TAXON 57 (1) • February 2008: 1–17 Khan & al. • Tribal and generic circumscriptions of Sabiceeae Sabiceeae and Virectarieae (Rubiaceae, Ixoroideae): one or two tribes? New tribal and generic circumscriptions of Sabiceeae and biogeography of Sabicea s.l. Saleh A. Khan1,3, Sylvain G. Razafimandimbison2, Birgitta Bremer2 & Sigrid Liede-Schumann1 Department of Plant Systematics, University of Bayreuth, Universitätstr. 30, 95440 Bayreuth, Germany. sigrid.liede@uni-bayreuth.de (author for correspondence) 2 Bergius Foundation, Royal Swedish Academy of Sciences; Department of Botany, Stockholm University, 10691, Stockholm, Sweden 3 Department of Botany, Jahangirnagar University, Savar, Dhaka 1342, Bangladesh 1 The results of two recent phylogenetic studies led to the reinstatement of the tribe Sabiceeae, currently classified in the subfamily Ixoroideae s.l. (Rubiaceae) but with conflicting circumscriptions. In the present study, phylogenetic analyses based on nrITS and trnT-F sequence data of 78 taxa are performed to evaluate the different circumscriptions of Sabiceeae, the generic limits within Sabiceeae, and the biogeography of Sabicea. The polyphyly of Sabiceeae sensu Andersson is confirmed, and Pentaloncha and Temnopteryx are shown not to belong to Ixoroideae s.l. but to the subfamily Rubioideae. Our results favour a broad circumscription of Sabiceeae that includes Ecpoma, Hekistocarpa, Pseudosabicea, Sabicea, Schizostigma, Stipularia, Tamridaea and Virectaria. Sabicea sensu Wernham is not monophyletic unless Ecpoma, Pseudosabicea, Schizostigma, and Stipularia are included. We find no support for the monophyly of Stipularia, Sabicea and Pseudosabicea. Therefore, our newly circumscribed Sabiceeae contains only Hekistocarpa, Sabicea s.l. (Ecpoma, Pseudosabicea, Schizostigma, Stipularia), Tamridaea, and Virectaria. Finally, our analyses indicate several dispersal events of Sabicea species between African phytogeographical regions and continental African origins of the Malagasy, São Tomean, Asian, and Neotropical species of Sabicea via perhaps four independent dispersal events. KEYWORDS: biogeography, nrITS, Rubiaceae, Sabicea, Sabiceeae, trnT-F, Virectarieae INTRODUCTION Grisebach (1861) originally described the pantropical subtribe Sabiceinae (as “Sabicieae”) of the tribe Cinchoneae in the subfamily Cinchonoideae (Rubiaceae) to accommodate two genera, Sabicea Aubl. and Coccocypselum P. Br., both with valvate corolla aestivation. Bremekamp (1934) established a monogeneric tribe Sabiceeae Bremek. (as “Sabiceae”), but no other rubiaceous taxonomists (except Bremekamp, 1966) accepted its tribal status between 1934 and 1996 (see Table 1). The type genus Sabicea was classified in the tribes Mussaendeae Benth. & Hook. f. (Verdcourt, 1958; Hallé, 1961; Hallé, 1966; Steyermark, 1962, 1972, 1974) or Isertieae A. Rich. ex DC. (Kirkbride, 1979, 1982; Robbrecht, 1988, 1993). For tropical Africa Hallé (1961) classified Sabicea and its four traditionally associated genera—Ecpoma K. Schum. (Schumann, 1896), Pentaloncha Hook. f. (Hooker, 1873a), Stipularia P. Beauv. (Palisot-Beauvois, 1807), and Temnopteryx Hook. f. (Hooker, 1873a)—in Mussaendeae. Hallé (1966) placed Ecpoma and Pseudosabicea and Pentaloncha, respectively, in his new subtribes Mussandenae and Urophyllinae of Mussaendeae. Steyermark (1962) classified the Neotropical Pittierothamnus Steyerm. in Mussaendeae s.l. but later merged it with Amphidasya Standl., also endorsed by Kirkbride (1979, 1982) and Robbrecht (1988). Bremekamp (1966) made the last attempt to re-establish Sabiceeae based on simple stipules, axillary inflorescences, and very narrow testa cells rather than bifid stipules, terminal inflorescences, and large testa cells of Mussaendeae. It is notable that some authors, mentioned above, used the tribal name Mussaendeae, although Isertiae had priority over Mussaendeae, because the tribe Mussaendeae contained the type genus (Isertia Schreb.) of Isertiae (Darwin, 1976). Robbrecht (1988) transferred to Isertieae the Indo-Malesian genus Acranthera Arn. ex Meisn. (Meisner, 1838), previously placed by Bremekamp (1966) in its own tribe, and all above genera traditionally associated with Mussaendeae plus Schizostigma Arn. ex Meisn., with the exception of Pentaloncha, which was left unclassified in Rubiaceae (see Table 1). Sabiceeae was resurrected as a result of the morphological-based phylogeny of Isertieae sensu Robbrecht (1988) conducted by Andersson (1996). Stipularia was deeply nested within Sabiceeae (Andersson 1996: Fig. 5) but was not among the nine genera that he included 1 Khan & al. • Tribal and generic circumscriptions of Sabiceeae in his Sabiceeae (Table 1). Based on a rbcL phylogeny Bremer & Thulin (1998) showed that Sabiceeae sensu Andersson (1996) was highly polyphyletic and additionally postulated that Acranthera might perhaps belong to the subfamily Rubioideae, consistent with Alejandro & al.’s (2005: Fig. 1) trnT-F-based phylogeny. Bremer & Thulin (1998) demonstrated for the first time that the broadly delimited Mussaendeae (sensu Hallé, 1961; Hallé, 1966) or Isertieeae (sensu Robbrecht, 1988) was also highly polyphyletic. As a result, they re-established Mussaendeae to accommodate Mussaenda and its satellite genera (Aphaenandra Miq., Heinsia DC., Neomussaenda C. Tange, Pseudomussaenda Wernham, Schizomussaenda H.L. Li) and restricted Isertieae to include the type genus Isertia. They further showed that the African genus Virectaria Bremek., previously placed by Verdcourt (1958) in its own tribe Virectarieae Verdc., is closely related to Pseudosabicea and Sabicea. Accordingly, they tentatively proposed a new circumscription of Sabiceeae, which included Sabicea, Pseudosabicea, the monotypic genus Tamridaea Thulin & B. Bremer, and Virectaria. They considered Stipularia to be closely related to Pseudosabicea and Sabicea based on morphological grounds. On the other hand, Dessein & al. (2001b: 22) considered Virectaria to be an isolated genus within Sabiceeae sensu Bremer & Thulin (1998) based on a few morphological characters (e.g., internal indument and seed anatomy). TAXON 57 (1) • February 2008: 1–17 The rbcL jackknife tree of Dessein & al. (2001a) further confirmed the close relationships between Tamridaea and Virectaria and showed for the first time that the African monotypic genus Hekistocarpa Hook. f. (Hooker, 1873b) is closely related to these two genera. Dessein & al. (2001a: 75) additionally stressed that they “fail to find any morphological characteristics that are common to Hekistocarpa, Pseudosabicea, Sabicea, Tamridaea, and Virectaria of Sabiceeae in a broad sense.” As a result, they resurrected the tribe Virectarieae to accommodate Hekistocarpa, Tamridaea, and Virectaria and restricted Sabiceeae to include only Sabicea and four of its traditionally allied genera (Ecpoma, Pentaloncha, Pseudosabicea, Stipularia). They admitted that their emended Virectarieae was difficult to diagnose morphologically. More recently, Robbrecht & Manen (2006) adopted another broader circumscription of Sabiceeae including eight genera and recognized two subtribes (Table 1): Sabiceinae (Bremek.) Robbr. & Manen and Virectariinae (Verdc.) Robbr. & Manen (= Virectarieae sensu Dessein & al., 2001a). The above conflicting circumscriptions of Sabiceeae drew our attention to further investigations. Sabicea is the most species-rich genus of Sabiceeae with ca. 146 species of scandent shrubs, woody climbers, and scramblers or twiners. With two main centres of diversity, mainland Africa (ca. 82 species) and the Neotropics (ca. 54 species), Sabicea shows a trans-Atlantic Bremekamp (1966) Hallé (1966) Steyermark (1962) Steyermark (1972) Robbrecht (1988) Andersson (1996) Bremer & Thulin (1998) – – Acr – – – Ise Sab Rubi – This study Hallé (1961) – Robbrecht & Manen (2006) Verdcourt (1958) Acranthera Arn. ex Meisn. Dessein & al. (2001a) Genera Bremekamp (1934) Table 1. Previous and new tribal positions of Sabicea and its traditionally and presently allied genera. – Rubi Rubi Amphidasya Standl. – – – – – – Mus Ise Sab Rubi – Uro Ecpoma K. Schum. – – Mus – Mus – – Ise Sab – Sab SabS Sab Hekistocarpa Hook. f. – – – – – – – Hed – – Vir SabV Sab Pentaloncha Hook. f. – – Mus Pau Mus – – Ins Sab – Sab – Rubi Pittierothamnus Steyerm. – – – – – Mus – – Sab – – – – Pseudosabicea N. Hallé – – – – Mus – – Ise Sab Sab Sab SabS Sab Sabicea Aubl. Sab Mus Mus Sab Mus Mus Mus Ise Sab Sab Sab SabS Sab Schizostigma Arn. ex Meisn. – – – – – – – Ise Sab – – SabS Sab Stipularia P. Beauv. – – Mus – Mus – – Ise – Sab Sab SabS Sab Tamridaea Thulin & B. Bremer – – – – – – – – – Sab Vir SabV Sab Temnopteryx Hook. f. – – Mus Pau Mus – – Ise Sab – – – Virectaria Bremek. – Vir – Oph Hed – – Hed – Sab Vir SabV Sab Rubi Acr, tribe Acranthereae; Hed, Hedyotideae; Ins, Incertae sedis; Ise, Isertieae; Mus, Mussaendeae; Oph, Ophiorrhizeae; Pau, Pauridiantheae; Rubi, Rubioideae; Sab, Sabiceeae; SabS, Sabiceeae subtribe Sabiceinae; SabV, Sabiceeae subtribe Virectariinae; Vir, Virectarieae; Uro, Urophylleae s.l. (including Pauridiantheae); – , not mentioned. 