Phylogeny of Hedyotis L. (Rubiaceae: Spermacoceae): Redefining a ...

TAXON 62 (2) • April 2013: 357–374Wikström & al. • Phylogeny of HedyotisPhylogeny of Hedyotis L. (Rubiaceae: Spermacoceae): Redefininga complex Asian-Pacific assemblageNiklas Wikström,1 Suman Neupane,2 Jesper Kårehed,3 Timothy J. Motley2 & Birgitta Bremer11 Bergius Foundation, The Royal Swedish Academy of Sciences and Department of Botany, Stockholm University,10691 Stockholm, Sweden2 Department of Biological Sciences, Old Dominion University, Norfolk, Virginia 23529-0266, U.S.A.3 The Linnaean Gardens of Uppsala, Uppsala University, Villavägen 8, 752 36 Uppsala, SwedenAuthor for correspondance: Niklas Wikström, niklas.wikstrom@bergianska.seAbstract The genus Hedyotis (Rubiaceae: Spermacoceae) has long served as a repository for tropical herbaceous species that donot fit readily into other genera. Circumscribed broadly the genus becomes a highly heterogeneous assembly, but relationshipsof Hedyotis have been difficult to resolve and it has proven very difficult to circumscribe the genus in a more narrow sense.Here we present Bayesian phylogenetic analyses of Hedyotis using plastid (rps16, petD) and nuclear (ITS, ETS) sequence datato resolve monophyletic lineages, to test former taxonomic hypotheses, and to revise the taxa within a well-supported evolutionaryframework. Four hundred and sixty-seven sequences representing 129 accessions, never previously included in anyphylogenetic analyses, are newly reported. Hedyotis, as previously circumscribed, is polyphyletic, but all investigated species,except for Hedyotis coronaria, are resolved in one of three well-supported monophyletic groups. The largest clade includesall investigated species of Hedyotis from the Indian subcontinent as well as three groups of species with primarily Chinesedistributions. The type species of Hedyotis (H. fruticosa) is resolved with the Indian subcontinent species and following previoussuggestions this group is referred to as Hedyotis s.str. Species currently recognized under the generic names Metabolosand Pleiocraterium are resolved in Hedyotis s.str. The second-largest group comprises a series of smaller, but well-supported,clades including the Leptopetalum clade, the genus Kadua, an unnamed group distributed in Asia and the Pacific, and a largeAsian group referred to here as the Exallage/Dimetia clade. The third group includes a few SE Asian Hedyotis, as well as allinvestigated species of the genus Neanotis. Hedyotis coronaria is not closely related to other species from Asia and is resolvedwith Spermacoce hispida. The analyses indicate that diplophragmous capsules and “fruticosa-type” seeds occur outside ofHedyotis s.str., and several species suggested to have these features are resolved in the Exallage/Dimetia clade. Species suggestedto have indehiscent capsules, a feature used by Bremekamp to characterize the genus Exallage, are also resolved in boththe Exallage/Dimetia clade and in Hedyotis s.str., but a close examination indicates that the capsules are not truly indehiscentin the Hedyotis s.str. species. One species of Metabolos and one species of Pleiocraterium are given new species names, andone species of Pleiocraterium is transferred to Hedyotis and three species of Hedyotis are transferred to Neanotis.Keywords Asia; Hedyotis; phylogenetic analysis; Rubiaceae; SpermacoceaeSupplementary Material The Electronic Supplement (Table S1) and the alignment file are available in the SupplementaryData section of the online version of this article (http://www.ingentaconnect.com/content/iapt/tax).Submitted: 20 June 2012; revision received; 18 Nov. 2012; accepted: 28 Jan. 2013IntroductionIn its broadest circumscription (Fosberg, 1943; Fosberg& Sachet, 1991), the genus Hedyotis L. includes 500 to 600 species,making it one of the largest genera of the coffee family(Rubiaceae). However, the genus has long served as a repositoryfor tropical herbaceous species with multiovulate loculesthat do not fit readily into other genera, and circumscribedthis broadly, Hedyotis became a highly heterogeneous group.Consequently, a large number of segregate genera have beenrecognized, often based on species confined to particular geographicregions. Most African species, for example, have generallybeen treated as Oldenlandia L., but Bremekamp (1952)recognized an additional 20 segregate genera in this group.North American species have most often been treated as HoustoniaL. (Terrell, 1975, 1991, 1996, 2001b) and South Americanspecies as Arcytophyllum Willd. ex Schult. & Schult. f. (Mena,1990). Additional small genera such as Carterella Terrell, Stenaria(Raf.) Terrell and Stenotis Terrell have been recognizedfor species from southern U.S.A. and Mexico (Terrell, 1991,2001a, c), and the generic name Kadua Cham. & Schltdl. wasrecently resurrected for Pacific, mainly Hawaiian, species (Terrell& al., 2005).Phylogenetic analyses (Bremer, 1996; Andersson & Rova,1999; Bremer & Manen, 2000; Dessein & al., 2005) haveplaced Hedyotis, and other genera of the former tribe Hedyotideae,together with herbaceous taxa from the tribe Spermacoceae.This is a tribe that has been treated both as a ratherVersion of Record (identical to print version).357

Wikström & al. • Phylogeny of HedyotisTAXON 62 (2) • April 2013: 357–374small (Bremekamp, 1952, 1966; Verdcourt, 1958; Robbrecht,1988, 1993) or as a very large group (Bremer, 1996; Bremer& Manen, 2000), and sometimes including the tribes Hedyotideae,Knoxieae, Manettieae, and Triainolepideae. Followingthe results from the most recent phylogenetic analyses(Dessein, 2003; Kårehed & Bremer, 2007; Kårehed & al., 2008;Groeninckx & al., 2009) and the most recent classification ofthe family (Bremer & Eriksson, 2009), Spermacoceae are heretreated as a tribe of ca. 1000 species and 60 genera, includingManettieae and most genera of Hedyotideae, but excludingKnoxieae and Triainolepideae (Bremer & Eriksson, 2009).Using plastid DNA data and combined plastid and nuclearrDNA data Groeninckx & al. (2009) and Kårehed & al. (2008)recently analyzed the phylogenetic relationships of Spermacoceae.Contrary to previous analyses they adopted a globalperspective on Spermacoceae and their analyses included abroad sampling of taxa covering most of the geographic andtaxonomic diversity of the group. While their analyses found alarge number of well-supported groups within Spermacoceae,they (Kårehed & al., 2008; Groeninckx & al., 2009) alsohighlighted considerable problems with our contemporarytaxonomy. Many of the smaller segregate genera, confined torestricted geographic regions, were supported as monophyletic,but at the same time their analyses indicated problemswith some of the larger, and less easily characterized, generasuch as Oldenlandia and Hedyotis. Oldenlandia in particularwas problematic and representatives included in their analysesshowed a broad range of unexpected relationships (Kårehed& al., 2008; Groeninckx & al., 2009).There was no support in the analyses by Kårehed & al.(2008) and Groeninckx & al. (2009) for circumscribing Hedyotisin a way to also include American and Polynesian taxa (Merrill& Metcalf, 1942; Fosberg, 1943; Fosberg & Sachet, 1991; Dutta& Deb, 2004). Hedyotis fruticosa L., the type species of thegenus from Sri Lanka, was grouped with strong support witha few Asian and Micronesian species, and the authors arguedthat the generic name Hedyotis probably should be restrictedto this smaller group (Kårehed & al., 2008; Groeninckx & al.,2009). It is possible that this group of Asian and Micronesianspecies corresponds to a group of species recognized alreadyby Wight & Arnott (1834: 405–418) as Hedyotis sect. DiplophragmaWight & Arn. Surveying Asian and Pacific species ofthe genera Hedyotis and Exallage Bremek., a large number ofspecies from Sri Lanka, China, and Micronesia were associatedwith this group by Terrell & Robinson (2003), and this possibilitywas also discussed briefly by Groeninckx & al. (2009).The sampling in the two studies by Kårehed & al. (2008) andGroeninckx & al. (2009) was however limited and resolvingthe taxonomic problems in this group required a more exhaustivesampling. Pleiocraterium Bremek., for example, was notincluded in their analyses and may be closely related to otherAsian species of Hedyotis (Groeninckx & al., 2009). The generaExallage and Leptopetalum Hook. & Arn. were also absent intheir analyses (Kårehed & al., 2008; Groeninckx & al., 2009).Exallage was originally described by Bremekamp (1952), andincludes several species from Asia that have been suggestedto belong to Hedyotis (Ridsdale, 1998). Leptopetalum has alsobeen associated with Hedyotis, at least in a broad sense (seeFosberg & Sachet, 1991, for a discussion). Other unsampledgroups included the genera Neanotis W.H. Lewis (ca. 30 spp.)and Metabolos Blume (1 or 2 spp.).Building on the analyses by Kårehed & al. (2008) andGroeninckx & al. (2009) we initiated a phylogenetic analysisfocusing on the Asian and Micronesian species referred to asHedyotis s.str. in their studies. Our primary aims are to bettercircumscribe the genus Hedyotis, and to identify some of therelationships within this group.Materials and MethodsTaxon sampling. — Taxa were sampled with two primaryobjectives: (1) to include as many species as possiblefrom Hedyotis s.str. (Kårehed & al., 2008; Groeninckx & al.,2009); and (2) to include as many taxa as possible from the tribeSpermacoceae that have been suggested to belong to Hedyotis,but that have not been included in any previous phylogeneticanalysis, or have not in some other way been shown to be distantlyrelated to Hedyotis.In total 203 accessions were included. The backbone ofthe datasets from Groeninckx & al. (2009) and Kårehed & al.(2008) remain, but the number of taxa from groups not in thefocus of the present analyses was significantly reduced. Sixtytwoaccessions from the analyses by Kårehed & al. (2008),twelve accessions from the analyses by Guo & al. (2011) and129 accessions not previously included in any phylogeneticanalysis were selected. Comprehensive information on speciesnames, voucher information, and references is given forincluded accessions in Table S1 (Electr. Suppl.).Following the first description of Hedyotis by Linnaeus(1753), there have been continuous disagreements over thedelimitation of the genus, particularly with respect to Oldenlandia.The two genera have often been treated as congeneric,and to make things worse, they have variously been treated aseither Hedyotis (Wight & Arnott, 1834; Fosberg, 1943; Fosberg& Sachet, 1991; Fosberg & al., 1993) or Oldenlandia (Hiern,1877; Schumann, 1891). Consequently, species described underone generic name (i.e., Hedyotis) almost certainly have a synonymunder the other (Oldenlandia) and by what name an individualspecies should be referred to is an almost arbitrarychoice. In order to be consistent, and without any preconceivedideas of what is right or wrong, we have chosen to apply thegeneric names Hedyotis and Oldenlandia to individual taxafollowing Govaerts & al. (2011).DNA extraction, amplification, and sequencing. — Totalgenomic DNA was extracted from silica-dried material and/or herbarium specimens using a standard cetyl trimethyl ammoniumbromide (CTAB) protocol (Doyle & Doyle, 1987),combined with the QIAquick PCR cleaning kit (Qiagen, Hilden,Germany) following the protocol specified by the manufacturer.Amplification and sequencing of the plastid regions rps16and petD, and of the internal and external transcribed spacers(ITS, ETS) of the nuclear rDNA were carried out using primerslisted in Table 1. Amplification reactions were performed358 Version of Record (identical to print version).

