Annals 
of the 
Missouri 
Botanical 


Garden 
207 VG 


Volume 96 


umber 


Annals of the 
Missouri Botanical Garden 


Volume 96, Number 1 
April 2009 


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 
Allison M. Brock 
Associate Editor, 

Missouri Botanical Garden 
Tammy Charron 

Editorial Assistant, 
Missouri Botanical Garden 
Cirri Moran 

Press Coordinator, 
Missouri Botanical Garden 


Roy E. Gereau 

Latin Editor, 

Missouri Botanical Garden 
Ihsan A. Al-Shehbaz 
Missouri Botanical Garden 
Gerrit Davidse 

Missouri Botanical Garden 
Peter Goldblatt 

Missouri Botanical Garden 
Gordon McPherson 
Missouri Botanical Garden 
Charlotte Taylor 

Missouri Botanical Garden 
Henk van der Werff 


Missouri Botanical Garden 


For subscription information contact ANNALS 
or THE Missouri BoranicaL GARDEN, 96 Allen Mar- 
keting & Management, P.O. Box 1897, Lawrence, 
KS 66044-8897. Subscription price for 2009 is 
$175 per volume U.S., $185 Canada & Mexico, 
$210 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 Mrissourt BOTANICAL GARDEN, Yo 
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- 
stract/Global Health databases, ingenta, ISI® databases, JSTOR, Research Alert, and Sci Search®. 
The full-text of ANNALS or THE Missouri BOTANICAL GARDEN is available online though BioOne™ (http:// 
www.bioone.org). 
© Missouri Botanical Garden Press 2009 
The mission of the Missouri Botanical Garden is to discover and share knowledge about plants and 
their environment, in order Lo preserve and enrich life. 

© This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 


Volume 96 
Number 1 
2009 


Missouri 
Botanical 


Garden 


THIRD INTERNATIONAL 
RUBIACEAE CONFERENCE: 
INTRODUCTION! 


Petra De Block? Charlotte M. Taylor? and 
Suzy Huysmans* 


The 


flowering plants in terms of th 


family Rubiaceae is the fourth largest family of 
e number of species known, 
with worldwide distribution, but most of its diversity is 
concentrated in the highly threatened and rapidly 
disappearing moist ecosystems of tropical and subtropical 
regions. Rubiaceae are badly in need of study by 
systematists, ecologists, and conservationists at a basic 
level, and their important role in these tropical 
ecosystems together with the active threat to the existence 
of so many species adds urgency to this work. The pace 
and intensity of this research are sioe ds increased 
by conferences specifically targeting this fam 

The First International Rubiaceae CR held 
at the Missouri Botanical Garden in St. Louis in 1993, 
brought together students of Paleotropical and Neo- 
tropical groups for the first time; selected proceedings 
were published in 1995 in the Annals of the Missouri 


355-439). 

International Rubiaceae Conference, 
held at the National Botanie Garden of Belgium in 
Meise in 1995, focused on Rubiaceae as part of the 


Botanical Garden (volume 82, issue 3, pp. 
The Second 


Gentianales (then a fairly new consensus classifica- 
tion for the family) and delimitation of subfamilies and 
problematie tribal and generic complexes; the ful 
proceedings were published in 1996 in Opera 
Botanica Belgica (volume 7, pp. 1-432). For more 
than 10 years after that, no meeting was held until 
2005, when a half-day symposium focused on 
Rubiaceae during the XVII International Botanical 
Congress in Vienna (no proceedings were published). 
This symposium clearly showed interest in and need 
for a longer meeting. 

The Third International Rubiaceae Conference was 
subsequently co-organized by the Katholieke Uni- 


1 The editors of these proceedings express their gratitude to Victoria C. Hollowell, Scientific editor and head of Missouri 


Botanical Garden Press; Beth Parada, managin: 


publishing the proceedings of the Third International Rubiacea 


We thank additional members of th 


hospitality during the m oia We are grateful to the authors 


volume published and acknowledge the fi 


tific committees for their help in ee a success 
Birgitta Bremer, Helga Ochoterena, Elmar e Erik Smets, and St 
and Abstracts book (Scripta Bot. Belg. 40). W 


ir collaboration in 
The N 


teven Dessein. 


and reviewers for “their hel 


Research Foundation—Flanders 


ies) 


for 
sp 
FWO), the Systematics Association, the National Botanic Garden of Belgium, Easyware 


van Bouchout BE-1860 n Belgium. d fgov. be. 
ouis, Missouri 63166, WA. 


Lp 


^ Laboratory of Enc Systematics, Katholieke Universiteit Leuven, ed Arenberg 31, P.O. Box 2437, BE-3001 


Leuven, Belgi uzy.l 


um. 
doi: 10.3417/2008118 


kuleuven 


ANN. Missouni Bor. Garp. 96: 1-3. PUBLISHED ON 23 Arni, 2009. 


Annals of the 
Missouri Botanical Garden 


versiteit Leuven and the National Botanic Garden of 
Belgium from 18 to 21 September 2006 (De Block et 
al., 2006) an 


to provide a forum for all Rubiaceae and Gentianales 


d held in Leuven. The conference aimed 


researchers to present results achieved in the decade 
since the second conference, and a venue for 
discussions and networking. Six themes were sched- 


uled: systematies at the family level, systematies at 


Rubiaceae, studies of other members of the Gentia- 
nales and the order as a whole, studies of particular 
genera, and Rubiaceae checklists. In addition, two 
keynote lectures reviewed ine different but intercon- 
nected subjects. The present v e groups the two 
eynote lectures and 10 other ene from this 
congress. 

From the first keynote lecture (in order of 
presentation at the congress) Graham reviews the 
fossil record for Rubiaceae in detail, covering fossils 
of 125 taxa attributed to the family from deposits as 
old as the Late Cretaceous and Paleocene; this review 
concludes that the oldest “dependable” fossils (i.e., 
those useful for dating phylogenies) are from th 
middle and late Eocene and represent Emmenopterys 


Oliv., Faramea Aubl., Guettarda L., 


Lam. From the other keynote Joni Bremer here 
years 


and Canthium 
gives a historical overview of 15 of molecular 
phylogenetic studies of Rubiaceae, covering a period 
that saw tremendous advances in our understanding of 
the phylogeny of the family. 

The sessions addressing studies of the family, 
tribal, and genus levels were dominated by molecular 
contributions, but did include a few studies of 
morphological and chemical characters. This reflects 

e general recent trend in systematie work, lamented 
by some and applauded by others. This trend is 
clearly evident in the present volume, which includes 
nine papers from these sessions 

Addressing the tribal level, posue et 
al. provide new insights into the phylogeny of the 
large Paleotropical tribe Vanguerieae. Focusing on 
the dioecious faxa within this tribe, their results point 
to a single origin of functional dioecy from h ph 
roditism followed by subsequent reversals back to the 
dec M condition in certain genera. Here also, 

ortés- nfirm the monophyly of the 
cone genus Retiniphy llum Bonpl. its place- 
ment in the monotypic tribe Retiniphylleae, and that 
this tribe is sister to the core members of subfamily 
Ixoroideae (i.e., tribes Coffeeae, Gardenieae, E 
Octotro Van 
Delprete discusses the taxonomie history, i 


opideae, Pavetteae, and Vanguerieae). Also, 


netic evidence, 


oO 


and reproductive biology of th 
i their 
unusual pollen catapult mechanism as the character- 


Posoquerieae, and cites 


istic feature of both genera of this recently described 
"E e. 

ressing problems originally confined within 
de but that 
Mouly et al. present a molecular phylogenetic analysis 


t finally require tribal readjustments, 
of the species-rich pantropical genus Ixora L., which 
Broaden the 
gly 


hey show to be polyphyletic. ing 
to encompass 


potyphy. 

circumscription of Ixora accordin 
additional species also necessitates a redelimitation of 
the tribe Ixoreae, for which these authors adopt a 
narrow circumscription and describe two new tribes, 
Aleisantheae of Indomalaysia and Greeneeae of 
Southeast Asia. Also, Rova et al. present a molecular 
phylogeny of the large, morphologically variable 
Neotropical genus Rondeletia L., which they show to 
be polyphyletic. Delimiting monophyletic groups 
within this traditionally circumscribed i leads 
the authors to divide Rondeletia s.l. and propose new 
delimitations of the tribes Rondeletieae m Guettar- 
deae. 

Considering problems that lie within tribes, Groe- 
ninckx et al. address the phylogenetic relationships 
within the herbaceous tribe Spermacoceae s.l. base 
on a broad sampling across the major lineages of this 
tribe. Along with some delimitations of problematic 
groups for future study, these authors stress the nee 
for morphological data to support clades and relation- 
ships found in molecular analyses. Here also, 
Martínez-Cabrera et al. document the variation in 
leaf and petiole anatomical characters and evaluate 
their phylogenetic value within the Neotropical tribe 
Hamelieae. 

Considering individual LM Tosh et al. present a 
lo d ic ae o o-Malagasy genus 
a A. Ric h. (Ceffeese) pe conclude that its 


Ee Tricalysia and subgenus Empogona (Hook. 


phy 
Tricalys 


f£) Robbrecht do not form a monophyletic clade, 
leading them to restrict the genus circumscription to 
subgenus Tricalysia and return subgenus Empogona 
to generic rank. Here also, Cabral presents a revision 
of the Neotropical e Galianthe Griseb., which 
Le 39 species in two sections, with section 
Laxae E. L. Cabral newly ue here. 

Last pe not least, Davis et al. analyze several 
aspects of the data available in the Royal Botanic 


apps.kew.org/wesp//home 
wide, both Rubiaceae 2 and others, regularly 
use this database for inquiries of synonymy, correct 
spelling, correct ex d and place of publication 
of Rubiaceae names. However, the data compiled here 
also address distribution, diversity, and endemism of 
Rubiaceae and show the taxonomic efforts in this 
family. A notable analysis presented here shows that 
the number of new Rubiaceae species described each 


Volume 96, Number 1 
2009 


De Block et al. 3 
Introduction 


species of Rubiaceae by destruction of their habitat, 
there is no time to waste filling in the gaps in our 
knowledge of this family. We are called upon to 
continue and, if possible, intensify our efforts to study 
the Rubiaceae, and to develop strong collaborations 


amongst ourselves as well as with specialists of other 
disciplines to preserve as much Rubiaceae diversity 
as possible. 


Literature Cited 

De Block, P., S. Dessein & E. Robbrecht (editors). 2006. 
Third International Rubiaceae Conference, Programme 
and Abstracts. Scripta Bot. Belg. 40: 1-92. 


A REVIEW OF MOLECULAR 
PHYLOGENETIC STUDIES OF 
RUBIACEAE: 


Birgitta Bremer 


ABSTRACT 


Rubiaceae is one of the five largest families of flowering plants with over 13,000 species. We have seen a tremendous 


1991 to 2005; som tionships are com 


2005, ca. 50 Suissa reconstructions bos the family dd been published ba 
rbcL sequences, but 13 PA ipu: id bee 


studies are bas 


ed on ITS and 
(as of 2005) are ia rps16, irn(T)L-F, rbcL, and ITS. W. 


three subfamilies and over 43 tribes; subfamily lu onda (Chioe ococceae, Cinchoneae, Gi 
b EE o MR e meca A o ed d Bertier 
, Mussaendea pidea 


Cremaspo: , Gardenieae, Ixo 


subfamilies. Tuo of these mbes, Condens 
eentihésfoca However, we hav 


increase in our ed des of the phylogeny of the nd ier studies on molec 
relat pletely un 


ular data during the 15-year ru from 
traditional classification. At t 
sed eh more than 4 


ted and different from 


sed. M ank 
a fr: ee of the family phylogeny with support d 
melieae, Hillieae, 


eae, Coffeeae, Cond. eae, 
e, Posoquerieae, Retin iphylleae, Sabiceeae 
stemmateae, Coussareeae, Craterispermeae, 


is place 


e and Morindeae, are paraphyletic/poly phyletic. Only about half of the tribes have 
e seen increased interest in using Rubiaceae phylogenies E studies 

of ecology, puolutióm and biogeography, e.g., and also for morphological and anatomical investigations uti 

traits, flower types, and myrmecophytism has been investigated, an 


of fruit 
d biogeographic patterns for sociis. taxa in Mica the 


Caribbean, and the Pacific have been studied. In addition, distribution of pollen types, chemical substances, and wood 


characteristics have been compared with 


molecular phylo 


genies. 


ey words: p dE classification, ecology, Sinan ITS, morphological characters, phylogeny, rbcL, rpsl6, 


Rubiaceae review, irn(T)L-F. 


The Rubiaceae family, with more than 13,000 
species (Govaerts et al., 2006), has been the subject of 
many molecular phylogenetie studies during the 15- 
year period from 1991 to 2005. 
summarize the main conclusions from these studies. 


Here, I review and 


By 
from the late 1970s and early 1990s. In 1979, po first 
cladogram of Neurocalyx Hook. placed the genus in 
Argostemmateae (Bremer, 1979); in 1990 the first 
cladogram of Xanthophytum Reinw. ex Blume s 
1990). B 


phylogenies were published in association with m 


the genus in Hedyotideae (Axelius, 


generic revisions, and the trees were the result of 

with morphologica 

on and Persson 0991) 1n m a 
e Cinchonea 


Their analysis Hn in a new 


simple parsimony analyses 
characters. Andersso 
very early morphological analysis of 
and relatives. 
circumscription of Cinchoneae, a description of the 
new tribe Calycophylleae, and an emended tribe 


Coptosapelteae. The Cinchoneae tree has a low 


resolution with many odd relationships compared to 
later molecular analyses (Razafimandimbison 

Bremer, 2001, ; Rova et al., ; Andersson 
& Antonelli, 2005). The relationships in Neurocalyx 
and Xanthophytum have not yet been tested by 
molecular data, but both genera have been transferred 


to tribe Ophiorrhizeae based on sequence data 


— 


Bremer & Manen, 2000). Very soon after the analyses 
described above, molecular data (from 1991, see 
below), or combinations of iini and morpholog- 
ical data, analyzed with com drei replaced 
simple manual morphological ps . There is no 
evident difference in quality b puede 
and molecular data, but because higher numbers of 
characters can be produced from DNA, it is easier to 
get better-supported trees (e.g., Bremer et al., 
During 15 years of molecular phylogenetic analyses 
of Rubiaceae taxa, from the beginning of 1991 to the 
end of 2005, ca. 50 studies have been published, 
which cover many parts of the family and address 
questions at different taxonomic levels, from closely 


! T would like to thank Kåre Bremer and Aum D is f 


Aaron Davis and 


Rafael G in Rubi 


Or 
to use Figure 1, and the Swedish Research Council and the Knut and Alice eee a for financial s 


sten Eriksson for permission 
support. 


2 Bergius Foundation, Royal Swedish Academy of Sciences and Botany Department, Stockholm University, SE-106 91 


Stockholm, Sweden. birgitta.bremer@bergianska.se. 
doi: 10.3417/2006197 


ANN. Missouni Bor. GARD. 96: 4-26. PUBLISHED ON 23 ApriL 2009. 


Volume 96, Number 1 Bremer 
2009 Molecular Phylogenetic Studies of Rubiaceae 


Theligoneae 
Rubieae 
Putorieae 
Paederieae 
Argostemmateae 
Anthospermeae 
Spermacoceae C Spermacoceae 
alliance Knoxieae 
L Danaideae 


Craterispermeae 


==== “Morindeae” 


Psychotrieae 
alliance 


Schradereae 


Gaertnereae 
Psychotrieae 
Coussareeae 
Lasiantheae 
c 


Rubioideae 


p 
Urophylleae 
Sabiceeae 
C Sipaneeae 
Posoquerieae 
Condamineeae 
Mussaendeae 
Octotropideae 
Cremasporeae 
Bertiereae 
=== "Gardenieae" 
Pavetteae 
Coffeeae 
Rubiaceae Alberteae 
Retiniphylleae 


xoroidea 


Vanguerieae 
Ixoreae 


Rondeletieae 
Guettardeae 
ex Cinchonoideae | —— —— iertiege 


—_ Naucleeae 
Hymenodictyeae 


Cinchoneae 
—L. Hillieae 
Hamelieae 
Chiococceae 
r~ Coptosapelteae 
L— iuculia 
Outgroup 


ure 1. Simplified majon ale consensus tree from MrBayes 3.1.1 Saree of 538 Rubiaceae taxa and 9420 
All resolved nodes and tribes haye 0.95 to 1.0 clade credibility (except Gieitardene, 

h 0.92) and are accepted as monophyletic Schradereae, and T 
represented by single taxa and thus could not be tested for mono hyly) Two tribes, Gardenieae and Mo PER are 
paraphyletic/polyphyletic. Presented (slightly modified) at the Third International Rubiaceae Conference in Leuven in 2 


Annals of the 
Missouri Botanical Garden 


related species to the whole family. Except for the first 
analysis of restriction site data, all later studies have 
used sequence data, and the most popular markers 
(the largest number of studies) have been ITS and 
rbcL. Altogether, 13 different sequence markers have 
been used, seven from chloroplast DNA (cpDNA) 
(atpB-rbeL, ndhF, matK, rbcL, rps16, trn(T)L-F, trnS- 
G) and six nuclear DNA (ETS, ITS, nontranscribed 
spacer [NTS], pep-C large, pep-V small, Tpi). At the 
end of 2005, more than 4400 sequences from the 
family were available from GenBank/European Mo- 
lecular Biology Laboratory (EMBL) (excluding the 
double number of Coffea L. sequences produced for 
phylogenetics). these 4400, 

ost sequences are from rps16 (719), trn(T)L-F (672) 
rbcL (643), and ITS (323). In the future, we will see 


many more markers used in Rubiaceae, but of the 13 


purposes other than 
m 


that have been used so far, many underexplored 
(e.g., matK and ndhF for stes taxonomic levels and 
ETS and NTS for more closely related taxa). 

This paper is divided into two main parts. The first 
part focuses on phylogenetie reconstructions, studies 
covering the whole family, studies sorted under the 
three subfamily headings, first tribal studies, and 
finally genera studies. I have tried to discuss them in 
chronological order according to the first molecular 
study of the specifie group. Some studies have been 
difficult to classify according to taxonomic level 
unless the author(s) had indicated a focus on a 
specific rank. Studies including substantially new 
data, 
considered. The second part of this review is a 


not just reanalyzed data sets, have been 
presentation of studies in which a Rubiaceae 
phylogeny has been used to ask other questions about 

the family, c e.g., ecology, 
biogeography, anatomy/morphology, or chemistry. To 


concerning, evolution, 
assist the reader in navigating among all subfamilial 
and tribal names, refer to a phylogeny an 

classification (Fig. 1) presented at the Third Interna- 
tional Rubiaceae Conference in Leuven in 2006 
(Bremer & Eriksson, unpublished data). In the tree, 
three subfamilies and 43 tribes are well rp (all 
resolved nodes and tribes have 0.95 to 1.0 clade 


credibility, except Guettardeae, E 0.92; the 
molecular markers) and accepted as RA ANM. 


Schrader and 


Theligoneae are monotypic or represented ps bale 


Remigio llene, 


taxa and thus could not be tested for monophyly), and 
two tribes, Gardenieae and Morindeae, are paraphy- 
letic/polyphyletic. Representatives from all 43 of 
these tribes have been included in some of the 
analyses, but only 16 tribes have been the focus of 


specific studies. All genera discussed are listed in 


PHYLOGENETIC RECONSTRUCTIONS 
FAMILY RUBIACEAE 


The first attempt to reconstruct the Rubiaceae 
phylogeny based on molecular data was published in 

991 by Bremer and Jansen in the American Journa, 
of Botany. The data were from restriction site mapping 
of cpDNA. Included were 161 informative characters 
for 33 taxa and genera representing 17 different 
tribes. Unfortunately, no external outgroup was 
incorporated, which affected the De of the family. 
Several relationships Kp. in ea classifica- 
tions by Bremekamp (1 966), e t (1958), 
Bridson and Verdcourt e and Robbrecht (1988) 
were corroborated, but many new relationships 
disagreeing with earlier classifications were also 
proposed. The subfamily Rubioideae of Verdcourt 


— 


1958) was mostly monophyletic (including the tribes 
Rubieae, Anthosper' ypseleae, Hedyoti- 
eae, Psychotrieae, but excluding Hamelieae [Hame- 
lia Taca Hoffmannia Sw] and Ixoroideae fide 
Robbrecht [1988; including Coffeeae, Gardenieae, 


Pavetteae, and Vanguerieae but not Chiococceae [). 


meae, ee 


Several taxa earlier classified to Cinchonoideae (e.g., 

Calycophyllum DC., Diem L., Pinckneya Michx., 
nopus zsch) were shown to be closer to 

the subfamily nde It was also shown that the 

recircumscribed lirlichesidies (Robbrecht, 1988) was 

s lene the tribes Cephalantheae, Chio- 
and Van 


an ogo 


were not close to each other 
or to es Mns Comm. ex Juss., Guettarda 
L.). The subfamily Cinchonoideae was not supported 
as a monophyletic group in Bremer and Jansen (1991). 
New relationships included Chiococca P. Browne and 


Erithalis P. Browne of the Chiococceae as close to 


inchonoideae. It was also shown that 
Cephalantheae and Vanguerieae are closest to 
Naucleeae and Ixoroideae, respectively. 

uring the First International Conference on Ru- 
biaceae at the Miss ] Garden in 1993, an 


analysis of rbcL sequences from 49 Rubiaceae genera 


souri Botanica 


representing 23 x was presented (later published 
in Bremer et al, at study included out- 
groups from eden and also Oleaceae. Rubia- 
ceae came out as d group to the rest of Gen- 
with an rbeL study of the 

Panem (Olmstead e et a , 1993) and a a 
analysis of "ium and Gentianales (Bremer & 
Struwe, 1992). In the 1995 study, the family was 
classified into ies subfamilies: Rubioideae (includ- 
ing Rubieae, Anthospermeae, Hedyotideae, Morin- 
deae, Ophiorrhizeae, Psychotrieae, and Theligoneae), 
Ixoreae s.l. (including Coffeeae, Gardenieae, Pavet- 

teae, and Vanguerieae, as well as several genera of the 


Volume 96, Number 1 


Bremer 
Molecular Phylogenetic Studies of Rubiaceae 


former Cinchonoideae), and Cinchonoideae s. str. 
(including Cinchoneae, Chiococceae s.l., Guettardeae, 
Hamelieae, Hillieae, Naucleeae, and Rondeletieae). 
The genus Luculia Sweet was unresolved at the base of 
the family, and the genus Hintonia Bullock was 
Wer 
At about the same time, Ehrendorfer et al. 

published the first analysis of the atpB-rbcL spacer Me 


unresolved between Cinchonoideae an 


cpDNA in a short communication, foregoing a more 
co ud study of the Rubieae (Natali et al., 
1995; see below) the 199. 

oe in St. Louis. They showed results for eight 


that was presented at 


as Benth., Psychotri 
and Rubia L.) representing five tribes, and the saltas 
tree was concluded to be in agreement with the 
relationships based on the restriction site data, with 
Ixora and Coffea together as sister group to the rest. 
In a study investigating effects of the number of 
characters, the number of taxa, and the kind of data 
for bootstrap values within a phylogenetic tree, 
Bremer et al. 
i ubiaceae genera 
together with 11 outgroups representing the rest of 
the Gentianales were analyzed for rbcL and ndhF. lt 
was shown that the percentage of supported nodes 
within the trees positively correlated to the number of 
characters, but ue correlated to the number of 


taxa. Further, the ne subfamilies Rubioideae, 
Cinchonoideae, and Ix all m oo 
and highly a es d ap). Ther ere 


M two investigated genera, Luculia and pen 
pelta Korth., placed at the base i the Rubiaceae, that 
were left unclassified to subfa 

Rova et al. dde a phylogenetic 
analysis of trnL-F for a large data set including 154 
Rubiaceae sequences and 11 outgroups in a study 
to test what had been suggested to form a tight 
complex of the tribes Condamineeae, Rondeletieae, 
and Sipaneeae by Robbrecht (1988). Several earlier 
molecular studies had indicated that this SERRE 
E had no support (e.g., Bremer et al. 
Andersson & Rova, 1999). Rova et al. (2002) dial 
taxa n most parts of the family, and the results were 
very much in agreement with earlier molecular 
analyses. Their main conclusions were that most 
m ae a several Rondeletieae 
genera (Aleisanthia Ridl., Aleisanthiopsis Tange, 
Augusta Pohl, Greenea Wight & Arn., and Wendlandia 

) are members of the Ixoroideae, as are the 
Sipaneeae (Maguireothamnus Steyerm., Neobertiera 
Wernham, and Sipanea Aubl) and its sister clade 
(Gleasonia Standl., 


quena Aubl, the 


Molopanthera Turez., and Poso- 
atter two correspond to the 


circumscription of tribe Posoquerieae by Delprete et 


al. [2004]. Condamineeae (as the first Ixoroideae 


anthus Kuntze]} formed a supported but almost 


unresolved clade of Ixoroideae. Rova et al. (2 

Rondeletieae, and Gu 
letieae taxa, was paraphyletic. 
almost entirely Antillean in geographic distribution. 
Furthermore, there was support for i gre of 
everal genera from the genus Rondeletia L. (Arachno- 


E 
thryx Planch., Rogiera Planch. | Roigella. Borhi idi & M. 


position of Retiniphyllum Humb. . (Re 

phylleae) in the Ixoroideae (in Aaii fide 
Robbrecht, 1988) between Mussaendeae and the main 
part of Ixoroideae as proposed in Andersson and Rova 


presented new taxo 
sequenced for the first time: Allenanthus Standl. (close to 
rue EL Blepharidium Standl. (Ron- 
deletieae), C . (close to Hamelieae-Hillieae), 
Coutaportla D (Chiococceae), Dolichodelphys (close 
to Calycophyllum-Condaminea-Hippotis), Mazaea Krug 

Urb. ce E Neobertiera (Sipaneeae), Neo- 
blakea Standl. (close to Guettardeae—Rondeletieae), 
Phialanthus Griseb. e EUN Phylla- 
canthus Hook. f. (Chiococceae—Catesbaeeae), Phyllomelia 
Griseb. (Rondeletieae), Schmidtottia Urb. (Chiococceae— 
Catesbaeeae), and Suberanthus (Rondeletieae). 

The studies discussed above provide strong support 
for three large supported subclades corresponding to 
the subfamilies Rubioideae, Ixoroideae, and Cincho- 


ertain or unresolved 
nodes are under investigation by Rydin et al. [2009]. 
We still do not know how the genus Luculia and 

ribe Coptosapelteae are related to the three 
subfamilies, i example. To have a detailed phylo- 
genetic picture of the family and to understand 
circumscriptions of subgroups, we need sequence 
ata for all described genera, and, so far, more than 
200 genera have not been included in published 
molecular analyses. In most cases, morphological data 
or traditional classification can indicate a possible 
phylogenetic position, such as placing genera within 
tribes, but for some genera this is difficult. Further- 


Annals of the 
Missouri Botanical Garden 


Table 1. List of the 348 Rubiaceae genera discussed in the text, with tribal position. 


Genus Position Genus Position 

Acranthera Arn. ex Meisn. no tribe Ceratopyxis Hook. f. CHI 
Adin NAU Ceriscoides (Hook. f IE d GAR* 
es Ri dsdale NAU Chalepophyllum Hoo SIP 
Afrocanthium (Bridson) Lantz VAN Chassalia Poir. PSY. 

S Bremer Chazaliella E. M. A. Petit & Verdc. PSY 
Aidia me GAR* Chimarrhis Jacq. CON 
Alber ALB Chiococca P. Browne 
yee IXOR, no tribe Chione 


a Ridl. 
pe ome Tange 
Alibertia A. Rich. ex DC. 
Allenanthus Standl. 


asma Bremek 
pore Pen 
Ancylanthos D 


Anthorrhiza C. P Huxley & Jebb 
Anthospermum L. 
Antirhea Comm. ex Juss. 


Asemnaniha Hook. f. 

Asperula L. 

Atractocarpus Schltr. & K. Krause 
rre 


Bikkia 


Carterella Terrell 
oe A. Rich. 
Catesbaea L. 
een Wolf 


Cephalanthus L 


TXOR, no ae 
GAR 
c GUE/RON 
CON 


GAR* 


R 
GAR* 
* 


TXOR, no tribe 


Ciliosemina Antonelli 
na L. 

Cinchonopsis L. Andersson 
Coccocypselum P. Browne 
Coddia Verdc. 
Coelospermum Blume 

offea. L. 
Commitheca Bremek. 


daminea DC 
Conostomium Sut JC 
Coprosma J. R. Forst. X 3 Tori 
Coptosapelta Korth 
Corynanthe Welw. 
smibuena Ruiz & Pav. 


C. F. Chiu 


Dendrosipanea Ducke 


Deniella J. R. Forst. & G. Forst. 


idymaea Hook. 
Didymosalpinx Keay 
Diodia L. 

Dioicodendron Steyerm. 
Diplospora. 


Do lode oy K. Schum. & K. Krause 


Duperrea Pierre ex Pit. 


Duroia 


Ecpoma c 
Elaeagia Wedd. 
Emmenopterys Oliv. 
pics P. Browne 


nodea Sw 
eid Salisb. 
Exostema (Pers.) Bonpl. 


Lp 
Durringtonia R. J. F. Hend. & Guymer 
K. Schum 


CHI 
c HAM/HIL 
CIN 


Volume 96, Number 1 
2009 


Bremer 
Molecular Phylogenetic Studies of Rubiaceae 


Table 1. Continued. 


Genus Position Genus Position 
Fadogia Schweinf. VAN oya Cavaco VAN 
Faramea A COU Limnosipanea Hook. f. SIP 
Feretia Delile OCT Luculia Sweet no tribe 
Fernelia Comm. ex Lam. OCT Ludekia Ridsdale 
riri Lam. GAE Macbrideina Standl CON 
Gal RUB Macrosphyra Hoo GAR* 
DN Thunb. ANT Maguireothamnus Steyerm SIP 
Gardenia Ellis GAR* Manostachya Breme SPE 
ipa L. GAR* Margaritopsis C. Wright PSY 
Geophila D. Don PSY Maschalocorymbus Bremek URO 
leasonia Standl c POS ee he Schum.) Hoyle GAR* 
Glossostipula Lorence GAR* azaea g & Urb. RON 
Gomphocalyx Baker S Mela. an Colla GAR* 
Greenea Wight & Arn. IXOR, no tribe —— Metadina Bakh. f. NAU 
Guetiar GUE Meyna Roxb. ex Link VAN 
Gynochthodes Blume MOR* Mitchella L. MOR* 
Gyrostipula J.-F. Leroy NAU Mitracarpus Zucc. ex Schult. & Schult. f. SPE 
Haldina Rids NAU Mitrag Korth. N 
Hamelia Jacq HAM Mitriostigma Hochst GAR* 
Hedyotis L. SPE lopanthera Turez. S 
Heinsia DC. MUS Morelia. A. Rich. ex DC GAR* 
Heinsenia K. Schum. GAR* TURON Vieill. CHI 
Hekistocarpa Hook. f S MOR* 
eterophyllaea Hook. f. COU Multidentia S: VAN 
Hindsia Benth. ex Lindl. COU aenda MUS 
Hintonia Bullock CHI Masons Baill. CON 
Hippotis Ruiz & Pav. CON en Rein ARG 
Hoffmannia Sw. HAM s ex Juss. IXO 
Houstonia L. SPE Myrmecodia Jack PSY 
Huichinsonia Robyns VAN Myrmeconauclea Merr. NAU 
Hydnophytum Jack PSY Myrmephytum Becc. PSY 
Hydrophylax L. f. SPE Nauclea L. NAU 
menocoleus Robbr. PSY Neblinathamnus Steyerm. SIP-tent 
Hymenodictyon Wall. HYM enax Gaertn ANT 
Hyperacanthus E. Mey. ex Bridson GAR* Neobertiera Wein SIP 
Ibetralia Bremek. GAR* Neoblakea St c GUE/RON 
Isertia Schreb. ISE Neolamarckia Bosser NAU 
Isidorea A. Rich. ex DC. CHI Neolaugeria Nicolson GUE 
Ixora L. IXO Neoleroya Cavaco VAN 
Janotia J.-F. Leroy NAU eomussaenda Tange MUS-tent 
Joosia H. Karst CIN Neonauclea Merr NAU 
Kailarsenia Tirveng. GAR* Nertera Banks & Sol. ex Gaertn. ANT 
Keetia E. Phillips VAN Neurocalyx Hook. OPH 
Kelloggia Torr. ex Benth. & Hook. f. c RUB Normandia Hook. f. ANT 
ee J. H. Kirkbr. ISE Notopleura (Benth. & Hook. f.) Bremek. PSY 
Knoxi KNO Ochreinauclea Ridsdale & Bakh. f. NAU 
a Hi arv. OCT nlandia L. SPE 
Kutchubaea Fisch. ex DC. GAR* Oldenlandiopsis Terrell & W. H. Lewis 
LATA Klotzsch CIN Oligocodon Keay GAR* 
La, E. . ex Robyns VAN Opercularia Gaertn ANT 
DN M MUS Ophiorrhiza L. OPH 
Lasianthus Jack LAS Oreopolus Schltdl. COU 
Leptactina Hook. f. PAV Osa Aiello CHI 
Leptodermis Wall PAE Oti ra Zucc KNO 
Leptostigma Arn ANT Otomeria Benth KNO 
L OPH Oxyanthus DC. GAR* 


Annals of the 
Missouri Botanical Garden 


Table 1. Continued. 


Genus Position Genus Position 
Oxyceros Lour. GAR* b ii dd Humb. & Bonpl. RET 
Pachystigma Hochst. VAN ichardia L. SPE 
Paederia L PAE ogiera Planch. RON 
Pagamea Aubl GAE Roigella Borhidi & M. Fernandez Zeq. RON 
Palicourea Aubl PSY Rondeletia L RON 
Parachimarrhis Ducke CON Rosenbergiodendron Fagerl. GAR* 
Paracoffea J.-F. COF Rothmannia Thunb GAR* 
p d a HYM ubia L. RUB 
Paragenipa Ba OCT Rudgea Salisb. PSY 
Parapentas Bremek KNO Rustia Klotzsch CON 
Pauridiantha Hook. f. URO Rutidea DC PAV 
Pausinystalia Pierre ex Beille NAU Caan Blume VAN 
avetta L. PAV Sabicea Aubl SAB 
Peniagonia Benth. CON Salzmannia DC. CHI 
Pentaloncha Hook. f. SAB-tent prn Azfel. ex R. Br. NAU 
Pentanisia Harv. KNO Schizomussaenda H. L. Li MUS 
dde ue Rendle SPE Schizostigma Arn. ex Meisn. SAB-tent 
KNO Schmidtottia Ur CHI 
dun Gaill) A Arènes VAN Schradera Vahl SCH 
Pertusadina Ridsdal NAU Schumanniophyton Harms GAR* 
Phialanthus Cu CHI Scolosant. he CHI 
Phuopsis (G eds d f RUB Scyphiphora C. F. Gaertn. c IXO/VAN 
Bur M CHI Spa Balf. f. VAN 
ANT Serissa Comm. ex A. Juss. PAE 
Phyllomelia Griseb. RON Sherardia L RUB 
Phylohydra. f SPE Sherbournia G. Don GAR* 
Picardaea Ur CON Siemensia Url CHI 
Pimentelia Wedd CIN-tent Sinoadina Ridsdale NAU 
Pinckneya Michx. CO Sipanea Aubl. SIP 
Pittierothamnus Steyerm. SAB-tent Sipaneopsis Steyerm. SIP 
Placopoda Balf. f. KNO olenandra H f CHI 
Platycarpum Humb. & Bonpl. SIP Sommera Schltdl. CON 
Pogonopus Klotzsch CON acoce L SPE 
Pomax ANT spermadictyo xb PAE 
Porterandia R idl. GAR* Sphinctanthus Benth GAR* 
Portlandia P. Browne CHI quamellari c PSY 
Posoqueria Aubl POS Stachyarrhena Hook. f. GAR* 
Pouchetia OCT Stenaria (Raf.) Terrell SPE 
Praravinia Korth URO Stenostomum C. F. Gaertn. GUE 
Pravinaria Bremek URO teyermarkia Stan SIP-tent 
Preussiodora Kea GAR* Stilpnophyllum Hook. f. CIN 
Pseudocinchona A. Chev. ex Perrot NAU Stipularia P. Beauv SAB-tent 
Pseudomussaenda Wernham MUS Streblosa Korth PSY 
Pseudopeponidium Arènes VAN Strumpfia Jac c CHI 
Pseudosabicea N. Hallé SAB p E Borhidi & M. Fernandez Zeq. RON 
Psilanthus H COF uk . $m GAR* 
Psychotria L. PSY "uen ends & Sastre GAR* 
Psydrax Gaertn. VAN Tamridaea Thulin € B. Bremer SAB 
Psyllocarpus Mart. & Zucc. SPE Tapiphyllum Robyns VAN 
Pteridocalyx Wernham SIP-tent Tarenna Gaertn. PAV 
toria, Pers. PUT Tarennoidea Tirveng. & Sastre GAR* 
Pyrostria Comm. ex Juss. VAN Temnopteryx f SAB-tent 
Ramosmania Tirveng. & Verdc. OCT Theligonum L. THE 
Randia L. GAR* Timonius DC GUE 
Raritebe Wernham URO Tocoyena Aubl. GAR* 
eadea Gillespie PSY Tricalysia A. Rich. ex DC. COF 
Remijia DC CIN Trichostachys Hook. f. LAS 


Volume 96, Number 1 


Bremer 
Molecular Phylogenetic Studies of Rubiaceae 


2009 
Table 1. Continued. 
Genus Position Genus Position 

Trukia eh GAR* Wendland TXOR, no tribe 
Uncaria Schreb. NAU Versteegia Valeton IXO 
Urophyllum Wall. URO Virectaria Bremek. SAB 
Valanti RUB Wittm nthus Kuntze CON 
Vangueria Juss VAN Xanthophytum Reinw. ex Blume OPH 
Warszewiczia Klotzsch CON Yutajea Steyerm. ISE 


* Paraphyletic/ qus tribes. 


I PORIE, E E 


Abbreviations: e, without tribal position (taxon has I 


nvestigated, but has not been placed within any 


described tribe); c, Fe to (taxon is sister group to or close to one or two tribes); tent, tentatively (taxon is not molecularly 


investigated but h uggested to be include 


ubieae; SAB, Sabiceeae; SCH, Schradereae; SIP, Sipaneeae; 


SPE, perio: THE, Theligoneae: URO, Urophylleae: VAN, Vanguerieae. 


more, if Rubiaceae should become the perfect model 
family for ecological, evolutionary, biogeographic, or 
other studies, we must work hard over the coming 
years with the challenge to sequence all described 
genera and species. 


SUBFAMILY RUBIOIDEAE 


At the Second International Conference on Rubia 
ceae in Brussels in 1995, Bremer (1996) focused on 
subfamily Rubioideae; 59 taxa representing most 
tribes of the subfamily were investigated for rbcL. 
The analysis showed that Anthospermeae, Rubieae, 


Spermacoceae s.l. (including the Pentas group 
Knoxieae [Pentas, Carphalea Juss., Parapentas Bre- 


and Placopoda Balf. f.], 
Hedyotideae, and Spermacoceae s. str.), and Psycho- 


mek., Pentanisia Harv., 


— 


trieae s.l. (including also Morindeae and Gaertnereae) 
are monophyletie. Paederieae and Ar; burg 
yletie. Lasianthus Jac 


Gaertnera Lam. were shown not to belong to seen 


were shown to be polyp 
trieae s. str. The following genera from different tribes 
were represented by single species and thus could not 
be tested for mo 
phylogenetically: Coccocypselum P. Browne (Coussar- 
Faramea | Aubl. 
(Coussareeae), Mycetia Reinw., Ophiorrhiza L., Paur- 
idiantha Hook. f. (Urophylleae), and Theligonum L. 

Myceti h b to 


enus Mycetia was shown to be close 


nophyly, but could be positioned 


eeae), Danais Comm. ex Vent, 


Argostemma Wall. and not a member of the Isertieae 
(Robbrecht, 1988). 

A few years later, Andersson and Rova (1999) 
published an analysis of rps/6 sequences from 143 
Rubiaceae taxa and five outgroups, also focusing on 
subfamily Rubioideae. The results confirmed those 


based on rbcL data (Bremer, 1996) for the main groups 
of the family, but more taxa were included and the 
A few 
differences between the rpsl6 = the rbcL results 


support was stronger for several clades. 


a, Spermacoceae 
76% vackknife 
support including three of the tribes recognized by 


were revealed. In the rb 
forms one monophyletic FE ME 


Andersson and Rova (1999), Spermacoceae, Heyoti- 
deae, and Knoxieae. In the rps16 uie Knoxieae is 
instead sister to a larger group of acoceae, 
Heyotideae, and also Paederieae and lubens but 
without support. Morindeae (80% bootstrap support) 
is found to be monophyletie, which disagrees with the 
rbcL data. The included and supported tribes of the 
Rubioideae from the base of the tree were the 
following: Urophylleae m Wall., J 
10046]. e 
reeae (Coussarea Aubl., 
amea [7696], Coccocypseleae (100%) together er 
the two unclassified ar Hindsia Benth. ex Lindl. 
and Declieuxia Kunth, Cruckshanksieae (Heterophy: 
Hook. f., Oreopolus Schltdl. [7896 ), Gaertnereae 
esac, Pagamea Aubl. [10095], Schradereae 
(Schradera Vahl, single taxon), Morindeae (Morinda 


y 


S. 
F 


Psychotrieae (Psychotria, Chassalia Poir., 
liella E. M. A. Petit & Verdc., A pe D. Don, 
Hydnophytum, Margaritopsis C. ght, Myrmecodia 
Jack, Palicourea Aubl, Readea M oed Rudgea 
Salisb., Squamellaria Becc., Streblosa Korth. [99%]), 
J Otomeria 
[100%], Antho- 
rst. & G. Forst., Galopina 
Nertera 


Knoxieae (Knoxia L., Oti 
Benth., Pentas, 
spermeae (Coprosma J. R. 

Thunb., Leptostigma Arn., 


ophora Zucc., 
Pentanisia Harv. 


Nenax Gaertn., 


Annals of the 
Missouri Botanical Garden 


Banks & Sol. ex Gaertn., Opercularia Gaertn., Phyllis 
L. [5396], Rubieae (Rubia, Asperula L., Crucianella 
a L., Valantia L. [100%), and 
Spermacoceae CMM L., Borreria G. Mey., 
Crusea Cham. & Schltdl., Diodia L., Ernodea Sw., 
m Zucc. ex Schult. e 1 , Psyllocarpus 
Ma , Richardia L. 


PN and Wee were ee as in 


., Galium, Sherardi 


e tribes 


Bremer (1996). The genus d des is paraphyletic 
in agreement with e et al. yi 
new eny and a omprehensive 
classification bi Rubioideae were pium 
Bremer and Manen (2000). They analyzed 151 genera 
with three different molecular markers, rbcL, atpB- 
rbcL, and rps16 (latter data from Andersson & Rova, 
1999). The separate markers and combined analyses 
The 
(Ophiorrhiza, Neurocalyx, Lerc 


Ophiorrhizeae 
ea L., Xanthophytum), 
Urophylleae (Urophyllum, Amphidasya Standl., 


gave similar results. 


mitheca Bremek., Maschalocorymbus Bremek., Prar- 
avinia Korth., Pauridiantha), 
Lasiantheae (Lasianthus, Trichostachys Hook. f.), and 
Coussareeae formed a grade 


Pravinaria Bremek., 


to the rest of the family, 
which consisted of two newly established but informal 
groups (with 99% 
respectively): the Psychotrieae alliance (Psychotrieae, 
Crat } Benth.] 


and 100% bootstrap support, 


erispermeae [Craterispermum Benth.], Gaertner- 
eae, Morindeae [paraphyletic], Schradereae} and the 
A alliance De E donar dd 
eae, Arg mateae eae, Paederieae [para- 
Be reine ee Of the a 16 
Rubioideae tribes, 11 were in agreement with earlier 
Coussarieae, and 


circumscriptions. Ophiorrhizeae, 


permacoceae received wider circumscriptions, and 
Lasiantheae and Danaideae were described as new 
All monophyletic tribes received 100% bootstrap 


(except for Psychotrieae, with only 81% 


From the studies outlined above, there is support 
for most of the Rubioideae tribes and the many 
relationships between them. However, at the end of 

05, only seven of the tribes had been the aoe of 
detailed studies, It should be 
stressed that several tribes and also ELS 


presented below. 


between tribes (e.g., the basal clades Coussareae, 
Lasiantheae, 
within the P 

tion. Rubioideae is probably the best understood 


Ophiorrihizeae, Urophylleae, and clades 
Psychotrieae alliance) are under investiga- 
subfamily phylogenetically, but still only a minority of 
e been 


tant task for the coming years will be to analyze and 


its species hav investigated. The most impor- 
sequence most species of the large and problematic 
genera. Rubioideae contains 11 of the 20 largest 
genera of the family (Psychotria, Galium, Ophiorrhiza, 
Oldenlandia L., La- 


Palicourea, Spermacoceae, 


sianthus, Faramea, Asperula, Argostemma, and Cou. 

sarea). These genera together contain about 40% of al 
of these 
genera represent much of the Rubiaceae species 


species in the family and, because some 
diversity, understanding of their phylogeny would be 
an de asset for deeper evolutionary studies. 

e Rubieae was investigated by Manen et al. 
1994), PN used the atp 
the tribe. They found support for a monophyletic 


— 


B-rbcL spacer of 25 species of 


Rubieae, and the two investigated species of Rubia 
to be sister to the rest of the tribe. Manen 
and coworkers identified four further clades, but with 


were found 


low or moderate bootstrap support. The highest 
uU (87% bootstrap support) was for the Sherardia 

ade (Sherardia together wi 
(Griseb.) Hook. f) a 
support was found for the pad clade (Asperula 


Mon 


together with Galium elongatum C. Presl and G. 
re L.). The relationship between the four clades 

and was paraphyletic. Later, 
Natali et al. (1995) added more sequences to the 
Manen et al. (1994) data set, for a total of 70 Rubieae 
species and 25 taxa of 12 other tribes of Rubioideae. 
They got 100% bootstrap support for tribe Rubieae 
and subfamily Rubioideae. They excluded Ophior- 
rhizeae, and, with that circumscription, the subfamily 
was also characterized by a 204 bp deletion in the 
atpB-rbcL region. Natali et al. (1995) divided the 
Rubieae into the same five clades as in Manen et al. 
(1994), but with lower support still 
monophyletic (100% support) and sister to the rest. 
They showed that the genus Asperula is paraphyletic, 


Rubia is 


with all added species instead belonging to their 
and Natali (1996), i 


article about the deletion in the atpB-rbcL region los 


Sherardia clade. Manen 


of an atpB promoter) in the Rubioideae, investigated 
the atpB-rbcL spacer from representatives of the whole 
family, but with a main focus on subfamily Rubioi- 
deae. They presented a tree for 22 genera (e. refer 
to an analysis of 111 taxa, which was not presented in 
the article). They rooted the published tree between 
subfamily Ixoroideae (Coffea and Ixora) and the rest. 
The Cinchonoideae, including five genera, was sister 
to a clade including their ndis and Ophior- 
d strong support for Rubioideae 
o Poen and ee (including the 
Ru a Hook. £). Rubi 


rhiza. 


two g and Didymae ubieae 
was sister to Theligoneae and Putoria Pers. and these 
are sister to Paederia L.; other Rubioideae taxa in the 
analysis included Anthospermeae, Coccosypseleae, 
Hedyotideae, Morindeae, Psychotrieae, and Sperma- 
coceae. Their results agree with the rbcL data (Bremer 


& Jansen, 1991; Bremer et al., 1995) that Hamelieae 
oes not belong to Rubioideae but instead to the 


a 


Cinchonoideae. Their main conclusion is that the lack 


Volume 96, Number 1 


Bremer 
Molecular Phylogenetic Studies of Rubiaceae 


of the atpB promoter for the Rubioideae excluding the 
Ophiorrhizeae “gives strong evidences on the bound- 
ary between the subfamily Rubiodieae and the other 
(Manen & Natali, 
they do not suggest any taxonomic position, or to 
which subfamily Ophiorrhizeae belon 
article, Natali et al. ( ublished the same tree 
based on atpB-rbcL data for the 22 genera, but they 
also analyzed the Rubieae with a denser sampling of 
78 Rubieae taxa. The res 


analysis in Natali 


Rubiaceae" 56). However, 


gs. In another 


ult agrees with their earlier 
(1995) but divides the 
Rubieae into seven clades, now with Didymaea as 
sister to the rest, followed by the clades Rubia, 
sperula sect. Asperula, Asperula sect. Glabella, 
Sherardia, Cruciata Mill., and Galium sect. Galium 
Only Rubia was highly supported as monophyletic. 


et al. 


Despite the extended sampling, the relationships 
between the different groups were unresolve 
Kelloggia Torr. ex Benth. & S (Pacdericas 
fide Robbrecht [1988], but in Backlund et al. [2007] 
without tribal position}, a genus of two species with 
disjunct distribution in western North America and 
the western part of eastern Asia, was analyzed with 
three chloroplast markers (rbcL, atpB-rbeL, rps16) by 
Nie et al. (2005). They showed that the genus is 
monophyletic and sister to the Rubieae. Kelloggia was 
included in a Ph.D. thesis by Backlund (2005), 
and the same stn of the genus close to Rubieae 
well rted. It further demonstrated 
an " 2005) that ils clade o 


ubieae is sister g 


also 


Theligoneae— 
roup to a reestablishe 
tribe Putoricue (a position that makes the rest of the 
Paederieae monophyletic). 

The taxonomically complex tribe Psychotrieae and 
the very large genus Psychotria were molecularly 
investigated for the first time by Nepokroeff et al. 
(1999). They analyzed 85 taxa for ITS and rbcL. The 
results suggested that Psychotria is broadly para- 

yletic. Taxa earlier assigned to Psychotria, Psycho- 
tria sect. Notopleura Benth. & Hook. f., and subgenus 
O O plus Palicourea were 

trieae Mie to 
d to be 
restricted to a monophyletic group ae two 


closer to gene ycho 
o o poda was sugges 


subclades. One subclade is Pacific in distribution and 

udes the m ao o Hydnophytineae 
including Hydnophytum, Anthorrhiza C. R. Huxley & 
Jebb, Myrmecodia, i Bees) as a sub- 
group. The other subclade included Psychotria subg. 
Psychotria and subgenus Tetramerae . A. Petit. It 
was also shown that the genus beu was not a 
member of the Psychotrieae but closer to Coccosypse- 
lum. Later, Andersson (2002a) sequenced rps/6 for 
ex. The result 
was very much in agreement with Nepokroeff et al. 


111 species of the Psychotria comp 


(1999). Andersson also analyzed a combined data set 
the ITS sequences of Nepokroeff et al. [1999] and 
their rps16 sequences) for 15 taxa 


=> 


that were shared 
between the two studies. That analysis resulted in a 
tree with three well-supported clades, the outgroup 
Chassalia, Geo- 
phila, Hymenocoleus Robbr., Notopleura (Bent 

Hook. f.) Bremek., Rudgea, Palicouria), two Psycho- 


tria subclades, Psychotria s. str. (= 


— 


including, e.g., Carapichea Aubl., 


subgenus 

Psychotria, and subgenus Tetramerae in Nepokroeff 
et al. [1999], and a Pacific subclade (including 
several Psychotria species and also the Hydnophy- 
tineae). d s. str. is characterized by usually 

aving pyrenes with or without preformed germina- 
tion slits Pesa 2001), a 


furrowed 


plane or shallowly 
adaxial surface, and usually numerous 
distinct ridges on the abaxial = ne characters 
are discussed by Davis et al. e Pacific 
de e M" 


is characterize yrenes with 


e) 
-= 


marginal preformed germination slits. The main 
difference between the studies by Nepokroeff et al. 
(1999) and Andersson (2002a) is that Nepokroeff et 
al. included the Pacific clade in Psychotria s. str. 
while Andersson excluded it. 
Carapichea was reestablished as a genus 

Andersson seine for three piis of the d rique 
y based on rpsl6 dat of 


P cem P. jae aa Molina) C. M. pira 
E W. C. Burger (— Cephaelis affinis ar and 
P. ipecacuanha (Brot.) Stokes, 
by Nepokroeff et al. (1999) 
and sister to Geophila and Hymenocoleus; Andersson 
(2002b) found a third 
Rusby (described as Carapichea guianensis Aubl.), 
but 
These three taxa in- 


to be PN Ee 


species, P. guianensis 
that was distant from the Psychotria s. str. 
belonged to the 
cluded in the reestablished genus Carapichea were 
rled as a 
relationship within the Palicourea complex was 


same group. 


strongly  suppo group, but the exact 
unsupported. The genus was characterized “by having 
stipules that are not shed by formation of an 
abscission layer, leaves that dry greenish or greyish, 
aperturate pollen, and planoconvex pyrenes with an 
adaxial furrow and preformed germination slits on 
abaxial ridges, but not along the margins" (Andersson, 
2002b: 363) 


Phylogeny of the tribe Anthospermeae was estimat- 
ed based on ITS and rpsi6 data by Anderson et al. 
(2001). They first analyzed a set of taxa, including; 
Anthospermeae together with representatives of other 
Rubioideae tribes, to test if the tribe was monophy- 
letic. In a second analysis of 25 Anthospermeae taxa 
(all except two genera of the tribe), they investigated 
the internal relationships of the genera. Most genera 
of Anthospermeae formed a monophyletic but 


Annals of the 
Missouri Botanical Garden 


weakly supported clade, with Carpacoce Sond. 


excluded. The latter was instead sister to the 
Knoxieae. They found no support for a subdivision 
of the tribe into three subtribes and no support for a 
subdivision of Coprosma into two subgenera. They 
found support for a clade corresponding to Puff's 
(1982) subtribe Anthospermeae (Anthospermum L., 
Nenax, Galopina, and Phyllis with Carpacoce 
excluded) and moderate ug for he nie 
( LA Durringtonia R. J. F. d.& Pr 
Leptostigma, Nertera, and da Hoo 
with the jo nested within Coprosma), but [es 
DC. and Opercularia (Puff's subtribe Opercularinae) 
were placed unresolved in a trichotomy together 
with the Coprosm 
Thulin and Bremer (2004) studied parts of the tribe 
Spermacoceae s.l. to circumscribe the genera Am- 
phiasma Bremek. and Pentanopsis Rendle and to find 
the affinity of Phylohydrax Puff. They analyzed rbcL 
sequences of 34 tribal members and found that the 
African ae Amphiasma, Conostomium (Stapf) 
anostachya Bremek. together with 
Phylohydraa form a strongly supported clade distant 
from Hydrophylax L. f., which was placed close to 
When Phylohydrax was 
a new genus (Pu , it was 
om a differen stock than 
the genus Hydrophylax. This was also confirmed in a 
study by Thulin and Bremer (2004). Furthermore, 


Amphiasma was found to be paraphyletic and a new 


suggested to have evolved fro 


taxonomy was proposed. Pentanopsis was circum- 

scribed as a genus of two species from northeastern 

tropical Africa, whereas Amphiasma was treated in its 

original sense as a genus of about eight species in 
south-central tropical Africa. 

ES. dn alter Phylohydrax was positioned in the 

hiasm clade by Thulin and Bre- 

er iu Dessein et al. (2005) Pega a study of 

ker and Phylohydrax. They investi- 

gated e and compared it to results pus 


omphocalyx Ba 


molecular data (mainly sequences from GenBank). 
ey showed that there are many morphological 
similarities between the genera, and they concluded, 
based on the molecular results, “that the character 
states of the two genera are largely consistent with the 
here-proposed position in Hedyotideae" (Dessein et 
al., 2005: 91) 
The Andean genus Arcytophyllum Willd. ex ye 
& Schult. f. was investigated by ie and £ 
They bd 
(with A 


serpyllaceum (Schltdl. Terrell excluded, due to its 
closer relationship to Bouvardia) sister to a clade of 
American Hedyotis L. and Houstonia L. species. It is 
further suggested that these latter should be treated as 


a single genus, under the name of Houstonia. It was 
also suggested the ancestral area of the 
Arcytophyllum—Houstonia clade is the South Ameri- 

can plate. 
Houstonia, a North American genus, was investi- 
gated for nuclear (ITS) and e (trnL) 
urc e analyzed 
closely related genera (Carterella 
Terrell, Dentella H R. Forst. & G. Forst., Hedyotis, 
H 


Oldenlandia, Oldenlandiopsis Terrell & W. 
ewis, Stenaria (Raf.) Terrell), 30 taxa ip ids 
The phylogenetie results were compared to chromo- 


some numbers, breeding systems, id life forms. 
Houstonia was not monophyletic and could not be 
ept distinct from Stenaria and American 
Hedyotis. Within the North American lineage, it 
appeared that chromosomal changes have had an 
important role for history of diversification. The 
annual habit and a homostylous breeding system 
have originated several times and have probably not 
been major factors in the radiation of the s 
ater, Church and Taylor (2005) investigated a 
larger set of species and populations (74 populations 
from 17 species) of the Houstonia lineage for ITS, 
trnL, 
-a in the an 
de 


and irnS-G. dd found no evidence for 
estral species, but more 
d a wide degree of 


Peu ie and genetic variation both within and 


ently derived species containe 
among species. They found a clear association 
between hybridization and polyploidy in the Hous- 
tonia lineage, supporting the idea that polyploidy 
may break down species barriers and allow 
hybridization among lineages. 
aertnera of the tribe Gaertnereae is a Paleotropi- 
its highest 
diversity on Madagascar (25 species). The genus was 
investigated by Malcomber (2002; also Malcomber & 
Davis, 2005) 


Denda markers, and the genus was strongly 


cal genus of regional endemies with 


Malcomber (2002) used four usually 


supported as monophyletie. However, the genetic 
variation among species was insufficient to recon- 
struct well-supported subgeneric groups “counter to 
expectations based on the very distinct morphologies 
and widespread distribution of the genus” (Malcom- 
ber, 2002: 42). 

The tribe 


Paederieae w groups 
studied in a Ph 


as of the 
esis by “Backlund (2005). 
Earlier molecular analyses (Bremer, 1996; Anders- 


son & Rova, 1999) had indicated that the tribe 


could be polyphyletic, and Backlund ioni Pe 
investigated the tribe und 
= support hh Paederieae s. str. ratis 
Paederia, Leptodermis Wall., Serissa Comm. ex Juss., 


a Roxb.) and a reestablished tribe 


Putorieae 


Volume 96, Number 1 
2009 


Bremer 
Molecular Phylogenetic Studies of Rubiaceae 


SUBFAMILY IXOROIDEAE 


1996) investigated both 


Andreasen and Bremer ( 
morphological and molecular (rbcL) data o 


subfamily 
Ixoroideae s. str. They analyzed 40 ingroup taxa from 
Gardenieae (Gardenia Elis, Aidia Lour., Alibertia A. 
Rich. ex DC., Burchellia R. Br., 

Casasia A. Rich., Coddia Verde., 
Keay, Duperrea Pierre ex Pit., Fuclinia Salisb., Genipa 


Calochone Keay, 
Didymosalpinx 


L., Glossostipula Lorence, Heinsenia K. Schum., 


pees : Bridson, Kailarsenia 
irveng., o K P hum.) Hoyle, Mitriostigma 
Hochst., Oxyani DC., Oxyceros Lour., Porterandia 
Ridl., Randia i ee ee Fagerl., Roth- 


mannia Thunb., Sukunia A. Sm.) Pavetteae 
(Pavetta L. ees Welw. ex Hook. f., Leptactina 

ook. f., dea DC., Tarenna Gaertn.), Octotropi- 
deae ae Delile, 
Kraussia Harv., Paragenipa Baill., 
Ramosmania Tirveng. &  Verdc.), 
(Coffea, Diplospora DC., Paracoffea J.-F. Leroy, 
Psilanthus Hook. f., Tricalysia n Rich. ex DC.) with 
Mussaenda as outgroup. They found that Vanguerieae 


Fernelia Comm. ex Lam., 
Pouchetia DC., 
and Coffeeae 


(Canthium Lam., Vangueria Juss.) should be included 
in the subfamily. The Octotropideae, Pavetteae, and 
with different 


circumscriptions of the latter two compared to earlier 


Coffeeae were monophyletic although 


classifications. Ixora (together with Myonima Comm. 
. and Versteegia Valeton) was not part of 
Pavetteae, and Coffeeae should include Tricaly: 

ably Bowie era Aubl. as well. Subtribe 


ineae (Cremaspora Benth. and Tricalysia) and Poso- 


ex Juss 
sia and 
Diplospor- 


queria should be excluded from the tribe Gardenieae. 
Furthermore, they suggested that the Pur tetrad 
group within Gardenieae (Robbrecht & Puff, 1986) is 
not monophyletic and that the is of the 
pollen that is released in tetrads may have evolved 
several times. A few years later, Andreasen et al. 
ud analyzed and co 
nuclear ITS reg 


Sedis 


a 


mpared the utility of the 
ion with the cpDNA rbeL for the 
Variation. of ITS was extensive and 
informative, but the sequences were difficult to align. 
New phylogenetic positions of taxa (e.g., for Poso- 
queria, Bertiera, Ixora, and Vanguerieae) that had 
been reported from the rbcL analysis, but contradicted 
the A were corroborated by the ITS data. 

, Andreasen and Bremer (2000) presented 
E analyses of the subfamily based on 
combinations of rbcL, , and restriction fragment 
length polymorphism md data for 77 ingroup taxa. 
The results agreed with the 1996 and 1999 studies, 
but many groups received higher support. Further, 
Ma (Alberta E. Mey.) was shown to be part of 
the subfamily, and the mangrove genus Scyphiphora 
C. F. Gaertn. (Antirheoideae fide Robbrecht, 1988; or 


Gardenieae s.l. fide Puff & Rohrhofer, 1993) was 
shown to be close to Ixoreae. 

There is strong support for 12 of the 15 investigated 
tribes of this s 
eae and Retiniphylleae are monotypic or represented 


ubfamily as monophyletic (Cremaspor- 


by Dd taxa and could not be tested for monophyly), 
Gardenieae is polyphyletic/ 

the subfamily 

and the subgroup including Alberteae, Bertiereae, 


but arge tribe 
deus Despite strong support for 


Coffeeae, Cremasporeae, Gardenieae, Octotropideae, 
and Pavetteae, most relationships between tribes are 
unresolved and in need of further research. So far, five 
Ixoroideae tribes have been studied and are presented 
below, and several tribes are under investigation. The 
most important tasks for the future in this subfamily 
will be to investigate the large complex around the 
polyphyletic/paraphyletie Gardenieae and to investi- 
gate the difficult and large genera Ixora, Pavetta, and 
Tarenna 


Coffea of the tribe Coffeeae has been the focus of 
several eas studies (Lashermes et al., 1997; 
1998). T 

he 


tion to the cla 


Cros et al., e lead of Coffea was in 


n, particularly relative 


to the genus Psilanthus. However, there were 
correlations between clades and biogeography. It 
was also shown that Coffea uet a recent origin and 
MEINE on in África (Cros et al., 1998). 
Dialypetalanthus Kuhlm. ario tribal position) is 
an endemic Amazonian genus that has been treated as 
monotypic family Dialypetalanthaceae (Rizzini & 
Occhioni, 1949), but 


suggested, e.g., Myrtaceae and Rubiaceae (Kuhlmann, 


various affinities have been 


1925). It is an aberrant genus with free petals and an 
indefinite, extremely high number of stamens, char- 
acters that do not agree with Rubiaceae, but the genus 
shares many characteristics with taxa of Rubiaceae, 
e.g., opposite entire leaves with interpetiolar stipules, 
inferior 7 stigma, Re fruit, and 

d se l. (1997) presented 
ue and morphological data that support an 


ilo 
winge s. Breda et a 


x with E es, Rubiaceae in particular. 
ay published the first analysis of 
ae (ba b m in th 

belongs 


which they showed that the 


genus to Rubiaceae in the subfamily 
Ixoroideae s.l., but without tribal position. 
Persson 1996 started his studies of tribe Garden 
with analysis of 70 


dmi pud for 81 taxa. MOS nodes were 


rphological and 


ee 
g 


unresolved or unsupported, but he found support for 
several of Robbrecht's and Puffs (1986) informal 
groups of the Gardenieae (tetrad group and Alibertia 
group, but Aidia group and Gardenieae were not 
supported). Later, Persson (2000a) continued his 
rpsl6 and trnL-F data 


Gardenieae s.l. to try to resolve the more or less 


study o taxa of 


Annals of the 
Missouri Botanical Garden 


unresolved phylogeny of the group; he also wanted to 
evaluate the conflicts ween his morphological 
study (Persson, 1996) and the results from the rbe 
data (Andreasen & Bremer, 1996). Persson's molec- 
ular tree (2000a) was still unresolved, with few 
supported groups. However, the informal Alibertia 
group (in the study including Alibertia, Amaioua 
Aubl., Borojoa Cuatrec., Duroia L. f., Glossostipula, 
Ibetralia Bremek., Kutchubaea Fisch. ex DC., Mela- 
nopsidium Colla, and Stachyarrhena Hook. f.) was well 
supported (97% bootstrap} and agreed with earlier 
results (Andreasen & Bremer, 1996; Persson, 1996; 
1997). He further identified a core 
& K. Krause, 
If, Deccania Tir- 

veng., Morelia A. Rich. ex DC., Sherbournia G. Don, 
Tamilnadia us & Pus Trukia Kaneh., and 
noidea T . & Sas mong others, but 
deum subtribe | o “ Burchellia, Didy- 


Andreasen, 
Gardenieae group (Atractocarpus Schltr. 
Benkara Ada Catunaregam Wo 


mosalpinx, Schumanniophyton Harms, and several 
taxa belonging to other Ixoroideae tribes) with two 


hand, there was no support for an Aidia 
group or for a monophyletic tetrad group (Robbrecht & 
Puff, 1986), both proposed from morphological data 
(Persson, 
Perss 


n 


as further concluded from 
n's molecular data that the pollen release in 
tetrads had originated several times. It occurs in the 
large genus Gardenia, but not in its close relatives 
Aoranthe Somers, Ceriscoides (Hook. f.) Tirveng., 
Genipa, and Kailarsenia (a clade with 83% bootstrap 
support); most genera with tetrad pollen occur in a 
clade of Neotropical genera around Randia in which 
several genera also have monad pollen, e.g., Rosen- 
bergiodendron, Sphinctanthus Benth., and Tocoyena 
Aubl. Furthermore, outside the core Gardenieae there 
was also a clade of the genera Atractogyne Pierre, 
Mitriostigma, and Oxyanthus (86% bootstrap support) 
with tetrad pollen. 

Persson later (2000b) extended his study of the 
Alibertia group (Gardenieae), the group of taxa that 
th hetero- 
" (Persson, 
2000b: 1018). He sequenced two nuclear spacers (ITS 
and 5S-NTS) for 38 species (of the ca. 120) and found 
clades i 


several stron in group. 
r, Borojoa was paraphyletic and nested within 


gly supported 
ps (in a group close to the type species A. edulis 
A. Rich. ex DC.) with Borojoa included and A. 
hispida Dueke excluded. Aliber 


and distinctly divided into two main clades, one 


tia was monophyletic 


including the type species and one around A. sessilis 
(Vell.) K. Schum. In the combined analysis, Alibertia 
of Duroia, with t 
Amaioua nested within Duroia. Ibetralia, Kutchubaea, 


was sister to a clade e genus 


and A. hispida formed a well-supported clade at the 
unresolved base of the tree together with the rest of 


andia, a genus of ca. 90 Neotropical species, was 
investigated by Gustafsson and Persson (2002). They 
studied 38 taxa of the genus together with d 
tatives of eight other Gardenieae g 
molecular (ITS and 5S-NTS) a 
The molecular data do not support a monophyletic 
er but with ig arcing data added, Randia, 

ogether with C 
mus than 
Randia-Casasia group is an African clade (Calo- 


nera nalyz 
nd Eod D 


sia, akly supported 
50%) wee nies group. Basal to the 


chone, Macrosphyra Hook. f., Oligocodon Keay, 


(Rosenbergiodendron, Sphinctanthus, Tocoyena). With- 
in the Randia group, there are three geographically 
distinct clades: an Andean clade (less than 50% 
support), Central American Randia (58%), and South 
American Randia (85%). 

The first attempt to construct a molecular phylog- 
eny of the morphologically distinct P Vanguerieae 
hed by Lantz et e (20 

ITS f 


was publis 
gated the nuclear spac or 41 Vanguerieae 


. They investi- 
a V i 
species representing 19 ; genera. The taxa fall into 
several well-supported clades, of which they dis- 
cussed three informal groups: spiny group (Canthium, 
Meyna Roxb. ex Link), gueria group, d 
Fadogia-Rytigynia group. Based on the investigated 
taxa, Keetia E. Phillips, Lagynias E. Mey. ex Robyns, 
Multidentia Gilli, 


were monophyletic units, but Canthium, Fadogia 


and Pyrostria Comm. ex Juss. 
Schweinf, Rytigynia Blume, Tapiphyllum Robyns, 
and Vangueria were found to be polyphyletic or 
paraphyletic. The analysis clearly demonstrated that 

several genera are in need of new circumscriptions. 
Later, Lantz and Bremer (2004) analyzed data for 69 
ingroup taxa representing 23 of the 27 genera of the 
tribe (ITS, trn 


strong support for many groups, but these rarely 


T-F, and morphology). They found 


coincided with traditional genera in accordance with 


their earlier study (Lantz et al., 3 the 
gyni 


investigated taxa, Keetia, Lagynias, and Multidentia 
were bos with strong support and Psydrax 

aertn. was monophyletic with. weak support. 
Canthium su ae anthiu on was given 


roc m 
generic status as Afrocanthium (Bridson) Lantz & B. 
Bremer, and also new combinations were made for 
Canthium s. str. Another identified, well-supported 
clade was the dioecious group, including Pyrostria 
and Cyclophyllum Hook. and several genera 
restricted to Madagascar (Leroya Cavaco, Neoleroya 
Cavaco, Peponidium (Baill.) Arénes, Pseudopeponi- 
ia Bridson an: 


dium Arénes), Canthium subg. Bulloc 


Seyphochlamys Balf. f. The relationships between the 


Volume 96, Number 1 
2009 


Bremer 
Molecular Phylogenetic Studies of Rubiaceae 


taxa are not well understood and are in need of more 
tudy. The earlier proposed spiny group (Lantz et al., 
2002) identified by supra-axillary spines was one 
identical to Canthium s. str., and the large-flowered 
group including Vangueria group and Fadogia— 
Rytigynia group were further investigated i in a later 
study (Lantz & Bremer, 2005) 


estimated ca. 180 species were analyzed for the 


ixty-six of the 


nuclear ren the chloroplast markers trn7-F and 


data were analyzed in combination and 


p e 
Lr Several taxa (Ancylanthos rubiginosus 
Desf., Hutchinsonia barbata Robyns, R. beniensis 


(De Wild.) Robyns, R. decussata (K. Schum. 


explanation m the incongruence. These 
excluded from the taxonomic discussions. 
gia—Rytigynia 
ported as monophyletic entities. Most of the taxa of 


Vangueria and Fado, groups were sup- 
the Vangueria group were merged into Vangueria (the 
genera Ancylanthos Desf., Lagynias, Pachystigma 
Hochst., Tapiphyllum, and a few investigated species 
of Fadogia and Rytigynia). The genus is characterized 
in the tribe by domatia rarely present, inflorescences 
hich the 1 


fallen, smooth retrorse hairs in the corolla, and large 


usually borne at nodes from w eaves have 
fruits (more than 1 em long) with three to five locules. 
The relationships within the Fadogia and Rytigynia 
group could not be resolved and are in need of further 
study. However, the whole group could be distin- 
guished from the Vanguería group by presence of 
domatia and a calyx with or without poorly developed 
calyx lobes (with exceptions) 

Taxa of the tribes Mussaendeae, Isertieae (see 
under Cinchonoideae), and Sabiceeae have been 
understood as a complex even before molecular data 
came into use, but are treated differently by different 
authors (e.g., Bremekamp, 1966; Robbrecht, 1988; 
Andersson, 1996). In a study of rbcL data from 
Cinchonoideae and Ixoroideae taxa by Bremer and 
Thulin (1 


close to Cinchoneae of the subfamily Cinchonoideae; 


1998), Isertieae was found to be a small tribe 


however, Sabiceeae and Mussaendeae are two tribes 
that belong to subfamily Ixoroideae. A new aberrant 


tra, Tamridaea Thulin & B 


Bremer, was shown to be a sister genus to Virectaria 


endemic genus from 
Bremek. and placed in Sabiceeae together with Sabicea 
Aubl. and é 
deae was reestablished. 
Miq., Heinsia DC., 
were included as the component gen 

(2005) later td tribe 
Mussaendeae and tested the monophy 


Pseudosabicea N. Hallé. The tribe Mussaen- 
and Mussaenda, Aphaenandra 


and Pseudomussaenda Wernham 


Alejandro et al. 
y of the genus 
Mussaenda and the circumscription of Mussaendeae 


sensu Bremer and Thulin (1998; see under Isertieae— 
Cinchonoideae). Alejandro et al. included 25 species 
of Mussaenda and representatives of all 

genera of the tribe, except for Neomussaenda Tange, 
plus outgroups (the genus Mussaendopsis Baill. was 
also included, which was shown to belong to the 
Condamineeae clade). They analyzed trnT-F and ITS 
data and demonstrated that the tribe Mussaendeae 
S Mussaenda, Aphaenandra, Bremeria Raza- 
fim. & Alejandro, Heinsia, Landiopsis Bosser, Pseu- 
domussaenda, and Schizomussaenda H. L. Li) is 
monophyletic, but the genus Mussaenda s.l. is 
polyphyletic. The Malagasy species were found to be 
more bu related to e than to the African 
and Asian Mussaenda. They described a new genus 
Deer to accommodate 19 Indian Ocean species. 
The recircumscribed Mussaenda is characterized by 
reduplicate valvate aestivation and glabrous styles, in 


contrast to the reduplicate and induplicate aestivation 
and densely pubescent styles in Bremeria. 


(2001) published a study of 
Hekistocarpa Hook. f. and showed that it belongs in 


Dessein at al. 
the vicinity of Virectaria. They also performed 
jackknife analyses of two molecular data sets, one of 
6 (mainly sequences from 


~ 
> 


cL and one of rpsl 
GenBank). Their 


analysis and the mor 


conclusions from molecular 

phological investigation were 
that the emended tribe Sabiceeae of Bremer and 
Thulin (1998) could not be morphologically charac- 
terized and is better treated as two distinct tribes: (1) 
Sabiceeae (Sabicea and Pseudosabicea and also, 
although not included in the analyses, Ecpoma K. 
Schum., Pentaloncha Hook 
Beauv.); and (2) Virectarieae (including Virectaria, 


— 


f, and Stipularia P. 


Hekistocarpa, and Tamridaea). In a sense, the 
Sabiceeae is characterized by entire stipules, medium 
to large flowers, valvate aestivation, berries, and small 
angular seeds with thickened radial walls. According 
to Dessein et al. (2001: 75), it is more difficult “to 
diagnose the tribe Virecatrieae emended to include 
Hekistocarpa and Tamridaea.” 

Stimulated by the from Rova et al. (2002), 
Delprete and Cort 04) carried out a more 
detailed molecular ae nl and ITS) of tribe 
Sipaneeae with Platycarpum Humb. & Bonpl. as the 
outgroup and evaluated relationships and delimitations 
of genera. They confirmed that the tribe is monophyletic 
and belongs within the Ixoroideae. In the tribe, they 
included Sipanea, Chalepophyllum Hook. f., Dendrosi- 
panea Ducke, Limnosipanea Hook. f, Maguire- 
othamnus, Neobertiera, and Sipaneopsis Steyerm. All 
genera investigated were found to be monophyletic. It 
was inferred that the herbaceous habit of Sipanea and 
Limnosipanea had evolved twice in the tribe as these 
two genera are not sister groups. Delprete and Cortés-B. 


Annals of the 
Missouri Botanical Garden 


(2004) had no material of Neblinathamnus Steyerm., 
Steyermarkia Standl., but, 

morphological MUN they tentatively 
‘ainda e in the Sipan 


eee Wernham, and 


SUBFAMILY CINCHONOIDEAE 


o study has focused explicitly on the entire 
subfamily Cinchonoideae, but several studies on the 
whole family (Bremer et al., 1995; Rova et al., 2002) 
1998; 
Razafimandimbison & Bremer, 2001; Andersson & 
Antonelli, 2005) have contributed to the knowledge of 
the subfamily. Based on these studies, there is support 


or of specific groups (Bremer & Thulin, 


for nine tribes: Cinchoneae, Chiococceae, Guettar- 


Hillieae, 


tieae, Naucleeae, and Rondeletieae, six of which are 


deae, Hamelieae, Hymenodictyeae, Iser- 
discussed below. The relationships between the tribes 
in this subfamily are very poorly understood, except 
for a few sister group relations between Guettardeae 
and Rondeletieae, Hamelieae and Hillieae, an 
Hymenodictyeae and Naucleeae, respectively. Most 
species of Rondeletia, the largest genus of this 
subfamily, have not been investigated so far. It would 
be interesting to analyze all species in this mainly 
South American subfamily, particularly because there 
are several interesting biogeographic hoc of 
relations between South America a and t d World 
tropics, the Pacific, and the 

Early molecular data (Bremer j Jansen, 1991) 


indicated the tribe Chiococceae (Antirheoideae fide 


Bremer (1992) analyzed 20 morphological characters 
for 22 genera of Chiococceae and the Portlandia P. 
Browne group, and, as a result, the tribe Chiococceae 
was emended to include also subtribe Portlandiinae 
(Condamineeae) and some taxa of Cinchoneae, as 
there was no resolution or support for two distinct 
clades corresponding to Chiocceae s. str. and a 
Portlandia group. 

study, Delprete (1996) reexamined the 
circumscription of the Condamineeae, Chiococceae, 
1996: 165), with the 


purpose “to test the tribal redefinition of Chiococceae 


In his 
and Catesbaeeae (Delprete, 
proposed by Bremer (1992).” He analyzed 170 species 
of 44 genera for 44 morphological characters. His 
conclusion was that the Portlandia group (former 
Condamineeae) is closer to the Chiococceae s. str. (as 
suggested by Bremer, 1992) than to the rest of the 
Condamineeae. Because Chiococceae s. str. was 
monophyletic without the Portlandia group, he 
retained Chiococceae as a restricted tribe and instead 
included the Portlandia group in the tribe Cates- 
the rest of the 


baeeae. Therefore, Condamineeae 


(Condamineinae and Pinckneyinae) was merged with 
the Rondeletieae s. 

In several later MT studies, the Er ud 
by Delprete (1996) w 
contradicted, and it has instead been shown that an 


tion of the two tribes 
taxa are intermixed in one group approximately 
corresponding to an emen ise (Bremer 
et al., 1995; mie sson & p 1999; Rova et al., 
2002). Motley et al. (2005) reas most of the 
genera from the Catesbaeeae—Chiococceae complex to 
reevaluate the generic rd They found 
strong support for a group with Str a no as 
sister to the complex, but there was no support to 
separate the taxa into two clades or tribes. They found 
Catesbaea L., Erithalis, Hintonia, Isidorea A. Rich. ex 
DC., Phialanthus, Portlandia, and Scolosanthus Vahl 


Chiococca, Exostema 
paraphyletic/polyphyletic, and for several taxa, mono- 
phyly could not be tested (monotypic genera or single 
species investigated; ly ae Hook. f., Badusa A. 
Gray, Ceratopyxis Hook. f., Coutaportla, Coutarea, 
Cubanola Aiello, ae Vieill., Osa Aiello, 
Phyllacanthus, Salzmannia DC., Schmidtottia, and 
Siemensia Urb.). 

Exostema, a genus of 25 species that occurs from 
Bolivia to Mexico throughout the West Indies, 
represents one of the first molecular analyses of a 
genus within Rubiaceae. McDowell and Bremer 
(1998) investigated all species for 37 morphological 
characters an sequences of 18 species. All data 
sets (morphology, molecular, and combined) resolved 
three main species groups corresponding to sections 

combined trees place 
American species (E. corymbosum (Ruiz & Pav.) 
Spreng. and E. maynense Poepp. € Endl.) basal to the 
three retrieved clades. The genus was later reinves- 
tigated by McDowell et al. (2003), e used rbcL, ITS, 
and combined data in order to understand the 
biogeographic pattern of the genus in the Caribbean 
region. n l4 Exostema 
species and nine species from eight ine genera. 


The analyses were based o 


The data did not support Exostema as monophyletic. 
In the ITS analysis, which showed the best resolved 
trees, Coutarea, Chiococca, and Erithalis were nested 
within Exostema, making Exostema highly polyphy- 
letic or paraphyletic. Coutarea (from South or Central 
America) was placed close to the two South American 
species of Exostema (E. corymbosum and E. 
nense). 

Erithalis is an endemic Caribbean genus, the 
phylogeny and biogeography of which were studied 
by Negrón-Ortiz and Watson (2002). They investigat- 
ed seven of the eight to 10 species with two nuclear 


Volume 96, Number 1 


Bremer 
Molecular Phylogenetic Studies of Rubiaceae 


markers, ITS and ETS. They found the genus to be 
monophyletie relative to the genus Chiococca and 
Exostema longiflorum (Lamb.) Roem. «€ Schult. 
Surprisingly, there was no support for monophyly for 
any of those species (Erithalis fruticosa L., E. 
salmeoides Correll, E. odorifera Jacq.) that were 
sampled from more than one specimen. Due to low 
variation in the molecular markers, they hypothesized 
that the genus radiated rapidly within the Caribbean 
islands and that an initial colonization may have been 
from Central Am 

Tribe Ciichuness aud the complex around this 
tribe were first analyzed with morphological charac- 
ters by Andersson and Persson (1991) and Andersson 
(1995). They found the tribes Cinchoneae, Hillieae, 
and Calycophylleae to be monophyletie, and they 
of these 


proposed new circumscriptions tribes. 


Ro ). More recently, Andersson and 
Antonelli (2005) reinvestigated the relationships of 
inchoneae, making a thorough analysis based on 
five molecular markers for 51 Rubiaceae taxa sampled 
from the Cinchoneae and closely allied tribes 
(Chiococceae, Guettardeae [Guettarda], Hamelieae, 
Hillieae [Cosmibuena Ruiz & Pav.], Isertieae [/sertia 
Schreb., Kerianthera J. H. Kirkbr.], 


Rondeletieae) as well as other representatives of the 


Naucleeae, 


family. They found the tribe to be strongly support 
hylet 


as monophyletie including the monophyletie genera 
Cinchona L., Cinchonopsis L. Andersson (monotypic), 
Joosia H. Karst, Ladenbergia Klotzsch, Remijia DC., 
and DE pat ook. f. The monotypic Pimente- 
lia as not investigated, but due to morpho- 
a pees it was suggested to be close to 
Stilpnophyllum. Further, Antonelli (in Andersson & 
Antonelli, 2005 cribed a new genus Ciliosemina 
Antonelli within the tribe, 


(former species of m cee ce 


including two species 


characterized by “long-pedunculate, corym| o 


A 


subcorymbose inflorescences (fig. 3A), and the ‘ciliate 
to fimbriate wing margins of its seeds” (Andersson & 
Antonelli, 2005: 26) 

Tribe Isertieae was first analyzed by Andersson 
(1996) with e data. He a all 
except one of the Isertieae genera enumerated by 
Robbrecht (1988), 26 genera total with ae 
tives of other tribes. The analyses resulted in a 
new circumscription of the tribe including only 
seven genera: /sertia (including Yutajea Steyerm.), 
Aphaenandra, Heinsia, Mussaenda, Neomussaenda, 
Pseudomussaenda, and Schizomussaenda. Andersson 
recircumscribed tribe Sabiceeae to include Sabicea, 


Acranthera Arn. ex Meisn., Amphidasya, Ecpoma, 


Pentaloncha, Pittierothamnus Steyerm., Pseudosabi- 
cea, Schizostigma Arn. ex Meisn., and Temnopteryx 
Hook. f. 

data 
scriptions of Isertieae, tested the phylogeny present- 
ed by Andersson (1996), and also pinpointed the 
position of an aberrant endemic species from Socotra. 
Bremer and Thulin (1998) e rbcL for 


Cinchonoideae and Ixoroideae taxa plus seven out- 


Molecular showed contradicting circum- 


a Their conclusion was that Isertieae belongs 
the Cinchonoideae but should be restricted to 
ae (including Yutajea) and Kerianthera, and that 


Sabiceeae and Mussaendeae instead belong to 


Ixoroideae. 
Naucleeae s.l. was investigated by Razafi- 
Pa and Bremer (2001, 2002). They 


investigated molecular (ITS, rbcL, trnT-F) and mor- 
D characters for a total of ca. 50 taxa of the 
esented most 


tribe in the different o that repre 


genera. vie roader circumscription of 
the tribes, ellas not only Naucleeae sensu 
Ridsdale but also Cephalanthus L. (of Antirheoideae 
fide Robbrecht, 1988) and Mitragyna Korth. an 
Uncaria Schreb. (of Cinchoneae fide Robbrecht, 
1988), belong to the group. They also showed that 
Coptosapelteae sensu Andersson and Persson (1991) is 
paraphyletic. Twenty-four genera were accepted in 
Naucleeae, which was divided into six highly supported 
and morphologically distinct subtribes (Breoniinae: 
~ Ridsdale, Breonia A. Rich. 

Gyrostipula J.-F. Leroy. J.-F. Leroy; epa: 
rcm Coda Corynantheinae: Corynanthe 


, Janotia 


Welw., sn RE Pierre ex Beille, Pseudocinchona 
A. Chev. ex Perrot; Naucleinae: Nauclea L., Burttdavya 
Hoyle, pou Ridsdale & Bakh. f., Neola- 

marckia Bosser, Sarcocephalus Afzel. 
Mitragyninae: Mitragyna; and Uncarinae: 
and one paraphyletic or poorly sup 
Adininae (Adina Salisb., Adinauclea Ridsdale, Haldina 
Ridsdale, Ludekia Ridsdale, Metadina Bakh. f., Myrme- 
conauclea Merr., Neonauclea Merr., Pertusadina Rids- 
dale, Sinoadina Ridsdale). The Neonauclea clade, part 
of the subtribe Adinae, with many myrmecophytic taxa 
(see below) was further investigated in a study by 
Razafimandimbison et al. (2005). They analyzed ITS 
and ETS and found 
and supported; Ludekia is sister to the two monophy- 


the Neonauclea clade well resolved 
letie genera Myrmeconauclea and Neonauclea (the 
latter were earlier suggested to be paraphyletic; 
Razafimandimbison & Bremer, 2002). 

A new tribe Hymenodictyeae, sister group to the 
Naucleeae, was described for Hymenodictyon Wall. 
and Paracorynanthe Capuron (Razafimandimbison & 
Bremer, 2001). The 
species) and Hymenodictyon (22 species) are distrib- 


two genera Paracorynanthe (two 


20 


Annals of the 
Missouri Botanical Garden 


uted in Madagascar, and in Madagascar, mainland 
Africa, and tropical Asia, respectively. The sister 
group relationship to Naucleeae is highly supported 
(Razafimandimbison & Bremer, 2001) 

Neolaugeria Nicolson of the tribe Guettardeae, 
endemic to the West Indies, was studied by Moynihan 
and Watson (2001). Their data supported the genus of 
three species as monophyletic, but it was found to be 
only distantly related to Stenostomum C. F. Gaertn., a 
genus with which Neolaugeria sometimes has been 
merged as a section. Instead, it was closer to Timonius 
DC., although the support was very low. Moynihan and 
Watson (2001) also tested an earlier hypothesis 
regarding the origin of the genus in the Lesser 
Antilles by comparing vicariance with long-distance 
dispersal. The conclusion, albeit also with low 
support, was that N. resinosa (Vahl) Nicolson may 
occupy a basal phylogenetie position, supporting a 
pattern of speciation and colonization in a northwest- 
erly direction from Lesser Antilles to the Greater 
Antilles and the Bahamas. 


APPLIED SrUDIES BASED ON RUBIACEAE PHYLOGENIES 


The power of a phylogenetic tree is not only that it 
can be used for classification and systematics, but that 
it can be used for studies of diversity, anatomy, 
morphology, biogeography, ecology, etc., in whic 
evolution of taxa, genes, and characters can be used in 
a comparative context. With this species-rich and 
diverse family and with more and better phylogenetic 
trees from the family, we can probably foresee a strong 
increase in studies based on phylogenetic trees. So 
far, we have only seen a limited number of such 
studies, with interesting evolutionary questions being 
addresse 


PHYLOGENETIC TREES FOR ECOLOGICAL, EVOLUTIONARY, OR 
BIOGEOGRAPHICAL QUESTIONS 


In 1991 and 1992, the first da ecology 
papers were published (Eriksson € Bremer, 1991; 
Bremer & Eriksson, 1992) in 2 which a Rubiaceae 
phylogeny was used. These studies addressed hypoth- 
eses about evolution of fruit traits, animal versus 
abiotic modes of dispersal, life forms, and species 
richness. It was shown that fleshy fruits have evolved 
several times and that in many lineages the animal- 
dispersed fruits (drupes and berries) have remained 
argely unaltered since the time of origin. This is in 
contrast to the evolution of lineages with wind- 
dispersed seeds in capsules, or with pterophylls 
promoting wind dispersal of fruits, where traits have 
shifted more frequently during evolution. Animal 
dispersal was widespread among shrubs, whereas 


abiotic dispersal was most prevalent among herbs. 
Drupes were common in transoceanic taxa and on 
islands, indicating dispersal over long distances, 
probably by birds, but no evidence supported the 
view that animal dispersal in general enhances 
long-distance dispersal. No single trait explained 
variation in species richness. Instead, certain 
combinations of dispersal mode or life forms were 
shown to be associated with species richness. Genera 
with herbs and with abiotic dispersal, or with shrubs 
and with animal dispersal, or with shrubs and trees 
with winged seeds were all characterized by large 


species numbers, a result that implies association 


jej 


een seed dispersability and rate of species 
diversification. 
High host specificity of herbivorous insects and 
of diversity have been much 
n, 1982; Stork, 1993; Odegaard, 


compared a plant 


global estimates 
discussed (cf. Erwi 
2 Novotny et al. 
phylogeny of 51 tree species, including Rubiaceae, 
from New Guinea with more than 900 leaf-chewing 
d on these plants. Compared to earlier, 
i studies, they fou w hos 
specificity of the tropical herbivorous insects, and, 


insects foun 

more theoretical 

as a consequence, a global estimate on arthropod 

diversity was reduced from 31 million to 4 to 6 million 
cies 


un 
no] 


Razafimandimbison et al. (2004) identified high 
polymorphism of the ITS region in three Pons 
species (Adinauclea fagifolia (Teijs 
Havil.) Ridsdale, Haldina cordifolia Roxb. ) Ri 
and Mitragyna rubrostipulata (K. Schum.) Havil.). 
They found both intra-individual and een 
olymorphism in the three species, but no 
variation in the other 22 reten species of ile 
same tribe. Most of the variants were putative 
n They explored the potential utility of 
eudogenes in a phylogenetic analysis and found that 
m polymorphism does not transcend species bound- 
aries in this group (all variants within a species come 
together in the tree), so any of the pseudogenes could 
e of use in a phylogenetie analysis without con- 
tradicting the phylogenetic signal. 

McDowell and Bremer (1996) used a tree of 
Exostema (see above under Exostema) to optimize 
and investigate major trends in morphological diversi- 
fication of the genus, e.g., attributes for specializations to 
a xeric environment and for pollination biology. 
Xeromorphic traits had evolved in all three lineages, 
of vegetative characters, an 


e.g, reduction also 


reduction of reproduction traits such as send size and 
eed numbers. In the genus, two different major 
elitm syndromes occur, a long-flowered mot 
(Lepidoptera) type and a short-flowered bee pollination 


type. According to the analyses, both of these pollination 


Volume 96, Number 1 
2009 


Bremer 
Molecular Phylogenetic Studies of Rubiaceae 


types (with characteristic flower lengths, flower num- 

bers, and corolla color) have evolved more than once. 
Evolution of myrmecophytism 

study by Raza 

biologically interesting ant-plant association occurs 


was investigated in a 
al. (2005 


[zi 


imandimbison et . Thi 


in 22 genera and ca. 140 species of Rubiaceae, most 
of these in Southeast 


Asia, aua in the Malesian 
region. Razafimandimbison et al. 


investigated the 
Neonauclea clade of ou o including 25 taxa 
molecular 
phylogeny, they concluded that multiple origins of 


with myrmecophytism. Based 


myrmecophytism occurred in Borneo and that the low 
level of genetic variation indicates a rapid radiation in 
the Neonauclea (65 


Myrmeconauclea (3 species) was explained by the 


species); low radiation in 
different fruit and seed types and the ability to 
colonize different habitats. 

In their study of the 
complex, Motley et al. (20 

and fruit evolution and dis 
isle for the disjunction rw le Caribbean 


Catesbaeeae-Chiococceae 
m reconstructed flow 
eir 


and Pacific genera. According to their optimization on 
the tree, the ancestral fruit type for the group seems to 
be capsular; drupaceous fruits seem to have evolved 
twice and baccate fruits once or twice. The three types 
more or less n to d d 
oths and but 
Chiococca type by bees, and Portlandia type 5 md 


of fl 
pollinators: Exostema type by 


and bats. All types have evolved three or more times. 
Motley et al. (2005) also concluded that fleshy fruits 

v n very successful in dispersing between the 
Caribbean islands, and wind-dispersed seeds of the 
capsular-fruited taxa have been more successful for 
long-distance dispersal over the Pacific Ocean. 

The first biogeographic analysis of the family was 
based on a phylogeny of Anthospermeae (Anderson et 
al., 2001). The biogeographic implications were that 


the ancestral area of the tribe is Africa (including 
Madagascar) and that the genus spread by lon 


distance dispersal to northeastern Antarctica. It was 
also suggested that the occurrences in America, 
Hawaii, and Tristan da Cunha are due to long- 
distance dispersal. 

Other publications that discuss the biogeography of 
Africa are Malcolmer’s (2002) Gaertnera study and 
Alej et al/s (2005) study of Mussaenda. 
Malcolmer (2002) proposed that Gaertnera migrated 
to Africa during the early Tertiary, possibly via a 
boreotropical land bridge, and he further suggested 
that the genus started to radiate about 5.2 million 
years ago (Ma). The range of distribution is explained 
by a number of long-distance dispersal events. The 
molecular clock estimate gave a rapid diversification 


rate of 0.717 to O 


comparable to estimates of radiation on Oceanic 


.832 species/million years, which is 


(2005) concluded that 


as an African origin and that the 


islands. Alejandro et al. 
Mussaenda s. str. 
Asian Mussaenda species descended from an African 
species that migrated to Asia, where the major 
radiation has occurred (now 97 of 132 species). 
Despite the close phylogenetic relationship between 
the African i 


on both continents. 


and Asian clades, not one species occurs 
One of the most widespread 
has 


reached the Comoro Islands, Madagascar, and Mas- 


African Mussaenda species, M. arcuata Poir., 


carenes as suggested probably via stepping-stone 
dispersal. 

Nepokroeff at al. (2003) investigated the phylogeny 
and biogeography of the Hawaiian species of 
Psychotria to reconstruct the ancestral pattern of 
colonization and dispersal. Both parsimony and 


likelihood analysis gave highly co ngruent results, 


Hawaiian species. The analysis strongly supported 
the Hawaiian taxa as monophyletic and descended 
from a single introduction to the islands. The genus 
Kelloggia, with disjunct distribution in western North 
America and the western part of eastern Asia, was 
investigated by Nie et al. (2005), who found that the 
two species diverged from each other about 5.4 Ma; 
dispersal-vicariance analysis (DIVA) suggested an 
Asian origin of Kelloggia. Nie et al. (2005) further 
suggested that the disjunct distribution is a result of 
long-distance dispersal from Asia into western North 
America 


From the Caribbean region, Negrón-Ortiz and 
Watson (2003) used the phylogenies of the two 
ndemie genera Erithalis (Negrón-Ortiz € Watson, 
2002) and Ernodea (Negrón-Ortiz & Watson, unpubl. 
data) in a biogeographic study using Brooks Parsimo- 
ny Analysis (BPA) and Fitch parsimony methods. 
raphic association between Cuba 
, but the 
Hispaniola (Dominican Republig and Haiti) were 


They found a biogeo: 
and the Dominican Republic two countries of 
found in two places in the cladogram, suggesting 
— of geologic areas. The 
ter An tilles 
origin for Erithalis, in contrast to the Negrón-Ortiz and 


Hispaniola to be a 
Fitch analyses also supported a Great 


hes (2002) article in which they suggest coloni- 
zation of the genus from Central America. The present 
ino of the t 
product of dispersal for Ernodea and by a combination 


o genera was explained as a 


of vicariance and dispersal events for Erithalis. The 
mainly Caribbean genus Exostema (McDowell et al., 
2003) has also been analyzed biogeographically, but 
its distribution pattern was found to be far more 
complex than anticipated and no clear conclusions 
could be drawn except for a close affinity between the 
Cuban and Hispaniolan groups. 


22 


Annals of the 
Missouri Botanical Garden 


UNDERSTANDING DISTRIBUTION AND EVOLUTION OF 
MORPHOLOGICAL, ANATOMICAL, AND CHEMICAL CHARACTERS 
THROUGH PHYLOGENETIC TREES 


Molecular phylogenies have also been very useful 
for understanding morphological, anatomical, or 
chemical traits in various parts of the family. Jansen 
et al. (2001) performed a large survey of anatomical 
characters of woody Rubioideae taxa and compared 
the characters with recent phylogenetic insights in the 
Po group on the basis of molecular data. The idea of 
molecular 
stated 
explicitly. Jansen et al. (2001) presented anatomical 


study was based on results fro 
derit analysis, even if this was not 
data in illustrations and in a table of 26 different 
characters for 23 genera (and ca. 70 species) 
representing woody taxa of Coccocypseleae, Coussar- 
eeae, Lasianthus group, Morindeae s.l, Pauri- 
diantheae, Trianolepideae, and Urophylleae. It would 
have been even more Ware with a aa 


orpholo 


morphological-molecular aon 


data or a com 
but their snp 
nevertheless seem to be in agreement with most 


analysis of the m 


phylogenetic hypotheses presented from molecular 
data. S 


survey of wood anatomy of the whole family. 


oon thereafter, Jansen et al. (2002) presented a 


sa 


optimized the characters on a hypothetical supertree 
and found that the wood characters agreed with the 
phylogeny. Furthermore, they found that fiber types 
and axial parenc 
indeed had g 
they id 


yma distribution, for example, 


ood taxonomie values in the family, but 


that wood MuR data i 


5 


Rubiaceae is more useful in confirming or negating 
already proposed relationships rather than postulating 
new affinities for problematic taxa (Jansen et al., 
2002). 

Pollen morphology was investigated in 29 species of 
northwestern European representatives of Rubieae 
Crucianella, Cruciata, Galiu 
Sherardia) by Huysmans et al. (2003). 


(Rubia, Asperula, 
They a 
the combination of pollen characteristics to be unique 
within the family: several colpate apertures, 

perforate and microechinate tectum, a relatively small 
size, absence of endoapertures, a coarse nexine area 
beneath the ectocolpi, and absence of orbicules. The 
tribe Gardenieae also lacks orbicules (Huysmans et 
al, 1998, 2000). Huysmans et al. (2003) further 
optimized presence and absence of endoapertures on a 
i r and Manen (2000) and 
showed that only the Paederieae/Theligoneae/Rubieae 


Rubioideae tree from Breme 
totally lack the endoapertures, while the character is 


2003) a nd concentration of 
several metals in Rubiaceae. The most characteristic 


pattern was for aluminium, and pus was also a 


correlation with occurrence of silicon but not with any 
other metals. The aluminium accumulation was 
optimized on a molecular phylogenetie tree, and it 
was most characteristic of Rubioideae but occurs also 
in Coptosapelta and is partly present in taxa of 
Vanguerieae and Alberteae. 
re are a few examples of surveys of various 
traits from the family, chemical and morphological 
data, in which no tree approach has been used but for 
which analyses in relation to a phylogenetic tree 
would be very interesting. At the first Rubiaceae 
conference, Kiehn presented (1995) a survey of 
chromosome numbers of the family. Although he did 
not optimize his characters on a molecular phylogeny, 
many interesting results corroborate the molecular 
hypothesis about relationships, e.g., a close associa- 
tion of Hedyotideae and d. m (as in Bremer 
et al., 1995; Natali et E 
et al. m investi a set of 50 
individuals aA 36 t of Coprosma from 
New Zealand. They avena: patterns of hybrid- 
ETS 
nd high intra-individual hetero- 


ization and genotype mixing in ITS an 
sequences. They fou 
geneity, and the conclusion was that the widespread 
oecurrence of sequence mixture was a result of 


They also 


is hybridization. in genus. 
ncerted evolution in the 


uggested that con genus is 
[odi and that the mechanisms evolved to 
maintain a high level of heterogeneity as an adaptive 
value for Coprosma in the climatically unstable and 
physically complex New Zealand landsca T 
authors have sequenced many taxa, but they have 
not performed any phylogenetic analysis. It would be 
very interesting to investigate patterns of suggested 
hybridization in a phylogenetic framework 

Mitova et al. (2002) analyzed iridoid patterns within 
Galium with some phylogenetic considerations. They 
found differences in iridoid compounds and identified 
three lines of evolution: one that led to G. rivale 
(Sibth. & pur Griseb., a second that included G. 
mollugo L. and the G. incurvum group, and a third that 


s (eg 
op.). The hers eer 
be much improved if sampling and discussion are 
based on available phylogenetic data of the group 
). 


E 


e.g., Natali et al., 

Recently, Moserand et al. (2005) investigated 107 
Rubiaceae species for composition of leaf fatty acids. 
They used principal component analysis (PCA) and 
identified the tribes i 
Rubieae from their data. 
informative dis chemical puc are as the PCA 


, Psychotrieae, and 
is difficult to see how 


only shows similarities between species, which can 
completely contradict a phylogenetic relationship; 
furthermore, the results are compared to a nonphylo- 


Volume 96, Number 1 
2009 


Bremer 
Molecular Phylogenetic Studies of Rubiaceae 


genetic tribal classification (Robbrecht, 1993), so it is 
unfortunately very difficult to draw any conclusions 

about evolution and relationships of the fatty acids. 
Since the present review of molecular phylogenetic 
studies of Rubiaceae was presented at the Third 
International Rubiaceae Conference in 2006, ca. 10 
cular phylogenetic studies of Rubiaceae 

n publishe re not reviewed in this 
article, but the most important are as follows. 
Robbrecht and Manen (2006) have presented a super- 
tree construction of the family Rubiaceae. Several 
detailed M of tribes have a ublished, e.g., 
Coffeeae (Davis et al., 2007), Knexiene (Karehed & 
Bremer, 2007), Paederieae, Putorieae (Backlund et al., 
medmark et al., 2008) 


Further, molecular studies of enigmatic or problematic 


), and Urophylleae 


genera have been presented, e.g., Acranthera (Rydin et 
al, 2009), Eun rn et al., 2007), Guettarda 
(Achille et al., 2006), Psychotria (Sohmer & Davis, 
2007), and LL. (Mouly et al., 2007). 
From the ca. 50 molecular studies of the family 
reviewed in this article, we now have a good framework 
ily. We know that Rubiaceae 


of the phylogeny of the fam 
1 d there is high support for three 


are monophyletic an: 
subfamilies (Cinchonoideae, Ixoroideae, Rubioideae) 
Of these 
eae, Ard de Schra- 

nieae and 


and over 40 tribes. tribes, four are 
monogeneric (Cremaspor 
dereae, and Theligoneae). T es, Gar 


Morindeae, are pee sica m the base 


including subfamily Rubioideae. These three clades 
and the two clades corresponding to Cinchonoideae and 
Ixoroideae are highly supported. Subfamily Cinchonoi- 
deae includes nine tribes. Most interrelationships 
between these are unresolved. Subfamily Ixoroideae 
includes two monogeneric tribes (Retiniphylleae, 
Cremasporeae), 12 well-supported clades correspond- 
ing to tribes, and also several taxa referred to as a 
erly nM de e LM tribe Gardenieae. Subfamily 
monogeneric tribes (Schra- 
aa "B 
sponding to Mie and also taxa of a iy 


es corre- 


polyphyletic tribe Morindeae. Despite all these studies, 
there are stil roblems to be investigated in 
Rubiaceae Weisen [ee half of the tribes have been 
the focus of specific studies, and the large problematic 
genera are still in need of much investigation, e.g., 
Psychotria, Galium, Ixora, Pavetta, Ophiorrhiza, and 
Palicourea. Evolutionary investigations, biogeography, 
species richness, morphological traits, and other 
studies in Rubiaceae have just started, and with the 
diversity and disparity of the family, we can foresee an 
increased interest in the family and its phylogeny. 


Literature Cited 


Achille, F., T. J. Motley, P. P. Lowry II & J. Jérémie. 2006. 
Polyphyly i in CURE L. Ru. Guettardeae) based 


on nrDNA ITS sequence data. Ann. Missouri Bot. Gard. 


93: 103-121 
Alejandro, Razafimandimbison & S. Liede- 
Schumann. 2005. Polyphyly of Mussaenda inferred from 


ITS and trnT-F data and its ida for generic limits 
in Mussaendeae (Rubiaceae). A 

Anderson, C. L., J. H. E. Rova "& L Andersson : 
Molecular pliylogeriý of the tribe Anihospenae (Rubia- 
ceae): Systematic Es 7 piogeographic implications. Austral. 


5: Te and genera of the Cinchoneae 
plex (Rübiaceas): Ann. Missouri Bot. Gard. 82: 
—427. 


. Circumscription of the tribe Isertieae (Rubia- 
m E Bot. Belg. 7: 139-164. 

2002a. Relationships and generic circumscription 
in the a complex nai Psychotrieae). Syst. 
& a 


tablishmentobCarapich 
57: 363-374. 
Antonelli. 2005. Phylogeny of the tribe 
Es (Rubiaceae), its position in Cinchonoideae, 
ae description of a new genus, Ciliosemina. Taxon 54: 
—28. 


(Rubiaceae, 


ES kew Bull. 
— & 


C. Persson. 1991. Circumscription of the tribe 
"br Nest cladistic approach. Pl. Syst. 
Evol. 1 


T: i Rove, 1999. The rps16 intron and the 
phylogeny of the Rubioideae (Rubiaceae). Pl. Syst. Evol. 
214: 161-186. 
—, ——— . A. Guarin. 2002. Relationships, 
circumscription, and biogeography of Arcytophyllum 
(Rubiaceae) based on evidence from cpDNA. Brittonia 
49. 

Andrease 997. Phylogeny of Ixoroideae. Ph.D. 
ied “Uppsala De. Uppsala. 

— & B. Bremer. 1996. Phylogeny of T subfamily 
Ixoroideae (Rubiaceae). Opera Bot. Belg. 7: 
000. Combined -o analysis in 
Rubiaceae- Ixoroideae: 


9. 

uti ott region in su 
Bodas (Rubiaceae): E with cpDNA 
sequence data. Pl. Syst. Evol. 217: 

Axelius, B. 1990. The genus [E ANE (Rubiaceae)— 
Taxonomy, phylogeny, and biogeography. Blumea 34: 

—49T. 
Backlund, M. 2005. e ue Studies in the Gentia- 
nales—Approaches at Differ axonomic Levels. Ph.D. 
ee tation, Uppsala Uni ce Uppsala. 

———, B. Bie mer & M. Thulin. 2007. Paraphyly of 
Pusdesean. recognition 
Plocama ee 

Bremekamp, C. E. B. 1954. Les sous-familles et les tribus 
des Rubiacées. ate congrés international de botani- 
que, rapports et communications, Paris Sect. 
a 


emarks on the positions the delimitations 
"T in of the Rubiaceae. Acta Bot. Neexl. 15: 


Bremer, B. 1979. The genus Neurocalyx Ug ug 
Argostemmateae) in Ceylon. Bot. Not. 132: 399—407 


24 


Annals of the 
Missouri Botanical Garden 


— ———. 1992. Phylogeny of the Rubiaceae (Chiococceae) 
based on molecular and morphological data—Useful 
approaches for classification and comparative ecology. 

ae Bot. Gard. 79: 380-387. 

6. Phylogenetic studies within Rubiaceae and 
mea to other families based on molecular data. 
Opera Bot. Belg. 7 

& O. iiriksi 1992. 
istics and dispersal modes 

Rubiaceae. e J. Linn. Soc. 47: 7 

& R. K. Jansen. 1991. Gone restriction site 
phylogenetic 

J. Bot. 78: 


Evolution o d character- 
pical family 


mapping of chloroplast DNA implies new 
relationships within Rubiaceae. Amer 
198-213. 


& J. F. Minen. 7000. Phylogeny and classification of 
the subfamily R Pl. Syst. Evol. 225: 
43-72. 


& L. Struwe. Piin Phylogeny of the Rubiaceae and 
the Logania ngruence or conflict between mor- 
phological and ae data. Amer. J. 

1171- A 


——— € Thulin. 1998. Collapse of Isertieae, re- 
es e of M d genus o 
Sabiceeae (Rubiaceae: Phylogenetic idis based 

on rbcL data. Pl. Syst. Evol. 211: 71-92. 

, K. Andreasen & D. Olsson. 1995. Subfamilial and 
tribal us in the Rubiaceae based on rbcL 
sequence data. Ann. Missouri Bot. Gard. 82: 383-397. 

. Jansen, B. Oxelman, M. Backlund, H. Lantz & 
K. J. Kim 1999. More characters or more taxa for a robust 
na F wd from the coffee family (Rubia- 

ae). Syst. Biol. 48: 413-435. 

Bridson, D. M. oe e a 1988. Rubiaceae (Part 2). 
Pp. 415-747 in R. M. Polhill (editor), Flora of Tropical 
East Africa. A. A. Balkema, Rotterdam, The Netherlands, 
and Cee United Kingdom. 

Church, S. A. 2003. Molecular phylogenetics of Houstonia 
(Rubiaceae): pM. e aneuploidy and breeding system 
evolution in the radiation of the e across North 
America. Molec. P Evol. 27: 

& D. R. 12 or. 2005. i dud iis P paie 
among Houstonia (Rubiaceae) species: The influe of 

polyploidy on Rubus celular e 
1572-1380. 

Cros, J., M. C. Combes, P. Trouslot, F. p S. Hamon, 
A. Charri Phylogenetic ors 

L. Mole 


ssaendeae, and a new 


Bridson, C. Jarro & R. Goverts. 2001. The 
ded n and characterization of the genus Psychotria L. 
(Ru. E d J. Linn. Soc. 135: 35-42. 

r, O. Maurin & M. F. Fay. 2007. Searching 
for the 2 di Coffea (Rubiaceae, Trotoideae The 
circumscription and phylogeny of Coffeeae based on 
plastid — data and morphology. Amer. J. Bot. 94: 
313-329 

Delprete, P. G. 1996. Evaluation of the tribes Chiococceae, 

and Catesbaeeae ca based on 

morphological carats uw Bot. Belg. 7: 165-192. 

& R. Cortés-B. 2004. To Ms of the 
tribe Sipaneeae ‘Rubiaceae Ixoroidea irnL-F and 

ITS sequence data. Taxon 53: 347-356. 

. B. Smi R. M. Klein. 2004. Rubiaceae. I 

— Volume I—Géneros de A-G: 1. Alseis até 19 

Galium. Pp. 1-344 in A. Reis (editor), Flora Ilustrada 

Catarinense. Itajaí, Santa Catarina, Brazil. 


Condamineeae, 


Dessein, S., L. Andersson, E. Robbrecht & E. Smets. 2001. 
D s (Rubiaceae): A member of an emended tribe 
Virectarieae. PI. pM nx ol. 229: a —18. 

n, E. Smets & E. Robbrecht. 2005. 

"TM and Phylohydras (Hunbiicesg) Sister taxa 

permacoceae s.s., featuring a remarkable 

0 


ase of convergent E i Taxon 54: 91-107. 

: & A. Natali. 1994. cpDNA 

ences Rien rate restriction site data for 

ction Rubiaceae phylogeny. Pl. Syst. Evol. 190: 

Eriksson, O. & B. Bremer. 1991. Fruit characteristics, life 
forms, and species d in the plant family Rubiaceae. 
Amer. Nata dls 138: 7 


Erwin, T. 982. md fore sts: Their richness in 
Coleoptera: and other arthropod species. Coleopterists 
Bull. 36: 74-75. 

Fay, M. F., B. Bremer, G. T. Prance, M. van der Bank, D. 


Bridson & M . W. Chase. 2000. Plastid rbcL sequences 
show Dialypeialanthus to be a member of Rubiaceae. Kew 
864 


E; eo E. Robbrecht, D. Bridson, A. 
e. 2006. World Checklist of 
e. The Board of aon of the Royal Botanic 


Güstafssön. c. & C. Persson. 2002. Phylogenetic relation- 
ships among species of the aci id genus Randia 
(Rubiaceae, Gardenieae) inferred from molecular and 

i P 


El-Ghazaly & E. Smets. 1998. Wm in 
Anc bó “Mor phology, function, distribution, and 
relation with tapetum types. Rev. 64: 240-27 2. 

& bo d Eo a xs 


hidden secret of the 3 Pp. 

Nordenstam, G. El- Ghaz aly & M. > [x un. 

Systematics 35 the 21st Cent tury. Wenner-Gren Int 

tional Series, Vol. T. br Press, London. 

Robbrecht. 2003. Pollen 
s 


15: 
sbcipholdgy of NW 
monophyly of pad (Rubiaceae). Rev. 
Palynol. i. 

Jansen, S., e, F. Piesschaert, E. Robbrecht & 
E. Smets. E oan de to the wood anatomy of the 
jode (Rubiaceae). J. Pl. Res. 114: a 

Robbrecht, H. Beeckman & E. Sm a 

— d the systematic wood anatomy of iis piene 

Int. Assoc. Wood Anat. J. 23 a 


Palaeobot. 


ts . Robbrecht. is A 
metal pee a leaves of s Al- 
da Rubiaceae. Ann. Bot. 91: 657-663. - 


uu E 56: 1051-1076. 

995. Chromosome survey of the Rubiaceae. Ann. 
eee Bot. Md 82: 398-408. 

Kuhlmann, J. G. 1925. Contribuicáo para a conhecimento de 
algumas Tm novas, contendo tambem um trabalho de 
critica e novas combinações. Arch. Jard. Bot. Rio de 
Janeiro 4: 347—365. 

Lantz, H. B. Bremer. 2004. Phylogeny inferred from 
morphology and DNA oe ee well- ps 

J. Linn. Soc. 1 


groups in 


—. 2005. Phylogeny of the complex 
locis (Rubiaceae) genera Fadogia, Rytigynia, 
Vangueria with close relatives and a new circum- 
selon of Vangueria. Pl. Syst. Evol. 253: 159-183. 


Volume 96, Number 1 
2009 


Bremer 
Molecular Phylogenetic Studies of Rubiaceae 


——. K. Andreasen & B. Bremer. 2002. Nuclear rDNA 
ITS s sequence data used to construct the first phylogeny of 

73-187. 

uy P: . Com i 


rred 
ribosomal DNA. Theor. Appl. Genet. 94: 9 55 
Malcomber, S. T. 2002. o f Gaertnera La 
(Rubiaceae) based on multiple DNA markers: Evidence 
of a rapid radiation 


in a veu morphologically 
diverse pes oe. 56: 42-57. 

& A. P. s. 2005. Six new Gaertnera (Rubiaceae) 
O from il and phylogenetic analyses that 
support Hymenocnemis as a synonym of Gaertnera. In R. C. 
Kea, V. C. Hollowell & T. Croat (editors), Festschrift 
cy: The Legacy of a Taxonomist. 

371-397. 
. 1996. The pi Eas atpB-rbcL 
spacer in Rubiaceae. on ra Bot. Belg. 7: 51-57. 
& Ehrendorfer. 1994. cae of 
EU Rubieae inferred from the sequence of cp 
intergene region. Pl. Syst. Evol. 190: 195-211 
: P : 


Maurin, O., A. avis, M. Chester, pn? 
Kakim & M. F. Fay. 2007. Towards a phyl y of Coffea. 
(Rubiaceae): Identifying well-supported lineages d on 
nuclear and DNA sequences. Ann. Bot. 100: 
1565-1583 

McDowell, T. 1996. Exostema (Rubiaceae) Taxonomic 
history, nomenclature, peu 


and subgeneric classifi- 
—295. 


cation. Pls ra Bot. Belg. 7: 
B 


Bremer. i ylogeny, diversity, and 


mese in Exostema dud Implications of 
morphological and molecular analyses. Pl. Syst. Evol. 
212: 215-246. 


———À, sek & P. Manos. 2003. Biogeography of 
Bosena (Rubiaceae) 1 in the Caribbean region in light of 
molecular phylogenetic analyses. Syst. Bot. 28: 431—441. 
A. Badoc, B. Patouille, C. Lacomblez, M. 
& J. J. Bessoule. 2005. Chemotaxonomy of the 
e family based on leaf fatty acid composition. 
Phytochemistry 66: 549-559. 
Motley, T. J., K. J. Wurdack & 5 G. Belpiste 2005. 
Molecular systematics of the Catesba 
ubiaceae): Flower d fait aon and 
biogeographic implications. Amer. J. Bot. 92: 316-329. 
A. afimandimbison, F. Achille, T. Haeverman 
& B. Bremer: 2007. m D Cae of Rhopalo- 
brachium P (Rubiaceae): E from molecular 


(rps16 and trnT-F) and RR NNI data. Syst. Bot. 32: 
8 a 


Moynihan, J. & L. E. Watson. 2001. Ra RE generic 
allies, and nomenclature of Caribbean endemic genus 
Neolaugeria duce ee on internal inc bed 
spacer sequences. Int. J. Pl. Sci. 162: 393-401. 

Natali, A., J. F. ET. Ehrendorfer, 1995. Phylogeny of 


sequence. Ann. Missouri Bot. Gard. 82: 

, M. Kihn € F o 1996. Tribal, 
spécifie rolas in the Rubioideae— 
n sequence data of the 


denen and 


an Piogcosraphy di Erithalis ubiacene) a an en 
the Caribbean Basin. Pl. Syst. Evol. 234: 7 


— & ———. 2003. Hypotheses for the de rie of 
E Caribbean asin by two genera of the Rubiaceae: 
Erithalis and Ernodea. Syst Bot. 28: 442—451. 
Nepokroeff, M., Sytsma & 
Reorganization of the genus Psychotria and tribe Psycho- 
trieae (Rubiaceae) inferred from ITS and rbcL sequence 
data. Syst. Bot. 24: 5-27. 


Bremer. 1999. 


A. Zimme 


—— . Wagner & E. er 2 
Reċönstiucüng ancestral patterns of colonization and 
dispersal in the Hawaiian understory tree genus Psychotria 
(Rubiaceae): A comparison of parsimony and likelihood 
ora Syst. Biol. 52: 820-838. 

Z. W n & B. Bartholomew. 2005. 
jet of Kelloggia 159. ex Benth. (Rubi dun and 
evolution of its intercontinental disjunction between 
western North America and eastern Asia. Amer. J. Bot. 
92: 642-652 


Novotny, V., Y. Basset, S. E. Miller, D. Weiblen, B. 
Bremer, È Cizek & P. Drozd. 2002. Low host specificity of 
herbivorous insects in a tropical fee Nature 416: 

—844. 
e F. 2000. How us species of deest Erwin's 
mate revised. Biol. J. Linn. Soc. 71: 583-597 
Olmstead, R. G., B. Bremer, K. M. Scott & , 
993. A parsimony analysis of the Asienidas pee lato 
"ie on rbcL sequences. Ann. Missouri Bot. Gard. 80: 


Palmer. 


Persson, C. 1996. o E "x Gardenieae (Rubiaceae). 
Bot. J. Linn. Soc. 

———. 200 Oa. Polen T Sade (Rubiaceae) based 
on chloroplast DNA sequences from the rps16 intron and 
irnL(UA A)-F(GAA) winds spacer. Nord. J. Bot. 20: 
C 


00b. Phylogeny of the neotropical Alibertia group 

T with emphasis on the genus Alibertia, 

nds from ITS and 5S ee DNA sequences. 

r. J. Bot. 87: 1018-102 

a E: 2001. Cael and Pollen see se 
the PE paco ae). Ph.D. Dis 

—— “1997. Dialypetalanthus 

pmo Kuhlm. Dialspcliniuceac) The proble 


n endemic. 


matic 
taxonomic position of an 
Missouri Bot. Gard. 84: 201-223. 
Puff, C. 1982. The delimitation of the tribe Anthospermeae 
and its affinities to the Paederieae (Rubiaceae). Bot. J. 
Linn. Soc. 84: 355-377. 
—. 1986. Marie (Rubiaceae-Spermacoceae)—A 
ccommodate the African and 
“ios species. Pl. 
& U o h 


xenon pos 


. Bremer. 2001 [2002]. Tribal 
delimitation of Naucleeae (Cinchonoideae, Rubiaceae): 
Inference from molecular and morphologcal data. Syst. 
Geogr. Pl. 71: 515—538. 

& ———. 2002. Phylogeny and € of 
Naucleeae s.l. (Rubiaceae) inferred from ecular p 
rbcL, and trnT-F) and morphological dus Amer. J. B 
89: Nc orn 


, E. g & B. Bremer. 2004. Recent origin and 
m idle. of divergent ITS putative pseudogenes: 
A case study from Naucleeae (Rubiaceae). Syst. Biol. 53: 
17 "192. 


26 


Annals of the 
Missouri Botanical Garden 


oog, H. Lantz, U. Maschwitz & B. Bremer. 2005. 
Re- àpsesomioul of mon MA evolution ES myrmecophyt- 
ism, and rapid radiation in Neonauclea s.s. (Rubiaceae). 
Molec. Phylogenet Evol. 34: 334— 
Rizzini, C. & P. Occhioni. 1949. Dialypetalanthaceae. 
Lilloa 17: pues 
Robbrecht, E. 1988. Tropial woody Rubiaceae. Opera Bot. 
Belg. 1: 1-271. 
19 93. Supplement to the 1988 outline of the 
classification of the Rubiaceae. Index to genera. Opera 
Bot. Belg. 6: 173-196 
& C. Puff. 1986. A survey of the Gardenieae and 
related Hae Pd "Bat Jarb. Syst. 108: 63-137. 


en. lineages 
of the Eom a (Rubiaceae, Anei pans: Combined 


to E Me Posten li 
P 


new 


ubfamilies, n and 
6: 85-1 
R . G. Delpret E Andersson & V. A. Albert. 
2002. A immb-F npn sequence study of ine Eun 
Rondeleti 


on "m o of ie Rubiäceae Amer. i Bot. 89: 
145-159. 


c C., K. Kainulainen, S. G. Razafimandimbison, J. E. E. 

ark & B. Bremer. 2009. Deep divergences in the 

coffee f family and the systematic position of Acranthera. Pl. 
Syst. Evol. 278: Ed 


n, S. * Pu doe S. 


Bremer. 2008. A 
phylogeny of Urophylléae e based on rpsi6 


: 2007. The genus Psycho- 

iria. (Rubia dene). i in the Philippine Archipelago. Sida, 

Botanical Miscellany 27. Bot. Res. Inst. Texas, Fort 

Worth. 

Stork, N. E. 1993. How many species are there? Biodivers. & 
Conservation 2: 215— 

Thulin, M. & B. . Studies in the tribe 
Spermacoceae o. d The circumscrip- 
tions of Amphiasma and Pentanopsis and the affinities of 

233-239. 


Bremen: 


Phylohydrax. PL. Syst. Ev dl "AT. 
Verdcourt, B. 1958. ens on the P E of the 
Paw Bull. Bot. État 28: 209— 
Wi < a , S. D. Wright, E. K. eae D. J. Keelin, 
Carino 2002. Elevated genetic heterogeneity 
i Di esa pun instability: Inferences from 
New Zealand 


nrDNA in Coprosma (Rubiaceae). J. 
Biogeogr. 29: den 


REVISIÓN SINÓPTICA DE Elsa T Obra 
GALIANTHE SUBGEN. GALIANTHE 
(RUBIACEAE: SPERMACOCEAE), 

CON UNA SECCIÓN NUEVA! 


RESUMEN 


Galianthe subg. Galianthe Griseb. se caracteriza por el fruto de mericarpios dehiscentes, semillas rollizas o complanadas 
con bordes aliformes, inflorescencias generalmente amplias tirsoides o pleiotirsoides, flores heterostilas, hábito e 
genera SE con xilopodio y cromosomas x — 8. Está jo od por 39 especies sudaméricanas que se agrupan 
secciones: sect. Galianthe (30 especies) y una nueva . Laxae E. L. Cabral (nueve especies y dos sul Koen c 
PUE Mie para diferenciar E s y las — de cada sección, con resumen snes de las especies y 
mapas de distribución. Se designan n. s para cuatro nombres: Borreria angustifolia Cham. & Schltdl. [= G. angustifolia 
(Cham. & Schltdl.) E. L. Cabral], Be i " d es Cham. & Schltdl. [= G. equisetoides (Cham. € Schltdl.) E. L. Cabral], B. 
thalictroides K. Schum. [= G. thalictroides (K. Schum.) E. L. cane B. valerianoides Cham. & Schltdl. [= G. valerianoides 
(Cham. & Schltdl.) E. L. hr: Se designan lectotipos para cinco nombres: B. de f. glabrior Chodat & Hassl. 
[= G. centranthoides (Cham. & Schltdl.) E. L. Cabral], B. ericoides Cham. & Schltdl. [= Galianthe peruviana (Pers.) E. L. 
Cabral], B. leiophylla K. haa. | = G. fastigiata Griseb.], G. hassleriana (Chodat) E. L. Cabral y G. verbenoides (Cham. € 
Schltdl.) Griseb. 


ABSTRACT 


The species of the genus Galianthe subg. Galianthe Griseb. are revised. The pron Fc e ens from South 
America and is characterized by its fruits of dehiscent mericarps, plump or l with wing-like margin usually 
wide, boa or pleigthyrsoid ales an erect habit ud with xylopodium, and a Denm ome n = 
8.T ctio Galianthe (30 species) and the new section Laxae E. L. Cabral (nine species, including 
). A key to the sections and their species are RUN Ei The species unn FE. are not described in this 
k. Neotypes are here designated for four names: Borreria angustifolia Cham. & Sch = G. angustifolia (Cham. & 
Schltdl) E. L. Cabra al], B. equisetoides Cham. & Schltdl. [= G. equisetoides (Cham. & Sc ee al], B. thalictroides K. 
Schum. [= G. thaliciroides (K. Schum.) E. L. Cabral], and B. valerianoides Gan. & Schltdl. [= G. valerianoides (Cham. & 
Schltdl.) E. L. M Lectotypes are designated here for five names: B. centranthoides f. glabrior Chodat & Hassl. [= G. 
centranthoides (Cham. € Schltdl.) E. L. Cabral], B. ericoides Cham. & Schltdl. [= Galianthe peruviana (Pers.) E. L. Cabral], B. 
leiophylla K. Schum. | =G. fastigiata Griseb.], G. hassleriana (Chodat) E. L. Cabral, and G. verbenoides (Cham. & Schltdl.) 
Griseb. 
Key words: Galianthe subgenus Galianthe, Rubiaceae, Spermacoceae. 


" 


Galianthe Griseb. es un género americano de la M in 1879). Además lo relaciona por las 
tribu Spermacoceae (Robbrecht, 1988), representado scencias con el género Emmeorhiza Pohl, por 
por 49 especies de distribución tropical y subtropical. E déliscencia de los frutos con Borreria G. Mey. y 
En la descripción original Grisebach lo define por: con Galium L. Haciendo alusión a este último género 
“fructus dicoccus, coccix aequaliter secedentibus lo denomina Galianthe. Grisebach describió este 
apice et intus dehiscentibus. Semina oblonga, a dorso género con dos especies nuevas, sobre material 
compressa. Flores in cymas iterato-tri-dichotomas v. —colectado por Lorentz en Argentina: Galianthe 
apice breviter scorpioideas dispositi, alari ebracteato, fastigiata Griseb., designado como tipo del género 
lateralibus pedicellatis, folis floralibus minutis? por Cabral (1991) y G. clidemioides Griseb., sinónimo 


1 Agradezco a los curadores de las distintas instituciones que han facilitado material en préstamo, cuyas siglas de lo 
herbarios se citan a continuación: AS, B, B-W, BA, BAA, BAB, BACP, BAF, BHCB, BHMH, BM, BR, CEN, CEPEC, CORD, 
CTES, ESA, F, FCAB, FCQ, C, GB, G-DC, HAS, HB, HBR, IAC, IBGE, ICN, IPA, JPB, K, LIL, LP, LPB, MA, MBM, MCNS, 

, MVFA, MVM, NY, OUPR, P, PACA, PY, R, RB, SI, SP, SPF, TEX-LL, UB, UEC, UPCB, US, USZ. Agradezco a Nélida 
Bacigalino la lectura crítica del manuscrito y estímulo permanente; a Carmen Cristóbal y Antonio Krapovickas por sus 
valiosas sugerencias; a Otto F. Ferber, Roberto Salas y Walter Medina por la colaboración en la edición de ilustraciones y 
mapas. Agradezco a Laura Simón por las d 

? Los editores agradecen a Diana Gunter por su solahas ración en la redacción de este manuscrito. 

*Instituto de Botánica del Nordeste (UNNE. CONICET), Facultad de ee Exactas y Naturales y Agrimensura, UNNE, 
Casilla de Correo 209, 3400 Corrientes, Argentina. ecabral@agr.unne.edu. 

doi: 10.3417/2006193 


ANN. Missouni Bor. Garp. 96: 27-60. PUBLISHED ON 23 APRIL 2009. 


28 


Annals of the 
Missouri Botanical Garden 


de G. centranthoides (Cham. & Schltdl.) E. L. Cabral. 

En el mismo trabajo cs una nueva combinación 6. 

hltdl) Griseb. [= 

verbenoides Cham. & E |, material erróneamente 

o por tratarse de G. laxa (Cham. & Schltdl.) 
L. Cabral (Cabral, 1991). 


coe 
eba 


verbenoides (Cham 


n ds la propuesta de 
Grisebach de considerar a grupo de especies 
como un género ane ee y ES vuelve a incluir en 
Borreria como sect. Galianthe (Griseb.) K. Schum., 
diferenciándolo por sus flores dimorfas en inflore- 
scencias tirsoides, de Borreria sect. Borreria, con 
flores isomorfas en inflorescencias capitadas termi- 
nales y/o axilares. Los estudios floristicos realizados 
os generales, el 
criterio de Schumann, sólo con id cambios 


con posterioridad siguieron en tér 


menores. 

Con el análisis de abundante material americano de 
Borreria s.l., se observaron marcadas diferencias 
morfológicas entre las especies de las dos secciones, 
motivo por el cual se buscaron nuevos elementos de 
juicio que pudieran aportar una correcta valoración de 
Borreria sect. Galianthe. 

El estudio palinológico realizado por Pire y Cabral 
(1992) demostró la homogeneidad de los caracteres de 
los granos de polen (colporados, semitectados con 
retículo complejo) a diferencia de los de Borreria sect. 
Borreria en que los granos de polen son porados, 

colpados, colporados, tectado-perforado, foveolados 
(Cabral, 1985; Pire, 1997) 

Con respecto a los datos citológicos, los primeros 

recuentos en especies de Borreria ct 


& 


mostraron un nümero básico, x — 14 (Kiehn, 1985, 
1986, 1995). Posteriormente Davifia y Cabral (1991) 
obtuvieron el nümero básico, 
Galianthe, 
recuentos cromosómicos en el género Galianthe. 
inflore- 


scencia tirsoidea con flores distilas, polen, semilla 


, en especies de 
Borreria sect. "(— los primeros 
Los siguientes caracteres diferenciales: 


aladas o ápteras, cromosomas y distribución geográ- 
fica limitada esencialmente a América del sur, 
mientras que Borreria es pantropical, se consideraron 
suficientes para justificar la separación de las 
Ha de Borreria sect. Galianthe, en otro género. 

esta manera se rehabilitó el género Calianthe 
isl d 
es 


ecies "s pn cuales 


y se reconocieron en el mismo 20 
nuevas Ee ASA 
se incorporan nuevos sinónimos y se amplía el área de 
distribución geográfica de algunas especies. Del total 

e los taxones registrados por Schumann en Borreria 
sect. Galianthe, se excluyen sólo dos especies: P. 
cymosa (Spreng.) Cham. & Schltdl. y B. monodon K. 

chum., por no reunir los caracteres que definen a 
Galianthe, las que fueron incorporadas en un género 
nuevo, Scandentia E. L. Cabral & Bacigalupo (Cabral 


& Bacigalupo, 2001), 


ores homostilas, semillas complanadas 


caracterizado por hábito 
trepador, 
zonocolporados con 

Además Cabral y Buoni (1997) en el trans- 
curso de la revisión de la tribu Spermacoceae 
extrajeron unas especies de Diodia L. y Borreria sect. 
Borreria cuyos caracteres no se ajustaban a la 
definición de esos géneros y presentaban afinidad 
con Galianthe. Esta idea fue apoyada también por Pire 
(1997) porque sus granos de polen tienen exina de 
retículo doble, 
Galianthe. 
mericarpos indehiscentes, entonces se incluyeron en 
Galianthe subg. Ebelia (Rchb.) E. Cabral & 
Bacigalupo (Cabral & Bacigalupo, 1997). 


Los autores que han estudiado las 


característica muy particular de 
Sin embargo estas especies tienen los 


especies 
reconocidas en este género, han utilizado diversos 
caracteres para definirlas, como el tipo de inflo- 
rescencia (de Candolle, 1830), el dimorfismo floral 
(Schumann, 1888), dehiscencia de frutos, ramifica- 
ción, estructura subterránea y semillas (Cabral, 2002), 
caracteres del polen (Pire € Cabral, 1992; Pire, 1997; 
oe 2002), análisis embriológico (Galati, 1988, 

991) y recuentos eromosómicos (Kiehn, ; 
ee & Cabral, 1991; Cabral, 2002). ieee 
(2003) en un estudio filogenético molecular confirma 
las sinapomorfias morfológicas, palinológicas y cro- 
mosómicas indicadas por Cabral y Bacigalupo y 
concluye que Galianthe merece la categoría genérica. 


TRATAMIENTO SISTEMATICO 


Galianthe Griseb., Symb. Fl. Argent. n Kónigl. 

es. Wiss. Göttingen] 24: 156. . Borreria 

sect. Galianthe PL ed Sim. Fl. Bras. 
(Martius) 6(6): 40-42. 

Galantle cs aL Eum Field Colum- 

ot. Ser 31. TIPO: 

K. Schum. [= Galianthe 


Borreria subg. 


ode. "eiophylla 
fastigiata Griseb.]. 
Distribución y hábitat. Constituido por 49 espe- 
cies, sudamericanas, Argentina, Bolivia, Brasil, Para- 


m Uruguay y Perú, con excepción de un único 
ón que vive naturalizada en México, Guatemala, 
ides (Galianthe brasiliensis subsp. angulata 


(Benth.) E. L. Cabral & Bacigalupo). Habitan en 
campos rupestres, en campos bajos, en sabanas y 
laderas de cerros, en suelos lateríticos, arenosos o con 


Excepcionalmente son umbrófilas, en 
sotobosque E E y bosques (G. hispidula (A. 
Rich. ex . Cabral & Bacigalupo, €. 
brasiliensis y w. ie. (Cham. & Schltdl.) E. L. Cabral). 


Volume 96, Number 1 
2009 


Cabral 29 
Revisión Sinóptica de Galianthe 


CLAVE PARA DIFERENCIAR LOS SUBGÉNEROS DE GALIANTHE 


la. Fruto de mericarpos dehiscentes; semillas rollizas 
o complanadas, con margen aliforme; hábito erecto, 


con frecuencia xilopodio muy s Ado, tallos 


nunca alados; cromosomas x Map Sur 

A eadera d eA 2. Galianthe 
lb. Fruto de puc d indehiscentes; E roll- 

izas con margen liso; hábito o, postrado, 


erecto, trepador, sin 2 s M alados o no; 
15. Centro y Sudamérica 
subg. Ebelia 


cromosomas x = 12, 14, 
(20°N-35°S) 


Galianthe subg. Galianthe 


Sufrütices erectos o apoyantes, con o sin xilopodio, 
glabros o pubescentes; tallos de tetrágonos a sub- 
cilíndricos. Hojas sésiles o pseudopecioladas, opues- 
tas y decusadas, con frecuencia pseudoverticiladas 
por la presencia de braquiblastos, persistentes o 


excepcionalmente caducas (Galianthe equisetoides 


Cabral & Bacigalupo); estípulas persistentes, inter- 
peciolares y unidas a la base de la hoja en forma de 
l-multi- 


vain sta a se prolonga; borde 


és pro 
fim nado: con i apicales. Inflorescencias 


tirsoideas, a veces con inflorescencias parciales más o 


menos ac Flores bri actinomorfas, 
generalmente or vistilas un poco más 
grandes que las longistilas; cáliz 4- -mero, con 


frecuencia con dientes menores en los s senos, a veces 
también con coléteres; corola 4-mera, infundibuli- 
forme, blanca, excepcionalmente rosada o lilacina, 
externamente glabra, pilosa o pubescente y en su 
interior con pelos moniliformes de distribución igual o 
diferente 


en flores longistilas y brevistilas; disco 
nectarífero entero o bipa i 


rtido, tapizado por papilas 
estriadas; estambres fijos en la garganta de la corola o 
en filamentos de distinta longitud en flores longistilas 
o brevistilas, anteras dorsifijas, pulcre gineceo 2- 
carpelar, 2-locular, con 1 óvulo por lóculo, peltado, 
fijo al septo interlocular; estilo filiforme, estigma 
bífido. Cápsula septicida de mericarpos dehiscentes, 
con cáliz persistente; semillas rollizas o complanadas, 
de margen con reborde o ala muy estrecha, estrofíolo 
persistente o caduco; exotesta con fovéolas super- 
ficiales o profundas, isodiámétricas o poligonales. 
eños, medianos y grandes, 
6—7(8-10) colporos; 
exina semitectada-reticulada con retículo complejo 
(Pire «€ Cabral, 1992; Cabral, 2002); 
cromosómico x = 8 (Daviña & Cabral, 1991). 


ranos de polen peq 
prolato-esferoidal o subprolato, 
número 


Distribución geográfica y ecología. Las 39 espe- 


Paraguay, Perú y Uruguay (1°-35°S); la mayor 


concentración con 23 especies endémicas, en el 
planalto central y meridional de Brasil y en un sector 
del Paraguay oriental. Crecen en cerrados, campos 
rupestres os bajos, en suelos lateríticos, 
arenosos o con afloramientos rocosos (observ. pers. y 
referencias de etiquetas). Generalmente se hallan 
como plantas aisladas y sobreviven a incendios y 
suelos removidos por el xilopodio. 

Las especies se reconocen por los caracteres 
incluidos en la clave y son presentadas por orden 
alfabético. En los apéndices 1 y 2 se proveen una lista 
de las especies y un índice de colecciones. 


CLAVE PARA DIFERENCIAR LAS SECCIONES DEL SUBG. GALIANTHE 


la. Plantas con xilopodio; monocaules o pluricaules; 
tallos simples o con escasas ramas secundarias 
cortas; inflorescencia largamente pedunculada sólo 
en tallos primarios, o rara vez brevemente 
pedunculada en ramas secundarias 


I—sect. Galianthe (30 especies) 


lb. Plantas 
secundarias desarrolladas; 
1 ] 1 


sin xilopodio, pluricaules, ramas 


con 
inflorescencia general- 
] nu bi z 


F F i 
en ramas secundarias .... II—sect. Laxae (9 especies) 


I. Galianthe subg. Galianthe sect. Galianthe 


Sufrútice con xilopodio desarrollado, con 1-20- 
tallos simples o si presentan ramas secundarias las 
inflorescencias se ubican sólo en los tallos primarios; 
inflorescencia pluriflora largamente pedunculada, 
raro pauciflora brevemente pedunculada; flores dis- 
tilas, 4-meras, frutos capsulares; semillas frecuente- 
mente aladas. 

Comprende 30 especies que presentan su mayor 
concentración y mayor variabilidad en el planalto 
central y meridional de Brasil y en las serranías de 
Paraguay oriental, y gradualmente en menor cantidad 

e especies en Argentina, Bolivia, Perú y Uruguay, en 
campos, bajos, inundables o en campos altos hasta 
O m. Figura 1 (A—F). 


CLAVE PARA IDENTIFICAR LAS ESPECIES DE LA SECCIÓN GALIANTHE 


l; Plantas monocaules o o tallos 


simples o 
cias largamente pedunculadas m en tallos 


con ramas secundarias; orescen- 


primarios IS Sey ee deu E Saag d dE Dre AS ee aie 
l. Plantas eae tallos con ramas secun- 
darias; inflore ias brevemente peduncu- 
m en S primarios y en ramas secun- 
WAS» etes eis e lea E e ecd es Ets 27 
2(1). a con un tallo principal con ramas 
secundarias desarrolladas ............... 3 
Ze Sufrútices con un tallo o varios tallos 
vc tallos simples con escasos bra- 
quiblastós s zc acetate du het o anne eus 4 


30 Annals of the 
Missouri Botanical Garden 
3(2). ate con 1-2(—5) tallos de 0.4—1(-2.5) m mente paralelos; vaina estipular 3.5—5 mm 
"prre" ong., pubescente, con 9-11 lacinias lineares 
3. ns con (5-)10-20 tallos de 0.1—0.5(-1) o linear-subuladas ....... valerianoides 
male: ros etu e et to Cte feto 12(10). Hojas persistentes, lanceoladas u oblongo- 
4(2). Hojas lineares o linear-lanceoladas ........ lanceoladas, de uda largamente 
4. Hojas elípticas, elíptico-lanceoladas, elíptico- atenuada, levemente pubescentes o escábri- 
oblongas o elíptico-ovadas ......... llus. das, plegado-nervosas, con 4— res de 
5(4). Hojas pul pubérulas o escabrosas, (8—) nervios; vaina estipular levemente pubescente 
20-35 X (0.7-)3-5 mm; flores brevistilas: o escábrida, con 7-12 lacinias de 5-25 mm 
superficie interna de la corola solo con pelos long.; tallos subtetrágonos de 1-1.7 
enel tubo? dace dese aye ade aoe? semillas aS: coena 25. G. ps ee 
9 as pubescentes, 5-10 X 0.2-2 mm; flores 12. ojas ane lineares, filiformes o elípticas, 
IR con pelos en la superficie interna de base aguda, glabras, planas con nervios 
oe flores longistilas pelos en secundarios inconspicuos o con 2-3 nervios 
B tubo y en los pétalos ...... 17. G. linearifolia secundarios; vaina estipular glabra, con 5 
6(4). Hojas 8-25 X Dato mm, glabras; vain lacinias de m long.; tallos cilíndricos a 
estipular glabra o pubérula, con lacinias de subcilíndricos de 060-150 m alt., constrictos 
0.2-2.5 mm long.; hipanto y fruto glabros; en los nudos; semillas redondeadas ...... 
superficie externa de la corola micropapilosa, eee eee eee eee eee eee 1. G. aspe 
papilas más densas en el dorso de los 13(9. Hojas de 0.5-5(-7) mmlat.............. 
lóbulos via me ts 9. G. thaliciroides 13. Hojas de (2-)5 Bain) mila; 2d ie 
6. Hojas 20-35 X 3—4 mm, ue vaina 14(13). Nudos caulinares con pseudoverticilos hasta 
estipular pubescente, con lacinias de 3— 15- Mise hojas abr ras o pubérulas en 
ong.; hipanto y fruto pubescentes; eae ambas caras o solo en el envés, con 2-3(-4 
externa de la corola pubérula .... 22. G. montesi cab de nervios secundarios 1. G. angustifolia 
7(5). Hojas de 6-11(-18) mm ae glabras, nervios 14. Nudos caulinares con serdar igs 2-6- 
secundarios inconspicuos ....... 6. G. elegans foliolados; hojas glabras, con nervios secun- 
7. Hojas d 55(-70) mm i escabriúsculas, aros INCONSPICUOS a 2 os oe a ecra ras 15 
levemen d o pubescentes, nervios 1514). Tallos de 0.40-1.5 m alt.; a 12357) mm 
secundarios visibles ........ooooooooooo.. lat; vaina estipular 1-2.5 mm long. con 
8(7) Tallos casi ee con pelos ralos; vaina lacinias de 2-10 mm long.; pos 
ipular] prol la y ima d amplia, laxa, rain de 17-28 cm long.; 
la inserción del par de hojas correspondiente; corola 5-7.5 mm long. ....... wG: ms 
hojas con nervios terciarios inconspicuos; 15. Tallos de 0.40— o. 60 m alt.; hojas de 0.5-1 m 
hipanto papiloso o raro con pelos dispersos; lat.; vaina estipular 4.5-6.5 mm long., con 
fruto glabro 44.3 mm long. 4. G. chodatiana lacinias de 5-15 mm long.; 2 a 
8. Tallos pubescentes; vaina estipular truncada; comprimida, pluriflora, de 4— ke em go 
hojas con nervios terciarios conspicuos; hi- corola 3-4.5 mm long. ........ Ha longifolia 
panto pubescente; fruto pubescente 4 mm (13). Hojas con nerviación secundaria inconspicua o 
Iur MERE 3. C. pm el nervio primario conspicuo. .............- 17 
9(3). Plantas de 0.60-2.50 m alt.; en terrenos bajos 16. Hojas con nerviacién secundaria notoria .... 19 
inundables, pantanosos o en borde de arroyos, 17(16). Nudos caulinares con pseudoverticilos hasta 
ríos, esteros o Id a iaa TO-foliados: ia aa a ans 
9. Plantas de 0.20-1.80 m alt.; viven en campos 17. Nudos caulinares sin pseudoverticilos fo- 
altos no inundables ........oo.o.o.o.o.o.o.o.. TALES ecu UE a a a 8. G. fastigiata 
10(9) ^ Hojas oblongo-lanceoladas a lanceoladas con 18(17). Nudos foliares 3—4-foliolados; vaina estipular 
la base obtusa o truncada ....... llli. truncada 5-10 mm long. con 3-5 lacinias 
10. Hojas filiformes, el elipticas, lanceola- soldadas formando un solo diente principal; 
das u oblongo-lanceoladas, de base aguda o corola rosada acina ........ 28. G. souzae 
largamente at di boe uA VE 18. Nudos foliares icio foliolados; vaina estipular 
(10). Tallos 0.60—0.80 m alt., subtetrágonos, glab- prolo > argen ular con notable 
ros; hojas glabras con nerviación paralelo- lacinia central, triangular, acuminada y laci- 
ee vaina estipular 6-7 mm long., con nias laterales más cortas; corola blanca 
breve prolongación por encima de la inserción eee eee eee eee econ... 15. G ld 
de las hojas, pubérula en el margen, con 5-6 (16). Vaina estipular prolongada por encima de la 
lacinias filiformes .......... 20. G. macedoi inserción elg ar de hojas o RF 2200 
b Tallos 0.80—2.50 m alt., tetrágonos, escabro- 19. Vaina 1 
sos, ángulos con pelos retrorsos; hojas escáb- inserción del par de hojas conesgondien tes 
ridas, plegado-nervosas, con 2-3 pares de 20(19. Nudos caulinares con pseudoverticilos; in 


p 
nervios basales y 2—3 suprabasales, ligera- 


30)40—50 redondeada 


mm long., base 


Volume 96, Number 1 


Cabral 31 
Revisión Sinóptica de Galianthe 


2009 
] t dada; la de lóbulos y tubo + 28(27) Vaina estipular pubérula con una lacinia 
del mismo largo, con papilas largas en el dorso central subulada hasta de 1.5—1.8 mm long. y 
de los lóbulos ........... 11. G. guaranitica os apéndices laterales escasamente desarrolla- 

20. Nudos caulinares sin ET us dos de 0.5-0.7 mm long.; corola 5-6 mm long. 

55-90 n . G. reüzii 

de lóbulos. más PE que el tubo, extemamente 28. Vaina estipular levemente pubescente, pu- 
micropapilada, con papilas cortas en el ápice de érula o pubescente, con 3-6 lacinias sub- 
los lóbulos .............. 21. G. matogrossiana iguales, filiformes de 1-3.2 mm long.; corola 

21(19). Tallos de 0.60-1.80 m alt., enteramente js np older 24. G. peruviana 

érulos a diversifaciales; hojas (0.5324 c 29(27) Hojas plegado-nervosas, pubescentes; vaina 
lat., pubérulas a pubescentes ... 10. G. : grandifilia estipular ese con 5-7 lacinias; inte- 

2]. Tallos de 0.30—0.60 m alt., e a pubér- ror de la corola brevistila con pelos mon- 
ulos; hojas de 0.5-1.7 cm lat, glabras a iliformes ia canindeyuensis 
pubérlas. ces i rer ee otis 29 Hojas planas, glabras o pubérulas; vaina 

22(21) Hojas pubérulas, secas discoloras, de haz estipular pubérula, con 3—5 lacinias; interior 
ferrugínea; estípulas con 7-10 lacinias pu- e la corola brevistila con pelos moniliformes 
bérulas a pubescentes; hipanto papiloso . n el tubo y en los lóbulos ............. 
o 16. G. tiliifolia 30(29). Planta de 0.80-1 m alt.; hojas con 3 pares de 

22. Hojas glabras, concoloras; estípulas con 3—5 nervios d. cápsula 4-6 mm long., 
lacinias glabras; hipanto pubescente ...... turbinada, pubéru lal illa alada con 
See eee eee eee 12. G. hassleriana el dorso convexo y cara ventral plana, con 

23(3).  Sufrútices erectos, decumbentes o apoyantes, estrofíolo persistente ........ G. kempffiana 
de 3-5-caule; vaina estipular de borde ir- 30. Planta de 20-25(-40) cm alt; hojas con 4-5 

regular, terminado en 3-5 lóbulos, rii 1 1 i lari ápsula 2.5-3 mm 
lanceolados; corola rosado-lilacina o blanque- long., subglobosa, pilosa; semilla lisa, subcilín- 
CMA rr c geriu rica con estrofíolo caduco ...... 23. G. parvula. 

23: Sufrútices erectos, de 5-20-caule; vaina estipu- 
lar de borde regular, terminadas en 3-6 lacinias 1. Galianthe angustifolia (Cham. & Schltdl.) 
filiformes; corola blanca ................- . L. Cabral, Bol. Soc. Argent. Bot. 27(3—4): 

(23). Hojas filiformes, lineares o lanceoladas de 239. 199] [1992]. Basónimo: Borreria angusti- 
(0.3—0.5-0.9(-3) mm lat. con nervios secu folia. Cham. € Schltdl., Linnaea 3: 330. 1828. 
darios inconspicuos ........ . G. peruviana TIPO: Brasil. Minas Gerais: Minas Gerais, 

ZR Hojas “Oyadas, a y prance we Pocos de Caldas, Cristo Redentor, 14 ene 

ipis s MUR ADM 1980, A. Krapovickas & C. Cristóbal 35308 
(24). Hipanto MEN ADR cáliz con segmentos (neotipo, designado aquí, SP!; isotipo, CTES!). 
linear-sul os o triangulares, reflexos en el Figura 2 
fruto; an secas con nervadura obscura 
contrastante con el color de la lamina del Sufrútice con xilopodio, tallos erectos, de 20- 
o 19. G. longisepala 70 cm alt, simples o con escasas ramificaciones 

25. Hipanto pubescente, cáliz con segmentos secundarias, glabros o pubérulos. Hojas 
triangulares, no reflexos en el fruto; hojas 1-5 mm, filiformes, lineares o linear-lanceoladas, 
secas, con per caduna del mismo color de la Po : 

ápice acuminado, base atenuada, margen recurvo, 
lámina del envés ........o.o.o.o..oo ooo... 2 ? 

26(25). Plantas muy uu hasta con 20 ejes glabras, o pubérulas en ambas pore o sólo en Al 
principales, de 12-30 cm alle ; hojas pubes envés, 2 o 3 (raro 4) pares de nervios secundarios, 

surcados en la haz y prominentes en el envés; vaina 
ne bic i Pes mm long.; powers pe "i estipular 3-4 mm, pubérula o pubescente, 5 6 7 
5 mm long., pubescente ...... 26. C. ramosa lacinias, 1.5-7 mm. Inflorescencia terminal. Hi- 

26. Plantas ramificadas hasta con 6 ejes princi- panto 1.2-1.5 mm, glabro, cáliz con lóbulos de 0.7— 
pales, e (20-)50-80 cm alt; hojas con 1.2 mm, bor i subulados, glabros; corola 3- 
indumento variado entre la haz y el envés, 4m perficie externa papilosa e interna con 
pubérulas o glabras en la haz y levemente ae a delgados en el tubo y gruesos en los 
pubescente o pubescentes en el envés; vaina NF sy 
sonado mm lone coria de 4 ain lóbulos; disto entero. Flor brevistila: corola, lóbulos 
A a ee 8 mm. dem iguales o más cortos que el tubo; estambres exertos, 

85 cap ng.» 2 : 
levemente pubescente ...... 46 a anteras 1 mm, filamentos 0.5-0.7 mm; estilo 2— 
21(1) Hojas con nervios secundarios inconspicu mm. Flor longistila: corola con lóbulos tan largos 
27. Hojas con 3-5 s de nervios e como el tubo; anteras subsésiles, 0.7—1 mm; estilo 


conspicuos 


3.7-4.2 mm. Cápsula 3-4 mm, subcilindrica, glabra; 


Annals of the 
Missouri Botanical Garden 


Gs iss 
G. fastigiata 


E dia 


G. guerenitica 


I4voe.o 


I+veo 


G. souzae 
G. thalictroides 


G. valerianoides 


Distribución de especies de la sección Galiant he. — 


Figura 1. 
cranes, G. chodatiana, G. cype 


montesii, G. parvul 
G valeriano des. 


a, G. peruviana, G. pseudopeciolata. 


semillas 2.5-2.7 mm, complanadas dorsiventralmente, 
aladas, ala más desarrollada en los extremos. 
Distribución, hábitat y fenología. Brasil, en el 
planalto central (Goiás, Mato Grosso, Minas Gerais, 
Rondonia y São Paulo); frecuente en campos cerrados, 


ana 
—F. Uisitibstión: de G. ramosa, G reiizii, G. souzae, G. ihaliciroids, 


1000-1600 m, con suelos rocosos, sujetos a quema- 
zones periódicas; florece de octubre a enero, fructifica 
de febrero a junio. 

Discusión. En el protólogo del basónimo Borreria 
angustifolia es mencionado un solo ejemplar, el que 


Volume 96, Number 1 
2009 


Cabral 33 
Revisión Sinóptica de Galianthe 


fue destruido en el herbario B, razón por la cual se 
elige el neotipo de Brasil, Minas Gerais, A. Krapo- 
vickas & C. Cristóbal 35308. 


Material rep tat tud, ASIL. Goiás: Serra 
Dourada, m ene. 1967, A. Duarte 10235 m 
Grosso: 165 da Rédowa, Cuiabá-Santarem, 20 jun. 
1979, M. pr et al. 5029 (MG). São Paulo: Sao ‘Bento do 
Sapucai, 13 abr. 1995, J. Tamashiro et al. 873 (SP) 


2. Galianthe canindeyuensis E. L. Cabral, Bon- 
plandia (Corrientes) 7: 8. 1993. TIPO: Paraguay. 
Canindeyti: Colonia Fortuna, 8 km de Curuguaty, 
6 mayo 1974, P. Arenas 662 (holotipo, CTES!; 
isotipos, BACP!, SI!). 


Distribución, hábitat y fenología. Paraguay (Amam- 
bay, Caaguazú, Canindeyú y San Pedro), en campos 
rocosos, 200—400 m; florece de 


diciembre a febrero, fructifica de marzo a junio. 


cerrados, arenosos, 


Observaciones. Es una especie de fácil reconoci- 
miento, por ser totalmente pubescente, muy ramifi- 
cada, de hojas plegado-nervosas, inflorescencia más o 
afin a Galianthe 
tranthoides, porque ambas especies son ramificadas, 


menos comprimida. Es cen- 


semillas complanadas, 
centranthoides) y las inflorescencias más o menos 
comprimidas (vs. inflorescencias laxas). 

Chodat y Hassler (1904) citan un único ejemplar 
Hassler 5839 para Paraguay, como Borreria eupatori- 
oides Cham. & Schltdl. [= Galianthe eupatorioides 
(Cham. & Schltdl. E. L. Cabral], pero ese material 


corresponde a G. canindeyuensis. 


Material representativo estudiado. PARAGUAY. Amam- 
bay: 30 km E de Pedro J. Caballero, feb. 1980, S. Tadashi 
275 (MO). Caaguazú: In viciniis Caaguazú, 1905, E. Hassler 
9156 (G, . San Pedro: Yaguareté forest, 23°48'38"S, 
56°07 00W, 20 jun. 1995, E. Zardini et al. 42813 (CTES, 
MO, PY). 


3. Galianthe centranthoides (Cham. & Schltdl.) E. 
L. Cabral, Bol. Soc. Argent. Bot. 27(3—4): 240. 1991 
[1992]. Basónimo: Borreria centranthoides Cham 
Schltdl., 27. 1828. T 
Brasilia meridionali pluries lectam misit, 1829, 
Sellow s.n. (holotipo, HB no visto; isotipos, G!, LE!). 


Linnaea 3: : Brasil. 


Borreria Mini d var. a Cham. & Schltdl., 
Lin IPO: Brasil Sellow s.n 
loti "o no visto; a LE?. 
Borreria mud var. angustifolia Cham. & Schltdl., 
naea 3: 330. 1828. TIPO: Brasil Sellow 4993 
holoti tipo, HB no visto; isotipo, LE!). 
Borreria pohliana DC., Prodr. 4: 550. 1830. TIPO: Brasil. In 
rasilia, 1828, Pohl s.n. (holotipo, HB no visto; isotipo, 
cn. 


TE clidemioides Griseb., Symb. Fl. Argent. 24: 157. 
permacoce clidemioides (Griseb.) Niederl., Bol. 
us Mus. Prod. Argent 
Ar, ree entina. Entre Ríos: Palmas Gare, s.d., P. Lorentz 
804 (holotipo, HB no visto; isotipos, CORD!, K 
Borreria centranthoides f. glabrior Chodat $ Hassl., 
Herb. Boissier, sér. 2, 4: 188. 1904. TI 
"ed prope San Estanislao, ago. 
po, uM ado a 
Ea centran. 
Her Boisaier sér. 2 188. 1904. TIPO: Paraguay. 
ao rupes in co libus s prope Paraguay, dic., E. Hassler 
6 (holotipo, G!; isotipos, NY!, P). 
Ee du ides f. pu escens Chodat & Hassl., Bull. 
Herb. Boissier, sér. 2, 4: 188. 1904. TIPO: Paraguay. In 
mpos d prope Valenzuela, ene., E. Hassler 
6976 (holo otipo, E isotipos, K!, MO! 

a f. angustifolia Chodat & Hassl., Bull. 
Herb. Boissier, sér. 2, 4: 188. 1904. TIPO: Paraguay. In 
uliginosis Cordillera de Altos, oct., E. Hassler 5314 
(holotipo, G!; isotipo, K!) 


Observaciones. Se caracteriza por ser un sufrütice 
pubescente, con xilopodio voluminoso hasta de 1 m 

ong. 1 a 4 tallos primarios con ramas secundarias 
desarrolladas; hojas elípticas, pubescentes, plegado- 
nervosas; inflorescencias terminales sólo en los tallos 
primarios, largamente pedunculadas, ejes y brácteas 
pubescentes; corola externamente pubescente e 
internamente con pelos densos, en el tubo y en los 
lóbulos; cápsula pubescente y semillas aladas, 


comprimidas dorsiventralmente, estrofíolo caduco, 
plano, adosado a la cara ventral. 

Galianthe centranthoides presenta variabilidad en la 
morfología, tamafio de las hojas y densidad de la 
pubescencia de t ió su 


lanta, que confund 
verdadera 1 la e 


EN razón por ual se incluyen 
varios ps y numerosos nombres subespecíficos. 

En medicina popular en el nordeste de Argentina, 
el cocimiento de las raíces, es usada como abortivo, 
diurético, purgativo, para ictericia, el asm 
Martínez Crovetto, 1981). En el sur de Brasil es 
conocida como remedio popular para los males de 


— 


ígado y vías urinarias, contiene resina odorífica, 
tanino catéquico, saponina, óleo volátil aromático, 
fitosterina (Lucas & Machado, 1944). 
“Guaycurú”, 


“sabugueirinho do 


Nombres vulgares. “raíz gaycurú”, 


499 66. 


ee rapó" "raíz charrüa", 


campo": de la Pefia y Pensiero (2004), y extraído de 


las observaciones de etiquetas. 


Distribución, hábitat y fenología. Nordeste de la 
Argentina E. Chaco, Entre Ríos, Formosa, 
Fe), sur de Brasil (Goiás, Minas 
ie Grande do Sul, Santa Catarina, 
São Paulo), Paraguay oriental (Alto Paraná, Amambay, 


Paraná, 


Caaguazú, Canindeyú, Central, Caazapá, Concepción, 


Guairá, Itapúa, Misiones, San Pedro, Paraguarí) y 


Uruguay (Artigas, Colonia, Florida, Lavalleja, Mal- 


34 Annals of the 
Missouri Botanical Garden 


Volume 96, Number 1 
2009 


Cabral 35 
Revisión Sinóptica de Galianthe 


donado, Rivera), en lugares modificados, entre la 
vegetación secundaria, en la orilla de los caminos y 
pos de 


sujetos a quemazones periódicas, 50-1 


preferentemente en cam suelos arenosos, 
m; florece 


de septiembre a diciembre, fructifica de enero a mayo. 


bn erial representativo ks Rd ARGEN pr Corr- 

: Santa Catalina, . Riachuelo, 2 dic. 2000, E. 

Cabral 667 (CTES). Entre "Rios: Concepcién del iie 

27 nov. 1878, P. Lorentz 1? ipd Formosa: mar. 1918, 
M ela 


4 E Parana: Mans 
pone Fda. S. L t 4 
(B, MBM). Santa Catarina: Araranguá, 7 dic. 194. 
869 C (CTES, RB). Sáo dae 1821, A. Saint-Hilaire 1472 
P). Rio Grande do Sul: Rio Guaribova, 20 e 909, P. 
Dusén 7534 (MO). PARAGUAY: Alto ee Prima- 
vera, 8 ene. 1961, A. Woolston 1241 (NY, US). Amambay: 
Sierra de Amambay, de b 9820 (F 
z 1905, E. Hassler 9213 


a 1955, G 


Cuchilla de la Ballena, 14 nov. 1899, C. Osten 3958 (MVM). 


4. Galianthe chodatiana (Standl) E. L. Cabral, 
Bol. Soc. Argent. Bot. 27(3-4): 242. 1991 bs 
Basónimo: Borreria chodatiana Standl., 

Field Columbian Mus., Bot 

TIPO: Paraguay. Sierra de TR) E. Hassler 
5165 (holotipo, G!; isotipos, BM!, F!, G!, K!, P^). 
Figura 4 


Borreria thaliciroides var. latifolia Chodat € Hassl., Bull. 
Herb. Boissier, sér. 2, 4: 189. 1904. TIPO: ws s In 
campo Ipe hu (Sierra de Maracayú), oct, E. Has 

5168 (holotipo, G!; ee FY, Kt, NY!) 


Sufrútice con xilopodio, de 0.30-0.60(-1) m alt., 
1- 6 5-caule, tallos ooo glabros, raro udis 
dispersos. Hojas 5 elípticas, 
ápice M o atenuado, base 


largamente aguda en 


elíptico-oblongas, 
pseudopecíolo, discoloras, más 
oscuras en la haz, eb pubescentes 0 con 
pelos dispersos sobre los ios del envés, con 3(4) 
pares de nervios secundarios idis en el envés; 
vaina estipular con una breve d iM por 
encima de la separación del p ojas, de 

5 mm, pilosa o pubescente, con e 67 fusi 1.5-6 
(-10) mm, glabras. Inflorescencias tirsoideas termi- 
nales, sólo en los tallos primarios, largamente 
pedunculadas, 20-35 em. Hipanto 2-2.3 mm, turbi- 
nado, papiloso, con escasos pelos dispersos, lóbulos 
del cáliz triangular-subulados, glabros; 
corola 4.5—5 mm, longitud de lóbul 

tubo; disco bipartido. 


1-1.5 mm, 
os + igual que el 
Flor brevistila: superficie 
de los 
lóbulos y en el tercio inferior del tubo; anteras 1.2— 
1.5 mm, filamentos 1.5-2.3 mm; estilo 2.3-2.5 mm. 
Flor longistila: superficie interna de la corola con 


interna de la corola con pelos en la base 


pelos gruesos en los lóbulos y en anillo denso de pelos 

más finos en la mitad del tubo, anteras subsésiles, 
1.3-1.5 mm; estilo 3-5 mm. Cápsula 2.5—4.3 mm, 

glabra; semillas .5 mm, ee de borde 
alado, manifiesto en las polos 


Distribución, En Brasil, 


Paraná y localidades vecinas de Santa Catarina; 
frecuente en campos 


hábitat y fenología. 


cos o levemente húmedos 
de 700-900 m. De Paraguay se conoce hasta el 
momento, sólo las colecciones tipos de Hassler, que 
proceden del depto. Canindeyú; florece desde octu- 
bre, fructifica en enero y febrero. 


Material representativo estudiado. BRASIL. Paraná: 
un. Curitiba, Instituto de Biologia, 13 ene. 1966, J. C. 
Lindeman et al. 309 (CTES, RB). Santa 
Ere, 22 feb. 1964, A. Castellanos 24690 (RB 


Catarina: Campo 


5. Galianthe cyperoides (Chodat & Hassl.) E. L. 
Cabral, Bol. Soc. is id Bot. 27(3-4): 241. 1991 
1992]. Basónimo: Borreria cyperoides s & 
Hassl., Bull. Herb. Coase, sér. 2, 4 1904. 
TIPO: Paraguay. In campo Ape eee 

ago., E. Hassler 4338 ae designado por 


Figura 2. Galianthe angustifolia. —A. Planta. 
Hipanto, cáliz, estilo y estigm 
estilo y estigma. —I. Interior de la corola despleg; 
L, Krapovickas 35308; G-I, Hauff 20.) 
Figura 3. Galianthe cyperoides. —A. Planta. 


a. —]. Fru 


estilo y e; —H. Interior de la corola despleg; 
dorsa 


—B-C. Vaina estipular con lacinias. D—F. Flor ND —D. Flor. —E. 
ma. —F. Interior de > corola ip ce G-I. Flor brevistila. —G. F 


—H. Hipanto, cáliz, 
J- 


o. K-L. Semilla. —K. Cara ventral. a Cara dorsal. (A-F, 


—B. Vaina estipular con lacinias. C-E. Flor brevistila. —C. 
Hipanto, > estilo y estigma. —E. Interior de la corola desplegada. F-I. e rime —F. Flor 
ada. —I. Alabastro. —J. F: 
l; ^W pom transversal. (A—B, F-I, Masa 9028; C-E, Hassler 4338: T M Schinini 33309.) 


—G. = ae sliz 
. Semilla. —K. Cara ventral. —L. Car. 


36 


Annals of the 
Missouri Botanical Garden 


Cabral [199]: 241], Gt 
Figura 3. 


isotipos, Kl, PP). 


irs leiophylla var. expansa Chodat & Hassl., Bull. Herb. 
oissier, sér. 2, 4: 187. 1904. TIPO: Paraguay. In 
campo prop FU oct., E. Hassler 4829 (holotipo, 

G!; isotipos, GLK !, MO!, NY!, P!, US). 

Sufrútice con xilopodio de 0.40—1.50 m alt., tallos 
simples o escasamente ramificados, tetrágonos, gla- 
bros, entrenudos 1-7.5 cm, pseudoverticiladas. Hojas 
15-70 X 1-5(-7) mm, lineares o linear-lanceoladas, 

bras, con nervios secundarios i UN paa vaina 

1-2.5 pubérula, con 3 6 


lacinias de 2-10 mm. ANA ardido, amplia, 


gla 
estipular 1-2.5 mm, glabra 


laxa, a 17-28 cm. Flores con hipanto 1— 
1.5 mm, turbinado, glabro; cáliz con lóbulos triangu- 
lar- ondes 1-1.5 mm; corola externamente con 
papilas notables, densas, en el ápice dorsal de los 
el tubo; 


lóbulos, lóbulos iguales o más cortos que 
o. Flor eec: corola 


disco entero, papiloso. 
7.5 mm, superficie interna con pelos moniliformes, 
cortos, en la mitad del tubo y pelos dispersos en la 
mitad inferior de los lóbulos; anteras 1—1.2 mm, 
filamentos 1.5-2.3 mm; estilo 2-3.7 mm. Flor longi- 
stila: corola 5-6.2 mm, superficie interna con anillo 
cortos en el tubo 


e pelos moniliformes, delgados y 


arcos de pelos moniliformes en los lóbulos; anteras 1— 
0.7 mm; estilo 4-6 mm 


b 3-3.5 mm, subelipsoide, 


1.2 mm, filamentos ca. 


convexo y cara ventral plana, con surco duds 
del estrofiolo persistente. 


Distribución, hábitat y fenología. Paraguay (Caag- 


uazü, Canindeyü, Guairá, San Pedro) en campos 
altos, cerrados, 200—400 m, con frecuencia asociada a 


odr.) L. H. 


palmares de Butia paraguayensis (Barb. 
) Becc.; florece y fructifica de 


Bailey y B. yatay (Mart. 
septiembre a mayo. 


Material representativo. estudiado. 


PARAGUAY. Caag- 


0 (AS, CTES, MO). 
Canindeyú 46 km S de Katueté, 3 Pe N del rio Itambery, 

Stro ea del Guairá, 18 dic. 1982, A. 
Shinn 23212 (CT E Guairá: Cerro de Villarrica, Cerro 


Santa Rosa & San Barban ra, 


56° 23 39"W, E. Zardini et al. 15560 (AS, CTES, MO). 


6. Galianthe elegans E. L. Cabral, Bonplandia 
(Corrientes) 7: 10. 1993. TIPO: Brasil. Paraná: 
Vila os en campo O de la Iglesia, 15 

puli rapovickas & C. Cristóbal 40875 

a MBM!; a CTES!, MO!, SI). 


Observaciones. Se caracteriza por ser un sufrútice 
erecto de 60-65 cm 


elípticas glabras; vaina estipular con 3-4 lacinias; 


alt., 1—5-tallos glabros; con hojas 


inflorescencia largamente pedunculada; hipanto, cáliz 
y corola externamente micropapilado; disco entero; 
cápsula glabra y semillas complanadas con alas 
apicales. Por el porte Galianthe elegans es semejante 
a G. chodatiana, pero se diferencian porque G. elegans 
tiene hojas —6 mm, aoe nervios 
secundarios inconspicuos (vs. 304, 15 mm 
escabritisculas, 3—4 nervios paris conspicuos, 


G. chodatiana). 


Distribución, hábitat y fenología. Brasil, Paraná, 
todos los ejemplares conocidos son de Ponta Grossa y 
alrededores, en campos con suelo arenoso, pedregoso, 


5 O m; florece y fructifica de octubre a enero. 


Paraná: 


ASIL. 
1968, H. Moreira Filho et al. 472 


Material C estudiado. BR 
Ponta Grossa, 18 o 
(CTES, UPCB, US 


7. Galianthe equisetoides (Cham. & Schltdl.) E. L. 
Cabral, Bol. pou Argent. Bot. 27(3—4): 242. 1991 


chltdl.) Kuntze, Revis. 
Gen. Pl. 3: 123. 189 hipo Brasil. e Nea 
do Sul: Campo dos Barcelos, 9 nov. 8, 
Abruzzi 1644 (neotipo, designado aquí, m 
isotipo, CTES!). 


Distribución, hábitat y fenología. Brasil (Paraná y 
Rio Grande do Sul) y norte de Argentina (Corrientes, en 
el Sistema Iberá y lugares A al río Uruguay); en 
campos bajos e inundables, 000 m; florece de 
octubre a febrero, fructifica m marzo a junio (Borreria 


equisetoides, Cabral, 1981: fig. 2). 


Observaciones. Es una especie de facil reconoci- 
miento por ser un sufrútice de 0.60-1.50 m, de 
aspecto equisetoide, con tallo fistuloso, cilíndrico o 
subcilindrico, glabro, constricto en los nudos; con 
hojas 2—4 por verticilo, cartáceas, filiformes, lineares, 
nervios secundarios inconspicuos o 2-3 visibles en el 
envés; inflorescencia terminal, amplia; hipanto y cáliz 
glabro o pubérulo, corola externamente papilosa, 
blanca o blanco-lilácea; cápsula glabra o pubérula, 
con semillas subcilindricas. Los ejemplares de Brasil 
presentan tallos más robustos y hojas más anchas de 
los que viven en Argentina. 

El único ejemplar citado en el protologo del 
basónimo Borreria equisetoides se ha destruido en el 
herbario B, razón por la cual se elige un neotipo 
coleccionado en Brasil, Rio Grande do Su L 
Abruzzi 1644. 


Nombres vulgares. En Brasil “sabugueirinho”, 


“baicurú” (Porto et al., 1 


Material representativo estudiado. ARGENTINA. Corr- 
ientes: Ituzaingó, 11 km S de Ruta 12, desvío a Gdor. 


Volume 96, Number 1 
2009 


Cabral 37 
Revisión Sinóptica de Galianthe 


Virasoro, 29 nov. 1970, A. Krapovickas et al. 16568 (CTES, 
ASIL. Paraná: Volta Grande, dic. 1979, E. Oliveira 

i nde do Sul: S ía, Reserva 
Biológica do Ibicuí-Mirim, Barragem de Saturnino; 9 nov 
1988, N. Silveira 5929 (CTES, HAS). 


8. Galianthe fastigiata Griseb., Symb. Fl. Argent. 


Borreria fastigiata (Griseb.) K. Schum., Fl. E 
6(6): " 1888. Spermacoce fastigiata 
(Griseb.) Niederl. Bol. Mens. Mus. Prod. Argent. 
3(31): 306. 1890. TIPO: Argentina. Entre Ríos: 
Palmar grande, 3 feb. 1876, P. G. Lorentz 803 
(holotipo, HB no visto; isotipo, CORD!). 


Borreria leiophylla K. Schum., Fl. Bras. (Martius) 6(6): 66. 
188 be mp ue (K. Schum.) Kuntze, 
123. a TIPO: Brasil. Brasilia 
australi, in provincia Rio Grande do Sul, Joannes de S. 
Barbara, Sellow 1570 a designado aqui, F!, 
foto F 879 
Distribución, hábitat y fenología. Norte de Argen- 
tina a (Corrientes, Entre Ríos, Misiones), sur de Brasil 
rande do Sul, Santa Catarina), Paraguay 
oriental (Alto Paraná, Amambay, Yi e Caazapá, 
y (Cane- 
lones, Cerro Largo, Florida, Lavalleja, “Maldonado, 
Rocha, 


campos de suelos arenosos, rocosos, en palmares de 


Cordillera, Misiones, Paraguarí) y Uru 


Rivera, Tacuarembó, Treinta y Tres); en 
Butia y bordes de caminos, 0-1000 m; florece de 


octubre a febrero, fructifica de marzo a junio. 


Observaciones. El carácter sobresaliente de esta 


especie es su tallo simple de 0.25-1.25 m alt., con 
hojas opuestas, lanceoladas o lineares, Eun 
nervios secundarios inconspicuos; hipanto, cáliz y 
corola externamente glabros, interior de la corola con 
pelos densos en tubo y lóbulos; cápsula glabra con 
semillas subcilíndricas con estrofíolo persistente. Es 

fin a Galianthe equisetoides por los tallos simples, de 
hojas opuestas, pero se diferencia porque C. fastigiata 
tallos con nudos 
as de 2-25 mm lat., 
nudos inferiores ve hojas 2-7(— 13) 
mm lat. porci 


no tiene nudos constrictos (vs. 
constrictos en G. equisetoides), hoj 
caducas en los 


Spermacoce fastigiata (Griseb.) Kuntze (Kuntze, 
1898) es nom. illeg. 
usa en medicina popular contra inflamaciones 
hepaticas (Bacigalupo, 19 
Nombres dd En Brasil “sabugueirinho do 
campo” (Porto et al., 1977), en Argentina “sauquito” 
(de la Pefia & e 2004). 


Material (i can estudiado. 
ientes: ltuz 
1980, E. Cabral 159 (CTES). Entre Ríos: 
dic. 1957, A. L. Cabrera 12370 (CTES, SD. 
San Ignacio, 18 dic. 1981, E. Cabral et al. 181 (CTES). 


ARGENTINA. sor 
2 feb. 


BRASIL. Rio Grande do Sul: Morro Sapucaia, p. São 
Leopoldo, 3 feb. 1956, B. Rambo 59164 (H 
Catarina: Araranguá, Morro dos C 
M. Detoni 65 (ICN). PARAGUAY. Alto Paraná: 1910, K. 
Fiebrig 6491 (G). meee Cnia. Pedro J. pu 
CTES). c 55720 

ntre Itaquyry y Curuguati, i 
| oun et d "20098 (CTES). Cordillera: 
a jul. G. nO 11401 (CTES). 
Paraguarí: Tebicu nov. 1978, i 
18735 (MO). URUGUAY. Canelones: Puerto Jackson, R. 
Herter 86625 (G, MVM); río Santa Lucía, 22 feb. 1946, G. 
Gallinal et al. 5610 (SP). Cerro Largo: 18 e 908, G. 
Flossdorf 2 (BAF). Florida: Ayo. Meal dis 1946, 
B. Rosengurit et al. 5835 (SP). Lavalleja: Minas, feb. 1874, 
O. Gibert s.n. (BAF). Maldonado: Rincón de Minas, 23 dic. 
1928, C. Osten 20188 (BAF, MVM). Rocha: Cerca de la 
Sierra de las Rochas, 18 ene. 1965, O. Brescia et al. 3957 

A) 


oct. 1995, 
Valenzuela, 


E 
= 


9. Gali L. Cabral, Bonplandia 

eres row hs “13. 1993. TIPO: Brasil. 
Paraná: Mun. Campina Grande do Sul, Serra 
Ibitiraquire, 22 ene. 1970, G. Hatschbach 23388 
(holotipo, MBM!; isotipo, RB!). 


Distribución, hábitat y fenología. Brasil, Paraná y 

localidades próximas del estado de Santa Catarina, en 

campos altos, o en laderas rocosas de 1000-1700 m; 
fru 


ctifica de marzo a abrıl. 


Observaciones. EE geri se individualiza 


por la vaina estipu e borde irregular terminado en 
3-5 lóbulos E lanceolados y también por las 
inflorescencias tirsoides umbeliformes, paucifloras de 
flores rosado-lilacinas o blanquecinas. 

Esta especie se dedicó al botánico Gert Hatsch- 
bach, quien ha contribuido permanentemente con sus 
colecciones al conocimiento de las especies de 


Galianthe de Brasil. 


Material representativo estudiado. BRASIL. Paraná: 
Mun. i RR S 
O. S. Ribas 
do lucem, ure Alea 5 feb. 1958, R. Reitz & a 


6426 (B, HBR, MBM, US). 


10. eras grandifolia E. L. Cabral, Bonplandia 
: 14. 1993. . Brasil. 
os ca. 20 orinto, 3 mar. q 
S. Irwin et al. 20100 (holotipo, RB!; isotipos, F!, 
MO!, NY). 


Observaciones. Galianthe grandiflora se reconoce 
fácilmente por ser un sufrútice erecto con tallo de 


0.60-1.80 m alt., 


lanceoladas. La inflorescencia es terminal y larga- 


y con hojas elíptico-lanceoladas o 


mente pedunculada, las cápsulas pubescentes y las 
semillas complanadas aladas. 


plegado-nervosas, pero 


simples pubérulos (vs. tallos ramificados, pubes- 


38 


Annals of the 
Missouri Botanical Garden 


Volume 96, Number 1 
2009 


Cabral 39 
Revisión Sinóptica de Galianthe 


centes, G. centranthoides), hojas 40-105 X 540 mm, 
haz ~ ra o pubérula, envés pubescente (vs. hojas 


30-70 


X 3-28 mm, pubescentes). 

Distribución, hábitat y fenología. Brasil (Bahía, 
Distrito Federal, Goiás, Mato Grosso, Minas 
Pará, São Pau 
525 y 1300 m, rupestres, de suelos areno-pedregosos 


Gerais, 
o, Tocantins}; en campos cerrados entre 


con afloramientos cuarzíticos; florece de septiembre a 
febrero, fructifica de marzo a mayo. 


Material E C estudiado. BRASIL. Bahia: Near 
Rio Piau, 150 km of Barreiras, 14 abr. 1966, H. S. Irwin 
et al. 14782 (MO). dicia Federal: Fda. Agua Limpia, 13 
abr. 1976, J. A. Ratter et al. 2881 (UEC). Goiás: 1846, G. 
Gardner 3785 (G); Mun. Cristalina, 4 feb. 1987, J. Pirani et 
al. 1490 (CTES, MBM). Mato Grosso: Serra do Roncador, 
24 m S. Irwin et al. 15952 is: 


mayo 1979, H. C. de ru et al. 1 B). E 0°95'S, 
4°92'W, 4 ago. 1981, J. Strudwick et al. 4055 (IAN). São 
Paulo: Itirapina, feb. de iu 2517 (SP). 


Tocantins: Palmas, Serra do Lageado; 13 abr. 1994, A. E. 
Ramos et al. 646 (CTE 


11. Galianthe guaranitica (Chodat & Hassl.) E. L. 
Cabral, Bol. Soc. Argent. Bot. 27(3-4): 244. 
1991 [1992]. Basémmo: Borreria 
Chodat & Hassl, Bull. Herb. Boissier, sér. 2, 4: 
186. 1904. TIPO: Paraguay. In campo Ipe hi, Sierra 
de Maracayú, Dec., E. Hassler 5594 (holotipo, G!, 
foto F 6919; isotipos, Kt, NY). Figura 5 


0.5-1.5 m alt., tallos 


mente ramificados, tetrágonos, gla- 


guaranitica 


Sufrútice con xilopodio de 
simples o escasa 
bros o pubescentes, entrenudos (324—8 cm, pseudo- 
verticilados. Hojas (3040-50 X (1217-27 mm, 
elípticas, ápice de agudo a atenuado, base redondeada 
o levemente cordada, subglabras, con escasos pelos 
66 


pares de nervios secundarios, en relieve en el envés; 


dispersos sobre los nervios en el envés, con 5 


vaina estipular prolongada por encima de la separa- 
ción del par de hojas, 4-5 mm, pubérula, con 7-8 
m. Inflorescencias tirsoidea 


lacinias, de 2-6(- s 


m 
terminales, = congestas, de urifloras 
Hipanto ca. 1.5 mm, turbinado, glabro o pubérulo; 
cáliz con lóbulos de 1.5-2 mm, triangular-subulados, 


del 


largo, con notables papilas sobre el dorso de los 


glabros; corola 5.5-7 mm, lóbulos y tubo + mismo 


lóbulos, característica muy conspicua en los alabas- 
tros; disco entero, papiloso. Flor brevistila: corola, 


superficie interna con anillo de pelos moniliformes en 
l tu 


o 
superior de los lóbulos; anteras 1.2—1.3 mm. 


n] 


y pelos moniliformes más largos en la mitad 
, filamen- 
tos ca. 1.5 mm; estilo ca. 0.5 mm. Flor longistila: 
corola, superficie interna con pelos moniliformes en el 
tubo y escasos pelos moniliformes en la base de los 
lóbulos; anteras subsésiles, 1.2-1.5 mm; estilo ca. 
5 mm long. Fruto no visto. 

Distribución, hábitat y  fenología. 
(Amambay y Canindeyú [Cabral, 


Paraguay 


campos cerrados con suelos arenosos, graminosos, 
sujetos a quemazones periódicas, 500—700 m; florece 
y fructifica de diciembre a abril. 


Material — representativo estudiado. BRASIL. Mato 
Grosso do Sul: Mun. Sidrolandia, Santa Fe, 23 ene. 1971, 
G. d 26028 (CTES, MBM, NY, US). PARAGUAY. 
Amambay: E Estrella, 45 2 NW de P. J. Caballero, 
22° 185, 55750'W, 8 dic Schinini et al. 33581 
(CTES). Canin iy Entre Ype- ge Capitán Bado, a 10 km de 
Itanará, 5 feb. 1982, J. Fernández Casas et al. 5983 (MO, NY). 


2. Galianthe hassleriana (Chodat) E. L. Cabral 

Bol. Soc. Argent. Bot. 27(3—4): 244. 1991 [1992]. 
Bull. 
. Borreria 


Bull. 


Basónimo: Borreria hassleriana Chodat, 
. Boissier, sér. 

e Chodat E latifolia Chodar 
Herb. Boissier, sér. 2, 4: 189. 1904 “TIPO 
Paraguay. In campis prope flumen one 
ep., E. Hassler 4562 (lectotipo, designado aquí, 
Gl; isotipos, K!, NY!, P!, foto F 6918!). Figura 6. 
Borreria hassleriana Chodat f. C E Chodat, Bull. Herb. 
Boissier, sér. 2, 4: 188. 1 TIPO: Paraguay. In camp 
e flumen Jejuiguazú, Eid E. Hassler 5689 (holotipo, 

GI; isotipos, Kt, NY?) 


Sufrútice con xilopodio de 35-40 cm alt., tallos 
simples, glabros o pubérulos, entrenudos 5-10 cm 


long. Hojas 40-75 X 


elíptico-lanceoladas o lanceoladas, = agudo y 


5-17 mm, pseudoverticiladas, 
base atenuada, glabras, con 2 6 es de nervios 


secun en el envés; vaina estipular de 
3—4 mm, pubescente o pubérula, con pelos más largos 


hacia el borde, con 3-5 lacinias de (47-10 mm. 


Inflorescencia tirsoide, terminal. Hipanto 1-1.5 mm, 


arios marcados 


pubescente; cáliz con lóbulos de 1 mm, triangular- 
acuminados, pubérulos, pubescentes o con pelos 


dispersos; corola 5-6 mm, papilosa, pubérula o 


ura 4. ‘sees chodatiana. 
Hipanto, cáliz, estilo y 


Semilla. —L. Cara ventral. —M. Cara dorsal. (A, B, G— 
Figura 5. es iani nihe guaranitica. —A. Planta. — 
—D. Flor o, cáliz, estilo y estigm 

Interior de m ed Fc ada. 


—A. Planta con xilopodio 
r. estigma. —F. Interior de la corola 
—H. Tor E Hipanto, cáliz, disco, ARE y estigma. 


—F. Interior de la corola desplegada. 
—I. Hipanto, "liz, stale y estigma. 


. —B. Vaina estipular con lacinias. C—F. Flor longistila. —C. 
a desplegada. G-J. Flor brevistila. — 
—J. a de la corola desplegada. —K. Fruto. 
F, Cordeiro 890.) 
acinias. C—F. Flor E —C. Alabastro. 
G-I. Flor brevistila. —G. Flor. —H. 
(A-F, Hassler 5594; G-I, Schinini 33496.) 


3 


L-M. 


40 


Annals of the 
Missouri Botanical Garden 


pubescente, papilas densas en el ápice de los lóbulos, 
lóbulos más cortos que el tubo. Flor brevistila: corola, 
superficie interna con anillo de pelos moniliformes 
finos y cortos en la mitad del tubo; anteras ca 
filamentos ca. 1.5 mm; estilo 3-3.5 mm. Flor longi. 
stila: corola, superficie interna con anillo de pelos 
moniliformes, finos y cortos en la mitad del tubo y 
pelos más gruesos en la mitad de los lóbulos; anteras 
subsésiles, ca. 1 mm; estilo 3.5-4.5 mm. Cápsula 
2.5-2.8 mm, subelipsoide, pubescente. Semillas no 
aladas, ca. 2 mm, estrofíolo persistente. 


Distribución, hábitat y fenología. Paraguay, San 
Pedro; en campos próximos a los ríos Carimbatay y 
Jejui-Guazá, 190-250 m; florece y fructifica de 
septiembre a enero. 


Chodat  (Chodat & Hassler, 
1904: 189) describió dos formas para el basónimo 


Observaciones. 


Borreria hassleriana, que aquí se consideran sinóni- 
mos de Galianthe hassleriana. En el protologo de 
hassleriana f. latifolia mencionó el ejemplar “In 
campis prope flumen Carimbatay, Sept. E. Hassler 
4562". En el protologo de B. hassleriana f. angusti- 
Jolia citó el oc “In rope flumen 
r 5689” Selecciono como 
lectotipo E. Hassler 4562, por ser un material bien 
representativo en G y con isotipos vistos en los 


herbarios K, NY, P. 
13. d E. L. Cabral, Brittonia 
P 


olivia. Santa 


Killeen, H. González, F. Mama 
(holotipo, USZ!; isotipos, CTES!, MO!, SI). 


Distribución, hábitat y fenología. Bolivia (Santa 
Cruz), muy frecuente en el Parque Nacional Noel 
Kempff Mercado. Brasil (Goiás y Mato Grosso); en 
campos rupestres de las chapadas de 360-950 m; 


florece y ino de marzo a junio. 


Observaciones. Se reconoce por ser un sufrútice 
de 80-100 em alt., 


undarias desarrolladas que rematan en inflorescen- 


muy ramificado con ramas sec- 


cias congestas y por tener sus hojas con 3 pares de 
nervios secundarios conspicuos; las semillas son 
aladas con estrofíolo persistente. Por las ramifica- 
s y ubicación B las inflorescencias es afín a 
Colonia parvula E. L. Cabral, pero esta especie es 
una de las de menor altura, 20-25(40) cm, las hojas 
tienen 4—5 pares de nervios secundarios, las semillas 
no son aladas y tienen el estrofíolo caduco. 
Material (rien estudiado. E Santa 


Cruz: Velazco, Parque Nacional Noel Kempff M., serranía 
de Hietuchaens 9 jun. 1994 B. Minds et al. 2160 (MO, 


USZ). BRASIL. Goiás: be Dourada, 16 km S of Goiás 
s 950 m s.m., 11 mayo 1973, W. R. Anderson 10109 (F, 

Y, UB). Mato Grosso: ro da Chapada, 19 feb. 1903, G. 
Mis 3446 (F). 


14. Galianthe lanceifolia E. L. Cabral, Bol. Soc. 
Argent. Bot. 2 5 vido TIPO: Brasil. 
Mato Gro 


: 1 km oa Agu 
Quentes, p ene. n bu Kms & c. 
Cristóbal 43/55 (holotipo, MBM!;  isotipo, 
CTES!) 


Distribución, hábitat y fenología. Sur de Brasil 
(Goiás, Mato Grosso do Sul, Mato Grosso, Minas 
Gerais), en cerrado, en suelos arenosos o pedregosos, 
200-1000 m; florece y fructifica de diciembre a julio. 


Observaciones. Galianthe lanceifolia se reconoce 
por ser pluricaule, con hojas lanceoladas, inflore- 
scencia en todas las ramas, pero se individualiza 
cuando fructifica, porque los frutos son subglobosos, 
notables. Presenta variabilidad en el indumento de 
hojas, tallo, ejes de la inflorescencia, hipanto y 
segmentos del cáliz, se observan ejemplares com- 
pletamente glabros hasta pubescentes, y otros con 
ind o intermedio. Se asemeja a eupatori- 
oides, pero G. lanceifolia tiene ipod (vs. sin 
xilopodio, G. eupatorioides), dis ectarifero entero 
pubescente (vs. disco bilóbado papilado), interior de 
las flores brevistilas con un anillo de pelos en la mitad 
del tubo y algunos pelos dispersos en los lóbulos (vs. 
pelos densos en el tubo hasta la base de los lóbulos). 


Nombre vulgar. “Aroeirinha” (observaciones de 


etiquetas). 


, Material representativo estudiado. BRASIL. Goiás: Lu- 
Mar. zaj oara, 22 dic. 1990, F. Melo et a 

. do Rie Novo Arinos, feb. 1914, J. 

G. Kuhlmann 5.n. (SP 11818). vnu py" do Sul: Campo 

Grande, 16 jul. 1966, R. Goodlan 1 (MO, NY). Minas 

Gerais: BR 050, 15 km SE de er 29 ene. 1990, M. M. 

Arbo et al. 3038 (CTES, HRCB). 


15. Galianthe latistipula E. L. Cabral, Bonplandia 
(Corrientes) 7(1—4): 18. 1993. TIPO: Brasil. Rio 
Grande do Sul: Vila Oliva, prope Caixas, 8 feb. 
1955, B. Rambo s.n. (holotipo, PACA 566611; 
isotipos, B!, CTES!, SID. 


Distribución, hábitat y fenología. Sur de Brasil 
(Rio Grande do Sul y sur de Santa Catarina), en el 
planalto meridional (Fernandes & Bezerra, 1990), en 
campos altos, rocosos de 1000-1800 m, florece y 
fructifica de diciembre a marzo. 


Observaciones. Se individualiza por ser esencial- 
mente monocaule, con pseudoverticilos regularmente 
dispuestos, con hojas de similar tamaño y por la vaina 
estipular prolongada por encima de la separación del 
par de hojas. Es semejante a Galianthe fastigiata, por 


Volume 96, Number 1 
2009 


ral 41 


Cab 
Revisión Sinóptica de Galianthe 


el tallo simple, y hojas glabras, pero se diferencia 


s iguales en 
semillas complanadas con alas apicales (vs. semillas 
subcilindricas, ápteras). 


Material tudiado. BRASIL. Rio Grande 
do Sul: Due 13 feb. 1951, B. Rambo 50047 (B, CTES, 
LIL, PACA). atarina: Campos Novos, 31 ene. 1963 
R. Reitz 6416 (HBR, US). 


16. Galianthe liliifolia (Standl.) E. L. Cabral, Bol. 
Soc. Argent. Bot. 27(3-4): 245. 1991 [1992]. 
Basónimo: Borreria liliifolia Standl., Publ. Field 
Columbian Mus., Bot. Ser. 8(5) 392. 1931. 
TIPO: Brasil. Sáo Paulo: Ipiranga, 31 de 1911, 
Alex Brade 5266 (holotipo, S!; isotipos, F!, SP). 
Figura 7. 


Sufrútice con xilopodio, 50-60 em alt., 1- ó 2- 
caules, sin ramas secundarias desarrolladas, tallo 
tetrágono o subtetrágono, pubérulo. Hojas 20-65(-90) 
X 5-15 mm, pseudoverticilas, oblongo-lanceoladas, 
de ápice agudo a atenuado y base de aguda a obtusa, 
ie aia haz ferrugínea cuando secas 

vios secundarios — paralelos, impresos 
en el envés, uc en la haz; vaina estipular de 
3-6 mm, pubérula, con 7-10 lacinias pubérulas, 
de 4—12 mm. Inflorescencia terminal amplia. Hipanto 
1-2 mm, turbinado, papilo 
lóbulos de 1-2 mm, 
lóbulos más largos que el tubo, internamente con 
disco 


o pubérulo; cáliz con 
MM dC corola 
anillo de pelos moniliformes en el tubo; 
entero. Flor brevistila: corola ca. 3.5 mm; anteras 
0.6-1 mm, filamentos ca. 1 mm; estilo ca. 1.5 mm. 
Flor longistila: corola 2.5-3 mm; anteras subsésiles, 
ca. 1 mm; estilo ca. 2.5 mm. Cápsula 2.5-3 mm, 
papilosa; semillas 2-2.5 mm, plano convexas, 
reducidas en 


cubriendo la cara ventral. 


las 
el Apice, con |-estelitló persistente, 


Distribución, hábitat e Centro y sur de 
Brasil (Distrito Federal, 
planalto central, en el cerrado o terrenos modificados 


al borde de caminos, de 900-1370 m; florece de 


noviembre a enero, fructifica de febrero a abril. 


Minas Gerais y Sáo Paulo), 


Observaciones. Galianthe liliifolia se reconoce 
fácilmente por sus tallos simples, pubérulos, ferrugí- 

s, por sus hojas pseudoverticiladas, oblongo- 
lanceoladas, pubérulas, marcadamente discoloras y 
por su amplia inflorescencia terminal. 


Material representativo estudiado. BRASIL. Distrito 
Leda Santa, 3 

n. (P). Paulo: Itapetininga, 1-1960, S. M. 
Campos 148 (C, NY, SP, US). 


7. Galianthe linearifolia E. L. Cabral, Bonplandia 
(Corrientes) 7(1—4): 20. 1993. TIPO: Paraguay. 
to Paraná: Ea. Santa Elena, Pira Pyta, 
54^35'W, 25^17'S, 11 oct. 1990, A. Schinini & 
G. Caballero y iia 27227 (holotipo, CTES; 
isotipos, G!, M 


Distribución, hábitat y fenologia. Paraguay (Alto 


Paraná), poco frecuente en campos altos, 400—500 m, con 


colección reciente. Con respecto al material 
tina, corresponde a una colección principios del siglo XX 
en la provincia de Misiones, y atin no se ha vuelto a 
encontrar; florece y fructifica de septiembre a enero. 


Galianthe 
nombre lo indica se reconoce por ser un sufrütice 


Observaciones. linearifolia como su 


ramificado, formando matas, con hojas lineares, 
pseudoverticiladas, cuando secas son verde-amari- 
llentas y por su inflorescencia apical, largamente 
pedunculada, los frutos son glabros con semillas 
complanadas y aladas en los ápices. Por las ramas 
secundarias bns con braquiblastos muy reduci- 
s se parece a G. tha Erol (K. Schum.) E. L 
Cabral, pero G linearifolia tiene hojas pubescentes de 
5-10 X 0.2-2 mm (vs. glabras, (8-)20-35 x 0.7- 
n G. thalictroides), interior de corola de flores 
brevistilas con pelos en el tubo (vs. pelos en el tubo y 
en los lóbulos de corola brevistila). 
Material representativo estudiado. ARGENTINA. Misi- 


Azara, Sp ime s.n. (LP). PARA- 
á: Tatí Yu upí, 24 sep. 1980, G. Caballero 


ones: Apóst v de 
Marmori 860 (CTES 


18. Galianthe longifolia (Standl.) E. L. Cabral, Bol. 
Soc. Argent. Bot. 27(3-4): 245. 1991 [1992]. 
Basónimo: Pd. thalictroides var. longifolia 

tandl., . Field Columbian Mus., Bot. Ser 
8(5): Pd m TIPO: Brasil. Paraná: Serriha, in 


campo, 840 m, 7 dic. 1908, P. Dusén 7303 
(holotipo, S!; isotipo, US!). Figura 8. 

Sufrütice con xilopodio de 40-60 cm alt, tallos 

subtetrágonos, simples, glabros. Hojas de (1 5 —65 


mm, lineares, nervios secundarios inconspi- 
cuos, glabras; vaina estipular de 4.5—6.5 mm, pubérula, 
de borde irregular con 3—4 lacinias de 5-15 mm. 
Inflorescencia tirsoide, terminal, con pedúnculos de 2— 
4 cm. Hipanto 1.2-1.7 mm, turbinado, glabro; cáliz con 
lóbulos 1.2-1.5 mm, triangular-acuminados, glabros, 
con papilas antrorsas sobre el borde; corola, superficie 
interna con anillo de pelos moniliformes, delgados, en la 
mitad del tubo y pelos más gruesos, largos, en arco en 
los lóbulos; disco entero. Flor brevistila: corola 3.5— 
4 mm, lóbulos iguales o más cortos que el tubo; anteras 

5-0.7 mm, filamentos ca. 0.5 mm; estilo 1.2-2 mm. 
Flor longistila: corola 3.7—4.5 mm, lóbulos iguales o 


42 


Annals of the 
Missouri Botanical Garden 


Volume 96, Number 1 
2009 


Cabral 43 
Revisión Sinóptica de Galianthe 


más largos que el tubo; anteras subsésiles, ca. 0.7 mm; 
estilo 3.7—4(—5.2) mm. Fruto no visto. 


Sur de Brasil 
(Paraná y Mato Grosso do dod en ipia altos de 
8 


Distribución, hábitat y fenología. 


0-875 m, en áreas “pinheiros” 
Araucaria angustifolia riw [^ florece 
fructifica de noviembre a abril. 

Material — representativo estudiado. BRASIL. Mato 


Grosso do Sul: Rod. MF 642, 5 Km O, Tacuru, 16 dic. 
1983, G. Hatschbach 47302 , UB). Paraná: Palmeira, 
28 nov. 1948, G. Hatschbach 1113 (HBR, MBM, US). 


19. Galianthe longisepala E. L. Cabral, Bonplan- 
dia (Corrientes) 13: 15-17. 2004. TIPO: Brasil. 
Goiás: Serra dos Cristais, 2 km N of Cristalina, 
1250 m, 2 mar. 1966, H. 
(holotipo, UB!; isotipos, F!, MO!, NY!, RB!, USS. 


Centro de Brasil 
(Goiás y Minas Gerais), en campos cerrados, en suelos 


Distribución, hábitat y fenología. 
lateríticos y con afloramiento rocosos de 1050- 
0 m; florece y fructifica de enero a marzo. 


Observaciones. Esta especie se reconoce facil- 
mente por las inflorescencias congestas y por los 
frutos con sépalos largos y reflexos; en material 
herborizado se diferencia por el color de los nervios 
en el envés, castafio-negruzco, marcadamente con- 
trastante con el resto de la superficie foliar; es afín a 
Galianthe lanceifolia, de la que se diferencia porque 
G. longisepala tiene hojas glabras de 3-8 mm lat. (vs. 
ojas pubérulas o pubescentes, 8-18 mm lat., en G. 
lanceifolia). 


Material representativo estudiado. BRASIL. Minas Ger- 
ais: Morro das Pedras, ca. 25 km NE of Patrocinio, 1050 m, 
28 ene. 1970, H. S. Irwin et al. 25504 (NY, RB, UB, US). 


20. Galianthe macedoi E. L. Cabral, Ponp on 
o 100-4): 121- 123. 2000. 
"a Goiás: Jataí, Faz. Queixada, 10 dis 
8, A. Macedo 1468 (holotipo, SP!; isotipos, 

ud F!, IAC 282701, NY!, SP). 


Observaciones. Se caracteriza por ser un sufrütice 
monocaule, hojas oblongo-lanceoladas de base obtusa 


o m con nerviación paralelodroma y con fruto 


abro, semillas con alas muy breves; es 


. S. Irwin et al. 13307 


afín a Galianthe grandifolia por el tallo simple con 
Cus wd sin brotes axilares, pero G. macedoi 

0— alt. (vs. m alt. en G. 
22d NOS glabras de 7— 20 mm lat. (vs. hojas 
con haz glabra o pubérula, envés pubescente de 5— 


40 mm lat.). 


Distribución, hábitat y fenología. Centro de Brasil 
(Goiás) en campos bajos cerca de ríos, 900-1000 m; 


florece y fructifica de diciembre a mayo. 


Material representativo estudiado. BRASIL. Goiás: 
25 km SW of Caiaponia, 1 mayo 1973, W. R. Anderson 
9599 (UB, US 2774976, US 2774977). 


21. Galianthe matogrossiana E. L. Cabral, Bon- 
plandia (Corrientes) 13(1-4): 17-19. 2004. 
TIPO: Brasil. Mato Grosso do Sul: Sidrolandia, 
Agua Rica, 12 abr. 1972, G. pu eA 29439 
(holotipo, MBM!; isotipos, US 2745683!, US 
2835289!) 

Observaciones. Galianthe matogrossiana se reco- 

noce por ser un sufrútice erecto, 1-2-caules, de 0. 

1 m alt, con hojas de 55-90 X 1-26 mm, opuestas 

sin ile elípticas, glabras, 

estipular prolongada por encima de la separación 
el par de hojas, hipanto pubérulo, corola externa- 
mente micropapilosa; fruto pubescente con semillas 
complanadas. Es similar a G. guaranitica por los 
nervios secundarios marcados y por vaina prolongada, 
pero se diferencian porque G. matogrossiana tiene 
hojas con base atenuada (vs. hojas de base redon- 
deada o levemente cordada en G. guaranitica), ala- 
bastro con papilas largas sobre los lóbulos (vs. 
alabastro con papilas cortas sobre los lóbulos). 
Distribución, hábitat y fenología. Sur de Brasil 

(Mato Grosso do Sul), en campos, 200—460 m; florece 

y fructifica de diciembre a abril. 

Material ird BRA Mato 


SIL. 
Grosso do Ribas do Río Pardo, 25 ene. 
1979, A. Krapovickas et al. 34386 (CTES, SI). 


representativo 
Sul: 22 km W 


22. Galianthe montesii E. L. Cabral, Candollea 58: 
— TIPO: Paraguay. Alto Paraná: Ñacun- 

v. 1950, J. E. Montes 9764 (holotipo, 

FUR pa LIL). 


«— 


Figura 6. Galianthe hassleriana. —A. Plan 
Flor. —E. Interior de la corola desplegad 


; G-L, 5689.) 
7. Galianthe elio —A. Planta. 
Interior de la corola desplegada. 
desplegada. —H. Hipanto, cáliz, estilo y estigma. —L F 


12244; F-K, Brade 6783.) 


—E. Hipanto, cáliz, Lu F-G. 
o. J-K. Semilla. 


a. —B. Vaina estipular con lacinias. C—F. Flor brevistila. —C. Alabastro. —D. 
a. —F. [Nen cáliz, aa EE nl longistila. —G. F 
desplegada. —I. E o, cáliz, estilo y estigma. —J. Fru 


lor. —H. Interior de 


. K-L. Semilla. —K. Cara dorsal. —L. Cara ventral. 


— B. Vaina estipular con lacinias. C-E. Flor longistila. —C. Alabastro. —D. 


Flor brevistila. —F. Flor. —G. Interior de la corola 
. —J. Cara ventral. —K. Cara dorsal. (A-E, Irwin 


44 Annals of the 
Missouri Botanical Garden 


Volume 96, Number 1 Cabral 45 
2009 Revisión Sinóptica de Galianthe 
Observaciones. Esta especie se caracteriza por ser 1963, J. Correa Gomes Jr. 1437 (UB) PARAGUAY. 


un sufrútice ramificado de 0-80 cm alt., de hojas 


linear-lanceoladas de m lat, pu as o 
escabrosas, hipanto y frutos becomes semillas 
pea aladas. Es semejante a Galianthe 
thalictroides, 


desarrolladas y por las hojas lineares, pero esta especie 


por las ramas secundarias opuestas, 


es más robusta, forma matas densas, tiene las hojas de 


0.7-2 mm lat., glabras, el hipanto y fruto glabros. 
Nordeste 


(Misiones) y Paraguay oriental 


de 
(Alto 


Paraná), en campos altos, con suelo arenoso laterítico 


Distribución, hábitat y fenología. 
Argentina 


de 200-600 m; florece y fructifica de octubre a enero. 


Material representativo estudiado. ARGENTINA. Misi- 
Gral. Manuel Belgrano, San Antonio, 2 Oct. 1949, J. 
E. Montes 7036 (CTES, SI). 


23. Galianthe parvula E. L. Cabral, Bonplandia 
(Corrientes) 7(1—4): 24. 1993. TIPO: Paraguay. 
Amambay: Sierra de Amambay, Estrella, ene. 


1908, T. Rojas 10082 (holotipo, G!). 


Distribución, hábitat y fenología. Paraguay orien- 
tal (Amambay, San Pedro) y se ha encontrado también 
en el sur de Brasil (Mato Grosso do Sul), en campos 
cerrados, campos rocosos en laderas de colinas, de 


300-1000 


febrero a junio. 


m; florece de octubre a enero, fructifica de 


Observaciones. Se la reconoce por ser un sufrútice 
muy ramificado, 10-20-caule; hojas de 12-25 X 
10 mm, elípticas o elíptico-lanceoladas, glabras, por 
las inflorescencias paucifloras, breves, congestas en 
todas las ramas; hipanto irregularmente pubescente, 
corola externamente papilosa y por la cápsula globosa, 
pilosa. Por las ramificaciones y ubicación de las 
inflorescencias es afín a Galianthe kempffiana; pero 
G. parvula tiene tallos glabros 20-40 cm alt. (vs. 
tallos pubescentes, 80-100 em alt. en G. kempffiana); 
hojas 4-5 par 


secundarios (vs. s 
nervios De semillas ápteras con estrofíolo 


es de nervios 


caduco (vs. semillas aladas con estrofíolo persistente). 

Material — representativo estudiado. BRASIL. Mato 
Grosso do Sul: Mun. Ponta Pora, m do estrada a 
Campo Grande, 22732'0'S, 55^4l]'0"W, nuc 18 nov. 


Amambay: El Buracón, 30 km W de Pedro A Caballero, 
O, UB). San 

Pedro: Yaguareté Forest, 23748'38" S, 56^ 07 ow. 20 jun. 
1995, E. Zardini et al. 42813 (CTES, MO, PY). 


24. Galianthe peruviana (Pers.) E. L. Cabral, 
10(1-4) 123-124. 


acoce peruviana Pers., 


(Corrientes) 


. inmutat. Borreria 
peruviana (Pers.) L. B. Sm. & Downs, Sellowia 7: 
18. 1350, TIPO: Perú. In peruviae montibus ad 
in ruderatis & nunc Mene 
Cormillà eS TOhconads, Ruiz & 
(holotipo, MA!; isotipos, B-W!, F!). Figura . 
Borreria a DC., Prodr. 4: 550. 1830. TIPO: Peru. In 
Peruviae montibus, Ceron, Haenke s.n. (holotipo, G- 


DC». 

Borreria ericoides Cham. & Schltdl., Linnaea 3(4): 326. 1828. 
TIPO: In Brasilia aequinoctiali, Sellow s.n. (lectotipo, 
designado aquí, LE!; isotipo, G-DC?). 


Sufrütice con xilopodio, muy ramificado, pluricaule 
de 10-20-caule, de po n cm alt, tallos de 
dar: a glabros nudos .5— 
m. Hojas pd TAE 25) X (0.3 E 
05-0903) mm, filiformes, lineares o lanceoladas, 
margen revoluto, ápice y base agudos, glabras, 
pubérulas o pubescentes, nervio medio prominente 
en el envés, surcado en la haz, nervios secundarios 
inconspicuos; vaina estípular de 1-2.5 mm long., 
pubérula o pubescente, con 3-6 lacinias de l- 
3.2 mm, glabras, rojizas. Inflorescencias d 
congestas, terminales. Hipanto = urbi- 

nado, glabro o pubérulo; cáliz de (shila pode 
subulados AS 1.2-1.5 mm; 34 mm, 
papilas formando crestas sobre el dorso de los lóbulos, 


corola con 
lóbulos iguales o más cortos que el tubo; disco entero. 


Flor brevistila: interior de la corola con pelos 
moniliformes desde la base de los lóbulos hasta la 
mitad del tubo; anteras de 1-1.5 mm, filamentos ca. 
1 mm, estilo 1-2 mm. Flor longistila: interior de la 
corola co moniliformes más gruesos en los 
lóbulos y d delgados en la mitad del tubo; anteras 
subsésiles, ca. 1 mm; estilo 3-4 mm. Cápsula 2- 
2.5 mm, subglobosa, pubérula o glabra; semillas 2— 


2.2 mm, subcilíndricas, notablemente escrobiculadas, 


Figura 8. jr dd e d —A. Planta. —B. Vai 


Hipanto, cáliz, estilo y 


estilo y estigma. “HT rus la iu ee (A-E, Dus 
Figura 9. Galia e i pda Planta. —B. Vaina 
Flor. —E. Hipanto, cáliz, estilo y mon "ur Inte 


—H. Hipanto, ofl, estilo 


na estipular con lacinias. 
—E. Interior de la corola despleg 


n, 4167; F-H, 
epale con n C-F 
rior de la corola KR G-I. Flor brevistila. —G. Flor y alabastro. 
y estigma. "i Interior de la corola desp 


C-D. Flor on —C. Flor 
—G. Hipanto, 


—D. 
ada. F—H. Flor brevistila. — p 


E e pte 
. Flor 


ps 5) 
longistila. —C. Alabastro. 


—J. Fruto. —K. Cara dorsal. 


emilla. 


—L. Cara ventral. —M. Corte transversal. (A-B, J-M, Arbo 4167; C "e. dum 22949; G-I, me 24802.) 


46 


Annals of the 
Missouri Botanical Garden 


cara dorsal convexa, cara ventral plana con surco 
alrededor del estrofíolo persistente. 


Distribución, hábitat y fenología. Bolivia (Santa 
Cruz), Brasil (Goiás, Minas Gerais, Sáo Paulo), en 
Peri (Cuzco, Huancavelica, Huánuco, Junín) en 
campos altos con suelos rocosos; en Brasil en campos 
cerrados, en suelos con afloramientos rocosos en el 
planalto central de 900-1600 m, en Bolivia en 
laderas de los valles interandinos de 1550-2250 m 
y en Perú en el estrato herbáceo generalmente en 
zonas de pendiente con afloramientos rocosos, entre 
1500-3000 m; florece y fructifica de septiembre a 
mayo. 


Observaciones. Posiblemente debido a la extensa 
distribución en lugares altos, esta especie presenta 
una notable variabilidad en el porte, desde plantas 
ramas secundarias hasta ram 


es sin as mu 


ric y 
ar: que rematan en inflorescencias con- 
gestas, desde aspecto ericoide hasta con notables 
entrenudos. También las hojas presentan tamaño y 
pubescencia variables. 


Discusión. Por considerar que el original de 
Borreria ericoides ha desaparecido en el herbario B, 
se elige como lectotipo, el ejemplar del herbario LE, 
ado de conservación. 

Spermacoce corymbosa Ruiz & Pav., Fl. Peruv. 
[Ruiz & Pavon] 1: 60, tab. 91, fig. a. 1798 y Galianthe 
iz & Pav.) E. L. Cabral, Bol. Soc. 


por estar completo y en buen est 


corymbosa (Ru 


Argent. Bot. 27: 241, 1991 [non Spermacoce corym- 
osa L., Sp. Pl. (ed. 2) 1: 149. 1762] son nom. illeg. 
Material representativo estudiado. BOLIVIA. Santa 


Cruz: Prov. Caballero, 3 km NE of Abra de Quiñe, 

18°04'S, 64^19'30"W, 31 dic. 1995, M. Nee 46651 (CTES, 
LPB, NY). BRASIL. Goiás: Mun. de Alto Paraiso, Nova 
Roma, 20 feb. 1991, B. A. S. Pereira et al. 1476 (IBGE). 
Minas Gerais: Serra de Moeda, F. Sellow 1735 (LE); Lagoa 


Santa, 14 dic. 1971, J. Semir 561 (UEC). Sáo Paulo: 1884, 
A. Glaziou 17640 (P). uzco: Urubamba, Machu- 
pichu, 11-1938, L. Vargas Huancavelica 


Tayacaja, 14 abr. 1954, O. Tovar 1810 M O). Huánuco: 
Camino a Panao, 2650 m alt., 7 sep. 1948, R. Scolnik 104 
(CORD). Junín: Manto. 11 jul. 1961, F. Woytkowski 6536 
MO). 

25. Galianthe pseudopeciolata E. L. Cabral, 
Bonplandia (Corrientes) 7(1—4): 26. 1993. TIPO. 
Paraguay. Amambay: Sierra de Amambay, ene. 
1908, E. Hassler 10102 (holotipo, Ct isotipos, 
F!, LIL!, NY!, P). 


hábitat y — fenología. 
(Amambay, Caaguazú) y sur de Brasil (Mato Grosso 


Distribución, Paraguay 
do Sul, Paraná, Sáo Paulo), en lugares abiertos e 
inundables, en terrenos pantanosos o en borde de ríos 
y arroyos, de m; florece de noviembre a 
febrero, fructifica de marzo a abril. 


Observaciones. Galianthe pseudopeciolata se reco- 
noce por ser un sufrütice erecto de 1—1.70 m alt., de 
ojas de 3-13 cm 
de base 


ie od formando un pseudopeciolo, haz 


tallo gue P glabro; h 
eoladas u oblongo-lanceoladas, 
glabra, envés escabriúsculo sobre los nervios y 
márgenes; cápsula glabra y semillas aladas. Por su 
hábito robusto es muy afin a 6G. valerianoides (Cham 
& Schltdl.) E. L. Cabral, ambas viven en campos bajos, 
pero es fácil distinguirlas porque esta ültima especie 
tiene hojas de base truncada u obtusa y además tiene el 
tallo con pelos retrorsos en los ángulos. 
Material — representativo estudiado. Mato 
Grosso do Sul: Maracajú, 29 dic. 197 


Haischbach 33914 (MBM). Sáo po 

5, E. Leite 3468 (LIL). aera Amambay: 
Pedro Juan Caballero, 8 feb. 1951, G. J. Schwarz 11791 
(CTES, LIL). Caaguazú: Yhú, P. Trene 4922 (F, MO, 
NY, SL US 


m 


26. Galianthe ramo L. Cabral, Soc 
Argent. Bot. 29(3—4): p 1993. BO. ri 
Goiás: BR 040, 12 km al S de Luziania, ca. 
1000 m, cerrado sucio, 1 feb. 1990, M. M. Arbo, 
R. Monteiro, A. Schinini & A. Furlan 3366 
(holotipo, HRCB!; isotipo, CTES!). 


Distribución, hábitat y fenología. Centro de Brasil 
Distrito 


Sul) en campos cerrados, sujeto a incendios 


Goiás, Federal, Minas Gerais, Mato Grosso 


26 


pericos. 800-1205 m; florece de octubre a febrero, 
fructifica de marzo a mayo. 

Observaciones. Se caracteriza por ser un sufrútice 
de 12-30 em alt., 
hojas ovadas o elíptico-lanceoladas, 15-42 X 4.5- 


muy ramificado, 5-20-caule; 
mm; vaina estipular 2-3 mm, con 5-6 lacinias 
filiformes, pubérulas. Es afín a Galianthe cen- 
tranthoides, por la pubescencia en toda la planta y 
por las hojas plegado-nervosas, pero se diferencian 
undarias escasas a 
lladas en G. 


centranthoides), inflorescencia congesta (vs. inflores- 


orque G. ramosa tiene ramas sec 
nulas (vs. ramas secundarias muy desarro 


cencia amplia), corola 2-3 mm (vs. corola 4.5-6.5 mm), 
semillas subcilíndricas ápteras (vs. semillas complana- 


das, aladas) 


erial representativo estudiado. BRASIL. Distrito 
Reserva Ecológica de IBGE, 


o, 20 Nov. 1 894, C. Ciaziou 12906 
(P). M : Amambaí, 1979, W. G. García 
14024 (UEC). Minas Gerais: BR 050, 65 jn N de Uberaba, 
29 ene. 1990, M. M. Arbo 3024 (CTES, HRCB). 


27. Galianthe reitzii E. L. Cabral, aes 
(Corrientes) 10(1-4): 124—126. 2000. TIPO: 


Brasil. Santa Catarina: Municipio Urubici, pue 


Volume 96, Number 1 
2009 


Cabral 47 
Revisión Sinóptica de Galianthe 


do Oratorio, Bom Jardim, 9 dic. 1958, R. Reitz & 
R. Klein 7688 (holotipo, HBR!; isotipos, G!, US!). 


Observaciones. Se caracteriza por ser un sufrútice 
muy ramificado de 0.20-1 m alt., con hojas pseudo- 
verticiladas, lineares, coriáceas, glabras; vaina estip- 
ular 1-1.5 mm, pubérula, con 1 lacinia central de 
1.5-1.8 mm y dos ipeo de laterales poco desarro- 
a Galianthe 
peruviana, por ser ideni muy ramificados, con 


llados, con coléteres apicales; es afin 


hojas pseudoverticiladas, lineares, pero se diferencian 
orque G. reitzii tiene vaina estipular con 3 lacinias 
—6 lacinias filiformes 
en G. per 

corola 3—4 mm), 


brevistila dos anillos de pelos, uno en la mitad del 


subuladas desiguales vs. 
As n uales 
6 mm (vs. interior de la corola 


tubo y otro en la base de los lóbulos (vs. un solo anillo 
de pelos desde la base de los lóbulos hasta la mitad 
el tubo) 


Sur de Brasil 


anta atarina en la erra Pera en a oramien Os 
Santa Cat. , en la “S Peral", fl t 


Distribución, hábitat y fenología. 
rocosos con altitudes que varían entre 1000 a m, 
del P 


8 
arque Naciona áo Joaquim, e 


un 


considerada la región más fría y el único lugar donde 
se producen nevadas en Brasil (IBAMA, 1998); 


florece y fructifica de diciembre a abril. 


Motena Eu inti estudiado. BRASIL. Santa Cata- 
o Urubici, ine a Bom Jardim da Serra, 20 
al. 5361 (CTESN, SD); Morro da 
duca. 3 ene. ` 1949, R "Reis 2969 (B, HBR, US). 


28. Galianthe souzae E. L. Cabral & Bacigalupo, 
Bol. Soc. Argent. Bot. 34: 153-154. 2000. TIPO: 
Brasil. Sáo Paulo: Apiaí, Distr. Barra do Chapéu, 
ca. 8 km de Bonsucesso de Itararé, 3 jun. 1994, 
V. C. Souza, P. Miyagi, E. Moncaio 6112 

(holotipo, SP!; isotipos, CTES!, UEC!). 


Observaciones. Se caracteriza por ser un sufrútice 
de 50-80 em alt., 


glabros, con las hojas basales caducas, 


de tallos simples, fistulosos, 
la vaina 
estipular con 3-5 lacinias soldadas en la porción 
basal y las corolas rosadas o lilacinas, esta ültima 
característica es poco frecuente en el género. Por su 
tallo simple, forma y tamafio de las hojas es semejante 
a Galianthe valerianoides, pero se diferencian porque 
G. souzae tiene tallos subtetrágonos glabros (vs. tallos 
tetrágonos, escabrosos, con pelos retrorsos en los 
ángulos en G. valerianoides), hojas glabras, con 
nervios secundarios inconspicuos (vs. hojas escabro- 
sas, con 2-3 pares de nervios basales y 2-3 pares 
suprabasales) semillas subcilíndricas, ápteras (vs. 
semillas complanadas, irregularmente aladas). 


Sur de Brasil 


(Sáo Paulo), en campos graminosos de suelos con 


Distribución, hábitat y fenología. 


eruviana), corola 5— 


afloramientos rocosos de 600-1000 m; florece y 
fructifica de octubre a julio. 


Material — representativo estudiado. BRASIL. Sao 
Paulo: Municipio de Itararé, Faze: o Nicolau, 30 oct. 
1993, V. C. Souza 4431 ( , ESA); Pedreira, Cobastalco, 


24°18'S, 49°12'W, 17 ago. 1994, Barreto et al. 2976 (ESA, 
IAC). 


29. Galianthe thalictroides (K. Schum.) E. L. 
abral, Bol. Soc. Argent. Bot. 278-4): 246. 
9 Basónimo: Borreria  thalictroides K. 


Schum., FL 


3: 123. -l 
a (K. Schum.) P 
Inst. Centr. Bioci., Ser. Bot. 35: 81. 1977. 
Brasil. Sáo Paulo, 1821, A. Saint-Hilaire 1499 
(neotipo, designado aquí, P!). Figura 10. 
0.60-1.20 m  alt., 
xilopodio muy desarrollado, 3—5-caules primarios 


Sufrútice ramificado de con 
con 
ramas secundarias, opuestas, tallos subtetrágonos, 
glabros o papilosos. Hojas 8-25 X 0.7-2 mm, lineares 
o linear-lanceoladas, de margen d glabras o 
papilosas, con ral notable y nervios 
teur deris poco visibles; vaina rin de 1-1.25 
(2.5) mm long., glabra o pubérula, con 1—5-lacinias de 
0.2-2.5 mm. Inflorescencia terminal, sólo en los tallos 
primarios, pedúnculo de 6-10 cm. Hipanto turbinado, 
1 mm, glabro; cáliz con lóbulos de mm, 
triangulares, glabros; disco entero. Flor brevistila: 
corola 5.5-6.2 mm, de lóbulos más cortos que el tubo, 
interior con anillo de pelos moniliformes en el tubo y 
e dispersos en los lóbulos; estilo 3-3.2 mm; anteras 

2 mm, filamentos 1.2-1.5 mm. Flor longistila: 
3.7—5 mm, de lóbulos más largos que el tubo, 
interior con anillo de pelos moniliformes, delgados y 
densos en el tubo y pelos gruesos en los lóbulos; anteras 
subsésiles, 1.2-1.5 mm; estilo 3.7—5 mm. Cápsula 


5 mm, glabra; semillas 2-3.5 mm, irregularmente 
aladas, comprimidas dorsiventralmente, cara ventral 


cubierta parcialmente por el estrofíolo. 


Nordeste de 


Argentina (Misiones y Corrientes), Brasil (Paraná, Rio 


Distribución, hábitat y fenología. 


rande do Sul, Santa Catarina, São Paulo), Paraguay 
oriental (Amambay, Caaguazá) y Uruguay, en campos 
altos, con suelo arenoso-rojizo, también en terrenos 


modificados, en borde de caminos, de 100—850 m; florece 


de noviembre a enero, fructifica de febrero a marzo. 


Discusión. Los sintipos citados en el protologo del 
basónimo Borreria thalictroides han desaparecido en 
el herbario B, razón por la cual se elige un neotipo, de 
una colección clásica de Brasil, Sáo Paulo, A. Saint- 
Hilaire 1499 (P), que se encuentra completa y en buen 
estado de conservación. 


48 


Annals of the 
Missouri Botanical Garden 


Material representativo estudiado. ARGENTINA. Corr- 
ientes: Santo Tomé, Ayo. Chimiray, 23 nov. 1974, A. 
Krapovickas et al. 26151 (CTES). yogi Cone 
ruta 1, 5 km NW de Concepcién, 18 dic. 1983, E. Cale ei 
al. 470 (CTES). BRASIL. Paraná: CM cl 15 nov. 
1957, G. Haischbach 4288 (MBM). Rio Grande do Sul: 
Giruá, 18 abr. 1974, M. L. ud et al. 1154 (ICN). Sáo 
, J. Maitos 13973 (SP). 
ay, in campis 
siccis Punta Porá, dic. 1908, E. Pos a a 9942 (G). 
Caaguazú: Dans les campos, . 1874 Balansa 1746 
(G, P). URUGUAY. Ban da orienta e del iius 1816 a 
1821, voyage d'Auguste de Saini-Hilaire 2653 (P) 


30. Galianthe valerianoides (Cham. & Schltdl.) E. 


L. Cabral, Bol. Soc. Argent. Bot. 27(8-4): 246. 
1991 [1992]. Basónimo: Borreria valerianoides 
Cham. & $e Hdl ea 3 


designado aquí, gs HBR!, 


MO). Figura 11. 


isotipos, BL 


Borreria luteovirens s Publ. Field Mus. Nat. Hist., Bot. 
Ser. 8: 1. TIPO: Brasil. [Rio Grande do Sul:] 
Cachoeiravi in E 12 Jan. 1902, G. O. Malme 1060 
(holotipo, S!). 


Sufrütice erecto, con xilopodio, de 0.8—2.5 m alt., 
tallo escabroso, rara vez glabro, ángulos muy 
marcados con pelos retrorsos; a veces ramas secun- 
darias opuestas. Hojas 35-11 0-25 mm, lanceo- 
ladas u oblongo-lanceoladas, base obtusa o truncada, 
ápice agudo, escábridas, pelos más largos sobre los 
nervios en el envés; plegado-nervosas, con 2-3 
pares de nervios secundarios basales y 2-3 pares de 
nervios suprabasales, ligeramente paralelos, vaina 

3.5-5 mm, 
Inflorescencia 
10-20 
2.5 mm, a veces piloso; cáliz con p de 2.5- 


.5-6.2 mm, 


con papilas densas sobre el margen y ae media de 


pubescente de 
5-15 mm. 
pedúnculos de 


lacinias de 
amplia, cm. Hipanto 2- 
4 mm, triangular-subulados; corola de 


los lóbulos; disco bilobado. Flor brevistila: corola de 
lóbulos tan largos como el tubo, interior anillo ancho 
de pelos moniliformes en el tubo, hasta la base de los 
lóbulos; anteras 1.5-1.7 mm, filamentos 1.7—2.5 mm; 
estilo 2.5-3.5 mm. Flor longistila: corola, lóbulos más 
largos que el tubo, interior con anillo de pelos 
moniliformes, delgados, cortos, en e y más 
gruesos y largos en los lóbulos; anteras subsésiles, 
1.2-1.5 mm; estilo 4.2-6 mm. Cápsula 3.7—4.2 mm, 
subcilíndrica, glabra, a veces con pelos dispersos; 
2.5-3.5 mm, 


mente, cara ventral cubierta por estrofíolo membra- 


semillas complanadas  dorsiventral- 


náceo, cara dorsal finamente foveolada, desigual- 


mente aladas. 


Nordeste de 


Argentina Aus d ae Corrientes y n: d 


Distribución, hábitat y fenología. 


oriental (Alto Paraná, Amambay, Caaguazú, Caazap 
anindeyú, peg Cordillera) y Brasil | (Distrito Feder. 
al, Goiás, Mato Grosso, Minas Gerais, Paraná, Rio 
Grande do Sul, Santa Catarina, S&o Paulo), en pantanos, 
lugares bajos, esteros, bafiados y en curso de arroyos; 
florece de octubre a enero, fructifica de febrero a julio. 


Discusión. El ejemplar citado en el protologo del 
zenaleguneides had d 


basónimo B. a parecido en e 
herbario B, razón por la cual se elige un neotipo, con 
duplicados, que fue coleccionado en el área donde 
esta especie es frecuente. 

Observaciones. Galianthe valerianoides se puede 
reconocer fácilmente, por la altura de la planta hasta 
de 2.5 m, tallos tetrágonos fistulosos, de hojas muy 
desarrolladas de base obtusa o truncada con nervios 

cundarios + paralelos y amplia inflorescencia 
largamente pedunculada 

En Brasil tiene utilidades terapeüticas como 
ipecacuana (extraido de las etiquetas). 

e vulgar. “Sabuguerinho do campo” (Porto 
et al., 1977). 


Material representativo estudiado. ARGENTINA. Corr- 
ientes: Ituzaingó, Playadito, 20 km W de Apóstoles, 4 Feb. 
a . Schinini et al. 21818 (CTES). Misiones: Loreto, 

del cruce de Ruta 12, camino a ru 18 dic. 2000, 
ps Cabral 674 (CTES). BRASIL. Distrito Fe 
Ecologica IBGE, 13 abr. 1983 
Goiás: Morrinhos, 4 ene. 1971 


ind ene. 1932, C. Jürgens 460 (B). São Paulo: Bocaina, 
Serra da Bocaina, abr. 1972, J. H. Kirkbride 1745 (SP). 
Pauls Catarina: Cagados, 2 1962, R. Klein 3544 (HBR). 
PARAGUAY. Alto Paraná: 7 km al sur de Villa Fortuna, 
sas ei al. 5768 (G, MO, NY). 
907, E. Hassler D. PY Caaguazú: 
6, B. Balansa 1742 
b. 1989, N. PRA 1157 (CTES, 
Tavai estero pe al pueblo, 20 dic. 
1988, F. Mereles 2371 (FCQ, MO). Central: en ene. 1895, 
E. Hassler 1814 (CTES). Cordillera: Cordillezc de Piribe- 
buy, dans les prairies, abr. 1883, B. Balansa 4551 (P). 
Canindeyú: In palude Ipé-hú, Sierra Mbaracayú, oct. 1899, 
E. Hassler 5089 (F, G, NY, P). 


Il. Galianthe subg. Galianthe sect. Laxae E. L. 
abral, sect. nov. TIPO: Galianthe laxa (Cham. 
& Schltdl.) E. L. Cabral [= 

& Sehltdl.]. 


Borreria laxa Cham. 


sufrutex sine xilopodio, thyrsi in ramis principa- 

TM et elena 
Frútice o sufrútice sin xilopodio, con raíces 
axonomorfas profundas, pluricaules, con ramas se- 


Volume 96, Number 1 
2009 


Cabral 49 
Revisión Sinóptica de Galianthe 


cundarias desarrolladas, inflorescencias tirsiformes 
comprimidas, densas, excepcionalmente amplias, 
laxas, terminales o axilares en ramas primarias y 
secundarias; flores distilas, 4-meras; frutos capsu- 
lares; semillas con estrofíolo caduco o persistente, 
ápteras, raro alas inconspicuas. Comprende 9 espe- 
cies que habitan en Bolivia (5 sp.), Brasil (4 sp.), 
Paraguay (4 sp.), Argentina (3 sp.) y Uruguay (1 sp.), 
en bordes de arroyos, de caminos, sotobosque de 
selvas o bosques, campos altos, rupestres, laderas de 
cerros de 0-3000 m. Figura 12 (A, B). 

Se diferencia de la sección Galianthe porque la 
sección Laxae no tiene xilopodio, presenta tallo 
primario con numerosas ramas secundarias, desarro- 
ladas y divididas, inflorescencias terminales y 
axilares en las ramas primarias y secundarias (vs. 
con xilopodio, 1-20-tallos primarios con inflorescen- 
cias terminales, si también tienen ramas secundarias, 
las inflorescencias se ubican sólo en tallos primarios, 
sección Galianthe). 

Se eligió a Galianthe laxa como especie tipo, por 
ser su basónimo (Borreria laxa), una de las especies 
más antiguas y por su amplia distribución. 


CLAVE PARA IDENTIFICAR LAS ESPECIES DE GALIANTHE SECT. LAXAE 


l. Hojas de 1-4(-8) mm lat., con nervios secundar- 
ios inconspicuos o 2-5 pares visibles ......... 
l. Hojas de (2-)5-20(-35) mm lat., con (3-)4-6 
pares de nervios secundarios visibles ......... 
2(1). Hojas con nervios secundarios inconspicuos; 
inflorescencias más o menos contestas ........ 
Hojas con 2-5 pares de nervios secundarios 
visibles; X Hi Iu lax: 


3(2. Hojas lineares o UR levemente 


4-partido; 


ea ba aa . G. bisepala 
3. Hojas lineares, glabras; cáliz 4-partido; corola 

micropapilada, 4—5 mm long.; semillas no ala- 

das 


4(2). Tallos pubérulos o levemente pubescentes; hojas 

pubérulas a pubescentes, con 3 pares de nervios 
secundarios; vaina estipular pubescente, con 5-9 
(-10) lacinias de 5-9 mm long.; flor brevistila 
con pelos moniliformes en la superficie interio 
del tubo corolino .......... udyungensis 
4. Tallos glabros; hojas glabras, con 2-5 pares 


nervios secundarios; vaina estipular pubérula a 


formes en el tubo y en los lóbulos ... 32. G. aurelii 
51). Hoa INI 4 pal l ] 


J pag 
nervios del envés; corola de la flor brevistila y 


longistila con pelos moniliformes en la superficie 
b 39 


interna del tubo . verbenoides 


5: Hojas sin papilas sobre los nervios del envés; 
corola de la flor brevistila y longistila con pelos 
moniliformes en el interior del tubo pelos 

también en los lóbulos, éstos a veces ausentes en 

las flores brevistilas ..................... 

. Corola de las flores brevistilas y longistilas con 
pelos moniliformes en la superficie interna del 
tubo y en los ló 
estipular glabra, raro pubérula . 34. 

6. Corola de las flores 


moniliformes, delgados y cortos en el tubo y en 


bulos; hojas glabras; vaina 
. chiquitosiana 
hersia con pelos 
la flor longistila pelos en el tubo y en los lóbulos; 
hojas glabras, pubérulas o pubescentes; vaina 
estipular pubescente .. o.. llle 
. Vaina estipular con lacinias de 3.5-17 mm long.; 
corola 3—6. ong., con papilas muy desarro- 
lladas en el dorso apical y márgenes de los 
lóbulos, más notables en el alabastro; sufrútic 
siempre erecto 


ia 
7. Vaina estipular con lacinias de 1.7— 8 mm long 


corola 2-4 mm long., con papilas cortas sólo en el 


dorso apical de los lóbulos: sufrútice erecto o 
apOyarte anios s batts, oh sados ves 
8(7). Vaina estipul 
rojizas; corola rosa : 
8. Vaina estipular con lacinias glabras no rojizas; 
corola blanca 


ar con lacinias glabras o pubérulas, 


o-lilacina aride ona 


G. laxa 


31. Galianthe andersonii E. L. Cabral, Bonplandia 
(Corrientes) 10: 119—121. 2000. TIPO: Brasil. 
Minas Gerais: Serra do Espinhaco, 25 km by rd. 
NE of Diamantina, ca. 1.5 km from Jequiti, 790— 
900 m, 12 abr. 1973, W. R. Anderson 8717 
(holotipo, UB!; isotipo, NY!). 

Distribución, hábitat y fenologia. Brasil (Minas 

Gerais) frecuente en chapadas con afloramientos 

mpos rupestres de 700-1400 m, 


florece y fruetifien de febrero a abri 


rocosos oO 


Observaciones. Esta especie es muy particular 
por sus tallos rojizos, por sus hojas secas dis- 
coloras, con la haz más oscuro y brillante, las 
lacinias de la vaina estipular 2-6 mm, rojizas; corola 

e 2-3.5 mm, con un ligero tinte rosa. Es afín a 
Calianshe eupatorioides, porque ambas son sufrútices 
ramificados cerca de 1 m alt., con forma y tamaño de 
hojas similares, pero esta última especie no presenta 
tallos rojizos, las lacinias de la vaina estipular son 
verdosas de 6-12 mm y la corola blanca de 2.5- 
5.5 mm. 

Material representativo estudiado. BRASIL. Minas Ger- 
ais: Serra do Espinhaco, 10 km SW of Diamantina, 3 feb. 
1972, W. R. Anderson et al. 35231 (NY, UB 
32. Galianthe aurelii E. L. Cabral, Bonplandia 

(Corrientes) ^ 1. 1993. TIPO: Paraguay. Guairá: 
Colonia Independencia, Cerro Pelado, ene. 1967, 


50 Annals of the 
Missouri Botanical Garden 


Volume 96, Number 1 
2009 


Cabral 51 
Revisión Sinóptica de Galianthe 


G. andersonii 


o 6. eupatorioides 


Distribución de especies de la sección Laxa. 


—A. Distribución de Galianthe andersonii, G. aurelii, G. 


Figura 12. 
berala G. chiquitosiana, G. eupatorioides. —B. Distribución de G. krausei, G. laxa, G. sudyungensis, G. verbenoides. 


A. Schinini 2111 (holotipo, CTES!; isotipos, G!, 

LIL!, MO). 
Distribución, bitat y — fenología. Paragua 

(Amambay, eet Cordillera, one Paragu ad 


en laderas de cerros, entre rocas desnudas de 200— 


300 m; florece y fructifica de noviembre a julio. 


Observaciones. Se reconoce por ser un sufrütice 


muy ramificado de 30-80 cm alt., de tallos glabros, 
con hojas membranáceas, linear-lanceoladas o lan- 
ceoladas de base decurrente formando pseudopecíolo, 
ind vaina estipular pubérula o glabra, con 
orescencias congestas, brevemente pedunculadas; 

di isco entero, se asemeja a Galianthe laxa su 
paraguariensis (Chodat) E. L. Cabral, por el hábito, 
forma y tamafio de las hojas, pero esta ültima especie 
tiene tallos pubérulos o pubescentes, hojas subcori- 
áceas con la haz pubérula y envés pubescente, vaina 
estipular pubescente y el disco nectarífero bipartido. 
Material d r estudiado. PARAGUAY. Amam- 
bay: Nacional Cerro Corá, 9 feb. 1982, J. 
a Casas 6133 (G, MO). Central: Areguá, cerro de 
Areguá, 23 jul. 1972, A. Schinini 5038 (MO) 
Piribebuy, iin mi, 27 abr. 1983, Soria 920 (CTES, FCQ). 
Guairá: Cer 924, T. Rojas 4885 (SD). 
Paraguarí: T Alta, Tebicuary Mi, 17 nov. 1978, L. 
Bernardi 18756 (MO). 


illarrica, abr. 1 


33. Galianthe bisepala E. L. Cabral, Bonplandia 
(Corrientes) 7: 4. 1993. TIPO: Argentina. Salta: 
La Vifia, Quebrada de las Conchas, entre la 
Salamanca y el Hongo, Ruta 68, Km 77—78, alt. 


m, 1 . 1990, L. Novara & S. Bruno 
9630 (holotipo, MCNS!; isotipos, B!, CORD!, 
CTES!). 


Distribución, hábitat y fenología. Norte de Argen- 
tina (Salta), en un área muy restringida de 1200- 
1300 m y en Bolivia (Cochabamba y Potosí), en 
lugares áridos, ladera de cerros, borde de camino de 


(1200-)2500-2750 m; florece y fructifica de enero a 
abri 


Observaciones. Sela reconoce por ser un sufrütice 


muy ramificado de 0.4-1 m alt, tallos pubérulos, 
hojas lineares o linear-lanceoladas con indumento 
variado en haz y en envés, cáliz con 2 sépalos, raro 
4 y semillas complanadas, ligeramente aladas. Por e 
hábito y las hojas angostas es afín a Galianthe krausei 
(Suess.) E. L. Cabral, pero esta especie tiene tallos y 
hojas glabras, el cáliz siempre tiene 4 sépalos y las 
semillas no son aladas 

n las observaciones de las etiquetas de los 
ejemplares de Bolivia, se menciona que localmente 

s usada como forrajera. 


E, 


Figura 10. Galianthe thalictroides. —A. Plan 
—D. Alabastro. —E. Flor 
—H. Alabastro. —I. Flor. ME ng uec cáliz, dis 
Semilla. —M. Cara dorsal. —N. Cara ventral. (A- M 
igura Gali nthe valerianoides 


estilo y e 


I-J. Semilla. 
E 21818.) 


—I. Cara dorsal. —J. Cara ventral. 


a. —B. Xilopodio. —C. Vaina estipular con lacinias. D-G. Flor longistila. 
—F. Hipanto, cáliz, esio y estigma. —G. Interior de la corola desplegada. H-K 
estigma. —K. 1 terior de la corola desplega 
Schinini 23467; H. 

A. Planta con xilopodio 


. Flor AE 
a. —L. Fru 
NG ede 21139.) 
. B-D. Elo longistila. —B. Flor. —C. Interior de la corola 
Flor. un Interior de la corola desplegada. —G. 
(A—D, Tressens 1885; E-J, 


52 


Annals of the 
Missouri Botanical Garden 


Nombre vulgar. En Bolivia “tholita” (extraído de 
observaciones de aec bnc por ser de porte 
similar, pero más reduc “tholas”, usadas 
como combustible (Lepido hoya anclada 


(Meyen) Benth. & Hook. f. y Baccharis tola Phil.). 


ARGENTINA. Salta: 
1963, A. Correa 521 (MCNS). 
BOLIVIA. pone mba: On the ascent from the Río 
Mizque going ard Totora from Aiquile, 2600 m, 20 Mar. 
1994, J. R. Wood 18137 (CTES, MO). Potosí: José M. Linares 
Lizarazu, Jatum Palmar, 2750 m, 167 km al E de la ciudad, 
bajando el camino hacia la quebrada, 5 abr. 1993, G. Torrico 
et al. 335 (CTES, LPB). 


Material representativo estudiado. 
Guachipas, Alemania, 14 abr. 


34. Sire the chiquitosiana x L. Cabral, Brittonia 

57(2): 142-145. 2005. O: Bolivia. Santa 

Cruz: Chiquitos Prov., S ee of Serrania de 

Santiago, 5-10 km E of town of Santiago de 

Chiquitos, 18°23'S, 59°30’W, 20 July 1983, D. 

. Daly et al. 2175 (holotipo, USZ!; isotipos, 
CTES!, MO!, NY!, SI!). 

Bolivia 


Distribución, hábitat y fenologia. En 


nta z), en campos, sometidos quemas 
frecuentes, también en laderas rocosas y campos 
rupestres, de 230-900 m; florece y fructifica de 


noviembre a julio. 


Observaciones. Se reconoce por ser un arbusto 


muy ramificado cerca l m alt, tallos glabros de 
corteza oscura, con hojas elíptico-lanceoladas, sub- 
coriáceas, glabras, plegado-nervadas, 2-3 pares de 
nervios secundarios basales, adheridos al nervio 
central del que se separan gradualmente hacia e 
ápice; y el interior de la corola con pelos moniliformes 
en los lóbulos y en la mitad superior del tubo. Por su 
hábito es similar a Galianthe verbenoides, pero esta 
, pubescent 


hojas discoloras 
papilas aculeadas sobre 


especie tiene es o 


a 
escabritisculas, con visibles 


n 


los nervios del envés, 3 ó 6 pares de nervios 
secundarios subopuestos, en arcos convergentes hacia 
el ápice y el interior de la corola tiene un solo anillo 


de pelos moniliformes en el tubo. 


Material idi ic estudiado. BOLIVIA. Santa 
Cruz: Chiquitos, Santiago de Chiquitos, campo rupestre, 
18°40'S, 50° 15 "w. 13 nov. 1997 B Mamani ei al. 1254 
(CTES, MO, SL USZ). 
35. Galianthe eupatorioides (Cham. & Schltdl.) E. 
L. Cabral, Bol. Soc. Argent. Bot. 27(3—4): 242. 
92]. Basónimo: Borreria eupatorioides 
Linnaea 3: 327. 1828. 
Spermacoce eupatorioides (Cham. & Schltdl.) 
Kuntze, Revis. Gen. Pl. 3(3): 123. 1898. TIPO: 
Brasil. Minas Gerais: Campina Verde, 26 mayo 
9 acedo 67 (neotipo, designado por 


Cabral [2003: 396], SP!; isotipo, MO). 


Nordeste de 


Argentina (nordeste de Corrientes y Misiones), Bolivia 


Distribución, hábitat y fenología. 


=> 


Chuquisaca, Santa Cruz), Paraguay (Alto Paraguay, 
Alto Paraná, Amambay, Boquerón, Caaguazú, Central, 
Concepción, Guairá, Itapúa, Misiones, Nueva Asun- 
ción), sur y centro de Brasil (Goiás, Mato Grosso, Mato 
Grosso do Sul, Minas Gerais, Paraná, Santa Catarina, 
Sáo Paulo); frecuente en suelos arenosos, lateríticos, 
bordes de camino, selvas y bosques submontanos, en 
campos cerrados, y como fijadora o estabilizadora de 
dunas, de 50-1600 g 


fructifica de marzo a julio. 


m; florece de agosto a marzo, 


Observaciones. especie se reconoce por ser 
de 0.5-2 m al 


sta 
un sufrútice muy ramificado alt., con 


inflorescencias + laxas, amplias, las semillas 
tienen surco amplio en la cara ventral, cubierto 
sólo en la línea media por el estrofíolo caduco. La 
pubescencia de la planta, tamafio, forma de las 
hojas y longitud de las lacinias, son caracteres 
variables de menor valor taxonómico. Se observaron 
que dichos caracteres variaban en la misma planta, 
en población de Bolivia y de Paraguay, mientras 
que los de Argentina y Brasil se mantienen 
constantes. 

Para la flora de Argentina el material mencionado 
como Borreria eupatorioides por Bacigalupo (1974) 
corresponde a Galianthe centranthoides. Del material 

e G. eupatorioides de la flora boliviana, analizado por 
Kuntze (1898), una parte identificó como Spermacoce 
eupatorioides [= 
resto confundió con B. angustifolia Cham. & Schltdl. 
[^ G. angustifolia]. Cuando transfirió a Spermacoce 


G. eupatorioides], mientras que el 


creó un nuevo nombre, $. chamissonis Kuntze. De 
manera que esa cita para Bolivia corresponde a 


material de G. eupatorioides. 
Material representativo estudiado. ee Corr- 


ientes: Ituzaingó, 5 km NE de Ituzaingó, camino a Apipé, 10 
abr. 1978, O. Ahumada et al. 2458 (CTES 2 


Saravia et al. 11457 (CTES). Reed Cruz: ani, jun 
1892, O. Kunize n (NY). ASIL. Cou Triangulo 
Mineiro, 26 943, A. a 67 (MO). Sáo Paulo 


Sao José do Rio aor ene. 1963, G. de Morns 90 (SP). 
o bue Serra do Aguapey, 4 mar. 1977, J. Kirkbride 
T al. 3056 (MO). Mato pow d 
991, O. Tiritan al. 451 
mar. 
Catarina: Morro a Pedras, 3 miar 1 
B). PARAGUAY. 1890, T. 
Paraguay: abino Mendoza-Gral. Garay, 31 abr. 
1995, R. Degen et al. ae (CTES, FCQ). Alto Paraná: Ea. 
Santa Elena, 12 km NE de Hern i 
Schinini et al. 27404 (CTES, G, MBM, MO 
13-15 km S de ruta 5, Cerro Cora, Cnia. Picada Lorito, 11 
dic. 1997, A. Schinini et al. 33727 (CTES, G, GH, MO, SI). 
Boquerón: Delegación de Nueva Asunción, 18 nov. 1992, 


a 


D 
3 
c 
[3 
= 
D 
Ed 
pent 

Dau 
E 
= 
rem 
m$ 


Volume 96, Number 1 
2009 


Cabral 53 
Revisión Sinóptica de Galianthe 


egen et al. 2897 (FCQ). Caaguazú: 20 km de 
Caaguazú hacia Yhú, 6 abr. 1965, O. Brescia et al. 5035 


63 km NE de Concepción: 19 abr. 1995, A Schinini et al. 
29553 (CTES). Guairá: Reserve 
Hernandarias, 7 ene. 
16 km N de Cnel. Bogado, 56°10" W, 2777'S, 21 mar. 1993, 
m rM et al. 27642 e BAB, bid CTES, F, x MA 

M, MICH, NY, gs US, WIS). Misiones: Ea. La 
TEMA in feb. 1955, T. Pe oen 3211 (CTES). m eva 
Asunción: Gral. E. 7 ago. 1992, F. Mereles et al. 


4572 (FCQ). 
36. Galianthe krausei A . E. L. Cabral, Bol. 
oc. Argent. Bot. 27(3—4): 244. 1991 [1992]. 


Basónimo: Borreria krausei Suess., Mitt. 


Staatssamml. München 1: 19. 1950. TIPO 
Paraguay. Cordillera: cerros de Tobati, Cerro 
Penitente, aridis, Jan., Fiebrig 755 


n xisis 
(holotipo, ch Seaton F!, K!, P). Figura 13. 
Borreria corymbosa f. microphylla Chodat € Hassl., Bull. 
Herb. Boissier, sér. 2, 4: 187. 1904. TIPO: Paraguay. 
Cordillera: In colle Tobaty, Mar., E. Hassler 4022 
(holotipo, G!; isotipos, BM!, F!, G!, K!, P!). 
Sufrütice sin xilopodio, muy o de 0.50— 
lm alt, tallos glabros. Hojas 7-35 


pseudoverticiladas, lineares, de margen revoluto, 


mm, 


glabras, nervios secundarios inconspicuos; vaina 
estipular de E 2 mm, Los pubérula con 2-5 
lacinias de 0 -7 mm. Inflorescencias tirsoideas, 
congestas, i " Hipanto 0.7-1.2 mm, turbi- 
nado, papiloso; cáliz con lóbulos triangular-subulados, 
0.7-1 mm, glabros, disco entero. Flor brevistila: 
corola 4.2-5 mm, lóbulos más cortos que el tubo, 
superficie interna con pelos moniliformes en el tubo, 
hasta la base de los lóbulos; anteras 1—1.2 mm, 


filamentos 0.7-1 mm; estilo 2.2-2.5 mm. Flor longi- 


Rn 


stila: corola 4—4.7 mm, lóbulos iguales o más corto 

que el tubo, superficie interna con pelos moniliformes, 

más gruesos en los lóbulos, cortos y finos en el tubo; 

4—4.5 mm. 

aue semillas 2-2.5 mm, pla- 
s, d 


anteras subsésiles, 1-1.3 mm; estilo 
Cápsula 2.5-3 m 
no-convexas, e e cara ventral con un 
surco profundo y Me po me del estrofíolo 


caduco. 


Distribución, hábitat y fenología. Paraguay (Cor- 
dillera y Paraguart), en Cordillera de Altos, en campos 
con suelos pedregosos de 200-350 m; florece de 


noviembre a febrero, fructifica de marzo a octubre. 


Observación. Borreria saxicola K. Krause (Krause, 
1908) [non Borreria saxicola K. Schum. (Schumann, 


1901)] es nom. illeg. 
Material ae estudiado. PARAGUAY. Para- 
p 1 Ybye 


guarí: Parque cuí, 15 Sep. 1988, E. Zardini 
7281 (CTES, MO. P 


. Galianthe laxa (Cham. & Schtdl.) E. L. Cabral, 
Bol. Soc. Argent. Bot. 27: 244. 1991. Basónimo: 

Cham. € Schltdl., Linnaea 3(4): 
337. 1828. Spermacoce laxa (Cham chltdl.) 
Kuntze, Revis. Gen. Pl. 3(3): 123. 1898. TIPO: In 
Brasilia meridionali lectam transmisit, Sellow 


s.n. (holotipo, B!, foto F 8809). 


Borreria laxa 
T 


Sufrútice erecto, a veces apoyante de 0.5-1.5 m 
alt; tallos glabros, pubérulos, pubescentes, de 
ángulos marcados, papilosos o con pelos retrorsos. 
Hojas 2-6(—10) X (0.2—)0.6-2(-3) cm, pseudoverti- 
ciladas, elípticas u oval-elípticas, discoloras, ápice 
agudo o atenuado, base aguda en pseudopecíolo, 
glabras o con pelos dispersos en la haz, pubescentes 
en el envés, o pubescentes en ambas caras, con 4-4 
pares de nervios secundarios, surcados en la haz y 
prominentes en el envés; vaina estipular 1.5— 
2.5 mm, pubescente, con 5-12 lacinias de 3-8 mm. 
Inflorescencias tirsoideas laxas, raro comprimidas. 
Hipanto turbinado, 1-1.7 mm, glabro, pubérulo o 
pubescente; cáliz 4-partido, lóbulos 1—1.7 mm, 
glabros con margen piloso; corola con papilas 
desarrolladas en el extremo superior de los lóbulos, 
a veces pubérula; disco entero. Flor brevistila: corola 
3—4 mm, lóbulos y tubo aproximadamente de la 
misma longitud, interior con pelos moniliformes, 
delgados en el tubo; anteras 0.7—1 mm, filamentos 
ca. 1 mm, estilo 1.5-2 mm. Flor longistila: corola 
2.7-4 mm, lóbulos mayores o iguales que el tubo, 
interior con un anillo de pelos moniliformes, finos y 
cortos en la mitad del tubo y sendas bandas de pelos 
gruesos en la mitad de los lóbulos; anteras subsésiles 
0.7-1 mm; 
turbinada, 


estilo 3.7-5 mm. Cápsula 1.7-3 mm, 
brevemente pedicelada, glabra o con 
pubescencia rala; semilla 1.7-2.5 mm, elipsoide, 


foveolada, castafio-clara, con surco amplio y pro- 


= 


undo en la cara ventral, estrofiolo grueso, irregular, 
facilmente caedizo, a veces persistente sobre el 


tabique interlocular. 
CLAVE PARA IDENTIFICAR LAS SUBESPECIES DE GALIANTHE LAXA 
la. Hojas elípticas, oval-elípticas, lanceoladas u oval- 


lanceoladas de 6-30 mm lat., glabras, pubescentes 
c 


membranáceas; tallos erectos y a veces apoyantes; 
d 0 cm 


UAM US lax: ong.; varios 
hábitats, 03000 m ...... Ta. G. laxa subsp. laxa. 
lb. Hojas ee lanceolata de 2-7 mm lat, haz 


és pubescente con pelos más densos 


, subcoriáceas; tallos siempre 
erectos; inflorescencia + congesta de 4-8 cm 


long.; campos rupestres, 100—300 m 
37b. G. laxa subsp. paraguariensis 


54 Annals of the 
Missouri Botanical Garden 


Volume 96, Number 1 
2009 


Cabral 55 
Revisión Sinóptica de Galianthe 


37a. Galianthe laxa (Cham. & Schltdl.) E. L. 
Cabral subsp. laxa. 


Borreria cristata S. Moore, J. Bot. 42: 101. 1904. Syn. nov. 
Galianthe cristata (S. Moore) E. L. Cabral, Bol. Soc. 
Argent. Bot. 27(3-4): 241. 1991 [1992]. TIPO: Brasil. 
Mato Grosso: Santa Ana da na 28 jun. 1902, A. 
Robert 368 (holotipo, BM!; isotipo, K!). 

S dre Standl., Publ. Field Columbian Mus., 

er. 8(5): 395. 1931. Syn. nov. TIPO: Brasil. Mato 

rosso: Serra da Chapada, in silva, 2 jun. 1903, G. O. 
Malme s.n. (holotipo, S!). 

a hábitat oe Norte y centro 

de Ar: a (Buenos Aires, Corri 


rgen entes, Entre Rios, 
Misiones, po Bolivia (Chuquisaca La Paz, Santa 


Uruguay (Colonia, Montevideo, San José, Soriano), en 
el sudeste del área, en campos bajos, a orillas de 
arroyos y ríos, en lugares abiertos o en el interior de 
bosques o selvas a 0-3000 m, cambiando hacia el 
norte, en Paraguay oriental, en campos rupestres de 
100-300 m, hacia el noroeste en Salta, Argentina, en 
fe entre 50 

Bolivia. Es la especie del s 


m, hasta llegar alturas de 
ubg. Galianthe 
más b su área de distribución llega hasta el 
delta del Paraná, Buenos Aires; florece de noviembre 
a febrero, fructifica de marzo a agosto. 

Observaciones. El material citado para la Argen- 
tina por Grisebach (1879) como Galianthe verbenoides 
Griseb. corresponde a G. laxa us axa. 

icina popular usada 
afecciones del hígado UP HEBR 9 
"rallada" (de la 


mo diurético y contra 
1974 


Nombre vulgar. En Argentina 
Pefia & Pensiero, 2004). 


Material T estudiado. ARGENTINA. Bue- 
la Martín García, Hg rs 1946, R. Palacios 73 
NW de Playadito, 


nos Aires: Isla 
(CTES). Corrientes ws 
981 


ruta 39, 14 n + Carnevali 5122 (CTES). Entre 
Ríos: Federación, 20 e 946, T. E id 11024 (CTES). 
Misiones: San cercanías del Piray-guazú, 17 j 


de y- ul. 
1949, E. pres “1993 (CTES). Salta: Depto. Iruya, 


Rd 15 abr. 1947, S. Pierotti 6606 (LIL). BOLIVIA. 


7, O. Buchtien 765 (MO, NY). Santa bar 
P ncia Que on límite de frontera, s 1989, 
reles et al. 2873 (CTES, FCQ, G). Tar aes 22°12’ a 
pes 7'W, 27 abr. 1983, J. Solomon 10159 (MO). BRASIL. 
Goiás: ape ourada, ca. 15 km S of ra bass 10 mayo 
1973, . Ander. pie 10010 (UB). M Mun 
Pedra a Rod. BR-364, Serra da EM 16 m on 
1995, G. Hula et al. "s (CTES, M BM). M 
Grosso ul: Mun. Corumbá, Sub-regiáo da Nhecolán dia. 
" M 1985, a Pat 2023 us CTES). Minas Gerais: 
n. Santa Rita do Sapucai, Vintém de are 19 feb. 1996, 
^s "S Ribas 1288 (CTES, MBM). Paraná: 23 nov. 1972, L. 
Dombrowski et al. 4309 C Rio Grande p sul nes Ch. 
M EAR 1114 (P). Santa Catarina: s Guar- 
a, 3 ene. 1964, R. Reitz ou 16955 on res Panlo: 
Mun a ions Reserva Estadual de Porto Ferreira, 30 
jun. 1981, J. Bertoni 18652 (CTES, UEC). PARAGUAY. 
Alto Parana: 1910, K. Fiebrig 124 Amambay: Rio 
Apa, 1908, K. Fiebrig 4323 (B, G, P). Caaguazú: Yhú, 12 
dic. 1982, A. Schinini 22915 (CTES). EE Tavai, 5 km 
del destacamento, 26°10'S, 55%17'W, 15 mayo 1989, N. 
Soria 3857 (CTES, FCQ). Canaan: 46 km s de Katueté, 
18 dic. 1982, A. Schinini 23198 (CTES, G). Central: Lacus 
Ypacaray, 1913, E. Hassler 


BAB, NY). URUGUAY. Coloni 
1-9 mar. "1964, "B 
Arrillaga et al. 1898 (SI). Montevideo: M. Fruchard s.n. (B). 


37b. Bcc riw s paraguariensis (Ch 
t & Hassl. L. Cabral, Candollea 58(1— 2. 
p 03. Wc Borreria paraguariensis 
ee Hassl., Bull. Herb. Boissier, sér. 2, 4: 
186. Borreria laxa Cham. & Schltdl. var. 
vestita T Sm. & Downs, J. Washington Acad. 
Sci. 48: 284. 1958. Galianthe paraguariensis 
(Chodat & Hassl.) E. L. Cabral, Bol. Soc. Argent. 
Bot. 27(3—4): 245. 1991 [1992]. TIPO: Paraguay, 
Cordillera: In dumetis Cordillera de Altos, Sep., 
E. Hassler 3263 (lectotipo, designado por Cabral 
[1991: 245], G!; isotipo, P!). 


Borreria paraguariensis f. puberula Chodat & Hassl., Bull. 
Herb. Boissie 2, 4: 186. 1904. TIPO: Paraguay. 
mambay: In glareosis collium prope e Paraiay, Dec., 

E. Hassler 6508 (holo otipo, G!; isotipo, G!). 
Borreria paraguariensis f. latifolia Chodat € Hassl., Bull. 
Herb. Boissier, sér. 2, 4: 186. 1904. TIPO: Paraguay. 


= 
Figura 13. Galianthe krausei. —A. Rama con inflorescencia. —B. Raíz axonomorfa. 
D-F. Flor longistila. —D. Flor. —E. Interior de la corola desplegada 
dig e —G. Flor. = Interior de la corola d. 
K. Cara ventral. —L. Cara dorsal. Ae EE iaa A a L, D 
EU 14. Galianthe rabeno ie 
F. Flor longistila. —D. Flor. —E. 
—G. Flor. —H. Interior de la corola desplegada. —I. Hipa: 


esplegada. —I. ae cáliz, estilo y e a. —]. F 


n ae dms y bere —F. Interior de la corola RUN A G-I. 
nto, cáliz, estilo y estigma. —J. Fru 


— C. Vaina estipular con lacinias. 

Hipanto, ina dia a n G-I. Flor 

o. K-L. Semilla. 
n 480. 

aeu oh con lacinias. —C. Base foliar y parte del hipófilo. D- 

Flor brevistila. 

o. K-L. Semilla. —K. Cara 


dorsal. —L. Cara ventral. (A-I, Bueno 3642; J-L, Hagelund 15744.) 


56 


Annals of the 
Missouri Botanical Garden 


Amambay: ad marginem silvae prope Bellavista, ia 
Nov., E. Hassler 7999 (holotipo, G!; isotipos, G!, P!). 
Distribución, hábitat y fenologia. En Paraguay 

(Amambay, Central, Paraguarí y Cordillera), 
campos rupestres de 100-300 m; florece de septiem- 
bre a mayo, fructifica de junio a agosto. 


Material O estudiado. PARAGUAY. Cen- 
tr Ypac: 


and Nueva Colombia, 25 Jane 1992, E. Zardini et al. 32302 
(AS, CTES, MO). Paraguarí: National Park Ybycuí, 22 June 
1991, E. Zardini et al. 27865 (CTES, MO, PY). 


38. es udyungensis E. L. Cabral, Brittonia 
57(2): 145. 2005. TIPO: Bolivia. La Paz: Pro 
Sud Yungas, Río Jucumarini, 16°37’S, 67° HA 
14 abr. 1990, M. Lewis 37202 (holotipo, LPB!; 
isotipos, CTES!, P!) 


Distribución, hábitat y fenología. Bolivia (La Paz, 
Sud Yungas), en campos altos de 985-1500 m, florece 
de enero a abril y fructifica de abril a julio. 


Observaciones. Galianthe sudyungensis se carac- 
teriza por ser un sufrútice muy ramificado, de 30— 


50 cm alt., 


ón linear-lanceoladas a lanceoladas, 


de tallos tetrágonos, rojizos, con hojas 


discoloras; vaina estipular pubescente, con 5-9(— 
lacinias, glabras; inflorescencias comprimidas, las 
semillas subcomplanadas, finamente aladas en el ápice 
y con el estrofíolo caduco en la cara ventral. Galianthe 
sudyungensis es muy similar a G. aurelii, ambas 
especies son muy ramificadas y tienen hojas pat 
verticiladas, pero difieren por los siguientes caract 


G. sudyungensis es diferenciada por el tallo y las hojas 
pubérulas a pubescentes (vs. glabros en G. aurelii), 
lacinias nae rojizas de 5-9 mm (vs. verdosas de 

26 mm or de la corola con pelos solo en el tubo 
(vs. pelos en a pes y en los lóbulos). 


Material representativo estudiado. BOLIVIA. Sud Yun- 
gas: ridge of Pasto Grande, 16737'S, 67729" W, 12 abr. 1990, 
M. Lewis 37173 (CTES, LPB, MO). 


39. Galianthe verbenoides (Ghats S: E: 
riseb., Symb. Fl. Argent 1879 

Basónimo: Borreria verbenoides E prima Gham, 

& Schltdl., Linnaea 3(4): 333. 1828. Spermacoce 

verbenoides (Cham. & Schltdl.) Niederl., Bol. 


superfl. Borreria verbenoides Cham. & Schltdl. 
var. eupatorioides (Cham. & Schltdl.) L. B. Sm. & 
h. Acad. Sci. 48: 284. 1958. TIPO: 


owns, J. Was 
il. In Brasilia meridionali legit Sellow 


Brasil. 
(lectotipo, designado aquí, LE!). Figura 14. 


Sufrútice ramificado, de 0.6—1 m alt., 
laterales desarrolladas, opuestas, tallos tetrágonos a 


con ramas 


subcilíndricos, glabros a veces con pelos o papilas 
aculeadas, mameliformes, con frecuencia cerca de las 
15-120 3-17 mm, 


elípticas, de ápice agudo a atenuado, base aguda 


estípulas. Hojas oblongo- 
prolongada en pseudopecíolo, discoloras, levemente 
pubescentes o escabritisculas, con papilas aculeadas, 
de base gruesa sobre los nervios del envés, más 
notables en el nervio medio, margen revoluto, 
escabriüsculo, con 3-6 pares de nervios secundarios 
en relieve en el envés, levemente surcadas en la haz; 
vaina estipular de 2-4 mm, pubérula o pubescente 
10-11 lacinias, de 7-12 mm. Hipanto 1-1.5 mm, 
turbinado, glabro; cáliz de lóbulos 1—1.5 mm, triangu- 
lar-subulados, glabros; corola con papilas densas hacia 
el ápice de los lóbulos; disco entero. Flor brevistila: 
corola (2.5—)4.5—4.7 mm, lóbulos iguales o más cortos 
que el tubo, superficie interna con un anillo de pelos 
moniliformes en la mitad del tubo; anteras 1—1.3 mm, 
filamentos ca. 1.5 mm; estilo 1.5-2 mm. Flor long- 
istila: corola 3-3.2 mm, lóbulos iguales o más largos 
que el tubo, superficie interna con pelos moniliformes 
cortos, en la mitad del tubo y algunos pocos en la base 
de los lóbulos, anteras subsésiles, 0.7—1 mm; estilo 
2.5-3.7(-4.2) mm. Cápsula 1.7-2 mm, subglobosa, 
glabra; semillas 1.5-1.7 mm, er eee escrobi- 
culadas, con un surco en la cara ventral alrededor del 
estrofiolo persistente. 
Sur de Brasil 
Santa Catarina), 
yos y en 
m; florece y fructifica de 


Distribución, hábitat y fenologia. 
Rio Grande do Sul, 


frecuente en bordes de camino, de arro 


(Paraná, 


campos altos de 0-3 
noviembre a abril. 


Observación. Smith y Downs (1958) proponen 
para el basónimo Borreria verbenoides la varieda 
eupatorioides (Cham. & Schltdl.) L. B. Sm. & 
Downs, pero en este trabajo se trata cada una de 


las especies: x verbenoides (— Galianthe verbenoides) 


identificado como B. verbenoides var. eupatorioides 
fue consultado en el Herbario HBR y se comprobó 
que corresponde a G. centranthoides (Cham 
Schltdl.) E. L. Cabral. 

Este concepto fue seguido por Delprete et al. 
2004) en la Flora de Santa C 


además agregan 


atarina, Brasil, 


— 


de 
como sinónimos a e iib: 
ianoides y G. chodatiana (secc. Caliaatho) que aquí 
son tratados como especies diferentes. 
Por considerar que el original de Galianthe 
verbenoides ha desaparecido en el herbario B (foto 
890), se elige como lectotipo, el ejemplar duplicado 


Volume 96, Number 1 
2009 


Cabral 57 
Revisión Sinóptica de Galianthe 


del herbario LE y además está en buen estado de 
conservación. 


Material no o m BRASIL. Paraná: 
Lapa, 5 abr. 1960, R. al. 266 (NY). Rio Grande 
do Sul: 1833, C. dica 1 1 18 (P); Arroyo dos Ratos, 30 
ene. 1985, E. Cabral 486 (CTES). Santa Catarina: Irani, 27 
feb. 1964, Klein 4719a (B). 


Literatura Citada 


Bacigalupo, N. M. 1974. Rubiaceae. En A. Burkart a 

Flora Ilustrada de Entre Ríos. Colecc. Ci. Ins 
ecno u. 6(6): 17-27. 

81. Novedades en el géne 

(Rubiaceae). para la flora de Corrientes. 
5(16): 143-148. 

——. 1985. Valor taxonómico del polen en las especies 
prO del género Borreria (Rubiaceae). Bol. Soc 
Argent. e 4 24: 169-178. 

1. Rehabilitación del género T aM (Rubia- 

aab). A Soc. Argent. Bot. 27: 235-24 

2. Revisión del Género onm he (Rubiaceae). 

Tesis Doctoral, Universidad Nacional del Nordeste, Corr- 

ientes, Argentina. 

. 2003. Novedades en el género Galianthe Griseb. 
(Rubiaceae- Spermacoceae) para la flora de Paraguay, en 
L. Ramella & P. Perret, Notulae ad floram paraquaiensem. 
Candollea x 392-398. 

— & N ee 1997. a del género 
EU subg: Ebelia v. (Rubia cum 
Na = Missouri Bot. Card p 857- “877 

2001. Scandentia, nuevo Dp de 

a 39(1-2): 2 

an A. P. de. 1830. Prodromus systematics m 

regni vegetabilis. Treuttel € Wiirtz, Paris. 

E. Hassler. 1904; Plantae Hasslerianae II. 
189. 


ro Borreria 
Bonplandia 


maco- 


. Darwinian: 


Daviña, J. & E. L. Cabral. 1991. Recuent 
Galianthe (Rubiaceae). Bol. Soc. pues Bot. cal 
252. 


de la Pefia, . Pensiero. 2004. COR 
Argentinas. Catálogo de Nombres Comunes. L.O.L 
Buenos Aires. 

Delprete, P. G., L. B. Smith & R. M. Klein. 2004. Galianthe. 
Rubiáceas. Pp. 213-272 en R. Reitz (editor, Flora 
Ilustrada Catarinense. I Parte, Vol. I—Generos de A-C. 


n the Spermacoceae 
Katholieke Universiteit 


P. M 1990. Estudo Fitogeográfico do 
Brasil. . Siylus Comunicaçoes, Fortaleza. 
Galati, 88. Estudios Embriológicos en la Tribu 
a (Rubiaceae). un Doctoral, Universidad 
de Buenos Aires, Buenos Air 
. 1991. 


Estudios Bic en la tribu Spermaco- 
ceae (Rubiaceae). Parte I: Anatomia pon iaa cs 
esis oc. Argent. B -2)y 1- 


Grisebach, A. 1879. lolis ad je [oue pes 


al de Sao Joaquim. Superintendéncia de Santa 
Catarina, eee: is. 


Kiehn, N. M. 1985. Karyosystematic studies on Rubiaceae: 
Chromosome pos s from Africa, Madagascar, and Maur- 
itius. Pl. Syst. Evol. 149: 89-118. 

1986. “Karyolögische Untersuchungen und DNA 

Messungen an Rubiaceae und ihre Bedeutung für die 

Systematik dieser Familie. Diss. Form.-Naturwiss. Fak., 
Univ. i 


is survey of the Rubiaceae. Ann. 
Missouri i. Gard. 82: 398-408. 

Krause, K. 1908. XE m andinae. Bot. Jahrb. Syst. 40: 
348. 


4. Arthur Felix, Lei 
en V. & O. Machado. 1944. Contribucáo ao mu das 
Rubiaceas medicinais do Brasil. Revista Fl. Med. 11(1): 
1-34. 
Martinez Crovetto, R. 1981. Las plantas utilizadas en 
medicina popular en el noroeste de a (Repüblica 
Misc. Fund. Miguel Lillo 69: 1-139. 
1997. Género Galianthe subg. fux (Rubiaceae: 
Bocimicosome Estudio palinológico. Ann. Missouri Bot. 
Gard. 84: 87 ar 
— & E. L. Cabral. 1992. El valor del polen en E 
revalidación de 


D C. O. 1898. Revisio Generum Plantarum, Pt. 3: 123— 
eipzi 


Galianthe Rubia 
0. 


C. Jacques, S. T. Miotto, J. L. Waechter & 

E Detoni. 195 Tribu 2 m (Rubiaceae). En 
A. R. Schulz (editor), Fl. IT. Rio Grande do Sul. Bol. Inst. 
i Bioci. Univ. Fed. Rio Grande do Sul, Sér. Bot. 35: 


Robin E. 1988. Tropical Woody Rubiaceae. pagina 
stic features and progressions. ee ons 
subfamilial classification. Opera Bot. Belg. 1 Lens 
Schumann, 88. Sa En C. Martius, Flora 
Brasiliensis '6(6): 62— 
1899. pem Africanae, Beitrage zur Flora von 
Alca. XIX. Bot. Jahrb. Syst. 28: 112. 
Smith, L. B. 1958. Notes F South PR phanerogams. J. 
ash. Acad. Sci. 48: 


APÉNDICE 1. Lista d 
al Apéndice 2. 


a Galianthe andersonii E. L. Cab 
2. Galianthe angustifolia [m & El E. L. Cabral 
3. Galianthe aurelii E. L. Cabra 
i Galianthe bisepala E. L. c al 
5. Galianthe canindeyuensis E. L. Cab 
6. Galianthe centranthoides (Cham: & PO E. L. Cabral 
1. Galianthe chiquitosiana E. L. Cabral 
s Galianthe chodatiana (Standl.) E. L. Cabral 
9. Galianthe cyperoides (Chodat € Hassl.) E. L. Cabral 
10. Galianihe elegans E. L. Cabral 
11. Galianthe equisetoides (Cham. & Schltdl.) E. L. Cabral 
12. Galianthe eupatorioides (Cham. & Schlidl.) E. L. Cabral 


16. Galianthe guaranitica. (Chodat & Hassl.) E. L. Cabral 
17. Galianthe hassleriana (Chodat) E. L. Cabral 
oA Galianthe kempffiana E. L. Cabra 
9. Galianthe krausei (Suess.) E. L. Cabral 
abr " 


22. Galianthe laxa p & DD E. L. Cabral 
axa subsp. fax, 


G. laxa subsp. paraguariensis (Chodat) E. L. Cabral 


58 


Annals of the 
Missouri Botanical Garden 


23. Galianthe liliifolia (Standl.) E. L. Cabral 
24. Galianthe linearifolia E. L. Cabral 

25. Galianthe m o n 2 Cabral 
26. Galianthe longisep 

27. Galianthe macedoi I E A 
28. Galianthe matogrossiana E. L. Cabral 
29. Galianthe moniesii E. L. Cabral 

30. Galianthe parvula E. L. Cabral 

31. Galianthe peruviana (Pers.) E. L. Cabral 
32. Galianthe pseudopeciolata E. L. Cabral 
33. Galianthe ramosa E. L. Cabral 


mi 


38. Galianthe valerianoides (Cham. & Schltdl.) E x Cabral 
39. Galianthe verbenoides (Cham. & Schltdl.) Griseb. 


APÉNDICE 2. Indice de colecciones. Cada especímen es citado 
] $ le] i l 3j Zola 


Ua r ES 

colectores hayan = de la colección; si fueron dos 

kn ectores se citan ambos. Se indica entre paréntesis el 
úmero K orden del taxón a que se corresponde (véase 

pinos 


BBOTT, R. 16495 (35); ABRUZZI, M. L. 
oa O. 1002 (37a), 2458 (35), 3038 (14), 4048 (7), 
4068 (29), 4079 (8), 6261 (10), 6375 9. 6718 (37a), 6799 
(7), 7332 (37a); AHUMADA, 337 (30), 3361 (27); 
AL (26), 464 (26), 529 (16), 868 (16): 
ALMEIDA, S. P. o Qo ALVARENGA, D. 534 (26), 604 
(16), 712 (10); A 5 (10); ANDERSON, W. R. 
6261 (10), 8529 pi "8717 (31), 9599 (20), 9835 (37a), 
10010 (37a), Ter Qo. 10109 (13), 12436 (24), 12479 
(24), 35231 QU RE P. 1192 (1); ANISITS, J. sn. 
(3), 514 (3); A oM . 1893 (8), 1952 (35), 2071 (8), 
2138 (35), 2436 (8), 2450 ix 2875 (3), 3024 (26), 3038 (14), 
3366 (26), 4153 (24), 4167 (24), 4581 (10), 4597 (24), 4896 
(24), 5019 (10), 5197 (10), 5889 (37a), 6144 (8), 6468 (35); 
ARENAS, P. 291 (36), A (24), 1217 (37b), 1306 (37b), 
5 86) 2911 (37b); ARRILLAGA, B. 1898 (37a); ASSIS, 

08 (24); AZEVEDO, M. L. 582 (10). 

pos N. 1151 (32); BADINI, J. 2086 (1), 4416 
(30), 19677 (10), 20647 (30), 21996 (1), 21997 (1), 22703 
s 8 (1), 24029 (30), 25592 (1) 26193 

ALANSA, B. 1742 (30), 1742a (30), 1743b (3), 1743c 
cs 1744 (6), 1745 (2), 1746 (29), 1747 (5), 1748a (32), 1748 
(37b), 4551 (30 
B , G. 21 B 112 (35); BARRETO, K. 2976 (28), 
TTURA, N. 1157 (30); BECK, S. 6636 (35), 


~ 
E 
v 


ESCIA, O. et al. 3957 (8), 
4089 (15), 4454 (37b), 5035 (35); BRITEZ, R. M. s.n. (30); 
37a); BUCK, H. 28042 (39); BUENO, 
O. 1154 (39), 1234 (39), 3624 (39), 17897 (39); BURKART, 
A. 8367 (37a), 19740 (30), 23147 (37a); BUSELATO, L. 19 
(37a). 
CABALLERO MARMORI, G. 255 (35), 860 (17); 
nary PARDO, R. 24 (33); CABRAL, E. 159 no 
66 (8), 180 (35), 181 (8), 242 (35), 274 (35), 326 (35), 337 
a 356 (35), 387 (8), 461 (29), 470 (29), 473 (3), 478 (8), 


486 (39), 521 (35), 624 (35), 667 (3), 668 (35), 669 (8), 674 


57508 (3); CAMPOS, S. M. 148 (16); CAMPOS NOVAES, J. 
s.n. o CAMPOS PORTO, P. 3266 (1, 3267 (301) 
CAPELL, P. s.n. (1); CARNEVALI, R. 5122 (37a), 6401 
(8) CASTELLANOS, d 21761 (26), 24690 (4); CHARPIN, 
A. 21387 (35); C T, R. s.n. 8) 279 60. 378 (37b), 
799 (87b), 810 Eum C . 1839, s.n. (10), 1840, 
sn. (10), sn. (16), 17 (16), 205 (10), 1841, sa. (16); 
CORDEIRO, J. et al. 392 (4), 890 (4), 932 (9), Gg (18); 
CORREA, A. 521 (83); CORREA, M. . 7930 
(37a); CORREA GOMES JR., J. 1437 (23), 1655 do. 1747 
(35); CRUZ, J. M. 240 (4); CUSATO, L. 1327 (37a), 1429 


Ta). 


= 
oo 


. C. 2175 (34), 2240 (37a); DA SILVA 
1915, (26); DAVIDSE, G. 10523 (10), 11294 (35); DAVINA, 
J. 146 (35); DAVIS, P. H. 60014 (10), 60147 (26); DAWSON, 
E. Y. 14276 (24); DE BARROS, F. 2517 (10); DEGEN, R. 
154 (37b), 210 (37a), pe a 480 (36), 2897 (35), 3301 
35); DE LIMA, H. C. 1020 (10); DE LIMA, J. 48989 (29); 
DEL PUERTO, O. am A 7607 (3); DE MARINIS, M 90 
(35); DEMATTEIS, M. 720 (8); DESCOBERTO, S. 
(10); DE SOUZA, H. C. 10929 (1); DETONI, E. M. 65 v: DI 
FERNANDO, sn. (8); DOMEN J. sn. (24; DOM- 
30), 1711 (4), 2231 (4), 2920 
4), 3862 (4), 4309 (37a), 4553018) 5515 (4), 6462 (6), 6682 
18), 6772 (4), 6869 (18); DUARTE, A. P. s.n. (16), 661 (30), 
2336 (16), 2529 (24), 8475 (1), 8805 (24), 9622 (24), 9694 

1), 9950 (26), 10051 (16), 10235 (1), 10397 (30), 13073 
(16; DUARTE, a 10373 (10 
na R. 106 (2); DUSÉN, P 18 

; 7534 (3), i (30), 9446 (18), 10825 (4), 13384 
i (9), 16033 (29), 16269 (18). 

, G. s.n. (30); EGLER, W. 60000 (3); EGYDIO, 

L 2 (8); EKMAN, E. L. 1350 (9); EMYGDIO, L. 3111 as 

FALKENBERG, D. B. 4320 (9); FERNANDES, 
37a); FERNANDEZ, J. 668 (7); FERNANDEZ ae F. ki 
5768 (30), 5963 (5), 5983 (11), 6133 (32); FERRARO, L. 
2420 (37a), 2465 (37a), 2978 (35); FERRUCCI, S. 497 (37a), 
34 (81b), 750 (37b), 759 (37b); FIEBRIG, K. 124 (37a), 500 
(3), 650 (8), 755 (5), 2729 (37a), 4323 (37a), 6076 (30), 6426 
30), 6491 (8); FILGUEIRAS, T. S. 
LEIG 197 (39); FLOSSDORF, 
FONTANA, J. L. 259 (30), F55-83 (29); OTI J. 
1285 (4); FONTES VIEIRA, R. 622 (26); FRANCO, C. E. 4 
10); FRAZÁO, A. 1917 (16); FRIEDRICHS, L. 28807 (1); 
FROMM, T. 408 (6); FRUCHARD, M. s.n. (37a), 203 (37a), 
404 (8); FUENTES, A. 1782 (13), 2598 (35); FURLAN, A. 
SPF 32772 (24). 

GALLINAL, G. 5610 (8) us d rur (35); 
GARCÍA, A. 03 (85; GAR W. G. 4 (26); 
GARDNER, G. 3785 (10); CA BIET zn sa. (3), 
1114 (37a), 1115 (3), 1117 (8), 1118 (39); GEHRT, G. 3529 
(3); GILBERT, O. s.n. (8), 86625 (8); GINZBARG, S. 379 (8), 
480 (35), 521 (35), 560 (35); GIULIETTI, A. M. 13601 (1); 


= 


= Bes 


ges 
— 


4). 


= 


Au 


~ 


GUAGLIANONE, R. 2837 
GUIMARÃES, E. 47 (37a), 2 
HACKER, J. 1000 B (35); NKE, T. sn. (24); 
HAGELUND, K. 1408 (29), 3 E ie (3), 13744 
39), 14801 (8); HAHN, W. s.n. (37a), 848 (5), 1615 (35); 
HANDRO, O. 29 (3); HASSLER, E. e (3), 1814 (30), 


(7a) GUILLEN, R. 4774 es. 
e 


= 


Volume 96, Number 1 
2009 


Cabral 59 
Revisión Sinóptica de Galianthe 


3263 (37h), 3314 (3), 3714 (3), 4022 (36), 4249 (3), 4338 (5), 


(37a); HATSCHBACH, G. 1113 (o 1917 1 GTa) 4288 (29), 


58771 


a 70740 m 70867 09), 71162 


En 
SK 
e 

E 
S 
es 
Eg 
co 
[€ 

as 
ay 

E: 
=2 
=D 
ax 
H 
ll 
= 
v- 
ao 

¡qa 

© 
Ra) 
[s 

mns 

u 
am 
E 
2 
N 
fy 
E 
9o 
ne) 
y 
x 


15742 (26), 16726 (26), 21962 (24); HERTEL 
HERTER, R. 86625 ae UD 184 n S EC. 
n. (10), s.n. (30), s 19 (14), s.n. 14511 (14), 10860 
(30), 13513 (30), ie x pon s 209 ee 28542 (1); 
, W. s.n. (30), s.n. (37a; H 
(24); HOLMBERC, E. us (35); HOLMIGREN 0. 25450 
(35); HUIDOBRO, A. 4827 (8); H T, H. et al. 
30644 (24); AETA I. et al. 3356 Bs 
IBARROLA, T. 3774 (3), 4250 (35); IMAGUIRE, I. 1734 
(9), 5682 (4); INSFRÁN, P. 1128 (37b); IRGANG, B. 783 
(7); IRWIN, H. 1879 (37b), 1989 (1), 3263 (37b), 3783 (37b), 
4022 (36), 6508 (37b), 10223 (26), 10616 (26), 12244 (16), 


. 366 (24), 2242 (a) JIMENEZ, B. 2001b 
(5); na J. 1235 (13); JOLY, A. B. 932 (24), 2292 (24); 
JOLY, J. 866 (24); JÓRGENSEN, P. 2179 (3), 2634 (3), 3520 
(5), 3706 (37a), 3714 A (3), 3714 (30), 4308 e» o (29), 
4922 (25), 4926 e pum 77 (37a); 
JÜRGENS, G. 4 

KIRKBRIDE, T i25 (26), 1745 (30), 3056 (35), 3162 
(10), a (26), 4686 (26), 4765 (10), 4994 A (26), 5334 
(10); KIRKBRIDE, M. C. G. 1010 (26), 1077 (26); KILL- 
EEN, n A (13), 5230 (13), 6492 (13), 6530 (13), 7405 
(37a); KIRIZAWA, M. 403 (10); KLEIN, R. 587 (37a), 2185 
(37a), 3242 (4), 3360 (30), 3544 (30), 3709 (30), 3814 (o, 

19a (39); KÓRNER, H. 57332 (39); KRAPOVICKA 
3056 (35), 12033 (8), 12066 (8), 12413 (37b), 12478 Bis 
13382 (37b) 13477 (37b), 14117 (35), 15120 (8), 15239 
16388 (8), 16568 (7) 


24883 (8), 25039 (37a), 25186 (30), 25651 (8), 2575 
26151 (29), 28697 (35), 28791 (3), 28873 (37a), 31404 (37a), 
31562 (35), 31673 (35), 32281 (35), 34154 (8), 34214 (8), 


w 
A 
E 
co 
= 

pz 
o3 
NI 
e 

E 
w 
[e 
ne] 
co 
© 

ein 
we 
n 

EL 
w 
MI 
e 
L1 
o 

Pet 
[25 

we 
w 
AZ 
e 
Oe 
NI 

Ves 

= 
w 
MI 
co 
S 
> 

ux 
Y 
= 


ae 
iu! 

QS 
EN 
nN 
Oo 
= 
ws 


(30), 43020 (14), 43155 (14), 43486 (37a), 43925 (35), 


44664 (8), 45169 (31b), 45444 (35), 45529 (37b), 45672 (8); 
EFGER, G. 7638 (4); KRIEGER, L. 16223 (10); 
HLMANN, J. G. s.n. (14; KUHLMANN, M. 2487 (1), 

3582 (10); KUMMROW, R. 85 (4), 663 (18), 1401 (4), 1795 

(4), pecs (4), 3397 (9). 

NDRUM, L. 2807 (1); LANGERON, n sn. (8); 

M UM. P. 5980 (37a), 10194 (37a); LEG 

2015 (3); LEITÃO FILHO, H. F. 9293 (35), 9315 c 15319 

(1), 32986 (37a); LEITE, E. 3468 (25); LEITE, P. s.n. (1), 

2296 (30); LEWIS, M. 37173 (38), 37202 (38); LIMA, A. 131 

(37a); LIMA, S. 58-2979 (10); LINDEMAN, J. C. s.n. (30), 

309 (4), 6022 (10); LORENTZ, P. G. s.n. (37a), 504 (37a), 

803 (8), 804 (3), 1486 (3), 1729 (3); LOSSEN, W. 563 (7); 


LURVEY, E. 249 (8), 256 (3), 294 (37b). 
MACEDO, A. 67 (35), 1468 (20), 3506 (26); rc 
37a); MAGNANO 221 


MARTINS, F. R. 247 (35), 1677 (1); MARTINS, S. A. 4 
(3); MARTIUS, C. 997 (24); MATTHEWS, D. M. 1150 n 
MATTOS, J. 12240 (10), 13973 (29), 14323 (1; MATTOS 
FILHO, A. s.n. (4); MeDANIEL, S. 11968 (37b); MEDINA, 
B. 103 (22), 170 (8), 181 (35), 184 (8), 287 (35); MELLO, F. 
425 (14); MENACHO, M. 596 (35), 711 (35); MENDONCA, 
R. 1191 (19; MENEGES, J. 6236 (24); MENEZES, N. 6236 
24); MENHOFER, O. X1411 (35; MERELES, F. 1525 
31b), 2371 (30), 2779 (35), 2872 (35), 4572 (35), 5023 (35); 
MEYER, T. 920 (37a e du 11583 (35), 11691 (35); 
MICHEL, R. 195 (35); M 
MOLLURA, R. 
(35); MONTES, J. 575 (29), 723 (35), 1287 (8), 1475 (35), 
1633 (8), 1720 (8), 7036 (22), 9418 (35), 9448 (29), 9764 
(22), 10323 (8), 14717 (35), 15072 (35), 15159 (35), 15495 
(35), 15694 (35), 27241 (35), 27310 (35); MONTOVANI 
7792 (24); MOREIRA, I. 3 (6); MOREIRA FILHO, H. 472 
(6); MORONG, T. 610 (35; MORRETES, B. s.n 
MORRONE, O. 968 (37a), 1017 (35), 1023 (35), 1867 (35); 
MOSTACEDO, B. 1778 (13), 1853 (13), 2121 (13), 2160 
(13); MROGINSKI, L. 320 (35), 344 (37a); MUSCON, M. 


EAT 


oo 
eo 


DE ANDRADE, E. 0); NEE, M. 35410 

(37a), 35763 (35), 35805 (35), 36374 (35), 37729 (35), 
38584 (35), 39065 (37a), 39444 (35), 39719 (35), 40362 
35), 40489 (35), 41607 (35), m [o 43413 24), eo 
37a), 46651 (24); NEIFF, J. J. 593 (35); NETO, E. 2863 
16; NICORA, E. 4625 (37a); Torun E. n (26); 
NOVARA, L. 2777 (33), 9630 oe 

LIVEIRA, E. 173 (7); OLIVEIRA, J. SPF 36156 (24); 
OLIVEIRA, P. I. 173 (18), 331 (30), E (30); ORBIGNY, A. 
d', 550 (35), 882 (34), 950 (35); OSTEN, C. 3958 (3), 9085 
(37h), 18954. (37a), 20187 (3), 20188 

PABS, G. 5449 (7), 6423 (7); PALACIOS, A. 286 9. m 
(24); PALACIOS, M. 3861 (1); PALACIOS, R. 
PALACIOS-CUEZZO, O. 880 (39), 1318 (8), 1412 e, 2; 
(37a); PARTRIDGE, W. 60650 (7), 68467 (7); PAVÓN, J. 3 
(24); PEDERSEN, T. 181 (3), 801 (8), 3211 (35), 3233 (à, 
4781 (8), 5405 (7), 5418 (37a), 5432 (30), 5452 (29), 8192 
(37a), 9283 (32), 9519 (5), 12138 (14), 15616 (10), 15795 
(37a); PEREDO, I. s.n. (35); PEREIRA, B. 467 (30), 1476 
(24), 2054 (10); PEREIRA, E. 1392 (10), 2417 (1), 3211 (35), 
5157 (4), 5418 (87a), 7651 (37a), 7971 (3), 8087 (4), 8192 
(37a), 8376 (37a), 8486 (8), 8526 (7), 9283 (32), 15795 (37a), 
30819 (8); PEREIRA, G. 4807 (29); PEREIRA, S. C. 03509 


230 


60 


Annals of the 
Missouri Botanical Garden 


(10), 1504 (1); PERSSON, O. 298 (35); PHILCOX, D. 4305 
(16); PICKEL, D. 4463 (1); PIEROTTI, S. s.n. (37a), 6516 
Gia), 6606 (37a); PIRANI, J. R. 420 (4), 1490 (10); PIRES, J. 
31 (10); FEX Ten 22 * (24) PLK & TRANG us 04; 
PO n. (3); P 1, C. B. 70 (6; P 
347 a 00 (37a), is ns 1153 (8), 1154 a ES 7 
1305 (8), 1328 (8), 1332 (29), 1346 (39), 1347 (8), 1378 (7), 
m 2 ne. (8); POTT, A. 2033 (37a); pu V. 463 (8); 
n. (8; PROCTOR, G. 46852 

eared c 295 (37a), 2292 (3), 3517 D QUINTANA, 
M. 151 (35). 

RAMBO, B. 2233 (39), 8690 (39), 8763 (15), 11408 (15), 
28228 (15), 29388 (3), 31722 (30), 34859 (15), 38661 (8), 
39809 (7), 40482 (8), 40866 (3). 41243 (39), 42585 (8), 
44454 (3), 44966 (39), 45393 (39), 45843 (8), 46408 (37a), 
46549 (39), 49624 (30), 49838 (37a), 50047 (15), 50048 
(39), 50139 (80), 51549 (30), 52052 (15), 52053 (8), 53473 
(30), 56427 (30), 56578 (30), 56661 (15), 56730 (8), 56790 
G0. Lux 60), 58564 (30), 59164 (8), 60307 (29); 
RA 6 (10); RATTER, J. A. 2835 (26), 2881 
ae cb o. ae A. F. 1849, 1-175 (16), 1870, I- 
175 (30); REINECK, C. 6 (17), 98 (17), 486 (3); REITZ, R. 
553 (30), 869 C (3), 1026 (30), 2072 (39), 2969 (27), 4562 


2276 (9); RIBEIRO, J. 
, 1511 (16); RIZZO, A. 22 
10), 5840 (30); TORE A. aed (37a); RODRÍGUEZ, M. 

9 (35), 488 (37a), 573 (37a); RODRIGUEZ, R. 679 (8), 
766 (30); ROJAS, T. 1760 (37h), 4885 (32), 9820 (3), 10102 
(25), 10082 (23), 11059 35), 12705 (37b), 12724 (37b), 
12735 (8), 13951 (3), 14582 (37b); ROMANCZUK, C. 197 
(8), 356 (8); ROMBOUTS, J. 3717 (35); ROSENGURTT, B. 
5800 (3), 5835 (8), alas Gray 1 ROTH, | L. 1678 (24); RUIZ 

LÓPEZ, H. & PAVÓN, J. A. s USBY, H. 2107 
(37a). 

SAINT HILAIRE, A. 706 (1), 2349 (1), 2653 (29), C2 
2460 bis (8), CAT. C2 1841 bis (8), 1472 (3), 1499 P (29), 
2653 (29); SAKANE, M. 199 (30), 202 (1); SALDÍAS, M. 196 
(2 4), ANNO (2 SANO, S. M 
3019 (6; SANTOS, I. F. 31 E 1566 

TOLEDO, C. 


wos 
co 


HININL A 


22915 (37A), 23091 (37a), 23198 (37a), 23212 (5), 23312 
(3), 23424 (8), 23467 a 23587 (30), 23664 (3), 23767 
(37a), 23798 (37a), 2 , 24296 a y b (30), 24715 
(37a), 25997 (35), lo a 26123 (35), 27227 ay b (17), 
27404 (35), 27618 (30), 27642 (35), 27782 (37a), 27931 
(37a), 28111 (5), 28174 (35), 28236 (35), 28256 a y b (37a), 
29353 (35), 30098 (8), 30120 (29), 33309 (5), 33398 (37a), 
33496 (11), 33581 (11), iu (23); SCHMIT, L. 097 8); 
SCHREITER, R. , 5424 (33); SCHULTZ, A. s 

(30), 3990 (15), 5425 CUN 6879 (8); SCHWACKE, C. s.n. 


(1), s.n. (37a), 2709 (29); SCHWARZ, G. J. 595 (35), 1353 
(29), 1462 (8), 1498 (8), 1588 (35), 1772 (35), 1795 (29), 
1884 (35), 2008 (35), 2113 (29), 2206 (35), 2464 (35), 2569 
(35), 2648 (35), 2749 (35), 2767 (35), 2811 (35), 2859 (35), 
3570 (8), 3926 (35), 3965 (29), 4273 (37a), 5052 (29), 5204 


(8), 5299 (3), 5477 (35), 5880 (35), 5924 (35), 8736 (3), 
103 (35). 10432 a), dun Tn L9. 11791 (25), 
12129 (25); SCHW 3 (35), 160 (35), 478 (29), 


1472 (35), 1909 dh do d d DU 2292 (8), 2387 
(3), ed ES, 4663 or ES NIK, R. 1048 
EHN 1489 (39), 2 o 8003 (35); SEIDEL, 
R. et 5 du (35), 190 (35), ne (35), 3084 (35); SELLOW, 
F. 217 (29), 1116 (15), 1517 (3), 1570 (8), 1733 (30), 1735 
(24), 4993 (3); SEMIR, J. 561 (24), 601 (24), 791 (24), 2647 
(24), 2788 (24), 3695 (24), 4060 (24), 17493 (14), 20483 
Co EPHERD, G. J. 7099 (10), 7167 (16), 12218 (1); 
A, J. M. et al. 1413 (4), 2727 (9); SILVEIRA, N. 3824 
o 5929 (7), 7314 (39), 7339 (39), 
nina o SIMÁO, 5 O (1); 

9 (37a); SMITH 7 (14), 10151 (27), 11668 (39), 
nee (30), 14292 ce du n 1565 (30); SOARES, 
Z. 233 (8), 265 (8); SOBRAL, M. s.n. (39), 1463 (8), 1657 
(7), E (8), 4830 (27), 4939 po RU. NEFFA, V. 272 
(8; SOLOMON, J. 6920 (35), 7721 (34), 10159 (37a), 
13511 yc 14019 (35), a an, 18018 (37a), 18546 
(35), 18639 (35); SORIA, N. 200 (37b), 920 (32), 1378 
(31b), 1700 (35), 3857 74) Pn (35), 7573 (35); SOUZA 

s.m. (10); SOUZA, V. C. 2339 (30), 4747 (28), 4431 
(28), Eun (Gita) a (28), Tum Do SPARRE, B. 1996 

0); 1083 (36) 
ia P s.n. (17), dd (3). 16557 (29), 18947 
8); STEHMANN, G. J. 1295 (27); STEINBACH, R. F. 332 
35), 766 (35), 2005 (35), 7034 (89) STRIEDER 33020 
39); STUBBLEBINE, W. H. 565 (1); STUTZ, H. 2144 (35); 
SUCRE, D. 399 (26), 547 (10), 6804 (1), 10451 (14), 10459 
0), 10523 (25). 

TADASHL, S. 133 (2), 275 (2); TAMASHIRO, J. 873 (1); 
TEIXEIRA, B. 102 XEIRA, W. 
TESMANN, G. s 
613 (37a), 663 (37a); TORRICO, C. 335 (33); TOVAR, O. 
1810 (24); TRESSENS, S. 1535 (29), 1621 (37a), 1684 (3), 
1726 (8), 1885 (30), 1886 (29), 1890 (10), 1891 (8), 2610 
(35), 2687 a y b (30), 2812 (29), 2892 (29), 4693 (37a), 4735 
(37a), 5000 (35), 5544 (37a), 5555 (35), 5558 (35), 5599 
( P gr W. 1137 (33); TÜRPE, A. M. 4892 (37a 
. 1037 G0; UNGARETTI, I. 446 (8), 1818 (39); 


= 


P 


BR. 
— 
E 
== 


35 (1); VANNI, R. 328 (35), 422 (8), 597 
(35), 669 (29), a (35), n (8), 1329 (35), 2673 (31a); 
VARGAS, L. 790 (24); VAZQUEZ AVILA, M. 381 (35); 
VENTURI, A. 133 (3); VIDAL, R. 276 (1), 587 (3); VIEGAS, 
A. 3812 (10 

WALTER, B. M. 057 (10), 124 (10), 157 (10); SING 
E. sn. (1), sn. (16); WEDDELL, M. s.n. (24), 1408 (24); 
WIDGREN, L 1845 (30); WILLIAMS, L. 5568 (24), 6640 
24); WOOD, J. R. 2 63); WOOLSTON, A. 599 (3), 1241 
3); WOYTKOWSKI, 6 (24). 

ZARDINI, E. 3867 ae 7281 (36), 8730 (36), 10216 (36), 
11695 (36), 12655 (36), 25517 (5), 25670 (5), 26027 (5), 
26104 (37b), 26269 (36), 26902 (37b), 27302 (37b), 27863 
31b), 29020 (36), 30036 (37h), 32302 (37b), 36464 (37b), 
36493 (36), 36531 (36), 41991 (30), 42813 (2), 43013 (2), 
45560 (5), 47874 (11), 47903 (11), 47938 (5); ZULOAGA, F. 
1516 (35). 


E 


= 


PHYLOGENETIC PLACEMENT OF 
THE TRIBE RETINIPHYLLEAE 
AMONG THE SUBFAMILY 
IXOROIDEAE (RUBIACEA E)! 


Rocio Cortés-B.,?* Piero G. Delprete,? and 
Timothy J. Moiley*? 


ABSTRACT 


The: tribe DADDA and ES single genius hert phyllum Bonpl. contains 22 species of shrubs and small trees that grow 
i n the eric 


The circumscription of the tribe is based on the diagnostic 


caract kic two collateral pe pendulous ovules per 
a has bee 


son, its placement within the family 


r locule, a rare condition in 
n PORE rsi. ah The moi 


the Rubiaceae. However, for the same 
and systematic position o of the tribe 
he results 


MAT 


Retiniphylleae a Retiniphyllum were D based on a y 


confirm oe pu ui of the tribe and genus Retiniphylh um. The tribe is nd sister to the core membe 
e genera Botryarrhena Ducke and Seyphiphora C. F. Gaertn 


Ixoroidea 
formexly pu e 


ey words: po S Guayana Region, Ixoroideae, 


phora, irnL-F. 


Retiniphylleae, 


rs of the subfamily 
. are not related to the tribe Retiniphylleae, as 


Retiniphyllum, rpsl6, Rubiaceae, Scyphi- 


e tribe Retiniphylleae includes only the genus 
Retiniphyllum Bonpl. This genus consists of 22 
species of shrubs and small trees that grow on white 
sand soils in the Neotropics. Most species are 
distributed in the Guayana Region, and a few reach 

e Amazon Basin, eastern Andes, and central and 
eastern Brazil. The Retiniphylleae consists of shrubs 
or trees characterized by the abundant resin located at 
apical buds. Each flower is subtended by a bracteole 
located at the base of the pedicel and an involucel 
dor ane m at the top of the pedicel. Flowers 
have with contorted aestivation, stamens 
reflexed | in prés anthers with basal and apical 
sterile appendages, a (4 to)5(to 6 to 8)-locular ovary 


with two collateral pendulous ovules per locule, 
drupaceous fruits, and pyrenes normally containing 
one seed due to the abortion of one ovule. In addition, 
many species exhibit secondary pollen presentation. 
Some of these characters are not common in the 
Rubiaceae, especially the condition of two ovules per 
locule. In this family, most members have one or many 
the tribe 
Retiniphylleae has been clearly defined and isolated 


ovules per locule. As a consequence, 
in the family, but for the same reason, its placement 
within the Rubiaceae has been controversial. 

In the earliest systems of classification (Kunth, 
1818; Roemer & Schultes, 1818; Jussieu, 1820; 
Richard, 1830; de Candolle, 1830; Bentham, 1841), 


1 We are grateful to Conciencias-Fulbright-LASPAU, Lehman College, and The New York Botanical Garden for supporting 
the first author during her doctoral ~ at the City University of New York; The Lewis B. and Dorothy Cullman Foundation 


for supporting the work in the labor: 
directors of the herbaria BM, BR, C. COAH 


making their specimens available; E. J. anda (Urecht University, The Netherlands), A. Vogel (Botanical Garden Leiden 


University, The Netherl ands), T. NER IDEE and E 
h leaves at the ü f thei 


ir institutions; and G. Aymard, P. Ber 


echt (National Botanical Garden, Meise, Belgium) for help in 


ry, R. Evans, an as Mori for sharing 


their field observations, specimens, and photographs. This research was undertaken has a fellowship for Visiting Scientist 
p 


to Piero Delprete from the Conselho 


, Leuven, Belgium. 


t for the 


? Herbario Forestal, Universidad Distrital, Campus El aa Avenida Circunvalar-Venado de Oro, Bogotá, Colombia. 


tanini phyllum E lei 


3 CNPq Vis 2 oe Institute of e Sciences (ICB-1), Department of General Biology/Botany, Universidade 


Federal de H oiás, Campus IL 74001-970 Goiânia 


dress: Institut de Recherche pour le 


Développemeni as et Bioinformatique de VArchitectue des Plantes (AMAP), TA-A51/PS2, Blvd. de la Lironde, 


5, France. 


Old Dom 


“Department of Biological Sciences, 


minion ee ersity, Norfolk, Virginia 23529-0266, U.S.A. tmotley@odu.edu. 


š Lewis B. and rd Cullman Program for Molecular Systematics Studies, The New York Botanical Garden, Bronx, New 


York 10458-5126, U.S.A 
doi: 10. 3417/2006198 


ANN. Missouni Bor. Garp. 96: 61-67. PUBLISHED ON 23 APRIL 2009. 


62 


Annals of the 


Missouri Botanical Garden 


Table 1. 


Taxa sampled for the phylogenetic analyses of the trnL-F spacer and the rps/6 intron data sets. 


GenBank accession number 


Taxa Voucher irnL-F rps16 
Aleisanthia rupestris Ridl. Tange 45171 (AAU) AF152660* — 
Alibertia edulis (Rich.) x ae in DC. Jansen-Jacobs 3840 (GB) AF201029! — 
Alibertia edulis (Rich. ich. in D Rova 2288 (GB) — AF200975! 
mphidasya Se C Steyerm, Ståhl et al. 3542 (GB) AF152624? — 
Amphidasya ambigua (Standl.) Standl taxon — AF129271" 
Borojoa patinoi Cuatrec Persson et al. 2194 (GB) AF201034! AF200984! 
Botryarrhena pendula Ducke Campos 29 (NY EU821638 — 
Calochone er (De Wild.) Keay hase 335 AF201036! AF200986' 
Calycophyllum spruceanum (Benth.) Hook. f. ex K. Schum. e 62777 (NY) AY555080* EU821613 
Ceriscoides vexsliflora (Wall. ex Kurz) Tirveng. well 87-967 (AAU) AF201039" AF200989" 
inchona pitayensis Wedd. Pici et al. 2109 (GB) AF152684 — 
Cinchona pubescens Vahl taxon 50278 — AF004035° 
Coffea liberica Hiern Delpreie 7357 (NY) AY555081? EU821614 
Condaminea corymbosa (Ruiz & Pav.) DC. Rova et al. 2084 (S) AF102406* — 
Condaminea corymbosa (Ruiz & Pav.) DC. taxon 60042 — AF004039° 
Duperrea ae Pit. Delprete 7373 (NY) AY555082? EU821615 
adogía audruana J. M. Fay, J.-P. LeBrun & Stork Fay 8901 (NY EU821639 EU821616 
Pi cu e Pohl sp. Alves 2267 (NY) AY555083* EU821617 
Feretia aeruginescens Stapf Mwanyambo 154 (N AY555084? EU821618 
Gardenia taitensis Struwe & Albert 1208 (NY) AF102426" — 
Gardenia volkensii ii subsp spatulifolia Stapf & Hutch. 1 (GB) — AF200996: 
nipa a Delprete 6522 (NY) AF152665? — 
nipa american ersson & Gustafsson 342 (GB) — AF200997: 
Hippotis brevipes Spruce ex K. Schum. Woytkowski 5620 (NY) AF152636 — 
Hippotis scarlatina Krause ta: 7 — F331650* 
Ibeiralia surinamensis Bremek Persson et al. 1930 (GB) AF 201048" AF201000' 
Ixora finlaysoniana Wall. ex G. Do Delprete 7344 (NY) AY555085" EU821619 
Keetia multiflora (Schum. & Thonn.) Bridson Delprete 7384 (NY) AY555086* — 
Kutchubaea Fisch. ex D Rodriguez 59 (NY) AY555087* — 
Kutchubaea C.s riguez 8. Y) — EU821620 
Leptactina leopoldi-secundi Biittner Delprete 7364 (NY) AY555083? EU821621 
imnosipanea spruceana Hook. f. Jansen- i et al. 2615 (NY) AY555102? — 
Limnosipanea erythraeoides (Cham.) K. Schum. Macedo 7 (US) — EU821622 
Macrosphyra longistyla (DC.) Hook. F. ex Hiern ies e (BM) AF201051! AF201004* 
M a pubescens Buch.-Ham Delprete 7399 (NY) AY555089? EU821623 
Pavetia stenosepala K. Schu Delprete 7387 (NY Y555090* EU821624 
Platycarpum acreanum G. K. Rogers Cid Ferreira 10407 (NY) AY555100? — 
Polysphaeria Hook. f. sp Groves (K) 152655! AF201011' 
Posoqueria gracilis (Rudge) Roem. & Schult. Munzinger 504 (NY) EU821640 — 
Pouchetia baumanniana Büttner Delprete 7359 (NY) AY555091 EU821625 
Pseudomussaenda flava Verdc Andrews 857 (S AF152652? 
Psilanthus mannii Delprete 7349 (NY) AY555092* — 
Psychotri 054 (NY) AY555079 EU821612 
Psydrax schimperiana (A. Rich.) Bridson Delprete 7388 (NY) EU821641 EU821626 
Pyrostria media (A. d ex DC.) Cavaco Zarucchi 7424 (NY) EU821642 EU821627 
Randia nitida (Kunth) D Delprete 7358 (NY) AY555093 EU821628 
Retiniphyllum concolor m ex Benth.) Müll. Arg. Berry 7093 (NY EU8216 — 
Retiniphyllum concolor a ex Benth.) Miill. Arg. Berry 7422 (NY) — EU821629 
Retiniphyllum maguirei Stand 0 (MO) EU821646 EU821632 
Retiniphyllum deen me Cortés 1648 (NY) EU821644 EU821630 
Retiniphyllum schomburgkii (Spruce ex a Miill. Arg. Berry 7567 (MO) EU821645 EU821631 
etiniphyllum sec cundifi orum Bonpl. Berry 7457 (MO) 8216 EU821633 
Rondeletia inermis (Spreng.) PE & Urb. Acevedo et al. 7691 (NY) AF152745* — 
Rondeletia portoricensis Krug & U Taylor 11678 (MO) AF243015* 


Volume 96, Number 1 


Cortés et al. 


2009 a Placement of Retiniphylleae 
Table 1. Continued. 
GenBank accession number 

Taxa Voucher trnL-F rps16 
Rosenbergiodendron densiflorum (K. Schum.) Fagexl. Jansen-Jacobs et al. 3977 (GB) — AF201061! — 
Rosenbergiodendron densiflorum (K. Schum.) Fagexl. E et al. 1994 (GB) AF201014" 
Rytigynia senegalensis Blume Ma — U821637 
Scyphiphora hydrophyllacea C. F. Gaertn. pun 43134 (NY) EU821648 EU821634. 
o stahelii Br Rova et al. 2068 (GB) — AF243023* 

ipanea wilson-brownei R. S. Cowan Mori 25 NY) EU821649 

Sipaneopsis is Hen (Spruce ex K. e Steyerm. Wurdack & Adde 43253 (NY) AF152678* — 
Stachyarrhena harleyi J. H. Kirkbr Tho (NY) 821650 — 
Stachyarn Jansen-Jacobs et al. 4707 (GB) — AF201021' 
Tarenna dumis Bridson Delprete 7406 (NY) AY555097? EU821635 
Tocoyena williamsii Standl. Stahl 3028 (GB) AF201071' — 
Tocoyena Aubl. sp. Jansen-Jacobs et al. 3976 (GB) — F201016' 
Vangueria madagascariensis J. F. Gmel. duc 7383 (NY) AY555098 EU821636 


nces were originally published in ‘Persson (2000b), *Delprete & Cortés-B. (2004), *Rova et al. (2002), “Rova 
(unpublished), "Struwe et al. (1998), Andersson & Rova (1999), and “Piesschaert et al. (2000). 


the bi-ovulated locules of Retiniphyllum were misin- 
terpreted, resulting in its association with tribes 
currently placed in the subfamily Rubioideae. Hooker 
(1873) established the tribe Retiniphylleae to include 
Retiniphyllum and, by an incorrect interpretation of 
the fruit, the genus Kutchubaea Fisch. ex DC. In the 
classic system of classification propose 

(1891), the tribe Retiniphyl 

and Retiniphyllum was placed in the tribe Gardenieae 
of the subfamily Cinchonoideae. Verdcourt (1958) 
jr gu the tribe Retiniphylleae, and maintained it 
kamp (1966), 
wh proposed one of the most important systems of 
classification for the Rubiaceae in the 20th century, 


d by Schumann 
was not recognized, 


n the subfamily Cinchonoideae. Breme 


criticized previous placements of Retiniphyllum and 
simply called it an genus. According to 
Bremekamp (1966), the absence of secondary pollen 
does a defining character of the tribe Garden- 

and his subfamily Ixoroideae, was absent 
Renton Robbrecht (1988, 1993) maintained 
Retiniphyllum in the tribe Ret d placed 
it in ois subfamily Antirheoideae, a subfamily that 


in 
tiniphylleae an 


was not supported by molecular data (Bremer & 
Jansen, 1991; Bremer & Struwe, 1992; Bremer et al., 
995; Bremer, 1996; Rova, 1999; Rova et al., 2002 

Andersson and Rova (1999), in their phylogenetic 


NA ie 


study that focused on the subfamily Rubioideae using 
rps16, sampled Retiniphyllum for the first time and 
placed it in the zl e. This placement 
was also supported by R 99) and Rova et al. 
(2002) in their study of the uis A 
tieae-Sipaneeae complex. In these analyses, Retini- 
of the 


Ixoroideae related to Paleotropical representatives, 


phyllum was located in an isolated clade 


sister to a clade with members of the tribes Coffeeae, 


Gardenieae, Octotropideae, Pavetteae, Rondeletieae, 
ieae. In the iaceae 
the traditional Cinchonoideae an 


Ixoroideae were merged in a single subfamily, and 


an angue 
classification, 
the tribe Retiniphylleae was placed in the supertribe 
Ixoridinae of the subfamily Cinchonoideae (Robbrecht 
& Manen, 2006). 

The = i uia Ducke and d el 
C. F. Gae ere tentatively included in the tri 
espe ps Robbrecht (1988). r 
comprises two species distributed in the Amazon 
Basin and Guayana Region. Ducke (1932) pointed out 
because they 
racemose DA and two (rarely three or 

e. On the other hand, th 
genus Scyphiphora alles only S. do doplellaies C. 
F. Gaertn. the only mangrove species in the 
Rubiaceae. It has unique placentation with two ovules 


four) ovules per loc Asian 


per locule, one of which is pendulous and the other 
erect. Robbrecht (1988) fant that this peculiar 
placentation is perhaps a form derived from the 
condition in Retiniphyllum. 
this study are to: (1) test the 
monophyly of the tribe Retiniphylleae, (2) evaluate 
the phylogenetic position of the Retiniphylleae within 
amily 


The main goals of 


xoroideae, an evaluate the 
era Botryarrhena and 


phora with Raih ia 


the su 
relationship of the 


METHODS 
TAXON SAMPLING 


A total of 49 taxa representing most tribes or groups 
recognized in the subfamily Ixoroideae s.l. were used 


64 


Annals of the 
Missouri Botanical Garden 


to test the monophyly of the Retiniphylleae: the 


Condamineeae complex clade (Rova et al., 2002), 
enriquezieae (Rogers, , Posoquerieae (Del- 
prete et al., , Sipaneeae (Delprete ortés-B., 


2004), Mussaendeae M & Tulin, 1998), Ixoreae 


(Andreasen & B 000), Vanguerieae (Lantz et 
al., 2002), Onde My n Coffeeae, 
Pavetteae (Andreasen & Bremer, 2000), the Alibertia 


group (Persson, 3000), and ne i (Rob- 
brecht, 1988) 

In the phylogeny of Retiniphyllum (Cortés-B. et al., 
in prep.), most of the species are resolved in three 
main clades. In the present study, five representative 
species of Retiniphyllum were selected, including at 
least one from each clade: R. concolor (Spruce ex 
Benth.) Müll. Arg., R. maguirei Standl., R. rhabdoca- 
lyx Müll. Arg., R. schomburgkii (Benth.) Müll. Arg., 
and R. secundiflorum Bonpl. 

Leaf samples were collected in the Botanical 


Gardens of Bruxelles (BR), Leiden (L), and Wagenin- 


(cpDNA) sequences of the trnL-F intergenic spacer 
and the rps/6 intron. Of the total number of sequences 
used in the analyses, 41% were original; the rest were 
downloaded from GenBank from Persson (2000b) 
[22%], Delprete and Cortés-B. (2004) [20%], and 
Rova et al. (2002) [9%], and the remaining 8% from 
98), Andersson 


Piesschaert et al. 
Table 1 


Rova (unpublished), Struwe et al. (19 
an ova (1999), a 
ucher loman is presented in 


OUTGROUP SELECTION 


our members of the subfamilies Rubioideae and 


Cinchonoideae were selected as outgroup. Psychotria 
L. and Amphidasya Standl. have been shown to be part 
of the subfamily Rubioideae (Bremer & Manen, 2000; 
Rova et al., 2002), while Cinchona L. and Rondeletia 


L. are members of the subfamily Cinchonoideae 


(Bremer & Thulin, 1998; Rova et al., 2002) 


DNA EXTRACTION, AMPLIFICATION, AND SEQUENCING 


Total genomic DNA was isolated from approxi- 
mately 1 cm? of dried leaf tissue desiccated in silica 
gel, or from herbarium material, using a modified 
CTAB quu. w et al., 2005). 

s using the 
reaction (PCR) fol i Motley et al. 
amplification of the irnL-F spacer, the primers “e” 
Gr GCTICAAGTCOCTCTATOGC 3) and “P (5'- 
ATTTGAACTGGTGACACGAG-3’) of Taberlet et al. 
( en 


) were used. The rpsió intron was amplified 


polymerase chain 
(2005) 


Q 
5 


using the primers (5'-GTGGTAGAAAG- 


rpsF 


CAACGTGCGACTT-3’) and rpsR2 (5'-TCGGGATC- 
GAACATCAATTGCAAC-3/) designed by d et 
al. (1997). The PCR conditions were: 4°C 
for 3 min., 32 cycles of 94°C for 45 sec., 52° for 
30 sec., 72°C for 1 min. 30 sec., and hold 74°C for 
7 min., hold 4°C. Cross-contamination was controlled 
y using negative controls in the PCR reactions. In 
addition, DNA from two individuals per species was 
extracted, amplified, and sequenced, when possible. 


PCR purification kit (Qiagen, 


Valencia, pu U.S.A.) following protocols 


provided by the manufacturer. Cycle sequencing 
n gene cleaning using hydrated Sephadex 
G-50 DN ade F columns (Amersham Pharmacia 
adr Inc., Piscataway, New Jersey, U.S.A.), and the 
visualization separation of fragments were run on an 
ABI Prism 377 DNA sequencer (Applied Biosystems, 
Foster City, California, U.S.A.) following the protocols 
described in Motley et al. (2005). 


PHYLOGENETIC ANALYSIS 


The sequences were first edited in Sequencher 
3.1.2 (Gene Codes Corporation, Ann Arbor, Michigan, 
U.S.A.) and preliminarily aligned with ClustalX 
Thompson et al, 1997) using the default settings. 
They were then manually edited using BioEdit (Hall, 

99) 


— 


m 
Ke} 


Parsimony analyses with equal character weights 
nd unordered characters were performed with NONA 
o 1993) in concert with WinClada (Nixon, 
e analyses, gaps were treated as missin 


ue E ewan searches were performed 
holding a maximum of 100,000 trees per search. In 
each search, replications were carrie 
keeping five trees per replication under the sehen 
mult*max*. The trees obtained were used to calculate 
a strict consensus tree. In Bret to ou the relative 


knife (JK) 


analyses were executed itin 1000 replicates. 


support of the clades, b ) and jack 


RESULTS 


The combined analysis of the trnL-F and rps16 data 
of 52 taxa and 1428 characters, 
The 
heuristic ek resulted in 54 most parsimonious 
trees of 651 steps in length, with a mdr index 
CD of 0.61 en excluding uninform charac- 
ters) and a retention index (RI) of en Figure 1 
shows the strict consensus tree obtained in the 


matrices had a total 
282 of which were parsimony informative. 


=> 


heuristic search with BS and JK support values. 
In the 


maximum parsimony analysis, a well-supported clade 


strict consensus tree resulting from the 


Volume 96, Number 1 Cortés et al. 
2009 uci Placement of Retiniphylleae 


po i. 
mphidasya 
Cinch ond Outgroup 


E Tribe 
R. maouirei Retiniphylleae 
R ser undiorum 
A 
quur 
Ix 


fangueria 


R figynia 

t 
Polysphaeria 
Psydrax 
Pyrostria 


Cet ^u ^u ^u ty 
Qo SE 


Gardenia 
Ceriscoides 
Genipa 


Randia 

Kos enberpiodendon 
Tocoyen 

Calo Lon 
Macrosphyrá 


Figure 1. Strict consensus tree of the 54 most parsimonious trees from the combined irn£-F and rps16 analysis. —A. 
Condamineeae complex clade (Rova et al., 2002). —B. Henriquezieae (Rogers, 1984). —C. Posoquerieae (Delprete et al., 
2004). —D. Sipaneeae (Delprete & Cortés- B. , 2004). —E. Mussaendeae (Bremer & Thulin, 1998). —F. S e (Andreasen & 


Bremer, 2000). —G. Vanguerieae (Lantz et al. , 2002). —H. Ro verd (Robbrecht, 1988). —I. Coffeeae. —J. Alibertia 
group (Persson, 2000a). —K. Pavetteae (Andreasen & Bremer, 2000). — L. Gardenieae s.l. (Robbrecht, 1988). Bootstrap (BS) 
and jackknife (JK) support values are indicated above and below branches, respectively. 


66 Annals of th 
Missouri Botanical Garden 
(BS = 89%, JK = 94%) was retrieved containing — Alibertia A. Rich. ex DC., and Borojoa Cuatrec. 
genera Condaminea DC., Ferdinandusa Pohl, Caly- (Fig. 1 
cophyllum DC., and Hippotis Ruiz & Pa hi idi had 


been placed in the subfamily odios in 
previous systems of classification (Verdcourt, 1958; 
Bremekamp, ; Robbrecht, 19 
clade was sister to the rest of the sampled genera. 
Members of the tribes Sipaneeae, Henriquezieae, and 
upported clade 
Ps to the members of the tribe Mussaendeae (BS — 
56%, JK = 67%) and the clade that includes 
Retiniphyllum species. All the species sample 


Posoquerieae are together in an unsupport 


g 


Retiniphyllum are together in a strongly "nA 
clade (BS = 100%, JK = 100%), sister to a clade 
that includes Ixoreae, Vanguerieae, Octotropideae, 
Coffeeae, Pavetteae, Alibertia group, and Gardenieae s.l. 


DISCUSSION 


THE MONOPHYLY AND POSITION OF THE TRIBE RETINIPHYLLEAE 
IN THE SUBFAMILY IXOROIDEAE 


Because species of Retiniphyllum a a strongly 
supported monophyletic lineage (BS = 100%, JK = 

100%) in the independent (not shown) combined 
chloroplast analyses (Fig. 1), the monophyly of the 
genus Retiniphyllum is confirmed. Similarly, the 
isolated position of Retiniphyllum in the cladogram 
also confirms the monophyly of the tribe Retiniphyl- 


eae. 

The tribe Retiniphylleae is placed as a clade within 
the subfamily Ixoroideae s.l., sister to the tribes that 
correspond for the most part to the Ixoroideae sensu 
Bremekamp (Bremekamp, 1966). Although the sys- 
tematics of the subfamily Ixoroideae have been largely 
his tribes 


modified since Bremekamp's proposal, 
i Vanguerieae correspond to 


Gardenieae, Ixoreae, and 
the core of the 
tion. The results presented here also support those 
previously reported by Rova (1999) and Rova et al. 
(2002) 


xoroideae in its original circumscrip- 


Robbrecht (1988) included the tribes Retiniphyl- 
leae and Vanguerieae in the subfamily Antirheoideae, 
Ea that morphological similarities of their 

nd s supported this relationship. In 
Sdn Robbrecht and Manen (2006) Mu 
that these similarities were consistent with their 
placement in their supertribe Ixoridinae. According 
to our results, the tribe Retiniphylleae is not closely 
related to the Vanguerieae. 


THE PLACEMENT OF BOTRYARRHENA 


Botryarrhena was resolved within the Alibertia 
group, sister to Stachyarrhena Hook. f. in a clade 


with the genera /betralia Bremek., Kutchubaea, 


Ducke (1932) suggested an affinity between the 
genera Botryarrhena and Stachyarrhena. They share 
racemose inflorescences, but differ because Stachyar- 
rhena has unisexual flowers and a dioecious breeding 
system o . Stachyarrhena is now placed 
within the in the tribe 
O (Persson 2000a, b). Subsequently, uni- 


a group, a clade wit 


sexual flowers have been observed in Botryarrhena 
(Persson & Delprete, pers. 
support for the placement of Botryarrhena in this 


comm.), providing further 


lineage. 

The relationship between Botryarrhena and Retini- 
phyllum pc by Ducke (1932) was based on 
flower uality, inflorescence morphology, and 
number of pari. per eue However, it is important 
to note that Standley was not able to see the berry-like 
fruits of Botryarrhena, a common fruit type of many 
genera in the Ixoroideae but not in Retiniphyllum. 


THE PLACEMENT OF SCYPHIPHORA 


According to our results, the genera Retiniphyllum 
and Scyphiphora are resolved in two distinct clades, 
indicating that the bi- ovulate condition has evolved 
independently in the su 

uff and Rohrhofer 198) si the d - 
ora in detail and found no char: 
suggesting a close ono to puo) pu 


of Scyphiphor 


tentatively placed Seyphiphora in the subtribe Di- 


plosporinae, tribe Gardenieae s. ased on the 
presence of tracheidal idicblaste - in Seyphiphora, 
which are similar to the mesophyll sclereids in the 
Gardenieae. Andreasen and Bremer (2000), using 
Seyphi- 
and they 


morphological and molecular data, placed 
phora as sister to the tribe Ixoreae, 
tentatively included it in the Ixoreae. 
Our results indicate that Scyphiphora is sister to a 
clade that also 
Vanguerieae (Fig. 1). This indicates that Seyphiphora 


includes the tribe Ixoreae and 
is neither a member of the tribe Gardenieae s.l. as 
Puff and Rohrhofer (1993) suggested, nor a member of 
the tribe Ixoreae as Andreasen and Bremer (2000) 
hypothesized. 


Literature Cited 


Andersson, L. & J. H. E. Rova. 1999. The rps16 p “a 
the phylogeny of = Rukioideae (Rubiaceae). PI. 


n, K. & B. a emer. 2000. Combined rl in v 
ubiaceae-Ixoroideae: Morphology, nuclear and chlor 
48. 


Bentham, G. 184 L Gostibudams töwards a Flora of South 
America—Enumeration of plants collected by Mr. Schom- 
burgk in British Guiana. J. Bot. (Hooker) 3: 212—250. 


Volume 96, Number 1 
2009 


Cortés et al. 
E Placement of Retiniphylleae 


Bremekamp, C. E. B. 1966. Remarks on the position, the 
WARE, um the subdivision of the Rubiaceae. Act 
eil. 15: 1-33. 


n E 1996. Phylogenetic studies within Rubiaceae and 
relationships to pu families based on molecular data. 
Opera Bot lg. 7: 33-50. 

——— & R. K us. 1991. Comparative restriction site 

mapping of the chloroplast DNA implies new crop 

relationships within the Rubiaceae. Amer. J. 18: 

98-213. 


——— € L. Struwe. 1992. d of the acid and 
the Loganiaceae: Con, conflict bet morpho- 
beu and a. du pum J. Bot. 70: no 1184. 

M. Thulin. d Collapse a Bav re- 
me n of Mussaendeae, and us of 
Sabiceeae (Rubiaceae: aa Pa os based 
on rbcL a E d 211: 71- 

—& Ph p and classification of 

Pl. Sys 


t. Evol. 225: 


rs apu 
-12 


n & D. Olsson. 1995. Subfamilial e 


Naturalis Regni Vegetabilis, Vol. 4: 341 "362. endi & 
Wiirtz, Paris. 

Delprete, P. . Cortés-B. 2004. A preliminary 
phylogenetic study of the tribe Sipaneeae (Rubiaceae, 
B pos irnL-F and ITS sequence data. Taxon 

47 


vate & R. B. Klein. 2004. Rubiáceas, 1. 
Alseis até us Galium. Pp. 1-344 in A. 2l. vee n 
ustrada Catarinense, Vol. 1: Gén e A- ro 


. Neue poa aus der Hyalea Prasiliens: 
Notizbl. Bot. Gard. 11: 471—483. 
Goloboff, P. A. 1993. NONA version 2.0. 


Program and 


documentation distributed by the author, Tucumán, 
Argentina. 

Hall, T. A. 1999. BioEdit A user-friendly Misa 
seque alignment or and analysis program 
Windows 95/98/NT. Nucl. Acid p. Ser. 

Hooker, J. D. 1873. Ordo LXXXIV. Rubiaceae. Pp. in 


G. Bentham & J. D. Hooker edd Genera Plantarum, 
Vol. 2. Lovell Reeve & Co., London. 

Jussieu, A. L. de. 1820. Sur la famille des plantes Rubiacées. 
Mém. Mus. Hist. Nat. 6: 365—410. 

Kunth, C. S. 1818. Retiniphyllum. P. n F. H. A. 
Humboldt & A. J. de ÉS (editors), Nee genera et 
species plantarum,.... Librariae graeco-latini-germani- 
cae, Paris. 

Lantz, H., K. Andreasen & B. Bremer. 2002. Nuclear rDNA 
ITS sequence data used to construct the first phylogeny of 
uc ae (Rubiaceae). Pl. Syst. Evol. 230: 173-187. 

Motley, T. J., K. J. Wurdack & P. G. Dee. 2005. 
Molecular systematics f the Chi ae—Catesba 

r and fruit evolution aud 

. Amer. J. Bot. 92: 316—329. 

002. i Vers. 1.00.08. Published by 

author, o New Y 

Didius B., M. n& p. glad. 1997. Chloroplast 
pi 6 intron ‘cn of the tribe ‘ae (Caryophylla 

eae). Pl. Syst. Evol. 206: 393-4 


Persson, C. 2000a. Phylogeny of the Neotropical Alibertia 
group Ta E emphasis on the genus Alibertia, 
inferred fro E ^n osomal DNA sequences. 
Amet a Bot, gm PUES 
— 00b. TO = dea ee uit 
on Ki neo DNA s m the and 
irnL(UA A)-F(GA A) iue ae i dx 


257-269. 

Piesschaert, F., dersson, S. Jansen, S. Dessein, E. 
Robbrecht & E. Smets, 2000. a for the taxonomie 
position of the African pipes Ene aie rcd 
Morphology and anatomy compared to rps on 


Robbrecht, = 1988. Tropical Woody Rubiaceae: Opera Bot. 
Belg. 


——— E. = com Supplement to the 1988 outline of the 
— G the Rubiaceae. Index to genera. Pp. 
. Robbrecht (editor), Advances in Rubia- 

ceae d us oy Belg. Vol. 6. 


position of Coptosapelta an i nd supertree 
onstruction based at 
rbe. ata. A new classification in two subfamilies, 
oe and Rubioideae. Syst. Geogr. Pl. 76 
85-144 


Roemer, f J. & J. A. Schultes. 1818. Systema vegetabilium 
3: 255 


mineea. 

Dissertation, 
University of Góteborg, Góteborg, Sweden. 

— ———, P. G. Delprete, L. Andersson & V. A. Albert. 2002. 
A imL-F cpDNA o study of the Condamineeae— 
Rondeleti: mplex with implications on 
n phylogeny E die. ao Amer. J. Bot. 89: 

—159. 


B K. 1891. Rubiaceae. Pp. 1-156 in A. Engler & K. 
Prantl natürlichen Pflanzenfamilien, 
4(4) 


is related to a ae in lineage of Gentanascae, 
arvard Pap. Bot. 3: 199-21 
Taberlet, P., L. Gielly, G. mm Bouvet. 1991. Universal 
primers for amplification of three non-coding regions of 
i 1109. 


2 e classification of the 
Rubiaceae. Bull. Jard. Bot. Etat Bruxelles 28: 209— 


A GLOBAL ASSESSMENT OF 
DISTRIBUTION, DIVERSITY, 
ENDEMISM, AND TAXONOMIC 
EFFORT IN THE RUBIACEAE?’ 


Aaron P. Davis,? Rafaél Govaerts,? 
Diane M. Bridson,? Markus Ruhsam,? 
Justin Moat, and Neil A. Brummitt? 


ABSTRACT 


Analyses of distribution, diversity, endemism, and taxonomic effort for Rubiaceae are o reported, based on queries from a 


World Rubiaceae ivan 


are 
ceae species s an 


n (13.143 s 


ee areas of hig 
required. 
es have res oe distribu 
estricted distri 


showing that many spec 
with a range of other pd rs includin; 
at which 


tion, 


databases, Taxonomie Databases Working Group (TDWG), 


oup in the p with greatest diversity in low- t 


mism is generally high i in Rubiaceae, 


ude h 
p/611 genera), which confirms that this is the fou rth lar, rs 


and percenta b 
one "ur n e areas where dus f eld. collecting 
which supports data from recent Duc 


tions. Given the aspumed ecologic rod of Rubiacea mbination 
ibu n this fam 


ac 
ily are pari cs M ps er. rable 


new v species are oe described is inadegüate: more resources are required before the 


conservation, endemics, endemism, Rubiaceae, species diversity, taxonomic 
taxonomy. 


Target One of the Global Strategy for Plant 
Conservation (Secretariat of the Convention on Biolog- 
ical Diversity, 2002) is the production of “a widely 
accessible working list of known plant species, as a step 

world flora,” which is a funda 
tal requirement for plant conservation (Nic Lughadha, 


towards a complet men- 
2004). For some of the largest flowering plant families 
and for larger groups (i.e., monocotyledons), several 
important works have been completed that significantly 
improve our prospects for achieving Target One. For 
example, information on Euphorbiaceae (Govaerts et 
al., 2000; <http://www.kew.org/wesp/malpigiales/>) is 
available in print and on the Internet, and for 
monocotyledons, information is accessible only via 
the Internet (e.g., Govaerts, 2006; <http://www.kew. 
org/wesp/monocots>), as part of the World Checklist 
Series (<http://www. kow org/wesp>). In the p 


only category, a heckli 


2006; 


<http://www.kew.org/wesp/rubiaceae>). is worl 


ar become available (Govaerts et al., 


aw ¢ 


* 
i aes in the series, represents an amalgamation 


and synthesis of taxonomic work. 


These checklists are valuable as widely accessible 
working lists of accepted plant species, but they also 
enable broad- analysis of distribution and 
d to be undertaken (e.g., ovaerts, 
ese checklist are pain e complete 
a "M include distributions for each accepted 
species, they provide an interesting counterpoint to 
more detailed but less complete compilations on plant 

pecies diversity that have been ee published 
Dti et al., 1996, 1999; Kier et al., 2005; Mutke 
& Barthlott, 2005). Thee a have produced 
impressively precise maps of global plant species 
diversity, generally modeled from available species 
lists for different parts of the world (Barthlott et al., 

» ; Kier et al, , but often without 
having complete species distributions to underpin the 
estimates of diversity. Herein, we use queries from the 
World Checklist of Rubiaceae (Govaerts et al., 2006; 
<http://www.kew.org/wesp/rubiaceae>) to ae the 
distribution, diversity, endemism, nomic 


effort for Rubia 


analysis of the whole family and follows recent work 


and t 
eae. This represents the first such 


of biam Sou 
parer 


ae 


re grateful to the curators of L and P for making specimens available for study. The Global Biodiversity Information 


T) program, support ted part of the work 
ion concerning the Rubiaceae 


? University of Edinburgh. Tous n Evolutionary Biology, Ashworth Laboratories, West Main Road, Edinburgh, EH9 3JT, 


United Kingdom 
doi: 10. 3411/2006205 


ANN. Missouni Bor. Garp. 96: 68-78. PUBLISHED ON 23 APRIL 2009. 


Volume 96, Number 1 
2009 


Davis et al. 69 
A Global Assessment of Rubiaceae 


on a global analysis of plant genus distributions 
(Brummitt, 2005). 

Rubiaceae is a member of the Gentianales and 
shares many of the features common to other families 
f the order, particularly basic leaf and flora 
morphology (Davis & Bridson, 2007; and see below), 
the presence of colleters, and lack of internal phloem. 
It is the fourth largest flowering plant family and is 
estimated to contain around 600 genera and bet 


6000 a 


easy to identify by the presence of simple, opposite or 


13,000 species (see below). It is Wels 


whorled, entire leaves, interpetiolar stipules, and an 
inferior ovary. Rubiaceae has a cosmopolitan distri- 
diversity and biomass are 
and the 


subtropies and especially in lowland humid forest, 


bution, but species 


distinctly concentrated in the tropics 
where it is often the most species-abundant woody 
plant family. The family is less frequent and less 
diverse but still very widespread in the temperate 
regions. It is also found in the subpolar regions of the 
Arctic and Antarctic (Davis & Bridson, 2007). In the 
tropical regions, Rubiaceae species are sensitive to 
disturbance and are be found in secondary forest 
types (A. Davis 


species are small trees or shrubs, but nearly all life- 
forms are found, including large trees, annual and 
perennial herbaceous plants, woody monocaul dwarfs, 
lianas, epiphytes, geofrutices (more or less herbaceous 
stems with a woody rootstock), myrmecophiles (hollow 
stems or special chambered tubers, containing ants or 
ant colonies), and rarely succulent or aquatic life- 
forms (Robbrecht, 1988; Davis & Bridson, 2007). 
L.), which is by far 


the most important economic plant within the family 


Rubiaceae includes coffee (Coffea 


and the world's most important commodity after oil 


(Vega et al., 2003). 


METHODS 
DATABASE 


The production of the World Checklist of Rubia- 
ceae was made from a database encompassing 24 
fields, including basic nomenclatural data (genus, 
species, author, place and data of publication, 
basionym [if applicable], synonyms [if applicable], 
and accepted name), distribution data, and life-form 
The World Checklist of Rubiaceae (Govaerts et al, 
2006) does not include altitude data. Altitude does 
have a bearing on occurrence and diversity; long- 
e knowledge of the family (D. Bridson & A. 

is, pers. obs.) indicates that Rubiaceae 2 d 
diem. is higher at low to mid-altitudes, with most o 


the diversity occurring at altitudes of less pow 


1500 m. The data comply with the data standards 
y the International Organization for Plant 

Information a n" e 1994), in association with 
he ases Working Group (TDWG) 
E. et A em) Citation of authors follows 
Brummitt and Powell (1992); book S are Doe 
ated according to Stafleu and Cowan 88) and 
Stafleu and Mennega (1992-2000); aL are 
abbreviated according to Bridson and Smith (1991); 
and the number and three-letter codes used for areas 
e.g., 23 CON) follow the TDWG syste 
of the database was undertaken using 
(Microsoft, Redmond, Washington, U.S.A.), a dBASE- 
class database program for personal computers. The 
atabase was founded on the Index Kewensis 

database, held at the Royal Botanic Gardens, Kew 
K). The selection of accepted names and the 


a 


m. Compilation 
FoxBASE 


= 


designation of synonyms were made on the basis of 
published or otherwise publicly available taxonomic 
works. Further taxonomic input and accuracy were 
achieved by: (1) specialist taxonomic review; (2) a 
complete herbarium survey of the Rubiaceae collec- 
tions held at K; and (3) a survey of selected parts of 
the collections housed at L and P (abbreviations from 
Holmgren et al., 1990). Data collection for procedures 
(2) and (3) mainly used herbarium specimens cited in 
axonomic revisions or identified by specialists; these 
procedures added a further 2500 geographic records 
to the World Checklist of Rubiaceae database at 
TDWG Level 3. 


STANDARDIZING RUBIACEAE DIVERSITY FOR DIFFERENT- 
SIZED AREAS 


Counts of taxa for both species and genera for all 
areas at TDWG Level 3 (369 bud were e 


from n v» ES of Rubiaceae (Govaerts et 
al., 2006). T WG World amai Bene 
for Pi Pus Distributions is based on 


geopolitical units, which vary widely in size from 
the Antarctic Continent to tiny oceanic islands. In 
order to make counts of species and genera compa- 
rable between these units, the counts were rescaled to 
However, the 
ded by 


the size of that region to give a value comparable with 


make them independent of area. 
diversity of a region cannot be simply divi 


other differently sized regions because the relation- 
ship between diversity and area is a nonlinear, power- 
law relationship (Rosenzweig, 1995). Dividing by area 
overinflates the diversity of small regions and 
underestimates the diversity of large regions (Brum- 
mitt & Nic Lughadha, 2003). Instead, the power-law 
relationship S = cA* (where S = number of species, A 
= area, and c and z represent, respectively, the 


intercept and the slope of the regression in a log-log 


70 


Annals of the 
Missouri Botanical Garden 


space) can be rewritten as c = S/A to give a value for 
pe region that is independent of area (Rosenzweig, 

This value is then standardized to a size 
unen for the range of areas being studied, again 
using the exponent value z. 

An important consideration is the exact value of the 
exponent used to rescale diversity figures. Although 
small changes in z values do not give very different 
results (results not shown here), they can nevertheless 
influence the relative positions of regions close together 
in size or diversity (i.e., more or less diverse regions 
might move up or down the list of most diverse areas 
relative to other regions). The z value is known to vary 
between different regions, being lower for large, 


continental region 


(Rosenzweig, 1995). For this study, z values appropriate 
to each region in question could be estimated from the 
previous study by Kier et al. (2005), ee 
determined z values for each of the 1 s of the 
World Wildlife Fund (WWF) ecoregions (Olson et al., 
2001) from smaller-scale studies within each biome. A 
spatial overlay was used between the TDWG Level 3 
areas and the WWF ecoregions, and the mean z value for 

WCG regions was caleulated with a weighted average 
by area of the intersection between each TDWG level 
and the WWF ecoregions. 

In this study, the intercept values of relative 
diversity were standardized to a size of 10,000 km’, 
i TDWG Level 3 areas, 
similar to and facilitating comparison with the work by 
Barthlott et al. (1996, 1999) (Brummitt et al, 
unpublished). The values resulting from the rescaling 


roughly the median size of 


of species numbers in this way (S/10000) do not 
taxa, but they 
relative diversities to be compared for different 


reflect actual numbers o do allow 
regions that are independent of the size of that region. 
In this contribution, only the first 20 records for each 
database query/analysis are given. Complete results 
for all 369 TDWG i 


authors upon request. 


Level 3 areas are available from the 


RESULTS AND DISCUSSION 
THE NUMBER OF GENERA AND SPECIES IN RUBIACEAE 


Recent estimates as to the number of Rubiaceae 
species and -o are quite constant, apart from the 
estimates pecies numbers by Verdcourt (1976, 
1989) and en (1988). Estimates are as follows: 
Verdcourt (1976, 1989), 500 genera and 6000 spp.; 
Smith (1988), 500 genera and 6500-7000 spp.; 
Mabberley (1987), 630 genera and 10,400 spp.; Mab- 
berley (1997), 650 genera and 10,200 spp.; Robbrecht 
(19 637 genera and 10,700 spp.; 

(1992), 606 genera. 


and Brummitt 


According to the World Checklist of Rubiaceae 
database, the number of accepted Rubiaceae species is 
13,143 i 


have been added to Rubiaceae over the past 30 years or 


in 611 genera. Given the rate at which species 


so (see below), and assuming that more synonyms have 
been created than retrieved from synonymy, all 
previous estimates for species number are much lower 
than the actual figures would have been at that time, 
apart from estimates by Bridson and Verdcourt (2003: 
650 genera and 13,000 spp.) and Davis and Bridson 

(2007: 615 genera and 13,150 spp.), which were based 
on earlier versions of the data used here. This is 
particularly so for estimates in the 6000-7000 range 
(which are roughly half of the actual figures for species 
number presented here). In general, the spe 
diversity for Rubiaceae has been considerably (Verd- 
court, 1976, 1989; Smith. , 1988) to moderately (e.g.. 


cies 


ties in estimating species numbers in 
large families, particularly where the total number of 
published names is often considerable. The World 
Checklist of Rubiaceae database holds a total of 36,385 
published names, for example (Govaerts et al., 2006). 
Estimates as to the number of genera have been quite 
accurate, mostly because the number of names involved 
is much lower and presumably also due to the presence 
of generic indices and similar resources that exist in 
herbaria and libraries. 
At 13,143 species, Rubiaceae is the fourth largest 
e ur family (Robbrecht, s after Orchida- 
ae (25,158 spp., ca. 830 genera; Cribb & Govaerts, 
2005) Asteraceae (23,000-30,000 Spp.. 1535-1700 
1994; Funk et al, 2005), 
Leguminosae (19 genera; Lewis et al., 
2005); Poaceae is the fifth largest (ca. 11,591 spp., ca. 
Govaerts, 2006). The numbers for 


Asteraceae and 


genera; Bremer, 


700 genera; 
Leguminosae have been estimated 
and are not based on definitive counts of accepted 
species, although Leguminosae has been carefully 
calculated (Lewis et al., 2005). Parenthetically, of the 


five largest families, Orchidaceae, 


Leguminosae, 
woody and herbaceous taxa, cies oe 
has a greater proport 
Lughadha et al. ( 


most representative family for angiosperm diversity 


o eous species. Nic 
2005)- Ru rm a is the 


patterns. Rubiaceae is particularly well represented in 
umid 
Asterac 
diversity, is shown to comprise one of the most 


ropical forests and, coupled wit 
l 


n 
eae in a global analysis of angiosperm 
representative pairs of families at species level (N. 
Brummitt, unpublishe 

s with most other fuscis plant families, the 
number of accepted Rubiaceae species is still 


Volume 96, Number 1 Davis et al. 71 
2009 A Global Assessment of Rubiaceae 

140 m - = - = = - = = = — 
g [| m [| [| 

2 8 "E E N | L— — B5.— — — — —- 

3 — 
3 

S 60 

E 

3 

40 

20 i Ho HA H H H 


Figure 1. 


increasing, year by year, and the final number will be 
greater than it is today. The total number of Rubiaceae 
species is estimated at ca. 16,000 (A. Davis, M. 
Ruhsam & D. Zappi, unpublished}, based on a review 
of herbarium collections (K, P, and L) and an 
awareness of undescribed species (from databases, 
literature, fieldwork, and anecdotal evidence). The 
l of 16,000 


take into account potential synonymy and may be 


species pe to make a tota 


broken down in the following manner: unplaced 
names brought into accepted usage (500); undescribed 

species from tropical South America (500), tropical 
Africa (400), Madagascar (300; Davis & Bridson, 
2003), southern Asia (300), Malesia (500), Australia 
and the Pacific (100), and other regions (100). There 
are ca. 1000 unplaced names on the World Checklist 
of Rubiaceae (Govaerts et al., 2006), and we estimate 
that nomenclatural and taxonomic work will see at 
least 500 species added to the current species count 
from this source alone (Ruhsam et al., 2008). 

From 1976-2005, 1842 Rubiaceae species names 
have been validly published (Fig. 1), which gives an 
average of 63.5 species per year for this period. The 
minimum number of species published since 1976 
was 17 (in 1976) and the maximum 117 (in 1988). The 
average for this period does not give us the figure for 
the increase in accepted names, as the number of 
synonyms created per year is not included in the 
above calculations. In fact, it is very difficult to 
quantify the number of names placed in synonymy per 
year, as it is not recorded. If we extrapolate by looking 
at the total number of validly published names 
produced since the starting date for formal biological 
s (1753), over a 252-year period from 

53-2005, the average is 52 species per year, and 
this gives us some idea of increase of accepted names 


Number of new Rubiaceae species published each year from 1976-2005. 


per year. The number of new (validly published) 
genera added over the same 30-year perio 

making an average of 3.5 genera per year. The 
maximum number of genera published during this 
period was 12 (1978) and the minimum was 0 (1976, 
1977, 1991, 1992, 2002). If we extrapolate in the 
same way as we did for species, in order to get some 
idea of gross increase in new genera per year, the 
average is 2.4 since 1753. As with the species 
caleulation, the figures for the 30- and 252-year 
period are not that different. Based on the same 
evidence as for species (herbarium data, anecdotal 
evidence, fieldwork, etc.), and taking into account our 
assumption that estimates for genera are more 
accurate than those for species, the number of genera 
is not likely to increase significantly. 

Taxonomic effort at the species level for Rubiaceae, 
as estimated above, is probably much lower than that 
of most other large plant families (i.e., those with more 
than 10,000 species), although Orchidaceae is the 
only other large plant family for which accurate 
figures are available for comparative purposes. In 
Orchidaceae, from 1978-2002, the average number of 
newly described orchid species was over 280 per year, 
with the number surpassing 500 (per year) twice 
during this period (Cribb & Govaerts, 2005). Even 
considering the special interest given to orchids, and 
the fact that Orchidaceae has approximately twice the 
number of species of Rubiaceae, taxonomic effort for 
Rubiaceae is considerably lower. Based on our 
estimate that there are actually around 16,000 species 
of Rubiaceae (i.e., 2800 species still requiring 
scientific names), it will take around 45 years before 
the species diversity in the family is satisfactorily 
enumerate e continue to describe species at the 
current rate n above). 


72 


Annals of the 
Missouri Botanical Garden 


SIZE OF GENERA 


Table 1. 
concerning delimitation (Taylor, 1996, 2001; Nepo- 
kroeff et al, 1999; Davis et al., 2001; Andersson, 
2002), Psychotria L. is still the largest genus in 
Rubiaceae, with 4 species. Recent publications 
(Sohmer & Davis, 2007; Davis et al., 2007; Ruhsam et 
al, 2008) will bring the total number of Psychotria 
species close to 2000, as estimated by Sohmer (1988) 
and Davis et al. (2001). P. E 


sychotria is now the wor 


actions 


The 20 largest genera of the Rubiaceae are listed in 
Despite recent discussions and 


third largest genus, after Astragalus L. (Leguminosae) 
with ca. 3200 species and Bulbophyllum Thouars 
(Orchidaceae) with ca. 2000 species (Frodin, 2004). 
There are 30 Rubiaceae genera with over 100 
species, but most contain fewer than 10 species. There 
are 211 monotypic genera (34.5% of genera; 1.6% o 
species), 328 genera with three species or fewer 
(53.796 of genera), and 440 genera with 10 species or 
fewer (72% of genera). Although all large taxonomic 
groups have a greater number of small taxa (Clayton, 
1972, 1974; Cronk, 1989), the percentage of mono- 
typic genera in Rubiaceae is higher than that in both 
Orchidaceae, with 211 monotypic genera out of 849 
(Cribb & Govaerts, 2005) (24.9% of genera; 0.8% of 
species) or Leguminosae, with 192 monotypic genera 
2005) (26.4% of 


genera; 1% of species). Similar analyses of other large 


out of 727 genera (Lewis et al., 


angiosperm families are needed to understand wheth- 
er such a large number of monotypic genera in 
Rubiaceae is unusual, part of a natural phenomenon, 
or an artifact of our classification systems (Knapp et 
al., 2005). However, the fact 
skewed frequency distributions a 


that such d 
t just 

by taxon size but also by spatial poss pos 
narrowly distributed taxa and few very widely 
distributed taxa) (Colwell & Lees, 2000; Gaston, 
and also temporal taxon distribution. (many 
short-lived i 
(Rosenzweig, 1995) suggests that this is a natural 


taxa and few very long-lived taxa) 
phenomenon. 

In light of ongoing Rubiaceae research (De Block et 
al., 2006), it is evident that even over the next five 
years or so the size of many of the large genera will 
change quite considerably (in particular Ixora L., 
Spermacoce L., Oldenlandia L., Tarenna Gaertn., and 
Canthium Lam.), as their circumscriptions are altered 
in the light of new systematic data. Some genera will 
increase in size, owing to the necessary inclusion of 
other genera, most notably fxora, whereas others will 
decrease in size, such as Canthium (Lantz & Bremer, 
2004; Razafimandimbison et al., unpublished). Of the 
largest avetta L. 


O Rubiaceae genera, only 


(Bremekamp, 1934) has been monographed, and for 


Table 1. 


in Rubiaceae. 


The 20 largest (by species number) genera 


No. Genus No. of species 
1 Psychotria L. 1834 
2 Galium L 621 
3 Ixora L. 530 
4 Pavetia L 357 
5 Ophiorrhiza L 317 
6 Palicourea Aubl 313 
7 deletia 260 
8 Spermacoce L 257 
9 Oldenlandia L 249 

10 Lasianthus Jack 228 

11 Faramea Aub 208 

12 Tarenna Gaertn 203 

13 Mussaenda L 200 

14 Asperula L 182 

15 Timonius DC. 169 

16 rgosiemma Wall 162 

17 Guettarda L 159 

18 Gardenia Ellis 143 

19 Coussarea Aubl 133 

20 Canthium Lam. 130 


the largest 50 genera, there are only a few with 
complete taxonomic treatments, e.g., Sabicea Aubl. 
Wernham, , Manettia Mutis ex L. (Wernham, 
1918-1919), Coffea (Chevalier, 1947), and Leptoder- 
mis Wall. (Winkler, 1922), although contemporary 
monographs are now needed for these four genera. 


DISTRIBUTION OF RÜBIACEAE 


Rubiaceae occur in every region of the world (at 
TDWG Level 3), except for the Antarctic Continent, 
which only has two native vascular plant species 


os 


Deschampsia antarctica E. Desv. and Colobanthus 
quitensis (Kunth) Bartl.). Rubiaceae is a predominant- 
ly tropical family, with species diversity decreasing 


in subarctic regions; the entire Subarctic America 
region (TDWG 70) has only eight Galium L. species, 
for example. There are, however, specific areas in the 
tropical belt that do not have high numbers of species 
or high species diversity for Rubiaceae (see below). 


DISTRIBUTION OF SPECIES DIVERSITY 


Table 2 gives the 20 most species-rich regions for 
Rubiaceae based on gross number of indigenous 
species for each TDWG Level 3 area. This makes a 
useful comparison between TDWG Level 3 areas but 


Volume 96, Number 1 
2009 


Davis et al. 73 
A Global Assessment of Rubiaceae 


Table 2. The 20 most diverse regions (TDWG Level 3) for Rubiaceae, based on total species numbers and irrespective 


of area. 

Rank TDWG Level 3 code Area (narrative) No. of species Area (km?) 
1 CLM om 1026 1,140,598 
2 VEN Venezuela 185 914,096 
3 NWG ew Guinea 125 819,979 
4 BZN Brazil North 645 3,849,262 
5 ZAI Democratic Republic of Congo 644. 2,336,991 
6 BZL Brazil Southeast 619 926,896 
7 PER eru 594 1,296,128 
8 ECU Ecuador 583 249,014 
9 BOR Borneo 578 743,470 

10 MDG Madagascar 569 594,765 
11 TAN Tanzania 559 945,437 
12 CMN eroo 553 466,814 
13 PHI Philippines 535 295,856 
14 MLY Malaya 485 132,735 
15 CUB Cuba 438 110,269 
16 THA Thailand 400 514,630 
17 PAN Panama 391 14,845 
18 GAB abon 353 261,859 
19 CHC South-Central China 342 1,309,801 
20 SUM Sumatera 342 473,039 


does not give us a realistic idea of species richness 
because of the considerable differences in unit area 
(square kilometers). In simple terms, TDWG Level 3 
areas in the tropics with large unit areas will tend to 
hold higher numbers of Rubiaceae species than 
smaller ones, given that other factors (such as forest 
type and altitude) are comparable. It is expected that 
areas with a large percentage of low- to mid-altitude 
humid forest (e.g., Colombia [83 CLM], Venezuela [82 
VEN] New Guinea [43 NWG] will have large 
numbers of Rubiaceae species per unit areas, for 
example. Table 3 shows the 20 most diverse areas for 
Rubiaceae based on relative species richness (species 
number/area log-transformed [8/1000]; Brummitt & 
Nic Lughadha, 2003; see Methods), at TDWG Level 3 
(Fig. 2). In Table 3, Venezuela (82 VEN) and 
Colombia (83 CLM) are in comparable positions with 
those of Table 2 (gross species number), and many 
other areas remain in , but the order of areas 
changes considerably between tables. In Table 3, 
Brazil North (84 BZN), South-Central China - 
CHC), and Sumatera (4. are not among t 

20 most species-rich areas (cf. Table 2), but iue 
Costa Rica (80 COS), Gulf of Guinea islands (23 GGI), 
and New Caledonia (60 NWC) are present. The Gulf of 
Guinea islands are equatorial continental islands with 
appreciable amounts of primary lowland forest 
(Figueiredo, 2005; Davis & Figueiredo, 2007). All 
major tropical regions (South America, Africa, Indian 
Ocean, South Asia, Southeast Asia, and the Pacific) 
are represented in the 20 most species-rich areas, 


with no obvious bias to any one of these regions. Of 
the 20 most species-rich areas, 13 are continental, 
five are large islands, one is a large island archipelago 
(Philippines [42 PHI]), and one is a small island group 
(Gulf of Guinea islands [23 GGI]. 
Figure 2 shows the areas within the tropical regions 
is aes 
given the paucity of pre habitat for 
Rubiaceae (ie. low- to medium-altitude, humid 
forest), or eee Rwanda (23 RWA) and perhaps 
23 BUR), where high altitude excludes many 
species present in surrounding countries, and central 
and eastern Brazil (84 BZC, BZE). which is largely 
omposed of savanna vegetation (cerrado). Figure 2 
vel 3 areas where 


where low relative species richness 
ferred macro 


Burundi 


also clearly shows severa 
significant relative species richness is expected 
(proximity to the equator and a prevalence of low- 
altitude, humid forest) but is not prese 
analyses. These areas include Equatorial Guinea (23 
EQG), the Democratic Republic of Congo (23 CON), 
Cambodia (41 CBD), Laos (41 LAO), Sulawesi (42 


, and Suriname (8 ). For these areas, we 


nt in our 


= 
N 


assume that low relative species richness is due to low 
specimen-collecting density per unit area (A. Davis 
and D. Bridson, pers. obs.) and low levels of taxo- 
nomic effort including determination of specimens to 
species, although these activities are closely associ- 
ated. We assume that the relatively low collecting 
densities for 

N), pee (41 CB 


due to previous military Bee and resulting limited 


74 Annals of the 
Missouri Botanical Garden 


Table 3. "Twenty most diverse regions for Rubiaceae based on relative diversity (species number/area log-transformed at 


TDWG Level 3). 


Rank* TDWG Level 3 code Area (narrative) No. of species Area (km?) Mean ofz ce = SW $/10,000 


1 (2) VEN Venezuela 785 914,096 0.2331 32.0157 274.0075 
2 ( CLM Colombia 1026 1,140,598 0.2862 18.9494 264.4819 
3 (15) CUB Cuba 438 110,269 0.2116 37.5397 263.5670 
4 (10) MDG Madagascar 569 594,765 0.2029 38.3264 248.3691 
5 (14 MLY 485 132,735 0.2593 22.1694 248.0571 
6 (12) CMN Cameroon 553 466,814 0.2108 35.2915 245.9629 
7 (3 NWG ew Guinea T25 819,979 0.2486 24.5565 242.4190 
8 (11) TAN Tanzania 559 945,437 0.1840 44.4524 242.0446 
9 (8) ECU Ecuador 583 249,014 0.2931 15.2769 227.2141 
10 (13) PHI Philippines 535 295,856 0.2583 20.6621 223.0354 
11 (17) PAN 0.3058 12.6374 211.2797 


Panama 391 74,845 A 
2 (5) ZAI Democratic 644. 2,336,991 0.2093 29.9175 205.6487 


ong 

13 (6) BZL Brazil Southeast 619 926,896 0.2513 19.5964 198.3242 
14 (9) BOR Borneo 518 143,410 0.2526 19.0041 194.6465 
15 (25) COS Costa Rica 300 51,273 0.3070 10.7445 181.6291 
16 (274) GGI Gulf of Guinea 133 3,208 0.2400 19.1582 174.7251 

islands 

17 (41) NWC New Caledonia 203 19,283 0.2483 17.5181 172.4595 
18 (18) GAB Gabon 353 261,859 0.2254 21.2101 169.0995 
19 (7) PER eru 594 1,296,128 0.2675 13.7611 161.6788 
20 16) THA Thailand 400 514,630 0.2324 18.8239 160.0695 


A = area, c and z = intercept and slope, respectively, of the regression in a log-log space, 5 = number of species. 
For rank, numbers in parentheses represent rank based on gross species number per TDWG Level 3 area (Table 2). Areas 
listed in Table 2 but not appearing in Table 3, with ranking based on relative diversity in parentheses: Brazil North (44), 
South-Central China (29), and Sumatera (27). 


Legend 

TDWG Level 3 

Species richness per area 
Low 


High 


BEERS 


Figure 2. Relative species richness of Rubiaceae at TDWG Level 3 regions rescaled by the size of that region using a 
power-law species area relationship and standardized to 10,000 km’. 


Volume 96, Number 1 
2009 


Davis et al. 
A Global Assessment of Rubiaceae 


Table 4. The 20 highest areas for gross number of endemic Rubiaceae species. 


Rank TDWG Level 3 code Area (narrative) No. of species Endemic species, No. (%) 
1 NWG ew Gui 125 620 (86 
2 MDG adagasca 569 520 (91 
3 PHI Philippines 535 443 (83 
4 BOR Borneo 578 428 (74 
5 CUB Cuba 438 344 (76 
6 BZL Brazil Southeast 619 311 (50 
T CLM Colombia 1026 265 (26 
8 VEN Venezuela 785 252 (32 
9 MLY alaya 485 213 (44 
10 NWC New Caledonia 203 200 (9 
11 TAN Tanzania 559 190 (34 
12 THA Thailand 400 179 (45 
13 IND ndia 326 169 (52 
14 BZN Brazil Northeast 702 165 (24 
15 SUM Sumatera 342 161 (47 
16 CHC South-Central China 342 149 (44 
17 PER Peru 594 147 (25 
18 PAN Panama 391 136 (35 
19 FIJ iji 166 134 (81 
20 VIE Vietnam 443 129 (2 


access, although priorities set in the colonial era may 
have also played a role. 


ENDEMISM 


We provide two crude measures for ds 
endemism in Rubiaceae: total number of endemics 
(Table 4) and percentage of endemism (Table 5) for 
each TDWG Level 3 area. A few areas of known high 
endemism cannot be shown by analyses of our data 
= they are split between different TDWG Level 
where mountain 


e 


areas. This is particularly m 
ranges |Guineide with and Bandara: (e.g., Rwen- 
zori Mountains, split between the Democratic Repub- 
lic of ad [23 ZAI], oak [23 bug and Uganda 
[25 UGA]. For the gross number of endemic species 
at e Level 3, eight of the 20 highest areas are 
large islands or island groups, with the first five falling 
into the island category. The other 12 are continental 
areas. In terms of percentage of endemism, the first 27 
TDWG Level 3 areas are islands, both small and 
Table 5 shows the highest 20 areas for 
percentage of endemicity. High numbers of endemics 


large; 


and percentage of endemics are expected for islands 
owing to the specific evolutionary scenarios associated 
with island floras, and, in the case of Rubiaceae, 
recent and rapid radiations following dispersal 
(Malcomber, 2002; Maurin et al., 2007) h 
particularly rad Continental areas with a high 
percentage endem 44, e.g, Brazil 
Southeast M BZL], India [40 END “Thailand [41 

THA], South-Central China [36 CHC], Malaya [42 


ave been 


MLY]; Table 4) require further explanation on a case- 
nding to t 
present-day physiography, UN and part 
from the smaller islands, which bs a 100% 
endemicity based on very few species, New Caledonia 
(60 MWC), Hawaii (63 HAW), and Madagascar (29 
MDC) of percentage 
endemicity (Table 5). Low percentage endemicity is 
continental 


y-case basis correspon heir historical and 
ogy- 


are outstanding in terms 


iase 


ard areas withi 
including areas with relatively high number of species 


regions, 


but negligible levels of percentage of endemicity, ; such 


Malawi (26 MLW), with 213 s ae 
Uganda (25 UGA), with 212 micity; 
Central African Republic (23 CAP), s 242 spp./2% 
endemicity; Ivory Coast (22 IVO), with 311 spp./39 
enini. and Nigeria (22 NGA), with 360 spp./4% 
endemicity. 

Species endemism is generally high in Rubiaceae. 
Of the 13,143 species of Rubiaceae, there are 
endemies at TDWG Level 3, which means that 64% of 


Rubiaceae species are endemics at this area level. 


vaerts, 2005] but is much greater than other big 


families (e.g., Poaceae) that not have specie 
diversity concentrated in the tropical regions of the 
world (Govaerts et al., 2006). This may be partly due 
to the evolutionary history and dynamics of tropical 
forests but also because dispersal and diversification 


in Rubiaceae at the species level seem to have 


76 


Annals of the 
Missouri Botanical Garden 


Table 5. The 20 highest areas for gross percentage of endemic Rubiaceae species. 


No. of 
Total no. of | nonendemic Endemic Endemism, 
Rank TDWG Level 3 code Area (narrative) species species species 
1 ASC Ascension 1 0 E 100 
2 STH St. Helena 1 0 1 100 
3 NFK Norfolk Islands 9 0 9 100 
4 KER Kermadec Isla 2 0 2 100 
5 MXI Mexican Pacific islands 2 0 2 100 
6 CPI Central America Pacific 1 0 1 100 
islands 
T NWC New Caledonia 203 3 200 99 
8 HAW Hawaii 47 2 45 96 
9 MDG Madagascar 569 49 520 91 
10 NWG Guinea 125 105 620 86 
11 PHI Philippines 535 92 443 83 
12 FIJ Fiji 166 32 134 81 
13 CUB Cuba 438 94. 344 76 
14 MRQ Marquesas Islands 17 4 13 16 
15 BOR Borneo 578 150 428 74 
16 MAU ritius 54 16 38 70 
17 SCI Society Islands 46 14 32 70 
18 SOC Socotra 21 di 14 67 
19 ROD Rodrigues 9 3 6 67 
20 JNF Juan Fernández Islands 6 2 4 67 


occurred very recently in many groups (e.g., Mal- 
comber, 2002; Maurin et al., 2007). At the present 
time, we simply do not have enough data to make 
supportable assumptions regarding the causes of 
rapid diversification in Rubiaceae. Considerable 
levels of endemism occur on both large and small 
islands and also in continental areas. In studies 
where area of oc nee and extent of occurrence 
(IUCN, 2001) s je caleulated for Rubiaceae, 
it appears that many species are highly localized 
and an alarming number are restricted to area 
polygons (extent of occurrence) of less than 100 km? 
in Coffea [Davis et al, 20006]. 
the ee of 
extinction, and for groups where extinction threat 
has been calculated neh 2001), the number of 
Threatened taxa is very high, e.g., ca. 70% in Coffea 
(Davis et al., 2006) and 74% in Philippine a 
(including nearly 10% extinction; Sohmer & D 

2007 


e.g., ca. 
Restricted distributions increase 


CONCLUSION 


With 13,183 species in 611 genera, the importance 
of Rubiaceae in terms of species number is supported 
by our study, and its position as the fourth largest 
angiosperm family is confirmed (Robbrecht, 1988) 
after Orchidaceae, Asteraceae, 


an eguminosae. 


Based on estimates of total species number in 


Rubiaceae (i.e., 16,000), we estimate that with current 
resources it will take us 45 years to fully enumerate 
species diversity in Rubiaceae. This calculation is 
oversimplified, as it does not take into account other 
variables such as names added to or removed from 
synonymy, and extinction (we have no way of knowing 
how many species will become extinct before they are 
discovered), but it does give us some idea of what 
needs to be done and an indication of where to focus 
taxonomic resources. 

Our assessment of Rubiaceae species diversity for 
each of the 369 areas of TDWG Level 3 using a 
measure of relative species diversity (Table 3, Fig. 2) 
areas of relative species diversity family 
analyses confirm that species richness in Rubiaceae 
is greatest in the tropical regions, particularly in 
continental areas and larger islands (Table 3, Fig. 2). 
Practical applications of our species-level diversity 
analysis include the identification of areas that 
require further field collections and/or taxonomic 
study, and the targeting of areas for efficient sample 
ling). 
precise measure 


Future analyses 

s of diversity 

need finer division of area and measurement of 
Rubiaceae habitat, 
remaining primary vegetation. In addition, reanalysis 
be 


o 
knowledge of the family improves and progresses. 


collection (e.g, DNA sam 


requiring more will 


suitable particularly areas of 


Rubiaceae data woul required as our 


Volume 96, Number 1 
2009 


Davis et al. 
A Global Assessment of Rubiaceae 


Basic analyses of endemism show that species 


endemism in Rubiaceae is considerable, with 64% of 


species endemic at the level of T Level 3, and 
that percentage endemism is distinctly higher for 
islands, large and small. 

iven the ecologic sensitivity of Rubiaceae (e.g., in 
the tropical regions mostly requiring primary forest), 
coupled with the restricted distribution of species, it is 
evident that many species are vulnerable to extinc- 
tion, particul 
change 


level. 


arly in an era of global environmental 


and huge anthropogenic ence at the local 


Literature Cited 
Andersson, L. 2002. Relationships and generic circumscrip- 


tions in the Psychotria complex (Rubiaceae, Psychotrieae). 


Sy 

Barthlott, W., N. Biedinger, G. Brau F. Feig, G. Kier & J. 
Mutke. 1999. Terminological and Terie as 
of the mapping and 20 of global biodiversity. Acta 


Bot. Fenn. 162: 103-1 

— —— , V. Lauer & A. Made 1996. Global p of 
species diversity in vascular plants: Towards a world map 
of phytodiversity. Erdkunde 50: 317-327 

CON LAM C. E. B. 1934. A monograph of the genus 
Pavetia L. Repert. Spec. Nov. Regni Veg. 37: 1-208. 

Bremer, K. 19 Non Cladistics & Classification. 
Ti 


Howe m 2003. Rubiaceae. Pp. 379— 
120 in G. v. Pope (editor), el Zambesiaca, Vol. 5, Part 
3. Royal Puy Gardens, 

Bridson, G. D. R. & E. R. 


Period 


cim B-P-H/S: Botanic 
icum A Sad lementum. Hunt Institute m 
Botanical Documentation, Carnegie Mellon University, 
Pittsburgh. 
Brummitt, N. A. 2005. 


Patterns in the global distribution of 
flowering . Friis & 


plant genera. Pp. 539—564 in I 

nd toas Plant Diversity and Complexity pu 

cal, Regional, a Skr. 55. 
E 


Brummitt, R. K. 1992. Vascular Plant Families and Genera; 
Royal eae Gardens, Kew. 
—& owell. 1992. Authors of Plant Names. Royal 
Botanic ecu Kew. 
assistance from " he o. E. o 
Brummitt, et al). 2001. tabas 
Standards No. 2, ed. 2. vm Cou raphi al pres e for 
Recording Plant Distributions, ed. 2. Hunt Institute for 
Botanical Documentation, E. “Mellon University, 


——— (with 


, N. A. 


proje 4i 
Chevalier, A. 1947. Les LI du globe E pp in 
des caféiers et faux- x maladie; 
nuisibles. Encycl. a 29: 352. 
E W. D. 1972. Some ms of the genus concept. 
Kew Bull. 27: ar 81. 
. The dcc distribution of Angiosperm 
families. Kew Bull. 29: -279 
Colwell, R. K. & D. C. eR 2000. The mid-domain effect: 
Geometric constraints on the geography of species 
richness. TREE 15: 70-76. 


Cribb, P. & R. Govaerts. 2005. Just how many orchids are 


Contereneb: Orchidées/Naturalia Publications, Turriers, 
rance. 
Cronk, Q. C 89. Measurement of biological and 
historical influences in plant classifications. Taxon 38 
70. 


D. M. Bridson. 2003. Introduction to E 
Rubiaceae: Pp. 431-434 in S. M. Goodman 
Benstead (editors), The Natural History of nae 
University of Chicago Press, Chicago. 
—— . 2007. Rubiaceae. Pp. 284—286 in V. H. 
Heywood, R. K. Brummitt, Culham & O. 
(editors), Flowering Plants of the World. Royal Botanic 
Garden: Pu 
RE Figueiredo. 2007. A checklist of the Rubiaceae 
e family) of Bioko and Annobon (Equat 
Gulf of Guinea). Syst. Biodivers. 5: 159-186. 
.M ridson, C. Jarvis ZR a 2001. The 
ychotria L. 


torial Guinea, 


me tion and 
(Ru biaceae). Bot. J. Linn. Soc. 135: 
— — —, R. Govaerts, D. M. Bridson & P. a 2006. An 
otated taxonomic conspectus of the genus Coffea 
(Rubiaceae) Bot. J. Linn. Soc. 152: 465-512. 

—, & M. Briggs. 2007. Indian Ocean Mapouria 
species transferred x Psychotria (Rubiaceae—Psycho- 
trieae). E 52: 245-262. 

De Block, P., S. Hen & E. Robbrecht. 2006. Third 
ede P Rubiaceae eae Programme & Ab- 
stracts. ee Bot. Belg. 40: 1-92. 


Figueiredo, E he a of São Tomé e Kap 
(Gulf of Guinea) ‘Taxonomy and conservation. Bot. J. Linn. 
Soc. 149: 85-11 


Frodin, D. 2004. DM and concepts of big plant genera. 
Taxon rà Hee 152. 

Funk, V R. J. Bayer, S. Keeley, R. Chan, L. Watson, B. 
Ceci E. Schilling, I. : o B. G. Baldwin, N. 
Garcia-Jac A. R. Jansen. 2005. 
beside w Antarctica: is a supertree to under- 
stand the diversity and distribution of the Compositae. 
Biol. Skr. 55: jm 314 

Gaston, K. 003. The Structure and Dynamics of 
Geographic Rs Oxford University "Press, Oxford. 

Grae R. 2006. World Checklist of up M The 

i rss E the Pula nic Gardens, Kew. 
cessed 6 Decem- 


ber 2006. 
n & A. Radcliffe-Smith. 2000. World 


D. 
ET E E phy of Euphorbiaceae (and 


Pandaceae). R Botanic Gardens, Kew 
, M. Ruh: m, ndersson, E. Robbrecht D. M. 
Bridson, A. P. Davis, L Schanzer & B. Sonké. 2006. World 


oard of Trustees of the 
Royal Botanic — Kew. E iia kew.org/wesp/ 
laceae>, acce; e 15 December 200 

T. Holmere n & L. ^ Barnett. 1990. 
York Botanical Garden, Bronx. 


" 1. Prepared by the IUCN Species Survival Commission 
UCN, Gland, ae and Cambridge, Unite 

a dom. 

Kier, G., J. Mutke, E. Dinerstein, T. H. Ricketts, W. Küper, 


Kreft & W. Barthlott. 2005. Global patterns of plant 
o and floristice knowledge Biogeogr. 32: 
1107-1116. 


78 


Annals of the 
Missouri Botanical Garden 


Knapp, S., E. Nic Lughadha & A. J. Paton. 2005. Taxonomic 

co species concepts and global species lists. TREE 
0: 7-8. 

Bremer. 


m 2004. Phylogeny inferred from 


morphology and DNA data ` Characterizing well- pra 


groups in V ). Bot. J. Linn. Soc. 1 
251— 
Lewis, G., , B. Mackinder & M. d 2005. 


B 
Legumes of dis World. Royal Botanic Gardens, K 
Mabberley, D. J. 1987. The Plant-Book: A Port able is: 
tionary > the Higher Plants. Cambridge University Press, 


The Plant-Book: A Portable Dictionary of the 
Higher Pars, 2nd ed. Cambridge University Press, 
id 


T. 2002. prea of Gaerinera Lam. 

(Rubiaceae) based on multiple DNA markers: Evidence 

of rapid radiation in a desd morphologically diverse 
genus. Evolution 56: 

Maurin, O., A E A Chester, E. F. 


Y. 
Jaufeerally-Fakim & M. F. Fay. 2007. D 


Mutke, J. & W. Barthlott. 2005. Patterns of vascular plant 
diversity at continental to global scales. Biol. Skr. 55: 
521—538. 

Nepokroeff M., B. Bremer & . Sytsma. 1999. 
Reorganization of the genus Psychoiria and tribe Psycho- 
trieae ES Mos from ITS and rbcL sequence 

ata. Syst. B -21. 


Nic Lughadha, E Towards a working list of all known 
plant br Philos. Trans., Ser. B 359: 681—687. 

, J. Baillie, W. Barthlott, N. A. Brummitt, M. R 

Cheek, A. Farjon, R. Govaerts, K. A. Hardwick, C. Hilton- 


of plant diversity: Towards 
nd op 


rtunities for m action. Philos 
372. 


K. R. Kassem: 2001. Terrestrial 
ecoregions of the world: A new map of life on earth 
Bioscience 51: 933-938 


Ex E. 1988. Tropical woody Rubiaceae. Opera Bot. 
Belg. 1: 1— 
eiiis m M. L. 1995. Species Diversity in Space and 
Time. Cambridge y Press, Cambridge 
Davis. 2008. No mencla- 
World Rubiaceae 


Secretariat o onvention on 


f the C Biological Diversity. 2002. 
Global Strategy for Plan 


t Conservation fs etariat of the 


Flora Mun Nova, Vol. 4. Pacific 
nolulu. 

8. The eic species of the genus 
io (Rubiaceae) 4 in New PvE and the Bismarck 
Ao Bishop Mus. Bull. Bot. 1: 1-339. 

& A. Davis. e genus Psychotria 
spec in de Philippine Archipelago. Sida, Bot. 


Gan 5 . Cowan. 1976-1988. Taxonomic 

Literatur dus Selective p da o Botanical I ARE 

and Collection ns with D deine ntaries, and T: ex 

Vols. ed. 2. Regnum Veg. Vols. 94, 98, ins, 110, 12, 
115- es Scheltema and Holkema, Bohn, Utre 

& E. Mennega. 1 O. Taxonomic 
ture: A Selective Guide to Botanical Publications and 
Collections with D a Commentaries, and Typ up- 


num Veg. Vols. 12: 


Neotropical. Genus Noto- 
pleura (Rubiaceae: du with the description of 
some new species. Ann. Missouri Bot. Gard. 
E. Rosenquist & "W. Collins. 2003. Global 
project needed to tackle coffee crisis. Nature 425: 343. 
Verdcourt, B. 1976. Flora of Tropical East Afric 
(Part 1). Pp. 1414 in R. H. Polhill (editor), Flora of 
Tropical East Africa. Rubiaceae, Part 1. Whitefriars Press 
Lid., m 


Rubiaceae. Pp. 1-210 in E. Launert e 
Flora ee Vol. 5, Part 1. Royal Botanic Gardens 


e H. F. 1914. A monograph of the genus Sabicea. 
Trustees of E British Museum, London 
19 9. The genus Manettia. I Bot. 57(suppl.): 


um H. 1922. Monographische Übersicht der P 
Leptodermis. Repert. Spec. Nov. Regni Veg. 18: 1 


TAXONOMIC HISTORY, 
MORPHOLOGY, AND 
REPRODUCTIVE BIOLOGY OF 
THE TRIBE POSOQUERIEAE 
(RUBIACEAE, IXOROIDEAE)' 


Piero G. Delprete? 


ABSTRACT 


tribe Posoquerieae was recently described to include the genera Posoqueria Aubl. and Molopanthera Turez 


. based on 


e ventral stamen that springs forward 
criptions were published, Posoqueria and 


Molopanthera have been positioned in several distantly related tribes within the Rubiaceae. The. close ei between 


the two gener: 
Molo, 


as only ps reveal 


era ed by molecular 
wal cu ia eneral m 


words: Gardenieae, Molopanthera, Neotropics, pollen catapult mechanism, Posoque 


phylogen The tax 
morphological comparison (parti icular ly of stamen 


etic studies xonomie history o 
mor y 


a, Posoquerieae, Rubiaceae 


genus Posoqueria Aubl. was established by 
mo (1775) based on his material from French 
Guiana and on P. longiflora Aubl. He explained that 
the generic name was derived from the name used by 
because the 
fish Aymara eats the fruits of this plant. However, the 


the Galibi tribe, “Aymara-Posoqueri,” 


typical laterally bent floral buds of this genus were not 
depicted in the 

e (1830) Suid Posoqueria in the 
Cardeniene (as “Gardeniaceae”) and, more 
specifically, in the subtribe Gardeniinae (as *Garden- 
de Candolle, 1830: 368), among genera now 


"e E in several other tribes. In Posoqueria, de 


leae”; 


Candolle recognized seven species, namely P. long- 
iflora, P. latifolia Roem. € Schult., P. decora DC., P. 
trinitatis DC., P. havanensis DC., P. gracilis Roem. & 


published Molopanthera 


Turez., describing M. paniculata Turez., and treated 


it as a genus with uncertain tribal position, differing 
from all the Rubiaceae genera with a multi-ovulate 
ovary. He derived the generic name from the Greek 
uóAoWV- (molops- = bruise or weal) and -xvOnpa 
(anthera = anther) meaning bruised anthers, probably 
in allusion to the dark ends of the anthers, which are 
the points of fusion of the anthers in two pairs (the 
other anther remaining solita 

Karsten (1 a adas ‘he pd Stannia H. 

arst., based S. formo which he 
distinguished bn ib dud "esi on 
stamen length (all equal in dea ; three out of 
rved in Stan 
probably unaware of Karsten's Sranda, a 
the genus as Martha F. J. Müll, which he distin- 
guished from Posoqueria because of the unequal 
stamens. In this work, he 
great detail, the catapult mechanism for throwing the 


five longer and cu 


was the first to describe, in 


pollen onto the flower visitors, observed on plants 


1 This research was realized dur 
Científico e Tecnológico (CNPq) at the Institute of Bio! 
Goiás, 


Heritage Private Reserve (commonly known as the “Ca 


ing a Visiting Sd fellowship from the Conse 
gical Sciences e the Universidade Federal de Goiás (UFG), Goiánia, 
Brazil. Observations of di individuals Es Mologan nihera w 


elho Nacional de Desenvolviment 


hi 
International Rubiaceae Congress was provided by the Fund for Scientific OY Flanders (FWO N. WO.005.05) and the 


dere of Plan 


Federal EM Goiás, Campus IL 74001- 


t Systematics of the Katholieke Universiteit, Leuven, Belgiu 
2 CNP eens gs tist, Institute of e Sciences (IC 
970 Goiánia, Goiás, 


B-1), Doo of General Biology/Botany, Universidade 


dress: Institut de Recherche pour le 


Développement, Pisa us et S n de VArehitecture des Plantes (AMAP), TA-A51/PS2, Blvd. de la Lironde, 


34398 Montpellier Cedex 5, France. 


L 


doi: 10.3417/2006192 


ANN. Missouni Bor. Garp. 96: 79-89. PUBLISHED ON 23 APRIL 2009. 


80 


Annals of the 
Missouri Botanical Garden 


growing in his private garden at Desterro, Santa Catarina, 
southern Brazil. The peculiar mechanism was | 
commented on by Charles Darwin (1876, 1877), ea on 
the notes and material sent to him by Müller. 

Hooker (1873: 8-9) positioned Posoqueria in the 
Gardenieae, which he distinguished from the Cates- 
baeeae by having (translated from Latin) “corolla 

contorted, seeds few to many, large, 
"corolla valvate, 


narrowly 
compressed, or small, angled" (vs. 
seeds many, large to vey large, compressed” i in the 
informal groups flower sexuality, inflores- 
cence position, style morphology, number of locules, 
and ovules biseriate or multiseriate. Hooker posi- 
tioned Posoqueria in the group with terminal inflo- 
rescence and hermaphroditie flowers, rarely poly- 
gamo-dioecious. He further distinguished Posoqueria 
from the other genera of this group with its corymbose 
inflorescence, flowers with elongated corolla tube, 
corolla lobes gibbous in bud, five exserted anthers, 
and a bifid stigma. Under this genus, he synonymized 
Solena Willd., icd Schreb., Kyrtanthus J. F 
Gmel., and Stannia. However, Hooker did not include 
Martha i in the ist. of synonyms, probably unaware of 


eric 
de cribed the flower bud. as laterally ene ie the 
filaments as erect or curved, he did not mention the 
peculiar pollen catapult mechanism. At the same 
time, Hooker placed Molopanthera in the Cinchoneae 
in the group with imbricate corolla lobes (one or two 
exterior) and stamens inserted at the base of the 
corolla. In addition, he described a second species, M. 
burchellii Hook. f., which he distinguished from the 
typical species by having pubescent vegetative parts. 

Baillon (1880) included the Gardenieae into his 
broadly circumscribed Genipa Series or Genipeae, 
where he positioned Posoqueria (Baillon: 435; includ- 

a anthus, Kooma Raf. 


treated it as closely related to Oxyanthus DC. and 
Kutchubaea Fisch. ex DC 


maintained Molopanthera in the Cinchoneae, follow- 


. In the same work, he 


ing the position and definition suggested by Hooker 
(1873), and positioned it near Calycophyllum DC. 
because of the 4- to 5-merous flowers, 2-locular 
capsules, and seeds with unequally dentate wing. 
The generic eae eae - jos and 
Stannia were repea d 
(1849, 1856, 1860, pd a Pos (1850, a 
Hooker (1873) and Baillon (1880), resulting in a 
complex diatribe of heated arguments published in a 
series of publications. Karsten put forward that 
Stannia has unequal stamens (with three stamens 


curved and longer) and leathery or lignified berries, 
while Posoqueria has equal stamens and juicy or 
fleshy berries. The other authors considered these 
characters trivial and preferred to synonymize the two 


— 


1888: 351-360; see annotated translation by Del- 
2005a: 50 


most authors in synonymizing the two genera and 


prete et al., —58). Schumann agreed with 
stressed that in both taxa the stamens are unequal. In 
addition, he used the morphology of the anthers as the 
unifying character and stated that NUN from 
German; Delprete et al., 2005a: 55): “The 

have a wide, dorsal area that is slightly aci from 


nthers 


top to bottom and also from ips to y They are of 
rather firm consist and truly 
whereby the 
are flattened with inclined, slightly angled sides like 


introrse, 
two parallel dose touching thec 


normal ones. These are not tapered at the top nor at 
the base, but the anthers are rather bounded by firm, 
solid ends on both sides. There are stiff, short bristles 
located at the sides of the anthers, in addition to very 
ims papillae, which are only visible with a lens, 
which give the surface an iridescent appearance" 

riii 1888: 356 

Most importantly, Ecken (1888, 1891) was the 
first to notice the overall similarity of the pollen 
catapult mechanism of Posoqueria and Molopanthera 
and its uniqueness within the family. He observed the 
flower buds of Molopanthera in herbarium specimens, 
noticed that th. 


e anthers are initially united in an 
ellipsoidal ile 1 
d 


structure while in ower bud, and 
suspected the stamen catapult mechanism. He wrote, 
“I think it is not impossible that this movement is 
executed with certain vehemence. But this question 
can only be investigated with living material, which is 
something I should like to point out to those botanists, 
who are lucky enough to be in a position to do this" 
(Schumann, 1888: a 
Delprete et al., 200 

the floral mor bes sd willen catapult mechanism 
with that of Posoqueria and declared that they 
correspond entirely. md he compared the 


translated from German; 
umann also compared 


corolla aestivation of the two genera and described 
them with lobes variably MN ie stating that in 
Molopanthera “the two lowermost lobes overlap the 
two middle ones, and these two [in turn] cover the one 
on top. This aestivation is constant, and without a 
doubt the result of corolla genesis. Inconsistent is only 
the overlapping of the two front corolla lobes, and this 
I have also always found in ascending aestivation. 
Here the right lobe sometimes overlaps the one to the 


left and vice versa. Bearing these observations in 


Volume 96, Number 1 
2009 


Delprete 81 
Taxonomic History, Morphology, and 
Reproductive Biology of Posoquerieae 


mind, I examined, with some difficulty, the aestivation 
of Posoqueria, and I was finally able to definitely 
ascertain that it was exactly the same as in 
(Schumann, 1888: 359, translated 
from German; Delprete et al., 2005a: 58 

Schumann irum 9-10) positioned both genera 


Molopanthera” 


in the 


subfamily oe ever, pro- 
ably because of t be [im pS included 
Molopanthera in n PE Cinchoneae, subtribe 


Cinchoniineae (as *Cinchoninae-Cinchoneae"), near 
Coutarea Aubl., 
with zygomorphie flowers. 


the only other genus of this group 
On the other hand, he 
positioned Posoqueria in the tribe Gardenieae, 
subtribe Gardeniinae (as Eugardenieae), probably 
guished it 
from the other genera by having flower buds laterally 


because of its leathery berries, and distin. 


bent at the lobes ms 
4a) iure the classification 
891) and contributed 
DEUS In addition, he 


Bremekamp (193 
proposed by Schumann (18 
several importa 
proposed the ixoroid pollination syndrome (pollen 
presentation at the style apex) as a strong taxonomic 
character for the subfamily Ixoroideae. However, his 
classification was focused on genera occurring in 
Suriname, thus he did not discuss the systematic 
position of Molopanthera. He zu the Gardenieae 

within the Ixoroideae and, in the three works de- 
dítated to the Flora of Surinam m 1934a, 
b, 1937), maintained Posoqueria in the Gardenieae. 
However, follow chumann's observations E 
1889, 1891), ic pum that Posoqueria does 
belong to this tribe, but did 
position in the family. In fact, the pollen catapult 


not suggest any jun 


mechanism of Posoqueria (and Molopanthera) is not a 
form of secondary presentation, as the pollen is thrown 
direetly from the anther onto the pollinator. 
Verdcourt’s (1958) family classification was pro- 
T influenced by the taxonomic observations of 
a ha 


he agreed with that author that 


igher chromosome number 
been reported for any other Rubiaceae, peculiar horny 
anthers like some Apocynaceae and it does not show 
mechanism that other members of 
cur 1958: 246 

966: 25-26), in his last notes on 
Rubiaceae übiditeation- declared that "[t]he true 
Gardenieae are recognizable by their many seeded, 
comparatively large fruits, which are provided with a 
thick, leathery or more 

gelatinous endocarp in which the numerous seeds are 


or less woody pericarp and 


embedded. They are not rarely dioecious, in which 
case the male flowers are provided with a style of 
which the upper part serves as ‘receptaculum pollinis.’ 


Fruits of the kind described above are found in the 
genera Gardenia Ellis, Randia Houst. [Randia L-J, 
Rosenbergiodendron Fagerl., Tocoyena Aubl., 
L., Alibertia A. Rich., Ibetralia Bremek. [= 
ubaea], Duroia L. f. and perhaps some other ones." In 
addition, he maintained that Posoqueria is excluded 
rom the Gardenieae as a genus probably related to 
Cladoceras Bremek. because of the unique pollination 
mechanism. 

Robbrecht and Puff (1986) presented a com- 
prehensive survey of the Gardenieae-Ixoreae com- 
plex using data from morphology, anatomy, cytology, 
and reproductive biology. However, several Neotrop- 
ical genera traditionally positioned in the Gardenieae 
were not mentioned in the study, among them 
Melanopsidium Colla, Posoqueria, and Botryarrhena 


Ducke. 

Robbrecht (1988) proposed a system of classifica- 
i of Bremekamp (1966), 
complemented by a synthesis of all the data available 


tion highly influenced by that 


to him at that time. He divided the family into four 
subfamilies and 44: tribe 
ieae in the su 


es and positioned the Garden- 
e. He delimited the 
Gardenieae according to the conclusions presented by 
Robbrecht and Puff od and divided it into the 


e an 


ubfamily Ixoroidea 


eniinae, positioning Poso- 
n the second M without any i ad dinal 
comments. 
Andersson and Persson (1991) presented a phylo- 
efine the tribe 


genetic study with an attempt to de 
C 


inchoneae. In this work, Molopanthera was found in 
a basal position near Condaminea DC. in the two 
analyses using a hypothetical taxon (combining the 
characters of the Loganiaceae genera Antonia Pohl 
and Gelsemium Juss.) as outgroup, and as sister genus 
with Condaminea in the analyses using Gelsemium as 
outgroup. These results prompted the authors to 
include Molopanthera among the genera that they 
provisionally transferred to the Condamineeae. 

Delprete (1993) presented a preliminary phylogeny 
using panne characters focused on represen- 
tative genera e hiococ cceae, Catesbaeeae, 
Condamineeae, pt Rondelet Based th 
results of this study, he indicated dat the subtribe 
Portlandiinae of the Condamineeae should be sepa- 
rated as the informal Portlandia group in which the 
genera Catesbaea L. and Phyllacanthus Hook. f. 
Catesbaeeae) should be included. In addition, he 
indicate d that Molopanthera ed be tentatively 

ncluded within the Rondeletiea 

Robbrecht (1993), in a c" to his 1988 


classification and following Delprete's preliminary 


uta 


MEE separated the “genera associated with Port- 


o 


andia.” He positioned this informal group near the 
ae ae with the note “If the Catesbaeeae are 


82 


Annals of the 
Missouri Botanical Garden 


included (see tribus incertae), they will provide a 
tribal name for this group" (Robbrecht, 1993: 176). 
Within this informal group, he included Molopanthera 
without any further comments. 

Delprete (1996) published an expanded phyloge- 
netic study based on morphological characters pre- 
sented in ee 3. In the ape using Cinchona L. 

and Joo H. Karst. as outgroup, 
ecd at a basal n position. In the 


Molopanthera 


a using solely Coffea L. as outgroup and in 
that with Coffea, Cinchona, and Joosia as outgroup, 
oe was positioned in a clade with Para- 
chimarrhis Ducke and Simira Aubl. These results 
to propos 
member o Hu T ie ex 
(Rondeletieae s.1.). 

Persson (1996) published a phylogenetic study of 
the tribe Gardenieae using morphological characters. 


prompted e Molopanthera as a 
f the Co 


In the cladograms obtained, Posoqueria was posi- 
tioned within the outgroup, with the following 
parallelisms: P extension of connective absent, 
pulp present, and e 
tangential wall thicke 


placenta exotesta without radial 
nings. d on thes 
results, Persson excluded Posoqueria from the Gar- 


denieae, but did not further indicate its position in the 


amily. 
Delprete (1999) included Molopanthera in his 
widely delimited Rondeletieae (including Sipaneeae 
and Condamineeae p.p.) as related to Chimarrhis 
. because of its narrowly imbricate corollas an 


PARAR In his taxonomic revision, he maintained 
it as a monospecific genus, with the same two varieties 
recognized by Schumann (1889). 

Rova et al. (2002), with a molecular phylogeny 
using trnL-F sequences, were the first to demonstrate 


the close phylogenetic relationship between the 


o be c t tri 
Henriquezieae and Sipaneeae, as further supported 
by the molecular phylogenies of Delprete and Cortés- 
B. (2004) and Cortés-B. et al. (2005), using trnL-F and 
rps16 sequences. 

Delprete (2004), based on morphological, palyno- 
logical, and phylogenetic evidence, described the new 
tribe Posoquerieae, with these genera as shrubs or 
small to tall trees, with stipules triangular or oblong- 
lanceolate; terminal inflorescences; flower buds 
gibbous (Posoqueria) or curved (Molopanthera); corol- 
la zygomorphic, long-tubular (Posoqueria) or rotate, 
small Sn kin ie anthers apiculate, base agitate 
or caudate, organized in bud in two pairs with a single 
one bearing a pollen mass released by all the anthers; 
pollen grains 3-colporate; ovary bilocular; fruits 
baccate (Posoqueria) or capsular (Molopanthera); and 
seeds many, large, and wingless with testa coriaceous, 


perlaceous (Posoqueria) or seeds minute with wing 
lacerate-dentate and testa membranaceous (Molo- 
panthera). 

Robbrecht and Manen (2006) proposed a new 
family classification based on a phylogeny obtained 
DNA 


this new system, 


from nuclear and chloroplast 
(supertree eid In 
divided the Rubiaceae 
several a They re 
quezieae and Posoquerieae to subtribes of the 
expanded Henriquezieae (sister to the tribe Sipa- 
neeae) However, despite these recent results, I 

prefer to maintain the Henriquezieae and Posoquer- 
ieae as sister tribes. The remarkable floral morpho- 
logical characters and the pollen catapult system 
unique to the Posoquerieae warrant recognition at 
the tribal level. 
show some very different characters, e.g., 


the Henriquezieae 


the half- 


Furthermore, 
superior to superior ovaries. 


CHARACTERIZATION OF THE TRIBE PosoQUERIEAE 


Genera with zygomorphic flowers are uncommon in 
the Rubiaceae; however, the trademark of the tribe 
d is the peculiar pollen catapult mecha- 
nism tha uires a composition of Rd Kd and 
ton. characters in order to undergo the various 
stages of anthesis (see below). With the goal of rum 
a general view of the morphological variation in the 
tribe and a 
included, an itemized characterization is presented 


below 


comparison between the two genera 


GEOGRAPHIC DISTRIBUTION 


us of about 17 
distributed throughout the SR from Central 
America to southern Brazil (Macias, 1988; Taylor & 
Cortés-B., 1999; Boom & Delprete, 2002; Macias & 
Kinoshita, 2003; Delprete et al., 2005b 
panthera is a monospecific genus (Delprete, 1999) 


species 


olo- 


endemic to the Atlantic forests of Brazil, with two 
varieties distinguished by the type of vestiture of the 
vegetative parts. 


HABIT 


are large shrubs (1.5-)2-7 m 
tall or, ib ient trees to 15(-20) m tall (espe- 
lly the Amazonian with bark usually 
smooth, or rarely shall fissured in old trees. 


Posoqueria species 


cially t species), 
Molopanthera is represented by trees 5-10(-30) m 
tall, with the main trunk 15-30(-80) cm DBH, and 
the bark longitudinally fissured (Fig. 1B) and pale 
brown 


Volume 96, Number 1 Delprete 83 
2009 Taxonomic History, Morphology, and 
Reproductive Biology of Posoquerieae 


B 

Xx S 

Figure 1. Molopanthera and Posoqueria. A—D. neto a: pe Inflorescence. —B. Bark. —C. Detail of 
inflorescence with flower buds. —D. Pollen. E, F. Posoqueria longiflora. Inflorescence with flower buds and open 


flowers. —F. Pollen. (A-C photos by E: Delprete; D aa with permission ay Huysmans et al., 1999; E photo by L. 
Westra; F reproduced with permission from Persson, 1993.) 