2 TAXON 57 (1) • February 2008: 1–17 distribution shared with few other Rubiaceae genera. Six species are endemic to Madagascar (Razafimandimbison & Miller, 1999), three to São Tomé and Príncipe (Joffroy, 2001), and one, S. ceylanica Puff. (Puff & al., 1998), originally described as Schizostigma hirsutum Arn. ex Meisn. (Meisner, 1838), to Sri Lanka. Aublet (1775) originally described Sabicea from South America including two species, S. aspera Aubl. and S. cinerea Aubl., with twining habits and 3–5-locular ovaries. Wernham (1914) proposed a broad circumscription of Sabicea including 105 species from Africa and South America with usually shrubby, climbing or prostrate to scrambling habits, isophylly or anisophylly, entire to fimbriate or laciniate stipules, axillary inflorescences, (sub-) free bracts, valvate corolla lobes, and (2)4–5-locular ovaries. Additionally, he recognized two subgenera in Sabicea based on the combination of habit and leaf and stipule sizes: Sabicea subgen. Stipulariopsis Wernham with 9 species and Sabicea subgen. Eusabicea Wernham with 96 species. Wernham (1914), endorsed by Hiern (1877), Hallé (1961), Hallé (1963, 1966), Andersson (1996), Bremer & Thulin (1998), and Dessein & al. (2001a), recognized the African Stipularia as a distinct genus because of its large stipules and well-developed campanulate involucral bracts completely surrounding the entire inflorescence (Palisot-Beauvois, 1807). On the other hand, Hepper’s (1958) herbarium studies revealed that involucral bracts also occurred in few African Sabicea species (e.g., S. capitellata Benth, S. dewevrei De Wild. & T. Durand, S. cordata Hutch. & Dalziel, and S. urceolata Hepper) with variation in the degree of fusion. As a result, he merged the five described species of Stipularia (S. africana P. Beauv., S. efulenensis Hutch., S. elliptica Schweinf. ex Hiern, S. gabonica Hiern, and S. mollis Wernham) with Sabicea. Both Hallé (1961) and Hallé (1963, 1966) rejected Hepper’s (1958) circumscription of Sabicea and reinstated Stipularia as a distinct genus. Plus, Hallé (1963) viewed Sabicea sensu Wernham (1914) as morphologically heterogeneous and accordingly restricted the genus to include only species with usually lianescent, slender and twining habits, long corollas, (4–)5-locular ovaries, accrescent fleshy axis of ovary, narrow, thin and sessile placentas, and fleshy juicy fruits with often-red carmine pulp. He then described the genus Pseudosabicea to accommodate all the African Sabicea species with creeping or climbing but non-twining habit, short corollas, 2(–3)-locular ovaries, non-fleshy axis of ovary, oblong, peltate and fleshy placentas, and scantly fleshy fruits with colourless pulp. In addition, Hallé (1963) transferred five African Sabicea species (S. bicarpellata K. Schum., S. cauliflora Hiern, S. geantha Hiern, S. gigantostipula K. Schum., S. hierniana Wernham) to the African genus Ecpoma. Arnott (1839) viewed Schizostigma as closely related to Sabicea and more recently, Puff & al. (1998) merged Khan & al. • Tribal and generic circumscriptions of Sabiceeae Schizostigma in Sabicea, which they considered to be closely related to Ecpoma, Pseudosabicea, Stipularia, and Temnopteryx. Both Hallé (1961, 1966) and Puff & al. (1998) totally rejected Hiern’s (1877) attempt to merge Pentaloncha and Temnopteryx with Schizostigma. Although most Rubiaceae systematists seem to accept Sabicea sensu Hallé (1963, 1966), the monophyly of the above conflicting circumscriptions of Sabicea or its close allies have never been assessed before. Previous phylogenetic studies in some Rubiaceae groups based on the nrITS region of rDNA (e.g., Andreasen & al., 1999; Razafimandimbison & al., 2004; Motley & al., 2005) and the trnT-F region of chloroplast DNA (e.g., Razafimandimbison & Bremer, 2002; Alejandro & al., 2005) have demonstrated that both markers are useful for inferring phylogenetic relationships at tribal and generic levels in the family. The main objective of this study is to reconstruct phylogenies of Sabicea and its closely related genera using the sequence data from both the nuclear ribosomal internal transcribed spacer (nrITS1-5.8S-nrITS2 region) and the trnT-F regions of chloroplast DNA (trnT UGU- trnL UAA 5′ exon, trnLUAA 5′ exon-trnLUAA intron, trnLUAA intron-trnLUAA 3′ exon, trnLUAA 3′ exon-trnFGAA). The resulting phylogenies have been used to assess: (1) the conflicting circumscriptions of Sabiceeae, (2) the generic limits within Sabiceeae, and (3) the biogeography of Sabicea. MATERIALS AND METHODS Taxon selection. — A total of 36 species (38 individuals) belonging to Sabicea and 9 genera currently or traditionally associated with Sabiceeae and 19 genera presently placed in Cinchonoideae s.str., Ixoroideae s.l., and Rubioideae (Appendix) were included in the trnT-F analyses to assess the competing circumscriptions of Sabiceeae. Neither Acranthera nor Amphidasya were included in our analyses, as they have recently been shown to be related to Rubioideae (Bremer & Thulin 1998; Alejandro & al. 2005). No material was available for Pittierothamnus. The genus Luculia Sweet (L. grandifolia Ghose) was used as the outgroup taxon, in agreement with its basal position in Rubiaceae (Bremer & al., 1999; Rova & al., 2002). A total of 39 Sabicea species (40 individuals), 8 Pseudosabicea species (9 individuals), 2 species each of Stipularia and Virectaria, 1 Ecpoma species, and 1 individual each of the monotypic Hekistocarpa, Schizostigma, and Tamridaea were included in the the nrITS analyses and all of these accessions excluding Tamridaea were included in the combined nrITS + trnT-F analyses to assess the generic limits within Sabiceeae. One species each of Heinsia DC. (Mussaendeae sensu Bremer & Thulin, 1998), Canthium Lam. (Vanguerieae A. Rich. ex Dumort.), Ixora L. (Ixoreae sensu Andreasen & Bremer, 3 Khan & al. • Tribal and generic circumscriptions of Sabiceeae 2000), and Warszewiczia Klotzsch (Condamineeae sensu Rova & al., 2002), all currently classified in Ixoroideae s.l., were selected to root the nrITS and combined analyses (see Appendix). DNA isolation, amplification, and sequencing. — DNA isolation, amplification, and sequencing of the nrITS region were accomplished following the protocols described in Alejandro & al. (2005). The amplification and sequencing of the trnT-F region were performed following the protocols outlined in Razafimandimbison & Bremer (2002). For each 25 μL PCR reaction we added 15.8 μL dH2O, 2 μL MgCl2 (25 mM), 1.5 μL dNTP (2 mM), 1.0 μL each of forward (P17F, 5′-CTA CCG ATT GAA TGG TCC GGT GAA-3′) and reverse (26S-82R, 5′-TCC CGG TTC GCT CGC CGT TAC TA-3′) primers (10 pmol/μL), 2.5 μL PCR buffer (10×), 0.2 μL TAQ DNA polymerase, and 1.0 μL DNA sample. Sequence alignment and coding of indels. — Forward and reverse sequences generated for both the nrITS and trnT-F regions were assembled using the Perkin Elmer Sequence Navigator, version 1.0.1 and Sequencher 3.1.1 and aligned with the CLUSTAL-W (Thompson & al., 1994) to obtain preliminary alignments, which were subsequently edited manually. We coded all informative indels using the simple gap coding method (Simmons & Ochoterena, 2000) and assessed their effects on the results. Phylogenetic analyses. — Maximum parsimony analyses (hereafter MPA) of both the nrITS and nrITS + trnT-F data were performed with PAUP*, version 4.0b (Swofford, 2000) using the heuristic search settings: MulTrees option on, tree-bisection-reconnection (TBR) branch swapping, swap on best only in effect, 5,000 random addition sequences. We performed MPA of the trnT-F matrix using the same settings, but the searches were frequently terminated prematurely due to the limitation of computer memory. As a result, we analysed the trnT-F data using MulTrees option off, TBR branch swapping, swap on best only in effect, and 10,000 random addition sequences. To estimate homoplasy the consistency index (CI) and retention index (RI) were calculated. To assess the support of the retained clades the bootstrap values were computed using 1,000 replicates, MulTrees option on, TBR branch swapping, and five random addition sequences. We performed parsimony and bootstrap analyses of each of the trnT-F, nrITS and combined nrITS + trnT-F datasets with and without the coded indels to assess the effects of indel coding. No notable conflicts were found in the topologies of the trnT-F, nrITS, and combined nrITS + trnT-F trees or supports to the recognized clades for using the matrices with or without coded indels; therefore, finally we used the trnT-F, nrITS and combined nrITS + trnT-F matrices without indel coding. In all analyses, characters were of equal weight, gaps were treated as missing data, and only parsimony-informative characters were included. Visual 4 TAXON 57 (1) • February 2008: 1–17 comparisons between the trnT-F and nrITS trees from the preliminary parsimony analyses revealed topological conflicts regarding the position of Tamridaea (Figs. 1–2). The agreement on when the datasets should be combined is not generalized (Queiroz & al., 1995) and combinability tests have come under considerable criticism (Bremer 1996; Bayer & al. 2002). Therefore, we combined the trnT-F and nrITS data partitions examining the conflicting position of Tamridaea in the trnT-F and nrITS bootstrap trees. The supports for the conflicting positions of Tamridaea in the trnT-F and ITS trees (BS = 79–87 and BS = 65–78, respectively, depending on the alignment) were high, due to which finally we performed the combined nrITS + trnTF analyses excluding Tamridaea. We performed Bayesian analyses (hereafter BA) in MrBayes, version 3.1.2 (Huelsenbeck & Ronquist, 2001) using the substitution model parameters: Prset statefreqpr = dirichlet (1,1,1,1); Lset nst = 6 rates = equal; selected as best fit under Akaike Information Criterion (AIC) by MrModeltest, version 2.2 (Nylander, 2004) for the uncoded trnT-F, nrITS and combined nrITS + trnTF datasets. In all searches, we used the default settings (MrBayes, version 3.1.2) for all active parameters for the corresponding substitution models, as well as, for the heating scheme. Eight chains under two simultaneous runs, with 100 sample frequencies were executed and monitored up to 3.4–4.0 × 106 Markov chain Monte Carlo (mcmc) generations for arriving at the stationary phase (with average standard deviation of split frequencies < 0.01 and PSRF = about 1.0). After discarding 25% of the samples as burn-in, the graphical presentations of summarized resulting trees were generated in PAUP* and Tree View (Page, 1996.) program. Internodes with posterior probabilities of more than 0.95 were considered as reliable support. In this study we infer the biogeography of Sabicea s.l. based on our results of MPA and BA of combined nrITS + trnT-F datasets. RESULTS Sequence and alignment characteristics. — The characteristics of the non-aligned trnT-F and nrITS sequences of Sabiceeae s.l. and the aligned matrices of the trnT-F and nrITS datasets and the nrITS and trnT-F partitions of the combined datasets are summarized in Table 2. The characteristics of the nrITS sequences and alignment were nearly the same in the nrITS and combined nrITS + trnT-F matrices. The 5.8S subunit was constant in length (165 bp) for all sequenced taxa. trnT-F analyses (Fig. 1). — The trnT-F analyses included 58 sequences, of which 39 are newly published here. The MPA of the trnT-F sequences data resulted in 8,067 equally parsimonious trees (each 977 steps long [L], Khan & al. • Tribal and generic circumscriptions of Sabiceeae TAXON 57 (1) • February 2008: 1–17 CI = 0.679, RI = 0.879). All ingroup taxa were resolved in three strongly supported (BS = 100, PP = 1.00) major clades, corresponding to the subfamilies Rubioideae, Cinchonoideae s.str., and Ixoroideae s.l. (Bremer & al., 1999). The investigated members of Sabiceeae sensu Andersson (1996) were