TAXON 60 (2) • April 2011: 471–484 Bendiksby & al. • Lamioideae─a taxonomic update M O L E C U L A R PH Y LO G E N E T I C S A N D B I O G E O G R A PH Y An updated phylogeny and classification of Lamiaceae subfamily Lamioideae Mika Bendiksby,1 Lisbeth Thorbek,1 Anne-Cathrine Scheen, 2 Charlotte Lindqvist 3 & Olof Ryding4 1 2 3 4 National Centre for Biosystematics, Natural History Museum, University of Oslo, P.O. Box 1172 Blindern, 0318 Oslo, Norway Museum of Archaeology, University of Stavanger, 4036 Stavanger, Norway Department of Biological Sciences, University at Buffalo (SUNY), Buffalo, New York 14260, U.S.A. Botanical Garden and Museum, Natural History Museum of Denmark, University of Copenhagen, Gothersgade 130, 1123 Copenhagen K, Denmark Author for correspondence: Olof Ryding, OlofR@snm.ku.dk Abstract Lamioideae comprise the second-largest subfamily in Lamiaceae. Although considerable progress has recently been made in Lamioideae phylogenetics, the subfamily remains one of the most poorly investigated subfamilies in Lamiaceae. Here we present a taxonomic update of the subfamily based on earlier published data as well as 71 new DNA extracts from relevant in- and outgroup taxa, and DNA sequence data from four chloroplast regions (matK, rps16, trnL intron and trnL-F spacer). The phylogenetic positions of 10 out of 13 previously unplaced small or monotypic Asian lamioid genera and 37 additional lamioid species have been identified, and the classification is updated accordingly. Results from parsimony and Bayesian phylogenetic methods corroborate earlier results, but phylogenetic resolution as well as overall branch support are improved. All newly added genera are assigned to earlier established tribes or the new tribe Paraphlomideae Bendiksby, which includes Ajugoides, Matsumurella and Paraphlomis. Acanthoprasium is resurrected as a genus. Transfer of species is proposed to accommodate the monophyly of two genera (Lamium, Otostegia), whereas ten genera remain non-monophyletic (Ballota s.str., Lagopsis, Leonotis, Leonurus, Leucas, Microtoena, Phlomoides, Sideritis, Stachys, Thuspeinanta). Eriophyton and Stachyopsis have been included in Lamieae, Hypogomphia in Stachydeae, and Loxocalyx in Leonureae. Betonica, Colquhounia, Galeopsis, and Roylea remain unclassified at the tribal level. Lamium chinense and three East Asian Galeobdolon species are transferred to Matsumurella. Sulaimania and four Otostegia species are transferred to Moluccella. Alajja and three Lamium species are transferred to Eriophyton. In total, 14 new combinations are made, one at the rank of subgenus and 13 at the rank of species. Keywords Acanthoprasium ; classification; Eriophyton ; Lamiaceae; Lamioideae; Matsumurella ; molecular phylogenetics; Moluccella ; Paraphlomideae Supplementary Material The Appendix is available in the free Electronic Supplement to the online version of this article (http://www.ingentaconnect.com/content/iapt/tax). INTRODUCTION In Kubitzki’s family monograph, Harley & al. (2004) divided the angiosperm family Lamiaceae into seven subfamilies. The second largest subfamily, Lamioideae (including Pogostemonoideae), was considered to consist of 63 genera and about 1260 species. Since Harley & al. (2004), Rydingia has been established (Scheen & Albert, 2007), Betonica has been resurrected from synonymy under Stachys (Scheen & al., 2010), Phlomoides has been resurrected and Lamiophlomis Kudô, Notochaete Benth. and Pseuderemostachys Popov. have been subsumed into Phlomoides (Mathiesen & al., in press). Thus, 63 genera are currently included in subfamily Lamioideae. Recently, Scheen & al. (2010) produced a first, general phylogenetic framework for Lamioideae based on chloroplast DNA data, which has elucidated evolutionary relationships of many genera and clades and permitted a preliminary tribal classification system. For example, the molecular phylogeny presented by Scheen & al. (2010) confirmed that Lamioideae were non-monophyletic following earlier work by Cantino & al. (1992), but monophyletic as circumscribed by Harley & al. (2004), i.e., including former subfamily Pogostemonoideae. However, the exact circumscription of the subfamily within Lamiaceae still needs to be corroborated with better sampling in the family. In Cantino’s (1992a,b) morphological phylogeny of Lamiaceae, the subfamily Nepetoideae was nested within Lamioideae, and this group in turn emerged most closely related to Ajuga (in Ajugoideae) and the incertae sedis genera Cymaria, Acrymia, Holocheila and Garrettia. However, according to more recent molecular data, the two subfamilies Lamioideae and Nepetoideae are only remotely related, and Lamioideae are more closely related to Scutellarioideae than to Ajugoideae (Wink & Kaufmann, 1996; Wagstaff & Olmstead, 1997; Wagstaff & al., 1998). Scheen & al. (2010) included Cymaria, which emerged as the sister group of Lamioideae, while Scutellarioideae were shown to be the phylogenetic sister of the Cymaria-Lamioideae clade. However, a molecular survey of 471 Bendiksby & al. • Lamioideae─a taxonomic update the phylogenetic positions of Acrymia, Holocheila and Garrettia remains to be published. Scheen & al. (2010) also investigated phylogenetic relationships within Lamioideae. Based on their analyses, which were based on three plastid markers in 159 species from 50 genera, they discerned nine tribes within Lamioideae and discussed the non-monophyly of some genera. Although a majority of lamioid genera was included in their survey, some important groups were underrepresented and a number of small or monogeneric taxa left out. For example, 16 genera remained unclassified at the tribal level, either because they would have formed monogeneric tribes (Betonica, Colquhounia, Eriophyton, Galeopsis, Paraphlomis, Roylea), or because they could not be placed due to lack of molecular data (Ajugoides, Alajja, Hypogomphia, Loxocalyx, Matsumurella, Metastachydium, Paralamium, Pseudomarrubium, Stachyopsis, Sulaimania). Three genera were classified to tribal level, either based on morphology alone (Colebrookea) or on morphology and limited, unpublished trnL intron sequence data (Eurysolen, Lagopsis). Although considerable progress has recently been made in Lamioideae phylogenetics, the subfamily remains one of the most poorly investigated subfamilies in Lamiaceae. For example, only limited groups within Lamioideae have been subjected to phylogenetic studies: e.g., tribe Lamieae (Ryding 2003), tribe Leucadeae (Ryding 1998; Scheen & Albert, 2009), the indigenous Hawaiian labiates (Lindqvist & Albert, 2002; Lindqvist & al., 2003), tribe Phlomoideae (Ryding, 2008; Pan, 2009; Mathiesen & al., in press), Sideritis (Barber & al., 2000, 2002, 2007), and tribe Synandreae (Scheen & al., 2008). The main purpose of the present study is to determine the phylogenetic positions of lamioid genera that were omitted in the study of Scheen & al. (2010). A modified DNA-miniprep. and PCR protocol (Bendiksby & al., in prep) was used to obtain amplicons from old and presumably DNA-degraded plant tissues. Moreover, in order to increase phylogenetic resolution and branch support, some additional taxa and one additional marker (matK) were included. Among the 64 currently recognized genera of Lamioideae, only the monotypic genera Metastachydium, Paralamium, and Pseudomarrubium are lacking in this study. A few taxonomic and nomenclatural changes in accordance with the obtained results are proposed. MATERIALs AND METHODs Taxon sampling. — All taxon names in the present study follow the “World Checklist of Lamiaceae and Verbenaceae” (Govaerts & al., 2010), except for species belonging to Betonica and Phlomoides, for which the Checklist is not yet updated (see Scheen & al., 2010; Mathiesen & al., in press). Author names for taxa included in the present study are assembled in the Appendix (Electronic Supplement). A total of 402 DNA sequences were generated from specimens held at the following herbaria: A, BHO, C, E, GH, L, NY, O, S, TEX, UPS, US, and WU or, in a few cases, from silicadried leaves (vouchers held at O). A total of 238 accessions representing 208 species from 60 Lamioideae genera (all but 472 TAXON 60 (2) • April 2011: 471–484 three) were included as in-group, of which 164 accessions were also used by Scheen & al. (2010). Ten genera were included in the molecular phylogeny of subfamily Lamioideae for the first time: Ajugoides, Alajja, Colebrookea, Eurysolen, Hypogomphia, Lagopsis, Loxocalyx, Matsumurella, Stachyopsis and Sulaimania. Colebrookea, Eurysolen and Lagopsis were included in order to confirm their suggested