types (Dobeš et al., 
2004a). 


variation, polyploid individuals always were found to 


However, based on microsatellite allelic 
show some extent of heterozygosity (Dobeš et al. 
2004a). 


for autopolyploidy in 


Therefore. we have so far no strict evidence 
Boechera, neither on the 


genotypic nor on the species level. It may be 
speculated, then, that polyploidization can only be 
triggered. in diploid hybrids via the 
discussion below). It is 


unreduced gametes (see 


furthermore unknown whether reported ploidy levels 


higher than 3x are the result of similar phenomen: 
formation from unreduced gametes in 
holboellii) or if 


history. 


(i.e., zygote 
mixed populations of B. these 
cytotypes are of a different evolutionary 
Although based on much smaller datasets, cytological 


characteristics similar to those in B. holboellii are also 


exhibited by several other Boechera species (cf 


Table la | and lb; <http://ephedra.hip.uni-heidelberg. 
>) 


A second type of cytological differentiation. with 
respect to chromosome numbers was encountered in 
several species so far observed to be represented by 
euploid diploids only. We mention here explicitly two 
examples. First, Boechera stricta, which is also a good 
example to discuss possible limits of morphology- 
based species recognition in Boechera. The majority 
of chromosome counts identified this species as a 
However, triploids were also found 
Mitch- 


diploid one. 


among the investigated accessions (Sharbel € 


formation of 


morphological investigations (Windham & Al-Sheh- 
baz, unpublished data) have shown that the triploids 
identified as B. stricta have actually undergone 
hybridization with another species. Support for these 
hypotheses was recently found using microsatellite 
allelic 2004a). 
conspicuous example is represented by a certain 


morphotype of B. holboellii, B. holboellii var. 


markers (cf. Dobeš et al., Another 


retro- 


fracta (= B. retrofracta (Graham) A. Love & D. Lóve). 


A significant proportion of diploid specimens assigned 


showed covariation of characteristic 


» this variety 
plastid and nuclear sequence polymorphisms (Dobes 
2004a). 


1 basal evolutionary unit within a highly 


et al., These genotypes were interpreted to 
constitute a 
1 and taxonomically artificial B. holboellii. 
that 


heterogenous species may actually be an amalgam- 


These examples demonstrate karyologically 
ation of different taxa, at least some of which can be 


characterized by single chromosome numbers and 
a restricted number of genotypes, respectively. 
Finally. based on our recent evaluation of data, we 
may define a third type of karyological differentiation 
represented by taxa that so far proved to comprise 
triploids only. Solely triploid taxa seem to be extremely 
rare according to current taxonomic classification of 
Boechera. Nevertheless, future taxonomic concepts can 
integrate knowledge on values of apomictic reproduc- 
tion of triploids or the age of apomictic clones, which 
may ultimately lead to the recognition of eytologically 


more uniform species. 
DEVELOPMENT OF GAMETOPHYTES 
Our cytological knowledge on the development of 


gamelogenesis in Boechera 


Tyge W. 


embryosacs and male 


from the work of Bócher 
1954, 1969) and contributions. by 
(2001) and (2004). 


Apomictic Boechera are diplosporous (the embryo 


comes mainly 
(Bócher, 1951, 
Naumova et al. 


Tsakin et al. 


develops from an unreduced embryosac mother cell) 
similar to the Taraxacum L.-type, and it appears that 
that 
endosperm has been identified in both diploid: and 
triploids (Böcher, 1947, 1951; Naumova et al., 2001; 

2004). Based mostly on pollen analyses, 


pseudogamy is maintained given fertilized 


Tsakin et al., 


sexuality has been shown to be associated with 
complete chromosome bivalent formation at meiosis, 
followed by regular heterotypic and homotypic divi- 
sions, to produce gametes with reduced chromosome 
numbers while pollen was tetradie and usually of high 


(Bócher. 1951). In 


gametogenesis involved either complete asyndesis or 


viability contrast, apomictic 


irregular pairing of chromosomes during meiosis J. 


followed by the formation of a restitution nucleus to 


528 


Annals of the 
Missouri Botanical Garden 


with | unreduced chromosome 


1951). the 


dyadic or monadic type. Apomeiotic development was 


result in gametes 


numbers (Böcher, Pollen was either of 


often accompanied by pathways exhibiting approxi- 


mately normal meiosis with varying 


univalents, resulting in tetrad pollen with varying 
chromosome numbers and relatively low fertility 


(Bócher, 1951; Sharbel et al., 2005). 


latter developmental traits can be considered to be 


Although these 


associated with apomixis, the occurrence. of meiotic 


abnormalities probably cannot be equated with 
apomietie reproduction itself. 

Complete analyses of embryological developmental 
stages have been performed from a limited number of 
accessions of Boechera holboellii or infraspecific 
members of this complex (Rollins, 1993) originating 
(“Disko”: 


Copenhagen”: “Eqaluit”; the Sondre Stromfjord) and 


— 


from 


reenland “Botanical Garden of 
two localities in the western United States (Colorado: 
Washington state) (Bócher, 1951, 1969: Naumova et 
al., 2001). In addition to these studies, Tsakin et al. 
(2004) analyzed the embryology of one accession from 
Colorado identified as B. gunnisoniana (Rollins) W. 
A. Weber. Any 


obtained. from. karyological investigations of 


other indications of apomixis were 
` isolated 
stages of female gametogenesis, population genetic 


1983; 


reproductive biology 


studies (Rollins, Roy, 1995), observations on 


Johnson. 1970: Vorobik. 1985: 


Roy. 1995), and analysis of pollen formation (e.g.. 
Bócher, 1951; Dobeš et al, 2004a) Therefore. 


embryological studies have so far been performed on 
two species, B. holboellii s.l. and aff. B. g 


from a few geographic areas only. 


unnisontana, 


From these considerations we can estimate the 
importance and frequency of the particular reproduc- 
We 


karyology, 


modes in Boechera. have summarized 


data on pollen and seed 


formation, fertility, and within-population differentia- 


tion in Table la and Ib and in the supplementary 
database (available on request and at <http:// 
ephedra. hip.uni-heidelberg.de/brassicaceae>). Im- 


traits that enable the differentiation of 


including pollen formation, 


portant 
apomixis and sexuality, 


degree of chromosome pairing at meiosis, and 


chromosomal reduction, also were considered. al- 


though nuclei showing irregular chromosome pairing 
at meiosis indicative. of 


may not necessarily be 


apomelosis. Given this uncertainty, these irregular 
pairing. constellations were not considered. sufficient 
to judge the reproductive mode of a given individual. 
used the that stable 


Furthermore, we argument 


triploidy should be associated with 
Sharbel & Mitchell-Olds, 2001). 
sexual Boechera could explain the 


As Boechera 


apomixis (cf. 
because no known 
pathway in 


maintenance of uneven ploidy levels. 


numbers of 


are annuals, biennials, and short-lived perennials 
without any effective modes of vegetative propagation, 
agamospermy seems to be the only possible mecha- 
nism through which triploidy can be inherited. 
However, triploidy does not necessarily mean obligate 
apomixis, and these cytotypes can potentially give rise 
to offspring of different ploidy levels. 
Although apomixis is a female trait (ier, asexual 


it can be inferred 


reproduction through egg cells 
from observations of pollen hens because apo- 
meiosis is reflected. in both the male and female 
components of Boechera (Bócher, 
(1951: 33) 
between the modes of male and female gametogenesis 
“HBH 3x," “S.Str. 
"oon. 10”). This 


in specific inflores- 


1951). For example, 


Böcher demonstrated 


— 


correspondence 


in several accessions (“Disko 2. 
. Str. 9.“ and 


correlation was even observed 


3. sterile shoot in 
cences in facultative apomictic Boechera (i. e., those 
producing seed both through sex and apomixis). For 
instance, a high degree of chromosome pairing during 
female meiosis was identified in flowers from the same 
inflorescence in which highly synaptic meiosis was 
identified in pollen mother cells (“HBH 3x” and 
S. Str. 107 1951). 


high degrees of syndesis were correlated. with 


material) (Bócher, Furthermore, 
seed 
sterility and decreased pollen viability on a particular 
branch of the inf hori in the triploid “S.Str. 97 


accession (Bócher, 195 

Taking the add traits into. conside- 
ration, we were able to find evidence of apomictic 
reproduction in accessions assigned to 17 species. 
Apomixis was associated with (1) diploidy, triploidy, 
or rarely tetraploidy: (2) aneuploidy and euploidy: (3) 
formation of dyadic and monadic pollen or tetradic 
pollen of usually low viability; and (4) low levels of 
genetic or morphologic differentiation within popula- 
tion, Apomixis is rarely the only mode of reproduction 
in any Boechera species, and hence nine of the 17 
species for which apomixis was identified are likely 
amphi-apomictic complexes (ef. Table la and 1b), 
which presumably have to be split into separate 
taxonomic entities. However, amphi-apomictic lineages 


have be expected given the report of decreased 


= 


£ 


values of apospory in Boechera (cf. Naumova el 


2001 and discussion below). 


Despite the accumulated amount. o chiefly. de- 
seriptive data regarding reproduction in Boechera. we 
still f the 
inheritance of apomixis in this genus. A first approach 
t 


made by Sharbel et al. (2004), who hypothesized that 


lave knowledge and 


poor ge netics 


» infer the genetic basis of apomixis was recently 


the B chromosomes may somehow be associated with 
the expression of the apomictic trait. Considering that 
the B chromosome in geographically and genetically 


distinct apomictic lineages is composed of a contigu- 


Volume 93, Number 3 
2006 


Dobes et al. 529 


North American Boechera (Brassicaceae) 


ous genomic region, Sharbel et al. (2004) proposed 
that it is evolutionarily important because it has not 
been lost by genetie drift, as would be expected for 


a neutral or deleterious B. chromosome (Camacho et 


— 


al., 2000). However, cases of apomeiotic formation o 
gametophytes in the absence of extra chromosomes 
have been reported (Rollins, 1941; Bócher, 1951, 
1954), and thus Sharbel et al. (2004) argued that such 


o be 


e 


an apomixis factor could also be expected 
present in autosomes. It should be noted that several 
experts in Boechera taxonomy have some doubt on 
some counts provided by Rollins (1941). 

The gathered data (e.g., Johnson, 1970: Rollins, 
1941; Mulligan. 1996; Rollins € Rüdenberg, 1977. 
1979) provided evidence for the existence of sexual 


species that were mostly diploid, exhibited complete 


chromosome pairing and meiotic chromosomal 


duction, and were characterized by high pollen 


— 


viability and seed fertility. These taxa (e. g. Boechera 
stricta) are characterized by euploid and diploid 
chromosome numbers and produce Polygonum L. 
type embryosacs via the meiotic pathway (Naumova 
et al., 2001; Tsakin et al., 2004). Recent taxonomic 
studies estimate the actual number of sexual diploid 
Boechera to be at least 70 (Al-Shehbaz & Windham, 


. Evidence for what is possibly purely 


— 


pers. comm. 
sexual reproduction in polyploids has been provided 
for three accessions only (B. holboellii, 2n = 28: 28 + 
If, Johnson, 1970; B. stricta, 2n = 28, Bocher, 1969 


In contrast, and as expected from theoretical con- 


— 


siderations, the reduction of chromosome numbers 
associated with regular bivalent formation was rarely 
observed in triploids. 

Interestingly, syndesis in Boechera is not only 
genotypically determined, i.e. reflecting homology of 
chromosomes, but it also can be modificative. Bócher 
(1951) explained the facultative occurrence. of syn- 
desis to be controlled by an unknown physiologic 
factor. This is best shown by the fact that apomeiosis 
and meiotic pairing can occur side by side and in 
definite parts within a single inflorescence (Bócher, 
1951; Naumova et al., 2001). Additionally, changes in 
temperature or environmentally induced differences 
in rate of growth of plants were shown to make B. 
holboellii change from apomeiotic to meiotic de- 
velopment, or at least change from fertility to sterility. 


Therefore. the available data indicate that we can 


— 


expect various reproductive pathways O CO-OCCUF 
frequently among plants of a population and even 


within single individuals. 
POLLEN FORMATION 


Male gametogenesis has been favored over the 


study of the development of female gametophytes by 


g. Rollins, 1941; Easterly, 1963; 
Raven et al., 1965; Johnson, 1970; Vorobik, 1985), 


because pollen is easily and abundantly available. 


several authors (e.g. 


Pollen viability can routinely be inferred from simple 
staining procedures (Vorobik, 1985; Naumova et al., 
2001), pollen grain shape and size (Bócher, 1951; 
Roy, 1995; Dobeš et al., 2004a; Sharbel et al., 2005), 
and germination. experiments (Bócher, 1951). Fur- 
thermore, correlations between the size of pollen 


erains and their DNA content have repeatedly been 


— 


shown in several other genera (Ehrendorfer, 1949; 
Stepánková, 1993; Trávnícek & Vinter, 1999; Vilhar 
2002; Lihová € Marhold, 2003), and thus 


relative pollen ploidy can also be estimated. 


et al., 


Apomictic Boechera is a highly interesting and 
promising system to study pollen cytology, as gametes 
with highly variable chromosome numbers can be 
produced by single individuals (Böcher, 1951, 1969). 
For example, tetradic, dyadic, and monadic pollen 
have been observed within specific parts of a single 
inflorescence, with monad pollen mainly observed in 
the lowermost portion of the inflorescence, while 
dyads prevailed in its middle part (Bócher, 1951, 
1969). Both types of pollen resulted from apomeiotic 
gamete development, which was often followed by the 
formation of a restitution or contraction nucleus. In 
any of these apomeiotic pathways, one (heterotypic 
division: formation of dyads) or both (heterotypic and 
homotypic division; monads) of the divisions corre- 
sponding to those in normal meiosis were abandoned 
(cf. 1951: 21 


gametes can be produced with either unreduced 


=, 


Bócher, seq... As a consequence, 


chromosome numbers or with double the somatic 


chromosome number of their bearers, as has been 


uce both triploid 
1951; 


chenko et al., 2001). Double pollen grains possessing 


shown in somatic triploids that proc 
and hexaploid pollen (e.g.. Bócher, Kravt- 
four generative and two vegetative nuclei have also 
been observed (Bócher, 1951). 

tetradic pollen 


In contrast, the formation of 


generally involved some degree of chromosome 
pairing followed by a reductional division. Gametes 
produced via this pathway by individuals exhibiting 
evidence for apomixis were reported to contain highly 
variable chromosome numbers, the result of cytolog- 
ical irregularities during their formation (cf. Bócher, 
1951, 1969). Gametes containing chromosome num- 
that 


a scenario of either regular meiotic division or 


bers deviated from those expected | under 
complete (apomeiotic) nonreduction have been ob- 
served by several authors (Rollins, 1941; Bocher, 
1951: Johnson, 1970; Vorobik, 1985). 

A conspicuous case was described by Böcher 
(1969 


(i.e., syndetic) triploids based on pollen size data and 


, who identified two different classes of synaptic 


— 


Annals of the 
Missouri Botanical Garden 


chromosome configurations during meiosis J. includ- 
ing those that produced only large pollen and those 
that produced a bimodal distribution of smaller pollen 
size. Those triploids producing a single 
relatively large pollen often showed regular meta- 


phase I with 10 bivalents plus one univalent, 


than the seven trivalents or corresponding lower 


numbers of bivalents and univalents (Bócher, 1951). 
Such pairing formations have subsequently been 
observed by several authors (Bócher, 1951, 1969; 


Johnson, 1970) and have recently been supported by 
pollen analyses (Sharbel et al., 2005). Regular 
bivalent formation in triploids was hypothesized. to 
be the outcome of structural chromosome mutations 
Bócher, 


how 


— 


1951), although it is difficult to understand 


such mutations (a random process) could 
convergently lead to a similar pattern of synapsis in 
different triploid lineages (Sharbel & Mitchell-Olds, 
in prep.). Although these cytological irregularities led 
to a considerable decrease in pollen viability (Bócher. 


1951: 


were still fertile and capable of participating in the 


Vorobik, 1985), a limited number of gametes 

production of viable offspring (Bócher, 1969; see later 

discussion). 
Nogler (196 


apomixis may act as, or are linked to, recessive lethal 


34) suggested that alleles responsible for 


factors, as would be expected in an asexual organism 
(Kondrashov, 1985 


lethal mutations in apomicts over time would therefore 


. The accumulation of recessive 


lead to decreasing fertility of haploid gametes. Indeed, 


Naumova et al. (2001) observed a mixture of reduced 


and unreduced pollen grains in apomictic diploid 
Boechera holboellii, while only unreduced female and 
These 
that 
Alterna- 


proper dosage of an apomixis factor may be 


male gametes contributed to mature seed. 


results suggested, as stated by these authors. 
there was selection against haploid gametes. 
tively, 
important, and haploid gametes may simply be unable 
to express this trait. However, tetradic pollen pro- 
duced by sexuals was usually of high quality (Johnson, 


1970; Vorobik, 1985; Roy, 


formed by normal and undisturbed meiosis. 


1995) and obviously was 


Data on the chromosomal composition of male 


gametes have recently been supplemented with 
morphological analyses of pollen. Although the size 


of pollen grains appeared highly variable among 


species (Vorobik, 1985; Roy. 1995) and within 
individuals (Bócher, 1951), Dobeš et al. (20044) 


demonstrated that pollen size can be used to identify 
diploid sexuals in Boechera holboellii and B. stricta. 
Furthermore, plants capable of apomictic reproduc- 
tion were characterized by a multimodal distribution 
1951). The 


of pollen size corresponded to tetradic, 


of pollen sizes (Bócher, sse various classes 
dyadic, and 


monadic pollen, respectively, as well as to two distinct 


class of 


rather 


sizes of pollen mother cells. Bócher (1951) concluded 
that formation of pollen with unreduced chromosome 
number is frequent in apomictic B. holboellii, taking 
place in some diploids and all triploids. This obser- 
valion is also in agreement with the results of Roy 
(1995). Naumova et al. (2001), and Kravtchenko et al. 
(2001). who also found formation of dyads as well as of 


monads to be common. 
ESTABLISHMENT OF OFFSPRING 


Two main reproductive pathways allowing the 
stable and constant inheritance of chromosome 
number in specific eytotypes have been identified in 
Boechera: regular meiosis in euploid diploids and tet- 
raploids and agamospermy in a variety of eytotypes. 
Sexual Boechera appear to be highly self-compatible 
and hence inbreeding is very common, as has been 
inferred through isozyme patterns (Roy. 1995) and 
2004a). The floral 


f Johnson (1970) on nine 


microsatellites (Dobeš et al., 
biological investigations of 
biotypes assigned to B. holboellii and B. lyallii (S. 
Watson) Dorn, B. sparsiflora, and B. stricta demon- 
strated that each was self-pollinaling and was adapted 
1971) 


Johnson (1970) also identified different biotypes co- 


for some outcrossing by insects (cf. Rollins, 


occurring at single sites and discovered evidence for 
mechanical and ethological (i.e. preference of 
different insects to certain biotypes) isolation mecha- 
nisms between them. Although the above observations 
provide clear evidence for strong isolation. between 
lineages, repeated introgression between sexual taxa 
and also between sexuals and apomictics is likely 
frequent at both the intra- and interspecific levels 
Koch et al., 2003: Sharbel et al., 2004: Dobeš et al., 
2004a. 2006). between Boechera 
species have been repeatedly obtained in crossing 
1970: Vorobik. 1985: Roy, 
1995) and observed in the field (e.g.. Rollins, 1983, 
1993: Mulligan. 1996). At present, the importance of 


hybrid speciation in the radiation and diversification 


Thus, hybrids 


experiments (Johnson, 


of the genus is widely accepted (Windham et al., 
2004). While 


viously thought and demonstrated to be associated 


hybridization in Boechera was pre- 
with polyploidization and the apomictic stabilization 
(Böcher, 1951; Sharbel & Mitchell- 
2003; Dobeš et al., 2004a), 
diploid and sexual B. 


= 


of new genotypes 
Olds, 2001: 


the report of 


Koch et al.. 
divaricarpa, 

hybrid between B. holboellii (and probably other 
species) and B. stricta, may provide evidence thal 


homoploid sexual hybrids have played a similarly 


— 


important role as polyploidization in the establishment 
of new species. 
As with regular meiosis, apomixis potentially allows 


for the stable inheritance of chromosome number i 


Volume 93, Number 3 
2006 


Dobeš et a 531 


l. 
North American Boechera (Brassicaceae) 


the course of reproduction. Furthermore, agamo- 


spermy resembles autogamy, a special case of sexual 
reproduction often encountered in Boechera that 
leads to offspring having the same genotype as their 
1995). a fundamental differ- 


parent (Roy, However, 


ence between agamospermy and apomixis is that 
apomixis maintains heterozygosity and stabilizes 
polyploid, aneuploid, and hybrid genomes. The 


population genetic consequences of apomictic re- 
production include the coexistence of different and 


potentially crossable but clearly distinct morphotypes 
within populations (Rollins, 1983), in addition to low 
(Roy, 1995; 


Whereas agamospermy in 


intrapopulation level differentiation 
Naumova et al., 2001). 
Boechera is almost never obligate, facultative apo- 
mixis has frequently been demonstrated (Bócher, 
1951, 1969; Naumova et al., 2001), 
values for apomeiosis during female gametophyte 
73% to 98% in 

2001). Naumova et al. 


(2001) also observed apomictic embryos with somatic 


[s 


with average 


development ranging from 


holboellii (Naumova et al., 


ploidy levels double that of their mothers, the result of 
fusion between unreduced male and female gametes 
and therefore deviating from normal parthenogenetic 
development. The above-mentioned mechanisms, in 
addition to variable chromosomal synapsis (Bócher, 


1951. 


and 


1969), imply that apomixis plays a significant 


specific role in the establishment of new 
genotypes. Additional genetic variation can be ex- 
pected from crossing-over events between chromo- 


somes (i.e., automixis) at various pairings, which were 


reported to lake place even in apomictic plants 
(Bócher, 1951, 1969). 

The role of apomeiotic development of gametes in 
which most 


the origin of triploids, represent. the 


— 


common ploidy level in Boechera, is particularly wel 


— 


understood. Bócher (1951) previously pointed out that 
amphi-apomictic diploids could easily produce trip- 
loid offspring by fusion of reduced and unreduced 
gametes. The frequent and repeated establishment of 
triploid genotypes within and among various evolu- 
tionary lineages has been confirmed by Sharbel and 
Mitchell-Olds (2001) and Sharbel et al. (2005) 
through the use of chloroplast sequence variation. 
Dobeš et al. (2004a) and Sharbel et al. (2004) have 
furthermore established that these polyploid cytotypes 


are the outcome o 


re 


crossing events between similar to 
largely diverged genotypes. By analogy. agamosper- 
mous diploids might pass the apomictic trait without 


a change in ploidy levels to genetically recombined 


e 


offspring when occasional fusion of two reduced 


gametes takes place, and this may occur between 


two apomicts or even between an apomict and a sexual 


[om 


individual. However, the establishment of triploids 


may have also involved diploid pollen from tetra- 


ploids. Although somatic tetraploids were found to be 
extremely rare in Boechera, they may be evolutionari- 
ly significant. Thus Verduijn et al. (2004) demonstrat- 
ed that tetraploid dandelions (genus Taraxacum L., 
Asteraceae) give rise to mainly triploid offspring in 
crosses with diploid sexuals. A similar scenario may 
apply to Boechera, as was recently hypothesized by 
(2004a). Unlike 


tetraploids are the only cytotype putting out diploid 


Dobes et al. Taraxacum, in which 


— 


pollen with high frequency, unreduced gamete pro- 


duction appears to be relatively common among 
diploid Boechera. 

While the origin of triploids can be explained by 
the above processes, the establishment of diploid 
apomicts is much less obvious. However, an intriguing 
possibility is that triploids that produce bimodal 
pollen sizes (i.e., pollen with 14 + B or a complement 
of 7 chromosomes) may be the source of apomictic 
diploid 2C + B Boechera holboellii (Bócher, 1969; 
2005). This explanation would allow 
shifts between diploids and triploids and facilitate 


i 


Sharbel et al., 


regular genetic exchange between these cytotypes 
2005). 


Evidence for exemplary hybrid origin in Boechera 


pon 
^ 


Verduijn et al., 


is provided and discussed in Dobes et al. (2006). The 
origin of diploid and triploid eytotypes. both of which 
are the most frequently encountered in Boechera, is 
with the chromosomal 


concordant composition of 


gameles. However, the scarcity of tetraploids and 
higher ploidy levels in natural populations suggests 
that natural selection acts against their formation, with 
genomic balance or stabilization through apomixis. 
We can say so because it appears that selection does 
not have a significant impact on gametes, considering 
that dyad and monad pollen. were shown to have 
morphological integrity (Dobes et al., 2004a) and high 
1951; Roy, 1995). 


unreduced pollen was shown to successfully fertilize 


viability (Bócher, Furthermore, 
the central cell of embryosacs, which is required for 
pseudogamous endosperm development (Naumova et 
al., 2001; Tsakin et al., 2004) 


CONCLUSIONS 


We can expect that the processes described above 
would lead to a large variety of new genotypes and 
cytotypes with distinct chromosome numbers. Fur- 
thermore, these new types may potentially become 
fixed through asexual reproduction, as meiotic ir- 
chiefly 


carrying the apomictic trait. However, the reproduc- 


regularities were observed in individuals 


tive success of such genetically recombinant and 


genomically heterogeneous progenies has never been 
demonstrated in the long term, and their evolutionary 
It will be 


significance is not fully understood. yet. 


532 


Annals 
118 um Garden 


promising field of future experimental research to 


correlate structural changes of the genome in hybrid 
offspring with fitness measures over several genera- 
tions, which we presumably can do as soon as linkage 
2000). 


hybrid. genotypes 


maps become available (cf. Rieseberg et al., 
Estimation of the performance 
and cytotypes will enable us to assess whether hy- 
brids are merely byproducts of evolution or whether 
they themselves actively drive diversification in this 
genus. 

The North American Boechera group is a highly 
within Brassicaceae. 


complex species aggregate 


However, the group itself is clearly defined and 
restricted North America, which is also true for 


For 


realistic. that the ongoing. projects focusing on ils 


most of its close relatives. this reason, it is 


taxonomy, systematics, evolutionary history, and 
phylogeography will provide a good framework to 


understand speciation and evolution in this notori- 
ously difficult group. Additional research on molec- 
ular mechanisms of apomixis, genome structure, and 


chromosomal synteny and collinearity will provide 


data that bridge evidences from cytology and 

embryology to species diversity. 

Literature Cited 

Acarkan, A.. M. Roßberg. M. Koch & R. Schmidt. 2000. 
G omparalive genome analysis re reve als exte nsive conserva- 


tion of genome organisation for Arabidopsis 1 ind 
5 Lg Plant J. 23: 55-62. 

Al-Shehba 2003. Transfer of most North American 
5 cles of. 575 to eee (Brassicaceae). Novon 13: 
381—391. 
Böcher, T. W. 


borealen 


1936 


Zur Zylologie einiger arktisehen und 
Blütenpflanzen. 32: 


Svensk Bot. Tidskr. 


— 1047 


Hornem. 


Cytological studies of Arabis Holboellü 
3: 


Here . 33: 


= 1951. Cytological “a embryological studies in the 
apap tic rabis kines complex. | Kongel. 

Danske Vidensk. Selsk. Biol. Skr. 6: 1-59, 
1954. Experimental taxonomical studies in 


the irabis holboellii complex. Svensk. Bot. Tidskr. 48: 
3144. 

. 1969. Further studies in Arabis a and allied 
species. Saertryk Bot. Tidsskr. 64: 14 


1950. C 


— ——— & K. Larsen. Iromosome 1 1 “rs of some 


arctic or boreal flowering plants. Meddel. Grønland 147: 

32. 
Camacho, J. P. M., T. F. Sharbel & L. W. Beukeboom. 2000. B- 
Philos. Trans., Ser. B 355: 163-178. 


chromosome Bud 
V. 19% 


. Chromosome numbers in some vascular 


plants from de Disko Bugt area (West Greenland). 
Willdenowia 18: 243-252. 
Dobeš, Ch.. Th. Mitchell-Olds & M. Koch. 2004a. In- 


diversification in North American Boechera 


stricta (= Arabis drummondii), Boechera Xdivaricarpa. 


and Boechera holboellii (Brassicaceae) inferred from 


nuclear and chloroplast molecular markers—an integra- 
Amer. J. Bot. 98: 2087-210 


live approach. 


& —— . 2004b 


sis 2 extensively sympatric and highly dive 


Phylogeographic anal- 
rse chloro- 
GS) in North American 
and A. hol 


pl ast haplotypes sige intron-irnF 


Arabis drummondii, e d, fei, boellii 
34 


370 


11 understand- 


dio de ). Molec. (^ col. 


—— ———, T. Sharbel & M. 


ing 15 dynamics 


E. 

of hybridization and apomixis in the 
evolution of genus Boechera (Brassicaceae). Syst. Biodi- 
vers. In press. 

Easterly, N. W. 1963. 
northwestern Ohio Cruciferae. 


b 


MES of 
Castane: 39-42. 
1949. Zur Phylogenie der b. Galium. |. 
yploidie und geographiseh-ókologische Eir 


C hromosome some 


E re dorfer. F. 


eiten in 
der Gruppe des Galium pumilum Murray (Se is Leptoga- 
lium Lange sensu Rouy) im österreichischen Alpenraum. 
Osterr. Bot. Z. 96: 109—138. 

Heenan, P. B.. E: Mitchell & M. Koch. 2002. Molecular 
systematics of f the New Zealand Pachycladon (Brassic: 

s generic circumscription and re bii 
o Arabidopsis s.l. and Arabis s.l. New Zealand J. Bot. 
543—502. 


ceae) comple 


Johnson, T. F. 1970. Investigations in the ~ Biology of 
the Arabis dus puse 73. Ph. 

Seattle 

Molecular. phylogenetics, 


Dissertation, 
Iniv. Washington, 
Koch. M. 2003. 
population biology in Brassicaceae. Pp. 5 In 
Sharma & A. Sharma edion. 
ution, Vol. I. Par 
Publishers, Enfield. New Hampshir 


evolution and 
1-35 


Plant Genome: A 
A 


and Evo Phanerogams. Science 


—— —— & M. Kiefer. 2005. Genome evolution among 


cruciferous plants: A lecture from the integration of the 
gene lic n laps of thre 38 diploid spec ie c apse tla rubella, 
er lyrata subsp. petraea, and A. thaliana. Amer. 
J. Bot. 92: 761-707 

. Bishop & Th. Mitchell-Olds. 1999. Molecular 
aes and evolution of Arabidopsis and Arabis. Pl. 
Biol. 1: )-531. 

He bold & Th. 


systematics of the 


Mitchell- z 


Brassicaceae: 


2001. Molecular 
ence from coding 
plastidie matK and nuclear Chs sequences. Amer. J. Bot. 
88: 534—544. 

Ch. Dobeš & Th. 


formation in 


Mitchell-Olds. 


natural 


. 2003. Multiple 
hybrid populations: Concerted 
evolution of the internal transcribed spacer of nuclear 
ribosomal DNA (ITS) in es ge irabis divari- 
carpa (Brassicaceae). Molec. Biol. Evol. 20: 338-350. 
, M. Matsc ira T. = 18 ber. J. Vogel. M. 
Kiefer & Th. Mitehell-Olds. 2005. Evolution of the 
IrnF(GAA) gene in Arabidopsis relatives and the Brassi- 
origin and 


caceae family: Monophyletic subsequent 


of a plastidic Molec. Biol. 
. 22(4): 1032-1043 


Kondrashov, A. S. 1985 


5 qe 


Deleterious mutations as an 


actor. I Facultative apomixis and selfing. 
035-053. 

1076. 
United sue dicotyle 1 65 
Kowalski. S. P., T. H. La A. 

1994. b mapping of Arabidopsis thaliana and 
Brassica oleracea chromosomes reveals islands n 
served organization. Ge netics 138: 499-5 10. 
Kravichenko, . N. Naumova & J. van der Laak. 2001 
Formation of uso di ceed pollen in the natural apomi 
Arabis holboellii. Presented at: 
I APO2001: 


dec 


Genetics | 

numbers of miscellaneous 

Rhodora 80: 431—140 
Feldmann & . Paterson, 


Kovanda, Chromosome 


=> 


2nd Inte 1 
April 24-28, 2001: 


species 
Apomixis Conference 
Como, Italy. 


Dobeš et a 


Volume 93, Number 3 A 
2006 North American Boechera (Brassicaceae) 


533 


Kuittine an, II., A. A. de Haa „ Vogl, S. Oikarinen, J. 
eppälä, M. Koch, Th. Mite an Olds, Ch. Langley & O. 
ents 2004. Comparing the linkage maps of close üdenberg. 1971. Chromosome numbers of 
relatives 5 lyrata and Arabidopsis thaliana. Cruciferae. II. Contr. a Herb. 201: 117-13: 
Genetics 168: 1575-1584. & 
Lagercrantz, U. 1998. 


993. The Cruciferae of Continental North America. 
Sund Univ. Press, Stanford. 
& L. Rü 


77. Chromosome 1 d Cruci- 
Comparative mapping between ferae III. Contr. Cay Herb. 207: 101-116 
Arabidopsis thaliana and Brassica nigra indicates that 
rassica genomes have evolved through extensive genome 
replication aecompanied by chromosome fusions and 
frequent „„ Genetics 150: 
Lihová, d 200: 


979. Cosa les ue Cruci- 
1979: 79 


ferae IV. Bussey Inst. Harvard L 9: 79— 
Roy, " A. 1995. The breeding sy 1 aof six ies idt. Arabis 
59 


217-1228. (Brassicaceae). Amer. J. Bot 


Tarhold. Chromosome numbe 
vb of Cardamine | amara (Bras ssi 
Slovenia. E (Bratislava) 58: 37 

) 


Sharbel, T. F. & Th. i Olds. 2001. 
Love, A. 1969. IOPB oe bes reports. XXII. polyploid origins 
Taxon 18: au 
1975 


and c 


— 


& I RPM D Evidence 107 apomixis in 


Arabis. J. Hoiedity 80: 


caceae) in 


V! 
-e 


Te 


Recurrent 
and chloroplast: phyloge ography in the 


Arabis holboellii complex (Brassicaceae). Heredity 87: 
. IOPB chromosome number reports L. Taxon 59-68 
24: 67 1-67 8. ———— M. E o Th. Mitchell-Olds, L. Kantama € H. D 
—— ———. 1976. IOPB chromosome number reports LII. Jong. 2004. Is the 
Taxon 25: 483—500. 
1979. 


aneuploid chromosome in an apomictic 
Boechera oum a genuine B. chromosome? Cytogen 
(9. [OPB chromosome number reports LXHI. 


a s. 106: 173-183. 
"usan 3 a 


. TOPR NT num 


Mitche js Olds, Ch. Dobeš, L. Kantama & H. D. 
Taxon » 1 
& L 


Jong. ane Biogeographic distribution of polyploidy and B 


yer reports LXXIV. 


chromosomes in the apomictic Boechera holboellii com- 
). Lo 1975. Nomenclatural notes on arctic plex. Cytogen. Genome Res. 109: 283-292, 

1 Bot. Not. 128: 497—523. Smith, F. H. 1938. Some chromosome numbers in the 
sak, M. A., A . Recent Cruciferae. Amer. J. Bot. 25: 220-221. 
of Arabidopsis and Stace, C. A. 19 

11: 195-204. 


Ly Pecinka € L Schubert. 2003 


chromosome | painting 
relate d: species. Chromosome Res. 
M. Koch, A. Pecinka € I. Schubert. 2005. 


Chrómosome triplication found across the tribe Brassi- 


progress in 


989. IOPB omaso data 1. 


Newslett. Int. 
Organ. ue E st. 13: 


p IOPB 5 data 13. Newslett. Int. 
M P 0 st. 29: 17-24 


J 
ceae. Genome Res. 15: 516-525 )PB chromosome data IA. Newslett. Int. 
Mulligan, G. A. 1964 = Chrono d of the family dd 1 30: 10-15 
Cruciferae. I. Ca mad: J. Bot : 1509-1519 


un Ste 1 J. 1993 
1996. Synopsis of dien enus Ara a tei aceae) 
109— 


. Re mas on the karyogeography of 
32 
in 9 Alaska and e ene Rhodora 97: 
163. 


Myosotis 1 in Europe. Preslia 65: 325-334. 
Taylor, R. L. € R. P. Broc SERM 1966. Chromosome numbers 


of some western Canadian plants. Canad. J. Bot. 44: 
Naumova, T. N., J. van der Laak, J. Osadtchiy, F. Matzk, . 1093-1103. 
Kravtche 125 J. Bergervoet, K. S. Ramulu & K. Boutilier. . 1977. € hromosome numbers of vascular 
2001. Reproductive development in d 11 plants of British 9 Syes 125-138. 
r ao holboellii (Brassicaceae). Sexual Pl. Reprod. 14: 


nícek, B. & V. Vinter. 1999, uico der Beziehung der 


Ploidie zur Pollengrósse und Stomatalánge bei der Art 
11785 5 a P 984. Gametophytic apomixis. Pp. 475—518 in Pseudolysimachion maritimum (= Veronica longifolia 
5 M. Johri (edito Embryology of Angiosperms. Springer auct., Scrophulariaceae) ‘ta Ped herum. Nat. Univ. 
erlin a Olomucensis, Biol. 35—45. 
D. ane 185 L. Jr. & J. A. Al-Shehbaz. 1997. A synopsis of Tsakin, K. . Turgut & R. a Scott. 2004. Apomictic 
pe 1 Novon 1: 323-327. d sel in Arabis gunnisoniana. Israel J. Bot. 52: 
Packer, J. G. 1964. Chromosome numbers and taxonomic 155-160. 
notes on western Canadian and arctic plants. Canad. J. a M. H., 
Bot. 42: 5 


P. J. Van Dijk & J. M. M. Van Damme. 


05. The role « dois in the se 
Raven, P. H., D. W. Kyhos & A. J. Hill. 55. Chr 


pk m cycle 


se 
1965. Chromosome in a (Taraxacum). Heredity 93: 
numbers of musik mostly Californian. Aliso 6 Vilhar, B., T. Vidic, N. Jogan & M. Demata 2002. 
105-113. Genome size and the nucleolar number as estimators of 
Rieseberg, L. H., S. J. E. Baird & K. A. Gardner. 2000. ploidy level in Doo posi in the Slovenian Alps. 
5 8 and linkage evolution. Pl. Pl. Syst. Evol. -13. 
Molec. Biol. 42: 205-224. 


Vorobik, L. 
Rollins, R. C. E A monographic study of Arabis in 


Western North America. Rhodora 43: pou 325. 
425-481. 


A. pre 5. 


Hybridization and Reproductive 
Isolation Between 


Sympatric Arabis (Cruciferae) in 
348411, Southwestern Oregon. 245. Dissertation, Univ. 
Oregon, a ne. 
. 1960. Arabis perstellata in Tennessee. Rhodora 62: Ward, D. E. 1983. Chr 
242—244. 


ts s from New Mexico and 
2 309 


— 
Ww 
= 
Q 


southern Cola js 54: —309. 
—— ——. 196€ o numbers of Cruciferae. Contr. Warwick, & L Al-Shehbaz. 2006. Brassicaceae: 
Gray T I7: 43-65. uN 1 index and database on CD- 
19 y otogyny in the Cruciferae and notes on Rom. . 237-248 in M. A. Koch € K 
poe and 1 AS Contr. Gray Herb. 201: 3-10. (editors), Evolution and Phylogeny of 
. 1983. Interspecific hybridization axon unifor- Plant Systematics and 
mily is Arabis (Cruciferae). Amer. J. Bot. 70: 625-6: 


ler 


— 
- 


Evolution: 259 


2—4). NE 
Vienna 


Annals of the 
Missouri Botanical Garden 


Species chec list and database on CD-Rom. Pp. 2 
Coch & 


. Francis 


A. & LA, 


Al-Shehbaz. 2006. Brassicaceae: 
249—258 


K. Mummenhoff (editors), Evolution and 


Phylogeny of the Brassicaceae. Plant Systematics and 
Evolution: 259(2-4). Springer, Vienn 

Wieboldt, T. F. 1987. The shale barren n e dan mie, Arabis 
serotina (Brassicaceae). Sida 12: 381— 


Volume € 


. Number 3. 


pp. 371-334 of the 


ANNALS OF THE 
was published on 24 October 2006. 


W he M. D, I. A. Al-Shehbaz, C. D. Bailey & 

Loreen. 2004. A taxonomist's worst nightmare: 

" preliminary glimpse into ul systematic of 

Boechera (Brassicaceae). Abstract Presented at: 

Botany 2004. Alpine Diversity: \dapte d to the 

Peaks; July 31-August 5, 2004; Sal Lake City, 
Utah 


Missouri BOTANICAL GARDEN 


Missouri Botanic. 


TIT 


al Gar 


den Librar 


J 1753 00338 5504 


www.mbgpress.org 


CONTENTS 


A Taxonomic Revision of Caribbean Adiantopsis (Pteridaceae) 


Chloroplast and Nuclear Genes — 
Ji-Pei Yue, Hang Sun, Ihsan A. Al-Shehbaz & Jian-Hua Li 402 


Revisión del Género Acaena (Rosaceae) en Chile Alicia Marticorena 412 

Scale-dependent Classification of Xeric Limestone Prairies: Annual or Perennial Grass- 
lands? Patrick J. Lawless, Jerry M. Baskin & Carol C. Baskin 455 

Pollination Biology in a Tropical High-altitude Grassland in Brazil: Interactions at the Com- 
munity Level Leandro Freitas & Marlies Sazima 465 


Embryology, Karyology, and Modes of Reproduction in the North American Genus Boechera 
(Brassicaceae): A Compilation of Seven Decades of Research 


Christoph Dobeš, Marcus Koch & Timothy F. Sharbel 517 


Cover illustration. Capparis sicula subsp. mesopotamica Inocencio, D. Rivera, Obón & Alcaraz, 
drawn by Jose-Antonio Barreña. 


Annals 

OÍ the 
Missouri - 
Botanical 


Garden 
pa Y 


Volume 93 


Number 


Annals of the 
Missouri Botanical Garden 


Volume 93, Number 4 
December 2006 


The Annals, published quarterly, contains papers, primarily in systematic botany, 
contributed from the Missouri Botanical Garden, St. Louis. Papers originating outside 
the Garden will also be accepted. All manuscripts are peer-reviewed by qualified, 
independent reviewers. Authors should write the Managing Editor for information 
concerning arrangements for publishing in the Annals. Instructions to Authors are 
printed in the back of the last issue of each volume and are also available online at 


www.mbgpress.org. 


Editorial Committee 
Victoria C. Hollowell 
Scientific Editor, 

Missouri Botanical Garden 
Beth Parada 

Managing Editor, 
Missouri Botanical Garden 
Diana Gunter 

Associate Editor, 

Missouri Botanical Garden 
Barbara Mack 

Editorial Assistant, 
Missouri Botanical Garden 
Monica Anderson 

MBG Press Assistant, 


Missouri Botanical Garden 


Ihsan A. Al-Shehbaz 
Missouri Botanical Garden 
Gerrit Davidse 

Missouri Botanical Garden 
Roy E. Gereau 

Missouri Botanical Garden 
Peter Goldblatt 

Missouri Botanical Garden 
Gordon McPherson 
Missouri Botanical Garden 
P. Mick Richardson 
Missouri Botanical Garden 
Charlotte Taylor 

Missouri Botanical Garden 
Henk van der Werff 


Missouri Botanical Garden 


For subscription information contact ANNALS 
OF THE Missourt BOTANICAL GARDEN, Y Allen Mar- 
keting & Management, P.O. Box 1897, Lawrence, 
KS 66044-8897. Subscription price for 2007 is 
$165 per volume U.S., $175 Canada € Mexico, 
$200 all other countries. Four issues per volume. 
The journal Novon is included in the subscription 
price of the Annals. 


annals@mobot.org (editorial queries) 
http://www.mbgpress.org 


e ANNALS OF THE Missouri BOTANICAL GARDEN 
1 5 0026-6493) is published quarterly by the 
Missouri. Botanical Garden, 2345 Tower Grove 
Avenue, St. Louis, MO 63110. Periodicals post- 
age paid at St. Louis, MO and additional mail- 
ing offices. PosruasrER: Send address changes to 
ANNALS OF THE Missouri BOTANICAL GARDEN, % 
Allen Marketing € Management, P.O. Box 1897, 
Lawrence, KS 66044-8897. 


The Annals are abstracted and/or indexed in AGRICOLA (through 1994), APT Online, BIOSIS®, CAB Ab- 
strac/Global Health databases, ingenta, ISI databases, JSTOR, Research Alert®, and Sci Search®. 


Typesetting by Allen Press, Inc. using Advent3B2 automatic paging system. 


O Missouri Botanical Garden Press 2006 


The mission of the Missouri Botanical Garden is to discover and share knowledge about plants and 
their environment, in order to preserve and enrich life. 
This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 


Volume 93 
Number 4 
2006 


Annals 

of the 
Missouri 
Botanical 


Garden 


A SYSTEMATIC STUDY OF THE 
OLD WORLD GENUS FOCKEA 
(APOCYNACEAE- 
ASCLEPIADOIDEAE)'? 


P. V. Bruyns? and C. CUR 80 CN 


misso 
JAN 07 N- 
BRARY 


ABSTRACT 


— 


The six species of Fockea Endl. (Apocynaceae, 
southern Kenya to South Africa. A phylogeny for all six spec 


Ascley riadoide ae, 


GARDEN Ll 


Fockeeae) occur in Africa south of the equator from 
Dee ockea is inferred from data obtained from two chloroplast 
] 


markers (the trnL-F region and the psbA-trn 


nucieal mark« I (I TSI region), and morphology. Fockea i 18 


found to be 11 lic and sister to hs bitypie Chiu ind LEAL E Fockea is characterized by the deeply 


tubular outer corona that is filled by 


he erect, inflated sterile appendages of the anthers. Of the six species, the vid ly 


distributed but exclusively tropical a multiflora K. Schum. is sister the other five members of Fockea, and. among 


these fiv 


1 0 to southern Africa. It is shown that Fockea most probably originate ua in South Tropic e Africa. 


F 


Moore, F. schinzii N. 


o ae, 


AS 
— 
~ 
Q 
— 


JOCynaceae, 


M d Mick: taxonomy, 1 -F region. 


the widely distributed and mainly tropical F. angustifolia K 


e six species of Fockea, and their PM distribution is mapped. | 


Fockea, IIS] region, morphology, 


Schum. is sister to the remaining four, which are 
nplete taxonomic 
totypes are designated foi 

Schltr. and T. 


E angustifolia, and F. comaru (E. Mey.) N. E. Br. An 


r.. F. sessiliflora S n ae c Mey.) Druce. 
)N. E. B 


phylogeny, psbA-irnH intergenic spacer. 


The small genus Fockea Endl. is endemic to Africa 


— 


south of the equator. It was established in 1838 by 
Stephan L. Endlicher for the single species, F. capensis 
Endl. He based the description of this species on 
a specimen cultivated at Schónbrunn Garden in Vienna 
(here lectotypified by Schónbrunn Garden 488 (K, W)), 
which previously had been given the illegitimate name 
Cynanchum crispum Jacq. (non C. crispum Thunberg, 


1794). 


This famous specimen cultivated at Schónbrunn 


Garden under the number 488 was collected in the 


Cape Colony after 1786 by Franz Boos and Georg 
Scholl. It was still in cultivation in Vienna in 1930 
(Marloth, 1932) and, until about that time, was reputed 
to be the only surviving member of the species. This 
myth was exploded by Rudolf Marloth, who collected F. 
capensis near Prince Albert in South Africa. The plant 
collected by Boos and Scholl remains alive in Vienna to 
this day (Zecher, 1988). 


This work was supported in part by a grant from the University of Cape Town Research Fund. We thank Bill and Chris 


Ander University of Michigan, for nomenclatural assistance. 


"The cus of the Annals thank Sophia Balcomb for her editorial a to this paper 


? Bolus I 


uct.ac.za 


Herbarum, University of Cape Town, 


ANN. Missourt Bor. 


1101 Rondebosch, South A 


klak@botzoo. 


ca. an ıths.uct.ac.za 


GARD. 93: 535-564. PUBLISHED ON 15 DECEMBER 2006. 


Annals of the 
Missouri Botanical Garden 


The first recorded. gathering of. Fockea was made 
during the expedition of Simon van der Stel of 1685— 
1686 to 


South Africa. Material was found north of Clanwilliam 


the Copper Mountains of. Namaqualand 


between the Olifants River and the Doorn River and 
was figured in an unpublished manuscript by Jan 
This figure was later identified as F. edulis 
(Thunb.) K. Schum. (Wijnands et al., 1996). However, 
F. edulis does not occur anywhere near Clanwilliam, 
and this plant was F. comaru (E. Mey.) N. E. Br. An 
early 


Commelin. 


collection of F. edulis was made by C. P. 


Thunberg near the Gouritz River in southern South 


Africa in October 1772. He described this as 
Pergularia edulis Thunb. in 1794 and later placed il 


in the 


genus Echites P. Br. Early gatherings of the 


small and inconspicuous species F. comaru and F. 
sinuata (E. Mey.) Druce were made by Johann Franz 
Drege, probably between 1832 and 1833. Ernst H. F. 
Meyer, 


seem to have been quite sure what to do with these two 


who worked on Drege's collections, does not 


species and described them both as possible members 
1838). Only after 1900 
did they come to be associated i Fockea. 

Both tropical 
species, F. angustifolia K. Schum. and F. multiflora 
K. Schum. 
Schumann (Schumann, 1893). F. angustifolia has by 
the 


of Brachystelma Sims (Meyer, 


widely distributed and mainly 


were described rather later by Karl 


far most extensive synonymy of any species of 
Fockea, probably because of its wide distribution and 
the variability in the shape of the leaves and the size 
of the flowers. 
A further 
Chymocormus Harv., was established by William H. 
Harvey (1842). 965 (K), 
which he considered to represent Thunberg's Pergu- 
that I do 


genus involving species of Fockea, 


He based this on Zeyher 


— 


laria edulis, with a “corona so remarkable, 


not hesitate to propose it as the type of a new genus 


(1842: 24). 


However, although Harvey did mention 


“Pergularia 2 edulis, Thunb. (Zeyher, 965)" (1842: 
24) as belonging to Chymocormus, he made no valid 
new combinations (Greuter et al., 2000: [CEN Art. 


33.1) and published no new names in Chymocormus. 
Modern taxonomic investigations into Fockea began 
with an informal account by M. B. Bayer, where the 


various species were discussed and several previously 


unknown facts were brought light (Bayer, 1976). 
This account was based on material at the Karoo 


Worcester, South Africa. A was 
prepared by Court (1982) for a Master's degree al 
Rhodes South Africa. 


features were investigated in this account for 


Garden, revision 


University, Morphological 
their 
found that, whereas 


relevance and it was 


axonomic 


the flowers were useful for characterizing the genus, 


only features of the leaves could be used to 
distinguish the species. Five species (F. comaru, F. 


* 


crispa K. Schum., F. edulis, multiflora, and F. 
sinuata) were recognized, but this account was never 
published. A summary of some of the results of Court 
1982) appeared in 
(Court, 1987), where six species (F. angustifolia, F. 
F. multiflora, and F. 
This the 
account of 1982 in that F. ii E and F. comaru 


were treated as separate specie 


n 


the popular journal Asklepios 


comaru, F. crispa, F. edulis, 


sinuata) were discussed. differed from 


E. Brown (1902-1903: m mentioned that 
Fockea "appears to form a connecting link between 


the tribes Secamoneae and Marsdenieae.” but he 


placed it in the Marsdenieae because the pollinia 
lacked a differentiated margin and are horizontal in 
1902-1903: 1907-1909). 
Kunze (1993) pointed out that many peculiar features 


the anther thecas (Brown, 


of this genus suggested a more “basal” position in the 


Asclepiadoideae. Later a new tribe, Fockeeae, was 


published for the two genera Fockea and Cibirhiza 
1994). 


distinguished from the Marsdenieae and others by the 


Bruyns (Kunze et al., This tribe was said to be 


shared derived characters of the complex corona, the 


arge anther appendages, and certain characters of the 


translator, namely the absence of true caudicles, the 
attachment of the pollinia on the side of the corpuscle 
closest to the style head, and the lack of a floor in the 
corpuscle connecting its flanks. 

In fact, many of these characters do not distinguish 
the Fockeeae from other asclepiads. A complex 


corona is also found Eustegia R. Br. where the 
ontogeny has been found to follow a similar pattern to 
that of Fockea (Bruyns, 1999: 25, fig. 4 


Sarcostemma viminale (L.) R. B 


). The corona of 
r. is also of complex 
construction os ils ontogeny see Endress & Bruyns, 
2000: 26, fig. 


the anthers are unusually 


11). At 2-3 mm long, the appendages of 
all 
Fockea, but in both species of Cibirhiza they are 
1994: 370, 


ong) in many 


large in species of 


ess 


than 0.5 mm long (e.g.. Kunze et al., fig. 


4). Certainly they are larger (at 1 mm 
species of Microloma R. Br. (e.g.. Bruyns & Linder, 
1991: 505, fig. 17) than they are in either species of 
Cibirhiza. Furthermore, the appendages in Telosma 
africana (N. E. Br.) N. E. Br. 
in that they are longer than the fertile part of the 
As Kunze 


showed, the pollinarium of Fockea exhibits 


resemble those in Fockea 


anther and 
(1993) 


many features of species of Secamone R. Br.. 


lave somewhal inflated margins. 


and one 
of these is the lack of true 
both. « Sal process of the 
corpuscle. Verhoeven et al. (2003) found that the wall 
of the pollinia and the pollen grains differed markedly 
in Fockea and Cibirhiza: 
the 


caudicles, with the pollinia 


attached in cases [o a 


in Cibirhiza the wall is 


and 


resent grains are single (as in the 
| 8 8 
Asclepiadoideae), while in Fockea the wall is absent 


and the grains are in tetrads (as in the Secamonoi- 


Volume 93, Number 4 
2006 


Bruyns & Klak 
Systematic Study of Fockea 


Table 1. 


Vouchers for Fockea and other Apocynaceae used in molecular analyses with Genbank accession numbers for 


nuclear (ITS1) and chloroplast (psbA-trnH, trnL-F) gene regions. 


Accession Origin ITS] psbA-trnH trnL-F 
Cibirhiza albersiana Tanzania, near Gulwe, Bruyns 9652a (BOL) AM 233381  AM231756 | AM233363 
Cibirhiza dhofarensis Oman, Dhofar, Miller 7525 AM 233382 AM231757  AM233304 
Fockea angustifolia, acc.! Namibia, Van Zyl’s Pass, Pu ns 6054 (BOL) AM 233385 AM231760 | AM233367 
Fockea angustifolia, acc. 2 South Africa, Muden, Bruyns 9406a (BOL) AM235384  AM231759  AM233366 
Fockea angustifolia, acc. 3 Tanzania, near Kondoa, Bruyns 9632 (E) AM 233386 | AM231761 AM233368 
Fockea angustifolia, acc. 4 Zambia, Sinazongwe, Bruyns 9587 (MO) AM235383  AM231758 AM233365 
Fockea capensis South Africa, Plathuis, Bruyns 7340 (NBG) AM233387 M231762 | AM233369 
Fockea comaru, acc. 1 mibia, Numeis, Bruyns 8821 (MO) AM 233388 AM231763  AM233370 
Fockea comaru, acc. 2 South Africa, near Port Elizabeth, Dold 2381 (GRA) AM233389 AM231764 AM233371 
Fockea edulis South Africa, near Steytlerville, Bruyns 7051 (BOL) AM233390 AM231765 AM233372 
Fockea multiflora, acc. 1 Namibia, Okonguati, Bruyns 4089 (NBG) AM233391  AM231766 AM233373 
Fockea multiflora, acc. 2 Tanzania, near Gulwe, Bruyns 9651 (E) AM233392 AM2317607 | AM233374 
Fockea multiflora, acc. 3 Zambia, N of Monza, Bruyns 9593 (MO M 233393  AM231768 AM233375 
Fockea sinuata Namibia, near Maltahóhe, Bruyns 5670 (MO, WIND) A M233394 AM231769 AM233376 
Raphionacme monteiroae Tanzania, near Gulwe, m 9652 (K) AM233395 AM231770 AM233377 
Rhyssolobium dumosum South Africa, Alexander ruyns 3948 (BOL)  AM233396 AM231771 AM233378 
Secamone filiformis South Africa, 5 Duns 6990 (E) AM233397 AM231772 AM233379 
Telosma africana South Africa, Empangeni, Bruyns 9340 (BOL) — AM231773  AM233380 
deae). However, they still considered that the data. Using trnL-F data, Potgieter & Albert (2001) 


“presence of a gynostegial corona fused into an 
undulate annulus around the base of the gynostegium, 
in addition to a staminal corona of free segments” 
2003: 70) was one distinguishing 


This is despite a similar 


(Verhoeven et al., 
feature of the Fockeeae. 
construction being present in Sarcostemma viminale 
and the fact that the corona in Fockea does not form 
the base of the 


annulus around 


tubular structure thal exceeds 


an “undulate 


gynostegium" but is a 
the length of both the anthers and the style head and 
dominates the gynostegium. 

The 
confused. Consequently the Fockeeae was abandoned 
in Endress and Bruyns (2000), and the two genera 


morphological position is therefore quite 


were again referred to the Marsdenieae. 
Molecular results (Sennblad, 1997; Civeyrel & 
Rowe, 2001; Potgieter & Albert, 2001; Verhoeven et 


al., 2003) suggested that at least some members of 
Fockea and Cibirhiza are sisters to the remainder of 


the Asclepiadoideae. Sennblad (1997) found that, 
according to rbcL data, F. tugelensis N. E. Br. was 


nested within the Periplocoideae and C. albersiana H. 
Kunze, Liede & Meve was sister to the Asclepiadoi- 
deae. However the specimen (Sennblad 237) that had 
been determined by U. Meve as F. tugelensis was later 
found to be Petopentia natalensis (Schltr.) Bullock of 
the Periplocoideae (Sennblad. 1997, note added in 
proof) and the specimen (Specks 248) identified as C. 
albersiana was Fockea multiflora (Verhoeven et al., 
2003) so that, according to rbcL data F. multiflora is 
sister to the Asclepiadoideae. Civeyrel & Rowe (2001) 
found the same position for F. capensis with matK 


found that F. cylindrica R. A. F. edulis, and F. 
sinuata were sister to the Asclepiadoideae. However, 
the credibility of these results is again called into 


question by the misidentification of material, since 


yer, 


these authors found distinct positions on their clado- 
gram for Stapelia revoluta Masson and Tromotriche 


revoluta (Masson) Haw., although the former is the 


basionym of the latter! Using data from one 
chloroplast gene region, Verhoeven et al. (2003) 
showed that Cibirhiza dhofarensis Bruyns, Fockea 


edulis, and F. multiflora formed a monophyletic group 
that is sister to the remainder of the Asclepiadoideae. 

We build on the results of Verhoeven et al. (2003). 
Using both molecular and morphological data, we 
show that the two species of Cibirhiza and all six 
species of Fockea belong to a monophyletic tribe, the 
Fockeeae, and we elucidate the relationships among 
the species of Fockea. We give a detailed account of 
all species of Fockea, paying particular attention to 
the three widespread and variable species, F. 
angustifolia, F. comaru, and F. multiflora. Informa- 
tion on the distribution of the species has been much 


extended since Court (1982), and these data are 
included. 
MATERIALS AND METHODS 

Material from the herbaria BM. 19 5 , COL, GRA, K. 


KMG, LISC, M, NBG, NU, P, PRE. S. SAM. WIND. 
and Z was examined. Vouchers used ^» the molecular 
analyses are listed in Table 1. Table 2 provides the 


scoring for the morphological character states. 


Annals of the 
Missouri Botanical Garden 


Table 2. 


characters (1-13) (see App. 1 for definitions) used in the 


Character states (0, 1) for the 13 morphological 


cladistic analyses. A dash indicates that the character is nol 
mark that the i 


applicable and the question state ds 


unassignable. 


Morphological characters 


00000 00000 111 

Species 12345 67890 123 
Cibirhiza albersiana 11001 10111 010 
Cibirhiza dhofarensis 11001 LOLLI 010 
Fockea angustifolia, acc. 1. 11001 00110 000 
Fockea angustifolia, acc. 2 11001 00110 000 
Fockea angustifolia, acc. s 11001 00110 OOO 
Fockea angustifolia, acc. e 11001 00110 000 
Fockea capensis 10100 10110 000 
Fockea comaru, ace. | 11001 00110 000 
Fockea comaru, ace. 2 11001 00110 000 
Fockea edulis 1010? 0110 000 
Fockea multiflora, ace. | 0--01 10110 000 
Fockea multiflora, ace. 2 0--01 10110 000 
Fockea multiflora, ace. 3 0--01 10 10 000 
Fockea sinuata 11000 00110 000 
Raphionaeme monteiroae 11001 110-- -]- 
Rhyssolobium dumosum 0--11 000 - - -11 
Secamone filiformis 0--01 0110- 110 
Telosma africana 0--1I 1010- 101 


Illustrations were made from living material in some 


cases and in others from material preserved in 


alcohol. 


INGROUP AND OUTGROUP SAMPLING 


Raphionacme monteiroae (Oliv.) N. E. Br., Rhysso- 
lobium dumosum E. Mey.. Secamone filiformis (L. f.) J. 
H. Ross, 


outeroups (Civeyrel & Rowe, 2001: Verhoeven et 


and Telosma africana were taken as 


al., 2003), with R. dumosum and T. africana chosen 


from the Marsdenieae, where the Fockeeae had 


previously been placed. Secamone filiformis was 
chosen as a representative. of the Secamonoideae, 
while Raphionacme monteiroae, from the Periplocoi- 


t al., 


deae, was used to root the tree (Verhoeven « 
2003). We treated the eight species (12 accessions) of 
the Fockeeae as our ingroup. Representatives of all 
six species of Fockea were used, and several 
accessions were included for the widespread species 
multiflora. Both 


species of Cibirhiza, C. dhofarensis and C. albersiana., 


` 


F. angustifolia, F. comaru, and F. 
were incorporated. In total, 18 accessions were 
included in the le and psbA-trnl data-sets. 
There were 17 accessions in the ITSI data-set, which 


excluded. Telosma because a sequence could not be 


obtained. For vouchers and Genbank accession 


numbers, see Table 1. 


DNA EXTRACTION AND AMPLIFICATION OF TEMPLATE DNA 


Total DNA was isolated from fresh leaf material of 
the 18 samples by the method of Saghai-Maroof et al. 
(1984) as modified by Doyle and Doyle (1987). The 
three DNA regions were amplified from total DNA by 
polymerase chain reaction (PCR). The ul region 
(consisting of the adjacent tral. intron and trnL-F 
intergenic spacer) was amplified using primers c and f 
(Taberlet et al., 1991). The psbA-trnH intergenic spacer 
was amplified using primers psbAF and trnHR (Sang el 
al., 1997). The ITS region was amplified using primers 
ITS2 (Baldwin, 1992) and I8KRC (5 -GCACGCGCGC- 
TACACTGA-3’). Twenty-five microliter reactions were 
prepared, which contained 19 uL of. sterile water. 
2.5 uL of 10X PCR buffer, I uL of 10 mmol/L dNTPs 
in equimolar ratio, 0.75 uL of each 10 Umol/L primer, 
0.5 uL of 25 mmol/L. MgCl», 0.125 uL of Taq DNA 
polymerase (5 U/uL), and 0.5 uL of genomic DNA. In 
general, a 10X dilution of the extraction product was 
used as template DNA. Thermal reactions were carried 


out as follows: first 2 min. at 97 C, initial denaturation, 


E 


cycle: denaturing at 95 € for | min.: annealing at 
52 C for 45 sec.: extension at 72°C for 2 min; the 
three eycles repeated 30 times, and finally followed by 
72°C. The 


cleaned using the QlAquick PCR purification kit from 


extension for 8 min. al reactions were 
Qiagen and the purified products were eluted in 60 uL 


of sterile water. 


SEQUENCING AND ALIGNMENTS 


Both 


sequenced, using the ABI PRISM Dye Terminator 


strands of the PCR products were cycle 


Cycle Sequencing Ready Reaction Kit (Applied 
Biosystems, Foster City, California. U.S.A.). The 
primers used for amplification were also used for the 
sequencing reactions. The samples were run on an 
Applied Biosystems 377 automated DNA sequencer. 

Data assembled and edited 
GeneDoc version 2.6.002 (Nicholas & Nicholas, 
1997) and Chromas version 1.43 (McCarthy, 1996— 
1997). Sequences were aligned by eye. Ambiguous 
IUPAC 


retention of 


files were using 


positions were coded using appropriate 


ambiguity symbols so as to maximize 


information. The method of simple indel coding o 
Simmons and Ochoterena (2000) is followed, in which 
all gaps were treated as single sites, coded as binary 
characters and added to the data matrix. Áreas of 
ambiguity in the alignment were excluded from all 


analyses. 


Volume 93, Number 4 
2006 


Bruyns & Klak 
Systematic Study of Fockea 


539 


Table 3. Taxa used in the analysis of ancestral areas 
and their presence (1) or absence (0) in six regions: (1) Cape 
Region; (2) Succulent Karoo Region; (3) N 
(4) Tropical and a al Southern 
Tropical Africa; and (6) Arabia. 


ama Karoo Region; 
Africa; (5) South 


hegion 
Species | 2 3 1 5 6 
Secamone filiformis 0 0 0 | l 0 


Cibirhiza albersiana 0 0 0 0 l 0 
Cibirhiza dhofarensis 0 0 0 0 0 | 


Fockea angustifolia 0 0 I l ] 0 
Fockea capensis l | 0 0 0 0 
"ockea comaru l 1 1 0 0 0 
Fockea edulis i 0 0 l 0 0 
Fockea multiflora 0 0 0 | | 0 
Fockea sinuata 0 l l 0 0 0 


CLADISTIC ANALYSES 
Three 


of combined trnL-F and psbA-irnH sequences for 


matrices. were analyzed, one consisting 
all 18 accessions and a second of ITSI sequences for 
Due to 


inheritance of the plastid 


17 accessions (excluding Telosma). the 


uniparental mode of 
genome, we would not expect incongruent patterns 
of relationships from the different plastid regions, and 
so we analyzed only the combined matrices for the 
nF and psbA-trnH data. A third matrix consisted of 
irmL-F, ITSI, and psbA-trnH 


sequences for all 18 accessions (Telosma coded as 


combined spacer 

2%). No tests of congruence were performed before 
combining the data, because incongruent nodes were 
at most weakly supported (Jackknife support percent- 
age, JK, < 63) 

All cladistic analyses were performed using the 
parsimony algorithm of PAUP* version 4.0b4 (Swof- 
2000). 
1000 replicates of random taxon-addition to find local 
TBR 


reconnection) branch swapping with MULPARS on, 


ford, Each data matrix was analyzed using 


minima in the tree-space, (tree bisection- 
and all character transformations treated as equally 
likely (Fitch, 1971). At most two trees were saved on 
each replicate to minimize time spent swapping at 
local minima in the tree-space. All trees found in the 
initial 1000 replicates were then used as starting trees 
for a second round of TBR swapping with a tree limit 
of 30,000. To assess clade support, the data were 
analyzed using the jackknife option as implemented in 
PAUP*. Itis known that, with a removal probability of 
36.79%, JK 63 or 


supported by at least one uncontradicted character 


more corresponds to a node 
and may be regarded as supported by the data (Farris 
et al., 1996; Backlund & Bremer, 1997; Bakker et al., 


1998). Therefore, the percentage of deleted characters 


for all jackknife analyses in this study was fixed at 
36.19%. Only groups with JK greater than 50 were 
retained. 

The ancestral area method of Bremer (1992) was 
used to locate the geographic origin of Fockea. The 
analysis was performed using PAUP*, with Secamone 
filiformis and the two species of Cibirhiza as out- 


groups. The area over which the species occur was 
divided into six regions: (1) Cape Region, (2) 


Succulent Karoo Region, (3) Nama Karoo Region, 
(4) Tropical and subtropical Southern Africa, (5) 
South Tropical Africa, and (6) Arabia. The occurrence 


of the species in these areas is given in Table 3. 
ResuLTS 


Statistics, including numbers of variable positions 


and measures of consistency, are given for each 
analysis in Table 4. 
CHLOROPLAST £rnL-F AND psba-trni ANALYSIS 

The strict consensus (Fig. 1) of the four trees 


recovered shows a monophyletic Fockea (JK 52), 
which is sister to the two species of Cibirhiza (JK 78). 


Within Fockea, 


form a polytomy 


— 
— 


three clades are retrieved and these 
with F. sinuata 
The three clades are: (1) 
the four accessions of F. angustifolia (JK 98), (2) the 
specimens of F. multiflora (JK 99), and (3) F. 
capensis and F. edulis (JK 63) 


and the two 


accessions of F. comaru. 


three 


NUCLEAR (ITS1) ANALYSIS 


The strict consensus (Fig. 2) of the four trees 
recovered shows a highly supported monophyletic 
Fockea (JK 98), which is sister (JK 100) to Cibirhiza. 
The nuclear data achieved a higher resolution among 
the species of Fockea than the chloroplast data. Within 
Fockea, 


a monophyletic clade (JK 96), which is sister to the 


the three accessions of F. multiflora form 


remainder of Fockea (JK 70). The remaining taxa fall 


— 


into a polytomy consisting of two major clades, F. edulis 
and F. capensis. The two major clades are: (1) all four 
accessions of F. angustifolia (JK 71) and (2) F. sinuata 
and the two accessions of F. comaru (JK 75) 


COMBINED MOLECULAR ANALYSIS 


As 1n the previous analysis, the strict consensus of 
the two trees recovered (not shown) shows a well- 
“ockea (JK 90), which is 


sister to Cibirhiza. With both nuclear and chloroplast 


supported, monophyletic 


data, the relationships within Fockea are well resolved 


and mostly well supported. The three accessions of F. 


Annals of the 


Missouri Botanical Garden 


Table 4. 


Statistics for cladistic analyses and trees for Fockea and related taxa. Cl, Consistency index; RI, 


retention index; 


RC, rescaled consisteney index. 
Total Variable Informative No. of Tree 

Data partition characters characters characters shortest trees length Cl RI RC Indels 
trnlL-F 8064. 41 26 — — — — — 2 
psb A-trnH 401 36 P — — — == 8 
Chloroplast 1265 77 103 |. 165 0.752 0.823 0.619 10 
ITS] 637 113 69 4 134 0.761 0.799 0.608 3 
All molecules 1902 190 172 2 309 0.731 0.788 0.576 13 
Molecules and 1915 203 185 4 327 0.722 0.781 0.563 13 


morphology 


multiflora are again sister (JK 82) to the remaining 


species of Fochea. The remainder forms a well- 
supported clade (JK 100) consisting of the four 
accessions of F. angustifolia, which is sister to 
a weakly supported clade (JK 53) made up of two 
main clades. These two clades are (1) F. capensis and 


F. edulis (JK 83) and (2) F. 


accessions of F. comaru (JK 70). The two accessions 


sinuata and the two 


of F. comaru are not sister taxa, since one of them is 
sister to F. sinuata. However, this relationship gained 
only minimal support (JK 52) and may therefore be 
spurious. 

We show one of the two equally parsimonious trees 
as a phylogram (Fig. 3). The relatively long branches at 
the bases of the various accessions of F. angustifolia 
and F. multiflora show that the two tropical species are 
well separated from each other and from the remaining 
four species of Fockea, which have accumulated far 
fewer molecular changes. There are also numerous 
molecular changes among the four accessions of F. 
angustifolia. Similarly, in the other widespread species, 
F. multiflora, there are several molecular changes. In 
F. 


sinuata there is only one molecular change, which 


contrast, among the more local F. comaru and 
places one accession of F. comaru (PVB 8821) as sister 
to F. sinuata. These two, in turn, are sister to the second 


Dold 2381). 


— 


accession of F. comaru 


COMBINED MORPHOLOGY AND MOLECULAR ANALYSIS 


Here data for all 18 accessions were analyzed, 
including the 13 morphological characters listed in 
Appendix I. In the strict consensus (not shown) of the 
four trees recovered, Cibirhiza and Fockea are mono- 
phyletic and sister to one another (JK 95 & 88, 
respectively). Fockea multiflora remains sister to the 
rest of Fockea, which forms a strongly supported clade 
(JK 85). Less resolution is achieved among the other 


species of Fockea, with the weakly supported southern 


clade (F. comaru and F. sinuata) of Figure 3 
collapsing to a polytomy with F. angustifolia. 


BIOGEOGRAPHY 


Results of the ancestral area analysis (Table 5) 
show that the highest ratio of gain to loss occurs in the 
South Tropical Africa region (Area 5), followed by 
Tropical and subtropical Southern Africa (Area 4). 
The lowest ratios are in the Succulent Karoo Region 
(Area 2) and in Arabia (Area 6). A higher gain to loss 
ratio for a given area indicates a higher probability 


that this area is ancestral. 


DISCUSSION 
1. MONOPHYLY OF FOCKEA 


In all of our analyses Fockea is monophyletic and 
sister to a monophyletic Cibirhiza. Although the 
monophyly of Fockea is poorly supported in our 
JK 52. Fig. I). both the nuclear 


JK 98, Fig. 2) and the combined molecular analyses 
X o J 


chloroplast analysis ( 


(JK 90, Fig. 3) show high support for the monophyly of 
Fockea. The long branch subtending the Fockea clade. 
shown in Figure 3, indicates that Fockea is well 
separated from its sister Cibirhiza. Morphologically 
the two genera are well separated too. Fockea is 
characterized by the tubular outer corona enclosing 
the anthers and much exceeding them, with the inner 
lobes partially fused to the inside of this tube, and the 
erect, swollen appendages of the anthers that more or 
less completely fill the corona tube and much exceed 
the style head. Cibirhiza is characterized by a low 
outer corona ring surrounding the whole gynostegium, 
with a tooth behind each slender and terete inner lobe. 
The anther appendages are short and adpressed to the 


style head. 


2. RELATIONSHIPS AMONG THE SPECIES OF FOCKEA 


Fockea consists of two widespread, mainly tropical 


species (F. angustifolia and F. multiflora) found from 


— 


Kenya to Namibia and South Africa and four much 


more local taxa (F. capensis, F. comaru, F. edulis, and 


Volume 93, Number 4 
2006 


Bruyns & Klak 
Systematic Study of Fockea 


541 


— Cibirhiza albersiana 
92 | 


— Cibirhiza dhofarensis 


— Fockea angustifolia acc. 4 
— Fockea angustifolia acc. 2 
— Fockea angustifolia acc. 1 


— Fockea angustifolia acc. 3 


— Fockea capensis 
63 | 


— Fockea edulis 


Fockea comaru acc. 1 


Fockea comaru acc. 2 


Fockea multiflora acc. 1 
95 — Fockea multiflora acc. 2. 


— Fockea multiflora acc. 3 


100 


Fockea sinuata 


— Rhyssolobium dumosum 


— Telosma africana 


Secamone filiformis 


Figure 1. 


Raphionacme monteiroae 


Strict consensus of four trees obtained from analysis of combined data from the two chloroplast genes trnL-F 


and psbA-trnH. Numbers above branches indicate jackknife support percentages. 


F. sinuata) that are endemic to Namibia and South 
Africa. The starkest morphological contrast within the 
genus is beween F. multiflora, which is a massive, 
tropical liane without a tuber, and the others, which 
are relatively small climbers with a swollen, mostly 
subterranean tuber. 

Relationships within Fockea are well resolved and 
mostly well supported. In addition, the relationships 
suggested by the molecular results are frequently 


congruent with those from the morphological char- 


acters. F. multiflora is sister to the rest of Fockea. 
Among the tuberous species, the mostly tropical F. 
angustifolia, is, in turn, sister to the remaining taxa 
that are endemic to southern Africa. Among the 
southern African taxa, F. capensis and F. edulis are 
well supported sisters and both have large, slightly 
exposed tubers, substantial stems, and elliptic to 
nearly circular leaves. Relationships among the small 
species having slender stems and mainly linear leaves 


are more complex, with the two southern African taxa 


Annals 


542 the 
8 Botanical Garden 


mm Cibirhiza albersiana 


— Cibirhiza dhofarensis 
Fockea angustifolia acc. 4 


Fockea angustifolia acc. 2 


Fockea angustifolia acc. 1 


100 


Fockea angustifolia acc. 3 


Fockea capensis 


Fockea comaru acc. 1 


98 


Fockea comaru acc. 2 


Fockea sinuata 


Fockea edulis 


Fockea multiflora acc. | 


96 Fockea multiflora acc. 2 


Fockea multiflora acc. 3 


Rhyssolobium dumosum 


Secamone filiformis 


Figure 2. Strict 
jackknife support percentages 


(F. comaru and F. sinuata) grouping together but V. 


angustifolia remaining well outside this grouping. 
This is particularly interesting as F. angustifolia and 


F. comaru may occur together and are morphologi- 
cally very similar, 
The phylogram (Fig. 3) shows that both of the 


widespread, mainly tropical species, F. angustifolia 


and F. multiflora, lie on comparatively long branches, 


consensus of four trees obtained from analysis of nuclear ITS! data. 


Raphionacme monteiroae 


Numbers above branches indicate 


while the more local southern African endemics lie on 


shorter branches. This corroborates studies in other 


families, where extensive radiation in southern Africa 
has been accompanied by relatively small changes in 
eene Klak et al., 2003: 


As in Fockea, radiation in these other families 


the 
2004). 
is especially associated with the arid winter rainfall 


regions examined (e... 


zone and its margins. In addition, considerable 


Volume 93, Number 4 


Bruyns & Klak 543 


2006 Systematic Study of Fockea 
96 [ Cibirhiza albersiana 
Cibirhiza dhofarensis 
57 — Fockea angustifolia acc. 4 
85 | — Fockea angustifolia acc. l 
100 * Fockea angustifolia acc. 2 
98 
Fockea angustifolia acc. 3 
82 
— 8 Fockea capensis 
Fockea edulis 
53 
7 Fockea comaru acc. 1 
90 
70| — Fockea sinuata 
— Fockea comaru acc. 2 
Fockea multiflora acc. 1 
100 

Fockea multiflora acc. 2 
Fockea multiflora acc. 3 

Rhyssolobium dumosum 

Telosma africana 
Secamone filiformis 
Raphionacme monteiroae 
—— 5 changes 


Figure 3. 


