TAXON 59 (1) • February 2010: 125–133 Yuan & al. • Phylogeny of Clerodendrum and allied genera Further disintegration and redefinition of Clerodendrum (Lamiaceae): Implications for the understanding of the evolution of an intriguing breeding strategy Yao-Wu Yuan,1,2 David J. Mabberley, 3,4 Dorothy A. Steane5 & Richard G. Olmstead1 1 Department of Biology, University of Washington, Box 355325, Seattle, Washington 98195-5325, U.S.A. 2 Current address: 4504 Miller Plant Sciences Building, Department of Plant Biology, University of Georgia, Athens, Georgia 30602, U.S.A. 3 University of Washington Botanic Gardens, College of Forest Resources, Box 354115, Seattle, Washington 98195-4115, U.S.A. 4 Current address: Royal Botanic Gardens, Kew, Richmond, Surrey TW9 3AB, U.K. 5 School of Plant Science, University of Tasmania, Private Bag 55, Hobart, Tasmania 7001, Australia Author for correspondence: Yao-Wu Yuan, yyuan@plantbio.uga.edu Abstract The genus Clerodendrum s.l. is polyphyletic. Although recent studies have resulted in C. subg. Cyclonema and C. sect. Konocalyx being removed to the resurrected genus Rotheca, and the unispecific genus Huxleya being sunk into Clerodendrum, it has been unclear whether Clerodendrum as currently circumscribed is monophyletic, particularly in relation to the American genera Aegiphila, Amasonia, and Tetraclea. This phylogenetic study employs four relatively fast-evolving chloroplast DNA regions, trnT-L, trnL-F, trnD-T, and trnS-fM, to clarify the generic boundaries of Clerodendrum and its relationship to allied genera. The results corroborate previous studies that there are three well-supported clades in the currently recognized Clerodendrum: an Asian clade, an African clade, and a Pantropical Coastal clade. The Asian clade and African clade are sister groups and together form a monophyletic group. However, the Pantropical Coastal clade is more closely related to the three American genera than it is to the other two Clerodendrum clades. In addition, a Caribbean species, C. spinosum, is found to be more closely related to the American genera than it is to any of the three major Clerodendrum groups. These results indicate that Clerodendrum as currently circumscribed is not monophyletic. We propose to separate the Pantropical Coastal clade and C. spinosum by reviving the genera Volkameria (including Huxleya) and Ovieda, respectively for these, and to restrict Clerodendrum to the Asian and African clades. Brief descriptions of the genera to be recognized are provided. All Neotropical ‘Clerodendrum’ taxa are referred to other genera, necessitating six new combinations, which are also provided, where required, for two other well-studied Old World Volkameria species; all names ever used in Ovieda are given their modern placings (two placed newly in synonymy). The study also sheds light on the evolution of an intriguing breeding strategy that avoids self-pollination or/and sexual interference. This strategy involves presentation of pollen and stigma in the centre of the flower in a sequential fashion by moving the filaments and style. It appears to have evolved in the common ancestor of Clerodendrum, Volkameria, Ovieda, Amasonia, Tetraclea, Aegiphila and Kalaharia, and still occurs in all of these taxa except Aegiphila, where it has been succeeded by a heterostylous system. Keywords Aegiphila; breeding strategy; chloroplast DNA; Clerodendrum; Huxleya; Ovieda; phylogeny; Volkameria INTRODUCTION The genus Clerodendrum L. as delimited by nineteenthcentury botanists (Schauer, 1847; Briquet, 1895) is heterogeneous. However, this delimitation has been followed reasonably closely by subsequent authors, even though they recognized it to be problematic (Lam, 1919; Thomas, 1936; Moldenke, 1985; Verdcourt, 1992). Cladistic analyses of morphological data (Cantino, 1992; Rimpler & al., 1992) provided preliminary evidence that Clerodendrum sensu lato (s.l.) was not monophyletic. But these analyses, primarily focused at subfamily or family level, included relatively few Clerodendrum s.l. species and did not provide good resolution of relationships within Clerodendrum s.l. or between the genus and other related ones. Phylogenetic studies based on chloroplast DNA (cpDNA) restriction site data (Steane & al., 1997) and nuclear ITS sequences (Steane & al., 1999), with extensive sampling within Clerodendrum s.l. and related genera, strongly suggested that Clerodendrum s.l. is polyphyletic. Subsequently, a number of species comprising the C. subg. Cyclonema (Hochst.) Gürke and C. sect. Konocalyx (Thomas) Verdc. were removed to the resurrected genus Rotheca Raf. (Steane & Mabberley, 1998). In addition, the molecular studies divided Clerodendrum (sensu Steane & Mabberley, 1998) into three major clades that are in general associated with geographic distribution: an Asian clade, an African clade, and a Pantropical Coastal clade (Fig. 1A). A more recent study (Steane & al., 2004) which included three New World genera, Aegiphila Jacq., Amasonia L. f., and Tetraclea A.Gray, and a unispecific Australian genus Huxleya Ewart, put the delimitation of Clerodendrum (sensu Steane & Mabberley, 1998) into question again. Huxleya was found nested within the Pantropical Coastal clade and therefore was sunk into the genus Clerodendrum (Steane & al., 2004). The African and Asian clades were still recovered as sister groups and together formed a monophyletic group (Fig. 1B). The three New World genera, Aegiphila, Amasonia, and Tetraclea, each 125 TAXON 59 (1) • February 2010: 125–133 Yuan & al. • Phylogeny of Clerodendrum and allied genera represented by a single species