tribal position based on morphology or restricted and unpublished molecular data (see Scheen & al., 2010). Additional samples relevant for monophyly assessments and taxonomic circumscriptions were also included: (1) accessions of some monotypic genera or genera represented by only a single species in Scheen & al. (2010) (e.g., Alajja, Colebrookea, Eurysolen, Garrettia, Lagopsis, Loxocalyx, Microtoena, Stachyopsis, Synandra, and Roylea); (2) species from un-sampled geographic areas belonging to genera with disjunct distributions (e.g., Achyrospermum wallichianum and Pogostemon aquaticus); and (3) species that, based on morphology, were expected to be extraneous to their genera (e.g., Ballota frutescens, Lamium chinense, L. tuberosum, Otostegia bucharica and O. olgae). The outgroup comprised 42 accessions, of which 18 were extracted for the present study, including seven taxa from subfamily Nepetoideae, five taxa from subfamily Prostantheroideae, five taxa from subfamily Scutellarioideae, one taxon from subfamily Symphorematoideae, five taxa from subfamily Ajugoideae, five taxa from subfamily Viticoideae, and six taxa that have not been ascribed to a subfamily (Acrymia ajugiflora, Callicarpa americana, C. japonica, Cymaria dichotoma, Garrettia siamensis, Tectona grandis ; referred to as incertae sedis in Harley & al., 2004). Finally, five members of related families were included as a more distant outgroup. DNA sequence data of the trnL-F region (trnL intron and trnL-trnF intergenic spacer) and the rps16 intron from all but three accessions used by Scheen & al. (2010) were also included in the present study, and DNA sequence data of an additional genetic marker (matK) were generated for the same accessions, except for about 20 DNA extracts that were no longer available. Attempts were made to amplify and sequence all four chloroplast regions from 71 new DNA extracts in order to provide a near complete generic representation of Lamioideae, a more balanced outgroup for the phylogenetic analyses, and a better resolved phylogeny. Sequences that were not new to this study have been retrieved from GenBank and were originally published by Wallander & Albert (2000), Barber & al. (2002), Beardsley & Olmstead (2002), Bremer & al. (2002), Lindqvist & Albert (2002), Shi & al. (2003), Paton & al. (2004), Scheen & al. (2008), Yuan & Olmstead (2008), Scheen & Albert (2009), Scheen & al. (2010) and Mathiesen & al. (in press). Voucher information and GenBank accession numbers are provided in the Appendix. DNA extraction, PCR amplification and DNA sequencing. — Between 10 and 30 mg of dried plant material was crushed twice in a 2 ml plastic tube with two tungsten carbide beads for 1 minute at 30 Hz on a mixer mill (MM301, Retsch GmbH & Co., Haan, Germany). Total DNA from the crushed samples was extracted using the E.N.Z.A SP Plant DNA Mini Kit (Omega Bio-Tek Inc., Norcross, Georgia, U.S.A.) according to the TAXON 60 (2) • April 2011: 471–484 manufacturer’s manual. All four chloroplast regions (trnL intron, trnL-trnF intergenic spacer, rps16 intron, matK gene) were amplified and sequenced as described by Scheen & al. (2010) unless otherwise specified. Amplification of shorter fragments was attempted when long fragments did not amplify successfully, presumably due to low-quality template. For amplification of matK we used the following six primers that were developed from available Lamiaceae matK sequences in GenBank: matK-1Fa (5′-CAGGAGTATATTTATGCATTTGCTC-3′), matK-1Fb (5′-CTATATCCACTTATCTTTCAGGAGT-3′), matK-3F (5′-CATGTGGAAATCTTGGTTCAAATC-3′), matK-5Ra (5′-CAAGAAAGTCGAAGTATATACTTTA-3′), matK-5Rb (5′-TCGAAGTATATACTTTATTCGATAC-3′), and matK-3R (5′-TAATAAATGCAAGGAGGAAGCATC-3′). matK was amplified either as one fragment using the primer combination matK-1Fb and matK-5Ra, or as two shorter fragments using matK-1Fb or matK-1Fa and matK-3R (the 5′ end), and matK-3F and matK-5Ra or matK-5Rb (the 3′ end). Likewise, rps16 was amplified either as one fragment using the primer combination rpsF and rpsR2R (Oxelman & al., 1997), or as two shorter fragments using rpsF or rps-LamF and rpsLamR2 (the 5′ end), and rps-LamF2 and rpsR2R or rps-LamR (the 3′ end). The following four primers were developed for the present study from obtained Lamioideae rps16 sequences: rps-LamF (5′-GAARGACACGATCCGTTGTGGA-3′), rpsLamF2 (5′-GAAGTAATGTCTAAACCCAATG-3′), rps-LamR (5′-CGATTCGATAGATGGCTCATTG-3′), and rps-LamR2 (5′-ATCATTGGGTTTAGACATTACT-3′). The PCR-enzyme AmpliTaqGold DNA Polymerase (Applied Biosystems, Foster City, California, U.S.A.) was used for amplifying DNA obtained from old herbarium specimens or DNA extracts of reduced quality, whereas AmpliTaq DNA Polymerase (Applied Biosystems) was used for all high-quality DNA extracts. Samples from which amplicons were not obtained using regular DNA miniprep- and PCR procedures, were re-assessed using a “replicate” procedure as described by Bendiksby & al. (in prep.). Regularly obtained PCR products were purified using 2 µl of a 1-in-10 dilution of ExoSAP-IT (enzyme/buffer stock; USB Corporation, Cleveland, Ohio, U.S.A.) to 8 µl PCR product, incubated at 37°C for 45 minutes followed by 15 minutes at 80°C. See Bendiksby & al. (in prep.) for purification procedure of the “replicate” reactions. Cycle sequencing was performed by the CEES ABI-laboratory (http://www.bio.uio.no/ABI-lab/) using the ABI BigDye Terminator sequencing buffer and v3.1 Cycle Sequencing kit. Sequences were processed on an ABI 3730 DNA analyser (Applied Biosystems) and assembled and edited using SEQUENCHER v.4.1.4 (Gene Codes Corporation, Ann Arbor, Michigan, U.S.A.). All DNA extracts generated in the present study, as well as most of the DNA extracts included in Scheen & al. (2010), have been deposited in the DNA/tissue collection at Natural History Museum, Oslo (O). All sequences new to the present study have been deposited in GenBank and accession numbers are listed in the Appendix. Alignment and phylogeny reconstructions. — Sequences from 280 accessions were aligned manually using BioEdit v.7.0.9.0 (Hall, 1999). Insertions/deletions (indels) were coded Bendiksby & al. • Lamioideae─a taxonomic update as present/absent and added to the matrices as additional, unordered characters using the program SeqState (Müller, 2005) following the simple indel coding of Simmons & Ochoterena (2000). The risk of DNA and/or PCR contamination increases when using old and degraded DNA. Therefore, parsimony analyses of the separate regions were conducted (as described below) in order to detect potential errors during the process from leaf tissue to aligned sequence. Six contaminated sequences were identified and excluded from further analyses, and congruence of the resultant four gene-trees was confirmed prior to concatenation. Thus, testing for contamination was done in several ways: by separate gene tree analyses, by including multiple accessions of taxa, and by evaluating phylogenetic position against expectations from morphology. Optimal models of nucleotide substitution for the various markers were estimated using the Akaike information criterion (AIC) and the software MrModeltest (Nylander, 2004) at the Bioportal (www.bioportal.uio.no). A partitioned concatenated alignment of four genetic regions and 280 accessions was analyzed twice, with and without indels coded, using MrBayes v.3.1.2 (Ronquist & Huelsenbeck, 2003) at the Bioportal. Because long stretches of missing characters may confound phylogenetic results in different ways, phylogenetic analyses with and without indels coded were performed also on a “non-orphan” matrix, i.e., including only the 259 accessions for which at least three genetic markers were available. Posterior probabilities were determined twice by running one cold and four heated chains for six million generations in parallel mode, saving trees every 1000th generation. To test whether the Markov Chain converged, we monitored the standard deviation of split frequencies (SDSF), which did fall below 0.01 (in all analyses) when comparing two independent runs. The generations prior to the point when the SDSF permanently fell below 0.01 were discarded as burn-in. A 50% majority rule consensus tree was used to calculate posterior probabilities. Parsimony analyses and branch support obtained from parsimony jackknifing (Farris & al., 1996) were run using the freely available software TNT (Goloboff & al., 2003) as described by Scheen & al. (2010). Alignment and phylogeny from the present study are available as Supplementary Data to the online version of this article (http://www.ingentaconnect.com/content/iapt/tax). REsULTs