84 


Annals of the 
Missouri Botanical Garden 


LEAVES 


eaves of both genera are ovate, elliptic, or 
odr 


The 1 
oblong-elliptic and have brochidodromous venation 
(Fig. 1A), which is the common condition in the 


family. 


STIPULES 


As in most members of the family, the stipules of 
and free at the base, 


era are inte r 
iffer in several characters. In Poso- 


both g 

Mrs. 
queria, ilio. are ovate, narrowly triangular, oblong, 
ligulate, or lanceolat 


those of Mol. 


opanthera are broadly triangular at the 


e and are readily caducous, while 
base, acuminate at the apex, and persistent. 


INFLORESCENCE 


In both genera, the inflorescences are terminal. In 
Posoqueria, they are cymose or corymbose, and few- to 
many-flowered. On the other hand, those of Molo- 
idi are laxly paniculate, many-flowered, with 
nches subtended by leaf-like bracts 
phe tertiary branches thyrsoid, and with 1- 

o 3-flowered terminal units (Fig. 1A, 


econdary bra 


FLOWERS 


In both genera, the flowers are zygomorphie, 5- 
merous, bisexual, and protandrous, with a glabrous 
corolla that is white to cream-white during anthesis, 
commonly turning pale yellow to yellow after anthesis. 
However, the main contrast between the two taxa is 
the difference in corolla size and shape. In Poso- 
queria, the flower buds are narrowly cylindrical and 
laterally bent at the apex (corolla lobes); the corollas 
cm long (28-35[-38] 


cm long in P. longiflora; Fig. IE), with a long, 


are m. 7-35(- 


narrowly cylindrical tube, 5-32(-34) cm long; and the 
corolla lobes are equal or unequal, imbricate or left- 
contorted, ovate, oblong-ovate, or oblong to lanceo- 
late, and obtuse or round at apex. In Posoqueria, the 
flowers are odorless or slightly fragrant during the 
dusk to the 
middle of the night. They are visited and probably 


daytime, becoming strongly fragrant from 


pollinated by long-tongued sphingid moths (Bawa & 
Beach, 1983; and pers. obs.) 

In Molopanthera, the flower buds are curved, 
slightly wider medio-distally (at the anther position); 
the corollas are rotate, deeply lobed, 3—4 mm long, 
with a sh indrical tube, 0.3-0.5 
(Fig. 1C); and the corolla lobes are unequal in length 
(shorter on the ventral side of the bud), imbricate, 
oblong-linear, and round at apex. In Molopanthera, 


ort, cylindrica 


the flowers are sweet-scented and open during the 


daytime. They are visited and probably pollinated by 
small bees (pers. obs.). 


STAMENS 


most species of Posoqueria, the two dorsal 
filaments are the longest, the two lateral ones are of 
intermediate length, and the ventral one is the 
shortest. However, in a few species (e.g., P. tarairensis 
C. M. Taylor & Cortés-Ballén), the filaments are of 
equal length and during anthesis they separate 
independently. This species does not show the typical 
pollen catapult mechanism. Only a few species of 
Posoqueria have been closely examined for the 
unequal length of the filaments, and even fewer were 
directly observed during anthesis. 

n Molopanthera, stamen morphology and pollen 
release are very similar to those of Posoqueria. As in 
Posoqueria, the anthers are initially united in an 
ellipsoidal structure held slightly oblique (Fig. 2A) 
according to the flower bud curvature; the two dorsal 
stamens are the longest, the two lateral ones are of 
intermediate length, and the solitary one (inserted at 
the ventral portion) is the shortest. The anthers are all 
of equal length, although those connected in pairs are 
slightly narrower than the solitary one (responsible for 
throwing the pollen onto the pollinator). 


POLLEN PRESENTATION 


Most species of iu din display the character- 
catapult mec 
trademark feature of the E However, the 


istic pollen 


presence of this peculiar mechanism has not been 


observed in all the species of the genus, as som 


species apparently have equal or subequal stamens 
that separate freely, without performing the pollen 
catapult (as reported by Burk [1884] in “Posoqueria 
hirsuta,” a name of doubtful application). In other 
species, the separation of the stamens occurs only 
after the pollen dispersal. Therefore, even though most 
species have been reported to display the typical 
pollen catapult mechanism, this remains to be 
confirmed in a number of species. 

The 
latifolia was first described, in much detail, by Müller 
(1866) and later by Hallé (1967), Beach (1983), and 
Puff et al. (1995). In addition, I have personally 
observed P. latifolia in several natural 
d southern Brazil (states of 
Goiás, Minas Gerais, and Santa Catarina), and this 


pollen catapult mechanism of Posoqueria 


opulations in 
Costa Rica and central an 


species is used here as an example for the several 
stages in the anthesis. In P. latifolia, the five anthers 
are initially united into an ellipsoidal structure, which 
is held in an oblique position in relation to the corolla 


Volume 96, Number 1 Delprete 85 
2009 Taxonomic History, Morphology, and 
Reproductive Biology of Posoquerieae 


2. Stages of anthesis of Molopanthera. —A. Flower bud, with anthers held in ellipsoidal structure. —B. Pollen 


Figure 
catapult mechanism. 
stamen ini 
expanded and receptive. 


C. Stage following the pollen catapult, with the two stamen pairs folded backward and the solitary 
bove th lla mouth. —D. Final stage with solitary anther folded backward and the style branches 


tube, as it can be seen in the flower bud (Fig. 3A). two anther couples. The anthers dehisce longitudi- 
Each anther has two basal and two apical appendages, nally and, while they are still united, they release a 
which are sterile extensions of the thecae, usually loose pollen mass at the center of the structure (Puff et 


much darker. These function as a contact zone for the al., 1995: fig. 14E). Müller (1866) and Puff et al. 


86 


Annals of the 


Missouri Botanical Garden 


e 3. e cum 2m A um representation of the pollen catapult mechanism. —4A. Flow: 
B, C. An 


mo: in ellipsoidal str e. thers held in 


mouth. (Modified from F. Müller, 1866.) 
(1995) reported that the anther structure points 


according to personal observation, they might also 
point upward or sideways, especially in some other 
species that may also have erect flowers (e.g., P. 
ei (Rudge) Roem. & Schult.; pers. obs. i 

Suriname). The corolla may open prior to pollination, 
exposing the anther ellipsoid structure (Fig. 3B, C), or 
may remain closed, with the anther structure enclosed 
inside the corolla lobes; this variation is possible even 
within the same individual (e.g., P. latifolia; pers. obs. 


structure exposed, 
the anther structure, the ventral stamen makes a 
forward movement (Fig. 3D, E), throwing a dust (or 
minute clumps of grains) of pollen onto the visitor (pers. 
orted by Hallé red 
cluded, it w 


observed that some hawkmoths forced the entrance of 


+= 


obs.; not a globose mass, as re 


n flowers with the anther structure inc 
their proboscis at the top of the flower bud (pers. obs.), 
causing the corolla lobes to open and d in the 
nther, releasing a 


sudden movement of the cat g 
dust of pollen onto the heads J At the moment of 
throwing the pollen, the two lateral pairs of anthers 

remain momentarily erect, folding backward shortly 
after, with the solitary stamen remaining erect above 
the corolla mouth. This stage has the obvious function 
of preventing the visitation of potential pollinators, as 


the flower at this point is devoid of pollen and the 


ellipsoidal structure above the corolla. 
pollen catapult, with Ten two dre € pairs folded backward, and the solitary stamen 


r bud, with 
=D, E. Sage following the 
above the corolla 


stigma is not yet receptive. The third stage of anthesis is 
represented by the solitary anther folding backward (in 
ventral position), probably due to cell shrinkage, which 
liberates the mouth of the corolla. This is followed by 
the expansion of the style and a final receptive stage, 
with the style either remaining included (e.g., 
latifolia, pers. obs.) or further elongating and becoming 
exserted (e.g., P. longiflora, pers. obs. in Ecuador, 


wm 


uriname, an Brazilian state of Tocantins), 
i h 


t 
depending on the species. The same catapult mecha- 
nism described in P. latifolia was also observed and 
photographed in P. longiflora (Fig. 4A, B). 
Molopanthera paniculata is a species becoming 
quite rare in nature, as the Brazilian Atlantic forest is 
now almost completely destroyed. I was able to find a 
healthy population at the Feliciano Miguel Abdala 
atural Heritage Private Reserve (al 
“Caratinga i 
Brazil. Studies on the pollination biology of this genus 


iological Station”), in Minas 


are planned for the near future. The pollen catapult 
mechanism of Molopanthera was personally observed 
for the first time at this locality, and although the 
flowers are much smaller, the process is practically 
identical to that i e fiv 

initially united into an ellipsoidal structure held at an 
as the 
Fig. 2A). As in Posoqueria, each anther has basal 


n Posoqueria. e anthers are 


oblique position, 


— 


and apical appendages, which are sterile extensions of 
the thecae, much darker, and which function as a 


Volume 96, Number 1 
2009 


Delprete 87 
Taxonomic History, Morphology, and 
Reproductive Biology of Posoquerieae 


e 4. Posoqueria longiflora, flowers in two stag 


Figure 
anthesis. —A. Flower 


of 

with anthers held in ellipsoidal 

structure above t : 
—B. Later 


solitary stamen folded backward liberating the corolla mouth; 
the style is in the process of expansion and is still included 
within the corolla tube. (Photos by L. West y 


contact zone for the two anther pairs. While still 
united, the anthers dehisce longitudinally and release 
a loose pollen mass at the center of the anther 
structure. When the visitor (most probably a small 
bee) touches the o structure, the ventral stamen 
springs forward (Fig. 2B), throwing a dust of pollen, 
while the two lateral pairs of anthers remain 
momentarily erect. However, the flowers were not 
directly observed while visited by the pollinators, but 
the catapul i 


c t movement was stimulated by lightly 
touching the tip 


of the flower bud 


Shortly after the catapult movement, the two lateral 


s with a small pin. 
stamen pairs fold outward, and the solitary stamen 
remains erect above the corolla mouth (Fig. 2C). As in 
osoqueria, the erect stamen has the function of 
obstructing the visitation of possible pollinators, as 
the flower at this point is devoid of pollen and the 
stigma is not yet receptive. At the final stage of 
anthesis, the solitary stamen shortens and folds 
backward, Med the mouth of the 
th b 


e style and style branches expand, exposing the 


corolla, and 


receptive ur (Fig. 2D) 


POLLEN 


According to the Mars en by P 
(1993) a 
of scu (Fig. 1D) a Posoqueria (Fig. 1F 

milar. They 


v 
wmm and with ectocolpi acute 


ersson 
nd Huysmans et al. (1999), the pollen grains 
are spheroidal (or oblate), 

at both ends 
The exine is reticulate with lumina gradually 
decreasing in size toward the poles, and supratectal 
processes are absent. However, this combination of 
characters is one of the most common in the family. 
The main difference between the ge of the two 
genera is found in the size, 34-57 X 40-59 um in 
Posoqueria (Persson, 1993) and 14-16 X 14-17 um 
1999), which is 


positively correlated with the flower size in eac 


in Molopanthera (Huysmans et al., 


OVARY 


n both genera, the ovary is 2-locular (sometimes 
with incomplete placenta and 1-locular in Pos 
queria), the placenta has a basal stalk, elevating the 
portion where the ovules are attached to the central 
septum, and the ovules are numerous. The placental 
extensions are quite different, terminating with two 
lateral lamellas in Posoqueria, and with a globose 
structure in Molopanthera. 


STYLE 


The style of Posoqueria is bilobed with oblong-ovate 
stigmatic branches. In most species, it elongates in the 
female stage of the flower, becoming exserted and 
e after the anthers have folded backward. 

ever, in some species (e.g., P. tarairensis), the 
ne appa pn remains bud even during the 
receptive stag 

Similarly, i style of Molopanthera is bilobed, with 
g and slightly 


the stigmatic branches p oblon 


reflexed at maturity (Fig. 2D). It functions much in the 
same way as that of Posoqueria. However, with the 
first stage of anthesis, the style is the same 
ength as the corolla tube an 


anthers have folded backward, the style expands and 


ou 
d not receptive. After the 


— 
5 


the style branches eventually elongate and become 
receptive. 


FRUITS 


Aside from the flower size, the most impressive 
difference between Posoqueria and Molopanthera lies 
in fruit size, seed type, and dimensions. Obviously, 
this set of characters influenced most rubiologists in 
keeping the two genera far apart in all historical 
classifications. 


88 


Annals of the 
Missouri Botanical Garden 


The fruits of Posoqueria are leathery or woody 
berries, globose, ovoid to ellipsoid, and 2.5—5 cm in 
diameter. The large seeds are found in the central 
portion, immersed in a white, gelatinous pulp. 

The fruits of Molopanthera are capsular, 2— 
3.5 X 3.5-5 mm, thin, woody, strongly bilobed, with 
the two sides subglobose, and with loculicidal 
dehiscence 


SEEDS 


of Posoqueria are attached to the two 


seeds 
lamellar extensions of the placenta, which is some- 
what difficult to detect in mature fruits. They are 6— 
15 mm in diameter, round or ovate in outline, d 
angled or flattened. According to Persson (1995), t 
exotesta cells are isodiametrical to elongate an 
parenchymatic. Following personal observations, the 
seeds are perlaceous, with the outer portion of 
sweetish and edi and 


e, 
"us and 


atinous consistency, 
dee d by birds (e.g., parrots, pers. 
mammals (e.g., monkeys, capivaras, pers. obs. 

The seeds of Molopanthera are eee ales 
to the globose placental extensions. They are 0.8-3 
x mm, very irregular, deeply fringed in 
outline, with a central hylum, and wind dispersed. 
The testa is shallowly reticulate. Exotesta cells are 
elongated, with radial orientation, and interspaces 
have foveolate-reticulate thickenings (Delprete, 
1999: 38, fig. 11B). Because of these features, they 
i which 
explains Delprete’s (1999) positioning within the 
Ro 


are very similar to those of Chimarrhis, 


ndeletieae. 


CONCLUSION 


As disc 
Posoqueria display th 


ussed here, apparently not all the species of 
e catapult mechanism typical of 
this tribe. In the species reported to have stamens with 
2 e the flower buds are iid nds 

n be n the dere of P. 
(Taylor & RUN B., 1999 


bent, as in species on = ie catapult mecha- 


airensis 
1) and not eiu on 


nism. This feature definitely needs further morpho- 

logical, anatomical, and ur RUE studies in order 
itin the gen 

Additional field observations of | both Posoqueria 


to detect patterns of evolut 
Molopanthera are necessary for a complete under- 
standing of the pollination bloss and the identifica- 
tion of the pollinators of this peculiar tri 

This work corroborates that the pollen catapult 
mechanism of Molopanthera and Posoqueria is 
generally identical, as originally pointed out by 
(1888, 1889, 1891) i 


close relationship between the two genera. 


Schumann his confirms the 


Literature Cited 


Andersson, L. & C. Persson. 1991. Circumscription of the 
tribe Cinchoneae "E cladistic approach. Pl. 
Syst. Evol. 178: 6 

Aublet, J. B. C. F lime ducam Pp. 133-136, fig. 52 in 
a des Plantes de la Guiane Frangoise. P.-F. Didot 

ne, Paris. 

Baillon, H. E. 1880. Rubia . Pp. 257-50 
in Histoire Naturelle des pta Vol d. » Hachette et 
Cie, Leipzig, Paris 

Bawa, K. S. € J. H. Beach. 1983. Self-incompatibility 
systems in the Rubiaceae of a tropical lowland wet forest. 

Amer. J. Bot. 70: 1281-12 
Beach, J. 1983. Posoqueria latifolia (Boca de Vieja, Guayaba 
e Mico, Fruta de Mono). Pp. 307-308 in D. H. Janzen 
(editor, Costa Rican Natural History. University of 
o ja c 
& 


Boo C. Delprete. 2002. Rubiaceae. 

Pp. one in E AL Mori et al. (editors), Guide to 
the Vascular Plants of Central French Guiana, 
Part 2: Dicotyledons. Mem. New York Bot. Gard. Vol. 
76(2). 


Bremekamp, C. E. B. 1934a. Notes on the Rubiaceae of 
Surinam. Recueil Trav. Bot. Neerl. 31: 248-308. 
1934b. Rubiaceae. 2) 113 a in A. Pulle = 
Flora of Suriname, Vol. 4. J 
pg 


Rubiaceae (additions and corrections). 

Pp. i 491 in A. Pulle wee ie of Suriname, 

Vol. 4(1). J. H. de Bussy Ltd., 

emarks on the Ms on, d delimitati; 

and the subdivision of the Rubiaceae. Acta Bot. Neerl. 1 
-33 


4. Sur lorganization florale chez quelques 
Buitenzorg 4: 12-87. 


of some genera. XVII Inter 
decia and Abstracts. Austria Center (17-23 July 2005), 


Vien 
Darwin, n 1876. The Effects of Cross and um fertilisation in 
the Vegetable Kingdom. J. Murray, Lon 
—. 1877. The Different Forms E ear on Plants of 
the Sam e OPEM J. Murray, Lon 
Delprete, P. G. 1993. Proposed P of the tribes 
Chiococceae, Condam 
i A 
p- onference on the Systematics of the 
Hubiscens (4-6 Oct. 1993), St. Louis. 
996. Evaluation of E 
Condamineeae, and Catesbae 
morphological characters. 
192. 


tribes s in 
o Rubiaseac ) based o 
om Bot. Belg. 7: 165- 


999. n m (Rubiaceae), Part I. Fl. Neotr. 


Monogr. 77: 1-2 


. 2004. M P. 23 in P. G. Delprete, L. ré 
Smith & R. B. Klein, Rubiáceas, Vol. 1—Gêneros T 
G: l. Aleis até 19. Galium (A. Reis, editor), F 
Ilustrada Catanin ense. Herbário Barbosa Rodrigues, e 
Santa Catarina, Bra 

& R. Cortés- B 2004. 
tribe Sipaneeae (Rubiaceae, Ixoroideae), u 
TS sequence data. Taxon 53: 347-356. 


i cna Ms of the 
irnL-F and 


Volume 96, Number 1 
2009 


Delprete 
Taxonomic History, Morphology, and 
Reproductive Biology of Posoquerieae 


chu ster & P. Hiepko. 2005a. An annotated 


notes on the Rubiaceae type specimens kept at the Berlin 
Herbarium. Bot. Jahrb. Syst. 126(1), 3-69 
———À, Smith & R. B. Klein. 2005b. sects 


Vol. Era de G-Z: 20. gd até 46. Tocoy: 
(com Dp ecologicas por R. Klein, A. Reis 
Iza). Pp. 345-843 in Reis er "Plora Ilustrada 
Gummer Herbário Barbosa Rodrigues, Itajaí, Santa 
Catarina, Brazil. 

Hallé, E 1967. Étude biologique et morphologique E 3 

tribu des Gardeniées (Rubiacées). Mem. O.R.S.T.O.M 

- d 

Hooker, J. D. 1873. Ordo LXXXIV. Rubiaceae. Pp. 7-151 in 
G. Bentham & J. D. Hooker end Genera Plantarum, 


Vol. 2. Lovell Reeve & Pa London. 

Huysmans, S., E. Robbrecht, P. G. Delprete & E. Sm 
1999 [2000]. „Folen mole ES for the 
Catesba Exostema compl (Rubia- 
ceae). Grana am 325- 338. 

Karsten, H. 1849. Stannia. Pp. 27-30, tab. 9 in Auswahl 


neuer und sehon blühender [sie iusti s. Verlag 
der Deckerschen geheimen Ober-Hofbuchdruckerei, Ber- 
lin. 


— —— . 1856. Plantae columbianae. Linnaea 28: 241—546. 

. 1860. Stannia metensis. Pp. 51—52, pl. 25 in Florae Col- 

umbiae, Vol. 1. Ferdinandi Duemmleri successores, Berlin. 

— ——. 1887. Bentham-Hooker's “Genera Plantarum" un 

Florae Columbiae specimina selecta. Bot. Jahrb. Syst. 8: 
4—360. 


Macias, L. 1988. Revisáo Taxonómica do Género Posoqueria 
Aubl. (Rubiaceae). Master's Thesis, Universidade Esta- 
dual de M e e Paulo. 


oe S. Kinoshita. 2003 p new species of Posoqueria 

from P Brazil. Novon 13: 206-208. 
Müller, F. 1866. Uber die dum der Martha 
ot. Zeitung (Berlin) 23: 


(Posoqueria?) fragrans. 
A. 


Pollen eee of the Gardenieae— 
Gardeniinae (Rubiaceae). N ord. J. Bot. 13: 561-582. 
———. 1995. 

Cardoniinas (Rubiaceae). 

———. 1996. Phylogeny of ue Gardenieae (Rubiaceae). 
Bot. I. Linn. Soc. 121: 91-1 


Planchon, A E 1850. Posoqueria formosa. Pp. 169—171, pl. 
587 in . van Houtte ae ee de Serres et des 
Jardins E m urope, Vol. 6 uis Van Houtte, Gent. 


of Rubiaceae. Morphology, M 
their role in Nd ecology. Ann. Missouri Bot. Gar 
— gp. 
rq E. nee Tropical woody Rubiaceae. Poiana 
c features and progressions. Con o EAD a ne 
subfarilil peur Opera Bot. Belg. 1 “mn. 
3 [1994]. Supplement to D g caine of ~ 
classification of t iaceae. Index to Gen 
73-196 in E. Robbrecht (editor), Lo in a 
ids i Opera Bot: Belg. Vol. 6. 

& J. -F. Man n. 2006. T l i lineages 
of the coffee B (Rubiaceae, at gaper; Combined 
A) to QNT thie Po HDN 2 

bas 


xd trnL-irnF, and ap. rbcL data. A new 

classification in tw ous P diia and 

Rubioideae. sy St. Geogr. “PLT 6: 85-1 

uff. 1986. A survey of 7 ee and 
related Ad Bot. Jahrb. Syst. 108: 63-13 

Rova, J. H. E elprete, L. Pu A. Alber 
2002. E irl- F cpDNA sequence study of Hie Condani- 


on the ee of the Rubiaceae: Amer. J. Bot. 89: 
145-159. 
ens K. 1888. 
geka ie 
Jahrb. E 10: 302 
ade. tib, X-XIX. Pp. 125-466 in C. F 
Martius & G. Eichler (editors), Flora 
Brasiliensis, Vol. 6(6). VEN Leipzig. 
. 1891 - Rubiaceae. Pp. 1-156 in A. Engler & K. 
natürlichen Pflanzenfamilien, 


Über einige ieee te oder wening 
ca Rubiaceen Südamerikas. Bot. 


Seyermaik J.A 4. Rubiaceae; Pp. 1-2070 in T. Lasser 
& J. A. St a. (editors), Flora de Venezuela, Vol. 9 

Instituto Botánico, Caracas 

Taylor, C. M. & R. Cortés- Ballén. 1999. Una especie nueva 
e Posoqueria (Rubiaceae) de la Guyana Colombiana. 
Novon 9: 428-430. 

Turczaninow, N. S. 1848. We cbe ST Sees et 
Quinta, generum adhuc non m. Bull. Soc. Imp. 


58. Remarks on " classification of the 
Rubiaceae. Bull. Jard. Bot. Etat. Bruxelles 28: 209-28 


FOSSIL RECORD OF THE 
RUBIACEAE 


Alan Graham? 


ABSTRACT 


Fossils of 134 taxa attributed to the Rubiaceae are described or iani in 115 iun uiid hie from 1850 and n 
he 


deposits as old as t eous and Pa e. Close 
ed on, U.S.A., Faram 


on) and Canthium Lam. (as f 


e scrutiny of thes 
likely (accepted) tepresentatives of the family are four genera, genie: erys Oliv. from the Middle Eocene of Ore, 
re hiom thes "i mee of Panama, 


cords indicates, how that the oldest and m 


n a 
and Guettarda L. (f. as anes, T= fossil 
Late Eocene of Australia, and 1 genus, the alternate- 


pou TPaleorubiaceophyllum eocenicum dn the Middle Eocene of Tones Kentucky, U.S.A. The record represents three 


orst., Coprosma- 
TRetitricolporites Dr and Pinckne 
Puerto Rico. The period eatest diversifica 


Key words: Fossils, Rubiaceae. 


nera, Coprosma Forst. 


Aubl., Macrosphyra Hook. f. (as tTriporotetradites hoekeni), Mitragyna Korth. (as 
ya Michx. from Africa O Australia and Ne 
ation and radiation was Mi 1 


Deer study Aem 


Reconstructing the fossil history of plant families 
requires assembling reports often from widely scattered 
literature dating back centuries. For the Rubiaceae, 
there are approximately 134 taxa described or 
mentioned in 115 publications including the earliest 
ones of Unger (1850, Canthidium [Unger spelling], 
TCinchonidium, Croatia), Wessel and Weber 
+Rubiacites, Germany), Heer (1868, Galium L., Green- 
land), and others from later in the 1800s and early 
1900s. All are accounted for in the present summary, 
and none have been revised since the original 
publications. There are other reports in unpublished 
theses and dissertations, and there is casual, uncon- 
firmed mention of the family as possibly present in 
some putative Late Cretaceous and older deposits in 
the literature of the 1800s that also are not included. 


These reports must be filtered through at least a 
yield a database of 


plausible records. The procedure for evaluating the 


preliminary assessment to 
fossil pollen records of extant angiosperms used here 
is similar to that of Muller (1981). In this informal 
rating, “A” (accepted) means that: (1) the specimens 
have been reexamined and are considered to represent 
the Rubiaceae; or (2) the specimens are sufficiently 
diagnostic to allow recognition to genus or bi from 
the illustrations and/or descriptions (e.g., the dimor- 
phic pollen of Faramea Aubl.); and (3) k identifi- 
cation does not pose improbable age, phylogenetic, 
paleoecologic, or biogeographic problems to the extent 
this context information is available for the taxon and 
the locality. “P” (pending) means additional informa- 
tion (e.g., more accurate age determination, better 


cek, Gengwu Liu, Dieter Mai, Helen 
no d UE 


t is acknowledged in the 


provided many helpful suggestions that improved the 
particularly effective in locating 


rad ut Shirley A. Graham, Maria C. Zamaloa, and on: 


difficult-to-find pea and special thanks goes to Mary Sufi 
nymous reviewer. The 


e author is grateful for information on the classification and distribution of modern Rubiaceae provided by Elmar 


e project was iem by Piero Delprete. 


Missouri Botanical Garden, P.O. Box 299, St. Louis, Moda 63166-0299, U.S.A. alan.graham@mobot.org 


doi: 10.3417/2006165 


ANN. Missouni Bor. Garp. 96: 90-108. PUBLISHED ON 23 APRIL 2009. 


Volume 96, Number 1 


Fossil Record of the Rubiaceae 


preserved or more complete specimens), and/or 


confirmation through reexamination of the material 
is needed. “NA” (not accepted) means: (1) only casual 
reference is made to family affinities (e.g., the fossil 

pollen named Tricolporopollenites arnotiensis that 
Scholtz [1985] E to Anthospermum L., Nenax 

aertn., and ia L., as well as to the Euphorbiaceas 
and pe: or (2) features are present that are 
not found in or characteristic of the family. These 
unless 


assessments are the present author, 


otherwise indicated. The abbreviations are also 
explained the first time they are used in the Synopsis 
section. 
Another step in reconstructing an accurate geologic 
history for a family is to identify the location. of 
as pen 


specimens presently rate Finally, 
reexamination of this material will eventually be 
necessary by those familiar with leaf, floral, seed/fruit, 


holo. 


pollen mo xtant species, and 
e phylogenetic and biogeographic implications of 


F and 
the reports. Such pending material for the Rubiaceae 


e fruit Galium 
Tantiquum from the Paleocene of Cana and the 
pollen +Psilatricolpites coprosmoides (Coprosma J. R. 
t.) from the Oligocene to Recent of 
New Zealand (A) and reported from the Paleocene of 
Chile (P). In the following summary, abbreviations for 
the repositories of specimens (not Index Herbariorium 
abbreviations [Holmgren & Holmgren, 1998], or 


n 
of Natural History, Gainesville, Florida; V. 
J, Faculté Sciences Techniques St.-Jéróme, MM 
France; GIL, Geological Institute, Leiden, The 
Netherlands; GSC, Geological Survey of Ca 
Calgary, Canada; GSV, Geological Survey of Victoria, 
Victoria, Australia (as of 2004, GeoScience Victoria); 
S PLU a Vienna/Graz, MACN 
useo Arg entino de Ciencias Naturales “Bernardino 
E aces os Aires, Argentina; MO, Missouri 
Botanical Garden, St. Louis, Missouri, U.S.A. (the 


author’s modern spore and pollen reference collection, 


nada, 


Austria; 


fossil collection, literature collection, and associated 
aterials are c ing transferred to the 
ii d Tropical Research Institute, Panama); 
of Paleontology, uoo of 

due Berkeley, California, U.S.A.; NIGP, Nan- 
jing Institute of Geology and Pálscitlomr, People's 
Republic of China; NMP, National Museum, Prague, 
Czech Republic; NTU, National Taiwan University, 
Taipei, Taiwan; . New Zealand Geological 
Survey, Lower Hunt, New Zealand; RN-D, Realgym- 


nasium zu Neustadt-Dresden, Germany; RRNA 
Robertson Research (North America), Calgary, AL 
berta, Canada; 
town, South Africa; 
Frankfurt am Main, Germany; SOCDH, Shell Oil 
Company, Den Hague, The Netherlands; SUPA, 
Stanford University, Palo Alto, California, U.S. 
UA, University of Amsterdam, Amsterdam, The 
Netherlands; ULP, Université Louis Pasteur, Stras- 
bourg, France; UPMC, Université Pierre et Marie 
Curie, Paris, France; USGS, U.S. Geological Survey, 
Colorado, U.S.A.; USNM, U.S 
Museum M id Division), Washington, D.C., 
As South Wales, 


A, University of Vienna, 


Denver, National 


dou of New 
Kensington, Australia: 


uo Vermont, U. s. A; ZOOZ, Zoological 
Mus A sity of Z rich, Switzerland; 
ZCIB, Zen rales Geologi e pu tut, Berlin, 


many. ene for repositories are provided at 
the oo of each cinis followed by abbreviations 
for us. Superscri mbers in text denote the 
follo owing: 1, Early ee fide Berry (1938), Early 
Eocene fide Romero (1986: 454), age unsettled. 2, 
Rubiaceae affinities not cited by authors, only implied 


Presl, Chomelia, Guettarda, Terebraria et al. type. 4 
Eocene fide Romero (1986: 453), 
Palma-Heldt (1980) 


Paleocene fide 


A Synopsis or Reports or FossiL RUBIACEAE 
CRETACEOUS 


There are six fossil taxa from the Cretaceous 
referred to the Rubiaceae that represent four form 
genera and six species. None have been confirmed as 


belonging to the family. 


1. tRubiaephyllum gaylussaciae Bayer. Leaf, 
Bayer in Fritsch (1893: 131, fig. 
1 to Kvacek (pers. 
2006), the morphology and preservation of the 


Bohemia, 
192). According comm., 
specimens in this report make the family 
assignment uncertain. NMP. NA 


2-3. 1T. li arnotiensis Scholtz 
(1985: 71, ur 17d—h) and fT. brinkiae Scholtz 
(1985: 72, figs. l7a-c) Pollen, 
Africa. The beds range in age from 
Cretaceous) to 64 Ma (Early Paleocene), but 
because the samples came from the upper part of 


southwest 


the section, they are ara Early Paleocene. 
The specimens are only informally compared by 


Scholtz (1985) to the Rubiaceae (Anthospermum, 


Annals of the 
Missouri Botanical Garden 


Nenax, Rubia), as well as to the Euphorbiaceae 
and Bombacaceae. SAMC. NA 


4-5. TTriorites aspidatus and TT. megaporus 
(authors unknown, contact: G. Liu, pers. comm., 
2006). Pollen, People's Republic of China. Both 
sa records provided by G. Liu (pers. comm., 

. Only informal reference is made t 


nes by Liu (possibly Gardenia J. Ellis). d 
A. 


6. *Triporoietradites scabratus van Hoeken-Klink- 

enberg (1964: 226, fig. 16). Pollen, Nigeria, 
1 ia by Krutzsch (1970; as 
Gardenia type). According to Muller (1981), the 
specimens are too poorly preserved to be recog- 


nized as Gardenia. GIL. NA 
PALEOCENE 


Six fossil taxa are mentioned for the family from the 
Paleocene, representing five form and modern genera 
and five species. None have been confirmed as 
belonging to the Rubiaceae. Three warrant reexami- 
nation: Cinchonidium ovale Lesq., Galium antiquum 


Heer, and Psilatricolpites coprosmoides Couper. 


. TCinchonidium ovale Lesq. van 229, E 48, 
figs. 8-10b). Leaf, North Dakota f. 
Cinchona L. USNM (the specimen cannot be 
located in the USNM collections; S. Wing, pers. 
comm., 2006). P 


2. Galium Tantiquum Heer (1868: 119, pl. 17, figs. 
8, 8b; Heer, 1883: 114). Fruit, Greenland. 
ZOOZ. P 


3. tPsilatricolpites coprosmoides Couper. Pol- 
nin originally recognized as Coprosma R. 
t. & G. Forst. sp. by Couper (1953: 54, pl. 9, 
i. e Late Miocene to Recent; 1960: 59, pl. 9, 
figs. 1-3, Middle T to uus from New 
Zealand. This pollen as recognized by 
Doubinger and Chotin (19:3. 559.5 60, pl. 2, fig. 
13) from the Paleocene of Chile, and was said to 
resemble certain Coprosma. There is only a very 
brief description for the Chile record. Couper, 
NZGS or CDH, Oligocene to Recent, A; 
Douginger, ULP, Paleocene, 


4. tRetistephanocolpiies Leidelmeyer sp. Pollen, 
tetracolpate, described by Scholtz (1985: 76, fig. 
19e-g) from southwestern Africa. Scholtz (1985: 
76) notes that “No positive suggestions can be 
made regarding the affinity of Retistephanocol- 
pites sp. The pollen of Rubia (Rubiaceae) and 
Catastemma Benth. 


(Bombacaceae), amongst 


others, appear superficially similar to this fossil 
species." SAMC. NA 


5-6. tTricolporopollenites arnotiensis and tT. 
brinkiae from the Arnot Pipe sediments of Late 
Cretaceous to Early Paleocene age have been 
mentioned earlier. SAMC. NA 


EOCENE 


For the Eocene, 32 fossil taxa (including family 
reports for the Rubiaceae and the name TTricolporé 
reticulé) representing 22 form and modern genera and 
28 species are mentioned for the Rubiaceae. 

1. porta j| or Berry (1938: 

—6). Leaf, Argentina. 
ee o a de fossil to Cephala- 
nathus glabratus (Spreng.) K. Schum. growing 
today in dre Uruguay, and northern 


Argentina. USNM 


2. Coprosoma Tincerta! Berry. Leaf, Berry (1938: 
133, pl. 47, fig. 1), Argentina. USNM. P 


Tspathulatifolia! Berry. Leaf, 
8: 133, pl. 52, figs. 4, 5), Argentina. 


3. Coprosoma 
Berry (193 
USNM. P 


4. Coussarea Ttertiaria! Berry. Leaf, Berry (1938: 
131, pl. 53, fig. 5), Argentina. USNM. P 


5. tCricotriporites Sal.-Cheb. 
Pollen, Salard- Cheboldaetf (1978: 246, pl. 6, 

fig. 4), Late Eocene to Early Miocene, Cameroon. 
d BE o ee compared the fossil to 

Ran ut this is aia, 
puer » Muller TN UPMC. N 


6. cf. Emmenopterys Oliv. Fruit, Middle Eocene 
Republic flora (49-48 Ma), Washington, U.S.A. 
(Wehr & Manchester, 1996: 25, pl. 2, fig. 6). 
FLMNH. A 


7. Emmenopterys *dilcheri Manchester (Fig. 1A, 
B). Infructescence and fruit, Middle Eocene 
Clarno flora (44 Ma), Oregon, U.S.A. (Manches- 
ter, 1994: 80—81, pl. 36, figs. 1-11). The extant 
Emmenopterys henryi Oliv. of People's Republic 
of China is mentioned as similar by Manchester 


(1994) and Wehr and Manchester (1996). 
FLMNH. A. 


8. Exostema | Tpseudocaribaeum Berry (1916: 
349, pl 106, fig. 3). Leaf, Middle Eocene, 
Tennessee/Kentucky, U.S.A. When originally de- 
scribed, Berry (1916) thought the Wilcox Forma- 
tion was Early Eocene, but it is now considered 
Middle Eocene. Similarity is cited by Berry with 


Volume 96, Number 1 
2009 


Fossil Record of the Rubiaceae 


e l. A, B. Emmenopterys dilcheri Manchester from the Middle Eocene Clarno flora, Ore 


1g 
frutescence. —B. Fruit. A, 
Institution, Win New York, U.S.A., 
atuncillo , Panama, o Graham (1985). 
Solo fl co, from Graham (1976). —E 


ora, oa Mex 


s Michx., moder 


with penmissian of the University of California Press, ae ee U.S.A 
MAE A type a la the Late E Gatun flora, Panam 
e Gat 


m Graham (1991a). 
m pes Graham (1988) 


the extant Exostema caribaeum (Jacq.) Roem. & 


Schult. of the Caribbean region. USNM. P 


9. Faramea Aubl. Pollen, distinctive diporate form, 
Late Eocene Gatuncillo flora, Pa Graham 


(1985: 9-520, figs. 64, 65; Fig. 1C). MO. A 


10. Gardenia type (as TTriporotetradites nachter- 
stedtensis Krutzsch (1970: 412, pl. 48, figs. 27— 
32). Pollen, Late Eocene of ae ZGIB. A 
(by Muller, 1981; here P, confirmation needed). 


11. Mei ihe M qm Berry (1916: 348, pl. 
. 2). Leaf, Middle Eocene, Mis- 


—D. ee pollen 


US. 
m ey ae (1994), used with permission of the aonad Research 
the author. —C. 


Late Eocene 
Middle Pliocene Paraje 


Faramea pollen, diporate form, from the 
, triporate form, from the 


y E. Paleorubiaceophyllum eocenicum (Berry) Roth & Dilcher 

from the Middle Eocene Claiborne flora, MannessepiKentusly, U.S.A 

& pie. dt us the Oligocene Bridge 
eed, and 


ssion R^ Roth b Dilcher 
Creek flora, John Day Formation, 
Manchester o Im ae G, Hu 
., and the author (Manchester). ui 
a, from Graham (1991a). —J. Posqueria pollen from the 
—K. Sabicea EI from the early Miocene Culebra flora, 


Reprinted with permi 


from Meyer a 


sissippi/Tennessee, U.S.A., Berry (1916) com- 
pared to the extant Guettarda elliptica Sw. from 


the Caribbean region. USNM. P 

12-13. cf. TGuettardidites ivirensis mE (1976: 
763, fig. 24). 
tralia, fd. ANE 

14. Hoffmannia tprotogaea! Engelh. Leaf, Berry 


(1938: 131, pl. 53, figs. 1, 2), Argentina, P. Berry 
(1922: 86), Chile*. USNM (all). P 


Annals of the 
Missouri Botanical Garden 


15. flxorophyllum anceps Geyer (1887: 495, pl. 
35, figs. 1, 2). Leaf, Eocene, Borneo. Repository 
unknown. P 


16. tPaleorubiaceophyllum eocenicum (Berry) 
Roth & Dilcher (1979: 1203-1205, figs. 1-22; 
Fig. 1E, F; Paleorubiaceophyllum Ds MUN 
Lott, 2005: 17-18, fig. 9a, b). Leaf, Middle 
Eocene Claiborne flora of Sidi Me y, 
U.S.A. Additional fossil material (branches) has 
shown that the leaves are alternate (Manchester, 
pers. obs., 2006). FLMNH. A or P (because of 


alternate branching?). 


17. Psychotria Teogenica Berry (1929b: 166—167, 
pl. 3, figs. 17-21). Seed, Eocene, Peru. USNM. P 


18. Psychotria {grandifolia Engelh. Leaf, Berry 
(1916: 349—350, pl. 105, fig. 1), Middle Eocene, 
Tennessee/Kentucky, U.S.A. Berry (1916) com- 
pared this with Psychotria grandis Sw. USNM. P. 
It is also listed (as Psychotria grandifolia?) by 
Berry (1941: 84) but not described or illustrated, 
leaf, Kentucky, U.S.A. USNM. NA 


19. Psychotria Toregona Chaney & Sanborn 
(1933: 96, pl. 33, fig. 4). Leaf, Eocene, Oregon, 
U.S.A. Chaney and Sanborn (1933) compared it 
to the extant Psychotria undata Jacq. of the 
Caribbean region, but the record has not been 
verified. MPUC. P 


20-21. tRandiapollis microreticulatus Ke & Shi. 
Pollen, R. reticulatus Ke & Shi. Pollen, 
Eocene, People's Republic of China (G. Liu, 
pers. comm., 2002). Only general affinity to the 
family suggested by Liu. NIGP. NA 


22. Remijia *ttenuiflorifolia' Berry. Leaf, Ber- 
ry (1938: 132, pl. 54, figs. 1, 2), Argentina, 
USNM. P. 


23. Rondeletia tlongiflorifolia’ Berry. Leaf, Ber- 
ry (1938: 132, pl. 54, figs. 7, 8), Argentina. 
USNM. P 


24. Rubiaceae type 1 (Graham, 1985: 520, figs. 66, 
67). Pollen, type 2 (Graham, 1985: 520, figs. 68, 
69); pollen, Late Eocene, Panama. MO. A (as 
family). 


25. tRubiaceocarpum markgrafi Krausel (1939: 
108, pl. 1, figs. 19-24). Seed, Eocene, Egypt. 
SM. P (as family). 


Ec x chomeliifolia' eee Leaf, B 
S 133—134, pl. 55, figs. 
md P (as family). 


erry 
1, 2), Argentina. 


27. tRubiacites? pellicieraformis Berry (1930: 
134, pl. 49, fig. 19). Fruit, Middle Eocene, 
Tennessee, U.S.A., family but no generic af- 


finity mentioned by Berry (1930). USNM. P. 


28. piu sphericus Berry (1930: 133-134, 
45, figs. 9-11) Fruit, Middle Eocene, 
e. U.S.A., family but no 
generic affinity mentioned by Berry (1930). 
SNM. P 


29. tRubiacites wilcoxensis Berry (1930: 133, pl. 
45, fig. 8). Fruit, Middle Eocene, Tennessee/ 
Kentucky, U.S.A., family but no generic affinity 
mentioned by Berry (1930). USNM. P. 


30. tRubipollis oblatus (Pocknall & Mildenhall) 
Mildenhall & Pocknall. Pollen, MacPhail (1999: 
205, pl. 11, figs. 26, 27), Late Eocene to Early 
Pliocene, Murray Basin, Australia. MacPhail 
(1999) associated this with Canthium Lam. 
NZGS. A 


31. TTricolporé reticulé, without attribution in 
Gruas- Cavagnetto (1977). Pollen, from the late 


as Chomelia Jacq. 
(1978). It is not clear if it can be distinguished 
from the pollen of other rubiaceous genera fide 
Muller (1981). UPMC. NA 


omelia). 


(as Rubiaceae or 


TTriporoitetradites nachterstediensis (see Garde- 
nia) 


32. oe tertiaria Berry Er Pu pl. 3, 
g. 16). Seed, Eocene, Peru. USN 


OLIGOCENE 


For the Oligocene, 16 fossil taxa representing 17 
form and modern genera (including Tricolporé 
reticulé) and 13 species have been assigned or 


compared with the Rubiaceae. 


1. tCanthiumidites aff. bellus (Stover € Partridge) 
Mildenhall & Pocknall. Pollen, Argentina 
(Barreda, 1997: 286, pl. 1, figs. 10, 11). 
Barreda (1997) compares the fossil to Garde- 
nia (Old World tropics; see Miocene). MACN. 


2. tCinchonidium copeanum (Lesq.) Ettingsh. 
(1883: 130; Ettingshausen, 1888: . Leaf, 
Nevada, U.S.A. Ettingshausen (1883, 1888) 
compares the fossil to Cinchona (Andes). USNM 
(Lesquereux material), JVG (Ettinghausen mate- 
rial). P 


Volume 96, Number 1 
2009 


Fossil Record of the Rubiaceae 


3. tCircotroporites camerounensis (see Eocene). 


4. Coprosma type. Pollen, New Zealand (Couper, 

1960: 59, pl. 9, figs. 1-3). Includ 
Mildenhall (1980: 215), NZGS. A (fide Milden- 
hall, 1980). 


5. Faramea. Pollen (triporate form), Puerto Rico 


(Graham € Jarzen, 1969: 328, fig. 21). 


Faramea presently grows in tropical America. 


6. Guettarda Tintercalaris Hollick (1928: 225, pl. 
81, fig. 5b). 


Puerto Rico. Guettarda presently grows in New 


Line drawing of leaf fragment, 


Caledonia and tropical America. USNM. P. 
7. cf. tGuettardidites Khan (see Eocene). 
Macrosphyra Hook. f. (see Triporotetradites hoekeni) 


8. ¡Mitragynaxylon gevini Koeniguer & Lemoigne 
in Gevin et al. (1971: 386—393, text-figs. 1, 2; pl. 
23, figs. 1-8). Wood, Oligocene and Miocene, 
Algeria. Laboratorie Géologie Appliquée (Gev- 
in), Laboratorie de Paléobotanique (Lemoigne), 
Université de Lyon (Université Claude-Bernard); 
Laboratorie de Paléobotanique, Université de 
Paris (Koeniguer). P. 


9. TNaucleaphyllum | ovale? Louvet & Mouton 
(1970: 82-85, pl. 2). Leaf, Libya. Repository 


unknown. NA. 


10. tPalaeocoprosmadites zelandiae Pocknall. 
Pollen, MacPhail (1999: 205, pl. 9, figs. 16, 
17) Late Oligocene to Pleistocene, Murray 
Basin, Australia. MacPhail (1999) Ol the 
fossil to Coprosma—Opercularia. ASNU. A. 


11. Pinckneya Tdilcheri Meyer & Manchester 


Rubiaceae with one species (Pinckneya pubens 


Michx.; Fig. 1H) in the southeastern U.S.A. 
FLMNH. A 


12. tPsilatricolpites coprosmides (see Paleo- 


13. tRetitricolporites ue Sal.-Cheb. (1978: 
236—238, pl. 4, figs. 7-9). Pollen, Cameroon. 
Salard- Cheboldaeff m compares the fossil to 
Mitragyna inermis (Willd.) Kuntze, which grows 

in Cameroon. UPMC. A (fide Muller, 1981). 


icon naucleoides E. Hofm. (1952: 
pl. 13, fig. 3). Wood, Austria. Repository 
une (Vienna, Austria). P 


15. fTricolporé reticulé (see Eocene). 


16. P uei ee hoekeni Sal.-Cheb. (1978: 
252, pl. 7 . Pollen, Cameroon. Salard- 
Cheboldaeff (1978) compares the fossil to the 
extant ace or (DC.) Hiern that 
grows in Cameroon. UPMC. A (fide 


Muller, 


MIOCENE 


Note that fossils described by Berry (1925, 1938) 
from Laguna del Hunco and Río Pichileufu, Argentina, 
and referred by him to the Miocene, are now regarded as 
Eocene (Romero, 1986; Wilf et al, 2005; Zamaloa, 
pers. comm., 2008). For the Miocene, 58 fossil taxa 
representing 46 form and modern genera and 37 species 
have been assigned or compared with the Rubiaceae. 


Médus (1975: 


1. Borreria G. . Pollen, 576 
pl. 10 30-32), Sere 


, fig. à uL ll, figs. 
FSTS JA E 


2. Bothriospora Twitii Engelh. (1895: 30, pl. 6, fig. 
6). Leaf, Colombia (also Pons, 1985: 241). Pons 
(1985) compares the fossil with the extant 

Bothriospora corymbosa (Benth.) Hook. f. growing 
today in Colombia, Ecuador, Guyana, and Peru. 
RN-D (Engelhardt material), UPMC (Pons mate- 
rial). P (both). 


3. Canthidium tradobojanum Unger (1850: 429). 
Leaf, Croatia. The occurs today in the 
Paleotropics. UVA 


4. TCanthiumidites bellus (Stover & Partridge) 


Mildenhall &  Pocknall Pollen, MacPhail 
(1999: 205, pl. 5, figs. 5, 6), Early to Middle 
Miocene, Murray Basin, Australia; Canthiumi- 


dites cf. bellus Middle Miocene—Early Pliocene, 
Falkland Islands, MacPhail and Cantrill (2006: 
610, table 1; 613, table 3; pl. III, figs. 39, 40); 
Randia. ASNU. A 


5. Chiococca P. Browne. Leaf, Axelrod (1940; 
1979: 32), Mint Canyon flora, southern Califor- 
nia, U.S.A. The Mint Canyon flora is found on a 
terrane transported north from northwestern 
Mexico along the San Andreas Fault. MPUC. A 


6. Chomelia Jacq. type. Pollen, Graham (1991a: 
212-213, fig. 40; Fig. lE Graham, 1991b), 
Panama. MO. A (as the fani y). 


7. TCinchonidium racemosum Unger. Fruit, Un- 
ger (1850: 430; Unger, 1865: 11, pl. 3, figs. 1, 2, 
6), Croatia. UVA. P 


Annals of the 
Missouri Botanical Garden 


8. tCondaminea grandifolia Engelh. (1895: 34, pl. 
7, fig. 2; pl. 9, fig. 1). Leaf (as Rutaceae), 
Colombia. RN-D. P; Berry (1919: 293—294, pl. 
17), leaf, Peru. USNM. P 


9. Coprosma type. Pollen, Leopold (1969: x pl. 
311, fig. 37), Marshall Islands. USGS 
also +P. 


Rubiaceae; see 


wm 


A (as 
silatricolpites copros- 
moides, Paleocene, pollen). 


10. Cosmibuena Ruiz & Pav. Pollen, Graham 
(199 1a: 213, fig. 43; Graham, 1991b), Panama. 


MO. A. 


11. Coussarea fmembranacea!* Engelh. Leaf, 
ile (see Tertiary undifferentiated), Berry 
(1922: 86, listed only), Chile. RN-D (Engelhardt 


material). P. 
12. tCricotriporites camerounensis (see Eocene). 


13. tElaeagnites campanulatus Heer. Calyx, 
Heer (1876: 58, pl. 12, fig. 11), Spitsbergen. 
ZOOZ. P 


14. Endlicher(ija rhamnoides Engelh. Leaf, En- 
gelhardt (1895: 12, pl. 1, figs. 7 [as 17 in 
Ol 1895 text], 19, 20), Colombia; Berry 
(1929a: 91, listed only), Ecuador, under Rubia- 
ceae a genus belongs to the Lauraceae. RN-D 
(Engelhardt material). P 


15. Exostema tprecaribaeum Berry. Leaf, Berry 
(1939a: 132-133, pl. 18, figs. 4, 5), Cuba 
USNM. P. 


16. Faramea types 1, 2, pollen (triporate, tetra- 
porate forms), Graham (1991a: 213, figs. 41, 44, 
45, 48; 1991b), Panama. MO. A 


17. Faramea miocenica Berry. Leaf, Berry (1925: 
228—230, pl. 7, fig. 4), Argentina (Patagonia). 
USNM. P 

Faramea (see also TPsilatriporites corstanjei). 


18. tFavitricolporites magnus without attribution 
in Mandaokar (2003: 190, 


description, 


no illustrations, 

referred to Rubiaceae), Mizoram 

region, northeastern India. Lucknow, India. 

NA 

19. Galium L. Pollen, White & Ager (1994: 51, pl. 
4, fig. 22), Alaska, U.S.A. GSC, USGS. A. 


Gardenia (see also tTriporotetradites letouzeyi). 


. Gardenia cf. grievei Horne. a Leopold 
"d 969: 1175, pl. 310, figs. 16, 17), Marshall 
Islands. USGS. A (fide ng 1981). 


21. 


25. 


36. 


37 


. Gou a ttenuinervis* Eng 


. Morelia A. Rich. ex DC. 
) 


4. cf. Muss 


. Posoqueria 


Gardenia pterocalyx Valeton type. Pollen. 
Anderson & a um 307—308, listed "tX 
SOCD 


Borneo. 


` de c p taiwanensis Huang. Pollen, 


Huang (1978: 79, pl. 1, figs. 9, 10), Taiwan. 
NTU. A (fide Muller, 1981). 


elh. Leaf, 
selhardt (1891: 656-657, pl. 5, ig 6b), die P, 
Berry (1922: 86, listed only), Chile, P. RN-D 
(Engelhardt material). P 


. cf. Guettarda L. Pollen, Leopold (1969: 1175- 


1176, pl. 310, figs. 29, 30), Marshall Islands. 
USGS. A 


Guettarda *cookei Berry. Leaf, Berry (1921: 


USNM. P. Berry (1923b: 26), Oaxaca, Mexico (as 
?. fragmentary). USNM. P 


. cf. TGuettardidites Khan (see Eocene). 


. Ixora cf. Ixora casei Hance. Pollen, Leopold 


(1969: 1174, pl. 310, figs. 13-15), Marshall 
Islands. USGS. A (fide Muller, 1981). 


. Macrosphyra (see Oligocene). 
. TMitragynaxylon gevini (see Oligocene). 


. Mitragyna Korth. type (see Oligocene). 


(see Retitriporites 


boltenhagenii 
. tRetiiriporites  boltenhagenii — Sal.-Cheb. 
(1978: 247—248, pl. 6, fig. 2). Salard-Chebol- 


daeff (1978) compares the fossil to Morelia 
senegalensis A. Rich. ex DC., Cameroon. UPMC. 
A (as Morelia type pollen; Muller, 1981). 


. cf. Morinda  citrifolia "i Pollen, Leopold 


(1969: 1175, pl. 311, figs. 1, 2, 9, 10), Marshall 
Islands. USGS. A (fide Muller, 1981). 


aenda frondosa L. Pollen, Leopold 
(1969: 1149, pl. 311, figs. 3, 4), Marshall 
Islands. USGS. A 


. Palaeocoprosmadites. Pollen, Zamaloa (2000), 


Middle Tertiary, Tierra del F 
MACN. A (as family; see Oligocene). 


o, Argentina. 
Posoqueria Aubl. type. Pollen, Graham (1991a: 
213, figs. 51-53; 1991b; Fig. 1J), Panama. MO. A. 


Tcolombiana Engelh. Leaf, En- 
gelhardt (1895: 40, pl. 7, fig. 8), Colombia. P; 


Volume 96, Number 1 97 
2009 Fossil Record of the Rubiaceae 
Berry (1945: 148, listed only), Ecuador. P. Berry 51. Sabicea Aubl. Pollen, Graham (1987; 
(1936: 65-66, pl. 2, fig. 4, Colombia). RN-D 63, 64; Fig. IK), Panama. 


(Engelhardt material). USNM (Berry material). P 


e 
e 


. TPsilatricolpites coprosmoides (see Paleo- 
cene). 


ve 
© 


. TPsilatriporites corstanjei Hoorn. Pollen, 
Hoorn (1994a: 102, pl. 4, fig. 35; see also 
Colombia, compared to 


triporate form of Faramea. UA. 
Randia (see also }Triporopollenites bellus). 


40. Randia L. Pollen, Mildenhall (1980: 222), New 
Zealand. NZGS. A 


a 
a 


. cf. Randia cochinchinensis (Lour.) Merr. Pol- 
len, Leopold (1969: 1176—1177, pl. 310, figs. 31, 
32), Marshall Islands. USGS. A (Muller, 1981, but 
wording may suggest that Leopold compares the 
fossil specifically to R. chartacea F. Mue 


a 
N 


. Randia tmohavensis Axelrod. Leaf, Axelrod 
(1950: 156), California, U.S.A. MPUC. P 


A 
w 


. td aie! crassicostaius Hammen & 
Wijmstra. n, rn (1994a: 105, pl. 6, fig. 
63a, b), Colombia, Compared to Rubiaceae. UA. A. 


i 
A 


. Rondeletia L. Leaf, Berry com 26, pl. 7, fig. 
3), Oaxaca, Mexico. USNM 


a 
gt 


. Rondeletia tgoldmani Berry. Leaf, Berry 
(1918: 42-43, pl. 18, fig. 3), Panama. USNM. P 


a 
an 


. Rubia. Pollen, van Campo (1976), Spain. 
CNRS. A (fide Muller, 1981). 


A 
e] 


. Rubiaceae?. Pollen types 1, 2, Graham (1989: 
63, figs. 45, 46), Panama. MO. A 


ÉS 
ie) 


. TRubiacites asperuloides? Weber. x Wes 
sel and Weber (1855: 149, pl. 26, fig. 12), 
Germany. Repository unknown. P (as family). 


D 
© 


. TRubiacites ixoreoides 
(1918: 43, 44, pl. 18, figs. 
USNM. P (as family). 


Berry. Fruit, Berry 
9-12). Panama. 


g 
Iz 


. TRuboides lignita Perkins. Fruit, Perkins 
(1905: 193, pl. 78, figs. 80, 84). Vermont, 
U.S.A. VGS (specimens not at VGS, location 
unknown). The Brandon Lignite was later studied 
y Traverse D , 1994; d re i 
Traverse [1 as Rubiaceae’, an 
Not vica ??Rubiaceae??) ad pons 
(1977 et seq.; fruits, seeds; see also Tiffney & 
Traverse, 1994; R P (fruit 
and pollen as family). 


ubiaceae not listed). 


1988: 1456, figs. 
MO. A 


52. Sabicea tasperifolia Engelh. Leaf, Engelhardt 
(1895: 40-41, pl. 5, fig. 6; pl. 8, fig. 6), 
Colombia. RN-D. P 


53. Sabicea? telliptica* Engelh. (1891: 657, pl. 5, 
figs. 5, 7). Leaf, Chile. P; Berry (1922: 86, listed 
only), Chile. RN-D (Engelhardt material). 


54. Scyphiphora hydrophyllacea Gaertn. type 
pollen. Leopold (1969: 1149, pl. 34, figs. 15, 16), 
Marshall Islands. USGS. A (fide Muller, 1981). 


55. cf. Timonius DC. Pollen, Leopold (1969: 1176, 
pl. 310, figs. 21-23), Marshall Islands. USGS. A 
(fide Muller, 1981). 


56. ¡Triporopollenites bellus Stover & Partridge. 
ollen, Martin (1978: 191, figs. 7r, s), Australia, 
similar to Randia chartacea. UNSWK. A 


57. tTriporotetradites letouzeyi Sal.-Cheb. Pol- 
len, Salard-Cheboldaeff (1978: 253, pl. 8, fig. 1), 
Cameroon. Salard-Cheboldaeff (1978) compares 
the fossil to Gardenia. UPMC. A (fide Muller, 
1981) 


58. tTriporotetradites sp. Pollen, MacPhail (1999: 
205). Early to Middle Miocene, Murray Basin, 
Australia. a ido compares the fossil to 
Gardeni 


PLIOCENE 


For the Pliocene, 16 fossil taxa representing 14 
form and modern genera and six species have been 
assigned or compared to the Rubiaceae. 


1. cf. Alibertia A. Rich. ex DC. Pollen, Graham 
1976: 813, figs. io. 197, 200) Veracruz, 
Mexico. MO. A 


2. Borreria G. Mey. Pollen, Graham (1976: 813, 
fig. 195), Veracruz, Mexico. MO. A 


3. tCanthiumidites reticulatus Khan. Pollen, 
Khan (1976: 766, fig. 29), Papua New Guinea. 
Khan (1976) compares the fossil to Canthium 
obovatum Klotzsch ex Eckl. & Zeyh. 


(as Canthium type fide Muller, 1981). 


4. Cephalanihus occidentalis L. Leaf, Hannibal 
(1911: 335, 339) California, U.S.A. SUPA 
(possibly consolidated with collections at the 
Museum, UC-Berkeley). P. 


Annals of the 
Missouri Botanical Garden 


5. Faramea Aubl. Pollen (triporate form), Graham 
(1976: 813, = 179, 180; Fig. 1D), Veracruz, 


Mexico. 


an 


. cf. Galium L. Pollen, Menke (1976: 65-66, text- 
fig. 4b, pl. 36, figs. 9-11), Ge qua Geolo- 
gisches Landesamt, PE Holstein. P (fide 
Muller, 1981, *Menke (1976) has ite. 
identified Galium iit from the Pliocene of 
northwest Germany.”). 


al 


. cf. TGuettardidites ivirensis Khan (see Eocene). 


s 


Hoffmannia  tboliviana Berry. Leaf, Berry 
(1939b: 63—64, pl. 4, fig. 11), Bolivia. USNM. P. 


Laugeria L. (see Terebraria Kuntze). 


9. Nertera Banks & Sol. ex Gaertn. Pollen, 
Mildenhall (1980: 215, 228); listed only; lower 
Plio ? New Zealand. See also Mildenhall and 
Croshie (1979) 


10. tPsilatricolpites coprosmoides (see Paleo- 
cene) 

11. Psychotria L. Pollen, Wijninga (1996: 152, pl. 
4, fig. 40), Colombia. UA. A 


12. Rubiaceae. Pollen, Wijninga (1996: 152, pl. 4, 
fig. 39), Colombia. UA. A 


13. Rubiaceae. Pollen, stephanocolpate, Graham 
(1976, figs. 237, 238), Veracruz, Mexico. MO. A 
. TRubiacites nummularioides Berry. Leaf, 


Berry (1917: 161, pl. 18, fig. 15; Singewald & 
Berry, 1922: 42, 111—112, pl. 7, fig. 7), Bolivia. 
USNM. P (as family). 


15. Sabicea Aubl. en, Graham and Dilcher 
(1998: 1430, fig. Pu Costa Rica. MO. A 


16. Terebraria Kuntze in Post & Kuntze. Pollen, 
Graham (1976: 813, figs. 187, 188), Veracruz, 
Mexico. MO. A (as Terebraria type; now Laugeria 
E 


TERTIARY (UNDIFFERENTIATED; MOST NEOGENE, 
PROBABLY MIOCENE) 


For the Tertiary undifferentiated, 14 fossil taxa 
representing 11 form and modern genera and 14 
species have been assigned or compared to the 

jaceae 


1-4. tCoprosmaephyllum angustifolium Deane. 
Leaf, Deane (1904: 213, a 20, figs. 4-6); 
Coprosmaephyllum  attenuatum Deane. 

Leaf, Deane (1904: 213, pl. p figs. 9, 10); 


Coprosmaephyllum minus Deane. Leaf, 
Deane (1904: 213, pl. 20, figs. 7, 8); Copros- 

hyllum ovatum Deane. Leaf, Deane 
(1904: 212, pl. 20, figs. 1-3), Australia. Dean 
(1904) compares the fossils to the extant 
Coprosma. GSV. A 


5. Coussarea  Tmembranacea Engelh. Leaf, 


Engelhardt (1891: 656, pl. 5, fig. 2). Chile. 


6. Gouatteria Tlenuinervis Engelh. Leaf, Engel- 
hardt (1891: 656-657, pl. 5, fig. 6b). Chile. RN- 


7. Guettarda fcookei Berry. Leaf, Berry (1921: 
125-126, pl. 21, figs. 5, 6), Dominican Republic. 
USNM. P. 


Hoffmannia tprotogaea Engelh. Leaf, Engel- 
hardt (1891: 657, pl. 5, fig. 1), Chile. RN-D. P 


9. Psychotria tgrandifolia Engelh. Leaf, Engel- 
hardt (1891: 656, pl. 11, fig. 4), Chile. RN-D. P 


10. tPsychoiriphyllum attenuatum Deane. Leaf, 
Deane (1900: 60, pl. 15, fig. 2), Australia. Deane 
(1900) compares the fossil to Psychotria. GSV. A 


11. Rondeletia goldmani Berry. Leaf, Berry 
(1937: 72, 79), Trinidad. USNM. P 


12. tRubiaceaecarpum muliicarpellare Menzel. 
Fruit, Menzel (1913: 10, pl. 1, figs. 20-24), 
Germany. Dresden. P (as family). 


13. uu dana linearis Hector. Leaf, Hector 
(18 . fide Andrews, 1970), New 
Du] puris unknown. NA. 


14. Sabicea? telliptica Engelh. Leaf, Engelhardt 
(1891: 657, pl. 5, figs. 5, 7), Chile. RN-D. P 


QUATERNARY 


A number of Rubiaceae have been reported for the 
Quaternary, mostly Late Glacial and Holocene, 
representing taxa currently growing in the ps of 
modern (post 1975) 


the fossil locality, in the p 
imarily for NUR UE l 


literature, and used pri 
reconstructions. An are plausible and designated *A" 
(Accepted). Among examples of these records includ- 
ed here are 13 genera and 10 species: 


1. Borreria G. Mey. (Cuba: Moncana Ferrera et al., 
1990-1991; Belize: 1990; P. 


Hansen, 


hiemstra, 1984; Brazil: Ledru et al., 2001; de 


Volume 96, Number 1 
2009 


Fossil Record of the Rubiaceae 


N 


a 


ul 


I 


ee 


© 


10. 


T 


14. 


15. 


. Borreria latifolia (Aubl. K. Schum. 


. Cephalanthus 


. Faramea occidentalis (L.) A. Rich. 


1. 


. Psychot 
ae 1997b). 


. Randia L. 


Oliveira et al., 1999; Behling, 1997b; Bolivia: 
Paduano et al., 2003) 


. Borreria anthospermoides DC. (Colombia: Hoog- 


hiemstra, 1984). 


orreria laevis (Lam.) Griseb. (Colombia: Hoog- 
hiemstra, 1984). 


(Brazil: 
Behling, 1997b). 


. Canthium Lam. (Borneo: Anderson & Muller, 
5) 


. Cephalanthus L. (Louisiana, U.S.A.: Delcourt € 


Delcourt, 1977). 


occidentalis L. (Tennessee, 


U.S.A.: Berry, 1924). 


. Faramea Aubl. (Panama: Bartlett & Barghoorn, 
1973) 


(Cuba: 
Moncada Ferrera et al., 1990-1991). 


Galium L. (California, U.S.A.: Potbury, 1932; 
Washington, D.C., U.S.A.: Berry, 1924; Baja 
California, Mexico: Lozano-Garcia et al., 2002; 
Chile: Latorre et al., "rum China, Late Pliocene 
to Quaternary: G. ers. comm., 20 
included in Ma 1980: 215, Late rdum 
nary, New Zealand). 


Galium californicum Hook. & Arn. (California, 
U.S.A.: Mason, 1934) 


- Galium palustre L. (Germany: D. Mai, pers. 


comm., 2002; Mania & Mai, 1969). 


- Galium stellatum Kellogg (Chihuahuan Desert, 


Mexico/U.S.A.: van Devender, 1990). 


Guetiarda calyptrata A. Rich. (Cuba: Mon- 
cada Ferrera et al., 1990—1991). 


Jackia Wall. in Roxb. (Borneo: Anderson & 
Muller, 1975). 
. Machaonia Bonpl. (Venezuela: Salgado-La- 


bouriau, 19 


. Psychotria L. (Brazil: Behling, 1997b). 


otria cf. alba Ruiz & Pav. (Brazil: 


(Panama: 


Bartlett & | Barghoorn, 
1973). 


. Relbunium (Endl.) Hook. f. (Brazil: Behling, 
1997b; Colombia: Hooghiemstra, 1984). 


21. Rubiaceae (Argentina: Heusser, 1995; Mancini, 
1998; Prieto, 2000; Quattrocchio & Borromei, 
1998; Brazil: van der Hammen & Absy, 1994; 
Behling, 1997a; de Oliveira et al., 1999; Chile: 
aduano et al., Heusser et al, 1999; 
i 1984; Costa Rica: 
Hooghiemstra et al., 1992; Islebe & Hooghiem- 
stra, 1997; Kesel, 1983; Guatemala: Islebe et al., 
1996; Mexico: Brown & Jacobs, 1988; Lozano- 
2002; Panama: Bartlett & Barg- 


Colombia: Hooghiemstra, 


Garcia et al., 
oorn 


N 
N 


. Timonius DC. (Borneo: Anderson & Muller, 


N 
w 


. Warszewiczia Klotzsch (Colombia: Hooghiem- 
stra, 1984). 


DISCUSSION 


Based on the above synopsis, the accepted and 
pending records are summarized in Table 1, and the 
currently accepted records are arranged according to 
subfamilies, age, locality, and present distribution of 
rom this 


the closest moder analogs in Table 
u 


mmary, some generalizations emerge, as well a 
places where additional information would be espe- 
cially useful. Regarding the origin of the family, the 
presently accepted fossils are too young to provide a 
clear indication. The oldest are cf. Emmenopterys from 
a Middle 

Washington, USA. (W 


= 
z 


Eocene Republic flora o 
Wehr & Manchester, 1996; 
subfamily Ixoroideae), and E. dilcheri from the 
44 Ma Middle Eocene Clarno flora of Oregon, 
U.S.A. (Manchester, 1994). The genus is presently 
found in the People's Republie of China, wh 
genera similar to western American Tertiary plant 
W. C. Cheng, 
Ailanthus Desf., Cercidium Tul. complex, c» osos 
Lesch. ex Blume, ea See & Zucc., 
Pterocarya Kunth; Graham, 1999: 200— E The next 
oldest fossils are from the Late Hosea: and include 
Canthium (Australia, MacPhail, 1999; present distri- 
bution Old World tropics; subfamily Ixoroideae}, 


ere other 


fossils occur (e.g., Metasequoia Hu & 


Faramea T Graham, 1985; present distribu- 
tion tropical America; subfamily Rubioideae), and 
Guet Di (Australia, MacPhail, 1999; p 

bution New Caledonia and tropical America 


resent distri- 
) in 
addition to the alternate-leaved +Paleorubiaceophyl- 
lum from the Middle Eocene of the southeastern 

S.A. Thus, in the Eocene, the family is represente 
three subf 


by four or five genera, from amilies as 


presently circumscribed, in North America, Central 


100 Annals of the 
Missouri Botanical Garden 


Table 1. Genera ja oe (A) and pending (P) Rubiaceae reported in the fossil record arranged according to 
ubfamilies and trib ilial classification follows that used in Dessein et al. (2005; see also Jalaluddin et al., 2008; 
Kahan et al., 2008; Martínez-Cabrera et al., 2008; Smedmark et al., 2008). 


Taxon Age Locality! Status 
Subfamily Rubioideae 
Coussareeae 
Coussarea. Tertiary Chile E 
Argentina P 
Faramea Late Eocene anama A 
Middle Oligocene Puerto Rico A 
arly Miocene Pa A 
Eocene Argentina P 
Pliocene Panama A 
Faramea (as Psilatriporites corstanjei) cene mbia A 
Middle Pliocene Veracruz, Mexico A 
Psychotrieae 
Psychotria Eocene Peru P 
Eocene KY/TN P 
P 
Pliocene Colombia A 
Tertiary ile P 
Psychotria (as Psychotriphyllum attenuatum) Tertiary Australia A 
Morindeae 
orinda Miocene Marshall Islands A 
Spermacoceae 
orreria Pliocene Veracruz, Mexico A 
Miocene Senegal A 
Anthospermeae 
oprosma. Eocene Argentina P 
Coprosma (as Coprosmaephyllum ere etal.) Tertiary Australia A 
Coprosma (as Psilatricolpites coprosmoi aleocene ile P 
Coprosma—Opercularia (as 11 ae Oligocene to Pleistocene New Zealand A 
zelandiae 
Coprosma type Miocene Marshall Islands BE 
Rubieae 
ium Paleocene Greenland P 
Galium Miocene AK A 
cf. Galium Pliocene Germany P 
Rubia Miocene Spain A 
Subfamily Ixoroideae 
ineeae 
ondaminea Miocene Colombia, Peru E 
Elaeagia (as Elaeagnites campanulatus) Miocene Spitsbergen P 
Emmenopterys Middle Eocene A 
inckneya Oligocene OR A 
Sabiceeae 
Sabicea Miocene anama A 
Pliocene Costa Rica A 
(as Sabicea? elliptica) Tertiary ile E 
(as Sabicea asperifolia) Miocene Colombia E 
Ixoreae 
xora Miocene Marshall Islands A 
(as Ixorophyllum anceps) ocene omeo P 
Scyphiphora. Miocene Marshall Islands A 
Vanguerieae 
Canthium (as Rubipollis oblatus) Late Eocene to Early Pliocene Australia A 
Miocene roatia P 
Canthium (as Canthiumidites reticulatus) Pliocene Papua New Guinea A 


Volume 96, Number 1 Graham 101 
2009 Fossil Record of the Rubiaceae 
Table 1. Continued. 
Taxon Age Locality" Status 
Gardenieae 
cf. Alibertia Pliocene Veracruz, Mexico A 
Gardenia type (as Triporotetradites nachterstediensis) Late Eocene Germany P 
Gardenia Miocene Borneo A 
Miocene Marshall Islands A 
(as Gardeniapites iaiwanensis) Miocene Taiwan A 
(as Triporotetradites letouzeyi) Miocene Cameroon A 
(as Triporotetradites sp.) Early to Middle Miocene Australia A 
oe (as Triporotetradites hoekeni) Oligocene, Miocene Cameroon A 
ia (as ea bolienhagenii Miocene Cameroon A 
Subfamily Cinchonoidea 
Cinchoneae 
Cinchona (as Cinchonidium ovale) Paleocene ND P 
Cinchona (as Cinchonidium copeanum) Oligocene P 
Cinchona Cinch anadai um iu 1 Croatia P 
Portlandia-E a—Cat Chi ae (PECC) 
clade 
Chiococca Miocene CA A 
Exostema Middle Eocene TN/KY P 
(as Exostema precaribaeum) Miocene uba P 
mijia Eocene Argentina P 
Naucleeae 
Cephalanihus Eocene Argentina P 
Pliocene F 
Mitragyna (as Retitricolporites annulatus) Oligocene Cameroon A 
Mitragyna (as eye n Oligocene, Miocene Algeria P 
Hilleae-Hameliea 
osmibuena Miocene Panama A 
Hoffmannia Eocene, Tertiary Argentina, Chile P 
Pliocene Bolivia P 
Rondeletieae 
ondeletia Miocene Oaxaca, Mexico P 
Miocene Panama P 
Eocene Argentina P 
Tertiary Trinidad P 
Guettardeae 
Guettarda ocene N P 
(as cf. Guettardidites) Oligocene Puerto Rico P 
Eocene to Pliocene Australia A 
Oligocene Puerto Ric P 
cf. Guettarda Miocene Marshall Islands A 
Tertiary Dominican Republic P 
Miocene aiti P 
Miocene Oaxaca, Mexico P 
Terebraria (Laugeria; as Terebraria type) Pliocene Veracruz, Mexico P 
cf. Randia Miocene a Islands A 
andia loc eala A 
Randia (as Canthiumidites bellus) Early to Middle Miocene b A 
Randia (as Triporopollenites bellus) loc ustralia A 
Randia Miocene P 
Randia Miocene Marshall Islands A 
cf. Timonius Miocene Marshall Islands A 
Uncertain position/incertae sedis 
Bothriospora Miocene Colombia P 
Endlicheria (Lauraceae?) Miocene Colombia, Ecuador P 
Gouatteria Tertiary Chile P 
cf. Mussaenda Miocene Marshall Islands A 


102 


Annals of the 
Missouri Botanical Garden 


Table 1. Continued. 
Taxon Age Locality! Status 
Posoqueria type Miocene na A 
Colombia P 
Ecuador P 
anis ia Pliocene Veracruz, Mexico A 
ragoga tertiaria (subfam.?) Eocene Peru P 
Mu (family onl 
homelia Miocene Panama A 
ce Panama A 
Miocene Panama A 
Pliocene A 
Plioc Veracruz, Mexico A 
as wid No ud d uae locene ombi. P 
s Rubiacites asperuloi Miocene Germany P 
as Rubiacites chometifolia Eocene Argentina P 
as Rubiacites ixoreoide. locene na P 
as Rubiacites nummularioides Pliocene Bolivia P 
as Rubiacites? RM oce P 
as Rubiacites sphericus Eocene TN/KY P 
as Rubiacites wilcoxensis) Eocene TN/KY P 
as Rubiaceocarpum markgrafi) oce pt P 
as Rubiaceaecarpum multicarpellare) rti Germany P 
as Rubioxylon naucleoides Oligocene Austria P 
as Ruboides lignita loc P 
Subfamily assignment uncertain 
Paleorubiaceophyllum eocenicum Eocene TN/KY A 


! Abbreviations for states (U.S.A.) are: AK, Alaska; CA, California; KY, Kentucky; MS, Mississippi; ND, North Dakota; NV, 
g 3 


Nevada; OR, Oregon 
America, and Australia. This representation and 
distribution indicate an earlier Late Cretaceous or 
asize the need to: ( 

leocene material, 


ER 
— 


megafossil deposits, such as the Early Cretaceous 
(Aptian) Anfiteatro de Ticó flora of Argentina, the 
Aptian—Albian Crato flora of Brazil, the late Middle 
Cretaceous (Turonian) flora of New Jersey, the Late 
Cretaceous d flora of Sweden, e. elsewhere 
for plants of 


After the Cretaceous and Mies de numbers of 


rubiaceous-complex affin 
accepted genera (form generic and modern names, 
including cf. Mii s in subs e e are 
Eocene: four or five ocene: six, Miocene: 20, and 
Pliocene: seven. Even lo for some “pin 
of form- "genere and modern names (fPalaeocopros- 


Tri, domes nd 
Cumt dex Chula: and 
the different number and size of the floras studied 


madites, +P: 


oprosma; 


letouzeyi-Gardenia; 


for the various intervals, the figures suggest the 
Miocene (23.8—5.3 Ma) as a time of major diversifi- 
cation of the Rubiaceae. 

The radiation of the family based on accepted 
records i 


parallels the pattern of diversification. 


In the Eocene, fossils of Rubiaceae are known from 


; TN, Tennessee; VT, Vermont; WA, Washington 


three regions: North America north of Mexico 
(Washington, Oregon, U Emmenopterys; 
possibly the bu US.A eee 
lum); Mexico—Central diners ese on 


(Panama, F a and the Southeast Pacific Asia 
region (Australia, Canthium, Guettarda). In the 
Oligocene, North 
(Oregon, Pinckneya); 
Mexico—Central America—Caribbean region (the An- 
tilles, Faramea); the Southeast Pacific—Asia mica 
(Australia, Guettarda; New Zealand, , Co- 
prosma—Opercularia); and Africa (Cameroon. | Maco 
T Mitragyna}. By the Miocene, accepted reports 
f Rubiaceae include six E (with nine subre- 
RN North America (California, U.S.A., Chiococca; 
Alaska, U.S.A., Galium), Mexico-Central America— 
Caribbean region (Panama, Chomelia, Cosmibuena, 
Faramea, Posqueria type, Sabicea), South America 
(Colombia, Faramea), Southeast Pacifie-Asia region 
(Australia, Gardenia, Guettarda, Randia; New Zeal- 
and, Randia; Borneo, Gardenia; Marshall ya 
Gardenia, Falsch Ixora L., Mori 
saenda, cf. Randia, cf. Loris Taiwan, DONI 
Africa (Cameroon, Macrosphyra, Mitragyna, Morelia, 
Gardenia; Senegal, Borreria), and Europe (Spain, 
Rubia). 


Volume 96, Number 1 


Graham 


103 


2009 Fossil Record of the Rubiaceae 
Table 2. Summary of the “A” (accepted) records of the Rubiaceae. 
Taxon Age Fossil occurrence! Modern distribution! 
Subfamily Rubioideae? 
Faramea Eocene m tropical America 
Middle Oligocene Puerto Rico 
Early Mioce Panam 
a Colombia 
liocene Veracruz, Mexico 
Coprosma ee to Pleistocene New Zealand Southeast Asia, Chile (Juan Fernandez 
Islands 
Morinda Miocene Marshall Islands tropical 
Borreria Miocene enegal warm regions of the world 
Pliocene Veracruz, Mexico 
Galium locene AK nearly cosmopolitan 
ubia locene pain warm regions of the world 
Psychotria Pliocene Colombia warm regions of the world 
Tertiary Australia 
Subfamily Ixoroideae? 
Emmenopterys Middle Eocene People's Republic of China, Burma, Thailand 
Canthium Late Eocene to Early Australia Old World tropics 
Pliocene 
Pliocene Papua/New Guinea 
Pinckne ligocen Southeastern U.S.A. 
Macrosphyra Oligocene to Miocen Ca tropical Africa 
Gardeni arly to Middle Miocene Australia Old World tropics 
loc orneo 
Miocene Marshall Islands 
Miocene Taiwan 
Miocene Cameroon 
Ixora Miocene Marshall Islands New and Old World tropics 
cf. Posoqueria Miocene Panama New World tropics 
Morelia iocene Cameroon tropical Afri 
Mussaenda Miocene Marshall Islands Old World tropics 
Sabicea iocene tropical America, Africa, Madagascar 
liocene Costa Rica 
Seyphiphora Miocene Marshall Islands Indomalaysian coasts, India, Sri Lanka, 
Australia 
Alibertia Pliocene Veracruz, Mexico tropical America, West Indies 
Subfamily Cinchonoideae* 
Chiococca cene FL; West Indies, tropical America 
arda een to Pliocene Australia New Caledonia, tropical America 
loc Marshall Islands 
Randia Miocene Australia New and Old World tropics (present-day 
Miocene New Zealand definition, tropical America 
locene Marshall Islands 
Mitragyna Oligocene Cameroon tropical Africa, Asia 
Cosmibuena locene nama tropical southern Mexico, Central and South 
America 
cf. Timonius Miocene Marshall Islands Southeast Asia, Pacific islands 


oo = states (U.S.A.) ar 
? Eight g ven 


e: AK, Alaska; CA, California; OR, Oregon; WA, Washington. 
accepted; idest: Middle(?) to Late Eocene (Fara: 


mea), Panama. 


E Dg gencia, “2 accepted; oldest: Middle Eocene (Emmenopterys), WA/OR; Late Eocene to Early Pliocene (Canthium), 


Australia. 


* Thirteen genera, six accepted; 32; oldest: Late Eocene to Pliocene (Guettarda), Australia. 


104 


Annals of the 
Missouri Botanical Garden 


The earliest accepted appearance of modern 
Rubiaceae in the Eocene and its extensive diversifi- 
cation and radiation in the Miocene involve several 
factors. One is the length of the different epochs 
within the Tertiary Period: Paleocene (~65-54.8 
Ma; 10.2 Ma duration), Oligocene a 7-23.8 Ma; 
9.9 Ma), and Pliocene (5.3-1.8 Ma; 3.5 Ma), com- 
pared to the Eocene (54.8-33.7 Ma; 211 Ma) and the 
Miocene (23.8-5.3 Ma; 18.5 Ma). However, during 
these longer epochs, fundamental changes took place 
in the climate and landscape at more specific times 
(Graham, 1999, 2008). In the Eocene, 
a change from the Paleocene-Eocene Ther- 

Maximum (PETM) or the Early Eocene Climatic 
ma (EECO), during which 


were as warm or warmer than in all of the Phanerozoic 


lt was a 


time temperatures 


(last 500 Ma), to cooler temperatures in the Middle 
and Late Eocene, representing the transition from 
This 
interval near the Middle to Late Eocene boundary 


hothouse to eventual icehouse conditions. 


also corresponds to the time when an increasing 
number of plant fossils can be referred to modern 
genera; viz., a principal period in the modernization of 
the Earth’s vegetation. 

The time spanning the Miocene involved both 


+ 


climatic and physical changes in the landscape. A 
the beginning of the Miocene (23.8 Ma), Arctic glaciers 
were present but not extensive, whereas by the end of 
i Eden (5.3 Ma) and in the Mio-Pliocene, there were 

ntinental glaciers and sea ice that supplemented 


Antarctic glaciers in cooling the waters of the ocean 
basins. The effect, evident globally by the beginning of 
the Middle Miocene (—15 Ma), was another drop in 
temperature and a marked increase in seasonality. The 
vegelation responded with expanding versions of 
seasonally dry forests, caatingas, cerrado, steppe, 
savanna, grasslands, and near-deserts that would form 
communities modern in range and composition during 
the increasingly cold, dry intervals of the Pliocene and 
Quaternary. 

Climate changes in the Miocene were paralleled by 
pronounced alterations in the landscape. For example, 
the Transvolcanic Belt of Mexico and the Central Andes 
Mountains of South America (Gregory- Wodzicki, 2000; 
1998; Graham et al., 2001) 


their elevation, as did the Himalayas, 


Gregory-Wodzicki et al., 
attained most o 
during the 18 Ma interval of the Miocene. 

The evolution of character traits is difficult to 
reconstruct from the fossil record presently available. 

owever, the appearance of certain features can be 
documented by specifie points in time. The typical 
leaf arrangement in the Rubiaceae is opposite, with 
alternate leaves found in a few, more advanced genera 
(e.g., Didymochlamys, Sabicea, Theligonum). Howev- 
er, if tPaleorubiaceophyllum is correctly assigned to 


the family, as seems likely, this apomorphic character 
was present by the Middle Eocene (~45 Ma). 

Pollen polymorphism is often associated with 
heterostyly, usually expressed by differences in pollen 
size, exine thickness, or occasionally pollen sculpture 
patterns in the different peice lengths (e.g., Rudgea 


Faramea, however, distinctly diporate, triporate, and 
occasional tetraporate pollen occurs on the same 
reference slides of modern pollen. In some species, 
almost all the grains are triporate (e.g., F. talaman- 
carum Standl., Panama, Kirkbride et al. 496, MO) and 
in others they are mostly diporate (F. vaginata 
i alge po ior sca 
Panama, Davidso 

2r as mcd to o pollen variability or 


MO). if is pollen polar: 


sterility in Faramea, is associated with heterostyly, 
then the feature, or its early development, was already 
present by the Late Eocene as shown e diporate 
form in the Gatuncillo Formation of Panama. The 
presence of alternate leaves and pollen dimorphism in 


stribu- 


the Eocene, together with the diversity and di 
f e in that epoch, all suggest an earlier 


tion of Rubiacea 
origin for the family. 


CONCLUSION 


The 


assessments from the synopsis of all known reports. 


history of the Rubiaceae is based on 
Tables 1 (accepted and pending reports) and 2 


ore realistic 


family than implied by the older literature, and reveal 
stages in the evolution of the group after its origin. 
The oldest verified occurrences are in the Eocene, yet 
the number of likely taxa and their distribution, and 
the possible presence of apomorphic features such as 
alternate leaves and pollen polymorphism, suggest a 
pre-Eocene origin, probably in the Late Cretaceous or 
Early Paleocene. The climatic changes in the Miocene 
favoring the generation of novel genotypes and 
phenotypes, and pronounced landscape evolution 
providing an increased diversity of habitats, are 
compatible with the fossil record and further suggest 
the Miocene as a time of significant diversification 
and radiation of the family. 


Literature Cited 
Anderson, J. A. R. & J. Muller. 1975. Palynological study of 


Names of Fossil 


en 
b 1820— 1965. US. Geol. Surv. Bull. 1300: 1-354. 


Volume 96, Number 1 105 
2009 Fossil Record of the Rubiaceae 
Axelrod, D. I. 1940. The Mint Canyon flora of d Miocene plants from E ai South 


Race A preliminary statement. Amer. J. S 
—585. 
— ——— 1950. The Anaverde flora of southern California. 
Capo Inst. Wash. Publ. 590: 119-158. 
. Age and origin of Sonoran Desert vegetation. 
c Pap. Calif. Acad. Sci. 132: 1-74 
Baker, ü G. 1956. Pollen dimorphism in a the Rubiaceae. 
Evolution T RS 
Barreda, V. D dr HR de la dp San 
Julián en dd área de Playa La a (Provincia de Santa 
Cruz), oe de la cuenca re (pco 34: 


Bartlett, A. S. € E. S. Barghoorn. 1973. Phytogeographic 
history E the Isthmus of Panama during the past 12,000 
a history of vegetation, climate, and sea-level 
ham (editor), Vegetation 
Latin America. 


rdam 
Behling, H. 19972. Late Quaternary vegetation, climate, and 
fire hist: fth 


istory o: region from 


cane Parana ae (south Brazil). Rev. 

sae le 97: 109-121. 
97b. Late Quaternary vegetation, climate, and fire 

oe fen m tropical mountain region of Morro de 
Itapeva, P s Palaeogeogr. Palaeoclimatol. Palaeo- 
ecol. 129: 

Berry, E. W. dele The we Eocene floras of southeastern 
North America. U.S. Geol. Surv. Prof. Paper 91: 1481. 

. 1917. Fossil plants from Bolivia and their Ru 

of the Eastern Andes. Proc. U. S. N 


— ———. 1918. The fossil higher plants from the Canal Zone. 
Pp. 15-44 in MU MAE i to hi Geology and Paleontol- 
ogy of the Canal Zone, Pan and Geologically Related 
Areas in Central rod un us West Indies. Ball U.S. 
Natl. irr perk 03. 

Mr A fossil plants from northern Peru. 
Proc. i Sud M s. 55: 279—294. 

———. 1921. i fossil plants from the Dominican 
Republi. Proc. U. S. Nat 7-127. 

———. 1922. 


ge to the Pal obcany of Peru 

Chile, . The Johns Hopkins University Studies in 

a aren re. 

. 1923a. mee fossil plants the Republic of 
Haiti. Proc. U. S. Natl. Mus. 62: 1-1 
—. 1923b. Miocene plants from ae Mexico. Proc. 
UL S. Natl. Mus. 62: 1-27. 

———. 1924. The s t deposit at the Walker Hotel 
site, Washington, Or pue remains other than 
diatom pe ddr h. "ur Se. 12-25. 

Miocene io from Patagonia. Johns 
E im Stud. Geol. 6: 183-251. 
929a. The fossil flora of the Loja Basin in southern 
ES Pp. 79-136 in Contributions to the Ps ntology 
of Colombia, Ecuador, d Peru, No. 10. The Johns 
uno University Studies in Geology, ae ore. 

— ——. 1929hb. Tertiary fruits and seeds from northwestern 
Peru. “Pp. 137-182 in Contributions to the Paleontology of 
Colombia, Ecuador, and Peru, No. 10. The Johns Hopkins 
University Studies in Geology, Baltimore 

. Revision of the lower acai Wilcox Flora of 

the southeastern United States, with d of new 

species chiefly from Tennessee and Kentucky. U.S. Geol. 
96. 


Surv. Prof. Paper 156: 1-1 


1936. 
^ rs der Torrey Bot. Club 63: 53-6 
. 1937. A late Tertiary flora uen Tii, B.W. I. 
Jolis m Univ. Stud. Geol. 1 
H Teniary flora e js 


pss Pichileufu, 


of the 

Yumurí River, Matanzas, Cuba. Johns Hopkins Univ. Stud. 
Geol. 13: 95-135. 
1 


939b. The fossil flora of Potosi, Bolivia. Johns 
Hopkins Univ Stud. Geol. 13: 9-67. 

—— 1. Additions to ihig Wilcox Flora from Kentucky 
and d U.S. Geol. Surv. Prof. Paper 193-E: 83—99. 

. 1945. Fossil floras from southern Ecuador. Johns 
E" be Univ. Stud. Geol. 14: 93-150. 

Blazer, A. M. 1975. Index of generic names d fossil plants, 
1966—1973. rr Geol. a Bull. 1396: 1-54. 

Brown, R. B. & B. F. Jacobs. 1988. Aná lige e interpretacion 
del polen de dos ae del occidente de México. Palynolog 
& en —99. 

Chaney, R. W. & E. I. Sanborn. 1933. The Goshen flora of 
west central os Carnegie Inst. Washington Publ. 439: 
1-103. 


Colinvaux, P. A. & E. K. Schofield. 1976. Historical ecology 
in the Galápagos Islands. I. Holocene pollen record from 
El Junco Lake, Isla San Cristobal. J. Ecol. 64: 989-1012. 


and. New Zealand Geol. 
Surv. To Bull. 22: 

- New DM. Mte and Cainozoic plant 
En New Zealand Geol. Surv. Paleontol. Bull. 32: 
1-8 


Deane, H O. Note on fossil leaves from the Tertiary 
deposits of Wingello and Bungonia. Rec. Geol. Surv. New 
South m 7 Th l: dee Se 


Cai ic fl f Sentinel 


Rock, a Coast. a of Mines, Records of the 
a forest species. Quatern. Res. 7: 218-237 


middle Sáo Francisco ud Palaeogeogr. Palaeoclimatol. 
Palaeoecol. 152: 319 

" Eo P. De Block, F. Lens, E. 
P. Schols, Smets, S. Vickier & S. 
m 05. Palynological characters p their 
phylogenetic signal in Rubiaceae. Bot. Rev : 354- 
414 


Dilcher, D. L. & T. A. Lott. 2005. A middle Eocene fossil 
plant assemblage (Powers Clay Pit) from western Tennes- 
see. Florida E Nat. Hist. Bull. 45: 143. 

oe 1915. Étude palynologique de 

d'Arauco-Concepción (Chili). 

Revista Esp. Mieloma 1: 549-565. 

Engelhardt, H. 1891. Ueber Teniärpflanzon von Chile. Abh. 
a Naturf. Ges. 16: 629-692. 

Über neue pue Vs Süd-Amerikas. 
AT > d Naturf. Ges. 19: 1-47. 

Ettingshausen, C. 1883. Beitráge zur Kenntniss der Tertiár- 
flora Australiens. Kaiserl. Akad. Wiss. Wien, Math.- 
Naturwiss. Kl, Denkschr. 47: 101-148. 

1888. Contributions to the Tertiary flora of 

Australia: Geol. Surv. New South Wales, Men. Paleontol. 


106 


Annals of the 
Missouri Botanical Garden 


Fritsch, A. 1893. Studien in Gebiete der Bóhmisschen der 
kredeformation—Palaeontologische ee der 
einzelnen. Arch. Naturwiss. Landesdurchf. Bohmen 9: 
-134 

Gevin, P., J.-C. Koeniguer & Y. Lemoigne. 1971. Les bois 
fossils di Dalaat el Admia (région de Tindouf, Algérie). 
Bull. Soc. Geol. France, 7th ser. 1 3. 

7. Über fossile Pflanzen von Labuan. Vega- 

4: 475-507. 

Graham, A. 1976. Studies in neotropical paleobotany. II. The 
Miocene v e er of Verac Ann. Missouri 

84: 


Bot. d 63: 


ruz, Mexico. 


pom in neotropical paleobotany. IV. The 
nd communities of Panama. Ann. Missouri Bot. Gard. 
534. 


72: 504— 


. 1987. Fossil pollen of Sabicea cp an 2 the 
lower Miocene Culebra Formation of Pan Ann. 
Missouri Bot. Gard. 74: 868-870. 
1988. Studies in aa nm bs 
f 


The "owes Miocene communitie —The 
Culebra Formation. Ann. Missouri B Ca ard. 75: 
1440- 


9. Studies in neotropical paleobotany. VII. The 
unities of Panama—The La Boca 


la. Studies in sa ier ees IX. The 
eee communities of Panam perms (dicots). 


. X. The 


: pecus in duin paleobotany 
Pliocene 


communities of P. sition, numeri- 
cal Yd ations, and diss sonny paleoenviron- 
mental reconstructions. Ann. Missouri Bot. Gar 
465-475, 


1999. Late Cretaceous and Cenozoic History of 
North American Vegetation (north of Mexico). Oxford 
University Press, Oxford, U.K 
———. 2009. Late Ga us and Cenozoic History of 
Latin American Vegetation and Terrestrial Environments. 
Missouri Botanical Garden Press, St. Louis (in press). 
& D. L. n er. 1998. Studies in neotropical 
nd RA s d i the 
For C 


Pliocene Río 
h 


of Costa and the Neoge 
Hih of e je D Bot. 85: 1426-1438 
. Jarzen. 196 tudies in neotropical 


hdledbouns. I: “The Oligocene communities 
Ric n . Missouri Bot. Gard. 56: 308-357 


— ——, K. M. 
— in neotropical paleobotany. XV. A 
palynoflora from the Eastern Cordillera, Bolivia: Implica- 
tions for the uplift history of the Central Andes. Amer. J 
000. Relationships between leaf 
mate, Bolivia: Implications for esti- 
mating E from fossil floras. Paleobiology 26: 
—688. 


McIntosh & K. Velásquez. 1998. Climate and 
— platon of the late Miocene Jakokkota flora, 
Bolivian Altiplano. J. S. Amer Karih a 11: > os 
Gruas-Cavagnetto, C. 1977. P 
cées dans le Pon du Bassin Anglo-Parisien. 
Rev. Micropaleontol. 20: 72-76. 
. 1978. Etude IN de l’Eocéne du Bassin 


Anglo Parisien. Mém Géol. France, Nouv. Sér., 56, 
Mém Ms i 
Hannibal, A Pliocene flora from the Coast 


LE 
Ranges d bos Bull. Torrey Bot. Club 38: 329-342. 


Hansen, B. C. S. 1990. Pollen stratigraphy of Laguna de 


Hector, J. 1880. Fossil plants. Official Catalogos of the New 
E Court, International Exhibit, Sydney, 1879: 


47-5 
Heer, O. 1868 R Theil. Pp. 1—77 in Flora Fossilis 
Arctica, Vol. 1 toting nels Schulthess, Zurich. 
876. Betrá 


Pp. 1 
141 à in Flora Fossilis Arctica, Vol. 4, Pt. 1. J. Wurster & 
Co., ern 

— ———. 1883. Die Tertiáre Flora von Grönland: Die Fossil 
Flora Grönlands. Pp. in Flora Fossilis Arctica, 
Vol. 7. J. Wurster & Co., 

Heusser, C. J. 1995. Three on pollen diagrams 
from southern Patagonia and their palaeoecological impli- 
cations. Palaeogeogr. Palaeoclimatol. Palaeoecol. 118: 1-24 

. E. Heusser & T. V. Lowell. 1999. Paleoecology of 

the southern Chilean Lake District—Isla 


Grande de 


ommen von Pr. 
ae 92: 122-183. 

Hollick, 1928. Paleobotany of Porto Rico. Scientific 
Survey a "bs Rico and the Virgin Islands 7, Pt. 3. New 
York Academy of Sciences, New York. 

Holmgren, P. K. & N. H. Holmgren. 1998 [continuously 
updated]. Index Herbariorum: A global directory of public 
Botanical 
al Herbarium. <http://sweetgum.nybg.org/ 

ih/>, accessed 2 Octob 08. 
Hooghiemstra, H. 1984. Neuen odi and climatic history of 
the High Plain of Bogotá, Colombia: À continuous record 
68 


Upper Quaternary vegetation dynamics 
tology of the La Chonta > area (Cordillera de Talamanca, 
Costa Rica). J. Quatern. Sci T: 205-225 


vironments of Northw 


palaeo-Amazon River System ( 
dimid Palaeogeogr. Palaeoclimatol. Palasoecol. 112: 
38. 


994c. Fluvial palaeoenvironments i in the intracra- 


m— Amazonas Basin (earl ne-early middle 


Miocene, Colombia Palaeogeogr. Palaeoclimatol. Pa- 
laeoecol. 109: 1—5 
Huang, T.-C. 1978. NM cap of E II. 


Sin. 19: 

Hooghiemstra. d Veram n 
montane Costa m since the las 
16: 589- 

r, J. H. des & D. A. Hodell. 
1996. A A Baboon anes history of lowland Guate- 


climate history of 
glacial. Quatern. Sci. Rev. 
. Brei 


mala. Pp. 81-99 in G. A. Islebe (editor), Vegetation, 
Phytogeography, and Paleo-ecology of the Last 20,000 
o: e Central America. University of 


erdam 
Jalaluddin, S. & J. I Brühl 2008. Testing species limits in 
Rennellia Ve PENNE Rubioideae, Rubiaceae). 
Taxon 57: 43-52. 


Volume 96, Number 1 
2009 


107 
Fossil Record of the Rubiaceae 


Kahan, S. A., 5. G. Razafimandimbison, B. Bremer & S. 
Liede- Schuma ann. We Virectarieae 
(Rubiaceae, Ixoroidea or two tribes? New tri 

1 das of Sabiceeae and Boone 

hy of baro 3 Taxon 57: 7-23. 
3. Quaternary history of the Rio General 
Valley, ee Es Res. Rep. Natl. Natl. CN Soc. 15: 
339-358. 


Khan, A. M. 1976. ao of oe ieee cius 
w for Species fr 


Floren Ágyptens. Bayer. Akad. Wiss., Math.-Naturwiss. 
Kl. 140 

r Kenntnis x disperser Tetra- 
den ite ME AU. dd B, Paláobot. 3: 399-433 
A. Rylander J. Quade & O. 
m the arid prepuna 


Palaeogeogr. ER Palaeoecol. 194: 223-246. 


Ledru, M.-P., ampello Cordeiro R., J. M. Landim 
Deminguer, L. Mar Ti, p- Mourguiart, A. Sifeddine & B. 
Tureq. 2001. Late- Sd cooling in Amazonia inferred 


from pollen at Lagoa do Cacó, northern Brazil. Quatern. 
Res. 55: pae 
Leopold, E. 
Eniwetok Fem po shall Islands. 
Paper ae 1133-1185. 
1883. Me eae of Miocene species from 
in the so-called Bad-lands of Dakota. 
See to the Fossil Flora of the Western Territories, 
LUS: d 2 = 8: 221-237. 
Louvet, D SJ flore Oligocene du Djebel 
rg is ee ce Nat. Soc. Savantes 3: 
96. 


U.S. Geol. Surv. Prof. 


Lozano-García, M. S., B. Or is Guerrero & S. Sosa-Nájera. 
Mid- to late Wisconsin pollen record of San Felipe 
Basin, Baja eee Omir Res. 58: 84-92 
MacPhail, M. 9. Palynostratigraphy of ihe Murray 
Basin, iin mea ui Australia. Palynology 23: 
97-240. 
——— € D. J. Cantrill. 2006. Age and implications of the 
Forest Bed, Falkland Islands, southwest Atlantic Ocean: 
Evidence from fossil pollen and spores. Palaeogeogr. 
Plan Palaeoecol. 240: 602-629. 
R. 1994. Fruits and seeds of the middle 
Clarno Oregon 
Palaeontogr. a 58: 1-205. 
V. 1998. Vegetational changes during the 
Holocene in extra-Andean Patagonia, Santa Cruz Prov- 
ince, Argentina. Palaeogeogr. Palaeoclimatol. Palaeoecol. 
138: a 
Mandaokar, B. D. 2003. Palynology and palaeoecological 
€ of Middle Bhuban Formation = 
Lawngtlai, Mizoram, India. Gondwana Geol. Mag. sp. Vol 
193. 


Formation, 


6: pan 
Mania, D. & D. H c at eee Mollusken und 
P este ttelpleis m Geiseltals 


(südlich von Halle). pores 18: 6 
Martin, H. A. 1978. Evolution of e ee flora and 
vegetation through the Tertiary: Evidence from pollen. 
Alcheringa 2: 181-202. 
Martinez-Cabrera, D., T. Terrazas, H. Flores & H. Ochoto- 
na. 2008. M atole anatomy, and taxonomic position 
e! d i Brandegee (Rubi e a monospecif- 
s endemic to Mesoamerica. Taxon 57: 33-42. 


Miocene pollen and spore flora of 


4. Pleistocene flora of the Tomales 
Washington Publ. 415: 81- 


Mason, H. L. 193 
Formation. Carnegie Inst. 
179. 


Médus, J. 1975. Palynologie de sédiments Tertiaires du 
Sénégal par Pollen & Spores 17: 545-608. 
Menke, B. . Pliozáne und áltestquartáre Sporen- und 


Palco y von Schleswig-Holstein. Geol. Jahrb., A 32: 
3-1 


und 3. Beitrag zur Flora der niederrheinischen 

pud ier iba res Preuss. Geol. Landesanst. 34: 
8. 

Meyer, H. W. & S. R. E 1997. The Oligocene 

n Creek flora of the John Day Formation, Oregon. 
v. Calif. Publ. Geol. Sci. 141: 1 "195. 

Mildenhall, D. C. 1980. New Zealand Late Cretaceous and 
Cenozoic plant biogeography: A une Palaeogeogr. 
Palaeoclimatol. Palaeoecol. 31: 197—233. 

— & Y. M. Crosbie. 1979. = porate pollen from E 
per Tertiary of New Zealand. New Zealand J. Geo 
ris 22: 499-508. 

Moncada Ferrera, M., C. E. Hernández Fuentes & M 

Cabrera Castellanos, 1990-1991. Anális polínico de 
dim del p de la Isla de la Juventud 
Cuba). y» ta Bot. Hung. 36: 1 

Muller, J. 1981. Fossil pollen dum " extant angiosperms. 
Bot. Rev. 47: 1-142 

Paduano, C. M., M. B. Bus h, P. A. Baker, S. C. Fritz € G. O. 
Seltzer. 2003 history of Lake 

itic acial Maximum. Palaeogeogr. 
Palaeoclimatol. ee 194: 259-279. 
Palma- Heldt, S. 1 


marin 


0. Contribucién al conocimiento palino- 
b 


Wien (Tertiaire de la 


Boca. SS. 
Mateo Co., ue Carnegie Inst. Wash. Publ. 415: 
—44. 

n ALR. 2000. _Vegetational history bs Late Glacial- 
m o 
Palaeogeogr. Palaeoclimatol. nr n 167-188. 

«SA. rromei. 1998. Paleovegetational 

atern: ry in 


Province and southern Tierra 
del Fuego a Palynology 22: 67-82. 

Romero, E. J. 1986. Paleogene phytogeography Ms 
climatology of E pou Ann. Missouri Bot. Gar 
73: 449—461. 


Ss 
a 


r. . Dilcher. 1979. Investigations of 
angiosperis from "e fiseedo of North America: Stipulate 
se of the Rubiaceae including a probable polyploid 

pulation. Amer. J. Bot. 66: 1194—1207. 

Salad. Cheboldaeff, M. 1978. Sur la palynoflore Maestrich- 
tienne et Tertiaire du basin sédimentaire littoral du 
Cameroun. Pollen & Spores 20: 215-260 

Salgado-Labouriau, M. L. 1980. A pollen diagram of the 
Pleistocene-Holocene boundary of Lake Vale 
zuela. Rev. Palaeobot. Palynol. 30: 297-312 


ncia, Vene- 


108 


Annals of the 
Missouri Botanical Garden 


Scholtz, A. 1985. The palynology of the upper lacustrine 
sediments of the e t Pipe, Banke, Namaqualand. Ann 
S. Afr. Mus. 95: 1-109. 


Singewald, J. T. Jr. & E. W. Berry. 1922. The geology of the 
Corocoro Copper o of Bolivia. Johns Hopkins Univ. 
Stud. Geol. 1: 

Msn J-E; E, Ki Bydin, » G. jon eae S: 

Khan, S. ide Schumann & B 008. A 
A of o. lleae o hal on rpsló 

intron data. Taxon 57: 24-32. 

s H. 1977. Fruits and seeds of ae Brandon Lignite: 

noliaceae. Bot. J. Linn. Soc. 75: 299-323. 

A. Traverse. 1994. "The Ne Lignite ee 
is of A not Cretaceous, age! E. Geol. 
215-22 

Traverse, E 1955. Pollen analysis of the Brandon Lignite of 
dee US. we Interior, Washington, D.C., Bureau of 
Min m Invest. 5151. 


moe i of the Brandon 
Lignite É Vermont, USA. . Palaeobot. Palynol. 82: 
265-2 


Unger, 4 0 Genera et m Plantarum Fossilium. 8 
vols. Wilhelm Braumüller, Vienna. 

. 1865. oge Plantarum Fossilium—Sammlun, 

Fossiler Pflanzen besonders 


der Tertiar-Formation. 
6 

van Campo, E. 

L. Asby. 1994. Amazonia during 


the last glacial: Pálssogeogr Palaeoclimatol. Nene 
109: 247—261. 


van Devender, T. R. 1990. Late Quatemary vegetation and 
ens m a 
e R 


Last 40,000 nd of Biotic Change: University of Arizona 

Press, Tuc 

van Hoeken- “Klinkenberg P. M. J. 1964. A palynological 
investigation of some Upper-Cretaceous sediments in 
Nigeria. Pollen & CR 6: 209-231. 

Watt, A. D. 1982. Index of Generic Names el Fossil Plants, 
1974-1978. He 2 s Bull. 1517: 1-63. 

Wehr, W. C. & S. nchester. 1996. ps PR 
significance of a cene rien fruits, p seeds from 
Republic, Pes rr Geol. 24: 25-27. 

Wessel, P. se O. Weber n. ES 
Tertiárflora 
E E 4: 111-16 
hi 


euer oe zur 


8. 
te, J. M. & T. A. Ager. 1994. Palynology, paleoclimatology, 
and correlation of middle Miocene beds from Porcupine 
River uar RS a Alaska. md Int. 22/23: 43-77. 
ijning 996. Paleobot: and Palynology of 


th, B 


mi Aen 
e plant diversity at 


, M. 2000. Palinoflora y ambiente en el Terciario del 
nordeste de Tierra del Fuego, Argentina. Revista Mus. 
Argent. Ci. Nat., n.s. 2: 43-51. 


PHYLOGENY OF THE 
HERBACEOUS TRIBE 
SPERMACOCEAE (RUBIACEAE) 
BASED ON PLASTID DNA DATA! 


Inge Groeninckx,? Steven Dessein,?? 
Helga Ochoterena,* Claes Persson,? 
Timothy J. Motley," Jesper Kárehed,* 
Birgitta Bremer,” Suzy Huysmans,” 
and Erik Smets?? 


ABSTRACT 


n its current circumscription, the herbaceous tribe Spermacoceae s.l. (Rubiaceae, Rubioideae) unites the former tribes 


Spermacoceae s. str., 
Mi 


within the tribe. 


anettieae, and the Hedyotis-Oldenlandia group. Within Sperm 
Oldenlandia De the generic delimitations are problematic. Up until now, gae aa aer have fi 
This stud relat 


coceae, and particularly within the 
ocused 
hin Spermacoceae 


tionships wi 


specific taxonomic problem study is t 


iun a Mess e Seen af three plastid markers (atpB-rbcL, rps] (o 2 irnL- pub were es d as 


well as combined using parsimony and Bayesian approaches. 
str. fori 


Mu e lei c. The former tribe Spermacoceae s. 


Schltdl), or polypliyleti (Hedyotis L. and Oldenl 
xu to iw. support t for the many n 
the 


cla 
e major lingage of Spermacoceae that can be u 


t the expanded tribe Spermacoceae 
sted within the Hedyotis Oldenlandia 


pa de . Morpho logical investigations of the taxa are 


des and relationships detected. This Judy pues a na 
and 


sed to 


ey wor 
trnL-trn F. 
The systematic relationships of the Rubiaceae paper as Spermacoceae s. str., are characterized by 


herbaceous representatives are still unclear at the 
species and genus levels (Robbrecht & Manen, 
2006). Even the higher-level classification in tribes 
has been the subject of the last 
comprehensive classification based on morphology 


(Robbrecht, 1988, 1993), most herbaceous repre- 


ate. 


sentatives were assigned to one of the following 
tribes: Anthospermeae, Argostemmateae, Coccocyp 
seleae, Hedyotideae, Knoxieae, Rubieae, 


ae as de delimit ed 
; 1873; rene kamp, 1 ; Verdcourt, 
1958; RoBbrecht, 1988, iem referred to in this 


the presence of raphides, fimbriate stipules, uni- 
ovulate locules, seeds with an apparent adaxial 
groove, and the frequent occurrence of pluriapertur- 
ate pollen grains. However, molecular data show 
as s. str. to be deeply nested within the 

Hedyotideae, making the latter tribe paraphyletic 
Bremer, 1996; Andersson & Rova, 1999; Bremer & 
anen, 2000; Dessein et al., 2005a). Therefore, 
Bremer (1996) and later Bene: and Manen (2000) 


we a wider definition for Eom in 


ze 


ormer tri 
oe Manett 
pideae are merged. 


ermacoceae s. 
ttieae, eo ur Triainole- 


1 We thank the XM of the Third International Rubiaceae Conference and the editors of this volume for the invitation 


to participate in the proce 


Leuven, Belgium. jon for corresp: 


macoceae. This research wa: 


aphs of Sp 
0250. 05 and c. 0268. 04). Inge Groeninckx holds a Ph.D. research grant from the Fund for 


nt Fou Katholieke Universiteit Leuven, Kasteelpark Arenberg 31, P.O. Box 2437, BE-3001 
ondence: inge.groeninckx@bio.kuleuve 


n.be. 


? National Botanic Garden of Belgium. Domein van Bouchout, BE- 1860 Meise, Belgium. 


4510, Mexic 
461, SE- 405 30 Göteborg, Sweden 


e 
5 Department of b E Sciences, Old Dominion e 110 Mills Godwin. Building, 45th Street, Norfolk, Virginia 


23529-0266, 


* Bergius Foundation, Royal Swedish Academy of Sciences and Botany Department, Stockholm University, SE-106 91 


s e Swe 


s of the Netherlands, Leiden University Branch, P.O. Box 9514, NL-2300 RA Leiden, The Netherlands. 


nal 
e 10. 3417/2006201 


ANN. Missouni Bor. Garp. 96: 109-132. PUBLISHED on 23 APRIL 2009. 


110 


Annals of the 
Missouri Botanical Garden 


Based on rpsi6 intron data, Andersson and Rova 
(1999) also found that Hedyotideae is paraphyletic 
relative to Spermacoceae s. str. not accept 
the wide delimitation for Spermacoceae as proposed 
by Bremer (1996), but suggested an emended tribe 
Knoxieae that included a few genera of Hedyotideae 
(i.e., Otiophora Zucc., Otomeria Benth., and Pentas 
Benth.) as a more prudent taxonomic approach to 
handle the information from molecular-based analy- 
ses. The latter view was followed by Dessein (2003), 
who preferred to M e an emended tribe Knoxieae 
(including Knoxieae s. . Triainolepideae, Otio- 
phora, the Pentas group E a obio fide Dessein 
et al. [2000], 
Spermacoceae (including Spermacoceae s. str., 
ettieae, and most of Hedyotideae). Robbrecht and 
Manen (2006), based on a supertree analysis of the 


and a din Juss.) as a sister group of 
M 


amily, came to a similar conclusion and likewise 
d Spe he 


recognized Knoxieae s.l. an rmacoceae s.l. 
d of the former tribe has also been confirmed 
a subsequent molecular study by Kárehed and 
eun (2 their taxonomic conspectus, 
Robbrecht sel. Manen (2006) listed 33 genera of 
Spermacoceae s.l. for which molecular sequence data 
are available. Based on morphological data, we 
recognize 31 of these 33 genera and consider that 
the tribe should include 30 additional genera; these 
are listed in Table 1. For each genus, the number of 
species, the distribution, and the position in Rob- 
brecht's classification of 1988 are given. 
Spermacoceae s.l. forms a primarily herbaceous 
lineage that is generally characterized by fimbriate 
SE and 4-merous flowers. Floral characters 
Fig. 1), 


and fruits, are highly 
variable. M di. i main groups ca 


n be 
identified within Spermacoceae s.l. The first, the 
Hedyotis-Oldenlandia group, is characterized by 
multiovulate locules and comprises the large genera 
Hedyotis L. and Oldenlandia L. and their presumed 
relatives. Most of these taxa were formerly placed in 
the tribe Hedyotideae. The generic delimitations of 
the Hedyotis-Oldenlandia | group 


subject of controversy for many years. The main issue 


ave been 


oO 


is whether most species of the complex should be 
lumped into Hedyotis (advocated by inter alia Merrill 
& Metcalf, 1946; Wagner et al., 1989; Fosberg & 
Sachet, 1991; Dutta & Deb, 2004) or whether many 
small genera should be recognized in addition to a 
ow circumscription of Hedyotis and Oldenlandia 
ee for African taxa by Bremekamp, 1952; fo: 
Neotropical taxa by Terrell et al., 1986; Terrell, 1991, 
2001a, b, c; and for Asian taxa by Terrell & Robinson, 


The second well-marked group within Spermaco- 
ceae s.l. is Spermacoceae s. str., which is character- 


zed by uniovulate locules. According to Dessein 
r the largest with an estimated 
275 species. Within ae s. str., controversy 
has focused on the delimitation of its nominal genus 
Spermacoce. The main question is whether Spermacoce 
should be limited to species with the same type of fruit 
dehiscence as S. tenuior L., the type species of the 
genus. In this species, fruits open asymmetrically, 
resulting in one closed and one open fruit part. If this 
narrow delimitation for Spermacoce (referred to as 
Spermacoce s. str.) is accepted, most other species in 
the tribe e s. str. must be included in 
Borreria G. M 

A third s ined group within Spermacoceae 
s.l. comprises only two American genera, Bouvardia 
Salisb. and Manettia Mutis ex L. Bremekamp (1952) 
considered Bouvardia closely related to Heterophyl- 
aea Hook. f. Hindsia Benth. ex Lindl., and 
Lecanosperma Rusby. Robbrecht (1988) placed these 
genera together with inter alia Manettia in a group 
with uncertain affinities, because their winged seeds 
suggest a relation to Cinchoneae, while the presence 
of raphides indicates a relation to Hedyotideae. In the 
classification of Bremer and Manen (2000), only 
ettia belon 


to Spermacoceae s.l., 
because Hindsia and Heterophyllaea (including Le 


Bouvardia and Man 
nosperma) are included in Coussareeae. ~ is 
similar to Bouvardia in many characters, but its 
winding shoots and corneous endosperm separate it 

which ect and has fleshy 


endosperm. These differences were the basis for 


from Bouvardia, 


Bremekamp (1934) to place Manettia in its own tribe, 
anettieae 
Until now, molecular studies within Spermacoceae 
s.l. have focused on particular taxonomic problems, 
such as t 


e circumscription and biogeography of 
Arcytophyllum Willd. ex hult 


Schult. & 
(Andersson et al., 2002), the generic status of 
Houstonia L. (Church, 2003), the delimitation of 
Pentanopsis Rendle, the affinities of Phylohydrax 
Puff (Thulin & Bremer, 2004), 
phocalyx B 


the taxonomic 
position of Gom, Baker Ter ein et al., 
2005a). In the present paper, we aim to present a 
phylogenetie hypothesis of Spermacoceae s.l. based 
on the analysis of three es markers (atpB-rbcL, 
rps16, and trnL-trnF) with the 
date. More specifically, we want to address the 
following questions: (1) Is Spermacoceae s.l. as 
Robbrecht and Manen (2006) 
monophyletic? (2) What are the relationships among 


oadest sampling to 


circumscri e y 


members of Spermacoceae s. str. and genera of the 
former tribes Hedyotideae and Manettieae? (3) What 
are the major clades within the Hedyotis—Oldenlan- 
dia group? 


Volume 96, Number 1 
2009 


Groeninckx et 111 


Phylogeny of IOE 


MATERIAL AND METHODS 
PLANT MATERIAL AND SAMPLING 


The aim was to obtain a broad sampling covering 
most of the geographic and taxonomic diversity of 
Spermacoceae and to enable identification of the 
principal clades within the tribe. We included a total 
of 128 species representing 32 of the 61 genera within 
Spermacoceae. Three taxa belonging to the Knoxieae 
(Batopedina pulvinellata Robbr., Carphalea madagas- 
Stapf ex 
Verde.) were chosen as outgroup following Robbrecht 
and Manen (2006) and (2007). 
For rpsi6 and trnl-trnF, we used 40 and seven 


cariensis Lam., and Pentanisia parviflora 


Karehed and Bremer 


previously published sequences, respectively (Anders- 
son & Rova, 1999; Andersson et al., 2002; Dessein et 
al, 2005a). Two hundred seventy-two sequences are 
newly generated (100 atpB-rbcL sequences, 67 rpsl 
sequences, 105 trnL-trnF sequences) using dried silica 
and herbarium material. Appendix 1 lists all taxa 
included in this study with voucher information and 
GenBank accession numbers. 


DNA EXTRACTION, POLYMERASE CHAIN REACTION 
AMPLIFICATION, AND SEQUENCING 


DNA was extracted from silica-dried and herbarium 
material using the CTAB method as described by 
Janssens et al. (2006). Amplification of the atpB-rbcL 
spacer was done with loce two and five as 
primers (Manen et al., . Specific id 
s ine could be d with ouchdow 
polymerase chain reaction (PCR) with two ls with 
an annealing temperature of 53°C, then 12 cycles with 
an annealing temperature of 52.5°C declining 0.5°C 
every cycle, followed by 16 cycles with an annealing 
temperature of 47°C. The rps16 intron was amplified 
with the rpsl6F and rpsl6R2 primers described by 
Oxelman et al. (1997). For the trn£-trnF intergenic 
spacer, we used the primers e and f of Taberlet et al. 
(1991). Both rps16 and trnL-trnF were amplified using 
standard PCR techniques with an annealing temper- 
ature of 55°C. The PCR reaction mixture was cleaned 
using a Nucleospin Extraction II Kit (Machery-Nagel, 
Dren, Germany) according to the manufacturer's 
a Sequencing was mostly done on an ABI 

O Genetic a (Applied E Lennik, 
c PCR products were sequenced by 
Macrogen (Seoul, South Korea) sequencing facilities. 


SEQUENCE ASSEMBLY, ALIGNMENT, AND GAP CODING 


The assembling and editing of sequences were 
conducted using the Staden Package (Staden et al., 


1998). Sequences were initially aligned with ClustalX 
(Thompson et al., 1997) applying the default parame- 
ters. Further NN of the preliminary AT. 

data matrices were done manually with MacClade 4.04. 
(Maddison & Maddison, 2001). Parsimonious dise 
tive gaps were coded manually according to the 
conservative simple indel coding method described 

y Simmons and rm (2000). 


PHYLOGENETIC ANALYSES 


Phylogenetic analyses were conducted using both 
parsimony (MP) and Bayesian inference (BD. The 
three plastid regions were first analyzed separately 
and then combine 

Equally ees MP analyses were performed 
using Nona 2.0 (Goloboff, 1993) launched through 
WinClada 1.00.08 (Nixon, 2002). Heuristic searches 
for the shortest trees were performed using the 
parsimony ratchet (Nixon, 1999). 
200 iterations each, holding one tree per iteration and 
randomly weighting 10% of the potentially informa- 
tive characters, were carried out until the most 
parsimonious trees (MPTs) were repeatedly found. A 
strict consensus tree was calculated using the trees 
obtained in the parsimony ratchet analyses. In order to 
evaluate the relative support of the clades, jackknife 
and bootstrap analyses were executed using 1000 
replicates with 10 initial trees holding five trees per 
random addition, doing tree bisection-reconnection 
(TBR) to hold 1000 trees, and calculating a consensus 
on each repetition. Frequency values were plotted 
onto the consensus of the 

For the BI analyses, a substitution model was 
selected for each DNA region with Modeltest 3.06 
(Posada & Crandall, 1998) under the Akaike 
Information Criterion (AIC). Modeltest selected the 
GT el of evolution for the atpB-rbcL spacer 
and the GTR+G model for the two remaining markers. 
Indels were not included in the BI analyses. In the 
combined analysis, a mixed-model approach was used 
(Ronquist & Huelsenbeck, 2003). The combined data 
were partitioned and the same models of evolution 
were T on s ie as selected for the single 
analyses. nalyses were conducted with 
MrBayes 3 bd & Ronquist, 2001). Four 
Markov chains (one cold, three heated) starting with a 
random tree were run simultaneously for one million 
generations, sampling trees at every 100 generations. 
The first 2500 ia trees (25%) were regarded as 
burn-in and discarded. PAUP* version 4b10 (Swof- 
ford, 2002) was used to calculate a 50% majority rule 
tree and to report the posterior probabilities for each 
clade. Only posterior probabilities above 
considered (Suzuki et al., 2002) 


112 Annals of the 
Missouri Botanical Garden 
Table 1. List of genera associated with Sy l., their distribution, and fi E tal. 


— 


2006), except when stat 


a in boldface were listed by 


ted otherwis Robbrecht and Manen (2006); other genera are here 
based on morphological similarities. Synonymous taxa are as given by Robbrecht (1988), except when stated otherwise. 


Robbrecht, 0. € 
Genus 1988 Native distribution species Sampled 
Agathisanthemum Klotzsch Hed tropical and S Africa, Comoros 4 yes 
Amphiasma Bremek. Hed tropical and S Africa 7 yes 
Anthospermopsis (K. Schum.) J. H. a Spe razil 1 no 
Arcytophyllum Willd. ex Schult. ed Mexico to W South America 17 yes 
Schult. f. 
Astiella Jovet Hed Madagascar 1 no 
Bouvardia Salish. Cin/Hed S U.S.A., Mexico to C America 42 yes 
Bradea* buen ex Brade Hed ce Brazil 5 no 
Carterella Ten Hed 1 no 
Conostomium sp Cufod. Hed Set to S Africa 5 yes 
Crusea Cham. & Schlidl. Spe Arizona, New Mexico, Mexico to C America 14 yes 
Dentella J. R. Forst & G. Forst. Hed tropical and subtropical Asia to SW Pacific 8 yes 
Diacrodon Sprag Spe Brazil 1 no 
Dürachiodonylus Bremek. Hed E Tropical Africa 1 yes 
Denscantia E. L = ral & Bacigalupo Spe E Brazil 4 no 
Diodella Small * Spe S U.S.A. to S America 16 yes 
iodia L. ' Spe S U.S.A. to S America 5 no 
Dolichometra K. Schum. Hed Tanzania 1 no 
Emmeorhiza Pohl ex Endl. Spe S tropical America and Trinidad 1 yes 
rnodea Sw. ? Spe Florida, Mexico to C America, Caribbean 4 yes 
Galianthe Griseb. © Spe S and C America 50 yes 
Gomphocalyx Baker Spe Madagascar 1 yes 
Hedyotis L. Hed tropical and subtropical Asia to NW Pacific ca. 115 — yes 
b. due Hed C a E ra Africa 2 yes 
oustonia L. Hed Am 20 yes 
Hydrph lax Spe India, ES pum pM 1 no 
Kadua Cham. © um (incl. Gouldia Spe Hawaiian Islands to S Pacific 28 yes 
A. Gray and d ee n e 
Kohautia Cham. Hed Africa, Madagascar, and Asia 3l yes 
ya Bremek. Hed tropical Africa 1 yes 
Leptomischus* Drake Hed ssam to China 7 no 
Leptoscela Hook Hed NE Bra 1 no 
ucya Hed Caribbean 1 no 
Manettia Mutis ex L. Cin/Hed tropical America 124 yes 
nostachya Bremek. Hed C and E tropical Africa 3 yes 
Micrasepalum Urb. Spe Caribbean 2 no 
itracarpus Zucc. ex Schult. Spe tropical America, naturalized elsewhere 58 yes 
chult. f. 
Mitrasacmopsis Jovet Hed C and E tropical Africa and Madagascar l yes 
Neanotis W. H. p Hed tropical and subtropical Asia 33 no 
Neohym n* Benne Cin/Hed E Himalaya, Tibet, SC China, N Indo-China 3 no 
a isa o Bremek. ed St. Helena 1 yes 
Nodocarpaea A. Gray Spe 1 no 
Oldenlandia L. (incl. Eionitis Hed pantropical ca. 240 yes 
Bremek., Exallage Bremek., 
and Thecorchus Bremek.) 
Oleo Terrell & W. H. Lewis Hed tropical and Eo America 1 no 
Pentanopsis Rendle Hed Ethiopia 2 yes 
Pent odon. Hoc hst. Hed tropical ij S y Arabian Pen., W 2 yes 
Indian Ocean, naturalized in dc 
Phyllocrater Wernham Hed rneo 1 no 
Phylohydrax Puff Spe coastal Tanzania to 5 Africa, Madagascar 2 yes 


Volume 96, Number 1 Groeninckx et 113 
2009 Phylogeny of C — 
Table 1. Continued. 
Robbrecht, No. of 
Genus 1988 Native distribution species Sampled 
Pleiocraterium Bremek. Hed tropical Asia 4 no 
Polyura* Hook. f. Hed E Himalaya to Assam 1 no 
Sie QUE NO Tennant Hed Tanzania 1 no 
Psyllocarpus Mart. & Zucc. Spe Brazil 9 no 
Richardia L Spe tropical and subtropical America, 16 yes 
naturalized elsewhere 
Sacosperma* G. Taylo Hed W and C tropical Africa 2 no 
Schwendenera K. Shu. Spe razil 1 o 
Spermacoce L. (incl. Borreria G. Mey. Spe pantropical 250-300 yes 
and Hemidiodia K. Schum.) © 

Staelia Cham. & Schltdl. Spe Mexico and S tropical America 14 no 
Stenaria (Raf. | Terrell Hed C and E U.S.A. to Mexico, Bahamas 5 yes 

tenotis Terrell Hed Mexico (Baja California) T no 
Stephanococcus Bremek. Hed WC tropical Africa 1 no 
Synaptantha Hoo Hed Australia 2 yes 
Tobagoa Urb. Spe Panama to Tobago 1 no 
Tortuella Urb. Spe fle de la Tortue (Haiti) 1 no 


Hed, Hedyotideae; Spe, Sea s. str.; Cin, Cinchon 
® = Bacigalupo & Cabral (1999); © = Negró 
Terrell et al. (2005); © = ae (1985); 


* Tentatively include 


RESULTS 


Sequence data from the ee atpB-rbcL, rps16, 
and trnl-F regions independently 
Individual 


plastid sequence analyses were ud con- 


were an 
and in a combined analysis “Table 2). 


gruent. Therefore, only the results from the MP and 
BI ads of the combined matrix are presented 
(Figs. 2—4). Compared to the 
a plastid sequence 


topologies of the in- 
analyses, the combined 
plastid trees show increased resolution and branch 
support. The Bayesian tree is somewhat better 
resolved than the consensus of the MP analysis, 

ut more resolved lineages have low posterior 
probabilities. 

Spermacoceae s.l., as delimited in the introduction 
(Table 1), form a well-supported monophyletic group 
(jackknife support [JS] = 100, bootstrap support [BS] 
— 100, posterior probability [PP] — 1), as can be seen 
in Figure 2. A highly supported pentamerous-flowered 
clade including Dentella J. R. Forst. & G. Forst. and 
Pentodon Hochst. (JS = 100, BS = 
resolved as sister to the rest of the tribe (Fig. 2). The 
remaining ingroup taxa are part of a clade that lacks 
significant jackknife and bootstrap me and has 
0.84). Within 


this clade, stars with Roman numerals I to III are 


only weak posterior probability (P 


assigned to the three deeper internal nodes that have 
reasonable support. These three clades are discussed 
in the following paragraphs. 


n-Ortiz & Hickey (1996); 
= Dessein (2003). 


© = Cabral (1991); “ = Terrell (1996); © = 


Clade I in Figure 2 JS = 88, BS = 77, PP = 1) 
includes a Kohautia subg. Kohautia Verde. clade 
sister to a clade that includes Pentanopsis and 2 
genera. This Pentanopsis clade 95, BS = P 
= 1) is similar to that proposed by Thulin and b. 
(2004). However, our larger sampling resulted in a 
broader circumscription adding Ce Old- 
enlandia affinis (Roem. & Schult.) D 
(L.) Roxb., and O. rosulata K. 
support the monophyly of both Amphiasma Bremek. 
(JS — 98, BS — 98, PP — 1) and Phylohydrax (J8 — 
93, BS = 95, PP = 1). 

In clade IT (JS = 88, BS = 83, PP = 1) of Figure 2, 


all Asian and Micronesian Hedyotis species, except H. 


O. herbacea 


m m results 


tenelliflora Blume, are part of a strongly supported 
Hedyotis s. str. clade (JS = 100, BS = 100, PP = 1), 
which is sister to a clade including Agathisanthemum 
Klotzsch and its allies. This clad 


Micronesian Hedyotis species 


e of Asian and 
includes H. 
fruticosa L., the type species of the genus. Relation- 
ships within this Hedyotis s. str. clade remain mostly 


also 


unresolved. Within the Agathisanthemum clade, 
Agathisanthemum is gis to Lelya osteocarpa 
Bremek. (JS = 100, BS = PP = 1), both sister to a 
lineage of African "pcd angolensis K. Schum. 
and O. goreensis (DC.) Summerh.) and North Amer- 
ican (O. uniflora L.) Oldenlandia species (JS = 100, 
BS = 99, PP = 1) 


In the MP consensus, clade II is sister to clade III 
(Figs. 2A, 3A). However, this sister relationship lacks 


114 


Annals of the 
Missouri Botanical Garden 


significant jackknife and bootstrap support and is not 
recovered in the BI (Figs. 2B, 3B). 

Within clade III (Figs. 3, 4), the earlier o 
lack significant support values in the 
consensus (Figs. 3A, 4A) and are collapsed in ele e 
(Figs. 3B, 4B). Therefore, PR between the 
different subclades of clade III shoul 

he 


with caution. In the following parag 


d be interpreted 
graphs, these 
subclades are discussed individually. 

In Figure 3, the Ria genus Dibrachiono- 
stylus Bremek. is er to clade of African 
Oldenlandia species P nins K. Schum., O 
geophila Bremek., and O. nervosa Hiern). However, 
this sister ono lacks significant jackknife and 
bootstrap support (Fig. is not supported b 
the BI (Fig. 3B 


with respect to Mitrasacmopsis Jovet and its allies also 


an y 
). The sister relationship of this clade 


lacks support. Mitrasacmopsis, another monospecific 
genus in the Hedyotis-Oldenlandia group, is never- 
theless highly supported as sister to Hedythyrsus 
Bremek. (JS = 99, BS = 97, PP = 1), and both are 
sister to O. fastigiata Bremek. (JS = 99, BS = 99, PP 
=1) 


The genus Kadua Cham. & Schltdl. 
Oldenlandia biflora L.) is resolved as monophyletic 
with moderate jackknife and bootstrap support but 
= 87, BS 


e Hawaiian Kadua species are 


(including 


maximum id pde probability (J 
= 86, PP = 1). T 
unresolved with Bre to the French Polynesian 
he genus Kadua shares a 
ll sampled 
Australian taxa (O. galioides (F. Muell.) F. Muell., O. 
mitrasacmoides F. Muell., and Synaptantha tillaeacea 
(F. Muell.) Hook. da the leo: Asian ak Hedyotis 
tenelliflora, and the African species O. lancifolia 
(Schumach.) DC. (JS = 91, BS = 86, PP 
The genus Arcytophyllum is strongly E as 
3, BS — 92, PP 


e of North and Central 
un species of Houstonia, Oldenlandia, and 


B PDA by our analysis (. 

= 1). It is sister to a clad 
Stenaria (Raf.) Terrell. The Houstonia species plus S. 
nigricans .) Terrell form one clade, although 
without significant support 

In Figure 4, Spermacoceae s. str. is nested within 
the S pr es group. In the 
(Fig. 4A), i 
lacking DUANE jackknife s d bootstrap 
support), while in the BI tree (Fig. 4B), Moledor: 
arborea (Roxb.) Bremek. is neste 
Spermacoceae s. str. clade (although with low PP = 
0.77). In both MP and BI analy 


MP consensus 
orms a monophyletic e (although 
support 


within the 


sis, Spermacoceae s. 

str ncertain relationships with respect to 
Arcytophyllum serpyllaceum (Schltdl. ) Terrell, Bouvar- 
dia, Manettia, Nesohedyotis (Hook. f) Bremek. 
(Fig. 4A), O. tenuis K. Schum., 


and O. salzmannii 


(DC.) Benth. € Hook. f. ex B. D. Jacks. Sister to this 
polytomy is a clade with species of Kohautia subg. 
Oldenlandia 
including the type species O. corymbosa L. (JS = 
99, BS — 98, PP — 
genus Kohautia Cham. & Schltdl. fa 


Pachystigma Bremek. species, 
1). Consequently, species of the 
in two well- 
supported, not closely related clades, which corre- 
spond to the two described subgenera: subgenus 
Kohautia (JS = 99, BS = 99, PP = 1) and subgenus 
Pachystigma (JS = 96, BS = 96, PP = 1). 


DISCUSSION 


ur analysis corroborates the monophyly of Sper- 
macoceae s.l. (Table 1), a mainly herbaceous assem- 
blage distributed pantropically, with only a few genera 
penetrating into more temperate regions. The mor- 
phological variation is considerable, but the fimbriate 
stipules and tetramerous flowers are shared by most 
species. There are no clear morphological ud 
morphies that separate Spermacoceae s.l. from its 
sister tribe, the emended Knoxieae. The pus 
differences are listed in Table 3 
ur analyses show several major evolutionary 
lineages within Spermacoceae s.l. and allow us to 
evaluate the monophyly of a number of genera. 
Several genera that have been Domin within the 
RE ndia group are supported her 
c (Amphiasma, Arcyto, phylum. Deniela, 
Kadua, E  Phylohydrax), while others appear to be 
paraphyletic (e.g., Agathisanthemum), biphyletic (Ko- 
aa or polyphyletic (Hedyotis and Oldenlandia 
sensu Bremekamp). These groups are discussed in the 
oe paragraphs. 


SPERMACOCEAE S. STR. 


In our analyses, Spermacoceae s. str. is nested 
within the Hedyotis-Oldenlandia group, which no 
longer makes it possible to pu eed this bie ata 
tribal level. Additionally, Sperm str. as 
delimited by Robbrecht (1988) is not Mud d as 
monophyletic. Both MP and BI analyses show that it is 
necessary to exclude Gomphocalyx and Phylohydrax 
oceae s. str. to be mono i 
in agreement with Thulin 
Dessein et al. (2005a). 

In the BI analyses, Nesohedyotis arborea, a species 


phyletic, which is 
(2004) and 


for Spermac 
and Bremer 


previously included in Hedyotideae, is placed within 


permacoceae str. sister to mmeorhiza 
umbellata (Spreng. ) K. Schum., but lacking significant 
0.67). This position of 


Nesohedyotis within Spermacoceae s. 


posterior probability (PP = 
str. was not 
recovered in the MP analysis. Because no morpho- 


logical characters can be found to support Nesohed- 


Volume 96, Number 1 Groeninckx et al. 115 
2009 Phylogeny of Spermacoceae 


dl 


e l. Floral diversity among species of Spermacoceae. —A. Kohautia microcala Bremek. —B. Hedythyrsus 
du (K. kA remek. —C. id frigidus (Willd. ex Roem. & Schult.) K. Schum. —D. Spermacoce debili 
enth. . Oldenlandia a (L.) Roxb. —F. Gomphocalyx herniarioides Baker. —G. Manostachya N E. 
Martins. E Oldenlandia pnm (Schumach.) Dé —I. Phylohydrax mádigostásienais (Willd. ex Roem. & Schult.) a 

—J. Manettia ~ Mu. Benth. —K. M ones Coe ar globosum (Hochst. ex A. Rich.) Klotzsch. ia "rusa 
goreensis (DC.) S erh. —M. Kohautia coccinea Royle. . Oldenlandia biflora L. —O. Kadua acuminata Cham. € 
Schltdl. —P. Oldenlandia m Pit. 


yotis as part of Spermacoceae s. str., we suggest that With the deeper nodes unresolved or only weakly 
the difference between the MP and BI e could idu the relationships d Spermacoceae s. 
be the result of data sampling artifacts (only rps16 w unclear and should be the subject of 
sequenced for N. arborea), which Seba He tori eae studies eal more taxa an 

the BI more than the MP analysis. or characters. Nevertheless, our analyses corroborate 


116 


Annals of the 
Missouri Botanical Garden 


Table 2. Characteristics of each data matrix and the corresponding tree statistics. 
No. of taxa No. of char. No. of PI char. No. of PI indels No. of MPT MPT length CI RI 
atpB-rbcL 100 1237 175 31 1949 399 0.55 0.84 
rps 105 105 191 20 1351 525 0.56 0.82 
irnL-irnF 107 1053 184 29 343 423 0.62 0.88 
Combined 128 2995 550 80 4782 1385 0.56 0.84 


characters; CL, consistency index (Kluge & Farris, 1969); MPT, most parsimonious tree(s); PI, potentially informative; 


Char, 
RI, retention index (Farris, 1989). 


the monophyly of most of the commonly accepted 
genera within Spermacoceae s. str., notably Crusea 
Cham. & Schltdl., Mitracarpus Zucc. ex Schult. $ 
Schult. f., and Richardia L., 


sampled only with a few species. In contrast, the two 


although these were 


Galianthe Griseb. sampled species are paraphyletic 
to Diodia spicata Miq., a species that was recently 
excluded from Diodia 


s. str. and tr. 
. If the position of D. spicata is confirmed by 


ansferred to 
Borreria 
further phylogenetic studies, the generic circum- 
scription of Galianthe should be widened to include 
at least this species. Dessein (2003) already showed 
that palynological data (7-zonocolporate pollen, long 
dans double reticulum) vu a close relation 

ata an anthe. Diodia L. as 
m D dimisi including species referred to 
Diodella Small by Bacigalupo and Cabral (1999), is 
not supported as monophyletic. Also, Spermacoce 
s.l. 
phyletic. 


including Borreria, is not supported as mono- 


BOUVARDIA AND MANETTIA 


Manettia is strongly supported as monophyletic (JS 
= 100, BS = 100, PP = 1), whereas support for 
Bouvardia is moderate (JS = 85, BS = 87, PP = 
0 n accordance with Andersson et al. (2002), 
Arcytophyllum serpyllaceum is corroborated as sister 
to Bouvardia. This strongly supported relationship (JS 
= 99, BS = 99, PP = 1), with the fact 
that the remaining Arcytophyllum species are strongly 


in combination 


supported as a monophyletic and distinct lineage (see 
below), suggests that at least A. serpyllaceum should 
be included within Bouvardia. Altho 


cates coe as a genus of shrubs only, it 


ough Bouvardia is 


both subshrubs and perennial herbs 
(Blackwell p Xn ons it possible to fit in 


pyllaceum i 1S similar 


to e nen eu different from oiher Arcytophyllum 
species in many respects. First, the stipule margin of 


A. serpyllaceum is not dentate or fimbriate, as in most 
Arcytophyllum species (Mena, 1990), but consists ofa 
basal sheath and a trullate mucro most 
Bouvardia species (Blackwell, 1968). Second, where- 


as the seeds of Arcytophyllum are more or less 


cymbiform i nA those of A. x en are 
discoid with hilum ardia 
(Andersson et P 


ae 


between seeds of A. serpyllaceum and Bouvardia is 


The major ae 


that Bouvardia seeds are winged, whereas those of A. 
serpyllaceum are not. 


ARCYTOPHYLLUM-HOUSTONIA CLADE 


Previous studies based on plastid DNA sequences 
have shown Vbi Me to be da and 
closely related to the Nor 
(Andersson , 1999; i ende eta 
Our analyses Ed the monophyly of the Neoitop: 
ical genus Arcytophyllum (JS = 93, BS = 92, PP = 1) 
only if 


n Houstonia 


. serpyllaceum is excluded from the genus (see 
above). Sister to Arcytophyllum is a group of North and 
Central American species presently classified i in the 
genera Houston nlandia, and Sten By 
having its noes relin d in North America ie 
than in South America, Arcytophyllum may be one of 
the few examples within Rubiaceae that has reached 
the Andes by a southern migration (Andersson et al., 
2002). From this perspective, Mesoamerican species 
like O. microtheca (Cham. € Schltdl.) DC. may 
represent remnants of stepping-stone populations. 
The Arcytophyllum—Houstonia clade as defined by 
our results is thus restricted to the New World. Seeds 
of Arcytophyllum and Houstonia are generally more or 
less cymbiform. Our results thus support Schumann's 


— 


Ea d of genera with cymbiform seeds. So 


ee good phylogenetic 
marker, Neanotis i be the a non-Ám 


relative of the Arcytophyllum—Houstonia clade (Lewis, 


= 
o 


the 


much discussion ou 
level. In a 


ere has been 
recognition of Houstonia at the generic 
recent molecular study based on ITS and trnL intron 
data (Church, 200. 

paraphyletic with respec 
genus Stenaria. I. Church (2003) suggested 
that Houstonia and Stenaria are better treated as a 


3), hace a appeared to be 
to the North American 


single genus. As currently circumscribed (Terrell, 


1996), the genus Houstonia is composed of 20 species 


117 


Groeninckx et al. 


Volume 96, Number 1 


2009 


Phylogeny of Spermacoceae 


T$ ovooooeunodc ura Sope[o peuoddns A[qeuoseor 99r Əy? 01 pousisse 318 JJJ 9 J s[eournu urwoy 


YIM SIEIS "p pue e sem$rjg 0) 1940 senuruoo pue (moue ue Ag poyeorpuT se) g OMSL uo sues "T's oeooooeurrodg oqu aur -seqouexq o^oqe pojeorpur ere soniqeqord 3orrejsoq “IPP quar quat 
pue torsd: *goga-gdso pourquioo uo poseq oor uvrso&eq oy} jo ouo Weg "q— :soquouexq oAoqe pojeorpur ore (0G <) sonqea (3u) densyoog pue (91) oymoyppoe[ "(pao = [TH] xepur uouo *oco 
= [I5] xeput &ouejsrsuoo *qgeq = 'T) seouenboes yuyu pue *orsda *T2g4-gdro Suipn[our srsApeue pourquios oy) uro; SLAN Z8Lp 9u Jo een snsuosuoo jotns oy} Jo euo Weg ^"y— 


g 


‘Ts 9v2202tur1adg 


N 


#870 


ge ound 


VE omy ———_ 


= a 


SY y 110 4p9H 
96 o] t 


EE ISUSIOLION SIJO/ Yes 


espe sno&peH 


[pes ESOMJNAY SIJOÁPIH 


e139]S0J98u SILO pap] 


Srsuaaro3 PIpur [ua po 


Lioun erpuejuapjo 
edie909150 VÁT 


sueisey srsdouejuoq 
siue erpuepus pro 
soplopnzny eurserqdur 


T 1923903 
y T 
— — 


t E T 1 


&soureo xeipAyorAyd 


'ejejsor erpuepueprO 


€2ul2209 tI neu ON. 


'esojdsoeo erjneqoy 


ri Tor jns enne 
Co pee erneuoy 
€orquirAjeure erneuowy 


suadal ejo1uacq 
; snupuejued uopojued 


BULILASSE] 


srsuooro3 erpueJuapIo 


ovIUOSMET srioApoH pesce 


SISUSIOHIOY SI)OApo]T ==] 


SOPIONIIMS SOAPS] 


vsoonnag syo<payy 
BISO]SOLNVU SIO API] 


nara LO 
edie909150 LÁI] 


suergen sisdouvjueq 
SULJE PHEAC 
sapropnzny wusergdury 


il 1 I 86/96 


esouieo xeipKuo| uq 


'ejernso: erpuejuapiO 


&ouro200 trjneqow 
tsojIdsoeo erjneqow 


ong] uopojuoaq 


V 


‘z om3 


s ?£29209tur19dg 


Ll 


Botanical Garden 


issouri 


Annals of the 


M 


118 


1940 SINUQUOS pue g omy uo SHES "| 


TR ==... urtpr« Sopeo peuoddns Apquuoseor 99r ou 01 pousisse 318 JJJ 9 [ s[eroumu ueuroq YIM SIEIS "p pue € SoS] 01 
TIM opou [euroqur 1odoop 9u 01 pousisse SI [JT [E39urnu ueuroy 'soqpougdq 940qe pojeorpur ste soniiqeqoxd 


1 


101191807 "yep uu pue *grsdi *qoqa-gdm patas uo Bess oon p our jo om) Weg 'q— 'seuouerq o4oqe pojeorpur ore (QG <) sonpea (Ys) densjoog pue (yop epwppef 
“€ emsry 


“($80 = TH “990 = ID “S8£T = T) se»uenbos ,ru;-Tu4; pue *grsda *T2941-gdro Surpnqout sisÁ[eue pourquioo ou urox SLAN Z8Lp ei Jo 9917 snsuosuoo JOLOS ot Jo 041 Weg "y— 


g 


apo 


moan udojKory 


uo um &udoy Kory 

umsojas un Aydoy Kory 

unpu an 4ydoy sory 
umreae[ un Aydorsary 

wmonnurunj&ugdoj&ary 

ues sue zn &ugdoj&ory. 


[aa SUBOLIS[U ereue1s 
SEU et Seo E SOL 
Es 'eruojsnoq 


iputquopio 


VIPO 
quuÁg enpey 


vuer33oy enpey 
soployjuenuas enpey 
"an|| enpex 

sisuadel enpey 


'E99EH09 En pex 


PIO SHEER 
Hd d 


RIO) 
govem sariaideu íS 


BSOAIOU erpuejuapiO 
t epydoad erpuejuapjo 


vsopnurqoo erpuerua pjo 


az omg 


Vp oni — ——] V 


moan NT Aydoy sory | 
ab s à 6/86 
soproouo ung Agdo1&ory 
umsojos un <ydoy ory: 
umparu um <ydoyAouy 
WnreAer urnp&qdoj Kor 
uma win <ydoy sory 


SUROLISIU BLIBUN)S men] 


rapi pg H me 
&o[nzoeo eIUO}SNOF] | 


BOOT] OJON erpurua pjo 


ey opourl erpuejuapiO 
səprone3 erpuejuapjo 

RIO[J[9U3) erpuepuoprto 
eyueideuás 


BSOAIOU PIPUR[UOPIO 
ppydoos eipurjuopio 


esou? LIPULTUEPIO a] 


versn 


vz oanig — —— 


119 


Groeninckx et al. 


Volume 96, Number 1 


2009 


Phylogeny of Spermacoceae 


ou jo pud 


Tino 1 ie es pue y EC page gdm pourquios uo poseq oor uetsodeg (0 
"^ g= = TY *oc'Q = ID ‘S8ET = T) se»uenbos Juq-que pue *orsdi *Togi-gdm Surpnqour sreAqpeue pouquoo o uio SLAN 28 LP am. M adi snsuosuoo yu Eu p oonp eq 'y— 


d 


LL ge aN 


TL garoqre so<payosany 


smua) PIPUR[USP|O) 
E a 


vdreo»o[eseur eesnr) 
mu EE SE tasnr) 


9Uuit0J SU) JO soaryequosordoy “y Pue € Som31J 01 


“390 uersoÁeq Əy} pue 994 snsuosuoo jorns 
3, 


n P G omg U0 SMEJ 


I 


DOS tL 4 aL (MESE 


ESF 


epidsry sdoomuads 


Siexoyi eopourq 
eun: e»ootuuadg 
eqni o»oovuuods 
'esojuouues erporq 
'"uedsootne vrporq 


puta du— 


edieoo[esiour eosni) 
v[eqdooo[eo wasn) 


epidsiy oo0oguuods 


"uodsooe[ne erpoiq 


snprsug sndreaen]q 

'ejourar o2o2tods 
"jojo vrporptuod 
€1ejdeo o»ooeuuods 

eyensoid ooooeuuodg 

esor ad00RULIOdS 


esnjuo) ooootuuods 


SISU: 


umjsi3 A] 


eqle enoue 
PI OFLA] erpreAnog 
euge? erpreAnog 


rpuepuopio 


muosurqor erputquapio 


*10ulo1 2209tuuods 


Sup: SNTE 


2 
E 
[o 
E 
& 
9 
9 
E 
E 
S5 
f 
DB 


esoo 2209€uuods 


vsnjuoo aooomuiadg 


Suelos erpreqong 
"Jqeos erpreuong — k 


suq em weeg 


[e 
*je[oquin ezrgroauurq 


)senqea Í 
6 « [ BUE OD oH I 


"p amd Ul 


V 


TUNISISBA] enoue 

qp enjeuejq 

Voru erpreanog 
PULL GES erpreAnog 


muosurqor erpuepuopio 


Sisuonew vrpue[uoptO 


SISUSLOQE] BIPURUIP|O 


vezna Pone yoy 
eeso ?rineuoy 
*qo[ismqo enneqox 


sisuonea vrpuepuopiO 
sTsuaroge} erpuepuapro 


SINUd] erpuejuopiO. LE 


8/6 


ewa EIjneqoyx pex 


Tr mu enneqo»x CY 
*qo|rsn3go enneuoyx => 


VE omiy — —— 


120 


Annals of the 
Missouri Botanical Garden 


Table 3. Major morphological differences between Knoxieae and Spermacoceae s.l. 


Knoxieae s.l. 


Spermacoceae s.l. 


Merosity often 5-merous or derived from the 5-merous state 
Inflorescence terminal (including pseudoaxillary) 
Calyx lobes often 1 or more calyx lobes enlarged 
Pollen bireticulum not yet reported 
xotesta ITW often slightly thickened 
Distribution 


paleotropical, centered in Madagascar and 
continental Africa 


often 4-merous, rarely 5-merous 

terminal or axilla 

rarely enlarged calyx lobes 

bireticulum common, often associated with 
heterostyly 

ITW without thickenings 

pantropical, with a few taxa reaching outside 

tropics 


the 


ITW, inner tangential wall. 


restricted to North America. The genus contains both 

annual pe 2m herbs with either ee 
flowers, crateriform seeds, 

osome numbers are oie 


or 11. 
s only recently described (Terrell, 


hom 
le e Chro 
among species of ies genus with x — 6, 7, 8, 
Stenaria, a genu 
2001a), includes five species previously included in 
the North American Hedyotis. The genus contains only 
perennial, heterostylous herbs. Due to our incomplete 


= 


sampling of these two genera, and given tha 
Houstonia forms a polytomy with Stenaria, our results 
are not conclusive with respect to whether it is best to 
recognize Stenaria or consider it part of a more 
broadly delimited Houstonia. A more extensive 
sampling should focus further on this question. 
Sister to the Houstonia—Stenaria clade is Old- 
enlandia microtheca. The prevailing basic chromo- 
, which occurs 


some number in Oldenlandia is n = 


in the type species O. corymbosa and in many of the 
species native to North America, Asia, Africa, and 


Australia (Lewis, 1965), but not in O. microtheca, 


W 

1991), not included in this study, and in some 
i ntil now, 
Oldenlandia microtheca and Oldenlandiopsis were 
never considered to be closely related to Houstonia 
ecause of t 
(Lewis, 1965; Terrell, 199 

Oldenlandiopsis contains only one species, O. 
callitrichoides (Griseb.) Terrell € W. 
previously included in Oldenlandia. This small- 


of morphological similarities 
I). 


Lewis, 


eaved, small-flowered, creeping herb is native to 
the West Indies and southern Mexico. Based on its 
chromosome number E its distribution, a position of 
Olde iue E 
clade clos 


likely. e seeds of Oldenlandiopsis are non- 


rcytop mE 
ao microtheca s 

crateriform pollen are 8-colporate with a 
lalongate, slightly crassimarginate endoaperture (Ter- 


rell & Lewis, 1990). These types of seeds and pollen 


are unusual within the Arcytophyllum—Houstonia 
clade. 


exceptional within the rest o 
landia 


Plurizonocolporate pollen grains are also 
enlandia group, where t 
exceeds five. The Asian genus Neanotis (Lewis, 1966), 
the Malagasy endemic Gomphocalyx (Dessein et al., 
2005a), the Afro-Madagascan Phylohydrax (Puff, 
1986), and the West Indian monotypic genus Lucya 
DC. (Terrell & Lewis, 1990) are notable B ee 
i the Hedyotis-Oldenlandia group in hav 
plurizonocolporate pollen grains. Both Go EN 
and Phylohydras belong to the Pentanopsis clade 


relationship between any of these taxa and the 
Arcytophyllum—Houstonia clade or the Pentanopsis 
clade. Nevertheless, considering their distribution, the 
Caribbean-Mexican genera Lucya and Oldenlandiopis 
are more likely to fall in the pd MC NE 
clade, whereas the Asian genus Neanotis is more 
likely to have its clo 


Pentanopsis clade. 


sest iode within the 


Two closely related genera from Baja California, 
Stenotis Terrell (Terrell, 2001b) and Carterella Terrell 
(Terrell, 1987), may also belong to the Arcytophyllum— 
e the Mesoamerican species 
Oldenlandia microtheca, they may represent remnants 
of stepping-stone populations. The  monospecific 
genus Carterella was described based on Bouvardia 
alexanderae A. M. Carter. It resembles Bouvardia in 
having unusually long corolla tubes, but differs from 
Bouvardia in pet wingless seeds and e 

ber n 3. The genus Stenotis, on the 
hand, includes seven former Hedyotis species a 
to the Baja California peninsula (Terrell, due 
These heterostylous, annual or i M herbs also 
chromosome number 13. A to 
Terrell (1987, 2001h), Cardi and a have 
their closest relatives among the Baja California 
species of Houstonia (H. 

Terrell et al., 1986). 


mucronata group sensu 


Volume 96, Number 1 
2009 


Groeninckx et 121 


Phylogeny of o — 


KADUA 


Our results support the resurrection of the genu 
oti 


Kadua for the Polynesian Hedyo e (Hawaiian 


Islands and French Polynesia: Terrell « et al., 2005). 
This taxonomic change was previously suggested by 
unpublished molecular data (Motley et al., 1998; 


Motley, 2003) and by morphological studies of the 
seed anatomy of the Hawaiian species (Terrell et al., 
2005). The genus Kadua was treated as a distinct 
genus until Fosberg’s (1943) revision of the group. He 
included the genus within a broadly delimited 
edyotis, except for the fleshy-fruited species, whic 
he treate ouldia A. Gray (Fosberg, 1937). Kadua 
species can, however, easily be distinguished from 
other Hedyotis species by their salverform, fleshy 
corollas with appendaged lobes, and by their either 
tardy, often "RUE septicidal dehiscent ES 
hi us fruits (Terrell et al., 2005). 
The genus Kadua Er comprises 28 species; all 


or indehiscent drupace 
are indigenous to the Pacific Islands with the majority 
(21 species) ma on the Hawaiian Islands 
(Terrell et al., 2005). Seeds of these Hawaiian Kadua 
species fall into four groups, described by Terrell et 
al. (2005). Based on the chloroplast data alone, the 
relationships within the genus Kadua remain mostly 
unresolved. Only section Wiegmannia Meyen, W. L. 

agner & Lorence (represented in our sampling by K. 
cordata Cham. & Schltdl., K. degeneri (Fosberg) W. L. 
Wagner & Lorence, K. elatior (H. Mann) W. L. 
Wagner & Lorence, K. flynnii (W. L. Wagner & 
Lorence) W. L. Wagner & Lorence, K. laxiflora H. 
Mann, K. littoralis Hillebr., and K. parvula A. Gray) 
and section Gouldiopsis (Fosberg) W. L. Wagner & 
Lorence (represented in our sampling by Kadua 
Hook. Arn. 
(Fosberg) W. L. Wagner & Lorence) were recovered. 


centranthoides and K. foggiana 
A broader sampling including more Kadua species 
and more molecular markers is needed to discuss 
molecular evolution in the light of the seed morpho- 
logical observations of Terrell et al. (2005). 

Oldenlandia biflora is sister to the Kadua clade. Its 
distribution from (sub)tropical Asia to the western 
Pacific is consistent with the sister relationship to the 
Polynesian Kadua clade. Our results show that O. 
biflora can no longer be included within the genus 
Oldenlandia, but it is necessary to await further 
studies before transferring it to Kadua or describing a 
new genus. So far, we have not found morphological 
characters to support the transfer. 


HEDYOTIS S. STR. 


It seems appropriate to restrict the name Hedyotis to 
the Asian and Micronesian species of the genus, 


which includes the type species H. fruticosa (Sri 
Lanka). Several d nic considered the genus 
Hedyotis to be a distinct Asian taxon (Bremekamp, 
1952; Hallé, 1966; nat 1975, 1; Andersson et 
al., 1999). Hedyotis fruticosa dnd its Ásian relatives 
are not closely related to the American species of 
is (Houstonia lineage) o olynesian 
dua). The Asia 


species (Hedyotis s. str.) differ from the American and 


/yot1. to t 
species (Ka an and Micronesian Hedyotis 


Polynesian ones in their combination of a robust 
(sometimes shrubby) habit, small beaked and diplo- 
hragmous capsules, dorsiventrally compressed seeds 
with the hilum on a conspicuous central ridge (Terrell 
& Robinson, 2003), and 
(Kiehn, 
broad concept of Hedyotis, merging eral genera 
(Hedyotis s. str., Houstonia, Kadua, Kohautia, Old- 
enlandia, etc.), as was proposed by several research- 
1943; Merrill E Metcalf, 1946; bc 

agner et al., Fosberg & Sachet, 1991; 
Dutta & Deb, 2004), i is no b. supported. If this is 
confirmed with further sampling, all North American 


a high chromosome number 
1986). Our results clearly demonstrate that a 
He sev 


ers (Fosberg, 
1987; 


species now called Hedyotis would require new 
combinations under other generic names. 
Pleiocraterium Bremek. (not included in this study) 
is probably related to the Hedyotis s. str. clade. The 
genus was described by Bremekamp in 1939, including 
four species distributed in India, Sri Lanka, an 
Sumatra. The gen 
cups that are formed by the connate leaf bases. The 


eric name refers to the numerous 


type species of the genus, P. verticillare (Wall. ex Wight 
rn.) Bremek., was previously included in Hedyotis. 
s differs fr 


species in having distinctly beaked capsules and 


However, the genus rom other Hedyotis s. str. 
parallel-nerved, quaternate leaves. The internodes 
as a result of which the leaf whorls 
are clustered in rosettes. It will be necessary to wait, 


remain ve 
however, until molecular data of Pleiocraterium 
become available before a close relation of the genus 
to the Asian Hedyotis species is confirmed. 


AGATHISANTHEMUM CLADE (CLADE II) 


The African genus Agathisanthemum is not sup- 
ported as monophyletic by our analyses. The mono- 
typic A Lelya Br 
Agathisanthemum, making it paraphyletic as currently 


M 
frican genus emek. is nested within 


tens (1998) showed that Agathisanthemum and Lelya 
share the same pollen type, characterized by a distinet 
endocolpus or endocingulum, a mesoporus surrounded 
by a costa at the inside of the grain (described as a 
compound ora by Lewis, 1965), and a microreticulate 


122 


Annals of the 
Missouri Botanical Garden 


sexine with granules on the muri facing the lumina 
(bireticulum). 

A group of African Oldenlandia species is sister to 
Oldenlandia 
species, O. angolensis and O. goreensis, belong to 
Oldenlandia subg. Anotidopsis (Hook. f.) K. Schum. 
This subgenus, as described by Bremekamp (1952), 
includes three other putative species of which only O. 


Agathisanthemum. Two of the three 


cephalotes (Hochst.) Kuntze (not included in our 
sampling) i is currently recognized. Subgenus Anotidop- 


erized by distinctly beaked capsules, The New 
World taxon O. uniflora is sister to O. angolensis and O. 
goreensis. More detailed (molecular as well as morpho- 
logical) studies within the Agathisanthemum clade are 
n d to evaluate if the three Oldenlandia species, O. 
angolensis, O. goreensis, and O. uniflora, or the entire 
Oldenlandia subg. Anotidopsis, are to be transferred to 
a new genus or if these species are better treated as 
members of the genus Agathisanthemum. 

The Asian Hedyotis species are sister to the 
Agathisanthemum-Oldenlandia clade. This relation- 
ship is not unexpected as Bremekamp (1952) already 
suggested a close relationship between Agathisanthe- 
Hedyotis 
sect. Diplophragma) based on a similar type of 


mum and the Asian Hedyotis species (i.e., 
dehiscence of the capsules. 


PENTANOPSIS CLADE 


Our sampling resulted in a broader concept of the 
Pentanopsis clade than proposed by Thulin and 
Bremer ( ). They included Amphiasma, Conosto- 
mium (Stapf) Cufod., Manostachya Bremek., Penta- 
nopsis, and Phylohydrax. 

Oldenlandia affinis was not included i in the study of 

o 
Amphiasma by 
(1999) n Dessein et al. 
(2005a). Amphiasma, O. affinis, and Pentanopsis share 

ed capsules, 
nonpunctate testa cells 
(Bremekamp, 1952). However, a detailed study is 
needed to find more unambiguous morphological 
characters to support a relation among the three taxa. 

n the past, Gomphocalyx (a monospecific genus 
endemic to Madagascar) and Phylohydrax (a genus 
described i in 1986 by Puff to accommodate the East 


another and to the Pentanopsis clade (Dessein, 2003; 


Thulin & Bremer, 2004; Dessein et al., 2005a). The 


morphological characters (amphistomatic leaves, pluri- 
zonocolporate pollen, indehiscent fruits, and seeds 
with a weak, pale exotesta) as shown by Dessein et al. 
2005a). Almost all taxa in the Hedyotis-Oldenlandia 


group have multiovulate ovaries, and the number of 


— 


pollen apertures rarely exceeds five. The presence of 
uniovulate ovaries and plurizonocolporate pollen were 
the main reasons why Gomphocalyx and Phylohydrax 
were previously included in Spermacoceae s. str., 
where it is more common than in the rest of the 
permacoceae s.l. tri which 3-colporate pollen 
eds (Dessein Ši ri 2002, 2005b; Dessein, 
003). As mentioned above, the nme genus Neanotis 

is a notable exception in having plurizonocolporate 
pollen grains. The genus also shows a trend toward 
up dn in the number of seeds per locule. In mature 
fruits, only one or two seeds are present. However, 


genus it wo 

ee oe between Neanotis 

Phylohydrax. A few authors (Capuron, 1973; Pies- 

schaert, 2000 alo proposed. : a close i und 
and 


z 
O 


t 
a eaten to 2) 1 Althetgh porci. 
carpa is not a trailing herb like Gomphocalyx but a 
(sub)shrub, the two taxa share a calyx with eight lobes, 
uniovulate ovaries, and plurizonocolporate pollen. The 
last two characters also support a close relationship 
between Phylohydrax and Lathraeocarpa. However, 
several morphological characters distinguish La- 
thraeocarpa from Gomphocalyx, some of which might 
even point to an affinity with Triainolepis Hook. f. 
First, the (sub)shrubby habit of Lathraeocarpa is much 
more similar to the shrubby habit of Triainolepis than 
to the herbaceous habit of Gomphocalyx. Second, the 
pyrene of L. decaryi Bremek. is surrounded by eight 
strands of thin-walled cells, a condition id similar to 
that observed in some Triainolepis spec 
kamp, 1957; e e 2001). Lik 


pnis dnd Pisnolenis have a plurilocular ovary and 


reme- 
ewise, Mp 


fleshy fruits, whereas Gomphocalyx has a bilocular 
ovary and dry fruits, which has pro e authors 
(Karehed & Bremer, 2007) 


Lathraeocarpa in the emended tribe Knoxieae ae 


mpted so 

to tentatively inclu 
than in Spermacoceae s.l. However, we will have to 
wait until molecular data become available to assess 
the taxonomic position of Lathraeocarpa with more 
certainty (Dessein et al., 2005a) 

Species of Conostomium form a strongly supported 
clade JS = 99, BS = 99, PP = 1) together with 
Oldenlandia herbacea. The type of the genus Con- 


ostomium, C. natalense (Hochst.) Bremek., is unre- 


Volume 96, Number 1 
2009 


Groeninckx et 123 


Phylogeny of AOE 


solved with respect to the other species of Conosto- 
mium and to O. herbacea. Both Conostomium and O. 
herbacea have seeds with coarsely granulate testa 
cells (Bremekamp, 1952; Dessein, 2003) and pollen 
that is larger than that of most other genera within the 
1952; 


ese characters, however, are 


Hedyotis-Oldenlandia group (Bremekamp, 
Scheltens, 8 
homoplasious because granulate testa cells and large 
pollen grains also occur elsewhere in the Hedyotis— 
Oldenlandia group. We observed granulate testa cells 
in Kohautia subg. Pachystigma, O. corymbosa, and O. 
nematocaulis Bremek., whereas large pollen grains are 
characteristic of Amphiasma, Gomphocalyx, and 
Phylohydrax. The most striking feature of Conosto- 
mium pollen, namely the short ectocolpi (Scheltens, 
1998; Dessein et al, 2005a), is not found in O 
herbacea or in most other members of the Pentanopsis 
clade, but it is reported F Gomphocalyx and 
Phylohydrax (Dessein et al., 5a). 

The last additional species M in the Penta- 
nopsis clade is Oldenlandia rosulata, an African 
species named after its basal rosulate leaves. The 
relationship of O. rosulata to other members of the 
Pentanopsis clade remains unclear. 

Despite the strong support for the Pentanopsis clade 
(JS = 95, BS = 95, PP = 1) in our molecular analyses, 
the group is not easily morphologically characterized. 
The only u ns feature for the clade wou 
Thulin and Bremer (2004) 


Nevertheless, ihe ciu is not truly basal, but 


= 


e wha 
called basal placentation. 


rather axile with E dde or ovule attached near the 
base of the se ur observations show that this 
kind of lso found outside the 
Pentanopsis clade. Moreover, the basal placentation 


S Lan is 


character state is only vaguely defined, and more 
detailed placentation studies within Spermacoceae s.l. 
are needed before further conclusions can be drawn 


about the phylogenetic value of this character. 


MONOSPECIFIC GENERA WITHIN THE HEDYOTIS— 
OLDENLANDIA GROUP 


Besides the genus Gomphocalyx of the Pentanopsis 
clade, 


several 


the [xd e group comprises 
other monospecific genera. These monospe- 
cifie genera often have a zen characters, 
making it very difficult to discuss their relationship 
with other Spermacoceae 

In our sampling, for example, the Afro-Madagascan 
genus Mitrasacmopsis has s with undulating 
radial exotesta cell walls, distinctly stalked placentas 
with ovules positioned on the periphery of the 
placental tissue, pollen grains with a double reticu- 
um, and fruits with a conspicuous beak (Groeninckx 


et al., 2007). Our molecular results suggest a close 


relationship of this monospecific genus to Hedythyrsus 

Oldenlandia fastigiata. Our own observations 
have identified similar placentation types within these 
taxa. Moreover, Hedythyrsus and Mitrasacmopsis have 
the same type of capsule dehiscence (loculicidal 
followed by den dehiscence), seeds with testa 
cells that s 
pollen with a a double ad (NE 2005). 


w the same undulating radial walls, and 


The lus is sister to 


a clade of African idea ic a genus was 
separated from Oldenlandia large he basis of its 
capsule dehiscence (both loculicidal nd septicidal vs. 
only loculicidal in Oldenlandia). Bremekamp (1952) 
closely associated Dibrachionostylus with Agathisanthe- 
mum because of their similar fruit dehiscence. However, 
Dibrachionostylus differs markedly from Agathisanthe- 
mum in the pollen aperture morphology (Lewis, 1965). 
As mentioned above, Agathisanthemum has a distinct 
ectocolpus, an endocolpus or endocingulum, and a 
mesoporus surrounded by a costa at the inside of the 
grain (Lewis, 1965). Pollen grains of vi drei 
are also 3-colporate but do not have a costa on the i 

(Lewis, 1965) 
pollen are, therefore, more similar to those of 
Amphiasma, Oldenlandia, and Pentodon (Lewis, 1965). 


Nesohedyotis is another monospecific genus previ- 


e apertures of Dibrachionostydus 


ously included in the Hedyotideae. Its only species, N. 
arborea, shows a superficial resemblance to the East 
African genus Hedythyrsus; specimens of both taxa turn 
black when dried, and their leaf shape and inflorescence 
structure are similar (Bremekamp, 1952). However, our 
results show that Nesohedyotis is more closely related to 
the former tribes Spermacoceae s. str. and Manettieae 
of the Hedyotis- Oldenlandia group. 
Nesohedyotis has unisexual flow: 


than to members 
are unusual 
among Spermacoceae, and, in contrast to Hedythyrsus, 

its fruits open by a single loculiridal split. x d itis 
on 


e of the 
its small population size and pu Wide 
distribution make JVesohedyotis eae ac- 
cording to IUCN Red List criteria (IUCN, 2 
According to Verdcourt (1976 
nesohe 


, the mee 
'yotis Tennant, which i 


Tanzanian Pseu 
included in our sampling, is closely Ei to 


Nesohedyotis and Hedythyrsus. Pseudonesohedyotis 
has indeed the same leaf shape and inflorescence 
structure the latter. two taxa. In habi 


distribution, however, it resembles Hedythyrsus more 
Both  Pseudonesohedyotis 
2 e are (sub)shrubs, whereas Nesohedyotis is 

mall tree. Moreover, Pseudonesohedyotis differs 
a Nesohedyotis i in having hermaphroditie flowers. 


than Nesohedyotis. and 


Again, it is necessary to wait until molecular data 
become available to discuss the taxonomic position of 
Pseudonesohedyotis with more confidence. 


124 


Annals of the 
Missouri Botanical Garden 


Based on the presence of an apparently superior 
ovary, Jovet (1941) originally placed Mitrasacmopsis 
and Astiella Jovet, another monospecific genus of the 
Hedyotis-Oldenlandia 


(not included in this study), within Loganiaceae— 


group endemic to Madagascar 


Spigelieae. Members of Rubiaceae . are generally 
characterized by the presence of a rior ovary. 
Groeninckx et al. (2007) je pe flowers of 
Mitrasacmopsis are initially epigynous with inferior 
ovaries. Expansion of the upper part of the ovary in 
fruiting stage results in a change in the ovary position 
of Mitrasacmopsis from basically inferior to secondar- 
ily semi-inferior. The same kind of fruit development 
also most likely occurs in Astiella. In her morpholog 

ical study of the Rubioideae, bo gd o sated 
that some genera of Spermac - 
inferior fruits. According to io um "Mis 
statement is based on the strong expansion of the top 
of the nectary disc in the fruiting stage. However, we 
have not 
Spermacoceae s. str. Nevertheless, within Spermaco- 


bserved semi-inferior ovaries within 
ceae s.l. several other taxa, apart from Mitrasacmopsis 
and Astiella, are characterized by the presence of a 
eak at fruit stage (Conostomium spp., Hedythyrsus 
spp., Kohautia spp., Oldenlandia spp.). These beaks 
are not remnants of the nectary dise and probably 
originate in a similar way as in Mitrasacmopsis. 
However, the ovaries of these species do not undergo a 
remarkable reverse in shape in the fruiting stage as 
observed in Mitrasacmopsis and Astiella. Based on 
their fruit shape, Mitrasacmopsis and Astiella seem 
closely related. However, Jovet (1941) also suggested 
a close relationship between Astiella and the Asian 
Anotis DC. species, presently classified in the genus 
Neanotis (Lewis, 1966). Astiella differs from both 
genera in having only two calyx lobes, a character that 
so far has not been observed within the Hedyotis— 
Oldenlandia group, and uniovulate locules. Molecular 
sequence data of Astiella will allow us to discover the 
taxonomic position of the genus in the fut 
Other monospecific genera of the Hedyotis-Old- 
enlandia group are Carterella, Dolichometra K. 
Schum., Lelya, Leptoscela Hook. f., Lucya, Phyllocra- 
ter Wernham, Polyura Hook. 
Bremek., and Oldenlandiopsis. The genera Carterella, 
Lelya, 


discussed in previous sections. To date, the taxonomic 


. Stephanococcus 


Lucya, and Oldenlandiopsis were already 


position of most of these monospecific genera remains 
controversial because molecular data are lacking. 
KOHAUTIA 


Kohautia is a genus of 31 species (Mantell, 1985) 
distributed from the Indian subcontinent throug 
Pakistan, Iran, the Arabian Peninsula, Sinai, eastern 


Egypt, and throughout most of Africa south of the 

Sahara (including Socotra, Cape Verde, and Mada- 

gascar). us can easily be hos d fro 
edyo 


s Oldenlandia 
group by its unique flower holy The anthers 


€ gen 
other representatives of the 


and stigma are always included, with the stigma held 
well below the anthers or Padre just d 
h his monomorphic short-styled condition is, 
with the exception of a few cen of PAGE 
mium, unique among the African members of the 


former tribe Hedyotideae. For this reason, Kohautia 


not sister clades. Subgenus Kohautia is sister to the 
Pentanopsis clade, whereas subgenus Pachystigma is 
sister to an Oldenlandia clade containing the type 
species O. corymbosa. 
Despite the unifying floral architecture, there are 
us morphological differences between the two 
) 


1965; Mantell, 1985 


of stigmatic lobes is the most striking diagnostic 


numer 
subgenera (Lewis, he number 
character that allows identification of the subgenera 
even in the fie 
styles with two thin filiform stigma lobes 
Pachystigma is characterized by the presence of only 
a single, ovoid to cylindrical stigma lobe. Seeds are 
also different in the 
Kohautia 


shape with 5- or 6-angled testa cells, whereas in 


subgenera: subgen 
ubco: 


wo 
seeds are angular-conie to subconic in 
a the seeds are rounded with 
n of Kohautia can 


a be divide di 
coinciding with d two subgenera (Lewis, 1965). 


o easily Rn a groups 


Other differences between the two dde are 
found in floral architecture and chro 
Based on these differences, Mantell (1985) e 
sized that the two subgenera may have diverged and 
developed independently of one another fairly early 
on and she even tentatively proposed E e of 
the two subgenera to generic rank. that time, 
Mantell decided to maintain a widely pe genus 
Kohautia, mainly for practical reasons. However, our 
molecular data now clearly support the recognition of 
two genera. Sampling within the genus still needs to 
be improved before proposing new generie circum- 
scriptions. 


OLDENLANDIA 


Govaerts et al. (2006) currently accept 76 Old- 
enlandia species from Africa, 155 from Asia and 
Australia, 23 from America, and eight from the Pacific 
slands. However, as documented in previous molec- 


ular studies (Bremer, 1996; Andersson & Rova, 1999; 


Volume 96, Number 1 
2009 


Groeninckx et 
Phylogeny of ANEO 


125 


Bremer & Manen, 2000), Oldenlandia is shown to be 
polyphyletic. 

Bremekamp (1952) divided the 61 species that he 
recognized from Africa into 16 subgenera. Our results 
do not support the majority of these subgenera. Only 
the subgenus Hymenophyllum Bremek. (Oldenlandia 
echinulosa and O. nervosa) and subgenus ee 
O. angolensis and O. goreensis) are corroborated. 

The type species, Oldenlandia pa is sister 
to a clade with the African species O. capensis L. f., O. 
robinsonii Pit, O. nematocaulis, O. taborensis Bre- 
mek., and O. wauensis Schweinf. ex Hiern. The last 
species, O. wauensis, was segregated by Bremekamp 
(1952) in a new genus Thecorchus Bremek., which he 


> 
[5] 


proposed to be allied with Otomeria of the tri 
Knoxieae because of its distinctly elongated capsules 
and equal number of tetramerous and pentamerous 
flowers. However, Kàrehed and Bremer (2007) showed 
that Thecorchus is not related to O 


to Oldenland: 


ia. Our results, which place Thecorchus 


tomeria but is close 
in a clade comprised of the type species of Old- 
enlandia, support the transfer of wauensis 
(Schweinf. ex Hiern) Bremek. back into Oldenlaridici 
The type species O. corymbosa and O. capensis belong 
to Bremekamp’s (1952) subgenus Oldenlandia K. 
Schum. Besides these two species, subgenus Old- 
enlandia also includes O. fastigiata and O. herbacea. 
These species are apparently not related to O. 
corymbosa and its allies. Oldenlandia fastigiata is 
sister to Hedythyrsus and Mitrasacmopsis, whereas O. 
herbacea in the Pentanopsis clade is sister to a 
paraphyletic Conostomium. Bremekamp (1952) al- 
ready pointed out that O. herbacea differs from the rest 

the subgenus by the coarsely granulated walls of the 
testa cells, the rather large flowers, and the slender 
corolla tube. 

The Australian species of Oldenlandia, O. mitra- 
sacmoides and O. galioides, sampled here belong to a 
clade comprising the Australian Synaptantha tillaea- 
the Austro-Asian O. tenelliflora, the African 
species O. lancifolia, and the Kadua species (includ- 
ing O. biflora). Oldenlandia mitrasa 
the rest of the clade. Synaptantha tillaeacea is sister 
to a clade with Oldenlandia tenelliflora, O. galioides, 

and O. lancifolia. Synaptantha Hook. f. may be 


cea, 


cmoides is sister to 


distinguished from the other genera in the clade by its 
slightly connate corolla lobes, stamens with filaments 
attached to both the corolla and the ovary, depressed 
obconie or ovoid od and half-inferior ovaries 
(Halford, 1992). In of Australian Old- 
enlandia, pen rd do) a five groups 
mostly based on seed morphology. Oldenlandia 
galioides and O. tenelliflora are placed together in 
his group one, which is characterized by obconic 
seeds that are slightly laterally compressed and 


obtriangular in outline. Oldenlandia mitrasacmoides 
belongs to his group two, which is characterized by 
scutelliform seeds that are oblong or broadly elliptic 
in outline, with the hilum situated on a conspicuous 
central ridge. The African species O. lancifolia has 
seeds similar in shape to those of its sister O. galioides 
(Dessein, 1998). 

Not all American Oldenlandia species included in 
our sampling are placed within the Arcytophyllum— 
Houstonia clade (see discussion above). The remain- 

South ecies of Oldenlandia, O. 


salzmannii and O. tenuis, form a clade sister to the 


ing South American 


former tribes Spermaococeae s. str. and Manettieae. 
Terrell (1990) already reported that O. salzmannii is 
clearly distinct from Houstonia and Oldenlandia. In 
contrast to other Oldenlandia species, O. salzmannii 


number of chromosomes 
Moreover, it shares some unusual characters with 
Oldenlandiopsis: stipules are minute, not more than 
nlandia stipules are often 2-3 m 
long); few stiff hairs occur on the leaves Oldenlandia 
species usually have smaller, softer hairs}; and it has a 


0.5 mm long 


creeping habit (which is rare in Oldenlandia, the 
usual habit being erect to spreading or prostrate). It 
would be very informative to include Oldenlandiopsis 
in future studies to investigate its relationship to 
either O. otheca 


Pi clade above) or O. salzmannii. 


see discussion of the Arcyto- 


Furure RESEARCH PLANS AND CONCLUSIONS 


though our analyses found well-supported clades 
within Spermacoceae s.l, many relationships within 
and between these clades still remain unresolved. 
Furthermore, many relationships detected here are 
contradictory to previous taxonomic treatments and 
await morphological backup. This study was a multi- 
partner collaboration resulting in a framework for 
future Spe 


focus on FIR addition. 


ch. Further studies will 
al DNA markers 


nuclear DNA data) to provide better resolution vin 


coceae researc 
(Le., 


the tribe. Besides improving the character sampling, 
we also need to balance the on samplin. 
including more Ásian and American taxa. In addition, 
concerted studies will focus on the morphological 
characterization of monophyletic d ithin Sper- 
macoceae. This requires a morphological investigation 
across taxa to find character support Hor the many new 
phylogenetic relationships detected. 


Literature Cited 


Andersson, L. € J. H. E. Rova. 1999. The rps16 intron and 
the phylogeny of the Rubioideae (Rubiaceae). Pl. Syst. 
Evol. 214: 161-186. 


126 


Annals of the 
Missouri Botanical Garden 


— & F. G. Alzate. 2002. Relationships, circumscrip- 
tion, 2 Nae Es E br Mn a 
0—49. 


las especies 
aed del is Diodia (Rubiaceae, Spermacoceae). 
arwiniana 37: 15-165. 
ren W. H. a 1968. Revision of Bouvardia (Rubia- 
ceae). Ann. Missouri Bot. Gard. 55: 1-30. 


g v EM 
rav. Bot. Neerl. 36: 438-445. 

The African za a Oldenlandia L. sensu 

Hiern & K. Schumann. Es Ned. Akad. Wetensch. 

Afd. Natuurk. 48: 1-19 

1 s M M tribu nouvelle des 

Rübicidées (Rubiacées). Bull. Jard. Bot. État Bruxelles 

159-166. 


Hedyotidearum. Recueil T 
——. 1952. 


Pp, 
e 


6. Remarks on the position, the delimitation and 
sabuivision of the Rubiaceae. Acta Bot. Neexl. 15: 1-33 
& 


anen. ao un ylogeny and oe of 
the subfamily R ). PL Syst. Evol. 225: 
43-72. 
Bremer, B. 1 Phylogenetic studies within the Rubiaceae 
and relati Ri other fam ~ based on molecular 


Uem Opera Bot. Belg. 7 
Cabral, E. L. 1991. du med del ird QUU 
(Rubiaceae). Bol. Soc. Argent. Bot. 27: 234-2 
. 1973. Révision des Rubiacées de MN et 
s Comores. Unpublished manuscript. Notes regroupées 
et mises en forme par J. Bosser, m de F. 
Chauvet. e de Phanérogam 
003. Molecular eo of Houstonia 


America. Molec. Phylogen. Evol. 27: 223-238. 
Bw e 1998. Fylogenie van de Hedyotideae ed 
morfo P e en 
"e nse taxa. Mas 


Leuven, Leuven, Belgium. 


T 


atomische studie van de 
s Thesis, Katholieke Universiteit 


o 
(Rubiaceae). Ph.D 
Leuven, oe ae 

. Scheltens, s "Huys 


Studies in the Spermacoceae 
— Katholieke Universiteit 


, E. Robbrecht & E. 


s & E. Robbrecht: 
2005a. Gomphocalyx and puc Rubiaceae): Sister 
taxa excluded om the Spe ies Su 
T 


a eaturing a 
Joi] 


4: 91—107. 
H. Ochoterena, P. De Block, F. P E us 
E ET E. Smets, S. Vinckier & S. ns. 2005b. 


Palynological ue and their la signal in 
Rubiaceae. Bot. Rev. 71 

Dutta, R. Deb. mic Revision of 
Hedyotis L. (Rubiaceae) i in cane Subcontinent. Botanical 
Survey ns Kolkata, India. 

Farris, J. 1989. The retention ps a the rescaled 
Hee nee 2 Cladistics 5: 417— 

Fosherg, F. R. 1. The p cold ad! Bull. 

ernice P. un Mus. 147: 
—— ——. 1943. The Polynesian species of a (Rubia- 
m— Bull. Bernice P. Bishop Mus. 174: 1— 


Mantell, D. E 


& M. Sachet. 1991. pe in Indo-Pacific 
Rubiaceae. Alleronia 6: 191-27 
Goloboff, P. Nona Ver: n 2.0. Program and 
documentation ee by the author. Tucuman, 
Argentina. 


Govaerts, R., M. Ruhsam, L. O E. Robbr echt, D. 


«http: e kew.org/wesp/ 

m m Nov ember 200 

Groeninckx, I. 2005. detec naar de ds positie 
an Mitrasacmopsis (Rubiaceae) op basis van moleculaire 


morfologische data. Masters Thesis, Katholieke 
Universiteit Leuven, Leuven, Belgium. 
" rijdaghs, S. Huysmans, E. Smets & S. Dessein. 


supe 

Halford, D. A. 1992. Review of the genus s Oldenlandia L. 
(Rubiaceae) a related genera in Australia. Austrobai- 
leya 3: 683-72 

Hallé, N. 1966. N pt. 1. III. Hedyotidées. Pp. 75— 
124 in A. Aubréville (editor), Flore du Gabon. Mus. Hist. 
Nat. 

Hayden, s. M. . 1968. Systematic Morphological Study of 
New World a Seeds: Rubioideae Sensu Brem 
kamp. i i 


vemplavia im 


Huelsenbeck, J. & F. onu 2001. MRBAYES: Bayesian 

inference of phylogenetic trees. 17: 
754-7 

IUCN. 2001. IUCN Red List Categories and Criteria Version 
3.1. Prepared by the IUCN Species Survival Commission. 
IUCN, Gland, Switzerland, and Cambridge, United 
Kingdom. 

Janssens, S., K. Geuten, Y.-M. Yuan, Y. Song, P. Küpfer & 
E. Smets. 2006. Phylogenetics of Impatiens and Hydrocera 
(Balsaminaceae) using chloroplast aipB-rbeL spacer se- 
quences. Syst. Bot. 31: 171 "180. 

Jovet, P. 19 Aux confins des 
Loganiacées. Notul. Syst. (Paris) 10: 3 

Karehed, J. & B. Bremer. 2007. The une of Knoxieae 
(Rubiaceae}—Molecular data and their taxonomic conse- 
quences. Taxon 56: 1051-1076 

Kiehn, M. 1986. Karyosystematic studies on Rubiaceae. PI. 

Syst. Evol. ae 213-223. 

. S. Farris. 1969. Quantitative phyletics and 

s. Syst. Zool. 18: 1-32. 

1965. Coole study of African Hedyoti- 

dea eae icc Amn. uri Bot. Ga 52: 182-211. 

The pia genus Neanotis nomen novum 

m e allied taxa in the Americas (Rubiaceae). Ann. 

3246. 


Bioinformatics 


oe et des 


Maddison, W. P. Maddison. 2001. MacClade 4: 
Analysis Phylogeny and Character Evolution. Vers. 
4.01. s Associates, sans erland, Massachusetts. 

Manen, J. F., rendorfer. 1994. Phylogeny of 
Rubiacear Hübiede ise from the sequence of a 
cpDNA intergene region. Pl. Syst. Evol. 190: 195-211. 

. 1985. The Afro-South-west Asiatic Genus 

Kohautia Cham. € Schlechtd. (Rubiaceae-Rubioideae— 

Hedyotideae): Morphology, Anatomy, Taxonom 

ography, and e on. 

Wien, Vienna, Aus 


A. Natali 


y, Phytoge- 
Ph.D. Dissertation, Universitàt 


Volume 96, Number 1 
2009 


Groeninckx et 127 


Phylogeny of INER 


Mena, V. P. 1990. A revision of the genus Arcytophyllum 
i Hedyotideae). Mem. New York Bot. Gard. 60: 


ur D. & C. Metealf. 1946. Hedyotis L. versus 
oleada L. and the status of Hedyotis lancea Thunb. in 
relation to H. consanguinea Hance. J. Arnold Arbor. 23: 
6— o 
Motley, T. J. : 8. Phylogeny of Hawaiian and Pacific 
Hedyotis E Fruit ev im tion and the 1 Ra a 
for conservation and genomics. Abstracts of A 
I Mobile, up Pote 2003: 88-89. 
St & A. Albert. 1998. Molecular 
men of a “Hedyotis (Rubiaceae). Amer. J 


truwe rt. 


R. J. Hickey. 1996. The genus Ernodea 


analyses and n Syst. 

Nixon, K. C. 1999. The parsimony 
rapid nou analysis. Cladistics 15: 4 

2. WinClada Mn Vers. 1.00. a Published by 
the n Ithaca, N 

Oxelman, B., M. Liden & e DN 1997. Chloroplast 
rps16 intron ps ny of the ee ee (Caryophylla- 
ceae). Pl. Syst. Evol. 206 

Hiec den. P, 200. Carpolosy ps Pollen Morphology of 
the 


owards a 


ubioid 
bioideae). 
ee an a Generic Delimitation: Ph.D. Dissertation, 

Katholieke Universiteit Leuven, Leuven, Belgium 

Posada, D. & K. A. Crandall. 1998. Modeltest: Testing the 
model of DNA substitution. Bioinformatics 14: 817-818 
Puff, C. 1986. Phylohydrax ARENA D cA d 
the African and Madagascan 


1988. Tomea woody Rubiaceae. Character: 
and progressions. pee to a new 

subfamili al classifi cation. Ope . 1: 1-271. 

—. 1993. Supplement to do 1988 pe of the 

classification of the i 


Robbrecht (editor), S in Rubiaceae Macte: 
matics. s ot. Belg. 6 6. 

n. 2006. T j l lineages 
of the uo wes (nemico dub sio ms). Combined 


analysis oe and ne) to net He pon of 
C d Luculia based 
M irnL-irnF and. aipB-rbeL de A new 
Cinchonoideas and 
46. 


of Cinchonoideae (Rubia- 
ceae) in the ulusm. United States. J. Arnold Arbor. 
83. 


J. P. Huelsenbeck. 2003. MRBAYES 3: 


under mixed models. 


ollenmorfologische studie van de 

Ae oranges Hadyoddess (Rubiaceae). Licentiate Thesis, 

Katholieke Universiteit Leuven, Leuven, Belgium. 

iaceae. In A -Eng ler & K. Prantl 

(editors), Die ups Pflancenfamilien 4: 1-156. 

Simmons, M. P. & H. Ochoterena. 2000. Gaps as characters 
in sequence- Dascd phylogenetic analyses. Syst. Biol. 49: 
369-381. 

Staden, n ; Ke vado Bonfield. 1998. The Staden Package. 
Pp. 115-130 n S. Miseners & S. Krawetz (editors), 
Computer Matis in Molecular Biology. The Humana 
Press Inc., New York. 


Schumann, K. 1891. Rubi 


Suzuki, Y., G. V. Glazko & M. Nei. 2002. Over credibility of 
moleculas ur de obtained by Bayesian phyloge- 
s. Proc. Natl. Acad. Sci. U.S.A. 99: 16138-16143. 
Swofford, D. 2002. PAUP*: eene Analysis Using 
Parsimony (* and Other a ers. 4. Sinauer 
Associ Sunderland, Massachus 
Taberlet, P., G. ae y, G. Pautou ps E Bouvet. 
Universal primers for amplification of three non- 
a reg a E d chloroplast DNA. Pl. Mol. Biol. 17: 
1105-11 
Be E. E. 1975. Relationships of Hedyotis fruticosa L. to 
Hou. sal and Oldenlandia L. Phytologia 31: 418—421. 
987. Carterella (Rubiaceae), new genus from Baja 
California, Mexico. Brittonia 39: 248—252. 
. 1990. Synopsis of Oldenlandia (Rubiaceae) in the 
E n Tirolen 68: 125-1 ue 
91. Overview and anm ted. e 
nud species of Hedyotis, px Oldent 
Eu and related genera. Pistols 115 19. E 
996. Revision n Houstonia (Rubiaceae-Hedyoti- 
m m Bot. Monogr. 48: 1-118. 
2001a. Today of Stenaria (Rubiaceae; Hedyo- 
ET anew genus including Hedyotis nigricans. Sida 19: 
14. 


199]. 


of s 


— ——. 2001b. Stenotis (Rubiaceae), a new segregate genus 
from Baja O Mexico. Sida 19: 899-911. 
——— 2001 ic review o ia acerosa and 
H. palmeri ueri notes on oe and Oldenlandia 
(Rubiaceae) Sida 19: 913-922 
——— W. H. Lewis. 1990. Oldenlandiopsis a 
e Caribbean basin, based 
a aa Grisebach. Brittonia e. 
185-19 
—— & H. Robinson. 2003. Survey of Asian and Pacific 
ele of eo and Exallage (Rubiaceae) with 
menclatural on Hedyotis types. Taxon 52: 
82 


———, W. H. Lewis, H. Robinson & J. W. ju. 1986. 
Phylogenetic implicati ions of diverse seed types, chromo- 
some numbers, and pollen e ogy in iid 
p Amer. J. Bot. 73: 1 

E. Robinson, W. L. » agne 
2005. jp n ri of genus 
Hedyotidinae s nid with emphasis on seed and 
fruit dg and notes on South Pacific necie Syst. 
Bot. 30: 

ud LD 


T. J. Gibson, F. Plewniak, F. Jeanmougin & 

3 ustalX windows interface 

Fail strategies for multiple sequence alignment aided 

by quality analyses tools. Nucl. Acids Res. 25: 4876- 

4882. 

Thulin, M. & B. Bremer. 2004. pidis in _ the tribe 
Pale el (Rubiaceae-Rubioideae 
tions of Amphiasma and Pent d and t 
Pn Pas Pl. Syst. Evol. 247: 2 

Verdcourt, B. Remarks on the qm of 

Rubiaceae. Bull. Jard. Bot. État Bruxelles 28 
ie 
6. Rubiaceae (part 1). Pp. 1-414 in R. M. Polhill 
(editor), a of Tropical East Africa. Crown Agents for 
Overseas PAAA ts and Administrations, London. 

Wagner, W. L., D. R. du & S S 
Contributions to > flora of Hawaii: 2. Begonia 
Violaceae and the SAU ledens. Bishop Mus. Occas. 
Pap. 29: 88-130. 


ohmer. 


Annals of the 


128 


Missouri Botanical Garden 


O60EPSNA = = (ug) 6SST poommy “eres meg Airy 2901p "(q 
“3810 +31 5 » 715-10 Tq “Y "f Dnjo1uo(q 
«680€pG(5I «ScO£ FG *TEGZPSNA (S) ¿c8€ 22u114 ‘oorxoyy uos A, "S (exo ^y) odmoru 7) 
«880€PSNA = xOG6TPSNA (49) STE T? t» uossforsnz) ‘epeueyens) "OG »pnudaoopoo 7) 
"IPIS Y cum?) paste) 
xL80€pG = x6Z6ZPSNA (Sd) IESP P 12 4nwo4g *eoujy ymog “yourorg (yoworg) əsuəfqsuvdmoz 7) 
«980€PS/1H «PZOEPSNA 8267 PSNA (sat) ZEZIZ8 vsseqjoy y Hd dopa “pony puey) 22 n3unaponb 7) 
- - +LZGTPSNA (Sd) IFEF P 12 sawasg *eougy ymog 
«S80€£PSAA @09LZ00AV - (49) org  puvasppoq *eoujy ymos owog (39q90]H) asuapnyou 7) 
pom) (deis) UMIUO]SOUO/) 
*LESZPINA = is (AN) £9g "JP 12 422uads; *ootxo]y 
(ug iw 
= @8SLZ00AV - apo) POG eyBsequeyan) uny ‘umouyun TPHYPS Cae) vgofnaer q 
xP80SPSNA xCGOEPSNA xSG6CPS NA (S) "ws saquoy ^io "uoSuj pumndaqpys y 
“SITES Dnipapnanosgp 
«COE q99£eeev «EZ6TPSOA (49) 18Pp 14915 openoq "pueis (sed 3» zmg) wumofruap y 
@SOSeSedVv @SSLZ001V i (49) 961 I? 12 uossiopuy *viquiopor) PPS Caed 29 zin) wmsogs y 
= POLELLAV - (OIN) £0& 7" 12 paoffois “ooo [Io (PPS) tunaooy (dias “y 
o£9€£eee4v OTOLLELAV 2262 VSOA (49) 38838 uossiopuy ap Fuysnpy “1openog woy Y fnSuwq 179011 cy 
= @OSEEESAV - (49) 2969 T» 1» Aodig *v[onzouo A IPS (puny) umprmu y 
« [80er wPSLZO0AV + IZ6ZPSNA (49) S6I& P 2 uossiopuy *erquiopo7) "pues Cppea) umonmu y 
coLSEELEAV OSELELAV = (AN) £Z0I 3ovpang miod [PUBIS r2puqopu y 
qS6eeeev OPSLELEAV x (AN) 2888 18mbuo.;) “eoty visor) umos “Y UNIDAD] y 
o£S£eeedv cSeeeelv - (S) PE9g o 12 wimpy umouxun "pues (AMOS ay WOY xo "p[[UA) sapio2142 y 
to ISELELAV “OSELELAV - (AN) 56888 `P 19 pavoj 1openoq "pues umor y 
@OPELEEAV @OPEEeedv - (49) SELOL UPH P door] “penoy "pues umpmisup y 
g amos Y “TOYS xo "PILA 100 4y dor hoy 
xO80EPS1A xOZOEPSNA xOZ6TPSNA (48) ZITI P 12 wassog *viquiez “yourarg, ('ummog y) sapromzn] y 
x6LOSPSNA wgSLZ00AV x6L6ZPSNA (S) osgg suay ^vjoSuy “youtorg (worp) asuapanzuaq y 
pug vusorydwy 
¥8L0EPSN «6 LOGPSNA x8T6TPSNA (49) 103 1? 19 wassagy *eiquie Wqosziopy (wong "y xo asqoopp) tunsoqo]3 y 
xLLOEPSOM «8 TOES xLI6ZPSNA (UD) 129 T? t» wassagy *exquiez Wqosziopy uafoq "y 
qosziorqy umuoypsnpms3y 
qu qua uou grsdi q?94-gdm UOreurmojur Joon A uoxe[ 


"(jsep e ir pexreur ore soouonbos Sursstpy "xsttojse ue qr poxreur ore seouonbos 
MON “BGQOS “TB 19 USOA (e fz00c “TE 19 uossropuy o 5666] "e^o  uossropuy (y zud-Tu4 pue ‘uonut grsda “yoqi-gqdy sxoxxeur pnsv[d sory) oy) 107 soouonbos pousmqnd. Apsnoraoad. woaz 
suoni OIEI] pue *sroqumu uorssoooe *(umrreqrou *roqumu 10}99T[09 *10129][09 uuo orgdex$09$) uoreurmogur 1oqonoA ir sosAqeue orouosSo[Aud oi ur posn exe) jo asr] '[ xmpuoddy 


129 


Groeninckx et al. 


Volume 96, Number 1 


2009 


Phylogeny of Spermacoceae 


x9€STPONA = x (AN) £966 puy “WS 
«916004 V PS6Z PENA (OW) 67-96 yrsaarysing “WS SRD vyofiãuo] y 
x6OTEPSNA @OLEEEEAV «ES6TPSAA (gi) "ws siounum] 3p puy “WS "| Panam “y 
"I DIUOISNO p 
«SOTEPSAA - xCS6CFS NA (49) OFOE `P 12 uossiopuy *euvmz) qouoxq “UNS “Y (apos gg Woy xo PIM) 1jofnucoo “y 
“MIIPS "y PIporpruo y 
«LOTEPSOA «TE0EPSAA + IS6Z HSA (Ud) ZIOT P 1 wassog *viquiez “youtarg, (UMS y) smuoopuads “y 
pug SASAÁYIA po 
«SOTEPSAA xTEOEPSNA xOS6ZPSNA (S) gg upung wz S1oquoqoppy *eipu] ymog «eod sepieraams qr 
«POTEPSAH +6 P6ZPSNA (S) 3805 puros y “eque] ug UWHJ, xo soymewqg DypAydounys “Ty 
«£OTEPSO «8 VOZ PENA (S) £97 P 12 wag “equeT us somenu] saumbumnb H 
xZOTEPSNA qL9Le004 V xL POCESNA (19) 9 4epunjpoy “yeqes ^eis&epe]y jdeig nivaysosonu H 
«TOTEPSAA xOEOEPSNA x9V6ZPSNA (S) 899g puros y “eque] ug UWHJ, z soyenu p uiSput “IBA DUDNI] “Y 
xOOTEPSNA x6ZOEPSNA xSP6ZPSNA (S) Tp Sioquogov]y *euvT us SONEMI], DUDTHOSSD] “IBA "UV DUDILOGSSO] “Y 
«PrOTPSNA (S) 0663 Homuny 3p yaoquoy “equeT us Sry 2D1UOS0D] "Hr 
x660EPSNA - *EP6ZPSNA (S) 2692F 81295041 “Spue[s] ouroxe7) "xosor (uoge y) sisuasou0y H 
860€PSNA = «2 VOTPSNA (S) ZZ vyoppAg 3 u0ss107 “eque] ug “1 peony H 
= «TPOZPSNA (S) 13801 ng Bu nus ‘Suoy Suoy SULH DAUMBUDSUO) `H 
*] suodpayy 
E «6G EOLAV - (ug) 695 |] 1» 320]g 2q “eosesepea IJL sopionmmaon 72) 
d1oxeg xÁmooudwuio-) 
«L60€vS n1 ¥8Z0EPSNA 6267 PSA (49) 1186 qos 3p rununpog “eunuosry paqe "T A CIpipuos 29 ue) soproriormdna “y 
060£PSNA @O6ZPOLAV ¥86TPS 1A (49) 966 1oroupaopoy y 1UuUDA ‘euNUEsIy odnyesioeg a parqe "T 7p (Suexdg) sisuarpisnsg 72 
JIS oup) 
«S60€vS n1 t £9Lc004 V «LEGTPSAA (49) 98 "JP 12 vaoy “eqny "^g spon `A 
"MS Dapouti 
«b60EPSA (687 P9LAV x9£6c V6 nl (49) "ws jounung *peprum], “unos ^x (Suexds) omjquim “y 
"PYA xo Med pzy oou 
xE6OEPSNA xLZOEPSNA «SE6TPSAA (49) 1961 T? 12 uosiopuy *euvmz) qouoxq "buy omoids q 
(pc94c004V (go) TZ06 Pp 12 uosiopuy “fuero UENGA MG PSOJUJUWADS (a 
«660€ PS(13 x9G0EPSNA xFe6cvcna (Sd) 6206 vir] susy “UNIS “y puuadsoop]mp ‘q 
penunop Ap[euonipen se "g miporg 
= WT9LZO0AV EL67 PENA (49) 868 Puig 'v&uox “yourorg (o100]N `g) aussany q 
“90191 g snj&jsouonjoniqug 
+ T60L£PSNA GOLELELAV «TEGZPENA (49) gogg uossaopuy “erensny 09104 72) Y 18104 "Y (CD suadas q 

Ju qua uoqut grsda q?947-gdm UOTDEULOJUI TOON A UOXe 


Penunuo) <q xipuoddy 


Annals of the 


130 


Missouri Botanical Garden 


xOTLEPSNA 69LZ00A¥ xCLOZPSOA (49) 8c I& P 9 uossiopuy *erquiopor) “ag (CT) ums W 
992500 V + TL6GPSNA (49) ZI6I P 2 uosszopuy *vuemz) gourg wequo y (Tqny) 9m yr 
"T xo sım Digoupnpn 
SZTEPSNA - xOL6ZPSNA (ug) SISZ nouos) twriezue], xeurorg nda202]80 "T 
"Xouraaq DA 97] 
«PZTEPSNA 6967 PSN0A (UD) 6£€ T» 12 yoorg aq “1eosesepela "xoure1q- (PLLA) vm24a Y 
«£ZIEPSAH «TPOEPSNA +8967PS0H (ug) 027 T» t» urossoqg *erquiez [PE ‘q Cumyog 7x) »mjnoraoaqns Y 
xOPSTPONA (AN) OF6¢ vaspoyy “ose vun[mg ‘PUPS Y wey) sisuaynsouas Y 
Prats sci xOFOETGM +LOG67PSNA (sat) SE06 om 'e&uox owog (wory) vqopreniqo Y 
*TZTEPSNA 6£0€PSNA «9962 PS0H (UD) 6FII P 12 urssog “erquiez owog Dpo2042nu Y 
xOZISTSNA «8£0€ T6135 +S067PSNA (ug) 697 T» 12 uassa(y *eoujy qmos Od »enjoupuko Y 
+6 TTEPSNA «L£O£ TG «DOG PSNA (ug) [gz T? t» urossoq *erquiez op&og nou1202 "y 
«8 TTEPSNA 0€0€PSNA *£967PSNA (ug) ZEP To 1» wassacy *erquiez uyog vsondsana "y 
xL TIEPSAA «S£0€£ T6131 «696cV S14 (Sd) ZOSP I" 12 sawasg *eoujy ymog YAZ x PPA vorquiipum "y 
"IPHS Y “WEY pammoy 
xShSZPONT «TESTPINA (AN) ES6ZT wmujiog ‘ersoudjog youery *pue[sp edey ag “4 sisuado "x 
9TTEPSNA SLELLELAV «T96ZPSNA (49) €84E[ umupog “M Aexo) "y pynasnd y 
«STTEPSNA x PE0EPSNA x0967PSNA (MAN) EZ80Z6 ouan] 29 uyary YOON TEMP “VSA "qe[ $1704071] Y 
«PPSTPONA xOSSZPONT (HsIg) ZF99 uDUpiag *rexopo]y ‘Terep “VSA uue ‘H 40]fixp] "y 
vuoro 29 IOUS A 
x€PhSZPONT xO6SCP9ONA (AN) 2291 420 ‘nye *nexepp “ys “TA OSqUOH H a 19088 A 7T M) n$4ogsof y 
«PTTEPSNA 66 7PSNA (S) ZZ oumdg ‘reme “VSA oouoroT WY TOUR y "[^ AV (S1oqso,p) vuns 8of y 
99uoIoT m 
xhScrona x8SStPONA = (HSI@) [g9sT wumupog Tenes *reweg “VSA TouSe y T M uoo WY TOUTE T AJ) muuAmf y 
«TPSZPONA *LESTPONA i (AN) ZFZI pon neo emey “W'S SOCIO "N 7D) emmanjf "y 
xOPSTPONA «9c Sc TOt (HSID) 0S€9 224304 *renvy Teme H “ys oouoroT WY 19USE A, "[ “Ay (VUEN 7H) 401702 "y 
e€1 Teron @ILeeesav +86 7PS0H (19d) Z908 poop “no oouo10T Y TOUTE “T “AA (S10q504) 1ouosop "y 
xOSSTPONT xS@SCPONA (AN) £0z1 pon ‘Teme “YSA oouoroT WY TOUTE y "T “Ay (MUS 7 ^f) vavo? "x 
«2 TTEPSNA @OLESESAV xLS6ZPSNA (941d) 1308 dn epo TPS Y "um pmpa "y 
«TTIEPSAA +££0€PSNA 086 7PSNA (S) 8849 Buaqsnoys *ueweg “ys "UY Ip OOH saproyrunsquas “Y 
aSSZPON #PCSTPONA (AN) FeZT non ‘men teme “VSA 
@S9LZ00AV - (go) "ws 2u8029-uosiang TEME "ven oouo 29 1oUSE A "T^ A, mey) SLDPPIXD y 
«8 ESZP9NA *EZSZPONA (AN) £gz1 pon ‘Teme “ys TPS Y um) sufo y 
«OTIePGO «S66z l1 yg æ apo Teme H ^en PUPS Y cum pgmunanop -y 
"IPS Y “UR pnpox 
Ju qua uonut 9 [sde q?94-gdm UOTJEULIOJUI TON A UOXP], 


penunuo) <q xipuoddy 


131 


Groeninckx et al. 


Volume 96, Number 1 


2009 


Phylogeny of Spermacoceae 


5 @E6TPOLAV ¥866CPS 1A (san) rp 79 12 sqoonf'-uasunf’ “wesna “UNDE “Y sinua () 
90TEPSNA «ZODEPSAA «L6GTPSAA (ug 18 m9) gz 2usuaqueqop;) un, *uede[ eziuny (oum[g) 010/1991 `Q 
xOPLEPSOA ¥966CPS NA (ug) eror 7» 12 poospig “etuezueL, JPW srsua10Qp3 `Q 
x8PIEPSNA POPOLAY «S66TPSAA (sam PISST 4opmg 'pzexg "xoef (pg xo 3 OOH 2» pug (90) wuunuzyns Q 

x€POCPSNA (UD) Z6II T? 12 urassag “erquiez “unos "x DIPIASO “Y 
«LPTEPSOA xTOIOEPSNA i (ug) 9rg I» 1» urossaqg *erquiez "nq muosuiqoa “CQ 
TOLLELAV = (19) GEE uossp Y uossiopuy ‘uoqes WIH Dsodiou “(y 
x090€PSN «D66TPSAA (ug) FZ6 P 1» wrassacy *eiquiez owog synooomuau “y 
xOPTEPSNA xE66TPSNT (ug) 9TST poommwy *eqeusny TEON "4p Seprowonspsnu =p 
xSPIEPSNA x6SOEPSNA xZ66ZPSNA (S) 189 UAMH 2p wiguisopos ONKIN “OC (IPPS 29 Wep) vayron “CG 
x PPLerSod ¥8S0€PSN4 xT66ZPSNA (UND) 9SET P 12 urassag *viquiv7 “Od Cyprus) vyofioun] y 
«EPTEPSIA xLSOEPSNA xO66ZPSNT (ug) gor I» 1» urossaqg *eiquiez DIDY “XA *qxoy (7T) vaooquay y 
xCPlLersnd ¥9S0€PSNA «686c PSA (ug) SPR I» p urossaq *eiquiez owog 1277202 “Tea "qxo (T) vaenquey `Q 
«TEIgPSOM xSSOEPSNA x886cPSNA (4D) 9851 I" 12 urassag “erquiez apreumung (9q) si9u22405 “y 
+OPTEPSNA «PS0LPSNA xL86cV 613 (UD) SE6 I» 1» urossaq *erquiez "qeurerq opnjdoo “CQ 
«6g Ter x€S0EPSNA +986cPSNA (ug) IISI poonang *eqexsny TOYAN “A CIPON A) sapiens y 
+8£TEPSNA *ZS0€PSNA «S867PSO0H (ud) 6101 P 12 urassag “erquiez SUI DIDIS1SDf '() 
*LETEPSAA «V 86c EG (ug) £661 vjowuaqny 3 oquokny “eruezuel, “Opie A (WH) opironjjod rea ug "x »sopmungoo '() 
«9g Terent xTSOEPSNA xES6TPSNT (ug) 8&6 I» 1» wəssaq *erquiez “UNDE ^x psopmunoa “y 
«S£ IePS n xOSOEPSNA «C86c (1 (ug) Zgp P p urossaq *erquiez "| »soqui402 “Q 
«PETEPSNA *6POEPSA «I86c G5 (ug) "ws qo 12 Oquokoy *viuezue], owog oyodasora]d xea sisuadoo “CQ 
«£ETEPSNA «8 POEPSA «0g6c G1 (ug) SFE I» 1» urossaq *erquiez sisuadoo "es ^j "[ sisuadoo '() 
«TETEPSNA x6L6CPSNA (Hd 1 aquo) Eg aysuaquayany un, *uede[ "T 940]fiq “O 
«TETEPSNA «LVOgPS(M «8L6cVG13 (UD ZE6 I? 1» urossaq *erquiez "ums "x sisuajosun ~Q 
«O£TEPSna OVO0LPSNA *LLGGPSNA (ug) 229 I» p urossaq *erquiez “Od OPs y woy) suff <0 
"I mpupjuə pjo 
- @L09E00AV i GD «168 aspyg “uo "eurexq- (qxoy) mo4oqu “y 
"Xouraagq CJ "490H) suo&poqosoAr 
«6c LePS n «SbOerSnd «9L6TPSAA (UD) SZZ e 12 urassag “erquiez aof sapanponb “py 
1940[ sisdowuoDspapipy 
«PPOEPSOA «SL6ZPSNA (49) $827 P P sqooof-uasunf *euemo umqog "x smuuodsouonu “py 
«ge EPS 00217004V «bL6CPS1A (49) S66T 1P 1» uosszopuy *euemz) qouexg "unos ^N (IMYIS 29 “SOY xe PIA) $np1St4f W 
Jd “Us » “muss XƏ MELLA sndinonaqiyy 
«LZLEPSOA xChOerSnd «EL6TPSNA (ug) sog I» p wiassacy *erquiez SUMEN "S "qp pyofnaot py 
ypsuraq DÁYIDISOUDH] 

qu qua uomur gysda qoqi-gdm UOTJEULIOJUT IAMO A uoxe[ 


Penunuo) =“, xipuoddy 


Annals of the 


132 


Missouri Botanical Garden 


"opio A xo Jdvig 40]ftaapd ^q 


«ce [eren «PO0EPSNA +TOOEPSNA (Hd) 829 P 7» urassag “erquez 
“ATE DISIUDJUO d 
= xEZOEPSNA x9c6cV6[19 (ug) 825 I» 12 320]g aq “Ieosesepe "uw ssuoLmospSppput 7) 
‘ssn vapydiny 
«£80€PSH «TOSS «PZG6ZPENA (ug) F9g I» p urossaqg *erquiez "aqqoy vrnyauraynd ‘g 
"IPTA puipodomg 
VXVL d')042100 
xLOTEPSNA xSLOEPSNA «9 LOShSNA (D 323 upd 3p. sopianzv] “erpensny «oon CIPON 74) 227020771 ^g 
j 90H piupidvuAg 
x9OTEPSNA @ELEEEEAV «STO£ TG (OW) 6-96 ysaarysing “WS ][ox1o], (we) sunaiusie ‘S 
[pax L Cry) DLIDUIIS 
+SOTEPSNA x PLOEPSNA «PTOEPSNA (19) o6g& uossp Y uossiopuy “uoquo “OC erigens *g 
«POTEPSNA - «£TO£ TG (49) 9108 `P 12 uossiopuy *euvmz) qouoxq "ue P034 `G 
«E£9TEPSAA = *ZITOSTSAA (19) 8203 7? 12 uossiopuy *erquiopo7) qy vipysosd ‘S 
«CO T£ TS +€LOLPSNA «TTOEPSNA (S) 2993 "pe 1» dioyuup yy *exueq us "| opidsnj `Ș 
«09 T€ TS xTLOSPSNA x600EPSNA (sat) Zzo6 opm] susy "oproA (unos x) »qnnpf "s 
«TOTEPSNA TLOLPSNA «0 TO£ TG (ug) F6Z I» 12 Y20]9 aq txeoseSepv]y "rog stuuofinasoy "e 
*6STEPSNA 0L02PSNA «800€ T6135 (ug) SFIT poonaog *eqexsny poomuv[] ns042 ‘Ç 
= MOTIEOOLY x (19) PL0G JP 1» uossiopuy *erquiopo7) S9 x9 e[puoy nsnfuoo "S 
«86 T£ TS 6908S xLOOEPSNA (49) 8061 vossiapuy “euemo Youg "aug X zmy vpydo ‘ç 
"I a900puuodg 
«LSTEPSAA +890£PS0H 900€ PSNA (49) SFEgse poss *epeusny “pros (IPPS 2p "ure sugars Y 
OSTEPSNA ÍÚPTOL00AV «S00€PSNA (49) €20G T? 1» uossiopuy *erquiopo7) Up P4qpos “y 
"I DIPADY ITY 
«SG TIETS(M @COCPOLAV «PO0EPSAA (ug) 079 I» 12 Y20]g 2q “eosesepea Pod (pos a Wy xo PM) sisuenmosp2opprui “gq 
= «LODEPSA «£00€ T6135 (Sd) £gzg wag *voujy ymos Pod (SPPON) vrouw ‘q 
gug Papiyo Ayd 
xPSTEPSNA +990£PS0H «200€PSNA (ug) gos I» 1» urossaqg *erquiez MPA (Vuo, p "umog y) srapuppaad ‘q 
“spo uopojuoq 
+ESTEPSNA «S90€PSNA = (sam) 8epz I» 12 vagy ‘erdonngy Spuy sunny “q 
appuay sisdoupjuag 
* TSTEPSNA *E90€PSNA 000€PS0H (Sd) 39£8 2477 ap ar] esuoy “unos "N muupuopaum Q 
«89TEPSNA «OLOEPSOA xl TOEPSAA (Sd) 0983 I» p snag ‘erdonngy uror] xo “JUIOMUDS SISUINDA “CQ 
«O6 TE TS @S6TPOLAV x666ZPSNA (49) 89525 «apo “VSN "T v4ojfiun y 

AU ua uomur 9 rsda q?97-gdm UOTJEULIOJUT IAMO A uoxe], 


penunuo) ‘q xipueddy 


FOLIAR AND PETIOLE ANATOMY  Dorismilda Martínez-Cabrera,? Teresa Terrazas,? 


OF TRIBE HAMELIEAE AND 
OTHER RUBIACEAE? 


3 


and Helga Ochoterena 


ABSTRACT 


In this study, folia and y l n 
can be used to Mieres A» Pu D Nu and taxono 
of 36 species, which were sectioned u 


ing conventional embe 


everal genera was compared to determine whether there are characteristics that 
omic position 
ddin 


of Ham 


elieae (Rubiaceae). Our Vp included a total 
ng and staining methods. From 


these species, 23 represented 


six of the Pa s E A Hamacas sensu Bs meneame NÉ was included in n order to a its putative 
d 


inclusion wit ENG purposes, th 


melie: 
(Rubioi Ph de. Bor PA s 


the following characters: cuticle more than 3 um, do 
spongy p hyma, raphides, tannins, and vascular tissu 
Brandegee is un having fiber: 


orc 


f tribes 
p (Ci nerds), and A (Ixoroi deae). on M indicated that folia 
can be used in system: 
orsiventral m p a single palisade SOR M 
eo n the 


matic studies. Members 
cell layer, loose 
etiole. Plocaniophyllon 


types I, IL, or HI 1 midrib and p 


1 with major and minor veins. "The petiole vascular Pu s an open arc shape 


in all studied species except in Randia L., Sah has a closed cylinder. Hamelieae, Syringantha, and Psychotria L. have 
fr hav 


raphides, one layer o 


and loose spo 


ngy parenchyma, while the other taxa 


e druses, two layers of 


palisade parenchyma cells, and compact spongy, ae id Salisb. is unique, showing both raphides and druses. 
h 


Our results show 


and phylogenetic studies within Rubiaceae. . Ther e was anato: 


2n support 


] characters to be considered in taxonomic 
t for the inclusion of Syring i i 


while the presence of o and the mesophyll attributes suggest an understanding of why Hamelieae was morphologically 


treated as a member of R 


ords: Bouvardia. a fibers, 


leaf trace, Plocaniophyllon, 


raphides, Syringantha. 


The taxonomie usefulness of leaf and petiole 
anatomical features for recognizing genera or circum- 
scribing generic, tribal, or family level is well known 
Lr p Chalk, 1950; Smith & Stern, 1962; Baas 
& K ; Wilkinson, 1983; Dickison, 1989; 
Engel, o Mentink & Baas, 1992; Bane 1995; 


Aoy 
2003; Dos Reis et al., 2004; Enn et al., 2004; 
Andrés-Hernández & e 2006). However, in 
Rubiaceae, which includes nearly 13,100 species 
classified into 611 genera (Govaerts et al., 2006), only 
a few studies have examined foliar and petiole 
anatomy (Metcalfe 7 ~ 1950; Herman et al., 

6; Kocsis et al., 

Robbrecht (1988, pu proposed a classification 
that divides the family into four subfamilies: Anti- 
rheoideae, Cinchonoideae, Ixoroideae, and Rubioi- 
According to phylogenetic n 


: sed o 
molecular data, Ant ey ee is D while 
Rubioideae has been the most S d Verdcourt, 


1958; Bremekamp, 1966; Bremer, 1987; Robbrecht, 


1988, 1993; Rebbrecht & Manen, 2006). The tribe 
Hamelieae was traditionally placed into the subfamily 
on 
raphide 


largely because of the presence of 


ever, phylogenetie analyses based on 
c pee, which included a few represen- 
tative species of this tribe, indicated that Hamelieae 
would be more appropriately classified in Cinchonoi- 
deae (Bremer et al., 1995; Andersson & Rova, 1999; 
Robbrecht & Maren, 2006) 

The circumscription of the tribe Hamelieae itself 
has been controversial. Classifications have included 
between tw peus ri and Hoffmannia Sw. 
[Verdcourt, 1958, 6; Bremekamp, 1966; Elias, 
1976] and 11 genera (Alibertia A. Rich. ex DC., 
Axanthes Blume, 
Hamelia, Olosty 
Gaertn., . 
and Schradera Vahl [de Candolle, 1830; Endli che, 
1836]. More isa Bremer (1987) redefined the 
tribe on the phological characteristies, the 
most a of sich. were inflorescences ebracte- 


Evosmia Kunth, 


ate or with very small scale-like bracts, alternate or 


1 The senior author thanks the Consejo Nacional de Ciencia y Tecnología for a scholarship (159282) to conduct her doctoral 
studies. We are grateful to the curators of CHAPA and MEXU for loaning us material for this o 2 are also extended 


and Guerrero 


s.m 
* Departamento de Botánica, Instituto de Biología, Universidad Nacional Autónoma de México, 70-233 México D. F. 04510, 


Mexico. tterrazas@ibiologia.unam 


doi: 10.3417/2006196 


ANN. Missouni Bor. Garp. 96: 133-145. PUBLISHED ON 23 APRIL 2009. 


134 


Annals of the 
Missouri Botanical Garden 


right-rotated Dosen imbricate corolla aestivation 
and tetramerous flowers with usually 
bicarpellate pistil. "With this definition, the tribe 
includes Deppea Cham. & Schltdl, Hamelia, Hoff- 
mannia, Omiltemia Standl., and Pinarophyllon Bran- 
degee. This delimitation was aecepted by Robbrecht 
(1988). Subsequently, Robbrecht (1993) revised the 
classification of Rubiaceae and, based on comments 
by Lorence and Dwyer (1988), added to Hamelieae 
the genera Eizia Standl. and is phyllon Brande- 
11 (1996) revived the 

monotypic genus Syringantha Standl. (considered by 
Robbrecht [1993] as a synonym of Exostema (Pers.} 

Bonpl.), asserting that it is closely related to Hamelia, 
with which it shares the hides and 
floral morphological characteristics, which suggests 


gee. A few years later, McDow 


presence of rap 


that Syringantha merits inclusion into the tribe 
Hamelieae. More recently, based on a supertree 
analysis, Robbrecht and Manen (2006) amended the 
tribe Hamelieae to include Hillieae (Cosmibuena Ruiz 

av. and Hillia Jacq.), Chione DC. 


(formerly 
classified within Chiococceae = ] 


Portlandia group), 
and Cosmocalyx Standl. (formerly considered as 
incertae sedis). Nevertheless, in Robbrecht and 
Manen's (2006) proposal, 


analyzed DNA sequences 


only the genera with 
were mentioned, leaving 
out many taxa considered in previous classifications, 
among which are four ae included by Robbrecht 
3) as part izia, 
Pinarophyllon, 


Omiltemia, 
M e plus 
gantha. For this reason, the current most comprehen- 
sive proposal for the classification of the tribe is that 
of Robbrecht (1993) and this is the one followed in the 
present stu 

In this study, the utility of leaf and petiole anatomy 
in the systematics of Hamelieae is evaluated. Six 


meli 
genera classified by Robbrecht (1993) within the tribe 


e Cinchoneae, subfamily Cin- 
chonoideae sensu tese [1988]) and a synonym 
ht (1993) was included in 
order to reevaluate McDowells proposal (1996) to 


of Exostema sensu Robbre 


resurrect it and classify it within Hamelieae. For 
comparative purposes, other genera of different tribes 
n i d 


subfamilies were also considered in the study 


MATERIAL AND METHODS 


To evaluate the utility of leaf and petiole anatom 
for generic and tribal classification, complete leaves 
of 23 species representing six genera (Appendix 1) of 
Hamelieae sensu Robbrecht (1993) were described: 
Deppea (11/27 spp.), Hamelia (5/17 spp.), Hoffmannia 
(6/111 spp.), Omiltemia (2/2 spp.), Pinarophyllon (1 


sp.), and Plocaniophyllon (1 sp.). We sampled more 
than 25% of the recognized species for each genus in 
Hamelieae, except Hoffmannia, being careful to 
represent the described morphological diversity of 
each genus. For comparative purposes, 13 species of 
the following genera were also considered in the study 
(Appendix 1) Exostema, because of the previous 
consideration. of Syringantha as a synonym of it; 
Hintonia Bullock, as potentially related to Exostema 
both belonging to the informal Portlandia group 

within Cinchonoideae); Psychotria L. (Peychorieae), 
as a comparative stable member of Rubioideae; 
Randia L. 


member of Ixoroideae; and Bouvardia 


(Gardenieae), as a Mu ect stable 


s 
POR genus with unstable subfamilial po 
Rubioid 


ub1io1deae 


or Cinchoneae—Cinchon- 


oideae). 

Leaves of those species were collected in the field 
and fixed in formalin-acetic acid-alcoho A) 
(Ruzin, 1999). From Hamelieae sensu Robbrecht 
1993), only Eizia is not included in this study 
because we have not been able to find it in the field, 


— 


and this monotypic genus with a restricted distribution 
in the state of Chiapas, Mexico, is only known from 
the type collection. We attempted to include repre- 
sentative species of Cosmibuena, Cosmocalyx, an 
Hillia, which were considered as Hamelieae sensu 
Robbrecht and Manen (2006), B we have not been 
able to find them in the field and it was impossible to 
section leaves from herbarium materi 

Voucher specimens of the newly cotes material 
were deposited at MEXU and CHAP a 
(acronyms following Holmgren et al., 2004). 


a 


addition to the personal collections, mes Re 
herbarium specimens (Appendix 1) were used to 
complete the sampling and to increase the number of 
individuals sampled per species (two or more 
individuals were sampled per species). Leaves from 
the herbarium material were rehydrated in 5% Na 

at 60°C for 1h. and fixed in FAA for 24h. All 
dehydrated in i 
TP1020 c 


oned (transverse and parader: 


samples were 
Germany) automatic 
paraffin, sectio 

thick with a rotary microtome, stained with safranin- 


fast green, : s (Carolina 
oe E North Carolina, U.S.A.) synthet- 
. Terminology follows eae (1979) and 


Mentink and Baas (1992) for leaf, Wilkinson (1979) 
for cuticle, and Howard (1979) for petiole anatomy. It 
should be note 


vascular tissue was described have used the terms 


d that previous studies in which petiole 


“main vascular bundle” and “lateral vascular bundle” 


respectively (Metcalfe & Chalk, 1950; Herman et 


Volume 96, Number 1 
2009 


Martínez-Cabrera et al. 135 
Foliar and Petiole Anatomy of Hamelieae 


al., 1986; Kocsis et al., 2004). Because the position of 
the tissues is the same, we consider them homologous. 
In this paper, therefore, we use the terms "centra 
trace" and "lateral bundles" 


(1919) 


proposed by Howard 


RESULTS 
INDUMENT 


Simple unicellular and/or multicellular trichomes 
are present in most studie i Pl j 
phyllon has glabrous leaves. Within Hamelieae, the 


species, only ocadnio- 


simple multicellular trichomes are the most common, 
and unicellular trichomes are present only in Hamelia 
(Fig. 1A), where they are 27-48 um long and thin 
walled (< 2 um). Outside Hamelieae, Syringantha 
and Bouvardia also have simple unicellular ee 
being exclusively of this type in the first genus. The 

simple multicellular eee have up to jdm cells 
in most members he tribe, but in Pinarophyllon 
they have 10 to 11 cell (Fig. 1B), a range also present 
only in Hintonia. Trichomes are commonly present on 
lamina, midrib, and margins (Table 1). They have 
mostly thin walls (< 2 um), but within Hamelieae in 
Deppea, simple multicellular trichomes have thicker 
walls (3—5 um), and Syringantha and Bouvardia have 
the thickest walls (5-6 jm) (Table 1). 


CUTICLE 


The cuticle may be smooth (Fig. 1C), but cuticular 
striations occur frequently, and both can be found in 
8 enera of Hamelieae, as well as Deppea and 

Hoffmannia, and in Psychotria and Randia of other 
tribes (Table 1, Fig. 1D—F). Cuticle thickness is less 
Hamelieae; genera of other tribes have 


thicker cuticles (Table 1). 


EPIDERMAL CELLS 


In surface views, unspecialized epidermal cells 
generally show the entire range from straight to 
varying 
of Hamelieae ui other tribes. 


undulating anticlinal walls (Fig. 1G, H}, 
among the genera 
Straight walls are common on the adaxial surface, 
while bad walls dominate in the abaxial 

enera, y Hamelia and 
Pinarophyllon M: Hamclieae i T Hintonia, 
and Randia of the Portlandia group and Gardenieae. 
In transverse section, the epidermal cells are mostly 


square or rectangular on both surfaces, or exclusively 
ellipseid shaped with a short dome in the outer 
periclinal cell walls in the cases of Hamelia and 
Pinarophyllon within Hamelieae (Fig. 1A, B), and in 


Exostema, Hintonia, and Randia outside Hamelieae 


(Table 1), whereas Bouvardia has a nipple-shaped 
papillae in adaxial surface (Fig. 1F) 


THE STOMATAL COMPLEX 


omata occur on the lower epidermis and a 


e 
Mum Mostly paracytic stomata (Fig. 1D wih 
two to three subsidiary cells occur among the sampled 
o (Table 1), but DR stomata (Fig. 1J) 
also present in Hoffman Omiltemia, and 
Pinarophyllon and are el in ni Plocaniophyllon. 
s cells show straight or undulating walls 
(Fig. J). However, in Deppea, only the larger 
c cells have undulating walls. Stomatal size 
mostly varies Mn 30 and 45 de in length. Giant 
stomata are pres n fou a of Hamelieae 
(Hoffmannia, On ont Pinarophyllon, and Ploca- 
niophyllon) and in Psychotria (Table 1 


MESOPHYLL 


studied showed  dorsiventral 
e palisade tissue consists of one adaxial 


layer in Hamelieae (Fig. 2 


All the material 
mesophyll. T 
and Syringantha, one to 
two adaxial layers in Bouvardia and Psychotria, and 
two layers in the other genera (Table 1, Fig. 2B). The 
spongy zone varies from two to six cell layers. Cell 
arrangement is loose in all material studied from 
Hamelieae, Syringantha, and Psychotria, 2 compact 
in the remaining studied species (Table 


CELLULAR CONTENTS 
Crystals are present exclusively as raphides in the 
nd midrib in genera of Ham 


mesophyll elieae 


— 


Fig. 2C), Syringantha, and Psychotria. Druses occur 
in a cells, idioblasts, in the meso 
ema and Randia (Fig. 2D), while druses and 
de are present in Bouvardia (Fig. 2E). Dark- 
staining deposits are also common in the mesophyll 
1 and are sometimes associated with the 
phloem of major and minor veins in Hamelieae 
(Table 1); they are present but scarce in Syringantha 
and Psycho and dark- 


staining deposits to particular tribes suggests the 


tria. The restriction of crystals 


potential taxonomic value of these features. Only the 
leaves of Hintonia lack dark cell contents. 


MIDRIB 


Hamelieae, the midrib is commonly raised on the 
abaxial surface and grooved or raised adaxially as 
seen in transverse sections (Fig. 3). The cuticle has 
similar characteristics to those of the lamina, but 
reaches higher (up to 6 um) in Deppea and Hoffman- 


nia; outside Hamelieae it is also thin in Psychotria 


136 


Annals of the 
Missouri Botanical Garden 


e l. Foliar anatomy of the 
unicellular trichome. —B. Pinarophyllon (D. Martínez et al. 317, CHAPA), simple multicellular cy —C. De 


Ochoterena 335, MEXU), smooth cuticle in adaxial epidermi: 
mis. —E. Bouvardia (J. Rzedowski 30784, MEXU), cuticular striations in abaxial epidermis. 
Bouvardia (J. Rzedowski 59784, MEXU), cuticular striations in adaxial ae 


striations in adaxial 


epiderm 


tribe Hamelieae and other Rubiaceae. —A. Ham 


elia (H. Vibrans 5885, MEXU), s 


eppe 
. —D. Syringantha (D. Martínez 190, CHAPA), cur 


2198, MEXU), adaxial epidermis with straight anticlinal cell wall. —H. Hoffmannia A. og & M. A. Heath 154, MEXU) 

adaxial epidermis with undulating ec cell walls. —I. Exostema i Ochoterena 289, MEXU), paracytic stomata. —J 

Omiltemia (M. E Heath & A. Long 1173, MEXU), mac stomata. Scale bars: A = 50 um; B-F, RJ 
um; H = 30 um. *, cuticle; e, epidermis; p, palisade parenchyma; s, spongy parenchyma; t, trichome. 


and thicker and more variable in members of other 
tribes (Table 
cells are ellipsoid shaped with a short 


). In transverse sections, the epidermal 
ome in the 

outer periclinal cell walls in abaxial epidermis and 
periclinally elongated in adaxial epidermis in most 
genera, but exclusively ellipsoid shaped on both 
surfaces in Hoffmannia and Pinarophyllon, and in 
Psychotria and Randia. angular or lacunar 
collenchyma consists of two to eight cell layers below 
the abaxial epidermis and two to 10 cell layers in the 
adaxial epidermis (Fig. 3, Table 2). The vascular 
undles are collateral. The variation in the vascular 
tissue of the midrib was classified into five types 


(Table 2). In Deppea, Exostema, Hintonia, and 
Bouvardia, it is a simple open are DE I, Fig. 3A). 
In Hamelia, it is also a simple o re, but with 


invaginated EM and a pair of mes helo (type II, 
Fig. 3B). In the remaining genera of Hamelieae as 
well as in Syringantha, it is a simple open are with 
slightly curved ends and a pair of lateral bundles (type 
II, Fig. 3C). In Psychotria, the vascular pattern has a 
V-shaped open arc with invaginated ends and four to 
six lateral bundles below or between the invaginated 
ends (type IV, Fig. 3D, E). Randia is unique in having 
a closed ring (type V, Fig. 3F). The vascular tissue 
has xylem cells in radial rows separated by paren- 


137 


"juosqe *— *1uoso1d *4- fouroorn re[iqpoorun ‘n tsutoa Lenia 9 wys 9s ‘orenbs ‘bs ‘yjoours “ws tsura Árepuooos ‘s trepn3uejo91 ‘or ‘soprydes ‘ex :onAoo[o[[exed ‘d turoopud. ‘yd 
*on&oered ‘od squaprat ‘aur 5[[Agdosour ‘our sursreur *eur woy rem[[oo nur w pasoo] ‘oy *eurure] Y teurure[ pue stuttoprde usom3oq prosdr pe “e ‘stueprde *o tsosnap ‘ap 5oeduroo ‘oo ‘ore “e 


Foliar and Petiole Anatomy of Hamelieae 


Martinez-Cabrera et al. 


our Ip “ex - 02 

9 “qu Ip = 02 
ou el + o[ 

- - - 02 

ou Ip 2 02 
oul el + o[ 
yd ‘our D E o[ 
e ‘yd ‘our D + o[ 
e ‘yd ‘our el $ o[ 
e ‘yd ‘our D E o[ 
yd ‘our Bl ES o[ 
e ‘yd ‘our D + o[ 


o Sans 


=p 


od 


bs ‘Jo 


o1 ‘bs 
I? 
bs ‘fe 
bs ‘or 
I? 
bs ‘fe 


E 


pipipanog 


3BINSUIBH 


uonnquusi[ spsg = suey, eulyousred 
ASuodg 


(uni 
y susT 


= 


Volume 96, Number 1 


2009 


Sju91uo?) 


Eewo 


jerxeqe 


pue penepy 


odes jpe? 
[euixoprd^q 


LIU) 


"e1ouoS oAnvjuoso1do:r IJO pue *nurupsung *owor[oureg] A Jo Áujeue Jeroj Jo S1ojoereqD) T AJEL 


138 


Annals of the 
Missouri Botanical Garden 


Foliar anatomy of the Hamelieae transverse section: 


Pus saltus parenchyma cell layer. —B. Ra 


E. Bouvar us Tenorio & C. Rom 
et al. 316, CHAPA), tannins in spongy pai 


—A. Plocaniophyllon (D. Martinez et al. 312, CHAPA), 
ndia (H. Itis 29632, CHAPA), two palisade parenchyma cell layers. —C. 
m Ss Mora 5314, dd xo in mesophyll. — 


i (H. Ochoterena 224, MEXU), tertiary vein. —I. a (D. Martínez pe CHAPA), quaternary vein. Scale bars: 


,D, E = 10 um; C, H = 30 um; F = 40 um; 


m. bs, bundle sheath; dr, druse; e, epidermis; p, palisade 


te ieee ph, inen ra, aphid s, spongy AO UG ta, tannins; x, lbs 


chyma cells with dark-staining deposits. The vascular 
bundle is enclosed by an 


distinctive dark-staining deposits in Hamelieae and 


arc of parenchyma cells with 
Syringantha, except in the case of Plocaniophyllon, 

where the arc enclosing the vascular bundle is 
composed of fibers. In members of other tribes, the 
arc enclosing the vascular bundle is more variable, 
with collenchyma or sclerenchyma (Table 


MAJOR AND MINOR VEINS 
Major veins are mostly similar to the midrib in their 


anatomy, with a bundle sheath of parenchyma cells 
(Fig. 2G, H). The minor veins are also collateral and 


have one to two tracheary elements and a sheath of 
H 


exclusively parenchyma cells amelieae and 
Hintonia, Psychotria, and Randia (Fig. 21) and fibers 


in Plocaniophyllon. 


PETIOLE 


In transverse sections, the petiole is mostly round at 
the base and winged toward the lamina, except in 
Deppea in which the petiole is reniform at the base, 
and in Pinarophyllon in which it is winged along its 
entire length. Indument, cuticle, and epidermal cells 
were observed to be mostly similar to the lamina in all 


material studied (Table 2, Fig. 4A—E). The number of 


Volume 96, Number 1 Martínez-Cabrera et al. 139 
2009 Foliar and Petiole Anatomy of Hamelieae 


RY y 
IES Y 
E LS 


> 


22 
Le 


Y 
> A 
» s 
W AXA 

es SN 

MA GER 

A S 

WY io NN | 
ISR WM 

AA RANA NS ors 
IIA 


RE LIS 
AA HD 
4 SOS x 


es 
bes 
eS 
Bes 
be 
RS 


AS 
RAS 


FZ om > 
LEO ERO OOS 

ROLES ER S 
ORE IY BER = 


RSS 


> Á eoe 
N K fo 
A E Ke 
» 2e, 
b SEL 
NS E] REO 
Me edere NA REELS 
SSI NN MER 
SS ERR 
RSS EL SIRIA 
RSS iS RRS 
RSS euet 


ce 
3 


A 
2 POOR: 
SRE 
RES 
REE 
Xp 


Figure 3. Blade midrib of the Hamelieae and other representatives: vascular tissue shape and distribution types. —A. 
Exostema (D. Lorence 3036, CHAPA), type I. —B. Hamelia (Maya 1868, CHAPA), type II. —C. Omiltemia (A. Méndez 285, 
MEXU), type III. —D. Psychotria (Magallanes 3687, CHAPA), type IV. —E. Psychotria (S. Maya 1753, CHAPA), type IV. 

E : hate ine = collenc 


—F. Randia (E. Domínguez & H. Ochoterena 1775, MEXU), type V. Symbols: hatched line hyma; continuous, thick 


black line above phloem — fibers; white — parenchyma; narrow black line above phloem — parenchyma with tannins; dotted 
line — phloem; di ü black line above phloem = sclerenchyma; vertical lines = xylem. Scale bars: A-E = um; F 


= 100 um. 


angular or lacular collenchyma cell layers is variable Figs. 4H, 5F). The central leaf trace has a sheath of 
(Fig. 5A-C), with three to five layers being the most — sclereids in Psychotria, fibers in Randia (Fig. 4D, or 
common, but up to 12 layers are present in Deppea the sheath is absent in members of other studied tribes 
and Pinarophyllon. In some species of Hamelia and in (Table 2). Cellular contents are similar to those of the 
Syringantha and Randia, the vascular leaf trace was mesophyll and midrib (Tables 1, 2, Fig. 4J—L), except 
enclosed by collenchyma. The range of vascular leaf in Hintonia, which has druses in the collenchyma. 
trace patterns in the petiole was classified into six 

types and is schematically illustrated in Figure 5. In Discussion 

Deppea and Syringantha, the vascular pattern has only 

an open arc and a pair of lateral bundles (type I, Leaf and petiole anatomical characters are quite 
Fig. 5A) with a continuous parenchymatous sheath homogeneous among genera of the tribe Hamelieae. 
with dark-staining deposits occluding the cell lumina There are, however, certain exceptions, e.g., the genus 
(Fig. 4F). In Hamelia, the open arc has invaginated Hamelia, which is characterized by type II vascular 
ends and only one pair of lateral bundles (type II, tissue in the midrib and petiole, and unicellular and 
Figs. 4G, 5B). Toward the lamina, the central trace is multicellular trichomes, and the genus Plocaniophyl- 
more curved toward the base of the petiole, with the Zon, characterized by the straight anticlinal walls in 
lateral bundles in the same position. The most abaxial and adaxial epidermal cells, parallelocytic 
common leaf trace pattern in Hamelieae consists of stomata with three to four straight-walled subsidiary 
an open are with slightly curved ends, the central leaf cells arranged in a C shape, no trichomes, and fibers 
trace, and two pairs of lateral bundles (type III, sheathing the midrib and other veins. Our multiple 
Fig. 5C). Toward the lamina, the central trace is more sampling for the species studied indicates that 
curved than toward the base of the petiole and it has a characters described within the species are constant; 
continuous parenchymatous sheath with dark-staining however, interspecific variation was found for the 
deposits occluding the cell lumina. The leaf trace number of spongy parenchyma layers, as well as 
pattern in members of other tribes consists of an open parenchyma and collenchyma layers in the midrib and 
V-shaped are with straight or invaginated ends and the petiole. These features are not diagnostic at the 
one to two pairs of lateral bundles (types IV, VI, genus level. Moreover, trichome type and distribution 
Fig. 5D, E), and a closed ring in Randia (type V, sometimes varied among species of Deppea and 


140 Annals of the 
Missouri Botanical Garden 


Figure 4. Petiole anatomy within the Hamelieae and other representatives. —A. Hintonia (D. Martínez & E. Dominguez 
201, CHAPA), abaxial epidermis with smooth cuticle and angular collenchyma. —B. Exostema (E. Carranza 3362, CHAPA), 
abaxial epidermis with striated cuticle and lacunar collenchyma. —C. Omiliemia (D. Martínez 275, CHAPA), abaxial 
epidermis with smooth cuticle and lacunar collenchyma. —D. Syringantha (F. González Medrano et al. 4659, MEXU), simple 
unicellular trichomes. —E. Psychotria (E. Martínez et al. 27882, MEXU), simple multicellular trichomes. —F. Deppea (D. 
Martinez 254, CHAPA), parenchyma with tannins above phloem. —G. Hamelia (D. Martinez 167, CHAPA), central trace 


— 


forming an open arc. —H. Randia (E. Domínguez & H. Ochoterena 1775, MEXU), central trace forming a closed ring. —I. 
Randia (E. Domínguez & H. Ochoterena 1775, MEXU), fibers sheathing the vascular tissue. —J. Plocaniophyllon (D. Martínez 
et al. 311, CHAPA), raphides in parenchyma. —K. Exostema (E. Martinez et al. 29719, MEXU), druses in collenchyma. —L. 
Bouvardia (J. Rzedowski 38901, CHAP A), raphides and druses in parenchyma. Scale bars: A, B, D, I, J, L = 10 pm; C E = 
2 ; F = 100 um; G, H, K = 30 um. *, cuticle; c, collenchyma; dr, druse; e, epidermis; fi, fibers; pa, parenchyma; ph, 
phloem; ra, raphide; t, trichome; x, xylem. 


Volume 96, Number 1 
2009 


Martínez-Cabrera et al. 141 
Foliar and Petiole Anatomy of Hamelieae 


eS 
ee 
eet. 
SA 
Meet 
(QS 


SSA 
SS SOD 
e 
SESE 
SSS 


99659. SS 

CRS SOSH 

V MEINE 

RSS SISSI 
IS 

RSS RAR 
OSOS 


W 


W 


XY 
XxX 
xy 


S 
SE 


> 
2565. 2 
LSSI 
eee COSE ER 
SSS. 


Figure 5. Petiole: vascular tissue types. —A. Deppea (J. I. 
nia 


CHAPA), type II. —C. Hoffman 
G 


parenchyma; narrow black line above phloem = parenchyma with tannins; dotted line 
l n 


t . Calz 
(Cuevas & Guzmán 4185, CHAPA), typ z 
. —E. H. Vibrans 4932, CHAPA), type VI. —F. Randia (E. Domínguez & H. Ochoterena 
1775, MEXU), type V. Symbols: hatched line — collenchyma; continuous thick black line above phloem — fiber: 

k l h d — phl 


VES 


" 
Ax 


ns 


"e 
XO 
XXV 


TV 
ARR 
S 


ERA 
A 


2 


d 
xi 


^ 
X 
X 


KE 
we 
NE 


SY 
m 
X 


E 
K 
o 


x 


RY 
COP 
RR? 


7 767974 

SSIES 
POSSE 
RAILS 

LR 


EA SSX 
ERRATA LLX 27 TERRE ORE USER 
EERE S ZRII Lo RARA o ) 
ERE I RO BSR d ROR RE RRR 
ROR REEL REED ERY Y RRR 
SA RL Le LE > 
rae ei CER Ma F 


ONG 
x 
e 
2% 


e t, 


ROM 
PSL 
LOBEL 
COE 
LE2 
SRA 


ada 5470, MEXU), type I. —B. Hamelia (D. Martínez 167, 
ype III. —D. Psychotria (D. Martínez i 


ibers; white = 
oem; discontinuous black line 


above phloem = sclerenchyma; vertical lines = xylem. Scale bars: A-F = 250 


Hoffmannia, thus a larger sampling in those large — This, together with morphological characters such as 


genera such as Deppea and Hoffmannia needs to be the dry, capsular fruit, axial placentat 


studied. 


n, and 
numerous small seeds with foveolated-reticulated 


Several types of vascular tissue distribution in the testa, supports a relationship between Deppea, 
midrib and petiole were recognized, and they appear — Omiltemia, Pinarophyllon, and Plocaniophyllon (Lo- 
to be useful diagnostic characters at the generic level. rence & Dwyer, ). 


Types I, IL, and III were observed in different genera 


ha share lamina and petiol 


yringant, s many petiole 
of Hamelieae. Deppea has type I, in which the petiole features with Hamelieae, including the single palisade 
is formed by a multiple trace, with the central trace parenchyma layer, loose spongy parenchyma, raph- 
forming an open arc and a pair of lateral bundles. In ides, and vascular tissue type I in the petiole (as in 
the midrib, the lateral bundles are fused to the central Deppea); and it shares common features of the midrib 
trace. Hamelia has type IL, which is characterized bya with Hoffmannia, Omiltemia, Pinarophyllon, and 
multiple trace and with the central trace forming an  Plocaniophyllon. These characters support the pro- 
open arc, with invaginated ends and a pair of lateral posal from McDowell (1996) to include Syringantha in 
bundles. The vascular tissue has the same distribution the tribe. In addition, Syringantha shares with some 
in the petiole and the midrib. The most common species of Hamelia, as mentioned by McDowell (op. 


vascular tissue distribution type found in the tribe was — cit.), deciduous stipules, secundiflorous infloresce 
type III, observed in the genera Hoffmannia, Omilte- es, yellow flowers, stamens with flattened filaments, 


nc- 


mia, Pinarophyllon, and Plocaniophyllon. In the  introrse dehiscence, subconical nectariferous disc 
petiole, the tissue is formed by a multiple trace with shape, and smooth to reticulate exine of the pollen 


pairs of grain. Syringantha is the only member of the tribe to 


entral trace forming an arc, and tw 

lateral circular bundles. Unlike the petiole, the midrib inhabit drier environments, and its narrowly elliptical 
has only one pair of lateral bundles. In Deppea, leaves and thick cuticle are doubtlessly adaptations to 
Hamelia, Hoffmannia, Omiltemia, and Pinarophyllon, these environments. 


the vascular tissue contains an are formed by embers of Hamelieae share some lamina and 
parenchyma cells with tannins above the phloem. petiole characters with the studied representatives of 


Missouri Botanical Garden 


Annals of the 


142 


1x9] 998 *uonrugop edá 103 , 
quosqe *— 5uosoxd + ‘yyeoys Tepnosea 
‘sa fouroorn repipoorun ‘n *eur&ouoro[os ‘Jos ‘soprydes “ex wood ‘yd tewAyouered ‘ed ‘ore uodo *o touoqorn re[njpoor nur “ur texoqr ‘IT ‘sosnap ‘Ip *eur&gouoqoo “os topuk poso[o “a 


0o “ed ip “el x E IA OI-S = 99 I 9 8 “sp SI70I QIp4Danoq 
9vopnoápog/2eouoqour) 

99 Tp -= y A m E iJ A ? 9 “sp m PURT 
IBITIIPABA 

yd ed E - ps AI E (nu ps AI o 9-€ '9-€ e= vigopÁsq 
ILIO ÁSA 

99 Ip - = IA 97v = 99 I o 9 “sp SY PIMOT, 

oo “ed Ip = — IA ZI-S n 02 I o OL ‘2P 9-G DUAISOKT 

yd ‘oo ‘ed EI + - I 9-p n ed IH o OT ‘8 1-9 pupupsukc 

993uom]our 

sa “yd ‘oo “ed EI + ed Hi £x - y IH o c 'G es uoydydomns0}g 

sa “yd ‘oo “ed EI E ed Ill £x ur ed IH o OL He t= uopdydomurg 

sa “yd ‘oo “ed EI E ed Ill cm ur ed IH o 9-7 ‘9-Z es DIS Yu) 

sa “yd ‘oo “ed EI + ed Ill cu ur ed IH o EE 9-£ pruuvuuffogr 

yd ‘oo ‘ed EI + -= I £x ui in ed Hi o 9-6 “ce €> Down 

sa “yd oo ‘ed EI + ed I Zs (n) ui ed I o 1-G “0-€ 9-£ vaddagy 
SBITSUIB AL 

uonnqrus[ ssl  uruueg reos 206 (uni) euoqoup ory edk uroyed ([erxepe pue (um) sponny BIQUd+) 
puensenuo) owy pog pmo Je[noseA repose A yerxeqe) euicqoueqor) 
sponed qUpTIN 


"exouoS oAnvquoso1dor 19710 pue “byyunFuLids *oqui oeorpoureg] jo Auroyeue o[oriod pue eto] 10; Sropoereu? — "c [qe], 


Volume 96, Number 1 
2009 


Martínez-Cabrera et al. 143 
Foliar and Petiole Anatomy of Hamelieae 


Rubioideae, Cinchonoideae, and Ixoroideae, namely a 
simple epidermis, stomata restricted to the abaxial 
yll. In addition, they 
share with Psychotria (Rubioideae) multicellular 


surface, and dorsiventral mesop 


trichomes, a single layer of palisade parenchyma 
cells, loose spongy parenchyma, and occurrence of 
raphides and tannins. Psychotria differs from mem- 
bers of Hamelieae, however, by its lignified elements 
sheathing the phloem and the presence of vascular 
tissue type IV in the midrib and petiole. To date, type 
V is unique to Psychotria, as it has not been recorded 
in any other member of the family. It is important to 
study additional species of the genus and other 
members of tribe Psychotrieae to confirm the 
diagnostic potential of this characteristic. 

The genera Exostema and Hintonia of the Portlan- 
(Cinchonoideae) share the presence of 


layers of palisade parenchyma cells and compact 
spongy parenchyma with druses, and by the absence 
of a sheath in the petiole vascular tissue. Bouvardia 
i type VI 


in the petiole as in Exostema and Hintonia, but, 


shows vascular tissue type I in the midrib an 
unlike these genera, Bouvardia had raphides and 
h the la i ] 


previously been classified in the tribe 


druses in bot mina and petiole. Bouvardia had 
edyotideae 
(Rubioideae) because of reports of raphides. Its 
winged seeds, however, motivated Robbrecht (1988, 

1993) to consider it a member of the 
(Cinchonoideae). This has not been supported by 
molecular data (e.g., Robbrecht & Manen, 2006). The 


presence of raphides and druses in the leaves of this 


tribe Cinchoneae 


taxon adds support to the idea of an independent 
origin of raphides and to the exclusion of Bouvardia 
from Rubioideae. 
The genus Randia of the subfamily Ixoroideae was 
e only one to ha in the 
midrib and petiole (type V). Moreover, the vascular 


ve a closed vascular trace 


tissue was sheathed by a discontinuous layer of 
fibrous elements. These characteristies and the pres- 
ence of druses are a unique combination of characters 
that distinguish this genus from the other taxa 
studied. Further sampling should test the taxonomic 
value of this character and the rank at which it may be 
useful. 

Attributes observed in the studied genera, such as 
simple epidermis, hypostomatic leaves, and dorsiven- 
tral mesophyll, have also been recorded for other taxa 
of Rubiaceae (Metcalfe & Chalk, 1950). Most of the 
for the 
first time for the family; thus, vascular tissue diversity 


leaf trace types recognized here are described 
in the P is higher than that reported by Metcalfe 
and Cha 


Kiel by an open vascular trace in the genera 


1950). Except in Randia, the petiole was 


observed in the present study. Comparisons of 
photographs of petiole vascular tissues of Rondeletia 

. (Cinchonoideae, Kocsis et al., and Pavetta L. 
1986) enabled us to 
confirm that type I, described for Deppea of 


(Ixoroideae, Herman et al., 
Hamelieae, has a broader distribution. The differenc- 
es among vascular tissue types recognized in this 
study are related to the particular shape of the central 
trace and to whether its endings are curved or straight, 
and to the number of lateral bundles. Moreover, the 
number of lateral bundles decreases from the base of 
the petiole to the midrib in types I, MI, and VI. 
Howard (1979) reported that reduction in the number 
of traces is a common modification in petiole vascular 
tissue. Notably, no modification of the foliar trace was 
observed in types II, IV, and V 

Raphides have been considered a taxonomic 
marker within Rubiaceae, in particular in d 
the subfamily Rubioideae (Bremekamp, is 

erstandable that Hamelieae 

Es classified in this subfamily, as all genera of 


s aie 


Hamelieae have raphides. pss m of 
d owever, sugges at tribe 


ata, 
hould be MON to the subfa E d d 


aphi 
indicates that these crystals appeared deem iy 
in both subfamilies. We consider that it is therefore 
important to study their ontogeny and chemical 
composition to evaluate potential homologies within 
these subfamilies. 

The single palisade parenchyma layer, the 
distribution of vascular tissue in the midrib and 
petiole (type I, IL, or ID, and the presence of 


including Syringantha. Some of these ian 
also enable taxonomic delimitation 

e tribe. This and 
that anatomical leaf and petiole characters have 
taxonomic potential (Herman et al., 1986; Dessein 


, 2001; Piesschaert et al., 2001: Kocsis et al., 


other st A p 
et al. 


Literature Cited 


Andersson, L. & J. H. Rova. 1999. The 
phylogeny of the Rubiales (Rubiaceae). Pl. Syst. 
214: 161-186. 

Andrés-Hernández, R. & T. Terrazas. 2006. Anatomía foliar 
y del pecíolo de especies del género Rhus s. str 
(Anacardiaceae). Bol. Soc. Bot. Mex. 78: 95-106. 

Aoyama, E. M. & M. 
de Aechmea Ruiz & Pav. Subgénero Lamprococcus (Beer) 
Baker e espécies relacionadas (Bromeliaceae). Revista 
Brasil. Bot. 26: 461—473. 


rps16 intron and the 
Evol. 


das Graca Sajo 20083. E foli 


144 


Annals of the 
Missouri Botanical Garden 


. & R. Kool. 1983. Comparative leaf anatomy of 
i. (Olacaceae). Blumea 28: 367-388. 

Bremekamp, C. E. B. 1966. Remarks on the position, the 
delimitation, and the subdivision of the Rubiaceae. Act 
Bot. Neerl. 15: 1-33. 

Bremer, B. 1987. The sister group of the paleotropical tribe 
Argostemmateae: A redefined neotropical tribe Hamelieae 
(Rubiaceae). Cladistics 3: 35-51. 

ndreasen & D. Olsson. 1995. Subfamilial = 
sequence data. Ann 

Buijsen, J. R. M. 1995. Leaf anatomy of Harpullia, Majidea, 
and Conchopetalum (Sapindaceae). Blumea 40: 345—361 

Candolle, A. P de. 1830. Prodromus systematais naturalis 
regni vegetabilis. Paris. 

Dessein, S., S. Jansen, S. Huysmans, E. Robbrecht & E 
Smets. 2001. A nd survey of Virectaria 
(African Rubiaceae), with a discussion of its taxonomic 

Linn. Soc. 137: 1-29. 

989. Stem and leaf anatomy of the 
Alseunamineoas: Aliso 3: 567-578. 
& A. L. Weitzman. 1996. 
the young stem, node, and leaf of Bonnetiac 


Comparative anatomy of 
eae, including 


observations on a foliar endodermis. Amer. J. Bot. 83: 
405 | 
Dos Reis, C., S. L. Proença & M. Graças Sajo. 2004. 


Vascularizagáo foliar e anatomia do pecíolo de Melasto- 
mataceae do vc do Estado de Sao Paulo, Brasil. Acta 
Bot. Bras. 18: 987— 
Elias, T. e 1976. r ioci of the genus i 
ae). Mem. New York Bot. Gard. 26: 81— 
Genera Plantarum. sc 
Engel, orth pa erican 
Astragalus species (Fabaceae) RS persistent. petioles. 
Aliso 2: 339— 345. 
Fariña, A., D. Arr 
2003. Anatomía comparada de la lámina 


ieche, A. Boada-Sucre & D. Velásquez. 
foliar de las 
especies de Heliotropium L. (Boraginaceae) presentes en 
Venezuela. Interciencia 28: 68-74. 


Rubiaceae: The vas eT Trustees of the eval Botanic 
Gardens, Kew. <http: p/>, accessed 7 
February 2000. 

Herman, P. P. de P... d i N. Grobbelaar. 1986. 


rican Pavetta species. S. 


W. Keaken & E. K. Schofield. 2004. Index 


national Association for Plant Taxono- 
B 


Howard, R. A. 1979. The petiole. Pp. 88-96 in C. 
Metcalfe & L. Chalk (editors), Anatomy of the Dicotyle- 
dons, Vol. 1, 2nd ed. Clarendon Press, Oxford 

Kocsis, M., J. Darók & A. Borhidi. 2004. Comparative leaf 
anatomy and morphology of some n nnd p eletia 
(Rubiaceae) Pl. Syst. Evol. 248: 2 

Lorence, D. H. & J. D. Dwyer. 1988. A revision s Deeds 
S ETE d beu ue 4: 389-436. 

McDowell, ub dd iei coulteri (Hooker 
McD wi mbin: and remarks on 
relationships of the me pun genus Syringantha 
Standley ww Novon 6: 273-219. 

Mentink, H. & P. 1992. Leaf anatomy of the 
Melastomataceae, oí and Crypto 
Blumea 37: 189-225. 


PET 
the 


a new com 


niaceae. 


Metcalfe, C. R. 1979. The leaf: General rund and 


ontogeny of the tissues. Pp. 63—75 in C. R. Metcalfe & L. 
Chalk (editors), Anatomy of the Dicotyledons, Vol. 1, 2nd 
ed. Clarendon Press, Ox cord. 

& L. Chalk. 


1950. Anatomy of the Dicotyledons, 1st 

arendon Press, Oxford. 

Jansen, I. Jaimes, E. Robbrecht € E. 

Morphology, anatomy, and taxonomic 

wow. d EU (Rubiaceae- Robicie. Britto- 
nia 53: 490-504. 

Robbrecht, E ll Tropical woody Rubiaceae. Opera Bot. 


E pum to the 1988 n of classifi- 
cation of the Mp aed s Bot. Belg. 6: 173-193 

2006. The major “evolutionary 

lineages of ae ‘family (Rubiaceae, angiosperms). 

A 


rbcL data. 
Cinchonoideae and Rubioideae. Syst. 
85-146. 

Ruzin, E. * 1999. Plant Microtechnique and Microscopy. 
Oxford University Press, New York. 

oe -Zapotitla, T. Terrazas. 2001. Leaf anatomy of 

xa of the “Teighooeninin clade (Orchidaceae: 

a Lindleyana 16: 81-93 

Smith, A. € W. Stern. 1962. Leaf anatomy as an aid in the 
identification of two Fijian plant species. Brittonia 14: 
31247. 


Geogr. Pl. 76: 


Souza, L. As LS innt & J. Oliveira. 2004. 
d an nom i e leaf and stem 
of Peperomia dahlstedtii c. DC., Ottania martiana Miq., 
a oe diospyrifolium Kunth (Piperaceae). Gayana, Bot. 


coe B. 1958. Remarks on the classification of the 
Rubiaceae. Bull. Jard. Bot. Bruxelles 28: 2 

. 1976. Rubiaceae Part 1. Pp. 1-414 in R. M. Polhill 
(editor), Flora of Tropical East Africa. Royal Botanical 
Gardens, Kew. 
Wilkinson, H. P. Cuticle. Pp. 140-155 i 
Metcalfe & L. ih. (editors), Anatomy of the Dicotyle- 
dons, s 1, 2nd ed. Clarendon Press, Oxford 

1983. Leaf anatomy of Gluta (L.) Dine Hou 
ET Bot. J. Linn. Soc. 86: 375-403 


] 


APPENDIX 1. List of sp | 1 
slides. Species are arranged alphabetically. 

Bouvardia cordifolia DC. MEXICO. México: H. Vibrans 
4932 (CHAPA). Michoacán: J. Rzedowski 59784 (MEXU). 
XU). B. ae 
E Peur 
R. Torres 9951 (MEXU). México: J. A. López & S. 
. ternifolia (Cav.) Schlt E 
), D. Martínez 


Ochoterena & 2 

"s & Lorence. MEXICO. E 
APA), D. os 195 (CHAPA), H. Ochoierena & D. 

bay 335 (MEXU). D. hintonii Bullock. MEXICO. Chiapas: 

T. B. Croat 47234 (CHAPA). Oaxaca: A. Campos 1840 


Volume 96, Number 1 
2009 


Martínez-Cabrera et al. 145 
Foliar and Petiole Anatomy of Hamelieae 


MEXU), R. E. Gereau et al. 1075 (MEXU), T. B. Croat 
MEXU). D. miahuatlanica Lorence. C 
o, é 


(MEXU). D. microphylla XICO. Hidalgo: D. 
Lorence 4895 (ME XU), J. flne 23420 E Puebla: 
a h.) Benth. MEXICO. 
a: D. Lorence 4200 (MEXU), D. Lorence 4338 ded 

ru Rivera 0965 (MEXU). D. pubescens Hemsl. MEXIC 
Oaxaca: A. E & J. Reyes 1333 eran H. ana 
& D. Bailey 533 (ME ais R. e A ue XU). D. purpusii 
Standl. MEXICO. Hidal dndez 5767 (MEXU), D. 
Martínez 254 (MEXU). m iile J. T Calzada D dur 
D. tenuiflora Benth. MEXICO. Veracruz: I. A. as 189 
(MEXU), £ A. Vargas 243 (MEXU). D. ie Hem sl. 
bI Oaxaca: R. Torres 6546 axo. Veracruz: J. Fay 
I. Calzada 734 (MEXU), H. Ochoterena et al. 370 

o M. G. Zola 0390 (MEXU 

Exostema d end (Jacq.) Room. & ho: MEXICO. 
Campeche: C. n 7698 (CH APA). Oaxaca: D. Lorence 
3036 (CHAPA). pm erétaro: E. Carranza 3362 (CHAPA). E. 


exicanum A. Gray. E a Martínez 
29166 (CHAPA), E. Reus 29719 (MEXU). Ch lapas: A. 
Chamé et al. 159 (CHAPA). Oaxaca: H. Ochoterena et al. 289 
(ME 


XU). 
Hamelia axillaris Sw. MEXICO. Chiapas: E. Martínez 


H. patens peg 
Ventura P (CHAPA). pni : 


rovirosae Wernham. MEXIC 


Hernández 1508 (CHAPA). H. versicolor A. Gray. MEXICO. 
Guerrero: C. Catalán 3 (CHAPA). Jalisco: H. Ochoterena & 
EXU). Morelos: D. M. Arias & D. Mai 


riínez 


Eoi XU). H. xorullensis Kunth. MEXICO. Jalisco: H. 
EE & D. Bailey 220 (M n México: D. Martínez 
298 (CHAPA), H. Vibrans 5885 (MEXU). 


Hinionia latiflora (Sessé & Moc. ex DC.) Bullock. 
MEXICO. Jalisco: A. Flores 3655 (CHAPA), D. Martínez E 
terena 
Bailey 224 (ME Gutiérrez i 1 
e APA). H. octomera E IB NC Yucatán: 

H. Ochoterena & H. Flores 168 (ME 

Hoffmannia angustifolia Standl. MEX ICO. Chiapas: D. E. 
Breedlove 24801 (MEXU), D. E. Breedlove 35293 (CHAPA), 


A. Log & M. Heath ae (MEXU). A. conzattii B. L. Rob. 


. Bachem & R. Rojas 891 (CHAPA). 

Guerrero: F. ee et al. 5070 (MEXU). Hidalgo: J. 
Rzedowski 12541 (CHAPA). Oaxaca: T. Wendi & M. Ishiki 
A (CHAPA). H. Pu b qi Standl. & L. O. Williams. 

XICO. Hidalgo: D. Martínez 171 (CHAPA). Oaxaca: S. 
an 1173 (CHA E ja a He Torres. 11059 
MEXU). H. cuneatissima XICO. a 
E. Martínez 5471 & F. Barrie asa p oes R. Cue 
L. Guzmán 4185 (CHAPA, MEXU). México. D. Lorence ns 
Tejero 4887 (MEXU). H. excelsa (Kunth) K. Schum. 
MEXICO. Oaxaca: S. Maya 3305 (CHA 
Lorence 3890 (MEXU), G. Castillo e rd "SL H. 
nicotanaefolia (M. Martens b: ae 
MEXICO. Chiapas: M. Gonz 
Martinez 18864 (MEXU). ie T. Wendt et al. 4661 

E 


— 


CHAPA). 

Omiltemia aan e (Standl. C. V. Morton. MEXICO. 
Chiapas: M. Hea A. Long 771 n M. Heath & A. 
Long 1173 MEX, D. Martinez ae APA). O. d 

Lor 


Standl. MEXICO. a 2048 (MEXU), D. 

Martínez 335 (CHI e ps 285 Pun XU). 
Pinarophyllon a Mu e ICO. M D. 
E. Breedlove 30795 (ME APA), E. 


XU), D. ea 318 (CH 
Ventura & E. López 2053 (MEXU). 

Plocaniophyllon a P ME 
Heath & A. Long 899 (MEXU), D. Martínez 312 (CHAP 
Matuda 17763 TMEXU). 

i ps E MEXICO. p C: 
Catalán PA). Oaxaca: S. Maya 1691 
Gaps. 3 Misa 1753 nes P. faxlucens ii aes & 


XICO. Chiapas: M. 
A), E. 


EXU). Jalisco: 
(CHAPA) LAS. rds 3687 puso P. microdon 
DC) Urb. ME : mpeche: E. Martínez 27882 
(MEXU). dos E. Palacios 1995 (CHAPA, MEXU). 
Jalisco: D. Martínez 205 (CHAPA 
Randia aculeata L. MEXICO. Jalisco: E. Domínguez & H. 
Ochotorena 1775 


d RE (Hook. £) T. McDowell MEXIC 
Hidalgo: sree ce APA), F. drano et d 
9631 (MEXU). on najuato: É. Ventura |. E. López 7989 


(MEXU). Tamaulipas: Mora. 5314 (MEXU). 


PARAPHYLY OF /XORA AND NEW 
TRIBAL DELIMITATION OF 
IXOREAE (RUBIACEAE): 
INFERENCE FROM COMBINED 
CHLOROPLAST (RPS16, RBCL, 
AND TRNT-F) SEQUENCE DATA? 


Arnaud Mouly,” Sylvain G. Razafimandimbison,? 
Jacques Florence,* Joël Jérémie, and 
Birgitta Bremer? 


ABSTRACT 


We performed phylogenetic e of DNA fu 


hloropl | beL 6, and irnT-F, to rigorously test 


Consequently, b fide “Andreasen and Bremer (2000) is 
Robbrecht and M 2006) 1 


re more closely related to /xora and allies than the monotypic genus Scyphipho 
not monophyletic a an exclusion of Seyphiphora. bonne fide 
s not RR Ris P iat un and D 


umscriptions of the tribe Ixoreae. Several genera mu or currently associated with the 


ra are 


monophyly of a 


rice included. The m 
nd its rela tives Ale isanthia, Aleisanthiopsis, and Greenea is, 


ht n( 
morphologically NGA 


R 
ria Montrouz., Sideroxyloides, Thouarsiora, Tsiangia B 
dus 


ted for Al thia + Aleisanthiopsis and Greenea, sospeso The m Malesian 
and the Southeast Asian Greeneeae is sister to the Ixoreae—Aleisanthieae clade. 
Greeneeae, Ixora, Ixoreae, Ixoroideae, phylog 


s. str. are sister groups, a. 


Key words: Aleisanthieae, cpDNA, 


ue & P. T. Li, and Veršteegidi), and two new w tribes are 
l Ixoreae 


T 


eny, Rubiaceae. 


The large pantropical genus Ixora L. was earlier 
classified either in the tribe Pavetteae (Richard, 1829; 
Dumortier, 1829) or the tribe Coffeeae (Wight & 
Arnott, 1834) until Gray (1858) placed it in his new 
tribe Ixoreae (Table 1), which now belongs to the 
subfamily Ixoroideae sensu Bremer et al. (1999) 
Ixoreae initially consisted of two large Linnaean 
genera, Ixora (the type genus) and Pavetta L., with 
contorted aestivation and peltate or centrally affixed 
ovules (Gray, 1858), but the genus Coffea L., presently 
classified in Coffeeae, was later added (Hooker, 1873; 
Bremekamp, 1934). From 1979 to 2000, both /xora 


and Pavetta were placed either in Coffeeae (Darwin, 
1979; Verdcourt, 1989) or in Pavetteae (Bridson & 
Robbrecht, 1985; Robbrecht, Phylogenetic 


Bremer (1996, 2000) based on the combined molec- 
ular and morphological data demonstrated for the first 
time that Ixora and Pavetta were not closely related 
genera. The monotypic and mangrove genus Seyphi- 
F. 1806), formerly 


Dr a in the tribe Gardenieae sensu Robbrecht 


Gaertn. (Gaertner, 


phora 


988), was resolved with moderate support (jackknife 


Tus 70%, Andreasen & Bremer, 2000: 1740) as 


1 The authors wish to thank Anbar Khodabandeh and Torsten Eriksson for technical help and support with laboratory work 


ector and team of the 


l de Re 


by the Unité 
uséum Muséum National ea 
” the he Depa r iment o ries et Evolution," MNHN, Paris, and the European 


y; by the French else an d ent to J. Florence; and by e Swedish 


X 05, F 


Swedish Academy of Sciences and Botany Department, a tión SE-106 91 


^US 084 IRD BIODIVAL: Biodiversité E tropicale: connaissance et valorisation, Institut de Recherche pour le 
Développement, Antenne MNHN, Herbier National, CP 39, 16 rue Buffon, 75231 Paris CEDEX 05, France. 


doi: 10.3417/2006194. 


ANN. Missouni Bor. Garp. 96: 146-160. PUBLISHED on 23 ApriL 2009. 


147 


Mouly et al. 


Volume 96, Number 1 


2009 


Paraphyly of /xora 


QBOIOXT QBOIOX] SLOX] SBIMIABA Lye d SBOJOX] JOU 133915124 
QBOIOXT Seyed DLOISLDNOY f 
ABIJOX[ QBOIOXT sapropAxosapig 
8rpos "our QROIOX] ¿OBOLOX] sTpos "our avorodseurar’y ovopreyjons) vioydy dios 
LIX] ABOJOX] SLOX] ovo9go"7) ovo}oaed Lye d ¿ƏLƏIOX] ƏLƏOXJ pwunuol y] 
QBOIOXT Leye d 9v99gpo7) vong 
ILIU) ABOJOX] ovono[opuoy D2u9244) 
LOX] SLOX] ovo9go") IBINIABA Seyed T422140(T 
QBOIOXT LIX] aevo}oaed Leye d piyoooumydn’y 
ovoryquesioly BOTOX] ovotjo[opuoy sisdonpunsiayy 
ovolyquesioly SLOX] ovono[opuoy pnygupsiy 
aroo ov99gpo7) 9v99go7) aroo) 9v99gpo7) SB9IOX[ — 9€9IOX[ owoo9go' o?o9go") vaffoy 
oeopnoseq eed SBINIABA aroo oeo}oaeg Leye d 9v99gpo7) SB9IOX[ | 9€9IOX[| OEOIOX[ Logon eped Duoanq 
AB9JOX] LIX] ABOJOX] aroo LALJ Seyed 9v99gpo07) SB9IOX[ ^ 9€9IOX[| OEOIOX[ orson ovonoaeg DAOXJ 
uorjeorrsse[o 900c 0006 6861 8861 “M99:1qq0y S861 6L6T Uimteg ve6T €Z8T 8S8T PEST Noury 6781 
juoso1q “USUAL Y “IGUSIG Y JMOIPISA, “19991 9O0H *durexouro1q *1oxXoor “kero, b qst M “preyory 
Wpsiqqoy — uoseo1puy 9 uospugq Pg 6381 
*1onxoum(q 
oqu poreorpur oy] ur snues oi popn[our Ap[oAnejuo) SIOE ou jer ueour sxreur uorson() EXE poje[ox pue oeorox[ oqrn ou 10] sourouos uoneorisse[o jo uosrreduroo OIH '[ ATEL 


148 


Annals of the 
Missouri Botanical Garden 


sister to a strongly supported (JK = 100%, Andreasen 
2000: 
species of Ixora and one species each of the genera 
(Jussieu, 1789) and 

1911). Accordingly, 


& Bremer, 1740) clade consisting of four 
Myonima Comm. ex Juss. 
Versteegia Valeton (Valeton, 


, 1973), Doricera Verdc. 
1989), Ixora, Myonima, and Versteegia; they tenta- 
e" ge M s dE there. The tribe comprises 
ecies, some of which are economically 
S pos ‘the mei L coccinea L., I. 
ex G. Don, and I. hookeri 
It is cec by mbin 
k, 1998), pee 
inflorescences, and 4-merous flowers with aestivation 
contorted to the left (Andreasen & Bremer, 2000). 


Pavetteae and Coffeeae were recircumscribed in a 


finlaysoniana Wall. 
(Oudem.) B 
tion of articulated petioles (De Bloc 


remek.). 


narrow sense 
Andreasen and Bremer (2000) also revealed that 
Ixora, represented by four species, was paraphyletic 
or polyphyletic. About 500 species (98%) of Ixoreae 
were encountered in the pantropical fxora (Mouly, 
2007). Several described monotypic genera, including 
Becheria Ridl. (Ridley, 1912), Bemsetia Raf. (Rafin- 
u pes 1899), 

uz. (Montrouzier, 1860; nom. rej.), 
Patabea Aubl. (Aublet, 1775), Sideroxyloides Jacq. 
(Jacquin, 1763), and Thouarsiora Homolle ex Arènes 
rénes, 1960), are considered to fall within Ixora 
ased on morphology (Bentham, 0; Beauvisage, 
1901; Fosberg, 1937; Sandwith, 1937; Guédés, 1986; 
Smith & Darwin, 1988; De Block, Ba The 

otypic genus Charpentiera Vieill. (nom. illeg. 

he awe to the family se MA 
(Vieillard, 1865), was included in /xora (Rubiaceae) 
by Beauvisage (1901). The genus Tsiangia But, H. H. 
Hsue & P. T. Li (But et al., 1986) 
species based on a single specimen from Hong Kong, 
but Bridson (2000) stressed that this species was a 
deviant parasitized collection of I. chinensis Lam. 


was described for a 


Captaincookia, Doricera, and Scyphiphora remain 


monotypic genera, and Myonima and Versteegia 
contain only four and five species, respectively. 
Baillon (1879) merged Myonima with fxora as a 
section, but subsequent Rubiaceae taxonomists did 
not follow this. 

A recent phylogenetic study by Rova et al. (2002: 
149) based on the trnL-F chloroplast region revealed 
for the first time a close relationship of Ixora with 
three Southeast Asian genera, Aleisanthia Ridl. 
(Ridley, 1920; Tange, 1996a), Aleisanthiopsis Tange 
(Tange, 1996b), and Greenea Wight & Arn. (Wight & 
Arnott, 1834), which were formerly classified in the 
tribe Rondeletieae of the subfamily Cinchonoideae 


(Robbrecht, 1988). Malesian Aleisanthia and Indone- 
sian Aleisanthiopsis are two small genera, each with 

species. Greenea is a genus from Southeast Asia 
(Bul et al., 2005) with ca. 10 species. These three 
rainforest genera have contorted aestivation, scorpioid 
inflorescences, and capsular multiseeded fruits, 
features unknown in Ixoreae sensu Andreasen and 
Bremer (2000). More recently, Robbrecht and Manen 
(2006) proposed a broad cireumscription of Ixoreae 
including Ixora, Myonima, Seyphiphora, Versteegia, 
and the above three genera. Their Ixoreae is 
diagnosed by contorted corolla aestivation, a feature 
commonly found in Ixoroideae s.l. 

Sequencing data from the rps16, trnT-F, and rbcL 
chloroplast markers have been used separately and/or 
in combination with that of the ITS 
ribosomal DNA (nrDNA) for assessing E d 


of nuclear 


relationships in some m group , Raza- 
imandimbison & Bremer [ or "n sd ITS; 
Razafimandimbison & uei [2002], Lantz & 


Bremer [2004], Alejandro et al. [2005] for trnT-F 
and ITS). Here, we perform phylogenetic analyses 
using DNA sequencing data from the three chloroplast 
regions to reconstruct a robust phylogeny for Ixoreae. 
The resulting phylogeny from the combined data is 
used to rigorously test the monophyly of previous 
circumscriptions. 


MATERIAL AND METHODS 
TAXONOMIC SAMPLING 


We used 33 species (Appendix 1) comprising seven 
Ixora species with a representative geographical 
range, one individual each for the monotypic genera 
Captaincookia, Doricera, Hitoa, Scyphiphora, Side- 
roxyloides, and Thouarsiora, a specimen of one species 
each for Myonima, Versteegia, Aleisanthia, and Alei- 
santhiopsis (two accessions), five representatives of 
Greenea, two members of the tribe Vanguerieae 
(Peponidium cystiporon (Byn. ex Cavaco) Razafim., 
Lantz & B. Bremer and Cyclophyllum deplanchei Hook. 
f) and one species each for the nine formally 
recognized tribes of Ixoroideae s.l. (Andreasen & 
Bremer, 2000; Rova et al., 2002), notably Pavetta for 
Pavetteae. Despite available names under Ixora, the 
species of the genera Hitoa and Sideroxyloides are used 
in the study to clearly assess the circumscription of 
Ixoreae, with those of Ixora being addressed in another 
study. Luculia gratissima (Wall) Sweet was used as 
outgroup taxon (Appendix 1), in agreement with its 
basal position in Rubiaceae. Several specimens of 

reenea included in the analyses could not 
identified at species level. We were unable to ibas 
any representatives of the genera Bemsetia and Patabea 


Volume 96, Number 1 


Mouly et 149 


Piani. of Ixora 


as they are only known from their respective type 
illustrations, and failed in obtaining sequences from 
echeria, 


extracted samples of B 
ancheria. Several taxa were inclu int 


Charpentiera, and 
ded he stud 
using accessions from the GenBank database, and we 
were not always able to obtain complementary material 
to complete missing data within the data sets (Appendix 

1) 


DNA EXTRACTION, AMPLIFICATION, SEQUENCING, AND ALIGNMENT 


was extracted from dried material 
Chase & Hillis, 1991) or 

herbarium specimens following the mini-prep proce- 

dure of Saghai-Maroof et al. (1984), as o by 

Doyle and Doyle (1987). Extracted D 

with the QIAquick polym 


otal DNA 
preserved in silica gel 


as cleaned 
merase a reaction (PCR) 
purification kit (Qiagen, Solna, Sweden). PCR reac- 
tions were as follows: 27.25 ul of H20, 5 pl of PCR 
buffer, 5 pl of MgCls, 5 pl of 0.1 M tetramethylam- 
monium chloride (TMACI), 4 pl of dNTP, 0.25 pl of 
Taq DNA polymerase (AB-0192 & AB-0192/b; 
ABGene, ee U.K.), 0.5 ul of each primer, 
and 0.5 pl o 
PCR Merida 


1% of bovine serum albumin (B p 
performed in an Eppend 
Mastercycler (Applied Biosystems, Life 5 
Carlsbad, California, U.S.A.) gradient, started with an 
initial Mi: phase of 2 min. at 95°C, a 7 35 
o s of 30 sec. at 95°C, 1 m PC-55°C, 
and Bo min. at 72°C and ended vies a pw extension 
phase of 7 min. at 72°C. In all PCR runs, one reaction 
was run with water instead of DNA as a negative 
control to check for contamination. 

The rps/6 intron was amplified with the primer pair 
1997). For half of the 


species, we repeatedly failed to obtain Pu” 


rpsF/rpsR2 (Oxelman et al., 
in one reaction because of a problematic poly A 
the 3' end of the intron os et al. 
amplification was successful with the itera primer 
pair rpsF2/rpsR3 (Bremer et al., 2002), but resulted in 
sequences 50—70 bp shorter. The entire trnT-F region 


at 
However, 


(including the two trnL introns) was amplified in two 
parts. The trnT-trnL segment was amplified with the 
2002; Lantz & 
Bremer, 2 e second part. irnF, was 

b 1991). 


For irnL-trnF, sequencing reactions were ela! 


primer pair Ál/I (Bremer et al., 


amplified sil primers c/f (Ta SS et al, 
using the two external primers c/f and two internal 
1991) to produce 
complete sequences of the bones region trnT-F wit 
at least partial overlaps (from 20-150 bp). The rbcL 
coding region was amplified in two parts. The first 


primers d/e (Tarberlet et al, 


segment was amplified with the primer pair z1/1020R 
(Andreasen & Bremer, 2000) 
with the couple 3'/427BS (Olmstead et al., 


and the second segment 


1992; 


Andreasen & Bremer, 2000). Sequencing reactions 
were performed with z1/1020R for the first segment 
and with and two internal primers 1204R/ 

5 (Andreasen & Bremer, 2000) for the second part 
to produce complete sequences of the entire region of 
rbcL, with at least partial rd bes 60—90 bp). 

l sequencing reactions 
performed with Big Dye 
Sequencing kit or the Big Dye Terminator v3.1 Cycle 


arkers were 
ie v1.1 Cycle 


Sequencing kit (Applied Biosystems) and subsequent- 
a 3100 Genetic Analyzer (Applied 


The rpsi6, trnT-F, and rbcL sequences were 
assembled using the Staden Package version 1.6.0 
beta-test (Staden, 1996) or Sequencher 3.1.1 (Gene 
Codes Corporation, Ann Arbor, Michigan, U.S.A.) and 
edited manually. Sixty-seven accessions from 33 taxa 
are new to this paper (Appendix 1). All sequences 
were aligned manually with Se-Al v1.0al (Rambaut, 

96). The insertion of numerous gaps in nucleotide 
sequences was required during the alignment proce- 
dure for both trnT-F and rpsi6 (Table 2). Other 
studies have shown indels to be reliable characters 
Lloyd & Calder, 1991; Giribet & Wheeler, 1999; van 
Dijk et al., 1999; Delarbre et al., 0; Freudenstein 
& Chase, 2001; Rouhan et al., 2004). Unambiguous 


insertions and deletions ndel were e coded as 


— 


additional characters by u 0 an mbols for 
eletions and ins jore pora " Gwolford, 
1993 


PHYLOGENETIC ANALYSES 


Bayesian analyses were performed with MrBayes 
3.0b4 (Huelsenbeck & Ronquist, 2001). The Bayesian 
approach evaluates the posterior probability (PP) of a 
tree given the c 

at th 


haracter matrix, i.e., the ro: 


pro! 
t A —— Bayesi 


= 
2 


e tree is correc 
analysis was conducted to account for the pinos 
of molecular data with various evolution up and 
standard binary characters of g . For each 
partition, MrModeltest 2.0 (Nylander, 2004) was used 
to choose the model of nucleotide substitution that 
best fit the data, following Akaike's 
Criterion Calculation Method (Akaike, 


s were ge 


a 
1974). T 
neral time revel ( B 
with among-site substitution rate 
heterogeneity described by a gamma distribution 
and all sites constrained to be variable (GTR + 
for rps16 and the trnT-L segment, GTR with a fraction 
invariant site constraint (GTR + I) for the trnL-F 
segment, and GTR + G + I for rbcL. Unambiguous 
indels were analyzed as an additional partition and 
treated as binary characters. All analyses were 
conducted with four independent Markov chains run 


150 Annals of the 
Missouri Botanical Garden 


Table 2. Indication of the number of accessions included, the length of the aligned matrices, the number of informative 
characters, and parsimony indices in each data set. 


rps16 irnT-F rbcL Combined data 
No. of accessions 33 
Matrix length 773 1726 1373 3872 
Parsimony informative 72 (68 substitutions 148 (119 substitutions 68 (68 substitutions) 288 (255 substitutions 
characters + 4 indels) + 29 in 
Parsimony indices L = 228; = ps = E = 936 
CI — 0.851; CI — 0.855; CI — 0.712; CI — 0.809; 
RI-O RI — 0.888 RI- 5 RI = 


L, length; CI, consistency index; RI, retention index. 


for 2 X 10° Metropolis-coupled Markov chain Monte — sites is low (Yoder et al., 2001; Darlu & Lecointre, 
Carlo (MCMC) generations, with tree sampling every 1 2002), which is pres sently observed (T 

0° generations, and burn-in after 500 trees (as According to Bremer et al. (2002) and despite 
detected by plotting the log likelihood scores against — incongruence, we favored an analysis of combined 
generation number). The analyses were run three data sets. 
times using different random starting trees to evaluate 
he convergence of the likelihood values and posterior — Rgaurqa 
clade probabilities (Huelsenbeck et al., 2002). The 


consensus tree was built using selected trees from The results of the analyses conducted in this study 
each run. Groups characterized by a PP of more than are given in the following section, where the sequence 
0.95 were regarded as strongly supported. characteristics of the individual data sets (Fig. 1, 

To test the Rd inferred from Bayesian Table 2, unpublished figure for rbcL, available from 
analyses, Pop an were also conducted the author for correspondence) are presented, and 
using PAUP* 4.0b8b raa 2002). The maximum detailed outputs from the analysis of the combined 
parsimony p trees were found by heuristic data sets are given (Fig. 2, Table 2). 


search, tree Wir ee Qc» (TBR) branch 


swapping, using 1 replicates of random SEPARATE ANALYSES 

stepwise addition, sii cM L option on. 

Characters were given equal weight, gaps were treated The rue rps16 and trnT-F analyses (Fig. 1A 

as missing data, and > phylogenetically informative B), as well as the rbcL piel all pss ihe 
indels were coded. The consistency index (CI) (Kluge following monophyletic the 


— 


&F arris, 1969) and dc retention index (RI) (Farris, (forming seven /xora species, Hue. Tout o 
1989) were calculated to estimate homoplasy. To  Thouarsiora, Captaincookia, Doricera, Myonima, and 
assess relative support for the identified clades, ^ Versteegia), the Greenea group, the Aleisanthia— 
bootstrap values ) (Felsenstein, 1985) were Aleisanthiopsis group (not tested for monophyly in 

estimated from 1 X 10* replicates, the MULTREES the rbcL analysis due to the single representative), and 
option off, TBR branch swapping, and five random Lu Fu (represented by Peponidium cystiporon 


addition sequences. Groups characterized by Byn Cavaco and Cyclophyllum | deplanchei). 
bootstrap support of more than 9596 were ended Fiona: Ixora was shown to be polyphyletic 
as strongly supported. and Seyphiphora left unresolved in a basal position in 

To test the null hypothesis that our data sets were the Ixoreae alliance. While both the rps16 and trnT-F 
homogeneous with respect to phylogenetic informa- trees (Fig. 1A, B) identified a highly supported clade 


tion, we used the incongruence length difference test containing Aulacocalyx jasminiflora Hook. f., Augusta 
(ILD, also called partition homogeneity test; Farris et austrocaledonica (Brongn.) J. H. Kirkbr., and all 
al., 1994), as implemented in PAUP*. Invariant sites sequenced members of the tribes Pavetteae (sensu 
were removed (Cunningham, 1997) and 1 X 10* Andreasen & Bremer, 2000), Octotropideae, Crema- 
replications were performed. The results of the ^ sporeae, Coffeeae, and Alberteae, this clade was 
partition homogeneity test indicated that the rps76, unresolved in the rbcL tree (not included). The 
irnT-F, and rbcL trees were significantly incongruent  Aleisanthia—Aleisanthiopsis group was resolved with 
(P = 0.0009). However, the accuracy of the ILD test high support (BS 99%, PP 1.00) as sister to the Ixora 
tends to be low when the number of the informative group in the trnT-F tree (Fig. 1B). This sister-group 


= 


Volume 96, Number 1 
2009 


y et al. 
Paraphyly of /xora 


151 


Ixora brachypoda - IXO 


ora tanzaniensis - IXO 
eane 


Ixora finlaysoniana - IXO 
i i IXO 


Hitoa mooreensis - IXO 
Doricera trilocularis - IXO 
Myonima violacea - IXO 
Versteegia cauliflora - IXO 
Ixora coccinea - [X 
Greenea corymbosa - RON 
nea sp. indet. 1 - RON 
Greenea oblonga - RON 
Greenea sp. indet. 3 - RON 
ONIS " ae 


94/1,00] 
[981.00 
A 


leisanthia rupestris - RON 
Peponidiu eb dis VAN 


i- VAN 


Aulacocalyx jasminiflora - GAR 
Pavetta platyclada - PAV 
Fernelia buxifolia - OCT 
emaspora triflora - CRE 
arabica - COF 


e rps16 data set. 


from 


relationship was collapsed in the rps/6 tree (Fig. 1A), 
while the Aleisanthia—Aleisanthiopsis group (repre- 
sented by Aleisanthiopsis distantifolia accession 1) 
and the Greenea group were resolved with poor 
support (BS 5596, PP 0.70) as sister groups in the 
rbcL tree. Within the Ixora group, there were some 
poorly to weakly supported differences between the 
rps16, trnT-F (Fig. 1A, B), and rbcL trees 


COMBINED ANALYSES 


To perform phylogenetic analyses of combined 
data, we merged the three matrices for all 34 terminal 
samples (represented in at least two of the three 
chloroplast DNA [cpDNA] markers) in a combined 
data set of 3872 sites, including 288 (7.44% 
parsimony informative characters. Of the informative 
o) were nucleotide substitu- 


Im 


characters, 255 (88. 
tions and 33 (11.4596) were indels. In our study, the 
combined rpsI6—rbcL-trnT-F tree (Fig. 2) identified 


Ixora brachypoda - IXO 


Ixora finlaysoniana - IXO 

Ixora brunonis - XO 

Sideroxyloides ferreum -IXO 

eLo Doricera trilocularis - [XO 
Myonima violacea - IXO 


Captaincookia iio -IXO 


Hito 
Ixora caccinea - IXO 


Verstee aig cauliflora - 


Aleisanthia rupestris - RON 
corymbosa - RON 
det. 2 - RON 


Grei 
Greenea oblonga - 
a sp. indet. 3 - RON 


uen" Peponidium cystiporon - VAN 
Cyclophyll deplanchei - VAN 
IXO 


phyllum 
Scyphiphora I lac 


Aulacocalyx HE 


Coffea arabica - COF 
Alberta minor - ALB 
Alberta sambiranensis - ALB 


RON 


Majority consensus tree eR from Bayesian analyses of hs markers alignments (2 M tele E 
— B. Topolog t. Num 

the ur analysis of identifiable nodes (left) and consensus ol a 

50% or unresolved node: 


rated from the irnT-F data se 


ue Species recently included to Ixoroideae from Rondeletieae are | smod RON. Outgroups were remove 
es. 


the similar larger monophyletic groups retained in 
the separate analyses (Fig. with strong 
support. Internal nodes received high support 
Fig. 2). The overall tree topologies of the tree 
generated from both the parsimony and Bayesian 
combined analyses were, in general, similar to those 
of the rpsi6 and trnT-F trees (Fig. 1A, 
TP taxa were resolved in two main clades, noted 
as A and B s " a Clade A was highly supported 
is 100%, P and contained the Ixora group, 
the deco d mre group, the Greenea 
group, exemplar Vanguerieae, 
[D i ee C. F. Gaertn. (Fig. 2). The relation- 
p between the Greenea group (A3) and the fxora— 
jp nel are E clade (A1-A2) was 
strong (BS 100%, PP 1.00). 
parsimony analysis, Seyphiphora was resolved with 
poor support (BS less than 50%) as sister to a clade 
consisting of subclades Al, A2, and A3 (results not 
shown). Finally, clade B (Fig. 2) received high 


— 


In the al 


152 Annals of the 
Missouri Botanical Garden 


Ixora brachypoda - IXO 


Ixora tanzaniensis - IXO 


| 1 IX 
lA 


Ixora hookeri - IXO 
Thouarsiora littoralis - IXO 
Sideroxyloides ferreum - IXO 


Ixora finlaysoniana - IXO 


82/1.00 
Ixo reae Ixora brunonis - IXO Al 
(this study) 
<<. Doricera trilocularis - IXO 
100/1.00 Myonima violacea - IXO 


Captaincookia margaretae - IXO 


Hitoa mooreensis - IXO 


Aleisanthieae .. 10.88 


97/1.00 Ixora coccinea - IXO 


Versteegia cauliflora - IXO 


Aleisanthiopsis distantiflora 1 - RON 


Aleisanthiopsis distantiflora 2 - RON A2 


100/1.00 Aleisanthia rupestris - RON 


Greeneeae 
N Greenea corymbosa - RON 
Greenea sp 2 - RON 
Greenea sp 1 - RON A3 
blonga - RON 
Greenea sp 3 - RON 
cS EI cystiporon - VAN A4 
Cyclophyllum deplanchei - VAN 
| 
l 


100/1.00 
A 


Scyphiphora hydrophylacea - IXO 


10.63 Cremaspora triflora - CRE 


10.67 Pavetta platyclada - PAV 


Fernelia buxifolia - OCT 


—— Aulacocalyx jasminiflora - GAR 


Coffea arabica - COF 
Alberta minor - ALB 


Alberta sambiranensis - ALB 


Augusta austrocaledonica - RON 


H =m Luculia gratissima - OUT 
0.01 


E 14 


I ajority esian analyses (2 M generations) of Ixoroid d from th bined 
chloroplast data sets. Numbers above usd indicate ea support kom m ee a ies bera d E Se and 
50% o esolved ted. 


consensus ol Ba yesian posterior probabi liie: 
letters below the nod f the i and those on tl ] 


I cla s Oha groups ting 
recognized = es of the ee sensu je and Bremer (2000) are abbreviated as follows: ALB, Alberteae; COF, C 

RE, Cremasporeae; GAR, Gar ; IXO; Ixoreae; OCT, Octotropideae; PAV, Pavetteae; VAN, Vanguerieae. Species desir 
included to Ixoroideae from ido are dud RON. The outgroup is annotated OUT. Gray boxes represent the former 
delimitation of the tribe Ixoreae (sensu Andreasen & Bremer, 2000). 


Volume 96, Number 1 
2009 


Mouly et 
Pera of Ixora 


153 


support (BS 98%, PP 1.00), while it was not always 


the case for its internal nodes. 


DISCUSSION 


First, we compare the sequence characteristies of 
our rpsi6, rbcL, and trnT-F data sets with those 
published for other Rubiaceae groups. Second, 
discuss the paraphyly of Ixora and the competing 
circumscriptions of Ixoreae in light of the results of 
our phylogenetic analyses. Accordingly, a new tribal 
circumscription of Ixoreae and two new tribes are 
establishe 


SEQUENCE CHARACTERISTICS 


Our conclusions on the circumscriptions of Ixoreae 
are based on the combined tree (Fig. 2), as it is the 
best-supported hypothesis. Most of the informative 
characters are localized in the trnT-L and trnL-F 
spacers rather than the £trnL introns, consistent with 
on and Bremer (2002) and pid 


owever, many substitutions and m 


n. 
et al. (20! 
the Sn are within the relatively more conservative 
trnL introns and rbcL exon (Table 3); they appear as 
unambiguous synapomorphies for several clades. The 
rbcL data used in earlier studies about relationships 
within Rubiaceae are less informative than the rps16 
and irnT-F data, consistent with Razafimandimbison 
and Bremer (2001); rbcL sequence 
used for assessing interfamilial relationships (Chase et 
al, 1993; Clegg et al. 1994; Bremer et al., 2002; 
Shaw et al., 2005). 


e data are usually 


PARAPHYLY OF ¿XORA 


The present analyses demonstrate that Ixora as 
presently delimited is paraphyletic, unless Captain- 
cookia, Doricera, Hitoa, Myonima, Thouarsiora, and 
Versteegia are all included (Fig. 2). fxora coccinea 
seems more closely related to odo des Hitoa 
and Versteogia than it is to the six sequenced species 
of Ixora. Plus, Doricera and Myonima appear more 
closely related to the six sampled Ixora species than 
they are to I. coccinea. The inclusion of Hitoa, 
Sideroxyloides, and Thouarsiora with Ixora is support- 
ed by our studies. On the other hand, very few 
distinctive morphological features distinguish Ixora 
from its allied genera, such as hypocrateriform versus 

undibuliform corollas and soft versus bony fruits 
(Bridson € Robbrecht, 1985). With respect to Ixora, 
problematic generic circumscriptions were already 
observed in Rubiaceae, especially among large genera 
(e.g., Psychotria L., Galium L., Oldenlandia L.). 


Phylogenetic quis based on multiple nuclear 


and chloroplast markers and a much larger sampling 
of Ixora and affined taxa are being undertaken to 
specifically address the generic limits of Ixora. 


CIRCUMSCRIPTION OF IXOREAE 


The results of the present study do not support the 
two recently suggested circumscriptions of Ixoreae. 
Ixoreae sensu Andreasen and Bremer (2000) is not 
monophyletic, unless Seyphiphora is excluded and 


= 
E 
[e] 
a 
o 
a 
— 
pd 
$ 
(y) 
e 
oO 
un 
[i] 
5 
wm 
E 
a) 
e 
om 
om 
= 
(9) 
e 
2 
= 
w 
5 
a 
= 
w 
5 
[7] 
5 
bh 
A 

=> 


is polyphyletic, as it includes Seyphiphora and does 
not comprise both Captaincookia and Doricera. Our 


sensu Andreasen and Bremer (2000), represented 
here by Pavetta platyclada K. Schum. € Lauterb., and 
support the inclusion of the former two genera in a 

roadly 


circumscribed 
position of Aleis 


oreae. The phy joie 

nthia, en and Gree 
within the Ixoroideae, as revealed by Rova et al 
(2002), and its close relationship with /xora s.l., 
a corroborated by our results (Fi 

e phylogeny (Fig. 2) clearly ho that Ixoreae 
d a new circumscription. Whether Ixoreae should 
be recognized in a narrow sense (i.e., including only 


Ixora s.l) or in a broad sense (ie., Ixora sl. plus 
Aleisanthia, Aleisanthiopsis, and Greenea) is a matter 
of taxonomic judgment. 
narrow sense would 
m and Greenea groups to 

he tribal 


pueden morphologically (Table 4) but cause 


is would make Ixoreae s. str. 


nomenclatural novelties. In contrast, one may propose 
a broad circumscription of Ixoreae (including genera 
lto A3 in Fig i 


nomenclatural changes but perhaps necessitates 


of subclades . 2), which requires no 
descriptions of three new subtribal taxa. However, 
merging Aleisanthia, Aleisanthiopsis, and Greenea in 
Ixoreae woul e tribe morphologically hetero- 
geneous (Table 4) and diagnosable only by its 
aestivation contorted to the left (Robbrecht & Manen, 
2006) and its lack of raphides, two features commonly 
found throughout the Ixoroideae. We favor a narrow 
and well-defined circumscription of Ixoreae, which 
contains the following eight genera of subclade Al 


E 


g. 2): Captaincookia, Doricera, Hitoa, Ixora (as 
Soli Myonima, Sideroxyloides, Thouarsiora, 
and Versteegia. Accordingly, we recognize both the 
distinctive Aleisanthia—Aleisanthiopsis (subclade A2) 
and Greenea (subclade A3) at the tribal level. 
Ixoreae s. str. (e.g., subclade Al, Fig. 2) can easily 
be diagnosed by its articulated petioles, 4-merous 
flowers, corolla contorted to the left in bud, exserted 


154 


Annals of the 
Missouri Botanical Garden 


Table 3. Unambiguous molecular synapomorphies for recognized tribal and generic lineages. 


No. of synapomorphic 


Lineages substitutions Synapomorphic insertions Synapomorphic deletions 

Ixoreae irnT-L (spacer): 1 rps16 (intron): CTAAA irnT-L (spacer): T 
irnL-F (introns): 6 
rbcL (exon): 3 

Aleisanthieae trnL-F (intron): 2 irnT-L (spacer): CATAATCATATATTTCTA 

irnL-F (introns): CTTTTAATT 

Greeneeae trnT-L (spacer): 2 rps16 (intron): T irnT-L (spacer): GTATA 
trnL-F (introns): 5 irnL-F (introns): CAAAA 
rps16 (intron): 4 rps16 (intron): TTTTAT 
rbcL (exon): 3 

Scyphiphora — trnT-L (spacer): 7 irn T-L (spacer): AAACTA; T; T; TTTIT irnT-L (spacer): A; A 


trnL-F (introns): 10 


rps16 (intron): 5 


irnL-F (introns): GAAAATAT; T 


irnL-F (introns): TTAATGA; 
ATTCATTTAT 


rps16 (intron): T 


stamens, drupaceous fruits and uniovular locules, 
sclerified endocarp, and entire endosperm (Breme- 
kamp, 1934, 1937; De Block, 1997; Andreasen & 

mer, 2000; Table 4). The Aleisanthia—Alei- 
santhiopsis group (as “Aleisanthieae”; subclade A2, 
Fig. 2) is a well-defined clade that can be diagnosed 
by the woolly hairs on the abaxial leaf surface, 
scorpioid inflorescences, infundibuliform corollas, 
inserted stamens, multiovulate ovaries, and capsular 
fruits (Tange, 1996a, b; Table 4). Similarly, Greenea 

"Greeneeae"; subclade A3, Fi 
distinguished from 
Aleisanthiopsis group by protogyny, flowers without 


; a 
Ixoreae and the Aleisanthia— 
an obvious calyx tube, simply consisting of minute 
triangular free lobes (Puff et al., 2005; Table 4) and 
the lack of secondary pollen presentation (Tange, 


1996b) 


SCYPHIPHORA 


We find no support for the i placement 
of Scyphiphora as sister to Ixo str. shown b 
Andreasen and Bremer (2000). lts cos in Ixoreae 
sensu Robbrecht nd Manen (2006) is not corrobo- 
rated by our results, as it is placed in an unresolved 


ed 


trichotomy with Ixoreae s.l. and Vanguerieae within 
clade A (Fig. 2). For now, Seyphiphora should be 
considered to be unplaced but close to Ixoreae, 
Aleisanthieae, Greeneeae, an an ew 
distinct morphological faure separate ee d 
A (Fig. 2, Table 4). The 
, 1886; 
h 


ora is a rather 


from the taxa in clade A 
presence of t vules per locule (Baillon 
Puff & Rohrhofer, 1993) in Seyphiph 

unusual feature within this clade A, as well as within 
. Furthermore, fibrous and very thic 
pyrenes are also rare and probably represent an 


adaptation to sea dispersal, as S. hydrophylacea is 


restricted to mangrove habitats. The u 
contorted aestivation like Ixoreae (as here delimit- 
ed), Aleisanthieae, and Greeneeae, whereas Van- 
guerieae representatives (subclade A4, Fig. 2) have 
valvate aestivation. Co yes Scyphiphora resem- 
Alei Mines in its axillary 


inflorescences. Despite its scat position within 


bles Vanguerieae (and 


clade A (Fig. 2), Seyphiphora remains both unre- 


solved from molecular data a mbiguous in its 
putative affinities within clade A tribes according to 


morphology. 


TAXONOMIC SYNOPSIS 


The tribal concept of Ixoreae is better understood 
with results from molecular phylogenetic analyses. 
Bentham (1849) 
provisionally (Darwin, 1976: 
the tl 


used the name Ixoreae first but only 

he current use of 

name correctly attributes Ixoreae to Gray (1858). 

The following treatment presents an Bere de- 

scription of Ixoreae and recognition of two new tribes 

(for Aleisanthia—Aleisanthiopsis and for Greenea) for 
le 


no names are availa 


, Proc. Amer. Acad. Arts 4: 

9. 1858, non AME Griseb., nom. illeg., p.p., 
Fl. Brit India 37. 18 as subtribe 
[including Coffea], nec Ixoreae Hook. f., nom. 
illeg., p-p., Gen. PL, 2: 9, 22. 1873, as tribe 
[including Coffea], nec Ixoreae K. Schum., 
nom. illeg., p.p., Nat. Pflanzenfam. 4: 1891; as 
tribe [including Coffea and Pavetta]. TYPE: 
Ixora L 

Coffeeae DC. p.p., Ann. Mus. Natl. Hist. Nat. 9: 217. 1807, 


ls pue A. Gray. 


as tribe. 
Pavetteae A. Rich. ex Dumort. p.p., Anal. Fam. Pl: 33. 1829, 


Volume 96, Number 1 
2009 


Mouly et 
Pea ay of Ixora 


155 


Table 4. Summary of selected morphological characters 


relevant to Ixoreae alliance as presently circumscribed. 


Unambiguous morphological features that characterize the recognized tribes and Scyphiphora within the main clade A of 
in bol 


Figure 2 are in boldface 


Ixoreae Aleisanthieae Greeneeae Vanguerieae Seyphiphora 
Wooly hairs absent resent absent se 
Petioles articulate not articulate not articulate not articulate articulate 
Domatia absent presen present present absent 
Inflorescence terminal/cauliflorous | terminal/axillary terminal axillary axillary 
position 
Inflorescence type ymo: scorpioid scorpioid ymose mo 
4-merous 5-merous 5-merous 4- to 5-merous 4-merous 
Corolla aestivation contorted contorted contorted valvate contorted 
Calyx tube present present absent present present 
Stamens ert inserted inserted inserted/exserted sert 
Ovule numbers l-ovule/carpel many ovules/carpel many ovules/carpel ^ l-ovule/ca o 2-ovules/carpel 
Fruits ce capsular capsular drupaceo ace 
Sclerified endocarps sent sent 
Dispersal mode zoochorous anemochorous anemochorous zoochorous sea currents 
Secondary pollen present present present present 
resentation 
Pollen grains colporate pororate/colporate colporate po(ro)rate colporate 
Sexuality protandrous protandrous protogynous protandrous protandrous 
Ixoreae representatives are easily ia Gage hairs on the abaxial leaf surface; inflorescence 
by the following combination of characters: leaf scorpioid, axillary or terminal; flowers 5-merous, 


ae articulate; Msc c as Du to large 
cym erminal (sometimes on brachyblasts o 
cauliflorous); flowers 4-merous, protandrous; aestiva- 


E 


tion contorted to the left; corolla hypocrateriform 
(Captaincookia excepted); stamens exserted; second- 
ary pollen presentation; ovary )-locular; locule 
uniovulate; fruit drupaceous, 2(to 7)-locular; pyrene 
leathery to crustaceous; seed with circular adaxial 
excavation extending into a basal groove; endosperm 
entire 


Included genera: Becheria, Bemsetia, Captaincook- 
ia, Doricera, Hitoa, Ixora, Myonima, Patabea, Sider- 
oxyloides, Thouarsiora, Versteegia. 

Ixoreae (subclade Al, Fig. 2) as presently circum- 
scribed is supported by several unambiguous synapo- 

morphies a three epo markers (e.g., 
Table 3) and hom e orphology. The 
following genera are m. el as Íxora 
synonyms: Becheria, Bemsetia, Hitoa, Patabea, Sider- 
oxyloides, Thouarsiora, but the paraphyly of fxora 
questions generic circumscriptions within Ixoreae. 


2. Aleisanthieae Mouly, J. Florence & B. Bremer, 
tribus nov. TYPE: Aleisanthia Rid 


Tribus nova quae ab Ixoreis A. Gray lamina foliari subtus 
plerum inflorescentia scorpioidea, flor- 
ibus pentameris atque ovarii loculis multiovulatis praecipue 


differt 


que omnimo lanosa, 


Aleisanthieae is distinguished by the following 
combination of characters: blades covered with woolly 


small, protandrous; calyx lobes small and triangular 
to rounded; aestivation contorted to the left; corolla 


infundibuliform; stamens included; secondary po 


presentation; ovary 2-locular, multiovulate; E 
capsular. 
Included genera: Aleisanthia, — Aleisanthiopsis, 


Greeniopsis Merr. 
The Aleisanthieae clade comprises ihe duo Sena: 


east Ásian genera eisanthia 


an lis 
resolved as sister to the pantropical roe s. str. 
Aleisanthieae is well supported by unambiguous 
molecular evidence, with synapomorphic substitutions 
and indels (Table 3 i 
leaf surface (Table 4) represent a rare feature in 
of 
b) also associated the 
9), with 
seven species, with Aleisanthia and Aleisanthiopsis, 


). The woolly hairs on the abaxial 


Ixoroideae and may be regarded as a synapomorphy 
the Aleisanthieae. Tange (1996 
Philippine genus Greeniopsis (Merrill, 


but reported no presence of woolly hairs on the leaves. 
However, G. discolor Merr. does possess these typical 
hairs (Mouly, pers. obs.). In the genera Aleisanthia and 
Greeniopsis, stylar complexes are observed, similar to 
those described in Vanguerieae (Igersheim, 1993; 
, 2002), 
Aleisanthiopsis distantiflora (Merr.) Tange (Tange, 


Lantz et al but these were not observed in 
1996b). The pollen grains are 3-pororate in Aleisanthia 
and Greeniopsis, although the pollen stains in Alei- 
anthiopsis are 3-colporate. Inflorescences are axillary 
in Aleisanthia but are terminal in Aleisanthiopsis and 
Greeniopsis. Our tentative placement of Greeniopsis in 


156 


Annals of the 
Missouri Botanical Garden 


Aleisanthieae based on morphological resemblance 
needs to be confirmed from molecular data. 


3. Greeneeae Mouly, J. Florence & B. Bremer, 
tribus nov. TYPE: Greenea Wight & Arn. 


ribus nova quae ab Aleisanthieis floribus proterogynis, 
calyce sine tubo manifesto atque stigmatibus linearibus sine 
pollinis praesentatione secundaria praecipue differt. 

The tribe Greeneeae is characterized by the 
following combination of characters: leaves without 
abaxial woolly hairs P ds glabrous); inflorescence 
terminal, scorpioid; s 5-merous, small, proto- 
gynous 
torted A the left; corolla infundibuliform to campan- 


p 
calyx without pos tube; aestivation con- 
ulate; stamens included; primary pollen presentation; 
ovary 2-locular, multiovulate; fruit capsular. 

Included genera: Greenea, Spathichlamys R. Parker. 
he nce of secondary pollen presentation in 
Greenea (lange, 1996b) is uncommon within the 
studied lineage (clade A, Fig. 2). Furthermore, the 
protogyny of the Greenea species is also unique: the 
stigmatic lobes are partially to completely exserted 


pollinators only after o 
enlargement of the corolla tube. Otherwise, in all 
species with secondary pollen presentation, an initial 
ce male stage precedes the m cni 
(Nilsson et al., 


stage during anthesis 
96). The disappearance of s 


male 
Puff et al., 
pollen presentation might be linked to the adaptation 


eco ey 


to protogyny. Greeneeae is supported by many 
unambiguous molecular synapomorphies (see Ta- 
le 3). The species G. commersonii (Korth.) Boerl. 
(Boerlage, 1891), previously described in the genus 
Rhombospora Korth. (Korthals, 1850), also shares the 
diagnostic morphological features of Greenea, but was 
not investigated. The monotypie Indonesian genus 
e (Parker, 1931; Ridsdale, 1982; Tange, 
96b) is identical morphologically to Greenea from 
s it differs only by the corolla that splits and rolls 
up at anthesis (Ridsdale, 1982). 


CONCLUSIONS 


The present study clearly shows that Ixora as 
currently delimited is paraphyletie and Ixoreae sensu 
Andreasen and Bremer (2000) and sensu Robbrecht 
and Manen (2006) are non-monophyletic. The mono- 
phyly of a broadly circumscribed and morphologically 


heterogeneous Ixoreae alliance (including Ixora, 


Aleisanthiopsis, and Greenea) is strongly supported. 
In order to recognize monophyletic and morphologi- 


cally consistent groups, we adopt a narrowly circum- 
scribed Ixoreae ui nu Becheria, Bemsetia, 
taincookia, Doricera, Hitoa, the paraphyl 
Versteegia) and erect the Aleisanthia—Aleisanthiopsis 
and Greenea groups at tribal level, Aleisanthieae an 
Greeneeae, respectively. Finally, the Malesian Alei- 
santhieae and the pantropical Ixoreae s. str. are sister 
groups and the Southeast Asian Greeneeae is sister to 
th Aleisanthieae clade 


e Ixoreae 


Literature Cited 


Akaike, 74. A new look at the statistical model 
E IEEE Trans. Automat. Contr. AC-19: 
716-72. 

Alejandro, ; G. Razafimandimbison & S. Liede- 


ITS and # data and its pares for generic limits 
in Mussaen m eae (Rubiaceae). Amer. J. Bot. 92: 544—557. 
Andreasen, Bremer. 1996; Phylogeny of "n 


subfamily Ixoroideae (Rubiaceae). Opera Bot. Belg. 
—138. 
——— & —— 000. Combined phylogenetic analysis in 
Hublasesc: oiu Morphology, nuclear and 
Mo DNA 
Arènes, J. 1960. A propos de pee genres Malgaches de 
Rubiacées. Not. Syst. (Paris) 16: 6-19. 
Aublet, F. 1775. Histoire des plantes de la Guiane Françoise, 
vant la méthode sexuelle, avec plusieurs 
mémoires sur différens objets interessansy relatifs à la 
ure & au commerce de la Guiane Frangoi 
notice des plantes d'Ile-de-France, Vol. 1. P 
Didot j jeune, Librairie de la Faculté de Médecine, Londres 


1879. Sur les limites du genre Ixora. 
Adansonia 12: 213-219. 
ier Sur Woo ES Scyphiphora. Bull. 
ens pee "d 4-1 
a G. pO T NN plantarum 
Novae RM pU -B. a & uo Paris. 
Bentham, G. 1849. Rubiaceae. Pp. W. J. Hooker 
dion y n i cl Milit, Toten 
Plan 


anaes 


Suela Inferae—Heteromerae. E. 
Bremekamp, C. E. 


Pavetta L. Rep. Spec. Nov. R 


Lantz & Ke i Kim. 1999. More characters or more taxa for 
t eny in a case study from the Coffee family 

E pon Bot. 48: 413—435. 

, K. Bremer, N. Heidari, ES Erixon, R. G. 


Olmstead, 


coding DNA at pn taxonomic levels. Molec. Phylo- 
genet. Evol. 24: 274-301 


Volume 96, Number 1 Mouly et 157 
2009 ey, of Ixora 
Bridson, D. M. 2000. The aa of Tsiangia (Rubiaceae). ^ Giribel 1999. On gaps. Molec. 


Kew Bull. x 1011-1012 

——— € E. Robbrecht. 1985. Further notes on the tribe 
Pavetteae (Rubiaceae). Bull. Jard. Bot. Natl. Belg. 55: 
83-115. 


But, P., T H. Hsue & P. T. Li. 1986. Tsiangia, a new genus 
bas ay pen ee (Rubiaceae). Blumea 
31: CE 

Chase, M. W. & H. Hillis. 1991. Silica gel: An ideal material 
for field preservation of leaf samples for DNA studies. 
Taxon E 215-220. 

D. E. Soltis, R. G. Olmstead, D. Mor, 

B. D. Mishlen M M. R. ue R.A. 


a n, D. H. co 
2d H. C. Hills 


du 


J. Kim, C. F. ue J. F. Smith, G. 


E 1993. 
Phylogenetics E seed plants: An analysis of nucleotide 
fro 


sequences from the plastid gene rbcL. Ann. Missouri Bot. 
Gard. 80: 528— 

Clegg, M. T si G. H. Learn Jr. & B. R. Morton 
199 


Can three incongruence tests 
predict en data ‘should be combined? Molec. Biol. Evol. 

Darlu, P G. cione 2002. When does the incongruence 
length O ua fail? Molec. Biol. Evol. 19: 432-437. 

Darwin, S. 1976. The es tribal, and subtribal 
cis ds of S axon 25: 595—610. 

. À synopsis " the indigenous genera of Pacific 
Rubicone. Allertonia 2: 1—44. 

De Block, 1997. Biosystematic Studies in the Tribe 
Pavetteae p eg cn roideae). Ph.D. Thesis, Univer- 
siteit Antwerpen, Antwerp. 

1998. The African ea of Ixora (Rubiaceae— 
der d Opera B ot. Belg. 9: 1-218. 

. Escriva, C. Gallut, V. Barriel, P. Kourilsky, 

he M 


uence of 

petra flw viatilis. Bearin 

of cyclostomes. a Biol. Evol. 17: 

L. Doyle. 1987. A rapid DNA oe 

procedure io small quantities of fresh leaf tis 

Phytochem. Bull. 19: 11-15. 

i . J. 1829. Analyse des familles de plantes, 
avec Vindication des ae genres qui s’y rattachent. 
Casterman, Tournay, Fran 

Farris, J. 1989. 


al of the 
ngs on the es 
529 


The retention index and the rescaled 
19 
Kállersjó, A. G. Kluge € C. Bult. 1994. Testing 
e 5-319. 
5. Confidence limits on Due qun An 
approach d ins Bootstap: Evolution 39: Kd 

Fosberg, F. R. 19. a 

Occas. i des Pauahi os Mus. 13: 245-293. 
Freudenstei & M. W. Cha . 2001. e of 
merced adlb- -c intron ves n Orchidaceae: 
Utility and coding of length-change RPM Syst. B 

ra 


06. Supplementum carpologiae (seu 
nuati operas Josephi Gaertner de Fructibus et 
Seminibus Plantarum 3). Richter, Leipzig 


= 


: Ee 

pem Evol. 13: 132-143. 

oo A. 1858. Notes upon some Rubiaceae, collected in the 

nited States South S ploring Expedition under 
d Wilkes, with characters of new species 
: 39-40 

86. Ixoras malgaches nouveaux à fleurs 

DEN Mu n 60: 243—250. 

Hallé, N 3. Captaincookia, genre nouveau monotyp up 

recaló de Rubiaceae-Ixoreae. Adansonia, Sér. 2 
13:1 

Hear. Li a 1873. Ordo LXXXIV. Rubiaceae. Pp. 7-151 in 

ham & J. D. Hooker (editors), Genera plantarum ad 
e. imprimis in herbariis kewensibus servata 

. Reeve & Co., London. 
. & F. Ronquist. 2001. MrBayes: Bayesian 
inference of F phylogenetic trees. Bioinformatics 17: 754-755. 
B. Larget, R. E. Miller & F. Ronquist. 2002. 

Parental applications and ae of Bayesian inference of 

phylogeny. He Biol. 51: 673—688. 

Igersheim, A. 1993. Gynoecium development in Rubiaceae- 
Vanguerieae, with o reference to the 
head”—complex and secon 
Syst. Evol. o 175-190. 


“stylar 
pollen presentation. Pl. 


9 iA plantarum uud ordines 
iid Barroi ls, gs 
and the 


eae of anurans. Syst Zol: 18: 1 -32. 

Korthals, W. 1 Overzigt der Rubiaceén van de 
Y Oostindische Kolonien. Ned. Kruidk. Arch. 
2: Ea 

Lantz, H., K. An dreasen & B. Bremer. 2002. Nuclear rDNA 

ITS sequence data used to construct the first phylogeny of 

Vanguerieae Pome Pl. Syst. Evol. 230: 173-187. 

& B. Phylogeny inferred from 

morphology " ‘DNA um Puoi i M - well-supported 
groups in ). Bot. J. Linn. Soc. 146: 
257-283 

Lloyd, p. " a Y L. Calder. IM Multi; ras gaps, 
class character reliability lor 
When c J. Evol. Biol. 4; 9 

Merrill, E. 9. New or noteworthy ibas plants, 
VIL P e i ne 4: 247-330. 

Montrouzier, X. 1860. Flore de ile Art (prés de oo. 
Calédonie). Mém. Acad. Roy. Sci. Lyon, Sect. Lett. 


SS emer. 


: 7. Systématique de la Tribu des Ixoreae A. 
Gray (Rubiaccn Phylogénie, mace hie et Taxono- 
1 .P - Thesis, Muséum National d'Histoire Naturelle, 


ea J. 1899. Plantes nouvelles des Îles de la Société. J. 


T nandrianina, B. Pettersson & J 
11 


s platythyrsa (Rubiaceae). Pl. Syst Evol. 170 
165. 
e E J. 4. MrModeltest 2.0. Pro 
destete A Tu desk Evolutionary Biology Cube 
Uppsala University, Uppsala. 

Olmstead, R. G., H. J. Michaels, K. M. Scott & J. D. Palmer. 
1992. Monophyly of ie i emn be ilerici of 
their major | of rbcL 


Ann. Missouri Bot. ud 79: 249 s 


158 


Annal 
Missouri Botanical Garden 


Oxelman, B., M. Lidén & D. Berglund. 1997. Chloroplast 
rpsl6 intron phylogeny of the s EAE (Caryophylla- 


Pl. Syst. Evol. 206: 
Parker, R. N. 1931. VI a es ae Plantarum 
novarum in © regii conservatarum. Decas CXXV. 


herba 
1244. Spathichlamys. Bull. Misc. Inform. Kew 1931: 
42-43. 


Cha ca & V. Ch we d 2005 
to 


Park, Wildlife and Plant Conservation Department, Bangkok. 
, E. Robbrecht, R. Buchner & P. De Block. 1996. A 
survey of secondary pollen presentation in the Rubiaceae. 
Opera Bot. Belg. 7: 369—402. 

& U. E 1993. The character states and 
taxonomic position of the m 


aken o 
properly classified, now reduced by their natural affinities 
to the proper natural orders and tribes, Vol. 3. C. S. 
Rafinesque, Philadelphia. 
Rambaut, A. 1996. Se-Al, version 1.dl. Sequence alignment 
<http://tree.bio.ed.ac.uk/software/seal/>, ac- 
9. 


. & B. Bremer. 2001 [2002]. Tribal 
e (Rubiaceae): pr from 
d mo a data. Syst. Geogr. Pl. 71: 


—— 2. Phylogeny and classification of 

a s.l. ud. inferred from molecular s 

rbcL and trn T-F) and morphological data. Amer. J. Bot 
1027 ol. 

e A. 1829. Leman sur la faris: des Rubiano 
contenant | 
mp noni d pedcs Tastu, Paris 

iie N. 1912. New ps rare ei plants (ña VD. 

traits Branch Roy. Asi 1: 1-43. 
New 2 rare plants us the Malay 
Peninsula. n Fed. Malay States Mus. 10: 128-156. 
n C. E. 1982. oe. remarkable Rubia- 
e. Blumea 28: 143-1 

TA E. 1988. e woody Rubiaceae: Character- 
istic features and progressions. a to a new 
subfamilial classification. Opera B ot. Belg. 1 ud 

& J.-F. Manen. 2006. neages 
of the coffee family (Rubiaceae, angiösperms]: ae bined 
analysis (nDNA and cpDNA) to infer the position of 
Dd apelta and Luculia, and supert truction based 
on rbcL, rsl6, trnL-irnF and atpB-rbcL data 
classification in two sul pu us and 

ioideae. Syst. Geogr. Pl. 76: 85— 

Rouhan, E J.-Y. Dubuisson, F. Ratna Desk 
Motley, J. T. Mickel, J.-N. Labat & R. C. Moran. 2004. 
Molecular phylogeny of the fern genus i 
Era based gi chloroplast noni B DNA 


Indian Ocean 


. A new 


sequences: C 
area. Molec. 2 Evol. 33: 745—163. 
Rova, J. H. E., Delprete, L. Andersson & V. A. Albert. 
2002. A "n F cpDNA sequence study of the Condami- 
ae-Rondeletiea ae complex with implications 


on the phylogeny of Hulhaceno. Amer. J. Bot. 89: 145-159. 


Saghai-Maroof, K., M. Soliman, R. A. Jorgensen & R. 
Allard. 1984. Ribosomal DNA spacer length ae li 
phism in ndelian inheritance, chromosomal 
location, and population dynamics. Proc. Natl. Acad. Sci. 
U.S.A. 81: 8014-801 

Sañdwith; N. Y. 1937. Contribution to the flora of tropical 
America: XXX. New species and records from British 
Guiana. Bull. Misc. Inform. p 1937: 100-112. 


arley: 


Relative utility of 21 noncoding chloroplast DNA 

sequences for phylogenetic analysis. Amer. J. Bot. 92 
—166. 

Smith, A. C. & S. P. Darwin. 1988. Family 168. Rubiaceae. 
Pp. 143-362 in A. C. Smith (editor), Flora Vitiensis Nova, 
A New Flora of Fiji Sagan! tes Only), Vol. 4. Pacific 
Tropical ave Garden, Lawai, Kauai, Hawaii 

Staden, R. 1 The Staden oe analysis package. 
Molec. Binh 5: 233-24 

Swofford, D. L. 1993. PAUP: uds Analysis Using 
Parsimony, 3.1.1. Computer program listed by 

P*: Phylogenetic Analysis Us sing Parsi- 

, Sunderland, 


elly, G. Patou & J. Bouvet. 1991. 

Universal primers for amplification of three non-coding 
regions of chloroplast DNA. . Molec. Biol. 17: 
1105-1109 

Tange, C. 1996a [1997]. Studies in SE Asiatic Rondeletieae 
I: The West Malesian endemic genus Aleisanthia (Rubia- 
ceae). Nord. Bot. 16: 563—570. 

——— b [1997]. ee in SE Asiatic aaa ead 
II: Aleisanthiopsis (Ru new genus from Bor 
Nord. J. Bot. 16: 571-578. 

Valeton, T. 1911. Rubiaceae. Pp. 437—519 + illustrations in 
H. A. Lorenz (editor), E de l'expédi tion Mee 
que néerlandaise a la elle-Guinée en 1907 et 1909, 
Vol. 8, Botanique. one et Imprimerie Ci- Devant E. J. 
Brill. Leiden. 

van Dijk, M. A., E. Paradis, F. Catzeflis & W. W. De Jond. 
1999. The virtues of gaps: Xenarthran Eana) 
ebrii supported s a dd deletion in alpha A 

rystallin. Syst. Biol. 4. —106. 

Verdot B. 1983. ae on Mascarene Rubiaceae. Kew 

e —574. 

- 108; pe Pp. 1-135 in J. Bosser, T. 

& Marais (editors), Flore des 
Maurice odri 


Vieillard, E. 
recueillies par M. Eugè 
Marine. Bull. Soc. Linn. Normandie 9: 332-348. 

Wight, R. & G 2 . 1834. Order DER 
Rubiaceae Jus 443 in R. Wight & 

Arnott Gee: i5 florae ie ie 

orientalis: Containing abridged descriptions of the plants 

found in the peninsula of British India, arranged according 

to the natural system, Vol. 1. Parbury, Allen € Co., 
ndon 


Lo 
Yang, Z. 1994. Maximum e hae ne estimation 
from DNA sequences with variable ra sites 

m methode P Molec. Evol. 39: "306. 
a zh e Irwin & B. A. Payseur. 2001. ae of 
ermine data oe for slow Loris 


the 
Pu dis Pu Biol. 50: 408—424. 


159 


Mouly et al. 


Volume 96, Number 1 


2009 


Paraphyly of /xora 


«OcbLISQnd — «69pLI903 «SpELIS0 Teose3epe (OIN) 28381 T opuvopuvuaqoy 2p 72) uostouqopy seugry XƏ o[[ouro]] szjm4o7m] DIOISLDNOY p 
xt@PL 180d xSOpLI8N a ZPOGOTAA UBo quer) (OIN) £6911 “9 401401 “bowl umauof sopiojéxodopig 
xZEPLISNA xSLELISO xOSPLISAT vwyuvT ug (S) 96 P 19 y uig uyoe2) “q 7D) oonpajdospég vaoydrjdkog 
«O£PLISnd x S2PL 180 xOPPLISNA SOUGIBOSY IA] (d) 9351 H q 2242407 “Opie A (UET) naopjora DuruoÁ y 
«ScbLISQnd — «ILPLIS xLpELISO vruezue T, (S) rog6 `Ò oq] uospug SISUAUDZUD] `] 
«SCbLISQnd  +89PLISNA xPPELISO visou£[oq ouor (d) GPE T oousio]4 x "y 4 mopy "ourexg (wəpno) məyooy T 
«PORLISQnd — *L9_LI8MH «epPLIS0 vuv?) (OIN) I09S I? 19 "y "M n2429 "pueg smussumb T 
€3PL18NA x9OPLISN a EPOSOZAA vduoyy (sat GFO6 “O MT uo( 5 xo pea, pupruoséopaf T 

USpoms (sam) POLE “Y wag 
OP9E8X  *b0bLI8MH TP9SOZTA ‘ogy ur pereanno (sam 6123 “q “2249 (Sdn) 6ILE “Y Puig "p 22u12209 `] 
«L@pLiend — «OLPLISO «OPPL18NA puereug, (d) £üFgF "I? 19 "y ussav] uo(q 4) xo qe stuounsq T 
«L@pPL 1804 x€9PL 1804 xChpL 18nd uoquz) (OW ¿201 P 19 ^4 y Koppoag DA »podéjoniq niox] 
«OCRLISnd — «coPLIS3 «IPELISO0 viseuA[oq Yuen] (d) us f 220240], pneopew sisuaasoou Dony 
«LIRLISQH — «ISPLIS «LEPLISNA SOUGIBOSY [A (NVI) Te X T Prosa] "opio A (3 Heg) SD noo py va2210(q 
«SIPLISQnd  +9SPLISNA «O€PL 180 vruopo[e?) MON (d) 193 "4| aruasouu] ap y &mopy IPH ^N. 2022103411 pi000utpidpy) 
9€9IOXT 
«£IPLIS — «SepLIS0 6£950c43 vun) (OW) 3291 7? P H "H pros F AOH vaojfrumuspt x&qmooongny 
9eoruoprez) 

S661 “Te 19 1owerg 8661 
1g9¢8x GOPZOLAV 8€£0P004V umouyun “Te 19 OMG “6661 PAOY 29 uossropuy "I Dogon naffo) 
w90 

9661 “Suerg X 
999997 OPOTOZAV D6600ZAV umouyun uoseoIpuy ¿000% MICA | ¿000% ‘uossIod "umuos “Sf DO DIOASDUILAL”) 
ovorodseumo1) 
PIESZAC TCI x PLPLTENA SPOSOZAA Teose3epe (SAM 09€ `9 `g vosiqunpununfozoy ODBAB]) Xo o[[ouoH SISUGUDAQUIDS y 
xOTPLISNA TSPLT8NA LE9SOGAA 1eoseSepe ly (SdM) 859 9 `g vosiqunpununfozoy qeg 4ounu puoqpy 
obo oq [y 
«6cbLISQnd — «cLPLIS «gpPLIS03 umouyun (sat) #9008 LO yoomg (TEM) Duassmas oinomq 
dnoising 

PH SToqumu A Ju Si9oqunu 9rsda suis) SQOUBIOJOY /SIOTONO A Soroodc 


UOISS920V 


UOISS9200 Y 


SToquinu uOISS920V 


'SUOISSQOOP Xuequo:) 19778 3SLI9]s? ue YIM porejouue 


ore soouonbos pousiqnd Amoy "uorreuuojgur IoyonoOA Joye sosommored ur post] ore sur&uoroe umrreqiopp ‘ojus uou suotgor oruos je 01 19jo1 10 *o[dnpnur usya suorgor oruo$ jo oouereodde 


oq MOT[OF suoreorpur *uum[oo oouotopoy/roqonoA otn UT 'sosATeue re[noopour ur posn soouonbos 7294 pue *7- ue *grsda jo sroqumu uoresoooe pue ‘UISLIO ‘stoyonoa uourroodc 


1 xipueddy 


Annals of the 


160 


Missouri Botanical Garden 


JOUIOJQq "d m zyueT ^ungezeq 


+ PTPLIENA 62960710 SEPLT8NA sorowon (d) 0835 I? 1? "N- 1»qv] (ooeaer) xo udg) uosodysho wumipniodog 
x9TPLISNA TEOSOZAA OP9SOG AA AN (d) gzz “4 eruasouu] 3p y Amow 91904 1ayoundap und dop do 
QROLIQUSUR A 
- ¥8SPL 1909 «Sep 1804 pue[neur, (d) ZSZ P 12 uva ^4 7) woyasnag g “Opu "ds naua245) 
«SIPLISnd — 09PL 180 - pue[reug, (d) OPTEP T? 12 "y uos] g yepur "ds n2ua245) 
«6 TPLI8MH — «IOPLISQ xOPPLISNA puerro (d) 8LEEE 'S 'S uosamq P Y uos] T yepur "ds pauses) 
- 6€PLT8NA xOSPL T9013] pere, (d) ISPEE S'S usse] P Y UWT qme) o5uoggo 9 
x LEO@STAV TOGZPZAV puereup, (AVV) ZOIPP I" 12 "y uos] WIOA (pef) vsoquilio> nauaasy 
aqu 
«cIPLISQd  PSPLISNA 8£9S0c.11 vruopo[e) MON (d) LEG “q enio20uuj p y 4mopy "H f CuSuo1g) voruopopoooarenp oisnsny 
i 859c6 LIV €O6TPTAV oou1og (n VV) 2269F “9 ?3un :z uorssoooe DIOJfUDISIP Y 
«ITPLISQW — «SPL TSN «PEPLISNA ooug (D IP I? 19 "y f "d 49]592y :[ uorssooot Sue, (uow) paoyfiupisip sisdongupsioyy 
- 0996 LIV ZOGZPZAV viskepe]y (nvv) IZISP 2 Sun] "pra CIPD suasodna pngrupsiopy 
9eono[ppuowd 
ISr9Tefv i Z800Z€[V BOUIN) MON (San) "ws y tejo 9 "d. pzo4q "qioneT y "umuos y »ppojpmd oyad 
QBONOAR gd 
POLO8Z[V me c68TTcAlV uMouxun 0007 ‘wg m uoseoJpuy E tIqndun “Bao “ULISeO "d ne) pyofixng D1]2u42 4 
oeoprdomo12 
BISQUOpUT 
e€EPLIend  +9LPLISNA xISPLISNA ^1080g ur pojeanno (San) "ws y UAW 9 d PA uojep[eA n4opfijmoo IB 9918194 
PH sIoqunu ALUA sroquinu 9rsda SUISTIC) SQOUBIOJOY /SIOTONO A Soroodc 
UOISSQ00 V UODISSIDOY SToquimu UOISS9200 V 


'penunuo) <q xipuoddy 


EVOLUTIONARY TRENDS, Sylvain G. Razafimandimbison,? Henrik Lantz,” 
MAJOR LINEAGES, AND NEW Arnaud Mouly,? and Birgitta Bremer? 
GENERIC LIMITS IN THE 

DIOECIOUS GROUP OF THE 

TRIBE VANGUERIEAE 

(RUBIACEAE): INSIGHTS INTO 

THE EVOLUTION OF 

FUNCTIONAL DIOECY” 


ABSTRACT 


the Pal Vanguerieae in the subfamily Ixoroideae s.l. (Rubi )h u 
bien éxblished as a result of a series of cane stüdies conducted by Lantz and Bremer Tie genus Candia Lam. was 
shown in their incl to be ise polyphyletio, and a ılar; gely was for 
The di Fani 


Ir within Vanguencaes 
e ut 140 species classified in eight nthium (C 


Bullockia Bridson) D aca kion Pon , Leroya Cavaco, Néoleroyá Cavaco, Peponidium (Baill) m Pseudopeponidium 
Homeless ex Arenes, Prosa Comm. ex Juss., and Seyphochlamys Balf. f. We sequenced 79 Vanguerieae taxa and performed 
(ETS and ITS 1) pinpoint the phyloge netic t E 

the Comorean and Indian Ocean Canthium and the Southeast Asian Canthium confertum Korth. group in Vanguerieae; (2) 
evaluate the phylogenetic utility of three taxonomic P chanel: (brace type, locule annee and fruit shape) vicus and 
currently 1 for delimiti in the dioeci ip 1 ü MS tae in o The 
result further disi ion of Canthi L,a e a Ms Malagasy 1 hown for the first 
ume to be closely related 9 aia Similarly; e. onfertum apy h Veloso dii with Cyeophllun. "Hook. f Te 


1; p 


as genera: Bullockia (Bridson) Razafim., Lantz & B. Bremer, h levated fi E thi ubgen. Bullockia Bilson a as ellas as 
Cyclophyllum, Peponidium (ineludin; ng all Comorean, Malagasy, and Seychellean Canihium species), gua Tyro Pr. 
llo 


Doone as Ne oleroya, P. 


ckia 


F E 
Bull | R fi Tanit & B. Bremer. B. le ii (Brid 


mer. 
B.ü Brid Razafim., Lantz & B. Bremer, B. mombazensis E wur ud Bremer, B. EE 
(Bridson) as € & p Bremer, and B. setiflora (Hiern) Razafim., Lante & a r Burthormore, the sults seem to 
i followed by | I atl 

+) Eat dE rÍ T) e duet lad d P. iria s TI Mal TEA TN " } Li + 


whereas the Comorean Peponidium and the African Pyrostria appear to have originated from Malagasy progenitors. 
ey words: Biogeography, Bullockia, Canthium, Cyclophyllum, ETS, functional dioecy, ITS, Peponidium, Pyrostria, 
Rubiaceae, Vanguerieae. 


The tribe Vanguerieae comprises between 600 and unpublished talk) genera depending on the generic 
700 species of trees, shrubs, geofrutices, and limits used. Vanguerieae belongs to the subfamily 
imbers, which are classified in 17 (Lantz € Bremer, Ixoroideae s.l. (Bremer et al., 1999) of the coffee 
2004, 2005), 27 (Robbrecht, 1988), or 37 (Bridson, family (Rubiaceae) and is a monophyletic group, 


e thank Aaron Davis (Royal Botanic Gardens, Kew, England), Thierry Pailler (Université de la Réunion, cele and 
g vie uc e. (Missouri Botanical Garden Program kindly providing DNA material for this s 
Ministére des Eaux et Foréts (MEF) and Association Nationale pour la Gestion dis Aires E nd (ANGAP) for i issuing 
collecting permits for SGR; Missouri Botanical Garden Program, Madagascar (especially Lalao Andriamahefarivo) fo 
arranging collecting permits for SGR; Désiré Ravelonarivo for being an excellent field assistant in Marojejy National Park; ds 
following herbaria for providing loans and access to collections: BR, GB, as Universitaire de la Réunion, K, MO, P, S, 
1 1 1 type specimens of five 


m 
=} 
a 
un 
= 
3 
o 
3 
eo 
= 
i 
= 
iz] 
E 
5 
2 
= 
o 
a 
e 
Es 
A 
= 
= 
a 
‘an 
E 
=] 
a 
= 
e 
Uu 
e 
l^: 
a 
E 
i=] 
ge 
E 
E 
bed 
=] 
e 
[em 
= 
E 
la] 
de 
e 


Bergius Foundation, Royal Swedish Academy of Sciences, and Department of Botany, S m University, SE-10691, 
Stockholm, Sweden. MNA Mer CUN MCN se; e; birgitta.l e 
o mn Department, Evolutionary Biology Centre, Uppsala University, SE-75236, Uppsala, Sweden. henrik. 
lantz@ebc.u 
doi: 10. 3417/2006191 


ANN. Missouni Bor. Garp. 96: 161—181. PUBLISHED on 23 ApriL 2009. 


162 Annals of the 
Missouri Botanical Garden 
which can be recognized by a combination of the phyllum Hook. f. (Hooker, 1873) was, however, 


following eight characters: absence of raphides, 
axillary inflorescences, valvate corolla-lobe aestiva- 
tion, a unique type of pollen presenter, uniovulate 
locules, pendulous ovules attached toward the top of 
septa, drupaceous fruits, and large embryos wi 

superior radicles (Bridson, 1987, 1992, 1998; Lantz et 
al., 2002 


flower and fruit sizes and shapes, which attract 


). There is a considerable variation in both 


various pollinators and a range of animal dispersers, 
respectively. The members of the tribe inhabit various 
habitats, ranging from evergreen rainforests to dry 
deciduous forests to kd dun = an 
occupy ange extending fro 
tropical Ta E jd pM the iuam ae 
islands, to the Pacific region and northern Australia. 
The African mainland is the center of diversity with 
a. 350 species, and Southeast Asia and Madagascar 
iet ca. 150 and 120 
Madagascar, at least 20 new species of Peponidium 
wd Arénes (Arénes, 1960) and d Comm. 
ex Juss. (Ju 1789) are yet to be described 
(Razafimandimbison et al., unpubli nr ata). 


species, respectively. In 


ussieu, 


While a member of Vanguerieae is MEUM easy 
to recognize, the generic limits within the tribe have 
always been considered to be extremely problematic 
(e.g., Verdcourt, 1958; Bridson, 1987, 1992; Schatz, 
2001). A series of recent phylogenetic studies based 
on both molecular and morphological data conducted 
by two of the S (Lantz et al., 2002; Lantz € 
Bremer, 2004, 2005 to the b inca of 
largely new generic Hia dd of Vanguerieae. 
su Verdcourt and Bridson 


genera shown to be highly 
polyphyletic. Canthium subgen. Afr j i 


was resolved as monophyletic and was more closely 
related to Keetia E. Phillips than it was to the other 
three subgenera of Canthium described by Bridson 
(1987, 1992): Canthium subgen. Bullockia Bridson, 
Canthium subgen. Canthium, and Canthium subgen. 
d (Roem. & S 
Lantz and Bremer (2004) ra 


Pein to the generic level and restricted 


dson. As a result, 
sed pus subgen. 


Canthium to e» only species with paired supra- 

illary sp addition, Lantz and Bremer's 
(2004) s pin for the first time a strongly 
supported and mostly dioecious clade (hereafter the 
dioecious group), which contained all sequenced 
species of Canthium subgen. Bullockia, Dinocanthium 
Bremek. (Bremekamp, 1933), Leroya Cavaco (Cavaco, 
1970), Neoleroya Cavaco (Cavaco, 191a), Peponi- 


(Balfour, 1877; Verdcourt, 1983). The genus Cyclo- 


o included as part of the dioecious Aii 
n Lantz and Bremer’s (2004) figures 1 and 5. In total, 
the dioecious group sensu Lantz and Bremer (2004) 
comprises at least 22% (ca. 140 species) of the total 
species of Vanguerieae and has its center of species 
diversity in Madagascar with ca. 71% (at least 100 
species [Govaerts et al., 2006; Davis et al., 2007; 
Lantz et al., 2007; Razaf mand mbison et al. . 2007]. 
Mainland Afric has 
[Bridson, 1987). and 


Asian pum group B (sen 
currently unknown. The generic limits within the 
dioecious group were not addressed in Lantz and 


Bremer 
phylogenetic support and 


mainly due to lack of internal 
limited sampling of Pyro- 
stria and its allied genera. The establishment 2i he 
new narrowly circumscribed Canthium by Lantz and 
Bremer ge as left si a subgen. Bullockia 
and th groups Canthium species (one 
consisting pr the Las Ocean and Comorean 
Canthium Pax corresponding to the two-locular 
group V [sensu Bridson, 1987] and the 

other tend e = the Southeast Asian C. confertum 
Korth. group or group IV [sensu Bridson, 1987] 


species of 


unplaced within Vanguerieae. 

The circumscriptions of Pyrostria have always been 
controversial and unsettled (see Table 1), and there 
are at least two reasons for this situation. First, the 
circumscriptions of some of the closely related genera 
e.g., Di 933]; opepo- 
nidium [Arénes, able 2) largely Mie dn 


E 


inocanthium a 
1960; T 
with the early narrow circumscriptions of Pyrostria 
(Jussieu, 1789; Cavaco, 1967) based on the combina- 
tion of three characters, i.e., presence or absence of 
persistent, basally fused, and long acuminate paired 
bracts completely enclosing the young inflorescence, 
type of sexual systems, and number of locules per 
ovary (see Table 1). Second, previous authors pro- 
osed conflicting morphological concepts of Pyrostria 
as a result of the application of different weightings to 
the above characters (Tables 1, 2} used in combina- 
tion gee 789; Cavaco, 1967) or 
(Capuron, 1969; Bridson, 1987). 


iere. poo. paired of Pyrostria has 


as “cardinals” 
The monophyly o 


never been tested before using molecular phylogenies. 
In addition, Capuron (1969) 
identity of Pyrostria and viewed the Malagasy 


even questioned the 


Canthium, almost all erroneously described as 
hermaphrodite by Cavaco (1972), as dioecious. He 
intended to group all other dioecious Malagasy 


Vanguerieae genera (Peponidium, Pseudopeponidium, 


Volume 96, Number 1 


Razafimandimbison et al. 


2009 Functional Dioecy in the Tribe Vanguerieae 
Table 1. Previous and new circumscriptions of Pyrostria. 
Characters used for Genera not included 
Authors circumscribing Pyrosiria Variable characters Genera included in Pyrostria in Pyrostria 


Jussieu (1789) ^ plurilocular ova — 
paired bracts, dioecy 
Cavaco (1967)! — two-locular ovary 
paired bracts, dioecy 
Capuron (1969)! 


functional dioecy 


without paired 


Cavaco (1972)! — functional dioecy 


with or without paired 


bracts 


two to ndis ovary, 
wi 


rai 
two to mieni ovary, 


Peponidium and 
Pseudopeponidium 
Peponidium and — 
seudopeponidium 
Peponidium and Leroya and Neoleroya 
Pseudopeponidium 


Bridson (1987) presence of persistent 
and basall 
connate paired 
bracts 


Schatz (2001) 


functional dioecy 


two to plurilocular ovary, Dinocanthium, Leroya, 
sexual systems eoleroya, a 
(functionally dioecious seudopeponidium 
ermaphrodite 
two to plurilocular ovary,  Leroya, Neoleroya, 


with or without paired 


Cyclophyllum, 


eponidium, and 


Seyphochlamys 


Razafimandim- presence of persistent 


ison et al. and basally connate 


(this study) paired bracts 


racts 
two to plurilocular ovary 
breeding systems 
(functionally dioecious 
or hermaphrodite) 


Cyclophyllum and 
eoleroya, Pseudope- Peponidium 
ponidium, and 
Seyphochlamys 


! Taxa from Madagascar and the Comoro Islands are considered herein. 


and Pyrostria) under Canthium based on accumbent 
(as “accombants” [Capuron, 1969]) cotyledons d 
ledons lying against the radicle along one e Har 

& Harris, 1994]. Leroy (1972: 1683), n. 
influenced by (1969), the 
“Canthium—Pyrostria sensu lato” for the Malagasy 


Capuron used term 


Table 2. 


species (including both Leroya and — with 
the typical 
339) als 


dioecious but treated them as Canthium and main- 


aired bracts of Pyrostria. Schatz 
so recognized the Malagasy Ca: s as 
tained the generic status of Pyrostria (including 
Leroya, Neoleroya, Peponidium, and Pseudopeponi- 


Comparison between all described genera of the dioecious group sensu Lantz and Bremer (2004), Cyclophyllum, 


the Comorean and Indian Ocean Canthium, and the Southeast Asian Canthium confertum group 


No. of locules per 
y! 


Type of inflorescence 
bracts! 


Taxa Type of sexual systems! 
Canthium (Comoro Islands and Indian 2 cupular bracts functional dioecy 
cean 

Canthium confertum group 2 to 6 without bracts hermaphroditism 
Canthium subgen. Bullockia 2 without bracts functional dioecy 

Cyclophyllum 2 variable hermaphroditism and 

nctional dioecy 
Dinocanthium? 4to6 paired bracts functional dioecy 
roya? 4to6 paired bracts functional dioecy 
Neoleroya?? 2 paired bracts functional dioecy 
Peponidium?* 2 to 10 cupular bracts functional dioecy 
Pseudopeponidium^?^ 2 to 10 paired bracts functional dioecy 
Pyrostria sensu Cavaco (1967)* 2 paired bracts functional dioecy 
Pyrostria sensu Jussieu (1789)* 8 aired bracts functional dioecy 
Scyphochlamys 4to6 spathe-like bracts functional dioecy 


1 Characters traditionally and currently used in 
? Genera merged 


combination or as cardinals for delimiting Pyrostria. 


^ Genera considered by Capuron (1969) as oe with Canthium based on the presence of accumbent cotyledons. 


164 


Annals of the 
Missouri Botanical Garden 


Table 3. List of the known dioecious Rubiaceae. 


Subfamilies (sensu 


Bremer et al., 1999) Tribes Taxa References 
Cinchonoideae Guettardeae Antirhea borbonica J. F. Gmel.'? Litrico et al. (2005 
Bobea Gaudic Achille et al. (2006) 
Guetiarda L? Achille et al. (2006) 
Tinadendron Achille Achille (2006) 
Ixoroideae s.l. Bertiereae Bertiera borbonica A. Rich.” ailler et al. (1998) 
offeeae ricalysia A. Rich? Robbrecht (1979) 
Condamineeae Dioicodendron Steyerm Delprete (1999) 
icholobium A. Gray? Skottsberg (1944) 
Gardenieae Agouticarpa C. H. Perss Persson (20 
Alibertia A. Rich. group Persson (2000) 
Atraciocarpus Schltr. & K. Krause? Puttock (1999) 
Casasia A. Ric Robbrecht (1988) 
Gardenia arenas Puttock? Osunkoya (2003) 
Melanopsidium Colla Delprete (2000) 
Randia L Gustafsson & Persson (2002) 
Trukia Kanehira mith & Darwin (1988) 
Ixoreae Doricera Verde. Verdcourt (1 a 
Ixora pudica Baker!” Friedmann 
Mussaendeae Mussaenda parviflora Miq.'? ck (1883); o (1958); 
aiki & Kato (1999 
Octotropideae Kraussia Harv. E (1944) 
Vanguerieae Bullockia (Bridson) Razafim., Lantz & B. Bridson (1987); this study 
emer’ (this stu 
Canthium laeve Teijsm. & Binn.? Burck (1884) 
Cyclophyllum Hook. f£." Mouly et al. (2007 
Peponidium (Baill.) Arènes s.l! (this study) Arènes (1960); Bridson (1987); 
this study 
Pyrostria Comm. ex Juss. s.l! (this study) Bridson (1987); Verdcourt 
3) 
Rubioideae Anthospermeae Coprosma billardieri Hook. £? Skottsberg (1922) 
Normandia Hi Guillaumin (1930 
Coussareeae Coussarea latifolia Standl? Burger & Taylor (1993) 
Coussarea talamancana Standl.? Beach & Bawa (1980) 
Gaertnereae Gaerinera Lam? van Beus 
Morindeae s. str. Morin 


Psychotrieae 
Psychotria officinalis (Aubl.) Raeusch. ex 
Sandwith? 
Gouldia A. Gray? 
Hedyotis L? 
Spermacoceae Kadua Cham. € Schltdl? 


Nesohedyotis arborea (Roxb.) Bremek. 


Chassalia corallioides (Cordem.) Verdc.'? 


Hamilton (1990) 


Burck (1884) 

Baker & Cox (1984) 
Wagner & Lorence (1998) 
Percy & Cronk (1997) 


! Taxa reported to be functionally dioecious. 
? Genera known to have 


dium) (see Table 1). No species of the Indian Ocean 

or Comorean Canthium were included in Lantz and 
Bremer (2004, 2005). 

Breeding systems have been shown to A evolved 

independently numerous times throughout angio- 

.g., Renner € Ricklefs, 1995; Sakai et al., 

1995; Weiblen et al., 2000). As in 


plant families, hermaphroditism is the most common 


most flowering 


some or many hermaphroditic species. 


breeding system in Rubiaceae. Monoecy, separate 
Heilbuth, 
2000), has been recorded at least in the monogeneric 
tribe Theligoneae (Robbrecht, 1988), presently clas- 
sified in Rubioideae (Bremer & Manen, 2000). 


Dioecy, male or female flowers on separate individ- 


male and female flowers on each plant ( 


uals of the same species, has been reported in many 


genera in all three subfamilies (sensu Bremer et al., 


Volume 96, Number 1 


Razafimandimbison et al. 
Functional Dioecy in the Tribe Vanguerieae 


1999) of Rubiaceae (see Table 3). In the functionally 
dioecious Vanguerieae species, unisexual flowers are 
morphologically hermaphrodite but functionally dioe- 
cious; the functionally female flowers produce func- 
tional ovaries and stigmatic lobes but sterile anthers, 
whereas the functionally male flowers bear reduced or 
nonfunctional ovaries and stigmatic lobes but fertile 
anthers. The two forms are usually easy to distinguish 
because the functionally female plants have flowers 
with well-developed ovaries and one- to few-flowered 
inflorescences, rather than small ovaries and many- 
flowered inflorescences in the functional male plants 
(Bridson, 1987). In Vanguerieae, functional dioecy 
(also known as cryptic dioecy [Mayer & Charlesworth, 
1991] is commonly found in Pyrostria and its allied 
genera (Bridson, 1987) and was once E from 
the d Canthium laeve Teijsm. & Binn. (Burck, 

pela 


Bb gu] placement of C. laeve is 
kno and theref 


ae unknown. ore we do not know 
iiim the species represents a separate origin of 
functional dioecy in the tribe or belongs to the 
dioecious group. Two African and one Afro-Comor- 

—India cean Pyrostria species have been 
reported to be hermaphroditic (Bridson, 1987). In 
addition, gynodioecy, a regular coexistence of her- 
"t and female flowers on the same individual 
(Mayer & Charlesworth, 1991), was reported from 
Psydrax odorata (G. Forst.) A. C. Sm. & S. P. Darwin 
(as Canthium odoratum (G. Forst.} Seem. [Skottsberg, 
1945]) and from Afrocanthium gilfillanii (N. E. Br.) 
Lantz (Bridson, 1998: 324) and A. mundianum (Cham. 
& Schltdl.) Lantz (Balkwill et al., 1996). 

The ITS of the nuclear ribosomal DNA (nrDNA) 
have been shown to be useful for assessing oes 
netic relationships at the generic level 
rubiaceous groups (e.g., Gardenieae Persson, 2000]; 
Naucleeae s.l. Em & Bremer, 2001, 
2002]; Vanguerieae [Lantz et al., |]; Mussaendeae 
[Alejandro et al., 2005]. The is of the nrDNA, in 
combination with ITS data, have also recently been 
used in Rubiaceae for assessing phylogenetic rela- 
tionships of young lineages (e.g., n. et al., 
2003; D kon et al., 2005). The main 
objective of this study is to reconstruet a robust 
phylogeny of the dioecious group using b Meu 
data from two nuclear markers, ETS and 


ITS regions. 
The Mar nds pr will t (1 


hen be used to 

pinpoint the phylogenetic positions of the Comorean 
and the Southeast Asian 
up in Vanguerieae; (2) evaluate the 


and Indian Ocean Canthium 
C. confertum gro 
phylogenetic utility of three taxonomic characters 
(braet type, locule number, and fruit shape) previously 
d for deli 


dioecious group (Table 1); and (3) assess the evolution 


and currently use miting genera in the 


of functional dioecy in Vanguerieae. 


MATERIAL AND METHODS 
TAXON SAMPLING 


We tried to sequence as many representatives as 
possible of all formally described genera in the 
dioecious cay p (see ro 1) and the Comorean 
and Malagasy Canthium. We were unable to obtain 

any a aie of Pyrostria group B 
(Bridson, 1987). As there are currently no available 
identification keys for the Malagasy Canthium and 
Pyrostria, all determinations of the studied collections 
were made by comparisons to the type specimens 
79 taxa of 
from the 
dioecious group. Twenty-three of the 67 taxa were 


received on loan. In total, we sequenced 
hich 67 tax 


Vanguerieae, of whic a were 
new undescribed species of the Malagasy Canthium, 
Leroya, Peponidium, and Pyrostria (see Appendix 1). 
All formally published generic names in the dioe- 
cious group are used in the trees (Figs. 1-3), although 
we are aware that no single author accepts all o 
hem 


DNA EXTRACTION, AMPLIFICATION, AND SEQUENCING 


The total DNA, extracted from leaves dried in silica 
gel (ie & Hills, 1991) an 
was isolated falling the mini-prep procedure of 

aghai-Maroof et al. (1984), as modified by Doyle and 
Doyle 


d/or herbarium material, 


7). We amplified and sequenced parts of 
the ETS region of all investigated taxa with two 
primers, 18S-E (5'-GCAGGATCAACCAGGTGAA-3^), 
designed by Baldwin and iare es ps qe. at 
the 5’ border of 18S and ETS, ETS-HL (5'- 

EN uM. CGTC- 3 . pues by H. 
Lantz and located at the 3' border of ETS, following the 
protocols described in Razafimandimbison et al. 
(2005). The entire ITS region (including the 5.85 gene) 
of all newly studied taxa was amplified and sequenced 
using the protocols outlined in Razafimandimbison et 
al. (2004). In addition, we amplified the ETS and ITS 
regions of the DNA templates of both Peponidium sp. 
indet. 6 and Pseudopeponidium CAE Arénes using 
the 97°C/ribo: cribed in Raza- 


fimandimbison et al. (2004). Direct sequencing of their 


somal RNA primers d 


purified PCR products consistently produced multiple 
sequence signals for both markers, indicating the 
presence of intraindividual polymorphism. As a result, 
the PCR products of both taxa were cloned according to 
the TOPO TA cloning kit AE Paisley, Scot- 
land) (see Razafimandimbison et al., our white 
colonies from d ITS mun reactions 
were screened and amplified with two universal 
primers, T7 (5'-AAT ACG CTC ACT ATA G-3') and 
MI3R (5'-CAG GAA ACA GCT ATG AC-3’), which 
were included in the TOPO TA cloning kit. Their 


166 


Annals of the 
Missouri Botanical Garden 


1.00) 
pp em 


a 


Figure 1. 


2-million MCMC generation analysis, 


L 


0 
Pyrostria 


5 Ae tota 


Alberta magna’ 

Ixora coccinea* 
Psydrax obovata* 
Er arvilora* 
Psydrax kraussioides” 
canta confertum* — 
m aff. confertum* 

Do ud lum barbatum 


| outgroup taxa 


- |Canthium confertum group 


conn fum balansae J Cyclophyllum clade 


m sp. indet. 6-clone 1 
[- Poponidum sp. indet. 6-clone 2 
Popondum sp. idt es eu 
m Sp. In 


4 
Peponidium madagascariense 
e hs indet., aff. madagascariense 
Peponidum horridum 
Ea AA uum 


Peponidium buxifolium acc. 2 
Peponidium buxifolium acc. 1 
Peponidi n ifolium acc. 3 
Canthium ind et. 8 

Peponidun ^ indet., aff. quu lium 
Canthium buxifolium acc. 


Pei m sp. in et. 1 
Canthium majors 
Canthium cystiporoi 


ntum 
Canton psoe 
Canthium sp. indet. 5 
Canthium sp. indet. $ 
Pyrostria sp. indet. 1 
Scyphochlamys revoluta 
rostria viburnoides 
yrosiria o oo laris 


Bullockia clade 


m 


à pen 
Prosna Sp. inde st 2 
Pyrostia sarodranensis 
Pyrosta sp. indet. 3 
2 ndet. 


Leroya sp. indet., aff. richardiae 
Pesucopepaim longiflorum 


Pyrostria clade 


Dinocanthium 
lopeponidium ampijorense 
hondo 


Pseudopeponidium ol leifolium-clone 1 
posean al oleifolium-clone 2 


Afrocanthium s siebenlistii* 
rocan, zm pseudoverticillatum" 
ant 


Rytigynia senegalensis* 


| Peponidium clade 1 


Peponidium clade 2 


e 


A 


(poooz) JowWalg pue ziue1 


Fifty percent Bayesian majority rule consensus ITS tree under the GTR + I + G model of substitution from a 
showing mean branch lengths. Numbers on internodes indicate PP. 


s. A vertical bar 


delimits outgroup taxa. Brackets correspond to the major clades and the dioecious group. *Denotes hermaphroditic taxa. 


Volume 96, Number 1 Razafimandimbison et al. 
2009 Functional Dioecy in the Tribe Vanguerieae 


eetia lukei 


| outgroup taxa 


K 
1.00 Cyclophyllum barbatum Ed 
Cyclophyllum balansae cy clophyllum clade 


1.00 


» 
L 
r- 

— l i 

1.00 Mascarene Pyrostria subclade vlr p rostria a ubun noides 


— t2 
033 Malagasy Pyrostria subclade po Canthium pseudosetiftorum 
0.74 tor Canthium fadenii 


E 
AL d 
2 
=. 
= E 
S55 5.2 
"EE *$ 3 gam 
TH 
E 
3 
epej enpoojng 


OV. 
1.00 C Peponidium sp. indet. 6-clone 1 
Peponidium sp. ndet dida 
~~ Peponidiu Te indet. 
Canthium sp. indet. P 
Canthium nares 
1.00 p- Canthium sp. indet. 2 
lr Peponidium sp. indet. 2 
oag 1.00 Peponid m sp. indet. 1 
ium sp. i 


nsues dnoJb snoisoiq 


e 
dr 
pet] 
Zh 
2. 
E 
GH 
= 
zi 
epej| unipiuodag 


1 
a 


932 h.or Peponidium madagasc 

Peponidium sp. indet., at Dota 

0.87| r- Peponidium horridum 
1.00—- Peponidium velutinum zu 
'seudopeponi ORTA ampljor 

Psa eudopeponidium o xia ne 1 
Pseudopeponidium oleifolium-clone 2 
Dinocanthium pue strix 

et. 5 


y 
E 
& 
E 
pa 
(y00z) Jawalg pue zjue7 nsues dnoJB snoipaolg 


(Apnjs siuj) “|e 19 


0.97 rostria sp. indet. 


rostria sarod anensis 


Pi 

p 
Afro-Malagasy P i : 

ie 1.00 Pyrostri 
Pyrostria subclade > D, osig roaa 
Pseudopeponidium longiflorum 

Leroya sp. indet., aff. richardiae 
L sp. indet. 


Pyrostria media 

Pyrostria oes 
Pyrostria s| 

Pyrostria gum 
Pyrostria ae aen 
Pyrostria m 
PisLdopapondm asosa 
Pyrostria sp. indet. 4 


Figure 2. Fifty percent Bayesian majority rule consensus ETS tree under the GTR + G model of substitution from a 
2-million MCMC generation analysis, showing mean branch lengths. Numbers on internodes indicate PPs. Taxa highlighted in 
shaded boxes are subclades belonging to the Pyrostria group. The vertical bar delimits outgroup taxa. Brackets correspond to 
the major clades 


respective purified PCR products were sequenced with contamination. All sequencing reactions were per- 
the 185-E/ETS-HL and P17/26.82R (Popp € Oxelman, formed using the Big Dye Terminator v3.1 Cycle 
2001). In all PCRs, one reaction was run with water Sequencing kit and Big Dye Terminator v1.1 Cycle 
instead of DNA as a negative control to check for Sequencing kit (Applied Biosystems, Stockholm, 


168 Annals of the 
Missouri Botanical Garden 


H-Keetia ja lukei-AFR — B outgroup taxa 
100 [— FD4 pene barbatum-PAC CE dine clade | 
1.00 L— FD-Cyclophyllum were -PAC = Cyclo, 
FD-Canthium pseudosetiflorum-AFR 
94 98 FD-( anthiu m faden enii- ; 
1.00 76 [1.0L— FD-Canthium mombazense-AFR Bullockia clade 
1.00| 97 FD-Canthium sp. indet. 5 = Bullockia 
1.00 F thium sp. indet. 
— as — FD-Peponi nidium m sp. NS 6-clone 1 
1.00 FD-Pe eponidium sp. | -clone 2 
-Pe onidium s n E 
[100 | FD- Pep nidium jum sp. indet. 5 
1.00 FD-Canthiu indet d 
97 FD-Canl Sp 
1.00 FD-Cant iium oen CONAD 
FD-Canthiu /COM 
FD-Can am y AIZE 
FD-Canthium sp. indet. 1 
] [97 | FD-Ca um sp. in et 
deciduous, cupular 1.90 FD-Peponidium sp. indet a buxifolium OC 
)-Peponidium buxifolium acc. 2 


n| bracts never enclosing 


rep 
: = Peponidium s.l. 
ogs] the young inflorescence p 


)-Peponidium buxifolium à 1 

)-Peponidium buxifolium acc. 3 

)-Peponidium buxi tolum acc. 4 
P et. 4 


eponidium Sp. 


)-Peponidium rosca 

nthium sp. indet. 

-Canthium sp. indet. 7 
.2 


eponidium op. in 


mam 
»] O 
D 


FD-Peponidium s Sp. | indet. 1 


)-Peponidium horridum 


-Peponidium sp. indet., madegascanensa] 
)-Peponidium m madagasc arien 

Scyphochlamys revoluta- ROD a 

tria viburnoides-M 

)-Pyrostria orbicularis-REU 

i ~ 

! 

j 

$ 


[27 
2 
-— 

cs 


ostria commersonii-REU 


js] 


-Pseudopeponidi um Endo 
-Leroya sp. in Dosis aff. richardiae 
-Leroya sp. 

Paene re um ala 
| E andilanen 


(Apnys siuj) le je uosiquipueujeze» nsues dnoJD snol9eo!g 


80 
6 0.98 EN 
persistent paired 3g Fin 
8 


bracts enclosing 


idis eudo| d det. 1 
; seudopeponidium a indel 
inflorescence Poetas sarodranensis 
Pyrostria sp. indet. 6 Route E 
Pseudopeponidium ixorifolium - Pyros 
-Pseudopeponidium s indet. 2 


Pseudopeponidium oleifolium-clone 1 
Pseudopeponidium oleifolium-clone 2 
Dinocanthium Won trix-A 

Pyrostria sp. in 

)-Pyrostria ohyllanthoidea-AFR 
-Pyrostria serpentina 


Pse udopeponidium asosa 
- Pyrostria sp. indet. 4 


1 
J 
22 
L 


)-Pyrostria ankazobee 
D-Pyrostria ieee AFRICOMIOC 
)-Pyrostria anjouanensis-COM a 


ntmm 
T 


e 3. Parsimony strict consensus tree of 1555 MPTs of the combined ITS/ETS data (L = 822; CI = 0.421; RI = 
i sd the new classification of the dioecious group implemented. Numbers above internodes are JK support (> 50%), 
: : : t 


COM = Comoro Islands; IOC = Indian Ocean; MAD = Madagascar; MAU = Mauritius; REU = Réunion Island; and ROD = 
Rodrigues Island; the remaining ingroup taxa are endemic to Madagascar. 


Volume 96, Number 1 


Razafimandimbison et al. 
Functional Dioecy in the Tribe Vanguerieae 


Sweden) and subsequently analyzed with the 3100 
Genetic Analyzer (Applied Biosystems). 


DATA ANALYSES 


The ETS and ITS sequences were assembled using 
the Staden Package version 1.6.0 beta-test (Staden, 
1996) and aligned using ClustalX (Thompson et al., 
1997) to produce an initial alignment and manually 
1996). All newly 

submitted to the 
European Molecular Biology aded (E 
f the ITS P 
with the ped o MrBayes 3.0b (Huelsen- 
beck & Ronquis 01) The best-fit model of 
nucleotide o (GTR + I + G [Yang, Pa 
was se ing the computer program MrModeltes 
2.0 (Nylander 2004) and Akaike’s nn 
criterion (Akaike, 1974). The Bayesian ITS analysis 
a indels) was ous a e indepen- 
dent Markov for 2 X 


aligned using Se-Al (Rambaut, 
published sequences have been 


initially performed a Bayesian analysis o 


chains run ^ Metropolis- 

mee Markov chain Monte HC genera- 

tions, with tree sampling every 1 X 10? generations 

in after 1 X 10? trees (as detected b 

plotting the log likelihood scores against generation 

number). The analyses were repeated four times using 

different random starting trees to evaluate the 

of the likelihood value 
probabilis (PPs). AI 


independent runs were pooled for a consensus tree. 


es and posterior 
l saved trees from the four 
Groups characterized by PPs more than 95% were 
regarded as strongly supported. 

on the results of the Bayesian ITS analysis, 
we selected two closer outgroup taxa (Psydrax obovata 
(Klotzsch ex Eckl. & Zeyh.) Bridson and Keetia lukei 
Bridson) from within Vanguerieae to root the analyses 
of the i 
targeted taxa of the 


S and combined data from all 

ioecious group and both the 

Comorean and Malagasy Canthium. We subsequently 

carried out Bayesian analyses of the ETS and the 

combined ETS/ITS data using the same 

as above but with the 

substitution, selected as best-fit model for the two 

ETS and ITS data sets. We performed nel 

analyses of the four data sets (excluding uninformativ 
PAUP* b (Sw 


settings 
model of nucleotide 


characters) with version 4. "wo: sd 
2000), using heuristic searches, with the MULTREES 
option on, tree bisection-reconnection (TBR) branch 
swapping, swap on best tree only in effect, and 5000 
random addition sequences. In all analyses, charac- 
ters were given equal weight, gaps were treated as 
missing data, and phylogenetically informative indels 


index (CI; Kluge & Farris, 1969) and retention index 


fertum (PP 


(RI; Farris, 1989) were calculated to estimate 
homoplasy. Jackknife (JK; Farris et al., 1995) values 
were computed using heuristic a wit L- 
T on, TBR b 


additions, and 5000 replicates to assess relative 


ranch swapping, five random 


support of the retained clades. 

We statistically evaluated the congruence of the 
ETS and ITS data sets ws the x C B 
difference test (ILD 
implemented i Pope p Um a 2000; The " 
ristic search was set to 500 replicates with 10 random 
addition sequence and nearest-neighbor interchange 
(NNI) branch swapping. If the probability of obtaining 
a smaller sum of tree lengths from the random 
05) than that of 
the original data sets, the null hypothesis that the two 


generated data sets is lower (P 


data sets are homogenous is rejected and they are 
interpreted as incongruent (Farris et al., 1995 


RESULTS 


ITS ANALYSES 


A total of 84 ITS sequences were analyzed and 58 
ca. 69%) are newly published here. The ITS matrix 
contained 776 positions and 189 (24.35%) were 


parsimony informative. 


== 


Both the parsimony strict 
790, CI = 0.439, RI = 
result not shown) and Bayesian ITS trees had 


consensus (length [L] — 


similar overall tree topologies, which were not in 
conflict with those of Lantz and Bremer (2004). In the 
Bayesian majority rule consensus tree pooled from the 
Bayesian trees (burn-ins excluded) from the four 
independent runs shown in Figure 1, all sequenced 
taxa of the dioecious group sensu Lantz and Bremer 
(2004) formed a moderately supported monophyletic 
group (PP — 0.86), which was resolved as sister (PP — 
1.00) to a clade containing the two studied species of 
Cyclophyllum a and two accessions of Canthium con- 
= 1.00). The 


Psydrax species were unresolved, and the hermaph- 


three hermaphroditic 
roditic 12 remaining Vanguerieae taxa formed a 
moderately supported clade (PP = 0.93). The mem- 
bers of the dioecious group sensu Lantz and Bremer 


— 


2004) were resolved in four clades: the Peponidium 
clade 1 (consisting of the three clonal accessions of 
Peponidium sp. indet. 6; PP = 1.00); the Peponidium 
clade 2 (including the remaining sampled Peponidium 
species and all sequenced Comorean and Malagasy 
0.63); the Bullockia clade 


(including three species of Canthium subgen. Bul- 


Canthium species; PP = 


lockia and two undescribed new Malagasy species; PP 
= 1.00); and the 
Pyrostria, Scyphochlamys, Leroya, Neoleroya, Dino- 
PP = 


stria clade (including all 
l 


canthium, and Peeudopeponidium species; 


170 


Annals of the 
Missouri Botanical Garden 


0.79). The relationships between the four dioecious 
clades were largely unresolved. 


ETS ANALYSES 


A total of 68 ETS sequences were analyzed together 
and all are n ice: here. The ns ETS 
30.8%), 
p POMA infor- 
mative. In the Bayesian ETS tree shown in Figure 2, 
the Cyclophyllum clade was resolved with high 
support (PP = 


positions, and 1 
co a NI were 


1.00) as sister to the moderately 
supported (PP — 0.93) dioecious group sensu Lantz 
and Bremer (2004). Within the dioecious group, all 
deep nodes received only poor (PP = 0.60) to 
moderate (PP = 0.86) support (Fig. 2). The Pyrostria 
clade, weakly n the ITS tree (Fig. 1), 
collapsed in three highly supported subclades (PP 

0.97—1.00): a Mascarene Pyrostria subclade; a sub- 
clade containing two Malagasy Pyrostria species, P. 
pendula Lantz, Klack. & Razafim. and Pyrostria sp. 
indet. 


supported i 


2; and an Afro-Malagasy Pyrostria subclade 
(including all studied species of Dinocanthium, 
Leroya, Neoleroya, and  Pseudopeponidium). Th 
Peponidium clades 1 and 2 in the ITS tree (Fig. 1) 
ormed a well-supported monophyletic group (hereaf- 
ter called the Peponidium clade; PP 

Pyrostria sp. indet. 1, the Peponidium clade, and 
the Afro-Malagasy Pyrostria subclade were unre- 


[2] 


formed a highly supported clade (PP = 1.00), which 
in turn was sister to a well-supported clade corre- 
sponding to the Peponidium clade in Figure 1. The 
Cyclophyllum clade received high support (PP = 
1.00, while the Bullockia clade had only weak 
support (PP = 0.74 
A parsimony analysis of the ETS matrix resulted in 
52 ieri most parsimonious trees (MPTs; L — 409, 
CI = 0.430, RI = . In the strict consensus ETS 
tree bono ifii: the 52 MPTs (result not shown), 
the Pyrostria clade was resolved in two separate 
clades: the strongly supported (PP = 1.00) Mascar- 
ne ria clade and the poorly supported (PP < 
50) Afro-Malagasy Pyrostria clade (including Dino- 
a oe Neoleroya, Pseudopeponidium, Pyr- 
ostria sp. indet. 1, Pyrostria sp. indet. 2, and Pyrostria 
pendula). The duos Cyclophyllum, and Peponi- 
ium clades all received the same levels of support as 


in the Bayesian ETS tree (Fig. 2). 


COMBINED ANALYSES 


The results of the ILD test indicated that the two data 
sets were not significantly incongruent and therefore 


could be combined. Fusion of the ETS and ITS data sets 
of 68 taxa in one matrix yielded 1122 positions and 238 
parsimony-informative characters (including coded 
indels). Both Bayesian and parsimony analyses of the 
combined ETS/ITS data produced identical tree 
topologies. rsimony analysis of the combined 
ETS/ITS data resulted in 1555 equally MPTs (L = 822, 
CI — 0.421, RI — 0.735). In the strict consensus tree 
igure 3, the ingroup taxa were resolved in 

four major e (Fig. 3): A Cyclophyllum clade 
(= Cyclophyllum), the Bullockia clade (= Bullockia 
Bridson) Razafim., Lantz B. Bremer), and the 
Peponidium clade (= Peponidium s.1.), all with high 
support (JK = 94—100, PP = 1.00), and the Pyrostria 
clade (— pr sl; JK = 
oo supported M belad (JK = 100 
00) was ae as sister to the poorly supported (JK 

= 57) Afro-Malagasy Pyrostria subclade. The position 
of the a MT es as ims to the strongly 
supported 00) dioecious group 
sensu Lantz ut p Cis was further corroborat- 


shown in Figu 


— 


ed. However, the phylogenetic ae between 
the Bullockia, us a clades 
ained unresolved. and 
Pyrostria were shown to E paraphyletic or polyphylet- 
ic, while Leroya, dpud) by Leroya sp. indet., a 
richardiae, and Leroya 

hele 


stri 
oth Pseudopeponidium 


p. indet., formed a highly 
Finally both Leroya and Neoleroya 
were dece nested within the Pyrostria clade. 


DISCUSSION 
FURTHER DISINTEGRATION OF CANTHIUM S.L. 


The Bayesian ITS analysis (Fig. 1) reveals a further 
y ens of Canthium s.l., as the Southeast Asian 
rium (cu 

Bridson [1987] ere by two acces- 
sions of C. confertum) is for the first time shown to be 
closely related to Cyclophyllum. However, this does 
not necessarily apply to the whole gr 


rrently placed in group IV sensu 
and represented h 


tested before. In ETS and combined ETS/ITS trees 
(Figs. 2, = the sequenced Comorean and India 

Ocean Canthium species form a highly spe 
clade = sampled Peponidium (JK = 97, PP = 
alyses (Figs. E further 


addition, all our ana 
he phylogenetic position of Canthium 


pa ose the 
subgen. Bullockia in the dioecious group previously 
demonstrated by Lantz and Bremer (2004) 


EVOLUTIONARY TRENDS IN THE DIOECIOUS GROUP ACCORDING 
TO THIS STUDY 


Phylogenetie reconstructions can help us under- 
stand how morphological variation found within a 


Volume 96, Number 1 
2009 


Razafimandimbison et al. 
Functional Dioecy in the Tribe Vanguerieae 


monophyletie group has evolved and also highlight 
synapomorphies for group recognition. 


Bract type 


The combined ETS/ITS tree (Fig. 3) points to single 
origins of both the persistent, basally connate paired 
bracts of the Pyrostria clade (consistent with Lantz & 
Bremer, 2004) and the deci 
bracts of the Peponidium clade within the dioecious 
(see Fig. 3, Table 2). We consider these 


eciduous, small cupular 


group 
characters to be the potential E 
synapomorphies for the two clades 'yrostria, the 
connate paired bracts completely b. the young 


inflorescences and are located either at the ied or at 
the apex of the peduncles. Many New Caledonian 
species of Cyclophyllum are known to have small and 
connate paired bracts that never enclose the youn 
inflorescence, but the typical bracts of Pyrostria are 
reported from the New Caledonian C. calyculatum 
Guillaumin (Guillaumin, 1930). We suspect that this 
species belongs to Pyrostria. The Bullockia clade and 
Canthium confertum group have no bracts 


Locule number 


The number of locules per ovary, previously used 
2 dr (1789) and Cavaco (1967), in combination 
wo ters (e.g. type of breeding 
ae for delimiting Pyrostria in a narrow sense, 
is clearly shown to have evolved independently 


other charact 


are ce within the 
dioecious group (see also Table 2) and in the other 
parts of Vanguerieae; this is also E with Lantz 
and Bremer (2004). 


numerous times (i.e., 


Ovary/fruit shape 


Leroy (1972: table 1), as well as field observations 
and herbarium studies conducted by two of the 
authors (SGR and HL), demonstrate a great and 
continuous variation of both ovary and fruit shapes 
and sizes in the Pyrostria clade. Cavaco (1970, 1971a) 
appears to have overlooked or was unaware of the 
presence of many Pyrostria species with intermediate 
ovary and fruit shapes (Leroy, 1972) between those of 
Leroya and Neoleroya. The phylogenetic positions of 
both Leroya and Neoleroya deeply nested inside the 
Pyrostria clade with our field 
observations and herbarium studies. In addition, the 
in the 
Cyclophyllum further 
indicating that both fruit shape and size cannot be 


are congruent 


is also found 
clades, 


same degree of variation 


and  Peponidium 


used alone for recognizing genera in the dioecious 


group. The members of the Bullockia clade have two- 
locular and heart-shaped fruits, which are commonly 
found in the Peponidium and Pyrostria clades, and 


many other Vanguerieae genera 


EVOLUTION OF FUNCTIONAL DIOECY IN VANGUERIEAE 


The present study supports a single origin of 
functional dioecy in Vanguerieae (Figs. 1-3). The 
Asian Canthium laeve has also been reported by 
Burck (1884) to be functionally dioecious. For now, 
we cannot tell whether functional dioecy had single or 
multiple origins in the tribe, until the phylogenetic 
placement of this species within Vanguerieae is 
known. The Southeast Asian species are, in general, 


both the functionally male and female flowers 
indicates that the functional dioecy of Vanguerieae 
may have originated from hermaphroditism. This may 
also imply that the functionally dioecious flowers in 
Vanguerieae are probably in transition to morpholog- 
ically true dioecious ones (see also Mayer & Charles- 
worth, 1991). On the other hand, the combined 


within Pyrostria s.l. (Fig. 3) and the Cyclophyllum— 
Canthium confertum clade, according to Bridson 
(1987) (see also Table 2). Recent herbarium studies 
and field observations conducted by A. Mouly reveal 
that the two sequenced Cyclophyllum species and 
seven other New Caledonian species are actually 
functionally dioecious (Mouly et al., 2007), inconsis- 
tent with Bridson (1987), who considers the genus to 
be hermaphrodite. Similarly, Friedmann (1994) re- 
ported functionally male and female and hermaphro- 
ditic flowers from the Afro-Comorean—Indian Ocean 
which is 


Mayotte Island (France). This further suggests that 
functional dioecy in Vanguerieae is evolutionarily 
unstable and, accordingly, should not be used alone 
for diagnosing any genera within the dioecious group, 
inconsistent with Capuron (1969), Cavaco (1972), and 
Schatz (2001) (see Table 1). However, it can still be 
used for characterizing the genera in the newly 
group (this study) in 
combination with other characters. 

all described African, 


s Vanguerieae 


cirscumscribed dioecious 
Malagasy, and 


Fina g 
species, 


ys * 
Mascarene dio except 


the Mn Gomorenn Ilian Ocean polygamous Pyros- 


172 


Annals of the 
Missouri Botanical Garden 


tria bibracteata and the functionally dioecious 


Schatz, 2001), indicating an autochthonous evolution 
of functional dioecy in Vanguerieae throughout the 
n the other 
nd, we cannot yet rule out the origin of the 
ally dioecious Mascare 


m a functio 


geographic ranges of the dioecious group 
functiona carene Pyrostria species 
onally dioecious Malagasy or Southeast 
Asian colonist. 


NEW CIRCUMSCRIPTIONS, MAJOR LINEAGES, AND NEW GENERIC 
LIMITS OF THE DIOECIOUS GROUP 


New circumscriptions of the dioecious group 


The ITS tree (Fig. 1) resolves the Southeast Asian 
and hermaphroditic Canthium confertum group (sensu 
Bridson, 1987), represented here by two accessions of 
C. confertum, as sister to the mostly Pacific Cyclo- 

pue A much larger sampling of Cyclophyllum and 
the Can group is needed for testing 
their sister-group relationship. The 


hium confertum 
support for the 
sister-group relationships between the Cyclophyllum 
clade and the dioecious group sensu Lantz and 
Bremer (2004) is consistently high in the ETS and 
combined analyses (Figs. 2, 3) and is also consistent 
with Mouly et al. (2007). An we include 
Cyclophyllum in the dioecious group, w currently 
contains approximately over 200 species es one third 
of the total species of Vanguerieae). If the sister-group 
relationship between Cyclophyllum and the Canthium 
confertum group is further confirmed, the C. confertum 
group will have to be considered members of the 
dioecious group as well. The combined E 
(Fig. 3) indicate that the dioecious i n be 
subdivided int morpholo, aly distet 
lineages: the Crelosh li clade, Bullockia clade, 
Peponidium clade, and Pyrostria clade. 


Major lineages of the dioecious group 


Cyclophyllum clade. The ETS and combined 
ETS/ITS analyses (Figs. 2, 3) further corroborate with 
strong support (JK = 100, PP = 1.00) the sister-group 
-o between the us clade and the 
dioecious group sen and Bremer (2004). 
Ancor ne to Bridson Ero Cyclophyllum can be 
characterized by the combination of the slowing 

without paired bracts 


fasciculate inflorescences or occasionally with "d 


characters: inflorescences 


mentary inflorescence branches; large hypocrateri- 
form corollas; hermaphrodite flowers with two-locular 
ovaries and style widening at apex; and dorsal face of 


anthers (except margins) covered with dark connec- 


er, as we mentioned above, at least nine 


yi 

functionally dioecious (Mouly et al., 2007). In 
addition, at least one New Caledonian species is 
reported by Guillaumin (1930) to have connate paired 
bracts that enclose the young inflorescences. This 
indicates that the monophyly of Cyclophyllum as 
presently delimited needs to be tested. We find no 
support for Baillon’s (1879) suggestion of merging 
Cyclophyllum with Canthium, which is presently 
restricted to the African and Asian spiny species 
(see Lantz & Bremer, 2004). 


Bullockia clade. The Bullockia clade, retained in 
all analyses (Figs. 1-3), comprises three of the six 
species of Canthium subgen. Bullockia (Bridson, 
1987) and two undescribed new Malagasy species, 
consistent with Bridson’s (1987) Sis Bridson 
(1987) used the combination of several characters to 
distinguish Canthium subgen. Bullockia from the 
other subgenera of Canthium and Pyrostria and its 
allied genera: persistent leaves; inflorescences un- 
subtended by the typical Pyrostria paired bracts; 
umbellate or fasciculate inflorescences bearing func- 
tionally unisexual flowers; inside of the corolla tube 
with a well-defined ring of deflected hairs; corolla- 
throat pubescent but not congested with hairs, rarely 
fleshy corollas; dorsal face of the anthers with only the 
central area with dark connective or entirely without 
dark connective; and hollow stigmatic knob (at least 
near the base) with recessed style. Our close 
observations of the sequenced Malagasy species, 
however, show that both of the undescribed new 
Malagasy species (Canthium sp. indet. 5 and 6) have 
deciduous leaves 


Peponidium clade. In the ETS (Fig. 2) and 
combined trees (Fig. 3), all sequenced species of 
the Comorean 


Peponidium and and Malagasy 


Canthium form a strongly supported (JK = 95-97, 
PP = 1.00) clade. The Comorean P. comorense Arénes 
(Arènes, 1960) and the Seychellean C. carinatum 
(Baker) Summerh. (Friedmann, 1994) have recently 
been shown to belong to this clade (Avino et al., ETS, 
ITS, and trnT-F unpubl. data). This finding indicates 
that Peponidium sensu Arénes (1960) is not mono- 
phyletie, unless the Comorean and Indian Ocean 
Canthium are also included. The Peponidium clade 
(Fig. 3) corresponds to Bridson’s (1987) group V, a 
be artificial. In fact, the 
Peponidium clade is supported by at least one 


group she considered to 


potential morphological synapomorphy. The members 


located at the bases of the inflorescence peduncles 


Volume 96, Number 1 
2009 


Razafimandimbison et al. 
Functional Dioecy in the Tribe Vanguerieae 


and never enclose the young inflorescences (see also 
Table 2). Arénes (1960) referred to this type of bract 
as a false involucre. While the bracts are cup-shaped 
in most of the species, some species appear to have 
two to four basally fused and imbricate bracts. This 
indicates that the cupular bracts in the Peponidium 
clade are most likely to have derived from a complete 
fusion of more than two bracts. The cupular bracts are 
conspicuous and often remain intact in species with 
subsessile and pedunculate inflorescences bearing up 
to three flowers. In contrast, they are difficult to see 
and are often crushed in species with sessile 
inflorescences with four or more flowers, or with both 
sessile inflorescences and flowers. Finally, neither 
Lantz and Bremer (2004) nor the present study find 
support for Capuron’s (1969) attempt to merge 
Pyrostria s.l. (including Peponidium and Pseudopepo- 
nidium) in Canthium based on accumbent cotyledons. 


Pyrostria clade. One of the main goals of the 
present study is to test the monophyly of the different 
ae MA. of Pyrostria. The rd a 

eakly to moderately suppor 
of which indio is an 


sequenced functionally dioecious species of Dino- 


aaa Pyrostria 


net from the 
other genera in the dioecious group (sensu Razafi- 
this study) and the rest of 
Vanguerieae in having persistent and basally connate, 
. There 


the narrowly 


s easy to recogni 
mandimbison et al., 


long acuminate paired bracts (see also Table 2 
is no support for the monophyly of 
delimited Pyrostria sensu Jussieu (1789) or Cavaco 
(1967), as the two-locular or plurilocular species (see 
(Fig. - In 
ition, Pyrostria sensu Capuron (1969), 

(1972), and sensu ju (2001), all dela 
by (functional) dioecy (see Table 1), are shown to be 


Table 2) never form aos clades 
Cavaco 


paraphyletic because the polygamous Pyrostria spe- 
cies (P. bibracteata) is deeply nested in the almost 
functional dioecious Pyrostria clade, consistent with 
Bridson (1987), 


diagnosed by the presence of paired bracts, is para- 


Pyrostria sensu Bridson 


phyletic as it excludes Seyphochlamys. No members of 
the Asian Pyrostria group B (Bridson, 1987) are 
included in our analyses due to lack of material; on 
the other hand, the fact that they all bear the typical 
paired bracts of Pyrostria indicates that they belong to 
this genus. The present study further supports the 
anes of Bridson (1987) to merge the African genus 
inoca: in Pyrostria. Bridson (1987: 
oe a both Leroya and M 
to be little more than Pyrostria with marked 


Ex 
m 
7 


eoleroya “appeared 


elaboration of the fruit” and therefore were “margin- 
ally worth generic rank." Bridson's statement is 
further supported by our analyses, as Leroya and 
Neoleroya are both nested in the Pyrostria clade. In 
addition, 
Baill. (Baillon, ied ego here by Pyrostria 
media (A. Rich.) C and P. major (A. Rich.) 
Cavaco, in Pyrostria by mice (1967) is supported by 


the inclusion of Canthium sect. Psydracium 


our results. Furthermore, our results reveal no support 
for the two sections of Pyrostria, Pyrostria sect. 
Pyrostria and Pyrostria sect. Involucratae Cavaco, 
described um Cavaco (1971b). Ped andilanensis 

vaco, the two species Pyrostria sect. 
an (Cavaco, 1971b), is deeds nested in the 
members of Pyrostria sect. Pyrostria, consistent with 
morphological data. Finally, Bridson's (1987) consid- 
eration of grouping Pyrostria s. str., Leroya, ue a, 
Pyrostria group A, Pyrostria group B, Seyphochlamys, 
Cyclophyllum, her group IV (Bridson, e and 


supported as monophyletic unless Bullockia is also 
included. Pseudopeponidium (Bridson, 1987; Schatz, 
2001), Leroya, and Neoleroya have been synonymized 
under Pyrostria by Schatz (2001). 


New generic limits of the dioecious group 


The present analyses indicate that the generic limits 
in the dioecious group sensu Razafimandimbison et al. 
need to be recircumscribed. r intention 1s to 
recognize both monophyletic and morphologically dis- 
tinct genera and to minimize nomenclatural chan, 

At least three alternative solutions are possible. The 

first solution is to maintain Cyclophyllum at the generic 

level and recognize a broader circumscription of 

Pyrostria (the oldest name) including all species of 

the Bullockia, Peponidium, and Pyrostria clades 

without infragenerie subdivision. The second solution 
l 


is to merge the four clades with Pyrostria. Both alter- 
natives would make Pyrostria rather heterogenous 
morphologically and require a large number of new 
combinations. The third possibility is to recognize the 


Cyclophyllum, Bullockia, Peponidium, and 


clades all at the generic level. We favor this last option 


Pyrostria 


because it reflects the morphological distinctness of the 
four clades and would minimize nomenclatural changes 
in the last three clades. In addition, the present study 
clearly shows that the generic status of Dinocanthium, 
roya, Neoleroya, and Pseudopeponidium is unten- 
able. We disregard ide possibility of maintaining the 
current generic status of Seyphochlamys because this 
would force us not only to recognize Pyrostria in a 
narrow sense (i.e., including only the Mascarene spe- 
cies), but also to raise Dinocanthium to accommodate 


174 


Annals of the 
Missouri Botanical Garden 


all African and Malagasy Pyrostria species. There is no 
distinctive morphologie charaeter for separating the 
Mascarene Pyrostria from the Afro-Malagasy or Asian 


PRELIMINARY BIOGEOGRAPHIC HYPOTHESES OF THE 
DIOECIOUS CROUP 


We are unable to perform a proper biogeographic 
analysis, as the relationships between the major 
lineages of the dioecious group as delimited here 
n the other hand, 
still be 
Bullockia, Peponidium s.l., and Pyrostria s.l. A well- 


are largely unresolved. 
biogeographic facts can discussed for 
resolved phylogeny of the dioecious group based on a 
much larger sampling is needed to test all biogeo- 
graphic hypotheses put forward below. 

The newly delimited dioecious group is distributed 
throughout eastern. and southern Africa, Southeast 
Asia, the Indian Ocean islands, the Pacific region, 
and northern Australia bn also Bridson, 1987). 

Bullockia is an Afro-M de consistent with 
Bridson's (1987) lider. wide to have had 
an African origin and may well ps reached 
Madagascar via a single long dispersal event. The 
African species are restricted to eastern and southern 
Africa Noite 1987) and the two studied iip 
species are confined, respectively, to the dry dec 
duous fies of southwestern and pe 
Madagascar. In contrast, our circumscribed Peponi- 
dium is almost completely Malagasy, with only two to 
four species restricted to the Comoro Islands (Arénes, 
1960; Cavaco, 1972) and two species confined to the 
Seychelles (Friedmann, 1994). The Comorean and 
Seychellean species appear to have had Malagasy 
unpublished data). In 
addition, the newly delimited Pyrostria is predomi- 
nantly Malagasy, with 14 species confined to both 
eastern and southern Africa (Bridson, 1987), eight 
Mascarene-endemic species restricted to Rodri- 


origins (Fig. 3; Avino et al., 


gues Island, five restricted t uritius, and t 

endemic to Réunion Island [Verdcourt 1983], ea 
one, P. bibracteata, shared between the African 
mainland, the Comoro Islands, Madagascar, and the 
Seychelles. The number of species of Pyrostria in 


own. The results 


a, represented here by P. hystrix (Bremek.) 
phyllanthoidea (Baill.) Bridson, 
appear to have originated from more than one 
Malagasy ancestor, as the two WM African 
species do not form a clade (see Fig. bd ue 
the highly supported sister-group relationship 


100, PP = 1.00) between the strongly a 


99, PP = 1.00) and the 
Rodrigues Islands Seyphochlamys indicates that they 


Mascarene Pyrostria (JK = 


ad a common ancestor (Fig. 3). Geologic evidence 
suggests that Mauritius emerged as a result of a series 
of volcanic events, the earliest of which occurred ca. 
6.8-7.8 million years ago (Ma). Rodrigues Island did 
(McDougall & C 


1969), suggesting that the common ancestor of both 


not emerge until 1.5 Ma hamalaun, 
Seyphochlamys and the Mascarene Pyrostria should 
be at least 1.5 Ma. Based o 
here, 
Pyrostria and possibly Pyrostria s.l. had a relatively 


n the evidence presented 
it can be Edd that the Mascarene 
recent diversification. This may well 2 one age 
o for the low to moder rt for 

stria sl. compared to that of ibd. pos 
pes and Peponidium s.l. (Fig. 3). 


PHYLOGENETIC CONCLUSIONS 


The present study supports a further disintegration 
of Canthium s.l., as the Southeast Asian C. confertum 
and the Comorean and Malagasy Canthium species 
group are distantly related to Canthium s. str. (Lantz 
& Bremer, 2004) and are for the first time shown to 
belong to the new circumscribed dioecious ie The 
combined analyses show that the dioecious 
be subdivided into 
clades. According to the rules of priority, the following 


gro 
four morphologically sh 
generic names are applicable to the clades: Cyclo- 
phyllum, Bullockia 
clophyllum 


the dioecious group sensu Lantz and Bremer (2004). 


, Peponidium, and Pyrostria. Cy- 
is ése with high support as sister to 
The results further point to a single origin of 
functional dioecy from hermaphroditism for the newly 
circumscribed dioecious group followed by subse- 
quent reversals back to the hermaphroditic conditions 
at least in both Cyclophyllum and Pyrostria s.l. 
Furthermore, the Malagasy j 
to have a common African ancestor, which is most 
likely to have reached Madag 
di Du a and subse i Mà radiated there. In 


Bullockia species seem 
ascar via a 


contrast, the Comorean Pepon appears to 
have ao from a age common ancestor. 
Similarly, our results further support the Malagasy 
origins of the African Pyrostria and a single origin of 
the Mascarene Pyrostria. Finally, we conclude that 
Pyrostria s.l. is a relatively young group that seems to 


have had a recent diversification. 


TAXONOMIC IMPLICATIONS 


Below, we formally describe Bullockia (Bridson) 
Razafim., Lantz & B. Bremer to accommodate all six 
described African species of Canthium subgen. 
Bullockia and the two undescribed new Malagasy 


Volume 96, Number 1 


Razafimandimbison et al. 
Functional Dioecy in the Tribe Vanguerieae 


species (Canthium sp. indet. 5 and 6). We synonymize 
Seyphochlamys with Pyrostria. Dinocanthium, Leroya, 
Neoleroya, and Pseudopeponidium all have been 
synonymized under Pyrostria before (Bridson, 1987; 
Schatz, 2001). Razafimandimbison et al. (2007) have 


recently transferred 20 species of the Indian Ocean 


and Comorean Canthium to Peponidium and all 
described species of Leroya, Neoleroya, Pseudopepo- 
nidium, and Scyphochlamys to Pyrostria. The transfer 
of the two Seychellean Canthium to Peponidium will 
be published elsewhere. 


- 


Bullockia (Bridson) Razafim., & B. 
Bremer, stat. nov. Basionym: Canthium subgen. 
Bullockia Bridson, Kew Bull. 42: 630. 1987. 

ullockia setiflora (Hiern) Razafim., 


Lantz & B. Bremer 
For a description, see Bridson (1987: 630). The 


only new information we add is that species with 


Lantz 


persistent (all African members) and deciduous leaves 
the two new undescribed Malaga 
Figs. 1-3) are found in Bullockia. 


gasy species; see 


Number of species. There are at least eight species 
(six species in mainland Africa [Bridson, 1987] an 


two in Madagascar). 


