TAXON 61 (6) • December 2012: 1251–1268 Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis Delimitation of the genus Margaritopsis (Rubiaceae) in the Asian, Australasian and Pacific region, based on molecular phylogenetic inference and morphology Laure Barrabé,1,2 Sven Buerki,3 Arnaud Mouly,4 Aaron P. Davis,3 Jérôme Munzinger5 & Laurent Maggia1,2 1 2 3 4 5 CIRAD, UMR AGAP, 98800 Noumea, New Caledonia, France IAC, BP 18239, 98857 Noumea sud, New Caledonia, France Royal Botanic Gardens, Kew, Richmond, Surrey TW9 3AB, U.K. Université de Franche-Comté, UMR CNRS 6249 Chrono-Environnement, 16 route de Gray, 25030 Besançon cedex, France IRD, UMR AMAP, Laboratoire de Botanique et d’Écologie Végétale Appliquées, Herbarium NOU, 98848 Nouméa, NouvelleCalédonie; and IRD, UMR AMAP, 34000 Montpellier, France Authors for correspondence: Laure Barrabé, laure.barrabe@ird.fr; Aaron P. Davis, a.davis@kew.org Abstract In the past, the circumscription of the large genus Psychotria (Rubiaceae) was difficult, until molecular phylogenetic studies revealed its considerable paraphyly, enabling the delimitation of its major lineages and the grouping of related genera, and most notably the separation of Psychotria and its relatives (former Psychotrieae) into two tribes: Psychotrieae and Palicoureeae. The genus Margaritopsis, which is included in Palicoureeae, encompasses 27 Neotropical species, and in previous studies these have been shown to be close relatives of a group of eight Psychotria species that occur over a large region extending from South-East Asia to tropical South Pacific through Malesia (= the AMP region, defined as including South-East Asia, Malesia, tropical Australia, Melanesia, Micronesia, New Caledonia and Polynesia). A molecular phylogenetic study, using one nuclear DNA region (ITS) and four plastid DNA regions (ndhF, rps16, trnH-psbA, trnT-F), is undertaken in order to test the placement of 17 AMP Psychotria species within Palicoureeae. The phylogenetic results show that they form a monophyletic clade (= clade G), which also includes the monotypic Fijian genus Readea and Hodgkinsonia frutescens from Australia. Clade G is embedded in a well-supported grade with five Neotropical representatives of Margaritopsis. A morphological survey based on twenty characters, with the potential to circumscribe generic entities, shows that the monophyly of clade G is supported by a character combination that is similar to species of Neotropical Margaritopsis, confirming their inclusion in this genus. Taxonomic and nomenclatural work on these species is required to formalize nomenclatural implications. Based on the same set of morphological characters, but in absence of molecular data, 28 other Psychotria species from the AMP region were detected as likely candidates for inclusion in Margaritopsis, allowing the estimation of species richness of clade G to be a minimum of 47 species. Within clade G, five well-supported subclades and a Readea lineage are delimited and each of these is generally supported by unique morphological features. Phylogenetic topologies reveal biogeographical patterns, including a main route of dispersal from western to eastern parts of the AMP region, with subsequent dispersals between archipelagos in the region. Keywords biogeography; big genera; Malesia; Margaritopsis; Palicoureeae; phylogeny; Psychotria; South-East Asia; tropical South Pacific Supplementary Material Appendix 2 is available in the Electronic Supplement to the online version of this article (http:// www.ingentaconnect.com/content/iapt/tax). INTRODUCTION The pantropical genus Psychotria L. is by far the largest genus of Rubiaceae (tribe Psychotrieae Cham. & Schltdl.) with ca. 1800 species (Davis & al., 2009). It is also in the top five largest angiosperm genera and is probably the world’s largest predominantely woody genus (Frodin, 2004). Due to its size and apparent lack of defining characters Psychotria is usually perceived as a ‘botanical nightmare’, especially for those requiring species identifications. Its immense diversity, spanning the whole tropics, coupled with the difficulty in ascertaining useful morphological characters (Piesschaert, 2001) has confounded Rubiaceae systematists (e.g., Robbrecht, 1988; Taylor, 1996; Davis & al., 2001). These problems are nearly universal in big genera: their size rendering them difficult, if not impossible, to study in their entirety (Frodin, 2004). Progress in understanding the delimitation of Psychotria has been slow but progressive. The broad Psychotria concept was refuted as early as the 1960s in classical morphotaxonomic studies of the Afro-tropical representatives, including the works of Petit (1964, 1966), Robbrecht (1975) and Verdcourt (1975, 1977). These authors segregated several genera or established new criteria to distinguish related genera, mainly using fruit characters (Robbrecht & Manen, 2006). As soon as molecular analyses were applied to Psychotria and associated genera, it quickly became clear that Psychotria was conspicuously paraphyletic (Andersson & Rova, 1999; Nepokroeff & al., 1999; Bremer & Manen, 2000). Foreseeing Version of Record (identical to print version). 1251 Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis taxonomic and nomenclatural implications, Davis & al. (2001) typified Psychotria with a species from Hong Kong (P. asiatica L.). Andersson (2001) informally recognized two major groups, viz. a ‘Psychotria group’ (including Psychotria s.str.) and a ‘Palicourea group’ (including Psychotria spp. and morphologically related genera). This concept was expanded and elaborated by Anderson (2002a, b), with the addition of further sampling and morphological data, although in this case the two major groups were referred to as the ‘Psychotria complex’ and the ‘Palicourea complex’. Robbrecht & Manen (2006) then formally recognized two tribes, Psychotrieae and Palicoureeae Robbr. & Manen. According to Andersson (2001, 2002a) and Robbrecht & Manen (2006) Palicoureeae is set apart from Psychotrieae by the following morphological characters (characters for Psychotrieae in parentheses): stipules non-caducous (vs. caducous), preformed germination slits in pyrene mostly present (vs. mostly absent; except in Pacific representatives), endosperm non-ruminate (vs. mostly ruminate; except in Pacific representatives), red seed coat pigment [ethanol soluble] absent (vs. mostly present). This classification has been adopted (Razafimandimbison & al., 2008), although some authors prefer adopting a more conservative approach by recognizing a more inclusive tribe Psychotrieae (Bremer, 2009; Bremer & Eriksson, 2009). The aforementioned molecular studies have enabled a more focused approach for investigating Psychotria, and workers have started to provide either morphological synapomorphies or at least a combination of diagnostic characters to support phylogenetic entities for Psychotria and its major lineages, and related genera in Psychotrieae (Davis & Bridson, 2001, 2004; Piesschaert, 2001). Although new insights into the phylogenetic relationships within and outside Psychotria are acknowledged, major problems still persist. Significant amongst these is the considerable problem of identifying Psychotria species in the Old World that are in fact not members of the Psychotrieae lineage at all but instead belong to Palicoureeae. For the New World, progress has been made in resolving generic delimitation within the Palicoureeae clade, and assigning species formerly included in Psychotria into their respective Palicoureeae genera, including Carapichea Aubl. (Taylor & Zappi, 2006), Margaritopsis Sauvalle (Taylor, 2005), Notopleura (Hook. f.) Bremek. (Taylor, 2001), Palicourea Aubl. (Taylor, 1997), and Rudgea Salisb. (Zappi, 2003; Taylor & Zappi, 2006). Largely on the basis of molecular studies (Andersson & Rova, 1999; Nepokroeff & al., 1999; Andersson, 2001, 2002b) ten genera are assigned to Palicoureeae: Carapichea, Chassalia Comm. ex Poir., Chazaliella E.M.A. Petit & Verdc., Geophila D. Don, Hymenocoleus Robbr., Margaritopsis, Notopleura (= Psychotria sect. Notopleura sensu Hooker, 1873), Palicourea s.l. (including Palicourea s.str. and Psychotria subg. Heteropsychotria sensu Steyermark, 1972), Readea Gillespie and Rudgea. Phylogenetic relationships between these genera remain poorly understood, and other questions regarding monophyly, geographical distribution and species richness persist. In addition, important taxonomic treatments have been undertaken for several Old World Palicoureeae genera, or for the Old World representatives, including, e.g., Chassalia (Piesschaert, 2001; 1252 TAXON 61 (6) • December 2012: 1251–1268 Piesschaert & al., 2001), Chazaliella (Verdcourt, 1977), and Hymenocoleus (Robbrecht, 1975). Among the New World Palicoureeae genera, Margaritopsis was initially described to accommodate a species endemic to Cuba, M. nudiflora (Griseb.) K. Schum. (initially M. acuifolia C. Wright), based on the previous but illegitimate genus Margaris Griseb. (Sauvalle, 1869). This genus was later expanded to include 27 species that are widespread throughout the Neotropics (Andersson, 2001; Taylor, 2005). However, Andersson & Rova (1999) found a close relationship between Margaritopsis, the type species of the African genus Chazaliella (C. abrupta (Hiern) E.M.A. Petit & Verdc.), and the monospecific Fijian genus Readea. Andersson (2001) confirmed the former relationship based on a single plastid region (rps16), and enlarged the delimitation of the clade by adding two African Chazaliella, and one Caribbean Margaritopsis, and, for Psychotria, two South American species, and eight species from South-East Asia, Malesia and tropical South Pacific islands. These phylogenetic results are in agreement with the carpological study conducted by Piesschaert (2001) who found a similar and stable morphology of pyrenes for all species tested by Andersson (2001). Consequently, Andersson (2001) enlarged the generic circumscription of Margaritopsis by including the genera Chazaliella, Readea, and certain Psychotria species (see above), and estimated its diversity at 50 species. However, he only published nomenclatural combinations in Margaritopsis for the type species of Chazaliella and Readea, and included the characters of these former genera in an amended generic description of Margaritopsis (Andersson, 2001, 2002c). This morphological circumscription was largely completed by Taylor (2005) to accommodate all other Neotropical species that belong to the genus. Among them Taylor (2005) delineated three informal groups: the ‘Margaritopsis group’, including M. nudiflora, the ‘Chytropsia group’ and the ‘Chazaliella group’. The actions of Taylor (2005) implicitly raise the question of the inclusion of the African genus Chazaliella within Margaritopsis. Even though there are similar morphological features within the Margaritopsis/Chazaliella/Readea group of Andersson (2001), its internal phylogenetic relationships still remain poorly resolved. A study currently in preparation promises new insights into African Chazaliella (O. Lachenaud, pers. comm.). In this study we aim to investigate the Psychotria species that have been suggested as belonging to Palicoureeae (Andersson, 2001) from the AMP region, which we define here as: South-East Asia (= South-East Asian mainland), Malesia (Indonesian archipelago, New Guinea, the Philippines), tropical Australia and tropical South Pacific islands (Melanesia: Bismarck, Fijian, Solomons and Vanuatu archipelagos; Micronesia; New Caledonia; Polynesia: Austral, Gambier, Marquesas, Samoan, Society, Tongan and Tuamotu archipelagos). In the study of Margaritopsis by Andersson (2001) eight species were examined: one species from China (P. montana Blume), one from Indonesia (P. straminea Hutch. in C.S. Sargent), two from New Caledonia (P. collina Labill., P. oleoides (Baill.) Schltr.), one from Vanuatu (P. aneityensis Guillaumin), and three from Fiji (P. amoena A.C. Sm., P. archboldiana Fosberg, P. incompta Version of Record (identical to print version). Australia Fiji 1 2 1 1 Monotypic NA Clade G Clade G Hodgkinsonia F. Muell. Readea Gillespie AMP region (see Fig. 4 for distribution and species richness) Paraphyletic (monophyletic with inclusion 17 of Hodgkinsonia frutescens and Readea) Clade G AMP Psychotria L. 47 Neotropics 27 Africa NA 5 Neotropical Margaritopsis Margaritopsis Sauvalle 1 Paraphyletic (possibly monophyletic) Chazaliella Chazaliella E.M.A. Petit & Verdc. 