in that study, formed a New World clade. However, the relationships among the New World clade, the Pantropical Coastal Clerodendrum, and the remaining Clerodendrum species (Asian + African), were unresolved (Steane & al., 2004; Fig. 1B), leaving the possibility that Clerodendrum (sensu Steane & al., 2004) as currently circumscribed is paraphyletic in relation to the clade of New World genera. In addition, the phylogenetic framework presented in those studies (Steane & al., 1997, 1999, 2004) is mainly based on cpDNA restriction site data and nuclear ITS sequences, with only a few chloroplast ndhF sequences. It is difficult to add more data to a restriction site dataset, due to the nature of this type of marker. For many Clerodendrum species the ITS region is difficult to sequence directly without cloning, possibly because of their being polyploids, which is indicated by the high chromosome number, 2n = 46, 48, or 52, of most species (see the Index to Plant Chromosome Numbers Database, http://mobot.mobot .org/W3T/Search/ipcn.html). The major objectives of this paper are, therefore, to: (1) test the monophyly of Clerodendrum as currently circumscribed; (2) present a phylogenetic framework of Clerodendrum and its related genera based on cpDNA sequence data, to which additional sequence data may be added easily in future studies; and (3) use this phylogenetic framework to examine the evolution of morphological characters. MATERIALS AND METHODS Our sampling included 40 species of Clerodendrum (sensu Steane & al., 2004), representing the three major clades Fig. 1. Summarized phylogenies from previous studies. A, results from Steane & al. (1997, 1999), showing that Clerodendrum s.l. is polyphyletic; B, results from Steane & al. (2004), showing the unresolved relationship between Asian + African Clerodendrum, Pantropical Coastal Clerodendrum, and a clade comprising Aegiphila, Amasonia, and Tetraclea. A identified in previous studies, 13 species from six closely related genera (Aegiphila, Amasonia, Tetraclea, Kalaharia Baill., Oxera Labill., Faradaya F. Muell.), and three species from more distantly related genera in the Lamiaceae-Ajugoideae (Ajuga L., Teucrium L., Rotheca). Voucher information for these 56 samples is listed in Appendix 1. Total genomic DNA was extracted from either silica-gel dried leaf tissue or herbarium specimens using the modified CTAB method (Doyle & Doyle, 1987). Four relatively fast-evolving non-coding cpDNA regions (Shaw & al., 2005) were chosen for sequencing. These were trnD-trnT, trnT-trnL, trnL-trnF (trnL intron and trnL-F intergenic spacer), and trnS-trnfM. PCR and sequencing primers with corresponding references are listed in Appendix 2. Procedures for PCR and sequencing are described in Yuan & Olmstead (2008). Sequences of the two outgroup species, Verbena officinalis L., Aloysia virgata (Ruiz & Pav.) Pers., are from Yuan & Olmstead (2008), while sequences of all other species were generated in this study and have been deposited in GenBank (trnD-trnT : EU160617–EU160666, FJ951910–FJ951915; trnStrnfM : FJ951916–FJ951970; trnT-trnL: FJ951971–FJ952025; trnL-trnF : FJ952026–FJ952081). Sequence alignments were made manually using Se-Al v.2.0a11 (Rambaut, 1996) based on the similarity criterion (Simmons, 2004). The four cpDNA regions were combined as a single dataset for phylogenetic analyses because these regions are part of the haploid chloroplast genome and, therefore, share the same evolutionary history. Phylogenetically informative insertions/deletions (indels) were coded as binary characters using the simple gap coding method (Graham & al., 2000; Simmons & Ochoterena, 2000) and appended to the end Clerodendrum Asian Clerodendrum African Clerodendrum Pantropical Coastal B Clerodendrum Asian Clerodendrum African Aegiphila Amasonia Tetraclea Oxera Faradaya Clerodendrum Pantropical Coastal Other genera Kalaharia 126 Clerodendrum subg. Cyclonema and C. sect. Konocalyx Oxera Outgroup Outgroup Faradaya TAXON 59 (1) • February 2010: 125–133 of the dataset. Six poly-nucleotide or microsatellite regions (a microsatellite with “AT” repeats and a poly-T region in the trnD-trnT segment, two poly-T regions in the trnT-trnL segment, a poly-A region in the trnL-trnF segment, and a polyC/T/G region in the trnfM-trnS segment) have been excluded from analyses due to uncertainty of homology assessment. Both parsimony and Bayesian analyses were performed on the final dataset. Parsimony analysis was conducted using PAUP* v.4.0b10 (Swofford, 2002). Heuristic searches were performed with 1000 random stepwise addition replicates and TBR branch swapping with the MULTREES option in effect. Nodal support was determined by bootstrap analyses (Felsenstein, 1985) of 500 replicates, each with 20 random stepwise addition replicates and TBR branch swapping with MULTREES on. Bayesian analyses were conducted using MrBayes v.3.1.2 (Ronquist & Huelsenbeck, 2003). A mixed-model approach (Ronquist & Huelsenbeck, 2003) was employed to integrate the phylogenetically informative gaps as binary characters with nucleotide data. The final dataset was divided into two partitions, the “nucleotide” partition and “gap” partition. We used Akaike information criterion (AIC) implemented in Modeltest v.3.7 (Posada & Crandall, 1998) to determine the model of sequence evolution that best fits the “nucleotide” partition (GTR + G). The restriction site (binary) model in MrBayes v.3.1.2 (Ronquist & Huelsenbeck, 2003) was used for the “gap” partition, with ascertainment bias for gap characters incorporated (lset coding = informative). We performed two independent runs of 1,000,000 