Lengths in basepairs (bp) of the aligned DNA-regions were: 1278 bp for the trnL-F region, 1293 bp for the rps16 intron, and 1185 bp for matK. The concatenated alignment of the four genetic regions was 3756 bp long, and simple indel coding recognized 526 indels. The General Time Reversible model of nucleotide substitution with gamma distribution and sites invariant (GTR + G + I) was the estimated best-fit model for all genetic regions except matK, for which a simpler model GTR + G was selected. All phylogenies obtained from Bayesian and parsimony analyses of the full (280 accessions) and the non-orphan (259 accessions) datasets, with and without indel 473 Bendiksby & al. • Lamioideae─a taxonomic update coding, were congruent, but resolved to different extents. Indel coding increased overall branch support. The resolution present in the parsimony results, with or without indel coding, was also present in the Bayesian results. Overall, the Bayesian tree was better resolved. The phylogenetic result from the indel-coded Bayesian analysis of the full dataset is presented in Fig. 1, with ranges of parsimony jackknife support (JK) indicated. The topology was generally strongly supported, and removal of accessions with long stretches of missing data resulted in increased JK for branches of the clades concerned (Fig. 1). The number of most-parsimonious trees was reduced drastically when orphans were excluded from the matrix, as was computational time, for both the Bayesian and the parsimony analyses. The consistency- and retention indices were highly similar in all parsimony analyses (CI = 55–56 and RI = 86–87). Subfamily Lamioideae formed a clade including the ten genera that were new to this study: Alajja, Ajugoides, Colebrookea, Eurysolen, Hypogomphia, Lagopsis, Loxocalyx, Matsumurella, Stachyopsis, and Sulaimania (Fig. 1). Colebrookea and Eurysolen both grouped with other species of Pogostemoneae, Colebrookea as sister to a clade of Craniotome, Microtoena, Anisomeles, and Pogostemon, and Eurysolen as sister a clade of Leucosceptrum, Rostrinucula and Comanthosphace (Fig. 1A: clades a and b, respectively). Hypogomphia was nested within Stachydeae, as sister to Thuspeinanta brahuica (Fig. 1B). Both Ajugoides and Matsumurella formed a clade with species of Paraphlomis and Lamium chinense (Fig. 1C). Lagopsis and Loxocalyx were both nested within Leonureae (Fig. 1C). Sulaimania and two species of Otostegia were nested within Moluccella (Fig. 1C). Alajja and Stachyopsis grouped with Eriophyton and two specimens of Lamium tuberosum (Fig. 1C). Some genera were resolved as monophyletic with the inclusion of more accessions, e.g., Gomphostemma, Melittis and Roylea (Fig. 1). However, several genera were non-monophyletic, including Ballota, Lagopsis, Lamium, Leonotis, Leonurus, Leucas, Microtoena, Moluccella, Otostegia, Phlomoides, Sideritis, Stachys, and Thuspeinanta (Fig. 1B–C). Acrymia and Cymaria formed a supported clade that was sister to the Lamioideae clade (Fig. 1A). Subfamily Scutellarioideae was the sister of the Acrymia-Cymaria-Lamioideae clade and Garrettia was the sister of the Scutellarioideae-AcrymiaCymaria-Lamioideae clade (Fig. 1A). Subfamily Viticoideae did not form a monophyletic group (Fig. 1A). TAXON 60 (2) • April 2011: 471–484 DIsCUssION Our expanded molecular phylogeny of subfamily Lamioideae (Fig. 1) largely corroborates the results and the taxonomic changes proposed by Scheen & al. (2010) and Scheen & Albert (2007). Phylogenetic affinity and tribal position have been determined for the ten Lamioideae genera that were omitted in Scheen & al. (2010; Alajja, Ajugoides, Colebrookea, Eurysolen, Hypogomphia, Lagopsis, Loxocalyx, Matsumurella, Stachyopsis, Sulaimania) as well as Eriophyton and Paraphlomis, which were previously not assigned to tribal level (see below; Fig. 1). However, four genera remain unplaced in the updated tribal classification of subfamily Lamioideae: Betonica, Colquhounia, Galeopsis, and Roylea. Seven genera were shown to be para- or polyphyletic by Scheen & al. (2010): Ballota, Leonotis, Leonurus, Leucas, Phlomis, Sideritis, and Stachys. Taxonomic changes proposed by Mathiesen & al. (in press), which have been followed herein, render Phlomis monophyletic, but the remaining six genera are still para- or polyphyletic (Fig. 1B–C). Furthermore, our results show that non-monophyly also applies to Otostegia, Lagopsis, Lamium, Microtoena, Moluccella, Phlomoides, and Thuspeinanta, as currently circumscribed (i.e., by Govaerts & al., 2010). The inclusion of more data has provided increased phylogenetic resolution and stronger support for most of the clades within Lamioideae. For example, the clade that includes Roylea and the three tribes Marrubieae, Lamieae and Leucadeae receives considerably improved support (Fig. 1C). However, the phylogenetic position of Colquhounia receives less support in the present study (Fig. 1B) as compared to Scheen & al. (2010). Since most major clades were thoroughly discussed by Scheen & al. (2010), only new results will be discussed below. Updates on tribe Pogostemoneae. — The new molecular results confirm that two Asian monotypic genera, Colebrookea and Eurysolen, belong in tribe Pogostemoneae (Fig. 1A), a relationship previously suggested based on morphology and limited unpublished DNA-sequence data (Scheen & al., 2010). The very distinctive monotypic genus Colebrookea is resolved as the phylogenetic sister to the large subclade of Craniotome, Microtoena, Anisomeles and Pogostemon (Fig. 1A, clade a). This relationship receives a posterior probability of 0.99, but is not supported by parsimony jackknifing. Some morphological traits of Colebrookea are similar to the genera in its sister clade, for example, small nutlets, not much longer than broad, Fig. 1. The 50% majority rule consensus phylogram from a partitioned Bayesian analysis of an indel-coded, concatenated matrix of 280 accessions and 3756 bp from four regions of the chloroplast genome (trnL-intron, trnL-F spacer, rps16 intron, matK). The 3565 generations prior to the point when the SDSF permanently fell below 0.01 (0.0077 at termination) were discarded as burn-in. Ranges of parsimony jackknife support (JK) above 50% are indicated with dots at the branches (see inset box). Only Bayesian posterior probability values (PP) of more than 0.95 are reported (below branches and in bold face). Numbers in italics above branches indicate branch support obtained from parsimony jackknifing on the non-orphan dataset (259 accessions) when exceeding the range reported from the full dataset (280 accessions; i.e., values that were comparable between the two analyses were not reported). Accessions that were omitted from the non-orphan analysis are indicated with an asterisk. Multiple accessions of the same species are numbered according to the Appendix. Major lamioid clades are named following the suprageneric classification proposed by Scheen & al. (2010) and updates suggested in the present study. Lowercase letters (a–h) indicate some of the clades discussed in the text. The phylogeny is subdivided as follows: A, Outgroup that includes five non-Lamiaceae taxa, Lamiaceae subfamilies and genera incertae sedis, and the lamioid tribe Pogostemoneae. Gray shading is used to highlight currently recognized (sensu Harley & al., 2004) subfamilies and genera incertae sedis of the outgroup. B, The lamioid taxa Colquhounia, Gomphostemmateae, Synandreae, Galeopsis, Betonica and Stachydeae. C, The lamioid taxa Paraphlomideae, Phlomideae, Leonureae, Roylea, Marrubieae, Lamieae and Leucadeae. 