20 Neotropics Africa 12 12 3 Monophyletic 1 Carapichea Chassalia-Geophila-Hymenocoleus group NA Hymenocoleus Robbr. Carapichea Aubl. Africa, Madagascar, SE Asia, Indonesia Worldwide 27 120 Chassalia-Geophila-Hymenocoleus group Monophyletic Geophila D. Don 2 Chassalia-Geophila-Hymenocoleus group Monophyletic Chassalia Comm. ex Poir. 6 Neotropics Neotropics 100 120 1 Notopleura-Rudgea group Rudgea Salisb. Version of Record (identical to print version). NA NA Notopleura-Rudgea group Notopleura (Hook. f.) Bremek. 1 Neotropics Neotropics 310 2 Monophyletic Species sampled 2 Monophyletic Palicourea s.l. Psychotria L. subg. Heteropsychotria Bremek. Palicourea s.l. Palicourea Aubl. s.str. Taxon sampling. — The sampling comprises 56 species, including sequences of 13 species retrieved from GenBank. Voucher information and GenBank accession numbers are given in Appendix 1. The Palicoureeae sampling was centered on the AMP region, covering the species richness in each biogeographical area as defined by Mueller-Dombois & Fosberg (1998) and Keppel & al. (2009): South-East Asia (Vietnam), Malesia (New Guinea), tropical Australia and tropical South Pacific islands (Fiji, New Caledonia, Samoa, Society archipelago and Vanuatu). In addition to the 17 AMP Psychotria sampled, at least one species of each Palicoureeae genus (as enumerated in the introduction and summarized in Table 1) was included in the sampling. Hodgkinsonia frutescens C.T. White, which belongs to the bitypic genus Hodgkinsonia F. Muell., endemic to tropical Australia, was also included because its morphology is similar to some Palicoureeae. The samples Psychotria sp. ‘NC’ and P. sp. ‘V’ belong to new undescribed taxa from New Caledonia and Vanuatu, respectively. Psychotria sp. ‘FIJ’, P. sp. 1 [VIET] and P. sp. 2 [VIET] were identified as species of uncertain affinity but constitute distinct species. The identification of P. sp. [AUST] was not possible because no herbarium specimen was provided (cultivated sample, Appendix 1). Extraction of DNA from herbarium specimens from Indonesia, the Philippines and the Micronesian archipelagos was carried out but failed to provide useful amounts of DNA, with the exception of one species of Chassalia from Sumatra. To further confirm the relationships within Palicoureeae, several species of Psychotrieae from the Pacific clade and the Psychotria s.str. clade as defined by Andersson (2002a) were sampled. The type species of Psychotria (Davis & al., 2001; Sohmer & Davis, 2007), P. asiatica, was also included in this study. Craterispermum Benth. and Prismatomeris Thw. were chosen as outgroup taxa following Razafimandimbison & al. (2008) and Rydin & al. (2009). Phylogenetic status in this study MATERIALS AND METHODS Phylogenetic group A.C. Sm.). However, this sampling does not satisfactorily represent the distribution and species richness of the group in the AMP region, as there is no sample from Australia, large parts of South-East Asia, Malesia and other tropical South Pacific islands. This is not an issue for Hawaii, as all species occurring in the archipelago belong to a single clade in Psychotrieae (Nepokroeff & al., 1999, 2003). These considerations provide clear objectives for the present investigation, which are to: (1) construct a phylogenetic framework that will allow us to elucidate those Psychotria taxa from the AMP region (= AMP Psychotria) that are closely related to Neotropical Margaritopsis, based on plastid and nuclear DNA regions; (2) assess the species richness and geographical range of the AMP Psychotria group (i.e., belonging to Palicoureeae) based on a morphological survey of species occurring in this area; and (3) find morphological synapomorphies that support both the monophyly of clades within the AMP Psychotria group, and the inclusion of these taxa in Margaritopsis. Finally, we discuss biogeographical patterns that may have been involved in shaping current species diversity. Species estimated Distribution Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis Table 1. Genera of Palicoureeae and their phylogenetic status based on the data presented in this study. The distribution and number of species per genus are taken from Govaerts & al. (2011) and this study. TAXON 61 (6) • December 2012: 1251–1268 1253 Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis DNA regions. — Five plastid and nuclear DNA regions, which have been widely applied to retrieve phylogenetic relationships within Rubiaceae, were sequenced. Four plastid markers were used: the rps16 intron (Oxelman & al., 1997), the trnT-F region (including the trnT-trnL and trnL-trnF spacers and the trnL intron; Taberlet & al., 1991), the ndhF gene (Kim & Jansen, 1995), and the trnH-psbA spacer (Aldrich & al., 1988). A single nuclear marker was used: the ITS region, including ITS1, 5.8S and ITS2 (Baldwin & al., 1995). The primers used in polymerase chains reactions were the following: A1/IR and C/F for trnT-F (Razafimandimbison & Bremer, 2002; Taberlet & al., 1991); trnHGUG/psbA for trnH-psbA (Sang & al., 1997; Tate & Simpson, 2003); F/2R for rps16 (Oxelman & al., 1997); 2F/1000R, 720F/1700R and 1320F/2280R for ndhF (Rydin & al., 2008) and P17/26S-82R for ITS (Bolmgren & Oxelman in Popp & Oxelman, 2001). PCR protocols and DNA sequencing. — All DNA samples were extracted with the DNAeasy Plant Mini Kit (Qiagen, Crawley, U.K.) and DNA regions were amplified with the GoTaq Flexi DNA Polymerase (Promega, Madison, Wisconsin, U.S.A.). For the ndhF, rps16, trnH-psbA and trnT-F regions, the mix template included 5 µL of Buffer (5×), 1.5 µL of MgCl2 (25 mM), 0.25 µL of dNTP mix (10 mM of each dNTP), 0.75 µL of primers (10 µM, forward and reverse), 0.25 µL of polymerase (5 U/µL), 2 µL of DNA and the volume was adjusted to 25 µL with water. For the ITS region, the mix template was the same as for the plastid regions, but with the addition of 0.625 µL of BSA (10%) and 0.5 µL of DMSO (10%). The following PCR programs were used for the plastid markers: 3 min at 94°C, 35 cycles of 1 min at 94°C, 1 min at 50°C, 2 min at 72°C, followed by a final extension of 7 min at 72°C, and for the nuclear marker: 3 min at 94°C, 35 cycles of 30 s at 94°C, 1 min at 50°C, 1 min 30 s at 72°C, followed by a final extension of 7 min at 72°C. All PCR products were purified using DNA purification columns according to the manufacturer’s protocol (QIAquick PCR Purification Kit; Qiagen) or using a Sephadex column purification (Sephadex G-50 Superfine resin, Sigma-Aldrich, St. Louis, Missouri, U.S.A.). Dideoxy cycle sequencing was then performed using the chain termination method and ABI Prism Big Dye version 3.1 reaction kit, following the manufacturer’s protocol, but using 0.5 µL of reaction mix (Applied Biosystems, Warrington, U.K). The products were prepared for sequencing using a Sephadex column purification (Sephadex G-50 Superfine resin, Sigma-Aldrich) and visualised on an ABI 3730 Genetic Analyser, also according to the manufacturer’s protocol. Phylogenetic analyses. — The program Sequencher v.4.10.1 (Gene Codes, Ann Arbor, Michigan, U.S.A.) was used to assemble complementary strands and verify software basecalling. The regions were initially aligned individually with the online version of MUSCLE v.3.8.31 (www.ebi.ac.uk/tools/msa/ muscle), and subsequently manually adjusted with MEGA5 (Tamura & al., 2011). Sequences of Palicourea nitidella (Müll. Arg.) Standl. and Psychotria stachyoides Benth. were sometimes difficult to align for a large part of the trnH-psbA and the trnT-F regions. As both species are quite divergent from ingroup species, these parts were replaced with missing data 1254 TAXON 61 (6) • December 2012: 1251–1268 (trnH-psbA: positions 112–538; trnT-F: 94–429 for Palicourea nitidella and 94–407 for P. stachyoides). An inversion was found in some of the trnH-psbA sequences (88–93) and this portion was therefore not included in the phylogenetic analyses. The introduction of many indels in the rps16, trnH-psbA and trnT-F matrices and the occurrence of a large poly A/T region in the rps16 sequences (positions 399–566) induced problematic homology hypotheses (Shaw & al., 2005). The nrDNA ITS region has proven useful for reconstructing relationships of closely related species, due to high levels of polymorphism (Zarrei & al., 2012). However, the occurrence of multiple substitutions, and/or an incomplete process of concerted evolution (that normally allowed elimination of pseudogenes), can corrupt the phylogenetic signal and establish inaccurate relationships (Razafimandimbison & al., 2004; Sanderson & Doyle, 1992). To improve the quality of analyses, ambiguous regions were removed from all DNA alignments by using the BMGE software v.1.1 (Criscuolo & Gribaldo, 2010), which estimates entropy for each position and eliminates those for which the entropic values are too high. The BMGE analyses were set as follows: PAM250 similarity matrix; sliding window set to 1 (with the exception of the coding ndhF region that was set with a sliding window of 3); gap cut-off set to 0.75 for most of the plastid regions and to 0.2 for ndhF and the ITS region (Table 2). Single-gene and combined phylogenetic inferences were carried out employing both maximum parsimony (MP) and Bayesian Markov chain Monte Carlo (MCMC) analyses. MP analyses were performed using PAUP* (Swofford, 2002) with the following settings: random addition sequence (nreps = 10,000 and 20 trees saved per replicate), tree-bisection-reconnection branch swapping, STEEPEST and MULTREES options in effect, and an unlimited value for MAXTREES. A majorityrule consensus tree was constructed based on the most parsimonious trees. To estimate homoplasy, the consistency index (CI) and the retention index (RI) were calculated. To evaluate node support a bootstrap analysis (BS; Felsenstein, 1985) was performed using PAUP* (Swofford, 2002) with 500 replicates by using the same parameters as above. Best-fit models for each DNA region were identified using jModelTest v.0.1.1 (Posada, 2008) based on the Akaike criterion (see Table 2). In the combined Bayesian MCMC analysis, the dataset was divided into five partitions and each locus was allowed to have partition-specific model parameters (Ronquist & Huelsenbeck, 2003; Nylander & al., 2004). The single-gene and combined analyses were set as follows: four Metropolis-coupled Markov chains with an incremental heating temperature of 0.2 were run for 25 million generations and a tree was sampled every 1000th generation. The analysis was repeated three times, starting with random trees, and was run in MrBayes v.3.1.2 (Ronquist & Huelsenbeck, 2003). Bayesian MCMC analyses were carried out using the facility Bioportal provided by the University of Oslo (www.bioportal.uio .no, Norway). The MCMC sampling was considered sufficient when the effective sampling size (ESS) was higher than 200, as verified in Tracer v.1.5 (Rambaut & Drummond, 2007). After a burn-in period of 1 × 106 generations per run, the remaining trees were used to construct a half-compatible consensus tree Version of Record (identical to print version). Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis TAXON 61 (6) • December 2012: 1251–1268 (i.e., majority-rule consensus tree from MrBayes) and its associated Bayesian posterior probabilities (PP). A clade with a PP or a BS value higher than 0.95 and 95%, respectively, was considered as well supported. Hypothesis testing and ambiguous phylogenetic placement of Margaritopsis nudiflora. — Preliminary results suggested ambiguous placements for the type species of Margaritopsis (see the Results section below for more details). We failed to sequence M. nudiflora for the ndhF region. Therefore, to estimate the influence of missing data on the phylogenetic position of this taxon, we removed this DNA region from the dataset and conducted subsequent combined Bayesian MCMC analysis with the pruned matrix (= dataset without ndhF). These analyses were run with the same parameters as those for the whole dataset (= dataset with ndhF). Furthermore, because we are aiming at unravelling phylogenetic relationships between taxa of Margaritopsis and the AMP Psychotria species, two additional combined constrained Bayesian MCMC analyses for each dataset (with ndhF and without) were computed following the approach described by Bone & al. (2012). In the first constrained Bayesian MCMC analyses, the five Neotropical Margaritopsis (including M. nudiflora) were constrained as monophyletic. In the second constrained Bayesian MCMC analyses, four Neotropical Margaritopsis (M. astrellantha (Wernham) L. Andersson, M. boliviana (Standl.) C.M. Taylor, M. guianensis (Bremek.) C.M. Taylor, M. kappleri (Miq.) C.M. Taylor) and all AMP Psychotria species plus their relatives were constrained as constituting a monophyletic group. These constrained analyses were run with the same parameters as those in the unconstrained combined Bayesian MCMC analyses (see above). The Shimodaira-Hasegawa test (SH test, Shimodaira & Hasegawa, 1999) was employed to determine whether the half-compatible consensus topologies resulting from each constrained Bayesian MCMC analysis were statistically worse than unconstrained topologies. The SH test was carried out in PAUP* (Swofford, 2002) using the RELL method with 10,000 bootstrap replicates. Because only one set of model parameters can be implemented in PAUP*, each combined dataset was then considered as a single DNA partition. The best-fit model for each combined dataset was estimated to be GTR + G + I with jModelTest (Posada, 2008), and these estimated model parameters were used to set the SH tests. Examination of herbarium specimens. — To identify morphological synapomorphies supporting monophyletic lineages, a representative selection of herbarium specimens of the ingroup taxa (deposited at K, NOU and P) was examined and combined with field observations by the authors. Type specimens were systematically consulted where available; Appendix 2 (Electronic Suppl.) provides details on the herbarium specimens examined. In addition, local taxonomic treatments of Psychotria and associated genera occurring in the AMP region (Samoa: Whistler, 1986; Fiji: Smith & Darwin, 1988; New Guinea and Bismarck Archipelago: Sohmer, 1988; the Philippines: Sohmer & Davis, 2007) were used. The revisions of Fosberg & al. (1993) and Chen & Taylor (2011), respectively, for Micronesian and Chinese Psychotria taxa did not provide enough diagnostic features and were therefore not incorporated in this study. There is a lack of taxonomic treatments for South-East Asian and Indonesian taxa, which resulted in difficulties in assessing species richness and morphological diversity in these areas. Morphological characters of outgroup taxa were compiled based on available taxonomic revisions and carpological data (Petit, 1964; Robbrecht, 1975; Verdcourt, 1975, 1977; Taylor, 1997, 2001, 2005, 2010; Nepokroeff & al., 1999; Piesschaert & al., 1999, 2001; Piesschaert, 2001; Andersson, 2001, 2002b; Zappi, 2003; Taylor & Zappi, 2006). These characters must be considered cautiously because they reflect global tendencies within Palicoureeae genera, although they allow morphological comparison with Table 2. Characteristics of the DNA partitions and BMGE software settings used in the phylogenetic analyses of tribes Palicoureeae and Psychotrieae. trnT-F trnH-psbA rps16 ndhF ITS Total Cellular compartment Plastid Plastid Plastid Plastid Nuclear Number of taxa 52 39 53 53 58 58 7089 Number of aligned characters 2209 845 1224 2061 750 Similarity matrix PAM250 PAM250 PAM250 PAM250 PAM250 Sliding windows size 1 1 1 3 1 GAP rate cut-off 0.75 0.75 0.75 0.2 0.2 Number of included characters (A) 1770 324 863 2040 702 5699 Number of variable uninformative characters 173 38 105 166 72 554 Number of parsimony-informative characters (PICs) (B) 231 83 149 245 245 953 Number of constant characters 203 609 1629 385 4192 17 1366 % PICs within separate DNA region (B / A × 100) 13 26 17 12 35 Evolution model GTR + G GTR + I + G SYM + I + G GTR + I + G SYM + G Retention index (RI) 0.906883 0.877301 0.900510 0.922351 0.723106 0.834950 Consistency index (CI) 0.798246 0.782609 0.790323 0.811287 0.464253 0.658811 Length of most parsimonious trees 570 184 372 567 1105 2843 Version of Record (identical to print version). 1255 Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis Psychotria taxa from the AMP region. Two taxonomic levels were investigated in the morphological section. Firstly, at the species level, features that allowed morphological comparisons between the AMP Psychotria species include: stipule shape, inflorescence type, corolla aestivation, calyx and corolla shape, size and texture, fruit shape, and pyrene details. Secondly, at the generic level, the characters that allowed comparison between Palicoureeae genera include: general habit, vegetative features of the stem, leaves and stipules, general features of corolla (symmetry and colour), fruit colour, and features of pyrenes. In addition, herbarium specimens of other AMP Palicoureeae species (not previously included in the ingroup sampling of the phylogenetic study) were also examined to determine if they share morphological affinities with the ingroup (Appendix 2 in the Electronic Suppl.). This work allowed us to estimate the TAXON 61 (6) • December 2012: 1251–1268 putative species richness and endemism of AMP Psychotria taxa for each biogeographical area of the AMP region. The species names used in this study correspond to those of the World Checklist of Rubiaceae (Govaerts & al., 2011). RESULTS Sequences characteristic. — The trnT-F region and ndhF gene provide the longest DNA alignment lengths (with 2209 and 2061 base pairs, and 1770 and 2040 included sites, respectively, after the exclusion of ambiguous sites using BMGE analyses; Table 2). The combined DNA dataset is composed of 7089 base pairs and 5699 sites, retained after trimming the single gene matrices in BMGE (see Table 2 for further statistics). The ITS 1 / 100 1 / 100 1 / 93 1 / 80 1 / 100 1 / 96 0.63 / - 1 / 100 1 / 100 1 / 100 1 / 100 1 / 100 0.52 / - G 1 / 100 1 / 100 1 / 97 1 / 97 0.61 / 54 0.97 / 54 1 / 99 0.99 / 53 1 / 100 F 1 / 100 1 / 100 1 / 99 E 0.67 / - 1 / 100 D 1 / 100 1 / 100 1 / 99 1 / 56 1 / 98 0.74 / - 1 / 100 C 0.83 / - 0.73 / 56 1 / 90 1 / 92 PALICOUREEAE B 1 / 60 1 / 97 A 1 / 100 1 / 100 0.95 / - 1 / 100 Pacific Clade PSYCHOTRIEAE 1 / 100 1 / 100 0.93 / 79 1 / 100 0.94 / 66 Psychotria s.str. sensu Andersson 1 / 100 1 / 68 1 / 80 1 / 69 Craterispermum sp2 Craterispermum sp3 Prismatomeris albidiflora Prismatomeris beccariana Psychotria lepiniana Psychotria spFIJ Psychotria carnea 1 Psychotria archboldiana Psychotria carnea 2 Readea membranacea Psychotria forsteriana Psychotria samoana Psychotria cf trichostoma Hodgkinsonia frutescens Psychotria spAUST Psychotria leptothyrsa Psychotria aneityensis Psychotria lyciiflora Psychotria spNC Psychotria spV Psychotria oleoides Psychotria collina Psychotria sp1VIET Psychotria sp2VIET Margaritopsis nudiflora Margaritopsis kappleri Margaritopsis astrellantha Margaritopsis boliviana Margaritopsis guianensis Chazaliella abrupta 1 Chazaliella abrupta 2 Carapichea affinis Carapichea ligularis Carapichea ipecacuanha Geophila repens Geophila obvallata Chassalia sp1VIET Chassalia sp2VIET Chassalia spSUMA Chassalia catatii Chassalia sp2MADAG Chassalia sp1MADAG Hymenocoleus hirsutus Notopleura tapajozensis Rudgea stipulacea Palicourea nitidella Palicourea crocea Psychotria stachyoides Psychotria poeppigiana Amaracarpus nematopodus 1 Amaracarpus nematopodus 2 Hydnophytum cf longistylum Psychotria cf impercepta Psychotria schlechteriana Psychotria grandis Psychotria asiatica Psychotria camptopus Psychotria parkeri * clade G * * Neotropical Margaritopsis Chazaliella Carapichea * Chassalia / Geophila / Hymenocoleus * Notopleura / Rudgea Palicourea s.l. * Fig. 1. Bayesian half-compatible consensus tree of tribes Palicoureeae and Psychotrieae. Asterisks indicate generic type species. Taxa from the AMP region, Neotropics and Africa are shaded in dark, middle and light grey, respectively. Bayesian posterior probability (PP) and bootstrap (BS) support values for major clades are indicated and separated by a slash. 1256 Version of Record (identical to print version). Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis TAXON 61 (6) • December 2012: 1251–1268 and trnH-psbA regions provide most polymorphic characters, respectively, with 35% and 26% Parsimony Informative Characters (PICs), whereas ndhF and trnT-F regions yield 12% and 13% of PICs, respectively (Table 2). Phylogenetic analyses. — The MP and Bayesian MCMC single-gene phylogenetic topologies yield no incongruence (with a PP > 0.95 or BS > 95%) and are less resolved than the combined phylogenetic analyses. Both MP and Bayesian MCMC combined analyses provide no incongruence between topologies (data not shown). Since the Bayesian MCMC analysis allows us to take into account different models of evolution across the plastid and nuclear regions, and provides branch lengths, only this analysis method is discussed hereafter (Fig. 1). The analyses strongly support the monophyly of a group composed of all AMP Psychotria and Margaritopsis species (PP = 1; Figs. 1 & 2). Within Palicoureeae seven clades are highly supported (clades A to G), all with PP = 1 (Fig. 1). Both species of Palicourea are embedded in a well-supported Neotropical clade with Psychotria stachyoides and P. poeppigiana Müll. Arg. (clade A). This clade is placed as sister to all other Palicoureeae with high support (PP = 1). Clade B is composed of single species representatives of the Neotropical genera Rudgea and Notopleura. Clade C includes species samples of Chassalia, Geophila, and Hymenocoleus from Africa and the AMP region (the first two genera are retrieved as monophyletic, PP = 1). Clade D includes Neotropical taxa of Carapichea; clade E is composed of two African Chazaliella accessions. Four of the five sampled Neotropical Margaritopsis species form clade F that is sister to the type species of Margaritopsis (M. nudiflora) and clade G. Finally, clade G comprises all AMP Psychotria taxa together with Readea membranacea and Hodgkinsonia carnea subclade samoana subclade leptothyrsa subclade G collina subclade Vietnamese subclade frutescens. The placement of M. nudiflora as sister to clade G still remains uncertain (PP = 0.61), but the clade which includes this species and clades F and G is well supported (PP = 1). Further details on the phylogenetic placement of M. nudiflora are given in the following section. The subdivision of clade G into five well-supported monophyletic subclades is well supported (PP ≥ 0.95; Fig. 2). The ‘Vietnamese’ subclade is composed of Vietnamese Psychotria species (P. sp. 1 [VIET] and P. sp. 2 [VIET], PP = 0.97). The ‘collina’ subclade (PP = 1) includes all species from New Caledonia (P. collina, P. lyciiflora (Baill.) Schltr., P. oleoides, P. sp. ‘NC’) and one taxon from Vanuatu (P. sp. ‘V’, PP = 1). The ‘leptothyrsa’ subclade (PP = 1) comprises both species from tropical Australia (Hodgkinsonia frutescens is placed as sister to P. sp. [AUST]; PP = 1), one species from Vanuatu (P. aneityensis) and the widespread P. leptothyrsa Miq. (sampled from New Guinea). The ‘Samoana’ subclade (PP = 1) is composed of one Samoan species (P. samoana K. Schum.), one species occurring in Fiji, Samoa and Tonga (P. forsteriana A. Gray), and one species from Vanuatu (P. cf. trichostoma Merr. & L.M. Perry). Readea membranacea Gillespie endemic to Fiji is placed as sister to the ‘carnea’ subclade (PP = 1). This latter subclade (PP = 1) includes species from Fiji and Tonga (P. archboldiana, P. carnea (G. Forst.) A.C. Sm., P. sp. ‘FIJ’), and the single species from the Society Archipelago, P. lepiniana (Baill. ex Drake) Drake (PP = 1). The relationships between these five subclades are well supported (PP ≥ 0.95) with the exception of the relationship between a clade combining R. membranacea and the Samoana and carnea and leptothyrsa subclades (Fig. 2). Psychotria carnea is retrieved as polyphyletic (PP = 1), even if its two accessions are found in the same clade (carnea subclade). Psychotria lepiniana Psychotria spFIJ Psychotria carnea 1 Psychotria archboldiana Psychotria carnea 2 Readea membranacea Psychotria forsteriana Psychotria samoana Psychotria cf trichostoma Hodgkinsonia frutescens Psychotria spAUST Psychotria leptothyrsa Psychotria aneityensis Psychotria lyciiflora Psychotria spNC Psychotria spV Psychotria oleoides Psychotria collina Psychotria sp1VIET Psychotria sp2VIET Margaritopsis nudiflora SOCI FIJI FIJI / TONG FIJI FIJI / TONG FIJI FIJI / SAMO / TONG SAMO VANU AUST AUST BISM / INDO / MICR / PAPU / PHIL / SOLO / WAL VANU NCAL NCAL VANU NCAL NCAL VIET VIET Fig. 2. Bayesian half-compatible consensus tree showing relationships within clade G (Fig. 1). Abbreviations used for geographical distribution: AUST: tropical Australia, BISM: Bismarck Archipelago, FIJI: Fiji, INDO: Indonesia, MICR: Micronesia, NCAL: New Caledonia, PAPU: New Guinea, PHIL: the Philippines, SAMO: Samoa, SOCI, Society archipelago, SOLO: Solomons, TONG: Tonga, VANU: Vanuatu, VIET: Vietnam, WAL: Wallis. Areas where species were sampled for the molecular study are indicated in bold. Bayesian posterior probability (PP) support values for major clades are indicated for each node. Version of Record (identical to print version). 1257 Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis Hypothesis testing and ambiguous phylogenetic placement of Margaritopsis nudiflora. — The placement of Margaritopsis nudiflora in all Bayesian MCMC single-gene analyses is unresolved or unsupported (PP < 0.95, data not shown). Relationships between M. nudiflora, clade G and the four other Neotropical Margaritopsis remain uncertain based on the analysis of each trnH-psbA and rps16 regions. Data provided by the trnT-F region place M. nudiflora sister to clade G but with low support (PP = 0.63). The monophyly of the five Neotropical Margaritopsis species is weakly supported (PP = 0.54) by the ITS region analysis. Phylogenetic topologies resulting from the unconstrained combined Bayesian MCMC analysis of the dataset without ndhF yield exactly the same topology as that with ndhF (data not shown). The phylogenetic placement of M. nudiflora is the same, as sister to clade G, with weak support (PP = 0.70). For both datasets (with ndhF or without) the SH tests, which compared the topology resulting from the unconstrained analysis with one that enforced the monophyly of the five Margaritopsis species, are not significant (P > 0.1, see Table 3). The SH tests that compared the unconstrained topology with one that constrained the monophyly of the group which included AMP Psychotria plus their relatives plus four Margaritopsis species, are also not significant (P > 0.1, see Table 3). Morphological studies. — The results of the morphological investigation based on herbarium specimens and taxonomic treatments are found in Tables 4 and 5. Significant morphological traits for species within clade G (i.e., 17 AMP Psychotria taxa plus Hodgkinsonia frutescens and Readea membranacea) were selected at the species level after examination of herbarium specimens (Table 4). Morphological characters generally provide support for the recognition of each of the five subclades of clade G as well as the Readea lineage (characters in bold in Table 4). Morphological comparison at the generic level allows recognition of unique characters (characters in bold in Table 5), or at least a combination of characters, enabling delimitation of Palicoureeae genera. All of the 19 species belonging to clade G share the following combination of characters (Table 5): (1) shrubs or small trees; (2) vegetative parts TAXON 61 (6) • December 2012: 1251–1268 pale yellowish green when dried; (3) branchlets smooth (noncorky), young shoots and young internodes often flattened; (4) stipules semi-deciduous, entire, usually united at the base or calyptrate for a large part of their length, becoming indurated (sometimes inconspicuously for calyptrate species) and their upper parts falling off through fragmentation (Fig. 3D & G); (5) inflorescence axes always green (Fig. 3A, B, E, F, H, I); (6) calyx and corolla actinomorphic, corolla with a limited colour range (white, creamy to yellow-green) and corolla tube straight at base (i.e., never swollen; Fig. 3A, B, E, F, H); (7) fruits orange or red (Fig. 3J–L); (8) pyrenes hemispherical in cross-section, without a ventral depression or intrusion of the seed-coat; (9) pyrenes with two basal ventral marginal preformed germination slits (PGSs); (10) endosperm non-ruminate, often with a small inner central ventral invagination; and (11) seed-coat lacking a red ethanol-soluble pigment. In addition to the 19 species used in the phylogenetic sampling, 28 other Psychotria species from the AMP region were detected by the study of herbarium specimens and on the basis of taxonomic treatments (Table 6; Appendix 2 in the Electronic Suppl.). They are assigned to clade G based on the same set of morphological characters. Indeed, the diagnostic features of the 19 species of clade G (see above) are shared by these 28 species. This part of the study allows us to estimate the species richness of clade G to a minimum of 47 species (Table 6; Appendix 2), with the inclusion of the 28 species examined using morphology alone. We are able to show that each archipelago and landmass of South-East Asia, Malesia, tropical Australia and tropical South Pacific possesses at least one species, except Hawaii (Table 6; Fig. 4). According to our findings archipelagos and landmasses would at least possess: eight species each for Fiji and Samoa, seven species for the Bismarck Archipelago, five species for New Guinea, four species each for New Caledonia and Solomons, three species for Vanuatu, two species each for the Philippines, Tonga and tropical Australia, and one species for eastern Polynesia. Probably more than eight species occur in South-East Asia and Indonesia, and more than one species in Micronesia (Table 6; Fig. 4), although these estimates need to be reassessed pending future taxonomic treatments. Table 3. Results of the Shimodaira-Hasegawa tests for the half-compatible Bayesian consensus trees resulting from analyses of the dataset with or without ndhF. The best hypothesis is given in bold. ‘5 MAR’ refers to the constraint placed to enforce the monophyly of the five Neotropical species of Margaritopsis (including M. nudiflora (Griseb.) K. Schum.). ‘4 MAR + clade G’ refers to the constraint that enforced the monophyly of a group which includes AMP Psychotria species plus their relatives (= clade G) and four Neotropical species of Margaritopsis (M. astrellantha (Wernham) L. Andersson, M. boliviana (Standl.) C.M. Taylor, M. guianensis (Bremek.) C.M. Taylor and M. kappleri (Miq.) C.M. Taylor). Hypothesis 1 (H1) Hypothesis 2 (H2) −ln L H1 −ln L H2 Diff. −ln L P value Unconstrained Constrained 5 MAR 25253.75206 25254.12317 0.37111 0.3265 Unconstrained Constrained 4 MAR + clade G 25253.75206 25254.13672 0.38466 0.3176 Constrained 5 MAR Constrained 4 MAR + clade G 25254.12317 25254.13672 0.01355 0.4181 Unconstrained Constrained 5 MAR 18308.35663 18308.74398 0.38735 0.3216 Unconstrained Constrained 4 MAR + clade G 18308.35663 18308.57976 0.22313 0.4912 Constrained 5 MAR Constrained 4 MAR + clade G 18308.74398 18308.57976 0.16423 0.4804 With ndhF Without ndhF 1258 Version of Record (identical to print version). Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis TAXON 61 (6) • December 2012: 1251–1268 Table 4. Morphological characters of the five subclades of clade G and the Readea lineage that are common (roman letters) and unique (bold). Subclade / lineage Vietnamese collina leptothyrsa samoana Readea carnea Stipules Truncate / triangular Truncate / triangular Truncate / triangular Papyraceous and calyptrate Truncate / triangular Calyptrate / lanceolate Calyx type Reduced, with or without small teeth Reduced, with or without small teeth Reduced, with or without small teeth Reduced, with or without small teeth Funnel-shaped, with long reflexed lobes Inflated, enlarged, with erect lobes Calyx tube Not enclosing the corolla tube Not enclosing the corolla tube Not enclosing the corolla tube Not enclosing the corolla tube Enclosing the corolla tube Not enclosing the corolla tube Calyx texture Chartaceous Chartaceous Chartaceous Chartaceous Fleshy Chartaceous Corolla aestivation Rounded Rounded Rounded Rounded Rounded Reduplicate (winged) Corolla type Reduced, with short lobes Reduced, with short lobes Reduced to urseolate (bell-like), with short lobes Reduced, with short lobes Funnel-shaped, with long reflexed lobes Funnel-shaped, with long lobes Corolla texture Chartaceous Chartaceous Chartaceous Chartaceous Fleshy Chartaceous Inflorescence Compound cyme Compound cyme Cyme, verticillate at each node Panicule / thyrse Compound cyme Compound cyme / fascicule Fruit shape Rounded to ellipsoid Rounded to ellipsoid Rounded to ellipsoid Rounded to ellipsoid Oblongoid Ellipsoid Pyrenes ventral face Concave Flat Flat Flat Flat Flat DISCUSSION The results presented here corroborate previous conclusions by Nepokroeff & al. (1999), Andersson & Rova (1999) and Andersson (2001, 2002b) and support the monophyly of a group that includes the AMP Psychotria species plus Hodgkinsonia frutescens and Readea membranacea (i.e., clade G), and that falls in a clade together with Neotropical Margaritopsis (Fig. 1). The results support the same seven major monophyletic lineages of Palicoureeae as retrieved by Andersson (2001), and in addition increases our understanding of the phylogenetic relationships between and within these lineages on the basis of additional DNA regions and a broader sampling from the AMP region (Table 1; Fig. 1). More significantly, the close relationship of species of clade G with Neotropical Margaritopsis necessitates a re-investigation of the circumscription of Margaritopsis, based on morphological evidence. However, the unsupported position of the type species of Margaritopsis (M. nudiflora) does not allow us to draw any systematic conclusions to be made on the phylogenetic status of Neotropical species of Margaritopsis. Single-gene Bayesian MCMC analyses and combined analyses, with or without ndhF, did not resolve the placement of M. nudiflora. Results of the SH tests (Table 3) did not allow us to either exclude the monophyly of Neotropical Margaritopsis species or support their paraphyly (M. nudiflora placed as sister to species of clade G). Diversity and geographic range of clade G. — The species richness of clade G (19 species) as ascertained by the molecular investigation reported here is higher than the nine species previously suggested by Andersson (2001). Furthermore, our investigation of herbarium specimens allows us to increase the potential species richness of the lineage, here represented as clade G, to 47 species. These species are widely distributed throughout the AMP region, from South-East Asia to eastern Polynesia (Fig. 4), as observed in other Rubiaceae genera (e.g., Ixora L., Mouly & al., 2009; Gynochthodes Blume, Razafimandimbison & Bremer, 2011). This lineage has colonized all archipelagos and landmasses of the AMP region Version of Record (identical to print version). 