TAXON 64 (1) • February 2015: 79–95 De Block & al. • Molecular phylogenetics of Pavetteae (Rubiaceae) Molecular phylogenetics and generic assessment in the tribe Pavetteae (Rubiaceae) Petra De Block,1 Sylvain G. Razafimandimbison,2 Steven Janssens,1 Helga Ochoterena,3 Elmar Robbrecht1 & Birgitta Bremer2 1 Botanic Garden Meise, Nieuwelaan 38, 1860 Meise, Belgium 2 Bergius Foundation, Royal Swedish Academy of Sciences and Botany Department, Stockholm University, 10691 Stockholm, Sweden 3 Instituto de Biología, Universidad Nacional Autónoma de México, Apdo. Postal 70-367, CP 04510, Mexico City, Mexico Author for correspondence: Petra De Block, Petra.DeBlock@BR.fgov.be ORCID (http://orcid.org/): PDB, 0000-0001-7210-7868; SGR, 0000-0002-4545-3893; HO, 0000-0002-3830-4603 DOI http://dx.doi.org/10.12705/641.19 Abstract This is the first phylogenetic study focused on the Pavetteae, one of the most species-rich and morphologically diverse tribes within the coffee family (Rubiaceae). Fifteen of the 17 currently recognized genera, represented by 85 taxa, were sequenced for rps16, trnT-F and ITS and analysed using Bayesian inference and maximum likelihood methods. The monophyly of the Pavetteae is confirmed. Four major lineages are identified, but their phylogenetic relationships are not fully resolved. The continental African genera Rutidea, Nichallea and Tennantia, the Madagascan genera Homollea and Robbrechtia, and the paleotropical genus Pavetta are monophyletic. Other genera are paraphyletic in their current circumscriptions and the following changes are made: Homolliella is placed in synonymy with Paracephaelis, and Coleactina and Dictyandra with Leptactina, resulting in four new combinations. The large paleotropical genus Tarenna is shown not to be monophyletic. In the future, the name Tarenna should not be used for continental African species. Most of these could be transferred to the hitherto monospecific genus Cladoceras, but other species might constitute altogether new genera. The relationship between the monophyletic Asian-Pacific and Madagascan Tarenna species remains unclear. The phylogeny of the Madagascan genera of the Pavetteae is largely unresolved and the largest Madagascar-centred genus Coptosperma was not recovered as monophyletic. The low resolution for the Madagascan taxa can be considered as an indication of rapid radiation. Further molecular and morphological studies are necessary to clarify the phylogeny of the Pavetteae, especially regarding the African Tarenna species and the Madagascan genera of the tribe. Keywords Coleactina; Dictyandra; Homolliella; Madagascar; Tarenna Supplementary Material Electronic Supplement (Figs. S1–S2) and alignment are available in the Supplementary Data section of the online version of this article at http://www.ingentaconnect.com/iapt/tax INTRODUCTION With 17 genera and ca. 650 species (Table 1; Fig. 1), the Pavetteae is one of the largest tribes in the subfamily Ixoroideae of the Rubiaceae (coffee family). The two genera with the highest species numbers, Pavetta L. (ca. 360 spp.) and Tarenna Gaertn. (ca. 200 spp.), have a Paleotropical distribution. Pachystylus K.Schum. (2 spp.) is endemic to Melanesia, and Triflorensia S.T.Reynolds (3 spp.) to Australia. All other genera occur in Africa and/or Madagascar, with Madagascar as an important center of diversity for the tribe. Pavetteae show high morphological variation, especially regarding reproductive characters such as the number of seeds per fruit, seed and placentation type, etc. Representatives of the tribe occur throughout the Paleotropics in both humid and dry vegetation types. The taxonomic history of the tribe is long and complicated, with the name Pavetteae first used in 1829 (Dumortier, 1829; Richard, 1829) but quickly abandoned. The genera included in it, such as Ixora L. and Pavetta, were placed in the tribes Ixoreae or Coffeeae (Richard, 1830). More then a century later, Bremekamp (1934) suggested that within Ixoreae/Coffeeae the genera with terminal inflorescences formed a natural group. Robbrecht (1981, 1984) concurred with him, and elevated this group to tribal rank, resurrecting the name Pavetteae A.Rich. ex Dumort. In this tribe, Robbrecht not only included genera of the Ixoreae/Coffeeae, but also several genera of the Gardenieae (e.g., Tarenna, Leptactina Hook.f.). The Pavetteae are characterized by the following characters: terminal inflorescences; 3- or 4-colporate pollen grains with reticulate or perforate tectum; relatively small, bilocular drupes; one to many seeds per locule, freely surrounding the placenta; seeds with an adaxial hilar excavation, the exotesta often forming a thickened ring around this cavity; exotesta cells with or without thickenings, when present, then thickenings only occurring along the outer Received: 16 Jun 2014 | returned for (first) revision: 4 Aug 2014 | (last) revision received: 16 Dec 2014 | accepted: 16 Dec 2014 || publication date(s): online fast track, n/a; in print and online issues, 2 Mar 2015 || © International Association for Plant Taxonomy (IAPT) 2015 Version of Record 79 De Block & al. • Molecular phylogenetics of Pavetteae (Rubiaceae) TAXON 64 (1) • February 2015: 79–95 tangential wall and containing narrow channels. Shortly after the reinstatement of the tribe, as a result of a detailed morphological study, Bridson & Robbrecht (1985) redelimited the Pavetteae and distinguished two informal groups, one including the genera allied to Tarenna, and the other including the genera allied to Ixora. Molecular studies demonstrated the paraphyletic nature of Pavetteae sensu Bridson & Robbrecht (1985), with Ixora consistently placed separately from the rest of the tribe (e.g., Andreasen & Bremer, 1996, 2000; Andreasen, 1997). Bridson & Robbrecht’s (1985) informal groups were therefore elevated to tribal rank, which resulted in the reinstatement of a much-narrowed tribe Ixoreae A.Gray and the transfer of Ixora, Captaincookia N.Hallé, Doricera Verdc., Myonima Comm. ex A.Juss. and Versteegia Valeton to this newly resurrected tribe. The currently accepted circumscription of Pavetteae follows De Block (2003b). It includes the genera cited as representatives of the tribe by Bridson & Robbrecht (1985) and Robbrecht (1988, 1994) with the exceptions of Duperrea Pierre ex Pit. and Ixora and allied genera (Table 1). The only recent change was the description of two new genera: Robbrechtia De Block, endemic to Madagascar (De Block, 2003b) and Triflorensia, endemic to Australia (Reynolds & Forster, 2005). Molecular studies have shown the tribe Pavetteae in this delimitation to be a strongly supported group (Andreasen & Bremer, Table 1. Overview of the Pavetteae according to recent literature. Genera Geographical distribution No. of species Protologue Inclusion in tribe based on: Genera of Pavetteae in De Block & al., this paper Cladoceras Bremek. East Africa 1 Bremekamp, 1940 2;3;4;5;6;8;9 Cladoceras (including a number of continental African Tarenna) Coleactina N.Hallé West central Africa 1 Hallé, 1970 2;3;4;5;6;8;9 Coptosperma ª Hook.f. = Enterospermum Hiern East and South Africa, Comores, Mascarenes, Madagascar 19 Hooker, 1873 1;2;3;4;6;8;9 Dictyandra Welw. ex Hook.f. Continental tropical Africa 2 Hooker, 1873 1;2;3;4;5;6;8;9 Homollea Arènes Madagascar 3 Arènes, 1960 2;3;6;4;8 Homolliella Arènes Madagascar 1 Arènes, 1960 2;3;4;6;8 Leptactina Hook.f. Continental tropical Africa ca. 20 Hooker, 1871 1;2;3;4;5;6;8;9 Leptactina (including Coleactina and Dictyandra) Nichallea Bridson West and west central Africa 1 Bridson, 1978b 1;2;3;4;6;8 Nichallea Pachystylus K.Schum. New Guinea 2 Schumann, 1889 1;2;3;4;6;8 Pachystylus Paracephaelis Baill. East Africa, Madagascar 4 Baillon, 1879 2;3;4;6;8;9 Paracephaelis (including Homolliella) Pavetta L. Paleotropics (not in Madagascar) ca. 360 Linnaeus, 1753 1;2;3;4;5;6;8;9 Pavetta Robbrechtia De Block Madagascar 2 De Block, 2003b 6;8;9 Robbrechtia Rutidea DC. Continental tropical Africa 22 Candolle, 1807 1;2;3;4;5;6;8;9 Rutidea Schizenterospermum Homolle ex Arènes Madagascar 4 Arènes, 1960 2;3;4;6;8 Schizenterospermum TarennaB Gaertn. Paleotropics ca. 200 Gaertner, 1788 1;2;3;4;5;6;8;9 Tarenna (excluding all continental African species) Tennantia Verdc. East Africa 1 Verdcourt, 1981 2;3;4;5;6;8;9 Tennantia Triflorensia S.T.Reynolds Australia 3 Reynolds & Forster, 2005 7 Triflorensia Coptosperma Homollea Note: this table does not take into account genera that in the past were considered to belong to the Pavetteae but whose new position outside the tribe is generally accepted (such as Ixora and allies, Duperrea). 1, Robbrecht (1984); 2, Bridson & Robbrecht (1985); 3, Robbrecht (1988); 4, Robbrecht (1994); 5, Andreasen & Bremer (2000); 6, De Block (2003b); 7, Reynolds & Forster (2005); 8, Robbrecht & Manen (2006); 9, Bremer & Eriksson (2009). a, sensu De Block & al., 2001 (all species with a single, usually ruminate seed per fruit occurring in Africa, Madagascar, the Seychelles, Mascarenes and Comores, i.e., including Zygoon and Enterospermum). b, sensu De Block & al., 2001 (all species of Tarenna except those with a single, usually ruminate, seed per fruit). 80 Version of Record TAXON 64 (1) • February 2015: 79–95 De Block & al. • Molecular phylogenetics of Pavetteae (Rubiaceae) Fig. 1. Representative taxa of the tribe Pavetteae. A, Homollea longiflora, young fruits; B, Homolliella sericea, flowers; C, Paracephaelis cinerea, flowers; D, Paracephaelis tiliacea, young fruits; E, Rutidea sp., flowers; F, Pavetta schumanniana, flowers; G, Pavetta schumanniana, leaf with bacterial galls; H, Coptosperma sp., leaves and stipules; I, Schizenterospermum rotundifolium, leaves and stipules; J, Coptosperma nigrescens, fruits and flowers; K, Robbrechtia grandifolia, flowers; L, Tarenna grevei, flowers; M, Tarenna thouarsiana, flowers; N, Dictyandra arborescens, flowers; O, Leptactina sp., flowers; P, Leptactina benguelensis, fruit. — Photographs by Petra De Block (A–B, H–I, K, M–N), Steven Dessein (C–E, P) and Frank Van Caekenberghe (F–G, J, L, O). Version of Record 81 De Block & al. • Molecular phylogenetics of Pavetteae (Rubiaceae) TAXON 64 (1) • February 2015: 79–95 2000; Robbrecht & Manen, 2006; Bremer & Eriksson, 2009). However, in previous studies only a limited number of genera were sequenced (Andreasen & Bremer, 2000: 6 of 17 genera; Robbrecht & Manen, 2006: 6/17; Bremer & Eriksson, 2009: 9/17). Furthermore, relationships within the Pavetteae remain unclear and generic circumscription uncertain, especially for Tarenna (De Block & al., 2001) and the Madagascan representatives of the tribe. This is the first molecular phylogenetic study focusing on the Pavetteae, carried out within the framework of ongoing classical taxonomic studies on the tribes Pavetteae and Ixoreae (e.g., De Block, 1998, 2003a, 2005). Fifteen of the 17 genera currently ascribed to the tribe are represented. The main objectives of this study are: (1) to test the monophyly of the tribe Pavetteae, (2) to assess the monophyly of its currently recognized genera, and (3) to assess the phylogenetic relationships among its genera. The results of this study will be used to evaluate, and, if necessary, to improve existing tribal and generic delimitations. species from different tribes belonging to subfamily Ixoroideae (Alberteae, Mussaendeae, Vanguerieae, Ixoreae, Coffeeae, Gardenieae) in accordance with earlier analyses (Robbrecht & Manen, 2006, Bremer & Eriksson, 2009). Five Coptosperma Hook.f. accessions from Madagascar and one Pavetta accession from continental Africa represent undescribed new species. In addition, we were unable to identify two Asian Pavetta species, here referred to as P. sp. B and P. sp. C (despite consulting the relevant taxonomic literature: Bremekamp, 1934, 1939a, b). Accession data of ingroup and outgroup taxa is given in Appendix 1. DNA isolation, amplification and sequencing. — Total genomic DNA was extracted from silica-dried leaf material using either a modified version of the hot CTAB protocol (Saghai-Maroof & al., 1984, Doyle & Doyle, 1987) or the commercial E.Z.N.A.TM High Performance Plant DNA Mini Kit (OMEGA bio-tek, Norcross, Georgia, U.S.A.). Primers for chloroplast rps16, trnT-F and nuclear ribosomal ITS are listed in