resolved in three separate highly supported subclades (Fig. 1): the Pentaloncha clade (BS = 100, PP = 1.00) and the Temnopteryx clade (BS = 100, PP = 1.00) both nested in Rubioideae, and the Ecpoma-Pseudosabicea-Sabicea-Schizostigma-Stipularia clade (BS = 83, PP = 1.00; called Sabiceeae s.str. hereafter) nested in Ixoroideae s.l. Within Ixoroideae s.l. Virectarieae sensu Dessein & al. (2001b), represented by Hekistocarpa minutiflora Hook. f., Virectaria multiflora (Sm.) Bremek. and V. procumbens (Sm.) Bremek., and Tamridaea capsulifera (Balf. f.) Thulin & B. Bremer, was not resolved as monophyletic. Virectaria multiflora and V. procumbens formed a strongly supported (BS = 100, PP = 1.00) monophyletic group, whereas H. minutiflora was left unresolved. Tamridaea capsulifera was resolved with moderate (BS = 80) and high (PP = 1.00) support, respectively, in the MPA and BA as sister to Sabiceeae s.str. The non-monophyletic Virectarieae sensu Dessein & al. (2001b) and Sabiceeae s.str. together (hereafter called Sabiceeae s.l.) formed a highly supported (BS = 100, PP = 1.00) monophyletic group. All studied Neotropical Sabicea species, with the exception of S. mexicana Wernham, formed a weakly (BS = 62) or highly (PP = 0.96) supported clade, respectively, in the MPA and BA. nrITS analyses (Fig. 2). — A total of 61 nrITS sequences were included in the analyses and 56 are newly published here. The MPA of the nrITS data resulted in 210 equally parsimonious trees (L = 542, CI = 0.601, RI = 0.758). In the strict consensus tree (Fig. 2), Hekistocarpa minutiflora was resolved, with high support (BS = 100, PP = 1.00), as sister to a very large, moderately (BS = 83) and highly (PP = 0.98) supported clade, respectively, in the MPA and BA analyses. That clade contained all investigated members of Tamridaea, Virectaria, Stipularia, Pseudosabicea, Sabicea, Schizostigma, and Ecpoma. Tamridaea capsulifera and the two Virectaria species formed a moderately (BS = 70) and strongly (PP = 1.00) supported clade, respectively, in the MPA and BA anslyses. This Tamridaea-Virectaria clade was in turn resolved as sister to the strongly supported (BS = 100, PP = 1.00) Sabiceeae s.str. clade. Within the latter clade, Stipularia elliptica was resolved as sister to a moderately supported (BS = 75) clade containing Stipularia efulenensis and all sequenced species of Ecpoma, Pseudosabicea, Sabicea, and Schizostigma (hereafter called Pseudosabicea-Sabicea-Stipularia-SchizostigmaEcpoma clade). Within this large clade all Pseudosabicea species were resolved in two highly supported clades: one formed by five Pseudosabicea species (BS = 99, PP = 1.00) and the other by three Pseudosabicea species, including the type species (Good, 1923; Hallé, 1970) Pseudosabicea nobilis (R. Good) N. Hallé (BS = 98, PP = 1.00). The former Pseudosabicea clade was resolved Table 2. Characteristics of Sabiceeae sequences and the alignments used in the phylogenetic analyses. Markers Matrix Range of non-aligned sequence lengths in Sabiceeae s.l. (bp) trnT-F trnT-F 1,574–1,688 Range of GC contents in Sabiceeae s.l. Number of sequences (%) characters 28.9–32.5 2,348 Informative characters Informative characters in Sabiceeae s.l. 495 (21.08%) 273 (11.63%) trnT-L spacer trnT-F 684–788 21.2–27.6 1,165 291 (12.39%) 165 (7.03%) trnL intron trnT-F 544–616 36.7–38.5 761 108 (4.60%) 62 (2.64%) trnL-F spacer trnT-F 268–324 32.1–36.2 422 96 (4.09%) 46 (1.96%) ITS ITS 566–599 53.7–65.5 670 202 (30.15%) 157 (23.43%) ITS1 ITS 186–221 52.7–68.7 268 109 (16.27%) 78 (11.64%) S5.8 ITS 165 54.5–53.3 165 6 (0.90%) 6 (0.90%) ITS2 ITS 207–216 54.2–70.9 237 87 (12.98%) 73 (10.90%) ITS nrITS + trnT-F 589–599 53.7–65.5 670 201 (30.00%) 148 (22.09%) ITS1 nrITS + trnT-F 216–221 52.7–68.7 268 109 (16.27%) 76 (11.34%) S5.8 nrITS + trnT-F 165 54.5–53.3 165 6 (0.90%) 6 (0.90%) ITS2 nrITS + trnT-F 207–216 54.2–70.9 237 86 (12.83%) 66 (9.85%) trnT-F nrITS + trnT-F 1,295–1,673 28.2–38.7 1,927 143 (7.63%) 64 (3.42%) trnT-L spacer nrITS + trnT-F 699–773 24.3–26.5 948 80 (4.15%) 47 (2.44%) trnL intron nrITS + trnT-F 523–616 36.5–44.4 642 29 (1.50%) 17 (0.88%) trnL-F spacer nrITS + trnT-F 185–331 35.4–37.3 337 34 (1.76%) 10 (0.56%) 5 Khan & al. • Tribal and generic circumscriptions of Sabiceeae 100 1.00 100 1.00 100 1.00 74 73 87 0.99 100 1.00 100 1.00 100 1.00 100 1.00 94 1.00 70 100 1.00 69 77 1.00 99 1.00 82 100 1.00 100 1.00 1.00 100 1.00 82 1.00 100 1.00 100 1.00 80 1.00 TAXON 57 (1) • February 2008: 1–17 83 1.00 Sabicea s.l. 62 0.96 Outgroup RUBIOIDEAE CINCHONOIDEAE IXOROIDEAE s.l. SABICEEAE s.l. SABICEEAE s.str. Luculia grandifolia Ophiorrhiza mungos Pentaloncha humilis 1 Pentaloncha humilis 2 Temnopteryx sericea 1 Temnopteryx sericea 2 Coussarea sp. Psychotria amboniana Gynochthodes coriacea Normandia neocaledonica Pentas parvifolia Danais xanthorrhoea Nauclea orientalis Cinchona pubescens Isertia pittieri MUS Mussaenda pinatubensis MUS Heinsia zanzibarica CON Warszewiczia coccinea Ixora coccinea IXO VAN Canthium coromandelicum Tarenna neurophylla PAV ALB Alberta magna Coffea eugenioides COF GAR Euclinia longiflora Hekistocarpa minutiflora VIR VIR Virectaria multiflora Virectaria procumbens VIR VIR Tamridaea capsulifera Ecpoma hierniana Sabicea angolensis Sabicea discolor Sabicea diversifolia Sabicea congensis Sabicea ingrata Sabicea johnstonii Sabicea xanthotricha Sabicea harleyae Sabicea mexicana Sabicea gilletii Sabicea orientalis Sabicea rosea Sabicea capitellata Pseudosabicea aurifodinae Pseudosabicea medusula Pseudosabicea segregata Pseudosabicea nobilis Pseudosabicea proselyta Pseudosabicea arborea Schizostigma hirsutum Stipularia elliptica Sabicea aspera Sabicea chocoana Sabicea glabrescens Sabicea humilis Sabicea panamensis Sabicea amazonensis Sabicea grisea Sabicea villosa Fig. 1. Strict consensus tree generated from 8,067 equally parsimonious trees based on the phylogenetic analysis of the trnT-F data. The numbers above the branches represent bootstrap support values ( > 50%) and those below the branches Bayesian posterior probabilities ( > 0.95). ALB, Alberteae; COF, Coffeeae; CON, Condamineeae; GAR, Gardenieae; IXO, Ixoreae; MUS, Mussaendeae; PAV, Pavetteae; VAN, Vanguerieae; VIR, Virectarieae. Brackets delimit the three subfamilies (sensu Bremer & al., 1999), Sabiceeae s.l., and Sabiceeae s.str. The genera shown in boldface belong to Sabiceeae sensu Andersson (1996). 6 Khan & al. • Tribal and generic circumscriptions of Sabiceeae TAXON 57 (1) • February 2008: 1–17 52 70 1.00 100 1.00 83 0.98 99 1.00 97 1.00 59 99 1.00 100 1.00 Sabicea s.l. 100 1.00 75 63 65 100 1.00 78 58 100 1.00 98 1.00 65 72 1.00 85 1.00 63 100 1.00 85 1.00 VIRECTARIEAE Neotropical Sabicea 90 1.00 VIRECTARIEAE SABICEEAE s.l. 88 1.00 Outgroup SABICEEAE s.str. 67 0.97 Canthium coromandelicum Ixora coccinea Warszewiczia coccinea Heinsia zanzibarica LG Hekistocarpa minutiflora SOC Tamridaea capsulifera UG Virectaria multiflora UG Virectaria procumbens Stipularia elliptica CO Pseudosabicea arborea 1 LV Pseudosabicea arborea 2 LV Pseudosabicea mildbraedii LG Pseudosabicea aurifodinae LG Pseudosabicea medusula LG Pseudosabicea batesii LG Sabicea xanthotricha LG Sabicea brevipes UG/S Sabicea dewevrei LG Stipularia efulenensis LG CO Sabicea gilletii LG Sabicea najatrix LG Sabicea caminata Sabicea dinklagei ZA Sabicea speciosa LG Schizostigma hirsutum SL Sabicea johnstonii LG Sabicea capitellata LG Sabicea ingrata STP UG Sabicea ferruginea LG Sabicea congensis Sabicea fulva LG Sabicea rosea UG Sabicea exellii STP Sabicea thomensis STP LG Ecpoma hierniana UG Sabicea harleyae UG Sabicea vogelii MAD Sabicea diversifolia Sabicea seua MAD Pseudosabicea segregata LG LG Pseudosabicea nobilis Pseudosabicea proselyta LG UG Sabicea discolor LG Sabicea angolensis SM Sabicea orientalis CO Sabicea venosa Sabicea amazonensis Sabicea aspera Sabicea sp. Sabicea chocoana Sabicea cinerea Sabicea humilis Sabicea mattogrossensis Sabicea mexicana Sabicea panamensis Sabicea grisea Sabicea glabrescens Sabicea venezuelensis Sabicea pyramidalis Sabicea villosa 1 Sabicea villosa 2 Fig. 2. Strict consensus tree generated from 210 equally parsimonious trees based on the phylogenetic analysis of the ITS data. The numbers above the branches represent bootstrap support values ( > 50%) and those below the branches Bayesian posterior probabilities ( > 0.95). Brackets delimit the outgroup taxa, Sabiceeae s.l., Sabiceeae s.str., and Neotropical Sabicea. Vertical bars delimit the genera of Virectarieae sensu Dessein & al. (2001a). CO, Congolian; LG, Lower-Guinean; LV, Lake Victorian; MAD, Madagascan; SL, Sri Lankan (Indian); SM, Somali-Masai; SOC, Socotran (Yemen); STP, São Tomean; UG, Upper-Guinean; UG/S, Upper-Guinean/Sudanian; ZA, Zambezian (African phytochoria; White, 1979, 1993). The phytogeographic data are mentioned only for the sampled African specimens. Sequenced species of Ecpoma, Pseudosabicea, Schizostigma, and Stipularia are shown in boldface. All shadowed taxa are from mainland Africa. 7 Khan & al. • Tribal and generic circumscriptions of Sabiceeae 100 1.00 99 1.00 100 1.00 100 1.00 77 1.00 100 1.00 Sabicea s.l. 100 1.00 100 1.00 59 99 1.00 65 0.99 59 1.00 53 0.97 0.97 80 1.00 100 1.00 77 100 1.00 83 1.00 80 75 1.00 87 1.00 53 68 99 1.00 82 1.00 68 0.98 VIRECTARIEAE VIRECTARIEAE Neotropical Sabicea 74 0.99 Outgroup SABICEEAE s.l. 69 0.99 Canthium coromandelicum Ixora coccinea Warszewiczia coccinea Heinsia zanzibarica LG Hekistocarpa minutiflora UG Virectaria multiflora UG Virectaria procumbens CO Stipularia elliptica Pseudosabicea arborea 1 LV Pseudosabicea arborea 2 LV Pseudosabicea mildbraedii LG Pseudosabicea aurifodinae LG Pseudosabicea medusula LG Pseudosabicea batesii LG Sabicea xanthotricha LG UG/S Sabicea brevipes Sabicea caminata LG LG Sabicea dewevrei ZA Sabicea dinklagei LG Stipularia efulenensis UG Sabicea ferruginea CO Sabicea gilletii LG Sabicea najatrix Sabicea congensis LG Sabicea fulva LG Sabicea diversifolia MAD Sabicea seua MAD Sabicea speciosa LG Schizostigma hirsutum SL Sabicea johnstonii LG LG Sabicea capitellata STP Sabicea ingrata Sabicea rosea UG STP Sabicea exellii STP Sabicea thomensis LG Ecpoma hierniana UG Sabicea harleyae Sabicea vogelii UG Pseudosabicea segregata LG Pseudosabicea nobilis LG Pseudosabicea proselyta LG Sabicea discolor UG Sabicea angolensis LG Sabicea orientalis SM CO Sabicea venosa Sabicea mexicana Sabicea aspera Sabicea sp. Sabicea chocoana Sabicea cinerea Sabicea mattogrossensis Sabicea panamensis Sabicea grisea Sabicea amazonensis Sabicea humilis Sabicea glabrescens Sabicea venezuelensis Sabicea pyramidalis Sabicea villosa 1 Sabicea villosa 2 SABICEEAE s.str. 0.98 100 1.00 74 TAXON 57 (1) • February 2008: 1–17 Fig. 3. Strict consensus tree generated from 104,428 equally parsimonious trees based on the phylogenetic analysis of the ITS-trnT-F data. The numbers above the branches represent bootstrap support values ( > 50%) and those below the branches Bayesian posterior probabilities ( > 0.95). Brackets delimit the outgroup taxa, Sabiceeae s.l., Sabiceeae s.str., and Neotropical Sabicea. Vertical bars indicate the position of the genera of Virectarieae sensu Dessein & al. (2001a). CO, Congolian; LG, Lower-Guinean; LV, Lake Victorian; MAD, Madagascan; SL, Sri Lankan (Indian); SM, Somali-Masai; STP, São Tomean; UG, Upper-Guinean; UG/S, Upper-Guinean/Sudanian; ZA, Zambezian (African phytochoria; White, 1979, 1993). The phytogeographic data are mentioned only for the sampled African specimens. Sequenced species of Ecpoma, Pseudosabicea, Schizostigma, Stipularia, and Tamridaea are shown in boldface. All shadowed taxa are from mainland Africa. 