classification of apocynaceae s. l. according to a new .../menu/standar… ·...
TRANSCRIPT
Syst Biol 51(3)389ndash409 2002DOI 10108010635150290069869
Classication of Apocynaceae sl According to a New ApproachCombining Linnaean and Phylogenetic Taxonomy
BENGT SENNBLAD1 AND BIRGITTA BREMER
Department of Systematic Botany Evolutionary Biology Center (EBC) Uppsala UniversityNorbyv 18D SE-752 36 Uppsala Sweden
Abstractmdash A new approach to a nomenclatural system including elements from both Linnaean andphylogenetic nomenclature is proposed It is compatible with the existing Linnaean system includ-ing ldquostandard namesrdquo corresponding to principal and secondary ranks and uses a variant of thedenitions from the Phylocode system A new infrafamilial classication using this nomenclaturalapproach of the Apocynaceae sl (ie including the Asclepiadaceae) based mainly on analyses of rbcLand ndhF data is discussed Twenty-one tribes and four rankless taxa are dened [Apocynaceae slclassication Linnaean nomenclature ndhF nomenclature phylogenetic nomenclature phylogenyrbcL]
Discussion about different nomenclaturesystems in recent literature has been in-tense (eg de Queiroz and Gauthier 19921994 Schander and Thollesson 1995 Bryant1996 1997 de Queiroz 1996 1997a Lidenand Oxelman 1996 Reveal 1996 Cantinoet al 1997 Liden et al 1997 Moore 1998Schander 1998 Sereno 1999)A new nomen-clature system termed the Phylocode orphylogenetic taxonomy (de Queiroz andGauthier 1992 Cantino and de Queiroz2000) has been proposed challenging thetraditionally used ldquoLinnaean systemrdquo (egGreuter et al 1994)Applications of phyloge-netic taxonomy to plant systematics includefor example classications of the Lamiaceae(Cantino et al 1997) Ericaceae (Kron 1997)Malvaceae (Baum et al 1998) and Scrophu-lariaceae (Olmstead et al 2001)
In the Linnaean system nomenclatureand classication (circumscription) are notstrongly coupled Named taxa are xedto two reference points the type (typegenusspecies or in the case of species typespecimen) and the rank the only mandatoryrank is genus (Greuter et al 1994 in prac-tice further ranks areoften treated as manda-tory) An exact circumscription of the taxonis not required the only restriction is that ataxon cannot include another taxon with thesameor higher rank Nevertheless a listof in-cluded taxa is often added in classicationsof taxa above species level
1Current afliation and address for correspondenceStockholm Bioinformatics CenterCenter for Genomicsand Evolution Karolinska Institute SCFAB SE-106 91Stockholm Sweden E-mail BengtSennbladsbcsuse
The Phylocode system aims at explicitlybinding names to specic clades the nameof a taxon is tightly coupled to a specicclade (or a specic ancestor) through a def-inition Three different kinds of denitions(node-based stem-based and apomorphy-based eg de Queiroz and Gauthier 1992Cantino and de Queiroz 2000) have beenproposed Because the Phylocode systemabandons ranks it is (with its present de-scription) incompatible with the Linnaeansystem
Both systems have certain advantagesas a communication tool in biology TheLinnaean system has a more stable set ofnames in use (but the exact circumscrip-tions of their corresponding taxa may vary)whereas the phylogenetic system providesmore exact denitions thus reducing in-stability caused by subjective changes intaxon circumscriptions (eg splitters andlumpers) In addition the principal ranks ofthe Linnaean system provide universal stan-dard names that are important for examplefor textbooks databases and oras
We propose a compromise approach thatcombines the advantages of the two systemsIt is compatible with the system presently inusemdashthe Linnaean systemmdashin using a sys-tem of hierarchic standard names (comparewith ldquoprimary ranksrdquo in the Linnaean sys-tem) and types but it also adopts a variant ofthe denitions from the phylogenetic systemto reduce the impact of subjective changesin circumscriptions Because the main aimwith the present study is to present a newtribal classication of the angiosperm fam-ily Apocynaceae sl we will mainly concern
389
390 SYSTEMATIC BIOLOGY VOL 51
ourselves with familytribal-level plant clas-sication Likewise our use of the term theLinnaean system mainly refers to the In-ternational Code of Botanical Nomenclature(ICBN Greuter et al 1994) We will notdiscuss species-level nomenclature nor thatof extinct taxa (see eg de Queiroz andGauthier 1992 Eriksson et al 1998 Mishler1999)
The Apocynaceae sl belong to the or-der Gentianales and have a mainly pantrop-icalsubtropical distribution with a fewtemperate representatives The family com-prises many well-known ornamentals suchas Nerium (oleander) and Hoya (wax ower)The plants are typically laticiferous and pro-duce various alkaloids and cardenolidessome of which have medical properties Themost well-known example is Catharanthus(rose periwinkle) which contains vinblas-tine and vincristine compounds now usedworldwide to treat childhood leukemia
The family Asclepiadaceaewas segregatedfrom the Apocynaceae sl by Brown (1810)Although some authors have questioned thisseparation (eg Demeter 1922 Safwat 1962Judd et al 1994 Struwe et al 1994) mostclassications subsequent to that of Brownhave followed his ldquotwo-familyrdquo treatmentMolecular studies of the Apocynaceae sl(Sennblad and Bremer 1996 Civeyrel et al1998 Potgieter and Albert 2001) have indi-cated that the Asclepiadaceae form a sub-clade of the Apocynaceae sstr thus render-ing the Apocynaceae sstr nonmonophyleticThese studies also indicated major problemswith the infrafamilial classicationin the twofamilies A detailed discussion of differentclassications of the Apocynaceae sstr andthe Asclepiadaceae is given in Sennblad andBremer (1996) and only a short review willbe given here concentrating on the most re-cent classications and results from molecu-lar studies