Table 2. Amplification of rps16 and trnT-F followed protocols of Oxelman & al. (1997) and Razafimandimbison & Bremer (2002), respectively. For ITS, an optimized touchdown PCR program was used for species of Pavetteae in order to provide sufficient DNA. This temperature profile consisted of 3 min initial denaturation at 95°C; 10 cycles of 30 s denaturation at 95°C, 30 s primer annealing at 60°C, and 1 min extension at 72°C; 30 cycles of 30 s denaturation at 95°C, 30 s primer annealing at starting temperature 60°C lowering 0.3°C/cycle, and 1 min extension at 72°C; and a final extension of 7 min at 72°C. Amplification reactions were carried out on a Perkin Elmer GeneAMP 9700 thermocycler or Eppendorf Mastercycler. The PCR mixes for rps16 and trnT-F contained 1 µl genomic DNA, 1 µl of each primer (100 ng/µl), 2.5 µl of 10 mM dNTPs, 2.5 µl Taq Buffer, 0.2 µl KAPA Taq DNA polymerase adjusted with MilliQ water to 25 µl. The PCR mix for ITS was similar to that of rps16 and trnT-F except for the addition of 1 µl DMSO (v/v) for a total of 25 µl. Sequencing reactions were performed using the Big Dye Terminator 3.1 Cycle Sequencing kit (Applied Biosystems, Foster City, California, U.S.A.) on an Applied Biosystems 310 Genetic Analyzer or were sent to Macrogen Inc. (Seoul, Republic of Korea) for sequencing. MATERIALS AND METHODS Taxon sampling. — Fifteen of the 17 genera of the tribe Pavetteae were sampled. Pachystylus and Triflorensia, New Guinean and Australian endemics, respectively, were not included due to lack of material. Apart from the monospecific Cladoceras Bremek, Coleactina N.Hallé, Homolliella Arènes, Nichallea Bridson and Tennantia Verdc., and Dictyandra Welw. ex Hook.f. (comprising two species), all genera were represented by at least two species. For genera with infrageneric classifications, we sampled species from all subgenera (Pavetta) or sections (Rutidea DC.). In total, we included 85 Pavetteae taxa, covering the entire distribution range of the tribe. Our sampling is representative for continental Africa and Madagascar. However, certain regions are underrepresented, such as the western Indian Ocean islands, Malesia/Indochina, Papua New Guinea, Australia and the Pacific islands (with the exception of New Caledonia). Five genera were sequenced here for the first time: Coleactina, Homollea Arènes, Homolliella, Nichallea and Schizenterospermum Homolle ex Arènes. As outgroup, we chose eight Table 2. Amplification primers used. Region rps16 trnT-trnL trnL-trnF ITS 82 Primer Primer sequence (5′–3′) rps16-F GTGGTAGAAAGCAACGTGCGACTT rps16-2R TCGGGATCGAACATCAATTGCAAC trnL-C CGAAATCGGTAGACGCTACG trnL-F ATTTGAACTGGTGACACGAG trnL-A1 ACAAATGCGATGCTCTAACC trnL-I CCAACTCCATTTGTTAGAAC ITS-P17F CTACCGATTGAATGGTCCGGTGAA ITS-26S-82R TCCCGGTTCGCTCGCCGTTACTA ITS-P16F TCACTGAACCTTATCATTTAGAGGA ITS-P25R GGGTAGTCCCGCCTGACCTG Version of Record Reference Oxelman & al., 1997 Taberlet & al., 1991 Razafimandimbison & Bremer, 2002 Lidén & al., 1995 Popp & Oxelman, 2001 TAXON 64 (1) • February 2015: 79–95 De Block & al. • Molecular phylogenetics of Pavetteae (Rubiaceae) Sequence alignment and phylogenetic analyses. — Sequences were automatically aligned with MAFFT v.7.017 (Katoh & al., 2002) using the E-INS-I Algorithm with a scoring matrix of 100PAM/k=2 and a Gap open penalty of 1. The automatically aligned data matrix was subsequently finetuned by hand in the Geneious v.5.6.1 software package. Potentially informative indels were coded as separate characters, following the “simple indel coding” method (Simmons & Ochoterena, 2000). Newly generated sequences have been submitted to GenBank (Appendix 1). The methodology of Wang & al. (2014) was used to infer topological conflicts between different datasets. A threshold with a BS value ≥ 70% and a PP value ≥ 0.95 was applied as an indication of strongly supported incongruence between different datasets. The best-fit nucleotide substitution model for each plastid and nuclear dataset was determined using jModelTest v.2.1.4 (Posada, 2008) under the Akaike information criterion (AIC). For ITS, the GTR + I + G model was found as best fit, whereas the GTR + G model was shown to be the best substitution model for rps16 and trnT-F. Bayesian analyses were conducted with MrBayes v.3.1 (Huelsenbeck & Ronquist, 2001) on three individual data partitions and a combined data matrix. Each analysis was run two times for 5 million generations with trees sampled every 1000 generations. Convergence of the chains was examined with TRACER v.1.4 (Rambaut & Drummond, 2007). Maximum likelihood analyses were computed on the CIPRES web portal using RAxML v.7.2.8 (Stamatakis & al., 2008) under the GTRGAMMA model. Non-parametric bootstrapping was carried out with 1000 bootstrap replicates. RESULTS For this study, we generated 261 new sequences, which were complemented with 7 sequences from GenBank in order to obtain a dataset of 268 sequences, representing 93 taxa. Due to difficulties in amplification and/or sequencing of ITS, Coptosperma madagascariense (Baill.) De Block and Paracephaelis saxatilis (Scott-Elliot) De Block were represented by two different accessions, which were thoroughly examined to make sure they belonged to the same species. For 14 further taxa (both ingroup and outgroup) we were unable to generate ITS sequences. Table 3. Characteristics of individual and combined datasets. rps16 trnT-F ITS Total Number of sequences 94 93 79 94 Characters 918 2217 895 4030 Constant characters 738 1777 614 3129 Variable characters 180 440 281 901 Parsimony-informative characters 117 293 115 525 % Parsimony-informative characters 12.7 13.2 12.8 13.0 Indel-coding characters 51 30 89 6 In general, sequence variability of individual markers was rather low (Table 3). The percentage of potentially parsimonyinformative (PI) characters was ± equal in all gene markers used in this study (trnT-F: 13.2%; ITS: 12.8%; rps16: 12.7%). For the combined nuclear ribosomal ITS and chloroplast trnT-F/ rps16 matrix, 89 indel characters were added (Table 3). The majority-rule consensus topologies from the separate Bayesian inference (BI) and maximum likelihood (ML) bootstrap analyses of rps16, trnT-F, and ITS revealed similar topologies, yet they were largely unresolved. Nevertheless, several clades at generic level were always recovered, e.g., Rutidea/Nichallea, Leptactina/Coleactina/Dictyandra, etc. No supported incongruence (Bayesian posterior probabilities, BPP > 0.95; maximum likelihood bootstrap support, ML-BS > 70) was observed between the different datasets. Hence, they were concatenated for subsequent analyses. The combined chloroplast phylogeny and the ITS phylogeny are given in the Electr. Suppl.: Figs. S1–S2, respectively. BI and ML analyses of the combined dataset provided similar topologies with a similar level of support for most clades (Fig. 2). The monophyly of the tribe Pavetteae was strongly supported in both the BI and ML analyses (ML-BS = 100, BPP = 1.00). Within the tribe there was support for four clades (Fig. 2), but the relationships among these clades were unresolved. Clade I (ML-BS = 100, BPP = 1.0) comprised the continental African genera Nichallea and Rutidea, which were resolved as sister taxa. The monophyly of Rutidea was well supported (ML-BS = 100, BPP = 1.00). Rutidea sect. Rutidea (R. orientalis Bridson to R. decorticata Hiern) was recovered as monophyletic, but R. sect. Tetramera Bridson (R. membranacea Hiern and R. fuscescens Hiern) was not. Clade II (ML-BS = 100, BPP = 1.0) comprised the continental African genera Leptactina, Dictyandra and Coleactina. Leptactina was shown not to be monophyletic in its current delimitation since Coleactina was nested within it. Coleactina papalis N.Hallé was sister to Leptactina benguelensis (Welw. ex Benth. & Hook.f.) R.D.Good (ML-BS = 70, BPP = 1.0). Dictyandra arborescens Welw. ex Hook.f. formed a weakly supported clade (ML-BS = 66, BPP = 0.93) with the sister species L. pynaertii De Wild. and L. mannii Hook.f. (ML-BS = 62, PP: 0.99), the type of Leptactina. Clade III (ML-BS = 100, BPP = 1.0) consisted of the paleotropical Pavetta, the monospecific East African Cladoceras and the continental African species of Tarenna. Within this lineage, Tarenna jolinonii N.Hallé was sister to the weakly supported clade comprising the rest of the sampled taxa (ML-BS = 65, BPP = 0.66). This clade was resolved into two well-supported monophyletic subclades, but the relationships between the two subclades and T. jolinonii remained unclear due to poor support. The first subclade comprised Cladoceras subcapitatum (K.Schum. & K.Krause) Bremek. from East Africa and seven of the ten sampled continental African Tarenna species (MLBS = 93, BPP = 1.0). The second subclade (ML-BS = 92, BPP = 1.0) contained the paleotropical Pavetta with Tarenna bipindensis (K.Schum.) Bremek. and T. precidantenna N.Hallé, which together were sister (ML-BS = 92, BPP = 1.0) to Pavetta (MLBS = 100, BPP = 1.0). Pavetta was recovered as monophyletic Version of Record 83 TAXON 64 (1) • February 2015: 79–95 De Block & al. • Molecular phylogenetics of Pavetteae (Rubiaceae) Razafimandimbisonia humblotii 100/1.0 100/1.0 Mussaenda flava Vangueria madagascariensis Ixora sp. Tricalysia semidecidua Oxyanthus troupinii Gardenia rutenbergiana Euclinia longiflora Nichallea soyauxii 100/1.0 Rutidea membranacea 92/1.0 100/1.0 Clade I Rutidea fuscescens Rutidea orientalis 71/0.79 Rutidea parviflora 100/0.94 57/0.54 Rutidea seretii 94/1.0 Rutidea olenotricha 91/1.0 Rutidea dupuisii Rutidea ferruginea 72/1.0 98/1.0 Rutidea decorticata Leptactina leopoldisecundi Leptactina involucrata 100/1.0 66/0.93 Dictyandra arborescens Leptactina mannii Clade II 62/0.99 Leptactina pynaertii Pavetteae 100/1.0 Leptactina delagoensis */0.51 Leptactina papyrophloea Leptactina epinyctios 74/1.0 Leptactina benguelensis Continental 76/0.95 Africa 70/1.0 Coleactina papalis Tarenna jolinonii Tarenna fuscoflava 57/0.95 Tarenna lasiorachis 89/1.0 Tarenna nitidula 100/1.0 93/1.0 95/1.0 Tarenna pallidula Clade III Tarenna vignei Tarenna roseicosta 70/1.0 Cladoceras subcapitatum 81/1.0 Tarenna pembensis 65/0.66 98/1.0 76/1.0 Tarenna precidantenna Tarenna bipindensis 51/0.50 82/1.0 Pavetta tetramera 92/1.0 Pavetta suffruticosa 85/0.98 Pavetta ternifolia 100/1.0 Pavetta hymenophylla Pavetta sp. A (FTEA) 100/1.0 100/1.0 100/1.0 Pavetta tarennoides Pavetta sansibarica Pavetta schumanniana Pavetta stenosepala 67/0.78 55/0.91 Pavetta batesiana Pavetta vaga 74/1.0 Australia Pavetta agrostiphylla 65/0.53 Pavetta indica Asia 80/0.96 Pavetta sp. B 71/0.99 Pavetta sp. C Continental Tennantia sennii Africa Tarenna alpestris 54/0.95 * /0.98 100/1.0 Tarenna asiatica Tarenna flava Tarenna leioloba 88/1.0 75/0.97 Tarenna rhypalostigma Asia/Pacific Tarenna microcarpa 94/1.0 * /0.96 Tarenna sambucina 97/1.0 Tarenna gracilipes * /0.90 Tarenna attenuata Clade IV * /0.59 Tarenna depauperata 60/0.73 Coptosperma graveolens Continental Coptosperma peteri Africa 100/1.0 Homollea perrieri Homollea longiflora * /0.78 * /0.84 Paracephaelis tiliacea Homolliella sericea 96/1.0 Paracephaelis cinerea 58/0.66 * /0.55 Paracephaelis saxatilis 90/0.93 100/1.0 Robbrechtia grandifolia Madagascar Robbrechtia milleri * / 98/1.0 Tarenna spiranthera Tarenna grevei * /0.86 0.77 92/1.0 Tarenna capuroniana 93/1.0 Tarenna alleizettei Tarenna uniflora 51/0.81 Tarenna thouarsiana 97/1.0 Coptosperma supra-axillare Fig. 2. BI phylogram of Pavetteae and outContinental Africa Coptosperma nigrescens + Madagascar group dataset using rps16, trnT-F and ITS 90/0.71 Coptosperma nigrescens 100/1.0 Coptosperma sp. D sequences. Bayesian posterior probabili95/1.0 Coptosperma sp. E 97/1.0 Madagascar ties and maximum likelihood bootstrap Coptosperma madagascariense * /0.97 69/0.99 Coptosperma sp. A support are indicated. An asterisk (*) Coptosperma littorale Continental Africa 52/0.56 indicates nomenclatural types; names in Coptosperma sp. B 82/0.99 Coptosperma sp. C bold indicate type species of synonymized 52/0.64 Madagascar * /0.85 Schizenterospermum rotundifolium 76/1.0 genera. Abbreviations: FTEA, Flora of 91/1.0 Schizenterospermum grevei Mascarenes tropical East Africa. Coptosperma borbonica * /0.65 * * ** * * * * * ** * * ** * 84 Version of Record TAXON 64 (1) • February 2015: 79–95 