Phylogram of one of two trees obtained from combined analysis of chloroplast ¢rnL-F, pshA-trnH, and nuclear 


ITSI data. The lengths of the branches are in proportion to the number of changes in the nucleotides. The arrow indicates 
8 pror t 
a node that collapses in the strict consensus tree. Numbers above branches indicate jackknife support percentages. 


molecular variation is found among the different 
accessions of each of the widespread, mainly tropical 
species of Fockea. In fact, the phylogram shows that 
these accessions vary more among each other than do 
all the in the 


relatively few molecular changes have occurred. It is 


accessions in southern clade, where 


— 


possible that these tropical species of Fockea are 


much more ancient than the southern taxa. 


3. RELATIONSHIPS WITHIN SPECIES IN FOCKEA 


Our sampling contained multiple accessions of 
three of the most variable and widespread species. 

We included three accessions of Fockea multiflora. 
In each of the molecular analyses, they formed a clade 
that is strongly supported. In the combined chloro- 


plast analysis (Fig. 1) and the combined molecular 


Annals of the 
Missouri Botanical Garden 


Table 5. The number of Fockea species gained and lost for each of six geographic areas. The proportion of gains/losses 


was used in the ancestral area analysis (Bremer, 1992 


to determine the geographic origin of Fockea 


Gains Losses Gains/Losses 
Area l: Cape Region ) 5 0.6 
Area 2: Succulent Karoo Region 2 5 0.4 
Area 3: Nama Karoo Region 2 l 0.5 
Area 4: Tropical and subtropical Southern. Africa | 3 1.3 
Area 5: South Tropical Africa | 2 2.0 
Area 6: Arabia l 3 0.3 


analysis (Fig. 3), the southernmost collection is sister 


to the others. 

We included four accessions of Fockea angustifolia 
and found that they consistently formed a clade that is 
strongly supported in the combined chloroplast 


analysis (Fig. 1) and in the combined molecular 


analysis (Fig. 3), though less so in the ITSI analysis 
(Fig. 2). In the combined molecular analysis (Fig. 3), 


is sister to 


the northernmost accession (Bruyns 9632) 
the others, which form a strongly supported clade 
relative to it. The wide morphological variation known 
in F. 


any significant morphological differences between 


jam 


angustifolia in southern A 
these northern collections and those further south 
make the recognition of separate taxa here impossible, 
despite the strong support for clades. within F. 
angustifolia. 

We included two accessions of Fockea comaru. In 
the combined molecular analysis (Fig. 3), one acces- 
sion of F. comaru (Dold 2381) comes out as sister to 


a clade consisting of F. comaru and F. sinuata. 


Figure 3 shows that this is caused by a difference in 
a single amino acid, which is also reflected in the 


weak support for the F. comaru—F. sinuata clade (JK 


I. BIOGEOGRAPHY AND ECOLOGY 
Our ancestral area analysis suggests that Fockea 


South Africa 


Fockea is endemic to Africa south of the 


originated in Tropical and radiated 
southward. 
equator, and most found in southern 


Africa. 


multiflora) extend north of this. Both of these reach 


species are 


Only two species (F. angustifolia and F. 


Tanzania and are the most widely distributed 


members of the genus. Fockea angustifolia is the 
most widely distributed of all and is recorded from the 
subtropical Kimberley and Prieska area of central 
South Africa (in and to the east of the Nama Karoo 
1991) to southern Kenya. Fockea 
hand, 


Region of Jürgens, 
multiflora, 


species occurring in northern Namibia and southern 


on the other a purely tropical 


Angola, the basins of the Okavango and Zambezi 


rica and the lack of 


Rivers, and in central Tanzania. The distributions of 
both F. angustifolia (Fig. 5) and F. multiflora (Fig. 9) 
exhibit a marked disjunction between the valley of the 
Zambezi River and central Tanzania. These patterns 
coincide largely with the more arid woodland or scrub 


Mill., 


Delile (where they are 


of Acacia Commiphora Jacq., and Balanites 
found) and the more mesic 
woodland dominated by Brachystegia Benth., where 
they tend not to occur. 

The remaining four species are endemic to the semi- 
arid to arid, temperate to subtropical parts of South 


Africa Fockea 


restricted, occurring on the Little Karoo and in some of 


and Namibia. capensis is the most 
the drier surrounding mountains in southern South 


1991). 


Fockea edulis is mainly found in semi-arid patches in 


Africa in the Succulent Karoo Region (Jürgens, 


the coastal parts of southern and southeastern South 
Africa from the Succulent 
Fockea and F. sinuata are 
South Africa 


comaru is strongly (but not exclusively) associated 


Karoo Region eastward. 


comaru more widely 


distributed n and Namibia. Fockea 


with the western part of South Africa, receiving winter 
rainfall, and is particularly found in the Succulent 
Karoo Region. Hs distribution extends into the Nama 
Karoo Region (Jürgens, 1991) of central South Africa. 
Fockea sinuata is mainly found within the Nama Karoo 
and on the eastern margins of the Sueculent Karoo 
Region. The shorter branches subtending these four 
species suggest that this radiation may have been more 
recent than the radiation within F. angustifolia and F. 
multiflora, where the branches are mostly longer. 
Fockea angustifolia, F. comaru, and F. sinuata all 
occur close together in some areas on the Great Karoo 
of South Africa but, whereas F. angustifolia and F. 
comaru may actually grow in the same habitat, F. 


which F. 


angustifolia is rarely found and where F. 


sinuata occupies. a distinct niche in 
comaru 
never grows. Fockea comaru occurs together with F. 
edulis on the Worcester-Robertson Karoo (southwest- 


with F. 


capensis on the Little Karoo of southern South Africa. 


ern South Africa) and very occasionally 
Most species of Fockea occur on rocky hills, where 


the edible tubers are wedged tightly among stones and 


Volume 93, Number 4 
2006 


Bruyns & Klak 545 


Systematic Study of Fockea 


protected from excavation. Fockea angustifolia occa- 
sionally grows in flat, pan-like areas, and F. sinuata is 
exclusively found in calcareous, 


almost pan-like 


areas, which become seasonally but very intermit- 
tently moist. All species, except F. pe have 
a geophytic tendency, often with the ability to die 


back to the tuber under very dry bon 


5. POLLINATION 

Two facts suggest that the method of pollination in 
Fockea is not the same as in most other asclepiads. 
Firstly, the flowers are remarkable for the extremely 
restricted access to the pollinaria, and, unusually, the 
main organs causing these restrictions are the anther 
appendages, which block the corona tube. Between 
adjacent inner corona lobes there are three narrow 
channels leading to the base of the corona. 


these are on the sides of the long outer corona lobules. 


7 


and a fine hair inserted down either of these 
directed. exactly to the base of the corpuscle. The 
other channel lies behind the lowest part of the outer 
corona, and it provides access to the nectarial orifice 
below the corpuscle. Secondly, the margins of the 
anthers are schlerenchymatous only for a very small 
distance near their base or are not hardened at all. 
The schlerenchymatous part is found below the 
corpuscle but does not continue below it for some 
distance, as is customary. 

In. field-collected flowers, it has been found in 
Fockea that very few pollinaria are removed from their 
position on the style head. The only germinating 
pollinia have been found in the base of the longer 


channels reaching the nectarial orifices, and i 
appears that pollination is achieved if pollinia are 
deposited. sufficiently far down these channels. The 
rudimentary guide-rails appear to direct a probing 
This 


enables the pollinarium to be removed, but the guide- 


proboscis toward the base of the corpuscle. 


rails seem to play no role in the deposition of the 
It seems that the weak corpuscle partly 
breaks up and in so doing probably causes the 
pollinium to be wiped off the pollinator's proboscis in 
the narrower parts of the channels (the guide-rails may 
assist in wiping them off too). Evidence suggests that 
the same mechanism is involved in Cibirhiza (Kunze 
et al., 1994). 

Pollination in the Periplocoideae is achieved when 
pollen is wiped off the spoon-like receptacle of the 
translator onto the base of the style head (Bruyns € 
Klak, obs.). 


absent. In the Secamonoideae, guide-rails are present 


pers. Here guide-rails are altogether 


and they guide potential pollinators toward the 
corpuscle for removal, but it is unknown what role 
they play in the deposition of pollen. In the Fockeeae, 


rudimentary guide-rails are present and they guide 


Two of 


potential pollinators toward the corpuscle for removal, 
but they appear to play no role in the deposition of 
pollen. These three groups are successive sisters to 
the rest of the Asclepiadoideae, in which the guide- 
rails are better developed. In the rest of the 
Asclepiadoideae, the guide-rails assist in the removal 
of the pollinarium but have the unique additional 
function of holding the inserted pollinium as the 


pollen tubes grow. 


TAXONOMY 


Fockea Endl., in Endl. & Fenzl, Nov. Stirp. Dec. 3: 
17. 1839. TYPE: Fockea capensis Endl. 
: 23-24 (1842). TYPE: 


d ormus Harv., London J. Bot. 
Uitenhage, Zeyher 965 (K!). 


Erect to twining herb to massive liane often with 
underground to slightly exposed tuber tapering into 
small taproot anchoring plant, with fibrous roots, sap 
milky. Stems dextrorsely twining or erect, 80 mm— 
15 m long, 2-300 mm thick, slightly fleshy, finely 
pubescent when young. later covered with a shiny gray 
bark. Leaves opposite, spreading, subcoriaceous, gray- 
ereen to green, pubescent (with fine verrucose hairs) 
to glabrous; petiole 0--25(-45) mm long, spreading: 
blade 13-60(-150) X 1.5-20(-100) mm, 


elliptic, acute to obtuse, apiculate, margins slightly to 


linear to 


strongly revolute, non-succulent, deciduous, midrib 


paler than rest. Inflorescences extra-axillary, 1- to 20- 
flowered, umbellate, pubescent; peduncle 0-15 mm 
long, stout, with several short bracts at bases of 
pedicels; pedicels 1-15 mm long. Sepals adpressed to 
corolla tube, lanceolate, acute, 1.5—3 | mm, 
pubescent outside and glabrous EUM corolla rotate, 


8-22(-40) mm diam.; 


inside puberulous to glabrous, green to brown; tube 


outside pubescent, gray-green; 


1.5-5 mm long, ca. 2 mm broad at mouth, + glabrous 
lobes (3—)5-12(-18) X ca. 


spreading, usually spiralling to left, 


within; 2mm, linear, 


obtuse, often 


twisted to left when viewed along lobe; corona 
consisting of 2 series of lobes, white, glabrous: outer 
forming cylindical tube with several series of erect to 
spreading lobules arising at mouth of tube: inner 
consisting of 5 erect lobules partly fused to outer 
corona tube lower down and usually pressed to backs 
of anther appendages and exceeding them; anthers 
consisting of erect + quadrangular fertile part pressed 
to style head, above this with much inflated erect, 
transparent-white sterile appendage 2-3 mm long. 
these appendages connivent and pressed to one 
another, filling up area inside corona tube; style head 
less than 2 mm long, not exceeding fertile part of 
anthers, slightly bifid at truncate apex, below this 


widening to base of anthers then narrowing and sessile 


546 


Annals of the 
Missouri Botanical Garden 


on ovaries; pollinarium with 2 small translucent 
nearly flat and elliptical erect pollinia sessile on 
small brownish usually narrowly elliptical corpuscle. 
Follicles 50-220 X 8-30 mm, 


fusiform and narrowing into slender beak, 


pendulous, broadly 
uniformly 
glaucous green turning brown, glabrous, sometimes 
pustulate, single by abortion, containing 50 to 150 
seeds; seeds flattened, pear-shaped in outline, 6-14. X 
pale brown, with 


3-8 mm, pale beige or yellow 


usually somewhat paler margin, slightly papillate. 
with coma attached at one end except in F. sinuata 


where hairs extend on margin around seed. 


One of the defining features of the flower in Fockea 
is the complicated, white, tubular corona with many 


small lobules projecting inside and around its edges. 


This corona starts off as a series of vertical ridges 
below the lower margins of the anthers (forming the 
guide-rails), with an outward swelling also occurring 
below these ridges (Endress € Bruyns, 2000: 18, fig. 
9). Gradually three distinet parts appear to this outer 
corona: a horizontal lobule bridging the gap between 
from. the 

further 


the anthers, two diagonal ridges arising 


vertical ridges of the earliest. stages, and 
lobule behind and somewhat below each anther. Thus 


the outer corona consists of 15 lobules. An inner 
corona of five lobules appears slightly after the first 
parts of the outer series have begun to swell, and each 
of these lobes lies behind the anther and a little above 
the 


structures become obscured, and the various lobules 


outer series. In more advanced stages these 


of the outer corona fuse into a tube. The longest 
spreading lobules around the mouth of this tube are 
the ones in the outer series that arose directly behind 
These 


the anthers. longest lobules are flanked on 


either side by a somewhat channel-forming lobule. 


and these are derived from the two diagonal ridges of 


the outer series. The remaining lobule of the outer 
series forms a fairly low bridge between these 


channel-forming lobules to complete the tubular outer 
corona. The lobes of the inner corona are fused to the 
outer corona tube nearly to its apex, and the "dragged 
vascular traces of the anthers show that at least some 
of the vertical expansion of this tube is achieved by 
growth at the base of the anther. 

In the 


able basal fusion between the corona tube 


mature flower in Fockea, there is consider- 
and the 
corolla. As can be seen in Figure 4F and G, the join 


between the corolla and the corona tube is much 
higher than that between the corolla and the calyx. 
However, the ontogeny of the corona shows that this 


fusion is a later development, and the fact that the 


vascular trace of the corona branches off from that of 


the adjacent anther also suggests that this fusion is 


secondary and does not imply that the corona is 


contributed to by the corolla, though this was 


suggested as a possibility by Kunze (1996: 29). 
Kunze (1996: 553, figs. 


that the anther in Fockea edulis and F. 


3—5; 594, fig. 12) has shown 
sinuata contains 
endothecial tissue that almost completely surrounds the 
He also found that there is 
still 


two locules in each anther. 


some dorsal dehiscence of the anther, as if it 


contained four locules rather than two. 

Another defining feature of Fockea is the cluster of 
five soft, erect, inflated, translucent white sterile 
appendages of the anthers which fill the corona tube 


(Fig. AF. 


inflated form relatively late. 


These appendages only develop their 
In the early stages there 
is a distinct separation of the fertile and sterile parts 


of the anther by slight indentations on either side al 


the base of the sterile part (Endress & Bruyns, 2000: 
18. fig. 9), as is commonly found in the mature anther 


of the Asclepiadeae. Beneath each appendage the 


short, roughly square anther is pressed inward onto 


the style head. The lower margins of the anthers are 


slightly hardened and hyaline toward. their base, 


forming rudimentary guide-rails just below each 


corpuscle. 
The two ovaries in each flower are together about as 
broad as tall, and they contain many small ovules 


(Fig. 4F. € 


ovules continue somewhat below the level of fusion of 


G). They are only partly superior, because the 


the calyx and the corolla. The comparatively short style 


is sessile on the ovaries. It is broadest at the level 


heac 
at which the corpuscle is secreted, after which it 
narrows into a short, truncate, and slightly bifid apex 
that does not exceed the fertile parts of the anthers. 


The 


y flattened, clear to translucent white, elliptical 


pollinarium in Fockea consists of two small, 


usual 
pollinia and a a ol almost equal size (Fig. 6D, 
I; Fig. 8G, M. 
flat that 1 they make up no more than the 
thickness of the corpuscle. Kunze (1996: 553. figs. 3. 


5) has shown that in F. sinuata the pollinium | is mostly 


; Fig. 10H). The pollinia are often so 


made up of just two layers of pollen grains, while in F. 
edulis it would appear that only a single layer is 
present. The corpuscle is pale brown. (often only 
brown and shiny around the top), and it is generally 
broader when viewed from the side than when seen 
from above. It is very easily broken into two halves. 
Near the apex of the corpuscle on the side closest to 
the style head there is a small process to which both 
the pollinia are attached, and this may be quite easily 
visible with the light microscope in some plants. In all 
species a caudicle joining each pollinium to. the 
corpuscle is lacking. Around its base the structures by 
which the corpuscle is attached to the style head may 
be visible as slight protrusions. Kunze (1993) showed 
that the corpuscle has flanks and a Y-shaped floor 


that disappears toward the base of the corpuscle. 


Volume 93, Number 4 
2006 


Bruyns & Klak 
Systematic Study of Fockea 


547 


Figure 4. 
D. Face view of cotyledon. —E. 
ca, er co, corolla tube: a, anther; aa, sterile anther a appe 


style head; ov, ovary. Drawn from: A, C, E, 


(at G 


found that in F. 


(2001) 


angustifolia, F. multiflora, and F. sinuata the grains 


Verhoeven and Venter 


are grouped into tetrads and there is no outer envelope 


or pollinium wall. 


AE. Seedlings in Fockea, about 20 days old. F, G. Half flowers in Fockea. 


Face view of one of second e of a c, colyle 


comaru, Bruyns 3329 (BOL); G, F. edulis, Bayer 325 (NBG). Scales: A, C, 
). 


G 


3. Side view of plantlet. —C, 
don; h, a hypocotyl: p, primary stem: 


uler corona tube: 1 .inner corona 


ndage; 
F. sinuata, Baan 3023 d B, D, F. edulis Bruyns 4995a 977 5 F, 


ecd 


— 2 mm (at B); — | mm 


— 3 mm (at A); 


Seedlings (Fig. 4) have been observed in all species 
of Fockea except F. multiflora. Each plantlet consists 
of a swollen, fleshy, green hypocotyl (at least half of 


which projects from the soil) with two broad. green 


Annals of the 
Missouri Botanical Garden 


cotyledons at the apex. The primary leaves and stem 
emerge from between the cotyledons after one to three 


weeks, and the primary leaves are always much 


narrower than the cotyledons. Swelling of the base of 


the hypocotyl leads to the formation of the tuber below 
ground level, while the upper part of the hypocotyl 
above the soil does not thicken. In species such as F. 
comaru and F. sinuata, where the tuber can be up to 
20 cm below the surface of the ground, the base of the 
hypocotyl elongates as well as swells so that gradually 
the tuber develops well below the surface. 

A similarly succulent hypocotyl is common in 
many of the succulent Asclepiadoideae and is even 
the Periplocoideae in such species as 
(Costantin & | Gallaud) 
Klack., Sacleuxia salicina Baill., and Stomatostemma 
Venter & D. Field. In the 


Asclepiadoideae it is found in the Fockeeae in 


known in 
Ischnolepis graminifolia 
pendulina H. 


Fockea, in the Asclepiadeae in Sarcostemma R. Br. 
and succulent members of Cynanchum L., and it is 
most common in the Ceropegieae (Bruyns, 2000). A 
succulent hypocotyl is unknown in the Marsdenieae 


(Forster, 1995). 


KEY TO THE SPECIES OF FOCKEA (SOUTHERN TO EASTERN ÁFRICA) 


la. Leaves rarely more than 30 mm long and not more 
than 20 mm broad; stems arising via narrow neck 
from distinct basal (mostly subterranean) tuber, 
s short and erect or climbing to 1—2 m. 
eaves linear. not more than 4 mm broad. 
3a. Margins of leaves revolute, leaves with 
dense adpressed pubescence, outer co- 
rona without CERE flap behind 
each longe sl terete lobule. 
4a. Margins of leaves strongly undu- 
late; corpuscle relatively massive 
and nearly half as broad as 
ps EREEREER 6. F. sinuata 
oa not undulate: 
ender and much 


le ss than half as geet as 8 


o 
^m 


ig ot bie ae " comaru 


3b. eae of leaves not revolute, 
cence, 


dues 
with sparse. pube ouler corona 
usually with spreading [lap Bi 11 cach 
longest terete lobule angustifolia 
2b. Leaves elliptic to nearly circ "M 6—22 mm 
broad. 
5a. Outer corona without spreading flap 
behind each longest terete lobule; upper 
surface of tuber projecting from ground. 
6a. Leaves + "puoi or very sparse- 


LE edulis 


Ob. Leaves finely and. densely mile s- 
cent, margins strongly undulate . . 

F. capensis 

5b. Outer corona with spreading flap be- 
hind each longest terete lobule; tuber 

well below surface of ground . . . . . . 


Ib. Leaves usually more than 30 mm long and 25 mm 
broad; stems usually massive and often climbing 
on trees to 10 m or more, plant without distinct 
basal tuber [i ͤap iii.... 5. F. multiflora 
I. Fockea angustifolia K. Schum., Bot. Jahrb. Syst. 


17: 146. 1893. TYPE: South Africa, Cape, 
Griqualand West, 48 km W of Kimberley, G. J. 
Lewis 07535 (neotype, designated here, PRE! 
duplicate, SAM!). Figure 6A-D. 

Fockea 5 Se s Bot. Di Syst. 20, Beibl. 51: 44. 
1895. PE: South Afri 
45007, 7 1 1894. A 
(lectotype, NR ps d here, drawing at tW, not seen). 

. E. , Fl. Trop. Afr. 4 (J): 429. 1903. 

Kwebe Hills, 


Tr ansvaal, 


mu lugardii N 
E: Ngamiland. 
299 (holotype, 5) 
Fockea Ld Schltr., 
TYPE: imibia. 
ee "BMD. 
sage gee N. E. Br., Fl. Cap. 4 (1): 778. 1908. TYPE: 
uth Africa. Natal, Tugela, W. Gerrard 1310 
ae Kl: isotype, BOL?). 


YPE: bo 

f ugard 
Jahrb. Syst. 
9, 


38: 56. 1905. 
a 1879, T. J. G. Een 


Fockea mildbraedii as in Mildbr., Wiss. Erg. Deut. 
Zentr.-Afr. Exped., 545—546. 1913. TYPE: 
arm Lembeni, » Sep. 1910, H. J. P. Winkler 3803 
(holotype, K!). 

Fockea monroi S. Moore, J. Bot. 52: 149. 1914. Type: 
Zimbabwe. Melsetter distr., Victoria, Monro 628 

lolype, de 'signale d here, > ) 


Cynanchum omissum Bu loc 
TYPE: Ke me $ 
kinnon road, m, Aug. 1953, 


Hemsley 4045 1 in K!). 


2 


. Samburu 
. B. 


Small geophytic herb to climber with underground 


napiform tuber (sometimes somewhat irregularly 


shaped by pressure of surrounding rocks) up to 
50 em long and 20 em thick with central growth 
point on smooth upper surface from which slender 
perennial and aerial stems arise, stems erect to 
twining to 1(-2) m, often whitish gray, softer upper 
deciduous, tomentose. 


parts young growth finely 


Leaves with petiole 1-3 mm long, grayish to brownish 
green (often with distinct brownish to purplish hue 
when young), linear to narrowly elliptic (to circular), 
13-65(-105) X (0.5—)1.5-8(-22) mm, obtuse to acute. 
glabrous to sparsely and finely pubescent (more 
midrib translucent. white, 


sparsely below), margins 


slightly to not at all. revolute, ania undulate to 
straight. Inflorescence l- to 6-flowered, flowers 


opening in succession, covered n fine adpressed 


wid peduncle absent; pedicels | X 1 mm. Sepals 1.5— 


X l mm, lanceolate, outside pubescent, inside 
He corolla 1O—40(—65) mm diam.; outside 
grayish green, pubescent; inside green to brown, 


sparsely and finely pubescent; tube 1.5-2 mm deep, 
ca. 2.5 mm broad at mouth, cupular; lobes 4-18(228) 


X l-2 mm, linear, obtuse, spreading and often 


Volume 93, Number 4 


Bruyns & Klak 549 


y 
Systematic Study of Fockea 


spiralling to left, often twisted to left when viewed 
along length, with margins strongly reflexed; corona 
white, glabrous; outer corona forming cylindrical tube 
4—6 mm long, divided at mouth of tube into 5 longer 
slender terete and often spreading lobules up to 5 mm 
long each flanked by 2 flattened spreading lobules, 
these groups of 3 lobes alternating with 5 linear and 
usually spreading lobules, longest outer lobes each 
usually with another flattened spreading lanceolate 
lobule behind it; inner corona of 5 terete linear lobes 

3 mm long adpressed to backs of anthers then 
erect above them. Follicles 70-200 X 8-12 mm, gray- 
green often banded with purple, smooth; seeds 8-10 X 


4—6 mm, yellow-brown. 


Phenology. Flowering October to May. 

Distribution. Angola, Botswana, Kenya, Namibia, 
South Africa, Swaziland, Tanzania, Zambia, and 
Zimbabwe; 200-1700 m (Fig. 5). 

Ecology. Fockea angustifolia occurs in areas 
receiving rainfall in summer. It is mainly found ir 
stony places, usually on slopes but occasionally in flat 


areas. In parts of the Kalahari it occurs on stony 
calcareous flats, while in Tanzania it grows in loamy 
flat areas. 

Discussion. The type (South Africa, Griqualand 
West, Groot Boetsap, 1200 m, H. W. R. Marloth 1008) 
of Schumann (1893) is missing, and a neotype has 
been designated here. No material seen by Schumann 
has been found. Consequently Lewis 67535 was 
selected as a neotype, because it is from the same 
general area as Marloth’s missing specimen and i 
represented in two herbaria. Moore (1914) cited two 
specimens of Monro in the protologue of F. monroi. 
One of these has been located and is selected as the 
lectotype. In the case of F. sessiliflora, no specimens 
of Schlechter 4493 have been located, but a sketch of 
the flower exists at W and this is selected as lectotype. 

K. Schumann (1893, 1895) and Brown (1907-1909) 
mentioned the name Brachystelma circinatum Marloth 


— 


non B. circinatum E. Mey. This appeared in a list of 
Marloth's collections where, under the Asclepiada- 


circinatum E. 


ceae, the entry “Brachystelma 
Meyer....bisher von den Witbergen im Kapland 
bekannt" appeared and the specimen Marloth 1008 


1889: 244). 


nor 


was cited (Engler, No description. or 


diagnosis was provided was any 


publishing a new name expressed, so the epithet 
circinatum Marloth does not exist 
(Greuter et al., 2000: ICBN Art. 32.1, Art. 34.1 (a)). 

Court (1982) treated Fockea angustifolia and F. 
were treated as 


Brachystelma 


comaru as one species, and they 
closely related also in Meve (2002). 
Our that Fockea 


angustifolia and F. comaru occur together over a wide 


own investigations revealed 


intent of 


area in Namibia and South Africa; the full extent of 
occurrence is probably still not accurately reflected 
by the recorded collections. In several cases both F. 
angustifolia and F. comaru have been noted on the 
same farm and have even been seen near each other 
on the same hillside. So far they have been recorded 
together in Namibia in the Tiras Mountains and the 
Great and Little Karas Mountains, while in South 
Africa they are known to grow together in the hilly 
southwestern part of Griqualand West. This wide 
sympatry leads one to suspect that they may not be 
closely related at all. 

Analyses of the molecular data proved our suspi- 


cions to be correct. They show that Fockea angusti- 


folia and F. comaru are two distinct species that are 


not particularly closely related to one another and that 
F. angustifolia is sister to all the southern species of 
Fockea. 

In the areas where they occur together, Fockea 
angustifolia and F. comaru look superficially similar. 
Both are narrow-leaved with slender, often twining 


stems and a relatively small tuber (at least in 


comparison to that of F. edulis). A number of features 
they can be 


can be used to separate them, and 


distinguished mostly even without flowers. In F. 


angustifolia the main stem generally divides above 
In F. 


divides mostly well m the surface, so that the stems 


the surface of the soi comaru the main stem 
of a single plant often spread extensively underground 
and may emerge from the soil some distance away from 
the central stem. The young leaves of F. angustifolia 
have a noticeably purplish to brownish hue and older 
leaves are grayish green, while the foliage of F. comaru 
is generally distinctly bluish green. In pressed speci- 
mens one finds that the younger stems and leaves of F. 
angustifolia are often very pale to nearly white, while 
they are dark brown in F. comaru. The whole plant in F. 
comaru is covered with fine and often adpressed hairs, 
while in F. angustifolia the hairs are extremely sparse, 
if present at all. A further feature of the leaves is that 
their margins are revolute in F. comaru, while in F. 
angustifolia they are mostly not revolute (this feature is 
often distorted by pressing, so that it is not always 
helpful in herbarium material). Finally, if flowers are 
present, the two can often be separated on the extra flap 
in the corona behind each large outer lobe that is 
usually present in F. angustifolia and absent in . 


comuaru. 


Specimens 1 ANGOLA. Huila, Sá da Bandei ira. 
Santos 764 (L ; E of Oncocua, Bruyns 10355 (MO); 5 km 
W of Braga 10363 (E); lona, Bruyns 10384 
WIND); near 1 San, Bruyns 1046 16 (K); N of 
Cahama, Dm 1043. . 

BOTSWANA. Dobe, Lee D64/61 (SRG a 45 km W of 
Nokaneng, Bruyns 6479 (BOL); 128 km WNW of Francis- 
town, Drummond 5283 (SRGH); 81 km 15 of Kuke gate, 


550 Annals of the 
Missouri Botanical Garden 


15 20 25 30 35 40 


Figure 5. Map of central and southern Africa, showing distribution of Fockea angustifolia. 


Smith 3167 (K, PRE): 9 km NE of Ghanzi, Bruyns 6455 NAMIBIA. Otjitande, Rycroft 2460 (NBG): 3 km W 
(BOL): D'Kar, Bruyns 6 160 (BOL): 8 km W of Lethlekane, lengua, Bruyns 5577 (WIND): SW of Nzinzi, de Winter 3985 
Wild a e 7244 (M, PRE); Makunda. 5 144 (PRE. WIND): Okorosave, Owen-Smith 103 dis ID); Kaoko 


(WIND of Lephephe, Snyman & Noailles (PRE) Otavi, Bruyns 5572 | in Barnard. (SAM); Orumana, 
Content p» Ke ho 469 (SRGH): Thorn an T. ; 1 d 63 (M, WIND); 147 km N of m injab, Bruyns 
Bayer 1431 (N 4076 (BOL): Elosha Park, 1 Katspruit, le Roux 

5 E RM 5 iis E of Dakabuko, Luke & Robertson 2547 1099 m B k ur a Horn (WIND); 4 miles SW 
(K); betw E and Mackinnon Road, Drummond & of Okaukeujo, Le Roux (WIND): 22 km E of Okakuejo. 


HU 1045 » (K); Shimba Hills. Luke et al. 6188 (K). Smook 7575 (PRE): Aa 1 9 5 2339 (WIND): Kumkauas, 


Volume 93, Number 4 Bruyns & Klak 554 
2006 Systematic Study of Fockea 


Figure 6. A-D. Fockea angustifolia. EL. Fockea capensis. —A, E. Part of flowering branch. —F. Side view of flower. 
8 ESEN | 8 
—G. Side view of gynostegium. —B. Side view of dissected gynostegium. —C, H. Side view of anthers. —D. J. Pollinarium. 


y: E-I, Bruyns 2555 (NBG). Scales: A = 3 mm; B = 1 mm (at A); C = 0.5 mm (at A; D = 


Drawn from: A-D, Bayer 1431 (NBG) 
0.25 mm (at A); E = 5 mm; F = 2 mm (at E); G = 1 mm; H = 0.5 mm (at G); I = 0.25 mm (at G). 


Giess & Smook 10659 (M, PRE, WIND); Auros, Giess 12576 PRE): near Klein Dobe Camp, Maggs et al. 1079 (WIND); 

WIND): Gobis Water, Bruyns 5490 (WIND); 10 km road to Nyae-Nyae Pan, Giess et al. 11133 (PRE, S, WIND): 
toward Tsumkwe, Bruyns 4116 (BOL); Grootfontein, Schoen- road to Makuri, Hines 910 (PRE, WIND): Gautscha Pan, 
felder S 128 (PRE): Sikereti, Maguire 2267 (BOL. NBG, Maguire 2193 (BOL, NBG, PRE), Story 6216 (PRE, WIND): 


Annals of the 
Missouri Botanical Garden 


Ozondjache, Volk 540 (M); Otjenga, Volk 

Volk 844 (M); Okakarara, 
1 NE of E pata, Giess 9738 (M, WIND): 38 km 
NE of in ies 9751 (WIND) uis 1 3611 
(WIND); Omatako-Sicht, Winston in WI "i 
Immelman 492 (PRE); Omusema, Kers Ms 
Bradfield 349a (PRE); Quic kborn, Pur 349b (PRE): 
Schoongelegen, Seydel 2594 (M, WIND): Sturmve Jd, Tolken 
61 (WIND); Wilhelm: nih Bruyns 2273 (WIND), 5474 (K): 
64 km N of ml Basson 238 (PRE); Epukiro, Giess 9760 
(M, PRE, ND); Okomitundu, Seydel 1419 (M, PRE); 
Ongos, A p (S); Neudam, Gress 3921 (M. WIND): 
Ludwig, Kers 2535 (S); Orumbo-noord, Gress 8403 (M, PRE, 
WIND); Gammams, Wanntorp 91 (PRE); Windhoek, Dinter 
275 J): Hanekom 354 (WIND); Müller n (PRE, WIND), 
Seydel 1704 (WIND); Pehlemann 1243 (WIND); M aa 
Leippert 4355 (M, WIND); Finkenstein, wW cd 185 
Nabitsaus, van Vuuren 585 (M. PRE, WIND); Brack, 10 
11198 (M): Okasewa, Dinter 7450 (SAM); oo 
Merxmüller 1234 (M); 10 km W of Witvlei, Burgoyne et al. 
5221 (WIND); Witvlei, Mason «€ Boshoff 2579 (PRE): 

Kirschberg. Bruyns 5470 (WIND); Dawis, Merxmiller & 
Giess 1200 (M, PRE, WIND); Oanob Dam Nature Res., 


1370 (M); 
Each h 202 


1 


(WIND 


Thorntrees, 


Z. 


2 8 


ai 


Sievers 168 (WIND); Klein Nauas, Fleck 698 (Z); 64 km N of 


YI 
A 
c 
A 


Kalkrand, Acocks 18149 (PRI 
(WIND); Sandverhaar, Bruyns 
Bruyns 4148 (BOL); Kapokvlakte, 
(WIND); cd Volk 12660 (M); 


): Fersbegin, 4 
5462 (WIND); Naukluft, 
Naukluft, Giinsler 9491 
Rooiberg Suid, Bruyns 


5672 (WIND); Barbi, Bruyns 5709 (WIND); Lovedale, 
Bruyns 5727 11 5 Gaibis, Bruyns 5445 (WIND): 
Steinfeld, Bruyns 8143 (WIND); Dassiefontein, Mannheimer 


653 (WIND); Noachebeb, Bruyns 5753 (W Garies, 

Bruyns 5801 (PRE); Bruyns 3511 (BOL); Genadendal, 

da dd 5781 (BOL): Strohbach et al. 3349 (WIND); 30 km 
f Narubis, De Winter 3314 (PRE). 

SOU TH AFRICA. Langkloof, Leistner 2099 (KMG. PRE): 
Knockbarragh, Barkly West, Brueckner 248 (KMG, PRE); 
Andaluzia, Herre (NBG), Mueller-Stoll (M); Padkloof, Acocks 
2207 (PRE); 5 Hay, Cooke 6121 (BOL, KMG); 
auwfontein, Cooke 6420 (KMG); Hay, Acocks 2014 (PRE): 
ux sheim, Bruyns 4493 (BOL): Postmasburg. Esterhuysen 

9 (K); PE Acocks 18792 (PRE); Newlands, 
E 3165 (KMG); er View, Acocks 1503 (KMG, PRE); 
Best Pan, Tapscott e (KMG); Riverton, Tapscott (BOL); 
48 km W of Kimberley, Lewis 67535 (PRE, SAN ue i km W 
of Kimberley, Hall 652 (NBG), Hall 654 (NBG); Waver lev 
Bruyns 9425 (NBG); Asbestos Mtns., Bryant ae (PRE 
i A x Donald 77/85 (PRE); Rode Pan, near 

24 (PRE); Olifantsrug, Power 6682 (BOL. 
K); Driekop, Lie vu van Zyl 1046 (PRE): Magut, Gerstner 
3157 (ES PRE s 1] an tows ard pe Bruyns 4459 (BOL): 
, PRE): 
(Peine (NH): near "Midas "n, Po 4437 (BOL); Be lou. 
Gerstner 3894 (NH, PRE); Umfolozi Game Reserve, Mthonti 
16 (PRE); Mfule Valley, Venter : 5107 (PRE): Hluhluwe Game 
Reserve, Ward 2061 (PRE); M 
1691 (PRI); Alfred, Meo 6606 (PRE); Gravelotte, 
1920 (NBG); Manyeleti Game Reserve, Brede 17 1784 

(PRE); 3 km S of esci Leistner 3190 (K PRE 

Schwerin. Leistner ` 3192 (K, M, Z); 9.5 km SW of 11 h. 
Louw 3545 (PUC); Sterkrivierdam Nature 100 
2809 (PRE); Pire tersrus, Bolus 11014 (BOL, 
of Marble Hall, Vorster & Jackson 2160 (PRE); cp do from 
rm to. Middelburg, Meeuse 9600 (PRE. S) 

Park, Gertenbach 5446 (PRE); 6.5 km WNW of Rustenburg, 
Acocks 19169 (PRE); 1 Mogg 14864 (M, PRE); 
van der Schijf & Marais 5644 (PRE); 


Bangonom: m 


I 
— 
— 


Y 

co 
= 
2$ 


Kruger 


N of Gomandwane, 


Barberton, Rogers 20304 (S); Maquassie, Morris & Engel- 
brecht 6924/3 (PRE). 
SWAZILAND. Sicusha, Bayliss 564 (K, PRE), 2076 (NBG, 
Z); St. 1 0 Hlatikulu, Dlamini (M, PRE). 
PA Winkler 3803 (K); 
;dbaya Hill, Mkomazi Game Res., 
| Kitangiri, 1 35 


20 km W of 


1 0 nl, 


` 
TE 
POE 
um 
S 
e 
bo 
= 
A 


Vollesen 96/22 (Ky 
31 km 5 Bruyns 8723 (MO 
Land, Dodoma distr., Rufo € Magogo 27] (Ky Kiboriani 
Hills, Bruyns 9035 (MO). 


S of Kondoa. 


ZAMBIA. Gwembe, near Ntoboute village, Scudder 
(SRGH): Sinazongwe, Bruyns 9587 (MO). 
ZIMBABWE. Mwenda, Jarman (SRGH); Redlands, Leach 


& Gosden 15014 (SRGH): 8km W. of Salt Springs, 
Sinamatila, Rushworth 2550 one 68 km W of Kadoma, 
Bruyns 7454 (BOL); (K); 
Umkondo Mine, Dale Drummond 
5814 (K, PRE, SRGH). 


a 


Chase 7979 


Ds R.. 


Devure R. bridge, 


208 ce H); 


2. Fockea capensis Endl., in Endl. & Fenzl, Nov. 
Stirp. Dec. 3: 17. 1839. Fockea crispa K. Schum. 
in Engl. & Prantl, Pflanzenfam. 4 (1): 296. 1895. 
Fockea edulis var. capensis (Endl.) G. D. Rowley, 
Asklepios 75: 17. 1998. TYPE: South Africa. 
Cape, Boos & Scholl sub Schönbrunn Garden 488 
(lectotype, designated here, K!; duplicate, W, not 


seen). Figures 6E-I. 


Perennial herb with large underground to partially 
exposed tuber to 80 X 60 cm with central. growth 
point on upper warty surface from which a cluster of 
several more slender perennial and aerial stems arise, 
stems erect to twining to | m with some of more 


slender growth deciduous, young growth shortly 


pubescent. Leaves with minutely puberulous petiole 
1-2 mm elliptic, 15-30 X ©- 


12 mm, finely and densely pubescent, acute, margins 


long. gray-green, 


strongly undulate and erisped but not revolute. 
flowers opening in 
rapid succession, pubis: peduncle 1-2 X 2 mm; 
pedicels 26 Sepals 1.5-2 X 
anceolate, outside pubescent, inside glabrous; corolla 
15-22 mm diam.; 


inside 


Inflorescence l- to 5-flowered, 


1-2 mm. lo mm, 


outside grayish green, puberulous: 
tube 2-3 mm deep, ca. 
8-12 X 


2 mm, linear, obtuse, spreading and often spiralling to 


ereen, puberulous; 


2 mm broad at mouth, cupular; lobes 1.5- 


left, usually twisted to left when viewed along length, 
with margins strongly reflexed; corona white. glabrous; 
outer corona forming cylindrical tube 2.5—3 mm long, 


divided at mouth of s into 5 longer slender terete 


and spreading lobules 2-3 mm long each flanked by 2 
flattened spreading ne , these groups of 3 lobes 


alternating with 5 truncate to emarginate, usually 


spreading lobules; inner corona of 5 terete linear lobes 
adpressed to backs of anthers then erect above them 
and slightly spreading toward tips. Follicles 50-80 X 
10-12 mm, gray-green, often longitudinally wrinkled: 
seeds 8-10 X 4-6 mm, yellow-brown. 


Volume 93, Number 4 
2006 


Bruyns & K 
Systematic cm of Fockea 


553 


d 
SAGA EE LA 
Rb ÉS 


Figure 7. 

Phenology. Flowering October to May. 
Distribution. South Africa; 300-1200 m a ig. 7). 
Fockea capensis is found only in the southern 


portion of South Africa where it grows on Bs Little 
Karoo and in the dry mountains forming its northern 
border. It has been recorded from a little east of 
Prince Albert westward on the Little Karoo to the 
Warmwaterberg (Fig. 7) 

of Fockea 


always found among other succulents on rocky 


Ecology. Specimens capensis are 
slopes (mostly on sandstones and quartzites but 
occasionally on shales as well), where the tubers are 
tightly wedged among the rocks and are often much 
misshapen by pressure from their surroundings. 

Discussion. Endlicher based his description of 
Fockea capensis on the specimen in cultivation at 
Schónbrunn Gardens but did not cite any preserved 
malerial. As only one plant appears to have been 
grown there, this is assumed to be the same as the 
specimen Boos & Scholl sub Schónbrunn Garden 468, 
which, consequently, is selected as the lectotype of F. 
capensis. 

The name Fockea cited with 
"(Jacq.) K. Schum.” as the authors, with Cynanchum 
crispum Jacq. (1800-1809: 31, t. 34) as the basionym 
e.g., Court, 1982; Meve, 2002). However, C. crispum 


Jacq. is illegitimate, being a later homonym of C. 


crispa is usually 


— 


crispum Thunb. (of 1794, which is a synonym of 
Brachystelma thunbergii N. E. Br.). Thus, it cannot be 


the basionym for any later combination. The first valid 
publication of the epithet crispa for a species of 
Fockea was that of K. Schumann in 1895. 
Brachystelma crispum E. Mey. (Meyer, 1838: 196: 
1830) is 
somelimes cited as a synonym of Fockea capensis 


(Meve, 2002). However, Meyer (1838: 196) mentioned 


a later homonym of B. crispum Graham, 


Map of southern Africa, showing distribution of Fockea capensis (triangles) and F. edulis (circles). 


. folia 5—4 lineas 


name applies to a species o 


"tuber depressum ut in Br. tuberoso. . 
longa," so 
cM rather than a species of Fockea. The 
of B. collected by J. F. Drége 
Candoo: near Hamerkuil, Meyer, 1838), has not 
been located, but F. capensis does not occur so far to 
the east in South Africa 


+ 


crispum, 


~ 
< 


The plants of Fockea capensis are not generally 
as those of F. edulis. As in F. 
distinctly warty upper surface of 


as large edulis, 


the 


usually projects from the ground. 


the tuber 


Noteworthy ir 


— 


F. capensis (and in F. edulis) is the frequent presence 
of a fairly large root spreading outward superficially 
from near the top of the tuber, which must be 
important for absorbing moisture close to the surface 
of the soil. In F. capensis the stems are often grazed off 
so as to be erect, but they may occasionally twine on 
The 


edulis are discussed under that 


surrounding bushes, if any are available. 
differences from F. 


species. 
Specimens examined. SOUTH AFRICA.  Skerpkrans, 


Vrede, Bruyns 5314 (BOL); Keurfontein, Bruyns 2809 
(BOL); Klein Speeufontein, Bruyns 4566 (BOL); 5 


tein, Bayer 4675 (NBG); 43 km NW of Ladismith, Bruyr 
2555 (NBC); Buffelsrivier Poort, Bruyns 2554 (NBC): 
Patatsfontein, Bruyns 7518 (BOL); Plathuis, Bruyns 7340 


(NBG); nr V Bayer 4676 (NBG); 39 km SE of 
Laingsburg, Bruyns 2874 (BOL): Amalienstein, sub STE 
14888 (NBG); V 1 6423 (NBG); 
zeespoort, Bruyns 8217 (N); versdale t'off, 
2924 (BOL); 11 km W of Ladismith, Hal 2592 (NBG): 5 
S of Ladismith, Bayer & Bruyns 3648 (NBG); Naauwkloof 
Hilton-Taylor 900 (NBG); 8km S of 
Bayer & Bruyns ; 


— 


at. eserve, 
Vanwyksdorp, I 
; ) ; Tierberg, 
| m (PRE, NBG); Boomplaas, Hugo 21 
Moffett 455 (NBG); Droékloofberge, Bruyns 8188 


(MO); Brakpoort, Hugo 148 (NBG, PRE); N of Robinson 


554 Annals of the 

Missouri Botanical Garden 
Pass, Bayer 458 (NBG); Leeukloof. Bruyns 7085 (BOL): side of South Africa. where il always occurs with 
Boesmanspoortberge, Bruyns 6315 (BOL). plenty of other succulents. Nevertheless, it is also 


= 


Mey.) N. E. Br., Fl. Cap. : 
. Basionym: Brachystelma comaru 


Austr.: 195. 1838. 


3. Fockea comaru (E. 
(1): 781. 1908 
E. Mey., Comm. Pl. Afr. 
TYPE: South Africa. Cape, Nuwerus, N of 
Beaufort West, M. B. Bayer 938 (neolype. 
designated here, NBG! duplicate, PRE). Fig- 
ures BA-G. 


Fockea gracilis R. A. Dyer, Bull. Misc. Inform. 1933: 459. 
1934. TYPE: South Africa. Grahamstown, Dikkop 
‘lats, amos short Karoo bushes, R. A. er 1251 


— 


(holotype, ; isotypes, BOL! , GRA! KD). 


Small geophytic herb with underground napiform 
tuber (sometimes deformed by roe 30 X 
15 


5 em with central growth point on smooth upper 


s) up lo 


surface from which slender perennial and aerial stems 
arise, stems erect to twining to 50 em (~l m), softer 
upper parts deciduous, voung growth finely tomentose. 
Leaves with petiole | mm long or less, green to bluish 
linear, 25-65 X 1.5-2(-4) mm, obtuse, upper 
darker 


pubescent, lower finely pubescent, midrib translucent 


green, 
surface than lower and finely adpressed 


white, margins strongly revolute, weakly undulate to 


Inflorescence l- to 6-flowered, flowers 
opening in succession, covered with fine adpressed 
hairs; peduncle absent; pedicels 1-2 X 1 mm. Sepals 
1.5-2 X | mm, lanceolate, outside finely pubescent, 


8—27 mm 


grayish green, finely pubescent; inside gray-green to 


inside glabrous; corolla diam.; outside 


brown-green, sparsely and very finely pubescent; tube 


5-3 mm deep, ca. 2 mm broad at mouth, cupular: 
lobes 4—12 X 1—2 mm, linear, obtuse, spreading and 


often spiralling to left, often twisted to left when 


viewed along length, with margins strongly reflexed; 
glabrous: outer corona forming eylindri- 


cal tube 4—6 mm long, divided at mouth of tube into 5 


corona white, 


longer slender terete and strongly. spreading lobules 
flattened spreading 


2-2.5 mm long each flanked by 2 g 
5 


) lobes alternating with 
lob- 


ules; inner corona of 5 terete linear lobes adpressed to 


lobules, these groups of : 


linear to b usually spreading 


hem. 


backs of anthers then rising erect above 
Follicles 50-100 X 10-12 mm, gray-green, smooth: 
seeds 8-10 X 


4—6 mm, yel low -brow n. 


Phenology. Flowering October to May. 
Distribution. Namibia. South Africa; 100-2000 m 


(Fig. 11). 
Ecology. Fockea comaru is always found in stony 
Du J R 
places on hillsides and mountains, with the tubers 
often tightly wedged between rocks at a depth of up to 
gi B | | 
20 em or more. It is the only species of Fockea that is 


common in the winter rainfall region of the western 


found well beyond this. Many of the known localities 
outside the winter rainfall region are in elevated areas 
such as the Tiras Mountains and the Great and Little 
Namibia) that 
rainfall, but those in Griqualand 


Karas Mountains of receive small 
amounts of winter 
West (for example) do not receive any significant 
winter rainfall. 

The specimen cited by Meyer (1838: 
Uitvluet, 


J. F. Drege) is missing. 


Discussion. 
195, namely South Africa, Richmond distr., 
near Steelkloof, 4000-5000“, 
and no other suitable specimen of Drege has been 
Very 


collections exist from the central part of South Africa 


located that could serve as a lectotype. few 


where Drége’s specimen came from, and other early 


collections are neither from this area nor are 


suitable as 


Bayer 


— 


similar to be neotypes. 


938, is 


sufficiently 


Consequently a recent collection, 
selected as neotype. 

Although the stems in Fockea comaru may climb in 
surrounding bushes, they are often short and erect and 
frequently do not twine at all, forming instead small 
rosettes on the ground. The stems have a rhizomatous 
habit and often spread underground for some distance 
from the tuber before emerging from the soil. These 
subterranean stems appear to be perennial. In such 
situations, stems from the same plant may appear over 
a considerable distance, giving the appearance of 
several specimens growing near each other, though, on 
excavation, they prove to originate from the same 
tuber. 


As shown in Fig. 8A=G, the flowers are exception- 


ally variable in diameter and the lobes may be quite 


short and not twisted at all. 

Apart from the highly sueculent stapeliads, ascle- 
piads are not generally common in the winter rainfall 
parts of southern Africa. In this respect Fockea 
comaru (along with a few species of Microloma) is 
and 1 


unusual, is of very wide and quite frequent 


occurrence in these areas. The very narrow, linear 
leaves and the milkv sap make the plants of F. comaru 
unmistakable among the asclepiads in most areas 
where they occur. His often well known where it 
occurs, with the tubers being dug out and consumed, 
and the plants are known as kambroo or bergkambroo. 

The molecular data (e.g., Fig. 3) show that Fockea 
comaru and F. sinuata are very closely related and 
the 


investigated. Morphological differences between the 


scarcely differ at all in three gene regions 


two are discussed under F. sinuata. 


In our molecular sampling, we included an 


accession of the former Fockea gracilis (Dold 2381). 


and this did not group strongly with the other 
accession. of F. comaru. F. gracilis was mainly 


Volume 93, Number 4 Bruyns & Klak 
2006 Systematic Study of Fockea 


555 


Figure 8. A-G. Fockea comaru. H-M. Fockea edulis. NS. Fockea multiflora. —A, H. N. Portion of stem. —I, O. Bud. 
—B. Face view of flower. —C. D, P. Side view of flower. —E. J. O. Side view of gynostegium. —F, K, R. Side view of 
dissected gynostegium. —L. Side view of anthers. —G, M, S. Pollinarium. Drawn from: A-C, E. F, Bruyns 1700a (NBG); D, E, 
Bruyns 3129 (BOL): G, Bruyns 3107 (NBG); H-M, Bayer 325 (NBG); N-S, D. Laing (K). Scales: A = 2 mm; B-D = 3 mm (at 
C); E = 1 mm; F = I mm (at C); G = 0.25 mm (at C); H, I = 3 mm (at H); J, K = 1 mm (at H); L = 0.5 mm (at H); M = 
0.25 mm (at H); N = 3 mm; O, P = 2 mm (at N); Q, R = 1 mm (at N); S = 0.25 mm (at N). 


— 


556 


Annals of the 
Missouri Botanical Garden 


distinguished. from F. comaru by its much smaller 
flowers. However, the wide variation in this feature 
within F. 
recognize reliably, so that we are unable to distinguish 
it from F. Molecular 


distinctness from F. comaru is also negligible (Fig. 3). 


comaru makes F. gracilis impossible to 


comaru. support for its 


Specimens examined. NAMIBIA. Landsberg, Bruyns 
8112 (WIND); Naus, Bruyns 5698 (WIND); Lovedale, Bruyns 
5724 (WIND); Tschaukaib, Bruyns 3170 (NBG); Nooitge- 


. Bruyns 8348a (WIND); Nasapberg, Bruyns 7206 

: Sebrafontein, 1 3924 (BOL); Steinfeld. Bruyns 
8142 D. Lord Hill, Bruyns 8147 (WIND); Witmond, 
Bruyns 5759 (WIND): 1 1 5 et al. 3800 (WIND); 
Bruyns 5805 (PRE) Bru 3512 (WIND); Genadendal, 
Ec 11 8 DL 

AFRIC ^. Pear Bruyns 6278 (K); Van Der 

1 Bruyns 7281 (NBG); Mtn., 
Bruyns 3291 (NBG); rg, Bruyns 3304 
NBG): 6 km S of Eksteenfontein, Bruyns 8286 (MO); 
AA Herre STE 12334 (NBG); Rudesheim, Bruyns 
4493 (BOL); Ste inkopf. n 13300 7 RE); Hester Malan 
Reserve, Rösch & le Roux 729 (PRE); Gannapoort, Leistner 
(KK, KMG, PRE); 11 70 rley, Bruyns 9424 (MO): 
Stndenburg Hall 640 (NBG); 6 km N of Ou Dam, Bruyns 
6753 (BOL): 4 km W of Paulshoek, Bruyns 6107 (PRE); N of 
Kaams, PUE 6348 (BOL); Stofkraal, Bruyns 4767 (BOL): 
Witvlakte Oos, Bruyns 6089 (PRE): Kubiskow, Bruyns 6085 
(PRE); Loeriesfontein, Marloth 12848 (PRE); Nooiensberg, 
Bruyns 6645 (BOL); NW of Komkans, Bruyns 6255 (BOL 
Nuwerus, Bruyns 6786 (BOL); Kalkgat 5 Bruyns 6018 
) 16 km S of Klawer, Hall sub NBG 817/18 (NBG): 
Bruyns 6824 (BOL): 1 8 
i Calvinia, Marloth 7147b (PRE); Klip rug, Bruyns 
6056 (PRE); Botterkloof, Bruyns 6038 (PRE), Hall 700 
(NBG); Aarkolk, Bruyns 5980 (BOL); Grootfontein, Coetzer 
59 (PRE); Boe 1 rg, Juriesfontein, Bruyns 6688 (BOL): 
Nuwerus, Bayer 938 (NBG, PRE); Pakhuis, Hall sub NBG 45/ 
50 (NBG); € pida Dam, Bruyns 6034 (PRE); 11 km SE 
of Redelinghuis, Bruyns 4734 (BOL); 6 km N of Het Kruis, 
Bruyns 6175 (BOL); E of L 5 Bruyns 7502 NOGI 
Bossiesberg, Bruyns 7 7572 (BOL); 2 km N of Uitsig, Bruyns 
6708 (BOL); Aarfontein, Bruyns 6361 mu jh Klipfontein, 
Bruyns 3107 (NBG); Amandelboom, Bruyns 6695 (BOL): 
Doornhoek, Bruyns 399] (BOL); Prutkraal, Bruyns 3141 
(BOL), Snyman (NBG); Worcester Veld Reserve, Olivier 255 
(PRE); 19 km SE of Worcester, Dayer 319 (NBG); 4 km W of 
Robertson, Schwegmann (NBG); 2 km SW of 
Bayer 136 (NBG); Doornrivier, St sakes (PRE): Avondrust, 
Bayer 318 (NBG); Whitehill, Archer 18286 (BOL): Ou Tol. 
Bruyns 6304 (BOL); Plathuis, Bruyns 7341 (NBG): Coetzee- 
spoort, Bruyns 8216 (NBG); Calitzdorp, Oddie 559/36 (BOL); 
Albert, Marloth 7147a (PRE); Haggas, Vlok 2440 
Vie sikuil, Bruyns 6309 (BOL); Skietkuil, Bruyns 3129 
M De Hall 1645 (NBG); near Steytlerville, 
0 155 8 (NBG); Korhaan Vlakte, Addo, Dold 2381 
(GRA, PR 1 99 Flats, Dyer 1251 (BOL, GRA, PRE); 
4 km NW ij lena Pass, Bayer 329 (NBG 


~ 


Garies, 


summit Rosyntjie 
summit Cornellsb 


— 


— 


Robertson, 


Prince 
(NBG); 
(BOL) 


4. Fockea edulis (Thunb.) K. Schum., Bot. Jahrb. 
Syst. 17: 146. 1893. Basionym: Pergularia edulis 
Thunb., Prodr. Pl. Cap. 1:55. 1 
edulis (Thunb.) Thunb., 


"chites 


In genus Echitis observa- 


tiones; 5. 1819, TYPE: South Africa. 
Gouritz River, C. P. Thunberg 6141 (holotype. 
UPS, not seen). Figure 8H-M. 


Cape 


Brachystelma macrorrhizum E. 1 Comm. Pl. Afr. Austr.: 
197. 1838. TYPE: South Africa, Cape, Graaff-Reinet, 
Marloth 7048 (neotype, 7 Suae here, PRE). 

in DC. rodr. 8: 545. 1844. TYPE: 

1 1 239 (holotype, PL 


ds. e Decne. 
uth Africa. elt 

isolypes, C, not seen, W, not se 
Fockea a R. ri Dyer, Bull. d 
1934. TYPE: South Africa. Cape, 
amongst karroic 


K!; isotypes, GRAL, 


Tai, 1933: 459 
Fish River Valley, 
scrub on dry flats, R. 


PRE). 


lom 


near r ommitte 


A. Dyer 1635 holotype. 


Perennial herb with large underground to partially 
exposed tuber up to 50 em long and 1 m broad with 
central growth point on upper usually warty surface 
from which 1 to several more slender perennial and 
aerial stems arise, stems erect to twining to 2 m with 
some of more slender growth deciduous, young growth 
puberulous. Leaves with minutely puberulous petiole 

2-8 mm long, green often with paler midrib, elliptic. 
15-45 X 6-20 mm, glabrous to sparsely pubescent, 
often undulate but not revolute. 


acute, margins 


Inflorescence l- to 8-flowered, flowers opening in 
succession, puberulous; peduncle 1-3 X l mm; 
pedicels 1-3 X Sepals 1.5-2.5 X 
lanceolate, outside pubescent, inside glabrous; corolla 


10—22 mm 


puberulous; inside green, glabrous to puberulous; tube 


] mm. ] mm, 


diam.; outside grayish green, sparsely 


2-5 mm deep, ca. 3 mm broad at mouth, cupular; 
lobes 6—12 X 1.5-2 mm, linear, obtuse, spreading and 
often spiralling to left, usually twisted to left when 
viewed along length, with margins strongly reflexed: 
corona white, glabrous; outer corona forming cylindri- 
cal tube 


longer 


3—4 mm long, divided at mouth of tube into 5 


slender terete and ascending to strongly 
spreading lobules 2—3 mm long each flanked by 2 
flattened spreading lobules, these groups of 3 lobes 
alternating. with 5 linear to truncate to emarginate 
usually spreading lobules; inner corona of 5 terete 
linear lobes adpressed to backs of anthers then erect 


above them and slightly spreading toward tips. 


Follicles 50-80 X 10-12 mm, gray-green, often 
longitudinally wrinkled; seeds 8-10 X 4-6 mm, 


yellow-brown. 


Phenology. 


Distribution. 


Flowering October to May. 

South Africa: 10-1200 m (Fig. 7). 
Ecology. | Fockea edulis is found in dry bush in the 
southern portion of South Africa, from the Worcester- 
and the coastal 
East 


similar habitats in southern Natal around the Oribi 


Robertson Karoo 


Bredasdorp to around London also in 


~ 


zorge. Plants occur among rocks and bushes or trees 


on slopes. It is rather more rarely found in flat areas, 


Volume 93, Number 4 Bruyns & Klak 557 
2006 Systematic Study of Fockea 
as in the low-lying areas between Steytlerville and Specimens examined. SOUTH AFRICA. SW side of 
Port Elizabeth. Graaff-Reinet, Burchell 2916 (K); 1 s Bruyns 
3 ] c i E 2976 (BOL); Gr -Reinet, Marloth 7048 (PRE); Adelaide 
Discussion. Se 893: 1 ; À : : 

Tope E Aa s kimanin d : n) and oad, J.-Guillarmod 4929 (PRE); Keiskamma, | pues 323 
Schumann (1895: 296), “Fockea edulis (Thbg.) NBG); Gonubie uth, Bayer 326 (NBG); Klaasvoogds, 
Sch.” is mentioned, and no attempt is made to Bayer 325 (NBG); Rooiberg, Goudmyn, Bruyns 6793 (BOL); 
describe this as a new species. Consequently in Klipplaat, Bayer (NBG); Witkop, 4 Desmet 2240 (NBG); 
Schumann (1893: 146) a valid new combination was leinspoort, Baye Veni van Jaarsveld 

11410 (NBG); Addo Elephant Park. Boa 5800 (PRE); near 
made, because he made an indirect 


reference to 
a previously AP name (Greuter et al., 2000: 
ICBN Art. 32.4, . 32.5) and the basionym for t 
name is „ zin Thunb. as given in Brown 
(1907-1909: 7 


For Baz itelma macrorrhizum, Meyer (1838: 197) 


— 


ais 


T 


cited an unnumbered collection of Drege (South 
Africa, Oudeberg mountain), which is missing. The 
neotype selected here is from the same area as Drege's 
that 
a synonym of Fockea edulis. 

As mentioned above, Harvey (1842) made no new 
combinations in he 


this 


collection, so this name is unequivocally 


Chymocormus even though 
mentioned Pergularia edulis as belonging 10 
new edulis 


1907—1909; 
2000: 


genus. Consequently C. 
1895; 


is not valid (Greuter et al., 


Schumann, Brown, 


2002), 


33.1) and is not cited here. 


ICBN Art. 


In more bushy areas Fockea edulis is a vigorous 
climber to a height of 2 m. In more arid spots, where 
the vegetation is altogether shorter (such as on the 
Worcester-Robertson Karoo), F. edulis can be a smal 
and inconspicuous climber among bushlets of Carissa 
bispinosa (L.) Desf. ex Brenan and other species. Even 
smaller, tufted plants that scarcely twine at all are 
succulent 
vegetation between Steytlerville and Port Elizabeth, 
and one of these was described as F. cylindrica [= F. 
edulis]. 


characteristic in areas of low, mostly 


In Fockea edulis the distinctly warty upper surface 
of the tuber usually projects slightly from the ground. 
The tuber may be enormous and has been recorded up 
to | m in diameter (Bayer 326, NBG). 

Fockea edulis and F. capensis are similar in many 
respects. Both have comparatively broad leaves (6— 
20 mm broad in comparison to 1.5—4 mm broad in F. 
comaru) and large with upper 


projecting from or level with the surface of the ground. 


tubers the surface 


Fockea capensis is short and shrubby in habitat, but 
the stems will twine if provided with the opportunity. 
In F. capensis the leaves have distinctly undulate or 
crisped and are 


margins with 


adpressed hairs, while in F. edulis the leaves have 


densely covered 
at most slightly undulate margins and are more or less 
glabrous. This lends them a green color in F. edulis 
and a distinctly gray-green color in F. capensis. 
Florally there are no significant differences between 


them. This close relationship is reflected in Figure 3. 


Groendal, 
Steytlerville, 


Vlok 554 (PRE); 36 miles from Port Elizabeth to 
Long 1184 (K); Springs Reserve, 


Uitenhage, 


Olivier 2495 (PEU); Uitenhage, Zeyher 965 (K); near 
Uitenhage, Burchell 4450, 4443, 4457 (K); betw. Port 
Elizabeth and Uitenhage, Long 1469 (K); Addo, Acocks 


(PRE), Kreft 96 (B OL). Coega, Bayer 320 (NBG), 
Bursey INES) Olivier 1577 (PRE); Bluewater Bay, Olivier 

2 ; Swartkops, Hollack 3052 (PRE); Humewood, 
Danan 60 (PRE); Redhouse, 544 (BOL, Z): 

oes stn., Long 903 (K); Galpin 13250 (BOL, 
PRE). 13251 (P, PRE); Hyde’s Hill, Nicholas 912 
(PRE); betw. Tootabi and Alicedale, Archibald 5984 (PRE); 
10 miles from Grahamstown toward Alexandria, Dyer 956a 
(K); Pass, Bayer (NBG); Pluto's Vale, Bayer 322 (NBG); 
Trump s Drift, Story 2190 (PRE); 5 5 toad, Jocot 
1 4929 (PRE); Nook Boosaak Forest, Alex 
Britten 2506 (PRE); N of Stormsvlei Pass, Bayer (N 
Hoop, Bruyns 6656 (BOL : Kafferkuils R. mouth, Muir 149 
(PRE): Klein Brakrivier extension 4, Taylor 8341 (NBG): 
Gibraltar rock, Nicholson € Strey 1968 (PRE), Balkwill 439 
(NU). 


Paterson 
Hounslow, 
near 


BG ); De 


5. Fockea multiflora K. Schum., Bot. Jahrb. Syst. 
17: 145. 1893. TYPE: Tanzania. French mission 
in Ussambiro, F. L. Stuhlmann 848 (holotype, B, 
missing; isotype, K!). Figure 8N—-S. 


Fockea sonia N. E. Br., Bull. Misc. pn 1895: 259. 
1895. TYPE: Angola. F. bine 4194 (lectotype, 


de al here, K!; duplicate, 


— 
— 


Large climber to 15 m, with stout trunk (to 30 cm 
thick) sprawling on ground or twisting around trees for 
support, rarely shrub-like; stems fleshy and becoming 
swollen toward base but without distinct basal tuber, 
young stems tomentose and slightly fleshy, remaining 
fleshy but becoming covered with gray, shiny bar 
Leaves with petiole 8-25(—45) mm long, gray-green, 
(20-)80-150 X (10-)25— 


100 mm, often apiculate, upper surface tomentose to 


oblong to broadly elliptie, 


glabrous, lower tomentose with raised midrib and 
veins, margins not revolute or undulate. /nflorescence 
(6- to)10- to 30-flowered, flowers opening + simul- 
taneously or in rapid succession, pubescent; peduncle 
5-15 X 2-5 mm; pedicels 5-13 mm long. Sepals 2-3 
X ca. 


glabrous; corolla 10-15 mm diam 


I mm, lanceolate, outside pubescent, inside 
; outside grayish 
green, pubescent; inside yellowish green to brown, 
glabrous to sparsely pubescent; tube 1.5-2 mm deep, 
ca. 3 mm broad at mouth, shallowly cupular; lobes 5— 
10 X 2 mm, broadly linear, obtuse, spreading, with 
margins and apex slightly reflexed; 


corona white, 


Annals 
3 ae Garden 


glabrous; outer. corona forming tube 2-3 mm long, 


becoming distinctly narrower above anthers, divided 
at mouth of tube into 5 longer slender terete spreading 
each flanked by 2 flattened 


lobules 2-2.5 mm long 


slightly spreading lobules, these groups of 3 lobes 
alternating with 5 linear recurved lobules; inner 


corona of 5 flattened linear lobes adpressed to backs 
of anthers then intertwined above them. Follicles 100— 


220 X 15-30 mm: seeds 10 X 7-8 mm. 


Phenology. Flowering August to December. 
peo le o je) 


Distribution. Angola, Botswana, Moçambique, 

Namibia, Tanzania, Zambia, Zimbabwe: 600-1000 m 

(Fig. 9). 
Ecology. Fockea multiflora occurs in rocky areas 


on low hills or among rocks around the base of hills i in 


open, often deciduous woodland consisting an 
Acacia-Commiphora-Balanites association or Colopho- 
spermum mopane (Kirk ex Benth.) Kirk ex J. Léon. 
Plants are usually common. 

at B on which K. 
Schumann based Pus multiflora was destroyed 
during World War of the 
collection is mide at Kew and this is selected as 
(1895) cited 
1194 and an unnumbered collection o 


Welwitsch 


Discussion. The specimen 


However, some same 


an isotype. For fF. 
Welwitsch 


Schinz, 


schinzii, Brown 


which has not been located. The 
collection is selected as lectotype. 
Fockea multiflora is by far the largest species, 


and massive specimens are probably the largest 


known asclepiads. The big fleshy stems sprawl on 
the ground or twist up the trunks of surrounding trees 
as massive lianes and appear to strangle them, though 
there is no evidence that they have any detrimental 
effect on these trees. Damaged specimens sometimes 
form a huge mass of entangled stems without any 
longer climbing stems, but this is exceptional. In all 
the other species of Fockea, the stems arise from 
a tuber that is much broader than the main stem. In H. 
multiflora, although the young plants have a thickened 


base, that in 


this generally disappears with age, s 


mature specimens there is no basal tuber and the 


rootstock consists of a network of fleshy roots 


radiating from the base of the stem. Only the youngest 


shoots and the leaves are deciduous, while the 


remainder of the plant is perennial. The leaves are 
also much larger (to 15 X 10 em) than those of any 
other species of Fockea. The large size extends into 
the fruit and the seeds. 


n that the inflo- 


Fockea multiflora is unusual 
rescences arise in large numbers around the end of the 


dry season between August and October (more rarely 


December), on young growth produced before the 


true growing season begins. The sweetly scented 


flowers appear on the first young shoots of the new 


erowing season among small, under-developed leaves 
and are sometimes used as a vegetable (B. L. Burtt 


1520 (K): 


Tanzania). Several features of the flower are unique 


Bruyns & Klak, pers. obs., Gulwe district, 


They are produced dense, more or less simulta- 


neously opening clusters. The corolla lobes are 


comparatively short and relatively broad (unlike the 


other species where they are mostly slender and 


linear) and are not twisted along their length (Fig. 8N, 
noticeably 


Both the 


compare ative ly broad anthe HY 


P) The corona tube is often also 


constricted toward the mouth (Fig. 8P, Q). 
and the 


corona tube 


Ww 
— 


(Fig. 8R). 


appendages are also relatively short 


pecimens examined. ANGOLA, Bibala, Gossweiler 131. 
(L Sch. 4 Cambos, Borges 308 , PRE, SRGH). Pearson 
2131 (BOL): Humbe. Menez zes & Sousa 3 123 (LISC. PRE): 
Mupa, Menezes 3152 (LISC, PRE, SRGH); Nangongo, Santos 


& Barrosa 2825 (LISC, P RE) ; Ondjiva, Menezes 1017 (LISC). 
BOTSWANA. Ngoma on Chobe R.. Miller 1050 (PRE). 
Story 4809 (P, PRE). Munro sub PRE 6924 (PRE): Okavango, 
near Nokaneng, Tinley sub PRE 51262 (PRE). 
MOÇAMBIQUE. About 17 km 10 0 Má igoè Lowa 
Velho, Torre & Correia 16156 (LISC): 13 
border in Mazoe R. 


rd Mágoe 
m from Zimbabwe 
area, Wild 2584 (SRGH): about 6 km 
from Changara, Torre & Correia 18709 L ISC 
NAMIBIA 219 (WIND): 
nda, Rautanen (Z); Onda 
(PRE); Olukonda, 
Pienaar 20la ( )puwo, Kolberg et al. 
771 (WIND): Kaoko Otavi. a 128 (WIND); 7 km S 
< Wiss 3 37 AA PRE. 


Giess 0 


Omu— 


of Okaruwizu, 
Mission Stn.. de Wi 
Falls. Maggs 032 M ID 
Tsumeb, Ndgelsbach 18D Pur E 
* oe 10 km N of ou 


quem 15309 (WIND); 
Walter 563 
. Jaarsveld . 3093 (NBG): 7 km 
, WIND); Ombanje, Schinz 5 

n Tek Mın.. Hilbert 31. (WIND); 
Strijdfontein, Dinter 684 (SAM); Grootfontein, Schoenfelder 
205 (PRE): 1). 


TANZANIA. Kakesyo, Greenway 9050 (EA, PRE); S 
oa 9057 (EA. PRE end of Lake 
11626 (K): Lake Kitangiri, [Tm 13490 (K); 
HU 203 (K); 


of Kakesyo, 
Evassi, Bally 
Singida to Itigi, 
Burtt 1454 (K): 
S Johannson 1140 


12 miles from 
' of Kondoa, 


$ Uie ne ngule-usangu, Leedal 
; Lake lo 
Tanner 1085 
. Verdcourt 4014A (K): 
E Sing 1520 (K): Mtera Dam, 
1 5 9613 (P RE i Sof n Bruyns 9646 (NBG): S of 
ulwe, Bruyns 9651 (E). 

3A. Luangwa Valley, A 
1877 (EA): Bombwe Forest Office, 

Bombwe, Martin 318 (K); Nkoata Isl., 


Bally 10017 (EA): Mo. HN Mwanza, 
(K): E of Lake E 


N 
2 5 


, 5149 (SRGH), Trapnell 
Bombwe 318/32 (EA): 

Mkupa, Bullock 1186 
0/55 


(K): Gwembe Valley, Mazubuka distr., Bainbridge | 
(K): Sinazongwe. Bruyns 9586 (E); Monza, Bruyns 9595 
(MO). 

ZIMBABWE. Chirundu, Wild 4852 (PRE, SRGH); just 5 


of Zambesi R. on Mazabuka-Harare rd.. 30784 
PR ESE 


Levy sub PRE 


E): Mei 7 8 km of Chirundu Bridge, 
Da 5363 (PRE, SRGH); Copper Queen purchase 


area, Bingham 175 (SRGED: 35 km N of Gokwe, Goldsmith 


— 


Volume 93, Number 4 Bruyns & Klak 559 
2006 Systematic Study of Fockea 


15 20 25 30 35 40 


Figure 9. Map of central and southern Africa. showing distribution of Fockea multiflora. 


21097 (SRGH); Wankie, Levy 1123 (SRGH); betw. Sesami R. Fockea undulata N. E. Br., Bull. Misc. Inform. 1895: 260. 
and Gokwe. West 2998 (SRGH). 1895. TYPE: South Africa. Cape “Transvaal”, Rhenos- 
terkop, J. Burke (holotype, K!). 


6. Fockea sinuata (E. Mey.) Druce, Bot. Soc. Exch. Small geophytic herb with underground napiform 
Club Brit. Isles: 623. 1917. Basionym: Brachys- luber to 30 X 10 em with central growth point on 
telma sinuatum E. Mey., Comm. Pl. Afr. Austr: upper surface from which slender perennial and aerial 
196. 1838. TYPE: South Africa. Cape, Brakval- stems arise, stems erect to weakly twining to 45 cm, 


lei, betw. Kat River & Swart River, 3500', 15 softer upper parts deciduous, young growth tomentose. 
Jan. 1827, J. F. Drege 3439B (lectotype, Leaves sessile, brown- to gray-green, linear, 25-65 X 
designated here, P!; duplicate, K!). Figure 10. 2—4 mm, obtuse, upper surface darker and adpressed- 


560 


Annals of the 
Missouri Botanical Garden 


Figure LO. Fockea sinuata. —A. pi of flowering Mr h. 
ws view of dissected gynostegium. — 

V, Bruyns 1269 (BOL); C 
(at E); H = 0.25 mm (at E). 


Side view 


pubescent, lower surface pubescent mainly on midrib, 
margins strongly undulate and revolute. /nflorescence 


l- to 10-flowered, flowers opening in succession, 


covered with fine adpressed hairs; peduncle ca. 
| mm; pedicels 1—3 X Sepals 2 X Umm. 
lanceolate, outside pubescent, inside glabrous; corolla 
8-12 mm 


pubescent; inside green to brown, sparsely eie ent; 


| mm. 


diam.; outside grayish green, sparsely 


.5 mm deep, ca. 3 mm broad at mouth, 
cupular; lobes 3—4.5 X E 


spreading, often twisted to left when viewed along 


tube 1.5-2 


AN obtuse, 


length, with margins strongly reflexed: corona white, 
glabrous; outer. corona forming tube 4-6 mm long. 


becoming slightly narrower above anthers, divided at 


mouth of tube into 5 longer slender terete (often 
wrinkled) and strongly spreading lobules 2-2.5 mm 


long each flanked by 2 flattened erect lobules, these 


anthers. 


, E, G, Bayer 930 (NBG). e A = 3 mm: B-D = 2 mm (at A); E, 


us view of flowe 


Y. D, H, Bruyns oni 


—B. Face view of flower. —C, 
II. Pollinarium. Drawn from: i 
k= ] mm (at E); ( 


= 0.5 mm 


groups of 3 lobes alternating with 5 linear to truncate- 


emarginate usually erect lobules; inner corona of 5 


inear lobes adpressed to backs of anthers then 


lerete 
rising in erect column above them. Follicles 50-80 X 
seeds 8-10 X 4— 


6 mm, yellow-brown, with hairs around margin. 


10-12 mm, gray-green, often warty: 


Phenology. Flowering October to May. 

Distribution. Namibia, South Africa: 500-1200 m 
(Fig. 11) 

Fockea sinuata has a very scattered (and almost 


certainly not well-documented) distribution over the 
karroid parts of southern Africa from near Maltahóhe 
in Namibia to the Free State in South Africa and the 
southern portions of the Eastern Cape. 

Ecology. Fockea sinuata is always found in open, 
flat, often pan-like areas and usually in loamy, slightly 


Volume 93, Number 4 
2006 


Bruyns & Klak 
Systematic Study of Fockea 


Figure 11. 


sinuata (triangles). 


calcareous ground. Specimens occasionally grow 
among shrubs of Rhigozum trichotomum Burch. but 
more usually occur with various species of Salsola L. 
and Zygophyllum L. in areas of quite low diversity. 

(1838: 190) 


specimens collected by Drége, one from “Dwyka” 


Discussion. | Meyer cited two 
and the other from “Brakvalei,” both without numbers. 
Only one, Drége 3439B, from Brakvallei has been 
located. Here we assume this to have been seen by 
Meyer, and it is selected as a lectotype. 

In Fockea sinuata the plant mostly consists of only 
6-10 em tall, but 


occasionally the stems may twine on the lower 


a very small erect stem very 


branches of a nearby shrub. The narrow leaves have 
strongly undulate margins and are usually brown- 


green. Both of these features serve to make them 


Map of the southwestern part of southern Africa, showing distribution of Fockea comaru (circles) and Fockea 


exceedingly inconspicuous and lend them a strong 
resemblance to the stems of some of the other plants 
with which they grow. The leaves and some of the 
above-ground parts die of? during the dry season. The 
pendulous follicles dangle down onto the ground and, 
in contrast to the rest of the plant, are comparatively 
conspicuous. In fact, the follicle may often be all that 
one sees of the plant, especially if the leaves are small 
and are not well developed. 