8 TAXON 57 (1) • February 2008: 1–17 as sister to a weakly supported (BS = 63) and Sabicea dominated clade formed by Stipularia efulenensis, all studied species of Ecpoma, Sabicea and Schizostigma and the other Pseudosabicea clade (containing P. segregata (Hiern) N. Hallé, P. nobilis, and P. proselyta N. Hallé). Furthermore, two investigated Malagasy (S. diversifolia Pers. and S. seua Wernham) and two São Tomean Sabicea species (S. exellii G. Taylor and S. thomensis Joffroy) formed strongly supported (BS = 100, PP = 1.00) groups, respectively. Another São Tomean Sabicea (S. ingrata K. Schum.) formed an unsupported group with two continental African Sabicea (S. johnstonii K. Schum. and S. capitellata). Similarly, all sequenced Neotropical Sabicea formed a weakly supported (BS = 63) clade in the MPA. These three clades were nested within the largely African Pseudosabicea-Sabicea-StipulariaSchizostigma-Ecpoma clade. Combined nrITS-trnT-F analyses (Fig. 3). — Each of the nrITS and trnT-F partitions used in the combined analyses contained a total of 60 sequences including 55 new sequences. All of the 55 new sequences of nrITS partition were used in the nrITS analyses and 32 new sequences of trnT-F partition were used in the trnTF analyses. The MPA of the combined nrITS + trnT-F matrix, composed of a total of 2,597 positions and 344 (13.24%) parsimony-informative characters (Table 2), generated 104,428 equally parsimonious trees (L = 714, Cl = 0.674, RI = 0.796). The overall tree topologies and support values of the resolved nodes in the strict consensus combined tree (Fig. 3) were largely similar to those of the strict consensus nrITS tree (Fig. 2). The support values in the combined tree were higher for some nodes (e.g., the Neotropical Sabicea clade and the Sabicea speciosa-Schizostigma hirsutum clade). The poorly supported (BS = 58) sister-group relationships between the Malagasy Sabicea clade (S. diversifolia, S. seua) and the Pseudosabicea clade formed by P. segregata, P. nobilis and P. proselyta (Fig. 2) collapsed in the combined tree (Fig. 3). Plus, the monophyletic group of one São Tomean Sabicea (S. ingrata) and two continental African Sabicea (S. johnstonii and S. capitellata), unsupported in the nrITS tree, was weakly (BS = 53) or strongly (PP = 0.97) supported in the combined tree. DISCUSSION Firstly, we compare the sequence characteristics between the nrITS and trnT-F sequences of Sabiceeae and those of the some other rubiaceous tribes (e.g., Naucleeae s.l., Mussaendeae, and Vanguerieae). Secondly, we discuss the new tribal circumscription of Sabiceeae, proposed in the light of our results. Accordingly, we propose the new circumscriptions for the genera of the tribe. Thirdly, we Khan & al. • Tribal and generic circumscriptions of Sabiceeae assess the biogeography of our newly delimited Sabicea s.l. and finally provide the updated description for Sabicea s.l. and make six new combinations. Sequence characteristics. — Both the ranges of lengths (Table 2) and the average lengths of nrITS1 and nrITS2 of Sabiceeae taxa fall within the records for other angiosperms (Baldwin & al., 1995; Noyes, 2006). The constant length of 5.8S subunit is consistent with the reports for other Rubiaceae (e.g., Alejandro & al., 2005) and close to those for other angiosperms (Baldwin & al., 1995). The records of GC contents in nrITS1 and nrITS2 of Sabiceeae taxa coincide with the reports for Rubiaceae (Razafimandimbison & Bremer, 2001; Alejandro & al., 2005) and other angiosperms (Tate & al., 2005). The total lengths of the nrITS region of Sabiceeae (566–599 bp) are nearly similar to those of Mussaendeae sensu Bremer & Thulin (1998) (570–596 bp; Alejandro & al., 2005), shorter than those reported for the tribe Vanguerieae (611–671 bp; Lantz & Bremer, 2004), and fall within the known range for other Ixoroideae (565–654 bp; Andreasen & al., 1999). The parsimony informative characters (PIC) for the nrITS region of Sabiceeae (157) are higher than those reported for Mussaendeae (103; Alejandro & al., 2005) and other Ixoroideae tribe Gardenieae A. Rich. ex DC. (e.g., 121 for the Alibertia group; Persson, 2000). On the other hand, they are lower than the PIC recorded for Vanguerieae (188; Lantz & Bremer, 2004) and the Cinchonoidae tribe Naucleeae s.l. (210; Razafimandimbison & Bremer, 2002). So, there is a great variation of the lengths of nrITS regions and eventually the number of parsimony informative characters between the different rubiaceous tribes. The range of the lengths of the trnT-F region of Sabiceeae (1574–1688 bp) coincides with the records for Mussaendeae sensu Bremer & Thulin (1998) (1662–1793 bp; Alejandro & al. 2005) and Vanguerieae (1559–1785 bp; Lantz & Bremer, 2004) but is shorter than that of Naucleeae s.l. (1707–1785 bp; Razafimandimbison & Bremer, 2002). The lengths of the trnT-F region of the studied Sabiceeae are 2.8 times longer than those of their nrITS region. In contrast, the trnT-F region of the sequenced Sabiceeae is less informative (11.63%) than their nrITS region (23.43%), concurring with Liede & Kunze (2002), Razafimandimbison & Bremer (2002), and Alejandro & al. (2005). In the trnT-F matrix, the trnT-L spacer (684–788 bp) is more variable than trnL-F spacer (268–324 bp), also consistent with Razafimandimbison & Bremer (2002), but our record of the trnL intron as more variable than the trnL-F spacer (Table 2) is inconsistent with their reports. The variations shown by the nrITS, trnT-L, trnL and trnL-F regions further indicate their usefulness for assessing the phylogenetic relationships in Rubiaceae and other families in the order Gentianales (e.g., Meve & Liede, 2002). 9 Khan & al. • Tribal and generic circumscriptions of Sabiceeae Tribal circumscriptions of Sabiceeae. — The polyphyly of Sabiceeae sensu Andersson (1996), which includes Amphidasya, currently classified by Bremer & Manen (2000) in the tribe Urophylleae Bremek. ex Verdc. (Rubioideae), and Acranthera, recently shown by Alejandro & al. (2005) to be associated with Rubioideae, is further corroborated by the trnT-F tree (Fig. 1), as both Pentaloncha and Temnopteryx are also resolved with high support (BS = 100, PP = 1.00) in Rubioideae. This is the first molecular phylogenetic study to include these African rubiaceous monotypic genera. We find no support either for the close relationships of Temnopteryx and Pentaloncha with Ecpoma, Pseudosabicea, Sabicea, and Stipularia postulated, respectively, by Puff & al. (1998) and Dessein & al. (2001a) or Hiern’s (1877) attempt to merge both Pentaloncha and Temnopteryx with Schizostigma (= Sabicea; Puff & al., 1998). The combined nrITS + trnT-F tree (Fig. 3) shows that Sabiceeae sensu Bremer & Thulin (1998) is not monophyletic, unless Ecpoma, Hekistocarpa, and Schizostigma are also included. Dessein & al. (2001a) tentatively included Pentaloncha in Sabiceeae s.str. based on morphological grounds. But our results strongly support the exclusion of Pentaloncha from Sabiceeae. Our results clearly favour a broad circumscription of Sabiceeae, which should include the following eight genera: Ecpoma, Hekistocarpa, Pseudosabicea, Sabicea, Schizostigma, Stipularia, Tamridaea, and Virectaria (Figs. 1–2), consistent with Robbrecht & Manen (2006). In all our parsimony and Bayesian analyses, we perceive no support for the monophyly of Virectarieae sensu Dessein & al. (2001a), as Hekistocarpa, Virectaria, and Tamridaea (Figs. 1–2) or Hekistocarpa and Virectaria (Fig. 3) never form a clade, and therefore, its tribal status is untenable. For the same reason our results do not support the new subtribal classification of Sabiceeae (Sabiceinae and Virectariinae) by Robbrecht & Manen (2006). The discrepancies between our results and the rbcL or rps16 trees of Dessein & al. (2001a) are probably due to taxon sampling. In the rbcL tree of Dessein & al. (2001a), Sabiceeae, represented by one Sabicea species, is weakly resolved as sister to the strongly supported (BS = 87) Virectarieae sensu Dessein & al. (2001a). In their rps16 tree, the support for Virectarieae, represented by Hekistokarpa and Virectaria, is weak (BS = 59), while that of Sabiceeae s.str., represented by four Sabicea and two Pseudosabicea species, is high (BS = 87). In other words, the support for the monophyly of Virectarieae sensu Dessein & al. (2001a) seems to decrease when more species from Sabiceeae are included in the rps16 analysis. This is further confirmed by our ITS and combined nrITS + trnT-F analyses (Figs. 2–3), which contain a much larger sampling of Sabiceeae s.str. (51 species of Ecpoma, Pseudosabicea, Sabicea, Schizostigma, and Stipularia), in which Virectarieae sensu Dessein & al. (2001a) totally collapse. The Sabiceeae s.l. clade of our 10 TAXON 57 (1) • February 2008: 1–17 trnT-F tree (Fig. 1) is largely congruent with that of the nrITS (Fig. 2) and the combined nrITS + trnT-F trees (Fig. 3), with the exception of the position of Tamridaea. We were unable to include the Neotropical genus Pittierothamnus (Steyermark, 1962) due to lack of material. Therefore, its phylogenetic position in Sabiceeae postulated by Andersson (1996) has yet to be tested with molecular-based phylogenies. We have not been able to find any potential morphological synapomorphy to diagnose our newly delimited Sabiceeae s.l. Therefore, the monophyly of the tribe is entirely