Distribution and habitat. 
in eastern and southern Africa and Madagascar in 


The genus is distributed 


bushland, woodland, savannas, and dry, deciduous 
forests. 


Dia, agnostic gentes Bulge is distinct from the 
Vanguerieae by 
owing Fans the inside 
of the corolla tube with a well-defined ring of 
deflected hairs; corolla-throat pubescent but not 
congested with hairs, rarely fleshy corollas; dorsal 
face of the anthers with only the central area with a 
dark connective or entirely without a dark connective; 
and a hollow stigmatic knob (at least near the base) 
with a recessed style (Bridson, 1987). 


w combinatio Here, we transfer all described 
locki 


e ns. 
species of Canthium subgen. Bullockia. 


1. Bullockia dyseriton (Bullock) Razafim., Lantz & 


ym: Canthium 


: enya. Teita Distr.: 
Hills, 2500-3500 ft., s.d., H. M. Gardner 3000 
(holotype, K not seen, hota 


2. Bullockia fadenii (Bridson) Razafim., Lantz & B. 
Bremer, comb. nov. Basionym: Canthium fadenii 
Bridson, Kew Bull. 42: 632. 1987. TYPE: Kenya. 
Kiambu Distr.: “behind Blue Ports Hotel, 


Thika,” 4900 ft., 23 Mar. 1968, R. B. Faden 
68/012 (holotype, K not seen, photo!). 


3. Bullockia impressinervia (Bridson) Razafim., 
Lantz & B. 
Canthium impressinervium Bridson, Kew Bull. 

. TYPE: Tanzania. Lindi Distr., 
“Noto-Plateau,” ca. m, ar. , H. J 
Schieben 6109 (holotype, K not seen, photo!). 


Bremer, comb. nov. Basionym: 


4. Bullockia mombazensis oe Razafim., Lantz 
mer. : Canthium 

o Bai o dq 188. 1879. 
TYPE: Kenya. “Côte orientale d'Afrique: Zanzi- 
bar,” 1847-1852, L. H. Boivin s.n. (holotype, P!). 


com 


5. Bullockia p (Bridson) Razafim., 
& B. 


Lantz er, comb. nov. Basionym: 
Canthium pendium Bridson, Kew Bull. 
42: dcs 1987. : Kenya. “Among 

y watercourse,” 3400 ft., 1952— iow J. 


sb 
B ptu 14096 (holotype, K not seen, photo!). 


granite 


6. Bullockia setiflora (Hiern) Razafim., Lantz & B. 
Bremer, comb. nov. iv anthium seti- 
o aay Fl. Trop. Afr. [Oliver et al.] 3: 134 
1877. TYPE: Moxque *Between Tete and 
the sea s Mar.-Apr. 1860, J. Kirk s.n. 
(holotype, K not seen, photo!). 


IIl. Peponidium (Baill.) Arènes, Notul. Syst. (Paris) 
16: 25. 1960. Basionym: Canthium sect. Pepo- 
nidium Baill., Adansonia 12: 197. 1879. TYPE: 
Peponidium horridum (Baill.) Arénes. 


For a description, see Arénes (1960). 


Number of species. There are at least 45 species, 

all (except two to four Comorean species [Aréne, 

n and two Seychellean species [Friedmann, 

94]) endemie to Madagascar (Arénes, 1960; Raza- 
7 


"dan ds et al., 


Distribution and habitat. This genus occurs in 
Madagascar (e.g., Arénes, 1960; Cavaco, 1969b), the 
Comoro Islands (Arénes, 1960; Mouly, 2007), and the 
1994) in dry, deciduous 
forests, lowland rainforests, and mid- and high- 


Seychelles (Friedmann, 


altitude humid forests. 


Diagnostic features. Our newly circumscribed 
Peponidium is diagnosed by the presence of decidu- 
ous, cupular bracts at the bases of the inflorescence 


peduncles that never enclose the young inflorescence. 
III. Pyrostria Comm. ex. Juss., Gen. Pl. 206. 1789. 
TYPE: Pyrostria commersonii J. F. 


er Balf. £., J. Linn. Bot. 16: 14-15. 1877, syn. 
TYPE: Scyphochlamys revoluta Balf. f. 


176 


Annals of the 
Missouri Botanical Garden 


Canthium sect. Psydracium Baill., Adansonia 12: 199. 1879. 
TYPE: Pyrostria major (A. Rich.) Cavaco [= Psydrax 
A. Rich.]. 
Dinocanthium Bremek., Ann. Transvaal Mus. 15: 259. 1933. 
TYPE: Docta hystrix Bremek. 
e oe Homolle ex Arénes, Notul. Syst. (Paris) 
19. 1960. TYPE: Pseudopeponidium neriifolium 
omolle ex Arénes. 
Leroya Sat nsonia, n.s., 10: 335. 1970. TYPE: 
ya An Cavaco. 
quibos mass Adansonia, n.s., 
Neo opa verdcourti CARES. 
Pyrostria sect dan 
1971. TYPE: Pyrostria amporoforensis Cavaco. 


11: 122. 1971. TYPE: 


11: 393. 


onia, n 


For a description, see Bridson (1987). 

There are at least 80 species 
(55 species in Madagascar [Govaert et al. 2006; 
Davis et al., 2007; Lantz et al., 2007; Razafimandim- 
bison et al., 2007], plus at least eight new undescribed 


ber of species. 


species there, with 14 species on mainland Africa 

ipd 1987; Verdcourt & Bridson, 1991], 
eight on the Mascarenes [Verdcourt, 1983]. The 

S is of species in Southeast Ásia is currently 


E. 


unknown. 


Distribution and habitat. 
Comoro Islands, Madagascar, the Mascarenes, the 


The genus occurs on the 


Seychelles, eastern and southern Africa, and South- 
east Asia in dry, deciduous forests, thicket xerophyl- 
lous forests, lowland rainforests, and both mid- and 
high-altitude and humid forests. 


Diagnostic features. Our newly circumscribed Pyr- 
ostria can be easily distinguished from the other genera 
of the dioecious group by its persistent, basally connate, 
and long acuminate paired bracts, which are relatively 


large and spathe-like in Seyphochlamys revoluta. 


Literature Cited 


Achille, F. 2006. Tinadendron, a new genus of Rubiaceae, 
ettardeae from eastern Melanesia. Adansonia 28: 
> 


——, T. J. M e T Lowry I €: J. Jérémie. 2006. 

Polyphyly i in ds L. (Rubiaceae, Guettardeae) based 

on nrDNA ITS sequence data. Ann. Missouri Bot. Gard 
l. 


74. A new look at the statistical model 


limits in Mussaendeae (Rubiaceae). Amer. J. Bot. 92 
YA 
Andreasen, K., G. Baldwin & B. Bremer. 1999. 

Phylogenetic siis of the nuclear rDNA ITS region in 


A propos de quelques genres Malgaches de 
(Vanguériées et Gardeniées). Notul. Syst. 
(Paris) 16: 6-41. 


Baillon, H. 1879. Mémoire sur les genres Canthium et 
H 19-213. 


aker, H. G 8. Studies in vd d nm ee of 
West ae ee J. n Sei. Assoc. 4: 
9-24. 


— € P. A. Cox. 1984. Further ca on dioecism and 
islands. Ann. Missouri Bot. Gar 4-253. 

Baldwin, B o a Tien es of 
external transcribed spac of 188-268 rDNA 
Congruence of ETS mu 
(Compositae). Molec. Phylogen. Evol. 10: 4 

m I. B. 1877. Phanerogamic vegetation D Rodriguez. J. 

n. Bot. 16: 14-15. 
Sidi K., J. R. Sebola & E. 


R. Robinson. 1996. 


van Medenbach de Rooy (editors), The Biodiver- 
sity @ African Plants. Kluwer Academic Publishers, 
Dordrecht. 


Beach: 1 H. SS K. S. Bawa. 1720; Roles of pollinators in the 
from distyly. Evolution 34: 1138-1142. 


. E. Mene 2000. Phylogeny ad clasifica 
tion of the subfamily Rubioideae (Rubiaceae). Pl. Sys 
Evol. 211: 71-92. 

— — —, R. J. Jansen, B. Oxelman, M. Backlund, H. Lantz & 

ge -—— study from the coffee family (Rubia- 

pe 48: 413-435. 

7. Studies in African a: Vanguer- 

1eae: A new circ Mure ue i 

Canthium a Bullockia. Kew Bull. 42: 611-639. 

The genus Canthium al Vanguer- 
ieae) in dir Africa. Kew Bull. 47: 353-401 

— ———. 1998. Rubiaceae (tribe Vanguerieae). Pp. 211-377 
in e, V. Pope (editor), Flora Zambesiaca. Royal Botanic 
Gardens, Kew. 

Burck, M. W. 1883. Sur l'organisation florale chez quelques 
Rubiacées. Suite. Ann. Jard. Bot. Buitenzorg III: 109. 

1884. Sur josgauibatón florale che: 
Rubisséss: Suite. Ann. Jard. B 

Burger, W. & C. M. Taylor. lora Castanea: 
Family #202. Rubiaceae. Fieldiana 33: 1-333. 


— 


Capuron, R. 1969. A propos des Rubiacées-Vanguériées de 
Madagascar. Adansonia, n.s., 9: 
Cavaco, A. 1966. Contribution à l'étude des Vangueriées 


ede de Madagascar. Bull. Mus. Natl. Hist. Nat. 
38: 700—70 


«T t Poi d pret ME on el mcd 
nidium | anisalov espéc 


(Rubia uer 


nouvelles de eee Adansonia: n.8., 
1968. Espèces nouvelles de Rubiacées de Mada- 

gascar. Bull. Mus. Natl. Hist. Nat. 39: 1015-1019. 
1969a. Contribution à l'étude des 

(Rubiaceae) de Madagascar. Bull. Mus. 


Vangueriées 
Natl. Hist. Nat. 


. 196 P Contribution à l'étude des genres Pseudo- 
peponidium et Peponidium Vangueriées (Rubiaceae) de 
Madag Adansonia, n.s., 9: 43-46. 


970. roya, nouveau genre de Rubiaceae. 
Adansonia. n.s., 10: 333-337. 

197 1a. n nouveau genre de Rubiaceae- 
Vänguericae. Adansonia, n.s., 11: 119-123. 
————. 191b. ie S 

Madagascar. Adansonia, n.s., 


sur quelques Pyrostria de 
11: 393-396. 


Volume 96, Number 1 


Razafimandimbison et al. 


2009 Functional Dioecy in the Tribe Vanguerieae 

—— 2. Les Canthium et les Rytigynia (Rubiacées) de K. Andreasen & B. Bremer. 2002. Nuclear rDNA 
mb Affinités avec les espèces Africaines ITS « used to Par = s n of Vanguerieae 
Nouveaux taxa et combinaisons nouvelles. Puede De l. Syst. l. 230: —187. 


Acta m Sér. B, Sist. 11: 219—247. 
End W. & H. H. Hills. 1991. Silica gel: An ideal 
i for preservation of leaf samples for DNA studies. 
jon 40: es 220. 
Davis, . Govaerts bh M. Bridson. 2007. New 
naa ons in Madagas an Vangie reae (Rubiaceae) 2 
nh genera Psydrax, Pues and Rytigynia. Blumea 5 
139-145. 

Delprete, P. G. iu Rubiaceae (part 3) Fr YS 
Condamineeae. Pp. 5—53 in G. Harling & L. Andersson 
(editors), D. E do Vol. 62 (7). Góteborg Univer- 
sity, Góteborg; Riksmuseum, Stockholm; and Pontificia 
dapi ien Católica del Ecuador, Quito. 

0. Melanopsidium Colla (Rubiaceae, Garden- 

plex 


ieae): A mono pecie Brazilian genus with a com 
nomenclatural history. Brittonia z N 

Doyle, J. J. & J. L. Doyle. 1987. id DN lation 
procedure for small quantities P pus les tissue. 
mU Bull. 19: 11-15. 

S. 1989. The retention n ciu the rescaled 

consi i index. Cladistics 5: 4174 

Kállersjó, A. G. Kluge & C. ds 1995. Testing 
Sloane of incongruence. iliis 10: 315-319 

Friedmann, F. 1994. Flore des Seyc ne d Dicotylédones. 
Edition i lORSTOM, Institut Frangais de Recherche 
scientifique pour le Développement en A Paris. 

Govaerts, R., L. Andersson, E. zo D. Bridson, D. 
Davis, I. Schanzer & B. World Checklist of 
Rubiaceae. The Board of -N E the Royal Botanic 
Gardens, Kew. 

ucc A. 1930. Matériaux pour la Du E a P wa 
Calédonie. XXVII. Révi s Rubia P. 48 i 
rit (editor), pun de Bound, Tome 3, ee 


calles ta. Taxon 51: 661-674. 

. 1990. Vaden on a distylous theme in 
Mesoamerican Psychotria subgenus Psychotria (Rubia- 
m. New York Bot. Gard. 55: ae 

M. t Identification 
Termin ology. E Lake Publishing, Spring Lake, Utah. 


Heilbuth, J. Lower 2 richness in dioecious 
clades. pha pa aa : 
Hooker, J. D e soe Addenda 


Corrigenda. Pp. 535—536 in 5 E & J.D "Hocker 
(editors), Genera plantarum ad exemplaria imprimis in 
herbariis kewensibus servata defirmata, Vol. 2, Pt. 1. 
Reeve & Co., London. 
Huplseabeck, J E & F. Ronquist. 2001. MrBayes: Bayesian 
Bioinformatics 17: 754-7 
1789. Genera Plantarum Secundum Ordines 


Jussieu, A. 


& J.S. Farris. 1969. Quantitative phyletics and the 

evolutiori of anurans. Syst. Zool. 18: 1-32. 

Lantz, H. & B. Bremer. 2004. Phylogeny inferred from 
morphology ind DNA bw n cie well-supported 
groups 1 ). Bot. J. Linn. Soc. 146: 


— —————. 2005. Phylogeny of the complex 
Vanguerieae (Rubiaceae) genera Fadogia, Rytigynia, 
ngueria with close relatives new circum- 


and Van, and a 
scription of Vangueria. T Syst. Evol. 253: 159-183. 


od 2 s. G. Dui. mandimbison & A. 
— Mouly. bot “Three new species of Vanguerieae (Rubia. 
ceae). p sér. 3, 29: 129-136. 


Leroy, J. F. 1972. La notion de genre et l'évolution: Sur un 
an senarqalle de différenciation t chez les 
Rubia ueri à Madagascar. Compt. Rend. 


d. Sci. Paris = 1682-1685. 
T L, E er & J. D. Thompson. 2005. Gender 


Antirhea borbon: 


E s. S. € D. Caden 1991. rn dioecy in 
25. 


MeDouga all, L F.H. Clamalaai: 1969. Dos dating and 
P E Pw studies on volcanic rocks from 
Mauritius, n Ocean. Geol. Soc. Amer. Bull. 80 
1419- 


Mouly, A. 2007. Étude systématique des Rubiaceae de 
Mayotte et des Comores. Rapport d'expertise à la Direction 
de l'Agriculture et de la Forét, Préfecture de Mayotte 

———, S. G. Razafimandimbison, F. A 

mans & B. Bremer. 2007. 
Rhopalobrachium (Rubiaceae: Ixoroideae): E 
uis us (rps16 and irnT-F) and morphological data. 
Syst. 72-882. 

Naiki, ni s M. Kato. 1999. Pollination system and evolution 

of dioecy from distyly in Mussaenda parviflora (Rubia- 
ae). Pl. Spec. Biol. 14: 217 Fee 

ner & E. A 


Psychotria (Rubiaceae): A 
likelihood approaches. Syst. Biol. 52: 820. 

04. MrModeltest Program 
utionary E ogy Pius 


Os 3. Two-sex Perrier projection of the 
endemic and dioecious rainforest shrub, Gardenia actino- 
carpa Rad Biol. Conservation 114: 39-51 

Pailler, T., u, J. Figier & 
een trait variation in the functionally dioecious 
and i eae ces island endemic Chas- 
salia corallioides. B n. Soc. 64: 297-313 

& F is 1998. Cryptic dioecy i in 

Ec borbonica var. borbonica (Rubiaceae), an endem 

La Réunion Island. Acta Bot. Gallica 145: 


Species of 
29-38. 


Percy, D. M C. B. Cronk. 1997. Conservation in 
relation mating system in Nesohedyotis arborea 
(Rubiaceae), a rare endemic tree from St. Helena. Biol. 
s N 135-145. 

Persson, C. 0. Phylogeny of the Neotropical Alibertia 
group (Rubiaceae) e emphasis on the genus Alibertia, 
inferred fro 


5S ribosomal DNA sequences. 
28. 
———. 2003. Agouticarpa, a new dL genus of Tribe 
Gardenicae E E 55: 176-201. 
Popp, M. & B. Oxelman. 2001. Inferring the history of the 
polyploid Silene aegaea aro can using plasmid 
sequences. Molec. 


999. Revision of Atractocarpus (Rubiaceae: 
Gardenieae) in Australia and new combinations for some 
extra-Australian taxa. Austral. Syst. Bot. 12: 271-309 


178 


Annals of the 
Missouri Botanical Garden 


Rambaut, A. 1996. Se-Al: Sequence Alignment Editor (Ver: 
1.dl). Available at <http://tree.bio.ed.ac. uk/software/seal/>. >. 
Razafimandimbison, S. G. & B. Bremer. 2001 [2002]. Tribal 
delimitations of Naucleeae (Cinchonoideae, Rubiaceae): 
molecular and morphological data. Syst. & 
Geogr. Pl. 71: 515—538. 
——— & oe Pyle aa classification of 
Naucl (R ecular bans rbeL, 
and ae and ds did. dos. J. 89: 
1027-104. 
, E. ^ Kellogg & B. Bremer. 2004. Recent origin and 
putative pseudogenes: 
A case study from Naucleeae (Rubiaceae). Syst. Biol. 53: 
177-192. 


a 


of divergent ITS 


og, H. Lantz, U. Maschwitz & B. Bremer. 2005. 
mus of mon "An oa a mymecophyt- 
ism, and rapid radiation in uclea s.s. (Rubiaceae). 
Molec. E Evol. 34 au —354. 
, H. Lantz & B. Bremer. 2007. New combinations and 

names in puce and Pyrosiria (Rubiaceae, Vanguer- 
ieae). Novon 17: 516-521 
Renner, S. S. R. E. Riles 1995. E n zs its 
correlates in the flowering plants. Amer. J. Bot. 82: 
6—606. 


ide E. 1979. The African e da A. Rich 
(Rubia E ubgenus Empogoma. 

Bull. Nat t Blastentü Belg. 49-- 239-300. 

1988. Tr Ki woody Rubiaceae. acd 
essions. Contributions to 
subfamilial EC ite Opera Bot. Belg. 1: 1-272 

Saghai-Maroof, K. M., M. Soliman, R. A. Jorgensen & R. W. 
Allard. 1984. Ribosomal DNA spacer length polymor- 
phism in barley: Mendelian inheritance, chromosomal 
location, and population dynamics. Proc. Natl. Acad. Sci. 
U.S.A. 81: 8014-8018. 

Sakai, A. K., W. L. Wagner, D. M. Fergurson & D. R. Herbst. 
1995. Origins of dioecy in the Hawaiian flora. Ecology 76: 
a 

uv G.E 1. Generic Tree Flora of Madagascar. Royal 

otanic iade Kew, and Missouri Botanical Garden, St. 


features 


je 

Bises C. J. 1922. Catalogue of Easter Island phaner- 
ogams 4 in Natural History of Juan Fernandez 
and Easier Island, Vol. 2. Almquist & Wiksells, Uppsala. 


—————. 1944. On the flower dimorphism in Hawaiian 

Rubiaceae. Ark. Bot. 31A(4). 
945. The flower of Canthium. Ark. Bot. 32A(5). 
eei. M. P. & H. Ochoterena. 2000. Gaps as characters 
in sequence-based phylogenetic analyses. Syst. Biol. 49: 
369-381. 

Smith, A. C. & S. P. Darwin. 1988. Family 168. Rubiaceae. 
Pp. 146-362 in A. C. Smith (editor), Flora Vitiensis 
Nova. À New Flora of Fiji (spermatophytes only), Vol. 4— 
Angiospermae: MEL ne ies. Pacific Tropical 
Botanical Garden, Kauai, H. 

Staden, R. 1996. The 2 
Molec. Biotech. 5: Ge 

Swofford, D. L. 2000. P Sudan e B 
Parsimony (* and bno: ae thods), Vers. 4.0b. Sin 
dedo Massachussetts. 

Thompson, A. J. D 

L 


sequence analysis package. 


T. G. Gibson. 1997. 


ean specific gap penalties and weight matrix choice. 
ucl. Acids Res. e 4673-4680. 
van m C. F. 1967. A revision of the Malesian and 
P ae T the genus Gaertnera Lamk. (Rubia- 
ceae). Blum —391. 
Verdcourt, B. 
e Rubiaceae. 
209—281. 
. 1983. Notes on Mascarenes Rubiaceae. Kew Bull. 
37: 563-570 


. Remarks on the classification of 
Bull. Jard. Bot. État Bruxelles 28: 


991. Canthium. Pp. 861—885 in R. 
M. ki prey i of Tropical East Africa. Balkema, 
Leiden. 

Wagner, W. L. & D. H. Lor . 1998. A new, dioecious 
species of Hedyotis (Flaca from a Hawaiian 

nd the awe a Hedyotis schlechtendahliana 


ae & M. J. Donoghue. 2000. 

Phylogenetic ans of dioecy in monocotyledons. Amer. 
Naturalist 6-58. 

Wong, K. M ica ji synopsis of Morinda (Rubiaceae) in the 
Malay Peninsula, with two new species. Malayan Nat. J. 
38: 89-98. 

Yang, Z. 1994. Maximum likelihood phylogenetic estimation 
from DNA sequences with variable rates r sites: 
Approximate cn ds. J. Molec. Evol. 39: 306—314. 


179 


Razafimandimbison et al. 


Volume 96, Number 1 


2009 


Functional Dioecy in the Tribe Vanguerieae 


TIEPESNAA PLEPESOAH TeoseSepepy (S9 036 yp qo UOS LIST i ii ‘rowolg “Y av zyue'T ^ungezeq (oyeg) wunmgofixng wnpruodaq 
SOST8S04 ELEPFSSNA TeoseSepepy (S 616 yp 12 UOS LIST € b: “19U9IG a8 m zyue'T E Urtfezew (oyeg) wnyofixng wnpruodaq 
COEPSSOA OLEPSSNA TeoseSepepy (S 606 Pp qo UOS LIST ra » “19U9IG a8 m zyue'T 5 weze y (oyeg) wunmgofixng wnpruodaq 
€0€T8S04 IZeT8S(14M TeoseSepepy (S 806 yp qo UOS LIST T x “19U9IG "d m zjue'] ^urgezeq (oyeg) wunmgofixng wnpruodaq 
FOLLIOLV 8cevesna (007) 1oue1g 29 ziueT OBALY) rr«moopaoa, DÁo42]09N] 
O86FL ELV (GQ0Z) 1ouro1g zp ziue'T uospug (umuog "N) »dupo042]0$ visuopinpy 
O80STELV (2003) Te 19 ziueT "DPI9A 29 uospug (oo[Ng) sua2s242u02 DIUIPIN 
£9LLTOfV Ocevy8snu (FOOT) 1oue1g 29 zie] OBALY aptparuon "je “IPU "ds n£o42T 
99ZV8SNA LZSPSSNA TeoseSupe]y (ug) SFIT T" 12 org aq “apur ds ooeae) plow] 
Z9LLT9ÍV (POOT) 1oue1g 29 ziueT uospug (AO) psouza 0422Y 
9cgrcc[ V (6661) [e 19 uosvorpu y "[ P2419209 D4OX[ 
ZOTSTELV (PONT) 1owe1g I ziv (2007) "Te 19 ziuv'T sudqoy pinging nruosunjomgy 
OOLLIOLV oceresnd (007) 1oue1g 29 ziueT SUSO "f X APH ON (95104 75) ummquoq wumjpajdojok) 
06ZPSSN4A LSEPSSNA Pruopo[e;) MON (d) 181 “ron ununve[pm?) (gpeg) apsunyng unp4ydop ho 
Toeresna 89€ P88N0H pue[s] eno&v]y (d) szg1 ¡Puig "Meg xe utarog SCA bas umngun) 
683-8904 pue[reur, (AVY) 83122 paxo umpafuoo "qe opur ds umngup;) 
cocr9sna Tceresna TeoseSupe]y (D 0301 sang 8 opur “ds umnprum;) 
£9cr9sna eseresna TeoseSupe]y (D £611 sang L yepur "ds wnnpuvy 
S9cr9Sna Scerasna TeoseSupe]y (ug) Q0II T" 12 org oq 9 pur "ds umngup;) 
PStpesna tTceresna TeoseSupe]y (3D GLE ojosmuorosuy wz sianqg q yepur "ds wnnpuvy 
£6cr9Sna 6SEPSSNa TeoseSupe]y (sam) FOS `P 12 pouyoavy p opur "ds umngup;) 
8L3v8Sna creresna TeoseSupe]y (OIN) E901 "I? 12 s24oN € yepur "ds wnnpuvy 
c6cr9sna eseresna TeoseSupe]y (San) OIE T» 1» powoany g apur "ds wnnpuvy 
LOEPSSNA 9LEPSSNA Teose3epe (S) 0301 P 1? uossyug T aepur ds “wey umngup;) 
LGLLIOfV TZEPSSNA (FOOT) 9uo1g 29 ziueT uosprig wropfizasopnasd un pun 
Tocrgcna ozeresnd vduoyy (S) ezgg I» 12 vyoBuna y Peg asuaznquou ummy) 
9e0€r9cna LLEPESNA TeoseSepepy (S) 685 oaumuojoavy 2p uosiquapupunfozog ooeaer) asuaAfoforoui UNA YUDr) 
OZISTELV (FOOT) 1owe1g 9 ziuvT (ZOOZ) "Te 19 ze] ozymy auuaur umm) 
09cr9Sn 6Ter9sna vduoyy (san) FIO6 err] uospug nuapof unn) 
66ZV8SNA eoeresna pue[s] onoke (d) 09z€ I» 1» PPT 02v^e7) xo umuÁg uoodusko um un 
883-8904 puepreup, (n vv) 396€ jenxop $ pos PLOW quoy unnofuoo Un pun) 
LPLLIOLV (p007) 1euro1q zp ziueT (y ueurroods) -Jsoq snsouiZiqna soyyun¡hou y 
crgrcc[V (6661) [e 19 uosvorpuy LON q DUsDW Duaq]y 
6SLLIOLV (FOOT) Towerg 29 zme zjueT (osnury “y) nsmuaqois umnpnuooafy 
8SLLTO[V (rooz) Ioulelg p zjueT zjueT (91300 'S) wnyoyaysaaopnasd UN JUDIOS Y 
SLI SIA uguo jo &nuno?) UOTJEULIOTUI 1SYINO A 


exe 


"SISQUINU uorssoooe pue “surgrio Ánunoo *uoreurojur 1oqonoa ‘exe poouonbos poreSnsoAur oy} jo asr] ‘q xipuoddy 


Missouri Botanical Garden 


Annals of the 


OTePSSna 6LEPSSNA TeoseSepepy (S) 9g9 oauuojoavy 29 uosiquapupunfozog "urgezey ZW YWA jue] oynpuad vrysoidg 


L8Ch8SNA ¿Seres naa pue[sp uorungy eT (uorungg Y] op SMPNSIOAU()] 191G19H) GBT 4204 "gon cy sumgmorgao 145044 
olerssna eseresnd twosesepe lA (San) TTS vosiquapununfozny oowaey ('Qorg ^v) vapeur piso 
POEP8SNA TLEPSSNA Teosesepe yA (S) 816 P 1» uossqus 09848) (ong ^y) 40fmu pugsosk g 
FPIISISÍV erergcna (P003) 1euro1q zx ziueq uosprrg (‘youroig) xus ugso4q 

ecergena pue[sp uorumoy eT (uorungy eT op QMPJNSISATU() IAH) BZZ PVJ quie) 7p pof muosiounuoo 10443 
€TTSTELV COoergcna (FOOT) 1eurexg zx ziuvT (Z00Z) Te 19 ziueT OWALNY (1oxeg) v1D21004q1Q VISO Y 
T9cT9S0W 6rer8Snt TeoseSepepy OD 3301 sq OBAL) XƏ SOUQIY SI8UIIQOZDYUD DLUSOIÁS 
L6cv9S04 99epocn  (e»uerp pue[sp eno&epy (d) 8081 P P jpusiq OBAL Xo Souory smuaunnofun DLUSOIÉ A 

69€P8SNA JeoseSepepy (San) 297 vosrumupunfozoy ap uosiquipupunfozoyg ONAC SISUJUD]IPUD DIAISO4Á 
OTISIGÍV (P003) 1ouro1q 29 ziueT viopfiarod "dsqns uospug ('Tozyy) 240jfiaaod xvapÁsq 
GOTSTELV 9Terge na (5007) 19u19 29 ziueT uospug (1497 29 “POH xe qoszio[y) DIDAOGO xp4p&sq 
892219[V (P003) 1eure1g y zue uospug (uxor) sapiorssnpas xvapAsq 
86278804 poerssna Teosesepe yA (Sdn) prs wosiquipupuafozoy q pur ds umipruodadopnosq 
S6cvY8S(14M TOEP8SNA TeOSeSEPE] (Sdn) ZZZ 9 12 payasvy [ pur ‘ds souory xo opowoy wnrprucdadopnasg 
LLcVSS[Y4 Trerssna z euo[o 
OLZP8SNA OPeresna JeoseSepepy (NVI) 337 72 12 uosiquapununfozoyg [ UOP seugry xe e[[ouogg wnyofiajo wmpruododopnosq 
b6cv8S 04 T96v8S/[14 Teosesepe yA (San) G&S vosiquapununfozoy ooeaer) unsopfiduo] wmpruodadopnasq 
L6cPr8SNa SSEPSSNaA TeoseSepepy (Sdn) 993 9 1? payasvy souory xo o[[Ourogp wnyofiiox umripruododopnosq 
cLCV8S(1M LEEPSSNA (FO0Z) 19tuorg x zuvT seugry nsosp wmipruodadopnas y 
997 2TO[V Geergcna (poooz) 1otuarg zx ziueq sougry asuacioidun wnrpruodadopnas gy 
PISPssna Teeresna xeoseS epe] (OIN) TEE 4opuvuqgpo) sougry umumnjoa umipiuodaq 
P8ePrssna ZTEP8SNMH JeosegEpepy (Sdn) EZZ P 1 payasvy aSUaLIDISDADPDUL "ye Pi ‘ds wnipmodag 
OOSPSSNA lL9€er9cna JeoseSepepy (NVI) [9¢ vosiquipupunfozpy 1901919 “Y 29 zue ^urgezew (1exeg) wnyofixng "ge “opu "ds umipruodaq 
SLopesna € euo[o 
PLOPSSNa 6EEPF8SNH Z euo[o 
ELOpEsna 8Eersesna 1eosesepeIo (NVI) gp vurunomupupuy p uosiquapupunfozoy q 9uop» 9 neputi ds wunspruodag 
TLZP8SNA TECPSSOM xeoseg epe] (X) 8301 sang € oput ds wmipruodoq 
OLZPSSNA €£er8ena Teosesepe yal OD SEIT sug p pur "de rumipruodog 
692 V8SN0H ceeresna Teosesepe yA (San) 9oFg-oror To 19 owog € “opur ds wnipruodag 
96cV8SstYX c9€r8S Teose3epe (san) 093 T» 12 payaspy g aput "ds umipiuodaq 
90EP8SNA SLEP8SNA Teosesepe yA (S) 6001 "JP 12 uossqu T apur -ds souary wmnipiuodoq 
gocrgsna ogerssna teosesepe (GAN) [£p Pulunowunupuy 2p uosiquapupuafozoy ODPAB) osuaLmpospSppmu wnipruodag 
L9cvY8S(YX 6GEP8SNa Teosesepe (Sdn) gzp vosmpupunfozuy 3p uosiquipupunfozoyg souory (peg) wnpruoy unipriodoq 
€TEP8SnaA TSEresna JeoseSepepy (S) 8g9 oanmuojoany 3p uosiquapupunfozog "ungezey 29 opepy "eT unrofisspao wumipruodaq 

SLI SLA uiuo jo LQunor) UOTJEULIOJUT 19YIMO A, exel, 


180 


Penunuo) ‘q xipueddy 


181 


Razafimandimbison et al. 


Volume 96, Number 1 


2009 


Functional Dioecy in the Tribe Vanguerieae 


€60STEÍV (2003) ‘Te 19 ziuvT "omg »jennfur v122n8unA 
960STELV (5007) 1oue1g X ziueT (2007) Te 19 ziueT sudqoy (urerg]) suaospaouro wunjpegdidp y 
OLLLIOLV psepesna (pooz) owog 29 ziuvT 3 Weg vmnpoass shrunpyooydiog 
TOISISfV (poooz) 1eu1e1g y ZUR] (2007) Te 19 ziueT oum[g sisuappSauas muksuky 
PLLLIOLV (FO0Z) 1ewuo1g x zuvT Ymy 2749012 DISULGOY 
98TP8SNA TSEP8SNH sure (NVI) “u's v42uopojoquimapuy “Opto A (1exeg) sapioumqia vtgsoatq 
LECPESNA OSePresna 1eosesepeIo (Sd) [Sp vusunoppunupuy s uosiquipuvunfozoy L ‘ypu "ds o1gs044q 
¿83 v8S.0131 LVEPSSNA 1eosesepe (Hd) ZIP P 12 uosiquapupunfozoy 9 put "ds 145014 
T8crasna 9vergsnot xeoseSepe]y (ug) ZIII P 19 org eq G oput “ds nzgso4£q 
SIgTSS O8£P8SNA TeoseSepepy (Sg) p69 oauuojoavy op uosiquapupunfozog p pur "ds vrgsoJkg 
60EP8SNA 8LEF8SNAH Teose3epe (S) ZF6 "9 p uossqas € opur "ds o1gso4q 
GTePesna c9erocna TeoseSepepy (Sg) 879 oaupuojoavy 2p uosiquapupunfozog Z pur "ds D14gs01 Lg 
6LEPSSNA Serena Twosesepe lq (OW) 022 `P 12 oapunoyumuoquy T “pur ds "senf xo umor) 245044 
€9cv9S04 SPersessna TeoseSepepy OD Zorr sq "npezey z De] UET vurquadios oiisoiq 
09cV9SnM 9€tV8S(Y4 twosesepe lA (OW) ZZI£ vypunyisaqny 39 uosdumjq OBAL) SIfU2UDADOADS DIMSOL Y 
STISTELV Pprerssna (FO0Z) 19uo1g x ZUR] uospug (reg) vapioyzuny 4d oiisoaq 
SLI SLA uiuo jo LQuno) UOTJEULIOJUT 19M0 A exel, 


Penunuo) ‘q xipuoddy 


THE RONDELETIA COMPLEX Johan H. E. Rova,? Piero G. Delprete,* and 
(RUBIACEAE): AN ATTEMPT TO Birgitta Bremer 

USE ITS, RPSI6, AND TRNL-F 

SEQUENCE DATA TO DELIMIT 

GUETTARDEAE, RONDELETIEAE, 

AND SECTIONS WITHIN 

RONDELETIA! 


ABSTRACT 


n the present study, a molecular phylogeny of bu on L. ped (Rubiaceae, Rondeletieae) was constructed with 
" following main objectives: (1) t posed by Fernández aqua ©) to test if Pionai 
Poit. belongs to the Roadeleione s. str.; (3) to check if üs data fro deleti previous fr 
trnL-F p regarding circumscription of Rondeletieae; and (4) to xs if Hodgkinsonia F. Muell. belongs to DUE or 


recognized by Fernández ueira were included in the ITS analysis. Five of her sections RE be tested for mon 
upport was only found for Rondeletia sect. Leoninae M. Fernández Zeq., while representatives from section Chamaebuxifoliae 
M ández Zeq., section l ll., and section Nipense. rnández Ze rm a well-supported clade 


studies, the two species described under Cuatrecasasiodendron (C. spectabile Ste d C. colom dl. 
eyerm.) are treated as synonyms to the new combination Arachnothryx spectabilis dium Rova, Delprete & B. Bremer, 
which is proposed here. 
Key words: Arachnothryx, Cuatrecasasiodendron, Guettardeae, Hodgkinsonia, ITS, phylogeny, Rogiera, Rondeletia, 
Rondeletieae, rpsi6, Rubiaceae, Stevensia, irnL-F. 


ith id Rondeletia 


! We are grateful to Attila Borhidi (Janos Pannonius University, Pécs) for hel 
and related taxa and to Nahid Heidari and the staff at the Evolutionary Biology Centre ae a for technical 
on 


Hagaberg Baskarp, SE-566 92 Habo, Sweden. Author for correspondence: jorova@telia 

3 CNPq Visiting Scientist, Institute of Biological Sciences-ICB-1, Department of ue Biolsgy/Botany, Universidade 
Federal de Goiás, Campus I, 74001-970 Goiânia, Goiás, B hail Current address: Institut de Recherche pour le 
Développement, Bougie et EE PE de a des Plantes (AMAP), TA-A51/PS2, Blvd. de la Lironde, 
34398 Montpellier Cedex 5, France. 

* Bergius E at the Royal Swedish Academy of Sciences and Botany Department, Stockholm University, SE-106 91 
Stockholm, Swe 

doi: 10. as 


ANN. Missouni Bor. Garp. 96: 182-193. PUBLISHED ON 23 APRIL 2009. 


Volume 96, Number 1 


Rova et al. 
The Rondeletía Complex 


183 


The tribe Rondeletieae (Rubiaceae, Cinchonoi- 
deae) includes predominantly shrubs and trees and 
is mostly distributed in the New World tropics 
(Robbrecht, 1988; Delprete, 1999a), with the main 
center of diversity in the Greater Antilles. A thorough 
description of the taxonomic and systematic history of 
the tribe is found in Delprete (19992). 

The largest genus of the tribe, Rondeletia L., is 
mainly Antillean and comprises approximately 120 
species. roh (1918) divided Rondeletia into 15 
ased on morphological and distributional 

l and molecular 


studies in the Rondeletieae have argued about the 


. Since then, ae morphologica 


status of d sections and the circumscrip- 
of the genus Rondeletia 
d dodi be re 


scribed genus, separated from morphologically similar 


. One opinion is that 
ped as a narrowly circum- 


genera such as Arachnothryx Planch., Javorkaea 
Borhidi & Jarai-Koml., Rogiera Planch., Roigella 
Borhidi & M. Fernández Zeq., Reca Borhidi, 
and Suberanthus Borhidi & 
O 1967; Borhidi & Fernández Zequeira, 

a, b; Borhidi, 1982, 1989, 1994; Borhidi & Járai- 
cr 1983; Fernández Zequeira, 1994; Delprete, 
1999a, as Rondeletia complex sensu Delprete; Rova, 
1999; Rova et al., 2002; Borhidi et al., 2004; Rova 
unpublished). On the other hand, Lorence (1991) 


recognized Rondeletia as a widely circumscribed 


M. Fernández Zeq. 


genus, treating the names applied to Mexican and 
Central American taxa of the complex as synonyms. 


Antillean (especially the Cuban) Rondeletia species 
into 10 sections. Her classification comprise 
species, most of them endemic to Cuba. This means 
that a majority of the species of Rondeletia s. str. were 
d in her study. According to Fernández 


uished by 
ultistate) do e og- 


d 1994), the sections are distin 
ous combinations of (often m 

ical characters such as position and shape of 

inflorescence, flower merosity, calyx lobe shape, leaf 
mentum, and e axy 

te). How oc 

did little to bs the cn in the larger 


in pir opposite vs. 
verticillat on Cuban species 
Rondeletia complex. The first aim of the present 
study was to test if Fernández Zequeira's sections of 
Rondeletia are supported by phylogenies obtained 
from molecular sequences. The second aim was to test 
if Stevensia Poit. was closely related to Rondeletia or 
not. Stevensia has not been included previously in 
molecular phylogenies, but morphology suggests a 
close affinity between the genera. Earlier studies 
(Bremer et al., 1995; Bremer & Thulin, 1998; 
Andersson & Rova, 1999; Rova et al., 2002) have 
shown that the tribes Guettardeae and Rondeletieae 


are closely related, and this study also aimed to 
investigate if ITS data would suggest a similar 
circumscription of Rondeletieae as previous studies 
had. Finally, the study was aimed to investigate if ITS 
sequence data would support Bremer’s i o 
sion of the Australian genus Hodgkins 

in the Guettardeae or Delprete's (1996) vetus of de 
genus from the Guettardeae to the Chiococceae. 


MATERIALS AND METHODS 


For the ITS analyses, material was sampled from as 
many Rondeletia species and subspecies as possible. 
An effort was made to include representatives 
from all genera in Rondeletieae sensu Rova 
(2002). The outgroup consisted of Luculia Sweet 
(basal in Rubiaceae, e.g., Bremer et al, 1995), 
Catesbaea fuertesii Urb., Chiococca alba (L.) Hitche. 


Chiococceae s.].), and 12 accessions representing 11 


=> 


species in the following six genera of the tribe 
Guettardeae (based on available material and the 
2002): Arachnothryx, 


Cuatrecasasiodendron Steyerm., Gonzalagunia Ruiz 


results from Rova et al, 


av., Guettarda L., Rogiera, and Timonius DC. 
Authors of species names are given in Table 1, or 
otherwise when first mentioned in the text. 
resh or silica gel-dried leaves were used for DNA 
vailable, but ofi 
material had to be used. DNA was extracted using 
the CTAB method (Doyle & Doyle, 1987) and cleaned 
with the QlAquick PCR Purification Kit (QIAGEN 
GmbH, Hilden, Germany). The cocktail for polymer- 
ase chain reaction (PCR) amplification was mixed as 
follows (to ca. ul: 2.5 u X buffer, 2.5 ul 
Cl», 2 wl dNTP, 0.125 ul Taq DNA poly- 
merase, 0.625 ul 10 pM forward primer, 0.625 pl 


extraction when ten herbarium 


everse primer, 2.5 ul 0.1 M TMACI, 2 pl 
neo sulfoxide a 2 ^ aie: and 10 ul 
s, the amount of primer or template 


was doubled Gol some of the water). Primers 
P17 and 268-82R (Popp & Oxelman, 2001) were used 
for amplification. Sequencing reactions were realized 
using the DYEnamic ET terminator Cycle Sequencing 
kit (Amersham Biosciences, Buckinghamshire, Eng- 
land) following the T. of the manufacturer 

s added in the same concentration as in 
the PCR mix) and run on a MegaBACE 1000 DNA 
Analysis System (Amersham Biosciences). For se- 
quencing, the same primers were used as in the PCR 
amplification. 

For the ITS study, 50 new ITS sequences were 
produced, and five additional sequences were down- 
loaded from GenBank and included in the data matrix. 
and GenBank 


accession numbers are presented in Table 1. 


Taxon names, authors, vouchers, 


184 


Annals of the 
Missouri Botanical Garden 


Manual alignment and gapcoding of the ITS 
sequences were performed with the following criteria: 
(1) an effort was made to see if gaps/insertions could be 
interpreted as caused by repeats or inversions, and if 
so, sequences were aligned to fit these possible events; 
(2) gaps (i.e., 
introduced into the sequences to keep the num 


inferred insertion/deletion events) were 
ber of 
substitutions in an aligned region to a minimum; (3) 
insertions/deletions and substitutions were considered 
equally probable ue and (4) a of 
equal ena s or more taxa were inferred 
to be is ME binary coded. c of more than 
one position in length introduced due to multiplication 
of single nucleotides, e.g., poly-A, were not code 
Regions where alignment could not be unambiguously 
interpreted were removed from the analysis. After 
alignment, two ITS matrices were produced, one 


alyses, conducted with PAUP* 
version 4.0b10 (Swofford, 2000), were performed for 
each matrix. The first ITS analysis was a heuristic 
search (random addition sequence with 1000 repli- 
cates, tree bisection-reconnection [TBR] branch swap- 
ping, and MULTREES option in effect), and the second 
analysis was a jackknife search (faststep search option, 
10,000 replicates, and Jac resample emulation). 

For the combined analyses, the data matrix from the 
ITS study was combined with the entire matrices from 
the irnL-F study of Rova et al. (2002) and from an 
rps16 analysis (Rova, pig a er the indel 

codings from each matrix. Previous analyses of each 
separate data set m in similar ie which 
implied that the data sets were congruent. Taxa not 
included in the combined analyses were then deleted 
using the command DELETE in the PAUP block. The 
resulting set of sequences comprised 20 ingroup taxa. 
This set included all taxa where all three sequences 
were available and all Rondeletia species where at 
least ITS and rpsi6 sequences were available. 
Chiococca alba was used as outgroup, e 
ie eased it had been | ir not to 

of the et al., 
Er by a heuristic NN a addition 
sequence with 1000 replicates, TBR branch swap- 
ping, and MULTREES option in effect). 


group (Rova ae were 


RESULTS 


More than 50 DNA extractions were obtained from 

ut on 7 of these 
cies) were amplified by PCR 
Extractions that did not 


Rondeletia i d 
(representing 
and yielded indien 


collected in silica gel almost always worked for PCR 


and resulted in high-quality sequences. It was not 
possible to obtain sequences from section Lindenianae 

ández Zeq., although extractions were at- 
tempted from two different specimens. It was also not 
possible to obtain PCR products from Roigella 
correifolia (Griseb.) Borhidi & M. Fernández Zeq., 
which trnL-F data showed to be closely related to 
Rondeletia s. str. (Rova et al., 2002). For four sections 
(Rondeletia sect. Odoratae Standl., section Pedicel- 
lares A 


tandl., section Rigidae M. Fernández Zeq., and 
Chamaebuxi, 


section uxifoliae), it was only possible to 
sequence one species from each section. We were 
eu not able to sequence the type species of 

ondeletia, R. americana L. This species seems to be 
very eds collected, and extractions made from the 
herbarium material that we found in the Swedish 
Museum of Natural History Herbarium did not nd 
despite several attempts. 
establish conta 


with recently collected material from St. Vincent, 


we not to 


emp e 
ct with anyone who could assist us 


where the species is endemic, and it was not possible 
to do such fieldwork ourselves. Ten species that 
yielded sequences were not listed under any section in 
the work of Fernández Zequeira, but four of them 

on the key 
provided in her paper (Fernández Zequeira, 1994): R 
inermis (Spreng.) Krug & Urb. and R. pilosa Sw. 
belonging to section Leoninae, and R. hameliifolia 
Dwyer € M. V. Hayden and R. purdiei Hook. f. 
belonging to section Calophyllae M. Fernández Zeq. 


could be assigned to sections base 


Sectional affinities are indicated in Figure 1. 

The first ITS matrix, without indel coding, included 
699 characters, of which 
informative. The second ITS mat 

e dede inclu ded 123 ee ol which 198 
were parsimony informative. The combined ITS, 
rps16, and trnL-F matrix included 2751 characters, 
1451 of which were parsimony informative. 

The strict consensus tree obtained from the ITS 
analyses is presented in Figure 1. Heuristic searches 
of both data sets each resulted in 48 most parsimo- 
nious trees. Tree lengths were 768 (consistency index 
CI] = 0 RI] = 0.76) in the 
heuristic search of the data set without indel coding 
and 805 (CI = 0.65, RI = 0.77) m the heuristic 


search where indels were coded. 


= 


56, retention index [ 


ict consensus 
trees were identical ` both data sets. Jac 

support was not found for all clades in strict 
consensus trees from the heuristic searches, and 
jackknife support values for a clade could vary up to 
more than 10 units between the two data sets. Tree 
topologies differed only marginally between the two 
jackknife searches. The jackknife search without 
indel codings found one clade that was not found in 
the other jackknife search (or in the heuristic 


Volume 96, Number 1 
2009 


Rova et al. 
The Rondeletia Complex 


185 


searches), and the jackknife search with indel codings 
resulted in two clades not found in the jackknife 
search without indels coded (Fig. 1). Support for these 
clades was low in all cases. 
The analysis of the combined matrix resulted in 12 
a parsimonious trees (length 2046, CI = 0.91, RI 
strict consensus of these trees is shown in 
us 2, and branches from the consensus tree that also 
occur in the ITS analysis are marked in bold in Figure 1. 


DISCUSSION 


ECTIONAL CLA 


IFICATION OF RONDELETIA 


The main aim of our study was to test Fernández 
Zequeira's (1994) classification with 10 sections of 
Rondeletia using molecular phylogenetic analyses. 


t to fin 
silica gel-dried material that would work 5 PCR and 


extensive search, it was difficult 


sequencing. For five of the sections, only one 
representative of each could be sequenced. Further- 
more, ITS data are obviously not variable enough to 
provide resolution among sections Hypoleucae and 
Nipenses. Nevertheless, we obtained several interest- 
ing results with regard to the circumscription of 
Rondeletia and some of Fernández Zequeira's sec- 
tions. 

There i is strong support for the Rondeletia s. Str. 


y exceptions 
to this distribution are R. hameliifolia from Central 
America (Panama) and R. purdiei from South America 
(Ecuador). Neither R. hameliifolia nor R. purdiei were 
included in Fernández Zequeira's (1994) treatment, 
but according to her identification key, both species 
would belong to section Calophyllae. In our study, the 
two species form a clade with strong support. A third 
representative of this section is R. alaternoides A. 
Rich. from Cuba, which is found in clade F (Fig. 1). 
us, ITS s inn data do not support a monophy- 
letic section Calophyllae. 
Rondeletia deamii (Donn. Sm.} Standl. is found just 
clade. 


position of this Central American species has recently 


outside the Rondeletia s. str. The generic 
been under debate. Lorence (1999) supported its 
position in Rondeletia, but Borhidi (2001a) positioned 
itin Arachnothryx. Our ITS sequence data suggest that 
this species should be treated as a Mosi 
although support for this hypothesis is less than 
Rondeletia inermis and R. pilosa Sw. were not 
included in Fernández Zequeira's (1994) treatment of 
Cuban Rondeletia, as these species occur in Puerto 
Rico and the irgin Islands, respectively. 
However, according to her key to sections, they would 


both belong to Rondeletia sect. Leoninae. In our 
analysis, they form a clade with strong support. 
Rondeletia sect. Leoninae would thus be the only one 
of Fernández Zequeira's sections that is supported by 
our ITS sequence data. 

In all analyses, there is moderate support for a 
clade with Rondeletia alaternoides, R. odorata Jacq., 
and R. pachyphylla Krug & Urb. (Fig. 1, clade F), 
which represent sections Calophyllae, Odoratae, and 
Pedicellares, respectively. Following HE diagnostic 
table of sections in Fernández Zequeira (1994), w 
were unable to find any morphological uic. that 
support this group. 

Rondeletia intermixta Britton and R. ochracea Urb. 
form a clade with strong support. While R. intermixta 
belongs to section Rondeletia M. Fernández Zeq., 
ochracea has not been previously classified to any 
section. It is thus possible to argue that R. ochracea 
should also belong to section Rondeletia. The only 
n Rondeletia 

& 


other known representative of sectio 
. portoricensis 


included in our analysis, 
Urb., is placed in an pone relationship to the R. 
intermixta-R. ochracea clade, although jackknife 
support for this is below 

Our 


separation of sections Hypoleucae and Nipenses in 


study does not show any support for a 
Rondeletia. All sequenced representatives of these 
sections, except R. berteroana DC., are found in a 
strongly supported but unresolved clade (Fig. 1, clade 
H). No morphological character combination seems to 
be unique for these two sections as one group, 
according to the character list in Fernandez Zequeira 
994). Rondeletia berteroana differs from the other 
sequenced species of section Hypoleucae (and section 
Nipenses) in being from Hispaniola. This species is 
found as sister to clade H but with very low support 
ig. 1) 


= 


Rondeletia chamaebuxifolia Griseb., 


quenced representative of section Chamaebuxifoliae, 


the only se- 


is found closely related to the a from sections 
and Nipenses. Following the 


characters provided in Ferná ndez Zequeira 


Hypoleucae diagnostic 
(1994) for 

sections Chamaebuxifoliae, Hypoleucae, and Nipenses, 
this clade (Fig. 1, clade C) could be a 
owered 


rom other sections by having l- to 
inflorescences and retrorse-pilose flowers. 


STEVENSIA 


e second aim of our study was to investigate the 
relationships between Stevensia and Rondeletia. 
Stevensia is here for the first time included in a 
molecular phylogenetic study. According to ITS data, 
there is strong support for an inclusion of at least S. 


minutifolia Alain in Rondeletia s. str. The genus 


Annals of the 


186 


Missouri Botanical Garden 


pog 


OPLZSTAV O00EPZAY OTEOELAV (49) 8383 P 12 vaoy f equ) — x “bez zepuvure “A »qopajopaq “dsqns sapioutom]p viojapuoy 
90£0£LA V (san) $228 P 1» adq “9 ^q equ) sapiowiopp]o "dsqus “Yoly “Y saplowsarnyy pmojopuoy 
(s) 
SBELTSTAV LOOLPZAV  OGZOSLAV  GIZE tug Wyong upe [eoruejog snirag op je pereanno ipiipog (seSopue1g) suaosamuffns smprupopaoy 
16ZOELAV (o81d) 0268 2242107 `q “DELI 1e parano prog (1puog) snsosiys smprupopaoy 
PEZOELAV (AN) SO0G vorm] Y 29 Oas) T T epeuroyenz) "youe[g (Pug) vropaoo n2150] 
Gp 2986 2804) MIS *uopree) peoruejoq [e40y ye 
oSTLZSTAV 666ZPZAV  S8Z0ELAV pawan :9[sdi (go) ggg vossyuposyg 2p uossfo3sno *ejeuropenz) SLI “youl (Pug) vropaoo r2150] 
(s) 
¿PLHZOTAV O00EPZAY  98Z0€ELAWY 961 P 2 $207 “epeuroreno :97sdi (FD) 2868 vixopy OXON “SLI "youe[g puaourm var doy 
OLD 
oL PLESTAV ;£LOPODMV  TOEOSLAV ss [SAY Futuuny *99z-po OLA Vps perdor, proie je peyeanno "wong Cboef) puponaum siponjopy 
£0£0£LAV (Sd) £168 P 12 arasdjag 2) q equ) "qosuz) DIDUOLO) DIJO] Ád 
(49) P&E 
ASLESTAV TIGZPEAY  POS0£LAWV ‘P P vaoy “equo iors (ANNO) £661-989F 2809 L "equi SLI eroidpoq (Puas) t2fpis popzopy 
oOFLZSTAV O86ZPTAV  ZOEOELAV (49) F9GE "JP 1» vaoy T equ) "qi Y Sny (qesua) seproysunporyd popzopy 
,9689rc[V osoqz?) nqgofipupas Dinar] 
EOZOELAV (MSNA) 2993098 yoorng *eipensny "PUN “A 240jfiroao viuosuryspopy 
OLD 
POCOELAV S296 $479 LETS X OLA Topes eordory, prone 1e pereAnpno TPHYPS Y “wey sisuansnin ppapgonz) 
£96ctc4V  ¿I90LZLAV (49) 99g "JP 12 vaog “equy “wey (T) »4qp2s npsoyans 
G6ZOELAV (san) OSEE T9 Ja saweig *xopeno^q ua kag xo [pues sufo nrunsnyozu0y 
PEGTPTAY — L6ZOELAV (S) £607 "I? 12 Daoy T *viquio[or) "ULISÁSIS apiqpioods UOLPUIPOISDSDIALPON) 
NOPZOTAV vPEOPOOAY — 129£S0ZAV "wong (T) Pq» w9209014) 
119€ S0c A V "qun nsouanf pangsa) 
oSELZSTAV 916ZpZAV  2870£LAV (49) ZIZ T? 1? uossfoisnz) euwog "]pueis asuappurayons wnipiunydayg 
86c0£LA V (AN) 80123 Sujoy 9 “openog "oput de xénpyounjoniy 
oS LLZSLAV OL6ZPTAV  9670£LAV (49) 482€ "JP 12 vaoy T “equ "pued (puny) »ydosna] xnpowgopay 
ZOZOELAV (AN) 86€9 v5npaaA "y X opadjaq 79) `q '1openo^q "uuoKojg (PULIS) sisuaovaoquimjo xXnpiouwopay. 
66c0£LA V (S) rpg wmpqpung xy vaoy f *eureueq "poue[q (1pueg) saplotayppng x&nprouopay 
S8ZOELAV (sat) 8188 P 12 arasdjag 2) d equ) "qup smunjoaas smppupukso4oy 
Jua 9rsda SLI erep 1ogyonoA UOX? T, 


Joquinu UOISS9929 Juegue) 


"uoneorqnd snotaord e UIOIJ parto ST oouonbos 19410 ou sso[un Surouenbos 9 rsda pue SLI 409 107 posn st 1oqmonoa owes oq *peuoruour SI I9Y9NOA ouo Ayuo H pesn Ssoouonbos 


[[8 107 $1oqumu uorsso2oe xuequo) se [PM se *1oded Sı} UL peiuoso1d Aq[eursuo Soouonbos 10} UOTJEULIOJUT TOTONOA gysda pu? SLI SuIpnpout *Kpms 9u UL popn[our exe] jo AQEL “TI AQEL 


187 


Rova et al. 


Volume 96, Number 1 


2009 


The Rondeletía Complex 


"peusiqudun *ipupog ^y R uejediuezg "I ‘f Iqexef 74 “ISBUIZOUBNS ^S, Wor USAR] osje stom soouonboc "gooq “TE 19 e^nug, 
{ZOOS “TE 19 PAOY, 5666] “BAOY 29 uossropuy, ‘ZOOS Aeuretg H uosrquipueurgezey. ‘100Z ‘UOS A 79 UBUTUÁOTAL, ‘E007 “Te 19 [[240J9TA, :Sepnpour so2uenbos pousr[qnd 10] o1njexeji[ paro eur, 


LO0ETEAV LA uro snp snruoun y 
00€0€L A V (TWO) E66T-PE87 PROG 1 equ) ‘boz zepugureT [A 3 IPItog (qestro) endanolipopa smpupaoqng 
60€0€LAV (AN) £99€1 4218017 "y ‘oyqndoy ueorunuoq urepy oyofimunu visusioig 
80€0£AV (San) OFSZ P P ateadjag 72) q ‘oyqndoy uvorumuoq urepy oyofimunu visusioig 
6ZEOELAV (SAN) ££88 P 19 adq 2 q "equo prog 19 ‘baz zepuvuray ‘y suaosaupoqns puajopuoy 
OZEOELAV (San 1928 P 19 adq 2 ^q PMY prog 19 ‘bez zepuvuray ‘y suaosaupoqns puajopuoy 
6TSOELAV (Sdn) LPL T9 P opadjoq “9 q eoremer eonaq (T) supyndus vusjopuoy 
BTEOELAV (SAN) S061 puojog-o:90407) siopeno oon pand uspopuog 

EPLESLAV SIO£PGIV — SEEOELAV (OIN) 28917 401401 "jj 7) “ony ouong "qi Y Bury sisusoniond oyojopuoy 
8T£0€LAV (Sdn) 9728 P P atada 9 q equi) "qun vynwwoyd ousjopuoy 
68Z0EL AV (AN) 996€T 1213077 “ejoruedstyy ueu p quo pupoand ouojpopuoy 

sbELZGDAV PIOEPZAV  ZELOLLAV (AN) 988% zenSupoy-opeasoy `q ‘spurys] WBA "en "mg osopid puajopuoy 
¿TEOELAV (Sdn) 7298 P P aradeg 2 q emo op dkpond -dsqns ‘qın op Sur o tydkond vuspopuog 

ipipog 29 ‘boz zopueuroy IA 
TIOSPZAY  TESOELAV (49) ZEEE T? 19 vaoy f “equ sapioj nau -deqns ‘qan 9 Su» y dtood vuojopuoy 
(Sdn) posg wosnopuy 

oTRLZSTAV OTOEPZAV  ZOS0£LAV qp uag Wong upes jeonrejog snog eq qe pereAnqno ‘boef »1n40po vrojapuoy 
OTEOELAV (AN) 8199 P 1» sougogr 7) 4 oyendoy ueorumuoq "quf eaonatoo vriojopuog 

STEOELAV (San 0228 `P 1» erdjaq 2 d "equi sisuapour "deqns sisuadru miyojapuoy 

OESOELAV (Sdn) 7598 P P aradeg 9 q equ) "qun sisuadru mizajapuoy 

600€bcIV — TICOgLAV (89) 41&& P 19 Daoy T equ) ipipog 29 ‘boz zopuputoy "Ji sisuosogfoanu prajopuoy 

ETSOELAV (49) 9138 T? P vaoy f “equ "qup sisuouo] vuajapuoy 

scbLCGLAV — LLOPO0dV  TISOELAV (d9) SEE e 19 vay T “eqny pixquuana eqns uopug DIXTULLSTUL rajopuoy 

(AN) [692 `P 19 zondupoy-opoa2oy 

oS PLESTAV GIOEPZAY  STSOELAV ‘ONY ouenq :97sdi (DEA) 9628 2249407 `Q OQLA W PANNY ‘SLI "qu o Sur (Suerds) smaour vuojapuoy 
OZEOELAV (AN) FOT "?pAvH A “W'S Y puquy H f “eureueg uep£eg ^A "I 3g Anq vyofinjauny vuojopuoy 

S9OLOBLIV —— S0£0gLAV (AN) O9ST OJJHSDD [f 'epeureyenz "pues (ws nuog) mupep onajapuoy 

PZEOELAV (Sdn) 6242 P 19 atesdjag 9 ‘q “vowel "ds puojopuoy 

EZEOELAV (Sdn) £094 P P orudjoq 9 q eoremer "qOSLIS DIDULO viajopuox 

LZEOELAV (S) 9T-£0-S66I "ws ‘P 12 oppyonyy “equ "qosuz) vijofixngooumo vuajopuox 

ZZEOELAV (SAM) 2982 P P adq `) "q-'oqqndey uvorumuoq "DA Punosa aq Dipojopuoyg 

TZEOELAV (Sd) FESZ 7 P eradpoq 9 "q-'oqqndey uvorumuoq "qu SisuauoYyoing viopopuoyg 

LODEPZAV —- TLEOELAV (49) EFEE P 19 nao T equ) "qur emgnoido vuajapuoy 

A ua 9rsda SLI erep IAMO A uoxe[ 


Toqumu uoIS$929e Jue guos) 


ponunuog 'q Oqe, 


188 Annals of the 
Missouri Botanical Garden 


100 => ba alba * OUTGROUP 
100 == calesiasa: la 
Rondeletia pitrèana 
88 99 s cordata * : 
91 (xy Rogiera cordata - 
- 400 Ra amoena * ; 
- Arachnothryx chimboracensis 
Arachnothry: i 
Arachnothryx | lla * ? 
E Cuatrecasasiodendron spectabile * & 
Arachnothny buddleioides 2 
mn $ 
Guett arda s cabra e 
Guettarda uruguensis F 
odgkinsonia ovatiflora 
Timonius nitidus 
Blepharidium guatemalense * 
Rovaeanthus strigosus 
Rachicallis america 
Mazaea phialanthoides * 
Phyllomelia corona 
azaea shaferi * 
Rondeletia d ii* unknown 
Rondeletia hameliifolia Calophyllae 
Rondeleti rdi Calophyllae 
Rondeletia alaternoides Calophyllae 
Rondeletia alaternoides * Calophyll 
Rondeletia odorata " oratae 
Rondeletia pachyphylla * dicell 
Rondeletia pachyphyll zm Pedicellares 
Rondeletia portoricensis * 3 Rondeletia 
Rondeletia inter a Rondeletia 
Rondeletia ochracea EI unknown 
Rondeletia inermis * 8 Leoninae 
Rondeletia pi x p Leoninae 
ondeletia stipularis i 3 unknown 
Rondeletia sp. Jamaica 2 unknown 
ndeletia barahonensis e| unknown 
Rondeletia cincta $| Rigidae 
tevensia minutifolia 
Stevensia minutifolia 
Rondeletia ch k li: Chamaebuxifoliae 
Rondeletia berteroana Hypoleu 
Rondeleti [s ees 
Nipenses 
Rondeletia nipensis Nipenses 
Rondeletia nipensis Nipenses 
Rondeletia lomensis i 
Rondeletia miraflorensis * lypol 
Rondeletia apiculata * Hypoleucae 
Rondeletia plicatula Hypoleucae 


e l. Tree compiled E the strict consensus trees from the two s m LE Without and with mele coded, 
as not found i nl 


indica te p cla E Ps siga in the text: —A. Clade comprising representatives of the tribe eee —B. Clade 
d C 


in Rova et al. (2 T D. Clade c es ae Mazae, Phyllomelia, and Rachicallis. —E. The Rondeletia s. str. clade. —F. 
Clade including representatives of se s Odora o, Pedicellares, and the paraphyletic section Calophyllae. —G. Clade 
distinguished by having l- to A ee andl Ee eal oo flowers (corresponding to sections 
Chamaebuxifoliae, Hypoleucae, and Nip enses). —H. Clade T. n representatives of sections Hypoleucae and 

Nipenses. Taxa marked by an asteris k (5 are included in the combined ITS, ns and irnL-F analysis presented in Figure 2. 
The circled Roman numerals I-XI on bold branches refer to ne in the combined ITS, rps16, and trnL-F analysis (Fig. 2). 


Volume 96, Number 1 
2009 


Rova et al. 
The Rondeletía Complex 


189 


Strict consensus tree of the 12 equally parsimonious trees fr 


g 
irnL-F data matrix. Guetta; 


assigned to section based 
sectional assignments). 


Stevensia comprises 11 species endemic to Hispa- 
niola. It is recognized by triangular stipules connected 

illary fl 
calyx lobes, five to seven stamens attached in the 


to a sheath, solitary and axil ers, two to three 
corolla throat, glabrous style, and ovoid to oblong 
seeds (Borhidi, 


character states are also foun 
ccording to 


001b). However, several of these 

within Rondeletia s. 

Fernandez Zequeira (1994). W 

therefore suggest that Stevensia should be included 

within Rondeletia s. str., pending future studies with 
an extended sampling. 


GUETTARDEAE AND RONDELETIEAE 


Our third aim was to compare a nuclear ITS 
phylogeny of the Rondeletieae with the results from a 


previous trnl-F chloroplast DNA (cpDNA) study 


rda scabra, Cuatrecasasiodendron spe ectabile, Rondeletia deamii, R. pac. 


Chiococca alba 

Rogiera cordata 

Rogiera amoena 
Blepharidium guatemalense 
Guettarda scabra 
Rovaeanthus suffrutescens 
Arachnothryx leucophylla 
Cuatrecasasiodendron spectabile 
Rachicallis americana 
Mazaea phialanthoides 
Mazaea shaferi 

Rondeletia deamii 
Rondeletia alaternoides 
Rondeletia odorata 
Rondeletia pachyphylla 
Rondeletia intermixta 
Rondeletia inermis 
Rondeletia pilosa 
Rondeletia portoricensis 
Rondeletia miraflorensis 
Rondeletia apiculata 


unknown 
Calophyllae 
Odoratae 
Pedicellares 
Rondeletia 
Leoninae 


Hypoleucae 
Hypoleucae 


rom the analysis of the combined ITS, rps16, and 
ach 


deletia eee according to Fernandez Zequeira (1994), is 
n Fernández Zequeira (1994), but have been 


on her key. Rondeletia eum is not assigned to a section (cf. unknown among the 


(Rova et al., 2002). The fourth aim was to see if ITS 
sequence data would place Hodgkinsonia in Guettar- 
deae or elsewhere 
Acrosynanthus revolutus Urb. and Rondeletia pi- 
treana Urb. & Ekman (not classified to section) 
appear as early diversified lineages in the oe 
l anation f 


sequence in its entirety. Because of this, it is about 40 
bases shorter than the other Rondeletia sequences. 
Another possibility is that R. pitreana does not belong 
to Rondeletia. In any case, further studies are needed 
tion of 


to solve the position of a pitr The pos 
thus 


Acrosynan Urb. in a m differs 
markedly from the d of the trnZ-F study from 
Rova et al. (2002), where Acrosynanthus was found in 
a position equivalent to basal in clade C (Fig. 1). A 


ea 


Sectional assignment of Rondeletia species, ieu to Fernández Zequeira (1994), is listed in the rightmost column. 


Rondeletia hameliifolia, R. purdiei, R. inermis 
section based on her key (1994: 106). 


and R. pilosa were not listed by Fernández Zequeira, but have been assigned to 


190 


Annals of the 
Missouri Botanical Garden 


possible explanation would be that Acrosynanthus is 
not monophyletic: A. latifolius Standl. was included in 

e trnL-F study, 
the ITS analysis. However, the possible paraphyly of 


. revolutus was sequenced in 


Acrosynanthus must be left to another study when 
more material of this genus is availa 

e found that the well-supported D onse s.l. 
clades in the traL-F stud 
from Rova et al. (2002) have only weak support from 


and Rondeletieae s. str. 


£e 


ITS sequence data. 


With regard to the tribe Guettardeae, po i 
moderate support for a clade includin achno 
thryx, Gonzalagunia, m. Guet- 


tarda, Hodgkinsonia, and Timonius (Fig. 1, clade A). 
According to our results, Cuatrecasasiodendron should 
be synonymized with Arachnothryx, and this is also 
morphologically supported (see taxonomic treatment 
showed 
ogiera amoena Planc 
Planch. as members of the Guettardeae, the inclusion 
of Rogiera s. str. in the Guettardeae clade is not 
supported by ITS data alone. In the combined analysis 
(Fig. 2), Rogiera is found within Guettardeae, while 
Arachnothryx is found to be more closely related to 
Rondeletieae. 
In a recent molecular phylogenetic o Achille et 
l. (2006) supported the monophyly of the Guettar- 
a as oed here, although they do that 
Guettarda, Antirhea Comm. ex Juss., and Stenostomum 
aertn. are polyphyletic. However, more genera 
and more species need to be included in the study in 
order to test the monophyly and delimitation of this 


ough Rondeletieae sensu Rova et al. (2002) is 
recognized by a in the consensus trees (Fig. 1, 
clade B), there is no jackknife support for this clade. 
In both heuristic ITS searches, the genera Blephar- 
idium Standl. and Suberanthus were found basal in the 
Rondeletieae s. str. clade, b 
jackknife support for this. 


ut again, there is no 
However, this position 
corresponds to the results from the trnL-F study of 
Rova et al. (2002) 

The ITS data place Rovaeanthus strigosus (Benth.) 
Borhidi in the Rondeletieae, as sister taxon to R. 
suffrutescens (Brandegee) Borhidi. Just as in the trnL-F 
study (Rora T al , 2002), th ES us ies that R. 

g to the tr. (although 
ihis is coairadicied d in the combined analysis where R. 


suffrutescens is found as sister to Guettarda). In any 
case, R. suffrutescens always appears in a separate 
position from Rogiera, and our study thus supports the 
transfer of these species from Rogiera into a new genus 
as proposed by Borhidi et al. 

here is strong support for a close relationship 
between Rachicallis DC., Mazaea Krug & Urb., and 


Phyllomelia Griseb. (D in Fig. 1). Rachicallis and 


comprises two species. Based on 
argue that all three genera should be merged together. 
However, both Mazaea and Phyllomelia are easily 
distinguished by the peculiar fruit (pseudosamara, 
1999b) and 
prete, 1999b), and for this reason we prefer to regard 


sensu Delprete, calyx morphology (Del- 


them as separate genera. 


HODGKINSONIA 


The fourth aim was to see if ITS sequence data 
would place Hodgkinsonia in Guettardeae or else- 
where. Our study undoubtedly places Hodgkinsonia 
close to Timonius, which means within Guettardeae. 
This position is in accordance with the view of 
Mueller (1861) in the original description and Bremer 
1992), but contradicts the supposition of Delprete 
1996), who tentatively included the genus in tribe 
Chiococceae in agreement with Robbrecht (1988). 


— 


CONCLUSION 

The ITS sequence data support only one of 
Fernández Zequeira's (1994) Rondeletia sections as 
monophyletic: section Leoninae. Rondeletia sections 
Calophyllae and Rondeletia are paraphyletic accord- 
ear in mind 


that we were not able to sequence more than one 


ing to our analysis. However, one should 


species from several sections. When we compare our 
ITS phylogeny with the character lists in Fernandez 
Zequeira’s treatment of Rondeletia, we were unable to 
find morphological characters that correspond with 
our phylogenies. The sections described by Fernandez 
Zequeira are often defined by various combinations of 
overlapping character states, which makes compari- 
sons difficult. The only exception is a clade including 
| from Rondeletia sections Chamaebux- 

e, Hypoleucae, and Nipenses, which could be 
1- to 3-flowered inflorescenc- 
This clade could 


potentially be recognized as one section. 


olia 
n by having 
and retrorse-pilose flowers. 
Rondeletia s. str. (i.e., excluding Arachnothryx, 
Javorkaea, Rogiera, Roigella, Rovaeanthus, and Sub- 
eranthus) has strong support, although some species 
need to be further investigated for their generic 
affinity (e.g., R. pitreana and R. deamii). 

n important result from our study is that Stevensia 
minutifolia is included within Rondeletia s. str. 
reevaluation of the morphological characters in 
Rondeletia (including Stevensia) based on the results 
rom ITS and other sequence data is certainly needed. 
The present analysis clearly suggests that Stevensia 
should be recognized at most as a section of Rondeletia. 


Volume 96, Number 1 
2009 


Rova et al. 
The Rondeletía Complex 


191 


here is strong support for a division of the 
Rondeletieae-Guettardeae complex into the tribes 
i trnL-F 
data (Rova et al., 2002), but only weak support from 
ITS data. While part of the Guettardeae has moderate 
support, support for Rondeletieae in the sense of Rova 

et al. (2002) is weak in the ITS study, although the 
Rondeletieae s. str. 


Rondeletieae s. str. and Guettardeae s.l. from 


clade is found in the strict 
consensus trees of all of our AM both including 
and excluding indel codings. It was not possible to 
compile a sufficiently large ids set in order to test the 
delimitations of Guettardeae and Rondeletieae using a 
combined ITS, rps/6, and trnl-F sequence data 
matrix; however, we consider that the phylogenies 
available up to this point (Rova, 1999; Rova et al., 
2002; Delprete & Cortés-B., 2004; the present study) 
provide sufficient support f a re-delimitation of the 
tribe Rondeletieae. 

Based on the results from ITS sequence data, we 
also reconsider Delprete's (1996) tentative inclusion 
(based on morphology) of Hodgkinsonia in the Chio- 
cocceae, since the present ITS sequence data support 
Bremer's (1992) conclusion (also based on morpho- 
logical data) that Hodgkinsonia is part of the tribe 
Guettardeae. 


Taxonomic TREATMENT 


Based on the results from the present and other 
recent studies wae 1322p; Rova, 1999; Rova et 
002; Delprete & Cortés-B., 2004; Borhidi et al., 
2004), we propose the llos taxonomic dessin. 
tions and rearrangements. 


Tribe Rondeletieae (DC.) Miq., Flora Nederl. Indié 
30, 156. 1856. Rondeletiinae DC., r. 4: 

; s subtribe *Ro OS i 
tribe Hedyotideae. Rondeletieae DC. ex Rehb., 
Der Deutsche Botaniker 1: 77. 1841, stat. non 
indic. TYPE: Rondeletia L 


Shrubs or trees; wood whitish or yellowish; raphides 
absent; axillary thorns absent. Stipules free or connate 
at base, mostly entire, rarely bifid, mostly interpetio- 
lar, frequently with colleters on the adaxial side 
secreting resinous compounds, persistent to readily 
caducous; leaves opposite or verticillate, decussate, 
ssi a blades chart hi 


coriaceous; domatia variably present or absent. 


petiolate to se aceous to thick- 
Inflorescences terminal or ba cymose, panicu- 


late or thyrsoid, multiflorous or pauciflorous, or 
Flo h KR 


uniflorous ermaphroditic, mostly actino- 
morphic, (3- to)4- to 6-merous; calyx persistent or 
caducous; lobes often minute, sometimes foliose; 
calycophylls commonly absent or pterophyllous (green 


to greenish white), with all calyx lobes expanding into 


a rotate pterophyll after anthesis and present in all 
flowers in Phyllomelia; corolla hypocrateriform or 
narrowly infundibuliform, orifice with annular thick- 
ening, white, cream-white, red, green, or yellow, 
membranous to fleshy; aestivation valvate, contorted, 
or imbricate; stamens mostly as many as corolla lobes, 


the orifice of corolla tube; 
exserted, oblong to narrowly elliptic to button-shaped, 
2-locular, opening by Vide slits, dorsifixed 
; pollen 

exine 


ear the base or around the middle, introrse 
released as monads, e or colporate, 
reticulate or foveolate (not E So branches 
present, with sti tic surfac rucate; 
ovary inferior (half inferior in pones bct 
with a few to many ovules (1 to 2 in Mazaea) per 
locule attached to a central placenta, or exceptionally 
one ovule per locule basally attached (Phyllomelia). 
Fruits woody capsules, loculicidal or septicidal, or 
septicidal and loculicidal contemporaneously (Ble- 
pharidium, Mazaea), commonly dehiscing basipetally, 
i (Phyllo- 


melia); placenta central, rarely apically incomplete, or 


or exceptionally pseudosamaras, indehiscent 
shortly stalked; seeds horizontal, imbricate, peltate, 


convex, 
(Blepharidium, Mazaea), or basally inserted, ellip- 
soid-ovoid and fleshy (Phyllomelia). 


Genera included: Acrosynanthus, Acunaeanthus 


Rondeletia, 
Stevensia, 


melia, Rachicallis, Rogiera, Roigella, 
Rovaeanthus, Spathichlamys R. Parker, 
Suberanthus. 

The description and delimitation of the Rondele- 
tieae here POE are based on the results of the 
present study run with those of Rova 
1999) and 
basically a reduction 
(1999a), based on his wide circumscription of the 


— 


Rova et al. The description is 
d 


of m proposed by Delprete 


tribe to include the Condamineeae and the Sipaneeae, 
which was produced primarily for the floristic 
treatment and not based on a comprehensive phylo- 
genetic analysis. 

Rova (1999) and Rova et al. (2002) demonstrated 
that the Condamineeae (except the subtribe Portland- 
iinae, which belongs to the Chiococceae s.l.) shoul 
be transferred to the subfamily Ixoroideae, in a 
complex also including the Calycophylleae and the 
Hippotideae (more studies are needed to re-delimit 
these groups; Kainulainen & Bremer, unpublished). 
B. (2004) and Rova et al. (2002) 


also demonstrated that the Sipaneeae belongs to the 


Delprete and Cortés- 


subfamily Ixoroideae and is a monophyletic group that 
was positioned in the same clade as the tribes 


192 


Annals of the 
Missouri Botanical Garden 


Henriquezieae and Posoquerieae in their phylogenetic 
Er a et al., 2004 

The typic genus Rachicallis, endemic to the 
daa Basin, is added (not included in the tribe 
by Delprete, 1999a) to the present delimitation of the 
Rondeletieae, which was placed close to this tribe in 
Bre et a shown to belong t 
Rondeleti tieae d Rova et al. (2002) and in the present 
study. 
As a result of this study, Stevensia is perhaps best 
treated as synonymous with Rondeletia, because in the 
phylogenies obtained it is positioned within the 
Rondeletia. However, as only one species of Stevensia 
(S. minutifolia) was included in the analysis, we 
refrain from proposing the necessary new combina- 


Taxa TRANSFERRED TO THE TRIBE GUETTARDEAE 


Steyermark (1964) positioned Cuatrecasasioden- 
dron in the Rondeletieae because of its foliaceous 
calyx lobes, capsular fruits, horizontal seeds, ovary 
with many ovules in each locule, and corolla with 
imbricate lobes. At the same time, he treated it as 
closely related to Rondeletia because of the corolla 
lobes being subzygomorphic, as the most interior lobe 
is more pubescent Boss than the external ones, 
and glabrous or alm labrous externally, while the 
others are iine pam This genus was 
maintained in the Rondeletieae by Delprete (19992) 
because of the same characters as used by Steyer- 
mark. However, in the phylogenies produced in the 
present study, Cuatrecasasiodendron was found within 
the Arachnothryx clade of the tribe Guettardeae, and 
the two taxa are treated here as synonym 

In addition, a detailed analysis of the two species of 
Cuatrecasasiodendron described by Steyermark was 
undertaken. (1964) distinguished C. 
its leaf blades hirsute below (vs. adpressed-pilose to 
arachnoid-pubescent below), shorter petioles, shorter 
long (vs. ca. 28 mm 


sa corollas 17-20 mm ong, 
r 


pubescence), and longer and more 
R inflorescence branches among other 
characters. A comparison of the type specimens with 
recent collections revealed that the characters used by 
Steyermark to separate the two taxa fall into a 
morphologic (and geographic) gradient. 

The types of both taxa of Cuatrecasasiodendron 
were collected in the Valle 
(Colombia); however, 


elevation of the coastal region, while C. colombianum 


del Cauca Department 
C. spectabile is from a low 


intermediate characteristics. Therefore, the two spe- 


cies are treated here as synonymous to one another, 
and only one new combination in Arachnothryx is 
necessa 


Arachnothryx Planch., Fl. Serres Jard. Eur. 5: 442. 
1849. TYPE: Arachnothryx leucophylla (Kunth) 


Planch. (= Rondeletia leucophylla Kunt! 
b e erre a & E Acta Biol. Venez. 
4: 29. 4. Syn. YPE: Cuatrecasasiodendron 


as Standl. & con 


Arachnothryx spectabilis (Steyerm.) Rova, Delprete 
& remer, comb. nov. Basionym: Cuatrecasa- 
siodendron spectabile Steyerm., Acta Biol. Venez. 
4: 33. 1964. TYPE: Colombia. Valle del Cauca 
Department: Costa del Pacifico, Río Cajambre, 
Barco, 5-80 m, 21-30 Apr. 1944. (fL), J. Cua- 
trecasas 17165 (holotype, USt; isotype, VEN!). 


Cuatrecasasiodendron d cs Standl. & Steyerm., Acta 
Biol. Venez. 4: 30. 

Valle del Cauc 

um Occidental, Ls del Río Achicayá, Queb- 

El Retro, 300 19 Dec. 1942 (fL) J. 

Dane 13694 eos F!; isotype, US!). 


Additional specimen examined. COLOMBIA. Depto. 
Valle del Cauca: Mun. Bu rd. e- 2s 
Anchicayá, Km 35, ca. 03°37'N, 76753" W. 

Apr. 1994 (fl), J. E E dina. L. Andersson, C. p sen E 
C. Persson 2093 (GB). 


enaventura, 


3 


Literature Cited 


Achille, F., T. J. Motley, P. P. Lowry II & J. Jérémie. 2006. 
Polyphyly in Guettarda L. (Rubiaceae, Guettardeae) based 
on nrDNA ITS sequence data. Ann. Missouri Bot. Gard. 
93: 103-121. 


Andersson, L. & J. H. E. Rova. 1999. The rps16 ow E 
the nor of the Rubioideae (Rubiaceae). PI. 
Evol. 214: 161-186. 

Borhidi, h 9o Studies in Rondeletieae (Rubiaceae). III. 
The genera Rogiera and Arachnothryx. Acta Bot. Acad. 
Sci. Hun, x 28: 65-11. 

. 1989. Studies on Rondeletieae (Rubiaceae). XI. 
en notes on some Central American species of 
Rondeletia s.l. Acta Bot. Hung. 35: 309-312 

—————. 1993-1994 [1994]. Studies in Rondelen tieae. XII. 
New combinations of Eon and Central American taxa. 
Acta Bot. Hung. 38: 1 2. 

. 2001a y dd corrections to the *Nomen- 

diss of Mexican and Central American Rubiaceae” of D. 
H. Lorence. Acta Bot. Hung. 43: 37-78. 

. Revisión taxonómica del género Stevensia 

Poit. Acta Bot. Hung. 43: 287-298. 

& Fernández d nue 1981a. Studies in 

w genus: Roigella. Acta 


e 
=] 
par 


Rondeletieae A ce LAn 

ot. o x 309— 312. 

b. Studies in Rondeletieae rra 
e). II. A new gen Suberanthus. Acta Bot. Acad. Sci 

a e 313-316 

M. Jarai- Komlódi. 1983. Studies in Rondeletieae 
m ie IV. A new genus: Javorkaea. Acta Bot. Hung 
9: 13-27. 


Na 


Volume 96, Number 1 
2009 


Rova et al. 
The Rondeletia Complex 


193 


arók, M. Koksis, Sz. Stranczinger & F. 
Kaposvári. 2. El Rondeleia complejo en México. Acta 
Bot. Hung. 46: 91-135. 

Bremer, B. 1992. Phylogeny of the Rubiaceae (Chiococceae) 
based on molecular and morphological data—Useful 
(is e for classification and comparative ecology. 

n. Missouri Bot. Gard. 79: 380—387. 

—— € M. Thulin. 1998. Collapse he Isertieae, 
establishment of Mussaendeae, and new genus of 
Sabiceeae urs Phylogenetic relationships based 
on rbcL data. 1: 71-92. 

, K. oe : . 1995. pepe 

and tribal relationships in the Rubiaceae based 

rbcL sequence data. Ann. Missouri Bot. Gard. 82: 383- 

39 


re- 


Delprelés P. G: 


ceae, 


996. Evaluation of the t 
Co us eeae, 


ribes Chio 
and Catesbaeeae (Rubiaceae) baad 
on E e characters. Opera Bot. Belg. 7: 165-192. 
Rondeletieae (Rubiaceae)—Part I. Fl. Neo- 
trop. Ven 77: 1-226. 
— ———. 1999b. Morphological and taxonomical comparison 
of the n endemic taxa Ariadne, Mazaea, Acu- 
naeanihus, Phyllomelia (Rubiaceae, Rondeletisse) and 


santhe, with one new combination. Brittonia 51: 
230 


s-B. 2004. phylogenetic study of the 

"m" nba o a using irnL-F and 
ITS sequence data. Taxon 53: 347—356. 

— ——, L. B. Smith & R. B. Klein. 2004. [Description t v 

tribe Posoquerieae] Rubiáceas, Volume 1— 

G: 1. Alse eis até 19. 


&neros de A— 
Galium. (com observações cages 
, A. Reis & O. Iza}. Pp. 1-344 in A. Rei 
(editor), Flora Ilustrada Catarinense. Herbário Barbosa 
es, Itajaí, Santa Catarina, Brazil. 
&J L 1987. A rapid DNA isolation 
r "ei m of fresh leaf tissue. 
Phytochem. Bull. 19: 11-15. 

Fernández aus de . 1993-1994, [1994]. Estudio taxonó- 
mico de ero Ronde letia L. s.l. (Rubiaceae) en Cuba. 
Acta Bot. Hung 38: 47-138. 

: New species and combinations in 
Mexican and Central American Rondeletia (Rubiaceae). 
Novon 1: 135-157. 
——. 1999. A Nomenclator of Mexican 
Andria D PS Mongr. Syst. 
Gard. 73: 


and Cen 
Bot. Missouri E 


McDowell, T., M. Volovsek & P. Manos. 2003. Biogeography 


of Exostema (Rubiaceae) in the Caribbean ra in E 
of Bot. 28: 


molecular phylogenetic analyses. Sys 
431-441. 
Moynihan, J. € N 2001. Phylogsography, p 
allies, and nomenclature of Caribbean endem 


Neolaugeria (Rubiaceae) ol on TTS s sequences. bs i 

PL a 162: 393-401. 

, F. 1861. Fragmenta Phytographiae Australicae 2. 

Jo ar ae Melbourn 

Popp, B. Oxelman. 2001. Inferring the history of the 

S WR Silene aegaea (Caryophyllaceae) using plastid 

and homoeologous nuclear DNA sequences. Molec 
4481. 

. & B. Bremer. 2002. Phylogeny and 
classification of Nanelegas s.l. (Rubiaceae) inferred from 
molecular a T E ou F) and morphological data. 

Amer. J. B 

Robbrecht, E om. py SS Rubiaceae. Opera Bot. 
Belg. 1 

———. ET Bon Supplement to the 1988 classification 
of the Rubiaceae. Index to genera. Pp. 173-196 in E. 

Robbrecht (editor), Advances in Rubiaceae Maja 

matics. Opera Bot. Belg. Vol. 6. 

Reva, J. a E. 1999. US o MM CM 

(R ral Dissertation, Bot 
ma Institute, ii mars of Eu Góteb hae Sweden. 

. Delprete, L. Andersson & V. A. Albert. 2002. 

A iml-F oo sequence study of the pu 

wit h 


1 


implica 

ds ane of he Rubiaceae. Aimer. J. Bot. 30: 

Standley, P. C. 
Steyermark, 
adr o de Qüstyécnsus. Acta Biol. Venez. 4: 1-117. 

7. Rondeletia an ee d Pp. um 261 in 

aguire & collaborators (editors), TI y of the 

Gans Highland, Part VII. Mem. New York p Gard. 


v, J. W. Kadereit, A.-R. Pepper, T. J. 
. H. E. Rova, K. Potgeiter & V. A. 

cane 1998. Sacalm [aeei Lacoa an endemic of 

a de la Br. elan border. 

is p to a temperate alpine lineage of o 

Harvard Pap. Bot. 3: 199-214. 

Swofford, D. L. 2002. PAUP*: Phylogenetic Analysis Using 
Parsimony (*a and Other Methods), Version 4. Sinauer 
Associates, Sunderland, Massachusetts. 


PHYLOGENY OF TRICALYSIA 
(RUBIACEAE) AND ITS 
RELATIONSHIPS WITH ALLIED 
GENERA BASED ON PLASTID 
DNA DATA: RESURRECTION OF 
THE GENUS EMPOGONA! 


James Tosh, Aaron P. Davis,” Steven Dessein,* 
Petra De Block,* Suzy Huysmans,” Mike F. Fay? 
Erik Smets,?? and Elmar Robbrecht* 


ABSTRACT 


Recent studies on the circumscription of the tribe Coffeeae (Rubiaceae) revealed a weakly supported clade d 


ra Argocoffeopsis Lebrun, Calycosiphonia Pierre ex Robbr., 
Diplospora DC., Discospermum Dalzell, Nostolachma T. Durand, and 1 


Belo nophora 
onnea Pierre ex Pit. The phyl 


Tricalysia and these allied taxa are investigated further using pow dara kom four plastid regions (ale intron and 
ali PeiD). Rob! 


rpLI6 intr ron, accD- 


brecht 


is not monophyle let 


b AA is ie sister to the ae Diplospora and is recognized at the generic level. The 34 n 


E. ac 

ct m. ns e Robbr., E. 2 (Robbr.) J. 
(Hiern) J. Tosh & 
buxifolia subsp. australis (Robbr.) J. Tosh & Robbr., 
bbr., E. coriacea (Sond 


.) J. Tosh € Robbr., E. crepiniana (De Wi 


«pou: 
The genus name Tricalysia should be uy to taxa from subgenus Ses dE Uu dn. Empogona 


cessary new combinations for 


idophylla ae ) J. Tosh & i E. ne (Robbr. y J. Tosh El Robbr., E. africana 
osh & Robbr., E. bequae 
Robbr., E. breteleri (Robbr.) J. D & Robbr., E buxifolia (Hiern) J. Tosh & Robbr. subsp. buxifolia, E. 


E. cacondensis (Hi 


ii (De Wild.) J. Tosh €: Robbr., E. br 


acieata 


iern) J. Tosh & Robbr., E. concolor (N. Hallé) J. Tosh & 
Id. & T. Durand) J. Tosh & Robbr., E. deightonii (Brenan) J. 


Tosh € Robbr., E. discolor (Brenan) J. Tosh € Robbr., E. d d ae (De d qu subsp. filiformistipulata, E. 


ds cda subsp. epipsila a JJ. Tosh x Robbr., E. glabra (K. Sch 
J. Tosh € R 


E. kirkii Hoo 
i E D. & m R^ e oe (Bridson 
., E. nogueirae (R 


ae iv. a Tosh $ Robbr., E. ovalifolia var. taylorii (S. Moo: 


Xa. var. ivorensis (Ro bbr.) J. Tosh & Robbr., E. BA eer J. Tosh & 
and £E. e eis de Schum.) J. Tos 
Empog: 


& Robbr., 


Coffea, coffee, 


E. refl 
1 E. e Nd Tosh 
ords: sal, Coffeeae, 

^u 


Tricalysia. MUS 


A. E. van Wyk) J. 
obbr.) J. Tosh € Robbr., E. ovalifolia ee a 
re) J. T. 


m.) J. Tosh € Robbr., E. gossweileri (S. Mrd 
obbr., E. TM [AR ae s To. sh&R obbr., E. macrophylla 
a & Robbr., E. ngalaensis (Robbr) J. Tosh & 

Robbr. var. ovalifolia, E. rn var. glabrata 
E , E. reflexa (Hutch.) J. Tosh € Robbr. var. reflexa, 
A E. somaliensis (Robbr.) J. Tosh & 
rpll6, Rubiaceae, 


& Robbr. 
a, molecular e petD, 


The genus Tricalysia A. Rich. is one of the largest 
genera of Rubiaceae in Africa and occurs in 
continental de (ca. 95 species), lr. (12 


species), a moros (one s 


[e] 
E 


e Com pec n 
RN is possesses the distingu iling peek d 
of the tribe Coffeeae (Bridson & Verdcourt, 2003; 
2007). These include axillary inflores- 
cences paired at the nodes with obvious calyculi, 


Davis et al., 


flowers with left contorted corolla aestivation and a 
distinetly 2-lobed style, and relatively small and few- 
seeded fleshy fruits. Most Tricalysia species can be 


separated readily from other Coffeeae by the presence 
of stipules with needlelike awns, truncate to distinctly 
obed calyces, and seeds with a shallow hilum. 
Identification of Tricalysia at the spec 
notoriously difficult, as the genus contains 
number of species across a broad geographic and 
ecologic range, often separated by minor and 
continuous characters. 

In a series of papers, Robbrecht (1978, 1979, 1982, 
1983, 1987) conduc 


Tricalysia, with later contributions by Ali and 


a taxonomic revision of 


‘James Tosh would like to acknowledge all members of the conservation genetics, molecular systematic, and Rubiaceae 


aa groups at the 


authors would also like to 


ne aes Gardens, Kew, who provided help and O during my research visit in 2006. The 
viewers of the paper for their helpful e 


ments and suggestions. This P ia was 


est 
ae financially by i from the Fund for Scientific Research- En (FWO, G.0250.05 and G.0268 
? Laboratory of Plant Systematics, Ses Universiteit Leuven, Kasteelpark Arenberg 31, P.O. Box Di BE- 3001 
-be. 


E Belgium. Corresponding au 


r: james.tosh@bio.kuleu 


oyal Botanic Gardens, Kew, Rich ag Surrey, TW9 3AE, United Kingdom. 


Nil Botanic Garden of Belgium, Domein va 


n Bouchout, BE-1860 Meise, Belgium. 


Herbarium of The Netherlands, Leiden University Branch, P.O. Box 9514, NL- 2300 RA Leiden, The Netherlands. 


nal 
b: 10. 3417/2006202 


ANN. Missouni Bor. Garp. 96: 194—213. PUBLISHED ON 23 APRIL 2009. 


Volume 96, Number 1 


Tosh e 
B of Tricalysia 


195 


ie) (1991) and Ranarivelo-Randriamboavonjy 

l. (2007). Robbrecht (1979, 1982, 1983, 1987) 
ide and revised two subgenera: subgenera 
Tricalysia A. Rich with five sections (Probletostemon 


Robbr., Tricalysia, Rosea (Klotzsch) 


a 


and an unname 
Madagascan section) and subgenus Empogona (Hook. 
f.) Robbr. with two sections (Empogona Hook. f. and 
ae de Robbr.). Separation of the two subgenera 
support 

corolla throat 


calysia was rted by differences in cal 


pubescence, fruit 


color, and the presence/absence of a sterile append- 


vu morphology, 


age on the anther connective. 

Empogona Hook. f. was originally recognized at the 
generic level by Hooker (1873) based on a single 
Zambezian species, E. kirkii Hook. f. Brenan (1947) 
reduced the genus Empogona to a section of 
Tricalysia, containing six mainly eastern and south- 
n species. During his revision of Tricalysia 
in particular, his treatment subgenus 
other tropical African species, them with 

eo- oo distribution, n preme to this 
ae s. 
Robbrecht (1978) also investigated the or 
related genus Neorosea é, Ward of 1 


ncluded in 
Tricalysia. Two of these 17 species, including the type 


. Ha 
species, many of which were formerly 
species N. jasminiflora (Klotzsch) N. Hallé, e to 
e genuine Tricalysia species; a new us, Ser- 
icanthe Robbr., was described to ami the 
remaining species (Robbrecht, 1978). 

The close association between Diplospora DC. and 
Discospermum Dalzell with Tricalysia has long been 
recognized, with some authors (e.g., Schumann, 1891) 
considering Diplospora and Tricalysia to be synony- 
mous. Ali and Robbrecht i broadly surveyed 
Diplospora and Discospermum and enumerated a 

whole suite of characters s could be used to 
distinguish these Asian taxa from the closely related 
African Tricalysia species. They also justified Diplo- 
spora and D Wi ad as separate genera on the 
basis of fruit morpholog 

The most recent M work on Tricalysia, by 
Ranarivelo- dera et al. (2007), focused 
on the unnamed gascan section that was allude 
to, but not ben e Robbrecht (1987). Of the 12 
species of Tricalysia occurring in Madagascar, only 
one species belongs to subgenus Empogona (T. 
es ae Hiern). The other 11 species, characterized 
by the presence of unisexual flowers, belong to 
subgenus Tricalysia. Ranarivelo-Randriamboavonjy 
et al. (2007) observed that the Madagascan taxa could 
be accommodated within section Tricalysia were it not 
for the presence of unisexual flowers. As a result, they 


formally placed these 11 taxa in Androgyne Robbr., a 
new section within subgenus Tricalysia. 
netic investigations 


Recent phyloge incorporating 


morphological and molecular data sets have enabled us 
to improve our understanding of the systematic position 
of Tricalysia and its relationships with associated genera 
(Andreasen & Bremer, een Persson, 2000; Bridson & 
Verdcourt, 20 ht & Manen, 2006; Davis et 
al., 2007). Andreasen and Bremer (2000) assessed tribal 
and generic delimitation in subfamily Ixoroideae using 

morphology, plastid and nuc 
sequences, and restriction site (restri 
oe polymórphism) in Their results highlighted 
L. and Psilanthus Hook. 


. (Coffeeae s. str) and several members of the 
Gardenieae subtribe Diplosporinae (Diplospora and 
Tricalysia), resulting in an expanded circumscription of 
the tribe Coffeeae to include Tricalysia, Diplospora, 
Discospermum, | Sericanthe, Coffea, Psilanthus, an 
Bertiera Aubl. Bridson and Verdcourt (2003) further 
enlarged and modified the concept of Coffeeae on the 
ee. plastid data 

ed). In contrast to 
d study of the Rubiaceae (Robbrecht & Manen, 

, the genus Bertiera was excluded from Coffeeae 
tribe, Bert 


Davis et al. (2007) al i circumscription 


A don in its own tri 


and phylogeny of Coffeeae and Bertiera using sequence 
data from 
intergenic spacer, accD-psal, 


three plastid regions (trnZ-F intron and 
and rpl16) in combina- 
tion with morphological data. Their study confirmed 
the placement of Tricalysia and related taxa (Seri- 


Coffea and 


d Coffeeae to include six other 


canthe, Diplospora, and n 
Psilanthus, and expand 
genera (Argocoffeopsis i Belonophora Hook. f., 


obbr., and Xantonnea Pierre ex Pit.). 
study only surveyed a limited number of Tricalysia 
species, all of which belong to subgenus Tricalysia. 
Bertiera was excluded from Coffeeae and retained in 
Bertiereae, in agreement with Bridson and Verdcourt 
(2003), and Gardenieae subtribe Diplosporinae was 
placed in synonymy with Coffeeae. 

The current investigation uses DNA sequence data 
o tes ysia as currently 
circumscribed and to assess the accuracy of the 
subgeneric classification for the genus (Robbrecht, 
1979, 1982, 1983, pi This is the first molecular 

study to include 


representative 
ie of Trica a In addition, we reassess the 


espread 
phylogenetic relationships within the broadly cireum- 
scribed Coffeeae, with an expanded sampling from 
both subgenera of Tricalysia. Given the wealth of 
trnL-F, rpl16, and accD-psal sequence data already 


196 


Annals of the 


Missouri Botanical Garden 


Summary of species from Tricalysia subgen. Empogona and subgen. Tricalysia sampled in this study (following 


classification of Robbrecht, 1979, 1982, 1983, 


1987) 


A) Tricalysia subgen. Empogona (ca. 27 spp., 


Robbrecht, 1979) 


Section 


Species group 


Species 


Tricalysia sect. Empogona Hook. f. 12 spp. 


T. discolor group 


T. 


acidophylla Robbr. 


sensu Robbrecht, 1979 


T. junodii group 


T. junodii (Schinz) Brenan 


T. 


ngalaensis Robbr. 


. ovalifolia Hiern 


No known group affiliation within sect. 


Empogona 


concolor N. Hallé 


gossweileri S. Moore 


T. crepiniana group 


bequaertii De Wild. 


riso. sect. Kraussiopsis Robbr. 15 spp. 
ensu Robbrecht, 197 


talbotii (Wernham) Keay 


T. ruandensis group 


cacondensis Hiern 


- lanceolata (Sond.) Burtt Davy 


M[Sim[SmimISISIS 


. ruandensis Bremek. 


B) Tricalysia subgen. Tricalysia (ca. 75 spp., 


Robbrecht, 1982, 1983, 1987) 


Section 


Species group 


Species 


Tricalysia sect. Probletostemon (K. Schum.) 


. anomala E. A. Bruce 


Robbr. 4 spp. sensu Robbrecht, 1983 


DH 


elliottii (K. Schum.) Hutch. & 
Dalziel 


aciculiflora Robbr. 


Tricalysia sect. Ephedranihera Robbr. 9 spp. 
sensu Robbrecht, 1982 


acocantheroides K. Schum. 


. bridsoniana Robbr. 


Tricalysia sect. Tricalysia 40 spp. sensu 


T. angolensis group 


griseiflora K. Schum. 


Robbrecht, 1987 


Core group for sect. Tricalysia 


bagshawei S. Moore 


coriacea. (Benth.) Hiern 


microphylla Hiern 


okelensis Hiern 


pallens Hiern 


Tricalysia sect. Rosea (Klotzsch) Robbr. 9 spp. 
sensu Robbrecht, 1987 


jasminiflora (Klotzsch) Benth. 
ex Hiern 


. schliebenii Robbr. 


Tricalysia sect. Androgyne Robbr. 11 spp. sensu| 
Ranarivelo-Randriamboavonjy et al., 2007 


ambrensis Randriamb. & De 


. analamazaotrensis Homolle ex 


Randriamb. & De Block 


- cryptocalyx Baker 


RE 


- dauphinensis Randriamb. & 


De Block 


. leucocarpa (Baill.) Randriamb. 


T. 


perrieri Homolle 


Randriamb. & n PN 


Volume 96, Number 1 


Tosh e 
B of Tricalysia 


197 


Table 2. Amplification primers used in this study. 
Region Primer Primer sequence (5’-3’) Reference 
irnL-F Forward (c) CGA AAT CGG TAG ACG CTA CG Taberlet et al., 1991 
Reverse (f) AAT TGA ACT GGT GAC ACG AG 
rpl16 Forward (711) GCT ATG CTT AGT GTG TGA CTC GTT G Jordan et al., 1996 
Reverse (16611) CGT ACC CAT ATT TTT CCA CCA CGA C 
Reverse (1516r) CCC TTC ATT CTT CCT CTA TGT TG Shaw et al., 2005 
Internal forward GTA AGA AGT GAT GGG AAC GA Davis et al., 2007 
Internal reverse TCG TTC CCA TCA CTT CTT AC 
accD-psal Forward (769 F GGA AGT TTG AGC TIT ATG CAA ATG Mendenhall, 1994 
everse (75 AGA AGC CAT TGC AAT TGC CGG AAA 
petD Forward (1365) TTG ACY CGT TIT TAT AGT TTA C Lóhne & Borsch, 2004: 


Reverse (738) 


AAT TTA GCY CTT AAT ACA GG 


2007), we have 


focused on these three plastid regions in the current 


available for Coffeeae (Davis et al., 


investigation and included further sequence data from 
the plastid region petD. 


MATERIALS AND METHODS 
TAXON SAMPLING 


An expanded sampling of Tricalysia, Diplospora, 
Discospermum, Sericanthe, and Bertiera was combined 
with sequence data generated by Davis et al. (2007). 

p db both subgenera and 
the seven secti of the genus (Robbrecht, 
1979, 1982, 1983, 1987) were included in the 


analyses (Table 1). Representative taxa from Ixoreae, 


Tricalysia samples 
ll of 


ctotropideae, and Gardenieae were selected as the 
outgroup. A list of the 80 accessions used in the study 


is given in Appendix 1. 


DNA EXTRACTION, POLYMERASE CHAIN REACTION 
AMPLIFICATION, AND SEQUENCING 


Most DNA samples were obtained from silica gel 
collections or, alternatively, from seed, flower, or leaf 
samples taken from herbarium specimens (BR, K, 
MO, WAG). A small number of DNA samples were 
obtained from fresh leaf material collected from the 
living collections of the National Botanic Garden of 
Belgium 


For silica gel samples, DNA was isolated using a 
modified DNA Mini Extraction Protocol (Royal 
Botanic Gardens, Kew [K] protocol). DNA samples 
were obtained from herbarium material using the 2X 
CTAB protocol of Doyle and Doyle (1987), er the 
DNA subsequently purified using cesium chloride/ 
ethidium bromide gradients and concentrated by 
dialysis before inclusion in the DNA Bank at K. All 
D samples were purified using a NucleoSpin 
Bethlehem, 


purification column (Macherey-Nagel, 


Pennsylvania, U.S.A.) according to the manufacturer’s 
instructions in order to remove any potential poly- 
merase chain reaction (PCR) inhibitors. 

Amplification of the trnL-F, rpl16, petD, and accD- 
psal plastid regions was carried out using the primers 
listed in Table 2. Amplification of the rp//6 region 
was primarily carried out using the forward primer 71f 
and the reverse primers 1661r (Jordan et al., 1996) 
and 1516R (Shaw et al, 2005), although Coffeeae 
specific internal primers designed by K were also 
required for certain taxa (Davis et al., 2007). 

All and sequencing reactions were performed 
using a Perkin Elmer (Waltham, Massachusetts, 
U.S.A) GeneAmp 9700 Thermal Cycler machine. 
Amplification of trnZ-F was carried out using the 
following profile: 94°C for 3 min.; 32 cycles of 94°C 
for 1 min., 51°C for 1 min., 72°C for 2 min.; and a 
final extension of 72°C for 7 min. accD-psal and rpl16 
were amplified as follows: 94°C for 3 min.; 32 cycles 

of 94°C for 1 min., 52°C for 1 min., 72°C for 


30 sec.; and a final extension of 72°C for 7 min. 


1 min. 


a ee of petD was carried out as follows: 96°C 
for 2 min.; 34 cycles of 94°C for 1 min., 50°C for 
1 min., m C for 1 min. 30 sec.; and a final extension 
of 72°C for 10 min. 

For the trnL-F, ue and rpl16 regions, 25 pl PCR 
reactions were made using a commercial PCR master 

1 mM MgCl; ReddyMix; ABgene; 
Surrey, U.K.). accD-psal did not amplify successfully 


Epsom, 


with the commercial master mix, so 25 pl PCR master 
mixes were prepared using Biotaq DNA polymerase 
(Bioline, London, U.K.), 2.5 pl of 10x NH, reaction 
buffer (Bioline), 1.5 ul of 50 mM MgCl;, and 2.5 pl of 
n, Wisconsin, U.S.A.). All 


roduets were purified using Nucleo- 


NTP pug Madiso 
amplified PCR 
Spin feito columns following the manufactur- 
er's protocol. 
Cycle sequencing reactions were carried out using 
BigDye Terminator Mix version 3.1 (Applied Biosys- 
Warrington, Cheshire, U.K.). The cycle 


tems, Inc., 


198 


Annals of the 
Missouri Botanical Garden 


sequence reaction consisted of 26 cycles of 10 sec. at 
96°C, 5 sec. at 50°C, and 4 min. at 60°C 


sequencing products were cleaned with the MagneSil 


. Cycle 


Clean-Up System (Promega) using an automated robot 
(Biomek NX 58; Beckman Coulter, High Wycombe, 
Buckinghamshire, U.K.). Analysis of cycle sequenc- 
ing produets was performed using an 0 DNA 
Analyzer (Applied Biosystems). In addition, a number 
of the trnL-F and petD samples were sent to Macrogen 


(Seoul, South Korea) for sequencing. 


ALIGNMENT AND GAP CODING 


equences were assembled and edited using the 
Staden software package (Staden et a 11 
sequences were aligned manually in MacClade 
(version 4.04, Maddison & Maddison, 2002). Low 
levels of sequence variation enable 
aligned ov diffic 
alignment, 


d sequences to be 


sequences, were removed. The edited sequences were 
yzed with gaps treate missing data a 
phylogenetically informative indels (insertions and/or 
deletions) coded according to the “simple 
coding" method of Simmons and Ochoterena (2000). 


indel 


PHYLOGENETIC ANALYSES 


Phylogenetic analyses were performed on the four 
separate plastid data sets in addition to the combined 
four-region plastid matrix. 


Maximum parsimony. Heuristic tree searches were 
carried out in PAUP* version 4.0b10 (Swofford, 2003) 
using 10,000 replicates of random taxon sequence 
addition, holding 10 trees at each step, with tree 
bisection-reconnection (TBR) branch swapping, 
elayed transformation (DELTRAN) optimization, 
and MUL effect, and saving no more than 
10 trees per Sc Support values for clades 
recovered in the I were estimated using 
bootstrap analysis (Felsenstein, 1985). One thousand 
replicates of simple sequence additi Rs 


ping, and saving 10 trees per replicate were rdi 


tion, T 


in PAUP*. We interpreted bootstrap values greater 
than 85% as being 
being moderately supported, and 5096—7446 as having 


well supported, 


o as 
low support 
Bayesian mes Evolu utionary models for each 
plastid re si 
(Posada E Crandall, 1998) under the 
information criterion. In the case of accD-psal, petD, 
rpl16, the nucleotide substitution model that best 
fits the data was HKY + I + G. The HKY + I model 


was selected for the trnL-F sequence matrix. The 
combined data set was partitioned into five discrete 
units. In addition to the four plastid regions, there was 
a fifth partition for the phylogenetically informative 
indels. The restriction site (binary) model of evolution 
for the indel data, S the 
recommendation of Ronquist et al. 


was implemente 


independent Bayesian searches, each a of 
two simultaneous parallel analyses, were carried out 


a MrBayes 3.1 (Huelsenbeck & Ronquist, 2001). 
In each Bayesian analysis, four Markov chains (three 
E) one cold) were run simultaneously for 


2,000,000 generations, sampling trees every 100 
generations. Thei initial 25% of trees were discarded 

s a conservative burn-in. After confirming by eye that 
trees dier from separate analyses had consistent 
topologies, the “post-burn-in” trees from each analysis 
were pooled together, imported into PAUP* version 
4.0b10 (Swofford, 2003), and summarized by majority 
rule consensus, with values on the tree equating to 
posterior probabilities (PP). 


RESULTS 


This study generated 229 sequences, which were 
combined with the 75 sequences obtained by Davis et 
2007). In total, this study included 79 accD-psal 
sequences (53 newly generated), 80 trnL-F sequences 
54 newly P 78 rpl16 sequences (55 newly 
erated), etD 
um ted). n x 6 eee proved to 


a 
E 


sequences 


problematic region to amplify, due in part to two poly- 
" stretches B. 373 bp from the 5' end, the other 
66 bp from the 3' end). As a 


polit to dns sufficient overlap during 


result, it was often 
g sequence 
assembly. Internal primers, designed specifically for 
Coffeeae taxa (Davis et al., 2007), were used to obtain 
a complete sequence for rp//6 in problematic taxa. 
In general, the amount of genetic variability in all 
plastid regions was low (Table 3). A large proportion 
of the total genetic variation occurred between the 
ingroup (Coffeeae) taxa and outgroup (other Ixoroi- 
deae). We observed considerable length variability in 
the ee -psal region. As with all the plastid regions 
investigated, accD- a is A age ee AT-rich and 
mber 


informative indels. In 


subject to several re 

of potentially fe nee 
the case of Tricalysia subgen. Empogona, all taxa 
included in the study share a 250 bp deletion in the 


s, giving rise to a nu 


accD-psal region. Less length variation was observed 
in petD, rpl16, and t 


all four individual analyses were examined by eye and 


rnL-F. The gross tree topologies of 


found to be topologically consistent, and the four data 
sets were subsequently combined in all further 
analyses 


Volume 96, Number 1 Tosh et al. 199 
2009 Phylogeny of Tricalysia 


The aligned combined matrix had a total length of 
ke 4465 bp. There were 669 variable characters and, of 
E = these, 352 characters were parsimony informative 
T 22 ue uu S 3 E ce (7.9% of total number of characters). In total, the 
p ee ses ea, 1o SSS & matrix contained 50 parsimony informative indels, 
E "T consisting of repeat sequences in addition to insertion/ 
= eletion events. Exclusion of outgroup taxa (Ixoreae, 

Gardenieae, Octotropideae, and Bertierieae) revealed 
211 parsimony informative characters within Coffeeae. 
e 5 
ES PHYLOGENETIC RESULTS 
312278022833 
= Ks = The heuristic maximum parsimony (MP) analysis of 
the combined plastid data matrix generated 8853 most 
parsimonious trees with a length of 929 steps, a 
E consistency index (CI) of 0.816, and a retention index 
a Z2. Wo tsa (RI) of 0.908. Table 3 summarizes the tree statistics 
SI/SETEMPTABSES ee . 
a, 2509 RISSA for the individual and combined analyses. 

T The topologies of the MP strict consensus tree and 
the Bayesian majority rule tree (Fig. 1) were consis- 
tent with each other. Figure 2 displays one of the 

c c parsimonious trees and indicates both sees 

Slos- SA Iag support (BS) and branch rM Both MP and 
BLT ALS “e835 ] hyly of th 
. S KS Z ayesian ana e reconfi the nophyly of the 
El e ingroup (BS 99%, PP 1.00). cn is represented 
2 by its two Kd is monophyletic (BS 100%, PP 
st 
ES 1.00) and is sister to the ingroup (BS 7996, PP 1.00). 
E i = = The clade and Psilanthus is well 
E alo Bata ds EN a x supported (BS 10075, PP 1.00) and is sister to the 
P 3 CE o BNaANSSS remaining ingroup taxa (BS 93%, PP 1.00). There is 
mt 
a S " also strong support for the clade of Argocoffeopsis and 
3 Calycosiphonia (BS 99%, PP 1.00) The sister 
E relationship of Calycosiphonia and Argocoffeopsis to 
E the rest of the ingroup receives weak bootstrap support 
E (BS 50%), but is supported by a PP of 0.98. 
E n MP and Bayesian analyses recovered a clade 
c cluding bd MCA b D ns ermum, and 
E Tricaly sia subgen gona. Although there is no 
* bootstrap support - ios clade (BS « 50%), the clade 
R pe does receive support in the Bayesian analyses (PP 
e 2 0.98). Within this clade, there is strong support for the 
u E monophyly of Sericanthe (BS 99%, PP 1.00), Dis- 
= M cospermum (BS 100%, PP 1.00), and the group of 
& E g ode and Tricalysia subgen. Empogona (BS 
a e % 99%, PP 1.00). The monophyly of both po (BS 
B ee 90%, op : 00) and Tricalysia subgen. Empogona (BS 
3 B g 98%, PP 1.00) is confirmed. Within Tricalysia 
a ~ 8 # E E as d two re receive high levels 
9 = E E c E of su . cacondensis Hiern, T. 
EE SE E iz xO a Sud p Davy, and T. ruandensis 
E o 
o ¿rio E E E Bremek. (BS 85%, PP 1.00); and the group of 7. 
Wem det oS Uu m 
ed 33 3 23 FE junodii (Schinz) Brenan, T. za and T. acid- 
o 35882589858 eta (BS 98%, PP 1.00). 
3 wo T DECEM E E w 
E SHEETS gag ade of Belonophora and Tricalysia subgen. 
CEREALES Tricalysia is present in both the MP strict consensus 


200 Annals of the 
Missouri Botanical Garden 


Ixora hookeri 


Doricera trilocularis 
1.00 Gardenia thunbergia 
: 1.00 Hyperacanthus 5 
0.94 1.00 Hyperacanthus perrieri 5 
Fyperacanthus microphyllus 4 
1.00 Eri d nora I 
1.00 Canephor S 
Polysphaeri ia o 
1.00 Bertiera breviflora 
1.00 al Bertiera bicarpellata . 
loo — Bertiera Bertiera 
B ertiera iturensis s 
on 1.00 j 
3 rgocoffeopsis rupestris 
1.00 asoca an sdandens Argocoffeopsis & 
: 1.00 Cal pa prona macrochlamys Calycosiphonia 
— Calycosiphon 
1.00 Calycosi honia spathicalyx 
7 0.96 Belanophor ra coriacea 
— mne Belonopho ra coriacea Belonophora 
Belonophor: 
1.00 — Tri E 
— Trica sia a EPH 
1.00 [— Trica ocantheroides EPH 
Trical a a acocantheroides EPH 
[—— Tricalysia ambrensis A 
[100 — Tricalysia analar O E is AND 
0.82 Tricalysia analamazaotrensi is AND 
calyx AN 
1.00 
oo} [7 
Tricalysia subgen. 
Tricalysia 
icalysia fes] 
1.00 = Tricaly sia anomala PRO < 
Tricalysi, ysia elliottii PRO m 
io Tricalysia bagshawi TRI m 
0.98 n TRI 5 
d 0.84 Tricalysia pallens TRI 5 
Tricalysia pallens TRI 
Tricalysia pallens TRI 
1.00 œ Tricalysia coriacea TRI 
Tri ] TRI 
Dipi pora a 
Iplospor: 7 
[1-00] Polospora d eu ge ee 


Trica Hsia ovate lia EMP 
Tricalysia ovalifolia EMP 
Tricalysia junodii EMP 
Tricalysia concolor EMP ; a 

Ti rical ysia poe EME Tricalysia subgen. 
Tri ogona 


Ti o Ibotii KRAR 

Tricalysia cacondensis KRA 

Tricalysia lan eola KRA 
sia ruandensis KRA 


Tricaly: 
Tricalysia nealaensis EMP 
1.00 p— 4 iscospermum e A 
'—— Discospermum abnorme Discospermum 
1.00 Sericanthe andongensis 
pH er ede i D E is ; 
1.00 anthe jacfelici Sericanthe 
Lo er ricantha 
1.00 offea homollei 
0.98 Coffea mangoroensis 
1.00 Coffea n mora. n : Coffea & 
0.85 Psilanthus ebracteolatus Psi 
silanthus 
109 Psilanthus mannii 


Psilanthus semseii 


e l. Maximum parsimony strict consensus/Bayesian majority rule consensus tree. Bayesian posterior probabilities 
are indicated EA branches: Sectional groupings are annotated after species names: AND, Tricalysia sect. i n ; EMP, 
Tricalysia se mpogona; EPH, Tricalysia sect. Ephedranthera; KRA, Tricalysia sect. Kraussiopsis; PRO, Tricalysia sect. 
T ROS, Tricalysia sect. Rosea; TRI, Tricalysia sect. Tricalysia. See Table 1 for species authorities 


and provenance. 


Volume 96, Number 1 Tosh et al. 201 
2009 Phylogeny of Tricalysia 


Ixora hookeri 


Doricera trilocularis 
l a 
lyperacanthus 2 
ecu perrieri e 
Hyperácanthus microphy. llus, ; [o] 
d ES 
Canephora 5 
Polysphaeria —  — — , O 
100 Bertiera 
22 1oy = "eric iturensis ú 
—— Bertiera 2 j 
x —— Argocoffeopsis eketensis : Argocoffeopsis de 
2 "al ij Calycosiphonia 
EAN ` cs phon 
73 hlamy 
2 100 4Belonophora cori iacea 
T Mu al coriacea Belonophora 
BI Tric alysia aciculiflora EPH 
; d iculiflora EPH 
96 da ta acocantheroides EPH 
ALC Tricalvsia acocantheroides EPH 
Tricalysia ambrensis AND 
ri i la trensis AND 
m Tricalysia analamazaotrensis AND 
ia crypto ca: 
lysia cryptocalyx AND 
EH Tricalysia dauphinensis AND 
[79 | a a a is AND 
5 a dauphinensis AND 
Tricalys m asminiflora | ROS 
Dn lysia leucocarpa AND Tenaya sibeen 
Tric ly: sia tenco arna AND 
əri AND Tricalys 
OS 
m 
< 
fes] 
m 
oy 
S 
[50] o 
99 Diplospora 
13 
Tricalysia subgen 
Ta ia 
raha Empogona 
E “albo tii KRA 
Tricalysia uA KRA 
Ly Tri "icalysi ia lanceolata KRA 
Tricalysia ngalaensis EMP 
E Mr BURN Discospe ici 3 
Discospermum abnoi Discospermum 
95 Se calle donani 
E {= Sericanihe andongensis i 
Sericanthe 
n — Sericanthe 
95 Coffea homollei 
joo Coffe ee 
C moratii y o 
Psilanthus ebracteolatus Caff ado 
Psilanthus mannii Psilanthus 
Psilanthus semseii 
— — $ changes 


igure 2. One of the 8853 most parsimonious trees generated in the maximum parsimony analysis. Bootstrap values > 
50% are indicated above branches, and selected branch lengths are indicated below bra iius zm Pas gs are 
annotated after pene names: AND, Tricalysia sect. pm ne; ; ura sect. Emp Tricalysia sect. 


ri^ 
Tricalysia sect, Tricalysia. See Table 1 for species authorities and provenance. 


tree and the Bayesian majority rule tree, rc) Dalziel and 7. anomala E. A. Bruce (BS 95%, PP 
there is negligible support for this clade 0%, — 1.00), and a group of predominantly Madagascan taxa 
PP 0.82). However, the monophyly of B ae (BS with the inclusion of 7. jasminiflora (Klotzsch) Benth. 
100%, PP 1.00) and Tricalysia subgen. Tricalysia (BS. & Hook. f. ex Hiern (BS 97%, PP 1.00). There is also 
99%, PP 1.00) is strongly supported. Within Trica- moderate bootstrap (BS 75%) and high PP (PP 1.00) 
lysia subgen. Tricalysia, several groups receive strong for the clade of T. acocantheroides K. Schum., T. 
support: the group of T. elliottii (K. Schum.) Hutch. € — griseiffora K. Schum., T. bridsoniana Robbr., T. 


Annals of the 
Missouri Botanical Garden 


and the 


aforementioned Madagascan group together with 7. 


microphylla Hiern, T. schliebenii Robbr., 
jasminiflora. 


DISCUSSION 


Previous taxonomic work on Tricalysia has focused 
on the use of traditional morphological and anatomical 
characters to infer relationships within the genus. In 
the most recent classification of the genus, Robbrecht 
(1979, 1982, 1983, 1987) subdivided it into two 
subgenera and seven sections. Here, for the first time, 
we have addressed relationships in this group using 
molecular data. 

In the current investigation, we obtained sequence 
data from four plastid regions for both Pd and 

of Tricaly. and generated 
estimates of phylogeny using both MP and Bayesian 


all seven sections 


inference methods. The consensus tree topologies of 
both analyses b consensus for MP, majority rule 
consensus for Bayesian) were consistent. Ás is often 
observed, Bayesian PP were higher than bootstrap 
support values for any given node (Huelsenbeck et al., 


2002; Erixon et al., 2003; Randle et al., 2005) 


TESTING THE MONOPHYLY OF THE GENUS TRICALYSIA 


Our phylogenetic analyses indicate that Tricalysia, 
as currently ee s not monophyletic. The 


monophyly of subgenera abia and Empogona is 
confirmed, but they are not sister to each other. This 
though 


observation, which has implications for the taxonomy 


represents a new, perhaps unsurprising, 
of the group (see below). 
(2007) included five species of 


Tricalysia in their molecular and morphological 


Davis et al. 


e circumscription and p 


2 
‘a 


reassessment 
of Coffe 


subgenus Tricalysia. In both their combined molec- 


o yloge 
eae. All five species were representatives of 
ular and combined morphological-molecular phylog- 
enies, Tricalysia (subgen. Tricalysia) was placed in 

oorly supported and olved clade containing 

Sericanthe, Belonophora, and an Asian clade (includ- 

ing Diplospora and Discospermum). The study of Davis 

et al. (2007) incorporated molecular data from three 
ac 


plastid regions (trnL-F, accD-psal, and rpl16). In our 


investigation, we included an additional plastid 
rd the group II intron petD. The extra characters 
provided by this fourth plas 


sti still not 
xx to fully elucidate systematic relationships 


tid marker were 


within the clade containing Tricalysia subgen. 
Tricalysia, Sericanthe, Belonophora, Diplospora, and 
Discospermum. 

e inclusion of taxa from Tricalysia subgen. 


Empogona led to results conflicting with the study 


of Davis et al. (2007). First, we did not recover an 
Asian clade. Instead, Diplospora formed a well- 
orted monophyletie group with Tricalysia subgen. 

(BS 99%, PP 1.00). S the MP 


strict consensus tree and the Bayesian majority rule 


supp 
Empogona econd, both 
consensus tree indicated sister relationships between 
and Belonophora, and 
nthe, Discosper- 
mum, Diplospora, and Tricalysia subgen. Empogona. 


Tricalysia subgen. Tricalysia 
ed a clad 


recovere cl ade conta aining Serica 


The clade of Belonophora and subgenus Tricalysia 
received poor internal support (BS < 50%, PP 0.82), 
but there was support for the second clade in the 
Bayesian analysis (BS < 50%, PP 0.98). 


TAXONOMIC IMPLICATIONS FOR GENERIC CONCEPTS 


The revelation that Tricalysia sensu Robbrecht is 
not monophyletic calls for a reconsideration of the 
taxonomic delimitation of Tricalysia ae closely 
related taxa. One taxonomic option would o widen 
the genus Tricalysia to include Finch Diplo- 
spora, Discospermum, and Sericanthe. However, these 
genera are easily identified (e.g., by the use of a key) 
and are so diverse morphologically and anatomically 
that consolidating them into one genus 
justified (Table 4). A more logical option would be to 
separate these taxa into groups at the generic level, 
based on morphological and molecular synapomor- 

hies 


Robbrecht (1979) enumerated four potential field 
characters that distinguish the subgenera Empogona 
subgenus 


and Tricalysia. Taxa o Empogona are 


identified by the presence of distinctly lobed calyces 


tive (vs. blunt anthers, occasionally forming a short 
triangular appendage), and fruits that turn black at 
maturity (vs. red fruits). Robbrecht (1979) considered 
recognizing Empogona at the generic rank, but opte 

to incorporate it as a subgenus of Tricalysia, given the 
similarity in a number of other key characters 


c 


placentation, pollen morphology, fruit and see 

morphology, and seed coat anatomy). This decision 
was also pragmatic in terms of taxonomic stability, as 
it required the fewest nomenclatural changes (Rob- 
brecht, 1979). 

The revision of Sericanthe (Robbrecht, 1978) and 
the survey of the Asian relatives of Tricalysia (Ali & 
Robbrecht, 1991) provided ample morphological and 
anatomical evidence to justify the exclusion of these 


bacterial leaf galls (rare in Rubiaceae), wing-shaped 


Volume 96, Number 1 
2009 


et al. 
Phylogeny of Tricalysia 


colleters, and pollen with a verrucate sexine (in 
contrast to the reticulate sexine occurring in all other 
2007) also 
presented the following synapomorphic characters 
for the genus: 
basifixed anthers, and horizontal micropyle orienta- 


members of Coffeeae). Davis et al. ( 
7- to 9-merous flowers, distinctly 


tion. 
Diplospora and Discospermum consistently have 
tetramerous flowers, which occur only rarely in 
n d and the flowers of Asian taxa are 
than their African counterparts (Ali € 
ub ee addition, there is 


tendency toward eer flowers in Asian taxa, a 


a strong 


trait that is absent in all but a few representatives of 
Tricalysia confined to Madagascar (Ranarivelo-Ran- 
driamboavonjy et al., 2007). Ali and Robbrecht (1991) 
also justified maintaining Diplospora and Discosper- 
mum as separate genera on the basis of rather 
divergent fruit types (small, fleshy, and red fruits in 
Diplospora and large, leathery, and purplish black 
fruits in Discospermum). The decision to maintain 
Diplospora and Discospermum as separate genera is 
also supported by our molecular analyses 

The tribal position of Belonophora has been fairly 
unstable since its initial description. by Hooker 


observed that Belonophora species actually possess 
two collateral ovules per locule, on the inner surface 
of a pendulous placenta, but he felt it premature to 


assign the genus to a new tribe until a more 


p tribal classification within Rubiaceae 
recht RN E (1986) 
ea Misco Po QE IS in the e Aulaco 


calyceae, although the axillary eee of 
Belonophora contrasted with the terminal or subter- 

minal inflorescences possessed by other members of 
the tribe. ement of Belonophora in the tribe 
Coffeeae was first proposed by Bridson and Verdcourt 
(2003) and later supported by the study of Davis et al. 
(2007). Th 


synapomorphic for the genus in the study of Davis et 


The plac 


e imbricate calyx lobes of Belonophora were 


al. (2007), and the genus is also distinguished from 
other members of Coffeeae by the presence of a 
superior embryo radicle (Cheek & Dawson, 2000). 

In light of evidence from our own molecular 
investigation, and in combination with morphological 
and anatomical observations reported elsewhere, we 
is appro 


believe it iate a ully justified to 
) at 


n 
u Robbrec ht, 1979 


the inclusion of many former Tricalysia species in the 
genus Empogona are provided at the end of the 
Discussion. 


RECOGNITION OF INTRAGENERIC GROUPS IN TRICALYSIA 


n addition to testing the monophyly of Tricalysia 
sensu Robbrecht, we were able to assess the levels of 
support for his sectional groups within the genus. All 
seven sections were sampled in our analysis, although 
some were better represented. Low levels of genetic 
variation. between species limited the amount of 
resolution between taxa, but there are some provi- 
sional findings from this study. 

Tricalysia subgen. Tricalysia was subdivided into 
five sections by Robbrecht (1982, 1983, 1987) 
Tricalysia sect. Probletostemon, here represented in 
our molecular study by T. elliottii and T. anomala 
(Table 1), was thought to possess many morphological 
and anatomical features regarded as primitive for the 

oup. ese included free bracteoles, standard 
rcm (Robbrecht, 1988), large pleiomerous flowers 
with many ovules per placenta, and large fruits 
(Robbrecht, 1983). Our study confirms the monophyly 
of section Probletostemon (BS 95%, PP 1.00), but it 
remains unresolved in a basal polytomy. 

Tricalysia sect. Ephedranthera, here — by 
three species, is characterized by the presence of 
anthers that are sessile in the corolla throat and 
partly included within the corolla tube (Robbrecht, 
1982). The monophyly of this section is not supporte 
in our investigation. Tricalysia aciculiflora Robbr. 
falls within the basal polytomy, whereas 7. aco- 
cantheroides and T. bridsoniana are situated within 
the moderately to well-supported clade (BS 75%, PP 
1.00) containing all the remaining taxa of subgenus 
Tricalysia. 

The other three sections (Tricalysia, Rosea, and 
Androgyne) are very similar DUE MA Most 
species in subgenus Tricalysia belo section 
Tricalysia, which Robbrecht (1987) m subdivid- 
ed into four informal groups. Only two of these 
informal groups are included in this investigation. The 
core group of taxa within section Tricalysia, here 
represented by T. coriacea (Benth.) Hiern and the 
weakly supported clade of 7. pallens Hiern, T. 
okelensis Hiern, and bagshawei S. nr is 
unresolved i in the e Pa The 
angolensis A. Ric DC., repre 
pora K. E fall within the clade containing 
T. bridsoniana, T. microphylla, and representatives 
from sections Rosea and Androgyne. 

In section Rosea, species differ conspicuously from 
sence of a 
1987). In section 
Androgyne, which comprises the Madagascan repre- 


those in section Tricalysia due to the 
spathaceous calyx (Robbrecht, 


sentatives of subgenus Tricalysia, species are char- 
acterized by the presence of unisexual flowers. There 
is weak bootstrap and significant Bayesian support 


Annals of the 


204 


Missouri Botanical Garden 


O€-OL ST-0T 6-6 0c-6 01-8 1-6 08-07 (wu) imag 
unidos unidos ou) 
ou jo xode umjdos ou jo jo epppra: oui 
oj 0} pouoene xode eq 0} peqoene 0] payoene umjdos əy} umjdos umidos ei 
‘eyuooryd *eguoove[d. prosdrj[o *tequeove[d jo epppru: eu o1 eu jo eppprur eu JO epppra: eu o1 
prosdie -uoy $ re[nomno prosdie poyoene tejoore[d 0] peqoene tequooe]d poyoene tejueoe[d umjdos eq jo eppprur 
-uoy x -uoy snompued -uoy x 01 prosdippo-rureu + prosdio-rureq prosdio-rureu. + eq 01 poyoene 5ejueov[d 
Te[nodro-uro e e JO 92€] 1QUUL Teqnoro-ruroq 0] re[nomo-Turoq F7 0) re[nogro-ruroq 0] re[noz-rureq prosdij[e-rureq = 03 1e[noxto 
uo sopnao (c-)z(-T) uo So[nAO [erojye][OO g e uo so[naO g-z e uo sopnao Z[-T e UO so[nao GZ '&2-T e uo somao (Q-)E-T -muom ? uo so[n4o c[-c(-g) uoneuooe[qd 
,sasepusdde 
9xg-uoqqri o[suern 
oSepuodde ou o[Suern oSepuodde S[SUBLO oys snonoidsues oys Átoa ur o[Suern noys 9ATJOQUUO) 
tpeuepnep A[Suons poys ur Surpnjoad Jeorde poys ur Surpnujoid. Apsour ursurpnajoad — Surpnaoid soumjouros Aroa ur Surpnujoird sournouros Tonpnuy 
pouesxo 
“9[15sos 
10 yeon 
peuesxe ‘wom ur squoureg pouesxe *eonp pouesxo ponosxo pouesxo 
ur sjueure[r pepnqour ‘eqn gos ur squoure[g ‘eop ur squeure[g ‘ory ur squeuremg ‘eop ur squoure[g 
noys uo tpexrrseq UL o[rssos fpoxrrpour uo pexugrpeur — Suo[ uo :pexrjrpeur Suo[ uo tpexryrpeur jus uo tpexryrpeur jus uo ‘poxyrpow sro y 
Airey 10 snorqe[s snoxqve[s Airey Airey 01 enoxqe[o qpapaeoq Áppsusp Airey 10 snorqe[s poepreoq 03 snoxqe[$ jeor] v[[0107) 
i991 Teoul[ 
[199] omuru SUIT t 10 Te Suum Surdde[1240 u93jo 
yam *pedopo4op pedopo4op *pedopo^op *ouou soqo] pue pedopasp repue Ápsour juosqe 10 juosoxd 
-0m equa -[[9^ seqo[ ‘noys equi -][£*^ equi ípedo[e4^ep-[[BM equi -pom seqo[ ogs aqui seqo[ ‘uoys equi seqo[ pepunoz ‘uoys equi xÁ]e) 
(g-e) (CP) iv (¿Jo Er) (o-s) (=F Y Asoro 
A[Á18019101 yenxestun uoneziue210 
onrporqdeurreq tonporydeuuoy yenxesrun ,ourpordeurreq onporydewog Jo orporydeuney —— [enxesmun 10 onporydewoy JOMOLJI 
(umi) 
ez-c1(-8) (0t0€-0c(-oT) OI-S 05-38 LE) OI-S ST-8 i3uop eroro 
UOULA SOTA] GOL sisdoissnpay 
o1s0ddo *ooxj E 3098 *imoqeo 
sjoviq 1oddn *rmo£qeo i[no4qeo UL JEUNE 9917 oyu posnj ‘nuosoduy Tpnodyeo So[0o1oe1q 
T[naÁ[eo oyu pesnj  OJUT pasny sjoexq 19MO[ our posng ‘modes oyu pesnj "1098 *opeurej[e 99 OJUT pesnj 10 991} I[noÁ[eo ojur pəsny 10 991} pue sjoerg 
(dds og eo) (dds c) (dds 11) (dds gg eo) (dds 6z) (dds gq -e9) (dds 7 eo) 
amos pa401douojag 2u/So1puy au&So1puy "1998 nuosoduy psodsojdig winuttadsOoosiy 
"1098 DISÁJDIM] — Surpn[oxo “DISÁJDILL Y, 
"K[oxex Aroa — sosoyjuored o[qnop 
ur songy :porex — sosomuored o[Suts ut soms1] "puozoduig 105 y ent 1 'Jp[9q ur 8193oexeq I [2 PUB v15410214] Jo S1ooereuo [eorgo[oudrour quotes ‘p o[qeL 


Volume 96, Number 1 
2009 


osh et al. 
Phylogeny of Tricalysia 


Table 4. Continued. 


Tricalysia, excluding — Tricalysia sect. 


Belonophora Sericanthe 
(ca. 20 spp.) 


Androgyne 
(11 spp) 


Empogona sect. Androgyne 


Diplospora 


Discospermum 


(5 spp.) 


(29 spp.) (ca. 80 spp.) 


first white, turning 


(ca. 10 spp.) 


(ca. 7 spp.) 


purplish black 


orange 


yellow 


red, rarely orange 


turning from yellow 


Fruit color 


purple, then black 


fleshy 


and orange to red 


fleshy 


fleshy 


fleshy fleshy 


fleshy; rarely 


sclerified or leathery 


Pericarp 


mosily none; 


massive, surrounding 


none 


mostly none 


none, with weak 


mostly none 


massive, mostly surrounding 


Placental 


massive in some 


seeds 


outgrowths in some 


seeds 


outgrowth 


entire 


entire, ruminate in some entire 


entire 


entire 


entire, ruminate in some spp. entire or ruminate 


Endosperm 


inferior 


lateral 


superior 


inferior 


inferior 


inferior 


away from septum 


Embryo radicle 


* Heterostyly in section Ephedranthera. 


» Glabrous in a few species, e.g., Empogona concolor. 


* Some species with an inconspicuous appendix, e.g., Empogona welwitschii. 


(BS 61%, PP 0.98) for a clade containing these two 
sections. Tricalysia schliebenii (section Rosea) is sister 
BS 97%, PP 1.00) 


to a strongly e clade 
bers of section Androgyne and T. 


containing member 
jasminiflora of section Rosea. 

Robbrecht (1979) recognized two sections within 
subgenus Empogona: section Empogona is character- 
ized by free bracteoles and distinct non-overlapping 
calyx lobes; in contrast, the bracteoles in section 
Kraussiopsis are fused to form calyculi, and the calyx 


exception of Tricalysia bequaertii De Wi 
calyx lobes are not touching). Tricalysia ngalaensis 

Robbr., previously thought to be closely related to T. 
junodii (Schinz) Brenan (Robbrecht, 1979), is in an 


unresolved position (Figs. 1, 2). ere is weal 


monophyly of section os. (BS 56%, PP 0.99), 
and the informal gro T. ruandensis is also well 
supported (BS 85%, PP 100, The remaining taxa o 
section Empogona are weakly supported (BS 60%, PP 
0.96), although the clade of T. junodii, T. ovalifolia, 
and T. acidophylla is well supported (BS 98%, P 
1.00) 


OTHER RELATIONSHIPS WITHIN COFFEEAE AND THE 
RELATIONSHIP TO BERTIERA 


The sister relationship of Bertiera and Coffeeae is 
recovered with moderate bootstrap and significant 
Bayesian support (BS 79%, PP 100), although our 
outgroup sampling is not complete. In order to confirm 
this result, more extensive sampling of representative 
groups within subfamily Ixoroideae is needed. Rob- 
brecht and Manen (2006) opted to place Bertiera in 
subtribe Bertierinae, sister to Coffeinae, 
Coffeeae. Davis et al. (2007) found only weak bootstrap 
support for the sister relationship between Bertiera and 
Coffeeae (BS « 40%) based on molecular data alone, 


combined molecular-morphological analysis. 
on the decision of Bridson and Verdcourt (2003), they 
opted to place Bertiera in the monogeneric tribe 
Bertiereae. Whether Bertiera is recognized at the tribal 
or subtribal level is still open to debate, but we agree 
with Davis et al. (2007: 321) that “Coffeeae, with the 
addition of Bertiera, is 
both scientifically coherent and practical." 


of new genera and the remova 


In the three-region plastid analysis of Davis et al. 
(2007), Coffea and Psilanthus form a well-supported 
monophyletic clade supported by a bootstrap of 87%, 
and are placed sister to the rest of Coffeeae. This 
relationship is recovered in our four-region analysis, 


Annals of the 
Missouri Botanical Garden 


with increased support values (BS 93%, PP 1.00). 
There was also strong support for the sister relation- 
ship between the well-supported Argocoffeopsis and 
Calycosiphonia clade and the remaining ingroup taxa 
in our Bayesian analysis (PP 0.98), but weak support 
for this relationship in the MP analysis (BS 50%). 
This relationship was also recovered in the strict 
consensus tree of Davis et al. (20 


TAXONOMIC NOVELTIES RESULTING FROM THE GENERIC 
RESURRECTION OF EMPOCONA 


An outline of an emended infrageneric classifica- 
tion for Empogona is provided below. It contains a 

rmal new combination for one of the two sections 
recognized. The outline is followed by a checklist of 
species, providing all necessary new combinations at 
the species level and below. 


OUTLINE OF AN EMENDED CLASSIFICATION FOR ÉMPOGONA 


Empogona Hook. f., Hooker’s Icon. Pl. 11: 72, t. 
1091. 1871. TYPE: Empogona kirkii Hook. f. 
oic a subgen. Empogona rn f} Robbr., Bull. Jard. 
t. Natl. Belg. 49: 259. 
he n a of 
hie 


genus ent 


subgenus Empogona 
259) remains applicable to the 


Empogona Hook. f. sect. Empogona. Tricalysia 
Empogona (Hook. f) Robbr. 
Empogona (Hook. f.) Brenan. 


subgen. sect. 


EMPOCONA KIRKII SPECIES GROUP 


This e ag 2 me group of Tricalysia a 
(Robbrecht, 1 


species num 


. The group compri 

2, and 24 in the de 
below. The position of nos ngalaensis (species 
22 below) was not confirmed by our moleeular 
analysis. 


EMPOCONA DISCOLOR SPECIES GROUP 


This corresponds to the group of Tricalysia dis- 
color e 1979: e group com- 
numbered 1, 4, 6, and 14 in 
the penne below. The group is only represented 


prises the species 


by Empogona acidophylla in the analysis, which 

falls in a elade corresponding to the previous species 

group. 

Section Empogona further comprises the three 

e numbered 10, 17, and 25 in the checklist 
ey were considered to be of isolated position 

Ra 1979: 300). 


concolor and 17. E. gossweileri) are included in ihe 


Two of these species (10. 


analysis. They have a basal position in the clade 
corresponding to section Empogona. 


Empogona sect. Kraussiopsis (Robbr.) J. Tosh & 
Robbr., comb. nov. Basionym: Tricalysia subgen. 
poco sect. Kraussiopsis Robbr., Bull. Jard. 
Bot. Natl. Belg. 49: 309. 1979. TYPE: Empogona 
crepiniana (De Wild. € T. Durand) J. Tosh & 


EMPOGONA RUANDENSIS SPECIES GROUP 

This corresponds to the group of Tricalysia 
ruandensis (Robbrecht, 1979: 310). The group com- 
prises the species numbered 8, 9, 19, 26, and 27 in 
the checklist below. 


EMPOGONA GLABRA SPECIES GROUP 


Ar s to the group of Tricalysia glabra 
du 1979: 


species, numbers 


292). This small group comprises 
16 and 23 in the checklist 


below. 


EMPOGONA CREPINIANA SPECIES GROUP 


This corresponds to the gro 
crepiniana (Robbrecht, 1979: 329). I 
speciose group comprising 11 species, dies 2; 
5, 7, 12, 13, 15, 20, 21, 28, and 29 of the 2. 
below 


up e Tricalysia 
is 2 most 


CHECKLIST OF SPECIES AND INFRASPECIFIC TAXA, 
INCLUDING Taxonomic NOVELTIES 


The list below, ordered alphabetically, enumerates 
all known taxa of Fior ee including the four 
species (species numbered 3, 7, 21, and 27 below) 
treated or described after ue s (1979) revision. 
The infrageneric assignment of the species is given in 
the preceding section of the present paper. Taxa 
preceeded by an asterisk (*) were included in the 
molecular analysis (see Table 1). 

The checklist includes taxonomic novelties for all 
dian i.e., 34 new combinations and three modifi- 
cations of infraspecific status. In his revision, 
Rebbrecht (1979) used varietal status for all infra- 
specific taxa due ee Here we reconsider the 
ropriaten of that treatment in 


app ap 
Rietz’s e, 1991) for distin 


guishing subspecies and TM Therefore, when 


criteria de cited in Sta 


infraspecific taxa are allopatric and differing in 
several features, we propose subspecific ms than 


varietal status. 


(*) 1. Empogona acidophylla (Robbr.) J. Tosh & 
Robbr., b 


comb. nov. Basionym: Tricalysia acid- 


Volume 96, Number 1 
2009 


Tosh e 
E of Tricalysia 


ophylla Robbr., Bull. Jard. Bot. Natl. Belg. 49: 
292. 1979. TYPE: Tanzania. Eastern Usambaras, 
2 mi. E of Sigi railway station, 27 July 1953, R. 
B. Drummond & J. H. Hemsley 3490 (holotype, 
K!; isotypes, B!, BR!, LISU!, S1). 


2. Empogona aequatoria (Robbr) J. Tosh & 
Robbr., comb. nov. Basionym: Tricalysia aequa- 
toria Robbr., Bull. Jard. Bot. Natl. Belg. 48: 465. 
1978. TYPE: [Democratic Republic of the Congo. | 
Congo belge. Yangambi, 4 Dec. 1937, J. Louis 
6887 (holotype, BR 
COU, EA!, HBG!, Kl, 
WAG!) 


isotypes, B!, BR!, Cl, 
MO!, P!, PRE!, UPS!, 


3. puru africana (Sim) J. Tosh & Robbr., 
om Pap Diplospora africana Sim, 
ie FL Cape, 238. 1907. Tricalysia africana 
oland, Egossa Forest, 

Aug. 1899, T. R. Sim 2386 (holotype, NU!). 


4. b odo bg curd Mero J. Tosh & 
Robbr., nov. calysia aula- 
cosperma | Robhr., em d Bot. a Belg. 49: 
296. 1979. TYPE: [Democratic Republic of the 
Congo.] Congo belge. Musenge, 20 Dec. 1958, A. 
Léonard 2088 (holotype, BR!; isotypes, EA!, K!, 
MO!, WAG). 


(*) 2 ds ox or (De Wild.) J. Tosh & 
obbr., comb. Basionym: Tricalysia be- 
s De Wild. "m Bun 3: 157. 1925. 
TYPE: [Democratic Republic of the Congo.] 
Congo belge. [Kisangani] Stanleyville, Tshopo 
River, 25 Feb. 1915, J. Bequaert 6969 (holotype, 
BR!) 


6. Empogona bracteata (Hiern) J. Tosh & Robbr., 
comb. nov. Basionym: Tricalysia bracteata Hiern, 
Fl. Trop. Afr. [Oliver et al.] 3: 120. 1877. TYPE: 
[Guinea.] Senegambia. Karkandy, s.d., Heudelot 
855 (holotype, K!). 


7. Empogona breteleri (Robbr.) J. Tosh & Robbr., 
comb. nov. Basionym: Tricalysia breteleri Robbr., 
Bull. Jard. Bot. Natl. Belg. 51: 166. 1981. TYPE: 
Gabon. Moanda-Franceville Km 23, 12 Se 

, F. J. Breteler 6431 (holotype, WAG!; 
isotypes, BR!, P!) 


8. Empogona buxifolia (Hiern) J. Tosh & Robbr. 


8a. Empogona buxifolia (Hiern) J. Tosh & Robbr. 
subsp. buxifolia, comb. nov. Basionym: Trica- 
lysia P. Hiern, Fl. Trop. Afr. [Oliver et al. l 
3:119 7. TYPE: Angola. Ambriz, Nov. 1872 
J: mund s.n. (holotype, K!; isotype, W!). 


8b. Empogona buxifolia subsp. australis (Robbr.) 
J. Tosh & Robbr., comb. et stat. nov. Basionym: 
Tricalysia buxifolia var. australis pee Bull. 
Fon Bot. Natl. Belg. 48: e 918: TYPE: 
gola. Tchivinguiro, 13 Dec. 1, G. Barbosa 

a (holotype, LISC!; e col K!, LUAT!). 


(*) 9. Empogona cacondensis (Hiern) J. Tosh & 
Robbr., comb. nov. Basionym: di cacon- 
densis Hiern, Cat. Afr. Pl. (Hiern) 1(2): 46 
1898. TYPE: Angola. 
fortress near Ferá 
3112 (lectotype, designated by Robbrecht [1979: 
320], LISU!; duplicates, BM!, COI, K!). 


(*) 10. Empogona concolor b Hallé) J. Tosh & 
Robbr., comb. n ym: Tricalysia con- 
color N. Hallé, Fl. Gabon 17. 283. 1970. TYPE: 
Gabon. Bélinga, mine de fer, 21 July 1966, N. 
Hallé & A. Le Thomas 119 (holotype, P!; isotypes, 
K!, P!) 


11. a am coriacea (Sond.) J. Tosh & Robbr., 

. nov. Basionym: Kraussia coriacea Sond 

ine "98: 54. 1850. Tricalysia sonderiana 
Hiern, Fl. Trop. Afr. [Oliver et al.] 3: 119. 1877, 
replacement for Kraussia coriacea Sond., non 
Randia coriacea Benth., Niger Fl. [W. J. Hooker] 
387. 1849 [=  Tricalysia coriacea  (Benth.) 
Hiern]. TYPE: E Africa. KwaZulu-Natal:] 
Natal: Durban, s.d., W. Gueinzius 100 (holotype, 
Wt; isotypes, iw "C. K!, PRE!, S)). 


12. Empogona crepiniana (De Wild. & T. Durand) 
& 


J. Tos obbr, comb. nov. Basionym: 
Tricalysia crepiniana De Wild. & T. Durand, 
Ann. Mus. Congo Belg., Bot. ser. 3, 1: 120. 1901. 

: [Democratic Republic of the Congo.] 
Wangata, b. 1896, A. Dewévre 740 


(holotype, BR!; isotype, Cory. 


13. Empogona pe icd Dus J. Tosh & 
Robbr., comb. nov. Basionym: Tricalysia deight- 
onm i Brenan, Kew Ball. Fd 112. 1953. TYPE: 

Jama, 10 Mar. 1948, F. C. 

Da uh 4723 CUN Kt; isotype, P). 


14. Empogona discolor (Brenan) J. Tosh & Robbr., 
comb. nov. Basionym: Tricalysia discolor Brenan, 
Kew Bull. 2: 72. 1947. TYPE: [Ghana.] Gold 
Coast. Mampong Scarp, Feb. 1933, c Vigne 
2748 (holotype, K!; isotype, MO?). 


15. Empogona filiformistipulata (De Wild.) Bre- 


mek. 


15a. Empogona filiformistipulata (De Wild.) Bre- 
mek. subsp. filiformistipulata, Bot. Jahrb. 71: 


Annals of the 
Missouri Botanical Garden 


201, 222. 1940. Basionym: Pepe Dd 
Pl. Bequaert. 3: 211. 
C i os s in Wild. 
Kew 953. TY 


stipulatum De Wild., 
) Brenan, 

E: [Democratic 
sae e is a Congo DE Kisangani, 
Tshopo River, 12 Jan. 1915, J. Bequaert 6580 
(holotype, BR!; isotype, K not seen). 


15b. Empogona filiformistipulata subsp. epipsila 
(Robbr.) J. Tosh & Robbr., 


Basionym: 


comb. et stat. nov. 
Tricalysia i Dum (De 

Wild.) Brenan var. epipsila r., Bull. Jard. 
Bot. Natl. Belg. 48: 465. 1978. TYPE: [Demo- 
cratic Republic of the Congo.] Congo belge. 
Yangambi, Feb. 1933, J. Louis 14233 (holotype, 
BR; isotypes, COI!, K!, MO!, P!, WAG?). 


16. Empogona glabra (K. Schum.) J. Tosh & 


Robbr., comb. nov. Basionym: Tricalysia glabra 


A 
Ie] 
> 
z 
B 
irj 
$ 
= 
5 
S 


3 , F. Welwitsch 
3117 (holotype, LISU!; isotypes, BM!, C!, COL, 
P!) 


1 
t 


(*) 17. Empogona gossweileri (S. Moore) J. Tosh 
& Robbr., comb. nov. Basionym: Tricalysia 
Hed : Moore, J. Linn. Soc. Bot 37: 305. 
1906. E: Angola. Cuanza Norte, Cazengo, 
1903, h EE 688 (holotype, BM!; isotypes, 
Kt, Pb. 


18. Empogona kirkii Hook. f. 


18a. e kirkii Hook. f. subsp. kirkii, 

Pl. 11: 72, t 1091. 1871. 
Tricalysia junodii (Schinz) Brenan var. kirkii 
(Hook. eh Robbr., Bull. Jard. Bot. Natl. Belg. 49: 
271. 1979. TYPE: Malawi. Cape Maclear, Oct. 
1861, J. em s.n. (holotype, K!). 


Hookers Icon. 


Empogona allenii Stapf is the only species validly 
published in the genus Empogona not taken up as a 
result of the present study. It is a synonym of the 
present taxon (Robbrecht, 1979: 272). 


(*) 18b. Empogona kirkii subsp. junodii (Schinz) 
J. Tosh & Robbr., comb. et stat. nov. Basionym: 
Empogona inod Schinz, Mém. Herb. Boiss. 10: 
67. 1900. Tricalysia junodii (Schinz) Brenan, 
id Bull. 2: 60. 1947. TYPE: Mozambique. 

e Laurengo Marques (Delagoa Bay), s.d., 
p an p Z!). 

(*) 19. Empogona lanceolata (Sond.) J. Tosh & 
Robbr., comb. nov. Basionym: Kraussia lanceo- 

Sond., Linnaea 23: 53 ] 

lanceolata (Sond.) Burtt Davy, Ann. Transvaal 


lata O. Tricalysia 


Mus. 3: 122. 1912. TYPE: [South Africa. 
KwaZulu-Natal:] Natal: Durban, W. Gueinzius 
68 (lectotype, designated by Robbrecht [1979: 
313], W!; duplicates, P!, S!). 


20. Empogona ages lead (K. Schum.) J. Tosh & 
Robbr., comb. nov. Basionym: Tricalysia macro- 
phylla K. Schum., Bot. Jahrb. Syst. 28: 66. 1899. 
TYPE: Cameroon. Bipinde, Zenker 1569 (lecto- 
type, designated by Robbrecht [1979: 339], 
COL; duplicates, BM!, BR!, COI, El, G! 
GOET!, HBG!, L!, M!, MO!, P!, St, Wt, WAGI). 


21. Empogona maputenis (Bridson & A. E. van 
Wyk) J. Tosh & Robbr., comb. nov. Basionym: 
Tricalysia maputensis Ena & A. E. van Wyk, 
Fl. Zambes. 5(3): 475. 2003. TYPE: Mozambi- 

. Matutuine, 8 Abus: 1957, L. A. G. Barbosa 
& A L. de Lemos 7807 (holotype, LISC not 


seen). 


(*) 22. Empogona ngalaensis (Robbr.) J. Tosh & 
mb. nov. Basionym: Tricalysia nga- 
, Bull. Jard 
277. 1979. TYPE: Malas. North N 3 
N of Chilumba, 17 Dec. 1969, J. Pawek 3095 
(holotype, K!). 


23. Empogona nogueirae (Robbr.) J. Tosh & 
Robbr., i 
gee m 
466. TYPE: Angol 
1966, P a Teixeira 10701 (holotype, LISC!; 
isotype, COM). 


24. Empogona ovalifolia (Hiern) J. Tosh & Robbr. 


(*) 24a. d ovalifolia (Hiern) J. Tosh & 
obbr. var. ovalifolia, comb. nov. Basionym 
Tricalysia ovalifolia Hiern, Fl. Tröp. Afr. [Oliver 
et al.] 3: 119. 1877. TYPE: brine ] Zanzibar: 
c, s.d. [ace. K Sep. 1868], J. Kirk s 
(ecos, designated by Fo en 3391, 
KI). 


24b. Empogona ovalifolia var. eie Ls A 
Tosh € Robbr., comb. nov. Basionym: Empog 
kirkii dup f. var. glabrata Oliv. Trans. Linn 
, 2: 336. 1887. Tricalysia a 
Hiern var. . es (Oliv.) Brenan, Kew 
58. 1947. TYPE: Kenya or Tanzania. a mi. 
from coast, [1884], H. H. Johnston s.n. [Kiliman- 
jaro Exp.] (holotype, K?). 


Soc., 


24c. Empogona ovalifolia var. taylorii (S. Moore) 
J. Tosh & Robbr., comb. nov. 
Empogona taylorii S. Moore, J 
1925. Tricalysia ovalifolia Hiern var. taylorii (S. 


Volume 96, Number 1 
2009 


et al. 
Phylogeny of Tricalysia 


Moore) Brenan, Kew Bull. 2: 59. 1947. TYPE: 
Kenya. Giriama, Oct. 1887, W. E. Taylor s.n. 
(holotype, BM). 


25. Empogona reflexa (Hutch.) J. Tosh & Robbr. 


25a. lb Sati reflexa (Hutch.) J. Tosh & Robbr. 

reflexa, comb. nov. Basionym: Tricalysia 

ne Hutch., Kew Bull. 1915: 44. 1915. TYPE: 

ierra Leone. Kessewe, 17 Apr. 1913, C. E. 
Lane-Poole 131 (holotype, K!). 


25b. Empogona reflexa var. ivorensis (Robbr.) J. 
Tosh & Robbr., comb. nov. m A NC 
reflexa var. ivorensis Robbr., : 
Natl. Belg. 48: 466. 1978. E p^ Coast. 

of Niapidou, 20 Jan. 1959, A. J. 
Leeuwenberg 2500 (holotype, WAG!; 
K}). 


© 
= 


isotypes, 


(*)26. D ruandensis A ) J. Tosh & 


omb. nov. Basionym: Tricalysia ruan- 
pene leno. Bull. Jard. Bot ms Bruxelles 
26: 253. 1956. TYPE: [Rwanda.] Mayaga, 


Mutema, 19 May 1954, L. Liben 1416 (holotype, 
Ut; isotypes, BR!, WAG). 


27. Empogona somaliensis (Robbr) J. Tosh & 


Robbr., comb. nov. Basionym: Tricalysia soma- 
liensis Robbr., Bull. Jard. Bot. Natl. Belg. 56: 
149. 1986. TYPE: Somalia. 17 km W of Badade, 


30 June 1983, J. B. Gillett, C. F. Hemming, R. M. 
Watson & H. Julin 25153 (holotype, K!). 


(*) 28. Empogona talbotii (Wernham) J. Tosh € 
Pa comb. nov. ua a tal- 
Wer 913. 
Tricaba Br las Keay, Fm due 
Bot État Bruxelles 28: 291. 1958. TYPE: 
Nigeria. Southern. Nigeria, Oban, 1911, P. A. 
Talbot 287 (holotype, BM!; isotype, K). 


29. Empogona welwitschii (K. Schum.) J. Tosh € 
Robbr., comb. nov. Basionym: ann wel- 
witschii K. Schum., Bot. Jahrb. E a 
1897. TYPE: Angola. Near Ponte de Felix 
Simões, Apr. 1855, F. Welwitsch 3106 (holo- 
type, LISU!; duplicates, BM not seen, COI!, K!, 
P?) 


CONCLUSIONS AND FuTURE DIRECTIONS 


We have been able to demonstrate that the two 
subgenera comprising the large Afro-Malagasy genus 
Tricalysia do not form a monophyletic group an 
should be treated as separate genera. Empogona has 
been previously recognized at generic rank, and 


subsequent authors have considered reviving its 
On the 


evidence, it is now fully justified to revive Empogona 


generic status. basis of our molecular 
at the generic rank. The Asian genus Diplospora is 
sister to Empogona, with both genera forming a 
strongly supported monophyletic group. As a conse- 
quence, the weakly supported Asian clade reported by 
(2007) is not 
investigation. Further data are still required to fully 


Davis et al. recovered in this 
elucidate the phylogenetic relationships between 
Belonophora, Diplospora, Discospermum, Empogona, 
Sericanthe, and Tricalysia. There is increased support 
for the placement of a Coffea and Psilanthus clade as 
sister to the rest of Coffeeae. 

Future work requires the inclusion. of nuclear 
ribosomal and low copy nuclear DNA sequence data, as 
well as expanded taxon sampling, in an effort to improve 
resolution between terminal taxa within the genera 
ee and E It seems prudent to defer 


n on the biogeography of Tricalysia and 
Empogona u until we have a broader sampling and a more 
resolved phylogenetic hypothesis of both genera. 


Literature Cited 

E. Robbrecht. 1991. Remarks on the tropical 

tralian taxa included in Diplospora or 
(Rubiaceae-Ixoroideae-Gardenieae). Blumea 

05. 


Andreasen, K. & B. Bremer. 2000. Combined phylogenetic 


lysia DC. dum Bull. 1947 53-63. 

ubiaceae. Pp. 379— 

120 in G. Y. Pope Mu un Flora Zambesiaca, Vol. 5(3). 

as Botanic Gardens, K 

Cheek, & Dawson. 2000. À synoptic revision of 
po. ue Kew Bull. 55: 63— e 

Davis, A. P., Chester, O. Mau 2007. 
Searching x the relatives of Coffea ERU Ixo- 
raoideae): The circumscription and phylogeny of Coffeeae 

based e x p data and morphology. Amer. J. 

4:3 


oe E Doyle. 1987. A rapid DNA isolation 

small oo of fresh leaf tissue. 

e aes Bull. 19: 11-15. 

Erixon, P., B. Sy ol T. Britton e B A buda 
Reliability istrap 
Po in phylogenetics Syst. Biol. 52: eae 3. 

Felsenstein, J. Phylogenies and the comparative 

method. Amer E 125: 1- AS 

Hooker, J. D. 1873. Rubiaceae. Pp. 1 in G. Bentham € 
J. D. Hooker ed Genera Y Vol. 2(1). Lovell 
Reeve Obs 

Huelsenbeck, . Ronquist. 2001. MRBAYES. 
Bayesian iieri of phylogeny. Bioinformatics 17: 
eee 


——— Lar, R. E. Millr & F. Ronquist. 2002. 
wen o and ge of Bayesian inference of 
phylogeny. Syst. Biol. 51: 673—688. 


Annals of the 
Missouri Botanical Garden 


Jordan, W. C., M. W. Courtney & J. E. Neigel. 1996. Low 
levels of Tues CRM variation at a ra pidly 
evolving chloroplast DNA loc North American 
duckweed t Amer. i Bot. 83: 430—439. 

Keay, R. W. J. 1958. Notes on Rubiaceae for the un of 

West Tropical Africa, 2nd ed. Bull. Jard. Bot. Etat. 28: 

297-298. 


Lóhne, C. & T sch. 2005. Molecular dins d and 
phylogenetic Moe of the petD group II i : A case 
study in basal angiosperms. Molec. Biol. "Evol. 22: 


317-332. 
Maddison, D. R. & W. P. Maddison. 2002. MacClade 4: 
es of phylogeny and character evolution, version 
r d E 


1. Si maue: 
ode ul M. 1994. Phylog 


Nuclear Ribosomal DNA Sequences, Secondary Chemis- 
try, and Morphology. Ph.D. Dissertation, Üniver rsity of 
Texas, in. 

Persson, C. 2000. ew of bob Med (Rubiaceae) 
based on chloroplast DN 


and irnL(UAA)- -F (GAA) pua spacer. Nord. J. Bot. 
Posada, D : andall. 1998. Modeltest: Testing the 
model of DNA Sees Bioinformatics 14: 817-818 


andrianina . De B 


D. J. Crawford. 2005. Bayesian 
iuvenem of phylogenetics revisited: Developments and 
concerns. Taxon 54: 9-15. 
1978. Sericanthe, a new African genus of 
Rubiaceae (Coffeeae). Bull. Jard. Bot. Nad. Belg. 48: 3-78. 
1979. The African hos Tricalysia A. Rich. 
(Rubiaceae Coffeeae). 1. A r n of the species of sub- 
genus Empogona. Bull. Jard. Bot ‘Nad. Belg. 49: 239-360. 
2. The African genus Tricalysia A. Rich. 
(Rubiaceae: Coffeeae). 2. Ephedranihera, a new section 
of subgenus Tricalysia. Bull. Jard. Bot. Natl. Belg. 52: 
311 EC 


1983. The African genus Tricalysia A. Rich. (Rub- 
iaceae). 3. Probletostemon revived as a section of subgenus 


Tricalysia. Bull. Jard. Bot. Natl. Belg. 53: 299-320 


——— 1987. The African genus Tricalysia A. Rich. 

(Rubiaceae). 4. A revision of the species of section 
Tricalysia and section Rosea. Bull. Jard. Bot. Natl. Belg. 
57: 39-208. 


. 1988. Tropical woody Rubiaceae. Opera Bot. Belg. 
1: 1-271 


86. A survey of the Gardenieae 
and related tribes (Rubiaceae). Bot. ee Syst. 108: 
63-137. 


Manen. 2006. The major evolutionary 
lineages of the coffee family (Rubiaceae, angiosperm rms). 
A 


Fr d and Rubioideae. Syst. & Geogr. Pl. 76: 
85-1 


Ri F, J. P. Huelsenbeck & P. van der Mark. 2005. 
MrBayes 3. l manual. «http: ES E 
1_manual.pdf>, accessed 30 Octob 

. Rubiaceae. Pp. 1— pe in A. Engler & K. 

Prantl (editors), Die naturlichen Pflanzenfamilien, Vol. 4 

(4). Wilhelm ee bun Leipzig. 

y. J. n. n. 93 AGE NW. Liu, J 

Winder, E. E. Schi 

- The ded sind the hare II 


Schumann, K. 1891 


s as characters 
in sequence-based phylogenetic a Syst. Biol. 49 
69-381. 
Stace, C. A. 1991. Plant Md and Biosystematics, 2nd 
ed. Edward Arnold, L 
Staden, R., K. Beal & J. Boni e E a Package. 
15-130 in S. Mis & S awetz (editors), 
Compute Methods in Mise an Humana Press, 
ES D. i" 2003. PAUP* 4.0b10: Phylogenetic Analysis 
Usi other methods). Sinnauer 
h 


la DNA. PL. Molec. Biol. 17: 1005-1109 


211 


Tosh et al. 


Volume 96, Number 1 


2009 


Phylogeny of Tricalysia 


€LPOOONV  ZES66O6INV STPOOOINNV — LSTLL TINA viquiez (Ya) 2601 urossaq "aqqoy (WOH) sisuasopun oyrunoLas 
Tés0gTÓG ZSSO8SIÒA Y9TP666NV  ZZSOSTÒA viuezuv], (y) gopg poospig "aqqoy (WOH) sisuadopun oyrunoLos 
ossoslOd . @sso8TOd VN — LIS081ÓG viuvzuv], *(y) cp Sunay "pur "ds puomjdskjoq 
ZOLESIO = PPOSSTO — vIvP666NV  —— SoSESTOd vruezuey, (3) gzpp v4sty uospug yaswas smpruvpisq 
ZLV666INV  I€S666INV  £TP666INY — £ZO09TIAA ¿PUBYo (Ya) gz ayssaquayan) una "FOOLY muma smpunpisq 
ILT666]|NV OSS666NV ZLP666NV  Z6L666INV 419907) Aoa] (Y) FSOES PUNT uror] $170/0919D.2 empnipjisq 
TOP666NY 8IS666NV P6S666NV PZO09TINA Ieosesepe (Ya) g9or I? 12 «sog, “MOL 190/13 vaoxg 
£9P666INV — [cS666INV VN  0Z009TINA TeoseSepe]y (Y) 9853 Sang qopur ds emprupopaod&q] 
C9V666l!NV — 61S666INV VN — 6I909TI TeoseSepe]y O) PEGE sang sued q "V 9 "Seuooey (oxexq) mord smpupopnzod&g 
TOT666lNV OZS666NV VN  Z8£666WV 1eosesepe (Y) rage 4op&oo uospug (ummos y) snpapdoaonu snypunonsedd y 
egsoslOd  6rsoe Od VN — PLSOsltod Poy AS (5) £026z-1961 P 12 siaoq g 7p Paqu vruopiz) 
zesoslOd . ereogrÓd VN — &IS08TÓG (sonsupoy) sousieosela (5) 6gog wumupong "OPISA ($ Heg) Sn noo Duo 
OZFP666INV 62S666NV ITP666NV —— 06£666INV (rounrg) oourog (Y) OF SOT pmusp "pur “ds umuaodsoosit 
69T666INV  8ZS666NV OLP666NV¥  — 098€666INV (uejueunpey) oourog (Y) gp 1g PAPIS aqqoy Y UV fos CHoy) ouaouqp wmunodsoosiq 
OIS666]NV  LZES666NV 6O0P666NVY  688666NV (rung) oouxog ‘(y) 2£35] mysunyy "jopur “ds niodsojdig 
08S08TÓd  9FSO8TOd VN  TISO8TÓG (rounrg) ooulog (NM) ggge] waag "joput “ds vaodsojdig 
89FP666INV 9ZS666NV  SOPGGGINV  88£666INV AUL) 6F ayssaquayany un, "wese (purt) vignp psodsojdigy 
SOP6O6ONV ZZS666NV  S6S666NV  TZ909INA ¿OMQPQUIZ (YE) Z9 aysuaquayan;) UVA Avoy (uu&ng) on40u xuidposoupig 
698€ESIÓA — IS2e610d  LOP666INV  ZOSESIÓAU TeoseSepe]y 0) ozgg Sang "SPUOJONPY Y SIAP oq y xo ore "qo nmmaour pafon 
OL8ESTOd  ZSLESTOd  9OP666NY  £OSESTÓU TeoseSepe]y (Y) pp ojosnuogoqua SHO sisusososunu vaffory 
69LESTÒA  TS9ESTÓAU VN — corestod TeoseSepe]y (Y) cogg Sang kore] "47 f 22]jow01 vaffo;) 
6.908100  €29666INV VN — OIS08TÓG TeoseSepe]y 0) zzzg sang oput “ds vuoydauny 
82508104  =FrSO8TOd SOF666INV  — 60€0810d viuvzuv], (SN) 936% sq “1qqoy Cunuos y) x4monpmds pruoudisoo£qp7) 
SLOO8IOG  TPSO8IÓ  vOF666INV 90S08TOd uooxourer (N) oggg sang aqqoq (UNOS 7x) séump[o0400ut vruoudisooAqpo;) 
9LG0810d — crcogrOd VN — 20e0810d uoorourer (SN) ppgg sang aqqoq (UNOS 7x) séump[o0400ut vruoudisooAqpo) 
PLsoslOd  6gS08TOd VN X POSOsTOd uoo1oure") (Ss) 210€ Sq qopur "ds varag 
19V666INV SZS666NWYV 868666INV  ZZO0YTINA «Uoqeo (HE) Op eysuequayany una osnery “yl ssuomnp DLL 
99P666NV —— VcG666lNV —— L6£666IN V VN «Uoqeo (HE) rp eysuequayany ung uror] »40]fia24Q DIAS 
ZLSOBIO  ¿£SOBIÓU — 96£666WVv  zOSo8Idd uooxourer (SN) 190€ sang STH ^N. Cumypg 7x) »mjjademo1q vaag 
TzsoslOd  9£S08IÓU  £€OP666NY TOSO8TÓU uooxoure7) (sy) pap nomofpn “opur “ds nioydouojag 
OLSOSTOd SESOSIÒA  ZOP666NY ^ 00S0810G uoorourer) (SN) 67 UMD ƏJLoH na2n1402 nsoydouojag 
69S081Od = PESOBTOA  TOP666NY ^ 66 PORTO uoorowen ($) c UMD Y aso] na2n1402 a01douojog 
98950810d . €£50810d 00r666IV  s6rosTOd uooxourer (N) 910€ sang unqoq ('umuos y) suapupos sisdoaffooos1y 
1950810d  zESO8IÓU VN — 96F0810d eqqudeg uvongy [enue)) (y) 8978 Sawy — "1qqoq (unigo) 222uuouz “dsqns 1qqoy (wory) suasodna sisdoaffooosiy 
995081Ód I£S0810d  66€666NNV Z6T0810d uoorourer (SN) 750g sang "aqqoq (wequo m) sisuaraya sredoaffooosy 
Aqua ord: quad posd-qaan Ino A uoxer, 


"eyep uorssooo? pu? 1oqgonoA uoxe[ ‘g Xrpuoddy 


Annals of the 


212 


Missouri Botanical Garden 


SOS666NV 80909 TINA OSP666INV S7£6661NV Pueyo *(s) 617 spiunjog ulem] sisuajayo DISÁDO |, 
TOS666NV —— 209091IN4 — 6PP666lNV  PLEGG6INV viuezueT *(x) 996% poo?pig "qqoy sisuanyosu viskqpona], 
00S5666INV —— 9090014  SPPOGGINV  £LEGGGINV vf£uoy (ug) sor 120g 24 uxorg nj ajdoaonui DISAP], 
66P666INV  SO909TWA — LPT666INV — 2L£666INV reose3epepy (Mg) 86€ "I? 19 [so], POUT 9q  "quieupueq (Preg) odapoo2no] DIADH], 
96F666INV —— TOOO0TINA  OPPGGGINV  TZEGG6INV eosvSepe]y 0D) Epps 421mp2 xoopg oq Y “quenpuey (preg) odon] viskpoouy 
L6V666INV  O909TINA — SPP666INV OLE666IN'V vougy ymos (y) 6151 ssrsog £seq nmg (puos) omjoooup] viscmonur 
O96F666INV  ZO909TINA — TPPO66INV —— 69£666INV «9Mqequirz (Hg) ez ausdaquasgap;) UDA uvue1q (zumos) rpounf vis&poni y 
S6F666INY TO909IWA — €PE666INV — 89£666INV eeN OD gp nupky WOH xo “$ 1007 29 “weg (qoszio[y) 4ojfrunuspf vis£qpona], 
€6F666INV ^ 666091. — IPP666INV ——99£666INV viquiez (49) sog uressaq "umos ^y DOPISU DISÁJDOL y 
T6F666INV —— 00900114 ZPP666NV  ZOS6G6INV e1quez (Ya) pror utessaq umog ^N ?40jfiosius oiskqmonr 
c6V666]A V 986809 TIAA OTT666IN V S0£6661NV vous) perioyenbq *Tppp 2421u25 9100] `S 119298803 DISÁJDIL T, 
T6v666IWNV LoSO9TINT — 6t£T6606INNV — T9£666]NV vui?) *(q) 9081 puryBuor Peq 9 Y9Mp [UNY Y) oro DIADILI, 
O6P666NV  9GSOOTINA  SEPGGGINV  £OLOG6INV eosvSepe]y 0) erra pampoaaqnay Yo 9q Y "quieupuey sisuaurydnnp DÁIN], 
68T666INV S6SOOTNA  OEPGGGINV — c9£666lINV eosvSepe]y *(g) opg T9 12 so], 9019 9q Y "quieupuey sisuaurydnnp viskponi 
88P666NV  PGSOOTINA  OEPGGGINV —— I9£666IAV eosvSepe]y (4a) p69 12014 20 Yo 9q Y “qUELIpuey sisuaurydnnp viskponi 
L8V666lNV  £6S09TMA — S£FP666INV  0O9EG66INV eosvSepe]y (48) ZZE T9 12 YsoL ioxeg xqm201d(40 piskpoo1 I, 
98t666INV  Z6SO9TINA  PEPOOONVY  6SE£666MNV reose3epepy (Ma) ZES 19918 90 xexeg xqpooidí40 DISAJDOL Y 
PEP666AY O6SOOTNA  ZEPGGGINV — 25£666IAV viquez (yg) 6581 urossaq WOH ("puog) »220102 DISÁJDOL Y 
S8F666lNV  T6SO9TMA EEP666NV  SSEG6GINV viquez (yg) ggg “198827 WOH ("puog) »200102 vis4jmoug 
£8F666lNV 68SO9INA — I£P666INV  9ISEG6GINV uoqe? (ua) s6 Pasaq TPH ^N 40702409 DISÁ]D1LJ 
Z8POOONV S8SO9INA — O£T666INV  SSEG6GINV viquie7 (Ya) 1801 utessaq UXoTH sssuapuoono DISÁJDOLL, 
I8F666INV Z8SO9TINJ — 6ct666lNV  PSE666NV v£uoy (UD) 68€ 120g 24 “1 QOY pupmospuq oisAmong 
O8TP666INV 98S09TNA  STPOGGINV  £SLGOG6INV uoqeo (OIN) ZrO sono "PIA. 9q auaonbaq viscmou 
6LV666INV S8SO9TNA — Zcb666INV — cS£666INV 03u09 eu jo or[qndey onexooureqq (Y) ggg essmpopy ALO] `S 1000/82Dq vIskpoory 
S6S80810d 09S5081Ód  9cr666WNV 9zS08TOd uoorourer (N) Spog sang "aqqoq smuaoums "es oonig ^V H D[DULOUD DISÁJDIM J, 
PIS666NV P8SO9TINGA SCT666IN V ISE666NV TeoseSepep (UD) PLOT I» P 99018 aq 19018 Oq 2 “QUBLIPUBy xo o[[ouiog] szsuauzopzpumpup DISÁJDIL J, 
8LP666WV  £8S09TINA  PIPGGGINV  OSSOG6INV 1eosesepela (4g) 11 7? 1 YsoL YO eq Y "quieupueq xe e[ouroj] $1842JODZDULD]DUD DISAJDIL Y 
LLVOOONV  c8S09IIN4 EZP666NV  6DEG6GINV qwosesepe 4a) ETET 120g 24 xoo[g eq  "quieupue sisvasquen DISÁJPIH I, 
EIS666WV I8SO9INA ZZP666NV  SPEG6GINV ERN (6D) 03€ mununig UMY "N SOP1019YIUDIOID DISÁIDOA[ 
OLV666INV  LES666NY — Ict666INV — LP£6661NV viquez (yg) ZrZr urssoq UMY “SN SOP1019YIUDIOID DISÁIDOA[ 
ZISG66NV  O£S666NV  OZP666NV  9IVEGOGINV vruezue] (YO) GEL Pepu ""qqoy D7 Mydopi9o DISAP], 
PLPOOONVY PESOOONVY S8I1T666lNV — TVVt666lNV vruezue], (5) 1202 equ "qqoy 24032100 visAqp2t4T 
GLP666NV SES6O6NV  6IP666NV  SPS666INNV viuezueT *(x) STET6 mojorpunpy aqqoy DLOJf MOD DISÁJDIM y 
ITS666NV — $€9666lNV LIFO6ONV  T6S666INV uoqeg (DVM) O9TE PoquasmA "put “ds ayyunoriag 
Z6SO8TOd VN VN  £zS08IÓUd (oyorg) spuv[sp eeumz jo Jma (5) 691p ouppaa)) "qqoq (TPH ^N) smafovf anmouag 
A Tue 9rjda quod ]osd-qqoon Iono A uoxer, 


penunuo) <q xipuoddy 


213 


"umnueqrog Suoy Suo Aq umis[og jo uspes oruejog [euoneN 01 uoAI$ penou SurArT "uwouxun utsu() + 
"e[qe1 eu ur USATS urgrro Jo ATMO) umg Jo upea omuejog [Euonew jo suono Furs OY) uio; poao? SISYINOA pue [errojeur PT y 


Phylogeny of Tricalysia 


Tosh et al. 


60S5666INV 8T9090TIN4 O9P600NV  98E666MV PLOBIN (NM) #2929 07007 Loy (Weu A) 1109]01 DISÁJDOL I, 
80S666WV  LZT909IWA  6SPOGONVY  S8E66G6NWV vruezue], (5) E767 poospig "qqoy miegaros viskqpona], 
L1S$666I!NV | 9[909[IN4 — 98ST666IN V . T8t666INV vruvzue, (dg) £cgac yny "JOWA SISUSPUDNA misKqoonag 
L0S666INV | SI909[IN4 LSPO66INV €8So66INV reose3epepy (4a) 997 19918 oq xoo[g eq 9 "quieupueg xo o[pourogg 14214d viskqponag, 
90S5666INV <sI909TINT — TSTO66INV  6LE666NWV veuoqrg (y) reg suopy UIH suagynd visépmonag 
9I[S666INV  PI909TIAA — 9ST666IWV  T8E666NWV eique (Mg) £56 urossaq uxorg suogod DISÁJPIM I, 
S15666INV | €I909[IN4 — SST666INV T[8S666INV eique (E) 9931 urossoq uxorg suogmd visémonig 
c0S5666INV OT909TINT — IST666INV LLEO6OINV vruezue], (5) 60€ pg um »jofimao iiskmon 
£05666INV | 60909[IN4 — cST666INV 9LS666INV 1vosesepela (Mg) 0607 I? 19 99918 9q WIH DIOfIJDAO viskmou 
T0S6660INV TI909TINT — €ST666IWV  8LE666NWV Teose3epepy (Mg) 2201 "I? 19 9908 og WOH DIOfIJDAO viskmou 
Jua 9rjda quod Tosd-qqooo Ino A uoxe, 


Volume 96, Number 1 


2009 


Penunuo) ‘q xipuoddy 


Acknowledgment of Reviewers 


The following individuals are thanked for their Pete Lowry 
collegial reviews in 2008. This peer commitment of Martin Lysak 
time and effort is sincerely appreciated by the Annals. Jinshuang Ma 


Nélida Bacigalupo 
Anders Barfod 
Rainer Bussmann 
Daniel Caden: 
Christopher Camphell 


Friedrich Ehrendorfer 
Peter Endress 
Per Ericson 
Richard Field 
Tarciso Filgueiras 
Victor Finot 

e 


Laurent Gautier 
Diego Giraldo- Cañas 
Liliana Giussai 

Socorro cam Elizondo 
Alan Graham 

Richard Halse 

Anna Hersperger 
Henry Hooghiemstra 
John Janovec 

Iván Jiménez 

Peter Jørgensen 
Walter Judd 
Margaret Koopman 
John Kress 

Kathleen Kron 
Gwilym Lewis 
Gabriela Lopez 

Jon Lovett 


Vidal de Freitas Mansano 
Ligia Matias 

Gordon McPherson 
Robbin Moran 

Tim Motley 

Jesús Muñoz 

Jens Mutke 

Sachiko Nishida 


aría Planchuelo 
Nina Probatova 
Sylvain Razafimandimbison 
Renata Reinheimer 


Pavel O. Rodríguez Vásquez 
Zachary Rogers 

Jens G. Rohwer 

Alicia Rotman 

Andrew Rozefelds 

Kalle Ruokolainen 


Fátima Regina Gongalves Salimena 


Wendy Silk 
Peter Stevens 
Charlotte Taylor 
Hans ter Steege 


Maximilian Weigend 
Peter Wilf 

Kenneth Wurdack 
George Yatskievych 


—Votis gratias agamus. 


ANN. Missouni Bor. Garp. 96: 214. PUBLISHED on 23 ApriL 2009. 


Volume 96, Number 1, pp. 1-214 of ANNALS or THE MISSOURI BOTANICAL GARDEN 
was published on April 23, 2009. 


www.mbgpress.org 


CONTENTS 


Third International Rubiaceae Conference: Introduction 
Petra De Block, Charlotte M. Taylor € Suzy Huysmans 
A Review of Molecular Phylogenetic Studies of Rubiaceae —— — — < — — < — Birgitta Bremer 
Revisión a de Galianthe Subgen. Galianthe (Rubiaceae: Spermacoceae), con una 
Sección Nueva Elsa L. Cabral 
Phylogenetic Placement of the Tribe Retiniphylleae Among the Subfamily Ixoroideae 
(Rubiaceae) Rocio Cortés-B., Piero G. Delprete & Timothy J. Motley 

A Global Assessment of Distribution, Diversity, Endemism, and ‘Taxonomic Effort in Rubi- 
Aaron P. Davis, Rafaël Govaerts, Diane M. Bridson, Markus Ruhsam, 
Justin Moat & Neil A. Brummitt 

Taxonomic History, Morphology, and Reproductive Biology of the Tribe Posoquerieae (Ru- 
biaceae, Ixoroideae) Piero G. Delprete 
Fossil Record of the Rubiaceae Alan Graham 
Phylogeny of the Herbaceous Tribe Spermacoceae (Rubiaceae) Based on Plastid DNA Data 
_Inge Groeninckx, Steven Dessein, Helga Ochoterena, Claes Persson, Timothy J. 


aceae 


Motley, Jesper Kárehed, Birgitta Bremer, Suzy Huysmans & Erik Smets 
Foliar and Petiole Anatomy of Tribe Hamelieae and Other Rubiaceae 
AA Dorismilda Martínez-Cabrera, Teresa Terrazas € Helga Ochoterena 
Paraphyly of /xora and New Tribal Delimitation of Ixoreae (Rubiaceae): Inference from 
Combined Chloroplast (rps16, rbcL, trn-F) Sequence Data Arnaud Mouly, 

— Sylvain G. Razafimandimbison, Jacques Florence, Joël Jérémie € Birgitta Bremer 
Evolutionary Trends, Major Lineages, and New Generic Limits in the Dioecious Croup of the 


The Rondeletia Complex (Rubiaceae): An Attempt to Use ITS, rps76, and trnL-F Sequence 

Data to Delimit Guettardeae, Rondeletieae, and Sections Within Rondeletia — < 
ohan H. E. Rova, Piero G. Delprete & Birgitta Bremer 
Phylogeny of Tricalysia (Rubiaceae) and Its Relationships with Allied Genera Based on 

Plastid DNA Data: Resurrection of the Genus Empogona 
James Tosh, Aaron P. Davis, Steven Dessein, Petra De Block, 
Suzy Huysmans, Mike F. Fay, Erik Smets & Elmar Robbrecht 


Acknowledgment of Reviewers 


214 


Cover illustration. Galianthe longifolia (Standl.) E. L. Cabral, drawn by Laura Simón. 


Annals 
of the 
Missouri 
Dotanical 


Garden 
207 VG 


Volume 96 


umber 


Annals of the 
Missouri Botanical Garden 


Volume 96, Number 2 
July 2009 


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, in- 
dependent reviewers. 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 
Allison M. Brock 
Associate Editor, 

Missouri Botanical Garden 
Tammy Charron 

Editorial Assistant, 
Missouri Botanical Garden 
Cirri Moran 

Press Coordinator, 
Missouri Botanical Garden 


Roy E. Gereau 

Latin Editor, 

Missouri Botanical Garden 
Ihsan A. Al-Shehbaz 
Missouri Botanical Garden 
Gerrit Davidse 

Missouri Botanical Garden 
Peter Goldblatt 

Missouri Botanical Garden 
Gordon McPherson 
Missouri Botanical Garden 
Charlotte Taylor 

Missouri Botanical Garden 
Henk van der Werff 


Missouri Botanical Garden 


For subscription information contact ANNALS 
or THE Missouri BoranicaL GARDEN, Y Allen Mar- 
keting & Management, P.O. Box 1897, Lawrence, 
KS 66044-8897. Subscription price for 2009 is 
$175 per volume U.S., $185 Canada & Mexico, 
$210 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 


Tue 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 Mtssourt BOTANICAL (GARDEN, %o 
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- 
stract/Global Health databases, ingenta, ISI® databases, JSTOR, Research Alert®, and Sci Search®. 
The full-text of ANNALS or THE Missouri BOTANICAL GARDEN is available online though BioOne™ (http:// 
www.bioone.org). 
© Missouri Botanical Garden Press 2009 
The mission of the Missouri Botanical Garden is to discover and share knowledge about plants and 
their environment, in order Lo preserve and enrich life. 

© This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 


Volume 96 
Number 2 
2009 


Missouri 
Botanical 


Garden 


BIOGEOGRAPHY AND 
PHYLOGENY OF CARDAMINE 
(BRASSICACEAE)? 


Tor Carlsen,’ Walter Bleeker,’ Herbert Hurka,? 
Reidar Elven, and Christian Brochmann? 


ABSTRACT 


The biogeography and phylogeny of e L. were inferred based on sequences of the nuclear ribosomal ITS regions 
and 


and the plastid irnL intro 


traditional sectional classification vel not supported. We found evid 


Colonization of the Southern H d the 
distinct Arctic lineages, two bi Oceanian lineages, 
occurred independently many times during the 
m offer an Sauli for the co 
ords: Arctic, 


biogeography, RCM A Cardamine, 


of the largest and polyploid-rich genera of the 


mine. Recent divergence combined with widespread 


us. 
phylogeny, polyploidization. 


Cardamine L. is a taxonomically complex, cosmo- 
politan genus with at least 160 to 200 Arctic, alpine, 
and boreal species, and is one of the most species-rich 
genera of the Brassicaceae (Sjéstedt, 1975; Hewson, 
1982; Al-Shehbaz, 1988; Webb et al., 1988; Al- 
Shehbaz et al., 2006). The number of species accepted 
varies considerably among different authors, illustrat- 
ing the notorious taxonomic complexity of this genus. 
The center of diversity is clearly situated in Eurasia; 


according to conservative estimates (mainly based on 
Al-Shehbaz, 1988), approximately 95 species are 
Eurasian (ca. 48 of which are in China and ca. 25 in 
Europe including the Caucasus). There are 43 species 
in North and Central America (Al-Shehbaz, pers. 
comm.), and, of these, at least nine species extend into 
Arctic areas. There are fewer native species in the 
Southern Hemisphere: 20 in South America (Al- 
Shehbaz, pers. comm.), 10 in Australia and New 


! We thank all institutions and colleagues (cf. Appendix 1) for providing plant material for our st 
Tundra Northwest (TNW) 199! 


obtained via R. Elven's parti 


History M 


rway. Current address: Microbial Evolution Research Group, Biological Institute, University of Osl 
or for correspondence: tor.carlsen@bio.uio.no. 


uo NO-0316 Oslo, Norway. Autho 


udy. Some material was 


. Box 1172 Blindern, NO-0318 Oslo, 
o, P.O. Box 1172 


Department of Systematic Botany, University of Osnabrück, Barbarastr. 11, 49076 Osnabrück, Germany. 


doi: 10.3417/2007047 


ANN. Missouni Bor. Garp. 96: 215-236. PUBLISHED ON 7 JULY 2009. 


Annals of the 
Missouri Botanical Garden 


Zealand, four in New Guinea, and three in Africa. 
Some species are invasive cosmopolitan weeds, such 
as C. hirsuta L., C. impatiens L., C. flexuosa With., 
and C. panier L. 
In O. E. Schulz’s (1903) monograph of the genus, 116 
pecies were accepted and classified into 12 sections. 
Sh (1936) later extended his account to in 
species in 13 sections (Table 1). Of edes interest 
ina dbi a context are the three largest sections 
of Cardamine: section a L. (Schulzs Eu- 
cardamine Godr.), ur Dentaria L., and section 
Cardaminella Prantl. c s treatments (1903, 
1936), section duce includes ca. 74 species, has 
a global distribution, and encompasses a wide range of 
morphological variation. His section Dentaria contains 
16 perennial species from North America and Europe 
characterized by fleshy creeping rhizomes and typically 
large, showy flowers. Schulz’s section Cardaminella 
as a highly disjunct distribution, including four 
species from the Arctic (including Beringia), three 
species from alpine areas in Europe, one species from 
and thre 
small, E adapted plants. Also noteworthy are six 


Japan, species from Oceania, all of them 
E sections (Table 1 

new species have been described since 
E. s (1903, 1936) Wesen but species delimi- 
tation is difficult and the 
Cardamine | remains da 
partitioning of Schulz has been criticized by several 
authors for o a few morphological 
characters (Al-She 1988; Rashid 
1993). Previous e in Cardamine have shown that 
some d his sections (Cardamine, Dentaria, pu Ue 
chulz, pri vd ulz, and P. 
Med in Schulz) are not imenöphişleie (Rashid & 
Ohba, 1993; ae et al., 1998; Sweeney & Price, 
2000; Bleeker a al. 5 2002); Several species w in 

ly bas 


o 
molecular as well as eytalogical and morphological ds 
The C. pratensis L. complex, for example, has a history of 
recurrent polyploidization events and dispersals over 
relatively long distances (Marhold & Ancev, 1999; 
Franzke & Hurka, 2000; Marhold et al., 2002, 2004; 
Lihova et al., 2004; Marhold & Lihova, 2006). However, 
no genus-wide molecular analysis of Cardamine has 
been performed so far. 

Cardamine is probably a fairly young genus. 
Molecular data indicate that a clade comprising the 
"na Barbarea W. T. Aiton, Armoracia G. Gaertn., 
B. Mey. & Scherb., and Rorippa Scop. is sister to a 
Cardamine—Nasturtium W. T. Aiton clade (Franzke et 
al, 1998; Yang et al, 1999; Koch et al., 2001). 
Rorippa pollen is first found in sediments from the 
Pliocene (2.5-5 Ma) (Mai, 1995). Koch et al. (2000) 
used this time span to estimate that the lineages that 


A rise to Cardamine and Barbarea diverged 

O Ma. This was suggested to be an underestimate 
> Heads (2005), as he doubted the dating of the 
pos However, using the nuclear data set of Koch et 

l. (2000), Haubold and Wiehe (2001) performed a 
more thorough study under various evolutionary rate 
assumptions, all resulting in a divergence time of 


Most species of Cardamine are polyploid, and up to 
five basic chromosome numbers have been suggested 


for the majorit 
2005). For some species, such as the Beringian taxa in 
the most probable basic 
006). Diploids are 
only known with an - = and the highest recorded 


section Mau 
number is x — Elven et al., 
number is 2n. (C. concatenata (Michx.) 
O; Schwartz and " di ds w Alph. bus 
ucera et al., 2005; Warwick & A 
The seeds of Cardamine are ejected by curling of the 
silique walls, a typical short-distance mode of dispersal 
(Kimata, 1983). C 
continents except Antarctica. Under moist conditions, 
the seeds can become mucilaginous and adhere to 
animals (Al-Shehbaz, 1988). 


Cardamine species occur in moist habitats, this may 


ardamine is nevertheless found on all 


As the majority of 


be a common mode of dispersal, also across vast areas 
via birds. Dispersal between Eurasia and North 
epwise via the Tertiary 
Beringian land bridge that existed until 5.4—5.5 Ma 
(Marincovich & Gladenkov, 1999, 2001; Gladenkov et 


al, 2002), but dispersal over longer distances must 


America may have occurred st 


ie occurred between these Cardamine-rich conti- 
nents and Oceania, South America, and Africa 

In this paper, we particularly address the occur- 
rence of such long-distance colonization events, 
including the colonization of the biogeographically 
young Arctic region. Among the ca. nine species of 
oe occurring in the Arctic, two (C. bellidifolia 

and C. pratensis sl) have complete circumpolar 
ieu ions, and seven are restricted to the Beringian 
region (C. blaisdellii Eastw., C. digitata Richardson, 
C. purpurea. Cham. & Schltdl. C. pedata Regel & 


Tiling, C. microphylla Adams, C. victoris N. Busch, 


ER 


; ei Ove 
2000). Murray (1995) inr Pv ihe 
Arctic flm of today is composed of a mixture of 


Lear et al 


survivors from the Arctic Tertiary forest, Pleistocene 
immigrants from various mountain areas, and in situ— 
evolved Pleistocene taxa. 


Volume 96, Number 2 
2009 


Carlsen et al. 
SN and Phylogeny of Cardamine 


Here we attempt to reconstruct the phylogeny of 
Cardamine based on extensive, genus-wide species 
sampling and sequencing of several DNA regions 
(using the nuclear ribosomal ITS regions and the 
plastid trnL intron and trnL-F spacer regions in the 
final analysis). In particular, we address the infra- 

generic classification of Cardamine, especially in 
light of Schulz (1903, Ws and what is already 
known from Franzke et al. (1998), Sweeney and Price 
(2000), and Bleeker et al. Du We also examine 
biogeographic patterns in this widespread genus, and 
particularly address dispersals and source areas for 
colonization of the Arctic and Southern Hemisphere. 


MATERIALS AND METHODS 


Fresh leaf material was sampled and dried in silica 
gel in the field. Vouchers are deposited in the 
herbaria at the Natural History Museum, University 
of Oslo (0), and the University of Osnabrück (OSBU) 
Leaf material was also sampled from herbarium 
specimens in ALA, CAN, CANB, DAO, HBC, LE, 
O, OSBU, OSC, S, UPS, and WU (Appendix 1). 
Species of Cardamine representing all continents 
where Cardamine occurs an of the 13 sections 
described by Sehulz (1903, 1936) were included: 
exceptions are the monotypic sections Giraldiella O. 


E. Schulz, Lygophyllum O. E. Schulz, and Spirobolus 
1) 


O. E. Schulz (Table 
DNA was extracted using e uan Plant Mini 
Kit or DNeasy Plant 96 Kit (Qiagen, Hilden, 


Germany) following the od os protocol. We 
initially tested several DNA regions for a subset of 
species. The mitochondrial nad6 gene, the nuclear 5S 


Table 
of species in some of the sections has increased after 


non-transcribed spacer region, and the plastid regions 
irnT-trnL spacer, psbA-irnH spacer, trnS-trnG spacer, 
and ndhF gene were tested but found either not 
variable enough (nad6), too variable and difficult to 
align (the 5S non-transcribed spacer and psbA-trnH), 
or difficult to sequence (ndhF, trnS-trnG, and trnT- 
trnL). The only useful regions were found to be ITS 
and the plastid trnL intron and trnL-F spacer regions. 
amplification of ITS was performed with 
primers ITS-4 and ITS-5 (White et al., 1990) using 
30 cycles of 45 sec. at 94°C (first ele 5 min.), 
45 sec. at 55°C, and 90 sec. at 72°C (last cycle, 
10 min). The trnL intron was amplified with the 
primers c and d, and the trnL-trnF intergenic spacer 
region with the primers e and f (Taberlet et al., 1991 
using 30 cycles of 30 sec. at 94^C (first cycle, 5 min.), 
30 sec. at 55°C, and 90 sec. at 72°C (last cycle, 
10 min.). PCR products were purified with ExoSAP- 
IT (USB Corporation, Cleveland, Ohio, U.S.A.) before 
sequencing with BigDye d e Biosystems, 
Foster City, California, U.S.A.) using 25 cycles of 
10 sec. at 96°C, 5 sec. at 50°C, es 210 sec. at 60°C. 
Sequences were edited in Sequencher 4.1.4 (Gene 
Codes, Ann Arbor, Michigan, U.S.A.), and ambiguous 
positions were coded according to the International 
nion of Pure and Applied Chemistry (IUPAC 
standards. Sequences were translated to R 
analyzed in RNAfold (Hofacker et al., 
MARNA (Siebert & Backofen, 


ectly. The 
sequences quently aligned manually in 
BioEdit (Hall, 1999). The three regions corresponding 
to the hairpin loop in helix III in ITS-2 and to loops in 


regions were aligned corr 
e 


Sectional classification, geographic distribution, and number of species according to Schulz (1936). The number 
1936. 


Abbreviation 
Section (cf. Fig. 1) No. of spp. Geographic distribution 
L Dem Dent 16 Eurasia and Atlantic North America 
II. A um O. E. Schulz Eutr 2 Pacific N orth America 
III. Sphaerotorrhiza O. E. Schulz Sphae 1 Siber 
IV. Coriophyllum O. E. Schulz Corio 1 Middle Europe 
V. Giraldiella O. E. Schulz* Girar 1 China 
VI. Macrophyllum O. E. Schulz Mac-ph 7 Asia and North America 
VIL. Lygophyllum O. E. Schulz ygo 1 Himalaya 
VII. Papyrophyllum O. E. Schulz Papyro 8 tropical mountains 
IX. Eucardamine Godr. 
— Cardamine Card ca. 74 cosmopolitan 
X. Cardaminella Prantl C-nella 12 cold areas all over the world 
XL Pteroneurum (DC.) Nyma Ptero 5 st Mediterranean region 
XII. Spirobolus O. E. Schulz* Spiro 1 Maia s 
XIII. Macrocarpus O. E. Schulz Mac-ca 1 South Americ 


* Not included in our study. 


Annals of the 
Missouri Botanical Garden 


helices III and IV in ITS-1 could not be unambigu- 
ously aligned, and these regions were therefore 
excluded from the final matrix. For the trnL-F spacer 
region, non-homologous pseudogene replications 
were excluded from the matrix prior to all analyses 
(Koch et al., 2005). In addition, sey and nuclear 
sequences of Cardamine an ated genera from 
GenBank were imported into rA matrices and aligned 
manually, resulting in a data set including a total of 


lll species . Carda m Several 


genera were pa was 


amine a ix 
, but R 
chosen in EN end as the most aren oa 


ested as outgro 


because it was the closest related genus among the 
available genera. 

Parsimony analyses were perform in 
Analysis Using New Technology (ND es et 
al, 2003) with potential parsimony informative gaps 
coded as present/absent (Simmons a 

2000). Heuristic searches were performed with 1 
random addition sequences and tree bisection-recon- 
nection (TBR) branch swapping, saving 10 trees per 
replication. The resulting trees were swapped on with 

TBR saving up to 100,000 trees. Collapsing rule was 
set to minimum length = 0. Random seed was set to 
” Goodness of fit was calculated using consis- 

in ex (RD) and rescaled 
consistency index e & Farris, 1969; Farris 
1989). Bremer support ee 1994) was calculated 
by producing 120,000 trees that were up to 12 steps 


tency index (CD), a ind 


longer, starting with saving 10,000 trees one step 
longer, and successively saving 10,000 trees of up to 
one step longer in 11 steps. Jackknife (Farris et al., 
1996) and bootstrap (Felsenstein, 1985) resampling 
were performed with 1000 o (10 random entry 

d 10 trees ch repetition) and 
collapsing rule = TBR 


orders an saved i 

; Tackknifing was performed 
with 36% deletion. Bootstrap and jackknife were 
f 50% and absolute 
frequencies as output. Implied weighting (Goloboff, 
1993) was performed with K — 1, 3, 6, 8, 20, and 50. 
In addition to the analysis of all taxa, separate 


performed with a cut-off value o 


analyses were performed on diploid taxa, tetraploid 
taxa, and diploid and tetraploid taxa together. 

A Bayesian analysis was performed on the ITS data 
set in eai (Huelsenbeck & Ronquist, 2001; 

elsenbeck, 2003) wit d 

nen time Mons + gamma provided by 
MrAIC (Nylander, 2004). The analysis was run with 
the default d in MrBayes, random starting trees 
run for 3,000,000 generations with sampling of 
Markov alins for each 100th generation. The first 
25% of the trees were discarded as 


Ronquis the model 


and 


“burn-in” 
samples. A Bayesian analysis was also performed on 
the plastid data set, but did not provide any additional 
information and was thus excluded. 


RESULTS 


included 629 


matrix 
ch 188 were parsimony informative 


al aligned ITS 
whi 


he fin 
characters, of 


— 


186 when excluding the outgroup). There were four 
potential parsimony informative coded gaps in the 
ata matrix (2 gaps of length 1 bp and 2 gaps of length 
bp). 


Ng 


most parsimonious trees (MPTs) inferred 
from the ITS data set were 749 steps long with CI — 
0.530, RI — 0.692, and RC — 0.367; one of them, as 
strict consensus tree, is presented in 
and diploids 
together did not give better resolution or conflicting 
Fou (result not shown). Using implied weighing 
did not give better resolution or conflicting topologies 
(result not NH The Bayesian analysis of the ITS 
data set is presented in Figure 2. 

We partitioned the ITS trees into nine operational 
groups to simplify presentation (marked A-J in 
Figs. 1, 2) 

Group A was supported by a posterior probability 
(PP) of 0.96 and Bremer e (BR) = 1. This group 
nae only 16) European species 
NORD poc extending into the 


= 


eer area), four of them belonging to section 
Cardaminella (C. alpina Willd., C. bellidifolia, C. 
plumieri Vill., and C. resedifolia L.) and one (C. carnosa 
Wald Kit.) to section Pteroneurum (DC.) Nyman. 
Group B was supported by BR = 1 and comprised 
the East Asian Cardamine tenuifolia (Ledeb.) Turez. 
of the NUM section Sphaerotorrhiza O. E. Schulz 
and the African C. trichocarpa Hochst. ex A. Rich. of 
section cata 
Group C was supported by BR = 1 and included 
Eurasian and North American taxa. North American 
high polyploids om = 12% = 96 to 2n = 32x = 256) 
of section Dentaria (Cardamine angustata O. E. 
Schulz, C. concatenata, C. dissecta (Leavenw.) Al- 
Shehbaz, and C. diphylla) formed a clade P em 
2 E r a" = a bootstrap (BS) = 88%, BR 
nd P ean species ii section 
Demari " pene O. E. Schulz, C. bulbifera 
Crantz, C. vi nc O. Schwarz, C. ou 
m and nquefolia (M. Bieb.) Ben 
k. f. ex Schmalh.) together with the Asian dinde 
5 pre O. E. Schulz of section Macrophyllum 
formed a clade eS by JK = 91%, BS = 87%, 
BR = 4, and PP = 
Group D 
the East Asian polyploids Cardamine macrophylla 
Willd. and C. O. E. Schulz and the 
European diploid C. trifolia L., bb s to three 
different (Macrophy llum, Dentaria, and 
Coriophyllum O. E. Schulz, respectively}. 


was Bum by BR = 1 and included 
tangutorum 


sections 


Volume 96, Number 2 


Carlsen et al. 
Blegecgraphy and Phylogeny of Cardamine 


Group E was supported by BR = 2 and PP = 1.0. 
This group included the European Cardamine wald- 
steinii Dyer of section Dentaria in addition to two 
well-supported clades with East Asian and North 
American taxa, respectively. The East Asian clade (JK 
= 98%, BS = 97%, BR = = 1.0) included 


members of sections Cardaminella (C. nipponica 


5, and PP 


Franch. & Sav.) and Cardamine (C. microzyga O. E. 
Schulz). The North American clade (JK = 100%, BS 
= 99%, BR = 12, and PP = 1.0) also included the 
cosmopolitan weed C. hirsuta of section Cardam 

Group F was supported by BR = 1 and PP = o 
This group included species of section Cardamine 
from South America, East Asia, and Afri 

Gro was supported by BR — 1 ad Pp = = 0.98. 
This group included Eurasian, African, and South 
American taxa. Most Asian species grouped together, 
containing members of both sections Cardamine and 
Macrophyllum. Notably, the accessions of C. scutata 
Thunb. from Japan and Taiwan did not group together. 
The South American C. ovata Benth. muc with 
African and South American C. a L., bot 
belonging to section Papyrophyllum. “The ie iid 
numbers in this group spanned from 2n = 16 (diploid) 
to 2n = 56. 

iy H was T "i JK = 65%, BS = 50%, 

= 4, and PP . This group included a 
specimen from New E referred to as Cardamine 
africana; it did not group with the other C. africana 
accessions (group C) and most likely i dc a 
different taxon. ‘Al fo rom 
Guinea belonged to this group, which also included 


our species known 


rd species of section Dentaria (C. uec ie 
chulz, C. kitaibelii Bech., C 
2 e (L) Crantz). 

Group I was supported by BR = 1 and PP = 0.94. 
This group included the Cardamine pratensis species 
group and its closely related European species, which 
also formed a clade in the plastid analysis (Fig. 3; JK 
= 81%, BS = 81%, BR = 3). As this group has been 
extensively i earlier (Franzke et al., 
Franzke & H , 2000; Lihova & Marh old, 
and our e supported their 


998 
2003) 
indings ito 
adding new information, we pruned tid. taxa of this 
complex from our final analyses and retained only four 
species related to C. pratensis to simplify this 
presentation (C. acris Griseb., C. flaccida Cham. & 
Schltdl., C. pratensis, and C. tenera S. G. Gmel. ex C. 
A. Mey.). 

Group J had no support and was only present in the 
MPTs and a combinable components consensus tree, 
but included several supported subgroups. This group 
included most of the Beringian taxa, all of t 
Australian and New Zealand taxa, and many North 
American taxa in addition to one species from South 


America and East Asia, respectively. The Australian 
and New Zealand taxa, Cardamine debilis Banks ex 
DC., C. lacustris (Garn.-Jones 
Heenan, C. lilacina Hook., and C. paucijuga Turez., 
constituted a monophyletic group (JK = 55%, BS = 

57%, BR = 1, and PP = 0.98) with the inclusion of 
the South American C. glacialis (G. Forst.) DC. and 
the amphi-Beringian/Pacific C. ad Greene. 


The ITS e 


ohnson) 


set was inconclusive about the 
ophyly the remaining Beringian species (C. 
blaisdelli " digitata, C. purpurea, C. pedata, C. 
microphylla, C. viens and C. sphenophylla). How- 
ever, most of the MPTs supported the Beringian 
species as a monophyletic group with North American 
taxa as sister groups. The remaining trees supported 
the Beringian species as two separate groups, but both 
of them with North American species as sister groups. 
This led to the collapse of these branches in the strict 
consensus tree and the resampling analyses. 

The South American species appeared scattered in 
the tree. Cardamine glacialis was most closely related 
to the taxa from Oceania in group J, C. bonariensis 
Pers. and C. flaccida close to or nested within the C. 
pratensis group in group L and C. ecuadorensis Hieron. 
and C. rhizomata Rollins were resolved as a sister 
group to C. griffithii Hook. f. & 

The three African species did not form a mono- 


homson in group F 


phyletie group. Cardamine trichocarpa was found in 
group B, while C. obliqua Hochst. ex. A. Rich. was 
sister to C. lihengiana Al-Shehbaz in group F with BR 
= 2 and PP = 0.96. A specimen from South America 
referred to C. africana was more closely related to 
outh American C. ov J 
specimen from Kilimanjaro. 


ata than to a C. africana 

The Oceanian taxa occurred in two distinct clades. 
In group H JK = 65%, BS = 50%, BR = 4, and PP 
= 1.0), three European species were nested among 
four species from New Guinea (Cardamine sp. aff. 
africana L., C. altigena Schltr. ex O. E. Schulz, C. 
a O. E. Schulz, and C. papuana O. E. Schulz). 


> 


group J, four species from Australia and New 
Zealand | (C. paucijuga, C. lilacina, C. lacustris, and C. 
debilis) were most closely related (JK = 54%, BS = 
57%, and PP = 0.98) to the Arctic C. umbellata and 
the South American C. glacialis. 
The B 
different groups. Cardamine pratensis (group I) and C. 
eae ae A) En their closest relatives in 


eringian/circumpolar taxa occurred in three 


urope. p J, there were se 
species (C. blaisdells, » digitata, 
pedata, C. microphylla, C. victoris, and C. spheno- 


en Beringian 


M" purpurea, C. 


phylla) possibly having their closest relatives in 
orth America. 

The aligned trnL-F matrix included 765 characters, 
77 of which were parsimony informative (71 when 


Oceania and 


Annals of the 


220 


Missouri Botanical Garden 


eJep peusmqnd Ápsnoraexd uo peseq oureu sotoods ey) rege USATS ore srequmu ouiosoulomn[) '[euruire] oY) 107 e[orro uedo o[gurs e yym poejeorpur ere uoxe] owes ot] Jo suorssoooe o[dnpnur 
“uonejuosoid Áp[durs og, "310ddns demsjooquoddns opuxoef oyeorpur soyouerq Mo[oq sroqumw “1roddns 19turorg/sqiBuo] Youe1q ojeorpur SOYouBIq oAoqe sroqumwN 770£'0 = DH PUR “2690 
= TH *0£€'0 = [N exe songea 115 jo ssoupoos oy) pue sdojs gp, st YSU] oen au] oon snsuosuoo POIS oY) ur Woy so ouexq SULMOJS Jos EWP SLI eu uo peseq SLAN oup jo 9UQ ^[ omy 


[prea] $9 = uz SITequeproco "ng 
[pIe] 9T = UZ P3ns1Tg "D 


[3ued] 29 = ug unzojnbuej ʻo 

yoddns densjoog / yoddns ajruxxoer al [ud-oeg] 96- E: E de eto creme Zn 
yoddns wag / uiuo] your [or205] 9T = uz B Phos pa 
[qued] Zp = ug Px0bT[Npuero 
[3ueq] 96 = ug ezegrqing * 


saqougJq MOA PUE aAoqu SISQUINN 


aon snsuosuoo jars 907 01 yey q9uelq — [qud-2ew] 9T = uz eyqueonst “Je 


[P189] zE'9T = uc sueriedur 
291 Smnsuesuo2 pus yy ur day jou qaug1q ..n. STSUDUPSTIP 
[38801 08 = uz SoTTAYdeauue 
3 [3usg] e7oəssSTp '*2 
adomg B [1ued] 959Z = UZ 232493290402 "Do 
[1ueg] 8ZI = uz pens A - 
vore ueyodouuso, [3uedq] 952-96 = uz erIÁ4udrp € Ui Ec/se 
uum og eed XE o SH 
vuesso [E] [paeo] srsueueuunÁ - ot 
I90uP3suoo '2 — pos 
sv Eg [prea] "A prenn E 


gp = Ug Ejreguoo '2 T/E 
[P120] ZEʻ9T = ug edIe2cyoFI} ^20 
[seuds] ze = ug errogrnue3 -p SLA 


vousury HON [i 
vere 1ejodumozio/erSuueg | | 


Ba 


[eTTeu-5] 91 = uz PFIOSFPFITOG ‘Qo SE m 
[errteu-9] 9T = ug eurd[e 9 
soy mos Mg [erteu-a] 91 = UZ 14 erTojrpasez 
vmv [E [eTTƏu-9] 91 = uz z# eTIOJIDese: 


[eTI9Uu-2] 9I = uz rzerungjd 
[o1934] esouze2 
STI2SO8ATÁS p aad 

stzysnted eddrzow EI er ‘Bra 


Volume 96, Number 2 Carlse al. 
2009 fami and Phylogeny of Cardamine 


obliqua 2n = 36,72 [Card] 


[Card] F [E] Africa 
dorensis [Card] N 
3 C. rhizomata | | South America 
OC. amara 2n = 16132) [Card] Lc" "T 
FP c. torrentis 2n = 56 [E] Beringia/circumpolar area 
p 2 dd c fae 2n - m [Card] " 
2n = 16 [Papyro Bl North America 
i a a 42 2n = 16 [Pa pyro] ag 
vata [Papyro] | | sia 
Oceania 
- 32 [Card] G O 
= 32,48 M Cosmopolitan area 
= | 18, 20 [Card] 


| | Europe 


+ --- branch not kept in the strict consensus tree 


2n = 48 [Dent] H 
[Papyro] 
lii [Dent] Numbers ab branches 
^ "n 19 pas LÀ Branch length / Bremer supp 
16 [Card] Jackknife support / Bootstrap support 
2n = 16-118 [Card] 
16 [Card] 
matthioli 2n = 16 [Card] I 
uliginosa 2n = 16 [Card] 
acris 2n 16, r32 (Cara) 


laris 2n = 16,24 
i 16 


: C. castellana 2n - 16 — 
pe ues clematitis [Card] — 
r^ c. lo ongii 

i C. rotundifolia [Card] 
2/67 C. bulbosa 2n - 32-64 
aep... ieu mm m = OI 

OC. parviflor = 16 [Card] 
e — C. nuttallii [Eu E 
H- c. rupicola 
us C. breweri m - 84-96 [Card] 

= 24 [Card] 
gallia 2n = 32 [Eutr] 
= 40 [Mac-ph] 


El 
ii 2n = 28,42 [C-nella] J 


[C-nella] 
C-nella] 


Hs YY 
of 


Figure 1. Continued. 


excluding the outgroup). There were 11 potential of the plastid data set (Fig. 3) resulted in a poorly 
raw informative coded gaps in the data matrix resolved strict consensus tree, but several io ee 
engths Des from 1 bp to 11 bp. The MPTs gro ecovered in the analys 
ite from the trnL-F analysis were 258 steps long diploid Cardaminella taxa (Cardamine bellidifolia. c 
with CI — 0.694, a = 0.807, and RC = 0.560; one of alpina, and C. resedifolia) constituted a monophyletic 
them, as well as the strict consensus tree, is presented group with 91% JK support, 88% BS support, and a 
in Figure 3. In terms of initial similarity retained as BR support of 2 with C. ow Spreng. as sister (JK — 
synapomorphy (RI), the plastid characters were more 67%, BS = 55%, and BR = 1). There was also 


self-congruent than the ITS characters. The analysis support (BR = 1) = monophyly of the Oceanian 


Annals of the 


222 


Missouri Botanical Garden 


1x9) eu) ur Ho peiuoumoo sdnois oyeustsop [—y s1omeT '([ AQEL Jo) syoyovrq ur oureu uonoos oy} Aq poao[[oj ore pue (GQOZ) Te 19 exoony ur pean’ e1ep peusrqnd 


A[snorAo1d uo poseq Y} Joye UoAIS oTe { 


az ‘Sta 


sg0 


[qued] gp = uz sorrAydequed °p 
[3ued] FrTeqrezry '2 


[3ueq] gp = uz err4udeadeu ʻo 

H IzessÁowx * 
PUobI3[? 

Pueorige “¡je ‘ds * 

— [etreu-2] eoruoddru ‘pj 
[paeo] ebAzOzO pu "0 
PUPIjeu2ue1j 


o 


A [pze5] 


P9 = uz sTTequeptooo "9 
[pred] 91 = ug eansatTy 


= [3ueq] IIUIS3SDp[eM s 
F— [quedq] zp = uz unzo3nbue2 “D4 
[ud-oeW] 96-79 = uz eT 14udo22Pur y 00) 


a [Jua] 96 = UZ P18119TDq '2 
[aued] zv = uz EIOBEIHPURTD do] 


[ud-oeW] 91 = uz euaueoner 

[pred] z£/91 = uz suer3edu; *9 

STSUSUPSTIP '2 
[qued] 9Gz = uz e3euejeouoo | 

9 [queq] 832essrp WT 
[3ued] 8ZT = uz e3e3snbue oed 

[3ued] 9827-96 = uz err4udrzp *9 

[paeo] errogrrizebezg * 


D) 'seniqeqoxd 10ut9js0d 9jeorpur SOTOURIG 9A0qe SISQUINN “Jes HEP SLI PY uo poseq Ay I gq 


‘| om3 


wo 
reTeos 


ez 'br4 


Qh = UZ 2}7Ə7JU02 “y 
Ieouejsuoo 29 — HA 

[3ued] 08 = uz soprÁAudeeuue * 
rIepezq 


[p1e9] srsueueuunA 


ZE*9T = uz PdIeooyoTI3 ^2 
[eeuds] ze = uz errozrnuez * 

[erTeu-2] 91 = UZ PITOFTPITTOG '2 
[erteu-9] 9T = uz eurdpe * 
[erteu-9] 9T = uz TH PTTOJTPOSOI '*2 


[pze5] 


C 


[erreu-5] 9T = ue z4 PITOJTpesezr *) 
[ĮeTTƏU-9] 9T = uz Izeiun[d - 
[01334] esouzeo 77) 


Sr43S8A[ÁS eddrioy 
sti3snTed eddrioy 


Volume 96, Number 2 Carlse al. 
2009 debet and Phylogeny of Cardamine 


Es A [Card] 
C. acris 2n = 16,32 [Card] 
Cs qiwi vue is 2n = 16,24,32 
flaccida 2n = 16 [Card] 
C. uliginosa s = 16 [Card] 
C. matthioli 2n 
C 


16-118 [Card] 
16 
astellana 2n - 16 


0.96 lihengiana 
100 l . obliqua 2n = 36,72 [Card] 
. griffithii [Card] 


1.00 rc. a em [Card] 
rhi 


PME 
Y 
oY i 

oe 4 
ja 

is 

SE OG 

Eg 

m = 

ERE 

NYAR 

55.46 

1 7 


LL c. 
—C. amara 2n = mis [Card] 
— C. amporitana 2n = [Card] 
L-C. torrentis pe - oe 

(Es ende api 2n = 18, e ei 
utata 2n = 


[Mac-ph] 
2: Hen G 


0.55 


Es n ] 
. africana #1 [^ = 16 [Papyro] 
C. africana eo 2n = 16 [Papyro] 
ovata [Papyro] 
Y densiflora 
2,48 


flexuosa 2n = 32 [Card] 

Ce blaisdallii 2n = 28,42 d nella] — 
: rupicola 

Fig. 2b LC. purpurea, 2n = 80,96 [C-nella] 

| C. nuttallii [Eutr] 


c Tongia 
3 HC. digitata 2n = 28,42 [C-nella] 
scales L———— C. clematitis [Card] 
oot aliforni 2n = [E 


082 ezoensis 2n = 32 [Mac-ph] J 
0.73 C. sphenophylla 2n = 28 [C-nella] 
C. victoris 2n = 28 [C- nella] 
lia [Card] 


—<e. bulbosa 2n - 32-6 
1.00 A Poza 2n = 16 [Card] 
2 2,72 


C. debilis 2n = 48 [Card] 
C. corymbosa 2n - 48 [C-nella] 
E umbellata 2n = 32,48 [C-nella] 


48 
alis 2n = 72 [Card] 


Figure 2. Continued. 


Cardaminella species (C. corymbosa Hook. f, C. Discussion 
debilis, and C. lilacina) with the inclusion of the 
Beringian C. victoris and C. umbellata, the North 
American C. cordifolia A. Gray, and the South Although 29% of the characters in the ITS data set 


American C. glacialis. The four d species 


RAPID DIVERSIFICATION AND WIDESPREAD POLYPLOIDIZATION 


were MM informative, only a few of them 
previously shown to be related to the C. pratensis were useful for resolving deeper relationships in 
complex (C. pratensis, C. matthioli ~ ex Comolli, Cardamine. Even un the overall RI for the ITS 
C. penzesii Anéev & Marhold, and C. rivularis Schur} trees might suggest that there is substantial homoplasy 
formed a group supported by 81% in both JK and BS, in the data set, with 31% of all characters being 
and 3 in s The North American high polyploids (C. ^ retained as such, implied weighting did not affect the 

concatenata, C. diphylla, an - topology or improve the resolution of the deeper 
d mE formed a clade (JK = 82%, BS = relationships as one might expect when reducing the 
76%, and BR = 1). effect of homoplastic characters. The most likely 


Annals of the 
Missouri Botanical Garden 


Fige 38 


0 
EXE Rorippa palustris 
[ — Rorippa sylvestris 


.2, =i Fer d pua 


c. p a 

s1— C. diphylla North American high polyploids 

M 2/776L0 —C, dissecta 

00/100 L2—C. africana 
$c. glanduligera 

m= tenuifolia 


» - - - branch not kept in the strict consensus tree 
a 
4/1 C. carnosa — — branch kept in the strict consensus tree 
F^ auri c. occidentalis 
100/100 L4 pat j 
T 


Numbers above and below branches 


Branch length / Bremer support 
Jackknife support / Bootstrap support 


folia 


C; d 


C, glau ad 
C, bellidifolia European diploid 
alpina Cardaminella taxa 

e resedifolia — 

3 
1 C. sp. aff. ae 
m —-C. red ři 
(e: ni 


- 2C. n ua 
2. .]— c. microzyga 
2 2/1 C. delavayi 
ES 69/66 
ods 
sie 


C. waldsteinii 


C. bonariensis 


C. ecuadorensis 
Ce eit a q 


ta 
C. matthiolii 
e penzesrr | European taxa related to 
C i s 


the C. pratensis complex 


1--- Fig. 3b 


. One of the MPTs based on the trnL-F data set showing branches kept in the strict consensus tree. Numbers 
ds Tranches are Bremer support values. Numbers below branches are jackknife/bootstrap support value 


explanation for the lack of resolution is, therefore, that (Al-Shehbaz, 1988; Kucera et al., 2005; Beilstein et 
the initial diversification in Cardamine occurred al., 2006; Marhold €: Lihova, 2006), also could have 
rapidly. 
Polyploidization, a common mode of evolution in 


results imply that polyploidization has happened 
Cardamine as well as in the Brassicaceae in general 


independently in many lineages in Cardamine. We 


Volume 96, Number 2 Carlsen et al. 
2009 roda and Phylogeny of Cardamine 
rs 
Es edi fol lå 
is 
C. glacialis 
lilacina Oceanian Cardaminella taxa 
mbellata 
. corymbosa 
debilis 
c. e e 
" PLC. brewe 
im E—C. flagellifera 
PC. flexuo 
E—C. lihengiana » - - - branch not kept in the strict consensus tree 
ic, prorepens 
LB ——c. rup — — branch kept in the strict consensus tree 
EÉ— C. tanakae 
1A C. tenella 
65/74] C. californica 
TAON, C. nuttallii Numbers above and below branches 
sume tite C. paucijuga Branch length / Bremer support __ 
C. scutata Jackknife support / Bootstrap support 
C. microphylla 
2.. C. pedata 
2 — Ç. purpurea 
apc: a nM 
€. digit 
PE (977 seio TN 
2— C. angulata 
C. pensylvanica 
C. bulbosa 
C. clematitis 
C. dentipetala 
C. douglassii 
C. rotundifolia 
C. scutata 
C. torrentis 
Figure 3. Continued. 


found that nine of the 10 groups (noted A-J in Results 
and Figs. 1 and 2) contain diploid species, that on 

one group is exclusively diploid (group A), and s 
most groups contain species that are hexaploid or even 
higher. However, our separate analyses of the ITS 
region for diploids and we did 


not provide 
better resolution or conflic th 


ing topologies, nor di 
plastid tree show better picos. These results 
support our hypothesis that the lack of resolution is 
due to rapid speciation rather than breakdown of 
phylogenetic signal from frequent allopolyploidiza- 
tion. 
Untangling the hierarchical structure among rapid- 
erging lineages is difficult a ms a huge 
effort (Fishbein et al., 2001). We did an 
extensive survey of different regions in the prelimi- 
nary analyses for this study, and a search for more 
o uou molecular markers in 
Cari e prove fru 
Ones is MARKEN pel of Koch et al. (2000) 
and Haubold and Wiehe (2001), who used molecular 
dating to demonstrate that Cardamine is a relatively 
young genus. After the split between Cardamine and 
Barbarea as late as 6.2 Ma (2-8 Ma), Cardamine 


3 
sequencing e 


rapidly diversified into one of the most species-rich 
genera in the Brassicaceae. This is consistent with the 
pattern observed in other large genera of this family, 
such as Lepidium L. (2.14.2 Ma, ca. 175 species; 
Mummenhoff et al., 2001) and Draba L. (4.5-9 Ma, 
ca. 350 species; Koch & Al-Shehbaz, 2002). 

Several papers have addressed the importance and 
previous underestimation of long-distance dispersals 
to explain biogeographic patterns (Donoghue & Smith, 
2004; Givnish & O 2001; Thorne, 2001; Cook & 


causing 
vicariant speciation can be ruled out because the 
present constellation of continents was established 
millions of years before the origin of the genus. Based 
on the estimate of 2-8 Ma, it is likely that the large 
late Tertiary forest in the Northern Hemisphere 
(Lafontaine € Wood, 1988; Bennike € Bócher, 
1990; Matthews & Ovenden, 1990; Murray, 1995; 
Graham, 1999) provided the first habitat for estab- 
lishment, divergence, and spread of Cardamine. It is 
e that the later ku a of the Bering Land 

e (Gladenkov et al., the successive 
oe of the Holarctic ean (Zachos et al., 2001) 


Annals of the 
Missouri Botanical Garden 


has caused vicariant speciation in the genus. Despite 
its low overall resolution, our phylogeny nevertheless 
provides evidence for several extensive long-distance 
dispersal events (further discussed below). 


SECTIONAL PARTITIONING 


We found no support for monophyly of any of 
Schulz’s (1903, 1936) large sections (Figs. 1—3). 
Section Cardaminella was not monophyletic; the 
species of this section occurred in different supported 
of other sec 
(compare clades in aes A, E, and J; Figs. 1, 2). Our 


clades intermingled with species tions 


results also reject the monophyly » section Dentaria, 
in agreement with Franke. et al. (1998) and Sweeney 
and Price (2000). Furthermore, the species included 
in the largest section, Cardamine, are spread among 
different supported s and intermingled with 
speci J). This 


section ane apparently served to include species that 


roup 
s of other sections (e.g., groups G and 


did not fit morphologically into any of the other 12 
sections, as suggested by Sweeney and Price (2000). 
We have also shown that Macrophyllum and Papyro- 

are not monophyletic, in agreement with 


phyllum 
Pid d et al. (2002) and Sweeney and Price (2000; cf. 
1-3). 


RAPID COLONIZATION AND SUBSEQUENT DISPERSALS 


Because of its poor resolution, our phylogeny was 
not suitable for biogeographic analyses such as D 
1997). 
However, because of the DER higher Ard 


to reconstruct ancestral areas (Ronquist, 
of species and, in particular, diploid ones, Eurasia 

certainly the most likely area of origin of Codamine 
Despite its typically short- distance main mode of 


Hemisphere (Asia, Europe, and North America). 
Then, at a later stage, it spread across vast distances 
to the Southern Hemisphere as several distinct 
lineages (cf. also Bleeker et al. , 2002). Both Eurasia 
and North Ameri 


colonization episodes can be inferred based on 
supported groupings in our phylogeny and are shown 
in Figure 4. 


Oceania. One example of very long-distance 
colonization (arrow f in T n followed by rapid 
speciation is provided by th 

ealand taxa, which form a monoph 
together with one Beringian (Cardamine umbellata) 
as well as one South American species (C. glacialis; 
subelade in J, Figs. 1, 2). Notably, these Oceanian 
species are morphologically diverse and comprise 


lowland as well as alpine taxa, but appear genetically 
similar. The four Oceanian species from New Guinea 
(C. sp. aff. africana, C. altigena, C. keysseri, and C. 
papuana), on the other hand, belonged to another 
distinct clade, which also comprised Northern Hemi- 
sphere taxa (group H; Figs 1, 2). Thus, Oceania 
appears to have been colonized at least twice from the 
Northern Hemisphere. 


uth America. We can discern at least four 


different dispersals into South America, two of which 


ardamine africana are monophyletic only by the 
inclusion of the South American C. ovata, which were 
suggested to be conspecific with C. africana earlier 
(Sjéstedt, 1975) (see subclade in group C). This must 
represent an independent Epa event into South 
America, either from Africa or from the Northern 
Hemisphere with later el to Abs (arrow d in 
Fig. 4). Long-distance dispersal from Africa to South 
America is not an impossible scenario (Graham, 2006). 
As noted above, the 
glacialis is most closely related to the species from 


outh American Cardamine 


Australia, Tasmania, and New Zealand in both the 
trnL-F (JK = 52%, BR = 1) and ITS (JK = 55%, BS 
= 57%, BR = 1, PP = 


a a separate dispersal event into South America (arrow g 


0.98) tree, and must represent 


rdamine flaccida G 
suggested by Sjóstedt LE to be 
conspecific, form a clade with the C 


nariensis 
uropean 
pratensis complex and thus provide an dns of 
yet another dispersal event (arrow e in Fig. 4; group I 
in Figs. 1, m the parsimony ITS tree, the 
direction is impossible to determine as both ways are 
equally parsimonious (Cook & Crisp, 2005). We can 
infer either two dispersal events into South America or 
one old dispersal event to South America with a 
subsequent speciation event and then dispersal back 
to Europe. 

Evidence for a fourth dispersal event into South 
America is provided by the sister see ju ah 
P = 0.78, BR = 1) between the 

griffithii and the South American C. a dnd 
C. rhizomata, which form a separate clade within 


a 


group F and may have originated from Eurasia (not 
indicated in Fig. 4). Our results reject Sjéstedt’s 
1975) hypothesis that C. 
rhizomata are conspecific with C. africana. 

Africa. 


monophyletie group in our analyses. The sister group 


ecuadorensis and 


— 


The African species did not form a 


of the African Cardamine obliqua differs between the 
ITS tree (group F) and the plastid tree, as also found 
by Bleeker et al. (2002), but it is possible that it has a 


227 


Carlsen et al. 


Volume 96, Number 2 


2009 


Biogeography and Phylogeny of Cardamine 


“(qe poexreur) 1xo1 91] ur possnosrp s1uo4o Tesiodstp 1ofeur poxropur ojeorpur SMOIIY ^c pue [ S9MÍT ur pue 4x9} ou ur posn se seore Sururpop dey ‘p omsrq 


Annals of the 
Missouri Botanical Garden 


Eurasian origin (arrow e in Fig. 4). The African C. 
trichocarpa (see group B) is highly divergent with its 
a 7 trnL-F autapomorphies and has no 
unambiguous sister group. It is not possible to infer its 
origin except that it is not sister to the other African 
taxa. Finally, C. africana, which occurs both in Africa 
and South America (cf. above), is member of yet 
another clade, nested with European and Asian taxa 
(group G; PP — 0.98 and BR - 1). 

The Arctic. 
colonization episodes into the Beringian/circumpolar 
h the Arctic 
(e.g., C. macrophylla, C. conferta Jurtzev, C. tenuifo- 


We can infer at least three major 
region. Many species of Cardamine reac. 


lia, C. prorepens Fisch. ex DC., C. scutata, and C. 
amara L.), but here we focus on the nine taxa having 
the major part of their distribution in this region (C. 
bellidifolia, C. blaisdellii, C. digitata, C. pratensis, C. 
purpurea, C. pedata, C. microphylla, C. victoris, and C. 
sphenophylla). 

Both the chloroplast and nuclear data demonstrate 
that the broadly circumpolar diploid Cardamine 
o in group A is ces E 77%, BS = 

= 1.0, and BR = ITS and JK = 91%, 
ne = om. and BR = "i in 1 mLF) to European 
diploids, specifically in the Alps and Pyrenees (arrow 
b in Fig. 4). This clade is sister to other European 
Mediterranean-Alpine diploids such as C. glauca in 
the trnL-F tree (JK = 67%, BS = 55%, and BR = 1) 
and C. carnosa and C. plumieri in the ITS tree (PP = 
0.96 and BR = 1). 

Another example of European origin is provided by 
the Arctic cireumpolar Cardamine pratensis (group I). 
The Arctic specimens are regarded as a separate 
subspecies, C. pratensis subsp. angustifolia (Hook.) O. 
E. Schulz, which some authors regard as a separate 
species, C. nymanii Gand. (Franzke & Mummenhoff, 

rdamine p is a common boreal 
selis and belongs to a complicated species 
complex with many described diploids and low 
polyploids distributed throughout Europe. Our results 
are consistent with those of Franzke and Hurka 
(2000), who concluded that C. pratensis colonized the 
Arctic from southern Europe during the Holocene 
(arrow h in Fig. 4). 

The remaining seven Arctic species (group J), 
which are restricted to the amphi-Beringian region, 
may have originated from one or two colonization 
events from North America (arrow a in Fig. 4), as 
these taxa are nested with other North American 
species. However, this is only inferred from the MPTs 
and the combinable components tree. Because we use 
collapsing rules that do not allow zero length 
ranches, we did not get support for this hypothesis 
in the resampling analyses. 


Thus, there are two distinct examples of European 
origin of Arctic Cardamine, including two different spe- 
cies that have become broadly distributed in the Arctic 
without further diversification. In addition, there is one 
example of a probable North American origin followed 
by diversification into many species in Beringia, but 
without further expansion into the circumarctic 


Literature Cited 


Al-Shehbaz, I. A. 1988. The genera of Arabideae (Cruciferae, 
rr in the southeastern United States. J. Arnold 
66. 


Arb. 6 


E . Beilstein & E. A. Kellogg. 2006. Systematics 
d phylogeny of the Brassicaceae (Cruciferae): An 


Beilstein, M. A., I. A. Al- Shéhbaz & E. A. Kellogg. 2006. 

mm phylogeny and trichome evolution. Amer. J. 
t. 93: 6 

Rs O. P" M 1990. Forest-tundra neighbouring 
the North Pole—Plant and insect remains from the Plio- 
Pleistocene Kap København formation, North Greenland. 
Arctic 43: 331—338. 

Bleeker, W. & H. Hurka. 2001. br red RURSUM: in 
Rorippa Manes de da 
natural se nthropogenic C Dobis 
2013-20 


Molec, eol 10: 

, K. Pollmann, A. H. D. Brown & H 

Hurka. 2002. Lyn. and PA of Southern 
n Cardam es (Brassic 


Bremer, K. 1 o Support and tree stability. 
Cladistics 10: 295-30: 
Cook, L M 005. Directional asymmetry of 


could mislead 


Murray, V. Y 
006. Panarctic Eos Checklist. Vascular Plants Draft 
version (2006). National Centre for Bios 
University of Oslo, Oslo. <http:// bi 
B ms accessed 16 April 2009 
Farris, J. S. on and d rescaled 
ia. vid Cladisties 5: 417—419. 
Albert, M. nnd D. Lipscomb & A. G. 
Bige v > e outperforms neighbor- 
g. Cladistics 12: 99-1 
TM J. 1985. eines limits on phylogenies: An 
approach A the Hei Evolution 4: 783—191. 


ystematics, 
r 


e retentio 


ysis of a rapid, ancient radiation. Syst. 
Franzke, A. & K. Mummen hoff. P Recent hybrid 
speciation in Cardamine (Bra 
nuclear ribosomal ITS A in statu Sid Theor. 
m Genet. 98: 831—834. 
& H. Hurka. 2000. Molecular systematics and 
MD a E i pratensis complex (Brassi- 
eae). a Syst. Evol. 224: 213-234. 


, W. Pleks: R. Kohrt & H. 
1998. de a of Cardamine and allied 
genera (Brassicaceae): ITS and non-coding chloroplast 
DNA. Folia Geobot. 33: 225-240. 


Hurka. 


Volume 96, Number 2 
2009 


Carlsen et al. 
Bisdeography and Phylogeny of Cardamine 


Givnish, T. J. & S. S. Renner. 2004. Tropical intercontinental 
LE Gondwana up, immigration from the 
bore: and transoceanic dispersal. Int. J. Pl. Sci. 
165: SLsó. 

Gladenkov, A. Y., "d E. boues : L. inno genis 
Barinov. 2002. A refined a, arliest opening + 
Bering Strait. es jene Palaeoecol. 
183: 321 Yu 

Goloboff, P. 
tree sear EN cid stic 

— J. Fi & K. TE TNT (Tree analysis 
using New Tech Published by the authors, Tucu- 
mán, Argentina 

Graham, A. 1999. dud Cretaceous and Cenozoic History of 

North American Vegetation North of Mexico. Oxford 
University pr New York. 

—. 2006. Modern processes and historical factors in the 
origin of the African element in Latin America. Ann 
Missouri Bot. Gard. 93: 335-339 

Avs E it: A iiri biological 
equence align editor and analysis program for 

Windows 95/98/NT. Nucl. Acids bano e al: 95-98. 

65. The Genus Dentaria (Crucif ae in 
orth America. Ph.D. Dissertation, Vanderbilt 


993. ae character weights during 


* 


lec me r phylogenies: A 
énumus of oleae [GEN Cladistics 21: 62-78. 
Hewson, J. Flora of Australia. Australian 
Coxeuimeut iu s Canberra 
F. St 


adler, S. NEUE M- 


RNA secondary structures. -o Chem. E E n 

Huelsenbeck, J. P. . Ronquist. 2001. MRBAYES: 

ayesian inference of disent trees. e 
17: 754-755. 

Kimata, M. 198 3. Comparative studies on the reproductive 
systems of Card amine flexuosa, de impatiens, 
Cardamine scutata, and Cardamine lyrata, Cruciferae. 
Bot. Mag. (Tokyo) 96: 299-312. 

& J. S. Farris meee ewe phyletics and 

1-32. 


Cardamine 


A. Al- Shehbaz. ed wd data indicate 
complex intra- and intercontinental differentiation. of 
American Draba Cage Ann. Missouri Bot. Gard. 
89: 88-109. 
—, B. Haubold & T. Mitchell-Olds. 2000. Comparative 
EU. analysis of chalcone synthase and alcohol 
dehydrogenase loci in Arabidopsis, Arabis, and related 
genera es Molec. Bial Evol. 17: 1483-1498. 
2001. Molecular systematics of 
ET | Evidence from coding plastidic matK 


and nuclear Chs sequences. Amer. J. Bot. 88: 534—54. 
——, ob ie M TA W. Bleeker, J o M 
Kiefer & T. hell-Olds 


ite 005. E of the irnF 

GAA) gene in sued ds psis E the Brassicaceae 
family: Monophyletic origin and MXN diversifica- 
tion of a plastidic pseudogene. Molec. Biol. Evo 
1032-1043. 

Kucera, J., I. Mis K. Marhold. 2005. pale line database of 
the c numbers of E genus Cardamine 
E “Biologia 60: 473-4. 

Lafontaine, J. D. & D. M. Wood. Ww À zoogeographic 
nA of the Noctuidae (Lepidoptera) of Beringia, and 

ome inferences about past Beringian habitats. Mem 

in, Soc. Canad. 144: 109-123. 


Lear, C. H., H. Elderfield & P. A. Wilson. 2000. Cenozoic 
al ice volumes from Mg/Ca 
287: 269-272. 
Lihova, J. & K. Marhold. 2003. Taxonomy and distribution of 
the Cardamine ea group (Brassicaceae) in Slovenia. 
Phyton ae 241-261. 
———, J. Fuertes Aguilar, K. Marhold & G. Nieto Feliner. 
2004. a of the disjunct tetraploid Cardamine 


mporitana (Brassicaceae) o a nuclear 2 
chloroplast DNA sequence data. r. J. Bot 
Ec 

Mai, D. H. 1995. Tertiáre Vegetationsgeschichte Europas. 
Gustav ee o a. 

Marhold, K. cev. 19 damine penzesii, 


ar 
rediscovered taxon of P E pratensis group (Cruciferae) 
Ann. Bot. Fenn. 36: 1 
— & J. Lihova. F Polyploidy, hybridization and 
las evolution: Lessons from the Brassicaceae. Pl. 
ISBE Evol. V259: 1 m 174 
, M. , R. Grupe & B. Neuffer. 2002. 
^ Natural Ns e in dd amine (Brassicaceae) in the 
Pyrenees: Evidence from pe and molecular 
data. Bot. J. Linn. Soc. T 4. 


i eker. 2004. Comparative 
ETT and AFLP analysis of se Cardamine (Brassica- 
m taxa from deed related polyploid complexes. Ann. 

t. (Oxford) 93: 507—520. 

ee L. Pur New evidence for the age of Bering 
Strait. oe Sci. Rev. 20: 329-335. 

. Gladenkov. 1999. Er x an early 

-151. 


. Ovenden. 1990. Late e plant 
€ fot legalities in Arctic sub-arctic North 
America—A review of the data. Arctic 43: 364-392. 

McGlone, M. S. 2005. Goodbye Gondwana. J. Biogeogr. 32: 
139—140. 

Mitchell, b D: ge B Heenan 2000 S 1 l 

of N 


1 
E 
14 DNA 


TTS se a ius Syst. Bot. 25: 98-105. 


Mummenhoff, K., H. Bruggemann & J. L. Bowman. 2001. 
peo DNA phylogeny and cp of Lepidium 
(Brassic e r. J. Bot. 88: 2051-2063. 

Murray, DF 1995. Cine of Arctic plant diversity: Origin 


and evalutiom Ecol. Stud. 113: 21-32. 

Nylander, J. A. A. 2004. MrAIC.pl. Program distributed by 
the author. Evolutionary Biology Centre, Uppsala Univer- 
sity, Uppsala. 

Queiroz, A. de. 2005. The resurrection of oceanic dispersal 


loxostemonoides O. E. Schulz (Cruciferae). J. Japan. Bot. 
Ronquist, F. 1997. Dispersal-vicariance analys 

approach to the quantification of historical o 
Syst. Biol. 46: 195-203. 
——— & J. P. Huelsenbeck. 2003. MrBayes 3: Bayesi 
phylogenetic inference under mixed models. Bioinfor- 


903. erg c der Gattung Cardamine. 
t. Ja ih E E st. 32: 280-623. 
6. Cruciferae. Pp. 227-658 in A. Engler & H. 
Harms vi es Die natürlichen Pflanzenfamilien. En- 
gelmann, Leipzig. 
Siebert, S. & R. Backofen. 2005. MARNA: Multiple 
alignment and consensus structure prediction of RNAs 
based on sequence structure comparisons. Bioinformatics 


21: 3352-3359. 


230 Annals of the 
Missouri Botanical Garden 
Simmons, M. P. & H. Ochoterena. 2000. Gaps as characters Webb, C. J., W. R. e a Md Jones. 1908, Pileta 


in sequence-based phylogenetic analyses. Syst. Biol. 49: 
369-381. 
Sjöstedt, B. 1975. Revision of genus Cardam 
(Cruciferae) in South America and Central dente Bot. 
128: 8- 


Sweeney, P. W. i R. A. Price. 2000. Polyphyly of the e 
Dentaria ies aceae): Evidence from irnL intron 
ndhF s e data. Syst. Bot. 25: 468-478. 

ed P. n ‘Ciel ly, G. Panton &T. Bouvet. 1991. Universal 

aie or amplification of three noncoding regions of 

LM oplast DNA. PI. Molec. Biol. 17: 1105-1109. 

Thorne, R. 2004. Tropical plant disjunctions: A personal 
reflection. Int. J. Pl. Sci. 165: e 

Warwick, S. L & I. A. AL 2006. Brassicaceae: 
Chromosome number index and dicbus on CD-Rom. PI. 
Syst. Evol. 259: 237-248 


of New Zealand. Botany D 

a [aes ch. 

J. Taylor. 1990. Amplification 
direct sequencing of a wor dn n ii 

phylogenetics. Pp. 315-322 D. H. Gelfai 

J. J. Sninsky & T. J. White ene PCR Protocols 

Guide to Methods and Applications. Academic Press, E 

Diego 


Yang Y. W., K. N. Lai, P. Y. Tai, D. P. Ma & W. H. Li. 


1999. Molecular ccn he of d oa 
Arabidopsis, and allied genera based on the internal 
transcribed spacer yc of 185-255 TDNA. "Meler. 


462. 
. Pagani, a co E. Thomas & K. Billups 
2001. Trends, rhythms, and o d in Deo a 
65 Ma to present. Science 292: 686-6 


231 


Carlsen et al. 


Volume 96, Number 2 


2009 


Biogeography and Phylogeny of Cardamine 


90£6T8NA 8FI6T8n3 weap (0) uesreqsmdg 9I npjounuppap;) pQofipieq 2) 

ZZIS6LAV 0003 ‘ud x &ouoong *KEMION 9T npjounuppap;) "T ?ofipieq 2 

TIO9092A V FOOT “TE 19 BAOUFT *oooo15) ze aupumpan,) Ásoy[eH seproapsnging 77) 

OZ909TAV OOS “Te 19 eaoyr] “Aux 8v IUYUIDPAD;) "1 vipofimosn 7) 

66161803 cecolena (S) veme, PIe e" SISUJUDSUD 7) 

90€619n75 vm) ‘H (O) uedef ung agdoaoopy "ABS ID “YURI DIO norpuaddo ^7) 

IZL861L4V 0003 ‘ouq x louons “yg 8zl nuno pypysnBun 7) 

crSTOT[A OcGPOT[4 89PPOF[A 'SOPPOP[A Tex ON) VSA 831 pupa z[upog “YO PING UD 7) 

Orz6lgna 86161813 1xo[&e], 29 [LABS “Pe (S) VSN Ob un pardoioo py ommu 7) 
Iles £r 

GpGTOT[4 ccarov[4 LEPPOPLA "POPPOP[A (OIX) 91m eae a "v'sn OF np doo py POH vpnu 7) 

8090974 V Yo07 “Te 19 eaor Ape ZE aupumpan;) Duppaodum 7) 

G8S00ZAV T00c “Te 19 BAOUTT ‘uredg ZE GUID PID) neg m uouuog pupqaodum T 

LVTI61905 S1oquosoA (O) ÁBMION OL 2umuppan?) DA4DUID 77) 

OSS00ZAV 007 “Te 19 eaoyr] “uredg 9T 2umumpan?) vovusléd "dsqus papu 77) 

€£99971V 0003 BM 2 eyzuery ‘uredg 9T 2umumpan?) uouuog napuastd -dsqns ppw 77) 

6LS09GAV T00c€ “Te 19 eaor] ‘A OL oununpav) DADUD "dsqns DADUD 19 

GLOOFZAYV “S86SPZAV POOZ “TR 19 pjoure]y *er[eao[s 91 AUDI) apum -dsqns "[ vapum `) 
8661 

EPEGLOAV ZLOSLOAV 'TI08204V “Te yo exzuexj ‘round moy endeg ZMOS 4 O xe "npqos Duato 7) 

6ZSTOPLA 60STOT[4 O8PPOPlA “EOPPOPÍA gzz (naso) Apr 9T npjounuppap) PIEM Pudo 7) 

8S92P0AV ZPOLPOAV ZZOLPOAV TIOLPOAV ZOOS “Te 19 1exeo[g “Aopenog 91 umpopdoakdpgg TH vupoufo 7) 

GGOLPOAV 6€9LT0A V EZOLVOAV ZIOLPOAV Z007 “Te 19 1exoo[g ^eruezue], 91 wnpphydoskdog TH "p vupvoufo 7) 
ZOOS “Te 19 1039918 *966T 

0S9 T0A V Cr£6204 V OLO8Z0AV ‘60081204V “Te Jo MUR. ‘vous moy ended ot um] hydoskdog QUI Puponfo ye ds 7) 

ZEOOFTAV “ZOOOPZAV FOOT “TR 19 pye ‘299919 ot AUDI) SHOD 7) 

LOO9PTAV “LLOSPZAV OOS “Te 19 PPE *ox8ouoquo]y OL aupumpan;) "JOSUS stop 7) 

GOPPOPÍA “OLPPOPÍA (QM) 281099 pupa SHORAOD DaIsHYOgD 2umumpan) 
1003 ^exmg 9 1exeopg 

PLOZOLAV ZSE6LOAV PZOBLOAV “EZOBLOAV 8661 “TR 19 oyzue1g *&uwuuoz) 8r 'Ssog (T) suusoajs "Y 
1003 ^exmg 9 exeopg 

699£9€,1V TSL6LOAV ZZ08L0AV “TZ08L0AV 8661 “Te 19 oyzue1g ‘Aureus ce ssog (T) suasnynd pdduoy 

1ooeds Jue UOTjUI July SLI 1199 = UZ uonoos Soroodc 


(Re di no [/&nuno7) 


snp 107 poonpord seouenbos 10] srequmu uorssoooe xuequoz) "uoneoiqnd 
o[dnpnur oxou« soroods 10} sodor oxex 01 19jo1 sosot[mored ur syunoo buo 'perpdde usoq sey 11 oroym soroods 10] USATÍ sr uoroog “Apms snp ur popn[our suourroodc 


(d 


qr pus. srt] 


El 


"UOISo1 VNC 4989 Topun pue soroods oes 107 pojsI[ 91e uoneorqnd 
YIM pojst[ ole »puequo) uror uo»xe soouonbog "peuuogpod U99q savy sSjunoo 


I XIGNaady 


Annals of the 


232 


Missouri Botanical Garden 


0992T0A V EPOLPOAV SZ9LPOAV “PIOLPOAV ZOOS “Te 19 1exoo[g “Pue[eoz MON 8p AMD) “Od xe SUPg sipigep 7) 
SO909ZAV OOS “Te 19 e^oqrT “uredg 91 “Log DOSS 7) 
OP9LPOAVY ES9LPOAV PZOLPOAV “ETOLPOAV ZOOS “TE 19 1exoo[g ‘purpeoz MON 8r »jjounuvpap;) DSOQUIÁLO) 7) 
GPOLPOAV 6£€6104V T008L04V *£008204V ZOOS “Te 19 1exee[g ^ereusny 8r »jjounuppap;) ‘J oon »soquiaoo 7) 
LST6ISNA med (o) vs vz AMD) vijofipioo 7) 
JULIJE 
SPZ6I8NA 90Z6T8NA PZEGTSNA y Posoy waup (s) VSA vc aumurp.p») fex) “y vijofipaoo 7) 
P ZISA S0z61903l cc£619134 supy 9 Y ()'vsn Supo 122upi$u02 7) 
16261804 OST6T8na 

“e8z6 Tena *eGT6 T8912 TZE6ISNA euog "T (AT) essy 8r Aozun[ nuafuoo 7) 
ersporla IZSTOPA (0,4) epeueo 987 nuno DIDUAMINO) 7) 
0c€6I98/14 (0) epeue’) OSG DUD WIT DIDUBIDIUOD 2 
61€618903l soJemg “Y ‘q (S) epeueo 987 DIA zwaos “O (XYON) PIDUAPDIUO) 7) 
orzolgn POZ61804 81e61803 projpey 29 preuovT *(S) VSA AMD wosTe M "S xo "o[nnus 87770272 7) 
TST6T803 11904 "f (0) pair vo srjoqosds "p enopijep 7) 
8/809zA V OOS “Te 19 e^oqrT “uredg 91 POUL 29 LAr] pupjjarepo 7) 
ebz618na £0z61804 LIS6ISNA 198unpooy “H "Y (S) 299919 UNINSUOLA UN Y SPIEA PSOUIDO 7) 
Leto [4 Gcetov[4 O6PPOPLA “LOPPOP[A eng Agso) vs ce umpagdoidoamsg auaerg (Pny) voruofio 7) 
TcIS6LAV 0003 ‘tg x louons “WS T9-c€ »soqmq 7) 
€ST6T903l uoppA `O “q (0) WS vO-Z£ psoqng 7) 
Z Z6ISNA ZOZ6ISNA 9TEGISNA xnomueT sor (s) epeueg T9-c€ »soqmq 7) 

Orsrorl PESVOPÍA Tex (OW) ven T9-c€ "quo$80q x suis 
“uoytig (TUNJA xe "qexnpos) vsogyng 7) 
pesporla (CANIN) eurer 96 punuq nsafiqng 7) 
OPE6LOAV S1O8L0AV ‘21108104V 8661 “Te 19 oyzue1g *&ueuuoz) 96 DIU zerg Ddiofiging 7) 
£cI86L4V svud (v5) WS 96-78 SUNUDPAD) quamalg 7) 
10z618N SI£6I8(YH 19104 “MW 19104 "T 75 (8) ven 96-48 aunuppany) uoste y "G Hamas 7) 
81861804 Tousor] "[ Y (S) Rony #809 ZMOS ^q “O 19pD4q 7) 
TPZ6 TNA uvoun(q “OW) nioq 2ununpap?) SISUMIDUOQ 7) 
002619013 TIS6I903 zequeug-pv "v T (s) ropenoq AMD, Slog SISUALIDUO 7) 
so£6 1804 IST6IBNA ZIS6ISNA uoaa Y (0) VSA ZP ‘87 pjjourunpan) ngepsm]q 7) 
TOE6ISAA OST6I8013l £I£61813 uoaa ‘Y (0) VSA cv ‘87 Djjounuppap;) "ase TUJapsmq 7) 
Gcetop[4 elsporla 68vr9b[A 'ooppop[4 (QM) vrssn pupaq ZMOS 74 “O vypuurdiq 7) 
S0£61913l 6FI6T8Q3l OT£6T803l eaomsye (0) erssny 9T njjounuppap;) oyofipyyeq 7) 

qaoeds A ua uomur qua SLI I[?2 = uc uonoog soroedg 
(umurequat]) hae [/Ánuno7) 
"penuguo) ë ‘I XIGNdddy 


233 


Carlsen et al. 


Volume 96, Number 2 


2009 


Biogeography and Phylogeny of Cardamine 


19161905 of] "v (O) Aemioyy 9T 2umumpan?) suaimdun 7) 
97108204V *GT08204V 8661 “TE 19 exzuery *Kueuuoz) 91 aunuppany) "p suermdun 7) 
866LL0AV ‘L66LLOAV 8661 “Te 19 oxzue1g *&uwuuoz) 91 aupumpan;) "p omy 7) 
Erré EI ZIZ6ISNA Sorrorla ‘TLPPOPLA iprequisg (NASO) ATL 87 puva ZMOS 74 ^O CILA) 97 ydorday 7) 
PSZ6ISNA 6€£61803l 96F (OIN) euo SUYUDPAD) UOSUOYL, P J ooH MMS 7) 
POT6ISNA ueexofng Ņ ezig (0) eUeuloy 91-91 umanguo4944 "| 922248 7) 
ESZ6I8N A TIZ618M4 310q [9308 *(s) Apa] 9T umanauo42q “Suoldg ponp]s 7) 
TreTOT [4 6LSPOPLA (IN) erssny DunDquaqq DIAFINPUD]S 7) 
OTGTOT[A SISTOLA (IN) owen nuno DISFINPUD]S 7) 
(&ueurroz)) 
€9I6I9(143 8€€6I9/14 uoprez) [£?ruejoq SinqsuoSow GP DUDA ZIBMUOS "0 DLBSYNPUDIS X 
6P9LPOAV GEO! TOÀ V LCOLVOA V “OL9LPOAV 3003 “Te 19 1exoo[g “eunuoSiy ZL aunUDpsD:) supiopjs 7) 
SPOLPOLV PEOLPOAV OZOLPOAV “SIOLPOAV ZOOS “Te 19 1exoo[g “op ZL aumupan») “Od (15104 75) 91701018 7) 
£T909ZAV FOOT “Te 19 eaoyr] uredg 8r (ce) 097] NEN SeAry (zureT W) 021230108 7) 
Tec61803l L€€61903l Tc00c (OIN) Puno spiq vunraysuny’ 7) 
OSZ6T8na 9€£61803l LIS LOW) FUND SUYUDPAD) z[mpos ^q 0 Pyofinnsnsy 7) 
ZOOS “TE 19 Yoo 
PPOLPOAV LES6LOAV 0082L0AV ‘666LL0AV *866T “Te 19 exzuerg *Aueuuoz) ce AUDI) "pr psonxoyf 7) 
PEEGISNA S1eqsnoxs (0) epp) 91 aunuppany) TPS Y uq) pronif 7) 
LESPOPÍA STSPOPÍA POPPOVÍA TLVPOPÍA ZSIS (NSO) uomugy *eujsuy 08 pupaq zuer) (T) soprapdpouua 7) 
OTZ61804 eeeorsna ayyerg, ‘(S) ropenoy AUDI) "uorerp] SISUALOPDNIA 7) 
colergna yneureyy (0) epeuep ZLB üsspjgnop 7) 
LECOTSNA 60261812 c£e6T9na peure (s) "ven ZLB uonug 2SSD]$NOP 7) 
TESTOT[A gcerov [a €6VVOT[A “OLPPOPÍA (04) W's’ Duoga zequeqs-[v (musaeo"]) DPəSSP 7) 
6rcolgn 80261813 TELGTSNA neopuo[g *(s) epeueg 982-96 puva Poom "udpy (IN) Py 4ydip 7) 
EESPOPÍA ZLSPOPLA (IN) erssny cv 87 Djjourump.p») DIDYÍIP 7) 
COPPOLA '6OPPOT[A (IN) erssny cv 87 Djpourump.p») DIDYÍIP 7) 
z6z61804 Totergena 
‘POS6ISNA ‘O9T618N4 82261804 ‘LZEGISNA uoA[q Aq 04 (O) epeue) cv ‘87 »jjounuvpap;) npn 7) 
ZOS6TSNA 6ST6T8N a 6c£61803l uoaa ((Q) WS cv 87 Djjourump.p») DIDYÍIP 7) 
8ST6T8n3l 0€£61803l uoaa ((Q) WS cv 87 njjounuppan;) uospreqorg 2Dp2:p 7) 
LOZ6ISNA 9c€6190l vumenzi ‘W (s) uedef ums pjpjadmuop 7) 
levrov[A BOPPOPLA (IN) vaste, "gosmo) n4ogfisuop 7) 
Srzorsna Sce6rgnsa 6PSPI (OIN) Puno ppuedg Mpanjap 7) 
1oowds j-Tqu4 UOJ]UI July SLI 1199 = UZ uonoos sor pode 


(ener) liiis 1/400) 


"penuguo) “[ XIGNdddy 


Annals of the 


234 


Missouri Botanical Garden 


Scc619/13 zequeus-[v (S) topenog wing dqdoskdog pypao 7) 
PZZ6ISNA eseorsna zequeys-Ty *(s) ropenog um Aqdosádog "pueg pao 7) 
gocelgn cceorana ZzeLse OSO) ve T9 aumupan») smppuapioo0 7) 
69761804 TSE6ISNA LT6SEE Oso) VSA T9 aumupan») ZMO H “O (uosre M) s1Jpzuaproo 7) 
£cc6l803 pras “y Ys) euog TLE aunuppany) onbrqo 7) 
PSOLPOLV SE9LPOAV 6ZOLPOAV ‘8TOLPOAV ZO0Z T 19 1oxoo[g ^erdorrq TLE aumupan:) »nbijqo 7) 
ZSOLPOAV O€9LPOAV 8ZOLPOAV ‘LIOLPOAV ZOOS “Te 19 1exoo[g *eXuoy ZL-9E aumumpan:) "png ^y xe aeuoopp nbijqo 7) 
292619031 £cetor[a OSEGISNA rope, (OW) enuopipe) “VSN wunyjkydodoang 9uoo1?) npgmu 7) 
6v£61905 [P16 ‘emurey, qw (s) uedef Djjounuppan?) ABS 29 uouergy poruoddiu 77) 
992618131 0cc6T8n3 8F£6T803l yoy wf (S) tuo aururp.p») z[nuos 74 ^O PENZOL 7) 
ceetov[4 (MW) ersoy T0-96 »jjounuvpap;) ppajdoaonu 7) 
OLI618N a vAoK|q “(Q) essy 0-87 njjounuppan;) ppajdoaonu 7) 
06261814 €L1618N a 4981101 o9 Aouoeuo], (qT) erssuy t987 Djjourunpan) pjjAydosmnu 7) 
9PEGISNA vaoynyz (IT) erssny 87 njjoununpan;) ppajdoaonu 7) 
¿2161804 LY£61903l vAozo[|oq 29 AeqoeuroT, (AT) ersoy t987 pjjourunpan) swepy vjpapdouorut 7) 
LOS99TV OS€6L0AV 8661 “TE 19 oxzuexj *eppeao[s or AUDI) yonu 7) 
TLO907AV 0003 “BMP Y exzuexq ee png or AUDI) Yomo 7) 
90909cAV T00c “Te 19 BAOUTT “eIUQA0TS 9T SUTUIDPADO T[[9u107) xo MƏN JOHDU ~F 
86 vrob[A ‘SLPPOPLA CANIN) Prssny 96-49 np Crydo.on py nydydossput 7) 
E9Z618N4 SPE6ISNA zequous-[v (OW) uedef 96-49 np Crydo.on py PIE 97¡4ydovomu 7) 
ez£618na supy “DY (S) VSA aumumpan:) uosm A "S 217047 7) 
cocelgna LIZ6ISNA ereorana Suo 39 pewag (8) ven ppeuroq 1340] 7) 
LPOLPOAY OPE6L0AV 9008L0AV *6008204V 8661 “Te 19 exzuex] *erensuy 9r DUDA] 7) 
6S9LPOLV TPE6LOAV 8008L0AV ‘21008104V 8661 “Te 19 exzuex] “erensoy 9r "OOH, DUDA] 7) 
Tocol8na cPe6T9n3 LOZT (OI) vun) zequeqs-v vunisuayy 7) 
L6VVOVA “PLPVOPÍA CANIN) erssny ot umpyadoaon py mipupona] 7) 
Ote6I9n osuim4 “JA (s) uedef 91 umpyadoaon py DYYUDINA] 7) 
092619131 9TZ6T8NA vuigsnzry “JA “(S) uedef 9T umpdoaoo py zmag “YO Pyjunona] 7) 
0003 ueuoopr 
€8900 LAV *ueuooRq m Te eet *pue[eoz MON 9r (uosuuof 'N 'd m souof-" uwes) SLISNID] 2 
GIZ618N soJemg *(s) epeuen Duoga poo CQN) von] 7) 
¿Sz618na PIZ6ISNA UQUIS (s) puepozmg Duoga njqim 7) 
SESTOLA OLSPorl O6PPOPLA “ELPPOPÍA (NM) *eurAoSozio]]-erusog DILO "[oeg maqrony 7) 
TSOLPOAV PPE6LOAV PLOSLOAV “ETOBLOAV 8661 “TE 19 oxzuexq ‘Rous MoN Zog “H “O Hasskay 7) 
1oowds g-un UOJ]UI Tt.4 SLI 1199 = UZ uonoos soroedg 
(umurequat]) hee [/Ánuno7) 
"penuguo)) ë ‘D XIGNdddy 


235 


Carlsen et al. 


Volume 96, Number 2 


2009 


Biogeography and Phylogeny of Cardamine 


PLOTS TEZ6I8SNA 29861801 weqpo (e) vis AUPUDPAD) “MIT Dyofipunzos 7) 
0£Z618M4 99861804 yooy (8) eunp) aurumpav:) zpnuog “4 “O 115204 7) 
181393 1V ‘TOZSOZAV 0003 MH Y exzuexg Pue mg 9T suoman 7) 
G6€9971V 9€6204V Z86LLOAV “I86LLOAV 8661 “Te 19 exzuexp “erueoy 9T MUYOS suona 7) 
6666181 Sog6T80 zequeys-T¥ (Ss) topenog suo DIDUOZMpL 7) 
O£CTOT [4 OTSTOV[A VOSTOT[A TSPVOVÍA EZT (n8s0) Área 91 pjjounipao") pyofipasaa 7) 
€Lc6l803l 83261804 pOeOl8Na uossponueurg (s) eLusny 9T vjjeunuppt;) "] ?uofiposat 7) 

9€ESPIPÍA PLSPOPLA POSPOPLA ITepror [a (MIA) eurer DLIDIUA "j[eumos xo 7j 00H 
3 pug Cqeig T) ?rjofonbumb 7) 
TESPOPÍA TISVOPÍA (MW) ersoy 08 Djjoununpan) vomdind 7) 
P8T6T8na 65€618031 pung y uong (0) VSA 08 pyaunuDp.D’) vomdind 7) 
£6c61803l EST6ISNA 9S€618021 ospy ‘soz Keung (0) VSA 08 pyaunuDp.D’y vomdind +) 
82361804 z8161804 To£6T8na uss14 (0) VSN 08 ngjounuppan) naindand 7) 
T8T6T8903I 098618021 uss14 (0) VSN 08 Djjourupap) vomdind +) 
POZOTENA 98T6T8NA g9£6 1804 Aysaoneg (gp) erssng 296 Dyjaununp.n’) voandand 7) 
S9816T1901 c9e619n4 Aysaoneg (HT) erssny 296 pjjounuvpan,) "IP3pU?S p cue) vaundind 7) 
€0STOF[4 “O8SPPOPÍA (MW) ersoy Oc ‘ST aumumpan) suado404d `) 
081618903 vdysposueyyry "qv (ET) essy 0c “81 aumuppap?) suada404d °F 
6Z1618N4 Aosumy (HT) eissny 0c ‘SI 2umumpan?) "DG xe "uyosrq suadasoid 77) 
SGOOPZAV 'G66SYcA V poos “Te 19 pjourepy Tesmuog 981I-9I 2umumpan?) "Ty sisuamid 7) 
ZOSPOFLA “6LPPOPÍA (Lvo) Aper 9T vjjeunuppu;) TEA tenund 7) 
0099971 Y EP9997AV 0003 MH Y exzuexg teuesng 91 PEYEW BY AeQuy nsazuad 7) 

(ops jo umosnpy Át01STH 
LSEOTSAA Termuew *ueprez) peorumiog) £002-0£0£A 8r DUB ziuen (71) sopcydoquad 7) 
8ZT6T8NA Anq 3 ddey oq (VIV) ven v9 ‘ZE porupajÁsuad 7) 
22161804 Ang (VIV) VSN vo ‘ZE “PILA xo TIN porunazásuad 7) 
21361804 9ZTeT8(14 ÁxS^0rjoq 
“91361801 “GL161804 98£618NA a sosm y ‘waoynyz (qT) essy 0€ vjjeunuppu;) Sur], Y [o3oy vwpəd 7) 
9S0! 50A V OFOLTOAV OL9LPOAV "6192 pOA V ZOOS “Te 19 1oxoo[g Lensy “210, oSnfionpd 7) 
1LZ618N0H SSe618031 POESbs (Oso) VSN 'pueg 4 "T Huosiagmd 7) 
TOSPOPÍA BLPPOPÍA emy '(ngso) emeny 9T aununpan) nsopfiaand 7) 
PSE6L8NA yoours’y *(Q) Bryeaoys 9T 2umumpan?) T 240gfiaad 77) 
OOSPOFLA *12VFOT[A SOET (ANVI) rouma moy endeq um Aqdoskdog z[pnjog “q O Punndod 7) 
02261804 zequeus-[v (OI) 10penoj um Aqdoskdog pao 7) 
1oowds j-Tqu4 UOJ]UI July SLI 1199 = UZ uonoos sotoodg 


¡eo P [/&nuno7) 


"penuguo) ‘I XIGNdddy 


Annals of the 


236 


Missouri Botanical Garden 


'(o00c) zequeus 


-[V pu? xor«reA pue (2007) “Te 29 emy ur suore[iduroo OU} MO[[0J SMOS 9UIOSOUIOJQ) “YUNOY 9UIOSOUIOTI[O ƏY) I0J SE IOYONOA OUIES OU] ST o[dures VN 9U1 107 Toqonoa IMO Jey} Soyeorpuy , 
-roded sup ur "T DUDILISO ue ‘ds "9 pewa ‘uoxe) ojeredos e sjuosoxdox ueuiroods $m ey} 9jeorpur Ápms mo wo Sj[n $91 mq ‘punon fo SUMUDPAD() se porjuopr SEM ueuiroods sni q 
"UOAIG ole 99UO9I9JOI pue Axogroj/&nunoo Ayuo “yuequos) WOT po4ornuor uooq GARY Soouonbos 9J9UA » 


S8z619131 T9€61903 09€z1 (OIN) tur) SUYUDPID) "poueag smuaupuunk 7) 
SOSPOPLA 'SGPPOP[A CANIN) erssny c£ umppadoaon py "unxepy s18u20z2Á 7) 
96261813 €Te61803l quoy (s) euasny DUD WAT MUIaISP|DM 7) 
8€Z6T8NMA 1ouddoT *(g) eunsny DUDAT MUIASPIDN 7) 
SELS6L1V 000% “QoL m ÁQUIIMS DUD UAT TUIS pon 2 
6gcyov[4 LISTOV[4 LOSPOFLA "repro (a (NM) wursoSozrog-erusog nuno wok] murapsppoa 7) 
S616191 uoaa H Aq "IOA (0) essay 9c pjjourumpa?) 8110210. 7) 
66261913 €6161903l T8E618N4 aoydoroyy (qT) erssny 87 D]9UVUADPAD) 81401914 7) 
86261913 T6l6I913 c9£61913 BAONNYZ 29 aosymyA (qT) essy 87 D]QUVUADPAD) 81401914 7) 
S6I6I9(13 €8€6I8(13 ASSIM m AOXIUAOUZON “(TT ) "y BIssny 8% DI]9UPUDPAD yosng 'N 81401210. 2 
S6Z6TSNA T6161803 82€61813 usaq (0) 'V'e'n 8v ZE D]QURUDPAD) nmjoqum 7) 
L6561913 061618913 088618913 usaq (0) VSA 8h ZE D]9URUDPAD) nmjoqum 7) 
962619131 68161803 62€61803l uma (0) VSA 8v ‘ZE D]QURUDPAD) 9uoo1) DID]JOQUIn 7) 
ILO9PZAV ‘T86SPZAV OOS “TR 19 PEN *er1092) 91 aumunpan,) pig IN psow 7) 
ersrorla OcGTOT[A 90STOT[4 “ESPPOPÍA jprequiog *((ngso) sny 91 ung q doro) "T yof 7) 
£COLTOAV LEOLPOLV C£OLTOA V TZOLPOAV ZOOS “Te 19 1Texoo[g *eruezuv], c£ “OT aunupap) ndapooionat 7) 
LSOLPOAV TFOLTOA V I€OLPOAV “OZOLPOAV ZOOS “Te 19 1expo[g *erdonprg c£ “OT UMD) "pg ^y xe a9»o[p ndepoototat 7) 
PSZ6ISNA LEZ6ISNA LL£6T8NA vuigsnzry “JA (s) uedef 98 $Hu2110] 7) 
9€Z61804 vinumqso x T (s) uedef 9c TENEN $2224401 7) 
6PSPOPLA LESVOvÍA CANIN) erssny c£ Dznp410042Dt[ds pyofimua 7) 
£8z6 1804 vaod y Áxsaomog “(s) essy ce Dznp41010429Dt[ds m] (qoperT) »ijofimuar 7) 
OIO9PZA V ‘O86SPTAV POOZ “Te 19 p[ourepy ‘erssny 91 SUPUIDPAD) “ASIN Y 7) xe RUIN "2 "S 742423 7) 
S£2619/13 ysmbuory (s) VSA ung hydordosnsy zpujos 74 ^O (emd) 9772421 7) 
Z8Z618N 4 PEZ6ISNA 9LE6ISNA zmuos *(S) euno cv nuno z[npos ^q OQ umuomsupt 7) 
098261913 S£61913 enesomy (OW) uedef SUPUDPAD) apypuny 7) 
18ez6 1804 ££261813 PLEGISNA osnm TUSOÁTIA *(s) uedef aunupap;) “ABS 29 PULI apyount 7) 
19161903 €2£61903 A98]M[ 29 AOXNTUA9UZON (HT) erssnyy 27 Djjounuppan?) (od.josr) ^ozum[ ojpdouagds 77 
SUL SS Y "YT 
LOISZLAV 6661 “Te 19 Suey ‘uemrey, ras "S p (eyekepq) vupsouuof rea orypgmos 7) 
ZLEOISNA LECOISNA OLE6ISNA usq (s) uemte ce DIO mos 7) 
6LZOTSNA ZLEOTSNA (peogTOIN) uedef ce "qun Pomos 7) 
82261814 Zez618na 89861813 ysmbuory (s) VSA "pong (MYOS) vjooidna 7) 
1oowds 744 UOJ]UI T4 SLI 1199 = uc uonoos sotoodg 
ara dans [/&nuno7) 
"penuguo)) ‘I XIIN3ddV 


A TAXONOMIC REVISION OF THE Chen Jin-Yong,2** Zhang Zuo-Shuang,? and 


SYRINGA PUBESCENS COMPLEX 
(OLEACEAE)! 


Hong De-Yuan? 


ABSTRACT 


Over 19 species and numerous infraspecific taxa have been described in the Syringa pubescens Turcz. complex (Oleaceae), 


o has 
n across China and 14 p 
species is recognized, with thr 


new sony of S. p 
synonyms of 
Nakai are ordes ignated here 

Key words: Oleaceae, solid sampling, statistics, 


been Rug abun To pro 
pulations were sample 
components analyses and general statistical ee Solus the significance of characters for taxonomy. As 
ee subspecies: Syringa pubescens subsp. pubescens Turcz. 
Chang & X. L. Chen, an iaa ies patula. des C. Chan, 
nd eC. k Schn 

S. pres subsp. M M pol. Mrd microphylla Diels, S. dielsiana C. K. Schneid., and S. venosa 


ovide a rational taxonomic revision of the c complex, field 
ed. The principal coordinate and principal 

a result, one 
pea microphylla (Diels) M. C. 

g & X. L. Chen. Syr bon ri C. K. Schneid. is treated as a 
eid. and S. meyeri var. spontanea M. C. Chan, 


ng as new 


Syringa pubescens complex, Syringa, taxonomic revision. 


The Syringa pubescens Turcz. complex (Oleaceae) 
"e to section Syringa L. ser. Pubescentes (C. K. 
Schneid.) Linge Ish. 1920; Chang, 
1992). I 


(Lingelsheim, 
rs and verrucose fruits, and is 
widely distributed in China and the Korean Peninsula. 

Turezaninow (1840) briefly described Syringa 
pubescens as new, stating that it grew in the mountains 
of northern China. From 1900 (Palibin, 1900) to 1990 
(Chang & Chen, 1990), more than 17 species and 
additional rpm taxa were described. A total of 
19 new species and varieties have so far been 
described in the complex. The taxonomy of this 
complex has been controversial. McKelvey (1928) 
recognized seven species, whereas Chang (1992) and 
Chang and Green (1996) recognized two species, S. 
ri C. K. Schn 


species comprises four subspecies: subspecies pu- 


pubescens and S. meye eid. The former 
bescens Turez., subspecies patula (Palib.) M. C. Chang 
X. L. Chen, subspecies julianae (C. K. Schneid.) M. 
C. Chang & X. L. Chen, and subspecies microphylla 
(Diels) M. C. Chang & X. L. Chen, which overlap in 
distribution in China (Chang, 1992). Syringa pu- 
bescens subsp. microphylla includes three varieties: 
microphylla, flavoanthera (X. L. 
Schneid.) P. " ‘Geen &M 


yringa meyeri consists of two EIE, 


ang, 
and aa (C. K. 
Chang. $ 


variety meyeri and variety spontanea M. C. Chang. 
However, Qin (1998) raised S. meyeri var. spontanea 
to specific rank as S. spontanea (M. C. Chang) X. K. 
Qin, considering it to have smaller, suborbicular, and 
palmately 5-veined leaf blades and densely pubescent 
branchlets. Therefore, three species, four subspecies, 
and three varieties are recognized for the S. pubescens 
complex in the most recent literature. 

Indumentum on leaf surfaces, inflorescence rachis- 
used to describe 
new species in Syringa. For instance, Komarov (1901) 
and Schneider (1910) stated that S. velutina Kom. and 
s. potanini C. K. Schneid. differed from $. pubescens 

n being hairy on both sides of the leaves. Nakai 
(1913) and Skvortzov and Wang (1955) described S. 


palibiniana Nakai and 


es, and calyces has been frequently 


a 


Wang based on their sidbrom leaves, different from S. 
velutina. Other characters used by previous authors 
have included size of plants, inflorescences, and leaf 
blades; shape and venation pattern of leaf blades; 
length of corolla tubes; color of anthers; and position 
of anther insertion. For example, Chen et al. (1989) 
described S. microphylla var. flavoanthera X. L. Chen 
as new based on its yellow anthers. 

Taxonomists have described new taxa based on a 
limited number of specimens, which caused quite a few 
superfluous names. For instance, Schneider (1905, 


"The authors are grateful to the National Natural Science Foundation of China (grant 30500036) and the Beijing 


Administrative Bureau of Landscape p Ue for support. We thank the curators of horbana A, BM, E, FI, HIB, 
T, UPS, and WUK for 

i Evolutionary Botany, Institute of B. Chinese Academy of Sciences, 

hongdyGibcas.ac.cn 

g Botanical Garden, Wofosi Road, Beijing 100093, People's Republic of China. 


IFP, K, LE, Do P, SDFS, SDNU, SH 
2 State Key Laboratory of Systemat a 
Xiangshan, Beijing 100093, Pople’ s Republic of China. 
? Beijin 


HNWP, 


ecimen loans. 


aduate School, Chinese Academy of Sciences, Beijing 100039, People's Republic of China. 


^ Grad 
doi: 10.3417/2006072 


ANN. Missouni Bor. Garp. 96: 237—250. PUBLISHED oN 7 JULY 2009. 


Annals of the 
Missouri Botanical Garden 


1910) described Syringa dielsiana C. K. Schneid., S. 
potaninii, and S. giraldiana C. K. Schneid. as new from 
the Qinling Range in Shaanxi Province; these species 
and Chen 
(1990). New species in Syringa have po based on 


were later lumped into one taxon by Chan 
minor differences. For example, Tang (1941) described 
S. trichophylla Tang as new, although it is scarcely 
different from S. microphylla in the denser indumentum 
on leaves and anther nearer to corolla throat. The 
objective of the present study is to examine the 
variation of morphological characters based on field 
observations, populatio 


on sampling, and subsequent 
multivariate al 


q 
analysis, and finally to taxonomically 


revise the S. pubescens complex. 


MATERIALS AND METHODS 


A total of 14 populations were sampled across 
China (Table 1). In addition, a large number of 


examined from the following herbaria: A, BM, E, FL 

HIB, HNWP, IFP, K, NWFC, P, PE, SDFS, SDNU, 

SHM, TI, UPS, and WUK. The characters used for the 
1. 


analysis were coded as in Appendix 


PRINCIPAL COORDINATE AND PRINCIPAL COMPONENTS ANALYSES 


In addition to the 14 populations mentioned above, 
specimens from the Korean Peninsula were used to 
ba d the Korean population (KOR) and speci- 

ns from Hubei and ee to represent another 


un (OTH). The MR determined as 
Syringa me d a Y i 


D 
oO 


yeri were represented as M 
analysis. The type specimen and other herbarium 
specimens were incorporated. into the analysis to 


tion. By calculating Gower similarity coefficient, the 
principal coordinate analysis ce 3.0 software; 
Kovach Computing Service, Anglesey 
ot 
populations or data sets. By using principal compo- 
nents analysis (PCA), 
mostly to the first three principal components were 


the characters contributing 


selected for further variation analysis in order to 
evaluate their taxonomic value. 


CHARACTER VARIATION ANALYSIS 


For selected characters, the range, mean value, 
and standard deviation of each population were 
calculated to determine whether the morphological 


variation was continuous or discontinuous among the 
populations. 


CHARACTER ANALYSIS AND RESULTS 


The principal component analysis showed that the 
first coordinate accounted for 20% of the variation 
8.4%, 


respectively. The main characters contributing to the 


the next two coordinates, 14% an 


w 
= 


first component included the length and width of leaf 
blades, indumentum on inflorescence rachises and 
calyces, diameter of corolla throats, and anther 
position on corolla tubes. The characters including 
length and shape of corolla tubes (CTS), and distance 
between anthers and corolla throat contribute mostly 
to the second component. The third component 
contributes much less than the first two components. 
The characters contributing mostly to the first two 
principal components and those used by previous 
authors for describing new taxa are analyzed below. 


SIZE OF LEAF BLADES 


This is a major character in the PCA. According to 
our observations, leaf blades varied continuously in 
size among populations, but they were somewhat 
larger in populations KOR, DAN, and LIA (4.4-8.2 X 
2.5-5.5 em) than in other populations (1.5-6 X 1.1- 
3.3 em) (Fig. 1A, B). The group comprising KOR, 
DAN, and LIA corresponds to Green and Chang's 
(1995) S. pubescens subsp. patula. 


SHAPE OF LEAF BLADES 


Shape of leaf blades was converted into the ratio of 
length to width in our analysis for effective compar- 
ison. The character is less important as shown in the 
PCA, and its variation ranges overlap considerably 
among populations (Fig. 


INDUMENTUM ON LEAF SURFACES 


Because the indumentum is a very variable 
character, we used five grades to describe it 
(Appendix 1). The analysis of all population samples 
and herbarium specimens showed that the leaves in 
most populations are usually glabrous on the adaxial 


forms (Table 2). The abaxial surface of the leaves is 
usually sparsely pubescent (grade 1 or 2) except in 
KOR, QIN, OTH, and TIA, which also have densely 
pubescent forms, and in JIN and SHM, which also 
have glabrous forms (Table 2). Plants from the Qinling 
Range of China and the Korean Peninsula tend to 


Volume 96, Number 2 al. 
2009 Revision of the Syringa pubescens Complex 
Table 1. The Syringa pubescens complex in China. Fourteen populations and three data sets (*) are listed approximately 


from east to west in China and Korea. 


Population Population 
Population site Voucher collection size (n) 

KOR* Specimens from the Korean Peninsula E. H. Wilson 8602, K, PE 15 

DAN Liaoning Prov.: Dandong, Mt. Wulong, sunny slope or J. Y. Chen 04108, PE 8 
semi-shade ps forest, 400—600 m 

LIA Liaoning Prov.: Mt. Fenghuang J. Y. Chen 03246, PE 7 

ianshan, ee deciduous forest or 

on the summit, 500-600 m 

JIN Liaoning Prov.: Jinzhou, Mt. Dahei, sunny, open thicket J. Y. Chen 03118, PE 15 
at the summit, 400 m 

WUL Hebei Prov.: Mt. Wuling, semi-shade forest, 1000-1600 m J. Y. Chen 03180, PE 17 

BAI Beijing: Mt. Baihua, shady deciduous forest, 1200-1700 m J. Y. Chen 04135, PE 12 

HEN Shanxi Prov.: Mt. Hengshan, semi-shade deciduous J. Y. Chen 03208, PE 8 
forest or open place at the summit, 1400—2000 

SHM Shanxi Prov.: Taiyuan, Mt. Tianlong, semi-shade forest J. Y. Chen 04102, PE 3 
verge, 1300 m 

TAI Shandong « Mt. Taishan, sunny, open thicket or J. Y. Chen 04207, 04210, 4 
forest m 1400-1500 m PE 

OTH* Specimens from Hubei Prov. and Chongqing Z. E. Zhao 2055, HIB 11 

HUA Shaanxi Prov.: Mt. Huashan, sunny, open slope or J. Y. Chen 03143, 03153, 16 
semi-shade thicket, 1600-2000 m PE 

QIN Shaanxi Prov.: Mt. Taibai, shady deciduous forest or J. Y. Chen 03021, PE 5 
forest margin, 1200-1700 m 

ZHU Shaanx v.: Zhuque National Forest Park, moist, J. Y. Chen 03158, 03160, 11 
semi-shade deciduous forest, 1700-1900 m PE 

HUO Shaanxi Prov.: Ningshan, Huoditang, semi-shade J. Y. Chen 05091, PE 4 
deciduous forest verge, 1900-2000 m 

TIA Gansu Prov.: Tianshui, Putaoyuan, dry, sunny, open J. Y. Chen 04036, PE 10 
thicket, 

KON ansu Prov.: Mt. p shady deciduous forest, J. Y. Chen 03171, PE 6 
2000 m 

MEY* Specimen determined as Syringa meyeri F. N. Meyer 23032, A 4 


have denser indumentum on both sides of the leaves 
the 


indumentum density cannot be used to delimit taxa 


than those from northern China. However, 


in the complex because of its continuous variation. 


LEAF VENATION 


The leaf venation of Syringa is generally pinnate, 
although palmate venation has been described in 
meyeri var. spontanea. (Chan, Chen, 
1990; Green & Chang, 1995). Having. critically 
examined the type specimen of S. meyeri, we found 
that the basal two pairs of veins do not connect to form 
palmate venation, the veins are 
(Fig. 2A). Sim 


pinnate veins in the type specimen r s meyeri var. 


though 
ilarly, there are both palmate ae 
spontanea (Fig. 2B). Meanwhile, a venation similar to 
meyeri 
frequently exists in the populations from Shandong 


that of S. meyeri and S. var. spontanea 


and Shanxi provinces. Therefore, leaf venation is not a 


diagnostic character to differentiate S. meyeri and S. 


meyeri var. spontanea from the other species. 


INDUMENTUM ON INFLORESCENCE RACHISES AND CALYCES 


This was an important character as shown in the 
PCA. The populations KOR, SHM, and HUA have 
inflorescence rachises varying from glabrous to 
WUL, BAL and 
they are mostly glabrous and densely hairy in 
2). Similarly, the 
WUL, BAL, HEN, and MEY 
have almost glabrous calyces, TIA has hairy calyces, 


a hairy, whereas in DAN, 


" remaining populations (Tabl 
populations DAN, LIA, 


and the remaining populations have both glabrous and 
hairy forms (Table 2 
SHAPE OF INFLORESCENCE RACHISES 


The shape of inflorescence rachises is difficult to 
distinguish in practice, especially when inflorescence 


240 Annals of the 
Missouri Botanical Garden 


oo 
T 
i] 
1 


e 


OH 
ILr 
<0 
in 

1+ 

| Erj 

+ 
++ 

m mm 

Cnr 
cir 

1173 


Length of leaf blades (cm) 


0 1 1 
KOR DAN LIA JIN WUL BAI HEN SHM TAI OTH HUA QIN ZHU HUO TIA KON MEY 
Population 
A 
T 
EE: 
u 
$^ ya | 
5 
z4 HU 
Y 
$ 3 FUT a e a epa] sal == -7 
a HERA ABH ADAH 
Si 
E 1 L L L 1 L I i I 1 1 1 1 1 1 L i | 
0 


KOR DAN LIA JIN WUL BAI HEN SHM TAI OTH HUA QIN ZHU HUO TIA KON MEY 


Population 


N PN C) C2 


leaf blades 


L L L i 1 fi fi fi fi fi 1 fi fi fi L fi J 


had sobgyd bladds 


Ratio of length to width of 


KOR DAN LTA JIN WUL BAI HEN SHM TAI OTH HUA QIN ZHU HUO TIA KON MEY 


Population 


C 


Fig Box plots depicting morphological variability of leaf blades in the Syringa pubescens Turcz. complex. —A. 
ae of leaf blades. —B. Widths of leaf blades. —C. Length to width ratios of leaf al Character ranges includes means 
and standard deviations. Abbreviations of the 17 populations are explained in Table 


Volume 96, Number 2 
2009 


al. 
Revision of the Syringa pubescens Complex 


Table 2. The number of individuals of each indumentum coverage (see Appendix 1) on leaf surfaces, inflorescence 


rachises, and calyces among populations of the Syringa pubescens complex. 


Adaxial leaf surfaces 


Abaxial leaf surfaces 


Inflorescence rachises Calyces 


Population 0 1 2 3 4 0 1 2 3 


4 0 L 2 3 4 0 1 2 3 4 


KOR lo 1 4 2 3 17 4 
DAN 8 1 1 
LIA 9 2 7 
JIN 18 4 8 6 
WUL 19 1 3 15 
BAI 14 1 13 
HEN 9 5 4 
SHM 7 1 2 2 4 
TAI 18 1 1 7 11 
OTH 4 6 1 1 1 1 3 6 3 
HUA 18 1 6 10 1 
QIN T7 4 4 3 1 1 8 7 
ZHU 8 2 10 
HUO 9 3 6 
TIA ll 1 3 4 8 
KON 6 1 5 
MEY 4 1 3 


5 8 1 8 6 3 EZ cl 2 5 
6 2 8 
4 2 3 6 3 
3 15 3 1 4 10 
19 19 
10 c 3X. 1 13 1 
6 1 2 9 
4 l1 3 5 1 1 1 
1 3 7 8 13 1 1 4 
1 3 2 7 8 1 2 2 
11 2 5 14 1 1 2 
3 1 3 15 2 1 3 13 
3 4 3 7 2 1 
2 1 4 l 4 
3 2 5 8 L 2 5 T 
2 3 1 3 2 1 
3 1 4 


rachises are hairy. From our observations, inflores- 
cence rachises are obviously 4-angled in the popula- 
tions DAN, LIA, BAL, WUL, HEN, and SHM 
4-angled, in the 
populations. 


, and are 


terete or subterete remaining 


SHAPE OF COROLLA TUBES 


The shape of the corolla tube is quite subjective. 
We used an alternative method in the analysis so that 
different populations could be compared. First, we 
measured the diameter of corolla throats, which varied 
from 1.6 to 2.5 mm in the pee KOR, 
and LIA, 0.9-1.6 mm in JIN, S I, QIN, HUO. 
and TIA, and 1.3-2.2 mm in Pu. us (Fig. 3A). 
Then we calculated the os CTS (Appendix 1), which 
varied from .2] in the populations KOR, 
DAN, and LIA, and = than 0.12 in the others 
(Fig. 3B). Therefore, the populations KOR, DAN, and 


LIA are quite distinctive in the complex. 


COLOR OF ANTHERS 


hers are usually purple in this complex. Our 
extensive mina on 


ound three collec- 
tions with yellow anthers, i.e., G. Giraldi 740 (FD, 
T. N. Liou 190 (IFP), and S. patula cultivated in 
Beijing Botanical Garden, which has both yellow and 
purple anthers on the same individual. Thus, we 
assume that yellow anthers are occasionally variant 
and not a good character for the taxonomy of this 
complex. 


ANTHER INSERTION ON THE COROLLA TUBE 


our observations, the distance between 


anthers and corolla throat is less than 1 mm in the 


S 
and 0.4-2 mm in TAL, QIN, and TIA, 
showing Ha variation among the populations 
Fig. 3C) Therefore, the distance between anthers 
and corolla throat is not effective for delimiting taxa. 


— 


CAPSULES 
The Ta apices in k c (DAN, 
LIA, JIN, , HEN, OTH, ZHU) are ae 


obtuse by Een althoug die sete: Thus, 
capsule apex is insignificant in the taxonomy èf > 
complex. 

The results of principal coordinate analysis (PCO), 
s all indicate that 
only one species can be recognized in the complex, 


PCA, and character variation analysi 


since there are no obvious gaps among the populations 
i t (Fig. 4} and 


plo no character shows 
discrete variation among the population 


in the 
ns. However, 
three entities can be recognized in this complex. It 
can be seen from the first coordinates (Fig. 4) that 
the d dk DA orm a loose 
roup, WUL, BAI, and ms im a second, JIN, QIN, 
and TIA form a on while the others (including TAL, 
HUO, KON, and MEY) are intermediate. Specimens 
especially from SHM, HUA, ZHU, and OTH are 


oe 


scattered among the above three groups. 


Annals of the 


242 


Missouri Botanical Garden 


'(SMoxre 908) oyeurped. Ayioexo jou ST uoreuoa Jeo] OY} TUI NON "(o2urao1q SuruoerT *nouzut[() 
&yqeoop ed& ou woaz popapo (HA) IZISO uy) “A f Suero 7D “JA vaunquods wea uakaw punhs ^q— “(y *ed&oou) zeggg Aon N 7p "preupog ^w 7D Halow gung "y— gZ omy 


V 


Volume 96, Number 2 Chen et al. 


2009 Revision of the Syringa pubescens Complex 
3.0 r 

3 

c om 295 

o B 

ene 020 ERR 

eo CULTE Tij 
d XI B B H- a i 

mo dU. A a acta pesce ea 
og 

as 05 

a 0.0 1 1 1 1 1 1 1 1 1 1 1 1 J 1 1 1 l 

KOR DAN LIA JIN WUL BAI HEN SHM TAI OTH HUA QIN ZHU HUO TIA KON MEY 
Population 
A 
0.25 rp 
0.20 + | 


KOR DAN LIA JIN WUL BAI HEN SHM TAI OTH HUA QIN ZHU HUO TIA KON MEY 


Population 


w e 
l— 


: l 
“tot feed” 


Ep 


Distance between 
anthers and throat 
(mm) 

INO 


KOR DAN LIA JIN WUL BAT HEN SHM TAI OTH HUA QIN ZHU HUO TIA KON MEY 


Population 


C 


e 3. Box plots depicting morphological variability of corollas in the Syringa pubescens Turcz. complex. — 
i of corolla throats. —B. Shape of corolla tubes (CTS), expressed as [(diameter of corolla throat — diameter of corolla 
base) / length of corolla tube]. —C. Cempa rison of distance between anthers and corolla e oat. Character ranges include 
means and standard deviations. Abbreviations of the 17 populations are explained in Table 


Because the populations SHM and HUA show a altitudes and characters (Fig. 5). It is shown that the 
gradual transition from the third group to the second shape and indumentum of inflorescence rachises and 
in the PCO plot, we chose all the specimens from the length of leaf blades are related to altitudes in 
these regions and depicted various characters and both regions. In the population HUA, plants above 
altitudes on the plot to see the relationship between — 1800 m have inflorescence rachises 4-angled and 


Annals of the 


244 


Missouri Botanical Garden 


OX [x] 


V0 


"g SIXB 0] 9IMQLUOS JeOIY) P[[OJO9 pue SIO u99A^]joq 9QUBISTP 941 pue ON) e[[o109 Jo odes pue iuo JL 


I SIxy 
b U 6 0 T0 0 "0 Lt & O= € U= V'0- 
[ [ | [ | ] | 
* 
e C W 
$ | 
e ~ Das Vo Ke VE ON 
O 
avo Seige O + {i 
Lo 90.9 
é OK D X S. Vv 
BAB---BEoA ewe Xp a thy, 0 
+ XA % : E 1 » L e 
" m "s 
WxxEQx E 
X a x 
|. p zX € 
Xg 
x z d; 


e 


“sooATBO pue sosrqoer oouoosoro[rur uo umquoumpur opn[our Y sxe 07 Sunnqruquoo srojoereu") *xo[duroo suaosagnd nsUuLLAG ou 10] c IsuTeSe | sojeurprooo [edround jo j0[d 19Heog — "p INSTT 


G STXV 


Volume 96, Number 2 Chen et al. 
Revision of the Syringa pubescens Complex 


€ Shape of 
TOS aA inflorescence 
SN rachises 
aA hi 2 à W Indumentum on 
E Lx Rescue iren: --- inflorescence 
g A 1 A A Aa rachises 
A ALength of leaf 

oe to e Ro. blades (om) 


AAAá >: UN o AAA 
1600 1700 1800 1900 2000 2100 
Altitude (m) 


O 2 m 0 o& ARNO 
" 


OShape of 
.AÀ |. pesas inflorescence 
A A rachises 
[| E indumentum on 
inflorescence 
rachises 
ALength of leaf 
blades (cm) 


N wo A oO Q 


of "I 


à g 2 o o $--—$ 
> o 


[| 
Q8" s HN — HE 
1300 1400 1500 1600 1700 1800 1900 2000 
Altitude (m) 


B 


Figure 5. Scatter plots of morphological character against elevational change. —A. Leaf blade length, nape; and the 

pube cen ce of the inflorescence rachis (y axis) are plotted against denial. Duns (x axis) on Mt. Huashan, Shaanxi 

puc China (HUA). —B. Morphological characters as in A (y axis) are plotted against dodo change (x axis) for 
localities in central and southern Shanxi Province in China. 


glabrous and leaf blades often more than 3 cm long, Therefore, three subspecies are recognized in the 
whereas those below 1800 m have inflorescence complex. x us subspecies includes the. popula- 
rachises subterete and hairy and leaf blades com- tions KOR an , which have leaf blades 


monly less than 3 em long (Fig. 5A). In the population usually p^r p. cm long, corolla tubes funnel-shaped 
SHM (here the population was enlarged, including (CTS greater than 0.12), and throats usually 1.6— 
exsiccatae from central and southern Shanxi Prov- 2.5 mm in diameter. These characteristics are in 
ince), plants from 1600 m and up have inflorescence ccordance with those of Chang and Chen’s (1990) 

&ihisen almost glabrous and 4-angled and leaf blades subspecies patula. This subspecies is distributed in 
ide more than 4 em long, while those below the Korean Peninsula and northeastern China (Fig. 6). 
1600 m tend to have inflorescence rachises d and The second includes the populations WUL, BAI, 
ug and lea s ong and the high-altitude specimens from the 
(Fig. 5B). The results indicate n. sis lem higher populations SHM and HUA, which have inflorescence 
altitudes of eastern Shaanxi (HUA) and southern and rachises 4-angled and glabrous and leaves generally 
central Shanxi (SHM) resemble plants from northern 3.5-5.4 em long. This corresponds to subspecies 
China (WUL, BAI, HEN), whereas those from lower pubescens. It occurs in Inner Mongolia, Beijing, 
altitudes resemble plants from western China (KON, Hebei, northern Shanxi, and at higher altitudes in 
TIA, QIN, etc.) central and southern Shanxi and eastern Shaanxi 


Annals of the 
Missouri Botanical Garden 


uo 
Pa 


{130 5 


HE x 


ure 6. Distribution map of the 17 populations ir D iia pubescens complex. 


a subsp. microphylla; filled diamond, subsp. p 


provinces (Fig. 6). The third comprises the remaining 
e hairy, a or 4-angled 
orescence rachises, and usuall m long leaf 
blades. The posa ai a Chang 
Chen's (1990) 


an 
subspecies julianae ni Cha 


ia aed which hav 


ge A ylla an 

s (1990) S. meyeri 
var. spontanea. It grows at bu. de in central 
and southern Shanxi and in Shaanxi, as well as in 
Gansu, eastern Qinghai, Chongqing, western Hubei 
and Henan, Mt. Taishan of Shandong Province, and 
the Jinzhou region of Liaoning Province (Fig. 6). 

Western Hubei is the type locality of Syringa 
julianae. Plants corresponding to descriptions of S. 

julianae are very rare in the wild and were not found 
in our fieldwork. Most specimens from this region 

species e 
ot (Fig. 4) except three sheets of F. 
Wilson 2024. Thus, plants from this region are al 
recognized as subspecies microphylla even though 
some extremes occur. 

The population JIN was recognized as Syringa 
meyeri var. spontanea, by Chang (1990) because its 
venation was close to that of S. meyeri. According to 

our ation in as 

population JIN i The bas 

lateral veins are dm close and do not always run 

parallel to the tip (Fig. 2B). Similar venation patterns 
I 


dre the 
not always palmate. 


extensive 


a 


frequently appear in the populations E , an 
TIA. Furthermore, the population JIN is closely allied 
to the populations TAI, TIA, and QIN in the PCO plot 
(Fig. 4). Therefore the population JIN is here included 
in subspecies microphylla 


Filled circle, subsp. pubescens; filled 


Because Syringa meyeri was described from a 
cultivated plant in Beijing, 
E 


M. C. Chang (PE), and a cultivated plant from Beijing 
Botanical Garden as population MEY for analysis. In 
the PCO plot a 4), the type specimen of S. meyeri 
is closer to S. pubescens subsp. e ubescens (BAI, HEN), 
while the two specimens from PE have closer 
relationships with subspecies microphylla, Because 
no distinct characters can . meyeri from the 
others, we treat S. meyeri as a new synonym of $. 


pubescens subsp. pubescens. 


TAXONOMIC TREATMENT 


Three subspecies are recognized in Syringa pu- 


bescens: subspecies pubescens, subspecies micro- 


phylla, and subspecies patula. 


1. Syringa pubescens Turez., Bull. Soc. Imp. 
Naturalistes Moscou 13: 73. 1840. TYPE: China. 
Hebei Prov.: 1831, P. J. Kirilov s.n. (holotype, 


!, photo 


Leaves 1.2-10 X 0.7-6 cm, glabrous or pubescent, 
lateral veins in 3 to 5 pairs; petioles 0.3-1.5 cm 
i inflorescence rachises 4-angled or 
nt. Calyces glabrous or 
pubescent; corolla tubes cylindrical or funnel-shaped, 
4.5-16 mm, corolla throats 0.8—3 mm wide; anthers 


Panicles lateral; 
terete, glabrous or pubes 


purple or rarely yellow, 0—4 mm below corolla throat. 
Capsules 6-20 X 2-6 mm, verrucose. 


Volume 96, Number 2 
2009 


al. 
Revision of the Syringa pubescens Complex 


Syringa pubescens is distributed on the Korean 
Peninsula and in China. The taxon typically flowers in 
May. Chromosome number 2n = 46, 48. 


KEY To THE THREE SUBSPECIES OF SYRINGA PUBESCENS 


Carall Toat, ]-sha: uh lla throat (1 2-) 
1 3) mm diam; leaf blades (3.7) 
4.4-8.2(-10) em long ........... subsp. patula 


lb. Mun Es cylindrical with corolla throat (0.8—) 
2.8) mm diam.; leaf blades (1.2-)1.5-6(—7) 


= 


cm long. 
2a. Inflorescence rachises glabrous or rarely pubescent, 
obviously 4-angled; leaf blades usually (333.5-5.4 
—6) cm lon subsp. pubescens 
2b. Inflorescence rachises pubescent, faintly 4- need 
or . leaf blades usually (1.2-)1.5-6(-7 
m lon subsp. microphylla 


la. Syringa pubescens subsp. pubescens 


ds m C. K. Schneid., Pl. Wilson. (Sargent) l: nes 
2. Syn. nov. TYPE: China. Beijing: Fengt 
edicion, June 1910, F. N. Meyer 23032 (cme. 


Sion wulingensis Skvortsov & W. Wang, Ill. Fl. Ligneous 
uu E. China: 566. 1955. TYPE: China. Hebei Prov.: 
Wulingshan, 4 Sep. 1952, T. N. Liou 4737 


nes IFP!». 


eaves 3-6 X 1.7-4.2 cm, pope ages 
glabrous or sparsely pubescent abax 

pinnate or somewhat acrodromous. p 
rachises obviously 4-angled, glabrous or rarely 
pubescent; calyces glabrous; corolla tubes cylindrical, 
throats 1-2.8 mm wide; anthers 0.5— 


4 mm below corolla throat. 


7-16.5 mm, 


Distribution and. habitat. 
distributed in Nei Mongol (Inner Mongolia), Beijing, 
Hebei, Shanxi, Shaanxi, H 


China. It usually grows in semi-shaded, moist decid- 


Subspecies pubescens is 
enan, and Shandong in 


uous forests, commonly at high altitudes up to 2400 m. 


scussion. Syringa meyeri was described by 
Schneider (1913) as new, and was stated to differ 
from $. ue in having two pairs of lateral veins 
e leaf tip. Green and Chang (1995) 


pairs of veins 


running t 
further denied S. meyeri with tw 
more or less palmately arranged at ihe base. But our 
careful observation showed that venation in S. meyeri 
was not distinguished from that in S. pubescens subsp. 
pubescens (Fig. 2A). Furthermore, no distinct charac- 
ters differentiate the two taxa. We propose that S. 
meyeri is an extreme cultivated variety in S. 
sm E pubescens because S. meyeri was 
described from a cultivated plant in Beijing and no 
wild plants have been found so far. 


Representative specimens examined. 
Mentougou, Baihuashan, T. F. King 70 ros us yo ors 


ied 810 (PE). Hebei: Chicheng, s. coll 4164 (PE); 

aishui, K. M. Liou 2199 (PE); Laiyuan, K. M. Liou 2367 
n v X. Y. Liu 158 (PE); Wanping, C. G. Yan, 
( 


ngshan, K. M. 
Tio 2094 (PE); Yuxian, Xiaowutaishan, H. W. Kung 51 (PE); 
Zhuolu, Xilingshan, C. G. Yang 847 (PE); Zunhua, Dongling, 
T. Tang 1773 (PE). Henan: Linbao, Puchabiaoben 14261, 
14632 (PE) Lushi, Laojunshan, K. M. Liou 5047 (PE) 
Songxian, Wumasi, Henandui 1505 (PE); Xixia, Laojunshan, 
K. C Kuan & T. L. Dai 1584 (PE). Nei Mongol (Inner 
Mongolia): A. one 1797 (K, P), s.d., P. Arteselaer s.n. (K). 
Shaanxi: Hua uashan, J. Y. Chen 03149, 03157 
y Wax Hsia FO (PE), T. N. Liou 10833 (PE), Z. B. 
Wang 19646 i Chang'an, s B. Z. Guo 1108 
(PE). Shandon, i 
(PE), Zhongdodui 674 (PE). S 
T. P. Wang 2583 (PE), no K. C Ki er 
Chen 310 (PE), ee Yellow pow Exped. 785 (PE 

unyuan, Hengshan n 0321 : 


n, D. Thee 2073 (PE); “Sieg an, Xuehuashan, 
Yellow River Exped. 407 (PE); Yuanqu, ee $. X. 
Bao 2231 (PE). 


lb. Syringa pubescens subsp. microphylla 
e M. A Chang & X. L. Chen, Invest. Stud. 
Nat. 10: 1990. Basionym: 

Pus Dd. Bot. Jahrb. Syst. 2 o: S31. 

TYPE: China. Shaanxi: Tui kio A Lao yu ROM 

Oct. 1896, G. Giraldi 1644 (lectotype, designated 

here, FI). 


inga micro- 


Syringa an var. flavoanthera X. L. Chen, Bull. Bot. 
Res., Harbin 9(3): 41. 1989. Syringa pubescens subsp. 
iierophylla var. flavoanihera (X. L. Chen) M. 
Chang, Fl. Reipubl. Popularis Sin. 61: 68. 1992. TYPE: 
China. Shaanxi: Foping, Longcaoping, 1900 m, s.d., X. 


Syringa je Cine var. e 
Wilson. (Sargent) 1: 301. 1913. TYPE: “Chins. Hubei 
iao-uan-san, 1898, H. Scallan s.n. (holotype, A 
not seen). 
di o pubescens var. tibetica Batalin, Trudy 
eterburgsk. Bot. Sada 13 (18): 378. A "TYPE: 
China. Gansu: Huidui, 7200 ft., 7 May 1885, G. N. 


en) 
neid., Bot. Jahrb. He a 
82): 88. 1905. TYPE: China. Shaan in 
10 Jus 1900, G. Giraldi 7195 (lione. pO 
FI). 
dis to alona C. K. Schneid., Bot. Pe Syst. ies 
Beibl. 82): : : ophy a va 
giraldiana (C. K. Schneid.) S. Z. Qu & X Oe. 
Bull. Bot. Res., Harbin 9(3): 41. 1989. m China. 
anxi: Liu siu shan, Aug. 1899, G. Giraldi 4405 
(olo de: FI). 


Syringa potaninii C. K. Schneid., Repert. Spec. Nov. Regni 
Veg. 9 


a. Gansu: 
Tschi lo ku, 18 June 1885, G. N. Potanin s.n. (holotype, 
LE!, photo PE!). 


Annals of the 
Missouri Botanical Garden 


Syringa julianae C. K. Sehneid., Ill. Handb. Laubholzk. 2: 
LT 


1911. Syn. nov. Syringa pubescens subsp. julianae 
(C. K. Schneid.) M. C. Chang & X. L. Chen, Invest. 
Stud. Nat. 10: 34. 1990. TYPE: China. 


Fangxian, 26 May 1911, E. H. Wilson 1220A iE 
K?). 


Syringa verrucosa C. K. Schneid., Pl. Wilson. (Sargent): 1: 
298. 1912. T China. Hubei: Xingshan Co., Mt. 
Wentsao, 2300 m, 5 June 1907, E. H. Wilson 2579 
(holotype, A not seen; isotype, K!). 

Syringa did. Lingelsh., Pflanzenr. om je due x 

: China. Hubei 
E. 2024 (holotype, K!; jennie, an 

Syringa trichophylla Tang, Bull. Fan. Mem. Inst. Biol. Bot. 
10: 287. 1941. TYPE: China. Skansi: Taigu, Fengshan 
400 m, 11 May 1929, T. Tang s.n. (holotype, A not 


een 
doen meyeri var. spontanea. M. C. Chang, Invest. Stud. 
Nat. 10: 


33. pas Syn. nov. Syringa spontanea (M. 
Chan, gXK Acta Phytotax. Sin 36( 62. 1998 


TYPE: China. T Jinzhou, Heshang shan (Dahei 
shan), 500 m, 12 Sep. 1989, M. C. Chang & X. K. Qin 
12872 (holotype, SHM!, photos K!, PE!). 


Leaves 1.2-7 X 0.74 cm, glabrous or pubescent. 
Inflorescence rachises pubescent, 4-angled or sub- 
terete. Calyces rra or glabrous; corolla tubes 
4.5-14 mm, throats wide; anthers purple or 
rarely vello. 0-3 mm below corolla throat. 


Distribution and bitat. Syringa pubescens 
subsp. microphylla is distributed in Qinghai, Gansu, 
Shaanxi, Chongqing, Hubei, Shanxi, Henan, Shan- 
dong, and Liaoning provinces. It usually occurs in dry, 
open scrub or at the margins of semi-shade forest and 


commonly grows at lower altitudes than subspecies 


Discussion. Two lectotypes were designated here. 
Diels (1901) described Syringa microphylla as new and 
aldi 1644 and 1645, both of 
which were from Tui kio shan of Shaanxi Province. We 
chose G. Giraldi 1644 as the lectotype here. Schneider 
(1905) described S. dielsiana as new and cited two 
i ldi 7193 (with flowers) and 741 (with 
fruit) from Shuang Province. We selected the flowering 


cited two specimens, G. Gir 


im as the lectotype here. 

ringa julianae was described by Schneider 
doi based on a specimen from Hubei Province, 
which was treated as S. pubescens subsp. julianae by 


Chang and Chen (1990). In the PCO plot (Fig. 4), 


most specimens from the region of the type locality 


The characters previously used to delimit 
S. julianae, such as the shape of the corolla tubes and 
anther insertion on corolla tubes, are not discrete in 
the analysis. Thus, S. julianae is newly placed in 
synonymy with S. pubescens subsp. microphylla. 
Syringa meyeri var. spontanea was described by 
Chang and Chen (1990) on the basis of its venation 


being similar to that of S. meyeri. Qin (1998) even 
treated it as a separate species, S. spontanea, based on 
its palmate venation. According to our extensive 
observation, the venation in S. meyeri var. spontanea 
is not always palmate. Both palmate and pinnate 
venations were observed in the type locality (JIN), even 
type specimen (Fig. 2B). Similar venation 
patterns were observed in Mt. lone (SHM) and 
Mt. Tai (TAI). Thus, it is not reliable to establish the 


new taxon based on the venation. Sy 


in the 


ringa meyeri var. 
spontanea resembles S. pubescens subsp. microphylla in 
the size and shape of the leaf blades and in the 
indumentum on the inflorescence rachises and calyces. 
Breeding experiments in Beijing Botanical Garden 
showed that S. meyeri var. spontanea and S. pubescens 
subsp. microphylla are interfertile, indicating their 
close affinity. Therefore, S. meyeri var. spontanea is 
here included as a synonym of S. pubescens subsp. 
microphylla, which is in accordance with the PCO plot 
ig. 


xamined. 


Rep tat CHINA. Chongqing: 
Chengkou, P. E Farges 885 (K). Gansu: Pingliang, 
Kongtongshan, L Y. Chen 03171, 03175 (PE); Tianshui, 


a wore, Fete River Cn 41 87 


. Kuan & 
Yichuan, Pusat 20296, 21100 (PE). Huber Badong, 
A. F Henry Saiwud ang. 
Yunxi, Tianfengshan, J. X. Yang [m d pum 
Jinzhou, M. C. Chang & X. K. Qin 12878 (K, SHM), J. Y. Chen 
05117, 03129 (PE), T. N. Liou 192 (IFP). Qinghai: d E 
P. Wang 041 (ANWP); Minhe, Gushan, B. Z. Guo 
8461 (ANWP); s Huoshizhai, L. e Zhou 2888 (EN. 
Shaanxi: Chang'an, Cuihuashan, Gansu o 
Nanwutai, T. N. mu ; 199, 11050 d Huay à 
Chen 03143, 03144 (PE); Huxian, Laoyu, T Y Ge 03041 
03159, 03167 n b 
- Ni tang, 


Y. la ee 2807, 2847 
ai 1579 (PE), Haopingsi, J. 
X. Yang 206 (PE), J. Y. Chen o 03024 (PE), Jiaolongsi, J. 
X. Yang 281, 285 (PE), Liujiacun, K. T. Fu 82, 2458 (PE 
Xiabaiyun, J. X. Yang 154 (PE). Shandong: Tai'an, Taishan, 
J: hen reat 04209 (PE) Shandong Agriculture 
University 818 (PE 4 (SDNU), Zhong D 
(PE). Shanxi: run Qiliyu, Yellow River Exped. 7 
(PE); Linchuan, K. M. Liou 7667 (PE); Qingyuan, Paoquan, r4 
M. Liou 1591 (PE); Ruicheng, Baiquancun, S. Y. Bao 544 
(PE); Taiyuan, J. Y. Chen 04101, 04102 (PE), EE d 
Yabe s.n. (PE); Xiegan, Xuehuashan, Yellow River Exped. 5 
(PE); Yicheng, X. Y. Liu 20370 (PE); Yongji, T. W. Liu D 


0221 (PE); Yuanqu, Tongshan township, S. Y. Bao 66 

E). 

le. Syringa pubescens subsp. patula (Palib.) 
M. C. Chang & X. L. Chen, Invest. Stud. 


Volume 96, Number 2 
2009 


al. 
Revision of the Syringa pubescens Complex 


dul 10: 34. en vum Ligustrum d 
m Palib., Tru Peterburgsk. Bot. Sa 
um. 156 a a patula (Palib.) Nakai, 
J. Jap. Bot. 14: 638. 1938. TYPE: Korea. Kyong- 
wi, Seoul prope Tap-Tong, 20 May 1895, A. 

Sontag s.n. (holotype, LE not seen) 


Syringa velutina Kom., Trudy Imp. S. RE M a Sada 
18(6) 428. 1901. TYPE: Kor ari-pi 
May 1897, V. L. o s e es LE 


sK; Schneid. Il. Handb. Laubholzk. 2: 


Syringa koehneana C 
0 1912. TYPE: Fig. 627 in Schneider, 


ig. 627. 

1912: 1063. 
a e Nakai, Bot. Mag. Ao 27: 32. 1913. 
Korea. s. loc., Sep. U. Faurie s.n. 

idc type, TI not m 
Syringa micrantha Nakai, Bot. Mag. (Tokyo) 32: 129. 1918. 
: Kor eae 23 June 1917, M. Furumi 65 
(holotype, Tn not se 

Syringa kamibayashii Nakai ai, Bot. Mag. (Tokyo) 32: 130. 
1918. TYPE: Korea. Dohosan, s.d., K. Kamibayashi s.n. 
holotype, TI not seen). 

Syringa cue pe Bot. Mag. (Tokyo) 32: 130. 1918. 
TY a. Ooryongto, 3 m 1917, T. Nakai 4194 
lectotype, designe here, 

Syringa po : A Tiles: Gen. Syringa: 
48. 8, as P debeld eri. Syringa pubescens subsp. 

An n Lilac Soc. 33(4): 123. 

. Sorak National Park, 

1971, National Arboretum 41179 (holotype, US not 

en). 


Leaves 3.7-10 X 1.9-6 cm, glabrous or pubescent. 
Infloresce 
cent. Calyces 


nce rachises 4-angled, glabrous or pubes- 


toothed, glabrous or pubescent; corolla 
tubes funnel-formed, 5-13 mm, throats (1.2-)1.6— 
2.5(-3) mm wide; anthers (0—)0.1—0.9(-1) mm below 


corolla throat 


Distribution and habitat. The distribution of 
ringa pubescens subsp. patula is in the K 
and Med Chi 


ince). It grows in deciduous Das or along forest 


Korean 
Pd a (Liaoning Prov- 


margins. 


Discussion. One lectotype was designated here. 
Nakai (1918) described Syringa venosa as new and 
cited five specimens from Ooryongto. We chose one of 
Nakai's collections with flowers, T. Nakai 4194, as the 
lectotype here. 

Syringa debelderi J. L. Fiala was described from a 
cultivated plant that was stated to have small leaf 
blades (7 X 5 cm) (Fiala, 1988). The name was 
d as S. debelderorum J. L. Fiala by Green 
B The size of leaf blades and other characters of 

orum are scarcely different from 
es uc patula. 

a ri heo examined. CHINA. e 


90 (PE); Dandong, J. py. 4107, 
n 13 co eee J. Y. Chen 03248 (PE); dae. le 


C. Chu 576 (PE); Xiuyan, W. Wang 1556, 1576 (PE). 

KOREA. s. loc.: V. L. Komarov 1259 (K), J. Sato 4147 (PE), 
J. Sato 4150 (PE), E. H. aaa 11705 (PE); Doekyu-san, 
Hagman 329 (UPS); Kongo-san, E. H. Wilson 10491 (K); M 
Chiisan, J. Ohwi 6811 (UPS); Shuotsu, J. Ohwi 380 T 
Sorak-san, Elsik & Hey 915-77 (K); Chonranam-do, H. T. 
22226 » Dagelet island, E. H. Wilson 8527 (K) Keiki, 
Poukhan-san, E. H. Wilson 10741 (K); Ooryonto, T. Nakai 
4194 (TI). 


Literature Cited 


Chang, M. 2. Syringa. Pp. 50-84 in Flora Rei- 
la ae Sinicae, Vol. 61. Science Press, 
dis m 


—. L. Chen. 1990. je on Chinese Syringa 1. 
Invest. ET Nat. 10: 32-4 
. Green. ibd Syringa. Pp. 280-286 in 
Li Ve p & P. H. Raven (editors), Flora of China, 
Vol. 15: Myrsinaceae through Loganiaceae. Science 
Press, Beijing, and Missouri Botanical Garden Press, St. 
Louis. 
Chen, X. L., X. Y. Zhao & S. Z. Qu. 1989. New d for 
genus [o L. Bull. Bot. Res., Harbin 9(3): 39-4 
Di te L. 1901. Die Flora von Central-China. Bot. ao 
19: 531-532. 
Fiala, 1 L. 1988. Lilacs: The Genus Syringa. Timber Press, 
Portland. 
Green, P. S. 1989. e d The Genus Syringa [book review]. 
Kew Mag. 6(2): 9 
——— C. TN de taxonomic changes in 
(Oleaceae), pais a revision of series 
Pubescentes: Novon 5: 329-333. 
Komarov, V. L. 1901. Species novae Florae Asiae Orientalis 
(Manshuriae et Koreae borealis). Trudy Imp. S.-Peter- 
8: 428. 


Lingelsheim, A. 1920. Syring . 75-9. ngler 
(editor, Das Blanzscidh, E 721V Po pd 
Engelmann, Leipzig. 

McKelvey, S. D. 1928. The Lilac: 
Macmillan Company, New 

Nakai, T. 1913. Notu D ad Plantas Japoniae et Coreae. Bot. 
Mag. o 27: 32-33. 

8. Flora T Koreana 10. Bot. Mag. (Tokyo) 


A Monograph. The 


Tu 
Bine I. 1900. Conspectus Florae Koreae, Part 2. Trudy 
Imp. S.-Peterburgsk. Bot. Sada 18: 156 
K. 1998 inga meyeri Schneid. and its 
confused Cra Acta Phytotax. Si a 359-364. 
905. L. Diels, Beiträge 


von Central-China. Bot. Jahrb. Syst. 36(5, Beibl. 82): 
86-89. 


0. Species et formae novae generis Syringa. 
Regen. e Nov. Regni Veg. 9: 79-82. 
. Illustriertes Handbuch der Laubholzkunde 
— (Band m m —185. Gustav Fisher, Jena. 
———. Xs d Handbuch der Laubholzkunde 
Band E p Gus 
———. : Phan. in Plantae Wilson- 
lanae (Sargent), Vol. 1. Cambridge University Press, 
Cambridge 
Eoo B. V. & W. Wang. 1955. Oleaceae. a 471-567 
. N. Liou ae, Ill. Fl. Ligneous Pl. . China. 
e Press, Bei 


Annals of the 
Missouri Botanical Garden 


es T. 1941. A new species of Syringa in Shanxi. Bull. Fan 


m. Inst. Biol. Bot. 10: 288. 
Turczaninow, N 1840. Decades Quatuor Plantarum Hucu: 
que Descriptarum D Bull. Soc. Imp. ausis 
Moscou 13: 73. 


APPENDIX 1. The coding of morphological characters used in 
the analysis of the Syringa pubescens complex. 


ntinuous characters: 1. um of leaf blades (cm). 2. 
Width bs a blades (em). 3. Ratio of a width of leaf 
blades mber of lateral vein pairs. 5. Length of petioles 
(cm). "x pm gth of inflorescences d T. Diameter of 
inflorescences (cm). 8. Number of flowers per werd 
9. Length of calyces (mm). 10. Diameter RU es (mm). 11. 
Length of corolla tubes (mm). i m ameter of corolla throats 

3. Shape of corolla tul = (diameter of corolla 
thr iar m eter of corolla E E me of cor EE tube]. 14. 
Length of corolla lobes (mm). 15. Width of corolla lobes 
(mm). 16. Anther position at corolla tube from base (mm). 17. 


(mm). 1 


Distance between anthers and corolla throat (mm). 18. Pisul 
Es (mm 
inary charae ters: Adaxial leaf surface flat (0); 
concave (1). 2. Abaxial leaf surface flat (0); convex (1). 3. 
Calyx lobes truncate en E D. 
Ordinal char l. In 


= 


m coverage on adax 
leaf surface 0 (0: EE 0 Ed o 25-15% 
e) 75%-100% (4). 2 

eaf surface 0 (0); 0%. f "D. "ro 25 Ex 25% A 
à; 15%- "i. Me 3. n e Lone (1); connate (2); 
crodromous (3). um coverage on 
peti ioles 0 (0); (us s Pr "Oh 25%-15% (3) 
75%-—100% (4). 5. Shape of inflorescence rachises terete (1); 


ndumentum coverage on 
calyces 0 (0); 0%-5% (1); 595—256 (2); 25%-75% (3); 
75%-—100% (4). 


A SYNOPSIS OF SOUTH 
AMERICAN ECHEANDIA 
(ANTHERICACEAE)! 


Robert William Cruden? 


ABSTRACT 


Eight of the 81 recognized species in Echeandia Ortega (Anthericaceae) oceur in South America. Four species occur in 
fi 


Venezuela ee cono one in rend and three in Peru. 
America. f th subge: 


enus avead 


> 


South and Conte America. As many 
ering wa amarca, Peru, r: 


Venezuela, which was heretofore included in E 
theric 


(Poelln.} Cruden, are made, and An 


as five of the eight species are narrow endemics and four may 
e RA of e isotypes E E. ciliata rM Cruden with material uo Co M Venezuel 
ype gath made in Caj ather than uel 


in subgenus Echeandia are end 


a, and Peru prov 


e 
ela, as indicated in Kunth's 


ac 
. ciliata. Two new combinations, E. e (Killip) Cruden and E. weberbaueri 
m glareosum Ravenna = E. le 


oris Ravenna annii 


(= E. ciliata) an 


(Baker) Marais & Reilly) are newly ud A neotype for E. leucantha eee ll a lectotype for E. ciliata are 


designated. Anthericum peruvianum Willd. e 
Key words: Anthericaceae, 


x Kunth is an illegitimate name. 
Cajamarca, Echeandia, endemism, South America. 


ith 81 recognized species (Cruden, 1999 and 
references therein; this paper), Echeandia Ortega, 
which occurs from the southwestern United States to 
southern Peru, is the largest New World genus in the 
Anthericaceae. Fifty-nine species occur in Mexico, 
and the distributional ranges of three of those species 
C l and two into the 


Thirteen species are 


extend into Central America 
southwestern United States. 
endemic to Central America, and two species occur 
in Central America and northwestern South American. 


As presently construed, Echeandia includes many 
of the New World species originally placed in 
Anthericum L. (Cruden & McVaugh, 1989; Cruden, 
1994, 1999). Previous workers separated the New 
World taxa on the basis of connate (Echeandia) versus 


free anthers (Anthericum) (e.g., Baker, 1876; Green- 
man, 1898; Weatherby, 1910; Hutchinson, 1959; 
Ravenna, 1988). The incorporation of New World 
species with free anthers into Echeandia was based on 
unique traits that they shared, e.g., scaled (ie., 

squamate) filaments and polycarpic rhizomes. All But 
two or three of the species in Echeandia can be placed 
easily in one of two subgenera based on morphological 
and physiological traits (Cruden, 1999). In both 
subgenera, some species have free anthers while 
others ha onnat thers. For example, in 
subgenus Echeandia, E. ciliata (Kunth) Cruden has 
free anthers and E. lehmannii (Baker) Marais € Reilly 
has connate anthers; in subgenus Mscavea Cruden, E. 
bolivarensis Cruden has free anthers and E. leucantha 
Klotzsch has connate anthers. In essence, the use of 
this key character confused rather than clarified 


I thank the curators of the following arte oe ore me to examine material of Echeandia: SAU: B BH, BM, BR, 


Y, QCA, P; SMU, U, UC, US, VEN. I 


greatly 


vided additional si regarding the type gathering of E. ciliata. Likewi 
iability of the labels and the : 


1 by Madame Gan. (P). W. Cn euter m P ei and 
L. Constance 
authors of the anno e tions. Ia 


. Ma ui thanks to the 
F. Rodríguez a . R. Valencia Reyes Js. 

Heinrichs (GOET), P. Wilkin (&). and B. Wallnófer (W) for looking for type. acit of E. leucantha, and to P. Wilkin (K) for 
providing information on the isotype of Anthericum glareosum. The ma 
and observations of V. owe (MO), R. Gereau (MO), D 
prepare ed the map and pro 


during iy visits t 
Baker, Lyda and Nathalie Cruden, Diana Gannett, Diana Horton, Virginia and David Lyon, and ‘Van’ Vandemark during the 
ie ne of this study. 

partment of Biology, University of Iowa, Iowa City, Iowa 52242, U.S.A. robert-cruden@uiowa.edu. 
hs “o. 3417/2002129 


ANN. Missouni Bor. Garp. 96: 251-267. PUBLISHED ON 7 JULY 2009. 


Annals of the 
Missouri Botanical Garden 


relationships within Echeandia and between Echean- 

ia and related genera. The latter include the South 
American genera Diamena Ravenna, Diora Ravenna, 
Hagenbachia Nees € Mart., 
genus (Cruden, in prep.), as well as the Old World 


and an unpublished 


genera Anthericum and Chlorophytum Ker Gawl. 
cheandia and its New World relatives are 

separated by both floral and vegetative traits. T 
floral traits (yellow flowers, connate anthers, and/or 
scaled filaments) o all but four species of 
Echeandia from their merican relatives 
(Cruden, 1999; se 
1998). The four A have white flowers, free 
anthers, and smooth, terete filaments (Cruden, 1999). 
All Echeandia have a perennial, unsegmented, erect 


also Jen 1987; Conran, 


rhizome that gives rise to new roots, a basal rosette of 
leaves, and a flowering scape each year. Also, their 
flowers lack nectaries. In contrast, flowers of the 
related genera in South America are almost always 
white, very rarely yellowish or possibly bluish white, 
have free anthers, and have scaleless filaments. The 
these 
terminal segment produces a set of roots that appears 


rhizomes of enera are segmented and the 
to function through two, three, or more subsequent 


ing: T basal leaves, and a up 


T 
oO 


c. owers of Diora, Diamena 
Met genus have septal estalle (Cruden, 
pers. obs.; Ravenna, 1987, 1988). In addition, Diora is 
distinguished by red pollen and scabrescent capsules, 
and produces flowering scapes and basal leaves at 
different times of the year; Diamena has long, tubular 
corollas (Ravenna, 1987); Hagenbachia has small 
flowers, few ovules, small, globose capsules, and roots 
with no storage areas (Cruden, ; and the 
filaments of the unpublished genus are expanded 
above the middle (Cruden, pers. obs.). 


FLoRAL TRAITS AND POLLINATION RELATIONSHIPS 


There are substantial differences in floral traits 


among the South American species that reflect 
adaptation to different pollinators, different pollinator 
ehaviors, and/or pollination environments. The 


flowers of Echeandia lehmannii, E. leucantha, and 

E. pittieri Cruden are nutant, and their anthers taper 
pically and are united laterally, thus forming a 
arrow cone that is open T 

vibrated by bees to extraet the pollen (Bernhardt & 

Montalvo, 1979; Conran, 1998; Cruden, pers. obs.). 


e anthers o 


at its apex. e cone is 


olivarensis are nonversatile and 


shed pollen through apical openings that result from 
the spreading of the anther walls at the top of the line 
of dehiscence (Cruden, pers. obs.). Each filament is 
inserted in a deep pit, which holds the anther on the 
same axis as the filament. Such anthers in Mexican 


species, either singly or together, are vibrated by bees 
to obtain the pollen (Cruden, pers. obs.). The stamens 
bolivarensis are similar to those o 

campechiana Cruden, and flowers of the latter are 
nutant. Likewise, the anthers of E. weberbaueri 
(Poelln.) Cruden are nonversatile, but are held in 
line with the axes of the filaments by the strongly 
reflexed walls of the anther sacs and dehiscence is 
lateral. The flowers of E. ciliata and E. weberbaueri 
ave straight to somewhat arcuate or declinate styles 
that extend (223-6 mm h h 


usually turn upward just below the stigmas. This 


beyond the anthers an 


characterizes flowers that are cernuous or patent. The 
styles of E. denticulata Cruden are deflexed and 
usually bend forward below and parallel to the 


anthers, and the stigmas are exserted below and (1-) 


No 


—4(—5) mm in front of the anthers. The flowers are 
ee a possibly ce e flowers 
of Mex species with similar i. (Cruden & 
McVaugh, 100 194, 195). In contrast, the flowers of 


E. herrerae (Killip) Cruden appear to be erect or 


rnuous, as a 


nearly so, and their geniculate styles pass between the 
filaments. The stigmas are exserted up to 2(-2.5) mm 
lateral to the stamens or, occasionally, the styles are 
straight and the stigmas are surrounded by or barely 
xceed the anthers. A similar relationship between 
stamens and styles occurs in Nemastylis Nutt. 
(Iridaceae) and Xyris juncea R. Br. (Xyridaceae). 
Erect flowers are frequently associated with plants 
with short inflorescences that live in open and/or 
exposed sites. Finally, the failure of most flowers on 
most plants to produce capsules suggests the flowers 
of all the species are cross-pollinated and the plants 
outbre 


ENDEMISM 


Five of the eight South American Echeandia 
(Appendix 1) are known from two to four collections 
and/or have limited distributional ranges and three, E. 
leucantha, are 


denticulata, E. herrerae, and 


i up and relatively common (Fig. 1). Echean- 
di 


ia ciliata ndemie to southern Cajamarca, Peru, 
which is within the Amotape-Huancabamba zone, a 
region noted for its endemic species (e.g. Berry, 
982; Mist 1999; Sagástegui et al, ; 
e 02). This species has been collected 
ca. a km southwest of Cajamarca, i.e., ca. 
umbe Mayo, to 43 km east of the city and from a 
little southwest of Bambamarca southeast to Caja- 
bamba, a distance of ca. 120 km (Fig. 1). The three 
collections of E. lehmannii with reasonable collection 
data were made 20-30 km north and northeast o 
Quito (Ecuador) at sites ca. 30 km apart. These sites 
are within or close to the Pululahua Crater and the 


Volume 96, Number 2 Cruden 
2009 Synopsis of Echeandia 


Venezuela 


Colombia 


++ E. bolivarensis 
m E.ciliata 

€ E.denticulata 
9 E.herrerae 

* E. lehmannii 
à E. leucantha 
@ E. pittieri 

+ E. weberbaueri 


Les 


63° 


Figure 1. Distributional ranges of the South American species of Echeandia. 


nearby Volcán Mojanda-Fuya Fuya. The absence of | Venezuela, were made within a few kilometers of each 
specimens from Volcán Pichincha and similar, nearby other on the Serranía de Los Pijiguaos near Los 
sites suggests this species has a restricted range. The  Pijiguaos in the western part of the state of Bolivar. 
two collections of E. bolivarensis, which is endemic to The proximity of the two localities suggests a limited 


Annals of the 
Missouri Botanical Garden 


distributional range, but the collections probably 
reflect, at least in part, the development of a bauxite 
mine on the serranía, which provided easy access to 
the area. Echeandia bolivarensis might well be found 
elsewhere in the region. Likewise, E. weberbaueri may 
be more widely distributed than the two collections 
from Peru suggest. These were made 5-10 km apart in 
or close to the valley of the Río Mantaro ca. 60 km 
east of Huancay: 


o. eandia pittieri is known from a 
single locality in Colombia (Fig. 


1) and two in Panama 
(Cruden, 1986b). This suggests a rare species with a 
relatively large distributional range. 


TAXONOMIC TREATMENT 


I. E ow s Nov. Pl. Descr. Dec., 135, t. 18. 
1800 E: Echeandia reflexa (Ca) Rose, 
Conii n Natl. Herb. 10: 93. 1906 [— 


Anthericum reflexum Cav.]. 


Perennial herbs from short, erect, unsegmented 
polycarpic rhizomes associated with annual roots with 
distinct and obvious a areas, these un 
either close to the rhizome or 
distance; the roots, vun the storage areas, 
usually covered with root hairs. Basal leaves bifacial, 
very narrowly linear, oblong to elliptic, the bases 
usually surrounded by a fibrous collar composed of 
previous years' leaf bases. 
(sometimes more) cauline leaves below th 
b 


ranch or flower-bearing node, these reduced in size 


KEY TO THE SOUTH AMERICAN SPECIES OF ECHEANDIA 


acropetally and subtend the branches and flower- 
bearing nodes. Inflorescences racemose or paniculate 
with (1 or)2 to 4(or 5) flowers at a node, 
subtended 
orange, or white, very rarely pale yellow or cream, + 


each flower 
by a bracteole. Flowers yellow to yellow- 
erect, tepals and 


cernuous to patent, or nutant; 


stamens originate independently on t 
to broadly elliptic, 
spreading, 3-, rarely 5-veined, these loosely enclose 


e receptacle; 
tepals narrowly ed to 


developing capsules and wither prior to dehiscence; 
filaments free, + cylindrical to clavate, smooth, 
wrinkled and scaleless, or bearing transverse scales, 
insertion in a shallow pit (anthers versatile), deep pit 
or pocket (anthers nonversatile), or in a deep pocket 
(anthers connate); anthers yellow, dorsifixed near 
the base, rarely basifixed, free or connate laterally, if 
free + versatile and dehiscing laterally or nonversa- 
tile and either dehiscing apically through apical slits 
or dehiscing laterally, if connate the anthers Mas 
the frustrum of a cone (hereafter cone) whose shape 
varies from + cylindrical or barrel-shaped s 
parallel-sided) with a broadly lobed apex to narrowly 
conical (anthers tapered from base to apex) with a 
minutely lobed apex; ovary oblong, superior; ovules 8 
or more per carpel Fruit a loculicidal capsule, 
broadly to narrowly oblong, rarely + globose, 
shallowly 3-lobed; seeds irregularly compressed and 
Chromosome 


48, 64, 80, ca. 84 (see Cruden, 


olded; seed coat black, colliculose. 
numbers 


1994, 1999). 


The key was constructed using the specimens listed in Appendix 2, and most of the species were represented 
by fewer than 10 collections. If those specimens were a biased sample of the variation in a species, the key may 
not work. Also, because of the considerable overlap of variation in both floral and vegetative traits among 
species, a few specimens may not key out. However, because most of the species are allopatric, most specimens 


can be identified using geography (Fig. 1). The key includes the species in both subgenera. 


la. Anthers joined laterally, forming a cone; flowers mostly nutant. 
2a. Flowers yellow; basal leaves 8-20 cm long; scape 19-32(—40) cm high; 2800-2850 m, Ecuador ........ 
juan ibaa eae dae eae a al da ds EL ds DA E. lehmannii 
2b. n white; most basal leaves 27-83 cm long; scape (4065-115 cm high; 190-1500 m, Colombia and 
uela. 
i dien with transverse scales; storage roots enlarged 3-6 cm from the rhizome; northern Colombia and 
nortawester Venezuela: ¿aia dd ti a dire o RD s s ice ime TE. pucr 
3b. Filaments smooth; Hr roots enlarged 1-2 cm from the rhizome; western Colombia ....... 8. E. pitt 
lb. Anthers free; flowers erect, cernuous, patent, or possibly nutant. 


4a. Scape 98-118 cm high; cauline leaves 4 or 5, 
3 i 3-5 


the lowest 10-19 cm long; storage areas of roots xr (1 
the rhizome, most 3— la 


=) 

cm long; below 700 m, Venezuela E. bolivarensis 
4b. Scape to 70 cm high, rarely higher; cauline leaves O to 2, rarely 3, lowest 0.6—3.5(—5.1) cm long; o areas 
E roots enlarged 1-2 cm from the rhizome, 1-3 cm long, rarely longer; above 1000 m, Venezuela, Colombia, 


d EU erect; styles 2-6(-6.5) mm long, geniculate, occasionally straight and the stigmas equal to or 
[uud ME. s anthers; pedicels of flowers 2.5-6(—7) mm long; 2800-3824 m, Peru, Junín south to 
ANG UZ MR "EE 3. E. herrerae 
5b. Sae a to cernuous; styles (5.5—)6-11 mm long, straight to somewhat deflexed, exserted 2-6 m 
beyond the anthers; pedicels of flowers (4-)5-15 mm long. 


Volume 96, Number 2 Cruden 255 
2009 Synopsis of Echeandia 
6a. Filaments sal leas tire to denticulat Era 
short- Mea or ane ciliate; o: p —3500 m, a Colombia and Venezuela ...... . denticulata 
6b. Filaments smooth, wrinkled a pide or occasionally bearing small, narrow, transverse seules: 
margins » e basal ipe a e to long-ciliate, rarely UE Peru. 
Ta. Tepals usually twice or more than twice the length of the stamens; anthers 1.5-2.5 mm long, 
ecoming twisted or strongly curve ring or after anthesis; capsules 10-12 mm long; most 
basal leaves 2-7 mm wide, p falcate; 2500-3400 m, Cajamarca ............ 1. E. ciliata 
Tb. Tepals usually less than twice the length of the stamens; anthers (2—)2.5—4 mm long, most 
remaining straight Pe ess capsu. ules 12-13.5 mm long; most basal leaves 6-12 mm wide, 


THE SUBGENERA 


The two subgenera of Echeandia, both of which are 
represented in South America, are distinguished by 
the shape of their inner tepals, time of flower opening, 
and whether the flowers are primarily yellow or white 
and the anthers free or connate. The South American 
My are morphologically similar to their Mexican 
and upon which this 

1999). Species 
in subgenus Echeandia have elliptical to broadly 


ntral American relatives, 
ee is predicated (see Cruden, 


elliptical inner tepals that are usually equal to or 
greater than 4.5 mm wide. Forty of the 55 species 
have yellow to orange flowers, 10 have white flowers, 
and five species have populations that are either 
yellow-flowered or white-flowered. Approximately two 
thirds of the species in subgenus Echeandia have free 
anthers, and the anthers of the remainder are connate 
laterally. In most of the species with free anthers, the 
anthers dehisce laterally. In contrast, 24 of 26 species 
in subgenus Mscavea have narrowly elliptical tepals 
(maximum width, 4.5 mm), and 21 of the 26 species 
have white flowers, two have cream-colored flowers, 
two have orange flowers, and one has populations with 
either white or orange flowers. Twenty-one of the 26 
species have connate anthers and the other five have 
free anthers that dehisce apically. Further, based on 
the observation of ca. 50 Mexican and Guatemalan 
species in the field and/or greenhouse, the flowers of 
species in subgenus Echeandia open early in the 

orning and close by mid- or late afternoon whereas 
those in subgenus Mscavea open from late in the 
morning to early afternoon and EK in late afternoon 
or early evening. Also, species in subgenus Echeandia 
tend to occur at higher elevations in relatively mesic 
habitats with warm- to cold-temperate climates 
compared to species in subgenus Mscavea. The latter 
in drier habitats with 
With the 


See ui 


occur at lower elevations 
subtropical to warm-temperate climates. 
exception of a few decaploid (2n = 
cae in subgenus Echeandia, equivalent chro 
ome numbers occur in the two subgenera (Cruden, 
1994, 1999). In both subgenera, diploid species 
predominate in Mexico north of the Isthmus of 
Tehuantepec and polyploid species are more common 


to the south of the isthmus, i.e., in Central America 


A tM. 5. E. weha bauc 
(Cruden, 1994). No documented chromosome numbers 
are known for the South American species. Finally, 
illustrations and/or E M of species in the two 
m we are rdt and Montalvo 

1979: 68), uc. e  MeVaus " (1989: 187, 194, 

195), and Conran (1999: 116). 


The five South American species in subgenus 


a 


Echeandia are similar to their Mexican and Central 
American relatives, but they exhibit a relatively small 
subset of the variation exhibited by their northern 
relatives (Cruden € McVaugh, 1989; Cruden, 1994). 
Those five species, all of which are endemic to South 
America, have yellow flowers with elliptical to broadly 
elliptical inner tepals, four have free anthers, and all 
occur above 2000 m (but see E. denticulata). The 
storage areas of the roots develop 0.5-2 em from the 
rhizome, the scapes are mostly 15-50 cm high, and 
the main axes of the inflorescences support six to 15 
flower-bearing nodes. If the filaments bear transverse 
scales, they are quite narrow to relatively narrow 

I suggest the species in subgenus Echeandia might 
have had a com 
morphologically and the traits that distinguish them 
The latter include the 
relative numbers of plants with dentate versus ciliate 


mon ancestor because they are similar 
are primarily qualitative. 


leaf margins, scaled versus scaleless filaments, the 
length of the stamens vis-à-vis the length of the tepals, 
The 


unique traits that distinguish E. herrerae, e.g., its 


and/or glabrous versus scabrescent scapes. 


shorter stature, erect flowers, and geniculate styles, 
may reflect its growing in open, exposed habitats. The 
major difference between E. lehmannii and the other 


essence, 
somewhat different combinations of traits that en 
have characterized a com 
South Ameri 


only Central Ámerican species ae free anthers that 


mon ancestor. Finally, t 
rican species are similar to E. skinneri, ne 
is found south of Guatemala. There is no Central 
American species that is an obvious progenitor of E. 
lehmannii 

Only one of the three species in subgenus Mscavea, 
Echeandia bolivarensis, is endemic to South America, 
and it is just one of three species in the subgenus with 
yellow flowers. One of these, E. campechiana, which is 
endemic to the Yucatán Peninsula in Mexico, is also a 


Annals of the 
Missouri Botanical Garden 


robust species with tall scapes (0.9—1.5 m high), 
numerous branches, and smaller flowers (tepals 8— 

mm long) with anthers that dehisce apically 
"m 1994). T 


Mscavea have white flowers with narrowly elliptical 


he other two species in subgenus 


inner tepals and connate anthers. These species, E. 
leucantha and E. pittieri, occur in both northwestern 
South American and Central America below 1500 m. 


Ia. Echeandia subgen. Echeandia 


, Flowers yellow, occasionally white; inner tepals el- 
d y 1H 


ptic; fl ] mornin; g 
and closing in early to mid- aler don (Cruden, 1999). 


1. poc ciliata (Kunth) Cruden, Phytologia 59: 
380. 1986. Basionym: Phalangium ciliatum 
Kunth, pes Gen. Sp. [HBK] (quarto ed.) 1: 276, 
t. 676. 1815 [1816]. Anthericum ciliatum (Kunth) 


ciliatum (Kunth) Schult. & Schult. f., Syst. Veg. 
7(1): 466. 1829. Anthericum o Hemsl., 
Biol. Cent.-Amer., Bot. 3: 374 


4, replacement 


name for A. Ee ae Eu chult. f. 
Anthericum a ii em. Torrey Bot. 
Club 6 6, nom. illeg, replacement name 


for A. e (Kunth) Spreng., pro syn. A. 
humboldtii. TYPE: [Peru.] “Crescit prope Car- 


. (lectotype, 
P-Bonpl. not seen, photo!; 
isotypes, P!, B-W 6657/1 not seen, photo!). 


a glareosum Ravenna, Onira 1: 29. d Syn. nov. 
YPE: Peru. Cajamarca: Cajam 1 Cumbé, 

2000 m, 17 Apr. 1958, A. pe 1307 (holotype, Hb. 

UT 2622 not seen, K not 


seen, US!). 
theresa peruvianum Willd. ex Kunth, Index Kew. 1: 146. 
1895, nom. illeg. Based on Anthericum peruvianum 
4: 596, 1843, nom. nud. 


Enum. Pl. (see 


odio 

Storage areas of roots enlarged 0.5-1 cm from the 
rhizome, 1-2(-2.5) em long. Basal leaves 5 to 10(to 
14), 6-15(-19) em X (1-)2-7(-9) mm, falcate, rarely 
flat, ciliate to long-ciliate, occasionally short-ciliate to 
ciliate, rarely densely so; cauline leaves O or 1(or 2), if 
, (8321-50 em 
rely a few centimeters higher, height (1.2—) 
2.5—4 times the length of the longest basal leaf, 
glabrous or nearly so to minutely scabreseent toward 


present, the lowest 8-21 mm. Scape 


l6 flower-bearing n 
lowest or 2-flowered, upper l-flowered; flowers 
pedicels (4—5- 


yellow, most cernuous to patent; 


9 mm; tepals (12—)13-16.5(-18.5) mm, usually twice 
or more than twice the length of the stamens, probably 
spreading to somewhat reflexed, inner tepals 6 mm or 
more wide; filaments 5-7 mm, + straight, smooth or 
wrinkled and scaleless, occasionally bearing a few 
tiny, quite narrow to numerous small, narrowly 
transverse scales, inserted in a shallow pit; anthers 
free, versatile, 1.5-2.5(-3.5) mm, or longer if not 


Ww 


curled and/or twisted, dehiscence lateral; ovary 2— 
(-3.5) mm; styles 6-10 mm, straight to gently arcuate, 
turned upward below the stigma; the stigma exserted 
(2-)3-6 mm beyond the anthers. Capsules oblong, 
10-12 X 4.5-5 mm. 


Distribution and habitat. This slender-stemmed 
species is endemic to southern Cajamarca, Peru, 
between 2500 and 3400 m elevation, from south of 
Bambamarca southeast to near Cajabamba and from 
15 km southwest of Cajamarca (Cumbe Mayo) to ca. 
43 km east of Cajamarca (Fig. 1). It is usually a plant 
of dry and/or rocky places, including over-grazed 
hillsides where it grows among shrubs and other 
herbs. It was also found in disturbed roadsides and 
edges of cultivated fields. 


Phenology. In flower January to May. 


Discussion. A combination of floral and vegetative 
traits distinguishes Echeandia ciliata and separates it 
from other South American species in subgenus 
Echeandia. It is distinguished by the narrow, falcate, 
basal leaves whose margins are ciliate, slender scapes 
([8—]21—50 cm high [rarely a few centimeters higher]) 
that are frequently three to four times the length of the 
longest basal leaf, relatively large flowers (tepals [12—] 
13-16.5[-18.5] mm long), and tepals that are usually 
twice or more than twice as long as the stamens. 
cheandia ciliata is most likely to be confused with 
E. denticulata. The filaments of most specimens of E. 
ciliata were scaleless and the leaf margins ciliate, 
whereas the filaments of all the specimens of E 
denticulata were scaled and most (39 of 49 specimens) 
ad entire to denticulate leaf margins. However, 
occasional specimens of both E. ciliata (e.g., Dillon et 
al. 2862 and Sagástegui 15148) and E. denticulata 
may have scaled filaments and ciliate leaf margins. In 
addition, the capsules of E. ciliata were smaller than 
those of E. denticulata (10-12 mm vs. [9-]11-17 mm 
long). 

Most specimens of Echeandia ciliata were easily 
distinguished from those of E. herrerae and E. 
weberbaueri. In general, the tepals of E. ciliata were 
twice or more than twice the length of the stamens, 

whereas those of E. herrerae and E. weberbaueri were 
usually less than twice the length of their stamens. 
Further, the flowers of E. herrerae were generally 


Volume 96, Number 2 


Cruden 
Synopsis of Echeandia 


smaller (tepals 9-13[-15] mm vs. [12-]13-16.5 
[-18.5] mm long) with shorter pedicels (2.5-6[-7 
. [4-]5-9 mm long), shorter geniculate styles 
(2-6[-6.5] mm vs. 6-10 mm long) and generally 
shorter (4—26[-ca. 45] cm vs. [8-]21-50 em high, 


rarely a few centimeters higher) and noticeably 


= 


scabrescent to scabrous scapes that were usually less 
than twice the length of the longest basal leaf. The 
scapes of E. ciliata were usually three to four times 
the length of the longest basal leaf and smooth to 
minutely scabrescent toward the base. Compared to 
those of E. ciliata had 
somewhat shorter scapes ([8-]21-50 em vs. 3 

[-68] em high), falcate basal leaves that were shorter 
(6-15[-19] em vs. [6.5—]10-25[-33] em) and narrower 
([1-]2—7[-9] mm vs. [3-]6—12 mm), and, if present, 
shorter cauline leaves (8-21 mm vs. 16-35[-49 
long). Also, the anthers of most specimens of E. 
1.5- 


2.5 mm), usually nonversatile, and remained straight 


plants of E. weberbaueri, 


| mm 
weberbaueri were longer ([2-]2.5-4 mm vs. 


to somewhat curved during and after dehiscence, 
whereas a of E. ciliata were versatile and twisted 

hiscence. Finally, E. ciliata is 
both E. herrerae and E. 


weberbaueri (Fig. 1). The former occurs on the Pacific 


during r de 
RN dicunt from 


slopes of the Andes in southern Cajamarca, and the 
latter two occur 500 km or more to the south on the 
Atlantie slopes of the Andes from Junín south to 
Apurimac and Cuzco. 


Nomenclature. Here I compare the type gathering 


of dri ciliatum (= 
(Humboldt & Bonpland s.n.) with specimens from 
D MEN and Peru and show that the 
epithet ciliata should be applied to plants from Peru. 


Echeandia ciliata) 
Venezuela, 


The sheet bearing the lectotype of Phalangium 
ciliatum bears two labels, both in the lower left-hand 

corner. One is a printed label: *HERB. MUS. PARIS 
Herbier Humboldt & Bonpland. AMÉRIQUE ÉQUA- 
TORIALE," and the second bears an annotation by 
Kunth (L. Constance, pers. comm Greuter, pers. 
comm.): *L" *Phalangium ciliatum." The sheet bears 
right, which is the 
and parts of a second on the left. I 


a complete specimen on the 
lectotype, 
designated the complete specimen as the lectotype 
because it is nearly identical to the illustration (tab. 
676) 
undoubtedly served as the model for the illustration. 


that accompanied Kunth’s protologue an 


The isotypes bear Bonpland’s collection labels, which 
are virtually identical. The sheet at P bears two 
specimens, the scape of a third, and three labels. The 
label in the lower left-hand corner was annotated by 
Bonpland (M. Cusset, pers. comm.; also see Rankin 
Rodríguez & PNE 2001: fig. P “VIL “Om ithoga- 


lum?” in the upper left corner and “Perou” in the lower 


right corner. The label in the lower right-hand corner of 
the sheet has printed above: *HERB. MUS. PARIS" 
followed by an annotation (possibly by an older Kunth 
[W. Greuter, pers. comm.]: “Phalangium ciliatum HBK. 
N. Gen. I. 276.”; in a second hand: “Anthericum R. Sch."; 
below that, possibly by Kunth (W. Greuter, pers. comm.): 
Caracas, E printed across the bottom “Herbier de 

t donné par M. A. Bonpland.” The 
second hand niei well be that of Richard Schomburgk. 
A printed label “isotype” is affixed over the upper right- 

bel. 


hand comer of the second la 
The isotype described immediately above includes 
two flowers. The single filament that can be observed 
in one flower is smooth. The flower in the pocket has 
five filaments that are relatively smooth, and the 
sixth bears a few tiny, narrow scales. The scales were 
barely discernible with a 10 hand lens in good, 
artificial light, and thus might have appeared as 
glabrous to Kunth. The tepals are more than twice the 
length of the stamens. Finally, the leaf margins of one 
plant are mostly short-ciliate to ciliate, and those of 
the second are irregularly denticulate to short-ciliate 
with occasional longer enations. The scapes are 
minutely and sparsely scabrescent toward the base. 
With the exception of the leaf margins of one 
its are consistent with the 
description of Phalangium pos (Kunth in Hum- 
boldt et al., 1815). 
The sheet at B-W bears two plants, one intact with 


specimen, these 


roots and the other a scape and detached leaves. Both 
the sheet and a label were annotated by Schlechtendal 
and it bears Bonpland's collection label. The sheet is 
annotated “Anth. peruvianum 1.” in the upper right- 

and corner and “Humboldt. W." in the lower right- 
hand corner (both by Schlechtendal; cf. Rankin 
2001: fig. 3). The latter 
indicates that Humboldt donated the specimen and 
that it was d in the Willdenow herbarium (W. 
Greuter, per 


Rodríguez € Greuter, 


m.) Bonpland's collection label 
ind left) is al identical to the label on the 

at P: upper left; “VII. ioi ur lower 
x K umboldt)”; and lower right; “Perou. e 
“(Humboldt)” was added by Schlechtendal (W. Greu- 
ter, pers. comm.; cf. Rankin Rodriguez & Greuter, 
2001), another indication that Humboldt was the source 
of the specimen (L. Constance, pers. comm.; W. 
comm.; also see MeVaugh, 1955. The 
general aspect of the specimen (see photographs: types 


Greuter, pers. 


of the Berlin Herbarium at F, GH) is remarkably similar 
to the isotype (P) as are the flowers. Two of the 
filaments have noticeable transverse scales, which the 
others lack. The leaf margins are ciliate with most 
enations 0.10-0.25 mm long, and the scapes are 
minutely scabrescent toward the base (C. Oberprieler, 
pers. comm.). 


Annals of the 
Missouri Botanical Garden 


The plants on the sheet at B-W and the two at Paris 
surely constitute a single gathering. First, the 
specimens are similar morphologically. Second, the 
isotypes are linked by Bonpland's collection labels 
and Schlechtendal’s annotations indicating that the 
sheet at Berlin came from Humboldt. Third, the 
annotation “Caracas” on the isotype (P) links it with 
the protologue, hence the lectotype. There being 
single gathering would explain why Schultes and 
Schultes f. (1829: 466), based on information from the 
dicun Schlechtendal, included Anthericum peruvia- 

as of A. ciliatum and Kunth (1843: 
506), e b Pond knowledge of the 


included it as a synonym o 


specimens, 
Phalangium ciliatum 

The combination Anthericum peruvianum Willd. ex 
Kunth has no standing as it was not legitimately 
described. Jackson (1895: 147) attributed the name to 
Kunth, whose name was illegitimate. In his treatment 
of Phalangium ciliatum, Kunth identified the isotype 
at B-W as “Anthericum peruvianum Willd. herb. 
no. 6657" and included this brief passage as a 
synonym of P. ciliatum. In essence, Kunth identified 
the herbarium sheet as being the same as the plants 
included in P. ciliatum. The combination is not 
legitimately described because the name was included 
as a synonym of A. ciliatum (McNeill et al., 2006: Art. 
34.1.c) and no description was mal Earlier, 
Schultes and Schultes f. (1829: 466) in their treatment 

of A. ciliatum, immediately following their reference 
to Kunth's P. ciliatum, provided a diagnosis of A. 
peruvianum and wa Sd ou) notes: *Anther- 
icum peruvianum, ..scapo simplici. Reliqu. 
Willd. Ms. (fide de Fen fil. > ” Although a type 
specimen was not designated, it is clear that the name 
was associated with the isotype at B-W. Even so, this 
combination was not legitimately described as it was 
included as a synonym of A. cililatum. 

The holotype of Anthericum glareosum was unavail- 
able for examination, and, as of April 2009, K had not 
received the isotype (P. Wilkin, pers. comm.). The 
presence of the isotype at HUT was confirmed by E. F. 
Rodríguez 


For nearly 200 


years, systematists used Kunth's Phalangium ciliatum 


Provenance of Echeandia ciliata. 


to describe plants collected in Colombia or Venezuela 
(included here in Echeandia denticulata) as well as 
auman, 1917; 


The former is surely due to 


elsewhere in South America (e.g., 


1996). T 


Kunth’s observation in the protologue “Crescit prope 


Guaglianone, 
Caracas,...,” which was repeated by subsequent 
workers (e.g., Baker, 1876; Cruden, 1986a). However, 
based on the material I examined, this and other 
information in Kunth's protologue are not supported 
by the available data. First, Echeandia is unknown 


from the region around Caracas, and the nearest 
populations of E. denticulata are in the Cordillera de 
Mérida, 300—400 km to the southwest of Caracas, an 
area that Humboldt and Bonpland did not visit 
(Sandwith, 1925; Stearn, 1968; Nüfiez & Petersen, 
1970). Second, no population of either E. ciliata or E. 
470 hexameters 
the isotypes bear Bonpland's 


denticulata has been found as low as 
(ca. 916 m). Third, 
collection labels (W. Greuter, pers. comm.), indicating 
they were collected in Peru. Fourth, Kunth (in 
Humboldt et al., 1815) described the flow as 
white, but yellow is the only flower color e for 
subsequent collections of South American species in 
These 


ect the absence of fie 


subgenus Echeandia. contradicto 


ory observa- 
tions may re notes concerning 
these specimens. Neither the lectotype nor isotypes 
bear a number that would associate them with an entry 
in Humboldt p d s field notes (W. Greuter, 
pers. comm.), t the source of the information on 
location, its enis time of collection, and flower 
olor is unknown. The absence of field notes 
undoubtedly accounts for Mc Vaugh’s observation that 
Kunth included inaccurate collection data in his 
ee id with some frequency (R. MeVaugh, pers. 
m.). In contrast, Bonpland's collection labels were 
probably reliable (W. Greuter, pers. comm.). Below, I 
briefly discuss flower color, examine Humboldt and 
Bonpland’s itinerary, and compare a number of traits 
to establish a possible origin for the Humboldt and 
Bonpland gathering. 

Given the absence of field notes and no indication 
of flower color on Bonpland’s field labels, one can 
only speculate as to why Kunth (in Humboldt et al., 
1815) described the flowers as white, whereas all the 
other specimens of Echeandia ciliata I examined were 


yellow-flowered. Further, the only flower color 


representing four species). Today, the original color 
of the flowers on the isotype at P is problematic; thus, 
h had 


aded from yellow to appearing off-white. It is also 


it is possible the flowers examined by Kunt 


possible that Humboldt and Bonpland encountered a 
rare, white-flowered population. The former seems the 
more likely explanation. 

n examination Humboldt and Bonpland's 
itinerary (Humboldt et al, 1825; Sandwith, 1926; 
1926; Stearn, 1968; Nüfiez & Petersen, 
1970) shows that they could have encountered only 
iliata. They could 


not have encountered E. herrerae or E. weberbaueri 


Sprague, 
Echeandia denticulata and/or E. ciliata 


because they did not visit the east side of the Andes 
where those d occur. Humboldt and Bonpland 
were in or near Bogo lombia, between June and 


September of 1801 e 1926), when E. den- 


Volume 96, Number 2 
2009 


Cruden 
Synopsis of Echeandia 


ticulata i is typically in flower (see below), and visited 


Núñez & Petersen, 1970) and visited or were close 

to a number of sites where E. ciliata was subsequently 

collected, e.g., between Hualgayoc and Cajamarca 

(north of Cajamarca), Baños del Inca (east of 

Cajamarca) and Cumbe Mayo, and other sites between 
h 


S 
my 


Cajamarca and Magdalena to the southwest 
Cajamarca (cf. Humboldt and Bonpland’s itinerary 
[Humboldt et al., 1825; Sandwith, 1926; Stearn, 1968] 
with specimens listed below) 


A comparison of two traits establishes Cajamarca as 


Kunth (in Humboldt et al, 1815) described the 
filaments of P. ciliatum as glabrous, as were most of 
the filaments of the two isotypes and most of the 
specimens I examined from Cajamarca. Of the latter, 
the filaments of flowers from six populations “ 
ecimens) o 

The 


filaments of flowers from two populations (six plants) 


specimens). 


bore tiny to small, quite narrow to narrow, transverse 
scales. These populations were to the east (Dillon et 
al. 2863) and southeast of Cajamarca (Sagástegui 


specimens with visible filaments that I examined from 
Colombia and Venezuela (22 plants from nine 
populations) bore numerous, transverse scales, which, 
in general, were easily observed with a 10X hand lens 
in natural light. 

Second, the leaf margins of most of the plants from 
Peru were ciliate and most of those from Colombia and 
Venezuela were denticulate. Kunth described the leaf 
margins of Phalangium ciliatum as ciliate (Kunth in 
Humboldt et al., 1815). The leaf margins of three of 
the four plants that constitute the isotypes were short- 
long). 
The margins of the fourth plant (at P) were a mixture 


ciliate to ciliate (most enations 0.1-0.25 mm 


of tiny teeth and tiny, short cilia, with occasional 


ciliate to ciliate (23 of 26 specimens) or ciliate to 
long-ciliate (three of 26 specimens). In contrast, the 
eaf margins of most of the plants I examined from 
Colombia and Venezuela were entire to denticulate 
(39 of 49 plants) (enations to 0.14 mm long) and 
occasionally denticulate/short-ciliate (seven plants) or 
short-ciliate (three plants). 

ased on these comparisons, it seems reasonable to 
conclude that the type gathering was made in 


Cajamarca rather than Colombia. The specimens I 
examined from Cajamarca had ciliate leaf margins and 
the filaments were either scaleless or bore narrow to 
quite narrow, transverse scales, whereas the flowers of 
the plants in the Humboldt and Bonpland gathering 
had a mix of filaments that were either smooth or bore 
quite narrow, transverse scales as well as having 
ciliate leaf margins. In contrast, the plants from 
Colombia had noticeably scaled filaments and most 
had entire to denticulate leaf margins. 

Also, plants from Colombia and Venezuela were 
more likely to have smooth scapes, whereas those from 
Peru were more likely to be weakly scabrescent 
toward the base to scabrous in the lower half, whic 
consistent with the type gathering being made in Pi 
Kunth (in Humboldt et al., 1815) described the scapes 
of Phalangium ciliatum as smooth apically and 
2 n scabrescent toward the base. This was true 
of the isotypes and 14 of 
ens (the others being smooth or nearly so). In 


specimens from 


contrast, 56 of the 63 specimens from Colombia and 
Venezuela had smooth scapes, and the scapes of the 
other seven were weakly scabrescent toward the base 
The only information that is inconsistent with 
Cajamarca being the location of the Humboldt and 
Bonpland gathering is the ering peri 
Echeandia alina: Based on T nine EN I 
examined, E. ciliata flowers from late December to 
late May, whereas Humboldt and Bonpland were in 
Cajamarca in Septe . Thus, it might seem unlikely 
that Humboldt nd found E. ciliata in 


flower. However, flowering at atypical times occurs in 


mber. 
and Bonpla 


other species, e.g., E. leucantha (see below; see also 
Hofreiter & Rodríguez, 2005). Su 


be relatively common in tropical regions with seasonal 


ch occurrences may 
dry forests where flowering may be triggered by 
rainfall rather than De (e.g., Augspurger, 
1981; see also Janzen, 1966; Rathcke & Lacey, 1985). 

The available evidence eis two conclusions: 
1) the Humboldt and Bonpland gathering included 
the lectotype and the isotypes, and (2) the gathering 


— 


was made in Cajamarca. Several lines of evidence 
support the first conclusion. The specimens are 
morphologically similar, and Bonpland’s collection 
labels on the isotypes are virtually identical. The 
annotations on the isotype at P connect it with the 
lectotype, and the annotations on the isotype at B-W 
link it to that at P. In addition, the inclusion of 
cum a synonym of A. ciliatum by 
edhe: E ier 3 (1829) and as a synonym of 
Phalangium ciliatum by Kunth (1843) connects it 
with the lectotype. 
Likewise, several lines of evidence are consistent 
with the conclusion that the Humboldt and Bonpland 
gathering was made in Cajamarca. First, the Hum- 


Annals of the 
Missouri Botanical Garden 


boldt and Bonpland specimens are similar morpho- 
logically to specimens collected in Cajamarca and 
distinct from material collected in Colombia and 
elsewhere in Peru. Second, Bonpland’s collection 
labels on the putative isotypes indicate the specimens 
were collected in Peru, and Bonpland’s labels are 
considered to be relatively accurate (see above). 
Third, Humboldt and Bonpland were in Cajamarca but 
did not visit the east side of the Andes where 


to describe material from 

Colombia and Venezuela for nearly 200 years reflects 

three factors: (1) the repeated citation of Caracas as the 

type locality (es. ., Baker, 1876; Cruden, 1986a); (2) 
with th 


a few traits, Echeandia denticulata 


"e E. ciliata are remarkably similar, and without a 
clear alternative there was no reason to question the 
usage; (3) based on the material I examined, no 
collections of E. ciliata were made between 1802 and 
1958. In essence, the alternative was unknown 
Additional specimens examined. PERU. Cajamare 
Cajabamba, near Can Vitis nes "n MICH); 
Hualgayoc, Las Ventanill ull 0 km S of 
Bambamarca, e z eee et al. P 5 km S of 
m o the coast, Sánchez pM 3266 (F); xd. to 
TURN w of Cajamarca, Sdnchez Vega et al. 1846 (F); 
Baños del Inca, above Pullucana, carr. r. Cajamarca—Celendin, 
del Inca on rd. from 


Celendín, Km 20, e & Müller 9375 (LZ); 43 km E of 
Cajamarca, 27 km N of San Marcos, Dillon et al. 2862 (F, 
MO). 


2. Echeandia denticulata Cruden, sp. nov. TYPE: 


2750 m, 15 Aug. 1939, J. Cuatrecasas 6639 
(holotype, COL!; isotypes, F!, US!). Figure 1. 


species Echeandiae ciliatae (Kunth) Cruden et EF. 
Ps s (Poelln.) Cruden similis, sed ab eis filamentis 
manifeste squamosis et foliis basalibus integris usque 
denticulatis iu a E. idus (Killip) Cruden stylis 
longioribus et floribus cernuis usque patentibus differt; 
Colombia et eee indig, 

Storage areas of roots enlarged 0.5-1.5 cm from the 
rhizome, 1-4(-6) em long. E leaves (3 to)4 to 14, 
(8-)10-4.2(-57) em X (3-)4-10(-12) mm, straight and 
flat to falcate, entire to denticulate, rarely ciliate or 
enations to 0.15(-0.26) mm; 
leaves O to 2(or 3), if present, the lowest 1.3-3.4 
(-5.1) em. Scape l(or 3), (10-)15-55(-85) em high, 


glabrous throughout or occasionally glabrous apically 


long-ciliate, cauline 


and minutely scabrescent toward the base, O to 2(to 4) 
branches; main axis of inflorescence with (3 or)4 to 
11(to 14) flower-bearing nodes, the lowest 2-flowered, 
occasionally 1-flowered in small plants or 3-flowered 


in large plants. Flowers yellow, most cernuous to 
patent; pedicels (4—)5.5-10(-11) mm; tepals 12.5— 
18 mm, probably spreading to somewhat re- 
flexed, m 

straight, bearing numerous, narrow, transverse scales; 
(1.5-)2-3 mm, 
weakly curved, dehiscence lateral; ovary (2.534 
(5) mm; styles (5.5-)6-8.5(-9.5) mm, bent downward 
and then forward or weakly to strongly deflexed, rarely 


inner ide; filaments 5-7 mm 


ps 
yum 


anthers free, versatile, straight to 


straight, arcuate, occasionally turned upward below 
the stigma, the stigmas exserted (1—)2—4(—5) mm 
beyond and below the level of the anthers. Capsules 
oblong, (9211-17 X 4.5-7 mm 


Distribution and habitat. This species is found 
primarily between 2300 and 3500 m, rarely to 
3900 m, in the Cordillera Oriental in Colombia and 
Cordillera de Mérida in adjacent Venezuela (Fig. 1). 
Two collections made near Quetame, Colombia 
(Pennell 1856 [NY, US] and Lehmann 8842 [F, KJ), 
are elevationally and somewhat geographically dis- 
junct from the nearest populations of Echeandia 
denticulata. T 
m, a little southeast of the southeastern-most 
collection of E. denticulata, which was made south of 
Usme between 3000 and 3100 m. Echeandia denti- 


culata is a plant of savannas and páramo, and occurs 


ese were made between 1300 and 


occasionally in pastures and cultivated fields. 


Phenology. In flower (April to) mid-June to 
September (to mid-October). 


Discussion. In addition to being geographically 
disjunct from its Peruvian relatives (Fig. 1), most 
specimens of Echeandia denticulata were easily 
identified by their scaled filaments and/or entire to 
of the 
specimens of E. denticulata with visible filaments 


denticulate leaf margins. filaments 
(22 plants from nine populations) were scaled, and the 
leaf margins of most of the specimens were either 
entire (seven of 49 specimens), entire/denticulate (six 
(26 


relatively few were either denticulate/short-ciliate 


specimens), or denticulate specimens) 

(seven specimens) or short-ciliate (three specimens). 
In contrast, the filaments of most of the Peruvian 
specimens were either smooth (37 of 66 specimens) or 
wrinkled and scaleless (16 specimens), and relatively 
few specimens a of the 66) bore ae mea 
of the 
Peruvian plants were short-ciliate to long- Ee (79 


transverse scales. margins of m 
of 99 Poncii or denticulate/short-ciliate (18 
specimens), and the leaf margins of just two 
specimens were denticulate. Also, compared to those 
of E. denticulata, the capsules of E. ciliata were 
smaller (10-12 mm vs. [9-]1 1- 


flowers of E. PHA were erect rather than cernuous 


—]7 mm long) and the 


Volume 96, Number 2 
2009 


Cruden 
Synopsis of Echeandia 


or patent, with tepals shorter (9-13[-15] mm vs. 12.5— 
18[-20] mm) and styles geniculate and shorter (2-6 
—6.5] mm vs. [5.5-]6-8.5[-9.5] mm). Further, the 
anthers of E. denticulata were versatile and somewhat 
shorter than those of E. weberbaueri ([1.5—]2—3 mm vs. 
[2-|2.5-4 mm long) and became twisted during or 
after dehiscence, whereas those of E. weberbaueri 
were nonversatile and usually remained straight after 
dehiscence. 

Other traits differed qualitatively between Echean- 
dia denticulata and its Peruvian relatives. The scapes 
of 56 of 63 specimens of E. denticulata were glabrous 
and seven were minutely scabrescent toward the base, 
whereas the scapes of most of the Peruvian plants 
were minutely scabrescent to scabrous (86 of 115 
E. denticulata, there was 
length and 
length, but there were strong biases in the Peruvian 


lants). Among specimens o 
no relationship between tepal stamen 
species. The tepals of E. denticulata were more than 
twice the length of the stamens in nine plants, twice 
the le 


twice the length of the stamens in seven plants. In 


ngth of the stamens in six plants, and less than 


contrast, the tepals of E. ciliata were usually more 
than twice the length of the stamens (12 of 17 
specimens), and those of both E. herrerae and E. 
weberbaueri were usually less than twice the length of 
the stamens (40 of 41 plants from 18 populations and 
11 of 14 flowers from seven plants, one population, 
respectively). 


Etymology. The specific epithet describes the 
denticulate margins of the basal leaves, which help to 
differentiate this species from its Peruvian relatives. 
Tcu BIA. ne 1203 (BM) Boyacá: 
a de San Pablin, Grubb, 
Cury & F mum EP i (&). Grubb, Com ry & 
Ky W slope above Villa de Leyva, 
74 (MO). Cundinamarca: Bo: 


Parat GN 


Melamp 


Sarona a 031 (COL), 
Hia 01048 (COL) 30 a NW of Bogotá, Vereda de 
Rozo, 4 km S of Cota, Fassett 25660 (NY, US); Guasca, 
Arbelaez 1134 (COL, US); Ubaté, 100 km N of Bogotá, Kgie 
4521 (C, US); “Terreros” [Bosa], van der Hammen 469 (COL); 
S end of Suba Hill, near Bogotá, Schiefer 881 (DS, GH). 
VENEZUELA. Mérida: betw. Apartaderos & Santo Dom- 
r& i 4542 (VEN); bet cde Dn & 
22 (US); Mucurubá, um 95 (F, MO, 
cig es de 
Cacute, Aristeguietia 3284 (NY, VEN); Páramo de Mucubají, 
near Laguna Grande, Schulz, Rodriguez & Sánchez 105 o 
head of Río Santo Domingo, Finca La Corcavada, Schul. 
Rodríguez 681 (U) 


3. Echeandia herrerae (Killip) Cruden, comb. nov. 
Basionym: Anthericum herrerae Killip, J. Wash. 
Acad. Sci. 16: 566. 1926. TYPE: Peru. [Cuzco:] 


Paucartambo, Hacienda Churá, 3500 m, Jan. 
L. Herrera 1012a (holotype, US!). 


Storage areas of roots enlarged 0.5-1.5 cm from the 
rhizome, 1.5-3 cm long. Basal leaves (4 to)6 to 12(to 
14) 3-18(-22) em X 


spreading, 


mm, usually falcate, 


the margins densely short-ciliate to 
densely ciliate, or rarely densely denticulate or 
densely long-ciliate; cauline leaves 0 or 1, if present 
6-37 mm long. Scapes 1 or 2(to 5), 3-26(-ca. 45) cm 
high, weakly to strongly scabrescent toward the base 
with well-developed enations, occasionally scabrous 
or scabrescent throughout, rarely glabrous or nearly 

1 th of the 
leaf, unbranched or rarely l- o 


branched; main axis of inflorescence with (2 203 to 


longest 


ll(to 13) flowering nodes, lowest nodes l- or 2- 
ered, upper l-flowered. Flowers yellow, erect; 
—6(—7) mm; tepals 9—13(-15) mm, 


less, rarely more than twice the length of the stamens, 


flow 
pedicels usually 


probably spreading, inner 4—6 mm wide; filaments 
shallow 


> 


.5(-7) mm long, + dnx inserted in a s 
pit, üsüallş smooth or wrinkled and scaleless, 
occasionally bearing small, narrow, transverse scales; 
anthers free, 1-2(-2.5) mm, versatile, usually becom- 
g curved or twisted during or after eee. 
aae lateral; ovary 2—4.5(-5) mm; style 
(—6.5) mm, geniculate, the stigmas exserted to "RE 
mm lateral to the stamens, rarely straight and the 
gma equal to or barely exceeding the anthers. 
Qa Pr to narrowly oblong, 7-15 X 


Distribution and habitat. 
distributed on the east side of the Andes in Peru from 


This species is widely 


Junín south to Apurimac and Cuzco between 2800 and 
3824 m (Fig. 1). It inhabits dry, rocky, or gravelly 
hillsides, fallow | fields, and, 


occasionally, more mesic habitats. 


pampas, pastures, 


Phenology. In flower December to April. 


Discussion. Plants of Echeandia herrerae are 
distinguished by a number of floral and vegetative 
traits, which separate them from the other South 
American species of Echeandia. Most specimens were 
easily identified by their short (3-26[-ca. 45] cm 
high), scabrescent to scabrous scapes, which were 
usually one to two times the length of the longest basal 
leaf, and/or the densely short-ciliate to ciliate leaf 
margins (52 of 65 specimens from 27 collections). 
Occasional specimens had either long- ciliate (four 


rarely straight styles, and short pedicels (2.5—-6[— 
mm long). The latter were probably too short to 


Annals of the 
Missouri Botanical Garden 


accommodate the bending that occurred in the 
of the 
weberbaueri, which e flowers that are cernuou 
patent and pedicels that are (4—)5—-14(-15) mm ed 
The erect flowers and shorter (2-6[-6.5] mm vs. 
[5.5-]6—11 mm long), geniculate styles distinguished 


pedicels ciliata and E. 


this species from the other South American Echean- 
ia. Regardless of the inclination of their flowers, the 
styles of the other species were straight, arcuate or 
somewhat declinate, and exserted well beyond their 
E. herrerae were, in 
of the other Peruvian 
species (3-26[-ca. 45] em vs. [8-]15-62[-85] cm 
high), and those of half the specimens (42 of 80) were 


noticeably scabrescent to scabrous, whereas the 


anthers. Also, the scapes of 
general, shorter than those 


scapes of E. ciliata and E. weberbaueri were either 
smooth or minutely to weakly scabrescent near the 
base. Because they are geographically sympatric, Æ. 
herrerae and E. weberbaueri might be confused. In 
addition to its geniculate styles, E. herrerae had 
shorter scapes (3—26[-ca. 45] em vs. 34—59[-68] cm 
high) and shorter, versatile anthers (1-2[-2.5] mm vs. 
[2-]2.5-4 mm long) that 


dehiscence rather than 


twist during or after 

being nonversatile and 

Further, £. 
errerae occurred at higher elevations (2800-38 

vs. 2300-2700 m). 


remaining straight after dehiscence. 


Etymology. The specific epithet honors Fortunato 

i i 875-1945), well- 
known Peruvian botanist and professor of botany at 
the Universidad Nacional San Antonio Abad del 


Cuzco, who collected extensively in the region around 


Lucian Herrera y Garmendia 


uzco. 
Additional specimens examined. PERU. Jan. 1864, 
Pearce s.n. (K). Apurima f And: 


e: Quebrada, 2 km N 
072. 


O ; 
MO) Abancay, 
Curahua 929 Cuz Apurimac Valley, 
Herrera 3067 (US); E km W a Cusco, o. Elle nberg 1195 (U); 
Anta, Cillapuyu, El Chaccan, 28 Dec. 1972, Brunel 204 
(MO); Mar. 1973, Brunel 621 Cae Ancahuasi, NW Cus 
, carr. a Cusco, Núñez 7252 (MO); m 
vador, uds 152 o Cuzco, Herrera 
96 (F); Cuzco, pem 893 (F); Stafford 268 (K); Huasao, 
ds 3108 (US); near Pisac, Hunnewell 15905 (GH); 
Saqsa 


LZ); Paucartambo, near 


od Davis et al. 1 595 (E); jd. 


Yanacona, near Perga Kachun, Q 
7B (F, ine Valley of the Urubamba, 
Ollantaitambo, Herrera 3457 (F, G). Huancavelica: Huan- 
cavelica, Larmes, E of Conaica, Tovar 1 78 (MO, US). Junín: 
Jauja, Cerro Gloria Malca (serranias that siren’ Jauja), 


Ochoa 275 (BH, US); Huancayo, Soukup 3972 (COL, F, GH); 
Huancayo, Quebrada Occopilla, Soukup 3641 (US); near 


Pachacayo, Gutte 2993b (LZ); above the train station, Gutte 
9263a,b (LZ). 


4. Echeandia lehmannii (Baker) Marais & Reilly, 
Kew Bull. 32: 662. 1978. Basionym: due 
lehmannii Baker as “lehmanni,” Bot. Jahrb. Syst. 
8: 208. 1887. TYPE: Ecuador. Rare on do 

[soil] near Malchinguí, S slope of [Volcán] 

Mojanda, 2800 m, 28 Jan. 1881, F. C. Lehmann 

429a (holotype, BM!, fragm. K!; isotype, G!). 

Figure 1 


Echeandia po Ravenna, Dus i 327. 1985. 
Syn . TYPE: Ecuador. Pichi a Mitad del 
Mu dosel Hacienda a 7 ae 1979, R. 
Jaramillo & D. Silva 933 (holotype, AAU not seen; 
isotype, QCA not seen). 


Storage areas of roots enlarged 0.5-2 cm from 
rhizome, 1—2.5 cm long. Basal leaves 5 to 8, 8-20 cm 
x (3-)4-9[-10] 
o to short ciliate; cauline leaves 1 to 3, 

cm long. Scape [19-]20-32(-40) cm 
high, puis or with a few 


mm, margins undulate, minutely 
lowes 
small enations near 
the base, 0- to 2-branched; main axis of inflores- 
cence with (5 to)9 to ló[or 17] flower-bearing 
nodes, lowest 2- or 3-flowered. Flowers yellow, 
nutant; tepals 10-15 mm, probably reflexed, inner to 
5mm wide; filaments narrowly clavate, bearing 
mm; anthers con- 
d, 0.8- 
—|3-5 mm. Capsules oblong 8— 


numerous small pk [2.8-.]5 
nate, 4.5—5.5 mm, apex of cone deeply lobe 
l mm across; ovary y [2 
13 X 4-4.5[-5] mm. (Measurements in brackets are 
from Ravenna's [1985: 327] description of Echeandia 
aequatoris.) 


Distribution and habita 


Collection localities were 
provided for thre E 


e of the fon collections o 

dorian endemic. Two of these were made 20-25 km north 
of Quito in or close to the Pululahua Geobotanical Res- 
the northeast of 
Quito near Malchinguí on the southern/southwestern 


erve, and the third was made 25-30 km to 
slope of Volcán Mojanda—Fuya Fuya (Fig. 1). Elevations 
of 2800 and 2850 m 
The two areas are approximately 30 km apart. The plants 


were provided for two collections. 


were growing on alluvial soils (Lehmann 429a) and in 
open grassy areas (Humbles 6. 


Phenology. In flower late January to April. 


Discussion. The combination of yellow flowers and 
connate anthers distinguishes Echeandia lehmannii 
from the other South American species. 

There is no evidence that this species occurs other 
than in a small area north of Quito. The limited range 
and few collections suggest that Echeandia lehmannii 


Volume 96, Number 2 


Cruden 
Synopsis of Echeandia 


is narrow endemic and quite rare. The Lobb 
collection (Lobb 33 E i is ean mislabeled, 
as numbers of his s s from Peru and Ecuador 


were labeled “Columbia” ‘Killip, 1934). 


Nomenclature. I have included Echeandia aequa- 
toris as a synonym of E. lehmannii because there are 
no obvious and/or substantive differences between 
them. Ravenna’s observation (1985: 327) that “the 
species [aequatoris| represents the first record of 
Echeandia from Ecuador" suggests that he was 
unaware of Marais and Reilly's (1978) transfer of 
Baker's species into Echeandia. Given the geographic 
proximity of the few populations, there is reason to 
question whether E. aequatoris represents a distinct 
species. A comparison of the protologue of E. 
aequatoris with the few available specimens of E. 
lehmannii revealed no substantive differences be- 
tween the two. For most traits, the variation observe 
by Ravenna (1985) was included within the variation I 


observed, and the few exceptions provide minimal 


E 


extensions of the ranges of those traits, e.g., leaf widt 
m 3-9 to 3— 


might be filament length, but equivalent variation 


10 mm (see above ne exception 
exists in other species (e.g., E. leucantha). Also, it 
seems quite unlikely that two morphologically similar, 
quite rare species would occur within a few kilometers 
of one another. Finally, the holotype of E. aequatoris 
was on loan and unavailable for study as of June 2001 
and September 2004: (Nørgaard, pers. comm.; Balslev, 
pers. comm.). Renato Valencia Reyes confirmed the 


presence of the isotype at QCA. 


Etymology. The specific epithet honors Friedrich 
Carl Lehmann (1850-1903), who collected widely in 
Latin America and made the first collection of this 
and other species now included in Echeandia. 


Additional oa examined. COLUMBIA O a 
33 (K). ECUAD ncha: Pululahua Crater, ca. 
N of Quito, =o 6283. (F, MO [2]. 


5. Echeandia weberbaueri (Poelln.) Cruden, comb. 
nov. Basionym: Anthericum weberbaueri Poelln., 
Revista inim Bot. 7: 103. 19 
Peru. Huancavelica: Tayacaja, valley of the 
Mantaro River 
2400 m, 14 Mar. 1913, A 
(holotype, B! [image examined online]; isotypes, 


DS!, F!, GH! [2], NY!, US! [2]). Figure 1. 


Storage areas of roots enlarged 0.5—1.5 cm from the 
rhizome, 1.5-3 cm long. Basal leaves 7 to 12(to 16), 
(6.5—)10—25(-33) em X (36-12 mm, flat, occasion- 
ally falcate, margins ciliate to long-ciliate (Weberbauer 

81) or denticulate to short-ciliate (Tovar 4001); 
cauline leaves O to 2, if present, the lowest 16-35 


(49) mm. Scape l(to 3), 34-59(-68) cm high, 
frequently (8 of 12 plants) more than twice the length 
the longest basal leaf, smooth or nearly so, 


S 
za 


occasionally minutely scabrescent toward the base, 
sometimes with l(or 2) few-flowered branches; main 
axis of inflorescence with (4 to)7 to 15 flower-bearing 
nodes, lowest nodes 1-, 2(or 3)-flowered, upper 1- or 
2-flowered. 

pedicels 6-1 


Flowers low, cernuous to patent; 
4(-15) mm; tepals 11-16.5(-18.5) mm, 
usually less than twice the length of the stamens, 
probably spreading to somewhat reflexed, the inner 
tepals 5-8 mm wide; filaments 6—7 mm, - straight, 
smooth, occasionally wrinkled and sealeless or 
bearing tiny, narrow, transverse scales; anthers free, 
(2-)2.5-4 mm, straight to reniform, non-versatile, the 
anther walls strongly reflexed holding the anther on 

he same axis as the filament (cf. Weberbauer 6481 
mne NY, USp, rarely versatile or apparently so, 
occasionally twisted or strongly curved, dehiscence 
lateral; ovary 2-3(-3.5) mm; styles 7-11 mm, straight 
to gently deflexed (to 30°), turned upward below the 
stigma, the stigma exserted (2-)3-5(-6) mm beyond 
the anthers. Capsules narrowly oblong, 12-13.5 X 


Distribution and habitat. 
known from two collections made 5—10 km from one 


This Peruvian species is 


another in or near the valley of the Río Mantaro in 
Huancavelica, ca. 60 km east of Huancayo (Fig. 1), 
between 2300 and 2700 m. The available data suggest 
it is a species of open hillsides and cultivated land. 


Phenology. In flower February to April. 


Discussion. Plants of Echeandia weberbaueri were 
distinguished by their tall scapes (34—59[-68] cm 
high), flat, relatively broad ([3-]6-12 mm wide) basal 
eaves, some of which had long-ciliate ma 
(Weberbauer 6481) and long ([2-]2.5-4 mm), em 
nonversatile anthers, whose strongly reflexed walls 
held the anthers on the same axes as their filaments. 

The wide, flat basal leaves and long, nonversatile 
anthers of Echeandia weberbaueri helped to distin- 
guish it from E. denticulata and the other Peruvian 
species. The anthers usually remained straight rather 
than twisting during or after dehiscence and were 
longer than the versatile anthers of the other Peruvian 
species ([2-]2.5—4 mm vs. 1-2.5 mm $n whose 
anthers usually became twisted during or afte 
dehiscence. Compared to E. denticulata, a filaments 
of E. weberbaueri were mostly smooth and/or scaleless 
(13 of 14 plants) and those of one plant bore small, 
narrow, transverse scales, whereas the filaments of E. 
denticulata were us scaled. Compared to E. 

ciliata, the basal lea eber 


general, flat and Mus ~ the falcate basal leaves of 


aueri were, in 


Annals of the 
Missouri Botanical Garden 


E. ciliata ([3—]6-12 mm vs. [1-]|2-7[-9] mm wide). 
The tepals of E. weberbaueri were generally less than 
twice the length of the stamens (12 of 15 plants), 
whereas those of most specimens of E. ciliata were 
twice or more than twice the length of the stamens (13 
of 15 plants) Also, the scapes of most plants of E. 
ciliata were three to four times the length of the 
longest basal leaf, whereas those of E. weberbaueri 
were mostly less than twice the length of the longest 
basal leaf. Specimens of E. weberbaueri were easily 
distinguished from those of E. herrerae, with which it 
is geographically sympatric (Fig. 1). The scapes of E. 
herrerae were shorter (8-26[-ca. 45] cm vs. 34-59 
[-68] cm high), frequently scabrescent to scabrous, 
the anthers shorter (1-2[-2.5] mm vs. [2-]2.5—4 mm), 
the styles shorter (2-6[-6.5] mm vs. 7-11 mm) and 
geniculate, and it occurred at higher elevations 


(2800-3824 m vs. 2300-270 


The ud epithet honors August 
To (1871-1 
and professor of s at the Universidad Nacional 
Mayor de San Marcos in Lima, Peru, who made the 
first collection of this species. 


erman-educated botanist 


Additional specimen. examined. 


ERU. Huancavelica: 
Tayacaja, near Huachocolpa, Tovar 4001 (US). 


Ib. gp Mscavea Cruden, Novon 9: 326. 1999. 
Echeandia mevaughii Cruden, Contr. 
i Michigan Herb. 16: 129. 1987 


Flowers white, occasionally cream-colored, yellow, 


or orange; inner tepals narrowly elliptic; flowers 
opening in late morning or early afternoon and closing 
near dusk (see Cruden, 1999). The subgeneric epithet 
honors Marion Stilwell Cave (1904— isti 


guished embryologist and student of the Liliaceae s.l. 


, distin- 
(see Constance et al., 1996 


6. Echeandia bolivarensis Cruden, Ann. Missouri 
ot. Gard. 76: 350. 1989. TYPE: Venezuela. 
Bolívar: Igneous forested slopes, Serrania de 
Pijiguaos, 160 km SW of Caicara del Orinoco, 
6°35'N, 66745"W, 100- T m, 12 M Dr J. 


ims 


Steyermark, Holst & Man 761 
(holotype [2 ets ona an 
isotype, UC!). F 


Storage areas of roots enlarged (1—)3—6 cm from the 


rhizome, 3-5(- m long. Basal leaves 6 to 1 


€ > 
scending, narrowly lanceolate above the base, 
minutely denticulate-serrulate, mostly 45-60 cm X 
11-18 mm; cauline leaves 4 or 5, the lowest 10— 
19 em, long attenuate. Scape glabrous, 98-118 cm 
high, 1- to 5-branched; 


9 to 14 flower- beann nodes, the lower 4-flowered, 


main axis of inflorescence with 


the upper 2- or 3-flowered. Flowers yellow; tepals 10— 
ll mm, inner ca. 3.5 mm wide; filaments linear, 
6.5 mm long, sparsely to noticeably scaled with tiny to 
small scales, inserted + basally in a deep pit; anthers 
free, 2.4—4 mm, somewhat flared basally, nonversa- 
tile, dehiscence apical; ovary 2-2.5 mm; style 7— 

Capsule broadly oblong, 10-11 X 6.5- 


1.5 mm. Seeds 2.5-3 mm across. 


Distribution and habitat. This Venezuelan en- 
mic is known from just two collections made 
ua 100 and 600 m in the Serranía de Pijiguao 
at 6735'N, 6645'"W and 6734'N, 66°47'W, respec- 
tively, a little west and southwest of Los Pijiguaos 
(6735'N, 66744" W) in western Bolívar (Fig. 1). 
type gathering was made on an open rock face in the 
forest, and Gróger & Berg (1064) reported it growing 
in cushions of Selaginella P. Beauv. among Vellozia 
h 


tubiflora (A. Rich.) Kunth. 


Phenology. In flower late July to September. 
Discussion. This rand is easily distinguished 
from other South Am Echeandia by th 


combination of ie es relatively short tepals 
(10—11 mm long), free, nonversatile anthers, tall (98— 
118 em) scapes, and roots with long (4—5 em) storage 
areas that are enlarged 4—5 cm from the rhizome 

I have included “his species in Echeandia subg. 
Mscavea with some hesitation because the plants are 
yellow-flowered, which is rare in the subgenus (Cruden, 
1999). However, a number of other traits are consistent 
with this disposition. The insertion of the filaments is 
basal, or essentially so, in a deep pit. Basal insertion is 
uncommon in subgenus Mscavea (six of the other 25 
species) but rare in subgenus Echeandia (one of 55 
species). Second, the distance from the opening of the 
pit to the base of the anther sac is less than 0.25 mm, 
which is common in subgenus Mscavea (12 of 20 
species examined) and rare in subgenus Echeandia 
(one of 19 species examined). Third, the anthers 
dehisce apically, as do the anthers of the other four 
species in subgenus Mscavea with free anthers, whereas 
in subgenus Echeandia, apical dehiscence occurs in 
only four of the 32 species with free anthers. Also, the 
inner tepals are ca. 3.5 mm wide and the capsules are 
broadly oblong, which are typical of species in 
subgenus Mscavea but uncommon or rare in subgenus 
Echeandia. Finally, E. bolivarensis occurs below 
800 m, as do eight of the other 25 species in subgenus 
Mscavea, whereas only six of 53 species in subgenus 

cheandia occur below 800 m and two of these occur in 
northeastern Mexico and/or the adjacent United States, 
which has a temperate climate. 


Etymology. This species was named for the state 
in which the plants were collected. 


Volume 96, Number 2 
2009 


Cruden 
Synopsis of Echeandia 


Additional specimen examined. VENEZUELA. Bolivar: 
Los Pijiguaos, NE Campamento de Bauxiven, summit Cerro 
a Guacamaya, Gróger & Berg 1064 (MO) 


7. Echeandia leucantha Klotzsch, Allg. Gartenzei- 
tung (Otto & ir 8: 275. 1840. TYPE: 
Venezuela. colon Tovar, 3500 ft., 
1854-1855, 4 Fender 1549 (neotype, desse 
ed here, GH!; isotype, K!). Figure 1 

Echeandia prolixa Woodson, Ann. Missouri Bot. Gard. 29: 
3 . TYPE: Panama. Panamá: Vic. of Bejuco, 7 

Sep: 1942, P. H. Allen 2962 (holotype, MO! [2]; eines 

GH). 


Storage areas of roots 2-3.5 cm, enlarged 3-6 cm 
from the rhizome. Basal leaves 7 to 12, (15-27-83 cm 
X 6-20(-25) mm, narrowly lanceolate to broadly 
elliptical, frequently falcate, serrulate to short-ciliate; 
cauline leaves 1 to 6(or 8), lowest 2.5-17 cm. Scape 
(40—)65-112 em high, glabrous, 0- to 4(or 5)-branched; 
main axis of inflorescence with 7 to 15(or 17) flower- 
bearing nodes, the lowest 2- to 4-flowered, upper 1- to 
3-flowered. Flowers white, nutant; tepals 8.5-12 mm, 
(2.5-)4— 


mm long, cone 


bem. reflexed; filaments scaled, linear, 

anthers connate, (3.5)4-6(—7) 
more is 1.6 mm wide, 0.8-1.2 mm wide at the apex; 
ovary 1.5-2.2 mm. Capsules broadly n 7-9 X 
4.5-5.5 mm, rarely globose, ca. 5 X 


Distribution and habitat. 
northwestern Venezuela and adjacent 
(Fig. 1) north t onduras. It inhabits 
woods, pastures, and roadsides usually below 1200, 
rarely to 1500 m 


This species occurs from 
Colombia 
savanna, 


Phenology. In flower mid-June to early Septem 
ber, also Aiea in flower 21 January 1982 naa 
& Ortega 821). 


Echeandia leucantha is one of the two 
in subge 


also occurs in Central America. The South American 


Discussion. 
South American species nus Mscavea that 
populations were white-flowered, but those in Hon- 
duras and some in Nicaragua were yellow- or orange- 
flowered. Specimens from South America were easily 
identified by the combination of white flowers, 
connate anthers, scaled filaments, and storage areas 
of the roots that develop 3-6 cm from the rhizome. 
The scaled filaments and storage areas that develop 
away from the rhizome distinguish it from E. pittieri, 
which has smooth filaments and storage areas that 
develop 1-2 em from the rhizome. The Purdie (s.n.) 
collection from Colombia lacks open flowers and is 
included here with reservation. 


Nomenclature. 
ed to 
specimen that includes roots with storage areas, 


The Fendler specimen was select- 
e the neotype because it is a complete 


flowers (in the packet), and capsules. The collection 
was made less than 50 km from Maracay, i.e., the 
locality cited by Klotzsch (1840) 

The designation of a neotype was necessary 
because no specimen was cited by either Klotzsch 


1840) in his protologue or Otto (1840) in his Nachtrag 
to the protologue; the protologue was not accompanied 


a 


y an illustration; and a putative type specimen was 
not found among the specimens examined. Klotzsch’s 
(1840) nt was based on d growing in 
the Botan ar in rom materia 
collected by. Morita i in 1836 near Po Venezuela. 
The most likely places to m Moritz's specimen, if 


one wer ; 1 HAC, or possibly W 
(Stafleu & Cowan, 1981 2 . Such a mee was 
not found at B (C. Oberprieler, pers. c s 


pers. comm.), BM (R. Vickery, pers. MU K (P. 
Wilken, pers. comm.), or W (B. Wallnófer, pers. 


comm.), and I received no reply to enquiries sent to 
HAC 


At the time Echeandia leucantha was described, 
Otto was the gardener at B and it is possible that he 
made a collection of this species. If he did, it is 
. C. Oberprieler, 


pers. comm.), and no specimen was found at GOET (J. 


missing at B (T. Raus, pers. comm 
Heinrichs, pers. comm.). The latter institution was 
suggested as a source for Otto's duplicates (Stafleu 

Cowan, 1981: 858). Some of Moritzs Venezuelan 
destroyed during World War II (B. Wallnófer, pers. 


comm.J. 


Etymology. The specific epithet describes the 
white flowers, which are typical of species in 
subgenus Mscavea. 