TAXON 62 (2) • April 2013: 357–374Wikström & al. • Phylogeny of HedyotisTable 1. Primers used for amplification and sequencing of new sequences in this study.Region Primer Primer sequence from the 5′ end Referencerps16rps16_Frps16_R2GTG GTA GAA AGC AAC GTG CGA CTTTCG GGA TCG AAC ATC AAT TGC AACOxelman & al. (1997)Oxelman & al. (1997)petDPlpetB1365FPlpetD738RTTGACYCGTTTTTATAGTTTACAATTTAGCYCTTAATACAGGLöhne & Borsch (2005)Löhne & Borsch (2005)ITSITS_P17ITS_P2526S_82RCTA CCG ATT GAA TGG TCC GGT GAAGGG TAG TCC CGC CTG ACC TGTCC CGG TTC GCT CGC CGT TAC TAPopp & Oxelman (2001)Popp & Oxelman (2001)Popp & Oxelman (2001)ETS18S-E18S-ETSHedETS-EritGCA GGA TCA ACC AGG TAG CAACT TAC ACA TGC ATG GCT TAA TCTTGG WTA GCA CGG TTT GGT TGG ABaldwin & Markos (1998)Baldwin & Markos (1998)newly designedusing Phusion enzyme (Finnzymes, Espoo, Finland) in 20 µlreactions; template DNA 1 µl, 5× Phusion HF Buffer 4 µl,10 mM dNTP 0,5 µl each, 10 µM forward and reverse primers0,5 µl each, Phusion enzyme 0,2 µl. The thermal cyclingprofile was set to 98°C 2 min (98°C 15 s, 60°C 15 s, 72°C30 s) × 45, 72°C 7 min. PCR products were cleaned using theMultiScreen Separations System (Millipore, Billerica, Massachusetts,U.S.A.), sequenced with the BigDye terminator cyclesequencing kit using the amplification primers, and analyzedon an ABI PRISM 3100 Genetic Analyzer (Applied Biosystems,Foster City, California, U.S.A.).Sequence assembly, editing and phylogenetic analyses.— Sequences were edited and assembled using the StadenPackage (Staden, 1996; Staden & al., 2000) and Seaview v.4.3.1(Gouy & al., 2010). All four regions were aligned using theprogram MUSCLE v.3.8.31 (Edgar, 2004) with default settings.Before sequences were submitted to MUSCLE for alignmentthey were sorted by size using the program USEARCH v.5.2.32(Edgar, 2010).Markov chain Monte Carlo (MCMC) methods (Larget& Simon, 1999) within a Bayesian framework were used toapproximate the posterior distribution of trees using MrBayesv.3.1.2 (Ronquist & Huelsenbeck, 2003). In the MCMC the datawas partitioned into 3 partitions (rps16, petD, rDNA) and eachpartition was allowed partition-specific parameters (Ronquist& Huelsenbeck, 2003; Nylander & al., 2004). The nucleotidesubstitution model for each partition was selected based on acorrected Akaike information criterion (AICc) as calculatedusing MrAIC v.1.4.4 and PHYML v.2.4 (Guindon & Gascuel,2003; Nylander, 2004). The GTR + Γ model was selected forthe plastid partitions rps16 and petD whereas GTR + I + Γ wasselected for the rDNA partition.Bayesian analyses were run for 20 million generations withfour MCMC chains in two independent and parallel runs. Followinga “burn-in” phase of 10 million generations, we sampledtrees and parameters every 2000th generation leaving a finalposterior distribution of 5000 trees and parameter estimates ineach run. Convergence of the two runs was confirmed (standarddeviation of split frequencies below 0.01 during the last10 milj. generations) and the two posterior distributions of treesand parameters were pooled and used to calculate Bayesianposterior probabilities.ResultsSequence data from the plastid regions rps16 and petD andthe nuclear regions ITS and ETS were successfully generatedfor 129 accessions for which sequence data have not previouslybeen reported. In total 125 sequences of rps16, 106 sequencesof petD, 117 sequences of ITS, and 119 sequences of ETS werenewly generated. Sequences are deposited at the EMBL NucleotideSequence Database and their EMBL accession numbersare reported in Table S1 (Electr. Suppl.). The compiled datasetused in the phylogenetic analyses comprised 203 taxa and 4308characters and is available in the Supplementary Data sectionof the online version of this article.Phylogenetic relationships indicated by the MCMC analysesare summarized as a 50% majority-rule consensus tree inFigures 1 to 3, and posterior probability values greater than 0.50are reported below each node. Nodes with posterior probabilityvalues equal to or greater than 0.95 (the node appears in atleast 95% of the sampled trees) are considered well-supported(Alfaro & al., 2003). A large number of monophyletic groupsand relationships in Spermacoceae are identified and supportedby the phylogenetic analyses. Some correspond to those alreadyidentified by previous analyses, but others are new and identifiedfor the first time in the analyses presented here. Groupsidentified and supported are referred to by the names indicatedin Figures 1 to 3. A summary of groups identified (and possiblesubgroups), their habit, fruit and seed morphology, anddistribution is also given in Table 2.The genus Hedyotis as previously circumscribed is polyphyletic.All investigated species recognized under Hedyotisby Govaerts & al. (2011), except for Hedyotis coronaria (Kurz)Craib that resolves with Spermacoce hispida L. (Spermacoceclade; Fig. 3), are resolved in one of three monophyletic clades(clades A–C; Figs. 1–3; Table 2), all with Asian or Asian-Pacific distributions. The first clade (clade A; Figs. 1, 2) iswell-supported by the analyses (BPP = 1.00) and includes themajority of Hedyotis investigated and a large number of speciescurrently recognized under the generic name Oldenlandiaby Govaerts & al. (2011). We will from here on refer to thisgroup as Hedyotis s.str. Included in Hedyotis s.str. are all Indiansubcontinent species of Hedyotis that were investigated(including the type species H. fruticosa), as well as the majorityVersion of Record (identical to print version).359

Wikström & al. • Phylogeny of HedyotisTAXON 62 (2) • April 2013: 357–374Fig. 1. Phylogenetic relationships in theRubiaceae tribe Spermacoceae. Speciestraditionally recognized under the genericname Hedyotis are resolved in three differentclades: clade A (Hedyotis s.str.; Fig. 2);clade B (Fig. 3); and clade C (Neanotis). Thetree is a 50% majority-rule consensus treefrom a Bayesian Markov chain Monte Carlo(MCMC) analysis of a combined dataset ofplastid (rps16, petD) and nuclear (ITS, ETS)data. Bayesian posterior probabilities areindicated at the nodes..991.75.94.971.98.86.93Carphalea madagascariensisParaknoxia parvifloraBatopedina pulvinellataClade C11111111111111.77111 Neanotis indica111 Hedyotis pahompokae.72.89 1.99111.96.73.96Neanotis gracilis1.87 Neanotis hirsuta.59 Hedyotis lindleyanaFig. 3Clade A1Neanotis wightianaNeanotis nummulariformisNeanotis nummularia1Hedyotis trichocladaHedyotis nanaNeanotis monospermaNeanotis calycinaNeanotis formosanaOldenlandia rupicolaDibrachionostylus kaessneriOldenlandia nervosaOldenlandia geophilaOldenlandia echinulosa1Kohautia coccinea.94Oldenlandia fastigiata1 Agathisanthemum bojeri1Mitrasacmopsis quadrivalvisHedythyrsus spermacocinusOldenlandia microthecaStenaria nigricansHoustonia longifoliaHoustonia caeruleaArcytophyllum thymifoliumPentodon pentandrusArcytophyllum muticumHedyotis s.str. (Fig. 2)Dentella repensKohautia cynanchicaOldenlandia unifloraOldenlandia trinerviaOldenlandia rosulataOldenlandia herbaceaOldenlandia goreensisOldenlandia angolensisPentanopsis fragransAgathisanthemum quadricostatumAgathisanthemum globosumOldenlandia strictaOldenlandia affinis bZ039Oldenlandia affinisAgathisanthemum chlorophyllumLelya osteocarpaLelya prostrataAmphiasma merenkyanumArcytophyllum−Neanotis DibrachionostylusHoustonia clade Kohautia Pentanopsis clade Agathisanthemum clade Outgroups(Asia)(Africa)(Africa)(America)(Africa/Asia) (Africa/Asia)(Africa/Asia)360 Version of Record (identical to print version).

TAXON 62 (2) • April 2013: 357–374Wikström & al. • Phylogeny of HedyotisClade A1.91.911.62.951.93111.951.86.99.99.921.86Oldenlandia ternata1Oldenlandia tetrangularis98 Hedyotis parryiPleiocraterium verticillare"Pleiocraterium"Hedyotis purpurascensHedyotis swertioides1Hedyotis stylosaHedyotis articularisOldenlandia benguetensisHedyotis sp. cB011Hedyotis decoraHedyotis sp. cA050Oldenlandia pubescensOldenlandia nutans1.84.961 .98.6311.991.9811.971111Hedyotis valetonianaHedyotis schlechteriHedyotis pulchellaHedyotis macrostegia1Hedyotis rigidaHedyotis philippensisHedyotis fissistipulaMetabolos rugosusMetabolos decipiens cC0171Metabolos decipiens cC016Hedyotis nodulosaHedyotis rhinophyllaHedyotis fruticosa"Metabolos"Hedyotis trimeniiHedyotis obscuraHedyotis lawsoniaeHedyotis thwaitesiiHedyotis quinquenerviaHedyotis marginataHedyotis lessertiana var. marginataHedyotis lessertiana var. lessertianHedyotis gardneriiHedyotis flavescens1Hedyotis dendroidesHedyotis tridentata1 Hedyotis coprosmoidesHedyotis communisHedyotis sp. bZ0241 Hedyotis sp. bZ022Oldenlandia prostrataHedyotis minutopuberulaOldenlandia cryptanthaOldenlandia hainanensisHedyotis cheniana1Hedyotis cathayanaHedyotis cantoniensis cA056Oldenlandia lancea1.57 Hedyotis caudatifolia cB027.54 Hedyotis caudatifoliaOldenlandia paridifolia TM31671 Oldenlandia paridifolia TM3165Oldenlandia melliOldenlandia assimilis cB055Oldenlandia assimilis.981.90.801.741.561.90Oldenlandia consanguinea aT079Oldenlandia consanguinea.731 Hedyotis caudatifolia cB037Hedyotis effusaOldenlandia tenuipesOldenlandia tenuipes bZ037Hedyotis effusaOldenlandia pulcherrimaHedyotis yangchunensisHedyotis xinyiensis11.99.52.99.54.99.99Hedyotis korrorensisHedyotis shiuyingiaeHedyotis shenzhenensis1Hedyotis ovataHedyotis uncinellaHedyotis rivalisHedyotis sp. SKP946Hedyotis novoguinensisHedyotis megalanthaOldenlandia bodinieriOldenlandia vachelliiOldenlandia bracteosaHedyotis acutangulaHedyotis cantoniensis(Indian subc.) (Indian subc.) (China/Vietnam)(China)(China)Hedyotis membranaceaHedyotis s.str.(Asia)Fig. 2. Phylogenetic relationships of Hedyotis s.str. The two small genera Metabolos (2 spp.) and Pleiocraterium (4 spp.) are both resolved wellinside Hedyotis s.str. The results also indicate biogeographical patterns in Hedyotis s.str. Species from the Indian subcontinent are resolved intwo different groups, and three different groups with primarily Chinese distributions are indicated. The tree is a 50% majority-rule consensustree from a Bayesian Markov chain Monte Carlo (MCMC) analysis of a combined dataset of plastid (rps16, petD) and nuclear (ITS, ETS) data.Bayesian posterior probabilities are indicated at the nodes.Version of Record (identical to print version).361

Wikström & al. • Phylogeny of HedyotisTAXON 62 (2) • April 2013: 357–374.86Clade BFig. 3. Phylogenetic relationships in the Rubiaceaetribe Spermacoceae. A large proportion of speciestraditionally recognized under the generic nameHedyotis are resolved in the Exallage/Dimetia clade.The Asian species Hedyotis coronaria is unrelated toother Asian Hedyotis and groups with Spermacocehispida. The tree is a 50% majority-rule consensustree from a Bayesian Markov chain Monte Carlo(MCMC) analysis of a combined dataset of plastid(rps16, petD) and nuclear (ITS, ETS) data. Bayesianposterior probabilities are indicated at the nodes.11.70.56362 Version of Record (identical to print version)..991111111111.71.71Synaptantha tillaeaceaOldenlandia cherevensisOldenlandia mitrasacmoides11.991.97.98Oldenlandia pteritaOldenlandia biflora"Thecagonum"Oldenlandia biflora KarehedLeptopetalum grayiiLeptopetalum1 Leptopetalum foetidumKadua parvulaKadua fluviatilisKadua affinisOldenlandia verticillataOldenlandia pinifoliaOldenlandia gracilipesOldenlandia stocksiiOldenlandia lancifolia1Oldenlandia galioides.71Oldenlandia herbacea cB058.991 Oldenlandia diffusa1 Oldenlandia brachypodaOldenlandia tenelliflora cA089Oldenlandia tenelliflora1.74Oldenlandia angustifoliaOldenlandia ovatifoliaHedyotis pachycarpaOldenlandia chrysotricha bZ04011 Oldenlandia chrysotricha bR0991 Oldenlandia chrysotricha bR100Oldenlandia lineata cB0101 1 Oldenlandia lineata bS006.61 Oldenlandia lapeyrousii.901.59.991Oldenlandia auricularia cC021Oldenlandia auricularia cB013Oldenlandia auricularia cC020Hedyotis vestitaHedyotis costataHedyotis scandensHedyotis capitellata bZ031Hedyotis capitellata bR098Hedyotis dianxiensis1Oldenlandia contracta.99 .97Oldenlandia scabra cB025.92 Hedyotis sp. bZ0131 Hedyotis sp. bZ0201 Oldenlandia hedyotidea1 .54 Hedyotis sp. bZ025Hedyotis sp. bZ021.65 1Hedyotis pitardianaHedyotis sp. bZ0181 Hedyotis obliquinervisHedyotis amplifloraOldenlandia tenuis1Oldenlandia salzmanniiSpermacoce hispidaHedyotis coronariaOldenlandia umbellata cC0251 Oldenlandia umbellata SN84Oldenlandia wiedemanniiOldenlandia taborensis.99 Oldenlandia duemmeriOldenlandia wauensisOldenlandia erecta cB02411 1 Oldenlandia erecta SN2Oldenlandia corymbosa.95Oldenlandia nematocaulis1 Oldenlandia densa.77Oldenlandia capensisCordylostigma virgataCordylostigma microcalaExallageDimetia(Asia/Pacific) (Pacific)Leptopetalumclade KaduaSpermacoce clade"Unnamed group"Exallage/Dimetia cladeOldenlandia s.str.(Asia/Africa/(Asia/Pacific) (Asia) Australia/America) (Africa)(Africa)Cordylostigma