474 Bendiksby & al. • Lamioideae─a taxonomic update ≥ ≥ ≥ incertae sedis incertae sedis Ajugoideae incertae sedis incertae sedis TAXON 60 (2) • April 2011: 471–484 475 Bendiksby & al. • Lamioideae─a taxonomic update TAXON 60 (2) • April 2011: 471–484 ≥ ≥ ≥ Stachydeae C 476 TAXON 60 (2) • April 2011: 471–484 Bendiksby & al. • Lamioideae─a taxonomic update ≥ Leucadeae ≥ ≥ 477 Bendiksby & al. • Lamioideae─a taxonomic update with a very distinctive sclerenchyma region, and lack of glands, although the condition of having eglandular hairs on the nutlets agrees better with genera in clade b. Eurysolen is nested within the strongly supported clade of Leucosceptrum, Rostrinucula, Comanthosphace and Achyrospermum (Fig. 1A: clade b). It also shares morphological characteristics with the genera in this clade, i.e., matt and glandular nutlets, while the other main clade within Pogostemoneae (clade a) has glossy and glabrous nutlets. According to Ryding (1995), Eurysolen and most other lamioid taxa have a sclerenchyma region in the pericarp, while the schlerenchyma region is lacking in the other four genera of clade b. However, whereas the sclerenchyma forms a very distinct region in most other lamioids, it is barely distinct in Eurysolen. Hence, this clade (clade b) is supported by having the sclerenchyma region obsolete, indistinct or absent. Inclusion of the African Pogostemon aquaticus and the Asian Achyrospermum wallichianum supports the respective monophyly of these two genera across disjunct distributions. Pogostemon aquaticus has whorled leaves and forms a wellsupported clade with two Asian species with whorled leaves, while the other species have opposite leaves. The phylogenetic position of Achyrospermum wallichianum is unresolved with respect to its African relatives. The monotypic Asian genus Paralamium was not included in the current molecular phylogeny but presence of small glossy nutlets suggests that Paralamium belongs in clade a (Fig. 1A). However, until this relationship is tested using molecular data we list Paralamium as incertae sedis. Support for two Galeopsis subgenera. — The very distinctive genus Galeopsis, which is represented by nine accessions from eight of nine currently recognized species, forms a strongly supported clade (Fig. 1B). Its two subgenera, Galeopsis Rchb. and Ladanum Rchb., form two strongly supported clades (Fig. 1B: clades c and d, respectively). Phylogenetic affinity and tribal position of Hypogomphia. — The Central Asian genus Hypogomphia was listed as incertae sedis in the previous taxonomic treatment of subfamily Lamioideae (Scheen & al., 2010). In the current molecular phylogeny, Hypogomphia is nested within tribe Stachydeae and is strongly supported as the sister of Thuspeinanta brahuica (Fig. 1B). The genus resembles Thuspeinanta in being annual, having narrow leaves, 1–2-flowered cymes, and narrow oblong nutlets. The morphology of the genus does not conflict with the vague description of tribe Stachydeae in Scheen & al. (2010). Hence, Hypogomphia is included in Stachydeae. The two species of Thuspeinanta included in the present study do not group; T. persica forms a strongly supported clade with Chamaesphacos ilicifolius (Fig. 1B), rendering Thuspeinanta paraphyletic. Additional sampling of taxa within this Chamaesphacos-HypogomphiaThuspeinanta clade is needed to clarify generic delimitations. A new tribe, Paraphlomideae Bendiksby. — The E Asian incertae sedis and monotypic genera Matsumurella and Ajugoides as well as one Lamium species, L. chinense, form a clade together with Paraphlomis (Fig. 1C). Although Ajugoides, Matsumurella, Paraphlomis, and L. chinense are morphologically similar, it has not been possible to find morphological 478 TAXON 60 (2) • April 2011: 471–484 synapomorphies that support the whole group. Most of the shared morphological characteristics seem to constitute plesiomorphic character states. However, the group can be distinguished from the rest of Lamioideae by possessing a combination of features as mentioned in the description (see Taxonomic conclusions). In spite of the absence of morphological synapomorphies, we believe that the group deserves to be named on the account of support from the molecular data (Fig. 1C). A formal description of the new tribe Paraphlomideae is therefore given below (see Taxonomic conclusions). Within Paraphlomideae, Matsumurella tuberifera and two accessions of Lamium chinense form a supported group (the Matsumurella group; Fig. 1C), a relationship not only supported by morphology but also reflected in the taxonomic history. Makino (1915) described the genus Matsumurella with only one species, the E Asian M. tuberifera, but Ohwi (1965) and Murata & Yamazaki (1993) included it in Lamium. Wu & al. (1965) included both M. tuberifera and L. chinense in Galeobdolon, and described three new Chinese species G. kwangtungense, G. szechuanense and G. yangsoense. Their generic delimitation is followed in Chinese Floras, including Li & Hedge (1994). In his monograph, Mennema (1989) included the European and W Asian type species of Lamiastrum (syn. Galeobdolon) in Lamium, but omitted or excluded Wu & al.’s (1965) five E Asian Galeobdolon species. Ryding (in Harley & al., 2004) resurrected Matsumurella and suggested that all the E Asian Galeobdolon species may belong to this genus. The E Asian Galeobdolon species are very similar to each other and there is no reason to question that the five species belong in the same genus. However, this group is obviously extraneous to Lamium and Lamiastrum. It differs morphologically in having the lateral corolla lobes prominent and rounded versus triangular-acute or short and toothed in Lamium and Lamiastrum. In our molecular tree, the representatives of the Matsumurella group (Lamium chinense and Matsumurella tuberifera; Fig. 1C) do not group with the other species of Lamium (incl. Lamiastrum, here represented by L. galeobdolon), but group with Paraphlomis and Ajugoides. The Matsumurella group resembles Paraphlomis, but can be distinguished from this genus in having the calyx lobes over half as long as the tube versus less than half as long in Paraphlomis. The Matsumurella group is morphologically less distinct from the monotypic Japanese Ajugoides, with which it only groups in the non-orphan analysis and then only with low support (JK = 58%; not shown). Because Matsumurella and Ajugoides together do not form a well-supported monophyletic group in our molecular tree (Fig. 1C), we hesitate to amalgamate the two genera into one. Hence, we transfer the five E Asian species of Galeobdolon (or Lamium) to the genus Matsumurella (see Taxonomic conclusions). Since the two names Matsumurella and Ajugoides have equal priority, the genus Matsumurella may retain its name also if Ajugoides should be included based on a future study. Notes on tribe Phlomideae. — Both the phylogeny presented by Scheen & al. (2010) and the current phylogenetic update (Fig. 1) support the split of Phlomis s.l. into two genera, Phlomis s.str. and the resurrected Phlomoides, as recently suggested by Mathiesen & al. (in press). That study showed TAXON 60 (2) • April 2011: 471–484 Eremostachys, Lamiophlomis (sometimes recognized as Phlomis rotata), Notochaete hamosa and Pseuderemostachys to be nested within Phlomoides and consequently Lamiophlomis, Pseuderemostachys and the included species of Notochaete were transferred to Phlomoides (Mathiesen & al., in press). In our analyses, Eremostachys is represented by five species belonging to three sections (Phlomoides Bunge, Eremostachys and Moluccelloides Bunge), that together form a monophyletic group within Phlomoides (Fig. 1C). The fact that E. sogdiana (in E. sect. Eremostachys) and E. moluccelloides (in E. sect. Moluccelloides) form a subclade of this clade supports Ryding’s (2008) suggestion that the two sections are closely related. However, we concur with Mathiesen & al. (in press) that a more complete survey of Eremostachys is needed. Updates on tribe Leonureae: inclusion of Loxocalyx and paraphyletic Lagopsis. — The previously unplaced genus Loxocalyx is nested within a strongly supported tribe Leonureae (Fig. 1C). The two included species of Loxocalyx form a group (JK = 62% in the non-orphan analysis) with unresolved relationship to Lagopsis, Leonurus and Panzerina. The morphology of the genus does not provide much support to the molecular phylogeny, but hardly conflicts with it either. The calyces of Loxocalyx slightly resemble the calyces of many Leonureae in being zygomorphic with the abaxial lobes longer. However, the genus lacks the apparently apomorphic features that characterize many members of the tribe: the more or less palmate leaf venation, and the condition of having the stamens short or included in the corolla tube. Loxocalyx is here included in Leonureae on the basis of its molecular characters. The placement of Lagopsis in Leonureae is confirmed by adding an additional species and more DNA sequence data. However, the two species included do not group: Lagopsis marrubiastrum groups with two accessions of Panzerina lanata whereas Lagopsis supina groups with five Leonurus species (Fig. 1C). Thus, both Lagopsis and Leonurus appear to be poly-, or at the best, paraphyletic. The non-monophyly of Lagopsis and Leonurus is retained and supported in all analyses. Clades within the Chaiturus-Lagopsis-Leonurus-LoxocalyxPanzerina-group do not seem to be supported by differences in morphological characters, and the group as a whole, although morphologically rather homogeneous, does not receive strong support from molecular data. Because of this, we hesitate