Additional us examined. COLOMBIA. Magda- 
lena: Santa Marta, Purdie s.n. (K); of Manaure, 
Haughi 4342 (NY, UC, US). VENEZUELA. 1846, Funcke de 
Schlim 674 (BM, BR, G). Carabobo: near Valencia, Pütier 


Rabe 96473 (US, VEN); Guanare, terrenos de la UNE 
Stergios & Ortega 1983 (NY, VEN), Stergios & Aymard 5601 
(MO), Aymard & Ortega 821 (MO); Acarigua, Burkart 17116 
(VEN). Zulia: Perijá, Gines 1857 (US); Kunaria, Gines 1948 
(F). 


8. Echeandia pittieri Cruden, Phytologia 59: 379. 
1986. TYPE: Panama. 
Boquete toward D 2n in savanna near rocky 
creek, 2800 ft., 26 Aug 5, S. McDaniel 6810 
(holotype, MO!; ies ud 


Chiriqui: 5 mi. S of 


Storage areas of roots 1-1.5 em, enlarged 1-2 cm 
from the rhizome. Basal leaves 5 to 11, (18-) 
29-41 cm X (4-)11-20 mm, narrowly lanceolate, 


Annals of the 
Missouri Botanical Garden 


entire or short-ciliate; cauline leaves 3 to 5, the 
lowest 5-20.5 cm. Scape 80-115 cm high, glabrous, 
- to 2-branched; main axis of inflorescence with 9 
to 15 flower-bearing nodes. Flowers white, nutant; 
10-12.5 mm, reflexed; filaments smooth, 

anthers connate, 5.5—7 mm; ovary 
3 mm. Coals broadly oblong, 6.5—7.5 X ca. 
5 mm. 


Distribution and habitat. This species is known 
from two localities in Panama (Cruden, 1986b) and 
one in Colombia (Fig. 1). The two collections from 

anama were made at ca. 850 an 00 m and that 
from Colombia at 120 ne collection from 
Panama was made in a savanna near a creek, and in 
Colombia, the specimens were collected on a loma 
where it grew with shrubs. A collection from northern 
Colombia (Haught 4342) previously included in this 
species (Cruden, 1986b) is properly placed in 
Echeandia leucantha 


Phenology. In flower late July to August. 


Discussion. This species is characterized by its 
white flowers, connate anthers, smooth filaments, 
and storage areas of the roots that develop 1-2 cm 
from the se. 
that zome distinguish 
Echeandia pittieri from E. leucantha, which has 


ape. The smooth anthers and storage areas 
evelop close to the rhi 


scaled filaments and longer storage areas (2-3.5 cm 
long) that develop 3-6 cm from the 
rhizome. 


Etymology. The s epithet honors Henri 
Francois Pittier (1857-1950), a Swiss botantist who 
collected widely in ine Venezuela, and Central 


America. 


Additional specimens examined. COLOMBIA. Valle del 
Cauca: W slopes of Cordillera Occidental, valley of the Río 
Sanjuniquín, Naranjal, Cuatrecasas 15356 (F, US) 


Literature Cited 


Augspurger, C. K. 1981. Reproductive eus A 4 a 

tropical shrub: Experimental studies on effe of 

pollinators and seed predators on Hybanthus pesca 

8. 

. Revision of the genera and pus of 

hericeae nd Eriospermeae. J. Linn. Soc. . l5: 
253-363. 

piensa P; d E A. Montalvo. ada The E 

of E a3l: 


Liliaceae). Britto 


64—1 Zl 

Berry, P. E. 1982. The systematics and evolution of Fuchsia 
sect. Fuchsia (Onagraceae). Ann. Missouri Bot. Gard. 69: 
1-198. 

Conran, J. G. 1998. Anthericaceae. Pp. 114—121 in K. 
Kubitzki (editor), The Families and Genera of Vascular 
Plants, Vol. 3. dis; Plants: Monocotyledons. Spring- 
er-Verlag, Berlin 


Constance, L., D. R. Kaplan & R. Ornduff. 1996. Marion 


a. New combinations in Echeandia and 
us (Liliaceae) Phytologia 59: 379-380. 

1986b. New species of Echeandia (Liliaceae) from 
Gens America. Phytologia 59: 373-379 
enbachia, a misplaced genus of New 
World Liliaceae. Nord. J. Bot. 7: 255-260 
Echeandia. P. 2 7-30 in G. David se, M. S. 
. O. Chater (editors), Flora. Mesoamericana. 


genus and fifteen new species of 
Echeandia (Anthericaceae) from Mexico and the United 
States. Novon 9: 325-338. 

R. McVaugh. 1989. Echeandia. Pp. 178—197 in R. 
McVaugh (editor), Flora Novo- o Vol. 15. Univer- 


of 3 new and critical 
cd Io ae Daedalus. Proc. Amer. Acad. Arts 
33: 471—489. 
oe E. R. 1996. Liliaceae Juss. Pp. 228-240 in F. 
O. Zuloago & O. Morrone eg ds Catálogo de las Plantas 
a Argentina. I. Pteridophyta, 
(Monocotyledoneae). 
ard. 60. 


Monogr. Syst. Bot. Missouri Bot. 

Hauman, L. 1917. Notes floristiques, quelque 
games, gymnospermes et monocotylédones de ee 
Anales Mus. Nac. Hist. Nat. Buenos Aires 29: 391 
443. 

Hensold, N. 1999. T dé del Dpto. de 


Cajamarca, Perá. puse 6: 141— 
Hofreiter, A. & Distribution and 


E. F. Rodriguez. 2005. 
of Bomarea (Alstroemeriaceae) in the 


phenology relict 
ae of northwestern Peru. Revista Peru. Biol. 12: 
275-282. 
Humboldt, A. Bo nd & C. S. Kunth. 


npla 1815. N 
Genera et oe Plantarum, Vol. 1. Paris. (Facsimile 
ed., Weinheim, J. Cramer, 1963). 


. 1825. Nova Genera et iE 

Plantarum, Vol. VII. Paris. (Facsimile ed., Weinheim, J. 
Cramer, oe 

Hutchinson, J. 1 The deae. of the Flowering Plants. 
IL Lo. Clarendon Press, Oxford. 

Jackson, B. D. 1985. Index aa An Enumeration of the 
Genera and Species of Flowering Plants. Clarendon Press, 
Oxford, United Kingdom 

Janzen, D. H. 1966. ds of sexual reproduction 

eason in Central America. 


la 
Klotzsch, Dr. 1840. Echeandia ua aus oo in 
Sudamerika. Allg. Gartenzeitung (Otto & Kd 8: ie 
nth, » E . Enumeratio Plantarum, Vol. 


Marais, W. sl Reilly 1978. po and its related 
genera (Liliaceae). Kew Bull. 32: 63 
c Barre, H. M. den 


International Code of Botanical Nomenclature (Vienna 
Code). Regnum Veg. 146. 
McVaugh, R. 1955. The American collections of Humboldt 
and Bonpland, as described in the Systema Vegetabilium 
of Roemer and Schultes. Taxon 4: 78-86. 


Volume 96, Number 2 
2009 


Cruden 
Synopsis of Echeandia 


Núñez, E. € G. Petersen G. 1970. El Peru en 1 la obra de 
E de Humboldt. Libreria Studium, Li 

Otto, 1840. Nachtrag zur obigen Aang Allg. 
[e i & binc 8: p 

Rankin Rodrigue: & W. Gre Humboldt, 

Willden E sud op a “Taon 50: 

1231-1247. 
athcke, x EPa n d e Pee Pa of 
terrestrial plants. Ann col. Sys 

Ravenna, P. 1985. New . or etu m d pec 

57: 327-328. 

——. 1987. Diamena and Diora, two new genera of 

Aviheticocene from Peru. tie Bat 92: 185-193. 

— ——. 1988. Six new species of Anthericum q 

TT from Bolivia and Peru - Oni ra 1(3): 24— 

Rusby, H. H. 1896. On the collections of Mr. Miguel Bang in 
Bolivia—Part III. Mem. Torr. E Club 6 

Ent A., M. O. Dillon. Sánchez, ^j pe & P. 

a. 1999. Diversidad ics del Norte de Perú. I. 
World Witte ae Lima. 

Sandwith, N. 925. XXXIV—Humboldt and Qus iy 
itinerary in d xem Bull. Misc. Inform. pp. 


. XXIV—Humboldt and Bonpland's itinerary 
in Ecuador and Peru. Bull. Misc. Inform. pp. 181-190. 
Schultes, J. n a H. Schultes. 1829. Systema Vegetabil- 

ium, Vol. 7, Part 1. J. G. Cottae, Stuttgart. 
Sprague, T. A. ion Humboldt and Bee itinerary in 


h 
e "Nova 
Verlag von J. Cramer, 


Tropical American Botany. 
eae et Species Plantarum.” 
xd 
Weatherby, C. A. 1910. A preliminary synopsis of the genus 
Echeandia. Daedalus 45: 387-394. 
es M. 2002. Observations on the ch i a 
Amotape-Huancabamba zone in northern Peru 


68: 38.54, 


andia (in boldface) 
and md names un ae listed ~ genus and subgenus. 


Anthericum L. 


A. ciliatum (Kunth) Spreng., nom. illeg. = E. ciliata 
A. glareosum Ravenna = E. ciliata 

A. herrerae Killip = E. herrerae 

A. humboldtii Hemsley = E. ciliata 

A. lehmannii Baker = E. E hin mannii 

A. weberbaueri Poelln. = E. weberbaueri 


I. Echeandia Ortega 
Ia. Echeandia subgenus Echeandia 
aequatoris Ravenna = E. lehm: 

E. ciliata (Kunth) Cruden 

E. denticulata Cruden 

E. herrerae (Killip) Cruden 

E. lehmannii (Baker) Marais & Reilly 
. weberbaueri (Poelln.) Cruden 

Ib. Echeandia subgenus Mseavea Cruden 
6. E. b 


annii 


SAA paces 


P. ciliatum Kunth = E. ciliata 


APPENDIX 2. Index to GSBIECATGS, Holotypes, lectotypes, m 
neotypes are in boldface. T 
of the species as it appears in 1 the text (see Appendix 1). 


Allen 2962 Ge e 1134 (2); Aristequieta 3284 (2 
DE & Ortega 821 (1). Brunel 204 (3), 62 
17116 (1). Cuatrecasas 6639 (2); 15356 (8). Da 
1595 E Dillon, Molau & Matekaitis 2862 (1). Ellenberg 
1195 (3). Fassett 25660 (2); Fendler 1549 (1); Franque- 

Franquemont 257B (3); Funcke & Schlim 674 "7 : 
Gehriger 295 (2); Gines 1857 (7); 1948 (7); Gróger & Ber, 
1064 (6); Grubb, Curry & Fernandez-Perez 117 (2), 122 o: 
Guite ipt is S 3), 2993b (3); Guite a Gutte 1916b (3); Gutte 


& Müller 9375 (1). Haught 4342 (7); Herrera 1012a (3), 
(3), $50 (3), 3067 (3), d (3). 3457 (3); Humbles 
3 (4); Humboldt $ o s.n. (1); Hunnewell 
m (3). Mesa-Bernal & Sm ut 410 o 


ae eis ill 
Ln 1122 (2). Kill 


7252 (3). Ochoa 275 (3). Pennell 1856 (2); Pittier 7996 (1), 
8900 (7); Purdie s.n. (1). D HE 15148 (1); Sánchez Vega. 
3266 (1), 4278 (1); Sánchez Vega, Flores & Levia 5687 (1); 

1 iz Vi Pom Vega 1846 (1); Saravia 
01031 (2), 01048 (2); Schiefer 881 (2); Schulz & Rodriguez 
681 (2); Schulz, Rodriguez & Sdnchez 105 (2); Si 


3 (7); Steyermark & Rabe 
& Manara 131761 (6); 
Stork & Horton 10722 (3). Tovar 178 (3), 4001 (5); Tracey 
(2); Triana s.n. (2); Tupayachi 900 (3). van der Hammen 

9 (2); Vargas 120 (3), 904 (3), 1788 e Vogel 10 (7). 
D Ma: 481 (5); Weigend & Weigend 2000/154 (3); 
Weigend, Dostert & Driefile 97/427 (1). 


THREE NEW SPECIES AND A 
NOMENCLATURAL SYNOPSIS OF 
URERA (URTICACEAE) FROM 
MESOAMERICA? 


Alexandre K. Monro? and Alexander Rodríguez? 


ABSTRACT 


Gaudich. is unique among Mesoamerican Urticaceae in having bright, fleshy fruits. Within Mesoamerica, ther 
significant confusion over the application of many names, ud U. corallina (Liebm.) Wedd., U. elata (Sw.) Griseb., n 


w species, U. fenestrata A. K. M 


a & Al. Rodr. (OM Rica), and U. lianoides A. K. Monri 
Costa Rica, Panama, Peru, and Bolivia), are descri 
inflo ces, stem, o nd 


orescen 
associations for 


the gen 


recogniz 
Gaudich. ex Wedd., U. caracasana (Jacq.) Griseb., U. mitis, 
iq., U. eggersii, U. subpeltata Miq., and U. subpeltata var. morifolia 


900 exsiccatae from 13 e ia. 


RESUMEN 


Mesoamérica existe sión significativa en la aplicació 


corallina (Liebm. ) Wedd. U. elata (Sw.) Griseb. y U. eggersii Hieron. Tres n 
& A]. Rodr. (Costa Ric) yt P ii we d K. Monro & Al. Rodr. 


. K. Monro 


ist 
words: Flora e TUCN Red list, 


onro & Al. s 


d and toed < on the bas is 


uras, Pp Costa Ric a, Panama, 


Mesoamerica, 


Ea 


Costa Rica and Panami a), U. guanacastensis A. 
o & Al. Rodr. (Mexico, Guatemala, Hon dues Nicaragua, 
w species 


the a of an 
a. In din a key he 10 species of Urera 


zed for Mesoamerica nomenda review is given in a U. mitis Vell) Miq. is lectotypified; U. baccifera (L.) 
and Urtica nitida Vell. are epitypified: 


; and Urera denticulata 
a Miq. are neotypified; and a list is provided of more than 


Urera Gaudich. es un género único entre las Urticáceas de Mesoamérica por presentar frutos lustrosos y suculentos. En 
la: 


n de muchos de los nombres, especialmente en las especies U. 


nuevas especies, U. fenesirata A. K. Monro & Al. 


Wm son d bre la ase de sus 


n en la cua U. mitis 


Urera, oe 


The genus Urera Gaudich. comprises shrubs, trees, 
and vines that occur most frequently in riparian and 
disturbed vegetation in Mesoamerica. Within the 
Urticaceae, Urera is characterized by fleshy fruits 
(formed by the inflation of the tepals), penicillate or 
capitate stigmas, glabrous pistillodes, and hairs with 
bulbous bases that are stinging in some species. Urera 
has a nearly pantropical distribution (Neotropics, 
Africa, Australasia, and the Pacific Islands) but is 
absent from Asia (pers. obs.). Currently, a single 
species, U. kaalae Wawra, has a Critically Endan- 


gered (CR) status, according to IUCN Red List criteria 
(IUCN, 2001). 

Within Mesoamerican Urticaceae, Urera is unique 
in having bright fleshy fruits. It is also characterized 
by stems that frequently release a watery latex when 
cut (a trait shared with Myriocarpa Benth.) and, in 
some species, stinging, bulbous hairs. It is for these 
stinging hairs that the genus is most widely known, 
hence the widespread local name of “chichicaste,” 
which is derived from the Nahuat word ‘ bed 
meaning ^to vibrate" (Bonilla A., pers. comm.). I 


1 We are grateful to Norman Robson (BM) for help with the Latin diagnoses; Charlie Jarvis (BM) and Sandra Knapp (BM) for 


manuscript; Roy Gereau (MO) and 


comments on the 


(OXF) for the illustrations; Victor Steinmann for sending n 


INB, K, LL, MEXU, MO, NY, P, PMA, TEX, and US for th 
was collected with support from Darwin Initiative pn 15/02 
ee of Botany, The Natural Histo 


Melanie Wilmot-Dear (K 
s of type material at IEB; and the curators at BM, C, F, GH, 
es of, and access to, collections. Some of the paratype material 


eum, Lon ds SW 


mary Wise 


7 5BD, United Kingdom. a.monro@nhm.a 


stituto Nacional de Biodiversidad, Apartado Postal 22-3100, Santo Domingo de Heredia, Costa Rica. aredig Ginbig: ac.cr. 


Ins 
doi: 10.3417/2006121 


ANN. Missouni Bor. Garp. 96: 268-285. PUBLISHED ON 7 JULY 2009. 


Volume 96, Number 2 
2009 


Monro & Rodríguez 
New Species and Synopsis of Urera 


also from these stinging hairs that Urera derives its 


LE 


limited economic and medical import 
fera (L.) Gaudich. ex Wedd. is u 

in Guatemala and Costa Rica (Standley & Steyermark, 
1952; Burger, 1977) and is also the focus of research 
for its anti-inflamm 


js 


used as “living lens 


atory properties (Badilla et al., 
1999). In addition, Urera includes species that are 
used to treat arthritis (Gonzáles, 1994; House et al., 

95), fever (House et al., , hemorrhage 
(Guánchez, 1999), erysipelas (Guánchez, 1999), and 
syphilis (House et al., 1995), and species that are of 
moderate importance as food for Lepidoptera (Janzen 
& Hallwachs, 2005) 

The genus Urera was first described by Gaudi- 
chaud-Beaupré in 1830, for which he proposed the 
subtribe 
Urtieaceae, which was later raised to tribal rank 


(Urereae) by Weddell (1856). This was subsequently 


Urerinnae (as Urereae) of the family 


renamed Urticeae by Friis (1989). Urera was 
expanded to include the monospecific genus Scepo- 
carpus Wedd. by Frii s (1980), a nd serm DNA 


sequence ca (trnL-F) suggests that Urera could be 
sister to Poikilospermum Zipp. ex Miq. within the 
Urticaceae (Monro, 2006). 
Urera has attracted no monographic attention since 
Weddell (1856, 1869), and, to date, subgeneric 
classification has been published (although Weddell 
id divide the species into unnamed groups according 
to inflorescence structure and distribution). System- 
oa work in NE has Mc resulted from localized 
floristic treatm and a ing Mesoamerica 
(Standley & enn 19 52, pe 1977; Pool, 
2001; Steinmann, 2005). 
here are currently 138 published species epithets 
for the genus (The International Plant Names pa 
08), of which 133 appear legitimate. Of t 
epithets, 16 have since da us to iade 
b., Dendrocnide Miq., 
Boehmeria Jacq., and Girardinia Gaudich. by subse- 


audich., Gyrotaenia Grise 


quent authors. Of the remaining 117 names, Friis 
(1989) estimates that there are ca. 35 good species 
and Pool (2001) estimates 35 to 7 

ithin Mesoamerica, there are 25 specific epithets 
for 10 recognized species of Urera, and there is 


significant confusion over the application of man 


combined with a lack of regional keys, the similarity 
of form and habit, and the extent of overlapping 
variation in characters between species, has made the 
determination of collections difficult. The result is 
that a significant proportion of herbarium material 
from Mesoamerica is misidentified. 


MATERIALS AND METHODS 


This work was undertaken as part of the revision of 
Mesoamerican Urticaceae for the Flora Mesoamer- 
icana project. The definition of Mesoamerica is as 
, 1994): a region 
bounded to its north by the Mexican siates of Yucatán, 


given in the Flora (Davidse et al. 


Campeche, Tabasco, Quintana Roo, and Chiapas, and 
to its sout y 
ddi material from Mesoamerica and areas 
adjacent to Mesoamerica (Oaxaca and Veracruz 
[Mexico], Greater Antilles, Colombia, Venezuela, 
Ecuador, and Peru) from BM, C, F, GH, INB, K, LL 
MEXU, MO, NY, P, PMA, TEX, and US was 
examined, resulting in 995 collections that were 
examined and determined, 850 from Mesoamerica. 
Determinations are listed in ris l. The 
nomenclatural revision was based on the examination 
of the original published Scri for al 
accepted names, as wel ial. 

The macro-morphological Pneu used most 
frequently by previous authors for the delimitation of 
species are le 
inflorescence morphology and structure, 
distribution and morphology, fruit size and color, 
In this study, 
d on stem morphology and 


cystolith shape, and stigma shape. 
mphasis was also pla 

habit, together with D observations in the field 
a thin, watery latex when 
no EN to this was found in the 
descriptions on collection labels, a number of 
collections had a dark stain on the rim of the cut 

. Thi taken 


ste s was s an indication that these 
ei tee may have released a wate 


ry latex, and it is 
for this reason that this was noted in p observations 
made and in the descriptions below. All material was 
examined using a stereomicroscope at X64 to X40! 
magnification, and up to 138 observations were made 
for each specimen sampled. These observations were 
then used as a guide to delimit taxa within a 
morphological species concept and in the preparation 
of a key to the identification of the Mesoamerican 

pecies. Once species were delimited, they were then 
matched to type material 


TAXONOMIC TREATMENT 


Urera Gaudich., Voy. Uranie 496. 1826 [1830]. 
TYPE (designated by Britton & Wilson, 1924: 
243): Urera baccifera (L.) per ex Wedd., 
Ann. Sci. Nat., Bot., sér. 3, 18: 1 1852 


Scepocarpus Wedd., Prodr. (DC.) 16(1): 98. 1869. TYPE: 
cepo edd. 


carpus manni W 


Annals of the 
Missouri Botanical Garden 


KEY ro THE MESOAMERICAN SPECIES OF URERA 


Distributions for each species are given to country level within Mesoamerica, with the exception of Mexico 
and species that are only known from a single country, in which case they are given to state level. Global 
distributions in the key are given to regional level (e.g., South America, West Indies), following Flora 
Mesoamericana protocols (Davidse et al., 1994). Some of the characters in the following keys, e.g., cystoliths, are 
only visible in dried material. Several types of hairs (i-e., bulbous, straight, and curved) are found on the surface 
of the leaves and young stems of Mesoamerican Urera. Here I define bulbous hairs as those with a swollen and 
inflated base giving the base a bulbous or bottle-shaped appearance. 


la. Leaves lobed; stem releasing white latex when cut 


Tr "c PT sat o Gio gas 7. U. laciniata 
lb. Leaves not lobed; stem e latex or not; wien released, latex gray o po 


Stem and leaves with spine 


arsely 0.5-1 cm; acheni 


co 2-3.2 mm; wow 
covering the basal 1/4 or les of achetie prior to the oa of tepals 


EN LU. Bom 
2b. Stem and leaves lacking spines; leaves dentate, crenate, or entire, when toothed the teeth spaced P 0.5 c 
apart; achene 0.75—2 mm; tepals covering 3/4 or more of the achene prior to the inflation of the tepals. 
3a. Adaxial surface of leaves sparsely pubescent or glabrous. 
4a. 


Leaf margin shallowly crenate, crenate-serrate, sinuate, or entire; hairs never urticating. 
Shrub or small tree; leaf-bearing section of ste er hollow. 
Ratio of leaf length to width less than 4:1; KAY leaf surface 
the secondary veins; mature fleshy fruits pink or orange to orange- 
6b. Ratio of leaf length to width greater than 4:1; E 
secondary veins; mature fleshy fruits red ....... llis. . U. guanacastensis 
5b. Shrub, lax shrub, or vine; leaf-bearing section of stem hollow, ca. 5-10 m 
Ta. Young shoots pubescent to densely pubescent, the hairs 
25 m immediately prior to anthesis .... llle > pandilla 
Tb. Targ shoots i pun or glabrous, s hairs ed present 0.125-4 m mm; 
staminate flowers c 2.5 mm immediately prior to anthesis ......... . U. fenestrata 
4b. Leaf margin prominenily serrate, crenate, or FA te; hairs hi. or not urticating. 
8a. Young stem glabrous or sparsely pubescent, the hairs = mm; stem coarsely sulcate, 
frequently fenestrate; stems, leaves, and petioles lacking bulbous hairs; 
to 110 mm; mature fleshy fruits red-pmk ..........0..00.0.020000004. U. fenestrata 
8b. Young stem sparsely to densely pubescent, the hairs > 0.5 mm; stem coarsely sulcate a never 
fenestrate; stems, leaves, and petioles with i cm bulbous hairs; staminate inflorescence to 
0 mm; malam fleshy fruits orange or oran, 
9a. 


d a 


in the axils of 
nee ares 4. U. glabriuscula 
abaxial leaf surface wit domatia in the axils of 

5... 


ia 
0.125-1 mm; staminate flowers c. 
8. U. 


Tertiary venation of abaxial leaf owe cream to pale green, noticeably paler a eae 
stipules forked or not forked; stem sparsely pubescent ...............- U. killipiana 
Tertiary venation of abaxial leaf surface darker or rarely paler than the lamina, oen aler 

than. the IE pale brown to orange-brown in color; stipules not forked; stem a 
pubescent 

10a. 


e$ 
E 


uum ovate or cordiform, never obovate or lanceolate; hairs urticating ....... 

———— rrr . U. caracasana 

10b. Leaves obovate, lanceolate, elliptic, or ovate, never cordiform; hairs urticating or 
not 


-— ——————À—— ———— tere . U. simplex 
en 
lla. Leaves bullate; staminate pns ME T at base for 40-80 mm; pistillate a 
unbranc &efor 21708 Trim ee soos ede Rete eee tecta dr deseas a Geor U. verrucosa. 


llb. ie not bullate; staminate and. pistillate peduncles unbranched at base for < 25 mm. 
1 


2a. Leaves ovate or cordiform, never obovate or lanceolate; hairs urticatin, caracasana 
imp 


ng 2. U. 
12b. Leaves FOR el elliptic, or ovate, never cordiform; hairs cazo or not ... 9. U. simplex 


1. Urera baccifera (L.) Gaudich. ex Wedd., Ann. Urera denticulata Miq., Fl. Bras. (Martius) 4: 192. 1853. 
T 


: A à : i YPE: Brazil. Minas Gerais: Viçosa, rd. E from Chacha 
Ser Ms Bot; 2 " a on 2o Sepp. valley, Fazenda da Creciuma, 10 May 1930, Y. Mexia 
e accifera p: e = i 1 4679 (neotype, designated here, BM!; isotype, MO not 

umier, Pl. Amer.: tab. 176 seen). 
ectotype, esignate e ooi E rera baccifera var. horrida (Kunt e rc us. Hist. 
"5 ype, designated by d i j ns Ui E horrida (Kunth) Wedd., Arch. Mus. Hi 
302]. EPITYPE: Jamaica. Stony Hill, 13 Mar. e a 1856. ie nee horrida Kunth, 
E p i en. Sp. [HBK] (quarto ed.) 2: 41. 1817. Urera 
1898, Fawcett 7177 (epitype, d ted h P 
à ere (epit pe» designated: here, horrida (Kunth) Miq., Fl. Bras. (Martius) 4: 192 
). X Colombia. “Santander, Magdalena prope A 

Urera armigera (C. Presl) Miq., Fl. Bras. (Martius) 4: 192. qns de Carare,” Humboldt Bonpland 1639 

1853. Basionym: Urtica armigera C. Presl, Bot. lec 


3 type, designated by de Rooij [1975: 302], P5. 

pou (C. UE 110. 1844 br TYPE: Brazil. Urtica ida Vell., Fl. Flumin. Icon. 10: t. 20. 1827 [1831]. 
“near Rio de Janeiro," J. Lho s.n. (lectotype, TYPE: FI. Flumin. Icon. 10: tab. 20. 1827. EPITYPE 

designated by de Rooij T1975: vie PR not seen). Brazil. Mato Grosso do Sul: Mpio. Paraguai, Serra des 


Volume 96, Number 2 
2009 


Monro & Rodríguez 
New Species and Synopsis of Urera 


Araras, Fazenda Currupira, 15°10’S, 5675" W, 24 Jan. 
1995, Dubs 1770 (epitype, ru. d here, K*5 
duplicates, E not seen, ESA n en). 
Urtica (ds ntata s Danske Vidensk. 
Sel Dos Tos ser. 5, 2: 296. 1851. TYPE: 
Cartago: “Irasú,” Jan. 1848, A. Orsted 
14283 ean nao by de Rooij [1975: 302], 
C». 


An epitype is selected for Urera baccifera because 


the type illustration, although accurate, is not 
sufficient to support the unambiguous fixing of the 
Fawcett 7177 was chosen 


because it includes flowering and vegetative material 


name to this species. 


that conforms to Linnaeus’ (1763) description “Urtica 
foliis alternis, cordatis" and is from the West Indies, 
as was the Plumier material that formed the basis of 
the pe illustration. 
it s selected for Urtica nitida because the 
type cones although accurate, is not sufficient to 
support the unambiguous fixing of the name to this 
species. Dubs 1770 was chosen because it includes 
flowering and vegetative material that conforms to 
Vellozo’s description and is from Brazil, as was 
the material that formed the basis of the type 
illustration. 
otype is selected for Urera denticulata 
because, although dc Rooij (1975: 302) cites Martius 
s.n. at M as lectotype for U. denticulata, no such 
collections could be traced either at M (M. Esser, 
ers. comm.) or BR (P. Stoffelen, pers. comm.). The 
collections database from the Fi Museum's De- 
partment of Botany (Field Museum of Natural History, 
2006), which includes the negatives of photographs of 
European type material taken by Macbride, does, 
however, include a negative (no. 8845) cited as being 
from a Martius collection that represents type material 
of U. denticulata at M. Such a negative cannot be 
considered isotype material, but it does indicate the 


lectotypification on this photograph or that he did 
see the type collection at M prior to 1975, and that 
this collection has since been lost. The specimen 
designated as neotype was selected because it 
includes good leaf, stem, and fruiting material and 
is from the same country and state as the material 
cited in Miquel's description. 


names. Bringa mosa (Panama: C. White- 
Pew Edd ota BM), chichicaste (Guatemala: J. 
A. Steyermark 38770, V; Honduras: tandley 
20526, F; El e P. C. Standley 22344, F; 
Nicaragua: P. C. Standley 11202, F), chichicaste 
cuyanigua (El Salvador: P. C. Standley 21880, GH, 
), chichicaste nigra (El Salvador: $. Calderón 1539, 
NY), cow itch (Belize: P. H. Gentle 2781, F, GH, NY, 


US) nigua, niguilla (El Salvador: P. C. Standley 
22394, GH), ortiga (Costa Rica: R. Anderson & S. Mori 
147, F; Panama: P. C. Standley 30536, US), ortiga de 
los caballos (Mexico: A. Schott 796, BM, F), rascate 
bien (Honduras: A. Molina R. 868, F, GH). 


Habitat and. distribution. 
forest, riparian vegetation, from sea level to 1400 m, 


Evergreen or seasonal 


Mexico to Panama, Colombia, Peru, Bolivia, Brazil, 
Paraguay. 


Comments. This species most closely resembles 


morphology as follows: 
bearing stem releasing gray but never white latex 
when cut, leaf margin coarsely dentate; (2) for 
laciniata, leaf-bearing stem releasing white latex 
when eut, leaf margin deeply lobed or laciniate. 


Selected. specimens examined. BELIZE. Cayo: Chiqui- 
bul, Las Cuevas, 16°43'N, 88759" W, A. js bad 671 ad 
BRH, MO). pan Cochabamba: 

Britton usby 1209 (A, BM, a wo o 
RICA. ae im Sea ii at bridge on rd. to Colonia 
Virgen del Socorro, 9 mi. S 
10^16'N, 
SALVADOR. La Libertad: Cordillera de Balsamó, San 
Julián rd., towards the Pacific, 13741'00"N, 89°38'32"W, A. 


Izabal: Montañas del Mico, 


Atlántida: 15°42 N, 86° 51' 
NICARAGUA. Rio San s Near Caño Ch 


Sector Grapanazu, limite Par 
ojas, K. Meza, J. Linean. E. Camavilca de M. Villaran 
1832 (BM, MO). 


2. bs caracasana (Jacq.) Griseb., Fl. Brit. W. I. 
154. 1859. Basionym: Urtica caracasana Jacq., 

Pl. Hort. Schoenbr. 3: 71. 1798. TYPE: Jacquin, 
Pl. Hort. Schoenbr. 3: t. 386. 1798. EPITYPE: 
Venezuela. Araugua: Tovar, 4-1855, A. 
Fendler 1275 (epitype, designated here, K?). 

Urera e (Poir) Gaudich., Voy. Uranie, Bot. 497 
1826 [1830]. s Urtica alceifolia Poir., a 
perd ppl. 4: 227. 1816. TYPE: French Guy. 
Cayenne: s.d., uae s.n. (lectotype, designated ja " 
Rooij [1975: 304], n 

Urtica A Kunth, Nov. Gen. Sp. [HBK] o v 2: 

as d iE ” TYPE: Colom 

ipa . Humboldt & Bonpland rox ie 
type, designated by de Rooij [1975: 306], P5. 

Urera jacquinii var. ge lía (Kunth) Wedd., Arch. Mu 
Hi: at. 9: 145. 1856. Basionym: Urtica ulmifolia 


Annals of the 
Missouri Botanical Garden 


Kunth, Nov. Gen. Sp. [HBK] (quarto ed.) 2: 141. 1817. 
TYPE: Colombia. Bolívar, Humboldt & Bonpland 1427 
ctotype, designated by de Rooij [1975: 306], P5. 
ne milis (Vell.) Miq., Fl. Bras. (Martius) 4(1): 191. 1853. 
asionym: Urtica mitis Vell., Fl. Flumin. 10: tab. 19. 
1827 [1831]. Urera caracasana var. mitis (Vell.) Wedd., 
Prodr. (DC.) 16: 90. 1869. TYPE: Fl. Flumin. Icon. 10: 
tab. 19 (lectotype, Fili i M iid tab. 195. EPI- 
TYPE: Brazil. Amazonas: Marapata, Municipality of 
ary, 25 May 1933, B. Eolo 4568 (epitype, 
designated here, BM?). 
Urera corallina (Liebm.) Wedd., Prodr. (DC.) 16: 90. 1869. 
sionym: Urtica corallina Liebm., Kongel. Danske 


Vidensk. Selsk. Skr., sre s Math. Afd., sex. 5, 
2: 295. 1851. TYPE: Costa Rica. Alajuela: Monte 
uacate, Orsted 14282 (lecto B. M usn by de 


A 
Rooij [1975: 304], further designation here, C es 
“26/2003/2”!). 

Urera capitata Wedd., Ann. Sci. Nat., Bot., 
1852. TYPE: Bolivia. Yungas: Dec. 
4317 (holotype, P!). 

Urera subpeltata Miq., Fl. Bras. (Martius) 4(1): 189, pl- 66. 

53. Urera jacquinii var. subpeltata (Miq.) Wedd., 
Arch. Mus. Hist. Nat. 9: 145. 1856. Urera caracasana 
var. subpeltata (Miq.) Wedd., Prodr. (DC.) 16: 90. 1869. 
TY Brazil. Bahia: 1839, p indi kd (neotype, 
desi gated here, BM!; isotype, G n 

Urera cle var. CE Miq., Fl. zem lari) m js 
Brazil. Para: Santarem, July 1850, R. 

ruce s.n. i (nsum. c here, BM). 

Urera jacquinii var. miquelii Wedd., Arch. Mus. Hist. Nat. 9: 

. 1856. i caracasana var. miquelii Wedd., 
Prodr. (DC.) 16: 90. 1869. TYPE. Peru. Gay s.n. 

olotype, P not se 

Urera acuminata Miq., " Bias: (Martius) 4(1): 190. 1853, 

om. illeg. non Urera acuminata (Poir.) Gaudich. 


ser. 3, 18: 2 
1846, "ud 


Jacquin’s original description of Urera caracasana 
is based solely on staminate material. Original 
material for the name has not been located at BM or 
LINN, and 


probably been destroyed (material could not be traced 


any material that may have been at W has 
at W). Based on the type illustration, it is not possible 
to distinguish between U. caracasana and U. corallina 
(Liebm.) Wedd. on sterile characters alone. Materia 
examined that had been determined as U. corallina 
(including the holotype) by Weddel and as U. 
caracasana did not uncover any significant morpho- 
logical differences. An epitype is selected because the 
type wie although accurate, is not sufficient to 

support the unambiguous application of the name to 
this species. 

A lectotype is designated for Urtica mitis Vell. 
because Vellozo (1827) does not refer to the plate as 
s the ty 


e material as his description predate ype 
sidered original type 


X 

concept. Tabulae 19 can be con 
material. An epitype is selected for Urera mitis 
because the type illustration, although accurate, is 
not sufficient to support the unambiguous fixing of the 
name to this species. The epitype was selected 


because it includes both leaves and inflorescences 


and is from the same country and region as the 
holotype. 

A neotype is selected for Urera subpeltata var 
morifolia because although de Rooij (1975: 306) cites 
Spruce 633 (M) as lectotype, no Spruce collection fitting 
this description could be traced either at M (M. Esser, 
ers. comm.) or BR (P. Stoffelen, pers. comm.). Miquel 
1853: 190), in his original PW cites two 


— d 


collections from the Brazilian Amazon, “Martius in 


silvis amazonicis, ad Barra do Río Negro" and "Spruce 
ad Santarem," neither of which could be traced either 
at M (M. Esser, pers. comm.) or BR (P. Stoffelen, pers. 
comm.). A Spruce collection was, however, located at 
BM with the annotation, *In vicinibus Santarem, Prov. 

Para.” This is 


collection is from the same collector, country, and state 


selected as neotype on the basis that the 


and includes good leaf, pistillate, and immature fruit 
xe 

eotype is selected for Urera subpeltata because 
i de Rooij (1975: 306) cites Martius s.n. at 
as lectotype for U. denticulata, no such collections 
could be traced either at M (M. Esser, pers. comm.) or 
BR (P. Stoffelen, pers. comm.). The specimen desig- 
nated as neotype was selected because it includes 
good leaf, stem, and pistillate and staminate material 
and is from the same country and state as the material 
cited in Miquel's 1853 description. 

De Rooij (1973: 306) cites Makin s.n. (not Ee 
as type (holotype?) for Urera jacquinii var. mique 
Weddell (1856), however, cites only Claude p 
collection from Peru s description and this is 
maintained as hope 


names.  Chichicaste a Salvador: E. 

sandra 1854, BM; Nicaragua: P. C. Standley 
2, F), chichicaste blanco picante (El Salvador: 

A a & R. Chinchilla 504, MO), chichicaste 
cujanigua de altura (El Salvador: E. Sandoval & R. 
Chinchilla 1182, MO), chichicaste rojo picapica (El 
MO), pan 
caliente (Honduras n et al. ' 3055, BM), 
migirillo (Costa Rica: A. Sanchez 10, F), miguito 
(Costa Rica: J. A. Echeverria C. 268, V), ortiga (Costa 
Rica: i A. n C. 268, F), pan caliente (El 
Salvador van Due M 13, P), i 
(Mexico: A ee G. 8945, 


Habitat Cloud forest, shade 
coffee i pine forest, Quercus L.—Liquidambar 
L.—Pinus L. forest, from sea level to 2 m. Mexico 
(Chiapas), Belize, Guatemala, Honduras, El Salvador, 
Nicaragua, Costa Rica, Panama, Colombia, Vene- 


Salvador: Martinez s.n. 
C. Nelso 


zulsimtezla 


and distribution. 


zuela, Ecuador, Peru, Bolivia, Brazil, and Argentina. 
Comments. Herbarium material of this species is 
commonly determined as Urera corallina. This species 


Volume 96, Number 2 
2009 


Monro & Rodríguez 
New Species and Synopsis of Urera 


most closely resembles U. verrucosa (Liebm.) V. W. 
Steinm. The two species can be distinguished from 
each other based on leaf texture and inflorescence 
peduncle size as follows: (1) for U. caracasana, leaves 
chartaceous, staminate peduncle branched to base or 
unbranched at base for 2-13 mm, pistillate peduncle 
branched to base or unbranched at base for 2— 
(2 for U. bullate, 
peduncle unbranched at base for 40—80 mm and 
densely pubescent, pistillate peduncle unbranched at 


mm; 


verrucosa, leaves staminate 


base for 


Selected specimens examined. ARGENTINA. Jujuy: San 
Pedro, Sierra Sta. Barbara, S. Venturi 9633 (K). BELIZE. 

ayo: Caracol Maya ruins, 14km W of Las Cuevas, 
16745'N, 8807 W, T. Hawkins 1132 (BM, m BOLIVIA. 
Cochabamba: Carrasco, ca. 11 km below Sehuencas, J. R. 
I. Wood 10275 (BM, K). BRAZIL. Roraima: IIha de Maracá, 
SEMA Ecological COE S of rar dE de Fumaga, W. 
ud 546 (K). C RICA. P. : W. Haber & 

uchowski 9294 o ECUADOR. S. ote, Bees 15611 

d. e hincha: new Alluriquín-Quito rd., Km 4, P. J. M. 
Maas & L. Cobb 4770 (K). EL SALVADOR. La Libertad: 

pi na,” P. 
s.n. *WB-01217" (BM, LAGU, 
tiapa: D. Dunn et al. 23234 MO). HONDURAS. Ocotepe- 
que: Alrededores de Belén Gualcho, C. H. Nelso 
Romero, A. Rubio & M. Pereira 3955 (BM, MO). RUD. 
Chiapas: Rancho Puy Ukum, sobre la carr. a 2 km 
m No de Bochil, A Méndez G. 8945 (BM, ka 
NICA UA. D rn “El Recreo,” 4 km al N de 


Along Continental 
Divide on trail in Zona Palo Seco, 08°47’ N, 82%13'W, S. 
Knapp & J. Mallet 9169 (BM, MO, PMA, SCZ). PERU. 
Loreto: Rio Itaya, T. B. ee 18829 (K). VENEZUELA. 
Falcón: Sierra de San Luis, uente de Jobo entre 
Curimagua y San Luis, J. A. Seana 99249 (K). 


3. Urera fenestrata A. K. Monro & Al. Rodr., sp. 
Monte- 
verde, Cordillera de Tilarán, Pacific slope, above 
Quebrada Cuecha, 1540-1600 m, 6 May 1976, 
V. J. Dryer 179 (holotype, F!; isotype, CR!). 
Figure 1A—E. 


ta Rica. Guanacaste: 


Species nova Urerae caracasanae (Jacq.) Griseb. similis, 
sed ab ea floribus staminatis tetrapartitis, mot petiolisque 
inosis ramulis saepe fenestra 


lateralibus per 2/3 longitudinis visibilibus differt. 


tis atque foliorum nervis 


rub, lax shrub, vine, or small slender tree, 
dioecious (?). Main stems arching 2-10 m, stems not 
releasing white latex when cut; without spines; young 
shoots Joule or sparsely pubescent, the hairs 
0.125-0.2 


5mm, erect, straight to weakly curved; 
internodes of leaf-bearing sections of stem 9-30 X 4— 
10 mm, 


hollow, and ieee fenestrate when = 5 mm diam 


pale brown to red-brown, coarsely sulcate, 


Stipules 5 
petioles ares X 1.5-1.75 mm, glabrous or sparsely 


m, lanceolate, not forked, pubescent; 


pubescent, the hairs 0.25-0.5 mm; leaf blade 103— 
320 X 644 


elliptic, chartaceous; adaxial surface sparsely pubes- 


220 mm, ovate, broad-ovate, or broad- 


cent or glabrous, the hairs 0.375-0.675 mm, weakly 
appressed, weakly curved or r crooked; the cystoliths 


liths oblong, parallel to veins; primary veins 3, 
primary to quinternary or hexternary veins visible to 
the naked eye, the lateral primary veins visible for 2/3 
of the leaf length; domatia not present in the axils of 
the secondary veins; base cordate or obtuse; margins 
entire or weakly crenate to serrate; apex subcuspidate 
to cuspidate. Peduncular bracts 2-3 mm; bracteoles 
75-0.5 mm; staminate inflorescences 1 to 16 per 
stem, peduncle branched to base or unbranched from 
base for 2-21 mm, pubescent, the hairs to 0.125 mm, 
the whole inflorescence 15-110 mm, bearing 160 to 
350 flowers in a symmetrical cyme with 3 or 4 orders 
of dichotomous branching; the flowers borne 
clusters of 10 to 35, pedicellate to subsessile, he 
pedicels when present to 0 Pistillate 
inflorescences 1 to 4 per stem, the peduncle branched 
to base or unbranched at base for 2-26 mm, pub- 
escent, the hairs ca. 0.125 mm, erect, straight, the 
whole inflorescence 10-80 mm ( m in fruit), 
broader than long (or as long as broad), bearing 224: to 
750 flowers in a symmetrical cyme with 5 orders of 
dichotomous branching, the flowers borne in clusters 
of 3 or 4, pedicellate to subsessile, the pedicels when 
present to 0.25 X 0.125 mm, glabrous. Staminate 
flowers 2-2.5 X 1.25-1.5 mm immediately prior to 
anthesis; tepals 4, ca. 3 mm; stamens 2.5-3 m 
pistillode ca. 0. 
late flowers ca. 


mm diam., 


0.5-0.675 mm, asymmetrically ovate; the dorsal tepal 
0.50-0.675 mm, 
thickening, the ventral tepal ca. 0.375 mm 


ovate, with a subapical dorsal 
, ovate; 
stigma penicillate, erect. Fruit prior to inflation of 
tepals ca. 1.25 X 1 mm; basal 3/4 of achene obscured 
T e the laterals ca. 1 mm, the dorsal tepal 

67 the ventral ca. mm; achene 
oo. X 0.75 mm, imc elliptic, keel- 
shaped, surface smooth. Fruit when ripe with tepals 
inflated and berry-like, ca. 1.5 X 1 mm, red-pink, 


red-violet, pink, or orange when fresh. 


Ortíga (A. Smith 100, F, NY). 


Local names. 


Habitat and. distribution. 
in premontane, montane, and cloud forest, frequently in 
disturbed shaded ] small rivers, at 
800-3000 m. It oceurs on the Pacific and the Caribbean 
slopes of the Cordillera de Tilarán in Costa Rica and the 


Urera fenestrata is found 


Cordillera de Talamanca in Costa Rica and Panama. 


Annals o 
Missouri Botanical Garden 


su 
[NY]. F. 


Kemnay Aite, Pd F m — 
25mm 25 mm 
—A. Habit, with staminate inflorescences. —B. Leaf abaxial 
tillate inflorescence. —E. Pistillate flower. (A, B: i: ryer r 179 [F]; C-E: Kirkbride & Duke 770 
—G. Leaf adaxial surface. 


Figure 1 o E. Urera dee A. K. Monro & Al. Rodr. 
u^ m. —D. c 
acastensis a K. Monro & Al. Rodr. —F. Habit 
—H. IS cee —I. Immature fruit. TE —I: Delgado 24 [ [MO]. 


ace. — 


Volume 96, Number 2 
2009 


Monro & Rodríguez 
New Species and Synopsis of Urera 


IUCN Red List category. Conservation for Urera 
Least e e 
Red List criteria (IUCN, 2001), 
owing to the fact that the species has been E 
28 times in several localities in Costa Rica and 
Panama. 


fenestrata must be considered as 
according to IUCN 


Etymology. From the Latin “fenestratus,” mean- 
ing “windowed,” referring to the stems which, when 
greater than become 


or equal to 5 mm diam., 


characteristically windowed. 


Discussion. This species is characterized by the 
gnarled, windowed, leaf-bearing stem sections and the 
ovate, entire or subentire to weakly indented leaves. 
Urera fenestrata has most frequently been determined 
as U. elata or U. caracasana and may be distinguished 
from them as follows: (1) for U. caracasana, stems 
when greater than or equal to 5 mm ollow, but 
never fenestrate, lacking spines; nd lacking 
spines; the leaf margins serrate, serrate-dentate, or 
crenate-serrate to dentate; staminate flowers (4)5- 
parted; (2) for U. elata, us when greater than or 
equal to 5 mm .h , but never Ded 
small spines present; Uk ih small spines; the 
leaf margins serrate, serrate-dentate, or dentate; 
staminate flowers 4-parted; (3) for U. fenestrata, stems 
when greater than or equal to 5 mm 

fenestrate with prominent ca. 5 mm windows, lacking 
spines; petioles lacking small spines; the leaf margins 


entire or weakly crenate to serrate; staminate flowers 
arted 


Both authors have observed ants of the genera 
Crematogaster Lund. and Pheidole Westwood (For- 
micidae, Myrmicinae) un E the hollow stems of 
. Collections of these ant 
species are available 1 at oe Instituto Nacional de 
Biodiversidad entomology collections. 


this species in the 


M COSTA RICA. Alajuela: La Palma de San 
n, A. M. Brenes 136 (5544) F, NY); La Palma de 
208) (CR, D rd. to ae 
o, T. B. Cro 
43488 (MEXU, MO): Cordillera de and Y J. Dryer 44 
n, V. J. Dryer 179 (CR), 233 (CR), 


Cordillera de Tilarán, 100 m NW of Station, D. Penneys 19 
(INB); Vara Blanca a Sarapiqui, N slope of Central 
Cordillera, betw. Pods Barba volcanoes, A. F. Skutch 
3602 (NY); Cordillera ae near San Juan de Laja, ca. 
15 km N of Zarcero, L. O. Williams, A. Molina R., T. P. 
Williams & D. N. Gibson 28972 (F, NY); Reserva Forestal de 
i Herrera C., 

0); above Quaker 
settlement at ` Monteverde, J: rb & K. Utley 2552 (MO). 
Cartago: Cantón de Paraíso, Reserva Forestal Rio Macho, 
Río Pejivalle, E. Alfaro 1799 (NB); Tapantí Hydroelectric 
eee pe Rio Grande is ae T. Croat 36151 (CR); 
Orosi, g von Rio Mac westlicher Richtung zum 
RR E " Dübbeler 5162 (n Tapani LC.E. Reservation, 


= 


9 mi. from bridge over Rio Grande de Orosi, A. Geniry 2043 
(MO); a la vera del Río Turrialba, camino entre Pacayas y 
Santa Cruz de Turrialba, J. Gómez-L. 8078 (CR); Tausito, S of 
Tres Ríos, cerros de la Carpintera, R. Khan, M. Tebbs & A. R. 
Vickery 1318 (BM); ca. S of Tapanti, along Río Grande de 
Orosí, R. W. Leni 806 (CR); Tausito, R. W. Leni 3803 (F, NY); 
G. Mora 543 


da casa a diens. V. 


Turr ialba, camino a a Man avia, antes del Río aa L 
F, NY). 1 Heredia: Parque Nac. 
o, Zurqui station, just above Los Guarumos 
trail, B. Boyle 2450 (BM, CR); Parque Nac. Braulio Canillo, 
sector Zurquí, 7. A. Chacón 1881 (CR); La Palma, Río ra 
de "P qs Braulio Carrillo Natl. Park, N. Garwood, M. 
Gibby, R. J. Hampshire . Humphries 386 (BM, CR); 
Due atia el Mochote, borde de Corro Zurqui, L. D. Gómez, I. 
A. Chacón, 


o 
3 


(INB); S slope of Volcán Barva, W. H. Hatheway 1354 (CR); 
Parque Nac. Braulio Carrillo, Estación Barva, era, 258 
INB), G. Rivera 259 (INB); Rio Porrosatí, 50 m N de e parada 
de buses de Paso Llano, G. Rivera 432 (INB); S slopes of 
Volcán Barba betw. Río Cirueles & Sacramento, J. Utley & 
K. Utley 2319 (NY); Parque Nac. Braulio Carrillo, sendero 
hacia Quebrada Zrquí, Zamora 625 (CR). Limón: Cantón de 
Talamanca, Cordillera Talamanca, camp Río Lori, J. Bitiner 
1882 (INB), 1894 (INB). Puntarenas: Cantón de Buenos 
Aires, Estación Tres Colinas, finca Benito Acufia, E. Alfaro 
745 (INB); Cantón de Coto Brus, Zona Protectora Las Tablas, 
Las Tablas, sendero Echandi, E. Alfaro 1260 (INB); Cantón 
de i ui Brus, Zona Protectora Las Tablas, I Camp. ACLA, 
camino a Cerro echando, Æ. ~ 1569 (CR) near the 
Continental Py ca. 2-5 k E of b M W. c. 
urger L. Geniry 8639 (CR), e (CR). Pun 
Me border: 2-5 km E of Monteverde, " 
R. L. Liesner 8639 (V); Reserva Santa Elena, 


— 


= 


Coronado, Parque Nac. 
Estación Zurquí, L. Acosta 480 (CR, INB), ja Acosta 486 (CR, 

INB) valley of Río a (below La Palma), NE of San 
Jerónimo, W. C. Burger de R. G. Stolze 4912 (CR, F); Río 
Claro valley (Río La He drainage) below La Palma, NE 
of San Jerónimo, Burger & R. G. Stolze 7646 (F); 
bosquecillos residuales entre las Nubes y Cascajal de 


Coronado, J. Gómez-Laurito 5535 


UNAM A. Rodríg 
Palma, cà ca. de 6 km N 


quebrada Hig & Gutierr 
770 (NY); DUM slopes of Cerro Fábrega at foot ui "Falso 


Annals of the 
Missouri Botanical Garden 


Fábrega,” A. K. Monro & S. Cafferty 4855 (BM, INB, MO, 
shee 4856 (BM, INB, MO, a” Chiriquí: vic. of Gualaca 
. 10.7 mi. from Planes de Hornito, La Fortuna rd. to dam 

e, T. Antonio 5111 (BM); jane Bambit & Cerro Punta, T. 

B = 10595 (MO); betw. Palo Alto & top of ridge (divide) 
near Cerro Pate Macho above Río i Alto, W. D'Arcy et al. 
12658 (BM); Boquete, Bajo Chorro . Davidson 349 (E); 
near Fortuna Dam. . Hampshire Lo e Whitefoord 214 (BM), 
203 (BM); lower reaches of trail to Cerro Pando, A. K. Monro, 
Mallet, I. Mallet & V. Mallet 3520 (BM, 


= 


h of iue 
Chiquero, along Río Caldera, R. E. Woodson, P. H. Allen & R. 
J. Seibert 1005 (MO, NY). 


4. Urera dubie. V. W. Steinm., Acta Bot. 

. 2005. TYPE: Mexico. Veracruz: 

Mpio. San e Tuxtla, Voleán San Martín, 

1300 m, 2 Apr. 1985, Cedillo T. 3175 (holotype, 
IEB!; isotype, MEXU not seen). 


Local names. 


G. 9054, BM) 


K'anal zulzimtez (Mexico: A. Méndez 


Habitat and distribution. 
ous and evergreen forest, 
Rich. forest from 100-2800 


(Chiapas, Oaxaca, Tabasco, Veracruz), Guatemala. 


Montane forest, decidu- 
Liquidambar-Taxodium 
m. El Salvador, Mexico 


Comments. Urera glabriuscula is most similar to 
U. lianoides in its glabrous or sparsely pubescent 


d. The two 


species can be distinguished from each other based on 


leaves that are entire or discretely divide 
habit, stem, stipule, staminate flower, and achene 
morphology as follows: (1) for U. glabriuscula, shrub 
or small tree, leaf-bearing section of stem never 
hollow, with internodes less than 20 mm; stipule apex 
minutely forked; pistillode 0.75 mm diam.; achene 
for U. lianoides, 


scrambling shrub, leaf-bearing section of stem hollow 


surface verrucate; (2) vine or 
at ca. 5-10 mm diam. with internodes greater than 
20 mm; stipule apex not forked; pistillode 0.375 mm 
diam.; achene surface smooth. 


meee specimens examined. GUATEMALA. Quezalte- 
na ¿LAS 


eapa, Cerro 
Universidad 
alo, 


A. 
Peregrino 34699 (BM, MEXU). Veraeruz: A. Gentry, E. Lott 
& UNAM tropical botany class 32227 (BM, MO). 


5. Urera guanacastensis A. K. Monro & Al. Rodr., 
sp. nov. TYPE: Costa Rica. Guanacaste: Parque 
Nacional Guanacaste, Cantón de Liberia, Esta- 


ción Cacao, 10%55'45"N, 85728'15"W, 1100 m, 


3 June 1990, R. Delgado 24 (holotype, INB!; 

isotype, MO!). Figure 1F-I. 

ecies nova Urerae simplici Wedd. similis, sed ab e 
TUN glabris, foliis angustioribus supra parce Disce 
bus vel galbris subtus domatiis praeditis, pedicellis florum 
staminatorum breviorbus atque fructu rubro differt. 


Shrub to small tree, dioecious (?), main stems 
arching, 2—4 m, not releasing white latex, without 
spines; young shoots glabrous; internodes of leaf- 
bearing sections of stem 3-10 X ca. 2 mm, pale grey- 
brown, not hollow, lacking a dark stain on the cut 
portion of the stem. Stipules 5-12 mm, narrowly 
lanceolate, forked, sparsely pubescent; petioles 6—90 
X ca. 0.5 mm, glabrous to sparsely pubescent toward 
leaf base, the hairs 0.25-0.375 mm, strongly ap- 
pressed, straight; leaf blade 55-230 X 8-55 mm, 
narrowly lanceolate, lanceolate to oblanceolate, 
chartaceous to subcoriaceous; adaxial surface sparsely 
pubescent to glabrous, the hairs most frequent toward 
the leaf base, hairs 0.375-0.5 mm, weakly appressed, 
straight; the cystoliths punctiform, inflated, densely 
scattered; abaxial surface glabrous; the cystoliths 
punctiform to oblong, occasionally appearing inflated, 
scattered and parallel to the veins; primary veins 3, 
primary to tertiary and occasionally quarternary veins 
visible to the naked eye, the lateral primary vein 
finer than midrib and visible for 1/3 of leaf length; 
domatia present in the axils of the secondary veins, 
composed of hairs; base obtuse; margins asymmetri- 
cally discretely crenate to serrate; apex pungent to 
subacuminate. Peduncular bracts 1-2 mm; bracteoles 
—0.5 mm, staminate inflorescences ca. 15 per 
stem, peduncle unbranched from base for 6—9 mm, 
pubescent, the hairs to 0.25 mm, the whole inflores- 
cence 5-40 mm, bearing ca. 110 flowers in a weakly 
asymmetrical to symmetrical cyme with 4 or 5 orders 
of dichotomous branching; the flowers subsessile, 
borne in clusters of 5 to 7; pistillate inflorescences 2 
to 21 per stem (4 to 7 in fruit), the peduncle branched 
(6-17 mm 
in fruit), sparsely pubescent, the hairs ca. 0.1 mm, 


to base or unbranched at base for 2-12 mm 


erect, d the whole inflorescence 5-40 mm (23— 
mm in fruit), longer than broad (or as long as 
broad), Ned 30 to 250 flowers (59 to 67 in fruit) in 
a symmetrical cyme with 4 or 5 orders of dichotomous 
branching, flowers pedicellate, borne i in clusters of 3, 
rarely 2 to 4, the pedi 
0.175 mm, gibus. Staminate flowers 1— 
1.50-1.75 mm immediately prior to anthesis; tepals 
2mm; stamens and pistillode not seen. 
Pistillate flowers 0.4-0.6 X 0.25-0.6 mm, lateral 
tepals ca. X 0.25 mm, ovate; dorsal tepal ca. 0.4 
x 0.2 the ventral tepal ca. 0.4 X 
0.25 mm, ovate; stigma penicillate, erect. Fruit prior 


5 mm, ovate, 


Volume 96, Number 2 
2009 


Monro & Rodríguez 
New Species and Synopsis of Urera 


to inflation of tepals with achene entirely obscured by 
tepals, the laterals ca. 1.75 mm, the dorsal ca. 1 mm, 
mm, achene 1-1.375 X 0.75- 

m, elliptic, surface smooth, fruit when ripe with 
tepals inflated and berry-like, 1-1.25 X ca. 2 mm, 
red. 


the ventral ca. 0.675 


Habitat and. distribution. 
found in montane and cloud forest, disturbed and 
at 820-1350 m, known only 


illera de Guanacaste in 


Urera guanacastensis 1s 


undisturbed forest, 
from the Cordi 
de Liberia, Guanacaste Conservation Area in Costa 


Rica. 


IUCN Red List category. Conservation for Urera 
guanacastensis is considered as Near Threatened (NT) 
according to IUCN Red List criteria (IUCN, 2001). 
This is based on an evaluation of the potential 
threat to its 


distribution of anacastensis, the t 


& 
itat within ihat area, and t mber of existing 
records for the species. ae ee from the 12 
collection localities of U. guanacastensis plotted on 
oogle Earth, this iud pol gial distribution 
covers an area of ca. 680 km?. All 12 records are 
from forest localities (Google Earth, 2008; collection 
label | that form Bon of Costa Rica’s protected 
ork. Currently, 50% of the potential 
ute is did (Google Earth, 2008) and no 
records exist from deforested localities. We therefore 
assume that ca. 50% of the original population has 
been lost and that the future of this species is 
dependent on the maintenance of Costa Rica's 
Protected Areas Network. 


mology. Urera guanacastensis is named after 


M e Conservation Área, where all known 


ee of this species have been collected. 


Discussion. Urera guanacastensis is characterized 
by narrowly lanceolate, lanceolate, or oblanceolate, 
glabrous to sparsely pubescent leaves that frequently 
possess domatia on the abaxial surface composed of a 
cluster of hairs in the axils of the secondary veins, and 
by the red color of the mature fleshy fruits (not unique 
within the genus). It is most likely to be confused with 
U. simplex Wedd., from which it may be distinguished 


pedicels and color of the fruit as follows: (1) for U. 
guanacastensis, young stems glabrous; leaves 8— 
38 mm wide, adaxial surface very sparsely pubescent 
to glabrous, abaxial surface with domatia in the axils 
of the secondary veins; pedicels of staminate flowers 
less than 0.5 mm; mature, fleshy fruit red; (2) for U. 
simplex, young stems pubescent, frequently densely 
so; leaves wide, adaxial surface pubes- 
cent, frequently densely so, abaxial surface lacking 


domatia; pedicels of staminate flowers greater than 
1 mm; mature, fleshy fruit orange. 
e RICA. Guanacaste: 


Paratypes. Cantén de 


Liberia, Parque 


E. Bello 2237 (INB); Estación Cacao, C. Chávez 95 (NB): 
Parque Nac. Rincón de La Vieja, Cordillera de Guanacaste, 
Estación Las Pailas, R. Espinoza 771 (INB}; Estación Cacao, 

sendero a casa de Fran, B. Gamboa 45 (CR, INB); Estación 
Naranjos, W, B. 
Estación Mengo, sendero el Potrero, aes sur, II INBio 183 


(BM, INB); es Cacao, cerro Cacao, sendero a Casa 
Fran, W, A. Mora 37 (INB); Estaci ión iacu. cerro Cacao, M. 
ibus) 60 (IND) Estación Cacao, Cerro Cacao, 


Quesada 275 (INB); Sector Las Pailas, Río Colorado, Asta 
Arriba, G. Rivera 651 (INB); Cordillera de Guanacaste, 
sendero a laguna Santa María, W, G. Rivera n (INB). 


6. Urera jus ge E & Steyerm., Fieldiana, 
Bot. 24: 427. 1952. TYPE: Guatemala. Quetzal- 
tenango: Voleán ma 1700 m, 8 1934, A. 
F. Skutch 982 (holotype, F!; isotype, “CHD, 


Local names. Chichicaste (Guatemala: P. C. 
ead 76980, F), chichicaste común (Guatemala: 
. Standley 64714, F), nigüita (Guatemala: P. C. 
odia 75549, F 
Habitat and distribution. 
montane forest, riversides, from 900—300 
Oaxa 


Hidalgo, Queretaro, 
Guatemala, and El Salvador. 


aoe vegetation, 
O m, Mexico 


— 


Tabasco, Chiapas), 


omments. Urera killipiana is most frequently 
determined as U. eggersii (= U. simplex) or U. 
caracasana. These species can be distinguished from 
each other based on venation, stipule morphology, and 
stem indumentum as follows: (1) for U. oo. 
young stems sparsely pubescent, stipule apex forke 
not forked, the secondary veins of the abaxial leaf 
surface without domatia in their axils, tertiary venation 
of abaxial leaf surface noticeably paler than the lamina; 

f 


— 
N 
E 


U. caracasana, young stems densely pubescent, 
stipule apex not forked; the secondary veins of the 
abaxial leaf surface frequently with domatia (flap-like 
or tufts of hairs) in their axils, tertiary venation of 
abaxial leaf surface darker or of the same color tone as 
the lamina; (3) for U. simplex, young stems densely 
pubescent, stipule apex not forked; the secondary veins 
of the abaxial leaf surface without domatia in their 
axils, tertiary venation of abaxial leaf surface darker or 
of the same color tone as the lamina. 


Selected specimens examined. EL SALVADOR. San 
Salvador: S. Calderon 727 (GH, NY). GUATEMALA. San 
Marcos: San Marcos, Finca Armenia, Dwyer 15338 (MO). 
MEXICO. Chiapas: Onaracoaatla de Espinosa, T. B. Croat 


Annals of the 
Missouri Botanical Garden 


40599 (MO). Hidalgo: Puerto Obscuro near Km 328 on hwy. 

betw. Santa Ana € Chapulhuacán, Jacala Distr., H. E. Moo 

M. Merello 15457 (MO). 

Ee 557 (BM). Tabasco: C. L. Gilly & E. “Hernandez X 
20 (GH, MEXU). 


7. Urera laciniata Goudot. ex Wedd., Ann. Sci. 
Nat, Bot, sér. 4, 18: 203. 1854. TYPE: 
[Colombia] Nouvelle Grenade. “Quindui” [Quin- 
dío?], La Bolsa, 1844, J. Goudot s.n. (lectotype, 

designated by de [1975: 308], P 


#00281783). 


Urtica ea Seem., Bot. Voy. Herald 194. 1854. 
TYPE: Panama. 1846-1849, B. Seemann 494 (lecto- 
type, designated by de Realy [1975: 308], BM!). 


Rooij 


Both epithets, Urera laciniata and Urtica girardi- 
nioides, were published in 1854. According to Stafleu 
and Cowan (1988: 139), the probable month of 
publication for Weddell’s publication is March, while 
that for Seemann’s publication is July (Stafleu & Cowan, 
1985: 476), thereby giving priority to Urera laciniata. 


Local name. 


9266, M 


Pringamoza (Panama: J. A. Duke 


Habitat and distribution. Urera laciniata has been 
collected from riverside scrub and from sea level to 
200 m. Its range extends from Honduras, Nicaragua, 


Costa Rica, Panama, Colombia, and Peru to Bolivia. 


Comments. The deeply lobed laciniate leaves and 
large (2-2.5 mm) asymmetrical fruit are unique 
amongst Mesoamerican Urera, and this species is 


unlikely to be confused with any other. 


Selected specimens examined. BOLIVIA. La Paz: Franz 
c. Madidi, senda Azariamas—San Fermin 
Mutún, E Ticona, A. Araujo M., V. Torrez, C. Perez, G. 
Jove & A. Urbano 149 (BM, MO). COSTA RICA. San José: 
A. E Skuich 4266 (GH, MO, NY). HONDURAS. Gracias a 


O. M. Montiel J. 16650 (BM, MO). PANAMA. s. loc.: 
1846-1849, B. C. pod 494 (BM) PERU. ee 
Cataract El Tirol, A. K. o, R. T. Pennington & A. Daz 
3995 (BM, MOL). 


8. Urera lianoides A. K. Monro & Al. Rodr., sp. 
nov. : Costa Rica. Alajuela: San Miguel 
Oeste, Naranjo, subiendo por ladera sur del 
Cerro ene Santo hasta s residual en el 

o, 10705'20"N, 
84724/20"W.. 4000-1900 m, E Nov. 1988, G. 
Herrera 2326 (holotype, INB!; isotypes, BM!, 
MO). Figure 2A—F. 


Species nova Urerae glabriusculae V. W. Steinm. m 
sed is ea habitu lianiformi, parte foliifera caulis 5-10 m 


diametro cava, stipulis apice integris non furcatis, pistillodio 
minore atque achenio laevi 

Vine, scrambling shrub, dioecious (?); main stems 
cane-like, to 25 m, young shoots pubescent to densely 
pubescent, the hairs 0.25- 
appressed, weakly 


l mm, erect or weakly 
curved or erooked, occasionally 
straight; internodes of leaf-bearing sections of ste 
21-84 X 3.5-10 mm, red-brown, occasionally yellow- 
green, ie where 5-10 mm diam., lacking a dark 
stain on the cut portion of the stem. Stipules 4— 
mm, narrowly ovate or lanceolate, not forked, 
ube es 11-90 


0.75-1.5 mm, sparsely pubescent to pu- 


sparsely ese to pubescent; 
(-170) X 


bescent, the hairs 0.125-1 mm, appressed, occasion- 


petio 


ally erect, a occasionally crooked or straight; 
leaf blade 72-280 X 29-110 mm, narrowly ovate 
obovate, Le or oblanceolate, always longer than 
wide, chartaceous and occasionally bullate; adaxial 
surface sparsely pubescent, the mm, 
appressed, weakly curved or crooked; the cystoliths 
punctiform, oblong or occasionally fusiform, occa- 
sionally inflated, aces scattered and occasionally 
parallel to veins, rarely arranged radially around a 
hair base; abaxial surface pubescent, the hairs 0.125— 
0.5(1) mm, 
occasionally straight; the kie fusiform, occa- 
to t 
randomly scattered; primary on 3, occasionally 5, 


appressed or erect, weakly curved, 


sionally oblong, paralle veins, occasionally 
primary to quinternary veins visible to the naked eye, 
the lateral primary veins visible for 1/2—2/3 of the leaf 
length; domatia not present in the axils of the 
secondary veins; base subcordate or obtuse, occasion- 
ally cuneate; margins dentate, occasionally crenate- 
entate or 2 entire; apex cuspidate. Peduncular 
bracts 1.25— ; bracteoles 0.25-0.5 mm. Stami- 
nate Panes 1 to 12 per stem, peduncle 
branched to the base or unbranched from the base 
bescent, the hairs to 
0.25 mm, the whole inflorescence 


or 4.5-9 mm, densely pu r 
mm, bearing 
192 to 384 flowers in an eel cyme with 4 or 
5 orders of dichotomous branching; flowers borne in 
is of 6 to 12, sessile; Es e inflorescences ca. 

d base of peduncle 7-13 mm, 
5-11 mm in fruit, densely uoce the hairs 0.125— 


6 per stem, unbranche 


0.25 mm, erect, straight; the whole inflorescence 12— 
40 mm (12-45 m 


s in an as 


Staminate flowers ca. 1.25 X 1.25 mm immediately 
prior to anthesis; tepals 4, 1-1.5 mm; stamens ca. 
1.75 mm; pistillode ca. 0.375 mm diam., glabrous. 
Pistillate flowers 0.5-1 X 0.375-0.5 mm, white to 
yellow-green; lateral tepals 0.375-0.75 mm, asym- 


Volume 96, Number 2 Monro & Rodríguez 279 
New Species and Synopsis of Urera 


Figure 2. Urera Henares A. K. Monro & Al. Rodr. —A. Habit with staminate inflorescences. =P. Cluster sof staminate 
flowers. —C. S anthesis. —D. Leaf abaxial surface with cystoliths. —E. Habit, with pisti 
—F. Immature fruit. (A-D: Johnson 1600 [GH]; E, F: Shank & Molina R. 4405 [NY]. 


Annals of the 
Missouri Botanical Garden 


metrically ovate, dorsal tepal 0.25-0.75 mm, ovate, 
HER a subapical dorsal thickening; ventral tepal 0.25— 
0.75 mm, ovate, stigma penicillate. Fruit prior to 


0.75 mm; basal 3/4 of 


achene obscured by tepals, lateral tepals ca. 1 mm, 


a of tepals ca. 1.5 X 

dorsal tepals ca. 0.75 mm, ventral tepals ca. 0.5 mm; 
1.25-1.5 X 1-1.25 mm, 
elliptic, keel-shaped, surface smooth. Fruit when ripe, 
> tepals inflated and berry-like 1.5-2.25 X 1.25- 


5 mm, orange when fresh. 


achene asymmetrically 


Habitat and distribution. 
found in premontane, montane, and cloud forest, in 


The new species is 


disturbed and undisturbed forest, from sea level to 
1300(1900—2500) m. Mexico (Chiapas), Guatemala, 


Honduras, Nicaragua, Costa Rica, Panama, Peru, an 
ivia 


IUCN Red List category. Conservation for Urera 
lianoides must be considered as Least Concern (LC) 
according to IUCN Red List criteria (IUCN, 2001), 
owing to the fact that the species has been collected 
50 times in several localities across Central and South 
America. 


Etymology. From the English “liana” or “vine, 
desee from the French, which, in turn, is 
derived from the Latin "ligare" or “to tie," 


“lier,” 
referring to 
the liana-like habit of this species. 


Discussion. Urera lianoides is distinguished by its 
scandent habit, hollow stems, sparse non-urticating 
hairs, and orange, mature, fleshy fruit. Material o 

lianoides from Chiapas differs from other collections 
of this species in having densely pubescent young 
stem and abaxial leaf surfaces, with very shert hairs 
(0.5 mm or less). This species corresponds to “sp. A” 
in the treatment of Urera in the Flora de Nicaragua 
(Pool, 2001: 2495). Urera lianoides most closely 
resembles U. glabriuscula with which it shares a 
distribution in Chiapas, Mexico. The two species can 
be distinguished from each other as follows: x for U. 
ianoides, a vine or scrambling shrub, ca. 5-10 mm 
leaf-bearing stem section hollow, sh inter- 
nodes greater than 20 mm; stipule apex not forked; 
pistillode 0.375 mm diam.; achene surface smooth; (2) 


diam., 


or U. glabriuscula, a shrub or small tree, ca. 5— 
10 mm diam., leaf-bearing stem section never hollow, 
with 20 mm; 
minutely forked; pistillode 0.75 mm diam.; 
su 


internodes less stipule apex 
achene 


ace verrucate. 


Paratypes. BOLIVIA. Beni: Provincia Ballivian, Pilón 


Provincia Arce, 108 km de Tarija hacia Bermejo, R. Ehrich 
494 (K). COSTA RICA. Alajuela: Cantón de San Ramón, 


Reserva Biológica Monteverde, Cordillera de Tilarán, 
5181 (INB); La Palma de San R 

NY); Mata Cartago, La Palma de San Ram 
6308 (F, NY); quebrada Azul (San (nins) A M. Brenes 
23063 (NY); Guatuso, Parque Nac. Volcán Tenorio, El Pilón, 
colectado en bosque primario a orillas de sendero Los 
Misterios del Tenorio, J. L. Chávez 874 (INB); Cantón de Los 
Chiles, Refugio Natural de Vida Silvestre Caño Negro, Río 
Frío, Finca Betel, K. Flores £ M. Flores 130 (INB); Reserva 
Biológica Monteverde, Río Peñas Blancas, Finca Wilson 


neys & W. Haber 710 (INB); Parque Nac. 
Rincón de La Vieja, Colonia La Libertad, Finca Julio Soto, G. 
Rivera 1485 (INB); Colonia Blanca, Finca Río Negro, G. 
Rivera 1563 (INB); San Carlos, Parque Nac. Arenal, Cerro 
Chato, sendero que lleva a la Laguna, A. Rodríguez, V. H. 
O & G. Soto 6254 (INB) Colegio 
Agropecuario, o Y Cantón de San Carlos, A. 
Weston, D. F. Weston & J. Weston 3104 (MO). Cartago: 
Turrialba, Valle del Reventazón, Carolina, 
a de Chirripó, P. g 176 (INB); 1/2 km S of 
pun near rd. #CR R. W. Lent 237 
Guanacaste: Cantón de eN oe de Tilarán, i- 
of Tilarán, continental 
Volcán i enorio, W. Haber & W. 
Zuchowski 11 624 (INB). Heredia: L. R. Holdridge's Finca 
Selva, Río Puerto Viejo at quebrada El Sura & quebrada 
El Salto, ca. 1 mi. above jct i 
3690 (GH); roadside bank about 35 km 
Taylor 4536 (NY). Lim 
Río Pacuare, steep hills 5 of the railroad bridge over the Rio 
Pacuare, W. C. Burger & R. L. so. 6952 (NY); borde de 
lago Dabagri hasta Río Llei, L. D. Gómez et al. 23129 (BM); 
Costado Oeste de lago Dabagri hasta Río Llei, L. D. Gómez et 


Ramírez, G. de La 


km a e Guapiles, 
panitanosos -yolillosos de Suerre y Dos Bocas, drenajes de 
ns Tams smina y Reventazón, P. J. ha nk & A. nen 
yolillosos de Gol 

Rio Reventazón, P. J. Shank & A. s R. 2 = 
untarenas: us de Osa, e Forestal Golfo Dulce, 
Península de Osa, Los Mogos, Bahía Chal, entrada .o 
R. Aguilar 3591 a foothills of Cordillera de manca, 
just N of Santa Elena on Fila Cotón, S of Agua Eine G. 
Davidse, G. Herrera Ch. & M. 
Ballena, py por Pue: 


González 


Talamanca, Las Nubes, Santa Elena, E. Alfaro 329 (INB); W 
part of montafias Jamaica, ca. 3 km NE of Bijagual de 


Volume 96, Number 2 Monro & Rodríguez 281 
2009 New Species and Synopsis of Urera 
Turrubares, Carara reserve, M. H. Grayum et al. 5878 (BM) Portobelo hwy. to 4 km up Río Guanche, S. Knapp 1016 


ca. 100 m 
M, F) Suchitepequez: Finca 
Mocá, A F. Shutch 1479 C HONDURAS. Atlantida: 
Vic. of La Cei f Danto river, slopes of Mount 

e E K. A. Wagner 8454 


García de M. 
(BM, MO). Olancho; Nds del Río de la población de 
Culmí, C. Nelson E. Romero 4738 ee MEXICO. 
Chiapas: 2-4 km below es along rd. to Pichucalco, 
Muni. of Solosuchiapa, D. E. Breedlove 19906 (MO); long 
gravel rd. betw. Palenque & Bonampak, 60 mi. SE 
Palenque, T. B. Croat 40192 (MEXU, iri ue gravel Eo 
etw. Palenque & Bonampak, 88-90 Palenque, T. 
B. Croat 40230 (MO); Mong hwy. 195, pw: Chiapa de Corzo 
W of Pueblo Nuevo Solistahuacán, 


& D. P. Hannon 65180 (MO); Duces Corozal. camino 
ara Lacantum, i Ococingo, E. Martínez S. 
15439 (BM); a 14 km W de Crucero Corozal sobre el 
ie Bm Mpio. 
Martinez v 16642 (BM). Veracruz: 
pio. T R. Cedillo T. 3439 (BM); Ps 
Biológica po Tux ammel, M. Merello & S. Sin 


e 
De 
S 
= 
ge 
SP? 
SR by 


15492 (MO). Nec c On ridge s of 
Cordillera eee A. Geniry, W. D. Stevens, A. Grijalva P. 
& P. P. Moreno 43950 (MO). Granada: Vol Mombacho, 


Hacienda UPE- Bn es del cráter), A. Grijalva, O. 
a 2924 (BM); NE del Volcán 


P. P. Moreno 4103 (MO); NW slopes of Volcán Mombacho, 
10 km S of Granada, M. Nee & J. Miller 27693 (MO); NW de 
Volcán Mombacho, me de Finca Cutirre y camino que 
lleva S del volcán, J. C. Sandino 1273 (BM, MO). Zelaya 
[Región Autónoma del rp oe along the rd. betw. 


no 12397 (MO). PANAMA. s. loc.: $. Hayes 

en Dam, P. H. Allen 2008 
(F, NY); Barro Colorado Island, S. Aviles 106 (F); rd. S-11, 
NW of Escobal, T. B. Croat 12466 (BM, NY); rd. along Río 
Pifia-Río Media divide, NW part of Canal Zone (area W of 
Limon Bay, Gatun Locks & Gatun Lake), 7. M. Johnston 1600 
(GH); rd. along W side of Gatun lake, NW part of Canal Zone 
món Bay, E. Locks € Gatun Lake), 7. M. 
i. S of Colón, ie L. Tyson et al. 
N of Campamento 


Luchio, A. K. Monro & E. Alfaro 4506 (BM, INB, MEXU, 


, M. E. Denda 349 (MO). Coclé: 
Foothills of Cerro Pilós near El Valle, J. Duke & M. Correa 
14670(1) (MO); EI Valle from potato farm above village to 
Cerro Pilon, J. Dwyer & M. Correa 7923 (BM). Colón: From 


(BM); Río Guache, K. J. Sytsma 1617 (BM). Darién: Parque 


Nac. Darién, Serranía de Sapo 


2 
2, 
q 
a > 
= 
o 
= 
S 
m 
5 
3 


a: Basis. 558 (BM 
: Pa argie Nac. del Manu, Río Manu, Cocha 
aii rk J. Terborgh & R. B. Foster 6488 (K). 


9. Urera simplex Wedd., Prodr. (DC.) 16(1): 90. 
1869. TYPE. Colombia. Cundinamarca: “ad salto 
Mar. 1856, Triana s.n. (holo- 


equendama,” 
785! 


de Te 
type, P 00281 


Urera eggersii Hieron., Bot. Jahrb. Syst. 20: 3. 1895. TYPE: 
Ec 


€i 
- D. Pennington 51 50 (pe designated here, K!; 
types, B not seen, NY not s 
Urera reds . W. Stein Acta Bot. Mex. 71: 37. 2005. 
Mexico. Veracruz: Mpio. Andrés Tuxtla, 


18°28'N, 95°10'W, 350 m, 4 Apr. 1981, J. I. Calzada 
8105 (holotype, IEB not seen; UE ENCB not seen). 
Urera tuerckheimii Donn. ae Gaz. sr 14. 1897. 
TYPE Dos ur a Ve erapaz: Pansamalá, 1160 m, 
7, H. von Turckheim 1243 nae US not 

seen; e NY9. 


A neotype was selected for Urera eggersii because 
the type collection cited by Hieronymus, Eggers 14466 
(B), was destroyed in enemy action during World War 
II and only photographs of the holotype could be 
located (F, MO). The 


that it includes good leaf and fertile material and was 


neotype was selected on the basis 


from the same country as the holotype. 


Local names.  Bilsimtezla (Mexico: A. Méndez T. 
6738, BM), chenek'mut (Mexico: A. Méndez T. 4863, 
BM), chichicaste (Guatemala: P. C. Standley 68232, 
F; Mexico: M. Heath & A. Long MA 44, BM), 
chichicaste huevo de cangrejo (El Salvador: £. 
Sandoval & H. Rivera 1252, MO), sakil zulsimtez 
laa (Mexico: A. Méndez T. 6238, BM), tzotzniz zul 
simtez (A. Méndez T. 9066, BM), zulsimtezla (Mexico: 
A. Méndez T. 7022, BM). 


Habitat and distribution. Disturbed and undis- 
turbed forest, cloud forest, and humid scrub from sea 
level to 2500 m. Mexico (Chiapas, Tabasco, Vera- 
cruz), Belize, Guatemala, Honduras, El Salvador, 
Nicaragua, Costa Rica, Panama, Colombia, Ecuador. 
Peru, Bolivia, Brazil. 


mments. Material of this species has frequently 
been determined and referred to (Flora of Guatemala, 
Flora de Nicaragua, and Flora Costaricensis) as Urera 
elata, U. eggersii, or U. tuerckheimii. Examination of 
type material of these species indicates that U. elata is 
a species endemic to Jamaica, while U. eggersii and 
U. tuerckheimii are conspecific with U. simplex. Pool 


Annals of the 
Missouri Botanical Garden 


(2001) indicates U. tuerckheimii (— U. simplex) may 


correspond U. aurantiaca Wedd. = South 
erica (Argentina, Bolivia, Paraguay); ie 
comparison of the holotypes e that the two 


species are distinct, with U. aurantica ir by 
ovate or cordiform leaves and relatively short pistillate 
inflorescences and known only from South America 


(Argentina, Bolivia, Brazil, and Paraguay). Some 
collections of U. simplex from Costa Rica and Panama 


e.g, Kennedy 1939 (GH) and Folsom y Page 5984 


MA), are unusual in the possession of narrowly 


oblanceolate pubescent leaves, while some material 
from Chiapas (Purpus 7039, NY) is characterized 
by densely pubescent leaves. Urera simplex most 
closely resembles U. elata. The two species can be 
e aide from each other based on petiole 
and inflorescence peduncle size as fallows 
(D for U. Eorum petiole lacking small spines; 
staminate flowers 5-parted, occasionally 4-parted; (2) 
or U. elata, petiole with small spines; staminate 
flowers 4-parte 


elected specimens examined. Stann Creek: 
Middlesex, W. A. Schipp 400 (BM, F, GH, E BOLIVIA. La 
Paz: N Yungas, valle de Huarinillas, Estación Biológica 
Tunquini, $. G. Beck 24612 (K, LPB). BRAZIL. Amazonas: 
E. Ule 5465 (K). COLOMBIA. Putumayo: Umbria, G. Klug 
1741 (K). COSTA RICA. Puntarenas: Cantón de Golfito 
Dos Brazos de Río Tigre, Jiménez, orilla de Quebrada Pizote, 
G. Cordero 95 (BM, INB, MO). ECUADOR. Pichincha: P. 
Pennington 5073 (K). EL SALVADOR. 
achapán: Finca L’Esperanza, Jujutla, A. K Monro et al. 
2997 (BM, ITIC, LAGU, MO). odd Olancho: 
i ies Rio Catacamas, slope of Sierra de Agalto, S. 
Blackmore & G. L. A. Heath 1 916 (BM). MEX ICO. Chiapas: 
Finca Menu C. A. Purpus 7039 (BM, F, MO, NY). 
NICARAGUA. Matagalpa: Macizo de Peñas Blancas. Finca 
. Stevens, O. M. 
. Guzmán & D. Cine 5181 (BM, MEXU). 
PANAMA. Comarca de San Blas: Udirbi Reserve, along 
park eric F: inu et al. 257 (BM, MO). PERU. 
Huanuco: Vic. of Tingo María cliffs above Río Monzon, M. 
E. Mathias & D. Taylor 5343 (K). 


10. Urera verrucosa (Liebm.) V. W. Steinm., Acta 


ot. Mex. 05. Basionym: Urtica 
verrucosa Liebm., Kongel. Danske Vidensk 
Selsk. Skr., Naturvidensk. Math. Afd., ser. 5, 2 
295. 1851. Urera caracasana var. tomentosa 


a Wedd., Prodr. 16: 90. 1869, nom. illeg., 
osta Rica. 


puse 14284 os CD. 


Cartago: lrasü, 


Chichicaste (Mexico: E. Kerber 333, 


Local names. 
BM). 

Habitat and distribution. 
coffee farms, Pinus—Quercus—Liquidambar forest, from 


500-2800 m. Mexico (Chiapas, Veracruz), Guate- 


Montane forest, shade 


mala, Honduras, El Salvador, Costa Rica, Panama, 
Peru, and Bolivia 


Comments. Urera verrucosa most closely resembles 
. n 


U. caracasana. The t l ished from 


each other based on leaf texture and inflorescence 
peduncle size as follows: (1) for U. verrucosa, leaves 
bullate; staminate peduncle unbranched at base for 40— 
80 mm, densely pubescent; pistillate peduncle un- 
branched at base for 27-98 mm; (2) for U. caracasana, 
leaves chartaceous; staminate peduncle branched to base 
or unbranched at base for 2-13 mm; pistillate peduncle 
branched to base or unbranched at base for 2-20 mm. 


elected specimens examined. BOLIVIA. La Paz: M. 
Lewis 882155 (K). T RICA. Cartago: N of Ca sd 
Rí ntada, R. A. R. Vickery 959 (BM 
EL SALVADOR. Alcichardas ne de Ninfas, Cordillera 
ae de Apaneca, NW of Juayua, a Davidse, K. Sidwell, 
nro, M. A. Renderos & C. Cortez 37383 (BM RUN 
d MO). GUATEMALA. hinalia On hus. 
e turnoff to Patzám & Sololá, 14.8 mi. NNW of E to 


me 


5 
Hannon 64732 (BM, MO). M 
Boiteri 288 (BM). PANAMA, nes 3 km 
Punta, along dirt rd. on rte. MR Nubes, B. o 1363 
(BM, MO, NY). PERU. San : T. D. Pennington & A. 
Daza 16676 (K, MOL). 


Literature Cited 


Badilla, B., C. Mora, A. J. Lapa & J. A. Silva E. 1999. Anti- 
tory activity of Urera baccifera (Urticaceae) in 
Sprague-Dawley rats. Revista Biol. Trop. 47: 365—371. 
Britton, N. L. & P. Wilson. 1924. Botany of Porto Rico and 
the n Islands. Sci. Surv. Porto Rico & Virgin Islands 
5: 243. 


inflamma 


B W. 1977. Urera. In Flora Costaricensis. Fieldiana, 


0: 276—280. 
iio G., M. S. Sousa & A. O. Chater. 1994. Introducción 
enera xiv ii se, M. S. Sousa & A. O. 


Alis- 


013 09 m Lanjouw & 
A. L. Stoffers (editors). Flora o£ Suriname, Vol. 5(1). E. J. 
Brill, Leiden. 


Field Mu of Natural History. 2006. The Botany 
Collections Database. Department of Botany, Field 
Museum of Natural History, Chicago. <http://emuweb. 


feldmussumorg dete cq Php = accessed 3 Septem- 
ber 20 

Friis, I. 1989. The Urticaceae: A syst Pp. 285- 

P. R. Crane & S. Blackmore m Evolu- 

tion, PE and Fossil History of the 
lidae, 2. Systematics Association. Special Volume 
40B. a Saience Publications, Oxford, United King- 
dom 


Hamame- 


om. 
Gaudichaud-Beaupré, C. 1830. Urera. Pp. 496-497 in 
Voyage autour du M 
roi,...exécuté sur les corvettes de S.M 
Physicienne... p M. Louis de Freycinet. Botanique. 
Pp. 496-497 Pillet- ainé, Paris. 


Volume 96, Number 2 
2009 


Monro & Rodríguez 
New Species and Synopsis of Urera 


Gonzáles, J. C. 1994. Botánica Medicinal Popular, Etnobo- 

tánica Medicinal de El Sal nd Jardín Botánico La 
a, Cuscatlán, El Salvad 

Google Earth, 2008. «hug: Deere acolle com>, accessed 12 


Puane ig F. J. 1999. Planta 6 d d ly 
gico. ee Polar, Servicio oe para d 
Desarollo Ambiental del Amazonas, Car 
. Ochoa, C. Torres, E c 
Medicinales Comúnes de 


M. Riv 1995. Pl 

ondas. Litografía López, S. Uned 
Nacional Autónoma de Honduras/Comité  Intemacional de 
Medicina Natural en Honduras/Cooperación Internacional 


para el pn i (Programa de Cooperación Técnica 
Británica en. Hon Ri ane n cM für Tech- 
nische Zusammenaerbeit, Teguc 

International Plant Names d 
org>, accessed 19 June 2 

IUCN. 2001. IUCN Red List Uam es and Criteria, Version 
3.1. Prepared by the IUCN Species Survival Commission. 
IUCN, Gland, Switzerland, and Cambridge, United 
Kingdom. 

"o D. H. & W. Hallwachs. pa food plants of the 

a de Conservación Guanacaste, northwestern Costa 
<http://janzen.sas. des edu/Wadults/searchfood. 
lasso, accessed 9 November 2005. 

Linneaus, C. 1763. Bes peu ed. 2. Stockholm. 

Miquel, F. A. W. 1853. Urera. Pp. 188-193 in Flora 
Brasiliensis s Vol. 4(1). Fleischer, Leipzig. 

Monro, A. K. 2006. The revision of species-rich genera: A 
phylogenetic framework for the strategic revision of Pilea 
E pr on cpDNA, nrDNA, and morphology. 
Amer. J. Bot. 93: 426-441. 

Pool, A. 2001. En In W. D. Stevens, C. Ulloa, A. Pool & 
O. M. A (editors), Flora de Nicaragua. Monogr. Syst. 
Bot. pru 95. 

Stafleu, F. A. & R. S. Cowan. 1985. Taxonomic Literature, 2nd 
ed., Vol. 5. is Scheltema & Holkena, Utrecht/Antwer- 
pen, and Dr. W. Junk b.v. Publishers, The Hague/Boston. 

p< H. A. Weddell. 1988. Taxonomic 

Literature, 2nd ed., Vol. 7. Bohn, Scheltema & Holkena, 
Uuccht/Antwerpen and Dr. W. Junk b.v. Publishers, The 
ague/Bo. 

Standley, P. A. Steyermark. 1952. Urera. In Flora of 
Guatemala. EN Bot. 3: 424—428. 

Steinmann, V. W. 2005. Four new neotropical species and a 
new combination of Urera so. dus Bot. Mex. 71: 

—43. 


dex. E <http://www.ipni. 


Vellozo, J. M. C. 1827 [1831]. Florae Fluminensis 10. 
Typ ceca Nationali, Rio de Janeiro. 

Weddell, H. A. 1856. Urera. Pp. 143-162 in G. Baudry & J. 
e dion) eee de la Famille des Urtica- 
cées. G. a udry, P. 

2s Pp. “199. 203 in A. de Candolle 
ETTS Prodromus (DC), Vol. 16. Paris. 


APPENDIX 1. Index to Exsiccatae. Collections are listed 
Mri by collector name and then in ascending 
ume order. Species are numbered as in the list 


donde here. 
List or SPECIES 


l. Urera baccifera (L.) Gaudich. ex Wedd. 
2. U. caracasana (Jacq.) Griseb 
3. U. fenestrata A. K. Monro & ‘AL Rodr. 


4. U. AA V. W. Ste 

5. U. gi Ko Momo & Al. Rodr. 
6. U. llipiana. rion val & Ste 

7. U. laciniata Goudot. ex We E 

8. U. lianoides A. K. Monro & Al. Rodr. 

9. U. lex Wedd. 

10. U. verrucosa (Liebm.) V. W. Steinm. 


Acosta, L. 480 (3), 486 (3); Aguilar, I. 387 (10); Aguilar, R. 
et al. 3591 (8); Alclos S., J. 92 (9); Alerasquilla, L. 180 (10); 
faro, E. 329 (8), 745 (3), 1260 O, 1569 (3); Allen, P. H. 
911 (cf. 2), 926 (1), 1471 (10), 2008 > 3617 i 6315 mE 
6325 (9), 6530 (2), 6946 ee 855 (cf. 2); A 
Mori, S. 147 (1); Angulo, L. 88 (5); Antonio, T. rs D, iu 
o 1738 PN o w^ 4988 hr 5111 (3); Cu aee M. 
82 (9); Ai o, D. 1905 raquistain, 
Moreno. P. P NE m Y^ “2009 D 2285 AN 2285 (1), 
2558 be n s 2605 E 2790 (9), 2863 (1); me dele 
& M 04 (2), 4 icc m 06 (8). 


aS 


Walehea L 39579 (2); Bailey, FA &S 
(9); Barrelier, M. 11 (9); Bello, E. 2237 (5), 3143 o e 
(8); Bit d r, J. 1882 (3), 1894 (3); Blackmore, S. h, G. 
L.A.1 16 (9), ud v 288 “854” (10); Boyle, B. p (8); 
rue in D. E. 910 , 9994 (4), 10003 (5), 19906 (8), 
20207 (9), 23588 (4). y (9). 24249 (4), 24998 (9), 34570 
(4), 51467 (5); Breedlove, D 


(9), 31520 (4); rr D. E. & Thorne, R. F. 
5544) (3), 220 (4208) (3), 339 (9), 3480 
s 3596 (9). a O 5793 (8), 6039 (2), 6308 a 6764 O) 
063 (8); Brown, E. 6 (2); Bunting, G. S. & Licht, L. 113 
a Burch, D. 4603 (9); Burger, W. e 4161 (1); Burger, W. j 
& Antonio, T. 10956 (9), 10993 (1); Burger, W. C. & Bake 
R. 10013 (1); B 


6879 (1), jd e r, W. C. & Maita U., 
Burger, W. C. & Stolze, "n E 4912 (3), 5608 (9), 5904 
Burger, W. C. et al. 10502 (1), 10563 Y: 

o E. et al. 2666 (4); 

; Calderón, $ 727 (5), 850 a. Es [^ 1539 0, 1687 
i 1775 (2), 1809 (2); D MET J. et al. 21101 (4), 21242 
(4); Campos, P. oe Carlson, M. C. 391 (10), 2056 (1); 
3) Omnis U. 303 (2); Chavelas, P. et 
0 (9); hate C. 95 (5); Chávez, J. L. 874 (8); 
Ca A.M. 1 i 


8984 (9), 9092 (7); uu W. C. & Liesner, R. L. 6605 (. 
(9); 
(2) 


o 


49785 (7), 66252 (9), 66612 (2), 66795 (2), 68307 (1, 68317 
(9), 78493 En 78524 (9), 78545 (9); Croat, T. 
, 64732 (10), dues (8), 65332 (9), 65354 
. & Porter, D. M. 15663 (9); Croat, T. B. & 
Zhu, G. 76527 (9); E R. 31 (1), s.n. WB-1176' (10). 
9 (9), 10767 (9), 10998 (10); D'Arcy, W. 
Hammel, P 12235 (9); D'Arcy, W. G. & Sytsma, K. 
14516 (cf. 2); n W. et al. 12658 (3); Darío, M. 461 (9); 
arwin, S. et al. 2148 (1); Davidse, G. et E 28277 (8), - 
a0. 37493 M prm C. & Donahue, J. 8353 (9); 


Annals of the 
Missouri Botanical Garden 


Davidson, C. vr i Dawidson, M. E. 349 (3), 484 (9), 487 
85 (1); Delgado, R. 24 (5), 76 (9); Delprete, 
P. 5150 (9); ER P. 5162 (3); Douglas, W. & Krukoff, B. 
A. 3522 (2), 4001 (2); Dryer, V. J. 179 (3), 233 (3), 234 O, 
1646 (10); Duke, J. A. (10), 9266 ab 11844 (9), 12008 
(7); Dunlap, V. C. 174 n, D. et al. 23234 (2); 
a J. D. 2414 (2), 355 (9), 15338 (3 >, 15365 (2); Fr 
J. D. & Correa, M. 7923 (8), 14670 (1) (8); Dwyer, J. D. & 
Conca A M. D. 7503 (10), 7964 (9); Dwyer, J. D. et al. 
(9). 
Ebinger, J. E. 812 (9), 967 (2); Echevarria C., J. A 


4837 


(10), C. 268 (cf. 2); Edwards, J. B. 107 (1), P-107 i 
Espinoza, R. 771 (5). 
Fernández, a-Zamudio, N. 2204 (4); Flores, » 


& Flores, M. 130 (8); Folsom: J. P. 5932 (1); Folsom, J. P. 
al. 2240 ae 5571 As Fonseca Z., 5 79 (2), 113 (9); Fosberg, 
D kie, G. W. 79c (2), 138c (9). 
condo B. 45 6, Garcia, A. R. & Martínez, E. 47 (1); 
ood, N. n et al. 386 i 1134 (1), 2728 (9); Gaumer, G. 
F. 501 (1), 936 (1); Ge 
28 


Cönzále m & ¿varón C. 2327 (8); Cium M. H. & 
Evans 9866 T a M.H. Pe cf. e O M. el 
al. 5878 (8); Gi a, A. & Burgos, F. 1547 (9); Grijalva, A. 
1916 (2); Gri wale a, > et al. 2906 (2 ), 29. » (8); Guadalupe J., 
S. laa F., F. 33 2 
Haber, W. ex Bello, E. 6103 a Haber, W. & E. Bello 7403 
(8); Haber, W. & Zuchowski, W. 9293 (9), 9294 (2), 9377 (8), 
10889 (10), 10907 (9); Haber, W. a a 11279 (9), 11296 (cf. 
2); Hall, J. S. & Bockus, S. M. a Hamilton, C. & 
Stockwell, H. 3537 (9); oe e 58 (8); Hammel, B. & 
Chavarría, E. 17536 (5); mel, B. 1365 (10), 1578 (10), 
2149 (2), 2676 (9). 3016 qo. 4769 (9); Hammel, B. et a. 
dx C. 125 (9), 133 (1 » 
diera R. J. et al. 632 (9); D W. 
s-n. (678139) ee Harmon, W. E. & Dwyer, J. D. 3384 (9); 
Wilson, C. L. s.n. Den (9); Mathew. V 
H. 1354 (10); Hawkes, J. G. et al. 2150 (9); Hawkins, T. 983 
(9), 1119 (1), 1132 (2); Hayes, S. 84 (1), 86 (1), 750 (9); 
Heath, M. & Long, A. MA44 (9), MA53 (9); Henrich, J. E. & 
Stevens, W. D. 345 (2); He 
354 (9), 560 (1), 1295 (8), 2526 (8). d G., C. et al. 542 
(3), 2926 (2); Herrera, H. & Guillen, O. 627 (9); Herrera, a & 
gies J. 795 (8); ue H. 916 (9); 1. . T. de Lux. 
67 (10); Holm, R. W. & Itis, H. H. 124 (2); Howard, R. 
et a 460 (2); E M. 1933 (9). 
Ibáñez G., A. 30 (9); II INBio 183 (5). 
énez M., A. 1067 (8), 2215 (2); Jiménez M., A. & 
Rédrigues, R. 325 (3); Jiménez, Q. & Elizondo, L. H. 746 ( 9); 
Jiménez, Q. 8 e Johnston, I. M. 1600 (8), 1706 (8); 


as 


A 
VASE 
[me 
EE 
Bet 
a 
Ss 
8 
ps) 
- 


Eus 2 H. 1939 (9); Kernan, 

Jan. n et al. 915 (1), 959 (10), 1518 Pu 

Kindo de: J. H. 47 (10), 152 a Kirkbride Jr., J. H. & 

Duke, J. A. 770 (3); Knapp, S. 1016 (8), 1435 (9); Knapp, 5 & 

Mallet, E 9169 (2); Knapp, S. & Monro, A. K 9255 (9); Knees, 
. G. 2706 (9). 


Lao, P A. & ED. A. 450 (9); Lems, K. px (9); 
P. s.n. WB-1 Dor. s.n. "WB-1217 (2); Le W. 

228 (2), 257 (8), 806 is 2067 (9), 2500 2605 00, 2922 
(8), 3648 (9), 3803 (3); e J. 796 (9), 959 (9); Lewis, B. B. 
a W. H. ys 67 (8), 941 (1), 2087 (9); Liesner, 

„£L ood, R. 3 (1); Liesner, R. L. 1706 (9), 1735 

a 1929 D. 3038 o. 3127 (9): Liesner, R. L et al. 15437 


(9), 26275 (1); E n 38 (9); Llodge, C. W. s.n. (13 July 
61 (3); Long, L. E. p d 20m € EL 
Kennedy, H. 4 
Maas, P. J. M. 2736 (1); ME A a et E E 
QM AE 2’ (10), s.n. 28 Apr. 1964’ 
(1); Martínez S, E. M. 8491 (1), ai m 15439 o 16642 
(8), 20711 (2); Vrbe S, E. Téllez, O. 9 (9); 
Martínez S., E. M. et al. 1755 6 eee (9), eus (4), 
34824 (cf. 4); ET 0. 31 2 s.n. (ISF225) (2); Matuda, 
E. 115 (9), 3699 (4), 3956 (10), 4172 (9), 5097 (10), 15347 
(9), 15409 (2), 15547 (5), 16681 (9), 16825 (2), 17637 (1); 
Maxon, W. R. 7123 (2), McDonagh, J. F. et al. 257 (9), 550 (9 
pA oe (9), 609 (9 aff.); ur P. 11 (1); McPhe 
9), Méndez G., A. 8945 (2), 9054 (4); Méndez T. 
ae 9. 6238 (9), 6738 cw d o 7267 (4), 9066 (9) 
Mo G. S. 2001 (1); Miller, J. S. & Sandino, J. C. 1094 (8), 
143 (2); Molina R., A. 868 (1), ps (9), 5520 (9), 6928 (9), 
m (2), 12900 2), 15557 (9), 21969 (cf. 1); Molina R., A. & 
Molina R., A. 24719 ao 26673 (2), 30730 (2); Molina R., A. & 
Montalvo, E. Eos 3 (105 Molina R., A. et al. 16046 (10), 
Monro, A. K. 671 (1); Monro, A. K. & Alfaro, E. 4241 (2), 4346 
2), 4407 (8), us (2), (8); Monro, A. K. et al. 2997 D 3016 
1) 3520 (3 o E. A. 3850 (2); 
M. 60 (5); Morales, 
v 5866 A m (8) Morales, J. F. et al. 
38 (1), a 473 ior. 544 


S 


A 
EN 


* i Ps o 727 


19165 (9}, 19582 (2), 19821 (2), 

(1). 24263 (2); Moreno, P. P. & Hen 

Moreno, P. P. & Sandino, J. C. 1490 
Navarro, E. 211 (8); Nee, M. & a J. 27693 (8); Nee, 


yu 8410 (2), 8883 (1); 


B 
8 


ME PST 
(9); Nelson, C. et al. 29264 (8), 3535 (2), 3955 (2); Nilson, V. 
& Manfredi, R. 392 (3). 
Oersted, a 15112 (7), 21740 (10); Opler, P. A. 179 (2); 
Ortíz, J. J. 989 (4). 
Peck, M. E. 504 (1), Peck 868 (9); Peñate, V. et al. 1307 
(2); Pennell, M. et al. 58 (10); Penneys, D. S. 19 


Pittier, H. 2939 (9), 3804 (2), 3899 (2); Poveda, 
1101 (9), 1159 (T); mos C. A. 7039 (9), 7354 D 135 
Quesada, A. 1161 (3); Quesada, F. 167 (8), 275 (5). 
Ramírez, V. 194 "i Ramos & Cowan, C. 2685 (9); 
mos E., G. et al. 2859 (4); Raren, P. H. 20901 (9), 21623 
8) Renderos, M. A. 494 (2); Renson, C. 204 (1), 279 (1); 
Rivera, G. 258 (3), 259 (3), 432 (cf. 3), 651 (5), 1152 (5), 
1485 (8), 1563 (8); Robles, a E o» is e 2097 (9); 
Robleto, W. 153 (2), 935 (cf. 9 [cf. 1)), 9 [close to 1]); 
Rocha, V. 6 (2); Rodríguez, A y al. A i: on K. et al. 
731 9» 898 (4), 1259 (2); Rojas, S. & Rojas, L. M. 84 (9); 
. M. 840 (2), 1338 (2), 1427 (2), 1496 (2); Rossbach, 


Aue 
2 


9. 


CE 


(2). 
570 (1), 1273 a 1291 (2), 1369 (9), 1749 (1), 2035 (9), 
2060 (2), 2545 (1), 2736 (1), jd (8), 3384 (2), 3400 (9); 
Sandino, J. C. et al. 35694 ; Sandoval, E. 1854 ( 
Sandoval, E. & Chinchilla, R. Py 2), 1182 (2); Sandoval, E. 
& Rivera, H. 1252 (9); Sandoval, E. & Sandoval, M. 1373 (9); 
Saunders, J. 1204 (7); Schipp, W. A. 400 (9), 8111 (1); Schott, 


Volume 96, Number 2 
2009 


Monro & Rodríguez 
New Species and Synopsis of Urera 


A. Hd dr. y D B. G. & Rogerson, D. L. 846 
( DES ann, B. C. 1 e id (7), 495 a Segura, M. 4 
B. Seibert, R. 608 on r, F. C. 3101 (9); Shank, P. 
Molina R., A. 4265 (8. "4405 (8); Er T Ps 2579 
ee 2582 (4), 2585 (5), 4173 (4); Sidwell, K. e 35 (10), 
85 (2); Skutch, A. F. 966 (9), 982 (5), 1479 (8), pr a 3565 
ux 3602 (3), 3750 (9), 4266 (1), 4842 (9); Smith, A. A397 (10), 
A439 (9), A597 (9), H408 (9), H471 (9), P1988 (9), 100 ie 

158 (9), 1082 m 1256 (9), 1487 (9), 2788 (9); Smith, C. E. 

a 2 80 (9); ue a 23 ud a y F. o 

(1); S s Moreno, P. 76 (2); S I. 265 (2); 
NU. E C. 8354 (2), js (2), (s o. “11202 (1), 
20526 (1), 21880 (1), 22344 (1), 22394 (1), 22764 (2), 23982 
(1), 24114 (2), 30536 (1), 4136 (1), 52857 (2), 52918 (9), 
53265 (1), 54051 (9), 60769 (10), 64714 (5), 67884 (5), 68232 
(9), 75549 9 oC (1), 76501 (10), 76980 (5), 78294 (2), 
79579 (1), 79832 (10), 81009 (10), te E 84677 his 

86272 (5), s i 91395 (9); Stegger: 37 (5); St 

E E. A E ao 16919 (2), 25612 y Stevens, W. D. P 
44 (2), 6428 (9), 6601 Y P 2), is D» 
2 Dx us e 12415 (1), i e 08 (9); Stev 
L9: ; Steyermark, J. A. Pd pr. (4), 33491 
(9), rus o. 34292 (4), o D 36617 (10), 36619 (4), 
5), 38770 (1), 46612 (4), 47937 (9), 48785 


we 


Y 


Do 
SS 


E 


Stork, H. y dus n e 2673 (2); Syisma, K. J. et al. 1617 (8). 
1697 (9), 4 

Tate, R. s 315) Taylor, J. & Taylor, C. 11575 (9), 
11672 (9); Taylor, K. 118 (8); Taylor, R. J. 4536 (8); e 
O. ei al. 5181 (9), 7477 (4), 7821 (9); Téllez, O. & Pankhur. 


> 


R. 7277 (9); Terry, M. E. & Terry, R. A. 1403 (7); Todzia, C. et 

al. 2013 (9); Tomlin, S. 10 (2); Tonduz, A. 7167 dd d 

(10); Fre a M. 965 (1); Tún O., R. R. 383 (1), 1080 (9), 

89 (9); ps E. L. 882 (9), 1741 (2). Pu (cf. 
l. 4478 


J. & Utley, K. i 1173 (8), 2319 (3), 2352 (3), 
"ms ©. cs Po 534 
Valerio, M. nes van E. Werff, H. & Herrera, E 
6187 (9); van Dune M. L. 113 (2); Vanderveen, B. D. 6 
(9); Vargas, E. et al. 39. ; 
Vega, S. & Quezada, B. 167 (2); Vi un a ey (5) (9); 
Ventura A., F. 20686 (4), 20944 (cf. E. & 
López, E. 1901 (2); Verhoek, S. E. 5493 co von RM C& 
von Hagen, W. 2083 (2); von Tuerckheim, H. 1243 (9); von 
i : r 2 P 
. 12526 (2), 12688 (9); Weston, A. » 
2991 un Dey (10). 4742 (10), 5080 (9); Weston, A. S. et ai 
04 (8); Whitefoord, C. 1081 (9), 1615 (1), 1881 (9), us 
x 3250 (9); s C. & Eddy, A. 249 (1); Wilbur, R. L. 
et al. 13014 (10); Williams, L. O. & Córdoba, J. J. 4668 (3); 
Williams, L. O. et al. 21816 (10), 25249 (10), F (4), 
28917 (10), 28972 (3), 29009 (2); Williams, R. S. 692 (1), 
716 (10), 728 (9), 794 7 815 (10); Wilson, M. R. dd 
(10); Wilson, P. 2 (9), 579 (1); Woodson, R. E., Jr. & Schery, 
R. W. 257 (3), 593 (3), 863 v Woodson, R. E. Ex et al. 1005 
s v M Kus C. s.n. (1). 
4514 (9), 5044 (2); Yuncker, T. G. et al. 
is o pu (8). 8481 (9). 
Zamora, N. 625 (3); Zuniga, R. et al. 180 (9). 


= 


A REVIEW OF THE GENUS 
DISTICTELLA (BIGNONIACEAE)'? 


Amy Pool? 


ABSTRACT 


nized by their te: 


Distictella Kuntze is a genus of 18 species in the tribe LU a ae aro lianas or less Qu en shrubs, and can 
bifol 


ifoliolate or less fre 


be recog rete branchlets 


or oifoliclate pA tin with a trifid nd tendril; terminal or less a lateral bsc: usually glan 


uently unifoliolate 


propo: sbyi Britton e Rus 
amp., and Distictella negrensis C. V. Freire a A. Yon are Dios sente s as new s 
Distictella broadwayana U 
designated for Distictis racemosa Bureau & K. Schum. 
Key words: Bignoniaceae, Distictella, IUCN Red List, 


sa. 
rb. is presented as a new synonym of Disticiella racemosa var. translucida, and a lectotype is 
& K. Sch 


Pithecocteniinae. 


Distictella Kuntze is a tropical South American 
genus of 18 species in the tribe Bignonieae that can 
generally be recognized by the combination of a 
number of vegetative, floral, and fruiting characters. 
As most members of the tribe Bignonieae, Distictella 
has bilocular ovaries, and fruits that dehisce parallel 
to the septum (Gentry, 1980c), and most of the species 
(as most Bignonieae; Gentry, 1980c) are lianas with 
compound leaves with the terminal leaflet modified 
into a tendril. The exceptions are D. monophylla 
Sandwith and D. laevis (Sandwith) A. H. Gentry, 
which are always shrubs, erect to semi-scandent, and 
and D. cuneifolia (DC.) 


Sandwith, which are sometimes described as scandent 


campinae A. Samp. 


shrubs or as low lianas. Distictella s Eur always 
of a tendril, while 


D. laevis can have either lol or bifoliolate 


is unifoliolate leaves and n 


leaves. The bifoliolate leaves of D. laevis have a 
terminal scar or residual tendril. Distictella campinae 
usually has bifoliolate leaves with terminal tendrils 


common; however, the lowest leaves of a branchlet are 


sometimes unifoliolate. All material of D. cuneifolia 
observed in this study had bifoliolate leaves, with a 
common presence of tendrils. All other Distictella are 
lianas with bifoliolate leaves (D. reticulata A. H. 
Gentry is reported to occasionally have trifoliolate 
leaves; Gentry, 1978b), with a trifid terminal tendril. 

All species of Distictella have terete (to somewhat 
flattened) branchlets and lack interpetiolar gland 
fields. Most have inconspicuous pseudostipules that 
are lost early, but in D. arenaria A. H. Gentry and 
chocoensis A. H. Gentry, they are often foliaceous and 
more persistent. The leaflets are always lepidote and 
often have additional glands. The species vary in 
leaflet shape, especially the base, size, venation, and 
distribution and size of trichomes, and these charac- 
ters are often very useful in differentiating between 
the species. 

The inflorescence is usually a raceme or racemose 
panicle, frequently elongate, with numerous flowers, 
but with only one to t 
simultaneously. Distictella pauciflora A. 


blooming 
entry, 


ree flowers 


‘This paper is number 15 of the Gentry Invitation Series, in acknowledgment of the contributions to the study of the 


pee made by Alwyn H. Gentry. 


I thank the staff of the "uA herbaria for providing loans of herbarium specimens 
MICH, NY, RB, S, TEX, U, UPS, US, W, and Peter J. Stafford, Piet Stoffelen, Timothy 


CM, F, FTG, G, GH, GOET, K, L, M, 


Phillipson for searching - specific specimens and providing me with digitize 


s: AAU, B, BM, BR, C, CAS, CGE, 
mages. 
PU manteca Duan Bilis for pw in preparing ii species 
tion. I give particular thanks to 


than! D. Stevens for his advice and helpfu Comments 
distribution maps, and Bee Gunn for drawing and Fre 
tec ] staff at the Missouri Botanical Gard. 


large percen 
iron (grant 57298). 


whose work mad 
tage of the specimens into the Tropicos database. Financial support was provided by the National Science 


e the specimens ae for study and w 


souri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A. amy.pool@mobot.org. 


Mis 
doi: 10.3417/2006156 


ANN. Missouni Bor. Garp. 96: 286-323. PUBLISHED ON 7 JULY 2009. 


Volume 96, Number 2 
2009 


Pool 287 
Review of Distictella (Bignoniaceae) 


known only from the type, differs from the other 
species in its depauperate inflorescence, and a few 
species (D. campinae, D. laevis, and D. monophylla) 
tend to have smaller inflorescences but (except 
sometimes D. monophylla) produce several flowers 
that bloom over a period of time. The flowers of all 
species of Distictella are very similar, and no floral 
character was found useful in differentiating between 
d The calyces are campanulate, pubescent, + 
te and denticulate, and usually have 
The coroll 


curving, infundibular (rarely campanulate), pubescent 


pically trunca 
distinct glandular fields. as are strongly 
externally and internally above a glabrous, cylindric 

ase, and are usually white with a yellow throat. A few 
own to have purple corollas: the only 
collection of D. pauciflora has purple flowers, and 
purple flowers are common in D. monophylla and rare 
in D. mansoana (DC.) Urb. and D. racemosa (Bureau 
& K. Schum.) Urb. var. translucida A. Pool 


stamens and stigma are included, with the stamens 


The four 


and staminode inserted at about the same level in the 
corolla tube. The points of attachment are marked by 
villous clusters of trichomes or papillae. The nectary 
disc is large, annular-pulvinate and is sometimes 
topped with a narrow ovary stipe. The presence or 
absence of a stipe does not appear to be consistent 
enough to use to separate species. The ovary and style 
are covered with appressed trichomes and are difficult 
to differentiate from each other 
e capsules of Distictella are e oblong or elliptic (or 

rarely weakly spatulate), woody, non-echinate, an 
usually pubescent. Distictella riada (Kunth) 
Sandwith is unusual in having a glabrescent capsule. 

The valves may be convex or slightly to Rind 
compressed, or one valve convex and o 
resulting in a curved fruit. The shape of Le CM 
especially the apex, curving or not curving, drying 
color, and nature of the midrib are useful characters in 
recognizing the species. The seeds are bialate and 
brown with irregular ridges on both the wing and body. 
Gentry, 
and D. racemosa var. racemosa), the wings are slightly 


In some taxa (D. campinae, D. cremersii A. H. 


to greatly reduced, opaque, and subcoriaceous. More 
often, the seeds are distinctly transversely oblong with 
the wings well developed, membranous, and hyaline. 
The wing may be well demarcated from the seed body 
or not. Seed characters are very useful in recognizing 
the taxa. 

1864; Bentham, 1876) 
recognized this genus based on fruit and seed 


characters, but Schumann (1894), Melchior (1927), 
and Gentry (1976) placed more emphasis on vegeta- 


Early authors (Bureau, 


tive and floral characters; the terete stem combined 
with curved flowers serve to separate Distictella from 
its closest relatives, probably Pithecoctenium Mart. 


ex Meisn. and Distictis Mart. ex Meisn. (Gentry, 


= 
2 


Bureau and Schumann (1896) and Sampaio (1935) 
placed a great deal of emphasis on inflorescence type 
and number and arrangement of calyx glandular fields 
in distinguishing the species within Distictella. Sand- 
with en suggested that the inflorescence type was 
not a useful character in this genus, as in many 
species des is a full range from racemes to racemose 
panicles, and Gentry (1982) found the number of 
glandular fields on the calyx to be as variable in the 
flowers of an individual O as between 
species. Neither character was found to be E in 
this study. Sandwith (1957) and p. (1980b) both 
ecies of Distictella 
are very similar and are not useful in distinguishing 
the species. Sandwith (1963) commented that most 
species 


stressed that the flowers of all sp 


in the genus on vegetative 
characters and stated, “It is significant that the floral 
characters of all these plants seem to be essentially 
the same, apart from minor pee of measure- 
ment, nor does it seem probable that the fruits and 
ie will afford good taxonomic i ns" (1957: 
363). In 1963, he went further in suggesting that the 
differences in vegetative characters might “be in- 

uced by habitat and other causes” (Sandwith, 1963: 
49). This study has found a strong species specificity 
to particular habitats, but has also found strong fruit 
and seed characters that correlate with the vegetative 
ones. Gentry, in his new species descriptions (1978a, 

, 1980b) and in his keys in floristic works (1982, 
1997), placed great emphasis on vegetative charac- 
Gentry (1980b: 101) ES 


Distictella have very similar flowers and are distin- 


ters. “All species of 
guished primarily by type of pubescence of the 
vegetative parts and to a lesser extent by fruits. 
While pubescence characters are notoriously plastic 
in many genera of Bignoniaceae, in Distictella they are 
highly constant and correlated with distinctive 
ecologically and geographically defined entities." 
This study basically agrees with this statement, but 
places more importance in the fruit and seed 
characters, and includes other vegetative characters, 
especially habit and leaflet number, shape, and 
venation 


History 


Kuntze (Post & Kuntze, 1904) published the genus 
name Distictella, recognizing seven species (but not 
naming them) and citing in syno 
Bureau (non Meisn.) No description was provided; 


o 
nymy Distictis sensu 


Bureau's 1864 description of Distictis serves as the 
65, Sand- 


with chose Distictella mansoana as the lectotype. 


validating description of Distictella. In 


Annals of the 
Missouri Botanical Garden 


In 1864, Bureau described Pistons as having a 


Eg 
5 
a 
= 
pg 
o 
el 
= 
pg 
2 
= 
Q 
Q 
2 
o 
m 
= 
o 
4 
2 
2 
un 
a 
o 
un 
$ 
= 
T 
=r 
© 
B 
E 
y 
un 
E: 
= 
n 


with that of Meisner (1840), who described the 
capsule of Distictis as glabrous and nearly flat. Bureau 
(1864) also described the new genus Macrodiscus 
Bureau (= Distictis) and compared Distictis sensu 
Bureau (= Distictella) to Macrodiscus (= Distictis). 
Distictella (as ne was separated from Distictis 
(as Mae 


acro on fruit characters, the 
capsule of Distictella ee pubescent (vs. glabrous) 


iscus) p 
and curved with one convex and one concave valve 
(vs. uncurved) with raised midrib in place of a furrow, 

um ends attached to the septum (vs. free), seeds 
pubescent (vs. glabrous), with relatively longer wings, 
and seed scars on the septum linear (vs. punctiform) 
(Bureau, 1864). 

Bentham (1876) included both Macrodiscus and 
Distictis sensu Bureau in his concept of the genus 
Distictis, in which he identified three groups: (1) 
Distictis arthrerion (Mart.) DC. (= Arrabidaea arthrer- 
ion rt ureau ex K. S 
ex B. Verl 
mansoana) and Distictis elongata (Vahl) Bureau ex 
Benth. (= Distictella elongata (Vahl) Urb.); and (3) a 
Macrodiscus group, Distictis lactiflora (Vahl) DC., 

( ) DC. (= Distictis lactiflora), 
(= Distictis 
gnaphalantha (A. Rich.) Greenm.), noting the curved 


mansoana (DC.) Bureau 


E 
Distictis rigescens (Jacq. 
and Bignonia gnaphalantha A. Rich. 


versus uncurved e in distinguishing between 
groups two and thre 

Baillon (1891), A nem (1894), and Bureau and 
Schumann ae recognized and carried on Bureau’s 
misapplication of Distictis and Macrodiscus. Baillon 
(1891) t Re Distictella (as Distictis) and Distictis (as 
Macrodiscus) as closely related, Distictella differing 
from Distictis in having more coriaceous calyces 


co a urved fruits. In his concept 
Distictella des included: Bignonia arthrerion Mart. 

(= Arrabidaea — arthrerion) longata Vahl 
(= Distictella elongata), B. laurifolia Vahl (= Para- 
ureau; Gentry, 1977), 


Pithecoctenium cuneifolium DC. (= 


gonia pyramidata (Ric 
Distictella cunei- 


folia), B. magnoliifolia Bureau (ined., possibly 
referring to B. magnoliifolia Kunth = Distictella 
magnolifolia), B. mansoana = Distictella 


mansoana), , with question, B. kerere Aubl. 
(= Mansoa kerere (Aubl.) A. H. Gentry). 

Schumann (1894) separated Distictis (as Mae 
discus) from Pithecoctenium, Distictella (as Dist 
nd other related genera based on floral character: 
the membranous to subcoriaceous corolla not curving 
at a right angle versus coriaceous corolla bent at 
nearly a right angle. Distictella (as Distictis) was 
separated from Haplolophium Cham. and Pithecocte- 


nium based on its terete stem, which lacks ribs that 
separate with age, and by its curved fruit. Schumann 
1894) treated two species: Distictis elongata and D. 
mansoana. 

In Flora Brasiliensis, Bureau and Schumann (1896) 
recognized seven species in Distictis sensu Bureau: 
oana), Distictis 


Distictis mansoana (= tella mans 


Distict 
elongata (= Distictella elongata), Distictis racemosa 
Bureau € K. Schum. (= Distictella racemosa), 
is Klotzsch ex Bureau & K. Schum. 
(= Distictella parkeri =. Sprague & Sandwith), 
Distictis granulosa Bureau & K. Schum. (retained in 
Distictis; Pool, 2007), Discs crassa Bureau & K. 
1911; 
Sandwith, 1968), and Distictis glaziovii (Bureau ex 
K. Schum.) Bureau & K. Schum. (— 
glaziovii (Bureau ex K. Schum.) A. H. Gentry). In their 


Distictis guianens 


um. (referred to Arrabidaea; Sprague, 


Haplolophium 


key to species, Bureau and Schumann (1896) placed 
emphasis on the type of inflorescence as well as the 
presence and arrangement of calyx glandular fields, 
characters = were found, 
variable species. Bureau a 
(1896) n so me leaflet shape, base, and pubescence, 
characters that were also found to be useful in this 


stud 

Urban (1916a) made the combinations in Distictella 
for Distictis mansoana, Distictis guianensis (= is- 
tictella parkeri), Distictis crassa (referred to Arrabi- 
daea), Distictis elongata, Distictis granulosa (retained 
in Distictis), Distictis racemosa, Distictis kochii Pilg. 
(= Distictella magnoliifolia), aiid Distictis angustifolia 


K. Schum. ex Sprague (= Distictella racemosa var. 


racemosa) and published Distictella broadwayana (= 
Distictella racemosa var. transluci 

Melchior (1927) treated Distictella à in his subtribe 
Pithecocteniinae, which 
nium, Neves-armondia K. Schum. (— 


also included Pithecocte- 
Pithecoctenium), 


o NE Melch. (— entry, 
j Hoo 


described as having bifoliolate or trifoliolate leaves; 
tendrils usually trifid (rarely bifid), or additionally 
forked, and often with disk-like tips; the calyx 
thickly coriaceous and densely tomentose, with or 
without defined lobes; the corolla coriaceous, and, 
except in 


Amphilophium, dense 
h le and s 


pubescent, often 
bilabia 


ep 

curved at a right angle and sometimes — 

and the capsule oblong-elliptic, flattened, the surface 
smooth or iin to echinate, and the seed with a 
membranous octenium (including Neves- 
armondia) and Distictella were placed close together 
based on their similar pollen and calyces and 
separated from each other based on the terete (vs. 
hexangular) stem and curved (vs. not curved) fruit o 


Distictella. 


Volume 96, Number 2 


Pool 289 
Review of Distictella (Bignoniaceae) 


Sampaio (1935) published a key to the species of 
Distictella, as recognized by Urban (1916a), adding 
the new species D. campinae and the new combination 
D. rosea (Kraenzl.) A. Samp. (based on Distictis rosea 
Kraenzl. = Distictis granulosa Bureau & K. Schum.; 
Pool, 2007). Sampaio’s (1935) key, following Bureau 


and Schumann 3 


Eh 


placed a great deal o 
emphasis on inflorescence type and number and 
arrangement of calyx inc fields, with some 
base, and pubes 
published two 
new species in Distictella, D. lutescens C. V. Freire & 
A. Samp. and D. negrensis C. V. Freire & A. Samp. 
(both treated here as D. racemosa var. racemosa). 

ith (1932) scl the 


combination | Distictella parkeri, 


lesser emphasis on leaflet 
1936 


ape, 
cence. in , ampalo and Freire 


prague and Sandwith 
placing — Distictis 
guianensis in its synonymy. Sandwith went on to 
make two additional recombinations in Distictella, D. 
magnoliifolia (1938b) and D. cuneifolia (1953), and 
published five new species: D. pulverulenta Sandwith 
eas ide T ta pulverulenta (Sandwith} 

try), D. dasytricha Sandwith (Sandwith, 
1953), r3 monophylla Sandwith and D. obovata 
Sandwith (Sandwith, 1957), and D. porphyrotricha 
Sandwith (1963), and one new variety, D. monophylla 
(Sandwith, 1957; — Distictella 


laevis). Sandwith relied heavily on vegetative charac- 


aevis Sandwith 
ters in defining his new taxa, primarily habit, 
pubescence, and leaflet shape, size, and venation. 
He also provided an interesting comparison of 
Distictella, Pithecoctenium, and Anomoctenium Pichon 
(= Distictis), emphasizing for Distictella the terete 
branchlets with fine ribs that do not separate; simple 
trichomes; corolla tube lacking glandular fields; non- 
echinate, pubescent capsule; and seeds with irregular 
ridges and brown wings (Sandwith, 1965). 
Gentry Pom DEC new species of Distictella: D. 
s (Gen 
978a), D. a ee 
ne 1980a), and D. chocoensis (Gentry, 1980b). 


His species are again primarily recognized by 


D. arenaria (Gent 
97 


vegetative characters such as pubescence and leaflet 
venation and shape, and to a lesser extent by 
characteristics of the fruit, seed, and inflorescence. 
Gentry (1976) moved Distictella pulverulenta to 
Distictis and provided an interesting discussion about 
the differences between these two closely related 


enera. 
Gentry (1976) outlined the generic differences 

etween Distictis and Distictella described by Bureau 

(1864) and Schumann (1894) and fo 

described or transferred to Distictis post-Bureau 

(1864) were examined, the fruit differences did not 

correspond well wit 


und when species 


the vegetative and floral 


characteristics. For example, Distictis granulosa 


Bureau & K. Schum., D. scabriuscula (Mart. ex DC.) 
A. H. Gentry, and D. steyermarkii A. H. Gentry all 
have pubescent capsules (a character of Distictella), 
but the flowers and branchlets of Distictis and many 
species of Distictella (such as D. arenaria, D. 


campinae, D. cremersii, D. laevis, D. monoph 
u 


even a glabrescent fruit (D. magnoliifolia), character- 
istics previously associated with Distictis, but flowe 

and branchlets of Distictella. Gentry (1976) E 
that. Distictis and D 
best separated by iss and floral eiue 


a were closely relate 


ANATOMY 


Santos (1995) studied 31 genera of Bignonieae and 
divided them into four groups based on the type of 
cambial variant present. The two varieties of Dis- 
tictella racemosa (as D. magnoliifolia) were included 
in her study. Distictella fits into her “group 4," which 
is characterized by stems with discontinuous phloem 
wedges intersected with xylem and parenchyma 
(Santos, 1995: figs. 143, 144). The only other genus 
in this group is Pithecoctenium, which differs from 
Distictella (and the other genera studied) by having 

ary xylem in the bark (Santos, 1995). Two other 
genera in the Pit h 


secon 
hecocteniinae, Amphilophium and 
Haplolophium, as well as Distictis (among a number of 
other genera) fell into her “group 2,” having stems 
with multiples of four phloem wedges in transverse 
section. The remaining genus in Pithecocteniinae, 
Glaziovia, was not included in her study. 


PALYNOLOGY 


Gentry and Tomb (1979) included Distictella 
racemosa var. racemosa d dm as D. ae 
in their SEM survey of pollen a in 
& Tomb, 1979: fig. 6). lion 
this, they concluded that Distictella is iem 
coarse-reticulate pollen (Gentry & Tomb, 1979). This 
is supported by the earlier traditio 
ork of Gomes (1955) and Sandwith (1965), bot! 
n examined D. mansoana (Gomes, 1955: ui i 
fig. 4). Urban — similarly described and 


illustrated the polle . racemos 


Bignoniaceae (Gentry 


nal light 2 


a var. translucida 


ms 


as D. A UN but with bulges on the 
reticulation Ps s the areoles er 
1916b: tab. fig. 3). Sandwith (1965: 412) 
suggested that br Hel poca the “bulging 
sinuosities of the walls with outgrowths.” 

Gentry and Tomb (1979) found this pollen type in a 
number of other genera, which they suggested could 
be divided into two natural groups. The one including 
Distictella is characterized by having thick-textured 


Annals of the 
Missouri Botanical Garden 


white to purple or red flowers, trifid to multi-trifid 
tendrils, and multiseriate ovules and seeds. The other 
genera in the group are Pithecoctenium and Distictis, 
genera that Sandwith (1965) and Gentry (1976) had 
veg suggested as closely related. Amphilo- 
ium, another genus 
(Sa ndwith, 1965; Gentry 
different type: 


suggested as closely relate 
, 1974a), had pollen of a 
_stephanocolpate (zonocolpate) with 


ty laziovia and Haplolo, 
ing ton: all of which are characterized by 
a frilly outer calyx margin, unique in the family 


(Gentry € Tomb, 1979). 


GENERIC PLACEMENT 


Melchior (1927) established the subtribe Pithecoc- 
teniinae, which included Distictella, Pithecoctenium, 
Neves-armondia (= Pithecoctenium), Urbanolophium 
(= Haplolophium; Gentry, 1992), Haplolophium, 


Glaziovia, 


eo 
= 


and Amphilophium. Distictis was 
included in Melchiors concept of the subtribe, 
perhaps due to its less coriaceous corolla and shape 
of the capsule valves; the species that would have 
been considered to be in Distictis at that time, D. 
eel D. gnaphalantha, and D. laxiflora (DC.) 
Gre all have convex capsule valves that are not 
at all fattened. However, in other respects, Distictis 
does fit the description of the Pithecocteniinae. The 
molecular studies of Lohmann (2006) concur, placing 
all six genera in the Pithecoctenieae clade 

Distictella is probably most closely related to either 
Pithecoctenium or Distictis. The three genera share a 
common pollen type (Gentry & Tomb, 1979) and can 
be separated from the 
Pithecocteniinae by the presence of a simple calyx, 
without frill. Distictella differs from both Pithecocte- 

ium isticlis in its terete stem (the stem in 
Pithecoctenium and Distictis being hexangular with 
the ribs often detaching with age), inner surface of the 
corolla pubescent (vs. glabrous, except in Distictis 
pulverulenta), ovary poorly demarcated from style (vs. 
clearly demarcated), seed with brown wings (also in D. 
pulverulenta and D. occidentalis A. H. Gentry), and, in 
some species with hyaline seed wings, seeds with 
minute excrescences on the ridges appearing as dark- 
colored dots on both body and wing of seed (not known 
in Distictis or Pithecoctenium). 

Distictella and Pithecoctenium both have strongly 
curved corollas that are nearly always white and 
always lack glands, have onl 
chomes, and share a similar type of cambial variant 
(Santos, 1995), while Distictis has straight to arching 
corollas that vary in color from white to purple or red 


non-branched tri- 


and often have glandular fields, often have dendritic 


trichomes, and have a different type of cambial 
variant. 

Distictis and Distictella have non-echinate capsules 
d seeds 


transversely oblong, while 


that have the septa without raised edges an 
bialate or distinctly 
Pithecoctenium has echinate capsules (except for P. 
ii (Vell.) A. Pool) that have septa with raised 
ges and seeds with wings clearly on three sides. 


REPRODUCTIVE BIOLOGY 


Gentry (1974b, 1980c) 


recognized 10 
morphological floral types within th 


majo 
e Bignoniaceae, 
each with its own correlating pollination agent(s). 
Gentry (1974b) characterized the flowers of Distictella 

as falling into the Pithecoctenium-type or xylocopid 
flower. Flowers in this group tend to have calyces that 
are thick and often glandular; have corollas that are 
middle, 


red, externally pubescent (above 


al bud bent or curved below the 
white eam colo 
the cio. pss thick (especially at the base), with an 
internal sturdy ridge of thickened (perhaps glandular) 
trichomes at the level of the stamen insertion that 
nearly closes the tube; and produce abundant nectar. 
Gentry (1990: 122) suggested that Pithecoctenium- 
type flowers are pollinated by the same large and 
medium-sized bees as typical Bignoniaceae flowers 
but with the added advantage of the xylocopids being 
"converted from thieves to legitimate visitors." 
Gentry (1974b) identified five different patterns of 
phenology in I Bignoniaceae. Distictella is listed in 
table 1 (Cen ry, 1974b) as type 3 (2) and in table 2 
(Gentry, Iri as type 2(?). “Type 2 applies to steady 
state species; species that put out one or two flowers 
over a long period of time and that are probably 
pollinated by long-lived trap-line bees that follow a 
regular daily foraging route (Janzen, 1971; Gentry, 
1990). There is little or no correlation of flowering 
time with season (Gentry, 1974b). Type 3 


cornucopia phenology; large masses of flowers are 


is a 


se over a period of time lasting from a few 
eeks to over a month and there is strong correlation 
with season » Gent 1974b). Species in this g 
have a wide array of pollinators: reu de 
NN e and bees bound ini 
Flowers of Distictella are very u species 
to species, and generally fit the Pato type, 
though often with a thinner corolla and with the base 
of the tube less markedly blocked by glandular 
species, the calyx is also membranous (D. 
monophylla, and D. obovata). Distictella pauciflora, 
known only from the type, differs from other species in 
its depauperate inflorescence and purple flowers 
(purple flowers are also frequent in D. monophylla 


Volume 96, Number 2 
2009 


Pool 291 
Review of Distictella (Bignoniaceae) 


and infrequent in D. mansoana and D. racemosa var. 
translucida). Most of the species have terminal (or 
terminal and lateral), elongate racemes or racemose 
panicles, 10-40 cm long with 10 to 34 flowers, but 
with only 1 to 3 flowers open at any one time. A few 
species (D. campinae, D. laevis, 
tend to have smaller inflorescences, but (except 
sometimes D. monophylla) produce several flowers 
that bloom over a period of time. 

Phenology patterns are difficult to determine from a 
herbarium study, particularly for species where little 
or no flowering material known (Distictella 
cremersil, pauciflora, "s reticulata, and 
dasytricha). Distictella racemosa var. translucida, D. 
racemosa var. racemosa, and mansoana flower 
nearly throughout the year over their entire range. For 
purposes of comparison, flowering times were com- 


translucida; Amazonas, Brazil, for D. racemosa var. 
racemosa; and Minas Gerais, Goiás, and Distrito 
Federal, Brazil, for D. mansoana. Flowering for five 
species of Distictella (D. arenaria, D. chocoensis, D. 
laevis, D. magnoliifolia, and D. monophylla) does not 
appear to correlate with seasonality. All of these, 
except D. chocoensis, are found in ite sa 
savannas or shru s. Seven taxa of Distictella 
(D. lohmanniae A. Pool. P obovata, D. parkeri, D. 
cuneifolia, D. and D 


elongata, D. mansoana, 


sa . translucida) flower primarily during 

the rainy season. Three taxa (D. campinae, 

porphyrotricha, racemosa var. racemosa) 

flower primarily during the dry season or at a time 

of relatively less rainfall. No apparent correla- 
db 


tions were found between wet or dry flowering seasons 
and soil type, vegetation type, inundation, or eleva- 
tion. 

Most of the species of Distictella have seed wings 
that are well developed, membranous and hyaline (at 
least at the tips). However, three taxa (D. cremersii, D. 
campinae, and D. racemosa var. racemosa) have seed 
wings somewhat to greatly r 


and opaque. Dist 


riparian or seasonally inundated forests and D. 


educed, subcoriaceous, 
ictella cremersii is known from 
racemosa var. racemosa is found in seasonally 
inundated forests, and it is assumed that the reduction 
in the wings is 
water dispersal. However, D. campinae is probably not 


an evolutionary adaptation toward 


associated with a water pa o to label 
data, except for A. H. Gentry & A. Pinheira 13130, 
varzea edge) and D. alla aly nl 
inundated savannas, D. reticulata, from forests along 
riversides, and D. dasytricha, from swampy forested 
areas, have well-developed, membranous, and hyaline 
seed wings. 


DISTRIBUTION 


Distictella is found only in tropical South America 
with one species also found on Tobago. Some of the 
species are found primarily in areas of white sand: D. 


arenaria, D. campinae, D. cremersii, D. elongata, D. 


o 


is, D. magnoliüfolia, D. mono, a, a i 
parkeri. Three taxa are found in seasonally inundated 
areas: D. cuneifolia, D. dasytricha, and D. racemosa 
var. racemosa, and two others, D. cremersii and D. 
reticulata, are found on the margin of rivers. Many 
species are found primarily in savannas or areas with 
scrubby vegetation (campinas, cerrados, or pampas): 


obovata, while the rest are found in forest. Distictella 
cuneifolia is often found in areas with an abundance of 
termite mounds. Most of the species of Distictella are 
found below 400 m elevation. Exceptions to this are 
D. arenaria (100—700 m), D. mansoana (160-1400 m, 
usually over 400 m), D. monophylla (100-1500 m), 
D. obovata (450—1600 m), D. porphyrotricha (900— 


380 m), and D. racemosa var. translucida (100— 


Oe 
A 
o 


m). 

Most of the taxa of Distictella are known from 
limited areas, while three, D. mansoana, D. racemosa 
var. racemosa, and D. racemosa var. translucida, have 
much broader ranges. Distictella mansoana has the 
most southern range. It is found primarily in south- 
central Brazil, northern Bolivia, and southern Peru, 
with a few collections found in northern and central 
Amazonian Brazil (Fig. 1B). Distictella racemosa var. 
racemosa is found from northern central Amazon and 
the Guianas to western southern Amazon (Fig. 2B). 
Distictella racemosa var. translucida is found in 
Tobago and distributed in a western are from the 
Guianas through Venezuela, Colombia, Ecuador, and 
Peru, and peripherally in Brazil (Fig. 2B). The three 
may overlap in northeastern Bolivia and central- 
western Brazil, but D. mansoana is usually found in 
savannas (usually described as pampas or cerrados) at 
higher elevations while the other two are found in 
forests, often at lower elevations. Distictella racemosa 
var. racemosa and D. racemosa var. translucida have 
peripherally overlapping distributions, but D. race- 
mosa var. racemosa is found in seasonally inundated 
areas, while D. racemosa var. translucida is found in 
non-inundated areas. Distictella cuneifolia is found 
along the northern border of Bolivia with Brazil 


— 


Fig. 1A), where it may overlap with the three taxa 
mentioned above, but it is found in savannas or 
pampas, usually at lower elevations. 

everal species are primarily located in central to 
southern Amazon (Distictella lohmanniae, D. retic- 
ulata, D. campinae, D. laevis [also southeast Vene- 


292 Annals of the 
Missouri Botanical Garden 


gene 
P| 0?0'0" 
T I T | 
-70°0'0" -60°0'0" -50°0'0" -40*0'0" 
Q 
1B 
gni 
-10*00"— 
-20°0'0"— 
| | I | l 
-80°0'0" -70°0'0" -60°0'0" -50°0'0" -40°0'0" 


ure 1. Distribution of selected species of a ul m D. — s Samp. (0), D. cremersii A. H. Gentry (0), D. 
cue ola (DC.) Sandwith Mà D. = ... ) A ry ($), D. monophylla Sandwith (L1), D. obovata Sandwith ($>, 
reticulata A. H. Gentry ( noli ifo D. mr Sandwith (A), D. mansoana (DC.) Urb. (O), D. parkeri (DC.) 

X) 


es & Sandwith (W), m D. E ra H. Gentry ( 


found in white sand savannas or campinas, but the 
present distribution suggests that the two species are 
isolated from each other. Distictella magnoliifolia 
(Fig. 1B), found primarily along the southern border of 
Venezuela with Colombia, has also been collected in 


zuela] and D. elongata [also Guianas] where they 
overlap with D. mansoana and D. racemosa var. 
racemosa. Distictella mansoana is the only one of 
these to usually be found above 400 m elevation. 
Distictella campinae and D. laevis (Fig. 1A) are both 


Volume 96, Number 2 


ool 
Review of Distictella (Bignoniaceae) 


00*0'0'4 


10*0'04 


v 


| 
-80°0'0" 


T [ T 
-70°0'0" -60°0'0" -50°0'0" -40°0'0" 
10°0'0"- 
2B 
a 
00*0'0" 4 3 
yu 
A b AS 
A 
-10?0'0"4 
A 
Fr a 
| I 
-80°0'0" -70°0'0" 


Figure 2. 


T | 
-60°0'0" -50°0'0" 


D MEE of selected species of Disticiella. —A. D. arenaria A. H. Gentry (B), D. chocoensis A. H. 
us D. no Sandwith (Y), D. elongata (Vahl) Urb. (O), D. lohmanniae A. Po nd D. porphyrotricha Sandwith | mi 
a (Bureau € K. Schum.) Urb. var. racemosa (A) and D. racemosa var. translucida A. Pool (0). 


Amazonas, Brazil, where it is known from similar 
habitats as D. campinae and D. laevis. Distictella 
elongata (Fig. 2A) is unique in this group. to white 


niae (Fig. 2A) and D. reticulata (Fig. 1A) are found in 


non-inundated forests, apparently not associated with 
white sand, at low elevations. 

A number of species are found in northeastern 
Venezuela and/or the Guianas: im d i 
tricha (also north-Atlantie Brazil; , D. elon- 


gata (also Brazil, D. obovata s Ta, D. pee 


Annals of the 
Missouri Botanical Garden 


(also Amapá, Brazil; Fig. 1B), D. pauciflora (Fig. 1B), 
and D. cremersii (also Ilha de Marajó, Pará, Brazil; 
Fig. LA) where they overlap with the widespread 
varieties of D. racemosa. Of these, only D. obovata and 
D. racemosa var. translucida are found at elevations 
over 500 m. Distictella obovata is often found in the 
savannas of the Pakaraima Mountains of Guyana and 
Gran Sabana of Venezuela, while D. racemosa var. 
commer is found widespread in lowland and 
montane forests. Neither are associated with white 
m pius — D. elongata, and D. cremersii 
are found primarily in white sand forests, D. cremersii 
in riparian vegetation, probably on a peneplain of a 
E ata base (Granville, 1988), D. elongata often 

ear the coast (in the Guianas; 
" 1988) 
disturbed forests, probably associated with schist, 


will and D. parkeri often in low or 


conglomerate, and quartzite of the Orapu-Bonidoro 


series (Granville, 1988). Distictella racemosa var. 
racemosa is known from seasonally inundated forests, 
and D. ce NAT and D. racemosa var. translu- 
cida are both found in non-inundated forests. The 
latter two might be separated by phenology, D. 
porphyrotricha flowers in the dry season and D. 
acemosa var. translucida in the wet season. 1. Distictella 
pauciflora is known from only one collec 

Two species are found only in i or 
Venezuela, or southwestern Venezuela and eastern 
Colombia: Distictella a (Fig. 1A) and D. 
arenaria (Fig. 2A), where they overlap with the more 
widely distributed D. laevis, D. magnolüfolia, D. 
racemosa cemosa, 
translucida. Distictella laevis, 
non-seasonally inundated, white sand savannas, or 
show 
Distictella laevis and D. magnoliifolia have not been 
found above 400 m, while D. arenari 
100—700 m elevation and D. monophylla ranges from 
100—1500 m elevation. 

Distictella chocoensis (Fig. 2A) is the only species 
of Distictella known from the Chocó and Valle del 
Cauca departments of western Colombia, where it is 


campinas, and no seasonality in flowering. 


a ranges from 


endemic. 

I have only seen five collections of Distictella 
dasytricha. Three are from i central 
Amazon of Per m Goiás 
(Fig. 2A). It grows in swampy forests aie 500 m 


u and Brazil and two 


elevation. Its distribution pattern is not clear from the 
collections seen, but appears to overlap with D 
acemosa var. racemosa, which is also found in 
inundated forests at low elevations, D. racemosa var. 
translucida, found in forests not seasonally inundated, 

. mansoana, generally found at higher elevations 


in non-inundated non-forest habitats. 


Economic AND ETHNOBOTANICAL USE 


I have found only one mention of use for 2. 
Schultes (1970) reported the use ictella 
racemosa as an ingredient in a type of curare ne 
by the Barasana Indians living on the Rio Apaporis, 
Colombia. The root is reported to be very toxic. 


MATERIALS AND METHODS 


All specimens examined by Gentry and as part of 
this study have been incorporated into the Missouri 
Botanical Garden 
Tropicos (<http://www.tropicos.org/>), 
n for all SS specimens 


database- ub wii 


contains label informatio 
housed at MO in addition to those received on loan 
from: AAU, B, BM, BR, C, CAS, CM, F, FTG, G, GH, 
GOET, K, L, M, MICH, NY, RB, S, TEX, U, UPS, US, 
and W. Specimens determined by this author were 
selected from this database to generate the distribu- 
tion maps and the Index to Numbered Exsiccatae 
(Appendix 1), the former with the assistance of Duan 
Bills. Geographie coordinates, bus not stated on 
labels, were estimated usin Internet sites: 
<http://earth-info.nga. Se html> and 

<http://www.fallingrain.com/world/>, and wet and 
dry season designations were made en on Brawer 
1991). H 


collectors as indicated on herbarium labels. 


— 


abitat descriptions are those of the plant 


TAXONOMIC TREATMENT 


Distictella Kuntze in T. t & Kuntze, Lex. Gen. 
an. 1904 an : Distictella 
mansoana (DC.) Urb. (lectotype, designate 
Sandwith, 1965: 412 


Lianas or rarely shrubs; branchlets terete, without 
interpetiolar glandular fields; pseudostipules usually 
inconspicuous, short and thick. Leaves usually 
bifoliolate with a terminal trifid tendril 
absent, or rarely unifoliolate or iilos Inflores- 


present or 


cence a terminal or lateral raceme or racemose 


panicle; calyx campanulate, truncate, sometimes 
with minute denticules, usually with glandular fields 
below the apical margin; corolla usually white with 
yellow throat (sometimes purple), infundibular (rarely 
campanulate) from a cylindric, glab 


m 
curving 1/5-1/2 


minutely pubescent above cylindric base, internally 


rous base, strongly 
2 length from base, externally densely 


usually a minutely et cent aN nearly 
glabrous ote) above cylindric 
base, and em with vina s Helene at bases 


en 
, sometimes also 


of stamens and staminode; stamens 4, inserted at 
approximately the same level in the corolla tube, 


dithecal, the thecae divaricate, glabrous, small 


Volume 96, Number 2 
2009 


Pool 295 
Review of Distictella (Bignoniaceae) 


staminode inserted at approximately the same level as 
the oblong, 
sericeous with 
times with a narrow stipe, ovules 4- to 8-seriate 


stamens; dise annular-pulvinate; ovary ol 
d h 


minute appressed trichomes, some- 


in each locule, style pubescent to apex, poorly 
demarcated from ovary. Capsule oblong, elliptic, or 
weakly spatulate, woody, non-echinate, both valves 
convex, one valve convex and the other concave 


ae in a curved fruit, or both valves slightly to 
para to es 
e ala shades of bro 


ngly compressed, 
ent rais B. edge 

with irregular eee (often appearing with ae 
dark-colored dots) on both wing and body, glabrous, 
ody, 
sometimes greatly reduced, often thin, hyaline to 


wings well to poorly demarcated from seed 


opaque. 


KEY TO SPECIES OF DISTICTELLA 


la. Leaves unifoliolate, or bifoliolate and petiolules absent to 2 mm lon 


2a. Un 


rder ti 
2b. Unifoliolate and er abaxial s 
and with the gher 


urfac 
ous and strongly raisal. forming a a fine, closed netwi 

urface of leaflets (or 

at or immersed, inconspicuous and loo. 


ifoli Ma abax ial s e of leaflets densely pubescent with small trichomes and with the tertiary and higher 
rk monophylla 
with trichomes at base of midrib) 
sely open-reticulate. 


and hi order venation 
Sa Lateral veins "initiating at a 70-90” angle with the midrib, extending in a uud sg line toward 
gin, bef lateral vein, strongly brochidodromous; seeds 7-8 X 2 mm, e 
greatly reduced, subcoriaceous And opague res pata t utate ito eem d d See GERTE ee . D. campinae 
3b. Lateral veins initiating at a 45° angle with the midrib, curving toward apex before anastomosing with 


25-28 mm, wings well developed, membranous 
8. 


tertiary veins, or weakly brochidodromous; seeds 10-14 X 
alin D. laevis 


and hy: 


lb. Leaves bifoliolate d petiolules 4 mm long or longer. 
4a. Abaxial surface of leaflets without trichomes or with trichomes restricted to axils of lateral veins with midrib 
idrib. 


sometimes also along midri 
Inflorescence greatly reduced, type with one terminal flower and one flower from axil of uppermost 
leaf reati eea tte pesa e oet ducat Ud cse c aaepe ita tur edsn De eect 15. D. pauciflora 


Inflorescence a raceme or racemose panicle. 
Low lianas or opa tas leaflets 4.7—11.5 X 1.6—5 cm; fruits 5-6.7 cm long; found in 
ae campinas, or pampas 
Ta. ral veins aie at a a 70°-90° angle with the midrib, extending in a nearly straight c 
margin, before looping to meet next lateral vein, strongly brochidodromous; seeds 1 
nd opaque; endemic to Amazonian dual 
d 2. D. campinae 
midrib, boa toward apex before 
mm wide, wings well 


pesi - 
12 mm wido, mae greatly reduced, subcoriaceous a 
whe found in savannas or campinas, on white san 
7b. Dios veins initiating at a 45°(60°) angle with t 
ertiary veins, or weakl pue du seeds 2 
embranous and hyaline; endemic to Bolivia, where it is found in savannas or 
s, on black, silty soils cuneifolia 
Lianas; leaflets 8-24 X 3.3-12.5 cm; fruits 7-22 cm long; found in forests, or D. magnoliijolia found 
y or low forests or savannas. 
perde coriaceous with 2 to 4 pairs of lateral veins, tertiary veins raised on both surfaces; fruit 
drying black, nearly without trichomes, narrowly oblong, 2-2.5 cm wide, apex ae attenuate 
wall 1-2 mm thick; growing in scrubby or low forests Bi E na 0. D. ma, gnolifolia 
8b. Leaflets ees to subcoriaceous with 4 to 9 pairs of lateral veins, e veins flat 
immersed on both surf: ellowish m l1 gold, golden brown, or m 
brown, densely pubescent, n elliptic to oblon oblanceolate, 3.5-7 cm wide, apex acute, 
obtuse, rounded, or obtuse and apiculate, wall NEN mm thick, growing in forests. 
9a. T usually curved with on other convex; seeds 2. E 4.5 times 
s wide as long, wings membranous and hyaline; growing in non-inundated are: 
— ám . D. racemosa var. translucida 
9b. Fruits not curved, valves convex or slightly compressed; seeds less than 2 times as wide as 
ong, wings subcoriaceous to coriaceous and opaque; growing in seasonally inundated areas 
or riparian habitats. 
10 e surface of leaflets with clusters of cun in the axils of the lateral veins 
the midrib, glandular fields absent; seeds 14-16 X 12-20 mm, wings nearly 
es reduced, reddish brown or reddish black: endemic to the Guianas and m de 


e 
E 


in sc 


8a. 


avan! 


e, 


e valve concave and the 


Marajó, n Brazil Se ieee mua usos OE PAR D. cremersii 
10b. Abaxials e of le, usually with glandular fields at the apex 
and in (iis dh of the iod lateral veins with the Sieb: seeds 28 X 24-45 mm, 


wings developed, brown or gray; widespread throughout greater Amazonian South 
America 17a. D. racemosa var. racemosa 


4b. Abaxial surface of leaflets pubescent over entire surface. 


lla. Longest trichomes of d a and petioles 1-2 mm lon, 
12a. Ad 


ng. 
xial surface of leaflets flat, lateral veins initiating a ngle with midrib, higher order 
venation na a very fine, closed network; e. of branchlets, petioles, petiolules, and 


Annals of the 
Missouri Botanical Garden 


ae a in color; Mire lost prior to anthesis; fruit curved with obtuse to rounde 


ent midrib; fou areas below 


d 
——Ó 6. D. dasytricha 


ampy 
12b. 151 su dics “of leaflets ballate, jaiei al veins initiating at a 45° angle with midrib, higher order 
venation forming a | loosely i sed network; trichomes of branchlets, petioles, petiolules, and 


bra fr 
without a prominent midrib; bu in terra firme forests between 900 a 


eoles persistent in flower; fruit not curved, E ais apex and 
nd 1 


Sie D. porphyrotricha 
llb. Longest und 7 E and pre 0.03-0.7 m 
13a. Ter sed to flat 


m lon, 
abaxial surface of leafle t. 


gm dos Vos. he adaxial surface blackish green, abaxial aaa covered with 


tomentose with trichomes to 0.15—0.2 mm long, lateral veins initiating ata 


de trichomes to 
lary s 


0.03 mm long, sn veins initiating at a 20° angle with midrib, 
veins generally parallel, closely and non-branching; do. slightly curved 
with wings well pene, from seed body; found in terr: 
BA octies alee a edd olt tees i tte eee oS 9. D aa ie 
g a similar color on both surfaces, + brownish green, abaxial Bud 
45^ angle di 


midrib, tertiary veins generally prp | not e SAA closely set, and parallel; 
all 


capsule strongly curved with midrib, n 


with wings poorly 


demarcated from seed body; found sumas in white sand forests, the Guianas, Bolívar, 


uela, and Amapá, Brazi 


Venez Á, Brazil . D. parkeri 
13b. Tertiary veins raised on abaxial surface of leaflet. 
15 


ets obovate or oblanceolate (rarely elliptic) with the base cuneate or shortly attenuate; 


capsule with an acuminate 
6 


ex 
artaceous to m oriaceous with a flat margin, lateral veins initiating at a 


E inent midrib; found between 0 and 200 m in French Guiana, Surina PE and 


D. elongata 


with midrib, adaxial surface drying glos 


Sy; willuesconta raceme; fruit not 


1660 m in the region of the Pakaraima 
untains of Guyana and adjacent Bolívar, Venezuela 13. D. 


Vu e ati AUR obovata. 


15b. Leaflets ovate, lanceolate, or elliptic e the base rounded to cordate na + acute 


in 


17a. 


lib. 


1. Distictella arenaria A. H. Ge 
York Bot. Gard. 29: 273. 1978. TYPE: Vene- 
zuela. Amazonas: 18 km S 
Río Sipapo, 125 


m, 29 June 1975, A. H. Gentry de ovate, thick, 3-10 X 3-5 
P. Berry 14615 (holotype, MO; isotype, NY not 
seen, digitized image!, VEN not seen). 
Liana; young branchlets terete, drying dull brown or 
black, usually solid, lepidote, pilose or puberulent or 


; a us a di with apex acute or obtuse (not known in D. choc 
rface 


nsis). 
a of leaflets with tertiary and higher order venation similarly 


Abasia ae of leaflets with tertiary veins more raised than higher order 


"Jh. 


. e alan 


"Tertiary and higher order venation wo ee and usually raised on 
adaxial surface, usually inconspicuous and always raised on abaxial surface 
1 loosely clo etwork; a narrow (width 1/5-1/3 
ees rarely present; fruit curving with a prominent midrib; seeds with 
1 ell demarcated from seed body; known from Brazil, Ed and 
Shs Gude ats eit Cet etiete E ee eS pao EE T Oe 1. D. mansoana 
Tertiary and higher order venation apie in adaxial surface, conspicuous 
and raised on abaxial surface formin; medium to fine closed network; 
pseudostipules b (width 1/2 to Sul Teng, often present; fruit not 
cu urving, midrib not evident or subevident (not known in D. chocoensis); seeds 
with wings poorly BERI from seed body (not known in D. chocoensis); 
known from Venezuela Colombia. 
19 


ey 


Petiolules longer m petioles; tertiary veins branching; often growing 
on white sand in disturbed areas, known from southern Venezuela and 
eastern Colombia elec emx 1. D. arenaria 
Petiolules shorter than petioles; tertiary veins not ES kno 
rom wet forests of western Colombia 


E 
© 
c 


OW: 
usu Vaan rE Ne . D. chocoensis 


ntry, Mem. New with erect or ascending trichomes; trichomes white, 


an, or ferruginous, 0.1-0.7 
of Samariapo toward ually p 


+ 


mm; pseudostipules 
resent, asymmetrically orbicular or broadly 
, the tip usually curving 
strongly downward. Leaves bifoliolate, rifid tendril 
sometimes present; petiole 5-15 mm, with pubes- 
cence like that of branchlets; petiolules 8-27 mm, 
with pubescence like that of branchlets; leaflets 


Volume 96, Number 2 
2009 


Pool 297 
Review of Distictella (Bignoniaceae) 


lanceolate or ovate, 7-19 X 4.4—11.5 cm, chartaceous 
to subcoriaceous, 4 or 5 pairs of lateral veins, each 
lateral vein initiating at a 45°-60° angle with the 
midrib and gradually curving upward toward margin 
and anastomosing with tertiary veins, tertiary veins 
conspicuous, anastomosing with higher order venation 
to form a fine, closed network, all veins immersed 
adaxially and raised abaxially, both surfaces drying 
yellow-green or olive, adaxial surface with midrib, and 
often lateral veins, pubescent, inconspicuously but 
abundantly lepidote, with or without scattered glands, 
abaxial surface pilose or with erect to spreading, white 
trichomes, 0.1-0.3 mm, abundantly lepidote, usually 
the lateral 
veins with the midrib, margin flat, base cordate, 


with fields of glands in most of the axils o 


subcordate, or rounded, apex acute or obtuse with tip 
cuspidate or apiculate. Inflorescence a raceme Pd 
racemose D peduncle with rachis 9.5-2 

with 27 flowers, peduncle ca. 3 mm c at 
base, Eme rachis, and pedicels drying reddish 
brown, with pubescence like that of branchlets or 
minutely puberulent, lepidote, bracteoles absent or 
not seen; calyx 8— mm, subcoriaceous, with 
minute, appressed, ferruginous trichomes and densely 
lepidote, glandular fields 2 or 2.5 pairs; corolla white 
with yellow throat, infundibular from cylindric base, 
curved 1/3 distance from base, 4.5-5.5 cm, charta- 
ceous, tube 3.2-4.2 em with cylindric base 9-12 mm 
long and mouth 14-15 mm 


internally pubescent with minute trichomes above 


wide, externally and 
glabrous base, internally with clusters of glandular 
papillae or villous at bases of stamens and staminode, 
lobes 11-15 X 11-12 mm; stamens and staminode 
inserted 9-10 mm from base of corolla tube, anthers 
4—4.5 mm, longer filaments 24—28 mm, shorter fila- 
ments 18-25 mm, staminode ca. 4 mm; disc ca. 1 X 
3.5—5 mm; pistil 3.4—4 cm, ovary 3.5—5 X 2-2.5 mm, 
stipe 0.5-1.5 mm, stigma lanceolate. Capsule oblong, 
not curved, 7.5-11 X 2.53 cm, ca. 0.3 cm diam., 
base acute and truncate, apex acute, valves fairly 
strongly compressed, drying reddish or golden brown, 
puberulent with minute, white or ferruginous tri- 
chomes, glandular and warty, midrib subevident or not 
at all evident, wall ca. 2 mm thick. Seeds 10-18 X 
32-55 mm, wings poorly demarcated from seed body, 
reddish brown with tips lighter brown, membranous, 
hyaline, especially at tips, the veins contrasting in 
color to surface 


Distribution. | Distictella arenaria is known from 
southern Venezuela (Bolívar and Amazonas) and 
eastern Colombia (Vaupés and Vichada). It has been 
reported from thickets on white sand, shrubby islands 
in savanna, and road and stream sides, at elevations 


ranging from 100 to 700 m. Figure 2A. 


Phenology. Flowering June, December, and Jan- 
uary; fruiting November and March. 


Discussion. — Distictella arenaria is most similar to D. 


mansoana and D. chocoensis. Distictella mansoana 
differs from D. arenaria in having narrow (3-7 X l- 
early caducous ipia un leaflets with 


on the xial surface, 


petioles usually longer ilum petiolules. (vs. petiolules 


.5 mm), 
the tertiary veins usually 


usually longer than petioles), capsule much pa (1.2- 
2.5 cm diam.) with a prominent ape eeds with 
pera Disticilla 
chocoensis differs from D. pua in having the petioles 


wings well demarcated from se 


longer than the petiolules, tertiary veins not branching, 
and ovaries without a stipe. The last character may not be 
reliable, as few flowering specimens of either species are 
available. Fruiting material of D. chocoensis is not known. 
Distictella arenaria is also separated by geography and 
habitat from D. chocoensis, which is endemic to the wet 
forests of Chocó and Valle del Cauca, Colombia, and by 
geography from D. mansoana of Brazil, Bolivia, and 
southern Peru. 


Additional specimens examined. COLOMBIA. Vaupés: 


trativo on T to El Tapón (due S of Cerro Peinilla), J. L. 


Zarucchi & C. E. Barbosa 3437-a (MO). VENEZUELA. 
azonas: a x ado sur del río Ventuari, 
Steyermark 


Atures Cane Mor coy, via Pue 
3502 (MO); Dpto. Ribas, Cerr 
Liesner 25720 (MO), 25767 (MO. T Río Caura, ariba 
del Salto Para, en las isles 2-3 km arriba del camp, Las 
Pavas, J. A. Steyermark et al. 113086 (MO). 


mp., Ann. Acad. 
TYPE: Brazil. Pará: 
e 1926, as “21 June 1927,” A. 

ucke s.n. (RB 22688) (E RB!; isotype, 
B ). 


2. Distictella campinae A. 


ana or scan ndent shrub to oun, 
g dull Re pes or 
black, solid, lepidote, a or glabrescent; 
trichomes colorless or ferruginous to ca. 0.5 mm; 
DN UNE caducous, clavate, thick, 1-3 X 0.4— 
lm eaves bifoliolat 


id on branchlet unifoliolate, trifid tendril present 


te, or sometimes first pair of 


(rarely residual to 1 mm long) or absent; petiole 4— 
12 mm, with trichomes like those of branchlets; 
petiolules (1)3—7 mm, with trichomes like those of 
branchlets; leaflets elliptic (rarely spatulate), 5.2— 


Annals of the 
Missouri Botanical Garden 


11.5 X 1.6—5 cm, coriaceous, (7)8(9) pairs of lateral 
veins, each lateral vein initiating at a 70-90” angle 
with the midrib and extending in a nearly straight line 
toward margin before looping to meet next lateral vein, 
strongly brochidodromous, tertiary veins inconspicu- 
ous, loosely open-retieulate, all veins immersed 
adaxially, midrib raised abaxially, and other veins 
immersed to flat (lateral veins sometimes slightly 
raised), both surfaces drying pale gray-green, adaxial 
abundantly lepidote, 
sometimes with scattered glands, abaxial surface 
without trichomes, abundantly lepidote, often with 
glandular fields at base and sometimes at apex 
glands scattered, in flat (rarely d red 
base rounded, or euneate and then rounded, apex 
acute. i ae a raceme or racemose panicle, 
peduncle with rachis 3-13 cm with 5 to 11 flowers, 
peduncle 1-2.5 mm wide at base, peduncle, rachis, 
and pedicels drying dull yellowish brown or grayish 

rown, pubescent with trichomes like those of 
branchlets, bracteoles ca. 1.75 X 0.5 mm, usually 
lost before anthesis; calyx 7-10 X 8-9 mm, membra- 
nous or subcoriaceous, pubescent with minute red or 
colorless trichomes tightly appressed, glandular fields 

to pairs; corolla white, infundibular from 

dde base, curved 1/3-1/2 distance from base, 
5.5-6.5 cm, membranous, tube 4—4.2 cm with cylin- 
dric base 10-11 mm long and mouth 13-18 mm wide, 
externally densely puberulent with minute trichomes 

ove glabrous base, internally RAE and 
lepidote, with villous clusters of trichomes at bases 
of stamens and staminode, lobes 15-20 X 1553 mm; 
stamens and staminode inserted 9-11 mm from base 
of corolla tube, anthers 3—4.2 mm, longer filaments 

9 mm, Va ur 15.5-17 mm, staminode 

2-3 mm; disc 1 a. 3.5 mm; pistil 3.3-3.5 c 
ovary ca. 3 X D mm, stipe 0.5-2 mm, sonia 
lanceolate. Capsule elliptic, not curved, ca. 5 X 
2.2 em, ca. 0.4 em diam., base acute, apex obtuse or 
truncate, valves inm drying brown, densel 
pubescent with minute, appressed, dull white tri- 
chomes, glands scattered, midrib not evident, wall ca. 
2.2 mm thick. Seeds 7-8 X 11-12 mm, wings greatly 
reduced and poorly demarcated from seed body, 
reddish brown, subcoriaceous, opaque, the veins not 
contrasting in color to surface 

Distribution. — Distictella campinae is known from 
the Brazilian states of Amazonas and Para with one 
collection from Rondónia. It is often found growing in 
white sand savanna or campinas, between 50 an 
100 m. Figure 1A 


Phenology. Flowering March, August, September, 
and December; fruiting Mare 


Sandwith (1962) treated Distictella 


a as a synonym of D. cuneifolia. Distictella 


Discussion. 

very similar to both D. cuneifolia an 
bifoliolate s specimens of D. laevis. However, the lateral 
veins in the leaflets of D. cuneifolia and D. laevis are 
initiated at about a 45^ angle before curving toward the 
apex and finally anastomosing with the higher order 
venation. They also differ markedly in the seeds; the 
wings are hyaline, well developed (seeds 9-14 X 25- 
28 mm), and well demarcated from the seed body in D. 
cuneifolia and D. laevis. Distictella campinae and D. 
laevis have similar geographie distributions and are 
found in similar habitats, but D. cuneifolia is endemic to 
Bolivia, where it is found in savannas or pampas with 
black silty soils, which are often described as seasonally 
inundated, or with numerous termite mounds 


Selected specimens examined. BRAZIL. Amazonas: Rio 
ü 


tumá, de ltapiranga, próximo à cachoeira do 
Tucumari, Cid Ferreira et al. 497 (MO). Para: 10 km 
of Portel, G. T. Pra i. 1291 (MO); Mun. de Vigia, 


36 km SE of Vigia, along hwy. PA-140 to Belém, G. Davidse 
et al. 17568 (MO). Rondônia: Mun. Costa Marques, BR- 
123 km de Conia Marques, entrando 06 km 

na Faz., Três Irmãos, C. A. Cid Ferreira 


m num vicinal 


As us 


et al. 8664 (MO). 


3. peus pee A. H. Gentry, o 

4T: 1980. TYPE: Colombia. Cho 

10 e W of Istmo de San Pablo on Power 

Hwy., W of Las Animas, 110 m, 12 Jan 9, A. 

H. Gentry & E. Renteria A. 24089 ee Es 
MO). 


not seen; isotype, ! 


Liana; young rus terete, sometimes flattened 
nodes, dryin 


ES pilose m tangled, ferruginous trichom 


ull brown or reddish brown, solid, 
ca. 03 mm long or puberulent with erect or 
0.05-0. A mm; 
4—7 X 
aves bifoliolate, trifid tendril sometimes 


ascending, ferruginous trichomes, 


sd) du often present, oblong, thick, 
4 mm. L 
um petiole 25-60 mm, with pubescence like that 
of branchlets; petiolules 10-30 mm, with pubescence 
like that of branchlets; leaflets lanceolate, ovate, or 
elliptic, (6-)13-24  (3.5-)7-14.4 cm, subcoriaceous 
to coriaceous, 5 or 6 pairs of lateral veins, each lateral 
vein initiating at a 45°—60° angle with the midrib and 
gradually curving upward toward margin, tertiary 
veins conspicuous, connecting midrib and lateral 
veins, higher order venation forming a medium to fine, 
closed network, all veins immersed adaxially and 
raised abaxially, adaxial surface drying brownish 
green, somewhat bullate in more coriaceous leaflets, 
midrib and often lateral veins and surface pubescent, 
without scattered glands, abaxial 
grayish green, pilose or with erect to spreading, white 


Volume 96, Number 2 
2009 


Pool 299 
Review of Distictella (Bignoniaceae) 


trichomes, 0.2-0.3 mm, abundantly lepidote, often 
with fields of glands in the axils of most of the lateral 
veins with the midrib, margin flat, base cordate, 
subcordate, or rounded, apex acute or obtuse. 
Inflorescence a raceme or racemose panicle, peduncle 
with rachis 15-25 em with 20 to 30 flowers, peduncle 
3-5 mm wide at base, peduncle, rachis, and pedicels 
drying reddish brown with pubescence like that of 
branchlets, bracteoles absent or not seen; calyx (7—) 
10-12 X (7211-12 


appressed, ferruginous trichomes and densely lepi- 


mm, coriaceous, with minute, 


dete, glandular fields 2 or 2.5 pairs; corolla white, 

undibular from cylindrie base, curved 1/3 distance 
e, 4-5 cm, subcoriaceous, tube 3—4 cm with 
ong and mouth 12-15 mm 


wide, externally and internally pubescent with minute 


from 
indc base 9-10 mm 


trichomes above glabrous base and internally with 
clusters of glandular papillae or villous at bases of 
X 8-14 mm; 


stamens and staminode inserted 6—8 mm from base of 


stamens and staminode, lobes (5—)10-1 


corolla tube, anthers ca. 4 mm, longer filaments 22— 
23 mm, shorter filaments 17-20 mm, staminode 5.5— 
7 mm; dise 1-2 X 4—5 mm; pistil ca. 3.6 cm, ovary 
(3-4—4.5 X (1.5-)2.2-2.5 mm, stipe absent, stigma 


orbicular. Fruit unknown. 


Distribution. | Distictella chocoensis is endemic to 
the Chocó and Valle del Cauca departments of 
Colombia. It has been found in wet forests from sea 


level to 110 m elevation. Figure 


Phenology. Flowering July, October, and Decem- 
ber; fruiting material not known 


Discussion. — Distictella chocoensis is most similar 
to D. arenaria. Distictella arenaria differs from D. 
chocoensis in having the petiolules longer than the 
petioles, tertiary veins branching, and the ovary on a 
stipe (see discussion of D. arenaria). Distictella 
arenaria is known from southern Venezu and 
eastern Colombia, a it is usually found in white 
sand thickets, shrubby islands in savannas, and 


disturbed areas. 


Additional specimens examined. COLOMBIA. Chocó: 
501 m, 2 i Triana 4124-10 (BM). Valle del Cauca: Bajo 
Cal to Juanchaco Palmeras, A. H. Geniry et al. 47844 
(MO), po (MO); Bajo Calima, ow: “of Carton de 
Colombia, near entrance to Dindo area, A. H. Gentry et al. 
E (MO); Bajo Calima, ese Pulpapel/Buenaven- 

ura, M. Monsalve B. 1916 (MO). 


4. Distictella cremersii A. H. Gentry, Phytologia 
6: 209. 1980. TYPE: French Guiana. Haut 
Tampoc, le = des Criques pres de la Crique 
Alice, 1 Apr. 1977, : Cremers 4589 (holotype, 
MO; isotype, »* not seen). 


Liana; young branchlets terete or flattened, drying 
reddish brown, hollow or solid, d lepidote, 
sparsely pubescent; trichomes to mm; 
trifid 


tendril sometimes present; petiole 15-17 mm, with 


pseudostipules not seen. Leaves p feliolate, 


trichomes like those of branchlets, or sometimes 
longer to 0.3 mm; petiolules 8-12 mm, with trichomes 
like those of branchlets, or sometimes longer to 

mm; leaflets elliptic, 10.3-16 X 5-7.5 cm, 
chartaceous, 4 to 7 pairs of lateral veins, each lateral 
vein initiating at a 20°—45° angle with the midrib and 


curving toward the apex and fading, tertiary veins 


extent aay anastomosing with higher or 

o form an open or loosely closed network, all 
veins pee adaxially or all but midrib flat, midi 
and lateral veins raised abaxially, and other veins 
immersed to flat, both surfaces drying olive-green or 
surface without trich 


brownish green, adaxial omes or 


midrib with minute appressed trichomes, scattered 
lepidote, glands not seen, abaxial surface with axils of 
lateral veins with midrib pilose, and some trichomes 
often along midrib, scattered lepidote, usually with 
scattered glands, margin flat, base cuneate, cuneate 
and then rounded, or inequilateral with one side acute 
and the other obtuse, apex acuminate or obtuse and 
then briefly acuminate, tip 
apiculate. Capsule broadly elliptic or oblong, not 
curved, 7-11 X 4-5.5 cm, 34 cm 


and truncate, apex ol 


iam., base acute 
obtuse, valves strongly convex, 
drying greenish gold or golden brown, densely 
pubescent with minute, appressed, golden trichomes, 
midrib not evident, wall 4-5 mm thick. Seeds 14-16 
X 12-20 mm, wings nearly totally reduced, poorly 
seed body, reddish brown or reddish 


lack, subcoriaceous, opaque, the veins not contrast- 


demarcated from 


ing in color to surface 


Distribution. | Distictella a is known from 


Suriname, French Guiana, an a de Marajó, Pará, 
Brazil. It is found in riparian Sx en the elevation 
indicated on only one specimen is 25 m. This 


& G. Lewis 1869) also described 


the habitat as a seasonally inundated, secondary, 


specimen (R. Evans 


white sand forest. Figure 


Phenology. Fruiting March and April. Flowering 
material not seen. 


Discussion. Flowering material of Distictella cre- 
mersii is not known, but the fruit and vegetative 
material suggest a relationship with D. racemosa. The 
leaflets are similar to those of D. racemosa var. 
translucida: the lateral vein axils are pilose and there 
are no glandular fields at the leaflet apex or base. The 


fruits and seeds suggest D. racemosa var. racemosa, 


Annals of the 
Missouri Botanical Garden 


except that the fruits of D. racemosa var. racemosa are 
relatively narrower and thinner (8-16 X 3.5-7 em, 

cm diam.) and often compressed and split 
along the midrib at maturity, and the seeds have a 
much larger wing (seeds 15-28 X 24—45 mm). Label 
information currently available suggests that D. 
cremersii is found in riparian vegetation oou 
nundated forests on white sand), D. racemosa var. 
racemosa in seasonally inundated ee associated 
with black water, and D. racemosa var. translucida in 
forests that are not seasonally floode 


Additional us examined. NA Pará: Ilha de 
Marajó, rio Mucunas, afluente do rio Anajás, em frente a 
cidade de Anajás, A. S. Tavares "s (MO). FRENCH 
GUIANA. Rivière Camopi, en Amont du Saut Yaniwé, J.-J. 
de e 2080 (MO); Haute Approuague, crique Matar- 
R. A. A. O); Rivière Tampoc, saut 
raton, Moretti 636 (MO). SURINAME. Para: Along rd. 
from Zanderij to Kraka, 4.9 km from intersection w/ Zanderij 
Hwy., 30 m before bridge over Sabakoe Creek, R. Evans & G. 
[PM 1869 (MO). 


5. Distictella cuneifolia (DC.) Sandwith, Kew Bull. 
1953: 476. 1953 aa Basionym: Pithecocte- 
nium cuneifolium D rodr. 96 


1845. TYPE: Without eee [Bolivia, cf. dise.]. 


“Para,” s.d., s. coll., s.n. (holotype, P not seen, 
o F ain Mela photo K not seen; isotypes, 
seen, microfiche C 


9.196. 16, 6 je. not seen). 


liana or scandent shrub to 
ull black or brown, solid, 


w 3m; young 
branchlets terete, drying d 

lepidote, pubescent; trichomes dull white, to 
0.06 mm; ipsutis not seen. Leaves sina. 
6-16 mm, with 
trichomes like those of branchlets; petiolules 4— 


trifid tendril often present; petiole 


mm, with trichomes like those of branchlets; 
leaflets narrowly spatulate, almost oblong, or rarely 
4.7-8(111) X 1.8-3.3(4) em, 
coriaceous or subcoriaceous, 5 to 7 pairs of lateral 
°) angle 


oblong or elliptic, 


veins, each lateral vein initiating at a 45°(6 
with the midrib and curving toward the apex, 
anastomosing with tertiary veins and weakly brochi- 
dodromous, tertiary veins inconspicuous, loosely 
open-reticulate, all veins immersed adaxially, midrib 
raised abaxially, and other veins immersed to flat 
(lateral veins sometimes slightly raised), both surfaces 
drying gray-green or brownish green, adaxial surface 
without trichomes or midrib with minute appresse 

trichomes, abundantly lepidote, usually with scattered 
glands, abaxial surface without trichomes, abundantly 
lepidote, usually with scattered glands, margin flat 
(rarely slightly revolute), base cuneate or cuneate and 
rounded (rarely acute) and 


then rounded, apex 


emarginated or apiculate. Inflorescence a raceme or 


racemose panicle, peduncle with rachis 7-12.5 cm 
with 9 to 17 flowers, peduncle 2-4 mm wide at base, 
peduncle, rachis, and pedicels drying dull grayish 
green, i like those of 
branchlets, bracteoles absent or not seen; calyx 8— 
10 x 
white trichomes tightly appressed, glandular fields 2 


pubescent with trichomes 
8-9 mm, coriaceous, pubescent with minute, 


pairs; corolla white, campanulate from cylindric base, 

ved 1/5 distance from base, 5-8 cm, membranous, 
tube 3.8-6.5 cm with cylindric base 9-13 mm long 
and mouth 20-29 mm wide, 


puberulent with minute trichomes 


externally densely 

above glabrous 

base, m minutely puberulent, with villous 
of trichomes at bases of stamens and 

Peri lobes 10-15 X 15-20 mm 

staminode inserted 8-9 mm from base of corolla tube, 

anthers ca. 5 mm, longer filaments 25-27 mm, shorter 


cluster: 
; stamens and 


filaments 18-21 mm, staminode 5-7 mm; disc ca. 1 

—4 mm; pistil 3.8—4.7 em, ovary 3-4 X 1.5- 
2 mm, stipe 0.3-1 mm, stigma lanceolate or subulate. 
Capsule weakly spatulate, nearly oblong, slightly 
curved, 6—6.7 X 2.5-2.7 cm, ca. 0.4-0.6 cm 


base cuneate and truncate, apex obtuse and apiculate, 


diam., 


one valve concave and one convex, drying red-brown, 
densely pubescent with minute, appressed, golden 
trichomes, midrib conspicuously prominent, wall 1.5— 
2 mm thick. Seeds 9-12 X 26-28 mm, wings well 
demarcated from seed body, light reddish brown or tan 
with tips dirty white, membranous, hyaline especially 
at tips, the veins contrasting in color to surface 


Distribution. Distictella cuneifolia is known from 
the Santa Cruz and Beni departments of Bolivia. It 
is often found growing in savannas or pampas with 


black, 


seasonally inundated or as 


fine, silty soil, frequently described as 
with numerous large 
termite mounds, between 100 and 400 m elevation. 


Figure 1A. 


Phenology. Flowering January, February, and 
May; fruiting July and October. 


Discussion. The only collection information on the 
type of Pithecoctenium cuneifolium is the word, 

abbreviation, “Para.” De Candolle (1845) and Sandwith 
1953) assumed this to be the state of Pará in 
I questioned this, given the 


— 


northeastern Brazil. 
currently known distribution of this taxon. A number 
the Rio 


e department of Santa Cruz, Bolivia, and I 


of recent collections have been made alon 
Paraguá int 
thought that the *Para." on the type specimen might be 
an abbreviated form for this river. Discussions with 
James Solomon and consultations with Goodman (1972) 
and Funk and Mori (1989) led me to conclude that the 
most likely person to have collected along this river 
prior to 1845 was A. D. d'Orbigny. However, while 


Volume 96, Number 2 
2009 


Pool 301 
Review of Distictella (Bignoniaceae) 


Orbigny did travel in this general vicinity, I found no 

mention of the Río Paraguá in Orbigny (1846), and the 

type locality of Distictella cuneifolia remains dubious. 
Distictella cuneifolia i is m similar to D. campinae 


and bifoliolat 


D. laevis. Gentry, following 
Sandwith (1962), identified specimens of D. campinae 
confusion in th 


thor has not 


0 id, whic may cause 


oO 


cunei 
Tropicos database for specimens this aut 
also studied. The most striking difference between D. 
campinae and D. cuneifolia 1s the lateral veins in the 
70°— 


inae and extend in almost a straight line 


leaflets, which are initiated at nearly a right angle ( 
90") 1 

to the margin before looping up to meet the next lateral 

vein. They also differ markedly in the seeds; the wings 

aah greatly reduced, and poorly demarcated 

> the seed body in 


most dissimslar to D. cuneifolia in its very short petioles 
(1—4 mm) and short to absent (to 1 or rarely 2 mm) 


inae. Dicta laevis is 


petiolules. It also appears to never o E 
which are common in D. cuneifolia. Fro 

limited ded observed it ala s seems that D. pu 
has smaller corollas (4.5—5.5 em) than D. cuneifolia. 
Both D. ca 


in habitat; they are found in white sand savannas or 


mpinae and D. laevis differ from D. cuneifolia 


campinas. 


Selected specimens examined. BOLIVIA. Beni: Prov. 
Gral. Ballivián, Riberalta, 160 km hacia Santa Rosa, S. G. 
Beck 20589 (MO). S 

0 


Puesto Pasto, R. Guillén & S. Coria 2111 (MO); Velasco 
Prov., Reserva e UE ^ dr 1500 m al SE pe la 
junta del río Par uillén & V. Roca 3092 
(MO); Velasco, Pargue ae Kempff M. camp. La Tone, R. 
Quevedo et al. 2611 (MO). 


6. Distictella dasytricha Sandwith, Kew Bull. 
1953: 476. 1953 [1954]. TYPE: Brazil. Goiás: 
Mun. Yataí (Jataf), Queixada, 8 July 1949, A. 
Macédo 1906 (holotype, K!; isotype, G!, MO!, US 

!). 


not seen, digitized image! 


Liana; young branchlets terete, drying dull black or 
olden brown, hollow (rarely solid), pubescent with 
es; trichomes golden, t 
mm; pseudostipules occasionally present, oblong to 
falcate, thick, 4-5 X 
trifid tendril uae dal present; petiole 10-20 mm, 
with pubescence like of branchlets; petiolules 7— 
20 mm, with o like that of branchlets; 
leaflets oblanceolate to obovate, 9-15.5 X 5.5- 
cm, subcoriaceous to coriaceous, 4 or 5(6 or 7) 


50. 
spreading trichom e longest l- 


ca. 1 mm. Leaves bifoliolate, 


pairs of lateral veins, each lateral vein initiating at a 
20° angle with the midrib and extending toward apex 
before curving and fading, tertiary veins conspicuous, 
connecting midrib and lateral veins, the higher order 
venation strongly reticulate and forming a fine, closed 


network, all veins immersed adaxially and raised 
abaxially, both surfaces drying brownish green, 
adaxial surface with golden, erect trichomes 

mm, dense to scattered on main veins and 
scattered on surface, abundantly lepidote, with or 
without scattered glands, abaxial surface with dense, 
golden trichomes 0.6-1.1 mm, erect to spreading, or 
pilose, abundantly lepidote, without visible glands, 
margin flat, base cuneate, apex rounded and cuspidate 
or emarginate, or obtuse. Inflorescence a raceme or 
racemose panicle, peduncle with rachis 3-36 cm with 
4 to 18 flowers, peduncle 2-4.5 mm wide at base, 
peduncle, rachis, and pedicels drying brown or golden 
i i branchlets, 


wn, pubescent with trichomes like 
x , lost before anthesis; calyx 


o 
bracteoles ca. 
8-13 x 


dense golden trichomes to 1 mm, erect to ascending, 


10-12 mm, subcoriaceous, pubescent with 


glandular fields 2 pairs; corolla white with yellow 
throat, infundibular from aa base, curved 1/4 
distan rom base, 5-7 cm, chartaceous, tube 3— 
6 cm with cylindric base 10-15 mm long and mouth 
15-25 mm wide, externally and internally densely 
pubescent with minute trichomes above glabrous base, 
internally with villous clusters of trichomes at bases of 
stamens and staminode, lobes 10-17 X 15-25 mm; 
stamens and staminode inserted 10—15 mm from base 
of corolla tube, anthers 4.5—5 mm, longer filaments 
20-25 mm, shorter filaments 16-22 mm, staminode 
ca. 6 mm; disc ca. 2 X 4 mm; pistil ca. 4.1 cm, ovary 
4-5 X 
Capsule elliptic-oblong, curved, 6.8-7.5 


ca. 1 mm, stigma subulate. 
X 33.5 cm, 


0.8-1 cm diam., base obtuse, apex obtuse to rounded, 


2-2.5 mm, stipe 


one valve concave and the other convex, subveluti- 
nous with minute, golden trichomes, somewhat warty, 
especially at base, midrib prominent, wall 1-1.5 mm 
thi eeds 13-15 X 27-35 mm, wings poorly 
demarcated from seed body, golden brown, membra- 
nous, hyaline only at tips, the veins not contrasting in 
color to the surface. 


Distictella dasytricha is known from 
Brazil (Goiás and Acre) and Peru (Loreto). It is 
reportedly found in swampy forested areas at 240 m 


Distribution. 


elevation. Figure 2A. 


Phenology. Flowering July and November; fruit- 
ing October. 

ussion. Gentry identified numerous sterile 

the Amazon (in Col 

Peru, Bolivia, and Venemels) a as Distictella aff. 

and included D. dasytricha in his 


unpublished treatment of Bignoniaceae for the Flora 


aci. throughout ombia, 


dasytricha, 
of Colombia based on one of these collections (A. H. 
087, MO he Fl 


Gentry et al. 9087, ) and in the Flora of the 
Venezuelan Guayana (Gentry, 1997), based on R. 


Annals of the 
Missouri Botanical Garden 


Liesner & B. Holst 21319 (MO). These sterile 
collections differ from the known fertile collections 
are not included in the concept of D. dasytricha 
employed here) in having much larger, membranous 
en sa are much ee pubescent, with the 
or only slightly raised 

ed pem slightly ase adaxially, and form 
very loose network. The leaflet apices also inii to o be 
(1997) suggested that the 


very long att enuate. Gentry 
lection, in ps to several Peruvian 


ry 
Venezuelan col 
collections, might » specifically distinet from true D. 
dasytricha 

Distictella dasytricha is similar to D. reticulata in its 
venation (four or five pairs of lateral veins initiating at a 
20° angle with midrib, higher order venation strongly 
reticulate, forming a very fine, closed network, all 


venation immersed adaxially and raised prd but 
D. reticu. (to c 
and capsules not eurved and witho 


lata has smaller 
ut a uie 
idrib. The only species of Distictella with trichomes of 
similar length those of D. dasytricha is D 
porphyrotricha. Distictella porphyrotricha differs from 
D. dasytricha most markedly in its bullate leaflets with 
much looser venation reticulation. It also has lateral 
veins diverging from the midrib at a wider angle (45°), 
trichomes ferruginous, red, or white, inflorescence with 
persistent bracteoles, and a capsule that is longer (ca. 
curved, with an acuminate apex, 
without a prominent midrib, and is found at higher 
elevations (900-1380 m) in terra firme forests. 


Additional specimens examined. BRAZIL. Acre: Proj. 
RADAM, sub-base de Cruzeiro do Su 


oil pipeline, A. H. Gentry & C. Díaz S. 28225 (MO), 28234 
(MO). 


7. Distictella elongata (Vahl) Urb., Repert. Spec. 
Nov. Regni Veg. 14: 310. 1916. Basionym: 
Bignonia elongata Vahl, Eclog. Amer. 2: 45, t. 
16. 1798. Pithecoctenium | elongatum pe 
Klotzsch in M. R. Schomb., Reis. Br.-Guiana 3 

. Distictis 20 Wah) 
in Benth. & Hook. f., Gen 


1038. 1876. TYPE: French Cuna ae 
d., J. P. B. von Rohr 2001 (lectotype, designated 
by Gentry, 1982: 174, C!, photo F neg. 221301). 


Pithecoctenium obovatum Mart. ex DC. in A. DC., Prodr. 9 

196. 1 a TYPE: French Guiana. s. loc., s.d., J. 

artin s.n. (holotype, BR “Herb. Martius” 
digitized image [UON P not seen). 


not seen, 


Liana; young branchlets terete, flattened at nodes, 
drying dull brown, usually hollow, densely lepidote, 


densely tomentose or with the trichomes erect then 
weakly curving or ascending; trichomes tan, 0.1— 
.3 mm; pseudostipules rarely present, obovate, ca. 5 
X 3.5 mm. Leaves bifoliolate, trifid tendril sometimes 
present; petiole 5-22 mm, with pubescence like that of 
branchlets; petiolules 5-11 mm, with pubescence like 
that of branchlets; leaflets irregularly obovate, 6.5-13.7 
3.5-8.2 cm, chartaceous to subcoriaceous, 3 to 5 
pairs of lateral veins, each lateral vein initiating at a 
20° angle with the midrib and extending toward apex 
before curving and anastomosing with tertiary veins, 


tertiary vein generally closely spaced and 
connecting midrib and lateral veins, to lesser extent 
anastomosing with higher order venation, which form a 
fine, closed network, all veins immersed adaxially and 
raised abaxially, adaxial surface drying brownish 
yellow-green, without trichomes except for midrib and 
apex and sometimes lateral veins, abundantly lepidote, 
with or without scattered glands, abaxial surface 
grayish yellow-green, densely tomentose with white, 
trichomes to 0.1 mm 
tangled, abundantly lepidote, usually w 


often translucent, T and 
fields of 
eral veins with the 
base shortly 
attenuate or cuneate, apex obtuse, rounded, or obtuse 


glands in the axils of most of the lat 

midrib (rarely absent), margin flat, 
nd cuspidate. Inflorescence a racemos d 

(perhaps rarely raceme), peduncle with rachis 

40 cm with 20 to 25 flowers, peduncle 3—4 mm s at 

base, peduncle, rachis, and pedicels drying dull brown, 


es lil 


pubescent with trichom bracteoles ca. 
X 1 mm, lost before anthesis; calyx 10-12 X 10- 
1l mm, coriaceous, pubescent with trichomes tan to 
ferruginous, like those of branchlets, and lepidote, 
glandular fields 2 pairs; corolla white with yellow throat, 
infundibular from cylindric base, curved 1/3 distance 
(4-)5.5-7.2 em, chartaceous, tube 4.5— 
5.2 em long with cylindric base 9-11 mm long and 
19-20 mm ally and internally 
pubescent with minute trichomes above glabro 


from base, 

mouth wide, ext 

us base, 

internally with clusters of glandular papillae (rarely 

villous) at bases of stamens and staminode, lobes 10—20 

x minode inserted 10— 
h 


mm; stamens and sta 
12 mm from base of corolla tube, anthers 


ca. 4 mm, 
longer filaments 22-25 mm, shorter filaments 20— 
2] mm, staminode ca. 6 mm; disc 1-1.5 X ca. 4 mm; 
pistil 3.9-4.7 cm, ovary 4—5 X ca. 2 mm, stipe absent 


] Mint 


to 1 mm, stigma lanceolate. Cap ptic, curved, ca. 


8 X 2.8 em, ca. 1 em diam., base acute and truncate, 
apex acuminate, one is concave and the other 
convex, drying golden-brown, densely pubescent with 
trichomes like those of branchlets, slightly warty, midrib 
prominent, wall ca. 1.25 mm thick. Seeds 8-10 X 25- 
28 mm, wings poorly demarcated from seed body, 
reddish brown with tips lighter brown, membranous, 
hyaline, the veins contrasting in color to surface. 


Volume 96, Number 2 Pool 303 
2009 Review of Distictella (Bignoniaceae) 
Distribution. Distictella elongata is known from specimens of D. elongata are sometimes identified as 


French Guiana, 
slightly atypical collection from Amazonas, Brazil, and 


Suriname, and Pará, Brazil, with one 


another from Rondónia, Brazil. It is found in secondary 
forests frequently characterized as growing on white 
sand and as being on terra firme, at elevations ranging 
from sea level to 200 m. Figure 2A. 


Phenology. Flowering January, February, June, 
and December; fruiting June. 


Discussion. | Pithecoctenium obovatum is placed in 
synonymy of Distictella elongata based on observation 
of the digitized image of the holotype (barcode BR 
880344) and following the synonymy of Sandwith 
(1937, 19382) 

Gentry's concept of the species Distictella elongata 
His 


ns indicates a 


is very different from that a here. 
identification of herbarium spec 
ve concept. Gentry identified as D. elongata 
Pie that are treated here as three different 
species, D. elon, mansoana, and D. parkeri. 


that he cited for D. 


neral ride tion range 
flects this usage and in the 


The 
elongata (1982, 1997) re 
Flora de Venezuela (1982), Gentry listed D. mansoana 
in synonymy biis D. elongata. However, Gentry's 
escriptions and keys in the Flora de Venezuela 
(1982) and Flora e the Venezuelan Guayan 
appear to be 
Distictella a differs from D. elongata most 
markedly s leaflets, which have a rounded to 
cordate f Vie veins immersed or flat on the 
abaxial surface, and lateral veins forming a wider 
angle (45^) to the midrib. In addition, the midrib of 
the fruit of D. parkeri is not at all prominent. The two 
species are found in similar habitats and have 
overlapping ranges of distribution. Distictella man- 
soana differs from D. elongata in generally having 
leaflets with bases usually rounded to subcordate, and 
the tertiary veins usually raised on the adaxial surface 
and bM. with the ou Ape venation to 
s of D. mansoana 
are in e larger (9.5-16 cm long) D and 


form a very loose network. The fru 


warty, and the seeds are much larger (13-20 X 32- 
55 mm) with well-demarcated wings. Lohmann and 
Pirani a and ae Ps A P used 
the . elongata for oana, fo 
Corin gm s. e specimen E 
tions. The range of D. mansoana is generally to the 
south of D. elongata in Brazil, Bolivia, and Peru, and 
it is usually found at higher elevations. 

Distictella elongata is similar to D. dasytricha in 
overall leaflet shape and venation, but the latter has 
much longer trichomes on the branchlets, petioles, 
petiolules, and inflorescence (the longest 1-2 mm) an 
capsules with the apex obtuse to rounded. Small-leaved 


D. obovata, but the latter has coriaceous leaflets with a 
recurved margin and lateral veins at a wider angle (45°) 
with the midrib, flowers in a raceme, and the fruit dull 
black without a prominent midrib. Distictella obovata is 
found at higher elevations (450-1660 m) and appears 
to be restricted to the raima Mountains of Guyana 
and adjacent areas of Bolivar, Venezuela. 

Selected specimens examined. BRAZIL. Amazonas: Rio 


anios ` 4703 MO); 


Ariramba, 


Mun. de Ce campos de 


ndónia: Vilhena, 

4917 (MO). FRENCH GUIANA. Rte. RN2 Cayenne-Regina, 

. Billiet & B. Jadin 5762 (BR); Savane Renner- 

Région littorale, G. Cremers & J.-J. de Granville 14427 (MO); 

Kourou, center s patial, C. Feuillet 1588 (MO). SURINAME. 

Betw. Tawajari-weg & de Crane-weg, W of Lelydorp, N. M. 
Heyde 591 ^N 0). 


8. Distictella laevis (Sandwith) A. H. Gentry, Mem. 
ew York Bot. Gard. 29: 274. 19 
Distictella monophylla var. 
Mem. New York Bot. Gard. 9: 362. 1957. TYPE: 
Venezuela. Amazonas: Cerro Yapacana, Río 
Orinoco, 100 m, 7 Jan. 1951, B. Maguire, R. S. 
Cowan & J. Wurdack 30788 (holotype, NY not 
seen, microfiche 947.A8!, digitized image!). 


Shrub; erect or semi-scandent, 0.5-3 m; young 
branchlets terete (rarely flattened at nodes), drying 
dull brown or black, solid, 


puberulent or glabrescent; 


sometimes lepidote, 
trichomes colorless or 
ferruginous, less than 0.1 mm; pseudostipules cadu- 
cous, clavate, X 0.7-1.2 mm. Leaves 
unifoliolate or bifoliolate, bifoliolate leaves with 
terminal sca , or residual tendril ca. 1 mm; 
petiole 1—4 mm, puberulent with trichomes like those 
of branchlets; petiolules absent to 1(-2) mm, puber- 
ulent with trichomes like those of branchlets; leaflets 
oblong, or sometimes in bifoliolate leaves, oblanceo- 
late to spatulate, 4.2-12 X 


to 7 pairs of lateral veins initiating at a 45° angle with 
b 


1.3-8.3 cm, coriaceous, 4 


the midrib and curving upward toward margin, 
anastomosing with tertiary veins or weakly brochido- 

romous, tertiary veins inconspicuous, loosely open- 
reticulate, all veins immersed adaxially, midrib raised 
abaxially, and other veins immersed to flat (lateral 
veins sometimes slightly raised in older leaves), both 
surfaces 
surface 


puberulent at base of midrib, 


drying grayish or yello na green, adaxial 
sometimes glossy, without trichomes or 
ul scattered 
lepidote, sometimes with scattered glands, abaxial 
surface puberulent at base of midrib or without 


trichomes, scattered lepidote, often with glandular 


Annals of the 
Missouri Botanical Garden 


fields at base and sometimes at apex or glands 
scattered, margin flat (rarely slightly revolute), base 
rounded, subcordate, or cuneate and rounded, apex 
rounded or obtuse, often with small mucro or 
sometimes emarginate. Inflorescence a raceme, or 
rarely with lowest pedicels branching, peduncle with 
rachis 2.5-9 cm with 6 to 16 flowers, peduncle 2— 

mm wide at base, peduncle, rachis, and pedicels 
drying dull reddish or grayish brown, puberulent with 
trichomes like those of branchlets, bracteoles 1-7 X 
5-6 X 5-8 mm, 
membranous, puberulent with minute red or colorless 


0.5—4 mm, early caducous; calyx 


trichomes, glandular fields 1 to 5 pairs; corolla white, 
sometimes with yellow lobes and throat, infundibular 
from cylindric base, curved 1/3 distance from base, 
tube 3.24.5 cm with 
cylindric base 6-9 mm long and mouth 12-20 mm 


4.5—5.5 cm, membranous, 


wide, externally and internally densely puberulent 
with minute trichomes above glabrous base, internally 
with villous clusters of trichomes at bases of stamens 
and staminode, lobes 8-15 X 9-15 mm; stamens and 
staminode inserted 8-10 mm from base of corolla 
tube, anthers 3-5 mm, longer filaments 18-20 mm, 
shorter filaments 12-17 mm, staminode 2.2-5 mm; 
1.54: mm; pistil 2.9-3.3 em, ovary 3-4 X 
1-2 mm, stipe 0.5-2 mm, stigma ovate. Capsule 
oblong, not curved, 5—6.3 X ca. 2.5 cm, ca. 0.3 cm 
thick, base rounded, apex obtuse, valves compressed, 


disc 


drying brown, densely puberulent with minute white 
or golden trichomes, glands scattered, often granular, 
midrib not evident to subevidently raised, wall 1— 
2 mm thick. Seeds 10-14 X 25-28 mm, wings well 
demarcated from seed body, reddish brown with tips 
yellowish tan, membranous, hyaline, T at 
tips, the veins contrasting in color to su 


Distribution. Distictella laevis is known from 
Venezuela (Amazonas, the type) and Brazil (Pará and 
Amazonas). It is often found growing in open campina 


in white sandy soil, below 400 m elevation. Figure 1A. 


Phenology. Flowering January, February, April, 
June, and September; fruiting February, May, and 
September 


Discussion. Unifoliolate specimens of Distictella 
laevis are similar to D. monophylla, which differs 
primarily in the abaxial surface of the leaflets, which 


higher 


closed network. One 


is pubescent with all venation raised and 
venation forming a fine, 
specimen (O. Huber et al. 3771) was seen that seemed 
to be intermediate between D. laevis and D. mono- 


ifoliolate specimens of Distictella laevis are 
both D 


is 
similar to campinae and D. cuneifolia. The 


most striking difference between D. campinae and D. 


laevis is the lateral veins in the leaflets of D. 
campinae, which are initiated at nearly a right angle 
(707—90^) to the midrib and extend in almost a straight 
line to the margin before looping up to meet the next 
lateral vein. They also differ markedly in the dis 
the wings are e greatly reduced, an rly 
demarcated in D. campinae. Dist iella eine 
differs from D. laevis in habitat (black silty soils vs. 
white sand), geography (endemic to Bolivia), longer 
petioles mm) and petiolules (4-10 mm), larger 
corollas (5-8 cm long), and the presence of leaf 
tendrils. 


Selected examined. BRAZIL. Amazonas: 


specimens > 
Humaitá-Ttaituba, Km 135 de 

2 do 

o, M. Alvarenga s.n. (INPA d 


R: Anderson 10928 (MO); Misso 
). 


Cururú, N. A. Rosa & M. R. Sanios 1862 (MO 


9. Distictella lohmanniae A. Pool, sp. nov. TYPE: 


Brazil. Amazonas: Reserva Florestal Ducke, 
anaus-ltacoatiara, Km m, 22535, 
59758' W, 14 July 1995, L. C. Lohmann & C. 


F. da Silva 20 (holotype, MO!; isotype, INPA not 
seen). Figure 3A-C. 


c species Distictellae parkeri affinis, sed ab ea foliolis 


arum 
seminibus alis a corpore bene distinctis praeditis distingui- 
tur. 


Liana; young branchlets terete, flattened at nodes, 
drying grayish green to dull brown, solid, lepidote, 
densely pubescent; trichomes white or ferruginous, to 
0.03 mm long, tightly appressed; pseudostipules 
rarely present, elliptic, thick, ca. 3 X 1 mm. Leaves 
bifoliolate, trifid tendril sometimes present; petiole 6— 

O mm, with pubescence like that of branchlets; 
petiolules 10-31 mm, with pubescence like that of 
ranchlets; leaflets ovate-oblong or obovate-oblong, 
8-15.3 X 4 subcoriaceous, 4 to 6 pairs of 
lateral veins, each lateral vein initiating at a 20^ angle 
with the midrib (except sometimes the basal) and 
extending toward apex, anastomosing with tertiary 
veins and fading, tertiary veins fairly conspicuous, 
connecting midrib and lateral veins, higher order 
venation loosely open-reticulate, midrib and lateral 
veins immersed and tertiary veins slightly raised to 
slightly immersed adaxially, midrib and lateral veins 
raised abaxially, tertiary veins immersed, adaxial 
surface drying dark blackish green, without trichomes, 
abundantly lepidote, without glands, abaxial surface 


Volume 96, Number 2 
2009 


Pool 305 
Review of Distictella (Bignoniaceae) 


drying grayish yellow-green, densely covered with 
minute appressed trichomes like those of branchlets, 
not noticeably lepidote, usually with a few glands in 
the axils of basal lateral veins with the midrib, margin 
flat, base rounded or cuneate, apex rounded or obtuse 
and apiculate to cuspidate. Inflorescence a racemose 
panicle, peduncle with rachis 6-11 cm with 10 to 15 
flowers, peduncle 4-6 mm wide at base, peduncle, 
rachis, and pedicels drying dull brown, pubescent 
with trichomes like branchlets, bracteoles ca. 2.5 X 
10-11 mm, 


ferruginous, 


lmm, lost early; calyx ca. 10 


coriaceous, pubescent with minute, 

appressed trichomes and lepidote, glandular fields 

with yellow throat, infundi 
4 distance con 


say rom cylindric base, curved 1/4: 


pairs; corolla white 
yli 

base, 5-7 cm, eae tube 4—5.2 em with 
cylindric base 9-10 mm long and mouth ca. 20 mm 
wide, externally densely pubescent with minute 
trichomes above glabrous base, internally densely 
lepidote with a few minute appressed trichomes and 
villous clusters of trichomes at bases of stamens and 
staminode, lobes 10-20 X 10-20 mm; stamens and 
staminode inserted 8-9 mm from base of corolla 
tube, anthers ca. 4 mm long, longer filaments 22— 
shorter filaments 19-24. mm 

5 mm; dise ca. 1.5 X 3 mm; pistil ca. 4.5 em, ovary 


staminode ca. 


ca. 5 X 2 mm, stipe apparently absent, stigma 
lanceolate. Capsule elliptic, slightly curved, ca. 9 X 
4 cm, ca. 1.5 cm diam., base acute and truncate, 

x acute, one valve slightly concave and the other 
slightly convex, both somewhat compressed, drying 
blackish green, densely M a with minute, 
appressed, ferruginous trichomes, somewhat glandu- 
lar and warty, midrib prominent, capsule splitting 
along pu at maturity, wall ca. 4 mm thick. Seeds 
4-15 40-48 mm, wings well demarcated from 
seed oa reddish brown with a halo of much lighter 
brown immediately around body and at wing tips, 
a hyaline, the veins contrasting in color 


surface (Figure: Lohmann € Hopkins, 1999: 620 
as i Distctella parkeri 


Distribution. Distictella lohmanniae is known 
from Amazonas, Brazil, where it is found in terra 
firme forests between 50 an Om elevation. 


Figure 2A 


Phenology. Flowering November and December; 
fruiting July. 


IUCN Red List category. Distictella lohmanniae is 
currently known from seven collections. The data are 
deficient (DD) to determine IUCN Red List status. It 

robably not Critically Endangered (CR), as its 


extent of occurrence is estimated to be more than 


100 km? and its area of occupancy is estimated to be 


more than 10 km? (IUCN, 2001) 


Discussion. At least one specimen of Distictella 
lohmanniae, A. H. Gentry 12976, was determined by 
Gentry and distributed to NY as the unpublished 
name D. Gentry. Most of the 
specimens cited as D. d oe were previously 


subincanum A. H. 


determine H. Gen . parkeri, and 
Lohmann and Hopkins T d this taxon under 
that name in their treatment of Bignoniaceae for the 
Flora da Reserva Ducke. The species is named for 
Lúcia Lohmann, who collected the type specimen and 
helped me to recognize D. lohmanniae as distinct from 
similar species 
Distictella lohm 
other species of Distictella by its minute and strongly 


anniae can be recognized from the 


appressed foliar trichomes, similar to those of Distictis 
pulverulenta. It is most similar to Distictella parkeri, 
which has leaflets drying the same color on both 
surfaces with lateral veins at a 45° angle and tertiary 
veins generally branching, and the abaxial surface 
s to 0.15 o 


ger inflorescences (peduncle si rachis 15- 


tomentose with trichome .2 mm long, 


— 
[a] 


35 cm), more strongly curving fruits with the midrib 
not at all evident to subevident, and seeds with the 
wings poorly demarcated. Distictella parkeri is found 
farther north than D. lohmanniae, in the Guianas, 
often in 


Bolívar, Venezuela, and Amapá, Brazil, 


disturbed, white sand forests. 


Additional specimens examined. BRAZIL. Amazonas: 
319, Km 60, Manaus—Porto Velho Rd., betw. Rios 
Castanho & Araga, G. T. Prance et al. 23051 MO); Manaus— 
Igarapé Leão Rd., 5 km from Manaus—Caracarai E 
Prance ei al. 11414 (MO); Km 130, Manaus—Caracarai Rd. 
(BR 174), A. H. Geniry 12976 (MO); Carauari, Poço Juruá, A. 
S. L. da Silva et al. 547 (M 


INPA Overseas Devel- 
opment Admin., s. coll. 3994-24 MO). 


10. L magnoliifolia (Kunth) Sandwith, 
illoa 3: . 1938. Basionym: Bignonia mag- 
noliifolia ae in Humb., Bonpl. & Kunth, Nov. 
Gen. Sp. (quarto ane : 136. 1818 [1819], a 
Bier a : Venezuela. Javita, E 
ipam flum a hadi et Temi (Missiones del 
Orinoco), "Bonpland 973 (lectotype, designated 
by Sandwith, 1938b: 460, P not seen, F neg. 
39415!) 


uso ird Pilg. in Koch-Grünberg, Zwei Jahre unter 
Indianern 2: 372. 1910. Distictella kochii (Pilg.) Urb., 
fs Spec. Nov. Regni Veg. 14: 310. 1916. TYPE: 
Brazil. Hylaea, Rio Aiary, Maloka, Dec. 1903, T. Koch- 
Grünberg 74 (holotype, B presumed destroyed, K 
photo!). 


Annals of the 
Missouri Botanical Garden 


A-C. Distictella lohmanniae A. Pool. 
F. da Silva 20, MO). D, E. Di. 


Figure 3. 
tictella racemo 


Hernandez 535, MO 


Liana; young branchlets terete, the nodes often 
flattened, drying dull black or brown, solid, densely 
lepidote, without trichomes or with few trichomes 
appressed at nodes; trichomes ferruginous, minute; 
pseudostipules rarely present, linear, thick, ca. 
trifid tendril often 
without trichomes or 


Leaves bifoliolate, 
petiole 25-55 mm, 
nearly so, densely lepidote; petiolules 12-40 mm, 
without trichomes or nearly 


present; 


so, densely lepidote; 
leaflets ovate or elliptic, often broadly so, 8-24 X 


—A. Branchlet. —B. Fruit. —C. Seed (A, B, C from L. G. Lohmann & C. 

sa (Bureau & K. Schum.) Urb. var. racemosa. —D. Fru 

T. Prance ei al. 3366, MO). F, G. Distictella n racemosa var. iranslucida A. Po 
)- 


it. —E. Seed (D, E from G. 
ol. —F. Fruit. —G. Seed (F, G from S. Hoyos & J. 


4.5-12.5 cm, coriaceous, 2 to 4 pairs of lateral veins, 
each lateral vein (above basal) initiating at a 45^ angle 
with the midrib and curving toward the apex, 
anastomosing with tertiary veins and fading, tertiary 
veins connecting midrib and lateral veins and weakl 

reticulate with higher order venation forming a loosely 
closed network, all veins raised adaxially or midrib 
and lateral veins flat to immersed, all raised abaxially, 
both surfaces drying yellow-green or brownish green, 
adaxial surface without trichomes or base of midrib 


Volume 96, Number 2 


Pool 307 
Review of Distictella (Bignoniaceae) 


with minute appressed trichomes, abundantly lepi- 
dote, sometimes with scattered glands and glandular 
fields at apex and in the axils of the basal lateral veins 

with the midrib, abaxial surface without trichomes, 
a lepidote, usually with glandular fields at 
apex and in the axils of the basal lateral veins with the 
midrib, sometimes with scattered glands, margin 
slightly recurved, base rounded or subcordate, apex 
acute, obtuse, or rounded and cuspidate. Inflores- 
cence a racemose panicle, peduncle with rachis 12— 
32 cm with 10 to 40 flowers, peduncle 2—4 mm wide 
at base, peduncle, rachis, and pedicels drying black 
or brown, pubescent with minute ferruginous, or 
white, appressed os bracteoles absent or not 
seen; calyx 7-10 coriaceous, pubescent 
with minute, ferruginous, Ae trichomes, glan- 
dular fields 1 to 2.5 pairs; corolla white with yellow 
throat, infundibular from cylindric base, curved 1/3 
distance from base, cm, a tube 3.5- 
5 em with e idos bise 9-12 m 
10695 mm wide, externally and dumis densely 


g and mouth 


pubescent with minute trichomes above glabrous base, 
internally with villous clusters of trichomes at bases of 
stamens and staminode, lobes 12-16 X 12-20 mm; 
stamens and staminode inserted 8-10 mm from 
base of corolla tube, anthers 3—4 mm, longer fila- 
21-25 mm, 17-20 mm, 
staminode 4—7 mm; dise 1-1.5 X 4-5 mm; m 
3.3—4.1 cm, ovary 4-5 X 2-2.5 mm, stipe 0.5-1 m 

stigma subulate. Capsule narrowly oblong, clightly 
curved, 8-13.5 X 2-2.5 cm, 0.5-1.2 cm 


base acute and truncate, apex very long attenuate, 


ments shorter filaments 


diam., 


with one valve slightly concave and one slightly 
convex, eee black, densely Pd with few 
e, ferruginous trichom s, numerous 
glands a granular and wrinkle "n sub- 
der wall 1-2 mm thick. Seeds ca. Te x 42— 
O mm, wings poorly demarcated from seed body, 
light brown with tips dirty white, membranous, hyaline 


scattered, minu 


especially at tips, the veins not contrasting in color to 


Distribution. Distictella magnoliifolia is known 
from Colombia (Vaupés and Guainía), Brazil (Amazo- 
nas), and Venezuela (Amazonas). It is usually found in 
scrubby or low forest, or savanna, on white sand, 


between 75 and 150(300) m elevation. Figure 1B. 


Phenology. Flowering February, April, May, July, 
September, November, and December; fruiting July 
and December 


Discussion. Sandwith (1938b) selected A. Bon- 
pland 973 (P) as the lectotype of Bignonia magnolii- 
folia Kunth. This specimen was not seen in this study, 


but a photo, F neg. 39415, of a P specimen labeled 


“Distictis magnoliaefolia Bur.; Bignonia magnoliae- 
folia Bonpl.; Echantillon del m Bonpland" was 
studied. ile the photo is not labeled with the 
collection number, Sandwith (1953: 478) cites that 
this is a photo of the type. 
Distictis kochii is placed in synonymy of Distictella 
c based on examination of the digitized 
the photo at K of T. Koch-Grünberg 74 (B), 
a synonymy previously proposed by Sandwith 


(1953) 
The concept of Distictella magnoliifolia employed 
here is narrower than that used by Gentry in his 


various publications (Gentry, 1973, 1977, 1982, 


1997) and that used by Burger and Gentry (2000), 
b» follows Sandwith (1953). The pop labele 
. TU ifolia in Gentry (1973, 1977, 1982) depicts 


d flowering branch of F. (a 5140 and 
capsule of B. A. Krukoff 6728, both treated here as D. 
racemosa var. racemosa. The descriptions for the most 
part also pertain to that taxon except for some of the 
vegetative characteristics, which seem to be a mixture 
of this, D. racemosa var. translucida, and a third, 
unidentified species. The specimens cited in the 
Flora of Panama (Gentry, 1973), all of which are 
sterile, are of this latter species. Both varieties of D. 
racemosa differ from D. magnoliifolia in having 
membranous to subcoriaceous leaflets, with (4)5 to 9 
pairs of lateral veins and tertiary and higher order 
venation immersed to flat on the adaxial surface, and 
capsules that are relatively broader (8-22 X 3.5- 
cm), densely pubescent, and with a thicker wall (3— 
5 mm) and an acute to rounded apex. In addition, D. 
racemosa var. racemosa differs from D. magnoliifolia 
in having fruits that are not curved and seeds with the 
wings coriaceous or subcoriaceous, and opaque. 
Distictella racemosa var. translucida differs from D. 
i n e 
trichom 
lateral vane with the midri rarely having 
distinctive glandular fields at both the leaflet apex 
and in the basal lateral vein axils. Both varieties of D. 
racemosa are found in forests, D. racemosa var. 
racemosa in seasonally inundated forests associated 
with black water, and D. racemosa var. translucida in 
non- inundated forest. 
ted in the Flora of Ecuador (Gentry, 
1977) P» “Distictela  magnolifolia (G. Harling & L. 
Andersson 11947 and H. Lugo 3121) pertain to D. 
racemosa var. translucida, and o cited in the 
a de Venezuela y" ry, 1982) are a mixture of D. 
Murena (P. E. Berry Des 1426, 1496; A. H. 
Gentry & S. F 10900; G. Morillo et al. 3904, 
3908, 4199), D. racemosa var. racemosa (L. Williams 
15309; J. J. Wurdack & L. S. Adderley 43008), and D. 
racemosa var. translucida (A. H. Gentry et al. 10608?; 


Annals of the 
Missouri Botanical Garden 


G. Morillo et al. 4134 [cited as 4143]; J. A. Steyermark 
107474). 

The specimens cited in the Flora of Panama 
(Gentry, 1973) for Distictella magnoliifolia (T. B. 
Croat 13183; A. H. Gentry 737, 1861, 1882-b, 4125, 
4265, 7400) have very large membranous to subco- 
riaceous leaflets with the tertiary veins raised on the 
abaxial surface and the midrib, lateral veins, and 
lateral vein axils pubescent with additional tissue 
domatia at the axils of the lateral veins with the 
midrib. The same taxon is known from Costa Rica (A. 
H. Gentry 1116 [MO], 71770-a [MO], and J. F. 

and is treated by Burger and 
00) as D. magnoliifolia. All of this material 


3 
S 
RU 
d 
w 
to 
M 
to 
— 
= 
c 


is a s parts of the valves of three old 
fruits, one with seeds. The fruits are broadly oblong, 
15.5-23.5 X 


brown, pubescent with very small white 


5-5.5 em, slightly curving, drying dull 
trichomes, the 
base and apex acute, the midrib subevident, and the 
wall 5-7 mm thick. The seeds are ca. 22 X 90 mm, 
the wings poorly demarcated, light brown, membra- 
nous, hyaline especially at tips, and without irregular 
ridges and dots of darker color. It is uncertain what 
genus these specimens belong to; the absence of 
ridges and dots on the membranous seed wing would 
be unique in the genus Distictella. 


Selected specimens examined. BRAZIL. Amazonas: Km 
130, Man: s ud d Rd. (BR 174), A. H. Geniry 12969 
(MO); Presidente Figueiredo, La e Balbina on Rio 


Uatumá, ca. 4 im NW . W. Thomas et 


of San Carlos de Rio Negro, R. Liesner 3784 (MO); 8 km from 
San Carlos de Río Negro on rd. to Solano, P. E. Berry 1426-a 
(MO). 


11. Distictella mansoana (DC.) Urb., Repert. Spec. 


157. 1845. Distictis mansoana (DC.) Bureau ex 
B. Verl., Rev. Hort. 40: 154. 1868, as “men- 
soana.” TYPE: Brazil. Near Cuiabá, s.d., A. Silva 
Manso s.n. (holotype, G-DC not seen, microfiche 
9:157.791). 


Liana; young branchlets terete, sometimes flattened 
at nodes, drying dull brown, black, or gray, usually 
solid, densely lepidote, pilose, tomentose, or tri- 
chomes erect to ascending; trichomes dull white, 
ferruginous, or tan, 0.1-0.7 mm; pseudostipules rarely 
present, obovate or linear, thick, 3—7 X 1-1.5 mm 
Leaves bifoliolate, trifid tendril sometimes present; 
i ike that of 
branchlets; petiolules 5—25 mm, with pubescence like 


petiole mm, with pubescence like 


that of branchlets; leaflets ovate, 
elliptic, 5-17 X 
of lateral veins, each lateral vein initiating at a 207— 
30*(—45^) angle with the midrib and extending toward 
apex before curving and anastomosing with tertiary 


lanceolate, or 
3-12 cm, chartaceous, 3 to 6 pairs 


veins, tertiary veins inconspicuous, generally anasto- 
mosing with higher order venation to form a loosely 
closed network, midrib and lateral veins immersed 
adaxially and tertiary veins usually raised, all veins 
raised abaxially, both surfaces drying olive-green or 
yellowish green, adaxial surface with trichomes dense 
along midrib, dense to scattered on lateral veins, and 
sometimes scattered over surface, abundantly lepi- 
dote, with or without scattered glands, abaxial surface 
pilose to tomentose with white trichomes mm, 
abundantly lepidote, usually with fields " aln: in 
the axils of the basal lateral veins with the midrib 


— 


pi absent) and sometimes in additional m 


margin flat, base rounded, or nearly so, or subcordat 
hardly somewhat acute), apex acute, E Nid 
or obtuse-cuspidate. Inflorescence a racemose panicle 
(rarely raceme), peduncle with rachis 6—45 cm with 4 
to owers, peduncle 2-5 mm wide at base, 
peduncle, rachis, and pedicels drying dull brown, 
black, or gray, minutely puberulent with trichomes 
erect to ascending, ferruginous or w sometimes 
with trichomes like those of en rie 4— 
8 X 1-3 mm, lost before anthesis; calyx 10-15 X 9— 
12 mm, Mite pubescent with minute, ap- 
sed, o ferruginous trichomes and lepidote, 
RA p or 2.5 pairs; corolla white with 


yellow throat (corolla reportedly purple, F. 


coriaceous, tube 3.5-6.7 em with cylindric 
base 9-17 mm long and mouth 17-30 mm wide, 
externally and internally pubescent with minute 
trichomes above glabrous base, internally per at 
ases of s ns and staminode, lobes 13-24 X 13- 
25 mm; stamens and staminode inserted 8-16 mm 
from base of corolla tube, anthers 4—5 mm, longer 
filaments 24—32 mm, shorter filaments 18—26 mm, 
staminode 6—9 mm; disc 1-2 X 5 mm; pistil 3.7— 
2-2.8 mm, stipe 0.5-1.5 mm, 


stigma lanceolate or o Capsule elliptic or 
6 X 2. 5 


4.9 cm, ovary 3-6 X 


oblong, curve .5— ./ em, 1.2-2.5 cm 
diam., base acidic and truncate, acute, or rounded, 
apex obtuse, acute, or acute or obtuse and apiculate, 
one valve concave and the other convex, drying golden 
brown, densely tomentose with minute white, ferrugi- 
nous, or tan trichomes, glandular and usually warty, 
midrib prominent, wall 2-3 mm thick. Seeds 13-20 X 
8-55 mm, wings well demarcated from seed body, 
reddish brown with tips lighter brown, membranous, 


hyaline, the veins contrasting in color to surface 


Volume 96, Number 2 
2009 


Pool 309 
Review of Distictella (Bignoniaceae) 


(Figure: Lohmann & Hopkins, 1999: 620 as Distictella 
elongata; Lohmann & Pirani, 1998: fig. 12f-l as 
Distictella elongata; Bureau & Schumann, 1896: fig 
87) 


Distribution. | Distictella mansoana has primarily 
been collected in central Brazil (Minas Gerais, Sao 

aulo, Goiás, Distrito Federal, and Mato Grosso) and 
Bolivia (along the border with Brazil in Santa Cruz 
and Beni, and also La Paz) Additional collections 
have been found in Brazil in Bahia, Pará, Amazonas, 
Roraima, and Rondónia and in Peru along the border 
with Bolivia in Madre de Dios and Puno. It is most 
often reported from cerrado or pampas, but it is also 
reported from disturbed areas, gallery forest, and 
moist forest. It is generally found at elevations over 
400 m, but has been collected from 160 to 1400 m 


elevation. Figure 


Phenology. Flowering primarily from November 
to February. There are also collections in flower from 
March, May, June, August, and October; fruiting from 
April to November. 


(1868) published “Distictis 
” This could be interpreted as a 


Discussion. Verlot 
Mensoana, Ed. Bur. 
new species based on J. Correa de Méllo specimens 
collected in Sáo Paulo, Brazil, and housed at 
However, digitized images of J. Correa Mello 13 (P) 
were seen in this study, and the specimens were 
annotated by Bureau as Bignonia mansoana DC. and 
Distictis mansoana "Bur." From this it can be deduced 
that “Mensoana” 1 ographic error for man- 
redit Bureau with 
the transfer of Bignonia mansoana (based on A. Silva 


a o 
soana and Verlot's intention was to c 


Manso s.n.) to. Distictis. This transfer is often cited 
Distictis mansoana (DC.) 
f., Gen. Pl. 2: 
1 or Distictis mansoana (DC.) Bureau, 
Vidensk. Meddel. Dansk Naturhist. Foren. Kjgben- 
havn 1893: 112. 1894. 

In Flora de Venezuela, Gentry (1982) treated 


Distictella mansoana as a synonym of D. elongata. 


from later publications: 
Bureau ex Benth. in Benth. 00 
38. 187 


Specimens identified by Gentry as D. elongata are 
treated here as three different species, D. elongata, D. 
mansoana, and D. parkeri. However, Gentry's use of 
the name D. elongata in the descriptions and keys in 
the Flora de Venezuela (1982) and Flora ia the 
Venezuelan di es (1997) appears to be 
on specimens of D. parkeri. Distitella m differs 
from D. mansoana in the tertiary veins of the leaflets 
being immersed in the adaxial surface and flat to 
immersed on the abaxial surface, the midrib of the 
capsule not evident or only subevident, and the wings 
of the seed 


differs from D. mansoana in having obovate leaflets 


poorly demarcated. Distictella elongata 


with bases shortly attenuate or cuneate, the tertiary 
veins immersed in the adaxial surface and generally 
closely spaced and connecting midribs and lateral 
veins without branching, while the higher order 
venation forms a fine, closed reticulation. The fruits 
of D. elongata are somewhat smaller (ca. 8 X 2.8 cm), 
eglandular, and oe slightly warty, and the seeds are 
smaller (8-10 
wings. Both D. elongata and D. parkeri are found 


X 25-28 mm) with poorly demarcated 


north of D. mansoana in second 
forests on white sand. Lohmann and Pirani (1998) and 
Lohmann and Hopkins (1999) u: the name D. 
elongata for D. mansoana, following Gentry’s (1982) 


ary or disturbed 


synonymy and specimen identifications. 

Distictella mansoana is also similar to, and 
sometimes confused with, D. arenaria. Distictella 
arenaria differs from D. mansoana in having conspic- 
uous, broader, and more persistent pseudostipules (3— 


= 
o 


3-5 mm), leaflets with the tertiary veins 
immersed on the adaxial surface, petiolules usually 
longer than the petioles, capsules thinner (ca. 0.3 e 
diam.) with the midrib not or only slightly evident, and 
seeds with the wings poorly demarcated. It is also 
found north of D. mansoana on white san 
Distictella mansoana and D. magnolia are the 
only species in the genus with the tertiary veins raised 
on both the adaxial and abaxial leaflet surfaces. 
e two are quite different in other n 
Distictella magnoliifolia has coriaceous leaflets with 
the abaxial id the 
capsules dry black, are nearly without trichomes, 


surface without trichomes 


and have long attenuate apices. 


Selected specimens examined. BOLIVIA. Meg 15 km 
SW of Guajará-Mirim on rd. to Riberalta, W. R. An 

11965 (MO). La Paz: Prov. Larecaja, 19 km 
Guanay por el camino a Tipuani, J. C. Solomon 17685 (MO); 
Nor Yungas, 21.1 km al NO del camino entre Yolosa y 
Caranavi por el camino a 


Bahia: Mpio 


us et al. 11445 (MO); Corrego Jeriva, E of Lagôa Pavano, 
H. S. Irwin et al. 15389 (MO). Goiás: Rio Cristal 44 km by 


rd. c of Cristalina, W. R. Anderson et al. 8273 (MO); Rio S. 
, Mun. ud E Haischbach & T. P. Rama- 
pao 38187 (MO). Grosso: ca. 20 km S of 


Xavantina, H. S. Irwin et P 16821 (MO); 10 km W de 
Chapada dos Guimaráes, S. Ferrucci 795 (MO); Porto XV, 
Mun. Bataguagu, G. Hatschbach 23542 (MO). Minas Gerais: 
ca. 40 km NE 


J. H. Kirkbride Jr. & E. Mees 2860 (MO); Alto Tapajés, Rio 
Cururú, region of Missão Velha, a Mundurukú village, ca. 


2 km N of the Rio Cururú, W. R. Anderson et al. 10945 (MO). 


Annals of the 
Missouri Botanical Garden 


Rondônia: Mun. de Colorado do Oeste, BR 364, Porto 
Velho-Cuiabá, estrada p Colorado do Oeste, Km 25, C. A. 
Cid Ferreira et al. 4307 (MO); Estrada Vilhena-Pimenta 
Bueno entre os Kms 640 e 645, Mun. de Vilhena, M. G. 
Vieira et al. 990 (MO). Roraima: Indian trail from Surucucu 
to Uaicá, betw. Maitá & Paramiteri "n s y T. 
Prance ei al. 10659 (MO). Sao Paulo: 

Cachoeira, Mun. de Piracicaba, F. R. Fosberg hor (MO). 


o, R. Pearce s.n. nou 


12. Distictella monophylla Sandwith, Mem. New 
York Bot. Gard. 9: 361. 1957. TYPE: Venezuela. 
Amazonas: Cerro Sipapo (Paráque) Camp S 
vanna, 1500 m, 15 Dec. 1948, B. Maguire & L. 
Politi 27717 (holotype, K not seen; isotype, NY 
not seen, microfiche 947.A71, digitized image!). 


rub, erect to semi-scandent, 1-2.5 m; young 
branchlets terete, drying dull brown or black, solid, 
sometimes lepidote, densely pubescent with trichomes 
usually tightly appressed (rarely ascending) tri- 
chomes dull white, ca. 0.3 mm; pseudostipules absent 
or not seen. Leaves unifoliolate, without tendril or 
tendril scar; petiole 3-10 mm, pubescence like that o 
branchlets; leaflets oblong (rarely ovate), 3-9.5 X 
1.7—1.5 em, coriaceous, 3 to 5(6) pairs of lateral veins 
initiating at a 45^ angle with midrib or less and 
extending toward apex, anastomosing with tertiary 
veins and fading, tertiary veins conspicuous abaxially, 
anastomosing with high order venation to form a fine, 
closed network, midrib and lateral veins usually flat 
adaxially and tertiary veins raised, occasionally 
tertiary veins flat or all immersed, all raised abaxially 
(except sometimes flat in young leaflets), both 
surfaces drying blackish green (rarely pale green), 
adaxial surface with scattered, appressed trichomes, 
or trichomes restricted to main veins, or without 
trichomes, scattered lepidote, sometimes with scat- 
tered glands, and glands concentrated at apex and/or 
base, abaxial surface with small, i white 
trichome an im eins 

veinlets, per REANO into areoles SP ee 
also in areoles), lepidote, without obvious glands, 
margin revolute, base rounded or subcordate, apex 


Inflorescence 
panicle), peduncle with rachis 2-4.5 cm with 1 to 


rounded or obtuse, often apiculate, sometimes emar- 
a raceme (rarely racemose 


ginate. 


6(20) flowers, peduncle 1-2 mm wide at base, 
peduncle, rachis, and pedicels drying gray, pubescent 
with trichomes like those of stem, bracteoles absent or 
not seen; calyx 6-8 X ca. 7 mm, membranous, 
pubescent with minute white, appressed trichomes, 
glandular fields 2 to 3.5 pairs; corolla white, pale 
purple, or white with purple lobes, sometimes with 


yellow throat, infundibular from cylindric base, 


weakly curved ca. 1/2 distance from base, 4.5-6 cm, 
membranous, tube 3.6—4.1 cm with cylindric base 5— 

mm long and mouth 15-20 mm wide, externally 
densely puberulent with minute trichomes above 
glabrous base, internally puberulent and lepidote 
with villous clusters of trichomes at bases of stamens 
and staminode, lobes 0—15 mm; stamens 
and staminode inserted 4.5-8 mm from base of corolla 
tube, anthers 3.5—4 mm, longer filaments 22-23 mm, 
shorter filaments 17-18 mm, staminode 2-3 mm; disc 
3 X ca. 
1.2 mm, stipe not distinct, stigma lanceolate. Capsule 


X 2-2.6 cm, ca. 0.2- 


iam., base rounded, apex attenuate, valves 


A. mm; pistil 3-3.3 cm, ovary 2.5— 


elliptic, not curved, 4.2-7 
0.3 cm 
compressed, s 


chomes, scattered glands, often with granular appear- 
ance, midrib not evident, wall 1—1.5 mm thick. Seeds 
10-15 x 23-32 
seed body, reddish brown with tips yellowish tan, 


mm, wings well demarcated from 
membranous, hyaline at least at tips, the veins 
contrasting in color to surface 


Distribution. — Distictella monophylla is known 
from Amazonas, Venezuela, and (reported by Sand- 
with, 1957) from adjacent areas in Colombia. It is 
often found in white sand savannas, between 100 an 


1500 m. Figure 1A. 


Phenology. Flowering February, March, May- 
July, November, and December; fruiting February, 


March, July, and October. 


Discussion. The only other species of Distictella 
known to have unifoliolate leaves is D. laevis (first pair 
unifoliolate in D. campinae) 


These are also the only two species of Distictella that 


of leaves sometimes 


are ever erect shrubs. They differ from each other 
primarily in the abaxial surface of the leaves, without 
trichomes in D. laevis with the tertiary veins immersed 
to flat and loosely open-reticulate and the margin flat. 
Their geographic ranges and Re overlap, but D. 
laevis has not been collected abov 

Leaves of Distictella dime are asse to the 
leaflets of the bifoliolate liana, D. obovata. However, the 
species differ not only in habit and leaflet number but 
also leaflet shape (generally obovate or oblanceolate in 
D. obovata), base (cuneate in D. obovata), surface drying 
color (bicolored and glossy adaxially in D. obovata), and 
capsule drying color (dull black in D. obovata). 


Selected specimens examined. VENEZUELA. Amazo- 
of Budare, o on S bank 
Berry et 


o Basso, G. Aymard & L. Delgado 
8385 (MO); Dpto. does 10 km al S i Río Autana y 15 km 
al SW del Cerro Autana, O. Huber 4061 (MO); Dpto. Río 


Volume 96, Number 2 
2009 


Pool 311 
Review of Distictella (Bignoniaceae) 


Negro, vertiente oriental del Macizo Aracamuni, O. Huber & 


E. Medina 5892 (MO 


13. Distictella eis pap Mem. New York 
Bot. Gard. 9: 362. 1957. TYPE: Guyana. Upper 
Mazaruni River, Kataima, 500 m, 17 Nov. 1951, 
B. Mni & D. B. Fanshawe 32637 (holotype, 
K not seen; is , NY not seen, microfiche 
lol mu pud US not seen, digitized 
image!). 


Liana; young branchlets terete, drying black (rarely 
brown-green), solid, lepidote, pubescent with tri- 
chomes appressed to spreading; trichomes white- 
Peer 0.1-0.3 mm; pseudostipules rarely pres- 
e or clavate, 4-6 X 0.5-1 
hilt trifid tendril sometimes es petiole 3— 

b that of branchlets; 
ae 6-12 mm, with de ae like that of 


ent, su mm. Leaves 


m, with pubescence like 


branchlets; leaflets obovate or oblanceolate (rarely 
elliptic), 2.8-9(11.2) X 1.5-4(4.8) em, coriaceous, 3 
or 4(5) pairs of lateral veins, each lateral vein 
initiating at a 45° angle with the midrib and curving 
toward the apex before fading, tertiary veins conspic- 
uous, but not always distinct from higher orders of 
venation, strongly reticulate with higher order vena- 
tion forming a fine, closed network, all veins 
immersed adaxially and raised abaxially, adaxial 
surface drying glossy, black or brown-green, generally 
without trichomes or nearly so except along midrib 
(and sometimes trichomes scattered on surface), 
scattered lepidote, usually without glands, abaxial 
surface grayish brown-green or grayish yellow-green, 
with d of small white-translucent trichomes 
erect then curving primarily in areoles (rarely evenly 
distributed), pero lepidote, without glands, or 
glands scattered in basal area (rarely concentrated 
into fields), 
rounded, rounded and apiculate (rarely obtuse and 


margin recurved, base cuneate, apex 
cuspidate). Inflorescence a raceme, peduncle wit 
rachis 15-20 em with ca. 10 flowers, peduncle ca. 
2 mm wide at base, peduncle, rachis, and pedicels 
drying black, pubescent with trichomes like those of 
branchlets, bracteoles absent or not seen; calyx 8-10 
mm, membranous, pubescent with minute, 
white, appressed trichomes and lepidote, glandular 
fields 2 or 2.5 pairs; corolla white with yellow throat, 
infundibular from cylindric base, curved 1/3 distance 
from base, cm, chartaceous, tube (3.5— 
5 em with cylindric base 8-15 mm 
(1015-25 mm wide, 


densely pubescent with 


long and mouth 
externally and internally 


minute trichomes above 


glabrous base (or rarely internally lepidote), internally 
with villous clusters of trichomes at bases of stamens 


and staminode, lobes 15-20 X 


15-22 mm; stamens 


and staminode inserted 7-10 mm from base of corolla 
tube, anthers 3—4 mm, longer filaments 20-24 mm, 
shorter filaments 16—20 mm, staminode 5—6 mm; disc 
4 x 
2.5 mm, stipe not distinct, o lanceolate. Capsule 
(4389 X (1.892.8- 


3 cm, ca. 0.4 cm diam., base acute and truncate, apex 


ca. X 4 mm; pistil ca. 3.7 ovary ca. 


narrow-elliptic, not curved, 
acuminate, both valves somewhat compressed, drying 


dull bla 


trichomes, very warty and somewhat wrinkled, midrib 


ck, pubescent with minute, appressed, white 


subevident, capsule splitting along midrib at maturity, 
wall 1.5-2 mm thick. Seeds (8210-18 X 30—40 mm, 
wings poorly demarcated from seed body, reddish 

rown with tips lighter, membranous, hyaline, espe- 
cially at tips, the veins contrasting in color to surface. 


Distribution. | Distictella obovata is known from the 
region of the Pakaraima Mountains of Guyana, and 
across the border into Bolívar, Venezuela. It is found 
in savannas and disturbed or low forests from 450— 


1660 m elevation. Figure 1A. 


Phenology. Flowering May, June, and October; 
fruiting May. 


Discussion. Leaflets of Distictella obovata strongly 
in ieee 


1957) 


obovata might prove to be a 


resemble the leaves of D. monophy 
pubescence, margin, and apex, and Sandwith 


simple-leaved, 
shrubby D. monophylla. However, the two species 
vary not only in leaflet number and in habit but also 
leaflet shape (oblong or ovate in D. monophylla), base 


— 


rounded or subcordate in D. monophylla), surface 
drying color (both surfaces similar in D. monophylla), 
and capsule color (brown in D. monophylla). Di- 
stictella monophylla is endemic to Amazonas, Vene- 
zuela, and adjacent areas in Colombia (Sandwith, 
1957) and is found in white sand savannas 

Sandwith (1957) also compared Distictella obovata 
to D. cuneifolia based on their similar leaflet shapes. 
Distictella cuneifolia has leaflets without trichomes 
and with more lateral veins (5 to 7 pairs), the higher 
order venation flat to immersed in the abaxial surface, 
and a generally flat margin and reddish brown 
capsules with a prominent midrib. Distictella cunei- 
folia is endemic to Bolivia where it is found in pampas 
or savannas with black, silty so 

The shape of the leaflets of Dde obovata and 

D. elongata can be similar, and these specimens are 
frequently misidentified. The leaflets of D. elongata 
are chartaceous to subcoriaceous, tomentose on the 
abaxial surface, the margin always flat, and the lateral 
p (20°) angle. In addition, the inflores- 
its are golden-brown, 


veins at a stee 
cence is de and the fru 
curved, and have a prominent midrib. Distictella 


Annals of the 
Missouri Botanical Garden 


elongata is also found at lower elevations (sea level to 
200 m), east of the range of D. obovata, in secondary 
forests growing on white san 


Selected specimens examined. GUYANA.  Pakaraima 
Mtns., Mt. Aymatoi, M. Maas et x: 5711 (MO); 
Pakarina Mtns., Mora forest; N of Kamarang, P. J. M. d 
et al. 4307 (MO); upper Mazaruni River basin, Mer 
Mins., Imbaimadai savannahs, outh of Partang River, 
S. S. Tillett & C. L. Tillett 13837 (MO); Cuyuni-Mazaruni 
Region, vic. of Utshe River, T. McDowell & D. Gopaul 2864 
(MO). VENEZUELA. Bolívar: Quebrada El Cajón, Puen P. 
Luis Raúl Vásquez Z., km E de Icabará, J. 
ru et al. 11 7818 (o; "a Piar, Río Apa 
Kambay-mert rapids, of 


tepui, R. Liesner de B. EN Dm 0). 


aramán, 
corner Amaruay 


14. Distictella parkeri (DC.) Sprague & Sandwith, 
1932: 90. 1932. 


rara, Parker s.n. 


P 
seen, miičrofičhë G-DC 9.157. "t ches F neg. 
76311). 


dr Laggan Klotzsch ex Bureau & K. Schum 
Fl. Bras. 8(2): 176. 1896; Dida guianensis 


. Schomburgk 1709 
Über B eure destroyed). 
Liana; young branchlets terete, flattened at nodes, 
drying brownish green to dull brown, usually hollow, 
lepidote, densely pubescent with trichomes ascend- 
ite or fe 


ing; trichomes dull w erruginous, to 0.06— 


.l mm; pseudostipules rarely present, falcate or 
clavate, thick, 4.5-8 X 2- 


trifid tendril sometimes present; petiole 13-35 


4 mm. Leaves bifoliolate, 
50) 


branchlets; 


=~ 


mm, with pubescence like 
petiolules 13-25 mm, with pubescence like that of 
branchlets; leaflets ovate, obovate, or elliptic-oblong, 
9.7223 X 4.8-12.5 em, 


chartaceous or coriaceous), 4 to 6 pairs of lateral 


subcoriaceous (rarely 
veins, each lateral vein initiating at a 45° angle with 
the midrib and curving toward the apex, anastomosing 
with tertiary veins and fading, tertiary veins incon- 
spicuous, weakly connecting cordel and lateral veins 
and loosely reticulate with higher order venation, all 
veins immersed adaxially, midrib and lateral veins 
raised abaxially, tertiary veins flat to immersed, 
adaxial surface drying brown-green, brownish yel- 
low-green, or dark brownish green, without trichomes 
except for midrib and sometimes lateral veins (rarely 
scattered trichomes on surface), abundantly lepidote, 
usually with scattered glands, abaxial surface drying 
brownish green, brownish yellow-green, or grayish 
yellow-green, densely tomentose with white trichomes 
to 0.15 or 0.2 mm, abundantly lepidote, with fields of 


glands in the axils of the basal lateral veins with the 
midrib, often also some scattered glands especially at 
apex, margin flat, base rounded, subcordate, or 
cordate, apex obtuse and cuspidate (rarely apiculate 
or rounded and apiculate). Inflorescence a racemose 
panicle, peduncle with rachis 15—35 em with 17 to 34 
flowers, peduncle 3-4 mm wide at base, peduncle, 
rachis, and pedicels drying dull brown or brownish 
green, pubescent with trichomes like those of 
branchlets or trichomes smaller, bracteoles 2-5 X 
ca. 1 mm, lost prior to anthesis; calyx 10-13 X 10- 
11 mm, coriaceous, pubescent with minute, ferrugi- 
nous or white, appressed trichomes and lepidote, 
glandular 


throat, fiv from cylindric base, curve 


s 2 pairs; corolla white with yellow 

d 1/4— 
1/2 distance from base, 5.5-8 cm, subcoriaceous, tube 
4.3-6 cm with cylindric base 10-13 mm long and 
mouth 20—25 mm wide, 
densely pubescent with minute trichomes above 


externally and d 


om 
glabrous base, nis with e e of 
trichomes at bases of stamens and staminode, lobes 
12-20 X 13-20 mm; stamens and staminode inserted 
8-10 mm from base of corolla tube, anthers 4—5 mm, 
longer filaments 23-29 mm, shorter filaments 19— 
mm, staminode 2-6 mm; disc ca. 
pistil 3.94.6 cm, ovary 4—4.2 X 1.5-2.5 mm, stipe 
0.5-1.5 mm, stigma lanceolate. Capsule oblong, 
strongly curved, 12-12.5 X 3.5-3.6 cm, ca. 1.3 cm 
diam., base acute and truncate, apex acute or acute and 
rounded, with one valve concave and the other convex, 
drying reddish brown or golden, densely pubescent 
with trichomes like leaflets or golden or red, with 
numerous glands overall or glands numerous but 
restricted to margins, warty or granular (sometimes 
only on margins), midrib not at all or only subevident, 
capsule sometimes splitting along midrib at maturity, 
wall 2—4 mm thick. Seeds 15-20 X 55-62 mm, wings 
poorly demarcated from seed body, reddish brown with 
tips lighter brown, m hyaline, the veins 
surface (Figure: Gentry, 1997: 
). 


contrasting in color 
fig. 375, as Dom Pm a 


Distribution.  Distictella parkeri is known from the 
Guianas, Bolívar, Venezuela, and Amapá, Brazil. It is 
generally found in disturbed forest, often described as 
on white sand, at elevations between sea level and 


500 m. Figure 1B. 


Phenology. Flowering January, March, June, Sep- 
tember, October, and December; fruiting March and 
May. 

Discussion. Sprague and Sandwith d studied 
the holotype of Distictis guianensis an 

onym of Distictella rue This 
wie is supported by the original 


guianensis as a 
placement, 


Volume 96, Number 2 
2009 


Pool 313 
Review of Distictella (Bignoniaceae) 


description of Distictis guianensis (Bureau & Schu- 
mann, 1896), is followed here. 

ictell | is treated in fee 
Venezuela (Gentry, 1982; based . Gentry et 
al. 10660 and J. A. Steyermark Ra not ie and 
the Flora of the Venezuelan Guayana (Gentry, 1997) 
leaflets with 


attenuate or cuneate bases, lateral veins at a steep 


as D. elongata, a species that has 


(20°) angle to the midrib, abaxial surfaces with a 
inely closed network of veins, and 

with a prominent midrib 
Distictella parkeri is closest to D. lohmanniae, 
which differs in having leaflets that are bicolored, the 
adaxial surface drying dark blackish green, the 
abaxial surface covered with appressed, minute 
trichomes to 0.03 mm long, lateral veins forming a 
steep (20°) angle with the midrib, 
aa) parallel, closely set, and nonbranching, 
maller inflorescences (peduncle with rachis 6— 
i cm), slightly curved fruit with a prominent midrib, 


tertiary veins 


and seeds with the wings well demarcated. Distictella 
lohmanniae is endemic to Amazonas, Brazil. 

Another similar species is Distictella mansoana, 
treated by Gentry as a synonym of D. elongata 
(Gentry, 1982). Distictella mansoana differs from D. 
of the leaflets being 
usually raised on both surfaces, the very elevated and 


parkeri in the tertiary veins 


swollen midrib of the capsule, and the well-demar- 
cated w of the seeds. Distictella mansoana is 
pe ie eu of D. parkeri 
Distictella parkeri can also be compared with D. 
racemosa var. translucida. The latter has the abaxial 
surface of the leaflets generally with trichomes 
ils of the | i 


midrib and occasionally with trichomes also on the 


restricted to the axils ateral veins with the 
major veins, or rarely without trichomes, capsules with 
a prominent midrib, and leaflet bases acute, obtuse, or 
rounded 


Selected d. BRAZIL. Amapá: Mun. de 
Calcoene, BR156 i in vic. of government rd. c 


crossing of Riviere La Comte, A. H. 
50256 (MO); Bord de la rte. de l'Est, Km 38, M.-F. Prévost 
428 (MO). uyuni—Mazaruni Region, alone 
Mazaruni River, confluence with 
3 km upstream, T. McDowell & D. Gopaul 2575 (MO); upper 
Demerara-Berbice Region, Fairview Landing E bank of 
Essequibo River, near end of Mabura rd., T. MeDowell 3250 


n, H. ier Steege et al. 398 (MO); Potaro— 
Siparuni Region, ices Natl. Park, betw. airstrip 

e L. J. Gillespie 902 (MO). ON Maro- 
acca Village, 25 km S of 


istr.: Vic. of Patam 


Mön ngo, B. Hoffman & M. van inaa 5355 (MO). 
VENEZUELA. Bolívar: Tumeremo to Anacoco (N side of 
Cuyuni River), 61 | at Anacoco, A. H. 


Gentry et al. 10660 A 


Flora of 


15. Distictella pauciflora A. H. Gentry, Ann. 
Missouri Bot. 


106343 (holotype, MO). 


Liana; young branchlets terete, drying grayish 
green, solid, densely lepidote, pubescent with ap- 
o trichomes or glabrescent; trichomes dull 
white mm; pseudostipules absent. Leaves 
bifelilate, i of tendril not seen; petiole 0 


with dense, white and ferruginous trichomes to 
0.03 mm, lepidote; petiolules 4—7 mm, with pubes- 
cence like petioles; leaflets elliptie or 
lanceolate-elliptie, 5.5213 X 1.44 


ceous, 4 to 7 pairs of lateral veins, each lateral vein 


.2 em, subcoria- 


initiating at a 45^ angle with the midrib and curving 
toward the apex, anastomosing with tertiary veins and 
fading, tertiary veins inconspicuous, connecting 
midrib and lateral veins and weakly reticulate with 
igher order venation forming a loosely closed 
network, all veins immersed adaxially, midrib and 
lateral veins raised, and tertiary veins flat to immersed 
abaxially, both surfaces drying brownish green, 
adaxial surface, without trichomes, abundantly lepi- 
dote, without scattered glands, abaxial surface without 
trichomes, densely lepidote, with scattered glands, 


glandular fields sometimes in basal lateral vein axils 


flower from axil of uppermost leaf, peduncle with 
pedicel 2.5-5 cm, peduncle ca. 1 mm wide at base, 
peduncle and pedicels drying grayish green, pubes- 
cent with trichomes like branchlets, bracteoles absent 
or not seen; calyx 8-9 X 7-8 mm, coriaceous, 
pubescent with scattered minute, ferruginous, ap- 
pressed trichomes and densely lepidote, glandular 
fields at least 1 pair; corolla white with pale lavender 
lobes, infundibular from cylindric base, curved 1/4 
distance from base, 5-6 cm, chartaceous, tube 3.5— 
4.5 em with cylindric base ca. 10 mm long and mouth 


staminode inserted 7-10 mm from base of corolla 
tube, anthers ca. 4 mm, longer filaments 19-21 mm, 
shorter filaments 17-19 mm, staminode at least 2 mm 
(broken); disc ca. 1 X 3 mm; pistil 
ca. 4 X 1.5 mm, stipe ca. 1 mm, stigma lanceolate. 


3.6-3.7 cm, ovary 


Capsule not known. 
Distribution. — Distictella pauciflora is known only 
from the type (Bolivar, Venezuela). Figure 1B. 


Annals of th 
Missouri Botanical Garden 


Phenology. Flowering July; fruiting not known. 


Discussion. — Distictella pauciflora is most similar 


to narrow-leaved specimens of D. racemosa var. 
racemosa (such as the types of its synonyms Bignonia 
and Distictis angustifolia) but 
and delicate 


inflorescence. Distictella pauciflora is unusual in this 


rusbyi Britton ex Rusby 
differs from these in its greatly reduce 
generally E mis genus in having 
corola lobes, though D. 
described as Mes pale purple corollas (at least in 


lavender 
monophylla is frequently 


part), one specimen of D. racemosa var. translucida 
was reported to have corollas with lavender lobes, and 
one specimen of D. mansoana was described as 
having purple corollas. Curiously, the illustration in 
the Flora of the Venezuelan Guayana (Gentry, 

with at least 
seven flowers. It is possible that D. pauciflora is a 


depicts a plant with an elongate raceme 
depauperate specimen of D. racemosa. 


16. Distictella ig a a PE dwith, Bol. 
Soc 


Venez. Ci. Nat. 25(106): 48. 1963. TYPE: 
Venezuela. Bolívar: Rio a yuni drainage, 
5 km S of El Dorado, NE of Luepa, 800- 


1200 m, 6-11 Mar. 1962, J. A. Steyermark & L. 

Aristeguieta 98 (holotype, K not seen; isotypes 

he 947/A11!, 
) 


image!, US not seen, digitized image!). 


not seen, microfic digitized 


Liana; young branchlets terete, drying dark red, 
solid, pubescent with erect to spreading trichomes; 
trichomes ferruginous, the longest 1.5-2 mm; pseu- 
e occasionally present, falcate, thick, 2.5-3 

—1.2 mm. Leaves bifoliolate, trifid tendril usually 
pubescence like that 


m, with pubescence 


ise petiole 3-10 m 
of branchlets; cri 7- 20 mm, 
like that of Ru leaflets elliptic to oblanceo- 
late, 4-12.5 X 


(sterile specimens sometimes chartaceous), 4 to 6(7) 


m, subcoriaceous to coriaceous 


pairs of lateral veins, each lateral vein initiating at a 
45° angle with the midrib and curving toward the apex 
before anastomosing with the tertiary veins and 
fading, tertiary veins conspicuous, anastomosing with 
the higher order venation to form a loosely closed 
network, all veins immersed adaxially and raised 
abaxially, both surfaces drying brownish green, 
adaxial surface bullate (not always so in sterile 


lepidote, without visible glands, margin recurved, 
base rounded, subcordate, or cuneate, apex obtuse or 
rounded, apiculate (sometimes on sterile collections 
long attenuate, or cuspidate). Inflorescence a raceme 


or racemose panicle, peduncle with rachis 4.5-12 cm 
with 10 to 20 flowers, peduncle 3—4 mm wide at base, 
peduncle, rachis, and pedicels drying reddish brown 
to dark reddish purple, pubescent with trichomes like 
branchlets, or trichomes smaller and ascending, 
bracteoles 2-5 1-1.5 mm, usually present in 
calyx 8-14 X 6-14 mm, 


pubescent with dense, red, spreading to ascending 


flower; subcoriaceous, 


or appressed trichomes to mm, glandular fields 
2.5 pairs; corolla white with yellow throat, infundib- 
ular from cylindric base, curved 1/5-1/3 distance 
from base, 4—5.3 cm, subcoriaceous, tube 3—4 cm 
with cylindric base 10-13 mm long and mouth 15- 
20 mm wide, externally and internally densely 
pubescent with minute trichomes above glabrous 
base, internally with villous clusters of trichomes at 
bases of stamens and staminode, lobes 8-18 x 12- 


om base of corolla tube, anthers 4—5 mm, 
filaments 21—24 mm, shorter MR 16-21 m 
staminode 5—7 mm; disc 1-2 X 3.5-6 mm; pistil a. 
X 2-2.5 mm, stipe 0.5-1 mm 
(rarely absent), stigma subulate or es ee 


4.5 cm, ovary 34 
obovate-oblong, not curved, , base 
obtuse 
pressed, densely Sulieseeht with ferruginous tri- 


x 
acuminate, on diste io com- 
chomes, with few scattered glands, midrib inconspic- 
uous, capsule splitting along midrib at maturity. Seeds 
ca. 20 X 40 mm, wings poorly demarcated from seed 
body, brown, membranous. 


Distribution. — Distictella porphyrotricha is known 

from eastern Bolívar, Venezuela, and Amapá, Brazil. 

It is generally reported from forests on terra firme at 
2A 


elevations between 900 and 1380 m. Figure 


Phenology. Flowering January and March; fruit- 
ing May. 

Discussion. Fruiting material was not seen in this 
study; the description here is based on Sandwith 
(1963), Gentry (1982), and notes by Gentry on the MO 
photocopy of the K photo 7. Lasser 1793 (VEN) 

e only other species of Distictella with such long 
trichomes as D. porphyrotricha (1.5-2 mm long) on 
branchlets, petioles, petiolules, peduncle, and pedi- 
cels is D. dasytricha (trichomes 1-2 mm long). The 
other species have trichomes on these parts ranging in 

ngth from —0.7 mm.  Distictella | dasytricha 
differs from D. — most markedly in its 
leaflets, which are non-bullate and have a fine, closed 
network formed by the higher order venation. 
Distictella dasytricha m has lateral veins that 
diverge from the midri a much steeper angle 
(20°), the branchlets, 


petioles, petiolules, and inflorescences, bracteoles 


golden-colored ae on 


Volume 96, Number 2 
2009 


Pool 315 
Review of Distictella (Bignoniaceae) 


that are lost prior to anthesis, and capsules that are 
smaller (6.8—7.5 cm), with an obtuse to 
rounded apex and 


curved, 
prominent midrib, and it is found 
at lower elevations (ca. 240 m), often in swamp 


forests. 


Additional specimens examined. BRAZIL. Amapá: Mun. 
Mazagáo, m o s esquerda do rio Jari, morro do Filipe 
UL M. J. P. Pires et al. 819 (MO) 819-a (MO). 
VENEZUELA. Bolívar: Alrededor del Km 123, carr. El 
Dorado—Santa Elena de Uairen, L. Marcano-Berti et al. 101- 
981 (MO); Km 122 S of El Dorado, A. H. Gentry et al. 10571 


m E of El Pauji by rd. & 6 

N of hwy., Río Las aie R. Liesner 19154 
W of Karaurin Tepui at jct. 
of Río Karaurin & Río Asadon (Río Sanpa), R. Liesner 23864 
de galleria a lo largo de las 


Teresita de Kavanayén, J. A. Steyermark et al. 115553 (MO); 
Kavanayén, T. Lasser 1793 (MO, photocopy of K's photo of 
sheet at VEN). 


17. Distictella racemosa (Bureau & K. Schum.) 


8(2) 179. 189. 

S River, s.d., R 

Pu ue 1033 (lectotype, Er ES 
BR 00 not seen, digitized image! 


Liana; young branchlets terete, the nodes swollen or 
flattened, drying dull reddish brown, black, or grayish 
ellow-green, solid a hollow), pane lepidote, 
sed trichomes or glabrescent; 

trichomes dull white or ae to 0.05 mm; 
oblong 


mm. Leaves bellas 


ye 
pubescent wit! 


to elliptic, 
trifid 


mm, with 


pseudostipules rarely present, 
thick, 3-6 X 1.54 
tendril sometimes present; petiole 7— 
pubescence like that of branchlets; petiolules 5- 
25 mm, with pubescence like that of branchlets; 
leaflets elliptic, 8-20 X 3. 
subcoriaceous, (4)5 
lateral 


3-11.5 cm, membranous to 
to 9 pairs of lateral veins, each 
20°-60° angle with the 


midrib and curving toward the apex oy anasto- 


vein initiating at a 


mosing with the tertiary veins and fading, tertiary 
veins inconspicuous, connecting midrib and lateral 
veins and weakly reticulate with higher order venation 
forming a loosely closed network, all veins immersed 
to flat adaxially, midrib raised, lateral veins raised to 
flat, and tertiary veins immersed abaxially, both 
surfaces rying brownish green or yellow-green, 
or midrib with 


minute appressed trichomes, abundantly lepidote, 


8 
adaxial surface without trichomes o 


sometimes with scattered glands, abaxial surface with 
axils of lateral veins with midrib puberulent or pilose, 
sometimes with additional trichomes on midrib, or 
without trichomes, abundantly lepidote, sometimes 


with glandular fields at apex and/or in basal lateral 
vein axils with midrib, and sometimes with scattered 
glands, margin flat, base acute, obtuse, or rounded, 


raceme, sometimes lateral, peduncle with rachis 7— 
40 cm with 3 to 30 flowers, peduncle 2—6 mm wide at 
base, peduncle, rachis, and pedicels drying grayish 
yellow-green, reddish brown, brown, or black, pubes- 
cent with trichomes like branchlets, bracteoles absent 
mm, coriaceous, 
or dull white, 
appressed trichomes and lepidote, glandular fields 2 


or not seen; calyx 6-15 X 6-12 
pubescent with minute, ferruginous 


to 5 pairs; corolla white (rarely pale lavender) with 

yellow throat, infundibular from cylindric base, 
curved 1/5-1/2 distance from base, 4.2-8 cm, 
membranous to coriaceous, tube 3-6 cm with cylin- 
dric base 7-15 mm long and es 16-25 mm wide, 
nt with 
with 
villous clusters of trichomes at bases of stamens and 
staminode, lobes 10-25 X 13-25 mm; stamens and 
staminode inserted 6—15 mm from base of corolla 


externally and internally densely pubesce 
abr 


minute trichomes above glabrous ur internally 


tube, anthers 3-5 mm, longer filaments 20-30 mm, 
shorter filaments 15-22 mm, staminode 2—9 mm; 
dise 1-2 X 3-6 mm; pistil 3.4—5 cm, ovary 3-6 X 
1-3 mm, stipe 0.5-1 mm (rarely absent), stigma ovate 


or subulate. Capsule broadly Le oblong, or 
oblanceolate, curved or not, 8-22 X 3.5—7 cm, 0.5- 
iam., base acute, obtuse, or rounded and 


truncate, apex acute, rounded, obtuse, or obtuse and 
apiculate, both valves convex to slightly compressed or 
one valve concave and the other convex, drying dull 
grayish yellow-green or reddish brown, densely 
pubescent with minute, appressed white or ferruginous 
trichomes, with or without glands, granular or not, 
midrib not evident to prominent, capsule sometimes 
splitting along midrib at maturity, wall 3-5 mm thick. 

eeds 15-28 X 24-85 mm, wings poorly to well 
demarcated from seed body, brown or gray, or reddish 
brown with tips yellowish brown, subcoriaceous to 
coriaceous or membranous, opaque or hyaline, the 
veins contrasting in color to surface or not. 


Distribution. — Distictella racemosa is known from 
northern South America to southern Peru, northern 
Bolivia, and Amazonian Brazil. It is usually found in 


forests between sea level and 850 m. Figure 2B. 


Phenology. Flowering and fruiting throughout the 
year. 
Discussion. Bureau and Schumann (1896) cited 
two kb quede H. R. Wullschlaegel 1033 and C. F. 
s.n. as Distictis racemosa. Images of two 


specimens of H. R. Wullschlaegel 1033 at BR were 


Annals of the 
Missouri Botanical Garden 


and the sheet barcoded BR 880400 was 


annotated by Schumann and is therefore selected 


seen, 


here as the lectotype. The duplicate barcode 
880502 was apparently not annotated by Bureau or 
Schumann and is better considered an isosyntype. The 
Martius collection ers M-88916) also does not 
appear to be annotated by Bureau or Schumann and 
may, at best, be considered an isosyntype. 

Gentry (1973, 1977, 1982, 1997) and Burger and 
Gentry (2000) treated Distictella racemosa as a 
synonym o nolüfolia, a syno as 
suggested by pede. (1954). Ea aa niin 
(1953) had recommended that the two be kept as 
separate species, emphasizing the elevated tertiary 
leaf venation and the large distinctive glandular fields 
at the apex and base of the leaflets of D. magnolii- 
folia. Gentry (1973) found both these characters to be 
variable and inadequate for specific separation. In the 
current study, the venation character was found to be 
consistent, but the glandular field character was not 
reliable. The leaflets of D. magnoliifolia are quite 
distinet from those of D. racemosa. They are heavily 
coriaceous with only two to four pairs of lateral veins 
that curve up toward the apex of the leaflet but follow 
ie for a considerable distance before fading 

capsules of D 
UM from those of D. racemosa: the 
oblong (8-13.5 X 2-2.5 em), with a js attenuate 
apex, with few scattered "e minute trichomes, dry 
black, and have a thin wall (1-2 mm). In addition, the 
habitat is different. Distictella dep de is fou 
in scrubby or low forests on white sand. Distictella 
magnoliifolia is further separated from D. racemosa 
var. racemosa by its curved capsules and membranous 
and hyaline seed wings 

Two varieties of Distictella racemosa are recognized 
in this study: D. racemosa var. racemosa, which is 
found in seasonally inundated forests below 200 m 
elevation and has seeds with the wings opaque and 
subcoriaceous to coriaceous and fruits not curved, 
with both valves convex to slightly compressed; and D. 

osa var. translucida, which is found in non- 
inundated forests hemeen 100 and 950 m elevation 
and has seeds with the wings membranous and hyaline 
and fruits beers curved with one valve convex and 
ne c The two varieties are very difficult to 
Distictella 
racemosa var. racemosa is always totally without 
trichomes on the abaxial surface of the leaflets and 
nearly always has glands at the leaflet apex and in the 


o 
separate E fruits are resent. 


axils of the basal lateral veins. Distictella racemosa 
var. translucida often has trichomes in the axils of the 
lateral veins with the midrib and rarely has glandular 
fields in both the basal lateral vein axils and at the 

apex. In addition, the lateral veins of D. racemosa var. 


translucida are often at a steeper angle (207—40^) than 


= 
2 


cemosa var. racemosa (45°-60°) 
Unfortunately, as the types of the pertinent names 
are flowering collections and the habitats are not 
indicated, the names are applied with some uncer- 
tainty. 


17a. Distictella racemosa (Bureau & K. Schum.) 
Urb. var. racemosa. Figure 3D, E. 


Bignonia rd Britton ex Rusby, Bull. Torrey Bot. Club 27: 
zl. E: Bolivia. Jet. of rivers Beni & 
i de Dios, Aug. 1886, H. H. Rusby 1140 ne 
NY not seen, digitized image!; isotypes, MICH 
een, MON. 

Distictis angustifolia K. hum ague rh. 
Vereins Prov. Brandenburg 1908: 120. bo en 


mellos, Rio 
Madeira, Mar. 1902, E. Ule 6111 (holotype, enr 
at B not seen, F neg. 18440; isotypes, HB not seen, 
not seen, photo neg. 4218 not seen, photocopy!, L not 

G not seen). 
Disicel Ls C. V. Freire & A. Samp. in A. Samp & C 
e, Ann. "ed pu Sci. 8(1): 31. 

nov. "TYPE: Brazil. Amazonas: 
Negro), 8 Nov. 1932, A. Ducke 8022-a (holotype, R 
28725 [as “R 28726"] not seen; isotype, RB 24885). 

Distictella negrensis C. V. Freire & A. Samp. in A. Sam 
C. 


gro), m do rio, 8 N 
oloye T 28726 EM “R 287257] not seen; isotype, 
80221). 


Liana; young branchlets terete, the nodes usually 
swollen (rarely flattened). Leaflets with lateral veins 
initiating at a 45°—60° angle with the midrib, abaxial 
surface without trichomes, 


usually with glandular 


fields at apex and in axils of basal lateral veins with 
midrib. Calyx glandular fields 2 to 5 pairs. Capsule 
br elliptic, oblong, or oblanceolate, not curved, 
16 m, 0.5-4.5 cm diam., base acute, 
ee or rounded and truncate, apex acute, rounded, 
obtuse, or obtuse and apiculate, both valves convex to 
d compressed, drying dull a yellow-green 
reddish brown, densely pul nt with minute, 
2 white or mue. ae usually 
with abundant glands, often granular, midrib not 
evident to evident, capsule often oe along 
idrib at maturity. Seeds 15-28 X 24. 
poorly demarcated from seed body, ais or gray, 
the veins not 
surface (Figure: Gentry, 1973: 
fig. 12, as ee  magnolifolia). 


subcoriaceous to e opaque, 
contrasting in color 


Distribution. | Distictella racemosa var. racemosa is 


ound primarily along the Amazon and its tributaries 


in Brazil (Pará, Amazonas, Horaima, and Mato 


Volume 96, Number 2 
2009 


Pool 317 
Review of Distictella (Bignoniaceae) 


Grosso), Peru (Loreto), Colombia (Caquetá, Vaupés, 
and Amazonas), and Bolivia us [type of Bignonia 
rusbyi] and La Paz) It is also found in southern 
Venezuela (Amazonas) and e Guyana, Suri- 
name, and French Guiana. It is found in seasonally 
inundated forests (igapó or várzea), generally associ- 
ated wit ack water, usually at elevations below 
200 m elevation (one collection from Guyana was 


collected at 500 m). Figure 2B. 


Phenology. Flowering and fruiting throughout the 
year. 


iscussion. Bignonia rusbyi is placed in synony- 
my of Distictella racemosa var. racemosa based o 
examination of type material. Because the type of B. 
rusbyi lacks fruits, the varietal placement is not 
absolute, but is supported by 
trichomes and with the presence of glands at 
apex and in the basal lateral vein axils and the wide 
angle of the lateral veins with the midrib (ca. 45°). 

Distictis angustifolia is placed in synonymy with 
some uncertainty. The leaflets of the type collection 
are narrowly lanceolate and only about one third as 
wide as long, conditions rarely seen in Distictella 
racemosa var. rac 

The captions of lle TE of res lutescens 
and D. negrensis were swapped i 
publication (Sampaio & Freire, m Distictella 
lutescens can be confidently placed in synonymy o 
. racemosa var. racemosa based on examination of 
the flowering isotype (RB) and illustration of the 
holotype in fruit. Examination of the isotype (RB) and 
illustration of the holotype of D. negrensis supports its 
placement in synonymy of D. racemosa, but the 
absence of fruit makes the varietal placement less 
certain. Its placement in purus of D. racemosa var. 
acemosa is supporte secondary varietal 
characteristies (leaflets MONA trichomes and with 
glands at both the apex and in the basal heal vein 
a 45^ angle with 
midrib) and in the collection of the type along the 
margins of the Rio Negro with the type 


axils and lateral veins forming 


of D. lutescens. 

Specimens from Venezuela and Guyana have 
capsules that are thicker and relatively broader than 
other collections of Distictella racemosa var. racemosa 
and approach D. cremersii. However, the seeds in 
these collections have much larger wings than those of 
D. cremersii, which has the wings nearly completely 
reduced, and the | 
on the abaxial surface 


eaflets are totally without trichomes 
(vs. pilose in the lateral vein 
axils in D. cremersii). 

Individuals of Distictella racemosa var. racemosa 
eaflets are e confused with D. 
4.78 X 


specimens of D. cuneifolia ed can be recognized 


with smaller 


cuneifolia (leaflets .3 cm). However, 


by the leaflets, which are often spatulate and lack 
apical P and the fruit, which are spatulate and 

6.7 X 2.5-2.7 cm) than t 
racemosa var. racemosa and have smaller seeds 
(9-12 X 26-28 mm) with membranous and hyaline 


wings. Distictella cuneifolia is also found in savan- 
s, while D. ra 


smaller ose o 


cemosa var. racemosa is found in 


forests. 
Selecied specimens examined. BOLIVIA. La Paz: Prov. 

Iturralde, Luisita, W del Rí 

Muqui, S. e 


Amazonas 


m 1475 


= 
E 
B 
a 
o 
e 
= 
A 
4] 
B 
a 
EN 
d. 
o 
"- 
a 
e 
8 
5 
O 
ES 
& 
E 
C» 
UR 
E 


20442 (MO). CO BIA. [ua zonas: 


2109 (MO). FRENCH GUYANA. Piste de Risque Tout, Km 
10 (environ) à partir de Montsinéry, F. Billiet & B. Jadi 
2037 (MO). GUYANA. Potaro-Siparuni Region, Kaieteur 
Falls Natl. Park, W. e et al. 4761 (MO). PERU. Loreto: 
Banks of Río N F. 
Prov., Río Tone. as E 
h.p. moi above Requena, om. et al. 21293 (MO). 
SURINAME. High creek m along bs Creek at crossing 
d. 


arriba dede la confluencia con el río Cunucunuma, J. A 


Steyermark ei al. 126235 (MO). 


17b. Distictella racemosa (Bureau & K. Schum.) 
Urb. var. translucida A. Pool, var. nov. TYPE 
Colombia. Antioquia: autopista Medellín-Bo- 
gotá, vereda La Josefina, camino hacia El Pitál, 
Mpio. de San Luis, 800 m, 29 Nov. 1983 (fr.), S. 
z 535 (holotype, MO!; 


Hoyos & J. Hernande 
isotype, JAUM not seen). Figure 


Distictella broadwayana Urb. E E 2l inco Md 
14: 310 


Tobag r Men 
EB siue 4753 ed BM n not 
seen, diia iù image!). 


Haec varietas a varietate typica seminibus alis membra- 
naceis ala praeditis et capsula plerumque curva 
distinguitur. 

Liana; young branchlets terete, the nodes usually 
flattened. Leaflets with lateral veins initiating at a 

20°-45° angle with the midrib, abaxial surface with 
axils of lateral veins with midrib puberulent c or 
rarely without trichomes, and sometimes 
tional trichomes along midrib (rarely also Pm ee 
veins), sometimes with glandular fields at apex, 
occasionally in basal lateral vein axils, s in both 
locations. Calyx glandular fields .5 pairs. 
Capsule oblong, curved, 11-22 x a cm, 0.5- 


Annals of the 
Missouri Botanical Garden 


2.5 cm diam., base rounded and truncate, apex acute, 
or acute or obtuse and rounded, one valve concave 
and the other convex (rarely not curved, and both 
valves somewhat flattened), drying reddish brown, 
densely pubescent with minute, appressed, ferrugi- 
nous trichomes, with or without glands, not granular, 
15-25 X 55-85 
poorly to well demarcated from seed body, reddish 


midrib prominent. Seeds mm, wings 


brown with tips yellowish brown, membranous, 
hyaline, especially at tips, the veins contrasting in 
color to surface 


Distribution. | Distictella racemosa var. translucida 
is known from Tobago (type of D. broadwayana), 
antander, 


western Colombia (Antioquia, Chocó, 


uetá, Amazonas), Ecuador (Napo), central Peru 
(Amazonas, Loreto, San Martín, Junín, Madre de Dios, 
Puno), southern and eastern Venezuela (Amazonas, 
Bolívar, Delta Amacuro, Sucre), northwestern Brazil 
mazonas, Acre, Pará, Rondónia, Roraima), 
Guianas. It is usually found in forest in mon nonien 


areas between 100 and 950 m. Figure 2B. 


Phenology. Flowering August to May; fruiting 
April, May, July, September to December. 


Red List category. The number of collec- 


tions of Distictella racemosa var. translucida studied 
(88) and the wide range of its distribution suggest that 
this taxon is of Least Concern (LC) status over its 
ntire range, according to ist criteria 


e 
(IUCN, 2001). Thre 


cannot be determined in the present study. 


ats to particular populations 


on. Examination of type material of Di- 
a uen ayana places it in synonymy of D 
racemosa at the species level. This treatment follows 
that of Sandwith (1938a, 1954, 1965). Because the type 
of D. broadwayana lacks fruits, the varietal placement, 
nslucida, is highly 


probable but not absolute. The name D. broadwayana 


as a synonym of D. racemosa var. tra 
has been little employed; the only specimen observed 
annotated with this name was the type, and the only 


of Urban 
(1916a, b). It was this lack of name usage in addition to 


publications accepting this name were thos 


its ambiguity, due to absence of capsule and seed, that 

o the decision to publish D. racemosa va 
translucida as ew variety as opposed to a 
changing the me of D. broadwayana. 

Vegetatively, Distictella racemosa var. translucida 
is very similar to D. cremersii, which differs in having 
fruits that are relatively broad and thick (7-11 X 4- 
5.5 em, 3-4 em diam.) and having seeds with the 
wings almost entirely reduced. Flowering material has 
not been associated with D. cremersii; it is possible 
that some of the flowering material cited here for D. 
racemosa var. translucida is actually D. cremersii. 


Distictella racemosa var. translucida can also be 
i. The latter has the 
undersurface of the leaflets densely tomentose a 
] homes. The midrib of the 


parkeri is not, or only slightly, prominent. In 2 


compared with 


small tric capsule 

the leaflet base of D. parkeri is often cordate or 
subcordate and never acute or obtuse. Both species 
can have rounded leaflet bases. 


Selected DT examined. AZIL. Aere: Vic. of 
Serra da Moa, G. T. Prance et al. 12191 (MO). Amazonas: 
nean near Livramento, on Rio Livramento, B. A. 

Krukoff 6915 (MO) Para: 
K 


Serrinha, N. 
Uraricoeira, G. 5 Frame et al. 1090. 
Amazo: 


Samaná—Río Claro, 
hacia la vereda Tulipán, Mpio. de San Luis, A. Cogollo & 
C. Estrada 200 ee Caqueta: 28 km E of Morelia toward 
Río Pescado, A. H. Gentry et al. 9076 (MO). Chocó: New rd. 
1 oro, A. H. Gentry & M. Fallen 
Las Colonias (Carare Opón 
Santander, E. Renteria et al. f 545 MON pag d 
Napo: Rd. Coca, Puerto F: Q Curaray, 20— 
& L. Andersson 11947 (MO). 
FRENCH GUIANA. Parcele Arbocel, piste de St. Elie, Km 
14, recur après coupe, M. F. Prévost 590 (MO). GUYANA. 
ot dE region, Chenapou, 
a, 50 km u dn. prm 


Amerindian village, 
L. 


Falls, £. P. Kvist 


Maynas, Almendras (Centro de In 
estan Forestal d Universidad Nacional de la 
ae Y a [UNAP)), R. Vásquez & N. Jaramillo 6517 
i, R. J. Seibert 2233 (MO). Madre de Dios: 

m upriver from Explorer's Inn, near 


confluence of Río Tambopata and Río La Torre, 39 km SW of 


et 9 "76030 (MO). San Martín: Mariscal Caceres, drach 
Nuevo, trail up Río Huallaga Valley toward Limón, A. H. 
Gentry et al. 25571 (MO). SURINAME. Tapanahoni R., 

Rapids, H. E. Rombouis 655 (MO). VENEZUELA. 


E Alrededores de San Simón de Cocuy, ca. 2- 


Blanes Sai Martí tín de 


González 16294 (MO). Suere: Peninsula de Paria, above Las 
Melenas, N of Río Grande Arriba, SE of Cerro de Humo, J. 
Steyermark & R. Liesner 120938 (MO). 


Volume 96, Number 2 
2009 


Pool 319 
Review of Distictella (Bignoniaceae) 


18. Distictella reticulata A. H. Gentry, Ann 
Missouri Bot. Gard. 65: 728. 1978 [1979]. TYPE: 
Brazil. Amazonas: Manaus, Igarapé de Cachoeira 
Alta do Tarumã, 28 Aug. 1962, W. Rodrigues de 
J. Chagas 4610 (holotype, INPA not seen; 
isotype, MO!). 

Liana; young branchlets terete, drying dull brown, 
hollow or solid, pubescent with trichomes weakly 
0.1-0.4 mm; 
pseudostipules caducous, spatulate, 3.5-5 X 1- 


ascending; trichomes tan or white, 


2 mm. Leaves bifoliolate or trifoliolate, trifid tendril 

m, with pubes- 
STEM petiolules 5-15 mm, 
leaflets 
elliptic to ovate, 5-13.5 X 3-8 cm, coriaceous, 4 or 


RDUM nis petio 
cence like of bra 
with ee like that of branchlets; 


5 pairs of lateral veins, each lateral vein initiating at 
20°-45° 


toward apex 


angle midrib and extending 
dus. ium with the tertiary 
veins, tertiary veins conspicuous but not distin- 
guishable from the higher order venation, together 
uel FREUE and forming a fine, closed 
ns immersed adaxially and raised 
a wally, s th surfaces dryin 


adaxial surface nearly without trichomes, with a 


brownish green 


few trichomes at the base along the midrib, scattered 
lepidote, without glands, abaxial surface wit 
ense covering of trichomes like those of branchlet 
or slightly smaller, mainly on venation reticulation 
and pointing in toward the areoles, abundantly 
lepidote, without visible glands, margin recurved, 
base rounded, apex acute to rounded. Inflorescence 
a raceme or racemose panicle, 
5-29 cm with 5 to 1 
wide at base, peduncle, rachis, and pedicels drying 


reddish like 


branchlets, bracteoles absent or not seen; calyx 


peduncle with rachis 
6 flowers, peduncle 2.5 mm 
pubescent with trichomes 
-14 X —]2 mm, coriaceous, pubescent with 
dense, tan, strongly appressed trichomes to 0.3 mm, 
glandular fields 2 pairs; corolla white, infundibular 
from cylindric base, curved 1/4-1/3 distance from 
base, 4.5-6.5 cm, tube 3—4.5 cm with cylindric base 
9-12 mm long and mouth 15-25 mm wide, exter- 
nally and internally pubescent with minute tri- 
chomes above glabrous base, internally slightly 
glandular-villous at stamen insertion, lobes 15-20 
X 15-20 mm; stamens and staminode inserted 12— 
anthers 3.5— 

orter fila- 


13 mm from base of corolla tube, 
4.2 mm, longer filaments ca. 22 mm, 
ments ca. 18 mm, staminode ca. 4 mm; dis 2-2.5 X 


2 mm, 
stipe absent, stigma d oblong. Capsule elliptic, 
2.5 


4—4.5 mm; pistil ca. 3.5 em, ovary ca. 4 X 
not curved, ca. 2.8 cm, ca. 

base acute and truncate, apex acute and apiculate, 
both valves slightly compressed or sometimes one 


very slightly convex and the other slightly concave, 
drying reddish brown, densely pubescent with 
appressed, ferruginous trichomes, E 


thick. Seeds ca. 


35-50 mm, wings poorly d from pus 


minute, 


not evident, wa 


body, reddish brown with tips tan, membranous, 
hyaline, especially at tips, the veins not contrasting 
in color to surface 


Distribution. Distictella reticulata is known only 

from Amazonas, Brazil, in the vicinity of Manaus. The 

elevation was not indicated on the labels; most 

collections were made along the margins of narrow 
1A 


rivers. Figure 


Phenology. Flowering May to July; fruiting Au- 
gust. 


Discussion. The tertiary veins of the leaflets of 
Distictella reticulata cannot be distinguished from 
the higher order venation and form a strongly raised, 
very fine reticulation on the abaxial surfaces of the 
leaflets, similar to that found on the simple leaves of 
D. monophylla and the leaflets of D. dasytricha. 
Distictella elongata and D. obovata also have a very 
fine and strongly raised network of veins on the 
abaxial leaflet surfaces, but usually with the tertiary 
veins distinct from the higher orders of venation. 
Distictella reticulata is easily distinguished from D. 
monophylla, a shrub with simple leaves, and D. 
dasytricha, which has (1-2 mm) 
trichomes. Distictella elongata differs from D. 


much longer 


reticulata in its leaflets, which have cuneate to 
attenuate bases, fruits with an acuminate apex and 


x 25- 


28 mm). Distictella obovata can be separated from D. 


prominent midrib, and small seeds (8-10 
reticulata by its leaflets, which are obovate or 
oblanceolate with a cuneate base and a 
adaxial surface, and its fruits, which dry dull "black 
and are quite warty. 


Selected specimens | exognmed, BRAZIL. Amazonas 
anaus, margem é da cachoeira Alta, Estrada 
Forquilha, J. Chagas 1 703 o, 1086 (K); Manaus, mar- 


argem do da 
Taruma, J. d 1330 a on Francisco & Dionisio 
3940 (K). 


Literature Cited 


Baillon, H. E. 1891 [1888]. Histoire des Plantes. Bignonia- 
cées, . 10: chette et ME o 

026-1053 in G. 

Hooker o p Plantarum, 

Vol. ae Lovell Reeve & Co., ams & Norgat 

London 


Annals of the 
Missouri Botanical Garden 


Brawer, M. 1991. Atlas of South America. The MacMillan 
Press Ltd., London and Basingstoke 


Bureau, E. 1864. Monographie des Bignoniacées. J.-B. 
Bailliére et Fils, Paris. 
K. Schumann. 1896. Bignoniaceae. i. Pp. 1-230 in 


C. F. P. von Martius (editor), Flora eens Vol. 8(2, 

e x2 Lipsiae apud Fried. Fleischer in Comm., 

We ve A. H. Gentry. 2000. Bignoniaceae. Pp. 77-162 
in W. Burger (editor), "ors Costaricensis. ae Bot., 
n.$., 

Eu A. P. de. 1845. Bignoniaceae. Pp. 142-248 in A. 

e Candolle (editor), Prodromus, Vol. 9. Sumptius 


Vi cds Paris. 

Funk, . & S. A. Mori. 1989. A e a 
ctr in Bolivia. Smithsonian Contr. —iil, 
12 


e A. H. 1973 [1974]. ju as Family 172. Bignonia 
ceae. In R. E. Woodson Jr. & R. W. Schery (editors), oe 
of Panama. Ann. Missouri Gard. 60: 781-977. 

1974a. e in Bignoniaceae 
noteworthy species of South 
ion uri Bu. Gard. 61: 872 
74b. Coevolutionary patterns in t Central American 
Bionic. Ann. Missouri Bot. Gard. 61: 728—759. 
— ———. 1976. Studies in Bignoniaceae 19: 
and new species of South 
ieee is Gard. 63: 4 
. 1977. Family 178. Bignoniaceae. In G. Harling € E 
DUM du. Flora of Ecuador. Opera Bot. 7: l- 
173. 


American Bignoniaceae. 


[d mergers 
American Bignoniaceae. Ann 
6-80. 


———. 1978a. Bignoniaceae. In B. Maguire ri ed 
pem of the ium Highlands—Part X. 
York Bot. Gard. 29: 245-283. 
1978b [1979]. Studies in See 31: New 
and combinations from Amazonian Peru and 
Oa. Studies in Bignoniaceae 37: New species of 
Bignoniaceae from eastern South America. Phytologia 46: 
201-215. 
980b. New species of Apocynaceae, Bignoniaceae, 
Massiilosieeae. and Piperaceae from coastal Colombia and 
Ecuador. Phytologia 47: 97-107. 
1980c. Bignoniaceae—Part I. Fl. Neotrop. Monogr. 
25(1): e 2s 


982. Bignoniaceae. Flora de Venezuela 8(4). 
9 de Educación Ambiental, Caracas. 
0. ea patterns in a Bigno- 
niaceae. Mem. New Bot. Gard. 55: —129. 
. 19 ix new species of un Med from upper 
acens. Novon 2: 159-166. 


——. 1997. dod or Pp. n P. E. Berry 
K. Holst & K. Yatskievych “sion Flora of. d 
Venezuelan Guayana, Vol. e to Cactaceae. 
Missouri Botanical Garden Press, St. Louis. 
& A. S. Tomb. 1979 [1980]. seat oe 
of Bignoniaceae palynology. Ann. Missouri Bot. Gar 
56-777. 


Gomes, J. C. 1955. Contribuição à sistemática das 
Bignoniaceae Bids Arq. Serv. Florest. 9: 261-296, 
L 14 


pl. : 
Goodman, E. J. 1972. The Explorers of South America. The 
Macmillan Company, New York, and Collier-Macmillan 
n 


1988. A daar characteristics 
of the Eus fiesta. Taxon 37: 578-594. 


IUCN. 2001. IUCN Red List Categories and Criteria, Version 
3.1. Prepared by the IUCN Species Survival Commission. 
IUCN, and, Switzerland, and Cambridge, United 
Kingdom. 

Janzen, D. H. 1971. Euglossine bees as long-distance 
pollinators of tropical e Science 171: 203-205 

Lohmann, L. G. 20 Untangling the phylogeny of 
Neotropical lianas (Bignonieae, Bignoniaceae). Amer. J. 
Bot. 93: 304—318. 


. R. Pirani. 1998. Flora da Serra do Cipó, Minas 
Gerais: Bignoniaceae. Bol. Bot. Univ. Sáo Paulo 17: 
127-153. 

——— . J. G. Hopkins. 1999. Bignoniaceae. Pp. 608— 
623 in J. E. L. da S. Ribeiro et al. Flora da Reserva Ducke: 
Guia de Identificacao das Plantas Vasculares de Uma 
Floresta de Terra-Firme na Amazonia Central. INPA- 
DFID, Man 

Meisner, C. 134 O. Plantarum Vascularium Genera. 9(1- 
Tab. De) 285-312, a Commentarius): 193—224. 
Libraria Weidmannia, Lei 

927. Der nidis Formenkreis der Pithe- 

Bignoniaceae. 
2. 


Orbigny, A. D. g. s Pisos ue Méridio- 
nale D Vol. 3, Part 2 (Geom io P. Bertand, 


Pool, d 2007. A review of the genus Distictis (Bignoniaceae). 
Ann. Missouri Bot. Gard. 94: 791-820. 

Post, T. € 0. Kuntze. 1904 [1903]. Lexicon 
Generum Phanerogamarum. Deutsche Verlags-Anstalt, 
Stuttgart 

= A. J. de. 1935. m especies de Bignoniaceas. 

cad. Brasil. Sa. 7: —127. 

E C 


Freire. 


s novas especies 
Amazonicas de “Distictella Ciena) Ann. A 
-32 


cal American 


Née. 34: 205- 


Sandwith, N. Y. 1937. Notes on trop 
Bignoniaceae. Hecueil Trav. Bot. 
232. 


a. Bignoniaceae. Pp. 1-86 in A. Pulle E 
Flora of Suriname, Vol. 4(2). J. H. de Bussy 
ap. am. 
938b. Notes on unidentified tropical South 
mde Bi Humboldt and Bonpland. Lilloa 
3: 457—465. 
— ——. 1938c. Three new South American plants. Brittonia 
3(1): 91-94. 
. 1953 [1954]. Contributions to the flora of tropical 
América: LVI. iad ther studies in Bignoniaceae. Kew Bull. 


1953: 4514 


354 in Flora 


noniaceae. Pp. 316— o 
ee of Agriculture, 


Trinidad and 4 Vol. 2. 
Government Printer, Port-of-Spa 

1957. Bignoniaceae. /n B. cee & J. J. Wurdack 
e The bd of e Guayana P RA II. 

k Bot. Gard. 9(3): 359-366 

D. to the flora of topical America: 
ud pon on Bignoniaceae XXVI. Kew Bull. 15: 
459-466. 


————. 1963. Bi gnoniaceae. In J. A. Steyermark & S. 
Nilsson eio e. tanical Novelties in the Region of 
Sierra de Lem sado Bolivar -2. Bol. Soc. Venez. 
Nat. 25(106): 48— 48 

65. oum to the flora " M me America: 

LXXII. Notes on Bignoniaceae: XXVIII. The oo of 

Anemopaegma nigrescens. Kew Bull. s 409— 414 


Volume 96, Number 2 
2009 


Pool 321 
Review of Distictella (Bignoniaceae) 


— — —. 1968. Contributions to the flora of tropical America: 
m Notes on Bignoniaceae: XXIX. Arrabidaea in 
Martius's ‘Flora Brasiliensis’ and subsequently. Kew Bull. 
22: 403—420. 

Santos, G. dos. 1995. Wood Anatomy, Chloroplast DNA, and 
Flavonoids of the Tribe Bignonieae (Bignoniaceae). Ph.D. 
e The University of Reading, Reading, United 


ngdom 
st R. E. 1970. De plantis toxicariis e Mundo Nov 
cale commentationes VII. Several pears ie 
s from the Colombian Amazon. Bot. Mus. Leafl. 22 


345-352. 
dre K. 1894 [1894—1895]. Bignoniaceae. Pp. 189— 
252 in A. Engler (editor), Die Naturli a Pu WE 
lien, Vol 4(3b). Wilhelm Engelmann, Lei 
RO ae T. A. 1911. Arrabidaea crassa. Hooker’ Icon. Pl. 
Bl 


N. Y. Sandwith. 1932. Contributions to the flora of 

tro 21 America: X. New 

from British Guiana, mainly collec 

University Expedition, 1929. Bull. Misc. Inform. Kew 
1-93. 


Urban, I. . Bignoniaceae trinitensis, nonullis aliis 
antillanis novis adjectis. Repert. Spec. Nov. Regni Veg 
14: 300-314. 


916b. Über Ranken = Pollen der A 
r. Deutsch. Bot. 158, ta 
Verlot B. 1868. Teenie ert ene la Rev. 
Hort. 40: 152-154. 


APPENDIX 1. Index to Numbered Exsiccatae. 


Collections are listed aeons by collector and 
collection number with each collec mber followed by a 
The index includes all 

Tropicos that were identified to 
es by A. Pool and are without qualifiers (aff., cf., p 
excluding those records lacking collector name. Boldfac 
indicates type collections. 


List oF SPECIES 


Distictella arenaria A. H. Gentry 

Distictella campinae A. Samp. 

Distictella chocoensis A. H. Gentry 

Distictella cremersii A. H. Gentry 

Distictella cuneifolia (DC. E MR 

Distictella dasytricha San 

Distictella elongata. xr 

Distictella laevis a hi HL Gentry 

. Distictella lohmanniae A. Pool 

"mon a uM Sandwith 
tella mansoana (DC.) U 


TORNADO No 


eee eee 
DUE DIN 
S 
qos 
ER 
ivi 
& 
a 
E 
8 
> 
8 
S m 
a 
a. 
pis 
yg 
e 
e 
Uu 
3 
- 
I 
ge 
E 
o 
Re 
wm 
A 
il 
a 
É. 
= 
= 


E 
ea] 
e 


. Disticiella racemosa (Bureau & K Schum.) Urb. var. 
racemosa 
17b. Distictella racemosa var. doe A. Pool 
18. Distictella reticulata A. H. Gen 
2 P. H. 3210 no pud D. & E. Curado Lopes 
39 (11). Alvarenga, D. & F. M. Paixão 651 (11). Alvarenga, 
Fi s.n. (INPA 120328) al ae, 1 L. et al. 545 (17a), 932 


(8), 955 (8). Amaral, M. C. et al. 7151 e Anderson, W. R. 
10539 O, AA (8), 11965 (11). Anderson, W. R. et al. 6555 
11), 8798 (11), 10216 "TN 10945 (11), 10999 
D. Alvarenga 933 (11 
, . C. S Ferreira 3111 (11 
Aquilante, D. 6 (11). Arbo, M. M. et al. 4878 (11), 4999 (11). 
Aronson, J. & F. Rodriges V. 861 (17b). Ayala, F. 257 (17a). 
Ayala, F. et al. 3366 (17a). Aymard, G. 7993 (17a). Aymard, 
G. & L uo p O mene G. et al. 4057 (17b). 
Azevedo, M. 

Barreto es »- el va E T 2354 cn aoe t i 
S. G. 4941 (11), 20589 (5). Beck, S. G. ase 1000 
(11), 10152 (17a). d E. 1382 00). 1s (10 T a 
(10), 1496 (10). Berry, P. E. & E. Melgueiro 5315 ITY, 
P. E. et al. 6026 (12), 6173 m 6250 (12), 6542 do p 
a Besse, L. etal. 1794 (11 

F. & B. Jadin 1604 (17a), 2037 (17a 
Black, G. ei iul 57-19576 (1). Boapland, A. 973 
(10). Broadway, W. E. 4753 (17b). 

Calderon, C. E. ^ al. 2703 (8). gcc D. et al. 2940 
(17b). Cardiel, J. M. et al. CHIN-292 (17b). ns A. 
4292 (lTa), 4995 (17a), 5308 (17a), p^ (17a). Castillo, A. 
et al. 5502 (1). Chagas, J. 1086 (18), 1330 (18), 1705 18), 


= 


(10), s.n. 


al. 497 2), 820 (17a), 1264 (1), 1389 (17a), 2210 (128), 
Fu ir a n wi dd 2 n Bx A. & 
Ü o m. 0 (79). 
Cora a ps a o. pM L sn. 
(cress n ee de ol J. 13 (11). Costa, F. N. e 
9(4 2 5281 (17b), 8377 (14. 
9528 o Cremers, G. & Ex de Granville 14427 (1). 
Cremers, G. & M. Hoff 11311 (17b), 11344 (17b). Cremers, 
G. et al. 12507 (14). Croat, T. B. 19905 (17a), 20049 (17b). 
us a S. de la 2734 (14). 

y, D. C. et E ie (2). Damião, C. 3001 (17a), 3089 
A. C. González 16294 (Y 7h). G. Davidse 
68 (2). ds C. & 


C. et ui 256 (17a), 605 (Ta), 1029 (17b). Dodson, C. H. & 
Torres 2940 (17h). Ducke, A. 8022 (17a), 8022-a (L7a), 
22688) (2). Dulmen, A. van € N. Matapi 42 


Eten, G. et al 5797 (11), 5813 (11), 6013 (11). 
Encarnacion, F. 80 la Evans, R. & G. Lewis i (4). 
. Feuillet, C. 1588 (7). Fosberg, F. R. 
Ps 1). Pad nus 


9). 
Gentry A H. Vin as "um E 12976 (9), 13062 
(17a). Geniry, A. H. & P. Berry 14615 (1), 14618 ap 
" Gentry, A. H. & D. Daly 18322 (17a). Gentry, 
H. & C. Díaz S. 28199 (17b), 28225 (6), 28254 (6). ci. 
. H. & M. Fallen 17815 (11b). Gentry, A. H. & R. Foster 
70871 QD. Gentry, A. H. & C. Feuillet 63205 (7). Gentry 
N. Jaramillo 57925 (17b). bu ug H. & P. Núñez 
s (11), 69789 iv Gentry, & R. Ortiz E 
A. Dodd do (2). Gentry, A. H. 


5 Stein 46381 (10), 46685 Qu) 47159 (17a). Gentry, A 
H. & S. Tillett 10874 (1), 10900 (10). e & K. 
ung 32010 (17b). Geniry, A. H. & Elsa 

(14), 50342 (7). Geniry, A. H et al. 9076 (us. "10484 (13), 

10565 (16), 10566 (13), 10571 (16), 10660 (14), 15637 

(17b), 15844 (17a), 18557 (17a), 20335 (17a), 21293 (17a), 

25571 (17b), 42096 (17b), 44250 (11), 44335 (11), 45995 


Annals of the 
Missouri Botanical Garden 


(17b), 47844 (3), 47957 (3), 49589 (11), 51285 (17b), 
53286 (3), 59175 (11), 63329 a 74053 (11), 76651 
(17b), 76930 (17b). Gillespie, L. J. 9 otisberger, I. 
S. 220 (11), 867 (11). Gounelle, E. s.n. E ‘Granville, J.-J. 
de 2080 (4). Grifo, F. & J. Solomon 773 (11). Grotta, A. S. 
s.n. (SPF15220) (i) s.n. (SPF37459) (11). Guanchez, F. 
1119 (12). Guillén, R. & R. Choré 5441 (5). Guillén, R. & S. 
Coria 15224 (5), 2111 (5). Guillón, R. & V. Roca 3092 (5), 
3212 (5), 3501 (5). Guillén, e E al. 3897 (11). Gutiérrez, E. 
et al. 558 (11), 767 (11), 1 

ahn, W. & S. Tiwari 2n (14). Hahn, W. et al. 4761 
(17a). Harling, G. € L. Andersson 11947 (17b). Hatschbach, 
G. 23542 (11), 28074 (11), 35915 a 39444 (11), 40808 
(11), 45956 (11). Hatschbach, G. P. Ramamoorthy 
38187 (11). eae G. et al. pes (11), 60189 a 


Heringer, E. P. 52 (11), 15412 i D " P. e 
3095 (11), dios rts 3522 (11). He 91 (1), la 
(17a). Hil, S. y 13199 (17a). Hoffman, r a M. van 


3 
e 
: 
= 
2 
E 
i] 
3 
E 
ea 
ea 

2 


14). Hoffman, B. et an 542 (14), 5304 (14). 
onda, M. & F. Mello 35996 (17a). He ernandez 
535 (17b). Huber, O. 2514 (12), 3173 13), 3953 (12), 4061 
ree 6007 (12). Huber, O. & E. Medina 5892 (12). Huber, O. 
. S. Tillett 2920 (12), 2938 (12), 5282 (12), 5355 (12), 
ape (12), 5587 (12). Huber O. ei al. 5687 (10). Humbert, H. 


27479 (10). 
INPA 3994 did pur J. C. 41 (17a). Irwin, H. S. 235 
(11). Irwin, H S. e 6291 (11), 11177 (11), 11445 (11 
a 


12021 (11), 13415 cm 15389 (11), 15653 (11), 16821 
25428 (11), 25684 (11), 27429 (11). 
a J. 9689 px 


1) 


65 10). Killeen, T. et al. 4873 (11), 5545 
d 6045 (11), Bos (11). Kirkbride Jr., J. H. & E. Lleras 
2860 (11). Knapp, S. & J. oa 2894 (14). Knob, A. et al. 
640 (17a). Koch-Grünberg, T. 74 (10). Krapovickas, A. & C. 
L. Cristobal 42946 (11). Krieger, P. L. 7828 (11). Krukoff, B. 
A. 1475 (17a), 6266 (17a), 6728 p ten (17b), 8767 
(17b), om d Mr L. P. et oe 
Lan 15 (11). Lasser. 93 (16). Liesner, R. 
o oS Y 3784 (10), ia T 6723 (10), 15682 
(17a), 16256 (17a), 19154 (16), 25864 (16), 24601 (17a), 
25720 o 25767 (1). Liesner, R. & F. 2). 
V. Funk 15847 (17a). Liesner, R. & B. Holst 
20668 (8), 20751 a uos L. G. & C. F. da Silva 20 
mann, L. G. e 8 (11), 70 s : A. et al. 
s.n. INPS 5 10. p H. 5121 
Maa. . M. et al. 4307 (13), Pp rm Macédo, A. 
1906 (6), 1906 [a] (6). Machado, F. & S. Assumpção S. s.n. 
(SP8976) (11). Maguire, B. & D. B. Fanshawe 32637 (13). 
i i uire, E m Politi 
27717 (12), 27799 (12). Maguire, B. et al. 30662 oo 
30788 (8), 56936 (11). WE L ei " 101-981 
7 17a). Martinelli, G. 6850 
Ta). We S. & M. Rimachi Y. 
19581 (17b), 25799 tie) do (17a). McDaniel, S. et a. 
24861 (17a). McDowell, T. 3250 (14), 3776 (17a). Wb, 
T. & D. Gopaul 2575 (14), pa (13). Mello-Silva, R. 
Pirani CFCR10874 (11). Mendonca, R. C. & F. qo > 
a Mendonça, R. C. et al. 2020 pu uu (1). M 
B., M. 1916 (3). Morawetz, W. Inófer 14-2 
qm, Moretti 656 (4). en S. & B. Pon, 14832 (14). pus 
S. & C. Gracie 21824 (17a). Mori, E i al. 14933 (14), 20442 
(17a), 21649 (17b). Morillo. G. e 3904 (10), 3908 (10), 
4134 ee e T Mose 4 & R. Abbott 2866 (11). 
Muniz, C. F. et al. s.n. CFSC78 
Ne M. Ps (5), 34758 ae 41162 (11). 
Oldeman, R. A. A. 2689 (17a), B-997 (4). Oliveira, P. & W. 
R. Anderson 428 (11). 


Pabst, G. 9098 (11). Parker s.n. (14). Pearce, R. s.n. (11). 
Pena, M. et al. 202 (5). Pereira, B. A. S. 957 (11). Pereira, B. A. 
S. de 


(11). Pipoly, J. J. & R. B 
AT 10191 (14). Pipoly, J. 
Pipoly, J. J. et al. 14909 (17a), 15005 (17b). Pires, M. J. P. et 
al. 819 > 819-a (16), 14109 eae Pivari, M. O. D. & D. 
Pifano - Plowman, T. unke V. 11670A (179). 
Poole, J. a 1771 (Vía), 1856 (173). pm G. T. 13876 (17a). 
Prance, G. T. & T. D. Pennington 1765 (17a). Prance, G. T. 
al. 1291 (2), 2560 (17a), 3366 (17a), 3730 (10), 4862 (10), 
6678 (11), 10659 (11), 10903 (17b), 11414 (9), 12191 (17b), 
14493 (17a), 14614 (17a), 15224 (17a), 20676 (17a), 23051 
(9), 24653 (17a), 28860 (2). Prévost, M.-F. 428 (14), 495 (17b), 
590 ers 1202 ee i Ai 3481 (17b), 4100 (7), 4109 


dd 8 (14). Ramírez C., R. 22 (17b). 
Ramirez, J. G. € a 1972 (17b). Renteria, E. et al. 
1545 (17b). Restrepo, D. & A. Matapi 484 (17a). Revilla, J. 
156 O 1353 (17a), 1746 (17a), 1790 (17a), 1865 (17a). 
Rimachi Y., M. 663 (17a), 6423 (17b), 7230 (17b), 8163 
(17a), 10382 a Robertson, K. R. & D. F. Austin 288 (14). 

Rodrigues, W. 9430 (8). Rodrigues, W. & J. Chagas 4610 
18). Rodrigues, W. & D. Coelho 4125 (10). Rodrigues, W. et 
al. 10254 (17a). Rohr, J. P. B. von 2001 (7). Rombouts, H. 
E. 655 (17b). Romero-Castañeda, R. 1229 (17a). Rosa, N. A. 

& M. R. Santos 1862 (8). s L. 1745 (11). Rudas, A. et al. 
5151 (Vib). Rusby, H. H. 11 

Saldías, M. 3790 (11). ae L. et "T 37 (11). Santos, 
G. dos et al. 240 (17a), 247 (17a), 249 (17a), 250 (17a), 272 
17a). Sastre, C. 6436 (17b). Sastre, C. 
Schomburgk, M. R 


— 


Lr R. & M. Schulte aa (11). Shepherd, A D7 
x wa, A. S. L d. . Silva, F. C. 2 

F. et al. 58192 a, [^ a. Silva, M. P F Pinheiro 
ue (11), 4452 (11). Silva, M. G. de C. Rosario 4917 (7). Silva, 

T. 1481 (17b), 4356 (1). v " T. & U. Brazáo o 

a Silva, N. T. & M. R. Sani 03 (1). Silva Manso, A. s 
(11). Simoes, J. 4 (11). Smith, J. x. x p 4164 (1). Smith, S. P 
et al. 1628 (17b). Soejarto & Cardozo (17a). 
Solomon, J. C. 17685 (11), D (11). Souza, V. C. & L 
Ferraro 2574 (11). m a 2 ^ al. 9846 (11). Spruce, R. 1721 
(17a). Steege, H. ter ei 8 (14). Stein, B. A. 2410 (17b). 
Stergios, B. 10318 8 A7). ie B.&G.A 

233 (1 


— 


0). Steyermark, J. A. 9. 
107474 qux rr 

Steyermark, J. A. & G. S. Bunting I 031 08 q odd J. 
A. & R. Liesner po (1 Pa Puy J. A. & S. Nilsson 
324 (13) Steyermark, e y monad 19 (1). 
Steyermark, > 3), D (1), 115553 (16), 
117818 (13), 122158 d 124541 (12), 126235 (17a). 


. Tessmann, G. 5150 (17b). Thomas, 
al. 5336 (10), 5591 (11). Tillett, S. S. & C. L. Tillett 
48837 (13) 45870 (17b). Triana, J. J. 4124-10 (3). 

1 7a) 


Vargas C., I. G. et al. 5790 . Vásquez, R. & N. 
Jaramillo 194 (17a), 1190 (17b), 6517 (17b), 10211 (17a), 
10212 (l7a) 10275 (17b), 11756 (17a), 13039 (17a). 
Vásquez, R. et al. 2749 (17b), 18604 (17b), 19529 (17b). 
Vicentini, A. et al. 818 (11). Vieira, G. et al. 108 (8). Vieira, 
M. G. et al. 990 (11) 


Volume 96, Number 2 Pool 323 
2009 Review of Distictella (Bignoniaceae) 


Walter, B. M. T. et al. 621 (11). Williams, L. 15309 ne Zarucchi, E L. 1293 (17a), 2109 uu Zarucchi, EA "s 
Woythowsi, F. 5140 Pi z AW were C. E. Barbosa 3437-a (1). Zarucchi, J. L & R. E. 
1033 (172). Wurd y 43008 n 1027 7 (17a). nia J. L. et al. FE ^3 3010 s bu 
Yanagizawa, Y. s.n. (7 e s.n. oes (11). ). 


A REVISION OF MALAGASY Zachary S. Rogers? 
GNIDIA (THYMELAEACEAE, 
THY MELAEOIDEAE)! 


ABSTRACT 


A systematic revision of Gnidia L. is presented based on an analysis of DE UE The O of the s 
adopte ed here includes Lasiosiphon Fresen. a excludes Atemnosiphon n Leandri and Dais L. Six mbin: s are made fo 
species previously recognized as Lastosiphon: G. ambondrombensis (Boiteau) Z. S. Rogers, G. " hibbe ae es "3 ore) Z. s. 
Rogers, G. humbertii (Leandri) Z. S. m n linearis (Leandri) Z. : oe G. occidentalis rs LoS. hee and G. 
perrieri (Lean dri) Z. S. Rogers. Two names, G. daphnifolia L. f. and G. kua ris, are resu urrected from synonymy with L. 
madagascariensis (Lam.) Decne. and L. decaryi Leandri, respectively, ada pertain to more broadly circumscri bed species. 
One new species, G. neglecta Z. S Rogers, is described. These changes res s in the recognition of 14 Pk cies, a endemic, 
making Gnidia the largest genus of ibo agasy M Ee p are m for 15 names: Dais gnidioides 
Baker, G. danguyana Leandri, L. bojerianus Decne., L. decaryi, L. dec ectus Leandri, L. danh var. Coa uk Leandri 
Le decur i var. tenerifolia p de o Loan, L "ildebrandtii CE Elliot, £. humbertii Leandri, L 
madagascariensis var. angustifolius Leandri, L. madagascariensis var. mandrarensis Leandri, L. occidentalis Leandri, L. 

errieri Leandri, and L. pubescens (Lam.) Decne. var. multifolius Leandri. Each species is illustrated, mapped, and assigned a 
preliminary IUCN conservation status. 


RÉSUME 


Le genre Gnidia L. est révisé sur la base d'une analyse de données morphologiques. Le genre est traité ici en n 

Lasiosiphon Fresen. mais en excluant Atemnosiphon Leandri et Dais L. Six nouvelles combinaisons sont effectuées pour des 

o déjà reconnues comme Lasiosiphon: Gnidia ambondrombensis (Boiteau) Z. S. Rogers, G. hibbertioides (S. Moore) Z.S: 

s, G. humbertii o Z. S. Rogers, G. linearis S ^ : s I: occidentalis (Leandri) Z. S. Rogers, et G. 

perrieri (Leandri) Z. S. Rogers. Deux noms, G. daphnifolia t G. lin sont retirés de la synonymie sous L. 

madagascariensis (Lam.) Decne. et L. decaryi Tenin. oe, et ils conecten désormais deux espèces plus 
en 


caryi var. erectus Lean 
hildebrandtii Scott-Elliot, L. humbertii Lean ri, L ma esis var. angustifolius Leandri, L. madagascariensis var. 
mandrarensis Leandri, L. occidentalis Leandri, L. perrieri Leandri, et L. pubescens (Lam.) Decne. var. multifolius Leandri. 
Chaque espéce est illustrée, cartographiée et assignée à un statut AN conservation provisoire de l'UICN. 
ey words: Africa, Atemnosiphon, Gnidia, IUCN Red List, Lasiosiphon, Madagascar, Thymelaeaceae, Thymelaeoideae. 


Gnidia L. (ca. 140—160 species; n 2003;  (Fresen.) Gilg, reaches as far east as Arabia, India, 
Peterson, 2006) is the largest genus in the Thym and Sri Lanka (Townsend, 1981; Herber, 2003 
laeaceae, belongs to the largest ou Thyme- Peterson, 2006). Gnidia was last revised in its entirety 
laeoideae, and is almost completely restricted to by Meisner (1857), but several significant regional 


* 


Africa and Madagascar. More than 100 species occur — treatments have been published in the important 
in South Africa's Western Cape Province alone African floras of the last century (Pearson, 1910; 

(Beaumont et al, 2001a; Bredenkamp € Beyers, Wright, 1915; Staner, 1935; Aymonin, 1966a, b; 
2003), and a single widespread species, G. glauca ^ Gastaldo, 1969; Robyns, 1975; Peterson, 1978). The 


! The author thanks the curators and support staff F the AA herbaria for providing loan material: A, B, BM, BR, F, G 

, NY, P, TAN, TEF, US, WAG. mber of other people should be acknowledged for their various 

contributions: S. Andriambalalonera and J. ee entered specimen data; F. Rakotonasolo and R. Razakamalala 
assisted with fieldwork; G. Schatz, M. Merello, G. McPherson, and P. Phillipson pur helpful comments on early drafts of 


the manuscript; M. Spencer assisted with the typification of the Linnaean names; L. Andriamiarisoa provided the peas 
P. Stevens translated the diagnosis into Latin; V. Malécot and M. Callmander en with the French abstract; B. Rye an 
anonymous reviewer provided the ews. The 2003 fieldwork was supported by the John Denver Memorial Scholarship 


granted from the International Center for Tropical Ecology at the University of Missouri-St. Louis. Funding for the 2006 
"i came from the Botanical Research Institute of Idaho. 
Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A. zac! I 
doi: 10.3417/2006114 


ANN. Missouni Bor. Garp. 96: 324-368. PUBLISHED ON 7 JULY 2009. 


Volume 96, Number 2 


Rogers 32 
Revision of Malagasy Gnidia (Thymelaeaceae) 


genus was last treated for Madagascar by Leandri 
(1950) in the Flore du Madagascar et des Comores. 
Generic delimitation of Gnidia and putative relatives 
has proven to be particularly problematic due to the 
considerable morphological variation in the characters 
that define the groups (see also Peterson, 1959; 
Aymonin, 1965; Rogers, 2006). A preliminary molec- 
ular study (Van der Bank et al., 2002) based on rbcL 
and trnL-F data suggests that the eed is para- 
h broader 
ore the taxonomic circum- 
scription of Gnidia can be adequately addressed using 
molecular data. 

Domke (1934), in the most comprehensive morpho- 
logical classification of Thymelaeaceae to date, placed 
Gnidia in subtribe Gnidiinae of tribe Gnidieae, along 
with six other African genera: Craspedostoma Domke, 
Cryptadenia Meisn., Dais L., Lachnaea L., Lasiosiphon 
Fresen., and Struthiola L. Tribal and generic limits 
have been defined traditionally by unique combina- 
tions of floral characters and, in some groups, 
supplemented by secondary leaf, bract, inflorescence, 
and fruit features (Meisner, 1857; Baillon, 1875; Gilg, 
1894; Domke, 1934). Herber (2003), the first author to 


exhaustively update Domke’s generic classification, 


Le 

led groups, 
each one corresponding roughly to one of the four 
tribes of Thymelaeoideae recognized by Domke 
(1934). The Gnidia group is defined n an puer 
hypanthium, a lateral style, a eed 
endosperm and relatively thin obi ae 
exceptions in these urs character states are 
known to exist for some of the genera in t 

e.g., the cei is reportedly rarely E 
in Dais (Peterson, 2006) and lacking in Kelleria Endl. 
(Heads, 1990), Atemnosiphon 
species of Gnidia (Rogers, pers. obs.), and the style 
is terminal in Drapetes Banks ex Lam. (Heads, 1990). 
Gnidia i is distinguished from other members within the 


Leandri, and two 


shorter than the tube, sessile or subsessile anthers, 
the petaloid scales (when present) borne near the 
uth of the articulated hypa 
relatively small or absent disk surrounding the base of 
the ovary (Herber, 2003). 
Historically, 


nthium, and 


the overlapping and inconsistent 
variation in the diagnostic characters of Gnidia led 
to the description of several segregate genera, namely 
Lasiosiphon (1838), Arthrosolen C. A. Mey. 
Gnidiopsis Tiegh. (1893), Rhytidosolen Tiegh. (1893), 
Englerodaphne Gilg (1894), Craspedostoma (1934), 
Basutica E. Phillips (1944), Pseudognidia E. Phillips 
(1944), Struthiolopsis E. Phillips (1944), m Atemno- 


siphon (1947). Most segregates were reduced to 
synonymy with Gnidia shortly after their publication 
without much controversy except for two genera, 
Lasiosiphon and Atemnosiphon. Lasiosiphon, originally 
segregated because of its 5- rather than 4-merous 
flowers, was maintained through the first half of the 
20th century in African floras (e.g., Pearson, 1910; 
Wright, 1915) and in the most important morpholog- 
ical classification (Domke, 1934). Leandri (1950) in 
the Flore des Madagascar des Comores was probably 
the last author to uphold Lasiosiphon as a distinct 
genus, recognizing 15 species of Lasiosiphon and four 
species of Gnidia from Madagascar (neither genus 
recorded for the Comoro Islands). Peterson (1959) 
argued strongly to synonymize siosiphon with 
Gnidia, a position based mainly on his observations 
of the morphological inconsistencies in the floral 
merosity of several African 
been follow 
also Rogers, 2006). Atemnosiphon, the newest pub- 


species, and his view has 
ed by subsequent authors (for a review, see 


lished uon (Leandri, 1947), includes a single 

MU DA (originally described in 
Lasiosiphon s L. coriaceus Leandri) and was only 
recently ed into synonymy with Gnidia by Herber 


— 


Ra Atemnosiphon does not fit very well within 


erber’s circumscription of Gnidia because of its 
Ta exserted filaments and overall uncharacter- 


istic appearance, e.g, leaves with a marginal vein, 


ruit that splits laterally 
through the side of the hypanthium during develop- 
ment. Aymonin (1965) noted similarities between 
Atemnosiphon and the widespread G. glauca, and the 
two do indeed share a similar inflorescence structure 
and lack an articulated hypanthium. Unfortunately, 
Atemnosiphon and G. glauca were not sampled in the 
molecular phylogeny of Van der Bank et al. (2002). 
ne other genus of Herber's Gnidia group occurs in 
Malagasy 
Unlike Lasiosiphon and  Atemnosiphon, 


Madagascar (Dais, one African and one 
species). 
however, Dais has always been retained as a distinct 
D. cotinifolia L., 
included in the molecular phylogeny of Van der Bank 
(2002) and formed a clade with Phaleria 
capitata Jack (99% bootstrap uate but the 


genu he African species, 


et al. 
in a larger 


e was unresolved w 
the m sampled 


Samoan i lands Rye, 1990) was 
Phalerieae (sensu Domke, 
Phaleria group (sensu Herber, 2003) because of the 
substantial morphological differences with Dais. 
Specifically, Phaleria differs fi 


rom Dais by having a 
2- versus l-locular gynoecium, a terminal versus 


Annals of the 
Missouri Botanical Garden 


lateral style, and a fibrous versus membranous or 
eshy pericarp. Morphology suggests that Dais is 
ied with Gnidia than Phaleria, but 
Dais can still be separated easily from Gnidia by the 


more closely a 


same androecial characters that separate Atemnosi- 
phon from Gnidia. Given the np ede d DE of 
the Gnidia group (Van der Bank al., des 
Rautenbach € Van der E ilc moe 
circumscription of Gnidia adopted for this revision 
includes P and provisionally excludes 
Atemnosiphon and D 

Since the Diea of the Flore de Madagascar et 
des Comores (Leandri, 1950), many new herbarium 
collections of Gnidia are now accessible, which, along 
with recent field observations, permit a reevaluation of 
the taxonomy of the Malagasy members of the genus. 
All taxa previously recognized by Leandri (1950) as 
Lasiosiphon are treated here as species of Gnidia, thus 
six new combination names are required. Aymonin 
(1962, 1965) previously indicated that several Mala- 
gasy names dii Pasto should be treated as Gnidia, 
b 


ut hi as valid transfers 


e regarded 
vui a to Article 33.4 of the International Co 
Botanical Nomenclature (McNeill et al., 2006) bu 
he did not provide full and direct references to the 
basionyms. A broader circumscription is adopted below 
for two geographically widespread and morphologically 
variable species, G. daphnifolia L. f. and G. linearis 
(Leandri) Z. S. Rogers, which Leandri (1950) previ- 
ously recognized by the names L. madagascariensis 
(Lam.) Deene. and L. decaryi Leandri, respectively. 
One new species, G. neglecta Z. S. Rogers, is described. 
These changes result in the recognition of 14 species 
(Appendix 1), all endemic, making Gnidia the largest 
Malagasy genus of Thymelaeaceae. 

Besides Gnidia, Atemnosiphon, and Dais, there are 
four additional genera and 15 species of Thymelaea- 
ceae occurring on Madagascar and in the adjacent 
Comoro Island archipelago: Peddiea Harv. ex Hook. 
(ca. 11 species: one Malagasy and ca. 10 African), 
DE Oliv. o o five Malagasy and one 

African; Rogers, 2005), S 
species: eight Malia and one Comorian; Rogers, 
2004), and Synaptolepis Oliv. 
Malagasy and about five African). 


D. Baill (nine 


(six species: 


MATERIALS AND METHODS 


Herbarium specimens of 530 collections 
ix 2) were examined from the following 
B G, GH, K, LINN 


ppendi 
institutions: A, B, yd G; » K, 


the author for eight of the 14 species recognized here 
during four collecting trips to Madagascar (January— 


March 2003, March-July 2004, January—February 
2006, and November 2006). Field trips were planned 
so that plants from as many populations as possible 
could be observed and sampled. Populations of Gnidia 
occurring in the northwestern and far western regions 
of the island were unable to be reached due to 
logistical constraints. 

The species concept and criteria follow those 
discussed in Rogers (2004). No 


are recognized. All examined specimens were data- 


infraspecific taxa 


based and are available on the Missouri Botanical 
Garden's Tropicos website (<http://www.tropicos. 
org/>) along with images of types and representative 
herbarium vouchers. Geographic coordinates an 
elevations were assigned, whenever possible, using 
the Gazetteer to Malagasy Botanical Collecting 
Localities (Schatz & Lescot, 2009). Full specimen 
data including post-facto coordinates and elevations 
are available from the author by request. Species 
distributions are mapped over the outlines of the five 
simplified bioclimatic zones of Madagascar as dis- 
cussed in Schatz (2000; following Cornet, 1974). Maps 
were created using ArcGIS 9 software (ESRI, Red- 
lands, California, U.S.A.). 

summary of all names treated in the taxonomic 
section is provided in Appendix 3. 


MORPHOLOGICAL CHARACTERS AND VARIATION IN 
MALAGASY GNIDIA 
HABIT 


The majority of species are erect, weakly to D 
branched, 
habit of a few species (e.g., Gnidia daphnifolia, c 


shrubs or treelets from 1-2 m ta 


linearis) can vary greatly, some individuals being 
small weakly branched shrubs, while others becoming 
trees up to 6 m tall with a diameter of 25 cm at breast 
height. Gnidia humbertii (Leandri) Z. S. Rogers differs 
from other Malagasy species by its rounded, densely 
branched habit, which rarely reaches 60 em in height, 
and also has a well-developed underground root 
system that probably links nearby plants (presumably 
clones). Some species (e.g., G. danguyana Leandri, G. 
linearis) are reported to resprout from a common 
rootstock, allowing them to survive local harvesting 
and the periodic burning (M. Madeleine & R. 
Ramiandrisoa, pers. comm.). It is quite evident that 
some populations of G. daphnifolia remain small and 
shrubby only because of frequent burning. Specimens 
from one particularly stunted individual served as the 
basis for the name Zasiosiphon suffrutescens Leandri 
(1947). M 


primary root with several smaller secondary roots, 


ost species, at l m tall, have a single 


each one possessing a few to many thin, fibrous roots. 


Volume 96, Number 2 
2009 


Rogers 32 
Revision of Malagasy Gnidia (Thymelaeaceae) 


Malagasy species generally exhibit some form of 
Du. is nd species have obvious ee 
ual mous branching ni- 

dioides (Bake) Dens (Fig 


known to possess a ess branching pattern 


8 
7A) is the oily s species 


composed of trichotomous orthotropic shoots that was 
discussed by Hallé et al. (1978) and referred to in that 
work by the illegitimate name “Gnidia bakeri Gilg.” 
The same branching pattern occurs in two closely 
related African species, G. bambutana Ed 
. and G. mollis C. H. 

and the architecture of all three was carefully dd 
and described in detail in Aymonin (1966c). 

have spiral phyllotaxy. Gnidia 
i (Fig 5A) and G. neglecta 
(Fig. 11A) differ markedly by their opposite to 
decussate leaves, but leaves on vigorous shoots may 


EM 


Ledermann ex Engl 


sometimes be subopposite or more rarely alternate. 
Internode a while quite variable in most species, 
m ambondrombensis (Boiteau) Z. S. 
G. humbertii (Fig. ie that 
adjacent leaf a us overlap along m 
stem. In both species, as well as G. n c UA Z. 
S. Rogers (Fig. 14), branches are covered by conspic- 
uous ré Scars. 
all Thymelaeaceae, the bark of Gnidia is 
Ps p fibrous, and in Malagasy Gnidia, it is 
usually longitudinally striate on younger branches. 
Bark on older branches of G. ambondrombensis and 
often G. humbertii frequently exfoliates. Most species 
have gray, brown, or black bark. Gnidia neglecta and 
G. razakamalalana have dark red bark. In G. perrieri 
(Leandri) Z. S. Rogers, bark on the young branches 
ries a characteristic red-purple or uid 
Lenticels are densely arranged on 
danguyana, G. decary 


and G. perrieri, 


the 
ana, G. bojeriana (Decne) Cil, 
whereas the lenticels are more 
sparsely spaced (when present) on the bark of G 
m ies G. linearis, and G. occidentalis ene 


INDUMENT 


unbranched, short, 
ed. Indument 


presence or been density, type, position, and 


richomes are unicellular, 
silver-colored, and generally appress 
length are taxonomically important characters. Spe- 
cific patterns of variation present on particular organs 
are discussed in subsequent sections. 


LEAVES 


eaf shape is p often obovate or elliptic. Gnidia 
gnidioides (Fig. 
linearis (Fig. 10B) have small needle-shaped ieaves, 


7A) and some populations of 


which are also prevalent in continental African 
species of the genus (e.g., G. mollis, G. pinifolia L.). 

he blades of some species with larger obovate leaves 
are often asymmetrical (e.g, G. daphnifolia, 
gilbertae Drake, and G. occidentalis). 

For a few species, the size and, to a lesser degree, 
the shape, of leaves may vary substantially between 
individuals growing in the same population, and even 
between leaves on the same branch (e.g., Gnidia 
daphnifolia). Beaumont et al. (2001a) conducted a 
morphometric oe 9 nidia species 
exhibiting a of variation in leaves (and 
bracts), lagasy species, G. dan- 
guyana and G. daphnifolia (as *G. mada, a 
sic), and found that the length to 
(lw) ratio of leaves and bracts are dues 


including hii "Ma 
var. baronii,” 


important characters, and that the l:w ratios of the 
leaves compared to the bracts are not correlated. Their 
results (Beaumont et al., 2001a, 
include all of the other bracteate Malagasy species. 


) can be extended to 


Leaf l:w ratios are more consistent and Redi. 
useful than linear size measurements i 
e.g., G. daphnifolia (Fig. 4A—C) versus " M ues 
(Fig. 12A, B). The largest leaf low ratios reach 14— 
15:1 in G. gnidioides (Fig. 7A) and G. linearis 
(Fig. 10A 

In live plants and sometimes on dry material of 
Gnidia decaryana, the leaves are appressed adaxially 
against the stems, thereby concealing most, if not all, 
of the sessile or subsessile few-flowered inflores- 
cences (Fig. For most species, leaves are 
generally persistent along most of the branch on a 


persistent in G. ambondrombensis (Fig. 

usually in G. gilbertae (Fig. GA) and G. humbertii 
(Fig. 9A). Leaf texture is oe chartaceous to 
coriaceous. whe esh, are usual 
lade surfaces 


of G. neglecta and G. perrieri are often glaucous, at 


arker in color adaxially. The abaxial 
least after drying. Leaves are shortly petiolate « or 
sessile, but this distinction becomes 
species with long-attenuate leaf bases. 

Most species have at least some trichomes on their 
leaves, at least initially. Leaves (and stems) are 
completely glabrous in four species: Gnidia dan- 
guyana, G. decaryana, G. neglecta, and 
Trichomes on leaves are generally appressed. The 


G. perrieri. 


subappressed trichomes of G. hibbertioides (S. Moore) 
Z. S. Rogers (Fig. 8A) may simply be a drying artifact 
of the 
Leaves of some species, e.g., G. bojeriana (Fig. 2A), G. 
gilbertae (Vig. 6A), and sometimes G. daphnifolia 

ig. 4D), may have 


present on the only known specimen species. 


persistent trichomes 
remaining on both surfaces of the blade. In these 


Annals of the 
Missouri Botanical Garden 


Figure 1. 


Gnidia 
gynoecium drawn from isotype, Boiteau (Hb. Jard. Bot. a E o tee dun Xen Mosen & Rakotonasolo 
706 (MO). 


ambondrombensis (Boiteau) Z. S. 


cases, however, the trichomes are faint and imper- 
ceptible by the unaided eye, and the indument does 
not generally obscure the leaf surface and venation 
pattern, at least not rep Alternatively, the 
o dense on both leaf 
surfaces in G. ciet (Fig. 1A) and G. 


sericeous indument remains 


Flow . Gynoecium. Habit 


humbertii (Fig. 9A), and on the abaxial leaf surface of 
G. razakamalalana (Fig. 14B) and rarely G. daphni- 
folia (Fig. 4D), that the surface and venation pattern 
remain completely hidden. Leaf trichomes o 

humbertii are generally matted and ca. 0.3 mm long, 
whereas those of G. andondendienat, G. razakama- 


Volume 96, Number 2 


Rogers 32 
Revision of Malagasy Gnidia (Thymelaeaceae) 


. Gnidia bojeriana (Decne.) Gilg. —A. Habit. —B, C. Flower. Habit drawn from Rogers de Randrianaivo 175 
(MO). Flower drawn from Rogers & Randrianaivo 183 (MO). 


lalana, and G. daphnifolia are straight and exceed 
1 mm in length. 

Leaf bases are usually cuneate to attenuate and 
ut the bases of Gnidia 
A) and G. neglecta (Fig. 11B) are at 
least slightly cordate. Within a species, leaf apices 
vary from obtuse to acute and the tips are usually 


decurrent along the petiole, 
danguyana (Fig. 


rounded or mucronate. Leaf margins are entire, a 
family characteristic, but may vary between flat to 
somewhat revolute, the latter character state occurring 
more frequently on wider leaves and especially near 
the base of the blade. 

Most Malagasy species of Gnidia exhibit some form 
of brochidodromous venation. The venation pattern 


Annals of the 


Missouri Botanical Garden 


Figure 3. 


consists of one, or sometimes two, submarginal loops 
interconnected with an anastomosing network near the 
margin (e.g., G. danguyana, Fig. 3A), or of strongly 
arcuate secondaries meeting the margin in the upper 
1/4 to 1/2 of the blade (e.g., G. daphnifolia, Vig. AB), 
or by a pattern only represented on the abaxial surface 
by a midrib and two or four longitudinal plicate veins 
(G. gnidioides, Fig. 7A). Most species have concolor- 
ous venation compared to the blade. Fine venation 
anastomoses in an irregular pattern when visible. 
Generally, venation is slightly raised and more 
prominently so abaxially. 


Gnidia danguyana Leandri. —A. Habit. —B, C. 


Flower. Drawn from Rogers et al. 76 (MO). 


INFLORESCENCES 


Species with opposite or decussate phyllotaxy (e.g., 
Gnidia decaryana) consistently have terminal inflo- 
rescences, whereas species with spiral phyllotaxy 
have axillary or pseudoterminal inflorescences, the 
latter condition being the most pronounced when 
branch internodes are short. In a few pedunculate 
species (e.g., G. danguyana, sometimes G. daphnifo- 
lia), the inflorescence can be extra-axillary with the 
peduncle being carried several millimeters above its 
axil as the stem develops. 


Volume 96, Number 2 Rogers 33 
2009 Revision of Malagasy Gnidia (Thymelaeaceae) 


H 
en Nds daphnifolia L. f. —A—D. Habits. Note large amount of variation in leaves, length of peduncles, and the 

Persist of 

hypanthium. —H. Gynoecium. Habits drawn from Rogers & Rakotonasolo 133 (part A, MO), Service Forestier (Rabevohiira) 


34925 (pait | B, TE, sin de ined Qat C, type of oa suffrutescens, TAN), and Humbert 18836 (part D, TAN). 
23 (part E, TEF), Rogers & Rakotonasolo 133 (part F, MO), Humbert 


ly 
18836 ‘(part T TAN), id [e Tae (part H, a 


Annals of the 
Missouri Botanical Garden 


Figure 5. 


There is substantial variation in the inflorescence 
structure of Malagasy Gnidia. Nine of the 14 species 
have many-flowered, long-pedunculate, involucrate 
inflorescences (e.g., G. linearis, Fig. 10A, C). The 
remaining species possess one of the following kinds 
of inflorescences: (1) 1-flowered, ebracteate, terminal, 
i (2) few- 
flowered, ebracteate, and long-pedunculate (G. ne- 
glecta, Fig. 11C); (3) few-flowered, bracteate, and 
sessile or subsessile (G. decaryana, Fig. 5B); (4) 


and sessile (G. razakamalalana, 


many-flowered, bracteate, and long-pedunculate with 
flowers racemosely arranged along a short fertile 
portion at the tip of an otherwise sterile peduncle (6. 
danguyana, Fig. 3A); (5) many-flowered, bracteate, 
composite-like heads (G. gnidioides, Fig. 7A). 
Peduncles are usually erect and sparsely pubescent 
with the same kind and density of indument as that 


Gnidia decaryana Leandri. —A. Habit. —B. 
Flower. —D. Dissection of caducous portion of hy 

(below). —E. Fruit. Habit drawn from Service Forestier (Capuron) 28650 (P). Inflorescence and flower drawn from Rogers & 
Rakotonasolo 108 (MO). Fruit drawn from Decary 4332 (TAN). 


salon geet ceils 
Dm 
Ed 


des ES 
AZ 


M 
" 


RNC 


i Inflorescence. Note small bracteole subtending flower. —C. 
panthium (above), and the persistent portion surrounding the gynoecium 


found on the leaves. Leandri (1931a, b, 1947, 1950) 


considered peduncle length to be an important 
taxonomic character to distinguish between species 
possessing involucrate inflorescences. His observa- 


tions were based, in most cases, on one or two 


majority of the species have highly variable peduncle 
length. For example, on a single specimen of G. 
daphnifolia the flowering peduncles can be 3-5 cm 
long, while others even in the fruiting stage might only 
reach 1.5 em long. Nonetheless, species can generally 
the length of their 
peduncles as follows: (1) completely absent (e.g., G. 
ambondrombensis), or nearly so (e.g., G. gnidioides); 


be classified according to 


(2) short, no more than 8 mm long (e.g., G. hibber- 


Volume 96, Number 2 
2009 


Rogers 33 
Revision of Malagasy Gnidia (Thymelaeaceae) 


tioides, G. perrieri); (3) long, reaching 5 cm or more in 
length (e.g., G. bojeriana, G. danguyana, G. daphni- 
folia). 

Involucrate inflorescences are always many flow- 
rarely six, 
of Gnidia 


gilbertae are composed of only four, or rarely five, 


imbricate bracts, 


bracts. Involucral bracts may or may not be strongly 
differentiated from the leaves; however, in many 
species the bracts are noticeably different from the 
leaves in shape and size and sometimes in texture and 
indument. In general, involucral bracts are substan- 
tially smaller than the leaves. The bracts of most 
species are 2 to 3.5 times longer than wide. Bracts of 
G. linearis are typically almost orbicular (ca. 1-1.5:1), 
whereas those of G. occidentalis are more lanceolate 
(ca. 3—7:1). The bracts within an involucre may differ 
noticeably, as in G. daphnifolia and G. perrieri, where 
usually one or two of the innermost bracts will be 
narrower, shorter, and not as strongly acuminate as the 
rest. 

Involucral bracts are usually broadly rounded at the 
base and acuminate or rostrate at the apex. Bracts 
rarely have conspicuous nervation, and the midrib is 
faint on both surfaces or only perceptible abaxially. 
Generally, bracts are appressed to the flowers and 
fruits, at least in their lower half, and remain 

e fruits have fallen off. The 
bracts usually fall off earlier in Gnidi 


persistent until all of t 
ia linearis and 
sometimes G. daphnifolia, exposing relatively long 
trichomes on the pedicel and lower portion of the 
hypanthium that resemble bristles on a brush (e.g., 
Fig. 4A) 


Generally, involucral bracts are coriaceous and 
indument typically remains densest on the abaxial 
surface and also on the upper half of the adaxial 
surface. Even after drying, bracts of Gnidia perrieri 
remain semi-succulent and become glaucous, as do 
the leaves. Trichome densities on the leaves and 
bracts are almost always positively correlated, but the 
indument on the bracts in G. bojeriana is much denser 
compared to the leaves. Gnidia decaryana (Fig. 5B) is 
the only Malagasy species with bracteoles. 


FLOWERS 


The flowers in Gnidia are hermaphroditic, diplo- 
stemonous, tubular, and either 4- (e.g., G. danguyana, 
Fig. 3C) or 5-merous (G. razakamalalana, Vig. 14 
Based on herbarium labels and personal observations, 
flower color is most often yellow. The character is 
usually always consistent within a single population; 
however, the feature appears to be more variable 
between populations of G. daphnifolia and G. linearis 
(yellow, orange, or red), G. decaryana (red-green, 


yellow, or greenish white), and G. gnidioides (pink, 
red, white, 
inflorescence of G. bojeriana change from yellow or 


or yellow). Flowers within a single 
orange to red in late anthesis. Similar color changes 
may be occurring in other species (e.g., G. daphnifo- 
lia) given the variation reported on specimen labels. 
Conspicuous yellow blisters usually develop on the 
flowers (and fruits) of G. danguyana after drying, 
similar to those found in other distantly related 
Thymelaeoideae (e.g., ae Lam., ia). 

ers are sessile to sho vedicellat e. The 
pedicels of species wien involucrate inflorescences 
are glabrous to moderately pubescent with short 
trichomes, and similar to those borne on the leaves 
and peduncles. Pedicels of species with involucrate 
inflorescences are almost always 0-1.5 mm long and 
hidden by long, silver, pedicel trichomes. The flowers 
of Gnidia bojeriana are borne on conspicuous 1.5— 
3.1 mm long pedicels (Fig. 2A) that are not always 
obvious in early anthesis without dissection. The 
pedicel trichomes may be similar in length to those 
borne on the persis 


stent portion of the hypanthium 


on 


e.g., G. bojeriana, G. perrieri), or much longer (e.g., 


gnidioides, G. linearis). In the latter case, the 


trichomes reach up to 4 mm long and resemble a 
dense, silver-colored brush (Fig. 7D). 

e floral tube is treated here as a hypanthium 
following Gilg (1894) and Herber (2003). For a 
thorough review concerning the various interpreta- 
tions of the structure, see Heinig (1951). In Malagasy 
Gnidia, the hypanthium is cylindrical or nearly so an 
the tube is almost always articulated slightly above 
the ov 
mene ae line falls away, while the p 


ry. The portion of the hypanthium above the 
ortion below 
the line persists and surrounds the loo fruit. In 
the remainder of this paper, the part of the tube above 
the line will be referred to as the caducous portion, 


genus are known to completely lack the articulated 
hypanthium, the continental African G. glauca and 
the Malagasy G. gilbertae (Fig. 6B). In G. gilbertae, 
the lower one third of the tube tears irregularly across 
as the fruit develops. In species with both articulated 
and unarticulated flowers, the fruit remains inside the 
persistent portion through dispersal. 

Leandri (1931a, b, 1947, 1950) considered the 
length of the hypanthium to be an important 
taxonomic character, and certainly there is some 
taxonomically distinctive variation in several species. 
Gnidia razakamalalana, with its 5 cm long hypanthi- 
um, has the longest flowers i in the genus and perhaps 
even in the family (Rogers, 2006). Tubes of the 
remaining Malagasy species approach but never reach 
2cm in length. In a few species, such as €. 


Annals of the 


Missouri Botanical Garden 


Figure 6. Gnidia gilberiae Drake. —A. Habit. —B. Flower. —C. Gynoecium. Drawn from Hb. Inst. Sci. Madag. 


482 (TAN). 


daphnifolia, the hypanthium may vary greatly (6.5— 
15 mm long) among populations and even between 
individuals within the same population. This obser- 
vation cannot always be explained by differences in 
the developmental stage of the inflorescences or 
flowers. The persistent portion of the hypanthium is 
usually 2-4(—5) mm long, except in G. hibbertioides, 
where it is 5-7 mm long. 

The density and length of trichomes on the 
hypanthium of Malagasy Gnidia are taxonomically 


important characters. Gnidia neglecta is the only 
species with a completely glabrous hypanthium, 
inside and out (Fig. 11C). For species with internally 
pubescent tubes, the indument is generally most 
prevalent just above the articulation. The inner 
surface of the hypanthia in G. ambondrombensis and 
G. danguyana is the most obviously pubescent of any 
species in Madagascar, but the indument itself usually 
remains only faintly visible with magnification (these 
trichomes are sometimes impossible to see when 


Volume 96, Number 2 
2009 


Rogers 33 
Revision of Malagasy Gnidia (Thymelaeaceae) 


flowers are wet). In most species, however, the outer 
surface of the hypanthium, above and below the 
articulation, is covered with a conspicuous dense, 
appressed or subappressed indument (e.g., G. hum- 
bertii, Fig. 9B). The trichomes on the outer surface of 
the hypanthium are either of uniform length (e.g., G. 
decaryana, Fig. 5C) or noticeably (i.e., ca. 2-5 times) 
longer on the persistent portion (e.g., G. linearis, 
Fig. 10D, E). In these cases, the trichomes borne on 
the persistent portion of the hypanthium are generally 

i lower half and become shorte 


approaching the articulation. The ratio of trichome 


A 


length on the upper versus lower portion of the 
hypanthium varies substantially between populations 
of G. daphnifolia, where rn populatio: 
flowers with trichome ratios of 3.5—5:1 (above:below 
the articulation) and southeastern populations fall into 
the 1 


northe ons have 


unique condition exists in Gnidia gnidioides 
where the hypanthium on the abaxial surface is 
densely pubescent above the articulation and com- 
pletely DA below (Fig. 7B, note the long brush 
of trichomes belongs to the pedicel). The long 


trichomes on the persistent portion of G. polycephala 
(E. Mey. ex Meisn.) Gilg, a mainland African species, 
aid in wind dispersal. Many other African species 
possess long trichomes on the persistent portion of the 
hypanthium and those probably serve a similar 
function. 

The number of calyx lobes in Malagasy Gnidia is 
consistently either four or five, with twice the number 
of stamens, respectively. Five pes have flowers 
with four lobes is s stamens): G. d. 
decaryana, G. gilb , G. gnidioi 
Aestivation of the pes lobes is always imbricate. 
During anthesis, lobes spread and normally become 
adaxially convex. For most species, the apex of the 
lobe is usually rounded (G. decaryana, Fig. 5C) or 
Fig. 10D, E) Lobes are 
generally obovate or oblong, ca. 1.5-3 X 1.5-3 mm, 
o 3 times larger than the petaloid scales 


emarginate (G. linearis, 


and ca 
(when those are present; discussed in next paragraph). 
Sometimes one lobe (when 5-merous) or two lobes 
4 the size of the 
y. calyx lobes are glabrous adaxially 


(when 4-merous) are ca. 
others. Generally 
and densely pubescent abaxially with an indument 
similar to, or slightly longer than, that present on the 
outer surface of the hypanthium. 

More than half of the Malagasy species have small 
scale-like structures located in the sinuses of adjacent 


calyx lobes (e.g., Fig. 14C, F). There has been much 
controversy regarding the interpretation. of these 
organs, which are regarded here as petaloid scales 
(see review in Heinig, 1951). Petaloid scales are 
membranous, glabrous, free, and equal in number to 
the calyx lobes. Six species lack these petaloid scales: 
Gnidia ambondrombensis, G. danguyana, G. decary- 
ana, G. gnidioides, G. neglecta, and usually €. 
humbertii. Scales generally fall into the 0.5-2 X 
.5-] mm range. Some species have a large amount of 
size and shape variation within a species (e.g., 
daphnifolia, G. linearis). Most Malagasy species have 
scales that are usually apically rounded, emarginate, 
ounded 


lobes. Petaloid scales are generally = same color or 


acute, or only rarely possess up to a few ro 


slightly lighter than the calyx lobes. No scale 
vascularization was obvious with a dissecting scope 
or any species. When fresh, the scales of some 
populations of G. linearis are weakly carinate and this 
ridge becomes darker after drying. 

The androecium of Gnidia consists of two equal 
whorls of four or five stam cycle with each 
whorl ned at different aie within the tube. 
No rudimentary or aborted stamens were observed, as 
is known to occur in a few continental CH species 
" Gnidia (e.g., - aberrans C. H. W 

eisn.). In a 
^ lower a were rarely observed as being about 
half the size of those in the upper whorl, but these 


t, G. anomala 
w Malagasy species, ae anthers of 


smaller anthers still produced pollen. Stamens are 
glabrous and introrse. Filaments are narrow, membra- 
nous, and usually almost completely adnate to the 
inner wall of the hypanthium (sometimes only weakly 
fused}, which results in sessile or subsessile anthers. 
The upper whorl of stamens is opposite the sepals 
mouth or partially 


and generally borne near the 


exserted. Rarely, the entire length of the anthers is 
exserted. The lower whorl al 
and i is borne ca. 0-2(-4) mm below the upper whorl. 
Anthers are basifixed, bithecal, and longitudinally 


ternates with the sepals 


dehiscent. 

small annular or cupuliform disk surrounds the 
base of the ovary in most Malagasy species and is 
referred to as the subgynoecial disk in the taxonomic 
treatment. The disk is glabrous, fleshy, and usually ca. 
0. in Gnidia 


3 mm tall, but reaches 0.7 mm tall 


ambondrombensis (Fig. 1D). Disks are completely 
absent or inconspicuous (i.e., less than 0.1 mm tall) 
nguyana and G. decaryana. 
disk is most often moot or shal- 
lowly lobed. Lobing is often irregular, and individual 
lobes rarely attain one third of the total height of the 

isk. 

The gynoecium is composed of a pseudomonomer- 
ous ovary (only one of the two uniovulate carpels 


Volume 96, Number 2 
2009 


Rogers 33 
Revision of Malagasy Gnidia (Thymelaeaceae) 


develops), a lateral style, and a capitate stigma. 
Ovules are anatropous and suspended. The superior 
ovary is completely free from the hypanthium and 
shortly stipitate at s base in most species. Most 
variation of the ovary is related to the density, 
position, and length of trichomes on the surface. 
Typically, the entire ovary is either sparsely pubes- 
cent or glabrescent. When pubescent, the indument 
on most of the surface is caducous with only a few 
trichomes remaining at the apex in mature fruit. In a 
few species, the apex of the ovary is covered by a 
conspicuous brush of relatively long, silver trichomes 
(e.g., Fig. 3C) 
through fruit. 


Gnidia danguyana, that persists 
The style is filiform, straight, and persistent in fruit. 
Styles are glabrous, but rarely a few trichomes are 
irregularly found along its length. Style length may vary 
significantly between populations and sometimes indi- 
viduals, but this variation is not associated with 
differences in the position of anthers within the tube, 
and thus the flowers are probably not truly heterostylous. 
Stigmas are always papillate and usually globose 
(Fig. 2C). At anthesis, in a single population or 
species, the stigma can be at the height of either whorl 
of anthers, several millimeters below the lowest whorl, 
or slightly above the articulation. Stigmas never 
surpass the upper whorl in any Malagasy species. 


FRUITS 


Fruits are small single-seeded achenes that remain 
completely surrounded by the lower portion of the 
hypanthium when mature (Fig. 5D, E). Most species 
have ellipsoid fruits from 2—4 mm long. Two species 
have distinctly ovoid fruits, Gnidia bojeriana and G. 
danguyana. In fruit, the persistent styles in species 
with articulated hypanthia will protrude through the 
apical circular aperture of the persistent portion 
(Fig. 1C) 


membranous, except in 


. The pericarp is dry, thin, and usually 
bojeriana where it is 
Most of the 
indument on the ovary is lost as the fruit develops, but 


commonly more fleshy and opaque. 
the fruits of two species, G. danguyana and G 
decaryana, retain the distinctive brush of apical 
trichomes from the ova 


SEEDS 


Seeds show minor variation with the shape and size 
en matching that of the fruit. The seed coat is 
crustaceous, thin, and can be either black or dark 
brown. The embryo is fleshy and the endosperm is 
absent from mature seeds. Cotyledons are slightly 
flattened and the radicle and 
plumule are minute. 


in cross-section, 


DISTRIBUTION AND ECOLOGY 


l4 species are Malagasy endemics and their 
vided in Figures 15-17. Given 
the wide distribution of some species of Gnidia in 


distributions are pro 


Madagascar, it is rather surprising that the genus has 
not been found in the nearby Comoro Archipelago. 
The genus occurs in all five provinces (Antananarivo, 
Fianarantsoa, Mahajanga, Toamasina, Toliara) and in 
each of five simplified bioclimatic zones (humid, 
umid, montane, dry, subarid) as discussed in 
ae (2000). The general bioclimate transitions from 
wet to dry from east to west across the island, but the 
far north is particularly arid. 
Clearly, 
elevations (ca. -2 m} along 
lateau in the subhumid and 
zones, e.g., Gnidia ambondrombensis, G. gnidioides, G. 


some species are restricted to higher 
entral 
montane bioclimatic 


perrieri. Gnidia bojeriana is basically limited to mid- 


elevations in the subhumid zone. Gnidia danguyana 

nd G. neglecta are endemic to fragments of coastal 
littoral forest near sea level in the island's humid 
bioclimate. Gnidia gilbertae and G. occidentalis occur 
lateau in 


rier western side of the central 


Gnidia linearis is almost completely confined to the 
subarid zone in southern and southwestern Madagas- 
car. Five species appear to be narrow endemics: G. 
Ambondrombe), G. hibber- 


tioides (exact locality unknown), G. neglecta (Ande- 


ambondrombensis (Mt. 


voranto), G. perrieri (Andringitra), G. razakamalalana 


— 


Ivohibe Forest). Two species, G. neglecta and G. 
hibbertioides, are only known from their type collec- 
ion. 


+ 


Po ed Malagasy Gnidia are associated with 
sunlit areas and found in habitats area 
disturbed by fire. Gnidia gnidioides grow he 


and other moist places. Numerous as revised 


Malagasy genera in er families, a Buxus L. 
(Schatz & Lowry, Ehretia P. Browne (Miller, 
2002), and Leptolaena a (Schatz hal 2001) 


are ped associated with, and xd to, 
particular soil and rock types. This does not appear to 
be the case for most species of Cnidia on the island. 
Gnidia daphnifolia has been recorded on a number of 
diverse substrates including sand, laterite, limestone, 
granite and gneiss, and perhaps marble, whereas G. 
linearis, mostly confined to the southern and south- 
western part of the island, occurs on sand, sandstone, 
and calcareous limestone. Based on label data and 
personal observations, four species apparently show 
substrate-specific distributions: G. danguyana and G. 
neglecta only occur on sand; G. ambondrombensis and 
G. razakamalalana grow on black soil on granite rock. 


Annals of the 
Missouri Botanical Garden 


In regards to phenology, many Malagasy species, 
Ree widespread ones, may flower and fruit year 
und, at least within some part of their geographic 


range. 


VERNACULAR NAMES AND USES 


According to dozens of herbarium specimen labels 
and several literature sources (e.g. ud 


1998), the most 


in Madagascar is havoa (also written as “avoha” i 


common vernacular name for Gni 
“havoha”), a Malagasy word for fiber. Fibrous bark of 
some Gnidia species (e.g., G. linearis, and probably G. 
danguyana and G. daphnifolia) is harvested as the 
raw material used in the fabrication of papier 
mora, a coarse paper that was eae gathered 
cally en 1933), but 
part of a much larger handicraft baile and sold to 
tourists (M. Madeleine & R. ndrisoa, pers. 
co pers. obs.). (1998) 
reported that the bark of Dais glaucescens Decne. in 


E used lo is use 


ani 
m.; Rogers, Randriatavy 
C. A. Mey., which superficially resembles some other 
species of Gnidia, is also known as havoa and used to 
make papier Antaimora. Several other species of 
Thymelaeaceae are jn by havoa or one of its 
phic variants in Madagascar, such as Octole- 
pis dioica Capuron (Rogers, 2005), three species of 
Stephanodaphne (Rogers, 2004), and Peddiea involu- 


crata Baker. 


orthogra 


In addition to paper, local artisans have reported that 
the bark of Gnidia is used to make ceremonial clothing 
(M. Madeleine & R. Ramiandrisoa, sien comm.), and 
DM of G. daphnifolia is used 

226). The leaves of G. gilbertae are a to induce 
ae (Randrianaivo et al. 614) 


cordage (Luckow 


CONSERVATION STATUS 


At least some Malagasy species of Gnidia survive 
annual burning regimes and grazing by livestock 
through resprouting from underground rootstocks 
(Rogers, pers. obs.). 
harvested for the papier Antaimora trade are reported 
above 
ground growth is cut (M. Madeleine & R. Ramian- 
drisoa, pers. comm.). Vigorous populations of G. 
G. daphnifolia, 


linearis were observed in Us disturbed unprotected 


few species commonly 


to regenerate from a xylopodium after the 


bojeriana, gnidioides, and 
areas, including cow pastures, cultivated fields, Pinus 
. stands, and along the edges of towns. Altogether, 
these factors suggest that most Malagasy Gnidia are 
not particularly threatened, and thus seven of the 14 


species are considered species of Least Concern (LC) 


by IUCN (2001) criteria. Gnidia neglecta and G. 

hibbertioides are the o 

tus of Critically Endangered (CE), 
lan G. ambond. 


y two species assigned a 


a 
zakamalalana and rombensis are 
Endangered (EN). Gnidia danguyana and G. perrieri 


are considered Vulnerable (VU) to extinction. 


TAXONOMIC TREATMENT 


Gnidia L., Sp. Pl. 
pinifolia L. 


1: 358. 1753. TYPE: Gnidia 


Dessenia Adans., Fam. Pl. 2: 285. 1763, nom. su 
Lasiosiphon Fresen., Flora 21: 602. 1838. TYPE: MN 
glaw = Gnidia glauca (Fresen.) Gilg]. 


Arthrosolen spicatus (L. f.) C. A. Mey. [= Passerina 
i Lf]: 


spicata 


Shrubs or small trees. Leaves alternate to opposite, 
sometimes decussate; venation RU brochido- 
often 
leaves. Inflorescences terminal or axillary, usually 


romous, conspicuous in edle-shape 

composed of many-flowered heads borne on elongat- 
ing peduncles; bracts usually involucral, less often 
Flowers 


foliose, poorly differentiated, or absent. 


hermaphroditic, tubular, 4- or 5-merous, actinomor- 
phic; hypanthium = cylindrical (in Madagascar) to 
funnel-shaped, articulated above the ovary (unartic- 
ulated in Gnidia glauca in Africa, G. gilbertae in 
Madagascar, and also ae G. razakamalalana) 
upper portion o panthium caducous late 
anthesis; lower botica of UM persistent; 
calyx lobes 4 or 5, imbricate, spreading; petaloid 
scales usually small or absent, alternisepalous (adnate 
to the sinus between adjacent calyx lobes), free, thin, 
membranous; androecium diplostemonous; stamens 8 
or 10, arranged in 2 alternating whorls at different 
heights, introrse, included or rarely only the upper 


anthers basifixed, bithecal, longitudinally dehiscent, 
sessile or subsessile; subgynoecial disk small or 
absent, annular or cupuliform when present; gynoeci- 
um pseudomonomerous; ovary sessile to shortly 
stipitate at base; ovule l, apical, anatropous; style 
lateral, filiform, persistent; stigma usually capitate 
and globose, rarely fusiform or di shaped, papillate, 
included or nearly so. Fr small, single-seeded 
achenes, enclosed by the imn lower portion of 
the hypanthium; pericarp thin, dry, membranous, 
rarely coriaceous or semi-fleshy. Seeds with a thin, 
crustaceous coat; endosperm absent. 


Volume 96, Number 2 


Rogers 33 
2009 Revision of Malagasy Gnidia (Thymelaeaceae) 


KEY TO THE MALAGASY SPECIES OF GNIDIA 


la. eee lobes 4; stamens 8. 
. Phyllotaxy opposite to decussate, sometimes subopposite or less often alternate on vigorous shoots; 
inflorescence bracts 0 to 4, not imbricate. 
3a. Leaf blades 148.3 X 0.6—5 cm; petioles 1-4 mm long; inflorescences 6- to 23-flowered, flowers 
racemosely arranged along a short, fertile portion of a much longer, mostly sterile peduncle 


uL gee cepa aia D DR MM EM as A a EE . G. danguyana 
3b. Leaf blades 0.4-1.8 X 0.2-1.2 cm; petioles to 1 mm long; inflorescences 2- to 4-flowered, flowers 
arranged in a sessile, subsessile, or long-pedunculate cluster. 
4a. 


Leaves broadly ovate, leaf base sonans fine venation clearly visible and uniformly reticulate, densely 


congested and darker than the blade; inflorescences with peduncles to 2.5 cm long; hypanthium 
abr l 11. G. neg 


IS 
c 


y 
. Leaves obovate to suborbicular, leaf base cuneate to attenuate; fine venation usually inconspicuous, 


when yisihle deal anastomosing and concolorous with the blade; inflorescences sessile or subsessile; 
5. 


————— CREER . decaryana 
2b. EN alternate, rarely subopposite; inflorescence bracts 4 to many, imbric 
5a. Le a ais Le. -shaped, rarely very narrowly obovate or ovate; POP 50- to 100- ip p d; 
in composite: elike A le A n eM: Ue 
5b. Leaf [ne ae obovate to nearly elliptic; n to 28-flowered; flowers armed in 
pitulum and surrounded by 4 or 5 involucral bra 
lb. PE um 5. stamens 10. 
6a. Inflorescences 1-flowered; hypanthium ca. 5 cm long ......... llle. 14. G. razakamalalana 
6b. LC EE 6- to 40-flowered; pati m to 1.9 cm long. 


Ace ales dirais eu tatem eels eee pes 6. G. um 


x.lobes: 728:3 X AA e ete ates OA deae es Rc ed etae 1. G. ambondrombensis 
d Pu lobes rarely to 5.2 x 2. Boh 
8a. Petaloid scales 4.6-6 X A 84. 7 mm, upper half irregularly lacerate or sinuate; persistent portion 
livpanihniim-5—0 mm lone A rre a ek ae e nas . hibbertioides 
8b. Petaloid scales rarely to 3 7 X 2 mm, upper half acute to rounded, emarginate, or rarely with a few 
rounded lobes; ta portion of hypanthium rarely to 5 mm long. 
unded, compact, densely ramified subshrubs; both leaf surfaces hidden by a dense sericeous 
WCUMONE e a o A A Sa a weed ts 9. G. humbertii 
9b. 


Laxly branched shrubs to small trees; both leaf surfaces usually visible, or only the abaxial leaf 
surface rarely obari red by dense strigose or tomentose indument. 
l0a. In 


dri pi ne on living 


KE involucral bracts ie recurved (most obvious 

on fresh material, ‘the er e (698-1 6 mm wide; pedice m long; Don 

portion of hypanthium at least twice the diameter of the 25 ucous portion ... 2. G. bojeriana 

Inflorescences erect on pu plants; involucral bracts erect or spreading, and planar or 

reflexed near the midpoint (but never recurved), the largest rarely to 8 mm wide; pedicels 

0-1. long; persistent and caducous portion of the hypanthium roughly the same 

diameter 

Broadest leaf blades 7-21 mm wide; bracts usually long-acuminate or rostrate, 

often a reflexed upper half, less often short- acuminate o or acute and + erect. 

12a. cal bracts generally broadly ovate (lw ratios 2—4:1), usually dens 
black or brown in the lower half; eee 654 2(- - mm lon 
widespread throughout Madagas CANE eA tice Hara ft aie tae Oa fo 4. G. daphnifolia 

12b. Involucral bracts narrowly lanceolate or elliptic-ovate (l:w ratios 3—7:1), 
ee light green, green- -red, or yellow-brown tg ii 12.5- 


= 
= 
c 


m long; restricted to northwestern Madagascar ...... 2. G. occidentalis 
iude leaf blades 2-5(-9) mm wide; bracts um a 1-2 mm bos strongly 
decurved apicule (usually G. linearis), short-acuminate and spreading, or less often 
acute and + erect. 
13a. Plants with completely glabrous stems and leaves; young stems red-purple or 

orange-red when dry; longest peduncles 2-5(-8) mm long; involucral bracts 
ith l:w ratios of 2-3.5:1; subhumid region, endemic to Ed. a, 2000— 


2550 m elevation nmm e pomier 


black when dry; longest de (5510-50 mm long; involucral bracts ed 
I:w ratios of 1-1.5(-2):1; dry and subarid regions in southern and western 
1 


Madagascar, 0-1400 m elevation ....oooooccccccccccccccccn.. 0. G. linearis 


1. Gnidia ambondrombensis (Boiteau) Z. S. Rogers, drombe, rocky summit, 1900 m, 11 Apr. 1941, P. 
comb. nov. Basionym: Lasiosiphon ambondrom- Boiteau (Hb. Jard. Bot. Tananarive) 4643 (holo- 

bensis Boiteau, Bull. Trimestriel Acad. Malgache, type, P5; isotypes, MO!, TAN). Figure 1. 

n.s., 24: 83. 1941 [1942], as “ambondrombense.” 


Sparsely branched subshrubs to 60 cm tall; inter- 
TYPE: Madagascar. Fianarantsoa: Mt. Ambon- 


nodes inconspicuous; young branches densely seri- 


Annals of the 
Missouri Botanical Garden 


ceous to tomentose, covered with prominent leaf scars; 
bark of mature branches often exfoliating. Leaves 
alternate, spirally arranged, sessile, very closely 
spaced (adjacent leaf bases overlapping), persistent 
only at tips of branches; blades broadly obovate or 
elliptic, 1.5—2.3 X 0.5-1.1 em, l:w ratios ca. 2-3:1, 
silver-green, both surfaces completely obscured with a 
dense sericeous indument (trichomes 1-1.5 mm), 
apex apiculate or obtuse, base long-attenuate; midrib 
and venation obscured by indument. Inflorescences 
terminal, erect, capitate, 8- to 15-flowered, sessile, 
surrounded by a rosette of leaves; involucral bracts 5, 
similar to leaves (ca. 3/4 the size and less obovate 
X 3.3-3.8 mm, 
nsely pubescent in upper half 


compared to adjacent leaves), 8.3— 
l:w ratios 2.3-2.7:1, de 
adaxially, sparsely pubescent to glabrescent in lower 
half abaxially; midrib strongly raised adaxially, 
inconspicuous abaxially; nervation inconspicuous on 
both surfaces, or very faint in lower half rd 
Flowers 5-merous, yellow; pedicels 0.7-1 
covered by 0.3-0.8 mm trichomes; E 4. 
16 mm, articulate (line sometimes faint), coriaceous; 
caducous portion covered by dense indument exter- 
nally, trichomes ca. 1 mm, sparsely to moderately 
pubescent internally; persistent portion 3—4 mm, 
covered with dense indument externally, trichomes 
1-2 mm, moderately pubescent internally; calyx lobes 
5, spreading or reflexed, broadly elliptic to obovate, 
7-8.3 X 3.54.2 mm, more membranous than hypan- 
thium, glabrous adaxially, densely to moderately 
pubescent abaxially, apex emarginate; petaloid scales 
absent; stamens 10, upper whorl of anthers ca. half- 
exserted, lower whorl 1.5-2 mm below upper whorl; 
anthers oblong, 1.2-1.5 X 0.4-0.6 mm, subsessile; 
subgynoecial disk cupuliform, 0.4—0.6 mm tall, gla- 
brous, fleshy, apex irregularly lobed, sinuses shallow 
to deep; ovary ellipsoid, ca. 1.6 X 1 mm, stipitate 
(stipe 0.3-0.4 mm), lower half glabrescent or sparsely 
pubescent, upper half moderately pubescent, apex 
densely pubescent, trichomes 1-1.5 mm; style 3- 

l mm, glabrous; stigma inserted, ca. 2 mm below 
lower whorl of anthers. Fruits not seen. 


Distribution and habitat. Gnidia ambondrombensis 
is endemic the windswept summit of Mount 
Ambondrombe from 1800-1900 m elevation (Fig. 15). 
The site is situated along the a separating the 
ubhumid and hi imid biocli 


population grows on scattered patches of soil on a large 


The only known 


outcrop of weathered granite that caps the peak. 
Phenology. The species has been collected in 
flower in April, May, and October. 
Vernacular name. 


Tananarive] 4643). 


Borona (Boiteau [Hb. Jard. Bot. 


IUCN Red List category. More than 100 individ- 
uals of Gnidia ambondrombensis were observed on the 
Mountain in 2004 


summit of Ambondrombe (Rogers, 


pers. obs.). While the site is not formally protected, 
the remaining forest on the mountain, located between 
1400 and 1900 m elevation, are considered sacred by 
the local people. As a result, woodcutting and burning 
inside the forest are regarded as taboo and are not 


permitted by leaders living in the villages below the 


ro more protection than 

government- NN lands. Nevertheless, ine 
al fields border the edge of the forest and some trees 
are still selectively cut down for local use. This 
pe is assigned a conservation status of Endan- 

ed (EN) according to IUCN (2001) criteria because 
uc species is known from a single unprotected 
population with an estimated area of occurrence 


AOO) of 1 km? (Blab + 2ab; D1). Our attempt to 


bring this attractive species into cultivation in 


— 


proved unsuccessful, but additional efforts should be 
made to ensure the longevity of the species. 


Discussion. Gnidia ambondrombensis is 
nized by its sparsely branched habit, closely a 
leaves that are covered on both surfaces by a dense 
sericeous indument (trichomes 1-1.5 mm long), its 
X 3.5- 


sessile inflorescences, and by its large (7-8.3 
.2mm) e another 


alyx lobes. Gnidia humbertii, 
species with very dense vegetative indument and 
closely arranged leaves, differs by its distinctive 


rou 
leaves (1.9-3 vs. 5-11 mm wide) 
sericeous indument (trichomes 0.2-0.5 mm long), 
and its much smaller (24.5 X 1.7-2.5 mm) calyx 
lobes. 
Selected rs examined. MADAGASCAR. Fianar- 
mbondro 


antsoa: Mt. mbe, summit, Rogers & Rakotonasolo 
706 (K, MO e P, TAN). 


2. Gnidia bojeriana (Decne.) Gilg, Nat. Pflanzen- 
2 


m. aj 


-—- 
w 


94. Basionym: a 
bojerianus Deene., oy. Inde 4: 
TYPE: Miden 
mtns., s.d., W. Bojer s.n. (lectotype, designated 
here, P 00370315!; isotypes, BM!, K!, P [2]). 
Figure 2 


Antananarivo: seen 


Gnidia m Baill., Hist. Phys. E 35(5) [Atlas 
3], pl. 312 1895, nom. illeg. TYPE: *Madagascar" 
(type, pl. 3121, Baillon in Grandidier, 1895). 

Shrubs to 1.5 m tall; bark dark gray-brown, usually 
lenticellate, densely pubescent on young branches. 
Leaves alternate, persistent on older branchlets, 
subsessile or petiolate; petioles to 4 mm, densely 


Volume 96, Number 2 
2009 


Rogers 34 
Revision of Malagasy Gnidia (Thymelaeaceae) 


pubescent; blades narrowly elliptic or obovate, 1.6— 
7(-8.4) X 0.3-1.1(-1.5) em, l:w ratios ca. 4—7:1, both 
surfaces densely to moderately strigose (trichomes ca. 
1-1.5 mm), apex apiculate, base long-attenuate; 
midrib plane or slightly depressed adaxially, raised 
and lighter than blade abaxially; venation raised on 
both surfaces, more pronounced abaxially. Inflores- 
cences axillary, pecie drooping when living, 
involucrate, (22- t0)30- 
(0.4—)1—5 cm, densely Aiea involucral bracts 5, 
broadly ovate, (9-)13-25 X (6-)8-16 mm, l:w ratios 
2-3:1, persistent, 


40-flowered; peduncles 


green-yellow, becoming strongly 
recurved (often only obvious on fresh material), both 
surfaces obscured by dense pubescence, apex apic- 
ulate or acute, base rounded to truncate; midrib plane 
or inconspicuous adaxially, raised abaxially; nervation 
usually conspicuous. Flowers 5-merous, yellow 
becoming orange-red in late anthesis, pedicellate; 
1.5-3.1 mm, light green, densely puberulent, 
trichomes 0.2-0.4 mm, slightly longer near flower; 


pedicels 


hypanthium 9-11 mm, articulate, coriaceous, densely 
0.2-0.5 mm 


sometimes slightly longer and denser below articula- 


pubescent externally, trichomes ca. 


tion, glabrescent to moderately puberulent between 
anthers and ovary internally; caducous portion ca. 
0.5 mm wide, becoming light brown after anthesis; 
persistent portion 3—4 mm, ca. 2 mm diam. (i.e., at 
least twice the diam. of the caducous portion); B 
lobes 5, spreading, broadly elliptic or obovate, 1.9— 
2.8 X 1-2.1 
cent abaxially, apex emarginate or rarely rounded; 
0.5-0.9 0.2— 

0.3 mm, membranous, glabrous, lighter yellow than 
the calyx lo 


mm, glabrous adaxially, densely pubes- 
petaloid scales 5, ovate-oblong, 


es when fresh, apex emarginate or 
rounded; stamens 10, upper whorl of anthers ca. 1/2 
to 3/4. exserted, lower whorl 0.2-0.5 mm below upper 
whorl; anthers oblong, 0.8—1.2 X 0.2-0.4 mm, sessile 
or subsessile; subgynoecial disk cupuliform, t 

mm tall, glabrous, fleshy, 
irregularly lobed; ovary ovoid to ellipsoid, 1.5-1.7 
X 0.6-0.7 mm, stipitate (stipe to 0.3 mm), moderately 
to sparsely pubescent, denser in upper half, trichomes 


[a] 


0.4—0.7 mm; style 2.7—5.6 mm, glabrous; stigma near 
mouth or at height of lower whorl of anthers. Fruits 
ovoid, 3.2-3.5 X 1.5-1.7 mm, sparsely to moderately 
pubescent, denser in upper half; pericarp membra- 
nous or flesh 


Distribution and habitat. Gnidia bojeriana occurs 
along the central plateau of Madagascar from 
Antananarivo to the Mahafaly Plateau and Isalo from 
800-1700 m elevations (Fig. 15). The species is 
commonl sunlit a 


found in open eas and notably 


present in Madagascars tapia forests, woodlands 


dominated by species of Uapaca bojeri Baill. (Eu- 


phorbiaceae). Gnidia bojeriana grows on gneiss, 
Wenn and sandstone rock oe in the subhumid 


and subarid bioclimatic zon 


Phenology. The species flowers and fruits from 
January through June. 
Vernacular name. 

Bot. Tananarive] 50. 


Kelimafana (Boiteau [Hb. Jard. 


IUCN Red List category. Gnidia bojeriana is 
widespread and has been recorded in one protected 
area (Isalo). Populations dating back to 1928 at Isalo 
and 1959 near Arivonimamo were found again in 2006 
and 2003, respectively, and more than 75 healthy 
individuals were seen in the latter population pecs, 
pers. obs). The habitat in both areas is burn 
periodically, and the species obviously E 
frequent disturbances over time. Gni ojeriana 

hou nsidered a species of Least Concern (LC) 
by IUCN (2001) criteria. 


Discussion. Gnidia bojeriana is easy to recognize 
by its broad, strongly recurved involucral bracts 
measuring (68-16 mm wide, its distinctly pedicel- 
ate flowers (pedicels 1.5-3. ong) and its 
hypanthium with a persistent portion at least two 
times the diameter of the caducous portion. It is also 
worth noting that the recurved a 


arance of 
bracts is usually lost on dried material, and that 


the 
at the 
peduncles on live plants droop distinctively. 

Nomenclature and typification. The provenance in 
the protologue (Decaisne, 1844: 149) was cited as 
*Hab. E ou in montibus provinciae Emirnae 

s. Paris." without mention of a specific 

collector or End number. However, judging from 
the epithet it seems likely that the name was based on 
material that was either SE d or provided by 
Wenceslas Bojer (1795-1856). Three 
sheets attributed to o. were found at P bearing 
labels with the typewritten script Herb. Mus. Paris. 
Decaisne's handwriting does not appear to be present 
on any of the specimens, but one sheet (P 00370315) 
has the exact locality information cited in the 
protologue (except for the final “e” mi 
old provincial name Emirnae). Specimens on al 
sheets match the original description, and sheet P 
00370315 is designated as the lectotype. Two other 
Bojer sheets of Gnidia bojeriana at BM and K are 
regarded as isolectotypes. 


unnumbered 


missing from the 


Original material of the later validly published 
a ae Bai 
1894 


nym Gnidi 5) non 6. 
bojeriana (Decne.) G 


189 
, 18 represented i in the 
protologue by a nice pia illustration bearing 
the inscriptions “Madagascar” and the name of the 
species, the latter of which did not include any 
authorship or other attribution to an earlier basionym 


Annals of the 
Missouri Botanical Garden 


. The plate 
taxonomically to 


name published by Decaisne or Gilg 
without doubt corresponds 
bojeriana (Decne.) Gilg and must be regarded as the 
type of Baillon's name. 


Selected specimens examined. MADAGASCAR. Antana- 
arivo: Imerintsiatosika, Km 22 along Natl. Rte. 
(Antananarivo—Imerintsiatosika), Rogers Randrianaivo 
175 (BR, G, K, MO, P, TAN); Mahevelona, rte. Majunga, 
Km 47, Rauh 1677 (TAN); Mt. Antongona, W of Antananar- 
ivo, Perrier la m 18459 (G, P ne TAN). 
Fianarantsoa: Isalo Natl. Park, Tsimanabaro Tapia forest, 
Rogers et al. 821 (K, MO, P, TAN, US); Itremo, Perrier de la 
Báthie 12472 (P). 


3. Esa rcs Ep Bull. Soc. Bot. France 
: 35. 1930. E: 


Aa fore "i 


Madagascar. Toamasina: 
po 1923, M. Louvel 118 
(lectotype, designated here, P!). Figure 3. 


Shrubs or small trees to 6 m tall; young branches 
light green, densely to moderately pubescent; mature 
branches light brown-red to dark black-brown; bark 
often lenticellate. Leaves o 


osite to can pairs 
rrange 


sometimes suboppositely or alternately 
vigorous shoots, imd petiolate; sales 1-4 mm, 
glabrous; blades broadly ovate or ovate-elliptic, 1-8.3 
X 0.6—4.5(—5) em, l:w ratios ca. 1.5-2.5:1, lighter in 
color abaxially, glabrous on both surfaces, apex 
acuminate, apiculate, or acute, rarely emarginate, 
margin with a distinet vein, base cordate or slightly 
cordate; midrib depressed adaxially, raised abaxially, 
glabrous, lighter green than blade on both surfaces; 
venation raised on both surfaces, more pronounced 
abaxially. Inflorescences terminal, drooping, 6- to 
3-flowered, those racemosely arranged at the tip of a 
long and otherwise sterile peduncle; peduncles to 
7 em; sterile portion to 6.5 em, glabrous; fertile portion 
to 1.3 em, sparsely to moderately strigose or tomentose; 
inflorescence bracts 2 or 4, foliaceous, much smaller 
than leaves, 4-15 X 1-6 mm, lw ratios 3—6:1, + 
membranous, glabrous on both surfaces, apex acute or 
acuminate, base eR or cordate, often persistent 
and nervation 


in fruit; bract stalk O mm; midri 


usually visible. fice 4-merous, yellow, often 
covered with yellowish blisters when dry, distinctly 
pedicellate; pedicels 3—4 mm, moderately to sparsely 
pubescent; hypanthium 7.5-9 mm, articulate, + 
membranous, sparsely to moderately strigose externally 
(surface still clearly visible), trichomes 0.2-0.5 mm, 
glabrous internally; persistent portion 3-3.2 mm; calyx 
4, spreading, broadly elliptic or orbicular, 2-3.9 

3 mm, glabrous adaxially, moderately to 
fee pubescent abaxially, smaller pair of lobes less 
pubescent, apex rounded; petaloid scales absent; 
stamens 8, upper whorl of anthers slightly below mouth 


or up to 1/4 exserted, lower whorl ca. 0.5-1 mm below 


upper whorl; anthers elliptic, 0.6-0.8 X 0.2-0.3 mm; 
subgynoecial disk absent or cupuliform, composed of 
several Po hs bcn when present, each 
segment to 0.2 m . glabrous, apex irregularly 

lobed, sinuses rm to deep; ovary ovoid to ellipsoid, 
0.7-1 X 0.4-0.6 mm, stipitate (stipe ca. 0.1 mm), 
glabrous or sparsely strigose on lower half, becoming 
1-1.5 mm; style 
2.5-3.6 mm, glabrous; stigma at height of lower whorl 


densely strigose at apex, trichomes 
of anthers or just above articulation. Fruits ovoid, 2.9— 
3 


mm, lower half glabrous, upper half 
glabrescent to sparsely strigose, apex densely pubes- 


cent; pericarp membranous. 


Distribution and habitat. Gnidia danguyana is 
distributed along most of Madag 

littoral forest from the Masoala saa to around 
Fort Dauphin from 0-15 m elevation (Fig. 16). The 


species grows on sandy substrates in open, often 


ascars east coast 


disturbed, sunlit areas in the humid bioclimatic zone. 


Phenology. The species flowers and fruits year 


Vernacular names. Avoha (Réserves Naturelles 
[Pierre] 8874, Plantes de Madagascar 5871); havoa 
5100); 


ervice Forestier havoa 


hafotra (Louvel 118, 


— 


e uud 
19 


IUCN Red List category. Gnidia danguyana has 
not yet been recorded within Madagascar's protected 
area network. The extent of occurrence (EOO) of the 
species is 18,200 km? and the AOO is 1600 kn? 
given a 10 X 10 km grid cell size. Nine subpopula- 
tions occur in small patches of highly fragmented and 
severely threatened littoral forest (Consiglio et al., 
2006). The species is assigned a prelimina 
(2001) conservation assessment of Vulnerable (VU) to 
extinction (Blab + 2ab). 


Discussion. Gnidia danguyana is easily recog- 
nized by its large (1-8.3 X 0.6—5 cm), ovate or ovate- 
elliptic leaf blades with cordate or bal cordate 
ases an its distinctive 6- 23-flowere 


Gnidia 


its 2- to 4-flowered E and smaller leaves 


oneness eglecta can E separated by 
(8-17 X 4-12 mm) with conspicuous densely con- 
gested fine venation. 


Typification. The original material for Gnidia 
danguyana was cited incorrectly in the protologue 
as “M. Humbert 118, 197” (Leandri, 1930a: 35). Both 
syntypes were actually collected by Louvel, rather 
than Humbert, a mistake that Leandri corrected years 
later in the Flore de Madagascar el des Comores 
Leandri, 1950). One sheet of each syntype is 
deposited at P. Both are annotated by Leandri, closely 


— 


match the description, and bear the same handwritten 


Volume 96, Number 2 
2009 


Rogers 34 
Revision of Malagasy Gnidia (Thymelaeaceae) 


locality on the label. Louvel 118 (P) is chosen as the 
lectotype because it is LT and in better physical 
condition than Louvel 197 (P). 


Selected specimens examined. MADAGASCAR. Antisir- 
anana: Masoala Nail. Park, near Cap Est, Schmidt et al. 4402 
(BR, G, K, MO, P, TAN, US, É 
MR E uar ahy, N of Nosy Varika, Ambolo ie 
Razakamalala et al. 1445 (MO, P, TEF); Mahabo forest, 

, BOL, K, T 
bw. Manampano, 
Rakotom 


Fianarantsoa: 


Toamasinat Ile Sainte Marie: o Sahasifotra, Ambo- 


Mandena ll Station, ea et al. 891 (BM, G, 

TAN, US); Mandromondro N of Fori t-Dauphin, Service 

F d (Capuron) 28646 e. TER: Sainte Luce, near Fort- 
Dauphin, Falinianina et al. 29 (BM, BR, MO, P, TEF). 


4. Gnidia daphnifolia L. f., Suppl. Pl. 225. 1782, as 
“daphnaefolia.” Gnidia pau L. f. var. 
glabra L. f., Suppl. Pl. 1782. Dessenia 
ee (L. D Raf., FL pM 4: 106. 1838, 

hnefolia." TYPE: Madagascar. Hb. Smith 

No. en (lectas te; jeden by Rogers in 

Rogers & Spencer, 2006: 486, LINN-SM!). 
Figure 


Dais madagascariensis Lam., Encycl. 2: 254. 1786. Syn. nov. 
Lasiosiphon madagascarensis (Lam.) Decne., Voy. Inde 
4: 148. Gnidia madagascariensis (Lam.) Gilg, 
Nat. Pflanzenfam. 3(6a): 228. 1894. TYPE: Madagas- 
car, s.d., P. Commerson s.n. (holotype, P-LA!; isotype, 
phy: 

Dais pubescens Lam., Encycl 2: 255. 1786. Syn. nov. 

Wu cud aH ty (Lam.) Decne., Voy. Inde 4: 148. 
1844. E: Madagascar, ommerson s.n. 
aoe a LA!; isotypes, G [2]!, MA!, P [3])). 
Gnidia rostrata Drake, Hist. Phys. Pe a 35(5) [Atlas 
pl. 315. 1896. Syn. nov. TYPE: *Madagascar" (type, 
En Drake in Grandidier, 1896). 

Lasiosiphon rostratus Meisn., Prodr. 14: 597. 1857. Syn. nov. 
Lasiosiphon madagascariensis (Lam.) Decne. var. 
rostratus (Meisn.) Leandri, Bull. Mus. Natl. Hist. Nat. 
(P. He sér. 2, 3: o 1931. TYPE: Madagascar. 
Antsiranana: Port Leven, Mar.—Apr. 1849, L.-H. Boivin 
2384 alone, CDC! inoype 1257 

Lasiosiphon baronii Baker, J. Linn. Soc., Bot. 25: 342. 1890. 
yn. nov. Lassen madagascariensis (Lam.) Decne. 
var. baronii (Baker) Leandri, Bull. Mus. Natl. Hist. 
Nat, sér. 2, 3: 151. 1931. TYPE: Madagascar. NW 
Madsgesca s.d., R. Baron 5770 (holotype, K!; isotype, 


Lasso hildebrandtii Scott-Elliot, J. Linn. Soc 
47. 


(Scou- Elliot) Cilg, Nat. Pflanzenfam. 3(6a): 228. 1894. 
Lasiosiphon a Case var. O ia 
Elliot) Leandri, Bull. Mus. Natl. Hist. Nat., 

151. 1931. TYPE: Madagascar. ec Mosis 


d'Ambre, May 1880, J. Hildebrandt c Cil 
designated here, BM!; isotypes, G [3]!, P!, US»). 
Lasiosiphon saxatilis Scott- Elliot, n Linn. E 
1891. Syn. nov. TYPE: Madagascar. Toliara Sainte 
Luce, rocky places near Fort- ais s.d., G. Seo 
Elliot 3030 ate K!; isotype, P!). 
Gnidia pubescens Baill., Hist. Phys. Madagascar 35(5) [Atlas 
3], pl. 314. 1895, nom. illes, non Gnidia pub 
rgius, Descr. Pl. Cap. 124. 
“Madagascar” a pl. 314!, Baillon in Grandidier, 
1895). 


= 


Lasiosiphon pubescens (Lam.) Decne. var. multifolius Leandri, 
Bull. Soc. Bot. France 76: 1042. 1929 [1930]. Syn. nov. 
Lasiosiphon multifolius (Leandri) Leandri, Notul. Syst. 
(Paris) 13: 51. 1947. TYPE: Madagascar. Toliara: 
Madagascar, T 1900, C. Alluaud 85 (lectotype, 
designated here, P. 

Lasiosiphon m (Lam.) Decne. var. parvifolius 
Leandri, Bull. Mus. Natl. Hist. Nat 436. 
1929 [1930]. Syn. nov. TYPE: dele Toliara: 
Ambovombe (Androy), 27 Apr. 1924, R. Decary 2605 
(holotype, P 00380375; isotype, P!). 


Decne. var. carinatus Leandri 


Leandri, Notul. Syst. (Paris) 13: 50. 1947. 

: Ambovombe, along ocean, cal- 
careous limestone and sand, 8 May 1924, R. Decary 
2694 (holotype, P!; isotypes, P!, TAN). 


ERE es eur m (Lam.) Decne. var. DES] 
li a 


eandri, Bull. Mus tl. Hist. Nat., sér. 2, 3: 151 
. Syn. nov ascar. 
Vohémar, 1840, J. Richard 580 (lectotype, dicun 


PS). 
mou waterlotii Leandri, Bull. Mus. Natl. Hist. Nat., 
, 3: 153. 1931 . nov. TYPE: Ma hdi caps 
pera Ambilobe, s.d., E. Waterlot 331 (holotype, 
Ph). 


se dumetorum Leandri, Notul. de ent 13: 52. 
7. Syn YPE dagascar. Toliara: Manam- 

X Valley, Mandrare basin, near the Isomono 
(confluence of the Sakamalio), Mt. Morahariva, 1000— 


400 m, Dec. H. Mr vida ries 
gnated here, P!; isotypes LGLK US?. 
pss madagascariensis ar 


13154 (lectotype, designated here, P 
003803444; isotypes, BM!, G!, K!, P [2], TAN!, US). 


Lasto. iphon uffrutescens Leandri, Notul. S 
194 


(holotype; P 00370336; isotypes, 
85. 


Mi, G [2]!, K!, MOI, P!, TANI, 


Shrubs or trees to 6 m tall; young branches densely 
to moderately pubescent; mature branches sometimes 
lenticellate. Leaves alternate, rarely subopposite, 
petiolate; petioles 1-2(-3) mm, densely to moderately 
pubescent; blades broadly obovate or obovate-elliptic, 


rarely narrowly elliptic, 7-61 X 4-21 mm, l:w ratios 


344 Annals of the 
Missouri Botanical Garden 
ca. 3-5(-7):1, often slightly inequilateral, both ^ southeast (Fig. 15). Several disjunct populations (e.g., 


surfaces usually strigose or tomentose initially before 
becoming glabrescent, sometimes densely pubescent 
but indument never dense enough to hide the adaxial 
surface), apex acute, rounded, or obtuse, tip usually 
apiculate, base attenuate, less often cuneate; midrib 


venation often discolorous, 
surfaces, more pronounced 
abaxially. Inflorescences terminal or axillary, some- 
times extra-axillary, erect, involucrate, (8- to)11- to 
35-flowered; peduncles 3-50 mm, densely to sparsely 
pubescent, rarely glabrescent; involucral bracts 5, 
generally broadly ovate (often 1 or 2 bracts within 
inflorescence with a broader base and shorter acumen 
or rostra), (83-)5-19 X (2-)5-6 mm, l:w ratios ca. 2— 
4:1, persistent, lower half of bracts appressed to 
flowers, glabrous to densely pubescent on bot 
surfaces, apex usually rostrate or long-acuminate, 
less often short-acuminate, acumen or rostra to 1.1 cm 
when present, upper half often reflexed when longer 
(but never apically decurved), base rounded-truncate; 
rs 5- 
merous, yellow, orange, or red; pedicels (0.2—)0.5— 


nervation conspicuous near margin. Flower 
1.5 mm, covered with 1.5-3 mm trichomes; hypan- 
thium 6.5-12(-15) mm, articulate, densely pubescent 
externally, usually glabrescent internally; caducous 
portion densely covered externally with ca. 0.5 mm 
trichomes; persistent portion (2.5—)3—4 mm, densely 
covered externally with (1.5—)2.5—4 mm trichomes that 
obscure the surface in fruit; calyx lobes 5, broadly 
1.2-3.5 1-2.3 mm, 
glabrous adaxially, densely pubescent abaxially, apex 


elliptic, obovate, or ovate, 
emarginate or rounded; petaloid scales 5, narrowly 
ovate-elliptie or obovate, sometimes approaching linear 
or orbicular, 0.8-1.4 X 0.2- 


glabrous, apex acute to rounded, emarginate, or with a 


1.8 mm, membranous, 


few rounded lobes, sometimes with a conspicuous dark 
midvein; stamens 10, upper whorl of anthers located 
just below mouth or to 3/4 exserted, lower whorl 1— 
1.5 mm below upper whorl; anthers elliptic, 0.6-1.1 X 
0.2-0.3 mm, subsessile; subgynoecial disk cupuliform, 
0.1—0.4 mm tall, 


sessile or shortly stipitate (stipe to 0.2 mm), mostly 
glabrescent, apex sometimes moderately pubescent; 
style to 12 mm, glabrous; stigma ca. 2 mm below or at 
LE of lower whorl of anthers. Fruits ellipsoid, 2.34 

.5 mm, usually glabrous, sometimes with a few 
sparse trichomes near apex. 


Distribution and habitat. Gnidia daphnifolia is 
widespread in Madagascar from 0—1400 m elevations, 
with most populations occurring either in the drie 


areas of the far north, or in the humid regions of the 


at Ankara Plateau, Tampoketsa d'Ankazobe, Tsiribi- 
hina Valley, Fanjahira) scattered along the central 
and occidental side of the high plateau link the 
The 


species occurs in degraded open savannas on sandy 


disjunct northern and southern populations. 


or lateritic soils, and is less frequently found on rocky 
slopes of granite, gneiss, limestone, and possibly 
marble. 


Phenology. The species flowers and fruits year 
ound 


Vernacular names. | Avoha (Decary 4030; Humbert 
20403); avoha madinika (Cloisel 135); havoa (Ran- 
driatafika 349); havoha (Réserves Naturelles 13004); 
mandrakieka (Service Forestier [Serrado] 1282); man- 
dreankaine (Bernier 157); mandriankiaka (Guittou et 
al. 4; Luckow 4226); tsifoladrivotra (Rogers 


Rakotonasolo 147). 
IUCN Red List category. Gnidia daphnifolia is 


widespread and common throughout Madagascar and 
as been recorded in at least several protected areas 
(e.g., Andohahela, 


Tsaratanana). e 
species is assigned a provisional IUCN (2001) 


Ankarana, 


conservation status of Least Concern 


Discussion. Gnidia daphnifolia is the most geo- 
graphically ess and morphologically variable 
species ni n Madagascar. The substantial 
morphological variation in this taxon is probably 

to the wide array of biophysical parameters pn 
elevation, UU bioclimate, habitat) affecting the 
popu. 
geographic eee 


that occur throughout a broad 
(Fig. 15, A). Plants 


within a single population, are capable of exhibiting a 


various 
, even 


large amount of nod A variation, poscis in 
those features (e.g., lea ract size and pu 

cence, peduncle length, bonn length) that were 
previously considered taxonomically important in 
various treatments of the group (Leandri, 1931a, b, 
1947, 1950 


based on the examination of relatively few specimens 


. Furthermore, most of Leandri's taxa were 
+ 


made from disjunct localities that did not show the 
morphological overlap and continuous variation in the 
tic. As a result, 13 of 
eandri’s names (five at the specific and eight at the 


characters he deemed diagnos 


varietal rank) are now placed into synonymy with G. 
daphnifolia. 

Gnidia daphnifolia can be distinguished from 
similar Malagasy species (e.g., gilbertae, G. 


hibbertioides, G. linearis, G. 


combination of features, including its relatively wide 


occidentalis) by a 


leaves (reaching 21 mm wide) that are generally 
broader above the midpoint, the involucrate inflores- 
cences usually borne on long peduncles (to 5 cm 


Volume 96, Number 2 
2009 


Rogers 34 
Revision of Malagasy Gnidia (Thymelaeaceae) 


long), the involucral bracts typically with a long 
rostrate or acuminate apex, the articulated hypanthi- 
um with five calyx lobes, and the petaloid scales 
measuring 0.8-1. .2-1.3 mm. 

Outside of Madagascar, Gnidia daphnifolia differs 
from the widespread G. glauca, a species distributed 
oughly from central Africa to southern India 
(Peterson, 2006), by its inflorescences with five (vs. 
six to the articulated (vs. 
unarticulated) hypanthium, and the leaves with fewer 


12) involueral braets, 


secondary and intersecondary veins. 


Nomenclature and typification. The validly pub- 
lished name, Gnidia daphnifolia (Linnaeus, 1782) 
has usually been treated in the literature as a synonym 
of either Lasiosiphon madagascariensis or L. pub- 
escens, combinations both based on Dais basionyms 
originally published by Lamarck (1786) t 

later transferred to Lasiosiphon by Decaisne (1 
Meisner (1857: 597) was probably the first author to 
formally treat G. daphnifolia as pro parte synonyms of 


hat were 


od and L. pubescens, respectively, 


s ac to have been the first 
r Lin n G. dap 
Leandri (1931b: o in his pu revision E the 
Malagasy Gnidia, incorrectly placed G. daphnifolia L. 
f. into synonymy with G. bojeriana. These two taxa are 
difficult to confuse and thus he might have actually 
been referring to “G. daphnaefolia,” a manuscript 
name of Bojer’s that appeared as a synonym in the 
protologue of L. bojerianus (Decaisne, 1844: 149). 
Strangely, Leandri did not mention G. daphnifolia L. f. 
in his second revision of the group (1947) nor in his 
Flore de Madagascar et des Comores treatment (1950). 
Whatever the case, Leandri certainly did not consult 
the original Linnaean material used to describe C. 
s that is still extant in the Smith Herbarium 
was recently designated as the lectotype 

of the name in Rogers and Spencer (2006). 
Two collections were cited in the protologue of 
Lasiosiphon hildebrandtii (Scott Elliot, 1891): Hilde- 
brandt 3369 (BM, G [3], K, 


P, US) from e northern 


K, P) fro 
ies Hildebrand coL 
1 A re 


and the BM sheet is chosen as the leete: 

Two collections, Alluaud 85 and Alluaud 106, were 
cited in the protologue of Lasiosiphon pubescens var. 
multifolius, both of w are deposited at P an 
match the original description (Leandri, 19303). 
Alluaud 85 (P) is selected as the lectotype because 
the original label bears Leandri's handwritten anno- 


sc 
extreme southern Madagasca 


lection more closely matches 


tation of the name. The other syntype is in equally 
good condition and instead bears Leandri's annotation 
on a typewritten Paris herbarium label. 


Two collections are cited in the protologue of 
Lasiosiphon | madagascariensi. angus see 
(Leandri, 1931b): Perrier de a, "Bathie 1276 a 
Richard 580. Later, Leandri (1947) went on to use 
Perrier de la Báthie 1276 as one of five syntypes for L. 
Gnidia occidentalis). In the Flore, 
Leandri (1950) did not recognize his variety angusti- 
folius and instead cited Richard 580 under the species 
is. Thus, Richard 580 more closely 


ndri’s concept 


occidentalis (= 


L. madagascariensis 
matched Lea of L. madagascariensis var. 
angustifolius, and the sheet deposited at P is chosen 
as the lectotype. 

ive collections of Lasiosiphon dumetorum were 
cited in the protologue (Leandri, 1947): Humbert 6742 
(G [2], P), 1281 2bis (P), 13010 (P), 13242 (BM, C, K, 

1 3800 (P). examined ns imens 
match ihe description, but Humbert 13242 is the most 
widely M M collection and the P sheet is 
selected as the lectotype. 

Three co e were cited in the protologue of 
wi 
an 1947) Humbert 13154 (BM, C, P [3], 
N, US), 13860 (P), 14053 (P). All three closely 
2 the description. Humbert 13154 is the most 
widely distributed collection and P 00380344 is 
selected as the lectotype. 


Lasiosiphon madagascariensis var. 


— 


Selected specimens examined. MADAGASCAR. Antan 
narivo: Angavo massif, near Ankazobe, Decary 7353 (P. 
oe rd., Km 184, Descoings 3283 (MO, 
TAN); Vohimbohitra massif, ti r Manakana, Cours (Hb. St. 
Agric. Alaotra) 1518 (MO, P, TAN. Antsiranana: jer be, 
SW of Ambilobe, Humbert & Capuron 25579 (P: Analabe 
forest, near village of Analabe & Lac Sa 
Vohémar, Razakamalala el al. 538 (BR, MO, P 
an part of 


ee Meyers 17 


m E o & Rasanga 598 (M 0, 
forest, Seigler 12881 (MO, TAN 


Orangea r 


P, TAN); a 
; Ivovona, Diego Suarez- 

, Rogers & Rakotonasolo 149 (G, MO, P, TAN); 

Joffreville, 2 km SE of Joffreville on rd. from Diego Suarez to 
Montagne d’Ambre Natl. Park, Rogers & Rakotonasolo 147 

O [3]; Montagne d'Ambre, Les Roussettes at Ankazobe, 
Homolle 167 (P); Montagne des Francais, 6-8 km N of Diego 
S & Rakotonasolo 133 (MO, TAN); Port Leven, 


Béthie 8552 (P [2]: Masokoamena, Bemarivo, yd de la 
nm. 8548 (P) Toliara: EU lkm E of town, 

et al. 914 (G, K, MO, P, TAN, US); Ambovombe, 
e 8391 (MO, P); Asael Natl. Park (Parcel #1), 


K, MO, TAN, TEF, WAG}; Andriambe, above Belavenoky 
RIP Rogers et al. 954 (B, G, K, MO, P, TAN, US); Ebakika, 
of Fort-Dauphin, Decary 10102 (MO, P, US); Emena 


Annals of the 
Missouri Botanical Garden 


village (Andohahela Natl. Park [Parcel #1]), Réserves 
Naturelles (Randriamiera) e (TEF); Fanjahira, plateau 
& valley of Isalo, Humbert 2755 (P); Fort-Dauphin, Decary 
4030 (P, TAN); Fort-Dauphin, rocky places in open country, 
I Elliot 2568 (BM, K, P); Kotriha, Mandrare basin, Mtns. 
otriha Isomonobe, 12bis (Py; Imonty, 
nara basin, betw. the & Elakelaka, 


éserves Nanirelles (Ramarokoto) oa (P [2] 
Minus , Manana asin, Humbert 13800 (P); Mahata- 
laky, 3-4 S of Mahatalaky village, Randrianasolo 571 


(BM, G, MO, P, TAN); Manambolo Valley, Mandrare basin, 
Humbert 6742 (G [2], P); Manantenina, N of Fort-Dauphin, 
Decary 3897 (P, TAN) Morahariva, Manambolo Valley, 
Mandrare basin, Humbert 13154 (BM, G [3], TAN, US), 
15242 (BM, G, K, P, TAN, ee 13860 (P); Nosibe, ca. 5 km 

E of Manambaro, Rogers ei al. 906 (G, K, MO, P, TAN); Pic 
Saint Louis, near Fort- Daas Rogers & Rakotonasolo 106 


106 (Py Tsiribihina, betw. Seda & Tsiribihina, Perrier 
de la Báthie 8551 (P). 