1259 Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis without exception (Fig. 4), and each of them possesses at least one of the species we attribute to the lineage. Clade G does not seem to have radiated significantly, because species richness does not exceed eight species in each archipelago and tropical landmass of the AMP region (up to seven species for the Bismarck Archipelago, five for New Guinea, and eight each for Fiji and for Samoa; Fig. 4; Whistler, 1986; Smith & Darwin, 1988; Sohmer, 1988). Most of the species are endemic to a specific archipelago (Fig. 4), but some are widely distributed throughout South-East Asia and Malesia (e.g., P. leptothyrsa, P. membranifolia Bartl. ex DC., P. montana; Sohmer, 1988; Sohmer & Davis, 2007). Interestingly, species of clade G are able to occupy a large range of ecological niches (in tropical to subtropical ecosystems). They grow on a wide range of geological substrates (limestone, clay, metamorphic rocks, ultramafic rocks, etc.), in most natural vegetation types and in degraded and secondary vegetation, and from sea level to high elevations (pers. obs. of the first author; Whistler, 1986; Sohmer, 1988; TAXON 61 (6) • December 2012: 1251–1268 Smith & Darwin, 1988; Sohmer & Davis, 2007). The estimation of species richness provided in this study requires confirmation, as some species remain poorly known (e.g., in South-East Asia and Indonesia) and cases of morphological convergence have been widely recorded in Rubiaceae (e.g., in Ixora, Mouly & al., 2009; and Coffea, Davis & al., 2011). Indeed, a taxonomic revision for the species of clade G, throughout the whole AMP region, is urgently required. Diagnostic features of the internal lineages of clade G. — Our morphological study at the species level reveals that each of the five subclades of clade G and the Readea lineage can be characterized by a diagnostic combination of morphological features, and generally possesses unique characters that further support their monophyly (characters in bold in Table 4; see below). The exception is the collina subclade which does not show any distinctive features (Fig. 3F). In contrast, the Vietnamese subclade is the only group where the ventral face of the pyrenes (Table 4) is conspicuously concave. The inflorescences of the Table 5. Morphological characters of the genera of Palicoureeae. The characters that are similar among all genera are shaded in grey, commonly … Generic group Palicourea s.l. Notopleura Habit Shrubs Shrubs / epiphytes / Shrubs succulents/ climbers Shrubs Vegetative colour when dried Greyish brown Greyish silver Greenish Greyish brown Stem indurate, corky No No No No Stem flattened No No No No Stipule type Entire or bilobed, truncate, sheathing Entire, bilobed or fimbriate, sheathing, sometimes with an interpetiolar outer appendage Entire or fimbriate, Entire or fimbriate, generally with sheathing an interpetiolar appendage Stipules not deciduous or semi-deciduous Yes Yes Yes Yes Stipules indurate No No No No Stipules becoming fragmented No No No No Inflorescence axes coloured Yes / no No No No Rudgea Carapichea Corolla zygomorphic Yes / no No No No Corolla colour White / brightly coloured White, creamy to yellow-green White, creamy to yellow-green White, creamy to yellow-green Corolla base swollen and gibbous Yes / no No No No Fruit colour Variable Orange / red Variable Variable Pyrene outline shape in transverse section Hemispherical / spherical Hemispherical / triangular Hemispherical / spherical Hemispherical Pyrene-coat with ventral depression Yes / no No Yes / no Yes Preformed germination slits (PGSs) present Yes / no Yes Yes / no Yes / no PGSs presence/absence; location Lacking / 1–2 on dorsal ribs (sometimes marginal) 2, basal (marginal / ventral) + 1 medioventral Lacking / 2 basal marginal (+ some dorsal slits) Lacking / 1 dorsal Seed-coat with T- or Y- shaped ventral intrusion No No Yes No Seed-coat lacking a red ethanol-soluble pigment (SCP) Yes Yes Yes Yes Endosperm non-ruminate Yes Yes Yes Yes Endosperm with small inner central ventral invagination No No No No 1260 Version of Record (identical to print version). Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis TAXON 61 (6) • December 2012: 1251–1268 leptothyrsa subclade have a verticillate arrangement at each node (Fig. 3H). The morphology of Hodgkinsonia frutescens mostly matches that of Psychotria leptothyrsa and P. aneityensis. Unfortunately, the living specimen representing P. sp. [AUST] was unavailable and no information is available regarding its morphology. The Samoana subclade representatives have stipules that are papyraceous and calyptrate (Table 4). Readea membranacea Gillespie has the following autapomorphies that might explain its previous recognition as a distinct genus (Gillespie, 1930; Smith & Darwin, 1988): corolla and calyx fleshy (Fig. 3C); corolla and calyx tubes funnel-shaped and their lobes long and reflexed (Gillespie, 1930: fig. 49); calyx tube enclosing the corolla tube completely in bud and almost completely at anthesis (Fig. 3A; Gillespie, 1930: fig. 49); fruits long and oblongoid (Table 4; Fig. 3L). The carnea subclade is the most spectacular subclade because it contains species that possess: stipules calyptrate or lanceolate (Fig. 3G); aestivation reduplicate (Fig. 3B & E); calyx well developed, inflated (bowl-shaped) or enlarged (plate-shaped), with erect lobes; and corolla funnel-shaped with long lobes (Table 4; Fig. 3B & E). The species are often morphologically similar and have the same general appearance, and this may explain difficulties encountered by taxonomists to circumscribe species, as exemplified by the polyphyletic Psychotria carnea (Fig. 2). The morphological study at the generic level reveals that species of clade G share a common set of characters (see the aforementioned list in the Results section and in Table 5), and these support its monophyly. Generic relationships between Neotropical Margaritopsis and clade G. — As defined by Andersson (2001, 2002a) and Robbrecht & Manen (2006), the species of clade G show the typical combination of characters found in Palicoureeae (characters shaded grey in Table 5). These taxa are closely related to Neotropical Margaritopsis, as shown by our phylogenetic analyses, which reveals the inclusion of clade G in Margaritopsis (Fig. 1; Table 5). The generic description of Margaritopsis … observed within the tribe in roman, and unique or discriminative for one or more genera in bold. Chassalia Geophila Hymenocoleus Chazaliella Neotropical Margaritopsis Clade G species Shrubs / trees / epiphytes / climbers Creeping herbaceous plants with stolons Creeping herbaceous plants with stolons / herbaceous Shrubs Shrubs Shrubs / small trees Greenish brown Greenish brown Greyish black Greyish Pale yellow greenish Pale yellow greenish No No No Yes Yes / no No No No No Yes / no Yes / no Yes / no Entire or bilobed, free or sheathing Bilobed or entire Bilobed, with a membranous sheath inside Entire, bilobed or lacerate Entire, bilobed or dentate, sometimes sheathing Entire, sheathing or calyptrate Yes Yes Yes Yes Yes Yes Yes (corky) No No Yes (corky) Yes Yes Yes No No Yes Yes Yes Yes No No No No No Yes / no No No No No No White / brightly coloured White, creamy to yellow-green White, creamy to yellow-green White, creamy to yellow-green White, creamy to yellow-green White, creamy to yellow-green No No No No No No Red / blue / black Red / blue / black Orange / red Orange / red Orange / red Orange / red Hemispherical / spherical Hemispherical Hemispherical Hemispherical Hemispherical Hemispherical Yes No Yes No No No Yes No Yes Yes Yes Yes 1 basal median dorsal + 2 marginal Lacking 2 basal dorsal + a funicular pore 2 basal marginal ventral 2 basal marginal ventral (+ some dorsal slits) 2 basal marginal ventral No No No No No No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No No No Yes / no Yes / no Yes Version of Record (identical to print version). 1261 Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis 1262 Version of Record (identical to print version). TAXON 61 (6) • December 2012: 1251–1268 TAXON 61 (6) • December 2012: 1251–1268 Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis Table 6. Species that belong to or are attributed to clade G based on molecular evidence and morphological investigation. Type specimens that were consulted in herbaria are indicated in bold, and those observed as images in roman. Region Australia Fiji Vanuatu Fiji The Philippines Fiji / Tonga Samoa New Caledonia Bismarck Archipelago Fiji Borneo / Java / Sumatera Java to Lesser Sunda Islands Fiji / Samoa / Tonga Bismarck Archipelago, New Guinea Samoa New Guinea Java Fiji Society Archipelago Malesia to West Pacific New Caledonia New Guinea, the Philippines Solomons Bismarck Archipelago SE China to tropical Asia / (New Guinea?) Bismarck Archipelago New Caledonia Fiji / Samoa New Guinea Indo-China to West Malesia Bismarck Archipelago Samoa Samoa Bismarck Archipelago, Solomons Samoa South China to Vietnam Solomons / (Vanuatu?) Samoa Indo-China New Caledonia Vanuatu Fiji Name Hodgkinsonia frutescens C.T. White Psychotria amoena A.C. Sm. Psychotria aneityensis Guillaumin Psychotria archboldiana Fosberg Psychotria balabacensis Merr. Psychotria carnea (G. Forst.) A.C. Sm. Psychotria chlorocalyx K. Schum. Psychotria collina Labill. Psychotria damasiana Sohmer Psychotria evansensis A.C. Sm. Psychotria expansa Blume Psychotria extensa Miq. Psychotria forsteriana A. Gray Psychotria galorei Sohmer Psychotria geminodens K. Schum. Psychotria goodenoughiensis Sohmer Psychotria horsfieldiana Miq. Psychotria incompta A.C. Sm. Psychotria lepiniana (Baill. ex Drake) Drake Psychotria leptothyrsa Miq. Psychotria lyciiflora (Baill.) Schltr. Psychotria membranifolia Bartl. ex DC. Psychotria merrilliana Sohmer Psychotria monopedicellata Sohmer Psychotria montana Blume Psychotria novohiberiensis Sohmer Psychotria oleoides (Baill.) Schltr. Psychotria oncocarpa K. Schum. Psychotria purariensis Sohmer Psychotria rostrata Blume Psychotria saidoriensis Sohmer Psychotria samoana K. Schum. Psychotria savaiiensis Rech. Psychotria schmielei Warb. Psychotria sclerocarpa Whistler Psychotria straminea Hutch. in C.S. Sargent Psychotria trichostoma Merr. & L.M. Perry Psychotria vaupelii Whistler Psychotria vietnamensis Ruhsam Psychotria sp. NC Psychotria sp. V Readea membranacea Gillespie Type seen X X X X X X Taxonomic treatment Smith & Darwin, 1988 Smith & Darwin, 1988 Sohmer & Davis, 2007 Whistler, 1986, Smith & Darwin, 1988 Whistler, 1986 X Sohmer, 1988 Smith & Darwin, 1988 X X X X Whistler, 1986, Smith & Darwin, 1988 Sohmer, 1988 Whistler, 1986 Sohmer, 1988 X X X X Smith & Darwin, 1988 X Sohmer, 1988 Sohmer, 1988, Sohmer & Davis, 2007 Sohmer, 1988 Sohmer, 1988 X X X X Sohmer, 1988 X X X X X Sohmer, 1988 Whistler, 1986 Whistler, 1986 Sohmer, 1988 Whistler, 1986 X Sohmer, 1988 Whistler, 1986 Whistler, 1986, Smith, 1988 Sohmer, 1988 X X Smith & Darwin, 1988 ◄ Fig. 3. Examples of species of the five subclades of clade G and the Readea lineage. A, Readea membranacea Gillespie (Barrabé & al. 1128, NOU); B, Psychotria sp., carnea subclade (Barrabé 1060, NOU); C, Readea membranacea (Barrabé & al. 1128, NOU); D, Readea membranacea (Barrabé & al. 1128bis, NOU); E, Psychotria archboldiana Fosberg (Barrabé & al. 1123a, NOU); F, Psychotria lyciiflora (Baill.) Schltr. (Barrabé & Bariole 747, NOU); G, Psychotria sp., carnea subclade (Barrabé & al. 1126, NOU); H, Hodgkinsonia frutescens C.T. White (Baba & al. 786, NOU); I, Psychotria forsteriana A. Gray (Barrabé & al. 1065, NOU); J, Psychotria sp., carnea subclade (Barrabé & al. 1123bis, NOU); K, Psychotria forsteriana (Buerki & al. 162, NOU); L, Readea membranacea (Barrabé & al. 1128bis). — Credits: A–G, I–L: L. Barrabé; H: F. Rigault. Version of Record (identical to print version). 