generations from a random starting tree using the default priors and four Markov chains (one cold and three heated chains), sampling one tree every 100 generations. Plots of log likelihood scores were used to determine stationarity and trees from the first 100,000 generations were discarded as burn-in. Yuan & al. • Phylogeny of Clerodendrum and allied genera clades are strongly supported sister groups (100%/1.0, BB/ PP), whereas the Pantropical Coastal Clerodendrum clade is sister to the New World clade (97%/1.0, BS/PP). Also consistent with previous studies (Steane & al., 2004), Kalaharia, a unispecific African genus, is recovered as sister group of the larger clade that includes all three Clerodendrum groups and the New World clade. The Oxera/Faradaya clade is sister to the even more inclusive clade including Kalaharia (Fig. 2). Within the Asian clade, relationships are fairly well resolved. One strongly supported monophyletic group, in particular, is worth mentioning. It consists of species (C. floribundum, C. indicum [type of Siphonanthus L.], C. minahassae, C. quadriloculare, C. tomentosum) that are characterized by an extremely long and narrow corolla tube (99%/1.0, BS/PP; Fig. 2), probably an adaptation to a particular type of pollinator. Within the African clade, however, relationships are poorly resolved but it is noticeable that one species, C. hildebrandtii, is strongly supported as sister to the rest of the African group (Fig. 2). Within the Pantropical Coastal clade, branches are short, indicating little sequence diversification between species. Within the New World clade, both Aegiphila and Amasonia are strongly supported monophyletic groups (Fig. 2). Tetraclea is resolved to be the sister lineage of Amasonia, but this relationship is only weakly supported (52%/0.49, BS/PP, these values are not shown in Fig. 2). Likewise, Clerodendrum spinosum is recovered as sister to the Amasonia/Tetraclea clade, but weakly supported (52%/0.78, BS/PP). Bayesian analyses gave very similar results. The only difference between the Bayesian majority consensus tree and parsimony tree shown in Fig. 2 is on the relationship between Aegiphila anomala, A. alba, the A. hassleri + A. brachiata clade, and the A. elata + A. martinicensis clade, but neither the relationship suggested by parsimony analyses nor that indicated by Bayesian inference is well supported (BS < 50%, PP < 0.7). RESULTS DISCUSSION The final dataset consisted of 44 scored gap characters and 4002 aligned nucleotides, of which 127 from the six polynucleotide or microsatellite regions were excluded due to uncertainty of homology assessment. One of the eight maximum parsimony trees resulting from parsimony analysis is shown in Fig. 2. The results are consistent with previous studies (Steane & al., 1997, 1999) in that Clerodendrum sensu Steane & al. (2004) is divided into three major clades: an African clade, an Asian clade, and a Pantropical Coastal clade. All three clades are supported by 99%–100% bootstrap (BS) values and posterior probabilities (PP) of 1.0. However, our results reveal that the Pantropical Coastal clade is more closely related to the New World genera (Aegiphila, Amasonia, Tetraclea) than it is to the Asian or African Clerodendrum clades (Fig. 2), a matter not revealed by earlier studies. In addition, a Caribbean species, Clerodendrum spinosum (L.) Spreng., is found not to belong to any of the three major clades. Together with Aegiphila, Amasonia, Tetraclea, it forms a well-supported (84%/1.0, BS/ PP) New World clade. The African and Asian Clerodendrum Redefinition of Clerodendrum and revival of Volkameria and Ovieda. — Clerodendrum as currently circumscribed (Steane & Mabberley, 1998; Steane & al., 2004) is certainly not monophyletic (Fig. 2). In order to delimit Clerodendrum as a monophyletic group, either the New World clade should be incorporated in Clerodendrum or the Pantropical Coastal clade should be removed from Clerodendrum and raised to generic level. Renaming species will therefore be inevitable. We choose the second option here for two reasons: (1) Aegiphila, Amasonia, and Tetraclea, have ca. 120, 8, and 2 species, respectively, whereas the Pantropical Coastal clade comprises ca. 30 species. To minimize the number of name changes, it is more sensible to separate the Pantropical Coastal clade into a distinct genus, for which the earliest name is Volkameria L. in which a number of the germane names have already been published. (2) Retaining the generic distinction for Aegiphila, Amasonia, and Tetraclea while resurrecting Volkameria provides increased information about evolutionary relationships 127 TAXON 59 (1) • February 2010: 125–133 Yuan & al. • Phylogeny of Clerodendrum and allied genera in the classification of this group. Besides separating the Pantropical Coastal clade as the revived genus Volkameria, Clerodendrum spinosum also needs to be removed from Clerodendrum. Its original name, Ovieda spinosa L., is revived for it. Therefore, the newly delimited Clerodendrum is restricted to the Asian and African clades. This is of no little historical Fig. 2. One of eight most parsimonious trees. The Bayesian consensus tree is very similar to this in topology. Only bootstrap values (BS) and posterior probabilities (PP) greater than 80%/0.95 are shown along the branches to avoid being overcrowded. Branches that collapse in the strict consensus are marked by black dots. Geographic distributions are shown on the right. The long-corolla-tube clade is indicated by a thickened arrow. 