De Block & al. • Molecular phylogenetics of Pavetteae (Rubiaceae) and so was P. subg. Dizygoon Bremek. (P. tetramera (Hiern) Bremek., P. suffruticosa K.Schum.: ML-BS = 82, BPP = 1.0). Pavetta subg. Baconia Bremek. (P. hymenophylla Bremek., P. ternifolia Hiern: ML-BS = 85, BPP = 0.98) and subg. Pavetta were also recovered as monophyletic in this study, but only few representatives of these subgenera were included so their monophyly is not certain. The Asian species of Pavetta (P. agrostiphylla Bremek., P. indica L., P. sp. B, P. sp. C) formed a monophyletic group (ML-BS = 80, BPP = 0.96). The Australian P. vaga S.T.Reynolds formed a weakly supported clade with P. batesiana Bremek. from continental Africa (ML-BS = 55, BPP = 0.91). Clade IV (ML-BS = 97, BPP = 1.0) comprised the East African monospecific Tennantia, the Paleotropical Tarenna, the Madagascan endemics Homollea, Homolliella, Robbrechtia and Schizenterospermum and the Madagascar-centered Paracephaelis Baill. and Coptosperma (also present in East Africa and the islands of the western Indian Ocean). Mostly, the nodes in this clade were poorly supported and the relationships between subclades remained unclear. In the first subclade, Tennantia was weakly supported as sister to the rest of the subclade (ML-BS = 54, BPP = 0.95), comprising the Asian and Pacific Tarenna species. The second subclade consisted of Madagascan and Madagascar-centred genera (ML-BS < 50, BPP = 0.78). Homollea (ML-BS = 100, BPP = 1.0) and Robbrechtia (ML-BS = 100, BPP = 1.0) were each supported as monophyletic. Homolliella and Paracephaelis formed a wellsupported clade (ML-BS = 96, BPP = 1.0), with Homolliella sericea Arènes nested within Paracephaelis. The Madagascan Tarenna species were a well-supported monophyletic group (ML-BS = 93, BPP = 1.0), comprising three main lineages: T. spiranthera (Drake) Homolle, T. capuroniana De Block and T. grevei (Drake) Homolle (ML-BS = 98, BPP = 1.0); T. uniflora (Drake) Homolle and T. thouarsiana (Drake) Homolle (ML-BS = 97, BPP = 1.0); and, T. alleizettei (Dubard & Dop) De Block, weakly supported as sister to the latter two species (ML-BS = 51, BPP = 0.81). The Afro-Madagascan Coptosperma was not recovered as monophyletic. The BI analysis showed that the type of the genus name, C. nigrescens Hook.f., formed a monophyletic group with all sampled Madagascan Coptosperma species, C. littorale (Hiern) Degreef from East Africa and C. borbonicum (Hend. & Andr.Hend.) De Block from the Comores (BPP = 0.97). In the ML analysis, the same group was recovered, with the exceptions of C. supra-axillare (Hemsl.) Degreef and C. borbonicum (ML-BS = 69, BPP = 0.99). In both analyses the Madagascan endemic Schizenterospermum was nested within the Coptosperma group comprising C. nigrescens. The continental African C. graveolens (S.Moore) Degreef and C. peteri (Bridson) Degreef were unresolved. Our results clearly showed the polyphyly of the large genus Tarenna. The type, T. asiatica (L.) Kuntze ex K.Schum., was situated in the Asian-Pacific subclade of clade IV. The Madagascan representatives of Tarenna were more closely related to Coptosperma, Paracephaelis and other Madagascan genera in the second subclade of clade IV than to the Asian-Pacific Tarenna. The continental African Tarenna were more closely related to Cladoceras and Pavetta in clade III. DISCUSSION Monophyly and main lineages of the Pavetteae. — Our study, the first to include all but two of the currently recognized genera of the Pavetteae (see Table 1 for a complete list), supports a monophyletic tribe Pavetteae in agreement with earlier studies (Andreasen & Bremer, 2000 ; Robbrecht & Manen, 2006; Bremer & Eriksson, 2009). It also supports the current circumscription of the Pavetteae, notably sensu De Block (2003b) (Table 1). While the circumscription of the tribe as a whole is supported, several genera are not monophyletic as currently delimited. We retrieved four main clades, all strongly supported in both BI and ML analyses. However, phylogenetic relationships among these lineages remained largely unresolved. Clades I and II were restricted to continental Africa and respectively comprised the genera Rutidea/Nichallea and Coleactina/Dictyandra/Leptactina. Clade III comprised the continental African Cladoceras and Tarenna species and the paleotropical Pavetta. Clade IV was the most heterogeneous with respect to geographic distribution; it comprised eastern and southern African (Tennantia, Paracephaelis, Coptosperma), Asian-Pacific (Tarenna), Madagascan (Coptosperma, Homollea, Homolliella, Paracephaelis, Robbrechtia, Schizenterospermum, Tarenna) and western Indian Ocean island elements (Tarenna, Paracephaelis, Coptosperma). Finding morphological synapomorphies for these four clades was not simple. Clade I comprised species with an incompletely bilocular ovary with two basally attached ovules and a unilocular fruit with a single, partially or completely ruminate seed. While fruits with a single ruminate seed also occur in some Afro-Madagascan taxa such as Coptosperma, the placentation in the latter is always different. Taxa in clade II were chacterized by large stipules, large and accrescent calyx lobes, well-developed flowers with pubescent corollas, massive placentas with numerous ovules, numerous small angular seeds with entire endosperm and (3–)4-colporate pollen grains. Within the Pavetteae, 4-colporate pollen grains are only present in one other species, Cladoceras subcapitatum. Clades III and IV were more difficult to characterize. Species of clade III were characterized by two to several seeds per fruit, seeds with a large adaxial hilar cavity, entire endosperm and exotestal cells with thickenings along the outer tangential wall. This combination of characters is by no means exclusive for clade III, but also present in Tennantia and in the Asian/ Pacific and Madagascan species of Tarenna in clade IV. Clade IV encompassed most of the morphological variation of the tribe, with seed number per fruit ranging from one to numerous, seed shape varying from angular to spherical or laterally flattened, exotesta cells parenchymatic or with thickenings, etc. No morphological synapomorphies can be identified at this point. It is premature to propose a subtribal classification for the Pavetteae, since the relationships between the four main clades retrieved in our analysis cannot yet be ascertained and the clades not unambiguously characterized by morphological features. Version of Record 85 De Block & al. • Molecular phylogenetics of Pavetteae (Rubiaceae) TAXON 64 (1) • February 2015: 79–95 Relationships and circumscriptions of the genera of the Pavetteae. — Of the 15 genera studied, 6 were supported as Coleactina, Dictyandra and Leptactina (clade II). — This study included 8 of approximately 20 species attributed to Leptactina, 1 of the 2 Dictyandra species, as well as Coleactina papalis. Coleactina, Dictyandra and Leptactina formed a strongly supported monophyletic group. This corroborates earlier morphological observations that considered Coleactina, Dictyandra and Leptactina as closely related (Robbrecht, 1984; De Block & Robbrecht, 1997, 1998). In our study, Coleactina was nested within Leptactina. It can therefore not be retained as a separate genus, but must be considered a synonym of Leptactina. As for Dictyandra, while the type D. arborescens grouped with the type of Leptactina, the support was not very high (ML-BS = 66, BPP = 0.93). Therefore, its exact position within Leptactina could not be ascertained. Based on morphological characters, however, Dictyandra is here placed in synonymy with Leptactina. Originally, Dictyandra was distinguished from Leptactina based on a number of flower characters such as, e.g., multilocellate anthers (Hooker, 1871, 1873). With the inclusion of more species in Leptactina and Dictyandra, it became evident that the flower characters used to distinguish the two genera were not clear-cut. As a pragmatic solution, Robbrecht (1984) retained Dictyandra as a small satellite genus of Leptactina based on differences in corolla tube length (much longer than the lobes in Leptactina, shorter than or equal to the lobes in Dictyandra). Coleactina was described as a separate genus because of its axillary, opposite, uniflorous inflorescences, 6–7-merous flowers, well-developed calyx tube (ca. 2 cm long), large stipules (≥ 4 cm long) forming an involucre together with the inflorescence-supporting leaves, and the presence of a second—small—involucre surrounding the ovary (Hallé, 1970). In our study, C. papalis, a monocaul dwarf up to 1.5 m tall of the humid lowland forests of west central Africa, grouped together with L. benguelensis and L. epinyctios Bullock ex Verdc., which are geofrutices up to 75 cm tall from the dry savanna woodland (miombo) in eastern Africa. The inclusion of Coleactina and Dictyandra in Leptactina is supported by stipule, placentation (De Block & Robbrecht, 1997), seed, seed-coat and pollen characters (Hallé, 1970; Robbrecht, 1984; De Block & Robbrecht, 1998). Furthermore, several characters used to differentiate between the genera are no longer considered valid. For Dictyandra, this is the case for the length difference between corolla tube and lobes and the multilocellate anthers (Neuba, 2006). For Coleactina, the axillary, uniflorous inflorescences are also present in, e.g., L. deblockiae Neuba & Sonké (Neuba & al., 2006). Other characters, such as the well-developed calyx tube and the involucre surrounding the ovary, can be considered apomorphies for C. papalis. Leptactina as newly delimited here can easily be recognized by large stipules, large pubescent corollas, welldeveloped calyces, massive placentas with numerous ovules, large bilocular fruits (up to 3 cm long) with many small angular seeds and (3–)4-colporate pollen grains. monophyletic and their generic delimitation was corroborated. Of these, only Rutidea and Nichallea (clade I) belonged to a well-resolved clade. Pavetta (clade III), Homollea, Robbrechtia and Tennantia (clade IV) belonged to less-resolved clades. Nevertheless, support for these genera was high and they are well characterized morphologically. Leptactina and Paracephaelis were shown to be paraphyletic unless smaller genera are included: Coleactina in Leptactina, Homolliella in Paracephaelis. Tarenna was shown to be polyphyletic and the delimitation of the Afro-Madagascan Coptosperma remained unclear with the genus not recovered as monophyletic. Rutidea and Nichallea (clade I). — This study investigated 10 of 22 species of the tropical African Rutidea as well as the widespread West and west central African Nichallea soyauxii (Hiern) Bridson, the only representative of Nichallea. Rutidea was supported as monophyletic and as sister to Nichallea. This corroborates the close relationship between the two genera as already proposed by Bridson & Robbrecht (1985) and De Block (1995) using morphological characters. Nichallea and Rutidea share a unique type of placentation within Pavetteae (De Block, 1995). The ovary is bicarpellate but incompletely bilocular due to an incomplete septum. Each locule has a single, basally attached, semi-campylotropous ovule. Because of the abortion of one ovule/locule, the drupaceous fruits have a single ovoid, ruminate seed. The exotesta cells are parenchymatic. This fruit and seed type is common in the Madagascan representatives of Pavetteae (clade IV), and occurs, e.g., in Coptosperma, which is also present in continental Africa. While closely related, there are enough morphological differences to keep Nichallea and Rutidea separate, notably lianescent (Rutidea) versus shrubby (Nichallea) habit, pubescent (Rutidea) versus glabrous (Nichallea) leaves and inflorescences, compact (Rutidea) versus lax (Nichallea) inflorescences, small (Rutidea) versus larger (Nichallea) flowers, completely ruminate (Rutidea) versus incompletely ruminate seeds (Nichallea: rumination restricted to the area around the hilar cavity; Bridson, 1978a, b) and a somewhat different pollen type (sexine rugulate to microreticulate in both genera except in R. dupuisii De Wild., the sexine of which is reticulate, but muri grooved longitudinally and columellae equal in height in Rutidea versus muri not grooved longitudinally but undulating as a result of height difference in the columellae in Nichallea: De Block & Robbrecht, 1998). Two sections, Rutidea sect. Tetramera (R. fuscescens, R. membranacea) and sect. Rutidea (all other species included in this study), are recognized within Rutidea. They are distinguished based on, e.g., flower merosity (4-merous in sect. Tetramera versus 5-merous in sect. Rutidea) and stigma type (ovate to globose in sect. Tetramera versus fusiform to clavate in sect. Rutidea; Bridson, 1978a). Our analysis showed sect. Rutidea to be monophyletic but this was not the case for sect. Tetramera. A more in-depth study of Rutidea comprising the third species of sect. Tetramera is necessary to clarify the status of this section. 