With its mostly erect and non-twining growth and 
narrow leaves, Fockea sinuata somewhat resembles F 
comaru. The two species never grow together, though, 
for F. comaru does not occur in low-lying places in the 
areas where F. sinuata occurs. The leaves of F. 
sinuata always have strongly undulate margins, and 


this is usually enough to distinguish them. Florally 


Annals of the 
3 Botanical Garden 


they are mostly easily recognizable. In F. sinuata the 


corolla is always small, with short and twisted lobes 


and a comparatively large and conspicuous. white 


corona. ln some flowers the corona lobes have 


a pustulate inner surface, but that has not proved to 


be constant (Fig. 10). The pollinarium is also 


distinctive, with fairly swollen pollinia and a larger 


darker 


pollinium is always rather flat in F. comaru, and the 


and corpuscle than in the others. The 


corpuscle Is also slender 
Figs. 8G and 10H 


Perhaps one of the most unusual features of Fockea 


particularly (compare 


sinuata is the fact that each seed has a row of hairs all 
1976). 
phenomenon that is otherwise only known in Raphio- 
Venter 
The seedling of V. 


the way around the margin (Bayer, This is a rare 


nacme namibiana & Verhoeven 
1994). 


its comparatively long and fleshy hypocotyl, which has 


(Bruyns, 
sinuata is also unique for 
always been observed to have a distinctly warty 
surface around the middle (Fig. 4A). 

Over its extremely wide distribution within South 
Africa, this species is generally well known where it 
occurs. The tubers are often excavated and eaten. and 


it is locally usually known as vlak-kambroo. 


Specimens examined. NAMIBIA. 6 km NE 
Bruyns 5667 (BOL); Rooiberg Suid, Bruyns 5670 (MO, 
WIND); Happy Days, Mostert (WIND): 62 km S of 1 
inghause in, Mueller 58 (PRU); Kubub, Gress et al. 5300 (M, 
Vane imer 800 


of Nuwerus, 


Aus and Rosh P p 


„ WIND): betw. 
(WIND): Sebrafontein, Bruyns 5914. (BOL); Grabwasser 
Bruyns 3041 (BOL, WIND): Vrede, o 1555 (WIND). 
SOUTH AFRICA. , Acocks 142604 


(BOL) 
1 
(NBG); Bulberg, 
Bruyns 6396 (BOL 
BG): Boe ak re, 


Kenhardt. 


VN of Prision 
Britstown commonage, Bruyns 302. 
Bulthoue rsfontein, 16 km N of T 
Sisters, Bru 5106 (BOL): Se hoppe Imaaikraal, 
1269 (BOI ) Re nosterkop, Burke (K); Prutkraal, 


Bayer 328 


BG). Snyman ( 165 Rooidam, Bruyns 2950 (BOL) 
Rietbron, Bri (NBC): Tierberg, Bruyns 3095 
(BOL), Marloth, p Du ): Willowmore., pen sub STE 
19941 (NBG); Palmietfontein, Smith 5344 RE): Rose- 


maryn, Henrict 3659 (PRE) 


Literature Cited 
dem V. & B. Bremer. 1997. Phylogeny of the 
ins d on Te se dn nees, pax pud s ar reference 
lo ae ur ales. Pl. Syst. Evol. 207: 2 254. 
Bakker. D. He 11 E Calle & ve Gibby. 1998. 


ee netice relationships within Pelargonium sect. 


Asteridae 


Peristera (Geraniaceae) inferred from nrDNA and epDNA 


sequence comparisons. Pl. Syst. Evol. 211: 273-287. 
Baldwin, B. G. 1992. Phylogenetic utility of the internal 
transcribed spacers of nuclear ribosomal DNA in plants: 


Molec. 


An example from the 
Evol. 1: 3-16. 


Compositae. br 


Jacquin, N. 18 


Jürgens, N. 1991. 


Bayer, M. B. 1976. 


Notes on species of the genera Fockea 
and Kahne Excelsa 6: 87-91 
K. 


Bremer, 1992. Ancestral areas: A cladistic reinterpreta- 

E on of the center of origin concept. Syst. Bot. dH: 
—445. 

Ma n, N. E. 1895. Diagnoses Africanae VIII. Bull. Misc. 
€ 19 247-265. 

— 1003. Asclepiadeae. /n W. T. Thistleton-Dyer 
"rs (s Tropical Africa 40): fou 503. Lovell 
Reeve ondon 

. 1907 7-1909. Ascle Um ae. n W. T. Thistleton- ción T 
(editor), Flora Capensis 4(1): 518- a Lovell Reey 


London. 
Bruyns, P. Y. 1994, 
(Ascle ae eae—Pe sriploc cae). Aloe : 
90 


A note on Md namibiana 

O: 017 lo 

— l The dem am of Eustegia. Bot. 

mus Syst. 121: Ad, 
1 0 


0. T RR ny and biogsographi of 9 stape- 
y st. —22 


Evol. 221: 199 

P. Linder. 199], A revision e ed R. Br 
Bob, Jahn. Syst. 112: 453-527. 

Civeyrel, L. 


liads. . Phylogeny. 


— a L 


. Rowe. 2001. Phylogenetic relationships of 


Sec 1 “ae based on the plastid gene math, nta 
ogy, and biomechanics. Ann. Missouri Bot. Gard. 88: 
283-002 


: 5 2: 
ko ;. D. 1982. 
Asc y 19 1 eae). 


Genus Foc thea Endl. 


Rhodes 


A Revision of the 
Unpublished M.Sc. thes 


p 


niv.. Grahamstown, South Africa. 


1987. Fockea Endl.—An African Genus. Asklepios 


yle, J. J. & J. L. Doyle. 1987. A 
procedure for Vm 
Bull. aoe 
E. 


Jo DNA isolation 
of fresh leaf tissue. 
Phytoc harn, Amer. 19: ad 
Endress, 3 revised 
s ation of 3 vnaceae s.l. Bot. Rev. 66: 
1889. Plantae Marlothianae. 
Südafrikas. Bot. 


Eneler, Ein Beitrag zur 


Kenntnis der Flora Jahrb. Syst. 10: 


242—285. 
"arris, J. S., V. A. Albe l. Kállersjó, D. Lipscomb & A. ( 


Kluge. 1996. Cie ie ac d tg outperforms ne he 
IE l 
g 


bor- joining. ‘ listics 12: 
Kitch, W. 

1 0 17 do» a spec ific tree topology. Sy st. Zool. 2 
1064 

Forster, a L 1905. Circumscription of Marsde nia (Ascle pia- 
Marsdenieae), with 


. Toward de TA m course of iu 


daceae: genus in 
Bot. 8: 703-933 

830. 1 of several new or rare 
lately 


Australia and Papuasia. / Syst. 


Graham, R. plants 


which have flowered in the neighbourhood of 


Edinburgh. and chiefly in the Royal Botanie Garden. 
Edinburgh New Philos. J. 23: 170-178 


Greuter, W.. : McNeill, F. R. Barrie, H. M. Burdet, V 
Demoulin, T. S bog ur. D. H. Nic 1 P. C. Silva, J. 
E. Skog. b Prehane, N. J. Turland & D. L. Hawksworth 
(editors). 2000. Inte ation Code of lai al Nomen- 


clature (S (S saint Louis Code). Regnum Ve 
. H. 1842. Descriptions of seve vn new genera of 
South African Plants. London J. Bot. 1: 18-29. 


00-1809. Fragmenta Botanica. M. 


V. Schmidt, 
Vienna. 

A new approach to the Namib Region. 
Vegetatio 97: 21-38 


Klak, C., A. Khunou, G. Reeves & T. Hedderson. 2003. A 
phylogenetic hypothesis for the Aizoaceae (Caryophyl- 
lale a based on four plastid DNA regions. Amer. J. Bot. 90 


1433-1445 


Volume 93, Number 4 
2006 


Bruyns & Klak 563 


Systematic Study of Fockea 


2004. 
African semi-desert ice 


G. Reeves & T. Hedderson. Unmatched 
tempo of evolution in Southern 
plants. Nature 427: 63-65. 

Kunze, H. 1993. E volution of the rod in Periplocaceae 
and Asc 15 e ae. Pl. Syst. Evol. 185: 99-122. 

1996. Morphology of the stamen in the Ascle pa 

dac 'eae ee its systematic relevance. Bot. Jahrb. Syst. 118 

47-57 

ET ve & S. Liede. 1994. Cibirhiza albersiana, 
a new species of Asclepiadaceae, 5 0 establishment of the 
tribe » oc ds eae. Taxon 43: 367 

Marloth, . R. 1932. The la a South Africa 3(2), 
Jarter br Ve . London. 

McCarthy, 
as and Biomedical Science, Brisbane. 
lia. 

Meve, U. 20 
Ilustrated 
ceae. Springer, | erlin, 

Meyer, E. H. F. 1838 
australiorum. Fasc. 2. 

Moore, S. 1914. Alabastra 


— 


13. School of 
Austra- 


08. =i Town; W. 
996-1 


2G uns Ve son l. 


& U. Meve (editors), 


Asclepiada- 


002. Fockea. In V. Albers 
Handbook E Succulent 
'idelberg. 


Plants. 


e sntatorium de plantis africae 
/oss, Leipzig. 

diversa XXIV. J. Bot. 52: 
A tool for 


alignments. 


1997. 


sequence 


B. € H. B. Nicholas. GeneDoc: 
editing and ip multiple 
Distributed 2 | uthors. 

Potgieter, K Albert. 2001. 
ships i pee s.l. based on tral. intron and 

Ann. 


46—151 
Nicholas, K 


Phylogenetic relation- 
irnL-F spacer se rut x  propagule characters. 
Missouri Bot. Gard. 88: : 549. 

Saghai-Maroof, M. 
W. lard. 198 " 


us in barley: 


\. Jorgensen & R. 


spacer-length poly- 


5 ea R. 
Ribosomal DNA 
endelian inheritance, c 5 


> 


loc RON; and dua dynamics. Proc. Natl. 
A. 8. 


U.S.A : 8014-801 


mane, T.. D. J. Crawford & T. F. Stuessy. 1997. Chloroplast 
N reticulate evolution PY biogeography of 
Amer. J. Bot. 1120-1136. 

Jahrb. 


Paeonia (Paeoniaceae ce 
O K. 1893 
st. 17 AN 

; 189 5. 

Die natürlic la 

Engelmann, Leipzig. 

Sennblad, B. 1997. 
Comprehensive Summaries of . 


sclepiadaceae africanae. Bot. 
Asclepiadaceae. /n H. A. Engler (editor), 
W. 


ym 189-306. 


Pflanzenfamilien 


T the 
sala Dissertations TM 


Phylogeny Apocynaceae 


the Faculty of Science and es Acta Universitatis 
Upsaliensis, . Sala. 

Simmons, M. P. & H. Ochoterena. 2000. Gaps as characters 
in sequence-based phylogenetic analyses. Syst. Biol. 49: 
369-381. 

Swofford, D. L. 2000. PAUP*. Phylogenetic Analysis Using 
Parsimony (and Other Methods). Version 4.0b4. Sinauer, 
Sunderland, Massachusetts. 

Taberlet, P., L. Pautou & J. 1991, 
Universal prime rs for PES ation r e non-coding 
0 


chloroplast DNA. PI. Molec. Biol. 17: 


Gielly, G. Bouvet. 
regions 
1105-110 

Thunberg, 3 P. 1794. 
Joh. Edman, U T 

Verhoeven, R. L., S. Liede & M. E. Endress. 2003. 
position of Hohen and Cibirhiza (Apocynaceae: 
doideae): Evidence from pollinium structure and epDN? 


Prodromus Plantarum Capensium 1. 


The tribal 
Asclepia- 


sequence data. Grana 12: 70-81. 

. Venter. 2001. Pollen morphology of the 

Secamonoideae, and Ascle no ae 

us Missouri Bo. Gard. 88: 

M. L. Wilson é Toussaint van a 

o4 Jan ^ ommelin's Monograph on Cape Flora. 
Univ. Stellenbosch Printers, Stellenbosch. 

Zecher, E. 1988. The old lady of Schoenbrunn. Asklepios 43: 
88-93. 


& 
eee 
(Apocynac eh 

Wijnands, D. O., 


(editor). 


Annals of the 
Missouri Botanical Garden 


APPENDIX 1 


THE 13 MORPHOLOGICAL CHARACTERS AND THEIR CODING AS USED IN 
THE ANALYSES (SEE TABLE 2 FOR ASSIGNMENT OF CHARACTER STATES 
TO SPECIES) 
Rootstock 


Plants without tuber = 0; with tuber = 


2. Upper surface of tuber flush with a of ground = 0; 


well below surface of ground = 


Upper surface of tuber 5 warts = 0; warty = J. 


The rootstock in all species of Fockea other than F. 
multiflora consists of a swollen, fleshy, edible, roughly 
turnip-shaped stem-tuber that either projects slightly 
from the ground (in F. capensis and F. edulis) or is well 
below the surface (in F. angustifolia, F. comaru, and F. 
sinuata). The upper surface of this tuber is distinctly 
warty in F. capensis and F. edulis and is smooth in the 

others. 
Stems 
4. Sap milky — 0: sap clear — 1. 
Leaves 


2. 


eaves frequently narrowly linear = O: never linear = I. 


6. Margins of leaves strongly undulating = 0; not undulating 


Very narrowly linear leaves are characteristic of 


Fockea comaru and F. sinuata and are also prevalent in V. 
| 


angustifolia, where, however, they are extremely. vari- 
able. The margins of the leaves are very strongly 
undul; F. sinuata and F. capensis, and they are 


strongly revolute in F. coma 


Gynostegium 


Corolline corona absent = es present = . 

8. Gynostegial corona absent = 0; present = 1. 

9. Gynoste a corona of one series of lobes behind anthers 
only ): with two series of lobes = 1. 

10. Outer series of gynostegial corona fused into tube = 0; 


not forming tube — 


11. Lobes of inner series of eee corona long, slender, 
): J: i 


Ne much ej g = 0; not e g 


12. aie ith swollen appendages filling up tube of corona 
= (0; without swollen appendages = I. 


Pollinarium 


= 0: with caudicles = 1. 


3. Pollinarium without caudicles 


IS MANDEVILLA (APOCYNACEAE, André O. Simóes,? Mary E. Endress.“ 
MESECHITE AE) Timotheüs van der Mel, Luiza S. Kinoshita,* 
MONOPHYLETIC? EVIDENCE Bened 

FROM FIVE PLASTID DNA LOCI 

AND MORPHOLOGY! 


ABSTRACT 


In order to test the monophyly of Mandevilla Lindl., the largest genus in tribe Mesechiteae (Apocynaceae, Apocynoideae 
and its affinities to other genera in the tribe, maximum E anal vsis was conducted on a data set e 
sequences from five plastid loci (rp/ 16, rps16, and trnK in AS arn inter rgenic spacer; and n ene), as well as 
morphologic val data for 65 taxa of Mesechiteae (48, Hand and nine taxa from 11 8 r tribes of the e Mandevilla, as 

circumscribed by Pichon, was found to be monophyletic, whereas Woodson's circumscription proved to be polyphyletic. Thus 
defined, Mandevilla forms a st ngly supported clade bo can be divided into six clades of species groups. Most of the 
infrageneric taxa of Mandevilla. proposed by Woodson ids Pichon are polyphyletie. Many of the diagnostic characters 
them unsuitable for 


— — 


18 01 used to define taxonomic groups shown to have arisen. multiple times, rendering 
classificatory purposes. The similar pec form pur bud flowe TS of 1 Macrosiphonia Müll. Arg. and ene (Woodson) 
ly disjunct segr iun es, represent the most extreme case of parallel evolution within Mandevilla, with 


Henr., two geographica 
their striking g similarities most likely correlated to colonization of open, dry habitats and pollination by hawkmoths. 
Key wc dd 'ynaceae, Mandevilla. matK, Mesechiteae, 5 phylogenetic systematics, rplló, rpsl6, trnK, 


In S C pur 


Mandevilla Lindl., a member of tribe Mesechiteae, flowers up to 9 em long. The genus is traditionally 
is the largest Neotropical genus in Apocynaceae and characterized by the following set of traits: racemose 
comprises about 150 species (Simões et al., 2004; inflorescence; leaf blade with one to many colleters on 


Sales et al., 2006). It is distributed throughout the the adaxial surface, sometimes extending along the 
Neotropics, from Mexico to Argentina, in a wide midrib; and style head with five strongly protruding, 
variety of habitats such as deserts, savannas, tepuis, well-developed longitudinal ribs (Woodson, 1933; 
open grasslands, and forests. Morphological variation Pichon, 1948; Henrickson, 1996; Morales, 1998; 
is remarkable in the genus in both vegetative and Simões & Kinoshita, 2002; Simões et al., 2004). 

reproductive parts. Most species are vines, but erect A combination of high morphological diversity and 
shrubs are also common, while unbranched subshrubs broad geographic distribution makes Mandevilla one 
f the most challenging and complex genera of 


and epiphytes occur less frequently. Flower size and 
structure are also very diverse, ranging from in- Neotropical Apocynaceae, a fact that is reflected in 
conspicuous white, tubular flowers less than l em its taxonomic history. The currently accepted circum- 
long to brightly colored, showy infundibuliform scription of Mandevilla was defined by Woodson in 


The authors thank the following persons for providing plant material: | eonardo. Alvarado Cardenas, Alberto ds z, Mark 
Fishbein, Andréia Silva Flores, Paul Forster, Leonardo Ga etlo, Günter Gerlach, Sueli Maria Gomes, Phil Jenkins, Stephanie 
Lieberherr Sigrid Liede, Ulrich Meve, Alicia Marticorena, Stefan Matezki, Francisco Morales, E: merson cM Led 
Nilsson, Marie Francoise Prévost, Ana Lilia Reina, Rodrigo Se de Rodrigues, Jürg Schönenberger, Tom Van Devender, 

anoviak and Scott Zona; without their generosity, this study would not have been possible. We are especially grateful to pd 
Wendt and Lindsay Woodruff of the Plant Resources Center at the University of Texas at Austin, for allowing us to remove leaf 
tissue from one of the only three known nes 'clions of Tintinnabularia mortonii. This study was partly supported by grants to A. 


Simóes from CAPES a ordenagac erfeicoamento de Pessoal de Nivel Superior), FAPESP (Fundação de Amparo à 
Pesquisa do Estado de Sao Paulo) 1 m nd the “Stipendien für Doktoriende aus Entwicklungslündern" from the 


Ressort Internationale Beziehungen and the Unifr duello of the University of Zürich, Switzerland. 

Programa de Pós- ae ão em Biologia k ege ES 5 de Biologie: Universidade Estadual de Campinas, Caixa Postal 
6109, Cep 13083-970, Campinas, Sao Paulo, Brazil. Current address: Institute of Systematic Botany, University of Zurich, 
Zollikerstrasse 107, 8008 Zürich, Switzerland. 5 unizh. ch, andresimoes0090 yahoo.com.br 

Institute of Systematic Botany, University of Zurich, Zollikerstrasse 107, 8008 Zürich, Switzerland. mendress@systbot. 
unizh.ch; niet@systbot.unizh.ch; ContiElena@access.unizh.ch 

Departamento de Botanica, Instituto de Biologia, Universidade Estadual de Campinas, Caixa Postal 6109, Cep 13083-970, 


Campinas, Sao Paulo, Brazil. luizakin@unicamp.br 


ANN. Missouni Bor. Garp. 93: 565—591. PUBLISHED ON 15 DECEMBER 2006. 


566 


Annals 
ie san 1 Garden 


1933. In a broad taxonomic study of the Neotropical 
species of subfamily Apo: ynoideae, he made signif- 
Mandevilla, 
including in its synonymy such genera as Exothoste- 
Woodson, Dipladenta A. 


icant changes i 


l the CIPCUMSC ription of 


mon (G. Don DC., Laseguea 


A. DC., Amblyanthera Müll. Arg.. Heterothrix Müll. 
s and part of Echites P. Browne. Macrosiphonia 
Müll. Arg., a small group of shrubby species with long, 
white, tubular flowers and a disjunct distribution in 


S. . and Mexico 


savannas of southern South 


the arid zones of the southwestern. | 


and the America, was 


Woodson as a separate genus. He 
that the 


Macrosiphonia and Mandevilla, based on plant habit. 


maintained. by 


admitted, however, distinctions between 


flowering time, and style head structure, were 


cautiously recognized 
(Woodson. 1933: 
(18), Telosiphonia Woodson and Eumacrosiphonia 


extremely tenuous. He also 


two subgenera in Macrosiphonia 


Woodson (= Macrosiphonia), comprising the species 
that occur in the Northern and Southern Hemispheres, 
respectively. 

In addition to broadening the limits of Mandevilla, 
Woodson (1933) also proposed a morphologically 
based infrageneric classification of the genus, with the 
subgenera Exothostemon and Eumandevilla Woodson 

= Mandevilla). 


ed based on the 


The two subgenera were differentiat- 
following suite of morphological 
characters: species of subgenus Exothostemon have 
leaf colleters distributed along the entire length of the 
midrib, 


calycine colleters with an opposite arrange- 


ment, and a curved corolla tube. whereas species of 


subgenus Mandevilla have leaf colleters restricted to 
the base of the midrib, calycine colleters with an 


alternate or continuous arrangement, and a straight 
corolla tube. 


Within 


proposed a 


Mandevilla, 


subdivision 


(1933) 


five sections: 


subgenus Woodson 


further with 


Laxae Woodson, Montanae Woodson, Tenuifoliae 


Woodson, Torosae Woodson, and Tubiflorae Woodson, 


which were differentiated based mainly on corolla 


shape, anther base shape, and number 


nectaries. The largest section, Laxae, included 46 


species distributed throughout South America and was 
characterized by infundibuliform corollas. Section 
Montanae consisted of 16 species also distributed in 
South America and was characterized by flowers with 
salverform to tubular-salverform corollas, anthers with 
a truncate base, and nectaries varying in number from 
two to five or even absent in 


some species, The 


smallest section, 
South American M. bd (Taub.) 


Woodson and M. tenuifolia (J. C. Mikan) Woodson, 


was distinguished from section ee by having 


Tenuifoliae, comprising only two 


species, 


anthers with auriculate bases and two nectaries. The 


two remaining sections, Torosae and Tubiflorae. have 


and size 0 


five and eight species, respectively, and are distrib- 


uted in Mexico and Mesoamerica. Both of these 
sections were characterized by flowers with salverform 
to tubular-salverform corollas. anthers with auriculate 
bases, and five nectaries surrounding the ovary, but 
differed from one another in the size of the nectaries. 
which were said to be equal to or taller than the ovary 


in section Tubiflorae and shorter than the ovary in 


section Torosae. 
A revised. classification of Mandevilla was pub- 
lished by 


(1933) circumscription by including Macrosiphonia, 


Pichon in 1948. He expanded Woodson's 
which he justified by arguing that the characters used 
by Woodson to differentiate between the two genera 
were inconsistent and arbitrary. He did not consider 
Woodson’s subgenera Macrosiphonia and Telosiphonia 
however, and 


to be each others closest. relatives. 


placed them in two distinct sections, based on the 
absence of a pedicel, longer staminal filaments, and 
larger pollen grains of the former. Within Mandevilla, 
Mandevilla 


and subgenus Exothostemon as valid groups but did 


Pichon recognized. Woodson’s subgenus 


not recognize his five sections within subgenus 


Mandevilla. 


supporting these two subgenera were reliable. whereas 


According to Pichon, the characters 
those supporting the sections were highly inconsis- 
tent, with no real diagnostic states to define them. 
Pichon (1948) proposed 

Wandevilla. 


(Orthocaulon Pichon, Exothostemon Pichon, 


new infragenerie classifi- 


cation. within recognizing four sections 
Megasi- 
phon Pichon, and Telosiphonia Pichon). A summarized 
comparison between the infrageneric classification of 


Woodson (1933) and Pichon (1948) is provided in 


Table 
Since Pichon’s work (1948, 1950), very few studies 
have investigated the taxonomy of Mandevilla and 


1991, 


nia Zarucchi as a monotypic genus morphologically 


related genera. In Zarucchi described Quiota- 
very similar to Mandevilla, and Woodson's subgenus 
Telosiphonia was later elevated to generic status by 
1996). Another relevant work 
synopsis of the Mexican and Central American 
f Mandevilla bv Morales (1998 


taxonomic combinations involving the species from 


P 


Henrickson was 


e 


E 


species 0 with new 


Woodson's sections Tubiflorae and Torosae. many of 


these reduced to synonymy. In addition, a large 
number of new species of Mandevilla have been 


leseribed in the past few decades, increasing the 


number of published Spe cies from the 108 recognized 


^ 
= 


by Woodson in o about 150 at present. 
Although new information has been accumulating for 
the genus. no overall classification within Mandevilla 


as a whole has been proposed since Pichon (1948). 


Taxonomic. difficulties involving both generic and 
infrageneric concepts have persisted for the past 


Volume 93, Number 4 Simóes et al. 567 
2006 Is Mandevilla Monophyletic? 
Table 1. Comparison of Woodson's (1933) subgenera and DNA sequence data from five chloroplast DNA 


and sections of Mandevilla and their corresponding ranks in 


Pichon's (1948) classification. 


Woodson (1933) Pichon (1948) 


Mandevilla sect. 
Exothostemon 

Mandevilla sect. 
Orthocaulon 


Mandevilla subg. 
Exothostemon 
Mandevilla sube. 
Wandevilla 
Section Laxae 
Section Montanae 
Section Tenuifoliae 
Section Torosae 
Section Tubiflorae 
Genus Macrosiphonia 
Subgenus Macrosiphonia Mandevilla sect. Megasiphon 


Subgenus Telosiphonia Mandevilla sect. Telosiphonia 


seven decades and 
Zarucchi 


generic 


still remain, as pointed out by 
(199]: 35): “The 


limits of the Mandevilla—Mesechites-Macro- 


ast word concerning 
siphonia complex gue near relatives has obviously not 
yel been written.’ 

The 


successfully applied in Apoeynaceae to solve contro- 


use of phylogenetic methods has been 


versial aspects of classification within the family. 
Previous studies have addressed the circumscription 
of Apocynaceae s. str. and their relationships with the 
Ascl | (e.g. Judd et al., 1994; 
Sennblad & Bre mer, 1996. 2002: Potgieter & Albert, 


2001). but a growing number of works have focused on 


former 


relationships within Apocynaceae s. str. Examples 
(1996) and 
Sennblad et al. (1998) for tribe Wrightieae, a study 


by Endress et al. (2007) for Alvxieae, and a larger- 
P \ B o 


include overviews by Endress et al. 


scale study of subfamily Apocynoideae by Livshultz 
et al. (2007). Phylogenetic studies based on morpho- 
logical characters were also published by Sidiyasa 
1998) for A/stonia R. Br.. van der Ham (2001 
Alyxieae, and Williams (2004) for Echites. 


Simões et al. (2004) provided the first phylogenetic 


for 


— 
— 


study of tribe Mesechiteae, with suggestions for 


taxonomic improvements in tribal and intergeneric 
delimitations. Preliminary results were obtained for 


Mandevilla and related genera, but due to the limited 
taxon sampling within Mandevilla, no firm conclu- 
sions could be drawn as to infrageneric relationships. 
Our present study represents the subsequent second 
step in interpreting the phylogeny of Mesechiteae by 
focusing on the intergeneric and infrageneric relation- 
ships of its largest genus, Mandevilla. 

The aims of the present article are to test the 
monophyly of Mandevilla and determine its relation- 
ships to the putatively affined genera Macrosiphonia, 


Telosiphonia, and Quiotania, using both morphology 


loci. The resulting phylogenetic hypotheses of mono- 
phyly and infrageneric relationships of Mandevilla are 
compared with the classifications of Woodson (1933) 
and Pichon (1948). Morphological features consistent 
the 


taxonomic 


with retrieved clades and/or used to define 


ranks are also discussed. 
MATERIALS AND METHODS 
TAXON SAMPLING 


Sixty-five taxa of Mesechiteae, including represen- 


latives from all genera of the tribe recognized by 


Simões et al. (2004) (Allomarkgrafia Woodson, 
Forsteronia G. Mey.. Macrosiphonia, Mandevilla, 
Mesechites Müll. Arg., Telosiphonia, and Tintinnabu- 


laria Woodson), were included in this study (Appen- 
dix 1). In order to test the infrageneric classifications 
of Mandevilla proposed by Woodson (1933) 
(1948) (Table 1). 48 (from 47 


species) of Mandevilla, representing all subgenera and 


and 


Pichon 


sections, were sampled (Table 2). Nine outgroup taxa 


representing all but the basalmost tribe (Wrightieae) 


of subfamily Apocynoideae were chosen, based 


argely on previous studies suggesting that the closest 
either Apocyneae or 
Echiteae (Sennblad et al.. 1998: Sennblad & Bremer, 
2002; Simões et al, 2004). Two genera from Echiteae 
(Prestonia R. Br. and Rhodocalyx Müll.“ 


from Apocyneae (Beaumontia Wall., 


relatives of Mesechiteae are 


Are.) and five 
Chonenorpha G. 
Benth.. Secondatia A. DC.. 


Trachelospermum Lem.) were included. Two species of 


Don, Odontadenia and 
Pachypodium Lindl. (Malouetieae) were used to root 
the cladograms. 
DNA EXTRACTION, AMPLIFICATION, AND SEQUENCING 

Total genomic DNA was extracted from silica dried 
leaf material or from herbarium 
DNeasy Plant Mini Kits (Qiagen, 


U.S.A.) following the manufacturer's. protocol. 


specimens using 
Valencia, California, 
Five 
and trnK introns: /S 
math 
Double-stranded DNA was amplified by 


(PCR) on a 


Tgradient machine (Biometra, Göttingen, 


plastid loci (rp! 16, rps 16, 
rn 


amplified. 


intergenic spacer; and gene) were 


polymerase chain reaction Biometra 


Germany). 
applying a thermal DE a program of 34 cycles of 
annealing al 


ee at 95 30 seconds., 


for l minute, and extension at 72°C for 
90 seconds. The K intron and matK gene were 
co-amplified in a single PCR reaction, and the 


thermal cycling program was modified in the following 


steps: denaturation at 94°C for 30 seconds and 
annealing at 54 C for | minute. For some taxa, 
amplification of the entire trnK intron/matk locus 


568 Annals of the 
Missouri Botanical Garden 


Table 2. List of the sampled taxa of Macrosiphonia, Mandevilla, and Telosiphonia and their placement in the 
classification of Woodson (1933) and the clades observed in the present study. 


Taxon name and current classification This study 


Genus Macrosiphonia Müll. 


Macrosiphonia longiflora ( 185 sl ) Müll. Arg. Clade | 
Macrosiphonia martii Müll. Arg. Clade I 
Macrosiphonia velame (A. St.-Hil.) Müll. Arg. Clade I 


Genus Mandevilla Lindl. 
Subgenus Mandevilla Woodson, as “Eumandevilla? 


Section Laxae Woodson 


Mandevilla atroviolacea (Stadelm.) Woodson Clade IH 
Mandevilla callista Woodson Clade | 
Mandevilla coccinea (Hook. & Arn.) Woodson Clade HI 
Mandevilla convolvulacea (A. DC.) Hemsl. Clade IM 
Mandevilla duartei Markgr. Clade IH 
Mandevilla fragrans (Stadelm.) Woodson Clade HI 
Mandevilla funiformis (Vell.) K. Schum. Clade I 
Mandevilla o (Ruiz & Pav.) Woodson Clade IV 
Mandevilla harleyi M. F. Sales, Kin.-Gouv. & A. O, Simões Clade | 
Mandevilla 11 8 (Vell.) Woodson Clade III 
Mandevilla laxa (Ruiz & Pav.) Woodson Clade IV 
Mandevilla ligustriflora Woodson Clade IV 
Mandevilla martiana apr Woodson Clade HI 
Mandevilla moricandiana (A. DG.) Woodson Clade HI 
Vandevilla oaxacana (A. DC. y d Clade IN 
Mandevilla pendula (Ule) Woodson Clade HI 
Mandevilla pohliana (Stadelm.) A. H. Gentry Clade IH 
Mandevilla sancta (Stadelm.) Woodson Clade HI 
Mandevilla sellowii (Müll. Arg.) Woodson Clade IH 
Mandevilla spigeltiflora (Stadelm.) Woodson Clade IH 
Mandevilla splendens (Hook. f.) Woodson Clade HI 
Mandevilla urophylla (Mook. f.) Woodson Clade HI 
Mandevilla venulosa (Müll. Arg.) Woodson Clade HI 
Mandevilla reraguasensis (Seem.) Hemsl. Clade [V 
Section Montanae Woodson 
Mandevilla cercophylla Woodson Clade IV 
Mandevilla emarginata (Vell.) C. Ezcurra Clade IV 
Mandevilla jamesonii n Clade IV 
Mandevilla pentlandiana (A. DC.) W ng Clade IV 
Mandevilla pyenantha (Ste m ex A. DC.) Woodson Clade IH 
Mandevilla tricolor Woodson Clade I\ 
Section Tenuifoliae Woodson 
Mandevilla myriophyllum (‘Taub.) Woodson Clade HI 
Mandevilla tenuifolia (J. C. Mikan) W oodson Clade III 
Section Torosae Woodson 
Mandevilla foliosa (Müll. Arg.) Hemsl. Clade IV 
Mandevilla karwinskii (Müll. Arg.) Hemsl. Clade IV 
Section Tubiflorae Woodson 
Mandevilla syrinx Woodson Clade IV 
Vandevilla tubiflora (M. Martens & Galeotti) Woodson Clade IV 
Subgenus Exothostemon (G. Don) Woodson 
Mandevilla anceps Woodson Clade I 
Mandevilla dodsonii A. H. Gentry Clade I 
Mandevilla hirsuta (Rich.) K. Schum. Clade | 
Mandevilla krukovii Woodson Clade I 
Mandevilla lancifolia Woodson Clade I 
Mandevilla leptophylla (A. DC.) K. Schum. Clade I 
Mandevilla nerioides Woodson Clade | 


Volume 93, Number 4 Simóes et al. 569 
2006 Is Mandevilla Monophyletic? 
Table 2. Continued. 
Taxon name and current classification This study 
Mandevilla rugellosa (Rich.) L. Allorge Clade I 
Mandevilla rugosa (Benth.) Woodson Clade 1 
Mandevilla scabra (Hoffmanns. ex Roem. & Schult.) K. Schum. Clade I 
Mandevilla subsagittata (Ruiz & Pav.) Woodson Clade I 
Genus Telosiphonia (Woodson) Henr. 
Telosiphonia brachysiphon (Vorr.) Henr. Clade IV 
Telosiphonia hypoleuca (Benth.) Henr. Clade IV 
Telosiphonia nacalpulensis Felger & Henr. Clade IV 
was only possible by fragmenting that region into two columns (Amersham Pharmacia Biotech Europe. 


parts, using a combination of external and internal 


primers. Reactions were terminated with a final 


extension of 4 minutes at 72°C for the rp//6 intron, 


rps16 intron, and trnS°-trnG'*© spacer, and 7 min- 
utes for the trnK intron and matk gene. All PCR 


reactions were performed in a total volume of 25 UL. 
2.5 mM MgCl» 10X PCR* buffer ( 
Piscataway, New Jersey, U.S.A.), 
0.25 mM of dNTP, 0.5 units of Taq DNA polymerase 
lot 17544), 1 to 4 ul of 
(BSA: 


Germany). and 0.1 mM of each primer. 


using 2 Amersham 


Biosciences, 


(Amersham Biosciences, 


bovine serum albumin Sigma, Steinheim, 


Table 3. For 


some taxa, internal primers were also used to amplify 


Primer information is presented 


the rp/16 intron and trnS ^ -trnG' © intergenic spacer. 
with the following changes in the thermal cycling 


program: 40 instead of 34 cycles and extension time 


shortened to 1 minute. Successfully amplified PCR 
products were then purified using GFX PCR DNA and 
Gel Band Purification Kit (Amersham Biosciences). 
For some taxa with DNA extracted from herbarium 
vouchers, no amplified products were obtained with 
our initial PCR protocols and primer sets. In those 
cases, amplification was only successful using a two- 
NOR 


amplification was performed using the total genomic 


step amplification procedure. A first round of 


DNA as a template, followed by a second round using 
one external and one internal primer and a 10% 
dilution of the product of the first amplification. as 
a template. The two-round amplification. procedure 
was used to obtain products of the trnK intron and 
alk gene for six species (Mandevilla anceps 
Woodson, M. krukovii Woodson, M. lancifolia Wood- 
son, M. leptophylla (A. DC.) K. Schum., M. 


Woodson. 


nerioides 
M. tricolor Woodson, and Tintinnabularia 
mortonit Woodson). 

Cycle-sequencing reactions were carried out using 
an ABI Prism Big Dye Terminator Cycle Sequencing 
Ready Extraction Kit (Perkin Elmer, Applied Biosys- 
Rotkreuz, 
Sequence products were purified on MicroSpin 6-50 


tems, Switzerland). 


Applera Europe BV, 


Dubencort, Switzerland) and loaded on an ABI Prism 
7 DNA sequencer (Perkin Elmer). Complementary 
one were edited and assembled with Sequencher 


3.1.1 (Gene Codes, Ann Arbor, Michigan, U.S.A.). 


DATA MATRIX COMPOSITION AND PARSIMONY ANALYSES 


Nucleotide sequences of the studied plastid loci 
were aligned using Clustal W version 1.8 (Thompson 
f 


mononucleotide repeats with variable length between 


— 


el al., 1994) and adjusted visually. Regions « 


taxa, as well as those composed of nested gaps 


resulting from ambiguous alignment, were excluded 
from the analysis. Aligned gaps were manually coded 
as presence/absence characters by applying the single 
indel coding method described by Simmons and 
Ochoterena (2000) for the matK gene and by using 
the software Gapcoder (Young « Healy, 2003) for the 
All coded 
sequence matrix and used in further analyses. 


Thirt 


using a combination of herbarium and fresh specimens, 


other loci. gaps were then added to the 


y-two morphological characters were coded 


pickled flowers, and, when available, flower sections 


prepared by the second author. For some taxa, the 


literature was also consulted (e.g., Woodson, 1933; 
Pichon, 1950; Leeuwenberg, 1997; Morales, 1996, 
1997, 1998). The morphological matrix and the 


characters and character states, including some 


xplanatory notes on characters, are given in Appen- 
dices 2 and 3, respectively. 

A total of six data sets were subjected to 
phylogenetic analysis, corresponding to the five loci 
sequenced plus morphology. Because simultaneous 
analysis of combined data sets has been proposed as 
the best approach to phylogenetic inference (Nixon & 
1996), 


by searching for incongruence between 


Carpenter, we tested the combinability of all 
partitions 
individual data sets. For this, we compared the results 
on a node-to-node basis of all individual data sets with 
respect to levels of resolution and bootstrap support 


(BS), as applied by other authors (e.g., Wiens, 1998; 


570 Annals of the 
Missouri Botanical Garden 
Table 3. DNA sequences of the primers used for amplification and sequencing of the five plastid loci used in this study. 


Primers designed for the two-step amplification are indicated by an asterisk 


m. 


cpDNA locus 


Primers Primer source 


rpl l6 intron Fl 5 n UPATGCTPAGTGTGTGACTC( ien 3! Baum & Wendel, 1998 
RIS1O "-GCCTTCATTCTPCCTCTATGTTG- Baum & Wendel, 1998 
513K 5.666 AACGATGGAAGCTG 15 qe Simoes el al., 2004. 
542R 5'-CGCGGGCGAATATTTACTCTTC-3' Simões et al., 2004 
*F73 5'-CYCATTACTTCGCATTATCTC-3' Phis study 
* 1] R582 5’-CGACCAGTGAATCATTAAGAT-3' This study 
[14,709 5'-ACAAATTTCATTATGAGCTCC-3' Phis study 
R 1060 5 -GOGAATAAAAGAATIMAAA-3' Phis study 
rps 16 intron rpsk 5'-TGGTAGAAAGCAACGTGCGACTT-3’ Oxelman et al., 1997 
rpsR2 5'-TCGGGATCGAACATCAATTGCAAG-3 Oxelman et al., 1997 
87 5'-GCACCGAAGTAATGCCTAAACC-23' Simões el al., 2004 
197R 5 -GGATTCTRAAGTCTGGCCCAG-3/ Simoes el al., 2004 
11486 5'-WAACTGGGCCAGACTTMAGAA-2' Phis study 
ER 768 5'-CGAATAAATTACATAAAAGG-3' Phis study 
*R782 5'-ATGGAATTCGAATAAATTACA-3' Phis study 
ras trn G spacer tri 5 -GCCGCTPTAGTCCACTCAGC-3! Hamilton, 1999 
mo SGAACGAATCACACTTTTACCAC-3" Hamilton, 1999 
3091 5'-GATGATTTTTCATTTATMTGA-3' Simões et al., 2004 
527R 5'-GTGCTWAAATATTTCY YATTMAC-3' Simões et al., 2004 
in intron + math irnlk 3914F 5 -GGGGTTGCTAACTCAACGG-3" Civeyrel & Rowe, 2001 
eene math BF 5'-AATTTCAAATGGAAGAAATC-3' Civeyrel & Rowe, 2001 
math 174R 5'-GCGAKTAATPAAMCGTTTCAC-3' Civeyrel & Rowe, 2001 
math 8F 5'-AATTTCAAATGGAAGAAATC-3' Civeyrel € Rowe, 2001 
math 503R 5'-GCATCTTTTACCCAATAGCGG-3' Civeyrel & Rowe, 2001 
math 5031 5'-TCGCTATTGGGTAAAAGATGC-3' Civeyrel € Rowe, 2001 
math 681E 5 -GTGAATACGAATCYATPPTC-2 Civeyrel € Rowe, 2001 
math 900F 5'-TGCAAATPTTACCTTGTCAA-23' Civeyrel & Rowe, 2001 
math 1628R 5'-GCATGCTACATCAACATPTGCAG-3 Civeyrel & Rowe, 2001 


math 1309F 


ink 2R 


5'-GACTTTCTTGTGCTAGAACT-2' 
5'-AACTAGTCGGATGGAGTA-2' 


Civeyrel & Rowe, 2001 
2001 


Civeyrel € Rowe, 


Sheahan & Chase, ; Reeves 


et al., 2 


2000; Whitten et al., 2000 
001). Because the trees generated. from the 
individual data sets did not show any topological 
conflict when supported by bootstrap values greater 
data partitions. were then combined as 
follows: all molecular data sets combined (molecular 
combined) and all molecular data sets combined with 
morphology (total evidence). 
Maximum parsimony analys performed 
PAUP* 4.0b (Swofford, 2000). 


were unordered and equally weighted. Polymorphisms 


ses were 


using All characters 
in the data matrix were treated as such, rather than as 
uncertainties. A heuristic search for most parsimoni- 
ous trees (MPT) included (1) an initial round of tree 
searches with 1000 
replicates. (RASR), 
(2) 
swapping with MULTREES and steepest descent in 
effect, 


random addition 
10 


bisection-reconnection (TBR) branch 


sequence 


holding trees at each step. 


and tree 


saving a maximum of 50 trees at each replicate. 


All shortest trees retained in P memory were then 


included in a second round of searches involving 


exhaustive TBR branch. swapping. Relative support 


for each node was estimated using the bootstrap 


resampling procedure (Felsenstein, 1985) as imple- 
mented in PAUP, employing a heuristic search with 
500 replicates, 250 RASR with three 
each step, and TBR branch swapping with steepest 
descent and MULTREES in effect. 
each RASR. 

Morphological characters were mapped onto the two 
the total 


trees held at 


saving 10 trees al 


most parsimonious trees resulting from 
evidence analysis using MacClade 4.0 (Maddison & 
Maddison, 2 


phies that are congruent with each of the major clades 


000) in order to identify the synapomor- 


of Mandevilla retrieved in our analyses. Unambiguous 
maximum 
parsimony applying both accelerated. (ACCTRAN) 
and delayed (DELTRAN) character state optimizations. 


changes were then reconstructed with 


RESULTS 


Amplification of the five selected loci was routine 
for most taxa. The two-round amplification of the trn K/ 


matk locus was only partially successful for Mande- 


Volume 93, Number 4 


Simões et al. 571 


2006 Is Mandevilla Monophyletic? 

- 9 — villa anceps, M. leptophylla, and  Tintinnabularia 
G = S ＋ Y wa FEM ` : f M 

E ES S. TANE. mortonit and failed completely for M. krukovii, M. 

— [ooa N oco 2" . f 

5 D S S FJS c S nerioides, and M. tricolor. 

LE en 15 — s 

— 2 2 = m Multiple sequence alignment for the matK gene 

= = required only a few gaps that, without exception, 
8 occurred in multiples of three. Alignment was 
S E 5 also unproblematic for the trnK and s introns 
T * S InN 

E "e = el but proved to be somewhat more difficult for the 

s . 2 <<< c — OS ! E : .4QGCU. ,44(UUC aa ! > 

E S S 2 = = rplló intron and traS®-trnG''© spacer due to the 

E = = larger number of gaps and mononucleotide repeats. 

3 A total of 165 characters, including nucleotides 
8 a ae and gaps, were thus excluded from further analyses 
> == S A : . : ] ier 

= = 8 > ed of the noncoding loci, mainly from the rp/16 intron. 

a a 2 ) . j 

= Sen] + O 10 N mos PME 

E Z2 |=%Y0uw e in Manual verification of the coded gap characters 

= S ^ D — = Ko) 

P. — ^ eo — x 2 

E = 3| ` e showed that Gapcoder performed well, even in 

s 00 , ' . 0 

50 cases of overlapping gaps with different start and/or 
o ` Fi . 

z ending positions, and no further adjustments in the 
z : + e M id matrices were necessary. More detailed. information 
c O xU Ox eb aac for the individual and combined data sets is given in 
4. T c — e^ 

E AR E a. O an m 

2 zx E " l'able 4. 

= I <} = 

— = zs N 

E: PARSIMONY ANALYSES 

m = S ON] rA : . : : 

ES E S EN = l'ree length, consistency index (CI). and retention 

p [STARS 8 = index (RI) for the cladograms resulting from the 
i 84 8 E N l 1 

D Ep 2 analyses of the individual and combined data sets are 
rs ~ summarized in Table 4. From the individual molec- 

ze ular data sets, the best-resolved cladogram was 

— 8 8 provided by the matK gene, with most of the nodes 
= E = = e E 85 receiving greater than 50% BS (Fig. 1). The highest 

ze Sa) Te s nos proportion. of parsimony informative characters, as 
z 2 © N 3 = TA . N 

E * E well as the highest Cl and RI values, were also 
E 5 E provided by this data set. Of the other data sets, the 
m trnK intron (Fig. 1) and rpll6 and %s introns 

S (Fig. 2) had similar levels of resolution; the least 

8 Z o e — un resolved cladogram was provided by trus  -tirnG* * 
m Z a3 N 2 2 à : : ` . 

— — | = i ut intergenic spacer, with the lowest number of 
E S 8 " ar e i E e fme 
5 DIN ER N N nodes supported by at least 50% BS (Fig. 3). Of 
) e. — e i RAS T d 
E ll these cladograms, only the matK and trnS** -ÉrnG ute 

E trees defined a clade containing all species of 
25 Mandevilla, Macrosiphonia, and Telosiphonia with 

2 z ci e CR BS higher than 50%. Because no strongly supported 
D H D ~ in — Ro a b ris 

m Elalao e >z (> 75%) incongruent clades were found between 
E us | mri `O re oO A "e c 

S. EIER bs 7 individual partitions, all molecular data sets were 

I "3 e — . PR . . 
"EE E ^ combined. Their analyses vielded the tree shown in 
Zg Figure 4. 

MEE - Analysis of the morphological data set resulted in 
0 d qe — E : 

= = gs 9 E a poorly resolved cladogram with only a few groups 
2 ne 5r — — z ` 7 . E TEE 

E ov B a ARE x supported by a BS value higher than 50% (Fig. 3). No 
2 En 4 . S : i . " us 

= 8 2 2 5 ps E: incongruent clades with BS greater than 75% were 
4 E Ss 2 n i F . 5 
As s 3 8 S eg detected when comparing the morphological tree with 

— Sp y ww VU . . ~ . ai Se 

0 = 7 Boa: = either the strict consensus of the individual or 
Iu — — 5 — C E . mn ~ 

w S 8 8 583 8 = combined molecular trees. Therefore, the morpholog- 

m EN Cm ical and combined molecular data sets were combined 

8. * S2 2 E E S to form a total evidence data set. 


572 Annals of the 
Missouri Botanical Garden 
tmk intron r 
atk gene 
xus math gene 95 , brenesiana 
plum rit ‘flo T xls , plumenitlora 
e. Dupuis E; Mansoanus 
fe. e. trifidus 
M 1 5 97 100 Ae. minimus 
e.trifidus dei 
Ti. gratissima II. gratissima 
i. mortonil 1. mortoni 
ouci couci 
refracta refracta 
llozia velloziana 
c.longifora c. longiflora 
c. velame C. 9 1 
1. men 4 pred 
dodsonii à Tog 
War A subsagital 
a, rugos 
E a 98 | L us 
à. rugellosa scabra 
a. subsagittata 1, anceps 
allista callista 
a. funiformis a. funiformis 
. harleyi a. harleyi 
a. atroviolacea a. atroviolacea 
a. martiana a. pendula 
. pendula a. Sancta 
à. sancta à. sellowii 
à. sellowii a. urophylla 
84 : Se a. coccinea 
ophylla a. illustris 
T iana a. pohliana 
nulosa a. spigeliiflora 
a. aie [76 | a. martiana 
a. illustris a. moricandiana 
a. pohliana a. splender 
a. em a. venulos: 
a. dua a. fragrans 
a. p 1 5 a. duartei 
d. myriophy lun: a. myrio hyllun 
a. tenuifolial 75 
a. pyenantha ie 
a. tenuifolia 
a. cercophylla 
a. emarginata 
= a. glandulosa 
m~~ 92 74. [—— —— Ma. jamesonii 
L— Ma. veraguasensis 66 
— a. laxa 
a. ligustriflor: 
a. pent Handiana 
purse | pom a. convolvulacea 
a. bel 100 
a. folios — — 
100 = — a 100 
[ 1| 65 i 
e. hypoleuca hy 
Te. nacalpulensis "bnt 
la. oaxacana e. naca 
fa. tubiflora la. folios 
79 O. lutea ). due 
— 11 mM 100 mE 
iss | riedelii — Dini ral 
—— n tundif olius nsiflora 
oa ß B. grandiflora 93 o a 
100 Ar agrans Li [ — “Tragrans 
Tr jasminoides jasminoides 
Da, geayi . veavi 
Pa. Tamerei 5 
Figure . Strict consensus of the most parsimonious trees generated by the sra K intron and math gene data sets. Bootstrap 
values > 50% are indicated above the branches. Full taxon names are given in Appene 


The total evidence tree (Fig. 5) contains a strongly 


supported clade (BS = 100%) including representa- 


lives of Allomarkgrafia, Forsteronia, Macrosiphonia, 


Mandevilla, | Mesechites, (the 


Mesechiteae clade). Within this clade, three strongly 


and Tintinnabularia 
supported clades were recovered: (1) the Mesechites 
clade (BS = 100%). 
Mesechites, and Tintinnabularia; 
clade (BS = 


comprising Allomarkgrafia, 
(2) the Forsteronia 


99%). formed by the three sampled 


species of this genus; and (3) the Mandevilla clade 
(BS = 94%), comprising species of Macrosiphonia, 
Mandevilla, 
consistent wilh our earlier findings (Simões et a 


2004). 


and result that is 


Telosiphonia, a 


Within the Mandevilla clade, two major, strongly 
supported clades (Fig. 5; BS 100%) were re- 


— 


covered: (1) one formed by Macrosiphonia, 11 species 


of Mandevilla subg. Exothostemon, and three species 


Volume 93, Number 4 


Simoes et al. 573 


2006 Is Mandevilla Monophyletic? 
. l ; 100 . brenesiana 
rpl16 intron 99 brenesiana rpsl6 intron ons 
l . plumeriiflora lc. ma 
le .mansoanus le. UN 
Ae. trifidus Me. minimus 
le .minimus 3 roseus 
e. roseus 1. gratissima 
1. gratissima i. mortonii 
1. mortọnii . àcouci 
Que refracta 
acta velloziana 
velloziana 89 c. longiflora 
a prc 
'elame . martii 
als — — 
allista B nerioides 
funiformis lancifolia 
arley 54 odsonii 
dodsónii 99 - krikovii 
anceps —— . leptophylla 
nerioides 93 hirsuta 
rsuta — 1 rugosa 
gosa à 
a. subsagittata d ata 
rugellosa rugellosa 
scabra 88 callista 
— — [—— Ma. lancifolia — 100 ` funiformis 
Loo Ma. krukovil harlevi 
a. leptophylla ao geen 
coccinea 
duartei 
IS m fragrans 
sellowii 
58 55 
— 93 . splendens 
Da 
. pendula 
. pohliana 
— à. sancta 
illustris 
. spigeliiflora 
. urophylla 
à. venulosa 
. cercophylla 
58 . emarginata 
. laxa 
sol — . pentlandiana 
87 . gla a n 
73 D i mesoni 
trico 15 s 
| — Ma. veras guasensis 
. ligustriflora 
100 US 
_oaxacan 
ou 
50 ; tubiflora 
= karwinskil 
00 29. a. oaxacana hypoleuca 
— à. svrinx . nacalpule 
a. tubiflora brachysiphon 
a. foliosa à. foliosa 
a. kanvinskii T 81 a, myriophyllum 
9.0 e. brachysiphon 65 £ a. tenuifolial 
99 Te. e — Ma. tenuifolia2 
: e. hypo Ma. pyenantha 
20 r— Q. lutea 65 SO . grandiflora 
— Sd nsiflora E a Fragrans 
100 — Pr. riedelii Tr. jasminoides 
L— R. rotundifolius 90 0 lutea 
73 EE ; grandiflora L— — S. densiflor 
Mores Tragrans 99 — Pr. riedelii 
Tr jasminoides R. rotundifolius 
a. geayi Pa. geay 
P Tamerei Pa al 


Figure 2. Strict 


consensus of the most parsimonious trees ge 'nerated by the rpl l6 intron and rps lO intron data sets. 
ndix 1. 


8 
Bootstrap values > 50% are indicated above the branches. Full taxon names are given in Apper 


© 


subgenus Mandevilla (M. callista Woodson, M. 
funiformis (Vell.) K. Schum., and M. harleyi M. F. 
Sales. Kin.-Gouv. & A. O. Simões), hereafter referred 


( 


to as Clade L and (2) another formed by Telosiphonia 
and all remaining species of Mandevilla sampled. 
hereafter referred to as Clade II. 

Within Clade II. 


diverse than Clade I and more extensively sampled in 


which is more morphologically 


our study, two strongly supported clades were re- 


FS 


covered: (1) a clade comprised of species « 
Mandevilla mostly from central to southern South 
America, hereafter referred to as Clade I, and (2) 
a clade consisting of Telosiphonia and species of 


Mandevilla with a wide range of distribution from 


Mexico to southern South America, hereafter referred 
to as Clade IV. Clade | 


into two smaller clades: (1) Clade V, a heterogeneous 


V can, in turn, be subdivided 


assemblage composed of the South American species 
of Mandevilla and (2) 


Clade VI. formed by Telosiphonia and all Mexican 


with truncate anther bases, 
species of Mandevilla. 
Mapping morphological characters onto the Man- 


devilla that Clades I, IV, and V are 


supported by unambiguous changes in character state 


clade shows 


(Fig. 6). In Clade I, opposite calycine colleters are 
derived from colleters with an alternate to continuous 


arrangement (character no. 15, Fig. 6), with a reversal 


to the ancestral state in the subclade formed by 


574 Annals of the 
Missouri Botanical Garden 
G Inferoenie «nace 
tm5-G intergenic spacer 91 A brenesiana Morphology 
A. plumertiflora A. brenesiana 
87 Me. mà insoanus A. plumeruflora 
87 le. trifidus Ne mansoanus 
86 le. minimus Me, e, minimus 
Me. roseus lc. roseus 
|, gratissima e le tias 
1. mortonil 1. gratissima 
64 r B. grandiflora 7 1. mortont 
95 C. fragrans 81 19 acouci 
. 2m voides da M. 
72 F. velloziana 
2 A Mec. longiflora 
A ” 
1 na | | 0 n 
c. longiflora j m 
c. martii de lend! 
c. velame E À = iphon 
à anceps On i 
d. Nerloldes . anceps 
a. lancifolia a. ve 
1 dont : alee 
a. dodsonii 
a 115 a. hirsuta. 
a Sis a cb l 
a. scabra ellos. 
a. subsagitlata a ea 
A -krukovi 1 
cl E ll 
a. funiformis Ma. neroides 
a. harley! Ma. mi 
a. atroviolacea moo 
a. coccinea i debis acea 
67 s hylla 
a. illustris - exophyl 
a. pohliana ; 4 | à glandulosa 
a. spigelitflora i apaa 
53 a. fragrans a. ligustriflora 
100 a. martiana Ma. tricolor 
am ricandiana Ma. en 
endula Ma. cocti 
a, Sancta a. convolulacea 
sp ll E== à 1 
à. splendens |. LEM 
à. venulosa à tübilor 
a. duartei 
3 ame 10 [— —— Ma. emarginata 
ea 
a à myrigphyllum a. EUM 
tenuilo a. foliosa 
à pycnantha 1 ue 
a. tenuifolia2 i E. 
a. cercophylla à 171 Il 
a. emarginata ib 
— a. glandulosa 100 a. mate " 
E - | A in Sen A myriophyllun 
i tricolor 60 ang 
aye | a. tenuitolial 
5 jm flora a. 0 
Ma. pentlandiana i M lu 
Ma. Veraguasensis a. pohliana | 
D d à. pycnantha 
iun inskii 4. Sancta 
vole cuca a. sellowii 
a «nmoelntlor 
e. qus hysiphon O 
e . 8 ^ 
. nacalpu 15 INN 1 sp 1 
a. oaxacana ! i "m T 
SVIIDN . venu 
übitlora ; prandiflora 
E foliosa e. Iragran 
utea lutea 
reel = p— Pr. riedeli 
ol 3 rotundifolius 
pe 1 5 S. densiflora 
S. densiflora 4 
a. geayi Tr, 7 es 
a. lamerei Pag Quy 


Figure 3. Strict consensus of 
sae a ul data sets. 


Appendix 


50% 


Bootstrap values 


species of Macrosiphonia. In Clade IV. a shift from 
short to long style head appendages was noted on all 
terminal branches, except in the subelade formed by 
species of Telosiphonia (character no. 29, Fig. 6). In 
Clade V, anthers with a truncate base evolved un- 
ambiguously from the ancestral state of a cordate base 
(character no. 20, Fig. 6). No unequivocal morpholog- 
ical synapomorphies were found to support Clades II. 
HI, and IV, as ACCTRAN and DELTRAN optimizations 


resulted in different reconstructions of character state 


are indic ated above the branches. 


changes 


(Fig 


the most parsimonious trees generated by the ins! -trnG 


Full 


s0); 


Utt 


laxon 


intergenic 
names 


1d. IdIICICI 


spacer and 
are given in 


A more detailed explanation of 


morphological characters and changes of state is given 


in the discussion of each individual clade. 


DISCUSSION 
PHYLOGENETIC 


In our 


Mandevilla, 


study, 


HYPOTHESIS AND CURRENT CLASSIFICATIONS 


the clade 


formed 


by 


pro 


species 0 


Macrosiphonia, and Telosiphonia (Man- 


devilla clade) largely corresponds to the cireumscrip- 


Volume 93, Number 4 
2006 


Simoes et a 
Is Mandevilla Monophyletic? 


575 


Molecular combined 


99 


A. brenesiana 


= plumeriiflora 
Me. mansoanus 
Me. trifidus 


BD g Me. a 
Me. roseu 


E 
II. 8 
Lo 


i. mortonii 
`. acouci 
. refracta 


F. velloziana 


Mc. longiflora 
Vic. velame 


a. leptophylla 
a 15 lsonũ 


SSSS5558558 
5 A5 3 c9 


a. scabra 

a. subsagittata 

a. anceps 

à. nerioides 

callista 

à nulos 

1. harleyi 

a atroviolacea 

a. pendula 

a. sellowii 
san 


a. s 

a. moricandiana 
a. venulosa 

à. coccinea 

a. illustris 

* pohliana 

i dins a 
ra 


a. dua 
3 oc vium 
a. tenuifolial 
a. tenuifo ee 


: pyen 
a. ec 
A m ginata 


3 pentlandiana 
a. glandulos 

a. remi 

A mor 
eraguasensis 
3 ligustriflora 
a. es 


SSSSSSSSSSSSSSSSSSSSSSS SSS SSSS585 
8 S c3 


a. Syrinx 

a. oaxacana 

a. tubiflor: 

a. 5 
e. brachysiphon 
e. nacalpulensis 
e. hypoleuca 


S. densiflora 


P riedelii 
R. rotundifolius 


'agrans 


T. jasminoides 


eayi 
pa. Perd 


Figure 4. Strict consensus of the most parsimonious trees generated by the molecular combined data set. Bootstrap values 


> 50% are indicated above the branches. Full taxon names are given in Appendix 


tion of Mandevilla proposed by Pichon (1948) but only 
partially corresponds to that of Woodson (1933). The 
main difference between the two classifications 
concerns the rank of Macrosiphonia and Telosiphonia. 


Woodson (1933) recognized Macrosiphonia as a dis- 


Macrosiphonia 


the synonymy 


tinct genus with two disjunct subgenera, subgenus 
Hemisphere and 
subgenus Telosiphonia in the Northern Hemisphere. 
Pichon (1948), in contrast, included Macrosiphonia in 


cited a set of 


576 Annals of the 
Missouri Botanical Garden 


100 A. ee 
plumeriiflora 
Me 99 97 Me 5 RE 
e. trifidus 
100 100 Me. minimus 
Me. roseus 


L gratissima 
i. mortoni 
. acouc . 
; ape eem Wood Pich 
F. velloziana 
c. longiflora 
c. velame 
c. martii 

a. anceps 

a. nerioides 
a. roe 


100 


1. kre 
a o ie 


f . dodsonii 
100 A anun 
69 a. sca 


a. subsagittat 1 


a. fun dormis 
a. harley1 
a. atr oviolacea 


1c 
a. . 
a. urophylla 
a. venulc 
a. coccinea 
a. illus 
a. aaa 
a. d e 
1. frag mans 
udit 
s a: myriophyllum 
a. tenuifolial 
a. tenuifolia? 
a. pycnantha 
a. cercophylla 
a. emarginata 
a. laxa 
a. pee 
a. glandulosa 
a. 5 0 a nii 


9] 


_ veraguasensis = 
100 a. 2 eraguasen Ps 


a. eo eee 
a. Oaxac 


a. See 

a. muton, 

a. karwi 

E € rachysiphon 


. hac 9 ens 
i. euca 


Aa. fo vis 


-densiflora 


Pa amerei 


W: Macrosiphonia subg. Macrosiphonia; P: Mandevilla sect. Megasiphon 
W: Mandevilla subg. Exothostemon; P: Mandevilla sect. Exothostemon 
W: Mandevilla sect. Laxae; P: Mandevilla section Orthocaulon 

W: Macrosiphonia subg. Telosiphonia; P: Mandevilla sect. Telosiphonia 
W: Mandevilla sect. Tenuifoliae 

W: Mandevilla sect. Montaneae 


W: Mandevilla sect. Tubiflorae 


E 
= 
[I 


W: Mandevilla sect. Torosae 


Fig Strict consensus of the ipsc AP. trees gener: dba op) the total evidence data set. Bootstrap values > 
50% are adi 'aled above the branches Mandevilla lade: Mesechites clade; F = Forsteronia clade. The six 


Volume 93, Number 4 
2006 


Simóes et al. 
Is Mandevilla Monophyletic? 


morphological characters to differentiate between 

oodson's subgenus Macrosiphonia and subgenus 
Telosiphonia (presence vs. absence of a pedicel, 
structure of staminal filaments, and pollen size) and 
recognized them as two different sections of Mande- 
villa, section Megasiphon and section Telosiphonia, 
respectively. Our results suggest that Telosiphonia and 
Macrosiphonia are not closely related to each other, 
although both are clearly nested within Mandevilla, 
and confirm the preliminary results from our previous 
2004). 

Most of the jubaseneno of Mandevilla 
proposed by Woodson (1933) are not monophyletic. 


study (Simóes et al., 
groups 


The two subgenera he proposed, subgenus Exothoste- 
mon and subgenus Mandevilla, correspond for the 
most part to the two major clades within Mandevilla 
identified in our analyses, Clades I II. r 
spectively (Fig. 5). To render Woodson's subgenera 


monophyletic, the following new classifications must 


and e- 


be made: (1) Macrosiphonia must be included in 
subgenus Exothostemon, (2) Telosiphonia must be 


included in subgenus Mandevilla, and (3) Mandevilla 
callista, M. funiformis, 
transferred from subgenus Mandevilla to subgenus 
Exothostemon. Of the 
Mandevilla proposed by Woodson (1933), only the 
smallest, section Tenuifoliae, containing two species, 
constitutes a monophyletic group (BS = 100%) in our 
All of the other sections are polyphyletic, with 
the 


and M. harleyi must be 


five sections of subgenus 


study. 
their constituent taxa scattered throughout 
Mandevilla clade (Fig. 5). The most extreme case of 
polyphyly is found in Woodson’s section Laxae, the 
largest of subgenus Mandevilla, which he character- 
ized by having infundibuliform corollas. In our study, 
the 24 species sampled from this section are scattered 
among all larger subclades of the Mandevilla clade 
ig. 5, Table 

With regard to the infrageneric ranks proposed by 
Pichon (1948), our results support the monophyly of 
two of his sections, namely, Megasiphon and Telosi- 


phonia. His other sections, Orthocaulon and Exothos- 


"i 
da 


temon, correspond to Woodson's subgenus Mandevilla 
and subgenus Exothostemon, respectively, and do not 
constitute monophyletic groups as indicated above. 
Despite their strongly supported monophyly, recogni- 
tion of 
untenable both taxonomically and morphologically, 


due to the considerable number of additional sections 


sections Megasiphon and Telosiphonia is 


without morphological synapomorphies that would 
Mandevilla. 


not 


need to be recognized in The same 


justification can be applied for recognizing 
Woodson’s section Tenuifoliae, despite its monophyly. 

After a detailed examination of herbarium vouchers 
and phototypes, we have concluded that Mandevilla 
and Quiotania colombiana 
Zarucchi are conspecific. As Q. 
only described species of the genus and M. ligustri- 
flora is nested within Clade IV with a strong bootstrap 


Woodson 


ligustriflora 
colombiana is the 


support (see Fig. 5), Quiotania cannot be recognized 
as a valid genus and should, therefore, be included in 
the synonymy of Mandevilla. The required nomencla- 
tural changes have been undertaken in a separate 


paper (Simoes et al, 2007). 


CLADE I 


Clade I contains representatives from three dispa- 
rate taxonomic groups of Woodson's (1933) classifi- 
cation. Of the 17 species included in this clade, the 
majority (11) belong to Woodson's subgenus Exothos- 
temon. Al 11 sampled species of Exothostemon in our 
study are within this clade. Of the six remaining taxa, 
three (Macrosiphonia longiflora (Desf.) Müll. Arg., M. 
martii Müll. Arg., and M. velame (A. St.-Hil.) Müll. 
Arg.) belong to Woodson's genus Macrosiphonia, and 
the other three (Mandevilla callista, 
and M. harleyi) fall under the circumscription of his 
subgenus Mandevilla. Clade I is characterized by one 
the calycine colleters 


M. funiformis, 


morphological synapomorphy: 
have an opposite arrangement in relation to the calyx 
lobes (character no. 15, Fig. 6; see Appendix 3). 
Subgenus Exothostemon forms a morphologically 
distinctive group within Mandevilla. Flower structure 
is quite homogeneous, of three 
character states considered by both Woodson (1933) 
and Pichon (1948) as diagnostic for the group: (1) leaf 
colleters distributed along the 


= 


wi the presence 


surface with many 
midrib on the adaxial surface; (2) opposite calycine 
colleters; and (3) corolla lower tube more or less 
gibbous or arcuate. Variation in vegetative characters 
and geographic distribution is, however, remarkable 
in the subgenus, and groups within Clade I can be 
discerned based on morphology. The first group, 
represented in our study by eight species (M. dodsonii 

H. Gentry, M. hirsuta (Rich.) K. Schum., M. 
M. leptophylla, M. rugellosa (Rich.) L. 
M. rugosa (Benth.) Woodson, M. 


krukovii, 


Allorge, scabra 


e 
subclades within the Mandevilla clade are indicated as 1, H, 
ligustriflora, which is conspecific with Quiotania Eun A 


id Pichon (1948) is illustrated in the two columns on the right of the cni 


— 


, IV, V, aña VI. The arrow indicates the position of Mandevilla 
A com n (1933) 


ac ch patte rn of f the columns i 18 48806 1 lo 


LI 


rison between the classifications of Woods 


its corresponding taxonomic rank in Woodson's (W) and Pichon's (P) classification. Full taxon names are given in Appendix 1. 


578 Annals of the 
Missouri Botanical Garden 


Mc. longiflora 
0 Mc. marti 
ae a 


I Hd VIO 


E > 
2 Ex 0 


Va. d 

Ma. scabra 

Ma. 1 05 8 

Ma. 1 
call 


m 


AAA 
Doo 
zii 
AI 
e 
om | 
=| 
= 
un 


en 
Ma. mor 1 


III HC VI 


a. 19 5 iophyllum 
a. tenuifolia l 

a. i 

à. pycnantha 

a. cercophylla 
a. emarginata 
a. laxa 

a. pentlandiana 
a. glandulosa 

a. jamesonil 

. tricolor 

a. ls 


a. ligust 
a. convolvulacea 


| 
TEARS 


AI adv IO 


EE. EA A A AA A A A a N 
a agi eat = ae unb ae LE a 


IA3QV IO | AJHAVIO 


7 5 


+ Unambiguous Ambiguous, ACCTRAN 
X Subsequent reversals | Ambiguous, DELTRAN 


Figure 6. One of the two most parsimonious total evidence trees showing optimized morphological character-state e hanges 
within the Mandevilla clade. Optimizations were identical in both trees. Diamonds show unambiguously optimized character- 


state changes diagnostic for clades IV and V. with subsequent reversals indicated by an X. Bars denote ambiguously optimized 
lire lerstate changes. Character-state changes using DELTRAN optimization are 1 d by solid bars; character-state 
changes using ACCTRAN pe e are indicated by dotted bars. All numbers above the symbols correspond to the 
character number as indicated in Appendix 3. The directions of character-state changes are reported below the symbols. 


Volume 93, Number 4 
2006 


Simoes et al. 579 


Is Mandevilla Monophyletic? 


(Hoffmanns. ex Roem. & Schult.) K. Schum., and M. 


subsagittata (Ruiz & Pav.) Woodson), is composed of 
taxa that show the most common morphological 
pattern in the subgenus: vines with terete stems and 
yellow flowers, often with a red center (white flowers 
in M. rugosa), that occur mainly in forests and their 
bordering zones throughout the Neotropics. The 
second group. represented in our study by three 


species (M. anceps, M. lancifolia, and M. nertoides), is 
composed of taxa that have a unique set of characters 
within the subgenus: they are shrubs or woody lianas 
with strongly angled to winged stems (tetragonal 


flowers of various colors that are 


[om 


cross section) ane 
found mainly in the open habitats of white sand 


f South 


Neither of these two groups is monophyletic, however. 


savannas and tepuis of northern America. 


All species from the first group form a strongly 


supported clade together with one species from the 


second group, M. lancifolia. The two remaining 
species of the second group, M. anceps and M. 
nerioides, form a strongly supported clade (BS = 


100%), but this clade’s 


species of Clade I is not resolved in our analysis. 


relationship to the remaining 


Therefore, no further conclusions on relationships and 


patterns of evolution within Exothostemon can be 


drawn at this time. Broader taxon sampling, especially 
including representatives from the poorly collected 


species of the winged-stem is needed to 


l 


group, 
address these questions. 

The inclusion of three species from Mandevilla 
subg. Mandevilla (M. callista, M. funiformis, and M. 
harleyi) Clade I 
character evolution within the genus, because they 
possess characteristics of both of Woodson's (1933) 


According to Woodson, M. funiformis has 


in is central to understanding 


subgenera. 
five calycine colleters in an opposite arrangement, as 
is characteristic for subgenus Exothostemon, but this 

f 
the midrib, a key character of subgenus Mandevilla. 
M. 


colleters are spread along the length of the midrib, as 


species also has leaf colleters restricted to the base o 


Conversely, in callista, Woodson noted that leaf 


is characteristic for taxa in subgenus Exothostemon. 
while calycine colleters form a continuous ring. 
Woodson (1933) recognized this “intermediate” status 
of M. callista and M. funiformis but justified their 
Mandevilla 


calycine 


the 
the 


subgenus based on 


— 


inclusion i 


presence of continuous colleters in 


former and leaf colleters restricted to the base of the 


examined vouchers of the 


We 


1 order to compare morphological 


midrib in the latter. 


three species i 


variation with the taxonomic descriptions provided by 
Woodson (1933) and Sales et al. (2006). We found 
that specimens of M. funiformis have leaf colleters 


restricted to the base of the midrib and calycine 


colleters in an opposite arrangement, confirming 
The same set of characters 


of M. Our 


however, do not agree 


Woodson's observations. 


was also seen in specimens harleyi. 


observations for M. callista, 
completely with Woodson's original description. We 
found that leaf colleters are spread along the entire 
midrib, but, in the specimens we studied, calycine 
colleters had the opposite arrangement typical for 
subgenus Exothostemon. 

The well-supported inclusion of representatives of 


Macrosiphonia in Clade | is surprising and somewhat 


unexpected from a morphological standpoint. In 
Macrosiphonia, the leef blade is covered by white 


woolly trichomes abaxially, the leaf colleters are 
restricted to the adaxial base of the midrib, the flowers 
l 
in a continuous ring (Woodson, 1933; Ezcurra et al., 
1992; Henrickson, 1996). The other species in Cl 


I, in contrast, have glabrous to tomentose (but never 


lack a pedicel, and the calycine colleters are arranged 


ade 


woolly) leaves with colleters spread along the midrib 


except in Mandevilla funiformis and M. harley, 


where the colleters are restricted 


pedicellate flowers, and calycine colleters with an 
opposite arrangement. Increased taxon sampling and 
that have 


Mandevilla 


palynology, floral ontogeny. 


additional studies focused on features 
scarcely been addressed previously in 
and Macrosiphonia (e.g.. 
and anatomy) could provide useful information to 
support relationships within this clade. 

The arrangement of leaf and calycine colleters are 


key 


tionships and morphological evolution in Mandevilla. 


characters to understanding phylogenetic rela- 


f the 
This 


character state is unique and has never been reported 


Colleters distributed along the entire length of 


midrib were observed in species of Clade l. 


Calycine 


in any other group within Apocynaceae. 

colleters with opposite arrangement were also ob- 
served only in species of Clade I within the 
Mandevilla clade, but the same state has been 


reported in other groups of Apocynaceae, as in the 
Neotropical genera of tribe Echiteae (e.g.. Thenardia 


HBK, Prestonia R. Br., and Temnadenia Miers) 


"haracter numbers and states are as follows: eaf colleter 


8. Le 
.15 l] 
5. 


Heter arrat 
Proportion between the a 


bi 


I 


1germ«c nt: 0, 


midrib of the leaf blade Calycine co 


2, truncate. 29 


appendages with the same size or 


pical appendages and main body 


osition: O, clustered at the base of the midrib; | 
alternate to continuous; J. 1 20. Anther b 
of the 


, spread along the 


ase: l, cordate: 


e head: 0, e shorter: 


digger than the main body. 30. Nectary number: 0, five; l, two 


Annals of the 
Missouri Botanical Garden 


(Pichon, 1950; Ezcurra et al. 1992; Simões & 
Kinoshita, 2002). Opposite calycine colleters were 


unambiguously reconstructed as the ancestral state of 


Clade I (character no. 15, Fig. 6), but two equally 
parsimonious hypotheses could explain the evolution 
of leaf colleters in this clade (character no. 8, Fig. 6). 
ACCTRAN the 


colleters midrib 


Using; optimization, 
the 


a synapomorphy for Clade I, with two subsequent 


distributed along represents 
reversals to colleters clustered at the leaf base in both 
the Mandevilla 
funiformis/M. harleyi subclade. DELTRAN optimiza- 


species of Macrosiphonia and in 


tion, in contrast, suggests three parallel origins of 
leaf colleters distributed along the midrib: in 


Mandevilla callista, i 


subclade, 


the M. anceps/M. 


and in the largest subclade of Clade 


nerioides 


Given the unique status of this feature in Apocyna- 


ceae and its occurrence only in species of Clade I, 


a single origin of this state seems more likely than 
three. parallel. changes character state, in. which 
case it would represent another synapomorphy for 
Clade 
our results. Future studies focusing on the morphology 
and ontogeny of leaf colleters in Mandevilla could 


help to clarify the evolution of this character in the 


No further conclusions can be drawn from 


genus. 


CLADE H 

This clade, which comprises Telosiphonia and the 
Mandevilla 
Pichon’s (1948) sections Orthocaulon and Telosipho- 
nia, and for the main part to Woodsor's (1933) 


subgenus Mandevilla and subgenus Telosiphonia. 


majority of species, corresponds te 


From a morphological standpoint, Clade I spans 
almost the entire spectrum of morphological variation 


Mandevilla, 


showy, lilac 


found in from subshrubs with large, 
pink infundibuliform flowers, as in 
M. sancta (Stadelm.) Woodson, to vines with small, 
inconspicuous, white tubular flowers, as in M. 
ligustriflora. This clade is also represented throughout 
the entire geographic range of Mandevilla, from the 
southwestern United States and Mexico to subtropical 
Argentina. 

All species from this clade share two morphological 
character states: leaf colleters restricted to the base of 
the midrib and calycine colleters with an arrangement 
that varies from alternate to continuous. These are, 
however, plesiomorphic states within the Mandevilla 
clade and therefore cannot be recognized as synapo- 
morphies of Clade II. Simões et al. (2004) showed that 
colleters restricted to the leaf base is one of the four 
morphological synapomorphies that characterize the 
tribe Mesechiteae, and calycine colleters with alter- 


nate to continuous arrangement are found in the two 


presence of 


other clades of Mesechiteae (Mesechites and Forstero- 


nia clades), as well as in some outgroup taxa. 