5. Gnidia decaryana Leandri, Bull. Mus. Natl. Hist. 
Nat., 2, 1: 436. 1929 [1930] TYPE: 
s ascar. Toliara: Fort-Dauphin, 3 Jul 

6, R. Decary 4332 lps P 003734261; 

E Pt, TAN). 


sér. 


Figure 5 


rubs to 2 m tall; young branches reddish when 
E flattened (more pronounced near internodes), 


Leaves opposite to decussate, pairs rarely alternate 
on vigorous shoots, appressed adaxially against stems, 
caducous on older branchlets; petioles 0.4—0.8 mm, 
glabrescent to moderately pubescent; blades obovate 
to suborbicular, (4.1-)8-15(-18) X (225-11 mm, l:w 
ratios ca. 1-2.5:1, both surfaces glabrous, lighter 
green or brown abasaliý, apex apiculate, obtuse or 
rounded, margin with a distinct vein that appears red 
on young leaves when fresh, base cuneate to 
attenuate; midrib slightly depressed adaxially, gla- 
rous or sparsely pubescent, raised and glabrescent 
abaxially, or wit arse strigose trichomes, 
lighter green than blade abaxially; secondary venation 
glabrous, raised or inconspicuous adaxially, usually 
conspicuous and more pronounced abaxially. Inflo- 
rescences terminal or axillary, erect, capitate, 2- to 
4-flowered, sessile or subsessile, subtended by 2 pairs 
of closely spaced decussate bracts, the upper pair 
smaller and caducous. Fl 4-merous, red-green, 
l 


0.4- 
0.8 mm, densely pubescent; hypanthium 6.1-9 mm, 


wers 
green-white, or yellow?, subsessile; pedicels 
articulate, = membranous, densely pubescent exter- 
nally, trichomes to 0.3 mm, trichomes of similar 
length on both portions or slightly longer on the 
persistent portion, glabrescent or sparsely pubescent 


in lower half near articulation internally; caducous 


broadly elliptic or orbicular, 1.4-1.9 X 1.4-1 
one opposing pair smaller, glabrescent or sparsely 
puberulent adaxially, tomentose on lower half abaxi- 
ally, otherwise glabrous, apex rounded or obtuse; 
petaloid scales absent; stamens 8, upper whorl of 
anthers ca. 1/4 exserted, lower whorl 0.2-0.5 mm 
below upper whorl; anthers elliptic, 0.6-0.8 X 0.2 
—0.3) mm, subsessile; subgynoecial disk cupuliform 


—~ 


or absent, to 0.1 mm tall, glabrous, apex irregularly 
lobed to nearly entire; ovary ovoid to ellipsoid, 0.9— 
l.l X 0.4-0.5 mm, sessile or very shortly stipitate 
(stipe to 0.1 mm), lower half densely to sparsely 
pubescent, rarely glabrescent, upper half densely to 
sparsely pubescent, apex always densely pubescent 


with 1-1.5(-2) style 2.2-6 mm, 
sparsely to moderately strigose, rarely glabrescent; 


mm trichomes; 
stigma ca. 1.5 mm below lower whorl of anthers. 
Fruits ovoid-ellipsoid or ellipsoid, 3.1-3.4 X 1.3- 
1.5 mm, sparsely to moderately pubescent, apex 
usually densely pubescent. 

Distribution and habitat. 


in southeastern Madagascar near Fort Dauphin (Toliara 
Province) from sea level to 950 m elevation (Fig. hi 


Gnidia decaryana occurs 


species grows in open sunlit areas and has been recorded 
on sand and gneiss in the humid bioclimatic zone. 


Phenology. The species has been found in flower 


and fruit in January, March, July, October, and 


December. 


IUCN Red List category. Gnidia decaryana occurs 
within the bound (A 
hela). In 2003, a small, unprotected population of G. 


aries of one protected area (Andoha- 


decaryana was located growing on the outskirts of Fort 

auphin along the base of the Pic St. Louis, bordering 
agricultural fields and no more than 25 m away from 
the edge of the town (Rogers, pers. obs.). The species 
is assigned a provisional IUCN (2001) conservation 
status of Near Threatened 


Discussion. Gnidia decaryana is distinguished 
from G. e the most morphologically similar 
Ma lagasy spec y 

blades with siete to attenuate bases (vs. broa 
ovate blades with cordate bases), a different venation 


obovate to suborbicular 


pattern, its sessile or subsessile (vs. pedunculate) 
inflorescences, and by the dense (vs. completely 
absent) pubescence on the outer surface of the 
hypanthium. In addition, populations of G. decaryana 
occur more than 600 km to the south of G. neglecta. 

Gnidia subcordata Meisn., an African species in the 
closely related segregate genus Englerodaphne (as E. 


subcordata (Meisn.) Gilg), differs from G. decaryana by 


Volume 96, Number 2 
2009 


Rogers 34 
Revision of Malagasy Gnidia (Thymelaeaceae) 


its pedunculate (vs. sessile or subsessile) inflorescences, 
longer flowers (11-15 mm vs. 6.1-9 mm long), and 
sparsely pubescent (vs. de 
shaped (vs. cylindrical) hypanthium 


The inflorescence structure " Gnidia decaryana 


nsely pubescent), + funnel- 
thi 


(Fig. 5B) is unique within the Malagasy species, with 
its two pairs of decussately arranged foliose bracts, 
located directly below two to four sessile or subsessile 
flowers. The bracts closest to the flowers are usually 
about half ue size of the lower pair and fall off earlier. 


l li tion between leaf and bract, if one can be 


med is definitely blurred in this species. One small 
and lacks 


obvious nervation and will rarely persist until anthesis. 


membranous bracteole subtends each flower 


Typification. Two sheets of Decary 4332 are 
deposited at P. Both sheets closely tch the 
” 1930: 436) 
and were annotated by Leandri. The sheet bearing the 
accession number P 00373426 has the locality and date 
cited in the protologue and is regarded as the holotype. 


description in the protologue (Leandri 


Selected specimens examined. Elend ue Toliara: 
Andohahela, Mananara bas Andohahela & 
Elakelaka, Mahamavo, Hambert 13804 P P [2], e 
Mananara basin, mtns. betw. Andohahela 
Elakelaka, S of Imonty, Humbert 14084 (P); betw. e 
motra & Lokaro, N of Fort-Dauphin, Service Forestier 
ausi 28650 (MO, P. TEF); Pic Saint Louis, Rogers & 
Rakotonasolo 108 (BM, G, MO, P, TAN, WAG). 


6. Gnidia gilbertae Drake, Bull. Mens. Soc. Linn. 
Paris 2: 1218. 1896. TYPE: Madagascar. Maha- 
janga: betw. “Madounga et Antsalahanki,” 1876, 
A. Grandidier s.n. (holotype, P!). Figure 6. 


hrubs or trees to 4 m tall; young branches 
pubescent; mature branches not lenticellate. Leaves 
alternate, rarely subopposite, leaves crowded near 
branch tips, rarely persistent on older branchlets; 
petioles 0-2(-3) mm, densely pubescent or glabres- 
cent; blades broadly obovate to nearly elliptic, 1.4— 
5.5 X 0.4—1.7 em, lw ratios 2.3—3.6:1, both surfaces 
moderately to sparsely pubescent, rarely densely 
pubescent, apex rounded or emarginate, tip usually 
apiculate, base attenuate or less often cuneate; midrib 
ssed or rarely plane adaxially, raised and 
densely pubescent abaxially; venation often discolor- 
ous, usually raised on both surfaces, more pronounced 
abaxially. Inflorescences terminal or rarely axillary, 
erect, involucrate, 16- to 28-flowered; peduncles 0.5— 
5(-9) mm, densely pubescent; involucral bracts 4(5), 
broadly ovate or orbicular, 7-12 X 4-7 mm, l:w ratios 
1.1-2.6:1, 


glabrescent adaxially, rarely sparsely or densely 


chartaceous, erect, persistent, usually 


strigose in upper half abaxially, apex short-acuminate 


or acute, acumen to 3 mm, spreading (i.e., not 


reflexed), base rounded-truncate; midrib inconspicu- 
ous or visible o 


oth surfaces, more pronounced 
abaxially; both 


nervation often conspicuous 

surfaces and represented by 2 or 4 veins, those more 
obvious near the margin. Flowers 4-merous, yellow or 
orange; pedicels 0.5-1(-1.5) mm, densely covered 
0.5-1 mm trichomes; hypanthium (10-)12.5— 


17 mm, unarticulated, coriaceous, densely pubescent 


with 
externally, trichomes (0.5-)1-1.5 mm, usually 
brescent internally; caducous portion torn irregularly 
in the lower 1/3 by the developing fruit; calyx lobes 4, 
broadly elliptic-oblong or obovate, 2.1-2.7  1.2- 
mm, glabrescent or sparsely puberulous adaxially, 
densely pubescent abaxially, apex emarginate or 
rounded; petaloid scales 4, narrowly ovate-elliptic, 
0.9-1.2 X 0.2-0.3(-0.5 


brous, apex acute or with 1 to 


mm, membranous, gla- 
several irregular 
rounded lobes, often with a conspicuous midnerve 


m below upper 
0.8-0.9 X ca. 
0.2 mm, subsessile; subgynoecial disk cupuliform, 

0.2-0.4 mm tall, glabrous, fleshy, apex smooth or 
slightly s ovary ellipseid, 1.2-1.6 X 0.4 
0.7 mm, stipitate (stipe ca. 0.2 mm), glabrous; style 
3.4—5.7 mm, glabro 
whorl of anthers. 


elliptic-oblong, 


rous; stigma 0—7 mm below lower 
Fruits aioi, 2.9-3.1 X 1.1- 
1.2 mm, glabrous or rarely with a few sparsely 
spaced trichomes at the apex. 


Distribution and habitat. Gnidia gilbertae is 
endemic to semi-deciduous gallery forests in north- 
western Madagascar (Mahajanga Province) from ca. 
10 m ae (Fig. 16). The species occurs in 
open sunlit areas on sandy substrates in Madagascar’s 


dry bioclimatic zone. 


Phenology. The species flowers and fruits from 
April through November. 


Vernacular names. Famoty (Réserves Naturelles 
[Harizo] 1022); sisitry (Réserves Naturelles b i ad 
4234); tzomangamena (Randrianaivo et al. 


IUCN Red d category. Gnidia gilbertae has 
been recorde 
(Ankarafantsika, Namoroka). The s 
been collected, both frequently A recently, at 
les is as- 


Ampijoroa and Ankarafantsika. The spec 


signed 01) conservation status 


a provisional IU 
of Least Concern (LC) 


ion. Gnidia gilbertae is P. Mm from 
5-)merous flowers, the 


G. puis by the 4- (vs. 
hypanthium that lacks an mcus and is instead 
torn rei in the lower 1/3 by the developing 
fruit, the leaves that usually only remain persistent 
near ide tips of the branches (vs. leaves persistent 


Annals of the 
Missouri Botanical Garden 


along most of the stem), the four or rarely five (vs. five) 
involucral braets, and the generally shorter peduncles 


0.5—5(-9) mm vs. 3-50 mm long. 


Selected specimens examined. MADAGASCAR. Maha- 
janga: Ampijoroa Forest Station, Phillipson 1936 (K, MO, P, 
TAN, WAG); Anjiafitatra [Tsitontroina], Réserves Naturelles 
(Randrianasolo) 2180 (P, TA 
ambe, Ranaiv 


TY Bisset M3 (K) 
EF); Port Bergé, 
2 


Ra d eat. 


orest of Marosely. 


EF, US); Tsitampiky [Sitampiky], Decary 8181 
7. Gnidia Te (Baker) Domke, Biblioth. Bot. 
2 1): 1l : 


bakeri Gilg, Nat. Pflanzenfam. 3(6a): 227. 1894, 
nom. superfl Arthrosolen gnidioides (Baker) 
Leandri, Bull. Soc. Bot. France 76: 1043. 1929 
Madagascar. Antananarivo: 
grassy hills of the province of Imerina," s.d., 
R. Baron 2061 (lectotype, designated here, K!; 
isotype, P!). Figure 7 


Shrubs to 80 cm tall, branching subequally (di-) 
trichotomously, sometimes one shoot in trichotomy 
does not elongate; branches not lenticellate, orange- 
red when fresh, usually dark red after drying, densely 
to sparsely strigose or strigose-tomentose, indument 
denser near branch tips. Leaves alternate, rarely 
subopposite, usually persistent on older branchlets, 
erect, drying subappressed or appressed to stems 
adaxially, subsessile; petioles to 0.8 mm, densely to 
sparsely strigose; blades needle-shaped, rarely very 
narrowly obovate or ovate, 7-19 X 0.75-2(-3) mm, 
l:w ratios (4.5—) coriaceous, both 
surfaces sparsely strigose to glabrescent (indument 
more obvious near base and along midrib), apex acute 
or short-acuminate, tip rounded or rarely apiculate, 
base 
slightly raised adaxially, plane or slightly raised 


:1, involute, + 


cuneate or attenuate; midrib inconspicuous or 


abaxially; venation inconspicuous adaxially, abaxially 
represented by 2 or 4 longitudinal plicate veins when 

ry. Inflorescences terminal, erect, subsessile, 
globose, resembling composite-like a (composed 
pa b clusters of bracteate 
flowers), ca. 100-flow "i to 2.8 em diam.; 
peduncles 0.5-3 mm, ier to moderately pubes- 


cent; bracts surrounding the base of the head 5 to 7, 


of many 


imbricate, spreading, persistent, broadly ovate or less 
often narrowly elliptic-ovate, 5-10. .5-4.6 mm 
s dimensions within the inflorescence), 


coriaceous or membranous, densely pide to 


glabrescent, apex acuminate to aristate, acumen or 
arista to 4 mm, margin densely ciliate, base cuneate 
or attenuate-truncate; midrib and nervation conspic- 
uous; bracts within the head 50 or more, subtending 
floral clusters, imbricate, persistent, broadly elliptic- 

3.2-5.2 0.5-2 mm 
the inflorescence), char- 


obovate to linear-elliptic, 
various dimensions within 


— 


taceous-membranous, scabrous, densely to sparsely 
pubescent adaxially, glabrous to densely pubescent 
abaxially, apex acute or slightly acuminate, margin 
densely ciliate, base cuneate or attenuate; midrib and 
nervation conspicuous. Flowers 
white?, or yellow?; pedicels 0.4-0.7 mm, covered by a 
dense brush of 3—4 mm trichomes; hypanthium 6.5— 
O mm, articulate; caducous portion = membranou 
ith erect indument externally, 


4-merous, pink, red, 


densely covere 
trichomes 1-1.5 mm, glabrous internally; persistent 
portion ca. 2 mm, membranous, glabrous externally 
and internally; calyx lobes 4, erect, white-pink, red, or 
yellow, broadly oblong 2 X 0.5- 
1.3 mm, glabrous adaxially, densely to moderately 


or elliptic, 
pubescent abaxially, apex rounded, truncate, or 
emarginate; petaloid scales absent; stamens 8, upper 
whorl of anthers just below mouth, lower whorl 0.1— 
0.5 mm below upper whorl; anthers elliptic, 0.4-0.6 
X 0.15-0.2 mm, subsessile; subgynoecial disk cupu- 
liform or absent, less than 0.1 mm ta n present, 
glabrous, fleshy, apex smooth or irregularly lobed; 
ovary ellipsoid, 0.7—1.1 —0.5 mm, sessile, 
style 3—4.7 mm, glabrous; 
elow the lower whorl of anthers. Fruits 
ellipsoid, 1.6-1.9 X mm, glabrous 9 
(Rabakonandrianina & Carr, 1987). 


glabrous; stigma 0- 


Distribution and habitat. 
curs on the central plateau of Madagascar from 1 
2500 m elevation (Fig. 17). Populations occur from 
the Tampoketsa of Ankazobe (a region located to the 
Andringitra massif. 


Gnidia gnidioides oc- 
00! 


north of Antananarivo) to the 
Gnidia gnidioides is associated with marshy areas and 
other moist places, and grows in open sunlit areas, 
including prairies and Tapia woodlands, in the 
subhumid bioclimatie zone. Some specimen labels 
mention that the species occurs on gneiss or quartzitic 


rocks. 


Phenology. The species flowers and fruits year 


acular name. Bambola (Réserves Naturelles 
[Razafindrakoto| 3055, Réserves Naturelles [Rabeva- 
zaha] 10387; Service Forestier [Marlange] 1950). 


IUCN Red List category. Gnidia gnidioides is 
common in both protected (Ambohitantely, Andringi- 
tra, Analamazoatra) and unprotected areas, at least 
some of which are periodically disturbed by fire 


Volume 96, Number 2 
2009 


Rogers 34 
Revision of Malagasy Gnidia (Thymelaeaceae) 


(Rogers, pers. obs). The species is assigned a 
provisional IUCN (2001) status of Least Concern (LC). 


Discussion. Gnidia gnidioides is easily distin- 


guished by its needle-shaped leaves and 50- to 100- 


flowered, terminal, subsessile inflorescences that 
To resemble composite-like heads. The 
s s is mos orphologically similar to two 


M RAS ee species, G. bambutana (Came- 
roon, Nigeria) and G. mollis (Congo-Kinshasa, Malawi, 
Mozambique, Tanzania, Zambia). Based on a study of 
the inflorescence architecture of all three species, 
Aymonin (1966c) found that G. gnidioides differs from 
G. bambutana by its fewer-flowered floral clusters 
within the head and its more coriaceous and thicker 
leaves. Examined specimens of G. mollis are similar to 
G. gnidioides, but tend to have broader leaves, larger 


inflorescences, ers with. petaloid scales. 


noted that the three species are closely related, but 
considered each one to belong to a distinct species. It 
should be noted that the basionym of G. gnidioides 
(Dais gnidioides Baker, 1883) would Fou priority 
over G. mollis (Wright, 1906) and bambutana 
(Engler, 1921) if future studies B that these 
three taxa are conspecific. 


Nomenclature and typification. The superfluous 
name Gnidia bakeri Gilg has been used incorrectly to 
refer to G. gnidioides as recently as the Flore de 
Madagascar et des Comores (Leandri, 1950). Baker 
(1883) originally described the species as Dais 
gnidioides, but Gilg (1894) chose not to retain the 
pi 
Leandri (1930a) transferred the species using the 
correct epithet (gnidioides) into Arthrosolen, a genus 
now widely considered to be synonymous with Gnidia. 
Later Domke (1934) created a combination for the 
species in Gnidia also using Baker's original epithet. 
In the Flore, Leandri (1950) placed both A. gnidioides 
and D. gnidioides into synonymy under the illegiti- 
mate name G. bakeri Gilg. 

HN collections (Baron 665, Baron 2061) were 


the lectotype to avoid any potential ambiguity 
caused by the fact that the only material of Baron 
665 at K was mounted on the same sheet as Baron 
1894, and consequently might represent a mixed 
gathering. 


Selected specimens examined. MADAGASCAR. Antana- 
6 


MO, 
8541 (P); betw. Ambatolampy & Tsinjoarivo, Viguier 1800 (P 


[2]; Central Madagascar, Baron 665 (K); Km 30 on rd. betw. 
Antananarivo & Ambatolampy, Keraudren & Aymonin 25160 
(P); Mandraka forests, D’Alleizetie 1110 (P). Fianarantsoa: 
Andringitra massif, Rés Naturelles (Razafindra, 
271 (GH, K, MO, P y. Ankarabe, Rauh 79 (TAN) 
Fenoarivo, Bosser 413 (TAN) Ibity massif, Fosberg 52380 
(US); Itremo massif, van der Werff & McPherson 13577 (G [2], 

, TAN, WAG). Toamasina: Antsahapandrano (Ankar- 
alia), Decary 1 7646 (K, P). 


8 


Gnidia hibbertioides (S. Moore) Z. S. Rogers, 
comb. nov. Basionym: Lasiosiphon hibbertioides 
S. Moore, J. Bot. 58: 189. 1920. TYPE: 
Madagascar; s.d., J. Thompson & J. Forbes s.n. 
(holotype, BM!). Figure 8. 

Shrub?; 


Leaves alternate, closely arranged, caducous on older 


young branches densely pubescent. 
branchlets; petioles 1-2 mm, densely pubescent; 
1.5- 
3—4 mm, l:w ratios ca. 4—6:1, involute, both 


blades narrowly elliptic or slightly obovate, 
2.2 cm X 
surfaces covered with a dense uniform indument of ca. 

mm, erect to subappressed, trichomes, apex 
apiculate, base long-attenuate; midrib inconspicuous 
adaxially, raised abaxially, densely to moderately 
pubescent on both surfaces; venation + inconspicu- 
ous. Inflorescences terminal, erect, involucrate, ca. 


7-flow 


involueral ru acts 5, Sa m elliptic 


; peduncles to ely Dues 
X 3-4 mm, l:w ratios ca. 3:1, similar in texture to 
leaves, erect, glabrous adaxially, densely pubescent 
abaxially, apex acute or apiculate, base obtuse- 
rounded; midrib and nervation inconspicuous on both 
surfaces. Flowers 5-merous, sessile or subsessile; 
pedicels to 0.5 mm, densely covered with ca. 2 mm 
trichomes; hypanthium 1.7—1.9 em, articulate; cadu- 
cous portion densely pubescent externally, trichomes 
ca. 0.5-1 mm, glabrous or with a few appressed 
trichomes near articulation internally; persistent 


1.5-2 mm 


trichomes, those erect and nearly perpendicular to the 


portion 5—7 mm, externally covered by 


surface of the tube, glabrous internally; calyx lobes 5, 
X 1.5-1.6 mm, 
glabrous adaxially, densely pubescent abaxially, apex 


narrowly ovate-subtriangular, 4—5.2 


acute; petaloid scales 5, suborbicular or very broadly 
ovate, 4.6-6 X 3.8-4.7 mm, membranous, glabrous, 
upper half with an irregularly lacerate or sinuate 
margin; stamens 10, upper whorl of anthers located 
elow 
1.2-1.4 X ca. 0.25 mm, 


subsessile; subgynoecial disk cupuliform, ca. 0.5 mm 


just below th, lower whorl ca. 3.5 mm 


m 
upper whorl; anthers oblong, 


tall, glabrous, apex smooth or irregularly lobed; ovary 
ellipsoid, ca. 1.2 X 0.5 mm, stipitate (stipe ca. 

.4 mm), mostly glabrous, apex with a few trichomes; 
style and stigma not seen in good condition. Fruits 
not seen. 


1mm 


WX SS 


BM). 


( 


Annals of the 


350 


Missouri Botanical Garden 


Figure 8. Gnidia hibbertioides (S. Moore) Z. S. Rogers. —A. Habit. Note the large petaloid scales. —B. Persistent portion 


of hypanthium surrounding the gynoecium. Drawn from holotype, Thompson & Forbes s.n. 


Volume 96, Number 2 
2009 


Rogers 35 
Revision of Malagasy Gnidia (Thymelaeaceae) 


Distribution and habitat. Gnidia hibbertioides is 
endemic to Madagascar and known only from a poorly 
labeled type. More specific distribution or habitat 
information is not available. 


nology. The type FER is in flower, but 
no collection date is recorded on the specimen label 


or in the protologue (Moore, i0 
IUCN Red List category. Gnidia hibbertioides was 


collected in Madagascar on one occasion in the early 
1800s and is represented by a single herbarium 
specimen with an unspecified locality. The species is 
assigned a provisional IUCN (2001) conservation 
status of Critically Endangered (CR) (Blab + 2ab). At 
the present time, the species cannot be assigned to an 
Extinet category (EX or EW) because exhaustive 
surveys searching for additional individuals have not 
been conducted. 


Discussion. Gnidia hibbertioides is distinguished 
from G. daphnifolia by its 1.7—1.9 cm long hypanthi- 
um (vs. rarely to 1.5 em long) with a 5—7 mm long 
persistent portion (vs. to 4 mm long) and by its large 
(4.6-6 X 3.8—4.7 mm) petaloid scales, which surpass 
the calyx lobes. The petaloid scales of G. hibbertioides 
(Fig. 8A) are generally three to four times larger than 
the scales of most other species of Malagasy Gnidia, 

and may prove to be the largest in the genus. Another 
diiit feature of the scales is the irregularly 
lacerate or sinuate apical margin. Other Malagasy 
species have petaloid scales with entire margins that 
are rounded, emarginate, or lobed at the apex. 


Gnidia hibbertioides was based on a 
scantly labeled sheet deposited at BM. “Madagascar” 


Typification. 


is the only locality information mentioned on the 
original material and cited in the protologue (Moore, 
1920: 189). 


herbarium sheet, the names of the collectors were 


In unidentified handwriting on the 


written as "Vaughn, Thompson orbes. is 
information is possibly a distortion of the names of 
two men known to have collected in Madagascar, John 
Vaughn Thompson and John Forbes (Dorr, 1997: 485). 
However, it is impossible for men to have 
collected the original material jointly, because 
Thompson visited the island in 1814 and Forbes did 
the only extant materia rticular 
sheet must be considered the holotype. 


9. Gnidia humbertii (Leandri) Z. S. Rogers, comb. 
ov. Bas : Lasiosiphon humbertii Leandri, 
Bull. Soc. Bot. France 76: 1039. 1929 [1930]. 
TYPE: Madagascar. Fianarantsoa: Isalo, mouth 
of Sakamarekely 
1000 m, 19 Oct. 


ambalinieto rivers, 500— 


1924, H. Humbert 2844 


(lectotype, designated here, P!; isotype, GN. 


Figure 9 
Rounded, compact, densely ramified subshrubs to 
60 em tall; 


covered with prominent leaf scars; mature branches 


young branches densely pubescent, 


densely pubescent, — exfoliating. Leaves alternate, 
spirally arranged, closely spaced, rarely with inter- 
nodes to 1.5 mm, persistent only at the tips of 
2 young leaves adaxially appressed against 
s; petioles 0—0. , densely pubescent; E 
elite slightly oboväte, or linear, 7-17.1 
3 mm, lw ratios ca. 4-8:1, silver-green, both e 
obscured by a dense sericeous-tomentose indument of 
0.2-0.3(-0.5) mm trichomes, apex acute or short- 
acuminate, base cuneate or attenuate; midrib ob- 
scured by indument adaxially, obscured or nearly so 
surfaces. 


abaxially; venation obscured on bot 


Inflorescences terminal, erect, involucrate, (6- 
to)8- to 15-flowered, sessile or subsessile; peduncles 
to 1.5 mm, densely pubescent; involucral bracts 5, 


broadly ovate, 6-9 X 


erect, persistent, densely pubescent adaxially, densely 


2-4 mm, l:w ratios ca. 2:1, 


pubescent on upper half abaxially, lower half sparsely 
pubescent or glabrescent, apex obtuse or rounded, 
base rounded-truncate; midrib inconspicuous adaxi- 
ally, inconspicuous or slightly raised abaxially; 
nervation inconspicuous on both surfaces. Flowers 
5-merous, yellow, sessile or subsessile; pedicels to 
5(-l) mm 
mm, ticulale, coria- 
ceous; caducous portion eer pubescent externally, 
trichomes 0.5-1 mm, glabrous internally; persistent 
portion 2-3 mm, obscured by 3—4.5 mm trichomes, 
glabrous internally; calyx lobes 5, spreading, broadly 
obovate or elliptic, 24.5 X 1.7-2.5 
adaxially, densely pubescent abaxially, apex emar- 


mm, glabrous 


ginate or rounded; petaloid scales absent or 5; scales 
when present narrowly to broadly ovate or elliptic- 
linear, to 1.2 X 0.4 mm, membranous, glabrous, apex 
rounded; stamens 10, upper whorl of anthers ca. 1/2 


subgynoecial disk cupuliform, 0.3-0.5 mm tall, gla- 
brous, membranous, apex smooth or irregularly lobed; 
ovary ovoid-ellipsoid, ca. 1.4 X 
(stipe 0.2-0.3 


0.5 mm, stipitate 
mm), lower half glabrous or sparsely 
pubescent, apex usually densely pubescent, trichomes 
to 1 mm; style 3.8-7 mm, glabrous or with a few 
trichomes on lower half; stigma near mouth or at 
height of lower whorl of anthers. Fruits ellipsoid, 2.8— 
3.2 X 1-1.2 mm, mostly glabrous, glabrescent or only 
densely pubescent near the apex. 

Distribution and habitat. Gnidia humbertii is 
restricted to the subarid bioclimatic zone in southern 


Annals of the 
Missouri Botanical Garden 


Gnidia humbertii (Leandri) Z. S. Rogers. —A. Habit. 


and southwestern Madagascar (Fig. 17). Populations 
occur from the Mahafaly Plateau and Isalo to Cap 
Sainte Marie and as far west as La Table (a mesa-like 
mountain on the west coast near Toliara). The species 
grows in open sunlit areas on sand, sandstone, and 
calcareous limestone (Rogers, pers. obs.). 


Phenology. The species flowers and fruits year 


ETT E EE A 


ee 


Flower. —D. Fruit. Habit and fruit drawn from 
Pu & Peltier 2469 (TAN). Flower drawn from Rogers & Rakotonasolo 399 (MO). 


IUCN Red List category. Gnidia humbertii has been 
recorded in two formally protected areas (Cap Sainte 
Marie, Isalo). Several large populations, obviously 
resistant to periodic burning, were observed at Isalo in 
2004 and again in 2006. The species is assigned a 


preliminary IUCN (2001) status of Least Concern (LC). 
Discussion. Gnidia humbertii is the only Malagasy 
species with a rounded, compact, densely ramified, 


Volume 96, Number 2 
2009 


Rogers 35 
Revision of Malagasy Gnidia (Thymelaeaceae) 


subshrub habit. The leaves of G. humbertii and G. 
ambondrombensis are covered on both surfaces by a 
dense sericeous indument, but the pubescence of the 
former is composed of much shorter trichomes 


(trichomes 0.2-0.5 mm vs. 1-1.5 mm long). 


Typification. Two collections were cited in the 
protologue of Lasiosiphon humbertii Do 19302): 
Douillot [Douliot| s.n. and Humbert 2844. 
tions are annotated by Leandri and match the protologue 
description. The P sheet of Humbert 2844 is in better 
physical condition and is thus selected as the lectotype. 


Both collec- 


Selected specimens examined. MADAGASCAR. Fianar- 
antsoa: Isalo Natl. Park, 2-3 km N of Nail. Rte. #7, Rogers & 
e 399 (BM, BR, G, K, MO [2], P, TAN, TEF, US). 
Toliara: La Table, near Tuléar, 15 Feb. ba Afzelius s.n. (P); 
Mahafaly Plateau, Men aran la Báthie 8555 
(K, P; Mus, 1892, Douillo iDoulior s.n. (P). 


10. Gnidia linearis (Leandri) Z. S. Rogers, comb. nov. 
Basionym: Lasiosiphon linearis Leandri, Bull. Soc. 
Bot. France 76: 1040. 1929 [1930]. Lasiosiphon 
decaryi Leandri var. linearis (Le one eon 
Bull. Mus. Natl. Hist. Nat., sér. 2, 3: 154. 1931. 
TYPE: Madagascar. Toliara/Fianarantsoa: savanna 
betw. Bemketa [Bereketa] & Malio, 15 June 1923, 
H. Poisson 692 (holotype, P!). Figure 10. 


Lasiosiphon decaryi Leandri, Bull. Soc. Bot. France 76: 1041. 
[1930]. Syn. nov. TYPE: Madagascar. Toliara: 
mbe (Amboasary), dunes, 21 Ma , R. 
Decary 2785 (lectotype, designated here, P!; isotypes. 
!, TANS). 
Yasiésphon decaryi Leandri var. erectus Leandri, Bull. Soc. 
France 76: 1041. 1929 [1930]. : 
osiphon erectus (Leandri) E 
o is) 13: 50. 1947. TYPE: Madag 
ivo: Tananarive, May 1916, E. o n. (lectotype, 
designated here, P 00373449! 
Lasiosiphon decaryi Leandri var. littoralis Leandri, Notul. 
Syst. ewe a. 49. 1947. Syn. nov. TYPE: Madagas- 
car. Toliara: betw. Tuléar & Manombo, May 1910, 
Perrier de la Bå thie 8553 (lectotype, dad here, P 


003734531). 
ME x decaryi uet var. tenerifolia Leandri, Notul. 
Syst. (Paris) 13: 48. 1947. Syn. nov. TYPE: Madagas- 


car. Toliara: i de distr., Antanimora, 6 Au 
1924, R. Decary 2971 (lectotype, a here, P 
00373439!; isotypes, BM!, K!, TAN!). 


Shrubs to 3 m tall; young branches glabrescent to 
pubescent; mature branches usually lacking lenticels. 
Leaves alternate, subsessile or short petiolate; 
petioles to 2 mm, glabrescent to pubescent; blades 
linear, narrowly elliptic-oblong or obovate, 6-50 X 
1.5-9 mm, lw ratios ca. 3-14:1, both surfaces 
glabrescent to densely tomentose-strigose (indument 
gradually becoming denser moving from SE to SW 
populations), apex cuspidate or apiculate, base long- 
attenuate or cuneate; midrib inconspicuous or raised 


on both surfaces, glabrescent; venation visible in 
larger leaves, otherwise usually inconspicuous. Inflo- 
escences axillary or terminal, erect, involucrate, (9- 
to)12- to 
50 mm, glabrescent to sparsely pubescent, 


18-flowered, pedunculate; peduncles 5- 
rarely 
moderately pubescent; involucral bracts 5, suborbic- 
ular or broadly ovate, 5-8 X 3.8—5.1 mm, l:w ratios 
1-1.5(-2):1, often caducous, chartaceous or coria- 
ceous, both surfaces glabrescent to moderately pubes- 
cent (southwestern coastal populations), apex apiculate 


~ 


apicule 1-2 mm) with a strongly decurved tip, short- 
acuminate, or acute, base rounded-truncate; midrib 
nearly inconspicuous on both surfaces, more pronounced 
near apex abaxially; nervation absent or represented by 


1 to 5 longitudinal veins. Flowers 5-merous, yellow o 

rarely red; pedicels ca. 0.5 mm, densely 
3—4 mm; hypanthium 7.1— 
15 mm, articulate, semimembranous to + 
rnally, gla 
rnally covered wit 


orange, 
pubescent, trichomes 
coriaceous, 
ent exte 


densely pu brous a ees 
th ca. 0.5 mi 


caducous portion exte 
trichomes; persistent portion 3—4.5 mm, eu 
covered with (1-)3-5 mm trichomes; calyx lobes 5, 
broadly oblong-elliptic, 222.5 X 0.9-1.5 mm, membra- 


nous, spreading, eo adaxially, densely pubescent 


abaxially, apex petaloid scales 5, 
yellow, suborbicular or obovate, 0.9-3. 7 X 0.5-1.5 mm, 
membranous, glabrous, spreading, apex most often 
emarginate, sometimes rounded, truncate, or with 3 to 
5 irregularly rounded lobes; stamens 10, upper whorl of 
whorl 0.5-1 mm 
0.8-1 X ca. 
subgynoecial disk cupuliform, 


anthers ca. to 3/4 exserted, lower 
below upper whorl; anthers elliptic, 
0.3 mm, subsessile; 
0.1-0.3 mm tall, glabrous, fleshy, apex smooth or 
X 0.6- 
.7 mm, stipitate (stipe to 0.3 mm), cd style 
5.2-9 m whorl 
of anthers, or rarely exserted by 1 mm. Fruits ellipscid, 
3.1-3.4 X 1.1-1.4 mm, glabrous. 


irregularly lobed; ovary ellipsoid, 1.2-1.4 


m, glabrous; stigma 0-2 mm below lowe 


Distribution and habitat. Gnidia linearis is almost 
entirely restricted to the subarid bioclimatic zone of 
southwest Madagascar (Toliara Province) from 0- 
1000 m elevation (Fig. 16). The disjunct population at 
Bemaraha extends the distribution by ca. 200 km to the 
north of other populations, and into the island’s dry 
bioclimatic zone. Several other collections along the 
central plateau in the humid bioclimatie zone (1200— 
1400 m elevation) may have been made from cultivated 
plants or au that escaped eule anion: aod are 


enote g marker in tl map 


Phenology. The species flowers and fruits year 


Vernacular names. Hafodramena (Boiteau 


3097; Decary s.n. [30 Aug. 1917]; Humbert 20259); 


354 Annals of the 
Missouri Botanical Garden 


Figure 10. on linearis (Leandri) Z. a ea —A, B. Habits. Note variation in indument, peduncle length, and apex 

of involucral bract. —C. Inflorescence. Flower dissections. One sepal, one petaloid scale, and one anther removed 

from part D. Habits drawn from Rogers ey i 523 (part A, MO) pe Rogers & Rakotonasolo 419 (part B, MO). 

S MUN drawn from Rogers & a 523 (part C, MO). Flowers drawn from Rogers & Rakotonasolo 523 (part D, 
MO) and Rogers & Rakotonasolo 419 (part E, MO). 


hafotra mena (Cours 5275); roinisa (Réserves Natur- IUCN Red List category. Naturally occurring 
elles [Rakotoniania] ; Réserves Naturelles populations of Gnidia linearis are widespread in 
[Ravelonanahary| 4285; Rogers & Rakotonasolo southwest Madagascar and have been recorded inside 

several protected areas (Andohahela, Isalo, Lac 


Volume 96, Number 2 


Rogers 35 
Revision of Malagasy Gnidia (Thymelaeaceae) 


Tsimanampetsotsa). The species is also cultivated in 
obs.) Gnidia 
linearis is assigned to t nservation categ 


c gory o 
Least Concern (LC) wow to IUCN (2001) criteria. 


central Madagascar o nd pers. 


Discussion. Leandri (19302) used two collections 
ME s.n. [June 1915] and Waterlot s.n. [May 
16]. both 
gascars capital, in the description of Lasiosiphon 
Later, Leandri (1947, 1950) 
decided to treat the taxon at the species rank as L. 


noted to be from Antananarivo, Mada- 
decaryi var. erectus. 


erectus, distinguishing it from what he considered to 
be related species (L. dumetorum and L. multifolia, 
both treated here as synonyms of Gnidia daphnifolia) 
by having inflorescences of ca. 15, rather than seven 
to 10 flowers. 
variable in many species of Gnidia with involucrate 


Flower number can be extremely 


inflorescences, and Leandri’s distinction breaks down 
once additional material is consulted. Nevertheless, 
the two Waterlot collections, along with several other 
collections from Madagascar’s central plateau, differ 
somewhat from other populations of G. linearis by 
having more fleshy leaves and glaucous bracts when 
dry. The differences observed in the high plateau 
populations are possibly caused by their substantial 
geographic disjunction and from variation due to 
diverse biophysical parameters, e.g., 1200-1400 m 
elevations in the subhumid bioclimatic zone versus u 
to 1000 m elevation in the dry and subarid bioclimatic 
zone. It is also plausible that the high plateau 
specimens could represent cultivated plants of G. 
linearis that hybridized with another high plateau 
species, such as G. perrieri. Plants of G. linearis 
continue to be brought from southwest Madagascar to 
paper factories operating in several cities on the 
central plateau (e.g., Antananarivo, Ambositra, Fia- 
narantsoa, Ambalavao), and cultivated plants of G. 
linearis (e.g., Rogers et al. 734, not mapped) were 
found growing at a factory in the town of Ambalavao as 
recently as 2008 (Rogers, pers. obs.). Despite several 
attempts, the naturally occurring populations on the 
high plateau have not been located, and additional 
collections and information from the area are neede 
to better understand the observed variation. 

In the Flore de Madagascar et des Comores, Leandri 
(1950) recognized Gnidia linearis as a variety of the 
name L. decaryi, along with two other varieties (L. 


decaryi var. littoralis and L. decaryi var. tenerifolia). 


ey 
a 


His varieties were distinguished by differences in lea 
size and shape (e.g., leaves broader and larger in 
variety littoralis; leaves longer, but not broader in 
variety o flower length (10-12 mm in 
s. 8-10 mm in the other two), bract 
y (eg. apex strongly decurved in variety 
littoralis), and petaloid scales (narrower in variety 


linearis). With additional material, Leandri's distin- 
guishing px d exhibit continuous, overlapping 
een found to 


variation. er differences have 


justify the ina of Leandri's varieties 
Lasiosiphon linearis and L. decaryi were validly 
published simultaneously (Leandri, 1930a: 1040- 
1041), the former name being most recently recog- 
nized by Leandri (1950) in the Flore de Madagascar et 
des Comores as a variety of L. decaryi. Both names are 
considered synonymous here, and if the decaryi 
epithet were transferred into Gnidia the resulting 
combination would cause unnecessary confusion with 
the nomenclaturally similar, but taxonomically differ- 
ent, G. decarya 1 
linearis epithet is adopted to form the new combina- 


na. Therefore, Leandri's less familiar 


tion. 
The more broadly circumscribed Gnidia linearis is 


orange-red), the less fleshy and membranous leaves, 
the bracts that are usually apiculate and strongly 
decurved at the tips (vs. usually long-acuminate or 
acute and with erect tips), narrower leaves, longer 
peduncles, and the longer hypanthium trichomes ([1-] 
3-5 mm vs. 0.3-0.5[-0.7] mm long). 


Typification. 
protologue of Lasiosiphon decaryi (Leandri, 19302): 
Decary s.n. (P), Decary 2759 (P), 2785 (G, P, TAN), 
3183 (MO, P, TAN, US), 3741 (P [2]. All examined 


syntypes are annotated in Leandri's handwriting and 


Five collections were cited in the 


correspond closely to the description. The P sheet of 
Decary 2785 is in particularly good condition and is 
chosen as the lectotype. 

Two collections were cited in the protologue of 


(Leandri, 19302): 


Lasiosiphon decaryi var. erectus 


Waterlot s.n. (June 1915) and Waterlot s.n m 
6). One sheet of each collection has been located 
at P. The sheet dated May 1916 (P 0037 a is in 


better physical condition and is chosen as the 
lectotype. 

Nine collections were cited in the protologue of 
Lasiosiphon decaryi var. littoralis (Leandri, 1947): A. 
Grandidier s.n. (Nov. 1868-Jan. 1869) (P), Humbert 
2489 (G, P, TAN), Humbert & Swingle 5170 (P), 5294 
(G, GH [2], MO, P, US, WAG), 5294bis (P), 5414bis 
(P), Lam & Meeuse 5439 (P, WAG), Perrier de la 
dide 8553 see 12807 (K, P [2]). The Paris sheets of 

all nine s bear Leandri’s own handwritten 
iuis jos de la Báthie 8553 (P) has bracts 
and leaves most closely matching the description and 
is designated as the lectotype. 
ive collections were cited in the protologue of 
Lasiosiphon decaryi var. tenerifolia (Leandri, 1947): 
Decary 2971 (BM, P, TAN), 3783 (MO, P, TAN, US), 


Annals of the 
Missouri Botanical Garden 


8341 (P), 8966 (MO, P), Geay 6328 (P). Decary 2971 
closely matches the protologue and P 00373439 is 
chosen as the lectotype. 


Selected specimens MADAGASCAR. Antana- 
narivo: Anta ssibly cultivated], June 1915, 
la rus bly cultivated], D’Alleiz- 
ma [possibly cultivated], 
, Tapia forest, Peltier 2181 (Pj; Horombe 
Plateau, Jacque 1129 (P); 
7, Rogers & Bieb 419 (K, MO, P, TAN); 
Itremo unm cultivated], Perrier de la Báthie 12471 (P). 
Dm Bemaraha Plateau, Hb. Jard. Bot. Tananarive 
a: Amboasary, Decary 3183 (MO, P, TAN, 


m 


Swingle 5527 (P [2]; Ampanihy to 
W of Ampanihy, Labat et al. 2075 (K, 
re Andohahela Parcel D pon from Tsimelahy W to 
Vohimainty, Birkinshaw GRA, MO, P, TAN); 
Andohahela = #3), 2 p (GRA, MO, P, TAN); 
Andrevo, 35 k of Toliara s coast rd., rud et al. 
1 


744 (K, MO, P. a Androka, Ampanihy on the Linta, on 
Etrobeke, Allorge 2298 (P) dn dm 38 im SW of 
Ampanihy, on rd. to Androka, oa et ii 3447 (G, K, 
MO, TAN, WAG); Ankalitany, 95 W of Fort-Dauphin on 


we E. #13, Miller & ado 6180 (K, MO, P, 
TAN); Ankaroabato (Tul zx Peltier & Moniagnac 
o (P, TAN); Ankilizato, Morondava basin, Perrier de la 
Béthie 8559 (P [2]; la forest station near 
Anta animora, dd Forestier Diu i £u (P, TEE); 
Antanimora—Ambovombe rd., mbovombe, 
Dorr et oan "3064 (K, ‘MO. TAN, is D 2 15 jn 
ENE of Beloha, Fosberg 52485 (MO); Beheloka "NOR near 
beach, Rogers & Rakotonasolo 523 (BM, G, K, MO [2], P. 
S to Tranovako; near Barabay, 
n &  Miljoona 3625 (G, K, M : 
(K); wo Esiva, m 
AN); Fiherenana Valley, 


MO); Itampolo, Phillipson et al. 3742 (G [2], K, MO, TAN, 
WAG); ra Manambolo Valley, Mandrare basin, around 
Isomono, Mtns. Kotriha € Isomonobe, Humbert 12812 (G, 
MO, P); La Table, Dequaire (Hb. St. 2 Alaotra) 27333 


— 


e (G, GH n MO, P, 2. WAG), 5294bis (P), near 
Itampolo, Humbert & Swi 14bis (P); Tsivonoakely, 
18-30 km N of Tuléar, n & o 6104 (K [2], 


Báthie 12807 (K, P [2]; Tuléar, delta of Fiherenaha, 
umbert & Swingle 5170 (P); Tuléar-Manombo, Perrier de 
la Báthie 8553 (P). 


11. Gnidia neglecta Z. S. Rogers, sp. nov. TYPE: 


Madagascar. Toamasina: Andevorante [Andevor- 


salo Natl. Park, 2-3 km N of 


anto], Moramanga, sandy plain, 3 Oct. 1912, K. 
Afzelius s.n. (holotype, P!). Figure 11 


pecies nova quae a Gnidia decaryana Leandri lamina 
| late ovata past cordata (haid obovata vel suborhiculan 


bas p 

rm ato (haud Us cuo conco Le 
inflorescentiis celia. dan sessililibus vel subsessi- 
liibus) et hypanthio glabro (haud extus dense pubescenti) 
differt. 


Shrub; young branches glabrous, flattened (espe- 
cially near internodes); branches not lenticellate, 


pairs 
petioles bu mm, glabrescent; blades broadly ovate, 
8-17 4-12 mm, l:w ratios ca. 1.3—2:1, glabrous, 
abaxial ru slightly glaucous, apex rounded or 
ate, margin with 


obtuse, tip apiculate or rarely emargin 


a distinct vein, base cordate; midrib depressed 
adaxially, raised abaxially, glabrescent and darker 
than 
blade, both surfaces, 


pronounced abaxially; secondary veins forming a 


ade on both surfaces; venation darker than 


glabrous, raised on more 
brochidodromous loop near the margin; fine venation 
uniformly reticulate, densely congested, more pro- 
nounced abaxially, smallest areolae ca. 

0.2 mm. Inflorescences terminal, capitate, 2- to 4- 
flowered; peduncles to 2.5 cm, flattened, glabrous, 
subtended at base by leaves ca. 1/2 the size of next 
lower leaf pair (i.e., modified inflorescence bracts 
7.8- 
abrous; persis- 
tent portion 2.8-3.4 mm; calyx lobes 4, broadly 
elliptic or orbicular, 2.1-2.5 X 1.5-2 mm, glabrous, 
apex rounded, 


absent). Flowers 4- os sessile; hypanthium 
8.2 mm, articulate, + membr 


obtuse, or less often emarginate; 
petaloid scales absent; stamens 8, upper whorl of 
anthers 1/4 to 1/2 exserted, lower whorl 0.7-1 mm 
0.7-0.8 X ca. 


ial disk cupuliform, to 


er upper whorl; anthers elliptic, 
; subgynoec 
i mm tall, glabrous, membranous, apex irregularly 
lobed; ovary ellipsoid, 1.1-1.2 X ca. 0.6 mm, not 
obviously stipitate, mostly glabrous, apex with a few 
0.5(-0.8) mm trichomes; style ca. 2.7 mm, glabrous; 
stigma ca. 2.7 mm below lower whorl of anthers. 
Fruits not seen. 


5 mm, subsessile 


Distribution and habitat. 
eastern littoral species known from a single specimen, 


Gnidia neglecta is an 


which was collected on a sandy plain near Andevor- 
Ambi 


anto and ila-Lemaitso in Madagascar’s humid 


bioclimatic zone near sea level (Fig. 17). 
Phenology. The species flowers in October. 


IUCN Red List category. Gnidia neglecta was 
collected on one occasion in 1912. No populations 
were located at the type locality in February 6 
Due to the apparent rarity of the species, the period of 


Rogers 357 


Volume 96, Number 2 


2009 


Revision of Malagasy Gnidia (Thymelaeaceae) 


p). 


( 


Figure ll. Gnidia neglecta Z. S. Rogers. —A. Habit. —B. Leaf, abaxial surface. —C. Inflorescence. Drawn from 


holotype, Afzelius s.n. 


Annals of the 
Missouri Botanical Garden 


time since it was last collected, an AAO estimated to 
be less than 10 km’, and the d threatened littoral 
forest habitat, it is sae assign ecta a 
preliminary IU 2001) conservation status of 


Critically Endangered (CR) (Blab + 2ab). 


Discussion. Vegetatively, this new taxon might be 
arge genus Wikstroemia 
Endl. (distinguished from Gnidia by the unarticulated 


hypanthium and a relatively well-developed subgy- 


mistaken as a species of the 


noecial disk), or even confused with the small, usually 
lianescent, genus Synaptolepis (tube also unarticulat- 
ed) The flowers on the type of G. neglecta are 
obviously artieulated and lack or possess c a 


es which s 
that this new taxon best fits morphologically Sos 


minute disk up to 0.2 mm tall, featur: 


Gnidia. No species of Wikstroemia (nor the closely 
related Daphne L.) have been collected on Madagas- 
car, and the island's sole species of Synaptolepis (S. 
perrieri Leandri) in no way resembles the new species. 

Gnidia neglecta differs from G. decaryana by its 
broadly ovate leaves with cordate bases (vs. obovate to 
suborbicular leaves with cuneate to attenuate bases), its 
akly arcuate 


more numerous, subparallel or more we 


secondary veins that form a discrete brochidodromous 
loop (vs. de concolorous, strongly arcuate second- 
aries lacking a distinet loop), its conspicuous, dense, 
dark, fine venation (vs. inconspicuous or faint, irregular 
fine venation), its pedunculate (vs. sessile or subsessile) 
inflorescences, and its completely glabrous (vs. exter- 
nally densely pubescent) hypanthium. 

Gnidia neglecta superficially resembles the south- 
east African species G. subcordata o treated 
as Englerodaphne subcordata), b n be disti 
guished by the same venation e e separate it 
from G. decaryana, in addition to the shorter (7.8— 

2 mm vs. 11-15 mm long), completely glabrous (vs. 
sparsely pubescent), cylindrical (vs. = funnel-shaped) 
hypanthium. 


ology. The epithet neglecta draws attention 
to the fact that this distinctive new species was 
overlooked by botanists for almost 100 years, having 
been originally misidentified anonymously as Gnidia 
decaryana. 


2. Gnidia occidentalis (Leandri) Z. S. Rogers, 

comb. nov. Basionym: Lasiosiphon occidentalis 
Leandri, Notul. Syst. (Paris) 13: 47. 1947. TYPE: 
Madagascar. Mahajanga: Kamakama forest, An- 
kara plateau, 14 July 1901, H. Perrier de la Báthie 
1276 (lectotype, designated here, P!). Figure 12. 


Shrubs to lm tall, 


branches densely to moderately pubescent; mature 


weakly branched; young 


branches lacking lenticels or only sparsely lenticel- 


late. Leaves alternate, persistent on older branchlets; 
petioles ca. 1(-2) mm, sparsely pubescent or glabres- 
cent; blades narrowly elliptic or slightly obovate, 2.1— 
4.5(—6.2) X 0.5-1.6 cm, l:w ratios ca. 3.5-6:1, light 
green when dry, both surfaces sparsely to moderately 
pubescent, trichomes ca. 0.5-1 mm, apex apiculate or 
acute, base cuneate to long-attenuate; midrib de- 
pressed adaxially, raised abaxially, lighter green than 
blade abaxially when dry, glabrescent to moderately 
pubescent on both surfaces; venation usually raised 

on both surfaces, more pronounced abaxially. Inflo- 
rescences axillary or terminal, involucrate, 7- to 14- 
flowered; peduncles 2-20(-32) 


moderately pubescent; involucral bracts 5(6), narrow- 


mm, densely to 


ly lanceolate or elliptic-ovate, 8-15 X 1.8—4 mm, l:w 
ratios 3—7:1, spreading, persistent, apex acute, base 
rounded-truncate, e rA to sparsely pubescent 
n bot ces; nspicuous adaxially, 
usually conspicuous in upper half abaxially; nervation 
faint or inconspicuous on both surfaces. Flowers 5- 
merous, Br pedicels 0.4—1.5 mm, densely covered 
with (1.5— 
1l 


mm, Pr 


.5-3.5 mm trichomes; hypanthium 12.5— 
coriaceous; caducous portion 
densely pubescent externally with 0.5-1(-1.5) mm 
trichomes, glabrous or sparsely pubescent near artic- 
ulation intern 
ly pubes 
glabrous internally; calyx lobes 5, broadly elliptic- 
oblong or obovate, 1.7-2.9 X 1.2-1.7 mm, glabrous 
adaxially, densely pubescent abaxially. 

l ded; pe 
narrowly elliptic, 0.5-1 X 


y, apex emargin- 
ate or rarely roun taloid scales 5, linear or 
0.1-0.3 mm, membranous, 
glabrous, apex rounded arginate; sta- 
mens 10, upper whorl of anthers located just below 
mouth, lower whorl 1-1.5 mm below upper whorl; 
anthers oblong, 1-1.5 X ca. 0.3 mm, subsessile; 


subgynoecial disk cupuliform, 0.1-0.3 mm tall, gla- 


0.6 mm, sessile or shortly 
stipitate (stipe to ca. 0.2 mm), glabrous; style 4— 
7.6 mm, glabrous; stigma 0—4 mm below lower whorl of 
anthers. Fruits ovoid, 3-3.4 X 1.3-1.6 mm, glabrous. 
Distribution and habitat. Gnidia occidentalis is 
known from several collections from northwest Mada- 
F (Mahajanga Province} in the dry bioclimatic zone 
17). 


No elevation data are available 


3 


ecimen 
labels, but the estimated range for the species is 150— 
400 m based on an ArcGIS digital elevation model 
(DEM). The species is noted on labels as occurring in 
forested habitats on calcareous limestone, basalt, and 
granite. Given what is known about the ecological 
preferences of other species of Malagasy Gnidia, it 
seems likely that G. occidentalis grows along forest 
edges or in open sunlit patches within scrub forest. 


Volume 96, Number 2 


Rogers 35 
Revision of Malagasy Gnidia (Thymelaeaceae) 


Mn re 12. 
ough caducous portion of hypanthium. —D. lona dinal se section oe persistent portion of hypanthium. Drawn from 
Moras 4550 (TAN). 


Gnidia occidentalis (Leandri) Z. S. Rogers. 


Phenology. The species flowers and fruits in May, 
October, and November. 


war name. Fafitao (Réserves Naturelles 


Vernac 
[Rakotovao] 5393). 

IUCN Red List category. Gnidia occidentalis has 
been recorded from one protected area (Namoroka). 
This species is provisionally assigned a conservation 
status of Least Concern (LC) based on IUCN (2001) 


criteria. 


—B. Inflorescence. —C. Flower, longitudinal section 


Discussion. At the time of the original description, 
Leandri (1947: 47-48) cited five collections of Gnidia 


occidentalis, which were informally categorized as 
el 


Bât ), or “probablement des formes de la 
sme espèce” (Decary 8181, Perrier de la Báthie 998, 
8549, rue The flowers of the species were 

n the protologue as 5-merous. Strangely, 
Duk (1947, 1950) failed to notice that two of his 
syntypes (Decary 8181 and Perrier de la Báthie 998) 


Annals of the 
Missouri Botanical Garden 


consistently have 4-merous flowers and that the leaves 

and braets closely matched those E the similarly 

distributed species, G. gilbertae. collections 
94 


without a doubt represent that species. tee ri 


T 


1950) mentioned that another one of the syntypes, 
Perrier de la Báthie 16324, possessed certain characters 
of G. daphnifolia, and the collection is identified here as 
that species. The two remaining syntypes, Perrier de la 
Báthie 1276 and Perrier de la Báthie 8549, differ 
consistently from G. daphnifolia and G. gilbertae by 
several characters, and thus are treated here as 
belonging to the recireumscribed G. occidentalis. 

nidia occidentalis is distinguished from G. 
daphnifolia by its bracts drying light green, green- 
red, or yellow-brown with 3—7:1 l:w ratios (vs. drying 
black or brown at least in the lower half and with 2— 
4:1 l:w ratios), and its 12.5-16 mm (vs. 6.5-12[-15] 
mm) long hypanthia. 

ompared to Gnidia gilbertae, G. occidentalis 
differs by its 5- (vs. 4-)merous flowers, its articulate 
(vs. unarticulate) hypanthium, and its narrower bracts 
measuring 1.8—4 mm (vs. 4—7 mm) wide. 

When sterile, Gnidia occidentalis may be difficult 
to distinguish from ojeriana, but it can be 
recognize its narrower spreading (vs. recurved) 
bracts with sparser indument. 


Nomenclature and typification. The earlier invalid 
homonym Gnidia ut Rege (1860) was origi- 
nally published as à nomen nudum and presumably 
based on an Asian species of Diarthron Turcz. or Stellera 
L. Gnidia occidentalis Regel was never validated later by 
a description or diagnosis, making the epithet available 
for use in the new combination proposed her 

As 2 above, Leandri (1947, 1950) consid- 
ered Perrier de la Báthie 1276 to represent the most 
common form of Gnidia occidentalis, and now only two 
of the five collections cited in the protologue (Leandri, 
1947) belong to the recircumscribed G. occidentalis: 
Perrier de la Báthie 1276 (P) and 8549 (P). The sheet 
of Perrier de la Báthie 1276 (P 00373473) i 
designated as the lectotype. 


Selected specimens examined. MADAGASCAR. Maha- 
janga: Manasamody, betw. Port Bergé & Antsohihy, Morat 
4550 (P. TAN); Mt. Ambohibenga, Milanja, near Cap 
d'Andre, Perrier de la Báthie 8549 (P); Namoroka (Natural 
Reserve #8), Soalala, Réserves Naturelles (Rakotovao) 5393 
(P). 


13. Gnidia perrieri (Leandri) Z. S. Rogers, comb. 
nov. Basion 
Notul. 

scar. Fianarantsoa: Andringitra Massif 
(Iratsy), valley of Riambava & Antsifotra, 2000— 
5 27 Nov. 1924, H. Humbert 3827 


(lectotype, designated here, P!; isotypes, BM!, 
G [2]!, K!, MO!, TAN!, US). Figure 13. 


Shrubs to 1.5 m tall, 


branches glabrous, red-purple or orange-red when dry; 


weakly branched; young 


mature branches usually lenticellate. Leaves alternate, 
rarely subopposite, usually persistent on older branch- 
lets, sessile or subsessile; petioles to m, glabrous; 
blades broadly elliptie or slightly e 9-35 X 4— 

8 mm, l:w ratios ca. 2-5:1, usually drying dark green, 
sometimes glaucous, membranous and somewhat fleshy, 
both surfa 


altenuate or cuneate; midrib raised or plane on both 


ces glabrous, apex apiculate, base long- 
surfaces, glabrous; venation discolorous, usually incon- 
spicuous adaxially, inconspicuous or raised abaxially. 
Inflorescences axillary or terminal, involucrate, 6- to 
19-flowered, subsessile to short-pedunculate; peduncles 
2-5(-8) mm, glabrous, drying dark; involucral es 55 
persistent through fruiting phase, ovate, 4.9—10.8 X 

5.4. mm (of various sizes in the same Ee. lw 
ratios 2-3.5:1, usually dark 
somewhat fleshy, erect, glabrous adaxially, densely to 


red-purple when dry, 


moderately pubescent abaxially, apex acuminate, acute, 
less often apiculate, base rounded or rounded-truncate; 
midrib raised or inconspicuous on both surfaces; 
nervation inconspicuous or with 2(4) veins diverging 
longitudinally from midvein near base. Flowers 5- 
merous, yellow or orange, short-pedicellate; pedicels 
0.3-0.9 mm, densely puberulent, trichomes to 0.3 mm 
(i.e., not reaching base of ovary); hypanthium 6-89 
densely 


internally between anthers and above articulation; 
caducous portion with trichomes to 0.3 mm; persistent 
portion 2.5-3.5 mm, trichomes 0.3-0.5(-0.7) mm; calyx 
lobes 5, broadly elliptic or obovate, 1.4—2.9 X (0.8-) 
1.2-2.1 mm, semi-membranous, glabrous and papillate 
adaxially, densely pubescent abaxially, apex emarginate 
or rounded; petaloid scales 5, oblong-elliptic or obovate 
to suborbicular, 0.3-0.6 X 0.2-0.5 


glabrous, apex emarginate or rounded; stamens 10, 


mm, membranous, 


upper whorl of anthers ca. 1/4 to 3/4 exserted, lower 
whorl 0.1—0.5 mm below upper whorl; anthers elliptic or 
elliptic-oblong, 0.6-0.9 X 0.2-0.3 
gynoecial disk cupuliform, 0.1—0.2 mm tall, glabrous, 


mm, sessile; sub- 


fleshy, apex smooth or with a few irregular lobes; ovary 
ellipsoid, 1.1-1.5 X 0.5-0.6 


sometimes with a few trichomes to 0.3 mm near apex; 


mm, sessile, glabrous, 


style 3.2-7.8 mm, portion below articulation persistent, 
glabrous; stigma located just below or at height of upper 
hers. Fruits ellipsoid, 3.3-4.1 X 1.4— 
1.5 mm, glabrous. 


whorl of ant 


Distribution and habitat. 
demic to the Andringitra massif from 2 


Gnidia perrieri is en- 
0-2550 m 


elevations (Fig. 16). The species occurs in sunlit, 


Volume 96, Number 2 
2009 


ogers 36 
Revision of Malagasy Gnidia (Thymelaeaceae) 


13. Gnidia perrieri (Leandri) Z. S. Rogers. 
Persistent portion of the 


ericoid scrubland and on rocky slopes in the 
subhumid bioclimatic zone. 
Phenology. The species has been collected in 
flower and fruit in May, October, and November. 
IUCN Red List category. Gnidia perrieri is en- 
demic to one formally protected area (Andringitra). 
The species must be somewhat resistant to fire as it 


—A. it. 
hypanthium surrounding the gynoecium. —E. 
Humbert 3827 (P). Seed drawn from Guillaumet 3550 (TAN) 


Habit. —B, C. Caducous portion of the hypanthium. —D. 
Seed. Habit and floral parts drawn from isotype, 


grows in ericoid serub along a few of the higher 
plateaus and peaks in the area. The species is 
assigned a provisional IUCN (2001) conservation 
assessment of Vulnerable (VU) to extinction (Blab + 
2ab). 
Discussion. Gnidia perrieri is distinguished from 
linearis and G. daphnifolia, the two most 


G. 


Annal 
Missouri Botanical Garden 


morphologically similar Malagasy species, by its red- 
purple or orange-red (vs. green, brown, o 
young stems after drying, its membranou 

(vs. coriaceous, chartaceous, or rarely slaty fleshy) 
blades and bracts, its 2-5(-8) mm (vs. up to 50 mm) 
long peduncles, and its 0.3-0.50.7) mm (vs. [1-]8— 
5 mm) long trichomes on the persistent portion of the 
hypanthium. In the dried state, collections of G. 


perriert d have slightly glaucous leaves, bracts, 


al to continental African Gnidia, C. 
perrieri is most similar to G. macropetala Meisn. 
Gnidia perrieri differs by its LA glabrous (vs. 
densely pubescent) stems and lea h 
leaves and bracts, its five "| 


es, its more 
eight) involucral 


gaia and its shorter hypanthia (6-9 mm vs. 12— 
mm). 


Typification. Four collections were cited in 
the pec Humbert 3827 (BM, G [2], K, MO, 
P, TAN, US), Perrier de la Báthie 8554 (P), 13700 
0, i (P). Each one of these was annotated 


lectotype as it is in the best physical condition 
and its duplicates are the most widely distri- 
buted 


Selected specimens examined. MADAGASCAR. Fianar- 

oo massif, Perrier de la Bathie 8554 (P), 

0 XP). 8 (P n ringitra, Pic Bory [Boby], 
CURRY. mo pm P, TAN). 


4. Gnidia razakamalalana Z. S. Rogers, Ad- 
ansonia, sér. 3, 28: 156. 2006. TYPE: Mada- 
Fivondronona Fort-Dauphin, 


2005, R. 


(holotype, MO!; 


ure 14 


> K!, Pt TAN). Fig- 


Treelets 2 m tall, with dichotomous branching; 
internodes very short (ca. 1 mm long near branch 
tips); branches glabrous, covered by prominent leaf 
scars. Leaves alternate, spirally arranged, sessile or 
blades 
ovate-lanceolate or elliptic, 4.5-8.4 X 1-1.2 em, l:w 


subsessile, persistent only at branch tips; 


ratios 4—7:1, adaxially ig iru densely 
sericeous, apex a acute, base obtuse-shortly decurrent 
or truncate; i ic "adaxially, raised 
abaxially; venation brochidodromous, inconspicuous 
o ly faintly visible, more obvious abaxially; 
petioles 0-0.3 mm, densely sericeous. Inflores- 
cences terminal, erect, 1-flowered, leaving prominent 
scars on older branches. Flowers 5-merous, reddish 


white, sessile or subsessile, surrounded by several 


involute leaves (each at least partly appressed to the 
the 


becoming smaller in size and lighter in color); 


lower portion of the floral tube, distal ones 

hypanthium ca. 5 cm, ca. 1 mm diam. near base, 

ca. 3mm diam. at mouth, articulation not seen, 

externally covered by 
i 


ens ous-tomentose 
iaa with longer straight trichomes, 


serice 
trichomes 
indument + uniform along the length of the tube, 
glabrous er RA lobes 5, spreading, lance- 
olate-ellipti 2c 3.9 

branous than hypanthium, papillate and sparsely 


mm, more mem- 


tomentose-sericeous trichomes shorter 
n the ifle abaxially, apex acute or 
apiculate, with a dense tuft of straight trichomes; 
petaloid scales 5, ii ovate, or -n 1.8- 

1-1 m membranous, glabrous, 
irregular l- to " eed, with each lobe of vary rying 


apex 


subgynoecial disk cupuliform, to 0.7 mm tall, glabrous, 
membranous, apex irregularly lobed, sinuses mostly 
shallow; ovary ellipsoid, ca. mm, shortly 
stipitate, apex covered with a tuft of 1.5-2.3 mm 
trichomes, otherwise glabrous; style 1.5-1.8 em, 
glabrous; stigma 


(Le., 


located near middle of tube), fusiform, ca. 4 mm 


+ flattened, ca. mm wide, 


ca. 1.5 mm below jum whorl of anthers 
long, ca. 0.5 mm wide, densely papillate. Fruits not 
seen. 

Distribution and habitat. Gnidia razakamalalana 
is apparently a narrow endemic, with only one known 
population occurring in a subcoastal forest in 
a ooa at ca. 100m elevation 
i e es pe ws slope among 
boulder on black pes soil. (R. 


Razakamalala, pers. oum 


Ai granite 


Phenology. The species has been collected in 
flower in February and November. 
IUCN Red List category. Gnidia razakamalalana 
is w from a single ap i s 
e AOO is cer d be more than 10 km? 
given a grid cell of the same size. im. the 
restricted range of the species and unprotected 
habitat, 
preliminary conservation status of Endangered (EN 


(Blab + B2ab). 


the species is assigned an IUCN 1) 


Discussion. Gnidia razakamalalana is easily dis- 
tinguished from all other Malagasy species by the ca. 
ong flowers arranged in 1-flowered terminal 


inflorescences, and the 1.7-2.2 cm X 3.5-5 mm 


Volume 96, Number 2 
2009 


Rogers 36 
Revision of Malagasy Gnidia (Thymelaeaceae) 


14. Gnidia razakamalalana Z. S. Roge 
disk. —E. Stigma. —F. Petaloid scales. Drawn in holotype, Razakamalala et al. 2670 (MO). 


Figure 


calyx lobes. Sterile specimens are also easy to identify 
because the leaves are densel 
and completely glabrous adaxially. 


y sericeous abaxiall 


The large fusiform stigma (Fig. 14C) is unique 
within the Malagasy members of the genus. At 
present, this species is only known from flowering 


material and the hypanthium does not show any signs 


s. —A. Habit. 


—B. Leaf. —C. Flower. —D. Gynoecium and subgynoecial 


of articulation. Presumably, the articulation develops 
in late anthesis, but additional collections are neede 


to confirm this dopethesis 


Selected specimen examined. MADAGASCAR. Toliara: 
Anosy m Fort Dauphin, Commune labakoho, Quartier 
Antsotso, Ivohibe- und. Rabenantoandro et al. 1725 
(MO, P, TAN. 


Annals of the 
Missouri Botanical Garden 


Figure 15. d "RT of Gnidia ambon- 
Tro Oy (Boiteau) Z. S. Rogers (BI), G. bojeriana (Decne. 
Gilg (@), and G. daphnifolia E 1 (A). 


EXCLUDED NAMES 


Arthrosolen madagascariensis Endl., Gen. Pl. Suppl. 
4(2): 63. 1848, nom. illeg. [= Phaleria octandra 
(L.) Baill., Adansonia 11: 321. 1875] 


Endlicher published d madagascariensis 
with reference to Lamarck’s D : 254) description of 
Dais octandra L has Reed recognized by 
Rye (1990) as a synonym d the widespread Pacific 


species Phaleria octandra (L.) Baill. 


Dais Pm F Baill., Hist. v. SR ac 35(5) 
, pl. 31 8. 1895 : “Madagascar” 
fone. pl. 318!, 1895) 
[= Dais glaucescens Decne. in C. A. Mey., Ann. 
Sci. Nat. Bot., sér. 2, 20: 51. 1843]. 


Baillon in UE 


The protologue of Dais rhamnifolia Baill. con- 
sisted of a AC Rd illustrated diagnostic plate 
“Madagascar” an 
iodo Dats 


the species name 
rhamnifolia taxonomi- 
cally belongs to the earlier named species D. 


glaucescens. 


er E L. £ var. hirsuta L. £., Suppl. Pl. 
1782. Dessenia hirsuta (L. f. ) Raf., Fl. 


A), and G. 


ed from cultivation 


possibly cultivated or escap 
perrieri (Leandri) Z. S. Rogers (sh). 


a 


Tellur. 4: 106. 1838. TYPE: Ab. Smith No. 688.1 
(lectotype, designated by Rogers in Rogers & 
pencer, 6: , LINN-SM!) [= Gnidia 
capitata L. f., Suppl. Pl. 224. 1782]. 


Gnidia daphnifolia var. hirsuta, and most likely the 
material on which it was based, was mistakenly 
attributed to Madagascar by Linnaeus (1782: 225) and 
was apparently used as the original material in the 
description of G. capitata L f. an African species 
described on the previous page (Linnaeus, 1782: 224). 
The name was therefore designated as the lectotype of 
G. capitata in Rogers and Spencer (2006). Gnidia 
daphnifolia var. glabra pertains to a Malagasy species 
and is based on the same original material as the 
binomial, G. daphnifolia (Rogers & Spencer, 2006; 


see above under that species). 
Lasiosiphon cuneatus Decne., Voy. Inde 4: 149. 1844. 
[= Gnidia sp.?]. 


Decaisne (1844: 149) published a scant description 


in the protologue with reference to “Dais cuneata 


Volume 96, Number 2 
2009 


Rogers 36 
Revision of Malagasy Gnidia (Thymelaeaceae) 


Figure 17. Geographic distribution of Gnidia decaryana 
G. gnidioides (Baker) Domke (@), G. humbertii 
ers (qk), G. 
occidentalis (Leandri) Z. S. Rogers (3K), and G. razakamala- 
lana Z. S. Rogers (X). 


Lamk., Le." which supposedly referred to a name 
given on page 255 of the second volume of Lamarck's 
Encyclopédie Méthodique (1786). Apparently, La- 
marck never actually published the Dais name, and 
Decaisne might have instead been trying to cite a 
Lamarckian manuscript name. Lasiosiphon cuneatus 
Decne. was treated as an insufficiently known species 
i : 599) and Leandri Mies 1042; 
1931a: 676), and I have not been able to find any 
relevant original material for the name. Te proto- 
logue description lacks sufficient diagnostic charac- 
ters to determine the affinity of this taxon within the 
Thymelaeaceae. 
E x di ier Baker, J. Linn. Soc., Bot. 25: 
as “Lasiosiphon? rhamnifolius.” 
n ae Vonizongo distr., s.d., Baron 
5115 (holotype, K!; isotypes, P [2]) [= Dais 
glaucescens Decne. in C. A. Mey., Ann. Sci. Nat. 
Bot., sér. 2, 20: 51. 1843]. 


The original material of Lasiosiphon rhamnifolius 
belongs taxonomically to Dais glaucescens. 


Literature Cited 

Aymonin, G. G. 1962. Quelques Thyméléacées de rocailles 
montagnardes, ui Wistron et Gnidia. Pl. Mont. 3: 
80-185. 

— — —. 1965. Diversification, ré titi t endémi h 


quelques groupes de Thméléscées de la flore cu 
. Sommaire Séances Soc. Bio- 


géogr., no. 365: 6-21. 

966a. Thyméléacées. Pp. 35-95 in A. Aubréville 

(dion Flore du nk Vol. 11. Muséum National 

d'Histoire Naturelle, 

———. 1966b. Thy es Pp. 3-86 in A. Aubréville 
lied: Flore du Cameroun, Vol. 5. Muséum National 
d'Histoire Naturelle, Pari 

.1 Sur un Grilla (Thyméléacées) á inflores- 

cence complexe du Cameroun. Bull. Soc. Bot. France 112: 
321-325. 

Baillon, H. 1875. Histoire des Plantes, Vol. 6. L. Librairie 
Hachette et Cie, P. 

Baker, J. G. 1883. ee to the flora of Madagascar, 

20: 237-303. 


E 


Edwards & F. R. Smith. 2001a. Leaf 
bract o in Gnidia (Thymelaeaceae): Patterns 
and taxonomic value. Syst. Geogr. Pl. 71: 399-418. 
& . 2001b. Patterns of diversity 
mm. involucral bracts, inflorescences and flowers in 
Gnidia (Thymelaeaceae). Syst. Geogr. Pl. 71: 419-431. 
ce xs C. L. & J. B. P. Beyers. 2003. Thymelaeaceae. 
Pp. 928-935 in G. Germishuizen &N.L.M 
Plants of Southern Africa: 
Strelitzia 14, rt Botanical Institute, Pretoria. 
E. Schatz, G. McPh 


Lis S. ers, R. 

Rabehevitra. 2006. Deforestation and plant diversity of 
Madagascar’s littoral forests. Conservation Biol. 20: 
1799-1803. 

Cornet, A. 1974. Essai de N bioclimatique à 
Madagascar. Not. id ORSTOM No. 55. 

i ws J. 1844. Plantae Rario nl 
Pp. 57-183 in V. sequeron (dios, TOS m "Pinde, 
Vol. 4. Firmin-Didot, Paris. 

Domke, W. 1934. Untersuchungen über die systematische 
und d irs Gliederung der Thymelaeaceen. Bib- 
lioth. Bot. 27(111): 1-151. 

Dorr, L. J. 1 997. Plant Collectors in Madagascar and the 
Comoro Islands. Royal Botanic Gardens, Kew. 

Thymelaeaceae. Pp. 625-642 in 

Drude (editors), Die we der 

eft 2 


Lei eipz ig. 
Estragon, M. 1933. Les industries locales, la fabrication du 
apier Antaimoro. Rev. Madagasc. 4: 59-62. 
Gastaldo, P. 1969. na Florae Aethiopicae, 19. 
Thymelaeaceae. Webbia 389. 
Gilg, E. 1894 s b aoe Pp. 216-245 in A. Engler & 
K. Prantl (editors), Die natürlichen Pflanzenfamilien, 
Vol. 3, 6a. W. Engelmann, Leipzig. 
Grandidier, A. Histoire Physique, Naturelle et 
hg (Atlas 3), part 3, fasc. 38. Imprimerie 
nationals, Par 
1896. Histoire Physique, Naturelle et Politique de 
Mada ascar, Histoire Naturelle des Plantes, Vol. 35, 
Tome 5 (Atlas 3), part 3, fasc. 40. Imprimerie sitial. 
aris. 


366 Annals of the 
Missouri Botanical Garden 
Hallé, F., R. A. A. Oldeman & P. B. Tomlinson. 1978. Regel, E. 1860. Catalogus Plantarum quae 


Tropical Trees and Forests: An cu Analysis. 


M. J. 1990. A revision of the genera Kelleria and 

peo (Thymelaeaceae). Austral. a Bot. 3: 595- 
652. 

Heinig, K. H. 1951. Studies in the floral morphology of the 
Thymelaeaceae. Amer. J. Bot. 38: 113-132. 
Herber, B. E. 2003. Thymelaeaceae. Pp. 373-396 in K. 
Kubitzki (editor), The Families and Genera of Vascular 
Plants, Vol. 5: Flowering Plants. Dicotyledons. Malvales, 
Fs M and Non-betalain Caryophyllales. Springer, 
Ber 

IUCN. 2001. IUCN Red List Categories and Criteria, Version 
3.1. Prepared by the IUCN Species Survival Commission. 
IUCN, Gland, Switzerland, and Cambridge, United 


Kingdom. 
Lamarck, rae B. ue Encyclopédie Méthodique, Botanique, 
Vol. 2 i 


Leandri, r1 . Descriptions d hyméléacées 
Madagascar (Lasiosiphon, prune Bull. Soc. 
France 76: 1039-1043. 

1930b. Thyméléacées nouvelles de Madagascar. 
Bull. pe Natl. Hist. Nat., sér. 2, 1: 435-437. 

a. Révision des Thyméléacées de Madagascar. 

Bull. D Natl. Hist. Nat., sér. 2, 2: 668—676. 

1931b. Révision des Thyméléacées de Madagascar. 

Bull. Mus. Natl. Hist. Nat., sér. 2, 3: 148-160. 

1947. Nouvelles rien sur - de pacess 
de a Notul. Syst. (Paris) 13: 3 

méléacées (family 146). = js 40 in H. 
Humbert da Flore de oun et y Comores 
(Plantes Vasculaires). Firmin-Didot, 

Linnaeus, C. (filius). 1782. m edle Plantarum, 
Impensis ud n wick. 

McNeill, J., F. Burdet, V. Demoulin, D. L. 
Hawksworth, Marheld, > H. Nicolson, J. Prado, P. C. 

, J. H. Wiersema & N. J. Turland (editors). 

2006. [e dem Code of Botanical Nomenclature 
(Vienna Code). Regnum Veg. 146. 

uci P F. 1857. Ordo CLXVII. Thymelaeaceae. Pp. 493— 

A. P. de Candolle, Prodromus a Made 


de 
Bot. 


2. A revision of Ehretia «is naue for 
Madagascar and the Comoro Islands. Adansonia, sér. 3, 
24: 137-157. 
Moore, S. L. M. 1920. Plantarum Mascarensium Pugillus. J. 
0 


W. 1910. Order CXVIII. Thymelaeaceae. 
Pp. 212-255 in. W. T. Thiselton-Dyer, Flora of Tropical 
Africa, Vol. 6. L. Reeve & Co., London. 

Peterson, B. 1959. Some interesting species of Gnidia. Bot. 

Not. 112: 465—480. 

. 1978. Thymelaeaceae. Pp. 1-35 in R. M. Polhill 
(edito), Flora of Tropical East Africa. Whitelriars Press, 
Lon 


. Thymelaeaceae. Pp. 85-117 in G. W. Pope, 
R. M. Polhill & E. S. Martins (editors), Fon Zambesiaca, 
Vol. 9, Part 3. Royal Botanic Gardens, Kew. 

Rabakonandrianina, E. & G. D. Carr. ca Fiere d 
numbers of Madagascar plants. Ann. Missouri Bot. Gar 

23-125. 

ala L. 9 
glaucescens (Ihymelusacene) M 
Antananarivo, Ántananarivo. 


98. Ey de la Plante Dais 
Thesis, University of 


in 
Aksakoviano Coluntur. Typography of State a. Se 
Petersbur 
Robyns, A. 1975. Thymelaeaceae. Pp. 1-68 in P. Bamps 
Pa Flore d'Afrique Centrale (Zaire- -Rwanda-Bur- 
undi). Jardin PE nique National de Belgique, Brussels. 
Rogers, Z. S. A revision of Wer depline Baill. 
oen Adans onia, a 3, 26: 7-35. 
. À revision of Octolepis D (Thymelaeaceae, 
m Adan onia. sér. 3, 27: 89-111. 

.An PE of Malagasy Gnidia and the 
— of Octolepis decalepis (Thymelaeaceae). 
nsonia, sér. 3, 28: 155-160 

ae 2006 Typifiċation of Linnaean and 
Linnaeus filius plant names in Thymelaeaceae. Taxon 55: 


Rye, B. L. 1990. Thymelaeaceae e ia 
Pp. 122-215 in A. S. George (editor), F of Aus 
Vol. 18. Australian Pen xri eran 


Canberra. 

Schatz, G. E. 2000. Endemism in re Malagasy tree flor 
Pp. 1-9 in W. R. Lourengo & S. M. Goodman E 
Diversity pg Endemism in e Mém. Soc. 

iogéogr., Paris. 
& P. P. Lowry II. 2002. A synoptic revision of the 

usd Buxus L. ena in Mei x and the Comoro 
ands. Adansonia, sé 96. 
M. Lescot. 2009. ae to Malagasy Botanical 
era Localities. Missouri Botanical Gard en web- 


w 
iw) 


a accessed 13 April 20 

P. P. Lowry II & A. e Tu 2001. Endemic 
families of Madagascar. VII. A synoptic revision of 
e na Thouars sensu stricto (Sarcolaenaceae). Adan- 

, sér. 3, 23: 171-189. 


Seii. Eli. E " oe Lasiosiphon hildebrandtii. J. Linn. 
Soc 


Staner, 4 peux es Thyméléacées de la Flore du Congo 
Belge. Bull. m ars État 
Townsend, C. Thy 
D.D " d ) to t] 
of Pur Vol. A ae Publishing Co., New Delhi. 

Van der Bank, M. Fay & M W. Chase. 2002. 
a cee te af Tisinclaseas with particular 
reference to African and Australian genera. Taxon 51: 
329-339. 

Wright, C. H. 1906. Gnidia mollis. Bull. Misc. Inform. 1906: 
2 


13: 321-372 
melaeaceae. Pp. e 511 in M. 
sed Flor: 


1915. Order CXVIII. Thymelaeaceae. Pp. 1-81 in 
W. T. Thiselton-Dyer (editor, Flora NS Vol. 5, 
Sect. 2, Pt. 1. L. Reeve & Co., London 


APPENDIX 1. List of recognized species of Malagasy Gnidia. 


Gnidia ambondrombensis (Boiteau) Z. S. Rogers 
Gnidia bojeriana (Decne.) Gilg 
idia danguyana Lea 


Gnidia gnidioides (Baker) Domke 
Gnidia hibbertioides (S. Moore) Z. S. Rogers 
Gnidia humbertii (Leandri) Z. S. Rogers 
Gnidia linearis (Leandri) Z. S. Rogers 
. Gnidia neglecta Z. S. Rogers 
12. Gnidia occidentalis (Leandri) Z. S. Rogers 
13. Gnidia perrieri (Leandri) Z. S. Rogers 


Volume 96, Number 2 
2009 


Rogers 36 
Revision of Malagasy Gnidia (Thymelaeaceae) 


14. Gnidia razakamalalana Z. S. Rogers 


APPENDIX 2. Index to Exsiccatae. 


e jr are e M dune oe by a first oo E 
wed by number or date of collec 


and syntypes, are see in boldface for all names. 


Afzelius s.n. [3 Oct. 1912] (11); s.n. [15 Feb. 1913] (9). 
Allorge 2208 (10) 2308 10). Alluaud 36 (4); 85 (4); 106 
( ndri pp 186 (4). 

94 (7); 2061 (7); 3309 (7); ud 
4281 (2); 5254 O. So (4); 6191 (4). Basse s.n. [18 M 


~ 
m 


I 
— 

= 
Ea 
= 

E 


1931] (10). Bernier 157 (4). Birkinshaw 435 
00) Bisset M3 (6). Es 3097 (10); 4643 [= Hb. ee 
3] (1). Boivin 2384 (4); 53] 
a pe s.n. (2); ig) Bosser 250 (10); 3701 aU. 4066 
(10); 413 (7); 4313 d 4320 (10); 5929 (4); 6065 (7); 7269 
(7); 8525 (7); 9936 (10); 10000 (7); 10229 (10); 10275 (10); 
13002 (2); 14048 ds 14315 (4); 14328 (4); 14329 (5); 
19176 (10); 19457 
mpenon s.n. a Catat 4327 ~ ew 2 na 29 (10); 
135 p 218 (4). Commerson s.n. n. (4; s 
(4). Cours 1518 [= Hb. St. Agric. o Ma dx 1797 (7); 
3143 ae 4564 (10); 5108 xb 5176 A 5275 (10); 5479 (4). 
Cremers 1632 (7). Croat 30900 (10); 31514 (10); 31647 (10); 
31654 10x 31660 (10); 31661 (10); 31681 (10); 31731 (10); 
31930 (10); 31945 (4); 32018 (10) 
D'Alleizette 394M (7); 769M (10); 1110 (7). D'Arcy 15430 


— 


3741 (10); 3897 (4); 4030 (4); 4332 (5); 4336 (4); 7353 
(4); 7399 (7); 7676 (4); 8181 (6); 8340 (10); 8341 (10); 
8391 (4); 8437 (4); 8966 (10); 9186 (10); 9831 (4); 9971 
(5); 9972 (4); 10038 (4); 10049 (5); 10102 (4); 10111 (4); 
FE me Ss a a i pur we "ees (7); s.n. 
2] (5). Ys aire 

s (10); 203 do 27547 ur Ton 283 (4); 431 
(4); 641 (4); 653 (4); 941 (7); 1251 (10); 1373 ae 1395 (10); 
1526 (1 ee quie 0); 2616 (4); 2744 (10); 2845 (7); 
2990 (2); 3283 (4). p 2949 (6); 3964 (10); 4055 (10). 
Douliot s.n. [19 Oct. 1891] m s.n. [1892] (9). Dumetz 
614 (4); 615 (4). Dupuy, B. MB197 (4). 

Eboroke 940 (10). ir 95-36 (9); 95-51 (10); 95- 
56 (4). 

Falinianina 29 (3). Fosberg 52380 (7); 52485 (10); 52495 
(10). 


CES 
E 


Geay 6328 (10); 7538 (3); a (3); 7540 (3); 7541 (3); 
7542 (3). Gentry 11909 (4); 62151 (6). Gereau 3242 (4). 
Goudot s.n. Pee Hs. s.n. Iis pos (7); s.n. [1840] (10). 
Grandidier, A. Nov. 1868-Jan. 1869] (10); s.n. 
[1876] (6). Guillaume e y 550 (13). Guittou 4 (4). 

Harder 1607 (4); 1665 (4). Heim s.n. [2 Oct. 1934] (1). Hb. 
Inst. Sci. Madag. 482 i Hb. Jard. Bot. Tananarive 2639 (4 
4643 [= Boiteau 4643] (1); 5051 (2); 5142 (7); 6151 (10); 
6513 (3); 6615 (4). Hb. Smith No. 688.2 (4); 688.3 (4); 688.4 
a 688.5 (4); Mens ). Hb. Mus. is 4297 (4); s.n. (7); 

n. (4). Hb. St. Agric. Alaotra 1518 [= Cours 1518] (4); E.37 
D. Hildebrandt 3369 (4); 3883 (7). Homole 167 (4); 1467 
(10); s.n. [24 Feb. 1945] (7). Hong Wa 130 (3); 143 (7); 204 
(4); 358 a 414 (6). Humbert 2489 (10); 2744 (4); 2755 
(4); 2844 (9); 2936 (10); 3772 (7); 3827 (13); 4661 (7); 


= 


4962 (10); 5039 (9); 5097 (10); 5170 (10); 5294 (10); 
5294bis (10); 54 14bis (10); 5527 (10); 5593 (10); 5720 
(4); 5720bis (10); 5890 (4); 5916 (5); 6742 (4); 12812 (10); 
12812bis (4); 12844 (4); 13010 (4); 13154 (4); 13242 
(4); 13800 (4); 13860 (4); 13884 (5); 14053 (4); 14084 
5); 18836 (4); 19104 (4); 19145 (4); 19337 (10); 19493 (10); 
20259 (10); 20401 (5); 20403 (4); 25579 (4); 28058 (7) 
28640 (2); 28846 (10); 29125 (4); 29140 (4); 29290 (9); 
29681 (9); 29941 ee 29967 (2); 32321 (4); s.n. [June 1928] 
(6); s.n. [3 Feb. 1960] (4). 

Jacquemin n» ri 0). 

Keraudren 928 (10); 987 (10); 1064 (5); 1214 (6); 24575 
7); 25160 (7); 25280 (7 

Labat 2073 (10). Lam 5439 e 6075 (4). Lavauden s.n. 
June 1928] (6 P ps zy 10). Lorence 2055 (10). 
Louvel 118 (3 AN in 4105 (4); 4226 (4). 
Ludovic 261 (3); A Y 699 (3); 722 (3). 

Mabberley 967 (10). Manjakahery 171 . McPherson 
14954 (10); 14959 (10). McWhirter 130 (10). Meyers 17 (4). 
Miege 142 (10); e a Miller 6097 (9); 6104 (10); 6180 
(10); 10687 (4); 10701 (4). Morat 1288 (7); 1442 (4); 2920 
(10); 2933 (10); um (9); 4550 (4). 

O'Connor 112 
Paulian 1 do a ii Peltier 1892 (7); 2181 (10); 2469 
M 2802 (10); 3194 (10); 4350 (7); 4477 (2). Pernet 33 (10); 


= 


— 


T 
~ 


8555 5 (9): 12471 (10); 12472 (2; 12807 (10); 13700 (13); 
14488 (13); 16324 (4); 16581 (10); 16690 (9); 18459 (2). 
Phillipson 1772 (10); 1936 (6); 1948 (4); 2676 (10); 3447 
(10); 3625 (10); 3742 (10); 4048 (7) 5929 (10). Plantes 
"a 5871 (3). Poisson 147 (4); 692 (10); 3455 (2); s.n. 


Rabehevitra 409 (3); 994 (4); 994A (4); 1186 (3). 
Rabenantoandro 598 (4); 735 (3); 987 (3); 1602 (3); 1725 
(14). Rabevohitra 4119 (3); 4455 (4); 5071 (3). Rakotoma- 
Rakotondrajaona 322 (4). Rakotonandrasana 

oza: a 1447 1025 (7) 


(10 . Randrianaivo 614 (6); 894 
10} BoA (4); 1043 (10). Pie olo, A. 143 (10); 529 
1 (4). Randriatafika 349 (4). Rasoja 1191 (7). Rauh 79 
7); 12 (10); 1229 (9); 1677 (2); 10524 (10). Razakamalala 


ZAR 
Q s 
> 
— 


9 2670 (14). Réserves Naturelles 1022 (6); 1130 (10); 1421 
(9); 1437 (3); 1437b (10); 1899 (6); 2180 (6); 2271 

7; 2312 (4); 2777 (10); 3055 (7); 3818 (4); 3930 (10); 4028 
7); 4234 (6); 4285 (10); 5183 (10); 5393 (4); 5742 (4); 6357 
4 [= Service Forestier 17671] (3); 9701 

6); 10361 (9); 10387 (7); 11484 (7); 11557 (7); 11571 (13); 
13004 (4). Richard 65 (4); 580 (4). Rogers 76 (3); 77 (3); 
4); 107 (5); 108 (5); 109 (4); 110 (4); 126 (7); 133 (4); 


pru TEN 


911 (4); 914 (4); 928 (10); 929 (10); 930 un n (10); 934. 
(10); 949 (4); 954 (4); 978 (5); 979 (4); 1 

Saboureau 49 (6). Schatz 1744 (10). A 4402 (3). 
Scott- € 1962 ub an (4); 3030 (4). poe 12872 (4). 
Seligson 638 (4). e Forestier 17 (6); 6); 341 (10); 
356 (10); m u- "1950 (7); 2271 (7); ee (6); 5100 (3); 


Annals of the 
Missouri Botanical Garden 


9900 (3); 17671 [= Réserves Naturelles 8874] (3); 18572 

7 x m i rd UT 26490 (4); a O -— A 
1] (4). Straka s.n. [18 Sep. 

n s.n. T. Seine 7 (4). 


s Mt n. (8). Thouars s.n. (3). 
van Nek 1942 (4). v van der Werff 13577 (7). Vesco s.n. 
[1850] (4). Moy 1800 (7). 


Waterlot 331 (4); s.n. [June 1915] (10); s.n. [May 
1916] (10). ae s.n. "lis Sep. 1929] (10). 
Zarucchi 7572 (10). 


APPENDIX 3. List of names and synonyms (including excluded 
names) Accepted names are presented in boldface and 
synonyms are italicized. 


Arthrosolen C. A. M di 
pM (Baker) Leandri = = cals gnidioides (Baker) 
mke 


Magarin Endl. [= 
xcluded name] 
Dais gnidioides Baker = Gnidia gnidioides (Baker) Domke 
madagascariensis Lam. = Gnidia daphnifolia L. f. 
Lam. = Gnidia daphnifolia L 


= Dais glaucescens Decne. in C. A. 


Phaleria octandra (L.) Baill., 


se 
joi lia | (L F D n. = Cait: daphnifolia 
hirsuta (L. f.) Raf. [= Gnidia capitata L. f., ve axo] 
Gnidia 
ambondrombensis (Boitean) ZS.R 
bakeri Gilg — a gnidioides (Baker Domke 
os i = = Gnidia bojeriana (Decne.) Gilg 
bojerian ilg 


glabra L. f. = Gnidia daphnifolia L 
var. i m Dt [= Gnidia capitata L. f., dicam taxon] 


Drake 
gnidioides (Baker) Domke 
hibbertioides (S: Moore 4 m E ded 
Elliot) G 
humbertii (Leandri) Z. S. = 
linearis (Leand ri) Z . 5. Rogers 
dec riensis (Lam .) Gilg = Gnidia daphnifolia L. f. 
seule cta .R 
ir den (Leandri) Z. S. Rogers 
perrieri a dri) Z. S. Rogers 
pubescens Baill. — = Gni idis daphnifolia L. f. 
ers 


dui Ric S. Roge 


rostrata (Meisn.) Drake = Gnidia daphnifolia L. f. 
end Fresen. — Gnidia L 
mbondrombense d cR - Gnidia ambondrombensis 
(Boi teau) Z. S. R 
baronii Baker = ‘Gnidia daphnifolia L. f. 
bojerianus Decne. = Gnidia bojeriana (Decne.) Gilg 
carinatus (Leandri) Leandri = e S uem uU Lit 
cuneaius Decne. [excluded name] = Hue 
decaryi Leandri = Gnidia linearis se 7. al Rogers 
var. erectus Leandri = Gnidia linearis (Leandri) Z. S. 


ogers 
var. linearis Leandri = Gnidia linearis (Leandri) Z. S. 
ogers 


Roger 
var. littoralis Leandri = Gnidia linearis (Leandri) Z. S. 


ogers 
var. oe Leandri = Gnidia linearis (Leandri) Z. 
S. Roge 
2. Tea = Gnidia daphnifolia L. f. 
erecius (Leandri) Leandr = Gnidia linearis (Leandri) Z. 
S. Rogers 
2 S. Moore = Gnidia hibbertioides (S. Moore) 
Z. S. Rog 


hildebrandi Scott-Elliot = 
hu ila tii Leandri = Gni 
Ro 


Gnidia daphnifolia 
dia humbertii ua ZS: 


oger 
nios Leandri = Gnidia Büeusie (Leandri) 4 Z. = ey 
madagascariensis (Lam.) Decne. f 
var. angustifolius Leandri = Gnidia daphnifolia E f. 
var. baronii (Baker) Leandri = Gnidia daphnifolia 
LE 


var. hildebrandtii (Scott-Elliot) Leandri = Gnidia 
e a a 

va rensis Leandri = Gnidia daphnifolia L. f. 
var. parda lius Leandri = Gnidia daphnifolia L. f. 
var. rostratus (Meisn.) Leandri = Gnidia daphnifolia 


multifolius (Leandri) Leandri = Gnidia daphnifolia L. f. 
occidentalis Leandri = Gnidia occidentalis (Leandri) Z. 
ogers 
perrieri Leandri = Gnidia perrieri P Z. A bn 
pubescens nidia da 
— Gnidia d ad E a 
var. multifolius ue — Gnidia daphnifolia L. f. 
rhamnifolius Baker [= 
ey., excluded name]. 
rostratus Meisn. = ia daphnifolia 
saxatilis Scott-Elliot = Gnidia a ESA 
suffrutescens Leandri = Gnidia daphnifolia L. f. 
waierlotii Leandri = Gnidia daphnifolia L. f. 


Dais glaucescens Decne. in C. A. 


Volume 96, Number 2, pp. 215-368 of ANNALS or THE MISSOURI BOTANICAL GARDEN 
was published on July 7, 2009. 


www.mbgpress.org 


CONTENTS 


Biogeography and Phylogeny of Cardamine (Brassicaceae) 
Tor Carlsen, Walter Bleeker, Herbert Hurka, Reidar Elven & Christian Brochmann 
A Taxonomic Revision of the Syringa pubescens Complex (Oleaceae) 


Chen Jin-Yong, Zhang Zuo-Shuang & Hong De-Yuan 
Robert William Cruden 


Three New Species and a Nomenclatural Synopsis of Urera (Urticaceae) from Mesoamerica 


A Synopsis of South American Echeandia (Anthericaceae) _ 


Alexandre K. Monro & Alexander Rodriguez 


A Review of the Genus Distictella (Bignoniaceae) Amy Pool 
Zachary S. Rogers 


A Revision of Malagasy Gnidia (Thymelaeaceae, Thymelaeoideae) 


Cover illustration. Galianthe longifolia (Standl.) E. L. Cabral, drawn by Laura Simón. 


I po 


20 


Volume 96 
Number 3 


Annals of the 
Missouri Botanical Garden 


Volume 96, Number 3 
September 2009 


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, in- 
dependent reviewers. 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 
Allison M. Brock 
Associate Editor, 

Missouri Botanical Garden 
Tammy Charron 

Editorial Assistant, 
Missouri Botanical Garden 
Cirri Moran 

Press Coordinator, 
Missouri Botanical Garden 


Roy E. Gereau 

Latin Editor, 

Missouri Botanical Garden 
Thsan A. Al-Shehbaz 
Missouri Botanical Garden 
Gerrit Davidse 

Missouri Botanical Garden 
Peter Goldblatt 

Missouri Botanical Garden 
Gordon McPherson 
Missourt Botanical Garden 
Charlotte Taylor 

Missouri Botanical Garden 
Henk van der Werff 


Missouri Botanical Garden 


For subscription information contact ANNALS 
or tHE Missouri BoranicaL GARDEN, 96 Allen Mar- 
keting & Management, P.O. Box 1897, Lawrence, 
KS 66044-8897. Subscription price for 2009 is 
$175 per volume U.S., $185 Canada & Mexico, 
$210 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 


Tue ANNAIS 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 Mtssourt BOTANICAL GARDEN, %o 
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- 
stract/Global Health databases, ingenta, ISI) databases, JSTOR, Research Alert, and Sci Search® 
The full-text of ANNALS or THE Missouri BOTANICAL GARDEN is available online though BioOne™ (http:// 
www.bioone.org). 
© Missouri Botanical Garden Press 2009 
The mission of the Missouri Botanical Garden is to discover and share knowledge about plants and 
their environment, in order lo preserve and enrich life. 

© This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 


Volume 96 
Number 3 
2009 


Missouri 
Botanical 


Garden 


BIODIVERSITY AND 
CONSERVATION IN THE ANDES: 
INTRODUCTION? 


Peter Moller Jørgensen? 


e Andes presents us with long list 
outstanding features. The adjectives, phe wm 
ly qualified, that have been used to describe this 
mountain range are: longest, highest, deepest, rough- 
est, flattest, steepest, wettest, driest, warmest, coldest, 
richest, poorest, youngest, and oldest. How do we 
present the biodiversity and conservation of such a 
varied region in a single day? 

The 54th Annual Systematics Symposium of the 
Missouri Botanical Garden, “Biodiversity and Con- 
servation in the Andes,” made an attempt on 12-13 
Octobe 07; however, with the nine 
publi Jed here as t 
were only able to aeh the surface of the Andes. An 


papers 
symposium proceedings, we 


understanding of the critical questions—what species 


scale and complexity that biodiversity displays in the 


ndes; these are the same qualities that make the area 
so fascinating. We present here a series of interesting 
papers that will serve as inspiration for the develop- 
ment of further hypotheses, questions, and research on 
the subject of biodiversity and conservation in the 
Andes, which in turn, we hope, will lead to the 
implementation of successful conservation plans in 
the area. 

The Andes has an extremely varied landscape with 
a diversity of geological features that developed in 
slow motion or rapid jumps over millions of years, as 
lan Grah 


arden) in the first morning lecture. On a geo 


presented by am (Missouri Botanical 
ogical 
timescale, the diverse and highly endemic páramo 
vegetation came into existence only recently (3.5 
million years ago [Ma]. The combination of slow lift 
and violent voleanism and, consequently, young and 
old substrates often in close proximity, as well as the 


‘This and the following nine papers are the proceedings of ihe 54th Annual Systematics Da ae of the Missouri 
the Andes 


Botanical Garden, “Biodiversity and Conservation in 


—13 October 2007 at the 


The symposium was held 1 


den in St. Tode Missouri, U.S.A. Peter Jorgensen pokes a editor a the ate 8. 
?'This was the 52nd Missouri Botanical Garden Annual Systema 


a grant from the National 


Science Foundation (grant DEB-0515933). Many persons from the 


and memorable day. Peter 
responsible for the program 
s ed with registration; Zubin Ch 

omputers and projectors; Mateo Unda 


put together the coffee break slide shows; Barbara Alon 


Missouri aren Garden pe contributed to a successful 


gi made the program 


laos auc Vie toria e Hollewell, Beth Parada, Allison Brock, and Tammy Charron (Missouri Botanical Garden Press) 
the y proceedings. I also want to thank the speakers and reviewers for making 
this series of most interesting papers available—an important step in reaching a better understanding and ultimately a better 


g oe of the biodiversity of the Andes. 


ouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A. peter.jorgensen@mobot.org. 


Mis 
doi: 10. 3417/2009065 


ANN. Missouri Bor. Garp. 96: 369—370. PUBLISHED ON 28 SEPTEMBER 2009. 


Annals of the 
Missouri Botanical Garden 


separations that are created by mountains and valleys, 
makes for a dynamic landscape on many scales and 
along numerous gradients that are immensely impor- 
tant for our understanding of biodiversity and 
conservation. 

The Andes have created the driest (the Atacama 
Desert) and wettest (the Chocó rainforests) places on 
earth. The Atacama has been 

million years, while the Altiplano, just east of it, has 

ycles 


dry for more than 10 
ndergone cycle eriods as Christa 
Plaezek (Purdue University) and un collaborators 
indicated in the second talk. It is fascinating that it 
never rains in northern Chile, while the salt flats of 
the nearby id have had a fluctuating precip- 
itation regime for thousands of years linked to the El 
p nodus Oscillation (ENSO) p 
Ou hyl 


ogeny and distribution, 


henomenon. 


Bae in d Andes, is limited for many clades. 
One exception is the 
species to potentially reach a general notion 
reflects both mechanistic ecological and historical 
Jon 
Copenhagen) and collaborator did just that 


(University of 
in the last 
morning session. They concluded that the upper forest 


evolutionary theory Fjeldsá 


limit plays a very important role in the diversification 
process, while small climatically stable areas near o 
human settlements have accumulated particularly 
high numbers of endemic birds. 

Valleys, rivers, and ridges—are they all equally 
important as barriers for the distribution, isolation, 
and evolution of species? We include in these 
by Jason Weir 


proceedings an invited paper 


Dd. of a who was not able to 
participat symposium. In his 


tion units, using genetic and netic e 
inf d 


ylogen 
on both inter- and infra-specifie variation. Lowlan 
barriers were found to display the highest nae of 
genetic differentiation, whereas above-tree-line bar- 
riers show less variation. Conservation recommenda- 
tions were to place special emphasis on endemism 
areas separated by lowland barriers and deep 
interandean river valleys 

res float on the air and can cross barriers 
with an ease unlike vascular plants and even birds; 
they therefore are not governed by the same rules. 
Steven Churchi 
the second invited paper about moss diversity in the 


ill (Missouri Botanical Garden) wrote 


Andes. The group has traditionally been overde- 


scribed, and many names have recently been reduced 


to synonymy. Endemism is still high and total richness 


wid to be greatest at elevations of 2500—3000 m. 

gen mosses contradict the usual negative 
ur x between latitude and species richness, 
but the narrow strip of Andean highlands and cloud 
forests harbors enough species to counteract the 
general tendency. 

The Andean landscape has a nearly infinite amount 
of spatial barriers and provides ample M cca 
for allopatric and sympatric speciation, but the 
barriers do not work with the sa 
and space or for different taxonomic groups. 


me efficiency in time 
hu pee pollinated D group is the subject of 
the paper by Lena Stru 

iod. Using Mine developed software called 


e (Rutgers University) and 


Spatial a and Ecological Vicariance Anal- 
ysis (SE 
of t 


VA), the md history and geography 

e group are tea 
Why are there so many species in the tropics? This 
is a question that has occupied ecologists for several 
decades. Numerous hypotheses have been proposed, 
and few believe that there is a single explanation for 
the patterns that we detect. Trisha Distler (lecture 
presented by Iván Jiménez, both from the Missouri 


Botanical Garden) and collaborators tested the 


that un plant richness across 
large areas of South and Central America, using a 

large data set of collection data. 
e Andes is one of the world's cradles of 
civilization; the Tiwanaku, the Inca, and many other 
were living in well-structured | societies 


T 
ro RU the Andes long panis 
h 


efore the 
cian. Kenneth Young (University of Texas) 
discussed the impact humans ~ had on biodiversity 
in the past and will continue to have in the future, 
with a focus on climatic changes in this region where 
large areas of the landscape matrix are devoted to 
agriculture. 
onservation of biodiversity without science is 
impossible, but how is scientific knowledge used in 
conservation? Carolina Murcia and Gustavo Kattan 
outlined a process where 
communication between the ales is critical. It 


— 


Fundación EcoAndina) 


will be crucial to establish dynamic triangles of park 
managers, scientists, and conservation nongovern- 
mental organizations, to encourage their cooperation 
so that scientific results can be quickly and efficiently 
applied in the field, and to secure that we all do the 
best possible job at conserving the biodiversity of the 
Andes 


THE ANDES: A GEOLOGICAL 
OVERVIEW FROM A BIOLOGICAL 
PERSPECTIVE 


Alan Graham}? 


ABSTRACT 


tod events of biological importance in the history E Hie Andes include their impact on global climates ee an 


oe cire 
aces is rou n páram 


anes s ranging 
discontinuous (páramo Bine ts), barriers (to east-west t migrations}, e selective pathways (via | the d 


from 


tmospheric concentration of COs; 
fro m decon ouou: dens ea highly 


timing of these effects is a function of the uplift history fth 
mostly lowland swamp and fluvial environments in the Creta 


yd 


(2) d pl 
} 


Eocene (ca. 40 million years ago [Ma], 
Oligocene (ca. 30 Ma), (4) d of the bound dill 
Middle Miocene (ca. 15 Ma), and (5) uplift of the remain 


ocene. V 


and Pale 
(3) n of an offshore volcanic chain (the pus Dorian Occidental) i in the 
era wies and the Altip 
half within pene E the past 10 million year: 
appearance of a biological community S = the Atacama Desert is estimated at ca. 15 Ma, and 


páramo at ca. 3.5 Ma. Longer-term ( 


ansgaard-Oeschger [D-O] climatic events, lon dudas from the high 


1 (Younger Dryas, Medieval warm/dry period, Little Ice Age, 
lati E 


y 
throughout Latin America, including the Andes. They document a dynamic physical environment from the Cretaceous through 


the Holocene and on all timescales 
e 


ords: Altiplano, Andes, 


Cretaceous, 


Eocene, Paleocene, páramo. 


The An 
features of the Earth. They are the longest mountain 
chain at nearly 9000 km and at the widest are 750 km 
in Bolivia. Only the Himalayas with Mount Everest 


des are one of the major physiographic 


, 2007) has modified 
global and regional climate by affecting atmospheric 
circulation and the distribution of rainfall, and by 
contributing larger or lesser amounts of COz to the 
atmosphere, correlated with the waxing and waning of 
orogeny. More than 90% of the world's seismic energy 
is released at convergent margins (Oncken et al., 
2006), and the Andean plate boundaries are a major 
outlet 
altitudinally zoned habitats from rainforest to páramo 


this energy. The Andes also provide 


to glaciated peaks, an east-west differentiation into 
wetter and drier sides through rain shadow effects, a 
discontinuous north-south migratory pathway for mid- 
elevation mesic biotas, an even more discontinuous 
route for higher Andean and páramo elements, an 
east-west pathway for interchange of arid elements 


through the dry Andean valleys, and an extensive 


range of edaphie diversity. Their rise during the 
Cretaceous and particularly in the Cenozoic also 
constituted a gradual but ultimately formidable 
vicariant event. The influence of altitudinal gradients 
on distribution patterns and sie processes is 
c t a first step 
of the 


recognize two categories of physical changes that 


plex hieving a better 
Ru E NE ae ship has been to 
occur with differences in altitude. One is altitude- 
specific, such as 
and clear-sky turbidity (cloudiness). The other is more 


atmospheric pressure, temperature, 
variable but, nonetheless, affects the first category: 
moisture, hours of sunshine, wind, season length, and 
edaphic factors. The distinction will be important 
eventually for more precisely modeling the effect of 
altitude on biotas, for example, like those on the 
Altiplano (Kórner, 2007). Uplift of a feature of this 
magnitude is also affecting the CO concentration of 
the atmosphere, and hence climates, by increased 
che 


iis where, for example, wollastonite (CaSiO3) is 


erosion of silicate rocks through mical transfor- 

nsformed to calcium bei 3) and silicon 
ic s (SiOz) while capturing COz from the atmo- 
sphere. The process runs in the direction of SiO; 


1 The author gratefully acknowledges Peter Raven and the organizers of the symposium, Olga Mar tha Montiel and Peter 


ry Hooghiem 


Peter Jorgensen, and 


Anthony Orme. 


ouri Botanical Garden, P.O. Box 299.. St. Louis, Missouri 63166-0200, U.S.A. alan.graham@mobot.org. 


Mis 
doi: 10. 3417/2007146 


ANN. Missouri Bor. Garp. 96: 371—385. PUBLISHED ON 28 SEPTEMBER 2009. 


Annals of the 
Missouri Botanical Garden 


80700"W 


60°0'0"W 


40"00'W 


COCOS PLATE 


0%0'0"- 


GULF OF 7 
GUAYAQUILSy 


BOLIVIAN 


20°0'0"S4 OROCLINE 
NAZCA PLATE 
CERRO 
ACONCAGUA 
CONCEPCION 
40?0'0"$4 
CHILOÉ 1. E 
47 S 
GULF OF 
PENAS 
ANTARCTICA 
PLATE 


OCA js LT SYSTEM: . 


0 500 1,000 2,000 
EE EE Eum Kilometers 


55° S 
SCOTIA SEA 


Physiographic map of South America with place mme ered i in n text n the U.S. Geological Survey 
g>). 


Figure 1. 
Sau: Radar Topographic Mission 90 m Digital Elevation Data (<http:/ 


during periods of mountain building and erosion, 
while in periods of metamorphism the process runs in 
the direction of CaSiOs and other silicate rocks. That 
is, more silicate rock is eroded at a faster pace with 
increased mountain building, resulting in an en- 
hanced drawdown of CO, and augmenting the trend 
toward cooler climates of the Late Cenozoic 

he continuous curve of the Andes extending along 
the entire length of western South America belies the 
heterogeneity of the system. An early appreciation of 


the complexity and an overview of the processes 
involved in their formation, ushering in the modern 
era of geologic studies, were provided by Dewey and 
Bird (1970). S 


that various segments of the Andes differ in time of 


ubsequent investigations documented 


origin, the type and intensity of forces acting on them 
during various intervals, differential response to these 
forces depending on the thickness and composition of 
the rocks, i 
availability of lubricating sediments delivered into 


the angle and rate of subduction, 


Volume 96, Number 3 
2009 


Graham 
The Andes: A Geological Overview 


the subduction zone (a function of climate), and the 
variable thickness of the underlying lithosphere. 
Gansser (1973) suggested the presently recognized 
subdivision of the mountains into the southern, 
central, and northern Andes, although some authors 
recognize seven sections (Moores & Twiss, 1995). 
These are convenient geographical separations, but 
they also reflect the varied tectonic forces that have 
defined the system (Fig. 


SEGMENTS AND FORCES 
THE SOUTHERN ANDES 


The southern Andes extend from the islands off 
southern South America in the Scotia Sea at 55°S to 
the Gulf of Penas at 47^8. lle are the lowest in 
their p 
arly Mesozoic ao is “slightly older y 


elevation at just o 
Paleozo oic 
the mountains to the north, reflecting to some extent 
the opening of the Atlantic Ocean along the Mid- 
Atlantic Ridge. The opening began in the south at ca. 
120 million years ago (M 
until South America separated from at ca. 
90 Ma at the latitude of eastern Brazil ds ca. 07S 
and 1078. Approximately l km of uplift in the 
southern. Ande 

(Blisniuk et al., 
adjacent A a were also 


a) and progressed northward 
Afric 


the earliest to 
be glaciated. Tills (unsorted glacial sediments) occur 
interbedded with basalts of Miocene to Quaternary 
age, and Mercer et al. (1975) found evidence of 
glaciation by 3. 

he principal tectonic forces in the southern Andes 
are generated by the Antarctic Plate subducting 
beneath the South American Plate, but the region is 
also affected by jostling at the triple junction between 
the Nazca, South American, and Antarctic plates, and 
it is influenced to some extent by the Scotia Plate. 
Fragmented lands were accreted to the coast, and 
subduction of the Nazca Plate, mostly in the Cenozoic 
and ^ dd between 50 and 42 Ma and 25 and 
1 2000; prior to the 14-10 Ma 
uh of the Chile Ridge with the subduction zone 
[Blisniuk et al., 2006], has acted to pull the coast 
downward. Relative sea-level rise at various times, 


a (Suárez et al., 


including the present interglacial, has contributed 
further to defining the archipelago extending south- 
from Chiloé Island (42°S 


suggests that the Late Cenozoic era in the southern 


ward ecent research 
Andes has witnessed 18 to 20 glacial advances and 
retreats, under the influence of the 100,000-year 
eccentricity cycle of the Milankovitch variations, over 
he ew million years (Giles, 2005). There are 


also variations on a finer timescale (see later 


discussion). This view of climate change, under a 
a Antarctic bu (high- ee 
system cov 


double 


zones northward by 10° latitude, with accompanying 


ring the glacial s 
"n in Hm laa and shifting e 


fluctuations in sea level and water tables, along with 
tectonic events, reveal a dynamic physical environ- 
of the southern Andes region, especially in the 
Late Cenozoic. 

The biotic response has been equally dynamic as 
recorded in numerous spore and pollen diagrams from 
southern Chile and Argentina (e.g., Heusser, 2003; 
see summary in Graham, 2009). Nothofagus Blume 
pollen is present throughout most profiles associated 
with herbs, especially grasses, and there was tundra 
Sphagnum L., Poaceae, Acaena Mutis ex L., Aster- 
aceae, Caryophyllaceae, Cyperaceae, Empetrum L.) in 
the terminal Pleistocene. A grassy steppe developed 
in the early Holocene (Chiliotrichum Cass., Festuca 
L., Mulinum Pers., Stipa L., Verbena L.), followed by 
Nothofagus (14,000 years BP) and closed Nothofagus 
forest (10,000 years BP). In the coldest intervals, 
temperatures are estimated to have been at least 5"C— 
6°C colder than at present in the lowlands and 
progressively colder in the highlands, assuming a 
km. The 
earliest occupation of Tierra del Fuego by people was 

9 


minimal global average lapse rate of 6°C/ 


ca. 13,000 years ago (Coronato et al., 
suggested by an increase in charcoal in the 
stratigraphic records at about this time. The impact 
of these people on the vegetation, other than by fire, 
was minimal because they were primarily hunters of 
Lama guanicoe (Miotti & Salemme, 1999). 


THE CENTRAL ANDES 


The central Andes is the longest segment of the 
mountain range and comprises 5200 km of the 9000- 
km length. It is the widest at 750 km and the highest 
at 3.5—4 km, including Cerro Aconcagua at 6962 m. 
The central Andes extend from the Gulf of Penas 
northward to around a megashear zone called the 
Amotope Cross at ca. 2^8 near the Gulf of Guayaquil. 

Uplift of the central Andes during the Cenozoic is 
direetly related to the subduction of the Nazca Plate. 
This plate can be subdivided into as many as seven 
smaller plates, each exerting related but separate 
subduction 
different histories. 


mechanies and each with somewhat 


Bolivian section is well known due to two unrelated 
circumstances. One circumstance is military, dating 
from a war between Bolivia and Chile (1879-1884) 
and subsequent border disputes that continue to the 
relating to the 


present; the other is economic, 


Annals of the 


374 


Missouri Botanical Garden 


Aysan PIOJXO Jo uorssruurod ym posn (P's (zz 'sSu :q'o0g ‘ZI “SIF ose 998 ‘TI “SIE 1007) ouuo wotg 


'Sopuy oy} pue eonoury ymog uo Sunoe sojye[d pedrouuq 


j \ 
Binsuiuag 913918u 
eur iL id ONE Uy Js (AAs) UOHOWY ABI Jos 
J 1 v 7 d 9 I 1 9 g jods jou euun? ep ue]st1] wos Aeme 
39diH VILOOS 1Sv3 l V UIOS jo YORI 
sjods joy pajsjag  €9 
auoz jeous [e19j81 SR 
V OZ 5 d les 
e BIW Id EAM 
ee HOIMONYS -< ^ — SEUMEJI de a 
Modo) 
elBld El 9p Ol 0 zepueulay ven 
= 
[—.0€ Beak 208 | 
U02109 HE 
SeJy zin], ^ TA 
P 
[72] 
y 
uoisusosy 
oe "og " opuewas@ 3 d V 7 d 
se SuoZ 8422; ayoueuwoy 
o Ea] S/Ned 1S o-| 
ES 
—» OldiOVd 
VIA Nv3aaiuvo 
A 
> y 
00 a 09 «06 Eu os — 
\ | | L L Ì 


“PlOFXO eso 


‘| om3 


Volume 96, Number 3 
2009 


Graham 
The Andes: A Geological Overview 


extensive deposits of copper, gold, iron, silver, and 
tin, and the current exploration for oil in the eastern 
lowlands of Bolivia. These activities have generated 
topographic and geologic maps, aerial photographs, 
seismic profiles, satellite images, and collaborative 
projects between 
scale) 
and other organizations. A digital rus map of the 
perd and the Cordillera Decenal” is available at 
ile/05-404/ 


0- er.usg; Tile/ 


Mee 
ae faq.html>, ME the economie resources are 
discussed in a U.S. Geological Survey and Servicio 
Geológico de Bolivia (1992) publication with an 
i i ng geologic I du d Andes ede 
(<h I>) 


and the Deformation Biscewsee | in the Andes p a 
collaborative research program of did German 
universities and institutes, generate and track i 
eology. Lamb (2004) has dod field 
experiences involved in Lado these data in Devil in 
the Mountain: A Search for the Origin of the Andes. 

In Bolivia, the mountains consist of three morpho- 
tectonic units: the Cordillera Occidental, the 250-km- 
wide Altiplano, and the Cordillera Oriental (Fig. 3), in 
addition to undulating eastern slopes of the 
Cordillera Oriental called the Yungas. Subduction of 
the Nazca Plate into the Peru-Chile trench has varied 
from relatively rapid and steeply angled at ca. 35° 


tion on Andean g 


prior to the Pliocene, generating volcanism, to a 
slower, lower-angled, flat-slab subduction of ca. 15° 
after 5 Ma with less volcanism. The angle of 
subduction also varies spatially along the Andes from 
steep-slab subduction in northern Peru and Ecuador 
to shallow flat-slab subduction in southern Peru and 
hile. This helps explain the 
volcanic and nonvolcanie zones in the present Andes. 


spatial occurrence of 


As the subducting plate reaches a depth of ca. 
200 km, i 


magma reaches the surface as lava through cracks, 


the leading edge becomes molten and the 


fissures, and volcanoes. Some volcanic activity still 
continues. In January 1835, Charles Darwin saw a 
"great glare of light” from the eruption of Corcovado 
Volcano in southern Chile and also observed an 
earthquake at Valdivia, Chile, that elevated the crust 
3 m in a matter of s s (Darwin, 1845: 277). In 
May 1960, the [e phar e ever iecorded in 
South America occurred near Concepción with a 
magnitude of 9.5 (Richter scale). The last eruption of 
Llaima was on 2 January 2008 (Associated Press, 
2008) 


In addition to subduction, which accounts for 
about one fifth of the height of the Andes, another 
force is compression generated by westward move- 
Mid-Atlantic Ridge with the 


western edge of the continent encountering resistance 


ment away from the 


against the Nazca Plate and where it encounters 
eastward-directed force from subduction. Compres- 
sion accounts for approximately four fifths of the 
height of the central Andes. A third factor is a 
thinning of the underlying lithosphere, as in the 
vicinity of Cerro Aconcagua, which produces heating, 
weakening, and buckling of the crust to exceptionally 
high elevations. This thinning often involves the 
sudden detachment of unstable lower lithosphere and 
results in a punctuated history of surfa 
(Garzione et al., 2008). In other p 
thickening is important, for ae beneath the 
i where the lithosphere exceeds 100 km, and 
di in e contribute to the 
diem ee along the 
Once substantial heights el is attained, other 
mechanisms are necessary to sustain the elevations. 
If compression was the only force operating on the 
central Andes, it is estimated they would average ca. 
2 km in altitude rather than the actual 4 km. Lamb 
and Davis (2003) have proposed an explanation that 
involves changes in climate. The rate of slippage of a 
plate into a subduction zone depends on the amount 
of eroded sediments brought in by rivers and that act 
a lubricant. If these sediments are absent or 
cae oe reduced river flow, subduction 
on builds up, and high 
mes “The e id of the cold Humboldt 


slows, altitudes are 
Current, sometime after 30 Ma (opening of the Drake 

oen in combination with altitudes attained by 
the central Andes that were sufficient t 
ius to the west at ca. 15 Ma, caused prr along 


ast a rain 


the coast and minimized transport of sediment into 
the subduction zone. Compression from the westward 
drift of South America, 
friction, accounts for ca. ui km of crustal shorten- 


combined with greater 


ing in the central Andes during the past ca. 35 
million years (Sobolev & ee 2005). Compara- 


y and the southern Patagonian 
Andes (Blisniuk et al., 2006). Isostacy (rebound of 
the crust from removal of overlying material by 
erosion) is another factor. 

In the Middle Cretaceous, about 100 Ma, the 
western border of South America in the vicinity of 
the cent la ak 
periodically inundated by marine waters but without 


tral Andes comprised swampland and lakes 


substantial highlands. This is shown by the horizontal 
Cretaceous strata, the paucity of eroded terrigenous 
sediments, and the presence of shallow-water lime- 
stone, including lime and mud with great concentra- 
tions of dinosaur footprints. Nearby to the west was a 
voleanic island are contributing lava and ash to the 
Cretaceous sediments. 


Annals of the 
Missouri Botanical Garden 


Sw 


ww | 


Cordillera Occidental 


Altiplano. ————————— 
Rio us nd 


Magmatic Rocks 


[vvv] Rhyolite Formation (Pliocene) 


« Granodiorite” (Miocene) 


Precambrian 


. Sectional view of the Bolivian Andes showing approximate elevations, faults, rock types, a 
Cordillera. Note the co rrugated topography 


Fig 
Western [em Altiplano, and Eastern 


Cordillera Quimsa Cruz 


Sedimentary Rocks 
bog Late Paleozoic 
E Early Paleozoic 


Zona Subandina 
Cordillera Oriental 
Llanuras Beni 


HH Río Cotacajes Cord. de 
~ d d e um Río Sécure 6000m 
Pto. Marguez| 
4 1 M 
yy E am Z | [3000 m 
V 0m 
T T T 
n A 62° — sp? 
E Qe MON 10? 
N " 
horizontal to vertical 1:5 E i ms 
a w- 
y Profiles 4 
oJ Quaternary A € 
[2 yer 
=.y 
Tertiary A ( 
E V y E c SN 22° 
Cretaceous . 
W. BARTH 1971 
and ages for the 


of the eastern slopes (Yungas) 


designated Zona Subandina on the figure. From Zeil (1979); used with permission of Borntraeger, Ber 


Around 40 Ma (Late Eocene), terrigenous sand and 
silt 
rapid emerging land 


mp, indicating relatively 
n of the 
Potoco Formation that contains a palynoflora of low- 
altitude terrestrial plants (Horton et al., 2001). Part of 
the reason for the uplift ca. 40 Ma is that m 
f the Naze 


possibly due in part to ad p the 


began covering the swa 
. One result was depositio 


a Plate slowed from ca. 15 t 


intensifying collision of India with Asia at this time 
(Lamb, 2004: 2 
cooled and hardened, friction bu 


5). As subduction slowed, sediments 
ilt up, 
compression forces resulted in the creation of 
highlands. This uplift of the central Andes ca. 


and continuing 


a was a biologically important event providing 
rising uplands where there had been coastal swamps 
earlier. 

By 30 Ma, segments of the line of voleanoes had 
ecome appressed to the continental margin, with 
associated uplift and crustal shortening, notably in the 
mostly volcanic proto—Cordillera Occidental of south- 
ern Peru and Bolivia. The sediments continued to 
accumulate in the intervening lowlands, which would 
eventually be elevated as the Bolivian and Peruvian 
Altiplano. At this time, the Altiplano was a river basin 
near sea level. Subduction, compression, and appres- 
sion forces continued, and by ca. 15 Ma (Middle 
Miocene) approximately half (1500-2000 km) of the 
present m average altitude of the central Andes 
had been attained. Uplift of the Cordillera Oriental 
also occurred but involved an additional source of 
compression. As the Brazilian Shield, the original 
Precambrian craton around which the continent was 
built, shifted westward subsurface, it underthrust the 


Cordillera Oriental. This caused crustal o of 

- Cordillera Oriental of more than 100 km between 
d a, n 100 km between 10 Ma 

a the present, as these mountains rode up and over 

the western edge of the Brazilian Shield. This 

high altitudes an 

corrugated topography of the eastern slopes 

213): 


Several paleobotanical studies have contributed to 


and more tha 


contributed to the distinctive 
Zon. 


Subandina on Fig 


econstructing the paleoenvironment, vegetation, and 
ndes. One is from 
Pislepampa (17°11’S), 20 km to the northeast of 
Cochabamba in the Cordillera Oriental of Bolivia at 
an elevation of 3600 m (Graham et al., 2001). The age 
of the flora is 6-7 Ma. Amon. nomorphs are 
spores and pollen grains of Isoetes L., Lycopodium L., 
Cnemidaria C. Presl, Cyathea Sm., Hymenophyllum 
Sm., Pteri 
Cavalli Ruiz 
Ber 
a leaves characteristic of lowland rainforest 
Killeen et al, 1993) The 
microfossils also include pollen of Pons Phil. 

f. Oreopa. ne. & Planch., 
grow together in the “cloud forest ne húmedo; 
Fig. 4). The plant fossil evidence indicates deposition 


g the pal 


L., Danaea Sm., a 
av. Megafossils illustrated by 
22) are mostly large and all have entire- 


bosque amazónico; 


a 


nax which presently 


near the contact between rainforest and the lower 
This 
implies an uplift of ca. 2400 m since deposition of the 
Pislepampa flora ca. 6—7 Ma. Other studies on the 


limits of the cloud forest, which is at ca. 


ology of the central Andes 
1998; G 
2001; Ghosh et al., 


paleobotany and ge 
(Gregory-Wodzicki et al., 
2000; Montgomery et al., 


regory-Wodzicki, 
2006; 


Volume 96, Number 3 
2009 


Graham 377 
The Andes: A Geological Overview 


igure 4. Cloud forest vegetation near Comarapa, Bolivia. The large palmate-leaved plant in the center is O and 
the tall plant to the far right is Prumnopitys. The pollen grains are ces (left) and Prumnopitys/Podocarpus L'Hér 
Pers. (right) recovered from the Pislepampa flora (6—7 million years old). 


Rowley & Garzione, 2007) suggest an uplift of one 
third to half the present altitude after 10 Ma; that is, a 
2000 m to the present 4000 m. The 
highest elevation would 


rise from ca. 
e the latest achieved, 
meaning that páramo habitats are the youngest in 
the Andean system. The latest cooling ushering in the 
Pleistocene glaciations began in the Late Pliocene, so 
the combination of maximum altitudes and globally 
cooling temperatures date the beginning of páramo 
vegetation to ca. 3—5 Ma, making it the youngest of 
the Latin American natural ecosystems and affording 
opportunities for rapid island-like radiation of mid- 
altitude species (e.g., Lupinus L.; Hughes & Eastwood, 
2006). Ribas et al. (2007: 2399) attribute existing 


specifically to three lineages "transported passively to 
high elevations by mountain building, and that 
subsequent diversification within the Andes was 
driven primarily by Pleistocene climatic oscillations 
and their large-scale effects on habitat change 

The establishment of a general chronology for the 
history of the central Andes makes it possible to 
address two other questions of biological interest. One 
question is when did uplift reach a point that it 
constituted a major vicariant event? This obviously 
with the 


varies different biotic lineages, their 


ecological parameters, and their distribution poten- 
tials, although after the Middle Miocene at ca. 15 Ma 
an increasing number of species to the east and west 
of the mountains would find it increasingly difficult to 
interchange. 

nother question is when did altitudes of the 
central Andes of northern Chile, Bolivia, and Peru 
reach an elevation that rain shadows began forming to 
the west and contributing to the formation of the 
Atacama Desert? The gradual separation of Antarctica 
from South America through the Drake Passage took 
place between 50 and 30 Ma (Livermore et al., 2005). 
This event isolated Antarctica from warm marine 
waters and contributed to the expansion of glaciers 


cooled at about this time, and winds blowing across 


the water lost moisture and arrived on the western 
coast as a dry and desiccating wind. A second factor 
was the barrier to moisture from the east that also 
developed in the Middle Miocene at ca. 15 Ma as 
revealed by arid lacustrine (lake) sediments in the 
hile (Sáez et a 


The Middle Miocene was an important finie in the 


foreare regions of northern 


modernization of the Earth's plant communities 
because Late Cenozoic cooling had reached a stage 
when water evaporating from the ocean surface 
became reduced, with an increase in seasonality and 


Annals of the 
Missouri Botanical Garden 


the spread of grasslands and dry forests (Graham, 
1999: chapter 7). In western South America ca. 
15 Ma, the combination of dry westerly winds and rain 
shadow allowed elements pre-adapted to aridity from 
slope, exposure, and edaphic conditions, as well as 
from immigration, to coalesce into the beginnings of 
an ecosystem recognizable as desert. 

Among the myriad features of the Andes relevant to 
the biology of the region is the Bolivian Orocline or 
“elbow of the Andes" at ca. 18°S. Killeen et al. (2007) 
suggested that this bend presents a face more at a 
right angle than other sections to winds coming off the 
Amazon Basin and augments rainfall on the already 
wet lower slopes. If so, this may have perpetuated wet 
conditions during dry intervals of the Late Cenozoic 
and served as a refugium for rainforest taxa (see 
discussion in Graham, 2009: chapter 8). Another 
physical feature of biological interest is the east- 
facing concave, almost semicircular shape of the 
Andes toward the Amazon between ca. 8°S and 10^N. 
This feature traps humidity and further augments wet 
conditions on the slopes and in the lowlands. 

ocus of current environmental and vegetation 
history studies in the central Andes is on the 
Quaternary period of the Altiplano. In the northern 
Altiplano around Lake Titicaca (Baker, 2001a, b; 
Paduano et al., i 
were estimated to be 3*C—5^C cooler than at present 
until 21 thousand years (Kyr) BP (around the late 
glacial maximum) and the snow line was ca. 500 m 
below the current 4850 m. Then 
began with higher lake levels and wet conditions until 


003; Tapia et al., 2003), temperatures 


a warming trend 


ca. 10 Kyr BP, followed by a e warm dry 
period (9000-3100 years BP, 
7960 and 3100 Kyr BP) when 
by 100 m; salinity, charcoal, and dust levels in- 
creased; and temperatures were an estimated 1°C-—3°C 


especially between 
water esi dropped 


warmer than at present. At the Salar de Uyuni on the 
Altiplano, moisture may have varied with Heinrich 
events 1 and 2 and with the Younger Dryas cold event 
between 13 and 11.5 Kyr (Baker et al., 2001a; see 
below). Lakes started to rise at ca. 4000 years BP and 
reached modern levels at ca. 1500 years BP (Abbott et 
al., 19972, b). 

In the southern Altiplano, at xs o limits of the 
Atacama Desert, Placzek et al. 
53 new uranium-thori 
that provide a more e chronology for climate 
change. The basins or salars are now mostly dry, and 
arren and rocky landscape of lava, 
shifting dunes, and some of the largest salt flats on 
Earth. However, between 120 and 98 Kyr BP there 
were s m deep on the Altiplano. r: 
intermittently shallowed between 98 and 
BP, then at 18.1-14.1 Kyr BP, a lake ca. 140 m Ded 


the region is 


developed that was the largest and deepest of the past 
120 Kyr (Placzek et al., 2006). Subsequently, there 
were smaller changes on a finer timescale that reflect 
climatie fluctuations of the Holocene. The m 
intervals and higher lake levels correspond to periods 
of human occupation on the Altiplano, whereas during 
the driest intervals e g., 9500-4500 years the 
sites were abandoned (Grosjean, 1994; Grosjean & 
Núñez A., 1994; Nites et al., 2002; ee et al., 
2005). 


THE NORTHERN ANDES 


The northern Andes extend from 5°S at about the 
Amotape Cross at 2°S to the Oca, Romeral, and other 
fault systems across northern South America at the 
aribbean Plate at ca. 12°N (Fig. 5). 
The Andes o in southwestern Colombia to s 
the Cordilleras Occidental, Central, and Oriental 
rme, 2007a: fig. e Cordillera Oriental 

a Nevada de 


contact with th 


© 


branches again into the western Sierr: 

Santa Marta and Sierra de Perijá m the eastern 
Cordillera de s the two branches enclosing 
ke Mar e latter is a region of major 
nee reserves p like the mineral deposits of 
the Altiplano, it has been the impetus for numerous 
A paleoenvironmental, and vegetational his- 

tudies (Germeraad et al., 1968; Jaramillo & 
nie 01). The And along the 


northern coast as the Caribbean Mountain system 


quine 


with portions submerged to form the Netherlands 
Antilles and other islands of the southern Caribbean 
ea. 

The northern Andes are as complex as the central 
Andes, but their uplift and other features are due to a 
different combination of tectonic forces ee 
et al., 1 
restriction between the Amaz 


. In the Late Sine ste there was som 
nd the Pacific 
Ocean, but passages existed issus which the proto- 


n Basin an 


Amazon and proto-Orinoco rivers drained into the 


Pacific. During the Paleocene, marine sediments were 
still being deposited at sites in the present northern 


1997). By the Middle 


5 Ma), the mountains had reached 


central Andes e et al., 
Miocene (ca. 
sufficient height so that there was substantial warping, 
ridgin, some tilting of the Amazon Basi 

eastward, and the rivers iem to flow toward the 
Atlantie Ocean. Evidence for the timing of this event 
comes from varied sources. Van der eg s (1952), 
Hoorn (19942, b, 2006), and Hoorn et al. (1995) n 
change along the Solimóes River from west-dipping 


ote a 


cross-bedding (Marifiame unit) to east-dipping cross- 
bedding (Solimées and Pebas formations) between the 
Early and L 
appears at the mouth of the Amazon River in the 


ate Miocene. A submarine fan first 


Volume 96, Number 3 


Graham 
The Andes: A Geological Overview 


-129 


-10° 


COLOMBIA 


-8° 


749 


100 km 


72° 


Fig 
Bloc n 
Elsevier, Amsterdam 


Middle Miocene, and delta deposits of the same age 
begin to form along the coasts of the Guianas. 

The tectonic factors include subduction from the 
Nazca Plate beneath northwestern South America at 
a present rate of 3-6 em/yr., with a particularly 
complicated section at the junction with Cocos Plate. 
There is compression from the westward movement of 
the continent meen the ae shearing along the 

coast as South America moves westward 
relative to the Caribbean Plate, and appression that 
has brought volcanic island ares and terranes onto the 
northwestern coast. One of these is the Cordillera 
Occidental, which is composed of accreted terranes of 


e 5. Portion of the North Andean Block in eastern Colombia 
ma the Perija and Merida Andes, with principal structural features. From Clapperton (1993); used with permission of 


and western Venezuela, including the Santa Marta 


voleanic origin, while the Cordillera Oriental and 

ordillera Central are prim nonvolcanic moun- 
tains. pro to Cordillera Dent al was accreted 
onto the coast in the Late Cretaceous (Winsemann, 
1994) as indicated by a line of ophiolites (oceanic 
basalts caught between suturing plates). By the 
Middle to Late Miocene, the northern Andes were 
about half their present average elevation, which is 
slightly later than the southern. Andes. Uplift of the 
Cordillera Oriental to its present height occurred 
between 6 and 3 Ma, and it is estimated that the 
Sierra de Perijá has risen 11-16 mm/yr. since the 
Late Pliocene (Clapperton, 1993). 


Annals of the 
Missouri Botanical Garden 


The rise of megamountain systems has many effects 


illera Oriental of the 
northern Andes blocked de flow of the Maracaibo 
River and formed Lake Maracaibo. This event created 
conditions favorable to the formation of extensive oil 
reserves in the basin where the need for accurate 
stratigraphic correlation, zonation, and environmental 
reconstruction pons in extensive palynological 
e Internationale Petroleum 


aafs 
ado om ‘Shell Oil), 


of the first extensive survey of Tertiary vegetation from 


and the publication 
the tropical areas of South America (Colombia, 
inidad, the Gu 
work is still of systematic, usarla, 


e. Trini lanas; Germeraad et al., 
and dice value, because microfossils were 
identified according to their biological affinities rather 
than by an artificial stratigraphic nomenclature. 

Paleodrainage patterns in the Amazon Basin, and 
changes in these patterns induced in large measure by 
the rise of the Andes Mountains, are important 
because they fragment the biota into geographically 
and thereby reproductively isolated populations. 
Deltaic deposits allow tracing the meandering course 
of rivers through long intervals of time. In the Middle 
Eocene, the Misoa paleodelta identifies a large river 
flowing north into the Maracaibo Basin that drained 
the Cordillera Central and the TH Highlands. 
Uplift of the mountains shifted delta formation 
southward as shown by the Late Eocene to Oligocene 
Carbonera Formation in the Llanos Orientales region 
of Colombia and Venezuela. By the Middle to Early 
Miocene, there was a delta in the northwestern part of 
the Falcón Basin to the east of Maracaibo, and in the 
late part of the Middle Miocene it had shifted to its 
present position in the Venezuelan Basin south of 
Trinidad and Tobago (Díaz de Gamero, 1996). These 
record the positioning of the Amazon and Orinoco 
rivers between the Late Eocene and the Middle 
Miocene, and similar changes affected all the rivers 
and tributaries in the Amazon and Orinoco basins 
during the Tertiary 

Alfred Russell Wallace (1849, 1876) used these 


strong nced by river barriers, and this has 
generally proven to be true, especially for mammals 
such as the tamarinds. For other animals, such as 
some rodents, other factors were involved, and these 
were the result of the uplift of the northern Andes. 
Patton et al. (1994; Patton & Da Silva, 1998, 2005) 
found that some populations, although morphological- 
istinct mitochondrial 


ly similar, differed in having 


DNA (mtDNA) signatures. The populations were not 


segregated on different sides of the most obvious 
barrier, the Juruá River, but rather between the upper 


from Morell, 1996: 1497) noted that “What is really 
is that all 11 [of the 17 species] are 
separated at almost the same geographical point on 


striking ... 


the river, although there is nothing remarkable about 
the spot—no bend, no hill, no valley.” The demarca- 
tion is at the ancient Iquitos Arch, which is one of 
everal arches formed during mountain uplift and that 
bunits. The 
basin floor is now covered in places by more than 
4000 m of eroded sediment that obscures the arches, 
but they were once effective barrier: 


subdivided the basin into a number of su 


s to the inter- 
change of some groups of organisms, and this history 
still resides in their molecular signature. The arches 
also affect the distribution of sediments and, thereby, 
edaphic patterns in the Amazon Basin (Dunne et al., 
1998 


Another effect of Andean uplift, combined with sea- 


(Hoorn et al., 1995), primarily from the Caribbean Sea 
through the Maracaibo Basin. Fish faunas and 
mangroves are ae m of the time and extent 


of the flooding w 


reunited. The 2 reveals several such 


e biota are fragmented and 


flooding events in the interior, persisting until the 

locene, as well as periods of deeper or 
shallower, and possibly torrential waters (Monsch, 
1998). The Amazon Basin includes the greatest 
number of freshwater fish species in the world, and 
Lovejoy et al. (1998), on the basis of the molecular 
phylogeny of stingrays, believe that this diversity 
resulted from the repeated fragmentation and reas- 
sembly of populations through intervening brackish or 
salt water. Volcanic ash deposits in eastern lowland 
Peru contain paleofaunas from 9.01 Ma and 3.12 Ma 
Campbell et al., 2001), and the latter date marks the 


last marine incursion around the margins of the basin. 


=> 


r mountainous volcanic settings, there is 


othe 
Bde of the impact of local orogenic history on 
speciation rans On Isla Isabela in the Galápagos 
Islands, five taxa of the large tortoise occur, one on 
each of the five volcanoes of the island. Geochelone 
nigra vandenburghi on Alcedo Volcano has consider- 
ably less matrilineal diversity in its mtDNA. Although 
all the volcanoes are about the same age (ca. 500,000 
years), Alcedo Volcano last erupted ca. 100,000 years 
ago. The suggestion is that the event drastically 
reduced the population diversity from which this 
particular lineage was derived (Beheregaray et al., 
imilar reductions may have occurred among 
other slow-dispersing organisms isolated in mid- to 


Volume 96, Number 3 
2009 


Graham 
The Andes: A Geological Overview 


Table 1. 


Glacial/climatic events first discovered in the Northern Hemisphere and now recognized in South America." 


Location 


Glacial/climatic event 


Reference 


Venezuela (9°N-10°N) 
Little Ice 
Colombia (4^N—5^N) Younger Dryas 
Brazil (37N—4^N) 
events, Little Ice Age 
Ecuador (3*5-4^S) Younger Dry: 
Peru (8%) 
Bolivia (15°S) 
Heinrich events 1 and 2 
Chile (Lake District; 
approximately 30°S—42°S) 
Southern Argentina—Chile 
5) 


(Patagonia; 45° Little 


Younger Dryas, mid-Holocene warm/dry period, 
Ag 
Younger Dryas, Heinrich and Dansgaard-Oeschger 


mid-Holocene warm/dry period, Little Ice Age 
Younger Dryas, mid-Holocene warm/dry period, 


Younger Dryas, Heinrich events 


mid- ae warm/dry period, Younger Dryas, 
Age 


Haug et al., 2001 


Thouret et al., 1996; van der 
Hammen & iaa 1995 
Jennerjahn et al., 2004; Cohen et al., 


2005 
Osborn et al., 1995; Heine & Heine, 
; Clapperton et al., 
Osborn et al., 
Osborn et al., 1995; Abbott et al., 
1997a, b, 2000 
Lowell et al., 1995 


Mancini et al., 2005 


Arrangement i is from north (Venezuela, 9?N—10?N) to south (southern 


°S}. Evidence for the 


A 
45 
Younger Dryas in the tropical Andes is considered equivocal by um and aa C00) and a (2007). 


higher-elevation habitats along the volcanically active 
Andes where comparable studies could be made 

ithin the mountains, there are many basins now 

filled with eroded sediments. The high plain of Bogotá 

is the floor of a drained lake. The Bogotá Basin formed 


re a 


ca. 6 Ma when the Cordillera Oriental was beginning 

iod of m. uplift. It was raised to its present 
altitude of 2 
continuous iene since then. The Bogota 


0 m between 4 an Ma, with almost 
Basin has been the site of important paleoenviron- 
mental and vegetational history studies beginning with 
those of van der Hammen (e.g., van der Hammen et al., 
1973) and continuing with those of Hooghiemstra (1984) 
and Hooghiemstra et al. (2006). Among the results has 
been the early recognition of downward shifts in high 
Andean vegetation during cold intervals that were 
suspected to have affected the lowland Neotropical 
rainforest. Later studies have shown this to be true (e.g., 
Hooghiemstra & van der Hammen, 1998; van der 
Hammen & Hooghiemstra, 2000). Another result was 
the demonstration of climatic changes and vegetation 
responses throughout the entire 3.5 million year interval 
represented by the basin sediments. Some of these 
correlate with regional to global events elsewhere (e.g., 
the Younger Dryas; see below). 

uaternary climatic and vegetation changes 
throughout the Andean chain are important for 
interpreting the regional vegetation history, but they 
have broader implications for environmental changes 
throughout South America. Many of the temperature 
and precipitation patterns of the Quaternary were 
initially documented in the northern latitudes and 
especially around the North Atlantic Ocean. In 
addition to the longer-term fluctuations of 100,000, 


41,000, and 23,000 years of the Milankovitch 
variations (with subcycles), there were warm and co 
intervals on a much finer scale. These intervals 
include the cold reversal during the terminal Pleisto- 
cene/Holocene called the Younger Dryas (13, Pa 
v e years BP), the Medieval Warm Period (80 

O CE), and the Little Ice Age (1300-1850 A 
aaah events are cold periods of a few to several 
thousand years’ duration, while Dansgaard-Oeschger 
(D-O) events occur on a scale of a few thousand to 
several hundred years. These events involve fluctua- 
tions in the thermohaline Sp dA belt) transport of 
heat in the form o waters from the 
tropies to the North Atlanti; cooling of these waters; 
and transport back to the tropics as cold Atlantic deep 
water, where it warms and returns again northward. 


gation, Mapping 
1976, 1981, 1984) and the Cooperative Holocene 
Mapping Project (COHMAP, 1988) were published, 
they suggested that while temperatures were colder by 
°C-14°C in the high northern latitudes at the Last 
Glacial Maximum, equatorial temperatures remained 
2°C. The 
work previously cited for the High Andes was difficult 


msn 


the same or cooled by a maximum of ca 


to reconcile with the climatic and the implied 
vegetation stasis of the tropics, but definitive evidence 
i . (1993; Watts 
& Hansen, 1994) found fluctuations in pine pollen 


was not yet available. Then Grimm et a 


a corresponded with Heinrich events 1 to 5 in bogs 

Lake Tulane, Florida (28^N). Guilderson et al. 
ied proposed a lowering of 5°C at the Last Glacial 
Maximum based on strontium:calcium ratios from 
corals on Barbados at a latitude ca. 12^N, and several 


Annals of the 
Missouri Botanical Garden 


studies now document a similar cooling for the 
Mg Webb et al., 


ere is now 


Brazilian lowlands (Stute et al., 
schbach-Hertig et al., 
B INO and indepen nüent Ee evidence for 
rapid temperature and precipitation changes from 
several places in the tropics (Table 1; see Graham, 
2009 for additional references). 

One reason these kinds of evidence are significant 
is that for a long time, through about the 1960s, it 
was thought that tropical climates and the tropical 
us were un a ee (e Ma . Corner, 

58). This se o be n. by th 
dona 1981, o id COHM P (19 88) results of 
the 1970s and 1980s. In turn, "i could be used to 
rationalize the view that the tropical biota could be 
Mes with impunity and would recover because 
ey had endured unchanged for millions of years. 
a was suspected not to be true (see reviews, e. B^ 


Pliocene Paraje Solo mierofossil floras in southern 
ovided direct 


evidence that in an area where rainforest is the 


ux obotanical 


dominant vegetation today, it was not present in the 
recent geologic past (Graham, 1976a, b). Subsequent 
studies have extended this finding to other commu- 
that 
ecosystems throughout the equatorial latitudes, from 


nities and other locations, so today all 
owland rainforest to the paramo of the High Andes, 
and from Brazil, 
Argentina and Chile, are recognized as delicately 
balanced assemblages that have taken millions of 
years to develop, have had a dynamic past on all 
timescales, and deserve the conservation and sus- 


tainable development efforts that fact alone implies. 


Literature Cited 


Abbott, M. Binford, M. Brenner & K. R 
1997a. A 2200 "s yr high-resolution record of water- level 
changes in Lake Titicaca, Bolivia/Peru. Quatern. Res. 47: 
169-180. 

, G. O. Seltzer, K. R. s & J. Southon. 1997b. 
elses paleo 7 d i Andes from lake 
records. Quatern. Res. 47: 70-80. 

, B. BW. R. Aravena, A. P. e&G.O 
Seltzer. 2000. Holocene hydrological E s from 
stable isotopes and paleolimnology, Cordillera Real, 
Bolivia. Quatern. Sci. Rev. 19: 1801-1820. 

ones -Hertig, W., F. Peeters, U. Ma Gia e f Pipe 


gn al water takin ng into account equilibration with 
entrapped air. en 405. 1040—1043. 

Alonso, R. N., B. Bookhagen, B. Carrapa, I. Coutand, M. 

ctonics, 


The Argentine Puna Plateau and adjacent 


Chong, G. Franz, P. Giese, H. -J. Götze, V. A. Ramos, M. 


R. Strecker & Peter Wigger (editors). The Andes: Active 
Subduction Orogeny. Frontiers in Earth Sciences. Springer 
Verlag 
POM Press. 2008. Villagers flee volcano eruption in 
Giles dne Guardian [online editio n]. E Janua ay 2008. 


, Berlin. 


chile>, "ara 13 July 2 


Baker, P. Lb. d ‘0. Seltzer, 5. C. Fritz, T. K. 
Lowen N. P. Bacher a E Veliz. 2001a. Tropical 
at millennial orbital timescales on the 


Sues Sli. Nature 400: p 701. 
, G. O. Seltzer, S. a us R. B. Fr. M. J. Grove, 
apia, S. L pne H. D. Rowe & J. P. Broda. 2001b. 
The history. of South American ol precipitation for 
the past 25,000 years. Science 291: 640—643. 
Beheregaray, L. B., C. Ciofi, 
Caccone & J. R. Powell. 2003. G 


Fp. Ne in O. Oncken, G. Chong, G. Franz, P. Class, 
, V. A. Ramos, M. R. Strecker & P. Wigger 
(dior) "The Andes: Active Subduction Orogeny Fron- 
tiers in Earth Sciences. Springer Verlag, B 
E e K E. ae aie Heizler, E De Filey, L . Romero- 
R. Prothero. 2001. er pra 
cromo of the o pa Basin. 
Geology 2 a eng 
Clapperton, C. Quaternary Geology and Geomorphol- 
of South o Elsevier, Amsterdam 
i Clayton, D. I. Benn, C. J. Marden & J. Argollo. 
1997. "mers ternary glacier advances and palaeolake 
heo: i the Bolivian Altiplano. nist Int. 38/39: 
9-59. 


Pittman 


CLIMAP Project Members. 1976. The surface of the ice-age 
Earth. Science 191: 1131-1137. 

. Seasonal reconstruction of the Earth’s surface 

at the last m maximum. Map Chart Series MC-36. 

bus Societ a, Boulder. 

The p Dee ocean. a Res. 21: 


123 zr 

Cohen, M. C. L., H. Behling & R. J. Lara. 2005. Amazonian 
mangrove dynamics during the last millennium: The 
relative sea-level and the Little Ice Age. Rev. Palaeobot. 
Palynol. 136: 93-1 
"T MAP Members. 1988. Climatic changes of the last 

years: Observations and model simulations. 

B. 241: 1043— O 


J 

6-17 August Fd Geological pod of Amena, d 

Comer, E. J. H 
6 in J. un A. C. Hardy & E. B. Ford oem 
Evolution a a Process. Allen and Unwin, London 

Coronato, A., M. Salemme & J. as ee 1999. Palas 
vonmontal- conditions during the early peopling of 
southernmost South America (Late Glacia ]— arly Holo- 
cene, 14-8 ka BP). Quatern. Int. 53/54: 7 

Darwin, C. 1845. Voyage of the Beagle m ee 
Wordsworth Editions, Ware, Hertfordshire, Unite 
Kingdom 


Volume 96, Number 3 
2009 


Graham 
The Andes: A Geological Overview 


Dewey, J. F. & J. M. Bird. 1970. Mountain "n on the new 
m an tectonies. J. Geophys. Res. 75: 2 
z de Gamero, M. L. 1996 de rasa course a the 
y 5 River during the Neogene: A review. Palaeogeogr. 
Palaeoclimatol. Pie 123: 385-402 
de, J. E. Richey & B. 


nt between the 


Garzione, C.N sG: D. Hoke, J. C. Libarkin, S. Withers, B. 
oo n, J. Eiler, P. Ghosh & A. Mulch. 2008. Rise of 
the Andes. boe ce 320: 1304—1307 

Germeraad, H., A. Hopping «d J. Muller. 1968. 
Palynology of Tertiary sediments from tropical areas. Rev. 
o Palynol. 6: 189-348. 

C. N. Garzione & J. M. Eiler. 2006. Rapid uplift of 
the Altiplano revealed "ra *®C-180 bonds in paleosol 
carbonates. Science 311: 511 

e J- fo 5. Geologists call time on dating dispute. Nature 


a. A 1976a. Studies in neotropical paleobotany. IL 
he Miocene communities of Veracruz, Mexico. Ann 
Missouri Por Gard. 63: 787-842 
———. 1976b. Late Cenozoic avalon of tropical lowland 
vegetation in Veracruz, Mexico. Evolution 29: 723—735 
. 1999. Late Cretaceous and Cenozoic p of North 
American Vegetation. O niversity Press, Oxford. 
—— - 2009, Late Cretaceous and C 
Latin Am 1 
Missouri Botanica G: 
re; ory-Wodz 


——— e & K 

e Eastern Cordillera, Bolivia: pud 

r the uplift history of the Central Andes. Amer. J. 
Bot. 88: 1545-1557. 

Gregory-Wodzicki, K. M. 2000. a between leaf 

orphology and climate, Bolivia: Implications for esti- 
mating paleoclimate hen fossil pa Paleobiology 26: 
668-688. 

———, V. C. McIntosh & K. Velásquez. 1998. Climate and 

tectonic Po ns of the late Miocene Jakokkota flora, 
Bolivian o no. J. S. Amer. Earth Sci. 11: 533—560 

a E., C: Hanseni 

A 50,000-year record of climate 
o from Florida and its temporal correlation with 
Heinrich e ue 261: 198-200. 

Grosjean, M Paleohydrology of the Laguna Lejia 
(north Chi i pe o) and climatic implications for 
late-glacial times. Palaeogeogr. Palaeoclimatol. Palaeo- 
ecol. 109: 89-100. 

— —— & L. Núñez A. 1994. Late glacial, early and middle 
Holocene environments, human occupation, and resource 
use in the Atacama a Chile). a 9: 

86 


. Cartajena. 2005. Palaeoindian 
tapaisi of the Atacama Desert, northern Chile. J. 


LGS . L. Rubenstone. 1994. 

Tropical temperature variations since 20,000 years ago: 

E. interhemispheric climate change. Science 

263: 663—665. 

UU Cr K. A. Hughen, D. M. Sigman, L. C. Peterson & 
U. Röhl. 2001. Bowen migration of the Intertropical 
Convergence Zone through the Holocene. Science 293: 


1304—1308. 


Heine, K. . T. Heine. 1996. Late glacial climate 
na in Ecuador: Glacial retreat during Younger 
ryas time. Arctic a us Res. 28: 496—501. 
Ice Age Southern Andes: A Chronicle of 
Palssestilcotedl Events. Elsevier, Amsterdam 


Hooghiemstra, H. 1984. Vegetational and climatic history of 


68. [Reprinted from Dissertationes Botanicae, 
Band 79, J. Cramer, Vaduz, S 
& A. M. Cleef, 2006. 
paleohotanieal record of Colombia: Implications i 
biogeography and biodiversity. Ann. Missouri Bot. Gar 
93: 0 PM 


Wijinga 


——— & T. Hy van dèi Hamnen. as Bu and 


ernary d n forest: The 
forest refugia ip. and a literature overview. Earth- 
Sci. R -183. 


Hoorn, C. Mob. Fluvial palaeoenvironments in the 
intracratonic Amazonas Basin (early Miocene-early middle 
Miocene, Colombia). Palaeogeogr. Palaeoclimatol. Pa- 
laeoecol. 109: 1-54. 

———. 1994b. An environmental reconstruction of the 

palaes Amazon River system (middle-late Miocene, NW 
ae. Palaeogeogr. Palaeoclimatol. Palaeoecol. 112: 
187-238. 

——. 2006. The birth of the mighty Amazon. Sci. Am. 
May: 52-59. 

———À,J. psi errero, G. A. Sarmiento & M. A. Lorente. 1995. 
Andean tectonics as a cause for changing drainage 
patterns in n Miocene South America. Geology 23: 237-240. 

Horton, B. K., B. Hampton & G. L. Waanders. 2001. 
Paleogene synorogenic sedimentation in the Altiplano 
Plateau and implications for initial mountain building 
in n x Andes. Bull. Geol. Soc. Amer. 113: 
1387- 

Hughes, E EF R. Eastwood. 2006. Island radiation on a 
continental scale: Exceptional rates of plant es 
after uplift of the Andes. Proc. Natl. Acad. U.S.A 
10334—10. 

Jaramillo, C. A. & D. L. Dilcher. 2001. Middle Paleogene 

palynology of central Colómbia, South America: A stud 

of pollen and spores from tropical ae Palaeonto- 

graphica, Abt. B, is Um 258: 87— a 
Ittekkot, Behling, J. 

Asynchronous canes and marine 
o of climate change during Heinrich events. Science 
306: 2 

Killeen, A Jis E. Garda E: & S. G. Beck (editors). 1993. Guía 
de Árboles de Bolivia. Herbario dem de Bolivia, La 
Paz; Missouri B ical Garden, St. ut 


7. The BA of ‘altitude’ in ecological research. 

se T) 2: 56 pur 

tes, S. B., . Bakker € M. Van der Wiel. 

. Late Ciok ns and paleogeography of the 
Colombian Andes: epus on the development of 
high-Andean biota. Geol. & Mijnb. 69: 279—290. 

Lamb, S. 2004. Devil in the Meanie A Search for the 
Origin of the Andes. Princeton University Press, Prince- 
ton, New Jersey 

— & P. Davis: 2003. Cenozoic climate change as a 

possible cause for the rise of the Andes. Nature 425: 

192—191. 


Annal 
Missouri Botanical Garden 


Livermore, R., A. Nankivell, G. Eagles & P. Morris. 2005. 
Paleogene Rupe: of Drake Passage. Earth Planet. Sci. 
Lett. 236: 10. 

Lovejoy, N. ^i E Bermingham & A. P. Martin. 1998. Marine 
incursion into South America. Nature 396: 421—422 

ersen, I. Moreno, A 

R. M 


Prie 

I. Vi e nova. pos Mid- “Holocen e climati ity 

reconstruction from pollen records (32°%—52°S, Pier 

t. 132: 47—59. 

. Mankinen & W. Sander. 
1915. Southern Patagonia: T events even 4 my. 

m.y. ago. Pp. 223-230 . P. Su; M. M. 
Cresswell (editors). cere ae "o Society of 
New Zealand, Wellington. 

Miotti, L. & M. Salemme. 1999. Biodiversity, taxonomic 
richness and Epa : during late Pleistocene/ 
early Holocene times in Pampa and Patagonia (Argentina 

[5 


e 
Brazil), with 
evidence of marine Voy Paliéogoam: Palaeoclima- 
tol. Palaeoecol. 143: 31— 
d D. R., G. Balco & S. D. Willett. 2001. Climate, 
tectonics, and the morphology of the Andes. Geology 29: 
579-582. 
Moores, vs & R. J. Twiss. 1995. Tectonics. W. H. Freeman, 
New = 


Morell, P. 1996. pr a rsity: A river doesn’t run 
through it. Science 273: 6—1497. 
Núñez, L., M. Grosjean " i Cartajena. 2002. 
occupations and climate [n in the Puna de Atacama, 
Chile. Science 298: 821-824 
Oncken, O., E. hong, G. Mana, P. Giese, H.-J. Götze, V. A. 
trecker & E. Mode ane 2006. The 


Frontiers in Earth 


Human 


a miu a xo da Pus 


a. 
rme Tn p EL o ia p E 
America n. University Press, Oxford. 
07b. Tectonism, climate, and landscape change. 
n T. T. Veblen, K. R. Young & A. R. Orme 
m The Physical Geography of South America. 
Oxford University Press, Oxfor 
ws G., C. Chalmers, R. AE s, M. A. Reasoner, D. T. 
Rodbell, G. O. Seltzer & G. “Zielinski. 1995. Potential 
glacial evidence for the Younger Dryas event in the 
Cordillera S North and South America. Quatern. Sci. Rev. 
Paduano, G. a , M. B. Bush, P. jo Baker S. C. Fritz & G. O. 
Selzer. 2003. A vegetation tory of Lake Titicaca 
since the Last Glacial bu B Palaeocli- 
matol. beue pi 259- 
Patton, J. L F. Da E Eon Rivers, refuges, and 
ridges: ae geography of i of Amazonian 
mammals. Pp. 202-213 in D. Howard & S. H. Berlocher 
(editors). Endless Forms: Species ind Speciation. Oxford 
University Press, m 
& 


2005. The history of Amazonian 
: Mechanisms 


(ed itors). Tropical Rainforests: foh a and Future. 
University of Chicago Press, Chica 


—. J. R. Malcolm. 1994. Gene genealogy and 
direcion among arboreal spiny rats (Rodentia: 
mydae) of the Amazon Basin: A o the riverine 

eR oo Pp 48: 1314- 
Placzek, C., J. Qua P. J. Patchett. e | m 


a Js 2 late e lake cycles on the 
uthem Bolivian Altiplano: Implications for causes of 
Soc. Amer. 118: 


dorea diae au Bull. Geol. 
—532. 


. L. Betancourt, P. J. Patchett, 5 A. Rech, 

olmgren & N. B. English. 

dry central Andes over n 

nd Painal edo Ann. Missouri Bot. 
97. 


Ramos, V. A. & A. Aleman. 2000. Tectonic evolution of the 
Andes. Pp. 635-685 in U. G. Cordani, E. J. Milani, : 
Filho & D. A. Campos (editors): v Evolution of 
South America. 31st International Geologi 
Rio de Janeiro, 6-17 August 2000. Geological Society of 
America, Boulder. 

Ribas, C. C., R. G. Meus S ie Miyaki & J- Ciera 2007. 
The assem mbl f 
and climatic dd to independent no 

c. London, 


: Linking 


áez, A. L. Cabrera, A. Jensen & G. Chon, 
Neogene lacustrine record and palacogeography i in the 
Quillagua-Llamara Basin, central A (north 
Chile). Palaeogeogr. Palaeoclimatol. Pastel, 151: 5 37. 
Seltzer, G. O. 2007. 

tropical Andes. Pp. T e T. T. Veblen, K. R. Young & 
A. R. Orme nies Physical oe of South 
America. Oxford University Press, Oxfi 


empre, T., R. F. Bien D.R Richards: E E Marshall, W. 
a C. Swisher m 1997. Stratigraphy and 
chronology Upper Cretaceous-lower Paleocene strata 


of 
in Bolivia and northwest Argentina. Bull. Geol. Soc. Amer. 
109: 709-727. 

RE S.V.& bur rs ES What drives orogeny 

n the Andes? cM 33: 6 

e, M., M. Forster, H. A IM A. potis J. F. Clark, P. 
Schlosse er, W. S. Broecker & G. Bonani. 1995. Cooling a 
a Brazil e p. oe the Last Glacial Maximum. 
Science 269: 379. 

iue M. R . de la A & C. M. Bell. 2000. Timing and 
origin of deformation along the Patagonian fold and thrust 
belt. Geol. Mag. 137: 345-353. 

apia, P. M., S. C. Fritz, P. A. Baker, G. O. Seltzer & R. B. 
Dunbar. 2003. A late Quaternary diatom record of tropical 
climatic history from Lake Titicaca (Peru and Bolivia). 
Palaeogeogr. Palaeoclimatol. a 194: 139-164. 

Thouret, J.-C. , B. Salomons & E. 
Juvigné. 1996. ee changes and glacial 

rs in the Cordillera Cental, 
Colombia. Quatern. Res. 46: 1-18. 

U.S. i. Survey & Servicio Geológico de Bolivia. 
1992. Geology and mineral resources of the Altiplano and 
Cordillera po dental, Bolivia. Bull. U.S. Geol. Su 
1975: 1-365 

n der Homem: T. 1952. Geología del Río Apaporis entre 
p ma y ea La Playa. Informe Servicio Geológica 
Nacional, Bog 


. van der Ham 


Volume 96, Number 3 
2009 


Graham 
The Andes: A Geological Overview 


——— & H. Hooghiemstra. 1995. The El Abra stadial, a 
Younger Dryas equivalent in Colombia. Quatern. Sci. Rev 
14: 841-851. 

——— & —— 2000. Neogene and Quaternary O : 

d Jr and e diversity in Amazoni 

Quatern. Sci. 19-7 725-7 

b H. ads & H. n Dommelen. 1973. Paly- 

nological record of the e upheaval of the Northern Andes: A 

study of the Plioce ower Quaternary of the 
Colombian Eastern Cordillera ui the early evolution of 
its Wr E A Rev. ats Palynol. 16: 1-122. 

Veblen, ung & A. R. Orme (editors). RM 

The tubi QE d South America. Oxfor 


ord. 
, B. S. 1971. Pleistocene changes in the fauna 
and ‘flora " un America. Science 173: 771—780. 
ce ALR. 9. On the eae of the Amazon. Proc. 
ool. Soc. nee 20: 107-110 


1876. The Geographical Distribution of Animals. 
TO New York. 
Watts, W. A. & B. C. S. Hansen. 1994. Pre-Holocene and 
Biene pollen i vegetation history from the Florida 
sula and their climatic xri Palaeogeogr. 
baleen linac Paleo 109: 163-176. 
Webb, R d, S. J. Lehman, R. J. Healy & D. 
Sigman. ros ence of ocean heat dug on the 
climate 7 = Last Glacial Maximum. Nature 385: 695- 
699. 


d J. 1994. Pia and tectonic d of d 
em CE M Pp. 1-15 in H. Seyfried & 

Hellmann (editors). DE of an sd kind ine 

s of m Nicaragua. a, and 
ma. Vol. 7. Institu ms Gel und 
uer oa Stuttgart, Stuttga 

Zeil, W. 1979. The Andes, A Geological Review. Gebrüder 
Borntraeger, Berlin 


CLIMATE IN THE DRY CENTRAL Christa Placzek,? Jay Quade,? 


ANDES OVER GEOLOGIC, 
MILLENNIAL, AND 
INTERANNUAL TIMESCALES’ 


Julio L. Betancourt,’ P. Jonathan Patchett,” 
Jason A. Rech,? Claudio Latorre, Ari Matmon,* 
Camille Holmgren,? and Nathan B. English? 


ABSTRACT 


Over the last eight years, we have developed several paleoenvironmental records from a broad geographic region spanning 


the Mise cs in Bolivia (188-22 
mogenic n nuclide concentrations in surface 
its, and plant macrofossils from urine-encrusted x 


°S} and continuing south along the western Andean flank to ca. 26°S. These records include: 
deposits, ever nitrate d rin iid die d Ap de n from wetland 
odent middens. Arid e ofte I 


ensitive to 


more than 10 O railor, years and de resulted fon de ined of p dde Mi formation of the Altiplano plateau. New 


1 multiple terrestrial co. e 
e rainía Ie events that likely originate Un dn 


ate records across e ee allows reconstruction of the spatial oe tempor ral com 
e Pleistocene wet events the modern sitio odes 


l 
id zone, climate history from multiple proxies 


d the hyperari 


ponents of 
n between two mo nterannual precipitation 


variability, and rena ee oe for de enl Ardeni Pluvial Event (CAPE; ca. 1-8 ka) points toward similar 


st millennial-scale cli 


Key words: Altiplano, Amazon Basin, Andes, 


mate variability. The north-northeast m 
El Ni io-Soathern a (ENSO) variability, and the southeast mode is linked to aridity in the Chac 


ode of climate niis is 
region of 


CAPE, ENSO, middens. 


The dry central Andes is the tripartite region 
western Andean 
(Fig. 1) and is a critical region for understanding the 
drivers of tropical climate change at multiple time- 
scales. Arid environments are often uniquely sensitive 
to climate change, and today modern interannual 
climate variability in the region is pronounced and 
influenced by both tropical climate phenomena, such 
as El Nifio-Southern Oscillation (ENSO), and moisture 
levels in the extratropical lowlands east of the Andes 
(Fig. 2) (Vuille & Keimig, 2004). Here, we compare the 


timing and likely drivers of precipitation changes 
across the dry central Andes over geologic, millennial, 
and interannual timescales. Understanding how re- 
gional climate is sensitive to processes like mountain 
building, the reorganization of global atmospheric 
circulation that occurs over glacial-interglacial and 
millennial timescales, and decadal to interannual 
changes due to processes such as the phenom- 
enon is a step toward assessing where and how this 
region is sensitive to global climate change. 

The Atacama Desert, located along the western 
Andean slope between ca. 18°S and 26°S (Fig. 1), is 


1 We thank Sohrab Tawakholi and Servicio Nacional de Geologia y Mineria (SERGEOMIN) for field logistical mou in 


Bolivia. This work was supported by the 


National Science Foundation (grant EAR-0207850 to J.Q. and J.P., and grant 02- 


13657 to J.Q. and J.B.) and by grants. from the Geological Society of America, the Arizona Geological Survey, qm and 


University of Arizona 
received gram 


peek as well as the Fondo de Desarrollo de Areas Prioritarias grant 1501 (to the Center for Advance 
rollo Cientifi 
nd Department of Earth and Atmospheric Sciences, Purdue University, 


and Biodiversity) and the Fondo Nacional de Desar 
2 Pardue Rare Isotope M 


^U.S. Geological Survey, Desert Laboratory, 1675 Anklam 


ts Proyecto Fondo Basal-23 and the Iniciativa Científica Milenio P05-002 (to the Institute of Ecolog 


ional Science Foundation grant 01-01249. C 
y and 
ed Studies in 


ico y Tecnológico grant 1060496. 


sity of Arizona, Dra Arizona 8572], 


, Tucson, pius 85745, U.S.A. 


5 Department of Geology, 123 Shideler Hall, Miami o Oxford, Ohio 45056, 


* CASEB/Departmento de Ec 
Casilla 653, Santiago, 6513677, Chile. 
“Institute of Earth Sciences, The H 


ologia, Pontíficia Universidad Católica de Chile and m of Ecology and Biodiversity, 


brew University of Jerusalem, Givat Ram, Jerusalem, I 


srael. 
ography and Planning Department, Buffalo State College, 1300 Elmwood Ave., Buffalo, New York 14222, U.S.A. 


£ Geograph 
doi: 10.3417/2008019 


ANN. Missouni Bor. Garp. 96: 386-397. PUBLISHED ON 28 SEPTEMBER 2009. 


Volume 96, Number 3 
2009 


Placzek et al. 
Climate in the Dry Central Andes 


Pacific 
Ocean 


r24^S 


r28^S 


74°W 


Figure 1. 


the driest and perhaps oldest desert on earth (Hart- 
ley et al., 2005). Hyperaridity requires orographic ex- 
clusion of Atlantic moisture by the Andes and 
exclusion of Pacific moisture by the Coastal Cordillera 
and subsiding air resulting from the cold, northward- 
flowing Humboldt Current. The stability and timing of 
moisture exclusion from these two sources are critical 
to determining if the Andean uplift created the 
Atacama or if this aridity results from changes along 


Location of relevant sites and geographic features in the dry central Andes. 


the Pacific coast (Lamb & Davis, 2003). Despite this 
prolonged aridity, major changes have occurred in the 
boundary conditions that contribute to hyperaridity 

the Andes acqui ough elevation to 
constitute a significant orographie barrier. These 
changes include uplift of the Coastal Cordillera 
(e.g., Clift & Hartley, 2007) and changes in the 
intensity of the Humboldt Current (e.g., Molnar & 
Cane, 2007) related to expansion of Antarctic ice 


Annals of the 
Missouri Botanical Garden 


Modern climate systems controlling central Andean rainfall. O haded zones show the 
onal functions, identified. "s Vuille and Keimig (2004). The north- 


gure 2. 
modern precipitation, as major rotated empirical ortho 


two modes of 


northwest mode is modulated by El Niño-Southern Oscillation (E i with strong westerly winds producing drought on the 


ears. The southeast mode is correlated with lowland humidity 


in the Chaco region of Argentina. The 


Intertropical Convergence Zone (ITCZ) is shown in its es (summer) Ses, 


sheets (e.g., Hartley & Chong, 2002) and/or closing of 
the Isthmus of Panama (Ibaraki, 1997). 

The evidence for prolonged hyperaridity in the core 
of the Atacama Desert n matched by paleoecological 
(e.g., Grosjean et al., Betancourt et al. j 


Latorre et al., 2002, ier and valedheolbeiosd (e.g.. 


Betancourt et al., 2000; Bobst et al., 2001; Rech et al., 
2002; Quade et al., 2008) evidence for dramatic 
millennial scale changes in climate along the fringes 
of the Atacama Desert. Thus, the boundaries of the 
Atacama Desert fluctuate in response to these climatic 
events, and the distribution and stability of these 


Volume 96, Number 3 
2009 


Placzek et 
Climate in ils Dry Central Andes 


boundaries through time can give insights into the 
causes of these shifts. Recent evidence (e.g., Quade et 
al, 2008) suggests that ancient millennial scale 
variability had two geographically distinct. modes, 
similar in distribution to two distinct modes of modern 
interannual rainfall variability. 


ATACAMA HYPERARIDITY 


The hyperaridity of the Atacama Desert is due to a 
combination of: (1) the extreme rain shadow created 
igh Andes and Altiplano, which excludes 
moisture from the Amazon Basin; (2) a strong 
emperature inversion along the Pacific coast, which 
effectively blocks Pacific moisture at ca. m 
elevation along the western flank of the Coastal 
Cordillera; and (3) the northern limit of the southern 
Westerlies (Houston & Hartley, 2003). Over millions 
of year: high Andes 
and/or Altiplano plateau was primarily responsible for 
the prolonged aridity of the Atacama Desert. An 
Andean elevation of at least 2000 m is considered 


s, the rain shadow created by the 


high enough to exclude much of the moisture 
originating in the Amazon Basin from the Atacama 


(e.g., Masek et al., 1994; Rech et al., 


formation of the Altiplano plateau and the interaction 
between climate and tectonies in the central Andes 
remain PED dE (e.g.. uos et al., 2006; Garzione 
et al., 2006; Ghosh et al., 
ne of the primary um d T for both a 
landscape and climate that has remained stable and 
hyperarid over the entire Pliocene and Pleistocene is 
extremely high cosmogenic nuclide concentrations 
rom ancient geomorphic surfaces. Cosmogenic nu- 
clides are produced by secondary cosmic rays in the 
uppermost few meters of the earth’s surface and can 
record the age of material suddenly exposed or 
constrain erosion rates (Lal, osmogenic 
nuclide concentrations from stable geomorphic sur- 
faces in the Atacama result in some of the oldest 
exposure ages found anywhere on earth, ranging 
between 9 and 37 million years ago (Ma) (Dunai et al., 
2005; Nishiizumi et al., 2005; Kober et al., 2007). 
Indeed, the Atacama is one of the few locations where 
must be verified b 
exposure ti 


exposure ages 
measurements, as long 
significant quantities of the radionuclides "Be and 
^A] produced during early exposure have decayed. 
Constraints on the rates of sediment production and 
transport in the Atacama also come from cosmogenic 
nuclide concentrations in multiple components of the 
landscape (Placzek et al., 2007) and deposition rates 

erred from dated a sh- fall tuffs (Placzek 


2009). Together, these indicate that overall sion 


et al. 


rates are some of the slowest in the world—a direct 
result of a prolonged arid climate. 

Additional evidence for the onset of aridity prior to 
10 Ma includes: accumulation of nitrate soils in 
ee an end of 

ergene mineralization (e.g., s & Brimhall, 
1988; Sillitoe & McKee, 1996; edis et al., 2006), 
nd changes in stream morphology on the Anden 
flank (Hoke et al., 2006). Ancient nitrate tescik, with a 
firm minimum age of 9.4 Ma from an overlying 
volcanic ignimbrite, attest to this ancient aridity as 


ancient deposits (Rech et al., 


nitrates require hyperarid conditions and today only 
accumulate in the driest portions of the Atacama 
Desert. These nitrate soils, however, probably repre- 
sent several million years of accumulation, and Rec 
et al. (2006) place the minimum age for the onset of 
hyperaridity at ca. 13 M 

At odds with all dis evidence for prolonged 
hyperaridit 
degree of aridity and the deposition of fluviolacus- 


is an inferred association between the 


trine, alluvial fan or evaporite deposits, which leads to 
the conclusion that Pliocene sediments suggest a 
transition from arid to hyperarid conditions as recently 
as 3 Ma (Hartley & Chong, 2002; Allmendinger et al., 
2005). Today, all of these depositional environments 

ceur both in the wetter Andean highlands and across 
the “absolute desert,” a bro. 
Desert completely devoid of precipitation and vascu- 
lar plants, thus d interpretation of modern 
Cosmogenic 


ad expanse of the Atacama 


or ancient ari from such sediments. 


nuclide concentration from the active components of 
the landscape (surface gravel, active alluvial fan 
deposits, and active channels) appears to be eroding 
at a rate that is at least an order of magnitude faster 
than relict geomorphic surfaces (Placzek et al., 
Furthermore, new ?!Ne, Be, and *A1 measurements 
from relict boulders indicate that many of these 
boulders have ages less than 3 Ma (Placzek et al., 
2008), long after the onset of aridity. This movement 
and erosion of all size classes of sediment after 3 Ma 
suggest that periodic rainfall and flood events 
continue to impact the Atacama. Furthermore, it 
suggests that the Atacama Desert, traditionally viewed 
as isolated from rainfall over geologic intervals, has a 

modern landscape that is sha y rare, but recent, 
rain events and is therefore not fully isolated from 
future global i Ms change. 


MILLENNIAL-SCALE CLIMATE CHANGE 


Climate proxies from lakes, wetland deposits, and 


urine-encrusted rodent middens reveal dramatic 
precipitation changes throughout the Pleistocene over 
a broad geographic region. Here, we focus on the 


paleolake record from the Altiplano and what it 


Annals of the 
Missouri Botanical Garden 


reveals about climate variability over the Pleistocene. 
We also compare this lake record to other types of 
roxies across this region during the post late 
ral Andean Pluvial Event (CAPE) 


concluding with an example of how a multiproxy 


climate p 
glacial-age Cent 


approach allows tracking of the source of moisture 
during wet intervals. 


LAKE RECORDS 


Four large lake basins (Fig. 1: Titicaca, Poopé, 
Coipasa, and Uyuni) dominate the Altiplano, and the 
size of the lakes has undergone periodie changes as a 
result of changes in precipitation. In the north, Lake 


Titicaca (3806 m elevation, 8560 km?) is a freshwater 


Desaguadero (Roche et al., 1992). The Río Desagua- 
ero e into the ao Lake Poopó 
(3685 m, 2500 km?) which is ted by a 
ee divide, the Laka sill (3700 al from the 
Salar de Coipasa (3656 m, 2530 km”) and Salar de 
Uyuni (3653 m, 12,100 k 
flats are connected and filled with shallow water 
(< 4 m) (Argollo € Mourguiart, 2000). 

Within these basins, multiple sites were studied 


m?). In wet years these salt 
11 


nd replicate records of lake-level change from 
multiple localities in all three major basins. Particular 
effort was directed toward sedimentary deposits 
associated with various visible roms This 
approach to reconstructing lake-level ry allows 
for direct determination of lake level, A of 
stratigraphy, and dating by two geochronologic 
methods ("C and U-Th, Placzek et al., 2006b). More 
than 170 dates are available from paleclake deposits 
within the basins, and the of both -Th and 
radiocarbon methods xis us to ae our record 
beyond the limit of radiocarbon dating (ca. 45 ka). 
The focus of this dating effort is sedimentary deposits 
indicative of a near-shore environment and the 
massive encrustations of calcium carbonate (tufas) 
found in the paleolake basins. Tufas and aquatic 
ae shells generally form in nearshore environ- 
and incorporate !^ 
me they precipitate. For samples younger than 


and uranium from water in 


45 ka, the quantity of remaining radioactive “C can 
be used to calculate a sample’s age. For older samples 
(greater than 25 ka), however, the very small quantity 
of remaining “C renders samples susceptible to errors 
introduced by contamination with very small amounts 
of modern carbon. Thus, reliable ages greater than 
method, which is 


based on the premise that uranium is incorporated 


25 ka come from the U-Th dating 


into carbonates precipitated from water, but thorium, a 
daughter of uranium decay, is largely not incorporated 
s. Sediments that are clearly associated with 
both 


deep and shallow lake events place constraints on 


into tufa 
ake shorelines or sedimentary units showing 


absolute paleolake elevation. The potential incom- 
pleteness of any single exposure is redressed by 
replication of stratigraphy at multiple locations 
(Placzek et al., 2006a). 

On the Altiplano, extensive natural exposures 
reveal evidence of two deep-lake and several minor- 
lake cycles over the past 120 ka (Fig. 3) in an area 
where today there are mostly barren salt flats or 
shallow saline lakes. The Ouki lake cycle was ca. 
80 m deep, and 19 U-Th dates place this deep-lake 
cycle between 120 and 98 ka (Placzek et al., 2006a). 
Old shoreline and sedimentary deposits from the Ouki 
lake cycle are extensively exposed in the Poopó Basin, 
but no deep lakes are apparent in the subsequent 
record between 98 and ka. Evidence of shallow 
lakes is present in the Uyuni Basin between 95 an 
80 ka (Salinas lake cycle), at ca. 46 ka (Inca Huasi 
lake UE and between 24 and 20.5 ka (Sajsi lake 
ia The Tauca 

een dn 1 sd f 
a 140 m) and largest lake in the basin over the past 
120 ka. Multiple '*C and U-Th dates constrain the 


cle occurred 


a lake eye 
4.1 ka, resulting in the deepest 


ighe ycle along a 
topographically conspicuous shoreline pisa 16.4 
d 14.1 ka. The Coipasa lake cycle produced a 
* 55 m deep lake E e between ca. 13 and 11 ka 
(Placzek et al., . Together, the Tauca = 
Coipasa lake a e the CA 
on the Bolivian Altiplano (Fig. 4). 


est elevation of the Tauca lake 


e occurrence of 


RODENT MIDDENS 


Urine-encrusted rodent middens (henceforth, ro- 
dent middens) are complex nests of local vegetation 
and feces encased in crystallized rodent urine. In arid 
eath 


climates, rodent m 
labs nt remains encased 


iddens are preserved undern 
rock slabs and within caves. Pla 

in middens reflect former vegetation cover within the 
rodent's foraging range, which is usually less than 
200 m (cf. Salinas & Latorre, 2007) the d 
central Andes, middens are produced by at least four 
different rodent families: Abrocomidae (Abrocoma 
cinerea, Thomas, chinchilla rats), Chinchillidae (La- 
gidium viscacia Molina and Lagidium peruanum 
Meyen, southern mountain viscacha), Muridae (Phyl- 
lotis spp., leaf-eared mice), and Octodontidae (Octo- 
dontomys gliroides Gervais € d'Orbigny [1884], 
mountain degu [Latorre et al., 2005]). These rodents 
collect plants for consumption and nest building, and 
studies of modem Phyllotis, Lagidium, and Abrocoma 


Volume 96, Number 3 
2009 


Placzek et 
Climate in ihe Dry Central Andes 


Surface area (1000 Km?) 


—A. Ao lake 


January solar insolation y a00- 
5°S (gray) Ouki N 
48074 T 
£ 
| J Je 
|i fec 
: = 
| É E 
dre H 4 /440 O 
Salinas; 4 | E 
Inca Huasi — "X EV 490, 
LI T L] | T L] LI Í LI I LI LI T T Í T T T 
60 80 100 120 140 
B 
estimated temp change 44 
at Vostok (gray) a 
Jo 
cold and dry o.1. 
o 
2 
© 
o 
a 
5 
E 
LI Li Li | LI L] T | LI LI LI | LI LI LI | LI I LI | LI T T | T T 
0 20 40 60 80 100 120 140 


Figure 3. 
is given in gray. —B. I 
Mix, 1999) fe inad lemper rature change at Vostok (gray) (Petit et al., 


indicate that they are ex generalists (cf. Cortés et 
an they are not likely to 
ine large selective biases into the midden 
record. 


oy A as suc 


Due to the abundance of plant macrofossils, rodent 

midden rich snapshots of local paleoecology at 
the ine (and datable) time they were deposited. 
Rodent middens deposited within the last 45 ka are 
dated using standard “C techniques. Analysis of 
ancient vegetation assemblages is most effective when 
coupled with surveys of modern vegetation in and 
around a collection site. The most basic analysis of 
rodent middens typically involves assessment of the 
percent of extra-local plant species contained in a 
midden and some interpretation of the relative climate 
(wetter, dryer, warmer, colder) represented by that 


xide ooo tión Mee [goethite + hemat tite]) of sediments derived from 


. January insolation (in watts/m’) at 15°S (Laskar, 1990) 
the Amazon (Harris & 
1999). X axis values denote time in ka. 


assemblage. At the outer edges of the Atacama Desert, 
middens containing abundant vegetation are found on 
landscapes that are currently too dry to support plants 
(Betancourt et al, 2000; Latorre et al., 2002). A 
simple proxy for precipitation amount from the central 
Andean midden record is the relative abundance of 
grass, as modern grasslands are currently found where 
there is higher precipitation present in fossil middens 
near the boundary of the Atacama Desert (Latorre et 
al, 2003, 2005, 2006). The O age of grass 
abundance from rodent middens on the fringes of the 
absolute desert in the Salar de Punta Negra region is 
generally high during the CAPE (Latorre et al., 2002). 
Here, rodent middens from the second phase of CAPE 
have a higher percentage of grass abundance than the 


first phase of CAPE (Fig. 4). 


392 Annals of the 
Missouri Botanical Garden 


Sajsi 


Elevation (m above sea level) 


La Nina 


ppm) 
e 
l 


( 


A 
Pco2 
5 e B 
L | 


C2 
1 


T eouepunqe sseb 


meters above 
local water table 
N 
O 
| 


0 5 10 15 20 25 30 


Figure 4. Comparison of pe and climate proxies during the Central Andean Pluvial Event (CAPE). — 
Reconstructed lake-level curve. —B. Change in pCOg in the Western Equatorial Pacific inferred from boron isotope analyses 
of planktonic foraminifera, in which increas al pCOs is deenciaed mun stronger RM. and la Nina- like Conditions 
(Palmer & Pearson, 2003). —C. Reconstructed water table | 2008) 
rodent middens in the Salar de Punta Negra area ae et al. "005. X axis dos e time in ue 


Volume 96, Number 3 
2009 


Placzek et 
Climate in > Dry Central Andes 


PALEOWETLANDS 


Wetlands form where the water table intersects the 
found either within steep- 
walled washes or in less confined settings where small 


land surface and can be 


local closures allow pooling of shallow freshwater and 
the formation of wetland deposits (Rech et al., ae 

003; Grosjean et al. e et al., ys 
Pálessetiond deposits generalis consist of bed 
silt, and biogenic deposits such as organic-rich mats, 
diatomites, and tufa. The abundance of organic 
material in these deposits makes them relatively easy 
to date using radiocarbon, and multiple stratigraphic 
levels within a deposit can often be dated. Further- 
more, the abundance of these deposits in the Atacama 
allows replication of results both within and between 
sites. Questions regarding hydrologic response time 
can be resolved by comparison of wetlands from 
several different settings; in the Atacama we find that 
increased precipitation in the high Andes is very 
rapidly translated into water table rise at multiple 
locations across the Atacama (Rech et al., 2002, 2003; 
Quade et al., 2008). High water tables in the Salar de 
Punta Negra region indicate that CAPE began in this 
region at ca. 17 ka, but may have terminated as late as 


8 ka (Fig. 4) 
SPATIAL AND TEMPORAL EXTENT OF THE CAPE 


Evidence from the CAPE is relatively recent and 
well preserved, allowing evaluation of the spatial and 
temporal distribution of climate change over the entire 
dry central Andes. The CAPE is divided into two 
phases, and the depths of the Tauca and Coipasa lake 
cycles suggest that the first phase of CAPE on the 
Altiplano This 


contrasts with climate records from wetlands in the 


o was the wettest and began at ca. 18 ka. T 
Punta Negra region (ca. 4^S of the Uyuni Basin), 
where high water tables indicate that the second 
phase of CAPE began ca. years after the 
transgression of Lake Tauca. In both areas, the first 
phase of CAPE terminates abruptly at ca. 14.1 ka and 
is soon followed by a second wet interval (Fig. 4). The 
second phase of CAPE created the shallower Lake 
Coipasa on the Altiplano, but the midden record from 
the Salar de Punta Negra region has a higher 
percentage of grass abundance during the second 
phase of CAPE, an indication that this second phase 


N 
i 


was wetter toward the south (Latorre et al., 
While the termination of both the Coipasa lake cycle 
and CAPE in the Punta Negra region is poorly 
constrained in time, the second phase also seems to be 
longer lived to the south (Quade et al., 20 

Paleolake shoreline evidence from the Altiplano 
also supports the assertion that the Coipasa lake cycle 


was sustained mainly from precipitation in the south- 
ern Coipasa and Uyuni basins. Climate affects lake 
levels in closed basins by altering the hydrologic 
balance between runoff, precipitation, and evapora- 
tion while basin topography influences lake levels by 
altering the surface area:volume ratio. In large lake 
systems elsewhere (e.g., Bonneville, Lahontan, Lisan), 
well-developed shorelines correspond to periods when 
a lake level was stabilized as a result of spilling over 
into an arid receiving basin at a lower level (Curry 

Oviatt, 1985; Benson & Paillet, 1989; Benson et al., 
1990; Bartov et al., 2002). Thus, a lake system is 
buffered to climate fluctuations at the level of a 
spillway because the receiving basin must fill before 
the lake in the spillo rise. The 
degree of buffering depends on the relative size of the 


over basin can again 


two basins. In the" case of the Poopó-Coipasa-Uyuni 
system, the Basin is erg ead smaller 
(< 1/3 the size) SN the combined Coipasa-Uyuni 
basins (Fig. 5). Thus, if a lake filled these basins with 
water from the north (the Titicaca and Poopó basins), 
then such a lake would have a relatively long period of 
stability at the level of the Laka sill (the spillway 
between Poopó and Coipasa). This should result in a 
prominent shoreline in the Poopó Basin at ca. 3700 m, 
the elevation of the Laka sill. In contrast, if a lake 
filled the larger and more southern Coipasa and Uyuni 
basins first, then the percentage of change in surface 
area at the level of the Laka sill is much smaller, so 
pronounced shorelines would not develop (Fig. 5). 
The maximum elevation of the Coipasa lake cycle 
remains poorly constrained because a prominent 


shoreline is not visible. Chronological evidence, 
however, suggests that at its maximum extent the 
Coipasa lake cycle approximated the elevation of the 


Laka sill. 


MODERN CLIMATE VARIABILITY 


Today, the sources, timing, and variability of 
precipitation are different for the northern Altiplano, 
i flank, and the 
iplano, more than 80% 


the southern Altiplano, western Andean 
Atacama. On the northern Alt 
of total annual precipitation falls in the austral 
summer (December to March) ers id and this 
moisture traverses the Amazon n the summer 
months when the Intertropical ae Zone 
(ITCZ) is displaced southward and convection is most 
m in the Amazon Basin (Lenters & Cook, 1997) 
Fig. 2). This moisture source to the north and east of 


= 


the Altiplano produces a pronounced north-south 
gradient and is referred to as the South American 
Summer Monsoon (SASM) (e.g., Zhou & Lau, 1998). 
The SASM on the northern Altiplano is modulated by 
ENSO variability, and the strength of the trade winds 


Annals of the 
Missouri Botanical Garden 


Titicaca 


Titicaca 


Figure 5. Schematic cross section of the T 


Laka sill 
Coipasa 


Uyuni 


drographic s Vertical exaggeration is ca. 
uth. 


: ii hy 
830X. —A. Filling of the basins from the Pea LB. p si the vim pur from the 


is strong in La Nifia years, resulting in increased 
precipitation. Conversely, during El Niño years, 
aridity dominates in upland Peru and Bolivia, but 
torrential rains occur along the Pacific coast (Acei- 
Vuille et al., 1998, 1999; Garreaud & 
; Vuille & Keimig, 2004). 

summer rainfall on the southern 
Altiplano and western Andean flank has a mode of 
variability that is closely tied to precipitation 
anomalies and humidity levels over the Chaco region 
of Argentina (Vuille & Keimig, 2004). Thus, today 
there are two distinet modes of variability in summer 
rainfall (Vuille & Keimig, 2004) (Fig. 2). The north- 
northeast mode is tied to ENSO variability, and the 
southeast mode i is tied to extratropical Leod 
anomalies the lowlands east es. 


Unfortun. 


southern mode of modern climate in the dry central 


o 
ately, a more complete eae d this 


Andes is hampered by a lack of reliable and 
continuous precipitation data. Recent advances in 
the isotope hydroecology of columnar cacti and their 
spines (English et al., 2007) and tropical dendrochro- 
nology (e.g., Evans & Schrag, 2004; Anchukaitis et 
al., 2008) should produce more detailed records of 
recent climate throughout South America. 

In contrast to the Atlantic moisture falling on the 
Altiplano and Andes, the Pacific is likely the source 
of the scant precipitation that falls today in the 


Atacama. Pacific moisture is effectively excluded 
from the dry central Andes by the descending limb of 
the southeast Pacific anticyclone under the influence 
of the cold Humboldt Current (Vuille, 1999), which 
has likely been active since the early Tertiary (ca. 
5 Ma) (Keller et al., 
Cordillera also limits the inland penetration of Pacific 
to a narrow elevation band (500-1000 m). 
Although the Coastal Cordillera largely blocks Pacific 
storms, rare ee events may penetrate the 
Atacama Desert (May through 
uly). These storm bm typically migrate northward 


e steep Coastal 


> 


the austral winter 


from the westerly precipitation belt that forms the 
southern boundary of the Atacama at ca. 26% (Vuille 
1997). Today, Pacific sea surface 
temperature gradients modulate penetration of these 
Pacific fronts in stern 

flank, and El Niño years are associated with increased 


Ammann, 


nto the Atacama and we Andean 
precipitation and/or fog intensity in the Atacama 


(Dillon € Rundel, 1990). 


CLIMATE CHANCE IN THE DRY CENTRAL ÁNDES: 
MECHANISMS AND IMPLICATIONS 


Potential causes of climate change in the dry 
central Andes include: (1) changes in seasonality, 
e.g., Baker et al., 


2004); (2) 


especially local summer insolation 


2001a, b; Rowe et al., 2002; Fritz et al., 


Volume 96, Number 3 
2009 


Placzek et 
Climate in ilio Dry Central Andes 


changes in global temperature (e.g., Blodgett et al., 
1997; 2003); (3) changes in adds 
(e.g., Mourguiart 


Garreaud et al., 
edru, 
2003); and (4) changes in sea surface temperature 
gradients (e.g., Betancourt et al., 2000; Garreaud et 
al., 2003; Placzek et al., 2006b; Quade et al., 2008). 
Our lake chronology strongly argues against simple 
forcing of summer precipitation by summer insolation, 
and we rule out local January insolation as the 
primary driver of lake cycles; both deep lakes occur 
during periods of low oderate local summer 
insolation. The Tauca lake cycle reached a maximum 
between 16.4 and 14.1 ka, ca. 5 ka after the 
insolation peak at ca. 20 ka (Fig. 3), and the Ouki 
ke cycle spans the most profound minimum (10 
100 ka) in January insolation in the past 200 ka. 
Similarly, the Ouki lake cycle and the CAPE occur 
uring periods of moderate global temperature, 
indicating no direct link between precipitation 
changes and temperature. Past, present, and possibly 
future climate changes in aridity over the region are, 
however, likely linked to changes in ENSO variability 
and moisture level in the eastern lowlands 
CAPE 


between ENSO and p anomalies over the 


allows examination of the interaction 


ran Chaco lowlands during past wet events over the 
dry central Andes. E and climate proxy data 
for CAPE suggest a temporal offset between the 
Altiplano lake record and the Salar de Punta Negra 
wetland and rodent midden record. We attribute this 
to the operation of two separate modes of rainfall over 
the northern and southern portions of the central 
Andes during CAPE. The timing of the first phase of 
CAPE coincides with evidence for intense upwelling 
(La Nifia) in the central Pacific between 18 and 13 ka 
(Palmer & Pearson, 2003) (Fig. 4). La Nifia—like 
conditions today result in wet years on the Altiplano, 
and important ancient links may exist between central 
Andean moisture and Pacific sea surface temperature 
gradients during the Pleistocene. The modern link 
between ENSO anomalies and precipitation S m 
is weaker farther south where CAPE starts — 1000 
years later. The second phase of CAPE created the 
shallower Coipasa lake cycle, but was the more 
Regu precipitation event farther south (Fig. 4). 
ly, modern precipitation anomalies on the 
western Andean flank to the south are tied more 
closely to circulation anomalies over the Gran Chaco. 


CONCLUSIONS 


Hyperaridity in the core of the Atacama Desert 
dominates over a period greater than 10 Ma, in 
flank and the 


Altiplano, where evidence from a variety of climate 


contrast to the western Andean 


proxies points toward significant changes in paleo- 
precipitation during the Pleistocene. Over long 
periods of time (7 10 Ma), the uplift of the Andes 
and the formation of the Altiplano plateau are critical 
in the formation of the Andean rain shadow, making 
the Atacama Desert uniquely long-lived and arid. 
Conversely, summer insolation over the Altiplano 
appea 


n to drive changes in 
precipitation over the Altiplano 


plateau does 
or Ámazonia over 
millennial and glacial-interglacial timescales. In- 
stead, 
paleorecords, increasingly points to ENSO-like vari- 


evidence, from both modern climate 
ability and extratropical moisture over the Gran Chaco 
region of Argentina as causal mechanisms for climate 
variability on the Altiplano and western Andean flank. 
These two modes of modern central Andean climate 
variability appear to operate over different geographic 
and at different ENSO 


variability is cocoa more _ significant on the 


regions time intervals. 


(18.1-14.1 ka), 
prolonged La Niña-like conditions. In contrast, 
extratropical moisture is today more significant on 
the western Andean flank and may play a greater role 


during the latter phase of CAPE (after 13 ka). T 


modern climate variability and past millennial scale 


hus, 


variability appear to be forced by the same mecha- 
nisms and suggest that future climate changes in the 
region will not come as a direct result of temperature 
shifts, but rather from 


teleconnections to globa 
circulation patterns such as ENSO 


Literature Cited 


Aceituno, P. On the functioning of the Southern 
Oscillation i in e South America sector. Part I: Surface 
24 


ndinger, R. W., G. González, J. Yu, G. Hoke & B 
Isacks. 2005. ieee parallel shortening in the 
Chilean Forearc: Tectonic and climatic implications. Bull. 
Geol. Soc. Amer. 117: 89-104. 

Alpers, C. N. & G. H. Brimhall. 1988. Middle Miocene 
climatic ae in the Atacama Desert, northern Chile: 
Evidence from ne dalla at La Escondita. 
Bull. Geol. Soc. Anes 100: 1640-1656. 

Anchukaitis, K. J., Ns Evans, N. T Wheelwright & D. r 


Northern 


calibration in al montane cloud forest trees. A 
Geophys. Res. 113: G03030. 

Arancibia, G., tthews & C. Perez de Arce. 2006. 
K-Ar and M Geochronology of supergene processes 
in a Atacama Desert, northern Chile: Tec 
climatic bas J. Geol. S 


history of the Bolivian Altiplano. Quatern. Int. 72: 36-51. 


illennial and orbital timescales on the 
Bolivian Altiplano. Nature 409: 698—701. 


396 Annals of the 
Missouri Botanical Garden 
— ———, G. O. Seltzer, S. C. Fritz, R. B. Dunbar, M. J. Gro Fritz, S. C., P. A. Baker, T. K. Lowenstein, G. O. Seltzer, C. 
P. M. “Tapi a, S. L. Cross, H. D. Rowe & J. P. Broda. 20015. A. Rigsby, G. S. Dwyer, P. M. Tapia, K. K. Arnold, T.-L. 
The mou of South American tropical precipitation for Ku & S. Luo. 2004. Hydrologic variation E the last 
the past 25,000 years. Science 291: 640—643. 170,000 years in the southern | of South 
Barnes, J. B., T. A. Ehlers, N. McQuarrie, P. B. O’Sullivan & America. Quatern. Res. 61: 95-104 
Pelletier. 2006. Eocene to recent variations in Garreaud, R. D. € P. Aceituno. 2001. Interannual rainfall 


erosion across 
northern Bolivia: Implications for plateau evolution. Earth 
Planet. Sci. Lett. 248: 118-133. 

Bartov, Y., M. Stein, Y. Enzel, A. Agnon & Z. Reches. 2002. 
Lake levels and sequence stratigraphy of Lake Lisan, the 
oo precursor of the Dead Sea. Quatern. Res. 57: 

21. 


TN L. V. & F. L. Paillet. 1989. The use of total lake- 
surface area as an indicator of climatic change: Ex- 
a from the Lahontan Basin. Quatern. Res. 32: 

62-275. 


y, R. I. Dorn, K. R. Lajoie, C. G. Ovi 
Si W. Robinson, "e. I. Smith ES. Stine. 1990. CRM E 


t 
Palasoclimaid]. Palaeoecol. 78: 2 - 
Betancourt, J. L., C. Latorre, J. A. R K. A. Rylander & J. 
Quade. 2000. A 22,000-year SW of monsoonal pre- 
cipitation from northern Chile's Atacama Desert. Science 
289: 1542-1546 
Blodgett, T. A., J. D. Lenters & B. L. Isacks. 1997. 
E on the origin of paleolake expansions in the 
entral Aies, Kairit Horace J. 1: 1- 2 n e 
Ai npr t 
35627», s 1 May 2009. 
s ., T. K. Lowenstein, T. E. Jordan, L. Godfrey, 
E p T.-L. Ku & S. Luo. 2001. 106 ka 
en record from drill core of the ‘in de 
tacama, oss a Palaeogeogr. Palaeoclimatol. 
Palaeoecol. 1 1-4 
Clift, P. D. & " Je am 2007 
erosion and coastal underplating. Geology 
Cortés, A., J. R. Rau, E. Miranda & J. E. Jiménez. 2002. 
Hábitos alimenticios de Lagidium viscacia y Abrocoma 


. Slow rates of subduction 
: 503—506. 


stillstands, and contractions during the last deep-lake 
cycle, 32,000 to 10,000 years ago. Pp. 9—24 in P. A. Ka 
& . Diaz (editors), Problems of and Prospects for 
Predicting Great Salt Lake Levels. University of Utah 
Center for Public Affairs and Administration, Salt Lake 


1Ly. 
Dillon, M. O. € P. W. Rundel. 1990. The botanical response 
of the Atacama and Peruvian desert floras to the 1982-83 
El Niño event. Pp. 487—504 in P. W. Glynn (editor), 
Global Ecological Consequences of the 1982-83 El Niño- 
Southern ew n. Elsevier, Amsterdam. 
Dunai, T. J., pl López & J. 2005. 
Oligocene- eee age of aridity in the Atacama Desert 
revealed by exposure dating of erosion- ue land- 


forms. Geology 33: 321-32 


Juez-Larré. 


. Sandquist & D. C. 
b-annual variations of 850, 
of a columnar cactus, 
Carnegiea gigantea. Oecologia 154: 247—258. 

Evans, M. N. & D chrag. 2004. A stable isotope-based 
approach to tropical dendroclimatology. Geochim. Cosmo- 
chim. Acta 68: 3295-3305. 


variability over the South oe Altiplano. Monthly 
Weath. Rev. 125: 3157-317 

——— Vuille & A. C. ce 2003. The climat E 
Altiplano: Observed current conditions and mechanisms of 
past changes. Palaeogeogr. Palaeoclimatol. Palaeoecol. 
194: 5-22. 

Garzione, C. N., P. Molnar, J. C. Libarkin & B. 
MacFadden. 2006. o late Miocene rise of the Bolivian 
Altiplano: Evidence for removal of mantle lithosphere. 

Earth Pla a o Let. 241: 54. 3- S 

Ghosh, P. ne & J. M. Eiler. 2006. Rapid uplift of 
the ee revealed Ti i m bonds in paleosol 
carbonates. MA ore 311: 511-515 

m goan M TR Cartajona o B. Messerli. 1997. 
Mid lim and the Atacama 
Desert, northern Chile. p = 48: 239-246. 

& . 2005. Palaeoindian occupation 

of the Atacama Desert, orthern Chile. J. Quatern. Sci. 20: 
643—653. 
Harris, S. E. 


. C. Mix. 1999. Pleistocene a 
bi ae Amazon Basin recorded in deep sea 
sediments. Qu atern. Res. 51: 14-26. 

Hartley, A. J. & G. Chong. 2002. Late Pliocene age for 
Atacama Desert: Implications for the lesecaton of 
western South America. Geology 30: 434: 

, ——, J. Hou sión & A E. m a 150 


HR TUO E 


Desert, d Chile. J. Geol. Soc. London 162: "wh 
e, G. D Isacks, T. E. Jordan & J. S. Yu. 2006. 


potential contribution to the 
tacama Desert. Int. J. 


Ibaraki, M. 1997. Closing of the Central American Seaway 
and Neogene coastal upwelling along the Pacific coast of 
South America. Tectonophysics 281: 99-104. 

Keller, G., E Adatte, C. Hollis, p p I. Zambrano 


Jimene: WE Sinne sbec ck, A. Hale-Er E 

997. Th > day event in Ecuador: 
Reduced biotic effects due to eastern boundary current 
setting. Mar. Micropaleontol. 31: 97-133. 


Kober, F., S. Ivy-Ochs. F. Schlunegger, H. Baur, P. W. Kubik 
& R. Wieler. 


rfaces: In situ 

roduction-rates and erosion “odele Earth 
Planet. Sci. Lett. 104: 424—439. 

Lamb, S. & P. Davis. 2003. Cenozoic jars change as a 
Mog cause for the rise of the Andes. Nature 425: 


192-1 
Laskar, J. E The clupas motion of the solar s A 
88: 266-291. 


Latorre, E Ld. dos K. A. o & J. Qua del 
20 


de Atacama Basins, northern Chile (22-245). Bull. Geol. 
Soc. Amer. 114: 349-366 


Volume 96, Number 3 
2009 


Placzek et 
Climate in ino Dry Central Andes 


—— —— O. Matthel. 2003. 
vegetation TTE from the arid prepuna of northern ate 
23°S) over the last 13,500 e Palaeogeogr. 
pub A rip 194: 223-246 
sA; h, J. Quade, C. Holmgren, C. 
E e M M Vuille & x A. Rylander. 
MES e Quaternary history of the Atacama Desert. 
. 13 ES in M. Smith & P. Hesse cium 23° S: The 
a and Environmental History of the Southern 
Deserts. National Museum of Australia Press, Canberra. 
& M. royo. 2006. Late Quaternary 
vegetation and climate hio of a perennial river canyon 
in the Rio Pes basin (22°S) of northern Chile. Quatern. 
Res. 65: 4504 
Lenters, J. D. & va H. Cook. 1997. On the origin of the 
Bolivian High and related circulation features of the South 
American climate. J. Atmos. Sci. 54: 656-677. 
, B. L. Isacks, T. L. Gubbels & E. J. Fielding. 
1994. Er rosion and tectonics at the margins of continental 
s. Res. 99: 13941-13956. 
M. A. eas 2007. Early Pliocene (pre-Ice Age) 
El Ning like global climate: Which El Nifio? Geosphere 3: 
5. 


Mourguiart, P. & M. P. Ledru. 2003. Last glacial maximum in 
an Andean cloud forest environment (Eastern Cordillera, 
Bolivia). et 31: 195-19. 

Nishiizumi, K., M. W. Caffee, R. C. Finkel, G. Brimhall & T. 

5. Qt E a fossil alluvial fan landscape of 

Miocene = in the Atacama Desert of northern Chile 

using c enic velit exposure age dating. Earth 

Planet. Sai. doit 237: 499-507. 

seit Quies 23,000-year record 


Palmer, M. R. & P. N. 
of s n the Western Equatorial 
482. 


urface water pH an 

Pacific Ocean. Science 300: 480— 
Petit, J. R., J. Jouzel, D. Raynaud, N. I. Barkov, J.-M. 
Mus n i o J: Po ie M. Davis, G. Delaygue, 
M e, V. M. Kotlyakov, M. Legrand, Y. 
m enkor, °C. ice L. Pépin, "c Ritz, E. Saltzman & 
M. Stievenard. 1999. Climate ad atmospheric history of 
he Vostock ice core, 


tropical climate change. Bull. Geol. Soc. Amer. 
ET 
—— ichett, J. Quade & J. D. M. Wagner. 2006b. 
MEUM i uu d U-Th dating of Qu 
carbonate dep An example from the BM 
Altiplano. Dn Geophys. Geosyst. 7: Q050 
—, A. Matmon, D. Granger, J. Quade € M. i on 
2007. Erosion rates in the Atacama Desert, northern Chile 
(24^S) from multiple cosmogenic nuclides. Geological 
Society of America Abstracts with Programs 39: 513. 


, S. Niedermann, D. E. Granger, A. Matmon & J. 
Quad 2008. een 2INe,'"Be and “Al in boulders 
from the central Atacama Desert, northern Chile. 

a “Union 89(53), Fall Meet. BE 
Abstract ie ae 
J. 


n ES RUN P. J. Patchett & C. Pérez de Arce. 
- 3009. eee chronology and stratigraphy of 


Neogene tuffs of the Central coe region. Quate 

Geochronol. 4: 22 = 

uade, J., J. , J. L. Betancourt, C. Latorre € T. 

Fisher. 2008. Paleovetlands and regional climate eod 
the central A a Desert, northern Chile. Quatern 

Res. 69: 343— 

ech, J., J El &]. L. Betancourt. 2002. Late Quaternary 

eMe dm of the central Atacama Desert (22-24°S), 


CE Betancout 2 003. Re- 
osits at Quebrad. 
Puripica, northern a Passes icon. 
ES 194: 207. 
B.S.C 


urrie, G. Et & A. M. Cowan. 2006. 
li h and uplift in the aa Desert, 

Chile. alo 34: 761 a 
Roche, es, J. s & R. Mat 1992. 


TE 

Claas and ivi a Ns Lake Titicaca Basin. 
63-88 in C. Dejo Itis (editors), Lake 
a A Synthesis of Lisi Knowledge. Kluwer 

Academic a Dordrec 
m H. D., ea Mision G. O. Seltzer, 
Le kor S. Fritz. ae Ben moisture balance 

merican Altipl 

> S ae Maximum. ae Change 52: 175-199. 
Salinas, M. . Latorre. 2007. Un Estudio Tafonémico 
sobre js cui de la Diversidad de Especies 
Vegetales en Paleomadrigueras de Roedores del Norte de 
Chil la I eunión Binacional de 


ile. Resúmenes 
Ecologia, La m Chile. 
Sillitoe, R. H. & E. H. McKee. 1996. Age of supergene 
oxi nd enrichment in the Chilean m copper 
rovince. Econ. Geol. 91: 164—179. 
Valles M; 1397: Aunosphone circulation over the Bolivia 


the erus Mu Int. a men 19: 1579— 1600. 
— & C. Ammann. 1997. ee snowfall ET in 
ES Apes arid FTN Climatic Change 36: 413-42 
& F. Keimig. 2004. iaa! variability of 
mertime convective cloudiness and precipitation in 
ntral Andes derived from ISCCP-B3 data. J. Clim 
e 
——— , C. Bra 
1998. Bn ea circulation 


aun, F. Keimig & R. S. Bradley. 
o associated with 


hys. Res. 103: 11191-11204. 
Z 8. Does a monsoon climate exist 
over South en, J. Clim. 11: 1020-1040. 


DIVERSIFICATION OF THE 
SOUTH AMERICAN AVIFAUNA: 
PATTERNS AND IMPLICATIONS 
FOR CONSERVATION IN THE 
ANDES! 


Jon Fjeldsá? and Martin Trestedt* 


ABSTRACT 
By combining distributions and phylogenies for large groups of birds, it is now possible to disentangle the relative. roles of 
ns 
During the upper 2 PR E was most intense in the tropical Andes m with recruitment back into 
m (me 


relan 
ihe ironical iow lands and into South Am open biomes. Within the tropical Andes, endem: nverse range size) 
and 1 mean branch length ae of phylogenetic nodes on lineages) i increase from the foothills up to m tree line and then 


n highland 


,the LI of e of 
nt to maintain the proc 
opulate: 


Key words: ara po 


1e plays a 
ee is locally aggregated, often with mar Led: peaks in areas immediately adja 
new species is linked with local factors that, over a Ina 
ople. If we ess of diversification, it becom 
ach. of peas MU NE in wilderness areas with few 


conservation, South America, 


special role in the PAPE pro cess. The 
edi nt to ancient population 

time MN CUN were 
es essential to supplement the 

people with efforts to support sustainable 


speciation. 


The tropical Andes region is recognized as one of the 
ee ee d for E and conserva- 
. Since the “hot was first launched 
rois 1989, ib. it ds oe a widely used 
buzzword, applied indiscriminately on different spatial 
scales, often combining species richness and endemism, 
although more critical data analysis reveals that spatial 
variation in species richness is often not congruent with 
endemism or with the occurrence of threatened species 
(Orme et al., 
biodiversity are idiosyncratic, we cannot use them 


2005). Because different measures of 


indly as a “currency of biodiversity.” 

Significant strides have been made to analyze to 
what extent the geographical variation in biodiversity 
measures reflects ecology, such as contemporary 
climate and landscape complexity. Ecological models 
perform quite well in explaining species richness, but 
closer examination reveals that the results are driven 
by the many data points representing widespread 
species, and that the models do not explain local 
aggregates of species with small distributions, which 
may instead reflect speciation history (Jetz et al., 


2004; Rahbek et al., 2006). With the rapid increase in 
DNA-based eae it is now possible to begin to 
link together the macroecological and historical 
approaches to ae (Fjeldsá & Rahbek, 
2006; Hawkins et al., 2006; Fjeldsá et al., 2007b). 
In this paper, we will illustrate how one can analyze 
large amounts of phylogenetic and distributional data 
to describe historical components 
ill 


also dis 


past history and present ecology in explaining the 


biodiversity 
patterns. We wi scuss the S roles of 
Andean biodiversity hotspot. To place the Andes in a 
regional context, we will first take a continent-wide 
overview and then scale down to examine the hot 
points within the Andean hotspot. Based on interpre- 
tations of distinctive patterns of endemism in the 
Andes, we will finally discuss where to focus 
conservation actions. 


MATERIALS AND METHODS 


We will base this analysis on bird data. Birds are 
not perfect indicators of wholesale biodiversity (see 


! P. Williams kindly provided the WORLDM 


huti ] dat: 


Ft +) t 
han i. years ue collaboration with nume 


and to generate Figures 14. The 
rous institutio ons and individuals. In 


E Museum, Universitetsparken 15, DK-2100 Copenhagen Denmark. jfjeldsaa@snm.ku.dk. 
ecular Systematics Laboratory, Swedish Museum of Natural History, P.O. Box 5007, SE-104 05 Stockholm, Sweden. 


m En 3417/2007148 


ANN. Missouni Bor. Garp. 96: 398-409. PUBLISHED ON 28 SEPTEMBER 2009. 


Volume 96, Number 3 
2009 


Fjeldsá & lrestedt 
Diversification of South American Avifauna 


Moore et al., 2003) but are useful in terms of the 


quality of ilie available data. The WR ADR 
taxonomy is more or less complete, an 
avian groups there are now relatively ee 


molecular phylogenies available. Furthermore, 
enough is known about species distributions to make 


fairly realistic distribution maps. 


DISTRIBUTIONAL DATA 


We used two databases of South American bird 
AP software 
based on extensive review of 


distributions digitized in the 
(Williams, 1998) 
museum collections and literature and comprehensive 
fieldwork by the first author in the Andes region (see 
primarily Fjeldså & Krabbe, 1990). These databases 
are (l) a continent-wide database over all resident 
birds in a geographical grid of 1 X 1 geographical 
degrees described, e.g., by Fjeldså and Rahbek (1998) 
and Rahbek et al. (2006); and (2) a database over the 
tropical Andes region, in a grid of 15 X 15 geo- 
1999, 20052). 
The latter is the most finely resolved data set of its 
kind for the Andes. Such high 
obtained in highly structured landsca 


graphical minutes (see Fjeldsa et al., 


resolution can only be 
pes, so this 
database includes only species that breed at eleva- 
tions higher s 2500 m (at least in some parts of 
their (cdi us lowland species ae ng the 

5 uth b. birds wit most restricted 
RUD e ranges. This database contains 300,000 
in-grid-cell records, of which 87,000 are confirmed 
present in a grid cell; the rest represents conservative 
interpolation in which careful scrutiny of topographic 
maps and satellite images, including Google Earth 
(<http://earth.google.com/>), suggests that the spe- 
cies should be present between AU. points. The 
interpolation is mostly used for species that are 
generally widespread or ubiquitous (and pec are 
rarely reported). For species considered to be rare or 
local, the use of interpolation is restrictive. 

For each grid cell, species richness is defined as 
the number of species (as accepted by the American 
Ornithologists’ Union) and endemism is defined as 
range-size rarity, which is the inverse range size 
(number of grid cells in the species’ range). This can 
be expressed as a mean value per species represented 
in a cell (nean endemism) or as a sum of inverse 
range-size scores 

To illustrate variation at the local scale, we used 
bird data from the 
is a local hot point at 16°S, where the species richness 
is almost as high as near the equator (98.596 


Yungas of Cochabamba, Bolivia. It 


compared with the average of four grid cells 
representing humid Andean slopes adjacent to the 
equator). The data are based on several ornithological 


surveys taken between 1991 and 2000 along three 
elevational transects in the Carrasco National Park 


F 


western, central, and eastern parts), from the 
lowlands over the top ridge of the Tunari Range at 
ca. 4000 m and into the adjacent rain shadow in the 
Cochabamba Basin. Because of undersampling in 
coca-growing areas in the foothills and in some very 
steep areas at mid-elevation, we found it appropriate 
to use interpolation to connect species records at 
different elevations along transects. We also added a 
y of 
to the 


o. in species less, we examined the 


historical records (mainl 
d but rare species additi 


variation in range-size rarity (endemism), as calculat- 
ed from the continent-wide database. 


HISTORICAL DATA 


The study of avian evolution by come and Ahlquist 
(1990) has, despite problems with their 
method of molecular pheneties e on DNA-DNA 
hybridization, already served as a basis for describing 


serious 


the general pattern of accumulation of ancient taxa in 
the tropical lowland rainforests and more 
diversification in montane 
and at high latitudes (Fjeldsà, 1995; Hawkins et al., 
2006). 

Today, better data based on DNA sequences are 
available. We obtained phylogenetic data by literature 
h collabor 


between the institutions of the two authors, where we 


review and through researc ation, primarily 
aim to generate global molecular phylogenies for the 
asserine birds (order 
e will present geographical distri- 
bution patterns for some clades defined through 
molecular phylogenetic studies, including. the histor- 
ical diversification of the suboscine y Furnar- 
iidae (Fjeldsà et al., 2005b, 2007b; ic et al, 
2006, unpublished data). This d was chosen as 
an example because it is endemic and diverse (302 
species) and has adapted to all terrestrial environ- 
ments in the continent and is a prominent component 
of the avifauna of even the harshest environments 
(Remsen, 2003). Only 60% of the furnariid species 
have so far been included in molecular phylogenies, 
lr and the 


solved, 
sampling gaps in the terminal parts of the phylogeny 


but the deeper branches are wel 


can therefore be reasonably filled using published 
judgments about relationships within these subgroups. 
In some terminal branches with many closely related 
species, which are left unresolved, branch lengths 
were assigned to each species by assuming a regular 
spacing of nodes (thus, a trichotomy translates into 1.6 
steps: four species in a clade represent two steps; 
eight species, three steps; 16 species, four steps, etc.). 


Annals of the 
Missouri Botanical Garden 


e l. Species richness patterns for ancient avian Eo in South America. —A. S 


Figur 
avian clades (of one to four species) dating back to the 


branches of New World flycatchers, altogether 248 species Den UE 


some small clades, all of whi 
greatest (black cells) to least (white cel 


Because a chronogram with local calibration points is 
not yet published for the Furnariidae family, we 
analyzed the diversification of the group by giving 
each species a branch-length score, which is the 
number of nodes to the root of the phylogeny. For a 
simple illustration, we divided the species into 
quartiles of branch lengths, with the first quartile 
being the 25% of the species representing the fewest 
apparent splitting events since the origin of the 

and the fourt 


representing the most splitting events. 


gr 
h quartile being the 2546 of the species 


RESULTS 
VARIATION IN SPECIES RICHNESS IN SPACE AND TIME 


By reviewing the published molecular evidence of 
avian evolutionary relationships, we identified 65 
species representing em clades (of one to four 
species) dating back to the Eocene or even earlier 
(e.g.. 
Opisthocomus hoazin; broad-billed Sapayoa, Sapayoa 


rheas, i pna screamers, Ánhimidae; hoazin, 


aenigma). These species represent lineages in which 
there was little or no subsequent speciation, or wher 

s had been erased by extinction (Ricklefs, 
2003); we see them here as surviving representatives 


such events 


IS 


Mete = 


Hi 
or before (as judged from n RE ev idence). —B. The deep 
Cotingidae, Pipridae, Tityrinae, pipromorphines, and 


h had their origin in the Oligocene (Ohlson et al., 2008). Species buen is represented from 
s). 


of the endemic South American avifauna of the Early 
Tertiary climatic maximum. As illustrated in Fig- 
ure 1A, these species are today mainly found in the 
tropies outside the Andes, and to some extent in the 
d with m concentrations in Nis 
floodplains with mosaics of swampy and s 

habitats (Sncumbíos/Napo/Pastaza i in Ecuador, mer 
and locally near Pucallpa and Madre de Dios, Peru; 
Bolivian savannas and adjacent Pantanal; the middle 
reaches of the Amazon River of Pará and savanna 
Brazil; Cuyuni-Mazaruni in 


mosaics of Amapá, 


Gu ims This richness pattern is almost identical to 


ecosystem productivity (Rahbek et al., 
only difference is a slightly higher representation of 
ancient species in the hydrologically unstable Chaco. 

Our next examples are from the largest endemic 
South American avian radiation of suboscine passer- 
ine birds (more than 1100 species), which is now 
subject to detailed phylogenetic study (Irestedt et al. 

» A ; Fjeldsà et 007a; Tello 

Bates, 2007; Ohlson et al., 2008). The early (Middle 
Tertiary) suboscine radiations (antbirds and gnat- 
catchers in the tracheophone radiation and cotingas, 
manakins, tityrines, and pipromorphines in the New 


Volume 96, Number 3 
2009 


Fjeldsá & lrestedt 401 
Diversification of South American Avifauna 


— 


Comparison of species E. patterns for three E that all E ind the early Miocene. —A. Cor 


igure 2. 
ovenbirds, Furnariinae (Irestedt et al., 2006). —B. New W 


tchers, Tyranni on et al., 2008). —C. Tanagers, 


rt a a larger group 
). 


colonizing from North America. Siis richness is represented from greatest (black cells} to least (white cells 


World flycatcher radiation) are all typical groups of 
the tropical lowland forests. Because a long branch 


separates the early radiations of New W 


et al., 2008), we present these early flycatcher clades in 

map (248 - Fig. 1B) that illustrates the species 
richness pattern of a group whose diversification started 
during the Oligocene and continued to o in the 
s high 
throughout the huncid tropical forest liane i the 


same (forest) habitat. The species richnes 

peak concentration of species in Ecuador and Peru on 

the transition from the Amazon lowlands toward the 
ndes. 

We illustrate the species richness pattern of groups 
whose diversification started well into the Miocene, 
with three Pw -rich groups as Een of these, 

nae (core ovenbirds, and the 
Fig. 2B) ar EIE of old South 


medicam groups, while the tanagers (Thraupidae, 


Tyrannidae. " str., 


2C) represent the large group of nine-primaried 
lonized from North America. These 
groups are more broadly distributed s the old 


oscines that colo 


groups, including those in the savanna biomes. In 
particular, all three have their peak concentration in 
the Andes. The er of patterns shows that the 
intensive radiation ew taxonomie group in the 
Andes (Fjeldsà & RE 2006) did not prevent the 
other two groups from diversifying in this same area. 
The tropical Andes region stands out clearly as the 
center of avian diversification during the Neogene. 


old South 
gh the Neogene, 


To illustrate how diversification in an 
American group proceeded up throu, 
we subdivide the 302 furnariid species into quartiles 
of different branch lengths (Fig. 3). The first quartile 
of species (Fig. 


. many of them in monotypic 


genera, represents lineages with low speciation rates 


or past extinction filters and is mainly found in the 
forested tropics, notably in upper Amazonian terra 
firme forest and on the adjacent humid Andean slopes 
and the northern part of the Andes region, which was 
more or less isolated by marine transgressions in the 
early Neogene. The second quartile (Fig. 3B) gives a 
similar pattern, although with more species associated 
with areas that once comprised extensive swamp 
forests, mangroves, and palm savannas fringing marine 
incursions in the Amazon Basin. The more derived 
species (third and fourth branch-length quartiles, 
Fig. 3C, D) are particularly strongly represented in 
the grassland biomes of the La Plata and Rio Negro 
basins and along the Andes, with the third quartile 
mainly in the dry-based southern and central Andean 
highlands and the most terminal branches (fourth 
quartile) concentrated in the tropical East Andean 
ote, in Figure 3D, that 
umid Andes to 
river island habitats along the main Amazonian river 


cloud-forest zone. We also 
there is a certain spillover from the 


channels. 


THE ANDEAN HOTSPOT AND ITS LOCAL HOT POINTS 


The species richness pattern for range-restricted 
Andean highland birds (Fig. 4) shows marked hot 
points within the tropical Andean hotspot. This i is less 
apparent in other studies of avian endemism in the 


Andes (e.g., Stattersfield e et al., 1998), which ae 


species but p 
aggregated (nested) distribution. Particularly strong 
aggregates of range-restricted species are found on the 
humid slopes in southern Chocó (Colombia), near the 
the East Andean cloud forest is 
deeply intersected by valleys (e.g., to either side of the 


equator and where 


Annals of the 
Missouri Botanical Garden 


Piene B Species Homes variation of Furnariidae tomar and woodcreepers) divided by branch-length quartiles, 


pecies that 


est nodes away from the root of the phylogeny and the fourth 


25% of t t are the fe 
being the 25% of the species i aaa he longest (terminal bandes Red cells represent maximum species richness; 


white cells represent no species prese 


North Peru Low, and in places where the eastern 
Andean oe is ee intersected by valleys in 
Huánuco [Pe d southern La Paz 
[Bolivia]. Fairly ‘high values are also found deeper 
into the Andes, where distinctive mist zones and cloud 
forest can be found on the transition between warm 
highlands (e.g., around the 
y, Colombia, on ms eastern slope of 
Cordillera lilius in Apurimac—Cuzco, Peru; and 
around the Cochabamba Basin, Bolivia). Many of the 
ndemie species in these areas are associated with 
environments near the tree line and with relict patches 
of high-altitude woodland deeper into the highland. 
Typically, the most closely related species inhabit 


no obvious barrier), while species replacing each other 
in different elevation zones on the same slopes (see 
Krabbe & Schulenberg, 1997) are often more distantly 
related (review in García-Moreno & Fjeldsá, 1999). 
THE VARIATION IN AVIAN DIVERSITY ON AN ANDEAN 
ELEVATIONAL GRADIENT 


The local variation in biodiversity parameters 
within an Andean hot point is illustrated in Figure 5. 


This elevational transect, from the lowlands up to the 
ridge of the Cordillera Tunari and over into the rain 

hadow area in the Cochabamba Basin, has altogether 
731 species of birds: 668 recorded on the 
and 123 on the slopes of the adjacent rain shadow 


humid slope 


basin. The species richness curve (full line in Fig. 5) 
suggests a maximum of 411 species around 500 m 
elevation and a gradual decline, following a concave 
curve up through the massively forested midslope to a 
relative plateau, or shoulder, d to the 
ere the 
species richness drops sharply up to tie barren top 


itat mosaics around the tree line 


ridge. In the rain shadow zone, the species richness is 
lower. The kind of pattern with a diversity peak on the 
lower slope seems typical of elevation gradients 
(Rahbek, 1995), although there is some variation 
between groups in the position of the diversity peak on 
the slope (Jørgensen € León-Yánez, 1999; Kessler et 
al., 2001) 

The endemism (mean range-size rarity per species, 
a different pattern, 


stippled line in Fig. 5) describes 


with very low levels (widespread species) at the 

of the Andes and high levels in the montane forest. 
The peak is in the tree-line zone, which is at 3400 m 
n a few places without human impact but is mostly 
depressed to below 3000 m as a consequence of 


Volume 96, Number 3 Fjeldsá & lrestedt 
2009 Diversification of South American Avifauna 


e 4. Species richness in Hn p Dies E eeoempiucel grid) for 470 species representing the lower range-size 
Mes “25% of species with h American birds. Red cells represent maximum species richness; 
white cells represent no species present. 


Annals of the 
Missouri Botanical Garden 


Humid Andean slope 


| 
| 


length 


100 Mean 200 endemism 300 


5 Mean branch 10 


Adjacent 
rain shadow slope 


ridge 


1000 m 2000 m 


3000 m 


4000 m 3000 m 


Figure 5. Species richness and endemism along an i hd gradient in the Andes: Carrasco National Park in Bolivia, 


8 
from lowlands to the highest ridge of Cordillera Tunari and in 


-line on , with 


o the adjacent Cochabamba Basin. The tree A 


and 


K. Herzog, Min d To compensate for undersampling in coca areas in the 


ope, we use x inte ope aa dd de on hi. storical records 
" d 


local survey results compiled by S. 

foothill ery p terrain with difficult access on TE 

rom the area and the known elevational | p 

size md of grid cells in th , as per s 
T: er of n 


by the er author. 


incessant burning to create fresh pasture (Kessler & 
Herzog, 1998). A high mean endemism is also found 
at 3000-3500 m in the adjacent rain shadow area, 
where tiny woodland patches can still be found in a 
highly degraded matrix of bushy habitats 
bunchgrass 


and 


To assess whether the high endemism score at the 
tree line is a simple consequence of the limited areas 
that are available or reflects a higher rate of 
speciation, we examined the variation in branch 
length (number of nodes from the root, for Furnari- 
idae) along the elevational gradient (thin line in 
Fig. 5). Species representing recent radiations are 
present both low and high. Nevertheless, the mean 
branch-length values produce a clear pattern with a 
predominance of deep branches in the foothills and 
terminal branches representing more recent radiations 
in the upper montane forest and on the transition 
toward the barren highland. Only 54 Furnariidae 
species were recorded along this gradient, and the 
sample size is particularly low (five species, with 
moderately long branches) at the barren top ridge. 
However, scrutiny of other large taxonomic groups 


MI DUE CAS the e length (thin line) is the 
ped 


odes on the phylogenetic branches of the a furnariid species, based on phylogenies develo 


(with good, although not complete, phylogenetic data} 
suggests that the depicted curve may be fairly typical. 
Thus, the many range-restricted biological species 
and sharp spatial replacements (with very few 
reported cases of hybridization) in the tree-line zone 
reflect exceptionally high rates of completed specia- 
tion in this zone 


DISCUSSION 


In agreement with other studies (e.g., Fjeldså, 1995; 
Hawkins et al., 6), we 


lowland rainforest was dominated by species repre- 


ound that the tropical 
senting basal lineages with a history of little apparent 
diversification. Paleobotanical evidence suggests that 
a species richness comparable with contemporary 
tropical rainforests extended to Patagonia during the 
(Wilf et al, 2003), 


suggesting that the current diversity pattern is a 


Eocene climatic optimum 


consequence of range contraction and niche conser- 
vatism in old clades. Thus, extinction filters (e.g., 
during the rapid Eocene/Oligocene or Plio-Pleisto- 
cene cooling) may explain why these groups are now 


Volume 96, Number 3 
2009 


Fjeldsá & lrestedt 
Diversification of South American Avifauna 


mainly restricted o the tropics, but lack of opportu- 


widespread species (Rahbek et al., 2006: fig. le) and 
ancient relictual taxa (Fig. 1A) are found in flat and 
swampy landscapes, the distribution of which has 
shifted much in geological time as a consequence of 
changes in drainage pattern and marine ingressions 
(Albert et al., 2006) and rapid sedimentation cycles. 
We must therefore assume that these species persisted 
by moving around and taking advantage of the patch 
dynamics and high productivity of the floodplains. 
Other rainforest groups continued to diversify, but 
ird in terra firme forest, and with slow speciation 
rates 0 
ned with small distributions follow | other 
geographical patterns, with remarkable aggregates in 
topographically structured parts of the tropics (Orme 
; Hawkins et al, 2006; Rahbek et al., 
2006). Because of a lack of a well-calibrated local 
to link 
distributions and evolutionary history are imperfect; 
we still et b 


molecular clock, our present attempts 

use node numbers as a relative 

measure of recency of speciation. Yet our data provide 

strong evidence of high rates of speciation in areas of 

igh endemism, both on the large scale (Fig. 3) and on 
the elevation gradient in the A (Fig. 5) 

The strong radiation of tanagers (and other colonists 
from North America) in the Andes did not seem to 
constrain the rate of radiation on the old endemic 
outh American groups (Fig. 2). Similar uM of 


E 
a 


endemism in other well-charted groups (e.8., w 
potatoes, Hijmans & Spooner, 2001) Me ih 
nters of intensive spec 

wed by oe 
and establishment of dense niche packing, for 


existence of local c 
the Andes. ae is follo 


instance on elevation gradients (García-Moreno & 
Fjeldsa, 1999), which 
accruing more and more species 


means that the ecosystem is 


Some students of lowland faunas may object to this 
simple picture and claim that the use of a broad 
biological species concept nithology underesti- 
mates the degree of differentiation of the 
avifauna. The variation in mitochondrial DNA of some 


Amazonian 


widespread South American birds an bs 
evolutionary entities (e.g., Joseph et al., o- 
vette, 2004) 


geographical populations represent species, or wheth- 


ut it is still unclear to what extent mes 


er the results reflect the pattern of sam pling or an 
unusual evolution: 
and the influence of selective sweep 
al., 2006). Admittedly, some widespread lowland 
species should be split into component phylogenetic 
species, but a full revision of t outh American 
avifauna would certainly also reveal a need for 


splitting Andean species. It is unlikely that such a 
revision would alter the overall pattern. 

Intensive speciation in mountains is found only at 
low latitudes (Rahbek et al., 2006; Hawkins & Diniz- 
Filho, 2006; Storch et al., 2006; Weir, 2006; Davies et 
al, 2007). Even at the moderate latitude of 27°N- 
31°N in the Himalayas, species richness is mainly 
built up by colonization from outside rather than by 
local differentiation (Johansson et al., 2007). At high 
latitudes, similar broad distributions are found in 
mountainous and flat areas (Hawkins & Diniz-Filho, 

006). Low seasonality in the tropics means reduced 
seasonal overlap in thermal regimes between low- and 
high-altitude sites, which is why the 

species could adapt to local conditions (Ghalambor 


individual 


et al., . This leads to dense altitudinal stacking 
of species (Krabbe & Schulenberg, 1997; 
Moreno & Fjeldsá, 1999) and complex patterns of 


García- 


endemism (vicariance, leading to added diversity on 
i k 


coarse-scale geographical analyses; Rahbe 
Graves, 2001), in contrast to the wide distributions 


on high latitudes—even for highland birds (Weir, 
2006) 


A pertinent question now is whether speciation in 
the Andes (or other tropical mountains; Fjeldsá et al., 
2007b) is (1) a specific Plio-Pleistocene phenomenon 
or (2) a general tendency for diversification to start in 
topographically structured landscapes but being 

by redistribution as climate-driven range 
dynamics gradually push the species diversity toward 
the lowlands, leading to the general correlation with 
available energy, water, and area (Storch et al., 2006). 
Figure 3D seems to indicate a spread of furnariids 
from the Andes along the Amazon (see García-Moreno 
et al., 1999, for a case of an Andean radiation leading 
to oun of Amazonian river island habitats). 
Similarly, the large radiation of Tangara Brisson in 
South American rainforests started in the Andes 
(Burns & Naoki, 2004). The furnariid genus Cinclodes 

. R. Gray, which radiated in the barren Andean 
highlands, gave rise to two independent colonizations 
(with speciation) of coastal habitats (Chesser, 2004). 
Yet, the evidence for montane areas acting as species 
pumps is still scant, and well-resolved phylogenies of 
species-rich groups are needed to fully evaluate this 
interpretation. 


FACTORS DRIVING THE DIVERSIFICATION PROCESS 


Traditionally, speciation in the Andes has been 


explained as vicariance (involving observed or 
hypothesized past barriers; e.g., Vuilleumier, 1980; 
grgensen et al., 1995). Weir (2006) suggests that the 


intense Bone speciation in tropical mountains 


al 


could be a result of the marked dispersal barriers 


Annals of the 
Missouri Botanical Garden 


created by glaciers and severe elevation shifts of 
montane vegetation zones in connection with Pleisto- 
cene glacial cycles. Weir’s analysis may have been 
biased by the fact that most of the molecular 
phylogenies he used covered species-rich genera, 
and he overlooked the fact that many birds of the 
barren puna are widespread and represent small or 
even monotypic genera, suggesting that the range 
dy namics caused by glacial cycles may erase 
orical population structure rather than promote 
eue (see Jansson & Dynesius, 2002). 


Figure 5 suggests that the highest incidence of 


line, and he linked this with the high risks of random 
extinction of local populations within this very narrow 
habitat band. It should be emphasized here that the 
tree-line zone is mostly 1 km wide but extends along 
more than 4000 km. We may not need glaciers to 
break up the populations of birds in such a habitat 
band. General environmental instability is enough to 
lead to relict distributions of some lineages, and thus 
inevitable divergence of remnant populations by 
genetic drift. 


To understand the mechanisms of speciation, we 


a 


should ea focus less on physical barriers an 
onditions that allow loca 
ia to eer despite global climate instabil- 


source 
ity. This could, in turn, allow the development of 
complex and co-adapted biological communities in 
places with a benign local climate (as opposed to 
highlands, where communities change incessantly as a 
consequence of range dynamies caused by climate 
change). Fjeldsá et ki (1999) demonstrated that peak 
concentrations of endemic birds at the East Andean 
tree line correlate with low interannual variation in 
ground-level climatic conditions, as documented by 
long series of satellite images. Unfortunately, remotely 
sensed data of montane habitats are difficult to 
pad and climate models (based on n us 
o a from an insufficient. networ weat 

n are far too crude to identify PY local 
environments, such as dco mist zones on the 


us. A geographical correspondence between 

lues of young endemic species and older 
relictual forms in the Andes (Fjeldsa, 1995) pro- 
vides indirect evidence for linking speciation with 
local persistence in stable environments. Still, a 
thorough evaluation must await detailed analysis of 
historical population structures, but a sufficiently 
dense sampling is still lacking for Andean birds 
et al, 2005, 


(see Bowie for Afromontane birds, 


however). 


In this context it is interesting to note that high 


suppor ing data from 
sampled sites only) with ancient cultural centers in 
the Andes (e.g. Tomebamba [Cuenca] Chavín, 
Ayacucho—Cuzco, Cochabamba) (Fjeldsà, 2007). A 


positive correlation between biodiversity and popula- 


strong statistical 
i l 


tion is well known at coarse geographical scales (e.g., 
Cincotta et al., 2000; Balmford et al., 2001; Araújo, 
2003; van Rensburg et al., 2004; Luck, 2007), but the 

ean study is on a finer scale (15° X 15^), and it is 
interesting to note that the biodiversity/population 
relationship is stronger for pre-Colombian population 
centers than for present population patterns, which 
may be more influenced by external drivers. 

The simplest explanation would be that special 
local conditions in some Andean valleys governed 
geographical patterns of persistence of bird popula- 
tions as well as patterns of population growth, albeit at 
different timescales e 2007). It is a that 
the local conditions that stimulated dev of 
rich (and ic co-adapted) biological communities 
F thus the cladogenetic process, see Jetz et al., 
2004; Rahbek et al. 
eio This may, in turn, have been 


, 2006) may also have meant crop 
a major 
prerequisite for the transition from hunter-gatherers to 
resident farming systems, and for the further advance- 
ment of agriculture and development of inis 
centers in certain mountain basins, such as Tom 

bamba, Cochabamba, and the Ayacucho Valley 
(Fjeldsa et al., 1999; Fjeldsà, 2007). The distinct 
altitudinal zonation 
blies) may also 


characteristic of Andean cultures, with specific crops 


(with distinct biological assem- 
have facilitated the vertical ecology so 


in specific zones. 


IMPLICATIONS FOR CONSERVATION 


The effort to preserve biodiversity has B 
focused on the wilderness; in other words, the 
bi Fjeldsa, 007). 0 
highest neces in 


unspoiled nature as concep 
a coarse spatial scale, the 
South America is found in the nen of the sub- 
Andean zone and on the humid Andean slopes that are 
relatively unaltered by humans due to their unsuit- 
ability for agriculture, grazing, and forestry (Ortiz, 
1975; Young & Leén, 1999). Because the cloud-forest 
zone is universally accepted by planners to be 
inappropriate for large-scale RUN 
national parks were established 
the effect of a moderate dishusbance by scattered 
small farms may not be very different from the patch 
dynamies caused by frequent landslides on these 
steep slopes. The conservation situation may only 
critical in the Andean forelands and foothills, which 


Volume 96, Number 3 
2009 


Fjeldsá & lrestedt 
Diversification of South American Avifauna 


are Pn to intensive exploitation, notably the 

g of c and "p us ogging in Peru 
cm 1998; Fielded et al., 
highest alpha diversity, oe on ME öiker hand most of 


is zone has the 


the species are widespread and therefore likely to 
survive bn local En loss (Fig. 5). 
Our r 


a on inac 


sults suggest that a conservation strate 
Rec tracts of Andea 
will not be enough. If we are to succeed in slowing the 


n cloud forest 


loss of open we also need actions where people 
live, in the montane valleys, and in the transition 
toward > 


Here, the 


ys 
cloud forest, where people graze their 


cattle. primary vegetation is almost 


of fire to renew pastures (Kessler & Herzog, 1998). 
Fortunately, many endemie birds persist remarkably 
well in small remaining woodlands, and range- 
restricted plants are also to a large extent associated 
with degraded habitat or small patches of special 
(azonal) habitat (Kessler, 1999). 

Finding the right solutions requires, first, that we 
understand the processes underlying the observed 
patterns of biodiversity and human settlement. The 
idea of speciation through formation of persistent 


cant implications. Notably this implies deterministic 
processes and development of functional assemblages 
of organisms that may, in turn, mold the ecological 
processes (Daily, 1997). One of the most prominent 
ecosystem services provided by Andean biodiversity 
is the water supply from the elfin forests in specific 
misty places, situated e e zones identifie 


at hotspots for avian pice ut Suc in forests 


represent a distinctive sclerophyllic cu type, 


d that have the ability to comb water out of the mist. 
Once established, such forests affect the surrounding 
environment and the livelihood for people living in 
adjacent valleys (e.g., Campos & Calvo, 2000; Becker 
et al, 2005). During the past 15 years, a rapidly 
growing research field suggests that increased species 
richness may enhance ecosystem productivity, 
drought resistance, and resilience (e.g., Kinzig et al., 
2001). Although this evidence is still mostly based on 
small experimental plots, - o seems 

ig elevant in relation e management of 
A parts of the Andean e which now appear 
as overgrazed bunchgrass steppe. 

If the correlation between population and biodiver- 
sity in the Andes is more than just two independent 
effects of suitable environmental settings (Luck, 2007) 


and if functional links exist (biodiversity enhancing 


livelihood for people), then we have a potential win- 
win situation for biodiversity conservation and 
development. However, in order for this to be realized, 
conservationists, developers, and local communities 
must come to terms; biologists must provide precise 
advice about how nature can help people achieve a 
better life; and politicians must be willing to pay the 
cost of setting aside essential tracts of land and 
support innovative development for ae production 
on some parts of the land (Green et a ). 
Unfortunately, the discussion pum os to focus 
conservation efforts is full of rhetoric of little scientific 
value when organizations focus on funding points 


a 


e.g., for the Andean Dispersal Corridor, or before that 
for the “green lungs” of the Amazon rainforest). In this 
context, the small pockets of remaining tree-line 
habitat in the Andes have been rejected as “the living 

ead" and therefore not worth saving (Ibisch et al., 
2005: 84). We have tried to illustrate that these areas 
are indeed sites of important processes, both with 
respect to the diversification process and to suitable 
conditions for people. T omotion of conservation 
in these areas is outside the scope of our paper but 
nonetheless will benefit from insights developed from 
consideration of patterns of biodiversity and their 
underlying processes 


Literature Cited 


Albert, J. S., N. R. Lovejoy & W. G. R - Crampton. 2006. 
Miocene t ict and the separation of cis- and trans- 
Andean river basins: Evidence from Neotropical fishes. J. 
S. Amer. Earth Sci. 21: 14-27. 

Araújo, M e coincidence of people and 
Biadivanign? in Europë: Global Ecol. Biogeogr. n 5-12. 

Balmford, A., J. C. Moo 


xd Science 312: 570—572. 


and head at Loma 


D 14: 2695-2707. 
1 . J. Fjeldsá, S. J. Hackett, J. M. Bates & T. M. 


Crowe. 2005. Coalescent models reveal the relative roles 


2000. Corren for 
environmental apo Hrom moa forests. Pp. 26-27 in 
M. Agenda (editor), M orld: equ Forest 
and Sustainable Devel Mountain Forum, Bern. 

Chesser, R. T. 2004. Systematics, evolution, p biogeogra- 
phy of the South American ovenbird genus Cinclodes. Auk 
121: 752-766. 


Annals of the 
Missouri Botanical Garden 


Cincotta, R. P., J. Wisnewski & R. Engelman. 2000. Human 
population in the biodiversity hotspots. Nature 404: 
990—992. 

Daily, G. C. 1997. Nature's Services—Societal Dependence 
on Natural Ecosystems. Island Press, Washington, D.C. 


, T.-S. Ding, P. C. Rasmusse : 

Owens, T. M. Blackburn & K. J. Gaston. 

2007. Topography, energy E the global distribution of 
bird species richness p . Soc. London, Ser. B, 

Biol. Sci. 274: 1189-1] 

Fjeldsá, J. 1995. al patterns of neoendemic and 
older relict species of Andea a birds: nod gd 
canc ecolo y stable areas. Pp. 89— n. S. P. 

Churchill, H. Balslev, E. Forero ay L e. (editors), 

Parag te and Conservation of ees Montane 

Forests. New a ical Garden, B 

"3007. w broad-scale studies of patterns and 
processes mande conservation planning in Africa 
Conserv. Biol. 21: 659-667. 

& N. Krabbe. 1990. Birds of the High Andes 
Zoological Museum, University of Copenhagen, and Apollo 
Books, Svendborg, Denmark. 

. Rahbek. 1998. Continent-wide conservation 

priorities and diversification processes. Pp. 139-160 in G. 

M. Mace, A. Balmford & J. R. Ginsberg (editors), 

Conservation i anging World. Cambridge University 

Press, pe Ped Kingdom. 

& 2006. Diversification of tanagers, à 
species rich bise group, from lowlands to montane regions 
of South America. Integr. Comp. Biol 46: 72-81. 

B. Mertens. 1999. Correlation between 

éndenism and local ecoclimatic stability documented by 

comparing Andean bird dis lue s remotely sensed 

land surface data. hod 22: 6 

. Álvarez, J. M. joies m lon. 2005a. Mlicit 


ronx. 


. Lambin & 


: Molecular 

data reveal some major jr adaptational shifts in the early 

dw on of the most diverse avian family, the Furnari- 
-13. 


e. J. Ornithol. oe 


A. Jonsson, J. 1. Ohlson & P. G. 
Bison. 2 2007. ` Phylogeny of the ovenbird genus 
Upucerthia: A case of independent a ations for 


puc 

terrestrial life. Zool. Scripta 36: "133-M 
ohansson, L. G. S. L Parm & R. C. K. 
2007b. Dire eon of African greenbuls in 

and time: Linking ecological and historical 
processes. J. Ornithol. Mu 2): 359-367. 

García-Moreno, e & J. Fjeldsá. 1999. Phylogeny and re- 
evaluation of species Me in the genus Atlapetes based on 

mtDNA sequence Hs Ibis 141: 191-207 

Arctander & J. Fjeldsá. 1999. A case of rapid 
diversifioation i in the Neotropics: Phylogenetic relation- 
ships among Cranioleuca pcs (Aves, Furnariidae). 
Molec. De cee. Evol. 12: 

Bea C. K., R. B. Hue 


Bowie: 
Space 


a R. Martin, ]3J; Tewksbury 
. Wang. 2006. Are tain passes higher in the 
iic Janzen's hypothesis revisited. Integr. Comp. Biol. 
: 5-1 
Graves, G.R . 1985. Elevational correlates of speciation and 
pes eographic variation in plumage in Andea 
forest n Auk 102: 556-579. 
988. Linearity of geographic range and its possible 
effect on the population structure of Andean birds. Auk 
105: 47-52. 


Green, R. E., S. J. Cornell, J. Scharlemann & A. Balmford. 
2005. Farming and the fate of wild nature. Science 207: 
550-555. 

Hawkins, B. A. F. Diniz- € 2006. Beyond 
Rapoport’s rule: Evaluating range size patterns of New 
World birds in a Sp nn joie cg Global Ecol. 
ae 15: 4 

m m A. Soeller. 2006. Post- 

ange, niche conservatism, and the 

E. dinal diversity gradient of New World birds. J. 
Biogeogr. 33: 770-780. 

Hijmans, R. J. & D. M. Spooner. 2001. Geographic 
ESTO of wild potato species. Amer. J. Bot. 88: 


ocene "EM p 


e p a d owic cki, N. Araujo, R. Müller & S. Reichle. 


e Gap: Creating Ecologically 
Representative Protected ru Systems. CBD Secretariat, 
Montreal. 


E M., J. F a, U. S. Johansson & P. G. P. Ericson. 

peau pareri and biopeders phy of the 

cR en e ae (Aves: Passeriformes). Molec. 
Phylogen. Evol. 2. 512. 


———À ae G. P. Ericson. 2004. 

Phylogenetic relationships of typical antbirds (Thamno- 

rae n test of incongruence based on Bayes factors. 

vol. Bi - 4, 23: 1-16. 

——— . P. Ericson. 2006. Evolution of the 

ovenbird- e in assemblage (Aves: Furnariidae)— 
ajor shifts in nest architecture and adaptive radiation. J. 

Avian Biol. 37: 261—272. 


Jansson, nesius. 2002. The fate of clades in a 
world of recurrent climatic change: Milankovitch oscilla- 
tions and 1 ] Rev. Ecol. S 


Johansson, U. S., 
Sundberg & T. B oe 2007. Build- up of the Himalayan 
Aa, through immigration: A biogeographical analysis 
of the Phylloscopus and Seicercus warblers. Evolution 61 

Jønsson, K. A. & J. Fjeldså. 2006. A phylogenetic supertree 
of oscine passerine birds (Aves: Passeri). Zool. Scripta 35: 
149-186. 

Jørgensen, P. M. & S. León-Yánez (editors). 1999. Sea d 
of the Vascular Plants of Ecuador. Monogr. Syst. Bot. 
Missouri Bot. Gard. 7 

, C. Ulloa, R. Valida & J. E. Madsen. 1995. A 
foris analysis of the high Andes of Ecuador. Pp. 221— 
237 i urchill, H. Balslev, E. Forero & J. L. 
Lu e (edit ors), Biodiversity and Conservation of the 
Neotropical Montane Forests. New York Botanical Garden, 
aA 

Joseph, L., T. Wilke, E. Bermingham, D. Alpers & R. 
anie 2001. Towards a phylogenetic framework for the 
evolution of shakes, rattles, and rolls in Myiarchus tyrant- 
flycatchers (Aves: Passeriformes: Tyrannidae). Molec 
Phylogen. Evol. 31: 139-152. 

Kessler, M. 1999. Plant species richness and endemism 
during natural landslide succession in a perhumid 
montane forest in the Bolivian Andes. Ecotropica 5: 


123-136 

K. Herzog. 1998. Conservation status in Bolivia 
of timberline habitats, elfin forest and their birds. Cotinga 
10: 50-54. 


Volume 96, Number 3 
2009 


Fjeldsá & lrestedt 
Diversification of South American Avifauna 


———, — J oa & » Bach. 2001. Species 
Heine and endem of plant and bird co mmunities 
along two gradients of ae ation, mers , and lan 
the D Andes. Diversity Disibuios 7: 61— 77. 

Kinzig, A. S. W. Pacala & Tilm: 2001. The 

Fu ae Consequences of Biodiversity. Pence Uni- 


e, N. & T. S. Schulenberg. 1997. Species s an 
natural history of Scytalopus tapaculos (Rhino mid 
m descriptions of the cua cs taxa, done thre 

species. Ornithol. M 48: 47-88. 
MA I. J. 2004. Molecul P a teeny and plumage signal 
evolution in a trans Andean and 
pet es Molec. Phylogen. Evol. 32: 512—523. 
,G. W. . The relation ships between net primary 
oOo po an population density and species 
2 


circum Ámazonian avian 


212. 

3 : N. D. Burgess, L. A. 

TEN C. Rahbek & P. H. Williams. 2003. Performance of 
ub-Saharan vertebrates as ir a € for vu 

was areas for conservation. Conserv. Biol. 17: 218. 

989. bord ES Hotspots" 
forests. ‘(ony ironmenta list 8: 1-2 

— ——. 1990. The Advent ies Expanded hot- 
spots analysis. d Up 10: 243-256. 

po J., J. Fjeldså G. P. Ericson. 2008. Tyrant 

ycat teher rs ep out in p open: Ecological radiation in 


in P 


Passeriforme: ool Scripta 37. 5 
Orme, C. 2 , R. G. Davies, M. Burgess, F. Eigenbrod, N 
Pickup, son, A. J. Webster, Te . Ding, P. 


rn, K. J. Gaston & I. E Owens: 2005. 


ess are not congruent with 


patterns of avian species melee Proc. Natl. Acad. Sci. 
U.S.A. a 4534-4539. 
N. J. Gotelli, R. K. Colwell, G. L. Entsminger, T. F. 
L. v. B. Rangel & G. R. Graves. 2006. Predicting 
continental- era po E bird species richness with 
spatially explicit models. Proc. Roy. Soc. London, Ser. B, 
4 


Biol. Sci. 274: 1 E 
Remsen, J. V. 2003. Family Pu ers 
Pp. 162-357 in J. del Hoyo, Elliot & D. Chris 


(editors), Handbook of the RR e the World, Vol. d 
BirdLife ro PER Cambridge, United Kingdom, and 
Lynx, Barcelon 


Ricklefs, R. E. 2003. Global diversification rates of 
TE birds. Proc. Roy. Soc. London, Ser. B, Biol. 
Sei. 2 85-2291. 

Sible Ahlquist. 1990. Phylogeny and 


Classification of Bi rds: A Study of Molecular Evolution. 
Yale University a New Haven 
Stattersfield, A. J., M. J. Crosby, A. J. e D. C. Wege. 

. Endemic Bird Areas of the World: Priorities for 
Biodiversity Conservation. BirdLife Mica Cam- 
bridge, United Kingdom. 

Storch, D., X. Davies, R. C. Dajícek, C. D. L. Orme, V. 
Olson, G. H. Thomas, T.-S. Din: Rasmussen, R. S 
-0 


Energy. d range dynamics and global 
species ute patterns: Reconciling mid-domain effects 
tal determinants of avian diversity. Ecol. 


in montane avifaunas in eotropics: 
161- 180 j in G. 
Mace, A. Balmford € J. R. Ginsberg (editors), Conserva- 
tion in a Changing sien Cambridge University Press, 


. Bates. 2007. Molecular phylogenetics of 
the tody- tyrant and flatbill d of tyrant flycatch- 
ers (Aves: Tyrannidae). Auk 124: 


van Rensburg, B. J., B. F. N. inii A. S. val an Jaarsveld, 


Afr. J. Sci. 100: 266-2 
a F. 1980. Speciation in birds of the mg Andes 
2 E Int. Ornithol. Berlin 2: ai 

m y 006. Divergent timing and pati un cies 

Bier in lowland and highland eed birds. 
AA 60: 842 

Wilf, P., N. R. E ane: K. R. Johnson, J. F. Hicks, S. L. Wing 

&J. D. Obradovich 2003. High plant diversity i in Eocene 

South America: Evidence from Patagonia. Science 300: 
122-125. 

Williams, P. H. 1998. Key sites for conservation: Area- 
selection methods for biodiversity. Pp. 211-249 in G. M 

ce, A. Balmford & J. Ginsberg (edit 

n a Changing World. Cambrid, 
Lr United Kingdom. 

Young, K. R. & B. León. 1999. Peru's Humid Eastern 

rests: An Overview of Their Physical Settings, 

Biological Diversity, Human U d Settlement, an 

Conservation Needs. No. 5, 

National Environmental Research Institute, Rønde, Den- 

mark. 


ors), Conservation 
ge Uis Press 


Montane > 


IMPLICATIONS OF GENETIC 
DIFFERENTIATION IN 
NEOTROPICAL MONTANE 
FOREST BIRDS! 


Jason T. Weir? 


ABSTRACT 


complex geography of the Neotropical montane system is a natural laboratory for population divergence. Understanding 


which geographic barriers (lowland barriers, arid river valle: 


diversity and endemism within the r 
endemism for 43 co-distributed l 


ys, and montane barriers n the tree li te tl 
of endemism) are instrumental in promoti ng and maintaining populat: 
d region. in I analy. ze pattern: 


r tic 
ween 16 predefined regions of 
ntane forest birds. The analysis shows that 


enc 
ns ote genetic 
otropical mo 


lowland barriers ous the highest Mm of genetic reno while barriers above the tree line in the Andes show the 
n divergen he 


least. Within the Andes, arid river valleys promote populat 
greatest effect, but the nee Apurimac and Rio Q 
genetic s ae across other river valleys is generally w 
porting a nonce history 


eak. Most barr 
of dispersal obi udis barrier formation. en rsity 
of endemism would help to ensure ve continued survival of ELS 


vergence to varying degrees. Río Marafión shows the 


If the 


= 
É 
o 
a 
E 
E 
e 
Si 


n Neotropical servation 


SEE lineages within P oci t ] 
E P : " 


d with special emphasis placed on endemism regio 


2 by lowland barriers and by deep intermontane river valle 


Key words: Andes, birds, 


endemism, Neotropical, 


A E river barriers. 


The Neotropies possess the world's most species- 
diverse floras and faunas (Rosenzweig, 1995; Newton, 
2003; Kreft & Jetz, ) In birds, ca. 300 
continental species occur (Newton, 2003), far more 
than in any other biogeographic realm. Almost half of 
Neotropical avian species occur in a complex chain of 
highland regions extending from Mexico to Argentina. 

e chief components of the Neotropical highland 
system are the Andes, extending the entire length of 
South America from Tierra del Fuego in the south to 
Venezuela in the north. A chain of isolated highland 
regions extends northward through Central America to 
Mexico, and additional isolated highland regions 
occur in northern South America (Fig. 1). Thi 
highland system is the largest and geographically 
most complex at tropical latitudes, and its ongoing 
uplift has promoted extensive diversification opportu- 
nities (Weir, 2006; Ribas et al., 2007 

ver the past 10 million years, extensive orogeny 
occurred throughout the Neotropics (Gregory-Wod- 
. 2002; Audemard, 2003; 


05). Decio occurred in 


zicki, 2 


Neotropical ud birds throughout this period, 


with little indication of a slowdown in diversification 
rates (Weir, owever, Neotropical lowland 
faunas experienced a slowdown in diversification rates 
over the same time period (Weir, 2006). The contrast 
suggests that highland regions are key cradles of 
diversity in the Neotropical region and their preser- 
vation is of utmost importance. 

Unlike lowland wet forest species, which generally 
high degree of local 


endemism has resulted because of the rugged terrain 


occupy large geographic ranges, a 


and geographic complexity of the Neotropical high- 
land system (Gentry, 1992; Stattersfield et al., 1998; 
Valencia et al., o-distributed species com- 
plexes often exhibit similar patterns of endemism. 
This finding has led several authors to define common 
patterns of endemism in ias i montane birds 

taxa (Cracraft, 1985; 
1996; Stattersfield et al., 1998). The 


boundaries of many highland regions of endemism 


shared across a wide diversity of 
totz et al., 


coincide with known or suspected geographic barriers 
believed by most authors to reduce or prevent gene 
1969; O'Neill, 1992). These 


intervening lowland 


flow (Vuilleumier, 


boundaries include regions 


1A special thanks to Trevor Price, Peter Jorgensen, and two anonymous reviewers whose suggestions helped improve this 
d Ocht 


ena and J. G. 


eca, respectively. 


eering Research Council postdoctoral fellowship 


2 Department of Ecology and Evolution, University of Chicago, 1101 E. 57th Street Chicago, Illinois 60637, U.S.A. 


jtweir@uchicago.edu. 
doi: 10.3417/2008011 


ANN. Missouni Bor. Garp. 96: 410—433. PUBLISHED ON 28 SEPTEMBER 2009. 


Volume 96, Number 3 
2009 


Weir 
Neotropical Montane Forest Birds 


El 
Fast Mexican 0 to 1000 meters 
d 1000 to 2000 meters 
NUR AC western sub-region > 2000 meters 
A 3 * — 4 eastern sub-region 
West Mexican Aij 
x cf eastern sub-region Santa Marta » 
N db ga 
Guatemalan C > sub-region x A Parian 
t AE . LL 
= 1 RD 
m 
Talamancan r \ 
— v PP ux. 
Darien > Rio Quinimart oa la E 
East Colombian V 
> Rio Caquetá y 
West Andean 
Rio Marañón [ 
East Ecuadorian 
Rio Huallaga 
North Peruvian Rio Apurimac 
Central Peruvian 
South Peruvian - Rio Granda 
Austral 
Figure 1. Geographic ranges of regions of endemism are ana lyzed. River valleys and lowland isthmuses separating 


Andean regions of endemism are illustrated. See text for other non-riverine barriers separating regions of endemism. 


(which support a variety of habitat types that differ 
floristically from montane regions), arid inter-Andean 
river valleys, and high-elevation barriers above the 
tree line. However, the exact role of these barriers in 
promoting population divergence and speciation is 

ebatable. Some authors suggested that montane 
forest diversification was primarily associated with 
ecologically stable regions that survived intact during 
recent glacial episodes, rather than as a consequence 
of geologic barriers (Fjeldsa, 1995). While speciation 


in a select group of highland taxa appears to have 
occurred on opposite sides of these barriers, giving 
rise to the observed patterns of endemism, many other 
species are distributed across barriers and occupy 
multiple regions of endemism. Thus, it remains 
unknown what extent these barriers promote 
population divergence in general. 

Here, I analyzed phylogenetic patterns in a set of 
43 Neotropical humid montane forest species com- 


plexes in order to understand how proposed barriers 


Annals of the 
Missouri Botanical Garden 


promote and maintain genetic differentiation between 
netic differenti- 
t high- 


land regions of endemism for both species- level taxa 


regions of endemism. I com 
D 


ation in mitochondrial 


and intraspecific populations in order to assess the 
general usefulness of treating regions of endemism as 
distinet conservation regions. 


MATERIALS AND METHODS 
TAXON SAMPLING 


used ur M species complexes (Mayr & 
2. 1970; Fjeldsá & Krabbe, 1990; Mayr & 
Diamond, 200 y to analyze pl ros 
within Neotropical highland regions. graphic 
species, as used in this study, include mo omen (at 
least in mitochondrial DNA phylogenies) clades 
composed of one or more allopatric or parapatric taxa, 
but do not contain taxa with sympatric distributions 
(other than highly local sympatry along contact zones). 
In a phylogenetic context, zoogeographic species 
represent the largest inclusive clades of taxa within 
genera that lac m 


within zoogeographic species lack sympatry, they 


patric overlap. Because taxa 
retain the geographic signature of the diversification 
process and are ideal units to analyze the role o 
dispersal barriers in promoting diversification. In most 
cases, zoogeographie species include very closely 
related taxa that are considered to form superspecies 
complexes and are likely capable of at least limited 
interbreeding, should migration occur between them. 

species occupying Neotropical 
Table 1), for which 


mitochondrial protein-coding DNA sequences repre- 


All zoogeographic 
montane forest were included ( 


senting two or more highland regions (Fig. l) were 
i zoogeographic 


available. The sample included bot! 
of a single and multiple allopatric 


species compose 
species and spanned a wide taxonomic diversity 
(Table 1). Sampling within zoogeographic species was 
not random geographically, with the greatest effort 
occurring in Bolivia, Peru, Ecuador, Panama, and 
Costa Rica. Samples from Colombia, Venezuela, and 
the tepuis were available for only a few zoogeographic 
species (Table 1). 


ENDEMISM REGIONS 


Areas of highland endemism have been defined for 
Neotropical birds by Cracraft (1985) for South 
America and i 
(1998) Stotz et 
Neotropical region. The regions of endemism defined 
by the three studies correlate closely in most cases, 
although Stattersfield et al. (1998) split a number of 
Cracraft's regions more finely. I studied differentiation 


across a modified classification of these regions for 
montane humid forest (Fig. 1). I did not include some 
i ce data were unavailable, 


regions for which sequen 


and several regions of endemism were split more 
finely into subregions of endemism (see Fig. 1). 
Montane humid forest (also referred to as montane 
evergreen forest) ranges ca. 1000-1500 m (locally 
lower, especially along the Pacific slope of the nort 
Andes and in the Darién Highlands; Gentry, 1986) to 
the tree line, which at tropical latitudes generally 
varies from 3000-3500 m. Humid montane forest is 
characterized by a profusion of moss and other 
epiphytes, which liberally blanket Puri and 
1996). Humid montane 
forest is strongly bisected in the Neotropical region by 


trunks of trees (Stotz et al., 


a series of lowland barriers, arid river valleys, and 
Adjacent 
regions of endemism are almost always separated by 


highland barriers above the tree line. 
one of these barriers. 


PHYLOGENETIC ANALYSIS 


Mitochondrial protein-coding DNA sequences for 
Neotropical montane forest birds were obtained from 
previously published phylogenetic or phylogeographic 
studies or sequenced for this project (Tables 1, 2) 
using standard protocols (Weir et al., 2008). Model- 
corrected genetic distances were compared between 
adjacent highland regions. Parameters of the general 
time reversible (CTR)-I' model of sequence evolution 
were estimated using maximum likelihood in PAUP 
4.0b10 D de s from a neighboring tree 
rooted with o L. Parameters were estimated 
separately for HR gene. Maximum likelihood param- 
eter estimates were used to obtain GTR-I' distances 
between sampled individuals. 

Barriers may promote genetic differentiation. be- 
tween populations in several ways. Barrier formation 
may result in the vicariant fragmentation of widespread 
species into populations on either side of a barrier. If 
gene flow is prevented or greatly reduced, then the 
populations will begin to differentiate in mitochondrial 
DNA markers on opposite sides of the barrier. In such 
cases, mitochondrial sequence divergence can be 
transformed into time estimates of barrier formation 
using an appropriate molecular clock. For species that 
did not span a barrier when it formed, dispersal across a 
barrier might result in the formation of founder 
populations any time after barrier formation. Again, 
provided that gene flow is prevented or greatly reduced 
by the barrier, genetic differentiation will occur on 
either side of the barrier and sequence divergence can 
be converted into a time estimate of the founder event 
using a molecular clock. Caveats arise when leaky 
barriers allow limited amounts of gene flow, which may 


Volume 96, Number 3 
2009 


Weir 
Neotropical Montane Forest Birds 


prevent mitochondrial differentiation after vicariance 
or founder events. The degree of genetic differentiation, 
however, provides an approximate timescale for t 


oO 


cessation of gene flow and may postdate the initial 
vicariant or founder events. 

Based on 84 avian calibrations eee by cross 
ation, an average molecular rate of close to 2% was 
strongly supported (Weir & Schlut e, 2008) for model 
(GTR-I)-eorrected distances of FE mitochondrial 
e b gene. The ae rate applies to iet uri 


validat 


es spanning the past 12 million yea eir 

Selle ter, 2008) and is ie throughout ds study to 
convert Pu -corrected sequence divergence (using 
the same model that is used in the clock calibrations) 
into approximate time estimates of when populations 
began to differentiate in protein-coding mitochondrial 
DNA markers. DNA from cytochrome b was used in all 
but eight of the 44 zoogeographie species complexes 
analyzed here. The remaining eight used either ATPase 
6 and 8 (Henicorhina P. L. Sclater & Salvin) or ND2 
(Hemispingus Cabanis, Ochthoeca Cabanis, Troglodytes 
Vieillot), two protein-coding qid rial DNA genes 
that evolve at similar rates 

Lovette, 2004; Weir et al., 200 


Sequence divergence actually measures coalescent 


o cytochrome b (e.g., 


times, the point at which haplotypes begin to diverge. 
Coalescent times and population splitting times will be 
equivalent if no standing variation occurred at the time 
of population splitting. If ancestral polymorphism in 


variation. If daughter populations randomly fix different 
alleles, then the date at which the fixed haplotypes 
coalesce will predate the time of population splitting. 

e discrepancy between coalescent and population 
splitting dates can be estimated by assuming that 
current and past levels of standing genetic variation 
within populations are 
highland birds, 
haplotype divergence averages 0.4%, suggesting that 


the same eotropical 
protein-coding sishende DN 


coalescent times predate population splitting times by 
only 0.2 million years (Ma) (Weir, 2006). Given that the 
discrepancy is slight, uncorrected coalescent dates are 
ion here. 

methods were used to construct consensus area 
lag of highland regions. The first used standard 
approaches for constructing supertrees (Baum, 1992; 
Bininda- Emonds, 2004). Supertrees are generally 
constructed from different gene phylogenies for a 
contain all taxa. Here, I used phylogenies of co- 
distributed zoogeographie species to generate a super- 
tree for a common set of geographic regions (supertree 
area cladogram [SAC]. Input phylogenies in the 
method may show discordant topologies just as they do 


in the regular supertree method when using topologies 
The SAC method finds the 


topology with the fewest steps required under parsi- 


from different genes. 


mony criteria. This topology should be viewed as a 
compromise that finds the strongest overriding phylo- 
genetic signal in the data set. 

aximum likelihood or Bayesian molecular bp sel 
enies of montane forest taxa were obta from 
published molecular phylogenies or were puse 
from unpublished data sets using standard methods 
Weir, 


species with samples from three or 


=~ 


2006). Trees were included for all Ming ir 
e highland 


regions. Next, I converted dirimere Bes to 
matrix Ves ntation. (Bininda-Emonds 
1998). A 


to a matrix with n rows and n-1 columns (polytomies 


Bryant, 
fully bifurcating tree with n tips is converted 


can also be coded but result in n-1-p columns, where p 
is the number of nodes collapsed into polytomies). Each 
row represents a geographic region and each column a 
node on the phylogeny. All descendent regions for a 
given node are coded as 1, and all outgroup regions to 
that node are coded as 0. Geo raphic regions not 
sampled are coded either as missing if the zoogeo- 
graphic species occupies the region but was not 
sampled, or as O if the zoogeographic species is absent 
from the region in question 

he SAC method may perform poorly by giving 
equal weight to each column in the supermatrix 
irrespective of whether nodes in input phylogenies 
received strong support. One option is to weight each 
column (i.e. 


its nodal support. This option requires that the same 


, input node) in the supermatrix based on 

nodal support metric be used consistently across input 

trees. In this analysis, most input topologies were 
ublish 


m likelihood and 


Bayesian phylogenies, which use different measures 


es. i 
erived from pu ed maximu 


of support, and no weighting was implemented. 
Matrices for each zoogeographic species were 
concatenated, and parsimony analysis was performed 
on the combined matrix. Several types of parsimony 
md have traditionally been performed in supertree 


8 


These ana | ew perform similarly 
(Bixinda- Emonds & San n, 2001; Bininda-Emonds 

2004). However, when pas supertree methods to 
generate consensus area cladograms, reversible parsi- 
mony methods are preferable because they make no 
ed standard 
parsimony (standard matrix representation parsimony; 
Baum, 1992) which allows reversals and local 
extinction. heuristic sea 
erformed in PAUP 4.0b10 

stepwise addition, 


assumptions regarding extinction. I perform 


re tree space was 

(Swofford, 2002) with 
random sequence addition (25 
replicates), tree bisection-reconnection swapping, and 
time. Strict 


a maximum o O trees stored at any 


consensus cladograms 1 from all equally 


Annals of the 


414 


Missouri Botanical Garden 


666T 


“Te 19 ouexopy y xe[duroo 
-eneg tÁpmis sL T I I 0 I I I 0 0 0 I I 0 0 Q9Bquemqorey 09r2]0nID47) 
ZOOS “HPPA T 0 I 0 I 0 0 0 0 0 0 0 0 0 0 0 QXepduroo uosmnenyrog $7724) 
800% ^p? 19M OT I I I I I 0 I I I I I I I [ duo? stueqey snzuidso10py) 
(19Seue] 
-ysnq pereonp-^oqpo4) 
8003 "Te 19 TOM I 0 0 I 0 0 0 T BPPS "['q soman 7) 
(1oSeue}-ysnq poyeor)-Ayse) 
oKgusov] 
8003 “Te 19 TOM 0 0 0 I I 0 I supqmsupo snsuidso10nj» 
(miqewáos 
OS 
Ápms sj, 0 I I 0 I 0 I "Td smpisnd smjdumpiopkdum;) 
(onbroeo urejunour) 
TO0c AUSIqIO,P Y 
tuo&uv'] Ip oud T 0 I 0 I 0 0 oKgusoie'] snjouosKao snoi27) 
SOOT WAS A 
“7007 SON 
*eooc “TE 19 UMUYEN I I 0 I 0 ¡Xodos prog nuasipuy 
(qouy-ysniq popeoy-odins) 
Lup 
L00c “Te 19 vuope) T I I I I I I I T T 0 T I 3 efeusexeT sajonba07 "y 
(uy 
-qsniq peddeo-inujsoqo) 
oKgusov] 
L00c “Te 19 vuopen I I I I I I I T T I T T I puonuiouunig uouoity 
pem SUOLS9A IIOU 10 39M, 
oouo1opor > g o z a E E z E y z y = z e e E a xo[duroo soroods onjpdersoo$007 
onjousso[ Aug E E a & g E E 8 a 2 E s. 5 E 5 E A g 
S = S ae ni ti oO S * = =. as 5 5 3 3 = 
= - = = > S 5 E z TA E E E 3 z 
= E = o 3 £ = 5 5 Zh Ui = 3 3 5 5 a = 
E P y P Ro: 3 F È B £ 8 8 3 5 bk B £ 
E z E z S = > 5 S Ss = 5 5 S S 
MER NEM E M M M 4 "^s y Bg 
E É S PN 2 E > ui 5 o ES = pem 
— 5 = > 5 a S9. T z E! 2 a B 
> Pu. > = T T B a Da pial ds Ea 
E X a E ? 8 £ E 


(ES 01 [S) suoigor UBILISUY imog porejosr pue (gy 01 [y) sepuy otn 


“(LO 01 19) vorreury oppprjy pue oorxopy 10] UMOYs ore p omy ur posn se sopoo ursmuopuo jo suorsoy "| Aq poioso1dor ore po[dures orom yey) oso ‘0 Aq poiuosoxdox ore Apnys snp ur po[dures 


jou mq soroods oryderso9300z e Aq pordnooo uisruropuo jo suorgoy 'Apms snp ur popnqour soxo[duroo soroods otyde1so93007 15910; outjuour peordorooyv Jo suonnqunsrp orjdezsooz) 


T SIREL 


415 


Weir 


Volume 96, Number 3 


2009 


Neotropical Montane Forest Birds 


8661 


"Te 19 ouo -enres I I I 0 I 0 0 0 ,,X9[duroo stueqey vooy 
9661 (querd1-18uo poxoeq-umoxq) 
“Te 19 ouo10p-ero1e2) I 0 0 I I 0 0 IPPS "Tq 40]0onumf y 
8661 Querq-qeyo 
“Te 19 ouo10p-Ero1e2) I I 0 I I 0 pouworo) oKeusoipe srpguo4f O 
(querdy-yeyo 
8661 peypeq-ejs) o&eusoxreT 
“Te 19 ouo10p-erore2 T 0 0 I I 0 0 siuguaajuoumuui DIGOYIYIO 
(uo1wque Ayers) 
ewp oouorurT 
poysyqndun * ea — 0 I 0 0 I 0 0 0 0 0 I 0 I 0 0 4ojoonsnjs DNIMIO UL jy 
(urejspox poyeo1q1-o1e]s) 
GOOG "ueurpzereq 0 I I 0 I I 0 0 0 0 0 I 0 I 0 I 0 I HOSUIEMS SIBIULUI QNT 
GOOG 'ueurpzoroq T I I I I I I I I I T 0 I ,X9[duroo pareg "p^ snaoqonqy 
(oxrenos poyoeq-umorq) 
1007 “Te 19 JEN 0 I I I 1oSou[oeig smppuapi220 "Wy 
200% E X OMN 0 I I 0 I I 0 0 I I I sXepduoo uosuremg $2I59PDA py 
6661 (maeu wenk) 
“Te 19 ouo10p-erore2 I I 0 I I 0 0 0 seSTpporT punprupiaK? DANOJ 
900% 
“Te 19 ouo10p-Ero1e2) I I I I Q,xepduroo uosumeAG snuodumT 
(prrqSuruumq 
9006 pereonpas&moure) uosur Ac 
“Te 19 ouor -enres I I I I smuisapgoum stusoduny 
(U9IM-poom pojseoxq-Ae13) 
9007 “Te 19 puq 0 0 I 0 I I 0 0 0 0 0 0 0 I 0 0 0 Q Ipnyos], sCajdoono] nunjooruo] 
100% 
“Te 19 ouo- — 0 I I 0 I 0 ¿9 xepduroo snguidsnuag] 
1007 sV xepduroo 
“Te 19 ouo- — I I 0 I 0 0 smeqer) snsuidsnuazy 
S66[ NAH 0 0 0 I 0 0 0 0 0 0 0 I T 0 0 0 I ,Xopduoo 1o[8e A 0550/51 
oouo1opor > v Q z a = = < yp F d y = = e e E 4 xo[duroo soroods onpdvexgooS007 
oneuoSo[4uq a & 3 = 4 & & z 2 = £ =. z & 2 2 a 2 
S zd ES ai t a S x 5 2: EN 3 El a 2 = 
= nz = nz > $ o E = pos TA = E 5 5 S zZ 
= 5 5 S 5 E = = = Iz EE 5 5 = = E E 
2 = z I e & 3 S = > ~ = Ed S = = = a 
e E = El $ E = 3 E mm M a 5 5 S S 
S RE NE E M M a 3 E f p 2 A $ 
2 3 2 E 5E 2 UE g a - È 2 QQ 3 
538738 sS 8 9 g * 


penunuo) 'q oqe, 


Annals of the 


416 


Missouri Botanical Garden 


6661 


“Te 19 ouo10p-Ero1e2 0 0 0 T I 0 pg Xe[dumoo nonojoruma) 
(1oSeue) Moj[oK 
L66T -pue-ypv[q) utes W 19S 
‘suing t&pms sty], I I 0 "['d spjawoskiys sidénpos&£n 
(ysna 
9007 -ope3uny3u perq-eSuexo) 
‘Honig Y IJU M qneprey 
“8007 “TP 19 memo 0 0 0 0 0 0 0 0 I 0 0 I STIS OAM PUDO STUND) 
pəojduws suoraa quəvefpe omy 
(pug 
Apmis sI, 0 0 0 I 0 0 0 0 I I I 0 0 0 Q 41818) rpnipos], vonsna ozidsojdog 
(1odooxopoon. 
2006 ponods) 190S "Tq 
“oxto[y ‘Apmis st, I I 0 I 0 0 0 0 smakdonpao smpouaioudiy 
7003 
“Te 19 IPASA (1odooxopoon. 
*e00c ‘xoy perrq-3uons) uossoT 
"peusrqndun roM 0 0 0 0 0 I 0 0 0 0 0 0 0 I 0 I Q  smjupadoí0uo4d soqdojooogdiy 
6661 “TR MH O0 0 I 0 0 0 0 0 0 I 0 0 I 0 I gU X9[duroo 3o[[rotA, sa7ApopFory 
(exuqsque 
$007 “Te 19 jessnr) UTA[ES 29 IPPS 
IPSI] “Apis eni, 0 0 I I I 0 I 0 I 0 0 "pd OGOUN Som 
00 “PORN (1eSvuv] 
39 sumg ‘Apms stu], 0 0 0 0 0 I I I 0 pepyaeds) srueqe?) 070713 I 
v00Z TLN (19Seue] preu) 
29 sumg pms eni, I I I I ULES Y IPS "pd Ppuojf L 
TOOG “PORN 
Ņ sumg ‘Apmis sy], 0 I 0 I 0 0 0 0 I I I a4 V xo[dwoo uossug DIDÍUD 
Qoxed porq 
200€ “Te 19 seqty 0 0 T I 0 I I I I -poi suowuurT snpipios snuoiq 
oouo1opol > yp Q z 4 5 E z v 5 d y = = e e E z xo[duroo soroods omdergo93007 
oneuoSo[4uq El & 3 = pa 2 2 5 El = = B. Em z 2 2 2 E 
E al B a E o Š È E = S 5 5 3 3 
E xe] E 5 > $ o E z 7A m E E 3 3 = = 
— T = T z E = o, Ey BR ui zs 5 > E] E g o 
> Z 3 5 a = z S a N z m S S = = o E. 
Es E g = 2 a El E = ~ d 5 5 S S 
NA I. E Iz. S T EN E S 
= 3. S z =A a EN T & a m a 3 5 
a B =. = E 5 2 Y 5 E E 2 TO 3 
> vnd > e = Em B = prs tid Y = 


“penuryuor) 


‘T 9meL 


417 


Weir 


Volume 96, Number 3 


2009 


Neotropical Montane Forest Birds 


'(qsunp Ajoos) smueqer) suaosa4ru ^p (ysn »xoepq) e&eusoxieT smmosnfui snpang gy 
‘(raSeue) poyon- i ac UTATES 29 IIPS "T “q sasuafostsun ^p '(oSvuei peddeo-xov[q) srueqer) 194194 DIDÍUD ,, 


ds peyeoyo- ou) 191e[9S "Td SISUASHUD. ey a peMorq-Yse) 19P]98 "T "d DIVINO DINAOMDA) yy 

‘(ueim mdoj) stueqeg 
sngnfna > y *(uoxe urequnour) xoje[og "T ^q. 81701745708 y (USIM ee BURG) sgueg DJO91JUOW * y (USIM SNOSIRIYIO) ABMSPIY snaonaugoo p “(USIM pewoxq-snoga) edreug sngpipoofna so1pog2o4y o 
«(193eue] pe[gueds-[Aroq) oXeusoxeT srpraaoastu ^p *(oSeuey pedeu-uoo13) uospo. vsoonf :, ‘(193eue) poysoyo-o[sueds) UTAES map DIDÍUD Y a 
“‘Quer4-jeyp peqpeq-^oqpe4) qepe Duespoip ‘Q (uex&-1eqo peuoxq-uop[03) ULATES W IAJ "T ^q PAPE vooounpop) 1, 

"(espar pojseoxq-uoagres) sdjoyg 29 rounurz muopaw "py *(ueyspox mdoj) stueqey snjymdbooooupispo "pj espor pooej-onym) sdpoyg X sdjoyg s2120/i970 

"Hi ‘Qseyspor eweg) sdpouq X sdpouq apund ‘py *(xeispox pouwoxo-»o[[o4) ugeg xoz:222207/ "py ‘espor posuoxr-uop[o$) neouuosstoq snuro "py *(yreyspox po[oeroods) pros, snjypudaooupjoui 
JW espor peruog-erq) uteg p xeje[og "T "qp suo4fiqpo py (Qxersper peddeo-umorq) Ausiqig.p zm eAeusexe' sdooruunaq "py '(uisper pexeqpo) preg ^p ^g sn;pnbio; snioqo1Apy o 
'(oxreyos peovj-xov[q) utApeg sdounjaw "py *(oxrey[os poues) uospov $712040702 "py *(oxreyos uvopuy) AUSIIO P $9P10]]D4 saysapvATy « 

'(uroS- po ie peprei-&e18) oouorwwT nprmD21242u12 "T '(uroS-urequnour poyeoax-o[damd) urapez Maa iin 7] (uos 
-urejunour pojseoxq-uoo18) ueurpoo) 29 UES 90772q4s "T ‘(UE ij-uoo19) ques 29 1oroznog suogmdipina 7T “(ua UI-911YM) pmo») s: smaodum] y, 
iom ee BA aspe! Ņ oyelg supmnoaodnsofna "gr NUM Opel: 9810) e Siponen snéuidsnuayy , 
‘(snsurdstmap] pemoiq-esuer10) UTATeS W IPS "J `q stuydopna ‘H ‘(sn3uidstwop peddeo-xov[q) eXeusoxvT snondoam snguidsnuag , 
"(199101d1o Mo] Á8n1) oXeusarey 29 KuStqiQ,p sapi01ms ‘q “(1eo1aidiamo[] Leys) stueqey vaqunjd ‘q '((1so1ordlio Mog per[eq-uoureuuto) 198e M Poq 0590121] s 

‘(reyoutds pooey-por) rerepog "Tq Sdoyo 7) *([reyourds mda) uewpos 29 urapeg Dssmuap 7) ‘(Treyourds 

peijsoro-Kureoro) stueqey sjpdpoigpo 7) *(pejourds pouworo-odrnus) jo[[rorA vrydowud 7) (pejoeurds uvrarpog) esprely 29 Ife anomuay 7) (sotoods pue[wo[ '[rejourds osgo) yoequeyotoy 
132j0sqo 7) ‘(satoeds puv[uor *jrejourds paxoeq-Ásn1) upezjeg vuidina 7) *(Teyoutds poumoxo-181]) o&eusoxjeT 29 ÁUBIquo p sdaoiq] 77) “prersuids eyedeo1e]a) 1eumurz anyndnoinw Donojoruvan) , 
‘Gaddip poyeory}-snoyns) erueqer) 1272s. 7) *(roddip poddeo-oyym) rpniposT, snjp1da202n2j 77) *(xoddip ueotroury) UOSUTEMS snupoixaw snjout) e 

'(roSvuv)-usnq oug) uospoN SNIDULOUI 
7) '(198euvj-usnq eunoreoe) uroosu«) IDUNIADID] 7) '(reSeuvj-usnq Aysnp) uta[eg ze IWP "[ "qp snosnfnuas 7) ((x98euvj-usnq uourutoo) sosis op sng nq snonuppunjdo snsuidso40pj7) z 
'(ueonoj-urejunour pepooy) pos o7077on2 "y *(uwonoi-urejunout poiseo1q-Av12) pmog 220/20d44| "y *(ueonoy-urejunour poqpq-o1e[d) pos sugsosrunun] vuazipuy , 


= 


2006 “TP 19 JOX[90A I 0 T 0 0 grXo[duroo snoeuurT snpan 

p007 “PORN 79 sumg 0 0 0 T T 0 0 0 0 ng xepduroo voun], 
0003 

*uo&uvT 29 oyeg I 0 I qX9[duroo puounjory smuouuoag 


,LXo[duroo AemSpry snuoydoyjasy 


oouo1opol xo[duroo soroods onpdvexgooS007 
onoeuoSo[Auq 


(gy) [ensny 
(LV) uelanieg mog 
(gy) ueianioq YON 
(pv) uvopuy 15994 
(gy) ueuopenog 18e 
(cy) ueiquiopor) 15e 
(Ty) ueponzouo A 
(gS) enen vues 
(cs) veneg 
(TS) mdoL 
(29) uoueq 

(99) 1587 e»ueurepe, | 
(pp) 1887 uepeuroyenz) 
(ED) 1994 ueeuropenz) 
(ZO) wvorxo]jy seg 
(T9) ueorxojy 15994 


(oy) ueranzoq- enuon 
(c) 1599 eouvurepe], |— 


penunuo) "Tq e[qep, 


418 Annals of the 
Missouri Botanical Garden 
Table 2. Collecting localities and GenBank accession numbers of samples sequenced for this project. 
Accession 
Taxon (tissue number) Locality number 
Campylorhamphus pusillus LSUMZ B33822 Peru, Cajamarca, Cordillera del Condor, Picorama  FJ222626 
C. pusillus LSUMZ B11879 Ecuador, Esmeraldas, El Place FJ222627 
C. pusillus LSUMZ B14 Panama, Darién, Cerro Pire, 9 km NW of Cana RI222028 
C. pusillus STRI JTWO94 Panama, Bocas del Toro, Chiriquí to Chiriquí FJ222629 
ande Rd. at continental divide 
Chlorospingus canigularis LSUMZ B34949 Ecuador, Napo FJ222652 
C. canigularis LSUMZ B34918 Ecuador, Pichincha FJ222653 
C. canigularis LSUMZ B35820 Costa Rica, Cartago FJ222654 
C. flavigularis FMNH 430078 Peru, Cusco, Paucartambo FJ222655 
C: a STRI JTW 018 Panama, Veraguas, Santa Fe FJ222656 
C. flavigular LSUMZ B11936 Ecuador, Esmeraldas, El Placer FJ222657 
Clirysothlypis ise STRI JTWO16 Panama, Veraguas, Santa Fe FJ222631 
ranioleuca erythrops LSUMZ B21 Panama, Darién FJ222630 
Haplospiza rustica. STRI EC-HRU-514 Ecuador, Carchi FJ22264 
H. rust non LSUMZ B16173 Costa Rica, San José, La Georgina FJ222648 
H. rusti LSUMZ B74: Venezuela, Amazonas, e de la Neblina FJ222649 
Pe S schisticolor LSUMZ B16048 Costa Rica, Heredia, 4 km SE of Virgen del Socorro FJ222632 
M. schisticolor LSUMZ B2124 Panama, Darién, 6 km s of Cana J222633 
M. schisticolor FMNH 429987 Peru, Cusco, Paucartambo FJ222634 
M. schist LSUMZ B11979 Ecuador, Esmeraldas, El Placer FJ222635 
Sa dris luteoviridis STRI JTW 533 Panama, Chiriquí, Cerro Colorado FJ222659 
Tangara florid STRI JTW 169 Panama, Bocas del Toro, Chiriqui to Chiriqui FJ222636 
Grande Rd. at continental divide 
T. florida STRI PA-TAL-1014 Panama, Veraguas, Santa Fe FJ222637 
T. guttata STRI JTW 013 Panama, Veraguas, Santa Fe FJ222638 
. guttata AMNH 8807 Venezuela, Amazonas, Tamac FJ222639 
Thamnistes anabatinus LSUMZ B6152 Ecuador, Morona-Santiago, W ue Cordillera FJ222640 
del Cutucu 
T. anabatinus LSUMZ B2154 Panama, Darién, 6 km NW of Cana FJ222641 
T. anabatinus FMNH JTW179 Panama, Bocas del Toro, Chiriquí to Chiriquí FJ222642 
Grande Rd. at continental divide 
Xiphocolaptes promeropirhynchus FMNH 394013 Mexico, Hidalgo, 5 km E of Tlanchinol FJ222643 
X. promeropirhynchus M 56169 Nicaragua, 10 km N of Matagalpa FJ222644 
d ed ae FMNH JTW596 Panama, El Copé, El — National Park FJ222645 
X. erythropygi NH JTW669 anama, Darién, Puerto Pin. J222646 
rae a LSUMZ B9941 Costa Rica, San José FJ222659 


merican Muse 


Natural History; FMNH, Field Muse 


eum of Natural History; LSUMZ, Louisiana Muse 


eum of 
Natal History; STRI, Smithsonian o Reech Institute (collected by author); UWBM, University of Washington 
urke Museum of Natural History and Culture 


most-parsimonious trees. A bootstrap analysis of nodal 
support with 1000 bootstrap replicates was perform 

on the entire matrix following B (1995). However, 
ch tree in ma 
nonindependent Mas Od the assumption of 


because nodes for ea rix form are store 


character independence is not met. Furthermore, 
zoogeographie species occupying many geographic 
regions will be represented in the matrix by a greater 
number of nodes and will thus contribute more to 
bootstrap analysis. Given the current lack of more robust 
statistics of nodal support for supertrees, I use bootstrap 
analysis to provide a rough measure of nodal support. 
Regions of the tree with poor support are interpreted as 
reflecting conflicting topologies in input trees. 


I constructed a second class of area cladograms 
directly from the sequence data for each zoogeographic 
species. Following Weir and Schluter (2004), sequences 
for each highland region were concatenated across a 
zoogeographie species complexes (sequence concate- 
nation area cladogram [SCAC]). This was done by 
combining a DNA sequence from a single individual for 
each zoogeographic species within a highland region 
into a composite sequence. When a zoogeographic 
species was absent from a particular highland region or 
if a sequence was not available for that region, a series 
of n’s of appropriate length was inserted to represent 
the missing sequence. This method allows both the 
relationships between geographic regions and the 


Volume 96, Number 3 
2009 


Weir 
Neotropical Montane Forest Birds 


average branch lengths connecting geographic regions 
to be estimated directly from the sequence data and in 
sense is more robust than supertree methods. 

difficulty arises in rooting the analysis without selecting 
any geographic region to be an outgroup. Maximum- 
likelihood methods that use a clock or relaxed clock 
assumption d root the analysis without the 
aid of a predefined outgrow 

A Bayesian brun was estimated in Bayesian 
Evolutionary Analysis Sampling Trees (BEAST) 
v1.4.3 (Drummond & Rambaut, 2007) using the log- 

rmal relaxed clock model with a Yule-prior for 
branch lengths. The coefficient of variation in rates 
(o; standard deviation in rates across lineages divided 
by mean rate) was significantly greater than zero in 
this analysis but less than one pe — 0.52, 9596; 
—0.66), indicating 
significant but minor rate variation across lineages. 


highest posterior density — 0.3 


The analysis was run for 10 million generations and a 
sample saved every 1000 generations. The first 2.5 
million generations were deleted as the burn-in 
perio a consensus phylogeny with mean branch 
lengths was generated in FigTree (Rambaut, 2007). 
A caveat of the SCAC method is that sequences from 
older zoogeographie species complexes will be more 
divergent than younger zoogeographic complexes. As a 
result, sequences from older taxa will contain more 
phylogenetically informative characters and will con- 
tribute more to the phylogenetic signal. One possibility 
is to weight each character for a given zoogeographic 
complex by the i inverse of the number of phylogenet- 
ically | t 


in that complex. Weighting 
each character would force each zoogeographie complex 
to contribute evenly to the phylogenetic signal. How- 
with branch 
length estimation and is not implemented here. 


ever, weighting characters will interfere 


The same zoogeographie species were included in 


netic analysis of Cinclus (Voelker, 2002) were not 
deposited in GenBank 


RESULTS 
GENETIC DISTANCE ANALYSIS 


GTR-I distances between adjacent highland re- 
gions are shown in Figures 2 and 3. Genetic distances 
are plotted separately for specifically distinct taxa 
versus conspecific populations. 


Genetic distances between 
. 2À) were 


available for six zoogeographic species EON) all 


Isthmus of Tehuantepec. 


exican an uatemalan regions (Fig 


but one considered conspecific by current taxonomic 
treatment. Genetic divergence ranged 0.6%-10% (or 
0.3-5 Ma using the standard 2% clock) across the 
intervening Isthmus of Tehuantepec, suggesting that this 
lowland barrier has promoted extensive population 
differentiation in a number of zoogeographic species. 
Interestingly, the least diverged complex was the only one 
considered to comprise distinct species. By contrast, 
populations of Arremon brunneinucha Lafresnaye and 
Chlorospingus o Du Bus de Gisignies 
separated for approximately 5 .6 Ma, respectively, 
are currently considered conspecific (American Or- 
nithologists’ Union, 1998) 


The 


Guatemalan region is isolated from the Talamancan 


Lowland isthmuses in lower Central America. 


region by extensive lowland barriers in Nicaragua. 
Genetic differentiation across this lowland gap ranged 
1%-9% (Fig. 2B) and had a similar distribution of 
divergence dates to the Isthmus of Tehuantepec. The 
three taxa with the greatest genetic differentiation 
across this gap are recognized as ic dnd distinct. 
Genetic differentiation in cons 

1%-4.5%, ag 


divergence within some species. Genetic differentiation 


pecifi ra 
ain demonstrating KHAN a 


across the lowland gaps that separate Talamanca, 
and the west Andes ranged 0.7%-11.6% for 


taxa considered specifically distinct in these regions 


Darién, 


and 4.9%-11.7% for taxa considered conspecific 
(Figs. 2C, 3A, B). 


The Santa Marta, 
Parian, and tepui regions are isolated from each other 
and the Andes by lowland barriers. The Santa Marta in 
northern Colombia possesses a larg 


South American lowland barriers. 


e number of 


endemic species (n = 18) and subspecies (n = 55) 


— 


Strewe et al., 2006), few of which have sequence data 
in GenBank. 
divergence  betw anta 

Colombia MID ped 2.6%-9.49% p pede 
taxa and 5.8%-11.3% for specifically distinct taxa 
(Fig. 3C). The Parian region represents an isolated 
highland re 


Pa 
Venezuela. It is isolated by expansive lowland barriers 


In s s o E bus 


gion in the Parian Peninsula of eastern 


from the closest highland region, the Venezuelan 
Andes. Across this gap, genetic distances for four 
zoogeographic species complexes ranged 5—6.59b 
(Fig. 3D). Only one of these is considered asocia 
distinct (Myioborus S. F. Baird; 6.2%). The tepuis, a 

series of ancient highland regions of the Guayana 
Shield, are isolated from the Andes and the Parian 
region by the extensive Amazon and Orinoco basins. 
Many Andean zoogeographic species complexes do not 


Annals of the 
Missouri Botanical Garden 


0246 8 1012 


ure 2. Ness of genetic distances (x-axis — 
d 


mism derived from protein-coding 


west 
versus eastern subregions within the Talamancan region. 


reach the tepuis. Of those that do, a large proportion 
of taxa represent endemic species or subspecies. 
Unfortunately, equences were available fo 
only one zoogeographie complex (Myioborus), which 
is differentiated by 496—596 from its relatives in the 
Parian (Fig. 3E), Venezuelan, and east Colombian 


E 


regions. This complex is ind distinct in each of 
these regions, and within the tepuis multiple iue 
are described, SEEN that the tepuis may con 

subregions of multiple endemism. Additional ns are 
necessary to determine the magnitude of genetic 
populations from other 


differentiation of tepui 


highland regions of endemism. 


Divergence across intermontane river valleys. A series 
of deep river valleys bisect montane forest patches 
along the eastern slope of the Andes and form barriers 
between adjacent regions of endemism. Each river 
valley promoted genetic differentiation, but to different 
extents. The Rio Quinimari and its tributaries, which 
carved out the deep and arid Tachira Depression, 
separate the east Colombian and Venezuelan regions 
(Fig. 1). Genetic differentiation across this depression 


o 
0246 8 1012 4 


Mexican versus east Mexican, (E) western versus eastern subregions within the Guatemalan region, 


2 
nr 


02468 1012 


0246 8 1012 


0 
024 6 8 101214 


0246 8 1012 


GTR-I pu e) princ between Mexican and Middle American 
DNA s ista axa 


Ie 
& 
Mei 
E 
e 
E 
E 
E 
to 
- 
S 
d 
[o 
Bs 
a 
= 


and (F) western 


in four zoogeographic species ranged 0%-6.2% 
Fig. 3F), suggesting that this barrier has effectively 


— 


promoted divergence in at least some species 


complexes. The Eastern Cordillera of s Colombian 


rapidly, primarily between 5 3 Ma 
ry-Wodzicki, 2000; Au. i 2003; ' Dhont et 
al., 2005). The oldest divergence suggests that 
population divergence between the east Colombian 
and adjacent regions of endemism began around the 
time of when uplift was completed 3 Ma. 

Either the Río Caquetá or a series of low-elevation 
montane passes may form the barrier between the east 
Colombian and east Ecuadorian regions. However, 
only limited genetic differentiation occurred in the 
two zoogeographic species complexes that span these 
potential barriers (Fig. 3G). One of these complexes 
has endemic species on either side of these barriers, 
yet showed only low levels of divergence. Additional 
sampling from the east Colombian region is necessary, 
but the two complexes available suggest that genetic 
differentiation is not great between these regions. 

e North Peruvian Low an ío Marañón 
represent a barrier that fully bisects the Eastern 


Volume 96, Number 3 
2009 


Weir 
Neotropical Montane Forest Birds 


Cordillera of the Andes. This region is believed to 
form one of the most important barriers for E 


1980), 


because a number of sister species occur north and 


birds and other groups (Vuilleumier, 1969, 


south of this region. Genetic differentiation across this 
barrier ranged as high as 6.196 (1196 when the north 
and central Peruvian regions are analyzed together) in 
seven zoogeographic species, confirming its impor- 
tance in promoting population divergence (Fig. 3L J). 

In contrast to the Río Marañón, five of the six species 
that span the Río Huallaga in Peru showed only limited 
genetic differentiation (Fig. 3K). However, Chlorospin- 
gus ophthalmicus showed 6% divergence across this 
river barrier, even though it showed no differentiation 
across the Río Marañón (Weir et al., 2008). Although 
Cracraft (1985) stated the importance of the Río 

Huallaga d in delimiting the ranges o 

mber of P. n endemics, he did 


recognize ile pon of the north and central 


a small 
not officially 


Peruvian regions. The available evidence suggests that 
these regions were important in promoting population 
divergence in relatively few species. Further sampling is 
necessary. 

Genetic differentiation across the Río Apurimac, 
which separates the central and south Peruvian 
regions, ranged 0.2%-14.3% (Fig. 3L). Like the Río 
Marañón, this river barrier has played a key role in 
promot ation divergence and 


ing and a populat 
speciation in the Andes. Sti 


l, a substantial number of 
zoogeographie species showed limited or no diver- 
gence across this barrier, implying recent range 
expansions or ongoing gene flow across this valley. 

Finally, the Río Grande and its tributaries have 
carved out an arid valley in Bolivia, isolating montane 
forest in the south Peruvian and Austral regions. Only 
three zoogeographic complexes spanned this barrier, 
and they showed moderate to large genetic differen- 
tiation ranging 1.3%-—6% (Fig. 3M). 


Di j b ontane 


foresta is distributed along both the eastern and western 
slopes of the Andes of Colombia and Ecuador south to the 
North Peruvian Low. High-elevation barriers above the 
tree line bisect montane forest along each slope. Genetic 
differentiation between the eastern and western slopes 
(Fig. 3H) ranged 0%-5.3% with nine of 12 taxa 
differentiated by less than 1%. The lack of strong o 
di ffe retinon in B copos d that at lea 

E the vs 
nc did not fi I li ]. The severe 
glacial episodes of the mid and late Plèistoceñe directly 


glaciated high elevations of the Andes, lowering montane 
forest zones (Bennett, 1990; Hooghiemstra et al., 1993; 
Van't Veer & Hooghiemstra, 2000; Hooghiemstra & van 
der Hammen, 2004). 


North of the Isthmus of Tehuantepec in Mexico, 
humid montane forest is distributed along the Sierra 
Madre Oriental in the east and the Sierra Madre 
Occidental and Sierra Madre Sur in the west (Fig. 1). 
These forests are separated by an arid high-elevation 
n the west and 


plateau. Population divergence betwee 
eas xic available for only three 


n regions 
zoogeographic species complexes, each considered 
conspecific by current taxonomic treatment. Genetic 
divergence between these regions was low for Lampor- 
(0.6%), but was between 6% 
and 7% for Arremon torquatus Lafresnaye & d'Orbigny 


and Chlorospingus ophthalmicus (Fig. 2D), suggesting 
that populations of the latter two species have been 
separated for more than 3 million years. 

Other barriers. I compared genetic differentiation 
between eastern and western subregions within both 
the Talamancan and Guatemalan regions. Three 
zoogeographic species complexes (Pselliophorus 
Ridgwa 20225 ornis Swainson, Selasp dd Swainson) 
have enia species in each of the Talamancan 
subregions, two of which are inched here  (Psellio- 


geographic species complexes exhibited minor to 
moderate genetic differentiation between the eastern 
and western subregions of the Talamanca (0.3%-3.8%; 
Fig. 2 wo complexes contain endemic species, but 
in each case, ‘ieee endemic species are separated by 
divergence values less than 1%. In addition, in Lam- 
pornis, samples from multiple individuals demonstrate 
that populations are not reciprocally monophyletic 
(Garcia-Moreno et al., 1999). 
Three zoogeo. c species complexes spanned the 
two Guatemalan hee including one with endemic 
species in each subregion (Lampornis). In contrast to the 
Talamancan subregions, genetic distances between the 
Guatemalan subregions were high, ranging 3.7%-6% 
and suggesting divergence dates of 1.6-3 Ma (Fig. 2E). 


AREA CLADOCRAM ANALYSIS 


Input topologies of zoogeographic species complex- 
es used in the SAC analysis are shown in Figure 4. 


Supertree area cladograms. The supermatrix obtained 
from phylogenies for 35 zoogeographic species had 1 

columns. Standard matrix representation parsimony 
resulted in a single most-parsimonious tree, requiring 


196 steps (Figs. 5A, 6A) 


IN t H area. ladog TOI. 


The topology 
uh corren in the relaxed- elodke. Bayesian phylogeny 
(SCAC) varied in important ways from the SAC topology 


422 Annals of the 
Missouri Botanical Garden 


4 D 
4 C 
2 45 E 
i ! 2- 
0 => 0246 8 1012 


02468 1012 0 


A 


"e 
hw 


0246 8 10 12 


8 
6 
0 

4 024681012 * 

2 

G 

2 

0 

02468 1012 

0 


0246 8 1012 


4 4 J 
I 
2 
au 
0 
o EE 0246 8 1012 


0246 8 1012 


4 


0 
0246 8 1012 
4 


0 
0246 8 10 12 14 


4 


0246 8 1012 


igure 3. Histograms of genetic distances (x-axis = GTR-T' corrected distances} between South American regions of 
E. derived from protein-coding mitochondrial DNA sequences. Distances between species-level taxa are shown in 
ifi ions i is is the numbe oogeo i 1 


versus west Andean, (B) Darién versus west Andean, (C) Santa Marta versus east Colombian, (D) Venezuelan versus Parian, 


Volume 96, Number 3 
2009 


Weir 
Neotropical Montane Forest Birds 


(Figs. 5, 6). The most striking difference was the 
grouping of the Talamanca and Darién regions with the 
Guatemalan and 

phyletic South American, 


Mexican regions to produce mono- 
Middle and 
Mexican clades, although posterior probability was low 


for these groug In the SAC topol 


o 


American, 


ogy, Talamanca and 


Darién were nested within the South American radiation. 

Topology in those parts of the SCAC and SAC trees 
with nodal d greater than 0.9 posterior proba- 
bility (SCAC) and 70% bootstrap support (SAC) 
generally agreed A 6). The main differences were 
the retention of a monophyletic Andean clade in the 
SCAC iis but not in the SAC analysis 

support for many nodes in both the SCAC and 

SAC aus highlights the general discordance in 
input phylogenies (Fig. 4) between many regions of 
endemism. Because of this conflict, it is not 
straightforward how to interpret branch-length infor- 
mation in the SCAC analysis, even for those parts of 
the cladogram that have high nodal support. 


DISCUSSION 
THE ROLE OF LOWLAND BARRIERS 

Populations within zoogeographic species complexes 
ibited a wide 
e between geographically 
0%-14%; 
Figs. 2, 3, 7, 8). Populations of at species 
7) exhibited the 


greatest genetic divergence Pd = 3.4%), while 


of Neotropical montane forest birds exhibite 
range of genetic divergenc 
adjacent regions of endemism (range, 


spanning lowland barriers 
populations spanning river valley barriers (median = 
2.0%) and highland barriers (median = 0.6%) were 
less diverged; differences were significant only between 
the ae ae highland barriers (Mann-Whitney U 
test, W = 451; 0.01 

surprising given dee low 


ese differences are not 

band barriers 
wider geographically than other barrier types. 

The lowland Isthmus of Panama distributed between 


are generally 


was inundated 
m 


-a 
oO 


the Talamanca and Darién highlands 
until 3—4 Ma when the final formation stage 
Central American land b ul was d pde et 
al, 1992; Coates & Oban d 
completion vied full ee D ie between 
North and South America in birds (Weir, bombe 
data) mammals (Simpson, 1980; Webb, 1985), and 
other groups (Stehli & We 
the Great American Biotic Interchange. A burst of 


5), an event known as 


interchange between the Talamanca and Darién and 


between the Talamanca and the west Andean regions at 
or shortly after the land bridge completion is evident for 
highland birds (676—896 sequence divergence; Figs. 2C, 
3A) 


A more recent burst during the last 1 million years 
(096—295 sequence divergence) coincides with the 
severe glaciations of the Andes 1990; 
Hooghiemstra et al., 19 i 
Hammen, 2004) Talamanca (Lachniet & Seltzer, 
2002), Guatemalan highlands (Anderson, 1969), and 
Mexican highlands (White & Valastro, 1984; Heine, 
198 n montane forest was repeatedly lowered by 
as much as 1000 m in elevation (Vant Veer «€ 
Hooghiemstra, 2000; Hooghiemstra & van der Ham- 
men, 2004). This lowering allowed some montane 
forest elements to expand into pe owlan 
regions and may have increased the pro oe of 
dispersal of montane species across the isthmus 
(Hooghiemstra & Cleff, 1995). 

Montane interchange during the past 1 million 
years was also observed across other lowland barriers 
(Figs. 2, 3), but none of them exhibited a distinct 
interchange burst during this time as observed 
between the Talamanca, Darién, and west Andes. 
This find 


across other low 


ing may simply reflect the small sample sizes 
land barriers, and additional sampling 
may reveal that interchange at this time was more 
extensive than the current data suggest. Alternatively, 
the mid to late Pleistocene elevational lowering of 
montane forests may have provided the first opportu- 
nity for most Andean species to disperse into the 
recently form 
regions follow. an 
Pleistocene oe into other highland regions may 
have been hindered by competition with the species- 
diverse faunas that already inhabited those regions. 


THE ROLE OF ARID RIVER VALLEYS 


The role of 


population differentiation is controversial. 


arid river valleys in promoting 
These 
barriers may directly promote differentiation 
preventing or minimizing gene flow (Vuilleumier, 
1969). Alternatively, Fjeldsà and his coauthors 
(Fjeldsá, 1995; Fjeldsá € Lovett, 1997; Fjeldsà et 
al., 1997, 1999) proposed that river valleys were not 
gu for promoting diversification events 
despite the fact that they demarcate adjacent regions 
of endemism and the range limits of a large proportion 


=> 


(E) Parian versus tepul, (F) east 
Ecuadorian versus 


t Colombian versus Venezuelan, (G) eas 
west Ecuadorian, (I) east Ecuadorian versus north pcm (J) east Ecuadorian versus nor 


t Colombian versus east Ecuadorian, (H) east 
th Peruvian and 


central Peruvian, (K) north versus central Peruvian, (L) central versus south Peruvian, and (M) south Peruvian versus Austral. 


Annals of the 
Missouri Botanical Garden 


Arremon brunneinucha 


Cl 


Ro Obs SSS 


Arremon torquatus 


Myioborus miniatus 


Myioborus complex 


Chlorospingus complex 

cl Cl 
C3 | C2 
C5 C3 
C7 | = CA 
A3 C5 
A6 C6 
A7 C7 
A4 A4 
E A3 

AS 

A6 

AT 

A8 

Al 


Myadestes complex 


C5 CS 
C7 81 Có 
82 82 C7 
$3 83 A3 
A7 Al A4 
A8 A2 A6 
AI A3 A7 
A2 A4 
A4 A5 
A3 A6 
AS A7 
A6 
Cranioleuca complex Pionus sordidus Tangara complex 

Cs 
C? 83 C6 
A4 82 c7 
S1 Al 

A4 
A6 A2 
A7 A4 A6 
A8 A6 


a cladograms of Neotropical montane zoogeographic peo complexes based on Bayesian or maximum 
le 1. 


Are; 
rent eee Regions 


of Andean birds. In their view, regions of endemism 
along the Andes are associated with areas of climatic 
and ecologic stability that persisted over long time 
spans and throughout the Pleistocene climatic cycles. 
These areas of stability are seen to promote parapatric 
or sympatric speciation, without the need for geo- 
graphic barriers to gene flow. While the model may 
apply commonly to taxa with parapatric or sympatric 
speciation, it is unlikely to apply commonly in birds 
for which sympatric speciation is rare and allopatric 
speciation is the predominant mode of pes 
(Coyne & Price, 2000; Phillimore et al., 

ecologically stable areas may help maintain species 
diversity in birds, they are unlikely to have commonly 


of endemism are labeled following codes in Tab 


promoted speciation without the aid of intervening 
geologic barriers to gene 

Rather than forming opposing models, ecologically 
stable areas and arid river valleys may have worked in 
concert to promote population divergence between 
regions of endemism. During periods of intense 
glaciation, montane forests were lowered by 
or more in altitude (Van’t Veer & Hooghiemstra, 
2000; Hooghiemstra & van der Hammen, 2004), 
potentially forming a continuous forest belt across 
river valleys, or at least i epi the possibility of 
dispersal across them ring interglacial periods, 
montane forest rationed to is altitudes, becom 
once again fragmented by intervening river valleys. 


Volume 96, Number 3 
2009 


Weir 
Neotropical Montane Forest Birds 


Andigena complex 


>> > 
Sh 


Cacicus chrysonotus 


>> > 
ta 5 


4 
A3 


zag 


Chlorospingus canigularis 


nig 


Chlorospingus flavigularis 


nig 


Cinclus complex 


> 
Do 


Diglossa complex 


EEE 
[e 
C Un 


Hemispingus complex A 


> > > 
uo 4 


Figure 4. Continued. 


Hemispingus complex B 


> > > 
aA tn Ww 


Henicorhina leucophrys 


ET 


Lampornis amethystinus 


Cl 


ana 
w ND 


a 
E 


Lampornis complex 


Qoo 
UA BR OW 


Metallura tyrianthina 


>> > 
aOR 


Myadestes occidentalis 


Qoo 
Sá 


Myrmotherula schisticolor 


Ochthoeca cinnamomeiventris 


Z È 


Ochthoeca fumicolor 
A4 


f 


Ochthoeca frontalis 
A4 


> > 
D ur 


Ochthoeca complex 


> > > > 
Nad b 


Tangara florida 


Oana 
IH 


Tangara guttata 


no 
= d 


Thamnistes anabatinus 


229 


AS 


Troglodytes complex 


reo 


Xiphocolaptes 
promeropirhynchus 


aga 
BN 


A3 


Xiphorhynchus erythropygius 


oo 
Y 


A4 


Ecologically stable areas are represented by the 
intervening slopes between river valleys that remained 
forested throughout glacial and interglacial cycles. 
Imost 50% (16 of 34) of populations from adjacent 
endemism regions separated by river valleys date to the 


last 1 million years (Fig. 7), suggesting that the intense 
glacial cycles during the mid and late Pleistocene may 
have resulted in repeated episodes of range expansion 
across river valley barriers, followed by isolation by 
those barriers. However, a large proportion (11 of 36) of 


Annals of the 
Missouri Botanical Garden 


West Mexican 


East Mexican 


Guatemalan West 
Guatemalan East 
Talamancan West 
Talamancan East 


Darien 


Venezuelan 
West Andean 
East Ecuadorian 


East Colombian 
North Peruvian 
Central Peruvian 


South Peruvian 


Austral 


Tepui 
L [o^ A Santa Marta 


| oU 


Parian 


West Mexican 


East Mexican 


i d 1 
Guatemalan West 


Guatemalan East 


E ——— Talamancan West 


L Talamancan East 


Venezuelan 
Austral 


South Peruvian 


North Peruvian 


Central Peruvian 


East Colombian 


0.03 0.02 : 
branch length (substitutions per site) 


e 5. Supertree area cladogram (A) and sequence concatenation area cladogram (B) for Neotropical montane forest 


1gur 
ceo species. Supp 
c 


ort indices are shown to the right of nodes. Bootstrap values are shown 
Poor Doo for the quu con aterianon area cladogram 


for the supertree area 
. Posterior probability o 1.0 shown by 


only for th 


ea cladogram. 


Lowland Parisis indicated by L, highland barriers by H, and river ee barriers by R. 


divergence events across river valley barriers predate 
the late Pliocene and Pleistocene glacial cycles 
altogether (2.4 Ma to recent; Fig. 7), demonstrating 


that glacial cycles alone were not responsible for all 


divergence events across river valleys. These results 
are consistent with the long-term role played by arid 
river valleys in promoting differentiation between 


regions of endemism 


Volume 96, Number 3 
2009 


Weir 
Neotropical Montane Forest Birds 


West Mexican 
i: DR East Mexican 


| Guatemalan West 
I Guatemalan East 

Talamancan West 
¡AL Talamancan East 


| Darien 


Venezuelan 


West Andean 
—— East Ecuadorían 
East Colombian 


North Peruvian 
Eo Central Peruvian 
E +h ha 


South Peruvian 


Austral 


Tepui 


[| Santa Marta 
——— Parian 


[ West Mexican 


l East Mexican 


| Guatemalan West 


| Guatemalan East 


m~ Talamancan West 


L— Talamancan East 


Darien 


Darien 


Santa Marta 


ae 


Cladograms showing 


o 


Venezuelan 
Austral 


South Peruvian 


Narf) 


North Peruvian 


Central Peruvian 


East Ecuadorian 


East Colombian 


only nodes with greater than 70% bootstrap support for the supertree area cladogram (A) 


Figure 
and greater than 0.95 posterior probability for the sequence concatenation area cladogram (B). 


THE ROLE OF DISPERSAL AND VICARIANCE 


The role of dispersal and vicariance in promoting 
the formation of patterns of endemism is widely 
debated (e.g., Zink et al., 2000}. A number of highland 
species or species complexes are derived from 
lowland source faunas. This has most often been 
interpreted as speciation following dispersal from 


lowland to highland regions (i.e. Weir, 2006; 
Brumfield & Edwards, 2007). However, Ribas et al. 
(2007) a suggested that highland populations 
were originally distributed at low elevations, but were 
gradually uplifted and vicariantly isolated from 
adjacent lowland populations during orogeny of the 
Eastern Cordillera of the 
vicariant scenario may have played a role in the origin 


though such a 


Annals of the 
Missouri Botanical Garden 


Lowland Barriers 


8 10 12 14 16 
pes distance (%) 


River Valley Barriers 


"B 2 4 6 8 10 12 14 16 
genetic distance (%) 


Non-forested Montane Barriers 


12706 


Count 


0 
O 2 4 6 8 10 12 14 16 


genetic distance (%) 
Figure 7. Genetic distances between adjacent regions of 
iem d pie = Em river valley, and nonfor- 
ted m Mri iers. Median values shown by arrows. 
pala nces PME n species-ley, a taxa are shown in black and 
between ipu. populations in gray. 


of certain highland clades, most highland zoogeo- 
graphic species complexes lack lowland components, 
suggesting they have diversified solely within Neo- 
tropical highland regions. Vicariant isolation. from 


lowland ancestors is highly unlikely for divergence 
events within the Mn RE species complexes 
included in this 

Other possible H of vicariance include the 
development of river valleys or uplift of highland 
regions above the tree line. Formation of such barriers 
in the Andes may have simultaneously fragmented the 
ranges of co-distributed species, resulting in popula- 
tions with similar ages along either side of the barrier. 
A burst of a events should date to the time of 
barrier formation. For taxa isolated on one side of a 
barrier, ora across the barrier could produce 
founder populations at any time following barrier 
formation. The combination of initial vicariance and 
subsequent founder events would result in a burst of 
events dating to the time of barrier formation followed 
by a series of subsequent events up to the present. 

Unfortunately, the geologic literature does not 
provide specific dates for the formation of Andean 
river valleys. Given that extensive uplift of highland 
regions along the eastern edge of the Andes occurred 
throughout the past 10 million years in Ecuador and 
Peru and the past 5 million years in Colombia (Gregory- 
Wodzicki, 2000} and Venezuela (Audemard, 2003; 

ont et al., 2005), river valleys in these regions almost 
certainly date to these time periods. The wide span of 
divergence dates for most river valley barriers suggests 
that a protracted history of dispersal was the predom- 
inant mode by which the current fauna in highland 
regions of endemism originated. Sample sizes are too 
small to detect o of divergence dates associated 
but this may be 


possible as other species are investigated. However, 


with formation of most river barriers, 


characterized by a burst of divergence events around 
3 Ma followed by a smaller number of divergence 
events up to the present. 

In contrast to the wide range of divergence dates 
associated with most river valley barriers, almost all 
populations isolated on either slope of the northern 
Andes date to the past 1 million years. This recent burst 
suggests either some ongoing gene flow or very recent 
isolation between these slopes. The latter seems likely, 
as Pleistocene glacial cycles were severe in the Andes 
only during the past 0.9 million years, and glaciation at 
high altitudes likely provided a hard barrier to gene 
flow, vicariantly separating populations along the 
eastern and western slopes (Weir, 20 


CONSENSUS AREA CLADOCRAMS 


The SAC and SCAC methods produced different 
topologies (Figs. 5, 6), and interpretation. of the 


patterns uncovered is not straightforward. Still, 


Volume 96, Number 3 
2009 


Weir 
Neotropical Montane Forest Birds 


Intraspecific 


Count 


0 2 4 6 8 10 12 14 16 
genetic distance (%) 


Figure 8. 
intraspecific ta 
conspecific populations in gray. 


branching patterns shared in common by the two 
methods highlight some key points. First, both 
methods suggest that the basal divergence event in 
montane forest taxa occurred the 
north Andes and Nicaragua. In the SAC topology, this 
break occurs north and south of the Nicaraguan 


somewhere between 


lowlands, while in the SCAC topology it occurs across 
the lowland barrier that separates the Darién from the 
Andes. This basal break suggests the presence of two 
or more distinct montane forest source faunas for the 
Neotropics, one in North America (ie., Mexico, 
Middle America) and one or more in South America. 
c placement of this basal break just 
nort AC) of the Isthmus of 
Panama suggests that the narrow isthmuses that 


The biogeographi 
rth (SAC) or just south (SC 
separate the North and South American continents 
ormed an important hland birds. This 
basal (or near basal) split is also reflected in the 
topologies of a large proportion of individual zoogeo- 


barrier for hi 


graphic species complexes (Fig. 4). A few of these 
dispersed between North and South America 
and bridge completion, when marine channels 
still bisected these isthmuses, while most dispersed at 
or after its final uplift (Figs. 2C, 3A). In source trees, 
Darién and Talamancan highland regions share affin- 
ities with highla he north some 
zoogeographie species and highland regions to the 
south in others (Fig. 4). In the two area cladogram 


nd regions to t 


methods, the conflicting placement of the Talamanca 
and Darién highlands ther the North or South 


American clades likely reflects this mixed ancestry for 


in eit 


zoogeographic species complexes invading these re- 
cently uplifted highland regions along the land bridge. 

oth analyses, the Parian, Santa Marta, and tepui 
regions—isolated from each other and the Andes by 


Genetic distances between adjacent regions of ende 
xa. Median values are shown by arrows. Distances between species-level taxa are shown in black and between 


Specific 


E m 
E 2 4 6 8 10 12 14 16 
genetic distance (%) 


mism for taxa differentiated at the species level and 


lowland regions—are basal within the predominantly 
South American clade. Within the Andes, relation- 
ships between regions of endemism were poorly 
resolved in the SAC method but well resolved in the 
SCAC method (Fig. 6). The SCAC topology supports a 
monophyletie Andean clade but suggests a curious 


o 
that four geographically intervening regions separate 
these regions. This unusual relationship is likely 
driven by the Chlorospingus ophthalmicus complex, 
ig 
quence similarity to representatives from the Austral 


whose Venezuelan representative shares se- 
region despite a series of genetically differentiated 
and geographically intermediate populations (Fig. 4; 
Weir et al., 2008). 

In t topology, the west Andean, east 
Ecuadorian, and east Colombian regions formed a 
well-supported clade in the north Andes. This clade 
was sister to a clade containing the north and central 
Peruvian posa: with the Río Marañón forming the 
barri n them. Basal to these was the clade 


ned m south Peruvian, Austral, and Venezue- 


er 


lan regions. Ignoring the unusual placement of the 


biogeographic history of northward expansion. Begin- 
ning with a source fauna in the southern Andes, 
northward expansion may have occurred in a stepwise 
fashion, first across the Ri 
across the 
younger ages for uplift of many highland regions in 
the north Andes, with uplift occurring in some regions 
(i.e., the east Colombian region) as recently as 3 Ma, 
expansion into these regions from a South Andean 


Annals of the 
Missouri Botanical Garden 


source fauna is not ble. Alternatively, a 


vicariant history may explain this pattern with 
formation of the Rio Apurimac forming the initial 
vicariant event followed by subsequent formation of 
the Río Marañón. While the burst of divergence 
events across the Rio Marañón near 3 Ma (Fig. 3L, J) 
may suggest a vicariant event, a clear burst is not 
detected for the Río Apurímac. 

Relationships in Middle America and Mexico are 
not well resolved for deeper nodes, but do strongly 
suggest that subregions within each highland block 
ormed monophyletic assemblages. Lowland barriers 
separate these monophyletic clades within highland 
regions from each other. 

Nodes represented by lowland barriers generally 
occurred deepest in the SAC and SCAC phylogenies 
(Fig. 5), highland 


barriers occurring in more derived positions. These 


with nodes spanning river and 


results further demonstrate that lowland barriers have 
t important role in diversification of 


yed the 
Neotropical highland birds. 


CONSERVATION IMPLICATIONS 


Highland montane forest regions of endemism were 
defined on the basis of endemi 
these regions (Cracraft, : 
Stattersfield et al, 1998). This study demonstrates 
that a large ee pulations wit! 
spread differentiated 
populations in many of these regions of endemism 


po 
ossess genetically 


(Fig. 8), emphasizing the value of preserving tracts of 
highland forest in each. Genetic differentiation was 
greatest between  highla regions ted 
lowland bus and by deep, arid river valleys. 
dun of endemism on either side of these barriers 
a high With the sole 


2 of the division between the east Colombian 


conservation priority. 


and east Ecuadorian regions, all geographically 


gre han 2 Ma (assuming 

icd 2% clo P and most had Ee e E 
greater than 2 Ma. Dates for intraspecific splits 
between adjacent highland regions ranged as high as 

.5 Ma. It is important to note that almost half (32 of 
70) of all intraspecific divergence events between 
adjacent highland regions (Fig. 8) occurred more than 
1 Ma and one third (23 of 70) occurred more than 
2 Ma, demonstrating substantial isolation of popula- 
tions in these regions 

The extensive genetic differentiation between 
conspecific populations in highland regions of 
endemism demonstrates that currently defined species 
boundaries do not adequately represent all evolution- 
arily distinct lineages. Under the biologic species 


concept (Price, 2007), this is not an issue because 

ogic species may contain multiple genetically 
differentiated taxa provided those populations are not 
reproductively isolated. Under the biologic species 
concept, genetically differentiated populations should 


species concept is currently favored (although species 
definitions are arbitrary for many allopatric popula- 
tions) by the South American Classification Commit- 
2007) for So 
and the American Ornithologists’ Union (American 
Ornithologists’ Union, 1998) for North and Central 


American birds, it is important that governments treat 


tee (Remsen et al., uth Ámerican birds 


evolutionarily ee units rather than species- 
level taxa as the basic units of conservation. 

Given the narrow altitudinal distributions of many 
montane species, it may be necessary to preserve 
large tracts of intact forest in order to maintain 
substantial population sizes. However, extensive 
deforestation threatens the continuity of forest tracts 
within montane regions. Continued development of 
road systems and slash-and-burn agricultural prac- 
tices are primarily responsible. Ironically, govern- 
ment control of illicit crops accelerates montane 
deforestation as drug growers are forced to clear 
iti i more inaccessible 
2005). Existing forest tracts 
are rapidly disappearing from many montane regions 


additional land along higher, 
slopes (Fjeldsa et al., 


of endemism. In the Andean region as a whole, less 
than 10% of montane forest is estimated to be intact 


— 


Henderson et al., he percentage is lower in 
some regions of endemism, with less than 4% of 
forest remaining intact in the west Andean region 


Dodson & Gentry, 1991). The key finding reported 


here, that many widespread 


— 


montane forest species 
are composed of multiple genetically differentiated 
populations in geographically localized of 
endemism, shows that a failure to ud 
preserve remaining montane forest tracts within each 
region of endemism will result in the loss of much 
greater diversity than predicted from a straightfor- 
ward taxonomic analysis. 


Literature Cited 


s A. 2002. Molecular systematics and the role of the 

cotone in the diversification of 

Xiphorhynchus woodcreepers (Aves: Dendrocolaptidae). 

Auk 1 1-640 

ee Omid Union. 1998. Check-list of North 
American Birds, 7th ed. American Ornithologists’ Union, 
Washington, D.C. 


a-firme" 


Volume 96, Number 3 
2009 


Weir 
Neotropical Montane Forest Birds 


Anderson, T. H. 1969. First evidence for glaciation » Br 
L north-western Guatemala. 
387. 


. Geomorphic and 1 id f 
and ne in the Merida Andes, 
Venezuela. da Int. 101: 43-65. 
Barker, F. K. € S. M. Lanyon. 2000. 

parsimony weighting schemes 


The impact of 
on inferred relationships 


among toucans and wwe barbets (Aves: Piciformes). 
Molec. id Evol. 15: 215-234. 
Baum, B 


2. Gombining trees as a way e combining 
T of 


& 


combining gene trees. Taxon 41: 
Buea, K D. 1990. Milankovitch TA aad their effects o 
species in ecological and evolutionary time. Paleobiology 
11-21. 
Bininda-Emonds, O. R. P. 2004. The evolution of supertrees. 
ends Ecol. Evol. 19: 315-322 
——— € 1998. Properties of matri 
representation with parsimony analyses. Syst. Biol. 47. 
497-508. 


J. Sanderson. 2001. Assessment of the 


accu f representation with parsimony analysis 
supertree d Syst. Biol. 50: 565-579. 
Brumfield, R V. Edwards. 2007. de into and 


out of the pes A Bayesian analysis of M wies 


diversification in Thamnophilus Ru Evolution 
346-367 


Burns, K. J. 1997. Molecular systematics of tanagers 
(Thraupinae): Evolution and biogeography of a diverse 
radiation of Neotropical birds. Molec. Phylogen. Evol. 8: 

348. 


— Naoki. 2004. Molecular Ee and 
biogeography of Nene tanager the genus 
angara. Molec. Phylogen. Evol. 32: 838-854. 
2 D; 


Evolutionary differentiation in the 

s, Emberizidae). Mol 

& J. A. Obando. 1996. The nd evolution 
—56 eC. 


of the entral American jars Pp. 
Jackson, A. F. Budd & A. G. Coates lea. Taub 


and Environment in Topical America. University of 
Mou: Press, Chicago. 
. B. C. Jackson, L. S. Collins, T. M. Cronin, H. J. 


E L. M. Bybell, P. Jung s A. onae 1992. 
Closure of the Isthmus of Panam: a 


s J. 1995. Geographic patterns of neoendemic and 
ndean forest birds: The signifi- 
89-102 in S. P 


and Patagonia, South America. Apollo Books, Copenha- 

en. 
——— & J. C. Lovett. 1997. Biodiversity and environmental 

dide Biodivers. & Conservation 6: 315—323. 
— Ehrlich, E. Lambin & 
biodiv pai ‘hotspots’ correlated w 
matic stability? A pilot study using the NOAA-AVHRR 
remote sensing data. Biodivers. & Conservation 6: 
i; 


———À n & B. Mertens. 1999. Correlation between 
ra iub local ecoclimatic stability documented by 
comparing Andean bird distributions E A sensed 
land E data. Lm 22: 6. 

———, M. D. Alva 

T e crops it js conflict 

biodiversity conservation in the Andes region. Ambio 34 

11 


Garcia-Moreno, J., P. Arctander & J. Fjeldsà. 1998. Pre- 
Pleistocene differentiation among chat-tyrants. Condor 
100: 629-640. 

———, —— € ——. 1999. A case of rapid 


diversification in the neotropics: Phylogenetic relation- 
g Cranioleuca spinetails (Aves, Furnariidae). 

Mies. mic dl Evol. 12: 273-281. 

, J. Ohlson & J. Fjeldsà. 2001. MtDNA sequences 

support monophyly of Hemispingus tanagers. Molec 

Phylogen. Evol 2]: 424—435. 

— — —, N. Cortes, G. M. Garcia-Deras € B. E. Hernandez- 
Banos: 2006. Local origin and diversification among 
Lampornis hummingbirds: A Mesoamerican taxon. Molec. 

98. 


Phylogen. Evol. 38: 488-4 


Gentry, A. H. 1986. Species richness and floristic composi- 
tion of Chocó region plant communities. Caldasia 15: 
11-91. 


. 1992. Tropical forest biodiversity—Distributional 
patterns and their conservational significance. Oikos 63: 
8. 


Grafe, K., W. Frisch, I. M. Villa & M. Meschede. 2002. 


record of Costa Rica and western Parama: a Geol. See. 
Amer. 104: 814-828. 
Coyne, J. & T. Price. 2000. Little evidence for sympatric 
dde in island birds. Evolution 54: 2166-2171. 
Cracraft, J. 1985. Historical ee and patterns of 


differentiation within the South American avifauna: Areas 
e ee 


: A. H. Gentry. 1991. Biological extinction in 
western Eciadir: Ann. Missouri Bot. Gard. 78: 273-295. 

. & A. Rambaut. 2007. BEAST: Bayesian 
P analyst by sampling trees. BMC Evol. Biol. 
7:2 


thermochronolosy. P 348: 187— 
Gregory-Wodzicki, K. M. 2000. Uplift d E the central 
and northern Andes: A review. Bull. I. Soc. Amer. 112: 
1091-1105. 
Hackett, S. J. 1995. Molecular systematics and zoogeography 
of flowerpiercers in the Diglossa baritula complex. Auk 
70. 


112: 156-1 

Heine, K. 1988. Late Quaternary glacial chronology of the 
Mexican volcanoes. Die Geowissenschaften 

Henderson, P. Churchill L. Luteya 1991. 


Neotropical plant diversity. Nature 351: 2 
Hooghiemstra, H. & A. M. Cleff. 1995. Miocene cli- 
cd in 


matic change and environment and generic 


Luieyn (editors), Biodiveisity and Conservation of Neo- 
tropical Montane Forests. New York Botanical Garden, 
Bronx 


Annals of the 
Missouri Botanical Garden 


——— van der Hammen. 2004. suaran Ice-Age 
dynamics s the Colombian Andes: Developing an 
understanding of our legacy. Philos. Trans., Ser. B 359: 
E Melice, A. Berger € N. J. Shackleton. 1993. 
Frequency-spectra and paleoclimatic variability of the 
high-resolution 30 1450 ka Funza I pollen record (Eastern 
Cordillera, "o an Sci. P. ev. 12: 141-156 


Irestedt, M., J. Fjeldsá & G. Ericson. 2004. 
m log ende relationships of De (Aves: Den- 
i tl 1 between molecular and 


dr laptina T 


morphological ne J. Avian Biol. 35: 280—288. 


Klicka, M. Spellman. 2007. A molecular evaluation 
of the North American “grassland” sparrow clade. A 
124: 537-551. 


Kreft, H. & W. Jetz. 2007. Global patterns and determinants 
of vascular plant diversity. Proc. Natl. Acad. Sci. U.S.A. 
104: 5925 50 30. 

Lachniet, M. S 


G. O. Seltzer. 2002. Late ru 
glaciation of Costa Rica. Bull. Geol 114: 
47-558. 


. Soc. Am 


de I. J. 2004. a EE dating and mixed support 
for the “2% rule" in birds. : 1-6 

Mayr, E & L. L. Short. Li . Species Taxa of North 
American Birds: A Contribution to Comparative System- 
atics. Nuttall Ornithological Club, Cambridge, Massachu- 


J. Diamond. 2001. The Birds of Northern 

Melanesia: Speciation, Ecology, and Biogeography. Oxford 
ee Press, Oxford. 

Miller, M. E. Bermingham & R. E. Ricklefs. 
"cl diera ography of the new world bites 
(Myadestes spp.). Auk 124: 868-885. 
oyle, R. G. : d of barbets (Aves: 
Piciformes) based on nuclear and m irochonilr ial DNA 

87-200. 


Nahum, , S. L. Pereira, F. Fernandes, S. R. Matioli 
& A tal. 200. versification f Ramphastinae 
(Aves, Ramphastidae) prior to the Dmm oi 
boundary as shown by molecular clock of mtDNA 

q es. Su olec. Biol. 26: 411—418. 


The Speciation and Biogeography of Birds. 
cademic "uin Amsterdam. 
in J. P. 1992. A general overview of the m 
avifauna of Peru. Mem. Mus. Hist. Nat. “Javier Prado” 2 
— - 
s , G. Voelker, B. Mila & D. a Girman. 2003. 
a on e long-distance migration and historical 
NE aphy of Catharus a ushes: A inb n phyloge- 
netic approach. us 120: 299-310 
Perez-Eman, 2005. a phylogenetics and 
biogeography of the Neotropical redstarts Voi OTUS; 
Aves, Du Molec. Phylogen. p 31: 511—528. 
Phillimore, A. B., C. e Orme, G a T. M. 
d die E Bennett, K. J. eut I. P. F Owe 
8, Sympatrió Fen in birds is rare: Insights from 
and simulations. Amer. Naturalist 171: 


M. Lanyon. 2004. Song and molecular data 


ovel affinities of the Green 


Price, T. 20 peciation in Birds. Roberts and Company 
Publishers, Greenwood Village, Colorado. 
Purvis, A. 1995. A modification to Baum and Ragans method 


for combining phylogenetic trees. Syst. Biol. 44: 251-255. 
Rambaut, A. igTree. <http://tree.bio.ed.ac.uk/software/ 
figt ree/>, a 23 June 2009. 


Remsen, J. V., C. D. Cadena, A. Jaramillo, M. Nores, J. F. 


bird species of South Aura. American Ornithologists’ 
Union, Washingt 
-Rensen SACC Baseline, hils pe 15 December 
2007. 


u.edu/ 


Ribas, C. C., R. G. Moyle, C x Miyaki & J. Ciao 2007. 
The assem mbl 
and climatic bee to independent regimes 
B in Pionus parrots. Proc. Royal Soc. London, 
Ser iol. Sci. 274: 2399-2408. 


+ 


: Linking 


Rice, t H, A. T. Peterson & G. Escalona-Segura. 1999. 
Phylogenetic patterns in montane eae wrens. 
Condor 101: 446-451. 

Rosenzweig, M. L 
Time. Ca 


199 ecies Diversity in Space and 
ambridge University Press 


, Cambridge, United 

Kingdom 
Simpson, C. G. 1980. Splendid It The Curious 
mals. Yale University 


g & D. C. We 


Stattersfield, A. rosby, A. J. L 
1 World. Bindlife 


E Endemic Bird Areas of = 

International, Cambridge, United Kingdom 

Stehli, F S. Webb. 1985. The Great American Biotic 
Interchange. Plenum Press, New York. 

Stotz, D. F., J. W. Fitzpatrick, T. A. . K. 
Moskovits. 1996. Neotropical Birds: Ecology and Conser- 
vation. University of Chicago Press, Chicago. 

Strewe, R., C. Navarro & C. J. Villa-De Leon. 2006. 
Conservation of the endemic birds of the x 2 de 
Santa Marta massif, Colombia. J. Ornithol. 147: 258-259. 

Swofford, D. L. 2002. PAUP* 4.0b10: Dea Analysis 
Using Parsimony (“and Other Methods). Sinauer Associ- 
ates, Sunderland, Massachusetts. 

i P. M. Jørgensen. 

s del Ecuador 
2000. Herbario QCA, Pontificia Universidad Católica del 

Ecuador, Quito. 

Van't Veer, R. € H. MR 2000. Montane forest 
evolution during the last 650 000 yr in Colombia: A 
multivariate approach es on pollen record Funza-I. J. 
Quatern. Sci. 15: 320-346. 

Voelker, G. 2002. Molecular phylogenetics and the historical 
biogeography of dippers (Cinclus). Ibis 144: 577-584. 

Rohwer, R. C. K. Bowie & D. C. Outlaw. 2007. 

Molecular systematics of a speciose, cosmopolitan song- 

bird genus: Defining the limits of, and relationships among, 

the Turdus thrushes. Molec. Phylogen. Evol. 42: 422— 

434. 


Vogler, . DeSalle. 1994. Diagnosing units of 
conservation management. Conseupidon Biol. 8: 354-363. 
Vuilleumier, F. 1969. Pleistocene speciation in birds living 
high Andes. Nature LRO, 
1980. S high Andes. Acta XVII 
Coi M v mE Omithologici 2: 1256-1261. 
Webb, S. e Ceno: mal xo between 
the Ane E 57-386 in F. Stehli & S. Webb 
(editors), The Great American Biotic Interchange. Plenum 
Press, New e 
Weckstein, J. D. 2005. Molecular phylogenetics of the ramphas- 
tos toucans: [enne ons for the evolution of morphology, 
vocalizations, and coloration. Auk 122: ae 09. 
J. T. 2006. Divergent timing and ns of species 


and highland Neotropical birds. 


in the 


accumulation in lowland 


Evolution 60: 842-855 


Volume 96, Number 3 
2009 


Weir 
Neotropical Montane Forest Birds 


— —— & D. Schluter. 2004. Ice sheets promote speciation 
in boreal birds. Proc. Royal Soc. London, Ser. B, Biol. Sci 
271: 1881-1887. 

———. 2008. EN the avian molecular 
ie n. Ecol. 17: 2321-2328 

Miller, J. Klick: M. A. 

oie 2008. aaa € a morphologically 
diverse Neotropical montane species, the common bush- 
tanager (Choros spina ophthalmicus). "Molec, Phylogen. 
Evol. 47: 650-61 


Bermingham, M. J. 


White, S. E. & S. Valastro. 1984. Pleistocene glaciation of 
Volcano M CEDE Mexico, and dr e Mea the 
standard M Quatem. Int 1-35. 

Winker, K. & c L. Pues 2006. Eu mea 
speciation, and morphological convergence in the genus 
uu E SR, Auk 123: 1052-1068. 

Zink, . C. Blackwell-Rago & F. Ronquist. 2000. 
E eh roles of dispersal and vicariance in 
biogeogr aphy. Proc. Royal Soc. London, Ser. B, Biol. 
Sci. 267: 497-503. 


MOSS DIVERSITY AND 
ENDEMISM OF THE TROPICAL 
ANDES! 


Steven P. Churchill? 


ABSTRACT 


E mosses of the tropical Andes are > examined to determine a conservative estimate of diversity, excluding a v 


of unconfirmed n: 
amilies are reco, 
f r 


trict 


zed. Within 


trópi x An des i is 


estimated at 31%. Regionally, the number 


to De th es (321 species) is po than the number 


restricted to the central Andes (241 species). Regional e aes exhibits a FM pattern: more endemics in the northern 
p 


Andes (155 spec 


words: 


ies) than in the central Andes 
Diversity, pean i mosses, 


(129 spec 


tropical du en 


The tropical Andes are widely acknowledged as one 
of the world's great centers of biodiversity (Rodríguez- 
Mahecha et al., 2004). Species richness is one of the 
criteria that serves to rank the tropical Andes as a major 
focal point of biodiversity. Other criteria include the 
level of endemism and past and current environmental 
degradation (Orme et al., 2005). The very foundation of 
biodiversity is our knowledge of the organisms. Precise 
estimates of diversity for most major groups of 
organisms are, however, elusive for the tropical Andes. 
It is very likely that diversity and distribution within 
this region are only well known for birds and mammals; 
all other estimates—for fungi, plants, and insects, for 
example—are only vague or approximate. This is due, 
in part, to required ongoing basic exploration, inven- 
tory, discovery of new species, and, most critical to our 
understanding of diversity, revisionary studies. 

Mosses represent just one group of organisms that 

e the tropical Andes one of the 
T E in the world. This region contains about 
15%-17% of the estimated 8000 to 9000 mosses in 
the world. Endemism is relatively high, with 31% of 
the species considered to be unique to the region (see 
below). Beyo 


another dimension that ranks these organisms as 


reat centers of 


nd high diversity and endemism, there is 


possibly one of the most important groups in the 
tropical Andes. Disproportionate to their small size, 
mosses, rather like the ants so eloquently described 


y the Harvard entomologist E. O. Wilson, play a 


major role in the ecosystem they occupy. Mosses, 
along with hepaties, are the major plant group 
responsible for the natural conservation of water and 
soil in the Andes 

The focus of this paper is an assessment of the 
diversity and endemism for the tropical Andean 
mosses. This present analysis is, in part, a reevalu- 
ation and update of a prior paper addressing moss 
"d of the d Andes (Churchill et pi 19 Ea 

There are several moss publications since 1995 t 
are specifically n pa the tropical nidi e» 
páramo mosses of Venezuela, Colombia, Ecuador, and 
Costa Rica were estimated at 543 species (Churchill 
& Griffin, 1999). The first checklist for the tropical 
Andean countries enumerated 2089 specific and 
infraspecific taxa distributed among 362 genera and 
76 families (Churchill et al., 2000). A descriptive 
treatment of the families and genera was provided for 
the Neotropics (Gradstein et al., 2001) and included 
an analysis of bryophyte regions and habitats. Various 
floristic papers for each of the Andean countries are 
provided on the "Overview of Regio 
and Countries" ete (ee mobot.org/W 2 T/Search/ 
andes/overviewintro.htm>) 


OVERVIEW OF THE TROPICAL ANDES 


The tropical Andes extend approximately 38 
degrees of latitude, from the coastal ranges and 


1 The author is much indebted to Bob Magill for his assistance in making us Tropicos database and the UR web page a 
d 


a classic 


reality. My interest in an 
My 


l mosses owes a gr t to 


article by Gentry (1982). This work was supported by grants from the National Science Foundation DEB- de DEB- 


0542422) and by the Taylor Fund for Ec 


preparing the 


2 Steven P. Churchill, Missouri Botanical Garden, P.O. Box 


ological Research through the Missouri Botanical Garden. 
Marshall nese 2 C. Hollowell, and Peter Jorgensen for their review of t 


The author is grateful to 
he manuscript, and to Eliana Calzadilla for 


299, St. Louis, Missouri 63166-0299, U.S.A., and Museo de 


Historia Natural Noel Kempff Mercado, Av. Irala 565, Casilla No. 2489, Santa Cruz, Bolivia. steve.churchill@mobot.org. 


doi: 10.3417/2008043 


ANN. Missouni Bor. Garp. 96: 434—449. PUBLISHED ON 28 SEPTEMBER 2009. 


Volume 96, Number 3 
2009 


Church 
Moss Diversity and Endemism 


Cordillera de Mérida of Venezuela to the puna and 
This 
the longest 


montane ranges in a Argentina. 
arched backbone of Sou 


mountain chain in the ne can be divided between 


America, 


the northern Andes (Venezuela to northernmost Peru) 
at 11°N to 4—5^$ and the central Andes ape central 
Peru to northwest Argentina) at 5°S to . The 
estimated area of the tropical Andes is co km? 
(Rodríguez-Mahecha et al., 2004). That figure is ca. 
39% of the total area of the tropical Andean countries 
and slightly less than 9% of the total land surface of 
South America. The Amazon Basin is nearly 4.5 times 
larger than the anes Andes. Useful overviews for 
e tropical Andes ovided on geography by 
n (1979) an nd ‘for anne by Luteyn and 
Churchill (2000). 


Tue Myrta or Tropica Moss DIVERSITY 


It is imperative to have a clear understanding of the 
misconceptions that impede our knowledge of moss 
diversity for the tropical Andes. The year 1801 marks 
two important events with regard to tropical American 
mosses. Hedwig’s opus, Species Muscorum Frondo- 
sorum, was published in 1801; in time this would be 
adopted as the official starting point of moss 
nomenclature, except Sphagnum This volume 
includes the first mosses collected in tropical 
America, those by Olof Swartz from the West un 

second event of that year occurred in 
i Andes; Alexander von Humboldt E 
Aimé Bonpland were the first to collect mosses from 
the Andes. These mosses were later described by 
William Hooker in Musci Exotici (1818-1820 

Historically, the most active period in which 
Andean mosses were collected and described oc- 
curred from about the mid-19th century to the first 
three decades of the 20th century. During that time 
period, several thousand new species were described 
by Europeans, beginning with Ernst Hampe, Carl 
Miiller, William Mitten, and later Viktor Brotherus 
and The 
describe species for the Andes in the first half of the 


odor Herzog. North Americans also began to 


20th century, most notably Robert Williams and 
Edwin Bartram. The quality of these authors varied 
considerably. ampe an itten, for the era 

deed reasonably sound descriptive treatments, 
but at the other end of the spectrum was Müller. No 
single individual described more species than Müller, 
w completely, or almost, indiscriminate in 
desoribing several thousand species; many of these 
are viewed as redundantly described species. Of the 
many collections sent to Müller from South America, 
almost all were described as new species. Unfortu- 
nately for bryology, Müller was blessed with a long life 


(1818-1899), and, most detrimental to bryology, his 

collections housed in the 

destroyed in World War 
The recognition of species by these 19th- and early 


Berlin herbarium were 


20th-century authors was based mostly on a very 
narrow species concept, often defined by minute or 

plant stature, 
1994. Many of 


the described mosses were based on a very limited 


trivial differences in morp A 
leaf shape, seta length, etc. “Pursell, 


number of specimens and incomplete knowledge of 
the species described at that time by different authors. 
Due to the paucity of specimens, there was almost no 
idea of how these plants varied or how they could be 
differentiated from other recently described species 
rom the same region. Other factors also contributed to 
the increasing confusion: communication was limited 
e and 


publications, duplicates were only later distributed if 


or long delayed with regard to correspondenc 


they existed, and in addition, there were various 
conflicts between the European nations (Kruijer, 
2002). 

The dilemma of excessive naming of species for the 
tropical regions has been discussed by Touw (1974) 
and Magill (1982). Those familiar with the state of 
tropical bryophytes stress the dire need for revisionary 
studies. A few examples from revisionary studies of 


tropical Andes: 
recognized 49 species of Campylopus Brid. for the 
tropical Andes, relegating 58 previously published 
(1995) recognized 12 
species of Bartramia Hedw., placing 22 into synon- 


species to synonymy; Fransén 


ymy; and Mufioz id recognized 15 species of 
rimm e tropical Andes, with 26 
birds published species relegated to synonymy. 
Among just these revisionary studies, 106 species 
were thus subtracted from the heretofore accepted 
mosses, or, viewed another way, the Andes lost 106 
endemic species. 
The compilation of checklists for mosses, either for 
individual countries or regions in the tropics, was an 
mportant initial phase during the late 20th century in 
ig development of floristic knowledge. In large part, 
such checklists were compilations from previous 
floristic and revisionary studies. Embedded within 
these checklists were numerous species for which the 
taxonomic status was unknown. Checklists for all the 
tropical Andean countries were compiled in the 20th 


e Bolivia (Herzog, 1916; Hermann, 1976), 
mbia (Florschütz-de Waard & Florschütz, 1979; 
em) 1989), Ecuador (Steere, 1948; Churchill, 


1994), Peru (Menzel, 1992), and Venezuela uw 
1936; Pursell, 1973). A complete summary 

tropical Andean countries was provided by Tene: 
et al. (2000). All of these checklists incorporated the 


Annals of the 
Missouri Botanical Garden 


many newly described species, as well as the earlier 
dubious species reports for the tropical Andes, 
although adjustments to recognized species were 

made based on revisionary studies that existed. 
A few previous studies have attempted to use data 
from checklists to analyze and provide generalizations 
and trends with regard to d (Churchi 1; 
Delgadillo. 994; Chu et al, 1995; Pala 
2003). Given the data e this has on occasion 
led to rather exaggerated species numbers. Delgadillo 
1994) examined moss diversity and endemism in the 
Neotropics for 24 countries. Three of these countries 
serve as examples of inflated species numbers: 

Bolivia, 1182 

5 species, with 


species, with 359 endemic; Brazil, 
endemic; and Paraguay, 148 
species, with 54 endemic. Although it is impressive to 
note that 49% of the Brazilian moss flora is endemic, 
36% is endemic to Paraguay, and 30% is endemic to 
Bolivia, it is far from the reality. Brazil has fewer than 
1000 species and probably fewer than 100 endemies, 
Bolivia has about 900 eee and 56 endemics (see 
ave on the order 


below), and, although as 
of 200 


pecies, it is hig f Ard that there is not a 
single endemic in this country (Churchill, pers. obs.). 
These assumptions are based on a greater knowledge 
gained through floristic and revisionary studies over 
the past two decades and on directions suggested 
these results that will impact our understanding of 
species diversity. 

There is a need then to develop a new generation of 
tropical countries 


One of the first 


checklists for bryophytes in the 
based on new stringent criteria. 


catalogs concerning bryophytes, in F case the 
hepatics for Bolivia by Grad dein et al. (2003), e 
more pragmatic approach in Au. doubtfu. 


names and reports. While very ~ of those eee 
names are now accepted, a greater portion has since 
been shown to be synonyms. This is a far better 
approach to take in future efforts of compiling 
bryophyte checklists particular to areas such as the 
Neotropics. The analysis of moss diversity and 
endemism of the tropical Andes will entail error, but 
it is better to err on the side of a conservative, realistic 
estimate than to err on the side of an embroidered 
fantasy of diversity. 

Data for this analysis are derived from four sources: 
(1) the taxonomic treatment of the tropical Andean 
mosses (<http://mobot. mebo org/W3T/Search/andes/ 


ndean countries in the Missouri Botanical 
Garden Tropicos system; (3) compiled world 
checklist for mosses by Crosby et al. (2000); and (4) 
recent floristic and taxonomic revisions not found on 

the aforementioned web page. Accepted taxa include 


only legitimate names at the level of species and 
above. Excluded from this analysis are all species 


whose taxonomic status is unknown or reports 
considered dubious. All nomina nuda are excluded. 
Infraspecific categories (subspecies, varieties) are not 
included in the analysis. The initial species list of the 
tropical Andean countries (Bolivia, Colombia, Ecua- 
dor, Peru, and Venezuela) and the northwestern 
departments of Argentina (Juyuy, Salta, Tucumán) 
Eros 1974 and included 1457 accepted species and 

excluded, most as status unknown or outside the 
o range. 

A second set of data was generated from this initial 
effort that includes only the Andean region (Fig. 1), 
with th imum elevation for the tropical Andes 
defined at 500 m. Outliers for the tropical Andes 
included in this analysis are the Cordillera de la Costa 

ierra de Santa Marta in 
the boundary is defined to 
include only the departments of Jujuy, Salta, and 


in Venezuela and the 
Colombia; in the sout 


Tucumán in northwestern Argentina. Geographically, 
nd Atlantic 
and Pacific islands (e.g., Galápagos) are excluded. 


the tepui regions (Guayana Highlands} a 


The adjoining coastal and interior lowlands in South 
America less than 500 m elevation (e.g., Amazon 
Basin and Chocó) are also exclude 

Data were arranged in an Excel Otters Word, 
Redmond, Washington, U.S.A.) 


included: name of family, genus, species, number of 


spreadsheet that 


synonyms for each genus, elevation range, endemic 
status, and species present in either the northern or 
central Andes or present in both, and finally in- 
dividual countries. This Excel spreadsheet is avail- 
able on i 
and Countries” under ‘tigpical Andes: «http: f mobo, 


ITQ 


Andean web page “Overview of Region 


mobot.org/W 


htm> 


Taxonomic DIVERSITY 


ee diversity for the tropical Andes is estimated 


ua 7 genera, a amilies 
vius 1) The Ps of species is subst 
tially lower than species Pin 


058 
1995), whose figures 
small fraction a the s 


the 
previously (Churchill et a 


ee 


nelude pecies present in 
the lowlands. Even the projected estimate of 1500 to 
1700 species (Churchill | et al., 1995), considered a 
more realistic figure, not substantiated. The 
umber of specific and infraspecific moss synonyms 


5 


MIRA recorded for the tropical Andes is 929 
(Appendix 1), many of which were recognized in the 
past three decades. 

most speciose families (Table 1) account for 
861 species, containing a significant portion (66%) of 


Volume 96, Number 3 
2009 


Churchill 
Moss Diversity and Endemism 


Ecuador / 
796 $ 
$ 


Pad 
ie 


g 


= 


Tropical Andes 


i 
2 NW Argentina 
122 


Figure 1. 


the total moss diversity for the tropical Andes. This is 
nearly identical to that estimated by Churchill et al. 
(1995). Differences include in part the recognition of 
the Macromitriaceae (segregated from Orthotricha- 
ceae, cf. Churchill & Linares C., 1995) and the 


significant increase in newly described species of 


Map of the tropical Andes with the numbers of species for each of the countries. 


Sphagnum. Other factors include the reduction of 
previously recognized names due to new synonymy for 
the Grimmiaceae, or names excluded in this study as 
status unknown, for example Mittenothamnium Henn., 
which contains an inordinate number of names, many 


likely referable to M. reptans (Hedw.) Cardot. 


Annals of the 
Missouri Botanical Garden 


Table 1. The 10 most diverse moss families and genera for the tropical Andes. 

Family No. of species Genus No. of species 
Pottiaceae 172 dis (Sphagnaceae) 61 
Bryacea 130 Campylop icranaceae) 49 
Dicrana 129 Fissidens sie ntaceae) 46 
Pilotrichaceae 109 Bryum (Bryac 44 
Bartramiaceae 64 Zygodon (TE 33 
Sphagnace 61 Macromitrium (Macromitaceae) 30 
Sematophyllaceae 52 Didymodon (Pottiaceae 26 
Orthotrichaceae 51 Syntrichia (Pottiaceae 26 

acromitriaceae 47 Lepidopilum ur 25 
Fissidentaceae 46 chizymentum* (Bryace, 24 


* Coequal with Sematophyllum (Sematophyllaceae). 


The 10 most diverse moss genera in the tropical 
Andes containing 20 or more species are listed in 
Table 1. Just these 10 of the 327 genera account for 
26% of the total species recorded. Eight of these 10 
genera have been revised or under current study so 
that these numbers seem to be a reliable estimate. 
This may be the case with Zygodon Hook. & Taylor 
(see families discussed below) but is less certain for 
Sematophyllum Mitt., which lacks 
studies. At the other end of the generic spectrum, 


modern revisionary 
159 
or 49% of the genera recorded for the tropical Andes 
are represented by a single species; 55 of the 159 
represent monospecific genera. 


eZ 


a 
elevational range, and distributional range by co 


ENDEMISM 


A taxon is considered endemic if it is only known 
from within the geographical range of the tropical 
Andes; that range may be restricted to a single locality 
or span the entire length of the tropical Andes. The 
number of endemic species estimated for the tropical 
Andes is 428 species distributed among 137 genera 
and 38 families (Appendix 1). The 
species is 3196 of the total recorded for the tropical 


number of endemic 


Andes. Twenty genera are endemic to the tropical 
Andes (Table 2). All are monospecific with the 


exception of Sciuroleskea Hampe ex Broth., with two 


Endemic genera of the tropical Andes. Provided for each genus are the family, associated Andean vegetation, 


Genus Family Vegetation Elevation (m) Country 
Aligrimmia Grimmiaceae 2250-2700 
Allioniellopsis Sematophyllaceae low montane forest 750-1400 cuador, Peru 
Callicostellopsis ilotrichaceae páramo/puna 3480—3620 Venezuela, Bolivia 
llidi rachytheciace low montane forest ca. 1400 olivia 
steinia Amblystegiaceae áramo 3650 Colombia 
iene Amblystegiaceae puna 4600 Bolivia 
Leskeadelphus Leskeaceae high montane, páramo/puna 1300-4000 Colombia, Bolivia 
Leptodontiella Pottiaceae open montane 600-4235 Per 
indigia Brachytheciaceae montane forest 1300-3400 all 
Mandoniella rachytheciaceae montane forest 1700-3350 Bolivia 
Polymerodon Dicranaceae puna 3600-4620 Bolivia 
Porotrichopsis Neckeraceae mid to high montane forest 2000—3800 Colombia, Bolivia 
Pseudohyophila Dicranaceae puna ca. 3820 Peru 
Schroeterella Sematophyllaceae montane forest ca. 2200 Bolivia 
Sciuroleskea. Stereophyllacea montane forest 1300-3160 Ecuador, Peru 
tenocarpidiopsis achytheciaceae montane forest 1400-2150 Ecuador, Peru 
tenodesmus Pilotrichaceae montane forest 700-925 Colombia, Ecuador 
Streptotrichum laceae high montane forest 3140-3400 Bolivia 
imotimius Sematophyllaceae montane forest ca. 2350 cuador 
Trachyodontium Pottiaceae montane forest 2650 Ecuador 


Volume 96, Number 3 
2009 


Church 
Moss Diversity and Endemism 


Table 3. Families and genera with 10 or more endemic species for the tropical Andes. 


Family No. of species Genus (family) No. of species 
ottiaceae 60 Sphagnum (Sphagnaceae) 35 
Pilotrichaceae 49 Zygodon. (Orthotrichaceae 23 
ae 46 Lepidopilum (Pilotrichaceae) 15 
Dicranaceae 37 Schizymenium (Bryaceae 15 
Sphagnaceae 35 Sematophyllum (Sematophyllaceae) 14 
Orthotrichaceae 33 Campylopus (Dicranaceae) 12 
artramiaceae 23 acromi (Macromitaceae) 11 
Sematophyllaceae 21 Dalionia (Daltoniaceae) 11 
Daltoniaceae 18 Orthotrichum (Orthotrichaceae) 11 
Brachytheciaceae 13 Cyclodictyon. (Pilotrichaceae) 10 
Macromitriaceae 11 Didymodon (Pottiaceae) 10 
Ditrichaceae 10 Synirichia (Pottiaceae) 10 
Polytrichaceae 10 
spec There are a few genera that 


coul 
ee as subendemic, i.e., isolated outliers from 


the main Andean range. For example, the monospe- 
cific genus Gertrudiella Broth. Pater is primar- 


d in dry 


Peru to iso dun Argentina, 


nter-Andean valleys from southern 
but has been 
recorded from a single locality in the Bolivian Chaco 


ily foun 


forest near the sub-Andean range and also from 
northernmost Chil 

There are 13 families that include 10 or more 
endemic species (Table 3). Significantly, just these 13 
families of the total 69 Andean families account for 87% 
of species endemism. Within these, 12 genera contain 
10 or more endemic species, and these 12 encompass 


42% (177) of the 428 total Andean endemics. 


Norewortuy Moss FAMILIES OF THE TROPICAL ANDES 


Twenty of 69 families of the tropical Andes 
(Appendix 1) that are considered significant for 
reasons of diversity, endemism, ecology, and distri- 
bution are discussed below 


AMBLYSTEGIACEAE 
The majority of the genera and species are found in 
h humid 


the high montane to páramo and 
Ecologically, a number of the genera are a major 


puna. 


component and play a significant role, second only to 
Sphagnaceae, in the Andean aquatic systems (i.e., 


lakes and ponds, streams and rivers, bogs and 


marshes). Aquatic and semi-aquatic genera include 
oe u pruce, Drepanocladus (Müll. Hal.) 
G. Roth, P. Puudbcolliegon (Limpr.) Loeske, Scorpidium 


(Schimp.) Limpr., Straminergon Hedenás, and Warn- 
storfia Loeske. Endemic genera, all monospecific and 
Grad- 


steinia Ochyra, found on rocks in streams of the 


only known from the type collection, include 


Colombian Eastern Cordillera, and Koponenia Ochyra, 
found on rocks in springs or streams of the Real 
Cordillera of Bolivia (Table 2). The Amblystegiaceae 
were recently revised by Hedenàs (2003) for the 
Neotropics. 


ANDREAEACEAE 


Andreaeaceae is almost exclusively found in 
paramo and puna on rocks. A few species are semi- 
aquatic or aquatic, e.g., Acroschisma wilsonii (Hook. f. 
& Wilson) A. Jaeger, Andreaea nitida Hook. f. & 
Wilson, and A. subulata Harv. A major portion of the 
20 ecostate Andreaea Hedw. 
described from the Andes have not been reevaluated; 


species previously 


revisionary studies may add five to 10 species based 
on observed morphological variation of selected 
Andreaea types and general collections examined by 
the author, of which some will likely be endemic. 


BARTRAMIACEAE 


Bartramiaceae is the fifth largest family for the 
Andes, with 64 species, 23 of which are endemic. 
Nearly all of the genera of the Bartramiaceae are 
terrestrial, common in the open montane to páramo and 
puna. Leiomela (Mitt.) Broth. is the exception, mostly 
found in montane forest as an epiphyte. Breutelia 
(Bruch & Schimp.) Schimp. is a common component of 
bogs, and Philonotis Brid. is common along streams and 
seeps. Revisionary studies are required for Philonotis, 
where the status is unknown for 21 species and the 
nonclasping, leaf-based species of Breutelia. 


BRACHYTHECIACEAE 


The Brachytheciaceae is the tenth largest family, 
with most genera associated with the montane forest. 


Annals of the 
Missouri Botanical Garden 


The family, as now circumscribed, includes five 
genera previously associated sy z Meteoriacaeae: 
Aerolindigia M. Menzel, Lin , Meteor- 
(Müll. Hal.) Manuel, Hier m ill. Hal.) 


Broth., and Zelometeorium Manuel (Ignatov & Huttu- 


idium 


nen, 2002) All five genera generally occur as 
epiphytes, often common and abundant, in montane 
t the Andes. There are four Andean 
endemic monospecific genera (Table 2), all occurring 
Flabellidium Herzog, Mandoniella 
me Lindigia, and Stenocarpidiopsis M. Fleisch. 


forest throughou 
as epiphytes: 


only exclusively aquatic genus is 
Flei 


ns Ree he 10 M. sch., typ 


rocks in streams. The m unrevised taxa of 


ically occurring on 


the Brachytheciaceae involve the generic complex 
duae rd Bruch «€ Schimp. and Eurhynchium 

3; both have rather numerous names 
Mee for 19 Senden) but probably few species, and 
likely even fewer or no endemies. 


BRYACEAE 


Bryaceae is the second largest Andean family, with 
130 species and 46 endemies (Appendix 1). The 
ed P genera and species are terrestrial and 
found i open montane to páramo and puna 
pud genera of the high montane and páramo/ 
puna regions are Anomobryum Schimp., Pohlia 
Hedw., and most notably, Schizymenium Harv. 
Acidodontium Schwügr., with nine of ll species 
endemic, is exclusively epiphytic, often occurring as 
twig epiphytes, as are about half of the species of 
Brachymenium Schwügr. Genera with at least some 
forest species include HE Hedw., Epipterygium 

indb., Orthodontium Schwügr., a d Rhodobryw 
(Schimp.) Lim e resiente Ochi (1980, 
1981) for deb Anomob 
nium, Bryum 
important foundation for these diverse genera; how- 
ever, all would benefit by at least a regional revision. 
Revisionary studies are required for the taxa associ- 
ated with Mielichhoferia Nees & Hornsch. 
Schizymenium. 


and 


DALTONIACEAE 


This family is restricted exclusively to the montane 
forest of the tropical Andes. The most notable genus of 
this family is Daltonia Hook. & Taylor, with 11 of the 
17 species endemic to the tropical Andes, which also 
appears to be the center of diversity for the genus. 
Species are small and inconspicuous, characteristi- 
cally one or a few individuals are found on twigs o 
shrubs (e.g., Baccharis L.) and trees and are often 
present on nodes of bamboo (Chusquea Kunth). Only 


Calyptrochaeta Desv. and Leskeodon Broth. remain to 
be revised 


DICRANACEAE 


Rich in genera, the Dicranaceae is j the third largest 
the tropical Andes, 
distributed among 2 genera. Many of the species 


family for 


are terrestrial, found on soil, humus, rocks, and logs. 
However, a significant portion or all of the following 


genera dre epiphytic: Campylopus, Chorisodontium 


Symblepharis Mont. Campylopus is the second largest 
genus in the tropical Andes, with 49 species, amply 
diversified in most habitats (except aquatic). Schlie- 
phackea, with two species in the north ndes, is 
the only genus of this family with a pendent growth 
orm in the New World. Critical revisionary studies 
are needed for the generic complex that includes 
Dicranella (Müll. Hal.) Schimp. and Microdus Schimp. 
ex Besch. Pseudohyophila Hilp. is the only endemic 
genus for the family (Table 2); although it is placed in 
the Dicranaceae, its systematic position is unclear 


DITRICHACEAE 


Astomiopsis Müll. Hal., Bryomanginia Thér., Pleur- 
idium Rabenh., and Tristichium Müll. Hal. are very 
small- statured, cleistocarpic or gymnostomous genera. 

of the nine species of these genera are endemic, 
all eniin. ps the most part, to the páramo and 
puna. Fc s R. S. Williams and Distichium 
B mp., both with a single species, are also 
restricted to Us páramo and puna. Ditrichum Hampe 
may have a few additional species and some possibly 
endemic, but revisionary studies are needed. 


FISSIDENTACEAE 


The majority of the Fissidens Hedw. species 
belonging to this monogeneric family are small and 
inconspicuous. Species occur on nearly all substrates. 

any of the species are widespread in the Neotropics, 
although 62% of the 93 recognized species are 
endemie to the region (Pursell, 2007). It is rather 
surprising that of the 46 species present in the Andes, 
only four are endemic. Fissidens is the third largest 
genus for the tropical Andes (Table 1), with about 
50% of the recognized Neotropical species occurring 
within the Andean range. 


GRIMMIACEAE 


The majority of genera and species are found on 
rocks in the high montane to páramo and puna. 


Volume 96, Number 3 
2009 


Church 
Moss Diversity and Endemism 


Grimmia, as with Andreaea, is a typical component of 
the páramo and puna, with 15 species. Eight of the 40 
species of this family are endemic to the Andes 
Endemic monospecific genera restricted to the central 
Andes are Aligrimmia R. S. Williams and Coscino- 
dontella R. S. Williams. 
revisionary study is Schistidium Bruch & Schimp., 


The genus requiring 
which may have as many as 15 species, and some will 
certainly be endemic to the Andes; the status of 12 
species is unknown. 


HYPNACEAE 


The Hypnaceae contains 17 genera, 33 species, and 
only five endemic species. In terms of distribution, 
many of the genera and species are widespread and 
common throughout the Andes. Ecologically, several 
genera are very abundant and conspicuous in montane 
forests, often occurring in extensive mats, e.g., 
Ctenidium (Schimp.) Mitt., 
Mittenothamnium. The most problematic genus re- 


Hypnum Hedw., and 
quiring revisionary a is Mittenothamnium, for 


ich numerous names have been proposed (2 
names considered in ie study as status unknown); 


eo 


it is likely that fewer than 10 species will be 
recognized. 


MACROMITRIACEAE 


Macromitrium Brid. is one of the principal generic 
elements of the Andean montane forest. The center of 
species diversity for this genus will likely prove to be 
the tropical Andes. Most species are epiphytic and 
commonly present in the canopy of high montane 
forest. More species will be recognized, based on 
examined types and general collections from the 
tropical Andes, and may total up to 50, with as many 
as half endemic. Schlotheimia Brid. requires revision- 
ary studies; 14 species have been reported 


e 
at 


described from the central Andes, particularly in 
Bolivia, but fewer than seven will likely be recog- 
nized 


NECKERACEAE 


Despite having relatively few species (26) and only 
four endemics, the Neckeraceae is a Bon 
component of montane forest throughout the Andes. 
This is particularly true of the mostly epiphytic genera 
Neckera Hedw., Porotrichodendron M. Fleisch., 


Porotrichum (Brid.) Hampe, which can form extensive 


and 


dendroid tufts on trunks and branches of trees. 
. & Herzog, with one sp 
only endemie genus in the Neckeraceae for the 


Andes (Table 2). Porotrichopsis flacca Herzog is 


Porotrichopsis Broth ecies, 1s 
d 


rather small and inconspicuous, resembling a depau- 
perate species of Porotrichum. 


ORTHOTRICHACEAE 


Represented by two diverse genera, Orthotrichum 
Hedw. and Zygodon, the majority of the species are 
epiphytic and concentrated in the transitional high 
montane forest and páramo-puna zone. Both genera 
contain a significant number of endemie species: 
Orthotrichum with 11 of 18 species and Zygodon with 
22 of 33 species. Zygodon, monographed by Malta 
(1926), requires a reevaluation of the species, but will 
likely remain one of the most diverse genera for the 
tropical Andes as can be presumed from the revision 
of the southern South American taxa by Calabrese 
2006). Orthotrichum was revised by Lewinsky (1984, 

87) 


EN 
o 


PILOTRICHACEAE 


The family is the fourth largest for the tropical 
Andes, with 19 genera and 109 species (Appendix 1). 
The Pilotrichaceae is the second most diverse family 
in the number of endemies, with 49 species (Table 3). 
The center of diversification of the Pilotrichaceae is in 
the northern Andes and, to a great extent, in Central 
America. Many of the genera and species are 
associated with montane cloud forest. The combina- 
tion of diversity and endemism marks this family as 


the 


ecosystem. A numbe 


single most important in the cloud forest 
r of genera are typically found 
over leaf litter, humus, and logs; epiphytic genera 
include Actinodontium Schwágr., Lepidopilum (Brid.) 
Brid., Pilotrichum P. Beauv., and Stenodesmus (Mitt.) 
A. Jaeger. Crossomitrium Müll. Hal. is one of very few 
moss ini in which several species are commonly 
epiphy. Gene ring revisionary studies 
include Callicostella (Mal. Hal.) Mitt., Cyclodictyon 
Mitt., and Trachyxiphium W. R. Buck 


POLYTRICHACEAE 


The family is exclusively terrestrial. Nearly all of 
the species in the Ándes are found in open mid to high 
montane, páramo, and puna. The species of this family 
play a significant role in the colonization of disturbed 
montane slopes and are among the first ow to 
MA recent landslides and newly cut r 
Very few species are associated with montane e 
genera include Atrichum P. Beauv., Steereobryon G. L. 
Sm., and a few species of Pogonatum P. Beauv. 
Within the Neotropics, the tropical Andes contain the 
highest diversity of genera (9) and species (23) for this 
family. The genus Polytrichadelphus (Müll. Hal.) Mitt. 


Annals of the 
Missouri Botanical Garden 


has its center of diversity and names in the tropical 
Andes; it is the only genus of the Polytrichaceae that 
still requires a careful revisionary study. 


POTTIACEAE 


The Pottiaceae is the single most diverse family for 
the tropical Andes in terms of genera, species, and 
endemies (Appendix 1, Tables 1, 3). The majority of 
the genera are common in the wet and dry páramo and 
puna, and in the dry inter-Andean valleys occurring 
on soil and rocks. In the montane region, a number of 
genera are common in open forested areas and 
deforested sites. Genera with some or all species 
found as montane epiphytes include Streptopogon 
(Taylor) Wilson ex Mitt., Leptodontium (Müll. Hal 


Hampe ex Lindb. p.p., and Syntrichia Brid. p.p.; three 


iu 


other genera represent monospecific endemics, each 
known from a particular country: Leptodontiella R. 

Zander & E. H. Hegew. (Peru), Streptotrichum Herzog 
(Bolivia), and Trachyodontium Steere (Ecuador). The 
entire family is now being revised by bryologists from 
dad de Murcia (Cano et 


., Hennediella Paris (Cano, 2008), Syntri- 
chia, and Tortula Hedw. (Cano & Gallego, 2008). 


SEMATOPHYLLACEAE 


Represented by 14 genera, the Sematophyllaceae is 
the seventh largest family for the tropical Andes, with 
52 species (Appendix 1). The most diverse genus, 
Sematophyllum, is found throughout the Andean 
montane region, occurring as epiphytes in forested 
areas and equally associated with streams and rivers 
on rocks. Ámong the genera in critical need of 
revisionary study are both Sematophyllum and 
Trichosteleum Mitt. 
additional nn species and endemics; the 
status of some 17 Three 
monospecific genera, all epiphytic, are endemic to 


the tropical Andes (Table 2): Allioniellopsis Ochyra, 


known only from three localities in Ecuador and Peru; 


e former is likely to have 


species is unknown. 


Schroeterella Herzog, from a single locality in Bolivia; 


and Timotimius W. R. Buck, also from a single 


locality in Ecuador. 


SORAPILLACEAE 


This monogeneric family is of interest for its rather 
peculiar gametophytic morphology and distribution. 
Sorapilla Spruce & Mitt. is represented by two 
species, S. sprucei Mitt. and S. papuana Broth. & 

eh., the former from the Neotropics and the latter 


from Australasia. The genus exhibits a leaf structure 


that of Fissidens. 
relationship of this taxon is thought to be with the 


eckeraceae ae 
w 


similar to The phylogenetic 

Ll ee sprucet is onl 
collection y Richard Spruce in 
1857 from the lower montane forest of Abitagua at 
about 1850 m. 


SPHAGNACEAE 


Sphagnum is the single most important genus of the 
aquatic ecosystems of the Andes. It is a typical 
component of bogs, lake, and stream margins in the 
páramo and humid puna, but it is also found 
associated with seeps and springs in montane areas. 
Nearly all of the 35 of 61 species considered endemic 
ward Crum 


des were described by Ho 
. It seems likely that a 


to the An 

between 1985 and 1997 
reevaluation of these species will result in some being 
reduced to in som 


one. cases entailing 


emended species conce One are very likely to 
d 


p 
be considered distinet a endemic the very 
important role of this genus in the po ecosystem, 


a revision is imperative. 


REGIONAL AND COUNTRY DIVERSITY 


The tropical Andes may be divided geographically 
into two areas, the northern and central Andes. For 
norther 
Andes are defined a as duin the cordillera systems 


pragmatic p this analysi 


of Venezuela, Colombia, and Ecuador, and the central 
Andes are defined as including the cordillera a 

eru, Bolivia, and the northwestern portion of 
mm The generally acd division pe 
the two regions is the Huancabamba Depression in the 
extreme north of Peru (Duellman, 1979), but too little 
collection data exist for this area of Peru. The number 
of mosses common to both the northern and central 
Andes is 816 (59%). More than half of these species 
are common and widespread throughout the Andes, as 
By d by ad at rhynchophorum (Harv.) 

. J. Kop. (for this and other 


view maps in Tropicos); ae species appear as 


yne 
pecies discussed below, 


disjuncts such as the endemic Porotrichopsis flacca, 
ut this may simply be a collecting artifact. The 
northern Andes contain 321 unique species (23% of 
total), thus the total for this region is 1137 species. At 
least a number of the species are more restricted in 
their distribution, as r by Polytrichadelphus 
ciliatus (Hook. son) Mitt. The 
contain 241 unique species (1896), a total of 1057 
species. Again, some of the species are narrowly 
distributed, 
ramicola Herzog, which may be more widespread, 


central Andes 


such as the endemic Streptotrichum 


extending even into Peru, but the substrate this moss 


Volume 96, Number 3 
2009 


Church 
Moss Diversity and Endemism 


Table 4. 


individual countries and 


Summary of moss diversity for the five 


three provinces (Jujuy, Salta, and 


The distributi f 492 d 


Table 5. the 


tropical Andes, northern and central Andes, and individual 


ucumán) in northwestern Argentina. countries including northwestern d AA tina (endemics 
shared between the two regions is 144 spec 
Total in 
Total in Andean % of total No. of 

Country country portion Andes Area Country species — 96 of total 
Venezuela 734 681 53% Northern Andes 5l 12% 
Colombia 915 883 67% Venezuela 13 3% 
Ecuador 807 796 59% Colombi 49 11% 
ru 113 761 56% Ecuador 42 10% 
Bolivia 901 884 66% Central Andes 39 9% 
Northwestern Peru 31 Th 
Argentina 125 122 9% Bolivia 56 13% 

Northwestern 

Argentin: 3 >1% 


occupies, nodes of bamboo, is rarely sampled. A 
comparison of two families, Pilotrichaceae and 


Pottiaceae, serves to emphasize the difference be- 
two n ve 


montane cloud forest, many as quus while the 
Pottiaceae are most common open tane, 
páramo, and puna regions, HR al pu This 
is reflected in the differences between Colombia and 
Bolivia. The in Colombia has 78 
species, and Bolivia, 47. The Pottiaceae in Bolivia 
i his likely 


Pilotrichaceae 


18 species, and in Colombia, 69. 
reflects the more complex and extensive cordillera 
system of montane forest in Colombia as compared 
with a much smaller and less complex cordillera in 
Bolivia. Conversely, the Bolivian puna (humid, semi- 
humid, desert) and dry inter-Andean valleys are 
extensive compared with Colombia, which has 
distinctly isolated páramos and dry inter-Andean 
valleys. 

The number of species recorded for each country 
varies from 734 to 915, excluding northwestern 
p um (Table 4). Colombia is the most diverse, 

915 species, but nearly equal is Bolivia, with 
Bof. Ecuador, for its small size, is notably diverse, 
with 807 species. Both Venezuela and Peru have 
fewer species. The number of species recorded for 
Venezuela (734) po be due to the smaller area 
occupied by the es, or more 
inventory will rod additional species. 
little doubt that Peru is undercollected, with only 775 
species, and will be equal to or greater than the 
number of species recorded for Bolivia or Colom 

The Andean portion of each country (Table 4) Der 
that there are only a few species from the lowlands 
that are not present in the Andes. Venezuela has more 
non-Andean species than any of the countries; this 
may be due to the Caribbean coastal regions and the 


tepui regions. 


REGIONAL AND COUNTRY ENDEMISM 


Endemism for the regions and countries differs 
slightly from the overall results for the countries, 
n M us northwestern Argentina (Table 5 

er of shared endemics between the nort 
and scis Andes is 144 species. There are slightly 
more endemics recorded for the northern Andes (155) 
than for the central Andes (129). Among the countries, 
the difference is with Bolivia, which contains more 
endemies (56) than Colombia (49), followed by 
Ecuador (42), Peru (31), and Venezuela (13). 


ern 


TRENDS IN Moss DIVERSITY FOR THE TROPICAL ANDES 
PATTERNS OF DIVERSITY 


The patterns of diversity can be viewed at different 
levels. The supposition for the observed pattern of 
moss diversity in the tropical Andes is that alpha 
diversity may be comparable to other forested areas 
g., lowland forest) but beta diversity, species 
accounting for 


= 
2 


turnover, is significantly higher, 
cd differences within and between elevational 
gradients and Ew present in the tropical Andes 
len et al., 1995). The 

diversity is ute high for the tropical Andes. 
This scenario has not been tested. This explanation 
appears to be supported by the observed differences in 


result is that gamma 


moss diversity (gamma) seen in data comparing the 
Amazon Basin to the tropical Andes. The area of the 
Amazon Basin, slightly less than the size of the 
contiguous United States, is nearly 4.5 times larger 
than that of the tropical Andes. The mosses of the 
Amazon Basin are estimated at 311 species (Church- 
ill, 1998), compared to 1376 species for the tropical 
Andes; thus, the much smaller area of the Andes is 
4.4 times more diverse than the Amazon. 


Annals of the 
Missouri Botanical Garden 


LATITUDINAL CRADIENT 


The cordillera systems of Central America and the 
he onl 


Andes are the only reason mosses do not serve as a 
classic contrary example to the latitudinal gradient 

that diversity increases toward the equator 
Cii 1991; Churchill et al., 1995; Shaw et al., 

. Latitudinal gradien tisa ibd loose, broad 
Bacci with various “classic” examples (e.g., 
birds, trees), but in some cases this appears to be 
avoiding other patterns of distributional diversity. The 
contrast in moss diversity between the two largest 
(Bum re of tropical South America, the 
Amazon Basin and the tro 


d or (see discussion above) In the absence of 


pical Andes, could not 


elevated topography (i.e., the cordillera systems of 


Central and South ic moss n ud in the 


8 
5 


Neotropics would be the sai 

sin, 1.e., ca. 300 species e slighily i It is 
oversimplified but accurate to note that bryophytes 
generally reach their greatest diversity in environ- 
ments of overall cool temperatures and continual 
precipitation in the form of rain or mist fogs (e.g., in 
the montane forest) or pronounced seasonality of 
precipitation (e.g., in the páramo or puna)—essen- 
tially equivalent to the environment of temperate and 
boreal forest, and tundra of the Northern Hemisphere. 


ELEVATION 


One obvious trend of major interest for the tropical 
Andes is elevational gradient. Mosses, as well as 
liverworts, reach their highest diversity levels in the 
ns of the tropical Andes. T 
band of vegetation, including both forested and open 


montane regio is narrow 
montane areas, contains an estimated 60% or more of 
the mosses of the tropical Andes. An analysis of 
elevational distribution of Colombian mosses demon- 
strated a gradual increase in species diversity 
maximizing at 2500-3000 m, with the next highest 
level of species diversity at 2000-2500 m, and the 
3500-4000 m. The 


number of unique species found within an elevational 


third at a higher elevation, 
range d a similar pattern in Colombia; for 
example, the 2500—3000 m range also contained the 
greatest number of species not present at other 
elevation ranges. In a study of ferns and bryophytes 

om various tropical Andean localities, Kessler 
(2000) n ted lut Fer boundaries were well 
iru with major ecological changes, essentially 
two such zones. One major zone could be interpreted 
as the transition. between lowland forest and low 
montane forest ind The s 
change 


transitional high montane forest with páramo-puna. 


ond d pee 
he 


occurred at the highest ne domm 


Between these low and high zones, species composi- 
tion showed little change. 


ECOREGION DIVERSITY 


The diversity and composition of mosses present in 
the various ecoregions (essentially equal to vegetation 
gh it 


may be intuitively apparent based on casual observa- 
y y app 


types) have not been well documented. Althou 
tion that certain ecoregions are more diverse than 
others, for example that montane forests are more 
species-rich than dry inter-Andean valleys, there is a 
need to quantify such patterns. Bolivia, for which 
there are preliminary data (Churchill, Sanjines & 
Aldana, in prep.), is o as an example. There 
are se major regions recognized in Bolivia 
Fig. 2). Bolivia can be divided into two general areas: 
the highlands (Andes), occupying ca. 40% of the land 


surface, and the much larger lowlands (Oriente), 


— 


din in o. In the highlands, the Yungas 
ontane forest occupies only 5% of the country's 


bd surface but is disproportionately the most diverse 
ecoregion, containing 61% of the 901 mosses 


recorded for Bolivia. The puna is the second most 


se 
diverse ecoregion with 3046, followed by the Tucu- 
mán-Bolivian montane forest with 25%, and the dry 
inter-Andean valleys with 7%. In the lowlands, 
EA corresponds to a precipitation gradient of 
high i north to low in the south; the Amazon 
isa occupying 34% of the Bolivian land surface, is 
the most diverse region in the lowlands with 11% of 
the total number recorded for the country, followed by 
the Chiquitano forest, with 10%, and Chaco 
estimated 3%. 


ecoregions may be similar to other tropical Andean 


, with an 
The diversity of mosses for these 
Bolivia, 


countries; only the Chiquitano is unique to 
dC 


an aco is also present in northwestern Argentina. 
This also serves to demonstrate that the tropical 
Andes with 


significant contributions from the dry inter-Andean 


is more just montane forest, 
valleys, páramo, and puna that add to the rich moss 
diversity of the region. 


TRENDS IN ACTIVITIES 


There are some notable trends, both positive and 
negative, that have an impact on our developing 
knowledge of the tropical Andean mosses. The most 
encouraging development has been the progress of 
bryology in the Andean countries. Much of this has 
the past 
individuals in all the tropical Andean countries 
involved at some level of bryological research. These 
individuals have contributed significantly in develop- 
ing and expanding research collections that were 


Volume 96, Number 3 
2009 


Church 
Moss Diversity and Endemism 


00 


0 00 


| 


. Puna: 273 species 


1 
2 
3 
4. 
5. Amazon Forest: 102 species 

6. Chiquitano Forest: 86 species 
7 
Figure 2. 


heretofore in most cases meager or nonexistent. 
Colombia provides a good example of these trends 
as reflected in the number of publications. In three 6- 
year ye the number of bryological Pd 
by Colombians has sed significantly: 

1995 a, 1996-2000 o. 2001- 2005 (22). The m 
development has been a web page devoted to the 
mosses of the tropical Andes (Churchill & collabora- 
tors, 2008). This web site provides the following main 
sections: Overview of Region and Countries, Key to 
Andean Moss Families, Family Treatments and 


57° 


100 1 2 
— a QÓÁ— 
Proyección cónica conforme de Lambert 
AA I 
T 
60* 


. Yungas Montane Forest: 552 species 
Dry inter-Andean Valleys: 60 species 
Tucuman-Bolivian Montane Forest: 225 species 


. Chaco Forest: 17 (est. 30) species 


Moss diversity among the seven ecoregions of Bolivia. 


Checklists (English), Spanish Family Treatments and 

ecklists, Index of Synonyms and Other Names, 
Bibliography and Literature Cited, Author List, 
Collector List, and Bryophytes of Bolivia. Both family 
treatments are linked to the Tropicos database 
(<http://www.tropicos.org/>, Missouri Botanical Gar- 
den). 

The negative trend is the significant downturn of 
revisionary studies. Our understanding of Neotropical 
mosses has increased significantly in the past 3 
years. A notable number of bryologists from North 


Annals of the 
Missouri Botanical Garden 


America, Europe, and Japan focused their research in 
various regions of the Neotropies, both in terms of 
fieldwo 1 


form of publications, could be said to have increased 


rk and revisionary studies. This trend, in the 


almost exponentially beginning in the late 1960s and 
early 1970s, peaked in the 1980s, and by the late 
1990s began a sharp decline. This trend was a result 
of economic growth of the 1960s and the concurrent 
in universities and 


expansion faculty. As new 


bryologists up expertise, the number of a 
the 1980 most of the 1990s 


until an economic Mene university cutbacks, and 


tions increas sed i s an 


faculty attrition through retirement or death resulted 
B 


in a noticeable decrease publications. 
commencement of biodiversity ma in the 1 


g 


which can be coupled with the bi oes of 
Biodiversity (Wilson & Peter, 1988), was in fact the 
beginning of a decline in taxonomic expertise. What 
was accomplished in ca. 20 years, however, was 
almost certainly a 10-fold increase of new collections 
in the Neotropics, as well as equally important 
revisionary studies that advanced our understanding 
of 35% 45% of the 


understanding of diversity. 


taxa, providing a better 


CONCLUSION 


Given the state of knowledge of the tropical Andean 
mosses, estimates of the number of taxa, particularly 
species, are and will continue to be in a state of flux. 
There are species to be newly described, even more to 
be relegated to synonymy, and new species records for 
the region and the individual countries (particularly 
for Peru). Over the next few decades, there will be a 
significant level of uncertainty with regard to moss 
diversity and composition for the tropical Andes. The 
estimates presented at any one time for moss diversity 
are — And I cannot refrain from saying (after 
an 20 years studying the Andean mosses) that 
eee and relativity apply wonderfully well to 
our attempts to delineate species and vegetation. 

A priority for the tropical Andes must be to provide 
a better resolution of species through revisionary 
studies and floras to promote an understanding of 
moss diversity. This would allow a greater under- 
standing of distribution and ecology and provide a 


etter means of assessing rarity and conservation. 
There is urgency to this priority given the major role 
mosses and hepatics play in the Andean ecosystem. 
Degradation of the Andean landscape has had a major 

act on the ecosystem. The predicted loss of 
glaciers lw the dude in the next few decades 


. 2006; Vergara et al., 


This can only exacerbate the situation faced by the 


major Ándean cities (with ever-increasing population 
growth) that depend on water from these ecosystems. 
The continuing loss of glacial water will likely impact 
the montane forest, which is currently the second most 
important source of water. Any attempt to further alter 
these forests can only lead to a greater loss of water 
and soil to say nothing of the plant and animal 
diversity contained in this narrow band of forest that 
th of the tropical Andes. To ensure the 
continuing function of the montane forest, the area 


spans the leng 


must be recognized as an endangered ecosystem and 
protected, not only for the water it provides but also 


for the rich diversity it contains, including mosses. 


Literature Cited 


Allen, B. H. 1981. A reevaluation of the Sorapillaceae. 
Bryologist 84: 335—338. 

Bradley, R. S., M. Vuille, H. F. Diaz & W. Vergara. 2006. 
"Threats to water supplies in de epl Andes. Science 
312: 1755-17 e 

Calabrese, G. M. 2006. A taxonomic revision of Zygodon 
(Orthot richacesej i in southern South America. Bryologist 
109: 453-500. 

Cano, M. J. 2008. Taxonomic revision of Hennediella Paris 
(Pottiaceae, ur D Biblioth. 64: 1-142. 

& M. The genus Tortula 

Es Beyo s in es America. Bot. J. Linn. 


Soc. 156: 173-220 

—— uerra & : AE 2008. 
Pottiaceae. Integrated Taxonomic Information "aed 
Universidad de Murcia. Mii sti cede com>, ac- 
cessed 30 April 2009. 


Churchill, S. P. 1989. Bryologia Novo Granatensis. Estudios 
de los musgos de Colombia IV. Catálogo nuevo de los 
musgos de s Trop. Bryol. 1: 95-132. 

The floristic oe and elevational 

e EE. a Colombian mosses. Bryologist 94: 157—167 

Pate ~ mosses of peer Ecuador. AAU 


Rep. 35: 
Set CUN of Amazonian mosses. J. Hattori Bot. 
Lab. x Ex 42. 
D. Griffin M. 
Luteyn RS Páramos: Their Ph 
ical Distribution and Botanical Literature. Mem 
Bot. Gard. 84 
& E. L. Linares C. 1995. Prodromus E Novo 
Granatensis. Bibliot. Jose RE. Triana 1-924. 
. Griffin III & M. Lewi Moss divos of 
E iopical Andes. Pp. 335-346 in s. P. Churchill, H. 
Balslev, E. Forero & J. L. Luteyn (editors), Biodiversity 
and Conservation of b Montane Forests. New 
York Botanical Garden, Bro: 
— —— & J. Muñoz. 2000. A checklist of the mosses 
of the tropical Andean countries. Ruizia 17: 1-203. 
collaborators: P Moss of the Tropical Andes. 
To J y: ] x 1 


1999. Mosses. Pp. 53-64 in J. L. 
ytodiversity, Geograph- 
. New York 


jj 


accessed 30 April 20 
Crosby, M. R., R. E. pm B. Allen & S. He. 2000. A 
oad or the Mosses. Missouri Botanical Garden, St. 


Delgadillo, M. C. 1994. a E in the neotropical moss 
flora. Biotropica 26: 12-1 


Volume 96, Number 3 
2009 


Church 
Moss Diversity and Endemism 


Duellman, W. E. 1979. The herpetofauna of the Andes: 
Pii of distribution, origin, aged: ps ded 
communities. Monogr. Mus. Nat. Hist. Uni as 7 

—459. 

Florschütz-de Waard, J. & P. A. Florschütz. 1979. Estudios 
sobre Criptógamas Colombianas III. Lista doe de 
los nd de Colombia. Bryologist 82: 215-2 

Frahm, J.-P. 1991. Dicranaceae: —ÀM Para- 
Eu Fl. Neotrop. 54: 1 

003. Manual of tropical -o Trop. Bryol. 23: 
-19 

Fraisen 995. A taxonomic revision of Neotropical 
Bartramia section Vaginella C. Müll. Lindbergia 20: 
147-179. 

Gentry, A. 1982. Neotropical floristic diversity: Phytogeo- 
Mery cometio bas etween Central a nd South Amer- 
ica or an accident of the 

ndean orogeny? Ann. Missouri Bot. Gard. 69: 557—593. 

Gradstein, S. R., S. P. Churchill & N. Salazar Allen. 2001. A 
guide to the bryophytes of spe America. Mem. New 
Yor k Bot. Gard. 86: 1-571 

—, R. I. Meneses Q. € B. A. Arabe. 2003. Catalogue of 
the Hepaticas and js Re of Bolivia. J. Hattori Bot. 
Lab. 93: 1-67. 

Hedenás, L. 2003. Ambl i (M 
l- 


i). Fl. Neotrop. 89: 


1 

Hedwig, J. 1801. Species Muscorum Frondosorum. J. A. 
Barth, Leipzig. 

Hermann, F. J. 1976. Recopilación de los musgos de Bolivia. 
Bryologist 79: 125-171. 

: i phyten meiner zweiten Reise 

durch E Mt Bot. 87: 1-347. 

Hooker, W. J. 8-1820. Musci Exotici. Longman et al., 


London. 

Ignatov, M. S. € S. Huttunen. 2002 [2003]. MM MCN 
D ga family of sibling genera. Arctoa 11: 
245-2 


Kessler, "n 2000. Altitudinal zonation of Andean cryptogam 

communities. J. Biogeogr. 27: 275-282. 
2002. Hypoterygiaceae of the world. Blumea, 

Suppl. 13: 1-388. 

Lewinsky, J. 1984. adi d Hedw. in South America 1. 
Introduction and taxonomic revision of taxa with immersed 
stomata. Lindbergia 10: 65-94. 

. 1987. ale Un pra in Soul 

America 2. Tax i of t s i 


& S. P. Churchill. 2000. Vegetation of the 
tropical Andes. Pp. 281-310 in D. Lentz (editor) 
Imperfect Balance: Landscape Transformations in the 
Pre-Columbian Americas. Columbia University Press, New 
York. 
iow R. E. 1982. Exotic bryophytes. Beih. Nova Hedwigia 
1: 317-322. 
Malta, N. 1926. The genus Zygodon Hook. et Tayl. Latv. 
Bot. Darza Darbi 1: 1-185. 
Menzel, M. 1992. Preliminary checklist of the mosses of Peru 
(Studies on Peruvian Bryophytes IV). J. Hattori Bot. Lab. 
11: 175-254. 
Missouri Botanical Garden. Tropicos. <http://www.tropicos. 
org>, accessed 30 April 2009. 
Muñoz, i 1999. A revision e Grimmia (Musci: Grimmiaceae) 
n America. Ann. Missouri Bot. 


Univ. 