1263 Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis proposed by Andersson (2001) and Taylor (2005) mostly includes the morphology found in species of clade G (cf. generic characters shown in Table 5), even though some taxa have features that are unique or rare in Palicoureeae. For instance, the carnea subclade has winged buds (Fig. 3B & E), a character shared with several species of Chassalia (Piesschaert, 2001; Piesschaert & al., 2001). We could not find any morphological characters that would allow the species of clade G to be separated from Neotropical Margaritopsis (Table 5). Consequently, there is no reason to consider that the species of clade G constitute a genus distinct from Margaritopsis. On this basis we TAXON 61 (6) • December 2012: 1251–1268 suggest that the circumscription and characterization of Margaritopsis should be expanded to incorporate the 19 species of clade G as identified here, and include the 28 additional species we have attributed to the group (Table 6). The adjustments to the characterization of Margaritopsis would need to include the characters identified as unique for each internal group of clade G (see above; Table 4). We refrain from proposing a new formal description of the genus in the present study, in lieu of further taxonomic work, and especially for species from South-East Asia and Indonesia. If we accept that Margaritopsis is paraphyletic without the inclusion of clade G species, at A Neotropical Margaritopsis clade G 20 ? 27 Chazaliella 47 B 8?/?/2 South-East Asia Philippines 1?/?/0 2/1/0 7/4/0 4/1/0 Hawaii 3/2/3 Indonesia Micronesia Papua New Guinea 8/6/1 Bismarck Archipelago 1/1/1 Solomons Samoa Vanuatu 5/2/1 Tropical Australia New Caledonia Fiji Tonga Society archipelago 2/0/0 8/5/5 2/2/2 4/4/4 Fig. 4. A, Geographical distribution and species richness of clade G, Neotropical Margaritopsis and African Chazaliella. B, Species distribution of clade G in South-East Asia, Malesia, tropical Australia and tropical South Pacific. Areas where the species diversity is known are shaded in dark grey, where it is unknown in light grey. For each area the numbers indicate total number of species present / endemic species / species sampled in this study, respectively. Information for total number of species and number of endemics was generated from herbarium studies and published works (Whistler, 1986; Smith & Darwin, 1988; Sohmer, 1988; Fosberg & al., 1993; Sohmer & Davis, 2007; Chen & Taylor, 2011). 1264 Version of Record (identical to print version). TAXON 61 (6) • December 2012: 1251–1268 Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis least 46 new combinations or new names are required, as most of these species are presently placed in Psychotria (Table 6). Readea (Gillespie, 1930) and its type species have already been placed in Margaritopsis (Andersson, 2001, 2002c). Hodgkinsonia frutescens also needs to be formally placed in Margaritopsis. Phylogenetic analyses have placed the type species of Hodgkinsonia (H. ovatiflora F. Muell.) in tribe Guettardeae (Cinchonoideae subfamily; Manns & Bremer, 2010), and consequently this genus will not be synonymized with Margaritopsis. Taxonomic and nomenclatural implications related to these new findings will be formalized in a future publication (Barrabé & al., in prep.) once the nomenclatural details have been carefully checked. Major biogeographical patterns within clade G. — The phylogenetic topology suggests that Palicoureeae have colonized the AMP region at least three times in the following lineages: Chassalia, Geophila, and clade G (Fig. 1). Concerning the origin and diversification of clade G, three geographical levels can be examined. Firstly, at a large geographical scale, the close relationships between a clade of the AMP region and Neotropical species (composed of Margaritopsis taxa; Figs. 1 & 4) requires complex assumptions. Indeed, considering the shape of our phylogenetic tree, three hypotheses could be evaluated for the pantropical distribution of the Margaritopsis/clade G/Chazaliella clade, involving three different main dispersal routes: (1) East–West long-distance dispersal from the Neotropics to the AMP region; (2) continuous West–East dispersal from the Neotropics to the AMP region through Africa; (3) dispersal from Africa to both the Neotropics and the AMP region. Additional sampling of African Chazaliella and Neotropical Margaritopsis is necessary to test these three hypotheses by placing these species in future biogeographical analyses, as only 17 species of AMP Psychotria, Readea membranacea, Hodgkinsonia frutescens, one of 20 species of African Chazaliella, and five of 27 Neotropical Margaritopsis have so far been included in phylogenetic studies. Although uncommon in Rubiaceae, a similar pattern was observed in the Chiococeeae tribe, which includes two tropical South Pacific lineages, one consisting of Morierina Vieill. and Thiollierea Montrouz. (as Bikkia Reinw. p.p.) and the second associating Badusa A. Gray and Bikkia p.p., nested within a diversified Neotropical clade (Motley & al., 2005; Manns & Bremer, 2010). In Augusta Pohl two of the four species occur in the Neotropics, whereas the two others are found in New Caledonia and in Fiji (Kirkbride, 1997). Secondly, a biogeographical history can be postulated from the internal phylogenetic topology of clade G (Fig. 2). The Vietnamese subclade (from South-East Asia) is placed as sister to all remaining species of clade G. The grouping of the carnea and Samoana subclades and Readea membranacea allows us to delimit a Polynesian and Fijian clade. The collina subclade (from New Caledonia and Vanuatu) and leptothyrsa subclade (from Malesia, Melanesia and Australia) are inserted between these two latter monophyletic groups. This topology may imply the existence of a dispersal route between mainland South-East Asia (western part of the AMP region) and tropical South Pacific islands (eastern part) via Malesia and Melanesia (middle part), as commonly observed in other plant groups (Fig. 4; Keppel & al., 2009). Thirdly, and at a yet finer scale, within each of the subclades of clade G, subsequent dispersals to neighbouring islands seem to have occurred and contributed to the current geographic distribution shown in Fig. 4. The nested placement of the Tahitian Psychotria lepiniana within a Fijian/Tongan grade, composed by the carnea subclade and the Readea lineage, could be the result of dispersal from the western to the central Pacific Ocean (Fig. 2). As species from tropical Australia are nested within a Malesian/Melanesian grade (with P. aneityensis and P. leptothyrsa), only one colonization event seems to have occurred in tropical Australia from the Malesian or the Melanesian area (Fig. 2). The placement of P. sp. ‘V’ from Vanuatu nested among New Caledonian species in the ‘collina subclade’ (Fig. 2) is likely to be the result of dispersal from New Caledonia, a feature also observed in other groups (e.g., Geissois Labill. or Schefflera J.R. Forst. & G. Forst.; Pillon, 2011; Plunkett & Lowry, 2012). PERSPECTIVES The phylogenetic analyses and morphological framework presented here provides an improved knowledge for Palicoureeae representatives of the Old World, specifically from the AMP region and comprising the clade we have labelled as clade G. It is clear that species of this clade have strong affinities with the Neotropical Margaritopsis and should be merged with it when adequate nomenclatural work has been done. Different kinds of biogeographical processes seem to have been involved in shaping the current species richness in the AMP region for the group, and these deserve to be fully investigated by expanding taxon sampling, undertaking appropriate analyses, and by testing different biogeographical models (Buerki & al., 2011; Ronquist & Sanmartin, 2011). Otherwise the combination of morphological homogeneity, the large geographical range in the AMP region, the high ecological plasticity, and low local species richness, indicate a special adaptive strategy for the species of clade G, in which dispersal and establishment is more important than the ability to radiate significantly in each archipelago and landmass of the area. ACKNOWLEDGMENTS We are grateful to the members of the Laboratoire de Botanique et d’Ecologie Végétale Appliquées and the herbarium at IRD Nouméa (NOU), especially Frédéric Rigault, Gilles Dagostini, Williams Nigote, Tanguy Jaffré, Jacqueline Fambart-Tinel, Michèle Magat and Céline Grignon. We thank Yumiko Baba and Darren Crayn (Tropical Australian Herbarium of Cairns, James Cook University), Marika V. Tuiwawa, Alifereti Naikatini and Fiona Tuiwawa (South Pacific Regional Herbarium, University of the South Pacific of Suva), Isaac A. Rounds (Conservation International, Fiji), Christopher Davidson and Sharon Christoph, Martin Callmander and Pete Lowry (Missouri Botanical Garden, Saint Louis) for their unconditional assistance Version of Record (identical to print version). 1265 Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis with Australian and Fijian fieldworks. We thank Clarisse Majorel and Laure Hannibal (Laboratoire des Symbioses Tropicales Méditerranéennes, IRD Nouméa) for their help in the laboratory; Edith Kapinos and Laszlo Csiba (Royal Botanic Gardens [RBG], Kew), Sylvain Razafimandimbison, and Birgitta Bremer (Bergius Foundation, Stockholm) for assisting us with access to DNA samples, and Yohan Pillon (University of Hawaii at Hilo) for his comments and advice on this study and an earlier version of the manuscript. We are also grateful to the currators and the staff of the herbaria of the Museum National d’Histoire Naturelle of Paris (Jean-Noël Labat †), and RBG Kew (Nina M. Davies and Sally E. Dawson), for their help accessing collections. For permission to conduct fieldwork, collect and export specimens we thank the Direction de l’Environnement of the Province Sud and the Direction du Développement économique et de l’Environnement of the Province Nord in New Caledonia, the Department of Environment and Resource Management of the Queensland Government in Australia, and the Department of Environment of the Government of Fiji. Material from Santo (Vanuatu) was collected during the Santo2006 expedition. Part of this work was funded through the BIONEOCAL and ULTRABIO projects supported by a grant from the French Agence Nationale de la Recherche. We also wish to thank two anonymous reviewers for providing useful comments that have improved the manuscript. LITERATURE CITED Aldrich, J., Cherney, B.W., Merlin, E. & Christopherson, L. 1988. The role of insertions / deletions in the evolution of the intergenic region between psbA and trnH in the chloroplast genome. Curr. Genet. 