TAXON 62 (2) • April 2013: 357–374Wikström & al. • Phylogeny of Hedyotisof Hedyotis species from temperate and tropical Asia. Speciesfrom the Indian subcontinent are resolved in two separategroups. The first consists of four species (Hedyotis purpurascensHook. f., Hedyotis swertioides Hook. f., Hedyotis stylosaR. Br. ex G. Don, Hedyotis articularis R. Br. ex G. Don) and theonly representative of the genus Pleiocraterium (BPP = 1.00)included in the study. The second group includes the remainingIndian subcontinent species of Hedyotis, but also two speciesof the genus Metabolos (BPP = 1.00).Three separate clades with primarily Chinese distributionsare also identified in Hedyotis s.str. One clade includes sevenChinese species (H. communis–H. cathayana; Fig. 2) and twounidentified accessions from Vietnam (Hedyotis sp. bZ024and Hedyotis sp. bZ022). Together they form a well-supportedgroup (BPP = 1.00). The second and largest clade includes 20accessions (H. cantoniensis cA056–H. xinyiensis; BPP = 1.00;Fig. 2), representing at least 13 different species. The last cladeincludes six species (Hedyotis shiuyingiae–Hedyotis acutangula;BPP = 1.00; Fig. 2). All three of these groups also includespecies that Govaerts & al. (2011) currently recognizes underthe generic name Oldenlandia.Species placed in Hedyotis by Govaerts & al. (2011) arealso resolved in a second large group with an Asian and Pacificdistribution (clade B; Fig. 3). This lineage comprises three wellsupportedand monophyletic clades resolved in a trichotomy.The first is the Leptopetalum clade + the genus Kadua (BPP =1.00), the second is an unnamed group distributed in Asia andthe Pacific (BPP = 1.00), and the third is an Asian group (BPPTable 2. A summary of the three clades (A–C) in which Hedyotis species are resolved. The summary indicates groups identified by the analysesand that are indicated on Figures 1to 3, subgroups that are discussed in the text and their habit, fruit and seed morphology, and distribution.CladeClade A(Fig. 2)Clade B(Fig. 3)Clade C(Fig. 1)GroupsidentifiedHedyotis s.str.LeptopetalumcladeKaduaExallage/DimetiacladeNeanotisSubgroupsdiscussedLeptopetalum“Thecagonum”(see text)ExallageDimetiaHabit and othermorphology Fruit dehiscence Seed DistributionSuffrutescentherb to shurb orrarely small treeSlender, glabroussubshrubsSmall annual orperennial herbsShruby to smalltrees, corollasalverform,fleshy, longtubed, andappendagedSuffrutescentherb with axialinflorescencesHerbs or shrubs,lianescent,climbing andscandentAnnual herbs,pluriaperturatepollenCapsules apex not protudingbeyond calyx lobes; septicidaldehiscence usually followedby a partial apical loculicidaldehiscence, usually resultingin two semi-split valvesCapsules subglobose, firmand thick-walled; dehiscenceloculicidal from apexCapsules subglobose towinged with somewhatcompressed walls; thin andfragile or firm and thickwalled;dehiscence loculicidalfrom topAll the taxa in subg. Kadua,except sect. Oceanica, havecapsules with initial loculicidaldehiscence from the apexfollowed by septicidal dehiscenceat maturity; in subg.Gouldia and subg. Kaduasect. Oceanica the fruits arefleshy and indehiscentCapsules indehiscentCapsule apex protruding beyondcalyx lobes; dehiscenceloculicidally from apexfollowed by partial septicidaldehiscenceCapsule subglobose;dehiscence loculicidallyfrom apexDorsiventrallycompressed, with aconspicuous hilumSlightly compressedinto irregularpolyhedrons withrounded angles andcornersGlobose orsub-globose withdeeply pittedexotestaVarious: fan-shaped,ovoid, flat withbroad wing orbrick-likeTrigonous (oldenlandioid)Dorsiventrallycompressed, with anapical centric hilumand sometimeswinged marginCymbiformto shallowlycup-shapedSri Lanka, India,SE China,Indo-China,Malesia, Papuasia,NW PacificPacificTropical Asia,tropical Australiaand PacificHawaiian Islandsand French PolynesiaTropical Asia,tropical Australiaand PacificTropical Asia,Malesia, Papuasiaand PacificThe information was compiled using Fosberg & Sachet (1991), Terrell & Robinson (2003, 2007), Terrell & al. (2005), Dutta & Deb (2004), Chen& Taylor (2011), Neupane & al. (2009) and from our personal observations.Version of Record (identical to print version).363

Wikström & al. • Phylogeny of HedyotisTAXON 62 (2) • April 2013: 357–374= 1.00), from here on referred to as the Exallage/Dimetia clade.The Leptopetalum clade includes Asian and Pacific species ofOldenlandia, sometimes recognized under the generic nameThecagonum Babu (see Discussion), as well as two representativesof the genus Leptopetalum. The Exallage/Dimetia cladecomprises a number of Asian species recognized either underHedyotis or Oldenlandia by Govaerts & al. (2011). Sister tothe Exallage/Dimetia clade is the Asian species Oldenlandiaovatifolia (Fig. 3).The third, and last group (clade C; Fig. 1) includesHedyotis trichoclada Merr. & L.M. Perry and H. nana Merr.& L.M. Perry from New Guinea, H. pahompokae Fukuokafrom Thailand, and H. lindleyana Hook. ex Wight & Arn. fromAsia and they are resolved with representatives of the genusNeanotis (BPP = 1.0; Fig. 1). We will from here on refer to thisgroup as Neanotis (see Nomenclatural changes outside Hedyotiss.str. below). Sister to Neanotis is an African group includingthe monotypic Dibrachionostylus Bremek. and Oldenlandiarupicola (Sond.) Kuntze.The analyses also provide continued support for groupspreviously identified by Kårehed & al. (2008) and Groeninckx& al. (2009). These groups include the Arcytophyllum-Houstonia clade, Kohautia Cham. & Schltdl., the Pentanopsisclade, the Agathisanthemum clade (Fig. 1), CordylostigmaGroeninckx & Dessein and Oldenlandia s.str. (Fig. 3).DiscussionIn their phylogenetic analyses of the tribe Spermacoceaeboth Kårehed & al. (2008) and Groeninckx & al. (2009) retrievedand discussed a number of large and well-supportedclades: Kohautia, the Pentanopsis clade, the Agathisanthemum–Hedyotiss.str. clade, Kadua, the Arcytophyllum-Houstoniaclade, Oldenlandia s.str., “Pachystigma”, and the Spermacoceclade. With one exception, these groups are also retrievedand supported in the present analyses. Following Groeninckx& al., (2010c) Pachystigma is here referred to by the genericname Cordylostigma. The only group not retrieved in the presentstudy is the Aganthisanthemum–Hedyotis s.str. clade, aclade that comprised two main components, Hedyotis s.str. anda group of two African genera, Agathisanthemum Klotzsch andLelya Bremek., and three species of Oldenlandia. We will referto this latter assemblage as the Agathisanthemum clade (Fig. 1).Both Hedyotis s.str. (Fig. 2) and the Agathisanthemum cladeare retrieved and well-supported in our analyses, but there isno support for their sister-group relationship as indicated byKårehed & al. (2008) and Groeninckx & al. (2009). One reasonbehind this lack of support is that two plastid regions, thetrnL-F intergenic spacer and the atpB-rbcL intergenic spacer,were not included in the present analyses. Reanalysis of thedataset from Kårehed & al. (2008), excluding these two plastidregions, resulted in relationships consistent with those obtainedhere (i.e., with no support for the Agathisanthemum–Hedyotiss.str. clade).Three out of four species of Oldenlandia that resolve in theAgathisanthemum clade (O. trinervia Retz., O. goreensis (DC.)Summerh., O. angolensis K. Schum.), are from Oldenlandiasubg. Anotidopsis (Hook. f.) K. Schum., and this relationshipwas seen and discussed previously by Kårehed & al. (2008) andGroeninckx & al. (2009). They debated if the three species ofOldenlandia they resolved with Agathisanthemum and Lelyashould be transferred to a new genus or treated as membersof Agathisanthemum, and if this treatment should apply onlyto the three species or to the entire subgenus. Our analysesadded O. trinervia from O. subg. Anotidopsis that also cameout in the Agathisanthemum clade, but at the same time ourresults clearly indicate that the subgenus is polyphyletic. Furthercomplicating this issue is that other species included in ouranalyses from O. subg. Anotidopsis, such as Neanotis hirsuta,and Hedyotis lindleyana, (Schumann, 1891), are resolved in thegenus Neanotis (see below).Much uncertainty reported in the analyses by Kårehed& al. (2008) and Groeninckx & al. (2009) remains. In Kårehed& al. (2008), for example, Spermacoceae was resolved in a basaldichotomy, with a well-supported clade (their clade A), comprisingKohautia and the Pentanopsis clade, sister to a poorlysupported group (their clade B) including remaining taxa. Ouranalyses continue to support a sister relationship between Kohautiaand the Pentanopsis clade (PP 1.00), but clade B (sensuKårehed & al., 2008) is not retrieved at all in our analyses. Ouranalyses resolve Spermacoceae in a basal dichotomy with awell-supported group (BPP = 1.00) that corresponds to clade D(sensu Kårehed & al., 2008), and this group is sister to a poorlysupported group (BPP = 0.75) comprising four subgroups:(i) Hedyotis s.str. (clade A; BPP = 1.00); (ii) Kohautia + thePentanopsis clade (BPP = 1.00); (iii) the Agathisanthemumclade (BPP = 1.00); and (iv) Pentodon Hochst. + Dentella Forst.(BPP = 1.00). The second group, Kohautia + the Pentanopsisclade, corresponds to clade A (sensu Kårehed & al., 2008).Although these differences are substantial, they all concernrelationships that were poorly supported in the analyses byKårehed & al. (2008) and Groeninckx & al. (2009), and thatremain poorly supported in the present analyses. Additionaldata is required before these relationships can be resolved.Hedyotis s.str. — Hedyotis, as previously circumscribed,is resolved as polyphyletic. All investigated species recognizedunder Hedyotis by Govaerts & al. (2011), except for Hedyotiscoronaria, are grouped in one of three monophyletic groups(clades A–C; Figs. 1–3). With few exceptions these groups onlyinclude species from Asia and the Pacific. The first of thesegroups (clade A; Figs. 1, 2) includes the majority of the speciesin Hedyotis investigated, and more or less corresponds to agroup that has been recognized for a long time and diagnosedby fruit characteristics. Originally the group was describedby Wight & Arnott (1834) who characterized the capsules ashaving an “apex more or less 4-lobed and slightly producedbeyond the calyx-tube, at length septicidal, splitting to the baseinto two somewhat bony cocci”. They included seven speciesin this group, all from the Indian peninsula (Wight & Arnott,1834). The key feature of the capsules, separating them fromthose in other Hedyotis, is that they dehisce along the septum(septicidal dehiscence) and separate into two distinct valves orcocci (Terrell & Robinson, 2003; Neupane & al., 2009). Wight364 Version of Record (identical to print version).