0. A revision of the neotropical i aed 
Musci (first part). J. Fac. Educ. Tottori Univ., Nat. Sci 
—154. 


1981. A revision of the neotropical on 
Musci (second part). J. Fac. Educ. Tottori Univ., Nat. Sci 
0: 21-55. 


Orme, C. D. L., R. G. Davies, M. Burgess, F. Eigenbrod, N. 
Pickup, V. A. Olson, A. J. Webster, T.-S. Ding, 
Rasm n, R. Ri : J. Stattersfield, P. M. 
Bennett, T. M. Blackburn, K. J. Gaston & I. P. F. Owens. 


2005. Global hotspots of species richness are not 
congruent with endemism or threat. Nature 436: 
019. 


Pittier, H. 1936. Los musgos de Venezuela. Bol. Soc. Venez. 
Ci. Nat. 3: 353-389. 
Pursell, R. A. 1973. Un censo de los musgos de Venezuela. 
Bryologist 76: 473—500. 
19 ; Taxonomie notes on neotropical Fissidens. 
Buegin: 97: 253-271. 
. Fissidentaceae. Fl. Neotrop. 101: 1-278. 
cha, J. V., P. Salaman. gensen, T. 
ni 


ere EUR 


ux & a Yonsscd (editors), 
otspots Revisited: Earth's Biologically Ric and Most 
Endangered Terrestrial Ecoregions. C X, Mexico City. 


Steere, W. C. 1948. Contibunöns to the bryogeography of 
Ecuador. I. A review of the species of Musci previously 
reported. Bryologist 51: 65— 167. 


Touw, A. 1974. Some notes on taxonomic and floristic 
research on exotic mosses. J. Hattori Bot. Lab. 38: 
123-128. 

Vergara, W., A. M. Deeb, A. M. Valencia, R. S. Bra 


y, B. 
Francou, A. Zarzar, A. Grünwaldt & S. M. The 
2007. Economic impacts of rapid glacier retreat in the 
Andes. Eos Trans. Amer. Geophys. Union 88: 261-268. 
Wilson, E. O. & F. M. Peter. 1988. p National 
Academy Press, Washington, D. 


APPENDIX l. Summary of moss diversity for the tropical 
h encompasses 1376 species and 327 


are denoted by an aste 
infraspecific Andean synonyms (929) is in bracket 


Ew oe 19/33 (4) [16] 

Amblysteg CD n Bruch, Schimp. & W. Gümbel 4, 
Anacan podon B 1 [1], Calliergonella [jns e 1, Camp 
liadelphus (Kindb.) R. S. Chopra 1, Campylium (Sull.) Mit. 
T, ao (Schimp.) M. Fleisch. a a air 
5 ull.) Spr 1 [4] Drepanocladus (Mil G. Roth 

5 [5]. Gaiei Ochyra* 1 


— 


1) ets Hedenäs l, 


2 [l], Sanionia Loeske 1, Scorpidium (Schimp.) Limpr. 2 [2], 
Straminergon Hedenás 1, Vittia Ochyra 2 (1) [1], Warnstorfia 
Loeske 5 (1 
Andreaeaceae 2/9 (3) [2] 
Acroschisma Lindl. 1, Andreaea Hedw. 8 (3) [2] 
An a 2/2 [ 
Hook. & Taylor 1 [1], Herpetineuron (Müll. 
Hal) “Cand ot 1 
Aulacomniaceae 1/1 [2] 


Annals of the 
Missouri Botanical Garden 


Aulacomnium Schwagr. 1 [2] 

ror Ec 8/64. Er [46] 
pou T: [1] Bartramia T 12 (5) [21], 

purs (Bruch & Schimp.) Schim 0 (4) [13], Con- 
ostomum Swe ex E a Bo [2], Flowersia D. 
G. Griffin € W s Broth. 6 (4) [3], 
Philonotis Brid. E G a [sue Bri 
Brachytheciaceae 14/45 (11) [57] 

Aerolindigia M. Menzel 1 ales Brachythecium Schimp. 
& Schimp. 4 [2], a 
doniella Hog 1 (. Lindigia Hampe* 

1 (D [1]. nn (Müll. Hal.) Manuel 2 [4], scd 
cladium Py nidium M. Fleisch. 2 (1), 
Rhynchos oa. a & Schimp. 4 Besch. 2, 
Squamidium (Müll. Hal.) Broth. 7 [17], Sterócarpidiopsis M. 
Fleisch. ex Broth.* 1 (1), Zelometeorium Manuel 5 [4] 


Bruchiaceae 2/3 (1) 
obruchia W. R. B 1 (1), Trematodon Michx. 2 
Bryaceae 11/130 (46 P 
Acidodontium Schwägr. 11 (9) [5], n Schimp. 9 
(6) [3], Brachymenium Schwägr . 13 (2) [15], Bryum Hedw. 44 
(9) [31], Epipterygium Lindb. 1 [1]. leporum 1 


Wilson 2 [3], Mie E Nees & Hornsch. 4 (3) [1]. 
Orthodontium Schwügr. 3 (1) [3], Pohlia Hedw. 11 [6], 
Rhodobryum (Schimp) pes 8 (1) [2], Schizymenium Harv. 
24 (15) [23 

wo y a 2 

Calym F. Weber 8 [2], Leucophanes Brid. 1, 
Syrrh rum owed 19 [21] 
wo oc na 

Cai Müll. Hal. ex Broth. 2 
Ciyphaeaceae 5/16 (6) [12] 

Cryphaea D. Mohr € F. Weber 10 (4) [12], Dendrocry- 
phaea Paris & Schimp. ex Broth. 1 (1), Dendropogonella E 
Britton 1, Tau Ud Dozy & Molk. 2 (1), Sphaerothe- 
ciella M. Fleisch. 2 
Daltoniaceae 4/31 (18) [9] 

Adelothecium Mitt. 1, Calyptrochaeta Desv. 4 (3), Daltonia 

aylor 17 (11) [9], Leskeodon Broth. 9 (4) 
Dicranaceae 28/129 (37) [107] 
dium Schimp. 1, Aongstroemia Bruch & Schimp. 3 
H ier Mitt. 2 (1) [6], d P. de la 

Varde € Thér. 1 [2], Camptodontium Dusén 2 (2), 
Campylopodiella. Cardot 1 [1] Md Brid. 49 (12 
[5]. Dicranella 

m Bruch & 


Ss 


a 2 [1], D 
Hedw. 2 [3], Eucamptodontopsis Br oth. 1 (1), Holodontium 
(Mitt) Broth. 1 [1], Holomitrium Brid. 10 (1) [7], Hygro- 
dicranum | Cardot (1), Leucoloma E o [1]. 
Microcampylopus (Müll. Hal.) M. Fleisch. 1 [2], Microdus 
Schimp. ex Besch. 6 (2), Oreoweisia d & Schimp.) De 
Not. 4 (2) [4], Orthodicranum (Bruch & Schimp.) Loeske 3 
(1), Pilopogon Brid. 6 (3) [4], Polymerodon Herzog* 1 (1), 
Pseudohyophila Hilp.* 1 (1), Rhabdoweisia Bruc chimp. 

3 [1], Schliephackea Müll. Hal. 2 (1), Sphaerothecium Hampe 

] 


— 


Diphyscium D. Mohr 2 (1) [1] 
Dine 10/21 (10) [10] 

Astomiopsis Müll. Hal. 2 (1), Bryomanginia Thér. 
Ceratodon Brid. 2 [3], Chrysoblastella R. S. Williams i [2]. 
Distichium Bruch & Sc i l, Ditrichum Hampe 6 (3) [4]. 
Pleuridium Rabenh. 4 (4), R e EE Mitt. 1, Tristichium 
Müll. Hal. 2 (1), SUN Müll. Hal. 1 (1) 
neal ppiacenc 1/3 4] 

Encalypta Heic 3 (1) [4 


Re 3/11 (2) [8 
Entodon Müll. Hal. B E [4], Erythrodontium Hampe 2 [2], 


1/4 
m (Brid.) Brid. 4 
bas V1 


Eustichia oat Brid. 1 [1] 


[3 
hodon d 9 6 ) [28], Funaria Hedw. 2 [3]. 
caian (Brid.) Brid. 1 
Gigaspermaceae y) 
Loreniziella Müll. Hal. ; , Neosharpiella H. Rob. & Delgad. 1 
Grimmiaceae 8/40 (8) [4 
ss eee us Coscinodon Spreng. 
1 (1), Grimmia He m 
3 a 0 [1], Ptychomitrium 
ee Racomitrium Brid. 7 (1) [2], Schistidium Bruch 
= Schimp. 7 (2) [8 
E ad 3/12 (1) [9] 
Bruch & Schimp. 10 (1 ) [7], Hedwigia P. Beauv. 
1, Hedw igidium Bruch & Schimp. 1 [2] 
Helicophyllaceie 1/1 
Helicophyllum Brid. 1 


fiodkeniecas 1A 


o a Sm. 
Hylocomiaceae 2/3 (1) [1] 
S m M. Fleisch. ex Broth. 1, Pleurozium Mitt. 2 
[ 
Hypnaceae 17/33 (5) [24] 
Caribaeohypnum Ando & Higuchi 1, Chryso-hypnum 
Hampe 2 [1], Ctenidium (Schimp.) Mitt. 1 [2], Ectropothecium 
Miu. 2 (1) [1], diae Broth. 1 [1], Hypnum Hedw. P [2]. 
1) [7], Mitenlhamnium Henn. 8 (3) 
ygyriella Catdor IP 


1] os M M. 
Fleisch. 3, eles Müll. Hal.) Müll. Hal. 1 [1 
Hypopterygiaceae 1/1 [8] 


= 


[5] 
Orthostichella Müll. Hal. 4 
Besch. 3 (1) 
Leptodontaceae 1/3 
Forsstroemia Lindb. 3 
E ss 1/1 [2] 
Lep n Hampe E [2] 
E 6/11 (4 
Haplocladium a Hal.) rd Hal. 1, Leptopterigynan- 
drum Müll. Hal. 4 [2], Leskea Hedw. 3 (2) [1], Leskeadelphus 
Herzog* 1 (1) [3], Lindbergia Kindb. 1, Pseudoleskea Bruch 
& Schimp. 1 (1) [1] 
Leu e e 2/10 [3] 
Leuc m Hampe 7 [2], Ochrobryum Mitt. 3 [1] 
Layee 2/2 [3] 
on Schwágr. 1 [2], Pierogoniadelphus M. Fleisch. 


= 


5], Pilotrichella (Müll. Hal.) 


Leucomiaceae 2/4 5 
Leucomium Mitt. 1 [2], Rhynchostegiopsis Müll. Hal. 3 [1] 
Macromitriaceae 5/47 (11) [14] 


Volume 96, Number 3 
2009 


Church 
Moss Diversity and Endemism 


nct id l. Groutiella Steere 7 [1], Macrocoma 


(Horns Hal.) o [2], Macromitrium Brid. 30 
(11) O, pores Brid. 5 [2] 
Meesiaceae 1/2 

Me 


] 
Barbellopsis Broth. 1, Floribundaria M. Fleisch. 1, 
nea Broth. 1, Meteorium (Brid.) Dozy & Molk. 
mra mE Trachypus Reinw. & 
Hornsch. 1 
Mniaceae 2/2 [1] 

Mnium Hedw. 1, Plagiomnium T. J. Kop. 1 [1] 
Myriniaceae 1/1 [2 

Helicodontium (Mitt.) A. Jaeger 1 [2] 

Neckeraceae 9/26 (4) [34] 

Brid. e Isodrepanium Sr E. Britton 1, 
Neckera Hedw. 6 (2) [12] Neckeropsis Reichardt 2, 
Pinnatella M. Fleisch. 1 [1], Porot "ead ndron M. Fleisch. 

Porotrichopsis Broth. & Herzog* 1 (1), Porotrichum 
(Brid) id 10 zr [14], i e Nieuwl. 1 [1] 
Octoblepharace: 

Pur cim ee 
Orthotrichaceae 2/51 ra [37] 

D 18 (1D [16], Zygodon Hook. & Taylor 


uo 


33 (22 

o 1/3 [5] 
Phyllogonium Brid. 3 [5 

Pilotrichaceae 19/109 a [61] 

Actinodontium s 1, Amblytropis (Mitt.) Broth. 3 um 
Brymela Crosby Allen 7 (5) Callicostella (Müll. Hal.) 
Mitt. 9 [1]. oia Broth.* Crossomitrium 
Müll. Hal. 5 [2], Cyclodictyon Mitt. 15 (1 0) [5], Helicoble- 

th. 


A. Jaeger* 1 (1) Stenodictyon (Mitt.) A. Jaeger 1 E 

rn E tt.) M. Fleisch. 9 (3) [2], Trachyxiphium W. 

R. Buck 1 
Lo. 1/4 [5] 

Plagiothecium Bruch $ Schimp. 4 [5] 
Polytrichaceae 9/23 (10) [35] 

Atrichum P. Beauv. 2 [2], oe G.L 
Oligotrichum D Pogonatum P. Beauv. ] 
He o qo (Mill. Hal.) Mitt. 6 (6) [2], Pieces 

2d he M E 3 (1) [7], Psilopilum Brid. 
T2 


5 m Cardot 6 (5), 
Anoectangium Schwágr. 1 [1], Barbula Hedw. 10 (5 
nl, Bryoerythrophyllum P. C. Chen 10 
(2) b. Colyptapogon ud Broth. 1, Chenia 
(2), € m Jur. 1 (1), Didymodon Hedw. 26 (10) [12], 
eR H. Za nder 1, Erythrophyllastrum R. H. Zander 
1, Erythrophyllopsis 2 1 (1) [3], Gertradiella Broth. 1 (1), 

obulinella Steere 2 (1), Gymnostomiella M. Fleisch. 1, 
Gymnostomum Ne 


= 
— 
= 

E 
= 


Hal.) Hampe ex Lind). 17 (4) 
[1], Molendoa Lindb. 1 [2], Ploubelie Brid. 1, Bleuroclinere 


Lindb. 1 [1], Pseudocrossidium R. S. Williams 9 (3) [2], 
Pseudosymblepharis Broth. 1, Rhex Pi ail "s 1 (1), 
Sagenotoriula R. H. Zander 1 [1], Saitobryum R. H. Zander 1 
[1], Scopelophila (Mitt.) Lindb. 2 d Sepia e 
al 1, Sire E, or m ex Mitt. 
Prud e Her "iri Synir n Brid. 26 o ue 


d () DL yodontium Steere" 1 (D. 
omum Bruch 7 (2 ) n ae Hedw. 3 [1] 
Prionodontaceae 1/5 (1) [1 
Prionodon Müll. Hal. 5 a [16] 
Pterobryaceae 9/17 (3) [6] 
rus pu eue (Müll. 
indb. 1, Hal. Orthostichidium Müll. 
Hal. ex Dusén 1 [1], Due pe Broth. 3 [1], Pireella 
Cardot 5 (1) al Pierobryon Hornsch. 2 (1) [2]. Pterobryopsis 
M. Fleisch. 1 (1), SN Müll. Hal. ex Renauld 1 


Hal.) 


E sce 1/1 Tu 

Regmatodon Brid. 1 [1] 
Rhacocarpaceae 1 (1) [3] 

Rhacocarpus Lindb. 4 (1) [3] 
c ed 4/5 [2] 

Hymenodon Hook. f. & Wilson 1, n Schwagr. 1 
[1], | s Mitt. 2, Rhizogonium Brid. 1 [1] 

Rhytidiaceae 1/ 

Rhytidium (Sull.) Kindb. 1 
m cde 14 [3 

Rigodium Kunze ex Schwágr. 1 [3] 
oed 1/1 
eudocryphaea " Britton ex Broth. 1 

d m 2/3 

Bii Bruch ls ques 2 (1), Brachydontium Fürnr. 1 
or mE d (21 ] 

cropori 1) [2], Allioniellopsis Ochyra* 1 (1), 

— “Broth) 1 [4], Aptychopsis (Broth) M. 
Fleisch. TO Donnellia Austin 2, Heterophyllium e) 

Kindb. 1, Meiothecium Mitt. 2, Pierogonidium Müll. zb 
prn Haber 1 (1), Sematophyllum Mitt. 24 WS [4]. 
uce ex m 1 [1], Timotimius W. R. Buck* 
0 (1), Wikia H. A. Crum 3 (1) 


fi] 


vus 
a, 


ithelium Spr 


Sphagnum L. 61 (35) [5] 

p asi 5/12 (3) HR 

Brachymitrion Taylor 4 (1) [1], Splachnum Hedw. 2, 
Tayloria a (2) [1], Tetraplodon Bruch & Schimp. 1, 
Voitia Hornsch. 1 
Spl a 1/1 

Splachnobryum Müll. Hal. 1 
Stereophyllaceae 6/10 (2) [2] 

Entodoniopsis Broth. 4 [1], Eulacophyllum W. R. Buck & 
Ireland 1, Juraizkaea Lorentz 1, Pilosium (Müll. Hal.) i 
Fleisch. 1, ae Hampe ex Broth.* 2 (2) [ 
Stereophyllum Mitt 

saiiphgadenticeds a 1] 
odon E a n 
Thuidiaceae 3/15 
Pelekium Mitt. 8 o 
Bruch & Schimp. 5 [ 


aS 
X 


a Rauiella Reimers 2 [1], Thuidium 


ANDEAN SPECIATION AND Lena Struwe,*? Scott Haag,* Einar Heiberg,” and 
VICARIANCE IN NEOTROPICAL Jason R. Grant 

MACROCARPAEA 

(GENTIANACEAE-HELIEAE)! 


ABSTRACT 


The genus Macrocarpaea (Griseb. ) Gilg d Helieae) is among the largest woody genera of tropical gentians, with 
most of its species occurring in the wet mountainous forests of the Andes. Phylogenetic and dispersal-vicariance analyses 
(DIVA) of 57 of the 105 currently recognized species in the genus, using two data sets iis sd DNA (ITS and 55- NTS 
sequences) and morphology, show a single origin of the Andean species from an ral distribution that includes 


northern Ecuador, Colombia, and Venezuela; and (2) southern Ee of the Andean Amotape—Huancabamba Zone in 
Ecuador and Peru, as well as the Andes of central and souther and Bolivia. The Amotape-Huancabamba Zone is 


and three Mesoamerican species derive from the northern clade, as is the single sampled species from the Guayana Shield. 
The position of the sube lade of the three Caribbean species is less certain, but it currently nests among Andean species. An 
Atlantic coastal Brazilian clade is placed as sister group to all other Macrocarpaea, providing further support o an O 


refuge in southeastern Brazil for ies Helieae. The a ide showed that local speciation is more n than 
long-distance dispersal, and allopatric seta n than sympatric speciation. Using detailed, A dE; 
herbarium collection data, PR in environme id is es ades and sister species were analyzed with 
ial Evolutionary and mara Vicariance Analysis EVA), ped a Url system (GIS) an 
statistical methods. Sister clades an valuated for statistical l rainfall and 
temperature, elevation, temperature s rainfall seasonality, geological esc age, an il nie to Dind ecological 
vicariance between sister groups. The results indicate t e are no general patterns for each variable, but that there are 
many significant divergences in ecological bet h larger sister gro d sister species, and ecological niche 


pared. 
Key words: Biogeography, ecology, Gentianaceae, Macrocarpaea, Neotropics, niche, South America, speciation, 
vicariance. 


The tropical Andes are one of the most biologically due to recent geological uplift. Biodiversity in South 
diverse areas in the world and a biodiversity hotspot America and Andean biogeography has been reviewed 
(Myers et al., 2000; Rodríguez-Mahecha et al., 2004a, by Burnham and Graham (1999) and Young et al. 
b; Jørgensen et al., 2007; Morawetz € Raedig, 2007), — (2002), highlighting the complex history, topography, 
despite being a relatively young part of South America and vegetation patterns of the Andes. The high 


‘This study was funded by the National Science Foundation a 0317612) and USDA- ae ene (Hatch 
211) to L.S., and the Swiss National Science Foundation (g rants 3100-052885, 3100-065395) and Swiss Academy of 
Sciences ree to IR E The authors wish to express their sincere thanks to Peter Smouse and tis ard G. Latine for 
d collaboration with SEEVA development. We also thank the following herbaria and their staff for 
ee ee help in coe their collections and providing information: AAU, AFP (Herbario “Alvaro Fernandez Pérez,” 
Popayán, Cauca, Colombia), ALA, B, BM, BP, BR, BRIT, BSB. C, CAS, CAUP, CHOCO, CHRB, COAH, COL, CONN, CR, 
CUVC, CUZ, DAV, DUKE, E, EHH, F, FAUC, FI, FLAS, FMB, FR, G, GB, GH, GOET, HAC, HAL, HAM, HAO, HUA, 
HUCP, HUQ, HUT, IAN, INB, INPA, JAUM, JBSD, JE, K, L, LD, LINN, LOJA, LPB, LS, M, MA, MANCH, MARY, MBM, 
MEDEL, MER, MG, MICH, MIN, MO, MOL, MSB, MU, MY, NA, NEU, NO, NSW, NY, OXF, P, PH, PORT, PR, PRC, Q, 
QAP, QCA, QCNE, QPLS, QUSF, R, RB, RNG, 5, SBBG, SEL, SP, SPF, TEX, U, UC, UCWI, UDBC, UPCB, UPS, UPTC, US, 
M, VALLE, VEN, W, WIS, WU, YU, and Z. 
? Department of Eu volition and Natural Resources, Rutgers University, 14 College Farm Road, New Brunswick, 
New Jersey 08901, U.S.A. st edu 
? Department of Plant Biology and fay Rutgers University, 59 Dudley Road, New Brunswick, New Jersey 08901, 
A. 


* Center for ON Sensing and Spatial Analysis, Rutgers University, 14 College Farm Road, New Brunswick, New Jersey 
08901-8551, U 
3 Department a Clinical Physiology, Lund University Hospital, 221 85, Lund, Swe 
ê Laboratoire de botanique évolutive, Institut de Biologie, Université de ue rue Emile-Argand 11, CP 158, 2009 
Neuchátel, Switzerland. 
doi: 10.3417/2008040 


ANN. Missouni Bor. Garp. 96: 450-469. PUBLISHED ON 28 SEPTEMBER 2009. 


Volume 96, Number 3 
2009 


Struwe 
Andean Speciation in Macrocarpaea 


biodiversity has been explained through habitat 
diversity resulting from large differences in bedrock 
types, soils, climate, and elevation, which are also 
fragmented due to the dissected topography. Species 
have spread north and south along this jagged 
mountain chain, and also up and down in elevation 
during colder times through repeated glaciations. Both 
allopatric and sympatric speciation processes could 
have resulted in such diversity. 


arge species 


Generally speaking, the theory behind sympatric 
speciation suggests that competition among popula- 
tions led to subsequent ecological niche divergence, 
whereas allopatric speciation invoked a spatial barrier 
and not necessarily any ecological niche separation 
(Mayr, 1963). This is applicable not only to extant 
sister species pairs, but also to deeper clades in the 
phylogeny that represent ancient allopatric or sym- 
patric events 

Progress in iaa developments has now 
made it possible to evaluate phylogenetic patterns and 
pn ion pen. ses in a historical and spatial context 

ncluding d hi 


m and morphological data for phylogenetic 


cological data. In this study, we use 
reconstruction using parsimony criteria, dispersal- 
(DIVA) 
ancestral areas (Ronquist, 1997), and Spatial Evolu- 


vicariance analysis for reconstruction of 
tionary and Ecological Vicariance Analysis (SEEVA) 
— 

n this paper, we will evaluate several hypotheses 
based. on a comprehensive data set from Macrocarpaea 
A ) Gilg. 

Continental patterns: What is the relationship 

between s PP in the Andes to o in 
e Caribbean, southeastern Bra- 
zil, and tepuis i the Guayana em 


Mesoamerica, t 


2. Andean patterns: What is the ancestral distri- 
bution area in the Andes and what large-scale 
Andean biogeographic patterns are present? 

3. Speciation patterns: Is sympatric or allopatric 
speciation most common in Andean species? Is 
dispersal or vicariance most common in Andean 

ecies? 

4. Ecological niche patterns: Does allopatric spe- 
ciation indicate a relatively larger shift in 
ecological niche characteristics than sympatric 

tion that occurs w e same area? Do 

opatric) versus overlapping (sym- 
patric) sister taxa have similar or divergent 
ies oe niches? 

Our study group, Macrocarpaea (Gentianaceae, 
Helieae), occurs in mountainous regions of th 
Neotropics at (80—)1500-3000(-3800) m elevation. 
It is composed of woody shrubs, small trees (up to 
10 m), and herbs (one species; M. rubra Malme) with 
large (up to 7 cm long), campanulate, white, yellow, to 


green flowers that are visited diurnally by humming- 
birds, butterflies, and insects, and nocturnally by bats 
and moths (Grant & Struwe, 2001; Grant, pers. obs.). 
Its woody habit is uncommon in the Gentianaceae and 
led to a comparative study of its wood anatomy 
indicating secondary derivation of woodiness d 
herbaceous ancestors (Carlquist & Grant, 2005). W 

105 species currently recognized, Macrocarpaea is 7 
far the most species-rich genus in the Helieae, while 
the majority of genera in this tribe are either 
han 10 €g., 
Celiantha Maguire and Yanomamua J. R. Grant, P. 
J. M. Maas & Struwe (Struwe et al., 2002; Grant et al., 


monotypic or have less t species, 


During monographic studies on Macrocarpaea, 
3500 he 


more than arium sheets were examined, 
resulting in the wo of 70 new species (Grant 
& Struwe, 2001, 2003; Grant, 2003, 2004, 2005, 
2007, 2008; Grant & Weaver, 2003). Fieldwork by 
rant, RUN in the Huancabamba region of 
Ecuador and Peru, continues to uncover new species 
that are often a different from one another 
when seen as arium specimens, but are otherwise 
clearly a in the fiel Wee. 


2009) 


has given strong support for its monophyly, for the 


phylogenetics of Macrocarpaea (Struwe et al., 
establishment of an infrageneric classification (Grant, 

and to the understanding of difficult species 
oni. 

Macrocarpaea has a broad distribution in moun- 
tainous regions of the Neotropics and is comprised of 
generally narrowly endemic species. It occurs princi- 
pally in the Andes (87 species from Venezuela, 
Colombia, Ecuador, Peru, and Bolivia), with outlying 


thern Mesoamerica (five s 
Costa Rica and Panama), in the Atlantic forest of 
southeastern Brazil (four species comprising section 


Caribbean (three species: one 
Hispaniola, and Jamaica). The three main centers of 
diversity are in the Colombian Andes (section 


(section Choriophylla (Griseb.) J. R. Grant, 74% of 
its species), and the central Peruvian Andes (section 
Magnolifoliae Ewan, 50% of its species). The genus is 
absent from lowland Amazonia, Mexico, Mesoamerica 
north of Costa Rica, and southern temperate South 
America. Here, we use Macrocarpaea as a model for 
understanding South American, and particularly 


Andean, 


Because Macrocarpaea has many narrowly distributed 


speciation, biogeography, and dispersal. 


Andean endemics, it is an excellent candidate for this 


kind of study. 


Annals of the 
Missouri Botanical Garden 


MATERIALS AND METHODS 
PHYLOGENETIC DATA AND ANALYSES 
Fifty-seven of 105 total species (54%) of Macro- 
carpaea, representing all sections as defined b 
wer 


(2005), e included in the phyl 


(Table 1). Five outgroups were selected for orientation 


y Grant 
ogenetic analysis 


of the phylogenetie trees (Nixon & Carpenter, 1993), 
the Gentia- 


naceae and all of tribe Helieae using trnL and matK 


based on previous phylogenetic work in 


chloroplast data as E as nuclear ITS markers 
(Struwe et al., 2002, 2009). xa used include 
Symbola. niu jasonii a E. Molina & e from the 
Symbolanthus G. Don e of Helieae, 
Pa psychotrioides Ewan and three species 
of Tachia Aubl. (T. grandiflora Maguire & Weaver, T. 

guianensis Aubl., and T. occidentalis Maguire & 
Weaver) assigned to the Macrocarpaea subclade 


and 


(Struwe et a 

The DNA data in this study were developed for a 
detailed phylogenetic study of the genus by Grant. For 
molecular data we used two different nuclear genom 
areas, ITS and NTS (for 5S RNA [5S-NTS]. These 
regions have successfully resolved phylogenetic 
relationships in tribe Helieae on both generic and 
tribal scales in the past (Struwe et al., 2002, i 
Gould & Struwe, 2004. Additional sequences for 5 
species of Macrocarpaea and the selected s 
were downloaded from GenBank (see Table 1 for 
GenBank numbers). In total, 47 ITS and 56 5S-NTS 
taxon-specific sequences were obtained from Macro- 
carpaea 
(Thompson et al., 1997 
(Rambaut, 2002). 


In addition to molecular data, nine morphological 


equences were aligned using ClustalX 


and a Al v.2.0 for Macintosh 


characters were coded for all included species 
(Table 2), and these were primarily selected from 
pollen and seed surface structure (Grant & Struwe, 
2001, mer Grant, 2003, 2004, 2005, 2007; Grant & 
Weaver, our i 
identified e ‘Grant (2005) 


without wings (flattened and a and two with 


ed types were 
ncluding two types 


wings (perimetrically winged and winged). Character 
1 refers to the general shape of the seeds, varying from 
three-dimensional (rectangular to spheroid) to com- 
pletely flattened, as does character 4 separating seed 
types with eit 


er a square or long-linear outline. 
Character 2 refers to the presence or absence of 


seed 
wings, and character 3 is Pd based | on whether 
hese wings oc opposite ends as in a bowtie 
(winged type) or ueber n" E MEN (perimet- 
rically winged). Character 5 is the weight of an 
individual seed in PA vie! less than 

1 mg in Macrocarpaea and g r than 0.1 mg in 


Chorisepalum Gleason & ee Tachia, and 


Symbolanthus. Character 6 refers to features of the 
pollen exine mede and character 9 identifies 
s or as tetrads 
P dieere de axilla 
flower position of Tachia, as compared to all other 


or polyads. Character 


genera that have multiflowered terminal inflorescenc- 
es, and character 8 codes for the color of the corollas. 
Character states were treated as unordered, and 
characters were regarded of equal weight. The two 
DNA alignments were combined with the data matrix 
of nine morphological characters and analyzed 
together simultaneously. 

The phylogenetie, unconstrained parsimony analy- 
sis was performed in Winclada and NONA (Goloboff, 
1999: Nixon, 1999, 2002) using 500 random repli- 
cates, five starting trees per replicate, MaxTrees as 
10,000, and the tree-swapping algorithm was multiple 
tree ection: reconnection (TBR) + TBR. NA 

g was done with 300 replicates in PAUP* v.4 
(Swofford, 2000). 


COLLECTION DATA 


A total of 794 georeferenced herbarium collection 
records of the 57 Macrocarpaea species included were 
entered into an existing FileMaker Pro 7.0 (Claris Pro, 

anta Clara, California) database. Each locality was 
georeferenced using printed and online maps, atlases, 
and gazetteers. Identification of all collections was 
confirmed by Grant, and an exsiccatae list with the 
georeferenced coordinates is available upon request. 
Only recorded locations known to the nearest minute, 
nearest second, or label with global positioning system 
GPS) coordinates were included in the analysis. 
Coordinates were converted t t shapefile in 
ArcGIS v. 9.2 (ESRI, Redlands, “Califor ia), and 
distribution maps of each species and the genus were 
produced. 


BIOGEOGRAPHIC ANALYSIS 


Nodes present in the selected most parsimonious 
tree were classified as either representing sympatric, 
partial ly o or allopatric pd distributions 
based on data from individual species maps. This 
dissivaton was we in comparing PUES, indices 
and P values for environmental species and clade data 
during the SEEVA analysis. 

For the DIVA analysis, the 
was divided into 10 biogeo, 


spatial structure and geographic connectivity, and 


total genus distribution 
graphic units based on age, 


geological coherence. The identified areas were: A, 
s Brazil; B, Pantepui of the Guayana 
; C, Greater Antilles of the Caribbean; D, 


eadem E, Cordillera Central in Colombia and 


Volume 96, Number 3 
2009 


Struwe 
Andean Speciation in Macrocarpaea 


Cordillera Oriental in Ecuador; F, Chocó and 
Cordillera Occidental in Colombia and Ecuador; G, 
Cordillera Oriental and Mérida in Colombia and 
Venezuela; H, Amotape—Huancabamba region in 
Ecuador and Peru; I, Cordillera Central in central 
Peru; J, Bolivia and Cordillera Central in southern 
Peru; K, Amazon Basin in Brazil, Colombia, Peru, and 
Venezuela. A map of the areas in northwestern South 
America is presented in Figure 1 
Áreas were delimited based on spatial, historical, 
and geological features; some outliers were clea 
disjunct, such as the Caribbean (C), tepuis on the 
Guayana Shield (B), and southeastern Brazil (A). For 
areas that are part of the Andean mountain chain and 
the Mesoamerican land bridge, their distinction is less 
obvious. The boundary between Mesoamerica (D) and 
the Chocé and Cordillera Occidental in northwestern 
Colombia (F) follows evidence that lowland Darién in 
eastern Panama has a stronger geological and 
historical connection to northern South America than 
to western Panama (Clapperton, 1993). The three 
cordilleras in Colombia and two in northern Ecuador 
were treated as separate areas because deep valleys 
nad the mountain ranges. The Cordillera de 
Mérida n Venezuela is an extension of the 
Cordillera pem of Colombia and is included in the 
same a The division between the Andean area of 
the poner pine dE Zone (H) to the south and 
the Andean areas in northern Ecuador and Colombia 
is correlated with the Amotape Cross, a 
geological shear zone at ca. 3°S in the continental 
crust of the South American Plate VAR 1993). 
The southernmost limit of area H is from Río Chicama 
n the western side and 
side of the Ande 
zone found in other biogeographic studies (Weigend, 
2002). The two southernmost Andean areas (I, J) in 
Peru and Bolivia are separated at the latitude of Rio 


Rio Huallaga on the eastern 
s and corresponds to a disjunction 


Pisco and Rio Entero, which corresponds to the 

geological Abancay Deflection, another shear zone 

(Clapperton, 1993). 

ersal-vicariance analysis was run using the 
VA 1 (Ronquist, 1996), limiting the 

maximum ie pP optimized areas at each node to 


is 
software DI 


three (maxareas — 3) and optimizing dispersal and 
vicariance events selected most parsimonious 
tree, because DIVA requires a fully resolved tree for 
its analysis. This tree was selected from all most 
parsimonious trees by being overall the most similar to 
the majority-rule consensus tree. The DIVA distribu- 
tion matrix is shown in Table 3. The DIVA optimi- 
zation method uses the parsimony criterion to limit the 
numbers of events (— steps) per tree based on the 
distribution of each species. Ancestral areas were 
optimized at each internal node in a way that limits 


vicariance and dispersal events as far as possible (i.e., 
it provides the simplest E of the current 
distribution data given the phylogeny). At several 

no there were several equally likely area optimi- 
zations and we selected from among these based on 
additional data from geological and spatial informa- 
tion. 


SEEVA ANALYSIS 


The theoretical background and statistical expla- 
nations and justifications for are outlined in 
Struwe (2008) and Heiberg and Struwe (2008), with a 
short overview given here. Environmental data were 
extracted using GIS from all canes collection 
s divided 


into four or five categories that were either quantita- 


localities of each species. Each variable was 


tive (e.g., rainfall amounts, in equally sized quartiles 
for the data od or s (e.g.. soil types) 
depending on as prepared in 
Microsoft Excel eat E Washington) 
that lists the number of collections per species for 
each variable (e.g., annual rainfall) and within each 
variable, for each category (e.g., 1-340, 341-732, 
733-1277, 1278-4000 mm). The table was imported 
into the software SEEVA v. 0.4 (Heiberg, 2008) for 
statistical analysis. The selected phylogenetic tree 
was then imported into A. 

Environmental point data for nine variables were 
extracted for the 794 Macrocarpaea collections using 
the following base layers in ArcGIS 9.2: elevation 
(U.S. Geological Survey; unit: m; format: grid; scale: 
30 are second), soil type. (ESRI/ArcAtlas; format: 
vector; scale: 1:5-10,000,000), geology (unit: bedrock 


temperature seasonality (BIO4; standard deviation X 
100; format: grid; scale: 1 km), minimum temperature 
of coldest month (BIO6; unit: °C X 10; format: grid; 
1 km), annual precipitation (BIO12; unit: mm; 
: grid; scale: 1 km), precipitation of driest 
month (BIO14; unit: mm; format: grid; scale: 1 km), 
precipitation seasonality (BIO15; unit: mm; format: 
grid; scale: m), and precipitation seasonality 
(coefficient of variation; format: grid; scale: 1 km). 
All climate data were retrieved from WORLDCLIM 
(Hijmans et al., 2005). 

For each node in the phylogenetic tree, extracted 
environmental data were pooled to represent data for 
each monophyletie group (clade) and compared with 
their sister group for each node. Ecologie data are 
measured for statistically significant differences be- 
tween clades based on the distribution in the four or 
five categories. All variables were analyzed indepen- 


454 Annals of the 
Missouri Botanical Garden 


Table 1. Voucher and GenBank accession numbers for 5S-NTS and ITS sequences of Macrocarpaea and outgroups used 


for the phylogenetic analysis. N/A indicates sequences not available. 


Taxon Voucher 5S-NTS ITS 

Macrocarpaea angelliae J. R. Grant & Ecuador, J. R. Grant 4289 (NY) EU541681 AY397760, 

Struwe AY397761 
M. apparata J. R. Grant & Struwe Ecuador, J. R. Grant 4002 (NY) EU541683 — DQ401413 
M. arborescens Gilg Ecuador, J. R. Grant 4084 (NY) EU541686  EU528076 
M. auriculata Weaver & J. R. Grant Costa Ri . L. Wilbur & Almeda 16828 (F) EU541688 N/A 
M. cs ilg olivia, S. G. Beck 8745 (N 541690 | EU528078 
M. bubops J. R. Grant & Struwe Ecuador, J. R. Grant 4046 (NY) EU541692 . EU528081 
M. rane R. Grant Peru, V. Quipuscoa 2044 (F EU541694 N/A 
M. cinchonifolia (Gilg) Weaver Bolivia, $. G. Beck 24780 (NY) EU541696 | EU528084 
M. cochabambensis Gilg-Ben Bolivia, A. Geniry 44200 (NY) EU541697 528085 
M. densiflora (Benth.) Ewan Colombia, K. von Sneidern 2523 (S) EU541700 | EU528087 
M. dies-viridis J. R. Gran Ecuador, J. R. Grant 4352 541702 EU528089 
M. domingensis Urb. & Ekman Dominican Republic, D. Kolterman s.n. (JBSD) U541704 . EU528091 
M. eli Ecuador, G. Harling & dsdeisin 23442 (MO) | EU541706 4 EU528094 
M. ericii J. R. Grant Pera, E. Rodriquez 2926 ( 541707 528093 
M. fortisiana J. R. Grant eru, D. McCarroll 128 (NY) EU541709 . EU528095 
M. gatiaca J. R. Grant Ecuador, J. R. Grant 4209 (NY) EU541710 — DQ401414 
M. gaudialis J. R. Grant Colombia, R. E. Weaver 2650 (GH) EU541713  EU528097 
M. glabra (L. f.) Gilg Colombia, J. R. Grant 4310 (NY) EU541714  EU528098 
M. glaziovii Gilg Brazil, B. Rezende Silva 1318 (NEU) EU541774 
M. gondoloides J. R. Grant Ecuador, G. Tipaz 1051 (MO) EU541716 N/A 
M. harlingii J. S. Pringle Ecuador, J. R. Grant 4049 (NY) EU541721 = EU528104 
M. innarrabilis J. R. Grant Ecuador, F. Luisier 2 (LOJA) EU541723  EU528106 
M. jactans J. R. Grant Ecuador, J. Clark 8927 (NY) EU541725 . EU528108 
M. jensii al. R. Grant & Struwe a J. R. Grant 4047 (NY) EU541724  EU528107 
M. kuelap J. R. Grant u, J. R. Grani 3942 (NY EU541726 EU528109 
M. lenae J. R. Grant ador, J. R. Grani 4013 (NY) EU541727 EU528110 
M. luna-gentiana J. R. Grant & Struwe Ecuador, J. R. Grant 4028 (NY EU541728 EU528111 
M. luteynii J. R. Grant & Struwe mbia, £ Cabrera € H. van der Werff EU541730 N/A 

769 

M. macrophylla (Kunth) Gilg Colombia, J. R. Grant 4312 (NY) 541735 . EUS28113 
M. maguirei Weaver & J. R. Grant Peru, B. Maguire 61569 (NY) EU541736 EU528114 
M. micrantha Gilg , J. R. Grant 3966 EU541737 EU528116 
M. neblinae Maguire & Steyerm. ezuela, B. Maguire 36886 (NY) EU541739 
M. nicotianifolia Weaver & J. R. Grant Colombia, A. S. Barclay 3402 (US) EU541740 EU528118 
M. noctiluca J. R. Grant & Struwe Ecuador, J. R. Grant 4003 (NY) EU541742 = EUS28121 
M. normae J. R. Grant Peru, K. García 267 (NEU EU541743 — EU528122 
M. obiusifolia (Griseb.) Gilg Brazil, W. Thomas 14304 (NY) EU541775 = EU528125 
M. opulenta J. R. Grant Ecuad R. Grant 4347 (NY) EU541746 = EU528124 
M. ostentans J. R. Grant Peru, B. Wallnéfer 12968 (U) EU54174 EU528128 
M. pachyphylla Gilg Colombia, M. L. Bristol 1429 (GH) EU541748 | EU528129 
M. pachystyla Gilg Peru, J. Schunke-Vigo 5298 (NY) EU541751 . EU528130 
M. pajonalis J. R. Grant Peru, M. Weigend 545 EU541749 
M. papillosa Weaver & J. R. Grant Venezuela, Weaver 2629 (GH) EU541750 | EUS28131 

. pinetorum Alain Cuba, Bisse 49708 ( EU541753 .— EU528133 
M. pringleana J. R. Grant Ecuador, F. Luisier 1 (NY) EU541755 . EUS28134 
M. revoluta (Ruiz & Pav.) Gilg Peru, M. Weigend 5288 (NY) EU541756 528135 
M. robin-fosteri J. R. Grant Peru, M. Weigend 5777 ( EU541757  EU528136 
M. rubra Malme Brazil, M. Reginato 755 (NE N/A EU528138 
M. sodiroana Gilg Ecuador, J. R. Grant 4210 (NY) EU541758  EU528140 
M. stenophylla Gilg Peru, J. R. Grant 3932 EU541762 | EU528142 
M. subcaudata Ewan Costa Rica, R. L. Wilbur & Almeda 16828 EU541763 . EU528143 

(DUKE) 

M. subsessilis Weaver & J. R. Grant Ecuador, J. R. Grant 4020 (NY) EU541765 . EUS28144 
M. tahuantinsuyuana J. R. Grant Peru, F. Woythowski 6672A (MO) EU541766 N/A 


Volume 96, Number 3 Struwe 455 
2009 Andean Speciation in Macrocarpaea 
Table 1. Continued. 
Taxon Voucher 55-NTS ITS 

mnoides cr Gilg Jamaica, P. Acevedo- e 9700 (NY) EU541767 N/A 
M. ie Stan Costa Rica, S. Hill 17751 (NY) 541769 . EU528148 
M. viscosa (Ru i Pav.) Gilg Peru, M. Weigend i9 EU541770 | EU528149 
M. weigendiorum J. R. Grant Peru, M. Weigend 5363 (NY) EU541771 . EU528150 
M. zophoflora Weaver & J. R. Grant Peru, J. J. Wurdack 1618 (NY) EU541772 | N/A 
Chorisepalum psychotrioides Ewan EU709793 
Symbolanthus jasonii J. E. Molina Ecuador, J. R. Grant 4350 (NY) N/A EU528151 

truwe 

Tachia grandifolia Maguire & Weaver DQ401429 — DQ401418 
T. guianensis Au DQ401430  DQ401420 
T. occidentalis Maguire & Weaver DQ401427 | DQ401423 


dently through a modified chi-square analysis that 
nitude 
in trends and differences between clades on a scale 


gives an impact index (1) that measures the mag 


from O to 1 for each variable and each node. P values 
were caleulated using Fisher's exaet test. Statistical 
significance was established at P « 0.05 


RESULTS 


PHYLOGENETIC DATA AND ANALYSES 


The lengths of the alignments were 687 nucleotides 
for ITS 
combined molecular analysis yielded 160 most 


T 
[7] 


an 9 nucleotides for - 
parsimonious trees (1252 steps, consistency index 
[CI] = 0.65, retention index [RI] = 0.74; Fig. 2). Five 
major clades are identified as monophyletie groups 
(Fig. 2, I-VI. Clade I represents three species with 
flattened seeds and pollen of the Glabra 

southeastern Brazil (Macrocarpaea sect. Tabacifoliae 
sensu Grant, 2005; for distribution data, see Figs. 3, 
4) and is sister to the rest of the genus. Next is a large 


-type from 


dichotomy that is is A s between northern 

species with rimmed o metrically winged seeds 
and Glabra-type iuba (clades II and ID and 
southern species with winged or perimetrically winged 
seeds and Corymbosa- or Glabra-type pollen (clades V 
and VI). The Caribbean clade (clade IV) 
seeds and Glabra-type pollen is weakly supported as 


with rimmed 


sister to the clade with the more southern species. 
Clades II and HI are each monophyletic, potentially 
sister groups, and include species primarily from the 
northern Andes, the tepuis, and Mesoamerica. Clades 
IL, III, and IV correspond together to Macrocarpaea 
sect. Macrocarpaea sensu Grant (2005). The only 
sampled species of six from the tepuis, M. neblinae 
Maguire & Steyerm. is sister to M. gattaca J. R. Grant 
from the Andes in clade III, and all Mesoamerican 
species are placed in clade IL Within the southern 


clade, M. arborescens a is sister to all other species, 
wich are divided into lades V and VI. 


Macrocarpaea sect. PA E is paraphyletic in 


clades: c 


the sense of Grant (2005; including the viscosa clade 
of clade VI), and clade V. Clad 


ade VI corresponds to 


Macrocarpaea sect. Choriophylla. 


BIOGEOGRAPHIC ANALYSIS 


s on the selected most parsimonious tree were 

either allop 

partially sympatric (slightly I. or sympatric 
o 


Nodes 
classified | into atric (nonoverlapping), 


(overlapping) distributions based the extant, 
detailed spatial distribution of each species. Past 
distributions for common ancestors to extant species 
most likely are at least somewhat different from the 
current species distribution, but because such histor- 
ical information is unavailable, no better estimate is 

ssible from the cu 
addition, Pu DIVA 


dispersal and vicariance events that were taken into 


rrent species distributions. 
analysis dba ypotheses of 


account when nodes were e classified. The classification 


sympatric (= partially allopatric), and 24 allopatric 
nodes (Fig. 3). When this is done for the 19 species 
pairs in the phylogenetic tree, s are four sympatric, 
three partially sympatric, and 1 
this phylogeny 
Allopatry is more common than true sympatry with a 
ratio of 3:1 to 4:1, but this does not account for the 
partially sympatric/allopatric distributions. 


atric species 
lin, 


Op 
pairs, given a species sampling 


e analysis resulted in an exact solution 
during the optimization, and an optimal reconstruction 
for our data required a total of 29 dispersal events. 
Within Macrocarpaea, there are 17 dispersal events 
and six vicariance events. Ancestral area reconstruc- 
tions and invoked dispersal events are shown in 
Figure 4. 


Annals of the 
Missouri Botanical Garden 


Table 2. Morphological data matrix of 57 Macrocarpaea 


species and five CU from Chorisepalum, Symbolan- 
achia 


thus, and T. 

Taxa/character 123456789 
AE un dum 111111000 
Symb nii 00-011011 
Ata pn s 00-011100 

: guianensis 00-011100 
T. occidentalis 00-012100 
Macrocarpaea angelliae 111001000 
M. apparata 111001000 
M. arborescens 00-001000 
M. auriculata 00-001000 
M. bangiana 111001000 
M. bubops 111001000 
M. chthonotropa 111001000 
M. cinchonifolia 110100000 
M. cochabambensis 111001000 
M. densiflor 222221000 
M. dies-viridis 111001000 
M. domingensis 00-001000 
M. elix 111001000 
M. ericii 00-001000 
M. fortisiana. 22222000 
M. gatiaca 00-001000 
M. gaudialis 00-001000 
M. glabra 00-001000 
M. glaziovii 00-001000 
M. gondoloides 111001000 
M. harling 111000000 
M. innarrabilis 111001000 
M. jactans 110100000 
M. jensii 111001000 
M. kuelap 111001000 
M. lenae 111001000 
M. luna-gentiana 111001000 
M. luteynii 111001000 
M. macrophylla 00-001000 
M. maguirei 110101000 
M. micrantha 111001000 
M. neblinae 00-001000 
M. nicotianifolia 00-001000 
M. noctiluca 111001000 
M. normae ?9-220000 
M. obtusifolia 00-001000 
M. opulenta 111001000 
M. ostentans 22—-220000 
M. pachyphylla 00-001000 
M. pachystyla 110100000 
M. pajonalis 111001000 
M. papillosa 00-001000 
M. pinetorum 00-001000 
M. pringleana 111001000 
M. revoluta 111000000 
M. robin-fosteri 110100000 
M. rubra 00-001000 
M. sodiroana 111001000 
M. stenophylla 00-001000 


Table 2. Continued. 


Taxa/character 123456789 
M. subcaudata 111001000 
M. subsessilis 00-001000 
M. tahuantinsuyuana 110100000 
M. thamnoides 00-001000 
M. valerioi 00-001000 
M. viscosa 111001000 
M. weigendiorum 110100000 
M. zophoflora 222221000 


* Question marks indicate missing sone and hyphen 
indicate inapplicable pe acters. Characters a character 
states are: 1. Seed shape deno angular or 


ee o flattened D Seed wings: absent (0), present 
(1). 3 d wings: 2-sided (0), perimetrical (1). 4. Seed 
shape: tin jd o. mad -linear (1). 5. D weight: 


< 0.1 mg/seed (0), > mg/seed (1). 6. Pollen exine: 
verrucose Yd reticulate m smooth (2). 7. Flower position 
aa (0), axillary (1). 8. Corolla color: green, white, or 

'w (0), red or purple (1). 9. Pollen aggregation: monads 
s oan or polyads (1). 


The selected optimization (Fig. 4) shows an ances- 
tral area for Macrocarpaea in southeastern Brazil (A), 
Cordillera Occidental in Colombia and Ecuador (F), 
and the Amotape-Huancabamba Zone in northern 
Peru and Ecuador (H 


between the Brazilian species (clade I) and Andean- 


). The first vicariance event lies 


Caribbean-Mesoamerican species. The major division 
within the Andea 
north-south vicariant event a 


n species into bs subclades 
. The C m 
species (clade IV; C) are eid. as likely derived 
from Andean species (H). Dispersal patterns among 
ndean areas invoked from the DI analysis are 
outlined in Figure 1. Patterns within the northern 
clades ane 1 a III; E, F, and C) are complicated 
and also involve one dispersal to and one from 
Mesoamerica (D), and one to the Guayana Shield (B). 
Vicariance and dispersal between the northern 
Ecuadorian and Colombian cordilleras are detected 
ut represent only two dispersals: one from Cordillera 
Occidental (F) to Cordillera Central (E) and one from 
Cordillera Oriental (G) to Cordillera Central (E). 

The Amotape-Huancabamba Zone (H) is support- 
ed as the ancestral area for the southern clade. Clade 
V dispersed early southward from H to central Peru 
(D) and later to southern Peru and Bolivia (J) twice, 
with one back-dispersal northward to H. A similar 
pattern is found in e VL which | largely 
restricted to the Am 

and shows four dispersele southward (two al to I 
and J) and one back-dispersal from J to I. Northward 
dispersals to northern Ecuador and Colombia (E and 


e-Huancabamba Zone 


rom the southern clade are only found in two 
cases. 


Volume 96, Number 3 Struwe et al. 457 
2009 Andean Speciation in Macrocarpaea 


Venezuela 

6'00"N- ~ ( p 

2'00"N- 
0°0'0"- 


200"S4 


Brazil 


soos- 


&'oo's-] 


800" 


1000" 


1200" 


Bolivia 


M 
Elevation (m) 


= High : 6795 
14'00"S-] 


Low : 1 


16'00's-] 


0 180 360 si 
o ——3 


T T T 
82700"W 8070'0"W. 7800W 


Map of northwestern South America and southern Mesoamerica with the areas used for biogeographic analysis 


Cordillera Oriental in Ecuador; F, Chocó and Cordillera Occ dial, in «Colombia and A G, Cordillera Oriental and 

Mérida in Colombia and Venezuela; H, Amotape-Huancabamba region in Ecuador and Peru; I, Cordillera Central in central 

Peru; J, Bolivia and Cordillera Central in southern Peru; K, Amazon Basin in Brazil, Colombia, Pen. and Venezuela. Areas A, 
and K are not indicated on the m. 


Annals of the 
Missouri Botanical Garden 


Table 3. Distribution data matrix used for the DIVA 
analysis, with O indicating absent and 1 present for each 
specific area. See further information in Materials. and 
Methods and Figure 4 for area definitions. Taxon 
marked with 


further explanation. 


names 


* represent a larger clade; see text for 


Taxon/area ABCDEFGHIJK 

Chorisepalum 01000000000 
psychotrioides 

Symbolanthus jasonii 00000001000 
Tachia grandiflora 01000000001 
T. guianensis 01000000000 
T. occidentalis 00000001111 
Macrocarpaea arborescens 00000001000 
M. auriculata 00010000000 
M. bangiana 00000000010 
M. chthonotropa* 00000001000 
M. cinchonifolia 00000000010 
M. cochabambensis 00000000010 
M. ericti 00000001000 
M. fortisiana 00000000010 
M. gattaca 00000100000 
M. gaudialis* 00001000000 
M. glabra 00000010000 
M. gondoloides 00000100000 
M. jactans 00000001000 
M. jensii* 00000001000 
M. luna-gentiana 00000001000 
M. macrophylla 00001100000 
M. maguirei 00000000010 
M. neblinae 01000000000 
M. nicotianifolia 00000010000 
M. noctiluca* 00000001000 
M. normae 00000000010 
M. obtusifolia* 10000000000 
M. ostentans 00000000100 
M. pachystyla 00000000100 
M. pajonalis 00000000100 
M. papillosa 00000010000 
M. pinetorum* 00100000000 
M. pringleana 00001001000 
M. revoluta 00000000100 
M. robin-fosteri 00000000100 
M. sodiroana 00000100000 
M. subsessilis* 00000001000 
M. tahuantinsuyuana 00000000100 
M. valerioi 00010000000 
M. viscosa 00000000100 
M. weigendiorum 00000000100 
M. zophoflora 00000001000 


SEEVA ANALYSIS 


Nine environmental variables were analyzed for 56 


outgroup nodes 1 to 5; 
index and P values). Apa nodes had an average 


total impact index (i) of 0.39 (all variables and all 
nodes) whereas partly sympatric nodes had an 
average impact index of 0.26, and sympatric nodes 
0.41. The differences in impact numbers indicate 
some differences between the groups, but there are 
only slightly larger environmental differences between 
sister clades of sympatric nodes than between sister 
clades of allopatrie nodes. When terminal species 
pairs are analyzed, the results are more pronounced. 
The average impact index for allopatric species pairs 
was 0.49 (all des), for 
sympatric 0.40, and for sympatric 0.41. This finding 


variables and all no partly 
indicates that allopatric species pairs may have 
ecological niches more divergent from each other 
than both sympatric and partly allopatric species. 
Allopatric terminal species pairs are the more 
ecologically different than when allopatric nodes are 
compared. 
our Andean terminal sympatric species pairs were 
analyzed: Macrocarpaea densiflora (Benth.) Ewan 
versus M. pachyphylla Gilg in Cordillera Central of 
Colombia (Fig. 3, node 21), M. apparata J. R. Grant & 
Struwe versus M. elix J. R. Grant in southern Ecuador 
(node 37), M. bubops J. R. Grant & Struwe versus M. 
harlingii J. S. Pringle in southern Ecuador (node 38), 
and M. dies-viridis J. R. Grant versus M. lenae J. R. 
Grant in southern Bivsdoc (node 45). For node 21, 
only temperature seasonality showed significant 
differences between the two sister species (i — 0.54; 
< represent two 
sympatric species pairs that, in turn, are partially 
sympatric sister groups to each other, but the two 
species pairs show Pea differences in what 
t between 


variables are diffe o 
showed significant ae in all variables a 


species. 
precipitation seasonality, with the largest differences 
in soil type (i = 1.0; P < 0.001). In contrast, node 38 
only shows significant differences in three variables: 
elevation @ = 0.34; P = 0.0037) mean annual 
temperature (i = 0.29; P = 0.0028), and minimum 
0.25; P = 
has five significant variables: 
< 0.001), 
temperature seasonality (i = 0.51; P < 0.001), annual 
ae ema (i = 0.58; P = 0.0083), precipitation of 

driest month (i = 0.51; P < 0.0001) and 
um seasonality (i = 0.51; P < 0.001). A 
similar lack of a general pattern is seen when 


temperature of the coldest month (i — 
0.0047) 


bedrock geological age (1 = 0.50; P 


allopatric or partially sympatric species pairs are 
analyzed (T able 4). 
The environmental differences between the more 
> supported nodes are presented here (cf. 
g. 2). Clades II and III are in the northern Andes, 
D and Mesoamerica, and joi node 10 


(Fig. 3). SEEVA analysis of the two ps (Table 4) 


Volume 96, Number 3 Struwe et al. 459 
Andean Speciation in Macrocarpaea 


Symbolanthus jasonii 
Tachia guianensis 


Chorisepalum psychotrioides 
-—- Macrocarpaea obtusifolia - 
iovii sect. Tabacifoliae 


. gondoloides 
. luteynii 


M. g 

M. V cotianifalla 

M. gaudialis 

M. pachyphylla 
P densiflora 


M. "Er 
G1 M. domingensis 
M. thamnoides 
— M. aiborescens 
. cochabambensis 
revoluta 


= 
o 
S 
9 
e 
< 
S 
t 
sect. Macrocarpaea 


r 
[ 
1 
Pree ee eee 


| 


M. an 
rez L— M. jact 
— M. A huentasunials 
nct M. cinchonifolia 
Ey M. normae 


Es 
š 
3 
& 
Ie] 
S 
3 
sect. Magnolifoliae 


M. luna-gentiana 


ax NE ¡— M. maguire 
TM. erm 


M. bangian A i 
M. subsessilis 
mum M. stenophylla 


rlingii 
a. arab 
~L M. kuela 


sect. Choriophylla 


r 
D 
H 
1 


SEEER 
w 
333: 
235 
=3 
ES 
5 


— M. len 
199 -M diios-vitidis 


Figure 2. One of 160 most parsimonious trees from the phylogenetic analysis based on combined molecular (ITS and 5S- 
NTS) and morphological data. Outgroups are Chorisepalum, Symbolanthus, and Tachia. This tree was use ed in ue ioe VA and 
DIVA analyses. Dotted branches collapse in the strict consensus tree, and numbers below branches indicate b 
above 50%. Current infrageneric classification for Macrocarpaea is indicated on the right. Clades a I-VI E to 
nodes discussed in the text. The northern clade includes clades 11 + ITI, and the southern clade includes M. arborescens plus 
clades V + VI 


Annals of the 
Missouri Botanical Garden 


reveals that all variables show significant differences. 
Impact indices above 0.40 are found for four 
variables: elevation, soil type, annual mean temper- 
ature, and minimum temperature of coldest month. 
Node 27 (Fig. 3) divides clades V and VI from each 
other and represents a division between a northern 
(primarily area H, clade V) and southern clade (areas 
I + J, clade VI. Seven of nine variables show 
differences in ecological trends between these two 
clades (Table 4), and the largest impact indices (= 
differences) are found for temperature seasonality (i = 
0.32). Node 


42 separates a clade of three southern. Ecuadorian 


0.37) and precipitation seasonality (i = 


species from a clade that dispersed from this area (H) 
into northern Ecuador and seuthern Colombia (E, F, 
and H; Macrocarpaea pringleana J. R. Grant and M. 
sodiroana Gilg). This major difference in distribution 
is associated with significant environmental differ- 
i (67%): soil type, 
annual mean temperature, temperature seasonality, 


ences in six of the nine variables 


annual precipitation, precipitation of the driest month, 
and precipitation seasonality. The largest differences 


are found in annual precipitation (i = 
temperature seasonality (1 = 0.37 


DISCUSSION 
CONTINENTAL SCALE PATTERNS 


Similar biogeographic patterns can often be 
detected in taxa of a similar age and geographic 
origin. For South America, there are many different 
groups of this kind, from the Gondwanic relicts with 
connections to Africa, Antarctica, and A ore than 
100 million years ago (Ma) to recent mu from the 
northern temperate zone along the Rocky Mountains 
of North America and the Andes when the Isthmus of 
Panama closed only a few million years ago. Other 
groups have evolved in situ on the South American 
continent and spread to North America, the Caribbean 
islands, and even farther away to New Zealand, 
Australia, and Africa. South America was, in effect, an 
isolated continent for more than 90 million du (from 
ca. 9 mid-Cretaceous Cenomanian, to 
Miocene; dd & Graham, 1999). During this 
time, many organismal lineages were dispersing to the 
South American continent, evolved in situ, and 
dispersed within it. 

We do not know the exact age of Macrocarpaea or 
its tribe Helieae due to the lack of fossils and secure 
phylogenetic dating. Given the position of the tribe " 
the family phylogeny and an estimated age of 42— 

50 Ma for the whole Gentianaceae (Yuan et al., 2003), 
we can assume that the Helieae, and maybe the genus 
Macrocarpaea as well, has been in South America for 


at least 30 million years. Macrocarpaea is part of one 
of the more basal divergences in the tribe (the 
Macrocarpaea subclade), with only the southeast 
Brazilian genera Prepusa Mart. and Senaea Taub. 
positioned below it in the phylogeny (Struwe et al., 
2009). This minimum age estimate (30 Ma) needs to 
be tested both in a larger molecular dating analysis of 
the whole Gentianaceae family and in a detailed 
analysis of Helieae, when fossils or other additional 
oet become available. 
e large-scale Dedi d m found i 

muda fit a scenario that is partly consistent 
with the geological "nes of South America. The 
DIVA analysis maps the disjunct areas (A, F, and H; 
Fig. 4) at the 


disjunet ancestral area for the genus including the 


ase of Macrocarpaea, supporting a 


mountains of southeastern Brazil and northern Peru, 
Ecuador, and Colombia. It is too early to tell whether 
this disjunet area distribution is due to an ancient 
dispersal, or extinction in the in-between areas. The 
sister genus (Tachia) occurs in the connecting 

mazon Basin area and is absent from southeastern 
Brazil, and therefore ics some support for a 
broad distribution of a common ancestor of Tachia and 
Macrocarpaea. An ancient dispersal event from an 
ancestor restricted to southeastern Brazil to the Andes 
r DIVA result 
strongly supports the inclusion of the northern Andean 


is not strongly supported, because ou 


areas in the ancestral area for Macrocarpaea. The 
supported scenario does not include the Bolivian 
Andes as a dispersal corridor northward from Brazil. 
The first divergence in Macrocarpaea is between 
the three oo) in the coastal Atlantic S of 
outheastern ig 
biduo (Myers El al. 2000), d al nme species 
in the genus (Andes, tepuis, Mesoamerica, and the 
Greater ee Southeastern on is part of the 
ancient wanie crust in America, together 
with the ies Shield, m des 
relatively long-term geological stability, with the lack 


areas represent 


of historical sea 
include relictual, ancestrally placed lineages of 
gentians, not only in Macrocarpaea and Helieae, but 
also in Saccifolieae E Chironieae subtribe Coutou- 
beinae (Struwe et al., ). This southeastern Brazil— 
Andean pattern, b Fk Brazilian area being more 
ancestral, has also been found in other groups that 
show a large divergence in Neotropical forests (e.g., 
Fuchsia L. [Berry et al., 2004] and Gesneriaceae tribe 
Sinningieae [Perret et al., 2003, 2007]. 

Several hypotheses nos been proposed for the 


representing long-distance dispersal from the A 
derivation from lowland white-sand areas (Kubitzki, 


Volume 96, Number 3 
2009 


Struwe 
Andean Speciation in Macrocarpaea 


1989), and in situ relictual lineages (Maguire, 1970). 
Support for several of these hypotheses is found within 
the Helieae and related tribes, with some genera (e.g., 
Potalia Aubl.) supporting the lowland to highland 
theory (Struwe et al., 2002; Frasier et al., 2008). 

The derivation of Mesoamerican species (Macro- 
carpaea auriculata Weaver rant, M. 
subcaudata Ewan, and M. valerioi Standl.) from two 
lineages in the Colombian Andes is a common pattern 
in many plant groups and is likely due to dispersal 
northward along the Isthmus of Panama after als 
closing 3.1 Ma (Burnham & Graham, 1999). D 
the American interchange, many organismal pus 
moved either north or south, and other gentians that 
show a southern derivation include Potalia (Frasier et 
al, 2008) and Tachia (Struwe, unpublished data). 
According to preliminary data from DIVA-GIS 
analysis of available ecological niches (Struwe, pers. 
comm.), there are suitable habitats for Macrocarpaea 
north of Costa Rica, but these have not evidently been 
inhabited yet. 

Three species of Macrocarpaea occur on the islands 
of the Greater Antilles: M. domingensis Urb. & Ekman 
(Hispaniola), M. pinetorum Alain (Cuba) and M. 
thamnoides (Griseb.) Gilg (Jamaica), and are included 
in this study. Their biogeographic relationships are 
still uncertain because their position as being derived 
from an Ándean ancestor has poor branch support in 
the phylogenetic tree, but they are strongly supported 
as being closer to Andean lineages than to the 
Brazilian lineage or Central American species. Long- 


location of the Caribbean clade in the phylogenetic 
result, this did not happen relatively recently. The 


the Greater Antilles did n 
seeds, so it was most likely AME inadvertently by 
birds. Similar long-distance dispersal patterns be- 
tween the Andes and Hispaniola have been found 
4). The Caribbean 


area is a complex region of several different origins, 


earlier in Fuchsia (Berry, 1982, 


and its general biogeography and geological history 
are under debate and still not fully understood 
(Iturralde-Vinent & MacPhee, 

Hispaniola are older than Jamaica, ad 
connected during the early Oligocene (Iturralde- 
Vinent & MacPhee, 1999), but this was most likely 


long before the arrival of Macrocarpaea on these 


they were 


islands. 


ANDEAN PATTERNS 


The Andean uplift started in the Miocene (ca. 
20 Ma) and continued until the Holocene, generally 


moving from the southern part of the mountain range 


E 
E 


toward the northern part, with the 
upheaval ca. 2-8 Ma (Haffer, 1987). Other scientists 
support an earlier start at 40 Ma, with major northern 
uplift at 18 Ma, as suggested by Ghosh et al. (2006) 
and Gregory-Wodzicki (2000). The Andean forests of 
the northern Andes where Macrocarpaea occurs have 
been hypothesized to be of Miocene to lower Pliocene 
in age (van der Hammen, 1979), but their range in 
elevation shifted downward during the Pleistocene 
glaciations (Haffer, 1987). The Andean clade of 
Macrocarpaea is clearly separated into two subclades, 
a northern and southern one, divided at the Amotape— 

ncabamba Zone. The i ah division of these 
two groups follows the Amotape ss, a geological 
shear zone, which divides the p (and Ecuador) 
into two parts: the younger northern Andes (formed in 
Late Pliocene to Pleistocene) and older central Andes 
(Miocene to Pliocene; Young & Reynel, 1997). 

as on our DIVA analysis, the ancestral 
distribution within the Andes includes the Amo- 


the ancestral Andean area are the Cordillera Occi- 
: Colombia and Ecuador (F) and the Cordillera 
iental of Colombia and Cordillera de Mérida in 
ne (G). The deep split between the northern 
and southern clades of Macrocarpaea represents an 
ancient divergence that strongly correlates with 
current distribution patterns. The fact that only two 
dispersal events have crossed over from the south to 
the north, and none from the north to the south, is 
in central and northern 
of Tachia and 
anthus, two other Helieae genera that have 
suitable habitats in the bou Possible 
explanations for this deep divide might be the early 


remarkable. This zone 


Ecuador is also known for its absence 
ndary area. 


division of t 


divided by sea incursions (early Miocene), mountain 
uplift and creation of valleys, and/or volcanism 
(Clapperton, 1993; Burnham & Graham, 1999), or 
perhaps a westward flow of the Amazon into the 
Pacific Ocean (Mapes et al., 2006). 

Quaternary volcanism is absent from the Amotape— 
Huancabamba Zone (area H) and the central Peruvian 
area (H and I h 
south of these regions (Clapperton, 1993). Volcanism 


I), but occurs in areas both north and 


could affect extinction rates severely, and the lack of 
voleanism might partly explain the relatively higher 
species numbers in the Amotape-Huancabamba 
Zone. In addition, this area is characterized by humid 
isolated forest islands occurring in a highly dissected 
landscape (Jørgensen, pers. comm.). The species of 
the southern clade also have wind-dispersed, winged 


Annals of the 


462 


Missouri Botanical Garden 


0000'0 Z£0 00000 TEO  SIZ00 ETO 60000 030 00000 ££'0 S1780 | S00 T0000 IcO 0OSTOO PTO 67190 | 100 Sd -— T€ 
£0S0'0 PTO 18900 600 €2000 sro  S8rz0o 600 0900'0 STO T1200 TTO 925000 STO OSEZO OTO 61000 TO Sd - ££ 
T0000 9T'0 00000 610 00000 STO 00000 ccO £€T000 ¿ro 00000  IcO 00000 STO 00000 TO T0000 STO Sd - ct 
0000°0 ZS'0  S88v0 STO 00000 2350 £0000 ZTO 00000 OOT $€S££00 050 00000 OOT 00000 TZO S0870 TEO V A TE 
0000°0 vcO 00000 970 00000 T0 00000 cc0 00000 IcO 00000 TO T0000 TO 00000  7cO 00000 €70 Sd - 0€ 
0000°0 6T'0 00000 610  c97cO 200 €6000 OTO 00000 9t0 c0IS'0 S00 66100 OTO €690 S00 00001  c00 V -— 6% 
00000 9T'0 T0000 TO S0000 TTO 00000 330 c0000 ¿TO 310000 TO 69€t00 OTO 6100 OTO 00000 6T'0 Sd = 8% 
00000 Z¿£0 00000 670 871000 Oro 00000 T0 00000 2£0 TcO0V'O S00 Treo 900 00000 FTO 22600 600 Sd -— Lg 
0000°0 STO 00000 FTO 00000 0%0 00000 TO 00000 STO 00000 €70 96620  vYOO 00000 T70 00000 070 Sd - 9G 
TS900 PEO T£000 TISO OIOOUO TSO OLETO 8t0 = 10070 PEO = E 00001 900 V A SG 
09600 670 LSTO'O 9t0 FEOO;O TIFO IITOO PEO = 09800 TEO x "E = E £600 tro V >- ud 
0000'0 TO 10000 TO 00000 FTO 00000 TO 00000 OCO 00000 070 = = 70000 ¿TO 00000 2330 V =- $6 
0000°0 00 T 00000 OOT 00000 TZO 000001 Tro O0000T STO 0000°T 9800 00000 OOT 20000'0 OOT £S0TO 60O V A GG 
T0600 OFO  606t8t0 90  0000T €70 FTO TPO 00000 PTO 96560  0v0 00S5cO 1€ 0 0000°T oro 96sc0 OO S A Ic 
00001 TO 60180 SZO  2Z86V0 GeO OSTO 870 T6000 PSO ££9T'0 ££0 ¿SI0'0 950 STIOO VO ££9T0. 880 V >- 0c 
98000'0 470 3650000 STO €0000 vFO 20000 TIFO 00001 $00  9Tr000 STO  S8t0'0 St 0 c90L00 . c£0 SLITO0 830 V >- 6I 
ST00"0 OCO 60€0°O PTO 00000 SVO c0000 ££0 ££00"0 SEO 00000 ro 00000 ETO à Sc600  ÁScO €S0000 280 V -— SI 
0000'0 470 00000 :70 00000 SFO T0000 70 06000 6T'0 00000 SEO 00000 670 200000 TEO s8690 S00 Sd -— LI 
96800 Ic0 00000 ££0 00000 PTO rero STO 00000 90 PSSc'0 STO BE = 00000 TZO £6670 9TO V A 91 
$6500 8£0 29000 cro 00000 Tro  SƏTO 90  9I600 SEO SOLTO = 9€0 = = 0000°0 3820  6rs00 ro Sd A ST 
0000°0 8£0  TISO TTO T8000 0c0 99600 030 00000 970 905070 cc 0 06000  7c0 82000 130 V -— TI 
0000'0 ££0 69600 PTO Z¿£000 TO €0000 ETO 00000 TVO 80000 70 00001 FOO LETOO sTo £0000 t$cO V = €T 
0000°0 IZO 00000 IZ'0 ZZ300"0 090 TcvcO SVO = ázat Svo 00000 S0 00000 LEO V A cT 
€000'0 SZO OSO 600 99€00 TO EZIO TTO 00000  ££'0 vr60'0 STO  8£t00'0 cro cL0c0 600 66060 900 V -— TI 
0000'0 ScO 00000 9c0 00000 2£0 00000 SFO 00000 9€0 00000 €O 00000 €70 00000 PFO) 00000 €70 Sd = OI 
0000'0 ZTO  o6rOrO TOO 00000 oT 0 $0000 000 00000 Tro 00000 STO 00000 TO 00000 GEO 00000 TTO Sd -— 6 
0000°0 2£0 00000 ct0 00000 :€0 SIO0O0 €70 E TS000 S¢O 00000 ITO v69TO 610 Sd A 8 
00000 870 00000 2£0 8£10'0 OFO 00000 670 T = 39000 TO  STIO'O 030 00000 LS 0) ZITOO €70 Sd - L 
£T00"0 800 00000 2T'0 00000 7c0 00000 TO 00000 8£0 00000 Tro 00000 STO 00000 Tro 00000 080 V -— 9. 
d D d i d a d 1 d E d 1 d 2 d A d td dS PN 
Ayrpeuoseas puou 3s9t1p uoney droid (puou jsop[oo ÁJT[RUOSBOS omperodua] (93) KSoooc) addy [os uoneAopyr 
uonepdroo1q jo uonemdroo1q yenuuy jo omyeroduray omezoduroT, ueour penuuy 


umutu] 


"uni 9q jou p[noo sisATeue əy} os soroods uo9oMjoq SoLr0gojeo [fuoumodiáuo ur 99u9J9gIrp OU SEM 9.1917 ‘sje? £iduo 104 "929Jp[oq ur (coro > 


d) sones q Aq poyeorput ore 


Diels 


Aypeonsneys pue ace O < 2 SOJBOIPUT s&roquinu 999jp[oq ur xopur du] "'suoreue[dxo 


Id £ SM317 ur eypnso: Sura o[qor (g) 


ormedui&s 10 (gq) oueduiis Aypenred (y) sanders oyu sopou [[e sorrese[o q UUM[O”) "'oouosoxd 107 X e YIM sired soroods [euruuoj € ae (unes "pezíqeue orom so[qeureA SUTN ^e osrjg 01 


puodsoxioo 4So[odoj 901) pue &roquinu poy "AuoSo[Aud oy} ur opou goto 10] sonjes q pue (2) xopur pedun Sutmoys eyep v2pndapoo420py peordonooy jo srsA[eue vA^p4s Woy synsoy "p o[qe] 


463 


Struwe et al. 


Volume 96, Number 3 


2009 


Andean Speciation in Macrocarpaea 


= = 00001 FIO 39870 6¢0  c6cDO  €€£0 F m T6900 1£0 00000 6€0 ITTOO PSO ozro S60 V A T9 
= B 000001 ITO 26080 SGO T9790 60 00000 ITO 0000T sT0 38€0000 9£0 2O96t00 OSTO  I/cCO PEO V >- 09 
0000'0 IZO  8£STO  6t0 00000 ZEO 69200 680 00000 370 8EsTO 970 00001 600 = E 00001 STO V A 68 
0000°0 670 99700 PTO 6€9c00 06c0 00000 SOO OTIOO ct£0 S270 PLO  c00000 OFO S9S000 SVO $0cc0 [60 Sd -— 8S 
0000°0 £S5'0 0000'0 6v0 €IIOO TEO S8TIOO 0 8£000 VO 39050 STO 40s90 STO = 2 s9070  9DT0 Sd -— LG 
TVcO'0 120 00000 ct0 00000 830 00000 330 00000 8<0 c£200  0%0  0000'0 OFO 00001 ITO 00000 610 Sd -~ 9S 
Scyo9'0O OTO 63180 OTO 92990 PTO  S9SEO STO $9500 970  LSTOO 060 T0790 PTO ss9T0 SO 96090 PTO Sd - SG 
OEPTO PTO 769080 STO  cScTO 6f0  ÀSS6cO0 910  ÁS965cO0 TO 39880 VIO  IS00'0 670 servo PTO IS60'0 6T0 Sd -— vs 
00000 IZO 00000 IZ0 00000 8S0 00000 IZ0 00000 TZO 00000 TZO 5 m = = 00000 IZO V A €S 
00001 FEO 00000 ZEO LEEOO EFO 00000 380 00001 PO 000071 360 B = 00000 OOT 2S€0'0 ETO Sd -— GS 
TITO 2470 00000 TO  TIITO 80  Á00001 TO TIMTO 7vVO 00000 8S0 D = 0000'T 610 00000 830 V -— IS 
SI00"0 Z270 90000 7vVO  €8IT0 €00  TPOTO TEO Z0000 ZTO TII00O 970 00000 OOT 00000 1790 00000 SFO V -— 0S 
63v0'0 E0 96700 SEO 85770 STO 100P0 €70 OITOO EFO 0000T ITO = E T5200  ££0 00001 TTO Sd -— 6v 
00000 020 €000°0 OC'O TcOOO 7t0 00000 TO 00000 970 00000 990 00000 ¿90 8S000 680 10000 STO Sd A 8v 
0000°0 650 00000 120 00000 8S0 00000 TZO 96690 STO 00000 850 = » €6000 OSTO 20000 270 V A LY 
0000'0 PrO 00000 OFO 00000 7t0 00000 OVO ¿FOSO Tro 00000 SEO 80000 ££0 €8cVO ¿IO 81000 870 Sd -— 9v 
0000°0 ISCO 000000 TISO €800°0 80 00000 3320 00000 ISO £s8 10 co 00000 0S'0 00001 ca) 00080 880 S A SP 
0000°0 €70 $SS000 69'0 TOTOO OSO TS000 ZTO SS000 690 10000 850 cSIOO 290  I70  8TO 03000 950 S A vv 
00000 ISCO 00000 OFO 00000 €F0O 62000 ITO 00000 2<'0 vroviO EU LS100 [60 00000 9€0 TISE£0'0 FeO V A er 
8100°0 Sco 01000 970 c0000 SEO  0I670 200 00000 2£0 99000  cc0 09170 STO 00000  ££'0 £r9000 T0 Sd -— cv 
0000°0 vTcO €T1000 Tro 00000 620 T0000 [20 00000 c£0 00000 870  66c00 STO 00000 ZEO 8£000 28T0 Sd - Iv 
0000'0 00 T 00000 IZO0 00000 TZO 00000 T270 m E 0000'0 120 T p 299070 OSO 00000 IZO V A OF 
0000°0 8€'0 00000 €90 £t00TO0 TEO 98000 980 ST000 90 £9080 T0 00000 220 00000 680 28680 £€c0 V A 6€ 
00001 SOO  S1800 c0  It£tc0 OTO ZvvTOO ScO  2S0g0 T0 86000 670 S6070 ToO X 00001 OUO L€00°0 teo S A 98€ 
95600 8v0 382c00 S9S'0 00000 ETO 00000 TZO 00000 TZO 00000 1Z0 00000 990 00000 OOT 00000 89'0 S A LE 
0000°0 ITO 00000 PSO I6IOO ScO 00000 970 00000 OTO 00000  tVO  c0000 c£0 00000 680 208100 SO Sd - 9€ 
0000°0 cro 00000 ceo 00000 380 00000 E0 00000 ££0 00000 380 060000 IZ0 00000 FEO) 00000 Oro S -— SE 
d a d 1 d j d 1 d 2 d : d : d d d ? (1 d$ “PON 
Ayr [euoseos (puou IS9LIP uonejidioo1d (puou jsop[oo Ay[euoseos emexoduioj (oSv) ASojoos addy [ros uoneAo[^T 
uonepndioorq jo uonendworg yenuuy jo omyeroduray omeroduro, ueeur [enuuy 


umutu] 


"penunuo ^ epqep, 


Annals of the 
Missouri Botanical Garden 


Symbolanthus jasonii 


Chorisepalum psychotrioides 
crocarpa Er Ln SE Brazil 
8 M. gla zio 
M. rubra 


M. papillosa G, W Venezuela, Merida 
. & E Panama, Cor. Occi & Central 


ubcaudata D, pm be & W Panama 


M. glabra G, E Calamb bia, Cor. tal 
aeB,S Ven estela, a de la Neblina 


3 M. neblin 
a , N Ecuador, Cor. Occi 
M. nicotianifolia G, C Colombia, Cor. Orient 
is E, N Ecuador olom a Cor el 


M. gaudialis 
7 M. pachyphylla E, W Colombia, Cor. 
aur Ud E. W Colombia, Cor. Cent al 


M. pinetorum C,E 
35, M. domin ENA ja ma 


a 
habambensis J W Bolivia 
M revoluta |, C P 
M. weigehdtorum |; C n 
M. ostentans |, C P 
robin- fasten |, C Peru 


u 
M. cinchonifolia J, „W Bolivia 
ie om mae J,SP 


M. viscosa |, C P 
M. ne dentara H, S Ecuador 


eru 
M. pachystyla | C Peru 
M. Erud W Boliv 

M. subsessilis H. $ Ecuador 


M. stenophylla H, NW Peru 
M. chthonotropa H, NW Peru 


xe 
. opulenta H, S Ecuador 
ericii El 


O sympatric M. 
O partially sympatric M. pajonalis |, C P 
delicias M. noctiluca H, S Ecuador & bis Peru 
NOR ANS z7 M. apparata H, S Ecua 
. elix H, S Ecuador 
38 M. bubops H, S Ecuador & NW Peru 


M. harlingii H, S Ecuador & NW Peru 
. innarrabilis H, S Ecuador & NW Peru 


. pringleana E S Col., Te aut 
p od F, N Ec. -&S Col., Cor. Occi. 
nsii E S Ecuador 
T i lenae H, S Ecuedó 
M. des idis H,S Ecuador 


selected most parsimonious tree used in DIVA and SEEVA analyses, with numbered nodes and geographic 
(see 


Figure 3. The 
dido listed for each species. ee area coding for DIVA is indicated with letters A-K after species names 
text for definition of areas). Symbols (O, LJ, 5x) on nodes indicate sympatric, partially sympatric, or allopatric clades, based 
on exact species distribution within a (not DIVA area classification), which is further analyzed in the 


SEEVA analysis. 


Struwe 


Volume 96, Number 3 
Andean Speciation in Macrocarpaea 


2009 


Symbolanthus j jasonii H 


Chorisepalum psychotrioides B 

(OM Brazilian Macrocarpaea clade (3 spp.) A 

Macrocarpaea glabra G 

M. neblinae B 

M. gattaca F 

M. gaudialis clade (3 spp.) E 
M. nicotianifolia G 


BH 


AFH 


M. valerioi D 


M. macrophylla EF 

Caribbean Macrocarpaea clade (3 spp.) C 

== M. arborescens H 

M. cochabambensis J 

M. revoluta | 

M. weigendiorum | 

M. ostentans | 

M. robin-fosteri | 

M. jactans H 

M. tahuantinsuyuana | 
M. cinchonifolia J 
á J M. normae J 


M. viscosa | 
M. luna-gentiana H 


M. pachystyla | 


M. subsessilis+M. stenophylla H 
M. jensii group (7 spp.) H 
M. pringleana EH 


B Pantepui of the Guayana Shield ; 
C ik Antilles of the Sal H E 
D Mesoamerica (Costa Rica & Pan M. chthonotropa*opulenta H 
E Cord. Cental, Colombia & Cord. Oriental in Ecuador M. noctiluca clade 
F Cordillera Occidental in Colombia & Ecuador, & Chocó H 5 spp.) H 
G Cordillera Oriental in Colombia and Mérida in oe 
H Pei -Huancabamba zone in S Ecuador & N Per 
| Andes eru 
J Arden | in Bolivia & S Per 
Amazon Basin of Brazil, Colomba: Peru & Venezuela 
Figure 4. Results of the DIVA analysis when mapped onto the selected most parsimonious phylogenetic tree. Rs 
optimized onto each branch are marked with letters (see legend and Fig. 3 for coding); when se D optimizations were equally 
mal ur re constructions), 


optimal, one was select ed Based ye Ley = im gical CL UA bs Appendi 1 for all o 
f n E 1 


Circles (@) on bran P 
ia : ra: a ition. Thesch nas: to the left sh an of area 


: 
new areas, solid lines indicating vicariance events, and arrows 
paea. 


and X by 
relati jen tw with bold jr (D, G) indicating expansion to 
icu dispersal events. Vicariances and dispersals are only marked for Macrocar; 


Annals of the 
Missouri Botanical Garden 


seeds that presumably could lead to increased 
thereby increased 
hopping. 
of allopatric speciation are found in 


colonization of new areas and 
allopatric oe through islan Howev- 
er, more instance 
the northern clade than in the southern clade. 
Weigend (2002, e showed that the Huancabamba 
pression is no ispersal barrier to ean 
species of middle ds and this enum is 
upheld by our data. This finding is in contrast to data 
from moy e dr Duméril & Bibron bid 
(Youn ynel, 1997; Duellman 
1999); e it should be noted that this "e 
group represents an immense Neotropical radiation 


with a very different speciation history. 


SPECIATION PATTERNS 


Within Andean Macrocarpaea, allopatric speciation 
is much more common than sympatric speciation, and 

i E own p E 
ees Brazil e et al, n some 
cases, allopatric speciation is aen to es 
dispersal (e.g., M. neblinae on Sierra de Neblina in 
and M. valerioi in Costa Rica), 
but often allopatric sister taxa occur in more adjacent 


southern Venezuela, 


parts of the same mountain range (e.g., M. cinchoni- 
T (Gilg) Weaver in western Bolivia and M. n 

R. Grant in southern Peru) Suc obs so ben 
hi. are common and indicate Be dispersals 
along mountain ridges or historical splitting of larger 
ancestral populations into separate species after 
isolation. Both scenarios support the theory that 
mountain ranges such as the Andes represent virtual 


2002) In 


Macrocarpaea, most species occur on the slopes and 


islands (Young et al, the case of 
not on the summits, but slopes may serve as isolated 
units as well. Isolated páramos on Andean summits 
have long been considered analogs to island chains 
such as Hawaii and the Galápagos (Young et al, 
2002), but our data also support this for lower 
elevations. Many of the paramo plants are relatively 
recent immigrants from plant groups from the northern 
temperate zones, whereas the forested slopes include 
primarily Neotropical floristic elements that reach 
. Within Helieae, 
the highest species diversity is found in the Andes, 


high species diversity in the Andes 
but is largely represented by only two genera, 
Macrocarpaea and Symbolanthus (Molina & Struwe, 
2008). The Amazon lowlands and the Brazilian and 
Guayana Shields have much fewer species, but they 
represent the ancestral evolutionary lineages (Struwe 
et al., 2002, 2009). 

The seeds of Macrocarpaea can be divided into two 
major types: very small and winged or larger, angular, 
and heavier. Winged seeds, which promote dispersal 


over larger distances, are found in clades V and VI, 
which show repeated dispersal patterns north-south 
within the southern clade. Such seeds are also found 

the northern group at node 12 (Fig. 3), M. 
a J. R. Grant and M. luteynii J. R. Grant 


& Struwe, as an independently derived character trait. 


occasions of di 
sampled. Because it was not possible to include all 
species in our study due to unavailability of material, 
species from the southern area are overrepresented. 


pers. 
Pollinator segregation therefore does not support 
sympatric speciation in Macrocarpaea, and it is more 
likely that ecological niche divergence or population 
isolation and subsequent fixation of different traits in 
smaller populations have led to different species. 


ECOLOGICAL NICHE PATTERNS 


Results from the SEEVA analysis of sister species 
show that all species pairs have individual divergence 
patterns and environmental differences. Divergence in 
ecological niches is common both in allopatric and 
sympatric species in Macrocarpaea, and generally 
these divergences are not significantly different in 
size between the two speciation types, but these 
variables differ between species pairs. For example, 
divergences were found based on Rx altitudinal 
zones (M. apparata and M. elix; node 37, Fig. 3) and 
on different types of bedrock and in zones with 
different climate seasonality (M. dies-viridis and M. 
lenae; node 45, Fig. 3). Perret et al. (2007) also found 
a lack of increased divergence in sympatric species 
when compared with allopatric species. 

The species differences found with SEEVA show 
only patterns, not processes, and variables should not 
be seen as the probable cause for speciation, unless 
further studies can show adaptation to specific 
environments or changes linked to paleoclimatological 
or geological events. We know that species have 
moved around on a geographic scale, especially 
during the Pleistocene, but their ecological niches 
might have been more stable due to niche conserva- 
tism. Using a different approach, Peterson et al. 
(1999) showed that speciation through geographic 
separation often appears before ecological niche 
separation in vertebrates and butterflies in Mexico, 
and our results indicate the same. Extracting 
environmental data from current locations therefore 
most likely represents historical ecological niches, if 
not the historical location of the population. 


Volume 96, Number 3 
2009 


Struwe 
Andean Speciation in Macrocarpaea 


In future studies, it would be interesting to compare 
species pairs for particular areas from many different 
genera to determine whether ancestral populations 
reacted similarly to climatological and geological 
events in the past, and whether recurrent ecological 
niche divergence appeared in unrelated lineages. This 
would be possible, for example, in the Amotape— 
Huancabamba Zone, where we now have at least four 
different data sets from angiosperms (including the 
studies by Weigend, 2002, 2004). 

Our study shows t 


explanations of speciation 


events need to be sought disc for each species 
pair, and that generalities are not necessarily 
applicable across a larger species group distributed 
. If we 


patterns, the ecological niches of species, and threats 


over a large area want to understand speciation 
and means to the conservation of these species, much 
more pu need to be collected and analyzed (Young 
et al., -to- 


02). One major difficulty is the lack of u 
hylo, and georeferenced 


date revisionary, genetic, 
data for most plant genera. Species-level phylogenies 
only represent a small percentage of Neotropical 
biodiversity, and there is a dire need for more 
taxonomie work that can be integrated with biogeog- 
raphy, ecology, and conservation. 


Literature Cited 
Berry, P. E. 1982. The systematics and evolution of Fuchsia 


section Fuchsia (Onagraceae). Ann. Missouri Bot. Gard. 
9: Rud 


W. J. Hahn, K. J. Sytsma, J. C. Hall & A. Mast. 
— 2004. Phylogenetic relationships and biogeography of 


uchsia a based on noncoding nuclear and 
chloroplast DNA dat J. Bot. 4. 
Burnham, R Graham. 1999. T tory of 


ws C. 1993. Quaternary Cello and Geomorphol- 
of South America. Else d: 
Db , W. E. & J. B. Pramuk. 1999. Frogs of the genus 
Eleutherodaciylus (Anura: Leptodactylidae) i in the Andes of 
ci. Pap. Nat. Hist. Mus. Univ. Kansas 13: 1-78. 
Albert & L. Struwe. 2008. Amazonian Sic 
and areas as ancestral regions for Sou th American 
biodiversity: Pewa: and phyloge 


evier, Amsterdam. 


northern Peru. Sci 


., C. N. Garzione Rapid uplift of 
the ‘Altiplano revealed through C-O . nds in paleosol 
nates. Science 311: 511-515. 

Goloboft P. 1999. NONA (NO or, ver. 2. Published by 
the PIS Au ue Argen 

Gould pun E 2004 ‘Phylogeny and os of 
id ae-Helieae) 
in the Guayana Fini and andes. based on ribosomal 
ay a sequences. Missouri Bot. Gar 


Grant, J. R. . De Macrocarpaeae Grisebach (ex 
a speciebus novis II: Typification of the Ruiz 
von names. Harvard Pap. Bot. 7(2): 423-436 
— "2004. De Macrocarpaeae Grisebach (ex Gentiana- 
ceis) speciebus novis V: Twenty-three new species largely 
rom Peru, and Caan r all species in the genu: 
aig ee Bot. 9(1): 1 
ROME Grisebach (ex Gentiana- 
d mo 


species, largely from Colombia. Harvard Pap. Bot. 9(2): 
305—342. 


—. 2007. De Macrocarpaeae Grisebach (ex Gentiana- 
ceis) speciebus novis VII: Four new species and two 
natural Moe Harvard Pap. Bot. 11(2): 129-139. 
008. Macrocarpaeae Grisebach (ex Gentiana- 
m"T— spe sii novis VIII: p new species from Ecuador. 
Harvard Pap. Bot. 13(2): 2 
——— we. 2001. ds m Grisebach (ex 
e ins ins novis I: An introduction to the 
enus new species from Colom- 


t. 5: 489— 


—— De 


acrocarpaea and three 
bia, Ecuador, and Guyana. Harvard Pap. Bo 
498. 


——— —— e Macrocarpaeae Grisebach (ex 
Gentianaceis) eden novis III: Six 
Ecuador 


Parque pas r. Harvard 


Pap. Bot. d 
& 


x» Podocarpus, 


pom Jr. 2003. De Macrocarpaeae 
Grisebach = Centianaceis) speciebus novis IV: Eleven 


o 
A 
"a 


presence of stipules in the family. Harvard Pap. Bot. 8(1): 
83-109. 


. M. Maas & L. Struwe. 2006. Yanomamua araca 


a Region in Amazonas, 
Brazil. Harvard Pap. Bot. 11(1): 29-37 
Gregory-Wodzicki, K. M. 2000. Uplift history of the Central 
and Northern Andes: A review. Bull. Geol. Soc. Am 
112(7): 1091-1105. 
m J. 1987. Quaternary history of Ao as Americ 
. 1-18 in T. C. C. Whitmore € G. T. Prance ar 
B and TT History in Tropical Amer- 
ica. Clarendon Press, Oxford 
. SEEVA. Scltware for Spatial yr 
and Ecological Vicariance Analysis, ver. Lun 
University, n P <http://seeva. mum se>, 
accessed 1 November 2 008. 

—— & L. Struwe. 2008. SEEVA Manual. Rutgers 
University, Newark, New Jersey. <http://seeva.heiberg. 
sec and <hitp://www.rci.rutgers.edu/~struwe/seeva>, 
accessed 21 Da 008. 

Hijmans, R. J., S. E. Cameron, J. L. Parra, P. G. Jon 
Jarvis. 2005. po high RU. edad ke 
surfaces for global land areas. Int. J. Climatol. 25: 
1965-1978. 
Iturralde-Vinent. 


. MacPhee. 1999. 
Paleogeogra hy of the Caribbean 3 region: Implications for 
Cenozoic biogeography. Bull. Amer. Mus. Nat. Hist. 238: 
1-95. 

i sss P. M., C. Ulloa Ulloa & C. Maldonado. 2007. 

ueza de Uem vasculares. Pp. 37-50 in R. M 

L. P. Kvist, F. Borchsenius & H 


Centrales. Universidad Mayor de San Andrés, La Paz 


Annals of the 
Missouri Botanical Garden 


Kubitzki, K. 1989. The ecogeographical ua p » 
Ev 


nian inundation forests. Pl. Syst. ol. 
285-304. 
Maguire, B. 1970. On the flora of the Guayana Highland. 


Biotropica 2: 85-100 

Mapes, m R. Nogueira, D. S. Coleman & A. M. 
Leguizamon Vega. 2006. Evidence for a continent scale 
drainage inversion in the A n basin since the Late 
Cretaceous. Geological Society “of America Abstracts with 
a 38(7): 51 13, 

Mayr, E. 1963. Animal Species and Evolution. Belknap 
Press, Cambridge 

& L. Stru 2008. Revision of ring-gentians 

(Sy mbolanthus, , Gentianaceae) from Bolivia, Ecuador, and 

Peru, with a assessment of conservation status. pn 
Biodivers. 6: 477 k 

Morawetz, W. & C. diae 2007. Angiosperm od 
endemism and conservation in the Neotropics. Taxon 56: 
1245-1254. 

. A. Mittermeier, C. G. Mittermeier, G. A 
Fonseca & J. Kent. 2000. Biodiversity o E" 
conservation RA a 403: 853-858. 

Nixon, K. C. 1999— inClada, vers. 1.0. Published by 
the author, A 

Mens 1993. On outgroups. Cladistics 9: 


Chautems, E Rite 


R. „Spi chiger, 


analyses of six plastid DNA regions and nuclear ncpGS. 
Amer. J. Bot. 90: 445—460. 

; . ———, T. G. Barraclough € V. 
Savolainen. 2007. The zeoğraphical pattern of speciation 
and floral diversification in the Neotropics: The tribe 

Sinningieae (Gesneriaceae) as a case study. Evolution 61: 
E 
Peterson, E. T., J. Soberón & V. Sánchez-Cordero. 1999. 
ae of ecological niches in evolutionary time. 
67. 


. Se-Al: Sequence Alignment Editor, vers. 
2.0all. «http: /tree.bio.ed.ac.uk/software/seal/ >, accessed 
1 November 2008. 


Rodriguez- -Mahecha, J. V., gr P. we n, T 


Fonseca (editors), Hotspots Revisited: 
Biologically Richest Most Endangered 
dal 2nd ed. CEMEX, Mexico a 
a yE o, A. Telesca, L. 
, F. Arjona, F. ,R. d Smith & V. H. 
inchs 2004b. Tne Cc Magdalena Pp. 80-84 
n R. A. Mittermeier, P. Robles-Cil, M. Hoffmann, J. D. 
Pilgrim, T. M. Brooks, C. G. Mittermeier € G. Fonseca 
T dem dou Earth's ER Richest 
and Endangered Ecoregions, 2nd ed. CE 


Mex rus y- 
e F. 1996. DIVA, vers. 1.1. Published by the author, 
ppsala 


. 1997. Dispersal-vicariance analys 
proach to the quantification of historical n s 
y: 03. 
Suse; L. 2008. SEEVA: Spatial Evolutionary and Ecolog- 
ical M s Analysis. Rutgers b e New Bruns- 
ew Jersey. <http://www. .edu/~struwe/ 
es accessed 21 Ronde 20 08. 


wick, 


——— D Kadereit, J. Klackenberg, S. Nilsson, M. Thiv. 
B. von Hagen & V. A. Albert. 2002. Systematics, iden 
ean, and biogeography of Gentianaceae, including a 

tribal and subtribal classification. Pp. 21-309 m L 
Rotes & V. A. Albert (editors), Gentianaceae: Systemat- 


lcs and Natural History. Cambridge University Press, 


t, M. F. Calió, C. Frasier, K. B. Lepis, K. 
G. Mathe ews DR. E Grant. 2009. Evolutionary patterns in 
neotropical tribe Helieae (Gentianaceae): Evidence from 
morphology, chloroplast and nuclear DNA sequences. 
Taxon 58: 479—499. 

Swofford, D. L. 2000. PAUP.* Phylogenetic Analysis Using 

s. 4. Sinauer 


ultiple sequence alignment aided 

by quality analysis tools. Nucl. Acids Res. 25: 4876—4882. 

van der Hammen, T. 1979. History of flora, vegetation and 

climate in the ig Cordillera Oriental during the 

last five million —32 in K. Larsen & L. B 

Holm-Nielsen rod "Tropical Botany. Academic Press, 
ndon 


Lo 
Weigend, M. 2002. Observations on the LE Es the 
Amotape-Huancabamba zone in northern Peru. Bot 
68: 38-54. 
004. Additional observations on the biogeography 
of the Amotape-Huancabamba zone in northern Peru: 
Defining the south-eastern limits. Revista Peru Biol. 11: 
127-134 


E ., C. Ulloa d e L. s & S. Knapp. 2002. 
Plant evolution and e n Andean South America: 
An inks Bot. ee 68: DE 2]. 

Young, K. R. & C. Reynel. 1997. Huancabamba Rs 

S. D. Davis 


Hamilton (editors), Centres of Plant Diversity: A Guide 
and Strategy for Their Conservation, Vol. 3. 
IUCN Publications Unit, Cambridge. 


The Americas. 


ve M DNA sequences. NEM Phylogen. Evol. 28: 


APPENDIX 1 


Optimal due from the DIVA analysis listed for 


each node (an r of terminals [AOT]; node numbers in 
Fig. 3 pi ps to legend for Fig. 4. Note that 
optimizations in Figure 4 sometimes reflect a dispersal event 


here, and due 

areas (see Materials and Methods), some node numbers are 

not listed here. 

Node 6, AOT M. obtusifolia-M. normae: AB AC ABC AD 
ABD ACD ABE AF ABF ACF ADF AG ABG ACG ADG 
AFG AH ABH ACH ADH AFH AGH 

Es 9, e M. glabr ae: CD CF CDF CG CDG 

H CDH FH CFH DFH GH CGH DGH FGH 

m i AOT M. glabra-M. macrophylla: Y G DG FG DFG 


Volume 96, Number 3 Struwe 
2009 Andean Speciation in Macrocarpaea 


Node 11, AOT M. gondoloides-M. macrophylla: F DF FG Node 33, AOT M. chthonotropa-M. pajonalis: H 
DFG E Node 34, AOT M. noctiluca -M. Pa Mer H 
Node 13, AOT M. valerioi-M. macrophylla: D DF DEF DG Node 39, AOT M. ericii-M. pajonalis: HI 


EG DEG FG DFG EFG Node 43, AOT M. pringeana-M cone FH EFH 

Node 14, AOT M. auriculata-M. Top DE DF DEF Node 44, AOT M. jensii-M. sodiro 

Node 16, AOT M. valerioi-M. papillosa: DG Node 49, AOT M. viscosa—M. udis HI 

Node 17, AOT M. glabra-M. nicorianifolia: G FG BFG EFG Node 50, AOT M. e et A pachystyla: H 

M 18, AOT M. neblinae-M. nicotianifolia: EF BEF BG Node 51, AOT M. zophoflora-M. pachystyla: HJ 
BEG FG BFG EFG Node 52, AOT M. ps riistana—M. pachystyla: J 

Node 19, AOT M. gaudialis-M. nicotianifolia: EG Node 53, AOT M. maguirei-M. pac HMM IJ 

Node 22, AOT M. neblinae-M. gattaca: BF Node 54, AOT M. Se normae: IJ 

Node 23, AOT M. pinetorum-M. normae: CH Node 55, AOT M. revolut ae: I 

Node 26, AOT M. sn qud normae: H Node 56, AOT M. hone mM. normae: 1 

Node 27, AOT M. viscosa-M. normae: HI HJ HIJ Node 57, AOT M. ostentans-M. normae: 1 IJ 

Node 28, AOT M. Prae pajonalis: H Node 58, AOT M. robin-fosteri-M. normae: I 

Node 29, AOT M. bangiana-M. pajonalis: HJ Node 59, AOT M. robin-fosteri-M. jactans: HI 

Node 30, AOT M. subsessilis— M. aa H Node 60, AOT M. tahuantinsuyuana-M. normae: IJ 

Node 32, AOT M. jensii-M. pajonalis: Node 61, AOT M. cinchonifolia-M. normae: J 


DETERMINANTS AND 
PREDICTION OF BROAD-SCALE 
PLANT RICHNESS ACROSS THE 
WESTERN NEOTROPICS! 


Trisha Distler,? Peter M. Jørgensen,” 
Alan Graham,* Gerrit Davidse,? 
and Iván Jiménez? 


ABSTRACT 


Patterns of broad- oe plant Po richness are Us to be largely determined by (1) variation in energy and water 


availabilit 


nergy hypo 
sis). However, lack o c and distribut 
western Neotropics 


s al the ou nico Md hypothesi 
, the regional effects hypothesis. 


thesis), (2) habitat and topographic heterogeneity within sa 
nc s ud hypothesis), and s ) ia pla in geographi 
n data, particularly m 


and regression models a 
ach of the above three hypotheses. Variati ion in plant ric 
l 


mpling units 


merica, and northern Central America) was 
ibutions from the species energy hypothesis 
incorporated the relative contributions of all 


Al contr 


models that 


Chiapas, Cordillera de Tilarán, iuris a de Talam Cr ME s Cordillera Central, the Andes, and the Venezuelan 
Guayana); relatively low SS ental Mexico an n Yuc nS Llanos al Ven ez melee and in the Gran Chaco region of 
Bolivia low Amazonia i Central America, the Andes, 


;an 
We discussed the contrast between our r 


eds evio 


ayana. esu 
richness to be primarily determined by the species energy hypothesis and predicted dierent patterns of plant richness across 


the western Neotropics. 
Key words: 
plants. 


Broad-scale species richness, Neotropics, regional effects, 


spatial heterogeneity, species energy, vascular 


Patterns of spatial variation in the number of 
species co-occurring within broad geographic areas 
were discussed by naturalists of the 18th and 19th 


centuries (von Humboldt, 1808; Darwin, 1862; 
Wallace, 1878; see Hawkins, 2001), and more 
rigorous attempts to quantify these patterns began 


during the last century (e.g., lff, 1935; Simpson, 
1964; see Mutke & Barthlott, 2005) along with the 
formulation of numerous hypotheses to explain them 
(Hutchinson, 1959; Pianka, 1966; Huston, 1979, 

; Rohde, ; Palmer, 5 d 
1995; Schemske, 2002; 


understanding what denies ue nn a 


illig et al., 


a of species richness still remains a central 
ssue in ecology and biogeography (Gaston, 2000; 
Rik 2004; ‘Pennisi, 2005). Recent work in this 


a has focused on a few prominent hypotheses, 


notably the species energy (SE), spatial heterogeneity 
SH), and regional effects (RE) hypotheses. 
The SE, broadly defined, holds that gradients of 


energy and water availability across sampling units 


= 


create and maintain species-richness patterns (Hutch- 
inson, 1959; Pianka, 1966; Brown, 1981; O’Brie 

; Rosenzweig, 1995) 
nonexclusive mechanisms (Hawkins et al., 2003; 
Currie et al., 2004; Evans et al., 2005; Clarke & 
Gaston, 2006) including the effects of these gradients 
on extinction (Wrig oe evolutionary rates 
Rohde, 1992; Eva 2005), and 
filtering according to eA requirements (von 
Humboldt, 1808; Turner et al., 1987). The SH posits 


that variation in elevation and habitat within sampling 


— 


species 


units increases richness, again through various 
nonexclusive mechanisms, including species sorting 


! We are grateful to B. Magill for major assistance handling Tropicos data. R. Field, J. Mutke, and one anonymous reviewer 


* Missouri Botanical Garden, P. a Box 299, St. Louis, Missouri 63166-0299, U.S.A. a rad org. 
8- 


*Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A 


graham@mobot t.or; 


* Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S. A. ^i 1 
ê Center for Conservation and Sustainable Development, Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 


63166-0299 
doi: 10.3417/2008034 


, U.S.A. Author for correspondence: ivan.jimenez(?mobot.org 


ANN. Missouni Bor. Garp. 96: 470—491. PUBLISHED ON 28 SEPTEMBER 2009. 


Volume 96, Number 3 
2009 


Distler et al. 
Plant Richness across the Western Neotropics 


1995) or isolated 
patches of similar habitats (Simpson, 1964) and the 


among habitats (Rosenzweig, 
effect of spatial heterogeneity on diversification rates 
(Simpson, 1964; Jetz et al., 2004). The RE proposes 
that differences in richness between sampling units 
with similar environments but located in different 
regions (i.e., diversity anomalies) result from regional 
disparities in history and geographic configuration 
that, in turn, cause differences in the occurrence of 
particular lineages, time available for diversification, 
and rates of species production and extincti 
(Pianka, 1966; Latham & Ricklefs, 1993; Schluter & 
Ricklefs, 1993; Qian & Ricklefs, 2000). 

Many ecologists seem to agree that some combina- 
tion of the three hypotheses above provides the best, 
currently available, explanation of broad-scale spatial 
patterns of species richness (Rahbek & Graves, 2001; 
eh et al., 2004; Field et al., 2005; Kreft & Jetz, 

However, progress in establicking the relative 
Sa of these hypotheses as explanations of 
broad-scale Aik of plant 2 richness has 
been hinder the scarcity of tax m and 
spatially md data sets (Whittaker et al., 2005), 
particularly for tropical regions (Frodin, 2001; n see 
Kier et al., 2005). Thus, most studies are focused. ona 
few relatively well- kno own P forms (e d , Currie & 
Paquin, 1987; O'B Fie old eta | 2005) o 
taxonomic groups (e 2, eu et al. prs Other 
studies are based on floras, checklists; and other 
literature sources that have little information 
rt (e.g., Kreft & Jetz, 2007). Yet, failure 


to account je sampling effort can alter our perception 


sampling 


of spatial patterns of plant richness (Nelson et al., 
90; Parnell et al., 
previous efforts to overcome these problems (Jiménez 


3). Here, we expanded on 


et al., 2009) by assessing the relative importance of 
the SE, SH, and 


patterns of vascular plant species richness across the 


RE in determining broad-scale 


western Neotropics. In addition, we used results on 
the relative importance of different hypotheses to 
predict relative plant species richness in areas of the 


western Neotropics with poorly sampled floras 


METHODS 


To describe the spatial pattern of vascular plant 


species richness across the western Neotropics, we 
used 755,401 georeferenced herbarium specimen 
records from the Tropicos database (<http://www. 
uri Botanical Garden 


plant species collec 


tropicos.org/>, Misso 
senting 48,264 


Mexico, Belize, 


, repre- 
ted in central 
Guatemala, Honduras, Nicaragua, 
Costa Rica, Panama, French Guiana, Guyana, Vene- 
zuela, Colombia, Ecuador, Peru, Bolivia, Uruguay, 
and northern Argentina. We mapped these data on a 


Behrmann cylindrical equal-area projection of the 
study area and used rarefaction (Gotelli & Colwell, 
001) to estimate relative species richness in 10 
100 km sampling units as the expected number of 
species in a random sample of size n from a set of N 
herbarium specimen records. To choose the value of 
n at which relative richness was measured using 
rarefaction, we balanced the benefits of increasing n 
in terms of increased precision of relative richness 
estimates agains osts in a of 


st the correspondin 
decreased sample size (Fig. 1; Ji 


g co 
iménez et al., ). 
We judged n — 500 herbarium specimen iie to be 
a reasonable compromise and retained 255 sampling 
units with at least 500 specimen records for further 
analysis (Fig. 2 

Admittedly, ect is unlikely to account for all 

rror in estimates of relative richness due to 
differences ; in sampling effort, because the set of N 
herbarium specimens collected in a single sampling 
unit is unlikely to be a random sample from the 
individuals of all vascular plants occurring in that 
sampling unit (Jiménez et al., 2009). Therefore, our 
estimates of relative plant richness contain potentially 
large measurement errors that, nonetheless, can be 
reasonably subsumed in the error term of statistical 
models oe different hypotheses. Such mea- 
surement errors may affect the conclusions from our 
analysis nae in proportion to its correlation with 
the explanatory variables relevant p i three 


hypotheses of interest (Jiménez et al., . Asa 
starting point, we assumed such RA was 
negligible. 


We used a set of regression models to simulta- 
neously measure the extent to which the estimated 
pattern of relative plant richness across the western 
Neotropies (Fig. 2A) supported predictions from the 
SE, SH, and RE. For each regression model, we 
derived predictions from one or more hypotheses 
about the sign of regression coefficients relating 
explanatory variables to relative plant richness. Some 
regression models included higher-order terms (i.e., 
interaction or quadratic terms) that may account for 
small portions of the variation in the response variable 
when the range of the relevant explanatory variable is 
limited. Therefore, when the coefficients of higher- 
order terms were not statistically significant, we 
examined the performance of reduced models with 
no higher-order terms in an attempt to trade-off a 
tolerable bias for increased precision (Chatterjee & 

adi, 1988). In these reduced models, the predicted 
sign for regression coefficients may depend on the 
range and central tendency of explanatory variables. 
We derived predictions for the sign of regression 
coefficients in reduced models based on the range of 
explanatory variables in our sample (Fig. 3) and the 


472 Annals of the 
Missouri Botanical Garden 


A B 
e 
o. o» g- [9] 
— » v c O 
z a S o0 
* oo 
"T zx | 9 
E O- 3 
o o Oo 
E 3 21 
S $3 E: " 
= o e 
$94 3 3 8- ® 
a = 3 Q 
8. 
8 - 5^ 
o 1 Too 
o 2872 
: E 
o £ © 
D + Ap € Í-0 o 
o T T T T 1 © T T T 
0 1000 2000 3000 4000 © 1500 2000 2500 
Herbarium specimen records Relative rich t 5000 speci records 
Figure l. Trade-offl ision and sample size, e peanon E n coefficients (line and circles) and their 
| ated by rarefaction as the expected number 


95% confidence intervals (dotted lines) f fort 
of species in 5000 specimen records and richness estimated by rarefaction at various other numbers of specimen records (in 


triangles): the number of sampling units in our study area that would be available for analyst ET ric chinese estimates based on 
rarefaction at a given aumie af specimen records WS emed acceptable. By at 500 npecumen 


1 1 h 


records we adopted a 
d relatively precise richness estimates. —B. “Re lat ete between relative plant richness estimated by favefaction at 5000 

records (in the abscissa) and 500 herbarium specimen records (in the ordinate) across 41 sampling units 
i: 100 x 100 km, each with at least 5000 specimen records. 


values of regression coefficients obtained by previous Gentry, 1988; Table 1). We calculated variance in 
studies ae 1998; Francis & Currie, 2003; Field elevation within each sampling unit from elevation 
et al., at a resolution of 90 X m (Shuttle Radar 
The a was represented by three models thought to Topography Mission, USGS, 2004). We derive 
explain a major portion of broad-scale spatial variance spatial variation in climate within sampling units 
in plant richness worldwide (O'Brien, 1998; Francis & from 19 variables measuring various aspects of 
Currie, 2003; Field et al., 2005; Table 1). Importantly, temperature and precipitation at a resolution of 30 
we selected data sources to reflect the original arcseconds, obtained from WorldClim. Principal 
Pipe of each model as closely as possible. Thus, Soran analysis on these 19 variables captured 
models 1 and 2 (Francis & Currie, 2003; 89.4% of the total variation in the first three principal 
Table E we obtained annual potential evapotranspira- components. Because variance within sampling units 
tion and water deficit from Ahn and Tateishi (1994), in these three principal components was highly 
while for SE model 3 (O’Brien, 1998; Field et al., 2005; correlated (Pearson's r > 0.75; P < 0.05), we used 
Table 1) we calculated minimum monthly potential only the variance within sampling units in the first 
evapotranspiration using mean monthly temperature principal component as an explanatory variable. We 
and Thornthwaite's formula (Thornthwaite, 1948; see calculated variance in available water capacity and 
details in Jiménez et al., 2009). We obtained data for soil carbon density within each sampling unit from 
mean annual temperature, mean monthly temperature, data at a 5 X 5 min. resolution (Global Soil Data 
and annual precipitation from WorldClim (<http:// Task). 
www.worldclim.org/>; Hijmans et al., 2005). We also considered regression models that com- 
The SH was represented by three regression models bined terms representing the SE and SH. One model 
based on hypothesized effects on plant richness of was the Interim General Model second-generation 
within-sampling-unit spatial variation in elevation — (IGM2) (O'Brien et al, 2000; Field et al, 2005; 
(Currie € Paquin, 1987; O’Brien et al., 2000; Kreft & Table 1), regarded as one of the best ae models 
Jetz, 2007), climate (Currie € Paquin, 1987; Linder, to explain broad-scale patterns of woody plant 
2003), and soil (Linder, 2003; Tuomisto et al, 2003; richness (O'Brien et al., 2000; Field et al, 2005; 


473 


Distler et al. 


Volume 96, Number 3 


2009 


Plant Richness across the Western Neotropics 


"eououry YINOS N ‘YSN feoneury enuon y ‘yS teououry enuon N ‘VON ‘sones oaneSou tnra srun Surdures ur moso 0} payorpord ore syuepd Apoom 
ON “(G00Z ^T? ?9 PPI) ZINOT Pf) 107 siuororgooo eqo? ein Aq parorpord se ssouyou soroods yuejd Apoo “q— 'so[qerreA. Axroyeue[dxo jo oSuer opdures oq 01 pojornsox st uonorpozq “(q AQLI) 
spog [euorgox YHA (ZIAD) dd -puooes [opo] [exeuoz) urtroqu] jo uors1oA penedsuou ew 0) y ur ep ei JO VJ ei UO poseq sseuqor eAmne[ox pooIpelg 7)— 'spooer uounoods umueqxeq 
jo raquinyy 'q— qreu 00S jo sepdures porporex 0001 pounseour ssouqou posiosqo "y— -(onpq) mog 01 (pax) uar woaz *eo[riuoo1od 
9AIJ JO [eaout ue ‘squosardon 3o[oo yoy 'sordonoeN uo]s9A ƏY} Sso10e pomquemp uy 001 p 001 jo sim Sundurs coz ut More Sur[dures [eoruejoq pue ssouyou jue[d oanepy “7 em$rgy 


M.08 — M09 NO Ios d dd Meth 


L0£--£S5-  gecc- vcc EN i / 


Oziz- £96 OSr-S8Lv NEN 
a 9 


bbs - -009 20€ - LOZ m 


SLL - - 6701 
90FL - SSLL 
ELOL - ZOFL 


1VZ6€ - ¿282l Lob - Sey MM 
a v 


474 Annals of the 
Missouri Botanical Garden 


150 200 
150 200 


1 
50 
fi 


50 


0 


0 


100 x 100 km sampling units 
100 
I 
100 x 100 km sampling units 
100 
L 


T T T T 1 T T Y T T T 1 


[^] 
o 


T 
-20 2 
Mean annual temperature (°C) Annual potential evapotranspiration (mm) 


oa 
T T T T T T T T T T T T T T T 
290. 28. 
eo EN 
= * 5 
E3] 23 4 
Bo] E- 
ge $9. 
Eo] ET 
© N x 
- 287 
x" | x 
o o 
2 o 4 e o 1 1 T T T T 1 
Q 1000 3000 5000 7000 0 250 750 1250 1750 
Annual precipitation (mm) Water deficit (mm) 
E F 
alee. a Ce Án nese Gece: 
T T T T T T T T T T T T 
225 £9- 
3 d 35 4 
Eg. ey] 
BC Es] 
4 e. 
Bo | Es] 
E TP 
CER So] 
x 4 PE 
o o 7] 
20-71 T 1 SoA 
0 50 100 150 1 3.4 5 6 7 9 
Min. monthly potential evapotranspiration (mm) Ln (elevation range +1 [m]) 
Figure 3. f iables across the western Neotropics (gray bars), and our sample of the study area 
(bars with black outline). At the top of each histogram, a line pa a dot represent the range an di mean, respectively, for the 
western Neotropics (eras), our sample 7 the w area (b lack), a samples used t nergy (SE) models ao 
the globally specified Interim General Model second-generation oe (dashed line). The summary statistics for value 


mean a es ‘A, water deficit (D. and potential evapotranspiration (B) in the samples used to generate “SE 


Volume 96, Number 3 
2009 


Distler et al. 
Plant Richness across the Western Neotropics 


Clarke & Gaston, 2006). While the IGM2 was 
specifically developed for rd plants, its concep- 
to all 


plants, at least as an DRIN Rane of the 


tual basis is general and allows application 
SE and SH. To replicate as closely as possible the 
original formulation of the IGM2, we estimated 
g unit using the 
2004) 


km. We constructed adiitiondl 


elevation range for each samplin 
T O aresecond elevation data set (USGS, 
resampled to 10 X 10 
models combining terms representing the SE and 
SH that yielded significant (P < 0.05) regression 
coefficients that were consistent with the respective 
predictions in all previous regression models. We 
examined all combinations of these explanatory 
variables, except when they were highly correlated. 
For brevity, we presented only the models with more 
by the Akaike 


Information Criterion corrected (AICc) for small 


empirical support as measured 


sample sizes, recommended when the ratio of 
sampling units to model eee is less than 40 
(Burnham & Anderson, 2002). G 
used to construct these latter models, we refer to them 
as ad hoc models (Table 1). 

The RE was represented by regression models that 
also included terms representing both the SE and SH 
(Table 2). This approach follows from the idea that 


regional effects account for variation in species 


iven the procedure 


richness that remains after the effect of environmental 
conditions of the sampling units has been accounted 
for (Schluter & Rickles s, 1993). We 
effects to regression models using dummy variables 
(Draper & Smith, 1998) to code intercepts and slopes 
for each of three major paleophysiographic regions of 
1997): 
through northern. Costa Rica (N 


added regional 


the Neotropics (Graham, central Mexico 
Central America), 
southern Costa Rica and Panama (S Central America), 
and northern South America (N South America). There 
are at least two well-known major differences in the 
history of these paleophysiographic regions (Graham, 
1997; Burnham & Graham, 9): (1) N South 
America was isolated from N Central America for 
tens of millions of years, and (2) unlike N South 
America and N Central America, S Central America 
emerged from the sea just a few million years ago. In 
addition, the three regions were probably differential- 
ly influenced by orogenic activity, and by late 
Cenozoic fluctuations in climate and sea level 
associated with glacial advances and retreats that 
intensified during the Quaternary (Graham, 1997; 
Burnham € Graham, 1999). Following previous 


approaches to examine regional effects (Schulter & 
Ricklefs, 1993; Ricklefs et al., 
regions but not 

nces. Variables representing t. 
removed from regression models if they did n not reduce 
AlCe values, following model simplification proce- 
dures suggested by Crawley (2 : 
To confront the estimated pattern of relative plant 

h Neotropies (Fig. 2A) 

against the models representing different hypotheses, 


richness across the western 


we used quantile regression through the median (Cade 
et al., 5 Vu AED using R package quantreg 
(Koenkes 2005, ; R 


2006). We used regression through the median rather 


Pei qus Core Team, 


than least squares, because the response variable 
measures plant richness in an ordinal scale and 
meaningful hypotheses about variables in an ordinal 
scale focus on order statistics such as the median 
(Wolman, 2006). We gauged the extent to which the 
data supported different regression models examining 
the statistical significance of regression coefficients 
nd their concordance with the predictions derived 
from io respective hypotheses. We also pce 
support for different regression models using AlCc. 
We calculated model fit in terms of R?, the proportion 
of the 
the response variable that is accounted for by a 


regression model (Cade et al., 2005). 


All ue ud models incorporated a covariate 


sum of absolute deviations from the median o 


P of area} to account for reduced area in 


ampling units Mod the border of the spatial 
extent of the study. Pairwise correlations among 
predictors (Table 3) showed that no predictors 


included together in a single regression model were 
highly correlated and, thus, collinearity was unlikely 
to be an issue. Furthermore, collinearity due to 
higher-order regression terms was alleviated by mean- 
centering the explanatory variables of p with 
higher-order terms (Quinn & Keough, 
tested for spatial autocorrelation in regression -— 
uals using a permutation test for Moran's I (Fortin & 
Dale iv 
(Birand. 
a was RE E we used spatial eigenvector 
mapping (SEVM; Dormann et al., 2007) to construct 
spatial regression models that estimated the relation- 


implemented with R package spdep 
hen spatial dependence in the 


ship between relative species richness and explana- 
tory variables in the absence of spatial autocorrelation 
in regression residuals. We used a forward-selection 
procedure to first include in quantile regression 


ás 
models come from Francis and Currie ii Summary a. for values of annual mean precipitation (C) and minimum 
sed by 


monthly potential evapotranspiration (E) in the samples u 


ield et al. (2005) to mom the globally specified IGM2 


come from O'Brien (1998), and those for elevation range F) are fea O'Brien et al. (20 


476 Annals of the 
Missouri Botanical Garden 


Table 1. Results from median regression models representing the species energy (SE) and spatial heterogeneity 


(SH) hypotheses. 


Regression coefficient, 


Regression coefficient, 


Models and variables Pred nonspatial patial 
Model 1 SE, R? = 0.039 (0.212) 
Log10 area 55.34 (23.89) 67.15 (20.26) 
Water deficit = E. E (0. Ea —0.02 (0.02) 
Mean annual temperature + 3 (0.11) —0 0.07) 
Water deficit X mean annual temperature — 2x e (5 X 1075 1 x 10* (3 X 1075) 
Eigenvector —119.08 (40.63} 
Eigenvector 2 210.64 (51.58) 
Eigenvector 5 —109.42 (43.95} 
Eigenvector 138.61 (40.18) 
Eigenvector 12 186.20 (55.10) 
igenvector — 164.99 (52.71) 
Model 1 SE reduced, R? = 0.037 (0.221) 
Logl0 area 51.200 (24.81) 44.58 (23.11) 
Water deficit — —0.021 (0.02) —0.02 (0.01) 
Mean annual temperature + —0.151 (0.12) —0.12 (0.07) 
Eigenvector 2 176.65 (53.93) 
Figenvector 3 —131.52 (40.62) 
Figenvector 5 —98.36 (41.72) 
Figenvector 12 193.03 (52.09) 
igenvector — 164.89 (49.37) 
Eigenvector 15 —158.76 (47.53) 
Model 2 SE, R' = 0.084 (0.177) 
Logl0 area 47.80 (24.79) 42.72 (24.06) 
Water zs —0.01 (0.02) —0.002 (0.01) 
Annual potential evapotranspiration + 0.05 (0.02) 0.05 (0.02)* 
Annual potential evapotranspiration” = —3 X 10* (8 X 1075 —3 X 10% (3 x 1075 
Eigenvector 2 194.63 (49.93) 
Eigenvector 12 219.29 (52.03) 
Figenvector 1 1.18 (51.64) 
Eigenvector 3 —93.58 (44.48)* 
Model 3 SE, R! = 0.109 (0.200) 
Logl0 area 63.52 (12.68) 42.15 (17.37) 
Annual precipitation + 0.01 (2 X 107%) 0.01 (0.003)* 
Min. monthly potential evapotranspiration + 0.15 (0.13) 0.04 (0.15) 
Min. monthly potential evapotranspiration? A —0.01 (3 x 107°} —0.02 (0.01) 
Eigenvector 164.31 (47.71)? 
Eigenvector 8 117.62 (51.81) 
Eigenvector 12 234.95 (57.95) 
Model 1 SH, R? = 0.143 (0.298) 
Logl0 area 42.59 (15.44) 37.97 (10.86) 
Log10 variance in elevation + 16.477 (2.4y 19.84 (2.11) 
Eigenvector 2 158.58 (29.35) 
Eigenvector 3 — 139.58 (27.87) 
Eigenvector 12 125.16 (45.66) 
Eigenvector 11 100.16 (33.61) 
Eigenvector 10 —137.96 (33.53) 
Model 2 SH, R' = 0.132 (0.253) 
Logl0 area 44.80 (24.91) 33.88 (22.81) 
+ 66.718 (6.37) 60.75 (9.49) 


ogl0 variance in climate PC 1 
Eigenvector 2 
Eigenvector 3 


Volume 96, Number 3 Distler et al. 
2009 Plant Richness across the Western Neotropics 


Table 1. Continued. 


Regression coefficient, Regression coefficient, 
Models and variables Pred nonspatial spatial 
Fagenvector 12 147.93 (40.46) 
Eigenvector 15 101.07 (41.73) 
Model 3 SH, R? = 0.051 (0.220) 
Logl0 area 14.34 (23.89) 53.16 (17.08) 
Log10 variance in available water capacity + —10.21 (5.52) — 12.78 (3.89) 
Log10 variance in soil carbon density + 20.43 (8.33) 22.46 (6.097 
Eigenvector 2 126.86 (44.30) 
Eigenvector 3 —102.54 (37.80) 
Eigenvector 7 121.81 (37.62) 
Eigenvector 12 278.07 (48.86)* 
Eigenvector 6 —146.83 (40.59) 
IGM2, R! — 0.218 (0.274) 
Logl0 area 50.79 (8.43) 55.51 (9.93) 
Annual precipitation + 0.011 (2 X 107?) 0.01 (0.003) 
Min. monthly potential evapotranspiration + 0.298 (0.11)° 0.28 (0.12) 
Min. monthly potential evapotranspiration? m —0.003 (3 X 107?) —0.01 (0.003) 
Ln elevation range + 15.037 (1.99)° 15.06 (2.01) 
Eigenvector 2 117.08 (41.17) 
Eigenvector 8 90.23 (42.97 
Eigenvector 3 —93.55 (39.41) 
Model 1 ad hoc, R! = 0.254 (0.298) 
Log10 area 52.70 (8.83) 59.35 (8.40)* 
Annual precipitation + 0.01 (0.002)° 0.01 (0.003)° 
Min. monthly potential evapotranspiration + 0.29 (0.10) 0.27 (0.12) 
Min. monthly potential evapotranspiration? = —0.01 (0.003)* —0.002 (0.003) 
Log10 variance in elevation + 18.77 (1.98) 19.64 (2.49)° 
Eigenvector 2 110.38 (40.88) 
Eigenvector 3 —109.42 (38.23) 
Eigenvector 8 109.61 (43.72) 
Model 2 ad hoc, R? = 0.258 (0.300) 
Logl0 area 49.93 (6.29) 64.49 (7.37) 
Annual precipita + 0.01 (7 X 1075* 0.01 (0.003) 
Min. monthly eni evapotranspiration + 0.280 (0.10) 0.25 (0.11) 
Min. monthly potential evapotranspiration? = —0.01 (0.002) —0.003 (0.003) 
Log10 variance in soil carbon density + 5.19 (3.38) 4.54 (3.59) 
Log10 variance in elevation + 18.477 (1.9y 18.65 (2.35)° 
Eigenvector 2 94.64 (38.38) 
Eigenvector 3 —98.03 (34.85) 
Eigenvector 8 102.67 (40.44) 
Model 3 ad hoc, R? = 0.254 (0.300) 
Lo 43.43 (10.82) 60.95 (16.24) 
nnual precipitation + 0.01 (0.003)° 0.01 (0.002)° 
Potential evapotranspiration + —0.001 (0.02) 0.03 (0.02) 
Potential evapotranspiration? = -2xX]10*(-2 X lo“ —1 X lo“(1 X 10% 
Log10 variance in soil carbon density + 8.289 (3.90) 6.65 (4.45) 
Log10 variance in elevation + 16.025 (1.62) 16.88 (2.07) 
Eigenvector 2 136.09 (39.45) 
Eigenvector 3 —122.15 (31.55)° 
Eigenvector 8 90.70 (40.97* 


Model 4 ad hoc, R! = (0.303) 
Logl0 area 41.41 (15.19) 
Potential evapotranspiration + 0.02 (0.01) 


Annals of the 
Missouri Botanical Garden 


Table 1. 


Continued. 


Regression coefficient, Regression coefficient, 
nonspatial 


Models and variables Pred spatial 
Potential evapotranspiration" o =2 X 10* (2 x 107% 
Log10 variance in elevation + 16.19 (2.04)* 


Eigenvector 2 


174.42 (34.07 


e response variable was relative plant species richness. The first column shows names of models and explanatory 
variables, as well as goodness of fit (R?) for nonspatial regression models first and "en spatial models in parentheses. a 
n la ? shows t 1 regression 


significance coded as: *, P < 0.05; ^, P < 0.01 


models those spatial eigenvectors that most reduced 
the spatial autocorrelation in the regression residuals 
(see Jiménez et al., 2009). Given that the interpreta- 
tion of differences between coefficients derived from 
spatial and nonspatial regression models remains 
controversial (Dorman et a . we provided 
results from both types of models. 

To prediet relative plant richness across the 
western Neotropics, including areas that have been 
only sparingly collected, we used the predicted values 
d by the 
data—that is, models that yielded statistically signif- 


rom the regression models best supporte 
icant regression coefficients, were consistent with the 
respective predictions, and had the lowest AlCe 
values. To avoid extrapolation, we predicted richness 
only in sampling units that fell within the range of 
explanatory variables in our sample of the western 
Neotropies (Fig. 3) 


RESULTS 
NONSPATIAL MODELS 


The performance of nonspatial models representing 
a single hypothesis indicated that the SH had more 
empirical support than the SE. All three models 
representing the - yielded significant regression 


coefficients consisten 


with the respective predictions, 


while only two out of three models representing the SE 
did so (Table 1). The AICc values for SH models 1 
and 2 were notably lower than those for other models 
hypothesis, differing by > 17 
4) and indicating large differences in empirical 


representing a single 


support. SH models 1 and 2 also accounted for a 
higher proportion of the variation in relative species 
richness (R? = 0.143 and 0.132, respectively) than 
other models (Table 1). Thus, the relative perfor- 
mance of SH models 1 and 2 suggested a major role 
for variance (within sampling units) in elevation and 
determinants of plant 


climate, respectively, as 


richness across the study area. 


coefficients accor e respective hypothesis. The nex 


01. PC, principal component; IGM2, Interim General Model second- 
generation; Log10, logarithm base 10; Ln, natural logarithm. 


The performance of models combining terms 
representing the SE and SH was, with no exception, 
superior to that of models representing a single 
Each of these four models yielded 
significant regression coeffi As AO wit 
predictions from both the SE and e 1). AlCe 
values for the IGM2 and ad hoc i = to 3 were 


substantially smaller than those for models represent- 


hypothesis. 


ing an individual hypothesis, differing by > 41 
Fig. 4). In addition, the IGM2 and ad hoc models 1 to 


3 accounted for more : 
> 


— 


variation in relative richness (R 

0.21) than models representing a single hypothesis 
(Table 1). The higher performance of ad hoc models 1 
to 3 compared to models representing a single 
hypothesis and to the IGM2 could be partly due to 
over-fitting, given the ad hoe procedure used to build 
the former models (see Methods). However, we used 
the IGM2 as an a priori model and, therefore, at least 
in this case, the higher performance of a model that 
combines hypotheses relative to models that represent 
a single MM indicates the complementary 
nature of the and SH. 

All models n included terms representing the RE 
yielded significant regression coefficients consistent 
with predictions from both the SE and the SH. Each of 
these models also yielded significant coefficients 
representing regional effects (Table 2), but we had no 
predictions about their sign (see Methods). Models 
representing the RE did best overall, with lower AICe 
values differing by > 7 from models combining terms 
representing the SE and SH with no regional effects 
Fig. 4), although respective differences in model fit 
were slight (Tables 
in model performance accomplished by adding regional 


). In general, the improvement 


effects was smaller than that achieved by combining 
terms representing the SE and SH. This finding 
suggests that the RE, as represented in this study, 
does not complement the SE and SH to the same degree 
that the latter two hypotheses complement each other. 

Nonetheless, adding regional effects to models 
representing the SE and SH decreased AICe values 


Volume 96, Number 3 
2009 


Distler et al. 
Plant Richness across the Western Neotropics 


substantially. For example, adding regional effeets to 


the IGM2 decreased the AlCe value by 23. The 
resulting model suggested that, after PP for 
other environmental variables, relative richness 


higher in S Central res than in N South "abe 
and higher in N South America than in N Central 
merica, across range of minimum monthly 
potential evapotranspiration and annual precipitation 
in our sample (Fig. 5A, B). In addition, maximum 
plant richness was attained at higher values along the 
axis of minimum monthly pcr Mi pe ped 
in N Central America than in S Central 
N South America (Fig. 5A). Rear Er eae 
increased faster with elevation range (within sampling 
units) in N Central America than in S Central America 
and N South America. Relative richness in sampling 
units with narrow elevation ranges was lower in N 
Central America than in S Central America and N 
South America, but the difference in relative richness 
decreased as elevation range increased 
There were differences among ad hoc models in the 
strength and significance of regional effects, but all 
models revealed that N Central America, compared 
with S Central America and N South America, had a 
lower intercept and a steeper slope relating relative 
richness to variation in elevation within sampling 


units (Table 2). 


SPATIAL MODELS 


Similar to the results from nonspatial models, the 
performance of spatial models representing a single 
hypothesis indicated more empirical support for the SH 
than the SE. All three models representing the SH 
yielded significant regression coefficients consistent 
with the respective predictions (Table 1). However, SH 
model 3 also yielded a 


coefficient for variance in potential available water 


significant negative regression 


capacity, contrary to the respective prediction. Two of 
three models representing the SE, SE models 2 and 3, 
yielded significant regression coefficients consistent 
with the respective predictions, while SE model 1 
yielded a significant negative regression coefficient for 
m annual temperature, contrary to the respective 
prediction (Table 1). Among spatial models represent- 
ing a single hypothesis, SH models 1 and 2 explained 
more variation in the response variable (R? = 0.298 
and 0.253, respectively; Table 1) and had substantially 
lower AICe values 


results from nonspatial models and suggesting a 


than the rest (Fig. 4), corroborating 


primary role for variance in elevation or climate, 
within sampling units, as determinants of plant 
richness across the study regi 

ach of the five spatial ane combining terms 
representing the SE and SH yielded statistically 


significant regression coefficients that were consistent 
with the predictions derived from both hypotheses. In 
no case was there a F a coefficient 
models 


performed better than n de an n indiéd- 


opposite to any prediction (Ta 


ual hypothesis, with a notable exception: the spatial 
version of SH mo 
relationship between variance in 
sampling units and relative plant richness, explained 
ore variation in the response variable and had a 
dep lower AICe value than the IGM2, in both 
odels 1 to 4 (Table 1 
oe P This exception strengthens the previous 


ics similar to ad hoc m 


suggestion that variance in elevation within sampling 


units, a variable epa the SH, was a primary 


determinant of p ichness across the study region. 
It also suggests Le spatial eigenvectors accounted for 
variation in relative plant richness that correlated with 
variables representing the SE, but not for variation in 
relative plant richness that correlated with variables 
representing the SH 

All spatial models including terms representing the 
RE yielded statistically significant regression coeffi- 
cients consistent with predictions from both the SE 
an , as well as significant coefficients represent- 
ing regional effects. In only one case was there a 
statistically significant coefficient opposite to 
prediction: 
yield 
transpiration (Table 2). Spatial models including 
regional effects explained only a slightly higher 
proportion of the variation in the response variable 
than models including terms representing the SE and 
SH only (Tables 1, 2), and their AICc values were not 
consistently lower than those of other models (Fig. 4). 
his result contrasts with the respective comparison 
for nonspatial models and suggests that spatial 
eigenvectors accounted for variation in relative plant 
richness that correlated with variables representing 
the RE. 
regional effects yielded the lowest 
model 2; Fig. 4) and the highest proportion of 
explained variation in relative species richness (Rt 
— 0.326; Table 2). This model revealed similar 
regional effects to those described by nonspatial 


Nonetheless, a spatial model including 


models, whereby N Central America compared to 
other paleophysiographie regions had a lower inter- 
cept and a steeper slope relating relative richness to 
variance in elevation within sampling units (Table 2; 


Fig. 5D-F). 


PREDICTED PLANT RICHNESS MAPS 


The models best supported by the data yielded 
similar patterns of predicted relative species richness 


480 Annals of the 
Missouri Botanical Garden 


Table 2. Results from median regression models combining the species energy, spatial heterogeneity, and regional 
effects hypotheses. 


Regression coefficient, Regression coefficient, 
Models and variables Pred nonspatial spatial 
IGM2, R! — 0.263 (0.297) 
Logl0 area 55.54 (13.49) 64.749 (18.32) 
Annual precipitation + 0.01 (0.002)° 0.01 (0.003)° 
Min. monthly potential evapotranspiration + 0.15 (0.14) 0.07 (0.13) 
Min. monthly potential evapotranspiration? —0.01 (0.004)? —0.01 (0.004)? 
Ln elevation range 10.39 (1.95) 10.55 (1.91) 
NCA —112.35 (21.65) —111.42 (30.07) 
SCA 23.63 (11.96)* 29.22 (9.24) 
Min. monthly potential evapotranspiration X NCA 0.87 (0.28) 0.70 (0.357 
Ln elevation range X NCA 16.00 (3.23y 15.15 (4.34)° 
Eigenvector 3 —140.92 (31.76) 
Eigenvector 2 74.53 (53.74) 
Model 1 ad hoc, R! = 0.284 (0.308) 
Log10 area 50.87 (7.61)° 65.41 (7.34)° 
Annual precipitati + 0.01 (0.002)° 0.01 (0.002)° 
Min. monthly na evapotranspiration + 0.14 (0.13) 0.001 (0.12 
Min. monthly potential evapotranspiration" = —0.01 (0.004)° —0.01 (0.003)* 
Log10 variance in elevation + 14.73 (2.49) 14.13 (2.18) 
—70.49 (22.79) —60 8.04)* 
SCA 59.05 (50.46) 102.91 (53.73) 
Annual precipitation X SCA 0.03 (0.01) 0.04 (0.01)* 
Min. monthly potential evapotranspiration X NCA 0.72 (0.25) 0.69 (0.25) 
Log10 variance in elevation X NCA 14.06 (4.67) 10.95 (5.61) 
Log10 variance in elevation X SCA 24.58 (10.52) —33.43 (11.63) 
Eigenvector 3 —153.32 (29.98) 
Model 2 ad hoc, R! = 0.277 (0.326) 
Logl0 area 48.65 (9.44)° 50.08 (13.57)° 
Annual precipitation + 0.01 (0.002)° 0.01 (0.003) 
Min. monthly potential evapotranspiration + 0.31 (0.11) 0.33 (0.12) 
Min. monthly potential evapotranspiratior” E —0.01 (0.003) —0.01 (0.005) 
Log10 variance in soil carbon density + 10.15 (3.72) 3.77 (3.40) 
Log10 variance in elevation + 15.36 (2.42)° 15.36 (2.49)° 
—111.96 (30.75) —86.68 (36.12) 
Annual precipitation X NCA 0.01 (0.01) 2 X 10^* (0.01) 
Logl0 variance in elevation X NCA 17.84 (5.40y 17.90 (6.11) 
Eigenvector 8 98.63 (35.58) 
Figenvector 2 149.03 (40.37) 
Eigenvector 10 —112.11 (31.40y 
Model 3 ad hoc, R! = 0.291 (0.304) 
Log10 area 48.25 (8.73) 44.42 (11.41) 
Annual precipitation + 0.01 (0.003)° 0.01 (0.003)° 
Annual potential evapotranspiration + —0.04. (0.02) —0.06 (0.02) 
Annual potential evapotranspiration” 2 —3 x 10% (1 X 107%" —3 X 10% (1 X 107% 
Log10 variance in soil carbon density + 10.81 (3.85) 13.95 (4.55) 
Log10 variance in elevation + 12.22 (1.95) 11.24 (2.45y 
NCA —56.45 (17.41y —62.95 (22.46) 
SCA 25.56 (9.52) 26.78 (12.07) 
Annual potential evapotranspiration X NCA 0.14 (0.05 0.17 (0.04) 
Logl0 variance in elevation X NCA 10.74 (3.59? 12.21 (4.55) 
Figenvector 7 97.51 (42.37) 


Model 4 ad hoc, R? = (0.272) 
Logl0 area 57.32 (13.73y 


Volume 96, Number 3 Distler et al. 
2009 Plant Richness across the Western Neotropics 


Table 2. Continued. 


Regression coefficient, Regression coefficient, 
Models and variables Pred nonspatial 1 
Annual potential evapotranspiration + 0.02 (0.02) 
Annual potential evapotranspiration” = —2 X 10* (2 x lo“ 
Log10 variance in elevation + 15.11 (1.94)° 
NCA — 17.78 (7.04) 
32.58 (7.35)° 
Annual potential evapotranspiration X NCA 0.07 (0.05) 
Eigenvector 3 — 184.55 (35.16) 
Eigenvector 2 83.90 (33.53)? 

Th. Ta ] E bare ba i! pos ] } E del d l iables, 
as we ella as A of fit (R p) for nonspatial ier c models nur and lor ae modeli in parents: The column labeled 
“Pred” show: siie predi cted si coefficient hypothesis. T w 
M for nonspatial Bal sai models with standard errors in parentheses and statistic: ded 
as: *, P < 0.05; 5 P < 0.01; °, P < 0.001. IGM2, Interim General Model second- generation; Logl0, logit base 10; Ln, 


ned logarithm; NCA, dummy variable for N Central America; SCA, dummy variable for S Central Americ 


across the western Neotropics (Figs. 2C, 6A-D). western portion of the Andes in Peru (Fig. 6C), 
Generally, the highest species richness was predicted presumably because this latter model did not include 
in topographically complex areas such as the any variable explicitly measuring water availability. 


including the Sierra Madre de Chiapas, the mountain Mexico and Yucatán, Los Llanos of Venezuela, and in 
ranges extending from the Cordillera de Tilarán the Gran Chaco region of Bolivia, Paraguay, and 
southeast along the Cordillera de Talamanca into Argentina. All models also predicted a species 
Panama's Cordillera Central, the Andes, and the richness trough in lowland Amazonia relative to S 
Venezuelan Guayana. Areas predicted to have highest Central America, the Andes, and the Venezuelan 
richness formed a longitudinally broad band in uayana. These predicted patterns of relative plant 
Colombia, encompassing the Chocó region and all richness should be seen in the light of important 
three Andean cordilleras, and included both Andean differences between observed and predicted richness 
cordilleras in Ecuador. These areas were largely values. Specifically, even the models best supported 
restricted to the eastern Andes in Peru and Bolivia by the data accounted for relatively small portions of 
according to most models, with the exception of adhoc the variation in the response variable (R! = 0.258- 


model 4, which predicted high plant richness in the — 0.326; Tables 1, 2). 


Table 3. Pearson's correlation coefficients among variables used in regression models representing the species energy 
and spatial heterogeneity hypotheses 


Area PET WD Temp Precip mPET Elev Range Soil Pawe | PCAI 

PET 0.352 1 
WD 0.234 0.121 1.000 

0.301 0.489  —0.155 1.000 
Precip 0.184 0.448 —0.518 0.392 1.000 

0.321 0.611 —0.123 0.775 0.536 1.000 
Elev 0.137  —0.070 0.282 —0.596 —0.121  —0.382 1.000 
Range 2  —0.061 0.201 —0.563 —0.118  —0.375 0.951 1.00 
Soil —0.256 0.084  —0.050 —0.055 0.145 0.121 069 0.001 1 
Pawe —0.095 0.058 0.129 —0.083  —0.013 0.060 | 0.055 0.033 0.545 1.000 
PCAI 0.073 0.020 0.220 —0.444 0.031 —0.261 0.838 0.799 0.161 0.112 1.000 
Rich 0.166 0.149 —0.146  —0.094 0.301 0.105 0.899 0.346 0.130 | —0.008 0.401 

a, logl0 area; PET, potential evapotranspiration; WD, water deficit; Temp, annual mean temperature; Precip, annual 


> 
Ns mPET, minimum monthly potential DUNG Elev, logl0 variance in elevation; Range, In elevation 
range; Soil, log10 variance in soil carbon density; Pawe, lo E variance in available water capacity; PCA1, log10 variance in 

climate's first principal component; Rich, relative plant ri red by rarefaction at 500 herbarium specimen records. 


Annals of the 
Missouri Botanical Garden 


Model 1 SE Jj 

Model 2 SE Jj 

Model 3 SE - 
Model 1 SH 4j 

Model 2 SH 4j 

Model 3 SH 4j 

IGM2 4j 

Model 1 ad hoc - 
Model 2 ad hoc - 
Model 3 ad hoc - 
Model 4 ad hoc - 
IGM2+RE - 
Model 1 ad hoc+RE Jj 
Model 2 ad hoc*RE JA 
Model 3 ad hoc* RE 4j 
Model 4 ad hoc+RE + 


| | | 
50 100 150 


Delta AlCc 


upport for models representing the species energy (SE), spatial vou (SH), and regional effects (RE) 


rianel 


ran for small sample sizes (delta A 


ge 
c 
e 
Ej 
> 
yan 
o 
ds 
> 
(0) 
"8 
Q 
e 
ge, 
= 
= 
^x 
E 
= 
o 
— 
o 
= 
o 
E 
+= 
> 5 
= 
oo 
o 
= 
=| 
23 
o 
d 
© 
= 
N 
ne 
as 
= 
ES] 
E, 
u- 
ae 
EE 
o 
EE 
— 
= 
=} 
o 
o 
E 
O 
= E 
ES 
5 
S 
= 
o 


dle), and models 


i m ir e nonspati models 


Filled symbols RN regression models that p statistically aes regression coeffi cients consistent with the 
mb models 


respective predictions. Open symbols rep 


regression c aL that y were neon nt with 
quadratic ter significant, w 


s) 
"Therefore, some models are represented by two circles or end 


DISCUSSION 


Our results supported most a priori predictions 
based on previous studies about determinants of plant 
richness (Currie & Paquin, 1987; Gentry, 1988; 
O'Brien, 1998; O'Brien et al., 2000; Francis & Currie, 
2003; Linder, 2003; Tuomisto et al., 2003; Field et al., 
2005; Kreft & Jetz, 2007). Nine of the 13 predictions 


regarding the sign of regression coefficients relating 


that yielded no statistically significant. regression 
mark vein models that a Mose significant 
When higher-order terms (interaction or 
i in of reduced models with no higher-order terms. 


plant richness to explanatory variables were supported 
at least once (potential evapotranspiration an 


nthly pote 


and its square, annual precipitation, within-sampling- 


p its 
square, minimum mo: ntial evapotranspiration 
unit range in elevation, within-sampling-unit variance 
in elevation, within-sampling-unit variance in climate, 
and within-sampling-unit variance in soil carbon 
density); three predictions were o e (annual 

ean temperature, potential evapotranspiration, and 


Volume 96, Number 3 Distler e 
2009 Plant ae across the Western Neotropics 


ES E 
Ww) 
E 2 
c o 
S 94 287 
E E 
Ee 2 
ST mo 
5 E 
E | z 
£2 £ 
e " 7 2 4 
2 2 
Es] J so 
© $ 9. 
c 2 al o Xx i 
> o 
T T T T T T T T i T T T T T 
20 40 60 80 100 120 140 o 1000 00 O 7000 


Min. monthly potential evapotranspiration 


mm) Annual precipitation (mm 


C D 


B - : 3 
27 LM 
Oo oO 
= 3 
oo o 
€ 9 E] 
B" 9 
S S 
E m] c 4 
ge E IE E ES 
o o BSP [e] 
3 $ o € -a PE 2 6s 
£ ] £ o .. d" $00 S SY o 
9 O | [o] 
= = ® Qe 
2 2 | 2 3 | o, 2 0 
S! S o 
És Lo o 
o o Ss . 
? y . Q o 
l T T T T T T ! T T T T T 
2 8 120 140 
Ln(elevation range [m] ) Min. monthly potential evapotranspiration (mm) 
E F 
e 
&.3 
238 e” 
E FP 
2 38. 
9 oj g 
S Boe 
3 &' 
a a t 
go 3 
D o " 
£2 % E 
a — 7 o 00 2 
o ! Lo o oo 
Zo ie So 2°77 
SG D Es 
or} 9 © 
Yo " © gj? 
. "Y 
T T f T T E T T T T T T T 
0 1000 3000 5000 7000 1 2 8 4 5 6 
Annual precipitation (mm) Log10( variance in elevation [m?] ) 


Figure 5. Plots of partial residuals (Draper & Smith, 1998) for models that incorporate the species energy (SE), spatial 
En er (SH), and regional effects (RE) hypotheses: —A—C. Nonspatial version of Interim General Model second- 
generation (IGM2) with regional effects (Table SM —D-F. Spatial version of ad hoc model 2 with regional effects (see Table 2). 


variables in the model are statistically controlled except for regional effects. Open circles (and dashed lines) represent N 
Central America, open triangles (and solid lines) represent S Central America, and filled circles (and solid lines) represent N 
South America. 


Annals of the 


484 


Missouri Botanical Garden 


'Se[qerreA ÁAroyeue[dxo jo oSuer o[dures ou 01 PALISO st uonorpoiq (q opqe) 89O 


peuo ira g [opour ooy pe jo uorsroA [erreds oy} uo poseg ssouqou pojorpexq “G— “(T 9pqer) p [opour oou pe jo uorero4 [eryeds ou uo poseq ssouqou pajorpa1g ^)— “(q 9[qe) s199]79 [euotser 
UNA € popou ooy pe jo uorsIos [erredsuou oy} uo poseq ssouyoL pojorpo1q "q— `(T SAB) z [opour oou pe jo uorsroA [enedsuou oy} uo poseq ssounorr pojorpoxq "y— “(an1q) mo] o1 (pax) ySty wox 
‘Uy OOT x OOT Jo uonn[osor e ye poxnseour ssouyoL soroods jo so[nuoo:od oArj jo [eA1oqur ue squoso1dox 10[02 YOR *sordonoow utojs9A OY} SSOLDB ssougorr jue[d o4rpe[or pojorpo:q “9 omSry 


LZE- Ele 


LPP - - A 
a 


LLE - Lhe 


Ove - ooc NH 


ét -viv E 
a] 


6ct - 262 EH 


GSF -cer m 
v 


M04 M-08 Me06 M00 L M LL 


MOS i id 


Mor Ming AEG Nen Muah M08 M.09 


-8.0c 


-5.01 


=N.OL 


—N.0Z 


Volume 96, Number 3 
2009 


Distler et al. 
Plant Richness across the Western Neotropics 


in available water 
supported 
nor opposed (water deficit and the interaction between 


within-sampling-unit variance 


capacity); and two predictions were neither 


water deficit and a by 
deni d e also 
fo cally signi 

plant Cre among similar environments located in 


nt results (Tables 1, 
nificant differences in relative 


different ded Ric d i "od 2; E 5) 
defined accordin ork (Graham, 1997; 
Burnham & us 1999. These findings ied that 
we used a reasonable set of hypotheses and corre- 
sponding models to examine the relative importance of 
the major determinants of broad-scale plant richness 
across the western Neotropics. 

The performance of nonspatial and spatial regression 
models, measured by the consistency of regression 
coefficients with a priori predictions and by AICe and 
R? values, indicated that explanatory variables repre- 
senting the SH we 
richness across the western Neotropics, with comple- 


re the primary determinants of plant 


mentary contributions from variables representing the 
SE and, to a lesser extent, the RE. In particular, 
and variance in climate within 
e the main 
predictors of the estimated pattern "a liue plant 


variance in elevatio tio: 


sampling units, ot the SH 


richness across the study area. The models performing 
ion and not 
of these 
two variables was more important was difficult because 
they were highly correlated (Table 3). A third variable 


representing the SH, variance in soil carbon density 


best overall included variance in elevati 
variance in climate, but distinguishing whic. 


within sampling units, was included in some of the best 
performing models, but its role was secondary to that of 
variance in elevation or climate. Annual precipitation, 
minimum monthly potential iari rine and 
potential evapotranspiration, representing the SE, most 
effectively improved the performance of model 
representing the SH. Here, agen. determining whether 
minimum monthly potential am aa or 
potential ae aed was mo 
difficult se they were ie now 3) 
Finally, eee representing the RE sometimes 
increased the performance of models combining terms 
representing the SH and SE. Most frequent among 
hese were variables indicating a lower intercept for N 
Central America than for other regions, a higher 
intercept for S Central America than for other regions, 
and higher slopes relating plant richness to spatial 
heterogeneity within sampling units (variance and 
range in elevation) and to water (annual precipitation) 
and energy availability (potential evapotranspiration 
and minimum monthly potential evapotranspiration) in 
N Central Ámerica than in other regions 

One of the main findings emerging from our 
analysis was more empirical support for the SH than 


for the SE, consistent with a similar study of broad- 
scale plant richness across northwestern South 
America that found at least as much support for the 
SH as for the SE (Jiménez et al., 2009). This result 
would seem at odds with previous work indicating that 
variables representing the SE are more important 
determinants of broad-scale patterns of plant richness 
than those ip e SH (Curie & Paquin, 
1987; O'Brien et al., As et al., 2003; 
Bjorholm et al., 2005; idi de 2005; Moser et al., 
2005; Kreft et al, 2006; Kreft n Tes, 2007). We are 
aware of only one earlier plant study (Pausas et al., 
~ a ER: a end role for the SH ouside 
study regio explore potential 
POSSE Um Ps ile ieee between our results 
and those from previous studies, related to the 
characteristics of the response variable and the 
distribution of the explanatory variables. 
eographie patterns of plant richness measured in 
small sampling units (e.g, = lha. plots) are 
sometimes considered broad scale (e.g., Hawkins et 
al., 2003), but the importance of different hypotheses 
can be contingent on sampling unit size. For example, 
the relationship between plant richness and variables 
representing the SE may be most evident when 
nits, while other 
d thus be 


more important determinants of spatial variation in 


measured across large sampli 
factors may exhibit greater e an 


richness at smaller grains (Whittaker et al., 
This would seem to be at least as much of a 
variables representing the SH. Specifically, = 
proposes that spatial heterogeneity fosters species 
coexistence across habitats and isolated patches of 
similar habitat, or that it accelerates speciation rates 
by increasing opportunities for isolatio l 

Both of these effects are likely 
increasingly opposed by dispersal as sampling unit 
us, the 
may increase with 


n and ecolog- 
ical divergence. 


size decreases (cf. Moser et al., 

relative importance of the SH 

sampling unit size, as suggested by studies of bird 

richness (Rahbek & Graves, 2001; van Rensburg et 
002; Hulbert & Haskell, 2003). 

E s procedures to estimate the response 
variable, previous studies estimated richness by 
M M eographie range maps (e.g. O'Brien, 

1998; Francis & Currie, 2003; Bjorholm et. al., 2005), 
while our estimates are based solely on locality data 
from herbarium specimens. These two methods may 
yield different estimates of spatial richness patterns 
and correspondingly different rankings of the impor- 
tance of different hypotheses (Hulbert & White, 2005). 
Because species do not typically occur everywhere 
within the area delimited by range maps (Rondinini et 
al., 2006), richness estimates based on range maps may 
measure richness at a larger grain than estimates based 


Annals of the 
Missouri Botanical Garden 


Figu 
range and sll precip taton, (B) 


Annual precipitation (mm) 


y e 
a “7 ` Q 
- 

T Bu: .- 
o 24 E 
D 
E 
g 
= 
S 44 
T 
> 
o 
o 
2 ad 
[7 

o - 

T T T T 
0 50 100 150 

Min. monthly potential evapotranspiration (mm) 

E 


ts inclu 


MRS annual precipitation for min. monthly 


> 


potential evapotranspiration [mm/mm] 


4000 6000 
L L 


2000 
L 


Ln (elevation range *1 [m]) 


200 


100 


in. monthly potential evapotranspiration ( 


mm 


MRS annual precipitation for 
n (elevation range +1) [mm/Ln(m)] 


B 
7 EO : 
2) L1 e 
E © i A ? 
E | ` A 
= A 2 
BG * o8 $ 
5847 VON ; 
a + d E 
8 | 3 1 - 
38 O 
S oJ s Sy 
28 + 
< E ` 
od , / 
T T T T 
0 50 100 150 


Min. monthly potential evapotranspiration (mm) 


J 


5000 
| 


3000 


1000 
1 


0 
l 


Ln (elevation range +1 [m]) 


T 


min. monthly 
/mm] 


potential evapotranspiration [Ln(m). 


1) for 


A) fi 


0 50 


MRSL 


A 
150 
monthly potential evapotranspiration (mm) 


T 
100 
Min. 


nimum 


B, C. Predicted plant richness isopleths (lines) in the bivariate os of (A) logarithm of d 
)m 


nthly pot 


minimum mo: 


ential evapotranspiration 
ithm of ek tion act across Ne western Neotropics (gray nd showing 


nual precipitation, and ( 


ne). The continuous lines are isopleths based on 


global 


w of the Interim General Model second-generation (GM), The dashed lines are IE for N el jim 


Volume 96, Number 3 
2009 


Distler et al. 
Plant Richness across the Western Neotropics 


on locality data (Hulbert & Jetz, 2007). As such, 
richness estimates based on range maps may tend to 
favor variables representing the SH, related to broad- 
rt & White, 2005). This 


expectation, however, is opposite to the difference 


scale species turnover (Hulbe 


between our results and those from previous studies. 
Another important difference is that between 


estimates of plant richness based on floras and 


attempt to correct for spatial 
variation in floristic knowledge. Floras and checklists 
differ not only in their geographic extents, but in the 

mpling effort across these 


2005) Such 
knowledge can affect 


quality of the data mo sa 
2001; Kier et al, 


in floristic 


units (Frodin, 
heterogeneity 

estimates of spatial patterns of plant richness and 
estimates of the importance of different hypotheses. 
We attempted to account for this heterogeneity using 
number of specimen records as an estimate of 
1990), assuming that 
the number of species found in a i it 1 


sampling effort (Nelson et al., 
mpung unit 1s a 
function of the number of specimens collected in that 
sampling unit. Nonetheless, it is possible to imagine 
scenarios in which the number of specimens collected 
in a ~ unit is a function of the number of 
spec We think the 
a is unlikely because plant species inventories of 
100 x 


are invariably incomplete. Furthermore, the inventory 


es occurring in that sampling unit. 


100 km sampling units across our study area 


for each sampling unit was derived from several 
collecting trips that, together, are bound to obtain 
multiple specimens of species that are common in the 
sampling unit, even if some m collectors discrim- 
inate against common sp 

Studies also differ in d km life forms on which 
they focus and, thus, may favor one hypothesis over 
another because various life forms may respond 
differently to different factors (Richerson & Lum, 
1980; Laanisto et al., 2008). For example, tree species 
might be less responsive to spatial RC E than 
nonwoody plant species (Gentry, 19 an & 
Ricklefs, 2004). However, among audis measuring 
species richness of both woody and nonwoody life 


forms, some conclude that the SE is more important 
than the SH (Francis & Currie, 
opposite (Pausas e 
useful for future studies to specifically address this 
issue by confronting species richness data for different 
life forms against models representing the SE and SH. 
Differences in the central tendency of explanatory 
variables may also explain why our results differ from 
those of other studies in terms of the relative 
importance of the SE and SH. In tropical and 
subtropical regions, where energy input is high (> 
mm potential evapotranspiration, Jetz, 
2007), plant richness may be largely 
energy and mainly determined by water availability 
(Hawkins et al., 2003; Whittaker et al., 2006; Kreft & 
Jetz, 2007) and, to a lesser extent, spatial heterogeneity 
Kreft & Jetz, 2007). Despite the fact that potential 
evapotranspiration values were well above 505 mm in 


eit 
independent of 


= 


all of our sampling units (Fig. 3), we obtained 
statistically significant coefficients supporting predic- 
tions derived from the SE abou 
related to potential evapotranspiration and minimum 


t how plant richness is 


monthly potential evapotranspiration (Tables 1, 2; 
Fig. 5). Thus, our results suggest that both energy 
and water 2: o determine plant richness 
across the egions but to a lesser extent than 


udy r 
variables ee the SH. Our sample, however, 
includes few sampling units from some extreme 
environments in the study regions, particularly those 
with the lowest annual precipitation (< mm 
annual precipitation) and highest water deficit (> 
750 mm; Fig. 3). If the effect of annual precipitation or 
water deficit on plant richness decreases with increas- 
ing water availability (Gentry, 1988; Whittaker et al., 
, our results cou av derestimated the 
importance of water availability and, therefore, the 
Differences between our results and those of other 
studies in the relative importance of the SE and SH 
may relate to differences in the range of a 
= spanning global 
reft x Jetz, 2007) 
or aiming to develop global 2 (Field et al., 
2005), 


variables. 


the ranges of several variables measuring 


e 
and the dotted lines for N South America and S Central America, oium to the fit of the eMe onm Tegional effects to our 


sample (Table 2). — 
MRS of annual precipitation for minimum 
confidence intervals are base 


IGM2 to 


o our sample. T 


LESE. Marginal rates of substitution (MRS) of 
potential evapotranspiration (E), and MRS of pe logarithm of elevation 


em 
=. 
a 
ge 
E 
EB 
€ 
un 
ET 
ao 
= 
lei 
ge 
E 
SE 
= 
iz 


ased on the global 


annual precipitation for 1 range (D), 


e dashed lines represent 95% confidence ase of the 
South America and S Central America. These 


on parametric bootstrap samples (Efron & Pe dea of size 1,000 000, assuming a 
multivariate normal distribution of He iu NM erica with the variance-cov 
repr MRSs b 


matrix estimated from fitting the 
ie of the IGM2 and have no error 


estimates because the variance-covariance BUM E such coefficients was not available. 


Annals of the 
Missouri Botanical Garden 


energy and water availability in our sample were small 
(Fig. 3). Therefore, we may have underestimated the 
importance of the SE because the proportion of the 
variation in a response variable (e.g., relative plant 
richness) accounted for by an explanatory variable 
(e.g., mean annual temperature) can be a function of 
the sample range of the latter (Pedhazur, 1997). B 
the same token, in our sample, the range of a variable 
ara the i bad mec within pum 
similar of y of 
doble extent (Keefe & n. ps e So = ” 
of a study aiming to develop global models (Field et 
al, 2005; Fig. 3). Thus, the latter study may have 
underestimated the 


importance of the . If our 


sample is representative of the variation in energy and 
water availability across our study area (Fig. 3), the 
conclusion of more empirical support for the SH than 
or the SE would still be valid for this area. 

We found significant regional effects among N 
Central America, S Central America, and N South 
America, consistent with studies that have explicitly 
tested for, and commonly found, evidence supporting 
the RE (Schluter rà fücklefss 1993; Ricklefs, 2004; 
Kreft & Jetz, 2007). Identifying the underlying causes 
of differences in plant richness attributed to regional 
effects can be difi 
reflect differences in environ 
accounted for (Schluter & Ricklefs, 1993). However, 
two of the most prominent regional effects that we 

obtained were consistent with previous assessments of 
historical influences on Neotropical plant diversity 
(Gentry, 1982). First, the lower intercept for N Central 
America than for other regions may be due to limited 
northward movement of Gondwanan clades that 
compose most Neotropical plant diversity in the 
lowlands (ic e., Amazonian-centered taxa sensu Gentry, 
1982). Y a steeper slope relating plant richness 
to spatial heterogeneity in N Central America than in 
other regions may be due to limited southward 
movement of Laurasian clades that are more important 
of montane than lowland Neotropical 
1982). 


contribute more to the increase in richness associated 


components 
floras (Gentry, Laurasian clades may then 
with spatial heterogeneity in N Central America than 
in S Central America and N South America. 

ast, we discuss implications of our findings for the 
perception of the spatial pattern of plant richness 
across the western Neotropics. Our analysis predicted 
peaks of relative species richness mostly in topo- 
graphically complex areas (Figs. 2C, 6), consistent 
with a similar analysis for northwest South America 
(Jiménez et al, 2009) and another recent mapping 
effort (Kreft & Jetz, 2007: fig 3b), but contrasting with 
maps showing higher vascular plant richness in 
lowland areas than in the northern Andes (Barthlott 


et al., 2005; Mutke & Barthlott, 2005; Kreft & Jetz, 
2007: fig. 3c, d). The latter maps are similar to the 
pattern predicted by global coefficients for the IGM2 


— 


Fig. 2D), designed for estimating broad-scale rich- 
ness of woody plants worldwide (Field et al., 2005). 
Below, we suggest a potential explanation for the 
contrast between different plant richness maps of the 
Neotropies, based on differences between the global 
coefficients of the IGM2 (Fig. 2D) and the coefficients 
yielded 7 fitting the IGM2 to our sample (Fig. 2C). 
The contrast between the of pr 
ao “eed n global coefficients of 
(Fig. 2D) and that | based on the fit of the IGM2 to our 
sample (Fig. 2C) is underlain by a difference in the 


edicted plant 
the IGM2 


importance of the logarithm of elevation range relative 
to annual precipitation. This is illustrated by isopleths 
showing combinations of logarithm of elevation range 
and annual precipitation that, according to a given 
model and holding other variables constant, yield a 
constant predicted richness value (Fig. 7A). The slope 
of the isopleths based on our sample are more 
negative than the slope of the isopleth based on the 
global coefficients of the 2, suggesting that the 
importance of the logarithm of elevation range relative 
to annual bc is higher in the fit of the IGM2 
to our data than in the global coefficients of the IGM2. 
This ae is quantifie marginal rate of 
substitution (MRS; Caraco, 1979; Brown, 1988) of 


annual oe for the logarithm of elevation 


m the is in Fig 


amount of annual i a t needed to substitute 


a 


a logarithmic unit of elevation range and maintain the 
same species richness in any given sampling unit (see 
appendix S1 note A in Jiménez et al., 2009). The MRS 
of annual precipitation for the logarithm of elevation 
range is higher in the fit of the IGM2 to our vs than 
in the global coefficients of the IGM2 (Fig 

The foregoing contrast between "a of 
predicted plant richness (Fig. 2C vs. D) also results 


from differences in the relative importance of 
minimum monthly qr M a ow and 
annual em This can be opleths 


showing combinations of A of minimum pese 
potential evapotranspiration and annual oa 
that, holding other va 


riables sas yield a con 
predicted richness value (Fig. 7B). 


These dura 
bend because, according to the IGM2, the relationship 
between minimum monthly potential evapotranspira- 
tion and plant richness is quadratic (e.g., Fig. 

The negative of the slope of these isopleths is the MRS 
of annual precipitation for minimum monthly potential 
evapotranspiration. It measures the amount of annual 
precipitation needed to substitute a small increase in 
minimum monthly potential evapotranspiration and 


Volume 96, Number 3 
2009 


Distler et al. 
Plant Richness across the Western Neotropics 


maintain the same species richness in any given 
This MRS decreases as minimum 
evapotranspiration — increases 
(Fig. 7E) because, as pointed out previously, according 
to the IGM2 the relationship between minimum monthly 
potential evapotranspiration and plant richness is 
quadratic. As minimum monthly potential evapotrans- 
piration increases, the decrease in the MRS of annual 
precipitation for minimum monthly potential evapo- 
transpiration is steeper in the fit of the IGM2 to our data 
than in the global coefficients of the IGM2 (Fig. 7E). 
This finding suggests that the importance of minimum 
monthly potential eae s es to annual 
precipitation is higher in the fit of the IGM2 to our 
sample than in the global denis of the IG 

Finally, the contrast between the map of -— 
plant richness derived from global coefficients of the 
IGM2 (Fig. 2D) and that derived from the coefficients 
yielded by fitting the IGM 
does not seem to reflect differences in the importance 


2 to our sample (Fig. 


of the logarithm of elevation range relative to 

minimum monthly potential evapotranspiration. This 

notion is suggested by isopleths showing combinations 

of values of logarithm of elevation range and minimum 

holding 

nstant predicted 
ths b 


monthly potential Va er ipid that, 


ect end 
because the IGM2 portrays a quadratic relationship 
a minimum monthly potential evapotranspira- 

on and plant richness. The negative of the s 
dee isopleths is the MRS of logarithm 
range for minimum monthly potential evapotranspira- 


ope of 
of elevation 


tion, and measures the elevation range (in logarithmic 
units) needed to substitute a small increase in 
inimum monthly potential evapotranspiration and 
maintain the same species richness in any given 
sampling unit. Along the axis of minimum monthly 
potential evapotranspiration, the MRS of logarithm of 
elevation range for minimum monthly potential 
evapotranspiration derived from the fit of the IGM2 
to our data overlaps with that derived from the global 
IGM2 (Fig 
indicates that the importance of the logarithm of 


coefficients of the This finding 
elevation range relative to minimum monthly potential 
evapotranspiration is similar in both cases. 

bove cera 
of plan 


Neotropies may result from ere in estimates of 


e comparisons t differences 
among representations ee across the 
the relative importance of three main determinants of 
plant richness. Relative to maps showing higher 
than 
topographically complex areas, our maps of predicted 


vascular plant richness in lowland areas 


richness may assign a larger role to elevation range 
and minimum monthly potential evapotranspiration 
than to annual precipitation. 


Literature Cited 


Ahn, C. H. & R. Tateishi. 1994. Development of global 30- 
minute grid potential evapotranspiration data set. J. Jap 
Soc. wa p ote Sensing 33: 12-21 

E W., J. Mutke, D. Rafiqpoor, G. Kier & H. Kreft. 

eba l centers of vas Salar plant ees Nova 
nes Leop, 342: 61-83. 

Bivand, 08. spdep: Spatial dependence: weighting 
jme. EIER CS and models, ved e ion "an Il 
cran.r ml ac- 
cessed 25 ~ 2008. 

Bjorholm, S., J. C. Svenning, F. Skov & H. Balslev. 2005. 
qu dM and 2s controls of palm (Arecaceae) 

ci ichne cross the Americas. Global Ecol. 


m J. H. 1981. Two deade of homage to Santa Rosalia: 
oward a general theory of diversity. Amer. Zool. 21 
—888. 

Brown, J. S. 1988. Patch use as an indicator of habitat 
preference, n risk, and competition. Behav. Ecol. 
Sociobiol. 2 "aia 

Burnham, K. x & D. R. Anderson. 2002. Model Selection 

and Multimodel hine A Practical Information- 

Theoretic p Springer, New York 

Burnham, . Graham. 1999. The his ed a 

nedtropical vege etation: New dev velopment and s 

n. Missouri Bot. Gard. 86: 546— 

Cade. B.S: Noon & C. H. Ae 2005. Quantile 
regression reveals ae bias and uncertainty in habitat 
models. Ecology 86: 7 

Caraco, T. 1979. Time leas and group size: A theory. 
Ecology 60: ee 

ena S. & A. S. Hadi. 1988. Regression Analysis by 

Examp "i e. $t Hoboken, New Jersey. 

Clarke, . J. Gaston. 2006. Climate, energy, and 
iva. he Roy. Soc. London, Ser. B, Biol. Sci. 273: 
2257-2 

Crawley, El 2002. Statistical Computing. Wiley, New 
York. 


Currie, D. J. & V. Paquin. 1987. Large scale d un 
din of e richness of trees. Nature 329: 326 
, G. G. telbach, H. V. Cornell, R. Field, T 2 


Guégan, B. 2 uds D. A. Kaufman, J. T. Kerr, T. 
Oberdorf O'Brie Turner. 2004. 
sts of ¿limar based o A 


Predictions phe te: 
broad-scale variation in taxonomic richness. 
7: 1121-1134. 

Darwin, C. 1862. The Voyage of the Beagle. Doubleday, 
Garden City, New Yor| 

Dormann, C. F. 


Ecol. 


ee M. B. Araújo, R. Bivand, 


analysis of e distributional data: A review. Eco- 
a 30: 609-628. 

Draper, N. R. & H. Smith. 1998. Applied Regression 
Analysis. Wiley, N rk. 

Efron, B. & R 


- Tibishizani. 1993. An Introduction to the 


—— the exaluionats rates 
hypothesis ud species-energy S Telätionships? Fun 
915. 


Ecol. 19: 899— 


Annals of the 
Missouri Botanical Garden 


Field, R., E. M. O'Brien & R. J. Whittaker. 2005. Global 
models i ae woody plant Mie from B 
Developm aluation. iae 6: 2263-22 

Fortin, M. " & M. Dale. 2005. pem Anais ET 
Ecologists. Cambridge University Press, Cambridge 

Francis, A. P. & D. J. Currie. 2003. A globally consistent 
i Am 


richness-climate relationship for angiosperms 
Naturalist 161 6. 

Frodin, D. A p: in retrospect and for the future. 
Pl. Talk 25: 


3 
Gaston, K. J. du ie patterns in biodiversity. Nature 
405: 220-227. 


Gentry, A 2. Neotropical floristic diversity: o 
geograp. connections between tral and Sou 
America, Pleistocene as fluctuations, or an Moss 
of the Andean orogeny? Ann. Missouri Bot. Gard. 69: 
551—5 

1 hanges in plant aaa diversity 4 and 


Task. 
(IGBP- eran ee ae pro- 
gramme data and information services. < 
daac.ornl.gov>, accessed 15 June 2 
G Colwell 2001. Quantifying 


biodiversity: ‘Procedunss añd pitfalls in the measurement 
an co of species richness. Ecol. Lett. 
379-391 
Graham, A. ph Neotropical plant dynamics during the 
and E ordering of evolution- 
nd special on processes. t. 22: 139-150. 
AEN B. A. 2001. Ecology's ie oe Trends Ecol. 
Evol. 16: 470. 
, R. Field, H. V. Cornell, D. J. Currie, J. F. Guégan, 
Mittelback, E. M. 
. Energy, 
water, and broad-scale geographic patterns of species 
richness. pu 84: 3105-3117 
Hijmans, R. J., S. E. Cameron, J. L. Par a, P. G. Jones & A. 
Jarvis. 2005. Very high cole ol ce 
oo for global land areas t. J. Climatol. 2: 
1965-1968. 
Hulbert, A. H. & J. P. Haskell. 2003. The effect of energy 


tl 


Bí 


e. 2005. Dispádty Bees range ma 
and survey- 3: based ss of species richness: e. 
processes and ep ud Ecol. Lett. 8: 319 
W. 007. Species richness, m and the 
scale d of range 
conservation. Proc. Natl. 
13389. 

Huston, M. 1979. A general QUEEN of species diversity. 
Amer. Naturalist 113: 8-10 

94. Biological ae The Coexistence of 
Species on Changing Landscapes. Cambridge University 
Press, Cambridge. 

Hutchinson, G. E. 1959. Homage to Santa Rosalia or why are 
there so many kinds of animals? Amer. Naturalist 93: 
145-159. 

Jetz, W., C. Rahbek & R. K. Colwell. 2004. The coincidence 
of rarity and richness and the potential signature of history 
in centers of endemism. Eco 80-1191 

Jiménez, I., T. Distler & P. Jorgensen. 2009. Estimated plant 
richness e across northwest South America provides 
similar the species-energy and spatial 
Med [E Ecography 32: 433—448. 


aps in ecology and 


Acad. U.S.A. 104: 13384- 


Kier, G., J. Mutke, E. Dinerstein, T. H. Ricketts, W. Kuper, 
H. Kreft & W. Barthlott. 2005. Global patterns of plant 
diversity and floristic knowledge. J. Biogeogr. 32: 
1107-1116. 

Koenker, R. 2005. Quantile Regression: ic Societ 
—— Cambridge University Press, Cambridge 

—— 8. Quantile D x and Related. Methods. 


accessed 25 June 2 

Kreft, H. & W. Jetz. 2007. Global patterns and E wo 
of vascular plant diversity. Proc. Natl. Acad. U.S.A. 
5925—5930 


, J. H. Somm W. Barthlot. 2006. The 
significance of geographic range size for spatial diversity 
atterns in Neotropical plants. m 29: 21-30. 
Laanisto, L., P. Urbas & M. Partel. 2008. Why does the 
unimodal species richness-productivity relationship not 
apply to woody species: A lack of clonality or a legacy of 
tropical p history? Global Ecol. Biogeogr. 17: 

320-326 


Latham, R. E. & R. E. Ricklefs. 
comparisons of tem psi m zone tree species diversity 
Pp. 294-314 in R icklefs & D. Schulter (e ditor) 
Species Diversity in Plea Communities. University 
of Chicago Press, Chicago. 

Linder, H. P. 2003. The jediaion of the Cape flora, southern 
Africa. Biol. Rev. 78: 597-638. 

Missouri Botanical Garden. ee <http://www.tropicos. 
org>, accessed 15 November 2006. 

Moser, D., S. Dullinger, T. Englisch, H. Niklfeld, C. Plutzar, 
N. Saubere H. G. Zec G 


1993. Continental 


Mutke, J. & W. Barthlott. 2005. Patterns of vascular plant 

diversity at continental to global scales. Biol. Skr. 55: 
- < 

Nelson, W., C. A. Ferreira, M. F. da Silva & M. L. 

uen 1990. Endemism centres, refugia and botanical 
collection density in Brazilian Amazonia. Nature 345: 

714-716. 

O’Brien, E. M. 1993. Climatic je in woody plant 
species richness: Towards an explanation based on an 
analysis of southern Africa's ee ds. E Biogeogr. 20: 
181-198. 

. 1998. Water-energy ice climate, and predic- 
tion of woody plant species richne; 
odel. J. m 25: 379-398. 

SR. Field & R. J. pae 2000. Climatic gradients 
in woody p E (tree and shrub) diversity: Water-energy 
dynamics, al ato and topography. Oikos 89: 

—600 


An interim general 


Palmer, M. L. 1994. Variation in species richness: Towards a 
unification of D ies Geobot. 29: 511—530 
Parnell, J. A. N., D. A. Simpso 


Muasya, A. Prajak: C. oma, S. 
Suddee & P. Wilkin. 2003. Plant collectie spread and 
densities: o Pix tial impact on B up 
studies in Thailand. J. Biogeogr. 30: 193-2 
Pausas, J. G., J. es reras, A. Ferré & X. Font. pon Coarse- 
scale plant species richness in relation to environmental 
heterogeneity. J. Veg. Sci. 14: 661—668. 
E. J. 1997. bape ack E eur in Behavioral 
anation pr tion. Harcourt. Brace 
College Publishers, Fort Wor 


Volume 96, Number 3 
2009 


Distler et al. 
Plant Richness across the Western Neotropics 


Pennisi, E. 2005. What determines species diversity? 
Science 309: 90. 
ianka, E. R. 1966. Latitudinal gradients in species 
diversity: A review of concepts. Amer. Naturalist 100: 
33-46 


Qian, H. & R. E. Ricklefs. 2000. Large-scale processes and 
the Asian bias in temperate plant species diversity. Nature 
407: 180-182. 


4. Geographical distribution. and 
a conservatism of disjunct genera of vascular 
plants in eastern Asia and eastern North America. J. Ecol. 
92: pq 

Quinn, G M. J. Keough. 2003. Experimental design and 
data a for biologists. Cambridge University Press, 
Cambridge. 

R Development Core Team. 2006. R: A Language and 

Environment for Statistical Computing. R Foundation for 

Statistical Computing, Vienna, Austria; Vienna wo 

of Economics and Busines 

ww 


<http://www.R-project.org>, accessed 25 June 

Rahbek, C. . R. Graves. 2001. Multiscale assessment of 
patterns of avian species richness. Proc. Natl. Acad. 
U.S.A. 98: 4534—4539. 

x e P. J. & K. Lum . 1980. uds of plant i 
diversity in. Califor : Relatio on ather and topog 


mer. Naturalist 116: 504-536 
Ricklefs, R E. 2004. ive fi 
palteris ii in e Eel Lett. 7: 1-15. 

—, H. Qian e. 2004. The region effect on 
pon plant species m cs between eastern Asia 
and eastern North America. Ecography 27: 129-136. 

, K. 1992. Latitudinal gradients i in species diversity: 
d 


k for global 


ham & H. P. 
Possinighar: 2006. D of different types of species 


occurrenc ta fi e in systematic conservation 
planning. Ecol. Lett. 9: 1136-1145 
Rosenzweig 7 cies Diversity in Space and 


, Cambridge. 

colo ical and evolutionary 
perspectives on the origins of tropical diversity. Pp. 163— 
173 in R. Chazdon & T. Whitmore (ed Md i p 
of Tropical Forest Biology: Classic Papers w 
taries. Pant of Chicago Press, Chicago. 

Schluter, D. & R. E. Ricklefs. 1993. Convergence and the 
regional component of species tiw ity. Pp. 230-242 in 
R. E. Ricklefs & D. Schluter (edito 
Ecological Communities: H 
Perspectives. University of Chicago Press, Chicago 


Simpson, G. G. 1964. Species density of North American 


3: 57— 
EN approach toward a rational 
classification of ciate Geogr. Rev. 38: 55-94. 
uokolainen & M. Yli-Halla. 2003. 
tem 

M. Gatehouse & C. A. Corey. 1987. Doe: 

ty? Butterflies, i 
and the British climate. Oikos 48: 195-205. 

U.S. Geological 2 2004. o cd by the G 
(1,3,30) Are second SRTM elev , reprocessed to 
GeoTIFF, Mesi 1.0. USGS, Global qd pem Facility, 
College Park, Maryland. Mens /glcf.umiacs.umd.edu/ 
data/srtm>, acce E P June 2008. 

van Rerisbüřg, BJ. t K. J. Gaston. 2002. 
Species richness, environmental correlates, and spatial 
scale: A test using South African birds. Amer. Naturalist 
159: 566— m. 

von Humboldt, 1808. Ansichten der Natur mit wis- 
sen ie pos ` Erlauterungen. J. G. Cotta, Tübingen, 


Ger 
Wallace, bs R. 1878. Tropical Nature and Other Essays. 
Macmillan, ae York and London. 
E Willis & R. Field. 2001. Scale and 
ards a e hierarchical theory of 
species diversity. y ps pe —470. 


Climatic- n 
elites an pom sity: A macroscopic perspective. 
Pp. 107-129 in T. M. Blackburn & K. J. Gaston Gee 
Macroecology: Concepts and Consequences. Cambridge 
ee ee Cambridge 
- Araújo, P jon R. 
Watson d K J. Willis. 2005. [oen hicecographiy: 
Assessment and prospect. Mur d Distri 
, D. Nogués-Bravo & M. B. o. 2006. Ge ple 
ical gradients of species due Po test of the water- 
energy conjecture of WAS et al. (2003) using European 
data P five taxa. Global Ecol. moy Es 16: 76-89. 


Willig. : ae an & R Stevens. 2003. 
cO gradients « of biodiversity: B scale, and 
synthesis. Annua col. Syst. 34: 273-309. 

Wolman, A. G. 2 H poem t on meaningfulness 


in conservation science. Conservation Biol. 20: 1626— 
1634. 
Wright, D. H. 1983. Species-energy theory: An extension of 
species area p bat 41: 496-506. 
uch einer Einteilung der Pip 
der Erde in rude Gebiete auf Grund 
Artenzahl. Repert. Spec. Nov. Regni Veg. 12: 57-83. 


ANDEAN LAND USE AND 
BIODIVERSITY: HUMANIZED 
LANDSCAPES IN A TIME OF 
CHANGE 


Kenneth R. Young! 


ABSTRACT 


Some landscapes cannot be understood without references to the kinds, degrees, and histor 

Andes represent one such 

e. uM land use paca. boih min and extensive grazin, g a 
and humidity 


to the Earth's surface. The tropical latitudes of the 


olocen 
und across virtually th 


of human-caused Dd 

place, with Or. land-use system: 
and crop- or a 

regi men Biodiversity toina i in 


or adjacent to such humanized Emm will ge been S in a eed composition, 
d cover i M ein i M AERE or deren human land uses. In 
the sha izes 


m differences 


he native a to harvest, lan 
ion, the geometries of land c 


the shifting of Andea je 


M enn fes as the Andes E 6 l 


nd, oos mosa 


inued 
character 


conservation efforts w 


ords: Andes, 


climate change, human impact, 


land use, 


ially temperature and humidity regimes along 
ation for some species. Both land-use systems and 


land use/land cover change, landscape ecology. 


Humankind has lived on the Earth, and thus altered 
parts of it, for millennia—in the case of the Old World, 
for several million years. In such long-inhabited 
landscapes, the nature of the influence of people 
through their land uses on native plants and animals, 
and on the living land cover provided by vegetation, is 
important to evaluate. For example, it matters greatly if 
deforestation began 40 years ago due to the entrance of 
colonists with cham saws, or if current land cover 
modifications are in fact affecting a landscape reforest- 
ed follow 


wi us but ancient forest clearance. 
Often, detailed historical information on past land use 


a previo 
will not be readily available, but an awareness of 
possible consequences is feasible, desirable, and useful 
for calibrating efforts for biodiversity conservation and 
for planning for resource sustainability. This article 
begins to assemble the information needed to prepare 
for the consequences of changed future environments in 
the Andean landscapes. 

The conservation of biodiversity has spatial dimen- 
sions that rai 
particular subspecies, land races, or individuals. This 


nge from global concerns to the genes of 
multiscalar quintessence results from the fact that 
biodiversity (biological diversity) includes not only 
species but the populations and genetic systems that 
underlie those species, in addition to multispecies 


oe communities, and ecosystems (Noss, 

ranklin, scale, and also an 
"m level, of — relevance to lan 
use and to inhabited places in general is that of the 
landscape, typically taken in this context to refer to an 
area of the Earth's surface tens to several hundreds of 
square kilometers in size (Turner, 2005). This article 
examines Ándean landscapes where that amount of 
surface area includes the size range that accommodates 
the dimensions of landscapes used by a particular 
owner, family, community, or organization. In addition, 
landscapes are at a human scale with which to observe, 
organize, and conceptualize the Earth's surface, as seen, 
e.g. in the prevalence of landscape photographs and 
paintings. Landscape ecology provides a conceptual 
framework, a vocabulary, and a set of paradigmatic 
expectations that can be used to classify and quantify 
the shifting of land cover types in a particular place 
(Young & Aspinall, 2006). 

This article uses a landscape approach to consider 
some of the ob of human impacts, current 
and past, in the Andes. It begins with an overview o 
how the Earth's MEA can be evaluated in terms of 
landscape mosaics (Forman, 1995), especially in 
mountainous regions, taking examples from other 
places in the world when the relevant studies have yet 


1 Department of Geography and the Environment, University of Texas, Austin, Texas 78712, U.S.A. kryoung@austin. 


utexas.e 


edu. 
doi: 10.3417/2008035 


ANN. Missouni Bor. Garp. 96: 492-507. PUBLISHED ON 28 SEPTEMBER 2009. 


Volume 96, Number 3 
2009 


Young 
Andean Land Use and Biodiversity 


to be done in the Andes. Often, a landscape ecology 
approach is conceptualized strictly in terms o 
ecological processes, but this article also situates 
the Andean landscapes within regional and geological 
contexts, with their respective spatial and temporal 
scales. It is also important to identify landscape 
legacies that continue to alter rae conditions 
for species today (cf. Foster et al., These 
legacies and current land-use pud p to 
humanizing p of the Andes, rema pr. 
and inhabi E spaces en 1968; eh 

Finally, this information is used to propose a 
conceptual model of what kinds of places and what 
st threatened with fu 
of considerable importance for 


s of species are mo ture 
extinctions—a topic 
the Andes, where species loss is occurring (Pitman et 


al., 2002; Pounds et al., 2006). 


LANDSCAPE ECOLOGY IN THE ANDES 


Land cover varies from place to place in ways that 
can be measured, mapped, and modeled. In a given 
Andean landscape, or indeed for most any terrestrial 
landscape, underlying environmental heterogeneity in 
depth, and 


chemistry can alter the species composition, density, 


soil characteristics such as moisture, 


and life forms of plants dominating local tracts of 
1998; Svenning, 2001), 
in addition to the jue soil biota and ED 


vegetation (e.g., Clark et al., 
processes (Wardle, a oa " al. 


Over slight dcin extents, ther l be SEE, 
ponds, or E outcrops that AE or eee differing 
substrates that add additional spatial heterogeneity to 
er (e.g... 
; Dwire et al., 2 


land cov ac Manríquez & Martínez-Ramos, 
isturbances, as minor as 
those caused by the ea of limbs from a large tree 
or as intense as fires that burn plants and leaf litter 
down to mineral soil, create patches of open habitat. 
Destructive storms can landscape consequences 
over regions subjected to freezing rain (Stueve et al., 
2007); volcanoes can set the biophysical features that 
then control post-disturbance vegetation and ecosys- 
tem processes (Vitousek, 2004). Those open sites can 
be colonized by plant species that are easily dispersed 
or that grow in quickly from gap edges. 

Disturbance gaps that result from the physical 
removal of land co and the biogeochemical 
alteration of perturbed sites allow a variety of fugitive, 
successional, and other species to survive in a given 
andscape (Wilcox et al., ant diseases 
similarly both move in response to eT heteroge- 
neity and also help to create and maintain heteroge- 


2007). Disturbance and 
ics thus act upon the living land 


neity (Plantegenest et al., 
successional dynami 
covers formed by native vegetation types, creating 


landscape mosaics that can be characterized in terms 
of changes along spatial gradients in underlying 
biophysical constraints, in addition to the dynamism 
imposed by plant death and regrowth (e.g., Velázquez 
& Gómez-Sal, 2007). 

Important coupled feedbacks tie in other trophic 


— 


e n act to reinforce 2 spatial 
heterogeneity or, euis to les . Herbi- 
vores, for example, may remove E Tm of 
palatable species, altering dominance or even shifting 
one vegetation type to another (e.g., Dorrough et al., 

007). Often, plant species growing in a disturbance- 
caused gap lack the plethora of herbivore defenses 
found in undisturbed sites (Coley & Barone, 1996), so 
herbivore impact is spatially heterogeneous in itself 
(e.g., Forester et al., 2007). The same would be true of 
the degree and influence of mutualistic relationships 
such as those between flowering plants and their 
hala & Jarrin-V., 2002) or 
to birds, 
monkeys, or bats and their vertebrate dispersers 
(Palacios & Rodriguez, 2001; Rodriguez-Cabal et al., 
2007). 


Predators can exert top-down infl 


pollinators (e.g., Much 


between plants with fruits attractive 


uences with 
landscape consequences (Ripple et al., 2001; Smith et 
al., 2003; Schmitz, 2008). 

Many human impacts leave conspicuous alterations 
on landscapes, with cover modified into housing, 
roads, pastures, or tree plantations. Other influences, 
however, may be more subtle, with forests still 
dominated by native species, but with the see 
dispersers and seed predators rearranged in their 
abundances by hunting and harvesting: leaving 
behind what Redford (1992) called empty forests, 
which, as a result, will have altered future forest 
—| trajectories. Bodin et a 06) showed 
that even small forest patches in sien deforested 
areas of Madagascar provided important environmen- 
tal services for local people. 

these features of spatial heterogeneity and 
landscape dynamism characterize the Andes (Young 
et al., 2007; Young, 2008). Complete explanations of 
spatial and tempora i 


l change will necessarily need t 


o 
consider or control for all of these edaphic, vegeta- 
tional, and ecological processes, in addition to human 
land use itself. Consider, for example, the landscape 
in Figure 1. The spatial heterogeneity most visible is 
imparted by the patchwork of houses and agricultural 
fields. Because the dominant land cover type includes 
the different kinds of agricultural fields, this could be 
called the landscape matrix, with additio 


and of the 


mach: explanation of the details of this mosaic 


onal patches 
n that matrix of trees settlement. Clearly, 
landscape patterning would require data regarding 

uman decisions on when and where the residents 
chose to live and decided what to grow. The social 


Annals of the 
Missouri Botanical Garden 


Figure 1. Humanized landscape in the Peruvian Andes 


sciences would need to provide many of the research 


paradigms. However, even this intensively use 


) lands 
spatial heterogeneity that c 


(humanized scape has an important underlying 


omes from minor changes 
in elevation, topography, and soils. Landscape ecology 
increasingly draws from relevant social, behavioral, 
and economie theories in providing explanations of 
patterns and processes 
Global environmental change will alter connectivity 
among different land cover patches, along environ- 
mental gradients, and within the landscapes them- 
selves. Often, species are found in areas much larger 
e, so that their overall range 
to their 
environmental tolerances, with cold-adapted species 


an a landscap 
distributions will be changed in relation 
shifted to higher elevations and moisture-requiring 
species pushed into humid refugia (Golicher et al., 

08 articular landscapes may be relatively 
buffered from change. In others, however, species 
dominance can be expected to change, leading to 
numerous feedbacks and shifts to and through trophic 
systems, with consequences for ecosystem processes. 
There are also important scalar considerations that 
arise for the Andean landscapes of interest. 


showing houses, planted trees, and agricultural fields. 


TEMPORAL SCALES IN ANDEAN LANDSCAPES 


andscape dynamism includes temporal shifts in 
E different spatial elements that make up the 
mosaic: the patches, the matrix background, and the 
long, linear features that form corridors. Seasonality 
can cause minor shifts in those patterns, with leaf fall 
and flush coordinated with the start and end, 
respectively, of dry or cold seasons. Changing seasons 
also bring in their wake different biophysical stresses 
or cause different kinds or degrees of disturbances. In 
addition, long-term monitoring often reveals subdeca- 
dal oscillations—for example, the El Niño-Southern 
Oscillation causes such repeated effects through the 
increased rain or drought that occurs, depending on 
the location. Indeed, work in Argentina has revealed 
subdecadal variations in fire that result from El Niño 
rains but that occur one to two years later, because 
s is available as fuel for 


hat is when maximum biomas 


= 


fires causing lagged effects (Grau, 2001; see also 
Kitzberger et al., 2007) 

Temporal shifts cause a variety of ecological 
responses, from physiological adjustments in individ- 
ual plants to shifts in landscape patchiness over 


Volume 96, Number 3 
2009 


Young 
Andean Land Use and Biodiversity 


decades or centuries. Over evolutionary time, as 
measured in millennia or in relation to the number of 
generations per unit time (e.g., 1/100 years for a tree 
species), population sizes and their rates of interpop- 
ulation gene flows vary with the size of patches of 
habitat, which controls the number of individuals in 
particular patches, and with their interpatch distances, 
which affect the degree of connectivity and amount of 
dispersal. As such, there is a temporal scale to 
landscape phenomena connecting metapopulation dy- 

amics and ecological and evolutionary processes. 
Thus, ies differences in a tropical forest may give 
rise not only to different vegetation types as mediated 
through herbivores ological time, but over 
evolutionary time to different plant lineages (e.g., Fine 
et al., 2004). These considerations shape how different 


landscapes interact over evolutionary time, potentially 


in ec 


ci Earth system science and evolutionary E 
the landscape ecology of the Andes. In 
addition, the study of landscape genetics is me 
promising research approaches (Manel et al. a as 
is the use of graph (Brooks, 2006; Ferrari et al., 
and circuit theory (McRae & Beier, 2007). 
Geological history reveals shifts in connectivity and 
extent of different elevational or life zones in the 
ry log t with additional 
dynamism originating with global climate changes an 
(e.g., Mont- 
. 2001). The Andes include rocks laid 


Andes over ve e spans, w 
their local and regional consequences 
gomery et al. 
down as sediment more than 100 million years ago 
(Ma). As they rose and folded or faulted, these rocks 
were transformed over wide areas into their metamor- 
phie derivatives, often e additional rock 
material forming from the co ma and lava 
forced up from the M Pacific and Caribbean 
plates. Their rise was episodic, with long periods of 
stasis followed by relatively rapid periods of height- 
ening (Garzione et al. iven this extended and 
ancient history, no dende ihera are many legacies 
among the Andean flora and fauna resulting from 
these past geographies, distinctive connectivities, and 
altered climates that have shaped speciation, promot- 
ed endemism, and caused extinctions of some lineages 
(Young et al., 2002). There are a variety of dry forest 
gro with 
intermountain valleys and on both sides of the Andes, 
which have been shown (Pennington et al., 2006) to 
reflect this colorful geological past. At least some of 


plant ups found ragmented ranges in 


the genera of Annonaceae found on both sides of the 
ndes come from dispersal and speciation events 
estimated by Pirie et al. (2006) to be 10-60 Ma 
However, uplift, volcanism, and other results of 
tectonics continue to this day in the Andes (Veblen et 
al s a result, some biodiversity patterns are 
due to much more recent and even ongoing speciation 


processes. Andean landscapes often have spatial 


characteristics that promote the separation and 
subdivision of species habitat patches. For example, 
Hughes and Eastwood (2006) show that the Andean 
members of Lupinus L. (Fabaceae) may result from 
speciation occurring at some of the fastest rates ever 
documented (2.49 to 3.72 species/1,000,000 years). 
They suggest that this alacrity was fostered by the 
repeated development of newly available islandlike 
habitats at high elevations in the Andes, caused by 
Quate 


fastest uplift of the Andes has been in the recent 


rnary glaciation cycles. In addition, much of the 
geological past, increasing the elevations of the 
highest cordilleras by more than O m compared 
ith their maximum heights in the Tertiary (Graham et 
al., 2001). Hughes and Eastwood (2006) suggest that 
resulting rapid speciation is behind high diversity in a 
variety of Andean plant genera. Muellner et al. (2005) 
begin to sort the details explaining current genetic 
structure of an Andean composite, Hypochaeris 
palustris (Phil.) Wildeman, as caused by differential 
colonization and survival in refugia. Brumfield a 
Edwards (2007) suggest ere complex routes to 
speciation and occupat ndean forests by 
Thamnophilus Vieillot prp while Torres-Car- 
vajal (2007) shows relatively recent divergence and 
speciation out of the central Andes by Stenocercus 
Duméril & Bibron lizards. Andean landscapes and 
regions contain dd of these evolutio onary pro- 
n utside the historical range 


pel variation may cause es extinctions of 


o 


particular species. 


SPATIAL DIMENSIONS OPERATING IN ÁNDEAN LANDSCAPES 


There is also a spatial and scalar aspect to consider 
as the biodiversity found in a particular landscape will 
be a subset of that found in the surrounding region, 
unless they are equal, as can happen in species-poor 
biomes. The regional context matters; adjacent areas 
with high connectivity, past or present, share more 
species wit andscape of interest. Regional 
connectivity thus influences or controls the composi- 
tion and evolutionary dynamics of the species to be 
found in a particular place 
also affect 
2007). 


For mountains, regional context comes from upland to 


oad-scale and gradient variables 
landscapes and species populations (Talley, 


lowland connections and cross-mountain range influ- 
ences Td , 1982), in addition to along- cordra 
shifts (e.g GEVE, 1988; Rull N ) 
Called eun by Seastedt et al. (2004), thee spatial 
transitions provide differential amounts of available 
e continua delimit 


habit tat an source areas. 


possible configurations of embedded landscapes; they 


Annals of the 
Missouri Botanical Garden 


influence biodiversity in spatial ways, increasing beta 
diversity and other place-to-place heterogeneity (e.g., 
Kattan et al, 2006) and affecting genetic variation 
(Ohsawa & Ide, 2008). Change in elevation in the Andes 
produces a series of interrelated shifts in atmospheric 
gases, solar radiation, wind, temperature ranges, and 
soil moisture (Kórner, 1999), not to mention in species 
composition (e.g., Terborgh, 1971; La Torre-Cuadros et 
al., 2007; Sergio & Pedrini, 2007). 
connections of highlands to contiguous low- 

lands potentially allow those lowlands to serve as 
sources of colonizing species over long time periods. 
More important, however, are the long but narrow 
connections to mountains on the same cordillera, 
while continued orogenies create potential new 
habitat for dispersing montane or alpine species. Yet 
another characteristic of mountains is great site-to-site 
change in biophysical factors in elevation as well as 
different bedrocks and aspects. As a result, most any 
place in high mountains is only a short vertical 
distance from lowlands, a short horizontal distance 
from a wetter (or drier) microsite, and centimeters 
from a slightly different altitude. Distance and 
dispersal barriers among potential habitat patches 
become critical features (e.g., Graf et al., 2007). 

Barriers that restrict species ranges can split 
e NEIN and reduce gene flo ch species 2 
be found in metapopulations esten by subpo 
Glas linked by occasional dispersal events, slices 
by local extinctions in particular habitat patches, and 
with some individuals in marginal or sink habitats. 
Some of the resulting s become conservation 
priorities (Young et al., 2002). A recent example by 
Marín et al. (2007) eus the ui aul of 
vicufia in conservation terms, with evidence of a 
into Pod 


relatively recent northern 
Peru, but with older multiple lineas across the 
southern distribution limit in Chile and Argentina. 
Anciently surviving lineages are found among several of 
the genera endemic to the Andes; some rare Andean 
species are paleoendemics. These kinds of biologic 
uniqueness warrant special conservation efforts direct- 
ed toward 


centered on the places where they may be endemic. If 


taxa without close-living sister groups or 


recent events have ae range reductions or frag- 

ions, ere is rapid speciation into 
bien like pus then numerous neoendemics 
will be present. These species might best be protected 
by conservation directed at the habitats and places of 
concern, as presumably the processes behind diversi- 
fication would also be conserved under those condi- 
tions. However, an additional source of spatial 
heterogeneity comes from Andean land-use systems of 
intensively farmed plots and gardens, with extensively 
used rangelands and croplands. 


Humanized LANDSCAPES OF THE ANDES 


mans alter land cover and hence landscape 
mosaics. Thus, part of the key to understanding future 
changes in the Andes may come from deciphering 
anthropogenic influences on these kinds of mosaics, 
aic-influencing 
ould be 


those affecting species composition and a in 


mosaic-relate enomena, and mos 
processes. Ecologic effects in this context w 


particular patches and those that alter connectivity 
among patches. Connectivity for a particular species 
is affected by the distance, presence, type of dispersal 
corridor, and type of land cover that composes the 
2006; Nascimento et al., 
2006). The type and location of remaining habitat can 


matrix (e.g., Kupfer et al., 


be a critical feature for surviving native biodiversity 
(e.g., Devictor & Jiguet, 2007), as is the presence of 
corridors (Chetkiewiez et al., 2006). Over evolutionary 
time spans, mosaic landscapes would host variable 
population sizes and connectivities, which are prime 
precursors of both extinction and speciation processes 
for native plants and animals. Humans can cause 
species extinctions in a matter of years or decades in 
particular landscapes, while their effects on selection 
pressures will be a function of the intensities or 
directions of those pressures and the generation time 
of the species of concern. There has been inadvertent 
selection pressure brought to bear on thousands of 
species due to modifications of habitat through land 
use and anthropogenic climate change (e.g, Reusch & 
Wood, 2007; Araüjo et al., 2008) 

Humanized landscapes are those that are altered in 
ways that satisfy human needs and goals. Kareiva et 
al. (2007) postulated that these kinds of modifications 
made to inhabited landscapes can be considered 
reminiscent of species domestication by humans, with 
similar benefits of maximized food production, 
increased control, and reduced risks of time periods 
without food, fibers, or fuels. Vulnerability of people 
to natural hazards in inhabited landscapes of the 
Andes can be reduced through manipulation of slopes, 
redirection of rivers, and other farming or engineering 
2009). Control of 
predators through hunting or of pests through 


approaches (Young & León, 


pesticides may be augmented by removal of required 
habitat. Some insect pests of agricultural concern are 
kept controlled by native predators surviving in edge 
and matrix habitat (Rand et al., 2006) 

While it is commonplace to lament the destructive 
influences of people on biodiversity, in cultura 
contexts such as in regard to long-inhabited land- 
scapes of the Andes, there is in fact an increase in 
some kinds of diversity—for example, as measured in 
overall habitat diversity or in terms of the genotypes 
and phenotypes of domesticated plants and animals. 


Volume 96, Number 3 
2009 


Young 
Andean Land Use and Biodiversity 


Humans worldwide have brought several hundred 
species into domestication, shaping their genotypes to 
create desired phenotypes and products. Several kinds 
of wheatlike grasses from the eastern Mediterranean 
were crossed and bred to produce wheat that feeds 
millions of people today (Salamini et al., 2002). The 
availability of rice selected for different flooding 
regimes permits farming on a wide variety of terrains 
with dry land, wetland, or controlled paddy conditions 
Car 1991; Khush, 1997). The d extant 

ong domesticated guinea pigs that originate from 
"i Andes is useful in laboratories nk. the world 
(Spotorno et al., 2006). This kind of agrobiodiversity i is 
ighly threatened, as much of it 
maintained because of agricu 


considered 


ue. 
un 


ultural practices support- 
ed by eultural norms that may change in the future 
(Brush, 2000; Young, 2002). 

Efforts to promote sustainable land-use practices in 
the future no doubt would be better informed if they 
were based on considerations of which practices have 
proved long-lived and which practices have resulted 
in undesirable consequences. Agriculture in Europe 
arose from land-use practices and associated world- 
eh dp dps in the Mediterranean—from Greek, 

n, Árab, and other civilizations and influences 
LEE 1976). Mediterranean shrubla 
bee 


n derived from previously forested E that 


nd might have 


were deforested for timber and agriculture several 


millennia ago, then maintained in scrub through 
rning and grazing for livestock (Blondel, 2006; 
He nkin et al, 2007), with bouts of soil erosio 


(Butzer, 2005). In the case of the Andes, much change 
in land cover happened several millennia ago. 
Recently, an evaluation of current spatial patterns of 
vegetation in a site in France concluded that present- 
day edaphic conditions were not sufficient to explain 
the patterns found; instead, there were legacies dating 
back to land use during the Roman conquest that were 
affecting modern vegetation patterns (Dambrine et al., 
2007). A similar finding was made by Pärtel et al. 
(2007) for explanations of the amount of floristic 
diversity in caleareous iss of Europe in relation 
to settlements 100 
botanist would i to be conversant with the 


rs BP. It is perhaps ironic that a 


archaeology of classical antiquity and the Iron Age to 
understand the causalities at play. Similarly, the 
origi rr nd-use practices in Papua New 
Guinea can be traced back to 11,000-7000 years BP, 
when forests and wetlands were converted for croplands 
(Hope & Tulip, 1994). Thi 
settled more than 50,000 years ago with fire becoming a 


e Australian outback was 


tool of landscape change used by people, although 
details on the relative importance of the anthropogenic 
influence on the loss of native species are sti 


controversial (Bowman, 199 


In the Andes, Dillehay et al. (2007) showed that 
people were using field agriculture by 9200 years BP. 
By using dated plant remains from archaeological 
sites in northern Peru at sites at 500 m elevation, 
Dillehay et al. also showed cultivation of squash 
earlier than 9000 years BP, the peanut by at least ca. 
7800 years BP, and cotton by ca. 5500 years BP. The 
Andean landscapes began to be altered in terms of 
land cover, accompanied by the domestication of wild 
species and manipulation of land races. There are 
archaeological remains of fields in that study area 
with furrows and evidence of irrigation through canals, 
social ee of interdependent 
. Similar kinds of 


scapes in southern 


suggesting 
households by those time per 
ta are sai for land 
(Perry et al, 2006) and generally in the Ei 
Andes B 2001). Paduano et al. (2003} found 
many w lant species in the pollen record of Lake 
g 100 years BP, 
(Chepstow-Lusty & 


yp 
Titicaca pires begi 
while Chepstow-Lusty et al. 
Winfield, 2000; Chepstow-Lusty et al., 2003) report 
abundant Ambrosia L. and Chenopodiaceae pollen 
from 2200 BCE to 750 BCE in the Cusco area, along 


with evidence of ancient agroforestry. 


nning ca. 


Because there are multiple possible combinations 
of land use and land cover, often their study through 
time is referred to as research on land-use/land-cover 
change (LULCC) (Lambin et al., 2001, 2003; Young, 
2005). An agent of landscape change can be a person, 
household, town, or natural agent such as an insect 
pest or a windstorm. A driver of landscape change 
may be the economie system or financial event that 
motivates the people involved, or perhaps a shifting 
air circulation system that alters the trajectory of a 
storm that then 
landscape mosaic, 


results in windthrow of 
classifiable into land-use/land- 
cover units and spatially delineated into patches and 
corridors, can be tied ie to theories and 
data, in this particular n economies and on 
atmospheric physics. um i e science draws on a 


wide range of social, biological, and physical sciences 


> 


or paradigms and explanations (Gutman et al., 
Kintz et al. (2006) used this ap 
quantify change in landscape mosaics of northern 


proach to map Ss 
Peru over a 13-year period. Most change was related 
to altered land use as local population size double 

with in-migration. Another component of change was 
the national park management, and possibly some 
vegetation shifts were caused directly by climate 
change. Postigo et al. 08) recently also used 
satellite-derived mapping to locate how climate 
change in south-central Peru was altering land cover 
that is useful for high-elevation pastoralists. However, 
some land-use changes were motivated by prices for 
products from their alpaca and sheep, while still 


Annals of the 
Missouri Botanical Garden 


others are the result of partial out-migration of many 
members of the households and families. 


ANDEAN LANDSCAPE LEGACIES 


Legacies of a long history of changes in agriculture 
and land cover in the Andes can be found in many 
landscape characteristics (Fig. 1). Soils, for example, 
can show topographic and micromorphologic features 
caused by past cultivation and terracing (Sandor & 
Eash, 1995; Kemp et al. pu D. dated features 

Rigsby et al. (2003) reveal human 
occupation of flat topographic ieee beginning 


in a study by 


10,000 years BP and continuing in the evolving 


landscapes of southern on ent terrain 
features for the next 10 millennia. Without terracing 
or following its abandonment, cultivation often results 
in increased soil erosion (Inbar & Llerena, 2000). 
Fire is another source of landscape dynamism. Fire 
is a natural agent of change in the Andes; however, it 
is often provoked by lightning strikes, so ignition 
caused by nature may frequently be limited to the 
rainy season. As a result, much burning and extensive 
fires are instead due to human agency, which began 
several millennia ago. People arrived in South 
America in the Early Holocene carrying stone 
ing fire. Fire 


weapons, accompanied by dogs, and usin: 


B 
as recorded in carbonized wood particles in dated lake 
sediments can be documented back as far as records 
a in Lake Titicaca to 27,500 years BP (Paduano et 
al., . However, there is an obvious human 
D on the abundance of charcoal in some Andean 
lakes. For example, Bush et al. (2005) showed a time 
period ca. 3500 years BP when charcoal dramatically 
increased. Because they were able to separate out the 
principle climate-caused shifts caused by a multi- 
decadal precipitation. oscillation id 2 wavelet 


analysis, other changes, is charcoal 
increase, would o m up in 
causality 


Millennia of Andean agriculture were interrupted 
by the impact of contact and conquest by Europeans 
500 years BP, followed by abrupt declines in human 
population numbers and then the incorporation of new 
domesticated phos Se and the development of novel 
land u ral populations slowly recovered (Gade, 
1999; ae 2003). Colonialism has cont 
consequences for landscapes (Sluyter, 2001). Because 


tinued 


the ry were concerned with control of their 
colonies, they rted a n land use as 
pd dm households a into towns and 
centered on a plaza with a church and administrative 
These 


concentrated houses into small urban areas 


buildin S. settlements known as “reduc- 
8 > 
ciones,” 


and were accompanied by altered economic goals for 


the people, often involving the provision of taxes, 
products, or labor to regional or national governments. 
h land use and land cover change 

Today, Andean landscapes not on the wet external 
flanks of the Andes are dominated by shrubs, not by 
trees (Fig. 1; Young, 1998). Thus, many Andean 
landscapes have a shrubland matrix, with remaining 
forest patches accompanied by other patches contain- 
ing planted and fallowed fields. Species in those 
patches will have their local populations shaped by 
the area, number, and separation of the respective 
patches. Indeed, in many Andean landscapes, dom- 
inant trees are not native species, but are planted 
eucalypts or pines, which may eliminate some native 
species due to missing mutualists. Structural features 
of these humanized landscapes include rock walls, 
terraces, forest patches, trails, fallow fields, md 
and house gardens. These land cover types delim 
patches of varying value to people for the extraction " 
firewood and medicinal plants and the production of 
field crops, fruit trees, and livestock. Fire continues to 

e used to reduce woody cover and foment sprouting 
of palatable plants for grazing over extensive areas, 
even those far from settlements. Andean land use is 
typically intensive locally around houses and exten- 
sive over hillsides and on mountain peaks. 

Surviving native species of plants of these land- 
scapes often have small, well-dispersed seeds; can 
resprout following apical damage; may have thorns or 
are either resilient or inc 
, 1992; Young, 1998). Examples given 
by Young and Keating (2001) for a site in northern 


other protection; and onspic- 


uous (Laegaar 


Ecuador include Barnadesia arborea Kunth, Brachyo- 
tum ledifolium (Desr.) Triana, Coriaria ruscifolia L., 
Escallonia myrtilloides L. f., Gaultheria foliolosa 
Benth., Maytenus verticillata (Ruiz & Pav.) DC., 
Monnina obtusifolia Kunth, Piper barbatum Kunth, 
and Vallea stipularis L. f. Surviving wildlife species 
are tolerant of human presence, unattractive as game, 
and unthreatening to human interests (Young, 1997). 
Most Andean species of humanized landscapes thrive 
in open, edge, or successional habitats 


FUTURE GEOGRAPHIES OF THE ANDES: TOWARD A 
CONCEPTUAL MODEL 


Species persistence in the inhabited and utilized 
landscapes of the Andes will depend on species traits, 
including their sensitivity to the effects of future 
change in the biophysical factors that control their 
distribution and abundance (Scholze et al., 
Tewksbury et al., 2008). The effects on species also 
will be MAC by and e through changes 


5 


in land cover, so LULC l be critical. Given 


current global iei pud. and visible 


Volume 96, Number 3 
2009 


Young 
Andean Land Use and Biodiversity 


Biophysical factors: 
f Temperature 

f Carbon dioxide 

1 Precipitation 

f Height cloud bank 


Vegetation and land cover: 
Species distributions 

Species abundances 

Ecosystem productivity, fluxes 


Perceptions of climate change 
Capacity to respond individually 
Capacity to respond collectively 


Land use and livelihood practices: 


À conceptualization of pro 


osed interactions and feedbacks ee to connect climate change in the 


igure 2. 
Andes, as mediated through biophysical ds to shifting land use and veget 


consequences in the high Andes in terms of reduced 
permanent ice and altered land cover (Ramírez et al., 
2001; Vuille et al., 2003; 2006; 
Young, 2007), plus heed natural hazards cee 
2005; Vilímek et al., d 


mosaics will continue to shift P least over the next 


Thompson et al., 

n landscape 
couple of decades. In this section, I develop a 
conceptual model of how landscapes are being 
affected and of how species distributions may be 
changed as a result. 

Figure 2 connects the major biophysical factors that 
are changing in tropical latitudes to land cover and land 
use. Not only are temperatures increasing, driven by 
more ise cme such as COs, but many tropical 
areas are expec o have less one (Cook 
Vizy, ). In addition, there will be ard shifts in 
altitudes where moisture condenses, pon frequently 
forms, and cloud forests are found (Foster, 2001). One 
important way that these environmental controls affect 
plants is through altered soil properties, such as soil 
moisture and nutrient availability. 

However, the effects on land cover will not be 
unidirectional given the presence of land use, which 
can respond quickly throug 


reactions to cues in the landscape (Young & Lipton, 


h farmers’ actions and 


2006). Thus, Figure 2 includes arrows suggesting 
links and feedbacks. Shifts in th and 


composition alter grazing systems and cropping 


plant grow 


patterns. Land use will change, given the capacity to 
shift the locations of pastures and field, the knowledge 
and resources to modify planted crops, and collective 


or household capital that permits new livelihood 
strategies (Mayer, 02) and external networks 
(Bebbington, 2000). For the Andes, this means pe 
agriculture and pastoralism must change, with grazing 
activities moved to higher elevations on newly formed 
wetlands following glacial retreat and crop choice 
modified on thousands of agricultural fields; it also 
implies fewer water resources for the settlements and 
2006). 


Predicting shifts of wild species will require ome: 


cities located downslope (Bradley et al., 


on on the bioclimatic envelopes occupied and the 


+ 
Be 


additional degree of sensitivity to direct and indirect 
influences of land use. 

e complexity of the more traditional farming 
shoul rgins for rapid 


systems provide amp 
fa 


em 
adaptation for the farmers themselves, with house- 
holds often planting multiple crops and land races on 
12 or more different fields, locating them at different 
elevations and on different exposures (Brush, 1976; 
Zimmerer, 1996, 1999; Young, 2008). Of course, this 
resilience assumes that (1) the appropriate knowledge 
exists and can be transmitted, (2) the needed genetic 
diversity is available, (3) alternative useful plant or 
animal stock is on hand as needed, and (4) top-down 
restrictions on or incentives for certain land uses and 
production goals by regional or national governments 
do not constrain household and community adapta- 
tion. 

Change is constant, and much of the biological and 
cultural diversity to be found in the inhabited and 
uninhabited landscapes of the Andes is due to that 


Annals of the 
Missouri Botanical Garden 


Change (warmer, drier) 


Cold (wet) adapted species 


Priorities 


Peaks; Moist enclaves 


Matrix 


Figure 3. A conceptual d of spatial Ioplcanons of future climate change i in the tropical Andes for native species. 
ndt 


The species’ ranges will te 
humid conditions 


species with sm; ould need t 


especially for the species adapted either to relatively cold or relatively 
- The converse will take place for the species adapted to relatively warm and/or dry habitats, with their 
potential distributions both larger and less fragmented. Given 


these predictions, biodiversity conservation priorities for the 


maller ranges w 


high elevations and moist microsites, while the species with 


o 
larger potential ranges pe instead need gonna strategies focused on Andean landscape matrix habitats 


change, whether over centuries, millennia, or millions 
of years. Given that future change will alter biota, 
vegetation, and land-use systems relatively quickly 
E , and hence mosaic-related processes, there 

ht be ways to prediet and prepare for conse- 
pai For land use, this would require LULCC 
measurements that locate and calculate the new 
ng on landsc , while 


biop Lora constraints actin. ape 
ing 


also conduc evaluations capable of elueidane the 
external peas acting on human decision making, 
including market signals, land tenure, and alterna- 
tives to rural livelihoods. It can be hypothesized that 


& 


in sites with reduced precipitation or diminishe 
glaciers, less water will require shifts by farmers from 
rain-fed agricultural systems to more elaborate 
irrigation schemes, or perhaps to more dependence 
on livestock such as goats and sheep. 

Andean landscape geometries help set contexts and 
limit possibilities for the populations and metapopu- 
lations of native species, thus allowing some degree of 
predictability (e.g., Seabloom et al., 2002; Starzomski 

Srivastava, 20 mall habitat patches have 
significant edge effect, while isolated patches have 


reduced gene flow and increased rates of local 
population extinctions. The amount of core forest 
habitat and the shape of forest patches change as 
an populations increase use intensifies 

6 


(Kin et al, 


tat change in relation to 
of the 


type, 
intrinsic population increases cies of 
concern are parameters involved in the likelihood of 


local extinctions (Schrott et al., 


50 years but lost considerable canopy density, thus 
potentially negatively affecting a subset of the biota 
sensitive to more open conditions 

armer and drier conditions it is likely that 
native species needing relatively wet condi- 
tions will become less common, with nied ranges 
(Fig. 3). Note that the ranges not only become smaller 
overall, but in the rough topography of the Andes, they 
will often become more fragmented, with species 
found in smaller areas and many with local popula- 


Volume 96, Number 3 Young 501 
2009 Andean Land Use and Biodiversity 
tions in small, isolated habitat patches. Species would need to be designed to permit species to occupy 


ome extinct in 
, 2005), 


while those dependent on humid microsites (Killeen et 


restricted to high elevations may bec 
at least part of their ranges (e.g., Thuiller et al. 


al., 2007) may similarly be affected if those sites no 
longer exist. Biodiversity conservation priorities for 
these kinds of species should include the disappear- 
ing and unique habitats, as well as places that would 
permit conservation along long environmental gradi- 
ents. These a 


preferably PUR ANS to habitat along an elevational 


e here proposed to be mountain peaks, 


gradient, and isolated patches of cloud forest and 
other very humid habitats, again preferably embedded 
within larger areas designated for species conserva- 
tion efforts (Fig. 3). 

Intervention strategies for these species would need 
to be designed to cope with more fragmented habitats 
(Fig. 3), smaller populations, decreased gene flow 


and increased sink habitat incapable of maintaining 
local populations (Pulliam, 1988; Hanski & Ovaskai- 
nen, 2003; Ditto & Frey, 2007, Pompe et al., 2008). 
The los 
would increase total amounts of extinctions 

al, 2004; Rezende et al, 2007). As supporting 
evidence, Golicher et al. (2008) found that the future 
distributions of the montane forest flora in Chiapas, 


s of species with mutualistic relationships 
Koh et 


Mexico, could be evaluated in relation to two groups 
of species: those adapted to moist and those to cool 
climatic conditions. Pounds et al. (2006) assembled 
data suggesting that climate-caused change is the 
reason behind dramatie extinctions already taking 
place among Andean amphibians, which are exacer- 
bated by habitat fragmentation. 

ions of ranges of other 


Large shifts and expans 


species are to be expected (Fig. 3). Given species' 
preexisting adaptations to us biophysical condi- 
tions, their distributions would often become not only 
larger, but less fragmented, and previously isolated 
ies that 


need or are tolerant of warmer conditions would be 


populations may be brought into contact. Specie 
likely to be found in more sites and/or in sites 
occurring at higher elevations in the Andes. Species 
adapted to the dry exposed habitats of much of the 
inner Andes—the intermountain valleys—might also 
expand their distributions. Thus, some species will be 
winners in a sense, with larger potential distributions. 

Biodiversity conservation strategies will still be 
needed, however, even for species predicted to 
increase their range size (Fig. 3). For example, the 
rare or desirable species with adaptations that would 
give them mI larger ranges are here proposed 
to need conservation efforts that include the installa- 
tion of conservation corridors permitting dispersal and 
the management of landscape matrices such that those 
species can continue to exist. Intervention strategies 


new locales through dispersal, and perhaps translo- 
cation, and to minimize noxious effects of land use in 
the dominant part of the landscape—the matrix. 
Donald and Evans (2006) recently evaluated ways that 
ecological restoration of matrix would help improve 
some ecosystem functions in agricultural landscapes; 
existing approaches to Andean restoration efforts 
(Sarmiento, 2003) would be reinvigorated with such 
goals. 


CONCLUSIONS 
Climate change will likely produce novel habitats 


produce large expanses of land 
succession on substrates that were previously Eon 
glaciers in high mountains (Thompson et al., 

Change in the inhabited landscapes of the Andés a : 
ubset of the 
and nuisance species E and useful or beneficial 


acting on a su original biota, with dangerous 
species encouraged. Some of those species may be rare 
or absent due to historical events, not because of current 
biophysical conditions. Other species may be more 
common than expected because of their resilience to 
continued human impacts, such as cutting, trampling, or 
hunting. Still others may be present because of 
introduction with human migration or trade. In addition, 
the conceptual model in Figure 3 implies that different 
spatial outcomes are expected, and thus different 
strategies are needed for species that are predicted to 

lose potential distribution because they are temperature 
sensitive, or for species that are limited in distribution 
and abundan 


. Other species may increase 
their potential distributions in terms of their bioclimatic 


ce by moisture 
constraints (Fig. 3), but may still be of conservation 
concern because land-use and land-cover changes limit 
their actual habitat, especially in the humanized 
landscapes 

Conservation goals that include restoration of 
populations of original species and natural land cover 
types will need to consider what time period is to be 
recreated and whether that goal is desirable and 
attainable (Hopfensperger et al., 2007). In Europe, the 
use of heavy-handed management tools, such as 
deliberate Sd ps or frequent burning and 
mowing, is used to favor species restricted to open 
habitats or E. the maintenance of early succes- 
sional vegetation is a conservation objective (e.g., 
Moro & Gadal, 2007). In the northwestern United 
States, forest management is being used to alter 
disturbance regimes in ways meant to maintain native 
biodiversity 


the central United States (Van Dyke et al., 2007). The 


Annals of the 


Missouri Botanical Garden 


Figure 4. High Andean landscape in en on the Cordillera Negra of north Peru, with the glaciated Cordillera 
Blanca dominating the background to the eas 


equivalent management options for Andean land- 
scapes need to be designed and calibrated. 

Over regional environmental gradients of humidity, 
elevation, and latitude, these fragmented landscapes 
are linked or separated by habitat and topographic 
connectivity (Fischer & Lindenmeyer, 2006). Figure 4: 


It shows in the foreground a 
stand of 5-m jc Puya raimondii Harms 
growing on the Cordillera Negra, a mountain range 
with no permanent ice and with land-use systems 
dependent on seasonal rains. Similarly, mountain 
ranges elsewhere that are losing their glaciers are also 
becoming less productive for farmers who must deal 
with limited water (Young & Lipton, 2006). 

In the background of Figure 4, the Cordillera Blanca 
rises 2000 m higher, is capped by ice caps, and is often 
shrouded in clouds. People living on slopes below this 
range have access to water from glacial lakes and from 
ice melt; most use irrigation to extend growing seasons, 
increase agricultural productivity, and produce crops 
for distant markets. Changing land use will aet in 
complex ways on the shifting habitats that contain wild 
native species on these two mountain cordilleras. For 


example, the native plants of the Cordillera Negra will 
be affected by both warmer temperatures and drier 
conditions, and livestock grazing will likely become 
even more extensive along that mountain range. In turn, 
native species of the Cordillera Blanca will have their 
respective potential distributions changed by climate 
shifts, but they will also be relocated inside the higher 
elevations that are within the boundaries of Huascaran 
National Park. The presence of a conservation- 
protected area makes additional conservation strategies 
possible, although there are likely to be increases in 
pressure on the park from land use extending into the 
park from adjacent rural communities 

Andean pastoralists often will quickly shift live- 
stock to higher elevations as ice retreats (Young & 
Lipton, 2006; Postigo et al., 2008). Every planting 


u 
certain kinds of adaptation by individual honseholds 
or on lands managed communally will be quick, 
taking place in less than one year, unless other 


socioeconomic, legal, or political factors impede them 


3 


mall-scale farmers studied in China by 
Hageback et al. (2005} showed considerable potential 


Volume 96, Number 3 
2009 


Young 
Andean Land Use and Biodiversity 


adaptation to climate change, including the diversi- 
fication of livelihoods. Bates (2002) found that another 
common response is out-migration by people from 
areas under environmental stress. Relevant studies in 
the Andes are yet to be done. Morton (2007) worried 
that complex subsistence agriculture in general is not 
sufficiently studied in relation to climate change, and 
future influences on pasture species and on crops 
other than the major commercial species have not 
2 research foci (Tubiello et al., 2007; Lobell et al., 

olicies need to promote HO Ne (Howden 
2007), although top-down efforts may not be 
successful in developing countries. 


et F 


Additional m of Andean agrobiodiversity may 
also require monitoring and intervention. The geno- 
types of iE used plants are threatened 
because of market forces, government subsidies for 
only a subset of the varieties, and loss of traditional 
knowledge dr 2000). Climate - may further 

reduce the use of land races no longer as productive 
and yet a gene a puo useful 
under other conditions. Predicting which species or 
varieties will prosper under future biophysical 
conditions would allow institutions to foster their use 
and maintenance among farmers. 

Humanized Andean landscapes are used in ways 


The 


species in these 


that favor some wild species and not others. 
continued persistence of native 
landscapes may depend as much on the patterns and 
dynamies of lan 
characteristics themselves (Danielson, 1991; Ovas- 


kainen et al, 2002). Knowledge of curre 


patterns, land-use change, and historical legacies will 


nt mosaic 


permit better understanding of landscape processes 
and possible conservation goals (Lunt & Spooner, 
2005). For example, the 

fields will have value for carbon sequestration 
2007) giving additional 


g with Pe 
h 


carbon found in Andean 


payments (Antle et al., 
conservation reasons for worki 

landscapes. Global environmental change will shift 
some of the environmental continua e agricul- 
ture and biota, especially temperature and humidity 


s will need to respond 
spatially to these shifts in ihe ae at landscape and 
regional scales, and they should also be cognizant of 
the long temporal scales that gave rise to valued 
diversity and the changes that can cause its loss over 
relatively short periods. 


Literature Cited 


Antle, J. M., J. J. Stoorvogel & R. O. Valdivia. 2007. 
Assessing the economic impacts of agricultural carbon 
ee am nos and agroforestry in the Peruvian 


Andes. Agric. Eco-syst. Environm. 122: 435-445 


Araújo, M. B., D. Nogués-Bravo, I. Reginster, M. Rounsevell 

& R. J. Whitaker. 2008. Exposure of European 
AR to changes in human-induced pressures. 
Policy 11 


Bates, D 002. tomen refugees? e 
human migrations cause environmental chan 
Populat. Environm. 23: 465-477 

Bebbington, A. 2000. Reencountering development: Liveli- 
h 


A millennial. study of humans and ecological systems 
during the historic to Hu 
odin, O., M. Tengo, A. Norman, J. Lundberg & T. Elmqvist. 


2006. The value of small size: Loss of forest patches and 
ee vo in southern Madagascar. Ecol. 
Applic. ol. 
Bowman, a m m S. 1998. TI 
burning on the Australian biota. ‘New "o 140: DOMUM 
lle, H. F. Diaz & W. Vergara. 2006. 


Threats to water supplies in > cd 


Tu 


Brooks, C. P. 2006. Quantifying NE substructure: 
pog. thé graph-theoretic approach. Ecology 87: 
—872 


pes R. T. & S. V. Edwards. 2007. Evolution into and 

of the Andes: A Bayesian analysis of historical 

dcus db in Thamnophilus M Evolution 61: 
346-367. 

Brush, S. B. b ner use of an Andean ecosystem. 
Human Ecol. 4: —166. 

(editor). EU poe in the Field: On-Farm Conserva- 
tion of Crop Diversity. Lewis Publishers, Boca Raton. 

Bush, M. B., B. C. S. Hansen, D. T. Rodbell, G. O. Seltzer, » 

Young, B. León, M. B. Abbott, M. R. Silman & W 
Gosling. 2005. A 17,000- REA hm of Andean ds 
and vegetation change from L a de Chochos, Peru. J. 
Quatern. T 20: 703-714. 

2005. Environmental history in the Mediter- 
ranean world: Cross- -disciplinary investigation of cause- 
and-effect for degradation and soil erosion. J. Archaeol. 
Sci. 32: 1773-1800. 

Carey, M 5. Living and dying with glaciers: People’s 
historical vulnerability to P and outburst floods 

: 122-134. 

Carney, J. 1991 soil M water management in 
Se ae ee rice farming systems. Agric. Human Values 
8: 37-48. 


Chepstow-Lusty, A. & M. Winfield. 2000. Inca agroforestry: 
Lessons a the past. a 29: 322-328. 
. Frogley, B. S. Bauer, M. B. Bush & A. 
Eom ma 2003. A liis olaa record of arid 
events from the Cuzco region, Peru. J. Quatern. Sci. 18: 
-502 


Chetkiewiez, C.-L. . C. St. Clair & M. S. Boyce. 2006. 
Corrid i 


& J. M. Read. 1998. Edaphic 

variation and the ers eee of tree species in 
a a rain forest. J. E 2 

Coley, P. 96. Herbivory and plant 

defenses in epica forests. Aantal Rev. Ecol. Syst. 27: 
305-335. 

Cook, K. H. & E. K. Vizy. 2008. Effects of ME e 
century climate Dr M the Amazon rain forest. J. 
Climate 21: 542-560 


Annals of the 
Missouri Botanical Garden 


Dambrine, E., J.-L. Dupouey, L. Laüt, L. Humbert, M. 
Thinon, T. Beaufils & H. Richard. 2007. Present es 
Biodivar rsity patterns in France re Ps ted to former Rom 
agriculture. a 88: 1430-14. 

Danielson, B. J. 199 oid in a landscape: The 

i e interactions 

20. 


apes of Native 

Amaronta ad the Andes. Oxford University Press, New 

York. 

Devictor, V. & F. Jiguet. 2007. Community richness and 
stability in agricultural landscapes: The importance of 
surrounding habitats. Agric. Eco-syst. Environm. 120: 

84. 


Dillehay, T. D., J. Rossen, T. C. Andres & D. E. Williams. 
007. B adoption of p squash, and cotton 
in northern Pe S nis 316: 1890-1893. 


. K. Frey. 2007. d F ey ae 
i mammals: 
poe warming 


. Ash, S. Bruce & S. McIntyre. 2007. 
From plant neighbourhood to landscape scales: How 
grazing modifies native and exotic plant species richness 
in grassland. E Ecol. 191: 185-198. 

Dwire, K. A., . Kauffman, E. N. J. Brookshire & J. E 
Baham. 2004. roe biomass and species composition 
along an environmental gradient in montane riparian 


ecosystems in Pen America. J. Ecol. 67: 401—416. 
R. K. Didham. 2007. The effect of fragment 
shape and species' sensitivity to habitat edges on animal 


Ecol. 22: 1315-132 
Fine, P. V. A., D. Coley. 2004. Herbivores 

promote habitat — by trees in Amazonian 

forests. Pix 30: 663-665. 

B. Lindenmayer. 2006. Beyond fragmenta- 
ontinuum model for fauna research and 
RN in human-modified landscapes. Oikos 112: 

473-480. 

m J. D., A. R. Ives, M. G. Turner, D. P. Anderson, D. 
Fortin, H. L Be er, D. W. Smith & M. E M 2007. 
e pu models link elk mov patterns to 
landscape ns in Wile ei "National Park. 

Ecol. Monogr. 7 9: 

Forman, R. T a 1995. Land Mosaics: The Ecology of 
Landscapes and Regions. Cambridge University Press, 
Cambridge 

Foster, D., F. Swanson, J. Aber, I. Burke, N. Brokaw, D. 

E A. Knapp. 2003. The impanands of land-use 

legacies to ecology and conservation. BioScience 53: 

71-8 


Tilman 


Foster, P. 2001. The potential negative impacts of global 
climate change on tropical montane cloud forests. Earth- 
Sci. Rev. 55: dt us 

Franklin, J. F. 1993. Preserving uw rr bo 
ecosystems, or aud. Ecol. Applic. 3: 202-205. 
ade, . Nature and e in de Andes. 

University of Wisconsin Press, Madis 


Garzione, C. N., G. D. Hoke, J. C. Libarkin, S. Withers, 
MacFadden, i Eiler, P. Ghosh & A. Mulch. 2008. Rise e 
the Andes. Science 320: 1304-1307. 
Gentry, A. H. 1982. Ta a a 
d South America, Plei 
ee of the ees 


ae of 


Ecol. Brom. 17: 262-273. 


Graham 
dera in n Neotropical paleana tany. XV. A Mio- Pliocene 
palynoflora from the Eastern Cordillera, Bolivia: Implica- 
tions for the nas histo of the central Andes. Amer. J 
Bot. 88: 1545-1 

Grau, H. R. 2001. Re nh scale spatial patterns of fire in 

relation to rainfall gradients in sub-tropica PUN 
W Argentina. Clobal Ecol. Biogeogr. 10: 133-14. 

Graves, G. R. 1988. graphic range m its 

of Andea 


Linearity of geo 

opulation structure 

Gutman, G., A. Janetos, E Justice, E. Moran, J. Mustard, R. 

. L. Turner II (editors). 2004. 

Observing, Monitoring, and 

Understanding Trajectories of Change on the Earth's 
Surface. Kluwer, New York. 


Hageback, J., J. Sundberg, M. Ostwald, D. Chen, X. Yun & P. 


Science: 


an shed, 
farmers’ adaptation. Climatie Mrd 12: 189-212 
Hanski, I. € o. Ovaskainen. 2003. Metapopulation theory for 
CE E Theor. na Biol. 64: 119-127. 
Henkin, Hadar & I. Noy-Meir. 2007. Human-scale 
sca ae induced by grazing in a Mediterra- 
oodland landscape. Lan do Ecol. 22: 577-587. 
Hope, G E rE J. Tulip. 1994. A long vegetation history from 
lowland Irian-Jaya, Indonesia. Paleogeogr. Palaeoclimatol. 
Palaeoecol. 109: 385—398. 
Hopfensperger, M. Engelhardt & S. W. Seagle. 
2007. Ecological feasibility studies in "uo eee decision 


change. Proc. Natl. Acad: s U.S.A. 104: 19691-19696. 
Hughes, C. & R. Eastwood. 2006. Island radiation 
continental sale Exceptional rates of plant diversification 

after uplift of the Andes. Proc. Natl. Acad. Sci. U.S.A. 103 

10334-10339 
Ibarra-Manríquez, M. Martínez-Ramos. 2002. Land- 

scape variation of liana communities in a al rain 

forest. PL. | 160: 91-112. 
A. Llerena. 2000. Erosion i de in high 
ltural terraces in Peru. Mountain Res. 
Developm. 20: 72-79. 
. M. Ramsay. 2007. Changes i in we pd 
Polylepis forest cover and quality in the a de 

Vilcanota, Peru, 1956-2005. Biol. concern pes buds 
Kareiva, P., S. Watts, R. McDonal . Boucher. 2007. 

Domesticated nature: Shaping puras and ecosystems 

fo w 


elfare. Science 316: 1866-1869. 


on a 


r human 


Volume 96, Number 3 Young 505 

2009 Andean Land Use and Biodiversity 

Kattan, G. H., P. Franco, C. A. Saavedra-Rodriguez, C. Mayer, E. 2002. The Articulated Peasant: Household 
Valderrama, V. Rojas, D. Osorio Martínez. 2006. Economies in the Andes. Westview Press, Boulder. 
Spatial components of bird diversity in the Andes of | McRae, B. H. & P. Beier. 2007. Circuit theory predicts gene 


uthem Peruvian 
ndes. wr Int. 158: 13-22. 
S. 1997. Origin, dispersal, 
variation T rice. Pl. Molec. Biol. 35: 25 
un IN F i p E S ri P. M. ADU 
spot 


cultivation. and 
—34. 


. Dry and wi the And 
|. El 2 p 1357- 
Kintz, D. B., Young & 4 5 Crews-Meyer. 2006. 


aplico of land use/land cover change in the buffer 
zone of a national park in the tropical Andes. Environm. 


n, E. K. Heyerdahl, T. W. Swe tnam 


North America. Proc. Natl. Acad. Sci. U.S.A. 104: 543—548. 

Koh, L. P., R. R. Dunn, N. S. Sodhi, R. E Colwell, H. C. 

Proctor & V. S. Smith. 2004. Speci tions and the 
wr po crisis. Science 305: 

Alpine Plant Life: al Plant Ecology 

aa Berlin. 
Franklin. 2006. Not 
seeing the ocean for the islands: E mediating influence 


of matrix-based processes on forest fragmentation effects. 
Global Ecol. Bi p 15: 8-20 
Laegaard, S . Influence of Hg in the grass páramo 


Lambin, E. F., H. J. Geist & E. Lepers. 2003. Dynamics of 
land-use and land-cover change in tropical regions. Ann. 
Rev. Environm. Res. 28: 205-241. 

, B. L. r, H. J. Geist, S. B. Agbola, A. Angelsen, 

J. W. Bruce, o. T Cotes: R. Dirzo, G. Fischer, C. Folke, 


E and land-cover : “Moving beyond the 
269. 


Herrando-Pérez & K. R. Youn 


7. Diversity and structural patterns for tropical 
montane and premontane forests of central Peru, with an 
assessment of the use of higher iodivers. 
& Conservation w 2965-2988. 

Lobell, D. B., M. B. Burke, C. Tebaldi, M. D. Mastrandrea, 
w P. Falcon & R. L. : Naylor: 2008. per climate 
Science 

319: m 
Lunt, I. D. & P. C. Spooner. 2005. Using lee ecology to 


understand patterns of co in n 
tural landscapes. J. Biogeogr. 32: 

Manel, S., . Schwartz, G. Luikart "i P. EN 2003. 
Landscape genetics: Combining landscape ecology and 
population E rm Trends Ecol. Evol. 18: 189-197. 


ted agricul- 


Marín, J. C., C. S. Casey, M. Tm K. Yaya, D. Hoces, J. 
Olazabal, R. Rosadio, J. R , A. Spotorno, M. W. 
Bruford & J. C. Wheeler. ed "Mitochondrial phylogeo- 


graphy and demographic history of the Vicuña: Implica- 
tions for conservation. Heredity 99: 70-80. 


flow in plant and animal Ee Proc. Natl. Acad 
Sci. U.S.A. 104: 19885— pus 
n D. R., G. Bal &S. D. Willett. 2001. a 
ectonics, and the m n. of the Andes. Geology 
—582. 
Moro, D. & S. Gadal. 2007. Benefits of habitat. restoration to 
small mammal diversity and abundance a pastoral 


in mid-Wales. Biodivers, & 


00 Te The impact ot ie us change on 
c. Natl. Acad. 


Sci. U.S.A. 104: meer] 

Muchhala, N. & P. Janín-V. 2002. Flower visitation by bats 
in cloud forests of western Ecuador. Biotropica 34: 
387—395 


Muellner, A. N., K. Tremetsberger, T. Stuessy & C. M. Baeza. 
2005. Pleistocene rr: and recolonization routes in the 


southern Andes: ights from Hypochaeris | palustris 
(Asteraceae, Lance Molec. Ecol. 14: 203-212. 
Nascimento, H. E. M., A. C. S. Andrade, J. L. C. Camargo, W. 


F. Laurance, S. G. a E-L: Ribeiro; 2006. Effects 
of the surrounding matrix on t ma 


forest fragments. Conservation Biol. 20: 853-860. 


onian 


ecosystems of the Pacific Northwest. Forest Ecol. 
Managem. 246: 57-65. 
uus T. & Y. Ide. 2008. Clobal f ariation 


lant Species bs vertical and 21m | gradients on 
E Biogeogr. 

o, J. Bo &L di 2002. 
Metapopulatun P for extinction threshold in d 
= Eo merr J: Bs Biol. 215: 9 ; 

. Bush, P. A. Baker, S. C. F res O. 
Soles 2003. i Végëtätiðri a fire a “of Lake 
Titicaca since the Last Glacial Maximum. Palaeogeogr. 
Palaeoclimatol. Palaeoecol. 194: 259-279. 


e in a gro 

seniculus) in a TRUM Colombian rainforest. Amer. 
Primatol. a —251. 

Partel, M., A. Helm, T. Reitalu, J. Liira & M. Zobel. 2007. 
pene diversity related to the Late Iron Age human 
population density. J. E i 2. 

Pennington, &]. A. Ratter (editors). 2006. 
Neotr apical. mos ind ey ally Dry Forests: Plant 
Diversity, Biogeo graphy, and Conservation. The System- 
atics Association Special Volume Series 69. CRC, Boca 
Raton. 

Perry, L., D. H. Sandweiss, D. R. Piperno, K. Rademaker, M. 
A. Malpass, A. Umire & P. de la Vera. 2006. Early maize 
agriculture and interzonal interaction in southern Peru. 
Nature E A 

W. Chatrou, J. B. Mols, R. H. J. Erkens & J. 

ntered” genera in the short- 

Testing biogeographical 

hypotheses using phylogeny reconstruction and molecular 

—46. 


Pitman, N. C. A., P. M. Jørgensen, R. S. ms, S. 
& R. Valencia. 2002. P des rate 
estimates for a modern Neotropical flora. Conservation 
Biol. 16: 1427-1431. 


Annals of the 
Missouri Botanical Garden 


Plantegenest, M., C. Le May & F. Fabre. 2007. m 
epidemiology of e diseases. J. Royal Soc. Interface 4 
963-972. 


Pompe, S., Hanspach, F. Badeck, S. Klotz, W. Thuiller & I. 
Kühn. 2008. Climate and land us 
plant o in Germany 

Postigo, J., K. R. Young 
continuity in a pastoralist community in the high Peruvian 
Andes. Human Ecol. 36: 535-551 
nds, J. Bustamante, ma, J. 

Consuegra, M. P. L. Fogden, F. N. Foster, 3 p Marca, 7 

Merino-Viter, Es Puschendorf, S. R. Ron, 

2 A & B. E. Young. 2006. 
Widespread amphibian moe oe from epidemic disease 
driven by global warming. Nature 439: 161-167. 

Pulliam, H. R. 19 Sources, Pu and population 
regulation. Amer. dra 132: 652-661. 

Ramírez, E., , P. Ribstein, M. Descloitres, R. 
Guérin, J. Med za, in "Callsire p. Pouyaud & E. Jordan. 
2001. Small glaciers disappearing in the tropical Andes: A 
case study in e Glaciar Chacaltaya (16°S). J. 
y 47: m 

Rand, As. J: ioe & T. Tschamtke. 

T edge effect ts: The dispersal of eser 

subsidized insect natural Po into adjacent natural 

RA. os uo 9: 603-61 

wr . The ue forest. BioScience 42: 


Reusch, 7 B. H. & T. E. Wood. 2007. mnm ecology of 
global we jane Ecol. 16: 3973-399) 

Rezende, E. Lavabre, P. R a Jd E 
Jordano "a o 2007. Non-random coextinctions 
in me ua structured mutualistic networks. 
Nature 448: 925-928. 

. Baker & M. S. Aldenderfer. 2003. 

“Fluvial Hao of po Rio Ilave valley, Peru, and its 

a to climate and human history. Palaeogeogr. 
Palaeoclimatol. Pasa 194: 165-185. 

Ripple, W. J., n, R. A. Renkin & D. W. Smith. 

2001. Trophic al Aa wolves, elk and aspen on 

wstone National Park's notes range. Biol. Con- 

02: 227-234. 


, M. A. Miasen & A. J. Novaro. 2007. 

Habitat fragmentation a a plant-disperser mutual- 
ism in the temperate forest of South America. Biol. 
cra 139-1 95-202. 

Rull, V. € S. Nogué. 2007. Potential migration routes and 
in for va ur plants of the Quo hus 
highlands dur the Quaternary. J. Bio r. 34: 
132 7-1341 

sed FS au Ozkan, A. Brandolini, R. p -Pregl & W 

002. Genetics and geogra wild cora 


a in the Nea > Bt x Te Genet. 
44.1. 


Sandor, J. A. & ash. 1995. Ancient agricultural soils 
in the Andes t Run Peru. J. Soil Sci. Soc. Amer. 59: 
170-179 

Sarmiento, F.O I in the land 


ighland aoe Geogr. Rev. 92: 213-234. 
. 2008. Effects of predator hunting mode on 


grassland ecosystem function. Science 319: 952-954. 
Scholze, M., W. Knorr, N. W. Arnell & I. C. Prentice. 2006. 
À climate-change risk analysis for world ecosystems. Proc 


Natl. Acad. Sci. U.S.A. 103: 13116-13120. 
Schrott, G. R., K. A. With & A. W. King. 2005. On the 
importance of landscape history for assessing extinction 


risk. Ecol. Applic. 15: 493—506. 


Seabloom, E. W., P. Dobson & D. M. Stoms. 2002. 
Extinction rates under nonrandom patterns of habitat loss. 
Proc. Natl. Acad. Sci B 99: 11229-11234 

Seastedt, T. R., W. D. Bowm: t, 
Townsend & M. W. Williams. 2004. The inis 
continuum: A model for high-elevation ecosystems. 
BioScience 54: 111-121. 

Sergio, F. & P. Pedrini. 2007. Biodiversity gradients in the 
Alps: The overriding importance of elevation. Biodivers. & 
Conservation 16: 3243-3254. 

2 Colonialism 


and landscape in the 


tinuing consequences. Ann. Amer. Geogr. 91: 
de 

Smith, Peterson & D. En 2003. 
mie after wolves: BioScience a ae 

eo A. E. J. 

. Ric n ions 

p dii domestication of guinea pigs (Cavia porcellus 
L.). J. E 270: 57-62. 

Starzomski, & D. S. Srivastava. 2007. Landscape 
geometry vend community response to disturbance. 
Oikos I 690—699. 

Stueve, K. M., C. W. Lafon & R. E. Isaacs. 2007. Spatial 
patterns d ice storm AME on a forested landscape 

ns, Virginia. Area 39: 20-30. 

Environmental P d recruit- 

of palms in 

opical montane rain forest. ij Trop. Ecol. 17: 97-113. 

Talley, T S. 2007. Which spatial heterogeneity framework? 

onsequences for conclusions about patchy population 
489. 


patterns in the avifauna of the Cordillera Vilcabamba, 

Peru. Ecology 52: 23-40. 
Tewksbury, J. J., R. B. Huey & 

the heat on tropical animals. Science 320: 1296— 


C. A. Deutsch. 2008. Putting 
1291. 


Abrupt tropical climate change: Past 
and present. ch Natl. Acad. Sci. U.S.A. 103: 10536— 
10543. 


Thuiller, W., S. Lavorel, M. B. Araújo, M. T. Sykes & I. C. 
Prentice. 2005. Climate 2 threats to e diversity 
in Europe. Proc. Natl. uar Sci. U.S.A. 102: 8245-8250. 

Torres-Carvajal, O. 2007. Phylogeny and ee aphy of a 
large radiation of no tad (Iguania, Seid 
Zool. Scripta 36: 311 

Troll, C. 1968. The UE of the tropical Americas: 
Aspects of climatic, phytogeographical and agrarian 
ecology. Colloq. Geogr. 9: 15—56. 

Tubiello, ii N., J.-F. 2 n - S M iu ss 2007. Crop 

nd p c. Natl. Acad. 


Sci. USA 104: 19686- 19690. 


Landscape ecology: What is the state of 
344, 


7. Scale-specific d between habitat 
il div a landscape 
structure gradient. Oecologia 153: 713-05. 
hmeling, S. Starkenburg, S. Heun Yoo 
. Stewart. 2007. Responses of plant and bird 
communities to prescribed burning in tallgrass prairies. 
Biodi ivers. & Conservation 16: 827-839. 


Volume 96, Number 3 Young 507 
2009 Andean Land Use and Biodiversity 
Veblen, T. T., K. R. Young & A. R. Orme (editors). 2007. The 2005. Andes. Pp. 47—50 in H. Geist a Our 


Physical Geography of South America. Oxford University 
Press, Oxford. 

Velázquez, E. & A. Gómez-Sal. 2007. Environmental control 
of early succession on a large landslide in a tropical dry 
ecosystem (Casita volcano, Nicaragua). Biotropica 35: 
6 09. 


ata, J. Klimes, Z. Patzelt & N. Santillan. 


radley, M. Werner & F. Keimig. 2003. 


owground Components. Prince- 

ton University a Prin ceton. 

s & H. P. Possingham. 2006. The role 
of liabitat dane and recovery in metapopulation 
persistence. Lo 87: 855-863. 

Williams, J. W., S. T. Jackson & J. E. Kutzbach. 2007. 
Projected 25. of novel and Eu climates 
by 2100 on Proc. Natl. Acad. Sci. U.S.A. 104 
5738-5742. 

Young, K. 1 1997. Wildlife conservation in the cultural 
landscapes of the central Andes. Landscape Urb. Planning 


———. 1998. Deforestation in landscapes with humid 
forests in the central A 
15-99 in K. S. . Young (editors), Nature's E 
Geography: New Lessons lor Conservation in UAE 
Countries. on ai bes UE Ped Madiko 

. 2002. M f 


the future of Mom n ae Conservation Biol. 
16: 855-856. 


Earth's ere Land: An Encyclopedia of Land-Use and 
Land-Cover Change. Greenwood Publishing, Nüsse 
Comet: 

———. 2007. Causality E current environmental o dm in 
ar i. Geogr. Compass 1: 1299-1314 

— ——. 2008. Stasis a flux in pes ee locals: 

Change in rural Andean landscapes nA 

Millington & W. Jepson (edi tors), Land- Coss Paine in 

the Tropics: Changing Agricultural Landscapes. Springer, 

New York. 

— & P. L. Keating. 2001. Remnant forests of Volcán 

Cotacachi, ee Ecuador. Arctic Antarc. Alpine Res. 

33: 

a Aspinall. 2006. Kaleidoscoping A ane 

E perspectives. Profess. Geogr. 58: 436-44 

. Lipton. 2006. Adaptive governance and 
inate changes in the tropi ical highlands of western South 
America. es Change 78: 63-10 

——— n. 2009. Natural hazards in Peru: Causation 
and subo: n E. Latrubesse (editor, Natural 

Hazards and Humer ure aed Disasters in Latin 

America. Developments in Earth Surface Processes Series, 

Vol. 13 evier, Amsterdam. 

—, C. Ulloa Ulloa, J. L. Luteyn & S. Knapp. 2002. Plant 

evolution and endemism in Andean South America: An 

introduction. Bot. Rev. 68: 4-21. 

- León, P. M. Jørgensen & C. Ulloa Ulloa. 2007. 
Topical and subtropical landscapes of ae Andes 
E Pp. 200-216 in T. T. Veblen, K. R. Young 

A. R. Orme (editors), The Physical Dr of South 
America. Oxford doy a Press, Oxford 

Zimmerer, K. anging Fortunes: Biodivers sity and 
Peasant Qe ey in the Peruvian Andes. University of 

California Press, Berkeley 

—. 19 verlapping es of mountain agricul- 
ture in Peru and Bolivia: Toward a regional-global 

landscape model. Human Ecol. 27: 135-165. 


APPLICATION OF SCIENCE TO 
PROTECTED AREA 
MANAGEMENT: OVERCOMING 
THE BARRIERS" 


Carolina Murcia? and Gustavo Kattan? 


ABSTRACT 


No other line of y i pplication 


urgeni 
that must be put in ies to eain lines of communication between scientists and manage 


necessary that scientists are aware of the 
ilabl 


information needs of managers, that they produce the Eie e and that 
ond that not only know ho 


ntly than conservation. Here we explore several elements 
rs of protected areas. First, it is 


managers n ess, and 


need for that science and the cleat advantages vi e 


this Hermann d a availa o managers. Second, it = Hecesi 

tion, but that they al t 
it into o tieit practice. l hani ade equate flow 
dialogue between cu part ies, translators of science located both i 


organizations (NGOs), and execution of joint projects. In particular, 


Science an do 


Key words: 


n-the Nue conservation. We finish by descr ibing three case tudi 


25 ld science, Colombia, conservation, management 


of informatio 
in academia and government and E 


nted NGOs can play a major role 


: active 


168 in 


, nongovernment organizations. 


"Science does not provide the solutions, but it can help 
understand ihe consequences of different choices." 


—Lubchenco (1998) 


Science is expected to play a major role in informing 
and implementing decision-making processes that 
affect civil society, because science provides a rigorous 
and objective knowledge framework (May, 1998; Mills 
& Clark, 2001; Eagle et al., 2003; Manning, 2005; Roux 
et al., 2006). In this context, no other 
requires the application of science more urgently than 
the conservation of biodiversity, in which policies and 
decisions could have tremendous implications for all 


line of practice 


forms of life. Yet, the broken lines of communication to 
transfer this knowledge between scientists and conser- 
vation policy makers and managers is a permanent 
concern of parties involved in conservation worldwide 
(Pendergast et al, 1999; Stone, 2002; Pullin et al, 
004; Meijaard & Scheil, 2007). This concern is 
particularly worrisome in iropicald 
because they harbor the greatest proportion of biodi- 
versity but lag behind developed countries, both in 
absolute and relative terms, in monetary resources and 


highly trained scientists and managers. Lack of 
nication leaves valuable information trapped in 
specialized journals, with no application, and tho 
making important decisions are deprived of critical 
information. 


commu 


plied the 
of biodiversity at 
multiple levels along a great continuum—from the 


Scientific knowledge can be ap 


conservati on an managemen nt 


president and high-ranking politicians of a nation to 
the local park warden. Each level deals with issues 
that operate at different spatial scales, with different 
degrees of detail. The professional profiles of the users 
of scientific information are different, and so are the 
mechanisms for into the 


incorporating science 


middle managers, in particular those involved in the 
management of protected areas. These managers are 
typically staff members in national or state govern- 
ment environmental agencies and are the people who 
make the day-to-day decisions on how to manage the 
parks; they are responsible for devising and imple- 
menting management plans for areas and species. 


! We wish to acknowledge the Wildlife a Society and The John D. and Catherine T. eal Foundation for 
ral of the processes described. Th ppo 
by the Colombian National Parks Unit and a Autónoma Regional de Risaralda. manuscript benefited from the 


supporting the involvement of the authors in se 


e work presented here was also supported 


comments of Javier Mendoza, Victoria C. Hollowell, Peter Jørgensen, 
2 Fundación EcoAndina, AA 25527, Cali, Colombia, and a! for Tropical Blues Aparado 676-2050, San Pedro, 


Costa Rica. carolinamurcia0l@gmail.com. 


* Fundación EcoAndina, AA 25527, Cali, Colombia, and footie Universidad Javeriana, Departamento de Ciencias 


Naturales y Matemáticas, Cali, Colombia. gustavokattan@gmail.co 


doi: 10.3417/2008031 


ANN. Missouni Bor. Garp. 96: 508—520. PUBLISHED ON 28 SEPTEMBER 2009. 


Volume 96, Number 3 
2009 


Murcia & Kattan 
Application of Science to Protected Area 
Management 


509 


To incorporate scientific knowledge into the 
practice of conservation, several elements are neces- 
sary. First, the relevant knowledge needs to be 
available. This availability requires that scientists 
address specific conservation and management issues 
and that the 


accessible by managers. Second, managers must have 


in their research, information is 


some incentive, or mandate, to base their decisions on 


training and the ti to access the scientific 
literature, extract the evant information, and 
inco into their action plans, within an 


adaptive framework that would allow them to evaluate 
options, as well as revise and adjust their decisions. 
Finally, the two groups need to be able to communi- 
cate effectively, to ensure that information is properly 
applied, and to provide feedback to scientists on the 
information needs of decision makers and managers. 
In this article, we explore whether these conditions 
are being met, identify the elements that condition the 
generation and use of information, explore how these 
elements are connected by information pathways and 
affected by external factors, and find the gaps. We 
also illustrate several case studies in Colombia in 
which science has been applied to conservation 
planning using different strategies. We argue that 
science-oriented nongovernmental organizations 
(NGOs) can play a major role in bridging the gap 
between basic science and on-the-ground conserva- 
tion. Although this paper is based on our experience 
in Colombia, we believe that it reflects the reality of 
e focus on 
the biological sciences, but the situation is likely to be 


many other Latin American countries. 
similar for the social sciences. 


INFORMATION NEEDS OF MANAGERS 


There is a tremendous amount of scientific 
knowledge pertaining to biodiversity and conservation 
that is published in journals. This knowledge is 
valuable and potentially usable by the conservation 
community in developing countries. However, the 
successful application of this science to management 


needs to account for the following premises. 


SITE SPECIFICITY 


Conservation problems are site specific, involving 
unique biological landscapes, varied human cultures 
with their own local idiosyncrasies, and particular 
sets of economic conditions and interests. Ideally, 
solutions for a site’s conservation problems should be 
devised using locally produced information, but few 
sites across the tropics have a local research program. 
One option is to extrapolate data from similar sites or 


ecosystems, but that is not without risk. Doing so 
requires an in-depth knowledge of the particular 
ecosystems, landscapes, or organisms to which this 

ormation would apply, so that reasonably reliable 
predictions on how as system will respond can be 
made. The more ecologically different the sites are 
from those where data have been generated, the less 
reliable and effective that extrapolation will be. For 
example, extrapolating the impact of edge effects from 
what we know about lowland tropical rainforests in 
Amazonia to high-elevation humid forest in the Andes 
would be less accurate than an extrapolation to other 
lowland humid forest sites. Any extrapolation consti- 
tutes a working hypothesis that needs to be tested. 
Monitoring of any management implementation is a 


recorded and compared to the expected behavior, 
and corrective actions are redirect the 
management to yield the desired effects (Walters & 
Holling, 1990; Johnson, 1999). Therefore, the appli- 
cation of science generated under different circum- 
stances does require an additional component of 
monitoring. 


CONSERVATION PROBLEMS PERSIST BEYOND THEORETICAL FADS 


Some issues that are pertinent to management and 
conservation today were investigated several decades 
ago. However, before some key aspects were fully 
explored, scientists moved on to search for newer 


ideas and other knowledge frontiers. For example, 


temporal scales, to give i un the adequate tools 

to address specific issues such as habitat restoration 

or the effects of EU logging. While s 

are being revisited under new conceptual oe d 

(Guariguata, 1998; Chapman et al. 

practical questions remain unansw ed eph & 
002) and 


, 1999), numerous 
stertag, and managers dep on tho 
answers today to manage pou ud disturbed 
landscapes. For example, how do wet tropical forests 
change pn the middle to late Pod stages 
(Le., > 50 years old; Guariguata , 2002; 
Chazdon et " 2007), and how long do p advanced 
successional stages last (Chazdon et al., 2007)? What 
is the average turnover rate for different tropical 
ecosystems? Given natural rates of disturbance an 
regeneration, what would be the minimum conserva- 
tion area of a given tropical ecosystem to ensure long- 
term persistence? What are the successional patterns 
of different animal communities in middle to late 
stages? What are the 


successional barriers to 


Annals of the 
Missouri Botanical Garden 


recolonization for different animal groups? If these 
questions remained unanswered for lowland wet 
tropical forests, where most tropical research occurs, 
there is less information for other ecosystems, such as 
dry, montane, or high-elevation ecosystems. 


LONG-TERM COMMITMENT TO A SITE OR ISSUE 


Many conservation problems require long-term or 
continued monitoring of a system to understand the 
nature of its internal dynamies. Such is the case of 


patterns (Newstrom et al., 
1999) or of mammal and bird species with low 
fecundity and long life expectancy. Management plans 
that involve these species require information on the 
frequency of the flowering or fruiting events, plus 
some indication of the factors that trigger flowering or, 
in the case of the animals, the frequency of 
reproduction and number of reproductive years of an 
animal to estimate overall fecundity. Such information 
can be attained only after many years of monitoring. 
However, few sites in the tropies enjoy the benefits of 
razil, a 


a long-term research program. For example, 


rog 
country of 8 million ka has only 12 sites recognized 
by the International Long-term Ecological Research 
Network as long-term sites (<http://www.icb.ufmg.br/ 


peld/>). Of those sites, only one is in the Amazon 
forest. 


Factors THAT AFFECT THE GENERATION OF 
APPLIED RESEARCH 


A survey conducted among conservation leaders 
found a general perception that academic research 
does not meet the needs of conservation practitioners 
(Meijaard & Scheil, 2007). 


application of science into management for e 


The generation an 
nserva- 
tion has two main elements: the scientist and the 
manager. Both are affected by a number of factors that 
, transferred, 


modulate how the science is pro 
1). On 


and applied (Fig. the sienta 

factors that affect selection of research topics and 
where this research is published are: the sources of 
funding, the scientists" own intellectual interests, the 
scientists” institutional criteria for professional ad- 
vancement, and the policies of publishers of scientific 
papers. On the manager's side, the factors that affect 
whether they access and apply science to their own 
decisions are: the policies from their own institutions 
regarding how much they expect science to be a part 
of the process and how many resources they provide 
managers to that end (e.g., time, funds, library access), 
their own training, and the sociopolitical context in 
which these decisions are made 


SELECTION OF RESEARCH TOPIC 


veral internal and external factors affect how 
scientists choose the topics for their research. The 
first two factors are of a personal nature. First, and 
most obvious, is the researcher's own curiosity and 
passion to understand a specific system. While some 
researchers prefer to focus their research on more 
theoretical questions, others are increasingly more 
interested in addressing questions that could have 
direct application to conservation issues (e.g., see 
Kremen, . The second motivation is professional 


advancement. Successful research on cutting-edge 

topics that generates many highly visible papers can 
be a significant boost a career, involving 
promotions or better salaries. In public Colombian 
universities, a faculty member's salary within a given 
rank is determined in part by his or her publication 
record. Promotion to higher ranks is also determined 
by productivity. Thus, there is an incentive for 


searching for topics that can produce papers in top 
journals. 


A third factor that affects topic selection is funding. 
Funds can come as grants from private donors or 
science government agencies or as contracts for 
commissioned work from environmental government 
agencies. Funding is also driven by fierce competition 
and by the priorities set by funding agencies and other 
donors. To secure funding from science funding 
agencies, proposals need to show how much science 
will be advanced. Thus, proposals that address 
cutting-edge issues and push the frontiers of science 


the 


Colombian government research granting agency, 


are more successful; however, Colciencias, 


in proposals 
avoid sensitive questions that 


requires a statement of social pertinence 
Donors are likely to 
could generate bad press or conflict (Meijaard & 
Scheil, 2007). This tendency precludes research that 
threats 
identifiable corporations. Donors also tend to s 


addresses caused by highly visible or 
ifi 
their programmatic emphasis frequently enough to 
make it hard for scientists to sustain long-term 
monitoring or research. To ensure continuance of 
funding that would allow the compilation of long-term 
data, scientists must exhibit great ingenuity to keep 
presenting the same system under a novel conceptual 
light at each new funding cycle. 
A more recent source of funding that has the 
ntists in answerln 


potential to engage scie 


g applie 
questions through research is through contracts with 
environmental government agencies. Although a rare 
occurrence, progressive regional environmental agen- 
e.g., as Corpora- 
DER], 


Corporacién Auténoma Regional del Valle del Cauca 


cies in countries such as Colom 
ción Autónoma Regional de Risaralda [C 


Volume 96, Number 3 
2009 


Murcia & Kattan 
Application of Science to Protected Area 
Management 


511 


Transfer Application 


Generation 
r----- 

| | 

| 

: v 
| Access to ee ve oe Own Professional 
field sites dd nalis Sone curiosity | |advancement 

priorities priorities work 
| | | ] 


y 


Researcher 


| Direct 
| consultation 


ure ]. 


| Scientific 
paper — =] 


o 
Public 
+ conference 


Institutional 
policies 


t 


Civil society 


Pathways describing the three-step process of (1) generation of knowledge that may be applicable to 


18 
management of protected areas, a transfer of y EE and E beh of knowledge. The steps of generation and 
d of the ctors (r co 


g the field of a 


application are framed circumscr 


esearcher and manager). ontrast, the 


second step (anten has been lefi neon vio boundaries or an associated actor to illustrate that this is an area that 


s-land, and that each ac 


is currently a no 
show 


related responsibilities. Boxes in dark gray 


t 
the factors diui is ct each of the two actors. Tine thickness Tum our perceived strength of the relationship. Broken 


lines indicate weak lin 


[CVC], Corantioquia) routinely outsource short-term 
research projects to address specific information 
needs. 

Finally, limitations inherent to a field site also 
affect topic selection. To address conservation issues, 
scientists often need to work in protected areas, which 
involves obtaining research permits from the appro- 
priate agencies. A lack of clear procedures by the 
environmental authority or highly complex processes 
makes it difficult to obtain these permits and 
discourages researchers from these areas and research 
topics. When issues involve local communities and 
several implementing agencies, two additional levels 
of complexity are added. Seeking agreement amon 
actors, and accommodating their participation in the 
project, is a diplomatic art in which few scientists are 
trained, although a handful of new schools are 
gradually training scientists in these multidisciplin- 
al, 1994) 


Interestingly, some private landowners have become 


ary, multicultural exercises (Gibbons et 


interested in having research conducted on their 
properties (e.g., for conserving populations of red 
howler monkey Alouatta seniculus L. in the Cauca 
erive a direct 
benefit and often contribute resources in-kind, their 
curiosity about their own systems opens this unex- 
pected opportunity. The availability of a permit-free 
and conflict-free site, where sometimes the logistics 
are provided or facilitated, has recently lured a 
number of Colombian scientists to explore conserva- 
tion and ecological issues in agroecosystems. 


BRIDGING THE GAP BETWEEN SCIENCE AND MANAGEMENT 


Scientists in general can become rather self- 
absorbed. Most scientists are trained in the tradition 
ions from within the 
In this 


tradition, scientists often do not carie t outside of 


of defining research quest 
academie domain (Gibbons et al., 


Annals of the 
Missouri Botanical Garden 


academia about the types of questions that are 
relevant to pursue. Instead, they are driven by the 
pursuit of objectivity, replicability, and rigor in the 
experimental design. This approach, while commend- 
able, leaves scant room for scientists to investigate 
questions. essential for other interest groups, such as 
trend defining the 
focus of science is emerging (Gibbons et al., 1994) 
and involves a more multidisciplinary Bi that 
with the rest of the 
world, frames the question in a practical problem that 


integrates the academic domain 


needs a solution, and engages other actors in the 
formulation of questions. Although this model has 
potential risks (e.g., the researcher loses control of the 
project), it increases insight and applicability of the 
recovered information and also increases accountabil- 
ity to all parties involved. 

Nevertheless, the burden of reaching out to the 
other party cannot be placed solely on the scientists? 
shoulders. Communieation between scientists and 
managers has to flow in both directions, and managers 
need to reach out to scientists (Fig. 1). ri ah in 
developed nations are already addressing the n- 
tific community at their own forums (Ruth et s 
2003). Unfortunately, this is not happening yet in 
ate a culture 
ings and 
address the scientific community to present their 
needs for information. 

Funding agencies can also help to bridge this gap. 
By giving priority to proposals that have a significant 
applicable component, government funding agencies 
can stimulate intranational research that answers 
management questions. While funding research is not 
usually a priority for private donors, these donors are 
also playing a role in shaping the interests of 
scientists id funding research that is directly 

cted to conservation issues and has immediate 
RUN (Stone, 2). 

Although managers or government agencies typi- 
cally have no money and therefore no influence on 
how research is focused, a few progressive government 
agencies in Colombia that have enough resources are 
commissioning work from scientists on specific topics. 
Usually, those contracts are geared toward the 
assessments of biodiversity or the conservation status 
of species or ecosystems. As such, they are not aimed 
at testing cutting-edge scientific hypotheses, but at 
generating information required for immediate appli- 
cation by the agency. However, some scientists are 
able to design their studies in a way that merges with 
hypothesis- -testing exercises. In these cases, it is a 


the scientist obtains funding for 
publishable research. 


Can MANAGERS Access AND Use SCIENTIFIC KNOWLEDGE? 


Applying scientific information to management a 
conservation takes two distinct steps: the transfer of the 
information and the actual incorporation into manage- 
ment plans and actions. These steps have different 
limitations and potentially different actors. The first 
step begins with the scientists who produce documents 
that summarize the questions asked, the results, and the 
implications (theoretical or applied) of their results. 
Accessibility to the documents varies depending on the 
type of publication chosen by the scientist. 


VENUES FOR PRESENTING RESEARCH RESULTS 


Given a choice, scientists m? to publish in 


cin journals (Aarssen et al., to advance 
demicall hei 


p 

aca cally or increase their EUR Papers 
bud journals bring visibility, access to more 
funding, and prestige. However, the collegial compe- 
tition to get an article published in those journals has 
exponentially increased in the past few decades and 
has resulted in a narrowing of the editorial criteria for 
accepting manuscripts (Meffe, 2006; Aarssen et al., 
2008). The competition among international scientific 
journals is fierce (Olden, 2007), and only manuscripts 
that appeal to the greater scientific audience may 
accepted. Therefore, journals favor articles addressing 
cutting-edge issues that require a conceptual devel- 
opment of broad implications and new methods. In 
addition, journals may be especially biased against 
manuscripts with a local d written in 
languages other than English, or with descriptive or 
applied undertones (Meffe, 2006; “Olden, 2007). While 
this may seem unfair at first glance, one must not lose 
sight of the fact that these journals are responding to 
market forces of a distinct sector of the scientific 
population: academics in developed countries (Lawler 
et al., 2006). Al 
applied emphasis has increased exponentially in the 
past 20 years (Lawler et al., 2006), the applicability of 
research results to different sites, ecosystems, or 


though the number of journals with an 


situations that are not central to the articles has not 
necessarily increased (Stinchcombe et al., ; 
Pullin et al., 2004; Lawler et al., 2006; Armstrong & 
McCarthy, 2007). Nevertheless, the selection pressure 
generated by journals can reinforce a scientist's 
determination to focus on cutting-edge issues and 
stay away from more applied topics. 

Less frequently, scientists produce reports in lay 
language, such as in white papers, working papers, or 
gray literature. Reports are usually the result of 
commissioned work and are tailored to suit specific 
needs of 


government agencies. These types of 


documents are assigned a lower value as intellectual 


Volume 96, Number 3 
2009 


Murcia & Kattan 
Application of Science to Protected Area 
Management 


513 


production in academic institutions because they are 
not peer-reviewed documents. In addition, scientists 
require special skills to communicate effectively to a 
general audience. Therefore, there is little incentive 
for this type of work, except perhaps as extra income. 

Faculty are often invited to present their research to 
general audiences as guest speakers of public events, 
and this makes some of their research accessible to 
managers. However, this is often an opportunistic 
effort to 
systematically transfer the information to the conser- 
valion practitioners. Often, the interaction with the 
scientist ends at the end of the talk. 


activity and not part of a concerted 


ACCESSIBILITY OF THE INFORMATION 


Once the information is published, in theory it is 
available for people to use; in Latin America, however, 
this is not a valid assumption. Managers of protected 
areas in Latin America typically have a professional 

(B.Se. or e limited 


exposure to scientific research. While many are trained 


degree B.A. equivalents) and hav 
in biology or related disciplines, a large proportion have 
engineering, administration, or social sciences back- 
ounds. Normally, parks ha 


outline strategies to address their particular problems, 


ve management plans that 


but managers may have to take iis actions in their 
day-to-day jobs, ideally infor y science. Park 
managers face a myriad of conservation problems. For 
many of these issues, there may be sufficient research 
published in scientific journals that addresses these 
problems from theoretical or ee perspectives, but 


olve the 


locates and identifies le aan information and 


this does not res estion ow a manager 
applies it to specific cases. 

The first barrier faced by the manager is time to 
conduct a literature search. This issue affects managers 
even in developed countries (Pullin et al., 2004 
Government agencies are usually understaffed, and 
their managers are overcommitted. Therefore, there is 
little time to devote to an activity that, by its very nature, 
is time-consuming. Although web-based, free tools such 
as Google Scholar (<http://scholar.google.com/>) have 
made literature searches much easier, managers still 
require time to read through the potentially hundreds of 
results for any particular que 

Information needs to be tias 
linguistically, and conceptually. Physical accessibility 

as improved in the past decade. The widespread use 
of the Internet has lessened the physical gap between 
readers and journals, with information theoretically at 
anyone's fingertips. Although most electronic journal 
subscriptions remain expensive, managers in Colom- 
bia find their way around this obstacle by requesting 
articles directly from the authors or by obtaining them 


from colleagues with better library access. Posting 
articles on authors’ web pages certainly facilitates fast 
access to many papers. 

The second component of accessibility is language. 
Most scientific papers are written in English, and 
Latin American managers are rarely fluent in this 
language. A of the conservation-relevant 
scientific literature produced in Indonesia since 
1884 (284 documents, including 81 from the gray 
literature) showed that only. four articles were in 


2007). In time, electronic publishing should make the 
publishing of papers in different languages affordable. 
The third component is the oia language 
sed in 


the scientific literature. A m er who is 
ne to learn ab 


nag 
a specific ure faces the 
challenge of learning a new vocabulary, new concepts, 
and new methods. Costs of DONE. have forced 
rnals and authors to compres 


is 


papers, thus limiting 
ete to those veiled | in the field. Clearly, 
managers in Latin America and other non-English- 
speaking, developing countries require a medium of 
communication that eliminates all of these barriers. 


PROCESSING AND EXTRAPOLATION OF SCIENTIFIC INFORMATION 


Assuming the barrier of accessibility is solved, 

anagers need to process the informati 
extrapolate it to adjust it to the local conditions of 
their site. Managers usually receive complementary 
on-the-job training to improve their administrative 
skills and broaden their capacity to tackle a diverse 
tools for 
i mU such as business or natural resources 


array of problems. Courses may include 


administration, conflict resolution, or social sciences. 


Z 


one of these courses improve the manager's skills to 
survey the scientific literature in conservation biology, 
much less to critically assess what is most relevant or 
best qualified for their particular area or situation. 

hese limitations are solvable, but this solution 
requires the impetus of the government organizations 
and the political will of policymakers. 

More importantly, managers must have a serious 
motivation to incorporate science into their decision 
making. At this time, there is no mandate that requires 
managers to do so. In some Colombian environmental 
institutions such as the National Parks Unit and several 
corporations (regional government environmental agen- 
cies), there has been some gradual effort to ensure that 
management decisions are based on solid scientific 
work, but it is frequently done without the resources and 
expertise required to do it properly. Large discrepancies 
exist among agencies and even countries as to the 
amount of science they incorporate in their work. 


Annals of the 
Missouri Botanical Garden 


Although park siting still occurs in an opportunistic way 
in many countries, Mexico is now basing its conserva- 
tion planning on an extensive nationwide database of 
biodiversity distributions. The next big step is to oversee 
that science is correctly applied. 

In any market-based society, those who improve the 
quality of their product and its fit to society's needs are 
most likely to succeed. Concurrently, to the extent that 
managers incorporate science into their planning and 
decision making, they are more likely to yield a good 
produet for the ultimate consumer: society at large. 
However, results-based incentives to incorporate sci- 
ence may take one or two generations to have an effect. 
The difference in the results between those who apply 
science and those who do not will not be apparent for 
several decades, and there is no time to waste. 
Therefore, we need to seek mechanisms that force a 
change in mindset a 


managers and decision 


B 
makers, and at the same time provide the resources 


for managers to tap into the existing science. 


THE NEED FOR A SCIENCE TRANSLATOR 


Thus far, 
knowledge generation and application, but there is 
Whose 


it to ensure that science is trans- 


we have analyzed some aspects of 


one issue that requires further analysis. 
etc 
erred? Is it the scientists’ or the managers’ respon- 
sibility to es such translation (Stone, 2002)? Current- 
ly, both groups are suffering from overextending their 
reach. Scientists often struggle with guilt that they are 
not sufficiently reaching out to other groups to ensure 
that their science is applied (Whitten et al., 2001), but 
they rarely have the time, mechanisms, or training to 
ilson or Bernd 
Heinrich, are blessed with the natural ability to 


oso. A few scientists, such as E. O. 
communicate outside of their particular academic 
circle. Ironically, these scientists commonly reach out 

eneral publie, but rarely to the managers and 
policy makers, perhaps because of differences in 
receptivity among these groups. As we have seen, 
managers are also overextended and lack the training 
to find, read, process, and synthesize the literature to 
apply it to specific circumstances. Even in countries 
such as the United Kingdom, managers tend to rely on 
other sources of information, rather than the primary 
scientific literature, to develop their management 
plans (Pullin et al., 2004) 
We propose that there is a need for a science 
translator to bridge that gap; s n individual could 
be housed either in scientific institutions (as an 
eii element) or in government organizations or 
science-oriented NGOs (as an official science trans- 
lator and liaison with academia) To this end, a 
number of non-mutually exclusive mechanisms can 


be devised to ensure proper transfer of science into 
policy and management: 


E 


ate an internal 
tor. The 
primary responsibilities of this position would be 


Management organizations crea 


position of science advisor or coordina 


research is conducted in-house (in association 
with universities, research institutes, and re- 
search NGOs) and that the results of this and 
n other 


incoiporsied into 


other research produced outside (ie., i 
systems or countries) are 
GE plans. 

2. Management organizations hire consultants—on 
a need-to-know basis—among the scientific 

community to address specific issues 

3. Management organizations team with universi- 
ties and NGOs and jointly develop a research 
agenda or policy documents that incorporate 
science into planning and action. 


e 


Management organizations form consortia with 
universities, research institutes, and research 
s to address specific conservation or 
management issue 
5. Professional organizations, academia, or NGOs 
conduct training and discussion workshops that 
bring scientists and managers together. The goal 
of these workshops would be to bring new 
scientific advances to the attention of managers 
and to provide an opportunity for managers to 
indicate to scientists the m of issues for 
which they informa 
6. Dti e steer e students 
toward conducting applied research in concert 
with government agencies and NGOs. 
ariations of some of these models have been 
successfully tried in Colombia. We will describe three 
case studies in which such efforts are being made with 
a remarkable positive effect on the quality of the 
conservation decisions made later. 


CASE STUDY 1: ASSESSMENT OF INFORMATION AVAILABILITY 
AND NEEDS IN THE COLOMBIAN NATIONAL Parks 


In 2000, the Unit 
commissioned the design of a research strategy in 
conservation biology that addressed the needs of 
2001; a xus 


social sciences). 


Colombian National Parks 


n areas (Kattan E Murcia, 
process was conducted 
uen was — in two phases: a dapes 
ase an ective construction phase 
es e involved assessing the dos 
information available and its accessibility, as well as 
the infrastructure available for conducting research in 
national parks. The collective construction phase 


involved two componi (1) a training workshop to 


Volume 96, Number 3 
2009 


Murcia & Kattan 
Application of Science to Protected Area 
Management 


515 


update 


from all the parks in key concepts 
gi conservation biology and (2) a series of exercises to 
determine the most urgent information needs to address 
the most critical threats in their parks. 

The information diagnosties evaluated the historic 

ut of research produced in the Colombian 

National Parks between 1975 and 2000. A total of 
726 documents representing scientific papers (pub- 
lished in national or international journals) as well as 
gray literature (theses and technical reports) w 
analyzed. Of thes e cn % fell into ihe 
category of basic vi. 8% assessed effects of 
human and 1% addressed restoration of 
species and e tems 

We also MP the accessibility of the documents. 
All reports were stored in a main library at the National 
Parks Unit headquarters in Bogotá, but the collection of 
theses and other documents was incomplete at this 
library. Half of the published documents were acces- 
sible because they were in local journals or in 
international journals with wide distributions in Co- 
lombia. The remaining half were published in interna- 
tional jonas that were not easily accessible in 
Colombia. Half of the papers were in i the 
remainder were in English, s in Ger 

From the parks-wide consultation for the dp of 
management issues that they were facing, park staff 
identified seven main themes that urgently required 

earch: 


Basic information on composition, structure, and 
dynamies of key animal and plant communities 
(mainly vertebrates and some invertebrates for 
marine ecosystems) and on the dominant 
ecosystems in each park 


2. Basic information on habitat use, habitat 
een population status assessments, and 
demographies of focal (flagship, endemic, and 


d T species 


3. Rates of transformation. of landscapes and 
ecosystems 

4. mpact of resource extraction (timber and 
nontimber products, wildlife, and minin 

5. Carrying capacities of the systems to sustain 
ecotourism 

6. Impact of external threats such as development 


projects in the immediate vicinity of the parks 
7. Restoration (guidelines, protocols, and species 
recommended) 

This exercise revealed that the range of issues faced 
by the parks is wide and covered almost all of the 
current issues addressed in conservation biology at 
the supraorganism level. The information available at 
the time was not sufficient to address those issues 
(research output was an average of 0.6 articles per 
park per year. There was a dramatic decoupling 


between what was being researched and what the 
managers needed. Only the first category (basic 
information on communities and ecosystems) was 
being partially met, because there was little informa- 
tion on dynamic processes at the community and 


ecosystem level a of the ormation 
needed addressed by ensuring that 
researchers turned over the products of their inves- 
tigations and that the main and regional libraries were 
fully stocked with copies of these documents. 

As part of this activity, 40 Colombian park 
managers received their first-e ever training in basic 
principles of conservation 

allowed them to put t 


current AME frameworks and deir their 


. This Bar 
eir management c rns in 


understanding of jargon and concepts in the literature. 
The training also familiarized them with a selection of 
classie scientific papers that had become common 
knowledge for the average biology student in the 
E but were not accessible to the managers. 
e following recommendations ensued from this 
E on: 
Create a full-time position of scientific coordi- 
nator (preferably Ph.D. level) within the parks 
unit who would define research e and 
serve as a science translator for the managers. 
This position could be ny a permanent 
employee of the unit or a position filled by 
university faculty who would serve two- to three- 
year service cycles. 
2. Develop an outreach policy to engage university 
faculty with conservation expertise in planning 


universities and research NGOs will be necessary. 

3. Streamline and clarify the process to grant 
seta permits. 

4. velop a more efficient way to recover, store, 
zi access the information produced by scien- 
tists in the dd parks of Colombia. 

5. Improve facilities to attract scientists to conduct 
research that is relevant to the parks. 

6. reate several research stations in selected 
national parks. 

7. Change the National Parks Unit policy to 
include research as one of its mandates. 

Eight years later, some of these recommendations 
have been put into place. Research is now recognized 
as a more important activity, there is a science 
coordinator position (but not at the Ph.D. level), 

s have research 


several national par stations being 


run in coordination with other organizations, there are 
clearer procedures in place for processing research 
permits, and academies and other scientists are often 
invited by the parks unit to participate in defining new 
management plans. 


516 Annals of the 
Missouri Botanical Garden 
-80 -75 -70 
ND Caribbean 
154 v H15 
1040 A x 
Venezuela 
Pacific Ocean 

54 -5 

Colombia 
04 LO 

Ecuador 
-54 L-5 
Peru 


-80 -75 


-70 


Figure 2. Topographic map of Colombia showing the location of the SIRAP Eje Cafetero (dark line). 


CASE Srupy 2: DIRECT APPLICATION OF SCIENCE TO 
CONSERVATION PLANNING: THE CASE or THE SIRAP 
Eye CAFETERO 


One of th itment 


government in the Rio Convention was the creation of 


acquired by the Colombian 


a national network of protected areas (Sistema 
Nacional de Areas Protegidas [SINAP]). Colombia 
chose to do so in a modular fashion, with the creation 
of regional systems of protected areas (Sistema 
Regional de Areas Protegidas [SIRAP]) that would 


eventually make up the nationwide network. The first 
region to take on this task in 2000 was the Eje 
afetero (EC), the coffee-growing region located in 
central Colombia (Fig. 2). This area is 30,000 km? 
and encompasses portions of two Andean ranges and 
two inter-Andean "ie as well as the core of the 
coffee-growing landsc the country. Middle 
elevations (1200— 1800 a of this highly productive 
agricultural landscape include coffee plantations 
shaded and nonshaded), pastures (for small-scale, 
intensive cattle ranching), other minor crops (mostly 


Volume 96, Number 3 
2009 


Murcia & Kattan 
Application of Science to Protected Area 
Management 


517 


fruits and vegetables), and forest remnants, as well as 
four provincial capitals (each with 300,000 to 750,000 
inhabitants) and small towns. Although the area is 
called the Eje Cafetero, land use in the lower 
elevations (below 1200 m) includes cattle ranching 
and crops such as sugar cane. Upper elevations are 
less populated, and only a few crops, such as potatoes, 
and cattle ranching constitute the agricultural ele- 
ment. The rest of the area is covered with forest 
remnants, páramo (high Andean alpine-type vegeta- 
tion), snowcaps, and exposed soil above the tree line. 
The rich volcanic soils and pleasant weather have 
sustained a dense population since pre-Columbian 
ed by 


ect a prosperous network 


times, and the region is currently travers 
numerous roads that conn 
of rural settlements. In this context, the original forest 
has been reduced to a lattice of forest fragments, 
relegated mostly to elevations above 2 m. Less 
than 10% of the original forest is now protected within 
three national parks and a few regional and municipal 
protected areas. The general objectives of the SIRAP- 
EC project were to increase the area under protection 
and, by doing so, to conserve a representative and 
viable sample of the region's original biodiversity 
(Kattan, 2005, 2006) 

To accomplish this, a consortium of 12 institutions 
including five regional government environmental 
agencies (called Regional Autonomous Corporations 
in Colombia), the National Parks Unit, the Humboldt 
Institute. (the national biodiversity institute), three 
Colombian NGOs, and two international 
formed. Three NGOs (Fundación EcoAndina, Wildlife 
Conservation Society [WCS], and World Wildlife Fund 
[WWF] led this consortium on the techn 
the biological planning phase. This phase involved: (1) 
defining the goals and objectives of the SIRAP-EC (i.e., 


defining what and how much we wish to conserve), (2) 


ical aspects of 


compiling spatial information and generating potential 
and current vegetation maps, (3) compiling a database 
of all biodiversity records in the region, (4) conducting 
biological information and conservation gap analyses as 
well as some rapid biodiversity inventories, and 
recommending new protected areas. 

This work was collaborative and all of the aforemen- 
tioned vae d ue participated in the 

orkshops and m Eni er, different organiza- 
tions played 2 voles coAndina, WCS, an 

WF were in charge of ED steering the 
biological planning phase, compiling the information, 
performing the analyses, conducting a consultative 
process with experts, and proposing management 
strategies to gain functional regional connectivity. The 
WWF provided the training in geographic information 
system (GIS) to technicians at the corporations and 
NGOs, conducted some of the GIS analyses, and 


The role of the 


government organizations was to coordinate the project, 


supervised the final GIS steps. 


steer the project in the social and political arenas, 
process the geographic information in their areas (under 
in the collection of 
IRAP-EC to other 


ecoregional planning, social 


WWF's supervision), assist 
biological information, link the 
(e.g. 
initiatives, disaster prevention, and the SINAP), and 


regional processes 


generate support in the process at high levels of regional 
and national decision making in preparation for the final 
proposal of protected areas. In addition, experts from 
different universities, NGOs, and other government 
institutions were asked to contribute to several planning 
workshops, or hired to prepare some of the sociological 
analyses that framed the SIRAP-EC. The vision, project 
goals, selection process of candidate areas and focal 
species, ang identification of ey conservation areas 


y allo 
a consulatise exercise that also involved national mi 
regional scientists. 

Some accomplishments of this process include the 
following: 
l. A proposal to increase the cover of protected 

areas in the region of the EC, based on the four- 

R principle: representation, O redun- 

30-33). 
a for 
conservation that would PRORA p» area under 
protection from 200,000-700,000 ha. (or ap- 
o 23 
h 


cy, and restoration E 
] areas we 


aney, 
Sixty-four additiona 


% of the region). Several areas 
already been granted protected status. This 
ieri. QUAD E the latest scientific 


advances in reserve network design. 


po 


elopment of management plans for focal 
species. Initially, seven species were selected 
for developing science-based management 
plans. The first two, the Cauca guan (Penelope 
perspicax Bangs) 
(A 


and the red howler monkey 
louatia seniculus L.), no 


w have management 
plans that are based on research conducted by 
Fundación EcoAndina—WCS on population sta- 
tus, habitat requirements, and availability, as 
well as their current distribution through their 
Ee and ecological ranges (Kattan & 
Valderrama, 2006; Valderrama & Kattan, 2006). 


contain a MN MK Wei with opportuni- 


ties for carrying ou posed activities. 

3. Development of s management plans 
for 18 additional vertebrate species in one of the 
provinces (Va e 


the multicolored tanager (Chlorochryssa nitidis- 


Annals of the 
Missouri Botanical Garden 


sima Sclatter), which is endemic to west-central 
Colombia, and the pacarana (Dinomys branickii 
Peters), a unique Andean rodent. 
4. Updating of management plans for already- 
established protected areas and their surround- 
In the case of the Otun 
lan was expanded 


ing landscapes. 
watershed, the management p 
to a regional scale. Two corporations (CVC and 
CARDER) have commissioned research projects 
to produce information for the updating of 
management plans. 

These processes involved the incorporation of 
conservation biology concepts ranging from island 
biogeography and fragmentation to WA 
ecology and habitat restoration. This was accom- 
plished through a series of workshops in Ds of 
conservation biology that 1 to the design of 
networks of protected areas and through the produc- 
tion of educational materials (e.g., Kattan & Naranjo, 
2008). In addition, GIS technicians from all of the 
participant government organizations received train- 
ing on satellite image analysis and classification. 

An offshoot of this process was the transference of 
science to other regional pla 
Colombia. One example is the Sistema Departamental 
de Áreas Protegidas (SIDAP) Sonsón, a province-level 


anning exercises in 


network of protected areas for the southern portion of 
Antioquia. Here, government organizations led the 
process, using the SIRAP-EC as a template, and 
EcoAndina—WCS assisted the process by prepari 
habitat model for tacled bear cor 
ornatus F. G. Cuvier, the animal with the largest 


the spect 


habitat requirements), which served as a planning tool 
to define conservation priorities. 


CASE STUDY 3: CONSORTIUMS OF ORGANIZATIONS THAT 
PROMOTE COMMUNICATION: THE Case or REDBIO 


third model that 
transferring science to practice is Red de Investigación 
en Biodiversidad y Conservación (REDBIO), which is a 


consortium of 25 organizations including national and 


has proved successful 
I 


regional government organizations, NGOs, and univer- 
sities from western Colombia that was mort in 2002. 
The consortium's main objectives are to promote 
communication, share experiences, and foster research 
and conservation projects. In its six years of activity, 
REDBIO has organized annual regional symposia on 
research and conservation of biodiversity. These 
symposia provide a showcase for ongoing research 
conducted in the region and are well attended by local 
researchers, NGOs, and protected area managers. 

DBIO has also organized several courses and joined 
forces with Universidad de Antioquia, Colombia, to 
offer a graduate degree in protected areas for staff from 


parks, corporations, and NGOs. In addition, all partner 
organizations collectively compile a database on 
regional biodiversity that is accessible to a 


FACILITATION Is THE Key: THe ROLE or INDIVIDUALS 
AND NGOs 


The current transfer of science in Colombia is 
happening as a result of changes in the human and 
institutional dimensions. There is now a synergistic 
partnership between the government an 8. 
Government organizations have discovered the benefits 
that result from these partnerships and are increasingly 
engaging those in academia and civil society as advisors 
and contributors to the planning processes. This 
development is due in part to a number of highly 
committed individuals who, through their careers, have 
moved between government organizations, NGOs, and 
academia; because of their professional experience and 
mobility, they have created networks of connections in 
which they act as mobile links. 

A key element in this synergism is the emergence of 
original 
with 
conservation agencies. These NGOs conduct research 


local, science-oriented NGOs that generate 
d llaborati 


esearch and usually work in collaboration 


that addresses conservation needs. Although they face 
several of the same challenges as academics when it 
comes to raising funds, they are less pressured to 
publish in prestigious international journals. Thus, they 
have more freedom to choose their research questions 
and to focus on those that are most relevant to local 
conservation. They also produce materials that synthe- 
size current knowledge on certain issues, and by 
participating in planning and other conservation action 
processes, they assist in the transfer of science to the 
government agencies. These NGOs are sensitive to local 
needs and have the flexibility to adapt their agendas to 
serve those needs. They are also committed in the long- 
term to sites or regions, and this allows them to collect 
data over a long period. This commitment is vital for 
understanding the responses of ecosystems and species 
to stressors and to monitor the impact of management 
practices. However, NGOs depend on money from 
grants and donors, which makes them fragile (Dour- 
ojeanni, 2006). Therefore, ae efforts should be 

laced on strengthening national NGOs, because they 
can play a significant Rs in improving the transfer of 
basic science into actual conservation practice. 


CONCLUSIONS 


Environmental problems abound, and some are 
already sufficiently advanced to create effects of 
global proportions. Therefore, conservation practice 
must make use of all the information available to 


Volume 96, Number 3 
2009 


Murcia & Kattan 
Application of Science to Protected Area 
Management 


519 


curtail and potentially reverse the negative effects of 


human practices. This huge task cannot fall on the 
shoulders of scientists or managers alone. It requires 
an integrated approach in which all parties in- 
volved. vernments, scientists, NGOs, donors, the 
media, scientie journals, and the general public— 
join forces to ensure that the relevant information is 
generated and applied in creative and bold ways. 
Elsewhere, there are already a number of different 
mechanisms in place that allow the application of 
science to decision making, planning, and implemen- 
tation of conservation (e.g., the Science and Technol- 
ogy Awareness Network of Canada). These mecha- 
nisms can be adapted and should be disseminated 
throughout the developing world. so requires 
strengthening organizations, ecu n those at the 
national level, that are already serving as translators 
of science into practice, fostering the creation of new 
organizations that can serve in this 
overn 


role, mandating 
ments to incorporate science into their work, 
and further involving NGOs and academia in planning 
and decision-making processes. 
Notwithstanding, there is a need for government 
organizations to strengthen their own staff so that they 
can directly access and apply science. To the extent 
that government officers embrace the notion that it is 
important to apply scientific evidence 2 methodol- 
ogies to their decisions, the t 
ill occur more freq 


transfer science 
uently and ede This w 
only happen when there is a minimal, basic level of 
understanding of how science can be applied within 
these organizations. 

Although budgetary restrictions are a constant in 
government organizations in developing countries, local 
experience shows that even small investments can 
produce valuable information. In addition, communica- 
tion among those who generate the science and those 
who should apply it requires little monetary investment 
and can yield impressive returns. 


Literature Cited 


Aarssen, L. W., T. Tregenza, A. E. p C. J. Lortie, J. 
Koricheva & R. Leimu. 2008. for your buck: 
Rejection rates and impact pede in Woran journals. 
Open Ecol. J. 1: 14-19. 

M. A. McCarthy. 2007. Big decisions and 
Adapting pour. m to E needs of 
pr a A m Ecol. 2 
pue 5 T 2009. 
Chan man, C. A., R. W. ae L. J. Chapman, D. K. 

Kennard & A. E Zanne: 1999. Fruit and flower phenology 

at two sites in Kibale National Park, Uganda. J. Trop. Ecol. 

15: 18 "d 


JAF 


Chazdon, R. L., S. G. Letcher, M. van ceca M. Martínez- 
Ramos, F. m & B. Finegan. 2007. Rates of chan 
tree communities of secondary Neo 


tropical D pile ae 
major disturbances. Philos. Trans. Ser. B 362: 273-289. 


Dourojeanni, M. J. 2 ¿Organizaciones no gubernamen- 


ales e ax. o “transnacionales”? Ecol. Aplic. 5: 
157- n 


Eagle, K. . Garson m G. A. Beller & C. Sennett. 
2003. _. the ga ween science and practice: The 
a d professional Ee Health Affairs 22: 
oes p? "n De H. Nowotny, E Schwartzman, P. 
Scott & M. 1994. The 
d 

Groves, C. nt: A 
20 Cii to Planning for Biodiversity. Island 
Press, Washin D.C: 

Guariguata, M. R 1998. Consideraciones Ecológicas sobre la 
Regeneración Natural Aplicada al Manejo Forestal. 
CATIE, Turrialba, Costa Rica. 

& R. Ostertag. 2002. Sucesión secundaria. Pp. 5 
6 in M. R. Guariguata us Do Kattan Lo: 
e ye de es Neotopicaos Libro 
gional, n Eom Ric 


The role of adaptive i emen as an 

eperalional sah En TROU management agencies. 

Conserv. Ecol. 3: L.org/vol3/iss2/ 

art8/>, accessed 5 May 2 009. 

jen. 2005. Planificando el Edén: Principios funda- 
e 


Instituto de Sd rum de Recursos Biológicos Alex- 
ander von H t, Bogotá 
06. Reduce me and practice in designing 
a A cie reserve system: A Colombian case study. 
48-550 in M. J. Groom . Meffe & C. R. Carroll 
Rio Principles of Conservation Biology. Sinauer 
Associates, Sunderland, Massachuse 
— —— & C. Murcia. 2001. Desarrollo dé una Estrategia de 
Investigación en Biolog 
de P 


e la Conservación en el Sistema 
arques ees Dom Unidad de Parques 
Nacionales de Colombia, Bogot 

C. Valderrama (editors). 200 6. P 
ción de la Pava Caucana (P. 
d ee 


lan de Conserva- 

enelope e per rspicax). Instituto de 
Recursos Biológicos Alexander von 
mboldt and ics EcoAndina/Wildlife Conserva- 

tion Society, B 
& L. 


de 


——— 3 ai 2008 
Herramientas para | 
nales de Áreas 
EcoAndina, and Warid Wildlife Jnd Cali, Co 
Kremen, 005. M What do we 
need to know E their ol Ecol. "Lett 8: 468479. 
Aukem -Gr 


i Regiones Biodiversas: 
In. 


onser- 


+) Zu 


Manning, P. 2005. Bri 


gap bereen 
he science and n 


Science and Technol- 
ogy Network Conference, Toronto, Canada.<http://www. 
2 ] ] 1 1 E TURN SE y T J./D, 


pdf>, E 5 May 2009. 
May, R. M. 1998. The scientific investment of nations. 
Science zu md 


Me K. The success—and challenges—of 
čonserjation E Conserv. Biol. 20: 931—933. 


Annals of the 
Missouri Botanical Garden 


Meijaard, E. & D. Scheil. 2007. Is wildlife research useful for 
wildlife conservation in the tro 
with global implications. 
3053— 306 

Mills, T. J. & R. N. Clark. 2001. Roles of research scientists 
in natural resource decision-making. Forest Ecol. Man 
agem. 153: 189-198. 

Newstrom, L. E., G. W. Frankie, H. G. r & R. K. 
Colwell. 1994. Diversity P: long-term iere patterns. 
Pp. e in L. A. McDade, K. S. Bawa, A. 


ee A review for Borneo 
Biodiv Conserv. 16: 


E 
D. 2007. How do ecole gical journals stack up? 
anking of scientific quality according to the h index 
^ 376. 


R. M. Quinn & J. H. Lawton. 1999. The 
ry and practice in selecting nature 
reserves. Conserv. Biol. 13: 484—492. 


Pullin, A. S., T. M. Kuight, D. A. Stone & K. Charman. 2004. 
Do servation managers use scientific evidence to 
support their decision-making? Biol. Conserv. 119: 


245-252 

Roux, D. J., K. ogers, i e Biggs, pz. duy 
Sergeant. 2006. Brideing t 

Moving from a knowledge transfer to a 


edge interfacing and sharing. Ecol. Soc. 11: 4. <http:// 
www. es org/voll Vis] land: accessed 5 
Ma 


Ruth, 3 M D. E Petit, J. R. Sauer, M. D. us F. A 
oe E E. a & J. P. Bennett. 
: Priorities for the 
AI, 
a he B. R. Hudgens, P. L. Bloch, S. 
W. F. Morris. 2002. The influence of the 
E emic conservation. ed Mos on endangered 
m recovery pla . Conserv. Ecol. 6: 15. Bd 
consecol.or; dace 5» | 5M 
ione. D. 2002. Using knowledge: The dilemmas of cee 
research and policy.’ En 32: 285-296. 
Valderrama, C. & G. n (editors). 2006. Plan de 


Conservación del Me ps (Alouatia seniculus) en 


Sincheombe, E [e ea 


Humboldt and Fundación EcoAndina/Wildlife Conserva- 
tion Society, Bogotá. 

Walters, C. J. € C. S. Holling. 1990. Large-scale 
management experiments and learning by doing. Ecology 
71: 2060-2068. 

Whitten, T., D. Holmes & K. MacKinnon. 2001. Conserva- 
tion biology: A displacement behavior for academia. 
Conserv. Biol. 15: 1-3 


Volume 96, Number 3, pp. 369-520 of ANNALS or THE Missourt BOTANICAL GARDEN 
was published on September 28, 2009 


www.mbgpress.org 


CONTENTS 

Biodiversity and Conservation in the Andes: Introduction Peter Moller Jorgensen 

The Andes: A Geological Overview from a Biological Perspective | | Alan Graham 

Climate in the Dry Central Andes over Geologic, Millennial, and Interannual Timescales 
Christa Placzek, Jay Quade, Julio L. Betancourt, P. Jonathan Patchett, Jason A. Rech, 
Bisco Claudio Latorre, Ari Matmon, Camille Holmgren & Nathan B. English 

Diversification of the South American Avifauna: Patterns and Implications for Conservation 
in the Andes Jon Fjeldsà & Martin Irestedt 

Implications of Genetic Differentiation in Neotropical Montane Forest Birds 


Jason T. Weir 
Moss Diversity and Endemism of the Tropical Andes_ Steven P. Churchill 


Andean Speciation and Vicariance in Neotropical Masorarnate (Gentianaceae—Helieae) 
Lena Struwe, Scott Haag, Einar Heiberg & Jason R. Grant 
Determinants and Prediction of Broad-scale Plant Richness across the Western Neotropics 

PEN Trisha Distler, Peter M. Jorgensen, Alan Graham, Gerrit Davidse & Iván Jiménez 
Andean Land Use and Biodiversity: Humanized Landscapes in a Time of Change — 
Kenneth R. Young 
Application of Science to Protected Area Management: Overcoming the Barriers 


Carolina Murcia & Gustavo Kattan 


Cover illustration. A sword-billed hummingbird (Ensifera ensifera) pollinating Passiflora 


tarminiana, drawn by Barbara Alongi. 


Annals 
of the 
Missouri 
Botanical 


Garden 
207 VG 


Volume 96 


umber 


Annals of the 
Missouri Botanical Garden 


Volume 96, Number 4 
December 2009 


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, in- 
dependent reviewers. 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 
Allison M. Brock 
Associate Editor, 

Missouri Botanical Garden 
Tammy Charron 

Editorial Assistant, 
Missouri Botanical Garden 
Cirri Moran 

Press Coordinator, 
Missouri Botanical Garden 


Roy E. Gereau 

Latin Editor, 

Missouri Botanical Garden 
Ihsan A. Al-Shehbaz 
Missouri Botanical Garden 
Gerrit Davidse 

Missouri Botanical Garden 
Peter Goldblatt 

Missouri Botanical Garden 
Gordon McPherson 
Missouri Botanical Garden 
Charlotte Taylor 

Missouri Botanical Garden 
Henk van der Werff 


Missouri Botanical Garden 


For subscription information contact ANNALS 
or THE Missouri BoranicaL GARDEN, 96 Allen Mar- 
keting & Management, P.O. Box 1897, Lawrence, 
KS 66044-8897. Subscription price for 2009 is 
$175 per volume U.S., $185 Canada & Mexico, 
$210 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 


Tue ANNAIS 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 MissoURT BOTANICAL GARDEN, Ye 
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- 
stract/Global Health databases, ingenta, ISI) databases, JSTOR, Research Alert, and Sci Search®. 
The full-text of ANNALS or THE Missouri BOTANICAL GARDEN is available online though BioOne™ (http:// 
www.bioone.org). 
© Missouri Botanical Garden Press 2009 
The mission of the Missouri Botanical Garden is to discoy er and share knowledge about plants and 
their environment, in order lo preserve and enrich life 

© This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 


Volume 96 
Number 4 
2009 


Missouri 
Botanical 


Garden 


A REVISION OF THE MALAGASY 
ENDEMIC HELMIOPSIS 
(MALVACEAE S.L.) 


Wendy L. Applequist? 


ABSTRACT 


taxonomic revision of the endemic Malagasy genus Helmiopsis H. Perrier (Malvaceae s.l.) is presented, based on material 
that has been col omplete revision by A: 


Preliminary conseivatien iecur qs using I 


UCN criteria indicate that most taxa are of f potential conservation concern: four 
ould be considered Vulnerable, and on I 


sidered Data 


Deficient but likely to i at 


r 
ords: Dombeya, Helmiopsis, IUCN Red List, 


Madagascar, Malvaceae. 


Helmiopsis H. Perrier is among a considerable 
number of endemie Malagasy genera of Malvaceae 
s.l; it was formerly placed in Sterculiaceae and may 
have affinities to genera such as Dombeya Cav. or 
Nesogordonia Baill. It is characterized by features 
including a lepidote indument on inflorescences, sepals, 
ovaries, petioles, and usually twigs and leaves; biseriate 
perianth with well-developed, not ligulate or appendi- 
culate Due usually apically winged ovules and seeds; 

l corona, with groups of fertile stamens 
alternating m staminodes; and glandular sepals and 
often petals. At the time of the last complete taxonomic 


revision (Arénes, 1959), little material was available, 
with four of the eight species and one variety described 
rom a single specimen, and only a handful of 
collections available for most of the others. Many more 
specimens have since been collected, especially in the 
extreme north of Madagascar, to which several species 
are endemic. This additional material has enabled 
better understanding of the genuine range of morpho- 

ogical variation within the genus and of the deb an 
distributio 
were underrepresented in Arénes’ (1959) treatment. q 


n of the more common species, both of w 


revision is therefore presented in which nine species are 


! T thank the Muséum national d'Histoire naturelle (P) and the Parc Pu a et Zoologique de Tsimbazaza (TAN) an 


curators thereof for access to collections and for loan 
P and helpful discussion; R. Halse, M. Koopman 
editing; R. Gereau for 


assistance; and D. Bills and T. Distler for the maps. 


help with the Latin diagnoses; B. Alon 


ry Il and P. B. Phillipson for P e at 
well, B. Parada, and A. Brock for helpful comments and 
gi for illustrating new taxa; F. Keusenkothen for computer 


? William L. Brown Center for Plant Genetic Resources, Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 


63166-0299, U.S.A. wendy.applequist@mobot.org 
doi: 10. 3417/2007184 


ANN. Missouni Bor. Garp. 96: 521-540. PUBLISHED oN 30 DECEMBER 2009. 


Annals of the 
Missouri Botanical Garden 


recognized; one species is newly described, and one 
more placed into synonymy. 


Taxonomic History 


Perrier de la Bathie (1944) described Helmiopsis 
with a single species, H. inversa H. Perrier, and two 
named varieties (not validly published due to the lack 
of a Latin diagnosis, as required by Art. 36.1 of the 
International Code of Botanical Nomenclature [ICBN; 
McNeill et al., 2006]) from a total of four specimens 
he had collected in western Madagascar. Simulta- 
neously, he described three new species of Nesogor- 
donia Baill., the type of which had been doubtfully 

ae, and that the 
relate d both 
Sterculiaceae. (e 1952) proposed to transfer 


two genera 
better referred to 


two species pl placed within Dombeya (D. 
, D. s ins Baill) to 
mbinations were not 
validly published. He aoe ree uds place- 
ment of Helmiopsis and Nesogordonia within Stercu- 
liaceae, remarking on similarities between the wood of 
those genera and that of Mansonia J. R. Drumm. ex 
Prain (Sterculiaceae), while suggesting a possible 
affinity to Tiliaceae. Arénes (1956a, b, 1959) also 
recognized Helmiopsis as belonging to Sterculiaceae; 
that entire family is now included within Malvaceae 
s.l. (Bayer et al., 1999). 
Helmiopsis was last treated in full by Arénes (1956a, 
b, is dp oe Np p only two species, 
m based u of Perrier’s (1944) 
2. rin ed P Hi inversa; he rejected 
Capuron's (1952) view that Dombeya richardii and D. 
pseudopopulus should be 


one 


placed within Helmiopsis. 
However, Arénes (1956b) published a revision recog- 
nizing eight species, including both of Capuron’s 
additions as well as two other former Dombeya species. 
He placed the newly described H. hily Arénes alone in 
subgenus Phelmiosis Arénes due to its reduced number 
of T and the remainder in subgenus Mihelopsis 
s. Publication of the fo 
blish 


Arè ormer subgenus automat- 
ele esta 


ed the corresponding autonym 
subgenus Helmiopsis for the subgenus including the 
type (necessarily H. inversa) under Article 22.3 of the 
ICBN (McNeill et al., 2006), and *subgen. Mihelop. 

by including that type, was not validly published d 
Article 22.2. Within that group, section Nudipetalae 
Arénes—again, not validly pd GN Md four 
species Mey relatively narrow 

stamens (15 to 30), and coma. an while 
section de Arénes comprised three spe- 
cies, including H. richardii (Baill.) Capuron ex Arénes 
and H. pseudopopulus (Baill) Capuron ex Arénes, 
which are characterized by broad leaves, normally 10 


stamens, and — glandular petals. The latter group is 
confined to a few areas at the northern tip of 
Madagascar, whereas the remaining species are more 
broadly distributed. 

Arénes (1959) renamed his infrageneric groups to 
conform to the rules of nomenclature, subgenus 
Mihelopsis and section Nudipetalae becoming subge- 
nus and section Helmiopsis, respectively, and largely 
maintained the same treatment of species. In two 
works (Arénes, 1956a, 1959), he recognized the 
invalidly published Helmiopsis inversa var. arenicola 
H. Perrier as a distinct variety, while unfortunately 
never validating it. Barnett (1987) described a new 
northern species, H. sphaerocarpa L. C. Barnett, which 
has certain features unique within the genus (virtual 
absence of a seed wing, spheroid capsule) but clearly 
belongs to section Glandulipetalae. 

Dorr (2001) has also recently transferred Dombeya 
rigida Baill. to Helmiopsis. Dombeya rigida and its 
apparent sister species, D. linearifolia Hochr., were 
treated by Arénes (1958, 1959) as Dombeya subsect. 
Rigidae Arènes, characterized by scorpioid cymes and 
glandular petals. Subsection Rigidae i is placed within 


ochr. e 
scorpioid p but glabrous petals. The ovules and 
seeds of D. rigida bear a shallow marginal wing; 
according to Arénes’ (1959) treatment of Malagasy 
Sterculiaceae, winged seeds would place the species 
within the tribe Helmiopsideae Arénes, rather than in 
Dombeyeae Benth. & Hook. Characters that would then 
place D. rigida within Helmiopsis rather than a related 
genus include a monadelphous androecium, with 
stamens borne opposite the sepals and staminodes 
opposite the petals, and glabrous locules (Dorr, 


— 


Stamen position was treated by Arénes as being a 
reliable character at the generic level, but actually is 
more variable within certain genera than usually 
recognized, and indeed appears to vary within D. rigida 
itself.) This placement may be regarded as controversial. 
Dombeya rigida (and D. linearifolia) would be anomalous 
in Helmiopsis, lacking a number of characters common 
within the genus, and most of the characters that are 
shared with species of Helmiopsis can also be seen in 
some a species of Dombeya, although glandular 
petal 


p pical of the latter genus. Dombeya 
is an exceedingly variable genus, and the possibility that 
it is paraphyletic or polyphyletic relative to other 


Malagasy genera of Malvaceae s.l. cannot be discounted. 
MORPHOLOGICAL CHARACTERS 
VEGETATIVE CHARACTERS 


All species of Helmiopsis are woody plants, usually 
reported to be trees or large shrubs; some reach to 


Volume 96, Number 4 
2009 


Applequis 
Revision 7 Helmiopsis 


20 m in height. Twig bark is grayish to dark reddish 
brown, shallowly ridged, with small pale lenticels; in 
twigs are at least sparsely 
s are alternate, simple, and 
stipulate; the stipules, where known, are small and 
inconspicuous, lepidote, and always rapidly caducous. 
Petioles are always lepidote. Shape and size of the leaf 
blade and relative length of the petiole are quite 
variable, as is the shape of the leaf apex. Section 
Helmiopsis has relatively narrow leaves, with length 
usually well exceeding breadth, and petioles that, 
seldom exceed 1.5 e 
has 


leaves ranging from 


e species 


xcept 1 ong; 
section ‘Glandulip 


etalae 
broadly ovate to orbicular, in one species the breadth 
well exceeding the length, and with petioles frequent- 
leaf base 


nded-truncate or shallowly 


ly 2 em to several centimeters long. The 
ranges from rounded to rou 
cordate; the margins are usually entire to weakly 
undulate or weakly repand, but may be crenulate to 
crenate-serrate. Both leaf surfaces or at least the 
abaxial ace are lepidote in most species; one 
other 
has sparse stellate pubescence on the adaxial surface 


su sp 
species has entirely glabrous a while ano 


and typical scales abaxially. All but one species of 
section Helmiopsis have three or sometimes five basal 
veins, with the lateral veins weak and inconspicuous, 
so that venation is basically pinnate; species of 
section Glandulipetalae have five to seven well- 
developed palmate veins at the base. Most species 
can be differentiated solely by their leaf morphology. 


INFLORESCENCE 


Inflorescences are cymose or geminate; cymes may 
well de 


several orders of branchin 


veloped with elongated peduncles and 
g (a character seen only in 
section Glandulipetalae) or small, few-flowered, and 
sometimes umbelliform. Inflorescences are normally 
lateral; only one species seems to produce genuinely 
terminal inflorescences, although several species may 
appear to have terminal inflorescences due to tight 
clustering of cymes at the tips of twigs. Bracts and 
bracteoles, where known, are small and rapidly 


caducous. Pedicels are angular, except in one species 


with subterete pedicels, and are articulated at the 
point of insertion of the bracteoles, which may be 
immediately beneath the bud or well below it. AII 


parts of the inflorescence are lepidote. 


CALYX 


Sepals are more or less lanceolate, fused only at the 
base, and lepidote, and have a patch of glandular 
tissue on the inner surface near the base that varies in 
extent and visibility. This has been presumed to be 


absent in some species, including all of those that 
have glandular petals; Barnett (1987) was the first to 


ep 
described Helm 


species of Hino very small relict patches of 


psis sphaeroca. 


apparently glandular tissue remain within the basal- 
most portion of the calyx, where they are not readily 
observed. Sepals 
Glandulipetalae, but generally not so far as to display 


are sometimes reflexed in section 


the glandular tissue. 

Similar small basal spots of glandular tissue were 
noted by Hochreutiner (1926) in a species of Dombeya 
sect. Capricornua (D. urenoides Hochr.). Specimens of 
other species of that section (including D. borragi- 
and probably D. 


marivorahonensis Arénes and D. triumfettifolia Bojer) 


nopsis Hochr., D. greveana Baill., 


have been observed to have tiny glandular patches 
within the basalmost fused portion of the calyx, 
although D. rigida and D. linearifolia do not. The 
character is also more widely distributed both in 
Dombeya subg. Dombeya (e.g., D. cacuminum Hochr.) 
and in subgenus Xeropetalum, including virtually all 
species of section Decastemon (pers. obs.; in prep.) as 
well as some species of section Xeropetalum (e.g., D. 


Applequist, 2009]). However, 


rienanensis Áppleq. 


a 


species in both subgenera lack sepal glands. 
Similar sepal glands might represent evidence of 
affinities between Dombeya subg. Dombeya and 
Helmiopsis; however, many distantly related genera 
of Malvaceae also bear nectary tissue on the sepals 


(Vogel, 2000 


COROLLA 


Petals vary in size, caducousness, and in 
presence or absence of small rounded, easily ee 
surface glands on the basal portion of the adaxial 
are generally obovate, 


surface. The occasionally 


markedly asymmetrical apex and frequently seem to 
be caducous. 


ANDROECIUM 


The androecium varies in the depth of the fused 
basal corona, the number of stamens, and the relative 
length and shape of the staminodes, filaments, and 
anthers 


fertile stamens, whic 


. Staminodes may be shorter or longer than 
h 


are borne in five groups 
alternating with the staminodes; staminodes are longer 
than the stamens in section Glandulipetalae and one 


un 


pecies of section Helmiopsis, H. glaberrima Arénes. 
Traditionally, Helmiopsis is described as having 
oppositisepalous stamens, as do many other plants, 


Annals of the 
Missouri Botanical Garden 


and oppositipetalous staminodes; this is one of the 
features that is said to differentiate it from Dombeya, 
which, like a few other Mala: 
s.l, usually has oppositipetalous stamens and oppo- 


gasy genera of Malvaceae 


sitisepalous staminodes (Arénes, 1959). However, this 
character proves to be more variable than generally 
believed. Helmiopsis specimens referable to section 
Glandulipetalae and one other H. hil 


usually have more or less oppositisepalous stami- 


species, H. hily, 


nodes, peni da some degree of apparent variation in 
stamen and staminode position can be seen even 
within diis Intermediate positions of these 
organs, not clearly opposite either petals or sepals, do 
occur. 


OVARY AND FRUIT 


The ovary is lepidote and 5-carpellate except in one 
species, Helmiopsis hily, which has three or occasion- 
ally four carpels; ovary lobing, style length an 


ry 
scaliness, and conformation of stigma branches are 
variable. All but one species of section Helmiopsis 
have relatively long, e to recurved branches, 
P have 
la lo f 
loculicidal capsule, which varies in je a color. 
According to Arénes (1959), one of the distinctions 
s that 


the locules of the former are internally glabrous rather 


between Helmiopsis and Helmiopsiella Arénes i 
than pubescent. However, mature fruits of Helmiopsis 
boivinii re Arénes (including the type of H. 
pe of the 


pubescent cs on the basalmost portion of the 


inversa, the t genus) have small white- 
septa and the central portion of the base of the fruit. 
Ovules and seeds of Helmiopsis s. str. usually bear an 
elongated apical wing, at least as long as the seed 
proper; one species, H. sphaerocarpa, has at most an 
inconspicuous ventral ridge. Seeds may number one or 
two per locule; if two, one is generally larger and 
better developed. Fruits and mature seeds of some 
taxa are not 


TAXONOMIC TREATMENT 


he major taxonomic issue in defining Helmiopsis is 
whether it shall be circumscribed so as to encompass 
Dombeya rigida, as proposed by Dorr (2001), and 
presumably its sister species earifolia, or 
whether these species shall be excluded, following 
the traditional circumscription (sensu Arénes, 195 
Only the winged seed, often oppositisepalous stamens, 
and glandular petals supported placement within 
Helmiopsis rather than Dombeya. Winged ovules were 
(2001), 


but it is not clear that the curved, shallow marginal 


taken to be the most critical character by Dorr 


wing of D. rigida is homologous with the narrow, 
prominently elongated apical wing of most Helmiopsis 
species (excluding H. sphaerocarpa). The seeds of D. 
linearifolia are not winged. If the petal glands were 
homologous to those in Helmiopsis, it would reflect 
specific affinities to section Glandulipetalae, with 
which D. rigida and D. li 
that are shorter than the staminodes and normally 10 


. linearifolia also share stamens 


in number. One would therefore expect the latter 
species fo o some of the features that are 
otherwise common to all species of Helmiopsis, and 
presumably Snnt within the genus. However, 
they in fact lack several such features, including 
lepidote pubescence on at leas gans, 
glandular tissue on the sepals, and rapidly caducous 
bracteoles (Table 1). Furthermore, the members of 
section Glandulipetalae usually have oppositipetalous 
a rigida and D. 


stamens. inearifolia also 


ombey. 
display characters that are seen in other Dombeya 
species but not in Helmiopsis, such as cymes with 
scorpioid branches. Thus, no matter where these 
species are placed, considerable morphological di- 
t be postulated. As 


noted, patterns of morphological variation within 


vergence from close relatives mus 
Dombeya are complex and difficult to interpret, and 

e monophyly of the genus is not proven. Pending the 
availability of molecular data that may address these 
issues (C. Skema, pers. comm.) it seems most 
conservative to leave D. rigida and D. linearifolia 
within Dombeya subsect. Rigidae. It is plausible that 
to be embedded within 


Dombeya as presently defined, although the status of 


Helmiopsis may prove 


Helmiopsis as a natural group worthy of recognition at 
some level is not in question, given the presence of 
several apparent morphological synapomorphies. 
Arénes’ (1956, 1959) infrageneric classification is 
modestly rearranged herein. The distinction between 
section Glandulipetalae and section Helmiopsis (both 
within subgenus Helmiopsis, according to Arénes) is 
easily made by vegetative as well as reproductive 
morphology, and recognition of these sections is 
opsis hily, the single species placed 
s Phelmiosis, for th 


most part typical features of section Helmiopsis and is 


maintained. Helmi 
y Arènes within subgenu has the 
here included within that section; the reduction in 
carpe mber to usually three is certainly an 
cse AE there is no reason to suppose that 
this species is sister to the remainder of the genus. No 
division of the genus at the subgeneric level is 


data, species are defined pragmatically by the 
recognition of unique suites of morphological charac- 
ters. Nine species and one subspecies are recognized, 
of which one species and the subspecies are newly 


described. 


Volume 96, Number 4 
2009 


Applequis 
Revision A Helmiopsis 


Table 1. 


Morphological distinctions among Helmiopsis sect. Glandulipetalae (four species), Helmiopsis sect. Helmiopsis 


(five species), and Dombeya subsect. Rigidae (Dombeya rigida and D. linearifolia). 


Helmiopsis sect. Glandulipetalae 


Helmiopsis sect. Helmiopsis Dombeya subsect. Rigidae 


Vegetative, lepidote throughout 
inflorescence, and 


sepal pubescence 


Inflorescence Bion) branching 
cymes or geminate 
Bracteoles d. caducous 


Sepal glands present in small, inconspicuous 
patches inside base 

present 

+ opposite petals 


Petal glands 

Stamen position 

Androecial conformation fertile stamens 10, shorter than 
staminodes 


Ovary pubescence ote 
Seed morphology bearing an elongated apical 
(where known) wing, or a narrow keel in one 


species 


usually lepidote aor ies stellate throughout, or 
l sometimes with scurfy 
scales on twigs 
cymose, d or solitary, 


always small 
pec node caducous; 


cymes with scorpioid 


ranches 
often persistent in bud; 
e sometimes relatively large and 
a not enveloping oad, convex, sometimes 
bud enclosing 
present, sometimes in absent 
large noticeable patches 
absent present 
+ opposite d except + opposite sepals 
in one specie: 
fertile stamens ie to 30, fertile stamens 10, shorter 
longer than staminodes than staminodes 
except in one species 
lepidote stellate 
bearing an elongated bearing a shallow wing 
apical wing around most of seed 
margin, or wingless 


Helmiopsis H. Perrier, Bull. Soc. Bot. France 91: 
230. 1944. TYPE: Helmiopsis inversa H. Perrier. 


Trees or shrubs; twig bark grayish to dark reddish 
brown, shallowly ridged, with small pale lenticels; 
young twigs lepidote or occasionally glabrous. Leaves 
alternate, stipulate, petiolate, simple, variable 
shape; stipules small, = lanceolate, lepidote, rapidly 
caducous; petiole lepidote or occasionally glabrous; 
leaf base rounded to shallowly cordate or rounded- 
truncate; margins entire to weakly undulate, weakly 
repand, crenulate, or crenate-serrate; at least the 
abaxial surface usually lepidote, occasionally gla- 
brous or the adaxial surface sparsely stellate-pubes- 
cent. Inflorescences cymose, ranging from small and 
sometimes umbelliform to well developed with several 
orders of branching, or geminate, usually lateral, 
occasionally terminal or pseudoterminal; bracts where 
present small, rapidly caducous; all parts of inflores- 
cences lepidote. Flowers actinomorphic, hermaphro- 
ditic, pedicellate; pedicels angular or occasionally 
subterete, articulated at the point of insertion of 
bracteoles; bracteoles 3, small, lanceolate to ovate, 

idly cad medi aeh 
bud or well below bud on pedicel; sepals 5, 


le Kr 
below 


free except at extreme base, lanceolate, lepidote, 


ucous, inserte 


normally bearing at least small patches of glandular 
i r the base; petals 5, 
or orbicular, frequently 


tissue on the adaxial surface nea 
free, obovate to obdeltoid 


caducous, in section Glandulipetalae bearing small 


round glands on basal half of adaxial surface; 
androecium basally fused into a short corona; fertile 
stamens 1 (very rarely 5 in aberrant 
individuals), borne in 5 groups alternating with 
spatulate to oblong or oblong-obovate staminodes; 
stamens may be shorter than staminodes and + 
oppositipetalous or longer than staminodes and 
usually + oppositisepalous, with filaments of the 
latter group sometimes basally fused into fascicles; 
anthers linear to oblong; gynoecium (3- to)5-carpel- 
late; ovary syncarpous, superior, lepidote, sometimes 
lobed; style ieee basally or throughout; style lobes 
(3 to)5, short, erect, and clustered, or well developed 
and usually recurved to spreading. Fruit, where 


nown, a loculicidal capsule, variably shaped, 


sometimes lobed; seeds 1 or 2 per locule, apically 


winged or occasionally unwinged. 


Distribution and habitat. 
nine species, all endemic to Madagascar. The greatest 


Helmiopsis comprises 


diversity within the genus is at the northern tip of 
the island, but species of section Helmiopsis are native 


substrates include sand, limestone, alluvial soil, and 
asalt. 
Key To SECTIONS OF HELMIOPSIS IN MADAGASCAR 


la. Leaves elliptical to ovate, lanceolate, or oblong; 
petioles relatively short, less than 15(-28) mm 


Annal 
Missouri Botanical Garden 


long; inflorescences l- to 5-flowered; petals no 
ction Henig 
sely 


E 
- 
o 
b 
e 
[n] 
u 
- 
[em 
E 
is) 
m 
= 
9 t 
55 
£2 
E 
= 
E 
e 
E 
E 
m 
ixl 
o 
5 
c 
E 
m 
i] 
üu 


elliptical; petioles usually over 1 
centimeters long; DEM variable; petals 
glandular; stamens section eo edi 


I. Helmiopsis sect. Helmiopsis. 


dr on xc Phelmiosis Arénes, Bull. Mus. Natl. Hist. 
Nat., sér. 2, 28: 415. 1956. 

Trees to 10 m high or shrubs, sometimes large; twig 

bark grayish to dark reddish brown, shallowly ridged, 


with small pale DE young twigs lepidote or 


species; margins 
entire to weakly undulate; abaxial surface lepidote or 
glabrous; adaxial surface lepidote, glabrous, or sparsely 
pubescent with flat stellate hairs; stipules where known 
less than 1.5 mm, rapidly caducous. Inflorescences 
lateral, geminate or 3- to 4(to 5)-flowered cymes, 
sometimes umbelliform, or 1-flowered, often clustered 
near twig apices; occasional bracts very small, rapi 
caducous; bracteoles rapidly caducous, sometimes 
ovate and persisting in smaller buds; pedicels angular 
articulated; lepidote 
throughout. Sepals 
with sometimes dee ae patches of glandular tissue 
petals obovate to 
base of 


to subterete, inflorescences 


narrowly lanceolate, lepidote, 

n of adaxial surface; 
aiid a or biela not glandular; fused 
androecium coroniform, sometimes almost Den 
fertile stamens 15 to 30, usually + oppositisepalous, 


in one species frequently oppositipetalous; staminodes 
spatulate to oblong or oblong-obovate, shorter than 
ongest stamens; filaments free to corona or basa 

fused into fascicles; anthers linear to oblong; gynoecium 
(3- to)5-carpellate; ovary lepidote, sometimes lobed; 
style lepidote basally or throughout; style lobes well 
developed, recurved to spreading or erect and 
appressed in one species. Fruit where known, a capsule, 
variable in shape, sometimes lobed; seeds where known, 


1 or 2 per locule, apically winged. 


Distribution. Helmiopsis sect. Helmiopsis ranges 


northern end of Madagascar to the ari 


from 
ud (Fig. 


Discussion. | Helmiopsis glaberrima, a rare northern 
species for which only one specimen is known, 
complicates the distinction between section Helmiop- 
sis and section Glandulipetalae. lts affinities are 
clearly with section Helmiopsis, as it has narrow 
leaves, 15 oppositisepalous stamens, small inflores- 
cences, well-developed sepal gland patches, and long 
style lobes. 

ormally five basa 
petioles, fertile stamens shorter than the staminodes, 


However, it displa characters 
l le 


ys a fe 
veins, relatively long 


a 
5 


and erect, appressed style lobes) that are otherwise 
typical of section Glandulipetalae. The hypothesis 
tentatively favored A this author is that H. glaberrima 
is sister to the remainder of section Helmiopsis, and 
that the few a shared with section Glandu- 
lipetalae are plesiomorphic within Helmiopsis. How- 
ever, an alternate hypothesis is that section Helmiopsis 
is paraphyletic, with H. glaberrima actually being the 
sister taxon of section Glandulipetalae. 


KEY To THE SPECIES OF HELMIOPSIS SECT. HELMIOPSIS IN MADAGASCAR 


la. Leaves glabrous; stamens shorte 


r than staminodes; style branches erect 


3. H. glaberrima 


lb. Leaves lepidote at least ala stamens longer than staminodes; style branches a or recurved. 
b 


Ww 
WU 


abaxial surface 


stellate or MAU abaxial surface variably lepidote; stamens (15 to)20 to 30; ovary and fruit 5-carpellate, not 
lobed. 


conspicuo 
3 


INE 


[peus narrowly ovate to lanceolate, adaxial surface weakly ie n flat stellate hairs; sepals 11 


SH calccola 


4a. Leaves ovate to elliptical, broadly ovate, or ape pedicels (6-)9- un 19) mm r D much 
obovate 


elongated in 


4b. Leaves 2 elliptical to elliptical or lanceolate, rarely ovate; pedicels 3-6 mm long, in fruit up to 


m; petals 8.2-9.3 m 


ju. wroiden UR 


1. Helmiopsis boivinii (Baill.) Arènes, as “Boivini,” 
Bull. Mus. Natl. Hist. Nat., sér. 2, 28: 418. 1956. 
Basionym: Trochetia boivinii Baill., as *Boivini," 


Adansonia 10: 109-110. 1871. Dombeya boivinii 


m long, suborbicular; stamens 30; staminodes oblong-obovate, 2.7-3.1 mm 
5. 


H. polyandra 


(Baill.) Baill., as *Boivini," Bull. Mens. Soc. Linn. 
Paris 1: 496. 1885. TYPE: Madagascar. Maha- 
janga: Ambongo, 14 Feb. 1841 (fl), A. Pervillé 
642 (holotype, P!; isotypes, G not seen, P!). 


Volume 96, Number 4 Applequis 527 


2009 Revision a Helmiopsis 
42*00"E 43*0'0"E 44*0'0"E 45°0'0"E 46°0'0"E 47*0'0"E 48°0'0"E 49°0'0"E 50°0'0"E 51*0'0"E 
L iT L L L L L 1 L L 
12*00"s-] [-12700"S 
1300" 13200" 
14700" $4 14200" 
1500" 1500" 
1600" = 165°0'0"S 
1700'$4 17°0'0"S 
1800" -1300"8 
1900" =19°0'0"S 
20*00'$-] [-20700"S 
21*00'S-] -2100's 
22'00"$-] P22°0'0"S 
23'00'S-] =23°0'0"S 
240084 -2400"S 
25'00'$-] [-25^00"S 
i LU ' Li 1 J 1 I ' 1 
42°0'0"E 43°0'0"E 44*0'0"E 45"0'0"E 46°0'0"E 47°0'0"E 48°0'0"E 49*0'0"E 50*0'0"E 51*00"E 


Figure l. Approximat tribution of species of Helmiopsis sect. Helmiopsis ( = H. boivinii; ME = H. calcicola; + = H. 
glaberrima; A = H. hily subsp. hily; Wl = H. hily subsp. boinensis; € = H. polyandra). 


Helmiopsis ow s Pe Soc. Bot. France 91:230. caducous; petiole 7-15(-18) mm, densely lepidote; 
Madagascar. Mahajanga: 

Aranana sai E poo E] au N de Majunga 
[Mahajanga], 27 May [19327] (fr), H. Perrier de la : 
Báthie 17986 ee. designated here, P!; isotypes, roun ed tip, occasionally slightly i oe 
P!, TAN!). entire to slightly undulate; both surfaces lepidote in 


basal veins 3(to 5), the lateral weak; base rounded to 
very shallowly cordate; apex rounded or acute with 


Shrub to 4 m high or tree to 10 m high, 20 cm young leaves, Rs so in older on the adaxial 
diam.; bark grayish, sometimes splotched, with pink- ^ surface sometimes becoming glabrous; scales of leaves 
orange slash; twig bark grayish, shallowly ridged, with and petioles variable in size and shape, occasionally 
small pale lenticels; young twigs lepidote. Leaves  fimbriate; stipules less than 1 mm, lepidote, rapidly 
ovate to elliptical, broadly ovate, or lanceolate, (2.5-) caducous and very rarely present. Inflorescences up to 


3.6—7.3(-8.4) X (1.4-)1.7-3.8(-4.5) em, frequently — 4(to 5)-flowered cymes, often umbelliform, or gemi- 


Annals of the 
Missouri Botanical Garden 


nate or flowers occasionally single, borne laterally, 

n near twig apices; usually only 1 or 2 flowers 
r inflorescence maturing; bracts less than 1.2 mm, 

deltoid, l l 

articulated, (6-)8—16(-19) mm, becoming woody in 


rapidly caducous; pedicels + angular, 
fruit; bracteoles 3, borne near base of flower, deltoid, 
caducous, ca. ; inflorescences lepidote through- 
out. Sepals ae 8 lanceolate, (6.5-)8.5-11 mm, 
pale green to whitish, lepidote, the adaxial surface 
basally glandular; petals white, (6210-13 X (4—)6— 
mm, obovate to obdeltoid; androecial corona 0.5— 

1 mm; fertile stamens (15 to o 25, in groups of (3 
to)4 to 5, opposite sepals, white; filaments 2.5— 
4.8 mm, often free to corona, occasionally fused at the 
base into distinct fascicles; anthers linear-oblong, 
2.3-3.5(-4.1) mm; staminodes spatulate, 2.8—4(—5) 
mm, shorter than longest stamens; gynoecium 5- 
carpellate; ovary lepidote, white; style lepidote mostly 
; lobes erect and 


on lower portion, 3-5 mm; style 


spreading or rarely recurved, (0.6-)1-2.3 mm. Fruit a 
woody capsule, (1113-19 mm, ovoid to cylindrical, 
sometimes a apparently swollen base or constric- 
tion à e base, sometimes 5-ribbed or shallowly 

brown to blackish brown at 


maturity; central base of capsule inside and basalmost 


ove 
lobed, ecu pale 


with often asymmetrica 5 mm apical 


wing, 2 per locule, only 1 well developed. 


Distribution, phenology, and habitat. Most collec- 
f limited of 


tions are from a limited area orthwestern 
Madagascar, including localities around Mahajanga, 
the Ankarafantsika reserve, and the Ambongo region 
(Fig. 1); only a few clearly distinct populations have 
been collected. The habitat is on sand, including 
coastal dunes, in scrub or forest; reported altitudes 
range up to 200 m. Existing flowering collections are 
from November to February, with buds seen in May; 
fruiting may occur at almost any time of year, with 
developing or undehisced fruits collected in February, 
March, June, August, and November. 


IUCN Red List category. Provisional IUCN Red 
List category (IUCN, 2001) is noted as Least Concern 
(LC) 


Com. uses. Mainaty (Réserves 
Naturelle 2067, 2568, 2832, 2944; Service Forestier 
12666); Maintinaty (Service Forestier 14785); Amonty 
(Service Forestier 50); Selivato (Service Forestier 
82) e woo s been used in construction 


(Service Forestier 12666). 


Orthography. Under Article 60.11 of the ICBN 
(McNeill « et al., 2006), the original spelling (“boivini”) 
of the epithet of Helmiopsis boivinii is treated as an 


error to be corrected, as Arénes (1959) did in his Flore 
de Madagascar treatment. 


Discussion. There is no qualitative distinction 
between Helmiopsis boivinii and H. inversa, which 
are native to the same region of northwestern 
Madagascar. The more frequently collected popula- 
tions that have been called H. 
considerable variation in such characters as leaf size 


inversa display 


morphology, 
and fruit shape. However, these characters do not vary 


and shape, leaf pubescence and scale 


in concert, nor are they consistently correlated with 
locality of origin. In 1963, Capuron determined 
specimens of H. inversa at P as H. boivinii, indicating 
that he thought the two taxa should be combined, but 
does not appear to have published this treatment. 
Helmiopsis boivinii m the invalidly published H. 
inversa var. arenicola H. Perrier were each described 
based on single specimens from Ambongo, an area 
where classic H. inversa has not been collected, and 
are distinguished by their somewhat larger leaves and 
smaller flowers. However, the increased leaf size and 
reduced petal size seen at Ambongo do not signifi- 
cantly exceed the extremes from collections at other 
repeatedly collected localities. Moreover, the types of 
these two putative taxa appear to be in early rather 
than full flower, and it is possible that the availability 


apparent petal size 
inadequate reason to formally distinguish the Am- 
bongo population from other populations in the region 
at any taxonomic level. 

Although Helmiopsis inversa is the type species for 
the genus, the epithet of H. boivinii predates its 
publication and thus has precedence when the two are 
combined. Perrier de la Bathie (1944: oe duet two 
17986 a 
"[t]lypes du genre et de Pespéce [H. eee p d 


specimens, Perrier de la Báthie 


latter number seems to be a misprint, as the only 
similar Perrier specimen at P, from the same locality 
(Ampasimana au N de Majunga) and date (July 1921), 
is clearly numbered 13838. Arénes (1956b, 1959) 
listed Perrier de la Báthie 13838 and 17986 as types 
in view of this fact. It appears that neither specimen 
has ever been selected to serve as a single lectotype. 
Perrier de la Báthie 17986 is herein chosen because it 
with buds as well as old fruit, 
and three duplicates exist; this choice also avoids any 


is in better condition, 


argument over whether the specimen whose number 
was misstated can be considered original material. 


Representative specimens 
Mahajanga: several km 


examined. MADAGASCAR. 
N of Majunga [Mahajanga; 


15%43'S, 046°19’E], vic. Zaha Motel, Barnett et al. 47 
MO, TAN); Namakia-Dunes de Boeny Ampasy, S/P. 
Mitsinjo [15754/S, 045%50'El, Debray 1442 (Py dunes, 


Ampasimarina [15731'S, 046°30’E] au N de Majunga 


Volume 96, Number 4 
2009 


Applequis 
Revision 7 Helmiopsis 


[Mahajanga], Perrier de la Bathie 13838 (P); n Réserve 
forestière d'Ampijoroa, Jardin Botanique B, á VE du La 
eae 16°18'19"S, 046749'38"E, Rann et al. 399 
arafantsika, canton Tsaramandroso, distr. Ambato- 
Been Inve 16°10" S, ro El pos Maia 2067 
(MO, P, TAN) 78 Réserve [Ankarafantsika], Ligne Bépilo, 
Service Forestier 50 (P); W du village de Bevazaha [16°14’50"S, 
De canton de Tsaramandroso, distr. d' Ambato-Boeni, 
rvice Forestier 12666 (P); forêt Analalava, Maintirano, Service 
le xd (MO); forét tout prés de Maherivaratra, village 
le plus proche Ambodibonara, canton Ankerika [14^57'S, 
047° 52'E] distr. Antsohihy, Service Forestier 15782 (P, Ped 
prés d'Ampasimariny [15731'S, 046°30’E], au NE de Majunga, 
Service Forestier 24317 (MO, P) 


Io] 


Inst. Sci. 
. Vég. 7: 55—56. 1956. 


agascar. a forêt tropo- 


2. Helmiopsis pow d rd Mém. 
Madagascar, Sér. B, 
TYPE: Ma 


phile sur rocailles calcaires, Kama-Kama, sur le 
plateau d'Ankara (Boina) [16755'S, 46729'E], v 

m alt., Jan. (fL), Perrier de la Báthie 
1018 (holotype, P!; isotypes, P [2]!). 


, 2-8 m high, with hard wood; twig bark 
mes below. ridged, with small pale lenticels; 
young twigs yellowish, lepidote. Leaves narrowly ovate 
to lanceolate, (3.5—)6—8.5  (1.3-)2.3-3.5 cm, cadu- 
cous, membranous; petiole 4-12 mm, densely lepi- 
dote; basal veins 3, the lateral weak; base rounded; 
apex acute; margins entire; abaxial surface lepidote, 
adaxial surface sparsely pubescent with flat stellate 
hairs, especially along midrib; stipules ca. 1 mm 
deltoid, lepidote, usually caducous. Inflorescences 
few-flowered, axillary, probably mostly op with 

eduncle less than 2 mm; pedicels + 
articulated, ca. 10— 
base of ioner 


gular, 
6 mm; bracteoles nce below 
peduncles and pedicels lepidote. 
11-13.2 mm, lepidote, 
the basal half of the adaxial surface glandular and 
reddish; petals pale yellow with pale pink highlights, 


Sepals narrowly lanceolate, 


14-17 X 9-11 mm, obovate with asymmetrical apex; 
androecial corona 0.6—1 mm; fertile stamens ca. 30, 
in groups of (5 to)6(to 7), opposite sepals; mes 


staminodes 
oblong, 3.5—4.8 mm; gynoecium 5-carpellate; ovary 
lepidote, 5-ribbed; style pale yellow, lepidote through- 
out, 4—4.5 mm; style lobes recurved, 3-4 mm. Fruit 
unknown. 


Distribution, _ Phenology, and habitat. The only 
the Ankara region of 
n Madagascar (Fig. 1); the species was 
nue in fae 1900. 

IUCN Red List category. Provisional IUCN Red 
List category (2001) is noted as Data Deficient (DD). 
This area has been inadequately collected, so a lack of 
collections is not necessarily evidence of rarity. 


However, the fact that the region has suffered 


considerable habitat degradation is cause for concern. 


Discussion. This is the only species of Helmiopsis 
that has more or less stellate pubescence in addition 
to lepidote scales; the stellate pubescence is sparse 
e adaxial leaf surface. Arénes 

on H. inversa var. calcicola 
H. Perrier (Bull. Soc. Bot. France 91: 230. 1944), 
which was not validly published, as the description 
was given only in French, with no Latin diagnosis, 
which the IC equired for taxa published after 
935 (Art. 36.1; McNeill et al., 2006). A 


was accompanied by a full Latin description and so 


rénes’ name 


was validly published. 


3. Helmiopsis glaberrima Arénes, Bull. Mus. Natl. 
Hist. Nat., 2, 28: 417. 1956. TYPE: 
Madagascar. Antsiranana: Analamerana, Diégo- 
Suarez [12°45'S, 49°32’E], 16 Mar. 1954 (fl), 
Service Forestier 9423 (holotype, P!; 
MOI, P not seen). 


sér. 


isotypes, 


Woody plant; twigs dark reddish brown, shallowly 
ridged, glabrous, with pale tan oo lenticels. 
Leaves a 
petiole (4—)13-28 mm, glabrous, de. at ends of 
twigs much shorter; basal veins 5, the lateral ones 


lanceolate-oblong, 4-7. 


weak; base shallowly cordate; apex rounded-acute; 
both 


Inflorescences geminate or 3-flowered 


margins entire; surfaces glabrous; stipules 
caducous. 
cymes, lateral and pseudoterminal, clustered at twig 
ends; peduncle 11-17 mm, dark reddish brown, 
woody, sparsely lepidote; pedicels articulated, nearly 
terete, lepidote, ca. 6.5-8 mm; bracteoles inserted 
well below flower. Sepals narrowly lanceolate, 8.2— 
9 mm, lepidote, the basal half of the adaxial surface 
glandular; petals 12-14 X 9-10 mm, broadly obovate 
with clawed base; androecial corona 1.5-1.9 m 

15, in eon of (2 to)3(to 4), 
oppositisepalous; filaments 


fertile stamens ca. 

l mm, fused into 
2.5-3.1 mm; 
spatulate, 6.5—7.5 mm; gynoecium 5-carpellate; ovary 


ascicles; anthers linear, staminodes 
deeply 5-lobed, lepidote; style lepidote only at base, 
4.5—5.1 mm; style lobes erect and closely appressed, 
1.2-1.4(-2) mm; ovules elongated, apparently apically 
winged. Fruit unknown 


Distribution, phenology, and habitat. The only 


known collection is from the extreme north of 


Madagascar (Fig. 1); it was flowering in March. 


IUCN Red List category. Provisional IUCN Red 
List category (2001) is noted as Vulnerable (VU D2). 
Only a ion i 
Special Reserve, area of occupancy is likely 


to be very small Several remaining patches of 


Annals of the 
Missouri Botanical Garden 


northern forest are relatively accessible and have 
been repeatedly collected; multiple specimens of 


petalae have been collected from certain areas. Thus, 
the lack of additional collections of this species 
suggests that it is indeed relatively rare and of limited 
distribution. 


Discussion. Helmiopsis glaberrima is the only 
species of section Helmiopsis in which the staminodes 


are longer than the fertile stamens, as in section 


closely appressed, though much longer than those in 
section Glandulipetalae, and the only one with 
petioles mostly more than 1.5 cm long and leaves 
mostly 5-veined at the base. In addition, it is easily 
distinguished from the remainder of section Helmiop- 
sis by its completely glabrous leaves. 


4. Helmiopsis hily Arénes, Bull. Mus. Natl. Hist. 
Nat., sér. 2, 28: 415. 1956. TYPE: Madagascar. 
Toliara: Betsipotika, Analaiva, Morondava 
[20°17'S, 044°27'E], 5 Apr. 1953 (fl), Service 
Forestier 7226 (holotype, P!; isotype, TEF not 
seen) 


Tree to 10 m high (to 20 m high and 0.5 m diam. in 
subspecies boinensis), or occasionally large shrub to 

m high; twig bark dark grayish brown, shallowly 
ridged with small rounded pale lenticels; young twigs 
ferruginous, densely lepidote. Leaves borne directly 
on long twigs or singly or clustered on short shoots, 
narrowly lanceolate to lanceolate, (2.6—)3—7.3(—7.7) 
em X (7-)9-16(-21) mm; petiole (1.5-)5-12(-13.5) 
mm, densely lepidote; basal veins 3, the lateral weak; 
base rounded; apex acute with rounded tip, occasion- 
ally rounded or emarginate; margins entire, often 
slightly irregular; adaxial surface glabrous, abaxial 
surface densely lepidote, pale or silvery green, usually 
with scattered tiny brown spots due to pigmentation of 
some large scales; midrib conspicuous, secondary 
venation pinnate, weak; stipules less than 1 mm, 
deltoid, lepidote, 
rarely present. Inflorescences lateral, geminate or 


immediately caducous and very 


occasionally a 3-flowered cyme or flowers solitary; 
peduncle (12-11 mm; pedicels angular, articulated 
near bud, (4—)7-16 mm; bracteoles rapidly caducous; 
ao and pedicels lepidote; only 1 flower or all 2 

owers maturing to fruit. Sepals narrowly 
e 7-9.2(-10) mm, lepidote, base of ada- 
xial surface glandular-pubescent; petals creamy or 
yellow, (6.5-)8.5-13.5(-15) X (4.5-)5-7.5 mm, ob- 
ovate, caducous; androecial corona 1.2-2 mm; fertile 
stamens normally 15; filaments 2-4.5(-5) mm; anthers 
linear, 2.1-3.8(4.5) mm, often but not always clearly 


oppositipetalous; staminodes spatulate, 2.5—3.2(—4) 

mm, oppositisepalous or DM EU) tending toward 
oppositipetalous; gynoecium 3(to 4)-carpellate; ovary 
lepidote, 3(to 4)-lobed; style a for entire length, 
3.2-)4-5.8(—6.8) mm; style lobes strongly recurved, 


2-3 mm. Fruit a 3-valved, lepidote capsule, irregu- 


— 


larly elliptical to broadly oblong-elliptical or ovate- 
elliptical, deeply 3-lobed, (7.5-)8-15.6 mm; seeds 
where known per locule, ca. 3.6-3.8 mm, 
irregularly elliptical with a flattened ventral surface, 
with a tapering apical wing ca. 4.2-5.1 mm 


Distribution. | Helmiopsis hily is native to forests in 
the arid southwestern portion of Madagascar and 


(subspecies boinensis) the northwestern Boina region. 


IUCN Red List category. Each subspecies is 
assessed separately, below. 


Common names. Hily (Service Forestier 7226, 
9819); Pisopisonala (Service Forestier 3398, 4984); 
Andriambolafoty (Service Forestier 14294), Menahy 

). 


(Service Forestier 157 


Discussion. Almost all flowers, and all maturing 


fruits seen, are 3-carpellate; the presence of four 


ate; 
e lobes and style branches (as reported by Arénes, 

959) appears to be an abnormal condition Arénes 
"em 1959) placed this species alone within 
Helmiopsis subg. Phelmiosis due to its reduced carpel 
as the 


reduction in carpel number seems very likely to be 


umber. This treatment is unwarranted, 


5 


not a d character within the genus, but 
orphy, which has no 
question of EA to other species; H. hily 


an autapomo relevance to the 
otherwise has for the most part typical features of 
section Helmiopsis 


KEY TO THE SUBSPECIES OF HELMIOPSIS HILY IN MADAGASCAR 


la. Tree to 10 m high or large shrub; leaves (7— 9- 16 


mm long; capsule (7.5— E 10.6 mm e 

to southwestern Madagasca reus hily 
lb. Tree to 20 m high; mon to at east 2 A mm broad, 

often borne singly on short shoots 

often all flowers (2 to 3) of infl 


15.6 mm long; native to northwestern Madag 
c 4b. subspecies “oinensis 


4a. Helmiopsis hily Arènes subsp. hily. 


Tree to 10 m to 6m 


high. Leaves mostly borne directly on long twigs or 


high, or occasionally shrub 


clustered on short shoots, narrowly lanceolate to 
lanceolate, (2.623-7.3(-7.7) cm X (739-1618) 
(1.5-)5-12(-13.5) m 


flower per inflorescence maturing; EE (1-)3- 


mm; petiole 


Volume 96, Number 4 
2009 


Applequis 
Revision E Helmiopsis 


6(-11) mm; pedicels (4-)7-10(-12) mm. Sepals 
narrowly lanceolate, 7—9.2(-10) mm, lepidote, base 
of adaxial surface glandular-pubescent; petals creamy 
(6.528.5-13.5(-15) X (4.5-)5-7.5 mm, 
obovate, caducous; androecial corona 1.2-2 mm; 
filaments 2-4.5(-5) 


mm, often but not 


or yellow, 


always clearly oppositipetalous; staminodes spatulate, 
2.5—3.2(-4) mm, oppositisepalous or a 

ard oppositipetalous; gynoecium 3(to 4)- 
carpellate; ovary lepidote, 3(to 4)- lobed; pus lepidote 
for entire length, (3.2—)4—5.8(-6.8) mm; style lobes 
strongly recurved, 2-3 mm. Fruit elliptical to oblong, 
(7.5-)8-10.6 mm, deeply 3-lobed. 
ca. 3.6-3.8 mm 


ventral surface, with a tapering apical wing ca. 4.2— 


tending tow 


; seeds 1 per locule, 
, irregularly elliptical with a flattened 


1 mm. 


Distribution, phenology, and habitat. Helmiopsis 
hily subsp. hily is native to southwestern Madagascar 
(Fig. 1). One collection was noted to be on sand at 


600-850 m 


January to March (with one specimen flowering in 


altitude. Flowering usually occurs from 


October), as is typical in the arid southwest; immature 
fruit has been collected in March, and nearly mature 
fruit in 

IUCN Red List category. Provisional TUCN Red 
List category (2001) is noted as Least Concern (LC). 


Specimens examined. MADAGASCAR. Toliara: forét de 
Zombier [22°46'S, a forêt de l'Isalo, Humbert et al. 
29604 (MO, P); forét prés de l'Isahaina, entre Sakaraha et 
P. Service Forestier 527 (MO, P [2]: rubr 

rvice Forestier 5598 (P, TAN 


, Se 
2 00" S, 044° a E TAN). 


Troboampam Po Ma [21°31'S, 044919" E], 
Service reie 9819 (MO, P); forêt de Zombitsy [22°46'S, 
044°42'E], Serv Forestier 11908 (P [3] s 


[22746'8, OA IP E], S 
Betsako, Ankazoabo [22° 2l S, 044^42'E], Serv 
794 (Py; 


(P forét d 
Bekily et la rte. Beraketa—Antani- 
mora, Service Forestier 22531 (MO [2], P). 


4b. Helmiopsis hily Arènes subsp. boinensis 
Appleq., subsp. nov. TYPE: Madagascar. Maha- 
ka sur le plateau d'Anta- 
ni na) [16°28’S, 046°11’E], 12-14 
Nov. 1957 (fr.), Service Forestier 18423 (holotype, 
P». e 2: 


aec subspecies a Helmiopse hily Arènes subsp. hily 
statura majore, foliis leviter latioribus saepe singulatim ad 
ramunculos breves portatis, inflorescentiis leviter majoribus 
atque fructibus majoribus differt. 


Tree to 20 m high; many leaves borne singly on 


short shoots ca. 1 cm long; leaves lanceolate, 3.8— 
6.4 cm X 12-21 mm; petioles 7-13.5 mm. Usually 


both (or all three) fruits per inflorescence maturing; at 
11-16 mm. 
Flowers not known. Fruit broadly oblong-elliptical to 


maturity peduncles 2-11 mm, pedicels 


ovate-elliptical, 12-15.6 mm; seeds not known 


ape di Lupi d es habitat. The A 
known collec s from the northwestern Boi 
region of Nn d. ly mature fruits, n 
seeds already dispersed, were collected in November. 


IUCN Red List category. Provisional IUCN Red 
List category (2001) is noted as Data Deficient (DD). 


Discussion. Helmiopsis hily subsp. boinensis can 
be distinguished from H. hily subsp. hily by its larger 
size (trees to at least 20 m), slightly broader leaves 
frequently borne singly on short shoots, larger 
inflorescences, and larger fruits. However, it resem- 
bles subspecies hily in other morphological charac- 
ters; thus, lacking information on floral morphology or 
variation in vegetative morphology, the two are most 
conservatively treated as subspecies of a single 
species. A significant group of endemic species are 
confined to the area of northwestern Madagascar that 
includes the Boina region, e.g., H. calcicola, Dombeya 
ankarafantsikae Arénes and D. boeniensis Arénes 
li 1958), Perrierodendron boinense (H. Perrier) 

co (Lowry et al., 2000), Combretum boinensis 
Jongkind (Jongkind, 1995), and Dalbergia davidii 
Bosser & R. Rabev. (Bosser & poe aia a to 
name only a few. It would refore be worth 
investigating the possibility "a dis deum md 
distinetive population actually represents another 

ocally endemic species. 


5. Helmiopsis polyandra Appleq., sp. nov. TYPE: 

Madagascar. Prov. Antsiranana: forét d'Analafi- 

ana, au N de la basse Manambery (au SW de 

Vohém ir 13%21'30"S, 050° 00'30"ED, 

11 Mar 7 (fL) Service Forestier 27532 
(oloiype, zs isotype, MO!). Figure 3. 


pecies nova quae a Helmiopse boivinii (Baill.) Arénes 
foliis infra pallidis, pedicellis brevioribus, bracteolis long- 
ioribus tarde delapsis, petalis minus quam 10 mm longis 
suborbicularibus, 0 S staminodiis 
oblongo-obovatis 2.7-3.1 mm longis differt. 


staminibus ca. 


Tree to 8m high; 
shallowly ridged, with pale lenticels; young 


twig bark reddish brown, 


lanceolate, or rarely ova 
(1.1-}1.4—-2.5(-3.4) em; petiole 4-10(-13) mm, dense- 
ly lepidote; basal veins 3(to 5), the lateral weak; base 
rounded; apex r 


unded-acute or emarginate; margins 


entire to EM m abaxial surface lepidete, 


glabrous, sometimes bearing 
hairs of crystalline appearance, distinctly different 


Annals of the 
Missouri Botanical Garden 


E 
o 
N 


Helmiopsis hily Arènes subsp. boinensis Appleq. —A. Fruiting branch. —B. Old fruit. Based on the type, 


Figure 
Service Forestier 18423. 


from abaxial surface in color, usually with visible 
venation; stipules ca. 1.2-1.5 mm, lanceolate, lepi- 
dote, caducous and rarely present. Inflorescences 
geminate or 3- to 4-flowered cymes or flowers solitary, 
axillary; peduncles 2-6.5 mm; pedicels + angular, 
articulated, 3-6 mm, in fruit woody to 10 mm; bracts 
few, ca. 2.6-3.5 mm, caducous; bracteoles inserted 
slightly beneath bud, 1.3-1.7(-2.4) mm, 
caducous but often persisting on smaller buds; 
inflorescences lepidote throughout. Sepals narrowly 


ovate, 


lanceolate, 7.5-8.2 mm, lepidote, the basal portion of 
the adaxial surface glandular; petals 8.2-9.3 X 7.5— 
8.5 mm, suborbicular; androecial corona virtually 
absent; fertile stamens ca. 30, in groups of 6, 
oppositisepalous; filaments 2-3.5 mm; anthers ob- 
long, 2.3-3.1 mm; staminodes oblong-obovate, 2.7— 
3.1 mm; gynoecium 5-carpellate; ovary lepidote; style 
pale yellow, lepidote throughout, ca. 3 mm; style lobes 
erect with recurved apices, ca. 2.4—2.5 mm; ovule 
shortly winged. Mature capsule ca. 20 mm, roughly 


Volume 96, Number 4 
2009 


Applequist 


Revision of Helmiopsis 


Figure 3.  Helmiopsis polyandra Appleq. —A. Flowering branch. —B. Flower. —C. Cluster of stamens. —D. Gynoecium. 
Based on the type, Service Forestier 27532. 


ovoid or cylindrical, grayish brown, with irregular 
surface, lepidote; seeds unknown. 


Distribution, phenology, and habitat. "This new 
species is known only from a few sites in the extreme 
northeast of Madagascar (Fig. 1); one collection is 
recorded from 300 m altitude. Buds and flowers have 
been collected in March; the only fruit seen, also 
collected in March, was old fruit that remained on the 
plant from the previous flowering period. 


IUCN Red List category. Provisional IUCN Red 
List category (2001) is noted as Vulnerable (VU D2). 


Only three distinct populations have been collected; if 
area of occupancy were calculated using a 4 km? cell, 
as suggested by IUCN (2001), the range of this species 
would be only 12 km?. It seems to be limited to a 
relatively small portion of Madagascar that, like much 
of the country, has suffered significant deforestation 
(e.g., Du Puy & Moat, 2003). 

Discussion. Known specimens of this species have 
been previously identified as Helmiopsis hily, appar- 
ently due to their relatively narrow leaves and variable 
color differences between leaf surfaces. However, the 


Annals of the 
Missouri Botanical Garden 


affinities of H. polyandra to H. boivinii seem much 

greater; for example, H. hily has strongly lobed ovaries 

and relatively few stamens. The former two species 

also share a geographic range in the extreme north, 

whereas H. hily is a southern and western species. 

distinguished from H. 
l 


whose 


Helmiopsis boivinii ma 
b ovate 


y be 
polyandra its frequently 
abaxial surfaces are not so distinctly paler than the 
adaxial surfaces; its pedicels are (6—)8-16(-19) mm 
long, and its petals are (6— mm long, obovate to 
obdeltoid rather than suborbicular. Additionally, H. 
boivinii has only (15 to)20 to 25 stamens, and the 

taminodes are spatulate and 2.8—4(—5) mm long; the 
smaller (ca. 1 mm) and are more 


eaves, 


sen es may be 
frequently lost even on very small buds. 

Service Forestier 23078bis is a sterile specimen 
whose leaves, sparsely lepidote on both surfaces at 
maturity, are much larger than those of the three 
collections of Helmiopsis polyandra from closer to 
Iharana. As its appearance (including leaf shape, 
venation, bark, and stipule morphology) is diro 
consistent with that of the known specimens of H. 
polyandra, it is placed therein. 


Paratypes. MADAGASCAR. Antsiranana: lisiére supé- 
rieure de la forêt d'Andranomadiro [12738'S, 049°28’E] 
(rebord S du plateau de Sahafary, entre les bassins de la 
Saharenana et E Rodo [Irodo)), pm Forestier 23078bis 


e 
s [13° 3 n 049 ^54'E] (entre la Fanambana et la 
Manambery), es inférieures de la rive droite de 
l'Andilana, a Forestier 27540 (MO, P). 


Helmiopsis Meter urs Arénes, 
Bull. Mus . Hist. Nat, sér 8: 416. 
1956. TYPE: pon m o 
Capuron ex Arénes, designated here [= 


— 


II. 


sect. 


beya pseudopopulus Baill.]. 


Trees, sometimes large, or large shrubs; twig bark 
grayish to dark reddish brown, shallowly ridged, with 
small pale lenticels; young twigs usually lepidote or 
bearing scurfy scales. Leaves broad, ovate to broadly 
elliptical, orbicular, or occasionally obovate; ES 
lepidote; basa to 7, well developed; 

shallowly cordate to rounded-truncate or rounded; apex 


veins 


variable; margins entire to weakly repand, crenulate, or 
crenate-serrate; at least the abaxial surface lepidote; 
stipules small, lanceolate or scurfy, rapidly caducous. 
Inflorescences lateral and sometimes terminal, cymose 
with up to several orders of branching or geminate; 
bracts lanceolate, immediat tely caducous; bracteoles 
lanceolate, inserte ow bud, 
immediately CUR pedicels + angular, sometimes 
articulated; inflorescences lepidote throughout. Sepals 
lanceolate, lepidote, normally bearing small inconspic- 
uous glandular patches at the very base of the adaxial 


mediately or we 


surface; petals white to creamy or yellowish, obovate, + 

glandular on lower portion of adaxial surface; fused 
base of androecium coroniform; fertile stamens nor- 
mally 10 (rarely 5 or 15 in aberrant flowers), borne in 
pairs, + oppositipetalous; filaments short, anthers 
linear; staminodes spatulate, longer than fertile 
stamens; gynoecium 5-carpellate; ovary lepidote, often 
5-lobed; style lepidote basally to throughout; style 
lobes erect, clustered. Inflorescence oua usually 
sap g woody uit; fruit a 

apsule, ovoid or pica pd s. de seeds (1 

to)2 per locule, only 1 well developed, with a prominent 


and becomin: 


apical wing or unwinged in 1 species. 
Distribution. — Helmiopsis sect. Glandulipetalae is con- 

fined to forested areas in the extreme north of Madagascar 

(Fig. 4). A variety of substrate types are reported. 

Arénes (1956b, 1959) did not specify 

a type species for this section. Helmiopsis pseudopo- 


Discussion. 


pulus is therefore selected herein as type for section 
Glandulipetalae. 


KEY To THE SPECIES OF HELMIOPSIS SECT. GLANDULIPETALAE 
IN MADAGASCAR 


la. Leaves transversely elliptical, breadth well ex- 
ceeding length; inflorescences geminate or cymose 
with up to ek to 12) x petals 15— F mm 

m lon 6. H. bernieri 


ie ; capsule ca. 18-23 m 
lb. pe broadly ovate to orbicular, occasionally 


2a. Leaves broadly ovate to orbicular, basal veins 


apex acı minate; margin 


repa and or r rarely er es CET 
(-15) mm long, basal adaxial surface usually 
sparsely to moderately nue ovary and 
fruit not strongly lobed E H. richardii 


a 
= 


N 
o 
ER 
c 
S 
<= 
t0 
[3 
4 
E 
S 
S 
> 
2 
E 


adaxial surface densely glandular; ovary and 
fruit strongly 5-lobed. 
3a. Leaf margins weakly crenate to repand; 
basal veins (5 to)7; venation not promi- 
nently raised; style e only be- 
neath; ovules and seeds w a 
platon da EEEE pseudopopulus 
. Leaf margins crenate-serrate; i veins 
5 to 7; venation prominently raised; style 
sometimes lepidote throughout; ovules 
and seeds not winged .... 9. H. sphaerocarpa 


eo 
> 


6. Helmiopsis bernie Pow Arénes, Bull. Mus. 
416.1 


Lingvatou [12°26'S, 049730'E], (fL), J. Bernier 
i) 338 (holotype, P!; isotypes, BM not 
seen, G not seen, P!). 


(2 


envot 


Volume 96, Number 4 
2009 


Applequis 
Revision s Helmiopsis 


Small tree, to 7-8 m high, or large shrub; twig bark 
grayish, shallowly ridged, with small pale lenticels; 
young twigs ferruginous, densely lepidote. Leaves 
transversely elliptical to barely reniform, (2-)2.5— 
5.5(-7) X (2.2-)3.5-8(-11) em, coriaceous; petiole 
(1—)1.5-3(-4.8) em, densely lepidote; basal veins 5; 
base rounded-truncate to broadly rounded or very 
shallowly cordate; apex rounded, occasionally short- 
acuminate; margins entire to weakly repand; both 
surfaces lepidote, densely so in young leaves, with 
scales of variable size and coloration; stipules ca. 1.5— 
usually immediately 


3 mm, scaly to lanceolate, 


caducous. Inflorescences few-flowered, geminate or 
with up to 6(to 12) buds, mostly clustered at twig 
apices on short thick peduncles (usually cm), 
sometimes subtended by small leaves, or individual 
lateral inflorescences borne on peduncles to 5(-8.5) 
cm; only 1 flower per inflorescence maturing at a time; 


(8.2-4.4-6 


bracteoles lanceolate, 


bracts lanceolate, .3 mm, rapidly cadu- 
keeled, (3.54.2-5.5 


—6.8) mm, inserted just below bud, rapidly caducous; 


cous; 


pedicels thick, + angular, articulated, 2-6(-9) mm; 
inflorescences lepidote throughout. Sepals narrowly 
lanceolate, 14-17 mm, lepidote, reflexed at maturity, 
bearing small patches of glandular tissue at base E 
adaxial surface; petals whitish with a 

yellowish tinge, (15-)19-24(-26) X (13. Ed 2e: e 
mm, broadly obovate and often broader than long, 
adaxial surface basally glandular; androecial corona 
1-2 mm; fertile stamens 10, oppositipetalous in pairs; 
filaments (1.5-)2-2.5 mm; —6 
mm; staminodes spatulate, 8-11 mm, oppo Deere 


> 
E 


anthers linear, 


lous; gynoecium 5-carpellate; ovary lepidote, weakly 
5-sided; style lepidote, 4.5—5(—5.8) mm; style lobes 
erect, clustered, (1—)1.5-2.3 mm; ovules up to 6 per 
locule. Inflorescence becoming woody in fruit; fruit a 
woody capsule, ovoid, ca. 18-23 mm, shallowly 5- 
lobed with a short apical nipple, lepidote; young seeds 
with a long apical wing, mature seeds not 


Distribution, phenology, and habitat. | Helmiopsis 
bernieri is confined to a very small region of extreme 

rthern Madagascar in the vicinity of Antsiranana 
(Diégo-Suarez; Fig. 4) and grows on sand in some- 
times degraded forested areas. Flowering is recorded 
from November through March; the only fruiting 
collection known was made in November. 


IUCN Red List category. Provisional IUCN Red 
List category (2001) is noted as Vulnerable (VU D2). 

o more than five clearly distinct locations and two 
distinet populations are known; all collections of 
Helmiopsis bernieri have been from a limited range, 
such that a single natural disaster could affect all 
populations simultaneously, and most of the forest 
remaining in these areas is not formally protected. 


, Discussion. Helmiopsis bernieri is a very distinc- 
ve species, easily distinguished by its transverse- 
elliptica l leaf ud it also has the largest flowers and 


fruits in the genus. 


ecimens examined. MADAGASCAR. Antsiranana: 

Baie de Rigny [12°26’S, eee Bernier 2603 : = 
ic Raynaud, S rte. de la Baie de Rigny, Serv 

20361 (MO [2], P); forêt donors n» S, di "ET 

Service Forestier 20937 (MO, Py Diégo-Suarez, prés 

ha, pi w à la baie de Rigny, Service 

; forêt d'Orangea [12%15'S, 
049°23'30"E], Service prs 23274 (MO, P 


7. Helmiopsis pseudopopulus (Baill.) Capuron ex 
rénes, as “Pseudo-Populus,” Bull. Mus. Natl. 
Hist. Nat., 6. 1956. Basionym: 
Dombeya ane ae Baill., as “pseudo-Popu- 
lus,” B ens. Soc. Linn. Paris 1: . 1885. 
TYPE: Meer Prov. AS Lingva- 


tou [12°26’S, 049°30’E], s.d. (fl.), J. Bernier (2e 
envoi) 339 (holotype, P!; isotypes, G not seen, 
P?) 


Tree to 10-20) m high, to 20 em diam., or large 
rub to 6 m high; twig bark grayish to redd a brown, 
shallowly ridged, with small pale lenticels; young 
twigs pale, lepidote. Leaves broadly ovate to orbicular, 
sometimes To ovate, 3.2-10.5 
petiole 2.7—7.5(-1 
reddish; basal veins 7 or rarely 5; base cordate or 


0.5) em, densely lepidote, slightly 


occasionally rounded- truncate; 
acute, cuspidate 

shallowly crenate to shallowly repand; both surfaces 
lepidote, often pri so on pn leaves, the 
abaxial surface with pale 
nerves, the a B is Slightly Paine iue 
lanceolate to linear, 2-6 mm, immediately caducous 
and almost never present. Inflorescences cymose with 
few orders of branching, axillary and sometimes 
terminal or pseudoterminal, clustered near distal ends 
of branches; peduncle (2.4—)3.5—6.5(—8.8) cm, inflo- 
(4.2-)7-13(-14.6) em; s 
y caducous; pedicels 5-9 mm, + lar, 
rd bracteoles inserted well s ice 
immediately caducous; inflorescences and pedicels 
lepidote. Sepals lanceolate, 5-7 mm, lepidote, re- 
flexed at maturity, yellowish, bearing small glandular 
patches inside at base; petals creamy or white, (7—) 
8.5-11.5 X (4.5-)5-8 mm, broadly obovate, 
portion of adaxial surface usually densely glandular; 
androecial corona 1.1-2 mm; fertile stamens 10 


— 


rarely 15 in aberrant flowers), borne in pairs, usually 
X oppositipetalous; filaments (0.6—)1.5-2 mm, fused 
for at least half their length; anthers linear, 1.6-2.4 


—2.T) mm, orange; staminodes spatulate, 3.9-4.8 mm, 


m 


usually oppositisepalous, at least sometimes yellow at 
the apex; gynoecium 5-carpellate; ovary lepidote, 5- 


536 Annals of the 
Missouri Botanical Garden 


48°0'0"E 49°0'0"E 
L 
12°00" S- Poe 
' ^ 
2 
| 
é EN KT 
* 0: 
- 9 
a 
0 LJ 
+ + 
v + 
" + 
13°0'0"S4 
+ L13°0'0"S 
o Q 
ES ad 
T q 
49°0'0"E 50°0'0"E 


Figure 4. Approximate distribution of species of Helmiopsis sect. Glandulipetalae in northern Madagascar (@ = H. 
) 


bernieri; O = H. pseudopopulus; dg = H. richardii; A = H. sphaerocarpa). 


lobed; style lepidote on basal third to half, 3.5-5 mm, 
white; style lobes erect, clustered, 0.5-1.5 mm, 
orange. Inflorescence branches lengthening and 
becoming woody in fruit; capsule ovoid and deeply 


5-lobed, (9-)12-18 mm, immature fruit golden green- 
ish (tan when dried, possibly darkening with age), 
lepidote, persistent sepals often present; apex acute, 
sometimes beaked or rounded; inner surfaces of 


Volume 96, Number 4 
2009 


Applequis 
Revision A Helmiopsis 


locules glabrous; seeds 2 per locule, only 1 well- 
developed, at maturity ca. (3—)4.5-6 mm, with 6— 
mm apical win, 


Distribution, phenology, and habitat. Helmiopsis 
pseudopopulus is native to a limited region of northern 
4). Its collected habitat is forest 


forest; one collection 


Madagascar (Fig. 
including gallery is at the shore 
Altitude ranges to at least 300 m, and possibly as high 
as 450 m. No strong substrate preference is clear; 
collections have been made on limestone, on basalt, 
and at a locality with substrate including sand and 
rocks. Most flowering specimens are recorded from 
January to April (two November-flowering specimens 
have been collected from Montagne des Frangais); 
undehisced fruits have been collected from March to 
May, and mature fruit with seeds still present in 
August. Flowers have been noted to have a Malus-like 
scent (Nusbaumer & Ranirison 1177) 


IUCN Red List category. Provisional IUCN Red 
List category (2001) is noted as Least Concern (LC). 


Common names. Hafomena (Service Forestier 


9389). 


Nomenclature and orthography. The orthography 


a 
of the specific epithet of Helmiopsis pseudopopulus 
must be corrected to remove the hyphen, according to 
Article 60.9 of the ICBN (McNeill et al., 2006: cf. Ex. 

0). This combination was first made by Capuron in 
Vol. 14, Fascicle 4 of Notulae Systematicae (Paris), 
which was dated on the front cover *Décembre 1952” 
(and which was received in libraries early enough in 
1953 
publication). However, the back cover gives 1953 as 
the publication date, and this is believed to be the 
more oru source of information. Unfortunately, 
Article 33.4 of the ICBN (McNeill et al., 2006) 


required that new combinations published after 1 


to indicate there was no lengthy delay in 


January 1953 provide a full and direct reference to the 
basionym’s place of publication. Capuron’s combina- 
tion was therefore not validly published; it was 


validated by Arénes (1956) 


E ds xamined. MADAGASCAR. 
m massif de l'Ankarana, Ambiloma- 
godro partie ds e bordure, 13°01’ 20S: 
049°08'14"E, Bardot- Vas et al. 1366 (MO, TAN 
Daraina, sous-préfecture de Vohemar, forét de Bekaraoka, 
partie sud, en aval d'Andranotsimaty, en direction du point 
coté 112, 13°09'59"S, 049742'07"E, Gautier et al. 4385 
pentes du Lac Sacre [= An 
[12744/S, 049^15'E]. Py 
Montagne des Francais [12722'S, 049°21’E], Keraudren 
1672 (MO, P); Mantamena, part of Bekaroaka Range, 7 
NE Daraina (Vohémar), 13°08'S, 049°42’E, Meyers & Bolte 
hemar, commune 
06'11”S, 
049742'39"E, pente 22°, orientation 330°, bas de pente, 


M. eodd 
tsiranan: 


Nusbaumer & Ranirison 1177 (MO); Ankara, Diégo-Suarez, 
Service Forestier 5438 (MO, P, TAN); Ankara, Diégo-Suarez, 
Service oe 9389 (P); Montagne des Frangais, vallée de 
la'Andavakoera, Service Forestier 20922 (MO, P); Montagn 


calcaires lapiazés de l'Ankarana, prés d'Ambondromifehy 

[12753'S, 049°12'E], Service Forestier 24730 (MO, Py 

nt E du massif de l'Ankerana (partie S du massif de 

vo [ca. 13?18'S, 049°50’E), au N de Vohémar, Service 
Forestier 27412 (MO [2], P). 


8. Helmiopsis richardii (Baill.) Capuron ex Arénes, 
Hist. Nat., sér. 2, 28: 416. 1956. 
“Ri 


Basionym: Tiodhena richardii Baill., as 


chardi,” Adansonia 10 1871. Dombeya 
richardii (Baill) Baill., as "Richardi" Bull 
inn. Paris 885 


Be 13°17'S, 04815"E], (fL), J. M. C. Richard 
343 (holotype, P!; isotype, P!). 


Tree to 15 m high, at least 20 cm DBH; twig bark 
grayish to reddish brown, shallowly ridged, with small 
pale lenticels; young twigs pale, lepidote. Leaves 
"en ovate to orbicular, (3—)4.2-8.7(-11.3) X 3.1- 

8.4(-10.8) cm, thick-textured; petiole (1.7—)3—6.8 
(8.3) cm, densely lepidote; basal veins 5; base 
shallowly cordate to rounded-truncate or rounded; 
apex acuminate; margins entire to weakly repand or 
very rarely shallowly crenate; both surfaces lepidote, 
the aba 


lepidote and sometimes possibly dark-margined, 


xial surface pale green; stipules ca. 2-2.5 mm, 
immediately caducous. Inflorescences cymose with 
few orders of ee a clustered at distal 
ends of twigs; peduncle 2 cm; branches of 
inflorescence compact, A. in total 5.5- 
o 


articulated; bracteoles 


12 em; bracts immediately caducous; pedicels (3 
15(-20) 


inserted well below 


mm, + angular, 


flower, immediately caducous; 
inflorescences and pedicels lepidote. Sepals narrowly 
lanceolate, (4—)6.5-10.4 mm, lepidote, yellow-green 
with brownish scales, reflexed at maturity, usually 
bearing small glandular patches inside at base; petals 
white to creamy or greenish white or pale yellow, (7—) 
11-13(-15) X (5-)6-8(-10) mm, broadly obovate, 
lower portion of adaxial surface sparsely to moderately 
glandular or subglabrous; androecial corona 1.2-1.9 
—2.4) m o 

rarely singly, u y E opposi petalous filaments 
131.5-2 mm, fused for half their length, white; 
anthers linear, (1.6—)2.4—4(—4.3) mm, brown; stami- 
nodes spatulate, (4.5—)5—8.5 mm, usually oppositise- 


m 


m; fertile stamens rne in pairs or 
uall 


— 


palous; gynoecium 5-carpellate; ovary lepidote; style 
lepidote on basal portion to lower third, 4.2-6(—7.5) 


Annals of the 
Missouri Botanical Garden 


mm; style lobes erect, clustered, 1—1.8 mm. Inflores- 
cence branches lengthening and becoming woody in 
(1311-18 


shorter and truncate with or without a small beak, 


fruit; capsule ovoid, mm, tapering or 
internally glabrous, dehiscing from the apex, reddish 
rown, lepidote; seeds 2 per locule, only 1 well 
developed, in larger fruits ca. 4.5-5 mm with 7-8 m 
apical wing, in small fruits ca. 3—4 mm aw wing ca. 


3—5.4 mm 


Distribution, Pei and habitat. | Helmiopsis 
richardii is to a limited region of northe 
Madagascar m 9 Its collected habitat is forest on a 
variety of soil types, but id on calcareous 
substrates, sometimes near river e orte 
altitude range is from 90 to at qun 200 (possibly to 
350) m. The recorded flowering period is (November 
to)February to April; mature fruits with seeds have 
been collected in April (Ankarana), id (Montagne 
des Francais), and October (Baie de 
Flowers have been reported to smell o of fish 
(Cheek et al. B1408). 


IUCN Red List category. Provisional IUCN Red 
List category (2001) is noted as Least Concern (LC). 


iégo-Suarez). 


Com mes. iky (Service Forestier 9432); 
T "(Service 5 orestier 15061); Sel 
Forestier 14683). 


ely (Service 
Discussion. This species bears considerable gen- 
eral resemblance to Helmiopsis pseudopopulus, but can 
be distinguished by its unlobed ovary and fruit and by 
its more sparsely glandular petals; the basal portion of 
the petals of H. pseudopopulus bears a fairly dense and 
even coating of small glands, except where these have 
been lost to disturbance, whereas those of H. richardii 
have ar few, unevenly distributed glands. On 
rage, H. richardii has slightly larger adi with 
ie sepals, petals, and androecial parts (sepals 
mostly over 7 mm, pes wa over 2.4 mm, 


staminodes mostly over 5 mm) and longer pedicels 


Eh 


(usually over 9 mm). Un loin the type ol 
lest ext 


typ 
richardii represents the smal reme of that 
species’ floral size range, indi because it is from 
a geographically and genetically distinet population 
(the island of Nosy Be), although the 


in ideal condition. Little fruiting material referable to 


specimen is not 


H. richardii is known, but it encompasses fruits with 
two distinet morphologies. Fruits collected at Baie de 
Diégo-Suarez and Ankarana have a blunt apex with a 
short acute tip, while fruits from Montagne des 
Francais are longer and have a tapering apex. More 
collection is certainly desirable. 

Leaves of Helmiopsis pseudopopulus have shallowly 
crenate to undulate margins and apices that may range 
from acuminate to acute, cuspidate, or rounded. The 


breadth of the leaves in H. pseudopopulus not 
infrequently exceeds the length, which is very rare 
in H. richardii; the basal veins almost always number 


ue. 


seven, rarely five. Leaves of H. richardii are entire to 
shallowly undulate, but usually not shallowly crenate, 
relatively narrower and sometimes more graceful in 
appearance, with consistently acuminate apices; there 
are normally five basal veins. There is enough overlap 
in leaf morphology that sterile specimens may be very 
difficult to identify; the variation in basal leaf vein 
number may be the most useful character. 

Two specimens from Ankarana have particularly 
unusual morphology: Humbert 32551 (MO [2]} has 
typical Helmiopsis richardii floral morphology but 
prominently crenate-serrate leaves (with only five 
basal veins), while Service Forestier 29220 (MO) has 
unlobed, narrowly elongated young fruits but slightly 
crenate leaves (with five to nine basal in and very 
small flowers on short pedicels (to vs. 
minimum o: in e MID Typical "n 
richardii and H. pseudopopulus have both been 
collected from the Ankarana region, and the two 
species are clearly closely related. Hybridization 
between them m i 
these two intermediate specimens. 


ay be possible and may account for 


Nomenclature and orthography. As for Helmiopsis 
pseudopopulus, discussed above, the combination H. 
richardii was first made by Capuron (1952); an 
unfortunately timed publication delay rendered both 
combinations invalid under Article 33.4 of the ICBN 
(McNeill et al., 2006). The combination was validated 
by Arénes (1956), who simultaneously introduced the 
correct spelling of the epithet, the originally published 
Latin termination (“richardi”) being an error requiring 
correction under Article 60.11 of the ICBN (McNeill 
et al., 2006). 


Representative specimens examined. MADAGASCAR. 
Antsiranana: S of Anivorano Nord, env. Ambalabao, ca. 
12°48'S, 049^14'E, Cheek et al. B1408 (TAN); Ankarana 
Special Reserve, ca. k Village near 
O stream, 12757'16"S, 049%07'30"E, Harder et a 

1725 (MO) c et plateaux calcaires de VAnalamera; 
Humbert 19244 (P); plateaux calcaires de l'Ankarana du N 

ntre Ambilobe et Anivorano, Humbert & Capuron 25558 (P); 
o de Diégo-Suarez, Montagne des Francais [12°22’S, 
049°21'E], Keraudren 1646 (MO, P); Nosy Be [13%17'S, 
048^15' de udi 317 2 plateau calcaire de l'Ankarana, 
au NE d'Ambondrom '30"S, 049°12'30"E], 
Service ae 3040bis (P. Wow E [12°40’S, 
049733'E], Service uf 9432 (P) Baie Diégo-Suarez, 
Service Forestier 14883 (P); Montagne des Francais, Diégo- 
Suarez [12°22’S, 049°21’E], Service Forestier 15061 (MO, P); 
plateau calcaire de l'Ankarana, à I'W d'Ambondromifehy 
(J.B. 8) [12753'S, 049°12’E], Service Forestier 22694 (MO, 
Py a de l'Ankarana, rebord S du Plateau de Mahory 
(rive gauche du haut Rodo) [12%49'S, 049°14’E], Service 
Forestier 23144 (P); massif de l'Ambongoabo, entre la baie de 


Volume 96, Number 4 
2009 


Applequis 
Revision e Helmiopsis 


Diégo et celle du Courrier men 049°10’E], Service 
Forestier 24644 (MO [2], P 


9. poing, sphaerocarpa L. C. Barnett, Ann. 
Misso Bot. Gard. 74: 450. 1987. TYPE 
Madscascax Puls Antsiranana: massif de la 
Montagne d'Ambre [ca. 12^36'S, 049°09’E], crête 
entre les bassins de la riviére des Makis et de la 
riviére d'Ankazobe, 600—800 m, 26-27 May 1970 
(fL), Service Forestier 29194 (holotype, P!; iso- 
types, K not seen, MOI, P [2]!, TEF not seen). 


Tree to at least 14 m high, 11 em diam., or large 
shrub; twigs sturdy; bark fibrous, dark, somewhat 
reddish brown, shallow ly ridged, p small pale 
lenticels; youngest twigs sometimes lepidote, with 
small scurfy fimbriate scales. Leaves broadly ovate to 
ovate or elliptical, or occasionally obovate or 
suborbicular, 5-13 X 3.5-9 cm; petiole (0.8-)2-5 
(—7.8) em, sparsely lepidote; basal veins 5 to 7; base 
dari cordate or MER rounded; apex acute 
t ort-acuminate or rounde argins crenate to 
crenate-serrate; adaxial surface 21. or nearly so, 
abaxial surface sparsely lepidote, mostly along veins; 
venation conspicuous on both surfaces, prominent 
abaxially; stipules lanceolate, ca. 4—5 mm, rapidly 
caducous. Inflorescences cymose with several orders 
of branching, terminal and lateral, clustered near 
branch ends; peduncles 2-8 cm, inflorescences in 
total 6.5-15 cm, with lower internal branches well 
developed; bracts narrowly lanceolate, rapidly cadu- 
cous; pedicels 2.5-6(-8) mm, + angular; bracteoles 
inserted immediately below flower, rapidly caducous; 
inflorescences and pedicels lepidote. Sepals lanceo- 
late, 5.2-6.5 mm, i 


small glandular patches inside at base; pe 


lepidote, yellowish, usually bearing 
tals white, 
7-10 X 6-8.5 mm, broadly obovate with asymmetri- 
cal apex, lower third of adaxial surface densely 
ertile 
oppositipetalous; 


glandular; androecial corona 0.6-1.3 mm; 
stamens 10, borne in pairs, + 
filaments 0.7—2.2 mm, free to corona; anthers linear, 

-7-2.6 mm; staminodes spatulate, 5-5.8 mm, white, 
+ oppositisepalous; gynoecium 5-carpellate; ovary 
lepidote, 5-lobed; style lepidote basally or for entire 
5.5 mm, e 


length, 3.5- EN stigma lobes 
8 m nflore 


erect, clustered, scence 
branches becoming ae in fruit; Me subspher- 
ical, 5-lobed, 5.4—7 mm, dark brown, lepidote; apex 
rounded-truncate with small apical projection; fruit 
subtended by hardened persistent sepals; inner 
surfaces of locules a seeds 1 to 2 per locule, 
only 1 well developed, at maturity ca. 5-6 mm, 
irregularly ma triangular, + laterally com- 
pressed, wingless except for inconspicuous keel. 


Distribution, phenology, and habitat. | Helmiopsis 


sphaerocarpa is native to a limited range of sites in 


northern Madagascar (Fig. 4). The few known collec- 

tions are from ae areas, one noted to be in 
ry ed habitat is 

en rocky; oe altitudes range from 

to 600-800 


and June, with mature fruits with seeds still present in 


ciduous for and the preferr 
m. Flowers have been collected in May 


September and December. 


IUCN Red List category. Provisional IUCN Red 
List category (2001) is noted as Vulnerable (V 
Helmiopsis from the same 
widely and frequently collected, suggesting that this 
species is relatively rare. 


Common names. Selivato (Service Forestier 5673); 


Sely (Rakotondrafara et al. 268) 


ussion. Barnett (1987), in the original publi- 
at this 
species is likely the sister taxon of H. pseudopopulus, 
with which it 
inflorescence, and flor: 
The two 


as the leaves of H. sphaerocarpa are coarser in 


si 
cation of Helmiopsis sphaerocarpa, suggested th 


shares a number of vegetative, 
characters or tendencies. 
can be distinguished i in the vegetative state, 


2 d more ir NER crenate, and closer 
to glabrous. The midrib and s ary veins are 
pro ruis nice on s abaxial pees for virtually 
their entire length and often conspicuously paler or 
darker than the leaf surface, whereas in H. pseudo- 
populus the secondary veins are weaker, often scarcely 
raised toward the distal ends, and not strongly colored. 
The fruit of H. pseudopopulus is larger (usually 12— 
18 mm long) and the seeds have a substantial apical 
wing, which is not present in H. sphaerocarpa. The 
sepals of H. pseudopopulus often persist during fi 
maturation but never become as thick and woody as 
those of H. sphaerocarpa. Several small differences in 


I 


floral morphology also separate the two species; for 

ple, the filaments of stamen pairs in H. 
pseudopopulus are fused near the base, and the style 
is lepidote only on the lower portion. 


pecimens examined. MADAGASCAR. Antsiranana: 
Sous-préfecture Antsiranana II, 
Fokontany Andranomanitra, Ampitilia 
d'Andranomanitra, 12723" 18's, 049°22'57"E, Jahetondia- 
fara. et al. 268 (MO); Diégo-Suarez, Montagne des Francais 
[12°22’S, 049"21'E], Service Forestier 5673 (P); versant E du 
massif de l'Ankerana (partie S du massif de Mafokovo [ca. 
13718'S, 049*50'E] au N de Vohémar, Service Forestier 
27349 (MO, P). 


SPECIMINA INCERTAE SEDIS 
Melmiopsis sp. indet.: Madagascar. Prov. Antsiranana: 
for : 


ét d'Analamahitsy (PK. 84 
Ambilobe), entre Anivorano et Ambondromifehy 


e la rte. Diégo— 


Annals of the 
Missouri Botanical Garden 


[12°48’S, 049714'E], 22 Apr. 
Forestier 22669 (MO [2], P). 


1963, Service 


This collection, a white-flowered tree 15-18 m 
high, has been previously determined as Helmiopsis 
pseudopopulus; it has been determined twice as 
“H. angulata,” a name used first r Capuron in sched. 

by him or any 
succeeding author. The general appearance is consis- 


but apparently never e 


tent with H. pseudopopulus, except that the inflores- 
ong (to 18 em in flower). 
als are narrower (4-5 mm broad, 
with the breadth being less than half the length); the 
petal glands are very sparse rather than plentiful; the 
style is longer than usual (5.5-6.2 mm), bent or 
inked below the stigmas, and lepidote for over half 
its length; and the staminodes are somewhat longer 
than usual (4.7—5.7 mm). This collection may well 
o a distinci—and undoubtedly rare s 
endangered— species; pi, since its 
wit i. range of H. p. 
that it is a very meh individual or distinct variety 
of that Further 
investigation and collection of populations along 
Route Nationale 6 would be highly desirable. 


loc 
eudopopulus, the bos 


species cannot ruled out. 


Literature Cited 


Applequist, W. L. . Two new species of Dombeya 
(Malvaceae) PON Madagascar. Novon 19: 289—294. 
a. Contributions a l'étude des Sterculiacées 
genre Helmiop. rr 
scar, sér. B, 7 5-57 


H. ie 
—— —. 1958. [or n de Madagascar et des Comores. 
Candolloa 247—449 
D. Sterculiacess: Fl. Madagasc. 131: 1-537. 
Batic E a 1987. An unusual new species of Helmiopsis H. 
d ud from Madagascar. Ann. Missouri 
50— 


Bot. Gar 


Bayer, C., M. F. Fay, A. » de Bruijn, V. Savolainen, C. M. 
Morton, K. dun 
1999. Support 


xa et noms nouveau 
dans le genre e Dalbergia (Papilionaceae) à Madagiscar el et 
aux Comores. Bull. Mus. Natl. Hist. Nat., sér. 4, sect. B, 
Adansonia 18: 171-212. 
Eod R. 1952 P Identité des PME Nesogordonia 
. Bn. istanthera K. Schum. et description de deux 
espéces nouvelles de Madagascar. Notal. Syst. (Paris) 14: 


new combination in n the onde 


and the implications for planning biodiversity conserva- 
tion. Pp. 51-74 in S. M. Goodman & J. P. Benstead 
(editors), The Nanital History of Madagascar. University of 
Chicago Press, Chicago. 
eran . G. 1926. i ue des Dombeya de 
Madagascar. Candollea 3: 5-12 
IUCN. 2001. IUCN Red mo Cue and. rA Version 
3.1. Prepared by the IUCN Species Sur Commission. 
i Gland, al am aie United 


ngdom. 

cae e 95. Prodromus for a rev un "d 
Combretum (Combretaceac) for Madagascar. Bull. 
Natl. Hist. Nat., sér. 4, sect. B, Adansonia 17: a. 
200. 

Lowry, P. P. IL, T. Haevermans, J.-N. Labat, G. E. Schatz, 
J.-F. Leroy & A. P Wolf. 2000. Endemic families of 
Madagascar. V. synoptic revision of Eremolaena, 
Pentachlaena and ph (Sarcolaenaceae). Ad- 
ansonia, sér. 3, 22: 11-31. 

i t, V. Demoulin, D. L 


Code of Botanical Nomenclature (Vienna 
Code). Regnum Veg. 146. 
s de la Báthie, H. 1944. Les genres so An 
aillon et Helmiopsis g. n. de Madagascar. Bull. Soc 
A 91: 226-232. 
Vogel, S. 2000. The floral nectaries of Malvaceae sensu 
lato—A conspectus. Kurtziana 28: 155-171 


A REVISION OF NEOTROPICAL 
BONYUNIA (LOGANIACEAE: 
ANTONIEAE)! 


Jason R. Grant? 


ABSTRACT 


A revision of the Neotropical genus Bonyunia M. R. Schomb. ex Progel (Loganiaceae, Antonieae) is provided, including a 
key to species, Pap rd distribution, VID n List status, Ed od n n is pal and ecologically 
diverse with tax call hab d indument of 
the calyx, Duns p De pues occurs in lowland regions of n eee River watershed (Brazil and Bolivia), the 
Orinoco River watershed (Colombia and eto e of the p E and outliers ERAN Guyana, Venezuela, 
and Colombia), and lowland regions of Amazon n-facing Andes , throughout on white sands. Ten species are 


cinchonoides Gleason & Standl.), and B. i rba Progel. Six species are newly described: B. excelsa » E 
Grant (Colombia), B. magnifica J. R. Gra sed D. d pw) R. Grant (Colombia), B. pulchra Ricketson, J. R. Gra 


Liesner (Peru), B. spectabilis J. R. Grant (Carra: and B. venusta. J. R. Grant (Brazil). 
words: Antonieae, Bonyunia, IUCN Red List, Loganiaceae, Neotropics, 


South America. 


The Loganiaceae includes the formerly segregate 
families Antoniaceae, Gardneriaceae, Geniostomata- 
ceae, Spigeliaceae, and Strychnaceae, and comprises 
13 genera: Antonia Pohl, Bonyunia M. R. Schomb. ex 
Progel, Gardneria Wall., Geniostoma J. R. Forst. & G. 
Forst., Labordia Gaudich., Logania R. Br., Mitra- 


(Bremer & Struwe, 1992; Struwe et al., 1994, 1998; 
Backlund et al., 2000; Angiosperm Phylogeny Group, 
2003; Struwe & Motley, in press) The genus of 
is a Neotropical woody 


interest. here, Bonyunia, 


member closely related to AE (Mori et al., 
bas 


2002) and Usteria, 


and molecular characters, the rad genera (perhaps 


and, on both morphology 
also Norrisia) comprise the tribe Antonieae. While 
Bonyunia has been shown to be distinct from related 
genera (Struwe & Albert, 1997), 


screening at the Université de Neuchat 


a preliminary 
el of a series 
of adequate herbarium samples was unsuccessful in 


the amplification of DNA (chloroplast trnZ-F) from 
enough taxa to establish a genus-wide phylogeny. 
Bonyunia has essentially lain in obscurity, with the 
exception of several brief references and a revision 
nearly 40 years ago (Leeuwenberg, 1969). The 
accumulation of enough interesting material has 
finally resulted in the need for a full taxonomic 
revision 

Bonyunia occurs in generally lowland regions of the 
Amazon River watershed (Brazil and Bolivia), the 
Orinoco River watershed (Colombia and Venezuela), 
tepuis of the Guayana region and outliers (Brazil, 
o lombia), and low and 


(Fig. 1). It does not occur on the Brazilian plateau, 
though it is approached there by B. antoniifolia 
which has a broad distribution in the 
Amazonian lowlands along the Amazon River and its 


Progel, 


tributaries in Brazil and Bolivia. Two other species 
have somewhat broad distributions that overlap each 


! T thank the following institutions for the loan of material or data/scans/photos* of specimens in their collections: AAU, B* 


BM, BR, COL*, F, FMB, G, GH, HUT* 
and Z. Information on collections, some of which were 
Bernal 
Humberto Mendoza (FMB), M 


, TAN*, K, M*, MA, MG* 
e not requested on loan, was pro ovided by: fans Bemerguy (MG), Rodrigo 


, R*, RB*, S, SP*, U, UC, US, USZ*, W*, WAG, 


> > > 


Rodriguez (HUT), Franz Schuhwerk (M), Bruno Wallnófer (W), and the Tropicos. o of the Missouri Botanical Garden 


MO). Library research was carried out at the Cons 
Angell skillfully prepared the illustrations. The 


ervatoire et Jardins botanique 
geographie information Roche distribution map was prepared by 


Ville de Genéve, Switzerland. Bobbi 


Phillip Miarmi, Rutgers University (CHRB), and modified by Yves Maumary (Neuchatel). Neil Villard (Neuchatel) prepared 


the photographs of seeds and 
provided relevant literat ture. 
? Laboratoire de bot é 
2009 Neuchatel, Switzerland. jason. grani@unine:c 
doi: 10.3417/2006135 


herbarium specimens. Lena Struwe (CHRB) edited the manuscript, and 


Mary Endress (Z) 


tive, Institut de Biologie, Université de Neuchâtel, rue Émile-Argand 11, Case Postale 158, 
ch. 