CLADE IH 


— 


This clade is primarily composed of species « 
Mandevilla occurring in forests, savannas, and campo 
rupestre formations of northeastern to southern Brazil, 
also reaching Paraguay and Argentina. Most of these 
(1933) section. Laxae, 
myriophyllum and M. 
tenuifolia, both ascribed to his section Tenuifoliae, 
and M. pycnantha (Steud. Woodson, 
attributed to section Montanae. With the exception of 
M. pycnantha, 


species belong to Woodson’s 


with the exception of M. 


ex A. DC.) 


all species in this clade share one 
morphological character state: the presence of only 
two nectaries alternate to the carpels. 

Other morphological e 'harac ‘ters, however, are more 
variable within this clade, in both vegetative and 
reproductive parts. Some species, such as Mandevilla 
pendula (Ule) Woodson and M. urophylla (Hook. f.) 
Woodson, are 


vines from the Atlantic Rainforest 11 


southwestern Brazil, but others, including M. illustris 
(Vell.) Woodson, M. pohliana (Stadelm.) A. H. Gentry, 
and M. spigeluflora (Stadelm.) Woodson, are small, 
unbranched subshrubs of savannas and campo 
rupestre formations from central and southern South 
America. Branched, woody shrubs are also common, 
is M. duartei Marker. and M. 
Arg.) Woodson, endemic to specific 
Their 


flowers are showy and variously colored and, in most 


with some species, such : 
venulosa (Müll. 


mountain formations. of southwestern Brazil. 


cases, have an infundibuliform corolla. Woodson 
(1933) used corolla shape as a diagnostic character 


and defined his entire section Laxae according to the 
shared occurrence of infundibuliform corollas among 
its members. Even though species with an infundib- 
uliform corolla comprise a strongly supported sub- 
clade (BS = 100%) within Clade IH, this character 
state clearly appears to have arisen independently 
multiple times within Mandevilla, 
taxonomic utility. 


undermining its 


The number of floral nectaries is an easily defined 
character, with no intermediate states. Most members 
HI, the Mandevilla 
pycnantha, are characterized by the presence of two 
nectaries in the flower (character no. 30, Fig. 6). This 


state could thus be considered as a synapomorphy for 


of Clade with exception of 


Clade HI, with a later reversal to five nectaries in M. 


pycnantha. However, an equally parsimonious re- 


‘onstruction would involve 


^ 


a swilch from five to two 
nectaries occurring independently twice: once in the 
clade composed by M. myriophyllum and M. tenuifo- 
lia, and again in the clade composed by the remaining 


species of Clade HI. Given that the occurrence of two 


Volume 93, Number 4 
2006 


Simoes et al. 
Is Mandevilla Monophyletic? 


1. —— 


Figure 7. Diagrammatic drawings of the anther base form and style head apical appendages in Mandevilla. —A. M. 
tenuifolia, anther base cordate. — eraguasensis, anther base truncate. —C. M. tenuifolia, short style head appendages. 
—D. M. syrinx, long style head nire wie Scale bar — 1 mm. 


nectaries is restricted to species of Clade III within 
Mandevilla and the relatively small number of 


that exhibit this character state in Apocynaceae, the 


axa 


first hypothesis, of a single switch from five to two 
nectaries, seems more likely than two parallel changes 


within Clade TIL Further studies focusing on the 
structure and development of floral nectaries in 
Apocynaceae, however, are needed to test these 


alternative hypotheses. 


CLADE IV 


This clade comprises a heterogeneous group of 15 
species of Mandevilla from Woodson's (1933) sections 
Laxae, Montaneae, Torosae, and Tubiflorae, plus the 
three sampled species of Telosiphonia, with members 
distributed mainly from Mexico and the southwestern 
United States to northern South America, but also 
reaching southern Brazil and Argentina. 

A morphological synapomorphy for Clade IV is the 
presence of elongate apical appendages of the style 
head that are the same size or longer than its main 
body (character no. 29, Fig. 6; Fig. 7). 
found the exception of two 


subclades: 


These are 


all species, with 


the one formed by the three sampled 


in 
species of Telosiphonia and the one comprising 
Mandevilla emarginata (Vell.) C M. laxa 
(Ruiz & Pav.) Woodson, and M. pentlandiana (A. DC 


Woodson. Two equally parsimonious reconstructions 


Ezcurra, 


— 


of ancestral states are possible for this character 
(Fig. 6). In one optimization, elongate apical appen- 
dages evolved in the ancestor of Clade IV and were 
independently lost in the two subclades mentioned 
above. In the alternative optimization, the evolution of 
elongate apical appendages in the ancestor of Clade V 


was followed by a single reversal to short appendages 


in the clade comprising M. emarginata, M. laxa, and 


M. pentlandiana. 


CLADE \ 

This clade is mainly composed of species from 
Woodson's (1933) Mandevilla section Montanae, but 
three species (M. glandulosa (Ruiz € Pav.) Woodson, 
M. laxa, and M. veraguasensis (Seem.) Hemsl.) were 
assigned to his section Laxae. Most are vines, with the 
exception of M. emarginata, an unbranched subshrub, 
and M. pentlandiana, which has both vine and shrub 
forms. Flowers in this clade are generally salverform 
or tubular, white to greenish, but M. veraguasensis, M. 
glandulosa, and M. laxa have showy, infundibuliform 
to campanulate corollas. One morphological feature. 
Mandevilla 
unambiguously reconstructed as a synapomorphy of 
no 


found nowhere else in the clade, was 


this clade: the anther base is truncate, with 


discernible auricles or protruding extensions (charac- 
20, Fig. 6; Fig. 7). In M. emarginata, M. laxa, 


and M. pentlandiana, 


ler no. 
auriculate anther bases can 
occasionally be found in some individuals, but in most 
cases the base is truncate. The presence of truncate 
anthers was used by Woodson (1933) to distinguish 
his section Montanae, although M. pycnantha has 
conspicuously auriculate anthers. Interestingly, our 
parsimony analyses dic not support the inclusion of 
M. pycnantha in Clade V, while they firmly placed M. 
glandulosa, M. laxa, and M. veraguasensis in it. These 
latter species, included by Woodson (1933) in section 
Laxae, are also characterized by truncate anthers. 


A distinctive aspect of Clade V is its geographical 


distribution. In contrast to Clade VI, which is 
restricted to a single region, two geographically 


Clade V. 


isjunet groups can be distinguished in 
disjunct groug be dist hed 


Annals of the 
Missouri Botanical Garden 


The majority of its species are found in the forests of 


Central America and northwestern South America. 
Three species (M. emarginata, M. laxa, and M. 
the 


forest in southern and southeastern Brazil. as well as 


pentlandiana), however, occur in Atlantic Rain- 
in more arid habitats from southern Bolivia and Peru 
with M. 


emarginata. also reaching the savannas of central 


to Paraguay, Uruguay, and Argentina, 
Brazil. The fragmentary, cireum-Amazonian distribu- 
tion found in this clade has also been reported in other 
1979, i 


Brunfelsia, Solanaceae) and might be related to the 


Neotropical plant groups (e.g., Plowman, 


climatic fluctuations of the Quaternary, as well as the 
geologic history of the Andes during the Pliocene/ 
Pleistocene. 

From the Paleocene to the Miocene, the continuous 
occurrence of everwet climates in South America is 
thought to have resulted in the spread of tropical rain 
forests across the continent, forming a continuous belt 
from the Atlantic to the Pacific coast (Morley, 2000). 
V general cooling and/or overall reduction of pre- 
cipitation on the continent during the Late Miocene 
the 
forested areas and expansion of savannas (Prance. 
1982; Morley, 2000). With the simultaneous climatic 
fluctuations and the major uplift of the northern Andes 
(van der Hammen, 1974; Flenley, 1979; Morley, 


2000) during the Late Tertiary and Early Quaternary 


and Early Pliocene resulted in retraction of 


and the subsequent expansion of rain forest through 
the Amazon basin, the northwestern part of South 
the 


southern parts of the continent. Thus the geologic 


America became isolated from central and 
and climatic history of South America during the 


Cenozoic could explain. the close phylogenetic 
relationships among greatly disjunct species observed 


in Clade V (see Fie. 5). 
O 


CLADE VI 


This clade is composed of species from Woodson's 
Mandevilla sections Laxae (M. convolvulacea (A. DC. 
Hemsl.. M. oaxacana (A. DC.) Hemsl.), Torosae (M. 
Joliosa (Müll. Arg.) Hemsl., M. karwinskii (Müll. Arg. 
Hemsl.), and Tubiflorae (M. syrinx Woodson, M. 
tubiflora (M. Martens & Galeotti) Woodson), plus 


genus Telosiphonia, all of which occur in deserts and 


— — 


dry forests of Mexico and the southwestern United 
States. Morphological traits in this clade are extremely 
variable, especially those related to flower structure. 
Mandevilla syrinx and M. tubiflora 


flowered inflorescences bearing small, tubular. white 


have many- 
flowers, whereas M. convolvulacea and M. oaxacana 
have few-flowered inflorescences with showy, yellow, 
infundibuliform flowers. The most striking floral 


morphology of this clade is found in Telosiphonia, 


characterized by long, narrowly tubular, white flowers 
forming l- to few-flowered inflorescences. 

Although species of Woodson’s (1933) sections 
Torosae and Tubiflorae are restricted to this clade. 
they do not form monophyletic groups, and thus their 
continued recognition might be questionable. The 
distinction between these two sections is based on 
nectary height: in section Torosae, the nectaries are 
shorter than the ovary, whereas in section Tubiflorae 
they are the same size or taller than the ovary. We 
observed that species of section Torosae always have 
nectaries taller than the ovary, but the same condition 
occurs in three other species from this clade, all of 
which belong to different sections sensu Woodson 
(1933): Mandevilla foliosa (sect. Torosae). and M. 
convolvulacea and M. oaxacana (both in sect. Laxae). 
Nectaries shorter than the ovary are found in al 
Telosiphonia species and in M. karwinskit, which 
together form a strongly supported subclade (BS = 
100%). 

The sister relationship between Mandevilla kar- 
winskit and species of Telosiphonia, which has never 
been proposed before, is congruent with their geo- 
graphic distribution and habitat preferences. Both 
laxa are. rhizomatous shrubs occurring sympatrically 
in the deserts of Mexico and the southwestern United 
States. Apart from their short nectaries, morphological 
traits are quite different between M. karwinskii and 
species of Telosiphonia, especially leaf indument, 
flower size, and style head structure. 

The striking similarity in morphology between 
species of Telosiphonia and Macrosiphonia is the 
the 
Vandevilla clade. The two genera, each of which 


most extreme example of parallel evolution in 


comprises a well-supported subelade nested within 
the Mandevilla clade (Macrosiphonia in Clade | and 
Telosiphonia in Clade VI, see Fig. 5), occur in 
disjunct geographic areas that roughly coincide with 
the extreme northern and southern range of Mande- 
villa. Macrosiphonia is found in the savannas of 
southern South America in arid, usually sandy cerrado 
and campo rupestre vegetation from southern Bolivia 
and Peru to central Brazil, Paraguay, Uruguay, and 
the 


arid zones of Mexico and the southwestern United 


Argentina, whereas Telosiphonia is restricted. to 


States. Despite their geographic disjunction, the two 
genera share a suite of morphological characteristics. 
Both 


rhizomatous, 


are erect. shrubs or subshrubs. sometimes 


with a well-developed underground 
storage system and leaves covered by a dense, wooly 
indument on the abaxial surface. The most remarkable 


similarities; however, are related to flower structure. 


n both genera, flowers are white and tubular, with 
some of the longest corolla tubes in Apocynaceae. 


reaching up to 17 em in Macrosiphonia longiflora, 


Volume 93, Number 4 
2006 


Simoes et al. 583 


Is Mandevilla Monophyletic? 


and are only fully open at dusk, when they produce 
a distinctive scent, suggesting pollination by hawk- 
moths. 

The apparent parallelism in vegetative characters 


observed in these two groups could be explained as an 


> 


adaptation to similar environmental conditions. 
shrubby, erect habit, the presence of a dense indu- 
ment covering both vegetative and reproductive 


organs, and well-developed | underground storage 
organs such as tubers and xylopods are common trails 
of plants of open, seasonally dry habitats. (Rizzini. 


1997: Dallman, 1998). On the other hand, parallelism 


in floral structure is more likely driven by pollinator 


— 


preferences. The distinctive features shared by 
Macrosiphonia and Telosiphonia are typical of the 
(hawkmoth) pollination syndrome 
reported by many authors (e.g., Vogel, 1954; Faegri 
& van der Pijl, 1966; Baker & Hurd, 1968; Endress, 
1994; 


flowers of various Telosiphonia and Macrosiphonia 


sphingophilous 


Galetto, 1997). Reports of hawkmoths visiting 
species are congruent with the hypothesis of hawk- 
moth pollination. For example, two hawkmoth species, 
have 


Manduca sexta L. and Hyles lineata Fabricius, 


been observed visiting Telosiphonia nacapulensis 
Felger & Henr. and T. brachysiphon (Torr.) Henr. in 


the Sonoran Desert of southern Arizona (R. 
2004). In Cordoba Province and in the El 


Raguso, 


pers. comm., 


Palmar National Park, Entre Ríos Province, Argen- 
tina, Manduca sexta, M. rustica Fabricius, and Agrius 


cingulata Fabricius have been found carrying pollen 
St.-Hil) K. 
attached. only to the very tips of their proboscis (A. 
2004). Given the remarkable 


length of the floral tube of this plant (ca. 105 mm). it 


of Macrosiphonia petraea (A. Schum 


Cocucel, pers. comm., 
is reasonable to expect that only insects with a very 
long proboscis could reach the nectar (Marcela Moré, 


in prep.). 
CONCLUSIONS 


The phylogenetic results presented here show that 
Mandevilla, as circumscribed by Pichon (1948), is 
monophyletic, but Woodson's (1933) circumscription 
of the genus is paraphyletic. Quiotania, Macrosipho- 
nia, and Telosiphonia are nested within Mandevilla 
and therefore should be included in its synonymy. 
Representatives from Macrosiphonia and Telosiphonia 
form distinct clades embedded within Mandevilla, and 
their striking morphological similarities may have 
evolved in parallel, possibly as a result of similar 
selective pressures driven by colonization of open, dry 
habitats and hawkmoth pollination. 

All infrageneric taxa within Mandevilla proposed 
by both Woodson (1933) and Pichon (1948) were 
be paraphyletic or with the 


found polyphyletic, 


exception of Woodson's section Tenuifoliae and 


Pichon’s sections Megasiphon and Telosiphonia. 


Recognition of these three sections is. however, 
untenable for the moment, as this would require 


recognizing additional sections that lack clear mor- 
phological synapomorphies. Six major clades were 
recognized within Mandevilla in our study, although 
only three have unambiguous morphological synapo- 
morphies. It is hoped that more detailed morpholog- 
ical studies in Mandevilla could uncover additional 
characters that might prove useful for delimitation 
Until 


we think it is most prudent to withhold 


within this group. such evidence becomes 


available 
from erecting a new classification. of 


Wandevilla. 


intrageneric 


Literature Cited 
Baker, H. G. & P. T Hurd. Annual 


Rev. Entomol. 13: 385-414. 


1968. Intrafloral ecology. 


Baum, D. A. & l Po 1998. Biogeography and floral 
evolution of a (Adansonia, ) pons üs 
pee Hon multiple data sets. Syst. Biol. 47: 181—207. 

Civeyrel, v N. Rowe. 2001. Phylogenetic re 1 of 


Sec Ead based on the plastid gene matK, morphol- 
ogy, and biomechanics. Ann. Missouri Bot. Gard. 


583-60. 


Dallman, P. R. 1998. Plant Life in the World's Mediterra- 
nean Climates: The Mediterranean Basin, South Africa, 
Australia; Chile, and California. Oxford Univ. Press, 
Oxford. 

Endress, M. E., B. Sennblad, S. Nilss . Civey 


jj, M. W. 
1996. A 


and some 


re 

Chase. S. Huysmans, E. G n & eN Bremer. 

phylogenetic analysis of Apocynaceae s. str. 

related taxa in Ge nijanales: ij multidisciplinary approach. 

Opera Bot. PET 7: 59-102. 

. Van der ox s Nils I. Civeyrel, 
a Se nnb lac 


= 
me 
gi 


2007. A phylogenetic analysis of Alyxie 


trnL intron, trnL-K s 


> (Apocynaceae A 
based on rbcL, mat K, acer sequences 
and cd val characters. Ann. Missouri Bot. Card. 


9 4 press). 


Sa 


1994. 
Tropical Flows TS. a Univ. Press, Can 


E 1 P. Diversity and Evolutionary Biology o 
ridge. 
Ezcurra, C., M. E. Endress & / . Lee venere 1992. 


Flor dal 1 ruay: Apocynaceae, Vol 1 . Conservatoire 
| 


et Jardin botaniques de la Ville de Genès ve, Genève, and 

Missouri Botanical Garden, St. Louis. 

1966. The Principles of 
Pollination Ecology. Pergamon Press, Oxford. 

Felsenstein, J. 1985 
approach using g the bootstrap. Evolution 39: 783— 


"aegri, & L. van der Pijl. 
. Confidence limits on phylogenies: An 
79 

Flenley, J. R . The y 2 Rainforest: A Geological 
History. Butte ae ondon. 

Galetto, L. 1997. Flower 
three Arg cay pe maceae. Flora 192: | 

Hamilton, M. B. 
of 11 intergenic regions with intraspecific 
tion. Molec. Ecol. 8: 513-525. 

Henrickson, J. 


Generic 
95 


“structure and nectar c gmposition in 
192: 197-201. 


). Four primer pairs for the amplification 
varia- 


1996. Studies in Macrosiphonia (Apocyna- 


ceae): recognition of Telosiphonia. Aliso 14: 


179— 


584 


Annals of the 
Missouri Botanical Garden 


Judd, W. S., R. Sanders & M. J. 1994., 


Angiosperm family pairs: Preliminary phylogenetic anal- 


Donoghue. 


ysis. Harvard Pap. Bot. 5: 1—51. 

Lee suwenberg, A. J. M. 1 997. Series of revisions of 
Apocynaceae XLIV. Craspidospermum Boj. ex A. DC 
Gonioma E , Mascarenhasia / C., Petchia Livera, 


Mey 

; E [RU Sm Baill. Wageningen 

2 1-124. 

I. M M. E. Endress & J. K. 

e Phylogeny of Apocynoideae (Apocyna- 
ceae) and the APSA clade. Ann. Missouri Bot. Gard. 94 (in 

ss). 
Main, D. & W. Maddison. 2000. MacClade 4: 


phylogeny and character evolution, Sinauer Associates, 


q 
— 


Analysis 


9 1 rland, iva husetts 
Morales, J. F. 


. Una nueva especie de Tintinnabularia 
(/ oe ynace Me (n 6: 392—394. 
997. 


1997. / nopsis of the genus Allomarkgrafia 
( cape ynaceae). itn 49: 337 
1998. A synopsis of dies genus Mandevilla 


ae ynace aci in Mexico and Central jm rica. Brittonia 
50: 2 1550 
Morley, R. J. 2000. ed ier E volution of Tropical Rain 
e is Wiley & Sons, Chichester. 
Nixon, K. Y, M. 9 5 nter. 1996. 
analysis. R aida s 12(3): 221-241. 
B.. M. Liden & D. 


rps16 intron phy ogen, ny A ES tribe S 


On simultaneous 


I 


1997. 
sileneae (Carophylla- 


Oxelman, Berglund. Chloroplast 


ceae). P s st. Evol. 903—410. 
Pichon, M. 1948 T 1 'alion des Apocynacées: X. Gene 
Manilla Bull. Mus. Hist. Nat. (Paris), sér. 20: 


E Classification des P add XXV, 
Échitidses: Mém. Mus. Natl. Hist. Nat., sér. B, Bot. I: 


143. 


Plowman, T. 


)19. The genus 1 d A pue ctus of the 
p. 475-491 in J. C. Hawkes 


axonomy and biogeography. | 
A. 


ester & A. D. Skelding (editors), The Biology and 


‘axonomy of Solanaceae. innean Socie ty Symposium 


Series. Ac ademic " ress 
& V. 


. New Yo 
. Albert. ii 


e within ee vnaceae s.l. 


Potgieter, "Phyloge netic relation- 


ased on tral. intron and 
rnL-F spacer se fae nees and e characters. Ann 

Missouri Bot. Gard. 88: 523-549 
Prance, G. T. 1982. yi 1 o, raphic 
Pleistocene climate changes in the Neo 


review af the 


evidences for 


8 Ann. Missouri Bot. Gard. 69: men. 
TE G. Chase, P. Goldblatt, P. Rudall, ". Fay, 
Cox B Lejeune & T. Souza-Chies. 2001. ve «cular 
1 DNA 


systematics of Iridaceae: Evidence from four plastic 
Amer. E. Bot. 2087. 

19 Tratado de Fitogeografia do Brasil: 
Ámbito 


a 
N 
N 


Aspe clos p gic os, Soc ¡ológic os e Florísticos. 


Cultural liie Rio de Janeiro. 
Sales, M. F., shita & A. O. Simões. 2006. diam new 
species of Mondes Lindley (Apocynaceae, Ápocynol- 
Novon 16: 112-128. 


deae) e Brazil. 
& B 1996. 
familial cios of Apocynaceae and Asclepiadaceae 
evaluated with Thel data. Pl. Syst. Evol. 202: 153-175. 
& 20 102 


Classification of Apocynaceae s.l. 


Sennblad, Bremer. The familial and sub- 


according to a new approach combining Linnaean and 


g 
phylogenetic taxonomy. Syst. Biol. 51: 389—109. 


— — —. M. E. Endress & B. Bremer. 1998. dera de 15 
molecular yon in phylogenetic frate 


W d ae a Apocynaceae) revisited. nini T. ES 17 
1143-115 

Sheahan, al E & M. W. Chase. 2000. Phylogenetic 
relationships within Zygophyllaceae based on DNA 


sequences of three plastid regions with special emphasis 
on Zygophyllaceae. Syst. Bot. 71-384. 
Sidiyasa, K. 1998 
of Alstoni 
Simmons, M. 


. Taxonomy, Kin ny, and us ani n 


[o 


6 Apoc ip^ eae). Blumea, 2 I]: 
. & H. Ochoterena. 2000. 

in sequence A : phylogenetic analyses. Syst 

309-38 1. 

Simoes, X. O. & I 
sl 


Gaps as e ‘ters 
Biol. 49: 


„S. Kinoshita. 2002. The A 
Minas Gerais, 


pocynaceae s. 


he Carrancas region, Brazil. 
Darwiniana 40: add 

. Endress, T. van der Niet, L. 
2004. Tribal 
Mesechiteae (Api 


5. Kinoshita & E. 
Goan. and intergeneric. TE of 
cynaceae, vi 

from three none coding 1 5 DNA regions and morphol- 
ogy. Amer. J. Bot. 91: 1409-1418 

>. Kinoshita & M. E. 


and 


Endress. 


200 
Mandev illa 


L 7. New 
combinations synonyms in Lindley 
(Apocynaceae). Novon de pel 

: Phylogenetic analysis using 
parsimony, Vers. 4.0b. A r Associates, Sunde 1 
1991. Structural, 


aspects of the colleter. Ann. Bot. (O 


functional and phylogenetic 
ma 1) 68: 287-305. 


Thompson, J. D., G. D. Higgins & T. J. Gibson. 1994. Clustal 
W: roving the sensitivity n progresivo multiple 


ce 
gap ls nalties s weight matrix 11 Nucl. 

Acids Res 467368 
van der lad: qn 


i Y. M. | PETAN S. Nilsson & A. 
lee rshei sim. 2001. Pollen LM and phylogeny of the 
9-19 


Hammen, T. 1974. 15 Pleistocene changes of 
vegetation and climate in tropical South America. J. 
Biogeogr. 1: 3-2 


: >. 

Vogel, S. 1954. Blütenbiologische Typen als Element der 
F Botanische Studien. I. Gustav-Fisher 
Verlag, Jena. 

Whitten, W. M., N. H. 


Subtribal 


Willians & M. W. 2000. 


relationships of xillar 


Chase. 

and generic rie 

(Orchidaceae) with emphasis on Sani inae: c Combined 
-]85€ 


Amer. J. Bot. 87: 1842 


Combining x 


molecular evidence. 
Wiens, J. J. 1998. 
phylogenetic stories. Syst. Biol. 47: 
W ili ams, J. K. 2004. 
(i Apoc ynaceae: Apoc 
based on a morphological cladistic ee Sida 21: 
117-13 E 


Woodson, E. Jr. 


with 1 


508-581. 
Polyphyly of the genus Kehites 


sets 


ynoideae:  Echileae): Evidence 


1933. Studies in the 

The (uw rican genera of Echitoideae. 

yard. 22: 153-3 

Young, N. D. & J. Healy. 2003. Gapcoc 
of f inde! characters in phylogenetic analysi C 


Apocynaceae IV. 
Ann. Missouri Bolt. 


— 


er automate E p" use 
B. 


ioinf. 4: 6. 

Zarucchi, L. 
ynaceae-; Avon pairs from Northern Colombia. Novon 1: 
Be 


36. 


Quiotania: A new genus of Apoc 


Volume 93, Number 4 
2006 


Simoes et al. 


Is Mandevilla Monophyletic? 


585 


Appendix l. Vo 


selected for morphol uen analyses are indicated by an asterisk ( 


icher information and Genbank accession numbers for the taxa used in this paper. 
*j 


Additional vouchers 


maa trn Gee 


rpllo rps 16 Intergenic trn K matK 
Species Voucher/citation Intron Intron spacer Intron Gene 
Allomarkgrafia Costa Rica, Endress 97-06 AY597546 AY597580 AY597614 0522660 50522591 
brenesiana Woodson (Z) 
Allomarkgrafia Costa Rica, Morales 9338 DQ522730 50522812 50522771 50522661 50522592 
plumeriiflora NB) 
Woodson 
Beaumontia Cultivated, Bot. Gard. AY597547 AY597581 AY597615 550522662 Z98174 
grandiflora Wall. nich, G. Gerlach 5/06 
(M D: *[ndia, Feb. 1909, 
coll. Native collector s.n. (Z) 
„ fragrans Cultivated, Queensland, AY597548 AY597582 AY597616 50522663 DQ522593 
on) Alston ustralia, Forster 2009 
(BRD: * India, Ridsale 
109 (Z) 
Forsteronia acouci French Guiana, Prévost AY597549 AY597583 AY597617 50522664 DQ522594 
(Aubl.) A. DC 3720 (CAY); * Peru, 
Revilla 291 (Z): 
* Venezuela, Breteler 
5029 (Z) 
c „ Brazil. n 02/108 DQ522731 50522813 50522772 | DQ522665 DQ522595 
Müll. / (UE 
D velloziana Brazil, m 343 (UEC) AY597550 AY597584 4597618 50522666 DQ522596 
.) Woodson 
1 Brazil. Bos Rodrigues AY597551 AY597585 AY597619 50522667 DQ522597 
longiflora (Desf.) JEC); * Brazil, 
Müll. Arg Simões 47, 859, 930 (UEC) 
Macrosiphonia martii M Simões 1245 (UEC); AY597552 AY597586 AY597620 50522668 50522598 
Müll. Ar zl arte 2445 
115 ‘i * Brazil, Simoes 
1205, 1206 (UEC) 
Macrosiphonia Brazil, Simões 1199 (UEC); 50522732 50522814 0522773 50522660 50522590 
me (A. St.-Hil.) Brazil, Leitao-Filho 
Müll. Arg. 15307 (UEC); * 7000 
Kinoshita 2000/67 (UE 
Mandevilla anceps Venezuela, Huber & 1 15 50522733 50522815 50522774 0522670 0522600 
Woodson 5793 (Z) 
Mandevilla atroviolacea Brazil, Meireles 1290 (UEC) DQ522734 50522816 DQ522775 | DQ522671 DQ522601 
(Stadelm.) Woodson 
Mandevilla callista Ec tadon Webster & Castro DQ522735 DQ522817 50522776 50522672 DQ522602 
Woodson 9 (Z); * Ecuador, 
i 2874 (AAU); 
1 po 
es 30842 (Z) 
Mandevilla cercophylla Ecuador, n. 420 (7) DQ522736  DQ522818  DQ522777 0522673 DQ5220603 
Woodson 
Mandevilla coccinea Brazil, Flores 452 (UEC) DQ522737 0522810  DQ522778 50522674 DQ522604 
Arn. 
n on 
Thesei convolvulacea Mexico, Alvarado 162 DQ522738  DQ522820 9005227709 0522675 DQ522605 
(A. DC.) Hemsl. (MEXU) 
ie dp dodsonii Ecuador, Fallen 875 (Z) DQ522739 DQ522821 DQ522780 DQ522676 DQ522606 
A. H. Gentry 
Mar duartei Brazil, Simões 1281 (UEC) 0522740 DQ522822 DQ522781 DQ522677 DQ522607 


Markgr 


586 


Annals of the 
Missouri Botanical Garden 


Appendix . 


trn S -trnG 


UUC 


rplló rps16 Intergenic trn K math 
Species Intron Intron spacer Intron Gene 
Mandevilla emarginata DQ522741 50522823 50522782 505220678 00522008 
(Vell.) C. E 
Mandevilla foliosa 5052272 DQ522824 DQ522783 DQ522679 DQ522609 
(Müll. Arg.) Hemsl. 
Mandevilla fragrans 50522743 50522825 DQ522784 — 50522630 DQ522610 


(Stadelm.) Woodson 


Mandevilla di din 


(Vell K.S 


Mandevilla glandulosa 
5 K 


(Ruiz & 
Woodson 


Mandevilla harleyi 
M. F. pues d Kin-Gouv. 
0. Si 
17 ias 
(Rich) K. Schum. 
Mandevilla illustris 
(Vell.) Woodson 


Mandevilla jamesonii 


Woodson 


Mandevilla karwinskii 
(Müll. Arg.) Hemsl. 


Mandevilla krukovii 


M oodson 


Mandevilla lancifolia 


Woodson 


Mandevilla laxa 


(Ruiz & Pav. 


Woodson 


Mandevilla Ea 
(A. DC.) K. 5 
Mandevilla 1 


M oods son 


Mandevilla martiana 


Stadelm 


Mandevilla 


moricandiana 

(A. DC.) W 
Mandevilla myriophyllum 

(Taub.) Woodson 


Mandevilla nerioides 


Woodson 


Mandevilla oaxacana 
(A. DC.) Hemsl. 


E 


S 
Acl 
t 


t2 


DQ0522744 


DQ522746 


DQ522747 


DQ522748 
AY597553 
0522749 
DQ522750 


DQ522751 


DQ522752 
AY597554 


DQ522753 


DQ522754 


1Y597555 


DQ522755 


50522750 


50522820 


AY597593 


DQ522828 


DQ522829 


DQ522834 


AY 596586 


DQ522835 


DQ522836 


AY 596589 


DQ522785 


x 


DQ522786 


AY597027 


DQ52278 


DQ522788 


DQ522789 
AY597021 
DQ522790 
DQ522791 


DQ522792 


DQ522793 


AY597022 


— 

=~ 
A 

A 


DQ522795 


AY597023 
DQ522796 


DQ522797 


DQ522681 


0522082 


0522683 


0522684 


0522685 


10522686 


50522087 


— 50522688 


DQ522689 


DQ522690 


1000522691 


D 0522692 


DQ522693 


DQ522694 


DQ522011 


DO522612 


50522013 


DQ522614 


DQ522615 


DO522616 


DQ522617 


DQ522618 


DQ522019 


DQ522620 


DQ522621 


DQ522622 


DQ522623 


1522024 


Volume 93, Number 4 Simoes et al. 587 
2006 Is Mandevilla Monophyletic? 
Appendix l. Continued. 
trn SC -trnG € 
rplló rps LO Intergenic iR matK 
Species Voucher/citation Intron Intron Intron Gene 
Mandevilla pendula Brazil, Ribeiro 2520 (UEC) 50522757 DQ522839 50522798 50522605 DQ522625 
lle) Woodson 
uc pentlandiana dt Simões 1272 (UEC): 00522758 DQ522840 DQ522799 0522606 DQ522626 
C.) Woodson Brazil, Silva 1081 
(UEC); * Brazil, 
Lewinsohn 15901 (UEC) 
Mandevilla pohliana Brazil, Feres 98/49 (UEC) 50722759  DQ522841 50522800 0522697 50522627 
(Stadelm.) A. H 
Gentry 
Mandevilla pyenantha Brazil. Yamamoto 02/107 AY597556 AY596580 AY597625 DQ522698 DQ522628 
(Steud. ex A. DC.) (UEC) 
Woodson 
md de pus Seena th Cana Prévost AY597561 AY597595 AY597629 50522690 DQ522629 
(Ricl . Allor DAY): * Surinam, 
Lindeman 1976 (Z) 
Mandevilla rugosa Brazil, Simões 1204 (UEC) AY597557 4597591 4597625 50522700 50522630 
(Benth.) Woodson 
Mandevilla sancta Brazil, Simóes 1060 (UEC) DQ522700 50522842  DQ522801 DQ522701 DQ522631 
(Stadelm.) Woodson 
Mandevilla scabra Brazil, Simóes 1126 (UEC) AY597558 AY597592 AY597626 | DQ522702 50522632 
7 manns. ex Roem. 
Schult.) K. Schum. 
muse sellowii Brazil, Ribeiro 2522 (UEC) 50522761 50522843 50522802 DQ522703 DQ522633 
(Müll. Arg.) Woodson 
Mandevilla spigeliflora Brazil, Gomes 513 (UEC) DQ522762 = DQ522844  DQ522803 50522704 50522634 
(Stadelm.) Woodson 
Mandevilla splendens Brazil, Simées 1268 (UEC) AY597560 AY597594 AY597628 DQ522705 DQ522635 
(Hook. f.) Woodson 
Mandevilla subsagittata Mexico, Alvarado 288 D0522763 50522845  DQ522804 550522706 DQ522636 
(Ruiz & Pav.) SXU) 
Woodson 
Mandevilla syrinx Mexico, Calzada 21305 DQ522764  DQ522846  DQ522805 00522707 DQ522637 
Woods (MEXU) 
Hard 18 Brazil, Simóes 1171 (UEC) AY597502 AY597596 AY597630 10522708 DQ522638 
. Mikan) 
W 8 1 acc. J 
Mandevilla tenuifolia Brazil, Kinoshita & AY597563 AY597597  AY597031 DQ522709 DQ522639 
(J. C. Mikan) Matsumoto 00/609 
Woodson, acc. 2 (UEC) 
Mandevilla tricolor Ecuador, Jorgensen 1464 DQ522765 50522847  DQ522806 
V Z 
Mandevilla tubiflora Mexico, Alvarado 106 DQ522766 DQ522848  DQ522807 50522710 DQ522640 
M. Martens & (MEXU) 
tti) Woodson 
Mandevilla urophylla ie M. P. Quast 6 DQ522767  DQ522849 0522808 50522711 DQ522641 
(Hook. f.) Woodson (UEC) 
Mandevilla venulosa "€ Simões 1107 AY597564 AY597598 AY597632 . DQ522712 DQ522642 
Müll. Arg.) Woodson (UEC) 
Mandevilla veraguasensis Costa hioa Endress AY597505 AY597599 AY597633 50522713 DQ522643 
(Seem.) Hemsl. 97-76 (Z) 
Mesechites mansoanus Brazil, [ovd 1087 AY597567 AY597601 AY597635 0522714 DQ522644. 
(A. DC.) Woodson (UEC) 
Mesechites minimus Cuba, Feb. 2001, AY597568 AY597602 AY597636 | DQ522715 DQ522645 


age & P. Wilson) 


Woo ods son 


Nilsson s.n. (Z 


588 


Annals of the 
Missouri Botanical Garden 


Appendix J. Continued. 


trnS C rn Gee’ 


rplló rpsl6 Intergenic trn K math 
Species Voucher/citation Intron Intron spacer Intron Gene 
Mesechites roseus Cuba, Feb. 2001, AY597569 AY597603 4597637 50522710 DQ522646 
(A. DC.) Miers Nilsson s.n. (Z) 
Mesechites trifidus Ecuador, Liede & Meve DQ522768 50522850 50522800 50522717 DQ522647 
(Jacq.) Müll. Arg. 5471 (UBT) 
Odontadenia lutea Seal Kinoshita 2002/56 AY597570 AY597604 AY597638 50522718 00522048 
(Vell.) Markgr. (UEC) 
Pachypodium geayi Cultivated, Bot. Gard. AY597571 AY597605 AY597640 550522719 DQ522649 
Costantin & Bois Chevreloup, Lieberherr 
s.n. (unvouchered) 
Pachypodium lamerei Cultivated, Zürich Bot. AY597572 AY597606 4597639 0522720 DQ522650 
Drake Gart., Simões 1333 (Z) 
Prestonia riedelii Brazil, Simões 1274 (UEC) AY597573 AY597607 AY597641 DQ522721 DQ522651 
(Müll. Arg.) Marker. 
Rhodocalyx rotundifolius Brazil, Kinoshita 2000/66 AY597574 AY597608 AY597642 50522722 00522652 
Müll. Arg. (UEC) 
Secondatia densiflora Brazil, Simões 1218 (UEC) AY597575 AY597609 AY597643 50522723 50522653 
A. DC. 
Telosiphonia SAn Jen nO 00-185 AY597576 AY597610 AY597644 50522724 DQ522654 
brachysiphon (Torr.) E PUC 
Henr. W Nc 25068 
(TEX) 
Telosiphonia hypoleuca Mexico, Reina 2000-362 AY597579 AY597611 AY597645 550522725 00522655 
(Benth.) Henr. (Z); * Mexico, 
Richardson 1526 (TEX) 
Telosiphonia J.S.A., Arizona, July 50522769  DQ522851  DQ522810 50522720 DQ522656 
nacalpulensis 2000, Van Devender 
Felger € Henr. s.n. (Z) 
Tintinnabularia Mexico, Ventura 107 50522770 50522852  DQ52281l 50522727 50522657 
1 (ENCB) 
J. F. Morales 
1 1 Mexico, Breedlove 34900 AY597578 AY597612 AY597646 550522728 DQ522658 
mortonii Woodson (TEX) 
Trachelospermum Cultivated, Zürich Bot. AY597577 AY597613 AY597647 DQ522729 DQ522659 


— 


Lindl.) 


jasminoides 
I 


em. 


Gard., Simões 1334 (Z) 


Volume 93, Number 4 Simoes et al. 589 
Is Mandevilla Monophyletic? 


Appendix 2. Morphological matrix. ? signifies missing data. Polymorphic states are indicated by numbers in brackets. 


Characters and coding for character states as in Appendix 3. 


Taxon Character states for characters 1-32 
Allomarkgrafia brenesiana 10000111000001000001 101032100010 
Allomarkgrafia plumeriiflora 100001 1 1000001000001 101032100010 
Beaumontia grandiflora 10010100000101000000102021000001 
Chonemorpha fragrans 10020100000001000000101021000000 
Forsteronia acouci 100001110100010300111011{23}2200001 
Forsteronia ina 100001110100010300111011[23]2200001 
Forsteronia velloziana 100001110000010300111010(23]2200001 
Mac na 1 5 20000111101011010001101032200000 
Macrosiphonia martii 200001 11101011010001 101032290000 
Macrosiphonia velame 2000011110101 101000 1101032200000 
Mandevilla anceps 20100112001001100001101132200000 
Mandevilla pum 10000111001001000001101032200100 
Mandevilla callista 10000112001101100101101132200000 
Mandevilla cercophylla 10000111011001010002101032201000 
Mandevilla coccinea 200001 1 1001001010001 101032200100 
Mandevilla convolvulacea 100001 11001001000001 1010322901011 
Mandevilla dodsonii 100001 12001101100101 101132200000 

Mandevilla duartei 200001 11001001000001 101032290100 
Mandevilla emarginata 200001 11001 10102000{12} 101032200000 
Mandevilla foliosa 200001 11001001010001 101032201010 
Mandevilla fragrans 100001 1 1001001000001 1010322900100 
Mandevilla funiformis 10000111001001100101101132200000 
Mandevilla glandulosa 10000111001001000002101032201000 
Mandevilla harleyi 20000111001001100101101132200000 
Mandevilla hirsuta 10000111001101100101101132200000 
Mandevilla illustris 2002011 1001001000001 101032200100 

Mandevilla jamesonii 1000011 D E 
Mandevilla karwinskti 20000111001001 20 
5 krukovtt 10000112001 101 me 101 132200000 

la isis E (12]0100(12]12001001100001101032200000 

Mos A ^p B lax 1002111100100100000[(12]101032200000 
Mandevilla CN 10021112001001100101101132200000 
Mandevilla ligustriflora 10000111011001000002101032201000 
Mandevilla martiana (123002111 tine RAE e eg 
Mandevilla moricandiana 100211110010010 

Mandevilla myriophyllum 2002011 19 0 f 101032200100 
Mandevilla nerioides 2010011200100 110000 1101232200000 
Mandevilla oaxacana 1000011 1001001000001 o : 

Mandevilla pendula 1000011 1001001000001 101032206 
Mandevilla 9 (1210000111001 pets cu o ono 
177 5 illa pohliana 20020111001001000001 101032: 

Mandevilla a 20000111001001010001 157 
Mandevilla rugellosa 10000112001 101110101101 8 din 

9 lla rugosa 10000112001001100101101132200 

indevilla sancia {12}0021 111001001000001 57 

8 scabra 10000112001001100101101132200000 
Mandevilla sellowii 10021111001001000001 101032200100 
a 7 . 20000111001001000001 101032200100 

ndevilla splendens 10021111001001000001101132200101 

us indevilla 1 10020112001001 110101 101732200000 
Mandevilla syrin: 1000011 1001001020001 101032201011 
Mandevilla tenuifolia l 20020111001001010001 101032200100 
Mandevilla tenuifolia 2 20020111001001010001 101032200100 
Mandevilla tric 10000111001001010002101032201000 


olor 
Mandevilla tubiflora 10000111001001020001101032201010 


Annals of the 
Missouri Botanical Garden 


Appendix 2. Continued. 


Taxon 


Character states for characters 1-32 


Mandevilla urophylla 
Mandevilla venulosa 
Mandevilla veraguasensis 
Mesechites mansoanus 
Mesechites minimus 
Mesechites roseus 
Mesechites trifidus 
Odontadenia lutea 
Pachipodium geayt 
Pachipodium lamerei 
Prestonia riedelti 
Rhodocalyx rotundifolius 
Secondatia densiflora 
Telosiphonia brachysiphon 
Telosiphonia hypoleuca 
Telosiphonia nacalpulensis 
Tintinnabularia gratissima 
Tintinnabularia mortonit 


Trachelospermum jasminoides 


1002 1111001001000001101032200100 
200001 11001001000001 101032200100 
100001 11001001000002 102032201000 
100001 11000001010001 101032100010 
100001 11002001010001 101032100010 
100001 11007001010001 101032100010 
100001 1100000 1010001 101032100010 
1002010000000 1000000 E1102 1000000 
000-0000000000-000 10000-00000000 
000-0000000000-00000000-00000000 
11010100001 1011110001 11010010000 
21000100000101 11 1000101010010010 
10020 10000000 10100001 1102 1000000 
20000111 101001010001 101032200000 
20000111101001010001 101032200000 
20000111101001010001 101032200000 
1000011 1017001000001 101032100010 
100001110172101000001 102032100010 
10020 10000000 10 1000010102 1000000 


Volume 93, Number 4 Simoes et al. 591 
2006 Is Mandevilla Monophyletic? 


Appendix 3. Characters and character states for the 17. Annular corona: 0, absent: J. present 
morphological matrix used in the cladistic analyses. 9. Form of the lower corolla tube: 0, straight: 1, curved. 
19. Stamens: 0, completely included: 1, tips of the anthers 

l. Habit: 0, trees; l, lianas or vines; 2, erect shrubs or exserted, stamens + completely exsertec 
subshrubs, these often with a xylopod. 20. le base: 0, strongly sagittate; I. cordate; 2, truncate, 

2. Latex: 0, white: 1, tran: slucent. 2]. Anther guide-rails: 0, composed mainly of endothecial 

3. Stem in cross section: 0, circular; 1, pentagonal. thickenings; 1, composed mainly of sclerenchyma. 

4. Nodal colleters: n iso tiolar; I. intrapetiolar: 2. 22. Dorsal side of anthers: 0, completely glabrous; |. with 
continuous. Colleters are small glandular structures mos 
found on the margin or in axillary positic » both 23. Filament length: 0, anthers + sessile; 1, « 1 cm long; 

vegetative and reproductive organs in the ae ynaceae 1 long. 
(Thomas, 1991). Their number and organization have. 
( Hm od. ). i eo enon de 24. reste of filament and anther connection: 0. flat: 
been traditionally used as taxonomic characters in the l 
family. TH : C coll he | 1 l, with a globose swelling: 
amı 1€ arrangement of colleters on the branch 9r a 
PUE. POR ` 25. Anther/style head union: 0, anthers attached E 
nodes constitutes an easily coded character that has not 4 
. circular patch of irichome-like cells; 1, anthers attached 
received taxonomic scrutiny in Mandevilla and related ] 5 
by a horseshoe-shaped rim of hairs: = anthers attached 
genera. i 
2 by a horseshoe-shaped rim of hairs and a narrow 
5. Spiny ring of nodal colleters: 0, absent: 1, present. In 
: longitudinal strip: 3, anthers attached by ce oe fusion, 
some species of Mandevilla, the nodal colleters are . 
20. Style head shape in cross section: 0, circular or 
greatly expanded and form a somewhat spiny crown 
subcircular; I. pe Nldsomili 2 2. with five strongly project- 
around the nodes : : 
ing ribs. 

6. Phyllotaxis: 0, alte nate: l, opposite. T . . 

> . 27. Style head ribs: 0, absent; 1, restricted to the base: 2, 

7. Leaf colla ers: 0, absent: l, present. 10 re | taf dhe bore 4f red 

i N along the entire length o ody e style he: 

8. Leaf colleter ME 0, clustered at the base of the » ang Ihe entte engin 98 y of the sty : eat 
leaf blade adaxially; 1, spread along ds midrib of the 28. Collar or wreath at base of style head: O. absent: I. 
leaf blade adaxially. present. 

X . ; 5 = 8 

9. Abaxial leaf surface: 0. thick indument of white wooly 29. Proportion between the api p E ndages and 

main body of the style head: 1:1: J. 1:1 or 


trichomes absent; 1, thick indument of white wooly 
appe ndage 8 >s biggo r than the main 19 “the style head is 


tricl 
I0. Domatia: ( 


a: 0. 3 l. pr divided in two portions: two apical appendages and 

11 orescence type: 0, Loe hed (cymose): |, unbranched a massive main body. The appendages are variable in 

(racemose). size within different species of Mandevilla, and their 
12. Braets: O. scarious; 1, petaloid. size in proportion to the main body constitutes 
13. Pedicel: 0. present; Le absent. a character that has never been used before in the 
14. Calycine colleters: 0, absent; 1, present. genus. 
15. Calycine colleter arrangement: 0, alternate to contin- 30. Nectaries number: 0, five; 1, two. 

uous: |, 1 1 31. Nectaries height: 0, smaller than the ovary; 1, equal or 
16. Corolla s ). infundibuliform or campanulate to greater than the « 


tubular-c 9 1 l, salverform; 2. tubular: 3, rotate. 32. Ovary ar a E sent; l, present. 


REVISIÓN DE LAS ESPECIES DE Ana María Cialdella, Osvaldo Morrone y 
AXONOPUS (POACEAE, Fernando O. Zuloaga” 
PANICOIDEAE, PANICEAE), 

SERIE SUFFULTT'? 


RESUMEN 


El prese nte rabajo compre nde la revisión taxonómica de las especies del género Axonopus P. Beauv. serie Suffulti ( 
Black. Esta serie se caracteriza por incluir especies perennes, frecuentes en Mon y campos. sobre suelo arcilloso o arenoso, 
con oats rocosos, hasta los 3500 m. Las plantas son rizomatosas o estoloníferas, con vainas e D sadas y láminas 
desde angostamente lineares a linear-lanceoladas, planas o plegadas totalmente o bien solo en la base. Las inflorescencias son 
exertas, ubic 1 en el ápice de las cañas, compuestas por racimos espiciformes, en número vari ble mayormente An 


El raquis de los racimos es triquetro, sinuoso, con ángulos hispídulos, a veces con pelos largos y rígidos. La lemma a rior 


la pálea superior ambas endurecidas, son castaño oscuro, de supe i ie papilosa y lustrosa. Dentro de la serie Suffulti se 
consideran 16 especies, mayormente distribuíd: as en Sudamérica, con A. 1. ciliatifolius Swallen y A. o Ma Davidse e 1 5 “micas 
de Belice y Panamá respectivamente. Se incluyen descripciones de los caracteres morfológicos y de los microcaracteres de 


antecio superior, una clave para la identificación de las especies, y para cada una de ellas se proporciona una descripción, 
sinonimia, distribución geográfica e ilustraciones de los caracteres con valor taxonómico. F inalmente, se lectotipifica la serie 


Suffulti y A. ramboi G. A. Black es incluida en la sinonimia de A. argentinus Parodi. 
ABSTRACT 


The taxonomic revision of species of the genus Axonopus P. Beauv. series Suffulti G. A. Black is presented. This series 

icludes perennial species, frequently growing in savannas and in grasslands, on argillaceous or sandy ground with rocky 
outcrops, at altitu 1 0 up to 3500 m. The plants are rhizomatous or slalonife rous, with conduplicate sheaths and narrowly linear 
to linear-lanceolate blades, flat to folded along their entire length. or open toward the apex. The inflorescences are exserted 
from the apex of the culms; they consist of spiciform racemes of variable number, mainly digitate. The rachis of each raceme is 
triquetrous, sinuous, with hispidulous angles. sometimes with long and rigid hairs. The upper lemma and the upper palea are 
indurate, dark brown, „ and shiny. Sixteen species are here considered within Axonopus series Suffulti, principally 
distributed in South America, with A ciliatifolius Swallen and A. jeanyae Davidse endemic a ss of Belize and Panama, 
respectively. Dese d of Dub: td characters and microcharacters of the upper floret, a key to identify the species, 
descriptions, synonymy, geographical distribution, and illustrations of characters of eg omie value are given. Finally. 

l 


— 


lectotypification of series 5 all! ise stablishe d. and A. ramboi G. ack is included in A. argentinus Parodi as a synonym. 


Key words: Axonopus series Suffultt, Centroamérica, Poaceae, Sudamérica. 


Axonopus P. Beauv. es un género americano — et al., 2006). Este género habita frecuentemente por 
distribuído desde el sur de Estados Unidos de debajo de los 1000 m, constituyendo un importante 
América hasta el norte y centro de Argentina y elemento de las sabanas naturales. Desde el punto de 


Uruguay. Comprende alrededor de 110 especies vista económico, algunas de sus especies lienen 


(Black, 1963; Clayton & Renvoize, 1980), mayor- importancia como es el caso de A. compressus (Sw.) 
mente concentradas en la región tropical, siendo P. Beauv.. “carpel-grass”, utilizada para césped: 


Brasil, Venezuela y Colombia los países con mayor asimismo, ésta, junto con otras especies—A. fissifolius 
o-Cafias, 2000b: Zuloaga (Raddi) Kuhlm., A. scoparius (Flüggé) Kulhm.. 


riqueza en especies (Giral 


! Deseamos agradecer a Vladimiro Dudás y Francisco Rojas (Instituto REDE la elaboración de las ilustraciones, 
| Alejandra Garbini (Instituto Darwinion) la 0 ned de las fotografías. y a Alejandro Escobar (Instituto 5 
Ms vrian Esther Giordano y Alejandro Reynoso (CITEFA). su colaboración en las 8988 'rvaciones realizadas con el microscopio 
electrónico de barrido. Agrade 'cemos también a los He rectores y curadores de los herbarios consultados la 1 en el 
préstamo de material, al Board of Trustees of Royal Botanic Gardens, Kew por el envío de imágenes de ejemplares tipo 
5 en ese herbario y a todo el personal del Instituto Darwinion por 1 siempre su ayuda. Parte de este estudio 


e llevó a cabo con fondos de la National Ce graphie Society (subsidio 6825-00) y de la Myndel Botanica F 9 10 que 
permitieron realizar trabajos de campo, del Consejo Nac ional de | nvesligaciones Científicas y Técnicas (CONICET, PIP- 
02131) v de la Agencia Nac ional de 3 romocion Ce ntífica y Tecnológica (ANPCYT, PICT-1 9). 

? Los editores agradecen a Sophia Baleomb su colaboración en la redacción de este manuscrito 


B1642HYD), Argentina. anacialdella@ 


— 


? Instituto de Botánica Darwinion, I enia 200, Casilla de Correo 22, San Isidro 


darwin.edu.ar 


ANN. Missouri Bor. Garp. 93: 592-633. PUBLISHED ON 15 DECEMBER 2006. 


Volume 93, Number 4 
2006 


Cialdella et 593 


al. 
Axonopus Serie Suffulti 


suffultus (Mikan ex Trin.) 
Parodi—son forrajeras Ed (Nicora & Rúgolo de 
Agrasar, 1987; Giraldo-Cañas, 2000a; Anton, 1982). 


cual 


Chase, A. 


— 


purpusii (Mez 


El trabajo más integral sobre Axonopus, de 


derivaron los estudios florísticos posteriores, fue 
realizado por G. A. Black (1963), quien reconoció 
109 especies y 20 variedades dentro del género. Black 


agrupó las espec ies en secciones y se ries, siguie ndo 


— 


dásicamente la clasificación anteriormente propuesta 
por Chase (1911). La desaparición de Black en 1957 
tuvo como consecuencia que el manuscrito fuera 
finalmente ordenado por L. B. Smith, lo que justifica 
que el manejo del trabajo resulte complejo y que los 
límites entre los taxones sean poco claros. Por otra 
parte, Axonopus no cuenta con un volumen importante 
de colecciones en los herbarios, y, de hecho. algunas 
entidades sólo se encuentran representadas por el 
material tipo. Recientemente, el interés por trabajos 
florísticos en Centroamérica y América del Sur ha 
llevado a incrementar. considerablemente las colec- 
ciones del género. Por esta razón, se considera 
oportuno intensificar el estudio en este género, v 
lograr así una clasificación más natural. 

El objetivo del presente trabajo es contribuir al 
conocimiento de las especies de Axonopus incluidas 
serie Suffulti G. A. Black de 


identificación de las 


en la la sección 


Axonopus. Para la especies 


estudiadas, se utilizaron. principalmente. caracteres 
exomorfológicos vegetativos, tales como el hábito y el 
tamaño, forma y disposición de las hojas, así como 
también caracteres relacionados con la inflorescencia. 
entre ellos el indumento del raquis, el número de 
racimos, el tamaño de las espiguillas y la nerviación 
inferior. Se 


superior y la lemma 


de la gluma 


proporciona una clave para la identificación de las 
especies incluidas en la serie, descripción, sinonimia, 


distribución geográfica y observaciones de cada una 


de ellas e ilustraciones de los caracteres de valor 
taxonómico. 
MATERIALES Y MÉTODOS 

Para la identificación de las especies, se han 


seguido los métodos clásicos de la taxonomía. Se 
examinó el material depositado en los herbarios: BAA, 
K, LPB, MEXU, MO, NY, SI, US, USZ, VEN 
(Holmgren et al., 1990). Una lista de las especies 
tratadas en la serie Suffulti de Axonopus se incluyen 
en el Apéndice 1. Dentro del material examinado de 
cada especie sólo se citaron ejemplares representati- 
vos de cada país; una lista completa de los 
especímenes estudiados, ordenados alfabéticamente 
final del texto 


por coleccionista, se encuentran al 


(Apéndice 2). 


Se realizaron observaciones de la lemma y pálea del 


antecio superior, en vista topográfica mediante 


microscopia electrónica de barrido, en un equipo 
Zeiss DSM 940 A, que funciona en el Instituto de 
Botánica Darwinion. Previamente al metalizado, se 
procedió a la remoción parcial de las ceras epicuti- 
culares, mediante ultrasonido, con el material sumer- 
gido en xileno durante 60 minutos (Dávila € Clark, 
1990). Luego el material se dejó secar a temperatura 
ambiente, para su posterior metalizado con oro- 
paladio. La película utilizada para fotografiar esta 
superficie fue AGFAPAN APX 100 Professional 120. 
Por otra parle, se analizaron otras muestras en un 
ESEM Philips Electroscan 2010, Microscopio Elec- 
trónico de Barrido Ambiental (environmental scan- 
ning electron microscope), que funciona en el Edificio 
Instituto de 


de Materiales, del Investigaciones 


Científicas y Técnicas de las Fuerzas Armadas 


(CITEFA). En 


servadas sin preparación previa (no necesitan meta- 


este caso, las muestras fueron ob- 
lizado ni punto crítico) y en condiciones de humedad 
relativa por encima del 50%, a un presión de entre 3 y 
> Torr y una temperatura no mayor de 5 C. Para crear 
estas condiciones el. ESEM cuenta con una cámara 
para muestras, con presión ajustable que va desde 0.1 
a 20 Torr y una platina por efecto Peltier que va en + 
50 grados centígrados a partir de la temperatura 


ambiente. 


HISTORIA DEL GÉNERO 
El género Axonopus fue establecido por Palisot de 
Beauvois (1812), sobre la base de cuatro especies 
originalmente pertenecientes al género Milium L.— 
M. digitatum Sw., M. 
Sw.—y 


M. compressum Sw., cumicinum 


y M. 


Axonopus aureus P. 


paniceum una quinta especie, 


Beauv. Beauvois señala que 


Axonopus se halla estrechamente relacionado con 


Milium y Paspalum L., y lo distingue por las 
inflorescencias digitadas. 
Chase (1911) afirma que el carácter que identifica 


este género es la presencia de espiguillas solitarias, 


— 


además de la forma de su inflorescencia. Diversos 
autores en trabajos posteriores siguieron el criterio de 
reconocer Axonopus como género, tales como Roemer 
y Schultes (1817), Hooker (1896), Hitchcock (1908), 
Kuhlmann (1922), Henrard (1922), Parodi (1938), 
Pilger (1940), Swallen (1948), Black (1950) y 
Dedecca (1956). Otros autores, en cambio, considera- 
ron Axonopus como grupo o sección dentro del género 
Paspalum, entre los que pueden citarse Trinius (1826, 
1828-1836), Nees (1829), Hackel (1887), 

(1881), Bentham y Hooker (1883), Doell (187 
(1917, 1921). 


por primera 


Bentham 
7) y Mez 
Cabe señalar que fue Nees (1829) quien 
reversa de las 


vez notó la posición 


Annals 
1 B Garden 


espiguillas en Axonopus, carácter también observado 

por Hackel (1887) y utilizado por Hitchcock (1908, 

1909) y Chase (1911) para delimitar al género. 
Chase (1911), por otra parte, fue el primer autor 


Axonopus en tres 


que propuso la subdivisión de 
secciones: Lappagopsis (Steud.) Chase, Cabrera (Lag.) 
Chase Axonopus. Este criterio fue adoptado por 


Pilger (1940) y por Black (1963). Según este último 


autor, la sección Axonopus, en la cual incluye 96 
especies, se distingue por presentar raquis glabro. 


escabroso o piloso, rara vez con pelos papilosos 


blancos, espiguillas glabras o pilosas pero no 


papiloso-pilosas y antecios pajizos a castaño oscuro. 


a sección Lappagopsis, compuesta por tres especies. 
se caracteriza por sus inflorescencias con raquis y 
espiguillas cubiertos con pelos  papilosos tiesos 


blanquecinos y antecios castaño oscuro. Finalmente, 


a sección Cabrera, con 10 especies, agrupa entidades 
con raquis densamente piloso, los pelos rígidos, 


papilosos, dorados, frecuentemente fasciculados e 


a base de las espiguillas, espiguillas glabras o 


esparcidamente pilosas y anlecios castaño oscuro. 
Giraldo-Cañas (2000b) establece la sección Senescen- 
tia sobre la base de A. senescens (Doll) Henrard. y la 
distingue de las secciones Cabrera y Lappagopsis por 
tener antecio superior pajizo y pelos del raquis 
blanquecinos a hialinos. 


En la sección Axonopus, Black (1963) propuso cinco 


series: Axonopus, Barbigeri, Suffult, Capilares y 
Fastigiati, que se distinguen por el indumento. del 


raquis, el color del antecio superior, el número de 
racimos, la nerviación de gluma superior y lemma 
inferior, la forma de las hojas y el hábito de la planta. Al 
Black (1963) no designó 


especie lipo para cada una de ellas, siendo entonces 


describir estas series, a 
nombres inválidos de acuerdo al Código Internacional de 
Nomenclatura Botánica (art. 10.2: € 2000). 

Dentro de la serie Suffulti, Black (1963) chos 19 


especies. Estas se identifican por su hábito perenne, 


„reuter el al. 


sus inflorescencias con raquis hispídulo generalmente 
con escasos pelos blanquecinos no papilosos, dis- 
persos o concentrados en la base de los pedicelos, rara 
vez con raquis glabro, sus espiguillas provistas de 
eluma superior y lemma inferior con nerviación poco 
notable y generalmente sin nervio central, glabras o 
con pelos esparcidos, cortos, no papilosos y por sus 


antecios oseuros y lustrosos aún cuando inmaduros. 


CARACTERES MORFOLÓGICOS DE Las ESPECIES DE LA 
SERIE SUFFULTI 
HÁBITO 
Las especies de esta serie son perennes, rizomalo- 
eráciles o bien robustas. 


sas. rara vez estoloniferas, 


Generalmente, están provistas de un rizoma corto y 
grueso, de nudos muy próximos, carácter que genera 
plantas densamente cespitosas. Un caso interesante es 
el de Axonopus pressus (Nees ex Steud.) Parodi, con 
pajizas, notable- 


innovaciones cubiertas por valnas 


mente dísticas, ascendentes u horizontales, tempra- 
namente subterráneas, luego aéreas. En A. argentinus 
Parodi, el rizoma es estoloniforme, es decir subterrá- 
neo pero delgado. con entrenudos largos, dando lugai 
cuatrecasasii 
1. suffultus, se 


EON en uds un típico estolón, arraigado en los 


a una planta laxamente cespitosa. En h 


. Bla 'k, EX CE pe ¡onalme nte en 


nudos y a partir del eual se originan vástagos distantes 
entre sí. 
Las cañas son herbáceas o subleñosas. cilíndricas o 


con un surco 


levemente comprimidas, veces 


longitudinal, simples o ramificadas en la base o en 


la zona media, paucinodes o plurinodes, con en- 
trenudos mayormente glabros y nudos conspicuos, 


elabros a vilosos. 


CARACTERES FOLIARES 


Las vainas son abiertas, lateralmente comprimidas, 
carinadas o redondeadas en el dorso, glabras a hirsutas 
o densamente vilosas. En algunas especies (Axonopus 
flabelliformis Swallen, A. suffultus. A 


Giraldo-Cañas) la disposición de las vainas en la base 


magallanestae 


de la planta es conspicuamente dística, equitante, 
carácter que genera plantas lateralmente comprimidas 
de aspecto iridáceo. En A. suffultus además las vainas 
son glaucas, con tintes violáceos. La zona del cuello 


es de aspecto variable: puede ser inconspicuo, o bien 


notable por su color más intenso que la vaina o por la 
presencia de pelos, a veces concentrados hacia los 
bordes de la vaina. Axonopus succulentus G. A. Black 
se distingue del resto de las especies por la presencia 
de aurículas. 

La lígula es membranácea, corta, glabra o pubes- 


cente, y su borde es siempre densamente pestañoso. 


En las especies estudiadas no se encontraron 
diferencias significativas de valor taxonómico. 


— 


as láminas son conduplicadas, desde angosta- 
mente lineares hasta. linear-lanceoladas, planas o 
plegadas totalmente o bien solo en la 
agudo a obtuso, a veces retuso en Axonopus 
Hitehe., A. 


A. sueculentus y ocasionalmente bi-partido 


ápice 


— 


elegantulus (J. Presl) magallanestae, / 
pressus, 

1A. pressus. La consistencia varía desde herbácea 
Black, A. polydactylus (Steud.) 


rígida. El indumento. de las 


(A. pennellii G. 
Dedecca) 


láminas es 


hasta 


marcadamente variable. hallándose 


ejemplares con láminas glabras a densamente 


vilosas. 


Volume 93, Number 4 
2006 


Cialdella et al. 
Axonopus Serie Suffulti 


INFLORESCENCIAS 


Las inflorescencias son exertas, ubicadas en el 


ápice de las cañas, compuestas por racimos espici- 


formes, mayormente digitados, siendo los basales 
alternos. El número de racimos es variable en esta 


serie, siendo poco numerosos (2-6(-10)) en Axonopus 
arcuatus (Mez) G. A. "d A. elegantulus, A. 
Black, A. hoehnei G. 
A. Black, A. magallanesiae, A. ciliatifolius Swallen, A. 


cuatrecasasti, A. gracilis G. A. 


Jeanyae Davidse y A. ramosus Swallen; en cambio, en 

suffultus, A. polydactylus, A. pennellii, A. succu- 
lentus y A. pressus, el número de racimos es mayor (de 
11-34, excepcionalmente menos). En A. argentinus y 


A. flabelliformis Swallen el número de racimos varía 


entre cinco y 20. Este caracter es de utilidad para e 
reconocimiento de las especies, solo asociado a otros 
caracteres. 

El raquis de los racimos es triquetro, sinuoso, con 
cara dorsal plana y provista en general de un nervio 
central conspicuo, la cara ventral con una costilla 
central prominente. Generalmente los ángulos del 
raquis son hispídulos y pueden presentar pelos largos 
los pedicelos, 
Este 


'arácter ha sido tradicionalmente empleado con valor 


y rígidos a la altura de la base de 


ocasionalmente dispersos sobre los ángulos. 
diagnóstico a nivel específico. 


ESPIGUILLA 


Las espiguillas se disponen solitarias, unilateral- 
mente en dos hileras sobre el raquis, generalmente 
alternas, rara vez geminadas. Son múticas, la gluma 
inferior se encuentra ausente y la gluma superior se 
ubica externa con respecto al raquis. La flor inferior 
está reducida a la lemma inferior de aspecto 
glumiforme y ubicada hacia el raquis. Tanto la gluma 
superior como la lemma inferior son membranáceas, 
generalmente 2-3-nervias, más rara vez 4—5-nervias, 


Ambas son 


siendo sus nervios poco conspicuos. 
generalmente de igual longitud que el antecio, 


ocasionalmente mayores o bien la gluma superior 
puede ser menor, dejando ver el ápice de la lemma 
fértil. En cuanto al indumento, pueden ser glabras o 
con pelos dispersos, a veces concentrados a ambos 


lados de los nervios. 


ANTECIO SUPERIOR 
Textura 


La forma del antecio superior es sumamente 


constante, de contorno oblongo, dorsiventralmente 


comprimido, con ápice agudo u obtuso, compuesto por 


la lemma superior y la pálea superior ambas 


endurecidas, glabras, a veces con escasos pelitos 


cortos en el ápice de la lemma, castaño oscuro, de 
superficie papilosa y lustrosa (Fig. 1A). Los márgenes 
de la lemma se encuentran abrazando a la pálea, 
cubriendo los 2/3 de la superficie, encerrando el ápice 
de la pálea. La lemma y la pálea se hallan compuestas 
de células largas rectangulares, más de dos veces más 
largas que anchas, con las paredes anticlinales 
longitudinales y tranversales onduladas (Fig. 1B). 


Ornamentación 


La epidermis abaxial de la lemma y de la pálea 
presentan papilas simples, cuerpos de sílice, macro- 
pelos o micropelos. Las papilas son simples, una por 
célula, excéntricas, próx:mas a la pared transversal 
distal, distribuidas regularmente en hileras longi- 
tudinales (Figs. 1C, D, F, 2A, C, E, F). Las papilas 
poseen el ápice uniformemente redondeado, excep- 
cionalmente éstas pueden terminar en una punta 
como se observó en algunas 


aguda, muy pequeña, 


superficies de lemma y pálea de Axonopus polydacty- 


lus, A. ramosus y A. 1 his (Fi 


de eje corto, o cruciformes, 


g. 2E). Los cuerpos de 
sílice son halteriformes, 
equidimensionales, exfoliados, presentes en la zona 
apical de la lemma y de la pálea, como por ejemplo en 
A. flabelliformis, A. 
hoehnei, A. magallanesiae, A. pennellii, A. polydacty- 
lus, A. ramosus, A. succulentus y A. suffultus (Figs. 1B, 
2B, D). 


pueden estar presentes o ausentes, cuando presentes 


/ 


A. arcuatus, A. argentinus, 


a presencia ce macropelos es variable, 


— 


son unicelulares, escasos, simples, de paredes 


e 


— 


igrosadas y paralelas, y se localizan en la región 
apical de la lemma, en A. argentinus, A. jeanyae, A. 
magallanestae, A. pennellii, A. ramosus, A. polydacty- 
lus y A. suffultus (Fig. 1A, E, F); ocasionalmente, se 
observaron macropelos aislados en la base de la pálea 


en A. flabelliformis, A. suffultus 


(Fig. 2F). Los micropelos pueden estar presentes o 


pennellii y A. 
ausentes, cuando presentes éstos son bicelulares, 
fusiformes, del tipo *panicoide", con la célula distal 
dos a tres veces más larga que la célula basal, de 
ápice subagudo y se observaron en el ápice de la 
lemma y de la pálea de A. arcuatus, A. hoehnei, A. 
y I > 
pennellii, A. polydactylus, A. ramosus y A. succulentus. 
DISTRIBUCIÓN GEOGRÁFICA 


Las especies de Suffulti se distribuyen desde Belice 
hasta Uruguay y la provincia de Buenos Aires en 
Argentina. En Centro América crecen dos especies, 
y A. jeanyae, la primera 
Belice y la 


Panamá. El mayor nümero de especies se concentran 


Axonopus ciliatifolius y 


endémica de segunda endémica de 


en Sudamérica con una distribución restringida. En el 


Escudo Guayanés habitan A. gracilis, A. ramosus, A. 


596 Annals of the 
Missouri Botanical Garden 


w 


Figura |. Fotomicrografías de antecios superiores j Ivonopus P. Beauv. A-C. A. argentinus Parodi (de Burkart 25270). 
-A. Vista general de un antecio, del lado de la pálea. —B. Detalle de cuerpos de silice en el borde de la lemma. —C m alle 
de la superficie de la lemma con E is. —D. Tan lliformis Swalle n (del holotiy »» Hitchcock. 1727: 0 Detalle « la 
superficie de la lemma con papilas. — . Jeanyae D: widse (del holotipo Davidse & C. W. lamilton 23570). Detalle de y ápice 
del antecio visto del lado de la pálea, con macropelos y cuerpos de sílice en la lemma. —F. A. pennellii E V. Black (del 


holotipo Pennell 15.39). Detalle de l ápice del antecio visto es | lado de la lemma, con mac rope los, mic rope los y papi ilas 


Cialdella et al. 


Volume 93, Number 4 
2006 Axonopus Serie Suffuiti 


d 


ae 


E 
; 
Eb 
) 


fe 


* 
n 
* EFE uw. 


A EUR 


Dee ‘ 
tS 


» fh om 
2?» 5 


32 19449945 


riu ee. 
TILE 


Fotomicrografías de antecios superiores de Axonopus P. Bea A-C. A. magallanesiae Giraldo-Cañas (de 
holotipo Huber 12995). —A. Detalle del ápice del antecio 9 del lado de la ER a. —B. Detalle de cuerpos de sílice en el 
borde de la lemma. —C. Detalle de papilas de la lemma. D—E. A. polydactylus (Steud.) Dedecca (del isotipo Salzmann s.n., US 

50). —D. Detalle de cuerpos de sílice y papilas en el bite de la lemma. —E. Detalle de papilas de la lemma. —F. A. 
suffultus (Mikan ex Trin.) Parodi (del isotipo Mikan s.n., US 80029). Detalle de la base del antecio visto del lado de la pálea 


con macropelos y papilas. 


Figura 2 


598 Annals of the 
Missouri Botanical Garden 
Tabla 1. Características morfológicas de las series (sensu Black, 1963) de la sección Axonopus. 


Series 
Caracteres Axonopus Barbigeri Capillares Fastigiati Suffulti 
Ciclo de vida perenne perenne anual perenne perenne 
Antecio superior pajizo pajizo pajizo a caslaño pajizo castaño oscuro 
claro 
Gluma superior glabra o pilosa — glabra o glabra o largamente elabra o pilosa 


pubescente 


Nervios de la gluma nolorios prominentes 
superior y lemma 


inferior 


pubescente ciliada 
notorios a tenues 


notorios notorios 


Nervio medio generalmente presente generalmente presente ausente 
ausente nte 
Raquis glabro glabro glabro papiloso-piloso glabro a piloso 


pennellii y A. magallanesiae. Axonopus flabelliformis y 

cuatrecasasii presentan una distribución disyunta, 
con especies en el Escudo de las Guayanas y también 
en el centro de Bolivia. Axonopus elegantulus es 
endémica de Perú y crece en altitudes medias de los 


Andes. Axonopus succulentus es una especie endémica 


de la región oriental del Paraguay. Axonopus arcuatus. 
A. polydactylus y A. hoehnet se encuentran restringi- 
das al centro y este de Brasil. Axonopus pressus es la 
especie que posee más amplia distribución creciendo 
desde el noreste de Brasil hasta la región oriental de 
Paraguay y noreste de Bolivia. Axonopus argentinus y 
A. suffultus se encuentran restrigidas al sur de Brasil, 


este de Paraguay, Uruguay y noreste de Argentina. 


DISCUSIÓN 


Las especies de la serie Suffulti son fácilmente 


distinguibles de otros miembros del género Axonopus 


por la presencia del antecio superior castaño oscuro y 
lustroso. De acuerdo con los caracteres morfológicos 
estudiados sus especies se caracterizan por tener 
inflorescencias exertas y axilares, con racimos 
mayormente digitados, raquis de los racimos glabro 


o piloso, cuando piloso con pelos rígidos usualmente 


próximos a la base de los pedicelos, la lemma y la 
pálea del antecio superior son endurecidas, castaño 
oscuro y de superficie papilosa lustrosa. 


Dentro del género Axonopus las especies de la serie 


Suffulti comparten con las especies incluidas en la 
secciones Cabrera y Lappagopsis la presencia de 
antecio superior castaño oscuro. La sección Lappa- 
gopsis incluye tres especies, A. brasiliensis (Spreng.) 
Kuhlm., A. chaseae G. A. Black y A. herzogii (Hack.) 
Hitehe. (Black, 1963; Giraldo-Cañas, 2002). Esta 
sección posee una distribución restringida en América 
del Sur, en Bolivia, Brasil y Paraguay, y crece en 


áreas abiertas de serranías y cerros, sabanas de arenas 


blancas, en cerrados y en afloramientos rocosos del 
Escudo Brasileño. Sus especies se caracterizan por 
presentar el raquis y/o las espiguillas densamente 


dos, de base 


pilosas, con pelos blanquecinos, rig 
tuberculada. La sección Cabrera, de amplia distribu- 
ción en América desde el sur de México e islas del 
Caribe hasta Bolivia y centro y sudeste de Brasil, 
incluye dos especies, A. aureus y A. chrysoblepharis 


Lag.) Chase (Giraldo-Cañas, 2001). Esta sección se 


caracteriza por tener el raquis densamente papiloso- 
piloso, con pelos dorados, frecuentemente fascicula- 
dos y asociados a la base de la inserción de la 
espiguilla. 


Black (1963) incluye 


Axonopus y 


a serie Suffulti dentro de la 
sección caracteriza a las especies 
pertenecientes a esta sección por tener el raquis 
glabro, escabroso o escasamente piloso, los pelos de 
base no tuberculada, espiguillas glabras a laxamente 
pilosas. Esta sección es la más compleja del género e 
incluye ca. 60 especies de amplia distribución en 
América, alguna de sus especies habita en el 


A 


a Pabla | se presentan las difer- 


Paleotrópico. En 


encias entre las series de Axonopus sect. Axonopus. 


TRATAMIENTO TAXONOMICO 


Axonopus P. Beauv., Ess. Agrostogr. 12: 154. 1612. 

TIPO: Axonopus aureus P. Beauv. (lectotipo. 
designado por Hitchcock, Contr. U.S. Natl. Herb. 
12: 142. 1908). 


Plantas mayormente. perennes, rara vez anuales. 
cespilosas, rizomatosas o estoloníferas; cañas herbá- 
ceas o sublefiosas; vainas comprimidas lateralmente, 
carinadas o redondeadas en el dorso, abiertas: ligula 
membranácea breve, con borde generalmente ciliado, 


veces peslanoso; aurículas mayormente ausentes, 


c 
excepc ionalme nte presentes; láminas lineares a linear- 


lanceoladas. conduplicadas, raro convolutas, plegadas 


Volume 93, Number 4 
2006 


Cialdella et al. 
Axonopus Serie Suffulti 


599 


hasta el ápice o solo en la base, herbáceas a rígidas 
frecuentemente concentradas en la base de la planta, 
a veces dispuestas a lo largo de las cañas. o bien en la 
porción mediana de las mismas. Cañas floríferas 


exertas, con inflorescencias solitarias, terminales o 
bien terminales y axilares, naciendo del último nudo 
caulinar, compuestas por varios a numerosos racimos 
espiciformes, unilaterales, simples, digitados o sub- 
digitados, más rara vez ubicados a lo largo de un eje. 
paniculados: 


^ 


raquis de los racimos triquetro, cara 


dorsal plana, cara ventral con costilla central. 


escabroso o piloso, frecuentemente sinuoso: espigui- 
llas unifloras, solitarias, unilaterales, alternas en 2 
hileras a lo largo de la cara ventral del raquis a cada 
lado de la costilla central. Espiguillas elipsoides, de 
contorno oblongo. dorsalmente comprimidas, müticas. 
orientadas con el dorso de la lemma inferior hacia el 
raquis y la gluma superior hacia afuera del raquis. 
cortamente pediceladas, pedicelos escabriúsculos 

con algunos pelos largos: 


eluma inferior ausente: 


gluma superior y lemma inferior membranáceas. 
subiguales, tan largas como el antecio. a veces 


ligeramente mayores, o bien gluma superior más corta 
que el antecio, 2—5-nervias, de superficie papilosa. 
opacas 


lustrosas; lemma superior y pálea superior 


endurecidas, subiguales, castañas. la lemma con 


márgenes levemente involutos abrazando los bordes 
de la palea; lodículas 2; estambres 3; estilos 2. Fruto 
una cariopsis; embrión 1/3-1/2 de la longitud de la 
cariopsis; hilo elíptico o cortamente linear. Número 
básico de cromosomas x = 10 (Hickenbick et al.. 
1975; 1995). Estructura anatómica C4 
(Kranz del subtipo MS) (Brown, 1977). 