TAXON 62 (2) • April 2013: 357–374Wikström & al. • Phylogeny of Hedyotis& Arnott (1834) named the type of dehiscence diplophragmousand described the group as Hedyotis sect. Diplophragma.Following the initial description, the group has consistentlybeen recognized by later workers, although at various taxonomiclevels. Bentham & Hooker (1873) included 24 species intheir H. sect. Diplophragma and also divided the group furtherbased on features of the stipules. Hooker (1880), in his Flora ofBritish India, included 30 species under H. sect. Diplophragma.Schumann (1891), also recognized the group, but included itunder a broadly defined Oldenlandia as O. sect. Diplophragma.More recently the group has been elevated and recognizedunder Hedyotis at the rank of subgenus, as H. subg. Diplophragma(Wight & Arn.) Fosberg (Fosberg, 1943), or H. subg.Hedyotis (Terrell & Robinson, 2003), or at the rank of genus,Diplophragma (Wight & Arn.) Meisner (Devi & Raju, 1990).The name Diplophragma, irrespective of rank, can not be maintainedunder Hedyotis, and this was noted by Fosberg & Sachet(1991). Hedyotis auricularia L., the species originally proposedas lectotype for Hedyotis (Chamisso & Schlechtendal, 1829;Hitchcock, 1929), was for good reason dismissed by Bremekamp(1939, 1952) who instead suggested H. fruticosa as thetype species. Bremekamp’s suggestion was adopted by Jarvis(1992), who proposed H. fruticosa as the conserved type ofHedyotis. Their proposal was subsequently approved (Barrie,2006) and a nomenclatural consequence of this is that section(or subgenus) Diplophragma, if including H. fruticosa, willhave to be referred to as section (or subgenus) Hedyotis. Terrell& Robinson (2003) accepted this consequence and includedsection Diplophragma and subgenus Diplophragma as synonymsunder Hedyotis subg. Hedyotis. Kårehed & al. (2008)and Groeninckx & al. (2009) took this one step further and suggestedthat the generic name of Hedyotis should be restrictedto this smaller group. Following their suggestion we continueto refer to this group as Hedyotis s.str.In addition to having diplophragmous capsules the grouphas been characterized as having “fruticosa-type” seeds. Thename for this type of seeds was originally coined by Terrell& Robinson (2003) and refers to seeds that are dorsiventrallyflattened, lenticular, and with an irregularly narrow and winglikemargin, resembling those in H. fruticosa. Surveying Asianand Pacific Hedyotis, Terrell & Robinson (2003) found thatmost of the Sri Lankan and Micronesian species have bothdiplophragmous capsules and fruticosa-type seeds, and theyincluded these species under Hedyotis subg. Hedyotis. Also, alarge proportion of species recognized under Hedyotis in theFlora of China have been associated to this group, either by explicitlydescribing the capsules as having septicidal dehiscence(Chen & Taylor, 2011), or by including the species in Hedyotissect. Diplophragma (Lo & al., 1999).Our analyses provide strong support for the monophyly ofHedyotis s.str. (clade A; BPP = 1.00). However, species characterizedas having indehiscent fruits such as Oldenlandia paridifolia(Dunn) Chun, O. prostrata (Blume) Kuntze, O. cryptantha(Dunn) Chun, Hedyotis philippensis (Willd. ex Spreng.) Merr.ex C.B. Rob., and H. rigida (Blume) Walp., and that previouslyhave been associated to O. auricularia (L.) F. Muell.(see the Exallage/Dimetia clade below), are here resolved inHedyotis s.str. Upon close examination of the fruits in thesespecies we note that they are not truly indehiscent, as seenamong species in the Exallage/Dimetia clade, but retain suturesthat with mechanical pressure open septicidally. All Hedyotiss.str. species may therefore still satisfy the diplophragmouscharacter. A complication though is that some species, characterizedto have diplophragmous capsules (i.e., Hedyotisscandens Roxb., H. capitellata Wall. ex G. Don, H. amplifloraHance, H. dianxiensis W.C. Ko, Oldenlandia hedyotidea (DC.)Hand.-Mazz.), show close relationships to species with trulyindehiscent fruits, and resolve in the Exallage/Dimetia clade(Fig. 3). Most of these species were also explicitly stated tohave “fruticosa type” of seeds by Terrell & Robinson (2003),indicating a complication also in the seed character. Species,with diplophragmous capsules, and that group in the Exallage/Dimetia clade, were all included in Hedyotis sect. DimetiaWight & Arn. and not in Hedyotis sect. Diplophragma by Lo& al. (1999). Hedyotis sect. Dimetia has been characterized ashaving “tumescens”, a swollen capsular top, protruding beyondthe calyx (Terrell & Robinson, 2003). Perhaps this featurecan be used to help discriminate between species with diplophragmouscapsules that group in Hedyotis s.str., and thosethat group in the Exallage/Dimetia clade.The analyses indicate the presence of biogeographic patternswithin Hedyotis s.str. (Fig. 2). Species from the Indiansubcontinent are resolved in two different groups, both wellsupported(BPP = 1.00). The first and smallest group, the oneincluding “Pleiocraterium”, is placed as an early-diverginggroup in Hedyotis s.str., but relationships in this part of the treeare poorly supported. The second group is the largest of theIndian subcontinent groups and is placed sister to a group variouslydistributed in China, Indochina, Malesia, and Papuasia.Three distinct and well-supported groups with primarily Chinesedistributions are also retreived in our analyses.“Metabolos” and “Pleiocraterium”. — Included in Hedyotiss.str. are species currently recognized under the genericnames Metabolos and Pleiocraterium. Metabolos is a smallAsiatic genus whose relationship to other Rubiaceae has beendifficult to understand. Hooker (1873), for example, initiallyassociated Metabolos with Lasianthus Jack (tribe Lasiantheae;Bremer & Manen, 2000), but a few years later emended thisview and included the genus in Hedyotideae (Hooker, 1880).Schumann (1891) also found similarities with Lasianthus andincluded Metabolos in Psychotrieae, whereas Bremekamp(1939), like Hooker (1880), suggested a placement in Hedyotideae.More recently Puff & Igersheim (1994) surveyed thecharacter states and taxonomic position of Metabolos. Theyacknowledged a striking habitual similarity between Metabolosand species of Hedyotis, yet they found no support for associatingMetabolos with the tribe Hedyotideae (Puff & Igersheim,1994). Based on what they perceived as congruent characterstates in inflorescence arrangement, ovary structure and placentation,fruit structure and anatomy, and pollen morphology,they agreed with Schumann (1891) and argued for aclose relationship to Lasianthus (Psychotrieae). Piesschaert& al. (2001) on the other hand, considered a relationship toHedyotideae a possibility based on palynological data. TheyVersion of Record (identical to print version).365

Wikström & al. • Phylogeny of HedyotisTAXON 62 (2) • April 2013: 357–374looked at photographs of Metabolos pollen, as reported by Puff& Igersheim (1994), and found the pollen to show a complexreticulum, consisting of a psilate suprareticulum and spinulateinfrareticulum, a type of reticulum that they associated withSpermacoceae sensu Bremer (1996).The results presented here unequivocally resolve this issueand clearly place Metabolos well inside Hedyotis s.str.and as sister to Hedyotis nodulosa Arn., a species from SriLanka to which it bears a great resemblance (Puff & Igersheim,1994). Our analyses included three different accessionsof Metabolos, two from Sri Lanka representing M. decipiens(Thwaites) Ridsdale, and one from Java, Indonesia representingM. rugosus Blume, but whether or not the Sri Lankan and theJavanese Metabolos truly represent two different species hasbeen debated (Puff & Igersheim, 1994). Our analyses indicateconsiderable sequence divergence between the two Sri Lankanaccessions and the accession from Indonesia (Fig. 2) providingsome support for maintaining two species. Govaerts & al.(2011) accept two additional names, Metabolos angustifoliusDC. and Metabolos macrophyllus Zoll. & Moritzi, but bothwere excluded from Metabolos already by Hochreutiner (1934)and this was followed by Bremekamp (1952) who included bothspecies under the generic name Exallage.Piesschaert & al. (2001) investigated morphological andanatomical features in the small Neotropical genus Pagameopsis(Standl.) Steyerm. and found considerable similaritiesbetween Pagameopsis and Metabolos. They argued thatMetabolos should be included in future phylogenetic (e.g.,molecular) analyses as a possible relative of Pagameopsis(Piesschaert & al., 2001). Given that our analyses placedMetabolos well inside Hedyotis s.str. we will include Pagameopsisin future analyses of Hedyotis to further evaluate theseputative similarities.The genus Pleiocraterium was described by Bremekamp(1939) and included four species, and all are currentlyrecognized under Pleiocraterium by Govaerts & al. (2011).Pleiocraterium verticillare (Wall. ex Wight & Arn.) Bremek.from India, the type species, was originally described by Wight& Arnott (1834) as Hedyotis verticillaris Wight & Arn. They,as well as Hooker (1880), included H. verticillaris in H. sect.Diplophragma. Bremekamp (1939) considered this positionanomalous, and listed a series of features separating this speciesfrom others in H. sect. Diplophragma. Features listed includethe presence of an ovary protruding distinctly beyondthe insertion of the calyx, presence of axillary, and not terminal,inflorescences, and the occurrence of strikingly differentvegetative characters in H. verticillaris (Bremekamp,1939). Because none of these characters occur in species thatBremekamp associated with sect. Diplophragma, he consideredthe creation of a separate genus fully justified. Based on thenumerous cups that are formed by the connate leaf bases inH. verticillaris, he named the genus Pleiocraterium. One additionalspecies, P. plantaginifolium (Arn.) Bremek. from SriLanka, had previously been described under Hedyotis and wastransferred to Pleiocraterium by Bremekamp (1939), but he alsodescribed two additional species, P. sumatranum Bremek. andP. gentianifolium Bremek., both restricted to Sumatra.Our analyses clearly place Pleiocraterium in Hedyotiss.str., and the species included in our analyses, P. verticillare,groups with other species from India including H. purpurascens,H. articularis, H. swertioides, and H. stylosa (Fig. 2).This result is not entirely surprising and although Pleiocrateriumwas not included in their analyses, Groeninckx & al.(2009) discussed this possibility. Bremekamp (1939) found thedistribution of Pleiocraterium peculiar and compared it to thatof Metabolos, the only genus in Rubiaceae he could think ofwith a similar distribution. The relationships of Metabolos indicatethat the group possibly originated in Sri Lanka (Indiansubcontinent) and subsequently spread to SE Asia, and althoughno SE Asian representatives of Pleiocraterium were includedin our analyses, a corresponding pattern is hypothesized alsofor this group.“Hedyotis” resolved outside of Hedyotis s.str. — Speciespreviously included in Hedyotis are resolved in two additionalclades (clades B and C), both quite large and not closely relatedto Hedyotis s.str. Species in these clades, with few exceptions,are distributed in Asia and the Pacific.Clade B. — The largest clade (clade B; Fig. 3) comprises aseries of smaller and well-supported monophyletic groups including:the Leptopetalum clade (BPP = 1.00); the genus Kadua(BPP = 1.00); an unnamed group including Asian, Pacific, andone African species of Oldenlandia (BPP = 1.00); and a largeAsian group that we refer to as the Exallage/Dimetia clade(BPP = 1.00). Exceptions to the Asian-Pacific pattern of distributionare found in clade B and include Synaptantha tillacea(F. Muell.) Hook. f. and Oldenlandia mitrasacmoides(F. Muell.) F. Muell., both from Australia, and Oldenlandialancifolia (Schumach.) DC. from Africa.The Leptopetalum clade. — The genus Leptopetalum wasdescribed in 1838 by W.J. Hooker and G.A. Walker-Arnott intheir Botany of Captain Beechey’s Voyage (Hooker & Walker-Arnott, 1841). Their description was based on a specimencollected from the Bonin Islands, but the specimen was incorrectlysupposed to have been collected in Mexico and thespecies they described was given the rather misleading nameLeptopetalum mexicanum Hook. & Arn. Hooker (1873) andSchumann (1891), in their general treatments of Rubiaceae, recognizedthe group, but included the name Leptopetalum under amore broadly defined Hedyotis (Hooker, 1873) or Oldenlandia(Schumann, 1891). More recently, and while preparing a treatmentof the family Rubiaceae for the “Flora of Micronesia”,Fosberg & Sachet (1991) reviewed the taxonomic history ofLeptopetalum and undertook an evaluation of its generic status.Following this evaluation, they recognized Leptopetalumat the subgeneric level and included it under a broadly definedHedyotis (Fosberg & Sachet, 1991).Six species were included in their Hedyotis subg. Leptopetalumand the group was characterized as woody, with largesalverform, funnelform, or campanulate corollas, with stamensinserted on the corolla tube or near its base, and distributedin the south, central and western Pacific (Fosberg & Sachet,1991). The species that they included correspond, with oneexception, to those currently recognized under Leptopetalumby Govaerts & al. (2011).366 Version of Record (identical to print version).