100/1.0 100/1.0 100/1.0 100/1.0 128 C. bungei C. lindleyi C. canescens C. cyrtophyllum 100/1.0 C. chinense var. simplex C. fortunatum 100/1.0 C. trichotomum 95/1.0 100/1.0 C. japonicum 100/1.0 C. paniculatum C. sp. 100/1.0 100/1.0 ‘C. buchananii ’ C. speciosissimum 100/1.0 C. tomentosum C. floribundum long, narrow C. indicum corolla tube 99/1.0 C. minahassae C. quadriloculare C. schweinfurthii C. johnstonii 100/1.0 C. polycephalum C. splendens C. umbellatum C. bipindense 89/1.0 C. thomsoniae 96/ 1.0 C. volubile 97/ 1.0 C. capitatum C. cephalanthum subsp. mashariki C. cephalanthum subsp. montanum C. poggei 99/1.0 C. rotundifolium C. hildebrandtii 99/1.0 Aegiphila elata 100/1.0 Aegiphila martinicensis 100/1.0 Aegiphila hassleri 88/1.0 Aegiphila brachiata Aegiphila alba 100/1.0 Aegiphila anomala Aegiphila multiflora 84/ 1.0 100/1.0 Amasonia sp. Amasonia hirta 100/1.0 Tetraclea coulteri C. spinosum C. emirnense 100/1.0 C. glabrum 97/1.0 C. aggregatum 85/1.0 C. pittieri 100/1.0 C. aculeatum var. gracile C. ligustrinum 100/1.0 C. linifolium 100/1.0 C. inerme Kalaharia uncinata 100/1.0 Oxera pulchella Faradaya splendida Ajuga chamaepitys Teucrium pyrenaicum Rotheca sp. Aloysia virgata Verbena officinalis Pantropical Coastal New World African Asian 97/1.0 99/1.0 10 changes interest in that molecular work has confirmed three of the four generic concepts used by Linnaeus for this group (see Taxonomy). It is also noteworthy that the long narrow corolla tube has evolved at least twice independently in the group: once in Ovieda and also in the common ancestor of the Asian clade comprising C. indicum (which was independently described TAXON 59 (1) • February 2010: 125–133 Yuan & al. • Phylogeny of Clerodendrum and allied genera in Ovieda at least three times; see below), C. quadriloculare, and others, as mentioned above. Although it is difficult to find unique morphological synapomorphies to separate Clerodendrum, Volkameria, and Ovieda, no doubt a cardinal reason why a broad view of Clerodendrum has prevailed for so long, a combination of several characters, as listed in Table 1, can be readily used to distinguish the three genera. Phylogenetic position of C. hildebrandtii. — Within the African clade, one species, C. hildebrandtii, is sister to the rest of the African group. This species is distinguished from other African species by its large corolla (few flowers in each inflorescence) and large cylindrical calyx. In fact, C. hildebrandtii is the sole member of Verdcourt’s (1992) C. sect. Cylindrocalyx (Thomas) Verd. in his treatment of the genus in East Africa. However, approximately 20 species of Clerodendrum that are restricted to Madagascar closely resemble C. hildebrandtii in morphology. Unfortunately, we were unable to obtain any living material of these species or herbarium specimens of sufficient quality for DNA extraction. But we predict that this particular Madagascan group, together with C. hildebrandtii, will form a clade that is sister to the rest of the African clade. A future study with extensive sampling of this group will shed light on the evolution of this, perhaps the most beautiful, Clerodendrum group. Evolution of an intriguing breeding strategy. — An interesting breeding strategy has been reported in some species of Clerodendrum sensu stricto (s.str.) and Volkameria (Corner, 1940; Primack & al., 1981; Reddy & Reddi, 1995). The stamens and the style are curled upwards tightly inside the flower bud. When the flower opens, the filaments and style start uncoiling. While the filaments project to the centre, the style continues to bend down towards the lower side of the flower. The flower is strongly protandrous and this is the functional male phase. After pollen has been released and the anthers wither, the filaments curl back sideways and the style with its receptive stigma projects back to the centre, taking the position occupied by the stamens in the male phase (see Fig. 3; also see illustration and a detailed description in Reddy & Reddi, 1995). This strategy was first noted by Corner (1940: 700). He mentioned this as typical of species of the genus native in the Malay Peninsula except for C. serratum, which is indeed referable to another genus as Rotheca serrata (L.) Steane & Mabb., where “the stamens and style arch over the top of the flower and one of the petals is modified into a lower lip or landing platform” (Corner 1940: 700). Four decades later the same strategy was described by Primack & al. (1981) in Volkameria inermis (as Clerodendrum inerme). Then in 1995, a detailed description was made by Reddy & Reddi from their observation of C. infortunatum, an African species. Recently, it was observed in C. trichotomum, an Asian species, and C. thomsoniae, an African species, by the first author of this paper before he was aware of the earlier work. This presentation of pollen and stigma in the centre of the flower in a sequential fashion by moving the filaments and style is an elegant combination of dichogamy and herkogamy, that avoids self-fertilization or/and sexual interference (i.e., receiving pollen by stigmas and exporting pollen from anthers: Lloyd & Webb, 1986; Webb & Lloyd, 1986). A subsequent search for floral images has revealed that not only Clerodendrum s.str. and Volkameria, but also Ovieda, Amasonia, Tetraclea, and Kalaharia, all display this particular floral presentation with curled stamens and style at different stages (Fig. 3). Field observations of the pollination ecology of Oxera and Faradaya by de Kok (1997) found these two genera resemble the aforementioned taxa by having protandrous flowers, but not in displaying the alternating movement between the filaments and style. In addition, floral