Morrone et al., 


~ 


nero con aproximadamente 100 especies, de 


América cálida. desde Estados Unidos de América 


hasta la Argentina y Uruguay. 

(1911), 
compressus como especie lipo del género, ignorando 
(1908). 
probablemente porque A. aureus es una especie mus 


Observaciones. Chase eligió Axonopus 


deliberadamente la decisión de Hitchcock 


pobremente descrita, basada en un ejemplar tipo del 


cual S 


' desconoce su procedencia y no ha sido 
(Anton, 1982). Black 


siguiendo el concepto de Chase. consideró 


localizado Por esta razón. 


(1963). 


A. aureus como taxón dudoso y lo excluyó del género. 


Axonopus serie Suffulti G. A. 
Sci. 5: 127. 1963. TIPO: Paspalum suffultum 
Mikan 1821 [= Axonopus suffultus 
(Mikan ex Trin.) Parodi] (lectotipo, aquí desig- 


Black. Frontiers. PI. 


ex Trin., 


nado). 


Plantas perennes gráciles o robustas, rizomatosas o 
estoloniferas: 


vainas lateralmente comprimidas: lá- 


minas conduplicadas, lineares a linear-lanceoladas. 


planas o plegadas. Inflorescencias con pocos a nume- 
rosos racimos (2 a más ce 30): raquis de los racimos y 
pedicelos frecuentemente pilosos. Gluma superior y 
lemma inferior con nervios inconspicuos, el nervio 
medio usualmente ausente: 


antecio superior castano 


oscuro, finamente papiloso, lustroso. 


Black (1963) al deseribir la serie Suffulti no 
designó especie tipo. En el presente tratamiento se 
seleccionó como lectotipo a Paspalum suffultum por 


ajustarse al protólogo de 


a serie 


CLAVE PARA DISTINGUIR LAS ESPECIES DE AXONOPUS SERIE SUFFULTI 


Vainas basales conspicuamente dísticas, equi- 
tantes: plantas lateralmente comprimidas. de 
aspecto iridáceo, densamente cespitosas 2 
|. Vainas basales dispuestas alrededor de la 
caña, no dísticas; plantas no lateralmente 
comprimidas 
20). Gluma superior y lemma inferior 4—5-nervias: 
inflorescencias compuestas por 3—5 racimos de 
1 de largo, láminas de (3-)6-9(-10) cm 
de 30— 
espiguillas de 2.1—2.3(-2 


de largo: plantas 50 em de altura: 

.5) mm de largo .. 
r A. magallanesiae 
Gluma superior y lemma inferior 2(-3)-ne 
inflorescencias formadas por 12-23(— 10) 7 raci- 


de 


N 


mos de 5-22 cm largo; plantas de 5: 


150 em de altura; espiguillas de (1.3—)1.6— 
2.5 mm de largo 


3(2). 


P l: antas con vainas gene ralme nte glauc 'as con 


zonas violáceas, rara vez pajizas; láminas 


pre dominante mente basales: gluma superior 
bs 5 


m de largo, igual o le 'vemente mayor 


que a anlecio; sur de Brasil. 


Paraguay y noreste. y centro. de la 


i P 

3 2 Pleats con vainas pajizas. glabras: láminas 
dispuestas a largo de las cañas: 
superior 1.2-1.5(-2.2) mm de largo 


mente 


Brasil, ( 


menor que e. antecio: Bolivia, 


Guyana, ua 
1. flabelliformis 


4D). ins con aurículas; plantas 1 con ca- 


Guayana Francesa, 


simples, de 5 mm de diámetro .. A. succulentus 


pH. e sin aurículas; 


generalmente menores de 5 mm 113 


54). Plantas estoloníferas, gráciles, levemente e 


cualrecasasti 


wa 


Plantas con rizomas cortos y gruesos, o bien 


a 


con rizomas estoloniformes, todos subterráneos 
0(5). Canas profúsamente ramificadas en la porción 


media, con las hojas mayormente dispuestas en 


a porción apical de las cañas. o bien en las 


ramificaciones laterales . ramosus 


añas simples o laxamente ramificadas en los 


— 
o 


nudos basales. con las hojas distribuidas 


mayormente 


en la región basal de la planta 


1(60). — Plantas con rizomas estoloniformes, 1 


o bien rectos, con en renudos largos .. A. argentinus 
ic Plantas con rizomas cortos, gruesos a delgados. 
Con entrenudos cortos 


600 Annals of the 
Missouri Botanical Garden 
807). Plantas con innovaciones grues curvas, espiguilla junto al pedicelo, el re le la 
asce ndentes, originalme nte subte Tráne as, lige des su perfici 4c € Spang ¡dame nte pilosa; bete 
mente comprimidas y cubiertas por numerosos distribuidas más densamente sobre el raquis, 
caláfilos rígidos, pajizos, notablemente dísticos: las contiguas opuestas en 1/4 o más de su 
láminas rígidas, de ápice subagudo frecuente- longitud; pedicelos de 0.1-0.2 mm de largo: 
mente hendido cuando viejas . . . . . . . .. . pressus Brasil, unguem Re Wwe ae us 1. hoehnel 
8'. Plantas con innovaciones aéreas, macolladas; 
láminas generalmente herbáceas, de ápice l. Axonopus arcuatus (Mez) C. A. Black, Advanc- 
agudo a obtuso, no bipartido .... .... ...... ing Frontiers Pl. Sci. 5: 137. 1963. Basónimo: 
€ ro reme Em (121 4&1 S e ~ B 
(8) co acintadas, anchas, de (1—)1.4 cm T Paspalum arcuatum Mez, Repert. Spec. Nur. 
GE XAIIGILO- Suus arse We eux eere dI n oy A. pennetcttt D > - =~ fT . "m 
i m i I Regni Veg. 15: 60. 1917. TIPO: Brasil. Minas 
9 Láminas angostamente lineares, de mm . . . . BM 
Ue O “ 10 Gerais, sin loc., Ago. 1889, A. F. M. Glaziou 
10(9). Inflorescencias formadas por (501 117030) 17933 (holotipo, B no visto: isotipos, BAA 2033! 
racimos; oan me 0 ente robustas, lá- fragm. ex P. G-DEL no visto. KI. foto ex K en SI, 
minas de 20—40 em de i cañas : : 904 10021 f 
» P no visto, US 2941983! fragm. ex B). 
plurinodes (5-6 e —— . polydactylus i 
d 9 ^ B 
10 Inflorescencias con 2—6(—10) racimos: k Plantas perennes, gráciles, de 0.40-0.60 m de 
eráciles, láminas de (1-)12-25 em de largo: : : 
DT d 5 altura. rizomatosas, el rizoma corto. densamente 
cañas paucinodes o plurinodes .. . . . . . . . .. . . lead le 2 le diá 
cespitosas; cañas delgadas de : de diámetro, 
11010). ae de los racimos con pelos largos (l- pitosas; ce gaca mus amet 
mm de largo), dispersos sobre los ángulos o simples, levemente comprimidas, plurinodes, lisas, 
bien concentrados a la altura de la base de los con surco lateral conspicuo, glabras; nudos castaños, 
pedic AR eee 12 comprimidos. glabros; vainas menores que los en- 
11’. Raquis de los racimos glabros, con los : ; 2 
trenudos, conspicuamente estriadas, pajizas o casta- 
márgenes escabriusculos . . . . . . . . . . . . . . . N 
12(11). Hojas concentradas en la base de la planta ñas, glabras, con bordes lisos, hialinos. glabros: 
rectas, ascendentes; raquis con pelos dispersos nervios de las vainas se continúan en la lámina: 
a lo largo de los ángulos, más numerosos en la cuello levemente conspicuo, castaño, o bien incon- 
ase g dense : T : 
base de los pedicelos: espiguillas densamente spicuo, pajizo, glabro, con numerosos pelos hacia los 
distribuidas a lo largo del 9 0 de 0.0 d | j Koul | , ‘liad 
márgenes de la vaina; lígulas membranaceo-ciliadas, 
0.8 mm de ancho; gluma superior b PE Coe YAU, MULAS DICTADO AGA 
inferior iguales o mayores que antecio, de 0.3-0.4 mm, castañas, blanquecinas en el márgen; 
2(3-nervias s...... A. elegantulus — contralígula y aurículas ausentes: láminas arqueadas, 
o! as dispuesiis d " Ae veh : : z 
12 Hojas dispuestas a lo largo las cañas, rígidas, 6-10 X 0.2-0,45 em, concentradas en la base 
prine ipdime nte en la porc ión me «dia e infe rior, 
y zona media de la planta, plegadas en la base, luego 
con las vainas imbricadas; láminas arcuadas; : a 
raquis con pe ‘los conce ntrados sólo a la altura planas, con bordes planos, el apice agudo u obtuso, la 
de la base de los pedicelos; espiguillas base continuándose imperceptiblemente con la vaina, 
laxamente dispuestas a i largo del raquis, de longitudinalmente estriadas en ambas caras, hipofilo 
0.5-0.6 mm de ancho: gluma superior menor : . np MD 20 [qu ! 
elabro, a veces con pelos dispersos, más frecuentes 
que el antecio, lemma inferior igual al mismo, | | : li lel ápi fil 
ambas Uds dd. a rs I. arcuatus. Sobre el nervio medio cerca del ápice, epifilo 
3(11). Gluma superior y lemma inferior, 0.1—0.4 mm levemente escabriúsculo, con los margenes escabriús- 
más largas que el antecio, 4-5-nervias . . . .. l4 culos, a veces también con pelos delgados, ralamente 
di : 
I3. G luma 3 rior menor que el antecio, rara vez dispuestos, más cercanos y numerosos hacia la base. 
igual, : -nerv . le nima inferior igual O mayor "e "T — ( . . 
Cañas floríferas con 1 inflorescencia terminal sub- 
que el antecio, px em TNI > o. Ó . . 
14(13). Cafias 20 3. podes nudas lens epirus digitada, con pedúnculo cilíndrico, ligeramente com- 
de 1.9-2.5 mm de largo, gluma superior y primido, longitudinalmente estriado, con surco lateral 
lemma infe rior pre )-nervias, con pelitos conspicuo, glabro; inflorescencia formada por 4-6(8) 
Mgeados c i S d 8 VI > 3 1 qc 
delgad "oa umbos lados de los nervios racimos de 5.5-9(-12) em; raquis de los racimos recto 
PANAMA sa 0 4. CIUS A. jeanyae : 
" en ; aaa levemente ondulado, triquetro, cara dorsal plana, 
14“. Cañas plurinodes (10 ó más nudos); nudos f : 
glabros; espiguillas de 2.6-2.8(-3) mm de cara ventral con costilla conspicua, de 0.3-0.5 mm de 
largo, gluma superior y lemma inferior 4- ancho, castafio v con tintes violáceos, hispídulo sobre 
vias, glabras; Venezuela . gracilis los ángulos y con pelos largos de 1.5-2 mm en la base 
15(13). Espiguillas de 2-2.2 mm de largo, cortamente le ] . . : : ‘ 
de los pedicelos; pedicelos aplanados, 0.1—0.2 mm, 
pilosas, con pelos delicados, adpresos, sobre briúscul "o | | 
oda la sopera de: le guma aüpenor 3 escabriúsculos en los ángulos, con numerosos pelos 
lemma inferior; espiguillas distribuidas laxa- largos similares a los del raquis. Espiguillas de 
mente a lo largo del raquis. las contiguas contorno oblongo-elíptico, de 1.7-2.1(-2.3) X 0.5- 
kir = 8 -— : e 
oponen 0-1/4 de su EM La de 0.6 mm,  dorsiventralmente comprimidas, agudas, 
0.4-0.6 mm de largo; Belice A. ci ee A io. p en . 
El ocráceas, alternas y laxamente distribuidas en 2 
15% Espiguillas de 1.6—-1.7(-1.8) mm de largi : 


densamente pilosas, con pelos rígidos a "lo 


largo de los márgenes y hacia la base de la 


hileras a lo largo del raquis, de tal manera que cada 


espiguilla resulta opuesta a una de la otra hilera en 


Volume 93, Number 4 
2006 


Cialdella et al. 
Axonopus Serie Suffulti 


1/4-1/3 


inferior frecuentemente tan largas como el antecio, 


de su longitud: gluma superior y lemma 


dup. superior a veces algo menor, con apic € agudo, 
hiali 


IHCIHTDI alta o-nla 


ialinas, ambas 2(—4)-nervias, con los 
nervios marginales, glabras; lemma superior oblongo- 
elíptica, 1.6-2 mm, castaña, brillante, glabra, con 
densas y pequeñas papilas en toda la superficie. 
disminuyendo hacia los bordes, éstos blanquecinos y 
pálea 


lisos, encerrando los bordes de la pálea; 


superior similar a la lemma, ligeramente menor, 


glabra. Cariopsis no vista. 


Distribución geográfica y ecología. Conocida sólo 
para el estado de Minas Gerais, Brasil. 

Fenología. | Coleccionada en flor y/o fruto desde 
agosto hasta febrero. 

Black (1963) menciona que Axonopus arcuatus se 
halla relacionado a A. polydactylus, distinguiéndolas 
principalmente por sus caracteres vegetativos. Anton 
(1982) señala que A. arcuatus podría tratarse de un 
ejemplar empobrecido de A. polydactylus. Sin embar- 
go. es posible diferenciar a A. polydactylus por incluir 
5—)0.9-1.2 m de altura, 
)11-17(-30), éstos 


23) em de largo y sus láminas rectas, no 


plantas de mayor porte, de (0.5 
mayor cantidad de racimos, de (5— 
de 7.5-18(— 
curvas. Consecuentemente, en este trabajo se man- 
tienen como especies independientes. 

Axonopus arcuatus es también afín a A. elegantulus, 
coleccionada en los Andes de Perú, distinguiéndose 
esta última especie por presentar hojas ascendentes, 
rectas, concentradas en la base de la planta, 
espiguillas distribuidas más densamente a lo largo 
del raquis, éstas más anchas (0.6-0.8 mm de ancho) y 
por presentar pelos largos, dispersos sobre los ángulos 
del raquis y en la base de los pedicelos. 

Mez (1917) menciona 


colección, sin especificar localidad. En la foto del 


“Brasilia” como lugar de 


isotipo depositado en K se indica “Minas”, proba- 
blemente refiriéndose al estado de Minas Gerais, y en 
la etiqueta del isotipo depositado en BAA dice 


“campos de Itabirä, 21 Nov. 1888” 


Material examinado. BRASIL. Minas Gerais: Serra do 


ltabirito, « ca. 50 km SE of Belo Horizonte, Irwin et al. 19768 
(US). 


2. Axonopus argentinus Parodi, Notas Mus. La 
Plata, Bot. 3(17): 15. 1938. TIPO: Argentina. 
Entre Ríos: Dpto. Concordia, Concordia, Ene. 
1922, L. R. Parodi 4649 (holotipo, BAA 300!; 

IAN no visto, SI!, US 1723532). 


Figuras 1A-C, 3. 


isolipos, 


Axonopus argentinus var. glabriflorus Parodi, Notas i La 
Plata, (IPO: 
937, 


N 
ND 


iruguay. Salto, 

2 lala 60, Herb. Mus. Nac. 72 
5058 100 9 BAA 302!; isotipos, MVM no visto, US 
2236050! fragm. ex BAA). 


Axonopus un var. 7 9 5 Parodi, Notas Mus. La 
. 3(17): 19. 1938. / ate 1 nim subsp. 
dies Ù (a oe ng.. B. R. Arill. € I 
Uru TIPO: Argentina. Entre Rios Dpto. 
Con is lar 1 Mayo 1933. J. 
n rcm 60 (holotipo, BAA 303!; isotipo, TAN no 
visto). 


. Gram. 


hb 
2 X 
T 
2 w 


Axonopus argentinus f. hirsutus Parodi, Notas Mus. La A 
Bot. 307): 18. 1936. TIPO Ves 
Dpto. Concordia, Concordia, Feb. 1930, L R. Paridi 
9584. (holotipo, BAA 305!; isotipo, is no visto). 

Axonopus suffultus 99 ex Trin.) Parodi var. p 
Parodi, Notas Mus. La Plata, Bot. 3(17): 1938. 
Axonopus argentinus Parodi subsp. dir da " 
Roseng., B. R. po & lzag.. Gram. Urug. 293. 1970. 
TIPO: Uruguay. Montevideo: Sayago, 22 Dic. 1921, L. 
R. Parodi 91 (holotipo, BAA 3345; isotipos, 
BAA en SI!, IAN no visto, SI!, US 11277411). 


Axonopus hagenbeckianus (Kuntze) Parodi var. aa G. 
1. 196: 


foto ex 


A. Black, Advancing Frontiers Pl. Sci. 5: 1 
TIPO: Brasil. Rio ie do Sul: Porto Ale egre, Morro 


5. B. Rambo es 3 Mo e US 
1910224: solivos, i ex US e LIL no visto). 


HO paraguayens us G Blac a ire ing Frontiers 
: J. Sci. 13: ido TIPO: Paraguay. Paraguart: 


Caapucü, Estancia Barrerito, Ene. 1949, T. Rojas 
13103a oke US 2012988: isotipos, foto ex US en 
I!, KY) 


lxonopus 11 G. A. Black, Advancing Frontiers Pl. Sci. 5: 
134. 1963. Syn. nov. TIPO: Brasil. Rio Grande do Sul: 
5 30 Ene. 1948, B. Rambo E rolotipo, US 

537!; isotipos, foto ex US en SI! K). 


Plantas perennes, cespitosas, de 0.2-1 m de altura, 
con rizomas delgados, estoloniformes, falciformes o 
rectos, cubiertos por numerosos catáfilos coriáceos; 
cañas 


innovaciones intravaginales y extravaginales; 


delgadas a ligeramente robustas aplanadas, de 2— 
3 mm diámetro, simples, 2-3-nodes, pajizas, longi- 
tudinalmente estriadas, con surco lateral conspicuo, 
glabras o con pelos en el surco lateral, o ralamente 
vilosas en los entrenudos a densamente vilosas hacia 
los nudos, con pelos de 3-4 mm de largo: nudos 
castaños, levemente comprimidos, glabros a pubérulos, 
a veces vilosos; vainas menores que los entrenudos, 


conspicuamente estriadas, pajizas, glabras o con 
algunos pelos largos en el borde hacia el cuello, 
a densamente vilosas, generalmente más notable cerca 
del cuello; nervios de las vainas continúandose en la 
lámina; cuello distinguible, castaño, glabro a piloso; 
lígulas membranáceas, de 0.5-1 mm, con márgen 
densamente pestañoso, el resto de la superficie glabra: 
aurículas y contralígula ausentes; láminas lineares, 


rígidas, de 10-30 40) x 0.3-0.7 


mente basales en la planta, plegadas en la base, 


cm, predominante- 


planas hacia el ápice, la base continuándose im- 


la vaina, el ápice obtuso o 


perceptiblemente con 
agudo, con pequeño acúmen algo punzante, glabras o 
con escasos pelos sobre el nervio medio en el epifilo, 
los márgenes glabros o escabriúsculos, o con escasos 


pelos de 2.5-3 mm cerca de la base o en toda su 


602 Annals of the 
Missouri Botanical Garden 


Figura 3. Ax vonopus ar gentinus Parodi (de Schinini 14006. S). —A. Hábito. —B. Detalle de la región ligular. — 
Detalle de una porción del raquis. —D. Espiguilla vista del lado de la gluma superior. —E. Espiguilla vista del lado de 1 


lemma inferior, —F. Antecio superior visto del lado de la lemma. —G. Antecio superior visto de i lado de la pálea. 


Volume 93, Number 4 


Cialdella et al. 603 


2006 Axonopus Serie Suffulti 
extensión, nervio central y laterales conspicuos. Nombre vulgar. “KapVipe fiu" (Paraguay; Mor- 
Cañas floríferas con 2 6 3 inflorescencias de rone etal., 1994) 


desarrollo desigual, naciendo del ültimo nudo cauli- 
nar, cada una de ellas con pedúnculo aplanado, 
glabro, longitudinalmente estriado; inflorescencia más 
(35-10 (6—)10— 


em, siendo los basales alternos levemente 


desarrollada con 15) racimos de 


12(-18 


distantes entre sí y los apicales más cercanos; raquis 


— 


de los racimos notablemente ondulado, pajizo o con 
tintes violáceos, triquetro, cara dorsal plana, de 0.5— 
con costilla con- 


0.75 mm de ancho, cara ventral 


n hispídulo, más conspicuamente sobre los 
ángulos, generalmente con 1-4 pelos largos de 1.5- 
2 mm, blanquecinos, más o menos rígidos situados 
sobre los ángulos a ambos lados de la base de cada 


de 0.2-0.3 mm, 


escabriúsculos en los 


pedicelo; pedicelos muy breves, 


conspicuamente | aplanados, 


ángulos, ocasionalmente con algunos pelos largos 
similares a los del raquis. Espiguillas de contorno 


oblongo-elíptico, de 1.9-3 X 0.6-1 mm, dorsiven- 
tralmente comprimidas, ocráceas, más raramente con 
tintes purpúreos, glabras a cortamente pilosas, alter- 
nas, dispuestas en 2 hileras a lo largo del raquis, de 
tal forma que cada espiguilla se opone a otra de la otra 
hilera en 1/3-1/2 de su longitud; gluma superior y 
lemma inferior de 1.9-3 mm, membranáceo-hialinas, 
con el ápice cortamente acuminado a subagudo, 
elabras o con escasos pelos delgados dispersos, 
a veces más densos hacia la base, los bordes y el 


2(—4)-nervia, 


lemma superior elíptica, 1. 


ápice, con los nervios próximos al 


márgen; 8-2.4 mm, en- 
durecida, castaña, subigual o hasta 0.6 mm más corta 
ue la gluma superior y la lemma inferior, poco 
brillante, glabra, con escasos pelitos cortos en el 
ápice; lemma con densas y diminutas papilas en toda 
la superficie, disminuyendo hacia los bordes, éstos 
blanquecinos y de superficie lisa, encerrando los 
bordes de la pálea; pálea superior de igual consis- 
tencia y aspecto que la lemma superior, ligeramente 
menor que la lemma, glabra; estambres 3, anteras 
purpüreas de 1—1.5 mm. Cariopsis elipsoide, de 1.5 
0.6 mm, blanquecina: hilo oblongo, subbasal: rn 
1/2 del largo de la cariopsis. 

Habita el 
sudeste y sur de Brasil (estados de Paraná, Rio 


Distribución | geográfica ecología. 
Grande do Sul y Santa Catarina), Paraguay (en la 
región oriental y occidental), Uruguay (Departamentos 
Cerro Largo, Durazno, Lavalleja, 


de Flores, Florida, 


Montevideo, Paysandú, Salto, Tacuarembó y Treinta y 
Tres) y la Argentina (Provincias de Chaco, Corrientes, 
Entre Ríos, Misiones y Santa Fe). Esta especie es 
común en bordes de montes, campos semihúmedos o 
en suelos arenoso-rocosos más secos. 

Fenología. | Coleccionada en flor y/o en fruto entre 


los meses de octubre y mayo. 


10, 20 (Hickenbick et 


al., 1975, sub Axonopus ramboi). 


Número cromosómico. 


suffultus y A. 


Axonopus argentinus es afin a A. 


pressus. Axonopus suffultus se distingue por tener 
rizomas de entrenudos cortos, no hojosos, vainas 


elaucas y cuello indistinguible. Axonopus pressus se 
aparia por incluir plantas más robustas, de 0.7- 

5(-2) m de altura, con gruesas innovaciones, 
comprimidas, cubiertas por vainas dísticas y pajizas, 
de (10-20-45 X 0.8-1.2 em, 


rígidas, frecuentemente bipartidas en el ápice cuando 


de (1032045 cm y 


láminas acintadas, 


viejas, inflorescencias mayores, 
racimos de (10—)13-32 cm. 
Parodi (1938: 
var. glabriflor us indica como material tipo 
deo n° 5038, Salto, leg. Orihuela n 60, 1-1937”. En 
los herbarios BAA y SI se halló otro ejemplar, no 
cilado en el protólogo, en cuya etiqueta se indica: 
"Herb. Museo Nacional Montevideo 5038, J. Arecha- 


valeta s.n.” 


19) al describir Axonopus argentinus 


Black (1963) describe a Axonopus ramboi sobre la 
base del ejemplar Rambo 36460 y lo diferencia de 
otras epi de la serie por sus espiguillas mayores, 
de 2.7-2.9 mm y antecio superior 0.3—0.5 mm más 
corto que la gluma superior v lemma inferior. El 
análisis de la descripción original, del material tipo y 
del material de herbario de A. argentinus permitió 
confirmar que existe una marcada variación en el 
tamaño de la espiguilla (de 1.9-3 mm) y en el largo 
relativo de la gluma superior y la lemma inferior en 
relación al antecio superior, desde subiguales hasta 
0.6 mm más largas. Además, el material tipo de A. 
ramboi depositado en el herbario US posee rizomas 
delgados, alargados cubiertos por numerosos catáfilos 
coriáceos, lípicos de A. argentinus. Por todas estas 
razones, A. rambot se incluye en la sinonimia de A. 
argentinus. 

Longhi-Wagner et al. (2001), al tratar el género 
Axonopus para Sáo Paulo, Brasil, sigue el concepto de 
Black (1963) y considera a A. 
y señala que la especie posee la gluma 


J 


ramboi como una 
especie válida, 
superior y lemma inferior 5-nervia. Sin embargo el 
análisis del material tipo permitió comprobar que 
ambas son 2-nervias. 

Axonopus argentinus es una especie que presenta 
una marcada variabilidad en el porte de las plantas, 
en la pilosidad de las vainas, láminas y pedicelos y en 
el tamaño de las hojas y de los racimos de las 


inflorescencias, razón por la cual se han considerado 
bajo esta especie numerosos taxones infraespecificos 
(Parodi, 1938; Burkart, 1968; Rosengurtt et al., 1970; 
1982). El análisis de material de herbaro 


permitió confirmar que existe una continua gradación 


Anton, 


Annals of the 
Missouri Botanical Garden 


los caracteres anteriormente mencionados que no 
permite la diferenciación de taxones discretos. 
Rosengurtt et al. (1970) 


argentinus es un forrajera medianamente apetecida 


señalan que Axonopus 


por el ganado. 


ARGENTINA. Chaco: 
Schulz 11786 (BAA), 


Material adicional examinado. 
Dpto. ! de Mayo, Colonia Benítez, 
\ 

. 0); Dpto. Concepción, Est. Santillán, 
de p Pedersen 4482 (BAA); Dpto. Gral. Alvear, Alvear, 
(BAA): Dpto. Ituzaingó, Ruta 12 próximo a Ituzaigó, 
Vicora & Cámara Hernández 517 (BAA): 10 2 NNE de 
ltuzaingó, camino a Guardia Cué, Quarín et al. 1850 (LPB): 
. Pedersen 0426 (BAA à die Santa 
5676 (BAA): Dpto. Mercedes, 
Parodi 6191, 6359 (BAA); Est. Nu-Porá, Martínez Crovetto 
9940 (BAA); Dpto. Montes Caseros, sur de Combai, Nicora 
571 (BAA); Dpto. San Sarli 101 (BAA): Est 
Pe dersen / pa AA); Chavarría, junto a la Ruta 
7 (CTES, SD. Entre Ríos: Dpto. 
\ Palmar, A irl el Crespo 22922 (BAA, SI); 
Dpto. Concordia, Colonia Yeruá, Galli 54 (BAB, SD), Galli 
287 (SI); Dpto. Concordia, C 1 Baez 27 (BAA): um 
Ana, Burkart et Troncoso 26195 Cabrera 28208 (SI 
Burkart et 
0436): Dpto. 


Roque, 
Caaguart, 


Ww 


c Burkart et E 25270 (S: Dpto 
Uruguay er ión E uguay, Est. Río Uruguay, cerca 
de Gualeguaychú, Burkart 24076 1 SI). re 
Dpto. Apóstoles, Campo San Juan, Fontana F 111-0 (SD: 
Dpto. Capital, Posadas, Establecimiento Santa Inés, pue 
1632 (BAA): Dpto. Guaraní, Ruta Prov. 2, Reserva Ecológica 
Provincial. Morrone et al. 720 (Sl) ai. Fe: ups Capital 
Santo Tomé, Parodi 157 (BAA); Gral. Obligado, Camino de 


Reconquista a Dr. Barros Pasos, Job 760 (BAA 

BRASIL. Araujo 33 (BAA). Rio Grande 
do Sul: 5. Sebastiáo do Cai, Faz. Paquete, Araujo 175 (BAA, 
US); Monte negro, Araujo 102 (BA A); Pelotas, Swallen 9169 
(US). Santa Pedir Mun. Lajes, Rio Bandeirinhas, 23 km 
: of Lajes, > Klein 8237 ( 5 Mun. Porto União, 

nith & nra 862 17 (US); Mun. Iraní, Campo de Iraní, Smith 
p Klein 13990 (US). São Paulo: depósito de remonta de Sao 
Simao, del Mazzo 25 (BAA) 

PARAGUAY. Alto Paraná: Puerto Casado, Ramirez 040 
(US). Amambay: Pedro Juan Caballero, 

(1 


Paraná: Castro, 


2 


ra de Amambay, 
Central: San Lorenzo de Campo Grande, 


Mariscal Estigarribia, Rojas 14105 (LIL, US 
Paraguaria end in regione lacus Ypacaray”. Hassler 
8 | 


m 
IP 
- 
^ 
= 
- 
gg. t 
— 


rocas, Rojas 5 bs (BAA arrica, Rojas 
14076 (BAA, SI); urbe, Montes 12834 (BAA): Camino entre 
Villa Rica y Rua et al. 38 (BAA) 
Rojas 13129 (BAA, 
campo cerrado on Ayo. Minas 
11827 (MO). 
Presidente Hayes: Chaco, frente a Asunción, Rojas 12 95 
(BAA, SI). 
URUGUAY. 


Rosa-Mato 


Coronel Oviedo, 


Paraguarí: Est. Barrerito. Caápucú. 
e 

LIL); Parque Nacional Ybicuf, « 

basin, 5km N of adm. area, Zardini 


Canelones: Dpto. 1. de Mayo, Parque 
Centenario, 1463 (B. K AJ. Cerro Largo: Río 
Negro, entre losayos, Palleros y Areguá, 
3AA, US): Río rn „ Gallinal et al. 1491 (US). 
Durazno: Maestre Campo, sd 111 (B s Flores: Rio 
Ji, Rosengurtt 8.56 8 (BAA, NT 
Manzavillagra, 


. Florida: Joy! 
Rosengurtt ee (BAA); Cerro 


Colorado, Est. Sta. Elvira, Rosengurtt 305 (US S, Lavalleja: 


Est. Pororó, Pedersen 3638 (BAA). Montevideo: Montevi- 
deo, Arechavaleta s.n. (BAA 301, SI). Paysandú: 
Lombardo 3032 (BAA). Salto: Río Uruguay y 
Ayo. San Rosengurtt B-955 (BAA). Sin 
departamento: Rio Uruguay, Orihuela 03 (BAA). 


Costa de 
Río Uruguay. 


Antonio Grande, 


Acad. 


3. Axonopus 9 Swallen, J. Wash. 


Sci. 23: 458. 1933. TIPO: Belice. El Cayo Dist.: 
Mtn. Pine us 25 Feb. 1931, H. H. Bartlett 


11746 (holotipo, US 1503594*; isotipos, foto ex 
US en SI!, K!, MO 3701229!). Figura 4. 


Plantas perennes, conspicuamente cespitosas, de 
0.50-0.70 m de altura, 


dos y cubiertos por las catáfilas, con numerosas raíces 


con rizomas breves, engrosa- 


adventicias y vástagos aéreos; cañas gráciles de 2— 


3 mm diámetro, simples, levemente aplanadas, con 


surco lateral conspicuo, 2—3-nodes, pajizas. longi- 
tudinalmente estriadas, glabras; nudos castaños, 
elabros: vainas menores que los entrenudos, con- 


spicuamente estriadas, pajizas, totalmente glabras o 
con algunos pelitos dispersos sobre la cara externa, 
con los bordes lisos, glabros; nervios de las vainas se 
continúan en la lámina; cuello inconspicuo: ligulas 
membranáceas, 0.1-0.3 mm, pajizas, blanquecinas y 
densamente pestañosas hacia el borde: aurículas y 
contral ígula ausentes; láminas angostamente lineares, 
(5-)12-14(-25) X 0.1-0.2 


temente basales, plegadas en la base, luego planas, de 


rígidas, em, predominan- 
ápice acuminado, la base continuándose impercepti- 
blemente con la vaina, longitudinalmente estriadas en 
ambas caras, con nervio central y laterales conspi- 
cuos, con pelos delgados y largos, de base tubercu- 
lada, más densamente dispuestos hacia el borde. 
Cañas floríferas con 2 a 4 inflorescencias naciendo 
del último nudo caulinar, cada una de ellas con 


pedúnculo aplanado, glabro, longitudinalmente es- 
triado: inflorescencia formada por 2 a 6 racimos de 3— 


12 em: 


triquetro, 


raquis de los racimos levemente sinuoso, 
cara dorsal plana, de 0.2-0.5 mm de ancho, 
cara ventral con costilla conspicua, pajizo en la zona 
central y verdoso hacia los laterales, hispídulo, más 
conspicuamente sobre los ángulos, sin pelos largos; 
pedicelos aplanados, ensanchados hacia el ápice, 0.4— 
0.6 mm, hispídulos. Espiguillas de contorno oblongo- 


2-2.2 0.8-0.9 mm. 


comprimidas, ocraceas, alternas, laxamente dispues- 


elíptico, dorsiventralmente 
tas en 2 hileras sobre el raquis, de tal forma que cada 
espiguilla se opone a una de la otra hilera en 0-1/4 de 

longitud; gluma superior y lemma inferior mem- 
branáceo-hialinas, 2-2.2 mm, gluma superior menor 
que el antecio, rara vez igual o mavor, obtusa, lemma 
inferior igual o levemente mayor que el antecio, 
aguda, ambas 2-nervias, ocasionalmente 3-nervias, los 
nervios conspicuos, glabras o con pelitos delgados, 


delicados, dispersos sobre toda la superficie; lemma 


Volume 93, Number 4 Cialdella et al. 605 
2006 Axonopus Serie Suffulti 


Gi —— 
— vr a — — 
uade 


0,5 mm 


f 7 A x ^ ZE <= 
P4 y o2 Nx s 


Figura 4. Axonopus a Swallen (de Hiebert 2, MO). —A. Hábito. —B. Detalle de la región ligular. —C. Detalle 
de una porción del raquis guilla vista del lado de la lemma inferior. —E. Espiguilla vista del lado de la bine 
supe rior. —F. Antecio superior visto Y 3] lado de la pálea. —G. Antecio superior visto del lado de la lemma. —H. Cariopsis. 


vista hilar. —I. Cariopsis, vista escutelar 


Annals of the 
Missouri Botanical Garden 


superior elipsoide, ca. 2.1 mm, castaña, lustrosa, 
elabra, con escasos pelitos cortos en el ápice, con 
densas y diminutas papilas en toda la superficie, 
disminuyendo hacia los bordes, éstos blanquecinos y 
de superficie lisa, encerrando los bordes de la palea: 
pálea superior de igual consistencia y aspecto que la 
lemma superior, ligeramente menor que la lemma, 
elabra. Cariopsis no vista. 

Hasta el 
Belice. 


Habita en savanas montanas, ocasionalmente sujetas 


Distribución | geográfica y ecología. 


momento sólo es conocida para El Cayo, 


a fuegos periódicos, entre los 600-800 m. 
Fenología. | Mallada en flor y/o en fruto entre los 


meses de agosto y febrero. 
Axonopus ciliatifolius presenta, junto con A. elegan- 
tulus, A. 


hábito grácil, hojas con láminas angostas e inflores- 


arcuatus, A. jeanyae, A. gracilis y A. hoehnei, 
cencias con pocos racimos. Axonopus elegantulus y A. 
arcuatus se apartan por la presencia de pelos en el 
raquis; A. jeanyae y A. gracilis se distinguen por tener la 
gluma superior y lemma inferior mayores que el 


antecio, 4—5-nervias; finalmente, A. hoehner, hasta el 


momento especie endémica del norte de Brasil, 


diferencia de A. ciltatifolius por sus espiguillas 
menores (1.6-1.7 mm de largo) y sus pedicelos más 
cortos (0.1-0.2 


Black (1963) relaciona Axonopus ciliatifolius con A. 


mm de largo). 
potophyllus Chase, esta última perteneciente a la serie 


Axonopus, distinguiéndola principalmente por sus 
Axonopus poiophyllus además pre- 
2-A(-8) 


53.5 mm de largo y gluma 


antecios pálidos. 
senta hojas más anchas, de mm de ancho. 
espiguillas mayores, de 2 
superior y lemma inferior mayores que el antecio, 2— 
4-nervias. 

Material examinado. BELICE. Belize: Almost border of 
Stann Creek Distr.. g Manatee (Coastal) Hw 
of turnoff to Gales Point, Nee & Atha 7 55 (MO) ¢ 
along No. 2 B ine Rd., 


7.4 mi. S 


Pine Ridge, near Augustine, 


(MO): Mt. Pine Ridge, Blancaneaux Lodge, Wiley 475 (MO); 
Mt. Pine Ridge, along A-10, For. Res.. 200 yds. above Rio 
Frio Crossing, Hiebert 2 (MO). 


Black, Advancing 


4. Axonopus cuatrecasasii G. A. 


Frontiers Pl. Sei. 5: 147. 1963. TIPO: Colombia. 
Arauca: Los Llanos, Río Casanare, Esmeralda, 
130 m, woods sabana, 19-20 Oct. 1938. J. 


Cuatrecasas 3882 (holotipo, US 17979180: iso- 


tipo, foto ex US en Sl). Figura 5. 


Plantas perennes, de 50-120 em de altura, grá- 
ciles, ligeramente cespitosas, estoloniferas, con esto- 


lones delgados: cañas delgadas de 2 mm de diámetro, 


rectas o geniculadas, simples, cilíndricas, con surco 


lateral ligeramente conspicuo, 2- a 4-nodes, a veces 


ramificadas en los nudos his pajizas, longi- 


tudinalmente estriadas, glabras en los entrenudos: 


nudos castaños, levemente comprimidos, densamente 
vilosos en toda la superficie del nudo o bien solo en su 


base: vainas menores o de igual largo que los 


entrenudos, longitudinalmente estriadas, glabras a vi- 
losas en su cara externa, a veces solo los bordes y el 
nervio medio densamente vilosos hacia el cuello; 
nervios de las vainas se continúan en la lámina; cuello 
diferenciado, con una línea castaña, glabro, a veces 


viloso hacia los bordes; lígulas membranáceo-ciliadas, 


castañas, de 0.2—0.5(-1) mm, pubescentes, blanque- 
cinas; aurículas contralígula ausentes; láminas 


lineares, ríg 


152332) X 0.3-0.5 


minantemente basales en la caña, plegadas en la base. 


idas, cm, predo- 


luego planas, con bordes planos, ápice agudo, de base 


continuándose  imperceptiblemente con la vaina. 
longitudinalmente estriada en ambas caras. hipofilo 


glabro a densamente viloso, epifilo glabro o con pelos 
ralos a densos, los bordes densamente vilosos hacia la 
base de la lámina, luego hispídulos a escabriúsculos. 
Cañas floriferas con La 2 inflorescencias subdigitadas 
en la última vaina foliar, con pedúnculo cilíndrico, 
levemente comprimido, con surco lateral notable, 
longitudinalmente estriado o liso, glabro: inflorescen- 
cia compuesta por (3-)5-6(-9) racimos delgados, de 
10 


levemente ondulado, 


— 
— 


(5.5—)7.5-10.5 cm; raquis de los racimos recto 


a muy triquetro, cara dorsal 


plana, de 0.4-0.7 mm de ancho, cara ventral con 
costilla conspicua, pajizo, glabro o con pelitos cortos, 
escabriúsculo sobre los ángulos; pedicelos aplanados, 
0.1—0.2 
longo-elíptico, de (1.5—)1.7—2 X 0.7-0.8 mm, agudas, 


mm, glabros. Espiguillas de contorno. ob- 
dorsiventralmente comprimidas, ocraceas, espaciada- 
mente distribuídas en 2 hileras a lo largo del raquis, 
de tal manera que cada espiguilla se opone a una de la 
otra hilera en (1/2-)1/3 o menos de su longitud: gluma 
superior y lemma inferior subiguales, tan largas como 
la espiguilla, glabras o con escasos pelitos a los lados 
de los nervios y los márgenes, 2-nervias, con los 
nervios submarginales, ocasionalmente 4-nervias, con 
los nervios subiguales, tenues. muy poco manifiestos: 
1.6-1,9 X 0.6-0.7 mm, 


densas y 


lemma superior elíptica, 


castaña, brillante, glabra, con pequeñas 


papilas en toda la superficie, disminuyendo hacia los 
bordes, éstos blanquecinos y lisos, encerrando los 
bordes de la pálea: pálea superior similar a la lemma, 
anteras de I— 


levemente menor, estambres 3. 


1.2 mm. ( 


elabra: 


Ariopsis no vista. 


Distribución geográfica y ecología. Venezuela 


Estados Anzoátegui, Monagas, Yaracuy). Colombia 


— — 


Departamento de Arauca) y Bolivia (Departamento 

Santa Cruz). Crece en sabanas, en suelos arenosos, 

bien drenados, entre los 100 y los 1000 m. 
Fenología. | Coleceionada en flor y/o en fruto entre 


los meses de agosto y abril. 


607 


Cialdella et al. 


Volume 93, Number 4 


2006 


Axonopus Serie Suffulti 


ww g' 


— z= 


— —— = 


ZEEPZ 


ww so 


— — 


ww g' 


& 8 — 


— — 1 


— 


SSeS 


m 


de la región 


A. Black (de Renvoize & Cope 3994, MO). —A. Hábito. —B. Detalle 


A 
J. 


Axonopus cuatrecasasit ( 


Figura 5. 


guilla vista del 


L. Espi 


4 


l 


C. Detal 


ligular. 


D. Espiguilla vista del lado de la gluma superior. —1 


F. Antecio superior visto del lado de la lemma. —G. Antecio superior visto del lado de la pálea. 


ón del raquis. 


e de una porci 


lado de la lemma inferior. 


608 


Annals of the 
Missouri Botanical Garden 


Axonopus cuatrecasasti se caracteriza por 


plantas cespitosas con cañas estoloniferas, que 
arraigan en los nudos basales y originan nuevas 


plantas, los nudos son pilosos, el cuello distinguible 


con una línea castaña y el ápice de las láminas 


largamente agudo. El raquis de las ramificaciones es 


elabro, marcadamente triquetro con la quilla muy 
desarrollada protegiendo a las espiguillas. 

Black (1963) al describir la especie señala que las 
espiguillas son glabras. El análisis del material de 
herbario permitió observar una marcada variabilidad 
en la pilosidad de la espiguilla, encontrándose 
ejemplares con espiguillas glabras (Montes 1304a), 
cortamente pilosas, con pelos delicados dispuestos en 


dos líneas a lo largo de las nervaduras y los márgenes 


y lemma inferior (Renvoize & 
Cope 4029, Montes 1305-A) o glabras y pilosas en un 
Steinbach 


de la gluma superior y 
mismo ejemplar (Killeen 1577 y 1725, 
1979, Renvoize & Cope 3994). 

Renvoize (1998) menciona en la descripción de 
Axonopus cuatrecasasti que la gluma superior y lemma 
inferior son 4-nervias. El estudio del material tipo y 
de ejemplares de herbario (entre ellos el citado por el 
autor, Killeen 1725) permitió determinar que la gluma 
superior y la lemma inferior 


son 2- nervias, y muy 


ocasionalmente se hallaron espiguillas con la gluma 


£e 


superior y lemma inferior 4-nervias, los nervios 
siempre cercanos al margen, siendo los más externos 


poco conspicuos. 


BOLIVIA. Santa Cruz: 
. Cordillera, 50 km S of Santa Cruz, Renvoize & Cope 
3947 (LPB, US): Chiquitos, Est. : San Ignacio, 22 km N of San 
dois Killeen 1725 (LPB, SI. US); Andrés Ibanez. W of Santa 
Cruz to e ro Hwy, just S of 5 to Viru Viru Airport, 
a 36. (LPB): Andrés Ibañez. Las Lomas de Arena, 
10 km ! d YPFB refinery, Killeen 1577 (LPB. SI); Andrés 
Ibañez, large sand « lune s S of the town, Renvoize & Cope 4029 
(LPB, MO, US): Santa Cruz to po 0. 
. & Cope 3994 (LPB. MO, US 
NEZUELA. Anzoátegui: 
DINAR s.n. (US 17641950); Sabanas de Trac MI un con 
El Pinal, 3 km NE de San Diego de 
1304a (MO): Hato La 
Montes 1305a (MO). dace 
Purgatorio, Chase 12590 (US). 
Salom, Davidse et al. 20754 NO: US). 


Material adicional examinado. 


near Viru Viru. 


Y ec 2 id de Santo Tomé, 
árboles diseminados, 
Montes 
Iguana, Santa María de Ipire, 

at along Rio 


Yaracuy: 4 km N of 


Cabrutica, alrededores de 


sandy 


Axonopus elegantulus (J. Presl) Hitehe., Contr. 
US. Natl. Herb. 24: 433. 1927. Basónimo: 
Paspalum elegantulum J. Presl, Reliq. Haenk. 


211. 1830. TIPO: T. 
Haenke s.n. (holotipo, PR no visto: isotipos, B no 


visto, BAA 310!, US 2854680! fragm. ex B). 


Perú. sin loc., sin fecha, 


Paspalum attenuatum J. Presl, Reliq. Haenk. 1: 212. 1830. 
lxonopus attenuatus (J. Presl) Hitche., Contr. U.S. Natl. 
Herb. 22(6): 471. 1922. TIPO: Perú. Huánuco, 7. 


Haenke s.n. (holotipo, PR no visto: isotipos, MO 


incluir 


21010731, MO 939545!, US 


yrocedencia no indicada). 


2854681! 


fragm. de 


Jahrb. Syst. 56: 10. 


4. Huancavelica: Prov. 


Pas 110 gregoriense Mez, Bot. 
0 
Pe 


= 


Tayacaja, 
1 des Flusses Mantaro, über San Gregorio, 
1300-1400 m. 4 Abr. 1913, A. 
ene D no 
2854082 

14981 "m ). 


W eberbaue HW 


— 


visto: isolipos, foto ex 8 
. BAA 2150! fragm. ex B, US 2859692 US 


Plantas perennes, gráciles, de 0.25-0.75 m de 
altura, cespitosas, rizomalosas; canas delgadas de 


— 


2mm de diámetro, simples, 2- 6 3-nodes, longi- 


tudinalmente estriadas, levemente comprimidas, con 


surco lateral conspicuo, glabras; nudos castaños, 


comprimidos. glabros, a veces con algunos pelos: 
vainas menores que los entrenudos, conspicuamente 
estriadas, pajizas, cara externa con pelos blanquecinos, 
cortos y delgados, ralos, adpresos, cara interna glabra, 


con los bordes lisos. glabros a densamente vilosos: 


nervios de las vainas se continúan en la lámina; cuello 


con pelos hacia los márgenes: lig 
0.4-0.5 mm, 


en la base, luego pajizas, con borde blanquecino: 


conspicuo, ilas 


membranáceo-pestanosas, de castañas 
contralígula y aurículas ausentes; láminas lineares, 
rígidas, de 7-15 X 0.2-0.25(-0.6) cm, predominante- 
mente basales en la planta, plegadas en la base, luego 
planas, los bordes planos a levemente revolutos. el 
ápice agudo o cortamente acuminado, retuso, hendido 
en hojas viejas, longitudinalmente estriadas en ambas 
caras, 


elabras. con los márgenes escabriúsculos, en 


ocasiones pilosos hacia la base. Cañas floríferas con 
inflorescencia terminal subdigitada, con pedúnculo 
cilíndrico, ligeramente comprimido, longitudinalmente 
estriado, con surco lateral menos conspicuo que en las 
3-6(-10) 


6—11(—13) em, siendo los basales alternos y 


cañas, glabro; inflorescencia formada por 


“acimos de 


distantes, los apicales más próximos entre sí; raquis de 


los racimos ondulado, pajizo, briquetro, cara dorsal 


plana, de 0.4-0.5 mm de ancho, cara ventral con 


costilla conspicua, glabra o con algunos pelos cortos, 
escabriúseulo y con pelos largos, 1-2 mm, rígidos, 
dispersos sobre los ángulos, más densos cerca de la 
base de los pedicelos; pedicelos muy breves, ca. 
0.2 mm, aplanados, escabriúsculos en los ángulos, con 


similares a los del 


algunos pelos largos raquis. 
Espiguillas de contorno elíptico, de 1.8-2.1 0.6— 


0.8 mm. 


subagudas, ocráceas, alternas y densamente distribui- 


dorsiventralmente comprimidas, agudas o 
das en 2 hileras a lo largo del raquis, de tal forma que 
cada espiguilla resulta opuesta a una de la otra hilera 


en 1/3-1/2 


— 


| longitud; gluma superior y lemma 


inferior tan ioc como el antecio, con ápice 
cortamente acuminado, membranáceo-hialinas, con 


tintes violáceos, gluma superior 2—3-nervia, cuando 


3-nervia el nervio medio a veces solo notable hacia el 


Volume 93, Number 4 
2006 


Cialdella et al. 
Axonopus Serie Suffulti 


ápice, lemma inferior 2-nervia, ambas glabras; lemma 


superior oblongo-elíptica, 1.7-2.1 mm, castaña, bri- 
llante, glabra, escabrosa en el ápice, excepcionalmente 
sin escabrosidades, con densas y pequeñas papilas en 
toda la superficie, disminuyendo hacia los bordes, éstos 
blanquecinos y lisos, encerrando los bordes de la pálea: 
pálea superior similar a la lemma, ligeramente menor. 
glabra. Cariopsis no vista. 


Distribución geográfica y ecología. Habita en los 


andes de Perú entre los 1000-3400 m, 


descubiertos y rocosos. 


sobre suelos 
Fenología. Encontrada con flores y/o frutos entre 
los meses de diciembre y abril (junio). 
Axonopus elegantulus es similar a A. arcuatus. 
distinguiéndose esta última especie por tener las hojas 
arcuadas, dispuestas en la zona basal y media de la 
planta, por sus racimos gráciles, delgados y por las 
espiguillas de ápice agudo, más angostas (0.5-0.6 mm 
de ancho) y más espaciadas en el raquis, resultando 
opuestas solamente en 1/4—1/3 de su longitud. En A. 
elegantulus, en cambio, las hojas se encuentran 
concentradas en la base de la planta, los racimos 
son de aspecto más robusto M más compacto, por la 
densa disposición de las espiguillas a lo largo del 
raquis, opuestas en 1/3-1/2 de su longitud. 
Axonopus elegantulus también puede confundirse 
distinguiéndose esta última 


con A. polydactylus, 


especie por sus cañas plurinodes, sus láminas de 


20-60 em de largo, el raquis con frecuencia hispi— 


oa 


dulo. el número de los racimos frecuentemente entre 
11 y 17 
a lo largo del raquis. 
En el 


detalladas anotaciones de Chase, que además incluye 


y la distribución más laxa de las espiguillas 


herbario US existe una cartulina con 


un duplicado de un ejemplar depositado en B y otro 


— 


en PR con los siguientes datos “Paspalum pulchellum 


nov. sp., inéd., Peruvia, Haenke”. Este nombre no fue 
publicado válidamente por Presl. Chase indica que en 
Praga no encontró el material tipo determinado. por 
Presl como Paspalum | elegantulum. Sin embargo 
señala que los especímenes de B y PR determinados 
como P. pulchellum con letra de J. Presl concuerdan 
con el protólogo de P. elegantulum. De acuerdo con 
Chase. en este trabajo se considera que los ejemplares 
mencionados depositados en US corresponden a la 
colección tipo de P. elegantulum. Un isotipo de la 
misma colección se halla depositado en BAA. 


Mez (1921), 


considerado sinónimo de Axonopus elegantulus por 


al describir Paspalum gregoriense, 


Hitchcock (1927), señala que las espiguillas son de 


15 mm. El análisis del material tipo permitió 


determinar que las espiguillas miden 1.9-2 mm. 
Tovar (1993) considera a Axonopus attenuatus como 
distin- 


un taxón independiente de A. elegantulus, 


guiéndola por el mayor número de racimos (8-11) y 
las espiguillas de menor tamaño (1.6-2 mm de largo). 
El estudio del material tipo y de ejemplares de 
herbario de ambas especies permile confirmar que 
existe una continuidad en los caracteres mencionados 
y que se trata de una misma entidad. Asimismo se 
observó que la pilosidad del raquis es marcadamente 
variable, desde conspicuamente piloso a lo largo de 
como en los ejemplares Hutchinson & 

Tovar 1407 y Macbride 3324, 


con pelos dispuestos más 


los ángulos, 
Wright 3962, 


esparcidamente piloso. 


hasta 


densamente en la base de los pedicelos (en Kunkel 


623, Macbride 3760, Asplund 13556 y Fosberg 
27824): e€xcep« ionalmente el raquis puede prese ntar 


muy escasos pelos a la altura de la base de los 
pedicelos, como en el ejemplar van der Werff et al. 
6406 y Tovar 188. De acuerdo con Anton (1982), A. 
attenuatus representaría un extremo de A. elegantulus 


con espiguillas menores, más distanciadas y pedicelos 


menos pilosos. 
Hitchcock (1922, 1927) cita la presencia de 
Axonopus elegantulus (bajo A. attenuatus) para 


Guyana, sin embargo los materiales citados por este 
autor corresponden a A. flabelliformis. 

Jørgensen y Ulloa (1994) y Zuloaga y Morrone (1999) 
citan erróneamente a Axonopus elegantulus para Ecua- 
Solís 20497 


(US), material que por sus características pertenece a la 


dor sobre la base del ejemplar Acosta 


— 


serie Barbigeri (sensu Black, 1963), y corresponde a . 
Black (1963) cita 4. 


ejemplar 


elegantulus para Loja, 
Espinosa 308 


la 


scoparius. 


Ecuador. mencionando el 


depositado en el herbario de US. El mismo no 
podido ser localizado por lo que la presencia de esta 
especie en Ecuador debe ser confirmada. 

Material examinado. PERU. Amazonas: Bongara, betw. 
Río Uteubamba and Sh a tkm from Campo. 
Ingenio, Hutchison et Wright 3962 (MO, Us). Apurimac: 
Grau. Hda. 
\mpay, 
(MO). Cajamare a: Hda. Sondor, 
(US): on Rio 1 ere 20 k 
bamba. Fosberg 27824 (MO). Cuzco: 20 km N 

. Hitchcock 6100 (NY, US). 
Podar 1407 
E de 


Namora, 
Jaen, Tabaconas, 
Huancave ‘lie a: Mejo- 


; Prov. Aia 'avelica, 


a 
9 9 9 1905 (MO, ve) 


asco: den pampa, Palmazo, o. 1 Na avarra, van 
r Werff et al. 6406 (MO, US). Sin vie amento: Mito, 
7 3270 (US), 2 3324 (US), Macbride et 
Featherstone 1498 (US); Yanano. m 3700 (US). 
6. Axonopus flabelliformis Swallen, Bull. Torrey 
Bot. Club 75: 82. 1948. TIPO: Guyana. 
Wismar and Rockstone. Demerara 
in si 30 Dic. 1919—Ene. 1920, A. 
S. Hitchcock 17275 (holotipo, US 103857 
isotipo, foto ex US en Sl). Figuras 1D, 6. 


Halfway 
station betw. 


Hiver, 


t^ 


610 Annals of the 
Missouri Botanical Garden 
Ixonopus purpurellus Swallen, Bull. Torrey Bot. Club 75: 83. alternos levemente distantes entre sí y los apicales 
1948 TIPO: : Surinam. Km 68. vic. echie O, on wel simples, más cercanos: raquis de los racimos 
sandy soil, 19 O VVV notablemente ondulado a casi recto, pajizo en la zona 
: O ablemente i a casi recto, pajizo en la zon: 
24997 (holotipo, Us 1713731: isol pag t 


Ds foto 
! 


. loto ex NY en Us 


1 48 o TIP O: 
frequent, rocky ds Pme 14 May 1944, n. Maguire 
& D. B. Fanshawe 23454 (holotipo, US 171387! 

ven SIL F no visto, NY -3406104/, A 


loc all 


isolipos, foto ex US 

US 191499] !). 
Axonopus tamayonis Luces, 
2]. 1953. TIPO: 
i Elena, en el cerro s 11 Fel 
Tamayo 2747 (holotipo, VEN 937941; 
VEN en US!, F no MO!, US 


Bol. Soc. 


Venezuela. 


Venez. Ci. Nat. 15(80): 
Bolívar. Gran Sabana. 

» 1946. F. 
isotipos, foto ex 


1910959!, US 


vislo, 


— flabe 1 var. 1 v A. Black, Advanc- 
x Frontiers. Pl. > 145 Tipo: Guayana 
E rancesa. Ex. Sinamay lo Trac 0 km 13. campo 
29 Oct. 1954, Black & R. Klein 54-1739 


(holotipo, TAN no visto; 1 0 no visto. US 
23033051, . 
Axonopus flabe p ne var. a ire G. A. Bla 
Frontiers ña 6. 19 TIPO: 
15 co P T. ps 19 


Isollpos, 
loto ex US en SI!) 

ck, Advancing 
Surinam. Poika 
Savanna, 361 (holotipo. US 


2184143!) 


Plantas perennes, cespitosas, de 0.55—1(-2) m de 
altura, con rizoma corto y robusto, cubierto por vainas 
fibrosas; cañas generalmente aplanadas de 2-4 mm 
de diámetro, 2- 6 3-nodes, simples, longitudinalmente 
estriadas, glabras, cubiertas por las vainas basales 
equitantes; nudos castaños, levemente comprimidos, 
glabros a densamente vilosos; vainas conspicuamente 
estriadas, pajizas, glabras a vilosas, con frecuencia 
más densamente cerca del margen y hacia el cuello, 


dispuestas en forma notablemente dística, condupli- 


cadas, comprimidas lateralmente, de manera que las 
láminas quedan dispuestas en 2 hileras, el conjunto 
de aspecto flabeliforme; nervios de las vainas se 
continúan en la lámina: cue Mo i unpe rceptible: lí ru ilas 
0.5 mm. 


membranácea-pestanosas, brevisimas, Ca. 


castaño rojizo, con margen blanquecino: contraligula 


y aurículas ausentes; láminas lineares, rígidas, 10— 
20(-60) X 0.5-0.9 cm, dispuestas en la base v 


rein mediana de las cañas, plegadas en toda su 
preg 


longitud o solamente en la base y planas hacia el 
ápice, de ápice subagudo u obtuso, base continua con 
la vaina, longitudinalmente estriadas en ambas caras. 
con nervio central y laterales conspicuos, glabras en 
ambas caras o bien escabriúsculas en el epifilo 3 
densamente vilosas en el hipofilo, con los márgenes 


lisos, 


glabros o hispídulos hacia el ápice. o con 


escasos pelos de 2.5-3 mm cerca de la base o en toda 
su extensión. Cañas floríferas con 1-3 inflorescencias. 


naciendo en el último nudo caulinar, pedúnculos 


cilíndricos o aplanados, longitudinalmente estriados. 
elabros; inflorescencia con 6-20(—30) racimos de 5— 
1520) em. ramificados, 


los basales en ocasiones 


central, verdoso en los márgenes, triquetro, cara 
dorsal plana, (0.3-)0.6-0.8(—1.3) mm de ancho, cara 
ventral con costilla conspicua, hispídulo en toda su 
superficie, más conspicuamente sobre los ángulos, sin 
pelos largos, o con pelos de 1-1.5 mm, situados sobre 
os ángulos a ambos lados de la base de algunos 


pedicelos; pedicelos de (0.1—)0.2-1 mm, conspicua- 
mente aplanados, glabros a pubérulos. Espiguillas de 
(1.3)1.6-2.2 X 0.6- 


comprimidas, ocráceas, 


contorno oblongo-elíptico, 
| : | 4 
VIUULOSIVUIILIEGULITIEICIIEL 


0.7 mm. 
alternas, dispuestas en 2 hileras a lo largo del raquis, 
de tal forma que cada espiguilla se opone a otra de la 
otra hilera en 0-2/3 de su longitud: gluma superior de 


1.2-1.5( 


cio, obtusa, 


2.2) mm, generalmente menor que el ante- 


muy rara vez igual y aguda, 2-nervia, 


ocasionalmente 3-nervia, glabra o pubescente con 


algunos pelitos cortos a ambos lados de los nervios: 


lemma inferior de 1.3-2 mm. membranáceo-hialina. 


ligeramente menor o tan larga como el antecio, 2- 
nervia, glabra o pubescente; lemma superior de 1.6- 


1.8 mm, aguda, castaña, poco brillante, glabra. 
[e] I [e] 


ocasionalmente con escasos y cortos pelitos en su 


ápice. con densas y diminutas papilas en toda la 


superficie, disminuyendo hacia los bordes. éstos 


blanquecinos y de superficie lisa, encerrando los 
bordes de la pálea; pálea superior de igual consis- 
lencia y aspecto que la lemma superior, ligeramente 


menor que la lemma, glabra. Cariopsis no vista. 


— 20 (Hickenbick et al.. 
1975, sub Axonopus flabelliformis var. camporum). 
Bolivia (La 


Guyana, Guayana 


Vúmero cromosómico, 
Distribución geográfica ecología. 
Brasil 


Francesa, Surinam y 


Paz), (Amazonas y Roraima). 


— 


Venezuela (Bolivar y T.F. de 
Judziewicz (1990) la cita también para los 


Brasil, 


menciona para Colombia, 


Amazonas). 
estados de Pará y A mientras 
Giraldo-Cañas (1999) la 


en el departamento Meta. Crece en sabanas abiertas, 


que 


sobre suelo arenoso, entre 


los 100-1600 m. 


Coleccionada en flor y/o fruto durante 


o entre rocas emergentes, 


Fenología. 
lodo el año. 

Esta especie es fácilmente reconocible por la 
disposición marcadamente equitante de sus vainas 
basales. 

Axonopus flabelliformis es afín a A. polydactylus, 
distinguiéndose esta última especie por la disposición 
de las cañas, no 


de sus vainas basales alrededor 


equitante. Asimismo, en A. polydactylus la gluma 


superior es mayor que el antecio y el ápice de los 


antecios es usualmente cortamente escabroso, mien- 


Volume 93, Number 4 Cialdella et 
2006 Axonopus A Suffulti 


0,5 mm 


0,5 mm 


SOS oe 


mo T 


0,5 mm 


0,5 mm 


dn flabelliformis Sw: allen a Harrison & Persand 1049, NY). —A. Habito. —B. De Aalle de la región 
talle de una porción del raquis. —D. Es Me vista del lado de ji 1 superior. “spiguilla vista del 
Antecio superior visto del lado de la lemn ecio superior visto ve 1 lado de la pá 


Figura 6. 
ligular. —C. D 
lado de la lemma inferior. —F. ; 


.—G. | 


Annals 
151 1 ae Garden 


A. flabelliformis 


menor que el antecio y el antecio superior es liso. 


tras que en la gluma superior es 

Axonopus flabelliformis también se halla estrecha- 
mente relacionada a A. magallanesiae, principalmente 
por la disposición equitante de las vainas. Sin 
embargo, A. magallanesiae es una planta más pequeña 
(hasta de 50 em de altura), presenta inflorescencias 


de 4-8 em. sus 


compuestas por pocos racimos (3—5 


espiguillas son más grandes, de 2.2-2.4(-2.5) mm, y 
la gluma superior y la lemma inferior son 4-5-nervias, 
con los nervios laterales conspicuos. 

Black (1963) reconoce dentro de Axonopus flabelli- 
formis tres variedades: la variedad tipo, la variedad 
camporum. caracterizada por tener espiguillas espar- 
variedad 


cidamente  pilosas y la 


presentar las vainas ligeramente. equitantes, raquis 
de los racimos con dos a cinco pelos y espiguillas 


1.3-1.4 


tipo y de herbario en general permitió determinar que 


pequeñas (de mm). El análisis del material 


existe una marcada variabilidad en los caracteres 


morfológicos anteriormente mencionados y por tal 
razón eslas variedades se incluyen en la sinonimia de 
A. flabelliformis. 
Axonopus flabelliformis var. decipiens fue incluida 
en la sinonimia de A. polydactylus por Anton (1982). 
Al estudiar un fragmento del holotipo de la primera, 
se comprobó que el raquis es más ancho, de 0.9— 
| mm versus 0.3-0.4 mm en A 
que permite sinonimizarla con A. flabelliformis. 
Por otra parte, Axonopus purpurellus fue consider- 
ada sinónimo de A. polydactylus por Anton (1982). El 
holotipo de 4. permitió 


estudio del purpurellus 


comprobar que las vainas basales son fuertemente 
equitantes y el raquis de los racimos también es más 
ancho que en A. polydactylus (0.7—1 mm de ancho). 


Por las características anteriormente mencionadas se 


considera que A. purpurellus debe incluirse e 
Jlabelliformis, 
criterio de Judziewicz (1990) y de Zuloaga y Morrone 
(2003). 


Swallen (1948) al describir Axonopus katetukensis 


a 


sinonimia de A. coincidiendo: con el 


señala que se halla estrechamente relacionada a A. 
flabelliformis, diferenciandola por las espiguillas más 
grandes (2 mm vs. 1.6 mm), las láminas más largas 
(15-20 cm vs. 9-15 cm) y las vainas más cortas y con 
vilosos. Del análisis del 


márgenes densamente 


material tipo de A. kaletukensis se pudo comprobar 
que el rango de las espiguillas es de 1.6-2 mm. 
siendo este caracter continuo con lo observado en 4. 


flabelliformis. 


las dimensiones de las laminas y de 


Además, como se indicó anteriormente. 
las vainas. asi 
como la pilosidad de estas últimas no son caracteres 
válidos para separar ambos taxones. 
Black (1963) y Renvoize (1998) 


Axonopus elegantulus el ejemplar Buchtien 12 co- 


citan bajo 


decipiens por 


. polydactylus, carácter 


leccionado en La Paz, Bolivia, depositado en BAA, 
IAN v US. Sin embargo, el estudio de un duplicado de 
dicho ejemplar procedente de MO, permitió confirmar 
que las vainas basales son marcadamente equitantes, 
numerosos racimos, 
flabelliformis. Black 


(1963) menciona que en el mencionado material las 


carácter que. junto con los 


permite identificarlo como 
espiguillas inmaduras son pilosas, mientras que las 
maduras son glabras. Se ha podido observar que las 
espiguillas maduras presentan densos pelos tubercu- 
lados en la gluma superior y la lemma inferior y 
antecios castaños. 

Anton (1982) 


esta especie es marcadamente polimorfa en cuanto la 


De acuerdo con Judziewicz (1990) y 


pilosidad de las vainas, de los raquis y de las 


espiguillas. Un ejemplo de esta variabilidad, se 
observó en el ejemplar Williams 1000, procedente 
de La Paz, Bolivia, el cual consta de dos plantas, una 
de ellas con espiguillas glabras y la otra con 
espiguillas pilosas. 

flabelliformis 


ocasionalmente presentan ramificaciones en 


Las inflorescencias de Axonopus 
la base 
racimos inferiores, característica 
observada en Hoock 412, 
32621 y Boerboom 8764. 


de los que fue 


Maguire & Fanshawe 


BOLIVIA. La Paz: 


Uc s is a media 


Material adicional e xaminado. Prov. 
| Senda Apolo-San 5 de 
hora antes de llegar a Naranjal. Miranda et al. (LPB) 
Prov. Tamayo, near Apolo, Smith 1000 (NY); Prov. E arecaja, 
San Carlos, San José, Mapiri Region, Buchtien 12 (BAA, 


BR ASIL. Amazonas: Tunui, Rio leana, Luetzelburg 
2410 (R-50055); Icana. no Led da cachoeira, Luetzel- 
22954 (R) 
det D7 i 
ANA FRANCESA. Rte. Sinnamary, km 14- 
ink ue (US). ak 412 (US). Hoock 414a (US) M 
N). 


e Roraima: Serra Tepequem, Maguire et 


— 


f: > Ren Cuyuni-Mazaruni Region: 7 ¡km 
village, after descent from S face, Mt. Waleliwatipu, Me 
Dowell et Hughs 2972 (US); Pakaraima Mins.. 
8.0 km NE e on Partang R. tributary, 0.5 
Hoffman et al. 1781 
Cuthbert 
W Demerara Regie x S of 


N Paruima 


zx 


basecamp 
km N 
; Linden Hwy near turnoff to St. 
s Mission, legi et al. 3 (US). Essequibo Isl- 
Georgetown on £ 

Linden Hwy.. ca. 7 mi. ON of Loo River. 
7459 (US). Mahler: Berbice Region: St. 
cut-off. 2 km E of the 


Peterson & Gopaul 
Cuthbert's Mis- 
sion road Timahri-Linden Hwy., 
Judziewiez & Lall 5410 (US). 1 
Kaieteur N | km W from W end of airstrip. Henkel & 
Williams 2309 (US); Savanna at S base of Mt. Kopinang along 
trail from Kopinang to Orinduik, Hahn et al. 4428 (US). 
Berbice River, 


n 
Potaro-Siparuni 1 


Harrison & Persand 

1049 E 

SURINAM. Nat. Res. “Brinchheuvel”, 
5) 


Waranama Ranch: 
49 (NY, US) 


Teunissen et 1 
11548 (US); Sabakoe Savannah, Hoock s.n. k 
Savanna sectio O, Pulle 141 Lt 
Geyskes 128 (US): 
Blahawatra, near 
Poika -i Semple 381 (l 
1265 (US). 


schut 
Zanderij savanne, 
Zanderij l, Hoock 1268 (US); 
Yodensavanne, ari LBB8764 (US): 
S): Sabakoe ' 


westzijde, 


Savannah. Hoock 


Volume 93, Number 4 
2006 


Cialdella et al. 
Axonopus Serie Suffulti 


613 


VENEZUELA. 1 nd Dpto. Atures, al W del medio 
Río Asita, Huber 3 EN); Dpto. Atabapo, 10 km al W de 
la Serranía Asisa, pan 6217 (VEN). Bolívar: Rocío, entre 

San Francisco de Yuruaní y Chirimatá, ca. 10-15 5 k 
de San Ignacio de Yuruaní, 
1 € 
Liesner 127569 (VEN); Gran Sabana, Parq. Nac. Canaima, 
Kama-Meru, 109 km de E pasando por El Dorado, 
Río Kama, Sorone et al. 4761 (SI); 7 km N of La Ciudadela, 
Davidse et al. < 2 (VE Cumbre de Cerro „ 
a lo largo del 7 Szezerbanari, Steyermark & Du 
113164 (VEN); Mun. Sucre, bes 1108 (VEN). 


isterville 


7. Axonopus gracilis C. A. Black, Mem. New York 
Bot. Gard. 9: 254. 1957. TIPO: Venezuela. 
Amazonas: Cerro Sipapo (Paráque), SE ridge 


and savanna slopes, infrequent, 1700 m, 20 Dic. 
1948, B. Maguire & L. Politi 27818 (holotipo, 
NY no visto, foto SI! ex NY; US 
2236051! fragm. ex NY). 


isolipo, 


Plantas perennes, cespitosas, ca. 1 m de altura, con 


rizomas delgados y cortos: cañas aplanadas de 2.5— 


3 mm de ancho, simples, plurinodes, pajizas, lon- 
gitudinalmente estriadas, glabras; nudos castaños, 


elabros; vainas conspicuamente estriadas, pajizas, con 
tintes rosados hacia el ápice, los bordes hialinos, 
lisos, glabras; cuello inconspicuo; lígulas membraná- 


ceas, 0.2—0.4 mm, con márgen densamente pestañoso, 


jay] 


el resto de la superficie glabra; nervios de las vainas 


se continúan en la lámina; aurículas y lígula externa o 


contralígula ausentes; láminas lineares, rígidas, 1-30 
X 0.1—0.4 cm, predominantemente basales, plegadas, 
planas hacia el ápice, de ápice agudo, base continua 
con la vaina, sin pseudopecíolo, longitudinalmente 


estriadas en ambas caras, con nervio central v 


laterales conspicuos, glabras, con margen escabriüs- 
culo, viloso en la zona basal. Cañas floríferas con 1 62 
inflorescencias, cada una de ellas con pedúnculo 


aplanado, glabro, longitudinalmente estriado; inflo- 


rescencias con 3-4 racimos de 4-12 em, siendo los 


basales alternos levemente distantes entre sí los 
apicales más cercanos: raquis de los racimos 
ondulado, triquetro, violáceo o verdoso, cara dorsal 


plana, 0.3-0.4 mm de ancho, cara ventral con costilla 
conspicua, escabriúsculo, más conspicuamente sobre 
los ángulos; pedicelos muy breves, hasta de 0.2 mm 
de largo, aplanados, escabriúsculos en los ángulos. 
Espiguillas de contorno oblongo-elíptico, 2.6—2.0(—3) 

0.7-0.9 mm, dorsiventralmente comprimidas, ob- 
tusas o cortamente apiculadas, ocráceas, alternas, 
dispuestas en 2 hileras sobre el raquis: gluma superior 
y lemma inferior mayores que el antecio, 2.6—3 mm, 


membranáceo-hialinas, 4-nervias, los nervios cerca- 


nos a ambos márgenes, rara vez con nervio central 


2. 


castaña, poco brillante, glabra, con 


=~ 


apenas conspicuo, glabras; lemma superior ( 


y 


2.8 mm de largo, 


densas y diminutas papilas en toda la superficie, 


disminuyendo hacia los bordes, éstos blanquecinos y 
de superficie lisa, encerrando los bordes de la pálea; 
pálea superior de igual consistencia y aspecto que la 
lemma superior, ligeramente menor que la lemma, 
glabra. Cariopsis no vista. 
Distribución geográfica y ecología. | Especie poco 


frecuente, hasta el momento es conocida por la 


colección tipo del Cerro Sipapo (Venezuela), en 
Fenología. Hallada en flor y/o fruto en el mes de 
diciembre. 

Esta especie fue considerada bajo la sinonimia de 
Axonopus polydactylus (Anton, 1982; Zuloaga & 


Morrone, 2003). Sin embargo, A. gracilis posee 
inflorescencias E por 3—4 racimos, de 4— 
12 em, espiguillas de 2.6-2.8(—3) mm, gluma superior 
y lemma inferior o con el nervio medio 
ausente. Axonopus polydactylus, en cambio, tiene 
(5-)11—17(-30) racimos, de 7.5-18(223) cm, las 
espiguillas son de 1.2-1.6(-2.3) mm, la gluma 


superior es 2(3)-nervia y la lemma inferior es 2- 
nervia. 

Black (1963) considera a Axonopus gracilis dentro 
de la serie Axonopus. No obstante por las caracter- 
ísticas de su hábito y por la morfología de la 
espiguilla, esta especie debe incluirse dentro de la 
serie Suffulti, coincidiendo con el criterio adoptado 
por Giraldo-Cañas (2000a) 

8. Axonopus hoehnei G. A. 


Black, Advancing 


Frontiers Pl. Sci. 5: 143. 1963. TIPO: Brasil. 
Pará: Lageado, Rio Tapajóz, Feb. 1912, F. C. 
Hoehne 5310 (184) (holotipo, IAN no visto; 


isolipo, US 22366057! fragm. ex IAN 
Plantas perennes, 
0.60(—1.25) m 
delgados; 
2- ó 


longitudinalmente estradas, con surco lateral poco 


levemente cespitosas, de 0.50- 


de altura, con rizomas breves y 


cañas cilíndricas, de 1-2 mm de diámetro, 


3-nodes, simples, gráciles, pajizo verdoso, 


notable, glabras; nudos castaños, glabros; vainas 


menores que los entrenudos, conspicuamente estria- 
das, castañas, glabras, de bordes lisos; nervios de las 
vainas se continúan en la lámina; cuello conspicuo, 


elabro, con una línea triangular, castaña a ambos 


ados del márgen; aurículas y contralígula ausentes; 
ligulas membranáceas, cortamente pestañosas en el 
0.1—0.3 mm, 
blanquecino: láminas engostamente lineares, rígidas, 
10-15 0.3—0.4. 


plegadas en la base, la base continua con la vaina, 


borde. de castaño rojizo, el borde 


cm, predominantemente basales, 


planas hacia el ápice. éste largamente acuminado, 


longitudinalmente estriadas en ambas 


= 


caras, con 
nervio central y laterales conspicuos, glabras. Cañas 


floríferas con 1-2 inflorescencias naciendo del último 


Annals of the 
Missouri Botanical Garden 


nudo caulinar, cada una de ellas con pedúnculo 


delgado y aplanado, glabro, longitudinalmente es- 


handi inflorescencia formada por 4-8(213) racimos 


raquis sinuoso, pajizo, Iriquetro, cara 


dorsal plana, 0.7-0.8 mm de ancho, cara ventral con 


costilla conspicua, híspido, más conspicuamente 


sobre los ángulos, sin pelos largos; pedicelos de 
0.1—0.2 


el ápice. Espiguillas de contorno oblongo-elíptico, de 


1.6—1.7(—1.8) X 0.7-0.9 mm. dorsiventralmente com- 


mm, hispidos, aplanados, ensanchados hacia 


primidas, oeráceas, alternas, dispuestas en 2 hileras 
sobre el raquis, de tal forma que cada espiguilla se 
opone a otra de la otra hilera en 1/4 de su longitud: 


gluma superior y lemma inferior membranáceo- 


hialinas, 1.6-1.8 mm, gluma superior menor que el 


antecio, obtusa, rara vez igual, lemma inferior igual 


que el antecio, aguda, ambas 2(-3)-nervias, los 


nervios conspicuos, con pelitos delgados, rígidos, 


dispersos sobre toda la superficie, principalmente 


hacia ambos lados de los nervios y más densamente 
5-1.7 


castaña, lustrosa, glabra, con micropelos bicelulares 


hacia la base; lemma superior de !. mm. 


y con densas y diminutas papilas en toda la superficie. 
disminuye ndo hi ac la los bordes 7 éstos blanquecinos y 
de superficie lisa, encerrando los bordes de la pálea: 


yalea superior de igual consistencia y aspecto que la 
o E 


lemma superior, ligeramente menor que la lemma, 
glabra. Cariopsis no vista. 
Distribución geográfica y ecología. Hasta el 


momento sólo es conocida para estado de Pará, Brasil. 
1 lugares bajos, periodicamente inundados, 
sles arenosos, y asociados a arbustos y 
pequeños árboles entre los 400-500 m. 
Fenología. Coleccionada en flor y/o fruto entre 
los meses de diciembre y febrero. 
Zuloaga y Morrone (2003), siguiendo el criterio de 
Anton (1982), 


sinonimia de A, pennellii. Anton (1982) considera que 


tratan a Axonopus hoehnet bajo la 
ambas entidades serían dos variantes locales de una 
misma especie, meramente distintas en el grado de 
pubescencia que presentan las partes vegetativas. Sin 
embargo, A. hoehnet presenta hojas con láminas 
lineares de 10-15 em X 3-4 mm, rígidas, con vainas 
de dorso subredondeado, cuello distinguible con una 
región triangular castaña, inflorescencias con raquis 


(0.6-0.8 mm 


con pelos distribuidos principal- 


más ancho de ancho) y espiguillas 


obtusas, pilosas, 


mente hacia ambos lados de los nervios y más 


densamente hacia la base. Consecuentemente, en 
este trabajo se mantiene a A. hoehnet como un taxón 
independiente de A. pennellii. 