TAXON 62 (2) • April 2013: 357–374Wikström & al. • Phylogeny of HedyotisRecent phylogenetic analyses by Kårehed & al. (2008) andGroeninckx & al. (2009) indicated a close relationship betweenOldenlandia biflora L. and the genus Kadua, a genus thatTerrell & al. (2005) recently resurrected for Pacific species previouslyincluded in Hedyotis. Their results clearly showed thatO. biflora needed to be transferred from Oldenlandia s.str., butthey considered it premature to include it in a more broadly circumscribedKadua. Our analyses provide support for their hesitationto recircumscribe Kadua, and instead group O. biflorawith representatives of the genus Leptopetalum. Our resultsalso define an additional member of this lineage, O. pterita(Blume) Miq., and these species form a well-supported group(PP 1.00) that is resolved sister to the genus Kadua. Kaduacould of course still be recircumscribed to include O. biflora,but this would require that species currently recognized underLeptopetalum also be included.The two species, O. biflora and O. pterita have previouslybeen considered closely related (Fosberg & Sachet, 1991) andboth were included under the generic name Thecagonum Babuby Babu (1969) together with O. ovatifolia (Cav.) DC. and oneadditional Oldenlandia species not included in our analyses(O. parishii Hook. f.). Investigating seed and capsule morphologyin different genera in Hedyotideae Terrell & Robinson(2007) found support for this grouping, and they added oneadditional species (Oldenlandia strigulosa DC.) to the group.However, based on considerable differences in its seed morphology,they excluded O. ovatifolia. Our results lend supportfor this exclusion, and there is no support for a close relationshipbetween O. ovatifolia and remaining species included underThecagonum by Terrell & Robinson (2007). The analysesplace O. ovatifolia in a very isolated position, as sister to theExallage/Dimetia clade (Fig. 3).Additional analyses should be undertaken before taxonomicdecisions are made in the Leptopetalum clade. Morphologically,the small and herbaceous species that have beenrecognized under Thecagonum are quite distinct, and they havenever before been associated with the genus Leptopetalum.Recircumscribing Leptopetalum to also include these speciesis for this reason not necessarily a good option and would makeLeptopetalum more problematic to diagnose. An alternativesolution would be to recognize two genera, Leptopetalum andThecagonum, but at present there is no support for a monophyleticThecagonum (Fig. 3). In fact, the species Oldenlandiabiflora is resolved as paraphyletic indicating that there are taxonomicproblems in this species. These problems are not entirelyunexpected and Fosberg & Sachet (1991) discussed at lengththe extensive amount of morphological variation seen in thisspecies across its geographic distribution.The “unnamed group” from Asia and the Pacific. — Sisterto the Leptopetalum clade and Kadua is a heterogenic group thatis difficult to characterize. The group includes not only speciesthat never before have been considered closely related, butalso species commonly associated to other groups such as theExallage/Dimetia clade or Hedyotis s.str. For example, Oldenlandiadiffusa (Willd.) Roxb. and O. brachypoda DC. have beenconsidered closely related to O. corymbosa L. and O. erecta(Manilal & Sivar.) R.R. Mill (Sivarajan & Biju, 1990; Chen& Taylor, 2011), but our analyses clearly indicate that they arenot. Furthermore, Oldenlandia verticillata L. from Asia hasoften been associated to O. auricularia (Fosberg & Sachet, 1991),but our analyses group O. verticillata with O. pinifolia (Wall.ex G. Don) Kuntze, and not together with O. auricularia (in theExallage/Dimetia clade). Oldenlandia verticillata, together withO. pinifolia, is sister to six (possibly seven) species of Oldenlandia(O. gracilipes Craib, O. stocksii Hook. f., O. lancifolia(Schumach.) DC., O. galioides (F. Muell.) F. Muell., O. diffusa,O. brachypoda). The seventh species, Oldenlandia herbacea (L.)Roxb. (cB058), is a Chinese accession, determined as O. herbacea,which is resolved in this group. However, the accessiondoes not group with O. herbacea from Africa (in the Pentanopsisclade), and clearly represents some other species. Chen & Taylor(2011), in their treatment of Hedyotis for the Flora of China,pointed out that they had seen very few specimens of this taxonfrom China, or anywhere else east of Sri Lanka. They also notedthat the description of the species, given in the Chinese versionof the Flora of China by Lo & al. (1999), indicated features notin agreement with specimens of H. herbacea, but that are applicableto the similar species H. brachypoda and H. diffusa (Chen& Taylor, 2011). Our results are completely consistent with theirobservations and group the Chinese O. herbacea with O. brachypodaand O. diffusa. The possibility that O. herbacea does notoccur in China has to be considered, and Chinese specimens ofthis taxon should be carefully reexamined.The Exallage/Dimetia clade. — The generic name Exallagewas introduced by Bremekamp (1952) in his work on AfricanOldenlandia. Following a discussion concerning thetypification of the genus Hedyotis, he concluded that Hedyotisauricularia, a species primarily found in Asia, needed tobe transferred to another genus. He dismissed the possibilityof including it under some generic name already in use, anddescribed the genus Exallage, with Exallage auricularia (L.)Bremek. as the type species. Although primarily focused onAfrican taxa, he provided a preliminary list of an additional 23non-African species in this genus (Bremekamp, 1952). Ridsdale(1998), however, considered the genus ill-defined and arguedthat Bremekamp made his combinations without properly examiningthe taxa, and that many belong to Hedyotis s.str.Surveying Asian and Pacific species of Hedyotis andExallage, Terrell & Robinson (2003) summarized the infragenericclassification and species groups recognized by Ridsdale(1998) and Lo & al. (1999), and after having examined alarge number of species and specimens from the herbarium atthe Smithsonian Institution (US), they characterized Hedyotisand Exallage with respect to their fruit and seed morphology.Exallage was characterized as: large perennial herbs; with glomerulateinflorescences; with small globose indehiscent fruits;and with seeds similar to those found in Oldenlandia subg.Oldenlandia (with oldenlandioid seeds). Contrary to Bremekamp(1952) they included Exallage as a subgenus under Oldenlandia,and stressed that indehiscent fruits seemed to be themain defining character of the group, a feature not found inother Oldenlandia (Terrell & Robinson, 2003).Our analyses found strong support for a group of species(the Exallage/Dimetia clade), associated with OldenlandiaVersion of Record (identical to print version).367

Wikström & al. • Phylogeny of HedyotisTAXON 62 (2) • April 2013: 357–374(Hedyotis) auricularia and that do not group with neitherHedyotis s.str. nor Oldenlandia s.str. (Fig. 3). Correspondingrelationships were indicated already by Kårehed & al. (2008)who resolved H. capitellata outside of Hedyotis s.str. and inassociation with several species of Oldenlandia (their cladeH). The entire group can, however, not be unequivocally characterizedas having indehiscent fruits. Species such as Hedyotisscandens, H. capitellata, H. ampliflora, H. dianxiensis,and Oldenlandia hedyotidea have all been characterized ashaving diplophragmous capsules (Terrell & Robinson, 2003;Chen & Taylor, 2011). These capsular species are resolved asa monophyletic group (“Dimetia” Fig. 3; BPP = 0.99), but despitehaving diplophragmous capsules they were not includedin H. sect. Diplophragma by Lo & al. (1999), who classifiedthem under H. sect. Dimetia. Sister to this group is a wellsupportedgroup (“Exallage” Fig. 3; BPP = 1.00), characterizedby having indehiscent fruits, and with the exception of Oldenlandiachrysotricha (Palib.) Chun, all species in this group havepreviously been included under the generic name of Exallage.There are some problems in the Exallage/Dimetia cladethat involve the determination of specimens included in ouranalyses, and the synonymy of species by Govaerts & al. (2011).For example, Govaerts & al. (2011) include H. rigida (Blume)Walp. under the accepted name Oldenlandia cristata (Willd.ex Roem. & Schult.) ined. together with Hedyotis vestita R. Br.ex G. Don, Hedyotis costata (Roxb.) Kurz, and a large numberof additional names, and Chen & Taylor (2011) includeH. philippensis (Willd. ex Spreng.) Merr. ex C.B. Rob. as asynonym under H. prostrata Blume. Our analyses do placespecimens determined as Hedyotis vestita R. Br. ex G. Don andHedyotis costata (Roxb.) Kurz together, as implied by the synonymy.They are resolved inside the Exallage/Dimetia clade,as suggested by Terrell & Robinson (2003), but they are notclosely related to H. rigida (Blume) Walp.. Also, the specimendetermined as H. philippensis (Willd. ex Spreng.) Merr. exC.B. Rob., as well as the one determined as H. prostrata Blume(determined by C. Taylor), are both placed inside Hedyotiss.str., but they are not sister species, as would be expected.Instead, H. philippensis (Willd. ex Spreng.) Merr. ex C.B. Rob.is grouped together with H. rigida (Blume) Walp.Clade C (Neanotis). — The second clade, in which speciespreviously included in Hedyotis are resolved, is clade C (Fig. 1).Four species of Hedyotis, H. trichoclada and H. nana fromNew Guinea, H. pahompokae from Thailand, and H. lindleyanafrom Asia, are resolved together with all included representativesof the Asian genus Neanotis. Following the new combinationsmade (see Nomenclatural changes outside Hedyotis s.str.below) we will refer to this clade as the genus Neanotis.The generic name Neanotis originates from Lewis (1966)and was adopted as a substitute name for non-American taxapreviously included under the generic name Anotis DC. Reviewingthe taxonomic history of Anotis DC., Lewis (1966)noted that taxonomists in the 19th and 20th centuries also includedAsian species from the tribes Hedyotideae (Hooker,1873, 1880) and Oldenlandieae (Schumann, 1891) under thisname. Species included were characterized as herbaceous andwith few and peltate, cymbiform seeds. Going back to theoriginal description of Anotis, Lewis noted that all species ofAnotis (sensu Candolle, 1830) were native to the New Worldand concluded that there was no justification for an inclusion ofspecies from Asia. Following the taxonomists of the 19th and20th centuries, who had recognized the Asian Anotis as distinctfrom Hedyotis or Oldenlandia, he transferred and included theAsian species under the new name Neanotis.Approximately 30 species, distributed in temperate andtropical Asia, are recognized in the genus by Govaerts & al.(2011). They have cymbiform seeds (Schumann, 1891; Lewis,1966), similar to those found in the genus Houstonia (Terrell& Robinson, 2007), and the possibility of Neanotis being theclosest non-American relative of the Arcytophyllum-Houstoniaclade was discussed briefly by Groeninckx & al. (2009).Our analyses provide strong support for monophyly ofNeanotis (including the former Hedyotis species), but do notindicate a close relationship of this group to the Arcytophyllum-Houstonia clade. Sister to Neanotis is a group including twoAfrican species, Dibrachionostylus kaessneri (S. Moore)Bremek. and Oldenlandia rupicola (Sond.) Kuntze (see belowfor a discussion of Dibrachionostylus), and together they aresister to a large group in which the Arcytophyllum-Houstoniaclade is nested. However, the relationship to this large group, inwhich the Arcytophyllum-Houstonia clade is nested, is not wellsupported(BPP = 0.94). Sister to the Arcytophyllum-Houstoniaclade is a group comprising two small African genera, MitrasacmopsisJovet (2 spp.) and Hedythyrsus Bremek. (1 sp.), andfour African species of Oldenlandia (O. fastigiata Bremek.,O. echinulosa K. Schum., O. geophila Bremek., O. nervosaHiern). This last group corresponds to clade G sensu Kårehed& al. (2008) and relationships in this group, as well as in theArcytophyllum-Houstonia clade were discussed exhaustivelyby them.Hedyotis pahompokae was described by Fukuoka (1969)based on collections from Thailand, and is grouped with strongsupport within Neanotis (Fig. 1). Fukuoka (1970), in his “Contributionsto the Flora of Southeast Asia”, associated H. phompokaewith Hedyotis nalampooni Fukuoka and Oldenlandiakrewanhensis Pierre ex Pit. It is possible that all three speciesshould be transferred to Neanotis, but the relationshipsof H. nelampooni and H. krewanhensis should be confirmedin analyses before such a transfer is completed. Groupedwithin Neanotis are also two species of Hedyotis from NewGuinea, H. nana and H. trichoclada, both described by Merrill& Perry (1945), who considered them closely allied and readilydistinguished from other species of Hedyotis by their minuteleaves and prostrate habit. While discussing H. trichoclada,they commented that “although the habit of this plant suggestsAnotis, it has all the characters of Hedyotis”. Contrary to thisview, our analyses strongly support a grouping within Neanotis.Hedyotis lindleyana is indicated as a heterotypic synonym toNeanotis hirsuta by Govaerts & al. (2011), and our analysesalso group H. lindleyana (accession collected in Japan) withN. hirsuta from Nepal.Lewis (1966) characterized Neanotis palynologically ashaving brevicolporate pollen with (5)6–12 apertures, and sawa distinct line of demarcation in the pollen morphology of368 Version of Record (identical to print version).