to propose changes in the classification. Resurrection of the genus Acanthoprasium and paraphyly of Ballota s.str. — The genus Ballota is polyphyletic as currently circumscribed (Fig. 1C) (Scheen & al., 2010). Ballota frutescens and B. integrifolia form a clade separate from the remaining species of Ballota (Fig. 1C). The two species differ from other species of Ballota in having a woody habit and spiny bracteoles versus an herbaceous habit and herbaceous bracteoles, and are therefore recognized as B. sect. Acanthoprasium. The two species also differ from most other Ballota in lacking branched hairs and having the calyces internally glabrous. Scheen & al. (2010) suggested that the B. sect. Acanthoprasium should be placed in a separate genus, but hesitated to propose this taxonomic change as the European Ballota frutescens was not included in their analysis. However, in our expanded Bendiksby & al. • Lamioideae─a taxonomic update phylogeny, both species of B. sect. Acanthoprasium are included and form a supported clade sister to all other taxa within tribe Marrubieae (Fig. 1C). Hence, we find it appropriate at this point to resurrect Acanthoprasium as a genus (see Taxonomic conclusions). In his description of B. sect. Acanthoprasium, Bentham (1832–1835) also included a species called B. forsskalii Benth., which is the type of Elbunis Raf. The latter name is older than Acanthoprasium at the rank of genus. However, the Yemeni type of B. forsskalii (Forsskål 222 p.p. at C) belongs to Leucas. As mentioned by Sebald (1978) the species should be called L. alba (Forssk.) Sebald. Thus, the correct name of the resurrected genus is Acanthoprasium. An additional accession of Ballota nigra was included in order to test the robustness of the phylogenetic position of this species as sister to a clade consisting of Marrubium and the remainder of Ballota s.str. (Scheen & al., 2010). The two B. nigra accessions do group (Fig. 1C), and the species is strongly supported as sister to Marrubium, retaining Ballota s.str. paraphyletic with respect to Marrubium, even after the exclusion of Acanthoprasium. Marrubium appears monophyletic, also with the inclusion of M. friwaldskyanum, but a more thorough study of Ballota and Marrubium is needed to sort out the generic delimitations. Circumscriptions and subgeneric classifications of Moluccella, Otostegia and Sulaimania. — Sebald (1973) recognized five sections within Otostegia: Otostegia, Isocheilos Chiov. emend. Sebald, Holophyllon Kudr. emend. Sebald, Mucrophyllon Sebald, and Chartocalyx (Regel) Chiov. emend. Kudrjaschew (1939). However, Scheen & Albert (2007, 2009) transferred the species of the O. sect. Holophyllon and sect. Isocheilos to their new genus Rydingia and O. aucheri in the monotypic O. sect. Mucrophyllon to Moluccella and re-circumscribed Otostegia to include only O. sect. Otostegia and O. sect. Chartocalyx. In our expanded Lamioideae phylogeny, Otostegia sect. Chartocalyx is represented by O. bucharica and O. olgae. These two species do not group with species of O. sect. Otostegia (Fig. 1C), but form a strongly supported clade with the monotypic Sulaimania and two accessions of Moluccella aucheri (Fig. 1C: clade e). This clade is the sister group to Moluccella s.str. (Fig. 1C: clade f). Otostegia olgae also grouped with Moluccella in the morphology-based cladogram presented by Ryding (1998). The entire Moluccella clade (Fig. 1C: clade g) is strongly supported in the molecular phylogeny presented here. The clade is also supported by morphological characters, also when other species of O. sect. Chartocalyx are included. The group can be defined by having the nutlets apically truncate, the upper lip of the corolla hardly bearded at the margin (although sometimes stated to be bearded), and the calyx zygomorphic, more or less expanded at the mouth, internally glabrous, and usually lobed with both primary and secondary lobes. There are considerable differences between the species, but these differences do not seem to be as strong as suggested by other authors (Hedge & Lamond, 1968; Sebald, 1973; Rechinger, 1982; Hedge, 1990). Since the bracteoles are mostly smaller and softer in O. sect. Chartocalyx than in Moluccella and Sulaimania, the plants of this section are often considered as non-spiny, but this difference is small and hardly consistent. The calyces of Sulaimania 479 Bendiksby & al. • Lamioideae─a taxonomic update otostegioides diverge in being smaller and less expanded at the mouth, but they are otherwise not very different from the calyces of Moluccella and O. sect. Chartocalyx. Moreover, as mentioned by Prain (1890) they resemble the calyces of M. spinosa in shape. Although differences in bracts and calyces hardly offer good diagnostic characters within the Moluccella group (Fig. 1C: clade g), its members can be divided into two distinctive subgroups on the basis of differences in habit and leaf shape. One subgroup (Fig. 1C: clade f) corresponds to Moluccella s.str., and consists of annual or short-lived herbs with toothed or incised leaves. The other subgroup (Fig. 1C: clade e) consists of shrublets with the leaves entire and coriaceous to slightly fleshy, and contains M. aucheri, O. bucharica, O. olgae and Sulaimania as well as the rest of O. sect. Chartocalyx. Monophyly of both the larger group (clade g) and its subgroups (clades e and f) is supported by both molecular and morphological data and may deserve the rank of genus. We prefer, with some hesitation, to treat the large group as a genus and the two subgroups as subgenera. Hence, the four species of Otostegia sect. Chartocalyx and Sulaimania otostegioides are transferred to Moluccella, and all five species as well as M. aucheri are included in Moluccella subg. Chartocalyx (clade e), while the remaining two species are included in M. subg. Moluccella (clade f) (see Taxonomic conclusions). The exclusion of the divergent section Chartocalyx renders Otostegia monophyletic and much more morphologically homogeneous, and reduces its geographical distribution to a smaller and less fragmented area. Although the transfer of this section to Moluccella increases the variation and distribution area of Moluccella, it is only to a moderate extent. Generic delimitation and classification of Alajja, Eriophyton, and Stachyopsis. — In the present study, duplicate accessions of three Central Asian taxa, Alajja rhomboidea, Stachyopsis oblongata and Lamium tuberosum, form a well-supported group together with the Himalayan monotypic genus Eriophyton (Fig. 1C). This group is sister to a strongly supported group of Lamium in the majority rule consensus tree (Fig. 1C), but the sister relationship is poorly supported. Lamium tuberosum, L. staintonii, L. nepalense, and L. rhomboideum (syn. Alajja rhomboideum) differ from the other species of Lamium in lacking the characteristic short and dentate lateral lobes of the corolla. In his monograph of Lamium, Mennema (1989) excluded these four species from Lamium. Unfortunately, L. staintonii and L. nepalense could not be included in the present study, but similarities in morphological characters suggest that all four deviating Lamium species are related to each other and to Eriophyton and Stachyopsis. All members of the group except for Lamium tuberosum and some Stachyopsis have anthers hairy with eglandular hairs, and their hairs differ from the anther hairs in Lamium s.str. in being shorter and not concentrated to the apices of the thecae. Lack of parsimony jackknife support for the group consisting of Eriophyton, Alajja and Lamium tuberosum (the Eriophyton group; Fig. 1C: clade h) is likely due to large stretches of missing data in the two accessions of L. tuberosum because strong support is obtained for the remaining taxa when these are excluded. Moreover, the species of the Eriophyton group (clade h) all have particularly large corollas (20–40 mm long) and a corolla tube that is much longer than the calyx. Eriophyton, Alajja, L. tuberosum 480 TAXON 60 (2) • April 2011: 471–484 and L. staintonii all lack an annulus in the corolla tube. And, as mentioned by Ryding (2003), L. tuberosum resembles Alajja in having the lateral corolla lobes emarginate, but this character is not consistent in Alajja. As mentioned by Hedge (1963), these two species also share a “scree habit”. They have woody roots, and thin rhizomes/stolons with the leaves scale-like at their base and mostly congested at their apex. Eriophyton has a similar habit but grows in a different habitat. Eriophyton and some species of Stachyopsis have a lanate indumentum, and Alajja has a similar but shorter