86 Tarenna, Pachystylus, Triflorensia (clades III and IV). — Tarenna, as traditionally circumscribed (e.g., Bridson, 1979; Smith & Darwin, 1988), consists of approximately 200 species, occurring in continental Africa, Madagascar, the western Version of Record TAXON 64 (1) • February 2015: 79–95 De Block & al. • Molecular phylogenetics of Pavetteae (Rubiaceae) Indian Ocean islands, Asia and the Pacific region. Recently, the circumscription of Tarenna was emended by the transfer to Coptosperma of all African, Madagascan and western Indian Ocean island species possessing fruits with a single, usually ruminate seed (Degreef & al., 2001; De Block & al., 2001). Even in this narrower sense, Tarenna remains variable in certain morphological characters, such as flower, fruit, seed and pollen types (De Block & Robbrecht, 1998; De Block & al., 2001), which has raised questions as to its delimitation (e.g., Bridson, 1979; Smith & Darwin, 1988; De Block & al., 2001). In the present study, we included 26 Tarenna representatives, 10 from continental Africa, 6 from Madagascar, and 10 from Asia and the Pacific. Our phylogeny showed that Tarenna was not monophyletic as currently delimited, since representatives occurred in both clades III and IV. They formed three geographically separate groups. The Madagascan and AsianPacific representatives each formed a monophyletic group, with the latter group containing the type of the genus name, T. asiatica. The continental African representatives, however, were paraphyletic with respect to Pavetta and Cladoceras. The relationship between the Asian/Pacific and the Madagascan Tarenna species remained unresolved, but it was clear that the continental African species must be excluded from Tarenna. The polyphyly of Tarenna was previously suggested by Bridson (1979) and Smith & Darwin (1988), based on morphological characters. Over the years, Tarenna seems to have become the “dustbin genus” of Pavetteae, comprising species with different flower, fruit, seed, seed coat, placentation and pollen types. Because of the wide distribution and the large size of the genus, studies have been restricted to limited geographic regions (e.g., Jérémie, 1974; Bridson, 1988, 2003a; Reynolds & Forster, 2005; Degreef, 2006), with no one evaluating the variation in the entire genus. Our analysis clearly showed that in the future the name Tarenna should not be used for species from continental Africa. At least one morphological character supports this redelimitation of Tarenna: exotesta cell shape. The Asian/Pacific and Madagascan species have exotestal cells with straight walls, whereas (most) continental African species have exotestal cells shaped like puzzle pieces with the cell walls wavy near the surface and straight near the endosperm (Robbrecht & Bridson, 1984; De Block & al., 2001). More than 140 species of Tarenna occur in Asia and the Pacific region. Therefore, our sampling of 10 species was too limited to infer the delimitation of and the relationships between taxa (generic and infrageneric). Also, two small genera closely related to Tarenna, Pachystylus and Triflorensia, were not included in this study because of lack of sequence data. The New Guinean Pachystylus is unusual in having unisexual flowers but has been considered closely allied to (Bremekamp, 1934) or part of Tarenna (Valeton, 1926). The Australian Triflorensia comprises three species that were only recently transferred from Tarenna (Ali & Robbrecht, 1991) based on, amongst other characters, the two ovules/seeds per locule (versus 3 to many), the 4- or 5-merous flowers (versus 5-merous) and the inflorescence branches terminated by 3-flowered dichasial cymes (versus not terminated by 3-flowered dichasial cymes; Reynolds & Forster, 2005). According to Reynolds & Forster (2005) the genus may also be represented in New Caledonia and they suggest Tarenna microcarpa (Guillaumin) Jérémie (included in our study) as a possible species to be transferred to Triflorensia. Recently, in a study of the tribe Gardenieae, Mouly & al. (2014) found that Pelagodendron vitiense Seem. from Fiji was closely related to Tarenna. They suggested inclusion of Pelagodendron Seem. in the tribe Pavetteae. Previously, the genus was considered a synonym of Aidia Lour. and placed in the Gardenieae. Pelagodendron vitiense posesses pseudo-axillary inflorescences, a calyx that completely encloses the corolla in bud and eventually ruptures into two or three broad lobes, and seeds that are embedded in placental pulp, all characters absent or unusual in the Pavetteae. Since Mouly & al. (2014) only sampled one of the three species of Pelagodendron, and since we have not been able to examine material of the genus, we only mention Mouly & al.’s (2014) finding without commenting on it. The Madagascan Tarenna species in clade IV formed a monophyletic group (ML-BS = 93, BPP = 1.00) within the unresolved, predominantly Madagascan clade. Two subclades were well supported as monophyletic groups, each characterized by different morphological characters: T. thouarsiana and T. uniflora (ML-BS = 97, BPP = 1.00) have partially exserted stigmas and anthers and microreticulate or reticulate pollen without supratectal elements but with undulating sexine as a result of different columellae heights (De Block & al., 2001) whereas T. capuroniana, T. spiranthera and T. grevei (ML-BS = 98, BPP = 1.00) possess flowers with long-exserted stigmas and anthers. The pollen is microreticulate without supratectal elements in T. capuroniana and T. spiranthera and perforate with supratectal elements in T. grevei (De Block & al., 2001). This last species is also characterized by 4-merous flowers and a single ovule/seed per locule, characters which, within Tarenna, are only found in the closely related T. sechellensis (Baker) Summerh. from the Seychelles and the Comoros. Tarenna alleizettei differs from all other Madagascan species by its short flowers and microreticulate pollen without supratectal elements. The group of 10 continental African Tarenna species (of 42 species in total; clade III) was paraphyletic with respect to Cladoceras and Pavetta. Most species grouped with Cladoceras subcapitatum (ML-BS = 93, BPP = 1.00), the only species of the East African Cladoceras (Bremekamp, 1940; Verdcourt, 1988b). Cladoceras subcapitatum was considered closely related to Tarenna (Bridson, 1979; Robbrecht & Bridson, 1984) but was placed in a genus of its own because of the presence of spines and the absence of secondary pollen presentation (Bremekamp, 1940). However, the species is similar in habit to the East African Tarenna junodii (Schinz.) Bremek., wich has secondary pollen presentation but no spines. Furthermore, fruit and placentation characters of Cladoceras are identical to those of the continental African Tarenna species. Also, C. subcapitatum possesses the same puzzle-shaped exotestal cells as present in the continental African Tarenna species it grouped with. This seed-coat type is further only present in Leptactina and Pavetta (De Block & al., 2001). The identical fruit, placentation and exotestal cell characters certainly support the merging of Version of Record 87 De Block & al. • Molecular phylogenetics of Pavetteae (Rubiaceae) TAXON 64 (1) • February 2015: 79–95 continental African Tarenna with Cladoceras. The absence of secondary pollen presentation in Cladoceras subcapitatum should not contradict such merger. As Verdcourt (1988b) stated, “reduced forms of the secondary pollination mechanism could be evolved easily”. In fact, another Pavetteae genus, Leptactina, also comprises species with and without secondary pollen presentation (Puff & al., 1996). The only remaining difference is in the pollen, which is (3–)4-colporate in C. subcapitatum versus 3-colporate in the continental African Tarenna. We do not see this as an obstacle to transfer certain continental African Tarenna species to Cladoceras. However, other Tarenna species seem to represent separate lineages. This is the case for Tarenna jolinonii (Hallé, 1970), which fell separate from the other continental African Tarenna species and was weakly supported as sister to the rest of clade III (ML-BS = 65, BPP = 0.66). While most of its characters fit well within Tarenna (e.g., 5-merous flowers, large axillary placentas with 12–15 embedded ovules, angular seeds with a deep hilar cavity), T. jolinonii differs from all other continental African Tarenna species by having large foliaceous stipules (up to 3.5 cm long). These stipules are different from the smaller stipules of the other species that have triangular or ovate, truncate or aristate sheaths, which are often blackened in the central area when dry (Hallé, 1970; Bridson, 1988, 2003a; Degreef, 2006). Furthermore, the exotesta of T. jolinonii has straight-walled cells rather than the puzzle-shaped cells present in Cladoceras subcapitatum and the species grouping with it (De Block, pers. obs.). The last two Tarenna species included in this study are the sister species T. bipindensis and T. precidantenna, which, together, formed the sister group of Pavetta. This relationship was well supported (ML-BS = 92, BPP = 1.00) but puzzling, because the two species certainly do not resemble Pavetta in their characters (e.g., 5-merous versus 4-merous flowers in Pavetta, up to six seeds versus two seeds per fruit in Pavetta). Tarenna bipindensis and T. precidantenna possess exotestal cells with wavy cell walls as is the case for the other continental African Tarenna species (excepting T. jolinoni) and for the Pavetta species. At present, the position of these two continental African Tarenna species as sister to Pavetta cannot be explained. It should be noted, though, that three different samples of T. precidantenna were analysed in order to exclude sequencing errors. Based on the present results, the continental African species must be excluded from Tarenna. At least two options are possible. The first option would require the amalgamation of continental African Tarenna and Cladoceras with Pavetta. Similarities in certain characters would support this inclusion, e.g., both groups have seeds with an adaxial excavation surrounded by a thickened annulus and exotestal cells with mostly continuous thickenings along the outer tangential wall but containing channels to the cell lumina (Robbrecht, 1984). However, there are important differences between Pavetta and continental African Tarenna and Cladoceras: e.g., presence versus absence of bacterial leaf galls, 4-merous versus 5-merous flowers, one versus several ovules per locule (but exceptions do occur: e.g., Tarenna burttii Bridson possesses a single ovule per locule and species of Pavetta subg. Dizygoon possess two ovules per locule. However, together these two groups represent less than 10 species of a total of ca. 560 in Tarenna and Pavetta). The inclusion of continental African Tarenna and Cladoceras would make the genus Pavetta heterogeneous and is therefore not desirable. We prefer a second option in which most of the continental African Tarenna species are transferred to the hitherto monospecific Cladoceras. However, two additional continental African Tarenna lineages (T. jolinonii and T. bipinensis/T. precidantenna) were retrieved in our analysis. We therefore plan a more extensive molecular and morphological study of the continental African Tarenna species in order to clarify relationships within this group before making nomenclatural changes. Further study is also needed to clarify relationships