based on molecular data. Generic circumscriptions in Sabiceeae s.l. — Our newly circumscribed Sabiceeae contains the following four genera: Hekistocarpa, Sabicea s.l. (including Ecpoma, Pseudosabicea, Schizostigma, and Stipularia), Tamridaea, and Virectaria, and a total of ca. 180 species. The monotypic genus Hekistocarpa is restricted to Cameroon and Nigeria (Dessein & al 2001a: Fig. 37). This genus can be characterized by the following characters: herbaceous growth habit, lateral scorpioid cymes, laterally compressed fruits, exotesta cells with strongly thickened walls, tuberculate surface and one perforation, and tricolpate pollens (Dessein & al. 2001a). Hekistocarpa was classified for a long time in the tribe Hedyotideae Cham. & Schltdl. ex DC. (Rubioideae) because of its herbaceous habit, scorpioid inflorescences and many-seeded fruits (Hooker, 1873b). The study of Dessein & al. (2001a) is the first to place Hekistocarpa in Virectarieae. All the trnT-F, nrITS and combined nrITS + trnT-F trees (Figs. 1–3) strongly (BS = 100, PP = 1.00) favour its placement in Sabiceeae s.l. Furthermore, Hekistocarpa is resolved as sister to the rest of Sabiceeae s.l. (e.g., Fig. 2) and therefore, its current generic status should be maintained. All nrITS and combined analyses (Figs. 2–3) indicate that Sabicea sensu Wernham (1914) including S. hierniana Wernham (= Ecpoma hierniana (Wernham) N. Hallé & F. Hallé), S. segregata Wernham (= Pseudosabicea segregata), and S. nobilis R. Good (= Pseudosabicea nobilis) is only monophyletic if Pseudosabicea proselyta, Schizostigma, and Stipularia efulenensis are also included. The African genus Stipularia appears polyphyletic, as the two sequenced species, S. elliptica and S. efulenensis, are resolved in two separate clades (Figs. 2–3). The type species S. africana is not included in the present study, so the generic status of Stipularia could still be maintained if it turns out that S. africana forms a clade with S. elliptica. On the other hand, our results indicate that the generic concept of Stipularia based mainly on the presence of the large campanulate involucral bracts subtending the entire inflorescence is untenable, as the two sequenced Stipularia species bearing the same type of the involucral bracts (Hepper, 1958) do not form a clade. Plus, Hepper (1958: 289–291) convincingly explained that the involucral bracts of some African Sabicea species show a great range of the degree of fusion Khan & al. • Tribal and generic circumscriptions of Sabiceeae TAXON 57 (1) • February 2008: 1–17 (from inconspicuous to distinct and totally free to partly or completely fused bracts). Also, Hallé (1966) showed that some African Sabicea species (e.g., Sabicea duparquetiana H. Baillon ex Wernham and S. najatrix N. Hallé) have large and partly fused campanulate involucral bracts. Based on the above evidence presented we concur with Hepper’s (1958) decision to merge Stipularia with Sabicea. Our analyses further reveal the polyphyly of Sabicea sensu Hallé (1963), as Ecpoma, represented by E. hierniana, Pseudosabicea, represented by P. segregata, P. nobilis, and P. proselyta, and Schizostigma are all resolved in the largely Sabicea clade with weak and high support (e.g., BS = 59, PP = 1.00; Fig. 3), respectively, in the MPA and BA. Similarly, Pseudosabicea sensu Hallé (1963) is also shown to be para- or polyphyletic, as the sequenced Pseudosabicea species group in two separate clades (Figs. 2–3). Accordingly, we merge Pseudosabicea with Sabicea. The range of variation in the characters of Sabicea includes the diagnostic characters of Pseudosabicea sensu Hallé (1963). One could recognize the strongly supported clade of five Pseudosabicea species at generic level; however, we find no distinctive morphological character for diagnosing this clade, once P. nobilis, P. proselyta, and P. segregata were included in Sabicea. The African genus Ecpoma (Schumann, 1896) is comprised of six species and characterized by its shrubby habit, isophylly, colourless pulp of small fruits, bilocular ovaries, non-accrescent septa, rounded or twisted to peltate placentae (Hallé, 1963). Ecpoma was traditionally classified in Mussaendeae (Hallé, 1961; Hallé, 1963, 1966) or in Sabiceeae (Andersson, 1996; Robbrecht & Manen, 2006). In Andersson’s (1996) study, Ecpoma did not form a monophyletic group with Pseudosabicea, Sabicea, and Schizostigma. In our nrITS and nrITS + trnT-F trees (Figs. 2–3), Ecpoma, represented by E. hierniana, however, is consistently and deeply nested within the PseudosabiceaSabicea-Stipularia-Schizostigma-Ecpoma clade, inconsistent with Hallé (1961), Hallé (1963) and Andersson (1996). Accordingly, we merge Ecpoma with Sabicea even if the type species is not included in our analyses because its character states clearly fall within the range of variation in Sabicea s.l. Adopting the broadened circumscription of Sabicea including Ecpoma, Pseudosabicea, Schizostigma, and Stipularia requires only a maximum of six new combinations, as five of the six Ecpoma species (Hallé, 1963) and 8 of the 13 Pseudosabicea species (Hallé, 1963, 1966) were originally described as Sabicea (Wernham, 1914; Good, 1923). Plus, all five Stipularia species and Table 3. Morphological distinctive characters of Hekistocarpa, Sabicea s.l., Tamridaea, and Virectaria. Characters Hekistocarpa Sabicea s.l. Habit Herbs Lianas, vines, straggling to scram- Shrubs (ca. 1 m tall) bling herbs, scandent or erect shrubs (up to 4 m tall), rarely small trees Herbs Inflorescence position and types Axillary, scorpioid cymes Axillary, fasciculate or densely capitulate to paniculate or thyrsoid, simple to compound dichasial cymes or solitary flowers Terminal, usually dichasial corymbose cymes Terminal, dichasial thyrsoid to monochasial or simple cymes Flower types Homostylous Hetero- and homostylous Heterostylous Homostylous Corolla aestivation Reduplicate valvate True valvate Reduplicate valvate True valvate Corolla lobes Ovate to deltoid with Ovate with (sub-) acute apices (sub-) acute apices Obcordate corolla Lanceolate to deltoid lobes with emargin- with (sub-) acute ate-mucronate apices apices Anther fixation and position Dorsimedifixed, included Dorsimedifixed, included (shortstyled flowers) and slightly exserted (long-styled flowers) Dorsifixed, included Dorsimedifixed, (short-styled flowers) exserted and slightly exserted (long-styled flowers) Stigma branches 2, filiform 2–5(6), filiform to oblong or very narrowly elliptic or oblanceolate to widely spathulate or dilated 2, filiform-oblong No. of locules per ovary Tamridaea Virectaria Initially 2, eventually truncated, spherical ca. 10 2–5(7) 2 2 Fruit types Dry, indehiscent or tardily dehiscent Indehiscent berries Dry, dehiscent capsules Dry, dehiscent capsules with one caduceus valve Pollen type 3-colporate 3–4-colporate 4-colporate 3-colporate 11 Khan & al. • Tribal and generic circumscriptions of Sabiceeae Schizostigma have already been merged, respectively, by Hepper (1958) and Puff & al. (1998) in Sabicea. Sabicea s.l. is very distinct from the other three genera (Hekistocarpa, Tamridaea, Virectaria) of Sabiceeae s.l. in some aspects (see Table 3). All our nrITS and combined nrITS + trnT-F analyses contradict the monophyly of Wernham’s (1914) two subgenera of Sabicea based on habit and leaf and stipule sizes. The two sequenced species of Sabicea subgen. Stipulariopsis (Sabicea xanthotricha Wernham and S. hierniana Wernham [= Ecpoma hierniana]) do not form a clade. The sequenced species of Sabicea subgen. Eusabicea (e.g., S. batesii Wernham [= P. batesii], S. mildbraedii [= P. mildbraedii], S. segregata [= P. segregata], S. seua, S. speciosa K. Schum., S. vogelii Benth., S. angolensis Wernham, S. discolor Stapf, S. venosa Benth., and S. hirsuta H.B. & K. [= S. villosa Willd. ex Roem. & Schult.], etc.) do not form a clade unless Pseudosabicea (P. arborea (K. Schum.) N. Hallé and P. proselyta), Sabicea subgen. Stipulariopsis, Stipularia efulenensis, and Schizostigma are also included. New molecular phylogenetic investigations using multiple markers and a much broader sampling of Ecpoma, Pseudosabicea, Stipularia, and Sabicea will be performed in attempt to establish, if possible, new infrageneric classifications for our newly delimited Sabicea with ca. 170 species and also address some evolutionary questions. The monotypic genus Tamridaea, endemic to Socotra (Yemen), is characterized by its shrubby habit, reduplicate-valvate aestivation, terminal cymes, flat, ± obcordate corolla lobes with emarginate-mucronate apices, bilobed stigma, bilocular ovaries (Bremer & Thulin, 1998), exotesta cells with verrucose thickenings, and 4-colporate pollens (Dessein & al., 2001a). Bremer & Thulin (1998) originally described Tamridaea to accommodate Pseudomussaenda capsulifera (Balf. f.) Wernham, previously classified in Isertieae sensu Robbrecht (1988), and placed it in their Sabiceeae s.l. Dessein & al. (2001b) accept the generic status of Tamridaea and its placement in Sabiceeae sensu Bremer & Thulin (1998), though Dessein & al. (2001a) placed the genus in their emended Virectarieae. Tamridaea has conflicting positions in our results. In our trnT-F tree (Fig. 1), it is moderately (BS = 80) and highly (PP = 1.00) resolved, respectively, as sister to Sabiceeae s.str. in the MPA and BA analyses. In the nrITS tree, the genus and Virectaria form a moderately to highly (BS = 70, PP = 1.00; Fig. 2) supported clade, consistent with Bremer & Thulin (1998) and Dessein & al. (2001a). When included in a combined nrITS + trnT-F analysis Tamridaea is weakly resolved (BS = 58) as sister to Virectaria. The tropical African genus Virectaria comprises eight species, of which three species (V. major K. Schum., V. multiflora, V. procumbens) are Guineo-Congolian wide (Dessein & al., 2001b), while four species (V. herbacoursi N. Hallé, V. belingana N. Hallé, V. salicoides (C.H. Wright) 12 TAXON 57 (1) • February 2008: 1–17 Bremek., V. angustifolia (Hiern) Bremek.) are endemic to one of the domains of the Guineo-Congolian region (White, 1979), Lower Guinea and V. tenella J.B. Hall to Upper Guinea (Dessein & al. 2001a: Figs. 69–70). The genus can be characterized by its herbaceous to semi-woody habits, terminal inflorescences, truncated stigmas, flat trichomes of the corolla orifice or inside the corolla tubes, elongated floral disc, one persistent and one deciduous valve during fruit dehiscence, and exotesta cells of seeds with many small perforations (Dessein & al., 2001a). Our results support the placement of Virectaria in Ixoroideae s.l., also consistent with Bremer & Thulin (1998) and Dessein & al. (2001a) but inconsistent with Bremekamp (1952, 1966) who classified the genus in the tribe Ophiorrhizeae of his Cinchonoideae, and Verdcourt (1975) who placed it in Cinchonoideae as a monogeneric tribe Virectarieae. In both nrITS and combined nrITS + trnT-F trees (Figs. 2–3), Virectaria is strongly (BS = 100, PP = 1.00) resolved as a monophyletic group, which is moderately supported as sister to Tamridaea (Fig. 2), consistent with Dessein & al. (2001a) and Robbrecht & Manen (2006). However, our results are inconsistent with the placement of the genus pair and Hekistokarpa in a separate tribe Virectarieae (Dessein & al. 2001a) or subtribe Virectariinae (Robbrecht & Manen, 2006). Tamridaea and Virectaria are morphologically distinct (see Table 3) and therefore, their generic status can be maintained. Biogeography of Sabicea s.l. — We are unable to perform a proper biogeographic analysis, because the clade of Sabiceeae s.str. is largely unresolved in all trees (Fig. 1–3). However, some biogeographical facts can be discussed for Sabicea s.l. The combined tree (Fig. 3) shows that neither the Upper-Guinean, nor the Lower-Guinean, nor the Congolian (White, 1976; Robbrecht, 1996) Sabicea species form a monophyletic group, and in contrast, the species of different phytogeographical regions (e.g., Lake Victoria and Lower-Guinea or Somali-Masai and Congolia; White, 1976, 1993) form highly supported clades. These results indicate that Sabicea species of these phytogeographical domains and regions are not closely related and there seem to be several dispersal events of Sabicea species between them. The volcanic Island of São Tomé (Deruelle & al., 1991; Munhá & al., 2002) has three endemic Sabicea species (S. exellii, S. ingrata, S. thomensis; Joffroy, 2001), which are consistently nested in the almost continental African Sabicea clade (Figs. 2–3). One São Tomean Sabicea species (S. ingrata) groups together with the Lower-Guinean S. capitellata and S. johnstonii, and the other two São Tomean species (S. ingrata, S. thomensis) group with the Upper-Guinean S. rosea Hoyle (Fig. 3). These results indicate that the São Tomean species must have had two African ancestors, which appear to have reached the island via two independent dispersal events. Similarly, the two TAXON 57 (1) • February 2008: 1–17 sequenced Malagasy species of Sabicea, S. diversifolia and S. seua (Razafimandimbison & Miller, 1999), form a highly supported (BS = 100, PP = 1.00) clade, which is nested in the large Sabicea clade. No record of Sabicea s.l. is known from the neighbouring Islands of Madagascar. Madagascar is about 400 km off the southwestern coast of Mozambique, whereas São Tomé & Príncipe are only within 225 to 250 km off of the northwestern coast of Gabon. All sequenced Sabicea species of the Neotropics form a moderately supported (BS = 74) clade in the MPA and a highly supported (PP = 0.99) clade in the BA (Fig. 3), indicating a single origin of all Neotropical Sabicea. The Neotropical Sabicea additionally appear to have originated from an African common ancestor. Furthermore, our data (Fig. 3) indicate that the African common ancestors of the Malagasy, São Tomean, and Neotropical Sabicea, respectively, most likely reached Madagascar, São Tomé, and the Neotropics through four independent dispersal events either via wind and/or ocean currents or dispersal of seeds across the Mozambique Channel, the Gulf of Guinea, and the South Atlantic Ocean by birds. Sabicea s.l. produce fleshy and (sub-) globose or obovoid berries bearing many small seeds, which would presumably provide an important source of food for tropical frugivorous birds. This seems to favour a zoochorous mode of dispersal (but see Renner, 2004). The fact that the Neotropics and São Tomé do not share in common any Sabicea species seems to exclude stepping-stone long-distance dispersal (i.e., dispersal from the mainland Africa to the Neotropics via São Tomé) as the mode of dispersal responsible for the present transAtlantic distribution of Sabicea s.l. Our results (Fig. 3) further indicate that four African Sabicea species (S. angolensis, S. discolor, S. orientalis Wernham, S. venosa) are more closely related to each other than they are to the remaining Sabicea s.l. Plus, they appear to be most closely related to the Neotropical Sabicea, also consistent with morphological grounds. It is, however, important to note that these four African Sabicea species are presently either restricted to a domain of Guineo-Congolian region (e.g., S. angolensis and S. discolor occur in Lower- and Upper Guinea, repectively) or dispersed to two to three phytogeographical regions (e.g., S. orientalis occurs in GuineoCongolian, Zambezian and Somalia-Masai region, and S. venosa in Guineo-Congolian and Lake Victoria regions; White, 1993). Finally, Sabicea s.l. seems to have started to diversify in mainland Africa, where at least 106 species are presently known. A second major radiation of Sabicea appears to have occurred after the group began to colonize the Neotropics. The occurrence of the single Asian species Sabicea ceylanica (restricted to Sri Lanka) indicates that the genus seems to have failed to disperse to the rest of Asia. Khan & al. • Tribal and generic circumscriptions of Sabiceeae CONCLUSIONS The present phylogenetic analyses favour a broad circumscription of Sabiceeae, which includes the following four genera: Hekistocarpa, Sabicea s.l. (including Ecpoma, Pseudosabicea, Schizostigma, and Stipularia), Tamridaea, and Virectaria. Pentaloncha and Temnopteryx belong to subfamily Rubioideae. Sabiceeae sensu Bremer & Thulin (1998) is not monophyletic, unless Ecpoma, Hekistocarpa, and Schizostigma are also included. Virectarieae sensu Dessein & al. (2001a) appears to be para- or polyphyletic. Dessein & al.’s (2001a) Sabiceeae and Robbrecht & Manen’s (2006) subtribal classification of Sabiceeae are not supported by our results. Sabicea sensu Wernham (1914) is monophyletic only if Pseudosabicea proselyta, Stipularia efulenensis and Schizostigma are included. Finally, our analyses support the monophyly of Malagasy and Neotropical Sabicea, but not of Sabicea and Pseudosabicea both sensu Hallé (1963, 1966) and Stipularia. Our results indicate several dispersal events of Sabicea species between few African phytogeographical domains and regions. The São Tomean, Malagasy, Asian and Neotropical species of Sabicea all appear to have had African origins and perhaps dispersed via four independent dispersal events. TAXONOMIC IMPLICATIONS Sabicea Aubl. Hist. Pl. Guiane Françoise 1: 192, t. 75. JunDec 1775 – Lectotype: S. cinerea Aubl. designated by P.C. Standley, N. Amer. Fl. 32: 148. 10 May 1921. PHAN.-RUBIACEAE (75/104). = Cephaëlis Sw., Prodr. (Swartz) 3, 45 (‘Cephaelis’). 20 Jun-29 Jul 1788 (nom. cons.) – Type: C. muscosa (Jacq.) Sw. ≡ Morinda muscosa Jacq. (typ. cons.). = Paiva Vell., Fl. Flum.: 104. 7 Sep-28 Nov 1829 (‘1825’) – Type: P. verticillata Vell. = Stipularia P. Beauv., Fl. Owar. 2: 26. 1807 – Type: S. africana P. Beauv. – Holotype: South Nigeria, Palisot de Beauvois s.n. (G!), isotype (P, not seen) ≡ Sabicea africana (P. Beauv.) Hepper. = Ecpoma K. Schum., Bot. Jahrb. 23: 430. 1896, syn. nov. – Type: E. apocynaceum K. Schum. – Holotype: Cameroon, near Lolodorf, Staudt 204 (B, presumably destroyed; K, photo!). = Pseudosabicea N. Hallé, Adansonia ser 2, 3: 170. 1963, syn. nov. – Type: P. nobilis (R. Good) N. Hallé ≡ Sabicea nobilis R. Good – Syntypes: Angola, Belize, Maiombe, Gossweiler 7550, 7043 (BM, P). = Schizostigma Arn. ex Meisn., Pl. Vasc. Gen. 1: 164; 2: 115. 1838 – Type: S. hirsutum Arn. ex Meisn. (holotype or syntypes not designated) ≡ Sabicea ceylanica Puff. 13 Khan & al. • Tribal and generic circumscriptions of Sabiceeae Schwenkfeldia Wild. (Sp. Pl. 4 [post Reichardianum quinta]: 982. 1797) was described based on Schwenkfelda Schreb. (Gen. Pl. 1: 123. 1789), but the latter was described based on Sabicea Aubl. Therefore, Schwenk felda and Schwenkfeldia are illegitimate names. Lianas or woody vines, climbing or scrambling to erect herbs, scandent to erect shrubs, rarely small trees, stems rounded to shallowly quadrangular. Stipules interpetiolar, free, persistent, minute to vigorous, usually entire, sometimes fimbriate to deeply laciniate, usually with few to many colleters inside the base. Nodes isophyllous or anisophyllous. Leaves membranaceous to subcoriaceous. Inflorescences axillary, sessile to pedunculate, solitary to compactly capitate to lax thyrsoid and few to many flowered cymes, subtended by inconspicuous to distinct and free to completely united and variously lobed bracts with usually 2 to many colleters inside the base, with or without forming spreaded to deeply campanulate involucre, rarely followed by prophylls. Calyces shallowly to deeply campanulate to funnel-shaped, 3–5-lobed, lobes filiform to elliptic or obovate, antrorse to abruptly reflex, usually with 1–2 colleters in or below each sinus. Corollas hypocrateriform or broadly infundibuliform, usually white, occasionally pinkish, usually 5-lobed, lobes valvate, narrowly to widely ovate, margins entire, glabrous or papillate inside, (sub-) acute at apex. Stamens included to slightly exserted just beyond the corolla tubes, anthers linear to narrowly oblong, basally and apically acute to rounded, dehiscent by longitudinal slits, dorsifixed near the middle by the very short free part of filiform filaments, attached to the upper part of corolla tubes. Pollens colporate to pororate, apertures 3 or 4, exine surface minutely reticulate, released as monads. Styles filiform, usually glabrous and included to slightly exserted just beyond the corolla tubes, stigmatic lobes 2–5, filiform to oblong or very narrowly elliptic or oblanceolate to widely spathulate or dilated. Ovaries usually (sub-) globose, 2–7-locular with axile placentation and numerous ovules per locule. Fruits (sub-) globose, indehiscent berries. Seeds minute, usually numerous, variously angular, exotesta cells narrow and elongated, with few to many rounded pits, radial wall with verrucose thickenings. Indument of stem, branches, leaves, stipules, inflorescences, bracts, hypanthia and corolla tubes isolatedly to densely puberulous to hirsute or pilose, strigose or sericeous to villous, velutinous or arachnose and indument of corolla orifice or inside the corolla tubes usually moniliform. The karyologically reported taxa are tetraploid with basic chromosome numbers x = 9 or 11 (Kiehn, 1995). Number of species: ca. 170 species (106 confined to the African mainland, 54 restricted to the Neotropics, 6 endemic to Madagascar, 3 to São Tomé and Príncipe, and 1 to Sri lanka). 