Since the classication of Schumann(1895) the Apocynaceae have tradition-ally been divided into two subfamilies thePlumerioideae and the Apocynoideae andapproximately into 12 tribes and 27 sub-tribes (Pichon 1948abc1950Leeuwenberg1994a) The studies by Sennblad and Bremer(1996) and Endress et al (1996) showed thatboth the Plumerioideae and Apocynoideaeare nonmonophyletic Four of the tribesof the Plumerioideae have been shown tobe nonmonophyletic (Endress et al 1996
Sennblad and Bremer 1996 Civeyrel et al1998 Potgieter and Albert 2001) threetribes were monogeneric and the mono-phyly of one tribe has not been evalu-ated The study of Sennblad et al (1998)was aimed mainly at the tribe Wrightieae(Apocynoideae) sensu Leeuwenberg (1994a)but also included representatives of the othertribes of the Apocynoideae as well as of theAsclepiadaceae Sennblad et al showed thatnone of the described tribes within the Apoc-ynoideae is monophyletic and suggesteda reclassication of the Wrightieae sensuLeeuwenberg (1994a)into the Wrightieae theNerieae and the Malouetieae
The infrafamilial classication of Ascle-piadaceae proposed by Brown (1810) whendescribing the family has been followedby most subsequent authors (Liede andAlbers 1994) and comprises three subfam-ilies of which the Periplocoideae and Se-camonoideae are monotribal and the As-clepiadoideae usually comprise ve tribesAn exception from this traditional view isthe classication of Swarupanandan et al(1996) which reduced the number of tribesof the Asclepiadoideae to two and includedthe Secamonoideae as a third tribe in theAsclepiadoideae Furthermore a revision ofthe Periplocoideae by Venter and Verhoeven(1997) recognizes three tribes in this subfam-ily The monophyly of the Asclepiadaceaeis uncertain in the study of Sennblad andBremer (1996) however the subfamilyPeriplocoideae itself monophyletic did notform a monophyletic group with the othertwo subfamilies (sensu Liede and Albers1994) whereas in the study of Civeyrelet al (1998) the three subfamilies did form amonophyletic group The study of Civeyrelet al (1998) also indicated that the tradi-tional Secamonoideae is monophyletic witha position as sister to Asclepiadoideae (seealso Sennblad and Bremer 1996 Potgieterand Albert 2001) The traditional Asclepi-adoideae receive good support (Sennbladand Bremer 1996 Civeyrel et al 1998Potgieter and Albert 2001) Of the tribesof the Asclepiadoideae in the classicationof Liede and Albers (1994) Asclepiadeaeand Marsdenieae have been shown to benonmonophyletic
The ndings of the molecular studies indi-cated above and also from a previous versionof this paper (included in the PhD thesisof BS 1997 see also Sennblad and Bremer
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 391
2000) were summarized and developed fur-ther by use of morphology in the most re-cent classication of the Apocynaceae sl byEndress and Bruyns (2000) The ve tradi-tional subfamilies were kept but 19 newlyrecircumscribed tribes were also recognized(see Table 1)
However the results of the present studyindicate that although the classication ofEndress and Bruyns (2000) are a major ad-vance in making the classication congru-ent with current best estimates of phylogenyproblematic cases still need revision Fur-thermore for many of the recognized tribestaxon sampling for molecular data is weakand the morphological homologies in thesegroups are difcult to interpret (Endress andBruyns 2000) Further rearrangements oftribal circumscription will almost certainlybe needed in the future
The Apocynaceae sl may thus provide asuitable test for the classication system pro-posed in this study We therefore propose anew classication of the Apocynaceae basedmainly on an rbcL analysis that includesextended taxon sampling as well as addi-tional data from length variation of rbcL in its30-end and downstream Certain relation-ships of special interest have been tested witha smaller data set for the subfamily Apoc-ynoideae and the traditionalAsclepiadaceaeincluding additionalnucleotide sequences ofthe chloroplast gene ndhF This gene oftenhas a greater substitution rate than rbcL (seeOlmstead and Sweere 1994 Kim and Jansen1995) and may therefore form a complementto rbcL when the latter proves too conservedWe have used only the 3rsquo region of the genebecause that region appears to show the mostvariability (see Kim and Jansen 1995) Fur-thermore results from other recent molecu-lar studies have also been taken into accountin the classication
MATERIALS AND METHODS
We sampled 77 representatives of theApocynaceae sl (Table 1) All subfamiliesand tribes of the classication of Endress andBruyns (2000) are represented in the analy-sis Gelsemium Juss and Mostuea Didr of theGelsemiaceae were chosen as outgroup taxabecause they were indicated to be the closest
TABLE 1 Classication of the Apocynaceae sl andsampling of taxa Classication are according to Endressand Bruyns (2000)
RAUVOLFIOIDEAEAlstonieae Melodineae
Alstonia RBr Craspidospermum ADCAspidosperma Mart Diplorhynchus Ficalho
et Zucc et HiernVallesia Ruiz Melodinus JR Forst
et Pav et G ForstVinceae Hunterieae
Catharanthus G Don Picralima PierreKopsia Blume Pleiocarpa BenthOchrosia Juss PlumerieaeRauvola L Allamanda LVinca L Anechites Griseb
Willughbeieae Cameraria LAncylobotrys Pierre Cerbera LDictyophleba Pierre Plumeria LVahadenia Stapf Thevetia Adans
Tabernaemontaneae CarisseaeCarvalhoa K Schum Acokanthera G DonMolongum Pichon Carissa LSchizozygia Baill AlyxieaeTabernaemontana L Alyxia RBrTabernanthe Baill Chilocarpus Blume
Lepinia DecneAPOCYNOIDEAE
Wrightieae ApocyneaeAdenium Roem Aganosma (Blume)
et Schult G DonNerium L Apocynum LStephanostema K Schum Baissea ADCStrophanthus DC Beaumontia WallWrightia RBr Trachelospermum Lem