between the Madagascan and AsianPacific species of the genus and the relationships within the Asian-Pacific clade. Pavetta (clade III). — Pavetta, as traditionally delimited, includes ca. 360 species. It is the largest genus of the Pavetteae and has a paleotropical distribution, although it is absent from Madagascar. Pavetta species are easily recognised by their 4-merous flowers, stigmatic lobes fused over most of their length, single ovule per locule and hemispherical seeds with a large adaxial hilar cavity. Typical is the presence of bacterial leaf galls (in ca. 85% of the species; Miller, 1990; Lemaire & al., 2012). Pavetta was extensively studied by Bremekamp (e.g., 1934, 1939a, b), who recognized three subgenera. As presently circumscribed, subg. Pavetta, subg. Baconia and subg. Dizygoon are well defined morphologically, notably by flower, fruit (Bremekamp, 1934, 1939a, b) and pollen type (De Block & Robbrecht, 1998). Most distinct are the west central African species of subg. Dizygoon, which differ from all other Pavetta species by having short anthers (versus long anthers), two collateral ovules per locule (versus one ovule per locule), red fruits (versus black) and pollen with reticulate sexine and no supratectal elements (versus pollen with microreticulate to perforate sexine, with supratectal elements in Pavetta, without supratectal elements in Baconia; De Block & Robbrecht, 1998). Subgenus Baconia is restricted to continental Africa, subg. Pavetta is paleotropical (but absent in Madagascar). In our study we included 15 species of Pavetta: 10 from Africa, 1 from Australia (P. vaga) and 4 from Asia (P. agrostiphylla to P. sp. C; Fig. 2). Subgenus Pavetta was represented by 10 species of a total of approximately 300, subg. Baconia by 2 species of approximately 60 and subg. Dizygoon by 2 of 4 species. Pavetta was recovered as monophyletic in its current circumscription and so was subg. Dizygoon (P. tetramera, P. suffruticosa). Subgenus Baconia (P. ternifolia, P. hymenophylla) and subg. Pavetta (P. tarennoides to P. sp. C.) were recovered as monophyletic only if P. sp. A of Flora of Tropical East Africa (Bridson, 1988) is considered part of subg. Pavetta. However, the taxonomic placement of P. sp. A is difficult since no flowering material is available, which makes it impossible to score important distinguishing characters (corolla length, style length, throat bearded or not). Bridson (1988, 2001) listed P. sp. A as new and poorly known but closely related 88 Version of Record TAXON 64 (1) • February 2015: 79–95 De Block & al. • Molecular phylogenetics of Pavetteae (Rubiaceae) (morphologically) to P. roseostellata Bridson, a species which resembles those of subg. Baconia by the short corolla tube, the bearded throat and the short style (table 2 in Bridson, 2001). Therefore, Bridson placed the two species “very provisionally” (2001: 573) in subg. Baconia, while at the same time stressing that they differ from members of Baconia in inflorescence position: terminal on very short lateral branches versus terminal on normal-length lateral branches. Bridson further suggested the two species may eventually need to be accommodated in a new genus. Because of the incomplete knowledge of P. sp. A and because the subgenera Pavetta and Baconia were represented by only a few species in our study, it was impossible to confirm their monophyly. Within the paleotropical subg. Pavetta, the Asian taxa (P. agrostiphylla, P. indica, P. sp. B, P. sp. C) formed a monophyletic group. The Australian P. vaga was not included in the Asian clade but formed a weakly supported clade with the African P. batesiana. The monophyly of Pavetta has previously been questioned by Bremekamp (1934: 20) who stated that subg. Dizygoon comprises “a group of species which are so markedly different from all others [Pavetta] that it is questionable whether they would not be more at home in a genus of their own”. Our results showed that Dizygoon is the sister lineage to the rest of Pavetta and therefore support the distinction between subg. Dizygoon and the rest of the genus. A further study including more Pavetta representatives and more genetic markers, and taking into account morphology, is necessary to ascertain the monophyly of subgenera Baconia and Pavetta, the taxonomic status of P. sp. A and P. roseostellata and the taxonomic rank of Dizygoon. Tennantia (clade IV). — In our study, this East African monospecific genus was well-supported as a member of the predominantly Madagascan/Asian/Pacific clade IV and weakly supported as sister to the Asian-Pacific Tarenna species. In the past, Tennantia sennii (Chiov.) Verdc. & Bridson was considered closely related to Tarenna, but placed in a separate genus because of its distinctive characters, such as its clavate, winged stigma, corolla tube half as long as the corolla lobes and seeds shaped like segments of an orange, with a groove on each side of the hilar cavity (Verdcourt, 1981, 1988a). Bridson & Robbrecht (1985) also considered Tennantia to be closely related to Tarenna, mainly because of the similar seed-coat structure and exotestal thickenings. Our results do not support a close relationship between Tennantia and the continental African Tarenna species (clade III), but rather between Tennantia and the Asian-Pacific Tarenna species. This grouping is corroborated by the fact that both Tennantia (Verdcourt, 1981; Bridson & Robbrecht, 1985) and the Asian-Pacific Tarenna species (De Block & al., 2001) have exotesta cells with straight walls. Furthermore, Tennantia seeds show intrusions of the seedcoat in the endosperm on the lateral surfaces, which is a form of superficial rumination. Superficially ruminate seeds are characteristic for one group of Asian-Pacific Tarenna species, including T. asiatica, the type of the genus name. Further morphological and molecular studies are needed to unravel the precise relationship between Tennantia and the Asian-Pacific Tarenna species. Homollea, Homolliella and Paracephaelis (clade IV). — In this study we included two species of Homollea, the monospecific Homolliella and three Paracephaelis species. These three genera form a morphologically distinct group within the Pavetteae. They can readily be distinguished by laterally flattened seeds with a shallow linear hilum and entire endosperm, two to seven ovules arranged on the periphery of the placenta (Capuron, 1973; De Block, unpub. data) and by pollen with supratectal microgemmae. The occurrence of supratectal elements is rare within Pavetteae and otherwise only seen in Pavetta subg. Pavetta (absent from Madagascar), in the AfroMadagascan Coptosperma nigrescens and in a few Madagascan Tarenna species (De Block & Robbrecht, 1998; De Block & al., 2001). The placentation and seed types in the three genera are not found in any other Pavetteae. Bridson & Robbrecht (1985) tentatively included Homollea, Homolliella and Paracephaelis in the Pavetteae, a position which is corroborated by our study. Because of their distinctive morphological characters, the three genera were considered closely related (Bridson & Robbrecht, 1985; De Block, 1997). In our study, this close relationship was confirmed for Homolliella and Paracephaelis, which formed a monophyletic group. Homollea was also retrieved as a monophyletic lineage but its relationship with the Paracephaelis-Homolliella clade remained unresolved (a sister relationship is weakly supported in the BI analysis: BPP = 0.84). Our results suggest the inclusion of Homolliella in Paracephaelis, a finding that was already proposed (but not published) by Capuron (1973). However, Capuron considered Homollea, Homolliella and Paracephaelis synonymous with Tarenna in Madagascar because of the seeds with entire endosperm. Within Tarenna, he recognized five sections, with Homollea and Paracephaelis (including Homolliella) among them. Our study does not support the congenerity of Homollea and Paracephaelis (including Homolliella) with the Madagascan Tarenna species. We therefore keep them at generic rank. Next to a similar placentation and seed type, Homolliella and Paracephaelis share the following characters: vegetative parts pubescent; flowers densely pubescent, inflorescences terminal on lateral branches, sessile. The only differences are quantitative: a well-developed calyx with the calyx tube much longer than the lobes and large flowers and fruits in H. sericea versus a less-developed calyx with the lobes longer than or equal in length to the tube and small flowers and fruits in Paracephaelis. Based on morphological characters, Homolliella can easily be accommodated within Paracephaelis. Homollea differs from the Paracephaelis-Homolliella clade by the absence of pubescence in the vegetative and generative parts (except in H. perrieri Arènes), the pseudoaxillary position of the pedunculate inflorescences and the corolla tube, which is much longer than the corolla lobes. All three Homollea species have long, narrow calyx lobes but this character is also encountered in some, as yet undescribed, Paracephaelis species (De Block, pers. obs.). Coptosperma and Schizenterospermum (clade IV). — Coptosperma was recently reinstated (De Block & al., 2001; Degreef & al., 2001) to accomodate African, Madagascan and western Indian Ocean island species previously placed in Version of Record 89 De Block & al. • Molecular phylogenetics of Pavetteae (Rubiaceae) TAXON 64 (1) • February 2015: 79–95 Tarenna (Bridson, 1988), and characterized by a single, usually ruminate seed per fruit, Hitherto, Coptosperma was a monospecific genus, only comprising C. nigrescens. Upon reinstatement, however, it was emended to include Zygoon Hiern and Enterospermum Hiern, the species of which also possess fruits with a single, usually ruminate seed. In the present paper, the name Coptosperma is always used in this emended sense. Coptosperma comprises approximately 40 species endemic to Madagascar (mostly undescribed: De Block, unpub. data), 9 species endemic to eastern and southern Africa and 2 species endemic to the Mascarenes. A further 2 species, C. nigrescens, the type, and C. supra-axillare, occur in Madagascar, continental Africa and the western Indian Ocean islands (Bridson, 1979, 1988; De Block & al., 2001; Degreef & al., 2001). Schizenterospermum is a genus of 4 species, restricted to the dry forest and scrub vegetation of western and northwestern Madagascar (Arènes, 1960). Species of Coptosperma and Schizenterospermum typically have multiflorous, often compact inflorescences with small (4–)5-merous white flowers and unilocular fruits with a single ruminate seed. Morphological characters that show variation within this group of species are, e.g., stipule shape (triangular with apiculate tip, ovate with rounded tip, …), placentation (ovules immersed in or pendulous from placenta), and degree of rumination of the seed (endosperm completely or partly ruminate or entire). We included 12 species of Coptosperma and 2 of Schizenterospermum, all of which fell in the predominantly Madagascan clade IV. In the past, Coptosperma was often considered a synonym of Tarenna, notably by botanists focusing on continental African Pavetteae, where ruminate endosperm is rare (Hiern, 1877; Bremekamp, 1934; Bridson, 1979, 1988). However, more than 60% of the Madagascan species of the Pavetteae have seeds with ruminate endosperm. Botanists studying Madagascan Pavetteae therefore attributed great importance to this character, placing species with two or more seeds and entire endosperm in Tarenna and those with a single seed and ruminate endosperm in Coptosperma (as Enterospermum; Homolle, 1938; Capuron, 1973). Our study favours this last view and corroborates the decision to separate Coptosperma and Tarenna at genus level (De Block & al., 2001; Degreef & al., 2001). However, we did not recover Coptosperma as a monophyletic group and the relationships between certain of its species remained unclear, especially between the Madagascan and the continental African species. In the BI analysis, 10 of the 12 Coptosperma and the 2 Schizenterospermum species formed a predominantly Madagascan, monophyletic group (BPP = 0.97), which included the type C. nigrescens (henceforth: main Coptosperma clade). In the ML