indumentum. Hedge (1990) regarded the morphological similarities between Eriophyton and Alajja as superficial, but as the two taxa are also very similar in molecular characters, there are strong reasons to believe that the similarities reflect evolutionary relationship. Stachyopsis emerges as sister to the Eriophyton group in our molecular phylogeny (Fig. 1C), and seems to be sufficiently well-defined to be retained as a genus. However, classification of the Eriophyton group (Fig. 1C: clade h) is more problematic. It is not possible to divide the group into distinctive genera consisting of more than one species, and monotypic genera should be avoided (unless they are highly distinct), as such entities are redundant. We prefer to include all five species in one genus, although this group is morphologically rather heterogeneous. The genus is named Eriophyton as this is the oldest name in the group (see Taxonomic conclusions). Eriophyton s.l. and Stachyopsis together are the sister group to tribe Lamieae (as circumscribed in Scheen & al., 2010) (Fig. 1C). Most species of the three genera (Eriophyton s.l., Lamium s.str. and Stachyopsis) have hairy anthers. Species of Stachyopsis and Lamium s.str. all have a very broad and deeply emarginated mid-lobe of the lower lip of the corolla. Based on these morphological characteristics and the molecular data, Eriophyton s.l. (Fig. 1C: clade h) and Stachyopsis are included in the tribe Lamieae, which will now consist of three genera. Notes on Lamiaceae phylogeny. — Although Lamiaceae molecular phylogeny was not a prime target of the present study, a brief discussion is appropriate as the present study has unintentionally become the most comprehensive molecular phylogeny published to date in terms of a balanced taxon sample of Lamiaceae subfamilies as well as number of molecular markers used. Moreover, the four genetic markers used herein all differ from markers employed in previous molecular studies of Lamiaceae phylogeny (Wink & Kaufmann, 1996; Wagstaff & Olmstead, 1997; Wagstaff & al., 1998). As in these studies, the subfamilies Ajugoideae (Teucrioideae), Lamioideae, Nepetoideae and Scutellarioideae form strongly supported groups (Fig. 1A). Additionally, monophyly of the Prostantheroideae taxa included is strongly supported (Fig. 1A). Although only prostantheroid members of tribe Westringieae could be included here, monophyly of Prostantheroideae has been confirmed from molecular data by Olmstead & al. (1998). Viticoideae are non-monophyletic as currently circumscribed (Harley & al., 2004) (Fig. 1A), which has also been demonstrated by other molecular investigations (Wagstaff & Olmstead, 1997; Wagstaff & al., 1998; Bramley & al., 2009). Subfamily Symphorematoideae (here represented by Congea) is supported in all analyses as sister to a clade consisting of two viticoid genera (Petitia, Vitex) (Fig. 1A), whereas the TAXON 60 (2) • April 2011: 471–484 incertae sedis genus Tectona groups with a separate clade of viticoid genera (Fig. 1A). This corroborates the results of Bramley & al. (2009), who also included a broader viticoid taxon sample and the symphorematoid genus Sphenodesme. Lamioideae, Cymaria and the newly included incertae sedis genus Acrymia form a clade that is separated from the rest of Lamiaceae by a long and strongly supported branch (Fig. 1A). Subfamily Scutellarioideae is strongly supported as the phylogenetic sister of this clade. Surprisingly, the incertae sedis genus Garrettia emerges as the sister of the larger clade consisting of these four groups. Based on morphology, a close relationship between Cymaria, Acrymia, and Garrettia has previously been suggested (Cantino, 1992a; Harley & al., 2004), but the above-mentioned topology strongly suggests that only Cymaria and Acrymia may be included in Lamioideae. However, due to great differences in morphology, inclusion of the two genera will make the subfamily much more heterogeneous and difficult to define. Whereas Lamioideae have the ovary 4-lobed to the base, and the nutlets hardly reticulate, the two genera resemble many Ajugoideae in having the ovary shallowly 4-lobed, and the nutlets reticulate. Unfortunately, the incertae sedis genus Holocheila, which is more similar to Lamioideae in ovary and fruit characters, could not be included in our analysis. It differs from Lamioideae in having the ovary less deeply lobed, and the upper lip of the corolla shorter, but according to R.G. Olmstead (unpub.), it appears to be placed inside Lamioideae as sister to Colebrookea. Similar to earlier phylogenies of Lamiaceae (Wink & Kaufmann, 1996; Wagstaff & Olmstead, 1997; Wagstaff & al., 1998), the branches between Lamiaceae subfamilies outside of the Lamioideae-Cymaria-Acrymia-ScutellarioideaeGarrettia clade are generally considerably shorter and weaker supported than they are within this clade (Fig. 1A). Moreover, inter-relationships of subfamilies, as well as the phylogenetic position of Callicarpa, vary between analyses of the various genetic markers (not shown). As there presumably is no recombination in the chloroplast genome, this unstable result is most likely due to saturation or wrongly induced character homology, which is perhaps also indicated by the long branches leading to the terminals in this part of the phylogeny. We anticipate that increased taxon sampling will greatly improve the quality of the alignment and thereby provide a more reliable phylogeny among subfamilies outside of the Lamioideae-CymariaAcrymia-Scutellarioideae-Garrettia clade. TAxONOMIC CONCLUsIONs The following taxonomic changes are proposed as a result of our new phylogeny of Lamioideae (Fig. 1). Lamiaceae subfam. Lamioideae tribe Paraphlomideae Bendiksby, tr. nov. – Type: Paraphlomis Prain in J. Asiat. Soc. Bengal, Pt. 2, Nat. Hist. 74: 721. 1908. Herbae perennes, subfrutices vel forsan interdum herbae annuae, plerumque rhizomatosae vel in parte stoloniformes. Omnes pili non-ramosi, vix lanati. Calyces actinomorphi vel Bendiksby & al. • Lamioideae─a taxonomic update subactinomorphi, non dilatati ad orificium, 5-lobati sed lobi interdum brevissimi. Corollae valde 2-labiatae, 8–22 mm longae, labium superum extra pubescens, ad marginem vix barbatum, labium infernum 3-lobatum cum lobis prominentibus rotundatis integris. Stamina tubo corollae longiora, labio supero vix longiora, antherae glabrae vel glanduliferae cum glandulis sessilibus. Apex nucularum truncatus, subtruncatus vel inderdum rotundatus. Perennial herbs, subshrubs, or perhaps sometimes annuals, mostly rhizomatous or stoloniferous. Indumentum of unbranched hairs only, hardly lanate. Calyces actinomorphic or almost so, not expanded at the mouth, 5-lobed but lobes sometimes very short. Corolla strongly 2-lipped, 8–22 mm long, upper lip hairy, but hardly bearded along the margin, lower lip 3-lobed with the lateral lobes prominent, rounded and entire. Stamens protruding beyond the mouth of the corolla tube, but hardly longer than the upper lip, anthers glabrous or with sessile glands. Nutlets apically truncate, subtruncate or sometimes rounded. Included genera: Paraphlomis Prain, Matsumurella Makino, Ajugoides Makino. Matsumurella Makino in Bot. Mag. (Tokyo) 29: 279. 1915 ≡ Galeobdolon sect. Matsumurella (Makino) C.Y. Wu & Hsuan in Acta Phytotax. Sin. 10: 157. 1965 – Type: M. tuberifera (Makino) Makino. = Galeobdolon sect. Biflora C.Y. Wu & Hsuan in Acta Phytotax. Sin. 10: 159. 1965 – Type: G. kwantungense C.Y. Wu. Ryding’s description of Matsumurella in Harley & al. (2004) should be modified in the following respect: Perennial herbs, subshrubs or perhaps also annual herbs. Five species in China, Japan and Taiwan. Matsumurella chinensis (Benth.) Bendiksby, comb. nov. ≡ Lamium chinense Benth. in Candolle, Prodr. 12: 512. 1848 ≡ Galeobdolon chinense (Benth.) C.Y. Wu in Acta Phytotax. Sin. 10: 157. 1965. Matsumurella kwangtungensis (C.Y. Wu) Bendiksby, comb. nov. ≡ Galeobdolon kwangtungense C.Y. Wu in Acta Phytotax. Sin. 10: 160. 1965. Matsumurella szechuanensis (C.Y. Wu) Bendiksby, comb. nov. ≡ Galeobdolon szechuanense C.Y. Wu in Acta Phytotax. Sin. 10: 159. 1965. Matsumurella tuberifera (Makino) Makino in Bot. Mag. (Tokyo) 29: 279. 1915 ≡ Leonurus tuberiferus Makino in Bot. Mag. (Tokyo) 19: 146. 1905 ≡ Lamium tuberiferum (Makino) Ohwi in J. Jap. Bot. 12: 327. 1936 ≡ Lamium chinense Benth. var. tuberiferum (Makino) Murata in Acta. Phytotax. Geobot. 15: 176. 1954 ≡ Galeobdolon tuberiferum (Makino) C.Y. Wu in Acta Phytotax. Sin. 10: 158. 1965. = Lamium kelungense Hayata in Icon. Pl. Formosan. 8: 91. 