Malouetieae MesechiteaeFuntumia Stapf Mandevilla LindlHolarrhena RBr Mesechites Mull-ArgKibatalia GDon EchiteaeMascarenhasia ADC Parsonsia RBrPachypodium Lindl Peltastes Woodson
Prestonia RBrRhabdadenia Mull-Arg
PERIPLOCOIDEAEMondia Skeels Periploca LParquetina Baill Petopentia BullockPentopetia Decne Tacazzea Decne
SECAMONOIDEAESecamone RBr
ASCLEPIADOIDEAEMarsdenieae Asclepiadeae
Fockea Endl Araujia BrotHoya RBr Asclepias LMicholitzia NEBr Calotropis RBrStephanotis Thouars Orthosia Decne
Ceropegieae Fischeria DCCeropegia L Matelea AublStapelia L Schizostephanus Benth
et Hook fTweedia Hook et ArnTylophora RBrVincetoxicum Wolf
392 SYSTEMATIC BIOLOGY VOL 51
sister group to the Apocynaceae sl in therecent analysis of Gentianales by Backlundet al (2000)
Forty-one new sequences of rbcL are pub-lished in this study rbcL sequences forthe outgroup taxa and for 35 ingroup taxawere published previously (Table 2) ForndhF the 30 region of the gene was se-quenced for 18 taxa representing the tra-ditional Apocynoideae and AsclepiadaceaeTotal DNA was extracted from fresh leavesor herbarium material by using the meth-ods of Saghai-Maroof et al (1984) and Doyleand Doyle (1987) The extractions were pu-ried by ultracentrifugation in CsCl gra-dients or ethanol precipitation Additionalpurication with the Qiaquick PCR puri-cation kit (Qiagen Inc) was performed incases with problematic polymerase chain re-action (PCR) amplication Double-strandedDNA was amplied by PCR by using Taq-polymerase kit (Promega Corp) SyntheticPCR primer sequences for rbcL were takenfrom Olmstead et al (1992) ndhF primerswere taken from Olmstead and Sweere (1994)and Oxelman et al (1999) In cases wherePCR amplication proved difcult a PCRreaction using Taq extender PCR additive(Stratagene Inc) was performed followingthe protocol provided by the supplier A sec-ond round of PCR with only one of theprimers was performed to obtain single-stranded DNA (Kaltenboeck et al 1992)Single-stranded DNA was sequenced man-ually (Sanger et al 1977) by using internalprimers designed for rbcL by G Zurawski(DNAX Research Institute) and for ndhF byOlmstead and Sweere (1994) and Oxelmanet al (1999)
The length of the sequences of rbcL ex-cluding the 26 rst nucleotides but includ-ing positions just downstream from the geneobtained in this study varied between 1411and 1509 nt For all taxa except Parsonsia thesequence includes the stop codon In posi-tions 27ndash1425 no structural mutations occurin rbcL Alignment of this part of the geneis thus very simple However after position1425 that is just before the typical positionof the stop codon gaps of different sizes oc-cur complicating alignment Similarly gapsoccur in the coding region of the ndhF se-quences Because of these alignment prob-lems the sequences were truncated 23 basesbefore the typical position of the stop codonin Apocynaceae sl The reading frame was
inferred through comparisonswith completeApocynaceae ndhF-sequences from the studyof Backlund et al (2000)
Multiple sequence alignment was made byhand to reduce the number of gaps while in-creasing the percentage similarity We usedthe criteria of Golenberg et al (1993) with mi-nor modications
Two cladistic analyses were performedThe data matrix for the rst analysis coveredall included taxa and comprised characterscorresponding to the nucleotide positions27ndash1425 of the rbcL gene and characterscorresponding to gaps and informative nu-cleotide positions downstream of position1425 EMBL database accession numbersfor rbcL sequences are given in Table 2Only parsimony informative characters wereanalyzed
The second analysis was of a subset ofthe taxa included in the rst analysis andincludes 18 genera representing the Apoc-ynoideae sensu Leeuwenberg (1994a) andthe traditional Asclepiadaceae The data ma-trix consisted of two submatrices One sub-matrix comprised positions 27ndash1425 of therbcL gene The second submatrix comprised733 aligned nucleotide positions from the3rsquo region of ndhF EMBL database accessionnumbers for the ndhF sequences are given inTable 2 Only parsimony informative charac-ters were analyzed
The cladistic analyses were performed us-ing PAUP 40b2 (Swofford 1998) An initialheuristic search (PAUP settings HSEARCH[ADDSEQ D RANDOM NREPS D 200SWAP D TBR] other options withdefault set-tings) with all characters given a unit weightwas followed by a successiveweighting anal-ysis (Farris 1969 1989) also using heuris-tic searches (PAUP settings HSEARCH[ADDSEQ D RANDOM NREPS D 10SWAP D TBR] other options withdefault set-tings) and characters reweighted accordingto their rescaled consistency index
Bootstrap frequencies (Felsenstein 1985)were calculated with 10000 replicates on theunit-weighted (uwboot) and successiveweighted (swboot) data matrices (PAUPsettings BOOTSTRAP [NREPS D 10000METHOD D HEURISTIC CONLEVEL D50] [ADDSEQ D RANDOM NREPS D 1SWAP D SPR NOMULPARS] other optionswith default settings)
To simplify the discussion of the resultswe used a rough scale of the relative support
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 393
TABLE 2 Vouchers andEMBLaccession numbers for sequences published in this study Forpreviously publishedsequences and for sequences from previously published voucher specimens the reference is given
Taxon Vouchersourceb EMBL number rbcLndhF