analysis, the same group (ML-BS = 70) was recovered with the exclusion of C. supra-axillare (continental Africa, Indian Ocean islands and Madagascar) and C. borbonicum (Mascarenes), which were unresolved. In both analyses, two exclusively continental African species, C. graveolens and C. peteri, were also unresolved in clade IV, but a third, C. littorale, was nested within the main Coptosperma clade. The continental African species of Coptosperma are rather variable in their characters, e.g., inflorescences compact (e.g., C. neurophylla (S.Moore) Degreef) or lax (e.g., C. nigrescens), endosperm ruminate (e.g., C. nigrescens), partly ruminate (e.g., C. neurophylla) or entire (e.g., C. graveolens), inflorescences few-flowered (e.g., C. kibuwae (Bridson) Degreef) or multiflowered (e.g., C. graveolens), etc. Bridson (1979, 1988, 2003b) therefore recognized three informal infrageneric groups, more or less coinciding with the historically recognized genera Zygoon (Bridson’s group I), Enterospermum (Bridson’s group II) and Coptosperma s.str. (Bridson’s group III). Species of Bridson’s group I are characterized by chartaceous leaves, often immature at the time of flowering, small and compact inflorescences born on short leafless spurs and small placentas with 3–7 pendulous ovules (Bridson, 2003b). This group includes, e.g., C. zygoon (Zygoon graveolens Hiern) and C. peteri. Because of lack of material, we only included C. peteri, which was unresolved in our analysis. Bridson’s group II comprises species characterized by the following characters: leaves coriaceous or subcoriaceous, mature at the time of flowering; terminal or axillary inflorescences on leafy branches, usually compact with short pedicels; small placentas with 2–7 pendulous ovules or large placentas with 2–3(–6) impressed ovules; and ruminate, partly ruminate or entire seeds (Bridson, 2003b). Three species of Bridson’s group II were included in our analysis but C. graveolens, C. supra-axillare and C. littorale did not form a monophyletic group. Coptosperma graveolens fell alone at the base of the entire Madagascan–western Indian Ocean– continental African clade in clade IV. Coptosperma supraaxillare fell at the base of the main Coptosperma clade, and, C. littorale, type of Enterospermum (Hiern, 1877), within the main Coptosperma clade. Bridson’s group III comprises only C. nigrescens. This species is characterized by coriaceous, very shiny leaves, mature at the time of flowering, lax inflorescences, positioned terminally on leafy branches, long pedicels, large placentas with three impressed ovules and seeds with fully ruminate endosperm (Bridson, 2003b). In our analysis, C. nigrescens grouped with C. madagascariense, which fits well the characters of Bridson’s group III. However, these two species also grouped with C. sp. D and C. sp. E, which possess flowers with short pedicels and small placentas with three pendulous ovules (fitting in Bridson’s group II). Our analysis indicated that neither Bridson’s group II nor her group III are monophyletic. We cannot comment on Bridson’s group I since we only included a single species in our analysis. It is clear that further study of the continental African Coptosperma species is necessary to ascertain their taxonomic position and the relationships between them. Within the Madagascan representatives, C. madagascariense, type of the name of the historical genus Santalina Baill. fell within the main Coptosperma clade. This corroborates that Santalina is congeneric with Coptosperma as was already suggested based on morphological characters (Homolle, 1938; Capuron, 1973; Bridson & Robbrecht, 1985; De Block & al., 2001; Bridson, 2003; De Block, 2007). Schizenterospermum was tentatively included in the tribe Pavetteae by Bridson & Robbrecht (1985) and this taxonomic decision is consistent with the results of the present study. The two Schizenterospermum species formed a well-supported monophyletic group nested within the main Coptosperma 90 Version of Record TAXON 64 (1) • February 2015: 79–95 De Block & al. • Molecular phylogenetics of Pavetteae (Rubiaceae) clade. Capuron previously suggested merging Schizenterospermum with Coptosperma (as Enterospermum) in his unpublished treatment of the Madagascan Rubiaceae (Capuron, 1973). Morphological data also support this transfer since the main characters of the two genera, such as fruit and seed morphology, stipule morphology, placentation type, etc., are identical. The main differences between the two genera are found in leaf shape and texture, but seem mostly related to ecological factors, e.g., smaller, glabrous, persistent, coriaceous leaves in Coptosperma versus larger, orbicular, pubescent, caducous, papyraceous/ subcoriaceous leaves in Schizenterospermum. However, with overall support within clade IV mostly weak and Coptosperma not retrieved as monophyletic, we prefer to postpone a decision on the synonymisation of Schizenterospermum until more resolution and better support have been achieved. Further study is necessary to reach conclusive insights into the delimitation of Coptosperma, and the relationships between its Madagascan and continental African species. Robbrechtia (clade IV). — Robbrechtia comprises two species, both of which were included in our analysis. They were strongly supported as a monophyletic group. However, the relationship between Robbrechtia and the other Madagascan genera of the Pavetteae remained obscure because of weak resolution and support. Robbrechtia is characterized by a striking character combination, unique in the Pavetteae: sheathing stipules with intrapetiolar lobes, young shoots and inflorescence parts covered with copious colleter exudate, long-tubed flowers with a well-developed calyx, a bilobed stigma, unilocular, one-seeded fruits, and seeds with deeply ruminate endosperm (De Block, 2003b). In Madagascar, fruits with a single ruminate seed also occur in Coptosperma and Schizenterospermum. Robbrechtia differs from these genera by the stipule type, the bilobed stigma and the much larger corollas, ovaries, calyces and fruits. In fact, the characters of Robbrechtia are reminiscent of those of Homolliella sericea, with which it shares its habitat but from which it differs by the unilocular fruits with a single ruminate seed, the bilobed stigma and the sheathing intrapetiolar stipules. Stipules interpetiolar, large, triangular or ovate, erect or bent back, sometimes with short needle-like awn. Inflorescences terminal or axillary, trichotomously branched, uniflorous to multiflorous, congested or lax, bracteose or more rarely frondose, sometimes with bracts and bracteoles fused into involucres, often all inflorescence parts including the corolla densely pubescent. Flowers hermaphroditic, actinomorphic, 4–7-merous, pedicellate or sessile. Secondary pollen presentation present or rarely absent. Calyx well developed; tube much shorter or rarely longer than lobes; lobes elliptic or foliaceous, up to 6 cm long, or, rarely narrowly triangular, up to 7 mm long. Corolla white or pale green; tube narrowly cylindrical, 1–17 cm long; lobes contorted to the left in bud, spreading at anthesis, 1–11 cm long. Stamens sessile or with short filaments, inserted somewhat below or at the level of the throat, entirely or for most of their length included in the corolla tube, or rarely exserted and spreading at anthesis; anthers linear, sometimes multilocellate. Ovary bilocular, with numerous ovules per locule; placentation axillary; ovules born on massive, peltate placentas attached to the upper half of the septum. Style slender; stigma bilobed, exserted at anthesis or included at the level of the throat or deep in the corolla tube. Disc annular, fleshy, glabrous. Fruits drupaceous, well developed, globose, ellipsoid or oblong, bilocular with numerous small seeds per locule; calyx persistent. Seeds small, angular; hilar cavity round, deep, surrounded by thickened annulus; endosperm entire; exotesta cells thickened along the outer tangential wall, thickenings containing narrow channels. Pollen grains (3–)4-colporate, supratectal elements absent. Distribution: tropical continental Africa. Leptactina arborescens (Welw. ex Benth. & Hook.f.) De Block, comb. nov. ≡ Dictyandra arborescens Welw. ex Benth. & Hook.f., Gen. Pl. 2: 85. 1873 – Holotype: Angola, Golungo Alto, Welwitsch 2561 (BM!; isotypes: K!, P!). Leptactina congolana (Robbr.) De Block, comb. nov. ≡ Dictyandra congolana Robbr. in Pl. Syst. Evol. 145: 114. 1984 – Holoype: Congo, Nsah Plateau, Ngo, Makany 1768 (BR!; isotypes: BR!, P!). Based on the results presented in this study we formally propose four new combinations and amend generic descriptions for two genera. Three genera are placed into synonymy. Leptactina papalis (N.Hallé) De Block, comb. nov. ≡ Coleactina papalis N.Hallé in Aubréville & Leroy, Fl. Gabon 17: 83. 1970 – Holotype: Gabon, Moughimba, SW de Koulamotou, approximativement à mi-distance de Mbigou, Le Testu 8356 (P!). Leptactina Hook.f., Hooker’s Icon. Pl. 11: 73, t. 1092. 1871 – Type (designated by Robbrecht in Pl. Syst. Evol. 145: 106. 1984): Leptactina mannii Hook.f. = Dictyandra Welw. ex Benth. & Hook.f., Gen. Pl. 2: 85. 1873, syn. nov. – Type: Dictyandra arborescens Welw. ex Benth. & Hook.f. = Coleactina N.Hallé in Aubréville & Leroy, Fl. Gabon 17: 83. 1970, syn. nov. – Type: Coleactina papalis N.Hallé. Small trees, erect or scandent shrubs, monocaul dwarfs, prostrate or erect geofrutices. Leaves petiolate or subsessile; blades papyraceous to thinly coriaceous, domatia often present. Paracephaelis Baill. in Adansonia 12: 316. 1879 – Type: Paracephaelis tiliacea Baill. = Homolliella Arènes in Notul. Syst. (Paris) 16: 16. 1960, syn. nov. – Type: Homolliella sericea Arènes. Shrubs or trees, sometimes deciduous. Leaves petiolate; blades coriaceous or rarely papyraceous, both surfaces pubescent to densely pubescent; domatia often present as hairy tufts at the axils of secondary veins. Stipules interpetiolar, triangular or ovate, acute to acuminate at the tip. Inflorescences terminal, sessile, trichotomously branched, multi- or pauciflorous, congested or lax, bracteose; all inflorescence parts including NOMENCLATURAL CHANGES Version of Record 91 De Block & al. • Molecular phylogenetics of Pavetteae (Rubiaceae) TAXON 64 (1) • February 2015: 79–95 the corolla pubescent. Flowers hermaphroditic, actinomorphic, 5-merous, pedicellate or sessile. Secondary pollen presentation present. Calyx well developed or not; tube short and lobes as long as or longer than the tube, or tube much longer than lobes; lobes triangular or ovate. Corolla white; tube cylindrical or infundibuliform, ≤ 2 cm long; lobes contorted to the left in bud, spreading at anthesis. Stamens with short filaments inserted at the throat, exserted at anthesis; anthers linear. Ovary bilocular, with 4–8 ovules per locule; placentation axillary; ovules arranged along the margins of a peltate placenta attached to the upper half of the septum. Style slender; stigma lobes fused over most of their length, exserted at anthesis. Disc annular, fleshy, glabrous. Fruits drupaceous, globose or ellipsoid, bilocular with (1–)2–8 seeds per locule; calyx persistent. Seeds laterally flattened; hilar cavity linear, shallow; endosperm entire; exotesta cells thickened along the outer tangential wall, thickenings containing 2 narrow channels. Pollen grains 3-colporate, supratectal microgemmae present. Distribution: Madagascar. Scientific Research-Flanders (postdoctoral travel grant to first author and research project G.0250.05) and the Percy Sladen Memorial Fund. The Royal Swedish Academy of Sciences and the Swedish Research Council supported a major part of the molecular work (grant 621-20041059 to last author). Paracephaelis sericea (Arènes) De Block, comb. nov. ≡ Homolliella sericea Arènes in Notul. Syst. (Paris) 16: 17, fig. 4 (19–25). 