TAXON 62 (2) • April 2013: 357–374Wikström & al. • Phylogeny of HedyotisNeanotis on the one hand and Hedyotis and Oldenlandia onthe other. This indicates that pollen morphology could potentiallybe used to characterize the Neanotis clade, and speciesof Hedyotis that are resolved in the Neanotis clade should beinvestigated for their pollen morphology in order to evaluatethis possibility.Dibrachionostylus. — The genus Dibrachionostylus wasdescribed by Bremekamp (1952). He separated Oldenlandiakaessneri S. Moore from other species of Oldenlandia basedon its capsule dehiscence and suggested a possible relationshipof O. kaessneri to species in the genus Agathisanthemum.However, differences in style, corolla tube, and seed testa ledBremekamp to assign O. kaessneri to a segregate genus whichhe named Dibrachionostylus. Recent phylogenetic analyseshave not supported a close relationship between Dibrachionostylusand Agathisanthemum. Groeninckx & al. (2009) resolvedDibrachionostylus as sister to a clade of African Oldenlandiaincluding O. echinulosa, O. geophila, and O. nervosa. Correspondingrelationships were also reported by Groeninckx & al.(2010a, b), although a different relationship now was indicatedfor O. nervosa. In both analyses the support for this placementof Dibrachionostylus was weak. Kårehed & al. (2008) resolvedDibrachionostylus as sister to a large clade (their clade E), butthe two accessions of D. kaessneri were not closely associatedwith any other African taxon included in their analyses.Our analyses support a relationship of Dibrachionostylusas sister to the African species Oldenlandia rupicola. Togetherthese two species are resolved as sister to the Asian genusNeanotis (Fig. 1), but this relationship is not well-supported(BPP = 0.94). The relationship of Dibrachionostylus to Oldenlandiarupicola is interesting and provides the first supportfor associating Dibrachionostylus with any other Africantaxon. Oldenlandia rupicola was classified in Oldenlandiasubg. Orophilum Bremek. by Bremekamp (1952), and he specificallyassociated O. rupicola with O. greenwayi Bremek.,O. muscosa Bremek., and O. tenella Kuntze. However, noneof these have been included in any phylogenetic analyses, andO. geophila, the only other representative of Oldenlandia subg.Orophilum included here, do not show a close relationship toO. rupicola (Fig. 1).The grouping of O. rupicola with Dibrachionostylus addsfurther evidence to the problematic and polyphyletic natureof African Oldenlandia. In previous analyses (Kårehed & al.,2008; Groeninckx & al., 2009), as well as in the analyses presentedhere, Oldenlandia species from Africa are resolved in noless than five different groups. The association of O. rupicolato Dibrachionostylus and their relationships to Asian Neanotisadds a sixth lineage associated with African Oldenlandia.ConclusionsThe phylogenetic analyses presented resolve all investigatedspecies of Hedyotis, except H. coronaria, in one ofthree clades (clades A–C; Figs. 1–3), each well-supported as amonophyletic group, each with an Asian-Pacific distribution,and each including a fair number of species recognized undersome other generic name by Govaerts & al. (2011). Hedyotiscoronaria is unrelated to other Hedyotis and groups withSpermacoce hispida. A summary of the habit, fruit- and seedmorphology, and distribution of all three clades is given inTable 2. All the major clades and sub-groups listed in Table 2are currently being investigated further prior to formal taxonomicrecognition. Many of the clades and sub-groups in thetable will require generic status. In addition to Hedyotis s.str.and Neanotis defined herein, the clades representing Dimetia,Exallage, Leptopetalum, and “Thecagonum” will likely requireelevation to the generic level. Based on geographic samplinggaps in the present study we were concerned that hasty revisionscould lead to inaccurate classifications which have beenall too recurrent in past treatments of these lineages. Presentlydetailed morphological and anatomical studies are being conductedon fruit, seed, and pollen and will be combined withmore comprehensive and cosmopolitan sampling of species forphylogenetic analyses. These species-level investigation on ourproposed generic assemblages will allow us to better define therange, composition and morphological synapomorphies delineatingthe monophyletic lineages. While we have been able toresolve many relationships and new monophyletic lineages, thisdiverse and widespread lineage of Rubiaceae still requires moreinvestigation as many regions and taxa remain unsampled.Because this group contains species that can vary from treesto tiny herbs yet harbors cryptic species much effort will berequired before a clear picture can be drawn about the evolutionof the members of the Hedyotis-Oldenlandia complex.Nomenclatural treatmentSpecies recognized under Hedyotis s.str.The following list includes species names and their homotypicsynonyms that we, based on the results presented here,recognize under Hedyotis s.str.Hedyotis acutangula Champ. ex Benth. in Hooker’s J. Bot.Kew Gard. Misc. 4: 171. 1852 ≡ Oldenlandia acutangula(Champ. ex Benth.) Kuntze, Revis. Gen. Pl. 1: 292. 1891.Hedyotis articularis R. Br. ex Wight & Arn., Prodr. Fl. Ind.Orient. 1: 407. 1834 ≡ Hedyotis articularis R. Br. exG. Don, Gen. Hist. 3: 527. 1834.Hedyotis assimilis Tutcher in Rep. Bot. Dept. Hong Kong 1914:32. 1915 ≡ Oldenlandia assimilis (Tutcher) Chun in Sunyatsenia1: 310. 1934.Hedyotis benguetensis (Elmer) Elmer in Leafl. Philipp. Bot.3: 976. 1911 ≡ Oldenlandia benguetensis Elmer in Leafl.Philipp. Bot. 1: 18. 1906.Hedyotis bodinieri H. Lév. in Repert. Spec. Nov. Regni Veg.11: 64. 1912 ≡ Oldenlandia bodinieri (H. Lév.) Chun inSunyatsenia 1: 310. 1934.Version of Record (identical to print version).369

Wikström & al. • Phylogeny of HedyotisTAXON 62 (2) • April 2013: 357–374Hedyotis bracteosa Hance in J. Bot. 23: 323. 1885 ≡ Oldenlandiabracteosa (Hance) Kuntze, Revis. Gen. Pl. 1: 292. 1891.Hedyotis cantoniensis F.C. How ex W.C. Ko in J. S. ChinaAgric. Univ. 16: 42. 1995.Hedyotis cathayana W.C. Ko, Fl. Hainan. 3: 579. 1974.Hedyotis caudatifolia Merr. & F.P. Metcalf in J. Arnold Arbor.23: 228. 1942.Hedyotis ceylanica (Thwaites) N. Wikström & Neupane,nom. nov. ≡ Allaeophania decipiens Thwaites, Enum. Pl.Zeyl. 147. 1859 [non Hedyotis decipiens (Valeton) Merr.& L.M. Perry in J. Arnold Arbor. 26: 2. 1945] ≡ Metabolosrugosus var. decipiens (Thwaites) Hochr. in Candollea5: 280. 1934 ≡ Metabolos decipiens (Thwaites) Ridsdalein Blumea 41: 460. 1996 – Lectotype (designatedhere): Sri Lanka, without date, Thwaites CP-3093 (BRno. BR0000005575176 [image!]; isolectotypes: GH no.GH00092309 [n.v.], NY no. NY130803 [image!]).Hedyotis cheniana R.J. Wang in Novon 18: 264. 2008.Hedyotis communis W.C. Ko, Fl. Hainan. 3: 579. 1974.Hedyotis consanguinea Hance in Ann. Sci. Nat., Bot., ser.4, 18: 221. 1862 ≡ Oldenlandia consanguinea (Hance)Kuntze, Revis. Gen. Pl. 1: 292. 1891.Hedyotis coprosmoides Trimen, Syst. Cat. Fl. Pl. Ceylon: 42.1885.Hedyotis cryptantha Dunn in Bull. Misc. Inform. Kew 1912:367. 1912 ≡ Oldenlandia cryptantha (Dunn) Chun in Sunyatsenia1: 311. 1934.Hedyotis decora Geddes in Bull. Misc. Inform. Kew 1931:219. 1931.Hedyotis dendroides Alston in Trimen, Handb. Fl. Ceylon 6:147. 1931.Hedyotis effusa Hance in J. Bot. 17: 11. 1879 ≡ Oldenlandiaeffusa (Hance) Kuntze, Revis. Gen. Pl. 1: 292. 1891.Hedyotis fissistipula Merr. in J. Malayan Branch Roy. Asiat.Soc. 1: 38. 1923.Hedyotis flavescens Thwaites, Enum. Pl. Zeyl.: 141. 1859.Hedyotis fruticosa L., Sp. Pl.: 101. 1753 ≡ Oldenlandia fruticosa(L.) Kuntze, Revis. Gen. Pl. 1: 292. 1891.Hedyotis gardneri Thwaites, Enum. Pl. Zeyl.: 142. 1859 ≡ Oldenlandiagardneri (Thwaites) Kuntze, Revis. Gen. Pl. 1:292. 1891.Hedyotis gentianifolia (Bremek.) N. Wikström & Neupane,comb. nov. ≡ Pleiocraterium gentianifolium Bremek. inMeded. Bot. Mus. Herb. Rijks Univ. Utrecht 56: 441. 1939– Holotype: North Sumatra, Country of the Gajos, summitof Goh Lembuh, alt. 3000 m, 20 Feb 1937, van Steenis9038 (BO [n.v.]).Hedyotis hainanensis (Chun) W.C. Ko, Fl. Hainan. 3: 580.1974 ≡ Oldenlandia hainanensis Chun in Sunyatsenia 1:311. 1934 ≡ Oldenlandia oligantha Chun in Sunyatsenia1: 313. 1934 ≡ Hedyotis oligantha Merr. in Philipp. J. Sci.23: 266. 1923, nom. illeg. [non Hedyotis oligantha Merr.in Philipp. J. Sci. 17: 431. 1921 (‘1921’)].Hedyotis korrorensis (Valeton) Hosok. in Trans. Nat. Hist. Soc.Taiwan 24: 204. 1934 ≡ Oldenlandia korrorensis Valetonin Bot. Jahrb. Syst. 63: 294. 1930.Hedyotis lancea Thunb. ex Maxim. in Bull. Acad. Imp.Sci. Saint-Pétersbourg 29: 161. 1883 ≡ Oldenlandialancea (Thunb. ex Maxim.) Kuntze, Revis. Gen. Pl. 1: 292.1891.Hedyotis lawsoniae Wight & Arn., Prodr. Fl. Ind. Orient.1: 407. 1834 ≡ Oldenlandia lawsoniae (Wight & Arn.)Kuntze, Revis. Gen. Pl. 1: 292. 1891.Hedyotis lessertiana Arn. in Nova Acta Acad. Caes. Leop.-Carol. German. Nat. Cur. 18: 339. 1836 ≡ Oldenlandia lessertiana(Arn.) Kuntze, Revis. Gen. Pl. 1: 292. 1891.Hedyotis leuserensis N. Wikström & Neupane, nom. nov. ≡Pleiocraterium sumatranum Bremek. in Meded. Bot. Mus.Herb. Rijks Univ. Utrecht 56: 440. 1939 [non Hedyotissumatrana Merr. in Pap. Michigan Acad. Sci. 23: 197.1938] – Holotype: North Sumatra, Country of the Gajos,G. Losir, alt. 3250–3450 m, 1 Feb 1937, van Steenis 8560(BO [n.v.]; isotype: BM [image!]).Hedyotis macrostegia Stapf in Trans. Linn. Soc. London, Bot.4: 170. 1894.Hedyotis marginata (Thwaites ex Trimen) Alston in Trimen,Handb. Fl. Ceylon 6: 147. 1931 ≡ Hedyotis lessertiana var.marginata Thwaites ex Trimen, Handb. Fl. Ceylon 2: 309.1894.Hedyotis megalantha Merr. in Philipp. J. Sci., C 9: 143. 1914≡ Oldenlandia megalantha (Merr.) Valeton in Bot. Jahrb.Syst. 63: 298. 1930.Hedyotis mellii Tutcher in Rep. Bot. Dept. Hong Kong 1914: 32.1915 ≡ Oldenlandia mellii (Tutcher) Chun in Sunyatsenia1: 313. 1934.Hedyotis membranacea Thwaites, Enum. Pl. Zeyl.: 143. 1859.370 Version of Record (identical to print version).