14 TAXON 57 (1) • February 2008: 1–17 Sabicea s.l. can easily be distinguished from the other three genera of Sabiceeae s.l. by the combination of the following characters: axillary inflorescences usually composed of few to many flowered fascicles or densely capitulate to laxly paniculate cymes or solitary flowers, hypocrateriform or broadly infundibuliform corollas with ovate, (sub-) acute lobes, anthers and 2–6-lobed stigmata usually included in the corolla tubes, moniliform trichomes of corolla orifice or inside the corolla tubes, and narrow to elongated exotesta cells of seeds, with few to many rounded pits and verrucose thickenings on the radial wall (see also Table 3). New combinations Sabicea apocynaceum (K. Schum.) Razafim., B. Bremer, Liede & Khan, comb. nov. ≡ Ecpoma apocynaceum K. Schum. in Bot. Jahrb. 23: 430. 1897 – Type: Cameroon, Lolodorf, Feb (fl.), Staudt 208 (holotype, B, presumably destroyed; K, photo). Sabicea aurifodinae (N. Hallé) Razafim., B. Bremer, Liede & Khan, comb. nov. ≡ Pseudosabicea aurifodinae N. Hallé in Fl. Gabon 12: 201. 1966 – Type: Gabon, Moubigou-2, au bout de la route de Massima vers Moumba, région d’Etéké, N. Hallé & G. Cours 6137 (holotype, P). Sabicea becquetii (N. Hallé) Razafim., B. Bremer, Liede & Khan, comb. nov. ≡ Pseudosabicea becquetii N. Hallé in Bull. Jard. Bot. État Bruxelles 34: 400. 1964 – Type: Burundi, Bururi chefferi Arawe-territoire, alt. 1,600 m, Becquet 115 (holotype, P; isotype, K). Sabicea proselyta (N. Hallé) Razafim., B. Bremer, Liede & Khan, comb. nov. ≡ Pseudosabicea proselyta N. Hallé in Adansonia ser. 2, 3: 172. 1963 – Type: Gabon, la Nkoulounga, 11 Jul 1959, N. Hallé 748 (holotype, P). Sabicea sanguinosa (N. Hallé) Razafim., B. Bremer, Liede & Khan, comb. nov. ≡ Pseudosabicea sanguinosa N. Hallé in Adansonia ser. 2, 11: 313. 1971 – Type: Gabon, environs de la Station forestière du Petit Bam-Bam, 50 km SW de la base rivière Ramboué, au sud de l’Estuaire, pays de savanes, 21 Aug 1966 (fl.), N. Hallé & A. Le Thomas 573 (holotype, P). Sabicea sthenula (N. Hallé) Razafim., B. Bremer, Liede & Khan, comb. nov. ≡ Pseudosabicea sthenula N. Hallé in Fl. Gabon 12: 208. 1966 – Type: Gabon, Makokou, 27 Feb 1961, N. Hallé 1339 (holotype, P). TAXON 57 (1) • February 2008: 1–17 ACKNOWLEDGEMENTS The authors thank the curators of B, BM, BR, BRLU, F, G, K, L, M, MO, NY, P, S, TAN, TEF, U, UPS, US, VEN, W, and WAG for providing the loans and/or for permission to sample herbarium collections; Jan Wieringa, for providing silica-gel dried leaf samples; ANGAP (Association Nationale pour la Gestion des Aires Protégées) and MEF (Ministère des Eaux et Forêts, Madagascar) for issuing collecting permits for SGR; Missouri Botanical Garden Program (Madagascar) for kindly arranging collecting permits for SGR; Désiré Ravelonarivo (ANGAP Andapa) for field assistance in the Marojejy National Park, Madagascar; Elmar Robbrecht, Charlotte Taylor, Petra De Block and Piero Delprete for their helpful suggestions; Angelika Täuber and Anbar Khodabandeh for help with sequencing; and Ulrich Meve, Andreas Jürgens, Stefan Dötterl, Heidi Döring, and Grecebio D. Alejandro for their technical support. Funds for this study were provided by the Swedish Research Council and the Royal Swedish Academy of Sciences to BB and Jahangirnagar University and University of Bayreuth to SAK. LITERATURE CITED Alejandro, G.D., Razafimandimbison, S.G. & LiedeSchumann, S. 2005. Polyphyly of Mussaenda inferred from ITS and trnT-F data and its implication for generic limits in Mussaendeae (Rubiaceae). Amer. J. Bot. 92: 544–557. Andersson, L. 1996. Circumscription of the tribe Isertieae (Rubiaceae). Opera Bot. Belg. 7: 139–164. Andreasen, K., Baldwin, B.G. & Bremer, B. 1999. Phylogenetic utility of the nuclear rDNA ITS region in the subfamily Ixoroideae (Rubiaceae): comparisons with cpDNA rbcL sequence data. Pl. Syst. Evol. 217: 119–135. Andreasen, K. & Bremer, B. 2000. Combined phylogenetic analysis in the Rubiaceae-Ixoroideae: morphology, nuclear and chloroplast DNA data. Amer. J. Bot. 87: 1731–1748. Arnott, G.A.W. 1839. Descriptions of some new or rare Indian plants. Ann. Nat. Hist. 3: 20–23. Aublet, J.B.C.F. 1775. Sabicea, Pentandria, Monogynia. Pp. 192–196 in: Histoire des plantes de la Guiane Françoise, vol. 1. P.-F. Didot jeune, London & Paris. Baldwin, B.G., Anderson, M.J.S., 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. Bayer, R.J., Greber, D.G. & Bagnall, N.H. 2002. Phylogeny of Australian Gnaphalieae (Asteraceae) based on chloroplast and nuclear sequences, the trnL intron, trnL/trnF intergenic spacer, matK, and ETS. Syst. Bot. 27: 801–814. Bremekamp, C.E.B. 1934. Notes on Rubiaceae of Surinam. Recueil Trav. Bot. Néerl. 31: 248–308. Bremekamp, C.E.B. 1952. The African species of Oldenlandia L sensu Hiern et K. Schumann. Verh. Kon. Ned. Akad. Wetensch., Afd. Natuurk., Sect. 2., 48: 1–297. Bremekamp, C.E.B. 1966. Remarks on the position, the de- Khan & al. • Tribal and generic circumscriptions of Sabiceeae limitation and the subdivision of the Rubiaceae. Acta Bot. Neerl. 15: 1–33. Bremer, B. 1996. Combined and separate analyses of morphological and molecular data in the plant family Rubiaceae. Cladistics 12: 21–40. Bremer, B., Jansen, R.K., Oxelman, B., Backlund, M., Lantz, H. & Kim, K.J. 1999. More characters or more taxa for a robust phylogeny—case study from the coffee family (Rubiaceae). Syst. Biol. 48: 413–435. Bremer, B. & Manen, J.F. 2000. Phylogeny and classification of the subfamily Rubioideae (Rubiaceae). Pl. Syst. Evol. 225: 43–72. Bremer, B. & Thulin, M. 1998. Collapse of Isertieae, re-establishment of Mussaendeae, and a new genus of Sabiceeae (Rubiaceae): phylogenetic relationships based on rbcL data. Pl. Syst. Evol. 211: 71–92. Darwin, S.P. 1976. The subfamilial, tribal, and subtribal nomenclature of the Rubiaceae. Taxon 25: 595–610. Dessein, S., Andersson, L., Robbrecht, E. & Smets, E. 2001a. Hekistocarpa (Rubiaceae): a member of an emended tribe Virectarieae. Pl. Syst. Evol. 229: 59–78. Dessein, S., Jansen, S., Huysmans, S., Robbrecht, E. & Smets, E. 2001b. A morphological and anatomical survey of Virectaria (African Rubiaceae), with a discussion of its taxonomic position. Bot. J. Linn. Soc. 137: 1–29. Deruelle, B., Moreau, C., Nkoumbou, C., Kambou, R., Lissom, J., Njongfang, E., Ghogomu, R.T. & Nono, A. 1991. The Cameroon Line: a review. Pp. 274–327 in: Kampunzu, A.B. & Lubala, R.T. (eds.), Magmatism in Extensional Structural Settings. Springer Verlag, Berlin. Good, R. 1923. New Tropical African Rubiaceae. J. Bot., 61: 86. Grisebach, A.H.R. 1861. XCIV Rubiaceae. Pp. 316–351 in: Flora of the British West Indian Islands. L. Reeve & Co., London. Hallé, F. 1961. Contribution à l’étude biologique et taxonomique des Mussaendeae (Rubiaceae) d’Afrique tropicale. Adansonia 1: 266–298. Hallé, N. 1963. Délimitation des genres Sabicea Aubl. et Ecpoma K. Schum. en regard d’un genre nouveau: Pseudosabicea (Mussaendeae-Rubiaceae). Adansonia ser. 2, 3: 168–177. Hallé, N. 1966. Famille des Rubiacées (1re partie). Vol. 12 of: Aubréville, A. (ed.), Flore du Gabon. Muséum National d’Histoire Naturelle, Laboratoire de Phanérogamie, Paris. Hallé, N. 1970. Famille des Rubiacées (2e partie). Vol. 17 of: Aubréville, A. & Leroy, J.F. (eds.), Flore du Gabon. Muséum National d’Histoire Naturelle, Laboratoire de Phanérogamie, Paris. Hepper, F.N. 1958. Sabicea Aubl. and Stipularia Beauv. (Rubiaceae-Mussaendeae) in Tropical Africa. Kew Bull. 289–294. Hiern, W.P. 1877. Ordo LXX. Rubiaceae. Pp. 33–82 in: Oliver, D., Dyer, W.T.T., Prain, D. & Hill, A.W. (eds.), Flora of Tropical Africa, vol. 3. L. Reeve & Co., London. Hooker, J.D. 1873a. Ordo LXXXIV, Rubiaceae. Pp. 7–151 in: Bentham, G. & Hooker, J.D. (eds.), Genera Plantarum ad exemplaria imprimis in herbariis Kewensibus servanta defirmata, vol. 2. L. Reeve & Co., London. Hooker, J.D. 1873b. Hekistocarpa minutiflora. Icones Plantarum 12: 46. William Pamplin, London. 15 Khan & al. • Tribal and generic circumscriptions of Sabiceeae Huelsenbeck, J.P. & Ronquist, F. 2001. MrBayes: Bayesian inference of phylogeny. Bioinformatics 17: 754–755. Joffroy, G. 2001. Le genre Sabicea (Rubiaceae) à São Tomé (São Tomé et Príncipe). Syst. Geogr. Pl. 71: 383–390. Kiehn, M. 1995. Chromosome survey of the Rubiaceae. Ann. Missouri Bot. Gard. 82: 398–408. Kirkbride, M.C.G. 1979. Review of the Neotropical Isertieae (Rubiaceae). Brittonia 31: 313–332. Kirkbride, M.C.G. 1982. A preliminary phylogeny for Neotropical Rubiaceae. Pl. Syst. Evol. 141: 115–121. Lantz, H. & Bremer, B. 2004. Phylogeny inferred from morphology and DNA data: characterizing well-supported groups in Vanguerieae (Rubiaceae). Bot. J. Linn. Soc. 146: 257–283. Liede, S. & Kunze, H. 2002. Cynanchum and the Cynanchinae (Apocynaceae–Asclepiadoideae)—a molecular, anatomical and latex triterpenoid study. Organisms Divers. Evol. 2: 239–269. Meisner, C.F. 1838. Plantarum Vascularium Genera, Secundum Ordines Naturales Digesta, 1: 162. Libraria Weidmannia, Lipsiae (Leipzig). Meve, U. & Liede, S. 2002. A molecular phylogeny and generic rearrangement of the stapelioid Ceropegieae (Apocynaceae-Asclepiadoideae). Pl. Syst. Evol. 234: 171–209. Motley, T.J., Wurdack, K.J. & Delprete, P.G. 2005. Molecular systematics of the Catesbaeeae-Chiococceae complex (Rubiaceae): flower and fruit evolution and biogeographic implications. Amer. J. Bot. 92: 316–329. Munhá, J., Afonso, R., Caldeira, R. & Mata, J. 2002. Estudo geológico preliminar da região nordeste da Ilha de S Tomé (Folha n° 2, Ana Chaves). Garcia de Horta, ser. Geol. 18: 1–8. Noyes, R.D. 2006. Intraspecific nuclear ribosomal DNA divergence and reticulation in sexual diploid Erigeron strigosus (Asteraceae). Amer. J. Bot. 93: 470–479. Nylander, J.A.A. 2004. MrModeltest, v2. Program distributed by the author. Evolutionary Biology Centre, Uppsala University. Page, R.D.M. 1996. TREEVIEW: an application to display phylogenetic trees on personal computers. Comput. Applic. Biosci. 