1960 – Holotype: Madagascar, Province Toamasina, forêt d’Analamazaotra, Thouvenot in Perrier de la Bâthie 122 (P!; isotypes: BR!, P!). ACKNOWLEDGEMENTS We thank the herbarium curators of the G, K, MO, P, S, TAN, TEF, UPS and WAG herbaria for loans of specimens and for help extended during study visits; the K, MO, TAN, UPS and WAG herbaria for supplying silica gel material or allowing the use of specimen fragments for DNA study; Aaron Davis (K), Quentin Luke (Fairchild Tropical Botanic Garden and East African Plant Redlist Authority), Steven Dessein (BR), Geoffrey Mwachala (EA), Frank Van Caekenberghe (BR), Inge Groeninckx (KUL), Dominique Champluvier (BR), Ferdinand Niyongabo (BR), Jean Pierre Vande Weghe (Wildlife Conservation Society, Gabon), Salvator Ntore (BR), Olivier Lachenaud (BR), James Tosh (KUL), Arnaud Mouly (University of Franche-Comté), Jan Wieringa (WAG) and Katarina Andreasen (UPS) for collecting leaf material in silica gel; Anbar Khodanbadeh (Bergius Foundation) and Wim Baert (BR) for help with the lab work and the sequencing; Anja Vandeperre (KUL), Koen Geuten (KUL) and Erik Smets (KUL) for the first author’s initiation in molecular techniques; Steven Dessein and Frank Van Caekenberghe (BR) for the use of photographs; Liliane Tytens (BR) for making the photographic plate; Pete Lowry (MBG) for allowing field work in Madagascar to take place within the framework of the Madagascar Research and Conservation Program of Missouri Botanical Garden; the MBG office staff in Antananarivo for logistic support; Madagascar National Parks, Ministère des Eaux et Forêts and Parc Botanique et Zoologique de Tsimbazaza (PBZT) for permission to collect in protected areas in Madagascar; Lalao Andriamahefarivo (MBG) and Franck Rakotonasolo (TAN) for arranging collecting and export permits; Franck Rakotonasolo (TAN), Solo Rapanarivo (TAN) and Tiana Randriamboavonjy (K) for help during field work. This study was supported by the Fund for 92 LITERATURE CITED Ali, S.J. & Robbrecht, E. 1991. Remarks on the tropical Asian and Australian taxa included in Diplospora or Tricalysia (RubiaceaeIxoroideae-Gardenieae). Blumea 35: 279–305. Andreasen, K. 1997. Phylogeny of the Ixoroideae: Molecular systematics within Rubiaceae. Dissertation, University of Uppsala, Sweden. Andreasen, K. & Bremer, B. 1996. Phylogeny of the subfamily Ixoroideae (Rubiaceae). Opera Bot. Belg. 7: 119–138. Andreasen, K. & Bremer, B. 2000. Combined phylogenetic analysis in the Rubiaceae-Ixoroideae: Morphology, nuclear and chloroplast DNA data. Amer. J. Bot. 87: 1731–1748. http://dx.doi.org/10.2307/2656750 Arènes, J. 1960. A propos de quelques genres Malgaches de Rubiacées. Notul. Syst. (Paris) 16: 6–19. Baillon, H.N. 1879. Observations sur les Naucléées. Adansonia, ser. 1, 12: 311–316. Bremekamp, C.E.B. 1934. A monograph of the genus Pavetta L. Repert. Spec. Nov. Regni Veg. 37: 1–208. Bremekamp, C.E.B. 1939a. A monograph of the genus Pavetta L.: Additions and emendations. Repert. Spec. Nov. Regni Veg. 47: 12–28. Bremekamp, C.E.B. 1939b. A monograph of the genus Pavetta L.: Additions and emendations II. Repert. Spec. Nov. Regni Veg. 47: 81–98. Bremekamp, C.E.B. 1940. Cladoceras subcapitatum (K.Schum. & K.Krause) Bremek. Hooker’s Icon. Pl. 35: t. 3411. Bremer, B. & Eriksson, T. 2009. Timetree of Rubiaceae: Phylogeny and dating the family, subfamilies and tribes. Int. J. Pl. Sci. 170: 766–793. http://dx.doi.org/10.1086/599077 Bridson, D.M. 1978a. A short revision of Rutidea (Rubiaceae). Kew Bull. 33: 243–278. http://dx.doi.org/10.2307/4109578 Bridson, D.M. 1978b. Nichallea, a new tropical African genus in the Rubiaceae. Kew Bull. 33: 287–293. http://dx.doi.org/10.2307/4109581 Bridson, D.M. 1979. Studies in Tarenna sensu lato (Rubiaceae subfamily Cinchonoideae) for part 2 of ‘Flora of tropical East Africa: Rubiaceae’. Kew Bull. 34: 377–401. http://dx.doi.org/10.2307/4110002 Bridson, D.M. 1988. Tarenna. Pp. 584–603 in: Polhill, R.M. (ed.), Flora of tropical East Africa: Rubiaceae, pt. 2. Rotterdam: A.A. Balkema. Bridson, D.M. 2001. Additional notes on Pavetta (Rubiaceae: Pavetteae) from tropical eastern and southern Africa. Kew Bull. 56: 567–600. http://dx.doi.org/10.2307/4117685 Bridson, D.M. 2003a. Tarenna. Pp. 599–608. in: Pope, G.V. (ed.), Flora Zambesiaca, vol. 5(3). Richmond: Royal Botanic Gardens, Kew. Bridson, D.M. 2003b. Coptosperma. Pp. 608–619. in: Pope, G.V. (ed.), Flora Zambesiaca, vol. 5(3). Richmond: Royal Botanic Gardens, Kew. Bridson, D.M. & Robbrecht, E. 1985. Further notes on the tribe Pavetteae (Rubiaceae). Bull. Jard. Bot. Natl. Belg. 55: 83–115. http://dx.doi.org/10.2307/3668012 Candolle, A.P. de 1807. Mémoire sur le Cuviera: Genre nouveau de la famille des Rubiacées. Ann. Mus. Hist. Nat. 9: 216–222. Capuron, R. 1973. “Révision des Rubiacées de Madagascar et des Comores”, [handwritten addition:] Notes regroupées et mises en forme par J. Bosser, dactylographiées de F. Chauvet. Unpublished manuscript, Muséum national d’Histoire naturelle, Laboratoire de Phanérogamie, Paris. Cortes-B., R., Delprete, P.G. & Motley, T.J. 2009. Phylogenetic Version of Record TAXON 64 (1) • February 2015: 79–95 De Block & al. • Molecular phylogenetics of Pavetteae (Rubiaceae) placement of the tribe Retiniphylleae among the subfamily Ixoroideae (Rubiaceae). Ann. Missouri Bot. Gard. 96: 61–67. http://dx.doi.org/10.3417/2006198 De Block, P. 1995. Ovary, seed and fruit of Rutidea (RubiaceaePavetteae). Pl. Syst. Evol. 196: 1–17. http://dx.doi.org/10.1007/BF00985332 De Block, P. 1997. Biosystematic studies in the tribe Pavetteae (Rubiaceae – Pavetteae). Dissertation, University of Antwerp, Belgium. De Block, P. 1998. The African species of Ixora (Rubiaceae – Pavetteae). Opera Bot. Belg. 9: 1–218. De Block, P. 2003a. New combinations in the genus Paracephaelis (Pavetteae, Rubiaceae). Syst. Geogr. Pl. 73: 99–100. De Block, P. 2003b. Robbrechtia, a new Rubiaceae genus from Madagascar. Syst. Bot. 28: 145–156. De Block, P. 2005. Notes on Tarenna as a precursor to a checklist of the Rubiaceae of Madagascar and the Comores. Syst. Geogr. Pl. 75: 107–116. De Block, P. 2007. Notes on Coptosperma as a precursor to a checklist of the Rubiaceae of Madagascar and the Comores. Syst. Geogr. Pl. 77: 122–127. De Block, P. & Robbrecht, E. 1997. On the ovary structure of Dictyandra and Leptactina compared with other Pavetteae (RubiaceaeIxoroideae), or evolution from multiovulate to uniovulate placentas. Bot. Jahrb. Syst. 119: 99–113. De Block, P. & Robbrecht, E. 1998. Pollen morphology of the Pavetteae (Rubiaceae, Ixoroideae) and its taxonomic significance. Grana 37: 260–275. http://dx.doi.org/10.1080/00173139809362678 De Block, P., Degreef, J. & Robbrecht, E. 2001. Reinstatement of the Afro-Madagascan genus Coptosperma (Rubiaceae). Syst. Geogr. Pl. 71: 455–492. http://dx.doi.org/10.2307/3668694 Degreef, J. 2006. Revision of continental African Tarenna (RubiaceaePavetteae). Opera Bot. Belg. 14: 1–150. Degreef, J., De Block, P. & Robbrecht, E. 2001. A survey of continental African Coptosperma (Rubiaceae). Syst. Geogr. Pl. 71: 367–382. http://dx.doi.org/10.2307/3668685 Doyle, J.J. & Doyle, J.L. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochem. Bull. Bot. Soc. Amer. 19, 11–15. Dumortier, B.C.J. 1829. Analyse des familles de plantes, avec l’indication des principaux genres qui s’y rattachent. Tournay: imprimerie de J. Casterman. http://dx.doi.org/10.5962/bhl.title.48702 Gaertner, J. 1788. De fructibus et seminibus plantarum, vol. 1. Stutgardiae [Stuttgart]: typis Academiae Carolinae. http://dx.doi.org/10.5962/bhl.title.53838 Hallé, N. 1970. Famille des Rubiacées (2ième partie). Pp. 1–335 in: Aubréville, A. & Leroy, J.-F. (eds.), Flore du Gabon, vol. 17. Paris: Muséum national d’Histoire naturelle. Hiern, W.P. 1877. Rubiaceae. Pp. 33–247 in: Oliver, D. (ed.), Flora of tropical Africa, vol. 3. London: Reeve. http://dx.doi.org/10.5962/bhl.title.42 Homolle, A.M. 1938. Les genres Tarenna, Enterospermum, Santalina (Rubiacées) à Madagascar. Bull. Soc. Bot. France 85: 605–609. http://dx.doi.org/10.1080/00378941.1938.10834138 Hooker, J.D. 1871. Leptactina. Hooker’s Icon. Pl. 11: t. 1092. http://dx.doi.org/10.5962/bhl.title.16059 Hooker, J.D. 1873. Ordo LXXXIV. Rubiaceae. Pp. 7–151 in: Bentham, G. & Hooker, J.D. (eds.), Genera plantarum, vol. 2(1). London: Reeve. http://dx.doi.org/10.5962/bhl.title.747 Huelsenbeck, J.P. & Ronquist, F. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17: 754–755. http://dx.doi.org/10.1093/bioinformatics/17.8.754 Jérémie, J. 1974. A propos du genre Tarenna (Rubiaceae-Gardenieae) en Nouvelle-Calédonie. Adansonia, ser. 2, 14: 473–480. Kainulainen, K., Razafimandimbison, S.G. & Bremer, B. 2013. Phylogenetic relationships and new tribal delimitations in subfamily Ixoroideae (Rubiaceae). Bot. J. Linn. Soc. 173: 387–406. http://dx.doi.org/10.1111/boj.12038 Katoh, K., Misawa, K., Kuma, K. & Miyata, T. 2002. MAFFT: A novel method for rapid multiple sequence alignment based on Fourier transform. Nucl. Acids Res. 30: 3059–3066. http://dx.doi.org/10.1093/nar/gkf436 Lemaire, B., Van Oevelen, S., De Block, P., Verstraete, B., Smets, E., Prinsen, E. & Dessein, S. 2012. Identification of the bacterial endosymbionts in leaf nodules of Pavetta (Rubiaceae). Int. J. Syst. Evol. Microbiol. 62: 202–209. http://dx.doi.org/10.1099/ijs.0.028019-0 Lidén, M., Fukuhara, T. & Axberg, T. 1995. Phylogeny of Corydalis, ITS and morphology. Pl. Syst. Evol., Suppl. 9: 183–188. Linneaus, C. 1753. Species plantarum. Holmiae [Stockholm]: impensis Laurentii Salvii. http://dx.doi.org/10.5962/bhl.title.669 Miller, I.M. 1990. Bacterial leaf nodule symbiosis. Advances Bot. Res. 17: 163–243. http://dx.doi.org/10.1016/S0065-2296(08)60134-2 Mouly, A., Kainulainen, K., Persson, C., Davis, A.P., Wong, K.M., Razafimandimbison, S.G. & Bremer, B. 2014. Phylogenetic structure and clade circumscriptions in the Gardenieae complex (Rubiaceae). Taxon 63: 801–818. http://dx.doi.org/10.12705/634.4 Neuba, D.F.R. 2006. Révision systématique du genre Leptactina (Rubiaceae, Pavetteae). Dissertation, Université Libre de Bruxelles, Belgium. Neuba, D.F.R., Lejoly, J. & Sonké, B. 2006. Un Leptactina (Rubiaceae, Pavetteae) nouveau de République Centrafricaine: Leptactina deblockiae. Adansonia, sér. 3, 28: 373–378. Oxelman, B., Lidén, M. & Berglund, D. 1997. Chloroplast rps16 intron phylogeny of the tribe Sileneae (Carophyllaceae). Pl. Syst. Evol. 206: 393–410. http://dx.doi.org/10.1007/BF00987959 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. http://dx.doi.org/10.1006/mpev.2001.0977 Posada, D. 2008. jModelTest: Phylogenetic model averaging. Molec. Biol. Evol. 2000: 205–224. http://dx.doi.org/10.1093/molbev/msn083 Puff, C., Robbrecht, E., Buchner, R. & De Block, P. 1996. Survey of secondary pollen presentation in the Rubiaceae. Opera Bot. Belg. 7: 369–402. Rambaut, A. & Drummond, A.J. 2007. Tracer, version 1.4. http://beast.bio.ed.ad.uk/Tracer Razafimandimbison, S.G. & Bremer, B. 2002. Phylogeny and classification of Naucleeae (Rubiaceae) inferred from molecular (nrITS, rbcL, and trnT-F) and morphological data. Amer. J. Bot. 89: 1027– 1041. http://dx.doi.org/10.3732/ajb.89.7.1027 Reynolds, S.T. & Forster, P.I. 2005. A taxonomic revision of Tarenna Gaertn. and Triflorensis S.T.Reynolds (Rubiaceae; Ixoroideae: Pavetteae) in Australia. Austrobaileya 7: 29–55. Richard, A. 1829. Rubiacées. Pp. 3–5 in: Bory de Saint-Vincent, J.B. (ed.), Dictionnaire classique d’histoire naturelle, vol. 15. Paris: Rey et Gravier. http://dx.doi.org/10.5962/bhl.title.33901 Richard, A. 1830. Mémoire sur la famille des Rubiacées. Paris: imprimerie de J. Tastu. http://bibdigital.rjb.csic.es/ing/Libro.php?Libro=5906 Robbrecht, E. 1981. Studies in tropical African Rubiaceae II. Bull. Jard. Bot. Natl. Belg. 51: 359–378. http://dx.doi.org/10.2307/3668069 Robbrecht, E. 1984. The delimitation and taxonomic position of the tropical African genera Leptactina and Dictyandra (Rubiaceae). Pl. Syst. Evol. 145: 105–118. http://dx.doi.org/10.1007/BF00984034 Robbrecht, E. 1988. Tropical woody Rubiaceae: Characteristic features and progressions; Contributions to a new subfamilial classification. Opera Bot Belg. 1: 1–271. Robbrecht, E. 1994. Supplement to the 1988 outline of the classification of the Rubiaceae. Opera Bot. Belg. 6: 173–196. Robbrecht, E. & Bridson, D.M. 1984. The taxonomic position of the East African genus Cladoceras (Rubiaceae). Bull. Soc. Roy. Bot. Belgique 117: 247–251. Robbrecht, E. & Manen, J.F. 2006. The major evolutionary lineages Version of Record 93 De Block & al. • Molecular phylogenetics of Pavetteae (Rubiaceae) TAXON 64 (1) • February 2015: 79–95 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. 