Acokanthera oblongifolia (Hochst) Codd Sennblad and Bremer 1996 X91758-Adenium obesum (Forssk) Roem et Schult Sennblad et al 1998 AJ002880AJ420130Aganosma marginata (Roxb) G Don Ryding 490 UPS AJ419730-Allamanda cathartica L Sennblad and Bremer 1996 X91759-Alstonia scholaris RBr Sennblad and Bremer 1996 X91760-Alyxia ruscifolia RBr Sennblad 236 UPS AJ419731-Ancylobotrys petersiana Pierre Sennblad 201 UPS AJ419732-Anechites nerium Urb Bremer et al 3386 UPS AJ419733-Apocynum cannabinum L Sennblad and Bremer 1996 X91761AJ420113Araujia hortorum E Fourn Bremer 3006 UPS AJ419734-Asclepias curassavica L Sennblad and Bremer 1996 X91774-Aspidosperma triternatum Rojas Acosta Bremer 3029 UPS AJ419735-Baissea leonensis Benth Sennblad et al 1998 Yallah 112 UPSa AJ002881AJ420114Beaumontia grandiora Wall Sennblad et al 1998 AJ002882AJ420115Calotropis procera (Aiton) WT Aiton Sennblad 222 UPS AJ419736-Cameraria latifolia L Houghton 1085 FTG AJ419737-Carissa bispinosa (L) Merxm Sennblad 235 UPS AJ419738-Carvalhoa campanulata K Schum Sennblad 217 UPS AJ419739-Catharanthus roseus G Don Sennblad and Bremer 1996 X91757-Cerbera venenifera Steud No voucher AJ419740-Ceropegia woodii Schltr Sennblad and Bremer 1996 X91775-Chilocarpus suaveolens Blume Endress et al 1996 X92445-Fockea multiora K Schum Specks 248 cult MSTRc AJ419741-Craspidospermum verticillatum Bojer Pettersson and Nilsson 742 AJ419742-Dictyophleba lucida Pierre Sennblad and Bremer 1996 X91762-Diplorhynchus condylocarpon (Mull-Arg) Pichon Sennblad 203 UPS AJ419743-Fischeria stellata (Vell) E Fourn Clark and Watt 793 UPS AJ419744-Funtumia elastica Stapf Sennblad et al 1998 AJ002884-Gelsemium sempervirens Aiton Olmstead et al 1993 L14397-Holarrhena pubescens G Don Sennblad et al 1998 AJ002884-Hoya bella Hook Sennblad and Bremer 1996 X91776-Kibatalia gitingense (Elmer) Woodson Liede 3268 Z AJ419745-Kopsia fruticosa ADC Sennblad and Bremer 1996 X91763-Lepinia taitensis Decne Perlman 15071 NTBG AJ419746-Mandevilla sanderi (Hemsl) Woodson Sennblad and Bremer 1996 X91764AJ420116Mascarenhasia arborescens ADC Sennblad et al 1998 AJ002885AJ420117Matelea hirsuta (Vahl) Woodson Sennblad 263 UPS AJ419747-Melodinus monogynus Roxb Sennblad 261 UPS AJ419748-Mesechites trida Mull-Arg Bremer et al 3351 UPS AJ419749-Micholitzia obcordata NEBr Bremer 3010 UPS AJ419750-Molongum laxum (Benth) Pichon Sennblad and Bremer 1996 X91765-Mondia ecornuta (NEBr) Bullock Sennblad 215 UPS AJ419751-Mostuea brunonis Didr Olmstead et al 1993 L14404-Nerium oleander L Sennblad 265 UPS AJ002886AJ420118Ochrosia coccinea Miq v d Laan 389 WAG AJ419752-Orthosia serpyllifolia Decne Bremer et al 3372 UPS AJ419753-Pachypodium lamerei Drake Sennblad et al 1998 AJ002887AJ420119Parquetina nigrescens (Afzel) Bullock Sennblad and Bremer 1996 X91777-Parsonsia heterophylla A Cunn Sennblad et al 1998 AJ002888-Peltastes peltatum (Vell) Woodson Sennblad 262 UPS AJ419754-Pentopetia sp No voucher AJ419755-Periploca graeca L Sennblad et al 1998 AJ002889AJ420120Petopentia natalensis (Schltr) Bullock Sennblad 237a and sn cult MSTR AJ419756AJ420121Picralima nitida T et H Dur Sennblad and Bremer 1996 X91766-Pleiocarpa mutica Benth Bremer 3017 UPS AJ419757-Plumeria inodora Jacq Sennblad and Bremer 1996 X91767-Prestonia quinquangularis Spreng Sennblad and Bremer 1996 X91768AJ420122Rauvola mannii Stapf Sennblad and Bremer 1996 X91769-Rhabdadenia biora Mull-Arg Zona 616 FTG AJ419759AJ420123Schizostephanus alatus K Schum Cult MSTR AJ419758-Schizozygia coffaeoides Baill Sennblad 207 UPS AJ419760-Secamone afzelii (Schult) K Schum Sennblad and Bremer 1996 X91779AJ420124
394 SYSTEMATIC BIOLOGY VOL 51
TABLE 2 Continued
Taxon Vouchersource EMBL number rbcLndhF
Secamone geayi Constatin et Gallaud Civeyrel 1200 LC AJ419761-Stapelia leendertziae NEBr Sennblad and Bremer 1996 X91778-Stephanotis oribunda Brongn Sennblad 256 UPS AJ419762AJ420125Stephanostema stenocarpum K Schum Sennblad and Bremer 1996 X91770AJ420126Strophanthus eminii Pax Sennblad and Bremer 1996 X91771AJ420127Tabernaemontana divaricata Roem et Schult Sennblad and Bremer 1996 X91772-Tabernanthe iboga Baill Leeuwenberg 12544 WAG AJ419763-Tacazzea apiculata Oliver Venter 9188 cult MSTR AJ419764-Thevetia peruviana (Pers) K Schum Sennblad and Bremer 1996 X91773-Trachelospermum jasminoides (Lindl) Lem Sennblad et al 1998 AJ002890AJ420128Tweedia coerulea Sweet Sennblad 254 UPS AJ419765-Tylophora sylvatica Decne Sennblad and Bremer 1996 X91789-Vahadenia caillei (AChev) Hutch et Dalziel Leeuwenberg 12275 WAG AJ419766-Vallesia antillana Woodson Meagher 966 FTG AJ419767-Vinca minor L Sennblad 230 UPS AJ419768-Vincetoxicum hirundinaria Medik Sennblad 257 UPS AJ419769-Wrightia arborea (Dennst) Mabb Sennblad et al 1998 AJ002891AJ420129
aVoucher for rbcL sequence onlybHerbarium abbreviations are according to Index Herbariorum except LC D Private herbarium of Laure CiveyrelcThis specimen was originally determined as Cibirhiza albersiana in Kunze et al (1994) but has recently been redetermined as
Fockea multiora K Schum (Verhoeven et al 2002)
of the clades in the successive weightinganalysis Clades with a successive weightedbootstrap gt63 were considered well sup-ported those with a successive weightedbootstrap gt85 were considered stronglysupported In a theoretical context a cladesupported by one uncontradicted charactercorresponds to a bootstrap support of 63anda cladesupported by two uncontradictedcharacters to 85 (Harshman 1994 Farriset al 1996) Here however this scale merelyrelates to the support distribution within theanalysis
RESULTS
The unit weight analysis of the completerbcL data set (237 characters total 214 ofthe cells were scored as missing data) re-sulted in 19003 trees 828 steps long witha consistency index (ci) D 0378 and a re-tention index (ri) D 0678 The consensustree from the successive weighting analysisof 252 most-parsimonious trees (each 1773steps long ci D 0661 ri D 0880) is presentedin Figure 1 The length of these trees withunit weight characters is 830 steps thus thesetrees are not identical to any of the most-parsimonious trees from the unit weightanalysis Branches not present in the strictconsensus tree from the unit weight analysisare indicated on the combinable consensustree from the successive weighting analysis