Axonopus hoehnet resulta también afin a A. cilia- 
tifolius, distinguiéndose esta última especie por su 


hábito densamente cespitoso, sus pedicelos mayores 


(0.4-0.6 mm de largo) y sus espiguillas más grandes 


(2.1-2.2 mm de largo). Hasta el presente, ciliati- 
folius ha sido coleccionada sólo en Belice. 

Material a BRASIL. Para: Serra do Ca- 
chimbo, Pires et al. 61564, 0178 (US); region of Missão 


Mundurukú le ca. 2 km N of the Rio Curua, 


Velha 
ian 10908 (MO). 


Ann. 


TIPO: Panamá. Coclé. area 


9. po jeanyae Davidse, Missouri Bot. 


Gard. 74: 421. 1987. 
Blanco del Norte, 


betw. 1 Caño Sucio and 


Chorro del Río Tife, evergreen. forest, 200— 
100 m. spray basin of waterfall, 3 Feb. 1983, G. 


Davidse & C. W. Hamilton 23570 (holotipo, MO 
3475959!: isolipos, foto ex MO en SI! ISC no 
visto, PMA no visto, US 3260346!). Figura IE. 
cespilosas, de 50- 


Plantas eráciles, 
e 


E 


(5 em de altura, con rizomas cortos y delgados: 


perennes, 
cañas 


delgadas, 2(—3)-nodes, simples, pajizas, longitudinal- 


mente estriadas, glabras: nudos NEegruzcos, vilosos; 


vainas menores que los entrenudos, estriadas, pajizas, 


totalmente glabras, ciliadas en el margen: nervios de 


las vainas se continúan en la lámina: cuello incon- 


spicuo: ligulas membranáceo-ciliadas, ca. 0.5 mm; 


aurículas y contralígula ausentes: láminas lineares, 


rígidas, hasta de 25 0.2-0.3 em, predominante- 


mente basales, plegadas en toda su longitud o planas 


hacia el ápice, de ápice obtuso, escabroso, base 


continua con la vaina, longitudinalmente estriadas en 


ambas caras, glabras o papiloso-ciliadas hacia la base. 


Cañas florfferas generalmente con 2 inflorescencias 


en el último nudo caulinar, cada una de ellas con 


pedúnculo delgado, glabro, longitudinalmente es- 


triado; inflorescencia con 2-5 racimos de 4-11 em. 


siendo los basales alternos levemente distantes entre 
sí v los apicales más cercanos; 


raquis levemente 


ondulado, pajizo verdoso, triquetro, cara dorsal plana, 


0.1-0.6 mm de ancho, cara ventral con costilla 


conspicua, escabroso, sin pelos largos: pedicelos de 
Espiguillas de contorno 


lisos. elabros. 
o 


).2-0.5 mm. 
oblongo-elíptico, de 1.9-2.5 X 0.7-0.8 mm, dorsi- 
ventralmente comprimidas, ocráceas, alternas, dis- 
puestas laxamente en 2 hileras sobre el raquis, de tal 
forma que cada espiguilla no se opone a otra de la otra 
hilera; gluma superior y lemma inferior membranáceo- 
hialinas, 1.9-2.5 mm, ambas mayores que el antecio, 
4—5-nervias, con pelitos delgados a ambos lados de 
los nervios; lemma superior de 1.8-2.4 mm, castaña. 
brillante, con escasos pelitos cortos en el ápice de la 
lemma, con densas y diminutas papilas en toda la 
hacia los bordes. éstos 


superficie, disminuyendo 


blanquecinos y de superficie lisa, encerrando los 
bordes de la páleaz pálea superior de igual consis- 


tencia y aspecto que la lemma superior, ligeramente 


Volume 93, Number 4 


Cialdella et al. 
Axonopus Serie Suffulti 


menor que la lemma, glabra; estambres 3, anteras de 


11.5 mm, purpúreas. Cariopsis no vista. 


Distribución geográfica y ecología. Sólo ha sido 

encontrada en Panamá, entre los 200 y los 400 m, en 

lugares húmedos, asociados a caídas de agua. 
Fenología. | Mallada en flor y/o fruto en el mes de 

febrero. 

ciliatifolius, 

Esta 


última se distingue por sus nudos glabros, los nervios 


Axonopus jeanyae resulta afín a A. 
especie hasta el momento endémica de Belice. 
de la gluma superior y la lemma inferior son menos 
conspicuos, la gluma superior es 2-nervia y menor que 
el antecio, rara vez igual, mientras que la lemma 
inferior es 2(3)-nervia y es igual o mayor que el 
en A. ciliatifolius las láminas son 
14(-25) X 0.1-0.2 em, de ápice acumi- 


Asimismo, 
de (5—)12- 


nado. 


antecio. 


10. Axonopus magallanesiae Giraldo-Cañas, Cal- 


dasia 22: 23. 2000. TIPO: Venezuela. 


Distrito Cedeño, meseta de Jaua, sector centro- 


Bolívar: 


cabeceras del río Marajano, afluente 
del río Cácaro, 1750-1800 m. 20 Nov. 1989, 0. 
Huber 12995 (holotipo, COL no visto; isotipos, 
MO-5331971!, MYF no visto. =f ex MO en SII, 
SI!, VEN no visto). Figuras 24—C, 7. 


meridional, 


Plantas perennes, densamente cespilosas, de 30— 
50 em de altura, con rizomas gruesos y cortos; cafias 


aplanadas, de 2 mm de diámetro, 2-nodes, delgadas, 


simples, erguidas, pajizas, longitudinalmente estria- 


das, con surco lateral conspicuo, glabras: nudos 


castaño rojizo. vilosos; vainas levemente mayores 
que los entrenudos, conspicuamente estriadas, paji- 
zas, los bordes hialinos, glabras, con pelos largos 
más densamente hacia el cuello, 


hacia el borde. 


equitantes, de manera que las láminas resultan 


notablemente dísticas; nervios de las vainas se 
continúan en la lámina; cuello conspicuo con una 
región triangular castaño rojizo hacia los márgenes, 
elabro, con pelos largos hacia el borde, similares a los 
de la vaina; aurículas y contralígula ausentes; lígulas 
lineares, 
rígidas, (3-)6-9(-10) 


E 
[e] 
dispuestas en la 


membranáceo-ciliadas, 0.5 mm; láminas 
a veces levemente arqueadas, 
X 0.5-0.6 cm. 


mediana de las 


base y porción 


cañas, plegadas hasta el ápice, 


obtusas o subagudas, base continua con la vaina, 


longitudinalmente estriadas en ambas caras, con 
nervio central y laterales conspicuos, ambas caras 
glabras, con ápice escabriúsculo plegado, retuso, 
partido en hojas viejas, rara vez hipofilo con pelos 
ralos. Cañas floríferas con 1 inflorescencia naciendo 
del último nudo caulinar, con pedúnculo aplanado, 


con surco lateral notable, longittrdinalme nte estriado, 


v 


glabro; inflorescencia compuesta por 3-5 racimos de 


4-8 cm, siendo los basales alternos y los apicales 
opuestos, raquis ligeramente ondulado, triquetro, de 
0.4—0.5 mm de ancho, cara dorsal plana, cara ventral 
con costilla conspicua, hispídulo, más conspicua- 
mente sobre los ángulos, sin pelos largos o bien con 
pelos hasta de 1.5 mm de la altura de la inserción de 
los pedicelos; pedicelos de 0.2-0.5 mm, aplanados, 


hispídulos, glabros o laxamente pilosos, con 1-5 pelos 


hialinos hasta de 1.5 mm. Espiguillas de contorno 
oblongo-elíptico, 2.1-2.3(22.5) X 0.9-1 mm. dorsi- 


ventralmente comprimicas, agudas, ocráceas, alter- 
nas, dispuestas laxamente en 2 hileras sobre el raquis. 
de tal forma que cada espiguilla se opone a otra de la 
otra hilera en 0-1/4(-1/3) de su 


lemma 


longitud: gluma 


superior y inferior membranáceo-hialinas, 


gluma superior de 1.9—2.1(-2.3) mm, menor que el 
antecio, obtusa, 4—5-nervia, el nervio medio poco 
prominente o ausente, los restantes conspicuos, 


submarginales, glabra, lemma inferior de 2-2.1(-2.5) 
mm, glumiforme, igual al antecio, obtusa, 4—5-nervia, 
1(-2.5) mm, 


poco brillante, glabra, con densas y diminutas papilas 


glabra; lemma superior de 2-2. castaña, 
en toda la superficie, disminuyendo hacia los bordes, 
éstos blanquecinos y de superficie lisa, encerrando los 
bordes de la pálea; pálea superior de igual consis- 
tencia y aspecto que la lemma superior, ligeramente 


menor que la lemma, glabra. Cariopsis no vista. 


Distribución Crece ei 


geográfica y ecología. 


en el T.F. 


Atures, en los cerros Yaví y Coro Coro, y en el estado 


Venezuela Amazonas, Departamento 


de Bolivar en la meseta Jaua del Parque Nacional 


Jaua-Sarisariñama. Habita en áreas de bosques 


— 


ribereños, arbustales y nerbazales tepuyanos de los 
en la altiplanicie 


1750— 


alrededores del río Marajano, 


meridional de la Meseta de Jaua, entre los 
2200 m 

Fenología. Coleccionada en flor y/o fruto en los 
meses de octubre y noviembre. 

Axonopus magallanesiae es afín a A. flabelliformis, 
por presentar vainas equitantes y láminas notable- 
A. flabelliformis se distingue por 
incluir ejemplares de mayor porte, de 0.55-1.20 m 
10-20-60) cm y 
inflorescencias generalmente con 6-20(-30) racimos, 
de 5-15 
(1.3—)1.6— (2.2) mm, la gluma superior es 2(-3)-nervia 


mente dísticas; 


le altura, sus láminas de 


sus 
em, las espiguillas son de menor tamaño, de 


y la lemma inferior es 2-nervia. 

El análisis del material tipo permitió observar que 
la pilosidad de los pedicelos es variable dentro de una 
misma inflorescencia, hallándose pedicelos glabros y 
otros pilosos, con l a 5 pelos hialinos laxamente 


distribuidos a lo largo del pedicelo 


Material inad VENEZUELA. Amazonas: Allures, 
W del 


cumbre del Cerro Coro-Coro, en las cabeceras N 


616 Annals of the 
Missouri Botanical Garden 


103 ) 
<A b 


— FA 
— — ee 
Figura 7. ird V Giraldo-Cañas (dibujado del isotipo Huber 12995, SI). —A. Habito. —B. Detalle del 
ápice de una lámina. talle de la región ligular. —D. Detalle de una porción del raquis. —E. Espiguilla vista del lado 
de la s superio WE > spiguille vista del lado de la lemma inferior. —G. Antecio superior visto tel lado de la lemma. 


—H. Antecio superior visto del lado de la pale 


Volume 93, Number 4 
2006 


Cialdella et al. 
Axonopus Serie Suffulti 


Cumbre del 


Huber 11882 


Río Manapiare, Átures, 


Huber 12339 (Sl; 
Cerro Yavi, cabeceras del Río Parucito, 
11. Axonopus pennellii G. A. Black, Advancing 
Frontiers Pl. Sci. 5: 142. 1963. TIPO: Colombia. 
Meta, Villavicencio, sandy open hillside, near 
Río Guatiquia, 700—900 m, 26—31 Ago. 1917, F. 
W. Pennell 1539 (holotipo, US  1041805!: 
isotipos, foto ex US en SI!, NY). Figuras IF, 8. 


de 0.70-1.20 m 


cespitosas, con rizomas cortos; canas generalmente 


Plantas perennes de altura, 
aplanadas, de 3—4 mm de diámetro, con surco lateral 
2- 6 


estriadas, glabras, cubiertas por las vainas basales; 


notable, 3-nodes, simples, longitudinalmente 
nudos castaños, levemente engrosados, glabros o con 
escasos pelitos a vilosos; vainas menores que los 
entrenudos, longitudinalmente estriadas, pajizas, gla- 
bras: nervios de las vainas se continúan en la lámina; 
cuello imperceptible; contralígula y aurículas au- 
0.3-0.4 mm, 
castañas, con márgen blanquecino; láminas acintadas 
(1030-45 X (1-)1.4-1.5 cm, 


predominantemente basales en la planta, plegadas 


sentes; lígulas membranáceo-ciliadas, 


anchas, herbáceas, 
solamente en la base y planas hacia el ápice, éste 
conspicuamente acuminado, la base continua con la 
vaina, longitudinalmente estriadas y escabriúsculas 
en ambas nervio central y laterales 


Caras, con 


conspicuos y pelos largos, delgados y dispersos en 


toda la superficie de ambas caras, el márgen 


escabriúsculo y viloso en toda su extensión. Cañas 
floríferas con 1-2 inflorescencias naciendo del último 
nudo caulinar, pedúnculos ligeramente angulosos, con 
estrías longitudinales poco  conspicuas, glabros; 
inflorescencia con (12-)18-26 racimos de 9-15 em; 
raquis notablemente ondulado a recto, 0.3-0.5 mm de 
ancho, pajizo en la zona central, verdoso en los 
márgenes, triquetro, cara dorsal plana, cara ventral 
con costilla conspicua, hispídulo o escabriúsculo en 
toda su superficie, más conspicuamente sobre los 
ángulos, sin pelos largos; pedicelos de 0—0.3 mm, 
conspicuamente aplanados, hispídulos. Espiguillas de 
contorno oblongo-elíptico, de 1.6—1.9(-2.1) X 0.6— 
0.8 mm, 


alternas, dispuestas en 2 hileras sobre el raquis, de 


dorsiventralmente comprimidas, ocráceas, 
tal forma que cada espiguilla se opone a otra de la otra 
hilera en 1/3 de su longitud; gluma superior de 1.9— 
2.1 mm, mayor que el antecio, aguda, 2(—3)-nervia, 
con pelos largos y delgados, dispersos, más densa- 
mente en hilera a ambos lados de los nervios; lemma 
inferior de 1.8 mm, mayor que el antecio, 2-3-nervia, 
con igual indumento que la gluma superior; lemma 
superior de 1.8-1.9 mm, aguda, castaña, glabra, con 
escasos y cortos pelitos en su ápice, rara vez sin ellos, 


con densas y diminutas papilas en toda la superficie, 


disminuyendo hacia los bordes, éstos blanquecinos y 
de superficie lisa, encerrando los bordes de la pálea; 
pálea superior de igual consistencia y aspecto que la 
lemma superior, ligeramente menor que la lemma, 


glabra. Cariopsis no vista. 


Distribución | geográfica y ecología. Hasta el 
momento ha sido colecc:onada en Colombia (Meta) y 
en Venezuela (Barinas). Crece en lugares abiertos, 
sobre suelos arenosos e altitudes medias entre los 
200-900 m. 

Fenología. Coleccionada en flor y/o fruto entre 
los meses de agosto y noviembre. 
Número cromosómico, 


Davidse 318, US, VEN). 


Anton (1982) menciona que en Axonopus pennellii 


— 10 (nota en el ejemplar 


los nudos son pilosos. Sin embargo, en el material tipo 


de esta especie, los nudos son glabros « 


con muy 
escasos pelitos. 

Axonopus pennellii se distingue fácilmente por sus 
hojas anchas de 14—15 mm, con ápice notablemente 
acuminado. Por el hábito robusto de las plantas y las 
láminas anchas, A. pennellii se acerca a A. succulentus 
y A. pressus. Axonopus succulentus se aparta por tener 
aurículas y ápice de las láminas retuso; A. pressus por 
otra parte, se distingue por la presencia de innova- 
ciones comprimidas, pajizas, cubiertas con vainas de 
disposición dística en la base de la planta. Axonopus 
hoehnei es también afín a A. pennellii por presentar 
láminas foliares largamente acuminadas (véase ob- 


servaciones bajo A. hoehnel). 


Material examinado. COLOMBIA. Meta: Sierra de la 
escarpment, Philipson 2262 (US) VENE- 

> km al S érida intersection just 
outside ol 8 Davidse 3183 (US, VEN). 


12. Axonopus polydactylus (Steud.) Dedecca, 


Bragantia 15: 273. 1956. Basónimo: Paspalum 

polydactylum Steud., Syn. Pl. Glumac. 1: 19. 

1853, como “P. polydactylon”. TIPO: Brasil: 
pot) ; 


Bahia, sin fecha, P. Salzmann s.n. 1 P no 
visto; isotipos, BAA 2257! fragm. ex US 
2859750! fragm. ex P, US 558295!). Figuras ; 2D, 
E, 9. 


Plantas perennes, conspicuamente cespitosas, de 
(0.50—)0.90—1.20 m de 


robustas; cafias delgadas, gráciles, aplanadas en la 


altura, rizomatosas, poco 


base, luego cilíndricas, de 2-3 mm de diámetro, 5- ó 
6-nodes, simples, pajizas, longitudinalmente estria- 
das, con surco lateral poco conspicuo, glabras; nudos 
castaño rojizo, levemente comprimidos, glabros; 
vainas menores que los entrenudos, conspicuamente 
estriadas, pajizas, conspicuamente laciniadas y fibro- 
sas cuando viejas, glabras o con algunos pelos largos 
cuello im- 


en el borde hacia el cuello a vilosas: 


Annals of the 


618 


Missouri Botanical Garden 


TA 


TENTER 


=== 


— 


F 
: 


_ AR 


4 17 
. | 10 
AMA 


AMUA 


== OS 
== SSS SSS z 

aeaa 
555 


Ss 


t 


alle de la región 


B. Det 


A. Hábito. 


Figura 8. 


Axonopus pennellii G. A. Black (dibujado del isotipo Pennell 1539, NY). 


E 


D. Espiguilla vista del lado de la lemma inferior. 


C 


Antecio superior visto del lado de la pálea. 


ligular. 


Detalle de una porción del raquis eon espiguilla. 


emma. 


| 


Antecio superior visto del lado de la 


` 
4 n 


Volume 93, Number 4 Cialdella et al. 619 
2006 Axonopus Serie Suffulti 


CD IX AS RRB es 


R ZE RISE 


Dd 
tT = SS S Oe 
SEES HSA 
EUN 


— 
— , T: 


NS 
Ys 
\ 


TZ 


— EA y 
E 
V EL 
Sie 


LA etm 


Uu 


Figura 9. Axonopus polydactylus ( (Steud. A Dedecca (de Zuloaga et al. 4819, SI). —A. Hábito. —B. Detalle de la región 
ligular. —C. Detalle de una porción del raquis. —D. Espiguilla vista del lado de la gluma superior. —E. Espiguilla vista de 
lado de la lemma inferior. —F. Antecio superior visto del lado de la lemma. —G. Antecio superior visto del lado de la pálea. 


Annals of the 
Missouri Botanical Garden 


perceptible, pajizo, glabro; contralígula aurículas 


ausentes: ligula membranácea, de 0.4-0.5 mm, con 


márgen densamente pestañoso, el resto glabra; nervios 
láminas 
20—410(—60) X 0.4-0.6 cm. 


predominantemente basales en la planta, plegadas en 


de las vainas se continúan en la lámina: 


lineares, herbáceas, de 


la base, planas hacia el ápice, de ápice agudo, base 


continua con la vaina, sin pseudopecíolo, longi- 
tudinalmente estriadas en ambas caras, con nervio 


central y laterales conspicuos. glabras o escabriúscu- 


las en el epifilo, de márgen escabriúsculo, o con 
escasos pelos de 2.5-3 mm cerca de la base, 


espaciándose hacia el ápice de la lámina o con la 
1 6 2 


cara abaxial vilosa. Cañas floriferas con 2 
inflorescencias, cada una de ellas con pedúnculo 
aplanado, glabro, longitudinalmente estriado; inflo- 
rescencia con (5—)11—17(—30) racimos mayormente en 
verticilo, los basales alternos, 7.5-18(230) em. cada 


racimo compuesto por ssr sailles alternas, dispuestas 
en 2 hileras sobre el raquis, de tal forma que cada 
espiguilla se opone a otra de la otra hilera en 1/3-2/3 de 
su longitud; raquis desde apenas ondulado hasta 
notablemente sinuoso, 0.3-0.4 mm de ancho, triquetro, 
cara dorsal plana, pajiza en la zona central, verdosa 
hacia los márgenes, cara ventral con costilla conspicua, 
a veces pubérulo, hispídulo sobre los ángulos, sin pelos 
largos, excepcionalmente con escasos (14) pelos largos 


(1.5-2 mm). 


situados sobre los ángulos a ambos lados de 


blanquecinos, más o menos rígidos 
la base 
de cada pedicelo; pedicelos brevísimos (0.1—0.2 mm), 
más espaciados en la base del racimo y más cercanos 


del 


pubérulos. Espiguillas dorsiventral- 


hacia el ápice mismo, ligeramente aplanados, 
glabros, a veces 


mente comprimidas, de contorno oblongo-elíptico, 1.2— 
1.602 


lemma inferior membranáceo-hialinas, a veces violá- 


0.6-0.9 mm, ocraceas; gluma superior y 
ceas cerca de los nervios, agudas, glabras u ocasio- 
nalmente con algunos pelitos dispersos, gluma superior 
igual o superando en 0.2 mm el largo del antecio, 2- 
nervia, lemma inferior igual al antecio, 2-nervia; lemma 
superior endurecida, subaguda, ca. 1.3 mm, castaña, 
brillante, glabra, con escasos pelitos cortos en el ápice, 
con densas y diminutas papilas en toda la superficie, 
disminuyendo hacia los bordes, éstos blanquecinos y de 
superficie lisa, encerrando los bordes de la pálea; pálea 
superior de igual consistencia y aspecto que la lemma 
ligeramente menor 


superior, que la lemma. glabra. 


Cariopsis no vista. 
Distribución geográfica y ecología. | Éndémica del 
Brasil (Bahia, 


Minas Gerais, 


Goiás, 
Rio 


Grande do Norte). Habita en cerrados y restinga, sobre 


norte y noreste de Ceará, 
Maranhão, Pará, Pernambuco y 


— 


suelos rocosos o arenosos, entre 0-1200 m. formando 


densas matas. 


Fenología. Encontrada en flor y/o fruto entre los 
meses de diciembre y abril (junio). 

Axonopus polydactylus se caracteriza por su hábito 
densamente ce spitoso, con vainas basales senescentes 
fibrosas. 


laciniadas, cañas plurinodes (con 5 6 6 


nudos) láminas angostas (4 a O mm de ancho). 
plegadas, inflorescencias pluriracemosas, con raquis 
delgado (0.3—0.4. mm de ancho), espiguillas pequeñas 
(1.2-1.6 mm), distribuidas laxamente a lo largo del 
raquis y antecio superior en general cortamente 
escabroso. 

Esta especie se halla estrechamente. relacionada 
1 Axonopus suffultus por sus inflorescencias plurira- 
cemosas, distinguiéndose esta última por tener las 


vainas basales marcadamente comprimidas, glaucas, 


no fibrosas, el raquis de las inflorescencias más 
ancho (0.4-0.7 mm de ancho), las espiguillas 


mayores (1.8-2.5 mm) y antecio superior glabro. 


Además, A. suffultus crece en sur de Brasil, Paraguay. 


Uruguay y noreste y centro de Argentina. También es 


alín a A. elegantulus (veáse observaciones bajo esta 
especie). 

En la descripción original de Paspalum polydacty- 
lum Steudel (1853-1855) menciona que esta especie 


presenta las glumas ]-nervias a obscuramente 5- 
nervias. Sin embargo, en el material tipo se ha 


observado que la gluma superior y lemma inferior son 
2-nervias. 

Axonopus polydactylus es una especie polimorfa en 
relación a la pilosidad de las hojas e inflorescencias. 
Se hallaron ejemplares con laminas vilosas, como en 
Swallen 3612, Zuloaga et al. 4819 y 4855, Chase 
3737 y 3775, y otros glabros como en Swallen 4054. 
Con respecto a la pilosidad del raquis de los racimos, 
si bien es frecuente que el mismo sea glabro, en 
algunos ri como en Swallen 3612 y 4034, 
Zuloaga et al. 4855, Swallen 3737 y 4536 e Irwin el 
al. 14656, se 1515 aron 1—4 pelos hialinos hasta de 
2 mm, junto a la inserción de los pedicelos. Tambien 

observaron espiguillas con algunos pelitos dis- 
persos sobre la gluma superior y la lemma inferior en 
Zuloaga et al. 4655 y 4619. 

Se observaron ejemplares cuyas inflorescencias 
presentan racimos de 25-30 cm, en Swallen 
3555, 3737, 3750 1/2, 3761 y 4084. Asimismo, en los 
especimenes Swallen 4028 y Valls et al. 


encontraron racimos inferiores ramificados, 


COMO 


332 se 
MM 
poco frecuente en el género (véase observaciones bajo 
Axonopus flabelliformis). 


En el ejemplar Chase 8003 se encontraron en una 


misma inflorescencia espiguillas con la gluma 


superior y lemma inferior 3-nervias y otras 2-nervias. 


Axonopus purpurellus y flabelliformis var. 
decipiens fueron consideradas como sinónimos de A. 
polydactylus (Anton, 1982); en este trabajo, se 


Volume 93, Number 4 
2006 


Cialdella et al. 
Axonopus Serie Suffulti 


consideran en la sinonimia de A. flabelliformis (véase 


observaciónes bajo esta especie). 


Material adicional O BRASIL. Bahia: Para- 
fuso, 39 km NE of Bahia, Chase 8003 (NY, US); Mun. 
Mucugé, de Mucugé a Jussiape, 6 6 km del desvío de Jussiape 
de la ruta a Barra de Estiva, Zuloaga et al. 4819 (IBGE, MO. 

5D: Ilheus, ruta de Olivença a Una, 17 km de Olivença, 
Zuloaga a T E, MO, SI); ca. 4 km SW of Belmonte, on 
road to Itapel , Harley et al. 17324 (US). Ceará: Campo 
Grande, 2 4536 (US); Campo Sales to Crato, Su 11 80 
4327 (US) Brasilia, Riedel (U 
Goiás: Rio Piau, ca. 225 km SW of Barreiras on s to 
Posse, Irwin et al. 14656 (NY, US); Rio da Prata, ca. 6 km S 
of Posse, /rwin et al. 14512 (NY, US); Mun. Tocantinópolis, 
ribeirao do Corrego, 55 km SW of Estreito along Belém- 
Brasilia Hwy. (BR 153), Plowman et al. 9241 (US). 
Maranháo: Grajahú to Porto Franco, Swallen 3801 (US): 
Mun. Imperatriz, Belém-Brasília Hwy. (BR 010), 30 km S of 
Imperatriz, Plowman et al. 9283, 929] (US); Carolina to San 
Antonio de Balsas, Swallen 4034 (R, US); Barra do Corda to 
Grajahú, Swallen 3612 (NY, US), 3754 (BAA, US). Pará: 
Mun. Tucuruí, S of represa Tucuruí 
along. Hwy. al. 9663 (US); Mun. 
Campo de apiquara, ca. lkm E of 
Matipaquara, Davidse et aL 5 (US). Pernambuco: 
Tapéra, Pickel 2547 (US). Piauí: Beiras, Gardner 2365 (US). 
Rio Grande do Norte: Estremoz to Natal, Swallen 4764, 
4765, 4766 (US) 


B 


Breu Branco, 
422, Ed el 
Marapanim, 


13. Axonopus pressus (Nees ex Steud.) Parodi, 
Notas Mus. La Plata, Bot. 3(17): 23. 1938. 
Paspalum pressum Nees ex Steud., Syn. Pl. 
Glumac. 23. 1853. TIPO: Brasil, sin loc. 
1836, F. Sellow 5638 (holotipo, B no 
isotipos, BAA 2258! fragm. ex B, 
0512.01 fragm. ex B no visto, P!, 
fragm. ex B). Figura 10. 


visto; 
-TRIN- 
US 2942557! 


Axonopus derbyanus G. A. Black, Advancing Frontiers PL 
e : 127. 1963. TIPO: Brasil. 
nde, campo, 540—550 m, 9(7-11) Feb. : 
Ce | 10836 918 8 ne 1501229!; isotipos, Die ex 
en p , NY 346101 no visto, foto ex 
1 


75 


grs 1015 s 


e os deban G. A. Black var. D o G. A. 


1 
x 
T 


, Advanci me Frontiers Pl. Ps 5: 130. 1963. 
TIPO: Brasil. Goiás: Goian 7 5825 e Mar 
1930, A. Chase 11552 (holotipo. A 14489021: 9 
foto ex US en SI). 


Plantas perennes, frecuentemente robustas, de 0.7— 
152) m de 


formes, 1-2 cm de largo, 


altura, cespitosas, con rizomas falci- 


gruesos, con innovaciones 
ascendentes, falcadas, originalmente subterráneas y 
cubiertas por numerosos catáfilos rígidos, pajizos, 
ligeramente comprimidas y notablemente dísticas; 
cañas robustas de 3-4 mm de diámetro, leve 
a marcadamente comprimidas, 2- ó 4-nodes, simples, 
con surco lateral conspicuo; nudos pajizos a castaño 
rojizo, comprimidos, glabros; vainas superando los 
entrenudos en la base, luego menores que los mismos, 


predominantemente basales, las inferiores fascicula- 


das, conduplicadas, aquilladas, longitudinalmente 


estriadas, pajizas, glabras a densamente hirsutas 
hacia la porción superior o densamente vilosas en 
toda la superfice, los márgenes densamente pilosos en 
toda su longitud o solo hacia el ápice de la vaina, 
excepcionalmente glabros; nervios de las vainas se 


continúan en la lámina: 


cuello conspicuo, pajizo 
a rojizo, glabro a densamente viloso en toda su 


superficie o bien solo cerca de los bordes: lígulas 
membranáceo-pestañosas, 0.5-2.5 mm, pajizas, lus- 


~ 


trosas, con pelos hasta de 1.5 mm, rara vez el margen 


muy cortamente ciliado, glabras, a veces la cara 
abaxial cortamente pubescente; contralígula y aurí- 
culas ausentes; laminas acintadas anchas, rigidas, 


(10-)20-45 X 0.8-1.2 


sales en la planta, 


cm, predominantemente ba- 


poco divergentes del caule. 
plegadas en la base, luego planas, con sus bordes 
frecuentemente revolutos cuando secas, con ápice 


retuso, cortamente mucronado, frecuentemente hen- 


dido a la vejez, la base angosta continuándose 
imperceptiblemente con la vaina a redondeada, 
pajizas, longitudinalmente estriadas en el hipofilo, 
menos conspicuamente en el epifilo, hipofilo glabro 
con pelos ralos sobre el nervio medio, a densamente 
viloso en toda la superficie, principalmente hacia los 
bordes, epifilo glabro a ralamente viloso en los bordes, 
1-2 


cada una 


márgenes es 'abriüsculos. Cafias floríferas con 


— 


inflorescencias en el último nudo caulinar, 


fol 
e 


e ellas con pedúnculo cilíndrico, longitudinalmente 
estriado, con surco lateral poco conspicuo, glabro; 
inflorescencias formadas por (10-)26-34 racimos, 
(101332 cm, siendo 


los basales alternos y distantes, 


excepcionalmente menos, de 
los apicales más 
próximos entre sí; raquis ondulado o recto, pajizo, 
triquetro, cara dorsal plana, 0.6—0.7(—1) mm de ancho, 
pajiza en la zona central, verdosa hacia los bordes, 
cara ventral con costilla conspicua, hispídulo en los 
ángulos, a veces también hispídulo sobre las caras, sin 
pelos largos, ocasionalmente con 1—2 pelos de 4— 
6mm a cada lado de la base de cada pedicelo; 
pedicelos de 0.2-0.3 mm, aplanados, escabriúsculos 
en los ángulos, ocasionalmente con algunos pelos 
largos similares a los del raquis. Espiguillas oblongo- 
2.2-3 X (0.7—)0.9-1 mm, 


mente comprimidas, agudas, ocráceas o con tíntes 


elipsoides, dorsiventral- 
violáceos, alternas en 2 hileras sobre el raquis, de tal 
forma que cada espiguilla se opone a una de la otra 
hilera en 1/4-1/3 de su longitud; gluma superior y 
lemma inferior de 2.2-3 mm, en general levemente 
mayores que el antecio, a veces lemma inferior igual 
al antecio y gluma superior menor, ambas membra- 


náceo-hialinas, delicadas, glabras a vilosas, más 


densamente hacia ambos lados de los nervios; gluma 
superior aguda u obtusa, 2(-3)-nervia; lemma inferior 


con ápice cortamente acuminado, 2(—3)-nervia; lemma 


— 


622 Annals of the 
Missouri Botanical Garden 


Figura 10. Axonopus pressus (Nees ex Steud.) Parodi (de Solomon et al. 6926, SI). —A. Hábito. —B. Detalle de la región 
ligular. —C. Detalle de una porción del raquis. —D. Espiguilla vista del lado de la gluma superior. —E. Espiguilla vista del 


lado de la lemma inferior. —F. Antecio superior visto del lado de la lemma. 


G. Antecio superior visto del lado de la pálea. 


Volume 93, Number 4 
2006 


Cialdella et al. 623 


Axonopus Serie Suffuiti 


2.0 X 0.7-1 mm, 


brillante, papilosa en toda la super 


superior elipsoide de castaña, 


— 


icie, disminuyendo 
hacia los bordes, glabra o con macropelos y micro- 


pelos bicelulares aislados hacia el ápice: pálea 


superior similar en textura y ornamentación a la 


lemma, ligeramente menor, glabra: estambres 3. 
anteras de 1.4—2.5 mm. Cariopsis ovoide, 1.7-1.8 X 
0.6-0.8 mm, blanquecino; hilo oblongo. subbasal: 


embrión 1/2 del largo de la cariopsis. 


Distribución geográfica y ecología. Brasil central 


(Bahia. Distrito Federal. Goiás. Maranháo. Mato 
Grosso do Sul. Minas Gerais, Pará. Paraná. Rio 
Grande do Norte y Sáo Paulo). región oriental de 


(Amambay, Cordillera, San 
de Bolivia (Santa Cruz). 


campos abiertos, en barrancos y en bordes 


Paraguay Caaguazú, 


Pedro) y este Crece en 
cerrados, 
sobre suelos arenosos o modificados. 


entre los 200 y los 1300 m. 


Fenología. 


de caminos, 


Coleccionada en flor y/o fruto entre 


los meses de noviembre y julio. 
Nombres vulgares. 
1994). 

Numero cromosómico. n = ca. 20 (Morrone et al., 
1995). 


à presencia de rizomas falciformes con innova- 


"Kapfi misione (Paraguay, 


Morrone et al., 


ciones en principio subterráneas y cubiertas por 


caláfilas pajizas y dísticas es un carácter muy útil 
para identificar esta especie. Axonopus pressus es afín 
a A. suffultus, distinguiéndose 


esta ültima especie por 
el color glauco y violáceo de vainas y láminas, la 
disposición dística de las vainas, las láminas de mayor 
tamaño, de 40-55 cm, cuello indistinguible, el raquis 
de las inflorescencias ralamente a densamente piloso 

la ausencia de innovaciones comprimidas latera- 
mente cubiertas por numerosas vainas. También 
resulta afín a A. succulentus, la que se aparta por la 
Asimismo, 


presencia de aurículas foliares. A. pressus 


— 


tiene hojas anchas, similares a las encontradas en A. 
Jlabelliformis y A. pennellii, pero se distingue de ellas 
porque en éstas el cuello es inconspicuo y glabro y los 
racimos son más cortos, de 5-15 cm; asimismo, en A. 
flabelliformis la disposición de las vainas basales es 
dística y en A. pennellii las espiguillas son menores, 
de 1.6-1.9 mm. 
Steudel (1853-1855) 


Paspalum pressus señala que la gluma superior y la 


en su diagnosis original de 


lemma inferior son 3-nervias. El análisis del material 


tipo y de ejemplares de herbario permitió observar 


A eel 


que la gluma superior y la lemma inferior son 2- 
nervias, y sólo excepcionalmente se halló en espi- 
guillas aisladas, dentro de una misma inflorescencia, 
gluma y lemma con nervio medio desarrollado. 

En el herbario BAA se estudió un fragmento ex B del 


isolipo de Paspalum pressum. en cuya etiqueta se indica 


“F. Sellow 5688”, datos que probablemente se deban 
a un error de transcripción de la etiqueta original. 
Axonopus pressus es una especie polimorfa con una 
marcada variación en la pilosidad de las vainas. 
Steudel 


(1853-1855), al describir la especie. señala que las 


láminas, tamaño y pilosidad de la espiguillas. 


vainas son vilosas en los márgenes y las espiguillas 
son glabras. Black (1963) sigue el mismo criterio en 
relación a la pilosidad de las espiguillas, pero para las 
vainas señala que pueden ser escabrosas a densamente 
vilosas. Sin embargo. entre los ejemplares citados por 
Black bajo esta especie, algunos presentan espiguillas 


pilosas. como por ejemplo Glaziou 16609 y Chase 


11564. En la misma obra, Black describe A. 
derbyanus y señala su afinidad con A. pressus. 
distinguiéndola por tener espiguillas glabras. El 


análisis del material tipo y de material de herbario 
en general permitió determinar que existe una marcada 
variación en la pilosidad de las vainas, láminas y 
espiguillas aún dentro de un mismo individuo. Así, el 
ejemplar tipo Sellow 5638 presenta espiguillas glabras 
y otras esparcidamente pilosas: en Filgueiras & 
Zuloaga 2097, Chase 9092 y 9317 y Dusén 11686, 
las espiguillas son glabras, el raquis presenta pelos en 
la base de los pedicelos y las vainas son densamente 
pilosas: otros con espiguillas vilosas, raquis con 
escasos pelos y vainas densamente pilosas, como en 
Clayton 4828 y Swallen 3733: con espiguillas, raquis 
y vainas glabras, es el caso de Zuloaga «€ Morrone 
4633 y Morrone & Pensiero 381; 


plares con espiguillas pilosas, con raquis, vainas y 


finalmente, ejem- 


áminas glabras como en Clayton 4799 y Rojas 6716. 


a variación observada permite incluir A. derbyanus 
en la sinonimia de A. pressus. 
Black (1963) describe 


parvispicula y la distingue por su 


Axonopus derbyanus va 


vainas y láminas 


densamente vilosas y espiguillas menores de 2- 
2.2 mm, características que ane parte de la 
variación presente en la especie, y que justifica la 
inclusión también en este caso de 15 ha variedad bajo 
la sinonimia de A. pressus. 

Los ejemplares Krapovickas et al. 14266 y Morrone 
& Pensiero 394 


contralígula marcadamente diferenciada, membraná- 


~ 


se caracterizan por presentar 


cea en la porción basal y ciliada hacia el ápice. 


Ambos ejemplares han sido coleccionados en 
San 


caso del ejemplar Morrone & 


a 


misma localidad, Estancia La Manina. Pedro 


(Paraguay). y en el 
Pensiero 394 solamente se hallaron dos plantas con 
contralígula entre una población de Axonopus pressus. 
Por el momento se incluye a los mismos dentro de esta 
ültima especie. 

Material adicional examinado. pe IVIA. Santa Mee 
Prov. Velasco, Parq. Nac. Noel Kempff M.. serranias S y 
de la pista Noel Kempff M., Mostacedo et al. 18 20 (t i 


624 Annals of the 
Missouri Botanical Garden 
Prov. Velasco, San Ignacio, 2 km hacia SE, Seidel & Beck Plantas perennes, gráciles, de 15-80 em de 
411 (L a altura, cespitosas, con rizoma delgado, de entrenu- 
BRASIL. Bahia: 6 km N of Alagoinhas along Hwy. BR. gos menores de 0.5 em. y vástagos muy próximos, 


116. 1 et al. 11739 (MO); Mucugé, de Muc 


a ruta a Barra da Estiva, 


icugé a Jussi 


ape. 6 km del desvío a Jussiape, de | 


Zuloaga el 4816 (IBGE, MO, Sl); Espigao Mestre. ca. 
100 km W of Barreiras, Anderson et al. 36658 (US). 
Distrito 1 Ren Brasília, Clayton 4799, 462€ 


Sáo Bartolomeu, Filgueiras & Zuloaga 2120 ( 
Barragem da Paranoa, Filgueiras 2180 (IBGE, SI). 
Goiania, Macedo 4306 (BAA, US): Mun. Niquelândia, 
C E de Niquel Tocantins (CNT), « 

Alva arenga et al. 1188 (IBC E, 


mina de niq iel, 


2h betw. 


to Ao Swallen 37. 33, 376 (NY. S); Imperatriz, 
Bananal, 15 km S of Impe n dons Belém-Brasilia Hwy. 


(BR 010), Plowman et al 9345 (MO). Mato Grosso: close to 


base camp, ca. 270 km N of RR on the Xavantina—Sao 
Felix rd., Ratter et al. 2078 (US); Mun. Bataguaqu. Corrego 


Feio, Hatschbach 23582 (US). Mato Grosso do Sul: Campo 
Grande, a 9601 (US). Minas Gerais: Arascá, 
Macedo 42. 3A A, US): Serra de Lenhieiro, Glaziou 17428 
(BAA, P, E S; ien do 1 33 km NE of Francisco 
Sá, 2 (NY, US); Rod. de Serro 
a Diamantina, 5 km después ie desvío a Pedro Lessa, 
Zuloaga & Mortona 1633 (SD; 61 km W of Pará de Minas 
262 to Uberaba, Davidse & Ramamoorthy 10824 


W of town 


— 


. lo Salinas, /rwin et al. 23 


along Hwy. 
(NY). Para: Mun. Conceição do Araguaia, 4 km 


center along Hwy. PA-150 MN et al. 8988 (US). 
Paraná: Jaguariaiva, Ribeirao 5 Reis. Hatschbach 9078 


(US); Mun. Arapoti, Rio das . Barra do Perdizes, 
Hatschbach 6866 (US). Rio Grande do Norte: Entremoz to 
Natal, Swallen 4751, 4754, 4759, 4760. 4701 (US). Sao 
Paulo: Mun. Botucatu, 18 km N of Botucatu, 
a0 Manucl-Piracicaba Hwy., 


Manuel, along S Gottsberger 979 


(US); Mun. ltapetininga, Faz. ta Luzia do Campo Largo 
SSE de Itapetininga, Mor hado de 2 1 90 US): Sao 
José dos Campos, ca 7.3 km SSE em linha da praç 


Mimura 350 (NY, 
s.n. (SD; Mun. Sta. 
Simao, Sendulsky 149 
PARAGUAY. Amambay 
Corá, Rojas 6872 (BAA, Sl); Parq. Nac. Cerro Corá, próxi. al 
Río Aquidabán, 99 17 1 5 & Pensiero 463 (SI); Sierra de 
geri Cerro Cora, Rojas 6716, 6716A (BAA, SD: Parq. 
Na 6926 (MO. PY, SI). 
9269a (BAA, G): 
López et al. 207 
5112 (BAA, SD. 


S); Serra de Cunha, ee & Ge hn 


105 from Sta. Rita, along road Sao 


Sierra de Amambay, Cerro 


1c. Cerro Cora, Solomon et al. 


Hassler 


camino a Yhú, 


e Prope Caaguazú, 
10 km al N de Caaguazú, 
(CTS. SD; Est. Primas 
Cordillera: P nino a Paraguarí, Morrone 
is Pensiero 114, 394 (S1). San Pedro: Ruta 3, 17 m al N de 
camino a Gral. po e entrada a la Est. La Manina, 
ins rone & Pensiero 361 
Krapovickas et al. 14206 Aj *. 


Rojas 


suayabí, « 
| de San Estanislao, 


14. Axonopus ramosus Swallen, Contr. U.S. Natl. 
Herb. 29: 413. 1950. TIPO: Surinam. Joints in 
bedrock, Savanna I, Table Mountain (Tafelberg), 

19 Sep. 1944, B. 24810 

(holotipo, NY 346107 no visto; isotipos, foto ex 

NY en SI!, 19149981). 


Figura 11. 


505 m, Maguire 


F no visto, U no visto, US 


paralelos entre sí; cañas delgadas, de 0.5-0.8 mm 


de diámetro, levemente comprimidas, 4- ó 5-nodes, 
lisas o con estrías longitudinales casi inconspicuas 
y surco lateral notable, simples en la porción basal, 
luego ramificadas, principalmente en la zona media: 
nudos pajizos a castaños, ligeramente comprimidos, 
densamente vilosos, con pelos delgados. patentes, 
blanquecinos, ca. l mm, excepcionalmente nudos 
elabros: vainas menores que los entrenudos, longi- 
tudinalmente estriadas, pajizas, con pelos delgados 
ralamente dispersos sobre toda la superficie, más 


densamente hacia los bordes y hacia la zona del 


cuello; nervios de las vainas se continúan el a 
lámina: cuello conspicuo, pajizo, densamente viloso: 
contralígula y aurículas ausentes: lígulas membra- 
náceo-ciliadas, 0.2-0.5 mm, blanquecinas, lustrosas; 


(3-)10- 


densamente dispuestas 


áminas angoslz imente lineares, rígic as, 


ps (0.10. 


| la porción mediana de las cañas, plegadas en la 


2-0.5 em, 


ine luego planas, con ápice conspicuamente 


agudo, base continua con la vaina, longitudinal- 


mente estriadas en ambas caras, con pelos delgados, 


de base notablemente tuberculada, ralamente dis- 


persos sobre ambas caras y sobre los márgenes, 
ocasionalmente láminas glabras. Cañas floriferas con 


1-2 inflorescencias naciendo del último nudo 


caulinar, cada una de ellas con pedúnculo cilín- 


drico, delgado, glabro. con estrías longitudinales 


poco conspicuas y surco lateral notable: inflores- 


cencias formadas por 2—4 racimos delgados gráciles, 
de 3-8 em; 
cara dorsal plana, (0.1—)0.6—0.7(-1) mm 


raquis levemente ondulado o recto, 
Iriquetro, 
de ancho. pajiza en la zona central. verdosa hacia 
los bordes, hispídulo en toda la superficie, princi- 
palmente en los ángulos, sin pelos largos, ocasio- 
altura de la base 
0.1—0.3(-0.6) mm, 


contorno 


nalmente con 1-2(—4) pelos a la 
de algún pedicelo; pedicelos de 
Espiguillas de 

0.4-0.7 mm, 


obtusas, 


hispídulos. 
1.2-1.6 


tralmente compri midas, 


aplanados, 


oblongo-elíptico. dorsiven- 


ocráceas, allernas 


en 2 hileras sobre el raquis, de tal forma que cada 
espiguilla se opone a una de la otra hilera en 0-1/3 
de su longitud: gluma superior igual o menor que el 


antecio, obtusa, 2-3(4)-nervia, glabra: lemma in- 


ferior igual o levemente mayor que el antecio, 


aguda, 2—3(4)-nervia, glabra; lemma superior elip- 


soide, 1.1-1.5 mm, castaña, brillante, glabra, con 


densas y pequefias papilas en toda la superficie, 
disminuyendo hacia los bordes, glabra, ocasional- 
mente con escasos pelitos cortos en el ápice, bordes 
blanquecinos y lisos, encerrando los bordes de la 
ligera- 


pálea; pálea superior similar a la lemma, 


625 


Cialdella et al. 


Axonopus Serie Suffulti 


0,5 mm 


0,5 mm 


E. Espiguilla vista del 


G. Antecio superior visto del lado de la 


G24 z 


2 2 
. 2 
TES A e 


= — T 
S x 


SAS 


Volume 93, Number 4 


2006 


“gura : 
Fig 11 


A. Hábito. —B. Detalle de una porción de cañas floríferas ramificadas en la 


región media y superior. —C. Detalle de la región ligular. —D. Detalle de una porción del raquis. 


Axonopus ramosus Swallen. 


| 


lado de 


a gluma superior. —F. Espiguilla vista del lado de la lemma inferior. 


H. Antecio superior visto del lado de la pálea. A, C-H. Maguire et al. 54191 (NY); B, de Granville 3816 (MO). 


lemma. 


626 


Annals of the 
Missouri Botanical Garden 


mente menor, glabra. Androceo compuesto por 3 


estambres con anteras de 0.8 mm. Cariopsis no 


vista. 
Distribución Crece en 
Brasil 


(Amapá). Habita en zonas montañosas, sobre laderas 


geográfica y ecología. 


Surinam, Guayana Francesa y norte de 


rocosas, a bajas altitudes desde el nivel del mar hasta 


los 750 m. 


Es un elemento predominante en las 
sabanas. 
Fenología. | Mallada en flor y/o fruto durante todo 
el ano. 
Vombres vulgares. misione” 
Morrone et al., 1994 


Esta especie se distingue por su particular hábito 


“Kapri (Paraguay. 


fastigiado, ya que las hojas se concentran en la zona 


media de las cañas, siendo escasas en la base de las 
mismas. Las láminas son frecuentemente decíduas y 
presentan pelos de base conspicuamente tuberculada, 
ralos sobre la superficie y los márgenes, principal- 
mente en la zona basal de la lámina. Por otra parte, los 
pelos presentes en las vainas tienen base no 
tuberculada. 

Por su hábito ramificado se asemeja superficial- 
mente a Axonopus caulescens (Mez) Henrard, pertene- 
ciente a la subserie Scoparii, serie Barbigeri (sensu 
Black, 1963), pero ésta difiere por el tamaño mayor de 
las espiguillas, la gluma superior y la lemma inferior 


5-7-nervias, con los nervios marcados y el nervio 


medio usualmente bien desarrollado. 


BRASIL. 


ssif des Tumuc-Humac, 


Material. examinado. 
louaken, Mas 
(MO. US). 

GUAYANA FRANCESA. Haut mom. 
Pere, Hook 224 (MO, US): Mont St. Marec 
(MO); 10 0 Saint-Marcel, zone SE du n 


Amapá: Inselberg Ta- 
Granville et al. 12239 


Roche Mon 
. de Granville 2051 


15301 (US): Tumue Humac, Frontière In sil-Guyane, Gran- 
ville Ts 10 S). Cayenne: Camp n. kouba Booka goo 
Soula, Bassin du Ha Roche 1, coil 9800 (US); Haute 
Canopi, Mont Belvédere, entre. Orstom 0909 (MO) 
Massif des Emerillons, zone centrale S Savane roche exposée 
a l'est sur le flane coline P Granville 3816 (MO) 
Roche NS atou, bou de L'Ovapock, Cremers & Granville 
13985 (US) Montagne des 1 Dassin « Acatave 
Feuillet : 25 (US) Montagne del Nouragues, Ba 


PAratave Somnet, Sarthou 188 (US 


zone N savane roche au sommet d'une 2 sur d eauche de 
a Haute Approuagec, de Granville 3953 (ME) 

SURINAM. Forested hills 9 km N of Lucie ET 12 km V 
of Oost Riviers, Mauire et al. 54191 (NY, US): Poe rg, 5 
I. Schulz et Donsebar 10597 (US); 


Lucie Rivier, 2 km NE of 


island, near affluence of Oost Rivier, Maguire et al. 54170 
(US); Saramacca, Tafelberg, Kramer et Hekking 2907 (US). 


15. Axonopus sueculentus C. A. Black, Advancing 
i 134. 1963. TIPO: Paraguay. 
Ybytymf Cordillerita, Feb. 
T. Rojas 6176 (holotipo, US 2012951); 
foto ex US en SIL BAAD. 


Frontiers Pl. Se 
Campo serrania, 
1933. 


ISOLIpos, Figura 12. 


P 


rizomatosas, 


antas perennes, robustas, de 0.6-1.2 m de altura, 


con vástagos cubiertos por catáfilos 


rígidos; cañas robustas de 5mm de diámetro, 
levemente cilíndricas, 3- 6 4-nodes, simples, longi- 
tudinalmente estriadas, glabras: nudos castaños, 


conspicuos, e comprimidos, glabros a pubér- 
ulos, principalmente los cercanos a la base de la caña: 
vainas menores que los entrenudos, longitudinalmente 
estriadas, pajizas, glabras, bordes lisos, glabros: 


nervios de las vainas se continúan en la lámina: 
cuello conspicuo. pajizo, glabro: aurículas presentes: 
contralígula ausente; lígula membranáceo-ciliada, 1— 
1.5 mm, pajiza, lustrosa, con cara abaxial pubescente: 
0-55 X l- 


basales en la 


láminas acintadas, anchas, rígidas, de 3 


2 cm, 


plegadas en la base, luego planas, con ápice retuso, 


predominantemente planta, 


base continua con la vaina, pajizas, longitudinalmente 
estriadas y glabras en ambas caras, márgenes lisos, 


rara vez levemente escabriúsculos. Cañas floríferas 


con 2-3 inflorescencias naciendo del ultimo nudo 


caulinar, cada una de ellas con pedúnculo cilíndrico, 
longitudinalmente estriado, con surco lateral conspi- 
cuo, glabro; inflorescencias formadas por 15-1825) 


racimos de (8-)12-20 em, siendo los basales alternos 


Qu 
— 


y los apicales más próximos entre sí, verticilados: 
raquis levemente ondulado, triquetro, 0.4-0.8 mm de 
ancho, cara dorsal plana, pajizo en la zona central, 
luego verdoso, cara ventral con costilla conspicua, 
hispídulo en los ángulos, con 34 pelos de 1—1.2 mm 
d Ya lado de la base de cada pedicelo; pedicelos de 
(0.2-)0.3—0.5 mm, aplanados, glabros o cortamente 
Pu di 
2.1-2.4 


das, 


Espiguillas de contorno oblongo-elíptico, 


| i | | i 
UOLSIVEHU CHEE 


0.7-1 mm, comprimi- 


agudas, ocráceas, alternas en 2 hileras sobre e 
raquis, de tal forma que cada espiguilla se opone a una 


0-1/4 


superior de 2-2.4 mm, igual o levemente menor que 


de la otra hilera ei 


de su longitud; gluma 


el antecio, aguda u obtusa, 2-3-nervia, con escasos 
pelitos dispersos, más densos a ambos lados de los 
nervios laterales: lemma inferior de 222.4 mm, igual o 
ligeramente mayor que el antecio, aguda, 2(3)-nervia, 
con igual pilosidad que la gluma superior: lemma 
superior ca. 2-2.2 mm, castaña, brillante, glabra, con 
densas y pequeñas papilas en toda la superficie, 
disminuyendo los 


b 


pálea; pálea superior similar a la lemma, ligeramente 


hacia bordes, glabra, bordes 


anquecinos y lisos, encerrando los bordes de la 


menor, glabra; estambres 3, anteras ca. 1.5 mm. 


Cariopsis no vista. 


Distribución geográfica y ecología. Especie en- 
démica de la región oriental de Paraguay, en cl 


departamento Paraguarí. Crece en campos. 


Fenología. | Coleccionada en flor y/o fruto en e 


mes de febrero. 


Volume 93, Number 4 Cialdella et al. 627 
2006 Axonopus Serie Suffulti 


— 
— 
* 
ao 


r rese ERE m 


Figura 12. Axonopus succulentus G. A. Black (dibujado del holotipo Rojas 6176, US). —A. Porción superior de una caña 
florífera. —B. Detalle de la lámina, con una porción de la vaina. —C. Espiguilla vista del lado de la gluma superior, con 
pedicelo. —D. Espiguilla vista del lado de la lemma inferior, con pedicelo. —E. Antecio superior visto del lado de la lemma. 


. Antecio superior visto del lado de la pálea. 


Annals of the 
Missouri Botanical Garden 


Nombre vulgar. “Kapri kyra” (Paraguay, Morrone 
et al., 1994). 
sta especie fue considerada por Anton (1982 


— 


como sinónimo de Axonopus pressus, con la salvedad 
de que tal vez sea un extremo de la variación de la 
especie, ya que los pedicelos de las espiguillas son 


pestañosos. Sin embargo, la presencia de aurículas en 


a mayoría de las vainas, junto con las hojas más 


anchas, permite diferenciar ambos taxones. 


16. Axonopus suffultus (Mikan ex Trin.) Parodi, 
Notas Mus. La Plata, Bot. 3(17): 23. 1938. 
Basónimo: Paspalum iar Mikan ex Trin., 
Neue Entdeck. Pflanzenk. 2: 46. 1821. Panicum 
suffultus (Mikan ex Trin.) 1 5 Revis. Gen. 
3(3): 364. 1898. Axonopus scoparius var. suffultus 
(Mikan ex Trin.) Herter, Anales Mus. Hist. Nat. 
Montevideo, ser. 2, 3(1): 49. 1929. TIPO: Brasil. 
sin loc., J. S. Mikan s.n. (holotipo, LE-TRIN 
0537.07 no visto; TOTO: US 80029! fragm. ex 
LE). Figuras 2F, 


- 


LT 


Panicum hagenbeckianum Kuntze, Revis. Gen. Pl. 32): 3 
1898. Anthaenantia hagenbeckiana t ) Se 10 
aisle Bot. Jahre: sber. 261: 329. 900. natpis 


Bot. 3(17): 21. 1938 TIPO: Tus utc Bolivie: Gran 
Chaco, C. F. Hagenbech s.n. (holotipo, B no vi 
isotipos, BAA 317! fragm. ex B, US 80689! fragm. 
ex B). 
Pp Hn M Mez, Bot. Jahrb. Syst. 56 (Beibl. 125): 
. 1921. Axonopus un ME z) Hitche. & Chase ex 
Bou pues Jard. Bot. list. Nat. Paraguay 2: 
160. 1930. Axonopus gene (Kuntze) Parodi 
var. dern vg ) G. A. Black, Advancing Frontiers 
Sci. 5: 


neon B no visto: isotipos, BAA 317! fragm. ex B, 
Gl, MO- 2101060. US 7254141, US 2855289! ie 
ex B) 


Plantas perennes, de 0.5-1.5 m de altura, densa- 
mente cespitosas, con rizomas cortos, y macollos 
dísticos, de aspecto iridáceo; cañas robustas de 3— 
6 mm de diámetro, levemente comprimidas, 2- 6 3- 
nodes, simples, con surco lateral conspicuo, glabras: 
nervios de las vainas se continüan en la lámina; nudos 
castaño rojizo a negruzcos, comprimidos, glabros, 
a veces con algunos pelos largos a la altura del cuello: 
vainas aquilladas, lateralmente comprimidas, menores 
que los entrenudos, conspicuamente dísticas, estria- 
das, generalmente elaucas con zonas violáceas, rara 
vez pajizas, glabras a vilosas, con bordes lisos. 
glabros; cuello ineonspicuo, pajizo, glabro: lígula 
membranáceo-ciliada, 0.4-0.5 mm, castaña, contralí- 


gula y aurículas ausentes: láminas acintadas, rígidas, 


de 40-55 X 0.5-1.3 cm, predominantemente basales 
en la planta, plegadas en la base, aquilladas, luego 


planas, glaucas a pajizas, con sus bordes frecuente- 


mente revolutos cuando secas, con ápice obtuso, base 
continua con la vaina, longitudinalmente estriadas en 
el hipofilo, menos conspicuamente en el epifilo, 
glabras, a veces con márgenes escabriúsculos. Cañas 
floriferas con 1-2 inflorescencias naciendo del ultimo 
nudo caulinar, cada una de ellas con pedúnculo 
cilíndrico, longitudinalmente estriado, con surco 
lateral menos conspicuo que en las cañas, glabro: 
inflorescencias formadas por 12-23(-40) racimos de 

—) l 1—22 em, siendo los basales alternos y distantes, 
los apicales más próximos entre sí; raquis ondulado o 
recto, triquetro, cara dorsal plana, 0.4-0.7 mm de 
ancho, con nervio medio notable, purpúreo a blanque- 
cino, cara ventral con costilla conspicua, escabroso en 
los ángulos, generalmente con (2-)3-4(-6) pelos de 
1-2 mm, rígidos sobre los ángulos a cada lado de la 
base de los pedicelos, rara vez dispersos a lo largo de 
los ángulos; pedicelos de 0.2-0.5 mm, distantes en la 
base del racimo, siendo más cercanos hacia el ápice 
del mismo, notablemente aplanados, escabriúsculos 
en los ángulos, con algunos pelos largos similares a los 
del raquis. Espiguillas de contorno oblongo-elíptico, 
1.8—2.2(-2.5) X 0.6—1 mm, dorsiventralmente com- 
primidas, agudas, ocráceas, alternas en 2 hileras 
sobre el raquis, de tal forma que cada espiguilla se 
opone a una de la otra hilera en 1/2 de su longitud: 
eluma superior y lemma inferior iguales o levemente 
mayores que el antecio, con ápice cortamente 
acuminado, membranáceo-hialinas, delicadas, con 
tintes castaños, gluma superior 2(—3)-nervia, lemma 
inferior 2-nervia, glabra; lemma superior elipsoide, 
1.5-2.2(-2.4) mm, castana, brillante, glabra, con 
densas y pequeñas papilas en toda la superficie, 
disminuyendo hacia los bordes, éstos blanquecinos v 
lisos, encerrando los bordes de la pálea, micropelos 
bicelulares hacia el ápice; pálea superior similar a la 
lemma, ligeramente menor, glabra: estambres 3, 
anteras de 1-1.2 mm, negruzcas. Cariopsis elipsoide, 
de 16 X 0.7 mm; embrión 1/2 del largo de k 


cariopsis; hilo oblongo. 


Distribución geográfica y ecología. Sur de Brasil 
(Paraná, Rio Grande do Sul, Santa Catarina y Sáo 
Paulo), Uruguay (Canelones, Cerro Largo, Durazno, 


Florida, Paysandú, San José y Tacuarembó), Paraguay 


(Alto Paraná, Amambay, Caaguazú, Cordillera, Guairá, 
ltapuá, Misiones, Paraguarí y San Pedro) y noreste y 
centro de la Argentina (Buenos Aires, Chaco. Córdoba, 


Corrientes, Entre Ríos, Formosa, Misiones y Santa Fe). 


Crece en suelos bajos, arenoso-arcillosos, húmedos, « 
bien en suelos pedregosos, secos. 

Fenología. | Coleccionada en flor y/o en fruto entre 
los meses de septiembre y abril (julio). 

Nombre vulgar. "Palha-dura" (Smith & Wasshau- 
sen, 1982). 


Volume 93, Number 4 Cialdella et al. 629 
2006 Axonopus Serie Suffulti 


NY 


INN 
SANA 
P 
n 


Figura 13. 1 na (Mikan ex Trin.) Parodi (de Rojas 2304, SI). —A. Hábito. —B. Es p vista del lado de 
la n na inferior. talle de una porción del raquis, con una e 5 vista del lado de la gluma superior. — D. Antecio 
superior visto del T H pi at —E. Antecio superior visto del lado de la lemma. —F. Detalle de la región ligular. 


630 


Annals of the 
Missouri Botanical Garden 


Numero cromosómico. 2n = 20 (Parodi, 1946, sub 
Axonopus iridaceus, Nunez, 1952), 2n = 40 (de Wet & 
1956). n = 10 (Hunziker et al., 1998). 


Según Anton (1982) en el herbario de LE existe una 


Anderson, 


cartulina que lleva varios especímenes correspon- 
dientes a dos coleccionistas: tres de ellos colecciona- 
dos por Riedel, año 1831, Brasil, que corresponden 
Ixonopus ae e y 
Langsdorff * 
Brasil” 


último ejemplar fue citado bajo Paspalum suffultus 


uno 


1 sylvaticis umbrosiis pr. Mandiocam. 


que poor con A. polystachyus. Este 
por Trinius (1826: 90). y fue el ejemplar seleccionado 
Su dibujar el hábito de la especie en su obra de 
). Los detalles de las espiguillas de la 
lámina de la obra de Trinius corresponden al ejemplar 
tipo de P. suffultus (Mikan s.n.). 


originales de los detalles pegados en la 


encontrándose los 
misma 
cartulina del ejemplar tipo. 

Herter (1956) describe 


sobre la base de un ejemplar coleccionado por J. 


Axonopus chloridiformis 


Schroeder, El tipo no pudo localizarse, pero los 
caracteres dados en la diagnosis original se corres- 


suffultus. 


Axonopus suffultus se caracteriza por su hábito 


ponden con A. 


cespiloso, con rizomas de entrenudos cortos, cañas 
erectas, con las vainas basales marcadamente fasci- 
culadas, éstas usualmente glaucas y violáceas. Se ha 
observado que ocasionalmente puede presentar esto- 
lones de entrenudos largos, que arraigan y dan nuevos 
macollos, como en el ejemplar Morrone et al. 1592. 


Asimismo, se ha encontrado que ocasionalmente, la 


gluma superior, la lemma inferior y el raquis son 


densamente hirsutos. Esta característica se evidencia 


WS 


en el ejemplar Schwarz 807: 


ARGENTINA, Buenos 
Basualdo, Parodi 6135 (BAA, SD; 
dea 


Material adicional examinado. 
Aires: Pdo. P 
) 

_CTES); 


Stuckert 11813 
t. Tranque ra de Hierro. 66 km 


arcos duin 


rrie a Do C MECepción, 


al NE de € havarría camino a Co oncepción, Cerro Puitá, Arbo et 
al. 6951 (C TES, LPB); Dpto. Gral. Alvear, Ruta Nac. 14, de 

Alvear a Santo To mé. a 12 km de Alvear, Zuloaga et al. 3102 
(SI): Dpto. Ituzaingó, desembocadura del Avo. G arapé en el río 


Paraná, Quarín et al. 2838 (BAA): Dpto. Me Est. 
Ita Caabó, Pedersen 611 
Cruz, Burkart 7938 (SU); 
Dpto. Santo Tomé, Ayo. Chirimay y Ruta 
1081 (BAA. CES); Ruta 41, 5-6 km i rt de 
Ferrucci et al. 6347 , ES. SI); límite con Misiones. Avo. 
Chirimay. Zuloaga & Morrone 6430 (SI). Entre Ríos: Dpto. 
24728 (SI, US) Dpto. 
Concordia. Concordia. Parodi 6790 (BAA). Formosa: Dpto. 


DL. 


Tres Cerros. nr 2995 
. Tressens et al. 
Galarza. 


Federación, Fe dé ración. Burkart 
Pilcomayo, Laguna Primavera, Morel 9012 (US): sin departa- 
(5D. Misiones: Dpto. 
Apóstoles, San José, Escuela Agrotéenica Don Bosco. Renvoize 


mento, Las Palmas, Joergensen 2428 


coleccionado por 


3042 (US): Dpto. Cande de Santa Ana, 1 a 573 (BAA, 
SI): Ruta Prov. 3, 10 km del desvío de la Ruta Nac 


Cora. Morrone et al. 2751 SD: D 


Dpto. San Pedro. San 
Dpto. Garay y San Janvier, entre Saladero Cabal y San Javier, 
): Dpto. Gral. Obligado, ipei Venturi 
MA): Dpto. San Justo. | 712 (BAA). 
Paraná: Ponta Grossa, Swallen 


a Mora, Ragonese 2 
Serrinha, in campo subpaludosn,: 


Campini ide Bis Sul, Est. 


0 87 


. Rambo 53411 (BA 
20 (BAA): . 
S). Santa Catarina: Mun. Chapecó, 

Campo Sáo Vicente, 24 km W 
9307 (US). Smith & Klein 13 
12 km N of 1 2 un 13338 (SL US): 
Bom Retiro. by Fazenda Santo Antonio, Campo dos Padres. 
Smith & Reitz 10306 (NY, SD. São Paulo: Mun. São Carlos, 
Porto Pulador on Rio Moji Guacu, 8.9-9 km NNE of the RR 
station at Sta. Eudoxia, Kiten & Freitas Campos 3495 . 

PARAGUAY. 
(BAA. US): 


Amambay: Sierra de 


Rambo 565 
8 


\belardo Luz. 


Alto Paraná: Tacurú Pucú, Fiebrig 073 

dnd. d 5089 (BAA. C. LIL, sn. 

. Hassler 9729 (G). Caa- 

guazú: Palomares, camino > de lt: aqu: yry a Curugualy. Se hinini 
SI): 


ltaquyry, 


A 
=> 
= 
> 


CTES iei Oviedo, 
" E Cordille "ai i 


— 
C 
- 
Sy 
= 
ey 
a 
5 
2 
= 
S. 
5 


José, km 93, Burkart 18: 
Dona Juana, pres de Villa Rica, Balansa b. (MO); Santa 
Bárbara, Jorgensen 4101 (BAA, a SD; V 
9254 (BAA, LIL, SI); 


n ‘a, Rojas 


1 1 Rojas 9254a 


(BAA. SL US). Itapúa: Isla Yaciretá. Jiménez 14005, 
14655112 (Sl). Misiones: Tebicuary, Rojas 9254a (SI): 
Est. “La dad”. Pedersen 4255 (BAA, SD. Pedersen 7633 
(BAA. US). Paraguarí: 119 8 Est. Barrerito, orilla del 


LIL): National Ybiewi NW 
corner of park, Zardini 15970 (MO). San Pedro: Villa San 
Pedro, Rojas 2504 (SI); Alto Paraguay. Primavera. 
; 77 (SI. Chaco Septentrionalis, 
Fiebrig 5374 (G); sud Paraguay, Kuntze s.n. (NY). 
URUGUAY. Canelones: ca. del Pat p> me. Legrand 
991 (BAA). 1 cui Rio Negro, Est. Palleros. Gallinal 
et al. 1235 (BA . Durazno: Est. 
18745 (US). TON Santa C PA Gallinal et al. 
3035 (US). Meseta de Artigas. Rio 
Rosengurtt 0285 (BAA, SL US). 
. Rosengurtt 6015 (US). 
Cuer os. ll 


Vo. Yaquart. Ramirez 130 (BAA 


Woolston 


~ 


Sin de partamen 110: 


s Palmas. Schrader 


Timote. 


San José: 


Taodareubo: 
Morron 
Felippone 5365 (S1). 


km al N de Tacuarembó, 


(SI). Sin departamento: sin loc., 


ZsPECIE. EXCLUIDA DE LA SERIE SUFFULTI 

Advancing 
1963. TIPO: 
Villa Unión 
(I mi. E of Santa Lucia) on rd. from Villa Unión 
to Durango, 1280 m. 27 Sep. 1953, J. R. Reeder 
& C. G. Reeder 2445 (holotipo, TAN no visto: 
isolipos, MEXU!, NY no visto, foto ex NY en SIL 
SI! fragm. ex MEXU). 


Axonopus mexicanus G. A. Black. 
144. t 10. 


E of 


Frontiers Pl. Sci 5: 


México. Sinaloa: ca. 47 mi. 


Volume 93, Number 4 
2006 


Cialdella et al. 
Axonopus Serie Suffulti 


Black (1963: 34, 144) incluye Axonopus mexicanus 


dentro de la serie Suffulti. Sin embargo, al analizar 


material del isotipo, se observó que los antecios son 
pálidos, blanquecinos a pajizos, carácter que no se 
encuentra presente en especies de la serie Suffulti. La 
presencia de nervio medio marcado, antecio superior 
pajizo y láminas planas, anchas, permitiría ubicar A. 
mexicanus en la serie Barbigeri (sensu Black, 1963). 


Literatura Citada 


Anton, A. R. 1982. 


Las especies de Axonopus sección 


nuc ps morfológica y taxonómica. Tesis 
doctoral, iv. Nac. de odo: Argentina. 
1801. Notes on Gramineae. I Linn. Soc., Bot. 


Bentham, 
18: 14-134 

& J. D. 0 1883. Gen. PI. 32): 1-1258. 

Black, G. A. 19 ). Novas espécies 


de Paniceae (Gramineae) 
do Brasil. D en. 29-37. 
. 19 


nst. Agron. N. 20: 29- 


e 


,rasses of the genus 18 (a taxonomic 
1 1 Frontiers Pl. Sci. 186. 
Brown, W. V. 1977. The K 
grass systematics. Mem. 
Bb 1969. Gramíneas. Pp. 1-551 en A. ZU 
(editor), iste Ilustrada de Entre Rios (Argentina), ES 
Pt. 2. Instituto Nacional de Tecnologia 


ranz syndrome E ils ~ in 
Torrey Bot. Club Dx 


Agropecuaria, 
Buenos Aires. 

Chase, A. 
Soc. W ashington. 24: 

Clayton, W. D. & S 1986. Genera Graminum. 

Her es sly's Stationery Office, London. 

Dávila, P. & L. 


survey of the leaf E epidermis o a ghas strum (Poaceae: 


V. Proc. Biol. 


1911. Notes on genera of Paniceae. 
—160. 


eee. 


E 


. 1990. Scanning electron microscopy 


/ e dr 1 Bot. 7 )-: 

Dedecc M. rud espécies XT do 
a 5 Bragantia 15: 251—296 

De Wet, J. M. J. € L. J. 
numbers i in Pug aal grasses. Cytologia 21: 1— 

Doell, J. € . Paniceae, Pp. 98^ 35 90 ~ F. P. 
(edito en m Brasiliens sis 

Giraldo-Cañas, D. 


gênero 


Anderson. 1956. Chromosome 


Martius 


9. Nuevas c itas ao M (Poaceae: 
3 


Panicoideae: E anic e para Colombia. Hickenia 
47-50. 

. 2000a. Una nueva especie de ed dn (Poaceae: 
i de la Guayana venezolana e entario del 
género para Venezuela. Caldasia 22: 231— 

2000b. Una nueva sección del género Axonopus 
ions eae, Panic oideae, Paniceae). Revista Acad. Colomb. 


Ci. Ex 24: 183-191 


c. Sinopsis de la sección Cabrera del género 


Panicoideae, 
Colomb. Ci. Exact. 25: 207-223. 


2002. Las especies de la 


ne dianed Axonopus (Poaceae, uniceae). 


Revista Acad. 
Sect. Lappagopsis del 
género neotropical Axonopus (Poaceae, Fame nde "ae, 
Paniceae). Revista Pa Colomb. Ci. 


Burdet, V. 


Greuter, W., J. McNeill, R. Barrie, H. M. 
Demoulin, T. S. Sum D. H. Nicolson, E e Silva, 
E. Skog, P. Trehane, N. J. Turland & D. wksworth 


(editores). 2000. International Code of die Nome 
clature (Saint Louis Code). Regnum Veg. 138. 

Hackel, E. 1887. E Pp. 1-97 en A. Engler & K. 
Prantl (editores), Die natürlichen Pflanzenfamilien. Eng- 
lemann, pae 

Henrard, J.“ wer P new grasses from Paraguay. Meded. 
Ri 


jks e AT: 


Herter, W. G. 1956. 
Gr: aminee d unteren Uruguaybeckens. Revista Sudamer. 
Bot. 10: 

E Kenbiek Y : E M., J. F 


Moraes la 1975 


Axonopus chloridiformis, eine neue 


. Valls, F. M. Salzana & M. 

: P ytogenc etic and e bid 
landi in m genus . Axonopus (Gramineae). Cytc ;lo- 
gla « 

Hiteheoc A 


of the e rican species of grasses described by Linnaeus, 


1908. Types of American grasses: A study 


5 Sloan, Swartz 4 Michaux. Contr. U.S. Natl. 
Herb. 12: 113-158. 
a )9 merc of the Grasses of Cuba. Contr. U.S. 
Natl. Herb. : 183-258, vii-xi. 
p of 3ritish Guiana. Contr. U.S. Natl. 
Herb: 223 vir 


. 1921. nis grasses of 1 Peru, and Bolivia. 


Com. U.S. Natl. 24: 291— 
9 1 P. K., Holmgren P» L . Barnett. 1990, 
Index TA uk Par cde of the World. 
n si rr Gar: a Es w York. 


Hooker. J "lora of British. India, Vol. 7. W. 
1 ye 

Hunziker, Zuloaga, O. Morrone & A. 'obar. 
1998. Estudios cromosómicos en Paniceae asd 'anas 
(Poaceae: Panicoideae). Darwiniana 35: 29-36. 

Jørgensen, P. M. & C. Ulloa. 1994. Seed gue = the high 
Andes of Ecuador: 0 checklist. A. A. U. Rep. 34: 1-443. 

Judziewiez, E. J. ve 0. Poaceae inca A pane 


y 


a ES 
Kuhlmann, J. G. s Gramíneas. omm: Lin. 
67(Bot. 11): me 


| onghi- W agner 


vewslett. (Ser. 
Telegr., Bot. 


„ V. Bittrich, M. C. Lapa Wanderley & 
E 


She ue a Poaceae, F la Fanerogámica 


a Estado de São Paulo, Vol 1. Editora Hucitec, São 
Paulo. 
Mez, C. C. 1917. XIII. Generis Paspali species novae. 


Repert. Spec. Nov. Regni Veg. 16: 60-7 
92]. Neue E Bot. Jahrb. Syak 56 (Beibl. 


125): 1-12. 
O., A. 


Pp. 27666 


in Anton & F. O. Zuloaga. 1994. Axonopus. 
e 3 rer & L. 
de ae. Vol. 
Jardin 

Botanical Garden, St. Loui 
J. H: 8 F. o. Zuloaga & A. Escobar. 


Números c romosómicos en P 


Morrone, 
2 Ramella (editores), Flora 


3 (Gramine ae Conservatoire. et 
> | 


5 E Geneve, and Missouri 
1995. 
aniceae aa 'anas (Poa- 
ceae: Panicoideae). Darwiniana 33: 53-6 

Nees von Esenbeck, C. G. 1829. S oes brasiliensis. 
de ii- Ti en C. F. P. Martius (editor), Fl. bras. enum. pl., 


Z 
S < 
; 
m 


1987. Los Géneros 


e América Austral. Editoria 


1. Rúgolo de Agrasar. 


N 


de Gram 


Sur, Buenos Aires. 
O. 1952. 


lineas de Hemisferio 


Nunez, Investigaciones cariosistemálicas en las 
gramíneas a de la tribus Paniceae. Revista Fac. 

ila 28: 220-256 
Beauvois, A. M. F. J. 1812 "Pus 
Paris. 


Agron. Univ. La Plat 
Palisot de Agrostogr. 
Imprimerie L Fain, 

R. 