presentation is heterogeneous in Oxera and Faradaya (de Kok, 1997), by contrast with the uniform system found in Clerodendrum s.str. and allies. Clerodendrum s.str., Volkameria, Ovieda, Aegiphila, Amasonia, Tetraclea, and Kalaharia together form a strongly supported clade (Fig. 2). Lamiaceous taxa outside this clade do not show such floral presentation. Therefore, this breeding strategy appears to have evolved only once in the common ancestor of these lineages and is a synapomorphy defining this clade. It has also been lost once and been replaced by a heterostylous system on the path leading to the extant Aegiphila lineage (Fig. 3G–H). This shift has interesting implications for the understanding of evolutionary pathways from homostyly to heterostyly. Our examination of this intriguing breeding strategy in a phylogenetic context provides a striking example of how molecular phylogenetics can re-direct our effort in finding and re-interpreting overlooked morphological characters. The Table 1. Comparison of morphological characters of Clerodendrum s.str., Volkameria, and Ovieda. Clerodendrum Volkameria Ovieda Branches Not tuberculate Not tuberculate Tuberculate Leaf Blade frequently longer than 6 cm, never spiny; venation not camptodromous Blade usually shorter than 6 cm, never spiny; venation not camptodromous Blade longer than 6 cm, margin usually with spiny teeth; venation camptodromous and reticulate with pronounced abaxial relief Inflorescence Usually terminal Frequently axillary Terminal Fruiting calyx Accrescent, larger than fruits, brightly coloured Rarely accrescent, smaller than fruits, enclosing the fruit base, not brightly coloured Accrescent, enclosing fruit, not brightly coloured Fruits Often fleshy, with bright colour contrast- Usually dryish, not brightly coloured ing with calyx Not recorded recently (Burman in Plumier, 1760, has ‘bacca obovata’), apparently not brightly coloured 129 Yuan & al. • Phylogeny of Clerodendrum and allied genera TAXON 59 (1) • February 2010: 125–133 Fig. 3. Representative floral images of Clerodendrum s.str. and allied genera. A–B, Clerodendrum trichotomum Thunb. (photos by Y.-W. Yuan). A, flower at male phase. Note the stamens project to the centre, whereas the style bends down towards the lower side of the flower. B, flower at female phase. The filaments curl back sideways and the style with its receptive stigma projects back to the centre. C, Ovieda spinosa L. (photo courtesy of Jackeline Salazar). D, Volkameria inermis L. (photo courtesy of Forest & Kim Starr). E, Tetraclea coulteri A. Gray (photo courtesy of Burr Williams). The red and blue arrows indicate the flower at male and female phase, respectively. F, Amasonia campestris Moldenke (photo courtesy of Robin Foster). The red arrow indicates the flower at male phase. G–H, Aegiphila sp. (photos courtesy of Kevin Nixon, www.plantsystematics.org), showing the heterostylous system. G, thrum flower; H, pin flower. 130 TAXON 59 (1) • February 2010: 125–133 “Clerodendrum s.str. + Volkameria + Ovieda + Aegiphila + Amasonia + Tetraclea + Kalaharia” clade recovered by molecular data has never been recognized in traditional classification schemes. But, after finding that this complex breeding strategy is shared by these taxa, it becomes obvious that they are all closely related as indicated by DNA sequences. TAXONOMY Genera recognized (see also Steane & Mabberley, 1998 for Rotheca) as a result of this work: 1. Clerodendrum L., Sp. Pl. 2: 637. 1753 – Type: C. infortunatum L. = Siphonanthus L., Sp. Pl. 1: 109. 1753 – Type: S. indicus L. (‘indica’) = Clerodendrum indicum (L.) Kuntze = Cryptanthus Osbeck, Dagb. Ostind. Resa.: 215. 1757 – Type: C. chinensis Osbeck = Clerodendrum chinense (Osbeck) Mabb. Trees, shrubs sometimes suckering, lianes or subherbaceous perennials. Leaves simple (sometimes lobed), decussate or (rarely) whorled, never spiny. Inflorescences cymose, usually terminal. Flowers bisexual; calyx campanulate to tubular, variously lobed, often coloured, usually accrescent; corolla red to yellow, pink or white, with narrow tube, 5-lobed, the lobes usually unequal; stamens 4 (or 5), didynamous, inserted within corolla tube, usually long-exserted; ovary imperfectly 4-locular, each locule with 1 ovule, style terminal, elongate, shortly 2-lobed. Fruit a drupe, often 4-sulcate or 4-lobed; endocarp tough, separating into 4 (or 2 pairs of) pyrenes (sometimes only two maturing), each with one seed. Circa 150 species in tropical Old World with some species found as far south as Australia and as far north as China and Japan. 2. Ovieda L., Sp. Pl. 2: 637. 1753 – Type: O. spinosa L. ≡ Clerodendrum spinosum (L.) Spreng. Shrub to 1.5 m, sometimes subherbaceous; branches tuberculate. Leaves simple, decussate (occasionally some in spirals), coriaceous, margin toothed, teeth usually spiny, venation camptodromous, pinnate-reticulate, with conspicuous pronounced venation in relief on abaxial surface. Inflorescences corymbose, terminal. Flowers bisexual; calyx large, campanulate with 5 acute lobes, accrescent enclosing fruit, not brightly coloured; corolla white with long narrow tube, mouth 5-lobed; stamens 4, exserted; style solitary, as long as stamens; ovary globose. Fruit a drupe, globose to obovate, with 2 locules, each with one seed. One species, restricted to Hispaniola. Ovieda spinosa L., Sp. Pl. 2: 637. 1753 – Lectotype (designated here): Burman in Plumier, Pl. Amer.: t. 256. 