TAXON 62 (2) • April 2013: 357–374Wikström & al. • Phylogeny of HedyotisHedyotis minutopuberula Merr. & F.P. Metcalf in J. ArnoldArbor. 23: 229. 1942.Hedyotis nodulosa Arn. in Nova Acta Acad. Caes. Leop.-Carol. German. Nat. Cur. 18: 340. 1836.Hedyotis novoguineensis Merr. & L.M. Perry in J. ArnoldArbor. 26: 4. 1945.Hedyotis nutans (Valeton) P. Royen, Alpine Fl. New Guinea4: 2718. 1983 ≡ Oldenlandia nutans Valeton in Gibbs, Fl.Arfak Mts.: 180. 1917.Hedyotis obscura Thwaites, Enum. Pl. Zeyl.: 141. 1859 ≡ Oldenlandiaobscura (Thwaites) Kuntze, Revis. Gen. Pl. 1:292. 1891.Hedyotis ovata Thunb. ex Maxim., Bull. Acad. Imp. Sci. Saint-Pétersbourg 29: 161. 1883 ≡ Oldenlandia ovata (Thunb. exMaxim.) Kuntze, Revis. Gen. Pl. 1: 292. 1891, nom. illeg.[non Oldenlandia ovata S. Watson in Proc. Amer. Acad.Arts 18: 97. 1883].Hedyotis paridifolia Dunn in Bull. Misc. Inform. Kew 1912:366. 1912 ≡ Oldenlandia paridifolia (Dunn) Chun in Sunyatsenia1: 313. 1934.Hedyotis parryi Hance in Ann. Sci. Nat., Bot., ser. 4, 18: 221.1862 ≡ Oldenlandia parryi (Hance) Kuntze, Revis. Gen.Pl. 1: 292. 1891.Hedyotis philippensis (Willd. ex Spreng.) Merr. ex C.B. Rob.in Philipp. J. Sci., C 6: 222. 1911 ≡ Spermacoce philippensisWilld. ex Spreng., Syst. Veg. 1: 401. 1824 ≡ Exallagephilippensis (Willd. ex Spreng.) Bremek. in Verh. Kon.Ned. Akad. Wetensch., Afd. Natuurk., Sect. 2, 48(2): 142.1952.Hedyotis prostrata Blume, Catalogus: 40. 1823 ≡ Oldenlandiaprostrata (Blume) Kuntze, Revis. Gen. Pl. 1: 292. 1891.Hedyotis pubescens (Valeton) Merr. & L.M. Perry in J. ArnoldArbor. 26: 3. 1945 ≡ Oldenlandia pubescens Valeton inLorentz, Nova Guinea 8(1): 439. 1911 ≡ Exallage pubescens(Valeton) Bremek. in Verh. Kon. Ned. Akad. Wetensch.,Afd. Natuurk., Sect. 2, 48(2): 142. 1952.Hedyotis pulchella Stapf in Trans. Linn. Soc. London, Bot.4: 169. 1894.Hedyotis pulcherrima Dunn in Bull. Misc. Inform. Kew, Addit.Ser. 10: 127. 1912 ≡ Oldenlandia pulcherrima (Dunn) Chunin Sunyatsenia 1: 314. 1934.Hedyotis purpurascens Hook. f., Fl. Brit. India 3: 50. 1880.Hedyotis quinquenervia Thwaites, Enum. Pl. Zeyl.: 141. 1859≡ Oldenlandia quinquenervia (Thwaites) Kuntze, Revis.Gen. Pl. 1: 292. 1891.Hedyotis rhinophylla Thwaites ex Trimen in J. Bot. 23: 208.1885.Hedyotis rigida (Blume) Walp. in Ann. Bot. Syst. 2: 772. 1852≡ Metabolos rigidus Blume, Bijdr. Fl. Ned. Ind. 16: 992.1826 ≡ Hedyotis rigida (Blume) Miq., Fl. Ned. Ind. 2: 181.1857, nom. illeg. ≡ Scleromitrion rigidum (Blume) Kurz inJ. Asiat. Soc. Bengal, Pt. 2, Nat. Hist. 46(2): 136. 1877 ≡Oldenlandia rigida (Blume) Kuntze, Revis. Gen. Pl. 1: 292.1891, nom. illeg. ≡ Exallage rigida (Blume) Bremek. inVerh. Kon. Ned. Akad. Wetensch., Afd. Natuurk., Sect. 248(2): 142. 1952.Hedyotis rivalis Ridl. in J. Fed. Malay States Mus. 6: 153. 1915.Hedyotis rugosa (Blume) Korth. in Ned. Kruidk. Arch. 2(2):160. 1851 ≡ Metabolos rugosus Blume, Bijdr. Fl. Ned. Ind.16: 991. 1826 ≡ Scleromitrion rugosum (Korth.) Kurz inJ. Asiat. Soc. Bengal, Pt. 2, Nat. Hist. 46(2): 137. 1877 ≡Allaeophania rugosa (Blume) Boerl., Handl. Fl. Ned. Ind.2(1): 124. 1891 ≡ Oldenlandia rugosa (Blume) Kuntze,Revis. Gen. Pl. 1: 292. 1891.Hedyotis schlechteri Merr. & L.M. Perry in J. Arnold Arbor.26: 1. 1945.Hedyotis shenzhenensis Tao Chen in Edinburgh J. Bot. 64:331. 2007.Hedyotis shiuyingiae Tao Chen in Harvard Pap. Bot. 13: 283.2008.Hedyotis stylosa R. Br. ex Wight & Arn., Prodr. Fl. Ind. Orient.1: 407. 1834 ≡ Hedyotis stylosa R. Br. ex G. Don, Gen. Hist.3: 527. 1834, nom. illeg. ≡ Oldenlandia stylosa (R. Br. exG. Don) Kuntze, Revis. Gen. Pl. 1: 293. 1891.Hedyotis swertioides Hook. f., Fl. Brit. India 3: 51. 1880.Hedyotis tenuipes Hemsl. ex F.B. Forbes & Hemsl. in J. Linn.Soc., Bot. 23: 375. 1888 ≡ Oldenlandia tenuipes (Hemsl.ex F.B. Forbes & Hemsl.) Kuntze, Revis. Gen. Pl. 1: 292.1891.Hedyotis ternata (Pierre ex Pit.) P.H. Hô, Cayco Vietnam 3(1):133. 1993 ≡ Oldenlandia ternata Pierre ex Pit. in Lecomte,Fl. Indo-Chine 3: 122. 1922 ≡ Hedyotis ternata (Pierre exPit.) J.F. Maxwell in Nat. Hist. Bull. Siam. Soc. 50(1): 4.2002, nom. illeg.Hedyotis tetrangularis (Korth.) Walp. in Ann. Bot. Syst. 2(5):769. 1852 ≡ Diplophragma tetrangulare Korth. in Ned.Kruidk. Arch. 2(2): 149. 1851 ≡ Oldenlandia tetrangularis(Korth.) Merr. in Pap. Michigan Acad. Sci. 23: 193. 1938.Version of Record (identical to print version).371

Wikström & al. • Phylogeny of HedyotisTAXON 62 (2) • April 2013: 357–374Hedyotis thwaitesii Hook. f., Fl. Brit. India 3: 54. 1880 ≡Hedyotis macrophylla Thwaites, Enum. Pl. Zeyl.: 142.1859, nom. illeg. [non Hedyotis macrophylla Wall. exWight & Arn., Prodr. Fl. Ind. Orient.: 408. 1834] ≡ Oldenlandiathwaitesii (Hook. f.) Kuntze, Revis. Gen. Pl. 1:293. 1891.Hedyotis tridentata Ridsdale in Blumea 41: 456. 1996.Hedyotis trimenii Deb & Ratna Dutta in Taxon 34: 297. 1985.Hedyotis uncinella Hook. & Arn., Bot. Beechey Voy.: 192.1833 ≡ Oldenlandia uncinella (Hook. & Arn.) Kuntze,Revis. Gen. Pl. 1: 293. 1891.Hedyotis vachellii Hook. & Arn., Bot. Beechey Voy.: 194. 1837[non Hedyotis vachellii Benth., Fl. Hongk. 148. 1861, nom.illeg.] ≡ Oldenlandia vachellii (Hook. & Arn.) Kuntze,Revis. Gen. Pl. 1: 293. 1891.Hedyotis valetoniana Merr. & L.M. Perry in J. Arnold Arbor.26: 2. 1945.Hedyotis verticillaris Wall. ex Wight & Arn., Prodr. Fl. Ind.Orient. 1: 409. 1834 ≡ Pleiocraterium verticillare (Wall. exWight & Arn.) Bremek. in Meded. Bot. Mus. Herb. RijksUniv. Utrecht 56: 440. 1939.Hedyotis xinyiensis X. Guo & R.J. Wang in Ann. Bot. Fenn.48: 443. 2011.Hedyotis yangchunensis W.C. Ko & Zhang in J. S. ChinaAgric. Univ. 16(4): 45. 1995.Nomenclatural changes outside Hedyotis s.str.Neanotis nana (Merr. & L.M. Perry) N. Wikström & Neupane,comb. nov. ≡ Hedyotis nana Merr. & L.M. Perry inJ. Arnold Arbor. 26: 6. 1945 – Lectotype (designatedhere): British New Guinea: Murrey Pass, Wharton Range,alt. 2840 m, Aug 1933, Brass 4691 (A no. A00097100 [image!];isolectotype: NY no. NY00131808 [image!]).Neanotis pahompokae (Fukuoka) N. Wikström & Neupane,comb. nov. ≡ Hedyotis pahompokae Fukuoka in Acta Phytotax.Geobot. 24: 29. 1969 – Holotype: Thailand, ChiangRai: Doi Pa Hom Pok, NW of Phan, alt. 1800 m, 12/9–1967,Iwatsuki & al. 9562 (KYO [n.v.]; isotypes: A [image!], K[image!]).Neanotis trichoclada (Merr. & L.M. Perry) N. Wikström& Neupane, comb. nov. ≡ Hedyotis trichoclada Merr.& L.M. Perry in J. Arnold Arbor. 26: 5. 1945 – Lectotype(designated here): Netherlands New Guinea: LakeHabbema, alt. 3225 m, Aug 1938, Brass 9197 (A no.A00097103 [image!]; isolectotype: BM no. BM000945138[image!]).AcknowledgmentsThe authors thank two anonymous reviewers for their constructivecriticism on an earlier version of the paper, and Anbar Khodabandehfor assistance with sequencing. This work was supported by theMinistry of Education of China through its 111 Project (B08044, CUN985-3-3), a Conservation International grant for the Botanical Surveyof Louisiade Archipelago, Milne Bay Province, Papua New Guinea,the J. Robert Stiffler Endowment at ODU, The Swedish ResearchCouncil, and the Knut and Alice Wallenberg Foundation.Literature citedAlfaro, M.E., Zoller, S. & Lutzoni, F. 2003. Bayes or bootstrap? Asimulation study comparing the performance of Bayesian Markovchain Monte Carlo sampling and bootstrapping in assessing phylogeneticconfidence. Molec. Biol. Evol. 20: 255–266.http://dx.doi.org/10.1093/molbev/msg028Andersson, L. & Rova, J.H.E. 1999. The rps16 intron and the phylogenyof the Rubioideae (Rubiaceae). Pl. Syst. Evol. 214: 161–186.http://dx.doi.org/10.1007/BF00985737Babu, C.R. 1969. Thecagonum Babu—a new generic name in Rubiaceae.Bull. Bot. Surv. India 11: 213–214.Baldwin, B.G. & Markos, S. 1998. Phylogenetic utility of externaltranscribed spacer (ETS) of 18S–26S rDNA: Congruence of ETSand ITS trees of Calycadenia (Compositae). Molec. Phylogen. Evol.10: 449–463. http://dx.doi.org/10.1006/mpev.1998.0545Barrie, F.R. 2006. Report of the General Committee: 9. Taxon 55:795–800. http://dx.doi.org/10.2307/25065657Bentham, G. & Hooker, J.D. 1873. Genera plantarum, vol. 2. London:Reeve.Bremekamp, C.E.B. 1939. Pleiocraterium genus novum RubiacearumHedyotidearum. Recueil Trav. Bot. Néerl. 36: 438–445.Bremekamp, C.E.B. 1952. The African species of Oldenlandia L.sensu Hiern et K. Schumann. Verh. Kon. Ned. Akad. Wetensch.,Afd. Natuurk., Sect. 2 48(2): 1–298.Bremekamp, C.E.B. 1966. Remarks on the position, the delimitationand subdivision of the Rubioideae. Acta Bot. Neerl. 15: 1–33.Bremer, B. 1996. Phylogenetic studies within Rubiaceae and relationshipsto other families based on molecular data. Opera Bot. Belg.7: 33–50.Bremer, B. & Eriksson, T. 2009. Timetree of Rubiaceae: Phylogenyand dating the family, subfamilies and tribes. Int. J. Pl. Sci. 170:766–793. http://dx.doi.org/10.1086/599077Bremer, B. & Manen, J.F. 2000. Phylogeny and classification of thesubfamily Rubioideae (Rubiaceae). Pl. Syst. Evol. 225: 43–72.http://dx.doi.org/10.1007/BF00985458Candolle, A.P. de 1830. Prodromus systematis naturalis regni vegetabilis,vol. 4. Paris: Masson.Chamisso, A. & Schlechtendal, D. 1829. De plantis in expeditionespeculatoria Romanzoffiana observatis. Rubiaceae. Linnaea 4:129–202.Chen, T. & Taylor, C.M. 2011. Hedyotis. Pp. 147–174 in: Wu, Z.Y.,Raven, P.H. & Hong, D.Y. (eds.), Flora of China, vol. 19. St. Louis:Missouri Botanical Garden Press.Dessein, S. 2003. Systematic studies in the Spermacoceae (Rubiaceae).Ph.D. thesis, Katholieke Universiteit Leuven, Belgium.Dessein, S., Andersson, L., Geuten, K., Smets, E. & Robbrecht, E.2005. Gomphocalyx and Phylohydrax (Rubiaceae): Sister taxa excludedfrom the Spermacoceae s.s., featuring a remarkable caseof convergent evolution. Taxon 54: 91–107.http://dx.doi.org/10.2307/25065305Devi, H.M. & Raju, P.V.S.N.G.K. 1990. An embryological approach372 Version of Record (identical to print version).