Parodi, L. 38. Gramíneas ju Puy nuevas o 
críticas. a Mus. La Plata, 307): 15-33. 
1946. Gramíneas 5 Clave para la 


Enumeración de las 
Especies, 4ta. ed. Acme Agency. Buenos Aires 

| 
Pilger, R. K. F. 


Panicoideae. Pp. 1-208 en A. 


Determinación de los Géneros y 
Idee 
Engler & K. Prantl 
2nd ed. 


Gramineae HI: 


(editores), Die natürlichen Pflanzenfamilien, 


Englemann, Leipzig. 


Annals of the 
Missouri Botanical Garden 


Renvoize, S. A. 1998. Gramineas de Bolivia. Royal Botanic 
Gardens, Kew. 

Roemer, J. J. & J. 

Rosengurtt, B., B. 


Artucio. 


A. Schultes. 
Arrillaga de 
Gramíneas 


1817. Syst. veg. II: 1-964. 
Maffei & P. 


Uruguayas. 


b de 
Univ. de la 
República, Montevideo. 

D. 


Smith, L. B. C. Wasshausen & R. M. Klein. 1982. 
Gramineas Pe. 910-1407 en R. Reitz (editor), Fl. II. 
Catarinense, Fasc. Gram. 85. Paspalum até 115. Zea. 


853-1855. Synopsis plantarum graminearum. 
Pp. 1-475 en E G. Steudel, Synopsis Plantarum Glu- 
J. B. Metzler, Stuttgart. 

Swallen, J. R. 1 5 Gramineae. Kn. B. Maguire & colabo- 
radores. Plant explorations in Guiana in 1944, chiefly to 
the Tafelberg and the Kaieteur Plateau. Bull. Torrey Bot. 
Club 75: 81-91. 

Tovar, O. 1993, 
13: 347-378. 

Trinius, C. B. 


macearum, Vol. 


as gramíneas (Poaceae) del Perú. Ruizia 


Graminibus Paniceis. Disertatio 


1826. De 


botanica altera. Impensis. Academiae. Imperialis Scien- 
tiarum, Petropoli (St. Petersburg). 
——. 1828-1836. Species Gra iminum Iconibus et De- 


se dipifenibus lllustravit, Vo Impensis Academiae 
Imperialis Scientiarum, Pe poo (St. Petersburg). 
Zuloaga, F. O. & O. Morrone. 1999, Le En P. M. 


Jørgensen & S. León-Yánez (editores), Catalogue of the 


Vascular Plants V Ecuador. Monogr. Syst. Bot. Missouri 
Bot. Gard. 75: 809-810. 
& ——— p Axonopus. En V. O. Zuloaga, O. 


T. S. Filgueiras, P. M. Peterson, R. J. 
Catalogue. of. New 


G. Davies: 
Soreng & E. J. Judziewiez ene 
World Grasses (Poaceae): III. 


Morrone, 


Subfamilies Panicoideae. 


1 ae, ALS ae. and Danthonioideae. Contr. 
U.S. Natl. Herb. 16-134. 
———, 11 & J. S. Pennington. 2006. 


C MAT ation and biogeography of the Panicoideae in the 
New International 
Symposium on Grass Systematics and Evolution, Clare- 


World. Proceedings of the Fourth 


mont, California. En pren 


APÉNDICE LISTA DE ESPECIES ACEPTADAS DE LA SERIE SUFFULTI 
DE AXONOPUS 

l. dd arcuatus (Mez) G. A. Black 
2. A. argentin 
3. A. ciliatifolius Swallen 
4. A. cuatrecasasii G. A. Black 
5. A. elegantulus (J. Presl) Hitche 
6. A. flabe lliformis Swallen 
7 Black 


ius Parodi 


E 
— 


10. 4. magallane "siae Giraldo-Canas 
11. A. pennellii G. A. Black 
12. A. polydactylus (Steud.) Dedecca 
13. A. pressus (Nees ex Steud.) Parodi 
14. A. ramosus Swallen 

A. succulentus G 


15 A 
16. A. suffultus (Mikan . ex ve ) Parodi 


COLECTORES. CADA | ESPECIMEN EFS 
POR SU PRIMER COLECTOR. EL (4%) 
APPENDIX 1). UN (*) ES UTILIZADO 


APÉNDICE 2. ÍNDICE DE 
ORDENADO ALFABÉTICAMENTE 
CORRESPONDE A LA ESPECIE (CFR. 
PARA INDICAR MATERIAL TIPO. 
Alvarenga 1188 p Anderson. 9967 
35179 (13). 36471 ( 


(13). 10908 (8). 


. 36658 (13); Araujo 33 (2), 61 (2). 85 


(16), 702 (2). 175 (2). 590 (16); Arbo 1083 
Archer 4830 m Arechavaleta s.n. (2); As 
. (2), 27 (2); 1 97 (2), 99 (16); s 
Cavale anli i 160 (J. >: Barreto 1317 (2), 1358 (16), 4689 (16): 
Bazzi 14 (lo) 401 (lo) Black 46-696 (5); Boerboom 
LBB8764 (O): Buchtien 12 (6); Burkart 816 (2). 824 (2), 
7938 (16). 14350 (16). 18566 (16), 21023 (2). 2: 
22929 (16), 24078 (2). 24728 (16), 25269 (2). 25: 
pos (16), 26193 (2). 26205 (2). 
cabrera 28146 (16), 28268 (2), 28410 (16). 28728 (16): 
Tos 9082 1/2 (13), 9032 1/2 (13), 8914 1/2 (13), 8003 (12). 
8820 (13), 8914 (13), 8918 (13), 8922 (13), 9006 (13), 9019 
9032 (13), 9048 (13), 9079 (13), 9082 (13), 9083 (13), 
91. 32 (13), 9153 (13), 9134 (13), 9244 (13), 9517 (13), 9322 
(13), 9323 (13), 10487 (13), 11173 (13), 11302 (13), 11330 
(13), 11564 (13), 11592 (13), 11603 (13), 11865 (13). 
(4): Clayton 4313 (16), 4799 (13). 4828 (13); Cowan 1819 
(6). 1970 (6): Cremers 13985 (14); Cristophal 130 (6); Cusato 
s.n. (16), 1178 (2). 
da Costa Sacco 2 
(11), 4942 (0). 4942 (6), pu (13). 
12079 (13), 17900 (12), 
22660 (6); 2 10 (€ x 
Llames s.n. (2); Del Mazo s.1 1085 3 (2), 2. 
M Delgado 1055 (6): Dios ski 4353 
2 (13), 5954 (13), 11086 (13). 17669 


(16), 6951 (16 


250 (16); Dauber 111 (2); Davidse 3163 
10867 (13), 


b (12): Dusén s.n. 16): 
Dutra 466 (16 

Fiten 2857 a 2858 (13), 3771 (12), 3876 (12), 4368 
(12); Eskuche 2528-6 (2): Essed 1 »): Evans 3043 (14) 
Felippone 5363 (16); Ferrucci 6347 (M 
14): Fiebrig 5374 (16), 5689 (16), 6735 (16): 
2093 (12). 2097 (13), 2115 (13), 2120 (13), 2134 a 3), 2180 
(13); Foldats 314-a (6); Fontana à p 6 (2): F 180-36 (16). 
F 253-4 (16); Fosberg 27824 (5): ald 60 (2). 

Galli 54 (2), 287 (2): Gallinal 1235 ( 
(2). 1812 (2), 1937 (2), 1947 (2), 2081 es 
3035 3 1165 3565 (2), 3991 (2); Gardner 2. 

, Geyskes 128 (6); Gillespie 15 (0) lucio 16609 

(13). 17428 (13), * 17933 (2), 186 2 (13), 2 2484 (13); 
aere 906 (6); Gottsberger 979F d: 3); 979 (13), 982 
1 (13); Granville 1230 (14), 2631 (14), 3816 (14). 
1860 (14), 12239 (14), 15301 (14); Gregory 10431 


— 


395. E : "e 

(12). 

Hahn 4428 (6); Harley 11459 (12), 15878 (12). 17075 
2). 1732 1 5 2). 18563 (12): orion 1049 (6): P 158 
J eek (1 n 9269 (13), 92604 (1: 3), 1729 (16), 
12556 (2), 1 (2): Hatschbach 3568 (16), < 
(13), 9078 (13), 14010 (13), 14013 (13), 23582 (13); H 
2309 (6); Henz SJ 35057 (16): Herb. Centre Orstom 77 05 (14): 
Hertel 2467 (13); rn 2 (3). 56 (3: Hill 27213 (6): 
Hitchcock 17285 (6), 22515 (5); Hoc 158 (16): Hoehne 28092 
(12); Hoffman 1781 195 pe (6); Hoock s.n. (6), 224 (14), 
410a (6), 412 (6), 414a (6). 420a (6). 1258 (6), 1265 (6). 
1268 (6), 1297d (6); Huber 3523 (6), 6217 (6), 7293 (6), 8008 
(6). 8174 (6). 9162 (6). 10446 (6), 11882 (10), 12339 (10): 
Hutchison ` 3962 (5). 

. 12590 (12), 
14320 (13), 14512 (12), 14656 (12), 19768 (1), 
23506 (13). 27280 (13), 27690 (13). 28812 (13). 

Jiménez 140 (13). 14665 (16). 14655172 (16): Job 760 (2): 

Jorgensen 2428 (16). 4101 (16); Judziewiez 5410 (6). 


I 


pn 


13340 (13). 
25102 (13). 


12890 (13). 


Killeen 1577 (4). 1725 (4), 2290 (4); Kirkbride Jr. 3389 
(13); Klein 3549 (16), 3609 (2), 3657 (16), 3729 (16), 3798 


(16); Kramer 2907 (14); Krap 
44594 (16); Kuhlmann s.n (13); ¿ 
Kuntze s.n. (16). 


wickas 12272 (2), 14266 (13). 
39999 (13); Kunkel 623 (5); 


Volume 93, Number 4 
2006 


Cialdella et al. 
Axonopus Serie Suffulti 


2 1500 (6); Le E 99] (16): mee dd (13) Lima 
194 (13); Loferen 3838 ; Lombardo 3032 (2); López 207 
(13); 2 2911 (16); 1 a (6), 22954 (6). 

Macbride 1498 (5), 3270 (5), 3324 (5), 3760 (5); Macedo 
. 3150 (13), 4235 (13), 4306 (13), 4362 (13), 4370 
(13), 4534 (13), 4616 (13); Machado de Campo 188 (13), 189 
(13); Maguire 27 (6), 32477 (6), 32505 (6), 32621 (6), 33719 
(6), 54.170 (14), 54191 (14); Martínez nel 13305 (2), 9940 
(2); Mattos 7290 (16), 9841 (12); Mc Dowell 2972 (6); Mello 
Barreto 10637 (13); Menacho 333 (4); Mendes Magalhaes 
8904 (13); O s.n. 1 1170 32 (2), 67 (16), 89 (16), 367 (16); 
353 ye ira 732 (6); Montes 1304a 
( 834 (2), 15230 (16), 15442 (2), 
16457 (16), 1644 42 2 2); NE 9012 (16); Morrone 114 (13), 
381 (13), 594 (13), 425a (13), 432 (13), 463 (13), 720 ( 
1592 (16), 1751 (16), 1767 (16), 1817 (16), 4761 (6), 52 
(16); Mostacedo 238 (4), 1880 (13); Múlgura 2170 (16). 

Nee 36379 (4), 46906 (3); Nicora 517 (2), 571 (2). 5071 
(2), 5746c (2) 

Núñez 7230 (5). 

Oliveira 515 (1. 

1. (16). 

Panetti 9993 (16); Parodi 11 (16), 63 (16), * 91 (2). 157 
(2), 1571/2 (2), 4094 (16), 4187 (2), 436 (16), 4632 (2), 
* 4649 (2), 4780 (16), 4942 (16), 5645 (16), 6135 (16), 6191 
(2), 6294. (2), 6359 (2), 6790 (16), * 9584 (2), 11203 (16), 
49432 (16): Pedelaborde s.n. (2); Pedersen 1338 (16). 3038 
(2), 4255 (16), 4482 (2), 4712 (2), 5829 (2), 5876 (2), 6110 
(16), 6426 (2), 7633 (16): Pereira 3922 (13); Peterson han 
(6); Philipson 2282 (11); Pickel 2320 (12), 2547 (12), 
(13), 5884 (13), 5886 (13), 5943 (13); Pinto 1013 (12); Pu 
6156A (8), 6178 (8). 57983 (13); Plowman 8988 (13), 9152 
9595 Vd (12), 9283 (12), 9291 (12), 9292 (13), 9345 (13), 
96 12); Porta 102 (16); Pulle 141 (6) 

suni 1694 (16), 1850 (2), 2838 (16), 2959 (16), 4137 


E 


DER. 


— 
Qo 
CN 
> 
x! 
a 
E 
8 
No ^ 
| 
e 
— 
— 
QO 
T 


— 


3); * Orihuela 60 (2), 63 (2), 300 (2); Orth 


> 


r 2712 (16), 2801 (16); * Rambo 29313 (2), 30763 
(16), 35153 (2), 36459 (2), * 36460 (2), 43592 (16), 44003 


mu 7 (16), 45409 (16), 55411 (16), 53858 (16), 56010 
d 9 0 5 407 (16), 430 (16), 649 (2); Ratter 
13); Reitz 2633 (2), 3269 (16), 7798 (16), 
, 14147 (2), 17681 (16); Renvoize 
3047 (A), 3994 (4), 1029 (4); Rie Po 
12); Rodriguez 573 (16); Rojas 2 
(16), 5112 (13), 5254 (2), 6704 (2), 6716 ( 


" 3, pus (13). 


7964 (16), 9254 (16), 9254a (16), 
13202 (2), 13203 (2), 14076 (2), 14105 
2): 


6749 (13), 6872 (13), 
13129 (2), 13197 (2), 
(2. 14197 (16); * Rojas 13103a (2); Rosa- Mato 1463 
1 ke 305 (2), B-588 (2), B-715 (2), B-825 
(2). 1063 (2), B-9189 (16), B-5391 (2), B- 1 (16), 
(16), 9283 (16); Rua 38 (2), 123 p 3 156 

1 6 7122 (13); Sánchez Vega 177 (5); x EN s.n. (4), 
406 (6); Sarli 101 (2); Sarthou Lr 1); ya 14606 (2), 
22042 (16). 30059 (16); Schrader 18745 ; Schulz 3269 
2), 3326 (2), 3628 (16), 10597 (14), 177 0 20 11786 (2), 
6310 (16), 6379 (16), 6581 (16), 8073 (16), 8154 (16); ws 
111 (13); Semple 381 (6); Sendulsky 140 (13), 149 (13), 1 
(13), 170 (13); Simas 35457 (16); Sin Colector 2503 (6). 9770 
(6): : s.n - (2): Smith 1000 (6), 8053 (2). 8237 (2). 8629 (2). 
9307 (16), 10306 (16). 13338 (16), 13861 (16). 
13990 (2). 15623 (13), 15644 (16), 16039 (2), 13820 (16). 
14064 (16). 15473 (16), /2839 (16); Solomon 6926 (13). 
7071 (13); Soria 5982 (13); Soukup 1905 (5). 6 
Steinbach 1979 (4); Steye ormark 59242 (6). 93856 (6). 113164 
(6), 7 Laa ); Stuckert 14813 (16); Swallen 3504 (12), 3612 
3696 (12), 3733 bis (13), 3733 
(13), 3734 (12), 3735 (12), 3737 (12), 3738 (12), 3750 1/2 
(12), 3753 (12), 3754 (12), 3755 (12), 3757 (13), 3758 (12), 
3760 (12), 3761 (12), 3762 (13), 3800 (12), 3801 (12), 3836 
(13). 3836 bis (13), 3986 (12), 3987 (12), 3988 (12), 4028 
(12), 4032 (12), 4032 1/2 (12), 4034 (12), 4035 (12), 4035 1/ 
2 (12), 4043 (12). 4048 (12). 4049 ( te 4005 (12), 4084 (12), 
4095 (12), 4098 (12), 4142 1/2 (12), 4127 (13), 4142 (12), 
4149 (12), 4327 (12), 4536 (12), = (12), 4653 (12), 4749 
(13). 4751 (13), 4751 bis (13), 4754 (13), 4759 (13), 4760 
(13). 4760 bis (13), 4761 (13), (12), 4765 (12), 4766 
(12). 7463 (2), 7544. 1 (2), 7638 (2), 8110 (2), 6320 (16). 8661 
) 2 (13), 6756 (13). 8968 (13). 8986 (13). 
3). 9169 (2). 9601 (13). 16940 (6). 
95 A 9); Teunissen 11548 (6); Tovar 188 
7 (5); Tressens 1081 (16), 1457 (16), 1855 (16); Tutin 621 


=> 


— 


= 

— 
wn 

= 
n 


e 


s 72 (16), 8332 s 8364 (12), 8489 (12); van der 
1 ST (5); Vargas 57: e 5); Venturi 2 (16), 2384 (16). 
Whit tton 228 (6); Wiley 475 (3); Woolston G 73 pro parte 
(16), C >). 
DM 11827 (2), 15970 (16), 24075 (16), 24738 (16); 
Zuloaga 3102 (16). 4653 (13), 4641 (13), 4757 (13). 4816 
13), 4819 (12), 4855 (12), 6430 (16); 6458 (16), 6551 (16). 
6969 (13). 


CHROMOSOME REPORTS IN M. Cristina Acosta,” Adriana del V. Ordóñez,” 
SOUTH AMERICAN Andrea A. Cocucci,” 
NICOTIANEAE (SOLANACEAE), 

WITH PARTICULAR REFERENCE 

TO NIEREMBERGIA'? 


and Eduardo A. Moscone? 


ABSTRACT 


Chromosome counts and karyotype information from 4l populations of 20 species and 6 varieties of Nicolianeae 
(Solanaceae) from South America belonging to Bouc hetia Dunal (x = d Fabiana Ruiz & Pav. (x = 9), 1 Benth. (x = 
are given, including first reports from 14 taxa: F. densa 


c 
= 
= 
= 
a` 
7 
Y 

| 


10), Nierembergia Ruiz & Pav. (x = 8,9), and / f 
Rémy, N. ericoides Miers, and P. patagonica (Speg.) Millán, with 2n = 2x 8, and N. Hana cra Griseb., N. calycina 

c, N. graveolens A. St-Hil., N. linariifolia Graham var. EX SS a - A. A. Cocucci & Hun pampeana 
(Millan) A. X. Cocucci & Hunz., and var. „ s (Millán) A. A. Cocucci & Hunz., N. 15 lla pu var. p hella and 
var. macrocalyx (Millán) A. A. Cocucci & Hunz, N. rivularis Miers, N. tucumanensis Millán. and N. veitchii Hook., with 2n = 


= 16. All species studie i are nen e most have 2n = 16. In addition. it ues cific ey ll is recorded for the First 
time in V. rigida Miers (2n = 4x = 32) and confirmed in N. aristata D. Don (n = 3x = 24, 2n 48). In general, 
karyotypes of the examined species are symmetrical and compose a of e i inis ‘ly small and ind size zed c TIONS 8. 
mostly of the metacentric (m) type except in L. linifolia (Miers) Griseb., P. axillaris (Lam.) e Stern & Poggenb.. and 

patagonica, where submetacentric (sm) chromosomes are predominant. 1 1 are one or, rarely, two chromosomes bearing 


7 


nucleolar organizing regions per basic complement. In Nierembergia, karyotype data reflect species grouping. Results suggest 
that Bouchetia is the closest taxon to Nier embergia, while Leptoglossis, Petunia, and Fabiana appear more distant. Possible 
chromosome ke evolution in ihe tribe i o based on molecular phylogenetic studies by other authors. In 
Vierembergia, x = 9 is proposed to be ud from x = 8. Dysploid changes in chromosome number and conservation of 
c as morphology in the karvotypes appear as 12 0 evolutionary events in the tribe, while polyploidy is noteworthy 
nly Vicotiana L. 

Kr words: Bouchetia, chromosome numbers, evolution. Fabiana, karyotypes, Leptoglossis, Nierembergia, Petunia, 
poly ploidy, Solanaceae, systemali S 


The tribe Nicotianeae G. Don (Cestroideae Schltdl.. Nicotiana (Goodspeed, 1954; Japan Tobacco, 1994; 
Solanaceae) is composed of plants native to the New Chase et al., 2003) and Petunia (Wijsman et al., 1983; 
World, principally in South America, with Nicotiana Maizonnier, 1984: Mishiba et al., 2000), the remain- 
L. also extending to other continents. As recently ing genera of Nicolianeae have been poorly studied 
Fable 1). 


recognized: Nicotianinae Hunz., comprising Nicotiana The chromosome number and karyotype informa- 


proposed by Hunziker (2001), three subtribes are karyologically (see ' 
with O7 species (77 after Chase et al., 2003), Petunia tion of 20 South American species belonging to five 
Juss. with around 34 species (over 40 following genera of Nicotianeae, i.e. Bouchetia, Fabiana, 


Mishiba et al., 2000), and Fabiana Ruiz & Pav. (15 Leptoglossis, Petunia, and Nierembergia, are reported 


species); Nierembergiinae Hunz. & A. A. Cocucci, herein as a part of a broad cytogenetic study. Our work 


including Nierembergia Ruiz & Pav. (21) and in Nierembergia and related taxa has the following 
Bouchetia Dunal (3); and Leptoglossinae Hunz. objectives: (I) to report original counts and karyotype 
encompassing Leptoglossis Benth. (J). Hunzikeria analyses for karvologically unknown taxa in Solana- 
D'Arcy (3), and Plowmania Hunz. & Subils (J. P. — ceae, (2) to increase our knowledge of the systematic 
nyctaginoides (Standl.) Hunz. & Subils). Although relationships in the tribe Nicotianeae. and (3) to gain 


much cytogenetic work has been carried out in a better insight into chromosomal variation and 


The authors are grate ful to G. Barboza for helpful comments. M. C. Acosta thanks fellowship support from Secretaría de 
Ciencia y Tecnología, Universidad Nacional de Córdoba (SECYT-L 1 ). This study was supported by grant No. PIP 599/98 
from Conse jo Nac no de Investigaciones Científicas y Técnicas (CONICET, Argentina) to A. A. Cocucci and E. A. Moscone, 
and grants No. PID 194/00 and No. PID 62/03 from SECYT-UNC to E. A. Moscone 

“The editors of the Annals thank Sophia Baleomb for her editorial contribution to vis pi 

Instituto Multidisciplinario de Biología Vegetal (IMBIV) CONICET-Universidad Nac al de Córdoba. Casilla de Correo 
195, 5000 Córdoba, Argentina. mosconeCimbiv.unc.edu.ar 

! Facultad de Ciencias Agropecuarias, Universidad Nacional de Cordoba, C. C. 509, 5000 Córdoba, Argentina. 


ANN. Missouri Bor. GARD. 93: 634—646. PUBLISHED ON 15 DECEMBER 20006. 


Volume 93, Number 4 Acosta et al. 635 
2006 Chromosome Reports in Nicotianeae 
Table 1. Previously published chromosome counts in Nicotianeae (Solanaceae), excluding Nicotiana and Petunia. Valid 


names for Nierembergia species according to Cocucci & Hunziker (1995). Synonymous names under which chromosome counts 


were originally published are indicated in brackets. 


Taxon n 2n Reference 
Hunzikeria texana (Torr.) D'Arey [sub nom. Leptoglossis texana l6 Whalen (1979) 
(Torr.) A. Gray | 
Bouchetia anomala (Miers) Britton & Rusby 8 Di Fulvio (1978 
Bouchetia erecta Dunal 8 Whalen ur 
Fabiana denudata Miers 9 Moscone (1992) 
Fabiana imbricata Ruiz & Pav. 9 Desai (1970) 
18 a e - "d 33) 
Leptoglossis linifolia (Miers) Griseb. 10 Subils (197 
Leptoglossis schwenckioides Benth. [sub nom. Salpiglossis schwenckioides 20 Diers ( B 
(Benth.) Wettstein] 
Nierembergia aristata D. Don [sub nom. N. stricta Miers] 8 Di Fulvio (19762) 
24 Di Fulvio (1984) 
„ linariifolia Graham var. linarüfolia [sub nom. 8 Di Fulvio (197642) 
N. hippomanica Miers| 
9 Ratera (1952) 
16 Di Fulvio (1976a) 
Nierembergia repens Ruiz & Pav. 16 Shizukawa & Mii (1997) 
Nierembergia rigida Miers [sub nom. N. aristata Sweet] 8 Di Fulvio (1978) 
9 Ratera (1969) 
Nierembergia scoparia Sendtn. [sub nom. N. frutescens Durieu] 18 Goodspeed (1933) 


possible karyo-evolutionary trends, especially in the 
subtribe Nierembergiinae. 

Bouchetia is a small genus of hemi-cryptophytes 
displaying a disjunct distribution, with B. anomala 
(Miers 
Uruguay and north to central Argentina, and two other 
2001). On the 


is a western South American 


Britton & Rusby in southern Brazil, Paraguay, 
species in North America (Hunziker, 
other d Fabiana 
endemic taxon of microphyllous and resiniferous 
shrubs or chamaephytes distributed along the Andes 
from southern Peru to southern Chile and Argentina, 
which reaches lowland regions in Patagonia. Consid- 


ering the shrubby species examined here, F. densa 


Rémy inhabits southern Bolivia and northwestern 
Argentina, F. denudata Miers extends from northern 


Chile through northern and central Andean areas of 


Argentina reaching northern Patagonia, and F. 
imbricata Ruiz & Pav. grows in the central and 


southern Andean regions of 


1993) 


Leptoglossis comprises xerophytic species inhabit- 


Argentina and Chile 
(Barboza & Hunziker, 
ing Peru except for the one examined here, the 
perennial herb L. linifolia (Miers) Griseb., endemic to 
the central and western Argentine lowlands (Hunziker 
& Subils, 1979). Petunia is a South American genus, 
with one species also in Central and North America. 
Concerning the two entities here analyzed, P. axillaris 
(Lam.) Britton, 


Stern € Poggenb. is an annual herb 


growing in southern Brazil, Bolivia, Uruguay, and 
Argentina and P. patagonica (Speg.) Millán is 


a chamaephyte endemic to the Argentine province of 
Santa Cruz (Figueroa Romero, 1999; Hunziker, 2001). 
Finally, 


tion, 


Nierembergia, focused on in this contribu- 


18 South American 


pom 


except for one species in 
Mexico), with its principal distribution in the western 
and southern regions, and the center of diversification 
in Argentina where 15 species are found (Hunziker, 


2001). Its members have nectarless flowers and, 


— 
un [e] 


most cases, are unique in the family for producing oi 
t 1991) 
Several species have become noteworthy as poisonous 
cattle, 


genetic 


— 


nat attract oil-collecting bees (Cocucci, 


weeds toxic to valuable ornamentals, useful 


lants for engineering or for otential 
| Fe] 
cardenolides 


1995; Godo 


pharmacological uses with Pd to 
and their antitumoral activity (Gil et al., 
1997; 2001). 


Species presently studied include all life forms 


et al., Hunziker, 
known in Nierembergia; they grow in distinct habitats 
and may have either a widespread or restricted 
geographical distribution. Some are small shrubs that 
grow mainly in mountain regions, i.e., N. browallioides 
Griseb., N. ericoides Miers, N. linariifolia Graham var. 
glabriuscula (Dunal) A. A. Cocueci € Hunz. and var. 
pinifolioides (Millán) A. A. Cocucci & Hunz., N. 
pulchella Miers, and N. tucumanensis Millán, or in 
lowlands, i.e., N. graveolens A. St.-Hil., N. linariifolia 
var. linartifolia and var. pampeana (Millán) A. 

Cocucci & Hunz., and N. scoparia Sendtn. Others are 
rhizomatous, stoloniferous, or tuber-bearing, prostrate 


(rarely ascendent) herbs from meadows and marshes, 


Annals of the 
Missouri Botanical Garden 


i.e., N. aristata D. Don, 
Ruiz & Pav., N. rigida Miers, 


N. veitchii Hook. 


linariifolia exhibit the broadest geographical distri- 


V. calycina Hook., N. repens 
N. rivularis Miers, and 
and 


Nierembergia rigida 


bution, extending from southern Brazil southern 
Argentina, reaching Patagonia, while N. tucumanen- 


sis, N. veitchii, N. browallioides, and N. pulchella are 


mainly distributed in northwestern Argentina, the last 


two species also found in southern Bolivia. Nierem- 


bergia aristata, N. graveolens, N. scoparia, and 


rivularis inhabit southern Brazil, Uruguay, and 


northeastern Argentina, with the first species reaching 


Paraguay and the last one into Bolivia. Lastly, N. 


calycina, N. ericoides, and N. repens are endemic to 
eastern Argentina and Uruguay, Argentina (Buenos 


Aires province), and southern 
(Millán, 1941; Cosa de Gastiazoro, 


Hunziker, 1995). 


Chile, respectively 


1989; Cocucci & 


MATERIALS AND METHODS 


The 


presented 


provenance of material studied is 


| Table 2 


mens were ee in 


plant 
The respective voucher speci- 
he herbarium of Museo 
Argentina (CORD). 

Somatic on were observed in squashed 
The 


acid/ethanol 


— 


Botánico de Córdoba, 


root meristems obtained from seed germination. 
fixed in 1:3 


2 hours (h) after a pretreatment either in 


root apps were acetic 
mixture for | 
paradichlorobenzene-saturated solution. for 2h at 
room temperature or 


for 8 h at 8°C. 


stained according to Feulgen's technique by 


in 2 mM 8-hydroxyquinoline 
In most cases, rool lips were then 
using 
Schiffs reagent for 14 h in darkness, after hydrolysis 
in ON 


temperature (Jong, 


acid for 50 min at 
1997). 


were stained with alcoholic hydrochloric acid-carmine 


hydrochloric room 


In some Cases, meristems 


(Snow, 1963) for 3-4 days at room temperature. 
Meristem cells were isolated, macerated, and 
squashed a drop of 2% aceto-carmine after the 


first procedure or 45% acetic acid after the second 
staining method. Comparisons of metaphase chromo- 
some size on test preparations after the different 
pretreatment and staining conditions demonstrated no 
Mei- 


otic chromosomes were examined in pollen mother 


significant differences between the procedures. 


cells from squashed young anthers fixed as described 
for root apices and stained with aceto-carmine. After 
staining, slides were made permanent by freezing with 
liquid COs (Bowen, 1956), removing the coverslip and 
mounting in Euparal (Chroma, Germany), or without 
removing the coverslip (Bradley, 1948) and mounting 
in Rhenohistol (Merck, Darmstadt). 

Somatic 


and meiotic chromosomes were observed 


and photographed with a Leica DMLB > microscope 


equipped with a Leica DC 250 digital camera and the 
Leica IM 1000 image management system. Chromo- 
some numbers were determined from 14 to 177 cells 
of one to 65 seedlings of each species and variety; all 
counts were made from somatic metaphases, except 
for the accession Acosta et al. 32 (CORD), which was 
analyzed in meiosis (Table 2). Somatic chromosomes 


were measured and classified according to their arm 


ralio (r = long arm/short arm length) as recognized 
after Levan et al. (1964) with the modifications 


introduced by Schlarbaum and Tsuchiya (1984) 
(r — 1.00-1.29), 
metacentric (r = 1.30-1.69), sm—submetacentric (r 
= 1.70-2.99), st—subtelocentric (r = 3.00-6.99), 
and t—telocentric (r = 7.00 and higher). 

Leaf 
voucher specimens of N. aristata 
(Table 3). The Mann-Whitney U- 


0.01 significance level was used to test for significant 


m—metacentric msm-—meta-sub- 


made from 


and N. 


test to at least the 


and calyx measuremenls were 


rigida 


differences in the median values between vouchers of 
the same species with different ploidy level (Sokal « 


Rohlf, 


1995). Statistical analysis was performed by 


using INFOSTAT, version 1.1 (Infostat Group, 2002). 
RESULTS 
Chromosome counts from 4l populations of 20 


South American species of Nicotianeae (Solanaceae) 


belonging to Bouchetia (1 species), Fabiana (3), 
Leptoglossis (1), Petunia (2), and Nierembergia (13 
species, 6 varieties) are reported (Table 2). The 


somatic chromosome number 2n = 2x = 20 is found 


in Leptoglossis linifolia (Fig. LA), while two other 
figures are observed in Petunia, 2n = 2x = 14 in P. 


axillaris (Fig. 1B) and 2n = 2x = 18 in P. patagonica 
(Fig. 1C). In addition, 2n = 2x = 
the examined 
(Fig. 1D), F. 

(Fig. TF), 


18 is also present in 
of Fabiana, i.e.. F. densa 


denudata (Fig. LE 


species 


and F. imbricata 


and in two species of Nierembergia, N. 
ericoides (Fig. 1H) and N. scoparia (Fig. 11). On the 
2n = 2x =16 is recorded in Bouchetia 
Fig. IL) and the 
Nierembergia studied, i.e., N. aristata, N. 
lioides (Fig. 1G), N. calycina (Fig. 2A), N. graveolens 


(Fig. 2D), N. linarüfolia (Fig. 2E). N. pulchella (Fig. 2 


p 


other hand, 


anomala 


— 


remaining species of 


browal- 


B-C), N. repens (Fig. 2F), N. rivularis (Fig. 2G), N. 
rigida, N. tucumanensis (Fig. 2H), and N. veitchii 


(Fig. 21). In addition, two populations of N. aristata 
display 2n = 6x = 48 (Fig. 1K), and one sample of N. 
Ax = 32 (Fig. 1J). 

The chromosomes are small, 
3.92 Um. 
deviation) is 2.54 (0.27) um in Bouchetia anomala 
and 2.33 (0.24) um in Leptoglossis linifolia, and, 


genera where more than one species Were ne il 


rigida exhibits 2n = 
ranging from 1.52 1 


The average chromosome length ated 


Volume 93, Number 4 
2006 


Acosta et al. 
Chromosome Reports in Nicotianeae 


ranges from 2.30 (0.26)-2.90 (0.38) um in Fabiana, 
2.87 (0.30)-3.01 (0.21) um in 
value in P. patagonica), and 2.17 (0.23)-2.87 (0.30) 
um in Nierembergia. Species of the latter genus can be 
separated into two groups according to chromosome 
size. On the one hand, N. browallioides, N. linartifolia, 
N. pulchella, and N. 


chromosomes, with mean chromosome length values 


tucumanensis all have larger 


above 2.70 um. On the other hand, the remaining 


species display smaller chromosomes, with mean 
estimates below 2.40 um, except for N. 


(2.59 um) and N. ericoides (2.61 um). 


Species studied have one chromosome pair carrying 


calycina 


nucleolar organizing regions (NORs) plus attached 
satellites of variable size (e.g., Fig. 1B-E, G-H), 


— 
— 


except Bouchetia anomala (Fig. 1L), Nierembergia 
graveolens (Fig. 2D), N. linariifolia var. glabriuscula, 
N. repens, and the accession Di Fulvio 847 (CORD) of 
N. aristata, where two of such pairs were observed. In 
the examined sample of L. linifolia no NORs could be 
detected (Fig. 1A). Usually, NORs are visible in both 
members of the respective chromosome pairs, al- 
sample they 

¿ K; 2b) 


17 0 ; 


though in. several metaphases of each : 
were missing in one homologue (Figs. 1F, 
NORs appear on short arms, except in N. 
where they were seen on long arms (Fig. 2A) 

In general, the Nicotianeae studied have karyotypes 
with chromosomes that are homogeneous in size and 
predominantly of the m e The respective karyotype 
formulae are given in Table 2. Particularly, in 
browal li ioides, N. 
varieties of N. pulchella (2) 


Nierembergia, N. calycina, N., 


tucumanensis, and the 


and N. linariifolia (4) examined, only m chromosomes 
or a majority of m and one to two msm chromosomes 


were observed. In contrast, N. graveolens, N. repens, 
N. rigida, N. rivularis, and N. veitchii possessed sm 
(1-2) in addition to msm (1-2) pairs, while N. aristata 
showed three msm pairs. Leptoglossis linifolia (T sm 
pairs), P. axillaris (3 sm and 1 msm pairs), and P. 
patagonica (4 and 1), are distinct in having a minority 


of m chromosomes. Bouchetia anomala. the three 


Fabiana species, N. ericoides, and N. scoparia were 
the only taxa observed with one st pair (in addition to 
| sm and/or 1 msm pairs), which, in the last two 
species, was the smallest pair at half the length of the 
largest pair of the chromosomal | complement 

(Fig. 1H—D. 
Statistical comparison of leaf and calyx size yields 
16 and 2n 


= 6x = 48 accessions of Nierembergia aristata and 


significant differences between 2n = 2x = 


between 2n = 2x = 16 and 2n = 4x = 32 accessions 
of N. rigida, except for the calyx length in the first 
species, with polyploids exhibiting greater dimensions 
for these structures than diploids (Table 3). In 


aristata, all individuals examined of the accession 


Petunia (the higher 


Acosta et al. 32 (2n = 6x = 48) showed multivalents 
in diakinesis-metaphase I, with the 1 eee 
urations being the most frequent: 2 VI + + 16 II 
27% of the cells) and 2 VI + 18 II (1396) is not 


shown). 


DISCUSSION 
Chromosome counts ‘rom 41 populations of 20 
species and 6 varieties of Nicotianeae from South 
America belonging to Bouchetia, Fabiana, Leptoglos- 
is, Nierembergia, and Petunia are given, including 


Fabiana densa, Nierem- 


first reports from 14 taxa 
browallioides, N. 


graveolens, 


bergia calycina, N. ericoides, N. 


three out of the four varieties of N. 


linariifolia (vars. glabriuscula, pampeana, and pinifo- 


— 


lioides), two out of the three varieties of N. pulchella 
A 


(Millán) A. 


tucumanensis, N. 


(vars. pulchella and 
Cocucci & Hunz.), 


veitchii, and Petunia patagonica 


macrocalyx 


N. rivularis, N. 


and a new chromo- 


some number from V. rigida (see Table 2). 


pu 


Chromosome figures obtained in Bouchetia anom- 


ala, Fabiana denudata, F. imbricata, Mania" 


linifolia, Nierembergia aristata, N. repens, and | 
scoparia agree with previously published reports, most 
Further- 


more, our findings of 2n = 16 in five populations of N. 


of these made during meiosis (see Table 1). 


linariifolia var. linariifolia and one of N. rigida are in 
accordance with previous counts by Di Fulvio (1976a, 
1978) and suggest that n = 
by Ratera (1952, 1969), 


c observed in 


9. as cited for both species 


is unlikely. Finally, the 


— 


hromosome number Petunia axillaris 
matches previous records both in meiotic (Ferguson, 
1924; Ratera, 1952; Moscone, 1992; Stehmann et al., 
1997) and mitotic divisions (Steere, 1932: Sullivan, 
1947; Rangaswamy & Shivanna, 1967; Sink & Power, 
1978; Watanabe et al., 1996; Badr et al., 1997). 


BASIC CHROMOSOME NUMBERS 


Results confirm the basic chromosome numbers 
already known for the examined genera, i.e., Bouche- 
tia (x = 8), Fabiana (x = 9), Leptoglossis (x = 10), 
9), and Petunia (x = 7, 9) 


2001). According to the available chromo- 


Nierembergia (x = 
(Hunziker, 
some information, Nicotianeae exhibits a dysploid 


series from x — 7 to 12 and some other derived basic 
numbers, which as arranged in decreasing frequency 
are: x — 12, 9, 8, 7, 10, 23, 19, and 11 (Moscone, 
1992; Stehmann et al., 1997; Mishiba et al., 2000; 
Hunziker, 2001; Chase et al. 2003). All of these 
numbers are present in Nicotiana, where x — 12 is the 
most frequent, as it is the basic number of subgenera 
Rustica (G. Don) 


Goodsp., as well as of seven of the nine sections of 


Don) Goodsp. and Tabacum (G. 


638 Annals of the 
Missouri Botanical Garden 


Table 2. List of the taxa and samples studied (all accessions from. Argentina unless indicated otherwise). provenance, 
voucher number (all voucher specimens deposited al CORD). number of seedlings and somatic metaphases (diakinesis- 
metaphases | in pea el E x analvzed f r ipu somatic c 1 rand mo formula. ur viations of 
collectors’ name: MC; Acosta, JAA = J. A. Ambrosetti, GE > E. Barboza, i A. Cicarelli, = 
Cocucci, CC = C. Costa, i = de Dafni, EDF = Di Fulvio, LAD = L. 5 Díaz, FE = F. Ehrendorfer, | re =L. ; ale 
ATH = A. T. Hunziker, CAK = C. A. Kirkwood, YK = Y. Kurakami, MMe = M. Medina. ENS = M. Moré, E 10 = E. 
Moscone, JN = J. Nattero, SS = S. Schneckenburger, ANS = A. N. Sérsic, RS = R. Subils, MT = M. Tomoya. SV = S. Voge A 
^ In pd is indicated the number of individuals and metaphases utilized for chromosome measurements per species 
and variety. m, melacentric; msm, meta- submetacentrie; sm. ze lacentriez st, sublelocentric chromosome. * Haploid count 


in meiosis 1 sis-me taphase D. * First chromosome re port; * New chromosome number. 


No. of seedlings and 
o 


somatic metaphases 


Taxon Provenance and voucher number analyzed per sample 2n Karyotype formula’ 
Bouchetia anomala (Miers) Prov. Córdoba, Dept. San Justo. betw. 1. 3 l0 Om+lsm+lst 
Britton & Rusby Villa Concepción del Tío & Frontera (5. 6) 
Sur. EDF 635 
B. anomala Prov. Córdoba, Dept. San Justo, Marull, 30. 64 l6 — 6 m y / 
EDF 840 
B. anomala Prov. Córdoba, Dept. Punilla, El 34, 110 10 | 6m LEsm- ls 
Durazno, AAC, ANS & EAM 941. 
Fabiana densa Rémy* Prov. Tucumán, Dept. Tafí del Valle, El 12, 28 18 ny t 1 st 
Molle, ATH, GEB & EAM 24884. (3, 6) 
Fabiana denudata Miers Prov. Mendoza, Dept. Malargúe, Sierra 12, 39 18 Yom en + 1 st 
de Palauco, JAA, AC & EAM 1417. (5. 10) 
Fabiana imbricata Ruiz Prov. Chubut, Dept. Futaleufú, Mt. El 4. 18 18  Tm+lsms+ls 
& Pav. Dedal, AAC & SS 404. (J. 5) 
Leptoglossis linifolia Prov. Santiago del Estero; Dept. Ojo de 12,33 20 3m+7 sm 
(Miers) Griseb. Agua, betw. Loreto & Villa Ojo de (3. 5) 
Agua, AAC & ANS 7712. 
Nierembergia aristata Prov. Córdoba, Dept. San Justo, betw. . Ol 10 5% +3 msm 
D. Don Villa Concepción del Tío & Frontera (4. 10) 
Sur, EDF 647. 
V. aristata Prov. Santa Fe, Dept. Capital, betw. La 22. 33 48  ? 
Guardia & Colastiné Norte, RS, AAC 
& GEB 4215 
N. aristata Prov. Córdoba, Dept. Punilla, San PE 48 ? 
Roque, MCA, LAD & CAN 52", 
Nierembergia browallioides Prov. Tucumán, Dept. Tafi del Valle, 25. 8l 140 0% 2 msm 
Griseb.* betw P" a Bi lsa & EL Infiernillo, LG, (6, 10) 
AAC, ANS & SV 159. 
did calycina URUGU, i Dept. oo Paysandú, 8, 43 ló | Om+2 msm 
Hook.* AAC & ANS 109. (3, 5) 
Nierembergia ericoides Prov. Buenos Aires, Diss Tandil. 9. 2] 18 Y m E msm + 1 st 
Miers* Tandil, AAC & ANS 923. (4. 6) 
V. ericoides* URUGUAY. Dept. Maldonado, Pan de 33 18 4m d omsm + 1 st 
Azúcar. AAC 110 
V. ericotdes* URUGUAY. Dept. Lavalleja, Sierras de 2,4 18  Tm+lmimes 1 st 
Solís. AAC 1100. 
N. ericoides* Prov. Buenos Aires. Dist. Tandil, 10, 31 IE amc d msm + 1 st 
Tandil, AAC & ANS 1330. 
! ee "ns Prov. Entre Ríos, Dept. Federación. l. 14 10 C I msm l1 sm 
.-Hil.* Chajarf, AAC, ANS, MMo, JN & AD (1,5) 
3021. 
Nierembergia linariifolia Prov. Cordoba, Dept. Rio Primero, betw. 37, 85 10 Om +2 msm 
Graham var. /inartifolia Rio Primero & Santa Rosa de Rio (8, 10) 
Primero, EDF 642. 
N. linarüfolia var. Prov. Córdoba, Dept. P 0 a, Santa 7, 10 10 | Om + 2 msm 


linartifolia Maria de Punilla, EAM 


Volume 93, Number 4 Acosta et al. 639 
2006 Chromosome Reports in Nicotianeae 


Table 2. Continued. 


No. of seedlings and 


somatic metaphases 


Taxon Provenance and voucher number analyzed per sample 2n Karyotype formula’ 
N. linartifolia var. Prov. Córdoba, Dept. Capital, El l;3 l6 6m+2 msm 
linartifolia Infiernillo, AAC, Ms » EAM 944. 
N. abes var. Prov. Tucumán, Dept.“ 5 o betw. 1, 4 l6 Om+2 msm 
linartifo La Sala € Raco, m p S 1114 
y. linarifolia var. Prov. Santiago del Estero, Ee Ojo de 4, 24 l6 Om+2 msm 
linarttfolia Agua. betw. Loreto & Villa Ojo de 
Agua, AAC & ANS 7205. 
N. linariifolia var. Prov. Córdoba, Dept. Punilla, La Posta. 2. 5 l6 7m-1msm 
yu en ula (Dunal) AAC, ANS, FE & EAM 1010. (3, 6) 
Cocucci & Hunz.* 
V. oen var. Prov. Córdoba. Dept. Punilla, Dos 9, 39 0 7 m- msn 
glabriuscula* Puerta 1D 4. 
V. linartifolia var. Prov. Río Negro, Dept. Pichi Mahuida, 5. 20 l0 Im msm 
pampeana (Millán) Río Colorado. AAC & ANS 1265. (3. 5) 
A. X. Cocucci & Hunz.* 
N. linariifolia var. Prov. Córdoba, Dept. Ischilín, betw. 2,4 10 Om+2 msm 
ae asta (Millan) Dean Funes & Sauce Punco, AAC & (5. 10) 
A. A. Cocucci & Hunz.* LAD 1200. 
N. linariifolia var. "rov. Córdoba, Dept. Ischilín, betw. 9.4 16 6Om+2 msm 
pinifolioides* Dean Funes & Tulumba, AAC & ANS 
518. 
V. linariifolia var. Prov. Córdoba, Dept. Ischilín, Dean 2,3 ló Om+2 msm 
pinifolioides* Funes, MCA & LAD 13. 
N. linariifolia var. Prov. Córdoba, Dept. Punilla, La Toma, 6, 43 10 — Om +2 msm 
pinifolioides* AAC 1455. 
V. linartifolia var Prov. Córdoba, Dept. Punilla, betw. 1:3 l6 6 m 2 msm 
pinifolioides* Capilla del Monte € Ongamira, AAC 
1450. 
Vierembergia d Prov. Mendoza, Dept. Las Heras, 8, 91 ló — 7 m l msm 
Miers var. pulchella* Villavicencio, AAC, ANS & CC 1122. (4. 6) 
Nierembergia pulchella var. Prov. La Rioja, Dept. Famatina, Mt. 1, 36 l6 8m 
macrocalyx (Millán) Famatina, AAC & ANS 972, leg. (4, 6) 
A. A. Cocucci & Hunz.“ MMe. 
Nierembergia repens CHILE. Region VIII. Prov. Concepción, 10, 31 ló Om+ l msm  l sm 
Ruiz & Pav. betw. Lota € Concepción, AAC & (4, 5) 
ANS 961. 
Nierembergia rigida Prov. C m Dept. San Justo, Marull, 29, 53 l6 Sm+ lnmsm--25sm 
Miers EDF 834. (2. 6) 
N. rigida Prov. € Pm Dept. Punilla, Sant 8, 5 32 2 
María de Punilla, AAC, ANS & » AM 
943**. 
Vierembergia rivularis URUGUAY. Dept. Soriano, Villa 1. 19 10 5m-lmsm--2sm 
Miers* 1 AAC & ANS 1099, (J. 5) 
Vierembergia scoparia URUG MT 57 Paysandú, Tres 6. 36 l8 6 m * ] msm + 1 sm 
Sendtn. Boc oca E : ANS 1097. + l st (3, 5) 
Nierembergia Prov. Tuc 'umán, 55 pt. Tafi del Valle, 4. 17 l0 — 7 m msm 
tucumanensis Millán* betw. i 1555 rnillo & Amaichá del (3, 7) 
Valle, . YK & MT 1451. 
Nierembergia veitchii Prov. 11 15 Dept. Taff Viejo, San 7, 24 l6 5 % 2 msm * lsm 
Hook.* Javier, AAC & ANS 7113. (3. 5) 
Petunia axillaris (Lam.) Prov. Cordoba, Dept. Calamuchita, 6, 32 l4 3m + 1 msm +3 sm 
Britton, Stern & Poggenb. Falda del Sauce, EAM 189. (3. 6) 
Petunia patagonica Proy. Santa Cruz, Dept. Lago Pon nlino, 26, 125 18 Im + 1 msm + dsm 
(Speg.) Millán* Estancia La Victoria, AAC & (5. 6) 


471. 


640 Annals of the 
Missouri Botanical Garden 


Table 3. Comparison of leaf and calyx size between populations of different ploidy level in Nierembergia aristata and N. 
rigida. All measurements in mm. Range, mean (standard deviation), median, and sample size (n) are given for each variable. 
** Median values differ significantly between vouchers of the same species with different ploidy level (p minor to 0.01). 
Voucher specimens of accessions: N. aristata 2x, Di Fulvio 847, and 6x, Subils et al. 4215; N. rigida 2x, Di Fulvio 834, and 4x, 


Cocucci et al, 943. 


Leaf size Calyx size 
Species Ploidy level Length Width Length Calyx lobe width 
Nierembergia aristata 2x 9.00—30.00 0.50-2.00 8.00-15.00 0.50-1.20 
16.32 (5.90) 1.24 (0.36) 11.33 (2.12) 0.78 (0.18) 
14.50** 1.20** 12.00 n.s. 0.70** 
n = 50 n= 50 n= 21 n= 21 
N. aristata Ox 11.50—34.00 1.20—4.50 10.00-16.00 0.70-1.80 
21.83 (4.91) 2.49 (0.91) 12.32 (1.48) 1.10 (0.25) 
21.45** 2.10** 12.00 n.s. 1.00** 
n — 50 n — 50 n= 25 n= 25 
Nierembergia rigida 2x 6.00—24.00 0.50-0.90 7.00-22.00 0.60—1.50 
14.95 (4.28) 0.59 (0.10) 13.24 (3.79) 0.90 (0.25) 
15.00* 0.60% yes 0.80% 
n = 50 n = 50 n = 25 n = 25 
N. rigida 4x 8.00-37.00 0.50-1.20 12.50-23.50 0.70-2.00 
18.67 (7.28) 0.75 (0.19) 15.98 (2.97) 1.25 (0.29) 
18.25% 0.75 15.00% 1.20+* 
n = 50 n = 50 n = 25 n= 25 
subgenus Petunioides (G. Don) Goodsp.; however, x = Suaveolentes. and two of section Alatae. It should be 


12 is absent in the remaining genera of the tribe. In noted that the basic . numbers found by 


fact, x = 12 is by far the most common basic number us, with the exception of x = 7, are also cited in other 
in the whole family where it is recorded in more than tribes within subfamily cM A Le, x= in 
50% of the species karyologically examined. Al- — Cestreae G. Don (Cestrum L. and Vestia Wild =9 
though typical in subfamily Solanoideae, this figure is in Latueae Hunz. & Barboza (Latua pubiflora (Griseb. 
rare in subfamily Cestroideae, with the exception of — Baill.), and x = 10 in Browallieae Hunz. (Browallia 
Nicotiana, as it has been cited just in one species of L.) and Schwenckieae Hunz. (Schwenckia americana 
Schwenckia I, and in the monotypic Streptosolen IL) (Hunziker, 2001; Chiarini, 2003). 

Miers. For evolutionary considerations, a comprehensive 


— 


Both x = 9 and x = 8 appear in four out of the eight — scheme of possible chromosome number changes 
genera of Nicotianeae; thus, they are the basic during speciation in the Solanaceae is still lacking, 


chromosome numbers most widely distributed in the and even its original basic number is a matter to be 


tribe. The former number is found in Nicotiana sect. clarified. In this sense, Raven (1975) has speculated 
Alatae Goodsp. (five of nine species) and section that x = 7 could be the original number for Solanales, 
Suaveolentes Goodsp. (two species), Fabiana, Nier- while x = 12 may have derived from the tetraploid 
embergia (two species), and Petunia. In the latter level (a = 14) by aneuploid reduction in the early 
genus, x = 9 is the most common number, which history of Solanaceae, subsequently giving rise to the 


appears in those species transferred to Calibrachoa La lower numbers of the series in. Nicotianeae by the 
Llave & Lex. by Wijsman (1990) and some of Petunia same mechanism, ending in x = 7. Nevertheless, 
sensu Jussieu (1803). On the other hand, x = 8 is the chloroplast DNA sequence analyses support the 


only basic number reported for Bouchetia and ancestral placement of Cestroideae sensu D'Arcy 


Hunzikeria, the typical number of Nierembergia, and (1991) in the family and the apomorphic condition of x 
one of the numbers of Nicotiana sect. Suaveolentes, = 12 (Olmstead & Palmer, 1992; Olmstead et al.. 
where the complete series is present. With respect tox 1990). 

7, this is documented in all species of Petunia In the phylogenetic trees presented by Olmstead & 
sensu Wijsman (1990) and a few species of Nicotiana Palmer (1992) and Olmstead et al. (1999), where 
sect. Suaveolentes. Finally, x = 10 appears only in a broad sample of Solanaceae genera was included, 


Leptoglossis, several species of Nicotiana sect. one may speculate a hypothetical direction of 


Volume 93, Number 4 Acosta et al. 641 
2006 Chromosome Reports in Nicotianeae 


- 


Figure 1. Somatic metaphases of ui e Petunia, Fabiana, Nierembergia, and Bo Ds spec les. —A 5 7 

(2n = 20). —B. P. axillaris (2n — 1 uagonica (2n = 18). —D. F. densa (2n = —E. deudora (en 0 
. F. imbricata (2n — 18). —G. N. 7 allioides (2n — 16). —H. N. ericoides (2n — 18). i N. scoparia (2n — 18) t 

ri noia (2n = 32). —K. N. aristata (2n = 48). —L. B. anomala (2n = 16). Arrows ( T indicate nuc icle olar organizing regions 

(NORs). Less frequent chromosome types are indicated: m, metacentric; msm, mete sm, submetacentric; st, 

subtelocentric. Scale bar = 5 Um is the same for all figure 


— 


— 


Annals 
1 9 1 Garden 


Figure 2. Somatic metaphases of Nierembergia species s with 
C. N. pulchella var. macroc ar —D. N. graveolens. —E. N. lin 
=H, N. tucumanensis. veitchii. Arrows ( t) ) indicate nuc 


types are indicated: msm, meta-submetacentric; sm, submetacentric. Se sale bar = 


chromosome number changes in the series of 
Nicotianeae in the context of the family. The pivotal 
position of x = 11 is noteworthy: thus, this number 


could represent a plesiomorphic state in Solanaceae. 
This basic number is dominant in the basal branching 


of the family (subfam. Salpiglossoideae (Benth.) Hunz. 


and tribe Browallieae), except for the earliest. di- 


verging lineages of Schizanthus Ruiz & Pav. (subfam. 


Schizanthoideae (Miers) Hunz.) and Schwenckia (sub- 


fam. Schwenckieae), both genera displaying x = 10, 
which may be derived by a decrease in chromosome 


In 11 


suspected primitive Benthamielleae 


number. addition, x = is the number of the 


Hunz. 


e 


Hunz.) 


a 


e msm a 
sm ame 
sm 


2n = 16. —A. 


alycina. —B. N. „ var. pulchella. 
rartifolia var. pinifolioides. —F. ^ A lar 

leolar organizing regions (NORs). 
5 Um is the s 


. repens. 


rivularis 


Less frequent c sequ 
same for all figure 


(Cestroideae) (Moscone, 1989a; Hunziker. 


2001), 


a taxon not included in the molecular studies by 
Olmstead & Palmer (1992) and Olmstead et al. 
(1999). Reduction in chromosome number could also 
account for the appearance of x = 8 in all Cestreae, 
whose members are in a group sister to Browallieae, 
where x = ll is typical (see Olmstead et al., 1999: 
fies. 3, 4). 


Among the Nicotianeae, chromosomal reduction is 
also suggested by the presence of x = 9 in Fabiana 
and Petunia, both taxa forming a sister lineage to the 


10 in Leptoglossis, a genus molecularly not 


tribe Francisceae G. Don (Cestroideae) with x 


and of x 


Volume 93, Number 4 
2006 


Acosta et al. 
Chromosome Reports in Nicotianeae 


643 


examined by Olmstead & Palmer (1992) and Olm- 
stead et al. (1999). Further decrease from x — 9 to x 
= 7 8 in the monophyletic 
Bouchetia, 
Nierembergia (see Olmstead et al., 


in Petunia and to x 
group composed of Hunzikeria, and 
1999: figs. 1, 2) 
could have occurred. Under this scheme, the rare x 
9 in the latter 


genus may be a dysploid increase from x 


= 8. 


Finally, x = 12, almost universal in the mono- 


phyletic and derived cluster comprising Nicotiana, 


Solanoideae Schltdl., and Juanulloideae (Hunz.) 


Hunz. (Olmstead € Palmer, 1992; Olmstead al., 
1999). could have arisen by an increase from x = 11. 
Particularly in Nicotiana, all other basic numbers 


should be derived from x = because they are 
restricted to sections Suaveolentes and Alatae, with the 


former group being 


— 


a species assemblage geographi- 
cally isolated and most probably derived (Goodspeed, 


1954; Aoki & Ito, 2000). Additional chromosome and 


molecular data are needed to test these speculations. 


POLYPLOIDY 


All Nicotianeae species studied are diploids, 


although Nierembergia rigida and N. aristata also 
show tetraploid and hexaploid cytotypes, respective- 
ly. The presence of infraspecific polyploidy in the 
former species was unknown, while in the latter, Di 
Fulvio (1984) has cited the same levels of ploidy as 


reported here. Although polyploidy has played 
a significant role in the evolution of Nicotiana, 


where many species are amphidiploids (Chase et al.. 
2003), the 
remaining genera of Nicotianeae as so far reported 
just for Petunia spathulata L. B. Sm. & Downs (n 
18; Stehmann et al., 1997) and Hunzikeria texana 
(Torr.) D'Arcy (n. = 16; Whalen, 1979), in addition 
to the aforementioned species of Nierembergia. In the 
d 
f 


multivalent configurations (particularly hexavalents) 


it does not appear to be relevant in 


case of the hexaploid accession Acosta et al. 


(CORD) of Nierembergia aristata, the presence 


^ 


in meiosis suggests an autopolyploid origin. Con- 
cerning the polyploid samples Subils et al. 4215 
(CORD) of N. aristata and Cocucci et al. 943 


(C 920 of N. rigida, although they show a smaller 
number of NORs than expected (just one pair), they 
could have arisen by autopolyploidy, as suggested by 
Di Fulvio (1984) for another polyploid accession of 
the first species. The increased size of plant organs 
and structures in those populations compared | 
diploid counterparts both 
N. 
1984), corresponds to the 


— 


) 
in and also 


(Di 


gigas phenotype more 


species, 


reported previously for aristata Fulvio, 


frequently present in autopolyploids than in allopol- 


yploids (Singh, 1993). Further karyotypic and 


meiotic analyses of the polyploid samples here 
studied are required to cast light on their origin. 
Both Nierembergia species with polyploid cytotypes 
have wide geographical distributions, although the 
hexaploid accessions of N. aristata in particular grow 
in flooded regions, while the tetraploid population of 
N. rigida occur in hilly ha 


— 


vitats modified by man. The 
capability of polyploids to tolerate harsh ecological 
conditions and colonize disturbed habitats is well 
known (Steb 1966, 1980). In particular, the 
population Acosta et al. 32 of N. aristata seems to be 
of recent origin, because its ecological niche, at the 
border of a lake, arose about 100 years ago with the 
building of a dam in the region (Di Fulvio, 1984). The 


derived condition of polyploidy in Nierembergia is 


bins, 


suggested by the fact that diploid complements are 
widespread in the genus. 
SYSTEMATICS AND KARYOTYPE EVOLUTION 

The results of the present work suggest that the 
species of Nierembergia can be grouped according to 
their karyotype features. Among the entities with 2n 
16, two groups are distinguished. One of them, 
group J. is composed of V. browallioides, N. calycina, 


N. linaritfolia, N. pulchella, and N. tucumanensis, with 


comparatively large chromosomes and more symmet- 
rical karyotypes, composed only of m pairs or mainly 
of 
N. calycina is the only species with NORs on long 
chromosome arms, perhaps as a result of an inversion; 
thus, 


m and one to two msm pairs. It should be noted that 


it stands 


— y 


apart in group I. The other 
consists a N. aristata, N. graveolens, N. 


rigida, with 


group (IL 
N. 


smaller 


repens, 


rivularis, and N. veitchii, 


„ and more asymmetrical complements 


where one to two sm in addition lo one to two msm 
pairs or three msm pairs are present. Finally, the 


species having 2n = 18, N. ericoides and N. scoparia, 
form a third group (HI), characterized by a small st 
chromosome pair. This grouping is supported by 
external morphology, reproductive biology, habit, and 
habitat affinities. The taxa included in groups I and III 
are small shrubs (except N. calycina, which is an 
herb), usually of mountain habitats, with single pollen 
grains and a semilunar stigma. On the other hand, 
group II is composed of rhizomatous, stoloniferous or 
tuber-bearing, and creeping herbs, Pei for the 
shrubby N. graveolens. 1 lowland 
habitats and are characterized by aggregates sof pollen 


These taxa grow 


grains (except N. graveolens with single pollen grains) 
and a capitate stigma (Di Fulvio, 1976b; 
Hunziker, 1995). 


In group III. 


Cocucci & 


2n 


because 2n 


18 seems to be a derived 
16 most 
Nierembergia species and also in the related genus 


condition, = is present in 


Annals of the 
Missouri Botanical Garden 


Bouchetia. This variation in basic chromosome 
number could have arisen by a centric fission event 
during species diversification as has been postulated 


993). 


hypothesis is supported. by the presence 


This 
( 

additional small s¢ chromosome pair in the taxa an 
2n = 18. On the II. with 


morphological specializations such as much reduced 


in other Solanaceae (Moscone et : 


- 


other hand, group 


branching, more pronounced zygomorphy (in part), 
and increased karyolype asymmetry, appears to have 
an advanced status relative to group | (Cocucci & 
Hunziker, 19095). 

The similar karyotypes in the shrubby species of 
Fabiana studied is striking because the plants are 
morphologically very different from each other, i.e., F. 
imbricata, tall with comparatively imbricate leaves 
and conspicuous flowers, as well as F. densa and F. 
denudata, both comparatively small, but the former 
of linear leaves and the latter 
1993). In 


chromosome 


with a dense foliage 
nearly without leaves (Barboza & Hunziker, 
Petunia, both species studied differ 
number but not in karyotype asymmetry. In particular, 
P. axillaris has a similar karyotype to P. hybrida 
(Hook.) Vilm.. 
Wijsman (1990) (a group with x = 7). On the other 


hand, the chromosome complement of P. patagonica 


another member of Petunia sensu 


does not differ from that of P. linearis (Hook.) Paxt.. 


a species from the group with x = 9, which is now 


considered as Calibrachoa (Wijsman et al., 1983; 
Maizonnier, 1984). Chromosome features together 


with morphological traits, such as the corolla. not 
entirely single colored and calyx lobes pentafid and 
not deeply incised, indicate that P. patagonica should 
be included in the Calibrachoa group (Wijsman & de 
Jong, 1985; 1993; Mishiba et al.. 
2000). In general, Petunia sensu Wijsman displays 


advanced morphological features in 


Figueroa Romero, 


— 


comparison to 


Calibrachoa and the related genus Fabiana, as it 


comprises annual herbs without woody stems and 
brachyblasts (Barboza & Hunziker, 1993; Kulcheski 
2000). 


et al., 


The present results demonstrate that some genera of 


Nicotianeae, Fabiana, Nierembergia, and Petu- 
nia, display general karyotype patterns where unifor- 
mity in gross chromosome morphology (i.e., similar 
chromosome shape between taxa) is maintained even 
This occurs in 


1954). 


orthoselection concerning 


after changes in chromosome number. 
several sections of Nicotiana (Goodspeed, 
Species-wide karyotypic 
both chromosome number and morphology has been 
observed in other genera of Solanaceae (e.g.. Cestrum, 
Capsicum L., and Solanum L.) and other plant groups 
including an entire angiosperm family, as in the case 
1989b: Bernardello & 


of the Aloaceae (Moscone, 


Anderson, 1990; 1993; Brandham & 
Doherty, 1998). 
Karyotypic data suggest that Bouchetia, with x = 8 


case of B. 


anomala, is closely related to Nierembergia. On the 


Moscone et al., 


and a rather symmetrical karyotype in the 


other hand, Leptoglossis, Petunia, and Fabiana, 


without x = 8 and with more asymmetrical comple- 


ments except in the latter genus, appear more distant, 


in agreement with the suprageneric classification 
proposed by Hunziker (2001). Finally, the basic 
chromosome number recorded for Hunzikeria (x = 8) 


indicates affinity to Bouchetia and Nierembergia and 
Leptoglossis, in 


i more 


distant relationship to 
disagreement with Hunziker’s (2001) system. 


Literature Cited 


Aoki, S. & M. Ito. 2000. Molecular phylogeny of Nicotiana 
(Solanaceae l o d on the ı enw sequence of the 
matK gene. Pl. Biol. ( pan 2: 316-324. 

Badr, A., iif 1 a & M. M. Abou-El- 
Enain. : M ( jener criteria and taxonomic. rela- 


103-113. 


Estudios en 


tionships in the wa P Cytologia 62: 
Barboza, G. E. & T. 1993. 


Revisión taxonómica de 


Hunziker. 
Fabiana. 


1990). 
studies in Solanum section Basarthrum (Solanaceae). 
Amer. J. Bot. 77: 420431. 

Bowen, C. C. 
making slides permanent. Stain Technol. 31: 87- 

Bradley, M. V. 1948. A method for 11 5 acelo-carmine 

dao nt without removal of the cover slip. 


Anderson. Karyotypic 


1956. Freezing by liquid a dioxide 
90 


squashes 
Stain ae 41— 
2 "RI Doherty. 1998. 


aceae, an angiosperm family displaying 


Brandham. Genome size 


variation in e 


karyol ypic 9 755 lection. Ann. Bot. (Oxford) 82 (Suppl. 
A): 67-7: 
Chase, M. W.. 5. Knapp. A. V. Cox, " E gee ', Butsko, 
Joseph, Savolainen & A "arokonny. 2003 


Mole cular systematics, GISH and n origin of hybrid laxa 
| 


(Oxford) 92: 


Ann. ot. 


in E oti ana (Solanaceae 
11 "n 
Chiarini. ii E. 2003. Los 


cromosomas 
O americana (Solanaceae). Bol. Soe. 
. 38: 325-327 


somáticos de 
S Argent. 
ot. 38: 325 


1 d, A. A. 


(Solanaceae 
& A. 


e COLIMA NOS biology of Nierembergia 
yy 35 


Evol. 174: 17— 


Im 


-l4 in 


F lus Fa anerogámica Argentina, 
| 


. Nierembergia. 


ES 
zx 


T. S (editor). 
256 Solan: iceae, Nierembergiinae. 


Cosa de Gastiazoro, M. T. 1989. Mapu ación vegetativa en 
5 sta 


Solanaceae: Nierembergia stricta, N. arts Eras 
anomala v re linifolia. Kurtziana 9 

D'Arcy, W. 199]. The Solanaceae since nee er? 
a review " ils biogeography. Pp. 75-137 in J. G. Hawkes. 

. N. Lester, M. Nee & N. Estrada (editors), Solanaceae 

Hl: 1 Chemistry, Evolution. Royal Botanic 
Gardens, Kew. 

Desai, S. R. 197 Hi 


omosome numbers and E in 
Phytor y: 199-2 
19 )76a. 


three Solanaceae. uenos Aires): ! 
Di Fulvio, T. E. 


Nierembergia re Feo ). Kurtziana 9: 141— 
a 


y casi cromosómicos en 
DL 142. 


Volume 93, Number 4 
2006 


Acosta 645 


et al. 
Chromosome Reports in Nicotianeae 


976b. m el om de Nierembergia (Solana- 

ceae). Puis 9: 
1978 ee 

Kurtzia 


gamélicos de 


e 


. Cromosomas 
aristata y Bouchetia anomala (Solanaceae). 
11: 118 

984. Poliploidía en Nierembergia (Solanaceae). 
Pine an me 25-30. 


Diers. L. 1961. Der Anteil an Polyploiden in den 
In der Westkordillere Perus. Z. Bot. 49: 


Veget: 1 
437—488 
Ferguson, M. C. announcement of 
a cylological and of Petunia. Tus 
Rec. 29: 137. (Not seen: 9 from Goodspeed, 1954 
kl Mp N, e 


1924. 


a genetical 


Preliminary 
study 


Figueroa Romero, M. R. 


ceae), Morfología y Toni Ph. hesis, University 
of Tucumán, Arge 
. 1999, Eua Pp. 1084—1086 in F. O. Zuloaga & 


O. ene (editors), Catálogo de las Plantas Vasculares 


de la Rep P Argentina II. Monogr. Syst. Bot. Missouri 
Ws Cand 
Gil, I Chai, J. M. Pezzuto & C. 


B3 Li 
C 12 1995. € ade 1 from Nierembergia aristata. J. 


Nat. Prod. (Lloydia) 58: 848-850. 
Godo, T., O. Tsujii, K. Ishikawa € M. Mu. 1997. Fertile 
transgenic plants of Nierembergia scoparia Sendtner 


obtained by a eap a 11 strain of Agrobacterium 
Sci. Hort. 08: -]11. 


T. H. 1933. kt number and | morphol: 


rhizoge nes. 


Goodspeed, 


= 


hromosome mbers 
. U.S.A. ic ): 649-653. 
1-530. 


Genera of 


ogy in Nicotiana VI. orty 
species. 1 55 Natl. Acad. Sci 
. 195 900 Genus Nicotiana, üben. Bot. 16: 
A. - The ( 
Solanaceae 11 1 1 Arranged According to a 
System. Cantner Verlag. 
— R. Subils. 1979. Salpiglossis, Leptoglossis and 
9 (Solanaceae), a synoptical survey. Bot. Mus. Leafl. 
: 1-43. 
1 Group. 2002. INFOS TAT versión 1.1. Facultad de 


Ciencas Agropecuarias, "a srsidad Nacional 15 C órdoba. 
I 


1 . Genera Solanacearum. 


ew 
Liechtenstein. 


Editorial Brujas, Argentina. 

Japan Tobacco. 1994 155 Genus Nicotiana Illustrated. 
Japan Tobacco, Tokyo. 

Jong, K. 1997. Laboratory Manual of Plant ee 
Techniques. Royal Botanic ROS n, Edinburgl 

Jussieu, A. L. 1803. Sur le Petunia, genre nouveau de i 
famille des esl Ann. Mus. Natl. Hist. Nat. 
214-216 

Kulcheski I . Muschner, A. P. Lorenz-Lemke, J. 
Stehmann, S. L. Bonatto, F. M. Salzano & L. B. Paias 
006. Molecular yg netic Ge of Petunia Juss. 
(Solanaceae). Genetica 126: 3-14 

Levan, A., K. Fredga & Š A. be 1964. Nomenclature 


for oe pers on chromosomes. Hereditas 52: 
201— 

Maior, D. 1984. Pp. 21-33 in H. C. 

cu Springer-Verlag, Berlin 

. Revision de las especies de Nierembergia 

: 487-547. 

Mishiba K.-L. T. Ando, M. Mii J. Watanabe, H. 
x Hasliimólorde E Marchesi. 2000. N 
as an index character discriminating taxa in the 
Petunia sensu Jussieu (Solanaceae). Ann. Bot. (Oxf ord) 85: 

5-673. 

Moscone, E. A. 10989a. 
Patagonian and S. Andean endemic genera of Nicotianeae 
(Solanaceae). Pl. Syst. Evol. 166: 31-39. 


Cytology. Sink 


Kokubun, 


A content 


uclear DN 


Karyotype analyses in 


1989b. Estudios Citotaxonómicos en las Tribus 
ie) de América del Sur. 
. Argentina. 
 meióticos en 
61-297. 


Ehrendorfer. 


Estudios de cromosomas 
Solanaceae de Argentina. 
ambrou, A. T. Hunziker & F. 
1993. Gier msa C- bandet 1 in Capsicum (Solana- 
. Evol. 213-229. 
Olmstead, R. G. 55 pins 1992. A chloroplast DNA 
phylogeny of ilis Solanaceae: Subfamilial ou 
Bot. 


Darwiniana 31: 
Hunziker & 
ceae). PI. 


and character evolution. Ann. Missouri Gard. 
346—300. 

———, J. A. Sweere, R. E. Spangler, L. Bohs & J. 
Palen 1999, Phylogeny and provisions) puc 'alion 2 
the Solanaceae based on chloroplas 9 9 . Pp. 111—137 in 
M. Nee, D. E. Symon, R. N. Leste P 1 (editors), 
Solanac eae IV: Advances in o and Utilization. Royal 

anic Gardens, Kew. 

Induction of 

axenic 


. R. Shivanna. 1967. 
formation in 


une 
gamete com apatibility and 
cultures of a 1 5 sel Eincompatible species of Petunia. 
Nature 216: 937-930 

Ratera, E. L. 
Solanáceas argentinas (Tercera contribución). Revista Fac. 
Agron. o erin, 13: 280-284. 

969. Estudios 
Bot. 


The bu of angiosperm phyloge ny: 


seed 


052. c" de cromosomas de algunas 


— 
— 


Paneg ‘os en Solanáceas. 
Bake 1 Munic. 


haven. P. H. 


Cytology. Ann. Missouri Bot. Gard. 62: 724—704. 
Schlarbaum, S. E. & T. Tsuchiya. 1984. The chromosomes of 


konishii. C. 


Cunninghamia 
| (Taxodiaceae). Pl. 


lanceolata, and Taiwania 
cryptomerioides ). Syst. Evol. 145 
5 
Shizukawa, 
regene ralion system 


.&M. 1997. A simple and e ino ¡ent plant 
or leaf protoplasts of Nierembergia 
PI. 


repens by ee ing in shoots on the microcolonies. 


wal 
= 


Cell. Rep. " 


Singh, R. J. 993. p Cytogenetics. CRC Press, Boca 
Raton. 
Sink, K. C. & J. B. Power. 1978. Incongruity of interspecific 


and inle rgene ric crosses inv olv ang Nicotiana and Petunia 


species that for somatic hybridization. 
Euphytica 27: 725-730. 

Snow, R. 1963. Alcoholic 

Stain p chromosomes in squash preparations. Stain 


exhibit. potential 
72 


hydrochloric acid-carmine as 
Technol. —]: 

Sokal, R. R. FJ. Rohlf. 
New Yo 


INEW 0 


995. Biometry. 3rd ed. Freeman, 
Stebbins, C. 5 Chromosomal variation and evolution. 
Science ] "S 1463-1469, 
1980. loli in plants: Unsolved proble ms and 
prospects. Pp. 495-520 in W. H. Lewis (editor), Poly- 
ological Relevance. Ple num “Press, New York. 
932. Chromosome be 9070 ior in triploid Petunia 
bd ee J. Bot. 19: 340— 
Stehmann, J. R., J. A. H. Dutilh. J. Se mir 1 & E. R. Forni-Martins. 
97. Solanaceae. P. 24 in A. Stace Seis IOPB 
Chromosome Data 11. Int. Organ. PI. Biosyst. Newslett. 26/27. 
Cromosomas a de a 
—20. 


ploidy. 


Steere, 


Subils, R. 1979 
linifolia (Solanaceae). Lore ntzia 3: 19-2 
Sullivan, T. 1947. Soma 
hybrid Petunia. Bull. Toney Bot. Club 74: 453—47 


c 1 of pedigreed 


Watanabe, H., Ando, S. lida, A. Suzuki, K. Bio. T 
Tukamoto, G ashimoto & E. Marchesi. 1996. Cross 


compatibility of Petunia cultivars and P. axillaris with 
native taxa of Petunia in relation to their chromosome 
number. J. Hort. Assoc. Japan 65: 625-6: 


Annals of the 


646 

Missouri Botanical Garden 
Whalen, M. D. 1979. Solanaceae. P. 276 in A. Love — — & J. H. de Jong. 1985. On the interrelationships of 
(e 0 0 IOPB Chromosome Number Reports EXIT. certain species S Petunia IV. Hybridization between P. 
Taxon 28. linearis and P. calycina and nomene latorial consequences 

Wi ijsma in, H. J. W. 1990. On the inter-rel: . of certain in the Petunia group. Acta Bot. Neerl. 34: 337—349 
species of Petu ew names for the species of ———— & T. M. Pedersen. 1983. On the interre A 
Cali hoa forme " inc elude : into Petunia (Solan: ceace). ships as certain species of Petunia VI. The position of P. 

| Bot. Neerl. 39: 101-10 12 and P. calycina. Acta Bot. Neerl. 32 323.332 


CHROMOSOME STUDIES IN 
AMERICAN 
PANICOIDEAE (POACEAE)! 


Osvaldo Morrone,” Alejandro Escobar,” and 
Fernando O. Zuloaga” 


ABSTRACT 


Chromosome numbers and observations on meiotic chromosome behavior of 37 taxa of American Panicoideae are given. Th 
ied are: Axonopus d is (Flüggé) Kuhlm $ Kt anth, "pare aciculare var. 
(Desv. ex Poir.) Gould & C. A. Clark, D. acuminatum var. ) Gould & C. A Clark 
Michael ex Renvoize, Pu. hloa montevidensis Griseb.. 


. Genchrus myosuroides 
acuminatum (Sw 
seb.. /sachne arundinacea (Sw.) i 

Nees, Paspalum coi Chapm., P. buchtienti 17 P. pees us P. candidum 
Pu B ‘Tl i 
ekmanianum Hen 
pn ens Hack., 


P. ellipticum Döll. P. 


P. 


parc Kunth, P. ec (Hack.) Morrone a Zuloaga, 
vartwegtanum E. ED rn., P. inconstans Chase. P. indecorum Mez, P. 