(Fig 1) The unit weight analysis of the com-bined set of the rbcL and ndhF data (153 char-acters total 043 of the cells were scored asmissing data) resulted in four trees 345 stepslong (ci D 0539 ri D 0482) The successiveweighting analysis gave three trees 978 stepslong (ci D 0840 ri D 0849) The strict consen-sus tree of the successive weighting analysisis presented in Figure 2 These trees are onestep longer with unit weights than the treesfrom the unit weight analysis Unless other-wise stated the following discussion refersto the successive weighting analyses of thecomplete rbcL data set
Investigating the justication of a prioriweighting (Sennblad and Bremer 2000) wemade some additional analyses of our rbcLdata set including tests for base compositionbias and rate heterogeneity This did not pro-duce signicantly different results and allwell-supported groups were congruent withthose of the present study The general struc-ture and the well-supported groups of theresulting trees (Figs 1 and 2) are to a largedegree congruent with the molecular stud-ies of Sennblad and Bremer (1996) whichare based on rbcL data of Civeyrel et al(1998) based mainly on matK sequence dataand of Potgieter and Albert (2001) basedmainly on trnLndashF data Although there aredifferences in the weakly supported rela-tionships among the well-supported groupsin particular the exact position of the
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 395
FIGURE 1 Combinable component consensus tree of the 252 most-parsimonious trees from the successiveweighting analysis of the complete rbcL data set Subfamilial and tribal classication is according to Endress andBruyns (2000) For subfamilies a three-letter code is used RAU APO PER SEC and ASC denote RauvoloideaeApocynoideae Periplocoideae Secamonoideae and Asclepiadoideae respectively OUT denotes the outgroupNumbers below branches are unit-weighted bootstrap values successive weighted bootstrap values are indicatedabove branches Branches not present in the strict consensus tree from the unit weight analysis are indicated witha cross (dagger) For branches marked A or B the frequencies in the combinable component consensus are 95 or 50respectively all other branches have 100 frequency
396 SYSTEMATIC BIOLOGY VOL 51
FIGURE 2 Strict consensus tree from the successive weighting analysis of the combined ndhF and rbcL dataset Tribal classication sensu Endress and Bruyns (2000) is indicated Numbers below branches are unit-weightedbootstrap values whereas successive weighted bootstrap values are indicated above branches Branches not presentin the strict consensus tree from the unit weight analysis are indicated with a cross (dagger)
Periplocoideae and of the root of the in-group all well-supported groups in our anal-ysis are congruent with those studies Thecombined analysis using data from rbcL-and matK-sequences and oral and pollenmorphology in the study of the Apocy-naceae sstr by Endress et al (1996) andthe analysis of Sennblad et al (1998) ofthe tribe Wrightieae sensu Leeuwenberg(1994a) and related taxa using rbcL andmorphological data are largely congruentwith the result from the present study Fur-thermore preliminary results from collab-orative studies by Sennblad et al (aimedat the traditional Apocynoideae and As-clepiadaceae pers comm) and Endresset al (aimed at the traditional Alyxieaepers comm) based on combined analy-sis of several molecular and morphologi-cal datasets are also congruent with ourresults
Support is good for the split betweenthe ingroup and the two outgroup taxaGelsemium and Mostuea (uwboot D 96swboot D 100) Thus the initial assump-tion of monophyly of the Apocynaceaesl is not violated Conforming to ear-lier studies (eg Sennblad and Bremer1996 Civeyrel et al 1998 Sennblad et al1998 Potgieter and Albert 2001) the tra-ditional Asclepiadaceae are nested withinthe traditional subfamily Apocynoideae(eg sensu Endress and Bruyns 2000)
The ApocynoideaendashAsclepiadaceae cladewhich is strongly supported (uwboot D 86swboot D 99) is in turn nested within thesubfamily Rauvoloideae sensu Endress andBruyns (2000)
Comparison of our results to the mostrecent classications of the ApocynaceaeEndress and Bruyns (2000 see Table 1 andFig 1) shows that although their circum-scription of the subfamilies PeriplocoideaeSecamonoideae and Asclepiadoideae aremonophyletic in our analysis their sub-families Rauvoloideae and Apocynoideaeare nonmonophyletic Similarly the tribesof the Rauvoloideae except Melodineaeare monophyletic but three of the tribesof the Apocynoideaemdashthe Apocyneae theEchiteae and the Wrightieaemdashare nonmono-phyletic Of the tribes of the Asclepiadoideaethe Ceropegieae and Asclepiadeae are mono-phyletic The position of Fockea as sisterto the rest of the Asclepiadoideae ren-ders the Marsdenieae nonmonophyletic Allthe included Periplocoideae taxa belongto the tribe Periploceae sensu Venter andVerhoeven (1997) except Pentopetia whichthey placed in the tribe CryptolepideaeGiven this unresolved position of Pentopetiathe monophyly of these tribes cannot be eval-uated however the reduction of Parquetinato synonymy with Periploca proposed byVenter and Verhoeven (1997) is not sup-ported by the present study
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 397
DISCUSSION