TAXON 62 (2) • April 2013: 357–374Wikström & al. • Phylogeny of Hedyotisto the taxonomical status of Hedyotis Linn. Proc. Indian Acad.Sci. 100: 51–60.Doyle, J.J. & Doyle, J.L. 1987. A rapid DNA isolation procedure forsmall quantities of fresh leaf tissue. Phytochem. Bull. 19: 11–15.Dutta, R. & Deb, D.B. 2004. Taxonomic revision of Hedyotis L. (Rubi -aceae) in Indian subcontinent. Kolkata: Botanical Survey of India.Edgar, R.C. 2004. MUSCLE: Multiple sequence alignment with highaccuracy and high throughput. Nucl. Acids Res. 32: 1792–1797.http://dx.doi.org/10.1093/nar/gkh340Edgar, R.C. 2010. Search and clustering orders of magnitude fasterthan BLAST. Bioinformatics 26: 2460–2461.http://dx.doi.org/10.1093/bioinformatics/btq461Fosberg, F.R. 1943. The Polynesian species of Hedyotis (Rubiaceae).Bull. Bernice P. Bishop Mus. 147: 1–102.Fosberg, F.R. & Sachet, M.H. 1991. Studies in Indo-Pacific Rubiaceae.Allertonia 6: 191–278.Fosberg, F.R., Sachet, M.H. & Oliver, R.L. 1993. Flora of Micronesia,vol. 5, Bignoniaceae–Rubiaceae. Smithsonian Contributionsto Botany 81. Washington, D.C.: Smithsonian Institution Press.Fukuoka, N. 1969. Two new species of Hedyotis from Thailand. ActaPhytotax. Geobot. 24: 28–31.Fukuoka, N. 1970. Contributions to the Flora of Southeast Asia. S. E.Asian Stud. 8: 305–336.Gouy, M., Guindon, S. & Gascuel, O. 2010. Seaview version 4: Amultiplatform graphical user interface for sequence alignment andphylogenetic tree building. Molec. Biol. Evol. 27: 221–224.http://dx.doi.org/10.1093/molbev/msp259Govaerts, R., Andersson, L., Robbrecht, E., Bridson, D., Davis, A.,Schanzer, I. & Sonke, B. 2011. World checklist of Rubiaceae. Kew:The Board of Trustees of the Royal Botanic Gardens. http://apps.kew.org/wcsp/ (accessed 17 June 2011).Groeninckx, I., Dessein, S., Ochoterena, H., Persson, C., Motley,T.J., Kårehed, J., Bremer, B., Huysmans, S. & Smets, E. 2009.Phylogeny of the herbaceous tribe Spermacoceae (Rubiaceae)based on plastid DNA data. Ann. Missouri Bot. Gard. 96: 109–132.http://dx.doi.org/10.3417/2006201Groeninckx, I., Briggs, M., Davus, A., De Block, P., Robbrecht, E.,Smets, E. & Dessein, S. 2010a. A new herbaceous genus endemicto madagascar: Phialiphora (Spermacoceae, Rubiaceae). Taxon59: 1815–1829.Groeninckx, I., De Block, P., Robbrecht, E., Smets, E.E. & Dessein,S. 2010b. Amphistemon and Tham- noldenlandia, two new generaof Rubiaceae (Spermacoceae) endemic to Madagascar. Bot. J. Linn.Soc. 163: 447–472.http://dx.doi.org/10.1111/j.1095-8339.2010.01060.xGroeninckx, I., Ochoterena, H., Smets, E. & Dessein, S. 2010c.Molecular phylogenetic and morphological study of Kohautia(Spermacoceae, Rubiaceae), with the recognition of the new genusCordylostigma. Taxon 59: 1457–1471.Guindon, S. & Gascuel, O. 2003. A simple, fast, and accurate algorithmto estimate large phylogenies by maximum likelihood. Syst. Biol.52: 696–704. http://dx.doi.org/10.1080/10635150390235520Guo, X., Simmons, M.P., But, P.P.H., Shaw, P.C. & Wang, R.J. 2011.Application of DNA barcodes in Hedyotis L. (Spermacoceae, Rubiaceae).J. Syst. Evol. 49: 203–212.http://dx.doi.org/10.1111/j.1759-6831.2011.00130.xHiern, W.P. 1877. Order LXX. Rubiaceae. Pp. 33–247 in: Oliver, D.(ed.), Flora of Tropical Africa, vol. 3. London: Reeve.Hitchcock, A.S. 1929. Genera 1–500. Pp. 114–155 in: Hitchcock, A.S.& Green, M.L. (eds.), Standard species of Linnean genera of Phanerogamae(1753–54). Pp. 111–199 in: International BotanicalCongress Cambridge (England), 1930: Nomenclature; Proposalsby British Botanists. London: His Majesty’s Stationary Office.Hochreutiner, B.P.G. 1934. Plantae Hochreuterianae: Etude systématiqueet biologique des collections faites par l’auteur au cours deson voyage aux Indes néerlandaises et autour du monde pendantles années 1903 à 1905 (fasc. 3). Candollea 5: 175–341.Hooker, J.D. 1873. Ordo LXXXIV. Rubiaceae. Pp. 7–151 in: Bentham,G. & Hooker, J.D. (eds.), Genera plantarum ad exemplaria imprimisin herbariis kewensibus servata definata. London: Reeve.Hooker, J.D. 1880. Order LXXV. Rubiaceae. Pp. 17–210 in: Hooker,J.D. (ed.), Flora of British India, vol. 3. London: Reeve.Hooker, W.J. & Walker-Arnott, G.A. 1841. The Botany of CaptainBeechey’s voyage. London: Bohn.Jarvis, C.E. 1992. Seventy-Two proposals for the conservation of typesof selected Linnaean generic names, the report of subcommittee3C on the lectotypification of Linnaean generic names. Taxon 41:552–583. http://dx.doi.org/10.2307/1222833Kårehed, J. & Bremer, B. 2007. The systematics of Knoxieae (Rubiaceae)—Moleculardata and their taxonomic consequences. Taxon56: 1051–1076. http://dx.doi.org/10.2307/25065904Kårehed, J., Groeninckx, I., Dessein, S., Motley, T.J. & Bremer, B.2008. The phylogenetic utility of chloroplast and nuclear DNAmarkers and the phylogeny of the Rubiaceae tribe Spermacoceae.Molec. Phylogen. Evol. 49: 843–866.http://dx.doi.org/10.1016/j.ympev.2008.09.025Larget, B. & Simon, D.L. 1999. Markov chain Monte Carlo algorithmsfor the Bayesian analysis of phylogenetic trees. Molec. Biol. Evol.16: 750–759. http://dx.doi.org/10.1093/oxfordjournals.molbev.a026160Lewis, W.H. 1966. The Asian genus Neanotis nomen novum (Anotis)and allied taxa in the Americas (Rubiaceae). Ann. Missouri Bot.Gard. 53: 32–46. http://dx.doi.org/10.2307/2394973Linnaeus, C. 1753. Species plantarum. Stockholm: Laurentius Salvius.Lo, H., Ko, W., Chen, W., Hsue, H. & Wu, H. 1999. AngiospermaeDicotyledoneae Rubiaceae (1). Pp. 1–427 in: Lo, H. (ed.) FloraReipublicae Popularis Sinicae, vol. 71(1). Beijing: Science Press.Löhne, C. & Borsch, T. 2005. Molecular evolution and phylogeneticutility of the petD group II intron: A case study in basal angiosperms.Molec. Biol. Evol. 22: 317–332.http://dx.doi.org/10.1093/molbev/msi019Mena, P. 1990. A revision of the genus Arcytophyllum (Rubiaceae:Hedyotideae). Mem. New York Bot. Gard. 60: 1–26.Merrill, E.D. & Metcalf, F.P. 1942. Hedyotis Linnaeus versusOldenlandia Linnaeus and the status of Hedyotis lancea Thunbergin relation to H. consanguinea Hance. J. Arnold. Arbor. 23:226–230.Merrill, E.D. & Perry, L.M. 1945. Plantae Papuanae Archboldianae,XV. J. Arnold. Arbor. 26: 1–36.Neupane, S., Dessein, S. & Motley, T.J. 2009. The Hedyotis–Oldenlandia–Kohautiacomplex (Rubiaceae) in Nepal: A study offruit, seed and pollen characters and their taxonomic significance.Edinburgh J. Bot. 66: 371–390.http://dx.doi.org/10.1017/S0960428609990035Nylander, J.A.A., Ronquist, F., Huelsenbeck, J.P. & Nieves-Aldrey,J.L. 2004. Bayesian phylogenetic analyses of combined data. Syst.Biol. 53: 47–67. http://dx.doi.org/10.1080/10635150490264699Nylander, J.A.A. 2004. MrAIC.pl. Program distributed by the author.Uppsala University Evolutionary Biology Centre.Oxelman, B., Lidén, M. & Berglund, D. 1997. Chloroplast rps16 intronphylogeny of the tribe Sileneae (Caryophyllaceae). Pl. Syst. Evol.206: 393–410. http://dx.doi.org/10.1007/BF00987959Piesschaert, F., Jansen, S., Jaimes, I., Robbrecht, E. & Smets, E.2001. Morphology, anatomy, and taxonomic position of Pagameopsis(Rubiaceae–Rubioideae). Brittonia 53: 490–504.http://dx.doi.org/10.1007/BF02809649Popp, M. & Oxelman, B. 2001. Inferring the history of the polyploidSilene aegaea (Caryophyllaceae) using plastid and homoeologousnuclear DNA sequences. Molec. Phylogen. Evol. 20: 474–481.http://dx.doi.org/10.1006/mpev.2001.0977Puff, C. & Igersheim, A. 1994. The character states and taxonomicposition of Metabolos Bl. (syn. Allaeophania Thw.) (Rubiaceae).Bull. Jard. Bot. Natl. Belg. 63: 241–262.http://dx.doi.org/10.2307/3668251Version of Record (identical to print version).373

Wikström & al. • Phylogeny of HedyotisTAXON 62 (2) • April 2013: 357–374Ridsdale, C.E. 1998. Rubiaceae, Tribe Hedyotideae. Pp. 236–284 in:Dassanayake, M.D. (ed.), A revised handbook to the flora of Ceylon,vol. 12. Rotterdam: Balkema.Robbrecht, E. 1988. Tropical woody Rubiaceae. Opera Bot. Belg. 1:1–271.Robbrecht, E. 1993. Supplement to the 1988 outline of the classificationof the Rubiaceae. index to genera. Pp. 173–196 in: Robbrecht, E.(ed.), Advances in Rubiaceae macrosystematics. Opera Bot. Belg.6: 173–195.Ronquist, F. & Huelsenbeck, J.P. 2003. MrBayes 3: Bayesian phylogeneticinference under mixed models. Bioinformatics 19: 1572–1574.http://dx.doi.org/10.1093/bioinformatics/btg180Schumann, K. 1891. Rubiaceae. Pp. 1–156 in: Engler, A. & Prantl, K.(eds.), Die natürlichen Pflantzenfamilien, vol. 4. Leipzig.Simmons, M.P. & Ochoterena, H. 2000. Gaps as characters in sequence-basedphylogenetic analyses. Syst. Biol. 49: 369–381.http://dx.doi.org/10.1093/sysbio/49.2.369Sivarajan, V.V. & Biju, S.D. 1990. Taxonomic and nomenclatural noteson the Hedyotis corymbosa–diffusa complex (Rubiaceae) in India.Taxon 39: 665–674. http://dx.doi.org/10.2307/1223392Staden, R. 1996. The Staden sequence analysis package. Molec. Biotechnol.5: 233–241. http://dx.doi.org/10.1007/BF02900361Staden, R., Beal, K.F. & Bonfield, J.K. 2000. The Staden package,1998. Pp. 115–130 in: Misener, S. & Krawets, S.A. (eds.), Bioinformaticsmethods and protocols. Methods in Molecular Biology132. New Jersey, U.S.A.: Humana Press.Terrell, E.E. 1975. Relationships of Hedyotis fruticosa L. to HoustoniaL. and Oldenlandia L. Phytologia 31: 418–421.Terrell, E.E. 1991. Overview and annotated list of North Americanspecies of Hedyotis, Houstonia, Oldenlandia (Rubiaceae), andrelated genera. Phytologia 71: 212–243.Terrell, E.E. 1996. Revision of Houstonia (Rubiaceae–Hedyotideae).Syst. Bot. Monogr. 48: 1–118. http://dx.doi.org/10.2307/25027862Terrell, E.E. 2001a. Stenotis (Rubiaceae), a new segregate genus fromBaja California, Mexico. Sida 19: 899–911.Terrell, E.E. 2001b. Taxonomic review of Houstonia acerosa andH. palmeri, with notes on Hedyotis and Oldenlandia (Rubiaceae).Sida 19: 913–922.Terrell, E.E. 2001c. Taxonomy of Stenaria (Rubiaceae; Hedyotideae),a new genus including Hedyotis nigricans. Sida 19: 591–614.Terrell, E.E. & Robinson, H. 2003. Survey of Asian and Pacific speciesof Hedyotis and Exallage (Rubiaceae) with nomenclatural noteson Hedyotis types. Taxon 52: 775–782.http://dx.doi.org/10.2307/3647351Terrell, E.E. & Robinson, H. 2007. Seed and capsule morphology insix genera of Hedyotideae (Rubiaceae): Thecagonum, Neanotis,Dentella, Kohautia, Pentodon, and Oldenlandiopsis. J. Bot. Res.Inst. Texas 1: 373–384.Terrell, E.E., Robinson, H.E., Wagner, W.L. & Lorence, D.H. 2005.Resurrection of genus Kadua for Hawaiian Hedyotidinae (Rubiaceae),with emphasis on seed and fruit characters and notes onSouth Pacific species. Syst. Bot. 30: 818–833.http://dx.doi.org/10.1600/036364405775097716Verdcourt, B. 1958. Remarks on the classification of the Rubiaceae.Bull. Jard. Bot. État Bruxelles 28: 209–281.Wight, R. & Arnott, G.A.W. 1834. Prodromus Florae PeninsulaeIndiae Orientalis, vol. 1. London: Parbury, Allen & Co.374 Version of Record (identical to print version).

Phylogeny of Hedyotis L. (Rubiaceae: Spermacoceae): Redefining a ...

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?