12: 357–358. Palisot-Beauvois, A.M.F.J. 1810 [1807]. Stipulaire. Stipularia, Fam. des Rubiacées. P. 26 in: Flore d’Oware et de Bénin, en Afrique. Imprimerie de Fain et compagnie, Paris. Persson, C. 2000. Phylogeny of the Neotropical Alibertia group (Rubiaceae), with emphasis on the genus Alibertia, inferred from ITS and 5S Ribosomal DNA sequences. Amer. J. Bot. 87: 1018–1028. Puff, C., Igersheim, A. & Buchner, R. 1998. Character states and taxonomic position of the monotypic Sri Lankan Schizostigma (Rubiaceae–Isertieae). Pp. 187–203 in: Dransfield, J., Coode, M.J.E. & Simpson, D.A. (eds.), Plant Diversity in Malesia III. Royal Botanic Gardens, Kew, London. Queiroz, A. de, Donoghue, M.J. & Im, J.K. 1995. Separate versus combined analysis of phylogenetic evidence. Annu. Rev. Ecol. Syst. 26: 657–681. Razafimandimbison, S.G. & Bremer, B. 2001 [2002]. Tribal delimitation of Naucleeae (Cinchonoideae, Rubiaceae): inference from molecular and morphological data. Syst. Geogr. Pl. 71: 515–538. 16 TAXON 57 (1) • February 2008: 1–17 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., 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. & Miller, J.S. 1999. New taxa and nomenclatural notes on the flora of the Marojejy Massif, Madagascar. III. Rubiaceae. A new species of Sabicea. Adansonia ser. 3, 21: 41–45. Renner, S. 2004. Plant dispersal across the tropical Atlantic by wind and sea currents. Int. J. Pl. Sci. 165 (Supplement): S23–S33. Robbrecht, E. 1988. Tropical woody Rubiaceae. Opera Botanica Belgica 1. National Botanic Garden of Belgium, Meise. Robbrecht, E. 1993. Supplement to the 1988 outline of the classification of the Rubiaceae, index to genera. Pp. 173–196 in: Robbrecht, E. (ed.), Advances in Rubiaceae Macrosystematics. Opera Botanica Belgica 6. National Botanic Garden of Belgium, Meise. Robbrecht, E. 1996. Generic distribution patterns in subsaharan African Rubiaceae (Angiospermae). J. Biogeogr. 23: 311–328. Robbrecht, E. & Manen, J.F. 2006. The major evolutionary lineages of the coffee family (Rubiaceae, angiosperms). Combined analysis (nrDNA and cpDNA) to infer the position of Coptosapelta and Luculia, and supertree construction based on rbcL, rps16, trnL-trnF and atpB-rbcL data. A new classification in two subfamilies, Cinchonoideae and Rubioideae. Syst. Geogr. Pl. 76: 85–146. Rova, J.H.E., Delprete, P.G., Andersson, L. & Albert, V.A. 2002. A trnL-F cpDNA sequence study of the Condamineeae-Rondeletieae-Sipaneeae complex with implications on the phylogeny of the Rubiaceae. Amer. J. Bot. 89: 145–159. Schumann, K. 1896 [1897]. Beiträge zur Flora von Afrika, XIII, Rubiaceae africanae. Bot. Jahrb. Syst. 23: 412–470. Simmons, M.P. & Ochoterena, H. 2000. Gaps as characters in sequence-based phylogenetic analyses. Syst. Biol. 49: 369–381. Steyermark, J.A. 1962. Pittierothamnus, new genus of Rubiaceae. Bol. Soc. Venez. Ci. Nat. 23: 92–95. Steyermark, J.A. 1972. Rubiaceae. Pp. 227–832 in: Maguire, B. & Collaborators, The Botany of the Guayana Highland, Part IX. Mem. New York Bot. Gard. 23: 1–832. Steyermark, J.A. 1974. Rubiaceae. Pp. 7–539 in: Lasser, T., Flora de Venezuela, vol. IX, primera parte. Ministerio de Agricultura y Cría, Caracas. Swofford, D.L. 2000. PAUP*: Phylogenetic Analysis Using Parsimony (*and Other Methods), version 4.0b. Sinauer Associates, Sunderland, Massachusetts. Tate, J.A., Aguilar, J.F., Wagstaff, S.J., LaDuke, J.C., Bodo Slotta, T.A. & Simpson, B.B. 2005. Phylogenetic relationships within the tribe Malveae (Malvaceae, subfamily Malvoideae) as inferred from its sequence data. Amer. J. Bot. 92: 584–602. Thompson, J.D., Higgins, D.G. & Gibson, T.J. 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, posi- TAXON 57 (1) • February 2008: 1–17 Khan & al. • Tribal and generic circumscriptions of Sabiceeae tions-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22: 4673–4680. Verdcourt, B. 1958. Remarks on the classification of the Rubiaceae. Bull. Jard. Bot. État Bruxelles 28: 209–314. Verdcourt, B. 1975. New sectional names and a new tribe Virectarieae (Rubiaceae). Kew Bull. 30: 366. Wernham, H.F. 1914. A Monograph of the Genus Sabicea. British Museum (Natural History), London. White, F. 1979. The Guineo-Congolian region and its relationships to other phytochoria. Bull. Jard. Bot. Natl. Belg. 49: 11–55. White F. 1993. The AETFAT chorological classification of Africa: history, methods and applications. Bull. Jard. Bot. Natl. Belg. 62: 225–281. Appendix. Voucher information and accession numbers for all species included in this study. Species, country origins, voucher, trnT-F acc. no., ITS acc. no. Alberta magna E. Mey., GenBank, AJ620118, – ; A. magna, GenBank, -, AJ224842; Canthium coromandelicum (Burm. f.) Alston, GenBank, AJ847401, – ; C. coromandelicum, GenBank, -, AJ315081; Cinchona pubescens Vahl, GenBank, AJ346963, – ; Coffea eugenioides S. Moore, GenBank, AJ847402, – ; Coussarea sp., GenBank, AF152612, – ; Danais xanthorrhoea (K. Schum.) Bremek., GenBank, AM409329, – ; Ecpoma hierniana (Wernham) N. Hallé & F. Hallé, Thompson 1803 (K), AM409140, AM409055; Euclinia longiflora Salisb., GenBank, AJ847399, – ; Gynochthodes coriacea Blume, GenBank, AJ847407, – ; Heinsia zanzibarica (Boj.) Verdc., GenBank, AJ847377, AJ846880; Hekistocarpa minutiflora Hook. f., Cameroon, Sonké & al. 2708 (BR), AM409141, AM409056; Isertia pittieri (Standl.) Standl., GenBank, AJ847404, – ; Ixora coccinea L., GenBank, AJ620117, – ; I. coccinea, GenBank, -, AJ224826; Luculia grandifolia Ghose, GenBank, AJ346929, – ; Mussaenda pinatubensis Elmer, GenBank, AJ847365, – ; Nauclea orientalis (L.) L., GenBank, AJ346958, – ; Normandia neocaledonica Hook. f., New Caledonia, Munzinger 532 (MO), AM409177, – ; Ophiorrhiza mungos L., GenBank, AF152610, – ; Pentaloncha humilis Hook. f. (2), Gabon, Wilde & al. 10235 (WAG), AM409173, – ; P. humilis (1), Gabon, Breteler & al. 10985 (WAG), AM409174, – ; Pentas parvifolia Hiern, GenBank, AJ847406, – ; Pseudosabicea arborea (K. Schum.) N. Hallé (1), Burundi, Reekmans 11116 (K), AM409167, AM409049; P. arborea (2), Burundi, Reekmans 11116 (WAG), AM409138, AM409050; P. aurifodinae N. Hallé, Gabon, Wieringa & al. 5026 (WAG), AM409162, AM409046; P. batesii (Wernham) N. Hallé, Gabon, Valkenburg & al. 2569 (WAG), AM409139, AM409048; P. medusula (K. Schum. ex Wernham) N. Hallé, Cameroon, Andel & al. 3555 (WAG), AM409163, AM409047; P. mildbraedii (Wernham) N. Hallé, Gabon, Wieringa & al. 5032 (WAG), AM409137, AM409051; P. nobilis (R. Good) N. Hallé, Gabon, Valkenburg & al. 2604 (WAG), AM409165, AM409052; P. proselyta N. Hallé, Gabon, Valkenburg & al. 2646 (WAG), AM409166, AM409053; P. segregata (Hiern) N. Hallé, Gabon, Wieringa & al. 5025 (WAG), AM409164, AM409054; Psychotria amboniana K. Schum., GenBank, AJ847409, – ; Sabicea amazonensis Wernham, Brazil, Campbell & al. P22037 (MO), AM409157, AM409007; S. angolensis Wernham, Republic of the Congo, Lisowski B-7136 (BR), AM409142, AM409006; S. aspera Aubl., French Guiana, Andersson & al. 2003 (NY), AM409143, AM409008; S. brevipes Wernham, Ghana, Jongkind & Nieuwenhuis 2793 (WAG), AM409178, AM409009; S. caminata N. Hallé, Gabon, Wilde & Sosef 10311 (WAG), AM409118, AM409010; S. capitellata Benth., Equatorial Guinea, Sonké & Esono 2533 (BR), AM409161, AM409012; S. chocoana C.M. Taylor, Colombia, Delprete 6342 (NY), AM409144, AM409013; S. cinerea Aubl., French Guiana, Andersson & al. 1903 (NY), AM409120, AM409014; S. congensis Wernham, Gabon, Breteler 12428 (WAG), AM409146, AM409015; S. dewevrei De Wild. & T. Durand, Republic of the Congo, Lemaire 1393 (BR), AM409121, AM409016; S. dinklagei K. Schum., Malawi, Pawek 6510 (UPS), AM409122, AM409017; S. discolor Stapf, Ivory Coast, Jongkind & al. 4880 (WAG), AM409145, AM409018; S. diversifolia Pers., GenBank, AJ847396, AJ846883; S. exellii G. Taylor, São Tomé and Príncipe, Joffroy 188 (BRLU), AM409124, AM409020; S. ferruginea Benth., Liberia, Jongkind & al. 5683 (WAG), AM409125, AM409021; S. fulva Wernham, Gabon, Wieringa & al. 4094 (WAG), AM409126, AM409022; S. gilletii De Wild., Dem. Rep. of the Congo (Zaire), Lejoly 82/903 (BR), AM409154, AM409023; S. glabrescens Benth., Guyana, Gillespie & Tiwari 825 (NY), AM409147, AM409024; S. grisea Cham. & Schltdl., Brazil, Arbo & al. 7191 (NY), AM409159, AM409040; S. harleyae Hepper, Ivory Coast, Jongkind & al. 4867 (WAG), AM409152, AM409025; S. humilis S. Moore, Brazil, Malme 2684 (S), AM409148, AM409026; S. ingrata K. Schum., São Tomé and Príncipe, Ogonnovsky 10 (BRLU), AM409149, AM409027; S. johnstonii K. Schum. ex Wernham, Gabon, Wieringa & al. 4652 (WAG), AM409150, AM409028; S. mattogrossensis Wernham, Bolivia, Beck & Haase 9986 (NY), AM409127, AM409029; S. mexicana Wernham, Mexico, Hahn 639 (NY), AM409153, AM409030; S. najatrix N. Hallé, Gabon, Wieringa & al. 4653 (WAG), AM409128, AM409031; S. orientalis Wernham, Tanzania, Mhoro 443 (UPS), AM409155, AM409032; S. panamensis Wernham, Ecuador, Harling & Ståhl 26896 (S), AM409156, AM409033; S. pyramidalis L. Andersson, Ecuador, Burnham 1455 (F), AM409129, AM409034; S. rosea Hoyle, Ivory Coast, Jongkind 4550 (WAG), AM409158, AM409035; S. seua Wernham, Madagascar, Malcomber & al. 1085 (WAG), AM409130, AM409036; S. speciosa K. Schum., Nigeria, Meer 1623 (WAG), AM409131, AM409037; S. thomensis Joffroy, São Tomé and Príncipe, Ogonnovsky 18 (BRLU), AM409132, AM409038; S. venezuelensis Steyerm., Venezuela, Huber 4201 (NY), AM409133, AM409039; S. venosa Benth., Central Africa Republic, Sonké & Beina 2797 (WAG), AM409134, AM409041; S. villosa Willd. ex Roem. & Schult. (1), Costa Rica, Delprete 5102 (NY), AM409160, AM409042; S. villosa (2), Ecuador, Delprete & Verduga 6396 (NY), AM409135, AM409043; S. vogelii Benth., Ivory Coast, Jongkind & al. 4859 (WAG), AM409136, AM409044; S. xanthotricha Wernham, Cameroon, Sonké 1082 (BR), AM409151, AM409045; Sabicea sp., Bolivia, Nee 46014 (NY), AM409119, AM409011; Schizostigma hirsutum Arn. (= S. ceylanica Puff.), Sri Lanka, Iwarsson 576 (UPS), AM409168, AM409057; Stipularia efulenensis Hutch., Cameroon, Andel 3417 (WAG), AM409123, AM409019; S. elliptica Schweinf. ex Hiern, Dem. Rep. of the Congo (Zaire), Lisowski 56663 (BR), AM409169, AM409058; Tamridaea capsulifera (Balf. f.) Thulin & B. Bremer, Yemen, Miller & al. 10087 (UPS), AM409170, AM409059; Tarenna neurophylla (S. Moore) Bremek., GenBank, AJ847403, – ; Temnopteryx sericea Hook. f. (1), Equatorial Guinea, Wieringa & Haegens 2266 (WAG), AM409175, – ; T. sericea (2), Gabon, Tabak 99 (WAG), AM409176, – ; Virectaria multiflora (Sm.) Bremek., Ivory Coast, Leeuwenberg 2295 (UPS), AM409171, AM409060; V. procumbens (Sm.) Bremek., Liberia, Adams 453 (UPS), AM409172, AM409061; Warszewiczia coccinea Klotzsch, GenBank, AJ847397, AJ846884. 17