ANN. Missouni Bor. Garp. 96: 541—563. PUBLISHED ON 30 DECEMBER 2009. 


Annals of the 
Missouri Botanical Garden 


a 7OW Y e 
of 


Colombia 


Juegos Ea 


Species 


B] antoniifolia 


o pulchra 
Á spectabilis 
SB superba 
® venusta 


km 
0 200 400 


E 


nezuela 
Atlantic Ocean 
S 


10°S 


Figure 1. 


other: B. aquatica Ducke and B. minor N. E. Br. 
Bonyunia aquatica ranges principally along the 
lowlands of the Orinoco River watershed in 
Colombia and Venezuela but also extends to the 

Negro Bonyuni 


higher-elevation savanna and tepui habitats of the 


in razil. a minor occupies 


Guayana region in Venezuela and Guyana. The 


remaining seven species are known from only one to 
10 collections, and therefore their distribution is 


poorly known. These mostly occur in areas of high 


endemism (Sierra de Chiribiquete and Mount 
Roraima of the Guayana region, and the Andes) 
and are probably narrow endemics. Both B 


spectabilis J. R. Grant and B. superba M. R. Schomb. 
ex Progel occur on Mount Roraima along the 
Venezuela-Guyana border in the Guayana region 
and are possibly partly sympatric with B. minor. 


Chiribiquete, and B. magnifica J. R. Grant and B. 
venusta J. R. Grant occur in the Amazon lowlands of 
rant & 
Liesner occurs in low-elevation areas of Amazon- 


Brazil. Bonyunia pulchra Ricketson, J. R. 


facing low-Andean slopes in Peru. 


Map of the distribution of Bonyunia in South America. 


With its 10 species (Appendix 1), Bonyunia has an 
overall distribution that us mirrors that of two 
genera of the Gentianaceae: Potalia Aubl. with nine 
species (Struwe & Albert, 2004) a Tachia Aubl. with 
13 species (Maguire & Pa 1975; Peters et al., 
2004; Struwe et al., 2005). While each of these three 
genera is generally composed of narrow endemics, each 
also has one species that is generally wide ranging in 
lowland Amazonia: B. antoniifolia, P. resinifera Mart., 
& Weaver. Further 
comparative phylogeographic and geographic studies 


and 7. occidentalis Maguire 
on these groups are currently under investigation 
(Grant & Struwe, in prep.). In contrast to these less 
speciose genera, Macrocarpaea (Gentianaceae), with 
more than species, has recently n shown to 
occur almost exclusively in mountainous regions of the 
Neotropics (Struwe et al., 2009). 
Richard Schomburgk named Bonyunia in honor of 
P friend George Reginald Bonyun, M.D. (ca. 1811- 
. a medical doctor of Georgetown Guyana 
p 1848). In the same manner in which 
his brother Robert Schomburgk reported on his 
geographie surveys to the governor of British Guiana, 
Sir Henry Light, George Bonyun was commissioned by 
Light to report on the general health of immigrants 


Volume 96, Number 4 
2009 


Grant 
Revision of Bonyunia 


and the conditions in rural hospitals (Bonyun, 1848: 
22). Bonyun also published several articles on the 
ethnobotanical use of plants, including the Demerara 
pinkroot, Spigelia anthelmia L. (Loganiaceae) (Bon- 
yun, 1844 


MATERIALS AND METHODS 


All available types and specimens of Bonyunia 
from herbaria with large Neotropical collections were 
examined on loan or via Pec ce of 
specimens including AAU, B*, BM, BR L* 
FMB, G, GH, HUT*, K, M, "d MO, NY, P, R*, RB*, 
U, UC, US, USZ*, W*, WAG, and Z. 

information also extracted fro 

online CAN f the Missouri 
Botanical Garden (<http://www.tropicos.org>). One 
undred seventy-eight collections were examined, 
often with multiple duplicates, for a total of 402 
sheets, which is a nearly fourfold increase of 
collections since the revision by Leeuwenberg (1969). 


Taxonomic History 


Bonyunia was proposed as a monotypic genus in the 
Loganiaceae (Schomburgk, 1848). However, despite 
providing locality information, th 
tion, diagnosi 
nudum and invalid. Valid establishment of the genus 


ere was no descrip- 
s, or illustration, rendering it a nomen 


and species was only effected 20 years later by Progel 
(1868) in his treatment of the Log 
Flora Brasiliens 


however, 


aniaceae in Martius? 
is. Typification remained problematic, 
because although the genus had been 
originally proposed by Schomburgk as monotypic 
(for B. superba M. R. Sch 


added a second species that inadvertently resulted in 


omb. from Venezuela), Progel 


a simultaneous publication date for both species, B. 

superba M. R. Schomb. ex Progel, and B. antoniifolia 

ublished in the Mantissa 

(addendum or supplement). Nevertheless, Leeuwen- 

berg (1969) effectively lectotypified the genus on B. 

superba, whereby Schomburgk's original intent of the 
d 


from Brazil, which was pu 


genus was preserve 
After those first two species, three more were 
eventually added to the n Bonyunia minor from 
1901), B. cinchonoides Gleason & 
la (Glea on, 1931), and B. 
aquatica from Brazil (Ducke, 1935) Leeuwenberg 


(Brown, 
from Venezue 
(1969) revised Bonyunia and accepted four species: 
minor (placing B. 
B. mino 

superba. While this is the only revision of Bonyunia, 
Berry, in the Flora of the Venezuelan Guayana (2001), 
provided a brief synopsis largely based 


B. antonüfolia, B. aquatica, B. 
der 


cinchonoides in synonymy un r), and B 


on Leeuwen- 
berg’s work. Until now, these two pieces of literature 


have remained the primary sources of information on 
Bonyunia. 


MoRPHOLOGY 


Bonyunia is characterized by a white to yellowish 
white corolla at anthesis that turns pink, red, to purple 
2A); thin and double-winged seeds 


at maturity (Fig 
Fig. 2€); e that are opposite, entire, with many 


— 


variations in shape (Figs. 2B, 3, 5); and 5-merous 
flowers with a campanulate calyx (Figs. 4, 5). While 
having only 10 species, Bonyunia exhibits a rather 
extraordinary breadth of morphological diversity, 


especially in habit, leaves, calyces, and seeds 


HABIT 


onyunia is comprised of lowland rainforest 
riverside trees (B. aquatica), lowland rainforest terra 
firme canopy-level trees (B. magnifica, B. venusta), 
shrubs and trees of grasslands and open savanna (B. 
antoniifolia), savanna-tepui shrubs (B. minor), tepui- 


base trees (B. excelsa, B 
(B. spectabilis, B. superba), and Andean rainforest 


nobilis}, cloud forest trees 


trees (B. pulchra). It occupies a broad range of 
habitats reflecting its adaptation to the major 
ecosystems in northwestern South America. Inverse- 
ly, species have undergone morphological adaptation 
in habit to these habitats. Bonyunia antoniifolia is a 
5-9 m tall shrub to tree that occurs in grasslands, 
fields, or open savanna on terra firme, while 

aquatica is a 2—15 m tall lowland rainforest riverside 
tree, and B. minor is a 2-10 m tall savanna-tepui 
shrub to tree. These three species share their strong 
hairiness and seed morphology, and they also have a 
preference for open, sunny localities (river corridors 
and savannas). The remaining species occur in closed 
forests and tend to be glabrous and ues) 
decrease in height toward higher elevations: B. 
tall lowland 


rainforest terra firme canopy Ded ene trees, B. 


magnifica and B. venusta are 15-20 


pulchra comprise 15—40 m tall Andean rainforest 
trees, B. excelsa and B. nobilis are 4—35 m tall tepui- 
base trees, while B. spectabilis and B. superba are 
1 


—1.6 m tall cloud forest trees. 


LEAVES 


The leaves of each species of Bonyunia have a 
unique suite of morphological characters that renders 
leaf morphology very useful in species identification 
(Figs. 3, 5). The variable and taxonomically useful 
(ovate, elliptic, oval, 
(attenuate, cuneate, cordate, to rounded), and the leaf 


544 Annals of the 
Missouri Botanical Garden 


uli imm en aluidundii lu MT 


. Bo onyunia minor in flower. Corolla is white to yellowish white at anthesis, yet changes at maturity or 
perhaps once seegullinisted to pink, red, or purple. —B. B. minor in s notice Cerambycidae insect. —C. Seed morphology of 
Bonyunia. 1. B. antoniifolia (Dambros 556 [US]. 2. B. minor (Maguire & Wurdack 30525 [W AG]. 3. B. aquatica (Ducke 224 
[NY]. 4. B. superba (Pinkus 270 [US]). 5. B. magnifica. (Prance et n" 22804[ [WAC]) 6. B. nobilis (Palacios et al. 2393 [MO]. 
7. B. pulchra (Wallnófer 14-41088 [K]). A, B photos by Robin Foster. C photo by Neil Villard. 


apex (acuminate, acute, obtuse, rounded, to retuse), in classification of Leeuwenberg (1969) relied nearly 
addition to the texture (thin to thick) and indument on entirely on leaf morphology and is also important in 
the abaxial leaf surface, especially along the midvein this monograph, although calyx and seed morphology 
and secondary veins (glabrous to densely hispid). The are equally important in this treatment. 


Volume 96, Number 4 
2009 


Grant 
Revision of Bonyunia 


STEMS AND INFLORESCENCES 


Bonyunia is composed of branched shrubs and 
trees, and the apex of each branch has an 
inflorescence composed of a dichasium of paired 
cymes. The stems of Bonyunia are solid and round. 
In most species, the stems and peduncles are at 
least sparsely pubescent, though a few species are 
completely glabrous (B. magnifica, B. minor, and 
B. nobilis). In several notable cases, the stems are 
nearly glabrous, or sometimes with only one side 
pubescent, yet the branches of the inflorescence 
are contrastingly extremely pubescent, e.g., B. 
excelsa. 


FLOWERS 


The corolla of Bonyunia is white to yellowish white 
at anthesis and changes at maturity or, perhaps, once 
pollinated, to pink, red, or purple (Fig. 2A). Never- 
theless, the corolla, pistil, and stamens of Bonyunia 
have few taxonomically informative characters. In 

species wit aila on the 
herbariu 


not made because the measurements are nearly 


only few flowers e 
m specimens for study, full dissections were 
identical between species and destroying small 
amounts of precious material was not worthwhile. 
The outside of the corolla tube and corolla lobes is 
always appressed strigose, its hairs being shorter than 
. The inside of the reflexed 
corolla lobes is glabrous and therefore exposes the 


the hairs on the calyx 


color of the corolla (otherwise covered by hairs on the 
outside). 

The shape and indument of the calyx and calyx 
lobes are among the most important features in 
species identification (Figs. 4, 5). The calyx indu- 
mentum ranges from densely appressed strigose to 
hispid (in Bonyunia antoniifolia, B. aquatica, and B. 
spectabilis), hispidulous to glabrous (in B. Mud 
glabrous to hispidulous (in B. venusta am 
excelsa), to glabrous (in B. magnifica, B. minor, 3 
nobilis, and B. superba). Sometimes a few un 
hairs migrate toward the pedicel or peduncle (e.g., 

B. minor), and sometimes tufts of hairs appear on t ile 
apices of the calyx lobes (B. nobilis). The overall length 
of the cal s (2-4[-6] 


including 0.3-1.5 mm calyx lobes), yet one species, B. 


yx is rather similar in all species 


superba, is exceptional in having 5-10 mm calyces, 
including 2-6 mm calyx lobes. Bonyunia superba is also 
unique in having spatulate-shaped calyx lobes, while all 
lobes. The 


classification presented here is largely based on calyx 


other species have triangular-shaped 


characters. 


FRUITS AND SEEDS 


The pistil and mature fruit of all species of 
Bonyunia are extremely hispid throughout. The fruits 
are erect and appear to have a rather consistent 
morphology a 
dry 


s ovoid, ellipsoid to obovoid, bivalved, 
dehiscent capsules wit 


Fig. 2B), 
containing one to 20 thin, flattened, winged seeds per 


th each locule 
locule (e.g., two to 40 seeds per capsule). Bonyunia 
and its closely related genus Antonia have winged 
seeds in dry capsules that facilitate wind dispersal 
(Mori & Brown, 1994). 

There is a broad and previously unrecognized range 
of seed morphology in Bonyunia (Fig. 2C). In the 
absence of a molecular-based phylogenetic hypothe- 
sis, the seed characters provide a useful insight into 
the relationships in the genus. Seven of the 10 as 
have specimens with mature fruits and seeds: 
antoniifolia, B. aquatica, B. magnifica, B. minor, a 
nobilis, B. pulchra, and B. superba (mature seeds are 
missing for only B. excelsa, B. spectabilis, and B. 
venusta). After examination of seeds of several 
capsules on multiple specimens per species, seed 
size appears to be consistent within species. Based 
on seed morphology, these species can be divided 
into two groups based on morphology and color. 
Group I (Fig. 2C, 1-4) is comprised of five species 
that have rather thin, flattened seeds with a large, 
prominent, brown seed body and tan-colored wings: 
and B. 


superba; the seeds available from B. spectabilis are 


B. antonüfolia, B. aquatica, B. minor, 
immature, but the species clearly belongs here. 
ih II (Fig. 2C, 5-7) comprises three species with 

e three-dimensional aspect with slightly curled 
E wings, has a less prominent difference in color 
between the bodies and wings, and is rather orangish 
in color: B. magnifica, B. nobilis, and B. pulchra. 
Despite disjunct localities, the unique shared seed 
morphological characters of B. magnifica (Amazo- 
nian Brazil) and B. nobilis (Colombia) support their 
close relationship, as seen in their similar leaf 
morphologies. Seeds are not available for B. excelsa 
or B. venusta, but, based on general morphology, they 


divided into two groups based on seed 
morphology: 

Group I: B. antoniifolia, B. aquatica, B. minor, B. 
spectabilis, B. superba (Amazon River basin of Brazil 
and Bolivia, the Orinoco River basin of Colombia and 
Venezuela, and the Guayana region: Guyana, Vene- 
zuela) 

Group II: B. magnifica, B. nobilis (Brazil, Colom- 
bia), B. pulchra (lowland Andes of Peru) 

Unknown: B. excelsa, B. venusta (Colombia, Brazil) 


546 Annals of the 
Missouri Botanical Garden 


Bonyunia minor 


B. excelsa 


| 
B. antonifolia B. venusta 


Figure 3. Leaf morphology of Bonyunia. Bonyunia antoniifolia (drawn from Prance 5758 [S], B. aquatica (drawn from 
Huber 1932 [NY], B. excelsa (drawn from Restrepo 587 [MO]). B. magnifica (drawn from Prance et al. 22804 [NY], B. minor 
(drawn from Maguire 40482 [NY], left, and Maguire 46143 [NY], right), B. nobilis (drawn from Palacios 2393 [MO]. B. 
spectabilis (drawn from Hahn & Gopaul 5420 [F]). B. superba (drawn from Schomburgk 613 [BM]. and B. venusta (drawn from 
Ribeiro et al. 1103 [NY]. 


Volume 96, Number 4 Grant 547 
2009 Revision of Bonyunia 


| 


Bonyunia minor 


" 
> 


B. magnifica 


B. venusta 


Figure 4. Floral morphology of Bonyunia demonstrating morphological differences in the shape and indument of the 
calyx. pou antoniifolia (drawn from Ferreira 5688 [MO], B. aquatica (drawn from Foldais 9425 [NY ]), B. excelsa (drawn 
from Restrepo 387 [MO]), B. 1 E n from Prance et al. 22804 [NY]). B. minor (drawn from Maguire 46143A [NY]. B. 

nobilis (drawn from Mendoza 9558 [FMB], B. spectabilis (drawn from Hahn & Gopaul 5420 [US], B. superba (drawn from 
Schomburgk 613 [BM]. and B. in (drawn from Ribeiro et al. 1103 [NY]. 


Annals of the 
Missouri Botanical Garden 


2cm 


e 5. Floral and leaf morphology of Bonyunia 
NI [um from Rojas 478 [F, MOJ) from Peru. 


TAXONOMIC TREATMENT 


Bonyunia M. R. Schomb. ex Progel, Fl. Bras. 
(Martius) 6(1): 267, t. 72. 1868. TYPE: Bon- 


KEY To THE SPECIES OF BONYUNIA 


yunia superba M. R. Schomb. ex Progel, Fl. Bras. 
(Martius) 6(1): 267, tab. 72. y (lectotype, 
designated by Leeuwenberg, 1969: 158). 


Branched shrubs or trees, 2-40 m tall, glabrous to 
hispid; stems round, solid. Leaves opposite, short to 
long-petiolate; blades simple, pinnately veined, thin 
and membranous to thick and coriaceous, margins 
entire, variously shaped (ovate, oval, elliptic, orbic- 
ular, obelliptic, obovate, lanceolate, to oblanceolate). 
Inflorescence a terminal dichasium of paired cymes, 3 
to 7 flowers per cyme; bracts (organs on peduncles 
that subtend inflorescence branches, or flowers) leafy, 
often similar to the leaves, sessile to short-petiolate; 
bracteoles (organs on pedicels that subtend individual 
flowers) triangular to spatulate. Flowers sessile to 
pedicellate, 5-merous (calyx, corolla, and stamens), 
bracteolate; calyx green, campanulate to urceolate, 
glabrous to hispid, ecarinate to slightly keeled, calyx 
lobes triangular, spatulate to obovate, apices acute to 
acuminate; corolla white to yellowish white at 
anthesis, turning pink, red, to purple at maturity, 
trumpet-shaped with a long fused tube to 2/3 of the total 
length, and 5 reflexed linear-long lobes; outside of the 
corola tube and corolla lobes densely pubescent 
(appressed strigose), inside of the reflexed corolla 
lobes glabrous and therefore exposing the corolla color 
(otherwise obscured); stamens 5, included, equal, one 
on each of the corolla lobes inserted just below sinuses; 
filaments adnate to the corolla, the free portion less 

mm; anthers basifixed, linear-oblong, sagittate 
at the base; pistil included within the corolla tube, 
much shorter than the stamens, hispid; ovary superior, 
ovoid, placentation axile; style straight; stigma bilobed, 
he lobes inate patulate; style and 
Capsules septicidally 
dehiscent, ellipsoid to obovoid, pubescent, erect, tan 


the S acute, acum 


stigma deciduous in fruit. 


o 40 seeds per capsule); seeds oblong, 
flattened, yet m concave to one side, winged all 
around; seed body ellipsoid, brown; seed wings straw- 


ulate. 
n discrete suites of morpholog- 
ical ee vrl habit and leaf, calyx, and 
seed morpho ome ue char: 
ntum (hispid i 
aquatica, and B. spectabilis vs. glabrous in all other 


acters include 
indume n Bonyunia antonitfolia, 
species), calyx shape (urceolate in B. magnifica vs. 
campanulate in all other species), and calyx lobe 
shape (obovate in B. superba vs. triangular in all other 
species). 


la. ae d bows throughout a. nO to hispid). 
bro: 


aves co ently broadly to owly obovate; lo 


Nene pe Colombia, and done ihe upper Rio Negro in Brazi 


wlands along the Orinoco River and its tributaries in 
. aquatica 


Volume 96, Number 4 
2009 


Grant 
Revision of Bonyunia 


189) 
> 


. Leaves elliptic or ovate. 


3a. Eaves Phe uen oe 2—4 mm), elliptic to ovate, rarely obov 
cles hispid; calyx 


and i major tributaries in Brazil a nd Bolivia 


and peduni 


g the Amazon 


3b. Leaves es petiolate ‘pet 5-10 mm), elliptic to arro 
o acute; stems glabrous, peduncles hispid; calyx densely appressed le e to 
a 8. 


hispid: Mun n. p 


= 
za 


from the pedicel beneath, o 
promine ent, Ta 
4b. Calyx lobes 
le jas than the calyx. 
5a. Leaves epe -coriaceous; leaf bas 
thro with 


ughout to rarely tufts of hairs ai 


Uyana 2. ee eee 

Calyx glabrous or e so, a of a hairs may occur on calyx lobe AT or hairs may approach the caly 
parse may occur, especially in Peruvian material. 

to much RAM the len "i of the calyx e eid 

tulat qualing to exceeding the 

rian Br ds saline ter du the length of the calyx tube; b 


ate, base cuneate to rounded, apex 
ressed strigose to hispid; widespread in lowlands 
"— má . antoniifolia 
ong attenuate to 


nearly narrowly lanceolate, base lo 


B. ine 


calyx in MM Mies and cas superba. 
ba to a 


es rounded to s ihe E de. o to n 
t the 


t part under x; calyx lobes 


Leaves sadi JR m 4. a —8.6) cm long, adaxial pura even in EN se dem bos brown, wings straw- 


ld; savanna shn 


Chiribiquete in um 


up PET 

up to 10 cm long, adaxial surface minutely etched me a lighter color s small 
lire cv eg seed body and 

Calyx e, glabrous, striate, with tufts of hana on calyx lobe apices; Sena de 


2-10 m tall at higher-elevation savanna and tepui habitats of 


. minor 


wings dark orange; forest trees 7-20 m 


B. nobilis 


7b. Calyx urceolate to penat calyx lobe apices glabrous; terra firme of the Amazon lowlands 
of 


Brazil 


5b. Leaves thin coriaceous; leaf bases cuneate; pedicels pubescent throughout; calyx lobes E. p acute. 
; P 


8a. 


. B. magnifica 


B. pulchra 


l. 
Stems glabrous, to hispid along one side, but branches of the inflorescence always hispid; leaves 
with 5 to 7 pairs of nearly parallel to arching secondary veins; Sierra de E in 
bi 3. B 


Colombian e neis A Liu tte LL E . excelsa 
9b. Stems and bran fl qually lightly hispid; leaves with 3 to 6 pairs of ae 
0. B. venusta 


secondary veins; Amazon lowlands of Brazil 


1. Bonyunia antoniifolia Progel, Fl. Bras. (Martius) 
6(1): 288. 1868. TYPE: Brazil. Mato Grosso: “In 
saxosis S. da Chapada" E Ana da Chapada, 
near Cuiaba], Sep. . Riedel 1149 
(holotype, BR o 2 not seen, MO!, 
NY?). Figures 1, 2C, 3, 4, 6A, B 
Branched shrub to tree to 5-9 m tall, densely 
hispid throughout, glabrous only on adaxial leaf 
surfaces; trunk to 10-30 cm diam.; bark thick fissured 
to corky. Leaves oval, elliptic, to ovate, rarely obovate, 
Pi petiolate, 3.2—7.2 cm, petiole 2-4 mm; blades 
3.5-7.4 X 
ally, slightly glossy adaxially and more 
axially, adaxial surface smooth with some slightly 
impressed veins, abaxial surface with slightly raised 
ary veins; base cuneate to rounded; apex 
obtuse to rounded. Inflorescence 2-8 cm; branches 
1-6 cm; bracts oval, ovate, to obovate, short-petiolate, 
4-20 x 2-15 mm; base cuneate to rounded; apex 
obtuse to rounded; bract petioles 1— 
flower sessile to subsessile, secondary "n pedi- 


1.8-5 em, thin-coriaceous, darker adaxi- 
opaque 


cellate; pedicels 1-5 mm; bracteoles triangular, 0.75— 
X1.5-2 mm, 
appressed strigose to hispid, ecarinate; calyx lobes 
triangular, 0.3—1 X —2 mm, apex acute; corolla 
me nbc BO X I-L S min: s 7-9 X 
0.75—2 mm, apex rounded to obtuse; stamens includ- 


0.5-1 mm. Calyx campanulate, 2-3 


ed; filaments less than 0.5 mm; anthers 1.8-2 X 0.3— 
0.5 mm; pistil ca. 8 mm; ovary ovate, orbicular to 
obovate, 1—1.5 X ca. 1 mm; style 6—7 X 5mm; 
stigma bilobed, each lobe spatulate, 0.50—0.75 X ca. 
0.5 mm. Capsules ellipsoid to ovoid, 12-16 X 5-8 mm 
(excluding style base), tan, 4 to 12 seeds per locule (e.g., 
8 to 24 seeds per fruit); seeds 4.5—7.5 X 2-2.75 mm, 
seed body brown, seed wings straw-gold, reticulate. 


P, ; n DS E 


Morphology and similarities. 
is distinctive in being densely hispid throughout (as in B. 
aquatica and, to a lesser extent, B. spectabilis) and having 
similar-sized elliptic to ovate leaves with a cuneate to 
rounded base and apex. It appears to be most similar to B. 
aquatica, B. minor, B. superba, and B. spectabilis. These 
five species share flattened seeds with a large, prominent, 
brown seed body and tan-colored wings. 


Distribution and habitat. Bonyunia antoniifolia 
occurs in grasslands, fields, or open savanna on terra 
firme, throughout on white sandy to stony soils. It has 
a broad distribution in the Amazon River basin and its 
tributaries in Brazil and Bolivia (Fig. 1) at elevations 
of 80-800 m. It has the broadest distribution of all 
species in the genus and a rather consistent 
morphology across its range. 


IUCN Red List category. Bonyunia antoniifolia 
has a broad distribution in Brazil and Bolivia and has 


Annals of the 
Missouri Botanical Garden 


been collected in formally protected areas such as the 
Parque Nacional Noel jM Mercado Nela It is 

assigned a preliminary IUCN status of Least Con: 

(LC) as set forth in the IUCN Red List a d 

Criteria (IUCN, 2001). 


Etymology. The epithet is taken from Antonia and 
the Latin “folium,” named for the resemblance of its 
leaves to that of the genus Antonia (Loganiaceae). 


Typification. The disparity of the type locality as 
published in the protologue of Bonyunia antoniifolia 
and printed on the herbarium label of its type (L. 

Riedel 1149) i is resolved here. The 
Progel identifies the collection site as “Serra da 
apada, Prov. Minarum,” 


references to Botnia have listed the type or at least 


protologue of 
and accordingly, most 


B. antoniifolia as coming from Minas Gerais, Brazil 
Ducke, 
: apada. 
Sept 1827. Brazilia. Riedel Nro. 1149 Rubiacea." In 
the list of itineraries for Flora Brasiliensis, it is noted 
that el did not collect in 
ee er 1827, : 
Lourengo ad m (L-IX. 27), Serra Bou. (V., 
VL)" (Urban, 1906: 91). The collection site of Serra da 
Chapada probably refers to Santa Ana da Chapada, 
which is near Cuiabá in Mato Grosso. This fits in 
perfectly well with the distribution of B. antoniifolia 
as mapped here. Bon es not occur in Minas 
Gerais. A ai af Riedel 1149 should be at R since a 
full set of his material was deposited there; however, a 
recent search at R has uncovered a catalogue 
indicating that Riedel numbers 1147 to 1182 were 
never sent (A. Costa, pers. comm.). The MO and NY 
sheets are identified here for the first time as isotypes 
th match the 

labels are completely different. Ludwig 
Riedel, the collector of the type, deposited the main 
set of his collections at St. Petersburg, Russia n 
and the second in Rio de Janeiro. The MO and N 
sheets were distributed from LE, and the labels are 
standard-issue typeset labels in Russian, with only the 


material on the 


following handwritten information to identify the 
specimen: “Indetercei, No 1149, Brasilia, Riedel." 


Pr examined. BOLIVIA. Beni: Itenez, Serrania 
R. Quevedo et al. 975 (G, MO, NY, USZ. not seen); 


carr. a Riberalta, 
W 


cado, E. Gutiérrez et al. d (G, MO, NY, USZ not an 
T Killen 2749 (NY); Nuflo de Chavez, Mesetea de 
Caparuch, E. Gutiérrez et al. 1444 (MO, USZ not seen); 


Velasco, Parque Nacional Noel Kempff Mercado, Serranía de 
Caparuch, 750 m, T. Killeen e al. 6501 (F, MO, NY, USZ not 


NY, USZ not pun Parque Nacional Noel Ke anne Mercado, 
Serrania S y NE de la posta Noel Kempff Mercado, 
Mostacedo et al. 1861 (G, MO, NY, USZ not seen); ie 
Est. Flor de Oro, margen del Río Iténez (Guapaoré), 30 km N 
Serrania de Huanchaca, ca. 85 

R. B. Foster 170 (NY). BRAZIL s. 
56°06'W, B. Pena 2005 (MG not seen, RB). A 
Rio Negro infer., Bahia Boiassá, Camp Amelia, A. Ducke 379 

uc 


m], A. Ducke 738 (F), A Ducke 5738 (S. A. 
Ducke ae (BM, G, RB, S, U), A. Ducke 11534 m A. 
Ducke 12197 (BM, G, P, US); Coary, E A. Ducke 
12384 (BM, G, RB, US); Novo Aripuana, BR 150 a al. 
L de Humaitá e 30 km para o S na rodovia ds Me o, C. A. 
C. Ferreira 5603 (MO, NY, WAG); Humaitá Fazenda Paraiso 


dos Campos, A. ibis, 640 (M); Humaitá, Fazenda Arlindo 
ció Janssen & V le 306 (M); Humaitá, 
Campos at Km 15 d. us, K. Kubiizki & H. H. 


fca 79-47 (MG not seen, NY) Rd. Porto Velho— 
Humait 75, E. pr et al. P19452 Me Meo Rio 

argem dire O km acima de Manaus, Campo 
Amelia (Paz Belo Horizonte) B. Nelson et a "1368 (MO, 


W. Rodrigues et a 8547 7 (US) Humaitá, E Humaitá- 
Porto Velho, Km 38, L. VP 
O. A 


Teixeira et al. 1268 (MG not seen, MO, NY, RB, US, WAC): 
rodovia do nho, margem da rodovia 150 km de 
Humaitá, G. Vieira et al. 1493 (K, NY, US, WAG). Mato 
Grosso: Vale Cuapo 
Castanheira, T. B. Crean et al 2377 (K); ¿Cuirainga, 
Morro d mbrós 356 (US); Proc. Mpio. 
Paula 1907. z J; Novo 

od Transamazônica, 
. C. Ferreira 5688 (MO, NY, WAG); Cuiabá, MT. Río 
ur prox. a Cachoeira Véu de Noiva, C. A. C. 
Ferreira 6540 (MO, NY, WAG); Entre, J. G. Kuhlmann 2233 
(SP); Colider, estrada Santarém-Cui 63, Km 762, 
Serra do Cachimb 
Silva et al. 28 (MO, NY, WAG). Para: Itaituba, estada 
S 4, BR 163, Km 877, e do Cachimbo, L 
NY). R 


E. 
E 

ES 
E 


antarém—Cuiabá 
L. Amaral et al. 112495 
Madeiro, Km 215 
, G. T. Prance et al. 5758 (COL, F, GH, K, MG not 
seen, NY, R, S, U, US, Z) 


2. Bonyunia aquatica Ducke, Arq. Inst. Biol. Veg. 
I: 211. 19035. TYP. i “fre- 


quens in ripis profunde et permanenter inun- 


razil. Amazonas: 


dates fluminis Curicuriary inferioris (affluentis 
Rio Negro superioris, civitate Amazonas),” 21 
Dec. r ucke 23760 (holotype, RB!; 
isotypes, G!, K!, P!, RB [2], S! , US». 
Figures 1, 2C, 3, 4, 6C, D. 


Volume 96, Number 4 
2009 


Grant 
Revision of Bonyunia 


Branched shrub to tree to 2-8(-15) m tall, densely 
hispid throughout especially on the midvein of the 
undersides of leaves, petioles, peduncles, inflores- 
cences, calyces, and corollas (glabrous only 
adaxial leaf surfaces). Leaves broadly obovate to 
narrowly obelliptic, short-petiolate, (2.5—5—7(-9) em, 
petiole 2-6 mm; blades (2.355-7(-8.6) X (1.853- 
5.2 em, thin-coriaceous, darker adaxially, lighter 
abaxially, slightly glossy adaxially and more opaque 
abaxially, adaxial surface smooth with some slightly 
impressed veins, abaxial surface with slightly raised 
secondary veins; base aequilateral; apex rounded to 
retuse. Inflorescence 3-9 em; branches 2.5-6 cm; 
bracts rry to E sessile to short-petiolate, 
(3-)6-28 x (13-1 ; base a 


E to retuse; p EROS 0-2 mm; primary 


equilateral; apex 


flower sessile to subsessile, secondary flowers pedi- 
cellate; pedicels 0.5-3 mm; bracteoles triangular, 
0.75-1.75 X 0.5-0.75 m 
3.5 X 1.5-2.5 mm, appressed strigose 


m. Calyx campanulate, 2.5— 
to hispid, 
ecarinate; calyx lobes triangular, 0.3-1.2 X 1.5- 
2.5 mm, apex acute; corolla 17-32 mm; tube 8-19 X 
1.3-2 mm; lobes 9-13 X 0.75-1.2 mm, apex rounded 
to obtuse; piae ver filaments less than 
0.5 mm; anthers m; pistil 7— 
9 mm; ovary im 1 X ca. 1.5 mm; nes ca. 7 
X 0.5 mm; stigma bilobed, each lobe spatulate, 0.75— 
1 X 0.3-0.5 mm. Capsules ellipsoid to obovoid, 13— 
22 X 8-9 mm (excluding style base), tan, 3 to 12 
seeds per locule (e.g., 6 to 24 seeds per fruit); seeds 
7-9 X 3-4.5 mm, seed body brown, seed wings straw- 
gold, reticulate. 


Morphology and similarities. | Bonyunia aquatica 
is distinct in its obovate leaves with rounded to retuse 
apices. Its inflorescence and calyx are hispid as in B. 
antoniifolia and B. spectabilis, while other species are 
glabrous or nearly so. Bonyunia aquatica appears to 
be most similar to B. antonüfolia, B. minor, B. 
superba, and perhaps B. spectabilis. 


Distribution and habitat. Bonyunia aquatica is a 
facultative emergent tree of inundated forests along 
riverbanks and on white sand savannas. It ranges 
principally along the lowlands of the Orinoco River 
watershed in Colombia and Venezuela, but also 
extends to the upper Rio Negro in Brazil at elevations 


of 80-180(-350) m 


that of outlier populations of B. minor in southwestern 


(Fig. 1). Its range overlaps with 


Venezuela 


IUCN Red List category. Bonyunia aquatica 
occurs in lowland areas of Brazil, Colombia, and 
Venezuela and has been collected in several formally 

Florestal do 
Rio Negro (Brazil) and the Parque Nacional Natural 


protected areas such as the Reserva 


El Tuparro (Colombia). It is assigned a preliminary 

IUCN status of Least Concern (LC) as set forth in 

ps IUCN Red List Categories and Criteria (IUCN, 
01). 


Etymology. The epithet is taken from the Latin 
“aquaticus,” meaning “living in or near water,” for its 
facultatively aquatic habit, growing along river 


corridors and inundated forests. 


ecimens examined. . Amazonas: 


o Cur 
curiary affl. Rio Negro, A. Ducke p^ (NY [2], A. p 354 


M. Pires et al. 14155 (IAN, RB). heres Guianía: Río 
Guiania, Puerto Colombia (opposite Venezuelan town of 
e jer Sapo, R. E. Sulis et al. 18240 (CH, US 
2). hada: Cumaribo, ue Nacional Natural El 
ree "H Mendoza & A. Robles 15691 (FMB). VENE- 


5973 (MO, NY); Atab 
et al. 16912 (MO); Rio Negro, lower part of the Rio Baria, G. 
Davidse 27650 (WAG); Rio Negro, Rio Pasimoni, betw. its 


; Atures, Río Guayapo, E. Foldats 
& J. Velazco 9423 (NY); Río Negro, final de la aa enta” N 
1 (MO); Atures, 
riberas del Río Sipapo desde la boca del Río Guayapo, F. 
ete 2641 (MO); dn p E de Canaripó, 
nos 20 km al. E de l 
con el Río Orinoc 0. Huber 1932 (COL, K, NY 
WAG); Atabapo, e ubicada a unos 10 km al. NE 
del Cerro Moricha, en la ribera E del medio río Ventuarí, 
S de la 


la Serranía Unturan, 


apiare y Vena O. 
Huber 3470 (B, NY, WAG); Atabapo, bajo Río Ventuarí, a 


= 

I 

^ 

= 

2 

> 

a 

E 

o 

B E 

S2: 
E 

= 

Qe 

$ 

ON 

re 

No 


Río Ventuarí, B. Maguire et al. 31012 bier F, NY, S, US 

[2], WAG); Atabapo, en a Suelo Faa, E. Marin 1126 

E de Sta. Bárbara del Orinoco, 
Ri 


opposite mi 
S. Adderley 42814 (COL, F, NY, S, US, WAG); Río 

Cuin betw. Cano San Miguel & Maroa, J. J. Wurdack & 
L. S. Adderley 43261 (NY, WAG). 


3. Bonyunia excelsa J. R. Grant, sp. nov. TYPE: 
Colombia. Caquetá: Solano, Parque Nacional 
Natural Serranía de Chiribiquete, cuenca media 
del Río Cufiare, creciendo sobre suelo arenoso 

en donde 

predominan Pagamea thyrsiflora y Tepuianthus, 
00°29'55.32"N, 72%37'11"W, 350 m, 15 Nov 

2002, H. Mendoza, A. bar, S. Medina & M. 

Leptuama 9456 (holotype, FMB!). Figures 1, 3, 


en un sitio cercano a un tepul 


552 Annals of the 
Missouri Botanical Garden 


LOAN FROM BR TO 
DEI 


1 Types exemplar specimens of Bonyunia. —A. Isotype of B. antoniifolia Progel (Riedel 1149 [BR]. —B. 
E e B. os Progel (Prance et al. 5758 [US]). —C. ee of B. aquatica Ducke (Ducke 23760 [US]. —D. 
Specimen of B. aquatica Ducke “Huber 1932 [NY]. 


Volume 96, Number 4 
2009 


Grant 
Revision of Bonyunia 


Species nova Bonyunia antoniifolia Progel cui affinis, sed 
ab ea habitu arboris excelsae (4.5—35 m vs. 5-9 m), foliis 
Ed ege glabris vel hispidulis atque lobis calycis 

vel acutis differt; etiam Sierra de Chiribiquete 


Branched tree to 4.5—35 m tall, hispidulous on 
petioles, stems (along one side), peduncles, inflores- 
cences, calyces, and corollas. Leaves obovate to 
A 4.5-6.2 cm, aon) 3-6 mm; blades 4—5.8 

labro thin-cori darker 
PTUS lighter discs slightly E ely 
and more opaque abaxially, with 5 to 7 pairs of nearly 
parallel to arching secondary veins, smooth with 
slightly impressed veins adaxially, with slightly 
e cuneate; 
5-9 em; 
branches 2-5 em, always hispid; bracts obovate to 
spatulate, sessile to short-petiolate, 20-28 X 7— 
et 


ase attenuat cuneate; apex obtuse, 


m 
acute, to rounded; bract petioles 1-3 mm; primary 
flowers subsessile, secondary flowers pedicellate; 
pedicels 1—7 mm; bracteoles linear-triangular, 1.5— 
3 X 0.3-0.75 mm. Calyx campanulate, 2-3 X 2- 
2.5 mm, glabrous to hispidulous, ecarinate; calyx 
lobes triangular, 0.5-1 X 2-2.5 
nate to acute; corolla 12-14 mm; tube 6-7 X 1- 
1.5 mm; lobes 6-7 X 0.75-1 mm, apex rounded to 
than 
m; pistil ca. 


mm, apex acumi- 


obtuse; stamens included; filaments less 
0.5 mm; anthers 1-1.25 Xx 0 

10 mm; ovary ovate, 1—1.5 X ca. 1 mm; i 1-8 X 
0.3-0.5 mm; stigma bilobed, each lobe acuminate, 
0.75-1 X 0.3-0.5 mm. Capsules ellipsoid, 23-30 X 
8—9 mm 


bilocular; seeds unknown 


(excluding style base), brown, erect, 


Morphology and similarities. | Bonyunia excelsa is 
a distinctive species in the open diffuse branching 
pattern of the inflorescence. It appears to be most 
similar to B. venusta in generally diffuse inflores- 
cences, yet differs in having generally glabrous stems 
that are hispid along one side, branches of the 
inflorescence that are hispid all the way around, and 
leaves with five to seven pairs (vs. three to six pairs) of 
nearly parallel to arching secondary veins. It differs 
from B. antoniifolia in being a tall tree (4.5-35 m tall 


hispid), with 


acuminate to acute calyx lobes. 


dulous vs. glabrous calyces and 


Distribution and habitat. | Bonyunia excelsa occurs 
on plateaus and river basins on sandy soils. Both B. 
excelsa (230—350 m) and B. nobilis (350—400 m) are 
known from the Sierra de Chiribiquete, which is an 
isolated outlier or tepui of the Guayana region in 
Colombia (Fig. 1). Bonyunia excelsa has also been 


found in the Araracuara region and Río Mesay. 


IUCN Red List category. | Bonyunia excelsa is only 
known from three collections, one collected inside a 
formally protected area, the Parque Nacional Natural 
Serranía de Chiribiquete (Colombia). It is assigned a 
preliminary IUCN status of Vulnerable (VU) according 
to IUCN Red List Categories and Criteria (IUCN, 2001) 


Etymology. The epithet is taken from the Latin 
"excelsus," meaning “high” or “lofty.” 


Paratypes. COLOMBIA. dee Araracuara, meseta 
de areniscas, 200—300 m, D. Restrepo & A. Matapi Pid 
Solano, Río Mesay, pis de 1 Yavilla, 230 m, D. Cárdenas 
al. 6788 (COAH not seen, MO). 


4. Bonyunia magnifica J. R. Grant, sp. nov. TYPE: 
BR 319, Km 190, 
o hwy., forest on terra firme, 11 Oct 
1974, G. T. Prance, T. D. Pennington, M. 
Leppard, P. P. Monteiro & J. F. Ramos 22804 
(holotype, NY!; isotypes, INPA not seen, K!, MG 
St, Ut, USt, WAG). Figures 1, 


il. Amazonas: Manaus- 


not seen, MO!, 


2C, 3, 4, 7B. 


Species nova Bonyunia minor N. E. Br. et B. nobilis J. R. 
Grant cui affines, sed a hac eee lobis apice c in ab illa 
habitu arboris excelsae (usque ri 
9.] X 2.5—5.5 em) brad 
aurantiacis, ab ambabus calyce urceolato vel campanulato 
differt. 


Branched tree to 20 m tall, glabrous throughout, 
except for hairs on the midvein of the undersides of 
leaves, pedicels, corollas, and fruits; trunk to 20 cm 
diam. Leaves ovate to oval, short-petiolate, 4-10 cm, 
petiole 3-5 mm; blades 3—9.1 X 2.5-5.5 em, thick, 
coriaceous, glossy on both surfaces, distinctly discol- 
ored with the adaxial leaf surface distinetly olive- 
green and speckled, and the abaxial surface a solid 
etched revealing a lighter color 


yphic-s haped marks, 


gold-green, distinctly 
ith 


adaxia small 


vein and secondary veins; base rounded to slightly 
cordate to rarely cuneate; apex obtuse to acuminate. 
Inflorescence 5-9 cm; branches 2—7 cm; bracts ovate 
O X 14-22 mm; base 


oval, short- pal 
ightly cordate to rarely cuneate; apex 


rounded to 
obtuse to acuminate; Pie petioles 1-2 mm; primary 
flower sessile to subsessile, secondary flowers sessile 
i i 0-3 mm; bracteoles 


o short-pedicellate; pedicels 


triangular, 1.5-3 X 0.5-1 mm. Calyx urceolate to 
campanulate, 4-6 X 2.5—4 mm, glabrous, ecarinate; 
calyx lobes triangular, 0.5-1 X mm, apex 
acute; corolla 10-13 mm; tube 6-7 X 1-1.5 mm; 
lobes 5-6 X puis mm, apex rounded to obtuse; 
stamens and pistil unknown. Capsules ellipsoid to 
obovoid, 15-23 X 7— 
3 to 4 seeds per locule (e.g., 6 to 8 seeds per fruit); 


8 mm (excluding style base), tan, 


Annals of the 
Missouri Botanical Garden 


seeds 14—14.5 X 1.5-3 mm, seed body dark orange, 


seed wings dark orange, reticulate. 


Morphology and similarities. Bonyunia magnifica 
is unique in the genus in having an urceolate calyx. Its 


both 


surfaces, olive-green and specked adaxially and solid 


large, thick, coriaceous leaves are glossy on 


gold-green abaxially (at least when dried), with a 
brown, raised midvein adaxially and 
The adaxial leaf surface 
of both B. magnifica and B. nobilis is covered in light 


prominent, thick, 
rounded to cordate leaf base. 


specks caused by the deterioration or etching of the 

upper layer of cells, creating a uniform specked 
surface; the specks when examined under microscope 
have the appearance of alphabetic or hieroglyphic 
marks of differing shapes. The seeds of these two 
species have a three-dimensional aspect where the 
wings are slightly curled, have a less prominent 
difference in color between the bodies and wings, and 
are rather orangish in color. Bonyunia magnifica is 
related to B. nobilis, with which it shares its unique 
leaf and seed morphology as described above. It 
differs from B. nobilis in its urceolate to campanulate 
calyx with glabrous calyx lobes, and from B. minor in 
being a 7-20 m tall forest tree with larger discolored 
leaves (3-9.1 X 2.5-5.5 cm 


Distribution and habitat. 
occurs in primary forest on terra firme in the Amazon 


Bonyunia magnifica 


lowlands of Brazil. It is only known from its type 
collection found at Km 190 on BR 319, the Manaus— 
Pórto Velho Highway. Although the elevation is not 
specified, B. magnifica was certainly fou 

50 and 100 m, the gen 

Amazonia. It is disjunet a its morphologically most 


d between 
lan 


E. 


ral elevation of low 


similar species, B. nobilis of Colombia. 


IUCN Red List category. Bonyunia magnifica is 
only known from the type collection. It is assigned a 
preliminary IUCN status of Critically Endangered 
(CR) according to IUCN Red List Categories and 
Criteria (IUCN, 2001). 


Etymology. The epithet is taken from the Latin 
“magnificus,” meaning “magnificent.” 


5. Bonyunia minor N. E. Br., Trans. Linn. Soc. 
Mie Bot. ser. 2, 6: 49, p 9, figs. 1—5. 1901. 
TYPE: Guyana. Mt. Roraima Exped., Kotinga 
Va e autumn 1894, J. J. Quelch & F. 
McConnell 161 eats, designated by Leeu- 
wenberg, 1969: 156, K!). Figures 1, 2A—C, 3, 4, 
7C 


Bonyunia cinchonoides p & Standl., Bull. Torrey Bot. 
Club 58: 448. 1931. TYPE: Venezuela. Amaz ae 8: 
summit of Mt. Duida, ps ft, Savanna Hills, 
1928-Apr. 1929, G. H. H. fale 770 (elote, n 
isotype, F!). 


Branched shrub to tree to 2-10 m tall, glabrous 
throughout, except for hairs on the petioles, pedun- 
cles, pedicels, and corolla; trunk t cm 


(Maguire & Fanshawe 3252 


individuals in savanna habitat (smaller, broadly ovate, 


jam 


8). Leaves variable from 


ovate, to nearly orbicular or reniform and distinctly 
cordate at the base) to forest plants (larger, ovate to 
oval, and nearly cuneate at the base), short- PEE 
.6 cm, petiole 1-4 mm; blades (1 x 
0.8— 133-45 cm, thick coriaceous, darker I 


lighter abaxially, adaxial surface smooth with some 


=e 


slightly impressed veins, abaxial surface with prom- 
inently raised midvein and secondary veins; base 
cordate, to rounded to cuneate; apex rounded to 
obtuse, to acute to nearly acuminate. Inflorescence 
1.5-6 em; branches 1—5 em; bracts (generally as in 
the leaves) spatulate, obovate to ovate, short-petiolate, 
5-17 X mm; base attenuate to cordate; apex 
obtuse to rounded; bract petioles 1-2 mm; primary 
and secondary flowers generally sessile; pedicels 0— 
1 mm; bracteoles triangular, 1—4 X 0.5-1 
2-3.5 


sometimes receiving a few 


mm. Calyx 
campanulate, 2-2.5 mm, glabrous (to 
hispid hairs from the 
pedicel or peduncle), ecarinate; calyx lobes triangu- 
lar, 0.5-1.5 X 2-2.5 mm, apex acute; corolla 11— 
17 mm; tube 7-9 X 1.25-2.25 mm; lobes 4-8 X 
0.75-1 mm, apex rounded to obtuse; stamens includ- 
ed; filaments less than 0.5 mm; anthers 1.8-2 X 0.5— 

5 mm; pistil 5.5-6 mm; ovary ovate, 1-1.5 X ca. 
lmm; style 4—4.5 X 0.2-0.4 mm; stigma bilobed, 
each lobe spatulate, 0.5—0.75 X ca. 0.5 mm. Capsules 
ellipsoid to obovoid, 15-24 X 5-8 mm (excluding 
style base), tan to brown, 1 to 4 seeds per locule (e.g., 

to 8 seeds per fruit); seeds 8-10 X 2-3 mm, seed 
body brown, seed wings straw-gold, reticulate. 


Morphology and similarities. Bonyunia minor is 
not only the most morpho = ally variable species in 
the genus, but is also easily identifiable by its 
generally small cordate coriaceous leaves. Part of this 
variation led to the naming of B. cinchonoides Gleason 
& Standl., later reduced to synonymy under B. minor 
by Leeuwenberg ( : 156) and also accepted here. 
The leaf morphology of B. minor varies considerably 
depending on whether the plants occur in open 
savanna or closed forest. Specimens collected in 
savanna tend to have smaller, more closely bunched 
a tight, bushy i 

lected in 


have larger leaves in a lax branching pattern. In the 


leaves ranching pattern, 


nd have 
while specimens col closed canopy forest 
savanna, the leaves are smaller, broadly ovate, ovate, 
to nearly orbicular or reniform, distinctly cordate at 
the base and rounded to obtuse at the apex; in the 
forest the leaves are larger, ovate to oval, and cuneate 


to rounded at the base, to acute to nearly acuminate at 


Volume 96, Number 4 Grant 555 
2009 Revision of Bonyunia 


Figure 7. es and exemplar 1x of Bonyunia. —A. Holotype of B. excelsa. J. R. Grant (Mendoza et al. 9456 
). —B. mus of B. magnifica J. R. Grant (coe et al. 22804 [NY]). —C. ee of B. minor N. E. Br. (Quelch & 
McConnell 161 [K]. —D. Holotype of B. nobilis J. R. Grant (Palacios et al. 2393 


Annals of the 
Missouri Botanical Garden 


the apex. Bonyunia minor appears to be most similar 
to B. antonüfolia, B. aquatica, B. superba, and 


perhaps B. spectabilis. 


Distribution and. habitat. 
distinctive shrub or tree of savanna, forest edges, and 


Bonyunia minor is a 


in Venezuela and 
uyana, especially of the Gran Sabana (Fig. 1). Its 
range overlaps that of B. aquatica to the east and B. 


tepui habitats in the Guayana region 
G 


spectabilis and B. superba to the west. It also has the 
widest range in elevation in the genus, ranging from 
100-1450 m 


IUCN Red List category. Bonyunia minor occurs 
nd Guyana and 


has been collected in several formally protected areas 


in the Guayana region in Venezuela a 


such as the Parque Nacional Canaima, Parque 
Nacional Duida-Marahuaca, and Parque Nacional 
Yapacana of Venezuela. It is assigned a preliminary 
IUCN status of Least Concern (LC) as set forth in the 
IUCN Red List Categories and Criteria (IUCN, 2001). 


Erymo logy. The epithet is taken from the Latin 


“minor,” “less,” for the small stature of some 


individuals, at the time described, probably in 
comparison to the only two other species in the genus 


at the time: Bonyunia antoniifolia and B. superba. 


Sp ZIL. Roraima: Serra dol Sol, 
Ph. Lueizelburg 21503 (M) Quino Igarapé, Ph. 
Pr 21514 (M); Brazilian side of divide near Serra 

ae . Maguire & bs K le 40374 (NY, RB). 
GUYAN Mazaruni Distr., 
dices Forest Dept. “of Br rem puc 7921 (K, NY). 
taro-Siparuni: Pakariama Mtns., Upper Ireng River 
watershed, Malakwalai-Tipu, T. W. Henkel 5699 (US); 5 
Pakaraima Mins., margins of Chima 
et al. 46143A (F, 
Mazaruni River, Samwarakna-tipu (Holi-tipu), B. Maguire & 
Fanshawe 32528 (K, NY, P, UC); 1842, Robert Schomburgk 


ecimens examined. BRA 


pu Savanna 


(BM, P [not seen, but listed by van Dam, 2002]; Utshi 
River trail to San lena, H. D. Clarke 913 (NY, US); 
Paar 0.5-6 from trail to Youwang & Monke 


(e D. c. 1238 o Es US); Mt. Roraima Exped., 
J. Que F. McConnell 331 (K). 


NW ud B. Maguire et al. 30525 (NY, WAG); Cerro 


Moriche, Río Ventuarí, B. Maguir (NY); summit of 
Mt. Duida, Savanna Hills, J. A. Sema 58291 (F, NY); 
summit of Mt. Duida, Savanna Hills, Ti e 770 (F [isotype], 


Canaima, G. Agostini 358 (US); Gran Sabana, desde Santa 
Elena en el Km 274, C. Benitez & W. G. D'A 
Río Caroui, region de Urim 
Region de los ríos Icabari, Hacha, A. L. Bernardi 2624 (NY); 
2 res Ayavaparú, 10-15 km W of Wadakapiapuéte- 

P. E. Berry & L. Brako 5522 (MO, NY), P. E. Berry & L. 
Brako 5525 (MO, NY); Gran Sabana, Km 195 S of El Dorado, 
P. E. Berry & L. Brako 5530 (MO, NY); Luepa along unpaved 
rd. to Minicentral La Ciudadela, P. E. Berry et al. 6559 (MO); 
Ucaima, 600 m, J. Bogner 1080 (K, M) orillas del Río 


Uaipará, afluente del Ikabarú, Caroní, F. Cardona o 
o 


Chacon 595 (MO), L. Chacon 624 (MO), L. Chacon 679 MO); 
und entre Icabarú, altoplano del 


Sta. Elena e plan 
o, G. Colonnello -Aznar $45 (MO); Piar, descent from 


section of Río Ambutuir, al 
O. Huber 23067 (NY); Sifontes, Gran Sabana, Kavanayen, A. 
Fernandez & B. Bracamonte 3186 (MO); Atures, 1 km abajo 
derecha del río Autana, 


e Meseta del C 0. H b 


O. Huber 11954 (AAU, NY); Heres, Meseta del Guaiquinima, 
a lo largo del Río Carapo, aprox. 8 km al. N del Salto Carapo, 
O. Huber 12382 (AAU, NY, US); Roscio, cuenca del Rio 


n, 7. Lasser 
ae (US, N 
Río Las Ahallas, R. L. Liesner 19300 (MO); 0-4 km N of El 
dM 2 n e R. L. Liesner 1 E (NY); Gran 
Saban W of Karaurin Tepui at j Río Karaurin 
and Río PU VA Sanpa), R. L. Liesner 123030 (MO), R. L. 
MO) Gran Sabana, ca. l0 km NW of 


Asadon a cee E L. Liesner E (MO, NY); e 
Sabana, WSW of Karaurin Tepui, Quebrad: 

Tanuan, R L pus 24101 (MO); Ciani Sabana, 5 km S of 
San Ignacio de Yuraní, R. L. Liesner 24438 (MO); Piar, Río 
Acanán, 2-5 km SW of SW corner of Amaruay-tepui, R. L. 
Liesner & Holst 20485 (MO, NY); Gran Sabana, Ilu-Tepui, 
Gran Sabana at Kamarang Head, B. Maguire 33296 (NY, W); 
i ic & San Rafael, B. 


guire et al. 461494 (NY, US, W 


& C. K uu eee (IAN, NY, RB); betw. Caju 
& C. K. Maguire 40482 (NY); 
vanayén, 4 km E of Mission, B. 
Maguire & J. J. Wurdack 35994 (G, NY, U); Kavanayén, trail 
an Misión de Santa Teresita de Kavanayén to Río Pakairau, 
Moore, Jr. et al. 9618 (NY); Gran Sabana, E de 
e 1120 m , G. Picón Nava 1191 (US); Sabana de 
Medio Carrao, 8-10 km NNE of the Carrao-Churun conflu- 
ence, G. T. Prance & O. Huber 28457 (MO, NY, WAG); 
Ucaima, Río Carrao above Salto Hacha, G. T. Prance & O. 


Volume 96, Number 4 
2009 


rant 
Revision of Bonyunia 


Huber 28501 (MO, NY, US, WAG); Gran Sabana, 100 m NE 


D 
105489 (NY); Gran Sabana, 2 km al. N la Misión de Santa 


Km 146, al. S de 


eo 
5 
e 
$ 
= 
o 
$ 
* E 
ae 
E 
3 
= 
© 
E 
© 


J. A. Steyermark et al. 131853 (MO); Río Uarama below 
Uarama-tepui, NE of Luepa, Steyermark & L 
Aristegueta 68 (NY, US); Gran Sabana, formación Roraima, 
Río Aponguao, selva de galeria a lo largo del Arautá-parú, J. 
A. Steyermark et al. 104146 (NY, US); Roscio, 7.5 km al. NE 
de Santa Elena de Uairén, Steyermark & R. Liesner 
127575 (MO, NY); Gran Sabana, selvas de galería del Río 
Uari, F. Tamayo 3132 (US). 


6. Bonyunia nobilis 


. Grant, sp. 
Colom de 


TYPE: 
js Chixibiquete, 
del campamento, sobre la 
ladera de la meseta, 00^55'N, 72°45’W, 400 m, 
14 Dec. 1990, P. Palacios, J. Estrada, P. Franco 
& J. Fuertas 2393 (holotype, MO!; isotypes, COL 
not seen, MA not seen). Figures 1, 2C, 3, 4, 7D. 


Pun nova Bonyunia magnifica J. R. Grant et B. minor 
affines, sed a hac habitu arboris eiae e 
ien dbi s atque seminibus atro-auranti 
illa calyce anguste campanulato striato lobis apice qe 
differt. 


Branched tree to 7-8 m tall, glabrous throughout, 
except for tufts of hairs on the base of the pedicels and 
the apex of the calyx lobes. Leaves ovate to oval, 
short-petiolate, 3-9 cm, petiole 1-3 mm; blades 3— 
8.8 X 2-5 em, thick coriaceous, glossy on both 
surfaces, distinctly discolored (adaxial leaf surface 
distinetly olive-green and specked surface 
solid gold-green), distinetly etched, revealing a lighter 
color on the adaxial surface with small alphabet/ 
hieroglyphic-shaped marks, adaxial surface smoot! 
with some slightly impressed veins, abaxial surface 
with prominently raised midvein and secondary veins; 
base cordate to rarely rounded; apex obtuse to acute. 
Inflorescence m; branches 3-7 cm; bracts 
ovate to oval, sessile to short-petiolate, 11-18 X 4— 
10 mm; base cuneate to rounded; apex acute; bract 
petioles 0-1 mm; 


, abaxial 


primary and secondary flowers 


sessile to subsessile; pedicels 0—2 mm; bracteoles 


us except 
for tufts of hairs on apex of lobes, ecarinate, striated 
vertically; calyx lobes triangular, 0.3-0.75 X 2- 
2.5 mm, apex acute; corolla 13-16 mm; tube 7-9 X 
1-1.5 mm; lobes 6-7 X 0.5-0.75 mm, apex rounded 
to obtuse; stamens included; filaments less than 


0.5 mm; anthers ca. 2 X 0.5 mm; pistil ca. 4 mm; 
ovary ovate, ca. 1.5 X 1 mm; style ca. 2 X 0.2— 
0.4 mm; stigma bilobed, each lobe spatulate, ca. 0.5 
X 0.3-0.5 mm. Capsules ellipsoid to obovoid, 22-25 
X 8-10 mm (excluding style base), tan, 2 to 5 seeds 
per locule (e.g., 4 to 9 seeds per fruit); seeds 9-14 X 
1.5-3 mm, seed body dark orange, seed wings dark 
orange, reticulate. 


Morphology and similarities. Bonyunia nobilis is 
unique in its striated, narrowly campanulate calyx 
with tufts of hairs on its calyx lobe apices. Its large, 
thick, coriaceous leaves are glossy on both surfaces, 
i c adaxially, and solid gold- 
green abaxially (at least when dried), with a prominent, 
thick, brown, raised midvein abaxially and a rounded to 
"v leaf base. The leaves and seeds of B. nobilis 
gnifica are similar and described above 
a B. jeder ue It differs from B. magnifica in its 
narrowly campanulate striate calyx that has tufts of 
hairs on the calyx lobe apices, and from B. minor in its 
discolored leaves and dark orange seeds. 


ive-green and spe 


Distribution and habitat. | Bonyunia nobilis occurs 
in forests on slopes of the plateau of the Sierra de 
Chiribiquete, isolated outliers of the Guayana region 

n Colombia (Fig. 1), at 350-400 m, with another 


species, B. d 


IUCN Red List category. | Bonyunia nobilis is only 
known from two collections, both collected inside a 
formally protected area, the Parque Nacional Natural 
Serranía de Chiribiquete (Colombia). It is assigned a 
preliminary IUCN status of Vulnerable (VU) accord- 
ing to IUCN Red List Categories and Criteria (IUCN, 
2001). 


Etymology. The epithet is taken from the Latin 
“nobilis,” meaning “noble 


Paratypes. COLOMBIA. Caquetá: Solano, Parque Na- 
cional Natural Serranía de Chiribiquete, 350 m, H. Mendoza, 
A. Escobar, S. Medina & M. Leptuama 9558 (FMB [2]. 


7. Bonyunia pulchra Ricketson, J. R. Grant € 

Liesner, sp. nov. T eru. Amazonas: Bagua, 

pr de Wawas, 

15'25'S, 7821'41"W, 

. 1997 (fL), R. Rojas et al. 478 

(holotype, Mor isotypes, F!, G!, HUT!, NY!, US!, 
USM not seen). Figures 1, 2C, 5, 8A. 


Imaza, Tayi Mujsj, 
ima ap 5° 


Species nova quoad i f onyunia magnifica 
J. R. Grant et B. nobilis J. 
ab eis foliis tenuiter coriaceis basi cuneatis, pedicellis 
omnino pubescentibus NA calycis lobis acuminatis vel 
acutis differt. 


El 
R. KA ut cae cui affinis, sed 


Branched tree to 1540 m tall, sparsely hispidulous 
on petioles, stems, peduncles, inflorescences, calyces, 


Annals of the 
Missouri Botanical Garden 


and corollas (leaves mu uos random hairs on 
abaxial surface); trunk BH (Morawetz & 
Wallnófer V79-13888; ae E 121088). Leaves 
elliptic, obovate to ovate, 3-6.5(-9.3) em, petiole 2— 
6 mm; blades 2.8-6.2(-9) X 1.2-2.7(-5) cm, thin- 
coriaceous, darker adaxially, lighter abaxially, slightly 
glossy adaxially and more opaque abaxially, adaxial 
surface smooth with some slightly impressed veins, 
abaxial surface with slightly raised secondary veins; 
base cuneate; apex obtuse, rounded, to acute. 
Inflorescence 4-8 cm; branches 1.5-6 cm; bracts 

vate, spatulate to triangular, sessile to short- 
petiolate, 3—27(—42) X 0.75-10 mm; base attenuate 
to cuneate; apex acuminate, acute, obtuse, to rounded; 
bract petioles 0—4 mm; primary and secondary flowers 
essile to d pedicels 0—3 mm; bracteoles 
triangular, 1 5 m. Calyx campanulate, 
24 X 2-3 mm, ou to glabrous, ecarinate; 
calyx lobes triangular, 0.75-1.5 X 2-3 mm, apex 
acuminate to acute; corolla 14-18 mm; tube 7-12 X 
-2 mm; lobes 5— 0.75-1 mm, apex rounded to 
obtuse; stamens included; filaments less than 0.5 mm; 
anthers 1-1.5 X 0.2-0.5 mm; pistil 8-9 mm; ovary 
1-2 X ca. 1.5 mm; style 5.5-7.5 X 0.3- 
0.5 mm; stigma bilobed, each lobe acute to acumi- 
nate, 0.75-1 X 0.5-0.75 mm. Capsules ellipsoid to 
obovoid, 18-23 X 5-6 mm (excluding style base), tan, 
8 to 9 seeds per locule (e.g., 16 to 18 seeds per fruit); 
seeds 10-12 x 2 


seed wings dark orange, reticulate. 


ovate, 


—2.5 mm, seed body dark orange, 


Morphology and similarities. Bonyunia pulchra 
has generally small leaves that are bunched together 
at the branch apices. Based on its seed morphology, it 
appears to be related to both B. magnifica (Brazil) and 
B. nobilis (Colombia). 


Distribution. and. habitat. 
15-40 m tall tree known from primary forest in the 


Bonyunia pulchra is a 


Amazon Basin-facing Andes in Amazonas and Huánuco 
provinces, Peru, at elevations of 500—800 m (Fig. 1). 


IUCN Red List category. Bonyunia pulchra is 
known from a few collections in unprotected areas. It 
is x ad a preliminary IUCN status of Vulnerable 

VU) according Es IUCN Red List Categories and 
Criteria (IUCN, 2001). 

Etymology. The epithet is taken from the Latin 

“pulcher,” meaning “beautiful. 

Notes. 
persons, John Ricketson, Jason Grant, 


Bonyunia pulchra is co-authored by three 
and Ron 
Liesner, who each independently determined this to 
be a new species. 


Paratypes. PERU. Amazonas: Bagua, Imaza, Región 
del Marañon, Com. Yamayaket, Quebrada Kusu-Chapi, R. 


Imaza, 


Vásquez et al. 19847 (MO); Distr. Comunidad 
Yamayakat, bosque 10, transecto 2 X 500m 


30 m, estéril [sterile], 29 May 1997, R. Vásquez, A. Peña, & 
E. Chávez 23812 (G, HUT, MO); Bagua, Imaza, Tayu Mujaji, 
comunidad de Wawas, oe m, R. Vásquez et al. 24694 (DLF 
v seen, F not seen, G, H not seen); Bagua, 
ma aka t, trocha a e 500 R. Vásquez 
Joni 20312 (MO, WAG). Huámeo: a region of 
ountains” & adjacent lowland, 
m, B. Wallnifer 14- 
41088 (K, W); Pachitea, Pucallpa, W par of the Sira Mtns. 
and adjacent lowland, from ca. 20-24 km SE of Puerto Inca, 
5 Morawetz & B. Wallnófer V30-15888 (W), W. Morawetz & 
puo d V79-13888 , W. Moraweiz & B. Wallnófer 
15- 10288 (W), B. Wallnöfer V62-121088 (K, W). 


zÍ 


8. Bonyunia spectabilis J. R. Grant, sp. nov. TYPE: 

Guyana. Cuyuni-Mazaruni: 2-5 km NW of tip of 

N prow of Roraima, 5%15'N, 60%35'W, 800- 

a m, mixed upland and cloud forest on talus 

s of Roraima, 22 Feb. 1989, W. Hahn & D. 

ae 5420 (holotype, Ut; isotypes, CAY not 
seen, F!, MO!, NY!, US!). Figures 1, 3, 4, 8B. 


Species nova Bonyunia anioniifolia Progel cui affinis, 
sed ab ea caulibus glabris, pedunculis hispidis, foliis 
ellipticis lanceolatis oblanceolatisve longi-petiolatis basi 
longi-attenuatis apice acumi 
appresso- an vel hispido differt; etiam montem Rorai- 
mam habita 


minatis atque calyce 


Branched tree to 10 m tall, glabrous throughout, 
except for hispid hairs on the petioles, peduncles, 
pedicels, corollas, and fruits. Leaves elliptic, lance- 
olate, to oblanceolate, long-petiolate, 7-11 cm, peti- 
ole 5-10 mm; blades 6.5-10 X 23- 4.4 em, thin, 
Md adaxially, lighter abaxially, adaxial surface 
smooth with some slightly i 
mus with slightly 
attenuate to cuneate; 


mpressed veins, abaxial 

raised secondary veins; base 
apex acuminate to acute. 
branches 1-2 cm; bracts 
upper 
tioles are pubescent; the true 


Inflorescence 2.5-4 cm; 
elliptic, lanceolate, to oblanceolate (as in the 
leaves, except the peti 
leaves have glabrous petioles), petiolate, 27-39(-71) 
X 6-10(-30) mm; base attenuate to cuneate; apex 
mm; primary 
and secondary flowers pedicellate; pedicels 3-7 mm 


acuminate to acute; bract petioles 3— 


bracteoles triangular, 1-2 X 0.5- 
3-4.5 
strigose to hispid, ecarinate; calyx lobes triangular, 
0.75-1 X 2.5-4 mm, apex acuminate t 

corolla poorly known (a single immature corolla on 
Hahn & Gopaul 5420 [US] is illustrated, but 


measurements not taken); stamens and pisti 


1 mm. Calyx cam- 


panulate, X 2.5-4 mm, densely appressed 


un- 
known. Capsules somewhat immature, but descrip- 
tion still prepared, narrowly ellipsoid, 11-13 X 3- 
4 mm (excluding style base), tan, 14 to 20 seeds per 
locule (e.g., 28 to 40 seeds per fruit). Only immature 
seeds seen. 


Volume 96, Number 4 Grant 559 
2009 Revision of Bonyunia 


missoum 
BOTANICAL GARDEN 
"HIERBA. 


Figure 8. Types of Bonyunia. —A. Holotype of B. pulchra Ricketson, J. R. Grant & Liesner (Rojas et al. 478 [MO]. 


o: ]. —B. 
Holotype of B. spectabilis J. R. Grant (Hahn & Gopaul 5420 [U]). —C. Lectotype of B. superba M. R. Schomb. ex Progel 


(Schomburgk 614 (939) [K]. —D. Holotype of B. venusta J. R. Grant (Ribeiro et al. 1103 [NY]). 


Annals of the 
Missouri Botanical Garden 


Morphology and similarities. | Bonyunia spectabilis 
has several unique characters, notably in that the 
inflorescence is pubescent (peduncles, petioles of 
bracts, calyces, and corolla), yet the leaves and stems 
are glabrous below. It is similar to B. antoniifolia and 
B. aquatica in having pubescent calyces, but is much 
more densely hispid than the other two. It is also the 
only species in the genus with long petioles (5— 
O mm) and narrowly elliptic leaves. Bonyunia 
spectabilis appears to be most similar to B. antonii- 
folia, B. aquatica, B. minor, and B. superba. Even if 
only immature seeds are known from B. spectabilis and 
therefore not illustrated or described in full, they 
resemble those of these four species. 


Distribution and habitat. 
occurs on talus slopes of mixed upland cloud forest 


Bonyunia spectabilis 


on white sand and boulders. It is only known from its 
type specimen collected on Mount Roraima along the 
border of Brazil, Guyana, and Venezuela (Fig. 1). It 
may be sympatric with B. minor and B. superba, 
especially B. superba, which has been most often 
collected on Mount Roraima. 


IUCN Red List category. Bonyunia spectabilis is 
only known from the type collection, from a formally 
protected area, the Parque Nacional Canaima (Vene- 
zuela). It is assigned a preliminary IUCN status of 
Critically Endangered (CR) according to IUCN Red 
List Categories and Criteria (IUCN, 2001) 


Etymology. The epithet is taken from the Latin 


“spectabilis,” meaning “showy 


Bon M. R. Schomb. ex Progel, Fl. 
ne ous EM 261, E 72. 1868. TYPE: 

Venezuela. Bolivar: Mt. Roraima, “Our Village, 
near Canaupang settlement (van o 2002)," 
Oct —4 Dec. 1842] Robert 
Schomburgk 614 [= Richard sch rad 


(lectotype, designated by L enberg, 
158, K!; isotypes, BM [2]!, BRI, F [2]! a Di 
GH!, K [2]!, P!, Ul, W!). Figures 1, 2C, 3, 4, 8C. 


Branched shrub to tree to 1.8—7.6 m tall, hispid 
throughout especially on the midvein and secondary 
veins of the undersides of leaves, petioles, peduncles, 
inflorescences, and corollas, glabrous only on abaxial 
leaf surfaces and notably glabrous on the calyx; trunk 

to 20 em in diam. (Pinkus 270). ue ovate, oval, 

to Rr Me foal HE (3.5—)6— petiole 3— 
5 mm; blades (3.2-)5.5-8.5 X pe em, thick 
coriaceous, darker adaxially, lighter abaxially, slightly 
glossy adaxially and more opaque abaxially, adaxial 
surface smooth with some slightly impressed veins, 
abusi l surface with prominently raised midvein and 


secondary veins; base rounded to cuneate; apex 


obtuse to acuminate. Inflorescence 2.5-6 cm; branch- 

—6 cm; bracts ovate, elliptic, lanceolate, to 
spatulate (as in the bracteoles above), sessile to 
short-petiolate, 9-37 X 3-18 mm; base rounded to 
cuneate; apex obtuse to acute; bract petioles 0-2 mm; 
primary flower sessile to subsessile, secondary flowers 
pedicellate; pedicels 1-10 mm; bracteoles distinctly 
spatulate to obovate, with a prominent midvein, 
a n much DE. the length of the s 
lobes, 6-10 X 


2-2.5 mm, nv ecarinate to slightly keeled "un 


x campanulate, 5— 


midvein; calyx lobes spatulate to obovate, as in the 
bracteoles, 2 m, apex rounded to obtuse; 
Lire 15-17 mm; tube 10-17 X 1.5-2 mm; lobes 5— 

X 0.75-1 m 


iie filaments less than 0.5 mm; anthers 1.8-2 
x0 


mm, apex rounded to obtuse; stamens 


—0.5 mm; pistil 6-7 mm; ovary ovate, 1.5-2 X 
ca. T mm; style 4—5 X 0.2-0.4 mm; stigma bilobed, 
each lobe spatulate, 0.75-1 X 0.3-0.5 mm. Capsules 
fusiform to ellipsoid, 15-33 X 5-8 mm (excluding 
style base), tan, 3 to 10 seeds per locule (e.g., 6 to 20 
seeds per fruit); seeds 11-17 X 2.5-3 mm, seed body 
brown, seed wings straw-gold, reticulate. 


Morphology and similarities. | Bonyunia superba is 


exceptional in the genus in having long pedicels (1— 
10 mm), spatulate to obovate bracteoles equaling to 
much exceeding the length of the calyx lobes, and 
spatulate to obovate calyx lobes. In its hispid 
pubescence nearly throughout it is similar to both B. 
antonüfolia and B. Bonyunia superba 


aquatica, B. minor, and perhaps B. spectabilis. 


Distribution and habitat. 
curs with B. spectabilis in the forest of Mount Roraima 
on the Ve 
the Guayana region (Fig. 1). The two species are 


Bonyunia superba oc- 


nezuela-Guyana border on the Pantepui of 


minor. 
actually been seldom collected in comparison to B. 
minor. 


IUCN Red List category. Bonyunia superba is 
known from few collections, most from a formally 
protected area, the Parque Nacional Canaima (Vene- 
zuela). It is assigned a preliminary IUCN status of 
Vulnerable (VU) according to IUCN Red List 
Categories and Criteria (IUCN, 2001). 


Etymology. The epithet is taken from the Latin 
“superbus,” meaning “excellent.” 


Typification. The herbarium labels of Schomburgk 
614 (939) have little text, but collection locality can 
be pieced together from the labels and the proto- 
logues. The two separate collection *numbers" refer to 
the two numbering systems of Richard and Robert 


Volume 96, Number 4 
2009 


Grant 
Revision of Bonyunia 


Schomburgk (van Dam, 2002). The first number, 614, 
is from Roberts second collection series, which 
corresponds to number 939 in Richard’s series (van 
Dam, 2002). There are samples of both at K 

Richard Schomburgk (1848) gives the locality of 


ag ün 
p ds of Our Village, at mde Hm edge). Progel 
(1868) translates this text to Latin as “In Guyanae 
anglicae montibus Roraima, ad margines silvarum 
propre Our Village in formatione arenacea" (In British 

uia Mount Roraima, at forest edges near Our 
Village on sandy soils). When iden and Richard 
Schomburgk traveled to Mount Roraima, they set up 
camp on the Kukenaam River, naming their site *Our 
Village." From there, they made their ascent to Mount 
Roraima and, in this region, collected many plants 
including B. superba. According to van Dam (2002), 
however, Our Village was near the settlement of 
Canapang and is actually situated in present-day 
Venezuela rather than Guyana. Richard collected 
around Our Village and Mount Roraima from 28 
October-4 December 1842. The original material on 
which Schomburgk based his description was depos- 
ited at Berlin but destroyed during World War II, 
which probably led Leeuwenberg to lectotypify B. 
superba on material at K. 


Specimens ned. GUY uyuni-Mazaruni: 
rene Moment: ine shrub to d i in height [1.8 m] found 
in scrubby forest on the slope of Krabu Mountain, 6.11.1966, 

Field No.: R.B. 160, Record No.: Forest r of British 
Guiana 7993 (NY). VENEZUELA. wd 

El Dorado-Santa Elena Rd. on rd Ws den 1200- 
1250 m, Gran Sabana, gallery forest m grassland, A. Gentry 
et al. 10511 (MO, NY, US); Mt. Roraima Distr., vic. of 
Arabupu, 4200 ft. [1280 m], A. S. Pinkus 270 (F, G, NY [2], 
S, US). 


10. Bonyunia venusta J. R. Grant, sp. nov. TYPE: 
Reserva Florestal Ducke, 
Manaus-—Itacoatiara, Km 26, 2°53'S, 59°58'W, 
[ca. 50-100 m], Igarapé d Tinga, Floresta de 
s INPA No. rs 11 Aug. 1993, J. E. 

L. S. Rib M. J. G. Hopkins, J. F. Ramos & S. 
S. Sousa 1103 E. m isotypes, IAN not 
seen, INPA not seen, K!, MO not seen, SP!, U). 
Figures 1, 3, 4, 8D 


Brazil. Amazonas: 


Species nova Bonyunia antoniifolia Progel cui affinis, sed 
ab ea a foliis et calycibus glabris, lobis c 
ear difusa arque LE interdum 


longioribus (1-10 vs. 1-5 mm) differ 


alycis acuminatis vel 


Branched tree to 15 m tall, hispidulous on petioles, 
stems, peduncles, inflorescences, calyces, and corol- 
las (leaves glabrous adaxially and abaxially); bark 
violet on the outside, chestnut inside, albumen whitish 
cream. Leaves oval, elliptic, to ovate, short-petiolate, 


4-7.5 cm, with 3 to 6 pairs of arching secondary 
veins, petiole 4-6 mm; blades 3.5-7 X 2-3.3 em, 
thin-coriaceous, darker adaxially, lighter abaxially, 
slightly glossy adaxially and more opaque abaxially, 
adaxial surface smooth with some slightly impressed 
veins, abaxial surface with slightly igus secondary 
veins; base cuneate; apex obtuse to acute. Inflores- 
cence 7-12 cm; branches 2.5-7 cm; enis ovate, 
elliptic, to lanceolate, petiolate, 18—27 X 3-10 mm; 
base attenuate to cuneate; apex obtuse, acute, to 
rounded; bract petioles 3-6 mm; primary and sec- 
ondary flowers pedicellate; ies 1-10 mm; brac- 
0.3-0.75 mm. Cal 


teoles linear-triangular, 1. 
5 xX ee mm, glabrous to 


campanulate 
rarely hispidulous, ecarinate; calyx lobes triangular, 
0.3-1 


corolla, stamens, pistil, capsules, and seeds unknown. 


1.5-1.75 mm, apex acuminate to acute; 


Morphology and similarities. Bonyunia venusta 
appears to be most similar to B. excelsa and perhaps 
also to B. antoniifolia. It differs from B. excelsa in 
having stems and branches of the inflorescence equally 
hispidulous and leaves with three to six pairs of arching 
secondary veins. It differs from B. antoniifolia in bein 

much less pubescent throughout, with glabrous leaves 
and ane a more diffuse inflorescence, and longer 


pedicels (1-10 mm vs. 1-5 mm) 


Distribution and habitat. 
only known from its type material collected on terra 


Bonyunia venusta is 


firme in closed canopy forest in the Amazon lowlands of 
Brazil in the Reserva Florestal Ducke just north of 
Manaus, at elevations of 50-100 m (Fig. 1). Of the four 
types of terra firme forest in the park, it was collected 
on "Floresta de Vertente," which is a sloped transition 
zone between a higher and lower type forest. 
described and illustrated in the Flora da Reserva Ducke 
as B. aquatica (Ribeiro et al., 1999: 564) 


IUCN Red List category. | Bonyunia venusta is only 
known from the type collection, from a formally 
protected area, the Reserva Florestal Ducke (Brazil). 
It is assigned a preliminary IUCN status of Critically 
Endangered (CR) according to IUCN Red List 
Categories and Criteria (IUCN, 2001). 


Etymology. The epithet is taken from the Latin 
“venustus,” meaning “attractive” or “graceful.” 


Literature Cited 


Angiosperm Phylogeny Group. 2003. An update of the 
Angiosperm Phylogeny Group classification for the orders 
and families of flowering plants: APG II. Bot. J. Linn. Soc. 
141: ~ 

Backlund, M., B. Oxelman & B. Bremer. 2000. Le ue 
relationships within the Gentianales based on ndhF 
rbcL sequences, with particular reference to the e 
ceae. Amer. J. Bot. 87: 1029-1043. 


Annals of the 
Missouri Botanical Garden 


Berry, P. E. 2001. Bonyunia qur Pp. 24-26 in P. 
E. Berry, K. Yatskievych & B. K. Holst Mus dodi of 
the Venezuelan Guayana, Vol. 6: Liliaceae-Myrsinaceae. 
Missouri a Garden Press, St. Louis. 

Bonyun, G. R . On the Demerara Pink-root, or Spigelia 

nthe: ger ps Mag. Nat. Hist. 14: 461—463. 
Colonial Office and Predecessors: British 
Guiana, S erly Berbic 


. Struwe. 1992. i of the Rubiaceae 
and the Loganiaceae: uence or conflict. between 
morphological and ra data? Amer. J. Bot 
1171-1184. 

Brown, N. E. 1901. NE on two botanical collections ae 
y Mssrs. F. V. McConnell and J. J. Quelch of Mount 
Roraima in British Cuiana. Trans. Linn. Soc. London, Bor. 

07. 


l- 

Ducke, A. 1935. Plantes nouvelles ou s de la 
région amazoniene (VII** série). Arq. Inst. Biol. Mer l; 
205-212. 

Gleason, H. A. 1931. Botanical results of br Ex -Duida 
Expedition. Bull. Torrey Bot. Club 58(7): 4 


peu connue: 


otes on American Logania- 
ceae III: Revisioñ of Bonyunia Rich. Schomb. Acta Bot. 
Neerl. 2. 152-158. 

Maguire, R. E. Weaver Jr. 1975. The neotropical genus 
Tachia (oon J: (ue Axbor. 56(1): 103-125. 

Mori, S. A. & J. L. Brown. 1994. Report on wind dispersal in 

a lowland moist forest in central French Guiana. Brittonia 


46: 105-125. 


, G. Cremers, C. A. Gracie, J. J. de Granville, S. V. 
Heald, M. Hoff & J. D. Mitchell. 2002. Guide to us 
ue cular Plants of US French Guiana. Mem. New 
York Bot. Gard. 76 
Peters, W., M. O. da A. M. Pohlit & L. Struwe. 2004. 
e Field Guide. nl University-Cook College, 
w Brunswick, New Jers 
8. Loganiaceae. d C. F. P von Martius, Flora 
: 249-290. 
ir J. G. Hopkins, A. Vicentini, CaA 


. Lohmann, P. A. C. L. 
a da C. Pereira, C. F. da Silva, M. R. Mesquita 
& L. C. Procópio. 1999. Flora da Reserva Ducke: Guia de 
Identificagdo das Plantas Vasculares de uma Floresta de 


n den Jahren 1840-1844, Vols. 1-3. J. J. Weber, 
a eipzi 

Struwe, L & V. A. Albert. 1997. Floristics, cladistics, and 
classification: Three case studies in Gentianales 851 
352 in J. Dransfield, M. J. E. Coode & D. A. Simpson 
(editors), Plant Diversity in Malesia III. Royal Botanic 
Gardens, Kew, Richmond. 


— " A 
Poialia Aublet (Gontianieons: 
29(3): 670-701. 

T. J. Motley. E dx (including d 
Geniostomaceae, Spigeliaceae, and Strychnaceae). f 

W. Kadereit (editor), Families and Genera of ela 
Plants: Asteridaeá Gentianales. Springer Verlag, Berlin. In 

ess. 


monograph of neotropical 
Potalieae). Syst. Bot. 


———, ——— & B. Bremer. 1994. Cladistics and famil 
level classification of the Gentianales. Cladistics 10: 
175-2 


; a hiv, J. W. Kadereit, A. S.-R. Pepper, T. J. 
Motley, P. J. White, J. H. E. Rova, K. Potgieter & V. A. 
Albert. 1998. ias roo an endemic of 
Sierra de la Neblina Brazilian-Venezuelan border, 
related to a o line un of Gentianaceae. 
Harvard Pap. Bot. 3(2): 1 
— ———, M. P. Kinkade & P T E Maas. 2005. T 
Brazilian species of fant (Gentianaceae: 
Blumea 50: 457-462. 


new 


Helieac) 


, 5. Haa g E. M erg & J. R. Grant. 2009. Andean 
speciation iis ariance in Neotropical Macrocarpaea 
(Gentian aceae-H. Ji ae). Ann Missouri Bot. Gard 


1906. Vitae itineraque collectorum botanicorum, 
notae collaboratum biographicae, flora brasiliensis ratio 
edendi chonologica, systema, index familiarum. /n C. F. P 
Martius, Flora pr Vol. 1, Pt. 1: 1-268. 

van Dam, J. A. M. J. Jansen- jsobs (editors). 2002. 
Flora of the Guianas: Supplementary Series: Fasc. 3. The 
Guyanan Plant Collections of Robert and Richard 
Schomburgk. Kew Publishing, Kew, Richmond. 


eee 1: Numbered Collections of Bonyunia examined. 


T T 
The n mbers in brackets indicat , identifi ng 
List of Species. Collections are arranged by es Du 
. Type collections are indicated by asterisks. 


name. 


List oF SPECIES 


E 


. B. antoniifolia Progel 
aquatica Ducke 
excelsa J. R. Grant 
magnifica n: R. Grant 
minor N. E. B 
nobilis J. R. 
. pulchra m J. R. Grant & Liesner 
spectabilis J. R. Grant 
. superba M. R. Schomb. ex Progel 
10. B. venusta J. R. Grant 


Agostini, G. . M Amaral, B. L. 44 (RB [1]. 
Amaral, I. L. 1 


SODA AN 
as 


Di e [5]. 1946, n US, WAG [ [5]. 2245 e [5). 
Cavalcanti, G. P. [1]. Chacon, L. 595 (MO [5], 624 
(MO a 679 Mo rd eee H. D. 913 (NY, US [5]. 1238 
(MO, NY, US [5]. Colonnello-Aznar, G. 845 (MO [5]. 
Dambrós, L. A. 356 (US [1]. Davidse, G. 16912 (MO [2]. 


ie 12197 (BM, G, P, US [ E 12384 (BM, G 
0* (G, K, P, RB, S, U, U; 
5]. Ferreira, C. A. C. 5603 (MO 


Dept. of British Guiana 7921 (K, NY [5], 7995 (NY [9]. 
Fróes, R. L. 22260 (IAN, HT 2), o (IAN, U [2). 


Volume 96, Number 4 


Grant 
Revision of Bonyunia 


Gentry, A. 10511 (MO, NY, US [9]. Guánchez, F. 881 (MO 

KD ien D m 3633 (WAG [a Dod R. 2315 (MA, 

Z [1]. 2513 (G, MA, MO, N Z [1]. 4161 (F, MA, 

Mi Tu ED- RA E. 1323 — lo. NY, USZ [1]. 1444 
a mee [1). 

W. & D. Gopaul 5420* (CAY, F, MO, NY, U, US [8]. 

Hon T. E CN (US E We 15 (GH [5]. Huber, O. 

K, NY, US, en "pud (NY qw 3470 (B, 

WAG [5 


76 
AAU, NY, US[ en 11748 (US a Es 


, A. 640 (M). Janssen & Gemichujnicov 306 (M). 
Killeen, T. 2749 (NY Tip. 6501 (F, ane oe USZ [1], 
7080 (MO, NY, USZ [1]. Koyama, T. 7521 (NY [5]). Kral, R. 
72137 (NY id Pa K. 79- plu nae i 79-198 (M, 
NY, US [2]. K n, J. G. 2233 (SP 
won T. no (NY, "US I BD, p? oY [5). Liesner, R. L. 
19300 (MO n 19463 (NY n 20485 (MO, p a 23930 
(MO [5], 24014 (MO [5], 24016 (MO [5], 24028 (MO, NY 
D a A [5]. 24438 a [5]. Heras, ES P19452 (U, 
G [1]. Lueizelburg, Ph. 21503 (M n 21514 (M [5]. 
CN Pu 31012 


9163 
WAG [2], 12382 


a 


33994 pu [b]. 40281 (IAN, e RB[ oe 
RB [B], 40482 (NY [5], 461434 (F, NY, US, WAG [5) 
46149A WAG [5]. Marin, E. 1126 (MO [2], 1188 


15691 (FMB [2]). Moore, H. E. 961 
15-10288 (W [1], V30-13888 (W ED p 13888 (W 
Mostacedo, B. 1815 (G, NY, USZ [1]), 1861 (G, MO, NY, USZ 
1) 


Nelson, B. a mo WAG [1 

Palacios, P. 2393* L, MA, MO [ [6]. Paula, J. E. de 
1907 (Z [1]. Pena, B. n D RB [1]. Peña, M. 170 (NY 
[1]). Picón Nava, G. 119 S [5]. Pinkus, A. S. 270 (F, G, 
NY, S, US [9]. a E ^ dura (IAN, RB [2]). Prance, G. 
T. 5758 (COL, F, GH, K, MG, NY, R, S, U, US, Z [1]. 13796 
K, MG, Te e) U, US, WAG [1], 22804* (INPA, K, 

MG, MO, NY, U, US, WAG [4]. 28457 (MO, NY, WAG [5], 
28501 (MO, AE US, NE [5]. 30033 (MO, NY, WAG [1]. 
Pruski, J. 1405 (MO, N 

Quelch, P a 161* (K 
MO, NY, USZ [1 

Restrepo, 387 (MO rat eg J. E. L. S. 1103* (IAN, 


— 


i 331 (K [5]). Quevedo, R. 973 (C. 


INPA, K, MO, NY, SP, uc n 49* (BR, LE, 
MO, R, NY [1]. Rade p 8547 (US [1]. Rojas, R. 478* 
(F, G, HUT, MO, NY, US, USM 


Sastre, C. 8502 d [5). Sohombuspk: R. 614* (BM, BR, 

, GH, K, P, U, W [9], 939 (F [9]), 966 (BM, P [5]. 

coe RE. p = US [2)). ate a N. 28 (MO, e 

[1]. Steyermark, J. A. 68 (NY [5]. 58291 (F, 

En xb (NY [5]. 104146 (NY. US [ T 105489 (NY ed 

115510 (MO [5], 117557 (MO [5]. 117819 (F, MO [5]. 
127575 (MO, NY [5]. 131853 (MO [5]. 

Tamayo, F. 3132 (US [5]). Tate, G. H. H. 770 (F, NY [5]. 

A 


Teixeira, L. 266 (NY, WAG [1]. 1268 (MG, MO, NY, 
RB, US, W. 

Ule, E. 8469 K MG [5 

Vargas, L. 830 (WAG [1]). Vásquez, R. ee (MO [7]. 


20312 (MO, WAG [7], 23812 (G, HUT, MO [7], 24694 
` G, HUT, MO, USM [7]). Vieira, €. po (K, NY, US, 
WAG [I]. 
Wallnófer, B. 14-41088* (K, W [ y V62-121088 (K, W 
[7]. Wurdack, J. J. 42814 (COL, F, NY, S, US, WAG [2), 
43261 (NY, WAG [2]) 


PHYLOGENETIC POSITION AND 


Ahmad Reza Khosravi,?* Fernand Jacquemoud,” 


TAXONOMIC CLASSIFICATION OF a dn di : d Menke,* 


AETHIONEMA TRINERVIUM 
(BRASSICACEAE): A 
MORPHOLOGICALLY VARIABLE 
SUBSHRUB FROM 
SOUTHWESTERN ASIA! 


aus Mummenho 


ABSTRACT 


Due in part to its distinctive and variable morphology, traditional taxonomy has not resolved the systematic position 


of Aethionema trinervium Boiss., which has be 


order to clarify its piy geriet relationships. 
detho nema, but is instead hi ighly 


Aik | 


en previously p 


nested within Vania F. K. Mey 


laced in several different genera including Aethionema R. 
equence data fi e ITS-1 and ITS-2 ns of ribosomal 
d taxa in 


Our molecular em indicates that A. trinervium is not a member of 


segregate lineage of the oi Pic s.l. 
we transfer A. trinervium to Vania 


ES new combination V. trinervia (DC. ) Khosravi, Jacquemoud, Menke, Mumm. & Mohsenz. Furthermore, we have Eo 


the eee 


Aethionema, Brassicaceae, ITS sequence analysis, 


Vania. 


Thlaspi, 


Noccaea, phylogenetic relationships, Southwest Asia, 


Returning to previously analyzed molecular data 
sets with additional data is a valid and efficient 
method for increasing our overall phylogenetic 
knowledge and for giving additional emphasis to more 
narrowly focused questions, such as clarifying the 
taxonomic placement of a single problematic species 
(Mummenhoff et al., 1997a, b, 2001; Crawford et al., 
2001; Warwick et al., 2006a). 

In the present study, we demonstrate the utility of 
sequences from the ITS region of nuclear ribosomal 
DNA to resolve the generic affinities of a problematic 
species in the Brassicaceae, i.e., Aethionema triner- 
vium Boiss., a perennial subshrub. It is distributed in 
Afghanistan, Turkmenistan, Iran, Transcaucasia, Tur- 
Iraq (Hedge, 1965, 1968). The 
plants are essentially glabrous and woody at the base, 


key, and northern 


with the ascending stems remaining herbaceous 


— 


Hedge, 1965). Leaves are strongly palmately veined, 
amplexicaul, mostly 10—20 mm in length, and vari- 
ously shaped from lanceolate to ovate to oblong with 
the basal auricles varying from almost absent to 
sagittately elongate. Leaf shape may vary dramatically 
on a single individual between the proximal and distal 
The 
racemose, starting as a condensed head but elongating 
in fruit (Hedge, 1965, 1968). Sepals are erect, petals 


are white and 1-nerved at the base, filaments are free, 


leaves of the same shoot. inflorescence is 


and anthers are apiculate (Hedge, 1965). Silicles are 
angustiseptate and 4—10 mm in length, and they vary 
rom lanceolate to oblong. Furthermore, fruit valve 
y be prese t. If prese 
vary in shape from broadly obcordate or squarely 


wings ma nt or absen nt, the wings 


! We thank Shahin Zarre for critically reading the draft manuscript and Seid Mansoore Mirtadzadini for kindly providing 
leaf material from his collection. Gérard Aymonin (P) kindly D information about the collections of Michaux and Olivier 


and Bruguière and two digital images of Hutchinsia trine: 
reproduction of the BE of Hutchinsia trinervia DC. fro 


Patrick Perret, curator of the library at G, permitted the 
m Icones selectae plantarum, and Patricia Riedy (G) scanned the 


image. We also thank Victoria C. Hollowell for editorial advice and Beth Parada and Allison Brock for assistance preparing 


the manuscript for publication. 


? Department of Biology, ord of Science, Shiraz University, Shiraz 71454, I 


Ville de Geneve, Case postale 60, CH-1292 ‘Chambésy, Switzerl 
t. Louis, Missouri 63130, U 


an 
.S.A.; Missouri F Garden, P.O. Box 


"Department of Biology, Systematic Botany, da ^ Osnabrück, Barbarastrasse 11, 49076 Osnabrück, Germany. 


Author for correspondence: khosravi€biology.susc.ac.i 
doi: 10.3417/2007004. 


ANN. Missouni Bor. Garp. 96: 564—574. PUBLISHED ON 30 DECEMBER 2009. 


Volume 96, Number 4 
2009 


Khosravi et al. 
Classification for Aethionema trinervium 


oblong or reduced to a narrow margin (Hedge, 1965, 
1968). The seed surface is smooth (Hedge, 1965). 
Individuals collected at higher elevations (above 
3000 m) tend to have markedly smaller leaves and 
display a subpulvinate habit (Hedge, 1965, 1968). 
Morphologically, the evidence for placing the 
species in Aethionema R. Br. is equivocal, with many 
vegetative features of A. trinervium being atypical for 
the genus. However, there are examples in which 
some unusual morphological states are approached in 
at least one other species of Aethionema. For example, 
the prominent palmate leaf venation in the annual A. 
arabicum (L.) Andrz. ex DC. resembles that of A. 
trinervium. Also, while no other Aethionema species 
possesses auriculate or amplexicaul leaves, the 
cordate and short petiolate leaves of A. cordatum 
Boiss. approach such a condition. Fructal morphology 
is quite variable within Aethionema, as it is throughout 
much of Brassicaceae, but all core taxa within 
Aethionema possess fruit with membranously winged 
valves. As stated before, the presence of wings is a 
variable trait in A. trinervium, which has been a 
principal reason for classification problems. 
Taxonomists who did not aecept that Aethionema 
trinervium belongs to the genus Aethionema a the 


s Hut y andel 
1821), it was also later placed in the: genera a Iberidella 
(Boissier, 1841), Aethionema DC. (Boissier, 
1867), Eunomia DC. (Prantl, 1891), and Thlaspi L. 
(Mozaffarian, 1996). Él (DC.) Prantl 
was recombined on the basis of the presence of 
Schulz (1933) also later 


recognized E. trinervia in his own familial classifica- 


tchinsia trinervia D 
Boiss. 
unomia trinervia 
median nectary glands. 
tion. This conclusion is not easy to follow because in 


salmasium Boiss., median 
the lateral 


A. trinervium, as in A. 
nectary glands are absent. However, 
nectaries are somewhat connected, 
giving the impression of median g 
some morphological and cytological characters, Mo- 
zaffarian (1996) included A. trinervium in Thlaspi as 
T. trinervium (DC.) Mozaff., and this species was also 
treated as T. trinervium in the recent species checklist 
of the Brassicaceae (Warwick et al., 2006b). Appar- 
ently, Mozaffarian did e n. the main 
phylogenetic dom wit s.l. suggested 
by Meyer (1973, 1 EN d are based on the 
analysis of seed coat anatomy and later ids 
supported by molecular studies (Mummenhoff $ 
Koch. 4; Mummenhoff et al., a, b; Koch et 
al, 2001). Thus, it is unclear to which Thlaspi s.l. 
segregate A. trinervium should be assigned. Recent 
broadly sampled molecular studies of Brassicaceae 
(Khosravi, 2001; Koch et al., 2001; Beilstein et al., 
2006) also proved that Thlaspi s.l. is a polyphyletic 


taxon and that the members of Thlaspi s. str. are 
distantly related to the other Thlaspi s.l. lineages. In 


ct 


e most recent systematic treatment of the family, Al- 
Shehbaz et al. (2006) placed members of Thlaspi s. 
str., along with other genera (e.g., Alliaria Heist. ex 
Fabr., Pachyphragma Rehb., Peltaria Jacq.), in tribe 
Thlaspideae and remaining Thlaspi s.l. segregates in 
the tribe Noccaeeae. In summing up its taxonomic 
history, it is clear that A. trinervium has long been a 
poorly understood species for which previous taxo- 
nomie classification attempts have proven both vague 
and unstable. 
Furthermore, the morphology within Aethionema 
trinervium is exceptionally variable, and some of this 
variation appears to show geographic structure (Davis 
et al, 1965). Herbarium specimens may suggest a 
gradual reduction in fruit valve wing size and an 
increase in leaf auricle size as the species ranges from 
west to east. In addition, four separately described 
taxa are often either reduced to synonymy with A. 
trinervium or treated as a variety thereof. Two of the 


a ovalifolia Boiss. and I. 


ous names, /beridell. 
(later A sagittatum (B 


ynonym 

sagittata Boiss. oiss.) Boiss.), 

were described by Boissier in 1842. They apparently 

differ from each other in the relative prominence of 

the leaf auricles. Such variation has been treated 
e scope of A. trine in s 


Furthermore, Boissier (1867) € Aethionema 
salmasium from A. trinervium on a basis of fruit 
differ However, A. not 
pa b Hedge (1965, "1068, pisi p^ to 
overlapping patterns of morphological variation. 
However, other national floras do recognize both 
cles as n (Karjagin, 1953; Avetisian, 1966; 
Khintibid ze, 1979). 
Perhaps most different from typical Aethionema 
trinervium is A. apterocarpum Rech. f. & Aellen, 
which is described from m 


provinces of North Khorasan, South Khorasan, and 


aterial i Iranian 
Razavi Khorasan, where the taxon is endemic (Hedge, 
1968). It was up as A. trinervium var. 
f. & Aellen) Hedge by 


( .Itis i by wingless fruit valves that 


apterocarpum Hedge 
are further distinguished by the fruit shape itself being 
shorter and more ovate than in typical A. trinervium. 
edunel 


Also, the inflorescence p e usually elongates 
A 


earlier in flora a than typical A. triner- 
vium, with each flowers pedicel separated by 
approximately a centimeter of peduncle length at the 

e of anthesis. In contrast, typical A. trinervium 


inflorescences remain as compacted racemes at 


Annals of the 
Missouri Botanical Garden 


anthesis but elongate later during fruit development. 
The biogeographic and genetic relationships between 
the morphological variants of A. trinervium remain 
unclear. The problem is introduced here, but clear 
answers await completion of a more comprehensive 
population-level genetic analysis of what may poten- 
Thus, the goal of this 
present study is to use sequence variation of the 
nuclear ITS of ribosomal DNA (ITS-1 and ITS-2) of A. 


trinervium and representatives of putative related 


tially be a multispecies complex. 


genera, especially Aethionema and Thlaspi s.l. 
segregates, advocated generic rank by Meyer (1973, 
1979, 1991), to clarify the phylogenetic position of A. 
trinervium. 


MATERIALS AND METHODS 


SAMPLING 


In the current study, we have included repre- 
sentatives of Aethionema and Thlaspi sl. along 
with other taxa previously shown to be closely 
related to Thlaspi s.l. (Zunk et al., 1996; ls hoff 

al., Ta, b; Koch & Miralo 001). Because 
Aethionema is sister to all remaining Brassicaceae 
(Zunk et al., 1996, 1999; e. oway et al, 1998; 
Koch et al., 2001, 2003) and A. trinervium might be 
nested in edi, cn i. Hook., a member 
of the Cleomaceae (sensu Hall et al., 2002), the sister 
family of the Brassicaceae, was selected as the 
outgroup. The final data set comprised 39 taxa. 
ITS sequences were generated for A. trinervium. 
Other sequences were obtained from GenBank 


(Table 1 


DNA EXTRACTION, AMPLIFICATION, AND SEQUENCING 


Total DNA was isolated from dried leaf material of 
Aethionema trinervium (IRAN. Kerman Province: 
Ravar, Sood Kuh, 5 Sep. 2004, Mirtadzadini 24985, 

ucher in Herbarium of Shiraz University) using a 
modifi ed CTAB method (Aras et al., 2003). Double- 
stranded DNA of the complete ITS region, including 
the 5.88 ribosomal DNA (rDNA) gene, was amplified 
by 35 cycles of symmetric polymerase chain reaction 
(PCR) es ITS primers initially designed by White 
and modified by eed et al. 
t with 5 min. 

94^C, and 35 cycles of ere (1 min. 94°C, 
j min. 50°C, 1 min. 72°C), iu final elongation for 
10 min. at 72°C. PCR produc e purified using 
the High Pure PCR Produet pa Kit (Roche 
Diagnostics—Applied Science, Mannheim, Germany). 
377XL 
automated sequencer (Applied Biosystems, Weiter- 
stadt, Germany) 


Sequencing reactions were run on an ABI 


DATA ANALYSIS 


DNA sequences were aligned with those obtained 
from ss using Clustal W version 1.8 (Thompson 
et al, 1994). Multiple alignment el were 
set to 12 for gap opening penalty and 0.1 for gap 
penalty. Alignments — confirmed 


extensions were 


Maximum parsimony 
analyses of the aligned sequences of the ITS data 
set were conducted using the computer program 
PAUP*, version 4.0b10 (Swofford, 2000). The most 
parsimonious trees were generated using the heuristic 
with 
(TBR) branch swapping, equal-weighted characters, 
1000 random additions of the sampled taxa, and 100 


trees saved per replicate. 


search method, tree bisection-reconnection 


Sets of equally most 
parsimonious trees were summarized by a strict 
consensus tree. SE tests (Felsenstein, 1985 

performed using 500 replicates with heuristic 
p settings identical to those of the original 
search. Pairwise distance sequence divergence of ITS 
was calculated for each accession pair in PAUP* 
using the Kimura two-parameter model (Kimura, 


RESULTS 


The aligned data matrix contained 403 positions 
after removing regions with alignment ambiguities. Of 
these, 188 (46.6%) were potentially phylogenetically 
informative, 126 (31.3%) were invariant, and 89 sites 

5) were pc id phi e heuristic search 
rola in t parsimonious trees o steps 
with a pire Un of 0.5615. The results of the 
phylogenetic analysis are shown in Figure l. Two 
on the 
consensus tree (Fig. 1). The first lineage (referred to 


major clades can be recognized strict 
here as Thlaspi clade) comprises all the species 
traditionally subsumed under Thlaspi s.l. along with 
other genera recently shown to = e related to 
J. (Mummenhoff & K 
997a, b; Koch n 
Koch & Al-Shehbaz, 2004). Most representatives of 
this lineage belong to tribes Noccaeeae and Thlaspi- 
eae. Also, denupnemo trinervium clearly belongs 


here, and it is found in a strongly supported clade 


— 


100% en support) along with Vania campylo- 

phylla F. K. Mey. and V. kurdica (Hedge) F. K. Mey. 
in the tribe Noccaeeae. The second lineage represents 
the genus Aethionema and is referred to here as the 
Aethionema clade (tribe Aethionemeae). The genetic 
these 


divergence between two main clades is 


Volume 96, Number 4 
2009 


Khosravi et al. 
Classification for Aethionema trinervium 


Table 1. 


List of taxa used for the current study. 


Taxon 


pee arabicum (L.) Andrz. ex DC. 

A. elongatum Boiss 

A. eae Boiss: & Hohen. 

A. saxatile R Br. 

A. trinerviu 

Alliaria petiolata “OH. Bieb.} Cavara & Grande 
me ook. 


i Boi 
osperma (Maire) R. Vogt 
. sempervivum Boiss. & Balansa 
C. sintenisii ur ex pes 
Peltaria turkmena 
Teesdalia as ni R. Br. 
Thlaspi alliaceum L 
alpinum Crantz 
arvense L. 
bulbosum Boiss. 
caerulescens J. Presl & C. Presl 
calaminare Lej. & C 
cepaeifolium (Wallen) Es M J. Koch 
ceratocarpum N. Busch 
cilicicum (Schott & pun Hayek 
crassiusculum (F. K um ~ & Burdet 
densiflorum Boiss. & pus 
elegans Boiss 
goesingense Halácsy 
granatense Boiss. & Reut. 
hastulatum Steven ex DC. 
jankae ern. 
nmm Boiss. & A. Huet 
macranthum N. Busc 


= 


oxyceras 3 Hedge 

T. perfoliatu 

T. szowitsianum EN 

T. umbellatum Steven ex DC. 
Vania campylophylla F. K. Mey. 
V. kurdica (Hedge) F. K. Mey. 


Source GenBank accession number 
Hong et al 03 AY254539 
Koch et al., unpublished DQ518386 
Bailey et al., 2006 DQ452067 
Mummenhoff et al., 2005 AJ86269 
Present study 180 
ace al., 20 AJ862703/AJ862704 
O'Kane & Al- Sere 2003 
och & Mummenhoff, 2 AF336202/AF336203 

Koch & Mummenhoff, 2001 AF336208/AF336209 
Peer et al., 2003 AY261529 
Koch & Mummenhoff, 2001 AF336204/AF336205 
Koch & Mummenhoff, 2001 AF336212/AF336213 
Koch & Mummenhoff, 2001 AF336214/AF336215 
Koch & a 2001 oe 
Koch & Al-Shehbaz, 2004 Y154 
Koch & e 2001 pres a a F336153 
Koch & Mummenhoff, 2001 AF336200/AF336201 
Koch & Mummenhoff, 2001 AF336188/AF336189 
Koch & Racer 2001 AF336192/AF336193 
Koch & Al-Shehbaz, 2004 AF336198/AF336199 
Koch & d 2001 AF336154/AF336155 
Koch & Mummenhoff, 2001 AF336166/AF336167 
Koch & Mummenhoff, 2001 AF336206/AF336207 
Koch & Al-Shehbaz, 2004 
Koch & Mummenhoff, 2001 AF336160/AF336161 
Peer et al., 2003 AY261528 
Koch & Mummenhoff, 2001 AF336176/AF336177 

ch & Mummenhoff, 2001 AF336164/AF336165 
Koch & Al-Shehbaz, 2004 

ch & Mummenhoff, 2001 AF336162/AF336163 
Koch & Mummenhoff, 2001 AF336194 /AF336195 
Koch & Mummenhoff, 2001 AF336196/AF336197 
Koch & Mummenhoff, 2001 AF336184/AF336185 
Koch & Mummenhoff, 2001 AF336172/AF3361 73 
Koch & Mummenhoff, 2001 AF336158/AF336159 
Koch & Mummenhoff, 2001 AF336180/AF336181 
Koch & Mummenhoff, 2001 AF336174/AF336175 
Koch & Mummenhoff, 2001 AF336186/AF336187 
Koch & Mummenhoff, 2001 AF336168/AF336169 
Koch & Mummenhoff, 2001 AF336170/AF336171 


considerably high. ITS sequence divergence between 

trinervium and other Aethionema species ranges 
between 34.5% (A. trinervium vs. A. elongatum Boiss.) 
and 39.5% (A. trinervium vs. A. grandiflorum Boiss. & 
Hohen.). On the other hand, the sequence divergence 
between A. trinervium and Thlaspi s.l. species is 
significantly lower and ranges between 2% (A. 


V. campylophylla) and 21.5% (A. 
trinervium vs. T. hastulatum (Steven ex DC.) Hedge). 


trinervium vs. 


This is additional evidence that A. trinervium is more 


closely related to Thlaspi s.l. taxa than to Aethionema. 


DISCUSSION 


In the current study, we used evidence from nuclear 
ITS sequences to clarify the generic status and 
phylogenetic relationships of Aethionema trinervium, 
variously assigned to different genera in 
different traditional M ara (e.g., Hutchinsia R. 

: l eridella, Boissier, 1841; 
Aethionema, Boissier, oe m Prantl, 1891; 
Thlaspi s.l., Mozaffarian, 1996). These genera are all 


taxonomically critical, but for a better understanding 


species 


Annals of the 
Missouri Botanical Garden 


Cochlearia aucheri 


ochlearia sintenisii || M ASMENIA/PSEUDO: 


Cochlearia sempervivum| | SEMPERVIVUM 


Thlaspi caerulescens 
Thlaspi calami 

Thlaspi alpinui 

Thlaspi rcm 
NOCCAEA 


Noccaeeae 


H RAPARIA 


100 | 


evum Thlaspi 
clade 


| 


VANIA 


100 


consensus tree of 126 m 


(Koch et al., 


of the d discussion, these taxa need some 
closer inspec 
Meyer (1973, 1979) divided Thlaspi s.l. into 12 
segregate genera based primarily on differences in 
seed sculpture and seed-coat anatomy in contrast to 
morphological fruit characters prone to homoplasy and 
used in traditional treatments be Meyer, 1991; 
Mummenhoff et al., 1997b; 
1; Meyer et al., 
segregates varied from complete rejection (Greuter et 
al., 1986) to partial acceptance (Al-Shehbaz, us to 
See acceptance (Czerepanov, 1995). Thlaspi s.l. 
has been subjected to extensive molecular studies to 
test the validity of Meyer's (1973, 1979, 1991, 2001a, 
b) t e gates. Recently, Koch and Mummenhoff 
(2001) simimadeed a decade of phylogenetic studies 
on Thlaspi and pointed out that, with the exception of 
certain segregates (e.g., Microthlaspi F. K. Mey.), most 
Meyer's segregates represent monophyletic groups 
supported by molecular data. Furthermore, species 


Sui st parsimonious Fitch trees (tree length — 718 s 
dus 50% are prie above branches. Generic grouping by Meyer (1973, 1979, 1991) is indicated 
Alliaria petiolata. (M. Bieb.) Cavara & Grande and Peltaria turkmena Lipsk 
2001; Mummenhoff et al., 2001). Tribal grouping by Al-Shehbaz et al. (2006) i is catia p white 


MICROTHLASPI 


I THLASPI s.str. 


las 
Thlaspi ceratocarpum 
Peltaria turkmen 


[]rüurasPI s 
[E] KoTscHYELLA 


Thlaspideae 


Thlaspi hastulatum NOCCIDIUM 


Aethionema elongatum a 
Aethionema 
AETHIONEMA 
lade 


Cleome lutea OUTGROUP 


Aethionemeae 


teps). Bootstrap values more 
by grey vertical bars. 


were not reco by Meyer as um s. str. 


B placed in Cochlearia L. sect. Pseudosem- 
rvivum Boiss sempervivum Boiss. & Ba- 
lanza. C. aucheri Boiss., C. sintenisii Hauska. ex 
Bornm.) are nested within Thlaspi s.l. and are better 
treated either as Masmenia F. K or Pseudo- 
sempervivum (Boiss.) Grossh. (Fig. 1). Finally, Thlaspi 
s. str. is more closely related to Peltaria, Alliaria, and 
Thlaspi s.l. a Noccidium ey. 
Kotschyella . Me n to any other of Meyer's 
segregates of PM uL Fig 1). Except for Thlaspi s. 
str., Noccidium, and Kotschyella, the other Thlaspi s.l. 
segregates of Meyer (1973, 1979) group with Noccaea 
Moench in a well-supported clade. It was suggested by 
Al-Shehbaz et al. (2006) that only a few of Meyer's 
segregates might deserve recognition (Fig. a o 
include Thlaspi s. str., Neurotropis (DC.) F. 
and only part of Microthlaspi, and the remaining 
segregates should perhaps best be treated as syno- 
nyms of Noccaea (Al-Shehbaz et al., 2006). However, 
our data also indicate that Vania F. K. Mey. might 


ey. 


an 


Volume 96, Number 4 
2009 


Khosravi et al. 569 
Classification for Aethionema trinervium 


Table 2. Distribution of morphological and cytological data in Aethionema, A. trinervium, Thlaspi s. str., Vania, and Noccaea. 


Vania Noccaea 


Thlaspi s. str. 


A. trinervium 


Aethionema 


Character 


perennial 


perennial 


annual 


perennial 


annual or perennial 


Habit 


rosulate or subrosulate 


not rosulate 


not rosulate 


Basal leaves 


oblong or subulate 


spatulate or oblanceolate 


oblong or subulate 


linear or oblong 


> 


sessile 


oblong, ovate, or linear-lanceolate broad ovate, ovate, or oblong 


oblong, ovate, or linear-lanceolate oblong or lanceolate 


linear, oblong, or ovate 


exauriculate or + auriculate auriculate exauriculate or + auriculate auriculate or exauriculate 


very rarely auriculate 


Base of stem leaf blades 


Leaf attachment to stem 


amplexicaul or sagittate 


amplexicaul 


amplexicaul or sagittate 


amplexicaul 


very rarely amplexicaul 


gs 


not saccate 


not saccate 


not saccate 


not saccate 


bisaccate 


Sepals 


No. of veins on petal claws 3 


not apiculate 


apiculate 


not apiculate 


apiculate 


apiculate or not apiculate 


crescent 
smooth 


crescent 
smooth 


crescent 


crescent 
smooth 


semiglobose 


Shape of lateral nectaries 


Seed surface 


striate or reticulate 


smooth or minutely 


papillose 


absent or slightly present 


absent 


absent 


absent 


present or absent 


Seed mucilage 


x=7 


x= = x= 


mosily x = 11, 12 


Base chromosome number 


deserve recognition because this lineage is well 
separated from the core Noccaea group consisting of 
MasmenialPseudosempervivum, Noccaea s. str., Thlas- 
piceras F. K. Mey, Raparia F. K. Mey, and 
Callothlaspi V. K. Mey. (Fig. 1). In the most recent 
systematic treatment of the family, Al-Shehbaz et al. 
(2006) plac 


Thlaspideae and the remaining Thlaspi s.l. lineages in 


ed the members of Thlaspi s. str. in tribe 


tribe Noccaeeae. The primary difference between the 
Thlaspideae and Noccaeeae is the presence of striate 
or coarsely reticulate seeds and often palmately 
veined basal leaves in the former tribe. 

The genus Aethionema comprises approximately 60 
variation in habit 
(annual herbs to shrubs), floral structure (filaments 


species and shows tremendous 


with or without appendages) floral color, fruit 


morphology, and base chromosome numbers (n — 7, 


closely related to Thlaspi s.l. (Schulz, 1936; Al- 
Shehbaz, 1986), but molecular data clearly demon- 
strate that Aethionema is distantly related to Thlaspi 


references therein). Khosravi (1989) recognized two 
unrelated on of Aethionema species, one gro 
with o erve on the petal’s claw, half m ee 
a nectar glands, a a base chromosome number 
of x = 7. These taxa were previously assigned to 
Eunomia (e.g., Aethionema iberidium (Boi i 
Eunomia iberi 
oppositifolia DC., A 
rotundifolia C. A. Mey., A. caespitosum (Boiss.) Boiss. 
E A. 


. oppositifolium Boiss. = E. 


mundi doter Boiss. = EF. 
= E. caespitosa e . E. Schulz, and 

) Boiss. = E. trinervia (DC.) Prantl). 
The second group, he Aethionema core group 


trinervium (DC. 


(including remaining Aethionema species and Moriera 
Boiss.), is characterized by three nerves on the claw, 
semiglobose lateral nectaries, and a base chromosome 


number of x — 11, 12 (Table 2). This group belongs to 


ribe Aethionemeae is currently being 
investigated by one of the authors (M.M.). 

The inis Eunomia has been most recently treated 

a syno d (Appel & Al-Shehbaz, 

2003) or - Tbe L. (Hall et al., 2002), but the latter 

phylogenetic eed demonstrated that I. oppositifo- 

lia Pers. (= 


Fanaa, oppositifolia} is neither related 


recognized as an independent genus. Recent ndhF 
and irnL-F data (Khosravi, unpublished data; Menke, 
unpublished data) provided evidence that E. opposi- 
tifolia is a close relative of Noccaea. Summing up, the 
systematic position of Eunomia needs to be resolved, 


Annals of the 
Missouri Botanical Garden 


as does the taxonomic adm of the ca. 16 species 
ues assigned to i 

molecular study ee shows that Aethionema 
trinervium is outside the Aethionema clade and is 
instead well nested in the Thlaspi clade (Fig. 1). 
Aethionema trinervium is placed with 100% bootstrap 
support into a lineage along with Vania kurdica an 
campylophylla. Vania is on 2 segbetié 
genera defined by Meyer (1991) Members of the 
Vania lineage are xeromorphic, cushion-forming 
plants growing at altitudes between 30 
4000 m. Detailed morphological examination of A. 
trinervium demonstrates that it is very similar to the 
typical representatives of the Vania line 
2006). Van 


pulvinate or subpulvinate habit and simple glabrous 


age (for a 
description, see Meyer. ia species have a 
stems. The basal leaves are spatulate, entire, and 
show a firm texture, whi e stem leaves ar 
amplexicaul or sessile, oblong or oblong-lanceolate, 
the flowers have white petals and apiculate anthers, 
the fruits are oblong or obcordate, wingless or almost 
wingless, and the seeds are smooth and not mucilag- 
inous. The base chromosome number is x = 7. Meyer 
(1973) described three species of Vania: V. campylo- 
phylla, V. kurdica, and V. pulvinata F. K. Mey. All 
Vania species and A. trinervium have a connective 
tooth at the top of the anther never seen in Noccaea. 


NOMENCLATURE AND TYPIFICATION 


Morphological and molecular data strongly support 
the transfer of Aethionema trinervium to Vania, and we 
propose the following new combination: 


Vania trinervia (DC. Khosravi, Jacquemoud, 
enke, mm Mohsenz., comb. nov. Basio- 
nym: Hutchinsia trinervia DC., Reg. Veg. Syst. 


Nat. 2: 387. rn TYPE: A poe Kuh: 
"Hab. in Persiae monte Elwind. Michaux. 
Olivier," e Olivier 1822 (lectotype, designat- 
ed here, G-DC?). Figure 2 


mar 


There is a handwritten label on the lectotype, with 
the following detail: “Hutchinsia trinervia [scripsit A.- 
DC.] / Mont evlind [sic, scripsit Olivier?] / herb. 
Dine 1822 [scripsit x, non A.P. DC." The 
e is the plant specimen attached e that label, 
on the right side of the sheet (see Fig. 2, barcode 
600151559; It should be noted that "i “1822” 
annotation represents the year of accession by de 
andolle, not a date of gathering. Additionally, there 
is an isolectotype, a single sheet and single specimen, 
at Paris (P 00633350 photo!) with the following 
handwritten label: “Hutchinsia trinervia DC. [scripsit 
A.-P. D Iberidella trinervia Boiss. / Aethionema— 


Bss. Fl. Or. [seripsit x] / Amadan. / Mont Elvind. / 


Olivier et Bruguiére [scripsit y].” An annotation label 
y A.-P. de Candolle, *Hutchinsia trinervia DC.," also 
appears on the isolectotype sheet 
known to us for 
consideration. The first specimen (G-DC!, 1: 178, 
G 00131230, one sheet with one 
s a handwritten label: “Iberis / Perse 
A small envelope containing a dissected 


Three syntypes were 


n.6; barcode 
specimen) carrie 
[scripsit x ]." 
flower is attached to this sheet and is annotated by de 
Candolle “petala aequalia / an Lepia?" This was part 
of the db collection by Michaux, and this a 
is mounted the same physical shee 
designated jection. on the left part (see Fig. 2. 
The second candidate syntype at P (P 006633349, P 
otol, one sheet with two specimens) bears a 
handwritten label: “Hutchinsia trinervia DC. [scripsit 
A.-P. DC.] / Iberidella trinervia Boiss. / Aethionem— 
Bss. Fl. Or. [scripsit x] / Perse. Michaux [scripsit y].” 
Finally, a third candidate syntype was examined (C1, 
ex hb. Delessert, one sheet with one specimen), which 
, in Delessert's [cones 
selectae plantarum (Delessert, 1824: 16). This third 
yntype has the handwritten label: “Hutchin. 
trinervia DC. [scripsit x] / Aethionema trinervium 
Boiss. / J. Briquet 1912 [scripsit inn / Michaux 
The G s 


annotation label glu n the first, 


is represented by plate 2, tab. 53 


(Herb. De i [scripsit y].” e has an 
lab ued o Pr states 

“Thlaspi / trinervatum jen ?] / Hutchinsia 

trinervia DC. [scripsit A.-P. DC.].” 

is no doubt that Michaux was the first 

botanist who collected Hutchinsia trinervia. However. 


There 


the taxon was already considered morphologically 
variable by Boissier (1842: Dalar who regarded it 
a representative of the genus Aet a in his 

f Cruciferae (Boissier, 1867: 341-343). 


Therefore, because further taxonomic a may 


as 
treatment o 


perhaps lead to division of A. trinervium s.l. into 
different taxa, the exact ge ogra aphic origin of the 
lectotype needs special attention. Thus, preference 
was given to the collections by Olivier and Bruguiére 

where the label indications d fit the protologue 


locality (de Candolle 


but also young fruits. Although immature, these fruits 
enabled de Candolle to resolve the unsatisfactory 
floral dissection made on a Michaux collection (“an 
Lepia?” written on the envelope, see above) and to 
describe a new species of the genus Hutchinsia. 
urther, the iconograph of Hutchinsia trinervia by the 
French ler strator J. F 
ca. 500) engravings published in leones selectae 


plantarum (Delessert, 1824: 16, tab. 53) (Fig. 3) with 
descriptions by de Candolle. 


Turpin was one of the many 


= 


evertheless, we attempted to find more information 
about the Persian travels of Michaux, as well as of his 


Volume 96, Number 4 Khosravi et al. 571 
2009 Classification for Aethionema trinervium 


"i jn i E 


HERE.PRCDR (6-DC) 
ee 
Pray Mae 600131230 
t^ un. = 
E 
So E 


eterna n Prvepute BC m. ad 
M FPES 


E 4 TIP RUETUSA à (m 


(Ree. Cau) 
| p MUR Syst ME 2: = 
my a AER determ. anno 202.5. a cL tg. 
E E (£v eer E E s 


l p e 
prose an 
7 figere i determ. anno 20 2.57 


DE ole LR cm EDO.) Berks. 


E Mcr TREN at, 2 
7. <>.) 
z Het Po ra 
iesu. 


Pathade Linera e 


copyright reserved = oS 


Lectotype specimen for Vania trinervia (DC.) Khosravi, Jacquemoud, Menke, Mumm. & Mohsenz. (herb. Olivier 
e. 


2. ch 
1 5 rc DC). Photo courtesy of Conservatoire botanique de la Ville de Genèv 


572 Annals of the 
Missouri Botanical Garden 


DICOTYLEDONEA. Cruciferze. Lab. 58. 
ie qe eee 


Lupin del! ef drea f Dien ve 
HUTCIHNSIA trinervia. [De] 
Pe 


Rogn. veg. vod. a pag. 387) 


Figure 3. Engraving of the analytic illustration of Hutchinsia trinervia DC. by Turpin, in Mild lectae pl edited 
by Benjamin Delessert, with descriptions by A.-P. de Candolle. —1. Cauline leaf. —2. Flower. —3. Calyx x with ovary and 
stamens. —4. Ovary and stamens. —5. Petal. —6. Stamen (internal view). —7. Stamen ae us Courtesy of the 


Library, Conservatoire botanique de la Ville de Genéve. 

journey at Hamadan and collecting foray on “Mont information. More accurate data are found in Jaubert 
Elvind” [Alwand Kuh]. However, neither Deleuze and Spach (1842-1843), since the book includes a map 
(1804: 198), the first biographer of Michaux, Sargent with the itineraries of the first botanical collectors in 
(1889: 3), or Boissier (1867: xxvi) provide any relevant the Middle East. Particularly interesting are details of 


Volume 96, Number 4 
2009 


Khosravi et al. 
Classification for Aethionema trinervium 


Michaux's Persian itinerary, which were found in 
handwritten notes held by Delessert. i is unclear why 
aubert and 


absolute lack of reference to these notes by Laségue 
(1845: 61), whose close familiarity with Delessert’s 
herbarium, library, and archives is well known, is 
difficult to understand. Consequently, it is only known 
sa red traveled in Iran from 1783 to 1785 and 
went to Hamadan (according to the map in Jaubert & 
DN [1842-18 
Alwand Kuh, likely on a similar route followed by 


43], certainly crossing part of the 


Olivier and mn about 10 years later in 1796— 
1797 (Boissier, 1867: xxvi). e it should be noted 
that although ucher- Eloy (1843) botanized twice on 
“Elwend” (25 and 29 May 1835) and provided a 
detailed report of his expedition, he does not refer to his 
compatriots Michaux, Olivier, or Bruguiére, nor does he 

n Hutchinsia trinervia (or Aethionema triner- 
vium). 


Literature Cited 


Al-Shehbaz, I. A. 1986. The genera of Lepidieae (Cruciferae, 
Brassicaceae) in the southeastern United States. J. Arnold 
Axbor. 67: 265-311. 

02. Noccaea nepalensis, species from 

and four new combinations in Noccaea (Brassica- 

92. 


a new 


——— A, Bele & E. A. Kellogg. 2006. 2 m 

and phylogeny of the pus (Cruciferae): 
overview. Pl. Syst. Evol. 259: 89-120. 

Appel, O. & I. A. Al-Shehbaz. 2003. Cruciferae. Pp. 75-174 
in K. Kubitzki (editor), Families and Genera of Vascular 
Plants. Springer-Verlag, Berlin. 

Aras, S. Yenilmez. 2003. Isolation of 

DNA for RAPD analysis from dry leaf material of some 

Hesperis L. specimens. Pl. Molec. Biol. Rep. 21: 46la— 

461f. 


Aucher-Eloy, P. M. R. 1843. Relations de cd en Orient 
de 1830 à 1838. Premiére partie. Roret, Par 

Pest V. 1966. Brassicaceae (Aethionema). Pp. 255-269 

A. Takhtajan (editor), Flora of Armenia, Vol. V. Nauka, 
Mo oscow. 

Bailey, C. D., M. A. Koch, M. Mayer, K. ieee ae S. E, 
O'Kane, S. I. Warwick, M. D. Windham & I Al- 
Shehbaz. 2006. Towards a global nrDNA ITS e d of 
the "Ww NET AS . Biol. Evol. 23: 2142-2160. 

Beilstein, M. A., . Al-Shehbaz & E. A. Kellogg. 2006. 
Brassicaceae du ue and trichome evolution. Amer. J. 
Bot. 93: 607—619. 

de E. 1841. Nova Genera Cruciferarum. Ann. Sci. Nat., 

. sér. 2 16: 378-382. 

1842. Plantae Aucherianae Orientales. Ann. Sci. 

Nat. Bot, sér. 2 17: 150-205. 

7. Flora o Vol. 1. H. Georg, Basel. 
Dom. e P. i mls Systema 
Naturale, Vol. 2. ne & Wür z, Par 
Crawford, D. J., R. T. Kimball & M. less 2001. The 

generic placement ofa rue e a species 
in Aster P ce from ITS se es. Pl. Syst. 
Evol. 228: 63-6 


Czerepanov, S. 995. Vascular Plants of Russia and 
Adjacent ae fie Former USSR). Oxford University 
Press, Oxford 

Davis, P. H., J. Cullen, M. J. E. Coode & I. C. Hedge. 1965. 
Materials for a flora of Turkey: X. Notes Roy. Bot. Gard. 
Edinburgh 26: 1 


Delessert, B. 1824 [1823]. Icones Selectae Plantarum, 
. 2. Paris. 
Deleuze, J. P. i 1804. Notice historique sur André Michaux. 
Ann. Mus. Natl. Hist. Nat —221. 


Felsenstein, J. 1985. Confidence limits on nc An 
approach using the bootstrap. Evolution 39: 18319 

Galloway, G. L., R. L. Malmberg € R 
Phylogenetic utility of the nuclear gene arg 
boxylase: An example from Brassicaceae. Molec. Biol. 
Evol. p d "us 

Greuter, urdet & G. Long mae 1986. 
Med- acta "EIS nes. Conservatoire et Jardin 
Botaniques de la Ville de Genéve, Genéve 

Hall, J. C., K. J. Sytsma & H. H. x 2002. Phylogeny of 

sed on chloroplast 


Hedge, I. C. 1965. Aahionema. Pp. 314-330 in P. H. Davis, 
I. Cullen & M. J. E. Coode (editors), Flora of Turkey and 
the East Aegean Islan ds, Vol. 1. Edinburgh University 

Press, Edinburgh. 

—. 1968. Aethionema. Pp. 102-110 in K. H. Rechinger 
57. Akademische Druck- u. 


(editor), Flora Iranica No 
Verlagsanstalt p raz 
er mond. 1980. Cruciferae (Aethionema). 
Den c. Townsend & E. Guest (editors), Flora 
art 2. Ministry of Agriculture & Agrarian 


Hamaguchi, M. A. Busch & D. Weigel. 2003. 
Regulatory elements of the floral homeotic gene AGA- 
MOUS by Kei footprinting and shadowing. PI. 
Cell. 15: 1296-1309. 


Jaubert, H. F. & E. Spach. PR E Illustrationes 
Plantarum Orientalium, Vol. 1. Ror aris 
Karjagin, J. 1953. Aethionema. Po, i 182 in J. J. 


Karjagin (editor, Flora Azerbaijan, Vol. IV. Ministry of 
Culture aa Azerbaijan. 

Ruhid. L. 1979. Aethionema. Pp. 204—209 in N. N. 
Ketskhoveli (editor), Flora of Georgia, Vol. V. Metsnier- 


. Cytotaxonomy and Phylogeny of the 
Dissension, Tarbiat Modarres Univer- 


Comparative Pm Site Mapping of 
E eu e Implies New and Novel Phylogenetic 
Pu us Within Ce ide Ph.D. Dissertation, 
Delhi ve New Delhi. 
Kimura, M. 1980. hod imati l 
rates of base a through comparative studies of 
nucleotide sequences. J. Mp des 16: 111-120. 
1. Thlaspi i s.str. rn 
a os ogical and anatomical 
characters in the light of ITS nrDNA sequence Y PI. 
E Evol. re 209-225 
& . AL Shehbaz. 2004. Taxonomic and 
p evaluation of the American Thlaspí species: 
Identity and d ad to the Eurasian genus Noccaea 
(Brassicaceae). Syst. Bot. 29: 375-384. 
, B. Hau ad & T. Mitchell-Olds. 2001. Molecular 
Cruciferae: Evidence from codin 
J. Bot. 


systematics of the 
plastidic matK and nuclear Chs sequences. Amer. 
88: 534—544. 


Annals of the 
Missouri Botanical Garden 


. Al-Shehbaz & K. Mummenhoff. 2003. 
Molecular systematics, evolution, and population biology 
in the mustard family (Brassicaceae). Ann. Missouri Bot. 
Gard. 90: 151-171 
Lasègue, A. 1845. “Musée lu. de M. Benjamin 
Delessert. Fortin, Masson & Cie, 
Meyer, F. K. 1 Cons eons oe “Thlaspi’’—Arten 
Europas, Afrikas und Vorderasiens. Feddes Repert. 84: 
449—470. 


979. Kritische Revision der “Thlaspi”—Arten 
Europas, Afrikas und Vorderasiens. Feddes Repert. 90: 
129-154. 
. Seed-coat anatomy as a a a a new 
classification of E ed Lus Mundi 9 
ritsche Revision der ar Mn ee 
ein ore und ee Spezieller Teil. 1. 
ey L. Ma s tia 8: 3-42. 
2001b. sche Revision der “Thlaspi”—Arten 
uropas, a d Vorderasiens. Spezieller Teil. 2. 
Eds (DC. )F 
200 


EC pm und Vorderasiens, Teil 11. Haussknech- 

tia 11: 217-22 
Mozaffarian, V. 

species, n 


8. 
1996. Studies of the flora of Iran, new 
new combination and new records. Iranian J. Bot. 
-142 
DNA 
hylogenetic relationships 
s Thlaspi sensu lato o Syst. Bot. 


Mummenhoff, M. Koch. 1994. Chloroplast 


K. 
restriction site variation and 
in the genus 
es 


Franzke $ M. Koch. 1997a. Molecular 
du E Thlaspi s.l. Mic ier based on 
chloroplast DNA restriction site variation and sequences 
of is aa! m iur of Mae d 


DNA. Canad. J. Bot. 75: 469-482. 
——— E 997b. Molecular data reveal 
used in the classification of 

ur esas Bot. J. Linn. Soc. 125: 183-199. 
— U. Coja € H. 


ibd dE 


and pes poses 
indicate Ens A to Thlaspi s.str. Folia Geobot. 


—— Al-Shehbaz, F. T. Bakker, H. P. Linder & A. 
me I Phylogeny, morphological a 
and speciation of endemic Brassicaceae gener: the 
Gine Fofa of. southern Africa. Ann. Missouri Bot. "Card. 
92: 400—424. 

O'Kane, S. L. € I. Al-Shehbaz. 2003. Phylogenetic 
position and generic limits of Arabidopsis (Brassicaceae) 
based on sequences of nuclear ribosomal DNA. Ann 
Missouri Bot. Gard. 90: 603—612 


Peer, W. A., M. Mamoudian, B. Lahner, R. D. Reeves, A. S. 
urphy & D E. Salt. 2003. Identifying metal 
hyperaccumulating plants: Germplasm analysis of 20 
Brassicaceae accessions from a e d area. 


n A. Engler & K. Prantl 
(editors, Die natürlichen "Püanzenfamilien, 3(2). Engel- 
mann, Leipzig. 

Sar ue nt, C. S. 1889. Portions of the journal of André 

ichaux, botanist, written during his travels in the United 
2 tates and ro 1785 to 1796. Proc. Amer. Philos. Soc. 
26(129): 1-145. 

Schulz, O. E. 1933. Kurze Notizen über neue x 

E und Arten der Cruciferen. Bot. Jahrb. 66: 


6. Cruciferae. Pp. 227-658 in A. Engler & K. 
Prantl ei Die natürlichen Pflanzenfamilien, 2nd ed., 
17B. E = mann, Leipzig. 
Shishkin, B 1 T. Vasilchenko (editors). 1948, Flora 
o Vol. 3. Academy of Sciences, Ashkhabad, 
Turkmenistan 


on D. L. 2000. PAUP*. Phylogenetic a Using 


arsimony (“and oe jm Vers. 4. Sinauer, 
ies dut Massachus 
"Thompson, G. "Hisgin ns & T. J. Gibson. 1994. 
CLUST 


L LA Impro oving the sensitivity of progressive 
multiple sequence alignment through sequence weighting, 
positions-specific gap penalties and weight matrix choice. 
cids Res. 22: 4673—4680. 
Al-Shehbaz & C. Sauder. 2006a. 
phragmus (Brassicaceae) based x 
uclear eal DNA. Canad. J. Bot 


—————, A. Francis & I. A. Al-Shehbaz. 2006b. Lou 
ceae: EM Checklist and database on CD-Rom. PI. 
Syst. Evol. ne 249-258. 

1 ns, S. Lee & J. Taylor. 1990. ice ond 
and direct sequencing of por ribosoma. > A gen 

for phylogenetics. Pp. "31 5-322 in M. Innis, D. Gelfan 

1 J. Sninsky & T. J. White (editors}, d o 
A Guide to Methods and Applications. Academic Press, 


k, K., K. Mummenhoff, M. Koch € H. Hurka. 
. relationships of Thlaspi s.l. (subtribe Tans 
pidinae, Lepidieae) and allied genera based on chloroplast 


DNA restriction-site variation. Theor. Appl. Genet. 92: 
375-381. 

——— E pones ree b a rea 
in tribe I 


restriction site variation. Canad. n Bot. 77: 1504-1512. 


A SYSTEMATIC REVISION OF Simon T. Malcomber? and. Charlotte M. Taylor? 
GAERTNERA (RUBIACEAE, 
GAERTNEREAE)! 


ABSTRACT 


The Paleotropical genus Gaerinera Lam. (nom. cons.) comprises 69 species, plus one presumed natural hybrid, of shrubs 
and small trees found from West Africa to Sulawesi in southeastern Asia, with 13 of them newly described here. Gaert is 
ized in the Rubi ip m i i 


do in support infrageneric cM ded of Gaerinera. The morphology d Canaan js notably variable; molecular analyses 
is within regions, and that much of it i homoplasious. Gaertnera species of 

Africa, Madagascar, the bs c a and 5n Lanka are hermaphrodite: and usually iene cea distylous, while those 
bees Asia for which information ti rypt ically so m appar rently E within the 
All species JM ind ed p are vwd prs and D. regional keys are included here. In Africa, 12 

spss an S t gnized in this p work; G. paniculata Sind is he most t widespread of tiese; G. aurea 


ES 
= 


recognized here; G. junghuhniana Miq. is the most commonly collecte xe widespread; G. alstonii Ma Ico 


ispida Aug. inflexa Baill., G. Lege er oe . 0 uci King & Gamble, G. paniculata, G. 


ca. m specimens studie 

Adifipomomes. Africa, Angola, Benin, Borneo, Brunei, Burkina Faso, Cambodia, Cameroon, Central 
African Republic, Côte d'Ivoire, Democratic Republic of the Congo, dioecy, floristics, Fructesca, Gabo 
Gaertnereae, ana, Guinea, heterodistyly, Hymenocnemis, Indonesia, IUCN Red List, Liberia, Madagascar, Mali, 
Mascarene Islands, Mauritius, Nigeria, Pagamea, phylogenetic analysis, Pristidia, Republic of the Congo, Réunion, 
Rubiaceae, ad Sierra Leone, Southeast Asia, Sri Lanka, Sulawesi, Sumatra, Sykesia, taxonomy, Thailand, Togo, 
Vietnam. 


1 S.T.M. gratefully acknowledges the following institutions, agencies, and societies for funding: National Science Foundation 
(NSF 9701008), Missouri Botanical Garden, American Society of Plant Taxonomists, Andrew Mellon Foundation, and 
Washington University in Saint Louis Division of B E and Biomedical Sciences; the following herbaria for T material 

U, 


Holmgren M al. , 1990); L Roger and J. Myer n due and R. Gereau for eatin of Latin os We ackno o 
J.A 


dise, Y. MS B : 
Razafimandimbison, P. M. Richardson, Z. Rogers, Saw Len Gun G. Schatz, Solo, P. Stevens, J. B. 
Sumithraararachichi, S. Teo Ping, J. Thompson, R. Thorstrom, G. Walters, and bua our Meng. S.T.M. d lp 
the following governmental agencies for granting or aid in obtaining field research permits: Association Nationale pour la 
Gestion des Aires Protégées (ANGAP); Ministère des Eaux et Forêts (MEF); Ministère de la Recherche Scientifique Cane 
National de Recherche Appliquée au A de Rural (FOFIFA); National Botanical Garden, Peradeniya (Sri Lanka); 
epar i nd 


Program of "Evolutionary and Population Biology. Washington Dn. St. Louis, Missouri, U.S.A.; and Missouri 
Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A. Current address: Department of Biological Sciences, 
c e State University-Long Beach, 1250 Bellflower Blvd., Long Beach, California 20940- 3702, US: . 

ouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A. t 

d 10. 3417/2002161 


ANN. Missouni Bor. Garp. 96: 575-671. PUBLISHED ON 30 DECEMBER 2009. 


Annals of the 
Missouri Botanical Garden 


Gaertnera Lam. (Rubiaceae) is a widespread and 
morphologically variable genus of small trees and 
shrubs found throughout the moist tropics of Africa, 
Madagascar, the Mascarene islands of Réunion and 
Mauritius, Sri Lanka, continental Southeast Asia, 
Borneo, and Sulawesi (Fig. 1). The genus is found 
from sea level to over 2000 m in evergreen forests, 
and grows in a variety of habitats from upland 
vegetation on clay and loam soils, to low-altitude 
swamps and littoral forests on laterite, white sand, or 
other substrates, to ao altitude moss forests in Sri 
Lanka. This is one of t nera of the tribe 
Gaertnereae, of the subfamily Rubioideae (Bremer & 
Manen, 2000; Robbrecht & Manen, 2006); the other is 
of South America. Gaertnera 
recognized by its woody habit; opposite or rarely 


Pagamea Aubl. can be 
ternate leaves; generally well-developed tubular 
stipules; distinctive ridges or wings that encircle the 

ides and b tend 


upward onto the stipule sheath; cymose, few- to many- 


si ottom of the petiole and often ext 
flowered, bracteate inflorescences borne in a terminal, 


pseudoaxillary, axillary, and/or supra-axillary posi- 
limb 


tion; calyx with a Se developed tube; salver- 
form to funnelform, usually white to pink corollas with 
four or five valvate lobes; the ovary superior in flower; 
and also in the fleshy, purple-black drupaceous fruit 
(Fig. 2). The superior position of the ovary is unusual 
in the Rubiaceae and apparently De derived 
(Igersheim et al., 4). other Rubiaceae 
genera, mostly Nerea, also m ovaries that are 
partially to fully superior at least in fruit (Robbrecht, 
1988). Gaertnera is very similar to Pagamea, and 
to be 


these appear allopatrie sister genera as 
discussed further below. Gaertnera is also similar in 
general aspect to species of Chassalia Comm. ex Poir. 
Chassalia has smooth 


stipules and petiole bases and a fully inferior ovary 


in some regions; however, 
in flower and fruit. Gaertnera is also notable for the 
variation in its breeding o 
Africa, Madagascar, the Mas 
Lanka m which the biology is us all distylous but 


ith the species of 
arene Islands, and Sri 


the species of southeastern Asia apparently al 
with this latter condition apparently 
derived (Malcomber, 2002). Gaertnera is additionally 
notable for its wide variation- in morphological 
ited iati 


many other plants (Malcomber, 2002), with similar 
characters evolved in parallel in geographically local 
radiations. 

Gaertnera has not been monographed comprehen- 
sively since A. P. de Candolle’s work (1845), but 
several regional up ents have been published. 
Baker (1877), Drake (1899), and Verdcourt (1983, 
1989) treated the m i species, which are by far 


d. Drake (1899) a 

s. Petit qe 

1962) studied the African species, with a focus on 
those of the Congo region. Van Beusekom (1967) 
revised the Sri Lankan and Southeast Asian species, 
Madagas 
the Mascarene Islands that he considered related. 
Malcomber and Davis (2005) recently described 


several new species from Madagascar in a review of 


most intensively studie 
he Mad 


e 
treated the Scar specie 


along with some taxa from Africa, car, and 


the Gaertnera species with l- ew-flowered 
inflorescences found in that region. 

We present a comprehensive revision of the species 
of Gaertnera here, based primarily on study of 
herbarium specimens, molecular analyses, and field- 
work by S.T.M. The species circumscriptions are 
based on Malcanitier (2000) and should be credited 
solely to S.T.M. This present work will be supple- 
mented by some additional materials presented online 
(<http://www.tropicos.org/Project/Rubiaceae>). 

As with other tropical Rubiaceae, many species of 
Gaertnera are geographically restricted and/or poorly 
known. In fact, the Mascarene species G. calycina 
Bojer and G. crassiflora Bojer are considered extinct, 
and other Gaertnera species must be considered on 
the verge of extinction given the current knowledge of 
their geographic range. However, the Mascarene 
species G. longifolia Bojer was also considered extinct 


at one time (Walter & Gillett, 1998) but ha 


rediscovered in an area dominated by introduced 


s been 


species (Anonymous, 1997). Similarly, before this 
present study, G. fractiflexa Beusekom, G. globigera 
eusekom, and G. schizocalyx Bremek., all from 
Malaysia, as well as G. microphylla Capuron ex 
Malcomber & A. P. Davis from Madagascar had not 
been collected for 35 to 100 years, but all were 
M. in 1997-1998. With the 


current rate of deforestation in the tropics, additional 


located again 


quired if there is to be any long-term conservation of 
many of these species. 


METHODS 
DESCRIPTIVE DATA 


This revision of Gaertnera is based on the study of 
more than 3500 herbarium specimens from throughout 
its range, and on field studies conducted b 

over 12 months in Brunei, agascar, Malaysia, 
Mauritius, Singapore, and Sri Lanka, during which 
230 collections were made. Specimens were studied 
from the collections cited in the acknowledgments. 
Flower and fruit measurements are based on collec- 
tions preserved in aleohol or rehydrated herbarium 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


EN i e E 
nad Bhp EP 
[155 Poe. YS. J NX ka LU 
c ER e RA 
| É 4 
p $) 7 y ! 
NA. LE ks i A $ is 
no dl 


| AB 


` 30E 


Figure 1. 


specimens. All other measurements (e.g., leaf length, 
width, stipule length, cyme length and width) 
were taken from dried specimens. The species 
descriptions were initially generated using DELTA 
software (Dallwitz et al., 1999} and then edited. 
In the treatment M the morphological charac- 
ters of the 


description. In the individual species descriptions 


genus are summarized in the genus 
for several characters, only the uncommon conditions 
are described, e.g., the fenestrate corolla of one 
species and the pendulous inflorescences of a few 
species. Only selected specimens. are provided a as 


collections and their identifications is included as 
Appendix 1. All collectors are identified there by 
initials of first names where possible; only collectors 
of unnumbered specimens are thus identified in the 
specimens cited in the text. However, some collectors 
only use one name (e.g., Dorr, 1997), while a few first 
names have not been traced. For some relatively small 
areas (e.g., Réunion, Mauritius), internal political 
units are not separate 

An index of all taxa treated in the current revision 
is provided in Appendix 2, with accepted names in 
Gaertnera presented in boldface 

In a few cases, holotype specimens deposited at P 
The searchi 


specimens at that institution was extensive but not 


were not locate hing done for these 


Geographic range of Gaerinera. Points represent individual herbarium specimens. 


exhaustive, and did locate some materials. Thus, we 
are reluctant to conclude that the remaining missing 
specimens are lost and therefore have not yet 
lectotypified several names pending further searching, 
although lectotypification may ultimately be neces- 
sary. 


SPECIES CIRCUMSCRIPTION 


What constitutes a species is controversial, and at 
least 22 different ES concepts have now been 


published (Mishler, 1 


owever, most concepts 


inclusion within the category (De Queiroz, 1998). The 
diagnosable morphological units classified here as 
species are hypotheses recognized by non-overlapping 
character distributions. Units that are diagnosable at 
the limits of their morphological range but not 
diagnosable otherwise and overlapping geographically 
are classified here as varieties. The units recognized 
in this revision satisfy the criteria imposed by the 
general lineage concept of species, and therefore 
represent segments of population-level evolutionary 
1998, 1999). 


phylogenetic analyses of separate and combined 


lineages (De Queiroz, Preliminary 
nuclear DNA (nDNA) data sets for Gaertnera species 

using the nDNA markers ITS, PepC-Large (PepC-L), 
PopC- Small (PepC-S), and Tpi found that multiple 


accessions of the same species all formed a single 


578 Annals of the 
Missouri Botanical Garden 


n, d. 
Lts "y A 


Figure 2. Gaerinera schaizii Malcomber. —A. Flow sien: branch. —B. Portion of stem with leaf bases and stipules 
Long- m flower. —D. Long-styled flower in cross section. —E. Short-styled flower. —F. Short-styled flower in cross 


section. C—F to same l-cm scale. A, B based on Malcomber 2829; C-F based on Moise 9. 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


Table 1. 
GenBank accession number. 


List of sequences included in the molecular analyses. Herbarium voucher information is included within the 


Taxon ITS PepC-Large PepC-Small 
Gaerinera aphanodioica Malcomber AF333817 AY046371 AF333876 
G. arenaria Baker AY185543 AY185544 AY185545 
G. belumutensis Malcomber AF333818 AF333849 AF3338 

G. brevipedicellata Malcomber & A. P. Davis AF333819 AY046374 AY046344 
G. capitulata Malcombe: AY185546 AY185547 AY185548 
G. cooperi Hut M. B. Moss AF333820 AF333851 AF333879 
G. cuneifolia Bojer AF333821 AY046376 AY046346 
G. drakean DC. AF333822 AF333853 AY046348 
G. edeniata Bojer AF333823 AF333854 AF333882 

G. fractiflexa Beusekom 33382 333855 04635 
G. globigera Beusekom AF333825 AF333856 AY046333 
G. hispida A AF333826 AF333857 AF333885 
G. inflexa Baill. AF333827 AF333858 AY046335 
G. junghuhniana Miq. AY046326 AY0463 AY046365 

G. longivaginalis (Schweinf. ex Hiern) E. M. A. Petit 1855 Y18555 18555 
G. longifolia Bojer AF333829 AF333860 AY046337 
lowryi Malcomber AF333830 AY046379 AY046341 
G. macrostipula Ba AF333831 AF333862 AY046343 
madagascariensis “Hook f.) Malcomber & A. P. Davis AF333832 AF333863 AF333891 

C oblanceolata King & Gamble 333833 33386 3339 

G. paniculata Benth. AY046329 AF333865 AF333893 
auciflora Malcomber " A. P. Davis AF333835 AF333866 AF333894 
G ama a (DC.) B AF333835 AF333867 AY046362 
idl. AF333816 AF33384 AY046338 

" rosea Thwaites ex Benth. AF333837 — AF333896 
G. schaizii Maleomber AF333838 AF333868 AF333897 
. schizocalyx Bremek. AF333839 AF333869 AY046355 
G. ternifolia Thwaites AY046331 AF333870 AY046361 
G. vaginans (DC.) Merr. AF333841 — AY046350 
G. viminea Hook. f. ex C. B. Clarke AF333842 AF333871 AY046354 

G. walkeri (Arn.) Blume 333843 333872 04635 
Morinda citrifolia L. AF333844 AF333873 AF333903 
M. royoc L. AF333845 AF333874 AY046370 
Dee guianensis Aubl. AF333846 AF333905 


cluster, and thus represent phylogenetic species 
(Malcomber, 2002, unpublished data). 
MOLECULAR PHYLOGENETICS 


ITS, PepC-L, 


Gaertnera species, Pagamea guianensis Aubl. as a 


and PepC-S sequences for 29 
representative of the sister genus, and the outgroups 
orinda citrifolia L. and M. royoc e., t 
sequences previously analyzed by Malcomber & 
Davis, 2005) were supplemented with ITS, PepC-L, 
and PepC-S sequences for G. arenaria Baker and G. 
capitulata Malcomber (Table 1). The 31 Gaertnera 
species included in this analysis sample the entire 
geographie range of the genus (Gabon, Democratic 
Republic of the Congo, Madagascar, Mauritius, Sri 
Lanka, Malaysia, and Brunei) and include nine of the 
12 species recognized here within the G. vaginans 
complex (discussed below). 


Total genomic DNA was extracted from silica-dried 
leaves using the CTAB miniprep protocol of Doyle and 
reaction R) products of the 
amplified using the primers ITSLEUI and ITS4, and 
the two copies of the fourth intron of PepC were 
amplified using the primers PEPCX4F and PEPCXSR 
in 50 wl PCR reactions as described in Malcomber 
(2002). The ITS PCR reactions produced a single 
product of approximately 650 bp, whereas the Pep 
PCR reactions amplified two distinct products of 
approximately 450 bp and 900 bp. PCR products 
were gel purified using Qiaquick columns (Qiagen 
Inc., Valencia, California, U.S.A.) and 
using pGEM-T easy vector systems (Promega Corp., 
Madison, Wisconsin, U.S.A). Plasmid DNA was 
cleaned using an alkaline lysis/PEG precipitation 


sub-cloned 


protocol (Sambrook et al., 1989) prior to sequencing. 


ITS clones were kequemed using plasmid primers T7 


Annals of the 
Missouri Botanical Garden 


SP6 and internal primers ITS2 (White et al., 
1990 and ITS3B (Baum et al., 1994). To check for 
were sequenced and dimethyl 
sulfoxide was used in both the PCR and sequenc- 
ing reactions, following the recommendations of 
Buckler et a 
in both directions using plasmid primers T7 and SP6 
for the 450 bp (PepC-S) produet and primers T7, 
PEPCINTF, PEPCINTR, and SP6 for the 900 bp 
(PepC-L) product as described in Malcomber (2002). 
Dideoxy sequencing was conducted using the BigDye 


). PepC clones were sequenced 


dye terminator cycle sequencing protocol (from 
Applied Ws en cdd City, California, U.S.A.), 
and sequencing reac 

ABI3100 


sequencer. 


ns were analyzed an 
(Applied DR viden ‘DNA 


Preliminary alignment of the ITS, PepC-L, and 
PepC-S sequences was performed using Clustal W 
(Thompson et al., 1994), before being manually edited 
using MacClade 4.0 (Maddison & Maddison, 2003). 
Sequences were an 
ic methods as 
senbeck & Ronquist, 2001) using the general time 


alyzed using Bayesian phylogenet- 
implemented in MrBayes 3.2 (Huel- 


reversible (GTR) model of evolution and Gamma 

istributed rates (GTR + G), as estimated by 
Modeltest 3.06 (Posada & Crandall, 1998). Each 
Markov chain was started from a random tree and run 
twice for 5,000,000 cycles each, ee M 100th 
cycle from the chain. Four chains w un simulta- 
neously with the initial 2000 Ha discarded as 
burn-in. To verify that stationarity had been reached, 
the fluctuating value of the likelihood was monitored 
graphically. Default settings for the priors were used 
on the rate matrix, branch lengths, gamma shape 
parameter, and proportion of invariable sites. Poste- 
rior probabilities were used to evaluate the support of 
specific nodes 


Taxonomic HISTORY AND GENERIC RELATIONSHIPS 


Gaertnera was described by Lamarck (1792) as a 
genus of Rubiaceae based on a Commerson collection 
labeled as coming from Île de France, today called 
Mauritius. Verdcourt (1989) noted that the Commer- 

tion was undoubtedly made on Réunion 
Um called Bourbon) instead. Gaertnera was named 
in honor of German botanist Joseph Gaertner (1732— 
1791), who researched the seed and fruit structures of 
ribed one 


he places of publication, 


angiosperms. Lamarck subsequently desc 
species, G. vaginata Lam. 
dates, and authorship of these names have been 
widely (and variously) cited incorrectly. Several of 
also described 


Lamarck’s contemporary authors 


genera named for Gaertner, thus complicating the 


nomenclature of this genus. The name Gaertnera is 
formally conserved in the usage detailed here; the list 
of excluded names is relatively long here because 
several of these Gaertnera species were described in 
other, homonymic genera. As an earlier solution to the 
problem of the duplicated genus names, Kuntze 
(1891) replaced Gaertnera as the name used for the 
plants treated here with the name Sykesia Arn., as 
discussed below. Petit (1959a) has a much 
of the nomenclatural history of Gae 

ough Gaertner (1806) Meus id this genus 
in the Rubi 
ovary, Diss (1807) contemporaneously first con- 
cluded that Gaertnera did not belong in the Rubia- 
ceae. He noted that although the opposite leaves, 


Alth 
clearly Dy aceae despite its superior 


fused stipules, flowers arranged in a corymb, 

opposite branching pattern all suggested affinities 
between Gaertnera and members of the Rubiaceae, the 
superior ovary of the fruit instead suggested affinities 


between the Apocynaceae and Rubiaceae. n 
(1814) od x this, describing the family 
Eoo to include Logania 
e t. & G. Forst., 
Willd., Gaertnera, and provisionally 

FE. pu ee (1818) regarded has 


family as not completely P d 


Usteria 


graea Thunb., 
me was 
retained in the Loganiaceae by de Candolle (1845) 
and Bentham and Hooker a pem dy then 
returned Gaertnera to the biaceae ting that 
Pagamea and team fy not differ ciant 
from Psychotria L. and Chassalia because Gaertnera's 
ovary was not didi free, as in other Logania- 
ceae, but fused to the calyx at t 
(1890) quae that 


Rubiaceae based on the presence of raphides and 


e base. Blei 
Gaertnera pow in the 


the absence of intraxylary phloem, and concluded that 
it was related to Psychotria and Chassalia. Gaertnera 
was subsequently usually included s the Rubia- 
ann, ie Klett, 1924; mekamp, 
, 1958; PO T 1988) but 


Cordonoy, 1893; Ridley, 1908, 1915, 1934; Fischer, 
1927, 1928; Hutchinson & Dalziel, 1931). Recent 
molecular analyses have all confirmed the placement 
of Gaertnera and Pagamea in the Rubiaceae and 
agreed on a close relationship pee these two 
genera, M otria, and Morinda L. (Bremer, T 
Andersson & Rova, 1999; Nepokroetf et al., 

A “is relationship between ias He the 
South American genus Pagamea has been hypothe- 
sized since the early 19th century, based on their 
shared sheathing stipules and secondarily derived 
superior ovaries, and more recently based also on 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


their xylem with parenchyma bands and compound 
pollen apertures with crescent-shaped “costae” (Iger- 
shiem et al., 1994; Jansen et al., 6a, b, Je 
Molecular phylogenetic analyses of chloroplast DNA 
(cpDNA) (Andersson & Rova, 1999) found Gaertnera 
and Pagamea to form a well-supported clade related 
to the Morindeae and Psychotrieae, and subsequent 
authors have treated this clade as a tribe (Bremer & 
Manen, 2000; Robbrecht & Manen, 2006). 

Although molecular phylogenetic analyses of mul- 
tiple nDNA (M 
centini, 2007) regions suggest 


alcomber, 2002) and cpDN i- 
t that Gaertnera and 
Pagamea are both monophyletic, few morphological 
synapomorphies distinguish the genera. The only 


inside (ie., adaxial face) of the corolla lobes in 
Pagamea (Taylor et al, 2004) versus glabrous in 
Gaertnera. Another mods ca difference may be 
the development of the distinctive ridge or wing 
surrounding the base of the petiole in Gaertnera, 
which is apparently absent in Pagamea; however, this 
structure is reduced and difficult to detect in some 

ies of Gaert 


used to separate the two genera cannot be interpreted 


nera. Some characters that have been 


as synapomorphies. Gaertnera cannot be differentiated 
of the infl 


a the inflorescence is axillary, while 


from Pagamea based on the position ores- 
cence: in Pagame i 
in Gaertnera it is usually terminal but may be supra- 
axillary (e.g., G. diversifolia Ridl.) or axillary (e.g., G. 
i Schumann (1891) 


oblanceolata King a 
reported that these genera also differ i 


n the ruminate 
endosperm of Pagamea (Piesschaert, 2001) versus 
entire in Gaertnera, but a few species of Gaertnera (e.g., 
G. cooperi Hutch. & M. B. Moss, G. aurea Malcomber) 
also have ruminate endosperm. Wood anatomy and 
pollen morphology may provide additional synapomor- 
to support the monophyly of the two genera: 
) demonstrated that eight Gaertnera 
species from Africa and the Mascarene Islands differ 


phies 
Jansen et al. ( 


from four Pagamea species in wood anatomy and in 
d morpholo. n particular, they documented 
ood differ 


PONE and number of cells in the parenchyma ee 


ogy. 
rences in vessel outline, diameter, an 
and the distance between the bands, and that Pagamea 
an Gaertnera pollen, has a more 
and has 


margo. However, these conclusions are based on a 


pollen is smaller th 
rounded outline in polar view, a less-developed 
partial sampling of both genera and remain to be 
confirmed globa 


GENERIC CIRCUMSCRIPTION AND 
INFRAGENERIC CLASSIFICATION 


Arnott (1836) accepted the placement of Gaertnera 
in the Loganiaceae and established the tribe Gaert- 


nereae for this and Sykesia, which was separated from 
Gaertnera based on the presence of hairs in the corolla 


+ 


hroat and comprised three species from Sri Lanka. 
Endlicher (1838) subsequently included Sykesia in 
Gaertnera but did not nem nomenclatural combi- 
nations for the specie 

De Candolle (1915) Pic two genera again, 
with Gaertnera comprising only Mascarene species. 
He also recognized two sections of Gaertnera, section 
Aetheonoma A. DC. 
calycina, and section “ sectio 
Gaertnera Es to current rules of i T 
with the remaining 13 species. Section Aetheonoma 
was distinguished by its bracteoles paired and borne 
well below the base of the calyx, its broad calyx, and 
its anthers arranged with three alternate to the petals 
and two opposite them; this last condition has not 
been found by others (S.T.M., pers. obs.; Verdcourt, 
1983). Section Gaertnera was characterized by having 
bracts borne at the base of the calyx and the anthers 
all alternate to the corolla lobes, with no calyx 
condition ud The name *Aethonoma" appar- 

y was en fro blished peel 


name, a 205 Steud.,” cited in 
synonymy with G. calycina by de Candolle. No valid 


publication of that name, as a genus or species, has 
een foun 


As de Candolle also noted (1845: 32), Meisner had 
previously published the name Fructesca mauritiana 
DC. ex Meisn. for a plant that was probably a 
Gaertnera, but no description accompanied the name, 
thus it was not validly published and he reported that 
its specimen was so incomplete that its identity could 
ther. There was, h 


ot be confirmed fur 
limited description presented by Meisner for the new 


owever, a 
genus Fructesca DC. ex Meisn. with only one species 
listed in that genus, thus it can be considered validly 
published under our current rules. The only distin- 
guishing characters given were a 5-lobed calyx with 
the lobes linear, a stipule sheath entire or usually with 
two incisions, and a contracted inflorescence. The 
calyx and inflorescence characters suggest this name 
might apply to G. cuneifolia Bojer, but the stipules of 
that are calyptrate and deciduous; thus, the identity of 
this name is still unknown. 
De Candolle’s (1845) circumscription of Gaertnera 
t by (1850), who 


synonymized Sykesia and provided names for its 


was formally expanded firs Blume 
species in Gaertnera. Subsequently, Bentham (1857) 
ykesia and recognized Une 
sections of Gaertnera: sections Aetheonoma, 
gaertnera DC.,” and Sykesia (Arn.) Benth. These 
sections were separated by calyx size and presence 


also synonymized 
f 


versus absence of villous hairs inside the corolla tube. 


Bentham and Hooker (1876) later reduced these to 


Annals of the 
Missouri Botanical Garden 


two sections separated only by corolla size: sections 
the classification of 


Schumann (1891) was the last author to recognize 
an infrageneric classification of Gaertnera, comprising 
sections Aetheonoma and * c o (i.e., section 
a). Within 


scribed two series: series 


section Gaertnera he also 
Datos K. Schum. 


comprised the Mascarene species with sessile flowers 


aertne 


in congested to capitate inflorescences, and series 
Laxiflorae chum. comprised species of various 
regions with sessile to pedicellate flowers in branched 
cymes. Schumann listed G. vaginata, the type of the 
genus, as a member of series Laxiflorae; thus, under 
current rules of nomenclature this is correctly called 
series Gaertnera. 

The genus “Andersonia Willd." was mentioned 
(Roemer & Schultes, 1819) only in a note to the 
treatment of the Neotropical genus Exostema (Pers.) 
B se cript description 


onpl. This note discussed a manus 


of one species, “A. vaginata Willd." written by 
Willdenow based on a plant collected by Dupetit- 
Thouars in Madagascar and did not mention a superior 
ovary. Roemer and Schultes noted that this manu- 
d been assign ned a position 

a om. illeg., 
non Solena Lour.; s equivalent to 
Posoqueria Aubl.) but that its identity was not clear; 
thus, those authors were not treating it as either a 
synonym or a recognized genus and thus were not 
accepting it. They also noted that this genus name ha 
previously been used, as Andersonia R. Br. (Epacri- 
daceae). Thus, Willdenow's name cannot be consid- 
ered validly published and must be regarded as a 
nomen nudum at best. 

The monotypic Madagascar genus Hymenocnemis 
Hook. f. was described based on its combination of 
bilocular ovaries, iis ae e 4-merous flow- 
ly 1-flowered inflor 


lobes, and anthers a uin appendages. Baker 


ers, usually scences, unequal calyx 
(1877) also recognized Hymenocnemis and suggested 
it was closely related to Gaertnera. Schumann 

also recognized both genera and placed 

but 
particularly close relationships for them. The genera 
(1899). 


reevaluated their 


the tribe Psychotrieae, did not suggest any 


Recently, 


Kuntze (1891) noted in his review p feta plant 
nomenclature that the name Gaertnera had been 
published for other plants before the Rubiaceae genus 
was described and used the next available Rubiaceae 


name for the genus Sykesia. Kuntze provided 


combinations for all of the then-recognized Gaertnera 
species; some of these have been noted by previous 
authors (van Beusekom, 1967), but others have been 
overlooked. 

The genus Tsiangia But, H. H. Hsue & P. T. Li has 
been associated with Gaertnera in some lists because 
its sole species was apr described in this genus, 

3 , Bentham (1857) 
noted that this name was ru on a teratologically 


G. hongkongensis Seem. However. 


malformed specimen, and Bridson (2000) agreed and 
formally synonymized this name with fxora chinensis 
Lam. 

No infrageneric classifications have been accepted 
espite Gaertnera’s broad geographic range, 
morphological variability, and the partial correlation 
of some characters with biogeographic distribution, no 
species groups can be clearly delimited within 
Gaertnera by either morphological or molecular 
characters (Fig. 3 


MorPHOLOGY 


Plants of Gaertnera are terrestrial, usually shrubs or 
small trees up to 15 m tall, although a few species 


E 


ternifolia Thwaites, G. d are 
m tall o plants 
may be rarely clambering (e.g. G. elas 


eg, G 
subshrubs growing to only ca. 


Malcomber & A. P. Davis, G. cardiocarpa Boivin ex 
Baill., G. darcyana Malcomber & A. P. Davis). Most 
species are found in pis understory, although 
several species, notably G. obovata Baker and some 
members of the G. vaginans Een (G. o 
Malcomber, G. arenaria, G. junghuhniana Miq., G. 
paniculata Benth., and G. vaginans (DC.) Merr.) are 
often found in open disturbed sites such as forest 
edges. The plants are er well branched, 
although few species (G. obesa x oC. B 

tke, G. monstruosa Male MERE, are pers 
monocaulous. A few species produce differentiated 
reproductive branches with flexuous to pendulous 
uced leaves, often from a supra-axillary 
The 


developmental origin of these branches is not clear 


stems and re 
position (G. diversifolia, G. inflexa Baill.). 
from observation of whole plants and has not been 
investigated here. These reproductive branches vary 
widely in length and are not found on all plants of a 
species; they appear to have evolved in parallel within 
Gaertnera 


bark is smooth in all species except G. 


in Madagascar and in Southeast Asia. The 
ketensis 
Wernham, which has distinctive striated or fissured 


ln 
e 


r 
Plants of Gaertnera vary from glabrous throughout 
the vegetative structures and sometimes also the 
reproductive structures, to densely pubescent with 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


G. arenaria 


G. drakea 


G. hispida (M 
G.m 


(MAD)** 
G. brevipedicellata (MAD) 
na (MAD) 


AD) 
madagascariensis (MAD) 


G. ro 


G. tos (SL) 
G. walkeri ( d 
G. vaginans (SL)** 


) — 


G. lowryi (MAD) 


G.I 
G. cooperi (AFR) 


G. paniculata (AFR)** 
ongevaginalis (AFR)** 


=> Morinda royac 


Pagamea guianensis 


(SAM) 


Morinda citrifolia 


— 0.005 changes 


Figure 
species of Gaertnera, Pagamea guianensis, an 
data for ITS, PepC-L, ad Bae: Su 
analyzed and sequence information are listed in Table 1. 
percen 
ecies names 
circumscription of 


SL, Sri Lanka 


types of indumentum including shortly 


puberulent, tomentose, pilosulose, scabrous, hirtel- 


various 


lous, hirsute, and velutinous (sensu Lawrence, 1951). 
The pubescence of the petioles is not described 
separately here but is generally similar to that of the 
underside of the leaf midrib 


are pubescent when young but lose their vesture to 


. The stems of many plants 


become glabrescent with age. The presence, density, 
and type of pubescence may vary ape a species 
G. etit, G. 
phanerophlebia Baker), but it sometimes appears to 
distinguish species (e.g., G. bieleri (De Wild.) E. M. A. 
Petit vs. G. leucothyrsa (K. Krause) E. M. A. Petit). 
The pubescence colors described here are those of 


e.g. trachystyla (Hiern) E. 


dried specimens; in life, the trichomes are clear or 
whitened. This d 


distinctive for individual species. 


ried color is often consistent and 


Bayesian consensus a 7 hand m and Bayesian consensus cladogram (right-hand side) for 31 
d two oui d 


using the ind time Aaa model and 
Numbers above br: Je in the Bayesian consensus i ia 

ntages. Branches shown in bold the 95 
show, in bold. with double asterisks were tada ded i 
AFR, Africa; MAD, Madagascar; MAU, Mauritius; SEA, Southeast Asia; 


xa (Morinda citrifolia, M. royoc), based on combined sequence 


amma-distributed rates (GTR + G); taxa 


clade 
broad 


are nsidered w ale supported (= 0 


van Beusekom’s (1968) 


No chemical studies of Gaertnera have been seen, 
but some variation in secondary compounds is 
suggested by the colors and patterns found on dried 
specimens. In addition to the dried trichomes, in many 
species the entire specimen of Gaertnera has a 
characteristic brown or chestnut brown color, similar 
to that found in Psychotria subg. Psychotria (Hamil- 
ton, 1989). A few Gaertnera species have a charac- 
teristic orange drying color (G. belumutensis Mal- 
j G. 


Malcomber} that presumably is not an artifact of the 


comber, ochummenii Malcomber, schatzii 
specimen preparation; similar coloration is seen in 
individual species of other genera (e.g., P. stenosta- 
chya Standl). These distinctive dried colors have 
been assumed to be due to characteristic secondary 
compounds, but no actual chemical cause is known. 
When the drying color is not detailed here, it is green, 


Annals of the 
Missouri Botanical Garden 


dull green, brownish green, or gray-green. Many 
specimens of Gaertnera also show small inclusions, 
visible under 7.5% magnification, in the epidermal 
cells of the leaves. These inclusions have a dark 
drying color that contrasts with the rest of the cell (or 
other cells); 
unknown. As wit 


the character of these inclusions is 
other members of Rubioideae, 
Gaertnera has raphides, composed of calcium oxalate 
crystals, as inclusions in its tissues (Robbrecht, 1988). 

e wood anatomy was studied in detail by Jansen et 
al. (1996b 

The stems are variously terete to quadrangular or 
flattened, with the angled or flattened shape of the 


age. Several species have distinctive 
longitudinal ridges or ribs that extend from the middle 
of the interpetiolar base of the stipule down the 
middle of the internode, called “interpetiolar ridges” 
here. This feature appears to be generally consistent 
within a species, although it seems to vary within a 
few species (e.g., Gaertnera paniculata). Van Beuse- 
kom (1967) noted that the data for ae collections 

of G. that the plants 
presumably in the stems. An association ani ants was 


a mention arbored ants, 
also noted i in some populations of G. aphanodioica in 

runei and population of G. oblanceolata in 
Bars Malaysia, during field studies by S.T.M., 
with the ants accessing the stem via holes left by the 
removal of old branches. 

The leaves are decussate and opposite or occasion- 
ally ternate (Gaertnera ternifolia}. The arrangement 
seems to be consistent within a species, and the 
variation in this feature found in some Sri Lankan 
plants is considered indicative here of their hybrid 
origin (G. Xgardneri Thwaites). The leaves are 
subsessile to usually petiolate with a well-developed 
blade that varies from small and narrowly elliptic (6. 
d 0.5-2.5 X 0.1-0.3 

obovate or broadly elliptic (to 55 X 


cm) to y large 


monstruosa, G. obesa). In a few species, the o aid 
in distinguishing species, but in general, as noted by 
van Beusekom the leaves are so similar 
between species that they do not present many 
taxonomically informative characters. 
The leaf margins are entire and generally flat but 
y be markedly cris (e.g.. Caertnera. ianthina 
Malcomber), thinly revolute (e.g., G. ifolia), or 
thickened (e.g., G. diversifolia). The blades 1 range from 
papery in texture to stiffly leathery. This character 
distinguishes some species; the texture described here 
is that of the dried leaves, which is generally 
correlated with the living texture. The secondary 
veins are generally eucamptodromous and broadly 
arching. Most species bear acarodomatia on the 
undersides of the leaves, in the axils of the secondary 


veins with the costa (Robbrecht, 1988) These 
structures comprise small to well-developed tufts of 
hirtellous pubescence on the surface of the blade an 

sometimes also the surrounding veins (“tuft-domatia” 
of Robbrecht, 1988), usually situated within a shallow 
concavity, and additionally may be enclosed by an 
outgrowth of tissue of the veins (“hairy pocket- 
domatia") or deeply concave and surrounded by an 
outgrowth of tissue of the leaves (“ciliate pit-domatia" 
to “crypt-domatia”). Gaertne 


ra species freque 
have several of thes G. 


ntly 
phyllostachya 
Baker); in the descriptions below, only quite uncom- 


e forms (e.g., 


mon domatia types are described. The presence of 


domatia on an individua ant (or individual 
specimen, at least) appears to ae variable in many 
Gaertnera species. 

The form of stipules of Gaertnera is rather unusual 


within the Rubiaceae and distinctive for the genus 


developed tube or sheath that is usually subtruncate 
except with two or four lobes. Generally, similar 
stipules are found in the closely related genus 
Pagamea, but as noted above, Gaertnera differs from 
Pagamea in the wing or rib that extends down from 
the stipule to encircle the base of the petiole (e. d 5 


Figs. 4—6); th 


E These ridges vary from only a little al 


stem distal to the petiole is smoot 


(e.g, G. sralensis (Pierre ex Pit) Kerr) to well- 
o ridges or thin to broad wings. In a few 
nera species with relatively small stature (e.g., G. 
Melo G. walkeri (Arn.) Blume), the stipules are 
persistent and relatively small and similar to those of 
a number of other Rubiaceae genera. However, 
most Gaertnera species, the stipules have relatively 
well-developed tubular sheaths (5-75 mm long) that 


leaving a persistent truncate base. Tw 
stipule forms are found in Gaertnera: tubular stipules 
are cylindrical with an open top in bud, so the leaves 
push through without damaging the sheath, which 
widens as the stem increases in diameter; calyptrate 
stipules are fully fused to form a cap covering the bud, 
so the 


structure as they 


expanding young stem and leaves split this 


form. Both tubular and calyptrate stipules may be 
caducous or deciduous through fragmentation, with 
little evident correlation between persistence an 
form; only tubular stipules have been seen to regularly 
persist on most of the stem nodes. 

In most species, the stipule sheath is cylindrical or 
tubular, but a inflated stipule 
sheath or tube that often widens toward the top to form 


few species have an 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


a funnelform or vase-like structure (Gaertnera obesa, 
G. macrostipula Baker). The dried texture of the 
stipule tube varies from membranous to leathery and 
is considered characteristic. of individual species 
here. The stipule tube or sheath remains entire, or 
often splits along one side to form a spathaceous 
structure, or sometimes splits along two sides to form 
two segments; the pattern of splitting is described 
below in some detail, although as more collections of 
Gaertnera become available, this character appears to 
be less consistent within some species than previously 
thought. The top of the stipule sheath may be truncate, 
asymmetrically l-lobed, or bilobed and may have 
variously two interpetiolar lobes or filaments, four 
h lobes, o lobes. Addit 
Dear numerous setae or short filamentous appendages 
. G. bieleri, G. liberiensis E. M. A. Petit, G. 
hiniora Verde); the development of these setae 


or no itionally, some species 


ies within some spe so some individual plants 


may have these but others not (e.g., G. bieleri). And, 


cies, 


many species of Gaertnera have the stipules further 
ornamented by ridges or wings. The stipule ridges are 
usually four and extend or arise from the ridges or 
wings that encircle the petiole. In many species, these 
stipule ridges or ribs extend upward along the sheath 
to the bases of us stipule lobes, either straight up 

for the somewhat 


quadrate sheath s G. obovata), or angling to fuse in 


from base to apex to form corners 
the middle of the sheath and outline an ornamented 
triangle in the basal portion of the stipule, sometimes 
then extending to the top of the sheath as one rib or 
sometimes separating again (e.g. G. psychotrioides 
(DC.) Baker). Some species have six longitudinal ribs 
on the stipule sheath (G. brevipedicellata). 

The inflorescences are usually terminal on vegeta- 
tive stems but may be displaced to pseudoaxillary 
(Robbrecht, 1988) by subsequent growth, or they are 

sometimes axillary and paired (e.g., Gaertner 


p a ob- 
terminal on relatively short branches 


lanceolata), 
arising from a well-developed main stem (“terminal on 

axillary branches”; e.g., G. obovata, G. divaricata 
(Thwaites) Thwaites}, or terminal on flexuous, supra- 
axillary reproductive branches with reduced leaves 
(G. diversifolia; see discussion above). Van Beusekom 
(1967) diagrammed these various arrangements and 
postulated routes of developmental change that link 
them, although he posited no explan 
supra-axillary inflorescence position. Following Rob- 
brecht (1988), the 


terminal here when they are borne at the stem apex 


ation for the 
inflorescences are considered 


on one peduncle arising from the apical bud and also 
when additional inflorescence sections arise from the 
two axillary buds subtending the terminal bud. This 
latter arrangement has sometimes been called in 


Rubiaceae “tripartite” or “terminal and sessile.” It 


has also been referred to as “terminal and axillary” 
but is not here considered truly axillary in the sense of 
Robbrecht. i 


within a species (e.g., G. divaricata), and the range 


inflorescence position may vary 


of variation in this feature within a demonstrably 
iix genus is notable. 

orescence cymose or thyrsiform 
(Weherling, 197i in EARN ie erect or sometimes 
pendulous (e.g., Gaertnera trachystyla, G. diversifolia, 
G. pendula Bojer), and generally dichasial in 
ranching or, rarely, t igher-order axes may be 


b 
scorpioid (G. divaricata). “The inflo 


multiflowered but may be reduced to a single flower 


orescence is usually 


(e.g, G. E O G. madagascariensis (Hook. f) 
Malcomber & A 
(e.g. G. rosea Thwaites ex Benth., G. pauciflora 


P. Davis) or few-flowered fascicle 


Malcomber & A. P. Davis), or "mu into a 
subcapitate head (e.g., G. globigera, G. rotundifolia 


Bojer) or spike (G. spicata K. Schum.). Van Beusekom 


— 


1967) diagrammed the various arrangements and 
postulated routes of development that link them. It is 
noteworthy that inflorescence morphology in general 
is homoplastic, and the hypothesized series of steps 
leading to the different forms is not supported by 
our best phylogenetic estimate for the genus. The 
of the brane 


characterized below 


general shape hed portion is usually 
bifor 


species as corymbiform (i.e., 
rounded in outline), pyramidal (ie., thyrsiform or 
paniculiform, generally conical in outline), or sub- 
Anais (i.e., congested- uh to dau This 
ange of variation in inflorescence arrangement is 
sole but similar to D pa in some other 
Rubiaceae genera 
The inflorescences are sessile to usually peduncu- 
late and bracteate with the bracts generally relatively 
small (0.1-5 mm long) but sometimes well developed 
e.g. Caertnera cuneifolia) and showy (e.g. G. 
phyllostachya). As in many Rubiaceae, there is often 
variation in the development and form of the bracts 
from the lower to upper nodes of the inflorescence, 
with the bracts 
structures with generally elliptic to Miri Sa shape) 
at the dis 
reduced subsessile leaves and/or EE E stipules at 
the basal nodes. Van Beusekom (1967) illustrated 
everal disti 


“normally” shaped (i.e., reduced 


talmost nodes but frequently resembling 


tinctive braet s and postulated the 
developmental changes that ink them. The position of 
the bracteoles, at the base of the calyx and/or borne on 
the pedicel, does not appear to be taxonomically 
informative in Gaertnera, although it has been cited by 
some previous authors. The inflorescence axes and 
bracts are usually green but are attractive and bright 
white in a few species (e.g., G. phyllosepala Baker, G. 
phyllostachya). The flowers range from sessile or 
subsessile to usually at least shortly pedicellate. 


Annals of the 
Missouri Botanical Garden 


The flowers are similar in general aspect and most 
details to many other Rubiaceae, except they are 
notable in their superior ovaries and their unusual 
variation in reproductive biology, which is either 
bisexual and heterodistylous or dioecious with the 
flower forms quite similar (e.g., Gaertnera aphano- 
dioica) to distinctly different (e.g., G. junghuhniana). 
Van Beusekom (1967) illustrated much of the range of 
shape, 
anther position, and stigma and style shape and 
position. The breeding biology is discussed in more 
detail below. The flowers vary with unusual frequency 


variation in corolla size, pubescence, and 


among Rubiaceae within species, between popula- 
er of 


condition is used here to characterize most species, 

as done by previous authors. Detailed measurements 

are provided sepa arately here for long-styled and short- 
tyled flow or pistillat 


These salio: are combined into a single measure- 


owers e and staminate flowe 
ment in Rubiaceae treatments, on the assumption that 


the flowers are generally similar. However, ps 
ted e 


assumption is rarely tested in general an 
variation. in floral form is unusually broad p 
Gaertnera, so these measurements seem useful to 
present here. In most species, the flowers do seem 
essentially similar in the two forms, but in some 
species there is a small difference that may be found 
in future studies to be significant. 


The gamosepalous calyx is generally cup-shaped 


developed with relatively large, sometimes bright 
white, sometimes fewer lobes (Gaertnera d ur 
G. phyllostachya). The measurements giv e for 
width of the calyx describe the diameter j its top at 
anthesis, except for species with relatively large a 
lobes that are narrowed at the base, in which t 
diameter given is that of the top of the unlobed um 
of the calyx. The relatively well-developed calyx lobes 
of some species (G. phyllosepala, G. phyllostachya) 
of Robbreeht 
led 


calyco- 


oo. to the “enlarged calyx lobes” 
( y have sometimes been cal 
phylls” an 1996), “petaloid calyx lobes,” or 
“semaphylls” but lack the differentiated stipe, broad 

large size of most suc 
structures eae flowers. When present, the 
relatively well-developed calyx lobes may be gener- 


E 


lade, and relatively ve 
ubiac 
ally equal to markedly unequal, with sometimes only 
one or few lobes developed on an individual flower. 
These enlarged calyx lobes may be developed on all 
the flowers of an inflorescence or only on a few 
flowers, and their number may vary on an individual 
plant (e.g., G. phyllostachya). The exterior of the calyx 
may be glabrous or variously pubescent and often is 


similar in pubescence to the inflorescence axes. Its 
interior is usually glabrous but sometimes bears a ring 
“hair- 


of well-developed, uniform pubescence, or 


ring”; this structure was studied and is described for 
all species but does not seem to be taxonomically 
informative. 

e corollas are salverform to funnelform at 
anthesis, with four or five lobes that are valvate in 
bud. The corollas are usually white but may be pink 
(e.g., Gaertnera brevipedicellata, G. macrostipula), red 
to orange (G. spicata), or blue 

e in Gaertnera ranges from 2-30 mm 
eee may be glabrous (e.g. G. iE G. 
phyllosepala) or variously pubescent externally, and 
(e.g.. 
pauciflora, G. phyllosepala) or densely villous in the 
upper half or third, with this pubescence sometimes 


P dd Corolla 
The 


internally are usually either glabrous 


extending onto the basal part of the internal or adaxial 
face of the lobes (e.g., van Beusekom, 1967: fig. 4). 
However, most of the internal faces of the corolla 
lobes are glabrous or, in a few species, glaucous. In 
most Gaertnera species, the corolla lobes are generally 
triangular to ligulate and flat, but in a few species they 
are thickened at the apex and prolonged into a hooked 
adaxial protuberance, and sometimes these apices are 
thickened abaxially enough to produce a sharply 
angled, pyramidal top to the flower bud instead of the 
smooth-sided, rounded to obtuse buds found in most 
species. Gaertnera cooperi is unique in its corolla 
lobes that are enlarged and cucullate at the apex and 
orm a markedly enlarged and angled top on the flower 
bud. The corollas of G. cooperi are also unique within 
the genus in being fenestrate in the upper part of the 
tube. Fenestrate corolla tubes are known in several 
her Rubiaceae genera of various tribes and 
1988; Piesschaert et al., 


been reported ae 


continents (Robbrecht, 
2001; 


Pagamea). The function of this corolla form is 


ut they have 
unknown. 

The stamens are alternipetalous and inserted in the 
middle or a p of the 2 tube, or rarely in 
the lower part some ered species. 
Gaertnera a was said m CLE 1845) to 
have three alternipetalous anthers and the other two 
anthers borne opposite to the petals, but this has not 
been seen by us or remarked on by Verdcourt (1983), 
and perhaps was an artifact of the dried specimens 
studied. The anthers are bithecal, dorsifixed near the 
middle or sometimes the base, and open via lateral 
slits. No appendages on the connective or anthers 
have been seen by C.M.T., although Malcomber and 
Davis (2005) reported that these are found in the 
genus but did not mention any particular species. The 
size of the anthers ranges from 1.5— and is not 
described in detail here because it was not found to be 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


variable between floral forms of distylous species or to 
be informative for separating closely related species. 
The anthers may be exserted or included, and either 
position may be found in either flower form depending 
on the species. The long-styled and short-styled forms 
of distylous Gaertnera species do differ in the length 

the filaments, often also in the position of the 
anthers (included vs. exserted), and rarely in the point 
of filament insertion on the corolla. The position of the 
anthers in each flower form of distylous species is 
similar to the position of the stigmas in the other form 
of that species. The staminodes of pistillate flowers 
may be reduced or similar to the stamens of the 
staminate flowers. The pollen was studied by Jansen 
et al. 

The ovary of Gaertnera is bilocular and secondarily 
superior as recently shown by Igersheim et al. (1994), 
with this superior position arising early in the 
development through upward 
intercalary growth of the ovary tissue. No Gaertnera 
species has an inferior ovary position. Igersheim et al. 
suggested that other Rubiaceae with superior or semi- 
inferior ovaries follow the same developmental 
pathway, but this appears not to be the case at least 


in another eg of Rubiaceae that was also 
a usly plac Loganiaceae, Mitrasacmopsis 
Jov ecos et al., 2007). As noted above in 


the td sunmary af the classification of this 
im the homology of this unusual ovary position has 
not s bee interpreted correctly. The ovary is 
2 to ellipsoid and surrounded by a ring- 
shaped nectary. Each locule of the ovary has one erect 
basal ovule. The style is slender and glabrous or 
per portion. The two 
well developed and linear to 


sometimes pubescent in the u 
stigmas are short to 
clavate; they are sometimes reduced in staminate 
flowers. The long-styled and short-styled flowers differ 
also in the length of the style; as noted above, the 
position of the anthers in each floral form of distylous 
species is similar to the position of the stigmas in the 
other form of that species. The pistillodes of staminate 
flowers may be reduced or similar to the stigmas and 
style of the pistillate flowers. 

The fruits of Gaertnera are succulent, drupaceous, 
globose to ellipsoid or rather didymous, and so far as 
In a 
few species from Mauritius, the fruits have been 
reported to be whitened (Bojer, 1837), n this has not 
been seen by others (D. H. Lorence, pers. comm.). The 


observed, all black or violet-black at maturity. 
M 


fruits range in size from 4—2 mm, with size 
varying to some extent between species and also 
depending on the number of pyrenes produced, but as 
noted by van Beusekom (1967) the fruits do not show 
much infrageneric variation. The fruits contain one 


generally subglobose or two generally plano-convex 


pyrenes; the number of pyrenes per fruit varies among 
individual flowers and is assumed to depend on 
pollination and other developmental factors affecting 
the number of ovules that develop in the individual 
fruits. The shape of the fruit also depends on the 


smooth to ribbed, p or deeply fissured. This form 


is usually visible on dried specimens because the 
fl 


ruit wall de to the pyrene shape. T 

marginal preformed germination slits and a third, 
median slit on the ventral surface have been seen in 
s. obs.); this is a 
common pattern in Rubiaceae (Piesschaert, 2001). 


several Gaertnera species (S.T.M., per: 


The seed is generally elliptic to subglobose and is 


smooth to sometimes rugose or invaginated with 


ruminations that follow the shape of the pyrene wall 


a 


e.g., G. aurea, G. cooperi). These ruminations are 
shared with Pagamea, and thus considered plesio- 
morphic in Gaertnera. The endosperm is starchy. 


REPRODUCTIVE BIOLOGY 


The flowers of Gaertnera are insect pollinated, 
stly by bee 
ors. There is apparent 


parently m es and/or other generalist 


ap 
pollinato y wide variation in 
pollination mode within the genus as shown by the 
wide range of flower size (e.g., G. edentata Bojer with 


Bojer with 


corollas 38-40 mm long, G. divaricata with corollas 


corollas 17-35 mm long, G. crassiflora 
6.5-8 mm long), the variation in flower color (e.g., G. 
spicata with corollas red to orange-red, G. edentata 
with corollas white), and other aspects such as 
pendulous versus erect inflorescences, fenestrate 
versus entire corolla tubes, and truncate to large- 
lobed calyx limbs. The plants flower diurnally and last 
for a few days with usually only a few flowers open at 
one time on a particular inflorescence. Several species 
from Mauritius with d large, fleshy, white 
flowers and reportedly et odors (e.g., G. edentata, 
G. hirtiflora, G. umi resemble flowers of other 
Rubiaceae that are nocturnal or crepuscular and may 
be moth pollinated; however, they are odoriferous in 
the middle of the day (D. H. Lo 


The distylous species are protandrous. 


rence, pers. comm.). 

The fleshy, drupaceous fruits of Gaertnera are 
presumably similar to those of other Rubiaceae in 
being bird dispersed; however, no active dispersal of 
Gaertnera fruits was observed during this study, and 
no published dispersal observations have been found. 

Caertnera i is either distylous (Fig. 2) or dioecious 


homostylous and bisexual, but this has not been 


Annals of the 


Missouri Botanical Garden 


ure 4. A-F. Gaerinera oblanceolata King & Gamble. —A. Flowering branch. —B. Portion of stem with Sea bases 


and stem apex. zT Fruiting branch. C-F to 


confirmed by us. Gaertnera aphanodioica (Fig. 5A—F) 
is notable in being morphologically apparently 


homostylous, but functionally dioecious based on 


pollination studies by S.T.M. in Berakas Forest 


in iru 


population were regar 
hrodi 


collections from this 
van Beusekom (1967) 
hermaphroditic un homostylous forms of 


Reserve, Brunei, 


vaginans. All the other species for which breeding 
biology is known are either evidently dioecious or 
distylous. The breeding biology of several species 
cannot be determined from the information available; 
such species from Africa, Madagascar, the Mascarene 
Islands, and Sri Lanka are presumed to be probably 
distylous, and the unknown species from southeastern 
Ásia are expected to be eventually demonstrated to be 
dioecious. The evolutionary aspects of the breeding 
biology of Gaertnera are discussed further below. 


BIOGEOGRAPHY, HABITAT, AND DISTRIBUTION 


s circumscribed here, species of Gaertnera are 
regional endemics in Sri Lanka (five species and one 
hybrid), southeastern Asia (16 species), Africa (12 
species), Madagascar (26 species), and the Mascarene 
Islands (nine species in Mauritius, one species in 
Retinion). Accordingly, regional keys are presented to 


e flower in 
m ases, stipules, 


ole 
same 5-mm scale. A—F based on Mine 3039; G-I based on Malcomber 2875. 


the species of the genus. The relationships among the 
species are not entirely clear (Malcomber, 2002), but 
the geographie patterns suggest the genus has a 
history of radiation within newly colonized regions, 
s Madagascar, Mauritius, 


. Robbrecht (1996), in his 
rican Rubiaceae biogeography, charac- 


especially on islands such a 
ri Lanka, and Bor 
of Afri 


terized Gaertnera in the category of “12 Paleotropical 


survey 


Genera, 122 mainly present in Africa and/or Mada- 
gascar," apparently based in large part on the work of 
Petit (19592) and van Beusekom (1967). More species 
are recognized here in all of these regions but 
particularly in 
ithin a particular biogeographic region, the 

Gaertnera species for which information is available 
show a consistent breeding enun i el 

ost of the range of the genus and d 
Senes Asia. Van posiom (1967) pit 
species of broad geographic range with mixed 
distylous-dioecious breeding biology, but these cir- 
cumscriptions are not supported by the data or the 
species descriptions on which this treatment is based, 
as discussed in detail below for the G. vaginans 
complex. 

The 12 African species of Gaertnera are found from 
Senegal in the northwestern part of the continent to 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


Zambia in its central southeastern region, from sea 
level to 1720 m (Fig. 1). The most widespread species 
is G. paniculata, which is found throughout this range. 
Most of the species are restricted to West or Central 
Africa, and most are found below 1000 m elevation. 
Presumably, the apparently unes distributions of 
species that appear to “jump over” some countries are 
artifacts of limited specimen representation rather 
than the actual range (e.g. longivaginalis 
(Schweinf. ex Hiern) E. M. A. Petit documented in 
the Democratic Republic of the Congo, Cameroon, and 
Angola but not Gabon). In the phytogeographic 
classification of Takhtajan (1986), Gaertnera is found 
in ineo- pius and no-Zambesian 
regions, with all of its African species Wes in the 
first region and two additionally ranging into the 
second. In the more detailed phytogeographic classi- 
fication of White n all 12 of the African species 
of Gaertnera are fou “Guineo-Congolian 
with 10 
restricted to this region, while the other two species 


regional centre of ee species 


also range into the “Guineo-Congolia/Sudania a 
lo culata) 
and the “Guineo-Congolia/Zambesia sud transi- 


al transition zone" (G. longivaginalis, G. p 
tion zone" (G. longivaginalis, G. paniculata), and only 
G. paniculata further ranges into the “Sudanian” and 
“Zambezian regional centres of endemism.” The 
species that are only found in the “Guineo-Congolian 
regional centre of endemism” are all further restricted 
to the west (G. aurea, G. cooperi, G. liberiensis) or 
east (G. bieleri, G. eketensis, 
comber, G. letouzeyi Malcomber, G. leucothyrsa, G. 


G. gabonensis Mal- 


spicata, G. trachystyla) of the Dahomey Gap. 
Some changes to White’s phytogeographic classifica- 
tion have subsequently been proposed, but as 
detailed by Friis (1998), these affect mainly the 
Afromontane regions and do not touch on the African 
range of Gaertnera. Robbrecht (1996) recognized 
several generic distribution patterns of African 
Rubiaceae; Gaertnera generally agrees with his 
“Africa-wide” distribution except it is not present in 
East Afric 

In de six species of Gaertnera and one presumed 
natural hybrid are restricted to Sri Lanka, while 16 
species are found in southeastern continental Asia, 
Borneo, and Sumatra (Fig. 1). On Sri 


nera is found at 5 


Lanka, Gaert- 
0-2000 m ato in wet evergreen 
forests, generally in the southwestern and central part 
of the island; G. vaginans is the most commonly 
collected species. Sri Lanka was classified phytogeo- 
graphically by Takhtajan (1986) in the Indian region, 
which he considered a separate region from south- 
eastern Asia. In southeastern Asia, Gaertnera ranges 
from sea level to 2750 m elevation, and 14 species are 


restricted to the region of Peninsular Malaysia, 


Borneo, and Sumatra, classified by Takhtajan in the 
Malesian region. Gaertnera junghuhniana, the most 
commonly collected species in this region, is mainly 
distributed in the Malesian region but is also found in 
peninsular Thailand, which Takhtajan classified as 
part of the adjacent Indochinese region; in contrast, G. 
sralensis is mainly distributed in the Indochinese 
region but is also found in Peninsular Malaysia in the 
Malesian region. 
Gaertnera has its center of species richness (as well 
s morphological diversity) in Madag and t 
Munere Islands (Fig. 1), which di ecce ia. 
ly comprise most hs the Madagascan region of 
Mars ed (1986). Here, Gaertnera is found mainly 
Madagascar (26 or possibly 27 species) but is also 
well represented on Mauritius (nine species) and 
Réunion (one common species), with all the species 
currently each known from a single one of these 
islands. Several species of Gaertnera of Mauritius may 
have grown in Madagascar in historical times (e.g., G. 
crassiflora), or these reports may be labeling errors. 
n Mauritius, Gaertnera species are found in wet 
forests and heathlike, “groundwater laterite” vegeta- 
tion at 50-800 m, with G. psychotrioides the most 
commonly collected species. Gaertnera vaginata is the 
only species known from Réunion, where it grows in 
wet forests at 50-1800 m 
The geological m" of Madagascar were 
mapped and included in vegetation analyses by Du 
Puy and Moat (1998, 2003); Gaertnera species in 
Madagascar are generally found on metamorphic and 
igneous basement rocks, except a few species that 
grow additionally to exclusively on sedimentary 
mada- 
gascariensis) or unconsolidated sands (e.g., G. guillotii 
Hochr.). Du Puy and Moat also classified the 
vegetation types of Madagascar; Gaertnera is found 


substrates such as alluvial deposits (e.g., G. 


here in evergreen humid forests, which are distributed 
along the entire eastern part of the island, through the 
east-central highlands region (although very little 
natural vegetation remains here) and westward 
through the mountainous northern region. sl 
in Du and 


evergreen, Qus 


grows in Madagascar at 0—2000 m, 
Moat's 


forest: low alt 


"coastal forest (eastern)," * 


few Gaertnera species here are found 
only at ca. 300 m elevation or below (G. cardiocarpa, 
G. drakeana Aug. DC., G. hispida Aug. DC., 

schatzii), but most of the species are found at higher 
elevations. By far, the most commonly encountered 
species is the widespread, morphologically variable G. 
obovata, in darme. its variety sphaerocarpa (Baker) 
by Davis and Bridson (2003), 


Gaertnera is one of the most commonly encountered 


Malcomber. As noted 


Annals of the 
Missouri Botanical Garden 


genera of Rubiaceae in mature, well-preserved forests, 
although it is not often found in secondary vegetation 
and this is not a relatively species-rich Rubiaceae 
Madagascar. Gautier and Goodman (2003) 
recently reviewed the phytogeography of Madagascar 


genus in 


and noted that although the island's flora is known for 
its high degree of endemism, Paleotropical genera 
such as Gaertnera as well as pantropical genera also 
comprise an important part of the flora. Gautier and 
Goodman recognized five phytogeographic regions; 
Gaertnera is found in three of these: the Eastern 
region, the Central region, and the Sambirano region. 
Several Gaertnera species are restricted to only one of 
these regions and are known from only a small range 
here within the respective region (e.g., G. schatzii, G. 
brevipedicellata, G. bambusifolia Maleomber & A. P. 
Davis), while a few species are found in all three of 
these regions (e.g., G. obovata). 


DIVERSIFICATION AND EVOLUTION IN THE GENUS 
PHYLOGENETIC RELATIONSHIPS AND ORIGIN OF GAERTNERA 


The Bayesian phylogenetic analysis of 31 Gaertnera 
species with Pagamea guianensis and two Morinda 
species as outgroups estimated a well-supported (99% 
posterior probability [PP]) Gaertnera clade, but only 
b PP) clades within the 
equence diver- 
gence among sampled species in the ITS, PepC-L, and 
PepC-S data sets. Gaertnera cooperi is estimated as 
sister to all other sampled species, followed by another 
African species, G. longivaginalis. The position of the 
remaining sampled African species, G. paniculata, is 
unresolved in a clade containing sampled species 
from Madagascar, Mauritius, Southeast Asia, and Sri 
Lanka. A well-supported clade of Malagasy species 
(9896 PP), excluding G. lowryi Malcomber, is found to 
be sister to a clade of sampled Mauritian species. 
Gaertnera lowryi is found to be sister to a well- 
supported clade of all the sampled Southeast Asian 
and Sri Lankan species, suggesting that the progenitor 
of the 


originated in northeastern Madagascar (e.g., 


Sri Lankan and Southeast Asian species 


T 
[2] 


Masoala Peninsula). However, this sister relationship 
is not well supported (92% PP), 
characters and Gaertnera species will need to be 
included to test this hypothesis. Sampled Southeast 


so additional 


Asian species from Peninsular Malaysia, Singapore, 
and western Sarawak form a well-supported clade 
(99% PP) that is unresolved relative to the other Sri 
Lankan and Southeast Asian species. Sampled Sri 

species do not form a single c 
(97% PP) 


estimated between G. ternifolia and G. walkeri. 


ade, although a 
shi 


ankan 
well-supported sister relationship is 


Similar to the results of Malcomber (2002) and 
Malcomber and Davis (2005) with data sets with 


[e] 
o 


mon eotide sequence data, our phylogenetic 
analysis estimated an African origin for Gaertnera 
followed by a limited number of dispersal events from 
Africa to B a Madagasacar to Mauritius, and 
anka/Southeast Asia. Molecular 


clock analyses nee that these dispersal events 


Madagascar to 


and radiation of the genus occurred within the last 5— 


5.5 million years (Malcomber, 2002). 


MORPHOLOCICAL AND REPRODUCTIVE DIVERSIFICATION 


Gaertnera is notable in Rubiaceae and also in the 
higher plants in the plasticity of morphology and 
within a 


breeding systems g 
relatively recently derived species 


roup of ie relate 


. The p 


plasticity and the varied directions in uus selection 


"T 


has pushed it are remarkable for both the range and 
number of novelties (e.g., the fenestrate corollas of 

cooperi, the spiciform inflorescence of G. spicata, the 
enlarged calyx lobes of G. phyllostachya) and the 
repeated derivation of unusual morphological features 
in species of different regions (e.g., the Ped large 
campanulate stipules of G. obesa and G. monstruosa; 
the pendulous supra- axillary ieee bearing 


vary in Gaertnera as much as in some 
other Rubiaceae genera of comparable size, in 
particular number of ovary locules and stigmas, 
corola and fruit colors, and the size of the plants 
(all Gaertnera species are small trees or shrubs up to 
15 m tall). 

Notable morphological variation in Gaertnera 
includes leaf size, which ranges from 0.5-60 cm long, 
and stipule form. The stipules are always united into 
at least a short shea 
to coriaceous, calyptrate to tubular or funnelform, and 


th, but are variously membranous 


caducous to persistent, with only the basal portion to 
e entire stipule persis ab The stipule tube is 
d with ribs and/or well- 


developed wings, or Bd ae the stipule apex is 


variously ridged, ornamen 


entire, lobed, possessing a combination of lobes and 
setae, or with only setae. Most Gaertnera species have 
terminal inflorescences borne on principal branches, 
but G. inflexa, G. diversifolia, G. oblanceolata, and G. 
divaricata have inflorescences on axillary or supra- 
axillary branches. Inflorescence arrangement varies 
from corymbiform-cymose to lax and thyrsiform, 
capitate, or reduced to a few-flowered cyme or solitary 
flowers. Calyx lobes are usually relatively small and 
deltate to linear, but are sometimes filiform or 
expanded into petaloid white structures that function 
to attract pollinators. The corollas are generally white 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


in all but one species, and range in size from 2— 
30 mm long. 

Species of Gaertnera appear to hybridize, in some 
cases perhaps frequently, as in some other Rubiaceae 
genera (e.g. Cinchona L.; Andersson, 1998). In 
particular, a distinctive group of plants that have 


intermediate but somewhat variable in characters and 
grows in the contact zone of its two presumed parents, 
G. ternifolia and G. walkeri, on Adam's Peak. Several 
species from Madagascar also appear to hybridize 
(e.g., G. phanerophlebia, G. humblotii Drake). Whether 
these hybrids are fertile is unknown; if so, this process 
may be part of the reason for the diversification and 
close species relationships seen in Gaertnera. 

Ás discussed above and elsewhere (Malcomber, 
2002), species of Gaertnera may be either dioecious or 
M and the Ven is also one of only nine 

porte ses in loecy has ded 
followed a Ap are and evolved fro 
distylous ancestors (Muenchow & M 1989). 

The available evidence—biogeography, molecular 
sequence data, and morphology—shows that this 
notable morphological diversity has evolved relatively 
rapidly and also that several unusual features of 
individual Gaertnera species have evolved more than 
once, in parallel, within the genus, in different regions 
presumably in response to similar selection pressures. 
Thus, Gaertnera provides an example of the ways in 
which tropical Rubiaceae have diversified, shows the 
dynamie environment in which tropical plants live, 
and provides an object lesson for systematists about 
the nature and “informativeness” of both morpholog- 
ical and molecular characters. 


THE GAERTNERA VAGINANS COMPLEX 


Van Beusekom (1967), in his revision of Asian 
Gaertnera, adopted a broad species concept. In 
particular, he circumscribed G. vaginans as on 
widespread, morphologically and biologically aa 
species with two subspecies and Helka in it plants of 


Africa Madaga car. Sri Lanka Asia with 


cylindrical stipules that closely encircle the stems, 
subcapitate to diffusely cymose inflorescences, 5- 
merous flowers, reduced calyx lobes, and variously 
distylous, homostylous, and dioecious breeding sys- 
tems. His subspecies vaginans comprised all the 
distylous plants of this group and all the plants of 
Africa, Madagascar, and Sri Lanka; his subspecies 
junghuhniana (Miq.) Beusekom comprised all the 
dioecious plants of this group and all the plants of 
Southeast Asia including Borneo and the Indonesian 
archipelago. Thus, van Beusekom's circumscription of 


vaginans resulted in a single species that was 
distinguished by a set of characters that are widespread 
and perhaps ancestral within the genus. His 
vaginans additionally was distributed across sera 
broad biogeographic regions and notably polymorphic 
in breeding system, and thus questionable as to the 
reproductive cohesion of the regional populations. 
Such broadly diagnosed widespread taxa that are 
morphologically similar in a general way yet poly- 
morphic in several features are recognized in most 
taxonomic studies of plant groups. Such polymorphic 
axa are d when the observed 
morphe area variation is not eea significant 
(e.g., Vink [1970] grouped 39 local “entities” into a 
roadly circumscribed Drimys piperita Hook. f. 


= 


someti Imes recognize 


[Winteraceae]. and sometimes they result from a 
lack of data adequate to support further subdivision 
within the group (e.g., Maldonado aye reduced half 
the previously described species of Elae 


> 


Rubiaceae] to synonymy based on a a pane pal 
coordinates hs of characters of individual 
KR at failed to distinguish diagnosable 
ps; fitum [1998] re 


eta variable species of Cinchona diag. 


cognized several broadly 


nosed by a unique combination of common characters 


S 
za 


the genus). a such morphologically 
plants are d as 


variable sets sometimes treate 
numerous uM poids species, in some 


hybrid 
Andersson [1998] also recognized 


cases linked by intermediate or presume 
specimens (e.g., 
several morphologically well-marked, geographically 
restricted Cinchona species among populations pre- 
viously included in other species, and listed interme- 
diate specimens 
White (1998) addressed this situation and de- 
scribed such widespread polymorphie taxa as ochlo- 
species, a term he defined as: *A very variable 
olymorphic) species, whose variation, though partly 
with ecology and geography, is of such 
complex pattern that it cannot be satisfactorily 


correlated 


ochlospecies, compiled a list of 10 traits that 
characterize and help recognize ochlospecies, and 
concluded that it is not always clear which are 
phenotypically plastic species and which result from 
the synthesis of inadequate data. 

Gaertnera ae as circumscribed by van 
67) has 
ospecies identified 
geographically 
acter states seem to be only partially correlated with 


Beusekom (19 several of the traits of 
nk; in particular, it is 
ha 


y 
and ecologically widespread; its 


geography; similar variants appear to be found in 
widely separated localities; and it has a long 


taxonomic synonymy. Characterizing van Beusekom’s 


Annals of the 
Missouri Botanical Garden 


G. vaginans as an ochlospecies suggests that it 
deserves reconsideration and that additional data are 
needed for this. Here, this is reviewed using molecular 
data incorporated with newly compiled morphological 
data based on additional collections made since 1968, 
and a different, more closely defined species concept 
(outlined in the Methods section). 

The molecular sequence data indicate that plants of 
van Beusekom's “Gaertnera vaginans” do not form a 
single clade, but instead are a polyphyletic group with 
members distributed among several of the clades 
within Gaertnera (Fig. 3). This supports the classifi- 
cation based on morphology, which separates van 
Beusekom’s “G. vaginans” into 12 species, each with 
more limited geographic distribution and morpholog- 


. alstonii M 
dioica, G. arenaria, G. belumutensis, G. capitulata, G. 


junghuhniana, G. kochummenii, G. longivaginalis, G. 


KEY TO SPECIES OF THE GAERTNERA VAGINANS COMPLEX 


paniculata, G. ramosa Ridl., G. sralensis, and G. 


vaginans. A key to the species recognized within this 
group is presented below 

The majority of these species are found in Southeast 
Asia, where van Beusekom’s study was focused; thus, 
the morphological variation that was problematic for 
this classification is indeed greatest in this region, 
although Gaertnera arenaria of Madagascar as cir- 
cumscribed here is also widely variable and perhaps 
represents an ochlospecies on a smaller scale 
(resolution of this is beyond the scope of the present 
work). Thus, in this case the addition of more data 
leads to the segregation of this polymorphic taxon into 
more numerous and more narrowly delimited species 
that are more comparable in their distinguishing 
features and variability to other species of Gaertnera, 
and thus more broadly te 
new classification. better elucidates some of the 
evolutionary radiation within Gaertnera. 


stable across the genus. This 


la. Stems and stipules pilosulose to hirtellous, tomentulose, and villous or stipules sometimes puberulent; plants dioecious. 
2a. 


aves elliptic-oblong to oblanceolate or elliptic; 


E 
> 


oecious. 


2b. iene elliptic to oblanceolate; inflorescence subglobose, sessile or with peduncles to 1.4 cm long; Born 
11. 


inflorescence corymbiform; peduncles 3-5 c 


G. da 


3 ee and stipules glabrous or puberulent, or stipules sometimes pilosulose in G. longivaginalis; plants distylous or 


Sa Inflorescences subglobose, sessile or with peduncles to 3.5 cm long. 


4a. Plants mind with pen to grayish green cast; continental Southeast Asia 


ong; id Southeast Asia 


64. G. sralensis 


1. G. belumutensis 


35. G. kochummenii 


4b. Plants dry ange cast. 
5a. Gn. Tae or with lobes to 0.3 mm lo 
Calyx lobes 1-3.5 m m long; continental ee Asi 
3b. d a e iform to pyramidal, sessile or with peduncle dios 


Ta. Sup ules drying membranous, with lobes 1 
out i 


co 
lan T b unisexual flowers rien appearing ees in ce aphanodioica). 
—2.5 mm long; inflorescences several-flowered, with ca. 5 to 
5 


7. G. ramosa. 


a 
Tb. mu drying chartaceous, with lobes 1—7 mm long; inflorescences many-flowered, with more than 


ers. 
8a. Corolla white, tube 6-9 mm long; Bor 


neo 
orolla pale green to bi ~ 2.5-5 mm long; continental Southeast Asia 


8b. 
6b. S distylous, with bisexual flow 
Stipules 


3. G. aphanodioica 
ES G. junghuhniana 


drying membranous, d lobes 1-8 mm long; corolla tube 4-6 mm long; West and Cent 
40. 


ral 
G. longivaginalis 


9b. S G chartaceous, with lobes 0.5-5 mm lon, 


rolla tube 3—4 mm long, externally densely puberulent or tomentose; stipule lobes 0 
m long; West ad Central Africa 50. 
la tube 4-16 mm long, externally glabrous, scabrous, or sparsely puberulent; stipule lobes 


10b. 
0.5-5 m 


m lon, 
lla. Corolla m 10-16 mm long; stipule lobes 2.5—5 mm long; Madagascar ... 2 
llb. Corolla tube 4-6 mm long; stipule lobes 0.5-2.5 mm long; Sri Lanka . . . . 


TAXONOMIC TREATMENT 


Gaertnera Lam., Tabl. 
13 Feb. 1792, 


1: 379, t. 167, 


non Cinta 


Encycl. 
nom. 


Gentianaceae; nor Gaertnera Retz., 


1791, nom. 


Sykesia Arn., Nova 


Fructesca DC. ex Mei 
TYP 


ENS 
. paniculata 


G. arenaria 
7. G. vaginans 


rej., Campanulaceae. TYPE: Gaertnera vaginata 
Lam 


Acta Phys.-Med. Acad. Caes. Leop.- 
Carol. Nat. Cur. 18: 351. 1836. une subg. Sykesia 
(Am.) Benth., J. Proc. Linn. Soc. 1 V LYRE: 
Sykesia on Arn., nom. illeg. viec pis da 
vaginans wc Merr. 

Vasc. Gen. 1: o 2: EDD 1840. 


is Pl. 
E: Fructesca mauritiana DC. e 


593 


Malcomber & Taylor 


Volume 96, Number 4 


2009 


Revision of Gaertnera 


supuidpa pewop 
"deqns supuidpa 4) SUDUIÍDA 9) ou Suoj wu Gp opyeounn = 10 [eoruoo *Kueur ou snoxqe[$  ojeaoqo 03 ondje ou snoyAysi(y 
supuidpa 
"deqns supuidpa 4) D?pmorupd 79 ou Suo[ wu q—c'c ooma) F powop “tre ou snoxqe[$ ondi 01 Suo[qo ou sno[A1su(q 
supuispa Zuo] powop oyeppouejqo 
‘dsqns SUDUISDA 4) — smpuidpvaisuo) `£) ou Suoj wu g—p ww q 03 dn ‘reour ‘AUBUI 10 MOT ou snomqels 01 ondre ou snoyAysi(] 
supuidpa 
"deqns supuidpa 4) DIADUDID 2) ou Suoj utut 97-01 eyeoun F [eoruoo “Área ou snoxqe[$ — Suo[qo o) ondje ou sno[A1su(q 
ounuynysunt osoqo[s 
dsqns supuna *) SISUDDAS “9 ou Suoj wu q—g Temaueun *&ugur 01 MOT ou snoqe[s ondio ou snor 
pupnnpmsunt eje[oooue[qo 
dsqns supuisDa 2) DSOUIDA `£) ou Suo[ uu Q[-G ooun F powop ‘moz ou snoxqe[2 -ondje ou Snoro 
ounuynysunt umouyun 
dsqns supuidpa 9) Tu2unumqooy 79) e[[ozo9 oxnper è quoururoid *oyeAo pewop “tre ou quoosoqnd ondio sok umouyuf) 
pupnnpmgsunt powop snorqe[2 eje[oooue[qo 
dsqns supuidoa 9 vubruynySunl 9 ou Zuo wu -e'g aoma y *Kueur 10 Moy ou Ayjensn — -oyepoooury ‘eoun ou snorooo0HT 
pupnnpmsunt eje[oooue[qo 
dsqns supuidpa *) popmmndvo “4 ou Suo[ uiu cg-c'7, jews *remSuern osoqo[$ ‘Aueur sok juoosoqnd 01 ondre ou SNOLDOI(] 
ounuynysunt umouyun 
dsqns supuidpa *) sisuagnunjag 75 e[[o102 ome ejeounn F osoqo[$ ‘Aueur ou snomqeys ondio sok SNODOI(] 
pupnnpmgsunt eje[ooouv[qo 
dsqns supuidpa -9 paioipounydn e) sok Suo[ wu 9-9 ejeounn F pewop “tre ou snoaqe[s -ondre ou SNOIDOI(] 
ounmuynysunt umouyun 
‘dsqns SUDUISDA 4) — muojs]o DISULIDA) e[[o102 ome jews *remS uer pewop *&ueur sok juo9soqnd Suo[qo sok snor 
(1961) (poooz) i [enbo vuisns puol adeys oqor odes pue jTuoumu1d ;snoxmqe[s 10 adeys feo] jMO[[94 wass 
woyosneg UBA xequioopemy pue sioyuy eoon xÁ]e) oouoosero[jur sSutM juoosoqud 10 o8uv1O Surpeoxg 
Iod somo Jo oN  e[ndng jeo] Surp wel 
SJUQUIJP91] OMUOUOXE J, 
Jequioo[e]y JO 597848 19I98.IBYO) 
TS UON (OQ) suvuidoa ni2uj1202) jo squourjeer] omuouoxe] Sunsenuo) ‘Z AJEL 


Annals of the 
Missouri Botanical Garden 


Gaerinera sect. Áetheonoma A. DC., Prodr. 9: 34. 1845. 
eonoma Meisn. ex Steud., Nemenek Bot., ed. 2, 
841 nud., pro syn. TYPE: Cocinera 


Bojer. 

Pristidia - Thwai aites, Enum. Pl. Zeyl. 2: 149. 1859. TYPE: 
Pristidia darieu Thwaites = Gaerinera divaricata 
(Thwaites) Thwaites 

Hymenocnemis Hook. Ea in Benth. € Hook. f., Gen. Pl. 2: 

32. 1873. TYPE: Hymenocnemis madagascariensis 
po m eee (Hook. f. 
Malcomber & A. P. D 
do ser. Densiflorae K. pum ., Nat. Pflanzenfam. 4(4): 
. 1891. TYPE: Gaertnera trass fora Bojer 


ceo 
a 
2 
3 
= 
= 
o 
ge 
= 
=< 
as) 
a 
2 
m dr 
a 
o 
un 
= 
ge 
3 
FJ 
= 
o 
à 
— 
= 
E 
E 
us 


series explicitly included the type species of the genus, 
nd thus is correctly called Gaertnera ser. Gaertnera.) 
Trees or shrubs, up to 15(-20) m tall, glabrous to 
variously pubescent, sometimes drying grayish green, 
brown, gray, chestnut, reddish brown, or wit 
distinctive orange cast. Branches flattened to terete 
or quadrangular, 0.5-15 mm diam.; internodes 0.1— 
26 cm, smooth or with 2 to 4 longitudinal ribs, 
sometimes corky (Gaertnera alata Bremek. ex Mal- 
comber & A. P. Davis). Leaves sessile to petiolate, 
decussate, paired or rarely ternate, DM to 
slightly anisophyllous; blade 0.5-55 X 0.1-20 em, 
linear-lanceolate to elliptic, elliptic-oblong, or cune- 
iform, margins flat or infrequently crisped; secondary 
veins visible or sometimes not visible abaxially, 
eucamptodromous, spreading and arched or infre- 
quently nearly straight and ascending at an acute 
angle; domatia absent or present as pilosulose or 
hirtellous, tuft- or crypt-domatia in axils of secondary 
veins; petioles 0.2-10 cm, smooth to occasionally 
furrowed or rarely winged, smooth or usually encircled 
on sides and below b es or wings, these often 
extending onto stipules. Stipules 0.1—75 mm, tubular 
and open at top in bud to fused at top and calyptrate, 
cylindrical to infrequently funnel-shaped, entire or 
splitting along 1 side to form a spathaceous structure, 
splitting on 2 sides to form 2 usually intrapetiolar 
segments, or rarely splitting on 4 sides into 4 spatulate 
parts (G. furcellata (Baill. ex Vatke) Maleomber & A. 
P. Davis, G. microphylla), drying membranous to 
coriaceous, variously caducous, deciduous through 
fragmentation, or persistent, usually with 4 longitudi- 
nal ribs or wings arising below petioles and extending 
along sheath to lobes, sometimes with 2 additional ribs 
or wings; apex entire or with the 1 to 4 incisions 
described above, marcesent (i.e., becoming hardened) 
to fragmenting or infrequently persistent; lobes none or 
4, up to 20 mm, deltate to filiform; additional setae 
none or several to numerous, to 20 mm. Inflorescences 
terminal on developed stems, sometimes borne on short 
axillary stems, or rarely axillary or terminal on supra- 
axillary stems, cymose to compound-cymose, paniculi- 


form, subcapitate, reduced to 1 or a few flowers, or 
rarely spiciform, erect to occasionally pendulous, 
sessile to pedunculate, green, white, or pink, bracteate 
or rarely ebracteate (G. alata); peduncle to 15 em; 
branched portion up to 25 X 30 cm, pyramidal to 
corymbiform-rounded, subglobose, or rarely cylindri- 
cal, branched to as many as 6 orders, lax to congested; 
axes dichasial or rarely scorpioid (G. divaricata); 
bracts subtending basalmost axes triangular to linear 
or sometimes resembling reduced leaves or stipules (G. 
microphylla), rarely grouped into an involucre (G. 
cuneifolia, G. microphylla); bracts subtending higher- 
order axes up to 30 mm, deltate to linear, ovate, or 
trifid, usually grading in size and shape between 
asalmost bracts and bracteoles; bracteoles reduced to 
developed, inserted on pedicel and/or at base of calyx, 
rarely enlarged and bright white (G. phyllosepala, G. 
phyllostachya). Supra-axillary reproductive branches 
rarely present (G. diversifolia, G. inflexa), paired, 
flexuous, with leaves well developed to reduced. 
Flowers sessile to pedicellate, 4- to 5(6)-merous, 
dioecious or bisexual and heterodistylous. Calyx limb 
cup-shaped, urceolate, or campanulate, 1— 
wide, outside glabrous or variously pubescent, inside 
sometimes with a distinct ring of trichomes, truncate to 
denticulate or lobed, lobes rounded to triangular or 
linear, equal to markedly uneq 3 


sometimes petaloid in color and texture (G. phyllose- 


ual, up to 30 mm, 


pala, G. phyllostachya); corolla white or infrequently 
ue, clavate to rhomboidal or 
nfundibuli- 


form, or campanulate, 2-30 mm, pass glabrous or 


pink, red, orange, or 
obclavate in bud, at anthesis sa 


variously pubescent, inside glabrous or villous inside 
tube, tube entire or sed fenestrate (G. ae p 
vate-oblon 


brous adaxially, apically pre to acute or a 


—10 mm, narrowly triangular to o 
expanded and cucullate (G. cooperi), abaxially smooth 
or rarely with a thickened subapical appendage (G. 
edentata); stamens inserted in corolla tube, 


E 


.5-4 mm, narrowly oblong, dithecal, purius by 
longitudinal slits, sessile or with developed filaments 
to 8m tyled flowers positioned below the 
stigmas and included to 


mm, in long-st 
partially exserted, in short- 
styled flowers positioned above the stigmas and 
included to exserted, in staminate flowers positioned 
near middle of corolla tube to exserted, in pistillate 
flowers staminodes present or absent, positioned in 
lower part of corolla to corolla throat. Ovary superior, 
2-locular, 2-celled, with 1 erect basal ovule in each 
cell; style filiform, glabrous or pubescent, stigmas 2, 
linear-clavate, often flattened, in long-styled flowers 
these positioned above the anthers and exserted, in 
short-styled flowers these positioned below the anthers 
and included, in staminate flowers pistillodes devel- 
oped or reduced and positioned in lower part to middle 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


of corolla tube, in pistillate flowers these developed 
and Positioned i in lower to upper part of corolla tube. 
drupe, blue to violet-black or reportedly 
sometimes whitened, globose to ellipsoid, obovoid, or 
didymous, smooth or infrequently ridged, 5-28 X 5- 
mm, glabrous or rarely pubescent; pyrenes 2 per 
drupe or sometimes 1 apparently by abortion or 
incomplete | ad spherical or hemispherical to 
edge-shaped, = smooth to finely fissured, rugose, 
and/or deeply litur endosperm entire to invaginat- 
ed or ruminated. 


REGIONAL KEYS TO GAERTNERA SPECIES 


KEY TO GAERTNERA SPECIES IN AFRICA 


la. Inflorescence spiciform to narrowly pyramidal, unbranched or branched to 


Distribution. Sixty-nine species and one presumed 
n evergreen moist s rud in forest 

Om in Africa (Senegal, Sierra 
Liberia, Cóte pee 
Burkina Faso, Ghana, Togo, Nigeria, Cameroon, Central 


i 
understory, at 0—200 
Leone, Guinea, Benin, Mal 


African Republic, Democratic Republic of the Congo, 
Republic ofthe Congo, Equatorial Guinea, Gabon, Angola, 


a o the Mascarene Islands (Mauritius, 


(Tha iland, Cam bodia, Vietnam, Malaysi 


(Indonesia), and Bonen (Brunei, Malaysia, Indonesia). 


to 2 orders with axes short and 


congested; or sessile to subsessile; corolla red to orange-red outside; : lobes 5-15 mm long; c in 


coastal for 


= 
> 


a. Calyx limb lobed, lobes 0.8-4.5 mm long; stems pilosulóse and/or hirtellous at least when young 
3b. Calyx limb truncate to shortly lobed, lobes to 0.6 mm É 
or striations; stipules glabrous; inflorescence corymbiform, with 

l: 


4a. Bark with longitudinal fissure 
d ie) portion 3-5.5 X 3-7 cm 
4b. Bark ap 


re ce stipules pilosulose; 1 


3. G. spicata 


Mie cymose to paniculiform, pyramidal, or corymbiform, with axes and/or pedicels developed, Ware to 


8. G. bieleri p-p- 
long; stems puberulent to glabrous 


G. eketensis 
pyramidal, with FS portion 


ings not extending below t 


8. G. liberiensis 


into the petiole 


6a. Pedicels 1-9 mm long; inflorescences deflexed to pendulous; with axes mostly Pr at ca. 
d 


cu "n 


6. G. shed p-p- 


6b. Pedicels 0-1 mm long; inflorescences ascending, with axes mostly ascending. 


Ta. Corolla tube 4—6 mm long, lobes 
7b. Corolla tube 2 


gl 
> 


encircling petiole base. 
8a 


Corolla 


ed s evident abaxiall 


Corolla 


co 
> 


flatten a at apex or with an adaxial hook or flange, in bud the 
0.8-13.5 cm, with the tertiary venation not evident abaxially or evident and 


2.5-5 mm 2: stipules drying mod y 
-2.5 m 0. G. paniculat 
. Stipules with wings or (is extending below le as a pee, well- "dovslened 3 M or flang 


—4 mm long, lobes 1.5 


. 40. E o 


tube 8-11 mm long, lobes 3.5-6 mm long, the lobes inflated and llate at in bud the 
lobes me an urceolate cap; leaves 9-25 X 


3-10 cm, with the tertiary venation usually regularly 
. G. cooperi 


ong, lobes 0.8-5 mm long (corolla a Hnnc in G. letouzeyi), the lobes 


moothly tapered, acute 


16-33 X 7.5-13. E] cm, consisently large, with petioles rather thickened. 
10a. 2 limb densely puberulent to pilosulose outside; inflorescences a a X 4- 
8c 


10b. Cale limb densely puberulent outside; inflorescences 7-23 X 6.5-2 
Ta cm, at least some of the leaves smaller than in the 1 lead, with 


9b. Leaves 2.2223 X 
petioles generally slen 


s gabonensis 
6. G. letouzeyi 


lla. Calyx limb truncate or with lobes to 0.4 mm long. 


12a. Inflorescences, F and pedicels generally ascending; corolla tube 1.8-3.3 mm 
pedic m lon, 


m long; pedicels 0-2 m 


ong, lobes 1 


5. G. aurea 


12b. Inflorescences eee to pendulous, axes and pedicels spreading to ca. 90 


oM corolla tube 2.4—4.5 mm long, lobes 2.5—4 mm long; a am 
lon 


6. G. Boh p-p- 


llb. Calyx limb lobed, lobes 0.8-4.5 mm long. 


13a. Stems P or hirtellous 


13b. Stems glabrou 
KEY TO GAERTNERA SPECIES IN ASIA (INCLUDING SRI LANKA) 


la. Leaves ternate, 0.1—0.3 cm wide; Sri Lan! 


Lue AA 8. G. bieleri p.p. 


a iodo ota E 37. G. leucothyrsa 


eee Rea ete m ae 65. G. ternifolia 


lb. Leaves opposite and/or ternate, 0.4—15 cm wide; widespread. 


Annals of the 
Missouri Botanical Garden 


2a. Stipules, young stems, m undersides of leaves moderately to densely hirtellous, pilosulose, hispid, 


b 


velutinous, and/or 
3a. Stipules 


eo 


with Sla narrow wings encircling base of petiole; peduncles 3-5.5 cm long; inflorescences 

with branched portion rather lax, corymbiform, 5-12.5 X 5-10 cm. 

4a. ay wings oe below petioles and extending into ribs along the stipule tube; NC 
2-5.emwide;:Bornéó soras pare Rens Get baer a genet s A E Red HRS COSA Erbe d 2. G. alstonii 

4b. si wings see to area near petiole, the stipule tube generally smooth at least in pe 

4—12 cm wide; Singapore and nearby islands .................... 7. G. gne 


; leaves 
: Stipules p relatively broad wings encircling base of petiole; peduncles none or up to 2.5 cm "on 


inflorescences with branched portion congested-cymose to subcapitate, subglobose, 1.2-3 X 1.2-3 c 
5a. Flowers 4-merous; calyx limb truncate or with lobes up to 1 mm long; corolla tube 7.8-8. = mm p 
B 


ud C" "EET 1. G. en ila 
5b. Flowers 5-merous; oe limb lobed, lobes 0.5—4 mm long; corolla tube 4—5 mm pets continental 
Southeast pud at ii e uS uae dede ede stai eme s too aug . G. schizocalyx 
. Stipules, young stems, and teils of leaves glabrous to puberulent, or stems sparsely hirtellous to 


pilosulose becoming aes cent. 


a 
zd 


PE pea A on well-developed axillary or supra-axillary reproductive branches with usually 

reduced lea 

Ta. Plants eee calyx 1-2 mm wide; fruits 14-17 mm long; Sri Lanka ....... 7. G. divaricata p.p. 

Tb. cs ts dioecious; calyx 1.5—3.5 mm wide; fruits 7-8 mm long; continental South Asia 
—— ————— — AEA Ble nC eee TR diversifolia 

eL erect to pendulous on principal and/or axillary vegetative branches, or uer on leales 

supra-axillary peduncles. 

8a. Stipules with tubular portion 20-75 mm long, drying chartaceous to leathery, funnelform (i.e. Pos 


part spreading); leaves 20-55 X 7-19 cm, with well-developed cartilaginous margins 7. G. obesa 
8b. Stipules with tubular portion 127 long, drying membranous to n cylindr: ET ris 
ar, not spreading at apex); leaves 1-31 X 0.4-11 cm, with margins not to thinly cartilaginous. 


10a. Inflorescences lax, with axes nee Hu at basal nodes then markedly ae at more 
distal nd i 


nodon — ary leaf veins flat a ardly visible abaxially ..... 1. G. divaricata 
lOb. I ted to saints with axes reduced or all dichasial; pe 
nu promo abaxial yss sara ca eae nat anaes ada 48. G. oblanceolata 


¡al ly 
9h. Inflorescences terminal on principal and/or aion branches, or sometimes pseudoaxillary. 
nfl few-flowered (i.e., with 1 to 4 flowers), unbranched or branched to 1 order; 
nes 1 s 4x 04-4 em; Sri Lanka 
12a. Pai lobes 1.5-5.5 mm long; iie sessile or subsessile, with pedicels to 


m long; Eroll t a 12-239: mm longs s sns see ted 59. G. rosea 
12b. Sip lobes 0.1—1 ng; flow il ith pedicel 2 mm long; corolla 
e 8-12 m 
n Site tube 3i 2-3 mm long, with narrow wings encircling petioles and 
xtending onto s calyx lobes 0.8-2 mm long ......... 5. G. Xgardneri 
13b. Sie tubes 3-10 mm long, without wings around petioles or on tube; ale 
0.3-0.5 mm long... 2.2 ee eee 70. G. Mr p-p- 


llb. Inflorescences aN o many-flowered (i.e., with 5 to numerous flowers), subcapitate or 
branched to 1 to 6 orders; leaves 1.5-25 X 0. 4—9 cm; widespread. 
.. subcapitate to congested-cymose, in shap e subglobose to corymbi- 
form, m or with peduncle to 3.5 cm long, br i portion 0.5-3.5 X 
3. 


5 em; es oblanceolate to elliptic or lanceolate, usually narrowly so. 
15a. Plants drying with orange cast. 
l6a. Calyx e or with lobes to 0. M mm long; Asia G. belumutensis 


Calyx lobes 1-3.5 m Ld ve 
15b. Plants drying green or with gray 
Stipule ribs narrowly ed DT CH 0.5-1.5 X 0.5- n 
"TED ERR 4 C sralensis 
l7b. Stipule ribs broadly winged; inflorescences 1-3.5 X Tos cm. 
Leaves 4.3—20 cm long; stipule lobes 7-10 mm lon; 
verde E uova tae edat eel cuo Dads due 11. G. capitulata p.p. 
18b. Leaves 15-25 cm long; stipules truncate or with lobes up 


ebd nad dee DE 5. G. kochummenii 


3mm long less aaa pei. 26. G. globigera 
14b. Inflorescences rather to very laxly cymose, in shape corymbiform to pyramidal, 
sessile or usually at least some inflorescences with peduncles 0.9-7.5 cm long, 
branched portion 0.8-24  1.5-22 cm; leaves narrowly to broadly da. 
lanceolate, ovate, obovate, elliptic-oblong, or oblanceolate. 
9, orescences with axes dichasial 2 lowest nodes then markedly scorpioid at 
more distal nodes; eaves with secondary veins flat and rd not evident 
abaxiallysisn Lankas iia n E 1. G. divaricaia p.p. 


Volume 96, Number 4 Malcomber & Taylor 597 


2009 


Revision of Gaertnera 


19b. Inflorescences with axes regularly SS. raa with secondary veins 
evident and usually M E widesp 
20a. Flowers 4-merous; Peninsular Malaysia, bod: Singa 


gapor 
la. rim pyramidal, deflexed to pendulous, «n branched 
portion 3-11 X 3-9 cm; stipule lobes ca. 3 mm long ..... 
supe E TU S AE E EA DUREE 2. G. fractiflexa 
21b. Inflor d i 


erect, with branched portion 
0.8—7 X 2.5-6 cm; y lobes 0.5-1.5 mm long ... 69. G. viminea 
20b. Flowers 5-merous; widespread. 
2a. Plants diss) lata, xu one eee Sri Lanka. 


23a. Corolla tubes 4 m long; inflorescences many-flow 
ed, with bran eis portion B 5 24 X 1.1-17 cm Ton 
3-20 X L5-9 cm... ied 61. G. vaginans 


23b. Corolla tubes 8.5-12 mm long; inflorescences several- to 
few-flowered, with branched portion 1.5-5 X 1.5- 


4.5 em; leaves 1.5-9.4 X 0.4—4 cm. ... 70. G. walkeri p.p. 
22b. dabo e with flowers unisexual (ofi en apparently 
homostylous in G. aphanodioica); continental Southeast Asia, 


Sumatra, Home eo. 
24a. Stipules drying membranous, with tube 7-15 mm long 
and lobes 1-2.5 mm long; leaves 3-14.5 X 0.8-5 cm; 
lla Malaysia? 57. 
24b. Stipules drying chart with tube 3 zs mm uus and 
lobes 1—7 mm long; leaves 3: 5-24 X 
25a. Corolla white, with tube 6— em ln süpüle 
lobes 1—4 mm long; growing at 12—60 m above sea 
velan;Boreo--« ue eh rete 3. G. ap ns 
Corolla pale green to white, with tube 2. Sis m 


25b. 
s a lobes 1.5—7 mm long; growing at o- 
1500 m, Thailand through Peninsular Malays 
and in ae and Borneo .... 34. G. ee did 


KEY TO GAERTNERA SPECIES IN MADAGASCAR AND THE MASCARENE ISLANDS 


la. WEE, persistent as four "A to linear segments, these often overlapping due to shortened stem internodes; 
s 0.3-1.5 X 0.3-0.8 c 


E 


leav 
2a, 


2b. 
b. Stip 


Leaves elliptic; cale mum 0.5— d 5 mm long; corolla tube 2.5-3 mm lon 


[MN bi 3. G. furcellata 


li 


i A hee deciduous often fragmentation, or persistent, spathifo 
seine into 2 segments, overlapping or separated by well-developed stem internodes; leaves 0.3—51 X 0. 3-14. 5 em. 


bovate; calyx lobes 2.4—3 mm long; corolla tube 4—4.5 mm lon; i G. microphylla 


3a. Young stems, stipules, and leaves densely ed or hispid to pilose, hirtellous, pilosulose, velutinous, or 
to with spreading, very evident pubescence 
4a. At least the longest calyx lobes 3-14 mm is (do not confuse these with bracts 
5a. En lobes linear n narrowly triangular, pale green to green; secondary leaf veins 6 to 1 
pules with tube 11-30 mm long .......... nannaa anaana 53. G. Pare a p-p- 
5b. aiz lobes narrowly lanceolate to elliptic or ovate, white; secondary leaf veins 9 to 12 D stipules 
tübe 8-21 mmm long. cor xke Ip 9E eme NEP EPIDEI IS IS 4. G. phyllosepala 
4b. Calyx truncate or with lobes to 2.5 mm lon 
6a. a with tube 22-68 mm long; Padi tube 10-13 mm long; plants drying with an ues 
VE rures. opere EN REVIEW IR NUS DET NINIAROSUNEIDUTQISERTS 1. G. schatzii 
6b. Salles with tube 7-40 mm long; corolla tube 3-11 mm long; plants not drying as above 
7a. Corolla pale pink to le calyx lobes 0.5-2.5 mm long; pubescence prm gray- "white A 
AE AA IG E DES EE EE ea E E EE EEEE 32. G. ianthina 
Tb. Corolla white; calyx truncate or with lobes to 1.5 mm long; pubescence drying reddish brown to 
brown or gray 
8a. Leaves L 13.1 cm wide, with apex acute to shortly acuminate; corolla tube 6. ie es 
lonp leti cempore tes hopes Nae hati wee ens eee ced Senet 0. G. hispida 
8b. bes 1.9-8.5 cm wide, with apex acute to rounded then sometimes rand very cum 
uspidate; corolla tube 3-6 mm long ............ obo ar. sphaerocarpa p.p. 
3b. Young stems, las and leaves glabrous or sparsely to dus ely puberulent, villosulous, DE HA strigose, or 


sericeous with pubescence nest appressed and not strongly evident. 
9 


Flowers solitary or 2 to 4 and fasciculate to shortly cymose; leaves 0.3—7 X 0.3-3.2 cm. 
10a. Stem internodes longitudinally ridged, ribbed, or winged; leaves 0.3—3.8 X 0.3-2 cm. 
lla. Flowers pedicellate, with pedicels 2.5-19 mm long; corolla tube 4—7.5 mm long. 
1 Calyx lobes 0.4—4 mm long; stems 2-winged with corky bar 
12b. Calyx truncate or with lobes 
non-corky bark 


E Sei ee eee tie 2 alata. 
to 0.4 mm long; stems with rounded ridges and ow S 
51. 


se 
"m c rM os p-p- 


Annals of the 
Missouri Botanical Garden 


llb. Flowers sessile or subsessile, with pedicels to 1 mm long; corolla tube 2.5—5. > mm lon; 
Stipules caducous; caly lobes 1.1-42 mm long .............. 9. G. herpes 
13b. Stipules persistent or spl segments; calyx lobes 0.5-2. a mm long 
3. G. PRU DN p.p. 


10b. ram with internodes smooth; leaves 0.5—7 X 0.3-3.2 cm. 
Calyx limb truncate or with lobes to 0.4 mm long; stems glabrou 
p Poe tubular, with 3 to 4 filiform lobes; inflorescences see or with S s 9- 
6. G. 


milona: 5 4.0 sta aise ia Pha acere Ap tnc det a e freed a bambusifolia 
15b. Sigal cali with 2 or 4 triangular to filiform lobes; inflorescences Ine 
with pedunc Au PA 9 T3anr OnE soe gst ae E E a Se 51. G. pauciflora p.p. 
14b. ig With lobes 0 m long: stems glabrous to Dl villosulous, stripe, or sericeous. 
Stipules ie pus 4-6.5 cm long; stems glabrous ......... 16. G. darcyana p.p. 
To Sd les EAE ho 0.5—4 cm long; stems deser .... 44. G. madagascariensis 
9b. Flowers several to numerous, 5 o in heads or cymes; leaves 3-51 X 1.4—14.6 cm. 
l7a. Calyx lobes 1.5-15 mm e at ie t some lobes 1.5 mm long or longer. 


18a. Calyx lobes elliptic to narrowly elliptic or elliptic-oblong, markedly narrowed at base, 6-15 m 


long; flowers mostly or all subtended by 2 white bracte oles 8-16 mm long . 55. G. pita 
18b. Calyx lobes triangular to ligulate, eee near, elliptic-oblong, ovate, or lanceolat 
broadest or a little narrowed at base, 1.5-8 mm pes bracteoles when present green to Sie 
green and 8 mi r shorte 
19a. Calyx funnelform to eens or tubular, 7-20 mm wide at 
20a. Calyx lobes ligulate to linear, 6-8 mm long; Bo elliptic elo to obovate o 
cuneiform, 35.5 X 1.44 cm; flowers subsessile ............ -G cuneifolia 
20b. Calyx lobes broadly triangular to ligulate, 1.5-5 mm long; leaves elliptic to 


oblanceolate, elliptic-oblong, or öboväte, 8-17.5 X 3.5-9 cm; flowers 


21b. Inflorescence axes and leaves pubescent n... nananana anana 
aertnera m A of Verdcourt (see discussion under G. calycina) 
th. 


19b. Calyx limb tubular to campanulate, 1—4.2 mm wide at mout 
22a. Flowering stems and peduncles flexuous, with inflorescence pendulous; corollas 
'wintesto pale blus seumas nk aad eye nand pr er the eed es 2. G. pendula 
22b. Flowering stems and peduncles not flexuous, with inflorescences ascending; 
corollas w 
23a. 3 to 9 in lax fascicl 0.8-2 X 0.4-2.5 cm ... 16. G. darcyana p.p. 


23b. Flowers 10 or more in lax cymes to subcapitate heads 0.7-19 X 1-20 cm 
24a. Calyx lobes narrowly spatulate to elliptic-oblong, unequal on an 
individual flower, 5-7 mm long with at least some lobes more than 
madonna bres pita le Pase ec iota 31. G. humblotii 
Calyx Num broadly triangular or ovate to linear or narrowly 
spatulate, 1.5-5 mm long. 
25a. Leaves broadly clipi to obovate or ovate, drying thickly 
coriaceous, 1.5-1 0.8-5.5 cm; inflorescences subcapitate 
to congested c vocis cic elu Ei eR 0. G. de o p-p- 
25b. Leaves elliptic to elliptic-oblong, oblanceolate, or ob 
drying chartaceous to coriaceous, 2.5-24 X 1-145 e i5 


N 
NM 
> 


ose. 
calyptrate, caducous or fragmenting EUN 
ll- oe mm lone drying pend anous 
27a. E E itate t 
subsessile or with peduncle to 4.3 cm MEE suh 
branched portion 0.74.5 cm long ......... 
53. G. phanerophlebia p.p. 
il 


27b. T ET ] 1 
ed, pedunculate with peduncle 0.54 cm long, 
with branched portion 1.4-7.5 cm long ...... 
A id 58. G. raphaelii 
26b. Stipules ine usually regularly persistent at least 
on distalmost nodes or slowly deciduous by fragmen- 
x» 


tation, with tube 4—33 mm long, drying chartaceous. 
Stipules 10-33 mm long, cylindrical or usually 
inflated to O ... 43. G. macrostipula p.p. 
28b. Stipules 4-12 mm long, cylindrical. 
2 Corolla ms ca. 13 mm long ...... 


A ee ee 29. G. hirtiflora p.p. 
29b. Corolla tube 16-25 mm long ...... 
68 


Volume 96, Number 4 Malcomber & Taylor 599 


2009 


Revision of Gaertnera 


17b. ma loben all less: than 1.5 mm long. 
30a lin 


> 


drical to f lfi ed, persistent at least on distal ] nodes, 


inflat 
with tubes 10— 80 mm long with at least some stipules 25 mm long or longer. 
3la. Inflorescences pendulous, with branched portion pyramidal ........... Lus. 
31b. Inflorescences generally pane ith branched ned rounded-corymbiform. 
32a. Stipules entire (1.e., not cleft); corolla white anke sss eue 43. G. macrostipula 
Stipules deeply dei i into two segments; EIE jer (arta 4i 
Stipules calyptrate or tubular, cylindrical and rather uw aa the stem or caducous, with 
tubes 2-50 mm long, if 25 mm long or longer then calyp 
3. Plants generally slender; inflorescences Hr or UE to pendulous, subsessile or on 
stiff to flexuous ae or stems, sometimes with supra-axillary reproductive branches; 
corolla tubes 3-8 mm lon, 
34a. Infl } a to shortly pedunculate, peduncle to 2.7 cm long, without aise 
ondary axes or these 1 to 2 pairs mi up to 1 cm s flowers 5-merous ...... akeana 
34b. Dd nces pedunculate, peduncle 1.5-6 cm long, with 2 to 5 pairs of Bd 
secondary axes, these 0.5-2 cm long; flowe 
35a. Stipules with tubes 4—10 mm long; pes aly elliptic- uaa to elliptic- 
anceolate or 


41. G. lowryi 


2. G. cardiocarpa 


lan e VOLS "m 
35b. Stipules D tubes 2.3—4.5 mm long; leaves narrowly elliptic to v or 
oblanceolate: a «uem teach o aS ean Ba oe ate ed eed G. inflexa 
33b. Plants generally rol infl ascending on generally straight peduncles or vegetative 


stems; corolla tubes 3-32 mm lon 
a ma ules tubular and with At to 10 setae at apex in addition to 4 linear lobes 
Calyx truncate or with lobes to 0.6 mm long; stipules with n poriio n 2— 
y 0. G. edentaia p.p. 
37b. Calyx lobed, with lobes 1-2.5 mm long; s m with tubular rp 5-12 
long; corolla densely puberulent to velutinous externally ..... © hirtiflora p-p- 
36b. Stipules calyptrate or tubular, without setae, with lo oe none or 1 to 4. 
St D tubular, tubular portion 2-5 mm long. (This measurement arbitrary, 
with overlap als so lis ted in next coupler 
I m wide, with 1 to 2 pairs of developed 
secondary lla tube 10-25 mm long ............... 20. - edentata p.p. 
39b. Inflorescences with branched portion 3-15 cm wide, with 2 to 5 pairs of 
developed secondary axes; corolla tube 10-14 mm oe p dinate 
——É Em 6. G. e iE p-p- 
38b. Stipules tubular or calyptrate, with tubular Lo more a 5 mm long. 
40. rescences subcapitate to congested-cymose, without developed 
secondary axes or with 1 pair shortly dene and unbranched ...... 
Wc dE ENT THEE 60. G. r ea p-p- 
40b. era branched with 2 or more pairs of developed, branched 
secon: 


4la. Si, A tubular and regularly persistent or sometimes fragmenting 
when old. 
42a. Stipules with tubular portion 8-17 mm long; fruit subglobose, 
-11 X 7.5-9 mm ccoo 4. G. arenaria p.p. 
42b. Ee wath tubular portion 2-12 mm long; fruit ellipsoid, 
G Bam | ere eee ee 56. G. p d ioides p.p. 
Alb ag ee or a caducous or quickly foeni 
la with tube 17-32 mm long and lobes 8-12 mm Tong 
Ron Corolla tube 7-9 mm diam. Sab qu eU 4. G. crassiflora 
Corolla tube 3-4 mm diam. ......... P G. longifolia 


44b. 
43b. ee with tube 3-16 mm long and lobes 2-5.5 mm long. 
Corolla with tube 10-16 mm long € lobes 3.5- 
Dom Tongr Paris Sad a eens ye te 4. G. arenaria p.p. 
Corolla E tube 3-10 mm long and lobes 2-4 mm lon 
A d n 10 mm = then lobes 2-3 mm long. 
with base rounded to truncate and 
sonny veins Bt to 10 pairs; corolla tube 
6.5-10 mm long .......... 42. G. macrobotrys 
46b. Leaves with base acute to obtuse and second- 
ary veins 5 to 16 pairs; corolla tube 3-9 mm lon 
47a. Stipuleswith tubular rportion 11—49 eee 
ius white; in littoral forests on d sand 
LIDSETALES arase cas e N 8. G. Er 
47b. Stipules with tubular portion 8-32 mm long 
corolla white to pink or lilac; in vari- 
o habitats on a substra 


ÉS 
e 
c 


9. G. Eden p.p. 


Annals of the 
Missouri Botanical Garden 


1. Gaertnera alata Bremek. ex Malcomber & A. P. 
Bot. Missouri Bot. Gard. 
104: 379, fig E: Mada pod 
Toamasina: "e et d'Andisibe, bassin de l'Oni 
19*50'S, 47°51’E, Feb. 1925, H. Perrier de p 
Báthie 17098 (holotype, P!). 


Shrubs 


diam., glabrous, usually + 


Davis, Monogr. Syst. 


1-2 m tall; branchlets terete, 0.5-2 mm 
2-winged, + corky, pale 
beige to gray or whitish; internodes 0.3—2.7 cm, with 2 
Leaf blades 0.6-3.6 X 0.3- 


or 3 longitudinal rib 
1 o P oblong or = elliptic- 


8 cm, elliptic 
obovate, apex acute to shortly abruptly acuminate, 
base cuneate to rounded, drying chartaceous, gla- 
brous; seconda ns prominulous abaxially, 3 to 
5(to 8) pairs; do present; petioles 0.9— 
Stipules calyptrate, drying membranous, ie. 
caducous or persisting on a few vw tube 2- 
ed, 4 a 


petioles, 1 additional rib on each uc e side, all 


m, with ribs 6, narrowly wing g below 
these extending to apex, apex with 2 clefts, marces- 
lobes 2, 0.7-1.2 mm, deltate. Inflorescences 
terminal o 


cent, 
on principal and/or axillary branches, 
flowered or 2- to 3(4)-flowered and sells le, 
sessile, glabrous or puberulent, ebracteolate; pedicels 
2.5-19 
distylous. Long-styled flowers: unknown. i dis 
flowers: c 6m 


wide, glabrous except inside with sparse 2 


mm. Flowers 4-merous, presumably hetero- 


alyx cup-shaped to + urceolate, 

lobes 0.4—4 mm, triangular to narrowly triangular or 
pink, clavate in ps at 
, tube 5-6 mm, 
lobes 2.6- 


linear; corolla pale pink to 
sis salverform, outside glabrous 
1.82.2 incide. 


mm diam. glabrous 


.l mm, triangular to narrowly triangular, acute; 
anthers included, filaments inserted just below middle 
E 5 ; le 2.7— m, 


glabrous, stigma ca. 0.8 mm. Drupes unknown. 


Distribution and habitat. 
Madagascar, in the provinces of Antananarivo and 


This species grows in 


Toamasina, where it is known from humid evergreen 
escarpment forest south of Tsinjoarivo and southeast 
of Ambatolampy, at elevations of 1400-1550 m. Here, 
it grows on rocks in zones of metamorphic and igneous 
basement rock. 


Phenology. This species has been collected with 
flowers in the months of January and Februa 


Discussion. Gaertnera alata can be recognized by 
its combination of few-flowered or solitary-flowered 
inflorescences, 2-winged rather corky branchlets, 
pubescent leaf domatia, calyptrate 6-ribbed stipules, 
relatively long pedicels, and relatively large pink 
flowers. The measurement for the corolla width given 
is the width 


in the protologue, 4.2-12 mm wide, 


across the corolla lobes at anthesis. This species is 


of G. 
microphylla, and perhaps closely related to that 
This 


ar 
collections, and Malcomber and Davis (2005) provi- 


similar in general aspect to some plants 


species. species is so nown from few 
sionally considered the conservation status of this 


species to be Endangered (IUCN, 2001) based 


Antananarivo Province by Malco 
a an (2005), but this ledio is actually in 


Toamasina Province as cited here. 


Additional specimens examined. MADAGASCAR. Anta- 
nanarivo: 16.2 km SE de Tsinjoarivo, le long de la riviére 
d'Andrindrimbola, Messmer & Andriatsiferana NM 690 (G, 
K), NM 692 (G, K). 

2. Gaertnera alstonii Malcomber, sp. nov. TYPE: 
Malaysia. Sabah: Tongod, Kinabatangan, Melian 
Basin, Gunong Lotung, 830 m, 30 Mar. 1982, G. 
Amin SAN 95125 (holotype, SAN!; isotypes, K!, 
L!, SAR). 


Haec species Gaerinerae junghuhnianae Miq. similis, sed 
ab ea foliis oblongis atque stipulis pubescentibus in quoque 
latere tubi ac sub petiolo alis duabus longitudinalibus 
prominentibus munitis distinguitur. 

Trees, 6-12 m tall; branches terete to quadrangular, 
when young pilosulose with indumentum drying 
yellow, becoming glabrescent, 2-5 mm diam.; inter- 
nodes 3.2-5 em, smooth. Leaf blades 6-15 X 2.2- 
5 em, elliptic-oblong or oblanceolate to elliptic, apex 
shortly cuspidate or acuminate, base cuneate to 
obtuse, drying chartaceous, adaxially glabrous, abaxi- 
ally glabrous except pilosulose on principal veins with 
indumentum drying yellow to brown; secondary veins 
visible abaxially, 8 to 12 pairs; domatia usually 
present; petioles 3-10 mm. Stipules tubular, pilosu- 
lose, drying membranous, deciduous through frag- 
mentation, tube ca. 20 mm, with ribs 4, narrowly 
winged, arising and ie developed below petiole, 

apex entire or p 2 incisions, marcescent, 
lobes 4, gw to linear. Inflo 


m, branched to 4 to 5 orders, lax; bracts 1-6 mm; 
. Flowers 5- 
uo to a. pedicellate. 


s reduced; pedicels 1-3. 
merous, biology unknown, 
Calyx cup-shaped, 2-3 mm wide, outside pilosulose, 
inside glabrous, truncate or with lobes to 0.4 mm long, 
cape orolla, anthers, and stigmas unknown. 
rupes violet-black, eibilobsase or didymous, 5-7 X 
5-6 mm; pyrenes spherical or heumibeniesl rugose, 
finely fissured, endosperm entire. 


Distribution and habitat. 
in Southeast Asia, where it is known from Borneo, 


Gaertnera alstonii grows 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


5mm 


stipules. —C. Pist E 


Figure 5. AF. Gaertnera aphanodioica Malcomber. —A. Flowering bran 
late fl —D. Pistillate fl in cross section. —E. 5 
section. G, H. Gaerinera ianthina Malcomber. —G. Fruiting 


stem apex. C—F to same 5-mm scale. A-F based on Malcomber 2995; G, H based on 


specifically in its Kalimantan (Indonesia) and Sabah 
(Malaysia) sections. Here, it has been found in wet 
forests, at elevations of 50-275 


Phenology. This species has been collected in 
fruit in January and March through December. 


Discussion. This is a poorly known species 
represented only by fruiting collections, but it is 
easily recognized by its often elliptic-oblong leaves, 
dense short pubescence, and prominent longitudinal 
wings on the stipule tube that extend downward and 
encircle the petiole. Gaertnera alstonii is named in 
honor of Arthur Hugh Garfit Alston (1902-1958) who 
made the first collections of this species. This species 
belongs to the G. vaginans complex; see also the 
discussion of that group for related species and their 
distinctions. 


Paratypes. INDONESIA. Kalimantan: Kwala Kwajan, 
Permantang, Alston 13252 (A, L). MALAYSIA. Sabah: 
Kalabakan Forest Reserve, Sumbing SAN 101357 (SAN); 
Tawau, Luasong Camp, Madani SAN 107944 (K, KEP, L, 
SAN). 


3. Gaertnera aphanodioica Malcomber, sp. nov. 
Gaert 


nera vaginans subsp. junghuhniana f. 


hermaphroditica Beusekom, Blumea 15(2): 390, 
fig. 4E. 1967. TYPE: Brunei Darussalam. Belait: 
Seria, 18 Apr. 1957, S. Smythies, G. Wood & P. 
Ashton S 5909 (holotype, L!; isotypes, BRUN!, 
K!, KEP!, SING). Figure 5A—F. 


Haec species Gaerinerae junghuhnianae Miq. similis, sed 
ab ea tubo corollino 6-8 mm longo atque floribus ut videtur 
homostylis autem in effectu dioicis distinguitur. 

Trees or shrubs, 2-12 m tall; branches terete or 
quadrangular, glabrous, 2— i internodes 
1.5-9 cm, smooth. Leaf blades 6-24 X 2-9 cm, 
shortly 
acuminate or acute, base cuneate, drying chartaceous, 


elliptic to oblanceolate or ovate, apex 
glabrous; secondary veins prominulous abaxially, 3 to 
10 pairs; domatia usually present, hirtellous; petioles 
5-35 mm. Stipules tubular, glabrous, drying charta- 
ceous, caducous or deciduous through fragmentation, 
tube 10-25 mm, with ribs 4, narrowly winged, arising 
below petiole and sometimes extending to lobes, apex 
entire or usually with 2 incisions, marcescent, lobes 4, 
14, mm, deltate. Inflorescences cymose, terminal on 
axillary branches, many-flowered, puberulent to 
pilosulose, sessile to pedunculate; peduncle to 6 cm; 
branched portion corymbiform, 4.5-11 X 4-18 em, 


Annals of the 
Missouri Botanical Garden 


branched to 3 to 5 orders, lax; bracts 6-20 mm; 
. Flowers 5- 
ut apparently homostylous), 
sessile to pedicellate. Pistillate flowers: calyx cup- 
shaped, 2-3 mm wide, outside glabrous or puberulent, 
lobes to 0.3 mm, 
triangular; corolla white, clavate in bud, when open 


bracteoles reduced; pedicels to 1 mm 


merous, unise 


inside glabrous, truncate or 
salverform, outside glabrous or puberulent, tube 6— 
mm, mm diam., inside villous in upper third, 
lobes 2—4 mm, lists; ovate-oblong, or lanceolate, 
acute; staminodia included to partly exserted, fila- 
ments inserted in upper third of corolla tube, ca. 
3 mm; style 7-8.5 mm, glabrous or often pubescent in 
pper part, stigmas 0.5-1 mm. Staminate flowers: 
similar to pistillate except corolla with tube 2 
4.5 mm diam., lobes 3.54.5 mm; anthers included or 
shortly exserted, filaments ca. 2 mm; pistillode with 
style portion 7-8 mm, glabrous, stigmatic portion 1.5— 
mm. Drupes violet-black, didymous or subglobose 
5-7 X 5-7 


rugose, finely fissured, endosperm entire. 


mm; pyrenes spherical or hemispherical, 


Distribution and habitat. This species grows in 
outheast Asia, where it is wn from Borneo, 
specifically its Brunei and western Sabah (Malaysia) 
sections. It has been found in humid forests at 


elevations of 12-60 m 


Phenology. This species has been collected with 
flowers from April through August, and with fruits in 
August and September. 

Discussion. | Gaertne 


scribed here was previously treated by van Beusekom 


ra aphanodioica as circum- 


(1967) as a homostylous, bisexual form of G. vaginans 
subsp. junghuhniana. However, van Beusekom ac- 
knowledged that his form hermaphroditica Beusekom 
was heterogeneous and comprised several recogniz- 
able populations that he considered to have evolved in 
isolation. Gaertnera aphanodioica represents one of 
these populations from Brunei and western Sabah, 
which includes the type specimen of van Beusekom's 
form. Pollination studies in Berakas Forest Reserve in 
Brunei indicate that the plants are not bisexual as van 
Beusekom suggested, but functionally dioecious with 
the two flower forms almost indistinguishable (Mal- 
comber, unpublished data). The epithet aphanodioica 
refers to the cryptic dioecy of the species. Gaertnera 
aphanodioica can often be found in the same vicinity 
as G. junghuhniana but differs in the flowers with 
stigmas or the stigmatic part of the pistillode and 
stamens or staminodia positioned at the same height 
within the corolla tube. This species belongs to 

e G. vaginans complex; see also iscus- 
sion of that group for related species and thei 
distinctions. 


BRUNEI DARUSSALAM. Belait: Andulau 
ve, Compartment E ed 2998 (MO), 
ps MO. "3000 (MO). Brunei-Mua rakas Forest 
Reserve, e $2200 (K, SAR, $ SING), poem $4928 
BRUN, SAR), Malcomber 2 MO), 2986 (MO), 2987 
(MO), em (MO), 2989 (MO), a (MO), 2991 (MO), 2992 
(MO), 2993 (MO), 2994 (MO), 2995 (MO), 2996 (MO), 2997 
(MO); Universiti Brunei Darussalam campus, Malcomber 
2966 (MO). MALAYSIA. Sipitang: Menggalong FR, Meijer 
SAN 21809 (K, L, SAN). 


Pie eee 


4. Gaertnera arenaria Baker, J. Linn. Soc., Bot. 
20: 209. ae ae arenaria (Baker) Kuntze, 
Gen. Pl. 2: 425. 1 adagas- 

car. idis near amataves: July 1 CJ 


Meller s.n. (holotype, K!). Figure 6H-N. 


Revis. 


Gaerinera spathacea Drake, in Grandid., Hist. 


e, 
Madagascar 36 (6, Atlas 6): t 


1898 [1899], syn. nov. TYPE: 
Madagascar. Nosy-bé, Boivin 2074 (holotype, P not 
located). 


or shrubs, 3-10 m tall; branches terete to 
E libus 2-5 mm diam.; internodes 
2-5 cm, smooth to shallowly sulcate. Leaf blades 5.5— 


obtuse, drying chartaceous, glabrous; secondary veins 
prominulous abaxially, 6 to 9(to 11) pairs; domatia 


chartaceous, persistent or occasion- 
ally deciduous after distalmost 1 to 3 nodes, tube 8— 
17 mm, with ribs 4, narrowly winged, arising below 
petiole and extending to lobes, apex entire or with 1 or 
2 incisions, marcescent, lobes 4, 2.5—5 mm. Inflores- 
cences cymose, many-flowered, terminal on principal 
and/or axillary stems, puberule 
or peduncle to 9 cm; branched portion corymbiform to 
broadly pyramidal, 3.5-18.5 X 5-16 cm, branched to 
3 to 5 orders, lax; bracts 1-28(-65) mm; bracteoles 
triangular to ovate, 
Fl 


nt to glabrous, sessile 


mm long. 

rs 5-merous, heterodistylous, sessile to pedicel- 
late. Long-styled flowers: calyx cup-shaped, 2.5—4 mm 
wide, outside puberulent at base to glabrous at apex, 
inside with hair-ring, truncate or lobes 0.2-0.5 mm, 
rhomboidal in bud, at 
anthesis salverform, outside glabrous or scabrous, 
tube 10-16 mm, 1.84.5 mm diam., inside villous in 
upper third, lobes 


triangular; corolla white, 


3.5-5.3 mm, ligulate or oblong, 
acute; anthers included, filaments inserted in upper 
third of corolla tube, 1.2-1.5 mm; style 13-15 mm, 
glabrous, Short- pe flowers: 
similar to long styled except calyx 2-3.5 mm wide, 
lobes 0.1-0.3 mm; corolla tube 10-15 mm, L 54 mm 
diam., lobes 3.5—5.5 mm; anthers shortly exserted, 
filaments 2-3.5 mm; style 6 


stigma 


mm, stigma 


1.9 mm. Drupes violet-black, sabglobsee or globose to 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


S. 


E 


D 


-G. edi de e ie 


Stamina ower in lla 
Baker. —H. Flowering branc I. Fr bs pipa 
flower. —L. Short-styled suave cross Sei 


2028; H-N Based on a Moise 5. 


8-11 X 


ellipsoid or hemispherical, rugose, finely fissured, 


didymous, 7.5-9 mm; pyrenes spherical to 


endosperm entire. 


Distribution and habitat. This species grows in 


Madagascar, where it is known from the provinces of 
Antsiranana, Fianarantsoa, Toamasina, and Toliara. 
Here, it is observed to be widespread in humid forests, 
at elevations of 0—1400 m, often near oceans. 


Phenology. This species has been collected with 
flowers February through October, and with fruits 


January through March and September through 
December 
Discussion. | Gaertnera arenaria is a member of the 


vaginans complex of species; see also the 
discussion of that group for related species and their 
distinctions. It is one of the most commonly collected 
pecies aert in Madagascar. Gaertne 
arenaria differs from other species in this 


its relatively long corolla tube, 


The holotype collection of G. spathacea, Boivin 2074, 


—A. ma. branch. 


—J. Portion of stem with iow bases and stip ules 
—M. Long-styled flow 


ame 5-mm ie X Lto same l-cm 


—B. Por 


tion of stem Ey E bases, nid s 
with petiole bases and stipule. . Sta 


nate flow 


—N. Long-styled flower in cross section. 
scale. vu G based on Moiconiber 


o 


scale; M. Nt to same A -mm s 


was not located, but based on its description and 
several other specimens annotated with this name by 
Drake, it appears to be synonomous with G. arenaria. 


Representative 
tsiranana 


specimens examined. MADAGASCAR. 
osy-be, Pa Birkinshaw 117 (K, MO, 
AN); Manongarivo t i 
MO); Bekolosy, Mas 1946 (K, MO, P 
Ambalafary, Malcomber 2229 (K, MO, P, TAN) Jardin 
Botanique, Malcomber 9. 
Montagne d'Ambre National Park, Malcomber 1275 (MO, 
TAN), 1799 (K, MO, P, TAN); Marojejy Nature Reserve, 
Schatz 1542 (BR, G, P, TAN, WAG); Sambava, Maroambihy, 
a Silasy RN 9512 Ho, TEP) Ambilobe, Water- 
anjary, Geay 7370 x 
nte, remm 16082 (P; Mas 

Malcomber 2812 (MO), 2815 o. 


a A Fianarantsoa: 


Toa ul po: 
a Park, “toda 


al. 3236 (BR, MO, P, TAN, TEF, WAG); Andohahela Nature 
Reserve, Malcomber 1152 (K, MO, P, TAN); Eminiminy, 
lla: River, Malcomber 2173 (MO, P, TAN); N of Isaka 
Ivondry, Malcomber 2637 (K, MO, P, TAN); NW Manatenina, 
Ambalavoanio, Cours 3203 (P, TAN 


5. Gaertnera aurea Malcomber, sp. nov. TYPE: 
Ivory Coast [Cóte d'Ivoire]. Left bank of La Mé 


Annals of the 
Missouri Botanical Garden 


River, near dd to Mépé, 5737'N, deb a 
50m, 196 Leeuwenberg 


A. J. M. 
(holotype, WAGI; isotypes, BR!, K!, L!, MO!, » 


Haec species Gaerinerae cooperi Hutch. & M. B. Moss 
similis, sed ab ea ramis gracil ute scabro- 
pubescentibus trichomatibus pallide fuscis atque tubo 
corollino breviore (1.8-3.3 mm longo) apice integro distin- 
guitur. 


ibus in juventui 


Trees or shrubs, 2.5—4 m tall; branches flattened 
near apex, otherwise terete, young branches densely 
puberulent to short-pilosulose with indumentum 
drying yellow to brown, becoming glabrescent, 1.5— 
2.5(-6) mm diam.; internodes 0.8-5.5 em, smooth. 
Leaf blades 6.5-19 X 3-6.5 cm, elliptic to elliptic- 
oblong or obovate, apex cuspidate to acuminate, base 
cuneate to obtuse, drying chartaceous, adaxially 
glabrous, abaxially glabrous or puberulent to pilosu- 
lose and sometimes hirtellous on costa and principal 
veins; secondary veins prominent abaxially, 4 to 7 
pairs; domatia present; 4—9 mm. Stipules 
calyptrate, glabrous to densely puberulent or pilosu- 
lose, drying chartaceous or membranous, caducous, 


petioles 


winged, arising beneath petiole and sometimes 
extending to lobes, apex with a l-interpetiolar 
incision, marcescent, lobes 4, 4—5 mm, deltate to 
linear. Inflorescences cymose, many-flowered, terminal 
and/or axillary branches, densely puber- 
peduncle 1-5 em; branched 


on principal 
ulent to pilosulose; 
portion corymbiform, 1-7 X 1.5-9 cm, branched to 

to 4 orders, congested to lax; bracts deltate or linear, 

pedicels mm 

Flowers 5-merous, heterodistylous, sessile to pedicel- 
late. Long-styled flowers: calyx cup-shaped, 1.5- 
2.5 mm wide, glabrous outside, with hair-ring inside, 


14 mm; bracteoles reduced; 


truncate or lobes to 0.2 mm, triangular; corolla white, 
clavate in bud, when open salverform, outside 
1.8-3.3 mm, 1.7- 
2.2 mm diam., inside villous in upper third, lobes 
1.8-2.3 mm, 
included, filaments inserted at ca. middle of corolla 
tube, 0.3-0.6 mm; style 2.8-3.4 mm, glabrous, stig- 
mas 1.8-2.2 mm. Short-styled ac similar to long 
mm; corolla tube 


ligulate or oblong, acute; anthers 


styled ds calyx lobes up to 
2.3-3.2 2.2 mm diam lobes 2-2.7 mm; 
anthers fully "m filaments inserted in upper 
third of corolla tube, 0.3-0.7 mm; style 0.8-1.2 mm, 
stigma 0.9-1.2 mm. Drupes violet-black, subglobose 
to ellipsoid, 5-10 x 
plano-convex to hemi-ellipsoid, rugose, deeply fis- 


mm; pyrenes ellipsoid or 


sured, endosperm ruminated. 
Distribution and habitat. This species grows in 
West Africa, where it is known from Ghana and Cóte 


d'Ivoire. Here, it has been found in wet forests at 


elevations of 0—100 m 


Phenology. This species has been collected with 
flowers throughout the year, and with fruits January 
through October. 


Discussion. Gaertnera aurea has been confused 
with G. cooperi but is a more slender plant with a 
ntire corolla tube, 2.3-3.2 


mm long er fenestrate in G. 


shorter, e long and not 
fenestrate versus 8-11 
cooperi. Gaertnera aurea and G. cooperi share deeply 
fissured, ruminated pyrenes, an unusual condition. 
The specific epithet refers to the dense pale beige to 
brown drying color of the indumentum on the young 


branches. 
Paratypes. GHANA. Bimpong FR, Foso, Enti SP 599 
(BR, MO); Tarkw. Benso, Duah 5823 (BR, FHO); 


Enti CG42670 (MO). IVORY COAST [CÓTE D'IVOIRE]. 

SW of Guéyo, Leeuwenberg 3798 (BR, K, MO, P, 
WAG); 33 km along Abidjan—Adsopé rd., Leeuwenberg 7992 
(K, MO, WAG); Abidjan University garden, Aké Assi 673 (G, 
P); Abou-abou forest, po Abidjan Grand Bassam, 
Oldeman 240 (K, MO, P WAC), F. Hallé 409 (MO, P), J. de 
Wilde 3161 (K, WAG), Leeuwenberg 236 6 (BR, .FHO, K, L, 

m & A. 


Aouabo, 25 km N of Abidjan, Versteegh & Wi 
(MO, WAG); ee R. Wow 6. n. (P; Abidjan, Banco 
Arboretum, Aké Assi 654 (C), B 71 (P), Chatelain 2 (G), 
F. Hallé 217 (P), 284 ca 296 (Mio, a J. de Wilde 135 (WA x 
F. Hallé 4 (MO); Banco National Park, W. 


Adioukrou, Dabou, "Chevalier 17250 (P); Téké Forest, 12 kmN 
of Anyama, Beentje 568 (WAG); Toumanguié Forest, Aboisso— 
Grand Bassam rd., Bamps 2041 (BR). 


6. Gaertnera bambusifolia Malcomber & A. P. 


Davis, Monogr. Syst. Bot. Missouri Bot. Gard. 
104: 380, fig. 2. 2005. TYPE: Madagascar. 
Antsiranana: Reserve Naturelle de Marojejy, 


along trail to the summit of Marojejy Est, NW 
of Mandena, wet, evergreen forest above second 
, 49°15'E, 850- E 
1 Miller & P. P. Lowry Il 3978 


camp, 14^26' 
1 : 
Ms MO-5714849!; isotypes, K!, P!, TAN). 


Trees, to 3 m tall; branches terete to flattened, 
glabrous, 0.5-2 mm diam.; internodes 0.8—4 cm, 
smooth. Leaf blades 2.5-7 X 0.5-1.5 em, linear- 
apex acuminate to lon 


lanceolate to lanceolate, 
b obtuse, 


8 
acuminate, base cuneate to charta- 
ceous, glabrous; secondary veins flat to thinly 
prominulous abaxially, 4 to 6 pairs; domatia absent; 
petioles mm. Stipules tubular, glabrous, 


drying membranous, caducous, tube 3—5.5 mm, with 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


ribs none or forming thin ridges under petioles, apex 
entire, marcescent, lobes 3 or 4, 0.3—2 mm, filiform. 
Inflorescences reduced to a single flower, terminal on 
principal and/or axillary branches, pendulous, sessile 
or pedunculate; peduncles 9-30 mm, glabrous; brac- 
teoles reduced. Flowers 5-merous, biology unknown. 
Calyx cup-shaped, 1.9-2.5 mm wide, glabrous, trun- 
cate or lobes to 0.4 mm, triangular; corolla, anthers, 
and stigmas unknown. Immature drupes subglobose or 


didymous, 5.5-6 X 5-6 mm 
Distribution and habitat. This species grows in 
Madagascar, where it is known from the province of 


Here, it has been found in humid 
evergreen forests of the Marojejy National Park and 
Anjanaharibe-Sud Wildlife 
metamorphic and igneous rocks at elevations of 


850-1235 m 


eserve, growing on 


Phenology. This species has been collected with 
immature fruits in February and Mare 


Discussion. Gaertnera bambusifolia is only known 
from two immature fruiting collections. The species is 
named for its distinctive lanceolate leaves and can 

by its tubular 
stipules and its inflorescences reduced to a single, 


also be recognized relatively short 
often long-pedicellate flower. Maleomber and Davis 
(2005) considered the con 


species to be 


nservation status of this 
2001) based 


Endangered 
primarily on its small known range. 


Representative specimens examined. MADAGASCAR. 
tsiranana: Andapa, Ambodiangezoka, Am ape 


uo -Sud Special Reserve, Ravelonarivo 680 


A 


es 


7. Gaertnera belumutensis Malcomber, sp. nov 
TYP alaysia. Johor: Kluang, Kluang Forest 
Reserve, Gunong Belumut, summit trail, 2203'N, 
103^33'E, 1000 m, 25 May 1998, S. T. Mal- 
comber 3024 (holotype, MO!; isotypes, A!, AAU!, 
BO!, K!, KEP!, L!, QRS!, SAN!, SAR!, SING!). 


Haec species. Porn BOUT. Benth. similis, sd 


ab ea planta in 


capituliformi distinguitur. 


Trees or shrubs, 1-6 m tall, drying with orange 
cast; branches terete, glabrous, 2-3 mm diam; 
internodes 2.5—7 cm, smooth. Leaf blades 3.5-14 X 
1.2-3.5 em, elliptic or elliptic-lanceolate, apex short- 
ly cuspidate or acuminate, base attenuate or cuneate, 

rying chartaceous, glabrous; secondary veins evident 
abaxially, 5 to 8 pairs; domatia absent or present; 
petioles 3-14 mm. Stipules tubular, glabrous, drying 
chartaceous, caducous or deciduous through fragmen- 
tation, tube 6-13 mm, with ribs 4, narrowly winged, 
arising below petiole and extending to lobes, apex 


entire or with 2 opposite incisions, marcescent, lobes 
4, 2-5 mm, deltate to linear. Inflorescences congested- 
cymose, terminal on axillary branches, several- to 
many-flowered, glabrous, subsessile or peduncle to 
3.5 em; branched portion subglobose or corymbiform, 
1.1-3 X 1.53 em, branched to 1 to 3 orders, 
congested; bracts deltate or trifid, 1-4 mm; bracteoles 
reduced; pedicels to 2 mm. Flowers 5-merous, floral 
biology unknown, sessile to pedicellate. Cal up- 
shaped, 3-5 mm wide, outside glabrous or puberulent, 
with hair-ring inside, truncate or lobes to 0.3 mm, 
triangular; corolla, anthers, and stigmas unknown. 
Drupes violet-black, didymous or globose, 6-8 X 

8 mm; pyrenes spherical or hemispherical, rugose, 
finely fissured, endosperm entire. 


Distribution and habitat. 
southeastern Asia, where it is known from Peninsular 


This species grows in 


aysia. Here it has been found in humid forests at 
aes of 170-1000 m 

Phenology. This species has been collected with 
flowers February through May, and with fruits April 
through September. 

Discussion. Gaertner: poorly 
known species Au to southern Johor State of 
Malaysia, with most of the collections from Gunung 


a belumutensis is a 


Belumut. In this region, oe G. aa and G. 
kochummenii are known to dry with a distinctive 
orange cast; G. B io. from the latter 
species in its glabrous stems, leaves, and stipules, 
and its generally truncate calyx. is species 
belongs to the G. vaginans complex; see also the 
discussion of that group for related species and their 
distinctions. The specific epithet refers to the type 


locality 

Paratypes. MALAYSIA. Johor: Endau, South Plateau, 
Kiew 2192 (KEP); Gunong Blumut Rede aa 
ee 3004 G); Gunong t trail, 


of 
m 53 (D, Shah 2 
a T Sipura FRI 17830 (KEP, L, SINO), Whitmore FRI 
8726 (K, KEP, L), a FRI dioe d SING), FRI 
17844 (KEP), Walker F. S. KEP 33831 (KEP), Holttw 
10687 do Gunong A Mat 3723 (SING); Kluang, Canone 
Gua n, Jumali 3054 (SING); Labis Forest Reserve, 
dia F RI 32848 (KEP). 


a 


8. Gaertnera bieleri (De Wild.) E. M. A. Petit, 
Bull. Jard. Bot. État Bruxelles 29: 51. 1959. 
Basionym: Psychotria bieleri De Wild., Ann. 
Mus. Congo Belge, Bot., sér. 5, 2: 179, t. 64. 
1907. TYPE: Belgian Congo [Democratic Re- 
public of the Congo]. 
[Equateur-Orientale-Maniema], aut 


1904, R. Bieler s.n. (holotype, BR!). 


Distr. Forestier Central 
opori, 


WE fissistipula K. Schum Krause, Bot. Jahrb. 
t. 39: 563. 1907. Dui (K. Schum. & 


Annals of the 
Missouri Botanical Garden 


Krause) E. A. Petit, Bull. Jard. Bot. État 
Bruxelles 29: des 1959. TYPE: dre bei Bipende 
im Umwald der Ngabilandschaff, Mar. 1900, G. Zenker 
2252 ([holotype, m lectotype, pue here, K!; 
isotypes, BM!, K! [labeled as 2252a)). 


Trees or shrubs, 0.5-8 m tall; branches terete, when 
young pilosulose and/or hirtellous with indumentum 
drying brown or yellow, often becoming glabrescent, 

mm diam.; internodes cm, smooth. 
blades 2.2-17.5 X 
oblanceolate, or obovate, apex cuspidate or acumi- 

base acute to cuneate, dryin 
adaxially glabrous or pestes se on principal veins, 


1.1-7.8 em, elliptic to oblong, 
chartaceous, 


abaxially glabrous or usually hirtellous to pilosulose 


lous abaxially, 3 to 10 pairs; domatia absent or 
present; petioles 2-11 mm. Stipules tubular, ne 
hirtellous to E drying chartaceous, caducous 
tube m, wit s 4, narrowly winged, arising 
ie petiole and Nu to lobes, apex entire or 
sometimes wit 
lobes 4 


numerous, to 2 mm. Inflorescences cymose, several- to 


2 opposite incisions, marcescent, 
, 1-10 mm, deltate to filiform, setae none or 


many-flowered, terminal on principal and/or axillary 

branches, cu ana to hirtellous, sessile or 
le to 5.4 subglobos 

uno ded or narrowly ae 0.7-6.7 x x 


pedunc portion 
7.4 em, rather congested, branched to 3 to 4 orders; 
bracts linear to filiform, 3-21 mm; bracteoles 1— 
2 mm; pedicels to 3 mm. Flowers 5-merous, hetero- 
distylous, sessile to pedicellate. Long-styled flowers: 
calyx eup-shape m wide, outside glabrous to 
puberulent or Pilosilose; with hair-ring inside, lobes 


(0.8—)2—4.5 mm, 


lanceolate; corolla white, clavate in bud, when open 


linear to narrowly triangular or 


infundibuliform or salverform, outside glabrous, tube 
4—5 mm, 0.9-2.5 mm diam., 
third, lobes 2-3 mm, 
n acute; anthers Y filaments inserted 
in upper 4.5 

7 mm, pues or x stigmas 0.3-1 mm. 


inside villous in upper 
ligulate or ovate-oblong to 


ird of corolla tube, 


Short-styled flowers: similar to long ERA except calyx 
3—4.5 m 


lobes 0.9-3 mm; corolla tube mm 
anthers fully exerted, 


on lobes 1.4-3.5 mm; 
laments 2.2-3.6 mm; 


style 2-3.8 mm, glabrous, 


stigmas 1-1.4 m rai violet-black, globose or 
subglobose, 5— 12 x pyrenes spherical or 
hemispherical, rugose, finely fissured, endosperm 
entire. 


Distribution and habitat. This species grows in 
Central Africa, and also in West Africa east of the 
Dahomey Gap. It has been found in Cameroon, the 
Democratic Republic of the Congo, and Nigeria, in 
humid forests at elevations of 50-1270 m 


Phenology. This species has been collected with 
flowers throughout the year, and with fruits February 
throug! 


Discussion. Gaertnera bieleri is a morphologically 
variable species, particularly in inflorescence shape, 
which may be subglobose, corymbiform, or pyramidal, 
although the inflorescence axes are usually relatively 
short and congested in all of these. This species is 
recognizable by its puberulent to hirtellous indumen- 

m, well-developed linear-filiform calyx lobes, and 
stipule tubes with longitudinal ridges and four lobes 
plus sometimes numerous setae at the top. Gaertnera 
ee is similar to G. longivaginalis var. bracteata (E. 

Petit) Malcomber; see comments under that 
section of the paper as well. 

The type of Gaertnera fissistipula was destroyed 
with the general Rubiaceae collection in the Berlin 
herbarium; the isotype specimen at K is selected here 
as lectotype because a digital image is available on 
the Aluka web site: <http://www.aluka.org/page/ 
content/plants.jsp>. 
jp xamined. CAMEROON. 55 km 
S of Nyong pr ə km W of So 
(K, MO, P, W. 


Rep JE 
SE of Eséka, 
Leeuwenberg 5074 
(MO). DEMOCRATIC 
OF THE CONGO. 60 km N of Kisangani, 
Mee Lisowski 18904 (BR, K); Eala, Lebrun 331 (BR, 
FHO, G, P), Pynaert 1788 a Kiou, Shabunda, Bulumba, 
T Léonard 3743 (BR, FHO, G, K); Leopoldville, Inongo, 
betw. Selenge & Lukolela, Goossens 5017 (K). BON. 
Woleu Niem: 20 km E of Mitzie, Jeffrey 231 (K, P). 
NIGERIA. Stubbs Creek Forest Reserve, 30 km E of Eket, 
Van Meer 1185 (WAG 


9. Gaertnera brevipedicellata Malcomber & A. P. 
Davis, Monogr. Syst. Bot. Missouri Bot. Gard. 104: 
an fig. 3. 2005. TYPE: Madagascar. Fiana- 

antsoa: Ran 
2i 16'S, 41*25'E, 800—900 m, 5 Nov. 1997, $, 
Malcomber, A. Davis, D. Gower, J. Andriantiana $e 
A. Katozafy 2877 (holotype, MO; isotypes, BR!, 
G!, K!, P!, PRE!, TEF!, WAG!). 


omafana National Park, Talatakely, 


Shrubs, 2-6 m tall, sometimes clambering; branch- 
es terete, glabrous, 0.5-3 mm diam., with bark corky, 
pale gray; internodes 0.4—4.5 cm, with 2 longitudinal 
Leaf blades 0.8-3.8 X 0.4-2 em, 


elliptic or ovate to elliptic-oblong, apex shortly 


ribs or wings. 


acuminate, base obtuse to acute, drying chartaceous, 
glabrous; secondary veins thinly prominulous abaxi- 
ally, 3 1.2- 
2.8 mm. Stipules calyptrate, glabrous, drying mem- 


to 5 pairs; domatia absent; petioles 
branous, caducous, tube 1 .5 mm, with ribs 4, 
winged, arising Beneath petiole extending to 
n each side extending along 
middle of interpetiolar side, with wings under petioles 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


becoming indurated and somewhat enlarged with age, 
apex with 1 or 2 incisions, marcescent, lobes 4, 0.4— 
1 mm, linear. Inflorescences reduced to a single flower, 
terminal on axillary branches; bracteoles reduced. 
Flowers 5-merous, heterodistylous, subsessile. Long- 
styled flowers: calyx cup-shaped, 1.5-2 mm wide, 
glabrous, lobes 1.1—4.2 mm, narrowly triangular to 
linear; corolla pink, clavate in bud, when open 
sa verform, outside glabrous, tube 3-3.5 mm, 1.4— 
., inside villous at ca. middle, lobes 3.5— 
4.8 mm, anelar or ligulate, acute; anthers included, 
filaments inserted at ca. middle of corolla tube, ca. 
.2 mm; style 4.5-5.5 mm, glabrous, stigma 1-1. 
Short-styled flowers: similar to long ds) except nd 
lobes 1—4.5 mm; corolla tube 4—5.5 mm, 1-2.5 m 
diam., lobes 4-5 mm; anthers shortly excerted. fla- 
ments 2.5-3.5 mm; style 1.5— 
1.6 mm. Drupes violet-black, subglobose or didymous, 
4.5-5.5 X 4.5-7 mm; pyrenes spherical or hemispher- 


mm, stigma 1.5- 
ical, rugose, finely fissured, endosperm entire. 


Distribution and habitat. 
Madagascar, where it has been found in the province 


This species grows in 


of Fianarantsoa in humid evergreen forests in the 
Ranomafana National Park. Here, it grows on 


metamorphic and igneous rocks at elevations of 


Phenology. This species has been collected with 
flowers in November and once with fruits, but the 
month of that collection was not noted. 


Discussion. Gaertnera brevipedicellata is similar to 
G. madagascariensis in the inflorescence reduced to a 
single flower, but differs from that species in lack of 
indumentum, internodes and stipules with prominent 
longitudinal ridges or wings, and pink subsessile 
wings 
are distinctive; these do not extend onto the stipule 
sheath. Malcomber and Davis (2005) considered the 


conservation status of this species to be Endangered 


owers. The well-developed internode ridges or 


(IUCN, 2001) primarily due to its limited known 
range. 

Representative specimens examined. MADAGASCAR. 
ianarantsoa: Ranomafana National Park, Talatakely, 


Fi 

Kotozafy 1073 (MO, TEF), Malcomber 2876 (MO, TEF); 
omafana National Park, Vatoharanana, Malcomber 2867 

(MO, TEF 


10. Nea calycina Bojer, Hortus Maurit. 217. 
7. Sykesia erue (Bojer) Kuntze, Revis. 
d pl. 2: 425. 1891. TYPE: Mauritius. Grand 
rel, W. Bojer s.n. (holotype, P 

not located; mn BM!, MAU!). 


Bassin 


Gaerinera calycina var. variegata Bojer, Hortus Maurit. 217. 
1837. TYPE: Mauritius. Trois lots, Bojer s.n. (holotype, 
P not located). 


Gaerinera aetheonoma Steud., Nomencl. Bot. [Steudel], ed. 2, 
1: 651 


A. DC., Prodr. (DC.) 9: 35 (1845), nom. nud., 
TYPE: "Mauritius. s. loc., Sieber s.n. dle > hol 


located). 
Gaerinera Ro C. Presl, p Konigl. Bohm. Ges. Wiss. 5, 
Bd. . 1845, nom. nud. TYPE: Mauritius. s. loc. 


Sieber pe peres PR? not located). 


es or shrubs, 0.9-1.2 m tall; branches terete to 
No glabrous, 4—7 mm di nternodes 

—3.5 cm, smooth. Leaf blades 3-175 3.5— 
9 cm, elliptic to oblanceolate or elliptic-oblong, apex 
ded, base 


rying chartaceous, glabrous; secondary 


obtuse then shortly cuspidate or roun acute 
or cuneate, 
veins prominulous abaxially, 10 to 13 pairs; domatia 
absent or present; petioles 17-30(—40) mm. Stipules 
incompletely known, caducous or fragmenting leaving 
truncate basal portion 2-5 mm, glabrous, marcescent. 
Inflorescences cymose, many-flowered, terminal on 
gm and/or axillary branches, glabrous; peduncle 
c g; branched portion corymbiform, 2 

7-24 cm, "branched to 2 to 3 orders, lax; bracts 
ligulate to ovate, deltate, or trifid, 12-30 mm, 
glabrous; bracteoles 3-8 mm s 

sometimes articulated above middle. Flow 
merous, presumably heterodistylous. Long-styled flow- 
9-20 


lobes 2-5 mm, broadly triangular to rounded; corolla 


ers: calyx campanulate, mm wide, glabrous, 
apparently white, clavate in bud, when open salver- 
form, outside glabrous, tube 15-20 mm, E 5—4 mm 
., inside villous in upper third, lobes 10-14 mm 
alte to oblong, acute; anthers included, filaments 
inserted in upper third of corolla tube, ca. 1.5 mm; 
style 20-25 mm, glabrous, stigma 2.5-3 mm. Short- 
Drupes black E. Jem or 


tyi nknown. 
violet- a éllipseid to fusiform, 15-20 


pyrenes plano-convex or ellipsoid, rugose, xe 
fissured, endosperm entire. 
Distribution and habitat. | Gaertnera calycina 


grows in Mauritius and perhaps also in Madagascar. 
The localities, habitat, and elevation have not been 
recorded. 


Phenology. This collected with 
flowers in March and has been collected in fruit, but 


species was 


the months of those collections were not noted. 


Discussion. This unusual, probably showy and 
certainly striking coe is notable for its rela- 
tively large calyx and is incompletely known and 

eai extinct ros : Gillett 1998). The ed 
of its 


collection that is aa ‘labeled erroneously. The 


distribution in r is based on 


most recent collection of this species was made by 


Bouton in the early 20th century (exact date 


nknown). 


Annals of the 
Missouri Botanical Garden 


A similar but apparently distinct taxon, “Gaertnera 
sp. A” of Verdcourt (1983, 1989), was collected on 
Mauritius by Edgerley (as MAU 13414) in 1969. These 
plants differ from G. calycina in the pubescent 
inflorescence axes, smaller calyx, and larger pubescent 
leaves. Verdcourt (1989) noted that this collection is 
intermediate morphologically between G. calycina and 
a pubescent-leaved form of G. longifolia and could 
represent a hybrid between these two species, or could 
be equivalent to 


G. calycina var. variegata. Numerous 
TM. h 


E 
researchers including; ave revisited Bassin 
Blanc attempting to re-collect both G. calycina and “G. 
sp. A” without success. Rare or extinct Gaertnera 
species have recently b d in Mauritius 
(e.g., , 1997), 


even in areas now dominated by introduced species, so 


een rediscovere 
ngifolia var. pubescens; Anonymous 


*G. sp. A” and G. calycina might be rediscovered with 
continued exploration. 
ne variety of Gaertnera calycina has been 
described, based on the calyx white rather than 
green. Because this species has been seen 
recently, the status of this variety is difficult to 
evaluate, but some other Gaertnera species show 
similar within-population variation in calyx color (e.g., 
G. phyllostachya) that does not seem taxonomically 
informative; TR variety variegata is not 
recognized her 
In the imus for Gaertnera calycina, Bojer 
suggested tentatively that the name Chassalia divar- 
icata DC. might also apply to this species, but he di 
; thus, his 
treated here as validly pulished. Chassalia divaricata 


not clearly accept the synonymy name is 
more recently treated as a synonym of C 

ev. (Verdcourt, 1989). As 
Candolle 45) placed €. 


calycina into its own group, section Aetheonoma, 


lanceolata (Poir.) A. 
discussed above, de 
based on several characters including an irregular 
arrangement of the anthers, with three alternipetalous 
and the other two opposite the corolla lobes. This 
condition has not been confirmed (cf. Verdcourt, 
1989; S.T.M., pers. obs.) and may be an artifact of the 
preparation of some specimens. Steudel's name G. 

aetheonoma was apparently based on a specimen of 
Sieber's that was previously identified as G. vaginata. 


Representative specimens examined. MADAGASCAR. s. 
loc., L. A. Chapelier s.n. (P). MAURITIUS. Grand Bassin 
Nouvelle Decouverte, L. S. ur (K); s. loc., Sieber 52 
(E, G, L, MO, P, W), 188 (E, L, P). 


11. Gaertnera capitulata Malcomber, sp. nov. 
TYPE: Malaysia. d P. Aman, Selepong 
Barangan, 15 Aug. 1991, Y. P. C. Runi S 59673 
(holotype, SAN!; Duos Fi KEP!, L!, MO). 


Haec species Gaerinerae junghuhnianae Miq. similis, sed 
ab ea stipulis in quoque latere tubi ac sub petiolo alis duabus 


longitudinalibus prominentibus munitis atque inflorescentia 
congesta capituliformi distinguitur. 

Shrubs, up to 1.5 m tall; branches terete, when 
young hirtellous to puberulent with indumentum 
drying reddish O white, or brown, Ed n. 
E 1- lam.; quies es 2- 
h. Leaf TE 4.320 X 


to pire n m. acute to acuminate, base acute 


—5.8 em, elliptic 
to cuneate, chartaceous, adaxially glabrous, 
abaxially hirtellous to villous or tomentulose with 
indumentum drying brown or reddened; secondary 
veins prominulous abaxially, 4 to 10 pairs; domatia 
absent; petioles 3-20 mm. Stipules tubular, hirtellous 
to villous, drying chartaceous, persistent on distalmost 
nodes or deciduous leaving a truncate persistent base 
14 mm, tube 6.5-13 mm, with ribs 4, broadly 
winged, arising below petiole and extending to lobes, 
apex entire or with 2 incisions, marcescent, lobes 4, 
7-10 mm, deltate to linear. Inflorescences congested- 
cymose to subcapitate, many-flowered, terminal on 
principal and/or axillary branches, puberulent to 
hirtellous or villous, sessile or peduncle to 1.4 em; 
branched ag os Musae d 1.2-3 X 1.2-3 em, 
ranched to 1 to eds Mine deltate, 1-10 mm; 
bracteoles ae pedicels to 1.5 Flowers 4- 
merous, unisexual, ea to cde te Pistillate 


tulose, glabrous inside, truncate or lobes 
triangular; corolla white, clavate in bud, when open 
tube 7.5-8.5 mm, 
inside villous in upper third, 


salverform, outside puberulent, 
0.7-2.5 mm diam., 
lobes 2.5-3.5 mm, 
shortly exserted, filaments inserted in upper third of 


ligulate, apex acute; anthers 
corolla tube, ca. 0.3 mm; pistillode reduced. Drupes 


unknown. 


Distribution and. habitat. 
Southeast Asia, where it is known from Borneo, in 


This species grows in 


both its Sarawak (Malaysia) and Kalimantan (Indone- 
sia) sections. Here, it has been found in humid forests 
at elevations of 45—700 


Phenology. This species has been collected with 
flowers April through November. 


Discussion. | Gaertnera capitulata differs from most 
other Southeast Asian Gaertnera species by its 
prominent longitudinal wings arising beneath the 
petioles and extending indie the stipule tube, o 


s, and c 


cences. It is similar to P Res see "additional 


elliptie leave sted, subglobose inflores- 
comments under that species. Most collections are 
ensely pubescent but Lee S 45386 is puberulent. This 
species belongs to the G. vaginans complex; see also 
the discussion of that group for related species and 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


their distinctions. The specific epithet refers to the 
often subcapitate inflorescences. 


e ace Mc ar Kalimantan: Upper Katingan 
River, 50-10 WNW of T g Samba, Mogea 3457 
(K, KEP, L). A Saràwak: "pela a, Linau, Sung 
Iban, Lee S 45386 (K, KEP, SAN, SAR); Bukit Mersing, Sibot 
ak Luang S 21999 (K, SAN, SING); Sri Aman Distr., 
Selepong Barangan, Malcomber 5037 (MO), Rena George S 
58327 (K, L, SAR). 


. Gaertnera eardiocarpa Boivin ex Baill, 
Toamasina: Sainte Marie, Tan. o, Apr. 1851, 
.-H. Boivin s.n. (holotype, P!; NE GN. 


Trees, 3—6 m tall, or sometimes Dru branches 
ed to terete, eid 0.52. diam.; 

es 1.26 mooth. Leaf blades 4.5 8 
(13.5) X 1.4-4(-6) em, , elliptic lanceolate, elliptic- 
oblong, or elliptic to obovate, apex cuspidate or 
acuminate, base acute to cuneate, drying coriaceous 
or chartaceous, glabrous; secondary veins prominu- 
lous abaxially, (5 to)7 to 11 pairs; domatia absent or 
(2-4—11 mm. Stipules tubular, 


glabrous, drying chartaceous, caducous or rarely 


present; petioles 
persistent on distalmost 1 or 2 nodes, tube 4— 
10 mm, with ribs 4, narrowly winged, arising below 
petiole and extending to lobes, apex entire or with 1 
, deltate to 
linear. Inflorescences cymose, many-flowered, terminal 


incision, marcescent, lobes 4, 0.3-1.5 mm 


on principal and/or axillary branches, glabrous; 
peduncle 2.5-6 em; branched portion narrowly pyra- 
midal, 1.8-11.5 X 1.5-7 em, pendulous, branched to 
2 to 3 orders, dele 1.5-7 mm; 
bracteoles triangular to lanceolate, 0.5-1.5 mm; 
pedicels absent or to 
heterodistylous. 


lax; bracts 


4-merous, 
cup- 
shaped, 1.5-3 mm wide, glabrous, truncate or lobes 


mm. Flowers 
Long-styled flowers: calyx 


to 0.3 mm, triangular; corolla white, clavate in bud, 
when open — outside glabrous, tube 5- 
6 mm, id illous 

middle, lobes w mm, Baa to ligulate, acute; 


at ca. 


anthers included, filaments inserted at ca. middle of 
corolla tube, ca. mm long; style 

"m l. 5 —2 mm. Short-styled diesen: 
similar to long styled except calyx 1.8-3 mm wide, 
lobes to 0.2 mm; corolla tube 1.2-2.5 mm diam.; 
filaments 0.9-1.5 mm; style 1.3-2 long, stigmas 1— 
1.5 mm. Drupes violet-black, globose to subglobose or 
didymous, 5.5-6 X 


hemispherical, 


7-7.5 mm; pyrenes spherical or 


rugose, finely fissured, endosperm 


entire 
Distribution and habitat. 
Madagascar, where it is known from the province of 


This species grows in 


Toamasina. Here, it has been found in the Ile Sainte 


Marie, Mananara-Nord National Park, and the 


Masoala Peninsula, growing in humid forests at 


elevations of 0-300 m 


Phenology. This species has been collected with 
flowers May through November, and with fruits April 
through October. 


Discussion. Gaertnera cardiocarpa can be recog- 
nized by its chartaceous to coriaceous leaves, 
narrowly pyramidal inflorescences, and 4-merous 
distylous flowers. It is generally similar to G. inflexa, 
ut differs in its inflorescences usually borne 
terminally on axillary branches e its reduced 
stipule lobes. Gaertnera cardiocarpa often dries 
blue-gray when the collections are not preserved in 
alcohol prior to drying. A c ing habit is reported in 
the notes of Dumetz 914, ae other specimens are 


reported as trees. 


Representative specimens examined. MADAGASCAR. 


, P, TAN), L 
Lowry 4486A (MO, P), 4127 (MO, P), 
Malcomber 2803 (MO, TEF), Malcomber 2820 (MO, TEF). 


13. Gaertnera cooperi Hutch. & M. B. Moss, Fl. 
Afr. 2: 21. 1931. TYPE: Liber 
Dukwia Rive G. P. Cooper 287 (holotype, Ki, 
isotype, Al). 


Trees or shrubs, 4-8 m tall; branches flattened to 
quadrangular or terete, when young puberulent to 
pilosulose with indumentum drying reddish to brown, 
becoming glabrescent to pilosulose with indumentum 
m diam.; internodes 1.5— 


to oblong or ovate, apex shortly cuspidate to acute, 
base cuneate to obtuse, drying coriaceous to stiffly 
chartaceous, adaxially glabrous, abaxially glabrous or 
puberulent on principal veins to throughout with 
indumentum drying reddish to brown; secondary veins 
prominent abaxially, 5 to 10 pairs; domatia absent; 
petioles 8-22 mm. Stipules calyptrate, densely puber- 
ulent to pur drying membranous, soon decid- 
uous, tube 9— m, with ribs 4, narrowly winged, 
endis irn base and eyendns to lobes, after 
falling leaving a persistent tubular base 1.5-2 mm 
A apex with 1 or 2 incisions, marcescent, lobes 4, 

deltate. Inflorescences 
terminal on principal and/or axillary 


cymose, many- 
Fs 
branches, densely puberulent to pilosulose or strigil- 
ong; branched E uan 
1.5-16 em, branched t 


orders, jailed congested; bracts deltate or ioe to 


lose; P 0.5-9 cm 


corymbiform, 2-8 X 


lanceolate, 1-9(—40) mm; bracteoles reduced; pedi- 
cels absent or to 2.8 mm. Flowers 5-merous, hetero- 
distylous. Long-styled flowers: calyx cup-shaped, 3.5— 
4.5 mm wide, glabrous to densely puberulent outside, 


Annals of the 
Missouri Botanical Garden 


lobes 0.3—4 mm, 


wu ee corolla white, clavate-urceolate 


glabrous inside, triangular to 
de outside im to 

pee iu tube 8-11 mm, 

fenestrate in upper part, E villous in upper third, 


lobes 3.5-6 mm, 


filaments inserted in upper third of corolla tube, 

3.5 mm; style 12-15 mm, glabrous, stigma 1—1.5 mm. 
Short-styled flowers: similar to long styled except calyx 
0.1-0.5 mm; 


style 2.5-3 mm, 
stigma 1.8-2.5 mm. Drupes violet-black, globose to 


well exserted, filaments 3—4 mm; 
subglobose or didymous, 5-10 X 5-8 mm; pyrenes 
spherical to ellipsoid or hemispherical to hemi- 
ellipsoid, rugose, deeply fissured, endosperm rumi- 
nate 


Distribution and habitat. 
est Africa, where it has been found in Ghana, Céte 


This species grows in 


d'Ivoire, and Liberia. Here, it has been found growing 
in humid forests at elevations of 80-90 m 


Phenology. This species has been collected with 
flowers January, February, and May through Decem- 
ber, and with fruits January through May and October 
through December. 


Discussion. Gaertnera cooperi is 


to G. 


aurea, also of Ghana and Cóte d'Ivoire; G. cooperi 


similar 


differs in its leaves drying coriaceous to stiffly 
chartaceous, its larger flowers with corolla tubes 8— 
11 mm long, and its corolla that is fenestrate in the 
upper part of the tube. This last character is unique in 
Gaertnera. On dried specimens of this species, the 
tertiary leaf venation is usually regularly sublineolate 
and well marked, in contrast to this venation being not 
evident or irregularly areolate in most Gaertnera 
species 

Gaertnera cooperi was first published with an 
English description and then was republished with a 
Latin description by the same authors in Kew Bull. 
1937: 62 


of publication. However. 


, 1937, which is sometimes cited as its place 
, a Latin description is not 
required by the International Code of Botanical 
Nornienelaiure (ICBN; McNeil et al., 2006) for valid 


first publication was adequat Pet 
(1959a), the paratype specimens Dues 12420, 
Chevalier 12664, and Chevalier 12936 do not belong 
to the same species as holotype; these specimens are 
included here in G. longivaginalis var. bracteata. 
Representative specimens examined, ANA. Ateiku, 


Vigne 1948 (K). IVORY COAST [COTE D'IVOIRE]. Tai 
National Park, beside Gala trail, Gautier-Beguin 1032 (G), 


Aké Assi 17797 (G, MO). LIBERIA. Grand Bassa Co., 

e Baldwin 11121 (K, i Bodji Town, Adam 

7 (K, MO); pe 16 mi. N of University Forest, 

ende 1650 (K, MO, WAG); near Firestone Plantations, 
Dukwai River, Cooper d (BM, FHO, GH, K 


he 
Nn 
eo 


14. Gaertnera crassiflora Bojer, Hortus Maurit. 
216. 
d te (Bojer) Kuntze, Revis. Gen. 
Pl.-2 1. TYPE: Mauritius. Trois Ilots et 
autour px e Bassin, W. Bojer s.n. (holotype, 
P not located; isotype, G-DC!). 


aertnera crassifolia, orth. var. 


Trees, height not noted; branches terete, glabrous, 
3-5 mm diam.; internodes 1.5-2 cm, smooth. Leaf 
blades 12-19 X 5.8-7.5 em, oblong to elliptic, apex 
rounded then abruptly contracted into a short tip, base 
cuneate, drying coriaceous, glabrous; secondary veins 
visible abaxially, ca. 7 pairs; domatia absent; petioles 
10-15 mm. Stipules incompletely known, calyptrate, 
glabrous, caducous, tube length unknown, with ribs 4, 
narrowly winged, Rad below petiole and extending 
to lobes, apex marcescent, lobes unknown. Inflores- 
cences cymose, acd) terminal on principal 
stems, glabrous; peduncle 1.5-2.5 cm; branched 
portion subglobose or d , 288.5 X 2 

m, branched to 1 to 2 orders, aiher congested; 
bracts ligulate to ovate or trifid, 1-6 mm; bracteoles 
1-5 mm. Flowers 5-merous, heterodistylous, subses- 
sile. Long-styled flowers: calyx cup-shaped, 6—7 mm 
wide, outside glabrous, with hair-ring inside, truncate 
or lobes to 0.4 mm, triangular; corolla color unknown, 
in bud, 


salverform, glabrous outside, tube 


clavate when open infundibuliform to 


mm diam. in a 
ligulate or elliptic-oblong, acute to 
included, filaments inserted in upper third of corolla 
tube, 0.2-0.3 mm; style 20-21 mm, glabrous, stigma 

0.6-1.3 mm. Short-styled Pera long 


styled except corolla tube mm, 3—5 mm diam.; 


similar to 


anthers shortly exserted, oe ca. 1.5 mm; style 
15-16 mm, stigma 6.5—7 mm. Drupes unknown, said 
to be ellipsoid to fusiform or pyramidal and white at 
maturity (Bojer, 1837) 


Distribution and habitat. 
Mauritius and has been reported also to grow in 


This species grows in 


adagascar, but this report seems doubtful, as 
discussed below i 
have not been recorded. 

Phenology. 


mens of this species have not been note 


. The habitat and elevational range 
The collection dates for the speci- 
d. 


Discussion. This species is poorly known and 
probably extinct (Walter & Gillett, 1998). Gaertnera 
crassiflora can 


e recognized by its deciduous 


calyptrate stipules leaving persistent bases with 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


stem apex ae eae not fully detailed). " —H. Ga 

—D. Portion of stem w af bases and stipules. 
Short-styled flower. H Short-styled flower in cross section. E, 
Malcomber 2940; C-H based on Malcomber 2888. 


well-developed longitudinal wings and its UA DM 


to congested-corymbiform inflorescences. The ec- 


tion locality of one specimen of this species, zw 
n. (G), is not from Mauritius and is given as near 
( This 


species has no 


Tananarive (= Antananarivo), Madagascar. 
t been found again in this region, an 
the distributional record seems most likely a labeling 


error. 


Additional specimens examined. MADAGASCAR. Anta- 

ntananarivo], J. P. 
., Bélanger 101 (P); 
ceived Nov. 1875, 
Herb. Horne s.n. n (K); Herb. Roxburgh s.n. (BM 


15. Gaertnera cuneifolia Bojer, Hortus Maurit. 
216. 1837. E ae (Bojer) Kuntze, 
Revis. Gen. P 425. TY 

Forests of a at e Tiseouseste and on 
the peak in the middle of the island, W. Bojer s.n. 
(holotype, P not G-DC?). 
Figure 7A, B 


: Mauritius. 


located; isotype, 


Trees or shrubs, 1-2 m tall; branches terete to 
flattened, glabrous, 3—5 mm diam.; internodes 0.5— 
x 


Leaf blades 3-5.5 1.44 cm, 


elliptic-oblong or obovate to cuneiform, apex rounded 


cm, smoot 


uncle broken). 
cross section. —G. 
same l-cm scale. A, B based on 


flower in 


F to same 1- emacile G, H to 


to truncate and sometimes abruptly shortly cuspidate, 
base acute to obtuse, margins often thinly revolute, 
dryi ] dary 

prominulous abaxially, 5 to 7 pairs; domatia usually 


ob 
ing stiffly coriaceous, glabrous; secon veins 


membranous, caducous or deciduous through frag- 
mentation, tube 14-20 mm, with ribs 4, arising below 
petioles, not extending onto tube or uniting in basal 
part of interpetiolar side of tube, apex with 1 incision, 
marcescent, lobes 2 or 4, 2-2.5 mm, deltate to linear. 
Inflorescences subcapitate, many-flowered, terminal on 
principal and/or axillary branches, glabrous, subses- 
sile, subglobose, 2-4 X 2— 
basalmost (i.e., outermost) bracts often foliaceous, 


4 cm, densely congested; 


racts linear, 
m. Flowers 5-merous, heterodistylous, subsessile. 
EL UMS flowers: calyx campanulate, 7-8 mm wide, 
glabrous, lobes mm, linear to oo corolla 
rh clavate in bud, when open salverform, outside 
glabrous or pubescent, tube to ca. 6 mm, to ca. 4 mm 
diam., pubescence condition inside unknown, lobes 


Short-styled 


flowers: similar to long styled except corolla lobes 


7-10 mm, linear or ligulate, acute. 


Annals of the 
Missouri Botanical Garden 


linear to ovate-oblong. Drupes white (Bojer, 1837) or 
violet-black, ellipsoid, ca. 15 0 mm; pyrenes 
plano-convex or ellipsoid, smooth or rugose, finely 


fissured, endosperm entire 


Distribution and habitat. 
Mauritius. Here, it is known from heathlike vegetation 


This species grows in 


on lava flows (“groundwater laterite”) at elevations of 


550—650 m. 


Phenology. This species has been collected with 
flowers May through August, and with fruits January 
through May and in December. 


Baker (1877) treated Gaertnera cunei- 
folia as a synonym of G. rotundifolia, but Verdcourt 
(1983, 1989) recognized these as distinct species. 

oth species are 
heathlike vegetation found on local lava flows, and, 


Discussion. 


nown from the distinctive 


unlike most Gaertnera, species grow there Hd to 
full sun (D. Lorence, 


cuneifolia differs from G. rotundifolia in its xn 
obl 


pers. comm.) Gaertnera 
, obovate, or cuneiform leaves with the 

vendue flattened and invisible deste its a 
rescences with well-developed, usually involucral 
lobes 6-8 mm long 
G. rotundifolia has broadly elliptic to obovate or ovate 


bracts, and its calyx . In contrast, 
leaves with the tertiary venation sometimes elevated 
and evident adaxially, reduced bracts, and calyx lobes 
1-2 mm lon 


not appear to be fused on most specimens, but do 


g. The basalmost inflorescence bracts do 


overlap widely and largely enclose the inflorescences. 
Field studies indicate that the two species also differ 
in phenology: G. cuneifolia flowers May through 
August, whereas G. rotundifolia flowers November 
nera cuneifolia was listed as 
are t (1998) and is ee 
restricted to Blac 
Le Pétrin and environs. 


River Gorges National Park n 


Bojer suggested tentatively in the protologue that 
this species may be the same as Chassalia coffeoides 
DC., but 
Thus, this name seems validly published; C. coffeoides 


did not positively accept that synonymy. 


is treated here as a synonym of Gaertnera psycho- 
trioides. 


Mq e d specimens examined. MAURITIUS. Black 
River G 
2940 (M 
Ehumpusne: Tirvengadum 969 (K). 


11894 (K, MAU) Plaine 


16. Gaertnera dareyana Malcomber & A. P. Davis, 


fig. 2. 2005. TYPE: Madagascar. 
11.9 km N of Andasibe, Me National Park, 
18°49'S, 4826'E, 930 m v. 1997, S. T. 
Malcomber 2921 Delos E^ 5714808 iso- 
types, BR!, K!, P!, TEF!, WAG). 


Shrubs, to 5 m tall, sometimes clambering; branches 
terete to flattened, glabrous, 2—4 mm diam.; internodes 
1; cm, smooth or with a longitudinal Ln Leaf 
blades 4—6.5 X 0.7-3.2 cm, 


elliptic or lanceolate-elliptic, 


linear-lanceolate to 
apex cuspidate or 
acuminate, base cuneate to rounded, drying charta- 
ceous, glabrous; secondary veins flat to prominulous 
abaxially, 4 to 7 pairs; domatia absent; petioles 0.1— 
1.1 mm. Stipules tubular, glabrous, chartaceous, cadu- 
cous or deciduous through fragmentation, tube 1.7— 

4, narrowly winged, arising eem 
petiole and sometimes extending to apex, apex entire or 
with 1 or 2 incisions, marcescent, lobes 4, 1.4-2.2 mm, 


li - to 9-flowered, terminal 
on allang branches, alibrous, sessile or peduncle 1.2— 
2.1 em; branched portion when present corymbiform, 
0.8-2 X 0.4-2.5 em, branched to 1 to 2 orders, lax; 
bracts deltate to linear or trifid, 2.5-8.6 mm; bracteoles 
educed; pedicels 2-6 mm. Flowers 5-merous, presum- 
ably heterodistylous. Long-styled flowers: not seen. 
Short-styled flowers: calyx cup-shaped, 2-3.2 mm wide, 
glabrous, lobes unequal, 0.4—3.1 mm, triangular to 
narrowly spatulate or linear; corolla white, in bud 
clavate, when open tubular-funnelform, outside gla- 
brous; tube ca. 4 mm, ca. 1.5 mm diam., villous in 
upper part, lobes ca. 2.2 mm, triangular, acute; anthers 
exserted, filaments inserted in upper part of tube, ca. 
1.5 mm; style ca. 2.5 mm, stigmas ca. 1.2 mm. Drupes 
violet-black, subglobose to didymous, 6.5-7.5 X 5.2— 

2 mm; pyrenes spherical to hemispherical, rugose, 
finely fissured, endosperm ruminate. 


Distribution and habitat. 
Madagascar, where it is known from the provinces of 


This species grows in 


Toamasina and Fianarantsoa. It is so far known only 
as a rare species found in humid evergreen forests on 
metamorphic and igneous rocks, at elevations of 700— 
1400 m 


Phenology. This species has been collected with 
er and December and has been 

with fruits, but the 

collections have not been recorded. 


in Nove 


e 
collected months of those 


Discussion. Gaertnera darcyana differs from other 
few 


Gaertnera species in its few-flowered inflorescences, 
calyptrate stipules, and rather well-developed, narrow 
a lobes. Malcomber and Davis (2005) considered 
servation d of this species to be V 
01). 


the cons 
perle (IUCN, 2 


Representative specimens examined. MADAGASCAR. 
Fianarantsoa: Haute vallée de la Rienana, Bassin du 

Matitanana, Humbert 3566 (P). Toamasina: Ambodiniana, 
Cours 1926 (P). 


. Gaertnera divaricata (Thwaites) Thwaites, 
Enum. Pl. Zeyl. 425. 1864. Basionym: Pristidia 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


Figure 8. 
bases, stipules, a 


A-E. Gaertnera ue M EE d 
—C. 


and stem apex Portion of stem 


short-styled flower. —E. Shor- styled Joner i in cross iion 


scale). —G. 

flower. —J. Short-styled flower in cross section. "K 
same l-cm scale; D, E to same 5-mm le I-L t 
Malcomber 2900. 


divaricata Thwaites, Enum. Pl. Zeyl. 149. 1859. 
aertnera koenigii var. divaricata (Thwaites) C. 
B. Clarke, in Hook. f., Fl. Brit. India 4: 91. 1883. 
TYPE: Sri Lanka. Near Galle, G. H. K. Thwaites 
CP 2991 (holotype, K!; isotypes, BM!, BR!, 
CGE!, G!, KI, P!, PDA!, W!, WU!). Figure 8A—E. 


Trees, 1.5-3 m tall; branches flattened at apex, 
otherwise terete, glabrous, 1.5-3 mm diam.; inter- 
nodes 3-10.5 em, smooth. Leaf blades (4—)7.5-16 X 
(2-)2.8-8.5 cm, 


ovate, or obovate, apex acuminate to cuspidate, base 


lanceolate to elliptic, elliptic-ob- 


cuneate or obtuse, drying chartaceous, glabrous; 
r not visible 
ioles 3— 
12 mm. Stipules tubular, glabrous, drying chartaceous, 
e tube (36-9 mm, with ribs none or 4, 

arrowly winged, encircling base of petiole but not 
euenit up tube, apex entire, marcescent, lobes 4. 
0.5-1 mm, deltate. 
flowered, axillary or supra-axillary on principal 


Inflorescences cymose, many- 
branches, pendulous, glabrous; peduncle 3-7 cm; 
branched portion corymbiform to pyramidal, 1.3-9 
X 2-9(-12) em, lax, branched to 3 to 6 orders; axes 


petiole bases and old st 
FL. Gaerinera pone Baill. 


Long- „styled r. —L. Long-style 
to same 5-mm a A-E based on Malcomber 2762; F-L based on 


—A. Young ee branch. —B. Portion of stem with petiole 
pules h 


. —D. Portion of inflorescence axis wit 


ower in cross ection: B, C to 


dichasial at basal nodes becoming scorpioid at distal 
nodes; bracts deltate to linear, 0.5—4 mm; bracteoles 
reduced; pedicels absent or to 1 mm. Flowers 4- 
Long-styled flowers: calyx 
1-2 mm wide, glabrous outside, with 


merous, heterodistylous. 
cup-shaped, 
hair-ring inside, truncate to denticulate with teeth to 
0.2 mm; corolla white, clavate in bud, when open 
4—5.] mm, l- 
inside villous in upper third, lobes 


salverform, outside glabrous, tube 
2.5 mm diam., 
2.54 mm, 
shortly exserted, filaments inserted in upper third of 


ligulate or triangular, acute; anthers 
corolla tube, ca. 0.5 mm; style 5—5.5 mm, glabrous, 
stigmas 0.5-1 mm. Short-styled flowers: similar to long 
styled except corolla tube 4.5-5.5 mm, 1-2 mm 

diam., lobes 2 mm; filaments 0.5-1.5 mm; style 
2.5-3.5 mm, Gens 1-1.5 mm. Drupes violet-black, 
subglobose or obovoid, 14-17 X 10-12 mm; pyrenes 
spherical or hemispherical, rugose, finely fissured, 


endosperm entire 


Distribution and habitat. This species grows in Sri 
Lanka, where it is known from the district of Galle. 
Here, it has been found in humid forests at elevations 


of ca. 900 m 


Annals of the 
Missouri Botanical Garden 


Phenology. This species has been collected with 
flowers September through November, and with fruits 
January through August 


Discussion. Gaertnera divaricata is a distinctive 
species with axillary or supra-axillary inflorescences 
with scorpioid axes and 4-merous flowers. The species 
is locally common around Hiniduma, particularly in 
Kanneliya Forest Reserve. 


Representative specimens examined. SRI LANKA. Galle: 
iniduma, eae 

, Huber 582 PDA), 

can 1509 (K, MO, PDA), Kosterman 27135 a L). 


18. Gaertnera diversifolia Ridl., J. Fed. Malay 
States Mus. 6: 163. 1915. Gaertnera oblanceolata 
var. age ie .) Beusekom, Blumea 1 

Malaysia. Selangor: Bukit 
in. LA 7 E Kelsall 1995 (lectotype, 
designa kom, 1967 [1968]: 
383, e. EE Malaysia, H. N. Ridley 
7429 (SING!).] 


Gaerinera intermedia Ridl. |J. Fed. Malay States Mus. 6: 163. 
1915. Au. elangor: Hulu Semangko, 

4, H. N. Ridley 12080 olas SING). 
Gaerinera. i Ridl., J. Fed. Ls States Mus. 6: 162. 
1 hom. illeg., non Gaertne d Bouton ex 
> J. Linn. 
non Gene 


Malaysia, H. N. Ridley 16255 (syntype, SING!).] 
Gaerinera ovata, Ridl., J. Straits Branch Roy. Asiat. Soc. 
. 1922. TYPE: Wire. Selangor: ll 
Pass, H. N. Ridley s.n. (holotype, K!). 
Gaerinera rigida Ridl., J. Straits Branch Roy. Asiat. Soc. 86: 
> TYPE: Malaysia. Negeri Sembilan: Bukit 
Ta H. N. eae oe KS). 
ies "larfelia Ridl., Bot. 62: 299. 1924. TYPE: 
Malaysia. Selangor: Tus aser's Hill, Sep. 1922, I. Burkill 
& R. E. Holttum 8608 (holotype, SING!). 


Trees or shrubs, 0.5—3 m tall; branches terete to 
quadrangular, glabrous, 1-6 mm diam.; internodes 3— 
5.5 em, smooth. Leaf blades 10-30 X 2-11 em, 
oblanceolate to obovate, apex shortly acuminate or 
acute, base acute to attenuate to cuneate, drying 

coriaceous, glabrous; secondary ve prominent 
abaxially, 5 to 15 pairs; domatia a petioles 
12-40 mm. Stipules tubular, 


persistent, caducous, or deciduous through b 


glabrous, variously 
tation sometimes leaving a truncate base 1-8 m 

m, with ribs 4, 
winged, arising below petiole and extending along 


narrowly to broadly 
tube to lobes, apex with 2 incisions, marcescent, lobes 
ular to deltate. 
terminal on 


or 4, l-4 mm, narrowly triang 
Inflorescences cymose, many-flowered, 


axillary and/or supra-axillary branches, puberulent to 
pilosulose, pendulous; peduncle 1.3-5 em; branched 
portion narrowly due 2—1.5 X 4.6—6.5 cm, lax, 
branched to 3 o orders; bracts and bracteoles 
reduced; pedicel 1 5 mm. Supra-axillary branches 
when present pendulous, up to 45 cm, 1-2 mm diam., 
with reduced leaves (0.2-)4-8(-15) X (0.1-)0.9-2 
(-2.5) em, linear to narrowly elliptic or oblanceolate, 
drying chartaceous; secondary veins prominulous 
abaxially, 4 to 8 pairs; petiole 1-10 mm. Flowers 5- 
merous, unisexual. calyx cup- 
shaped, 
puberulent, with hair-ring inside, truncate or lobes 
to 0.2 mm, broadly triangular; corolla white, in bud 
mature 


clavate, outside glabrous or puberulent, 


corolla, staminodes, and mas unknown. Staminate 


sti 
rs: similar to pistillate as far as known, corolla 
villous inside in upper third. Drupes violet-black, 
didymous or globose, 7-8 X 7-9 mm; pyrenes 
spherical or hemispherical, rugose, finely fissured, 


endosperm entire. 


Distribution and habitat. 
Southeast Asia, where it is known in Peninsular 


This species grows in 


Brunei, Kalimantan (Malaysia), and Sarawak (Indo- 
nesia) sectors. Here, it has been found in humid 
forests at elevations of 130-1450 m. 


Phenology. This species has been collected with 
flowers January through September, and with fruits 
January through May and July through December. 


Discussion. Gaertnera diversifolia is a widespread, 
locally common species within Peninsular Malaysia 
and Borneo and is recognized by its narrowly 
pyramidal inflorescences often borne on pendulous 
supra-axillary branches and its 5-merous flowers. It is 
similar to G. inflexa of Madagascar, but differs from 
that by its 5-merous unisexual flowers. Gaertnera 
diversifolia was treated by van Beusekom (1967) as a 
variety of G. oblanceolata, but these are treated as two 


species here. 


Representatiwe specimens examined. BRUNEI DARUS- 
L 


Hill, Malcomber 3018 (MO), 3019 (MO), Neg FRI 1935 (K, 
KEP) Sabah: Ranau Distr, Tenompok Ridge paik 
headquarters, Beaman 8215 (L). Sarawak: Miri, Sua Ilias 
S. 39183 (K, L, SAN, SAR). Selangor: Ulu Terengganu, Ng 
FRI 22056 (KEP) 


19. Gaertnera drakeana Aug. DC., Bull. Herb. 
i é 586. 1901. TYPE: Madagas- 


car. Toamasina: Maroa [Maroantsetra], in forest, 


Boissier, Sér. 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


ure 9. A-G. Gaertnera drakeana Aug. DC. 


—A. Flowering branch. —B, C. Portions of stems with a bases and 
G. Lor 


stipule s. —D. er m flower. —E. Short-styled flower in cross section . Long-styled fl tyled flow 
in cross secti N. Gaerinera D ba —H. Flowering Rn (to es -cm a —L J. Patong of stems with pail 
and apical e —kK. E led flower. —L. Short-styled flower in cross se —M. Long-styled flow: 


"i flower in cross se 
Moise 2; H-N based on Malsonibe 2915. 
1897, A. Mocquerys 273 (holotype, G-DC!). 
Figure 9A-G 
Trees, 0.3-6 m tall; branches flattened to terete, 
glabrous, 1.5-4 mm diam.; internodes (0.3-)2.8—7 
(-11.3) em, smooth. Leaf blades 5.5-16 X — 
4.8 cm, to oblanceolate, 
acuminate or cuspidate, base acute or cuneate, drying 


narrowly elliptic apex 
chartaceous, glabrous; secondary veins prominulous 
abaxially, 9 to 12 pairs; domatia absent; aa 5- 
1l mm. Stipules tubular, glabrous, drying membra- 
nous, caducous or deciduous through ace ld 
tube 4.7-18 mm, with 23 4, rounded to narrowly 
winged, arising below petiole and extending to lobes, 
shallow to deep incisions, 


Inflorescences subcapitate to cymose, several-flowered, 
terminal on principal and/or axillary branches, 
deflexed to pendulous, glabrous, sessile or peduncle 
to 2.7 cm; branched portion subglobose, 1.2-2.5 X 

1.1-2.1 


bracts deltate, 0.5—1 mm; bracteoles ovate to trian- 


cm, branched to 2 to 3 orders, congested; 


gular, 0.5-1 mm. Flowers 5-merous, heterodistylous, 
subsessile. Long-styled flowers: calyx cup-shaped, 2— 


BC to same 1 cm scale; D— ‘6 w 


scale. AL GI ei d 


same 25: -mm s serie KEN to same a -mm s 


3.5 m 
corolla white, clavate 
outside glabrous, iube 6-7 mm, 1.5— qm diam., 


mm wide, i lobes 0.3-0.6 mm, triangular; 
n bud, when open salverform 
inside villous at ca. middle, lobes 4-7 mm, Bod 
or ligulate, acute; anthers included, Alsen es mme 
in upper third of corolla tube, 0.2-0.5 mm; style 9— 
10 mm, glabrous, stigma 1-1.5 mm. Short-styled 
flowers: similar to long styled except calyx 2.2— 
3.5 mm wide, lobes 0.3-0.7 mm; corolla tube 6.5— 
8 mm, 2-4.5 m 


shortly Boi. Bent: 
1.5-1.8 


lobes 5.5-7 mm; anthers 
3-4 mm; style 4-5 m 
stigma —].8 mm. Drupes viele black or Vine, 


m diam., 


globose to subglobose or didymous, 7-10 X 7- 
10 mm; pyrenes spherical or ene rugose, 
finely fissured, endosperm entir 


Distribution and habitat. 
Madagascar, where it has been found in the province of 


This species grows in 


Toamasina in the Mananara-Nord and Masoala Na- 
tional Parks. Here, it grows in humid forest at 0300 m 


Phenology. This species has been collected with 
flowers August through December and has been 
collected in fruit, but the dates of those collections 
were not noted. 


616 Annals of the 
Missouri Botanical Garden 
Discussion. | Gaertnera drakeana is an infrequently — Short-styled flowers: similar to long styled except 
collected but locally common species that is appar- corolla tube 11-25 mm, 3-5 mm diam., lobes 8- 
ently restricted to Mananara-Nord and Masoala mm; anthers shortly exserted, filaments inserted in 
National Parks in northeastern Madagascar. The upper third of corolla tube, ca. 2 mm; style 5-9 mm, 


species is diagnosed by its stipules drying membra- 
nous and its relatively small subglobose inflorescenc- 
es. The 
southeastern Asia, but differs from that in the absence 


species is similar to G. sralensis of 


of a hair-ring in the calyx tube and its larger flowers. 


Representative specimens examined. MADAGASCAR. 
Toamasina: Mananara-Nord Nature Reserve, Raharimalala 
1502 (P); Masoala National Park, Malcomber 2802 (MO), 
2822 (MO). 


20. M egi edentata Bojer, Hortus Maurit. 216. 
Sykesia hes (Bojer) Kuntze, Revis. 
en Pl. 2: 425. 1891. TYPE: Mauritius. Foréts 
de la edle Decouverte au Quartier- Militaire 
& montagnes Grand-Port, 
(holotype, P not pede isotypes, G!, MAUN, 
Figure 10C—L 


Gaerinera petrinensis Verde., Kew Bull. 37: 538. 1983, syn. 
nov. TYPE: Mauritius. Petrin, L. Bernardi 14794 
(holotype, K!; isotypes, G!, P5 

Trees or shrubs, (1—-)1.5-3 m tall; branches flat- 
tened to terete, glabrous, 2-7.5 mm diam.; internodes 

(0.4—)0.9—2.8 cm, smooth. Leaf blades 2— 10 x 1.19 

1.7-5 em, lbs to oblong, apex acuminate or acute, 

base acute to obtuse or rarely subcordate, drying 

fl 


= 


coriaceous, glabrous; secondary veins visible and flai 
to prominulous abaxially, 4 to 9 pairs; domatia absent; 
petioles 2.5-20(-30) mm. Stipules tubular, glabrous to 
puberulent, drying a eM or d 
tent on distalmost 1 t nodes, tube 2-5 mm, with 
ribs 4, narrowly NU arising pu netic and 
sometimes extending to lobes, apex entire, marces- 
cent, lobes 4, 3—4( 

1-3 mm. Inflorescences cymose to subcapitate, (few- 
on principal and/or 


mm, linear; setae absent or 2 to 
to) many-flowered, terminal 
nches, glabrous; peduncle 0.7-3.5 em; 
portion corymbiform, (1-)2-4 x (2-)3- 
6 cm, branched to 2 to 3 orders, congested; bracts 
deltate or trifid, 2-15 mm, glabrous; bracteoles 
inserted at base of calyx, ovate to triangular, 1l— 
2 mm; pedicels absent or to 1 mm. Flowers 5-merous, 
Long-styled flowers: 
ide, gla ie truncate or lobes to 
ite, 1 


heterodi e calyx 


ped, 3-5 m 


06 1 mm, nie corolla 


cup- 


d clavate, 
lobed m with linear to ‘elliptic appendages l- 
2m open ae iform or salverform, 
due e tube 10-25 mm, m diam., 

inside glabrous or villous in upper did, pes 7 
10 mm, ligulate to linear, acute; anthers included, 
filaments inserted in upper third of corolla tube, ca. 
l mm; style 9-25 mm, glabrous, stigma 1-4 mm. 


glabrous, stigma 1-2.5 mm. Drupes white (Bojer, 
1837) or vit black to blue, ellipsoid, 6-10 X 
5-7 mm; pyrenes ellipsoid or hemispherical to 
finely fissured, endosperm 


plano-convex, rugose, 


ruminate. 
Distribution and habitat. 
Mauritius, where it is found in evergreen wet forests 


This species grows in 


and in the heathlike vegetation found on lava 
(“groundwater laterite”) at elevations of 300—700 m. 


Phenology. This species has been collected with 
flowers January through May and October through 
December, and with fruits May through September. 


Discussion. Gaertnera edentata is an attractive 
species with the largest flowers in the genus. The apex 
of the stipule may or may not have setae, but four 
obes are consistently present. This species differs 
from G. hirtiflora in its externally glabrous corolla, 
and from G. psychotrioides in its congested rather than 
lax inflorescences and larger flowers. The label data of 
Lorence 1554 and 2201 note that the flowers are 
fragrant; these were collected in the middle of the day 
D. H. Lorence, pers. com 

Verdcourt ie E Gaertnera petrinensis as 
a M localized species or form found in the 
thlike se dim that grows on lava flora on 
Mauritius. He separated this based on its calyx 3— 
4 mm long and undulate to shortly lobed, with the 
lobes up m long versus ong and 
truncate to undulate in his circumscription of G. 
edentata. However, the variation in these features on 
the species studied appears to be continuous and not 
correlated with habitat. At least one other Gae 
erdeourt and in this 
treatment is found in both a forests and the low 


rinera 


species as circumscribed 
heathlike vegetation (G. psychotrioides), and this name 
is considered a synonym here. 


eu specimens examined. MAURITIUS. 
e 


Curepipe, Vaughan MAU 1636 (MAU); Perrier Nature 
Reserve. Tirvengadum (P), Vaughan MAU 22 


MO), 2942 (MO), 2943 (MO), 2944 (MO), "2945 (MO), 
MO), 2950 (MO), 2954 (MO), 2960 (MO), d (MO), 
; Petrin Reserve, Lorence 131 (MO), 
MO; 2376 (MO), 2615 (MO), abes ud 394/44 


ho 
KS) 
A 
eo eo 


be 
gt 
as 


21. Gaertnera eketensis Wernham, J. Bot. 52: 30. 
1914. TYPE: Nigeria. South "rA State [Akwa 
Ibom]: Eket Distr., , P. A. Talbot 
3391 (holotype, BM! ae je Ky 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


1, m. id NES, 
BSS Sy 


Figure 10. A, B. Gaerinera e ci 11 —A. Fruiting branch. —B. Portion of stem with p ees d 
(to cale). — 


and stem 


styled flower in cross section. F-I to 


Trees, height not noted; bark with longitudinal 
striations or fissures; branches terete, glabrous, 1.5— 
m.; internodes 1.6—5 mooth to shortly 

led Leaf blades 8-12 x 34 cm, elliptic to 
obovate, apex cuspidate or acuminate, base cuneate, 
drying vp qu glabrous; secondary veins visible 
abaxially, 4 pairs; domatia present; petioles 3— 
6 mm. Sous tubular, glabrous, drying membranous, 
mm, with ribs 6, 4 of them 


narrowly winged, arising below petiole and extending 


caducous, tube 4—10 m 


to lobes, 2 ribs extending from longitudinal ribs of 
internodes to apex between lobes, apex entire, 
marcescent, lobes 4, mm, linear to filiform; setae 
terminal on principal and/or axillary 
branches, sparsely puberulent to glabrescent, sessile 
or E e to 4.2 em; branched portion corymbiform, 
—5.5 —1 cm, branched to 3 to 4 orders, lax; bracts 
ue or trifid, 1—4.5 mm; bracteoles reduced; 
pedicels absent or to 0.8 mm. Flowers 5-merous, 
Long-styled flowers: 
, 1.1-2 mm wide, glabrous, truncate or lobes 
i in bud, 
when open salverform, externally glabrous, tube ca. 


3.5 mm, 0.8-2 mm diam., 


heterodistylous. calyx cup- 
shape 
0.6 mm, triangular; corolla white, clavate 

] 


inside villous in upper 


ojer. —C. Flowering branc 3-c E. Por of 
. Long-styled flower in cross section. —H. Short. Sei me i hos 
same l-cm scale. A, B based on Malcomber 3038; C-I based on Malcomber 2954. 


third, lobes 2.5-3 mm, triangular to ligulate, acute; 
anthers shortly exserted, filaments inserted in upper 
third of corolla tube, 0.1—0.3 mm; style 2.7-3 mm, 
stigma 0.6-1 mm. Short-styled flowers: 
unknown. Drupes unknown. 


glabrous, 


Distribution and habitat. This species grows in 
West Africa, where it is found to the east of the 
Dahomey n Nigeria. Here, it can be found in 
humid lowland ious. but the elevations at which it 
has been found have not been recorded. 


Phenology. This species has been collected with 
flowers in April and May but has not yet been found 


with fruits. 
Discussion. This poorly known species is so far 
cumented three collections. pees and 


Dalziel nee considered Gaertne 

culata, but, as er E Pun (19508), 
these are iinet The m with numerous setae and 
the brown-black bark with prominent longitudinal 
fissures deis this species from G. paniculata, 
which lacks both of those features. Gaertnera eketensis is 
similar to G. liberiensis, as also noted by Petit (1959b); 
see comments about their separation under this latter 
species. Petit also considered G. eketensis similar to G. 


Annals of the 
Missouri Botanical Garden 


stictophylla (Hiern) E. M. A. Petit; however, the identity 
of the latter has not been established here, and that 
name is treated here as doubtful as to application. 


Additional specimens e. NIGERIA. en Eastern 
State [Akwa Ibom]: 30 km E of Eket, Stubbs Creek Forest 
Reserve, van Meer 1174 (W. ne Onochie FIH 5 68 (WAC). 


22. Gaertnera fractiflexa Beusekom, Blumea 15: 
375, fig. SA—C. 1968. TYPE: Malaysia. Sarawak: 
Matang, Peakes Tea Plantation, 18 July 1890, M. 
R. Haviland s.n. (holotype, SING!). 


Shrubs, 2—5 m tall; branches terete, glabrous, 1.5— 


diam.; internodes 0.5-5 em, smooth. 

a x . lanceolate to elliptic or 
m elliptic- PO apex cuspidate or acuminate, 
base cuneate, drying chartaceous, glabrous; secondary 
veins prominulous abaxially, 4 to 8 pairs; domatia 
absent; petioles 3-13 mm. Stipules tubular, glabrous 
to puberulent, drying membranous, deciduous through 
fragmentation, tube 7-15 mm, with ribs 4, narrowly 
winged, arising below petiole and sometimes extend- 
ing to lobes; apex with 2 incisions, marcescent, lobes 
mm, deltat 


4, ca. iid cym 


flowered, terminal on axillary SIME MN prd 
pendulous; peduncle 2.5-7 cm; branched portion 
broadly pyramidal, 3-11 X 3-9 cm, lax, branched 
to 3 to 4 orders, bracts deltate or linear, 3-30 mm; 
bracteoles reduced; pedicels absent or to 3 mm. 
Flowers 4-merous, unisexual. Pistillate flowers: calyx 
cup-shaped, 1.5-2.5 mm wide, outside glabrous, with 
to 0.3 mm, 


triangular; corolla, staminodes, and stigmas unknown. 


hair-ring inside, truncate or lobes 
Staminate flowers: calyx similar to pistillate as far as 
known; corolla white, clavate in bud, when open 
salverform, externally glabrous, tube 2-4 mm, 0.75- 
1.5 mm diam., inside villous in upper third, lobes 2— 
3 mm, lolas, acute; anthers included, filaments 
inserted in upper third of corolla tube, ca. 2.5 mm; 
pistillode poorly developed. PDrupes violet-black, 
5-7 5-8 m 


spherical or hemispherical, rugose, finely fissured, 


globose or subglobose, mm; pyrenes 


endosperm entire. 


Distribution and habitat. This species grows 
in the Sarawak (Malaysia) sector. 
be found in humid forests at an elevation of ca. 
145 m. 

Phenology. This species has been collected with 
flowers l but the date of that collection was not noted; it 


has been collected with fruits in June. 


Discus. 
junghuhniana but differs in its pendulous, pyramidal 


sion. Gaertnera fractiflexa is similar to G. 


inflorescence and 4-merous flowers. Gaertnera fracti- 


Leaf 


flexa was described by van Beusekom (1967) based on 
only one collection, suggesting that the species might 
be rare and possibly now extinct, but recent fieldwork 
G. fractiflexa is still extant at the type 
locality, although the population was limited in size 


shows that 


and scattered. 


Additional specimens examined. MALAYSIA. Sarawak: 
Matang, water catchment area, trail to Sri Maha Mariamman 
temple, Malcomber 3033 (MO 


23. Gaertnera pes (Baill. ex Vatke) Mal- 
mber & A 


Missouri Bot. 
Basionym: Psychotria furcellata Baill. ex Vatke, 
Abh. Naturwiss. V. 
TYPE: Madaga 
Rutenberg s.n. (holotype, P!). 


ereine Bremen 
Nov 


1877, D. C. 


scar. s. loc., 2 


Shrubs, to 2 m tall; branches terete, glabrous, 1— 
2 mm diam.; internodes 0.5— , with 2 longitu- 
dinal ribs. Leaf blades 0.5-1.2 X 0.3—0.5 cm, elliptic, 
apex acute to shortly cuspidate, base cuneate, drying 
membranous or chartaceous, glabrous; secondary 
veins not or poorly visible abaxially, 3 to 4 pairs; 


domatia absent; petioles absent or to 0.3 mm. Stipules 


tubular, glabrous, drying membranous, persistent 
whole or as 4 spatulate shreds, tube absent or up to 
l mm, with ribs 4, narrowly winged, arising below 
petiole and extending to lobes, apex with 4 incisions, 
lobes 4, 3—4 mm, linear to filiform. Inflorescences 
terminal on axillary 
to l mm; 
bracteoles reduced. Flowers 4-merous, heterodisty- 


lous. Long 


2 mm wide, outside glabrous, with hair-ring inside, 


reduced to a single flower, 
branches, sessile or with peduncles 
-styled flowers: calyx campanulate, 1.5— 


lobes 0.5-2.5 mm, triangular to linear; corolla white, 


villous in upper third, lobes 2.2-2.8 mm, ligulate to 
linear, acute or rounded; anthers included, filaments 
inserted in upper third of corolla tube, ca. 0.3 mm; 
style ca. 5 mm, glabrous, stigma 0.5-0.7 mm. Short- 


styled flowers: unknown. Drupes unknown. 


Distribution and habitat. 
Madagascar, 


This species grows in 


Province. Here, it can be found in evergreen humid 
forests on metamorphic and igneous rocks at an 
elevation of ca. 900 m 


Phenology. This species has been collected with 


olo 
flowers in November, but has not yet been collected 


with fruits. 


Discussion. Gaertnera furcellata is similar to G. 


microphylla. These poorly known species share 


calyptrate stipules that usually persist as four mem- 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


branous shreds at the leaf nodes and inflorescences 
reduced to a single flower; for further discussion, see 
the section for this latter species. This species was 
treated under the unpublished name “Gaertnera 
lacerata" by Malcomber (2000). Malcomber and Davis 
(2005) considered the conservation status of this 
species to be Critically Endangered (IUCN, 2001). 


AGASCAR. Toa- 
Cours 1203 (P), 


Additional specimens examined. MA 
masina: Samalahaza, Am Eee RR 
Dequaire 27977 (K, P 


24. Gaertnera gabonensis Malcomber, sp. nov. 
TYPE: Gabon. Iwateki, 15 Dec. 1930, G. Le Testu 
8575 (holotype, P!; isotypes, BM!, BR!, MO!). 

ecies Gaerinerae spicatae K. Schum. similis, sed 


ab ea inflorescentia thyrsiformi atque calyce dense puberulo 
vel pilosulo distinguitur. 


Haec sp 


Trees or shrubs; branches terete to + tetragonal, 
hollow, densely puberulent or pilosulose to glabres- 
cent, 3-6 mm diam.; internodes 7-9 cm, smooth. Leaf 
blades 21-30 X 8-11 cm, oblanceolate to elliptic- 
oblong, apex rounded then abruptly shortly cuspidate, 
base cuneate to obtuse, drying stiffly chartaceous, 
glabrous; secondary veins prominulous abaxially, 8 to 
10 pairs; domatia absent; petioles 15-25 mm. Stipules 
tubular, densely hirtellous or pilosulose to glabrous, 
drying chartaceous, deciduous through fragmentation, 
tube 15-25 mm, with ribs 4, broadly winged, arising 
below petiole and extending to lobes, apex with 2 
, 4-8 mm, deltate to 
linear. Inflorescences cymose, many-flowered, terminal 


incisions, marcescent, lobes 
on principal branches, puberulent to pilosulose; 


broadly pyramidal, 4.5—1 
to 5 orders, rather e bracts deltate or trifid, 
5-12 mm; bracteoles reduced. Flowers 5-merous, 
bisexual, heterodistylous, subsessile. 
flowers: calyx cup-shaped, 2-3. 
densely puberulent to pilosulose, glabrous inside, 
truncate or lobes to 1.2 mm, triangular to linear; 
corolla white, clavate in bud, when open salverform, 
outside densely puberulent to pilosulose, tube 6— 
9 mm, 1.4-2.1 


third, lobes 3.5—4.5 mm, elliptic-oblong, acute; an- 


mm diam., inside villous in upper 
thers included, filaments inserted in upper third o 
corolla tube, ca mm, glabrous, 
stigma 1.5-2.5 mm. Short-styled flowers: similar to 
long styled except corolla glabrous outside, tube 7— 
8.5 mm, 1.1-2 mm 
third, lobes 2-3 mm, ligulate; anthers fully exserted, 


iam., inside villous in upper 


filaments 3—4 mm; style 7-8 mm, stigma 1-1.5 mm 

Drupes unknown. 
Distribution and. habitat. 

Central Africa, where it is known from Gabon. Here, it 


This species grows in 


can be found in humid forests at elevations of 30— 
O m. 


Phenology. This species has been collected with 
flowers September through December, and with fruits 
in January, November, and December. 


Discussion. Gaertnera gabonensis is often con- 
fused with G. spicata, but differs in the absence of 
stipule setae, its corymbiform to broadly pyramidal 
inflorescences, and its densely puberulent to pilosu- 
lose calyx tubes. The specific epithet refers to the 
geographic distribution, which is apparently restricted 
to Gabon. 


Paratypes. GABON ounié: Betw. Mouila & Yeno, 
60 km from Mouila, Brewler 8029 (K, WAG). Nyanga: 
Mayombe bayaka, ee Le Testu 1642 (BM, BR, P); Near 
Van Nek 569 (WAG). Ogooué Maritime: Rabi- 
Kounga, Breteler 10214 ale: 9 Nov. 1991, Schoenmaker 
117 (WAG), 4 km N of Shell camp, Wieringa 1654 (WAG). 


25. Gaertnera Xgardneri Thwaites, Enum. Pl. 


346 (lectotype, designated by van E 
1967 [1968]: 380, BM!; isotypes, G!, PDA!, W!). 
[SYNTYPE: Sri Lanka, G. H. K. urn CP 363 


(K5.] 

Trees, 2—4 m tall; branches terete to flattened or 
trigonous, when young puberulent or hirtellous to 
glabrous with indumentum drying yellow or gray- 
white, becoming glabrescent, 1-5 mm diam.; inter- 
nodes 0.6-1.5(-2.4) cm, smooth or with 2 or 3 
blades 1.5— 


longitudinal ribs. Leaves paired or ternate; 
A 1.2) em, narrowly elliptie to 


) X 0.4-0.8(- 
lisent laeso. apex cuspidate, base attenuate to 
cuneate, drying chartaceous, glabrous, margin flat to 
thinly revolute; secondary veins invisible or visible 

ut flat abaxially, 3 to 6 pairs; domatia absent; 
petioles 1-3 mm. Stipules tubular, glabrous to puber- 
ulent, drying chartaceous, caducous or infrequently 
slowly deciduous, tube 1.2-3 mm, with ribs 2 or 3 
with 1 rib on each side extending from internode rib to 
stipule lobe, sometimes also with narrow wings arising 
from base of petiole, apex with 2 or 3 incisions, 
marcescent, lobes 2(4) or 3(6), 0.2-1 mm, deltate to 


puberulent to hirtellous or glabrescent, pendulous, 
sessile or peduncle to 7 mm; branched portion when 
present corymbiform, 0.5-2.5 X 0.5-4.8 em, lax, 
branched to 1 to 2 orders; bracts deltate or linear, 1— 
5 mm; bracteoles reduced; pedicels 2-6. Flow- 


ers 5-merous, heterodistylous. Long-styled flowers: 


Annals of the 
Missouri Botanical Garden 


calyx cup-shaped, 3-3.5 mm wide, outside glabrous 
or bud sometimes with hair-ring inside, lobes 
0 m, triangular to linear; corolla white, clavate 
in "e Mic open salverform, outside glabrous, tube 
8-10 mm, 1.3-3.5 mm diam., 
middle, 3—4 mm, 


obes 
ile included, filaments inserted at c 


inside villous at ca. 
icd or qu acute; 


corolla tube, ca. 0.1 mm; style 9-11 mm, prre or 
t, stigma 0.5-1 mm. Short- pe flowers: 
g styled except calyx 2.5-3.5 mm wide, 
corolla tube 9.5-10.5 mm, 3.5- 
lobes 3.5—4 mm; 
style 5.5-6.5 mm, 
glabrous, stigma 1.5-2 mm. Drupes violet-black, 
6-9 mm; 
pyrenes spherical to hemispherical, rugose, finely 


pubescen 
similar to lon 
lobes 


5mm diam., 


1-2 mm; 
‘ anthers shortly 
exserted, filaments 1.5-2 mm; 


globose to subglobose or didymous, 6-7 X 


fissured, endosperm entire. 

Distribution and habitat. This natural hybrid is 
known only from Sri Lanka. Here, it can be found in 
wet montane forests at elevations of 1400—1630 m. 


Phenology. This natural hybrid has been collect- 
ed with flowers April through August and has been 
but the months of those 
collections were not noted. 


collected with fruits, 


Discussion. Although this has previously been 
treated as either a biological species ora cas 
ame, Gaertnera Xgardneri sidered a 
al hybrid between G. walkeri and c ternifolia. 
(1967) discussed the numerous 
morphologically irregular, apparently intermediate 

t link these two species and concluded that 
these are hybrids, but he did not formally recognize or 


is here coi 


Van Beusekom 


n the same plants and often on 
stems, and their leaves and stipules of intermediate 
orm. Whether these apparent hybrids are fertile is 
unknown. Gaertnera X gardneri is most often confused 
with G. ternifolia, but differs from that in the presence 
of both opposite and ternate leaves on a single stem, 
broader leaf blades, and salverform corollas. 


Representative specimens examined. SRI NKA. 
Kandy: Fairlawn Estate, Meriyakota, Peak S RD 
aud de 2841 (A, "a Nawara Eliya: Peak Wilderness 

nal Park, lower slopes of Adam’s Peak (Sri Pada), 
Mise 2768 (MO, PDA), 2769 (MO, PDA). 


26. Gaertnera globigera Beusekom, Blumea 15: 


Panjang to Teku, J. A. R. Anderson 
(holotype, SAR!). Figure 10A, B. 


Trees, 2-6 m tall; branches flattened near stem 


apex, otherwise terete, when young glabrous or 


puberulent or pilosulose with indumentum drying 
yellow to white, becoming glabrescent, 3-6 mm diam.; 
p cm, smooth or with 2 longitudinal 

s. Leaf Blades 15-25 X 2.5-5 em, lanceolate, 
ule > or e m lanceolate, apex 
, base acute 


acuminate riaceous, 


ra ing c 
glabrous, margin Tos to thinly pea secondary 
veins prominulous abaxially, 6 to 10 pairs; domatia 
absent; petioles 15-30 mm. Stipules tubular, glabrous, 
rying chartaceous, persistent, tube 5-25 mm, with 
ribs 4, broadly winged, arising below petiole and 


persistent or sometimes fragmenting, lobes 
4, 0.3 deltate. Inflorescences congested- 
cymose, many-flowered, terminal on principal branch- 
es, glabrous to densely hirtellous, sessile or sub- 
sessile, subglobose, 1-3. m, branched to 3 
o 4 orders; bracts deltate, 0.5-2 mm; bracteoles 
reduced. Flowers 5-merous, subsessile, floral biology 
unknown. Calyx cup-shaped, 2.2-3.5 mm wide, 
outside pilosulose to glabrous inside, 
truncate or lobes to 0.5 mm long, triangular; corolla, 
stamens, and stigmas unknown. Drupes unknown. 


Distribution and. habitat. 
Southeast Asia, where it is known from Borneo in the 


This species grows in 
Sarawak (Malaysia) sector. Here, it can be found in 
humid forests at an elevation of ca. 

Phenology. This species has been collected with 
flowers in June and has not been collected with fruits. 
Van Beusekom (1967) reported that 
Gaertnera globigera was found in both laterite forest 


Discussion. 


forests on white sand substrates in 
Bako National Park, Sarawak, Malaysia, on Borneo, 
based on a note on the type label. However, despite 
intensive documentation of the plants of this park, G. 
globigera has not been rediscovered there and overall 
is still only known from the type locality in Rantau 
Panjang Forest Reserve. Both collections seen of G. 
but the 


species can be easily recognized by its relatively long, 


globigera lack mature flowers and fruit, 


lanceolate to narrowly elliptic or narrowly lanceolate- 
elliptic leaves, its distinctive broadly winged stipules, 
and its subglobose inflorescences 
In the protologue, the collector of the type specimen 
was not listed but cited as “? collector, Sarawak 
63”; the name of this collector is J. A. R. Anderson 
ome on records from SAR. 


Additional specimen examined. AYSIA. Sarawak: 
Sibu Distr., 1.43 km ENE of Rantau Panjang Forest Reserve, 
Malcomber 3038 (MO). 


27. rur grisea Hook. f. ex C. B. Clarke, in 
ook. f., Brit. India 4: 92. 1855. age 
grisea (C. B. Clarke) Kuntze, Revis. Gen. Pl. 2 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


425. 1891. TYPE: Singapore, N. Wallich 8389 
(holotype, K!; isotypes, BM!, K) 


Trees or shrubs, 2-8 m tall; branches flattened near 
stem apex, otherwise terete to quadrangular, when 
young densely velutinous to pilosulose with indumen- 

drying pale yellow or gray-white becoming 
glabrescent, 3-5 mm diam.; internodes 2-10 cm, 
smooth. Leaf blades 9-30 x 4-12 cm, elliptic to 
oblanceolate or obovate, apex cuspidate or acuminate, 
base cuneate to obtuse, drying coriaceous, adaxially 
glabrous, abaxially densely velutinous to hirtellous 
with indumentum drying pale yellow or gray-white; 
secondary veins prominent abaxially, 7 to 11 pairs; 
domatia absent or present; petioles 7-18 mm. Stipules 
tubular, drying 
chartaceous, persistent at least on distalmost nodes, 
tube 12-22 mm, with ribs 4, narrowly winged, arising 
below petiole and sometimes extending partway along 
sheath to lobes, apex entire, marcescent or fragment- 
ing, lobes 4, 3-7 mm, deltate to linear. Inflorescences 
cymose, many-flowered, terminal on axillary branch- 
peduncle 3.5-5.5 em; 
branched portion corymbiform, 5-12.5 X 5-10 em, 
lax to congested, branched to 3 to 4 orders; bracts 
4-6 mm; bracteoles 


densely velutinous to pilosulose, 


es, ensely — velutinous; 


narrowly deltate or linear, 
reduced. Flowers 5-merous, unisexual, subsessile. 
Pistillate flowers: calyx cup-shaped, i 
outside densely puberulent to velutinous, with hair- 
ring inside, truncate or lobes to 0.5 mm, broadly 
triangular; corolla white, clavate in bud, when open 
salverform, outside densely puberulent, tube 5-7 mm, 
1.5-2.5 
2.5-4 mm, ligulate to ovate-oblong, acute; staminodia 
shortly exserted, filaments inserted in upper third of 
mm, glabrous, 


mm wide, 


mm diam., inside villous in upper third, lobes 


corolla tube, ca 
stigma 2-2.5 mm. Staminate flowers: similar to 
pistillate except corolla tube 6.5-8.5 mm, 2-3 mm 
diam., lobes 3—4 mm; anthers shortly exserted, 
neo 0.5-1 mm; style rudimentary. Drupes vio- 


let-black, 


pyrenes spherical or hemispherical, rugose, finely 


globose or didymous, 5-7 5-7 mm; 


fissured, endosperm entire. 


Distribution and habitat. This species grows in 
Southeast Asia, where it is known from Peninsular 
Malaysia, Singapore, and the Riau Archipelago of 
Indonesia, near Singapore. Here, it is found in humid 
forests at elevations of 0-150 m 


Phenology. This has been collected 
with flowers January through July and in November 


and with fruits January through 


species 


and December, 


August. 
Discussion. Gaertnera grisea is a locally common 
species. Most of the collections are from Bukit Timah 


in Singapore. Gaertnera grisea can be recognized by 
its pale yellow or gray-white dried pubescence, 
tubular stipules with narrow stipule wings encircling 
the petiole that do not extend through to the top of the 


tube, and truncate to very shortly lobed calyx. 


Additional specimens examined. INDONESIA. Rioux 
[Riau] Archipelago: Ge ea Bünnemeyer 6512 (K, L, 
SING). MALAYSIA. i Sembilan: Passir Panjang, 
Ridley 13337 (BM, SING). "SINGAPORE. Bukit Panjang, 
Ridley 12528 (BM, K, SING). 


28. Gaertnera guillotii Hochr., Annuaire Conserv. 
11-12: 1908, non 
3 


mandry prés d’Analatsara, 3 Oct. 1903, J. Guillot 
36 (holotype, G!; isotype, K!). Figure 9H-N. 


Trees, 2-8(-15) m tall; branches terete to quadran- 
gular, when dry usually dark purple-black, glabrous, 
internodes 0.5—8.5 cm, smooth. Leg 
blades 1.7-13.2 X 0.7—4.6 em, elliptic to oblanceo- 
late or elliptic-oblong, apex rounded then shortly 
base cuneate to obtuse, 


1.5-4 mm diam.; 


cuspidate to acuminate, 
drying chartaceous, glabrous; e veins visible 
and flat to pro: 


rominulous a (9) pairs; 


domatia present; petioles a calyptrate, 
often becoming ee prior to rupturing as tip 
elongates, glabrous outside, strigillose to sericeous 
inside, drying membranous, 
49 mm, with ES 4, rounded, arising above petiole 
and sometimes extending to top of sheath, wings under 


caducous, tube - 


petiole base usually tardily developing after stipule 
tube falls and without connection to stipule ridges, 
apex with 1 incision, lobes 2, 1.2-3.7 mm, del 

Inflorescences cymose, many-flowered, terminal on 
axillary branches, glabrous to puberulent or pilosulose 
with pubescence sometimes in lines; peduncle 0.5- 
2.2 em; branched portion corymbiform, 0.6-7 X 0.8- 
6.2 cm, branched to 3 to 4 orders, lax to congested; 
bracts deltate or linear, 0.8-2 mm; bracteoles trian- 
gular to i 
pedicels absent or up to 1.8 mm. Flowers 5-merous 
heterodistylous. Long-styled flowers: calyx cup- 
shaped, 1.5-2.5 mm wide, 


0.2 mm, triangular; corolla white, clavate in bud, 


rounded, 0.5-1.5 mm, often laciniate; 


glabrous, lobes 0.1- 


when open salverform, outside glabrous, tube 3.5- 
4.5 mm, 1.5-3 mm 
third, lobes 2.5-3 mm, triangular or ligulate, acute; 


diam., inside villous in upper 
anthers included, filaments inserted in upper third of 
corolla tube, 0.3—0.7 mm; style 4.5—5.5 
stigma 0.5-0.7 mm. Short-styled flowers: similar to 
long styled except corolla tube 1.3-3 mm diam., lobes 
2-3 mm; anthers shortly exserted, filaments 2— 
2.5 mm; style 2.4-3 mm, stigma mm. Drupes 
violet-black, subglobose or didymous, 4.5-5.5 X 4.5- 


mm, glabrous, 


Annals of the 
Missouri Botanical Garden 


5.5 mm; pyrenes spherical or hemispherical, rugose, 
finely fissured, endosperm entire. 

Distribution and. habitat. 
Madagascar, where it is known from Antsiranana, 


This species grows in 


Fianarantsoa, Toamasina, and Toliara provinces. 


Here, it is widespread along the east coast of this 
island continent in humid forests at elevations of 0— 
25 m but occasionally up to elevations of 750 m; it 
usually can be found in littoral forests on white sand. 


Phenology. This species has been collected with 
flowers May through December, and with fruits 


January through April and in November and Decem- 
er 


Discussion. Gaertnera guillotii is similar to G. 
obovata, but can be separated by its combination of 
elliptic to oblanceolate or elliptic-oblong leaves with 
at margins, inflorescences with often pubescent axes, 
white flowers, and a habitat on white sand substrates 
and usually in littoral vegetation. The young branches 
usually drying dark purple-black are distinctive and 
helpful to recognize this species but are not exclusive 
to it; similarly colored dry branches are found in 
several other taxa, notably G. ETE var. sphaer- 
ocarpa (e.g., Malcomber et al. 2916, 

A group of specimens from the Tem e in the 
northern part of Madagascar (Antsiranana and 
Toamasina) generally matches Gaertnera guillotii, 
but lacks domatia on the undersides of the leaves 
(e.g.. 


arger calyx lobes, to 1 mm lon 
4195). These specimens are 


Roheschina 
provisionally — 1 here in G. guillotii pending 
further study. 


ean specimens examined. MADAGASCAR. 
anana: Near Antalaha, Imbert 94 (MO, P), Ranjokiny 
RN 8860 (MO, TAN), Unknown Collector RN 8898 (TAN); 


Man: Geay 

ee 291 3 Do. TER), Ralarivohita SF 1103 (P, TAN, 

TEF); Ambodiriana, Cours 1945 AN); Mananara-Nord 

National Park, Raharimalala 317 à d at (MO). 
Toliara: Betw. Antandrainiminty sakona, Ran- 


drianasolo 288 (K, MO, TAN); Fort ae Dey 1 1066 


a 


(P) Humbert 5999 (P) Rabevohitra 2159 (MO, P, TAN), 
McPherson 14146 (MO, P, TAN, TEF) 
29. Wb n hirtiflora Verdc., Kew Bull. 37: 


E: Mauritius. Plaines Wilhems 


= Macabé, 540 m, Aug. 1980, C. Puff 
TU 1/10 (holotype, WU isotype, K!). 


Trees, 5—7.5 m tall; branches flattened to terete, 
1.5-3.5 em, 
1.5-4.5 em, elliptic or 


glabrous, 3-5 mm diam.; internodes 
smooth. Leaf blades 4-11 X 
oblanceolate to obovate, apex obtuse and shortly 


cuspidate or acuminate, base acute to cuneate, drying 


coriaceous or chartaceous, glabrous; secondary veins 
prominulous abaxially, 4 to 8 pairs; domatia present; 
petioles 2-12 mm. Stipules tubular, glabrous to 
densely puberulent, drying chartaceous, persistent at 
least on distalmost 3 to 6 nodes, tube 5-12 mm, with 
ribs 4, narrowly winged, arising below petiole, angling 
toward middle of interpetiolar side of tube then 
extending to lobes, apex entire, lobes 4, 4—6 mm, 
filiform; setae 4 to 10, linear, 3-7 mm. di di QR 
ymose, y-flowered, terminal, puberu to 
oa poses branched port 

corymbiform, 2.5-6 X 1.5-6 cm, branched to 2 to a 
orders, generally congested; bracts ovate, linear, or 
used and trifi —6 mm; bracteoles p to 

vate, 1-2 mm; pedicel absent or to 

merous, heterodistylous. Long-styled flowers: ir 
cup-shaped or campanulate, 2.5-4 mm wide, puber- 
inside d lobes 


ulent outside, mm, 


triangular to ovate; corolla te, clavate in bud, 
outside puberulent, tubular- Es when open, 
externally densely puberulent to velutinous, tube ca. 

mm, 1-1.5 mm diam., inside villous in throat, 
lobes narrowly elliptic-oblong, ca. 5.5 mm; stamens 

inserted in upper part of corolla tube, anthers 
de die included or p exserted; style ca. 
rted. Short-styled 


13 mm, stigmas ca. 2 mm 


flowers: unknown. Drupes ae 
Distribution and habitat. 
Mauritius, where it is found in medium to tall wet 


This species grows in 


evergreen forests (D. H. Lorence, pers. comm.) at 
elevations of 500—700 m. 


Phenology. This species has been collected with 
flowers August through December, but has not been 
collected with fruits. 


Discussion. Gaertnera hirtiflora is similar to both 


the cultivated gardenia; these were collected 
idday (D. pers. comm.), the 


flowers with these RESI ies of nocturnal Rubia- 


at midday Lorence, 


ceae flowers may be nocturnal or crepuscular as 
well. 


diuo specimens examined. MAURITIUS. Black 
River National Park, S. Malcomber 2957 (MO), 2934 
(MO), Vaughan MAU 13042 (MAU); Plaine Champagne, 
Lorence 1541 (MO). 


30. Gaertnera hispida Aug. DC., Bull Herb. 
1: 585. 1901. TYPE: Madagas- 


car. Toamasina: Maroa [Maroantsetra], A. Moc- 


Boissier, sér. 2, 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


A-F. Gaertnera hispida Aug. DC. —A. Flowering branch. 
. Short- styled flower in cross section. —E. Long- styled flow: 


—B. Portion of stem with petiole bases and stipules. 


er. —F. Long-style ed fl 


ame l-cm scale 


cale; J-M to s 


A-F based v on Moise 11; G, I-M based on Moise 10; H based on Malcomber 2830. 


querys oe 2 S designated here, G-DC!). 
Figure 1 


Trees, 2.5-12 m tall; branches terete, hispid with 
indumentum drying red-brown to brown, 2-5 m 
; internodes 1.2-8.5 cm, smooth. Leaf blades 
6.1-16.1 X 1.1-3.1 em, narrowly elliptic to elliptic- 
shortly 
acuminate, base acute or cuneate, drying chartaceous, 


oblong or oblanceolate, apex acute to 
adaxially glabrous or p hispid to hirsute on 
osta and sometimes se ary veins and lamina, 
abaxially hispid to koe with indumentum denser on 
veins and drying reddish brown to brown; secondary 
veins prominulous abaxially, 9 to 16 pairs; domatia 
absent; petioles 2-10 mm. Stipules calyptrate, densely 
hispid to strigose, drying membranous, caducous, tube 
1 without ribs, apex with 1 incision, lobes 
or 4, 0.5-2(-4) mm, deltate or linear. Inflorescences 
cymose, many-flowered, terminal on axillary branch- 
es, densely villous to villosulous or pilosulose, sessile 
or aoe e 5 em; branched portion corymbiform, 
1.8-8.5 
to rather ibd. bracts linear to ligulate or deltate, 
1.1-6 mm, 


4—8.5 cm, branched to 3 to 5 orders, lax 


sometimes glabrous above; bracteoles 


reduced; pedicels absent or to 2.5 mm. Flowers 5- 
Long-styled flowers: calyx 
puberulent outside, 


ous, heterodistylous. 
cup-shaped, 1.7—4 mm wide, 
glabrous inside, truncate or lobes to 0.9(-1.5) mm, 
triangular; corolla white, clavate in bud, when open 
salverform, outside glabrous to puberulent, tube 8- 
9.5 mm, 2-3.5 mm diam., inside villous at ca. middle, 
lobes 2-3.5 mm, "— or ligulate, acute; anthers 
included, filaments inserted in upper third of corolla 
tube, ca. 0.4 mm; style 9-11 mm, glabrous, stigma 
0.7—2 mm. Short-styled flowers: similar to long styled 
except calyx 1.5-3.5 mm wide; corolla puberulent 
outside, tube 6.5-11 mm, 1.5-3 mm diam., lobes 1.5— 
4 mm; anthers shortly exserted, filaments 3-4.5 mm; 
tyle 3-6 mm, stigma 1.5-2.5 mm. Drupes toed 
black. s to subglobose or didymous, 8-14 X 
; pyrenes spherical or hemispherical, rugose, 
finely f fissared, endosperm entire. 


Distribution and habitat. This species grows in 
Madagascar, where it has been found in the province 
of Toamasina in Mananara-Nord and Masoala Nation- 
al Parks. Here, it can be found in humid forests at 


elevations of 0—300 m 


Annals of the 
Missouri Botanical Garden 


Phenology. This species has been collected with 
flowers in October, and with fruits January through 
April and in December. 


Discussion. Gaertnera hispida can be recognized 


by its calyptrate stipules that dry membranous, its 
hispid indumentum dry 


and its generally lax corymbiform inflorescence. This 


rying reddish brown to brown, 


species is similar to G. phanerophlebia, which differs 
in its well-developed calyx lobes 2-5 mm long. 

Two syntypes were cited in the protologue of 
Gaertnera hispida, A. Mocquerys 155 (G-DC!) and A. 
Mocquerys 167 (G-DC!). The former is selected here as 
lectotype because it is a more complete and 
exemplary specimen. 


Representative specimens examined. MADAGASCAR. 
Toamasina: ananara National Park, 6.4km W of 
ntanambe, Malcomber 2897 (MO, TEF), Morat et al. 858 
(P), 8613 (MO, P); Masoala National Park, Malcomber 2824 
(MO, TEF), 2817 (MO, TEF), 2818 (MO, TEF), Schatz 3314 
(K, MO, P, TAN), 3346 (K, MO, P, QRS, TAN) 


31. Gaertnera humblotii Drake, Bull. Soc. Bot. 

52. 1899. TYPE: Madagascar. NE 

Madagascar, H. Humblot 655 (holotype, P!; 
isotypes, MO!, P!, WU?). 


Trees or shrubs, 2-10 m tall; branches flattened 
near stem apex, otherwise terete, glabrous, 1.5-4 mm 
EE internodes 1.5—7.5 cm, smooth. Leaf blades 4— 

6.5 X 1-5.5 em, elliptic-lanceolate or oblanceolate 
to ens apex cuspidate to acute, base attenuate or 
cuneate, drying coriaceous, cee secondary veins 
barely visible and flat to prominulous abaxially, 7 to 
10 pairs; domatia absent; petioles 3-20 mm. Stipules 
tubular, glabrous, drying chartaceous, deciduous or 
usually persistent on at least distalmost 2 to 5 nodes, 
tube 3-10 mm, with ribs 4, rounded to na 
winged, arising below petiole and sometimes extend- 


rrowly 


ing to lobes, apex entire, marcescent, lobes 4, 0.5— 
1.5 mm, filiform. Inflorescences cymose, many-flow- 
ered, terminal on axillary branches, glabrous or 
puberulent; peduncle 1.5-8.8 cm; branched portion 
corymbiform, 2.5-9(-14.5) X 3-11 em, branched to 3 
to 5 orders, lax to somewhat congested; Linie deltate 
white; 


bracteoles triangular, 1-2 mm, white; pedicels absent 


to narrowly spatulate or trifid, 5-35 mm, 


or up to 5mm. Flowers 5-merous, heterodistylous. 
Long-styled flowers: calyx campanulate, 2-3 mm wide, 
glabrous, lobes unequal, 5-7 mm, narrowly spatulate 


to narrowly ue ed white; corolla white, 


pad 


clavate in pe Mi oie to 
salverform, nen prd tube 10-17 mm, 

3 mm diam., inside villous in upper third, lobes 3.5— 
4 mm, ligulate to linear, acute or rounded; anthers 
included, filaments inserted in upper third of corolla 
tube, 0.2-1 mm; style 10-18 mm, glabrous, stigma 


1.4-2 mm. Short-styled flowers: similar to long styled 
—6 mm, linear; 
3.7—5 mm; 
anthers shortly exserted, filaments 2-4 mm; style 5— 
7 mm, stigma 4-5 mm. 2 violet-black, n 
bose or didymous, 6-8 mm; pyrene 
spherical or hemispherical, + rugose, finely fissured, 
endosperm entire. 


Distribution and habitat. 
Madagascar, where 


This species grows i 
it is known from the province of 
Toamasina. Here, it is found in humid forests at 


elevations of 0-600 m 


Phenology. This species has been collected with 
flowers in February and with fruits February through 
April. 


Discussion. Gaertnera humblotii is an attractive 
species with distinctive tubular stipules and relatively 
large, showy, white, mostly narrowly spatulate to 
narrowly kd blong inflorescence bracts 

calyx lobes. It is similar in general aspect and many 
details to G. raphaelii Malcomber; see additional 
comments under that species. Gaertnera humblotii 
may hybridize with G. phanerophlebia; see comments 


under that species. 


examined. MADAGASCAR. 
Toamasina: oala ara Gis 2620 (MO, P, 
TAN); Dac Peak o of Andrambola- 
hykely, Andranampony, = urs 4509 ri TAN) Marojejy 
Integral Reserve, ere 320 (K, MO); Maroantsetra, 
Schaiz 1875 (MO, P, TAN). 


id how 


Gaertnera ianthina Malcomber, sp. no 

: du pl Antsiranana: Marojejy Na- 

2 on summit trail, 

eos 4974 S'E. 80 o “1000 m, 17 Sep. 1997, 

S. T. Malcomber 2773 (holotype, MO!; isotypes, 

Al, BR!, G!, K!, P, PRE!, TEF!, WAG!) 
Figure 5G, H. 


Haec bud Gaert tnerae phanerophlebiae Baker similis, 
ed ab e ns uadrangulari, 
n sicco griseoalbo atque corolla pallide roseo- 
purpurea distinguitur. 


sectione transversali 


Trees or shrubs, 3—10 m tall; branches terete to 
quadrangular, when young densely strigose to pilosu- 
lose with indumentum drying gray-white, becoming 
glabrescent, 1.5— diam.; internodes 2.5—4 cm, 
smooth. Leaf blades 6-20 X 2-7 cm, elliptic to 
elliptic-oblong or obovate, apex cuspidate or shortly 
acuminate, margins crisped, base obtuse to usually 
cuneate or acute, drying chartaceous, adaxially 
glabrous or densely pilosulose on costa and sometimes 
secondary veins, abaxially strigose or hispidulous to 
hispid on principal veins or throughout with indu- 
mentum drying gray-white; secondary veins prominu- 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


lous abaxially, (8 to)14 to 20 pairs; domatia absent; 
petioles 4-9 mm, furrowed and with 2 
Stipules calyptrate, densely pilosulose to villosulous, 


teral ridges. 


drying membranous, caducous, tube 7-40 mm, with 
ribs 4, narrowly winged, arising beneath petiole ps 
sometimes extending to lobes, apex wit or 

lobes 


li b AE cymose, 


incisions, 4, 2-3 mm, deltate or o 


many-flowered, terminal on 
principal and/or axillary branches, villous or pilosu- 
lose to strigose or glabrous, sessile or peduncle to 
4 cm; branched portion Ms oa 2.5-6(-8) X 2- 
7(-9) em, branched to 3 to 4 orders, lax to congested; 
bracts ligulate, linear, or trifid, 2-30 mm; bracteoles 
ligulate, 0.1-2 mm; pedicels absent or to 2.5 mm. 
Flowers 5-merous, heterodistylous. Long-styled flow- 
ers: calyx cup-shaped, 1.5-2.5 mm wide, outside 
glabrous or strigillose, with hair-ring inside, lobes 
0.5-2.5 mm, triangular, ciliate; corolla pale pink to 
purple, clavate in bud, when open salverform, outside 
inside 
villous in upper third, lobes 3—4 mm, triangülar to 


glabrous, tube 7-9 mm, 2-2.5 mm diam., 
ligulate, acute; anthers included, filaments inserted in 
upper third of corolla tube, ca. 0.4 mm; style 7.5— 
8 mm, glabro stigmas mm. Short-styled 
flowers: similar to long styled except lobes linear; 
corolla tube 9-10 mm, 1.5-3 mm diam.; anthers 
shortly exserted, filaments 0.8-1 mm; style 5- 
stigmas 1-1.5 mm. Drupes violet-black or blue, 


globose or didymous, 6—7.5 X 7-10 mm; pyrenes 
spherical or hemispherical, faintly rugose, deeply 


fissured, endosperm entire. 


Distribution and habitat. This species grows in 
s known from the 


provinces of Antsiranana and Mia Here, it 


northern Madagascar, where it 
grows near Manongarivo, Tsaratanana, and bd 
in humid forests at elevations of 300—2000 


Phenology. This species has been collected with 
flowers January through May and in December, and 
with fruits January through June and in November and 


December. 

Discussion. Gaertnera ianthina is similar in gen- 
eral aspect to some plants of G. phanerophlebia, but 
differs in its generally quadrangular stems, indumen- 
tum drying grayish white, generally lax inflorescences, 
pink to purple flowers, and shorter calyx lobes 


atypes. MADAGASCAR.  Antsiranan Andap: 
ea RN 8044 d TAN); WS Ambodihasina 
Cours 3615 (P, TAN); Ankobahina Peak, 
Humbert 22024 (P). 21 997 (P), 22045 (P): Andapa, Marovato, 
known Collector SF 21632 


Nature Reserve, Lewis 1318 
; Ankaizina, Decary 1949 (P); Lokoho, near 
Feci ' Humbert 22796 (P), 22817 (P); Manongarivo 
Nature Reserve, Perrier de la Báthie 3836 (P), Malcomber 


2710 (MO, P, TAN), Miller & Lowry 4050 (MO), Nicoll 616 
MO, P, TAN), Miller 4189 (M E P), Humbert 23106 (P), 
Lewis 1253 (K, MO, P, TAN), 
& Randrianasolo 4667 (MO), ae 22429 (P); Massif du 


Tsaratanana, haut bassin de 


o 


evarano, crete dans le 
bassin superieur du ruisseau de Befosa, T SF 24978 
(P, TEF); Sambava, Zamanivato RN 2 (P, TAN); 
Tsaratanana Nature — ene T RN 4719 
(P, TAN). Mahajanga: Tsararatanana Massif, Magindrano up 
S ridge of emen eis ay 11575 (TAN). 


Adansonia 12: 237. 

Madagascar. amas S Mariam, 
nid Jan. 1848, L.-H. Boivin 1778 (lecto- 
type, i G). Fig- 
ure 8 


33. Halu inflexa Baill., 
E: 


designated here, P!; isotype, 
F-L 


cue ur Hochr., Annuaire Conserv. Jard. 
e 11-12: 111. 1908. wie oo nas 

Bremek, Verh. Kon. kad. ch., Afd. 

Naturrk., section 2, 54: 148. m S iles. non ii 

guillotii Hocht: 1908. TYPE: Madagascar. Toamasin: 

Vatomand Oct. 1908, J. Guillot 27 (holoipe, Ct 

isotype, K!). 

Trees, 1-6 m tall; branches flattened to terete, 
glabrous, 1.5—5.5 mm diam.; internodes 1 E 
smooth. Leaf blades 4.5—24 X 1.2-8 em, narrowly 
elliptic, elliptic, or oblanceolate, apex acuminate, 

ase acute or cuneate, drying Sane glabrous; 
secondary veins prominulous abaxiall pairs; 
domatia absent; petioles 4-55 mm. Stipules tubular, 
glabrous, drying chartaceous or membranous, cadu- 
cous or E persistent on distalmost 1 to 3 
nodes, tube .5 mm, with ribs 4, narrowly winged, 
arising vn or sometimes above petiole, often 
angling to connect in middle of TO ue 
then diverging n extending to lobes, apex entire, 

, 0.5-11 mm, deltate or Sfr 
a. to many-flowered, 


marcescent, lobes 

Inflorescences cymose, 
terminal on axillary and/or supra-axillary branches, 
pendulous, glabrous or puberulent; peduncle 1.5- 
6 cm nched tion narrowly cylindrical or 
1.1-3(-6) em, branched 


to 2 to 3 orders, lax; bracts deltate to linear, 1-14 mm; 


narrowly pyramidal, 3.5-7 X 


bracteoles triangular to ligulate, 0.5-1 mm. Supra- 
axillary branches pendulous, up to 40 cm long, 0.8— 
1.4 mm diam., glabrous; leaves 2-10 X 0.3-2.5 em, 
narrowly eii or linear, base attenuate; d 
4-12 mm. Flowers 4- 
merous, ER M d subsessile. ri fow- 


veins o 9 pairs; petiole 
ers: calyx cup-shaped, 1.2-1.5 mm wide, outside 
men or puberulent, glabrous inside, truncate or 
0.3 mm, 
Br ide open salverform, outside glabrous, tube 3— 
3.5 mm, 0.9-2 mm diam., 
third, lobes 1.5-2 mm, vm or e acute; 


= 
E 


triangular; corolla white, clavate in 
inside villous in upper 


anthers included, filaments inserted at c 
corolla tube, ca. 0.5 mm; style 2.5-3 mm, glabrous, 


Annals of the 
Missouri Botanical Garden 


stigma 0.5-0.7 mm. Short-styled flowers: similar to 
long ipe except a 1-1.5 mm wide; corolla tube 
3.5-4.5 m diam., inside villous at ca 
middle, aie 1.5- 25 mm; hor included, fila- 
ments inserted in upper third of corolla tube, 1.5— 
2 mm; style 0.8-1.3 mm, stigma 0.8-1.3 mm. Drupes 
violet-black, subglobose or didymous, 6—6.5 X 5- 

m; pyrenes spherical to hemispherical, rugose, 
finely fissured, endosperm entire. 


Distribution and. habitat. 
Madagascar, where it is known from the provinces of 


This species grows in 
Antsiranana and Toamasina. Here, it can be found in 
humid forests at elevations of 150—500 m. 


Phenology. This eras has been collected with 
hrough November and with fi 
February through March. 


flowers Tus ruits 


Discussion. With narrowly cylindrical or pyrami- 
dal inflorescences often borne at the apex of long, 
pendent supra-axillary SEEN Gaertnera inflexa is 
a distinctive species withi ascar. It is similar 

o G. diversifolia of B s fux but differs from 
ds in its bisexual 4-merous flowers. Gaertnera 
inflexa is also similar to G. cardiocarpa; see comments 
under that spec 

The plants sneha in this species here exhibit 
considerable variation in length of the stipule lobes 
and inflorescence shape and size. Collections from the 
southern part of its range (e.g., Guillot 27, type of 
Psychotria guillotii) might be better considered a 
distinct subspecies or fave but ee e. 
inflexa 


urrent paucity of stable ME 


are provisionally included here in Gae 


Two syntypes were cited in the protologue of 
Gaertnera inflexa, L.-H. Boivin 1778 (G!, P!) and J. M. 
C. Richard 5 (Madagascar, Bay of Antongil, 1837, P!). 
The former is selected as lectotype here because it has 
a duplicate at another herbarium. The duplicate at P 
is chosen as the lectotype because it is deposited in 
the herbarium where the author of the name worke 


oo specimens ned. MADAGASCAR. 
Antsiran: ntalaha, ENIMS 521 (TAN). Toama- 
sina: Ivo a, o nanara Pa Park, 
Malcomber 2886 


(MO), 2900 (MO, TEF), 2901 (MO, TEF), 
Rafamaniananisoa OUEM 31 (K, MO); eae Nord, 
Raharimalala 1502 e 2151 [o maj us Schatz 
1661 (MO) Sahavolamena, Ivongo, Unknown 

Collector SF 11060 (MO, TEF), Capuron m 25805 (P, TEF). 


34. Gaertnera junghuhniana Miq., Fl. Ned. Ind. 2: 
383.1 


Revis. Gen. P 


956. 2 (SU (Miq.) Kuntze, 

2: 425. 1. Gaertnera vaginans 
subsp. ene e Beusekom, Blumea 
15: 388. 1967 [1968]. TYPE: Indonesia. Suma- 
tra, F. W. Junghuhn s.n. (lectotype, designated 


by van Beusekom, 1967 [1968]: 385, L!; isotype, 


U not seen). [SYNTYPE: J. E. Teijsmann s.n. 
(syntype, BO not seen, GH!).] Figure 6A—F. 


Fl. Ned. Ind. 2: 382. 1857. 


Gaerinera zollingeriana Miq., 


Sykesia Pop i. (Miq.) Kuntze, Revis. Gen. Pl. 2: 
426. 1891. TYPE: Indonesia. Java, H. Zollinger 3051 
(holotype, "s isotypes, A!, G!). 


7, nom. nu 

Gaerinera koenigii var. a (Benth.) C. B. Clarke, 
in Hook. f., Fl. Brit India 4: 91. 1883. ncs 
oxyphylla (Benth.) e Revis. Gen. PI. 

1891. Gaertnera acuminata var. oxyphylla qu 
Ridl., Fl. Malay. Penin. 2: 428 

Singapore, N. Wallich 8374 (holotype, K!; isotypes, 
M!, K! 


Gaerinera acuminata Benth., J. Proc. Linn. Soc., Bot. 1: 112. 
1857. Sykesia eas (Benth.) Kuntze, Revis. Gen 
Pl. 2: 425. 1891 E Singapore; N. Wallich 8342 
(holotype, K!; Bs * 

Gaerinera oxyphylla var. angustifolia Ridl. , Fl. Malay. Penin. 
2: 428. 1932. TYPE: Malaysia. Kedak Kedah Peak, 

B A H. C. Robinson 5963 

(holotyp. e, SING). 

Gaerinera. ae) Ridl., Bull. Misc. Inform. Kew 1934: 
1 - TYPE: Malaysia. Sabah: Sandakan, C. V. 
Creo sn: (holotype, K*; isotype, K!). 

Gaerinera quus Kerr, Bull. Misc. Inform. Kew 1940: 180. 

YPE: Thailand. Trang: Kao Soi Dao, 300 m, A. 
Kerr 19137 (holotype, K!; isotypes, BM!, L5). 


Trees or shrubs, 1-10(-15) m tall; branches terete 
to flattened or quadrangular, glabrous to puberulent, 
1-6 mm diam. internodes (0.5-)1.3-7(-10) cm, 
smooth. Leaf blades 3-24 X (0.7-)1-9.5 cm, lance- 
olate to narrowly lanceolate, elliptic-oblong, elliptic, 
or oblanceolate, apex acuminate or cuspidate, base 
cuneate to acute (rounded), drying chartaceous, 
adaxially glabrous, abaxially glabrous to iam 
or pilosulose on principal veins with i 

rying gray-white; seconda 

abaxially, 3 to 9(to 11) pairs; domatia absent or 
present; petioles 2-25 mm. Stipules tubular, glabrous 
ent, drying chartaceous, caducous or 
fragmenting, tube (3—)8—23 mm, with ribs none or 4, 
narrowly to broadly winged, arising below petiole and 
sometimes extending to lobes, apex entire or with 1 or 
2 incisions, marcescent, lobes 4, 1.5—7 mm long, 
deltate. to many- 
flowered, terminal on axillary branches, glabrous to 


Inflorescences cymose, several- 
pee puberulent or pilosulose, sessile or peduncle 

6.5 em; branched portion corymbiform to pyrami- 
ia 1: i 15(-18) X 1.5-17(222) cm, lax, Pics to 


orders; 


N 


bracts deltate or trifid, 

dd reduced; pedicels absent or to 2 mm. 
Flowers 5-merous, unisexual. Pistillate flowers: calyx 
cup-shaped, 2-3 mm wide, outside glabrous or 
puberulent, with hair-ring inside, truncate or lobes 
to 0.4(-0.7) mm, triangular; corolla pale green or 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


white, clavate in bud, when open salverform, outside 
glabrous or puberulent, tube 2.5-5 mm, 1.5-2.2 mm 
1.5-3 m 


staminodia view 


diam., inside villous in upper third, lobes 


itam: to ligulate, acute; 
filaments inserted in upper third of corolla tube, ca. 
0.2 mm; style 3-6 mm, glabrous, stigma 2-2.5 mm. 
Staminate flowers: similar to pistillate except corolla 
tube 2-5 mm, diam.; anthers shortly 
exserted, filaments 0.5-2 mm longs pistillode with 
style portion absent or ca. 1 mm, glabrous, stigma 
absent or ca. 0.5 mm. Drupes violet-black, didymous 
or globose, 5-7 X 5-8 mm; pyrenes spherical or 
finely fissured, endosperm 


hemispherical, rugose, 


entire. 


Distribution and habitat. This species grows in 
southeastern Asia, where it is known from Thailand, 
Peninsular Malaysia, Sumatra (part of Indonesia), 
Borneo in the Brunei, Kalimantan (Indonesia), and 
Sarawak (Malaysia) sectors, and Sulawesi (part of 
Indonesia). Here, it is found in humid forests at 


elevations of 0-15 


Phenology. This species has been collected with 
flowers and fruits throughout the year. 


Discussion. This species belongs to the Gaertnera 
vaginans complex; see also the discussion of that 
group for related species and their distinctions. This 
neluded by van Beusekom (1967) in his 

ns, as G. vaginan. 


subsp. junghuhniana, but uc. species are consid- 


species was 1 
broad circumscription of G. v 


ered separate here as discussed in the introduction. 
ome specimens included by him in G. vaginans 
subsp. junghuhniana are treated here in G. alstonii, G. 
aphanodioica, G. belumutensis, G. capitulata, G. 
kochummenii, G. ramosa, and G. sralensis. Gaertnera 
junghuhniana as circumscribed here can be distin- 
guished by its tubular Le that dry chartaceous 
and closely surrou m, its several- to 
flowered cymes, and its one green or white wee 
with either a well-developed androecium and relictual 
gynoecium (staminate) or well-developed gynoecium 


and relictual androecium (pistillate 

dee dicum specimens examined. BRUNEI [DARUS- 
SALAM]. Belait: Badas Reserve, Malcomber 5001 (MO), 
3002 (MO), 3003 (MO), 3004 (MO), 3005 (MO), 3006 (MO). 


unei-Muara: Ri 


Kostermans 9252 (K, L, SING). Kilian dicun East 
Kalimant: i 


1 an]: Wanariset-Handil IL, Van oue 6062 (K); 
alili, Larona, 2^4í 121710'E, Meier 11241 ) 
Celebes. Sulawesi: betw: Soroaka & Wawondala, Van 


Balgooy 4052 (A, K, L). Sumatra. Aceh: Asahan, Poelock, 
Rahmati Si Boeea 5722 (A, G, K, L). Bangka: Lobok-besar, 
Kostermans 238 (A, BM, K, KEP, L, SING). Riau: Bukit 


Karampal, 0°46’S, 102°32’E, 100 | m, Burley 1157 (A, E, K, 


Palembang, Bigin Telok, Forbes 3214 (BM, K, L, SING, WU). 
MALAYSIA. Sabah: Sandakan, Sepilok Reserve 5, Nawas 
A. 834 (A, K, KEP, SING). Sarawak: Bako Park, Ashton 
17970 (K, L, SAR), Malcomber 2028 (MO); Lambir Park, 
Rena George S. 40507 (E, K, KEP, L, SAN, SAR). Selangor: 
T Gombak Reserve, a FRI 32515 (A, K, KEP, 
SAN). Terangganu: Bukit Bauk, Malcomber n (MO). 
NUN Trang: Kao Soi Dao, Kerr 19137 


a 


35. Gaertnera kochummenii Malcomber, sp. nov. 
TYPE: Malaysia. Terengganu: 10th Mile, Dun- 
guan-Bukit Besi Rd., Compt. 12, Bukit Bauak 
Forest Ti 4^46'N, 103° 10'E, 18 June 
1967, K. M. Kochummen FRI 2387 (holotype, 
KEP!; Paige p 


Haec species Gaerinerae junghuhnianae Miq. similis, sed 
ab ea planta in sicco aurantiaca, inflorescentia congesta 
Wr in atque lobulis calycinis bene evolutis usque ad 
3.5 mm longis distinguitur. 

Shrubs, to 1 m tall; plants drying with orange cast; 
branches terete, when young puberulent with indu- 
mentum drying brown, becoming glabrescent, 2-3 mm 
diam.; internodes 3.2-7 cm, smooth. Leaf blades 6.5— 
17 X 1.5-5 em, (narrowly) elliptic, apex acuminate, 
base attenuate, drying chartaceous, adaxially gla- 
brous, abaxially puberulent with indumentum drying 
reddish to orange; secondary veins distinct abaxially, 
7 to 9 pairs; domatia absent; petioles 3-5 mm. Stipules 
tubular, puberulent, drying chartaceous, tube 6— 
12 mm, with ribs 4, narrowly winged, arising below 

etiole and extending to s with 2 incisions, 
4, , deltate 
Inflorescences o — — ter- 


marcescent, lobes to linear. 


minal on axillary branches, pilosulose, sessile, 


subglobose, 1.5-2.5 X 
to 1 to 2 orders; bracts deltate, 0.5—2 mm, sometimes 


2-3 cm, congested, branched 


glabrous; bracteoles reduced; pedicels absent or to 
2mm. Flowers 5-merous, floral biology unknown. 
Calyx campanulate, 2.5-3.5 mm wide, outside puber- 
ulent or pubescent, glabrous inside, lobes 1-3.5 mm, 
ovate; corolla, unknown. 


stamens, and stigmas 


Immature drupes globose or subglobose, 5-7 X 5- 


Distribution and habitat. 
southeastern Asia, 


This species grows in 
where it has been found in 


The habitat and 


where it grows have not been recorded. 


Peninsular Malaysia. elevations 


Phenology. This species has been collected with 
flower buds in June; it has not been collected with 


mature flowers or fruits. 
Discussion. This is a poorly known but distinctive 
species. Gaertnera kochummenii can be recognized by 


Annals of the 
Missouri Botanical Garden 


the distinctive orange color of dried specimens; its 
pubescent branches, leaves, and stipules; its congest- 
d subglobo: its well-dev 
calyx lobes. The mature flowers and fruits are 
n honor of K. M. 
99), who collected the type 


se inflorescence; and its eloped 
unknown. This WDR is named i 
Kochummen (1931- 
specimen and un many important contributions 
to Malesian botany. This species belongs to the 
G. vaginans complex; see also the discussion of that 
group for related species and their distinctions. 


36. s letouzeyi Malcomber, sp. nov. 
YPE: Cameroon. Betw. e & Akoumayip 

River, 20 km W of M , 2 June 1975, R. 

e ee p isotypes, BR!, 
, MO!, P, WAG). 


Haec species Gaerinerae paniculatae Benth. similis, sed 

ab ea foliis oblongo-ovatis atque stipulis manifeste quad- 

ibus apice semel fissis in quoque latere tubi ac sub 

petiolo alis duabus longitudinalibus prominentibus munitis 
distinguitur. 


Shrubs, 
apex, otherwise terete to subquadrangular, 
ent with indu- 


4—5 m tall; branches flattened near stem 

when 

young branches glabrous 
yel 


mentum drying low, es ing glabrescent, 5— 
m diam.; us 4.5-16 cm, smooth 
des 16- 33 7.5-13.5 cm, elliptic-oblong 


to 
shortly cuspidate 
drying 
coriaceous or chartaceous, glabrous; secondary veins 


ovate, apex rounded then abruptly 
or acuminate, base cuneate or obtuse, 
prominulous abaxially, 7 to 12 pairs; domatia absent; 
petioles 10-45 mm. Stipules tubular, glabrous or 
puberulent, drying chartaceous, persistent on distal- 
with ribs 


most nodes or nun tube 5 mm, 
win ing below petiole and 


4, narro 
extending to lobes, apex oath 1 incision, lobes 2, 1- 


wly w 


7 mm, deltate. Inflorescences cymose, many-flowered, 


terminal on principal and/or axillary branches, 


densely pilosulose to glabrous; peduncle 2.5- 

2 cm; branched portion corymbiform, 7-23 X 
, branched to 3 to 5 

n or trifid, 3-10 


pedicels absent or up to 3.5 mm. Flowers 5-merous, 


orders, lax; bracts 


mm; bracteoles reduced; 
floral biology unknown. Calyx cup-shaped, 2.5- 
m wide, pale pinkish green, outside glabrous 
0.3-2 mm, 
triangular to oblong or rounded; corolla in bud white, 


or puberulent, with hair-ring inside, lobes 


clavate and densely puberulent outside with tube to 
m es to 3 mm, mature 

and stigmas unknown. Drupes violet-black, globose or 

subglobose 


spherical or hemispherical, rugose, finely fissured or 


corollas, stamens, 
or didymous, 6-9 X 6-9 mm; pyrenes 


with pale brown striations, endosperm entire. 


Distribution and. habitat. 
Central Africa, where it has been found in Cameroon. 


This species grows in 


Here, it is found in lowland humid forests, but the 
elevation where it grows has not been note 


Phenology. This species has been collected with 
flowers in May and June and with fruits in July. 


Discussion. Gaertnera letouzeyi is similar t 
lowryi of Madagascar, but differs in its elliptic- "m 
to ovate leaves with seven to 12 pairs of secondary veins 
and its pale pinkish green calyx tubes. This species is 
r of the botanist René Letouzey (191 
1989), who collected the type specimen in addition to 


named in hono 


numerous other notable plants from Cameroon. 


Paratypes. CAMEROON. e Mamfe rd., 
Nyang beme river, 20 km E of Mamfe, 31 v 
1975, Letouzey 14153 (P); Korup nd Park, Kenfack 763 
(MO). 


37. Gaertnera leucothyrsa (K. Krause) E. M. A. 
Petit, Bull. Jard. Bot. État Bruxelles 32: 186. 
1962. Basionym: Psychotria leucothyrsa K. 

rb. Syst. 57: 47. 1920. TYPE: 

Belgian Congo [Democratice Republic of the 


Krause, Bot. 