1760. Notes. – Hitherto type material has not been designated (Jarvis, 2007: 716). Linnaeus cited Plumier (1703, as Valdia) but that shows insufficient detail to be the basis for his description. A much more detailed description, together with a plate Yuan & al. • Phylogeny of Clerodendrum and allied genera that shows the inflorescence and leaf features that Linnaeus described in 1753 are shown in Burman’s 1760 work. Burman’s plate is based on a tracing of Plumier’s drawing that was made by Claude Aubriet for Herman Boerhaave, and later prepared for publication by Burman. Burman sent proof copies to Linnaeus ahead of their publication, and it seems clear that Linnaeus must have received this one prior to 1753, hence its being appropriate as type material, even though published after Linnaeus’s own work (Jarvis, 2007: 151). Note: Ovieda Spreng. (1824) = Lapeirousia Pourr. (Iridaceae). For Clerodendrum s.l., Baillon (1891) resurrected Ovieda L., which had been made a synonym of Clerodendrum by authors from the 1820s onwards, so his action was inadmissible, though Ovieda as the name for Clerodendrum s.l. (including Volkameria) gained some currency in American publications (*Ovieda aculeata (L.) Baill., Hist. Pl. 11: 95. 1891 non O. aculeata Klatt (1864 = Lapeirousia fabricii (de la Roche) Ker, Iridaceae) = Volkameria aculeata L.; *O. bracteosa (Kostel.) Baill., l.c. (C. bracteosum Kostel. [type (icon): Rheede, Hort. Malab. 4: t. 29. 1683] = Rotheca serrata (L.) Steane & Mabb. [syn. nov.]); O. fragrans (Willd.) Hitchcock (= C. chinense (Osbeck) Mabb.); O. inermis Burm. f. (= C. indicum (L.) Kuntze); *O. infortunata (L.) Baill., l.c. (= C. infortunatum L.); O. mitis L. (= C. indicum (L.) Kuntze); O. ovalifolia A. Juss. (C. ovalifolia (A. Juss.) Bakh. [type: India, Pondichéry, Commerson s.n. in Hb. Juss. (microfiche 347/19-P-JUSS)] = V. inermis L. [syn. nov.]); *O. trichotoma (Thunb.) Baill., l.c. (= C. trichotomum Thunb.); O. verticillata Roxb. ex D. Don [nom. in synon.] = C. indicum (L.) Kuntze). *Additions to Index kewensis and other standard lists. 3. Volkameria L., Sp. Pl. 2: 637. 1753 – Type: V. aculeata L. ≡ Clerodendrum aculeatum (L.) Schldl. = Huxleya Ewart in Proc. Roy. Soc. Victoria, ser. 2, 25: 109. 1912, syn. nov. – Type: H. linifolia Ewart & B. Rees, l.c. ≡ C. linifolium (Ewart & B. Rees) de Kok = V. linifolia (Ewart & B. Rees) Mabb. & Y.W. Yuan, comb. nov. – Type: Australia, Northern Territory, Darwin, 1892, N. Holtze 1322 (lecto MEL). Shrubs, sometimes subherbaceous, lianes, rarely small trees; branches ± tetragonal, usually ash-grey, nodes swollen. Leaves decussate (to ternate), subglabrous,with entire margin, never spiny, venation arcuate-reticulate. Inflorescences axillary to supra-axillary cymes. Flowers usually fragrant; calyx campanulate, only rarely accrescent, margin with 5 broadly triangular small teeth; corolla hypocrateriform, white, sometimes pink or purple, lobes unequal; stamens 4 (or 5), didynamous, inserted within corolla tube, exserted; ovary cylindrical; stigma shortly bifid. Fruits generally globose to obovoid, becoming black or brown and separating into 4 corky pyrenes, each with 1 seed. Approximately 25–30 species, pantropical but with apparently only one species in Asia (V. inermis L.) Notes. – Pending a critical review of Volkameria taxa of Madagascar, where they appear to be numerous (Moldenke, 1956), the germane species are not transferred from Clerodendrum here, though one at least already has a name in 131 Yuan & al. • Phylogeny of Clerodendrum and allied genera Volkameria: V. heterophylla Vent. (C. heterophyllum (Vent.) R. Br., Mascarenes, Madagascar), nor are those of Australia except C. linifolium above. However, all the five species native in Mesoamerica can now be placed as they have been monographed by Rueda (1993, see for species descriptions). To these can be safely added two widespread African species to add to those already with names in Volkameria, plus C. aggregatum used in our analysis: 1. Volkameria acerbiana Vis. ≡ Clerodendrum acerbianum (Vis.) B.D. Jacks. Tropical Africa. See Verdcourt (1992) for a description. 2. V. aculeata L. ≡ C. aculeatum (L.) Schldl. Tropical America. 3. Volkameria aggregata (Gürke) Mabb. & Y.W. Yuan, comb. nov. ≡ Clerodendrum aggregatum Gürke in Bot. Jahrb. Syst. 18: 177. 1894 – Type: Madagascar, Loko-be, Hildebrandt 3339 (B† holo; K iso, NY iso). Tropical Africa. 4. Volkameria costaricensis (Standl.) Mabb. & Y.W. Yuan, comb. nov. ≡ Clerodendrum costaricense Standl. in Publ. Field Mus. Nat. Hist., Bot. Ser. 18: 1002. 1938 – Type: Costa Rica, La Pena de Zarcero, Smith H588 (H holo; MO iso, NY fragm.). Caribbean. 5. Volkameria eriophylla (Gürke) Mabb. & Y.W. Yuan, comb. nov. ≡ Clerodendrum eriophyllum Gürke in Bot. Jahrb. Syst. 18: 178. 1894 – Type: [Tanzania] Fischer ser. I, 331 (B† holo). Tropical Africa. See Verdcourt (1992) for a description. 6. Volkameria glabra (E. Mey.) Mabb. & Y.W. Yuan, comb. nov. ≡ Clerodendrum glabrum E. Mey., Comm. Pl. Afr. Austr.: 273. 1837 – Type: South Africa, Cape Province, R. Basche, Drège s.n., (B† holo; K iso). See Verdcourt (1992) for a description. 7. V. inermis L. ≡ C. inerme (L.) Gaertn. Tropical Asia and Pacific. See Mabberley (2004) for a description. 8. V. ligustrina Jacq. ≡ C. ligustrinum (Jacq.) R. Br. Tropical America. 9. V. linifolia (Ewart & B. Rees) Mabb. & Y.W. Yuan (see above). Northern Australia. 10. Volkameria mollis (Kunth) Mabb. & Y.W. Yuan, comb. nov. ≡ Clerodendrum molle Kunth., Nov. Gen. Sp. 2, quarto ed.: 244. 1818 – Type: Ecuador, Bonpland 3387 (PHB 6209.46: III: 1 (microfiche). Tropical America. 132 TAXON 59 (1) • February 2010: 125–133 11. Volkameria pittieri (Moldenke) Mabb. & Y.W. Yuan, comb. nov. ≡ Clerodendrum pittieri Moldenke in Phytologia 1: 416. 