201 


. Fourn l.) ash. 


> 


P. squamulatum E. Fourn 


lorentziana (Mez) Morrone Zuloaga. 


72), Panicum stoloniferum (2n = 20), Paspalum puo Jhitienii (2 
60). 


E (2n = 20), P. inconstans (2n = ca. ; 

= previously published counts: Axonopus scoparius (2n = 
20), e DUM (2n — 20), P. 

P saccharoides 


0), P. squamulatum (2n = 20) 


Key words: pm chromosome numbers. 


js Michx.. P. e Pa ribn. & Merr. 
che. 


ı = 20) 

MEDIA (2n = 40), i 
ellipticum (2n = 40), P. 
and Urochloa loreniziana (2n = 36). 
Panicoideae. 


P. juergensii Hack., P. pow 
H. saccharoides Nees ex Trin., 
achi fusca (Sw.) B. F. 2 N & 
duda for dci iain chacoensis (An = ce 
20). p. 


Ten counts differed 


is hase, 
. P. ceresia (2n 40). P. ekmanianum (2n = 
and P. variable ^ = 40). 

72), Isachne arundinacea 2n 
40). P. hartwegianum (2n = 40), 


Cenchrus myosuroides (2n = ca. 


Poaceae. 


Knowledge of chromosome numbers and their 
meiotic behavior is important for understanding the 
phylogeny of vascular plants (Raven, 1975), particu- 
larly with regard to polyploid taxa (Davidse et al., 

1986). Polyploidy is a common feature in the Poaceae, 
but karyological studies are still lacking in many 


991), 


limiting our ability to understand the patterns. of 


— 


tropical and subtropical grasses (Honfi et al., 


chromosome number change across the family. In 
addition, knowledge of the relationship between 
polyploidy and apomixis in tropical genera of grasses, 
& Chase) Gould, 
Paspalum L., and Urochloa P. Beauv., 1s 


such as Dichanthelium (Hitche. 
Panicum L., 
necessary for plant breeding programs and for studies 


1979; Quarín & 


of reproductive. systems (Connor, 


Norrmann, 1987; Honfi et al., 1991: Norrmann et al., 
1994 

The purpose of this work is to increase the 
cytological knowledge of species of Panicoideae 


collected different regions of America. Chromo- 
some counts were made in taxa of the tribe Isachneae 
(Isachne) and Paniceae 


L., Dichanthelium, 


p 
E 


Ixonopus P. Beauv., Cenchrus 


Echinochloa P. 


Beauv.. Ertochloa 


Kunth, Panicum, Paspalum, and Urochloa). Chromo- 
some numbers were compared with previous results 
published in the Poaceae, including the following 


indexes: Ornduff (1967, 1968, 1969), Bolkhovskikh et 


al. (1969), Moore (1970, 1971, 1972, 1973, 1974, 
1977), Goldblatt (1981. 1984, 1985, 1988). and 
Goldblatt and Johnson (1990, 1991, 1994, 1990, 


1998, 2000). 
MATERIAL AND METHODS 


Meiosis was studied in five to 10 pollen mother 
cells per sample. All counts but one here reported 
were made from meiotic material fixed in the field 
with Carnoy’s solution (6 parts ethanol:3 chloroform: | 
acetic acid); chromosomes were stained with 296 iron 
1960: Núñez. 1968) 


Permanent slides were 


propionic haematoxylin (Sáez. 
and with 2% acetic carmine. 
obtained by the dry method or by Bradley's technique 
(i.e., slides are dehydrated in absolute alcohol and 
mounted in Euparal) (Bradley, 1948). A complete list 
of the species, chromosome numbers, and voucher 


specimens studied is given in Table 


Científica Técnicas (CA CET) for research. grant P 
í 


s “Instituto de Botanic a Darwinion. 


omorronc edu 


ANN. Missouri Bor. 


Labardén 200, Casilla de Correo 22, San Isidro B1642HYD 


GARD. 93: 647—657. PUBLISHED ON 


! We thank the staff of Instituto de Botánica Darwinion for support. We also thank the Consejo Nacional de Investigaciones 

IP-02131, ANPCyT (Agencia Nacional de Promoción Científica 
Técnica, bue ntina) for grants 11739 and 13374, and the National Ge 'ographic Society for research ¢ 
97 


grants 6825-00 and 6042 2. 


. Buenos Aires, Argentina. 


DECEMBER 2006. 


Annals 


of the 
Missouri Botanical Garden 


Table 1. 


previously cytologically studied, ** 


Chromosome numbers and configuration chromosome associations in taxa of American Panicoideae. 


l'axa with chromosome numbers that differ 


*Taxa not 


from previously published reports. 


Taxa 


Configurations observed in 
diakinesis of metaphase | 


and mitotic metaphase 


Origin and voucher 


**Axonopus scoparius (Flüggé) Kuhlm. 


* ( Kunth 


venchrus myosuroides 


Dic hanthe ac ic pida Var. 


ex Poir.) Gould & C. Clark 


Dichanthelium 5 var. aciculare (Desv. 


ex Poir.) Gould & C. A. Clark 


Dichanthelium acuminatum var. acuminatum 


(Sw.) Gould & C. A. Clark 


W. Michael ex 


* Echinochloa chacoensis P. 
Renvoize 
Eriochloa montevidensis Griseb. 
schine arundinacea (Sw.) Griseb. 
Panicum hirticaule J. Presl 
*Panicum stoloniferum Poir. 
Panicum trichanthum Nees 
Paspalum blodgettii Chapm. 
i 8 | 
*Paspalum buchtienti Hack. 
*Paspalum buchtienii Hack. 


Paspalum caespitum Vlüggé 


Paspalum candidum (Humb. & Bonpl. ex 
Vlüggé) Kunth 


*Paspalum ceresia (Kuntze) Chase 


Wright 


Paspalum clavuliferum C. 


aciculare (Desv. 


3011 


91 


9 II 


911 


ca. 36 II 


18 H 


10 H 


9 Il 


10 H 


18 H 


20 11 


2n = 2 


10 H 


10 H 


30 H 
28 11+ 41 
20 II 


10 I 


BOLIVIA. La Paz: Prov. 
Yungas, alrededores de Coroico, Morrone et 
al. 4209 (LPB, SI). 

ARGENTINA. Jujuy: Dpto. Dr. M. Belgrano, de 
Aeropuerto El Cadillal a SS de Juj 
Zuloaga & Morrone 7049 (MO, B 

MEXICO. Veracruz: Carretera 140 de 
a Perote, El Acajete, Zuloaga et al. 
(MEXU, SI). 

MEXICO, Oaxaca: Carretera de 5 
a Huautla, : 15 km la de 
al. 7395 (MEXU, S 

MEXICO. Die ‘a: Carretera de San Jerónimo 

2 km W de Huaulla, Zuloaga et 
al. 7392 (MEXU, SI). 

ARGENTINA. Jujuy: Dpto. Palpalá, Ruta el 
algarrobal a El Cucho, Zuloaga et al. 5848 
(MO, SI). 

ARGENTINA. Jujuy: Dpto. Dr. M. Belgrano. 

Zapla, Mina 9 de Octubre, Morrone et al. 
4422 (MO, SI). 

BOLIVIA. La Paz: Prov. Sud Yungas. camino Los 
Yungas, a 57 km de La Paz camino a Coroico, 

1192 (LPB. SI). 

Ruta 110, camino 

ora pasando Villarrica, Morrone & 
Giussani 3041 (MEXU, SI). 

BOLIVIA. Santa Cruz: 
58 km de 5. Cruz camino a Samaipata, 
Morrone & fs 4268 (LPB. SD. 

ARGENTINA. Jujuy: Dpto. Santa Bárbara, 

21 km de Palma Sola camino a Puesto Maíz 
Gordo. Morrone et al. 4451 (MO, SI). 

MEXICO. Quintana Roo: Ruta Mex-307, 43 km 
al 5 de Playa del Carmen a Tulum, 1 & 
Giussani 3058 (MEXU, SI). 

BOLIVIA. Santa Cruz: Prov. 
58 km de Santa Cruz camino a pou 
1 8 & Giussani 4258 (LPB, SD. 

BOLIVIA. La 
1 1 a 63 km de La ind camino a Coroico, 
Morrone et al. 4201 (LPB, 

MEXICO. a Roo: bs Mex-307, 43 km 
al S de Playa del Carmen a Emus Morrone & 
A 3656 (MEXU. 

MEXICO. Oaxaca: San Jerónimo, Zuloaga et al. 
7396 ve XU, SI). 

ARGENTINA. Jujuy: Dpto. Dr. M. Belgrano, 

10 km de León camino a Tiraxi, 
al. 4386 (MO. SI). 

BRAZIL. Bahia: Mun. 
Mucugé a Barra da Estiva, 6 a i4 
Zuloaga & Morrone 6078 (MO, 


Nor Yungas, camino Los 


peus 


San Jerónimo 


Huaulla, Zuloaga el 


a Huautla, 22 


N 


ai et al. 
MEXIC 


se 


. Mie de 'ün: 


Prov. Andrés Ibanez, 


Andrés Ibanez, 


Paz: Prov. Nor Yungas, camino Los 


Morrone et 


BA-142, de 
Mucugé, 


Mucugé, 


Volume 93, Number 4 
2006 


Morrone et al. 
Chromosome Studies 


649 
in Panicoideae 


Table J. Continued. 


Taxa 


Configurations observed in 
diakinesis of metaphase | 
and mitotic metaphase 


Origin and voucher 


Paspalum commune Lillo 


**Paspalum denticulatum Trin. 


*Paspalum ekmanianum Henrard 


** Paspalum ellipticum Doll 


** Paspalum fimbriatum Kunth 


*Paspalum glabrinode (Hack.) Morrone & 


Zuloaga 


Paspalum glaucescens Hack. 


**Paspalum hartwegianum E. Fourn. 


*Paspalum inconstans Chase 


Paspalum indecorum Mez 


Paspalum juergensii Hack. 


Paspalum pauciciliatum (Parodi) Herter 


*Paspalum penicillatum Hook. f. 


Paspalum plicatulum Michx. 


Paspalum prostratum Scribn. & Merr. 


**Paspalum saccharoides Nees ex Trin. 


Fourn. 


**Paspalum squamulatum E. 


Paspalum squamulatum E. Fourn. 


20 II 
19 11+21 


10 I 


10 H 


20 H 
1911 +21 
18 1+ 4I 

2 1V+ 131461 


20 Il 
ca. 30 H 
2n — 20 

10 H 

2n — ca. 38 

ca. 20 H 

18 +41 
10 H 
ca. 20 Il 
20 Il 
10 H 


20 M 


ARGENTINA. Jujuy: Dpto. Dr. M. Belgrano. 


Morrone et al. 


N 


apla, Mina 9 de Octubre. 
4420 (MO, si). 

MEXICO. Puebla: Carretera de Tehuacán 
a Zapotitlán, km 10, cerca £x ps Ta, 
510 0 et al. 7384 (MEX 

ARG NA. Jujuy: Dpto. x 2 borum 
UA Mina 9 de Octubre, Zuloaga et al. 
852 (MO, 5I). 

de Goiás: 
(IBGE, SI). 

MEXICO. Quirtana Roo: Mun. Cozumel, 
Cozumel, Laguna C hankan aab, Morrone & 
Giussani 3651 (MEXU, SI). 

ARGENTINA. Misiones: Dpto. 
Peñón de La Reina Victoria, Zuloaga & 

Morrone 6809 (MO, SI). 

ARGENTINA. Formosa: Ruta Nac. 11, de 
E a 10 Zuloaga & Morrone 


7 (M( 


Alto Paraiso, Filgueiras 3517 


Isla de 


San Ignacio, 


MEXICO. Michoacan: Ruta 15d, km 159, 
pasando el Mun. de Maravatio, Puente Satire, 
Morrone & Giussant 3612 (MEXU, SI). 

BOLIVIA. La Paz: Prov. Sud Yungas, camino Los 
Yungas, a 50 km de La Paz camino a Coroico, 
Morrone et al. 4190 (LPB, SI). 

ARGENTINA. Misiones: Dpto. Capital, Ruta 
Nac. 105, de 157 5 a San José, Zuloaga & 
Morrone 7142 (M 

BOLIVIA. La Paz: E pe Yungas, a 56 km de 
La Paz camino a Coroico. Morrone et al. 4188 
(LPB, SI). 

ARGENTINA. Misiones: Dpto. San Ignacio, 
Jardín América, Zuloaga & Morrone 7191 
(MO, SI). 

BOL VIA. La Paz: Prov. Sud Yungas, camino Los 
Yungas, a 57 km de La Paz camino a Coroico, 
Morrone « et "i 4200 (LPB, SI). 

MEXICO. Veracruz: Autopista de La Isla 
a Acayucán, km 66, Zuloaga et al. 7418 
(MEXU, SI). 

MEXICO. Puebla: Carretera de Teotitlan 
a Huauclan, 20 km de Teotitlán, Zuloaga et 

al. 7387 (MEXU, SD 

BOLIVIA. La Paz: Nor Yungas, a 23 km de 
Yolosa camino hacia Caranavi, Morrone et al. 
4211 (LPB, SI). 

MEXICO. Oaxaca: Carretera de 5 
a one 22 km W de Huautla, Zuloaga et 
al. 9 (MEXU, SI). 

11 5 Veracruz: Carretera 140 de 
a Perote, El Acajete, Zuloaga et al. 7443 
(MEXU, SI). 


San Jerónimo 


Xalapa 


Annals of the 
Missouri Botanical Garden 


Table 1 


Continued, 


Configurations observed in 


diakinesis of metaphase 


Taxa 


and mitotic metaphase 


Origin and voucher 


*Paspalum variabile (E. Fourn.) Nash 20 I MEXICO, . Carretera entre San Jerónimo 
191-21 y Huautla, 12 km de Huautla. laa et al. 
7397 (ME 8 SI). 
Paspalum wrightii Miche. & Chase 10 I ME XIC O. Oaxaca: Carretera Tuxtepec a Loma 
Bonita, 20 km de Tuxtepec, Zuloaga et al. 
7411 (MENU, SI 
Urochloa fusca (Sw.) B. F. Hansen & 9 Il ARGENTINA. Jujuy: Date: Ledesma, Ruta Prov. 
Wunderlin 31.7 km del desvío de la Ruta Prov. | camino 
a VM iat Morrone et al. 2914 (MO, SI). 
** Urochloa lorentziana (Mez) Morrone & 18 H ARGENTINA. Tucumán: Dpto. Trancas, Ruta 


Zuloag 


Nac. 9, 80 a de R. de la Frontera camino 
a S.M. de Tucumán, Morrone et al. 4653 (MO. 
Sp. 


Voucher specimens were deposited at the Instituto 
de Botánica Darwinion (SD), with duplicates at LPB. 
MEXU, and MO. 

RESULTS AND. DISCUSSION 

All studied species are summarized in Table 
including the examined configurations, voucher speci- 
mens, their provenance, and the respective collector. 
Collections are alphabetically ordered by genus and 
and are 


that order. 


Photomicrographs of selected species are shown in 


species discussed below in 
Figures | and 2 

Most results agreed with previously published 
chromosome numbers, but our counts differed for 10 
species: these taxa were Axonopus scoparius (Flüggé) 
Kuhlm. (2n = 60), Cenchrus myosuroides Kunth (2n = 
ca. 72), Isachne arundinacea (Sw.) Griseb. (2n = 20). 
Paspalum denticulatum Trin. (2n = 20), P. ellipticum 
Doll (2n = 40), P. fimbriatum Kunth (2n = ca. 40), P 
hartwegianum E. Fourn. (2n. = 40), P. 
Nees ex Trin. (2n = 40), P. squamulatum E. Fourn. 
(2n = 20), and Urochloa lorentziana (Mez) Morrone & 
Zuloaga (2n = 30). 


saccharoides 


Nine additional species were previously unknown 
eytogenetically. These first reports for chromosome 
“© Echinoc jaa chacoensis P. W. Michael ex 
Renvoize (2n. = ca. 
(2n 
ceresia (Kuntze) Chase (2n. = 
Henrard (2n = 20). P. 
Zuloaga (2n = 20). P. inconstans Chase (2n = 
P. penicillatum Wook. f. (2n. = 40), and P. 
(E. Fourn.) Nash (2n = 40). 


Axonopus scoparius, a 


number ar 
. Panicum stoloniferum Poir. 
= 20), lan las Hack. (2n = 20), P. 
40), P. ekmanianum 
glabrinode (Hack.) Morrone & 
ca. 60). 


variabile 


series 
Barbigeri, subseries Scoparii (sensu Black, 1963). is 


species of the 


an imporlant natural forage grass in the Andean 
region. The specimen studied, from Bolivia, is 


a hexaploid with a count of 2n = 60 (Fig. LA), which 
differs from previous counts published by Pohl and 
Davidse (1971) and Shibata (1962), who reported 2n 


— 20 for this species. 


The widespread genus Cenchrus, with approximate- 


ly 20 pantropical species, has base chromosome 
numbers of x = 9, 10, and 17 (Núñez. 1952: DeLisle, 
1963). Cenchrus myosuroides has 2n = ca. 72. 


a number which differs from pom Mn for this 


species: 2n = 54 (Brown, 1950), 2n = ca. 68 (Bowden 


& Senn, 1962), and 2n = 70 (Parodi, 17 5 Gould. 
1965: Reeder, 1968). 
The American genus Dichanthelium includes 


several polymorphic complexes of uncertain specific 
delimitation, with different foliar and floral dimorph- 
isms correlated with seasonal variability (Gould & 
Clark, 197 1993). A high percentage 
of diploids, approximately 80% 


8: Zuloaga et al., 
„ have been reported 
for North and Central American species of Dichanthe- 


lium (Dubcovsky & Zuloaga, 1991). The genus has 
a base chromosome number of x 9 (Brown, 1948; 


1966. 


D. aciculare (Desv. ex Poir.) 


Gould, 1968), and the chromosome counts for 
Gould & C. A. Clark and 
A. Clark. 
Mexican specimens, are diploid with 2n = 
that 
(Brown, 


D. acuminatum (Sw.) Gould & C. based on 
18. a result 
previous reports for these species 
1948: Gould. 1958: Dubcovsky & Zuloaga. 


1991, cited under the genus Panicum). 


confirms 


The genus Echinochloa includes nearly 40 species, 
many of them dominant in wetlands or in disturbed 
open areas, including cultivated fields, of tropical and 
temperate zones. Echinochloa chacoensis, a species 
restricted to eastern Bolivia and northern Argentina. 


is an octoploid with 2n = ca. 72, the first count for 


Volume 93, Number 4 Morrone et al. 651 
2006 Chromosome Studies in Panicoideae 


b o E „5 of meiotic chromosomes of American Panicoideae. -—A. Axonopus d 30 II. 

aphas . Eriochloa a 18 I, metaphase I. —C. Panicum pe bie, 10 IL, prometaphase 
Paspalum aa. 10 yi ase I. —E. Panicum trichanthum, 18 IL, prometaphase |. — Bees M 18 
Palm denticulatum, 10 II, prometaphase I. —H. Paspalum candidum, 30 I, diakinesis. Arrows 


E. 1 1 15 3 je — 20 


652 Annals of t 
Missouri 1 Garden 


G 


rure 2. ee ene of meiotic chromosomes of American Panicoideae. —A. Paspalum buchtienii, 10 
prometaphase I. —B. Paspalum ceresia, 20 H, metapha Um Paspalum commune, bridge at anaphase II. —D. don 
lorentziana 115 II. metaphase 1. —E. Paspalum 1 ens, + 2L prome ee I. —F. Paspalum ekmanianum, 10 I, 

3 I+ 2 


metaphase J. —G. Paspalum variabile, 18 +21, til A . —H. aspaluum prostratum, ca. 20 II, diakinesis. —l. 
CENE ME 1 10 UI. prometaphase L Arrows show univalents. Scale bar = 20 um 


Volume 93, Number 4 
2006 


Morrone et al. 653 


Chromosome Studies in Panicoideae 


this species. Polyploidy is common in Echinochloa, 
but octoploids had not been previously documented in 
the genus. 

Eriochloa is a cosmopolitan genus with ca. 30 
species present in open habitats of all continents; 
tetraploids, hexaploids, and octoploids have been 
reported in the genus. Eriochloa montevidensis Griseb. 
is a tetraploid with 2n. = 36 (Fig. 
count that agrees with a previous count for the species 


1B), a chromosome 


(Nünez in Parodi, 1946); it should be mentioned that 
there is also a diploid count of 2n = 18 for this 
species (Pohl & Davidse, 1971) 


Within the pantropical genus /sachne, the count of 


a Bolivian J. arundinacea specimen was diploid with 
2n = 20, 
for this species that found pup es with 2n = 40 
(Tateoka, 1962; Davidse & Pohl, 1974). 

Panicum hirticaule J. Presl, a species of subgenus 
Panicum (Zuloaga & Morrone, 1996), is a diploid with 


in agreement with previous results for this 


which differs from other published counts 


2n — 18, 
species (Fairbrothers, 1954; Gould, 1965; Davidse & 
Pohl, 1972b). 

Section Stolonifera Hitche. & Chase ex Pilg. of 
Panicum includes 13 American species found in the 
the edges of forests (Zuloaga, 1987 


1988). This section has been 


pen 


interior or at 


Zuloaga & Sendulsky, 


characterized as having a base chromosome number of 


a 


x = 10, based on studies of P. pulchellum Raddi 
(Gould & Soderstrom, 1970; Pohl & Davidse, 1971; 
Davidse & Pohl, 
count "n P. 


74). We here present the first 


chromosome stoloniferum, based on 


a specimen from Bolivia with a diploid number of 


2n = 20 (Fig. 1C). 

Section Parvifolia Hitche. 
Panicum has a base chromosome number of x = 9 
(Zuloaga, 1987). A total of 31 species, found in open 
and moist places, are recognized from America within 

(Aliscioni et al., 2003). 
trichanthum Nees is a tetraploid with 2n = 


(Fig. 1E). 


this section Panicum 


— 


which confirms previous counts for this 


species (Pohl & Davidse, 1971; Davidse & Pohl. 
1972a, b, 1978; Honfi et al., 1991). 
The genus Paspalum has approximately 310 


species that are primarily American, and most have 
= 10 (Burson, 1975). 
Many of these are polyploid, but most are tetraploid 


a base chromosome number of x 


A total of 25 species were 
14 are 


from previously reported counts. Of these 25 species, 


(Pagliarini et al., 2001). 
studied; of the counts recorded, new or differ 
13 were tetraploids, 10 were diploids, and two were 
Chase (1929, 
Morrone (2005) are followed for the infrageneric 


hexaploids. ined.) and Zuloaga and 


Diploidy (2n = 20), 


tetraploidy (2n Ji 
hexaploidy (2n = 60), and octoploidy (2n = 80) have 


& Chase ex Pilg. of 


been cited for subgenus Ceresia (Pers. Rchb. of 
1948; 1962: Gould € 
Soderstrom, 1967; Davidse & Pohl, 1972a, 1974. 
1978; Pohl, 1980; Killeen, 1990; Hunziker et al., 


1998). Species of this subgenus are distinguished 


Paspalum (Saura, Tateoka, 


— 


y 
a winged rachis of the inflorescence, pilose spikelets, 
and a pale upper anthecium that is usually hyaline or 
membranous, with the tip of the palea not enclosed by 
its lemma. Two diploid counts were made for P. 


buchtienii, the first. counts for this species; 10 


— 


bivalents were observed from microspore mother cells 


(Fig. 2A), and Zn 


metaphase of an anther wall cell. Also, the first count 


— 20 was observed from a mitotic 


was made of P. ceresia, a tetraploid species with 2n — 


40 (Fig. 2B). 


Paspalum  blodgettii Chapm P. caespitosum 
Flüggé, and P. indecorum Mez, species of group 


— 


raespitosa Chase, were diploid, agreeing with pre- 
viously reported counts of 2n = 20 (Davidse & Pohl, 
1972a; Quarín, 1977; Quarín & Burson, 1991; 
Hunziker et al., 1998; Pagliarini et al., 2001), and 
tetraploid eytotypes, agreeing with previously reported 
of 2n = 40 (Banks, 1964). Mexican P. 
blodgettii accessions are tetraploid with 2n = 40, 
Banks (1904). 


Mexican P. caespitosum is a diploid with 2n. = 20, 


E 


counts 


confirming a previous count by 
which agrees with determinations by Davidse and 
Pohl (1972a), and differs from the count reported by 
Banks (1964) that indicated a tetraploid eytotype with 
2n — 40. Paspalum indecorum, as previously reported 
y Quarín and Burson (1983) and Hunziker et al. 
(1998), is a diploid with 2n = 20. 


The group Racemosa Morrone, Zuloaga & Carbonó 


— 
— 


in Paspalum (Morrone et al., 1995) includes annual 
species of moist habitats up to 4600 m in elevation, 
flattened, 


foliaceous and the upper anthecium pale, 


all having the rachis deciduous, and 
cartilagi- 
nous, or crustaceous. Species are also anatomically 
characterized by having distinctive Kranz cells. Our 


P. candidum (Humb. & 


Bonpl. ex Flüggé) Kunth is hexaploid with 2n = 60 


count for the common species 


(Fig. 1H), and this agrees with several counts from 
Costa Rica and Hondures by Pohl and Davidse (1971) 
Pohl (1972a) but differs from 
previous diploid counts (2n. = 20) (Davidse & Pohl, 
= 40) (Hunziker et 


al., 199 “a Paspalum penicillatum is a tetraploid with 


Davidse and 


— 


and 
1974) and tetraploid counts (2n 


2n = . and some meiotic irregularities were 
observed in this 


univalents (Fig. IF). Thos 


for this species. Finally, P. prostratum Scribn. & Merr. 


species, with the presence of 


is the first published count 


has 2n = ca. 40 (Fig. 2H), which agrees with previous 
studies by Davidse and Pohl (1974) and Reeder 
(1968). However, Pohl and Davidse (1971) also cited 


a hexaploid count for a Costa Rican specimen. 


654 


Annals of the 
Missouri Botanical Garden 


The Parviflora Chase group in Paspalum (Chase, 
1929; Zuloaga & Morrone, 2005 


small caespitose annuals with minute and pilose 


consists mostly of 


spikelets. Paspalum  clavuliferum C. right was 
diploid, with 2n 20 and a regular meiosis. This 


count agrees with previous counts by Pohl and Davidse 
(1971). Davidse and Pohl. (1972a). and Reeder (1967). 
The Virgata Chase group in Paspalum (Chase, 1929; 


Zuloaga & Morrone, 2005) includes tall species with 


a dark upper anthecium. Accessions of P. commune 
Lillo from Argentina have 2n = 40, which confirms 


previous counts of this species as a tetraploid (Saura, 
1941. 1048: al.. 1998). 


anomalies resulting from an inversion heterozigosity 


Some meiotic 


Hunziker e 


were observed in this species, with the presence of 
univalents and bridges at anaphase II (Fig. 20). 

Two species were analyzed in the Livida Chase 
group (Chase, 1929; Zuloaga & Morrone, 2005). 


Paspalum denticulatum and P. hartwegianum. Pas- 


palum denticulatum is a widespread American species 
with a high degree of variability in its vegetative and 
led to the 
establishment of several different entities now recog- 
2005). 


cited as 


floral characters, which has taxonomic 


nized under this species (Zuloaga & Morrone, 


Paspalum denticulatum was previously 

a tetraploid with 27 = 40 (Reeder, 1967: Gould. 
1968: Davidse & Pohl, 1972b; Moraes Fernandes el 
al., 1974; Quarín, 1977; Quarín et al., 1982; Quarín & 
Burson, 1991; Hunziker et al., 1998) or a hexaploid 
wilh 2n = 60 (Burson, 1975), but our Mexican 
collection was diploid (a = 10) (Fig. 1G). We found 


Mexican P. hartwegianum to have a count of 2n = 40 
whereas other reports had a hexaploid eytotype with 
2n = 60 (Gould, 1958: Davidse & Pohl, 1972b). 
Paspalum ekmanianum, a member of the Lachnea 
Chase group (Chase, ined.: Zuloaga & Morrone, 2005). 
is an uncommon species characterized by its filiform 
inflorescences with usually a single 


blades and 


raceme and spikelets paired. It grows in Andean 
regions of southe rn. Bolivia and northwestern Argen- 


2200 m in 


Paspalum ekmanianum has 2n = 20 (Fig. 


elevation. 
2E) this is 


tina, in open 


the first reported count for this species. 
1929; 


2004) includes caespitose perennials 


The group Notata Chase of Paspalum (Chase, 
Zuloaga et al., 
with inflorescences usually with two racemes and pilose 

r glabrous paired spikelets. The South American P. 
ellipticum was previously reported from Brazil as an 
1974). 


Brazilian collection was tetraploid, with 2n = 40. 


but 


octoploid (Moraes Fernandes et < 


The group Fimbriata Chase in Paspalum (Chase, 


1929) includes species with fimbriate spikelets, the 


upper glume having wings up to 3 mm long. One 
collection of P. fimbriatum was tetraploid (2n = ca. 


40). whereas other plants have been reported with 2n 


our 


1964: & Soderstrom, 1907: 


1974). Meiotic irregularities, such as 


20 (Banks. Gould 
Davidse & Pohl. 
the presence of 2 to 8 univalents, were observed 
behavior was normal but 


several cells. The meiotic 


showed univalents in some cells, probably due lo 


asynapsis or desynapsis at the end of diakinesis. 
Asynapsis occurs randomly in most individuals and is 
failure form chiasmata, 
the 


accompanied by a pre- 
sumably caused by interaction between the 


genotype and the environment. 


The Decumbentes Chase group in Paspalum 
(Chase, 1929: Zuloaga & Morrone, 2005) is charac- 


terized by the presence of a first glume and axillary 


inflorescences and includes a total of 21 American 


species. Within it, the Argentinian species P. 
glabrinode is a diploid (2n = 20: Fig. 1D), the 
Mexican species P. variabile is a tetraploid (2n = 40; 
Fig. 2G), and the Bolivian species P. inconstans is 


a hexaploid (2n = ca. 60). None of these species have 
been previously studied cytologically. 

The highly polymorphic group. Plicatula Chase in 
1929: Zuloaga & Morrone, 2005) 


includes American species distributed from the United 


Paspalum (Chase, 


States to Argentina that are distinguished primarily by 
the presence of obovoid spikelets and a dark upper 
anthecium. Diploid accessions of P. plicatulum Michx. 
and P. wrightii Hitehe. & Chase, with 2n = 20 
(Fig. 21), were analyzed. Brown (1950), Reeder (1907), 
Moraes Fernandes et al. (1974). Davidse and Pohl 
(1972b), Honfi et al. (1991), and Martínez and Quarin 
(1999, under P. hydrophiliun Henrard) reported the 
same number, al. (2000) pub- 
lished counts for P. plicatulum of 2n = 40. This species 
floral 


— 


whereas Pozzobon et 
has a striking variation in its vegetative and 
characters, a fact that has led to the establishment of 
many entities now synonymized under this species 
(Zuloaga & Morrone, 2005). Several apomictic tetra- 
ploid cytotypes have been described for P. plicatulum 
(Quarín & Norrmann, 1987). Paspalum glaucescens 
Hack. is a tetraploid with 2n = 40, a count that is in 
agreement with the previous counts recorded by Moraes 
Fernandes et al. (1974, under P. 
Pozzobon et al. (2000). and Pritchard (1962, under P. 


yaguaronense Henrard). Quadrivalents, bivalents, and 


yaguaronense), 


univalents were observed as meiotic configurations in 
this species (Fig. 2E). The morphological and eytolog- 
ical complexities of this group indicate a need for 
exlensive and intensive study. 

The group Paniculata Chase in Paspalum (Chase, 
1929; Zuloaga € Morrone, 
arrangement of 10 caespitose perennial species with 
flat 


orbicular spikelets 


2005) includes a complex 


blades. multiflowered inflorescences, and sub- 
Diploid and tetraploid eytotypes 


The 


Bolivian material of P. juergensii Mack. was diploid 


have been published for the Paniculata eroup. 
I : | 


Volume 93, Number 4 
2006 


Morrone et al. 
Chromosome Studies in Panicoideae 


with Zn = 20, confirming the previous counts by 
Moraes Fernandes et al. (1974), Burson and Quarín 
(1982), Dandin and Chennaveeraiah (1983), Pitman et 
al. (1987), and Pozzobon et al. (2000). Two accessions 
of P. squamulatum from Mexico were studied. One 
accession is a tetraploid with 2n = 40, in agreement 
): the 
second is a diploid with 2n = 20, which differs from 


with a previous report by Pohl and Davidse (1971 
| | A 


previous reports for this species. 
The group Dilatata Chase in Paspalum (Chase. 1929: 
grou} ] 
Zuloaga & Morrone, 2005) 
oO 


forage species with multiflowered inflorescences and 


includes four important 


pilose spikelets that are ciliate along the margins 
Paspalum pauciciliatum has 2n = ca. 38 chromosomes: 
the cells observed showed restitution nuclei (Rosem- 
1927). 


include a diploid cytotype with 2n = 


berg, Previous counts for P. pauciciliatum 
20 (Moraes 
Fernandes et al., 1974) and a tetraploid cytotype with 
2n = 40 (Honfi et al., 1991; 2000). 
Material of Paspalum saccharoides from Bolivia, 
belonging to group Saccharoidea Chase (Chase, ined.: 
Zuloaga & Morrone, 2005), is tetraploid with 2n — 40, 
which differs from the count reported by Pohl and 
Davidse (1971) for a Costa Rican specimen (2n = 20). 


Paspalum saccharoides is an odd species wil 


— 


Pozzobon et al.. 


— 


un the 
genus Paspalum. The plants are tall and scrambling to 
2 m, with long and plumose inflorescences, spikelets 
that are long-ellipsoid and densely pilose, the upper 


glume and lower lemma 2-nerved and hyaline, and 
upper anthecium hyaline and open at the summit. 


Pohl (1971) 


between spikelets of P. 


and Davidse indicated | similarities 
saccharoides and those of 
the genus Digitaria Haller. Nevertheless, the genus 
Digitaria has a base chromosome number of x — 9, 
The genus Urochloa has approximately 110 species 
present in tropical and subtropical regions of the New 
and Old World; several species have been extensively 
cultivated in tropical regions. Common basic chromo- 
some numbers in the genus are x = 7, 8, an 
Urochloa fusca (Sw.) B. F. 
2n — 18, a count that is in agreement with previously 
reported counts by Gould (1958, 1966, 1968), Gould 
and Soderstrom (1970), Pohl and Davidse (1971), and 


Reeder (1971) under Panicum fasciculatum with 2n — 


Hansen & Wunderlin has 


18. Urochloa lorentziana is a diploid with 2n = 36 
(Fig. 2D). but previous counts indicated an arrange- 
ment of 2n = 54 (Morrone & Zuloaga, 1992). 


Literature Cited 


Aliscioni, S. S., L. M. Giussani, F. 
2003. A (Poaceae: 
Paniceae). Tests of monophyly and pny 1 5 netic 985 ement 
within the Panicoideae. J. Bot. 90: 796-821 

Banks, D. J. 1964. Chromosome counts for Pasndluri: 
Rhodora 66: 368-370. 


O. Zuloaga & E. A. Kellogg. 


1 phylogeny of Panicum 


mer. 


Black, G. A. 1963 
1 ae 
1-18 

lsh Z. V. Grif, T. Matvejeva & O. Zakharyeva. 


.C€ 1 Numbers of Flowering Plants. V. L. 


Grasses of the 
treatment). 


genus Axonopus (a 


Advancing Frontiers Pl. Sci. 


1 Botanical Institute, Russian Academy of 
Sciences, St. ES ALersburg. 
Bowden. W. M. & H. A. Senn. 1962. Chromosome numbers 


in 28 grass genera from South America. Canad. J. Bot. 40: 
1115-1124. 

Bradley, M. V. 1948. A method for making aceto-carmine 
squashes eee b removal of the cover slip. 


Stain Technol. 41-44 
Brown. dd \ ad study in the Gramineae. 
\mer. J 


: n 35: 382-306 
. 1950. A cytologic al study of some Texas Gramineae. 
1 5 Bot. Club 77: 63-76 
Mu B. L. 1975. € ohn of some apomictic Paspalum 
species, Crop Sci. (Madison) 15: e 

Quarín. 1982. Cytology of a 
virgatum and Neg mea with P erica 
l. J. Gene! 2]9-: 
A. 1929. The North ere 'an species s Pipas: 
. Herb. 23: 1-310. 
Revision of the South American 
Hitchcock and Chase Library, 
Department, Smithsonian Institution. " ert 
Connor, H. E. 1979 


& C. 


Jurgensit. C . Cytol. 219-22 


t 


Ined. species of 


genus op aspalum. Botany 
C. 
Breeding systems in the grasses: A 
survey Ne ew 5 Bot. 17: 13 uer as 
Dandin, S. B. & M. S. Chennaveeraiah. 


number and meiotic 


. Chromosome 


behaviour in o of basic 
c ener number 
Genet. 18: 26-33. 

& R. W. Pohl. 1972a. 


meiotic behavior and notes on some tro] 


in the genus Paspalum. J. Cyto 


E G. Chromosome numbers, 


cal grasses from 
Canad. J. Bot. 


Sed 1 rica and the West Indies. 


1441— 


2 — ———,. 1972b. Chromosome numbers, meiotic 
behavior and notes on some grasses from 


and the West 


Central America 
Indies. Canad. J. Bot. 50: 273-283. 
8 1974. pee eo numbers, meiotic 
behavior and notes on some n al — grasses 
Um j : 


(Gramineae). Canad. J. Bot. 371-32 
& 


1978. Caromosome 5 ers of tropical 

American grasses (Cenie ae). Ann. Missouri Bot. Gard. 
55: 637-040, 

ino & B. K. Simon. 1986. Chromosome 


) and an analys 
of poop in the grass flora of Zimbabwe. S. Afr 

Bot. 52: 521—528. 

DeLisle, D. G. 


counts ji Zimbabwe an grasses (Poaceae) 


= 


rican J. 


962. Taxonomy and B ls of the genus 
S 


Cenchrus. 1555 State Coll. J. Sci. 37: 259-351. 
Dubcovsky, J. & F. 0 o 1991 [1992]. Números 

cromosómicos de especies sudamericanas de Panicum 

(Poaceae: Paniceae). Bol. Soc. Argent. Bot. 27 1-2 


Fairbrothers, D. E. 1954. Relationships in the Ca 
Group of Panicum. Ph.D. Dissertation, Graduate School of 
Cornell University, 2 

plant chromosome 


1 numbers, 
5-1978. Menger Syst. Bot. Bot. Gard. 5 


Missouri 


rs, 1979— 
-421. 
82. 
1983. Mong Syst. Bot. : 1-224. 
)88. Index to plan: chromosome s 1984— 


1985. Monogr. Syst. Bot. Gard. 23: 1-264. 


. 1984. Index to den | chromosome ia qr 
un a 30 
. 1985. 


— 
e 
A 
ET 
— 
om 
zi 
5 
> 
> 
ae 
3 
E 
ME 
a eni 


Missouri Bot. 


656 


Annals of the 
Missouri Botanical Garden 


. Johnson. 1990. Index to plant chromosome 


us 1987. Missouri Bot. 
—243 


& 
mua Monogr. Syst. Bot. 
ard. 

1991. bh 


1 T 
1988-1989, Monogr. 


Lo plant chromosome 


Syst. Missouri. Bot. 


— 


sue 
ard.“ 


994. Index to plant chromosome 


bens imo pt. Monogr. Syst. Bot. Missouri Bot. 
ard. 267. 

S —————. 1996. Index to plant chromosome 

numbers, 1992-1993, Monogr. Syst. Bot. Missouri Bot. 


Gar d. 58: 1-276. 


———— & ———, 1998, Index to plant chromosome 


numbers, 1994-1995. Monogr. Syst. Bot. Missouri Bot. 
Gard. 69: 1-208. 

&& ———. 2000. Index to plant chromosome 
numbers, 1996-1997 Monogr. Syst. Bot. Missouri Bot. 
Gard. 81: 1-188. 

Gould, F 5 1958. Chromosome numbers in southwestern 


Amer. J. 


grasses Bot. 45: 757-161. 
1965. Chromosome. numbers in some Mexican 
grasses. Bol. Soc. Bot. México 29: 40-62 
1966. 1 numbers of some Mexican 
grasses. : anad. J. Bot. 44: 168: 
98. Chromosome numbers of Texas grasses. 
Dim : d 16: 1315-1325. 


— & . Clark. 1978. Dichanthelium (Poaceae) in 
the ind Sue and Canada. Ann. Missouri Bot. Gard. 
65: 1088- 

== e TEM 1967. Chromosome numbers of 
tropic al rom rican grasses. Amer. J. Bot. 54: 670-683. 
& —— 19 aa 1 of some 


p cu "us um grasses. Canad. J. Bot. 48: 


1633 . 
Honfi, A P L Q. L. Quarín & J. M. Valls. 


Estudios cariológicos en gramíneas sudamericanas. 


1991 |1990]. 

Dar- 
winiana i ui 

Hunziker, J. 


1998. E Wo cromosómicos en Par 


). Zuloaga, O. Morrone & A. Escobar. 

niceae sudamericanas 

(Poaceae: 1 oideae). Darwiniana 35 30 

Killeen, T. J. 1990. The grasses of C hiquitanía, S Santa Cruz, 
Bolivia. Ann. 1 Bot. Gard. 77: 125-201. 

Martínez, E. J. & C. L. Quarín. 1999, 
comportamiento reproductivo de un citotipo diploide de 
Paspalum 5 y sus híbridos con p. palustre 
Poa 


C on y 


(Poaceae uceae). Darwiniana 37: 243-251 

Moore, R. J. 197 0. a "x oe plant chromosome numbers for 
1968. es Veg. 68: 1-115. 

—— ndex to " chromosome numbers for 1969, 


Em 5 77: 1-11 

72. NR x Lo m chromosome numbers for 1970. 
a Veg. 84: 1-1: 
. 1973. Index to 1 11 chromosome numbers for 
1967-1971. Regnum Veg. 90: 1-539. 


974. n ex to eg chromosome numbers for 1972. 


: 1-108 


77. de lo dani 


ias Veg. 


chromosome numbers for 
=25 


197 3- 1988 = 96: 
Moraes ^ 1 77 M. „. J. à arreto, F. M. Salzano & / 
M. O. F. Sacchet. Ds tology and 8 


re TOR NE in Brazilian 19 5 Ñ Paspalum (Gramineae). 
Caryologia 27: 455—405 

Morrone, O. & F. 0. 88 1992. Revisión de las especies 
sudamericanas los 
Brachiaria y Urochloa (Poaceae: Panicoideae: 


nativas e introducidas de gé neros 


— 
— 


aniceae). 


Darwiniana 31: 43-109 


; Y 1995, Revisión del 
Racemosa del género Paspalum (Poaceae: Panicoideae: 
„ ee Bot. Gard. 2-116 

Norrma . Quarm € T. J. 1994. 

Chromosome 1 05 in 1 grasses. (Gramineae). 
i Bot. 


p Investigaciones 


— 


. Carbonó. grupo 


Ann. 
Killeen. 


Gard. 8 1768-114 


cariosistemáticas en las 
Revista Fac. 


2 
Si y 
* 
= 


gramíneas argentinas de la à 55 Paniceae. 
La Plata 28: 220-256 
1968. An acetic-haematoxilin squash method for 
aall chromosomes. 11 119 
out R. 1967. Index to plant chromosome m s for 
5 129. 


Agron. Univ. Nac. 


Caryologia 21: 


b Veg. 50: 1-12 
3. Index to plant chromosome numbers for 1966. 
3 n 55: 1-126. 
1969. les to plant chromosome numbers for 1967. 
Regnum Veg. 59: B 20. 
M. S., L. R. Carraro, 


Pagliarini P. M. de Freitas, E. V. 
Adamowski. L. A. Batista & 5 ps M. Valls. 2001. 


C yloge netic 1 le 1 r Homilia Paspalum acces- 

. Hereditas (Lund) 1 —3* 

Parodi, L. R. 1946 
determinación de los géneros y 
especies, 4th ed. Acme Agency, 

Pitman, M. W., B. L. Burson & E. 1987. 


Phylogenetic relationships among Paspalum species with 
48: 


Gramíneas Penales . Clave para la 


enumeración de 


Bas E aw. 


— 


1 arent base chromosome numbe Bot. Gaz 


1 
n T W. 
008 


980. Flora Costaricensis. Fieldiana, Bot.. 


— ——— & G. Davidse. 1971. € Ade numbers of Costa 


Rican grasses. Brittonia 23: 293— 


Pozzobon, M. T., J. F. M. Valls & S. dos Santos. 2000. 
Contagens cromossômicas em espécies brasileiras 
(Gramineae). Acta Bot. Brasil 14: 


de Paspalum L. 
151-162 
Pritehard, A. J. 1962. 
P. yaguaronense (Henr.). 
206-211. 
Quarín, C. 
de Argentina viii Hie 
— — — & B. L. Burson. 1983. 
Paspalum notatum, p pumilum, P. indecorum and 1 
vaginatum. Bot. 144: 433438 
& ——. Ero 
Paspalum. Cytologia 56: 223— 


The cytology and reproduction. of 
I 


Austral. J. Agric. Res. 13: 


. 1977. Recuentos cromosómicos en gramíneas 
enla 1: 73-78. 


Cytogene tic relations inp. 


C nim of sexual and apomictic 
& G. A. Norrmann. le 587. Relaciones. entre el 
número cromosómico, su comportamiento en la meiosis y 
productivo del género Paspalum. Anales del 
Botánica, 


els na re 
V pou Latinoamericano de 
34 


Bogotá, 


Colombia 3: 25 
W. V 


Fernández. Genetic 


. Hanna & A. 1982. 
diploid and tetraploid Paspalum species. 


aiea in 

Heredity 73: 254—250. 

The basis of angiosperm phylogeny: 
294-764 


Ann. 


Cytology. Missouri Bot. Gard. 62: 724—76 
Reeder: J. R. 1967. Notes on Mexican grasses VI. Mis- 
cellaneous chromosome numbers. Bull. Torrey Bot. Club 


94: A 
Notes on Mexican grasses VIII. Miscellaneous 
Bull Bot. Club 95 


c me numbers-2. Bu Torrey 
30. 

— ——, 1971. Notes on Mexican grasses. IX. 

Brittonia 23: 105-117. 


a in uninucleate 


Miscellaneous 
c ee ne "s. T 

927. Homeotypic 
ar Hereditas (Lund) 9: 


Rosemberg. 19 


eo 


pollen mother 


Volume 93, Number 4 
2006 


Morrone et al. 
Chromosome Studies in Panicoideae 


657 


Sáez, F. A. 1960. El empleo de la hematoxilina acética o 
propiónica para el estudio de los cromosomas con la 
técnica de aplastamiento. Comun. Soc. Biol. 
(mimeografiado). 

Saura, F. 194 


Montevideo 


. Cariología de algunas especie s del género 
Rev. Fac 


S Cariologia de gramíneas en 


a Agron. Veterin. 2: 41—48. 

———. Argentina. 
Rev dini Veterin. 12: 51-67. 

Shibata. K s aude citológicos de plantas colombianas 

stres y cultivadas ls ytological studies on some wild 

Re 1 plants of (Rabie): J. Agric. Sci. Tokyo 
Nogyo s Suppl. 8: 49-62 

Tateoka, 52 


ck 


A logical ee : some Mexican 
ut Torrey Bot. Club € 
Systematics S New World species of 
Panicum 1 5 Pp. å R 
Soderstrom, K. W. Hilu, C. S. 

Barkworth (editors). 


N 
c 

= 1 
e 

S 

Q 

— 

rr 

a 


ed Systematics and 
Smithsonian Institution Press, Washington, D.C. 


————— & O. Morrone. 1996. Revisión de 55 especies 


americanas de Panicum subgénero Panicum | sección 
Panicum (Panicoideae: Paniceae). Ann. Missouri Bot. 
Gard. . 200—280. 

2005. Revisión de las especies de 


niis para América del Sur Austral (Argentina, 
Bolivia, sur de Brasil, Chile, Paraguay y Uruguay). 
Monogr. Syst. Bot. Missouri Bot. Gard. 102: 1-297 

& T. Sendulsky. 1988. A 


revision of Panicum 


subgenus Phanopyrum section wd ls (Poaceae: 
Paniceae). Ann. Missouri Bot. Gard. 420—455. 


, O. Morrone & Ellis. 199 " revision of 
Panicum subg. Dichanthaliüm Sect. Dial 1 
ceae: Panicoideae: Paniceae) in Mesoameric Ve 
Indies, and South America. Ann. Missouri mn uh a 
119-190. 


Pensiero & O. Morrone. 


2004. Systematics of 
E Paspalum Group Notata (Poaceae-Panicoideae-Paniceae). 
. Bot. Monogr. 71: 1-75. 


ANNALS OF THE MISSOURI BOTANICAL GARDEN: CHECKLIST FOR AUTHORS 


The Annals publishes original articles in systematic 
botany and related fields. Papers whose purpose is the 


establishment of new nomenclatural entities in vascular 


plants and bryophytes are not accepted. Rather, they 
should be submitted to Novon for consideration. (Novon 
instructions are available on the Web or from the 


managing editor.) 
CONDITIONS FOR PUBLICATION 


Authors are urged to contact the managing editor prior 
to submitting manuscripts. Because the electronic file 
will be used by the printer for typesetting. it is important 
to adhere to the items listed in the format section of the 
checklist. i 
editors and sent to 
file 
the first author's 


Authors’ electronic files are prepared by the 
the 
by 


printer for formatting. Every 


electronic the author should be clearly 


labeled with last name, the first four 


letters of the taxon involved, and an indication of the 


files contents (i.e. text, figure, table, ete.). Manuscripts 


that 


subject to the full normal review process initiated here. 


have been reviewed before submission will be 


Manuscripts not properly prepared may be returned for 
revision prior to review. It is the author's responsibility to 
supply missing bibliographie information; the editor does 


not perform library research except by ns arrangement 


with authors not having access to crucial reference 


materials. Communication with the Ew at all points of 


the publication process 1s encouraged. 


Pace CHARGES 


Page costs are $80 per page, although charges may 


sometimes be reduced or waived. Changes in proof made 


by authors will be billed non-negotiably to the author at 
$3.5 


50 per line changed. 


CORRESPONDENCE 
Submissions should be addressed to: 


Beth Parada 
Managing Editor 
Annals of the Missouri Botanical Garden 
Missouri Botanical Garden Press 
P.O. Box 299, St. Louis; MO 63166-0299 U.S.A. 
Managing Editor 
Annals of the Missouri Botanical Garden 
(Street address) 
Missouri Botanical Garden Press 
Monsanto Center 
4500 Shaw Blvd. 
St. Louis, MO 63110-2291 U.S.A 
Tel: (314) 577-9473 ext. 6519 
Fax: (314) 577-9594. 
Email: annals@mobot.org 
l. General 


Text is in English or Spanish on numbered pages. 
8 | pag 


| Manuscript is submitted as Microsoft Word [prefera- 
ble] f 


annalsOmobot.org 


file via email if possible or by ftp [contact 


for alternate 


for 1 0 


medie 
file 


nante, 


electronic 


Electronic of manuscript is labeled. with first 


author's last first four letters of the taxon 
Mw d. m an epi ‘ation of what the file pas 
lis Cipu text.doc, Celis Cipu Figure tif. Celis 


11 9 el 


al 
the 


may 


Electronic scans of figures are included with 


submission, with the understanding that. these 


have to be revised based « on review. Low-resolution 


JPEG images are suitable for the y process, bu 


are not acceptable for final x submission to the printer. 
See the 


Go to httpAwww.mbepress.org/: click on 


printers electronic illustration requirements. 
the Publica- 
tion Submissions at upper right of opening page: click 


Allen 


according 


on Guide to Digital Art, 


Authorities, abbreviated to Brummit & 
Powells Authors of Plant Names, are provided for 
species’ names the first time they are mentioned 
the text; OR they are provided in a table that includes 
names of species. 

letter 


full mailing addres 


Email or is enclosed. containing the author's 


. tele 1 9 number, fax number, 


and e-mail address (require 
2. Format of Text 

Right-hand margin is not justified, and words are not 

hyphenated there. 

One space is inserted after periods, colons, question 

marks, and all other punctuation marks. 

First sentences of paragraphs are indented with a tab. 


Two hyphens with no space between them are used to 


indicate ranges of numbers, for example, page 
numbers, 20-23: measurements, 5-7 mm; altitudes, 
1500-2500 m. 


A single hyphen is used hyphenated words, for 
example, 9-nerved, oval-triangular. 

phrases are not italicized 
etc.). 
rank 


Common Latin words 
1. E., Sensu 
Only taxon names at the genus and below 
are italicized. 

Correct accents, umlauts, and other diacritical marks 
have been included. 

All figures and tables are cited in the text and are 


numbered in the order in which they are to appear. 


3. Style 
Recent issue of the Annals is used as a model. 
Chicago Manual of Style, latest edition, is used 


a reference. 


A. First Page 


Footnotes are typed as double-spaced paragraphs 
on the first page. The first footnote contains acknowl- 


edements, including information on granting agencies. 


Volume 93, Number 4 
2006 


659 


herbaria that loaned specimens, and the name of the 
artist. The second footnote is the author's address. 
Addresses for additional authors are each separate 
bles, 


footnotes. No other footnotes occur. except in ta 
where unavoidable 


5. Abstract & Key Words 


A one-paragraph abstract precedes the text. 


apers in 
Spanish have an English abstract in addition to 


a Spanish resumen. 


E 


The abstract is concise (1 paragraph p includes 


— 


brief statements about the papers intent, materials 
and methods, results, and significance of indi. 

A brief list of key 
Abstract. 


O 


words immediately follows the 


6. Taxonomic Treatment 


Species entries are organized as follows: Heading. 
vernacular name(s), Latin diagnosis (if necessary). 
description, distribution, summary, discussion, speci- 
mens examined. The discussions are parallel and follow 
the same order, e.g., diagnostic characteristics. distinc- 
tion from similar distribution. and 


species variation, 


ecology. nomenclature and typification, 


T 
2 


LJ One paragraph per basionym is used as follows: Taxon 
literature citation, hs cité Beilsch- 

(Nees) Sa. Nishida, dnm. Missouri 
: 680. 1999, hifi latifolia Nees, 

pos 1836. D Peru. 
1835 (fl). 19 5 
BMI. EI. K!, LE not seen. 


author. ition, E. g. 
miedia e a 
ot. Gard. 

Syst. 


indicated. 


Laur.: Locality 


1433 (holotype, El: 
OXF 


isolypes. not seen). 


Synonyms based on different types are placed in 
separate paragraphs, each beginning with the basio- 
nym, followed by other combinations (if appropriate). 
and citation of the type. 

LJ. A brief Latin diagnosis for each new taxon is provided 
rather than a complete Latin description. 

For species with infraspecific taxa: Desc OSA and 


discussion are composite (incorporating all inf apo 
cific 


scriptions. 


and other 


Descr 


parallel with one 1 (in the same species). 


laxa) n with species de- 


tions of infraspecific taxa are 


All synonyms are listed under the appropriate 
infraspecific taxon. 

Descriptions: Descriptions are parallel, within a 
given rank. All measurements are metric. Hyphens 
are used for parenthetical extremes: “peduncle 


(8.2—)14.3—28.0(-3 7 
values are not expected: ovary with (2)4(6) locules. 
Length X width are given in the 


lamina 36.1—82.8 X 9,1—16.8 cm 


When relevant. nomina nuda, misapplied names, and 


31.9) em long,” unless intermediate 


following manner: 


[] 


rfluous names are included 


— 


in the discussion 


following the description, but are not part. of the 


for mal synonymy. 


7. Citation of Types 
LJ. Exclamation points are used for specimens e idee 
and types not seen are indicated as such (e. 
US not seen). 


| Lectotype designations are included together with an 
indication of where they were designated [author, vear. 
page number, and herbarium of deposition; e.g.. C.. 


W. Schiede 159 (lectotype, designated. by even 
(2000: 256), Pl: isotypes....)]. This reference is listed 
in the Literature Cited. If the author of the paper 


submitted is making the lectotypification, the phrase 
"designated here” is use 


8. Tables 


neat, double-spaced. 


Tables are and clearly pre- 


sented. In most cases the printer will typeset these 
from hard copy. 


Captions are typed double-spaced as paragraphs al 
the tops of the tables. 


Fach table starts on a separate sheet. 


9. Abbreviations 


Periods are used after all abbreviations (which are 


minimized) e except metric measures, compass direc- 
tions. and herbarium designations. 


| When 


ormation, 


dates are given as of collection 


Ja in- 
nonth abbreviations are used, 


=> 


three-letter 
except for months with four letters, which are spelled 
out in full. 


States are not abbreviated, and cities are spelled out. 
le 


(St., as in St. Louis, is acceptab 
|] Periodicals are abbreviated according to B-P-H 
(Botanico-Periodicum-Huntianum) and to B-P-H/S 


(Botanico-Periodicum-Huntianum/Supplementum). 


a 


Authors’ names are abbreviated according to Brummit 
& Powell's Authors of Plant Names. 

Book titles 
Literature, edition 2, but with initial letters capitalized. 
Cited. 

If an item does not appear in B-P-H or TI-2, or if 
references are not available, its title is fully spelled out. 


are abbreviated according to Taxonomic 


Book titles are spelled out in the Literature 


these 


Herbaria are abbreviated according to the most recent 


edition of /ndex Herbariorum. 


Abbreviate : forms are not used for references i in the 


lext, exc Rn when citing the names of plants. If it is 


o cite a particular page in the text, the 
1 1 pe (1998: 12) is used. 


10. Specimens Examined 

[L] If many specimens were examined, those cited in the 
text are limited to ca. 12 manuscript pages. 

L] If there are a large number, an index to specimens 
examined is placed at the end of the paper, 
the Literature Cit 
collector, 


following, 
is arranged alphabetically by 


followed by collection number, followed by 


the number of the taxon in the text. Names (including 
e of first and second com ‘tor are provided, 
“et al.” if three or more 


- 


dal are cited in the text as follows: Additional 
specimens examined (or selected specimens examined). 
MEXICO xaca: Sierra San Pedro N 
12°37'N, 85 145 W. 950-1100 m, 3 

Jergensen 865 (BM. G, K. US). 


status are optional but are omitted from longer lists.) 


il 


Nolesco, Talea, 
Feb. 1987 (Il), 


[Dates and reproductive 


Annals of the 
Missouri Botanical Garden 


C dud s are run together in the same Ms ^ e.g. 
COUNTRY A. Major political division: ... COUN- 
TRY B. Major political division: 1 
paragraphs are used for major continental regions 


within major political divisions. 


Specimen Vouchers and Genetic Sequences 


If the paper presents original data, associated her- 


bariumvouchers are cited. [Vouchers for seed and/or 


other collections should be included where pertinent. 
Dependent on the paper, reference to the original wild 
source may be required.] Vouchers are also cited from 
common names, and uses are taken from specimen 
labels 

Herbarium vouchers state the collector and number, 
herbarium in which the voucher is located, and a clear 
annotation that the material represents the voucher 
for the study in questi 

Nucleic acid or protein sequences corresponding to 
equal or greater than 50 nucleotides are entered into 
an spo data bank, e.g., GenBank/EMBL. The 

1 


accession numbers are ied before publication. 


[Long sequences (exceeding two pages) will not be 
routinely published. } 
Author accepts responsibility for establishing. the 
accuracy of information provided. 

12. Keys 
Keys are clear and have been checked carefully for 
consistency with the descriptions. Leads of each 
couplet are parallel. 
Dichotomous keys are indented. 
Infraspecific taxa are keyed separately, not in species 
ceys. 


13. Literature Cited 


The Literature Cited contains full citations of a 
references cited in the text 

All entries in the Literature Cited are cited in the text. 
Spelling of author(s) name(s) and vears of publication 
have been double-checked. 

All entries have been verified against original sources, 
especially journal titles, accents, diacritical marks. 
and spelling in languages other than English 
Periodicals are listed as follows: author's last name, 
initial(s). Year. Full title of article. Journal abbrevi- 
aled as in B-P-H/S. Volume: pages. No parenthetical 
part numbers after volume numbers are given unless 
each part is paginated separately. 

For more than one author, this style is followed: 
author's last name, initial(s), second author's initi- 


al(s). last name & third author's initial(s), last name. 


— 


Books appear as follows: author's last name, initial(s). 
Year. Full Unabbreviated Title (edited by Editor). 3rd 
ed., Vol. 2. Pub 
For an article within a larger work, this style is 
followed: Author(s). Year. Name of the article. 
Pp. 00-00 in Name of the editor(s). Full Title of 
Larger Work. Publisher, City of Publication. 


— 


isher, City of Publication. 


Citations of work “in prep." unpublished theses and 
dissertations, and similar references to inaccessible 
sources have been eliminated or kept to a minimum. 
They are not necessarily included in the Literature 


Cited. 


J. Illustrations 
Electronic figures are labeled with the first author's last 
name, res first four letters of the taxon, and . .. Fig- 
urel.tf^, .Figure2.üf7, ete. The file extension 
clearly ndeas what type T file it is. 
Seale bars appear on illustrations, photographs, and 


maps. 


| Magnifications/reductions are not indicated in cap- 


tions. 

Electronic scans may be pre-sized to fit either column 
width (2-5/8 in. or ca. 68 mm) or full-page width 
(5-1/2 in. or ca. 140 mm). [Maximum size for printed 
illustrations is 5-1/2 X 8-1/4 in.] 


Figures are numbered in Arabic numerals in the order 


of their citation in the text. Parts of figures are labeled 
with capital letters 

Photographs are prepared according to the printers 
requirements. [See Guide to Digital Art, Allen Press 
through the Publication Submissions link on the 
opening page of http://www.mbgpress.org/. | 

Figures are grouped into composite plates when 


»ossible: edges of photographs are abutted. 
I 8 | gra] 


No stripping is inserted. between plate or figure 
segments. [Printer will insert stripping.] 

Edges of figures are squared. 

Maps include reference to latitude and longitude and 
are bounded by a fine border. 

Scanning electron micrographs are free of conspicu- 
ous charging. 

Axes on graphs are all labeled. 

Captions provide all explanatory text. Captions are 
separate from other text, one paragraph for each group 
of figures, and following the style in current issues of 
the Annals. 

Symbols on maps are legible, and reduction has been 


lake "n into e onside "ration. 


ANNALS OF THE 
MISSOURI BOTANICAL GARDEN 


VOLUME 93 
2006 


Colophon 


This volume of the ANNALS of the Missouri Botanical Garden has been set in APS Bodoni. The 
text is set in 9 point type while the figure legends and literature cited sections are set in 8 point 
type. 

This volume has been printed on 70% Sterling Ultra Gloss. This is an acid-free paper designed to 
have a shelf-life of over 100 years. 

Photographs used in the Áxxals are reproduced using 300 line screen halftones. The binding used 
in the production of the ANNALS is a proprietary method known as Permanent Binding. 
Lawrence, Kansas 00044, U.S.A. 


The Anais is printed and distributed by Allen Press, Ine. of 
© Missouri Botanical Garden 2006 


ISSN 0020-6493 


ACHILLE, FRÉDÉRIC, TimotHy J. Morley, Porter P. Lowry II & JoéL Jeremie. 
Polyphyly in Guettarda L. (Rubiaceae, Guettardeae) based on nrDNA ITS 
sequence data 


Acosta, M. CRISTINA, ADRIANA DEL V. ORDÓÑEZ, ANDREA A. Cocucci & EDUARDO A. 
Moscone. Chromosome Reports in South American Nicotianeae (Solana- 
ceae), with Particular Reference to Nierembergia — —. 

ALCARAZ, FRANCISCO (see Cristina Inocencio, Diego Rivera, M* Concepción 
)bón. Francisco Alcaraz € Jose-Antonio Barreña) 

AL-SHEHBAZ, [SAN A. (see Ji-Pei Yue, Hang Sun, Ihsan A. Al-Shehbaz & Jian- 
Hua Li) 


BALDWIN, Bruce G. Contrasting Patterns and Processes of Evolutionary Change 


in the Tarweed-Silversword Lineage: Revisiting Clausen, Keck, and 
Hiesey's ad 


BARKER, Micha, S. & R. James Hickey. A Taxonomic Revision of Caribbean 
Adiantopsis 1 

BARRENA, JOSE-ANTONIO (see Cristina Inocencio, Diego Rivera, M* Concepción 
Obón, Francisco Alcaraz & Jose-Antonio Barreña) = < 

Baskin, CAROL C. (see Patrick J. Lawless, Jerry M. Baskin & Carol C. Baskin) 

Baskin, Jerry M. (see Patrick J. Lawless, Jerry M. Baskin & Carol C. Baskin) 

BERNAL, RopRIGO (see Marcela Mora, Rodrigo Bernal, Thomas Croat & Jorge 
Jácome) 


Bruyns, P. V. & C. KLak. A Systematic Study of the Old World Genus Fockea 
(Apocynaceae—Asclepiadoideae) 


BUCKLEY, GREGORY A. (see David W. Krause, Patrick M. O'Connor, Kristina 
Curry Rogers, Scott D. Sampson, Gregory A. Buckley & Raymond R. 
Rogers) 

CHAUTEMS, ALAIN (see Mathieu Perret, Alain Chautems & Rodolphe $ Spichiger) 

CIALDELLA, ANA María, OsvaLDO Morrone & FERNANDO O. ZULOAGA. Revisión de 
las species de Axonopus (Poaceae, Panicoideae, Paniceae), Serie Suffulti 

CLEEF, ANTOINE M. (see Henry Hooghiemstra, Vincent M. Wijninga & Antoine 

Cleef) 

Cocucci, ANDREA A. (see M. Cristina Acosta, Adriana del V. Ordóñez. Andrea A. 

Cocucci & Eduardo A. Moscone) 


Conti, ELENA (see André O. Simões, Mary E. Endress, Timotheüs van der Niet, 
Luiza S. Kinoshita € Elena Conti) 
Croat, Thomas (see Marcela Mora, Rodrigo Bernal, Thomas Croat & Jorge 
Jácome) 


CULLEY, THERESA M. (see Ann K. Sakai, Stephen G. Weller, Warren L. Wagner, 
Molly Nepokroeff € Theresa M. Culley 

Doses, CHRISTOPH, Marcus Koch & Timothy F. SHARBEL. Embryology, Karyology, 
and Modes of Reproduction in the North American Genus Boechera 
(Brassicaceae): A Compilation of Seven Decades of Research 


Volume 93 
2006 


103 
034 
122 


402 


664 Annals of the 
Missouri Botanical Garden 


Express, Mary E. (see André O. Simões, Mary E. Endress, Timotheüs van der 
Niet, Luiza S. Kinoshita € Elena Conti) 

ESCOBAR, ALEJANDRO (see Osvaldo Morrone, Alejandro Escobar & Fernando 0. 
Zuloaga 

Frerras, LEANDRO & Marites Sazima. Pollination Biology in a Tropical High- 
altitude Grassland in Brazil: Interactions at the Community Level 

Gancta-JAcAs, NÚRIA (see Alfonso Susanna, Núria Garcia-Jacas, Oriane Hidalgo, 
Roser Vilatersana & Teresa Garnatje) 

GARNATIE, TERESA (see Alfonso Susanna, Núria Garcia-Jacas, Oriane Hidalgo, 
Roser Vilatersana & Teresa Garnatje) 

Gron, RICHARD E. (see Jonathan B. Losos, Richard E. Glor, Jason J. Kolbe & 
Kirsten Nicholson) 

GRAHAM, ALAN. Latin American Biogeography—Causes and Effects, the 51st 
Annual Systematics Symposium of the Missouri Botanical Garden. 
Introduction 

GRAHAM, ALAN. Paleobotanical Evidence and Molecular Data in Reconstructing 
the Historical Phytogeography of Rhizophoraceae 

GRAHAM, ALAN. Modern Processes and Historical Factors in the Origin of the 
African Element in Latin America 

Grant, Perer R. & B. Rosemary GRANT. Species Before Speciation is Complete 

Grant, B. Rosemary (see Peter R. Grant & B. Rosemary Grant) 

Hepces, S. BLAIR. Paleogeography of the Antilles and Origin of West Indian 
Terrestrial Vertebrates 

Hickey, R. James. (See Michael S. Barker & R. James Hickey) 

HibaLGO, ORIANE (see Alfonso Susanna, Núria Garcia-Jacas, Oriane Hidalgo, 
Roser Vilatersana & Teresa Garnatje) 

HoocureMsrRA, Henry, Vincent M. Wininca & Antoine M. Creer. The 
Palaeobotanical Record of Colombia: Implications for Biogeography and 
Biodiversity 000000 0 s 

INOCENCIO, CRISTINA, Dieco Rivera, M* CONCEPCIÓN OBON, FRANCISCO ALCARAZ & 
Josk-ANTONIO. BARREÑA. A Systematic Revision of Capparis Section 
Capparis (Capparaceae) 

Jácome, JoncE (see Marcela Mora, Rodrigo Bernal, Thomas Croat & Jorge 
Jácome) 

JÉRÉMIE, JOËL (see Frédéric Achille, Timothy J. Motley, Porter P. Lowry H € Joël 
Jérémie) 

KivosurrA, Luiza S. (see André O. Simões, Mary E. Endress, Timotheüs van der 
Niet, Luiza S. Kinoshita & Elena Conti) 

Krak, C. (see P. V. Bruyns & C. Klak) 

Koch, Marcus (see Christoph Dobeš, Marcus Koch & Timothy F - Sharbel) 

KorBE, Jason J. (see Jonathan B. Losos, Richard E. Glor, Jason J. Kolbe & 
Kirsten Nicholson) 


297 


122 


359 


Volume 93, Number 4 
2006 


665 


Krause, Davip W. Science with a Social Conscience: Digging for Dinosaurs and 
Helping Children in the Land that Time Forgot 


KRAUSE, DAVID W., Parrick M. O'Connor, Kristina CURRY Rocers, Scorr D. 
Sampson, GREGORY A. BuckLeY & Raymonb R. Rocers. Late Cretaceous 
Terrestrial Vertebrates from Madagascar: Implications for Latin American 
Biogeography - 


LawLEss, Patrick J., Jerry M. Baskin & Caron C. Baskin. Scale-dependent 
Classification of Xeric Limestone Prairies: Annual or Perennial Grass- 
lands? 


Li, Jian-HUA (see Ji-Pei Yue, oe Sun, Ihsan A. Al-Shehbaz & Jian-Hua Li) 


Losos, JONATHAN B., RICHARD E. Gron, Jason J. KOLBE & Kirsten NICHOLSON. 


daptation, Speciation, and Convergence: A Hierarchical Analysis of 


I] $ Joël Jérémie) 
MaciLL, RoBERT E. (see James C. Solomon & Robert E. Magill) 


MARTICORENA, ALICIA. Revisión del Género Acaena (Rosaceae) en Chile 


METCALFE, SARAH E. Late Quaternary Environments of the Northern Deserts and 
Central Transvolcanic Belt of Mexico 

Mora, ManckLA, Roprico BERNAL, Thomas Croar & Jorce JÁCOME. A 
Phytogeographic Analysis of Araceae of Cabo Corrientes (Chocó De- 
partment, Colombia) and Comparable Lowland Tropical American Floras 

Morrone, OsvALDO (see Ana María Cialdella, Osvaldo Morrone & Fernando O. 
Zuloaga) 


Morrone, OSVALDO, ALEJANDRO ESCOBAR & FERNANDO O. ZULOAGA. Chromosome 
Studies in American Panicoideae (Poaceae) 
Moscowk, EDUARDO A. (see M. Cristina Acosta, Adriana del V. Ordóñez, Andrea 
. Cocucci & Eduardo A. Moscone) 
Morey, Timorny J. (see Frédéric Achille, Timothy J. Motley, Porter P. Lowry H 
& Joel Jérémie) 


NEPOKROEFF, MoLLY (see Ann K. Sakai, on G. Weller, Warren L. Wagner, 
Molly Nepokroeff & Theresa M. Culley) 
NICHOLSON, KIRSTEN (see Jonathan B. Losos, Richard E. Glor, Jason J. Kolbe & 
Kirsten Nicholson) 


OBÓN, M* CONCEPCIÓN (see Cristina Inocencio, Diego Rivera, M’ T ion 
Obón, Francisco Alcaraz & Jose-Antonio Barreña) 


O’Connor, Patrick M. (see David W. Krause, Patrick M. O'Connor, Kristina 
Curry Rogers, Scott D. Sampson, Gregory A. Buckley & Raymond R. 
Rogers) 


ORDÓÑEZ, ADRIANA DEL V. (see M. Cristina Acosta, Adriana del V. Ordóñez, 
Andrea A. Cocucci & Eduardo A. Moscone) . 


Pascual, Rosenpo. Evolution and Geography: The Biogeographic History of 


South American Land Mammals |... 


367 


359 


592 


647 


634 


666 Annals of the 
Missouri Botanical Garden 


PERRET, MATHIEU, ALAIN CHAUTEMS & RODOLPHE SPICHIGER. Dispersal-Vicariance 
Analyses in the Tribe Sinningieae (Gesneriaceae) A Clue to Un- 
derstanding Biogeographical History of the Brazilian Atlantic Forest 

PiPERNO, DoLorEs R. Quaternary Environmental History and Agricultural Impact 
on Vegetation in Central America 

RICHARDSON, P. Mick. Species Reconsidered: Consequences for Biodiversity and 
Evolution, the 50th Annual Systematics Symposium of the Missouri 
Botanical Garden. Introduction 

RieseBERG, Lores H. Hybrid Speciation in Wild Sunflowers 

Rivera, Dieco (see Cristina Inocencio, Diego Rivera, M* Concepcion Obón, 
Francisco Alcaraz & Jose-Antonio Barreña) 

ROGERS, KRISTINA CURRY (see David W. Krause, Patrick M. O'Connor, Kristina 
Curry Rogers, Scott D. Sampson, Gregory A. Buckley & Raymond R. 

i 8 | mu : À 
Rogers) 

Rocers, Raymonp R. (see David W. Krause, Patrick M. O'Connor, Kristina 
Curry Rogers, Scott D. Sampson, Gregory A. Buckley & Raymond R. 
Rogers) 

SAKAL ANN K., STEPHEN G. WELLER, WARREN L. WAGNER, MOLLY NEPOKROEFF & 
TueresA M. Corey. Adaptive Radiation and Evolution of Breeding 
Systems in Schiedea (Caryophyllaceae), an Endemic Hawaiian Genus 

Sampson, Scorr D. (see David W. Krause, Patrick M. O'Connor, Kristina Curry 
Rogers, Scott D. Sampson, Gregory A. Buckley & Raymond R. Rogers) 

SAZIMA, MARLIES (see Leandro Freitas & Marlies Sazima) 

SHARBEL, Timotay F. (see Christoph Dobeš, Marcus Koch € Timothy F. Sharbel) 

SIMOES, ANDRE O., Mary E. ENpRess, TIMOTHEUS VAN per Niet, Luiza S. KINOSHITA 
& ELENA Conti. Is Mandevilla (Apocynaceae, Mesechiteae) Monophyletic? 
Evidence from Five Plastid DNA Loci and Morphology 

SOLOMON, James C. & Robert E. Macie. Statistical Summary of Some of the 
Activities in the Missouri Botanical Garden Herbarium, 2005 

SPICHIGER, RODOLPHE- (see Mathieu Perret, Alain Chautems & Rodolphe 
Spichiger) 

Sun, HANG (see Ji-Per Yue, Hang Sun, Ihsan A. Al-Shehbaz € Jian-Hua Li) 

SUSANNA, ALFONSO, NURIA GARCIA-JACAS, Oriane HIDALGO, ROSER VILATERSANA & 
Teresa GARNATIE. The Cardueae (Compositae) revisited: Insights from ITS, 
trnlb.-trnF, and matk nuclear and chloroplast DNA analysis 

VANDER Niet, TiMorTHEts (see André O. Simoes, Mary E. Endress, Timotheüs van 
der Niet, Luiza S. Kinoshita € Elena Conti) 

VILATERSANA, Roser (see Alfonso Susanna, Nuria Garcia-Jacas, Oriane Hidalgo, 
Roser Vilatersana & Teresa Garnatje) e 

WAGNER, WARREN L. (see Ann K. Sakai, Stephen G. Weller, Warren L. Wagner, 
Molly Nepokroeff & Theresa M. Culley) 

Wake, Davip B. Problems with Species: Patterns and Processes of Species 
Formation in Salamanders 


Wess, S. Davin. The Great American Biotic Interchange: Patterns and Processes 


340 


Volume 93, Number 4 
2006 


WELLER, STEPHEN G. (see Ann K. Sakai, Stephen G. Weller, Warren L. Wagner, 
Molly Nepokroeff & Theresa M. Culley) 


WuüuwiNGA, VINCENT M. (see Henry Hooghiemstra, Vincent M. Wijninga & 


Yue, Ji-Per Hang Sun, Insan A. AL-SHemBaz & Jian-Hua Li. Support for an 
Expanded Solms-laubachia (Brassicaceae): Evidence from Sequences of 


Volume 93, Number 4. pp. 535-668 of the NN IIS or THE Missouri BOTANICAL 
GARDEN. was published on December 15. 2006. 


TE 


www.mbgpress.org 


CONTENTS 


A Systematic Study of the Old World Genus Fockea (Apocynaceae-Asclepiadoideae) |. 

P. V. Bruyns & C. Klak 

Is Mandevilla 1 MUR Monoubytetiof Evidence from Five Plastid DNA 

Loci and Morphology André O. Simões, Mary E. Endress, 

Timotheüs van p Niet, Luiza S. Kinoshita & Elena Conti 

Revisión de las Especies de Axonopus (Poaceae, Panicoideae, Paniceae), Serie S Ati — 

Ana María Cialdella, Osvaldo Morrone y Fernando O. Zuloaga 

T A Reports in South American Nicotianeae (Solanaceae), with Particular Refer- 


ence to Nierembergia M. Cristina Acosta, Adriana del V. Ordóñez, 
Andrea A. Cocucci & Eduardo A. Moscone 
Chromosome Studies in American Panicoideae (Poaceae) Osvaldo Morrone, 


Alejandro Escobar & Fernando O. Zuloaga 
Checklist for authors 
Index for Volume 93 


Cover illustration. Capparis sicula subsp. mesopotamica Inocencio, D. Rivera, Obón & Alcaraz, 


drawn by Jose-Antonio Barreña.