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. http://dx.doi.org/10.1016/j.ympev.2008.04.006 Saghai-Maroof, K., Soliman, M., Jorgensen, R.A. & Allard, R.W. 1984. Ribosomal DNA spacer length polymorphism in barley: Mendelian inheritance, chromosomal location, and population dynamics. Proc. Natl. Acad. Sci. U.S.A. 81: 8014–8018. http://dx.doi.org/10.1073/pnas.81.24.8014 Schumann, K. 1889. Pachystylus nov. gen. Pp. 133–134 in: Schumann, K. & Hollrung, M., Die Flora von Kaiser Wilhelms Land. Berlin: Asher. Simmons, M.P. & Ochoterena, H. 2000. Gaps as characters in sequence-based phylogenetic analyses. Syst. Biol. 49: 369–381. http://dx.doi.org/10.1093/sysbio/49.2.369 Smith, A.C. & Darwin, S.P. 1988. Family 168. Rubiaceae. Pp. 143–362 in: Smith, A.C. (ed.), Flora vitiensis nova: A new flora of Fiji (Spermatophytes only), vol. 4. Lawaii: Pacific Tropical Botanical Garden. Stamatakis, A., Hoover, P. & Rougement, J. 2008. A fast bootstrapping algorithm for the RaxML web servers. Syst. Biol. 57: 758–771. http://dx.doi.org/10.1080/10635150802429642 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. http://dx.doi.org/10.1007/BF00037152 Valeton, T. 1926. Die Rubiaceae von Papuasien. Erster Teil: Cinchonoideae. Bot. Jahrb. Syst. 60: 1–104. Verdcourt, B. 1981. Notes on African Rubiaceae. Kew Bull. 36: 493–557. http://dx.doi.org/10.2307/4117584 Verdcourt, B. 1988a. Tennantia. Pp. 582–584 in: Polhill, R.M. (ed.), Flora of tropical East Africa: Rubiaceae, pt. 2. Rotterdam: A.A. Balkema. Verdcourt, B. 1988b. Cladoceras. Pp. 603–605 in: Polhill, R.M. (ed.), Flora of tropical East Africa: Rubiaceae, pt. 2. Rotterdam: A.A. Balkema. Wang, W., Lu, A.M., Ren Y., Endress, M.E. & Chen, Z.D. 2009. Phylogeny and classification of Ranunculales: Evidence from four molecular loci and morphological data. Perspect. Pl. Ecol. Evol. Syst. 11: 81–110. http://dx.doi.org/10.1016/j.ppees.2009.01.001 Appendix 1. List of taxa used in the phylogenetic analyses with voucher information (geographic origin, collection, herbarium), EMBL accession numbers and literature citations for previously published sequences for the two plastid markers, rps16 and trnT-trnF, and the nuclear marker ITS. FTEA: Flora of tropical East Africa; (1) = Cortes-B. & al., 2009; (2) = Rydin & al., 2008; (3) = Bremer & Eriksson, 2009; (4) = Kainulainen & al., 2013. New sequences are marked with *. Tribe Alberteae Hook.f.: Razafimandimbisonia Kainul. & B.Bremer: R. humblotii (Drake) Kainul. & B.Bremer, Madagascar, Tosh & al. 263 (BR), KM592238*, KM592145*, KM592324*. Tribe Coffeeae DC.: Tricalysia A.Rich. ex DC.: T. semidecidua Bridson, Zambia, Dessein & al. 1093 (BR), KM592279*, KM592185*, –. Tribe Ixoreae A.Gray: Ixora L.: I. sp., Thailand, Sudde 1487 (K), KM592208*, KM592115*, –. Tribe Gardenieae DC.: Euclinia Salisb.: E. longiflora Salisb., Africa (country unknown), Van Caekenberghe 348 (BR), KM592203*, KM592110*, –. Gardenia J.Ellis: G. rutenbergiana (Baill. ex Vatke) J.-F.Leroy, Madagascar, Groeninckx & al. 24 (BR), KM592204*, KM592111*, KM592295*. Oxyanthus DC.: O. troupinii Bridson, Burundi, Niyongabo 115 (BR), KM592219*, KM592126*, –. Tribe Mussaendeae Hook.f.: Mussaenda Wernham: M. flava Verdc., Africa (country unknown), Van Caekenberghe 60 (BR), KM592217*, KM592124*, KM592306*. Tribe Pavetteae A.Rich. ex Dumort.: Cladoceras Bremek.: C. subcapitatum (K.Schum. & K.Krause) Bremek., Tanzania, Luke & al. 8351 (UPS), AM117290(3), KM592094*, KM592281*. Coleactina N.Hallé: C. papalis N.Hallé, Gabon, Dessein & al. 2355 (BR), KM592188*, KM592095*, KM592282*. Coptosperma Hook.f.: C. borbonicum (Hend. & Andr.Hend.) De Block, Comores, De Block 1389 (BR), KM592189*, KM592096*, KM592283*; C. graveolens (S.Moore) Degreef, Kenya, Mwachala 3711 (BR), KM592200*, KM592107*, KM592293*; C. littorale (Hiern) Degreef, Mozambique, Luke & al. 9954 (UPS), KM592190*, KM592097*, KM592284*; C. madagascariense (Baill.) De Block, Madagascar, Razafimandimbison & al. 577 (UPS), KM592191*, KM592098*, –; C. madagascariense (Baill.) De Block, Madagascar, De Block & al. 2238 (BR), –, –, KM592285*; C. nigrescens Hook.f., Madagascar, De Block & al. 535 (BR), KM592192*, KM592099*, KM592286*; C. nigrescens Hook.f., Kenya, Luke & Luke 9030 (UPS), KM592193*, KM592100*, KM592287*; C. peteri (Bridson) Degreef, Tanzania, Lovett & Congdon 2991 (BR), KM592201*, KM592108*, KM592294*; C. supra-axillare (Hemsl.) Degreef, Madagascar, De Block & al. 1321 (BR), KM592194*, KM592101*, KM592288*; C. sp. nov. A, Madagascar, De Block & al. 720 (BR), KM592199*, KM592106*, KM592292*; C. sp. nov. B, Madagascar, De Block & al. 796 (BR), KM592195*, KM592102*, KM592289*; C. sp. nov. C, Madagascar, De Block & al. 1355 (BR), KM592196*, KM592103*, KM592290*; C. sp. nov. D, Madagascar, De Block & al. 704 (BR), KM592197*, KM592104*, KM592291*; C. sp. nov. E, Madagascar, De Block & al. 733 (BR), KM592198*, KM592105*, –. Dictyandra Welw. ex Hook.f.: D. arborescens Welw. ex Hook.f., Ghana, Schmidt & al. 1683 (MO), KM592202*, KM592109*, –. Homollea Arènes: H. longiflora Arènes, Madagascar, De Block & al. 767 (BR), KM592205*, KM592112*, KM592296*; H. perrieri Arènes, Madagascar, Morat 4700 (TAN), KM592206*, KM592113*, KM592297*. Homolliella Arènes: H. sericea Arènes, Madagascar, De Block & al. 849 (BR), KM592207*, KM592114*, KM592298*. Leptactina Hook.f.: L. benguelensis (Welw. ex Benth. & Hook.f.) R.D.Good, Zambia, Dessein & al. 1142 (BR), KM592209*, KM592116*, KM592299*; L. delagoensis K.Schum., Tanzania, Luke & Kibure 9744 (UPS), KM592210*, KM592117*, KM592300*; L. epinyctios Bullock ex Verdc., Zambia, Dessein & al. 1348 (BR), KM592211*, KM592118*, KM592301*; L. involucrata Hook.f., Cameroon, Davis 3028 (K), KM592212*, KM592119*, KM592302*; L. leopoldi-secundi Büttner, Republic of Congo, Champluvier 5248 (BR), KM592213*, KM592120*, –; L. mannii Hook.f., Gabon, Dessein & al. 2518 (BR), KM592214*, KM592121*, KM592302*; L. papyrophloea Verdc., Tanzania, Luke & Kibure 9838 (UPS), KM592215*, KM592122*, KM592304*; L. pynaertii De Wild., Republic of the Congo, Champluvier s.n. (BR), KM592216*, KM592123*, KM592305*. Nichallea Bridson: N. soyauxii (Hiern) Bridson, Cameroon, Dessein & al. 1402 (BR), KM592218*, KM592125*, KM592307*. Paracephaelis Baill.: P. cinerea (A.Rich. ex DC.) De Block, Madagascar, De Block & al. 2193 (BR), KM592220*, KM592127*, KM592308*; P. saxatilis (Scott-Elliot) De Block, Madagascar, Davis & al. 2731 (K), KM592221*, KM592128*, –; P. saxatilis (Scott-Elliot) De Block, Madagascar, De Block & al. 2401 (BR), –, –, KM592309*; P. tiliacea Baill., Madagascar, Groeninckx & al. 113 (BR), KM592222*, KM592129*, KM592310*. Pavetta L.: subg. Baconia Bremek.: P. hymenophylla Bremek., Tanzania, Luke & al. 9101 (UPS), KM592225*, KM592132*, –; P. ternifolia Hiern, Burundi, Ntore 19 (BR), KM592235*, KM592142*, KM592321*; subg. Dizygoon Bremek.: P. tetramera (Hiern) Bremek, Gabon, Van de Weghe 163 (BR), KM592236*, KM592143*, KM592322*; P. suffruticosa K.Schum., Cameroon, Lachenaud & al. 838 (BR), KM592231*, KM592138*, –; subg. Pavetta: P. agrostiphylla Bremek., Sri Lanka, Bremer B. & K. 936 (UPS), KM592223*, KM592130*, KM592311*; P. batesiana Bremek., Gabon, Dessein & al. 2071 (BR), KM592224*, KM592131*, KM592312*; P. indica L., Sri Lanka, Andreasen 202 (UPS), HM164217(4), HM164331(4), KM592313*; P. sansibarica K.Schum., Kenya, Luke & al. 8326 (UPS), KM592227*, KM592134*, KM592314*; P. schumanniana F.Hoffm. ex K.Schum., Zambia, Dessein & al. 911 (BR), KM592228*, KM592135*, KM592315*; P. stenosepala K.Schum., Kenya, Luke & al. 8318 (UPS), KM592233*, KM592140*, KM592319*; P. tarennoides S.Moore, Kenya, Luke & al. 8325 (UPS), KM592234*, KM592141*, KM592320*; P. sp. A of FTEA Bridson, Tanzania, Luke & al. 9134 (UPS), KM592232*, KM592139*, KM592318*; P. sp. B, Vietnam, Davis & al. 4082 (K), KM592229*, KM592136*, KM592316*; P. sp. C, Asia (country unknown), Van Caekenberghe 199 (BR), KM592230*, KM592137*, KM592317*; P. vaga 94 Version of Record TAXON 64 (1) • February 2015: 79–95 De Block & al. • Molecular phylogenetics of Pavetteae (Rubiaceae) Appendix 1. Continued. S.T.Reynolds, Australia, Harwood 1290 (DNA), KM592237*, KM592144*, KM592323*. Robbrechtia De Block: R. grandifolia De Block, Madagascar, Kårehed 311 (UPS), AM117339(3), AM117383(3), KM592325*; R. milleri De Block, Madagascar, Bremer & al. 5295 (S), KM592240*, KM592147*, KM592326*. Rutidea DC.: sect. Rutidea: R. decorticata Hiern, Cameroon, Maurin 14 (K), KM592241*, KM592148*, –; R. dupuisii De Wild., Gabon, Dessein & al. 1802 (BR), KM592242*, KM592149*, –; R. ferruginea Hiern, Cameroon, Dessein & al. 1674 (BR), KM592242*, KM592150*, KM592327*; R. olenotricha Hiern, Ghana, Schmidt & al. 1731 (MO), KM592246*, KM592153*, KM592329*; R. orientalis Bridson, Tanzania, Borhidi & al. 84008 (UPS), KM592247*, –, KM592330*; R. parviflora DC., Liberia, Adam 20156 (UPS), KM592248*, KM592154*, KM592331*; R. seretii De Wild., Cameroon, Gereau 5588 (UPS), KM592249*, KM592155*, KM592332*; sect. Tetramera Bridson: R. fuscenscens Hiern, Tanzania, Luke & al. 9124 (UPS), KM592244*, KM592151*, KM592328*; R. membranacea Hiern, Liberia, Adam 21433 (UPS), KM592245*, KM592152*, –. Schizenterospermum Homolle ex Arènes: S. grevei Homolle ex Arènes, Madagascar, De Block & al. 2167 (BR), KM592250*, KM592156*, KM592333*; S. rotundifolia Homolle ex Arènes, Madagascar, De Block & al. 771 (BR), KM592251*, KM592157*, KM592334*. Tarenna Gaertn.: T. alleizettei (Dubard & Dop) De Block, Madagascar, De Block & al. 1883 (BR), KM592272*, KM592178*, KM592353*; T. alpestris (Wight) N.P.Balakr., India, De Block 1474 (BR), KM592252*, KM592158*, KM592335*; T. asiatica (L.) Kuntze ex K.Schum., India, Auroville 998 (SBT), KM592253*, KM592159*, KM592336*; T. attenuata (Hook.f.) Hutch., Asia, country unknown, BR Living Collection 20031135-53 (BR), KM592254*, KM592160*, KM592337*; T. bipindensis (K.Schum.) Bremek., Liberia, Jongkind 8495 (BR), KM592255*, KM592161*, KM592338*; T. capuroniana De Block, Madagascar, De Block & al. 937 (BR), KM592273*, KM592179*, KM592354*; T. depauperata Hutch., China, Chow & Wan 79063 (UPS), KM592256*, KM592162*, KM592339*; T. flava Alston, Sri Lanka, Klackenberg 440 (S), KM592257*, KM592163*, KM592340*; T. fuscoflava (K.Schum.) S.Moore, Ghana, Schmidt & al. 2099 (MO), KM592258*, KM592164*, KM592341*; T. gracilipes (Hayata) Ohwi, Japan, Van Caekenberghe 149 (BR), KM592259*, KM592165*, –; T. grevei (Drake) Homolle, Madagascar, De Block & al. 959 (BR), KM592274*, KM592180*, KM592355*; T. jolinonii N.Hallé, Gabon, Champluvier 6098 (BR), KM592260*, KM592166*, KM592342*; T. lasiorachis (K.Schum. & K.Krause) Bremek., Gabon, Wieringa 4432 (WAG), KM592261*, KM592167*, –; T. leioloba (Guillaumin) S.Moore, New Caledonia, Mouly 174 (P), KM592262*, KM592168*, KM592343*; T. microcarpa (Guillaumin) Jérémie, New Caledonia, Mouly 297 (P), KM592263*, KM592169*, KM592344*; T. nitidula (Benth.) Hiern, Liberia, Jongkind 8000 (BR), KM592264*, KM592170*, KM592345*; T. pallidula Hiern, Gabon, Dessein & al. 2215 (BR), KM592265*, KM592171*, KM592346*; T. pembensis J.E.Burrows, Mozambique, Luke & al. 10136 (UPS), KM592266*, KM592172*, KM592347*; T. precidantenna N.Hallé, Gabon, Dessein & al. 2360 (BR), KM592267*, KM592173*, KM592348*; T. rhypalostigma (Schltr.) Bremek., New Caledonia, Mouly 182 (P), KM592268*, KM592174*, KM592349*; T. roseicosta Bridson, Tanzania, Luke & al. 9170 (UPS), KM592269*, KM592175*, KM592350*; T. sambucina (G.Forst.) T.Durand ex Drake, New Caledonia, Mouly & al. 364 (P), KM592270*, KM592176*, KM592351*; T. spiranthera (Drake) Homolle, Madagascar, De Block & al. 946 (BR), KM592275*, KM592181*, KM592356*; T. thouarsiana (Drake) Homolle, Madagascar, De Block & al. 655 (BR), KM592276*, KM592182*, KM592357*; T. uniflora (Drake) Homolle, Madagascar, Bremer & al. 5230 (S), KM592277*, KM592183*, KM592358*; T. vignei Hutch. & Dalziel, Republic of Guinea, Jongkind 8126 (BR), KM592271*, KM592177*, KM592352*. Tennantia Verdc.: T. sennii (Chiov.) Verdc. & Bridson, Kenya, Luke & al. 8357 (UPS), KM592278*, KM592184*, KM592359*. Tribe Vanguerieae Dumort.: Vangueria Juss.: V. madagascariensis J.F.Gmel., Africa (country unknown), Bremer 3077 (UPS), –, AJ620184(2), –; V. madagascariensis J.F.Gmel., Africa (country unknown), Delprete 7383 (NY), EU821636(1), –, –. Version of Record 95