One of the most important tasks for anomenclature system is to provide a sta-ble means to communicate classications(which we here will understand as represent-ing phylogenetic relationships) Althoughboth the Linnaean and the Phylocode sys-tems are able to communicate phylogeniesthey are sensitive to changes in phylogenetichypotheses which in both systems will de-mand taxonomic changes In addition theLinnaean system is subject to a potentialinstability relating to subjective changes incircumscription that is lumpers and split-ters may refer to different groups of organ-isms by the same name (see de Queiroz andGauthier 1992) For many taxa a consensuscircumscription has eventually been estab-lished but for other groups this is still a prob-lem A current example of this from Apoc-ynaceae is the genus Tabernaemontana sensuLeeuwenberg (1991 1994b) which in theclassication of Allorge (1985) is split intoeight genera The Phylocode system was de-signed to avoid this problem by use of strictdenitions of taxa subjective changes incircumscription are avoided
However this rigidity in denitions leadsto a potential instability and high turnoverof names associated with a group of nestedtaxa Relatively small changes in phyloge-netic views eg when a dening taxonldquospecierrdquo receives a less nested positionmaycause some(or all) of these nested namesto become synonymous (eg de Queirozand Gauthier 1994 Bryant 1996 Liden andOxelman 1996Sereno 1999)The junior syn-onyms should according to the Phylocodebe rejected and new names may then beneeded for the new system of nested clades(de Queiroz and Gauthier 1994) The namesassociated with this group of taxa are thusreplaced by new names This can be espe-cially problematic when careless denitionshave been made as have been noted bySchander and Thollesson (1995) and Cantinoet al (1997) among others Consider a hy-pothetical worst-case example The Ascle-piadaceae were traditionally considered asseparate from Apocynaceae A stem-baseddenition of Apocynaceae reecting thisview could be the largest clade that in-cludes Apocynum but not Asclepias On thepresent tree such a denition would limitthe Apocynaceae to be synonymous with the
tribe Apocyneae and a new name wouldbe needed for the taxon corresponding tothe current view of Apocynaceae (eg sensuEndress and Bruyns 2000) A node-baseddenition would handle this particular prob-lem but may have other problems as illus-trated here with another hypothetical worst-case scenario Early classications of theApocynaceae included Plocosperma BenthA node-based denition representing thisview could have included as reference taxasay Apocynum and Plocosperma The recentlyproposed position of Plocosperma close tothe Boraginaceae (Backlund et al 2000)would with this denition have made Apoc-ynaceae a very large taxon (containing atleast both Gentianales and Boraginales) pos-sibly synonymous with (a hypothetical def-inition of) the informal taxon EuasteridsII (Bremer et al 1998) and again a newname for the Apocynaceae sensu Endressand Bruyns (2000 ie excluding Plocosperma)would have been needed Because theLinnaean system allows recircumscription ofsynonymous names the potential turnoverin the set of names in use is reduced withthis system (albeit at the expense of allowingsubjectivity) Additionally the ranks of theLinnaean system provide a set of names thatwork as relatively stable universal standardnames in communication note that equalranks does not guarantee comparable evo-lutionary units however Continuity in us-age of names is important for example injournals literature databases oras teach-ing and inventory work and the need forsuch continuity has been recognized by pro-ponents of both nomenclature systems (deQueiroz and Gauthier 1992 Greuter et al1994 Reveal 1996)
We propose a compromise approach thatuses solutions relating to these aspects fromboth systems concentrating mainly onbotanical classication of extant plant taxaat the tribal and familial levels and onlyconsidering monophyletic taxa We recom-mend a system that uses ldquostandard namesrdquo(ie universally used communication unitsat convenient hierarchical levels) such asthose provided by the principal ranks of theLinnaean system (Greuter et al 1994) but inwhich sensitivity toward subjective changesin circumscription is reduced One way to dothis is to adopt the denitions of the phy-logenetic system but use constraints similar
398 SYSTEMATIC BIOLOGY VOL 51
to those that apply between ranks in theLinnaean system To reduce sensitivity to-ward changes in phylogenetic hypotheses(Schander and Thollesson 1995 Bryant1996 1997 Cantino et al 1997 Moore 1998)we propose using combined node C stem-based denitions that will provoke explicitincompatibilities between taxa in case ofunfortunate changes in circumscription (deQueiroz 1996 has proposed a similar type ofdenition for designating nonmonophyletictaxa somewhat similar variants are alsodiscussed in the Phylocode) A node-basedpart of the denition provides a minimalcircumscription whereas a stem-based partprovides boundaries to competing taxa(maximal circumscription) If on a particu-lar phylogeny the minimal (monophyletic)circumscriptions of competing taxa are over-lapping then the two taxa are incompatibleTo make this meaningful at least two taxashould be included in the node-based partof the denition this will also avoid unnec-essary monotypic taxa This incompatibilitywill provide an objective point when nomen-claturalreconsideration isneeded Returningto the rst of our examples above node Cstem-based denitions of the Apocynaceaeand Asclepiadaceae could be as follows
The family Apocynaceae is the most inclusive clade inthe order Gentianales including the type specimensof Apocynum and Dictyophleba but not the taxa Ascle-piadaceae [ Gentianaceae Rubiaceae ]
and
The family Asclepiadaceae is the most inclusive cladein the order Gentianales including the type speci-mens of Asclepias and Periploca but not the taxa Apoc-ynaceae [ Gentianaceae Rubiaceae ]