14: 137–146. Andersson, L. 2001. Margaritopsis (Rubiaceae, Psychotrieae) is a pantropical genus. Syst. & Geogr. Pl. 71: 73–85. Andersson, L. 2002a. Relationships and generic circumscription in the Psychotria complex (Rubiaceae, Psychotrieae). Syst. & Geogr. Pl. 72: 167–202. Andersson, L. 2002b. Re-establishment of Carapichea (Rubiaceae, Psychotrieae). Kew Bull. 57: 363–374. Andersson, L. 2002c. Validation of three new combinations in Margaritopsis (Rubiaceae, Psychotrieae). Syst. & Geogr. Pl. 72: 230. Andersson, L. & Rova, J.H.E. 1999. The rps16 intron and the phylogeny of Rubioideae (Rubiaceae). Pl. Syst. Evol. 214: 161–186. Baldwin, B.G., Sanderson, M.J., Porter, J.M., Wojciechowski, M.F., Campbell, C.S. & Donoghue, M.J. 1995. The ITS region of nuclear ribosomal DNA: A valuable source of evidence on angiosperm phylogeny. Ann. Missouri. Bot. Gard. 82: 247–277. Bone, R.E., Strijk, J.S., Fritsch, P.W., Buerki, S., Strasberg, D., Thebaud, C. & Hodkinson, T.R. 2012. Phylogenetic inference of Badula (Primulaceae), a rare and threatened genus endemic to the Mascarene Archipelago. Bot. J. Linn. Soc. 169: 284–296. Bremer, B. 2009. A review of molecular phylogenetic studies of Rubiaceae. Ann. Missouri. Bot. Gard. 96: 4–26. Bremer, B. & Eriksson, T. 2009. Timetree of Rubiaceae – Phylogeny and dating the family, subfamilies and tribes. Int. J. Pl. Sci. 170: 766–793. Bremer, B. & Manen, J.-F. 2000. Phylogeny and classification of the subfamily Rubioideae (Rubiaceae). Pl. Syst. Evol. 225: 43–72. Buerki, S., Forest, F., Alvarez, N., Nylander, J.A.A., Arrigo, N. & Sanmartín, I. 2011. An evaluation of new parsimony-based versus parametric inference methods in biogeography: A case study using the globally distributed plant family Sapindaceae. J. Biogeogr. 38: 531–550. 1266 TAXON 61 (6) • December 2012: 1251–1268 Chen, T. & Taylor, C.M. 2011. Rubiaceae. Pp. 5–368 in: Chen, T., Luo, X., Zhu, H., Taylor, C.M., Ehrendorfer, F., Lantz, H., Funston, M. & C. Puff (eds.), Flora of China, vol. 19. Beijing: Science Press; St. Louis: Missouri Botanical Garden Press. http://www.efloras .org/volume_page.aspx?volume_id=2019&flora_id=2. Criscuolo, A. & Gribaldo, S. 2010. BMGE (Block Mapping and Gathering with Entropy): A new software for selection of phylogenetic informative regions from multiple sequence alignments. B. M. C. Evol. Biol. 10: 210, doi: 10.1186/1471-2148-10-210. Davis, A.P. & Bridson, D.M. 2001. A taxonomic revision of the genus Dolianthus (Rubiaceae). Blumea 46: 421–446. Davis, A.P. & Bridson, D.M. 2004. A revision of the genus Amaracarpus (Rubiaceae, Psychotrieae). Blumea 49: 25–68. Davis, A.P., Bridson, D., Jarvis, C. & Govaerts, R. 2001.The typification and characterization of the genus Psychotria L. (Rubiaceae). Bot. J. Linn. Soc. 135: 35–42. Davis, A.P., Govaerts, R., Bridson, D.M., Ruhsam, R., Moat, J. & Brummitt, N.A. 2009. A global assessment of distribution, diversity, endemism and taxonomic effort in the Rubiaceae. Ann. Missouri Bot. Gard. 96: 68–78. Davis, A.P., Tosh, J., Ruch, N. & Fay, M.F. 2011. Growing coffee: Psilanthus (Rubiaceae) subsumed on the basis of molecular and morphological data; Implications for the size, morphology, distribution and evolutionary history of Coffea. Bot. J. Linn. Soc. 167: 357–377. Felsenstein, J. 1985. Confidence limits on phylogenies: An approach using the bootstrap. Evolution 39: 783–791. Fosberg, F.R., Sachet, M.-H. & Oliver, R.L. 1993. Flora of Micronesia, pt. 5, Bignoniaceae-Rubiaceae. Smithsonian Contr. Bot. 81: 1–135. Frodin, D. 2004. History and concepts of big plant genera. Taxon 53: 741–752. Gillespie, J.W. 1930. New plants from Fiji. I. Bernice P. Bishop Mus. Bull. 74: 1–41. Govaerts, R., Ruhsam, M., Andersson, L., Robbrecht, E., Bridson, D., Davis, A.P., Schanzer, I. & Sonke, B. 2011. World checklist of Rubiaceae. Facilitated by the Royal Botanic Gardens, Kew. http:// apps.kew.org/wcsp/ (accessed 1 Jan. 2012). Hooker, J.D. 1873. Rubiaceae. Pp. 7–151 in: Bentham, G. & Hooker, J.D. (eds.), Genera plantarum, vol. 2. London: Lovell Reeve, Williams & Norgate. Keppel, G., Lowe, A.J. & Possingham, H.P. 2009. Changing perspectives on the biogeography of the tropical South Pacific: Influences of dispersal, vicariance and extinction. J. Biogeogr. 36: 1035–1054. Kim, K.-J. & Jansen, R.K. 1995. ndhF sequence evolution and the major clades in the sunflower family. Proc. Natl. Acad. Sci. U.S.A. 92: 10379–10383. Kirkbride, J.H. 1997. Manipulus Rubiacearum – VI. Brittonia 49: 354–379. Manns, U. & Bremer, B. 2010. Towards a better understanding of intertribal relationships and stable tribal delimitations within Cinchonoideae s.s. (Rubiaceae). Molec. Phylogen. Evol. 56: 21–39. Motley, T.J., Wurdack, K.J. & Delprete, P.G. 2005. Molecular systematics of the Chiococceae-Catesbaeeae complex (Rubiaceae): Flower and fruit evolution and systematic implications. Amer. J. Bot. 92: 316–329. Mouly, A., Razafimandimbison, S.G., Khodabandeh, A. & Bremer, B. 2009. Phylogeny and classification of the species-rich pantropical showy genus Ixora (Rubiaceae-Ixoreae) with indications of geographical monophyletic units and hybrids. Amer. J. Bot. 96: 686–706. Mueller-Dombois, D. & Fosberg, F.R. 1998. Vegetation of the tropical Pacific islands. New York: Springer. Nepokroeff, M., Bremer, B. & Sytsma, K.J. 1999. Reorganization of the genus Psychotria and tribe Psychotrieae (Rubiaceae) inferred from ITS and rbcL sequence data. Syst. Bot. 24: 5–27. Nepokroeff, M., Sytsma, K.J., Wagner, W.L. & Zimmer, E.A. 2003. Reconstructing ancestral patterns of colonization and dispersal Version of Record (identical to print version). TAXON 61 (6) • December 2012: 1251–1268 Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis in the Hawaiian understory tree genus Psychotria (Rubiaceae): A comparison of parsimony and likelihood approaches. Syst. Biol. 52: 820–838. Nylander, J.A.A., Ronquist, F., Huelsenbeck, J.P. & Nieves-Aldrey, J.L. 2004. Bayesian phylogenetic analysis of combined data. Syst. Biol. 53: 47–67. Oxelman, B., Liden, M. & Berglund, D. 1997. Chloroplast rps16 intron phylogeny of the tribe Sileneae (Caryophyllaceae). Pl. Syst. Evol. 206: 393–410. Petit, E. 1964. Les espèces Africaines du genre Psychotria L. (Rubiaceae) – I. Bull. Jard. Bot. État Bruxelles 34: 1–229. Petit, E. 1966. Les espèces Africaines du genre Psychotria L. (Rubiaceae) – II. Bull. Jard. Bot. État Bruxelles 36: 65–190. Piesschaert, F. 2001. Carpology and pollen morphology of the Psychotrieae (Rubiaceae-Rubioideae). Dissertation, Katholieke Universiteit Leuven, Leuven, Belgium. Piesschaert, F., Robbrecht, E., Poulsen, A.D. & Smets, E. 1999. Pyrene and pollen observations in the pantropical genus Geophila (Rubiaceae-Psychotrieae). Nordic J. Bot. 19: 93–100. Piesschaert, F., Robbrecht, E. & Smets, E. 2001. Patterns in pyrenes: The systematic significance of pyrene morphology in Chassalia (Rubiaceae-Psychotrieae) and related genera. Flora 196: 121–131. Pillon, Y. 2011. Focus on Geissois (Cunoniaceae): Another example of the Melanesian connection. Pp. 93–94 in: Bouchet, P., Le Guyader, H. & Pascal, O. (eds.), The natural history of Santo. Paris, Marseille: Muséum National d’Histoire Naturelle, Institut de Recherche pour le Développement, Pro-Natura International. Plunkett, G.M. & Lowry, P.P., II 2012. Phylogeny and diversification in the Melanesian Schefflera clade (Araliaceae) based on evidence from nuclear rDNA spacers. Syst. Bot. 37: 279–291. Popp, M. & Oxelman, B. 2001. Inferring the history of the polyploid Silene aegaea (Caryophyllaceae) using plastid and homoeologous nuclear DNA sequences. Molec. Phylogen. Evol. 20: 474–481. Posada, D. 2008. jModelTest: Phylogenetic model averaging. Molec. Biol. Evol. 25: 1253–1256. Rambaut, A. & Drummond, A.J. 2007. Tracer, version 1.5. Program distributed by the authors. Institute of Evolutionary Biology, University of Edinburgh. http://tree.bio.ed.ac.uk/software/tracer/. Razafimandimbison, S.G. & Bremer, B. 2002. Phylogeny and classification of Naucleeae s.l. (Rubiaceae) inferred from molecular (ITS, rbcL, and trnT-F) and morphological data. Amer. J. Bot. 89: 1027–1041. Razafimandimbison, S.G. & Bremer, B. 2011. Nomenclatural changes and taxonomic notes in the tribe Morindeae (Rubiaceae). Adansonia 33: 283–309. Razafimandimbison, S.G., Kellogg, E.A. & Bremer, B. 2004. Recent origin and phylogenetic utility of divergent ITS putative pseudogenes: A case study from Naucleeae (Rubiaceae). Syst. Biol. 53: 177–192. Razafimandimbison, S.G., Rydin, C. & Bremer, B. 2008. Evolution and trends in the Psychotrieae alliance (Rubiaceae) – A rarely reported evolutionary change of many-seeded carpels from oneseeded carpels. Molec. Phylogen. Evol. 48: 207–223. Robbrecht, E. 1975. Hymenocoleus, a new genus of Psychotrieae (Rubiaceae) from tropical Africa. Bull. Jard. Bot. Natl. Belg. 45: 273–300. Robbrecht, E. 1988. Tropical woody Rubiaceae. Opera Bot. Belg. 1: 1–271. Robbrecht, E. & Manen, J.-F. 2006. The major evolutionary lineages of the coffee family (Rubiaceae, angiosperms). Combined analysis (nDNA and cpDNA) to infer the position of Coptosapelta and Luculia, and supertree construction based on rbcL, rps16, trnLtrnF and atpB-rbcL data. A new classification in two subfamilies, Cinchonoideae and Rubioideae. Syst. & Geogr. Pl. 76: 85–146. Ronquist, F. & Huelsenbeck, J.P. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572–1574. Ronquist, F. & Sanmartín, I. 2011. Phylogenetic methods in biogeography. Annual Rev. Ecol. Evol. Syst. 42: 441–464. Rydin, C., Razafimandimbison, S.G. & Bremer, B. 2008. Rare and enigmatic genera (Dunnia, Schizocolea, Colletoecema), sisters to species-rich clades: Phylogeny and aspects of conservation biology in the coffee family. Molec. Phylogen. Evol. 48: 74–83. Rydin, C., Razafimandimbison, S.G., Khodabandeh, A. & Bremer, B. 2009. Evolutionary relationships in the Spermacoceae alliance (Rubiaceae) using information from six loci: Insights into systematic affinities of Neohymenopogon and Mouretia. Taxon 58: 793–810. Sanderson, M.J. & Doyle, J.J. 1992. Reconstruction of organismal and gene phylogenies from data on multigene families: Concerted evolution, homoplasy, and confidence. Syst. Biol. 41: 4–17. Sang, T., Crawford, D.J. & Stuessy, T.F. 1997. Chloroplast DNA phylogeny, reticulate evolution, and biogeography of Paeonia (Paeoniaceae). Amer. J. Bot. 84: 1120–1136. Sauvalle, F.A. 1869. Flora Cubana. Revisio Catalogi Grisebachiani vel index plantarum cubensium. Anales Acad. Ci. Med. Habana 6: 146–147. Shaw, J., Lickey, E., Beck, J.T., Farmer, S.B., Liu, W., Miller, J., Siripun, K.C., Winder, C.T., Schilling, E.E. & Small, R.L. 2005. The tortoise and the hare II: Relative utility of 21 noncoding chloroplast DNA sequences for phylogenetic analysis. Amer. J. Bot. 92: 142–166. Shimodaira, H. & Hasegawa, M. 1999. Multiple comparisons of loglikelihoods with applications to phylogenetic inference. Molec. Biol. Evol. 16: 1114–1116. Smith, A.C. & Darwin, S.P. 1988. Rubiaceae. Pp. 143–376 in: Smith, A.C. (ed.), Flora vitiensis nova: A new flora of Fiji, vol. 4. Lawai: National Tropical Botanical Garden. Sohmer, S.H. 1988. The nonclimbing species of the genus Psychotria (Rubiaceae) in New Guinea and the Bismarck Archipelago. Bishop Mus. Bull. Bot. 1: 1–339. Sohmer, S.H. & Davis, A.P. 2007. The genus Psychotria (Rubiaceae) in the Philippine Archipelago. Sida Bot. Misc. 27: 1–247. Steyermark, J.A. 1972. Rubiaceae. Pp. 227–832 in: Maguire, B. & Collaborators (eds.), The botany of the Guayana Highland, Part IX. Mem. New York Bot. Gard. 23: 1–832. Swofford, D.L. 2002. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4.0 Beta. Sunderland, Massachusetts: Sinauer. Taberlet, P., Gielly, L., Pautou, G. & Bouvet, J. 1991. Universal primers for amplification of three non-coding regions of chloroplast DNA. Pl. Molec. Biol. 17: 1105–1109. Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M. & Kumar, S. 2011. MEGA5: Molecular Evolutionary Genetics Analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Molec. Biol. Evol. 28: 2731–2739. Tate, J.A. & Simpson, B.B. 2003. Paraphyly of Tarasa (Malvaceae) and diverse origins of the polyploid species. Syst. Bot. 28: 723–737. Taylor, C.M. 1996. Overview of the Psychotrieae (Rubiaceae) in the Neotropics. Opera Bot. Belg. 7: 261–270. Taylor, C.M. 1997. Conspectus of the genus Palicourea (Rubiaceae: Psychotrieae) with the description of some new species from Ecuador and Colombia. Ann. Missouri. Bot. Gard. 84: 224–262. Taylor, C.M. 2001. Overview of the Neotropical genus Notopleura (Rubiaceae: Psychotrieae), with the description of some new species. Ann. Missouri. Bot. Gard. 88: 478–515. Taylor, C.M. 2005. Margaritopsis (Rubiaceae, Psychotrieae) in the Neotropics. Syst. & Geogr. Pl. 75: 161–177. Taylor, C.M. 2010. The genus Carapichea (Rubiaceae, Psychotrieae). Fifth international Rubiaceae and Gentianales Conference, 6–10 September, Stockholm, Sweden. [Oral presentation.] Taylor, C.M. & Zappi, D. 2006. 162. Rubiaceae (part 5), Tribe 18. Psychotrieae (2). Vol. 79 of: Persson, C. & Ståhl, B. (eds.), Flora of Ecuador. Gothenburg: Department of Plant and Environmental Sciences, University of Gothenburg, in co-operation with Pontificia Universidad Católica del Ecuador, Quito. Version of Record (identical to print version). 1267 Barrabé & al. • Phylogenetic inference of Asian, Australasian and Paciic Margaritopsis Verdcourt, B. 1975. Studies in the Rubiaceae-Rubioideae for the ‘Flora of tropical East Africa’: I. Kew Bull. 30: 247–326. Verdcourt, B. 1977. A synopsis of the genus Chazaliella ( RubiaceaePsychotrieae). Kew Bull. 31: 788–818. Whistler, W.A. 1986. A revision of Psychotria (Rubiaceae) in Samoa. J. Arnold Arbor. 67: 341–370. TAXON 61 (6) • December 2012: 1251–1268 Zappi, D. 2003. Revision of Rudgea (Rubiaceae) in southeastern and southern Brazil. Kew Bull. 58: 513–596. Zarrei, M., Wilkin, P., Ingrouille, M.J., Leitch, I.J., Buerki, S., Fay, M.F. & Chase, M.W. 2012. Speciation and evolution in the Gagea reticulata species complex (Tulipeae; Liliaceae). Molec. Phylogen. Evol. 62: 624–639. Appendix 1. Sequences produced in this study and those previously published. Taxa, voucher information, and GenBank accession numbers. GenBank accession numbers are in the following order: ITS, ndhF, rps16, trnH-psbA, trnT-F. Newly generated sequences are indicated with an asterisk. The superscript numbers refer to the publication where the sequences were first published: 1 Nepokroeff & al. (1999), 2Andersson (2001), 3Andersson (2002b), 4 Rydin & al. (2008), 5 Razafimandimbison & al. (2008), 6Andersson & Rova (1999), 7 Andersson (GenBank, unpub.), 8Andersson & Taylor (GenBank, unpub.). Amaracarpus nematopodus (F. Muell.) P.I. Forst. 1, Australia, Queensland, Atherton Tableland, Barrabé & al. 1030 (NOU), JX155060*, JX155105*, JX155152*, JX155192*, JX155011*; 2, Vanuatu, Iles Torres, Hiu, Pillon 1072 (NOU), JX155074*, JX155119*, JX155166*, –, JX155026*. Carapichea affinis (Standl.) L. Andersson, America, –, AF0720211, –, AF3700393, –, –. Carapichea ipecacuanha (Brot.) L. Andersson, America, –, AF0720201, –, AF3700403, –, –. Carapichea ligularis (Rudge) Delprete, French Guiana, –, AF1493908, –, AF1475672, –, –. Chassalia catatii Drake ex Bremek., Madagascar, Razafimandimbison 480 (UPS), AM9452185, AM9452835, AM9453315, –, AM9453635. Chassalia sp. 1 MADAG, Madagascar, Antsiranana, Diana, Callmander 307 (K), JX155100*, JX155147*, –, –, JX155054*. Chassalia sp. 2 MADAG, Madagascar, Antsiranana Province, Montagne d’Ambre, 1030 m, Ramandimbimanana 140 (K), JX155092*, JX155138*, JX155181*, –, –. Chassalia sp. 1 VIET, Vietnam, Hoa Binh Province, Mai Chau District, Pa co, the forest alongate main road from HoaBinh to son La, 942 m, HNK 610 (K), JX155087*, JX155132*, JX155177*, JX155218*, JX155039*. Chassalia sp. 2 VIET, Vietnam, Quảng Bình Province, Phong Nha-Ke Bang National Park, in forest on the edge of agricultural land and Highway 1, 255 m, Briggs 114 (K), JX155090*, JX155135*, JX155179*, JX155219*, JX155042*. Chassalia sp. SUMA, Sumatra, Jambi Province, Batanghari District, SPAS, 50 m, De Kok 1297 (K), JX155088*, JX155133*, JX155178*, –, JX155040*. Chazaliella abrupta (Hiern) E.M.A. Petit & Verdc. 1, Tanzania, Kilombero District, Uzungwa Mountains, Kihanzi Gorge, 800 m, Davis 2924 (K), JX155089*, JX155134*, –, JX155217*, JX155041*; 2, Tanzania, Morogoro Province, Nguru Mountains, near Mhonda mission, along Chazi River, Bremer 3081 (UPS), JX155101*, JX155148*, –, JX155227*, JX155055*. Craterispermum sp. 2, Madagascar, Razafimandimbison & Ravelonarivo 656 (SPF), AM9452105, AM9452745, AM9453235, –, AM9453545. Craterispermum sp. 3, Madagascar, 24169-SF (P), AM9452115, AM9452755, AM9453245, –, AM9453555. Geophila obvallata Didr., Africa, –, AM9451965, AM9452595, AF3698452, –, EU1455694. Geophila repens (L.) I.M. Johnst., Vanuatu, Santo, Pénaorou, Munzinger 3649 (NOU), JX155079*, JX155124*, –, JX155210*, JX155031*. Hodgkinsonia frutescens C.T. White, Australia, Queensland, Wongabel State Forest, Baba & al. 786 (NOU), JX155061*, JX155106*, JX155153*, JX155194*, JX155012* / JX155013*. Hydnophytum cf. longistylum Becc., Vanuatu, Espiritu Santo Island, Sanama Province, McPherson 19437 (NOU), JX155078*, JX155123*, JX155170*, JX155209*, JX155030*. Hymenocoleus hirsutus (Benth.) Robbr., Africa, –, AF0720181, –, AF3698487, –, –. Margaritopsis astrellantha (Wernham) L. Andersson, Brazil, Parque Estadual Cristalino, Zappi 938 (K), JX155096*, JX155142*, JX155185*, JX155225*, JX155047* / JX155048*. Margaritopsis boliviana (Standl.) C.M. Taylor, Brazil, Parque Estadual Cristalino, Zappi 973 (K), JX155097*, JX155143*, JX155186* / JX155187*, JX155220*, JX155049* / JX155050*. Margaritopsis guianensis (Bremek.) C.M. Taylor, Brazil, Parque Estadual Cristalino, Zappi 980 (K), JX155098*, JX155144*, JX155188*, –, JX155051* / JX155052*. Margaritopsis kappleri (Miq.) C.M. Taylor, Brazil, Parque Estadual Cristalino, Zappi 921 (K), JX155095*, JX155141*, JX155184*, JX155224*, JX155046*. Margaritopsis nudiflora (Griseb.) K. Schum. in H.G.A. Engler & K.A.E. Prantl, Cuba, Oriente Province, Guantánamo, Monte Libanon ad San Fernandez, Ekman 10248 (UPS), AM9451985, –, AF0013406, JX155230*, EU1455684. Notopleura tapajozensis (Standl.) Bremek., Brazil, Parque Estadual Cristalino, Zappi 869 (K), JX155094*, JX155140*, JX155183*, JX155223*, JX155045*. Palicourea crocea (Sw.) Schult. in J.J. Roemer & J.A. Schultes, America, Cordiero 2736 (SP), AF1493228, AM9452805, AF1475108, –, AM9453595. Palicourea nitidella (Müll. Arg.) Standl., Brazil, Parque Estadual Cristalino, Sasaki 1828 (K), JX155093*, JX155139*, JX155182*, JX155222*, JX155044*. Prismatomeris albidiflora Thwaites, Cambodge, Marie 89 (P), AM9452055, AM9452705, AM9453205, –, AM9453515. Prismatomeris beccariana (Baill. ex K. Schum.) J.T. Johanss., Indonesia, Ridsdale 2461 (L), AM9452065, AM9452715, AF3316527, –, AM9453525. Psychotria aneityensis Guillaumin, Vanuatu, Iles Torres, Loh, Pillon 1127 (NOU), JX155075*, JX155120*, JX155167*, JX155207*, JX155027*. Psychotria archboldiana Fosberg, Fiji, Viti Levu, Nadroga Navosa Province, Nadrau district, along the road between Naqelewai and Monasavu, Buerki & al. 170 (NOU), JX155068*, JX155113*, JX155160*, JX155201*, JX155020*. Psychotria asiatica L., China, Hong-Kong (cultivated), Davis s.n. (K), JX155082*, JX155127*, –, –, JX155034*. Psychotria camptopus Verdc., Cameroon, Hong-Kong (cultivated), Maurin 38 (K), JX155084*, JX155129*, JX155174*, JX155214*, JX155036*. Psychotria carnea 1 (G. Forst.) A.C. Sm., Fiji, Vanua Levu, Cakaudrove Province, Wailevu District, West of Savu Savu, Barrabé & Tuiwawa 1113 (NOU), JX155065*, JX155110*, JX155157*, JX155198*, JX155017*. Psychotria carnea 2 (G. Forst.) A.C. Sm., Fiji, Taveuni, Cakaudrove Province, Somo Somo District, Naibili, road from Somo Somo to hydroelectricdam, Barrabé & al. 1123 (NOU), JX155066*, JX155111*, JX155158*, JX155199*, JX155018*. Psychotria collina Labill., New Caledonia, Province Nord, Ponérihouen, Aoupinié, Barrabé & al. 698 (NOU), JX155072*, JX155117*, JX155164*, JX155205*, JX155024*. Psychotria forsteriana A. Gray, Fiji, Viti Levu, Naitasiri Province, Vuna District, Colo-i-Suva Forest Park, Buerki & al. 162 (NOU), JX155067*, JX155112*, JX155159*, JX155200*, JX155019*. Psychotria grandis Sw., Caribbean, Puerto Rico, Municipio Río Grande: El Verde Research Station, Taylor 11745 (MO), JX155080*, JX155125*, JX155171*, JX155211*, JX155032*. Psychotria cf. impercepta A.C. Sm. & S.P. Darwin, Fiji, Viti Levu, Serua Province, West of Nabukeleva village, Barrabé & al. 1086 (NOU), JX155063*, JX155108*, JX155155*, JX155196*, JX155015*. Psychotria lepiniana (Baill. ex Drake) Drake, French Polynesia, Mont Mahutaa, vallée d’Orofero, Orohena, Nadeaud 345 (P), JX155059*, JX155104*, JX155151*, JX155193*, JX155010*. Psychotria leptothyrsa Miq., New Guinea, Madang Province, Drozd & Molem s.n (CBSF), JX155102*, JX155149*, JX155190*, JX155228*, JX155056*. Psychotria lyciiflora (Baill.) Schltr., New Caledonia, Province Sud, Sarraméa, Dogny, Barrabé & Bariole 747 (NOU), JX155069*, JX155114*, JX155161*, JX155202*, JX155021*. Psychotria oleoides (Baill.) Schltr., New Caledonia, Province Sud, Dumbéa, Montagne des Sources, Barrabé & Rigault 658 (NOU), JX155076*, JX155121*, JX155168*, JX155208*, JX155028*. Psychotria parkeri Baker, Madagascar, Toamasina Province, Alaotra-Mangoro Region, Moramanga District, Mantadia National Park, Maurin 132 (K), JX155085*, JX155130*, JX155175*, JX155215*, JX155037*. Psychotria poeppigiana Müll. Arg. in C.F.P.von Martius & auct. suc. (eds.), Ecuador, Bremer 3330 (UPS), AF1494008, AM9452795, AF0027486, –, –. Psychotria samoana K. Schum., Samoa, Savai’i, far inland from Aopo, Bristol 2128 (K), JX155091*, JX155136* / JX155137*, JX155180*, JX155221*, JX155043*. Psychotria schlechteriana K. Krause, New Caledonia, Province Nord, Poya, Vallée des Roches d’Adio, Barrabé & al. 712 (NOU), JX155070*, JX155115*, JX155162*, JX155203*, JX155022*. Psychotria stachyoides Benth., Brazil, Campos do Jordão, Parque Estadual, Trilha do Rio Sapucai, São Paulo, Pirani & al. 4897 (SPF), JX155103*, JX155150*, JX155191*, JX155229*, JX155057* / JX155058*. Psychotria cf. trichostoma Merr. & L.M. Perry, Vanuatu, Iles Torres, Hiu, Pillon 1057 (NOU), JX155073*, JX155118*, JX155165*, JX155206*, JX155025*. Psychotria sp. AUST, Australia, Sydney Botanical Garden (cultivated), Weston 888 (Syd. Bot. Gard. 862980), JX155086*, JX155131*, JX155176*, JX155216*, JX155038*. Psychotria sp. FIJ, Fiji, Viti Levu, Naitasiri Province, Vuna District, Colo-I-Suva, Forest Park, near Suva, Track between Rangers Station and Waisili Falls, Barrabé 1060 (NOU), JX155062*, JX155107*, JX155154*, JX155195*, JX155014*. Psychotria sp. NC, New Caledonia, Province Nord, Poya, Boulinda, Munzinger 4963 (NOU), JX155071*, JX155116*, JX155163*, JX155204*, JX155023*. Psychotria sp. V, Vanuatu, Espiritu Santo Island, Sanama Province, McPherson 19425 (NOU), JX155077*, JX155122*, JX155169*, –, JX155029*. Psychotria sp. 1 VIET, Vietnam, Dong Nai Province, Tan Phu District, Cat Tien National Park, Davis 4031 (K), JX155081*, JX155126*, JX155172*, JX155212*, JX155033*. Psychotria sp. 2 VIET, Vietnam, Hoa Binh Province, Mai Chau District, Cat Tien National Park, HNK 614 (K), JX155083*, JX155128*, JX155173*, JX155213*, JX155035*. Readea membranacea Gillespie, Fiji, Vanua Levu, Cakaudrove Province, Wailevu District, West of Savu Savu, Waisali Rainforest Reserve, Barrabé & al. 1093 (NOU), JX155064*, JX155109*, JX155156*, JX155197*, JX155016*. Rudgea stipulacea (DC.) Steyerm., Brazil, Parque Estadual Cristalino, Zappi 986 (K), JX155099*, JX155145* / JX155146*, JX155189*, JX155226*, JX155053*. For Appendix 2, see Electronic Supplement. 1268 Version of Record (identical to print version).