1940 – Type: Peru, Pittier 4965 (US holo; MO iso). Caribbean. ACKNOWLEDGEMENTS We would like to thank all the people who provided plant material; these include Shixiao Luo, Scot Zona, Rogier de Kok, Jun Wen, John Tan, and Robert Jansen. Thanks are also due to the directors/ curators of the following herbaria for the loan of herbarium specimens and permission to take tissue samples: MAPR (Gary Breckon), MO (Jim Solomon), L (Erik Smets), and US (Jun Wen, Vicki Funk). We are grateful to Rogier de Kok (Royal Botanic Gardens Kew) for communicating his observations on the pollination biology of Oxera and Faradaya, and to Kevin Nixon, Jackeline Salazar, Burr Williams, Robin Foster, Forest & Kim Starr for permission to use thei r floral images. We also thank James Wearn (Royal Botanic Gardens Kew) for help in preparing a generic description of Ovieda and its typification, and thank Phil Garnock-Jones and an anonymous reviewer for helpful comments on the manuscript. Funding for this research was provided by a McIntire-Stennis Formula Fund to DJM and a NSF Grant (DEB 0542493) to RGO. LITERATURE CITED Baillon, H. 1891. Histoire des plantes, vol. 11. Paris: Hachette. Briquet, J. 1895. Verbenaceae. Pp. 132–182 in: Engler, A. & Prantl, K. (eds.), Die natürlichen Pflanzenfamilien, vol. IV, 3a. Leipzig: Engelmann. Cantino, P.D. 1992. Evidence for a polyphyletic origin of the Labiatae. Ann. Missouri Bot. Gard. 79: 361–379. Corner, E.J.H. 1940. Wayside trees of Malaya, vol. 1. Singapore: Government Printing Office. De Kok, R.P.J. 1997. The biology and systematics of Oxera, Faradaya, and Hosea (Labiatae). D. Phil. Thesis, University of Oxford, U.K. Doyle, J.J. & Doyle, J.L. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochem. Bull. 19: 11–15. Felsenstein, J. 1985. Confidence limits on phylogenies: An approach using the bootstrap. Evolution 39: 783–791. Graham, S.W., Reeves, P.A., Burns A.C.E. & Olmstead, R.G. 2000. Microstructural changes in noncoding chloroplast DNA: Interpretation, evolution, and utility of indels and inversions in basal angiosperm phylogenetic inference. Int. J. Pl. Sci. 161: S83–S96. Jarvis, C. 2007. Order out of chaos. London: Linnean Society of London & Natural History Museum. Lam, H.J. 1919. The Verbenaceae of the Malayan archipelago: Together with those from the Malayan Peninsula, the Philippines, the Bismarck-archipelago, and the Palau, Marianne, and Caroline islands. Groningen: De Waal. Lloyd, D.J. & Webb, C.J. 1986. The avoidance of interference between the presentation of pollen and stigmas in angiosperms. I: Dichogamy. New Zealand J. Bot. 24: 135–162. Mabberley, D.J. 2004. Clerodendrum L. Pp. 53–57 in: Tirel, C. & Pisivin, C. (eds.), Flore de la Nouvelle-Calédonie et Dépendances, vol. 25. Paris: Muséum National d’Histoire Naturelle. Moldenke, H.N. 1956. Verbenaceae. Pp. 1–273 in: Humbert, H. (ed.), Flore de Madagascar et des Comores, vol. 174. Paris: Firmin-Didot. Moldenke, H.N. 1985. Notes on the genus Clerodendrum (Verbenaceae), IV. Phytologia 57: 334–365. TAXON 59 (1) • February 2010: 125–133 Plumier, C. 1703. Nova plantarum americanarum genera. Paris: Boudot. Plumier, C. 1760. Plantarum americanarum fasciculus decimus. Amsterdam: Schouten; Leiden: Potvliet & Haak. Posada, D. & Crandall, K.A. 1998. Modeltest: Testing the model of DNA substitution. Bioinformatics 14: 817–818. Primack, R.B., Duke, N.C. & Tomlinson, P.B. 1981. Floral morphology in relation to pollination ecology in five Queensland coastal plants. Austrobaileya 1: 346–355. Rambaut, A. 1996. Se-Al: Sequence alignment editor. Available at http://evolve.zoo.ox.ac.uk. University of Oxford. Reddy, T.B. & Reddi, C.S. 1995. Butterfly pollination of Clerodendrum infortunatum (Verbenaceae). J. Bombay Nat. Hist. Soc. 92: 166–173. Rimpler, H., Winterhalter, C. & Falk, U. 1992. Cladistic analysis of the subfamily Caryopteridoideae Briq. and related taxa of Verbenaceae and Lamiaceae using morphological and chemical characters. Pp. 39–54 in: Harley, R.M. & Reynolds, T. (eds.), Advances in Labiatae science. Kew: Royal Botanic Gardens. Ronquist, F. & Huelsenbeck, J.P. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572–1574. Rueda, R.M. 1993. The genus Clerodendrum (Verbenaceae) in Mesoamerica. Ann. Missouri Bot. Gard. 80: 870–890. Schauer, J.C. 1847. Verbenaceae. Pp. 658–676 in: de Candolle, A. (ed.), Prodromus systematis naturalis regni vegetabilis. Paris: Treuttel & Würtz. Shaw, J., Lickey, E.B., Beck, J.T., Farmer, S.B., Liu, W.S., Miller, J., Siripun, K.C., Winder, C.T., Schilling, E.E. & Small, R.L. 2005. The tortoise and the hare II: Relative utility of 21 noncoding chloroplast DNA sequences for phylogenetic analysis. Amer. J. Bot. 92: 142–166. Simmons, M.P. 2004. Independence of alignment and tree search. Molec. Phylog. Evol. 31: 874–879. Yuan & al. • Phylogeny of Clerodendrum and allied genera Simmons, M.P. & Ochoterena, H. 2000. Gaps as characters in sequence-based phylogenetic analyses. Syst. Biol. 49: 369–381. Steane, D.A., de Kok, R.P.J. & Olmstead, R.G. 2004. Phylogenetic relationships between Clerodendrum (Lamiaceae) and other Ajugoid genera inferred from nuclear and chloroplast DNA sequence data. Molec. Phylog. Evol. 32: 39–45. Steane, D.A. & Mabberley, D.J. 1998. Rotheca (Lamiaceae) revived. Novon 8: 204–206. Steane, D.A., Scotland, R.W., Mabberley, D.J. & Olmstead, R.G. 1999. Molecular systematics of Clerodendrum (Lamiaceae): ITS sequences and total evidence. Amer. J. Bot. 86: 98–107. Steane, D.A., Scotland, R.W., Mabberley, D.J., Wagstaff, S.J., Reeves, P.A. & Olmstead, R.G. 1997. Phylogenetic relationships of Clerodendrum s.l. (Lamiaceae) inferred from chloroplast DNA. Syst. Bot. 22: 229–243. Swofford, D.L. 2002. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4b10. Sunderland, Massachusetts: Sinauer. Taberlet, P., Gielly, L., Pautou, G. & Bouvet, J. 1991. Universal primers for amplification of three non-coding regions of chloroplast DNA. Pl. Molec. Biol. 17: 1105–1109. Thomas, B. 1936. Die Gattung Clerodendrum in Afrika. Bot. Jahrb. Syst. 68: 1–106. Verdcourt, B. 1992. Clerodendrum. Pp. 84–144 in: Polhill, R.M. (ed.), Flora of tropical East Africa: Verbenaceae. Rotterdam: Balkema. Webb, C.J. & Lloyd, D.G. 1986. The avoidance of interference between the presentation of pollen and stigmas in angiosperms II: Herkogamy. New Zealand J. Bot. 24: 163–178. Yuan, Y.W. & Olmstead, R.G. 2008. A species-level phylogeny of the Verbena (Verbenaceae) complex indicates two independent intergeneric chloroplast transfers. Molec. Phylog. Evol. 48: 23–33. Appendix 1. Plant material used in this study. Each entry consists of: taxon name, country in which it was collected, year of collection, collector with collection number, and herbarium (accession number when available) where voucher was deposited. Aegiphila alba Moldenke, Ecuador, 2004, J. Clark 7960, US; A. anomala Pittier, Costa Rica, 1995, J.F. Morales & V. Urena 3760, MO (5748376); A. brachiata Vell., Brazil, 2003, A. Kegler 1581, MO (5772239); A. elata Sw., Panama, 1986, G. McPherson 8473, MO (4290179); A. hassleri Briq., Argentina, 2000, De Romero & al. 2201, MO (5830202); A. martinicensis Jacq., Puerto Rico, 1980, J.C. Solomon 5732, MO (2897682); A. multiflora Ruiz & Pav., Bolivia, 1983, St. G. Beck 8688, MO (4275754); Ajuga chamaepitys (L.) Schreb., Cantino 185, K; Amasonia hirta Benth., Brazil, 1982, B.A.S. Perecira 251, MO (2994726); Amasonia sp., Brazil, 1997, Giulietti & al., PCD 6176, K; Clerodendrum aculeatum (L.) Schldl. var. gracile Griseb. ex Moldenke, Fairchild Tropical Botanic Garden, 2006, Zona 1100 & Gillis 9169, FTG; C. aggregatum Gürke, Madagascar, 1996, Gautier & S.T. Be, LG 2873, MO (5208533); C. bipindense Gürke, Guinea, 1987, Carvalho 3033, MO (4322999); C. buchananii auctt., non (Roxb.) Walp., Oxford Botanic Garden, Steane 76, FHO; C. bungei Steud., Oxford Botanic Garden, Steane 78, FHO; C. canescens Wall. ex Walp., China, S.X. Luo 242, IBSC; C. capitatum (Willd.) Schumach. & Thonn., Tanzania, 1998, O.A. Kibure 183, MO (5310537); C. cephalanthum Oliv. subsp. mashariki Verdc., Tanzania, 1999, M.A. Mwangoka 974, MO (5290487); C. cephalanthum subsp. montanum (Thomas) Verdc., Tanzania, 2003, O.A. Kibure 982, MO (04473598); C. chinense (Osbeck) Mabb. var. simplex (Moldenke) S.L.Chen, Fairchild Tropical Botanic Garden, 2006, Fantz 3419, FTG; C. cyrtophyllum Turcz., China, S.X. Luo 253, IBSC; C. emirnense Boj. ex Hook., Madagascar, 1990, P.B.Phillipson & al. 3414, MO (3842769); C. floribundum R. Br. Australia, 1989, C.R. Dunlop 8055, L(0625213); C. fortunatum L., China, S.X. Luo 252, IBSC; C. glabrum E. Mey., Oxford Botanic Garden, Steane 87, FHO; C. hildebrandtii Vatke var. puberula Verdc., Tanzania, 1999, C.J. Kayombo 2302, MO (5291257); C. indicum (L.) Kuntze, Singapore, RK 5394; C. inerme (L.) Gaertn., Oxford Botanic Garden, Steane 86, FHO; C. japonicum (Thunb.) Sweet, China, S.X. Luo 254, IBSC; C. johnstonii Oliv., Tanzania, 1999, P. Phillipson & J. Mlwangwa 2017, MO (5189370); C. ligustrinum (Jacq.) R. Br., Mexico, 1998, R. Novelo & Ay Ramos V.L. 2901, MO (5300788); C. lindleyi Decne. ex Planch. China, S.X. Luo 251, IBSC; C. linifolium (Ewart & B. Rees) de Kok, Australia, 1999, Cowie 8213, K; C. minahassae Teijsm. & Binn., Fairchild Tropical Botanic Garden, 2006, Houghton & White 1145, FTG; C. paniculatum L., Palau (in cultivation), 1966, N.H. Cheatham 32, US(3356180); C. pittieri Moldenke, Peru, 1976, D.C. Wasshausen & F. Encarnarium 686, US (2956667); C. poggei Gürke, Tanzania, 1999, G. Gobbo 281, MO (5290499); C. polycephalum Bak., Ghana, 1994, C.C.H. Jongkind & H.H. Schmidt 1735, MO (04667183); C. quadriloculare Merr. Fairchild Tropical Botanic Garden, 2006, Zona 1104, FTG; C. rotundifolium Oliv., Tanzania, 2000, S. Bidgood & al. 4819, MO (5658503); C. splendens G. Don, National Botanic Garden, Zimbabwe; C. schweinfurthii Gürke, Oxford Botanic Garden, Steane 82, FHO; C. sp., Indonesia, 2001, Ramadhanil & al. 461, L (0334140); C. speciosissimum Hort. Gand. ex Drapiez, Oxford Botanic Garden, Steane 90, FHO; C. spinosum (L.) Spreng. Dominican Republic, 2001, F. Jimenez 3332, MAPR (26103); C. thomsoniae Balf., UW Greenhouse, 2002, RGO 2002-07, WTU; C. tomentosum (Vent.) R. Br., Australia, 1984, S.J. Forbes 2510, L (0625219); C. trichotomum Thunb., R.G. Olmstead home garden, 2002, RGO 2002-134, WTU; C. umbellatum Poir., Tanzania, 1999, G. Gobbo 486, MO (5290496); C. volubile P. Beauv, Ghana, 1996, M. Merello & al. 1345, MO (05030536); Faradaya splendida F. Muell., H. Rimpler 2144, FB; Kalaharia uncinata (Schinz) Moldenke, Tanzania, 2002, N.A. Mwangulango 973, MO (5721485); Oxera pulchella Labill., H. Rimpler 1328, FB; Rotheca sp., Madagascar, Wen 9487, US; Tetraclea coulteri A. Gray, U.S.A., Ki-Joong Kim 100026, TEX; Teucrium pyrenaicum L., Voucher unknown. Appendix 2. Primers used for PCR and sequencing; the ones that were specifically designed for this study are marked by an asterisk (*). trnL-F: c, d, e, f (Taberlet & al., 1991); trnS-fM: trnSUGA & trnfMCAU (Shaw & al., 2005), psbZF(V) (Yuan & Olmstead, 2008), psbZR(C)* (5 -CATCAATCTTATTGATTAGCGTA-3 ); trnD-T: trnDGUCF & trnTGGU R (Shaw & al., 2005), trnD-TE2T* (5 -AATTCGAATCCCCGCTGCCTCC-3 ), trnD-TE2D* (5 -CATTCCATTATATTGACAATT-3 ); trnT-L: a (Taberlet & al., 1991), R(Cler.)* (5 -ACCTATAGGAAACCCATATT-3 ). 133