On the tree in Figure 1 these two denitionsare mutually exclusive and a decision onwhich name to keep must be made Mini-mizing the number of classicatory changesshould take precedence in this decision Thereason for this is to avoid the large incon-venient ldquostandard namerdquo changes that couldresult from relatively small rearrangementsRejected names are then ignored when oc-curring in denitions of other taxa This isbecause the reference taxa in the stem-basedpart of the denitions are ldquodenedrdquo taxarather than physical types (thus reliance ondened reference taxa which has been crit-icized for stem-based denitions by Sereno1999 is what we aim for on the other hand
the explicit reference of type specimens in theminimum circumscription provided by thenode-based part of the denitions circum-vents the tautology problem raised byBryant 1996) In the current example thiswould mean that regardless of which nameof Apocynaceae and Asclepiadaceae is re-jected the other will be dened as corre-sponding to the Apocynaceae sensu Endressand Bruyns (2000) The choice will there-fore simply be which name to keep In thiscase because Apocynaceae has priority un-der the Linnaean system we would chooseto keep Apocynaceae to promote consistencywith the current system Nevertheless out-side these recommendations choices of whatnames to choose will probably to a certaindegree be arbitrary We also propose thatthe rejected name should be reinstated withan emended denition if appropriate to po-tentially allow for more stability in the setof names in use Thus in our second ex-ample above a node C stem-based deni-tion reecting the inclusion of Plocospermain Apocynaceae (eg by exchanging Dic-tyophleba for Plocosperma in the denitionabove) would presumably be incongruentwith the denition of Gentianales as well asthe denitions of Rubiaceae Gentianaceaeand so forth and therefore would be rejectedIn such acasewhere the only real change is inthe position of a single taxon a reinstatementof the Apocynaceae with an emended def-inition is motivated However a drawbackis that this may also introduce elements ofsubjectivity in the circumscriptions of taxa
Not all named clades need to be providedwith standard names In fact for the purposeof a universal set used in journals databasesand so forth it may sufce to use standardnames corresponding to the principal ranksof the Linnaean system (Greuter et al 1994)such as species genus family and orderNevertheless to provide compatibility withthe present system we suggest that addi-tional names corresponding to commonlyused secondary ranks (eg tribes) may berecognized The hierarchical level of a stan-dard name needs to be indicated This doesnot imply that taxa of the same hierarchi-cal level are comparable evolutionary unitsbut simply indicates the hierarchical levelof a taxon relative to nested taxa Becausesuch misinterpretations will in any case bedifcult to prohibit one may as well pro-vide compatibility with the present Linnaean
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 399
system by using the names of the Linnaeanranks (eg species genus tribe familyorder) to indicate hierarchical level The prin-ciple of exhaustive subsidiary taxa is notaccepted so redundant taxa need not be rec-ognized Other ldquononstandardrdquo taxa couldbe dened by using Phylocode denitionsmdashthat is without the restraints discussed forstandard taxa above
One advantage of the described systemis that utilizing the correspondence betweenthe standard names and Linnaean ranks al-lows the system to be largely compatiblewiththe establishedLinnaean system A change ofsystem could therefore be gradual with thetwo systems coexisting during a conversiontime (see eg de Quieroz 1997b)
Even though the node C stem-baseddenitions reduce the sensitivity towardschanges in circumscriptions they may still besensitive towardldquounfortunaterdquo or ldquobadrdquo def-initions (see eg Cantino et al 1997Cantinoand de Queiroz 2000) Many of the recom-mendations for such things as choices of ref-erence taxa (speciers) in phylogenetic def-initions are applicable also to the node Cstem-based denitions (eg Schander andThollesson 1995 Bryant 1996 Cantino et al1997 Sereno 1999) Most likely the compro-mise system described above will have sev-eral further problems and we hope that thispaper will invite further discussion on thesubject
A New Classication of the Apocynaceae slOur results indicate that some problem-
atic taxa remain in the classication ofEndress and Bruyns (2000) Monophyly fortwo of the subfamilies and ve of the tribesis called into question If other recent molec-ular studies (eg Potgieter and Albert 2001)are taken into consideration a further fourtribes (Alstonieae Alyxieae Plumerieae andVinceae) may be nonmonophyletic This canin some cases be simply analysis artifactsbut as Endress and Bruyns (2000) themselvespoint out their classication is to be consid-ered preliminary Insufcient taxonsamplingin molecular studies and difcult homologydecisions relating to morphological charac-ters make the circumscription of many ofthe tribes for example in the Apocynoideaesensu Endress and Bruyns (2000) uncertainThere is thus risk for future rearrangementsin many of the tribes A classication using
node C stem-based denitions may be ableto subsume such rearrangements withoutnumerous explicit recircumscriptions Wewill here use the result fromthe present studyas a basis to propose such a classication ofthe Apocynaceae sl