1919. Matsumurella yangsoensis (Y.Z. Sun) Bendiksby, comb. nov. ≡ Galeobdolon yangsoense Y.Z. Sun in Acta Phytotax. Sin. 10: 160. 1965. 481 Bendiksby & al. • Lamioideae─a taxonomic update TAXON 60 (2) • April 2011: 471–484 Acanthoprasium (Benth.) Spenn. in T. Nees, Gen. Fl. Germ. 2: [no page number]. 1843 ≡ Ballota sect. Acanthoprasium Benth., Labiat. Gen. Spec.: 598. 1834 – Type: A. frutescens (L.) Spenn. Small shrubs, with simple hairs only. Inflorescences racemoid or thyrsoid, cymes 1–3-flowered, bracteoles in pairs, prominent, spinose. Calyx broadly campanulate, lobes spiny, 5 and subequal, or 6–10 with the secondary lobes smaller. Corolla strongly 2-lipped, tube annulate inside, upper lip entire or shallowly emarginate, hairy above, but not bearded with longer hairs at the margin, lower lip 3-lobed. Thecae ± distinct. Style-lobes slightly unequal. Nutlets apically rounded, hairy or smooth. Two species in the Maritime Alps and Cyprus. name of this labiate genus. However, the genus in Tiliaceae is called Chartacalyx, and it is questionable whether this name is similar enough to be regarded as a homonym of Chartocalyx. Here, we follow Sebald (1973), Govaerts & al. (2010) and other authors in regarding the name Chartocalyx as legitimate. Acanthoprasium frutescens (L.) Spenn. in T. Nees, Gen. Fl. Germ. 2: [no page number]. 1843 ≡ Moluccella frutescens L., Sp. Pl.: 587. 1753 ≡ Ballota spinosa Link, Handbuch 1: 475. 1829 ≡ Beringeria frutescens (L.) Rchb., Fl. Germ. Excurs.: 325. 1831 ≡ Ballota frutescens (L.) Woods, Tourist’s Fl.: 295. 1850. Moluccella fedtschenkoana (Kudr.) Ryding, comb. nov. ≡ Otostegia fedtschenkoana Kudr., Fragm. K Monogr. Otostegia: 28. 1939. Acanthoprasium integrifolium (Benth.) Ryding, comb. nov. ≡ Ballota integrifolia Benth., Labiat. Gen. Spec.: 599. 1834. = Ballota wettsteinii Rech. in Oesterr. Bot. Z. 40: 153. 1890. Moluccella L., Sp. Pl.: 587. 1753 – Type: M. laevis L. After five more species are included, Ryding’s description in Harley & al. (2004) has to be modified as follows: Perennial shrublets, or annual or short-lived herbs. Stems and leaves glabrescent or hairy with short hairs. Bracteoles more or less spiny. Calyx zygomorphic, mostly also 2-lipped, tube strongly or sometimes only slightly expanded (M. otostegioides). Eight species from S Europe to Central Asia, Pakistan and Kashmir. M. subg. Moluccella Annual or short-lived perennial herbs, glabrescent, leaves thin, almost palmately crenate, serrate or incised. Two species from southern Europe to Central Asia. Moluccella laevis L., Sp. Pl.: 587. 1753. Moluccella spinosa L., Sp. Pl.: 587. 1753. Moluccella subg. Chartocalyx (Regel) Ryding, comb. et stat. nov. ≡ Chartocalyx Regel in Trudy Imp. S.-Peterburgsk. Bot. Sada 6: 367. 1879, non Chartacalyx Maingay ex Mast. 1874 ≡ Harmsiella Briq. in Engler & Prantl, Nat. Pflanzenfam., Nachtr. 1: 291. 1897 ≡ Otostegia sect. Chartocalyx (Regel) Chiov. in Malphigia 34: 521. 1937 – Type: M. olgae (Regel) Ryding. = Sulaimania Hedge & Rech. f. in Rechinger, Fl. Iranica 150: 345. 1982 – Type: S. otostegioides (Prain) Hedge & Rech. f. Perennial shrublets, glabrescent or shortly hairy, leaves entire, coriaceous to slightly fleshy. Central Asia to Pakistan and Kashmir. Briquet (1897) regarded Chartocalyx as a younger homonym of a genus name in Tiliaceae, and proposed Harmsiella as a new 482 Moluccella aucheri (Boiss.) A.-C. Scheen in Syst. & Geogr. Pl. 77: 234. 2007 ≡ Otostegia aucheri Boiss., Diagn. Pl. Orient. 5: 40. 1844. Moluccella bucharica (B. Fedtsch.) Ryding, comb. nov. ≡ Otostegia bucharica B. Fedtsch. in Izv. Imp. Bot. Sada Petra Velikago 15: 2. 1915. Moluccella olgae (Regel) Ryding, comb. nov. ≡ Chartocalyx olgae Regel in Trudy Imp. S.-Peterburgsk. Bot. Sada 6: 368. 1879 ≡ Otostegia olgae (Regel) Korsh. in Zap. Imp. Akad. Nauk Fiz.-Mat. Otd., ser. 8, 4(4): 96. 1896 ≡ Harmsiella olgae (Regel) K. Schum. in Just’s Bot. Jahresber. 28(1): 484. 1902. Moluccella otostegioides Prain in J. Asiat. Soc. Bengal, Pt. 2, Nat. Hist. 59: 311. 1891 ≡ Sulaimania otostegioides (Prain) Hedge & Rech. f. in Rechinger, Fl. Iranica 150: 345. 1982. Moluccella sogdiana (Kudr.) Ryding, comb. nov. ≡ Otostegia sogdiana Kudr., Fragm. K Monogr. Otostegia: 24. 1939. Eriophyton Benth. in Wallich, Pl. Asiat. Rar. 1: 63. 1830 – Type: E. wallichii Benth. = Erianthera Benth. in Hooker’s J. Bot. Kew Gard. Misc. 3: 880. 1833, nom illeg., non Nees 1832 ≡ Alajja Ikonn. in Novosti Sist. Vyssh. Rast. 8: 274. 1971 ≡ Susilkumara Bennet, Indian Forester 107: 432. 1981 – Type: E. rhomboidea Benth. After four species are included in this earlier monotypic genus, Ryding’s description in Harley & al. (2004) has to be modified as follows: Perennial herbs with a woody root, and unbranched hairs. Calyx actinomorphic or almost so, subequally 5-lobed. Corolla 20–40 mm long, tube much longer than the calyx, exannulate or sometimes annulate, upper lip hairy above, not bearded along the margin, lower lip 3-lobed, mid-lobe slightly to much larger than the lateral lobes, emarginate to almost entire, lateral lobes prominent, rounded or emarginate. Nutlets apically truncate or subtruncate. Five or six species growing in alpine area at 2700–5000 m in Tadzhikistan, Afghanistan, N Pakistan, S China, Nepal and N India. Eriophyton nepalense (Hedge) Ryding, comb. nov. ≡ Lamium nepalense Hedge in Notes Roy. Bot. Gard. Edinburgh 29: 30. 1969. Eriophyton rhomboideum (Benth.) Ryding, comb. nov. ≡ Erianthera rhomboidea Benth. in Hooker’s J. Bot. Kew Gard. Misc. TAXON 60 (2) • April 2011: 471–484 3: 880. 1833 ≡ Lamium rhomboideum (Benth.) Benth., Labiat. Gen. Spec.: 509. 1834 ≡ Alajja rhomboidea (Benth.) Ikonn. in Novosti Sist. Vyssh. Rast. 8: 274. 1971 ≡ Susilkumara rhomboidea (Benth.) Bennet, Indian Forester 107: 433. 1981. = Erianthera anomala Juz. in Bot. Mater. Gerb. Bot. Inst. Komarova Akad. Nauk S.S.S.R. 15: 269. 1953 ≡ Alajja anomala (Juz.) Ikonn. in Novosti Sist. Vyssh. Rast. 8: 274. 1971. = Eriophyton afghanicum Rech. f. in Biol. Skr. 8(1): 58. 1955 ≡ Alajja afghanica (Rech. f.) Ikonn. in Novosti Sist. Vyssh. Rast. 8: 274. 1971. Govaerts & al. (2010) and some other authors treated Alajja anomala as a species, but Hedge (1990) included it as a synonym under E. rhomboideum. We tend to agree with Hedge (1990). Eriophyton staintonii (Hedge) Ryding, comb. nov. ≡ Lamium staintonii Hedge in Notes Roy. Bot. Gard. Edinburgh 29: 29. 1969. Eriophyton tuberosum (Hedge) Ryding, comb. nov. ≡ Lamium tuberosum Hedge in Notes Roy. Bot. Gard. Edinburgh 25: 49. 1963. = Lamium gilongense H.W. Li, Fl. Xizangica 4: 163. 1985 (syn. fide Yonekura, 2008). Eriophyton wallichii Benth. in Wallich, Pl. Asiat. Rar. 1: 63. 1830 ≡ E. wallichianum Hook. f., Fl. Brit. India 4: 694. 1885, orth. var. Otostegia Benth., Labiat. Gen. Spec.: 601. 1834. After four of the five sections are excluded from the genus, Budantsev’s description in Harley & al. (2004) has to be modified as follows: Bracteoles herbaceous to slightly spiny, spiny bracteoles sometimes extending to nodes of ordinary leaves. Calyx slightly to strongly zygomorphic, tube with an annulus of eglandular hairs near the mouth of its narrow proximal part. Corolla white, upper lip densely hairy and bearded with the hairs longer at the margin than on the upper surface. Nutlets rounded at the apex. About eight species, from NE Cameroun to W Saudi Arabia and Yemen, and in Egypt (Sinai). ACKNOWLEDGEMENTs The authors thank the curators at A, BHO, E, GH, L, NY, O, S, TEX, UPS, US, and WU for permission to sample from herbarium specimens used in this study, Richard G. Olmstead for a DNA sample of Gomphostemma javanicum, Janet Barber for DNAs of Sideritis, Philip D. Cantino for providing silica-dried material of Chelonopsis moschata, and Charlotte S. Bjorå for providing silica-dried material of Salvia nilotica and Leucas volkensii and for helping out with various tasks. Victor A. Albert is thanked for writing the proposal for the grant (no. 154145 from the Research Council of Norway) that has supported the present paper. Liv Borgen, Anne K. Brysting, Inger Nordal and Marte Holten Jørgensen are thanked for valuable comments on the manuscript. Finally, we are most grateful to Richard G. Olmstead, two anonymous reviewers and the editor, Mary Endress, for positive, useful and rapid feedback on our submitted manuscript. Bendiksby & al. • Lamioideae─a taxonomic update LITERATURE CITED Barber, J.C., Finch, C.C., Francisco-Ortega, J., Santos-Guerra, A. & Jansen, R.K. 2007. Hybridization in Macaronesian Sideritis (Lamiaceae): Evidence from incongruence of multiple independent nuclear and chloroplast sequence datasets. Taxon 56: 74–88. Barber, J.C., Ortega, J.F., Santos-Guerra, A., Marrero, A. & Jansen, R.K. 2000. Evolution of endemic Sideritis (Lamiaceae) in Macaronesia: Insights from a chloroplast DNA restriction site analysis. Syst. Bot. 25: 633–647. Barber, J.C., Ortega, J.F., Santos-Guerra, A., Turner, K.G. & Jansen, R.K. 2002. Origin of Macaronesian Sideritis L. (Lamioideae: Lamiaceae) inferred from nuclear and chloroplast sequence datasets. Molec. Phylog. Evol. 23: 293–306. Beardsley, P.M. & Olmstead, R.G. 2002. Redefining Phrymaceae: The placement of Mimulus, tribe Mimuleae and Phryma. Amer. J. Bot. 897: 1093–1102. Bentham, G. 1832–1835. Labiatarum genera et species. London: James Ridgway and Sons. Bramley, G.L.C., Forest, F. & de Kok, R.P.J. 2009. Troublesome tropical mints: Re-examining generic limits of Vitex and relations (Lamiaceae) in South East Asia. Taxon 58: 500–510. Bremer, B., Bremer, K., Heidari, N., Erixon, P., Olmstead, R.G., Anderberg, A.A., Källersjö, M. & Barkhordarian, E. 2002. Phylogenetics of asterids based on 3 coding and 3 non-coding chloroplast DNA markers and the utility of non-coding DNA at higher taxonomic levels. Molec. Phylog. Evol. 24: 274–301. Briquet, J. 1897. Labiatae. Pp. 290–291 in: Engler, A. & Prantl, K. (eds.), Die natürlichen Planzenfamilien, Nachträge zum II.