May 1908, W. J. Mildbr 

(lectotype, designated by Petit, 1962: 186, HBG 
not seen; isotypes, Bt, HB not seen). [SYNTYPE: 
G. W. J. Mildbraed 3206 (B1).] 


Gaerinera parvipaniculata E. M. A. Petit, Bull. Jard. Bot. État 
Bruxelles 29: 52. 1959. TYPE: 
D Republic of the Congo]. 

angambi, ca. 470 m, 25 June 1938, J. Louis 9979 
[ros pe, BR!). 


Trees or shrubs, 0.8-2.5 m tall; branches terete, 
glabrous, 1.5-5 mm 
smooth. Leaf blades 3.2-17 X 1.3— 


oblanceolate or obovate, apex cuspidate or acuminate, 


diam.; internodes 0.4—6.5 cm, 


5.5 em, elliptic to 


base attenuate or cuneate, drying e e 
glabrous; secondary veins prominulous abaxially, 
to 7 pairs; domatia absent, petioles 5-10 a 


tubular, glabrous, drying membranous, a tube 
10-30 mm, with ribs 4, narrowly winged, arising 
below petiole and sometimes extending to lobes, apex 
with 1 incision, marcescent, lobes 4, 1—4 mm, linear 
to filiform. Inflorescences cymose, several- to man 

owered, 
pilosulose or puberulent, 
4 cm; branched portion subglobose or corymbiform 
0.9-5 X l-4cm, branched to 1 to 4 orders, 


congested; bracts linear to lanceolate, 1—6.5 mm, 
glabrous; bracteoles reduced; pedicels absent or to 
n owers 5-merous, heterodistylous. Long- 


styled flowers: calyx cup-shaped, 1.5-3 mm wide, 
glabrous or puberulent outside, with hair-ring inside, 
obes 1—4 mm, linear to triangular or lanceolate; 


corolla white, clavate in bud, when open infundibu- 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


liform or salverform, outside glabrous, tube 3.5- 
5.5 mm, 1.5-2.5 mm diam., inside villous in upper 
included, filaments inserted in 
upper third of corolla tube, ca. 0.3 mm; style 4.5— 
5.5 mm, glabrous below and pubescent near apex 
stigmas 0.5-1. 


anthers 


mm. Short-styled flowers: similar to 
long styled except corolla tube 4-5 mm, 2.2-2.6 mm 
diam., lobes 3—4 mm, ligulate or ovate-oblong, acute; 
anthers fully exserted, filaments 1.5-2.5 mm; style 
2 mm, glabrous, stigmas 1—1.4 mm. Drupes 
6-10 x 6- 


10 mm; pyrenes spherical or hemispherical, smooth, 


violet-black, globose or subglobose, 


endosperm entire. 


Distribution and habitat. This species grows in 
Central Africa, where it has been collected in the 
Democratic Republic of the Congo and Gabon. Here, it 
is found in humid forests at elevations of 470-1 


Phenology. This species has been collected with 
flowers June through October and with fruits January 
through April and July through December. 


Discussion. | Gaertnera leucothyrsa is similar to and 
may be confused with G. bieleri in Central Africa, but 
G. leucothyrsa differs from that species in its glabrous 
stems and stipules and its caducous stipules with one 
incision. Gaertnera leucothyrsa is also similar in 
general aspect to G. longivaginalis var. bracteata; see 
comments under that species. 

As noted by Petit (1959b), the stipules of this 
species are distinctive but rarely observed on 
herbarium collections. Petit described Gaertnera 
vu PAL n his review of African Gaertnera 


(19592) a r this 


species me a review of African Psychotria (Petit, 


then discovered an older name fo 


1962). He selected as lectotype the only specimen he 
located of Krause’s original materials. 


Additional specimens examined. DEMOCRATIC RE- 
PUBLIC OF THE CONGO. Befale, Ifale River, Evrard 
2893 (BR, P, WAG); Epulu, Mambasa, Ituri Forest, Hart 519 
(BR) Liengola 157 (WAG) E (Monkoto Territory), 
Evrard 2827 (BR, K, WAG); Kole (Sankuru), Lebrun 6332 
(BR, K), 6414 (BR, FHO); AM Village (Lodja- dem 
Territory), Germain 7551 (BR); Yangambi, J. Loui 
(BR), 2305 EM, BR), 11211 (BR. MO), 11586 He 
GABON. Ngo : Chaillu Massif, A. M. Louis et al. 909 
(K, WAG) 

38. Gaertnera liberiensis E. E A. Petit, Bull. 
Jard. Bot. Etat Bruxelles 29: 4 1959, nom. nov. 
Gaertnera salicifolia Hutch. F Gillett, FL 
Trop. Afr. 2: 21. 1931, hom. illeg., non Cano 
salicifolia C. H. Wright ex Baker, 1903. TYPE 
Liberia. Monrovia, Dukwai River, 1928, G. P. 


Cooper 277 (holotype, K!; isotypes, A!, BM!, 
FHO!) 


Trees, up to 7 m tall; bark smooth; branches terete, 
puberulent with indumentum drying yellow or gray- 
white, 1-2 mm diam.; internodes 2.5—5.2 em, smooth. 
Leaf blades 6-12 X 1.5-3 em, elliptic to oblong or 
lanceolate, apex acuminate, e cuneate, drying 
chartaceous, adaxially glabrous, abaxially glabrous 
except pubescent on principal veins with indumentum 
drying reddish brown to brown; secondary veins 
distinct abaxially, 4 to 7 pairs; domatia present; 
petioles 2-8 mm. Stipules tubular, pubescent, drying 
chartaceous, caducous, tube 5 
narrowly winged, arising beneath petiole and extend- 
ing to the apex, apex entire, marcescent, lobes 4, 
ca. 9 mm, linear to filiform, setae numerous, 

9 mm. /nflorescences cymose, many-flowered, ied 
on principal and/or axillary branches, puberulent; 
peduncle 2-3.5 em; branched portion narrowly 
18-5 X 1.2-3 em, congested, 


pyramidal, to 
branched to 3 or 4 orders; bracts deltate or trifid, 3— 


8mm, sometimes glabrous; bracteoles reduced; 

pedicels absent or 0 m. Flowers 5-merous, 

nid po ous. c p-shaped, 
side puberulent, glabrous inside, 


bud white, 


corollas, stamens, and stigmas unknown. Drupes 
nknown 


clavate and glabrous outside, 


Distribution and habitat. 
est Africa, where i 
The habitat a 


been recorded. 


This species grows in 
it has been collected in Liberia. 
nd elevations where it grows have not 


Phenology. Nos 
flowers, d 


s species has been collected with 
ths have not been 


recorded; it a not been collected with fruits. 


collection mon 


Discussion. | Gaertnera liberiensis- is so far only 
similar to G. 
eketensis, but G. liberiensis lacks the diagnostic 
longitudinally striated or fissured bark of G. eketensis 
and has pubescent rather than glabrous stipules and a 
pyramidal rather than c biform inflorescence. 
Although Hutchinson and Gillett later provided a 
Latin diagnosis for their name G salicifolia (Bull. 
Misc. Inform. Kew 


diagnosis; the ICBN requirement (McNeil et al., 
2006) for a Latin diagnosis only applies to names 
published during or after 1935. 


39. es longifoli Hortus Maurit. 

Me d (Bojer) Kuntze, 

>S B Pl. 1: 425. 1891. TYPE: Mauritius. 

Nouvelle ¡vete and Quartier Militaire, W. 

Bojer s.n. P not located; 
1 


a Bojer, 


(holotype, isotype, 


Annals of the 
Missouri Botanical Garden 


Gaerinera ee var. bur wow Kew Bull. 37: 
542. syn. TYPE: Mauritius. Nouvelle 
fo and Forêts de Grand Bassin, L. S. Bouton 
s.n. (holotype, K!). 

Trees, 1.5-3 m tall; branches terete to quadrangu- 
lar, when young glabrous or puberulent with indu- 
yellow, becoming glabrescent, 8— 
internodes 0.7-2 cm, smooth. Leaf 
blades 12.5-42 X 3.5-11.5 cm, elliptic to elliptic- 
— or oblanceolate, apex short! 
acute, base c in 


e abaxially glabrous or puberulent with 


cu 
neate, g coriaceous, adaxially 
indumentum drying brown or yellow; secondary veins 
prominulous abaxially, 8 to 16(to 20) pairs; domatia 
absent; petioles 20—40 mm. Stipules calyptrate, pu- 
berulent or glabrous, drying chartaceous, caducous, 
tube 20-25 mm, with ribs 4, broadly winged, arising 
below petiole and extending to lobes, apex with 2 
incisions, marcescent, lobes 2, 2-3 mm, deltate to 
linear. Inflorescences cymose, many-flowered, terminal 
on principal and/or axillary branches, glabrous to 
densely puberulent; peduncle 1—4 cm; branched 
portion corymbiform, 8-10 X 12-13.5 em, branched 
to 4 to 5 orders, lax; bracts deltate or trifid, 3-12 mm; 
bracteoles triangular to ovate, 1-2 mm; pedicels 1.5— 
3 mm. Flowers 5-merous, heterodistylous. Long-styled 
flowers: calyx cup-shaped, 3-4 mm wide, outside 
glabrous or puberulent, um hair-ring inside, truncate 
or lobes to 0.5 mm, rounded; corolla white, clavate in 
bud, when open salverform, glabrous kie, tube 
17-18 mm, 3-4 mm di mm, elliptic- 
oblong, obtuse; anthers included, sena inserted in 
upper third of corolla tube, 0.5-1 mm; style 18- 
part, 


lam., lobes 


19 mm, glabrous or often pubescent in upper 
15m led fl mature 


stigma ca. m. ort-styte owers 
corolla, stamens, and stigmas not seen. Drupes 
whitened to blue, obovoid to ellipsoid, 23-28 X 12- 
mm, ribbed; pyrenes plano-convex, rugose, finely 
fissured, endosperm entire 


This species grows in 
in wet forests and 


Distribution and habitat. 
Mauritius, where it can be foun 


lower montane forests at elevations of 500—700 m. 


Phenology. 


flowers in June and December and with fruits in May. 


This species has been collected with 


Discussion. Gaertnera longifolia is a distinctive 
cies with relatively ae elliptic-oblong or 
M leaves that coriaceous and relatively 


ud obovoid to ellipsoid fruits. The label data of 
Lorence 2224 note a “slight” 
(1 083) a the pubescent-leaved plants 


fragrance. Verdcourt 

as G. 
longifolia var. pubescens Verdc. This variety was only 
known from an early 19th-century Bojer collection 
d D'Argent 


rediscovered three individual trees growing in an area 


and was considered extinct until Page an 


in Black River Gorges National Park dominated by 
invasive species Mir ae 1997). These varieties 
are not separated pubescence is 
the only character p distinguished them, and this 

Fee seems to show continuous variation rather than 


; the variation in 


separating two distinct populations. 


Representative specimens examined. MAURITIUS. Per- 
rer Nature. Reserve, Lorence 2224 (MO), Vaughan MAU 
10387 (MAU), Vaughan MAU 14123 (MAU); Black River 
Gorges National Park, Malcomber 2935 (MO), Page MAU 
22729 (MAU 


40. s cus (Schweinf. ex Hiern) 
tit, Bull. Jard. Bot. État Bruxelles 29: 

T 1959, 1 as “longevaginalis.” Basionym: Psy- 
chotria Aa r Schweinf. ex Hiern, FL 
Trop. Afr. 
TYPE: Belgian Congo [Democratic Republic of 


201. 1877, as “longevaginalis.” 


the Congo]. Forestier Central Distr. [Équateur?]: 
Kapili River, Apr. 1870, G. A. Schweinfurth 3552 
(holotype, K!; isotypes, BM!, P!). 


Trees or shrubs, 1-6(-8) m tall; branches terete, 
glabrous to puberulent, 2-3(-6) mm diam.; internodes 
1.5-6.5 em, smooth. Leaf blades (8-)5.5-18.5 X 
0.7-)2.3-8 cm, i 


( elliptic 
oblong, or oblong, 


to oblanceolate, 
apex cuspidate or acuminate 
acute to obtuse, drying chartaceous, adaxially gla- 
brous, abaxially glabrous or sparsely hirtellous on 
principal veins; secondary veins prominulous abaxi- 
ally, 5 to 
5-15 mm. Stipules tubular, glabrous to puberulent or 


pairs; domatia absent or present; petioles 
pilosulose, drying membranous, caducous or with 
Hue base 1-2 mm, tube 8-28 mm, with ribs 
nt or 3 or 4, narrowly winged, arising above to 
below petiole and sometimes extending to lobes, apex 
entire or with 1 incision, marcescent, lobes 4, 1— 
mm, deltate to filiform. Inflorescence 
several- to many-flowered, terminal 
and/or axillary branches, puberulent to pilosulose or 
hirtellous, sessile or peduncle to 6.5 cm; branched 
E corymbiform, 1-10 X 1.2-15 em, lax to 
ngested, branched to 3 or 4 orders; bracts deltate 
or ae 3-18 mm, entire or ciliate; bracteoles 0.5— 
2 mm; pedicels absent or to 1 mm. Flowers 5-merous, 
Long-styled flowers 


wide, glabrous outside, with hair- 


heterodistylous. calyx cup- 
shaped, 
ring inside, lobes 0.2—4.5 mm, triangular to lanceo- 
in bud, when open 


infundibuliform or salverform, outside glabrous, tube 
4 


late; corolla white, clavate 


mm, m diam., inside villous in upper 
third; lobes bes. mm, len to ligulate; anthers 
shortly exserted, filaments inserted in upper third of 
corolla tube, ca. 0.5 mm; style 5—7 mm, glabrous or 
pubescent in upper portion, stigmas 0.8-2 mm. Short- 


styled flowers: similar to long styled except calyx 1- 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


3.5 mm wide, lobes 0.5-2.9 mm, triangular to lance- 
olate; corolla tube 1—4 mm diam., lobes 3-5 mm, 
triangular to ligulate or ovate- oblong: anthers fully 
exserted, filaments 3—6 mm; style 3.5-5 mm, stigmas 
0.5-2 mm. Drupes violet-black, globose to subglobose 
or didymous, 5-10 X 5-12 mm; pyrenes spherical or 
hemispherical, finely fissured, 


rugose, endosperm 


entire. 


Distribution and habitat. 
West and Central 
Angola, Cameroon, the Democratic Republic of the 
Congo, the Republic of the Congo, Gabon, Guinea, 
Cóte d'Ivoire, Liberia, and Sierra Leone. Here, it is 
found in humid forests at elevations of 150—1600 m. 


This species grows in 
Africa, where it is known from 


Phenology. This species has been collected with 
flowers and fruits throughout the year. 


Discussion. This species belongs to the Gaertnera 
vaginans complex; see also the discussion of that 


distinctions. This 
“lon, 


group for related species and their 
name was originally published as vaginalis,” 

but the ICBN mandates a spelling correction for this 
variant (McNeil et al., 2006). Van Beusekom (1967) 
included 


. nein, i n G. vaginans subsp. 
hat 


vaginans, but t om more narrowly 
here; consequently, Pune taxa included by van 
Beusekom are separated here. Gaertnera longivagi- 
nalis differs from G. vaginans in its tubular stipules 
that dry membranous, closely envelop the stem, and 
have prominent filiform lobes at the apex. 

Petit (19592) recognized Gaertnera longivaginalis 
and G. bracteata as separate species, but wit 
more specimens available these are not com- 
pletely distinct morphologically and are treated here 
as varieties. These differ n ide as outlined 


now, t 


in the key below and also in ele nal an 

geographie ranges, with variety bnc docu- 
mented widely from West through Central Africa at 
150-1000 m versus variety bracteata documented 
from West a LOU the Republic of the Congo 


at 1100-160! 


la. Calyx with tube 1-1.75 mm long, i 1.2- 

4.5 mm long; inflorescence bracts deltate to 
usually linear- as 40a. ues bracteata 
Calyx with tube 1.5-2.5 mm long, truncate or wit 

lobes to 1 mm long; inflorescence bracts deltate to 
usually elliptic ........ 40b. variety longivaginalis 


gs 
> 


ÉS 


0a. 


Gaertnera longivaginalis var. bracteata (E. 

Petit) Malcomber, comb. et stat. nov. 
Baden Gaertnera bracteata E. M. A. Petit, 
Bull. Jard. Bot. Etat Bruxelles 29: 49. 1959. 
TYPE: Congo [Republic of the Congo]. Kasai 
Distr.: Butala, E. Laurent & M. 
Laurent s.n. (holotype, P!). 


SE P ds var. ud E M. A. Petit, du eid 
xelles 29: 50. 1959, 

ee pres [Democratic Republic in ie de 

Distr. Forestier Central [Equateur], llenge, Nov. 1921, 

V. G. Goossens 2688 (holotype, BR; isotypes, G!, P!). 


ees, 14 m tall; branches 2-3 mm diam. Leaf 
lades 3-15 X 0.7—6 cm, elliptic to elliptic-oblong; 
petioles 5-10 mm. Stipules with tu m, with 
ribs none or 4, lobes 3-8 mm, filiform to narrowly 
triangular. eae densely pilosulose to hirtel- 
su peduncle 1-6 cm; ; branched portion 1.2-15 cm 
aliut pedicels absent or to 0.3 mm. du 
dud calyx 2-3 mm wide, with tubular 
75 mm, lobes 1.2-4.5 mm, triangular to lanceolate; 
ain tube 4—5 mm, 1-2.5 mm diam; lobes 3-5 mm, 
linear to ligular or elliptic-oblong. Short-styled 
flowers: similar to long styled except calyx l- 
2.5 mm wide, lobes 1.2-2.9 mm; corolla tube 4— 
6 mm; filaments 4-6 mm; style 4-5 mm, stigmas 0.5— 
l mm. Drupes 5-10 X 5-12 mm. 


Distribution and habitat. This variety grows in 
dre and West Africa, where it has been found in 
e Democratic Republic of the e the Republic 
d the Congo, Guinea, Cóte d'Ivoire, Liberia, and 
Sierra Leone. Here, it can be found in tail fures at 


elevations of 1100—1600 m. 


Phenology. This variety has been collected with 
flowers and fruits throughout the year. 


Discussion. | Gaertnera longivaginalis var. bracteata 
has been confused with G. cooperi, and some specimens 
cited in the original description of that species are not 
us with its type and are included here (Petit, 
1959a). This variety is similar to G. bieleri and 
hie those can be distinguished by the wëll- 
W. stipular ridges or wings that surround the 
of the petiole, and G. bieleri e by its 
IEAS to hirtellous stems and stipules 

Petit (1959a) separated ae — var. 
glabrifolia E. M. A. Petit based on its abaxial leaf 
surfaces completely glabrous rather than sparsely 
hirtellous along the principal veins. However, the 
criteria applied in this current revision for distin- 
guishing taxa are different, and here this variation in 
pubescence is considered to be a character that varies 
among local populations or among individual plants, 
and thus is not indicative of distinct, 
evolving evolutio p lineages 
ered to in a dist 
taxonomic ie nis is synonymized here. 


variety is 


Representative specimens examined. DEMOCRATIC RE- 
PUBLIC OF THE CONGO. Ligasa, Lokombe River (T. 
Isangi), Germain 8501 (BR). GUINEA [GUINEA-CON- 

AKRY] Cxéte, Adam 28771 (MO, WAG); Mt. Nimba, Adam 


Annals of the 
Missouri Botanical Garden 


ce P ed D [CÓTE e Mt. Nimba, 
Gaut utier-Beguin 1263 (G, MO); Tonkoui 
Massi " pos 380 E LIBERIA. Bomi Hills, poe n 1515 (K, 
MO, ns Nimba, Breteler 5458 (MO, P, WAG). REPUB- 
LIC OF THE CONGO. Alima-Likouala Basin, e DS 
o m (P, WAG). SIERRA LEONE. Lom 
Jaeger 7158 c MO, P, WAG); Nongowa, End Hille 
Bakshi 184 (K, 

(Schweinf. ex 


40b. E longivaginalis 
em) E. M. 


A. Petit var. longivaginalis. 


d plagiocalyx K. Schum. ex Thonn. & H. Durand, 
Fl. T 373. 1909. [Westafr. 
Ex ped. 322 , no TYPE: Belgi 


ville nsh nley 1899 
sh chier 12586 ‘(he slot ype, i co, d 
e, BM isotypes, BR!, G!, 
CEN longivotinalis var. bien r3 a “A. Petit, Bull. 
ard. Bot. État Bruxelles 29: 48. 1959, syn. nov. TYPE: 
Belgian Congo [Democratic a of He Congo]. 
Distr. Tum Central, Orientale, 
environs, ca. 460 m, J. Louis 12577 (holotype, BR!). 


c 
= 


Trees or shrubs, 1-6(8) m tall; branches 2-3(-6) 
mm diam. Leaf blades (3-)5.5-18.5 X (0.8-)2.3- 
8 cm, elliptic to elliptic-oblong or oblanceolate; 
petioles 10-15 mm. Stipules with tube 10-28 mm, 
with ribs 3 or 4, lobes 1-5.5 mm, deltate to filiform. 
— densely puberulent to ier ea sessile 

or peduncle to 6.5 cm; branched n 1.5-15 ¢ 

wide; deeds elliptic to deltate, 10 mm, ee 
pedicels absent or to 1 mm. Long-styled flowers: calyx 
lobes triangular to 
; lobes 2.5-3.5 mm, ligulate. Short- 
styled flowers: ae to long styled except calyx lobes 
0.5-1 mm; corolla tube 4—5.5 mm, 1.4—4 mm diam 
filaments 3-3.5 mm; style 3.5—4.5 mm, stigmas 1.5— 
2 mm. Drupes 5-8 X 5-8 mm. 


Distribution and habitat. This variety grows in 
West and Central Africa, Angola, 
Democratic Republic of the Congo, the Republic of 
the Congo, Gabon, Guinea, Cóte d'Ivo 
Sierra Leone. Here, it can be found in humid forests at 


elevations of 150—1000 m. 


Cameroon, the 


ire, Liberia, and 


Phenology. This species has been collected with 
flowers and fruits throughout the year. 


19592) 


etit ( 
var. louisii based o 


Discussion. Petit 


USS usan Gaertnera 
longivaginalis its 


s inflorescence 
with the branched d 1-6 cm Hus with two or 
three ial leaf 


a s and its abax 
surfaces sparsely xu on the principal veins 


pairs of se 
or less often glabrous, versus inflorescences with the 
branched portion 1.5—9 em long with two to five pairs 
of secondary axes and the abaxial leaf surfaces 
glabrous or less often pubescent along the costa. 


The conditions that characterized each of Petit's 
variety are thus similar and do not separate two 


Thus, the 


criteria applied in this current revision to distinguish 


clearly distinguishable groups of plants. 


taxa, the clear separation of distinct groups, are not 
met and so this variety is not recognized here and is 
synonymized accordingly. 

The holotype of Gaertnera plagiocalyx was de- 
stroyed with the general Rubiaceae collection in the 

erlin herbarium; the isotype at BM is selected as the 
lectotype here because it has a digital image available 
on the Aluka web site and is the best representative of 
this species among the digitally imaged isotypes there. 


Representative Ns examined. ANGOLA. Rio Lua- 
chimo, near Vila e de Carvalho, Exell 708 (BM 
CAMEROON. oe po 42 km of Mbalmayo, 


eee 34 (BR 
Leiouzey 3780 (MO, P, WAG); M 


Edjune rivers, 
km S of 


1120 (BM, BR, FHO); Imbonga (Territory Irede), Evrard 
6083 (BR, FHO); Léopoldville, o 6197 (BR, FHO); 
Liaka (Basankusu Territory), ambi, Tutuku Island, 


Ogooué-Lolo: PE Le Testu 7622 (BM, BR, MÓ, 

A [GUINEA-CONAKRY]. 

27518 (BR, MO, WAG). IVORY COAST [CÔTE D’ IVOIRE]. 

Mont Tonkoui, Bamps 2263 (BR, P). LIBERIA. Mt. Wolagisi, 
d 


, Descoings 
8282 (P, WAG), 8664 (WAG); Brazzaville; Chevalier 27743 
(P). Harder 3751 (MO). SIERRA LEONE. Mt. Loma, Jaeger 
1166 (MO, P, WAG). 


41. Gaertnera lowryi Malcomber, sp. nov. TYPE: 
Madagascar. Toamasina: Masoala National Park, 
oast, near Antalavia, 0.8 km NE of 


es 
village, lowland forest beside Antalavia River, 


15^47'S, 50°02'E, 50-150 m, 12 Oct. 1997, 
. T. Malcomber 2827 (holotype, MO!; iso- 
types, BR!, G!, Pt PRE!, TEF!, WAG). 
Figure 7C-H. 


aec speci DC.) Merr. similis, sed 
ab ea fo liorum anguste ellipticorum vel oblanceolatorum 
13- ad 18-jugat 


sub petiolo alis duabus prominentibus munitis distinguitur. 


Trees, 2-8 m tall; branches + tetragonal, glabrous, 
4-8 mm diam.; internodes 5.8-12.5 cm, smooth. Leaf 
blades 29-40 x 5.3-8.4 cm, elliptic or oblanceolate, 
apex acuminate or acute, base cuneate to ro d, 
drying coriaceous, glabrous; secondary veins promi- 
nulous abaxially, 13 to 18 pairs; domatia absent; 
petioles 5-30 mm i 
chartaceous, deciduous after distalmost 1 to 3 nodes, 


. Stipules tubular, glabrous, drying 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


often by fragmentation, tube 29—80 mm, with ribs 4, 
broadly winged, arising below petiole and extending to 


lobes, apex with 1 deep incision, marcescent, lobes 4, 


m, branched to 3 to 4 orders, lax to congested; 

edid deltate or linear to ligulate, 2.3-18 mm; 
bracteoles triangular to ligulate or lanceolate, 0.5— 
i abs or to 2.5 mm. Flowers 5- 
Long-styled flowers: calyx 
outside glabrous or 


; pedic nt 
, heterodistylous. 
cup-shaped, 2-3 mm wide, 
puberulent, glabrous inside, lobes 0.1-0. mm, trian- 
gular; corolla white, clavate in bu en op 
salverform or infundibuliform, outside es ie 
18-21 mm, 1.54 mm diam., 


middle, lobes 4—5 mm, Mgr to ligulate, acute; 


inside villous at ca. 
anthers included, filaments O at ca. middle o 
corolla tube, 0.3-1 mm e mm, glabrous, 
stigmas 2.5-3 mm. Short-styled ee similar to long 
styled except calyx urceolate or cup-shaped, truncate 
.3 mm; corolla outside glabrous or 
puberulent; tube 17.5-20 mm, 1.5-5.5 mm diam., 
lobes 4.5—5 mm; anthers shortly exserted, filaments 
inserted in upper third of corolla tube, 2.4-3 mm; 


style 
known 


or lobes to 


—10 mm, stigmas 3-3.5 mm. Drupes un- 


Distribution and habitat. This species grows in 

Madagascar, where it is known from the a of 
Toamasina in the Mananara area and on Masoala 
Peninsula. Here, it can be found in um forests at 


elevations of 0-600 m. 

Phenology. This species has been collected with 
flowers January through August and October through 
December, and with fruits in October. 

Discussion. 
letouzeyi of Central Africa, but can be distinguished 


Gaertnera lowryi is similar to G. 


by its elliptic or oblanceolate leaves, 13 to 18 pairs of 
secondary veins, and pale greenish white calyx tube. 
In the field, G. lowryi 


iens inflorescence and reflexed uppermost 


is an impressive species with a 


leaves. Gaertnera lowryi is named for Porter Prescott 
“Pete” Lowry II (1956-), who made some of the first 
collections of the species and has made a number of 
important botanical collections from the Masoala 
Peninsula of Madagascar. 


Paratypes. MADAGASCAR. Toamasina: Mananara- 
Nord National Park, eee et al. 2888 (MO, TEF); 


N), 
Misa 2800 (MO, TER), Zjhra & Hutcheon 392 (MO). 


. Gaertnera macrobotrys Baker, J. Linn. Soc., 
Bot. 20: 208. 1883. pe macrobotrys Qu 
Gen. : 4 

Madagascar. Central ae R. Baron 


Kuntze, Revis. 
1945 
(holotype, K!; isotype, P!). 
Trees or shrubs, 3-10 m tall; branches quadrangu- 
lar, i 5-7 mm diam.; internodes 1.4—9.5 cm, 
h. Leaf blades * 7526 X 3 
de to ovate or occasionally oblanceolate, apex 
rounded then sometimes abruptly cuspidate, base 


8-13 em, elliptic- 


rounded to truncate, drying coriaceous or chartaceous, 
glabrous; secondary veins prominulous nr. 9 to 
10 pairs; domatia absent; petioles 10—25 mm. Stipules 
tubular, inflated to funnel-shaped, m drying 
chartaceous, persistent on distalmost 1 to 2 nodes to 
deciduous often through fragmentation, often leaving a 
well-developed annular scar, tube 7-20 mm, with ribs 
none or 4, rounded to narrowly winged, arising 
beneath petiole and extending to base of tube, apex 
lobes 4, 0.5-1 mm, 
deltate. Inflorescences cymose, many-flowered, termi- 
nal on principal and/or axillary branches or axillary, 

abrous or puberulent; peduncle  1.5-10 em; 
branched portion corymbiform, 5-15 X 4-22 em, 
branched to 4 to 6 orders, lax to congested; bracts 
deltate or trifid, 1-10 mm; bracteoles triangular to 


with 2 incisions, marcescent, 


ge 
— 


lanceolate, 0.5-1.5 mm; pedicels absent or to 1.8 mm 

owers 5-merous, heterodistylous. Long-styled flow- 
calyx cup-shaped, 2— wide, outside gla- 
brous, inside glabrous or usually with hair-ring, 
truncate or rarely with lobes to 0.3 mm, triangular; 


ers: 


corola white or pink, clavate in bud, when open 


salverform, outside glabrous, tube 7—10 mm, 2- 
3. iam., inside villous in upper third, lobes 
2-2.5 mm, triangular to ligulate, acute; anthers 


included, filaments inserted at ca. middle of corolla 
tube, 0.8-1 mm; style 7-9 mm, glabrous or often 
pubescent near apex, stigmas 0.5-1.1 mm. Short- 
styled flowers: similar to long styled except calyx 
1.5-3 mm wide, glabrous inside; corolla tube 6.5— 
10 mm, 1.5-3.5 mm diam., lobes ca. 3 mm; anthers 
shortly exserted, filaments inserted in upper third 


of corolla tube, 2-3 mm; style 4.5-5 mm long, stig- 


mas 1.5-2 mm. Drupes violet-black, globose or 
subglobose, 8-9 X 8-10 mm; pyrenes spherical or 
hemispherical, rugose, finely fissured, endosperm 
entire. 


Distribution and habitat. This species grows in 
Madagascar, where it is known from the provinces of 
Fianarantsoa and Toamasina. Here, it can be found in 


humid forests at elevations of 660—1100 m. 


Phenology. This species has been collected with 
flowers January through March and October through 


Annals of the 
Missouri Botanical Garden 


December, and with fruits January through August 
and October through December. 


Discussion. Gaertnera macrobotrys is similar to G. 
arenaria, G. monstruosa, and some plants of G. 
obovata var. sphaerocarpa with relatively large leaves 
(G. macrobotrys is easily separated from the Sain: 

of G. obovata var. sphaerocarpa by its 
larger leaves). However, G. macrobotrys can be 
distinguished from these by its elliptic-oblong to 
ovate or oblanceolate leaves, its inflated to funnel- 
shaped, shortly persistent to deciduous stipules that 
are partially split into two large, ovate lobes, and its 
usually terminal inflorescences. Gaertnera macrobo- 
irys is also similar to G. macrostipula, but differs from 
it in its stipules usually deciduous after the top one or 
two nodes and usually elliptic-oblong or ovate leaves. 


Representative specimens examined. MADAGASCAR. 
Fianarantsoa anomafana National Park, Malcomber 
2585 (K, MO, P, TAN), Kotozafy 267 Mo, Py; Vohipeno, 
Beaujard 183 (K), 387 (K). Toamasi ndasi 
Benoist 1520 (P, TAN), Dorr e (TAN), Lowry 4272 (P. 

TAN), Miller 3818 (MO, P, 
4250 (P, TEF); Mananara re Park, near Antanambe, 
Morat 8576 (P). 


43. Gaertnera macrostipula Baker, J. Linn. Soc., 
Bot. 20: 207. 1883. Sykesia pp d (Baker) 
Kuntze, Revis. Gen. Pl. 2: 425. 1891. TYPE: 
Madagascar. Central Madagascar, i^. Baron 1922 

i pude here, K!; isotypes, K!, P!). 


Trees or shrubs, 2-7 m tall; branches flattened or 
terete or quadrangular, glabrous or puberulent with 
indumentum drying brown, mm diam. inter- 
nodes 1.8-7 em, smooth. Leaf em 3.5-24 X 1.8- 
14.5 em, oblanceolate to obovate, elliptic, or ovate, 
apex rounded then shortly cuspidate or acute, base 
acute and often decurrent, drying chartaceous, 
adaxially glabrous, abaxially glabrous or hirtellous 
along principal veins; secondary veins prominulous 
abaxially, 7 to ll(to 13) pairs; domatia absent or 
rarely present; petioles 18-20 mm. Stipules bs 

or often Aei to ariel shaped, gla- 
brous or rarely puberulent, 


cylindrical o 
drying a 
persistent, tube 10-33 mm, with ribs 4, rounded to 
narrowly winged, arising below petiole and sometimes 
extending along tube to lobes, apex entire or with 2 
incisions, marcescent, lobes absent or 4, 1-2 mm, 
deltate. /nflorescences cymose, many-flowered, termi- 
nal on principal and/or axillary branches, sparsely to 
densely puberulent, pilosulose, or rarely glabrous; 
peduncle 2.5-11 em; branched portion corymbiform, 
(18-19 X (1-)4-16(-20) cm, branched to 3 to 5 
orders, lax to congested; bracts deltate or trifid, 3— 

mm; bracteoles triangular, 1—3.5 mm; pedicels 
absent or to 2 mm. Flowers 5-merous, heterodistylous. 


Long-styled flowers: 
wide, outside sparsely to densely puberulent or 


calyx campanulate, 2-3.5 mm 
glabrous, glabrous inside, truncate es l- 
3.5 mm, equal to markedly 


spatulate; corolla white to dark pink, clavate in bud, 


r lo 

nequal, linear to 
liform or salverform, outside 
.5-8.5 mm, 2-3.5 mm 


iam., inside villous in upper third; lobes 2.5-3 mm, 


when open infundibu 
puberulent (glabrous), tube 7 


ieee. acute; anthers included, filaments inserted in 
0.5-1 mm; style 9— 
10 mm, glabrous, stigmas 0.5-1 mm. Short-styled 


upper third of corolla tube, 


flowers: similar to long styled except calyx lobes 
absent or 1.5-3.5 mm; corolla tube 6.5-9 mm, 1.3— 
lobes 2-3 mm, 
ligulate or linear; anthers shortly exserted, filaments 
glabrous or pubescent. 
stigmas (0.2-)0.8-2 mm. Drupes violet-black, globose 


3.5 mm diam., inside glabrous, 


—4 mm; style 2.5-3 mm, 


to subglobose or didymous, 7-9 X 7-9.5 mm; pyrenes 
spherical or hemispherical, rugose, finely fissured, 
endosperm entire. 


Distribution and habitat. 
Madagascar, where it is widespread. It lives in humid 


This species grows in 


forests at elevations of 0-1800 m. 

Phenology. This species has been collected with 
flowers January through March and August through 
December, and with fruits January through May and in 
November and December 
This 
nized by its well-developed, often funnel-shaped 


Discussion. common species can be recog- 
stipules that loosely surround the stem; od 
on many stems the stipules are almost as long as the 
internodes and thus mostly cover the entire stem. The 
calyx lobes are usually linear to spatulate and well 
developed but rarely are absent. The flowers are 
usually dark pink but often bleach white in the sun. 
Gaertnera macrostipula is one of the most commonly 
collected Gaertnera species in Madagascar, growing in 
a variety of forest types from littoral forest near sea 
level to moss forest at 1 t is similar to G. 
macrobotrys; see also comments under that species. 

Two syntypes were cited in the protologue, R. Baron 
1922 (K!, P!) and W. T. Gerrard 54 (K!). The former is 
selected here as lectotype because it is a more 
complete and exemplary specimen and has duplicates; 
the lectotype specimen is selected because it is the 
most complete specimen of the set. 

b diee specimens examined. MADAGASCAR. 
njanaharimbe Massif, Humbert 24621 (P); 


serve, Morat 4150 (TA 
2709 (MO), 2774 i 2777 (MO), 2778 (MO), 2782 (MO). 
Fianaranisoa: diamontana, Ranomafana, Seigler 


12802 (MO); ae Basin, Rienana Valley, Humbert 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


3450 (MO, P) Toamasina: Mantadia National Park, 

0), p. 16553 (MO); Ambodiri- 

O, Pj; Andavikimenarana, 

Dorr 4484 (MO); Didy, Brickaville, Cours 4879 (P); Masoala 

National Park, Malcomber 2807 (MO), 2834 (MO); Masoala 
9 (P). 


> 
2 
3 
2 
I: 
a 
2 
a 
ies] 
OB E 
= 
z 
A 
ai 
E. 
A 
O 
Bs 


Peninsula, Perrier de la Báthie 3649 

44. Gaertnera madagascariensis (Hook. f.) Mal- 
4, 5. 
005. Basionym: Hymenocnemis madagascarien- 


sis Hook. f., Gen. Pl. 2: 132. 1873. TYPE: 
Madagascar. s. loc. W. Bojer s.n. 
K! 


Missouri Bot. 


(holotype, 


Trees, 2-5 m tall; branches terete, when young 
densely pilosulose to villosulous, strigose, or sericeous 


; internodes 0.4-2.5 cm, smooth. Leaf blades 
0.5—4 X 0.3-2.5 cm, ovate to elliptic or obovate, apex 
tuse), base cuneate to 

glabrous 


veins visible and flat to prominulous abaxially, 3 to 


acute to shortly acuminate (ob 
rounded, drying chartaceous, secondary 
5(6) pairs; domatia present; petioles 0.5-2.5 mm. 
Stipules eee densely villosulous to Lae or 
strigillose, drying membranous, caducous, persistent 
on distalmost E to 4 nodes, or a through 
fragmentation, tube 3.5-23 mm, with ribs 4, narrowly 
winged, arising beneath petiole and sometimes 
extending along tube, apex with 1 incision, lobes 2, 
0.3—0.6 mm, deltate. /nflorescences 1- to 3-flowered, 
fasciculate, terminal on principal and/or axillary 
branches; peduncles 2.5-12 mm, strigose to seri- 
ceous; bracts deltate, to 5 mm, glabrous or puberu- 
lent; bracteoles absent. Flowers 5-merous, heterodi- 
stylous. Long-styled flowers: calyx campanulate, 1.2— 
ide, glabrous, lobes mm, triangular to 
linear; corolla white, clavate in bud, when open 


salverform, throughout glabrous, tube 5-10 mm, 2- 
iam; lobes 3-5 mm, ligulat elliptic- 
oblong, acute; anthers included, filaments inserte 


in upper third of corolla tube, 0. E 0.4 mm; style 7— 
10 mm, glabrous, stigmas 0.3-0.7 mm. Short-styled 
flowers: unknown. Drupes violet-black, globose or 
subglobose, 5-6 X 5-6 mm; pyren 


hemispherical, 


es spherical to 


rugose, finely fissured, endosperm 


entire. 
Distribution and habitat. 
Madagascar, where it is widespread in the eastern 


This species grows in 


escarpment. Here, it is found in humid evergreen 
forests on metamorphic and igneous rocks and rarely 
on alluvial deposits, at elevations of 800-1400 m. 
Phenology. This 
January through Mare in November and 
December, and with fruits March through May. 


species has been collected 


and 


Discussion. This species was originally separated 
in a monotypic genus, but molecular analyses indicate 
that Hymenocnemis madagascariensis is nested within 
the Gaertnera clade (Malcomber, 2002; Malcomber & 
2005) and 
Gaertnera. Hymenocn 
scribed based on its superior 


Davis, support its inclusion within 


emis was originally circum- 
ovary in the fruit, 
“sheathing” stipules, and inflorescence reduced to a 
single flower. All of these characters are found within 
Gaertnera, 


and so Gaertnera's 
characteristic stipular wings BM the petiole. 


ecie 
Gaertnera madagascariensis can be recognized by its 
densely appressed-pubescent young branches and 
stipules, its membranous calyptrate stipules, and its 
inflorescence reduced to one or a few flowers. This is 
the most commonly collected species in Madagascar 
of the 


flowers discussed by Malco: 


solitary or few 
(2005). 


Gaertnera madagascariensis is similar to G. brevipedi- 


of Gaertnera species with 
mber and Davis 


cellata; see additional comments under that species. 
Malcomber and Davis (2005) considered the conser- 
vation status of this species to be Least Concern 


(IUCN) 


Representative specimens examined. MADAGASCAR. 
tananarivo: Anjozorabe, Bosser 12839 (K, P, TAN), 
8200 (TAN); bere nka, batolaona, 65 km E o 


Antananarivo, Leandri 3. 04 (P). Fianarantsoa: Andringitra 
Massif, Ambodipaiso cud Cours 2286 (P), Homolle 2286 
(P, TAN); Ranomafana National Park, Vatoharanana, Mal- 
comber 2865 (MO), 2870 (MO). Toamasina: Analamazoatra, 
Perrier de la Béihie 6892 (K, P) Bemainty Massif, 
Rahobevava, Cours 4177 (P, TAN). 
45. Gaertnera microphylla Capuron ex Malcom- 
& A. P. Davis, Monogr. Syst. Bot. Missouri 
Bot. Gard. 104: 390, figs. 2, 6. 2005. TYPE: 
r. Toamasina: 6.6 km SE of Pese 
small a beside RN2 at base of Andria 


vibe, 18°56'S, pa 23 Nov. 1997, s p 


Malcomber, A avis, D. Gower & J. 
Mec p dni MO; 2 
BR!, G!, K!, LE!, , MAU!, P!, PRE!, 
ol UPS!, WAG?). 
Tre r shrubs, 2-4 m tall; branches flattened to 
terete, m 0.4—3.4 mm diam.; internodes 0 


0.6 em, with 2 lengitudinal ribs. Leaf blades 0.3-1.5 


branous, glabrous; secondary veins hardly visible to 
prominulous abaxial pairs; domatia absent or 
2 


present and foveolate; petioles 0.5-1.2 mm. Stipules 


calyptrate, glabrous, drying membranous, persistent, 
quickly splitting into 4 narrowly spatulate to linear 
mm, with ribs 4, narrowly to 


segments, tube 


broadly winged, arising beneath petiole and extending 


Annals of the 
Missouri Botanical Garden 


along each persistent segment; apex with 2 or 4 
incisions, lobes 4, 1-1.5 mm, linear. Inflorescences 
reduced to 1 flower, terminal on axillary branches; 
bracts stipuliform, involucral; bracteoles | absent. 
Flowers 4-merous, heterodistylous, subsessile. Long- 
styled flowers: calyx cup-shaped, 1.5-2 mm wide, 
glabrous, lobes 2.4-3 mm, linear to oblanceolate or 
ligulate; corolla white, clavate in bud, when open 
salverform, outside glabrous, tube 4—4.5 mm, — 
3 mm diam., inside villous at ca. middle, lobes 2 
2.5 mm, triangular to ligulate, obtuse; anthers includ- 
ed, filaments inserted in upper third of corolla tube, 
0.2-0.4 mm; style 4.5-5 mm, glabrous, stigmas 0.9— 
1.3 mm. Short-styled flowers: unknown. 
didymous, 5-6.2 X 


cal, rugose, finely fissured, endosperm entire. 


rupes blue, 
6.5-7 mm; pyrenes hemispheri- 


Distribution and habitat. This species grows in 
Madagascar, where it is known from the province of 
Toamasina. Here, it can be found in wet evergreen 
forests on metamorphic and igneous rocks at eleva- 


tions of 790-1050 m 


Phenology. This species has been collected with 
flowers in November and December and with fruits in 
March through May. 


Discussion. This species is distinctive in its 
relatively small leaves that are broadest near the 
apex and its stipules that quickly split into four 
narrow, persistent segments. The stipules or stipule 
segments cover the stem and overlap due to the 
reduced internodes. The calyx with its well-developed 
narrow lobes resembles the stipules. This species is 
only known from around Andasibe and Analamaza- 
otra; it was n a conservation status of Critically 
E. UCN, 2001) by Malcomber and Davis 
(20 aertnera microphylla and G. furcellata are 
A and in particular share the unusual stipules 
and inflorescences reduced to a single flower; they 
differ primarily in their respective distinctive leaf 
shapes. 


eru — examined. MADAGASCAR. To- 

. Perinet & Anevoka, base de rocher de 
ho. n SF pe (MO, P, TEF), SF 
24770 (MO, P, TEF) re region of Saharanga mountain pass, NE 
of Perinet, Capuron SF 24022 (MO, P, TEF) Andasibe 
(Perinet), N of rd. from Antananarivo to Tametave, Lowry & 


Schaiz 4283 (MO). 


46. Gaertnera monstruosa Malcomber, sp. nov. 
TYPE: Madagascar. Toamasina: 6.4 km W of 
Antanambe, Mananara National Park, lowland 
n Wn 49°48'E (WGS 84 a 

8 . 1997, S. 

an; n "^ en. BRI, KI, P! "TEF! 

WAG)). Figure 12A-D. 


Haec me Gaerinerae obesae Hook. f. ex C. B. Clarke 
similis, sed ab ea caule in sectione transversali subqua- 
drangulari, m wA hermaphroditis distylis atque tubo 
corollino 15-20 mm longo distinguitur. 

Trees or shrubs, 1—5 m tall; branches quadrangular, 
glabrous, 2-15 mm diam.; internodes 3.5-26 cm, 
h. Leaf blades 35-51 X 10.5-14.6 cm, elliptic 


to oblanceolate or obovate, apex acuminate, base acute 


smoot! 


to cuneate, drying chartaceous to coriaceous, glabrous; 
secondary veins prominulous abaxially, 12 to 16 pairs; 
domatia absent; petioles 15-25 mm. Stipules tubular, 
funnel-shaped, glabrous, drying coriaceous, persistent 
at least on distalmost 3 or 4 nodes, tube 18-51 mm, 
with ribs 4, broadly winged, a below petiole and 
extending to lobes, apex with eep incisions 
producing 2 broadly oblong to p segments, apices 
acute or sometimes tardily splitting, without lobes. 
Inflorescences cymose, many-flowered, terminal on 
principal and/or axillary branches or sometimes on 
reduced axillary branches and appearing axillary, 
puberulent, sessile or s to 11 em; branche 
portion 8.5 X 6-26 cm, branched 
to31to4 pis lax to congested; bracts deltate or trifid, 
0.5-2.2 mm, sometimes ted O fe aes 
to rounded, 0.1-0.5 mm; pedicels absen 
Flowers 5-merous, m nad pM mid flowers: 
calyx cup-shaped, 1.8-3 mm wide, outside puberulent, 
with hair-ring inside, truncate; corolla white, clavate in 
bud, when xr salverform, glabrous throughout, tube 
15-20 m 1.5-2.5 mm diam., lobes 3—4.5 mm, 
ligulate or E c d acute; anthers included, 
l r third of corolla 
tube, 1-2 mm; style 16- 20 mm, Pee stigmas 2— 


filaments inserted at middle to u 
3 mm. Short-styled flowers: similar to long styled E 
calyx 1.5-2.5 mm wide; corolla tube 12-13.5 
1.8-2.5 mm diam., lobes 4-5 mm, ligulate; hee 
included, filaments 1.2-1.4 mm; style 3.54 mm, 
stigmas 1.5-2 mm. Drupes unknown. 


Distribution and habitat. This species grows in 
Madagascar, where it is known from the province of 
Toamasina, in the Soanierana-Ivongo and Mananara 
Nord National Parks. Here, it can be found in humid 
forests at elevations of 180—400 m. 


Phenology. This species has been collected with 
flowers August through November but has not been 
collected with fruits. 


Discussion. Gaertnera monstruosa is similar to G. 
obesa of Southeast Asia, but differs from that d 
in its more squared stems, larger flowers, and stam 

inserted in the middle or upper third of the ed 
tube. Additionally, bisexual 
distylous flowers, whereas G. obesa is dioecious. 


G. monstruosa has 


ithin Madagascar, G. monstruosa is similar to C. 
macrobotrys, but can be separated from it by its 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


aertnera monstruosa Malcomber. 


Figure 12. NC 
stipules. —C. jor styled flower. —D. Short-styled flowe 


flower in cross section. —I. Short-styled flower. —J. Short 

same 1-cm scale; I, J to same 1-cm scale. 

relatively large, elliptic to obovate or oblanceolate 

leaves, often a axillary inflorescences, and 
er flowers. The 

ai ten size ie the leaf blades, stipules, 


cific epithet refers to the 


and stem 


MADAGASCAR. Toamasi 


Paratypes. ina: Mananara- 
Nord National Park, Ibanda, Malcomber et al. 2909 (MO); 


S of Soanierana-Ivongo, Sahavolamena, Capuron SF 23800 
(TEF); Sahafotra, N of Navana (Bay of Antongil), Capuron SF 
28396 (TEF). 


. Gaertnera obesa Hook. f. ex C. B. Clarke, in 


Uragoga stipulacea Kuntze, Revis. Du: 

. 1891, nom. superfl. TYPE: Singapore: N. 

Wallich 8328 (holotype, K!; isotype, K!). 

Shrubs or trees, 2—5 m tall; branches terete to 
quadrate, glabrous, 5-10 mm diam ernodes 

1l em, smooth. T blades 20-55 X (7-)9-19 cm, 


obovate to oblong, apex acuminate to acute, base 


attenuate to cuneate or sometimes attenuate, drying 
coriaceous, glabrous, margin flat and with cartilagi- 
nous thickening; secondary veins prominulous abaxi- 
ally, 9 to 13 pairs; domatia absent; petioles 1-10 cm, 


er in cross secti 
Flowering branch. —F. Portion of stem with sedis bases, stipules, and stem dd. —G. Long-st 


—A. Fruiting a —B. Portion of stem with leaf ine and 
e —E. 


. E-J. Gaerinera walkeri (Arn.) Blum 
tyled flower. —H. line styled 


styled flower in cross section. C, D to same l-cm scale; G, H to 


A-D based on Maleomber 2887; E-J based on Malcomber 2771 


furrowed and sometimes with 2 lateral ridges formed 
by a e leaf 
shaped, glabrou: 


to coriaceous, ique via fragmentation, tube 20— 


se. Stipules tubular, funnel- 
e ing chartaceous 


50(—75) mm, with ribs 4, broadly winged, arising 
ng to lobes, apex 


nd extendin wit 
ucing 2 broadly oblong sections 


below petiole a 
deep incisions pro 
with acute to rounded apices, lobes 4, 3—4 mm, 
deltate to narrowly triangular, ciliolate. Inflorescences 
cymose, many-flowered, terminal on axillary branches 


branched to 3 to 5 orders, rather congested; braci 
0.5-12 mm, glabrous; bracteoles 
s 5- 


calyx cup- 


linear to ovate, 
reduced; pedicels absent or to 
Pistillate flowers: 
shaped, 2-3 mm wide, outside glabrous or puberulent, 


mm. 
merous, unisexual. 
with hair-ring inside, truncate; corolla white, clavate 
in bud, when open salverform, outside glabrous, tube 
7-9 m m insid pper 
third, ities 1.8-2.5 mm, ligulate. acute; staminodia 


e vi ous inu 


included, filaments inserted in lower third of corolla 
tube, ca. 0.3 mm; style 3-3.5 mm, glabrous, stigmas 

1.3-1.5 mm. Staminate flowers: similar to pistillate 
except corolla tube 7-10 mm, 2-3 mm diam., lobes 


Annals of the 
Missouri Botanical Garden 


—2.5 mm; anthers included, filaments ca. 0.5 m 
pistillode reduced or absent. PLU violet- black, 
globose to didymous, 8-10 X 10 
hemispherical or spherical, faintly rugose, deeply 


; pyrenes 


fissured, endosperm entire. 


Distribution and habitat. 
southeastern Asia, where it is known from Peninsular 
e the Sarawak (Malaysia) 

sector of B ere, it can be found 


This species grows in 


Malaysia, — 
.H in humid 
forests at dE of o m. 


Phenology. This species has been collected with 
flowers January through March and August through 
December, and with fruits in April. 


Discussion. Gaertnera obesa is similar to G. 
monstruosa of Madagascar, but can be distinguished 
by its cylindrical stems, smaller flowers, and stamens 


hird of the 
Additionally, G. obesa is dioecious, 


and staminodes inserted in the lower t 
corola tube. 

reas G. monstruosa is bisexual and distylous. Van 
Beusekom (1967) noted that the field notes of several 
collections of G. obesa mention that the plants 
harbored ants. One collection, Cantley's Collector 
3008 (SING), is intermediate between G. obesa and G. 

brid. 


grisea and presumably a natural hy 


Representative specimens examined. MALAYSIA. Johor: 
N of M 


= 
UB 


Gunong Lesong, 
Rompin, Shah 3090 (SING). Perak: 8. en Ridley 1440 e 
Sarawak: Sibu, Ulu Sg. Pasai, Bukit Tanggi, Yii S 64422 

(SAR). o Pu ru Botanic Gardens, Garden 
Jungle, 7 Mar. 1926, C. X. D. R. Furtado s.n. (SING); Upper 
Peirce Reservoir. Nature Reserve, Malcomber 3007 (MO, 


SING) 


48. Gaertnera oblanceolata King & Gamble, J. 
Asiat. Soc. Bengal, Pt. 2, Nat. Hist. 4: 92. 1907. 
T : Malaysia. Perak: Sungai Luat, L. Wray, 
Jr. 2283 (lectotype, designated here, SING!; 
isotype, GP). Figure 4A—F 


Trees, 1-2.5 m tall, often monocaulous; branches 
terete or quadrangular, 
h. Leaf blades 11-31 X 

cm, oblanceolate or obovate, apex cuspidate to 
acuminate, base attenuate to cuneate, drying coria- 
ceous, glabrous or puberulent abaxially on principal 
veins, margin flat and sometimes thinly cartilaginous; 
condary veins prominulous abaxially, 6 to 15 pairs; 
domalla absent or present; petioles 3-40 mm. Stipules 
ous, deciduous 


tubular, glabrous, drying chartace 


8 
persistent on distalmost nodes or deciduous leaving 
a persistent base 5-10 mm, tube 11-20 mm, with ribs 
4, broadly winged, arising below petiole and extend- 
ing to lobes, apex entire or with 1 


lobes 4, 1-3 mm, deltate to linear. 


incision, 
marcescent, 


Inflorescences | congested-cymose to subcapitate, 
many-flowered, axillary or paired and supra- prd 
densely puberulent to pilosulose; peduncle 1.5-8 e 

branched portion subglobose or narrowly ned 
2.5-11.5 X 1-6 cm, branched to 2 to 4 orders; bracts 
deltate or linear, 3-10 mm, sometimes glabrous; 
bracteoles reduced; pedicels absent or up to 3 mm. 
Flowers 5-merous, unisexual. Pistillate flowers: calyx 


cup-shaped, 1.5-3 mm wide, outside puberulent, with 


hair-ring inside, truncate or lobes to 0.2 mm, 
triangular; corolla white, clavate in bud, when open 
salverform, outside glabrous, tube 4-5 mm, 1.5- 


iam., inside villous in upper third, lobes 

1-1.4 mm, triangular to ligulate, acute; staminodia 
included, filaments inserted in upper third of corolla 
tube, ca. 0.3 mm; style 3.5—4 mm, glabrous, stigmas 
0.5-0.8 mm. Staminate flowers: similar to pistillate 
wide, outside glabrous or 

puberulent; lobes ca. 0.3 mm; corolla tube 2.5— 
4.5 mm, 1-2 mm lobes 1.3-1.7 mm; anthers 
shortly exserted, Bon 0.2-0.4 mm; pistillode 


except calyx 1.5— 
diam., 
with style portion 2-3.5 mm, stigmatic portions l- 
1.5 mm. Drupes violet-black, globose or subglobose, 
4-6 X 4-8 mm; pyrenes spherical or hemispherical, 


> rugose, finely fissured, endosperm entire. 


Distribution and habitat. 
southeastern Asia, where it has most commonly been 


This species grows in 


collected in Peninsular Malaysia and is also known 
from the Sarawak (Malaysia) sector of Borneo. Here, it 
is found in humid forests at elevations of 30-150 


Phenology. This species has been collected with 
flowers April through August and in November, and 
with fruits in August. 


Discussion. | Gaertnera oblanceolata is similar to G. 
diversifolia, but G. oblanceolata differs in its sub- 
globose, axillary and supra-axillary inflorescences, 
versus lax pyramidal inflorescences that are terminal 
on axillary and supra-axillary branches. Gaertnera 
diversifolia was treated as a variety of G. oblanceolata 
y van Beusekom (1967), one of several species 
circumscribed more broadly by him t 

Four syntypes were cited in the protologue: £. 
Wray, Jr. “2283 (Gl, SING), L. Wray, Jr. 1948 (G), 
Scortechini 253 (not located), and King's Collector 
8449 (not located). The first collection is here apii 
as lectotype because it is more complete an 

xe d because it has two 


deposited in different herbaria. The specimen at 


mplary, a 
SING is selected as the lectotype because it is more 
complete. 
Representative no Negeri 
Sembilan: Pasoh For , Wong FRI 28905 (KEP), 
Wong FRI 2 (KED), [Op 3021 (MO), 3022 (MO). 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


Perak: Bukit Blakang Parang, Hanifi SING 21055 (K, L, 
. Sarawak: Miri Distr., Niah National Park, Mal- 
comber 5059 Pun 3040 (MO), June 1894, M. R. Haviland 
r: Kanching, Bukit Batu. Berdinging, Nur 

SING 34357 A. n KEP, L, SING). 


49. Gaertnera obovata Baker, J. Bot. 20: 218. 
1882. Sykesia obovata (Baker) Kuntze, Revis. 
Gen. Pl. 2: 425. 1891. TYPE: Madagascar. 

Chiefly from Betsileo-land, R. Baron 149 

(holotype, K!; isotype, P!). Figure 13A—L. 


Trees or shrubs, 1.5-5(-12) m tall; 
flattened to quadrangular or terete, glabrous or 
puberulent, 2-9 mm diam.; internodes 0.3-6.5 cm, 
smooth or with 2 longitudinal p at least when young 
Leaf blades 1.2—7(-23) X 0.7-8.5 cm, elliptic to 


. d 


branches 


oblanceolate, elliptic-oblong, or 
ovate, apex acute to rounded then abruptly short- 
acuminate or cuspidate, base acute to obtuse, drying 
chartaceous, adaxially glabrous, abaxially glabrous or 
puberulent, margin flat or crisped; secondary veins 
visible and flat to prominulous, 5 to 16 pairs; domatia 
absent or present; Pid 2-25 mm. Stipules calyp- 
trate, glabrous, drying membranous or chartaceous, 
caducous or desiduus e distalmost 1 or 2 nodes, 

tube 8-32 mm, with ribs 4, narrowly winged, arising 
below petiole and sometimes extending to lobes, apex 
entire or with 1 incision, lobes 2, 2-3.5 mm, deltate to 
linear. Inflorescences cymose, several- to many- 
flowered, terminal on principal and/or axillary 
branches, glabrous or puberulent to pilosulose, sessile 
or peduncle 0.8-3 cm; branched portion corymbiform 
to pyramidal, 1.1-12.5 X 1.3-15.5 em, branched to 3 
to 5 orders, lax to congested; bracts deltate to ligulate or 
trifid, 0.8—40 mm; bracteoles lanceolate to triangular, 


rous, er d s x 
cup-shaped, glabrous or 
puberulent, sometimes with hair-ring inside, truncate 
or with lobes 0.1-1.1 mm, triangular; corolla white, 
pink, or lilac, clavate in bud, when open salverform, 
outside glabrous, tube 3-7 mm, 1.5-2.5 mm dia 
inside villous in upper third, lobes 2-3.5 mm, ligulate 
to elliptic-oblong, acute; anthers included or shortly 
exserted, filaments inserted in upper third of corolla 


m., 


tube, 0.3-1 mm; style 3.3-8.2 mm, glabrous, stigmas 
0.8-1.6 mm. Short-styled flowers: similar to long styled 
except corolla tube 4-9 mm, 1.7-3.5 mm diam., lobes 
2.34 mm; m shortly exserted, filaments 1.3— 
3.5 mm; style 2. glabrous, stigmas 1.3— 
3 mm. Drupes mie black, pee or didymous, 4-10 
X 5-10 mm; pyrenes spherical or hemispherical, 
rugose, finely fissured, endosperm entire. 

Distribution and habitat. This species grows in 
Madagascar, where it is known from the provinces of 


Antananarivo, Antsiranana, Fianarantsoa, Mahajanga, 
Toamasina, and Toliara. Here, it is found in moist 
0 


forests at elevations of 0—2 
Phenology. This an has been o with 
January through roug 
mber, and with fruits Jan anuary pum ue and 
rus through December. 


Discussion. Gaertner 


a obovata is 
monly encountered Gaertnera species within Mada- 


Psi 


the most com- 


gascar, in particular its variety o The 
species is distinguished by its ge 
caducous calyptrate stipules; genera 
ovate, obovate, or oblanceolate, acuminate to cuspi- 


nerally sq 
lly elliptic 


date leaves; and terminal cymose inflorescences. This 
i, but lacks the 


inflorescence axes often pubescent in lines that are 


species is similar to G. guillotii, 


characteristic of that species. 

Gaertnera obovata var. obovata and variety sphaer- 
ocarpa are easily distinguished at the extremes of their 
morphological ranges, but are linked by continuous 
morphological variation in several characters and thus 
are not completely distinct. The well-marked, most 
distinct morphological forms differ markedly in leaf 
shape, leaf margin, and flower color: variety obovata 
typically has elliptic-oblong to obovate leaves with flat 
margins and pink to lilac flowers, whereas variety 
sphaerocarpa usually has elliptic-oblong to oblanceo- 
late leaves with crisped margins and white flowers. 

e current circumscription of these varieties is not 
entirely satisfactory; these are separated here as 
outlined below pending additional research. 


la. Leaves pu to m oblong, D E 
obovate, vate, 1.2-8. 0.7-5.2 with 


areis flat coral pink to lilac or ue RA 
styled nim with a 5X 
DEO IRE than i deo dad e qn obovata 
lb. Leaves elliptic oblong to a 4.5-23 
1.9-8.5 cm, with margins flat or crisped; corolla 
white; long-styled [ with anthers shortly 
serted; drupes 6-10 X 5-10 


49a. Gaertnera obovata Baker var. obovata. 


Figure 13A—F. 
1.5-5(-12) m tall; branches 


diam.; 


Trees or shrubs 
flattened to terete, glabrous, 2-3 mm 
nodes 0.3—4.5 cm, smooth or with 2 Wesce ribs 
at least when young. Leaf blades 1.2-7(-8.5) X 0.7- 
5.2 cm, elliptic oblanceolate, 
obovate, or ovate, apex rounded then abruptly shortly 
cute to obtuse, drying chartaceous, 


inter- 


to elliptic-oblong, 


acuminate, base 
glabrous, margins flat; secondary veins 5 to 8 pairs; 
petioles 2-10 mm. Stipules drying membranous to 
chartaceous, tube 12-20 m lobes 5 

Inflorescences with peduncle 0.8-3 cm; branched 


mm. 


Annals of the 
Missouri Botanical Garden 


Figure 13. A-F. Gaertnera obovata Baker var. obovata. —A. Flowering branch (to E -cm o — B. Portion of stem with 


leaf E and stipules. —C. Short- -styled flower. —D. Short-styled flower in cross se 


ns sty! 


Long- ,-styled flower. —L. Long-style 
Malcomber 2882; G-L based on Malcomber 2916. 


portion corymbiform, 1.1-5 X 1.3—4.5 em; bracts 8— 
27 mm; pedicels absent or to 3 mm. Long-styled 
flowers: calyx mm wide, with hair-ring inside, 
lobes 0.1-1.1 mm; corolla white to pink or lilac, tube 
3.5-7 mm, 1.5-2.5 mm diam., lobes 2-3.5 mm, 
ligulate to elliptic-oblong; anthers included, filaments 


0.3-1 mm; style 5.5-8.2 mm, stigmas 1-1.6 mm. 
Short-styled flowers: corolla tube 4—9 m 
3.5 mm diam., lobes 2. mm, ligulate; tocius 


1.3-1.7 mm; style 2.8-5.2 mm, stigmas 1.7-3 mm. 

Drupes 4—5 X 5-6 mm; pyrenes hemispherical. 
Distribution and habitat. 

Madagascar, where it is known from the provinces of 


This variety grows in 


Antananarivo, Antsiranana, Fianarantsoa, Mahajanga, 
and Toamasina. Here, it is found in moist forests at 


elevations of 10-2000 m. 


Phenology. This variety has been collected with 
flowers January through March and May through 
December, and with fruits January through March and 
October through December. 

Representative Pie MADAGASCAR. 


Antananarivo: ESE Anjozorobe, Malcomber 
2838 (MO, TEF); Aa i Reserve, Bernardi 


specimens 
7 km 


ed flower in cross section. C—F to sa 


—E. Long-styled flower. —F 
ertnera obovata var. sphaerocarpa (Baker) Malcombe er. —G. Flowering branch. 
1 n. —K. 


me 5-mm scale; I-L to same 5-mm scale. A-F based on 


11097 (A, BR, G, P); Manjakandriana, Ambatolaona, e 
RN 1487 (P) Antsiranana: 
Reserve, Lewis 130 rojejy Nune 
Reserve, Malcomber 2776 (MO), 3793 (MO), 2799 (MO); 
lower valley of Androranga River, near Ant 
Anjenabe, Hien 24117 (P; M 
Schatz 3211 (MO, P, TAN), Malcomber 1460 (K, MO, TAN), 
2786 (MO); Tsaratanana Nature Reserve, Humbert 18201 (P). 
Fianarantsoa: Andrambovato, Capuron SF 255 (MO); 
Andringitra Integral Reserve, Lewis 1030 (K, MO, P, TAN); 
Antoetra, Jongkind 878 (K, MO, P, TAN); Ranomafana 
National Park, me E 106 (BR, G, K, MO, P, PRE, T 
O, TEF), 2881 (MO), 2882 (MO, TEF). 
. Du, .n. (P). To 
297 (P, TAN); 6.6 km SE of 
avibe, Malcomber 2926 (MO, TEF); Lac 
ka, Humbert 17480 (P); Mantadia 
National Park, Nicoll 170 (MO); Masoala National Park, 
Andranobe, Malcomber 2819 (MO). 


49b. Gaertnera obovata var. sphaerocarpa (Bak- 
er) Malcom e 
Gaertnera ar Baker, J. Linn. 
20: 208. 1883. Sykesia Pee de Need) 
Kuntze, Revis. Gen. PI. . 1891. TYPE: 
Madagascar. Central bus R i. 1243 
designated here, K!; 


nov. Basionym: 


Soc., Bot. 


, comb. et stat. 


(lectotype, isotype, P!). 


Figure 13G-L. 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


Trees 1.5-12 m tall; branches flattened or quadran- 
gular to terete, glabrous or puberulent, 3—9 mm diam.; 
h. Leaf blades 4.5-23 X 


oblanceolate to elliptic-oblong, apex 


internodes 
1.9-8.5 em, 
acute to ee then "ue shortly cuspidate, 


cm, smoot 


base acute to cuneate, drying chartaceous, adaxia 
glabrous, sbaxially glabrous or on erulent, margin flat 
or crisped; d veins 5 to lO(to 16) pairs; 
petioles 3-25 Me e membranous, tube 
8-25(-32) mm, Te 2-3 

or peduncle to 2.5 em; ire portion corymbiform 
to pyramidal, 3.5-8.5(-12.5) X 6.5-12(-15.5) em; 
bracts 0.8—40 


styled flowers: calyx 1.5-2.5 mm wide, glabrous or 


Inflorescences sessile 


mm; pedicels absent or to 2 mm. Long- 


with hair-ring inside, truncate or with lobes to 
.5 mm, triangular; corolla vos tube 3-5. 
2—2.5 mm lobes 2 m, ligulate; anthers 
shortly exserted, filaments 0.3- “05 mm; style 3.3- 
5.8 mm, stigmas 0.8-1.4 mm. Short-styled joue 
similar to long styled except calyx 1.7— 

corolla tube 4-6 mm, 1.7-2.5 mm diam., lobes 23- 
2.6 mm; filaments 1.4—3.5 mm; style 2.7-3.5 mm, 
Drupes 6-10 X 5-10 mm; 


stigmas mm. 
pyrenes phetical or hemispherical. 


Distribution and habitat. 
Madagascar, where it is known from the provinces of 


This variety grows in 


Antananarivo, Antsiranana, Fianarantsoa, Toamasina, 


and Toliara. Here, it grows in moist forests at 


elevations of 0-16 


Phenology. This variety has been collected with 
flowers in January and July through December, and 
with fruits January through April and in December. 

Discussion. ed Sere is similar to Gaertnera 


bat 
MaCTOvOL 


that species. 
Two syntypes were cited in the protologue: R. Bar 
1233 (Ki, P?) and R. Baron 1243 (K!, P!). The hie is 
selected as lectotype here because it is a more 
complete and exemplary specimen. The specimen at K 
is selected as the lectotype because the author of the 
name worked there. 


Representative specimens examined. MADAGASCAR. 
ananarivo: 5.7 km ESE of Anjozorobe, ric ded 
2836 (MO, TEF), 2837 (MO, TEF). Antsiranana: Andap: 


Lokoho Basin, Ankobahina Hills, Humbert 22019 (MO, P. 
Marojejy Nature Reserve, Malcomber 2783 (MO, TEF), 2788 
(MO, TEF) 2789 (MO). Fianarantsoa: Andrambovato, 
Humbert 28500 (MO, P); Farafangana, Ivongo, Rakotovao 
593 (TEF), 607 ea Ranomafana National Park, Vohipar- 
ara, Malcomber 2857 (MO, TEF), 2859 (MO, TEF); W o 
| Itremo Massif, Malcomber 2839 (MO, TEF), 2840 
(MO, TEF). T. ibe, Mantadia National 

Birkinshaw Malcomber 2917 (MO), 2918 
(MO, TEF), 2979 (MO, TEF), 2928 (MO, TEF), 2929 (MO, 
TEF); Analamay rd., Rakotomalaza 1317 (MO, TAN); Didy, 
Brickaville, Cours 4790 (MO, P) Toliara: Andohahela 
Massif, Ranohela Valley, Humbert 6250 (MO, P); Beampin- 


E 


'oamasina: near Andas 


a Massif, Maloto Valley, Humbert 6322 (MO, P); Fort 
uphin, Mandena, Bosser 14413 (TAN). 


50. Gaertnera paniculata Benth., Niger Fl. 459. 
1849. Sykesia paniculata (Benth.) Kuntze, Revis. 
Gen. Pl. 2: 425. 1891. TYPE: Nigeria. Grand 
Bassa, S. Vogel 71 (lectotype, designated here, 
K) 


Gaerinera occidentalis Baill., Bull. Mens. Soc. Linn. Paris 1: 
YPE: Gabon. s. loc., G. du Bellay et al. 

232 (holotype, P!). 
Trees or shrubs, (1.2-)2-9 m tall; branches flat- 


tened to terete, when young glabrous o or puberulent, 


3-9 em, elliptie to elliptic-oblong or oblanceolate, 
apex shortly cuspidate or acuminate, base cuneate to 
obtuse, drying chartaceous, adaxially glabrous, abaxi- 
ally glabrous or sparsely pilosulose to hirtellous on 
principal veins; secondary veins prominent abaxially, 
3 to 8 pairs; domatia absent or present; petioles 5— 
17 mm. Stipules tubular, glabrous, drying chartaceous, 
caducous, persisting on distalmost 1 to 3 nodes or 
deciduous leaving persistent base 1-5 mm, tube 8— 
mm, with ribs 4, narrowly winged, arising below 
petiole and extending to lobes, apex entire, marces- 
cent, lobes 4, 0.5-1.2 mm, deltate to narrowly 
triangular. Inflorescences cymose to paniculiform, 
many-flowered, terminal on principal and/or axillary 
branches, densely puberulent to glabrescent or 
pilosulose, sessile or peduncle to 8 em; branched 
portion corymbiform or narrowly pyramidal, 7-20 X 
4—25 em, rs to 4 to 6 orders, lax; bracts deltate 
or trifid, .2 mm, sometimes ad. Caan 
reduced; d absent o ers 
long-stoled. pai M 
cup-shaped, 1.2-2 mm wide, glabrous or puberulent 


merous, heterodistylous. 


outside, with hair-ring inside, truncate or lobes to 
n bud, when 
utside densely oe tube 3- 


1 mm, triangular; corolla white, clavate 
open salverform, o 
4 mm, 1-2.5 mm diam., inside villous in upper third, 
lobes 1.5-2.5 mm, ligulate to linear, acute; anthers 
shortly exserted, filaments inserted in upper third of 
corolla tube, ca. 0.3 mm; style 6—7 mm, ps or 
pubescent in upper part, pd 0.5-1.5 mm. Short- 
styled flowers: similar to long styled REM corolla 
outside tomentulose or puberulent, tube 2.5-4 mm 
lobes 1.5-2 mm, 
filaments 2-2.5 m mm, 7 
m. Drupes violet-black, pine or subglobose, 7— 


ligulate; Pes fully exserted, 


mm; style stigmas 
De x Pt 9 mm; pyrenes spherical or hemispherical, + 
rugose, finely fissured, endosperm entire. 


Distribution and habitat. 
West and Central Africa, where it is widespread and 


This species grows in 


Annals of the 
Missouri Botanical Garden 


known from Angola, Benin, Burkina Faso, Cameroon, 
the Central African Republic, the Democratic Republic 
of the Congo, Gabon, Ghana, Guinea, Cote d'Ivoire, 
Liberia, Mali, Nigeria, the Republic of the Congo, 
Senegal, Sierra Leone, Togo, and Zambia. Here, it is 
found in humid forests at elevations of 0-1720 m. 


Phen This 


enolog species has been collected with 
flowers = ae throughout the year. 


Discussion. Gaertnera paniculata is similar to G. 
but differs in its 


externally densely puberulent or tomentulose corolla 


vaginans and G. longivaginalis, 


ubes. This is the most commonly collected species of 
Gaertnera in Africa. 
Two syntypes were cited in the protologue of 
aertnera paniculata, S. Vogel 70 (K!) and S. Vogel 71 
(K!). The latter is selected as lectotype here because it 
is a more complete and exemplary specimen. 


Representative specimens examined. ANGOLA. Dundo, 
Gossweiler 14108 (K); Estrada, Raimundo 2846 (BR). 

NIN. Natitingau, Aubréville 81 (P); Tohoué, Porto-Novo, 
Chevalier 22774 (P). BURKINA FASO. Guenako, Toutain 
2408 (P). C 


> 
= 
m 
zal 
o 
O° 
Z 
E 
-1 
Pr 
E 
un 
e 
A 
z: 
o 
de 
m 
= 
de 
E 
3 
M 
= 


` P. 
CENTRAL AFRICAN REPUBLIC. Bouar, Audru 29996 (P); 
near Mbanza, Km 15 Nola—Salo rd., Leeuwenberg 7106 (BR, 
K, P, WAG). DEMOCRATIC REPUBLIC OF THE CONGO. 
Eala, Corbisier 1266 (FHO, K, P, WAG); Kiri, Lac Leopold, 
Goossens re (BR); Madimba, Mpese, M Ur 2067 (P, 
WAG). GABON. Haut-Ogooué: Km 23 Moanda-France- 
ville rd., Breteler 6275 (BR, MO, P, WAG). Ogooué-Lolo: 
Fatou ville, Boucimbi, 
ANA. Hahoe, Togo Plat 
(E, FHO, ee GUINEA [GUINEA- CONAKRY]. 
Friguiagbé, Mayou-Dowie, Chillou 800 (MO, P); Mt. Nimba, 
Schnell 1123 (p, IVORY COAST [CÓTE D'IVOIRE]. Aplati, 


(MO, 


ur dra Farak 21 012 MO, NI- 
GERIA. irme m Onitsha, Okeke FHI 38440 a 


LIC OF THE CONGO. Brazzaville, Chevalier 11206 (P); La 
Lifoula, de Néré 1355 (P, WAG); Stanley Pool, Schlechter 
12549 (BR, G, G-DC, L, P). SE oc., Perrottet 41 

(G-DC). SIERRA LEONE. Mt. Loma, Jaeger 7688 (G, WAG); 


Sherbro Island, Hunter 28 (BM). TOGO. Kloto, Pisch 7692 
(MO, WAG). ZAMBIA. Mwinilunga, White 3311 (BM, BR, 
FHO, K, WAG) 


51. Gaertnera pauciflora Malcomber & A. P. 
Davis, Monogr. Syst. Bot. Missouri Bot. Gard. 
104: 392, figs. 5, 7. 2005. TYPE: Madagascar. 
Antsiranana: wae Nature Reserve, 1300— 

T. Malcomber 2781 
ree MO- caren ones, Al, BR!, Gl, 
K!, LE!, MAL!, MAU!, MO-5714903!, P!, PRE!, 
TEF!, WAG). 


Trees or shrubs, 1.5-6 m tall; branches flattened to 
terete, glabrous, 1-3 mm diam., with bark becoming 
gray-white; internodes 0.4—2.6 em, smooth or with 2 
al ridges. Leaf blades 1-3 X 0.4- 


1.5 em, elliptic to elliptic-oblong or oblanceolate, 


low longitudin 


apex shortly cuspidate or acute, base acute to usually 
cuneate, drying chartaceous, glabrous; secondary 
veins visible and flat to prominulous abaxially, 4 to 
6 pairs; domatia absent or present; petioles 1.5-3 mm 
Stipules calyptrate, glabrous, drying membranous, 
caducous, tube 7-21 mm, with ribs 4, narrowly 
winged only beneath petiole, apex with 1 or 2 
incisions, lobes 2 or deltoid to filiform. 
Inflorescences reduced to 1 flower or 2-flowered and 
fasciculate, terminal on axillary branches or axillary, 
glabrous; peduncle 2.5-9.1 mm, often articulated in 
upper half; bracteoles elliptic to ovate, 1—4 mm. 
Flowers 4(5)-merous, heterodistylous. Long-styled 


flowers: calyx cup-shaped, 1-2 mm wide, glabrous, 


truncate or lobes 0.2—0.4 mm, broadly triangular; 
corolla white, clavate in bud, when open salverform, 
i , 2.53 
inside glabrous, lobes 6—8 mm, ligulate to linear, 


outside glabrous, tube 4—7.5 mm mm diam., 
acute; anthers included, filaments inserted in lower 
third 0.2-1 mm; style 5-6 
glabrous, stigmas 0.5-1 mm. Short-styled flowers: 


of corolla tube, mm, 

similar to long styled except calyx 2.2-2.8 mm wide, 

lobes 0.1-0.3 mm; corolla tube 5-7.5 mm, 2.5-3 mm 

diam., lobes 6-7 mm; filaments 1.5-2.5 mm; style 
1.8— m, stigmas 0.5-1.2 mm. Drupes violet- 

black, subglobose to didymous, 6-7 X 

cal or hemispherical, — rugose, finely 


—7.5 mm; 
pyrenes spheri 
fissured, endosperm entire. 


Distribution and habitat. 
Madagascar, where it is known from the provinces of 


This species grows in 
Antsiranana and Fianarantsoa. Here, it is found in 
humid evergreen forests on metamorphic and igneous 


rocks at elevations of 750-1800 m 


Phenology. This species has been collected with 
flowers January through March and September 
through December, and with fruits January through 
March and October through December. 


Discussion. Gaertnera pauciflora can be sep- 


arated from the other few-flowered  Gaertnera 
species in Madagascar by its smooth or only slightly 
ridged internodes, lack of pubescence, calyptrate 
stipules, d subtruncate to denticulate calyx 
Malcomber and Davis (2005) considered the conser- 


vation status of this species to be Endangered (IUCN, 
2001 


V aces specimens examined. MADAGASCAR. 
tsir of Andapa, Marojejy Nature Reserve (RNI 


#12), A camp 3 summit trail, Malcomber 2779 (MO, 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


TEF), 2780 (MO, TEF). Fianarantsoa: Ifanadiana—Tolon- 
goina, ngalanihelatra, Namorona River, Capuron SF 


23220 (P). 


52. Gaertnera pendula Bojer, Hortus Maurit. 217. 
7. Sykesia pendula (Bojer) Kuntze, Revis. 
25. 1891. TYPE: Mauritius. 
Montagne de Pouce, W. Bojer s.n. (holotype, P 
not located; isotype, G-DC!). 


ie lanceolaia Bouton ex A. DC., Prodr. 9: 33. 1845, 
aerinera lanceolata Ridl., 1915. TYPE: Maur- 
Tus La Nouvelle Decou "ed Bouton s.n. 
(holotype, G-DC!; isotype, K!). 
de 2 lla Drake, Bull. Soc. Bot. France 45: 354. 
, hom. illeg., non Gaertnera wd enth., 
D TYPE: acs s. loc., L. A. Chapelier s.n. 
(holotype, P5. 


Trees or shrubs, 1.5—4.5 m tall; branches flattened 
to terete, glabrous, 2—4 mm diam.; internodes 0.8— 
8 cm, smooth. Leaf blades 4.5-17 X 2-8 cm, elliptic 
to ovate or oblanceolate, apex shortly cuspidate or 
acuminate, base acute to obtuse or rounded, drying 
chartaceous to subcoriaceous, glabrous; secondary 


r pre sent, often 


ous, 
en aie 
winged, arising bios a and extending to lobes, 
apex with 1 or 2 incisions, marcescent, lobes 4, 1— 
3.3 mm, deltoid or filiform. Inflorescences cymose, 
many-flowered, terminal on flexuous axillary branch- 
es, glabrous or puberulent; peduncle 3.5-9(-13) em, 


lax to 


flexuous; branched ce gp aaa 1.8-3.5 X 
3.3 hed 2 3 orders, 


m, branche 
eine what congested; m e 3-12 mm; brac- 
teoles triangular, 1-2 mm; pedicels absent or to 
mm. Flowers 5-merous, heterodistylous. Long- 
styled flowers: calyx campanulate, 1.5-3 mm wide, 
ae to glabrous outside, glabrous inside, 
lobes 1.5-3.5 mm, narrowly triangular o 
E corolla white or pale blue, cla 
when open salverform, a glabrous, tube 16— 
24 mm, 2-3.5 mm 


ligulate or elliptic-oblong, acute; 


or oblong- 
vate in bud, 
iam., inside glabrous, lobes 8— 
9 mm, anthers 
included, filaments inserted in upper third of corolla 
tube, to 0.5 mm; style 18-19 mm, glabrous, stigmas 
2.5-3 mm. Short- pe dd similar to long styled 
except calyx sparsely escent outside; l 
tube 20-25 mm, e mm diam., lobes 8-10 mm, 
gulate or ovate-oblong; anthers shortly exserted, 
an 0.5-1 mm; style 9 m, stigmas 3.5— 
3.7 mm. Drupes black (Bojer, 1837) or violet-black, 
ellipsoid or fusiform, 20-25 X ca. 10 mm; pyrenes 
finely fissured, endosperm 


plano-convex, rugose, 


entire. 


Distribution and habitat. This species grows in 
Mauritius and may also grow in Madagascar, but that 

report is based o documented 
specimen and seems pe uritius, this 
species is found in wet to very wet forests at elevations 


of 50-650 m 


Phenology. This species has been collected with 
flowers in January, February, and October through 


December, and with fruits in June. 


Discussion. The collections of Gaertnera pendula 
attributed to Madagascar appear to have inaccurate 


labels. 


pendulous inflorescences borne at the ends of long 


This species can be recognized by its 


exuous branches, white or pale blue flowers, and 
ellipsoid or fusiform fruits. This is the only Gaertnera 
species reported to have blue flowers. 


Representative specimens examined. MADAGASCAR. s. 
loc., L. S. Bouton 


Park, Malcomber 2933 (MO), 2936 (MO); in Pidgeon wood 
opposite Les Mares, Page 94 (MAU). 


53. Gaertnera phanerophlebia Baker, J. Linn. 
oc., Bot. 21: 425. 1885. Tp uae Qu. 
(Baker) Kuntze, Revis. Gen. : 425. 1891. 
TYPE: Madagascar. Central d on R. 
Baron 2982 (lectotype, designated here, Kl; 
isotypes, BM!, P!). Figure 11G— 


ub crinita Drake, in Grandid., Hist. Phys. Madagascar 
36(6), Adas 4: t. 432. 1897 [1898], syn. nov. TYPE: 
Madagascar. [Ile] Sainte Marie, 1850 (f1), L-H. Boivin 
1780 (lectotype, designated here, P!; isotype, MO!). 
Shrubs and trees, 3-8 m tall; branches flattened to 
terete, 2s to hirsute (glabrous) with indumentum 
red-bro 1.5-5 mm diam.; internodes 1.7— 
BG cm, s pe blades 2.5-12.5 X 1-3.5 em, 
elliptic-oblong to oblanceolate, apex acute to acumi- 
nate, base obtuse to truncate, drying chartaceous, 
adaxially glabrous or densely strigose on costa and 
sometimes secondary veins and occasionally sparsely 
hispid on blade, abaxially densely strigose to hirsute 
(glabrous) on principal veins to throughout; secondary 
veins prominulous abaxially, 6 to 14 pairs; domatia 
absent; petioles 1—7 mm. Stipules calyptrate, densely 
hirsute or strigose (glabrous), drying membranous, 
caducous, tube 11-30 mm, without ribs, with narrow 
wings only developed below petiole, apex with 1 or 2 
incisions, marcescent, lobes 2 mm, deltate. 
Inflorescences subcapitate to Mc MEN many- 
flowered, terminal on principal and/or axillary 
branches, densely pilose to strigose (glabrous), sessile 
or peduncle to 4.3 em; branched portion subglobose 
or corymbiform, 0.7—4.5 X 1-8.5 cm, branched to 2 to 


Annals of the 
Missouri Botanical Garden 


4 orders, congested; bracts deltate to linear, 4— 
20 mm; bracteoles 1-2 mm; pedicels absent or to 
3 mm. Flowers 5-merous, heterodistylous. Long-styled 
flowers: calyx cup-shaped, wide, outside 
glabrous or pilosulose to strigose, glabrous inside, 
lobes equal to unequal, 2-5 mm, linear to narrowly 
triangular; corolla white, clavate in bud, when open 
salverform, outside pilose to Pilosulose (glabrous), 
m., inside villous at 


lobes 1.5-6 mm, 


triangular or ligulate, acute; anthers included, fila- 


tube 9.5—1/ mm, 2—4.5 mm 
ca. middle or in upper third, 


ments inserted at ca. middle of corolla tube, ca. 
0.4 mm; style 9.5-16.5 mm, glabrous, stigmas 1.4— 

mm. Short-styled flowers: 
except calyx 1.5-3.5 mm wide, outside pubescent to 


similar to long styled 


pilose or glabrous, lobes 5-13.9 mm, triangular to 
linear; corolla tube 1.8—4.5 mm diam., lobes 1.7— 
5 mm, filaments inserted in ae third of corolla 
tube, 1.2-3.5 mm; st —7 m, stigmas 1.5— 
-6 mm. Drupes violet- black. uhal bood or didymous, 
6-9 X 6-9 mm; pyrenes spherical or hemispherical, 


rugose, finely fissured, endosperm entire. 


Distribution and habitat. This species grows in 
Madagascar, where it is known widely through the 
northeastern region. Here, it is frequent in wet forests 


at elevations of 50-1650 m 


Phenology. This species has been collected with 
flowers January through March and June through 
December, and with fruits January through March and 
August through December. 


Discussion. Gaertnera phanerophlebia is a distinc- 
tive species that can be recognized by its subglobose 
to corymbiform, congested-cymose to subcapitate 
inflorescences, well-developed narrow calyx lobes, 
reddish brown to chestnut drying color, and usually 


dense indumentum. Occasionally, plants are glabrous 
but otherwise match this species (e.g., McPherson & 
van der Werff 16360, MO) and are included here; such 
variation in pubescence is found in some other species 
of Rubiaceae. This species is similar to G. hispida; 
see additional comments under that section of this 
paper. 

The protologue of Gaertnera crinita does not discuss 
the comparison or contrast of this species with the 
similar species G. phanerophlebia, so it is unknown if 

randidier was familiar with G. Viera d e The 
figure in the protologue of G. shows both 
flowers and fruits, and matches dais PN here 
in G. phanerophlebia in all details; accordingly, these 
names are synonymized here. No specimen was 


mentioned in G. crinita’s protologue, but several 
P. Thes 


specimens are annotated with this name at 


generally match the figure and are all identified as 


Boivin 1780 with the flowering collections recorded as 
collected in 1850 and the fruiting collections in 1849. 
No other collections were found annotated with this 
name. Because the fruits of most Gaertnera species are 
similar, the flowering specimens that show all the 
features that distinguish this species and are dated 
1850 are chosen as the lectotype. The specimen at P 
is chosen as the lectotype because the author of the 
name worked at that institution, and the specimen 
deposited at MO was distributed only recently. 

Two syntypes were cited in the protologue of 
Gaertnera phanerophlebia, R. Baron 2372 (K!, P!) and 
R. Baron 2982 (BM!, K!, P!). The latter is selected 
here as lectotype because it is a more complete an 

or The 


specimen at K is chosen as the lectotype because 


exemplary specimen with duplicates. 
the author of the name worked at that institution. 

Gaertnera phanerophlebia apparently hybridizes 
with some other Gaertnera species. In particular, one 
specimen, P. J. Rakotomalaza et al. adagas- 
car, Majunga, 1200 m, MO), seems to represent a 
hybrid between G. phanerophlebia and either G. 
humblotii or G. raphaelii. 

Representative specimens examined. MADAGASCAR. 
Anisiranana: Manongarivo N 


kolosy, Gachet SF 7531 (TEF); Marojejy Nature Reserve, 
e 22283 1 31467 (MO, P), 31669 (MO, P), Miller & 
Randrianasolo 4488 Toamasina: Analalva, W of 
Foulpointe, Capuron SF 2384 
eserve, Antsavokabe, Ravelonarivo 903 (MO, P, TAN); 
Didy, Brickaville, Cours 4893 (P); fle Saint-Marie, Boivin 
1780 (G-DC, MO, P); Vohimarongitra Reserve, Rakotoniaina 
RN 2448 (MO); Zahamena National Park, Humbert 17618 
(D), Botoalina RN 3192 (MO, TAN). 


54. Gaertnera phyllosepala Baker, J. Linn. Soc., 
Bot. 20: 207. 1883. Sykesia dou Pa 
Gen. Pl. 2: . 1891. TYPE: 
Madagascar. Central "eben R pun 
(holotype, K!; isotype, P). Figure 14A—F. 


Kuntze, Revis. 


Trees, 1.5-8 m tall; branches flattened to terete or 
subquadrate, when young densely hispidulous to 
strigose or pilose with indumentum drying yellow to 
gray, when older, trichomes often breaking off leaving 
prominent persistent basal portions, 3—7 mm diam.; 
internodes 2-8.5 cm, smooth. Leaf blades 6.5-20.5 X 

.5-7.5 em, lanceolate to elliptic, oblanceolate, or 
TM oblong, apex shortly acuminate, base d 
to truncate or EA drying chartaceous, adaxially 
glabrous except pilosulose to pilose on costa and 
sometimes secondary veins, abaxially pilose or 
dto with indumentum on principa 


9 to 12 
pairs; pine absent; petioles 3-11 mm. Stipules 


enser 
veins; secondary veins prominent abaxially, 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


Figure 14. A-F. Gaerinera phyllosepala Baker. 
aa —C. Short-styled flower. 


n cross sectio 
ules 


n. G-M. Gaerinera ternifolia Thwaites. 
stipules. —I. Portion of flowering bra 


Flowering branch. 


—G. 
—J. Short-styled flower. 


—A. Flowering branch. —B. Portion of stem with petiole bases and 


ort nd 
—K. Short- ver ee in cross section. —L. Long-styled 


flower. —M. Long-styled flower in cross section. C—F to same 1-cm scale; J-M to same 1-cm scale. A-F based on Moise 6; G— 


M based on Malcomber 2767. 


tubular, pilose to hirsute or strigose, drying charta- 

ceous, persistent at least on distalmost several nodes, 

tube 8-21 mm, with ribs 4, narrowly winged, arising 

below petiole and sometimes extending to os apex 
ith 1 o scent, lobes 

deltate to linear. Inflorescences cymose, pee 


or 21 incisions, marces 


ered, terminal on axillary branches, pilosulose to 
pilose; peduncle 1.8-8 em; branched portion corym- 
biform, 4.5-14 X 3-19 cm, branched to 3 to 5 orders, 
lax or somewhat congested; bracts lanceolate to 
elliptic or ovate, 5-30 mm, white, sometimes gla- 
brous; bracteoles 5-10 mm, white; pedicels absent or 
to 3 mm. Flowers 5-merous, heterodistylous. Long- 
styled flowers: calyx cup-shaped or campanulate, 2.4— 

mm wide, outside densely pilosulose or puberulent, 
glabrous inside, lobes ca. equal in size or 1 lobe 
larger, 5-9.4 mm, narrowly lanceolate to elliptic or 
ovate, white; corolla white, clavate in bud, when open 
infundibuliform or salverform, outside densely pilo- 
, tube 14-16 mm, 1.54 mm 
inside glabrous, lobes 4-6 mm, triangular to 


sulose or puberulent 
lam., 1 
ligulate, acute; anthers included, filaments inserted in 


upper third of corolla tube, 0.3-0.7 mm; style 14— 


16.5 mm, glabrous, stigmas 1.5-2.5 mm. Short-styled 
flowers: similar to long styled except calyx 2-3 mm 
wide, lobes 6 mm, lanceolate or elliptic; corolla 

tube 15-22 mm, 1.5-4.5 m bes 2.5 
acute; filaments 2-3.5 mm; style 8.5-11 mm, um 
2 


diam., lo! 


.5-3.5 mm. Drupes violet-black, subglobose or di- 
dymous, 6-7 X 7-8 mm; nes spherical o 


hemispherical, rugose, finely fissured, endosperm 


entire. 
Distribution and. habitat. 
Madagascar, where it is known from the provinces of 


This species grows in 
Antsiranana, Fianarantsoa, and Toamasina. Here, it is 
found in humid forests at elevations of 0-1200 m. 


Phenology. This species has been collected with 

owers in January, February, and October through 
December, and with fruits January through May and in 
December. 


Discussion. | Gaertnera phyllosepala is an attractive 
and commonly encountered species with showy, well- 
developed white calyx lobes and bracts. This species 
is similar to G. phyllostachya; G. phyllosepala can be 
recognized by its densely hispidulous to strigose or 


Annals of the 
Missouri Botanical Garden 


pilose branches, tubular stipules, shorter calyx lobes 
and bracteoles, 5-10 mm long, and regularly lobed 
calyx limbs with the lobes rounded to obtuse or shortly 
acuminate, versus glabrous to puberulent branches, 
calyptrate stipules, calyx lobes and bracteoles 6— 
16 mm long, and truncate or 1- to 3-lobed calyx limbs 
with the lobes sharply acute in G. phyllostachya. 


a specimens examined. MADAGASCAR. 

tsiranana: Antalaha, Ampanavoana, Ranjokiny RN- 
10845 (P). “Fiánaranfsoa: n 
Collector SF 13604 (P, 
Croat 32613 (MO, TAN), D'Arcy 15273 (MO, P); Mananara- 
Nord National Park, Malcomber 2910 (MO, TEF), Rahar- 
(P; Masoala National Park, Andranobe, 
Malcomber 2814 (MO, TEF), Schatz 1348 (MO, P, TAN, 
WAG), wo (MO, P, TAN), Van Nek 2107 (TAN), Zjhra 133 
(MO, TAN); Soanierana- p Fenerive, Unknown 
ee SF 1058 (P, EF); Zahamena (RNI 3), 
Boioalina RN 607 (MO), pucr 16786 (P). 


55. Gaertnera phyllostachya Baker, J. Linn. Soc., 
Bot. 21: 425. 1885. Swkesia phyllostachya 
n Kuntze, Revis. Gen. Pl. 2: 425. ale 
TYPE: Mada s. loc, R. Baron 2327 
Merlo: od here, K?). 


Trees, 5-10 m tall; branches terete to flattened or 
subquadrangular, when young glabrous or puberulent 
at least near nodes with indumentum drying yellow, 
becoming glabrescent, 2-4 mm diam.; internodes 2— 
7 em, smooth. Leaf blades 3.5-15(-17) X 1-5.5(-8) 
cm, elliptic to oblanceolate or oblong, a 
then shortly cusp 
cuneate aes drying chartaceous to subcoria- 


apex obtuse 
ate or acuminate, base attenuate or 
ceous, glabrous or pilosulose along midrib abaxially; 
secondary veins prominulous abaxially, 5 to 7(to 12) 
pairs; domatia absent or present, pilosulose to deep 
crypt-type; 
glabrous to densely pilosulose or puberulent, drying 


petioles 4-8 mm. Stipules calyptrate, 


membranous, caducous, deciduous through fragmen- 
5(-55) 


s 4, narrowly winged, arising beneath 


tation, or ae persistent, tube 10-3 
mm, wit 
petiole and. sometimes extending to apex, apex with 1 
incision, lobes 1 or 2, 1-2 mm, deltate to linear. 
terminal on 
rulent; 1 1.4—4.5 cm; 
2-10(-20) X 2.7- 
18.5 em, branched to 3 to 4 orders, lax or somewhat 


Inflorescences cymose, many-flowered, 
anches, pube 
d portion corymbiform. 
congested; bracts lanceolate to ovate or trifid, 15— 
20 mm, white, adaxially glabrous, abaxially glabrous 
; bracteoles lanceolate to ovate, 

5-merous, heterodistylous. Long-styled flowers: calyx 
1-2.5 mm 


glabrous inside, truncate or some flowers with 1 to 


campanulate, wide, outside puberulent, 


lobes, lobes equal to unequal, 6-15 mm, elliptic to 
narrowly elliptic or elliptic-oblong, white; corolla 


white, clavate in bud, when open salverform or 


infundibuliform, outside glabrous, tube 9-11.5 mm, 


ER 


—2.5 mm diam., inside glabrous, lobes 3.54: mm, 
ligulate or EUN M acute; anthers included, 
filaments inserted in upper third of corolla tube, 0.5— 
1 mm; style 11-13 mm, glabrous, stigmas 0.7-2 mm. 
Short-styled flowers: similar to long styled except calyx 
0.7-2.5 mm; corolla tube 10-16 mm, 2-4.5 mm 
diam., lobes 3—4 mm; anthers shortly exserted, 2.5— 
3 mm; style 9-10 mm, stigmas 3-3.5 mm. Drupes 
violet-black or blue, globose, 4-8 X 5—7 mm; pyrenes 
spherical or hemispherical, + rugose, finely fissured, 
endosperm entire. 


Distribution and habitat. 
Madagascar, where it is known from the provinces of 


This species grows in 


Antananarivo, Antsiranana, Fianarantsoa, Mahajanga, 
and Toamasi 


sina. Here, it grows in humid forests at 
elevations of 100-12 


Phenology. This species has been collected with 
flowers in January, February, November, and Decem- 
ber, and with fruits January through May and in 
November and December. 


Discussion. The calyx of Gaertnera phyllostachya 
appears at first glance to be regularly lobed, but the 
top of the limb is actually truncate or denticulate in 
most flowers; in a minority of the flowers it is truncate 
with one to three well-developed lobes. The flowers 
are also regularly subtended by two showy bracts that 
are fused to the base of the calyx and seem to have 
been confused with calyx lobes by some authors, but 
these clearly arise well below the top of the calyx 
limb. Similarly, though, smaller bracts are borne at 


variable, from slender plants with relatively small 
leaves and deciduous stipules (e.g., Miller & Miller 
3821 [MO]; Rasoavimbahoka et al. 69 ee to robust 


plants with relatively large leaves and larger persis- 
tent stipules (e.g., Schatz & Miller 2432 mo Nicoll 
202 [MO)). The presence, number, and form e leaf 


domatia also vary notably, apparently cod corre- 
lation to these other characters. 

This species is largely restricted to eastern forests 
of Madagascar at 100-1200 


has also been collected in western forests near 


m altitude; however, it 
Mahajunga, at an elevation not recorded (D'Alleizette 
1477, P). Gaertnera phyllostachya is similar to G. 
phyllosepala; see additional comments under that 
o 

Three syntypes were cited in the protologue, R. 
Baron 2327 (K D, R. Baron 2683 (K), and L. Humblot 
510 (K!, P!). The first specimen is selected here as 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


Figure 15. 


young leaves. —C. Short-styled flowe: 


A-F. Gaerinera Dp Pa Baker. e sa branch. —B. at d stem wiih i stipiles Es 
hor —D. Sho led flov 

. G-N. [oem inera vaginans (DC. ) Merr. E SS ben 
. Portion of stem with petiole EAE aa and stem apex 
flower in cross section. —M. Short-styled flower. —N 
to same 3-cm scale; I, J to same 1-cm scale; 


ur Tie braich. E 
E eae 


C-F 
K-N to same ce -cm sele. ru F based on fond 


2931; G, I-N based on | Malbondier 2760; H based on Malcomber 2761. 


lectotype because it is a more complete and 


exemplary specimen. 


Pp examined. MADAGASCAR. 
Anke eramadrinka, Scott-Elliot " Ae (K). 
Antsiranana: Wes ejy Nature ee Miller 3918 (MO, 
P), ia 70 (MO, T Humbert 22106 (P). 
Fianarantsoa: ature se iin Lewis 786 (MO, 
2 (MO, P, TAN), 
9 (K, MO, P, TAN), 2866 (MO), verdof 24 
(MO), Schatz & Miller 2432 (MO, P, WAG). Mahajanga: 
near Mahajanga, D'Alleizeite 1477 (P). Toamasina: Ambo- 
wr p 1843 (P, TAN); Analamazaotra, Capuron SF 
568 (P, TEF), C. D’Alleizetie s.n. (P), Perrier de la Báthie 
4014 A P); Andasibe (Perinet), Cours 4414 (P), 830 m, 
Croat 32246 (MO, TAN, WAG), Dorr 3080 (MO, P, WAG), 


kde c 
Antananariv 


— 


Decary 15297 (M 

x a 2578 (TEF), 6907 (P), 708 5 (Py Zahamena 

Nature Reserve, Ramanatsoavina RN 2814 (MO, TA 

56. Gaerinera psychotrioides (DC.) Baker, Fl. 
Mauritius 231. 1877. Basionym: Chassalia 
psychotrioides DC., Prodr. 4: 531. 1830. Sykesia 


ns (DC.) Kuntze, Revis. Gen. Pl. 2: 
425. 1891 : Mauritius. s. loc., F. W. Sieber 
57 a G!; isotypes, E!, K!, MO!, OXF!, P!, 
Wt, WU!). Figure 15A-F. 


Chassalia coffeoides DC., Prodr. 4, ws 1830. P ds 


chasalioides D. Dietr., Syn. Pl. 1: 777. 1839, 
superfl. illeg. TYPE: Maurits. 8. oa F. W. Sieber 335 
(holotype, G-DC not located). 


Gaerinera bifida Bojer, Hortus Maurit. 217. 1837. TYPE: 
Mauritius. Quartier Militaire and Moka, W. Bojer s.n. 
(holotype, P not located; isotype, BM!). 

Gaerinera parviflora Bojer, Hortus Maurit. 217. 1837. TYPE: 

auritius. Savanne and Trois llots, W. Bojer s.n. 
(holotype, P not ue 

Gaerinera quadriseta A. DC., Prodr. 9: 34. 1845, syn. n 
TYPE: Mauritius. s. ie, W. Bojer s.n. (holotype, G- DC 
microfiche!). 

Gaerinera cae var. "a bur A. DC., Prodr. 9: 34. 

1845, s v. TY uritius. s. loc., F. W. Sieber 

272 (rm uo o BM!, E!). 
Gaerinera quadriseta var.  platypoda A. DC., Prodr. 9: 34. 
4. . TYPE: Mauritius. s. loc., W. Bojer s.n. 


A. DC., Prodr. 9: 34. 
1845, s : Mauritius. s. loo, F. W. Sieber 
332 nee i ys nt. “isotypes, Et, P5. 


Annals of the 
Missouri Botanical Garden 


Gaerinera DUM var. 6 Ld A. DC., Prodr. 9: 34. 
845, “Tle de France ou Bourbon,” 
coll., C». 

Gaerinera truncata A. DC., Prodr. 9: 34. 1845, syn. nov. 
TYPE: Mauritius. s. loc., F. W. Sieber 54 (holotype, G- 
DC isotypes, E!, G!, K!, L!, P!, W. 


Trees or shrubs, (0.6—1.8-12 m tall; branches 
terete, glabrous, 2-8 mm diam. internodes l- 
mooth. Leaf blades 2.5-1 1.1- 


8.3 em, oblanceolate to elliptic-oblong, elliptic-lan- 


ceolate, or elliptie, apex rounde acute to 
acuminate, base cuneate to acute, drying coriaceous, 
glabrous; secondary veins visible and flat to promi- 


irs; domatia n sent; 


A 


mm. Stipules de gla o 
puberulent to pubescent, drying deu ps 
tent or sometimes fragmenting, tube 2-12 mm, with 
ribs 4, narrowly 
angling to meet in mid 


winged, arising beneath petiole, 
dle of interpetiolar side and 
sometimes extending to lobes, apex entire or with 2 


lobes 4, 1-5 mm, 


Inflorescences cymose, many-flowered, 


incisions, marcescent, linear. 
terminal on 
axillary branches, puberulent or glabrous; peduncle 
3.556 cm; branched e corymbiform, 1.5-10 X 
3-15 em, branched to orde 
er Hed. bracts ae to linear or trifid, 


12 


rs, lax or usually 


3-9 mm; bracteoles triangular to 


merous, heterodistylous. Long-styled flowers: calyx 
cup-shaped, 2—4 mm wide, glabrous, truncate or lobes 
to 0.2 mm, triangular; corolla white, clavate in bud, 

when open salverform, outside puberulent or glabrous, 
tube 10-12.5 mm, 2-2.5 mm diam., inside 
upper third, lobes 4.5—5.5 mm, ligulate to linear, 


villous in 

acute; anthers e exserted, filaments e in 
hi rolla tube, 5n yle 

17 mm, glabrous e near apex), stigmas (3 


upper third of c 


1.5 mm. Short-styled flowers: similar to long styled 
except corolla tube -l4 mm, 2.5-3 mm diam., 
lobes 4—5 mm; filaments ca. 2 mm; style 9-10 mm, 
stigmas T mm. iid violet-black or blue, 
ellipsoid, 7 X 6 
rugose, ae nd is entire. 


; pyrenes plano-convex, 


Distribution and. habitat. 
Mauritius, where it is found in wet evergreen forests 


This species grows in 


and in the heathlike vegetation. on lava flows 
("groundwater laterite"), at elevations of 200—812 m. 


Phenology. This species has been collected with 
flowers in January, May, November, and December, 


and with fruits February through June. 


Discussion. | Gaertnera psychotrioides is frequently 
encountered within the forests of Mauritius. Unlike 
the similar species G. hirtiflora, G. psychotrioides 


lacks setae at the top of the stipule tube and has 


externally glabrous corollas. Gaertnera psychotrioides 

might also be confused with G. edentata, but can be 

distinguished by its lax corymbiform inflorescences 
—15 em wide and its smaller flowers. 

Verdcourt (1989) separated Gaertnera quadriseta, 
noting that it is imperfectly known and distinguishing 
it by its corolla tube ca. 3 mm in diameter, versus 
what he characterized only as a narrow corolla tube in 
G. psychotrioides, and also in having relatively larger 
though overlapping conditions of several characters 
such as leaves and bracts. According to the species 
circumscription used here, these are not distinct taxa 
and are not separated here. Verdcourt (1989) also 
separated G. truncata, noting that this species was 
imperfectly known and separating it by its “truncate to 
very lightly lobed” calyx, in contrast to subtruncate or 
with lobes 0.5-0.8 mm long in G. psychotrioides of 
Verdcourt’s description. The characters of G. truncata 
thus fall within the variation found in G. psycho- 
trioides, and this species is not separated here. 


(ce a specimens examined. MAURITIUS. Bel 
D'Argeni MAU 22454 ds Lorence 2209 (MO), 

4449 (MO), 4490 (MO); Mt. Le Pouce, Bernardi 14714 (BM, 
G, K, MO, P), G. Gardner s.n. 
Tirvengadum 
Park, Malcomber 2938 (MO), 2939 (MO), Bernardi 14778 (G, 
P); Macabé, Bernardi 14714 (K), 14794B (K), Vaughan MAU 
13761 (MAU), Lorence M 281 (MO), 1495 (MAU, MO), 2125 
(MO), Malcomber 2946 (MO), 2947 (MO), Tirvengadum 396/ 
46 (MAU); Plaine Champagne, Lorence 1541 (MAU, MO), 
Coode 4765 (K, MAU, P), Malcomber 2950 (MO), 2931 (MO). 


57. Gaertnera ramosa Ridl., J. Linn. Soc., Bot. 38: 

908. TYPE: Malaysia. Gunong Tahan, 3 

July 1905, H. G. Robinson & L. Wray, Jr. 5488 
(holotype, SING!; isotypes, BM!, K!, SING!). 


Gaerinera aes Ridl., J. Straits Branch Roy. Asiat. Soc. 
19: 1918. TYPE: Malaysia. Selangor: Gunong 
Mi To 6 Feb. 1913, H. G. Robinson s.n. (holotype, 
K!; isotype, SING!). 

Gaerinera us Ridl., J. ard Branch Roy. Asiat. Soc. 
79: 99. . TYPE: Malaysia. H. G. Robinson s.n. 
(holotype, a A SINGH. 

Gaerinera caudai 2s Fed. Malay States Mus. 6: 51. 
1915. 


unong Kerbau, 14 Mar. 1913, H. G. Robinson s.n. 
(holotype, K!). 


Trees or shrubs, 3-5(-8) m tall; branches terete to 
flattened, glabrous, 1-4 mm diam.; internodes 0.6— 
7.5 em, smooth. Leaf blades 3-14.5 X 0.8-5 cm, 
elliptic-lanceolate to elliptic, oblanceolate, or linear- 
lanceolate, apex cuspidate to acuminate, base acute to 
cuneate, drying chartaceous, glabrous; secondary 
veins distinct abaxially, 3 to 9 pairs; domatia absent 
and sometimes foveolate; 


or present, hirtellous 


petioles 3-25 mm. Stipules tubular, glabrous, drying 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


membranous, marcescent or caducous, tube 7-15 mm, 
with ribs 4, narrowly winged, arising below petiole 
and sometimes extending to lobes, Ps usually entire 
or with 1 incision, marcescent, lobes 4, up to 2.5 mm, 
deltate to filiform. Inflorescences cymose, several- 
flowered, terminal on axillary branches, glabrous, 
sessile or dt to 4.5 em; branched Py 
corymbiform X 1-6 em, branched t o 3 
orders, lax etre in apical part); bracts ae or 
deltate, 1-5 mm; bracteoles reduced; pedicels absent 
or to 5 mm s 5-merous, unisexual. Pistillate 
flowers: calyx cup- eee .5-3.5 mm wide, gla- 
brous, truncate or lobes to 0.8 mm, os corolla 
white, clavate in bud, when open salverform, outside 
glabrous or puberulent, tube 5-10 mm, 1.54 mm 
diam., inside villous in upper third, lobes 2-4 mm, 
ligulate or ovate-oblong, acute; staminodia included, 
filaments inserted in upper third of corolla tube, ca. 
0.5 mm; style 6-10 mm, glabrous, stigmas 2-2.5 mm 
Staminate flowers: similar to pistillate except anthers 
shortly exserted, filaments 1-8 mm; pistillodes with 
style portion 4—6 mm, stigmatic portions 0.8-1.5 mm. 
ia violet-black or blue, globose or didymous, 7-8 
—11 mm; pyrenes spherical or hemispherical, 
rugose, finely fissured, endosperm entire. 
Distribution and habitat. 
southeastern Asia, where it 
Malaysia at 


particularly frequently tun in e a 


This species grows in 
it is known e Peninsular 
elevations of 450 


and Genting highlands. Here, it is found in humid 
forests at elevations of 1300-16 

Phenology. This species has been collected with 
May through 
December, and with fruits April through October. 


flowers January through April and 


Discussion. This species belongs to the Gaertnera 
vaginans complex; see also the discussion of that group 
for related species and their distinctions. Van Beuse- 
ko 


of G. 


vaginans subsp. junghuhniana. Although G. ramosa is 


m (1967) considered G. ramosa a synonym 


phenotypically similar to some of the fewer-flowered 
forms of G. junghuhniana, it is recognized here as a 
separate species based on its larger, often pedicellate 
flowers. Gaertnera ramosa might be confused with G. 
belumutensis, but differs in the lax inflorescence and in 
not drying with an orange cast. 


cu depu specimens examined. MALAYSIA. Pa- 
hang: Cam Highlands, Bukit Jasar, Wong FRI 32350 


NG); Fraser's Hill, Purseglove 4227 
Tree zm Nickille 4816 (K). 


a 


A, K, L); Pine 


58. a raphaelii Malcomber, sp. nov. 
adagascar. Fianarantsoa: Ranomafana 


ae Park, 21°13’S, 47°27'E, 900 m, Nov. 


1991, S. T. Malcomber et al. 1018 (holotype, MO- 
4570509!; BR!, G!, TANI, WAG). 
Figure 


isotypes, 


Haec species Gaerinerae obovatae Baker similis, sed ab ea 
inflorescentiae bracteis albis prominentibus atque in quoqu 
uno duobusve in calycophylla alba 
petaloidea expansis distinguitur. 

Trees or shrubs, 2-10 m tall; branches flattened or 
terete or quadrangular, glabrous, 2—4 mm diam.; 
internodes 1.5—4 em, smooth. Leaf blades 3.2-11 
(214) X 14(-5.5) em, elliptic to oblanceolate or 
elliptic-oblong, apex shortly cuspidate or acuminate, 

ase cuneate to obtuse, drying chartaceous, glabrous; 
secondary veins distinct, thinly prominulous, 5 to 8 
pairs; domatia usually present; petioles 3-11(-14) 
mm. Stipules calyptrate, glabrous, drying membra- 
nous, caducous or quickly fragmenting, tube 11-30 
(-55) mm, with ribs 4, rounded to 


arising beneath petiole and sometimes extending to 


narrowly winged, 


apex, apex with 1 incision, marcescent, lobes 2, 0.5— 
0.7 mm, deltate or linear. Inflorescences cymose, 
many-flowered, terminal on axillary branches, puber- 
ulent to id e peduncle 0.5—4 cm; branched 

mbiform, 1.4-7.5 X 1.4-7.5(-9) em, 


branched to 3 to 4. n. lax or sometimes congested 


portion corym 


near apex; bracts linear to ligulate or trifid, 5-15 mm, 
1-2.5 mm, 


white; 
white; pedicels absent or to 2 mm. Flowers 5-merous, 


bracteoles triangular to ovate, 
heterodistylous. Long-styled flowers: calyx cup-shaped 
or campanulate, 1.7-2.5 mm wide, outside glabrous, 
with hair-ring inside, lobes unequal with 1 or 2 larger, 
1-5 mm, linear to triangular or elliptic-oblong, white; 
corolla white, clavate with thickened abaxial append- 
ages at apex in bud, when open salverform, outside 
glabrous, tube 8-10 mm, 2-3 m 
villous in upper third, lobes 2-3 mm, triangular to 


m diam., inside 
ligulate, acute; anthers included, filaments inserted in 
upper third of corolla tube, 0.3-0.5 mm; style 6— 

mm, glabrous, stigmas 1—4 mm. Short-styled flowers: 
similar to long styled except calyx 1.5-3 mm wide; 
corolla tube 6.5-10 mm; anthers shortly exserted, 
1.5-2 mm; style 4-5 mm, stigmas 2-2.5 mm. Drupes 
violet-black, globose or didymous, 4-8 X 5-8 mm; 
pyrenes spherical or hemispherical, rugose, finely 
fave endosperm entire. 


Distribution and habitat. This species grows in 
Madagascar, where it is known from the provinces of 
Fianarantsoa and Toliara, in the Ranomafana, Andrin- 
gitra, and Andohahela Nature Reserves. Here, it is 
found in humid forests at elevations of 150-1150 m. 


Phenology. This species has been collected with 


flowers e A through December and with fruits 


January through April and in December. 


650 Annals of the 
Missouri Botanical Garden 
Discussion. Gaertnera raphaelii is similar to G. — domatia present, foveolate with hirtellous pubescence 


obovata, G. humblotii, and G. phyllostachya, but can 
lyx lobes 


with only one or two 


be distinguished by its elliptic unequal ca 
and 


enlarged calyx lobes per flower. Gaertnera humblotii 


inflorescence bracts, 
is similar to G. raphaelii, in particular in the 
distinctive bracts and calyx lobes, but differs in its 
longer calyx lobes, 5—7 mm long, larger leaves, and 
tubular stipules. Gaertnera raphaelii may hybridize 
with G. phanerophlebia; see additional comments 
under that species. 

Gaertnera raphaelii is named in honor of Malagasy 
botanist Raphael Rakoto, who made some of the early 
collections of this species from Ranomafana National 


Park before his untimely death in 1995. 


Paratypes. 
Humbert 6056 (M 


MADAGASCAR. nM 
P); Sakaleona 


Valley. 


(MO, P); Bernardi 11535 (K); Nil 106 (P, TAN). 
Fianarants mbalavao, Mahazory, Vohimary, Rakotovao 
537 (TEF); Andranomanaraka, 


Tolongoina, 2 Collec- 
tor SF 5231 (TAN); Andringitra Nature Reserv s 8l 
of Ambalavao, pus 756 (MO, 


, TAN); Valley of Ianlara, Ivohibe- 
Farafangana, Unknown Collector SF 1458 (P, TAN, TEF), SF 


1472 (P, TAN, TEF). Toliara: Andohahela Reserve, 
Leeuwenberg 14017 (WAG), Malcomber 1172 (MO, 
A chaiz & Miller 1228 (MO, P, TAN); NW of 


11018 cu 


59. Gaertnera rosea Thwaites ex Benth., J. Proc. 


Linn. Soc., Bot. 1: 111. 1857. Siena rosea 
(Thwaites ex T Kuntze, Revis. Gen. P 
425. 1891. TYPE: Sri Lanka. G. H. K. Thwaites 


CP 2673 dst designated by van Beuse- 
kom, 1967 Mt s K!; isotypes, BM!, BO 
not seen, BR !, G!, GH!, L!, P not seen, 
PDA!, W!, wu a Sri NN G. W. 
Walker s.n. (K not seen).] 


Treelets or trees, 2—5 m tall; branches terete, 
glabrous, 0.5-2(-3) mm diam. internodes — 
4.5 em, with 2 longitudinal ribs. T blades 1- " x 
0.5-3 em, elliptic-lanceolate, elliptic-oblong, or el- 
liptic, apex cuspidate or acuminate, base cuneate, 
drying chartaceous, glabrous; secondary veins evident 
and fla to prominulous abaxially, 3 to 6 pairs; 


inside; petioles 0.5-5 mm. Stipules tubular, glabrous 
to densely pilosulose, drying membranous, persistent, 
tube 1-5 mm, with 
above or beneath petiole and extending onto tube, 


ribs 4, narrowly ridged, arising 


rpetiolar side, these 
extending to lobes, apex entire, marcescent, lobes 2 
and deeply bifid or 4, 1.5—5.5 mm, filiform. Inflores- 


cences 3-flowered, congested-fasciculate to subcapi- 


fusing into 1 rib on each inte 


tate, terminal on axillary branches, glabrous, 0.2-2. 
.4-2.6 em, sessile or peduncle to 4 cm; bracts 
deltate or stipuliform, 1.2-2.5 mm; bracteoles re- 


duced; pedicels absen o 0.6 mm. Flowers 4- 


lobes to 0.9 mm, triangular; corolla pink on tube 
and white 
salverform, outside glabrous, tube 12-20 mm, 1- 


on lobes, clavate in bud, when open 
inside villous in upper third, lobes 7— 
acute; anthers included, fila- 

r third of corolla tube, 0.2— 
0.5 mm; style 13-17 m mm, glabrous, stigmas 2-5 mm. 
Short-styled flowers: similar to long styled except 
16-23 mm m diam., lobes 5- 
1.84 mm; style 6.5-9 mm, 
stigmas 2.5—4.5 mm. Drupes violet-black, globose or 
subglobose, 7-10 X 7-10 mm; pyrenes 
hemispherical, rugose, finely fissured, 


3 mm diam., 
10.5 mm, ligulate, 
ments inserted in up 


corolla tube 
11.5 mm; filaments 


spherical or 
endosperm 
entire. 


Distribution Ev habitat. 


ka, where n be foun 
elevations of ee m. 


ae species grows in Sri 
in humid forests at 


Phenology. This species has been collected with 
flowers May through August and with fruits June 
through October. 


Discussion. | Gaertnera rosea can be recognized by 
its ridged internodes, inflorescences reduced to a 3- 
flowered, congested cyme, and 4-merous flowers with 
pink corolla tubes and white corolla lobes. Yellow 
flowers are reported on several specimens (e.g., 
Kostermans 23597); from field observations, it appears 
that the flowers open with a pink corolla tube and 
white lobes and then turn yellow as they age as in 


some other species of Rubiaceae. 


Representative specimens examined. SRI LANKA. Galle, 
Kanneliya, Waas 1335 (K, MO, PDA). Ratnapura Distr.: 
Mannikkawatta, Waas 1768 (E, GH, K, L, MO, PDA). 


60. Gaertnera rotundifolia Bojer, Hortus Maurit. 

1837. Sykesia rotundifolia (Bojer) Kuntze, 

Revis. Gen. Pl. 2: 425. 1891. TYPE: Mauritius. 

Forests of Grand Port and Savanne and around 

rand Bassin, W. Bojer s.n. (holotype, P not 
located; isotype, G-DC?). 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


Trees or shrubs, 1-6 m tall; branches flattened, 
terete, or quadrangular, glabrous, 3-6 mm diam.; 
internodes 1.3—4 cm, smooth. Leaf blades L 5-12 x 

—5.5 em, broadly elliptie to obovate or ovate, apex 


obtuse and shortly acuminate to rounded, base 
cuneate to subcordate, drying thickly coriaceous, 
glabrous, margins often thinly revolute; secondary 
veins flat and visible to DR EM Deer 6 to 8 
pairs; domatia absent; petioles m. Stipules 
calyptrate, glabrous or ae ae membra- 
tube 15-20 mm, with ribs 4, 


narrowly winged, arising beneath petiole but not 


nous, marcescent, 
extending to top, sometimes uniting in basal part of 
interpetiolar portion, apex wit incision, marces- 
cent, lobes deltate or linear. Inflorescences 


congested-cymose to subcapitate, many-flowered, 
terminal on axillary branches, puberulent or glabrous; 
peduncle 0.5-2 em; branched portion subglobose or 
corymbiform, 1.3—4 X 1.5-4 cm, branched to 2 to 3 
orders, congested; bracts deltate to linear or trifid, 8— 
10 mm; bracteoles triangular to lanceolate, 2-4 mm; 
pedicels absent or to 2 mm. Flowers 5-merous, 
heterodistylous. Long-styled flowers: calyx cup-shaped 
or campanulate, 2.5-4 mm wide, outside glabrous, 
with hair-ring inside, lobes 1-2 mm, triangular to 
ovate; corolla white, clavate in bud, when open 
salverform, outside glabrous, tube 15-25 mm, 2- 
lobes 5-7 mm 


narrowly triangular or ligulate, acute; anthers includ- 


3.5 mm diam. glabrous inside, 
ed, filaments inserted in upper third of corolla tube, 
ca. 0.5 mm; style 15-25 mm, glabrous, stigmas 2— 
3 mm. Short-styled flowers: 
except calyx + campanulate, 2-3.5 mm wide, outside 


similar to long styled 


puberulent or glabrous, lobes 1-1.5 mm, triangular to 
ovate or oblong; corolla clavate in bud or sometimes 
with appendages at apex, tube 2-4 mm diam., lobes 
5-6 mm, narrowly triangular or ligulate to ovate- 
oblong; anthers shortly exserted, filaments 2-3 mm; 
style 9-12 mm, stigmas 3-5 mm. Drupes violet-black, 
ellipsoid to obovoid, 15-17 X 6-8 mm; pyrenes 
finely fissured, endosperm 


plano-convex, rugose, 


entire. 
Distribution and habitat. 
Mauritius, 


This species grows in 
eathlike 
vegetation on lava flows (“groundwater laterite”) at 
elevations of 500—700 m. 


where it is found in the 


Phenology. This species has been collected with 
flowers in February and December and with fruits in 
January, February, and May through December. 


Discussion. Gaertnera rotundifolia and G. cunei- 
folia are generally similar and were considered the 
aker (1877); see the discussion 


under G. cuneifolia. Gaertnera rotundifolia was treated 


same species 


as an unnamed variety "B [beta]? of Chassalia 
clusiifolia DC. by de Candolle (Prodr. 4: 532. 
based on the specimen 
s.n. (G-DC!). This variety has been cited b 
authors (e.g., Verdcourt, 1989), but because it was 
unnamed it is not validly published. 


Representative specimens examined. MAURITIUS. Cure- 
pipe, n MAU 1655 (MAU); Perrier Reserve, Vaughan 
MAU 12271 (MAU) Black Ri National Park, 
Petrin, a 1548 (MAU, MO), Malcomber 2949 (MO), 
2955 (MO), Richardson 4078 (K), Tirvengadum 947 (K); 
Plaine Champagne, Puff 800825—1/5 (K). 


61. Gaertnera schatzii Malcomber, sp. nov. TYPE 
Madagascar. Toamasina: Masoala National Park, 
Antalvia, 15%47'S, 50%02'E, Nov. 1989, 

Schatz 2808 (holotype, MO-3769600!; MN 
K!, P!, TAN!, WAG!). Figure 2. 


Haec to Gaerinerae griseae Hook. f. ex C. B. Clarke 
similis, sed ea planta in sicco aurantiaca, stipulis 
brunneis MR fissis 22-68 mm longis atque tubo corollino 
extus dense tomentoso distinguitur. 

rubs or trees, 3-9 m tall; plants drying with 
orange cast; branches terete, when young densely 
tomentulose to tomentose with indumentum drying 
reddened to orange, sometimes with indumentum 


ryi own or Sud oming a ent, 3-6 
diam.; internod 4-8 cm, smooth. Leaf blades ve 
29 5-10 cm, ee to m d. apex 


cuspidate to acuminate, base cuneate to rounded, 
rying chartaceous, adaxially glabrous, abaxially 
densely pilosulose to tomentose with indumentum 
drying reddened to yellow-orange; secondary veins 
prominulous abaxially, 9 to 12 pairs; domatia absent; 
petioles 5-20 mm. Stipules tubular, densely tomentu- 
lose to tomentose, drying chartaceous, generally 
tube 22- 


68 mm, with ribs 4, narrowly to broadly winged, 


persisting on distalmost 2 to 5 nodes, 


arising below petiole and usually extending to lobes, 
h l incision, 


mm, deltate to lin 


apex wit marcescent, lobes 

PES Inflorescences cymose, 
many-flowered, NN on n branches, densely 
tomentose; peduncle (absent) 4.9—7.8 cm; branched 
portion corymbiform, 5-11 X 6-13 cm, branched to 4 
to 5 orders, lax to congested; bracts deltate or linear, 
3-14 mm, 


sometimes glabrous above; bracteoles 


merous, heterodistylous. Long-styled flowers: calyx 
cup-shaped, 3.5-5. wide, outside densely to- 
mentose, glabrous inside, truncate or lobes to 0.5 m 


triangular; corolla white, clavate in bud, when open 
infundibuliform or salverform, outside densely stri- 
gose to tomentose, tube 11-13 mm, 2-5 mm diam., 
inside villous in upper third, lobes 4—5 mm, triangular 
or ligulate, acute; anthers included, filaments inserted 


Annals of the 
Missouri Botanical Garden 


in upper third of corolla tube, ca. 0.5 mm; style 10— 
12 mm, glabrous, stigmas 2-2.5 mm. Short-styled 
flowers: calyx 
—5.5 mm diam.; anthers shortly exserted, 
filaments 2-3.5 mm; style 3.5-5 mm, stigmas 2- 


12 mm, 


4 mm. Drupes unknown. 


Distribution and habitat. 
Madagascar, where i 


This species grows in 
it is known from the province of 
Toamasina in the Masoala National Park. Here, it is 


found in humid forests at elevations of 0-380 m 


Phenology. This species has been collected with 
flowers in October and November, but has not been 
collected with fruits. 


Discussion. A localized species, this is only 


known from the Antalavia River Valley on the west 
coast of the Masoala Peninsula. Gaertnera schatzii can 
be recognized by its dense pubescence that becomes 
orange on dried specimens, once-cleft stipules that 
dry brown, and densely tomentose corolla tubes. This 
species is named after George Schatz, who collected 
the type specimen and has made significant contri- 
butions to our knowledge of the flora of Madagascar. 


Paratypes. MADAGASCAR. Toamasina: Masoala Na- 
tional Park, Antalavia, Rahajasoa 923 (K, MO, P, TAN), 
1089 (MO, TAN), Malcomber 2825 (MO, P, TEF), 2829 (MO, 
P, TEF), [no initial] Moise 9 (MO) 

62. Gaertnera schizocalyx Bremek., Bull. Misc. 
Inform. Kew 1940: 193. 1940. TYPE: Malaysia. 
Sarawak: Matang, Nov. 1871, P. B. Beccari 1799 
(holotype, K!; isotype, K?). 

Shrubs or small trees, 2-3 m tall; branches terete, 
when young densely hispid to villous with indumen- 
tum drying yellow to white, becoming glabrescent, 
2.5-6 mm diam.; internodes 3-11 cm, smooth. Leaf 
blades 6-14 X 2 


obovate, apex cuspidate or acuminate, 


—5 em, elliptic to oblanceolate or 
obtuse to 
cuneate, drying chartaceous, adaxially glabrous or 
hirtellous on principal veins, abaxially densely hispid 
to villous or villosulous with indumentum drying 
yellow or gray-white; secondary veins prominulous 
abaxially, 3 to 7 pairs; domatia E petioles 3— 
7 mm. Stipules tubular, densely villous to hispid, 
rying chartaceous to membranous, persistent on 


m, with ribs 4, broadly winged, 
arising below petiole and extending to lobes, apex 
entire or with 2 incisions, marcescent, lobes 4, 2— 
12 mm, linear to deltate. /nflorescences several- to 
many-flowered, congested to subcapitate, terminal on 
axillary branches, densely ee to villosulous or 
pilosulose, sessile or pedun m; branched 
portion subglobose, 1.2-3 X les em, branched to 1 


to 2 orders; bracts d or trifid, 1-10 mm; 
bracteoles reduced; s absent or to 

Flowers 5-merous, i union Pistillate precem un- 
known. Staminate flowers: calyx campanulate, 1.5— 
2 mm wide, outside pilosulose to hispidulous, gla- 
brous inside, lobes 0.5—4 mm, narrowly triangular; 
corolla white, clavate in bud, when open salverform, 
outside glabrous, tube 4-5 mm, 1.5-2 mm diam., 
inside villous 


e or ovate-oblong, acute 


in upper third, lobes 1.5-2.5 mm 
ligulat ; anthers shortly 
exserted, filaments inserted in upper third of corolla 
tube, ca. 0.2 mm; pistillode reduced or absent. Drupes 
violet-black, globose or didymous, 5-8 X 5-10 mm; 
pyrenes spherical or hemispherical, + 


pe 


rugose, finely 
fissured, endosperm entire. 

Distribution and habitat. 
southeastern Asia, where it is known from Peninsular 
ici in the Sarawak (Malaysia) 


orne alaysi 
ctor. Here, it can be found in humid forests, usually 


This species grows in 
and 


at "e edges of swamps and riverbanks, at an elevation 
of ca. 10 m 


Phenology. This species has been collected with 
flowers May through July and with fruits in October 
and November 


Discussion. Gaertnera schizocalyx is not well 
known. In the Panti in Joh 
Malaysia, it appears to be restricted to swamp oe 
The species can be recognized by its subglobose, 
subcapitate, 


orest Reserve, 


sessile or very shortly pedunculate 
inflorescences, dense hispid to villous indumentum, 
and well-developed, narrowly triangular calyx lobes. It 
is similar to C. capitulata, which differs in its smaller 
4-merous flowers. 


Additional specimens examined. MALAYSIA. Johor: 
Panti Forest Reserve, 1.7 km W of Rte. 3, Malcomber 3026 
(MO); Sungai Kayu, Kiah SFN 32015 (A, L, SING); Sungai 
Kayu, Mawai-Jemaluang rd., Corner SFN 52507 (L), 7 July 
1935, E. Corner s.n. (SING). 


63. Gaertnera spicata K. Schum., Bot. Jahrb. Syst. 
33: 372. 13 Mar. 1903. TYPE: Gabon. Estuaire: 
Munda [Mondah], Sibange Farm, 21 Aug. 1879, 
H. Soyaux 24 ([holotype, Bt]; lectotype, desig- 
nated here, K!; isotype, P!). 


Gaerinera rhodantha Baker, in Oliver, Fl. Trop. Afr. 4 (1, pt. 
3): 543. 30 Mar. 1903. TYPE: Gabon. Estuaire: Munda 
[Mondah], Sibange Farm, 21 Aug. 1879, H. Soyaux 24 
(holotype, K*; isotype, P5. 

es or shrubs, 1.5-8 m tall; branches terete or 
mu when young puberulent with indumen- 
tum drying reddish to brown, when older puberulent to 
glabrescent, 3-6 mm diam.; internodes 1.5-10.5 cm, 

smooth. Leaf blades 18-29 x 5-9 


obovate, apex euspidate to acute, base attenuate or 


em, oblanceolate to 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


cuneate, drying chartaceous, adaxially glabrous, 
abaxially glabrous or puberulent on principal veins; 
secondary veins indistinct or prominulous abaxially, 8 
to 12 pairs 15-32 mm. 
Stipules hibulas densely pilosulose or strigillose to 


; domatia absent; petioles 
glabrescent, drying chartaceous, persistent on distal- 
most nodes, tube 10-17 mm, with ribs 4, broadly 
winged, arising below petiole and extending to lobes, 
apex entire or with 1 or 2 incisions, marcescent, lobes 
4, 5-15 mm, deltate to linear; setae numerous, 1.5— 
6 mm. Inflorescences many-flowered, terminal on 
principal and/or axillary branches, densely strigillose 
to puberulent; peduncle 2-3.5 cm; branched portion 
1.7-5.5 X 0.9- 
m, unbranched or branched to 1 to 2 orders, 


spiciform to narrowly pyramidal, 
1.5 


densely congested; bracts deltate or linear to 
lanceolate, 1-10 mm, sometimes glabrous; bracteoles 
reduced. Flowers sessile or subsessile, 5-merous, 
heterodistylous. led flowers: calyx cup- 
shaped, abrous or sometimes 
puberulent outside, truncate or lobes to 0.5 mm, 
triangular; corolla red to orange-red on tube and red or 
internally white on lobes, clavate in bud, when open 
salverform, outside puberulent or glabrous, tube 9— 
1l mm, m diam., inside villous in upper 
third, lobes rie mm, elare to lanceolate, acute; 
anthers included, filaments inserted in upper third of 
corolla tube, ca. 0.3 mm; style 10-14 mm, glabrous, 
stigmas 0.8-1.5 mm. Short-styled flowers: similar to 
long styled except corolla tube 9-10 mm, 1.5-3 mm 
diam., lobes 4-6 mm; anthers fully exserted, filaments 
inserted in upper third of tube, 1.5-2 mm; style 5.5— 
6 mm, stigmas 1.5-2 mm. Drupes reddish brown 
(Breteler & de Wilde 389) or perhaps violet-black, 
d M ed to exp po and + 
caniculat apex x mm; pyrenes 


a rugose, is Md 


Distribution and. habitat. 
Central Africa, where it is known from Gabon. Here, it 


This species grows in 


is found in coastal forests on white sands, at elevations 
at sea level or very near it. 


Phenology. This species has been collected with 
flowers August through October and with fruits in 
January, February, November, and December. 


Discussion. | Gaertnera spicata is a morphologically 
isolated and QUU restricted species that can 
be recognized by its narrow, usually cylindrical 
inflorescences and red to orange oven. This species 
is apparently endemic to coastal white sand forests 
near Cap Esterias, Gabon. Petit (1959b) provided the 
publication date for G. rhodantha Baker, described 
almost simultaneously with G. spicata from a duplicate 


set of specimens. 


Two syntypes were cited in the protologue of 
Gaertnera spicata, H. Soyaux 24 (B S K!, P?) 
d H. Soyaux 178 (B destroyed, P!). T 


"edi here as lectotype because it is a more 


ormer is 


exemplary specimen and has more duplicates. The 
bou deposited in the ipw museum € 
ogica be selected as the ype wa 
rd RR with the general fiir ee collection 
there; 
lectotype because it is a more complete specimen and 


the specimen at K is chosen here as the 


the selection a it makes the synonymy of these two 
names indisputable. 

The names po spicata and G. rhodantha 
were described independently and simultaneously, 
and apparently neither author knew about the other 
name so the publication of these two names based on 
the same plants is a coincidence. Each author based 
his name on a different duplicate from the set of 
specimens of Soyaux 24, thus each of these names has 
a different holotype or lectotype. Presumably each 
author intended to include all the duplicate specimens 
of that set in his species, as is the custom today for 
designating isotypes and as done here. 


Representative specimens examined. GABON. Estuaire: 
Cap Santa Clara & Cap Esterias, Reitsma 1336 (WAG); 
Modal Forest, 25 km along the rd. from Libreville to Cape 
Esterias, ca. 1 km from seashore, Breteler & J. J. F. E. de 


Wilde 389 (K, MO, P, WAG); Mt. Bouet, Leroy 8 (P). 


64. Gaertnera sralensis (Pierre ex Pit.) Kerr, Kew 
ull. 1940: 180. 1940. Basionym: Psychotria 
sralensis Pierre ex Pit, in Lecompte, Fl. 


Indochine 3(3): 344. 1924. Uragoga sralensis 
ierre ex Pit, in Lecomte, Fl. Indochine 3 
344. 1924, nom. nud., pro syn. TYPE: Cam boda. 
“In montibus Sral, prov. d Samrong-tong," Apr. 
1870, J. B. L. Pierre 1253 (lectotype, designated 
by van Beusekom, 1967 [1968]: 386, L!; 
isotypes, K!, 


Shrubs, to 3 m tall; branches terete, glabrous, 1.5— 
Leaf blades 


4.5 em, elliptic-lanceolate to elliptic or 


4 mm EN ; internodes 1-7 em, smooth. 
4-13 X 
T apex acuminate to cuspidate, base = 
attenuate to cuneate, drying chartaceous, glabrous to 
variously pubescent, drying green or grayish green or 
a pie gray, chestnut, reddish brown, or 


with ad ast; secondary veins distinct 


abaxially, 4 to 7 uo domatia absent; petioles 4— 
15 mm. Stipules tubular, glabrous, drying chartaceous, 
caducous or deciduous through fragmentation, tube 
ca. 16 mm, with ribs 4, narrowly winged, arising 
below petiole and extending to lobes, apex with 1 or 2 
incisions, marcescent, lobes 4, ca. 1.6 mm, filiform. 
Inflorescences congested-cymose, several- to some- 
times many-flowered, terminal on principal and/or 


Annals of the 
Missouri Botanical Garden 


axillary branches, glabrous, sessile or peduncle to 
0.3 em; branched portion subglobose, 0.5-1.5 X 0.5— 
2 cm, branched to 1 to 2 orders; bracts deltate to 
linear or trifid, 1-2 mm; bracteoles reduced; pedicels 
absent or to 1.5 mm. Flowers 5-merous, unisexual. 
Pistillate flowers: unknown. Staminate flowers: calyx 
cup-shaped, 3-3.5 mm wide, outside glabrous, with 
hair-ring inside, truncate or lobes to 0.4 mm, triangu- 
lar; corolla in white, outside glabrous, up to 3.5— 

m, 1.5-2 mm diam., inside villous in upper third, 
oben 2.5-3 mm, triangular to ligulate, acute; filaments 
inserted in upper third of corolla tube, ca. 0.5 mm 
stamens not seen; pistillode reduced. Drupes violet- 
black, globose, 6-8 i 


or spherical, + 


mm; pyrenes hemispherical 
rugose, finely fissured, endosperm 
entire. 


Distribution and habitat. This species grows in 


southeastern Asia, where it has been found in 
Cambodia, Peninsular Malaysia, Thailand, and Viet- 
nam. Here, it grows in humid forests at elevations of 


600-1800 m. 


Phenology. This species has been collected with 
flowers January through July and with fruits January 
through April and October through December. 


Discussion. Van Beusekom (1967) considered 
Gaertnera sralensis a synonym of his widely circum- 
scribed G. vaginans subsp. junghuhniana. With the 
narrower circumscription of G. junghuhniana adopted 
here, G. sralensis is recognized as a separate species 
distinguished by its subglobose inflorescences, gla- 

ous vegetative structures, and calyx with a hair-ring 
inside. This species belongs to the G. vaginans 
complex; see also the discussion of that group for 
related species and their distinctions. 

In the protologue, several syntype specimens were 
cited but without number. These can now be detailed: 
J. B. L. Pierre 1253 (the lectotype) (L), J. B. L. Pierre 
3247 (Pb), E. Poilane 237 (BM!, P!), 274 (P), [no 
initial; Vegter, 1988] Talmy s.n. (P), and C. Thorel 
1165 (P). Two names were published simultaneously 
for this species by Pitard in the protologue of 
included 


which was listed there as a 


Psychotria sralensis, P. sralensis which 
Uragoga sralensis, 
synonym and thus not validly published. The name 
P. sralensis was specifically cited as the pas for 
a sralensis, thus Kerr regarded i the 
ished name. Later, without pur 
van Beusekom (1968: 386) subsequently cited U. 
sralensis as the validly published name and P. 
sralensis as an invalid synonymous name. However, 
van Beusekom's usage is contrary to the presentation 
of these names in the protologue, where P. sralensis 
was clearly intended by Pitard to be the accepted 


flowers: calyx cup-shaped, 


name for this species: he included this species in the 
treatment of the genus Psychotria; the name P. 
sralensis is placed at the beginning of this treatment 
in boldface type similar to all the other accepted 
names there, while the name U. sralensis is placed 
second and in italic type similar to other synonyms 
there; and this species is called P. sralensis in the key 
there. 


Representative specimens examined. CAMBODIA. Mont 
de l'Eléphant, Poilane 237 (BM, P). MALAYSIA. Pahang: 
alay States 


Kuap, Kerr 17764 (BM, K, L. 17798 (BM, K. Put 2864 (BM, 
K, L), 2939 (BM, K, L). South: Khao Luang, Van Beusekom 
839 (K, L). VIETNAM. Prov. Baria [Ba Ria]: Mt. Dinh, 
Pierre 3247 (Py; Ti-tinh, Thorel 1165 (P). 


65. Gaertnera ternifolia Thwaites, Enum. Pl. Zeyl. 
202. 18 ykesia ternifolia (Thwaites) Kuntze, 
Revis. Gen. Pl. 2: 426. 1891. TYPE: Sri Lanka. 
Near Adam's Pw G. H. K. Thwaites CP 440 

(lectotype, cda by van Beusekom, 1967 

68]: 38 E a BM!, BR!, CGE!, G!, 

GH!, K!, P, W!, WUD. [SYNTYPE: G. H. K. 

Thwaites CP 457 (K?).] Figure 146-M. 


Gaerinera walkeri var. angustifolia Benth., J. Proc. Linn. 
x | 1: 111. 1857. TYPE: Sri Tanka. s. loc., G. H. 

K. tes CP 440 (lectotype, -o ated by van 

Beu E , 1967 [1968]: 381, Pl; isotypes, BM!, BR!, 
CGE!, Gl, GH!, KI, PL, W!, WU). 

Trees or shrubs, 1-5 m tall; branches terete, when 
young puberulent or glabrous with indumentum drying 
ud or Pad -white, becoming glabrescent, 1-5 mm 
dia des -1.2 mm, smooth or with 3 
Nue bs. Tea blades 0.5-2.5 X 
narrowly oblong or linear-lanceolate, apex acuminate 
to acute, base acute to cuneate, drying coriaceous, 
glabrous to pustulose-scaberulose, margin flat to 
revolute; secondary veins not visible; domatia absent; 
petioles 0.2-0.7 mm. Stipules shortly tubular to 
DUDEN nn pd to densely puberulent, 
drying c or membranous, persistent, tube 
0.1-0.3(-1) mm, ins ribs 6, arising below petioles 
then uniting in the middle of each interpetiolar side 
and extending to lobes, apex with 3 incisions, 
marcescent, lobes 3, 0.1—3 mm, deltate. Inflorescences 
reduced to 1 flower or 2 to 3 fasciculate flowers, 
al, pendulous; peduncles (1-)2—7 mm, puberu- 
lent; bracts deltate to linear, 1.5-3 
puberulent; bracteoles linear, ca. 1 


termin: 
mm, glabrous or 
mm, ciliolate. 
Long-styled 
outside gla- 


Flowers | 5(6)-merous, e 

mm 
brous to puberulent, glabrous m lobes 1.2-2 mm, 
triangular to linear; corolla white, clavate in bud, 
when open campanulate, outside glabrous, tube 8— 


11 mm, 1.5-6.5 mm diam., inside villous at ca. 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


middle, lobes 2—5 mm, triangular to ligulate, acute; 
anthers included, filaments inserted at ca. middle o 
corolla tube, 0.4—0.6 mm; style 7-12. 


stigmas 1—1.5 mm. Short-styled flowers: similar to long 


mm, glabrous, 


styled except corolla tube 8-12 mm, 1.4—6 mm diam., 

lobes 3—4 mm; pub ace exserted, filaments 3- 
4 mm; style 3.5-6 
or subglobose or D E 8 X 6-8 


s violet-black, globose 

mm; pyrenes 

spherical or hemispherical, rugose, finely fissured, 
endosperm entire. 


Distribution and habitat. 
Lanka, where 


This species grows in Sri 
it lives in wet premontane and montane 
forests at rn of 1350-2000 m 

Phenology. This species has been collected with 
flowers June through October and with fruits January 
through June and in November and December. 

Discussion. | Gaertnera 
and morphologically isolated species distinguished by 


ternifolia is an attractive 
its relatively small ternate leaves, short persistent 
stipules, few-flowered inflorescences, and campanu- 
late flowers. There has been some confusion of C. 
ternifolia with G. Xgardneri, presumed here to be a 
natural hybrid between G. ternifolia and G. walkeri. 


te and ternate leaves and salverform 
Van Beusekom (1967) cited the Thwaites CP 440 


olotype" of this 
species, without explanation. In general, Thwaites' 


collection deposited at P as the 
first set and his (de facto) holotypes are assumed to be 
eae at K (Stafleu & Cowan, 1986), and van 
sekom does cite an isotype of Gaertnera ternifolia 
at K so the basis for his selection of the P specimen is 
unclear; no one else seems to have previously 
specified any individual collection as the type of that 
species. In his next nomenclatural paragraph pan for 
G. iis var. angustifolia he cites the type 
“Lecto ifolia.” Thus, t ibo 


are P considered lectotypifications here, following 


pe: same as that of G. tern 


d 
(y) 


what at least appears to be van Beusekom's intent. 
The reason for his selection of the same lectotype 
specimen for both names is also not explained; these 
selections, if accepted by others, do definitively 
reduce these names to synonymy. 


d. SRILANKA. Kandy 
Sa E o Pro ]: Adam's Peak, Moray Estate, 
Kostermans 24210 (G, X. ga or (G, K, D, Sohmer 9884 
(GH, K, PDA), Burtt 88 (K, MO, PDA), ol 27020 (G, 
K, L); Ratnapura, Pinnawala, Balakrishna 
Central Province: Nuwara Eliya, Pe 
al Park, Malcomber 2766 (MO, PDA), 2767 (MO, PDA), 
Waas 1693 (E, GH, K, L, MO, PDA). 


= 


66. Gaertnera trachystyla (Hiern) E. M. A. Petit, 
Bull. Jard. Bot. Etat Bruxelles 29: 382. 1959. 


Basionym: Psychotria trachystyla Hiern, Fl. 
Trop. Afr. 3: 213. 1877. TYPE: W tropical 
Africa, Mt. John River, 1°N, 1862, G. Mann 
1791 (holotype, K!). 


Gaerinera ane K. Schum., Bot. Jahrb. Syst. 28: 88. 
, syn. nov. TYPE: Cameroon. Bipinde, 1898, G. 
Zenker 1763. (tone. designated here, MO!; isotypes 
BR!, E!, Gt, K!, L!, P, W!, WAG!, 
Gaerinera salicifolia C. H. Wright ex Baker, HL. Trop. Afr. 4 
543. 1903. TYPE: Gabon. Estuaire: Mfóa, G. Bates 516 
(holotype, K!; isotypes, BM!, BR!, G!, P). 


Trees or shrubs, 1.5-3.5 m tall; branches terete or 
flattened at apex, terete below, glabrous to puberulent, 
1-4 mm diam.; internodes 0.6— th. Leaf 
blades 5.4—23 X 1.8-7.6 cm, elliptic to obovate or 
ovate, apex cuspidate to acute, base acute to cuneate 

obtuse, dryin ly glabrous, 


o chartaceous, adaxial 
abaxially uberulent 


8 
glabrous or pu o pilosulose on 
principal veins; secondary veins prominulous abaxi- 
ally, 4 to 10 pairs; domatia absent or present; petioles 
2-15 mm. Stipules 


puberulent or pilosulose 


tubular, 
, drying 
cous or with persistent Bons l- 2 mm, tube 4—25 mm, 


glabrous to densely 


chartaceous, cadu- 
with ribs 4, narrowly winged, arising above to below 
g to lobes, 
escent, lobes 4. 


petiole and excu extendin, apex 
th 2i 


entire or wit cisions, marc 
0.3—5 mm, deltate to e Inflorescence cymose to 
paniculiform, many-flowered, terminal on principal 


and/or 


axillary ee dense 
pilosulose, 


deflexed to 
8.5 em long; branched dere narrowly pyramidal or 
corymbiform, 2723.5 X 3-19 cm, branched to 1 to 4 
orders, lax; axes usually divergent at ca. 90 degrees; 
bracts deltate or trifid, m 

cent; bracteoles reduced; pedicels 1-9 mm. Flowers 4- 


puberulent to 
endulous; cle 1.2- 


, glabrous or pubes- 


or 5-merous, heterodistylous. Long-styled flowers: calyx 
cup-shape m wide, outside glabrous to 
densely puberulent, with hair-ring inside, truncate or 
lobes to 0.4 mm, triangular; corolla white, clavate in 
bud, when open salverform, outside glabrous, tube 2.4— 
3.5 mm, 1-1.7 mm diam., inside villous in upper third, 
lobes 2.5-3.5 mm, ligulate to elliptic-oblong, acute; 
anthers included, filaments inserted in upper third of 
corolla tube, ca. 0.3 mm; style 5-6 mm, glabrous or 
pubescent near apex, stigmas 0.5-1.1 mm. Short-styled 
flowers: similar to long styled except corolla tube 
3—4.5 mm, 1-2 mm diam., lobes 3-4 mm, ligulate to 
ovate-oblong; anthers fully exserted, filaments 2.5— 
4 mm; style 2-2.5 mm, glabrous, stigmas 1-1.5 mm. 
6-9 X 6- 


9mm; pyrenes spherical or hemispherical, faintly 


Drupes violet-black, globose or subglobose, 


rugose, deeply fissured, endosperm entire. 


Distribution and habitat. 
Central Africa, where it is known from Cameroon, 


This species grows in 


Annals of the 
Missouri Botanical Garden 


Equatorial Guinea, and Gabon. Here, it is found in 
wet forests at elevations of 100-300 m 


Phenology. This species has been collected with 
flowers January through May and September through 
December, and with fruits in January, February, 


November, and December. 


Discussion. Gaertnera trachystyla is variable in 
leaf shape but can be recognized by its deflexed 
to pendulous, many-flowered, pyramidal cymes with the 
axes usually spreading at ca. 90 degrees, and its 
mai de- 
agei as having 4-merous flowers, but 
close inspection of the type and paratype specimens 
studied by him reveals that both 4- and 5-merous 
flowers are borne in the same inflorescence. No 
morpholog p . trachystyla 
from plants described as inklagei and G. salicifolia 
C. H. Wright ex Baker and these are considered 
synonyms here 


ical 3c separate 
G. d. 


The type collection of Gaertnera dinklagei has 
numerous duplicates that are all equally complete 
and exemplary specimens, as with most collections 
by Zenker. The holotype specimen was not explicitly 
designated in the protologue but would have been the 
one at B; however, it also would have been destroyed 
there along with the general Rubiaceae collection so it 
is presumably lost and a lectotype is needed. The M 
specimen is selected as the lectotype here for its good 
condition and for convenience. 


Representative specimens examined. CAMEROON. 2- 
8km S of Kribi, Bos 3166 (K, MO, P, WAG), Bos 
3982 (MO, P, WAG), Bos 4429 (P. WAG); Bipinde, Zenker 
2393 (BM, BR, E, G, K, L, P, W, WAG, WU), 4760 (BM 

P, W): EQUATORIAL GUINEA. 
Miton, Bank & Bion 3106 (MO). GABON. Estuaire: 
Mondah Forest, 25 km Libreville-Cap pU rd., Breteler 
& J J. E. de Wilde 382 (K, MO, WAG); Monts de 
Cristal, Kinguélé rapids, Mbei River, i Hallé 4437 (P), W. 
Hallé & Villiers 4664 (MO, P). Ngounié: 2-15 km SE of 
forestry Camp Waka, A. M. Louis 1344 (WAG), 1309 (K, 
WAG). 


67. Gaertnera vaginans (DC.) Merr., Enum. Born. 

]. 580. 1921. Basionym: Psychotria vaginans 
DC., Prodr. 4: 520. 1830. Ophioxylon arboreum 
J. Koenig ex DC., Prodr. 4: 520. 1830, nom. nud., 
pro syn. Sykesia koenigii Arn., Nova Acta Phys.- 
Med. Acad. Caes. ud zac Nat. Cur. 353. 
. superfl. illeg. Gaertnera koenigii 
(Arn.) Wight, Icon. Pl. Ind. Orient. 4: 6, t. 1318. 
1848, as “konegii,” 
vaginans (DC.) 


nom. superfl. illeg. Sykesia 
Gen. Pl. 2: 42 

in herb. Van 
(holotype, L!; 


Kuntze, Revis. 
" : Sri Lanka. s. loc., 
Royen 108, J. G. Koenig sn. 
isotype, L!). Figure 15G-N. 


MESI p cu Nova Acta Phys.-Med. Acad. Caes. 
Cu 1836. Gaerinera 


> Mus. Bot. 1: 174. 1850. 


by van i Heusckom, 1967 [1968]: 385, E!; isotype, E). 


or shrubs, 1-6 m tall; branches terete to 
flattened, glabrous, 1-10 mm diam.; internodes 1— 
5.5(-10.5) em, smooth. Leaf blades 3-20 X 1.5-9 em, 


narrowly elliptic to obovate, 


Trees 


broadly elliptic, or 


elliptic-oblong, apex acute to acuminate or rounded, 

base cuneate to acute, drying chartaceous, glabris: 
secondary veins prominent abaxially, 4 to 9 pairs; 
domatia absent or present; petioles 0.5-32 mm. 
Stipules tubular, glabrous, drying chartaceous, decid- 
m, with 


ribs none or 4, narrowly winged, arising below petiole 


uous through fragmentation, tube 7 


and sometimes extending to lobes, apex with 1 or 2 
lobes 2 or 4, 0.5-2.5 mm, 


eltate. Inflorescences cymose, 


incisions, marcescent, 


many-flowered, termi- 
nal on principal and/or axillary branches, puberulent 
or glabrous, sessile or peduncle 2-7.5 cm; branched 
(2.5-)10-24 X 


.1-)7-17 cm, branched to 2 to 5 orders, sometimes 


portion corymbiform or pyramidal, 


— 
ER 


congested near apex; bracts deltate or linear, 

10 mm; bracteoles reduced; pedicels absent or to 
5 mm. Flowers 5-merous, heterodistylous. Long-styled 
flowers: calyx cup-shaped, 2.5-4 mm wide, outside 
inside, lobes 

in M 
salverform, outside glabrous to Vae tube 4— 
5.5 mm, 1.5-3.5 mm diam., inside in upper 
third, lobes 3—7 mm, ligulate to elliptic- Milone acute; 


clavate 
villou: 


anthers included, filaments inserted in upper third of 
corolla tube, ca. 0.5 mm; style 6-10 mm, glabrous 


c 


pubescent near apex), stigmas 0.7—1.5 mm. Short- 
styled flowers: similar to long styled except calyx lobes 

4-l mm; corolla tube 4-6 mm, 1—3.5 mm diam., 
lobes ligulate; anthers shortly exserted, filaments 2— 
4 mm; style 2.5-4.5 mm, stigmas 1-2.5 mm. Drupes 
violet-black, subglobose or didymous, 4-8 X 4- 
9 mm; pyrenes spherical or hemispherical, faintly 
rugose, deeply fissured, endosperm entire. 


Distribution and habitat. 
Lanka, where it ca 


forests at elevations of 50—900 m. 


This species grows in Sri 
n be found in wet and premontane 


Phenology. This species has been collected with 
ugh May 


flowers donum thro and August through 
December, and with fruits January through April and 


August through December. 
Discussion. | Gaertnera vaginans is a morphologi- 


cally variable species, but can be recognized by its 


Volume 96, Number 4 Malcomber & Taylor 657 
2009 Revision of Gaertnera 

chartaceous tubular stipules that closely surround the merous, heterodistylous. Long-styled flowers: calyx 
stem, its corymbiform to pyramidal, erect, cymose campanulate, 2-4 mm wide, outside glabrous or 
inflorescences, and its bisexual flowers wik corolla puberulent, with hair-ring inside, lobes 1-2 mm, 


tubes glabrous to puberulent outside. This and a 
number of closely related species form the C. 
vaginans complex; see also the discussion of that 
group for related species and their distinctions and 
see the discussion of G. junghuhniana for further 
discussion of G. vaginans. 

dd eM i examined. SR NKA. Galle 
Dis Hiniduma Pattuwa, Tawalama, mls 11406 
(A, p E, K, L UCM pup Distr.: Pahingala, 
Sohmer 10231 (GH, K, MO, P, PDA, W). Kandy Distr.: 
Laxapana-Maskeliya, Dou Tuum Kostermans 24101 (A, 
BM, G, K, L). Ratnapura Distr.: ey, Sohmer 8797 
(BM, RV MO); Kudawe, Weddagala, Hoogland 11451 (BM, 
G, L, PDA). 


68. Gaertnera vaginata Lam., Tabl. Encycl. 2: 
273. 1797. Sykesia vaginata n 
Revis. Gen. Pl. 2: 426. 1891. 
France" [but 
1983], Commerson s.n. (holotype, FI not seen; 
isotypes, G!, P!, P-LA?) 


Kuntze, 
: “fle de 
surely er an 


Andersonia vaginaia Willd. ex Roem. & Schult., n Veg. 5: 
21. 1819. TYPE: Madagascar. s. loc., du Petit- 
Thouars s.n. (holotype, B-WILLD no 2 

Gaertnera laxiflora Cordem., Fl. Réunion (E. 2 "ds Cordemoy) 

1895. TYPE: La Réunion. Plaine des Caffres, 
Crm Tampon, E. J. de Cordemoy s.n. (holotype, 
ARS). 


ana Cordem., Fl. J. de 

4 895. TYPE: La Réunion. Fia 
Plaine des Palmistes, pres de la Grande Mon 
Propriété Godefroy, E. J. de Cordemoy s.n. (PM P 
not located; isotype, MARS!). 


Min godefroy: Réunion (E. 


17 m tall; 
internodes 


Trees or or branches terete, 


1.54 cm, 


—5 cm, oblanceolate 


glabrous, 2—5.5 mm diam.; 
smooth. Leaf blades 2 2.5-15.5 X 
to obovate, elliptic-oblong, or elliptic, apex acumi- 
nate, base acute, dryin 

glabrou 


condary veins AUR Se abaxially, 7 to 1 


chartaceous, adaxially 


glabrous, abaxially s and often glaucous; 
pairs; 
domatia e petioles 5-23 mm. Stipules tubular, 
glabrous, drying chartaceous, deciduous or usually 


be 4— 


, rounded to narrowly winged, 


persistent at least on distalmost 2 to 4 nodes, tul 
6 mm, with ribs 4 
arising beneath petiole and sometimes extending to 
ire, marcescent, lobes 4 


es, apex enti 
linear or filiform, setae 4 to 8, 


m. Inflorescences 
cymose, many-flowered, terminal on i Dic pal and/or 
axi na eee glabrous to densely Je RA 
pedun .5-6 em; branched porti bifor 
1o x ps cm, branched to 2 to 3 ein 
congested; bracts deltate or fused and trifid, 2— 
mm; bracteoles ovate to elliptic or triangular, 2— 
4 mm; pedicels absent or to 3.5 mm. Flowers 5- 


ovate to oblong or triangular; corolla white, clavate 
in bud, when open salverform, outside glabrous 
edu age 16-25 mm, 4—9 mm diam., inside 
glabrous, lobes 7-11 mm, triangular to ligulate or 
elliptic-oblong, included (shortly 
exserted), filaments inserted at ca. middle of corolla 
tube, 0.5-1.5 mm; style 15-21 m 
1-2.5 mm. Short-styled flowers: similar to long 


acute; anthers 
m, glabrous, stigmas 
styled 
except corolla tube 18-20 mm, 46mm diam.; 
anthers shortly exserted, filaments inserted in upper 
third of corolla tube, 4— 
3.5—4 mm. Drupes violet-black, ellipsoid to subglo- 
bose, 10-18 X 7-15 mm; pyrenes (1)2 per drupe, 
plano-convex, faintly rugose, deeply fissured, endo- 


mm; style 9-10 mm, stigmas 


sperm entire 


Distribution and habitat. 
the Mascarene Islands, where it is known from the 


This species grows in 


it can be found in wet 
lowland forests, humid leds: rers and 
cloud forests at elevations of 100-1800 m 


island of Réunion. Here, 


Phenology. This species has been collected with 
flowers January through May and August through 
December, and with fruits January through July and 
September through December. 


Discussion. | Gaertnera vaginata is widespread and 
Pe collected on Réunion and can be recog- 
d by its acuminate le 


niz aves, stipules with numerous 
sell developed setae, relatively 


large white flowers, 
and ellipsoid or subglobose drupes. 
specimen was attri 

court (1983) noted that this specimen is undoubtedly 
mislabeled and comes instead from La Réunion. The 
fleshy, relatively large, white flowers are said to be 
fragrant (e.g, Lorence 1425 [|M a these were 
collected in the middle of the day orence, 
pers. comm.) The authorship and ae of pub- 
lication of the vaginata have been 
variously and usually incorrectly attributed. The name 
Mussaenda borbonica Lapeyrère (Rev. Agric. Île 
Maurice 2: 85. 1889, as to type based on a plant 
collected at Petit Brule de Saint Denis, La Réunion) 


een cited as synonym of G. vaginata 


cen 1983), but because of the lack of a 
physical type, may not be included in formal 
synonymy 

Representative specimens n. o RÉ- 


UNION. Mare Longue, Barclay AU), Bernardi 
14503 (G, K, P), Cadet 1179 (Kj. dcus 11 (MAU, P); 
Plaine des Chicots, Billiet 832 (BR, K), ur SF 28160 
MAU, P); Plaine des Palmistes, Cadet 592 (K), Friedmann 
2247 (G, MO, P) Kramer 9220 (MO); Saint Philippe, 


— 


Annals of the 
Missouri Botanical Garden 


Bernardi 15035 (BM, K, P), Boyer 59 (P), Schlieben 10921 
(BR, MAU, MO). 


69. Gaertnera viminea Hook. f. ex C. B. Clarke, in 

ook. f., Fl. Brit. India 4: 91. 1883. Psychotria 

viminea Wall. Numer. List [Wallich] Cat. n. 

8354. 1847, nom. nud. Sykesia viminea Sod f 

ex C. B. Clarke) Kuntze, Revis. Gen. Pl. 25. 

1891. TYPE: e. N. Wallich 8354 
(holotype, K!; e. BM!, CGE!, K?. 


Trees or shrubs, 3-6 m tall; branches terete to 
flattened, puberulent or glabrous with indumentum 
drying brown, 1-2 mm diam.; internedes 2—5 cm, 
smooth. Leaf blades (3—4—10 X (0.75—1-4 em, 
linear-lanceolate to lanceolate or oblong, apex 
cuspidate or acuminate, base acute to cuneate, drying 
chartaceous, glabrous; secondary veins prominulous 
abaxially, 4 to 7 pairs; domatia absent; petioles 2— 
10 mm. Stipules tubular, glabrous, drying chartaceous, 


msn 


persistent or caducous, tube 4—10 mm, with ribs 4, 


narrowly winged, arising below petiole and sometimes 
extending to lobes, apex entire, marcescent, lobes (2 
or4, 0.5-1.5 mm, filiform. Inflorescences cymose, 
many-flowered, terminal on principal and/or axillary 
branches, puberulent to glabrous, sessile or peduncle 
1.2-2 em; branched portion corymbiform, 0.8-7 X 
2.5-6 cm, branched to 2 to 4 orders, congested at 
least near apex; bracts deltate or linear to ligulate, 
—3 mm; bracteoles reduced; pedicels absent or to 
2 mm. Flowers 4-merous, unisexual. Pistillate flowers: 
calyx cup-shaped, 1-2 mm wide, outside glabrous or 
puberulent, with hair-ring inside, lobes to 0.5 mm, 
triangular; corolla white, clavate in bud, when open 
, 0.75- 
inside villous in upper shied, lobes 


salverform, outside glabrous, tube 3.5-6 m 


1.5 mm 
0.5-1.5 mm, ligulate or ovate-oblong, acute or obtuse; 


diam., 


staminodia included, filaments inserted in upper third 
ca. 0.3 mm 


(pubescent near apex), 


of corolla tube, ; style 3—4 mm, glabrous 


stigmas ca. lmm long. 
Staminate flowers: similar to pistillate except corolla 
tube 3.5—4 mm, 0.8-1.2 mm diam., lobes ca. 1.5 mm, 
ligulate; anthers shortly exserted, filaments 0.3— 
0.4 mm; pistillode reduced or absent. Drupes violet- 
black, globose, 5-6 X 3-6 mm; pyrenes spherical or 
finely fissured, endosperm 


hemispherical, rugose, 


entire. 


Distribution and habitat. This species grows in 

southeastern Asia, where it is known from Peninsular 

Malaysia and Singapore. Here, it is found in humi 
90 m 


gap 
forests at elevations of 60—1 


Phenology. This species has been collected with 
flowers January through June and in November and 
December and has been collected with fruits, but the 
collection dates have not been recorded. 


Gaertnera viminea is very similar to 
but can be 
distinguished by its relatively slender branches, few- 


Discussion. 
ome plants of G. junghuhniana, 
flowered cymes, and 4-merous flowers. Van Beusekom 
(1967) noted that a report of this species from Borneo 
by Merrill (1921) is apparently erroneous and based 
on misidentification, probably of a plant treated here 
as G. junghuhniana. 


M dE e specimens examined. MALAYSIA. Johor: 


nti, Corner SING 29968 P Gunong Pulai, Md. 


FRI 22075 (K, KEP). SINGAPORE. 

4828b (BM, K, SING), 8923 SM Sinclair 5325 (E); 

National Botanic Gardens, Garden Jungle, Malcomber 3010 

(MO), Ridley 13 (SING) 

70. Gaertnera walkeri (Arn.) Blume, Ill. Ind. Bot. 2, t. 
156b. 1850. Basionym: Sykesia walkeri Arn., Pug. 
Pl. Ind. Or. 354. 1836. TYPE: Sri e s. loc., G. 
W. Walker 102 (lectotype, designated by van 

Beusekom, 1967 [1968]: 380, E). pU 12E-J. 


csi E 4 m tall; branches terete, glabrous, 0.7— 
jam.; interno .5-1 em, smooth. Leaf 

vie (1.5-)2.8-7.2(-9.4) X 0.4-2.9(-4) cm, lance- 
olate or elliptic, apex cuspidate or acuminate, base 
cuneate to acute, drying chartaceous, glabrous; 
secondary veins prominulous abaxially, 2 to 6 pairs; 
domatia present; petioles 2-12 mm. Stipules tubular, 
glabrous, drying chartaceous or membranous, cadu- 
cous or deciduous through fragmentation, tube 3— 
10 mm, smooth, apex entire, marcescent, truncate or 
with lobes 2 to 4, 0.1-0.5 mm, deltate. Inflorescences 
cymose, few- to several-flowered, terminal on princi- 
pal and/or axillary branches, glabrous or puberulent; 
peduncle 0.9—1.4(—2) em; branched portion corymbi- 
form, 1.5-5 X 1.5-4.5 cm, branched to 1(2) orders, 
lax; bracts deltate or linear, 2—4.5 mm; bracteoles 
reduced; pedicels absent or 0.8-11(—22) mm. Flowers 
-merous, heterodistylous. Long-styled flowers: calyx 
2.5-3.5 mm 


puberulent, glabrous inside, lobes 0.3-0.5 mm, trian- 


cup-shaped, wide, outside glabrous or 
clavate in bud, when open 
10-12 mm, 2- 
iam., inside villous at ca. middle, lobes 5— 


gular; corolla white, 


salverform, outside glabrous, tube 


7.5 mm, ligulate to elliptic-oblong, acute; anthers 
included, filaments inserted at ca. middle of corolla 


similar to long styled except corolla tube 8.5— 
1.9-5 
exserted, filaments 1-2.5 mm; style 4-5 mm, gla- 
1.5-2.5 mm. Drupes violet-black, 
9-9 X 8-9 mm; pyrenes 
spherical or hemispherical, rugose, finely fissured, 


mm diam., lobes 5-6.5 mm; anthers shortly 


brous, stigmas 


globose to subglobose, 


endosperm entire. 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


Distribution and habitat. 
Lanka, where it is found in premontane and montane 
6 


This species grows in Sri 


forests at elevations of 750— 
Phenology. This species has been collected with 
through September and with fruits 
April through November. 


owers Februa 


Discussion. | Gaertnera walkeri is generally similar 


to G. ramosa of western Malaysia and to G. rosea, also 
of Sri Lanka; G. ramosa differs in its unisexual flowers 
with shorter corolla lobes, and G. rosea differs in its 1- 
to 3-flowered and fasciculate inflorescences. In areas 
where G. walkeri and G. ternifolia grow sympatrically 
(e.g., Adam's Peak), the natural hybrid G. Xgardneri 
occasionally occurs. Gaertnera Xgardneri differs from 
alkeri in having mixed opposite and ternate leaves 

and ridged internodes. 
The lectotype was originally designated as a 
the G. Walker Arnott 


herbarium, but this collection has been moved to E 


specimen at GL in 


SRI LANKA. Kandy 
Distr.: Knuckles, Madulkelle, Kostermans 25017 (A, BM, K, 
PDA). Distr.: Enselwatte, Sinharaja, 


"d RT 1483 (K, 


Matara 


Soler 10461 (GH, K 
Eliya Dist 


(MO, PDA). Ratnapura Distr.: Dotalugála- Forest 
0). 


aas 1840 (E, L, M 
Names OF Unknown IDENTITY 
Fructesca mauritiana DC. ex Meisn., Pl. Vasc. Gen. 1: 


259, 2: 168. 1840. TYPE: Mauritius, 
designated. 


not 


As discussed in the introduction, this name applies 
to a species of Gaertnera from Mauritius, but which 
species is unclear. 

Gaertnera boivinii Drake, Bull. Soc. Bot. France, 45: 
354. 1898 [1899]. TYPE: Mauritius. s. loc., 
Boivin s.n. (holotype, P not located). 

Neither the 


specimens annotated by Drake with this name have 


Boivin specimen nor any other 
been located and the description is inadequate, thus 
the identity of this name is unknown. 


Gaertnera borneensis Valeton, Bot. Jahrb. Syst. 44: 
568. 1910. TY orneo. Betw. Buntok 


Djihi, Winkler 3321 (holotype, BO not seen). 


s name was treated as a synonym of Gaertnera 
j k 


i 
vaginans subsp. junghuhniana by van Beusekom 
(1967). However, no specimens identified with this 
name by Valeton have been seen, and G. junghuhni- 
ana is circumscribed much more narrowly here than 


by van Beusekom. The description suggests that this 


species probably falls within G. junghuhniana as 
circumscribed here, but because of the complicated 
morphological patterns in this group, this placement 
needs to be confirmed with the study of specimens. 


Gaertnera chapelieri Drake, Bull. Soc. Bot. France 45: 
355. 1898 [1899]. TYPE: Madagascar. s. loc., 
Chapelier s.n. (holotype, P not located). 


Neither the Chapelier specimen nor any other 
specimens annotated by Drake with this name have 
been located and the description is inadequate, thus 
the identity of this name is unknown. 


Gaertnera longiflora C. F. Gaertn., Suppl. Carp. 59, t. 
191, fig. 1. 1806. TYPE: Madagascar. s. loc., 
Commerson s.n. (holotype, TUB not seen). 

This species is described and illustrated as having 
an irregularly, distinctly lobed calyx, v und ovoid 
well-devel 


ig variously sessile to 


fruits, oped bracteoles, and flowers and 
shortly pedicellata: No 
specimen has been seen that can be connected 


conclusively to this name, and its identity is unclear. 


Gaertnera oxycarpa Drake, Bull. Soc. Bot. France 45: 
54. 1898 [1899]. TYPE: probably Mauritius, s 


(holotype, P not 


g w 


c., Dupetit-Thouars s.n. 
located). 


Neither the Dupetit-Thouars specimen nor any 
other specimens annotated by Drake with this name 
have been located and the description is inadequate, 
thus the identity of this name is unknown 


Gaertnera stictophylla (Hiern) E. M. A. Petit, Bull. 
État 29: 382. 1959 
Basionym: Psychotria stictophylla Hiern, in 
Oliver, de Trop. Afr. 3: 77. 
Gabon. Estuaire: Sierra del Crystal Mtns., 
Mann 1721 (holotype, K photo!). 


Bruxelles 


1862, 


Only a photograph of the type specimen at Kew has 
been seen. The identity of this species, including its 
generic placement, is difficult to confirm because the 
e poorly conserved and 
959b) was also 
unsure of the identity of this species, which he said 


stipules on the type specimen are 
the inflorescence is immature. Petit (1 
seemed to him similar to Gaertnera liberiensis but 
apparently lacks the distinctive stipules of that species. 


Gaertnera thouarsii Baill., Bull. Mens. Soc. Linn. 
209. 1879, nom. nud. TYPE: Mauritius. 
P. Dupetit-Thours s.n. (holotype, P not located). 


Paris 1 


This name was published with no description at all, 
simply the explicit citation of the name and Baillon's 
intent to name a species, and the comment that the 
name is based on a specimen of Dupetit-Thouars that 
is found in several herbaria and variously identified as 


Annals of the 
Missouri Botanical Garden 


Psychotria in some and Gaertnera in others. No 
specimens with this name have been located, and its 
identity is unclear. 


Sykesia lanceolata Kuntze, Revis. Gen. Pl. 2: 425. 
1 nud 


This name was published in Kuntze's summary of 
Sykesia, ii name he used to replace Gaertnera, which 
new to be a later homonym. This Sykesia name 
appears in a list of new nomenclatural combinations 
for species previously treated in Gaertnera and was 
said to be based on a Gaertnera name published by 
Bojer, but that previously published name has not 
been found. Kuntze may have intended to reference 
and make a new combination for G. lanceolata Bouton 
ex A. DC., which is otherwise not listed by him, but he 
did not identify the basionym well enough to exclude 
any names published in other genera with the same 
epithet by Bojer, so the identity of this is not clear. 


EXCLUDED NAMES 


Several names listed here were published in 
different, homonymic genera also named Gaertnera, 
which belong to other families as noted. Several names 
published 


neria (Asterace 


in the essentially homonymic genus Gae 
t listed n dile "a 
have sometimes been misspelled as “Gaertnera.” 


ae) are no 


Gaertnera australiana C. T. White, Proc. Roy. Soc. 
Queensland 53: 223. 1942 [= Psychotria sp. 
Gaertnera capitata Bojer, Hortus Maurit. 216. 1837 
— Chas C. (Verdcourt, 1989; he 


conde these two different names, based on 


salia capitata D 


two different types)]. 

Gaertnera caerulea Bojer, Maurit. a 1837. 
Gaertnera coerulea, ort Prodr. 

: 35. 1845 ain ew) DC. 

Verden, 1989, as “G. coerulea”)]. 

Gaertnera cymiflora Bojer, Hortus Maurit. 218. 1837 
[= Chassalia boryana DC. (Verdcourt, 1989)]. 

Gaertnera a. A. Chev., Explor. Bot. Afrique 
Occ. Franc. 1: 444. 1920, nom. nud. [= P. 
quadrifolia Schumach. & Thonn., Verbenaceae 
(Petit, 1959a)]. 


remna 


Gaertnera hongkongensis Seem., . Voy. Herald 
384. 1857. Sykesia Vorne (Seem.) 
Kuntze, Revis. Gen. . Tsiangia 


hongkongensis (Seem.) P. p. Fi But, H. H. Hsue 
& P. T. Li, Blumea 31(2): 311. 1986 [= Ixora 
chinensis Lam. e 2000)]. 

Gaertnera incarnata Bojer, Hortus Maurit. 217. 1837 
[= Chassalia Dos (Poir.) A. Chev. subsp. 
lanceolata (Verdcourt, 1983)]. 


Gaertnera ea E i Gmel., Syst. Nat., ed. 13[bis], 
2(1): 685. 
Nu NE 

Gaertnera lasianthoides C. E. C. Fisch., Bull. Misc. 
Inform. Kew 1927: 209. 1927 [— 
rhinocerotis Blume (van Beusekom, 1967)]. 

Gaertnera "i nas D. Good, J. Bot. mE 
2): 1 929 [= Psychotria gossweileri E. 

Petit p 1959a, b)]. 

Gaertnera lushaiensis C. E. C. Fisch., Bull. Misc. 
Inform. Kew 1928: 411. 1928 [= Chassalia 
m (C. E. C. Fisch.) C. p 2 Fisch., Bull. 

Inform. Kew 1931: 282. 1931]. 
bos Wound Baker, Bull. n Inform. Kew 


— Morinda morindoides (Baker) 


ir madablota Gaertn., 
40)]. 


Psychotria 


Gaertnera obtusifolia (DC.) Roxb., Fl. Ind., ed. 1832, 
2: 369. 1832 [— Hipage amets (Roxb.) DC. 
(Steudel, 1840)]. 

Gaertnera pangati Rheede ex Retz., Observ. Bot. 6: 
24. 1791. Gaertnera pongatii, orth. var. [— 
Sphenoclea zeylanica Gaertn., Campanulaceae 
or Lobeliaceae (Steudel, 1840)]. 

Gaertnera racemosa (Cav.) Roxb., Pl. Coromandel 1: 
19, t. is 1795. Basionym: Lc racemosa 

Cav ss. = Hiptage 

ea? L) Ku, UR s (Jacobs, 


Gaertnera richardii Drake, Bull. Soc. Bot. France 45: 
355. 1898 [1899] [probably — Psychotria sp.]. 


No adequate material is available to ier the 
identity of this species, but syntype specimens ; 
C. Richard 236 and J. M. C. Richard 657 (both Es dry 
with a red-brown cast and have distinet marginal leaf 
veins that are characteristic of Psychotria. 


Gaertnera aes Stapf, Trans. Linn. Soc. London, 


Bot 183. 1894 [= Psychotria sp. (van 
eee 1967 
Gaertnera violascens Ridl., J. Fed. Malay States Mus. 


6: 164. 1915 [= Psychotria sp. (van Beusekom, 
1967) 
Gaertnera zimme nii K. Krause & Gilg, Bot. 


z rman 

Jahrb. Syst. 48: 430. 1912 [= Strychnos sp., 
Loganiaceae (Petit, 1959a)]. 

Literature Cited 


Andersson, L. 1998. A revision of the genus Cinchona 
(Rubiaceae-Cinchoneae). Mem. New York Bot. Gard. 80: 
1-75. 


. E. Rova. 1999. The rps16 intron and the 
eleven of the Rubioideae (Rubiaceae). Pl. Syst. Evol. 
214: 161-186. 

Anonymous. 1997. Mauritius endemic tree refound. Pl. Talk 
8: 18. 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


Amott, G. A. W. 1836. Pugillus Plantarum Indiae oo 
Academia Caesarae Leopoldino-Carolinae, Breslai 


Losas Flora of Maüditius and the 

Seychelles, Lovell Reeve & Co., London 

Baum, I J. Sytsma & P. 
db EE uus of Epilobium (Onagracene ase 
on nuclear ribosomal DNA sequences. Syst. Biol. 
181-207 

Bentham, C. 1857. Notes on Loganiaceae. J. Proc. Linn. Soc., 
Bot. 1: 52-113. 

— & J. D. Hooker. e ha Plantarum, Vol. 2, 

Pad Zi Lovell Reeve & Co. 


Blume, C. L. 1850. Museum Pss anicum. E. J. Bril, Leiden. 

Bojer, W. 1837. Hortus Mauritianus. Aimé Mamarot, 
Mauritius 

Bremekam 1954. Les sous-familles et les tribus des 


Rubiacées. PF Congrès International de Botanique, 
Paris, Rapports et Communications, sect. 4: 113-114 
66. Remarks on the Bm s ae and 
subdivisión of the Rubiaceae. Act exl. 15: 1-33. 
Bremer, B. 1996. Phylogenetic orice MEA eee and 
relationships to other families based on molecular data. 


Ph ylogeny and aerate of 
the subfamily Rubioideae (Rubiaceae). Pl. Syst. Evol. 
ut 


Bridson, D. The identity of Tsiangia (Rubiaceae). 
Kew Bull. sit Ton 1012. 
Brown, R. General remarks, geo a and 


a on the botany of Terra Australis. Pp. 533— 
613 in M. O ene A Voyage to Terra Australis, 


Vol. I. G. & W. , London. 
18. Observations systema me and oe 
on Professor Christian Smith’s Collec of Plants fro 


the Vicinity of T River Congo. Pp. pow in J. H 

Tuckey (editor), Narrative of an Expedition to Explore dive 
River Zaire, Usua alled the Congo, in South Africa, in 
1816, under the inst of Ga J.K. [sic] Tuckey. 


Buckler, E. S., ippolt ito & T. P. Holtsford. 1997. The 
evolution of Viso DNA: Paralogues and phylogenetic 
32. 


Prodromus s Naturalis 
Regni Mu yu va 4. Treuttel € Wertz, Paris. 
Candolle, A. 5. Prodromus ces Naturalis 
Regni Vertis vol 9: “Fortim Masson & x Paris. 
ur m C. B. 1998. TI pt. 155-170 
ne E M. L oek & D. F. is (editors), 
Ti m Taxonomy and Ecol of the Floras of Africa 
nd Madagascar. is Botanical E Kew, Richmond. 
Dallwitz, M. J., T. A. Paine & E. J. Z 
to the DELTA Eae A Coral System for Proce: 
ing Taxonomic RR ad ed. <http://delta-intkey. 
com>, accessed 13 July 20 
is, A. P. & D. Bd 2003. Introduction to the 
odman & J. P 


and a new species, Chimarrhis geniryana. Brittonia 48(1): 
35-44. 


De Queiroz, K. 1998. The general li 
Species criteria, an e process of speciation: A 
conceptual unification and terminological 
tions. Ts 57-75 in D. J. Howard & S. Berlocher 
(editors), Endless Forms: ENS and Speci. Oxford 
ae Press, New Y 

The El eN concept of Du and 

EE n properties of the species concept. Pp. 49-89 

New Inerdincilinar 

e 


Comoro Tdantis: Royal Botanical Gardens, Kew, Rich- 
ond. 


DE J. J. & J. L. Doyle. 1987. A rapid isolation procedure 
for small quantities n fresh leaf tissue. Phytochem. Bull. 
: 11- 


Drake, E e sur deux genres de Rubiacées des Iles 
del liis orientale. Bull. Soc. Bat France 45: 344—356. 
Du Puy, D. J. & J. Moat. 1998. Vegetation mapping of 
classification i in Madagascar (using GIS): Implications and 


f the Floras 
and Madagascar. Royal Botanical Gardens, Kew 


n Chorology, Taxonomy and Ecology o 
of Afric 
Richmond. 
—— & — 03. Using geological substrate to 
identify and m ry vegetation in Madagascar 
and the dedos for planning biodiversity conserva- 
tion. Pp. 51-67 in S. M. Goodman & J. P. Benstead 
(editors), The Natural History of Madagascar. University of 
Chicago Press, Chicago. 
Endlicher, S. 1838. cs Plantarum Secundum Ordines 
Naturales. Fr. Beck, W 
Fischer, C. E. C. 1927. XXVIL neces to the a of 
Burma: III. Bull. Misc. Inform. Kew 3 
——. 1928 


à XXIII. Decades a o 
Novarum i in Herbar ie NE Conservatarum: Decas 
CXX. Bull. Misc. In [e w 1928: 141-147. 


Friis, I. 1998. Frank White = the development of African 
chorology. Pp. 25-52 in C. R. Huxley, J. M. Lock & D. F. 
Cutler (editors), Chorology, Taxonomy and Ecology of the 

ca and Madagascar. Royal Botanical 


1806. Supplementum carpologiae: Seu 
continuati operis Josephi Gaertner de Fructibus et 


seminibus Du C. F. E. Richter, Leipzig. 
Gautier, L. a mee . Goodman. E Introduction to the flora 
f Madagascar. Pp. 229-250 ii dman & J. P. 


Bensiead “editors The Natural History of Madagascar. 


smans, E. Sm 
fr MO en 


soamerican 
Psychotria subg. Psychotria (Rubiaceae). Ann. Missouri 
Bot. Gard. 76: 67-111, 386—429, 886-916. 
Huelsenbeck, J.P. & F: Ronquist. 2001. MrBayes: Bayesian 
55. 


1931. Flora a West Tropical 
rica. Crown Agents, Panden. 
— € . 1937. Tropical African plants XV. Kew 


— Bull 1937: 62— 63. 

Igersheim, A., C. Puff, P. e & C. Erbar. 1994. Gynoecial 
development of Gaertnera Lam. and of presumably allied 
taxa of the Betas (Rubiaceae): Secondarily “supe- 
rior” vs. inferior ovaries. Bot. Jahrb. Syst. 116: 401—414. 


Annals of the 
Missouri Botanical Garden 


IUCN. 2001. IUCN Red List Categories and Criteria, Version 
3.1. Prepared by the IUCN Species Survival Commission. 
m coe d, Switzerland, and Cambridge, United 


E M. 1955. o E Flora Malesiana, ser. 1, 
Spermatophyta 5: 125-143. 

Jansen, S., E. poca E. Smets. 1996a. The e 
value of endex namentation in some Psychotrieae 
pollen (Rubiaccac-Rubioideae). Grana 35: 129-137. 

H & E. Smets. 1996b. 
mom and Pa, amea: Genera within the Psychotrieae 
or constituting the tribe Gaertnereae? A wood anatomical 
palynological approach. Bot. Acta 109: 466-47 

& 997. Wood "— of 


Beeckman 


the predominantly African mM of the tribe 
Psychotrieae (Rubiaceae-Rubioideae). Int. Assoc. Wood 
Anat. J. 18: 169-196. 


Jussieu, A.-L. de. 1807. Sur les charactéres généraux des 
familles, tirés des eand, et confirmés ou rectifiés par les 
observations de Gaertner. Ann. Mus. Hist. Nat. 


Umfang ps Inhalt der Familie der 
—338. 


m ra 
itinere mundi collectarum, Vol. 


Lamarck, J. B. A. P. M. de. 1792. Tableau Encyclopédique et 
Méthodique des Trois Régnes de la Nature, Vol. 1. 
Panckoucke, Paris. 

Lawrence, G. H. M. 1951. Taxonomy of Vascular Plants. 
WC Co., New Yo s 

Maddison, & Maddison. 2003. MacClade: 
Analysis x Phylogeny B Character Evolution. Sinauer 
Associates, Ea Bede Massachusetts. 

S. Systematics and Evolution of 

Breeding P A Gaerinera (Rubiaceae). Ph.D. 

Dissertation, Washington University, St. Louis. 

2002. Phylogeny of Gaertnera Lam. (Rubiaceae) 
kers: Evidence of a rapi 
radiation in a E morphologically diverse genus. 

Evolution E 42— 

& 


"EN 2005. new species of 

Eug uin from Madagascar and phylogenetic 

nalyses that support Hymenocnemis as a synonym of 

Gan nera. Monogr. Syst. Bot. Missouri Bot. Gard. 104: 

371-398. 

Maldonado, G. C. B. 2005. A Revision of the G Elaeagia 

(Rubiaceae). Thesis oo of Systematic Botany, 
versity of Aarhus, Aarhus 


t, V. Wr. D. L. 

o, P. C. 

sade H; Wiersema & N. J. Tarland eden 

06. Dod Code V Botanical Nomenclature 
(Vienna Code). Regnum Veg. 146. 

Merrill, E. D. 1921. A bibliographic enumeration of Bornean 
plants. J. Straits Branch R. Asiat. Soc., special number, 
1-637. 


we B. D. 1999. Getting rid of species? Pp. 307—315 in 
New Interdisciplinary 


. The rollo of 
dioecy fons distyly: Reavaliisticn of a hypothesis of the 
loss of long-tongued pollinators. Amer. Naturalist 133: 
149-156. 


Nepokroeff, M., B. Bremer & K. Sytsma. 1999. Reorganiza- 
tion of the us Psychotria and the tribe Psychotrieae 
(Rubiaceae) ened from ITS and rbcL sequence data. 
Syst. Bot. 24: 5-27. 

Petit, E. 1959a. Les Gaertnera Lam. (Rubiaceae) de l'Afrique 
tropicale et spécialement du Congo belge. Bull. Jard. Bot. 
État cs 29: 37-53. 

Ob. Historique de la position systématique de 

m Lam. et remarques sur quelque espéces 

pr du genre. Bull. Jard. Bot. État Bruxelles 29: 
- m 

——— Rubiaceae Africanae IX. Notes sur les genres 
Ag. pin Aulacoc 
Morinda, Mussaenda. 


Tricalysia. Bull. Jard. Bot. État ‘Bruxelles 32: —1 
ral i ZU. Carpelgy: and Pollen n. of 
the Towards a 


Rubioideae), 
Tal a ae Delimitation. Dissertation, Catholic 
University of Leuven, Leuven, 

, S. Jansen, I. James, E. Robbrecht & E. Smets. 2001. 
Morphology, Renee and taxonom 


meopsis (Rubia one Fubioid» e). Brittonia 33: 490—504. 
osada, D. & dall. 1998. MODELTEST: Testing 
the model of DNA substitutions. Bioinformatics 14: 
817-818. 


Ridley, H. N. 1908. > a e S 2 made by H. C. 
obinson and L Tahan, Pehang. J. 
Linn. pe Bot. N 301 
915. Plants fom | Gunong Kerbau, Perak. J. Fed. 
Malay $ Sat Mus. 4: 4 
mue ren p a Td of British North 
EL Kew Bu 
m E. 1988. Tropical iod a N Opera Bot. 
Belg. 1: 1-271 


: Cenere distribution patterns in Subsaharan 
Afen Rubiaceae (Angiospermaceae). J. Biogeogr. 23: 
311-328. 


anen. 2006. The major evolutionary 
lineages of the coffee family (Rubiaceae, angiosperms). 


position of Coptosapelia and Luculia, supertree 
construction based 
rbcL data. new classification in two subfamilies, 


po e and Rubioideae. Syst. Geogr. Pl. 76 

85-1 

ene A J. & Moos 1819. Systema Vegetabilium, 

5. J. G. Catt Es 

eet TS & T. Maniats. 1989. 

Molecular us ^ Laboratory Manual. Cold Spring 
Harbor Press, New York. 

Schumann, K. 1891. Rubiaceae. Pp. 1-156 in A. Engler & K. 
Prantl (editors), Die Natiirlichen Pflanzenfamilien, IV, 
Teil, 4 P W. Engelmann, Leipzig. 

d e H. 1890. Studien uber die Tribus oo 

ook. Ber. Deutsch. Bot. Ges. 8, Gen. -Versam 
P ETIN 

Stafleu, F. A. & R. S. Cowan. 1986. Taxonomic Literature, 
2nd ed., Vol. VI: Si- Vuy. Regnum Veg. 115: 1-926. Bohn, 
Scheltema & Holkema, Utrecht. 


Steudel, E. G. 1840. Nomenclator botanicus, seu: Synonymia 
plantarum universalis... Vol. . G. Cott, Stuttgart. 
Takhtajan, A. Medsis Region of the World. 


University of joe Press, eley. 

A. Steyermark, P. c. Delprete, A. Vicentini, 
R. Cortés, D. Zappi, C. Persson, C. B. Costa & E. A. da 
Anunciagáo. 2004. Rubiaceae. Pp. 497-847 in J. A. 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


Steyermark, P. E. Berry, K. Yatskievych & B. K. Holst 
d Flora of the ku mm Vol. 8. 
Missouri Botanical Garden Press, St. 


18 
dijs J» T. Olsen & D. Tiene 1994. JEDE. Ww, 
1 


s Gaertnera Lamk. 


ndex Herbariorum Part II(4), Collectors 
475-576. Bohn, Scheltema & 


83. Index Herbariorum Part II(5), Collectors N— 
R. ios Veg. 109: 577-803. 

1988. Index Herbariorum Part II(7), Collectors T t/m 
Regnum Veg. 117: 987— 
eo B. 1958. Remarks on de classification of the 

Rubiaceae. Bull. Jard. Bot. État Bruxelles 28: 209-290. 
1983. Notes on Mascarene Rubiaceae. Kew Bull. 

37: 521-574. 

——. 1989. 108. Rubiaceae. Pp. 1-133 in J. Bosser, T. 
Dd bin on & Maras (editors), Flore des 
a Ré Rodrigues. Sugar 

Industry (edes Institute, “Port Louis, Mauritius. 
Vicentini, 


, Maurice, 


> 
R 
. Q 
aa 
78 
2g 
E 
Z 


Blumea 18: B e 


ceae- a Ber. D 
White, F. 1993. The AETFAT ehorolepical classification ha 
Africa: History, methods and applications. Bull. Jard. B. 
Nat. A 62: 225-281. 
98. The vegetative structure of African Ebena- 
ceae A the evolution E rheophytes and ring species. 
opkins, C. R. Huxley, 
Pannell, G. T. Pines & F. pos peo The Biological 
Monograph. E Botanical 2) Kew, Richmond. 
White, T. T. Bruns, 
Amplification and direct sequencing of bet ribosomal 
RNA genes for jap ide fe 315-322 in M. A. Innis, 
D. H. Gelfand, J. J. 5 & T. J. White (editors), PC R 
Protocols: A Guide. to T and Applications. 
Academic Press, San Diego 


Appendix 1. Index to numbered collections. 


ecimens are listed alphabetically according to the 
collectors excluded. 


indicate type collections. Dorr (1997) presented additional 
any collectors in Madagascar. Som 
"ndm no Migs 
others without initials here have not e trace 


& 
È 
© 


LIST OF GAERTNERA SPECIES AND VARIETIES RECOGNIZED IN 
This REVISION 


l. Gaertnera alata Bremek. ex Malcomber € A. P. Davis 
2. Gaertnera alstonii Malcomber 


. Gaerinera T Malcomber 


. Gaerinera belumutensis Malco od 

. Gaerinera bieleri (De Wild.) E 

. Gaerinera Mi di Une d & A. P. Davis 
10. Gaerinera calycina Boje 
ll. Gaerinera capitulata Malcomber 

. ert 


3 

4, 

5: ber 

6. onan ee Malcomber & A. P. Davis 
E: 

8 

9 


ojer 
16. Gaerinera darcyana Malcomber & A. P. Davis 
17. Gaerinera divaricata (Thwaites) Thwaites 
l; 


era fractiflexa Beusekom 
23. Gaerinera furcellata Baill ex Vatke) Malcomber & A. P. 
Davis 
24. Gaerinera gabonensis Malcomber 
25. Gaerinera Xgardneri Thwaites 


ra i a Baill 
34. Gaerinera junghuhniana Miq 
35. Gaerinera kochummenii o 


gifolia 
40. Gaerinera. D ur (Schein. ex Hiern) E. M. A. 
Petit 
40a. Gaerinera longivaginalis var. bracteata (E. M. A. 
Petit) Malcomber 
40b. SO longivaginalis a ex Hiern) E. 
M. A. Petit var. po ina 
41. Gaerinera lowryi Malcomber 
42. Gaerinera 


E 


45. nom microphylla Capuron ex Malcomber & A. P. 
Davis 


46. Gaerinera monstruosa Malcomber 
47. Gaerinera obesa Hook. f. ex C. B. Clarke 
48. Gaerinera oblanceolata King & Gamble 
49. FE M obovata Baker 
ertnera obovata Baker var. obov 
D» Gaerinera obovata. var. o (Baker) Mal- 


mber 
50. Gaerinera paniculata Benth. 
51. Gaerinera pauciflora Mica & A. P. Davis 
52. Gaerinera pendula Bojer 
53. Gaerinera phanerophlebia Baker 
54. Gaerinera phyllosepala Baker 
55. Gaerinera Pur em Baker 
56. Eun p (DC.) Baker 
57. Gaerinera ramosa Ridl. 
58. rine nae Malcomber 


Annals of the 
Missouri Botanical Garden 


59. Gaerinera rosea. Thwaites « ex Benth. 


um. 
64. Gaerinera sralensis a ex Pit.) Kerr 
65. Gaerinera ternifolia Thwa 
66. Gaerinera trachystyla (Hiem E E. M. A. Petit 
67. Gaerinera di Ap (DC) M 
68. Gaerinera vaginata Lam 
69. eee viminea Hook, f. ex C. B. Clarke 
10. Gaerinera walkeri (Arn.) Blume 


bang Mohtar S 41768 (18), S 52719 (18); Abraham, J. P. 
SF25874 (49a), SF7811 i SF7997 (49b); Abu Nawas A 
834 (34); Adam MAU 11606 (56); Adam, J.-G. 124 (40b), 
3658 (40b), 4697 (40a), 5997 (50), n im 16307 c 
20579 (40b), 22252 (50), 22405 (50), 27518 (40b), 28771 
(40a), 28815 (50), 29544 (50), 29749 eo 29876 (50); 

Adames, P. 568 (50 


26415 (58); Alphonse [no inital Dorr, 1997] RN 8677 (49a); 
Alston, a ls (2); Amin, G. 86356 (34), SAN 95125* (2), 
SAN 97444 (34); And ns . R. 8356 (34), SAR 9863* 
(26); Dur J. W. 19 dacianstiseta 158 GD: Armand, 
W. 6 (49b), 26 (58); Ar mange b EE 5 (68); Ashton, P. S. 
BRUN 165 (34), 192 (18), 916 (65), 16599 (18), 17970 (34); 
Asonganyi 161 (66), 34 db "die 11 D Aubréville, D. 81 
(50), 1134 (50), 733 (50); Audru, J. 29996 (50). 
Badré, F. J. 763 (68); Bakshi 184 D Balakrishna 

528 (17), 548 (65), 969 (67); Baldwin, J. T. 9833 (50), o 
(13), 10928 (13), 11121 (13), 11554 (50), 13057 (50); 
Bamps, P. 379 (40b), 2041 (5), 2263 (40b); Barker 1267 (13); 
Barkly 536 (68), 630 (68), 1156 (56), 1374 (56); Barnes, E. 
SING 10889 (57); Baron, R. 149* (49a), 158 (44), 366 (44), 


18), 8271 (18); Beaujard liil s Dorr, d 183 
ier 184 (43), 387 (42); Beceari, P. B. 1799* (62); Beddome, 
H. 5309 (70), 5310 (70); Beentje, H. 568 (5); ie 2911 


JEDE 
= 


F. M. MAU 1094, (60), MAU 1627 (10, MAU 2734 (52), 
MAU 2753 (10); Billet 437 (68), 452 (68), 479 (68), i. 
(68), 832 (68); Birkill, SING 234 (47); Birkinshaw, C. 


9 (43), 1780* (53), 2074* (4); 
; Bos, J. J. 2026 (13), 2984 (13), 3166 (66), 
ae (66), PS (66) si (66), 5766 (66), 6340 (8), 6570 
66); Bos 49b), 6722 (44), 8200 (44), 9453 
68), 11621 (68). p (44), 14413 (49b), 20811 (68); 
92 (53), RN 3734 (31), to 6077 (54); 

3); E Station Team 330 (50), 1974 


ZAR 


Bouet, Dr. 8 (6 
( n, L. S. MAU 1080 (20), 1085 c Boyer, G. 59 
(68); Bremer, B. 895 (59), 1006 is 1051 (70); Brenan, J. 
934. ; Breteler, F. 344 (66), 382 (66), 389 (63), 1626 
(40b), 5458 (40a), 5533 (50), in (50), 6275 (50), 7651 
(40b), 8029 (24), 10214 (24), 12496 (40b); Breyne 94 m 
3081 ims 3417 (40a); Brown, R. C. 958 (50); Brummitt, R 

13918 (50); Bünnemeyer, H. A. B. 1995 (27), 6512 (27, 


= 
[^u 


7686 (34); Burgess FRI 9100 (18); Burkill, I. 52 (18), 136 
(69), | ju SING 7613 (27), 7823 (18), 8521 (18), 8608* 
(18), 12902 (48); m y 1157 (34), 3066 (34); Burtt, B. D. 88 
(65); ela P.6 
adet, L. J. T a (68), 592 (68), 698 bis (68), 1179 (68), 
m A 3859 me Cadinouche MAU 2 (56); Caille, 
94 (50), 14992 (50); Callens, H. 3247 (8), 3409 (8); 
a - 67 (47), 159 (27); BLU Collector 164 (27); 
Capuron, R. [P.R.] SF 255 (49a), SF 568 (55), 1641 (44), SF 
20287 (44), SF 23220 (51), SF 23658 (53), SF 23658 bis 
(28), SF 23800 (46), SF 23805 (33), SF 23845 (53), SF 24022 
(45), SF 24027 (44), SF 24065 bis (33), SF 24355 (45), SF 
24413 (44), SF 24770 (45), SF 24978 (32), SF 28160 (68), SF 
28204 (68), SF 28260 (20 Me SF i (60), SF 28396 (46), SF 
28774 (44); Carmichael, 2 (68); Catat, L. D. M. 2524 
(33); Chalot, C. 29 (63); Chane SA 5 (27); Chapman, J. D 
3992 (50), 5247 (50); Chatelain, C. 2 (5), 538 (5); Chevalier. 
11206 (50), 12420 (40a), 12664 (40a), 12936 (40a), 
12984 (40a), 17142 (50), 17159 (40b), 17250 (5), 17542 (5), 
22774 (50), i ae Chew 735 Los es i A (50); 
Christiaense 37); Christo; 4 (32); 
Church 1768 Do Chon € 6101 (65); p a 40466 
(34); Comanor 1200 (67 merson, P. 371 (10); Coode, M. 
J. E. 4283 (56), 4765 (56), 4894 (52), eee Yo Cooper, G. 
P. 202 (13), 277* (38), 287* (13), 465 (13); Corbisier- 
Baland, A. 1120 (40b), 1266 (50), 1385 (40b), 1917 (40b), 


aS 


. RN 50 i Cours, G. 1129 (5 


( 
0), 94 (50), 118 (405). Cramer 4395 (6 
4506 (70), 4641 (70); Croat, T. B. 32246 (55), 32579 (2 
32604 (43), 32613 (54), 32647 

D'Alleizette, C. 1353 (19, 1477 (55); D'Arcy, W. C. 
15273 (54); D'Argent, G. MAU 21399 (56), MAU 22454 (56); 
Daramola, B. O. 451 (50), n 12488 (50); Debeaux, G. 412 
(50); Decary, R. 111 (33), 1949 (32), 5200 (43), 6907 (55), 
7085 (55), 10135 (49b), a (58) 11018 569) lO es, 
14342 (58), 14790 (49h), 15297 (55 6 (54 
(28), 16917 (42), 16935 (54), pr (42), p (42), ion 
(42), 16982 (4), 17694 (55), 17713 (28), 17770 (42), 17783 
(55), 18428 (53); Dechamps, R. 13067 (50); De Giorgi, S. 
258 (40b), 271 (50); De Graer, P. 348 (40a); Deighton, F. C. 
3 (40b); Demange, R. 3036 (50); de Néré [no initial] 1172 
( ; Dequaire 27657 (54), 21671 (55), 27126 
(51), 27977 d Derleth 118 (43); Up B. 180 (55), 


). 
8), 


1097 (50); Dinklage, M. J. 2986 (13), 3056 Q3; Dorr, L. 

3080 (55), 3222 (42), 4484 (43); Doumenge 411 (8); Dowsett- 

Lemaire 1936 (5 3 (4); Dumetz, N. 858 (28), 914 
0). 


REU 13871 68); Ekwuno [no initial] 284 
86 (50), GC 42614 (50), GC 42670 
9 Be Evrard, C. M. 2827 


E = R. 16/503 (17), 77/42 (67); Fanja 539 (55); Fay, J. 
M. 8106 (50), 8111 (50); dum SAN 94840 (34); Fleury, F. 
26727 a Florens, D. MAU 22551 Do MAU 22711 (56), 
MAU 22725 (15); Forbes 3214 (34); Forsyth-Major, C. I. 304 
(49a), 317 (44), 401a (44); Fosberg, R. 56500 (59); Fotius 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


2690 (50); Friedmann, F. 555 (68), 2: (68), 2247 (68), 2845 
(52), 2984 (52), ue 0), 3335 (6 


Gachet, C. SF 1 (53) Cds G. 581 (70), MAU 
19158 (56); MR. T 228 (68); Gau a em 2514 (43), 
2807 (4), 2817 (53); Gautier-Beguin, D. 2 (13), 1107 


(50), 1263 (40a); Geay, [M.]F. 7370 (4), n s 7719 E 
Geerling, C. 420 (50), 422 o 1689 (50); Gent 
11575 (32); Gereau, R. E. 3236 (4); Germain, R. G.A 
(50), 7551 (37), ino (40a); Gerrard, W. T. 37 (53), 54* ru 
Gilbert, G. C. L. 14173 (50), 14813 (50); Gillet, J. 2809 
(40a); eodeni J. S. 1630 (47), 4827a (69); Goossens: V. 
G. n (40a), 5017 (8), 6197 (40b), 6208 (50); Gossweiler, 
08 (50); Gueho, J. MAU id (56), MAU 17853 (52); 
EM J. 27* (33), 36* (28), 38 

Hallé, F. 4 (5), 146 (50), 217 e. 284 (5), 296 (5), 380 
(40a), 409 (4), 418 (13); a N. 2010 (18), 4437 (66), 4617 

(66), 4664 (66); Hallé, W. (66); Haniff, M. SIN 
(57), SING 21055 (48); e D. K. 3751 (40b), 3753 (50); 
Hardial 628 (27); EA R. M. 1035 (40b); Harley, W. 1910 
76 (40b); Hart, T. 319 (37); Hassan $ 
Haviland, G. D. 3460 (34); Henderson, 
A R. 11476 (18), 17769 (57), SING 23390 (57); Hepper, F. 
N. 1514 (50), 4545 (67); Hill, H. C. 383 (27), 397 (27), 424 
(27), 425 (27), 470 (27); Holttum, R. E. 53 (7), 10687 (7), 
SING 19944 (47); Homolle, A. M. 61 (42), 1843 (55), 1926 
i EL i p e 2373 (55), 2405 (53), 2543 (55); 
and, R ii 11451 ME n 619 (48), 649 
a p ber Rn qa 825 (70); Hullett, R. W. 113 (47), 756 
57); Humbert, J. H. 3366 (58), 3450 (43), 3566 (16), 3578 
(44), E s 5999 (28), 6056 (58), 6250 Pd 6322 
(49b), 17480 (49a), 1 9 (49a), 18201 (49a), 


e 


(55), 22045 (32), 22106 (55), 22114 (43), 22118 (49a), 


p- n E (54), 504 (54), 510* (55), 655* (31); Hunter, K. 


po d P. S 39183 (18); Imbert, T. 94 (28 
amal ae 36172 (57); Jacques-Félix, H. 3 
2444 (50), 8 , P. 1166 (40b), 1290 60), 1665 
oy 7158 a 7601 (50), 7688 (50), 7832 (50); Jansen, J. 
W. A. 754 (13), 1071 (13), 1119 (50), 1515 (40a), 1615 (13), 
1650 s s (13); Jaofety-Bosy [no initial; Dorr, 1997] 
RN 10096 (43); Jayasuriya 800 (17), 916 (70), 1509 (17), 
s D. 2961 OD 2970 (70); Jeffrey, C. 231 (8); Jolly, A 
4 (63); Jongkind, C. 878 n 2077 (4), 2085 (4), 4982 
Ho Jumali, K. 3004 C A3 
Kahindo SF 116 (40b); uc F. G. KEP 20457 (57); 
ee m p^ i n A J. 1995* (17); 
enfac 763 (36); Kerr, 4 (64), S pe 
= cM 17996 (64), 19137* s [en 
"ei 8 (7); Kiah, M. S. SFN 32015 (62), SING bonia uns 
Kiener, A. SF 50 > 28); Kiew 2192 (7), 2197 (18); wet 
Collector 8449* ; Klaine, R. P. 52 (63), 149 (63), 2 
(63), 203 (50), a (63), 421 n 516 (63), 872 (63), T 
(63), d (50); Kloss, C. B. 596 4); O 2 M. 
2311 (18), FRI 2387* 2 FRI 16671 (18), FRI 32515 
(34); Kostermans, A. J. G. H. 238 (34), 1315 (34), 2059 (34), 
23597 (59), 23659 A (70), ds B (70), 24101 (67), 24142 
(65), rem dE do BN ps ird 25017 2h 27020 
(65), 27135 (17), 27527 (17), 27573 (65), 28274 (10) 
a 7 267 e MS (9); DAR [F. or i H. de; 
which is unknown] 92 me 68). 
ibosaka, R. RN 10945 (31), RN 12597 (43), RN 10926 
(55); Lalouette, J.-A. MAU 10010 (60); Lam, H. J. 5530 (33), 


abe: 


a (68), 5628 (54), 5628A (49b), 5720 (33); Langlassé, E. 


J. M. 1814 (5), 2366 (5), 3798 (5), 4154* (5), 4840 (13), 5074 
(8), 5350 (8), 5534 (66), 5687 (66), 5710 (50), 7106 (50), 
7677 (50), 7992 (5), 8889 (50), 10352 (50), 11508 (40b), 
14017 (58); Léonard, J. 84 (50), 137 (50), 2914 (37), 3743 
(8); Leroy, J.-F.[P.] 8 (63); Lesmy FRI 35937 (47); Le Testu, 

42 (24), 7622 (40b), 8575* (24), 8642 (50), 9143 (40b), 
9572 (40b); Letouzey, R. 2116 (40b), 2498 (50), 3780 (40b), 
12457 (50), 12601 (40b), 12648 (60) 13702* (36), 14153 
60. 14901 (66), 15091 (50), 15223 (66); Lewis, B. 756 (58), 

86 (55), 819 (58), 1030 (49a), 1253 (32), 1306 (49a), 1318 
oo Liengola 157 (37); Linder, D. H. 286 (13), 758 (40a), 
1487A (13), 1487B (13); Lisowski, S. 16116 (8). A (8), 
45071 (8), 52450 (8), 90036 (8), 90039 (8); Lorence, D. H. 
1495 (56), 1541 (56), 1548 (60), 1597 (56), 1956 (56), 2125 
56), 2201 (20), 2209 (56), 2224 (39), 2415 (68), 2674 69. 


— 


= 


Low 
4262 (55), 4272 (42), 4263" (45), 4471 (12), 4486A (12): 
Lyall, R. 161 (55), "n 

Macrae, J. 602 (7 "Em SAN 107944 (2); Maingay, A. 
C. 2701 (47), 925 s Malcomber, S. T. 930 (4), 999 (58), 
1018* (58), 1065 (49a), 1152 (4), 1172 (58), 1275 (4), 1460 
(49a), 1799 (4), 1946 (4), 2027 (34), 2028 (34), 2155 (58), 
2173 (4), 2229 (4), 2585 (42), 2589 (55), 2594 (43), 2637 (4), 


2917 (49b), 2918 (49b), 2919 (49b), 2920 (43), 2921* (16), 
2922 (16), 2925* (45), 2926 (49a), 2927 (55), 2928 (49b), 
2929 (49b), 2930 (56), 2931 (56), 2932 (39), 2933 (52), 2934 
(29), 2935 (39), 2936 (52), 2937 (29), 2938 (56), 2939 (56), 


3025 (47), 3026 (62), 3031 (34), 2 (22), 3037 (11), 3038 


Annals of the 
Missouri Botanical Garden 


(26), 3039 (48), 3040 (48), 3044 (34), 3045 (34); Mann, C. 
1791* (66); Manning, S. Med ), 2200 (50); Manohiraza [no 
initial; Dorr, 1997] RN 8709 (55); Marmo, V. 18 (40a), 19 
(40b); Martin, S. 38199 (34); Mat [no initial; Vegter, 1976] 
3723 (7); Mat-Salleh, K. 3250 (34); Maxwell, J. F. 78-339 
5 (27) 81-159 (27), 985 (67); 


(50 

12 (4); Messmer, N. 690 (1), 
Mildbraed, G. W ae per (37), 3287* (37); Miller, J. S 
3818 (42), 3821 (55), 3918 (55), 3978* Pa is E 4189 
(32), 4488 ( 4667 (32), 4483 (51); M 155 
(30), 167* (30), 197 (4), 221 (53), 273* (19), 349 (54) oo 
3457 Moise [no iu A (28), 5 (34), 6 (54), 9 (61), 10 
(53), 11 (30); Morat, 6 (68), 3215 (49a), e A 
8576 (42), 8587 (30), m D 8613 (30); Morton. 
8473 (13); Moureau, J. ne Moysoy, L. one e 
(57); a i 334 (4 Te Kin bor 29840 (34); 
34); Muslim Plant "collec 

Ndjele i (8), 735 (8); Ng, F. S. P. FRI pos (18), FRI 
22056 (18), FRI 22075 (69), FRI 5233 (47), FRI 5893 (57), 
Nicoll, M. F. 195 (58), 


27); Noorebooti, H. P. 1526 (18), 

3122 (65), 3123 (70), ES. (17), 4324 (18); Nur, [n 1165 

(69), 7807 (47), 11174 (18), 11613 (18), SING 34357 (48). 
Okeke, R. FHI 38440 (50); Oldeman, R. 240 (5), 468 (50); 

Onochie, C. F. A. FIH 33168 (21); Othman, S. 19909 (34); 

D. 24 (55). 

.94 m. a (52), 227 (52), te 22729 (39); 
Paivo d RN 8866 (54); Pauwels, L. 4756 (40 

ve 6656 (8); Pais J. 2457 (4); Perrier de la Báthie, [J. 
M.] H. [A.] 3649 (43), 3661 (49b), 3744 (33), 3754 (33), 

3806 P 3836 (32), e (55), 4014 (55), 6892 (44), 6928 


(8), 2567 (40a); Pierre, E B.] L. 1253 rs Pisch m (50); 
Poilecot 3447 (5); Price 29 (34); Puff. dos 
800825-1/10* (29), 2 1/1 (34); ee iv 
(18), 4227 (57); Put, N. 2864 (64), 2939 (64); Pynaert, a o 
(40b), 1788 i 

ohitra, A. R. 2159 (28), 2578 (55), 2587 (43), 4195 
(28), SF 29901 E SF 29861 (4 us di T G. 
OUEM 31 (33); Pres G. 9 (61), 923 (61); 

"ES V. 145 (42), 193 en ce (28), 1502 (19), 
1502 (33), 2151 G3. Rahmat Si Boeea [or Rahmat Si Toroes, 
no pe DN 1983] 5418 (34), 5722 (34); Raimundo, ~ 
R. F. 6 (50); Rajanaparany [no initial; Ls 1997] R 
8162 e Rakoto, E. RN 1487 (49a); Rakotomalaza, P. j 
1317 (49b); Rakotoniaina RN 2426 oo. RN 2446 (54), RN 
2448 (53), RN 2451 (54), RN 3218 


9602 53) Rakotovao, C. 108 (58); 
593 (49b), 607 (49b); Rakotozafy, A- 521 (33); Ralaikoto 197 
RN 


5 r G 


= 


(53); Ralarivohita SF 1 
2814 (55); 69 (55); Randriamampio- 
IS ( ): Randrianasolo, ^ 70 (55), 288 (28), 320 


anjokiny, A. 0 (28), ae "x n 
acus F: E o Apart m 

(55), RN 1338 (55); Ratsirahonana, L J. p p as 
Ravelonarivo, D. 680 m 903 (53); Raynal, J. 21012 (50); 
Razafimanantsoa, A. RN 1563 (4); Razali Salam 109 (18); 
Reitsma, J. 1336 (63), ps (66), 2645 (50), 3131 (66); Rena, 
R. S 60874 (34); Rena George S 40507 La S 58327 (11); 
Reygaert, F. J. 624 (37); Richard, J. M. C. 5* (33), 26 (33); 
Richards, P. W. 3249 (50); Richardson, n B. K. 4046 (56), 


4078 (60), 4090 B (68), 4102 (68), 4148 (68); Ridley, H. N. 


= 
oN 
© 
= 
5 
3 
o 
m 
ron 
£ 
ES 
a 
d 
=) 
AN 
= 
E 
= 
Y 
A 
I 
o 
qa 
N 
J 
= 
Y 
A 
RI 
S 
I 
p 
N 
Jj 
= 
N 
A 
Go 
S 


( 
10943 (69), 11123 (27), 12080* (18), 12528 (27), 13337 
(27), P. (51), i 8) 2o C. 1934 co, 1935 
(70), 1943 (70), 204 ajanoparary SF2830 (42); 
hue. ra 12815 vs 2T (50), 15361 | (13) Robinson 
H. G. 5488* (57), 6275 (34); Robyns, F A. 
59673* (11), $ 62906 2 
Sajy, C. RN 884 ee Saw FRI 34247 = Schatz, G. E 


3346 (30), 3604 (42); Schlechter, R. 12549 (50), 12586* 
(40b); Schlieben, H.-J.E. 10921 (68), 8023 (43); Schnell, R. 
A. A. 1123 (50), 4093 o a (40b); Schoenmaker 117 
(24); Schweinfurth, G. A 40b); Scortechini 253* (48); 
Scott-Elliot, G. F AT ux ES ae 2171 (55); Seigler, 
D. S. 12802 (43); Service Forestier de Cote Ivoire 381 (13); 


RN 9512 (4) c: 
ao A. 88/29 (53); Simpson, N. 94.93 (67 
x, J. 4926 (27), 5125 a, F (69); Singh, 
3 (34): Smythies, S. 5909 ; Sohmer, S. 8671 (6 D 
Bul (65), 8769 (61), 8197 (67), 2 (67), 9884 (65), 10231 
(67), 10244 (67), 10288 (67), 10374 (17), 10461 (70), 10611 
(70); Sonké, B. 2687 (50), 2918 (66), 2919 (40b), 2920 (66), 
2921 (40b), 2923 (66), 2935 (66), Dd (66), 3044 (50), 3106 
(66), 3197 (66), 3251 (66), 3253 (66), a 3318 (6 


Ss 


(66), 4179 (66), 4486 (8). 4541 8): Syn 2 24* (63), 28 
(66), E (63); St. Clair- "DE G. 3. 50); St. John, 
29 (68); Staub, F. MAU 11986 > 11107 (68); 

Sue 186 (13), 353 (13), is 03 Suhaili Hj. Zinin BRUN 
a Pe aoe 
06 (7), FRI 17830 

8965 ae KEP 98992 (47); 
0 (57), KEP 31454 (57), s.n. (57), 


hun bin Bara KEP 33613 (48); Talbot, P. A. 3005 (50), 
3391* E Tan 8 (27; Tang 1170 (47) 1684 (34); 
Teijsmann, J. E. [also “Teysmann”] 10358 (27) 10732 
e). 19800 (27), 19428 (27); Theobald 2378 (17); Thollon, 
F.-R. 81 (40b); Thomas, D. 484 (66), 4398 (8), 6225 (40b), 
6856 a 8117 (8), 8649 a Thomas, N. 1033 (40b), 2945 
(50); Thwaites, G. H. K. CP 45* (65), CP 346* (25), CP 288 
(70), CP 346 (25), CP 363* (25), CP 440* (65), CP 457* be 
CP 2673* (59), CP 2991* (17); Tirvengadum, D. D. 11 (68 
19 (68), 28 (20), 379/29 (56), 387/37 = EM Pd 399 
e nc 947 (60), 969 (1 5); Toutain, B. 2408 (50); T 
34 (50); Troupin, G. 9165 (8), 9225 a 10931 (8): [s 

a i 58). 

Van dn pe 2 J. 4052 (34), 6062 (34); van 
Beusekom, C. F. 839 (64), s P v O i14 
(64), 2666 c van en Maesen 5454 (6 a n der 
13526 (43); Vanderyst, H. J. R. d 337 (3 7) van Dis ten, s 
M. 349 (13); van Meer, P. P. C. 110 (13), 121 Go) 1174 (21), 
1185 (8); van Nek, I. 569 (24), 2107 (54); Vasey 32 (53); 
Vaughan, R. E. 3031 (60), MAU 1084 (52), MAU 1636 (20), 


(20), MAU 12271 (60), MAU 12507 (20), MAU 13042 


— 


29), 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


MAU 13163 (15), MAU 13751 (56), d 13761 (56), MAU 
14123 (39), MAU 14191 e MAU 14222 (20); Vermeulen [J. 
or P., which is unknown] 794 ( Md C. 50 (5); Vigne, 
C. 1948 (13); Vigreux, M. 15373 (49b); iu J.-F. 247 (63); 
NS ^ 70* 2r 71* (50); Voorhoeve, 23 (13). 
, S. 52 (17), 888 (67), 1190 E e (17), 1335 
T 465 59). 1483 (70), 1511 a E a: b. a 
1799 Bu. fet oe e (59); Walker, F. S. KEP 338. 
(7); W: 8* (67), 102* co 212 An pt a 
vie P. ps A 8342* P 8354* (69), 8374* (34), 
8389* (27 D owe G. o a 0), 9 
E. G. 335 


— 


7 (57), FRI 15503 (57), FRI 
ee (D. un 15585 er Wiehe, P. O. 1712 (68), 2746 


Appendix 2. 


(68); Wieringa, J. 1654 (24), 4040 (50); Wilde, J. de 135 (5), 
572 (50), 637 (50), "UR 9 B (13), m (40b), 8827 
66), 11449 (40b); Wild 6 (5), 348 (5), 2917 (66), 
3945 (50); Wilks, W. ve y p n Wong, Y. K. ms 
(18), 1724 (34), FRI 28905 (48), FRI 28919 s FRI 28950 
lees FRI 30928 (18), FRI 32246 m m 50 (57), FRI 
323 2 Worthington, T. B. 1962 (70), oe (59), 3635 
e 83 (67); Wray, L. Jr. 25 (48), 1948* (48), 2283* (48), 
4122 c 5343* (18). 
s » 64422 n 

vato [no initial; Dorr, 1997] RN 8252 (32), RN 
e EN Zenker, c. 1392 (40b), 1763* (66), 1838 (66), 
2034 (66), 2252* (8). 2252a (8), 2393 (66), 3569 (66), 4207 
(66), 4419 (8), 4547 (66), 4760 (66); Zjhra, M. 133 (54), 392 
(41); Zollinger, H. 3051* (34). 


E 


Index to names treated in this revision. Names in boldface are accepted for species of Gaerinera Lam. 


Name 


Taxonomic identity 


Andersonia vaginaia Willd. ex Roem. & Schult. 
Chassalia clusiifolia var. P [beta] DC., nom. ined. 
C. coffeoides 

C. psychotri MN 

Coffea 


Fruciesca mauritiana DC. ex Meisn 


DC. 
chasalioides D. Dietr., nom. superfl. illeg. 


Gaerinera acuminata Benth. 
. acuminata var. mo Ridl. 
. acuminata var. es i Me Ridl. 
. aetheonoma Steud., nom 
. alata id ex Moa & A. P. Davis 
alstonii Maleom 
. ap hanodioica Maleomber 


. australiana C. T. 
. bambusifolia Malcomber & A. P. Davis 


belumutensis Maleomber 
bieleri (De Wild.) E. M. A. Petit 
bifida Bojer 


borneensis Valeton 
bracteata E. M. A. Petit 

bracteata var. glabrifolia E. M. A. Petit 
a lata Malcomber & A. P. Davis 
brevistylis R 

caerulea Bojer 

calycina Bojer 

calycina var. variegata Bojer 

capitata Bojer 

capitulata Malcomber 

cardiocarpa Boivin ex Baill. 

caudaia Ridl. 

chapelieri Drake 

co ERU. a var. 

Hutch. & M. B. Moss 

e Bojer 

crassifolia, orth. var. 

crinita Drake 


rr qi cU 
ix] 


= Gaerinera vaginata 


= G. junghuhniana 
= G. calycina 


= Psychotria L. sp. 


= G. psychotrioides 
= G. longivaginalis var. bracteata 
= G. longivaginalis var. bracteata 


= G. junghuhniana 
= Chassalia grandifolia DC. 


= G. calycina 


y 
= C. capitata DC. 


identity unclear 


grandifolia 


= G. crassiflora 
= G. phanerophlebia 


668 Annals of the 
Missouri Botanical Garden 


Appendix 2. Continued. 


Name Taxonomic identity 

G. cuneifolia Bojer 
G. cymiflora Bojer = C. boryana DC. 
G. s Malcomber & A. P. Davis 
G. dinklagei K. Sch — G. trachystyla 
G. divaricata: (Thwaites) Thwaites 
G. diversifolia Ridl. 
G. drakeana Aug. DC 
G. edentata Bojer 
G. eketensis Wernham 
G. ferruginea A. Che ta OS Schumach. & Thonn. (Verbenaceae) 
G. fissistipula (K. Scumi & K. Krause) E. M. A. Petit =Gb 
G. fractiflexa Beusekom 
G. furcellata (Baill. ex Vatke) Malcomber & A. P. Davis 
G. gabonensis Malcomber 
G. Xgardneri Thwaites 
G. globigera Beusekom 
G. godefroyana Cordem. = G. vaginata 
G. grisea Hook. f. ex C. B. Clarke 
G. guillotii (Hochr.) Bremek., hom. illeg. = G. inflexa 
G. guillotii Hochr 
G. hirtiflora Verde 
G. hispida . DC 
G. hongkongensis Seem. = [ chinensis Lam. 
G. humblotii Drake 
G. ianthina Malcomber 
G. incarnata Bojer = C. lanceolata (Poir.) A. Chev. subsp. lanceolata. 
G. indica J. F. Gmel. = Hiptage madablota Gaertn. (Malpighiaceae) 
G. inflexa Baill 
G. intermedia Ridl. = G. diversifolia 
G. junghuhniana Miq. 
G. kochummenii Malcomber 
G. koenigii (Arn.) Wight, nom. illeg. = G. vaginans 
G. koenigii var. divaricata (Thwaites) C. B. Clarke = G. divaricata 
G. koenigii var. oxyphylla (Benth.) C. B. Clarke, nom. illeg. = G. junghuhniana 
G. koenigii var. ihyrsiflora (Arn.) Thwaites, nom. illeg. — G. vaginans 
G. lacerata ined. — G. furcellata 
G. lanceolata Bouton ex A. DC. — G. pendula 
G. ie aia Ridl., hom. illeg. — G. diversifolia 
G. lasianthoides C. E. C. Fisch. = Psychoiria rhinocerotis Blume 
G. latifolia Ridl. = G. diversifolia 
G. laxiflora Cordem. = G. vaginata 
G. letouzeyi Malcomber 
G. leucothyrsa (K. Krause) E. M. A. Petit 
G. liberiensi A it 
G. longevaginalis, orth. var. = G. longivaginalis 
G. longiflora C. F. Gaertn. identity unclear 
G. longifolia Bojer 
G. longifolia var. pee Verdc. — G. longifolia 
G. longipetiolata R. D. Good = Psychoiria gossweileri E. M. A. Petit 
G. longivaginalis (Schweinf. ex Hiern) E. M. A. Petit 
G. longivaginalis var. bracteata (E. M. A. Mo Malcomber 
G. longivaginalis (Schweinf. ex Hiern) E. 

Petit var. longivaginalis 
G. longivaginalis var. louisii E. M. A. Petit — G. longivaginalis var. longivaginalis 


ryi Malcomber 


G. lushaiensis C. E. C. Fisch. — C. lushaiensis (C. E. C. Fisch.) C. E. C. Fisch. 


Volume 96, Number 4 Malcomber & Taylor 669 
2009 Revision of Gaertnera 


Appendix 2. Continued. 


Name Taxonomic identity 


macrobotrys Baker 

macrostipula Baker 

madagascariensis (Hook. f.) Maleomber & A. P. Davis 

microphylla Capuron ex Malcomber & A. P. Davis 

monstruosa Malcomber 

morindoides Baker — Morinda morindoides (Baker) Milne-Redh. 


oblanceolata King & Gamble 

oblanceolata Ridl., hom. illeg. — G. diversifolia 

oblanceolata var. densis (Ridl.) Beusekom — G. diversifolia 
vata Baker 


obovata Baker var. obovata 
obovata var. sphaerocarpa (Baker) Malcomber 


obtusifolia | Roxb. =H. oe (Roxb.) DC. (Malpighiaceae) 
occidentalis B — G. panicula 
ovata. Ridl. -G a 
oxycarpa Drake identity unclear 
oxyphylla Benth. = G. junghuhniana 
aie. Drake, hom. illeg =G perdia 
xyphylla var. dur d ddl — G. junghuhniana 
pangati Rheede etz. = Sphenoclea zeylanica Gaertn. (Lobeliaceae/Campanulaceae) 
paniculata Bent 
parviflora Bojer = G. psychotrioides 
parvipaniculata E. M. A. Petit — G. leucothyrsa 
pauciflora Malcomber & A. P. Davis 
pee Ridl. — G. ramosa. 
endula Bojer 
pow Verdc. = G. edentata 


phanerophlebia Baker 

phyllosepala Baker 

hyllostachya Baker 

plagiocalyx K. Schum. = G. longivaginalis var. longivaginalis 
pongatii, orth. var. = Sphenoclea zeylanica Gaertn. (Lobeliaceae/Campanulaceae) 
psychotrioides (DC.) Baker 

quadriseta A. DC. = G. psychotrioides 

quadriseta var. brevipes A. DC. = G. psychoirioides 

quadriseia var. hebepoda A. DC. = G. psychotrioides 

quadriseta var. petiolaris A. DC. = G. psychotrioides 

quadriseta var. platypoda A. DC. = G. psychotrioides 

racemosa (Cav.) Roxb. = H. benghalensis (L.) Kurz (Malpighiaceae) 


po uc qc c c uu um ne 
= [-] 


osa. Ridl. 
. raphaelii Malcomber 


G. rhodantha Baker = G. spicata 
G. richardii Drake = Psychotria sp. 
G. hated Ridl. = G. diversifolia 


a Thwaites ex Benth. 
je 


G vod ode B 

G. rufinervis Stap = Psychotria L. sp. 

G. salicifolia C. H. Wright ex Baker = G. trachystyla 

G. salicifolia Hutch. & Ge hom. illeg. = G. liberiensis 

G. schaizii Maleomber 

G. schizocalyx Bremek 

G. sessiliflora Ridl. = G. ramosa 

G. sieberi C. Presl — G. calycina 
Gaerinera sp. A Verde. probably = G. calycina 


G. spathacea Boivin ex Drake = G. arenaria 


670 Annals of the 
Missouri Botanical Garden 


Appendix 2. Continued. 


Name Taxonomic identity 
G. sphaerocarpa Baker = G. obovata var. sphaerocarpa 
E un K. Schum. 
ralensis (Pierre ex Pit.) Kerr 

à sonal (Hiern) E. M. A. Petit identity unclear 
G. taiensis Kerr — G. junghuhniana 
G. ternifolia Thwaites 
G. thyrsiflora (Arn.) Blume = G. vaginans 
G. S cum E. M. A. Petit 
G. thouarsii B identity unclear 
G. truncata rk oe = G. psychotrioides 
G. vaginans (DC.) M. 
G. vaginans subsp. junghuhniana (Miq.) Beusekom = G. junghuhniana 
G. vaginans subsp. junghuhniana f. hermaphroditica 

eusekom — G. aphanodioica 
G. vaginata Lam. 
G. viminea Hook. f. ex C. B. Clarke 
G. violascens Ridl. = Psychotria sp. 
G. walkeri (Arn.) Blume 
G. walkeri var. angustifolia = G. ternifolia 
G. walkeri var. gardneri oe C. B. Clarke = G. Xgardneri 
G. zimmermannii K. Krause ilg = Strychnos L. sp. (Loganiaceae) 
G. zollingeriana Mig. = G. junghuhniana 
Grumilea fissistipula K. Schum. & K. Krause = G. bieleri 

'ymenocnemis madagascariensis Hook. f. = G. madagascariensis 

Mussaenda borbonica Lapeyrère = G. vaginata 
Ophioxylon arboreum Koenig ex DC., nom. nud., pro syn. = G. vaginans 
Pristidia divaricata Thwaites = G. divaricata 
Psychotria bieleri De Wild. = G. bieleri 
P. furcellata Baill. ex Vatke = G. furcellata 
P. guillotii Hochr =G. i 
P. leucothyrsa K. Krause = G. leucothyrsa 
P. longivaginalis Schweinf. ex Hiern = G. longivaginalis 
P. obesa Wall., nom. nud. = G. obesa 
P. oxyphylla a Wall., nom. nud. = G. junghuhniana 
P. sralensis Pierre ex Pit. = G. sralensis 
P. stictophylla Hiern = G. stictophylla 
P. trachystyla Hiern = G. trachystyla 
P. vaginans D = G. vaginans 
P. viminea Wa IL, nom. nud — G. viminea 
Sykesia acuminata eth ones = G. junghuhniana 
S. arenaria (Baker) Kuntze = G. arenaria 
S. calycina (Bojer) Rimes = G. calycina 
S. crassiflora (Bojer) Kuntze = G. crassiflora 
S. cuneifolia (Bojer) Kuntze = G. cuneifolia 
S. edeniata (Bojer) Kunt: = G. edentata 
S. grisea (C. B. Clarke) Kuntze = G. grisea 
S. hongkongensis (Seem.) Kuntze = J. chinensis Lam. 
S. junghuhniana (Miq.) Kuntze = G. junghuhniana 
S. koenigii Arn., nom. superfl. illeg. — G. vaginans 
Ss a Kuntze, nom. nud. i unclear 
S. longifolia (Bojer) Kuntze . longifolia 
S. macroboirys (Baker) Kuntze = G. macrobotrys 
S. macrostipula (Baker) Kuntze = G. macrostipula 
S. obovata (Baker) Kuntze = G. obovata var. obovata 
S. oxyphylla (Benth.) Kuntze = G. junghuhniana 
S 


. paniculata (Benth.) Kuntze — G. paniculata 


Volume 96, Number 4 
2009 


Malcomber & Taylor 
Revision of Gaertnera 


Appendix 2. Continued. 


Name 


Taxonomic identity 


S. pendula (Bojer) Kuntze 
S. phanerophlebia (Baker) Kuntze 
S. phyllosepala (Baker) Kuntze 


ternifolia (Thwaites) Kuntze 
yrsiflora Arn. 
vaginans (DC.) Kuntze 
vaginata (Lam.) Kuntze 
. viminea (Hook. f. ex C. B. Clarke) Kuntze 


dau d qa di ue 4 
E SFY 


alkeri Arn 
colacao (Mig) K 
Diana ia hongkongensis Seem P. P. H. But, 
H. H. Hsue & P. T. L 


Uragoga sralensis Pierre ex Pit. 
U. stipulacea Kuntze 


rotundifolia 


ternifolia 


alkeri 
junghuhniana 


1 chinensis Lam. 


CRYPTIC DIOECY IN NYSSA 
YUNNANENSIS (NYSSACEAE), A 
CRITICALLY ENDANGERED 
SPECIES FROM TROPICAL 
EASTERN ASIA! 


Bao-Ling Sun,** Chang-Qin Zhang,” 
Porter P. Lowry I,” and Jun Wen? 


ABSTRACT 


Nyssa yunnanensis W. Q. Y 


known from only three small ue in a eyed forest area of southern Yunnan Provin 
nd the a with morphologically Mice flowers that produce both pollen 


of individuals occur, one 


N. Qin & Phengklai (Nyssaceae) is a critically a range-restricted tree species 


e, southwestern China. Two types 


and ee a an a. oe d and laboratory studies conducted between 2004 and 2007 
indicate, how that is functionally dios ious: polle n from the morphologically pe flowers is 
invi i fl ionall le. Field observations showed that the 


ipee system is supplem 
exhi 


ti 
bit parthenogenesis. The average sex ratio of individua 


15 days earlier than the QR dE female flowers but that fl 
cts wi 


rly 
ved and e dba té the flowers of N. yunnanensis, four 
ents Mi UR that the lee uera d 
N. yunnanensis is xenogamous and doe: 
BE three populations was in ale-bi en 


o 57. D. us the ratio among flowers was male- s ue 56: 3 ur flower um was higher in males. The 37 known 


trees of N. yu 
y human disturban 


nnanensis are Tikely the remnants of a 


a fate shared with m 


Androdioecy, critically endangered, 
blo 


Key words: 
yunnanensis, Nyssaceae, pollen viabilit 


aihen e A Southeast Asian 
dedicated conservatión EA informed by a o understanding of population structure and reproductive biolo 
z cryptic dioecy, floral phenology, IUCN Red List, 


despread, abundant ARD that has been heavily ea 
tax ntinued survival will requir 


8y- 
Nyssa, Nyssa 


The existence of dioecism in flowering plants has 
been universally acknowledged since Darwin (1877), 
yet dioeey is still poorly understood from both an 
ecologic and evolutionary point of view (Bawa & 
Opler, ade The reason for this may be the relatively 
ow proportion of dio taxa (Yamplosky 
Yamplosky, 1922), althoagh poa (1969) and Bawa 
and Opler (1975) reported a large number of dioecious 
tree species in tropical forest and dioecy appears to be 
favored in island environments (Baker & Cox, 1984; 
Sakai & Weller, 1999: Carpenter et al., 2003). So far, 


however, relatively few evolutionary studies have been 


conducted on tropical forest trees from eastern Asia 
(e.g., Ashton, 1969, 1977) 

The use of a functional perspective to study plant 
sexual systems has greatly advanced our understand- 
ing of sexual strategies in these organisms and has 


paved the way for testing relevant evolutionary 


Charlesworth, 1991; Sakai & Weller, 1999; Delph & 
Wolf, 2004; Dunthorn, 2004; Pannell, 2005). Dioe- 
cious species may vary in their sexual expression 
resulting in androdioecy, gynodioecy, trioecy, or 


! We are grateful to J. R. Pannell, S. A. Cunningham, S. Q. Huang, and J. W. Zhang for he aet comment on earlier versions 
F 


of this manuscript, X. K. Fan fo 


opy, L. 
L. Zhou for providing assistance and permission to work in the study populations, and F. Q. Shi 


Z.W ang f 


i and W. Tian for their support 


during our fieldwork. This study was funded by the National Natural Science Foundation of China (30770139, 30571137) and 


of Sciences 132 Lanhei Road, Kunming, Yunnan 650204, People’s 


Republic of China. The first and second authors contributed equally to the work reported here. Author for correspondence: 


zhangchangqin@ma ail.kib.ac.cn. 


Graduate School, Chinese Academy of Sciences, Beijing 100049, rn E P pam of China. 
LU 


* Missouri Botanical Garden, P.O. Box 299, St. 
5 Département Systématique et Evolution, Muséum Nat 
CEDEX 05, Franco. 


tional d'Histoire ERR 


Louis, Missouri 63166-02 


CP 39, 57 rue Cuvier, 75231 Paris 


oe of Botany, National Museum of Natural History, MRC166, Smithsonian Institution, Washington, D.C. 20013- 


7012, U.S.A. wenj@si.edu. 
i 10.3417/2008015 


ANN. Missouni Bor. Garp. 96: 672-684. PUBLISHED ON 30 DECEMBER 2009. 


Volume 96, Number 4 


Sun et al. 673 
Cryptic Dioecy in Nyssa yunnanensis 


subdioecy (Sakai & Weller, 1999). Mayer and 
Charlesworth (1991) defined cryptic or functional 
dioecy as a breeding system that has unisexual 
morphs, at least one of which appears to have perfect, 
hermaphrodite flowers. Flowers of such morphs in 
cryptically dioecious taxa retain nonfunctional organs, 
either as a gynoecium in functionally staminate 
flowers or an androecium in functionally pistillate 
ones. The sex ratio of functionally dioecious species is 
expected to be 1:1 in i e (Lloyd & Webb, 
1977), although a biased sex ratio is in fact often 
observed (e.g., Melampy & Howe, 1977; Mayer & 
Charlesworth, 1991; María & Ramón, 1995; Queen- 
borough et al., 
lyssa L. (Nyssaceae) comprises about 13 species, 
with four in North America, one in Costa Rica, one 
ranging from India to Indonesia, and seven in China 
(including six endemic species) (Qin & Phengklai, 
0 he genus exhibits a disjunct distribution 
between eastern Asia and North America (Eyde, 1966, 
1988; Wu & Fan, 1977; Fang et al., 1983; Wen & 
Stuessy, 1993; Wu et al., 2003) and has a rich fossil 
record in the Tertiary of the Northern Hemisphere 
(Eyde, 1963). The reproductive system of Nyssa has 
often been described to i e ry Mp Ud (e. a " 


y, 1993; Qin & n 2007, Pid it 
has not Les documented clearly for most species. 
Wangerin n presumed that Nyssa species were 
i. ed by wind and insects. Cipollini and Stiles 
1991) and Batra (1999) studied the cost of reproduc- 
tion and the behavior of native flower-visiting bees in 
the North American species N. sylvatica Marshall. 
Both studies indicated that N. sylvatica is dioecious 
and documented nat 


aquatica L., another North American species, is also 


ive bees as its pollinators. ue 
dioecious and may be pollinated by both insects and 
wind (Shea et al., 1993), but the pollination biology 
and breeding systems of Asian members of the genus 
are still poorly known. 

x H. N. Qin & 

Phengklai was originally (and iad) Nd b 

W n (1977) based 


Nyssa yunnanensis W. Q. Yin 


y 
u and F on tw ections, one 


bearing om staminate flowers and ls us in fruit. 
Wen and Stuessy (1993) reconstructed the phylogeny 
rs and 


of Nyssa using 18 morphological character: 
d N. 


— 


regarded N. yunnanensis an javanica (Blume 
Wangerin as closely related taxa with the two species 
possessing capitate male inflorescences. Because the 
morphology of N. yunnanensis was poorly document- 
ed, Wen and Stuessy (1993) treated it as belonging to 
the N. javanica complex and used N. javanica to 
represent this presumed lineage in their phylogenetic 
analysis in an attempt to reduce the impact of missing 
data. Based on comparisons of herbarium specimens 


and field observations, we found that N. yunnanensis 
differs from N. javanica in bearing thicker leathery 
leaves, but both appeared to be morphologically 
androdioecious, with structurally male and herma 
roditic flowers borne on different individuals coexist- 
ing in the same population (Sun & Zhang, 2007). 
Androdioecy is a rare breeding system in Mi run 
Darwin, 1977; B 1975; rae Ao 
worth, 1978; a & Beach, 1981; Ross, x 
Charlesworth, pus jure et i 1990; Fritsch & 
Rieseberg, 1992; Pannell, 2002), and many reported 


instances have proven to be functionally dioecious 


— 


when examined in detail, with the morphologically 
perfect flowers in fact being functionally female (e.g., 
Charlesworth, 1984; Anderson & Symon, 1989; 
1990; Mayer & Charlesworth, 


Schlessman et al., 


= 
2 


Nyssa yunnanensis is a canopy tree species whose 
range is located at the northern edge of the East Asian 
tropical zone. It is confined to mountainous bogs and 
marshes in southern Yunnan Province, China (Fu, 
1989), and is seriously threatened (Fu, 1992), p 
been s as Critically Endangered (CR) in 2004 
List criteria (IUCN, 200 
and ides for transfer from Rank III to Rank I 
protection in China (Wang & Xie, 2004). Understand- 
ing the reproductive biology of highly threatened 


based on 


species such as N. yunnanensis is important for 
developing conservation strategies, especially when 
they are known only from very few populations (see 
Falk & Holsinger, 1991; Bernardello et al., 1999). 

In this study, we documented the floral phenology, 
pollen morphology and germination, sex ratio, polli- 
nation biology, and mating system of Nyssa yunna- 
nensis to address the following questions: (1) Is N. 
yunnanensis functionally dioecious? (2) If so, what is 
its sex ratio? (3) Is it pollinated by wind, insects, or 

oth? (4) Can any relationships be detected among the 
breeding system, sex ratio, and the rarity of N. 
yunnanensis? 


MATERIAL AND METHODS 
STUDY SPECIES 


Nyssa yunnanensis is a critically endangered 
wetland tree up to 30 m in height. It has small 
green-yellow sessile flowers (ca. 2 mm long) arranged 
wide (Sun & Z 
2007). Its flowers, which are of two types, structurally 


in capitulae less than 5 mm hang, 
male and perfect, secrete abundant nectar with an 
applelike fragrance from the surface of the ovary disc. 
The single-seeded drupes have a low germination rate, 
and no seedlings have been observed in natural 


populations (Sun et al., 2007). Nyssa yunnanensis 


Annals of the 
Missouri Botanical Garden 


often grows in association with species of Quercus L., 
Juglans L., Persea Mill., and Laurus L., particularly in 
riparian vegetation along rivers (Fang et al., 198. 


STUDY SITE 


ME Rape is known from only three 
pulat wen Tropical Forest Station 
(or? 06! T. 99° 35 Nji in southern Yunnan, one along 
a stream at 842 m (hereafter referred to as the Stream 
population}, one around a well at 837 m (the Well 
population}, and the third at ca. 900 m adjacent to an 
old dam (the Dam population). The size of these three 
with 14, 9, and 14 individuals, 


respectively. Field eres were conducted 


populations is small, 


from 2004 to 2007. Experiments were carried out 
continuously on two trees bearing staminate flowers 
and three trees with morphologically perfect flowers 
(all others were so tall, reaching to almost 30 m 
height, as to make it difficult or impossible to access 
flowers on a regular basis). In order to assess sex 
ratios, we did, however, have local people climb all 
trees in each of the three populations to collect 
specimens 


FLORAL PHENOLOGY 


Floral phenology was studied in the wild popula- 
tions of Nyssa yunnanensis in 2006. Two tagged male 


trees were monitored to record information of 
phenological phases, such as morphological changes, 
flowering period, flower life span, anther dehiscence, 
nectar production, odor, and fruiting. Data were 
recorded from a total of 40 fertile branches bearing 
197 inflorescences and a total of 2811 flowers. The 
same types of information were collected from three 
trees bearing morphologically perfect flowers with a 
total of 40 branches bearing 82 inflorescences and 
590 flowers. Flower longevity was assessed by daily 
observation of 30 flowers randomly sampled and 
tagged before anthesis. Information on other stages of 
flowering and fruiting was obtained by weekly 


observations. 


POLLEN VIABILITY AND STIGMATIC RECEPTIVITY 


Pollen from both male and morphologically e 


flowers was examined with a light microscope 
(KYKY- 


and a scanning electron microscope odi. 
Science Apparatus Co. of the Chinese Academy of 
Sciences, Beijing, China) Pollen viability was 


estimated in vitro by d pollen germination in 
a gradient series of sucrose solutions (0.5%, 1%, 2%, 
5%, and 10%) with distilled water as the control (Hu, 
1993). Mature anthers were identified by color 


change, 10 of which were collected from male and 
from morphologically perfect flowers, respectively, 
light microscope once per 
times. The 
inflorescences were collected, taken to the lab, and 


and examined under t 
hour, with each experiment repeated five 


placed in a culture dish containing water to test pollen 
longevity. Observations were made every three hours. 
The data were analyzed using the software package 
SPSS 11.0 (SPSS Ine. Chicago, Illinois, U.S.A.). 
Stigmatic peas was checked by examining 
changes in an and then verified 


by ee (DAB) Dain 1992. 


SEX RATIO WITHIN POPULATION AND AMONG FLOWERS 


The within-population sex ratio of individual trees 
was determined for all 37 individuals at the three 
known sites based on observations made over the 
three years from 2005 to 2007. For each tree, a visual 
inspection of 30 to 100 flowers from approximately 
ive inflorescences allowed us to determine its sexual 
status. 

In order to determine whether Nyssa yunnanensis is 
functionally dioecious, consideration had to be given 
i rpholog- 
ically perfect flowers (the latter potentially being 


to the relative number of staminate and mo 


functionally female) in the populations (see Opler & 
Bawa, 1978; Webb & Lloyd, 1980). We therefore 
estimated the total within-population ratio o 
turally staminate to perfect flowers (hereafter id 
to as the population flower ratio [PFR]) according to a 
modified D of the method proposed by Opler and 
Bawa (1978). The PFR was calculated by E 
the ratios of i staminate and perfec 
flowers per inflorescence (the flowers per n 
cence ratio [FR]) by the ratios of inflorescences 
bearing structurally stam 
branch (the inflorescence ratio 
multiplying this in turn by the ratios of structurally 
staminate and perfect branches per tree (the branch 
ratio [BR]) and by the A sex ratios (PR). The 
h can be ressed by the 
FR x IR. X BR X PR, 


provides an estimate of the ratio of total number of 


resultant product, whic 
following formula, PFR = 


staminate to total idem perfect flowers in 
the three known populations of N. yunnanensis. For 
our estimate of the ratio of flowers per inflorescence, 
all 37 three 
populations and randomly selected 50 staminate 


we examined individuals in the 


staminate flowers and 40 branches from individuals 
with morphologically perfect flowers to determine the 
IR, and finally five trees with staminate flowers and 
five with morphologically perfect flowers to caleulate 


Volume 96, Number 4 
2009 


Sun et al. 675 
Cryptic Dioecy in Nyssa yunnanensis 


the BR. Sex ratios are consistently presented as 
male:female, with the second term always given as 
unity (1) such that ratios in which the first term is less 
than one indicate a female-biased sex ratio and those 
with the first term greater than one indicate a male- 
biased ratio. 


FLORAL VISITORS 


Floral visitors were recorded in the field at 
flowering time between 07:00 and 23:00 during 11 
days between 15 and 25 March 2006, for a total of 
176 hours of observation. Pollinator behavior and 
movement between the male and the morphologically 


= 
2 
un 
o 
a 
= 
< 
a 
un 
> 
e 
5 
un 
z 
o 
bed 
o 
o 
t5 

o] 
= 
= 
E 
o 
£ 
e 
= 
[em 
Ex 
= 
e 
E 

ga 
> 
= 
= 
© 
o 


laboratory of the Puwen Tropical Forest Station for 


ry 

identification and for further examination under the 
light microscope to detect any presence of pollen. 
Voucher insect specimens were deposited at the 


Kunming Institute of Botany. 


MATING SYSTEM 


To test whether individuals with morphologically 
n and were 
capable of bot 
experimental protocol was used during the 2006 and 
2007 flowering seasons involving a control and seven 
treatments, as follows: flowers were (1) untreated to 
serve as the control; (2) 

anthers on morpholo 


semis of self-pollination in the ele of pollina- 


ged to assess whether 
ce perfect flowers are 
tors; (3) emasculated and bagged to investigate 
possible parthenogenesis; (4) emasculated and artifi- 
i self- 
compatibility; (5) emasculated and artificially polli- 


cially self-pollinated to test for intra-flower 
nated with pollen from another morphologically 
perfect flower from the same individual to test for 
inter-flower self-compatibility; (6) emasculated and 
artificially | cross-pollinated with pollen from a male 
flower to test xenogamy and to assess whether gene 
flow is possible between the plants with staminate 
flowers and those with morphologically perfect 
wind 
pollination; and (8) directly emasculated to test the 


s; (7) emasculated and netted to test 


function of pollen on the morphologically perfect 
flower by comparison with the results of the control 
group. Bags of MF tracing paper p ely 20 X 

2 on fiber), tested 


to make sure they did not allow e penetration of 


O cm in size and reinforced with n 


airborne pollen, were placed on immature inflores- 
cences at the beginning of March and removed in mid- 
had withered. Artificial 
pollination was conducted by directly brushing the 


pril when all flowers 


stigmas of the recipient flowers with stamens from the 
donor. As indicated above, because many trees are too 
tall to carry out the various treatments, just three 
individuals with morphologically perfect flowers were 
used in 2006 and one in 2007 as the female parent, 
and two individuals were used as pollen donors. In 
general, we chose flowers from the highest part of 
these trees to serve as controls; flowers lower down 
were used to carry out the other treatments because 
they were easier to access and manipulate for artificial 
pollination. It should be noted that light intensity 
increases toward the top of these large trees. Mature 
fruits were collected from the control and the treated 
flowers in August 


RESULTS 
FLOWER PHENOLOGY 


Anthesis in Nyssa yunnanensis extended from 
February to April (Fig. 1). Staminate flowers opened 
10 to 15 days earlier than the morphologically perfect 
ones, but flowering ceased at nearly the same time. 
The staminate flowers had five to seven petals, 10 to 
14 stamens, and a central disc that secreted abundant 
nectar (Fig. 2A). The morphologically perfect flowers 

ad four to six petals, five to seven stamens, a nectar 
disc that also produced abundant nectar, and a bifid 
B). taminate and the 


style (Fig The stamens in the s 


morphologically perfect flowers were similar in 


wo a whorls with 
g. 3). Growth and the 


order of dehiscence of the id on the longer 


orming t 
cue P different length (F 


stamens were recorded on staminate flowers with five 


petals and 10 stamens. The order of dehiscence 


a 


shown on the left of Fig. 3) was sequentially from 
anthers number 1 to 5. By contrast, the short anthers 
in the staminate flowers did not develop and dehisce 
in a regular order. The stamens of the morphologically 
perfect flowers are equal in length, although they too 
can be divided into two groups. The first group 
includes most of the stamens, which develop early and 
are inserted around the disc. The second group, 
comprising just one or two stamens, is positioned 
directly on the disc and develops well after the flower 
opens. We refer to members of this second group as 
anaphase stamens (Fig. 3, right). 

The life span of staminate flowers was 10 to 14 
days, whereas that of the morphologically perfect 
flowers was age nid from 11 to 13 days. The 
nate flowers 
after the 
opening of the flower; those on the short stamens 


anthers o in sta 


ong stamens mi 
dehisced m pde) pollen four days 


676 Annals of the 
Missouri Botanical Garden 


—@— Staminate flowers © —&— Morphological perfect flowers 


m" 
> A oo e 
o o oO © 

1 


N 
© 


Flowering percentage (%) 


© 


25 Feb. ] Mar. 5 Mar. 10 Mar. 15 Mar. 20Mar. 25Mar 30 Mar. 5 Apr. 


Figure 1. The flowering period of Nyssa yunnanensis W. Q. Yin ex H. N. Qin & Phengklai in the studied population 


in 2006 


Figure 2. SEM images of flowers and pollen of bon yunnanensis sis W. Q. Yin ex H. N. Qin & Phengklai. —A. Sta 
flowers. —B. Morgkolosicaliy = flowers. —C. Pollen grain of staminate fine. —D. Pollen grain of don decal 
perfect flower. Scale bars: A, B = 1 mm; C; D = 10 um 


Volume 96, Number 4 
2009 


Cryptic Dioecy in Nyssa yunnanensis 


stamen 


anaphase stamen 


Figure Flower structure of Nyssa yunnanensis W. Q. Yin ex H. N. Qin & Phengklai, showi 
morphologically perfect flower (right) (the actual number of petals varies from five t in tl 


the morphologically perfect flower). 


released pollen one or two days later, after which the 
filaments of the short stamens elongated to reach the 
length of the long stamens. On the first day when the 
morphologically perfect flowers opened, the style was 
green and undivided, but after three or four days it 
began to become evidently bifid and the color 
changed from green to white, a process that took 
about five or six days to complete. The anthers of the 
morphologically perfect flowers dehisced seven or 
eight days after the opening of the flower. Nectar 
secretion began after the flower had been open two or 
three days and ceased when the stamens and style 
became brown. 


100% p 


90% + 


A 


80% - 


70% 


60% 


Germination percentage 


50% 


OSA 


E 
r 
B 


4096 ; 3 
0.096 0.596 


1.096 


(left) and 


and four to six in 


POLLEN VIABILITY, STICMATIC RECEPTIVITY, AND FLORAL 
SEXUAL FUNCTION 


The diameter of the pollen grains from staminate 
flowers ranged from 25-30 um. They were tricolporate 
and triangular in polar view, with scabrate and 
reticulate sculpturing (Fig. 2C). Morphologically per- 
fect flowers produced only inaperturate pollen grains, 
which were circular in polar view (Fig. 2D). The 
pollen from staminate flowers germinated in a gradient 
series of sucrose solutions ranging from 0%-10% 
(Fig. 4), with the 0.596 sucrose solution yielding the 
best result (up to 97.4%, F5 [Levene F statistic of the 


2.0% 


5.0% 


Different concentrations of sucrose solution 


Figure 4. Pollen i 
sucrose solution and different culture times. 


f Nyssa yunnanensis W. Q. Yin ex H. N. Qin & Phengklai in diffi 


T 
= 


Annals of the 
Missouri Botanical Garden 


Table 1. 


Sex ratio and flower sex ratio in three populations of Nyssa yunnanensis. 


Sex ratio (PR) 


Flower sex ratio (PFR) 


SP WP DP FR IR BR 
(M:H:J) (M:H:J) (M:H:J) (n = 50) (n = 40) (n = 5) 
atio 6:6:2 0:4:5 2:4:8 1.94:1 2.39:1 0.97:1 
— of three populations 0.57:1 2.56:1 


R, branches per tree; DP, Dam population; FR, flowers 
nerloloicall perce IR, e po pu 


noes jane ela morphologically perfect); J, juvenile; 


per inflorescence ratios C Ru wu E 
; M, sta 


PF 
population; SP, Stream population; WP, Well ped. ag 


five samples] = 1.654, P < 0.05, using distilled water 
as the control). The pollen germination rate increased 
more or less regularly with increased cultivation time, 
although a slight change occurred after three hours. 
By contrast, i 


no pollen from the morphologically 


perfect flowers germinated, regardless of sucrose 
concentration and culture time, indicating that these 
flowers are functionally female 

In ii phere ene perfect flowers, as indicated 
above, the styl 


from green to white and then turned brown following 


ecame receptive after changing color 
receptivity. The stigma was receptive before dehis- 

e of the first group of anthers, so the morpholog 
ically perfect (and functionally female) flowers were 


thus protogynous. 


SEX RATIO WITHIN POPULATIONS AND AMONG FLOWERS 


Three populations of Nyssa yunnanensis are present 
at the Puwen Tropical Forest 
population, the Well 
population. When juvenile individuals were excluded, 


Station: the Stream 
population, and the Dam 
the ratio of adult male to female trees was 1:1 in the 
Stream population (six male, six female, and two 
juvenile trees), 0:1 in the Well population (zero males, 
four females, and five M and 1:2 in the Dam 
(two male males, and eight 
ble 1). When ^ 37 Pal of N. 


yunnanensis known from the study area were consid- 


population 
juveniles) (Ta 
ered, the overall sex ratio was 0.57:1, showing a clear 
female bias. 


The FR = 


inflorescence 


1.94:1 (number of flowers per male 
14.66 + 3.27, n = 50, P = 0.05; 
number of flowers per female inflorescence = 7.56 + 
1.34, n = 50, P = 0.01). The IR = 2.39: x atera! of 
inflorescences per male branch = 4.90 
40, P = 0. "s number of ""— per female 
branch = 2.05 1.01, n = 40, P = 0.01). The BR = 
0.97:1 mae of branches per male tree = 18.80 + 
2.86, n = 5, P = 0.01; number of branches per female 
tree = 19.40 — 2.51, n = 5, P = 0.05). ee the 
overall PFR = 1.94:1 X 2.39:1 X 0.97:1 X 
2.56:1, which suggests a male-biased c sex ratio. 


* 


perfect) (PFR = FR X IR X BR X PR); PR sex ai of eee in 


FLORAL VISITORS 


Thirty-six insect species belonging to five families 
and 25 genera were collected as visitors of Nyssa 
yunnanensis flowers. Twenty species were observed on 
male and/or functionally female flowers (Appendix 1). 
Because male flowers opened before female flowers, 
insects were at first observed and captured visiting the 
former. Most insects visited flowers exclusively for 
nectar, and only four species were observed to carry 
pollen on their body and/or legs and could therefore 
be considered are pollinators: Chrysomyia 
megacephala Fabr. (a fly), Apis cerana Fabr. (a bee), 

sp., and Praestochrysis Linsenmaier sp. 
(both chalcidoids). Pollen grains were also noted on 
the bodies of Polistes Latreille sp. and Mutilla 
marginata Baer, but these species only visited flowers 


occasionally and were thus not considered to be 
effective Dens Other visiting insects were 
served, including syrphids, beetles, mosquitoes, 


and small MIR all apparently for nectar, but they 
neither carried pollen nor remained in the population 
of Nyssa for more than a short time, suggesting that 
they played little or no role in pollination. 


MATING SYSTEMS 


The results of the breeding experiments (Table 2) 
showed that fruit set from untreated s was higher 
than that from the other treatments in both of the years 
studied (49% in 2006 and 45% in 2007). Fruit set was 
zero for the following treatments: directly bagged, 
emasculated and bagged, emasculated and artificially 
self-pollinated with pollen from the same flower, and 
emasculated and artificially self-pollinated with 
pollen from another functionally female flower on 
the same tree. In functionally female flowers that were 
emasculated and artificially cross-pollinated with 
pollen from a staminate flower, fruit set at a rate only 
slightly lower than that in untreated flowers (47% in 
2006 and 43% in 2007). Fruit set in functionally 
emale flowers that were emasculated and netted was 


only 13% in 2006 and 15% in 2007, suggesting that 


Volume 96, Number 4 


Sun et a 


2009 Cryptic Dioecy in Nyssa yunnanensis 
Tabl Number of flowers and percentage fruit set in control and treated flowers of Nyssa yunnanensis over a two-year 
period (2006-2007). 
2006 2007 
Treatment No. of flowers Fruit set (%) No. of flowers Fruit set (%) 
Untreated (control) 49 49 71 45 
Directly bagged 27 0 24 0 
ma 2 nd bagged 13 0 32 0 
Emasculated and atificialy self- 
pollinated with pollen from the same 
ower 21 0 26 0 
Emasculated and artificially self- 
pollinated with pollen from another 
= lo; ud ically pm flower on the 
sal 29 0 34 0 
B e and artificially cross- 
pollinated with pollen from a 
aminate flower 43 16 63 19 
Emasculated and netted 15 13 46 15 
Emasculated 21 47 28 43 


some wind pollination can take place but that it is 
much less efficient than insect pollination. However, 
comparing fruit set between untreated (control) 
that 
artificially cross-pollinated with pollen from male 

owers, we found that the latter had lower fruit set 


(16% vs. 49% in 2006 and 19% vs. 45% in 2007). 


flowers and those were emasculated and 


DISCUSSION 
FUNCTIONAL OR CRYPTIC DIOECY 


Barrett. (2002) pointed out that it is important to 
consider the quantitative nature of gender and adopt 
functional rather than morphological criteria when 
ena plant sexuality. Our study shows that 

a yunnanensis is characterized by two e of 
individuals, those producing staminate flowers and 
those wit ally perfect flowers that 
produce sterile, inaperturate Es Such apparently 


rphologic 
androdioecious 


families were recognized by Mayer and Charlesworth 
(1991) to exhibit functional or cryptic a and 
include num 

ndiculatum a 
Specht (Levine & Anderson, 1986; Anderson & 
Symon, 1989). 
Why do the cnr kd perfect flowers of 
Nyssa yunnanensis r fiv se ns that 
produce inviable pollen? ae to attract and 


ven stam 


deceive pollinators, as suggested by the fact that the 
anthers of these flowers are otherwise similar in 


overall morphology to those of staminate flowers, a 


situation also reported in s: x^ Caledonian species 
pancheri — (Baill.) (Araliaceae) 
1990). Real ean reported that 


bees will visit artificial flowers that offer no rewa 


Polyscias 
Schlessman et al., 


— 


rd of 
any kind, and this type of visual deception can be 
important to the success of some mating systems 
(Dafni, 1984). Batra (1999) thought the bright yellow 
anthers of N. sylvatica may be attractive to pollinators, 
and our results show that fruit set is higher in 
untreated flowers of N. yunnanensis than in emascu- 
lated flowers, which supports this hypothesis. Fur- 
thermore, the anthers of the functionally female 
flowers may also provide a landing structure for 
pollinators in species whose flowers have very small 

| Thus, FO ie 
inaperturate pollen in functionally female flowers is 


als, as in yunnanensis. 
not viable, the presence of anthers that produce pollen 
(albeit sterile) may nevertheless be important for the 
reproductive success of the species. 

r observations indicate that Nyssa yunnanensis is 
pollinated by generalist visitors such as bees and flies, 
and that both nectar and pollen appear to serve as 
rewards for them. Dioecy may be favored in forest 
species if pollinators are primarily o and if 
there is increased selection dio e for se 
tion (Beach & Bawa, 1980). Batra (1999) RAN that 
pollinators were attracted to ns small, greenish, and 
odorless flowers of N. sylvatica by the sparkling nectar 
in addition to the pollen. Our field observations 
showed that the nectar of N. yunnanensis is fragrant, 
with an applelike smell, and that it serves as a reward 
for pollinators as well as other floral visitors. Hence, 
N. yunnanensis can be considered primarily a nectar- 


Annals of the 
Missouri Botanical Garden 


rewarding species in which pollination relies on a 
system of unidirectional exploitation (Dafni, 1984), 
with pollen serving as a secondary rewar 


PHENOLOGY 


Bawa and Opler (1975) reported on the phenolog- 
ical patterns of dioecious trees in tropical forests and 
contrasted them with those of hermaphroditic taxa. A 
small number of dioecious species flowered during the 
dry season (which occurs from January to March in 
Yunnan), whereas the majority flowered at the dry-wet 
interface (April and May). Nyssa yunnanensis fits this 
pattern well. At the population level, male flowers 
opened earlier than functionally female flowers but 
simultaneously. We 
all number 


attract a few early pollinators 


completed flowering almost 
suggest that the precocious opening of a sm 
of male flowers serves to 
ld de serve in turn to guide or attract others by 
e of their memory and learning ability (Cartwright 
& xs 1983), which may facilitate long-distance 
cross-pollination. Moreover, the 2-whorled androeci- 
um and the sequential dehiscence of anthers in male 
flowers of N. yunnanensis may prolong the functional 
life span of an individual flower and thereby increase 
the efficiency of out-crossing. It is possible, however, 
that pollen discounting may also be taking place 
(Harder & Wilson, 1998) because the sequence of 
anther maturation and dehiscence is such that the 
distance between a dehiscing anther and the one p» 
will reach maturity next is maximized, while 
length of time during which an anther bears ies is 
shortened. This may be an effective way of making the 
best use of a limited resource for the plant. 
nnanensis, the number of flowers in an 


In Nyssa 
inflorescence ween males and females. 
Anther size also differs strikingly; anthers of the male 
flowers are ca. 5 X 3 mm, whereas those of the female 


X 0.2 mm (Sun & Zhang, 2007). 


When differences in the total numbers of male and 


flowers are ca. 0.5 


female flowers within an inflorescence and on an 
entire plant are also taken into consideration, it is 
clear that male individuals produce more pollen 
grains than female plants. 


MATING SYSTEM 


Because male flowers produce many more pollen 
grains than functionally female flowers and pollen 
from the latt 


regarded as xenogamous. Our experiments showed 


ter is sterile, Nyssa yunnanensis can be 
that fruit set was higher in untreated (control) flowers 
than in flowers subjected to the various treatments in 
our experimental protocol, especially when female 
flowers were emasculated and artificially pollinated 


using stamens from male trees. This result may be 

explained by differences in the position of the flowers 

selected for the treatments. In general, flowers located 
d 


8 
on a tree were left as controls and those lower 
down were treated because they were physically 
easier to access and manipulate. Because flowers 
positioned higher on a tree almost certainly received 
ore light, this may have affected fruit set (Cipollini 
& Stiles, 1991). 


could, however, be ruled out in N. yunnanensis because 


Parthenocarpie fruit. development 


no fruit was produced by flowers that were emasculated 
and bagged. Also, flowers that were emasculated and 
artificially pollinated from male individuals had lower 
fruit set than untreated flowers, possibly indicating out- 
crossing depression (see Fischer & Matthies, 1997), 
although this should be tested further to ensure that no 
errors occurred during the process of artificial 
pollination. Lastly, Whitehead (1969) considered that 
wind pollination is probably uncommon or even absent 
Our study suggests, 
however, that both wind and insect M occur 


among tropical forest taxa. 


n N. yunnanensis because fruits were produced in 
treatments where flowers were emasculated and netted 


and fruit set was higher in emasculated flowers that 
were not netted. This may be related to the fact that 
adult trees of N. yunnanensis are very tall, extending 
above the denser parts of the forest canopy where wind 
speed is not significantly reduce 


SEX RATIO WITHIN POPULATION AND AMONG FLOWERS 


dioecious plant species, male and female 
individuals often show secondary intersexual differ- 
ences that can be related to the differential constraints 
and selection pressures imposed on male and female 
functions (Lloyd & Webb, 1977). Several studies have 
shown that a 1:1 primary sex ratio often occurs in 
dioecious species (Fisher, 1930), and deviations from 
this have been the focus of many theoretical and 
empirical ma (e.g., Melampy & Howe, 1977; Shea 
l., 1993; Marí 
00 


al., 7). Our results showed a female- 


eta 


ratio of 0.57:1 in Nyssa yunnanensis, adding to only a 
few oe reported cases in dioecious species 
mpy & Howe, 197 pler & Bawa, 1978; 

1). Opler and per (1978) study of 


tropical forest trees indicated that a female-biased sex 


as 199 


ratio tends to occur in species whose population 
densities are high. However, the ratio of the total 
number of male to female flowers in N. yunnanensis 
shows a different pattern, with a male bias of 2.56:1. 
This may be explained by the fact that about twice as 
many flowers are produced per inflorescence on male 
trees, which also bear more than twice as many 
inflorescences per branch compared to female trees, 


Volume 96, Number 4 
2009 


Sun et al. 681 
Cryptic Dioecy in Nyssa yunnanensis 


although the branch per tree ratio and the overall sex 
ratio are both less than 1:1. Differential resource 
allocation to reproductive function between males and 
f es may translate into differences in trade-offs 
between vegetative and reproductive activities (Maria 
& Ramón, 1995). Males of N. yunnanensis produce 
ore flowers than females, indirectly supporting the 
aes that female plants incur a higher cost of 
sexual reproduction e that this higher cost is 
measurable as reduced vegetative growth (male trees 
bear fewer branches than females) and p flowering 
frequency (also see Cipollini & Stiles, 1991). 

One of the populations at the Puwen Tropical 
Forestry Station, the Well population, comprises only 
functionally female individuals an cks male 
individuals altogether, yet no parthenogenesis was 
detected. Chase et al. (1996) reported that the longest 
intrapopulational gene-flow distance covered for 
tropical trees was approximately 350 m. If this is true 
for Nyssa yunnanensis then gene transmission must 
have taken place between the Stream population and 
the Well population, which are about 100 m from one 
another, although we cannot rule out the possibility 
that some pollen may also be transported from the 
Dam population located about 2.5 km away. 


CONSERVATION IMPLICATIONS 


Rarity in plants may be the result of intrinsic or 
extrinsic (environmental) factors and is mainly related 
to reproduction (Rabinowitz, 1981; Falk & Holsinger, 
1991). Ashton (1977) indicated that apomixes or gene 
fixation in small populations of self-compatible 
individuals must be regarded as an evolutionary dead 
end and would prelude inevitable extinction in a 
continuously changing biotic environment. However, 
in our study, despite the very small population sizes 
observed in Nyssa yunnanensis, we found no evidence 
of apomixes. 

Biological diversity is especially high in the humid 
tropics, but human interference in n tropical ecosystems 


biodiversity (Tumer et al., 1994). The highly restricted 
populations of Nyssa yunnanensis appear to have long 
suffered from various types of human disturbance such 
as the building of a dam to supply the water for people 
around the Puwen Tropical Forest Station and the 
conversion of large areas of forest to cultivate 
economically valuable species such as rubber, banan- 
as, and pineapples. Moreover, although the species has 
been reported as valuable for construction, furniture, 
decoration, and as an ornamental for landscaping (Wu 
& Fan, 1977; Song et al., 1989), the local people regard 


its wood as too soft to be of much use, even as fuel, so 


they simply fell the trees when clearing for agriculture. 
This suggests that the small population size of N. 
yunnanensis seen today may be relictual, i.e., that the 
37 trees we studied may be all that remain o a once 
more widely distributed and abundant species whose 
range has been severely reduced as a result of human 
disturbance. If this is the case, the bi 

g the remaining trees may exacerbate and furt 


iased sex ratio 
her 
UT the extrinsic threats faced by N. yunnanen- 
sis, regardless of whether the observed sex ratio has 
always been a characteristic of the species or has 
resulted directly or indirectly from human-caused 
reduction in population size. 

The long-term survival of critically endangered 
species such as Nyssa yunnanensis and hundreds of 

er Yunnan Province (Gong et al., 
elsewhere in China (Fu, 1989, 1992), a 
throughout eastern. Asia (IUCN, 2008), is uo 
dependent on developing and implementing sound in 


others in 


situ conservation policies (supplemented when nec- 
sary by ex situ measures such as seed banks and 
cuisine in botanical gardens) coupled with respon- 
sible management of natural resources and promoting 
improved environmental education and publie aware- 
ness campaigns. In the case of N. yunnanensis, 
germination and recruitment in the three populations 
at the Puwen Tropical Forest Station appear to be 
hampered by an unnaturally dense herbaceous layer 
that prevents sufficient light i 
ground. In an effort to ensure the long- 
of this rare local endemic, more than 200 individuals 
e now being grown at Puwen and at the Kunming 
Institute of Botany, which could be us 
coordinated program involving in situ reintroduction, 
reinforced local protection, and a simple program to 
monitor its survival and reproduction, coupled with ex 
situ efforts to maintain genetic diversity and promote 
the cultivation of trees in appropriate reserves an 
botanical gardens. As the present study shows, 
however, such efforts must be informed by research 
on the population structure and reproductive biology 
of individual species, which may also be of great 
importance and may become even more so as the 
impacts of projected global and regional climate 
change begin to alter many complex and often 
interrelated ecosystem attributes such as flowering 
and fruiting phenology, pollinator behavior, and 
microclimate. 


Literature Cited 


Anderson, G. J. & D. E. Symon. 1989. Functional dioecy and 
4-219. 


Ashton, P. S. 1969. Bn among tropical forest trees: 
Some deductions in the of recent evidence. Biol. J 
Linn. Soc. 1: 155-196. 


Annals of the 
Missouri Botanical Garden 


A ogee of rain forest research to 


er ie theo Ann. Missouri Bot. Gard. 64: 
—105. 

Baker, H. G. & P. A. Cox. 1984. Further thoughts on 
dioecism and islands. Ann. Missouri Bot. Gar 
244—253. 

Barrett, S. C. H. 


2002. The gn of plant sexual 
diversity. Nat. Rev. Genet. 3: 274—284. 
Batra, S. W. T. 1999. Native bees in ure Apoidea) in 
1 a sylvatica Marsh. (Cornaceae). Proc. 
d. 


Bawa, K. S. 1980. Evolution ka dioecy in flowering plants. 
Annual Rev. An Syst. 5-39. 
& P. A. Opler. s Dioecism in tropical forest 
trees. Evolution 29: 167-179. 
& J. H. Beach. 1981. Evolution of sexual systems in 
flow: un on Ann. Missouri Bot. Gard. 68: 254—274. 
Bose. s K. S. Baw: a o: Role of pollinators in the 
from distyly. Evolution 34: 1138-1142. 
d ur G. G., G. J. Anderson, L. S. Patricio, M. A. 
Cleland, T. F. Stuessy D. Crawford. 1999. 
nun biology of p fernandeiane (Lactor- 
r. J. 7 ys 829- ih 
& T. Jaffr . 2008. Reproductive 
traits of tropical rain- E trees in New Caledonia. 
Trop. Ecol. 19: 351—365. 
Cartwright, B. A. & T. S. Collett. 1983. Landmark learning in 
bees. J. Co mp. Physiol. 151: 521 343. 
as B. & D. Charlesworth. 1978. A model for the 
evolution of dioecy and gynodioecy. Amer. Naturalist 112: 
975-997. 
Mi D. 1984. Androdioecy and the evolution of 
y. E 


+ 


J. Linn. Soc. 23: re 


ncy and nutrient fl 
oe ae deception in pollination. Ann. 
Rev. Ecol a 15: 259-278. 
ni pue A Practical Approach. 
Oxfo rd os Press, New York. 
Darwin, C. 1877. The Different Forms of Flowers on Plants of 
the Same Species. Signy New Yor E 


es 
morphic breeding 


Dunthorn, M. 2004. Uer doriy in Mammea (Clusiaceae). 
Plant Syst. Evol. 249: 

Eyde, R. H. 1963. mco and paleobotanical studies 
of the Nyssaceae. I. A survey of the modern species and 
their Er J. Amold Arbor. 44: 1— 

66. The Nyssaceae in the southeastern United 

States. j Amold ee 47: 117-125. 

988. Comprehending Cornus: Puzzles and Pd 
in the die. of the dogwoods. Bot. Rev. 54: 2. : 
Falk, A. € E. Holsinger. 1991. Genetics and 

Conservation ri Hans Plants. Oxford University Press, 
Oxford. 


Fang, W. P., g & H. Y. Su. 1983. 
Pp. ibt in W. P. p & Z. E Zhang ee Ton 
Reipublicae Popularis Sinicae, Vol. 52(2). Science Press, 
A 

Fischer, Matthies. 1997. Mating structure. and 
Sens and outbreeding depressi rare plant 
M germanica. (Gentianaceae). Amer. J. Bot. 84: 


1685-169 


ion in the 


Fisher, R. A. 1930. The Genetical Theory of Natural 

Eq ua P University Press, Oxford. 
Fritsch, P. & L. H. Rieseberg. 1992. High o aed rates 
manta. su and hermaphrodite individuals in popula- 
plant Datisca glomerata. Naire, 356: 


Be E ri A China Plant Red Data Book, Vol. 1. Science 
Press, eee 
2. China Plant Red Data Book, Vol. 2. Science 


Press, eijing. 

Gong, X., Q. T. Zhang, G. D. Tao, Z. Z. Feng, Z. H. Yang & 
Y. T. Liu. 2006. Rare Plants of Yunnan in China I. 
Yunnan Science and Pa Press, K 

Harder, L. D. 998. A cla P of pollen 
discounting and its joint. ues with inbreeding depres- 
sion on mating system evolution. Amer. Naturalist 152: 
684—6 ES 

Hu, S. Y. 1993. Experimental methods in plant embryology 
o determination of pollen viability. Chin. Bull. Bot. 10: 


ae pm IUCN Red List of Threatened Species. IUCN, 


(editor), Solanaceae: Biology and ais: Columbia 
University Press, New York. 

Liston, A., L. H. Rieseberg & T. S. Elias. 1990. Functional 
androdioecy in the flowering plant Datisca glomerata 
Nature 343: o 

Lloyd, D Sexual strategies in plants. 
quantitative d for n the gender of plants. 
New Zealand J. Bot. 18: 103-10 


& C. J. Webb. 1 ¡Secondary sex characters in 

plants. Bot. Rev. 43: 

María, B. G. & J. A. rs m Sex ratio and sexual 
dimorphism in the dioecious Borderea pyrenaica (Dioscor- 
eaceae). e 101: 59-67. 

Mayer, S. S. & D. Charlesworth. 1991. Cryptic dioecy in 
flowering plants. Trends Ecol. Evol. 6: 320-325. 


Melampy, M. & H. Howe. 1977. Sex ratio in the A tree 
Triplaris americanus (Polygonaceae). Evolut 3l: 
867-872. 


Opler, P. & K. Bawa. 1978. Sex ratios in tropical forest trees. 
Evolution a Ren 
Pannell, J. 
cure udi 
———. 200 


The evolution and Sa of 
v. Ecol. Syst. 33: 397-4 
e. plasticity and a facon Es 
genetic Per pE of plant gender. New Phytol. 1 
506-510. 


Qin, H. N. & C. Phengklai. 2007. Nyssaceae. Pp. 300-303 in 
Z. Y. Wu, P. H. Raven & D. Y. Hong (editors), Flora of 
China, Vol. 13 (Clusiaceae through Araliaceae). Science 

dun Beijing, and Missouri. Botanical Garden Press, St. 

Loui 


Queen ugh, S. A., rslem, N. C. 
Valencia. 2007. (uen anes of biased sex ratios and 
inter-sex costs of reproduction in dioecious tropical forest 
trees. Amer. J. Bot. 94: 67-78. 

biais D. 1981. Seven forms of rarity. Pp. 205-217 in 
H. Synge (editor), The Biological Aspects of Rare Plant 
Conservation. Wile 

Real, L. 1981. MC and pollinator-plant interactions: 
The foraging behavi nd wasps on artificial 
flowers. E ie 20-26. 

Ross, M. D. 1982. Five evolutionary pathways to subdioecy. 

Amer. eri 1e 297—318. 


Volume 96, Number 4 
2009 


Sun et al. 683 
Cryptic Dioecy in Nyssa yunnanensis 


Sakai, & S. pS Weller. m O and sexual 
dim: s sm in flow plants: of terminology, 
biogeographic EHE Se dos i undas an s phylo- 
genetic wt Pp. 1-25 in M Um E. 
Dawson & L. Delph Pi bos er d Sexual 
1 m in icum Plants. inger, Ber 
991. Sex ratio ae in uk: biased 
sr" of M o iene 212: 169-179. 
Schles , P. P. Lowry II & D. G. Lloyd. 1990. 
Finetioral po w Caledonian endemic 
Pu x pandin ed Bat 22: 133-1 
Shea M., R. Sharitz. 1993. Size 
m sex ratio, e pu distribution of male and 
female water tupelo, Nyssa aquatica (Nyssaceae). Amer. J. 
vip 80: 26— 2 


in rus 


H. Zhang. 


& 1989. Rare and 
Ending [s in China. 


China Forestry Press, 


E B. L. & C. Q. Zhang. 2007. A revised d E 
v yunnanensis ete, Acta Bot. Yunnan. 
173-175. 

—— u, F. Q. Shi & Z. K. Wu. 


Seed. mòrphölogy d m of different treatments on 
germination o of the critically endangered Nyssa yunn 
sis (Nyssaceae). Acta. En d nns. 29: 351- 


nen- 


Turner, I. M., H. T. W. Tan, Y. C. Wee, A. B. Ibrahim, P. I. 
Chew & R. T. Corlett 1994. A study of plan t species 
in Singapore: Lessons for the conservation of 

tropical biodiversity. S Biol. 8: 705-712 


extinction 


Wang, . Xie. 2004. China Rm Red List, Vol. 1. 
Red List Higher Education ir: e ing. 
Wangerin, W Nyssaceae. Pp. pi 15 im A. En 


gler 
teo Das Pflanzenreich, Vol. 41. Wilhelm Engelmann, 
el zi 

Web C. J. & D. G. Lloyd. 1980. Sex ratios in New Zealand 

a. 18: 121-126. 

Wen, J. € T. F. Stuessy. 1993. 
biogeography 2 Nyssa (Cornaceae). Syst. Bot. 18: 68-79 

T D. R. 1969. Wind pollination in the angio- 

2r and environmental considerations. 
Pound 23: 28-35. 

Wu, C. Y. & J. R. Fan. 1977. Nyssaceae. Pp. 288-293 in C 
Y. Wu (editor) p Yunnanica, Vol. 1. Science Press, 
Beijing. 

—, A.M. Lu, Y. C. Tang, Z. D. Chen & D. Z. Li. 2003. 

The families and genera of angiosperms | in China: A 


H. Yamplosky. 1922. Distribution of 
x forms in the phanerogamic flora. Biblioth. Genet. 3: 
ps 


Appendix l. Pollinators and visitors of Nyssa yunnanensis. 
Observed on 
Species Family Voucher! Male tree Female tree Reward 

Nepitisnata adipala Moo: Nymphalidae Sun2006032806 + + nectar 
Cethosia bibles bibles (Drury) Nymphalidae Sun2006031803 + + nectar 

Delias hyparete L ieridae 2006031801 + + nectar 
Ypthima medusa Leech Satyridae Sun20060315 + + nectar 
Lethe Hübner sp Satyridae Sun2006031601 + + nectar 
Polyhachis Fr. Smith sp. Formicidae Sun2006032005 + + nectar 
Monomorium orientale Mayr Formicidae $un200003 1301 + = nectar 
Monomoriam chinense Santschi Formicidae Sun2006022901 + + nectar 
Apis cerana Fabr. pidae Sun2006030604 + + nectar, pollen 
Monomorium pharaonis (L.) Formicidae Sun2006022805 + + 
Polistes Latreille s Vespidae Sun2006030601 + + nectar, pollen 
Myrmosa melanocephala (Fabr.) Tiphiidae Sun2006031205 + pollen 
Mutilla marginata Baer Tiphiida n2006031304. + + nectar, pollen 
Praestochrysis Linsenmaier sp. Chrysididae Sun2006031603 + = nectar, pollen 
Eurytomidae s Eurytomidae Sun2006030608 + nectar, pollen 
Curculionidae sp Curculionidae $ 06032801 = + nectar 
Dinoderus nce sp. Bostrichidae Sun2006030603 + nectar 
Donacia Fabr. Chrysomelidae $ 060; + + nectar 
Sphenocorynes - ETT Curculionidae — Sun2006031206 + m: nectar 
Gibbium Scopoli s tinidae un2006030503 + nectar 
Tethina cinerea i Tethinidae Sun2006031707 + + nectar, pollen 
Chrysomyia megacephala Fabr. Calliphoridae — Sun20060307 + + nectar, pollen 
Iphiaulax imposter (Scopoli) Braconidae n2006032806 = + nectar 
Syrphidae s Syrphidae Sun2006031701 + + nectar 
Syrphidae sp xi) Syrphidae Sun2006031701 + + nectar 
Psorophory Gulicidae 2006030701 + nectar 
Psorophory sp2 Gulicidae Sun2006030706 + nectar 
Psorophory sp3 Gulicidae Sun2006030601 + = nectar 
Tipula L. sp. Tipulidae Sun2006031901 + nectar 
Fannia Robineau-Desvoidy sp. Fanniidae Sun2006030702 + + nectar, pollen 


684 Annals of the 
Missouri Botanical Garden 


Appendix l. Continued. 


Observed on 


Species Family Voucher! Male tree Female tree Reward 
Parasarcophaga Johnston & Tiegs sp. Sarcophagidae  Sun2006031204 + + nectar 
Euxesta Loew. sp. Ulidiidae Sun2006030610 + nectar 
Cobolidia i (Meigen) Scatopsidae Sun2006030602 + + nectar 
Stratiom: iis Stratiomyidae  Sun2006030504 + + nectar 
Stratio Stratiomyidae  Sun200603 + nectar 
Blattella ( cel sp. Blattaria Sun2006032004. + nectar 

+, visiting flower; —, not visiting flow 


! Vouchers deposited at the Herbarium of the Kunming Institute of Botany (KUN). 


SYSTEMATICS OF THE SOUTH Estrella Urtubey,? Tod F. Stuessy,? and 
AMERICAN HYPOCHAERIS Bann. Tremetsberger 
SESSILIFLORA COMPLEX 

(ASTERACEAE, CICHORIEA E)! 


ABSTRACT 


n" 


The URP sessiliflora complex (Asteraceae, Cichorie t f the genus from South America (all 
in section .) that have sessile or nearly ER fee heads surrounded by a rosette of d They occur 
at 1430-5100 m in elevation along the Andean chain from Venezuela to Chile and Argentina. Two species, H. sessiliflora 

unth and H. meyeniana (Walp.) Benth. & Hook. f. ex Griseb., are extremely polymorphic, and they vary conspicuously in the 


(Venezuela to central Peru, and Peru to northern Chile e northwestern Argentina, respectively), men M rA the 
year, and they also are primarily associated with dry and sunny habitats. rized by retrorsely 
vinnatifid leaves (rarely lobate) and slightly narrower “yl apice: s. Hypochaeris lohan (Sch. Bip) Done and H. 


taraxacoides Ball are glabrous, whereas H. acaulis (J. Rémy) Britton has et shaggy aes on the leaves; all three 
occur in humid places, such as seeps or bogs. H; oes eriolaena oe ip.) Reiche and H. mucida Domke are pubescent, 
with long whiplike trichomes on leaves and phyllaries, giving a niveous-tomentose appearance. Hypochaeris echegarayi 


Hieron. (white corollas) and H. eremo, ophila Cabrera (yellow cor lela are two related species with shaggy trichomes on the 
abaxial surfaces of the phyllaries, both with considerable ecological tolerance, that grow in dry as well as humid sites. 
Morphological cladistic analyses siiggcšt a hypothesis of Mundo: within the mde: Surprisingly, H. acaulis from Chile 
and Argentina, although fitting morphologically within the A. sessiliflora complex, based on amplified fragment length 
polymorphism (AFLP) data, clearly does not seem to belong to this group. Instead, the species ties to H. palustris (Phil.) De 
Wild. and H. tenuifolia (Hook. & Arn.) Griseb., also from the southern Andes. The aA leseont habit of H. acaulis seems best 
interpreted as a parallel adaptation to survival at high elevations. 


RESUMEN 
n qa 1 sdeSudamérnca (sec Ach h con cabezuelas 


E Hy 
sésiles o cortamente pedunculadas y con una roseta de hojas basales. Habitan entre ia 14302 5100 m a lo m de los a 
desde Venezuela hasta el centro de Chile y el centro oeste de Argentina. Dos especies, H. sessiliflora Kunth e H. meyen 
: : 7 


(Walp.) h. & Hook. f. ex Griseb., so remadamente polimórficas (f de los filarios, hojas e indumento) y Ken 
una amplia distribución, desde Venezuela al centro de Perá, y desde Perá al norte de Chile y noroeste de Argentina, 
respectivamente; principalmente crecen en ambientes secos y soleados, y pareeen puo a año. APUD meyeniana se 
caracteriza por las hojas pinnatífidas y retrorsas oa lobadas) y los aquenios | n el ápice. Hypochaeris 


hohenackeri (Sch. Bip.) Domke e H. taraxaco ides Ball son glabras, mientras que H. acaulis (J. Rémy) Britton tiene hojas 
híspidas; estas tres especies están asociadas a ambientes húmedos (e.g., vegas y mallines). n p (Sch. Bip.) 
Reiche e H. mucida Domke son níveo-tomentosas. Hypochaeris echegarayi Hieron. (corolas blancas) e H. eremophila Cabrera 
(corolas amarillas) son especies afines con beta híspidas, comúnmente con e ia y con 
considerable tolerancia ecológica, creciendo en ambientes secos o húmedos. Pon Hipótesis del analisi cladístico morfológico y 
los recientes estudios moleculares muestran al complejo H. sessiliflora la hipótesis basada en datos 
moleculares excluye a H. acaulis del complejo H. p donde el Wie acaule es RM MER una adaptación paralela 
como respuesta a la sobrevivencia en la alta montaña. 


Key words: | Achyrophorus, Asteraceae, Cichorieae, Hypochaeris, South America. 


1 We express thanks to John McNeill and Werner Greuter for help with the nomenclature of Hypochaeris taraxacoides; A 
Migoya and M. Theiller for illustrations of the species and H. Calvetti for help with ma pss the curators of BM, CONC, K, LP, 
LPB, M, MCNS, MO, NY, OS, P, SGO, SI, UC, US, W, and WU for loans or permission to consult herbarium material; M 
Muñoz (SGO) and L. Willemse (L) for digital i images of the types of Distoecha taraxacoides and H. ornata and H. parvifolia, 

i 8. 


manuscript; À. Luck for editorial help with various drafts of the manuscript; and financial support from the Consejo Nacional 
de iu iu EL "n Técnicas (CONICET, PEI and PIP6510), Myndel Botanica Foundation, a small grant from the 
University of Vienna, and the Austrian National Science Foundation (FWF, grant numbers P15225-BIO and P18446-B03). 

? [nstituto de Bos Darwinion, Labardén 200, San Isidro, B1642HYD, Casilla de Correo 22, Buenos Aires, Argentina. 
eurtubey @darwin.edu.ar 

? Department of a and rom Botany, Faculty Center Biodiversity, University of Vienna, Rennweg 14, A- 
1030 Vienna, Austria. 

‘Institute of Botany, Department of ns Biology and Biodiversity Research, University of Natural Resources and 
Applied Life Sciences, Vienna, Gregor-Mendel-Strafie 33, A-1180 Vienna, Austria. karin.tremetsbergerOboku.ac.at 

doi: 10.3417/2006136 


ANN. Missouni Bor. Garp. 96: 685-714. PUBLISHED ON 30 DECEMBER 2009. 


686 Annals of the 
Missouri Botanical Garden 
Hypochae CLADISTIC ANALYSIS 


eris L. consists of about 60 species of 
ial herbs in Asteraceae, tribe 
P DC., subtribe Hypochaeridinae 


annual or per 
Cichorieae Lam. 
Less., which are characterized by paleae on recepta- 
cles, a plumose pappus, and style branches papillose 
to well below the bifurcation (Fig. 1A-C). The genus 
12 species 
on, two in Central 

urope, one in Ásia, and the remainder (some 40 to 50 
species) in South America. Hoffmann (1893) reduced 
mber of rows of pappus bristles, 
shapes of heads, arrangements of phyllaries, and 
shapes of the cypselae. One of these sections, 

Achyrophorus Scop., is characterized primarily by 
having a uniseriate pappus; all species from South 
America fall into this group. 

Previous evolutionary studies have been completed 
on species of Hypochaeris from South America. 
Chromosome counts from many species and popula- 
tions have revealed most members of Hypochaeris to 
be diploid, with some infraspecific poly E and 
"n some species p y tetraploid (Ba in 

000; Weiss et al., ; Weiss- Sine wee et al., 
2 2008). es all diploids uniformly have 2n 
= 8, small karyotypic differences allow hypotheses 
regarding modes of odd evolution. (Weiss- 
Schneeweiss et al, 2003). Sequences of nuclear 
ribosomal w oris and jn oe DNA have 
revealed that the uth can 
iiie and e 


species are 
fe are closest to H. 
angustifolia Maire from Morocco (Samuel et al, 
2003; Tremetsberger et al., 2005). Sequence variation 
within species from South America is minimal, and no 
clades be recognized with high levels of 
confidence. Amplified ao cp polymorphism 
ped r 


( tudies have hel eal phyletic assem- 
blages within the South American taxa (Tremetsberger 
et al, 2006), as we phylogeographic 
trends at the populatigial level in H. acaulis (J. Rémy) 
Britton, H. DE (Phil) De Willd., 
m (Hook. 
al., 


s re et al., 


as indicate 


and H. 
Arn.) en enone et 


as ihe South ce taxa, a EET 
distinctive group is formed by species that have single 
sessile or nearly sessile capitula nestled among a 
rosette of basal leaves, the so-called ij ndi 
sessiliflora complex. These taxa occur from Venezuela 
south into Chile and Argentina and occur at 1430- 
5100 m, usually ry habitats or in seeps within 
these high An 


complex as a beginning for understanding detailed 


rm regions. This paper focuses on this 
morphological relationships within the South Ameri- 
can species of Hypochaeris. A revision of all taxa is 
included plus a morphological cladistie analysis. 


MATERIALS AND METHODS 


Morphological characters and states were obtained 
primarily from study of herbarium material from B 
CONC, K, LP, LPB, M, MCNS, MO, NY, OS, P, SGO, 
I, UC, US, W, and WU, 


purs and field observations. 


un 


plus our additional 
ro: an 
tal ty d terminolog sed ac 

re e "Chalk (1950, 1979, " Uphof du pe 
Harris and Harris (1994). For study of vascularization 
of corollas and cypselae, the clearing and staining 
(1963) was used. 
regarding color, habitat, and Dus 
taken fr ls 


Information 
of pla 


ogy 9 
e literature, and our 


method of Fuchs 
nts was 
rom specimen labels, 
field observations. Le numbers were ob- 
tained from the reviews of Weiss et al. (2003) and 
Weiss-Schneeweiss et al. (2007, 2008). 

e ingroup consisted ll nine South American 
species of the Hypochaeris sessiliflora complex: H. 
acaulis, H. echegarayi Hieron., H. eremophila Ca- 
brera, H. eriolaena (Sch. Bip.) Reiche, H. hohenackeri 
E Bip. Domke, H. meyeniana (Walp.) Benth. & 

k. f. ex Griseb., H. mucida Domke, H. sessiliflora 
Pu. and H. taraxacoides Ball. We included five 
additional South American species of Hypochaeris to 
serve as outgroup. These species were selected to 
represent morphological d within Hypochaeris 

ookeri Phil. and H. 


caespitosa Cabrera) are n to be close relatives 


on the continent, two of w 


of the H. sessiliflora complex. Hypochaeris hookeri is 

closely related morphologically to the complex 

(especially in its acaulescent habit). Its leaves are 
d h 


long, linear, and erect, the peduncle is shorter than 
the leaves, and the paleae have trichomes on the 
adaxial surface; it is distributed in central-western 
Argentina and central-eastern Chile. Hypochaeris 
caespitosa is also an acaulescent herb, the peduncle 
is longer than the leaves, and it inhabits the mountain 
ranges of central Argentina. We also examined three 
caulescent species, H. argentina Cabrera of central 
Argentina, and H. chillensis (Kunth) Hieron. and H. 
elata (Wedd.) Benth. & Hook 


south into Argentina. 


. f. ex Griseb. from Peru, 


To provide a context for selection of characters and 
states in the cladistic analysis, and to better clarify 


follows here (see also Fig. 1). All nine species of the 
Hypochaeris sessiliflora complex are dwarf perennial 
herbs, from only 1 em (H. mucida) to 13 cm (in H. 
sessiliflora) tall. They are acaulescent plants with 
thick rhizomes and rosette leaves. The capitula are 
always solitary, and they can be sessile or shortly 


Volume 96, Number 4 E E 687 
2009 a sessiliflora Complex 
pu raceae) 


A B C 
AAAVAVN 
| J G H 


7300-0-0- 0-0-0706 


K L M N O P 

Figure 1. Selected morphological features of t ic effi in the Hypochaeri ilifl A. Palea. —B. 
D. ius SEA —C. Papillose style. D-H. cu T Papillate. —E. Conical (many- EN m Whip. —G. 
Shag H. Glandular. I, J. Corolla vascularization. —J. Type 2. K-M. Anther tails. —K. Type 1. —L. Type 2. 


M Type e 3. m ny appendages. —N. Type 1 =. po 2. —P. Type e 3. 


Annals of the 
Missouri Botanical Garden 


pedunculate (never longer than the leaves). The leaves 
are entire or pinnatisect, the leaf base is attenuate in a 
long petiole and tomentose, and the blades are 
commonly lanceolate or linear to elliptic. The leaf 
margin is entire or dentate. The epidermal cells are 
E undulate, and the stomata are anomocytic. 
olucre is c only campanulate, campan 

ulate- «cylindrical (Hypo Gies eremophila), or GR 
dric (H. taraxacoides), and sometimes hemispheric 
(H. acaulis, H. meyeniana, and H. sessiliflora). The 
phyllaries are in three to six series. The outer ones 
are lanceolate to broadly ovate, often becoming 
gradually lighter toward the hyaline margin. The 
maximum expression of this character state occurs in 
some populations of H. sessiliflora in which the apical 
margin is broadly expanded, called “cucullate” 
(hooded), and it ean be entire or divided. The 
etimes 


aries are either glabrous or ciliat 
whi 


phy le, som 
lanuginous p trichomes) and/or setaceous (shaggy 
trichomes). The adaxial side of the paleae is 
glabrous 

corolla is of typical Lactuceae-type, i.e., 
ligulate and 5-lobed, with epidermal cells lightly 
undulate and ih epicutieular plated waxes (very 
conspicuous in Hypochaeris mucida), and it has short 


middle third 


corolas is characterized by the 


trichomes on the . The vascularization of 
adjacent marginal 
bundles united and fused at the apex of their lobes, 
and there are two different types of vascularization. 
The most common pattern has five marginal bundles 


1D; the 


other type comprises the same number of bundles, but 


united at the apex of the lobes (type 1, Fig. 


these are repeatedly divided in the throat and ligule of 
ig. 1J). The color of the corolla 
varies: white in H. echegarayi and H. taraxacoides; 


the corolla (type 2, F 


yellow in H. acaulis, H. eremophila, and H. hohenack- 
eri; or white, yellow, or orange in H. eriolaena and H. 
sessiliflora. Some plants with pink corollas in H. 
hohenackeri and H. sessiliflora have been collected. 
We have no information about the corolla color of H. 
ucida, because the available s 
faded, none of the 
and we have not collected the taxon. 


specimens are t 
arium labels mention the color, 


The five stamens have short anther tails, with basal 
appendages with pollen sacs extending almost to the 
end (1/2 to 3/4, Fig. 1K), into the middle or less (1/4 
to 1/2, Fig. 1L}, or basal appendages with only sterile 
cells (Fig. a The antheropodium consists of 

epidermal cells that are axially elongated with 
thickened ho lignified walls, and it can be s 

than, the same length as, or longer than the basal 
appendages (Fig. 1N—P). 

The cypselae are commonly erostrate or narrowed 
toward the apex; in Hypochaeris acaulis they are 
rostrate. The cypselae are usually 5-ribbed (some- 


times inconspicuous), and the surface is smooth or 
scaly and striate. 

Five types of trichomes are present, three being 
uniseriate: (1) papilla-like (Fig. 1D), a small unicel- 
lular epidermal cell thickened at the apex, being 
present on lobes of the corolla; (2) conical (Fig. 1E), 
2- to many-celled, e.g., on the margins of the palea or 
pappus bristles (2-celled), or comprising the cilia of 
the phyllaries (2- to many-celled); and (3) whip 
(Fig. 1F), many-celled, with a very long apical cell, 


trichomes are multiseriate: 
many-celled, ending in one to few cells, forming a 
setaceous indumentum (these trichomes are often 
brown or red, i.e., on phyllaries of H. echegarayi, H. 
eremophila, and H. sessiliflora); and (2) biseriate- 
glandular (Fig. 1H), present on the leaf surface of 
some populations of H. sessiliflora. We have used the 
term glabrous when trichomes are relatively few or 
totally absent. 

From this spectrum of morphological variation, 16 
morphological characters (five vegetative and 11 

uti EA 
e listed i 
Appendix 1, and selected oe are e highlighted i in 


oral) were i So a evolu 
. The 


relationships among the 
Figure 1. Autapomorphic characters were not includ- 
The data matrix of the 14 
and 16 characters is given in Appen 


ed in the analysis. 


are listed in Appendix 3. A complete listing of 
specimens us in this study is presented in 
20% 


ative within the 


Appendix 4. Polymorphisms represent nearly 
of the data. ibd 2 was uninform 
complex, but it was included because it is important 
for circumscribing all the caulescent species of 
'ypochaeris 

Phylogenetic analyses were performed with Pee- 
Wee 3.0 (Goloboff, 1998). Initial PeeWee runs were 
conducted using sequence commands holding 10000: 
rseed 0, hold/40, poly=, amb-; mult*500; fit* = fit 
rescaled. All characters were weighted equally and 
character states were treated as unordered. Jackknife 
support values for nodes were caleulated with 1000 
replicates, two search replicates (mult*2), and two 
starting trees per replication (hold/2). Character 
distributions were studied and strict consensus trees 
were calculated using Winclada (Nixon, 1999). 


RESULTS 


The analyses resulted in one most parsimonious 
tree of 40 steps, with fit = 110 and fit* = 57%; 
synapomorphies and homoplasies are mapped onto the 
phylogeny (Fig. 2). All species except Hypochaeris 
argentina are supported by the presence of type 1 


Volume 96, Number 4 
2009 


Rand 689 
ee sessiliflora Complex 


pen raceae) 


corolla vascularization (character 11, state 1). The 
pair H. chillensis-H. elata is supported by corollas 
with the same length as the 
state 1) and 
caespitosa and the rest of the species are supported by 


involucre (character 10, 
a jackknife value of 64. Hypochaeris 
md heil lanceolate leaf laminas (character 3, 


state 4), w 
phyllaries ue 9, state 1), and smooth cypselar 


richomes on the abaxial surface of the 
walls (character 15, state 1). Hypochaeris hookeri is 
the sister species of the ingroup, a clade supported by 
acaulescent plants or plants with shortly do. 
heads (character 1, state 1) and jackknife 2 rt o 

The H. sessiliflora complex is supported as a aus 
letic clade by glabrous phyllaries and paleae (character 


H. sessiliflora is the basal species 
The clade of the remaining species is supported by 
linear s Lus 3, state 1). SPACE 
oug aulis are support: 
ecu RA ue 8, state 1). The clade 
of H. mucida and the rest of the species is supported by 
sessile heads (character 7, state 1). The remaining three 
roups are supported primarily by homoplasies, the first 
homoplasy being a reversal (scaly cypsela 
character 15, state 0), followed by a parallelism (divided 
leaves, character 4, state 2). The unresolved p 


r wall, 


contas of H. . meyeniana, 

pported by a 

character 3, state 0) and one synapomorphy (shaggy 
) 


. echegarayi 
reversal (lanceolate dee 


phyllaries, character 9, state 2 


DISCUSSION 


Hypochaeris hookeri appears as the sister species of 
the i e Future studies, including the rest of the 
n species, must evaluate whether H. 
ae a properly belong to the H. sessiliflora 
complex or not. The results tentatively support 
monophyly of the H. sessiliflora complex, but for a 
critical morphological assessment of this question, a 
cladistie analysis of all South American taxa must 
completed. 

The basic structure of relationships (Fig. 2) makes 
considerable sense. Hypochaeris sessiliflora is a large, 
acaulescent herb that contains conspicuous morpho- 
logical variation (Fig. 3) and has a broad range in the 
northern Andes (Fig. 4A). Hypochaeris taraxacoides 
and H. hohenackeri are very close relatives morpho- 
logically (cf. Figs. 5, 6) and are found in similar seeps 
above 2600 m in the central Andes (Fig. 4B, squares 
and circles). Their evolutionary history must be 
closely entwined. 

Hypochaeris mucida and H. eriolaena are both very 
tomentose with sessile heads (Figs. 7, 8). They are 
easily distinguished from each other in features of the 


involucre, corolla vascularization, and cypselar wall. 


tall, densely pubescent, with shaggy 
trichomes on the backs of the phyllaries and leaves, 
and growing above in grassy steppes 
associated with Pyenophyllum J. Rémy (Caryophylla- 
i Peru and 


Bolivia on the eastern side of Lake Titicaca (Fig. 4B, 


ceae Juss.) in southeastern northwestern 
triangles). Hypochaeris eriolaena has whip trichomes 
on the backs of the phyllaries and leaves (or glabrous), 
2200 and 5100 m) 
from Peru to Bolivia (Fig. 4C, squares). 
The other rosette 
Hypochaeris meyen 


and it inhabits dry places (between 


taxa with divided a 
echegarayi, a 

eremophila, are o i Morphologically, > 
last two are similar (Figs. 9, 10), both with relatively 
short peduncles, very divided leaves, phyllaries with 
long trichomes on the midribs (Figs. 10C, D; 11C—E), 
and plumose pappus bristles. They have different and 
consistent floral colors: white versus yellow, respec- 
tively. They are also both found in the junction 

etween southern Peru, southwestern Bolivia, and 
northwestern Argentina. It is very likely that these two 
taxa share a common origin. Hypochaeris meyeniana is 
similar to the other two, but its heads are subsessile, 
leaves less divided, E nearly glabrous, and 
pappus bristles wit short side trichom 
Fig. 11). 

Hypochaeris acaulis, 
southern Andes (Figs. 12, 13), is geographically far 


— 


which occurs only in the 


removed from the other taxa of the H. sessiliflora 
complex. Although morphologically H. acaulis fits 
comfortably within this complex (it can be differen- 
tiated from the other species of this complex by its 
broader phyllaries and rostrate cypselae; Fig. 12; see 
also key) recent AFLP data do not support this 
inclusion. Based on a broad sampling of taxa of 
Pide x South America (but not including 

o lack of material), Tremetsberger et 
al (2006) a Ps that A. acaulis does not at all 
relate to the H. sessiliflora complex, but instead to H. 
palustris and H. tenuifolia, which are also found at 
high elevations along volcanoes of the southern Andes 
in Chile and Argentina (Tremetsberger et al., 2003a, 
b; Muellner et al., 2005). This makes sense geo- 
graphically. Hypochaeris acaulis appears to represent 
case of extreme morphological parallelism as 


e 


adaptation for survival against the harsh environment 


(especially low temperatures and wind). Hypochaeris 
a d here f l be 


caulis is include or convenience; it wil 
separated in a future comprehensive treatment of the 
entire South American complex of the genus. 
Considering the classical biogeographic division of 
the Andes into three portions (Morrain, 1984; Taylor, 
1991), the northern Andes (10°N to 3^S), central 


Annals of the 
Missouri Botanical Garden 


H. argentina 


10 64 H. chillensis 


H. elata 
H. caespitosa 


———— H. hookeri 


Mos o tree of 40 step: 


Figure 
states es below lin 


Andes (37S to 18°S), and southern Andes (18%S to 
54°S), only H. sessiliflora is present in the northern 
Andes (Fig. 4A). Hypochaeris acaulis is the only taxon 
in the southern Andes; as discussed above, it appears 
ry. Most 


of the species of the complex occur in the central 


to have had an independent evolutionary history 


es; the major evolutionary and biogeographic 
differentiation must have occurred in this general 
region. Exactly w 
and biogeography have been in this subgroup remain 


what the mechanisms of speciation 


to be determined; they likely have involved dispersal 
throughout the zone, isolation due to diverse topog- 
raphy, adaptation to local ecologic conditions, and 

o ridization 
between H. echegarayi and H. eld 


obs.) in western Bolivia emphasizes the evolutionary 


meyeniana (pers. fi 


s (fit = 110; fit* = 
see Appendix 1) ER clades. Closed circles indicate synapomo 
homoplasies re or reversals), The numbers in italics are jackknife support values greater than 50% taken from a 
strict consensus t 


H. sessiliflora 


H. taraxacoides 


H. hohenackeri 


H. mucida 
—— — H. eriolaena 
H. meyeniana 


H. echegarayi 


H. eremophila 


H. acaulis 


57%), showing all characters (numbers above line) and 
rphies; open circles indicate 


closeness of this group and perhaps also its recent 
evolutionary origins. 


TAXONOMIC TREATMENT 


Hypochaeris L., Sp. Pl. 2: 810. 1753. TYPE: 


Hypochaeris glabra L. 


The genus Hypochaeris is characterized by the 
presence of latex, ligulate corollas, receptacular 
de and a plumose pappus. It comprises approx- 
ecies with an importa 
America (where m 


E 


mately 60 sp 
veces in South 
2/3 of the species occur). The rest of the taxa inhabit 
the Old World and Asia. All South American species 
fall under section Achyrophorus, characterized by the 
presence of a uniseriate pappus. 


691 


Volume 96, Number 4 Urtubey et 
2009 The phe sessiliflora Complex 
(Asteraceae) 


> 
— 
== 


"EA 


===== 


A _ 


= 
SSS 


=== 


== 


— 


== 


<= == 


=== 


i 
Mi 
| Il E 
M E 
S 
R 
E 
E £ 
E o 
F G H I [e P 
Figure 3. Hypochaeris sessiliflora. —A. Habit. —B-D. ei in capitula and ras —E. Setaceous te —F-J. 
Phyllaries. ary Trichomes occurring on some phyllaries. —K. ar trichomes. —L. Shaggy trichomes. —M. Conical 
trichomes. —N. Corolla with anthers and style. —O. Palea, Eno and p: P. Smooth achenial wall. 0 R. Scaly 


appus 
cypselar wall. A, e F-I, N-P from Cuatrecasas et al. 25599 (US); B from ces et m 18538 (LP); and D, E, J-M, Q, R from 
Plowman 1955 


692 Annals of the 
Missouri Botanical Garden 


oL VH Calvo 80" 70° 60" oL VA Calvi 80 70° 60 


Figur —A. pa ee i DE eon in the northern and central Andes. —B. Dist rasos in the central 
and nn Andes of H. hohenackeri (BI), H. da (A), and H. taraxacoides (). —C. Distribution in the central Andes of 
H. echegarayi (A), H. Stm pa and H. pees (B). —D. Distribution of H. meyeniana in the kental Andes 


a 


The Hypochaeris sessiliflora complex REA leaves), which inhabit dry and humid places along the 
nine South American species of dwarf (1-13 cm tall), ndean mountains, from southern Venezuela to 
perennial, acaulescent herbs, with solitary Nene central-eastern Chile and northwestern Argentina, 
that are sessile or pedunculate (shorter than the between 1430 and 5100 m. 


KEY To SPECIES OF THE HYPOCHAERIS SESSILIFLORA COMPLEX 


la. Outer phyllaries broadly ovate (as wide as long); cypselae rostrate, proximally scaly .............. 9. H. acaulis 
lb. Outer phyllaries usually longer than wide; cypselae erostrate or narrower near apices, scaly or smooth. 
2a. pue cylindre: corollas white «ex Vosotros tac IB mida . H. taraxacoides 
2b. Involucre campanulate-cylindric, campanulate to hemispheric, corollas yellow (white in H. echegarayi, 
CHA white in H. sessiliflora). 
3a. Plants to 1.5 em tall; phyllaries with both whip and shaggy trichomes ................. 4. H. mucida 


3b. iari = cm tall; phyllaries glabrous or with only one type of trichome. 
es undivided (margins entire or dentate). 


Volume 96, Number 4 
2009 


laa he 693 
ie sessiliflora Complex 


pei raceae) 


5a. Outer phyllaries lanuginous (whip trichomes) toward the apex 
5b. ruri dar lario or setulose (shaggy trichomes). 
s linear-lanceolate or elliptic-lanceolate 


UTER 5. H. eriolaena 


1. H. sessiliflora 
H. hohenackeri 


d Lt oblon ur cT 
4b. Leaves pinnatipartite to pinnatisect. 
Tax Corollas white adn otek Sut eM ah en Be tO apa ee 1. H. echegarayi 
Tb. Corollas yellow. 
8a. Heads usually sessile; involucre campanulate ......o.ooooococonno.... . H. meyeniana 
8b. Heads usually pedunculate; involucre eylindric- campanulate ........... 8. H. eremophila 


1. Hypochaeris sessiliflora Kunth, Nov. Gen. Sp. 
[HBK], folio ed. 4: 2. 1818, quarto ed. 4: 2. 1820. 
Oreophila mom (Kunth) D. Don, Trans. 


in alta convalli Quitensi, juxta mo 
mum Pichincha, alt. 1500 hex." [ca. 3000 m], 14 
Apr. 1802 [Sandwith, 1926], F. W. H. A. von 
Humboldt & A. J. G. Bonpland s.n. (holotype, P!, 
P photo MO!). Figure 3. 


Hypochaeris tee Kunth, Nov. Gen. Sp. [HBK], quarto 
ed. 4: 2. 1818, quarto ed. 4: 2. 1820. Achyrophorus 
SÓ (Kunth) DC., Prodr. (DC.) 7: 95. 1838. 
Achyrophorus quitensis Sch. Bip. var. sonchoide. 

Ga Wedd., Chlor. And. 1: 219. 1857. TYPE: 


Ecuador. Prov. Pichincha: * p od cum bi Md 


14 Apr. 1802 [Sandwith, 1926], A. von 
Humboldt & A = G. Bonpland s.n. eres PeP 
photos LP!, MO 


E e Sehi: Bip., Jahresber. Pollichia 16-17: 
1859. Achyrophorus sessiliflorus (Kunth) DC. var. 

bara m ch. Bip.) A. Gray, Proc. Amer. Acad. Arts 5: 

. le DE barbata (Sch. Bip.) Reiche, 

ae ae Chile 116: 589. 1905. TYPE: Colombia. 


2, J. J. Linden 746 (lectotype, designated Hess. P: 
E BM [2]. K [2]. P!, W!, W photo MO}. 
Hypochaeris stuebelii Hieron., Bot. J 
1896. 


ntis Antisana,” Oct. 
(holotype, B [presumably destroyed], B photos LP!, 
M09. 


Herbs to 13 cm tall. Leaves linear-lanceolate or 
M oU 20-130 X 5-25 mm, attenuate in 
a broad petiole with 
ee to slightly dentate, both surfaces glabrous or 


conical trichomes, margins 


with some glandular trichomes. Capitula sessile or 
ue (to 10 em); bracts on peduncle linear, 
with long shaggy trichomes to 6 mm or glabrous; 
involucres campanulate to hemispheric, 13-25 X 10- 

5 mm; phyllaries 3- to 4-seriate; outer and middle 
phyllaries oblong-lanceolate to lanceolate, 7-15 X 2— 
4 mm, with shaggy trichomes on midribs to near apex 
or glabrous, at margins ciliate; inner phyllaries linear- 
lanceolate to linear, 11— mm, glabrous or 
sparsely ciliate at margins, often divided at the apex; 
paleae 12-23 mm; florets ca. 25. Corollas usually 


yellow or white, rarely orange or purple, 12-27 mm; 
tube 5.5— 
type l; stamens 7-19 mm; anthers 2.5-9 mm; basal 


mm; ligule m; vascularization 
appendages types 2 and 3, 0.5-1 mm; orn 5- 
14 mm; antheropodium types l and 9— 
with branches 1-2.5 


rostrate, or rr 2-7 mm; walls 


22 mm w 


semirostrate, 


mm. ue E ed 


scaly or end pappus 7-18 mm. Chromosome 
numbers 2 8, 16 (Jansen & Cid 1980; Olsen, 
1980; ens et ER 2003; Weiss-Schneeweiss et al., 
2007, 2008) 
Habitat and. distribution. 
fro 


has been foun m Venezuela 


places, between 2000 and 4500 m x. ma 


pud wis ck 
Boliv rier 


Phenology. Hypochaeris sessiliflora flowers 


throughout the year. 


Common names. 


“Achicoria” (Ecuador, Peñafiel 
ide & 


et ; “cachu-cachu” (Peru, Macbri 
Meses 1791); “chicorea” (Colombia, Soejarto 
494); "chicoria" (Venezuela, Jahn 146); “chicoria 
blanca” (Venezuela, Aristeguieta 2438). The white 
latex is reported as being used for toothache (Ecuador, 


Ellemann 91672). 


hological characters. Hypochaeris n 
fro f the x 


Morp 
is distinguished other species o 
by having leaves i are linear-lanceolate pa jen 
often elliptic-lanceolate), glabrous, and toothed or 
slightly toothed at the margin, and cypselae that are 
cylindrical 


Hypochaeris sessiliflora is the most polymorphic 
with respect to vegetative and reproductive characters 
(especially shape and pubescence of leaves and 
phyllaries, and variation in floral color) of all members 
of the H. sessiliflora complex (Fig. 3B-D). The heads 
can be pedunculate (to 10 em) or sessile, the outer 
phyllaries can have the margin hyaline or opaque, an 
in some populations (described as H. stuebelii by 
Hieronymus, 1896) the phyllaries are inflated with the 
apex either divided or entire. Another variable 
character is the color of corollas, with white, nnd 
purple often encountered in the sam 
arias 
(described as Achyrophorus barbatus, Schultz Biponti- 


Hypochaeris sessiliflora var. 


nus, was characterized by having linear 


peduncular bracts with long shaggy trichomes to 


Annals of the 
Missouri Botanical Garden 


0,1 mm 


01mm 


D E 


— B. Leaf. —C-E. Phyllaries. —F. Trichomes (in cilia) on phyllaries. — 


Figure 5. A. Ha 
G. Corolla with anthers and style. —H. Pales; S and pappus. —I. Smooth cypselar wall. From Stuessy et al. 18074 (LP) 


Volume 96, Number 4 
2009 


der ie 
de a sessiliflora Complex 
hes raceae) 


695 


6mm, setaceous phyllaries with yellow-green and 
purple trichomes, and outer and middle phyllaries 
that are divided at the apex (Fig. 3D). Despite these 
conspicuous morphological features, this form like- 
wise seems to have no geographic integrity; we deem it 
not worthy of formal recognition. We have seen 
variation in capitula as represented in Figure 3B—D 
within a ee population on Volcán Pichincha 
A et al. 18539) in Ecuador. Schultz Bipontinus 
(1845: 120) re struggled wi i 

variation within this species (as A. quitensis, nom. 
illeg.). He recognized a set a unnamed “Formae” in 
manner: = s gla hee Ila, 
IIb, hea ung as small; Illa, 


ax a IIb, ree Eu This e a 


with the morphological 


variable species. More populational sampling within 
its broad geographic range using morphometric and/or 
molecular analyses would be worthwhile. 


Observations. Hypochaeris sessiliflora and 
unth (1818) in ra 
same volume (even on the same page), and either 


could be selected for use (McNeill et al., 2006: Art. 


much more widely used name, 


sonchoides were published by K 


Schultz Bipontinus (1845: 120) püblished the 
names =o Sch. Bip., Scorzonera 


quitens sessilis Humb. under 


is Hum 
'ypochaeris sessiliflora. Because they were not 
accompanied by descriptions or specimens, T 
remain prosynonyms. Mc oi f. c 
lescens Hieron. (Bot. Jah t. 28: 658 190) i is 
another invalid name see it was published 
without descriptio 
chulz P (1859: 52) also published 
s Sch. Bip. and A. humboldtii 
Se th are nom. illeg., being superfluous 
uf de quitensis Sch. s which itself is 
illegitimate, 


Achyrophorus albiflo 
h. Bip., but bo 


names 
having been cited of 

'ypochaeris sessiliflora. Bortiri (1997 lectotypiied 
both albiflorus and A. humboldtii, 
illegitimate, this i is unnecessary. Hypochaeris ed 
ae var. albiflora Hieron. (Bot. Jahrb. ca 

1901), at is also an illegitimate n 

Reiche (1905) may have had a ice concept of 
Hypochaeris barbata, as he cited it as occurring in 
northern Chile (Prov. [— 
sessiliflora is not kno 


but being 


Region] i oe where 
n occur (Fig. 4A); he 
probably had material of H. md as he cited 
Distoecha taraxacoides Phil. in synonymy. 


Representative specimens. BOLIVIA. Cochabamba: Tu- 
naria [probably Cerro Tunari], Mü 
COLOMBIA. Boyacá: entre G: Zuloaga & 
Landono 4156 (SD). Caldas: Nevado del Cocuy, alto valle de 
Las Lagunillas, Cuatrecasas 1484 (US). Cauca: Cordillera 
Central, E slopes of páramo del Purace, around la laguna de 


San Rafael, Cuatrecasas & eae 26299 (US). Cundina- 
de Chisa: (NY). 


marca: Páramo Soderstrom 1254 

Magdalena: Sierra Nevada de pw Marta, valley descend- 

1 rom s Reina and E a, Cuatrecas & Romero 
El Galeras 


Rancho 


Carchi: Julio 


Occidental y Cordillera Oriental entre Oña y 
Ovejero, Barclay & Juajibioy 8472 (US). 


amba, ca. 20 km SW of pene King & Garvey 
6968 (US) Cotopaxi: 5.5 km E of Pujilí, Stuessy et al. 
18549 (LP, WU). Imbabura: Cotacachi Canton, Res. Ecol. 
Cotacachi-Cayapas, laguna de Cuicocha, Peñafiel et al. 382 


Ainchilibí y Río Portrero, E de 
Dd 9210 (US). Pichine ha: » km 

al. 18540 (WU). Tungurahua: Patate Des Parque Nac. 
Llanganates, lagun 
V nq 


Ch 
Wurdack 761 (US). Ancash: Huaylas, distr. Pamparnas, path 
arka to laguna Negra Huacanam, around the laguna itself, 
Weigend 2000/516 (NY). Celendín: Cel- 
endín, Smith & Cabanl las ote 
oo de las ruina 
Jun , Killip & Smith 23369 (US). Lima: Rio 
Blanco ps e Sm ith 21680 (US). Puno: Cavasaya, hda. 
Sojela, Vargas 21784 (LP). VENEZUELA. Apure: a lo largo 
del Río ies (Oirá) y sus afluentes, en paramos entre Alto 
de Cruces y Tierra Negra, Steyermark & G. C. K. & E 
Dials 101105 (NY). Mérida: Mérida, paramo at 
laguna de Mucubaji, near Apartaderos and San Rafael, 
35 km NE of Mérida, Maguire 39405 (NY). 


2. Hypochaeris taraxacoides n. Soc., 
Bot. 22: 48. 1885. Oreophila niu Meyen 
& Walp., Nov. Actorum Acad. Caes. Leop.-Carol. 
Nat. Cur. 19(suppl. 1): 291. 1843, non Oreophila 
1827. TYPE: “Peruvia: in 
planitie circa Tacoram," 14,000-17,000 ft., 
1831, F. J. F. Meyen s.n. (holotype, B [presumed 
destroyed, Stafleu & Cowan, 1981]. Figure 5. 


taraxacifolia Loisel., 


dg vie Te dm (A. Gray ex Wedd.) Kuntze var. 
erifolia ntze, Revis. Gen. Pl. 3(2): 160. 1898. 
Wedd.) Kuntze 


C. E. O. Kuntze s.n. (holotype, NY). 


Herbs to 7 em tall. Leaves lanceolate or oblong, 
1.5-11 X 0.3- 


at base attenuate. Capitula pedunculate to 40 mm; 


1.2 em, pinnatifid to lobate, glabrous, 


Annals of the 
Missouri Botanical Garden 


0,1 mm 


0,1 mm 


E 
Corolla with anthers 


B C D 
Figure 6. Hypochaeris hohenackeri. —A. Habit. —B—E. Phyllaries. —F. Cilia on phyllaries. —G. 
and style. —H. Palea, achene, and pappus. —I. Smooth cypselar wall. From Lewis 35168 (LP). 


o 697 
a sessiliflora Complex 
e) 


Volume 96, Number 4 
2009 
e raceae 


aw al Wan, 
" 


nu 


0,1 mm 


| "m 


B C D E F J 


abit. —B-E. Phyllaries. —F. Whip trichomes on phyllaries. —G. 


Hypochaeris mucida. 
mes cilia) on phyllaries. —I. Corolla with anthers and style. —J. Palea, cypselae, E 
(SD. 


Figure 7. 
viciome on phyllaries. —H. Tric m 
pappus. —K. Smooth cypselar wall. From Ceballos et al. 614 


Annals of the 
Missouri Botanical Garden 


involucre cylindrical, 14-23 X 5-10 mm; phyllaries 
3- to 4-seriate, lanceolate, 2 rounded, margin 
ciliate; outer hilados 8-11 X 2-3 mm; inner 


phyllaries 15-19 X 3-3.5 mm; paleae 12-18 mm; 


flowers ca. 30 per Parme Corollas white (upper 
surface) to dark blue-black at tips of ligules 
underneath, 15-21 mm; tube 7-9 mm; ligule 8- 


12 mm; vascularization type 2; stamens 10-17 mm; 
anthers 4—5 mm; basal appendages type 3, ca. 1 m 
filaments 6-12 mm; antheropodium type 1; style 12 
22 mm, with style branches 1.8-2 mm. Eas 
erostrate, 1.8-2 mm; wall smooth; E — 
mm. Chromosome number 2n = 8, FA 
1971; Weiss e et al, 2003 [as pde stenoce- 
phala]; Weiss- Ernte et al., 2007, 2008) 
Habitat 
coides wis bogs and seeps from Colombia to 


northern Chile and northwestern Argentina, 2640— 
5000 m (Fig. P circle). 


and distribution. Hypochaeris | taraxa- 


Phenology. Hypochaeris taraxacoides flowers 


pias the year. 


mmon names. “Achicoria” (Argentina, Haber 

142 = ‘chicoria’ Rod Mexia 4191); “diente de león” 

(Bolivia, Schulte 9); “flor de ciénaga” (Argentina, 

Cabrera et al. a “lechero” (Argentina, Haber 5); 

“mula-siki” an Wolstenholme 6); “pilly” (Peru, 
Mexia 419 


Morphological characters. Hypochaeris | taraxa- 
coides is readily distinguishable from other taxa by 
the combination of peduncular (rarely sessile) cylin- 
drical heads, glabrous phyllaries (ciliate on margins) 
and leaves, and corollas white on the upper surface to 


dark blue-black 


(abaxial surface), and being longer than the involucre. 


isotype has 
located, the detailed description and locality combine 


su 
at the tips of ligules dide. 


Observations. Although no been 
to leave no doubt as to the biologic affinity of this 
name. The locality of plains near “Tacoram” presum- 
ably refers either to the town of Tacora and/or Volcán 
Tacora, both of which are now within the boundaries 
of Chile, Re 


the border from Peru), Tacna Department. 


gion I, Parinacota Province (just across 
At the time 
of collection and for another half century, this region 

within Peru (e.g, Enock, 1908: cf. map). 
een (1843: xviii) tells us that Meyen went up 
from Arica, Chile, toward Lake Titicaca, and Tacora is 
on this route. This om is known from this general 
region (Fig. 4B, circle). 

In publishing pomo taraxacoides Walp. 
(Repert. Bot. Syst. 6: 336. 1846, nom. illeg.) [non A. 
taraxacoides (D. Don) Steud. (Steudel, 1841: 226)], 
Walpers (1846) purported to make a new 
based on Oreophila taraxacoides Meyen & Walp., but 


combination 


gave reference to the place of publication of O. 
taraxacifolia Meyen & Walp. Whether or not this 
incorrect epithet used in the combination should be 
treated as an error to be corrected to “A. taraxacifolia” 
or considered deliberate in view of the existence of A. 
taraxacifolia Moench (Moench, 1802) is unimportant 
ecause, in either case, the name is illegitimate, with 
Steudel having validly published the name A. 
taraxacoides five years earlier. deo taraxa- 
coides (Walp.) Benth. & Hook. f. (Gen. P 
1873) has also been used in some reports un Bortiri, 
1999, as synonym} and in much herbarium material, 
but this also uses the incorrect epithet in combination. 
More importantly, Bentham and Hooker (1873: 5 P 
did not actually make the combination. They 
listed ^... E 
Hypochaeris (as *Hypochoeris"), a particular technique 


in Achyrophoro taraxacoide, Walp. ines 


in this work that is specifically disallowed as a new 

the International Code 
of Botanical Nomenclature (and exemplified in 
McNeill et al., 2006: Art. 33, Ex. 2) The first 


publication of a legitimate name for the species in 


combination in Article 


Hypochaeris followed shortly thereafter: H. taraxa- 
coides Ball (J. Linn. Soc., Bot. 22: 48. 1885). It is 
curious that Walpers, one of the two original authors, 
cited the original epithet taraxacifolia incorrectly 
when transferring it into Achyrophorus, and in this was 
followed by Bentham and Hooker and Ball under 
Hypochaeris. They may have been aware of the earlier 
ompeting name, H. taraxacifolia Moench (Suppl. 
Meth. 224. 1802), now referable to Crepis albida Vill. 
a 1947: 310) or to the earlier O. taraxacoides 
which does not belong to the H. Ras 
Md but there is no evidence of t is 
fortunate for maintaining stability of usage p H. 
taraxacoides Ball can be treated as a new name in 
1885, for a new combination based on the epithet that 
was actually used would have resulted in th 
illegitimate H. taraxacifolia (Meyen & Walp.) Ball, 
a later homonym of H. taraxacifolia Moench (Moench, 
802). An earlier H. taraxacoides Pourr. ex Steud. 
(Nomencl. Bot. [Steudel], ed. 1, 422, 618. 1821) also 
exists, but appears as a synonym of Picridium albidum 
DC. (— Crepis albida Vill., Babcock, 1947: 310); the 
name is not validly published and so creates no 
competing difficulty. 

Achyrophorus stenocephalus A. Gray ex Wedd. is an 
invalid name because it was listed by Weddell (1857) 
as x P Achyr 
& Walp.) W 
Wedd.) Bind [Kuntze, 
illegitimate combination. 

characters used by Kuntze (1898: 160) to 


differentiate Hypochaeris stenocephala var. integrifolia 


ophorus pied qq (Meyen 
p. Hypochaeris stenocephala 


1898], 


A. Gray ex 


therefore, is an 


from variety taraxacoides are leaf margins entire or 


Volume 96, Number 4 


Urtubey et 699 
The obe m sessiliflora Complex 
(Asteraceae) 


[o D E G J K 
Figure 8. Hypochaeris a. A. Habit. eaf. —C-F. Phyllaries. —G. Trichomes a Ea on phyllaries. — 
Whip trichomes on phyllaries. —I. Corolla with a te E style. —J. Palea, cypselae, and pap —K. Smooth in 


wall. From Smith & Buddensick 11133 


( 


LPB). 


Annals of the 
Missouri Botanical Garden 


denticulate versus runcinate, respectively, which were 
essentially validating the listed, but not named, 
varieties in Achyrophorus fando Meyen & 
Walp.) Walp. by Weddell (1857). This variable 
feature (leaf margin) is insufficient for varietal 
recognition. 

The name Hypoch 
Walp.) Ball var. lanuginosa Herzog is a nomen 
nudum because it was only mentioned in a list of 


Bolivian plants published by Herzog (1923: 228). 


aeris taraxacoides (Meyen & 


ARGENTINA. Catamarca: 
de Antofagasta, Haber 5 

(SD. Jujuy: Capital, refugio del nevado de Chañi, Fabris, 
Cano & Tello 4035 (LP); Cochinoca, Abra Pampa, cerro 
Huancar, rea et al. 15256 (LP); Humahuaca, 31.4 km 

of Humahuaca on gravel rd. to El Aguilar, Stuessy, Urtubey & 
ces i 8089 (LP, WU): e 0.8 km NW of 
Chiapi Rodeo o 


Representative specimens. 


Rinconada, Mina Pir rquitas, Schwabe, Ancibor & Vizinis 641, 
818, 819 (LP); Tilcara, lecho de Río Grande cerca d 
Huacalera, Werner 820 (LP); Tumbaya, camino de El Angosto 
al Chañi, encrucijada Río Chañi, Cabrera et al. 22484 (LP); 
Valle Grande, Caspalá, Burkart & Troncoso 11822 (SI); Yavi, 
Quebrada de Toquero, A. L. & S. abrera, Malacalza 
von Schmiden 17656 (LP); Yes: l 2 km S RP 5, Km 13, 4— 
5 km SW de Yavi, Tolaba, Acufia, Arapa, Gutiérrez, n 
mallo & Da Siwa 1539 (MCNS). L 


pa 3155 (K, SD; Santa Victoria, 43.7 km from Santa 
Victoria on rd. to La Quiaca, Hawkes, Hjerting & Rahn 3900 
(K, LP). ee Do [un Máfioz, La Banda, Fabris 
1520 (LP). BOLIVIA. Cochabamba: Chapare, Quebrada de 
Colomi, Balls B-6253 (K, s Peu ae] 63 km de 
Cochabamba en dirección al poblado i, alrededores 
del cerro Pytuljata, meni po (LPB, US) Tapacarí, 
Qachi uñkata, cerca del apo 


Paz-Oruro, Krach 7492 (SI); cs Puerto Acosta, 10 km 
hacia La P ee al borde del R uaycho, cerca del L 
7712 (SI; Fray mut Pelechuco, Krach 


Viacha, ca m Viacha, Solomon & Nee 14233 (LPB, 
NY}; La Paz, Palca, zona basal del Illimani, Ceballos et al. 
548 (Sl); Pacala, La Paz e Corocoro y Topohoco, 


nO et al. 164 (SI); Los Andes, Kanton Peñas, Straße zur 

Fabulosa, Km 6.5 von der Abzweigung von der Strafe 

La Paz- figs Krach 8435 (SD; ela viciniis Sorata, 
inter Ancohuma et Turilaque, Mandon 276 ( 

Murillo, 4.6 km S jet. rd. to eae on rd. from M 

Alto (La Paz), Solomon 13177 (LPB); Nor Yungas, Canton 

Pocollo, La Cu mbre, Krach 8672 (SI); Omasuyos, Cantón 

Cruz 


roBes Moor oberhalb der Endnoranen oberhalb Canizaya, 
Freueler 4384 (SD; Tamayo, Ulla-Ulla, Okaria, Menhofer 
1600 (LPB). Oruro: Sajama, Cantón Lagunas, 


puna y 


vegetación alto andina, Loza de da Cruz 254 (LPB). Potosí: 
Frías, cerro Khare-Khare detrás de la ciu lte 9 
é María Avilez, P. 


Barclay & Juajibioy 7642 (US). P 
hwy. 105 to Chavin de Huantas, ca. D km E of Catac, King 
& Collins 9064 (US). Arequipa: Arequipa, just SW of Puno 
ep Paso del ar Iltis & pw 1489 (UC). Ayacucho: 
m N of Mataral, on trail to Rp b. 3663 (UC) 
d Üben. Zamalloa Díaz 78 (LP 


; Chumbivilcas, 


Hoogte & Rossel 2534 


(UC). Huancaya: 


Macbride & Featherstone 
oncepció 
Satipo, Saunders 1106 (K, UC); Junín, Capillacocha, a 20 im 
E of ue Tovar 390, 593 (LP); Llanos de Junín, 
0'S, Smith et al. 5652 (NY); Tarma, lago Junín, 


— 


do, Cajamar- 
guilla, López & Sagástegui did (LP); A Huia d 
Tajabamba, Ló 

Oise Jalca Quiruvilea, Lopez & Sagástegui 2888 (LP ). 
Lima: Canta, Carhuapampa (camino a VR Re ie 
Hear 162 (LP); Yauyos, Huancracha arriba de Tupe, Cer 

& Tovar 1174 (LP). Puno: Patanca, Fiebrig 3188 (BM, b 


3. Hypochaeris hohenackeri (Sch. Bip.) Domke, 
Notizbl. Bot. Gart. Berlin-Dahlem 13: 251. 1936. 
Basi Pic o Sch. Bip., 

Bonplandia 4: 54. 1856 : Peru. “Tobina in 
Cordill. sum. jug. os ii July 1854, W. 
Lechler 2111a (holotype, P!). Figure 6. 


MD m J. Kost., Blumea 5(3): 661, fig. 3c. 
945. a P. 


p : 
label “Teacota-Thal”], 4300 m, Oct. 1911, T. Herzog 
2425c (holotype, L, digital image!). 


Herbs to 4 em tall. Leaves oblong, 20-30 X 2- 


3 mm, base attenuate, margin entire or slightly 


dentate, glabrous. Capitula pedunculate to 20 mm 
rarely duh Involucre | d 11-15 X 6- 

m; phyllari - to 6-seriate, lanceolate, apex 
acute or dlightly gne margin ciliate; 


phyllaries 8-9 X 2 


com 
E 


m; paleae 11-1 ; florets 
per capitula. Corollas yellow, 
& Cabanillas 7212), 10-20 


7-12 mm; vascularization type 1; 


3-8 mm; ligule 


20.5 mm; anthers 8-12 mm; basal appendages type 
1 or 2, 0.50-0 3.5-8.5 mm 
opodium type 1 or 2; style 10-12 mm; style branches 


.75 mm; filaments ; anther- 


701 


Urtubey et al. 
The Hypochaeris sessiliflora Complex 
) 


Volume 96, Number 4 
(Asteraceae 


2009 


0,1mm 


H 


B C D 
Hypochaeris echegarayi. —A. Habit. —B—D. Phyllaries. —E. Shaggy trichomes on phyllaries. —F. Trichomes 
a, achene, and pappus. —I. Smooth cypselar wall. —J. 


Figure 9. 
(in cilia) on phyllaries. —G. Corolla with anthers and style. —H. 
Scaly cypselar wall. A-H, J from Valenzuela 998 (LPB); I from Hunziker & Caso 6039 (LP). 


Annals of the 
Missouri Botanical Garden 


2-2.5 mm; wall 
smooth; pappus 9.5-15 mm. Chromosome number 


1.5-2.5 mm. Cypselae erostrate, 


unknown 


Habitat and distribution. Hypochaeris hohenackeri 
occurs from Peru to Bolivia, 3200—4500 m, in humid 
grasslands, mountain pastures, riverbanks, and bogs 
associated with Juncus L. and Stipa L. (Fig. 4B, squares). 


Phenology. Hypochaeris hohenackeri flowers from 
January to September. 


Morphological characters. see hohenack- 


ri is characterized ong, sometimes dentate, 


E 
leaves, a ee head us € and dark, 


g I at the apex 
ervations. Achyrophorus hohenackeri [“Hohe- 
nakeri” | was published ces by Schultz Bipontinus 


(1855: 236) as a nomen nudum. Although without 
B mes e did cite a specimen (W. Lec 
la), which was later taken by him (1856: 54) a as 
valid 
description (1856: 54) in the same journal. Hypo- 


i. original specimen to accompany his 
chaeris hohenackeri Herzog Mia 1923) is an 
illegitimate name because no onym was provided. 

Although the type specimen of Hypochaeris parvi- 
flora J. Kost. is small, there is no doubt of the affinity 
of this 


accompanied by a clear illustration that shows the 


name; the original description is also 


diagnostic characters of this species. 


Representative 
Arani, near the 


specimens. BOLIVIA. | Cochabamba: 
summit of the pass through the cordillera 


EE "s d» betw. Poi 


Lewis 38124 (LPB); Saavedra, el camino de Charazani a 
Hayrapata, Menhofer X-1832 (LPB, SI). PERU. Amazonas: 
Chachapoyas, Chachapoyas-Celendín rd., cerro Calla-Calla, 
Smith & Cabanillas 7212 (US); cerro de Calla-Calla, betw. 
Leimebamba-Balsas rd. pass and the camino de herradura, 
Wurdack 1225 (LP). Huancavelica km 
from Conayca, Tovar 222 (LP). 


: Huancavelica, 4 


4. Hypochaeris mucida Domke, Notizbl. Bot. Gart. 
Berlin-Dahlem 13: 250. 1936. TYPE: Bolivia. La 
Paz: “Mittlere Anden: Südlich des Titicaca-Sees 
über Ancoraime,” 13,500 ft., 12 Feb. 1903 
W. Hill 299 (lectotype, designated here, K). 
Figure 7 


Hypochaeris mucida Domke var. integrifolia Cuatrec., Proc. 
Bi . Wash. 77: 156. 1964. TYPE: Peru. Puno: 
WSW fi Checayani, NE of Azangaro, oe 4150 m, 29 
Mar. 1957, H. Ellenber, e 495 bs lotype, U not seen; 
isotype, US not seen, Us photo IP) 


Herbs 1-1.5 em tall. Leaves oblong, 8-20 X 1.5- 
4 mm, slightly toothed or pinnatifid, adaxial sur- 


face sericeous with whip and some shaggy tri- 


chomes, abaxial surface with whip trichomes, 
petiole sericeous. Capitula sessile. Involucre cam- 
panulate, 8-12 X 9-10 mm; phyllaries 3- to 4-seriate; 
outer phyllaries lanceolate, to 12 mm, lanuginous 
with shagg p 
e indument or glabrous, margin ciliate. Paleae 


9-11 mm; florets ca. 12. Corolla 9.5-11.5 mm, color 


unknown; tube 3.5—5 mm; ligule 6—6.5 mm; vascu- 


trichomes; inne laries with the 
1 : 


larization type 1; stamens 9—10.2 mm, anthers 4— 
4.2 mm; basal appendages type 1, 0.8-1 mm; fila- 
ments 4.5-6 mm; antheropodium type l; style 8— 
10.5 mm, with branches 2-3 mm. Cypselae 5-ribbed, 
unbeaked, 2-3 mm; wall smooth; pappus ca. 8 mm. 
Chromosome number unknown. 


Habitat and distribution. Hypochaeris mucida is 
known from Peru (Puno) and Bolivia (La Paz). It 
inhabits the Andean steppe at 4150-4700 m, associ- 
ated with Pyenophyllum (Caryophyllaceae) (Fig. 4B, 
triangles). 


Phenology. nda mucida flowers from 


February to Apr 


Morphological characters. Hypochaeris mucida is 
the smallest species of Hypochaeris, growing to only 
1.5 em tall. The adaxial surfaces of the | 


phyllaries are tomentose (with whip and some shaggy 


eaves and 


trichomes), giving it a distinctive appearance. 


Observations. Cuatrecasas (1964) considered Hy- 
pochaeris mucida var. integrifolia to be unique in 
aving leaves entire to sinuate, but this admittedly 
extreme leaf shape sporadic and shows no 
geographic coherence. Hypochaeris mucida var. in- 
tegrifolia is not recognized. 
The holotype deposited at the Berlin herbarium was 
destroyed during World War Il; the isotype specimen 
at K is here designated as the lectotype. 


Specimens examined. BOLIVIA. La Paz: macho, 
Carabuco, Mina Matilde, Ceballos et al. 614 (SI); Tamayo, 
Ulla-Ulla, estribaciones de la cordillera de Apolobamba, 
Menhofer X-2197 (LPB). PERU. Puno: ENE of Checayani, 
Ellenberg 638 (US, photo LP). 


5. Hypochaeris eriolaena (Sch. Bip.) Reiche, 
116: 589. 1905. Basionym: 
Achyrophorus eriolaenus Sch. Bip., Bonplandia 4: 
54. 1856. TYPE: Peru. Puno: “Cordiller. pascuis 
sterilibus pr. Aza ” June chler 
1754 (lectotype, posuer by ports [1997: 
228], P!; isotypes, K!, P!). Figure 8. 


Anales Univ. Chile 


rg s cryptocephalus Sch. Bip., Bonplandia 4: 54. 

vH dd mu Ve (Sch. Bip.) Domke, 

dl we Dahlem a 251. a as 
E m . Puno: “pro 

abaya in cacumine NO A penn Tama? d a 


Volume 96, Number 4 Urtubey et 703 
2009 The d m sessiliflora Complex 
(Asteraceae) 


C D E E G l J 


Figure 10. h phila. —A. Habit. —B. Leaf. —C-E. Phyllaries. —F. Shaggy poesi on phyllaries. —G. 
Tite (in E on phyllaries. ET Corolla with anthers and style. —I. Palea, cypselae, and pappus. —J. Scaly cypselar 
A, C-J from Stuessy et al. 18064 (LP); B from Fabris 1369 (LP). 


Annals of the 
Missouri Botanical Garden 


1854, W. Lechler 1963 (lectotype, designated here, P'; 
w). 


isotypes, P!, 
eu spinneri Beauverd, Bull. Soc. Bot. Genéve, ser. 
14: 177. f. XIII. 1922 [1923] TYPE: Peru. 


2 s elica: Huancavelica, above Huancavelica “ 
pascuis ers " 4700 m, 1915, E. Godet 58 (holo- 
type, presumably NEU, not located). 


Herbs to 5 em tall. Leaves lanceolate or oblongate, 
25-70 X 5-16 mm, base attenuate in a broad petiole, 
margin toothed or entire, glabrous to sericeous (whip 
trichomes, to 35 mm) or with shaggy trichomes. 
Capitula sessile. Involucre campanulate or hemi- 
spheric, 10-30 15-25 mm; phyllaries 4- to 5- 
seriate; outer phyllaries ovate or lanceolate, lanugi- 
nous (with whip trichomes), 5 x mm; inner 

phyllaries ie to linear, T9 21 X 1.5-2.5 mm, 
lanuginous to glabrous, ciliate; paleae 15-28 mm; 
florets numerous. Corollas white or yellow to golden- 
yellow, 18-38 mm; tube 8-18 10-20 mm 


a type 2; stamens 11-22 mm; anthers 3— 


mm; ligule 


7 mm; basal appendages type 1 or 2, 0.7-1 mm; 

Ta 8-15 mm; pu PS type 1 or 2; style 

—26 mm, with branches 1.5-3 mm. Cypselae 5- 

a transversely MEN erostrate or slightly 

narrower at the apex, 1.5—4 mm; wall scaly; pappus 
9 mm. Chromosome number unknown 


Habitat and distribution. 
occurs in Peru and Bolivia in dry, stony Andean 


Hypochaeris eriolaena 


ridges, with scattered grass and limestone outcrops, 


2200—5100 m (Fig. 4C, squares). 


Phenology. Hypochaeris eriolaena flowers from 
May to November 


Co names. “Cebollana” (Peru, Angulo & 
López 1 358) “qachi tika" (Bolivia, Pestalozzi 606). 


Morphological characters. Hypochaeris eriolaena 
is clearly distinguishable from other members of the 
H. sessiliflora complex by long sericeous trichom 

broad petioles “This 


species often has no developed leaves when the 


ong 
plus lanuginous phyllaries. 


capitulum opens. 

Observations. Schultz Bipontius (1859) in his 
revision of Hypochaeris (as Achyrophorus) distin- 
guished H. 


abaxially sericeous phyllaries in the former and 


cryptocephala from H. eriolaena by 
niveous-tomentose phyllaries in the latter. Indument 
on phyllaries is a variable character in H. eriolaena, 
which can have eo with both lanuginous (with 
long whip trichomes) and matted indument; further- 
more, sometimes phyllaries in the inner series are 
glabrous. Because of this range of variation, we treat 


spinneri, the figure associated with the protologue 


leaves little doubt of the biological placement of this 
name 

Hypochaeris eriolaena var. hispida Herzog (Herzog, 
1923) is a nomen nudum appearing only in a list of 
plants from Bolivia. 


. Cochabamba: Ar- 
LPB); 


Representative gaa BOLIVIA 
a Tacopaya, Ibisch & Rojas 439 
Rancho, comunidad Ja 


a 


e la mina Kaluyo 
hacia la cumbre, Beck 11914 (LPB, SI). CHILE. Tarapacá: 
Cosapilla, Limani, Wevi 129 (LP). PERÚ. Ancash: Huaráz- 
Marcará, Velarde Núñez 3225 


rca: Cajamarca, Majada Pampa, Jalca Alta, Becker & 
Terrones 145 S ontumazá: ón (arriba de 
nt n Sagástegui, Alvitez & Mostacero 9006 (SI). 
Z nchis, en irecció al cementerio (parte 


ica: Huancavelica, Huan ta- 


nd Junín, Canne & Sole 232 (US) 
La Libertad: Otuzco, carretera a Shoreyo, ladera rocosa, 
Angulo & López 1358 (LP); Sánchez Carrión, summit above 
ht & 


a on rd. to Huamachuco, Hutchison, Wrig 


Capachica, Península Lake Titicaca, Tutin 1209, 1210 (BM). 


6. Hypochaeris meyeniana (Walp.) Benth. & 
Hook. f. ex Griseb., Abh. Kónigl. Ges. Wiss. 
Gottingen 19: 199. 1874. Basionym: Oreophila 
meyeniana Walp., Nov. Áctorum Aca aes 
Leop.-Carol. Nat. Cur. 19(suppl. 1): 292. 1843. 
Achyrophorus meyenianus (Walp.) Walp., Repert. 


1831 [one label on type, “4/31, ” presumably 
indicating Apr. 1831, but another label in 
another hand gives “1833”], F. J. F. Meyen 22 
(holotype, B [presumably destroyed], B photos 
LP!, MO!). Figure 11 


E x meyenianus (ap, Walp. var. ciliatus Wedd., 
And 2 


Calama,” June 1846, 
H. A. ‘Weddell 4017 (lectotype, designated by Bortiri 
[1997: 227], P!; isotype, P5. 

(Walp.) Benth. & Hook. f. ex Griseb. 


TS meyeniana 


4500 m, Feb. 1931 
isotype, LP). 
Hypochaeris meyeniana 


, E. Budin s.n. (holotype, LP!; 


(Walp.) Benth. & Hook. f. ex Griseb. 


var. brachylepis Cabrera, Fl. Prov. Jujuy 10: 677. 1978. 
Argentina. Jujuy: Yavi, "alrededores de La 
uiaca," T. Meyer, A. R. Cuezzo & V. Legname 21270 


(holotype, LP!; isotype, LP). 


Volume 96, Number 4 o 
2009 a sessiliflora Complex 
es raceae) 


5mm 


B C D E 

Figure 11. Hypochaeris meyeniana. —A. Habit. —B—F. Phyllaries. —G. Trichomes (in cilia) on phyllaries. —H. Cor 
with anthers and style. —I. Palea, cypselae, and pappus. —J. ~ cypselar wall. A-D, G-J from Stuessy et al. 18062 (LP); E 
from Meyer et al. 21306 (LP); F from Meyer et al. 21107 


Annals of the 
Missouri Botanical Garden 


Herbs to 8 em tall. Leaves lanceolate-elliptic or 
15-150 


pinnatifid), runcinate, base attenuate, margin toothed 


obovate, 5-40 mm, pinnatisect (rarely 


or entire, glabrous or hispid on the margin (or also on 

the midrib). ala sessile or rarely sim a very short 

peduncule. Involucre o to hemispheric, 8— 
x 8-25 mm; phyllar 


es 3- to 4-seriate; outer 
phyllaries lanceolate 7 


to pes 1-11 —7 mm, 
attenuate to rounded or acute, glabrous, hirsute 
(with shaggy trichomes) or lanuginous (with whip 


trichomes}, margin ciliate; mid e phyllaries lanceo- 
i 


late to ovate-lanceolate; inner phyllaries lanceolate, 
13-22 X 2-2.5 mm, Ei. slightly ciliate; paleae 
13-22 mm; florets ca. 25. Corollas yellow, yellow- 
orange, or golden-yellow, 13-24 mm; tube 7-11 mm; 
ligule 6-13 mm; vascularization types 1 and 2; 
stamens 10—22 mm; anthers 4—6 mm; basal append- 
ages types 2 and 3, 0.3-1 mm; filaments 6-16 mm; 
antheropodium type 1; style 11-22.5 mm, with 
branches 2-4.5 m 


apex, 2-10 mm; wall scaly or smooth; pappus 7— 


m. Cypselae slightly constricted at 


17 mm. Chromosome numbers 2n = 8, 16 (Diers, 
1961; Weiss et al., 2003; Weiss-Schneeweiss et al., 
2007) 


Habitat and. distribution. 
occurs from Peru to northwestern Argentina, on rocky 
humid places, 1700— 
4800 m (Fig. 4D). Hypochaeris meyeniana grows 


Hypochaeris meyeniana 
and/or sandy slopes, in dry and 
sympatrically with H. taraxacoides, but they occupy 
ifferent microhabitats, dry places versus bogs, 
respectivel 


Phenology. Hypochaeris  meyeniana flowers 


throughout the year. 


Common names. “Achicoria” (Argentina, Cabezas 
25, Cabrera 7723, 12121); “chambi” (Bolivia, Pesta- 
lozzi 206); “kawi kawi” (Bolivia, Pestalozzi 939). 


Morphological characters. Hypochaeris meyeniana 
is polymorphie, with phyllaries ranging from lanceo- 
ate to ovate, generally glabrous, sometimes with 
shaggy trichomes. Ás a result o se and other 
morphological variations, several varieties have been 
described within the species. Hypochaeris meyeniana 
var. brachylepis was described by Cabrera (1978) to 
have phyllaries wider (more than 5 mm) than those 
normally encountered in this species. Likewise, 
Hypochaeris (as Achyrophorus) meyeniana var. ciliata 
was regarded as distinct by Weddell (1857) for plants 
with pubescent phyllaries. Variety eriolaenoides was 
described by Cabrera (1957) for plants characterized 
by deeply runcinate leaves and tomentose phyllaries. 
None of these features shows a consistent geographic 
pattern to warrant varietal status, and hence we treat 


H. meyeniana as a taxon in a broad sense. 


Our field studies reveal that 
the conservation status of Hypochaeris meyeniana is 
very good. 
both dry and humid rocky places in puna and prepuna 
(with Trichocereus pasacana (F. A. C. i Britton 
& Rose, 


Conservation status. 


e found many and large populations, in 


“cardon,” Cactaceae Juss.) in Jujuy ipea 
Paz Department, Bolivia. 
Cochabamba Department, Bolivia, it occurs in e 


Argentina, and in La 


lepis tomentella Wedd. (“tabaquillo,” Rosaceae Juss.) 
ods. 


Observation. The type of Achyrophorus meyenia- 


nus was probably collected in Chile. See comments on 
this same locality un 'ypochaeris taraxacoides. 


Bortiri (1997) placed Hypochaeris 


var. ciliata into synonymy of H. 


meyeniana. 
sessiliflora, a 
referral with which we do not agree. Description of 
"feuilles roncinées" suggests placement here in 
meyeniana. 


Representative ee nae ARGENTINA. Catamarca: 
Andalgala, Subida. al cemo Yutuyaco desde Capillitas, 
Sleumer 2730 
río Blanco, arriba de Granadilla, Sleumer & Vervoorst 2589 
(LP, UC). d Cachi, camino a Mina Don Otto, 4—5 km al 
S de la ruta a Cachi, Novara 10644 (MCNS); Moor us 
Abra io. Vignati 445 (LP [2], NY}; Humahuaca, 20.7 km 
rer o 


18068 (LP 
Ancibo 


9 (LP); entre Cachi y San Carlos, camino a 
in id. cruce de la R 33 a la Care del Sant 
a Fi n Díaz 10905 (MCNS); San 
Isonza, 20 S de Piedra del Molino, rms 
9781 (MCNS); o dus e al pie del d Tuzgle, 
Werner 1, 103, 131 (LP); Tilcara, abra de Yala, Fabris, 
Crisci & Petriella el (LP); Tumbaya, Stuessy, Urtubey & 
duis xad 18062 (LP [3], WU); Valle Grande, Santa Ana, 
Cerro Ronque, Kiesling et a 
Cajas, lea 7826 (LP); Chalguamayor, 15-2 


a & Neumann 


Yavi, P [^ Km 35-38, Tolaba, neos Arapa, 
Gutiérrez, Quir mallo da $ 77 

CNS); Abra eee siete de RP 5, Km 40-45 
(camino a Santa Victoria pasando quebrada de Cajas, 
Tolaba, Acufia, Arapa, iérrez, Quiroga, Ragno, Ramallo & 
Da Sila 1652 (MCNS). La Rioja: Famatina, entre Los 


Corrales y Cueva de Pérez, Ghee et al. 27174 (SI). Salta: 
Iruya, Loman, alrededores del pueblo de Delfina Diaz, 2 
al W de San Isidro, 14 km al NW de Iruya, Tolaba, Ragno & 
Quiroga 1235 (MCNS); Pantipampa, — del puesto 
de Concepción Bustamante, filos del cerro al N de San Isidro, 
Iruya, Te Rae, & Quiroga 1198 
(MCNS); Orán, Caucillar, Hilgert & Lamas 1654 (MCNS); 
Poma, Abra Mufiafio, Cabrera 8982 (LP); Santa Victoria, RP 
7, Km 12, Deginani, Cialdella & Bortiri 806 (SI); Nazareno, 
200 m al S E pueblo, id a (MCNS); San Tas Arm 
mblayo, valle Ison: 0 km S de valle tado, 
Novara 6356 (MENS), onus Chuquisaca: TE, near 
Cochabamba, Balls 6232 (K). Cochabamba: Apopaya, 
cuenca Río Tambillo, estancia Pajchanti, Baar 386 (LPB); 
comunidad de Kutimarca y 
Ps e 2181 (K); Cha- 
pane Cerro Guakanquí, Steinbach 9696 (K); Cochabamba, 
as, odds 4347 (US); Quillacollo, camino Sipe-Sipe- 


a 


Volume 96, Number 4 
2009 


e 
ng aim sessiliflora Complex 
hee raceae) 


707 


Lipichi, Hensen 789 (LPB); irene Sacha loma, Ramirez 
203 (LPB). La Paz: about 50 km SE of 


uisivi, 


Illimani, o um ies 0 Rs Omasuyős, el camino 
principal a Peña; a Fabulosa, Hichukkota, 
Beck 2861 (LPB, is a Paz, aa, Asplund 4886 
a rd. t 


TA 


Stuessy, Tremeisberger «e Hinge 18504 (LP); Chayanta, 
Cruz Kasa, Zamora 156 (LPB); Frias, Potosi, “Las 
Lecherias,” Schulte 129 (LPB). Tarija: Aviléz, lagunas de 
Tajsara, Meyer, Cuezzo & Legname 21517 (LP); Méndez, 
cerca Pasajes, Bastián 334 (US); Tarija, de Tarija a Iscayada, 
Kiesling et al. 3801, 3826 (SI). CHILE. Región l: Arica, 
Chapiquiña, Ricardi, 


Werdermann 247 (LP [the label of this collection bears the 
herbarium name, m. choeris glacialis Weder., which was 
never published]. PERU. Ancash: d Ocros, m 
6025 (SI); Huaylas, p Pamparomas, Path Karka to lagun 


(arriba d azá) Sagástegui et (US) 
Huane ea: Castrovirreina r in Meical, 
30280 (US); Huancavelica, Occopampa, entre Laria 

Tambopata, a 25 km de Conaica, Tovar 842 (LP). Junín 


Huarochiri; Distr. San Mateo, Río Rimae, pe 822 K). 
Puno: Huarochiri, Distr. Matucana, 85 on hwy., 
Saunders 379 (BM). Tacna: 20 km Mss Candasave on 

Mazo Cruz just S of Volcán Tutupaca, Weigend & 
Fórther 97-790 (K). 


7. Hypochaeris echegarayi Hieron., Bol. 
Nac. Ci. 4(1): 51. 1881. TYPE: Atge ntina. ua 
Juan: Barriales de Leoncito, Dec. 1875—Jan. 
1876, D. S. Echegaray s.n. (holotype, CORD, not 
located; isotype, LP!). Figure 9. 


Achyrophorus setosus Wedd., Chlor. Andina 1: 220. 1857. 
ipochaeris setosa (Wedd.) Rusby, Bull. New York Bot. 

Gard. 4: 402. 1907, non Hypochaeris setosa Formánek, 
1897. TYPE: Bolivia. Potosí: “environs de Potosi," 15— 

28 Mar. 1833 [Papavero, 1971], A. C. V. D. d'Orbigny 
1425 (lectotype, e by Cabreta [1978: 674], P!; 


$ 
E 
se 
= 
a 


, Blumea id 660, fig. 3a, b. 
. TYPE: Bolivia Onno: “am cerro de Oruro auf 
” 3800 m, Ne 1911; T: 6. 
J. Herzog 2522c n: L, digital uU 
Hypochaeris meyeniana (Walp.) Benth. € Hook. f. ex Griseb. 
var. leucantha Cabrera, Revista Deni A 11(4): 


410. 1957. TYPE: wis Jujuy: Yavi, Quebrada de 
Cajas, 4000 m, b. 1943, A. L. Cabrera 7837 
(holotype, LP!). 


Herbs to 5 em tall. Leaves lanceolate or oblong, 20— 
70 X 3-10 mm, apex acute, base attenuate in a broad 


petiole, ee uid to pinnatisect, runcinate, margin 
entire toothed, hirsute to glabrous. Capiti la 
a (to 2 cm), rarely sessile. Involucre 


campanulate, 10-25 X 8-20 mm; phyllaries 3- to 4- 
seriate; outer phyllaries lanceolate, 7-11 X 2-2.5 mm, 
apex acute or semiobtuse, hirsute on middle nerves 
(many to few shaggy trichomes), margin hyaline; inner 
phyllaries oblongate or lanceolate, 12-14 X ca. 2 mm, 
hirsute (shaggy trichomes) to glabrous; paleae 7— 
13.5-20 mm; 


scularization type 


7 mm; florets 10 to orollas white, 
tube 7-10 mm; ligule 6.5-10 mm; va 
l; stamens 14-17 mm; 5 mm; basal 
appendages type 2, 0.6—0.8 mm; filaments 9-12 mm; 


anthers ca. 


antheropodium type 1; style 12-15 mm, branches 2— 
5-ribbed, unbeaked or slightly 


narrower at the apex, 2-6 mm; wall scaly or smooth; 


mm. Cypselae 


pappus 12-14 mm. Chromosome number unknown (a 
possible hybrid with Hypochaeris meyeniana, however, 
has been reported as 2n = 16; Weiss-Schneeweiss et 
al., 2007) 


Habitat and. distribution. 
occurs in t 


Hypochaeris echegarayi 
e mountains of Peru to northwestern- 
western Argentina (to San Juan Province), in bogs or 


dry places, 3100—4950 m (Fig. 4C, triangles). 


Phenology. Hypochaeris echegarayi flowers from 
December to Marc 


Common name. 


194). 


“Qausilla” (Bolivia, Pestalozzi 


Morphological characters. Hypochaeris echegarayi 
has shaggy trichomes on the leaves, phyllaries, and 
corollas. This species is closely related to H. 


eremophila, which has yellow corollas. 


Observations. | Achyrophorus sessiliflorus (Kunth) 
DC. var. subruncinatus A. Gray (1861: 146) is an 
"ar name that included both A. setosus We 
eriolaenus Sch. Bip. as synonyms. 
a by Gray (1861) corresponds more closely 
to the former than to the latter. 

During 1874 and 1883, Hieronymus worked in the 
herbarium CORD of Argentina (Stafleu & Cowan, 

9), and his main collection is deposited there. In a 
thorough search of that herbarium, we have been 
unable to locate holotype material. 

Cabrera (1957) originally described Hypochaeris 
meyeniana var. leucantha as new based on its white 
flowers. In 1978, he relegated this name to synonymy 
under H. echegarayi Hieron. 


Annals of the 
Missouri Botanical Garden 


The small specimen of Hypochaeris ornata, plus the 
clear figure that accompanies the protologue, showing 
diagnostic features of campanulate involucre, setulose 
phyllari 
HOMES The protologue (Koster, 1945) des not 


s, and runcinate leaves, allows 


mention the color of the corollas, and the colors of the 
specimen are faded. 


Representative specimens. ARGENTINA. Jujuy: Cochi- 
Mina Aguilar 


, Mina Pirquitas, Quebrada Conadera, Schwabe, 
Ancibor & usns 909 (LP); Santa Catalina, ciénaga cerca de 
Cienaguillas, Cabrera. et al. 15393 (LP, N 


. 21505 (LP); alrededor d 

Hunziker Caso 6039 (LP). San E Iglesia, E San 
Guillermo, Los Caserones, Cazal, Pujalie & Reca 17 (SI). 
Tucumán: Tafí, Cumbre de Chagtayil. Olea on (UC). 
BOLIVIA. Cochabamba: Tapacarí, 2 km al E de Japo K'asa 
(Km 125 d Iu cero Kampiyani, Pestalozzi 
194 (LPB); 42 km from Car n the rd. to Cochabamba, 
King & Bishop 7530 (US. us P 
(LPB); Juan Bautista Saavedra, Mergel- “Kalksandštein: Horn- 


2 
Pedro Domego 
Murillo, vic. of lago Zongo at the head of the ae valley, 
d 13153 (LPB, NY); Presa Incachaca, 10 


pide 18516 (LP 
rom La Paz), valle do ER Stuessy, Tremetsberger & 
Hóssinger 18525 (LP); Franz Tamayo, Puyo-Puyo (Ulla-Ulla), 
Menhofer X-1880 (LPB); Pacajes, entre Corocoro y Topohoco, 
Ce s et al. 180 (SI). Oruro: Sajama, Cantón Lagunas, 

de la Cruz 253 (LPB). Potosí: N de Potosí, entre Bolivar 
y ‘Gri. Casas 8025 (NY). PERU 
Huarichori, b aa d M rd. to Ratontay, 
Sanderman 855 (K). P , San Antonio de Esqui- 
lante, minas a San eae. D RA 3915 (K). 


. Lima: 


8. Hypochaeris eremophila Cabrera, Notas Mus. 
La Plata 13: 22. 1948. Replacement name for 
Distoecha taraxacoides Phil., Anales Mus. Nac. 
Santiago de Chile 8: 37, tab. 2, fig. 2. 1891, non 

Distoecha taraxacoides Ball, 1885. TYPE: Chile. 
pu III: Copiapó, Colorados  [26^59'S, 
6856'"W; Muñoz Pizarro, 1960: 171], 3600 m, 
15 Jan. 1885, A. Philippi s.n. (lectotype, 
designated here, SGO 65208!, SGO photo LP!, 
SGO digital image!). Figure 10. 


Herbs to 7 em tall. Leaves lanceolate or obovate, 
20-70 X 
glabrous or hispid principally on the midrib, base 


5-12 mm, generally deeply runcinate, 


attenuate in a broad petiole, margin toothed or entire. 
rarely 
13-18 


18 mm; phyllaries 3-seriate, lanceolate, apex acute, 


Capitula pedunculate to 30 mm, sessile. 
x 


Involucre campanulate-cylindrical, 


margin with shaggy trichomes to 7 mm long; outer 


phyllaries 10-14 X 2.5—4 mm; inner phyllaries acute 
or e obtuse, glabrous or only slight trichomes, 
11-14 mm; florets 15 to 25 
Corollas ellos to yellow-orange, 12-20 m 
10 mm; 
stamens 11-18 mm; anthers 5-6 mm; basal append- 

s type 2 or 3, 0.8-1 mm; filaments 6-12 mm; 
pm is type l; style 10-14 mm, with branch- 


-17 X mm; paleae 


ligule 7-10 mm; elas: type 1; 


es ca. 3 mm. Cypselae 5-ribbed, transversely wrin- 

kled, erostrate, narrowed at the apex, 2.3-4 mm; wall 
scaly; pappus 7— 
( 


mm. Chromosome number 2n = 8 
Weiss-Schneeweiss et al., 7 


Habitat and distribution. Hypochaeris eremophila 
occurs from southern Peru to northwestern Argentina, 
inhabiting dry places, bogs, and “mallines,” 28 


4700 m (Fig. 4C, stars). 


Phenology. i gn eremophila flowers from 
November to 


mes. “Achicoria” iei] Cabrera 


En “eéndor siki" (Bolivia, Meneses P602 


Morphological characters. Hypochaeris eremophila 
can be distinguished from the other acaulescent 
species of Hypochaeris by its peduncules with 
cylindric-campanulate with 


shaggy trichomes on the 


(sometimes yellow-orange) corollas. 


Cabrera (1948: 22) de the name 
choeris] eremophila when he moved 


involucres, phyllarie 
abaxial >. and ye 


Observation. 
Hypochaeris 
Distoecha la into paar because be 
taraxacoides Ball was already occupied. S 


comments under H. taraxacoides in this treatment. 

ufioz Pizarro (1960) listed two original specimens 
of Distoecha taraxacoides from the type locality in 
Philippi’s herbarium at SGO. We have selected 
$GO65208 as lectotype because it is the better of 
the two specimens. 


Representative specimens. ARGENTINA. o 
Belén, e a Antofi 
(SD. Jujuy: Susque: 
Tocomar, Cabrera 8278 (LP); al pie del Cerro Tuzgle, Cabrera 
8399 (LP); Tumbaya, 33.3 km W Purmamarca, Stuessy, 
Urtubey & Mae 18064 "ri ED; Valle Grande, subida 
a cerro Amarillo desde Alto de € , Kiesling, Ulibarri & 
López 1720 (SI). Salta: ae Com de Cachi, Spegazzini s.n. 
); Gua Von Pampa MEC s.n. (LP); Pastos 


Capella, Pisano & Venturelli 1726 (LP). PERU. Puno: Grau, 
Chuquibambilla, Pennell 13362 (US). 


9. Hypochaeris acaulis Rémy) Britton, Bull. 


(J. 
Torrey Bot. Club 19: 371. 1892. Basionym: 


Volume 96, Number 4 Urtubey et 709 
2009 The d um sessiliflora Complex 
(Asteraceae) 


2.5mm 


0.1 mm 


d. 


J 


G 


Figure 12. Hypochaeris acaulis. —A. Habit. —B—F. Phyllaries. —G. Tric 
mn and style. —I. Palea, cypselae, and pappus. —J. Scaly cypselar wall. From López 1825 


B Cc D 


homes (in cilia) on phyllaries. —H. Corolla with 
(LP). 


Annals of the 
Missouri Botanical Garden 


75° 


a 3 E SA, 

35 Pd AS de cr 
d ^ 
TM Jj 
a / 

NY / 
4 no m 
K 


80° 


75 


70° Delineó: V.H.Calvetti 


Figure 13. Distribution of Hypochaeris acaulis in the southern Andes (Chile and Argentina). 


Achyrophorus acaulis J. Rémy, in Gay, Fl. Chil. 
3: 448. 1847 [late 1848 or early 1849; Stafleu & 
Cowan, Toxo; TYPE: Chile. Region VI: Cacha- 

en los prados pantanosos de las 
cordilleras de Ten ue, provincia de 
Colchagua,” 5-7 Feb. 1831 [Muñoz Pizarro, 
1944], C. Gay s.n. (holotype, P not seen, P photo 
SI!; isotype, K!). Figure 1 


Herb to 5 em high. Leaves oblongate, 32-60 X 12— 

20 mm, pinnatisect, base 

petiole, margin entire or toothed, 
shaggy trichomes (also on 


with divisions antrose 
attenuate in a broad 
with 


some nerves) or 


glabrous. Capitula sessile. Involucre campanulate or 


margin ciliate; paleae 13-16 mm; florets ca. 40. 
Corollas yellow, 10-14 mm; tube 5—6.5 mm; ligule 5— 

mm; vascularization type 1; stamens 7-9 mm; 
anthers ca. 2 mm; filaments 5-7 mm; basal append- 
ages type 1, ca. 0.3 mm; antheropodium type 3; style 
7-10 mm, with branches ca. 1.5 mm. Cypselae 5- 
ribbed, transversely wrinkled, 3.5-8 mm, 
wall scaly; pappus 10-15 mm. Chromosome number 


2n = 8 (Wulff, 1998; Weiss et al., 2003). 


Habitat and distribution. Hypochaeris acaulis 
grows only in the southern Andean Cordillera across 
a range about 600 X 150 km, from the regions of 
Santiago Metropolitan to Araucania of Chile and 
provinces of 


rostrate; 


endoza to Neuquén of Argentina 


(Fig. 13). It occurs between 1430 and 3000 m in 


bogs, marshy places, humid meadows, stream banks, 
and Araucaria araucana (Molina) C. Koch woods. It 
also can be found occasionally in disturbed areas near 
roads and buildings. 


Phenology. Hypochaeris acaulis flowers from De- 
cember to March. 

Morphological characters. Hypochaeris acaulis can 
be clearly distinguished from the other acaulescent 


species of t s idely te outer 
phyllaries, yellow corollas slightly longer than th 
voluere rostrate cypselae metsber- 


ger et al. (20032) have completed a molecular (AFLP) 
study of seven populations of H. acaulis from Chile and 
Argentina. All data point to restricted gene flow among 
populations, probably due to limited dispersal capa- 
bility. There is also the expected positive correlation 
between genetic variation and population size. The 
species is postulated to be inbreeding (autogamous), 
based on pollen:ovule ratios, which would be consistent 
with the observed pattern of genetic variation. Based on 
other AFLP data (Tremetsberger et al., 2006), the 
evolutionary relationships of this taxon appear to reside 
with H. palustris and H. tenuifolia and not with the 
other acaulescent members of the South American 
species of the genus. appears to be a rather 
dramatic case of parallelism of ea features 
for survival in the high Andes. 


Representative specimens. 


ARGENTINA. Mendoza: Ma- 


extremo S laguna Varvarco Campos, cajón del e. 


Volume 96, Number 4 
2009 


pr 
egere sm sessiliflora Complex 
na raceae) 


Boelcke et al. 14111 (LP, A Norquín, Copahue, Cabrera 
1440 hue, Cabrera 6154 (LP [2]; 
1 chado, small laguna, 
Stuessy a Baeza 15593 (CONC, WU); Zapala, Comber 1257 
(K). CHILE. Región VI: Cachapoal, Rancagua, Cordillera 
P) A : 


mi 
Finot & ie 1825 (CONC). Región VIII: Ñuble, valle de 
Las Nieblas (near Termas de Chillán), Stuessy & Baeza 
ca. l km N of 


17 (CONC, WU); Cautin, M Las Raíces, 
oe Burkart 9528 (SI). R n Metropolitana: 
Santiago, Valle Largo, Mia; "bei Las 
Hochkordillera, C. & G. Grandjot s.n. (LP) 


Condes 


Literature Cited 

Babcock, E. B. 1947. The genus Crepis. Part two. Systematic 
treatment. Univ. Calif. Publ. Bot. 22: 199-1030 

Baeza, C. M., J ORE E. Vosyka, TE ea ES H. Weiss, 


L. (Asteraceae) de Chile. Gayana, Bot. 57: 105- 
Bentham, . Hooker. 1873. Hypochaeris. 19- 
7 D. Hooker (editors), Caan 
Plantarum, Vol. 2. Reeve & na London. 
i, E. ; 8 Hypochoeris (Compositae, 
Cichorieae) de la Argentina. “Hickenia 2: 223-232. 
1999. 280 


———. ED Asteraceae, parte 14. Tribu XIII 
actuceae: Hypochoeris. Pp. in Espinosa 
(coordinator) Flora Fanerogámica Argentina, Fasc. 6. 


Programa PROFLORA (CONICET), Córdoba. 

Cabrera, A. L. 1948. Compuestas nuevas del noroeste de la 
Argentina. Notas Mus. La Plata 13: 7-23. 

———. 1957. La vegetación de la República Argentina. VI. 
La vegetación de la Puna Argentina. ee Invest. Agric. 
11: 317 e 

—— dis A Pp. 671-677 in A. L. Cabrera 
(editor), s de Provincia de Jujuy, Parte X. 
Compositae. INTA, mu Aires. 

Cuatrecasas, J. 1964. a on Andean Compositae: VI. 
Proc. Biol. Soc. s 77: 127-156. 

Diers, L. 1961. Anteil an Polyploiden in den 
uc m pm westkordillere Perus. Z. Bot. 49: 
437-488. 

Division of Geography (Department of the Interior). 1955. 
NIS Gazeteer: Bolivia. Central Intelligence Agency, 

Washington, D.C. 

Enock, C. R. 1908. Peru: Its Former and Present Civilisation, 
History and ide a Topography and d 
Resour General Development. 

Fuchs, C. 1963. Fuchsin staining with NaOH clearing for 
lignified elements v whole plants or plant organs. Stain 
Technol. 

Goloboff, P. A. jn Pee-Wee m and 
documentation: Willi i Society web site. <http:// 
WWW , accessed 2 Septem- 


61. Characters of some Composi tae in the 


Wilkes; with observations, etc. 


Proc. Amer. Acad. Arts 5: 114—146. 


Harris, J. G. & M. W. Harris. 1994. Plant Identification 
d An oe Glossary. Spring Lake Pub- 
lishing, Spring Lake, Utal 

Herzog, T. 1923. Die an der bolivischen Anden 
und ihres östlichen Vorlandes. Wilhelm Engelmann, 


Leipzig. 

Hieronymus, 1896. Plantae Stuebelianae novae. Engl. 
Bot. Jahrb. Syst 21: 306-378. 

ou E 893. o Pp. 361-363 in A. Engler 


. Prantl (editors), D iirlichen Pflanzenfamilien, 
vol. 4 e Wilhelm ed Leipzig. 

Jansen, R. K. & T. F. Stuessy. 1980. Chromosome AO a 
Compositae from Latin America. Amer. J. Bot. 
585—594. 

Koster, J. T. 1945. Plantae a Th. Herzogio initinere eius 
Boliviensi frin annis 1910 et 1911 collectae, Composi- 

tae, Cichorieae. Blumea 5: 659—661. 

Kunth, C. S. 1818. Compositae: Hypochaeris. Pp. 1-2 i 
Humboldt, A. Bonpland & C. S. Kunth (editors), Nova 
Genera et Species Plantarum (quarto ed.) 2. Sumptibus 

aris 

Revisio Generum Platos Pt. 3(2). 

Arthur Felix, leipa: 

J., F. R. Barrie, H. M. Burdet, V. d D. L. 

o, P. C. 

sJ H: Wiersema & N. Je Tarland (editors). 

6. Intermadonal e of Botanical Nomenclature 
(Vienna Code). Regnum Veg. 146. 

Metcalfe, C. R Chalk. 1950. Anatomy of the 
Maus oua Vols. 1 and 2. Clarendon Press, 

— & — 1979. Anatomy of the D 
Vol. 1. Clarendon Press, Oxford. 
i. C. 1802. Supplementum ad Methodum Plantas. 
Marburgi Cattorum. 
Morrain, S. A. 1984. Systematic and Regional Biogeography. 
k. 


Oxford. 
icotyledons, ed. 


- Baeza. 
2005. Pleistocene refugia an recol 


uñoz a C: 4. El So de don Claudio Gay. 

Bol. Mus i st. M 2: 27 

Bs Espécie; RD tas s Denias por R. A. 
Philippi en el Siglo XIX. Universidad de Chile, Santiago. 

Nixon, K. C. 1999 . 0.99+ (BETA). <hitp:// 

cessed 2 September 2009. 


Olsen, J. 19 rd number reports LXVII. 
Taxon 29: 347— 
apavero, N . Essays on the History of Neotropical 


Dipterology, with, Special Reference to b Calenus (1750- 
ee Museu de Zoologia, Universidade de São Paulo, São 


mo id 1971. The Control of Recombination. Ph.D. 
hesis, Vire of Oxford, Oxford. 
. C. 1843. ea Lebenslauf. Nov. Actorum 
ur. 19(suppl. 1): xiii—xxxii. 
Reiche, K. 1905. Estudios críticos sobre n flora de Chile. 
Anal. pu Feed 116: 575-606. 
Samuel, R., T. F. Stuessy, K. Tremetsberger, C. M a € 
S. Sil ak iis lev. uM Ph ylogenetic Bt 
among species of Hyp s (Asteraceae, Cichorieae) 
ae on ITS, plastid las intron, irnL-F spacer, and matK 
ces. Amer. J. Bot. 90: 496—507. 
ni N. Y. 1926. Humboldt and AR itinerary in 
Ecuador and Peru. de Bull. 1926: 181-19 
d M UE E 
m Acad. Cae: 


1845. uem ae Nov. 
es. a -Carol. Nat. Cur. 21: 86-172. 


Annals of the 
Missouri Botanical Garden 


. 1855. Lechler's Plantae Peruvianae. Bonplandia 3: 


— ——. 1856. Lechler’s rud Sammlungen aus Peru und 
Chile. fem wa 0-54. 
. 1859. Revisio critica generis Achyrophori. Pollichia 
16-17: 45-13. 
Stafleu, F. A. & R. S. Cowan. 1976. Taxonomic Literature, 
ed. 2, Vol. 1: A-G. re aa Venid ema, Utrecht. 
& omic Literature, ed. 2, vol. 2 
H-Le. T ais Scena & “Holkema, Utrecht. 
axonomic Literature, ed. 2, vol. 3: 
ES O. E Dos & Holkema, Utrecht; W. Junk. 
The H 


axonomic Literature, ed. 2, vol. 7 


1988. T. 
us Z. Bohn, Sel Séheltema:& Holkema, Utrecht; W. Junk, The 


see, *. G. 1841. Nomenclator Botanicus, ed. 2, Vol. 2. J. 
G. Cottae, Stuttgart. 
ms D. d 991. Wed acne relationships of 


Andean of Cre us to Pliocene age. 
ees on matol. a 88: 69-84. 
Eon K., T. F. Stuessy, Y.-P. Guo 

Weiss & R. M. Samuel. 2003a. Amplified fragment length 

polymorphism (AFLP) variation within and among popu- 

lations of Hypochaeris acaulis (Asteraceae) of Andean 

southern South America. Taxon 52: 237-24, 

,R.M.S . M. Baeza & M. F. Fay. 
2003). Genetics of elector in Hypochaeris tenuifolia 
(Asteraceae, Lactuceae) on Vol 
Molec. Ecol. 12: 2649-2659. 
H. Weiss-Schneeweiss, T. Stues 

M. A. Ortiz 


amuel, 


cán Lonquimay, Chile. 


ssy, R. Samuel, 
. 2005. Nuclear 


origin o th American Hypochaeris (Asteraceae, 
Cichorieae}. Molec. Phylogen. Evol. 35: 1 A 

$ sy, G. Kadlec, E. Urtube M. Baeza, 
S. G. Beck, H. A. Valdebenito, C. de Fátima Ruas & N. I. 


Matzenbacker. 2006. AFLP phylogeny of South American 
species of Hypochaeris (Asteraceae, Lactuceae). Syst. Bot. 
10-626. 


Uphot J. C. 1962. arce Vol. 4(5). Pp. 1-206 in K. 
Linsbauer (editor, Handbüch der Pflanzenanatomie. 
Gebrüder Borntraeger, jo 

Walpers, W. G. 1846. Repertorium Botanices Systematicae, 
Vol. 6(2). F. Hofmeister, Leipzig 

Weddell, H. A. 1857. Chloris Andian, Vol. 1, Pt. 7. P. 
Bertrand, Paris. 

Weiss, H., T. F. , J. Gra C. Bae 3. 

hromosome ru 2 South pos pocitos 

(Asteraceae). Ann. Missouri Bot. Gard. 90: 56-63. 


Es p Hed dd S. Spe LAE. C. 
za & J. Parker, 2003. K in South 
Anerem ee es of EUR Tes Lactuceae). 


Pl. Syst. Evol. 241: 


. Stuessy, P d ee E. Urtubey, H. 
o. S: Beck & C. M. Baeza. 2007. Chromosome 
numbers and karyotypes of South American species and 
populations of Hypochaeris (Asteraceae). Bot. J. Linn. Soc. 
153: 49-60 


, K. Tremetsberger, G. M. Schneeweiss, J. S. Parker 
& T. F. Stuessy. 2008. Karyotype diversification and 


evolution in diploid 


and eae e Mp data. bene Bot. 101: 909-918. 
Wall, A. F. 1998. E n Asteraceae. VIII. 
nie 35: v 


APPENDIX 1. Morphological characters and character states 
used for cladistic analysis of the Hypochaeris sessiliflora 
complex. See also Figure 1. 


l. Habit: caulescent or peduncle longer than leaves (0); 
acaulescent or peduncle never longer than the leaves 
(1). 

2. Stem: ae or ne ud ramified (1). 

3. Leaf sha anceolate (0); linear (1); obovate (2); elliptic 
(3); linear- loreet (4); linear-filiform (5). 

4. Leaf margin: dentate (0); entire (1); divided (2). 

5. Leaf vestiture: shaggy (0); whip (1); shaggy and whip 
glabrous 


e 


2 


6. Tiealuerer: mpanulate (0); hemispheric (1); cylindric 
(2); aa: eae e (3). 
7. Capitulum: pedunculate (0); sessile 


e (D. 
llary shape: linear-lanceolate (0); lance- 


axial): whip and shaggy tri- 
0); whip trichomes (1); shaggy trichomes (2); 


— 


10. Length af corolla: longer than the involucre (0); as 
long as involucre 
11. Corolla vascularizatioi type 2 (0); type 1 


14. Cypselar shape: rostrate (0); semirostrate (1); erostrate 
15. Copeclis wall: scaly on all surfaces (0); smooth (1); 


scaly on main portion onl 
16. Palea surface: pubescent (0); glabrous (1). 


Volume 96, Number 4 


o 


2009 cs sessiliflora Complex 
hes raceae) 
APPENDIX 2. Data matrix used in the cladistic analysis of the Hypochaeris sessiliflora complex. Dashes indicate missing 


data or inapplicable states. 


1; 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 
Outgroup 
H. argentina 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 
H. caespitosa 0 - 4 01 3 0 0 0 1 0 1 0 1 1 1 0 
H. chillensis o X 0 02 13 3 0 0 013 1 1 1 1 0 0 0 
H. elata 0 01 0 0 3 0 0 0 0 HOR: oh 0 0 0 
H. hookeri 1 - 5 01 13 3 0 0 1 0 1 1 0 2 1 0 
Ingroup 
H. acaulis 1 - 1 2 03 01 1 2 3 0 1 1 0 2 1 
H. echegarayi 1 - 01 2 03 0 01 1 2 0 1 0 0 12 01 1 
H. eremophila 1 - 02 02 03 3 01 1 2 0 102 0 02 0 1 
H. eriolaena 1 - 01 Ol 013 OL 1 12 1 0 0.0102 12 O0 1 
H. hohenackeri 1 - 1 01 3 0 0 1 3 0 1 0102 2 1 1 
H. meyeni 1 - 023 2 03 01 01 12 123 0 0102 0 1 01 1 
H. mucida 1 - 1 02 2 0 1 1 0 0 1 1 0 2 ] 3 
H. sessiliflora 1 - 34 0 3 01 01 012 23 0 102 02 12 O 1 
H. taraxacoides 1 - 01 02 3 2 0 0 0 0 2 0 2 l1 1l 
APPENDIX 3. Vouchers for the cladistic analysis. 3. Hypochaeris hohenackeri Bip.) Domke 
4. Hypochaeris mucida Dom 
up 5. Hypochaeris a (Sch. Bip.) Reiche 
chaeris argentina: Burkart et al. 10486 (LP); Krapo- 6. ex 


vickas 7806 (LP); Urtubey & Tremetsberger 146 (LP, WU) 
H. caespitosa: Castellanos s.n. (LP); Hunziker 8649 (LP); 
Urtubey & Tremetsberger 145 (LP, WU 
chillensis: Cabrera 5688, 8595 (LP); Urtubey & 
Trenetsberger 114 (L 
a: Cabrera et al. 16971, 17657 (LP); Stuessy et al. 
Td 78083 (LP, WU) 

H. hookeri: Stuessy et al. 18019, 18044 (LP) 

Ingro 

H. od: Boelcke et al. 10379 (LP); Cabrera 6154 (LP); 
López 1825 (LP 

H. echegarayi: Hunziker & Caso 6039 (LP); Cabrera et al. 
15393 E Meyer, Cuezzo & Legname 21305 (LP); Stuessy et 
al. 1802. ds n 

H. : Burkart 5234 (LP); Cabrera 8278, 8399 
(LP); eee et 9i 18064 (LP. 

H. be Angulo & López 1558 (LP); Hutchison, Wright 
& Straw 6267 (US); Macbride & Featherstone 2487 (LP, US); 
Velarde pos 3225 (LP 

ackeri: Lewis 35168 (LPB); Menhofer X-1832 (SI, 
LPB) Smith & a 7212 (US). 
meyeniana: in (LP); Fabris 6514 (LP); 
pees 2029 (P. Mes 21270 (LP); Stuessy et al. 
18062 (LP. 

H. mucida: Ceballos et al. 614 (SI); Menhofer X-2197 ay 
sessiliflora: Cuatrecasas et al. 25599 (LP); Plowm 
1955 (US) Molina & Dau. 185388 (US); Stuessy el al 

18538 nil b 

H. taraxacoides: Meyer et al. 21040 (LP); Stuessy et al. 

18074, “18075 (LP) 


APPENDIX 4. Index to numbered collections cited. Numbers 
in Eo correspond to the species number in the List 


of Spe 


1. Hypochaeris sessiliflora Kunth 
2. Hypochaeris taraxacoides Ball 


Hypochaeris meyeniana (Walp.) Benth. € Hook. f. 
iseb. 


; o 
9. oe acaulis (J. ii Britton 
Adler Aristeguieta 
Asplund 2 (6). Baar 386 e p 6232 (6), B- m (2), 
B-6361 ae B-6711 (2); Tod & Juajibioy 7642 (2), 8472, 
9210 (1); B 3790 (9); Bastián 354 (6); Beck 2861 (6), 
10 (2), 8356 a), 9060 (6). 11914 (5), 17244 (7); Beck & 


Paniagua 27086 (2); Becke: 


Burkart & Troncoso 11822 
6154 (9), 7723, 7826 (6), 8278, 8399 (8), 8982 (6), 8993 (2), 
9051 (8), 12121 abrera et al. 15256 (2), 15393 (7), 
17656 (2), 19002 (7), 21505 (1), 22484, 22531 (2), 27174 (6), 


32555 (8); Canne & Schunke 232 (5); Cardenas 4347 (6); 
Cazal et al. 17 (7); Ceballos et al. 164 (2), 180 (7), 548 (2), 
614 (4); Cerrate 6025 (6); Cerrate e Tovar 1174 (2); Comber 


1257 (9); oe 1484 (1); Cuatrecasas & Romero 
Castaneda 24560 (1 2^ dpa ed A Willard 26299 (1). 
Deginani et al. 806 e la e & Díaz 10905 (6). 


Ellemann 91672 (1); ee 2 8 (4 Eyerdam 24839 (3), 
> 59 (6); Ewan 16324 (1). Fabris 1369 (8), 1520 (2); Fabris 
al. 4035 (2), 6514 (6); Fernández 6); Fernández 
an 8025 (7); Fiebrig 3188 (2); Finot & López 1825 (9); 
Fisel U-123 x. Freueler 4384 ~ Sosa & Múlgura 
157 (8); Grandjot, C. & Grandjot, G. s.n. (9); Guillén p i 
10784 (6). m 5, 142 (2); Haw a et al. 3900 (2); Hens 


6039 (7); VOCE et al. 5892 (2), 
ns (5). Ibish & as 439 (5); Htis € U, 6, 1489 (2), 
1495 (6). Jahn p Q. Kiesling et al. 581 TA p (8), 3801, 
3826 O Killip & Smith 17641, 21680 (1), a G 23369 
(1); King & Bishop 7504 (6), 7530 (7), 7538 (8); King & 
Collins 9064 (2); King & Garvey 6968 (1); ce 7492 (2), 


Annals of the 
Missouri Botanical Garden 


e (7), 8417 (6), 8435, 8672 (2), 9391 (2); Krapovickas 
3155 (2). "a el i 4995 (1); pom 35168 (3), 57146 ©) 
38124 (3); Liberman 2343 (2); López & Sagástegui 2888 
3223, 8245 (2); ie et al. 7421 5r López Veta one ©: 
Loza de la Cruz 72 (2), 253 (7), 254 (2); Luteyn & Dorr 
(6). Macbride & Feathersione 1791 (1), 2167 
Maguire t 7 DIM D er Marticorena & a 
343 (6); M. & Rodrig 0 (6); M. 6 (8); 
Menhofer oe a, 1600 n P3 HU (7), x 1832 (3), i 2197 
4) de 2181 (6); Metcalf 30280 (6); Meyer 21270 (6); 
age et al. 21040 (2), 21305 i 21517 (6); n 149 (6), 
2 09 (5); Mezar 162 (2); Molina & Barkley 18 5.338 (1); 
wd & Heinrichs hank (3); Mie 151 (p. d QA Ls 
6356, 10644 (6); Novara & Ne n 9781 (6). O 
(2); Olea 241 (7); uh ia 132 6). Pom nnell 13362 eN P i a 
et al. 382 (1); Pestalozzi 194 (1), 206 (6), 284 (2), 292 (6), 606 
(5), 939 (6); Pisano * Veniurelli ie (8); oe 1955 (1). 
Ramirez 203 (6); Ricardi et al. 343 (6); weder T-12 (2). 
Sagástegui et al. 9006 (5), 10006 He J; nt. 853, 3915 
(7); Saunders 379, 822 (6), 1106 (2); Schulte 9 (2), 129 (6); 


E 
riae 


— 


— 


Schwabe et al. 641 o 818, 819 T i a 1098 
[scu 2730 (6), 3507 (7); Sleumer st 2589 (6); 
Smith & Buddensie ee (5); Sm. e (uis 7212 (3), 
7328 (1); Smith et al. 5652 (2); Soejarto 494 (1); a 
1254 ax Solomon | 13155 O d (2); RON & N 


6); 


BES 


14233 (2 ; Stafford 865 (5); Stein om 

9696 (6); Ste k et al ee (1); ds dion 1090 

9. ud Qi Edi m 5565, 15571, 15593, 15703, 
7 (9); S al. 18062 (6), a (8), 18068 (6), 


A 18075 a, fais (2), 18505 D 18504 (6), 18515, 
18516, 18525 (7), 1 , 18540, E 9 (1). Tolaba et al. 
1198, 1235 (6), 1539 D. T 52, 16 Do Tovar 222 (3), 390, 
393 (2), 842 (6), 850 (2), 2512 (5); Tutin 1209, 1210 9. 
ue d od 11350 (p.p.) (5), 21784 (1); e. et 
2619, 2887 (1); Ve s Núñez 3225 (5); Vignati 445 a 
Velomam 247 (6), d 2. 2000/61. 5 P $4 
6 (1); Weigend & E 90 (6); Werner 89, 91, 1 

ES (6), 820 FO; West 3663 2 We vi 129 (5); Wurdack i 
1), 1225 (3). Zamalloa Díaz 78 (2); Zamora 156 (6); 
Zuloaga & Landono 4156 (1) 


E 


— 
= 
n 


ANNALS OF THE 
MISSOURI BOTANICAL GARDEN 


VOLUME 96 
2009 


716 


Annals of the 
Missouri Botanical Garden 


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 604 Opus Gloss Recycled No. 2. This is an acid-free paper 
designed to have a shelf-life of over 100 years. 


Photographs used in the ANNALS are reproduced using 300 line screen halftones. The binding used 
in the production of the ANNALS is a proprietary method known as Permanent Binding. 


The ANNALS is printed and distributed by Allen Press, Inc. of Lawrence, Kansas 66044, U.S.A. 
© Missouri Botanical Garden 2009 


ISSN 0026-6493 


Volume 96, Number 1, pp. 1-214 of ANNALS OF THE MISSOURI BOTANICAL GARDEN was published on 23 April 2009. 
Volume 96, Number 2, pp. 215-368 of ANNALS OF THE MISSOURI BOTANICAL GARDEN was published on 7 July 2009. 

Volume 96, Number 3, pp. 369—520 of ANNALS OF THE MISSOURI BOTANICAL GARDEN was published on 28 September 2009. 
Volume 96, Number 4, pp. 521—726 of ANNALS OF THE MISSOURI BOTANICAL GARDEN was published on 30 December 2009. 


www.mbgpress.org 


CONTENTS 


A Revision of the Malagasy Endemic Helmiopsis (Malvaceae s.l.) Wendy L. Applequist 


A Revision of Neotropical Bonyunia (Loganiaceae: Antonieae) — — —— _Jason R. Grant 
Phylogenetic Position and Taxonomic Classification of Aethionema trinervium (Brassi 


A Morphologically Variable Subshrub from Southwestern Asia 
Ahmad Reza Khosravi, Fernand Jacquemoud, Sasan Mohsenzadeh, Marck Menke & 
Klaus Mummenhoff 


A Systematic Revision of Gaertnera (Rubiaceae, Gaertnereae) 
Simon T. Malcomber & Charlotte M. Taylor 
Cryptic Dioecy in Nyssa yunnanensis (Nyssaceae), a Critically Endangered Species from 


Tropical Eastern Asia Bao-Ling Sun, Chang-Qin Zhang, 
Porter P. Lowry I] & Jun Wen 
Systematics of the South American Hypochaeris sessiliflora Complex (Asteraceae, Cichorieae) 
Estrella Urtubey, Tod F. Stuessy & Karin Tremetsberger 


Checklist for Authors 
Author Index 


Subject Index 


521 
541 


y 
N 


Cover illustration. Galianthe longifolia (Standl.) E. L. Cabral, drawn by Laura Simón.