–IV. Teil. Leipzig: Engelmann. Cantino, P.D. 1992a. Evidence for a polyphyletic origin of the Labiatae. Ann. Missouri Bot. Gard. 79: 361–379. Cantino, P.D. 1992b. Towards a phylogenetic classification of the Labiatae. Pp. 27–37 in: Harley, R.M. & Reynolds, T. (eds.), Advances in labiate science. Kew: Royal Botanic Gardens. Cantino, P.D., Harley, R.M. & Wagstaff S.J. 1992. Genera of Lamiaceae: Status and classification. Pp. 511–522 in: Harley, R.M. & Reynolds, T., Advances in labiate science. Kew: Royal Botanic Gardens. Farris, J.S., Albert, V.A., Kallersjo, M., Lipscomb, D. & Kluge, A.G. 1996. Parsimony jackknifing outperforms neighbor-joining. Cladistics 12: 99–124. Goloboff, P.A., Farris, J.S. & Nixon, K. 2003. TNT: Tree analysis using new technology. Version 1.0. http://www.cladistics.com/ aboutTNT.html. Govaerts, R., Paton, A., Harvey, Y. & Navarro, T. 2010. World Checklist of Lamiaceae & Verbenaceae. The Board of Trustees of the Royal Botanic Gardens, Kew, Richmond. http://www.kew.org/ wcsp/lamiaceae/ (accessed 10 Oct. 2010). Hall, T.A. 1999. BioEdit: A user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucl. Acids Symp. Ser. 41: 95–98. Harley, R.M., Atkins, S., Budantsev, A.L., Cantino, P.D., Conn, B.J., Grayer, R., Harley, M.M., de Kok, R., Krestovskaya, T., Morales, R., Paton, A.J., Ryding, O. & Upson, T. 2004. Labiatae. Pp. 167–275 in Kubitzki, K. & Kadereit, J.W. (eds.), The families and genera of vascular plants, vol. 7. Berlin & Heidelberg: Springer. Hedge, I.C. 1963. A new species of Lamium from Nepal. Notes Roy. Bot. Gard. Edinburgh 30: 49–50, pl. 2. Hedge, I.C. 1990. Labiatae. In: Ali, S.I. & Nasir, Y.J. (eds.), Flora of Pakistan, vol. 192. Edinburgh: Royal Botanic Gardens. Hedge, I.C. & Lamond, J.M. 1968. Studies in the flora of Afghanistan: VII. Labiatae: Lam.-end. Notes Roy. Bot. Gard. Edinburgh 28: 89–191. Kudrjaschew, S. 1939. Fragmenti k monografji roda Otostegia Benth. Trudi Bot. Sect. Kom. Nauk USSR Taschkent. Li, H. & Hedge, I.C. 1994. Lamiaceae. Pp. 50–299 in: Wu, Z & Raven, 483 Bendiksby & al. • Lamioideae─a taxonomic update P.H. (eds.), Flora of China, vol. 17. Beijing: Science Press; St. Louis: Missouri Botanical Garden Press. Lindqvist, C. & Albert, V.A. 2002. Origin of the Hawaiian endemic mints within the North American Stachys (Lamiaceae). Amer. J. Bot. 89: 1709–1724. Lindqvist, C., Motley, T.J., Jeffrey, J.J. & Albert, V.A. 2003. Cladogenesis and reticulation in the Hawaiian endemic mints (Lamiaceae). Cladistics 19: 480–495. Makino, T. 1915. Two new genera Matsumurella Makino and Ajugoides Makino. Bot. Mag. (Tokyo) 29: 379–383. Mathiesen, C., Scheen, A.-C. & Lindqvist, C. In press. Phylogeny and biogeography of the lamioid genus Phlomis (Lamiaceae). Kew Bull. Mennema, J. 1989. A taxonomic revision of Lamium (Lamiaceae). Leiden Bot. Ser. 11: 1–198. Müller, K. 2005. SeqState—primer design and sequence statistics for phylogenetic DNA data sets. Appl. Bioinformatics 4: 65–69. Murata, G. & Yamazaki, T. 1993. Lamiaceae (Labiatae). Pp. 272–321 in: Iwatsuki, K., Yamazaki, T., Bufford, D.E. & Ohba, H. (eds.), Flora of Japan, vol. 3a. Tokyo: Kodonsha. Nylander, J.A.A. 2004. MrModeltest, version 2. Program distributed by the author. Uppsala: Evolutionary Biology Centre, Uppsala University. Ohwi, J. 1965. Flora of Japan. Washington: Smithsonian Institution. Olmstead, R.G., Reeves, P.A. & Lepschi, B.J. 1998. Confirmation of a monophyletic Chloanthoideae (Lamiaceae) comprising tribes Chloantheae and Prostanthereae. Lamiales Newslett. 6: 7–10. Oxelman, B., Liden, M. & Berglund, D. 1997. Chloroplast rps16 intron phylogeny of the tribe Sileneae (Caryophyllaceae). Pl. Syst. Evol. 206: 393–410. Pan, Y.Z., Fang, L.Q., Hao, G., Cai, J. & Gong, X. 2009. Systematic positions of Lamiophlomis and Paraphlomis (Lamiaceae) based on nuclear and chloroplast sequences. J. Syst. Evol. 47: 535–542. Paton, A.J., Springate, D., Suddee, S., Otieno, D., Grayer, R.J., Harley, M.M., Willis, F., Simmonds, M.S.J., Powell, M.P. & Savolainen, V. 2004. Phylogeny and evolution of basils and allies (Ocimeae, Labiatae) based on three plastid DNA regions. Molec. Phylog. Evol. 31: 277–299. Prain, D. 1890. Some additional species of Labiatae. J. Asiat. Soc. Bengal, Pt. 2, Nat. Hist. 59: 311. Rechinger, K.H. 1982. Flora Iranica, vol. 150, Labiatae. Graz: Akademische Druck- u. Verlagsanstalt. Ronquist, F. & Huelsenbeck, J.P. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572–1574. Ryding, O. 1995. Pericarp structure and phylogeny of the LamiaceaeVerbenaceae-complex. Pl. Syst. Evol. 198: 101–141. Ryding, O. 1998. Phylogeny of the Leucas group (Lamiaceae). Syst. Bot. 13: 637–646. Ryding, O. 2003. Reconsideration of Wiedemannia and notes on the 484 TAXON 60 (2) • April 2011: 471–484 circumscription of Lamium (Lamiaceae). Bot. Jahrb. Syst. 124: 325–335. Ryding, O. 2008. Pericarp structure and phylogeny of the Phlomis group (Lamiaceae subfam. Lamioideae). Bot. Jahrb. Syst. 127: 299–316. Scheen, A.-C. & Albert V.A. 2007. Nomenclatural and taxonomic changes within the Leucas clade (Lamioideae, Lamiaceae). Syst. Geogr. Pl. 77: 229–238. Scheen, A.-C. & Albert V.A. 2009. Molecular phylogenetics of the Leucas group (Lamioideae: Lamiaceae). Syst. Bot. 34: 173–181. Scheen, A.-C., Bendiksby, M., Ryding, O., Mathiesen, C., Albert V.A. & Lindqvist, C. 2010. Molecular phylogenetics, character evolution and suprageneric classification of Lamioideae (Lamiaceae). Ann. Missouri Bot. Gard. 97: 191–219. Scheen, A.-C., Lindqvist C., Fossdal, C.G. & Albert V.A. 2008. Molecular phylogenetics of tribe Synandreae, a North American lineage of lamioid mints (Lamiaceae). Cladistics 24: 299–238. Sebald, O. 1973. Die Gattung Otostegia Bentham (Labiatae) in Afrika und auf der Arabischen Halbinsel. Stuttgarter Beitr. Naturk., A 263: 1–84. Sebald, O. 1978. Studien an afrikanischen Leucas-Arten (Labiatae). Stuttgarter Beitr. Naturk., A 308: 1–42. Shi, S.H., Du, Y.Q., Boufford, D.E., Gong, X., Huang, Y.L., He, H.H. & Zhong, Y. 2003. Phylogenetic position of Schnabelia, a genus endemic to China: Evidence from sequences of cpDNA matK gene and nrDNA ITS regions. Chin. Sci. Bull. 48: 1576–1580. Simmons, M.P. & Ochoterena, H. 2000. Gaps as characters in sequence-based phylogenetic analyses. Syst. Biol. 49: 369–381. Wagstaff, S.J., Hickerson, L., Spangler, R., Reeves, P.A. & Olmstead, R.G. 1998. Phylogeny in Labiatae s.l., inferred from cpDNA sequences. Pl. Syst. Evol. 209: 265–274. Wagstaff, S.J. & Olmstead, R.G. 1997. Phylogeny of Labiatae and Verbenaceae inferred from rbcL sequences. Syst. Bot. 22: 165–179. Wallander, E., & Albert, V.A. 2000. Phylogeny and classification of Oleaceae based on rps16 and trnL-F sequence data. Amer. J. Bot. 87: 1827–1841. Wink, M. & Kaufmann, M. 1996. Phylogenetic relationships between some members of the subfamily Lamioideae (family Labiatae) inferred from nucleotide sequences of the rbcL gene. Bot. Acta 109: 139–148. Wu, C., Li, H., Hsuan, S. & Huang, Y. 1965. Materiae ad Floram Labiatarum Sinensium (1). Acta Phytotax. Sin. 10: 143–166, pl. 29–38. Yonekura, K. 2008. Labiatae. Pp. 250–279 in: Ohba, H., Iokawa, Y. & Sharma, L.R. (eds.) Flora of Mustang, Nepal. Tokyo: Kodonsha. Yuan, Y.W., & Olmstead, R.G. 2008. A species-level phylogenetic study of the Verbena complex (Verbenaceae) indicates two independent intergeneric chloroplast transfers. Molec. Phylog. Evol. 48: 23–33.