Our aim is to base taxa on clades that arewell supported A second aim is to makeour classication compatible with the cur-rent system We will therefore primarily rec-ognize nonoverlapping tribes that are validunder the Linnaean system We will furtheradopt the principle of nested referencing (assuggested by Lee 1999 and Sereno 1999)from the Linnaean system Thus the rst ref-erence taxon in the node-based part of thedenition is the primary type which corre-sponds to the Linnaean type used for thename of the taxon Instead of the (optional)enumeration of subsumed taxa of a lowerrank (eg genera or subtribes) we will in-clude node C stem-based denitions as dis-cussed above Citations to primary types ofreference taxa in node-based part of de-nitions are also nested and thus refer toldquothe type specimen of the type species ofthe type genus of the dened taxonrdquo Aproblem is that the relationships between thetribes areweakly supported This may lead tovery cumbersome denitions enumeratingall competing tribes in the stem-based partof the denition (eg Moore 1998) One wayto reduce this problem is to dene interme-diate well-supported taxa and use them ascompeting taxaWe therefore will further rec-ognize four nonstandard taxa (note that anassignment of ldquononstandardrdquo taxa does notimply less reliability standard names relateto communication purposes only) The taxadiscussed are indicated in Figure 3 We willdiscuss the tree from the top of Figure 3 start-ing with the taxa of the traditional Plumeri-oideae Tribes and subtribes given in paren-theses in the text refer to the classicationsof Endress and Bruyns (2000) in other casesour classication is implied For some tribeswe have chosen a name with priority underLinnaean system that implies inclusion of agenus not sampled in the present study Forall but one (Willughbeieae) of these casesadditional molecular or morphological phy-logenetic analyses support the inclusion ofthe genus Even if too uncertain denitionsshould be avoided this is less critical withnode C stem-based denitions In the caseof Willughbeieae we have decided from
400 SYSTEMATIC BIOLOGY VOL 51
FIGURE 3 The tribal reclassication discussed in the text indicated on the combinable component consensustree from the successive weighting analysis of the complete rbcL data set Four nonstandard taxamdashApocynoidinaEuapocynoidina Asclepiadacina and Asclepiadoidinamdashare also indicated
discussions with a morphological expert(M Endress pers comm) that morphologi-cal support exists for including WillughbeiaRoxb in Willughbeieae If this is correctthe denition given below will be valid if
it is not our system allows redenition ofthe taxon We have also tried to providecomments on potential morphological char-acters taken from literature for the differ-ent taxa recognized However we have not
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 401
performed any morphological analysis butrely on information external to this studyfor these comments In most cases we cantherefore not identify synapomorphies forthe taxa
TAXONOMIC TREATMENT
Tribus Willughbeieae ADCThe Ancylobotrys Dictyophleba and Va-
hadenia clade is strongly supported (uwbootand swboot D 100) and corresponds to theWillughbeieae of Endress and Bruyns (2000)The present sampling contains closely re-lated genera that have been traditionallykept together Landolphia P Beauv sl in-cludes all these taxa (eg Schumann 1895)In addition to the taxa sampled here thestudy of Potgieter and Albert (2001) sup-ports the inclusion of Couma Aubl LacmelleaH Karst Pacouria Aubl and Saba (Pichon)Pichon in this clade Plants in this cladeall have a gynoecium that is congenitallysyncarpous and indehiscent and that con-tains a eshy pulp including numerous seedswith copious horny endosperm (Fallen 1986Persoon et al 1992) This probably con-stitutes synapomorphies for this clade(M Endress pers comm) although not un-ambiguously for example syncarpy is par-alleled in Carisseae We will therefore as-sume that the reference genus Willughbeia isincluded in this tribe
DenitionmdashTribe Willughbeieae is themost inclusive clade that includes the prim-ary types of Willughbeia and Dictyophleba butnot Alyxieae Aspidospermeae CarisseaeHunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Apocynoidina
Tribus Vinceae Bartl
Ochrosia Rauvola Vinca and Catharanthus(all Vinceae) form a strongly supported clade(uwboot D 89 swboot D 99) The associa-tion between Rauvola and Catharanthus wasindicated by Sennblad and Bremer (1996)The study of Potgieter and Albert (2001) sup-ports inclusion of the genera LaxoplumeriaMarkgr Neisosperma Raf and Tonduzia Pit-tier previously classied with AspidospermaThe Vinceae contain both herbaceous andfruticose taxa The fruits are apocarpousbut whereas the fruits of Catharanthus Lax-oplumeria Neisosperma Tonduzia and Vincahave dry fruit walls the fruits of Ochrosia andRauvola are drupes Plants in Vinceae share
a similar type of style head (ie the apicalenlarged part of the style) with a stigmatichollow and upper and lower hair wreaths(except in some species of Ochrosia) the pres-ence of a nectar disk and a linear hilum onthe seed (Pichon 1948b) On the present treeVinceae also include Kopsia however this po-sition of Kopsia is very weakly supported (seealso Sennblad and Bremer 1996)
DenitionmdashTribe Vinceae is the mostinclusive clade that includes the primarytypes of Vinca and Rauvola but not AlyxieaeAspidospermeae Carisseae HunterieaeMelodineae Plumerieae Tabernaemon-taneae Willughbeieae or Apocynoidina
Tribus Tabernaemontaneae G DonThe strongly supported association be-
tween Tabernaemontana Tabernanthe Carval-hoa Schizozygia and Molongum (Tabernae-montaneae uwboot D 84 swboot D 98)has been suggested previously (Boiteau et al1978 Fallen 1986 Endress et al 1996Sennblad and Bremer 1996) Inclusion of taxatraditionally associated with Tabernaemon-tana (eg Voacanga Thouars and CallichiliaStapf see also Tabernaemontaneae sensuLeeuwenberg 1994a) and Macoubea Aublis supported by the analysis of Potgieterand Albert (2001) The taxa share sclerenchy-matic anthers free from the style head anda characteristic band of heavily cutinizedcells at the insertion on the staminal rib(Endress et al 1996) The taxa tradition-ally associated with Tabernaemontana arecharacterized by their apocarpous fruitswith arillate seeds whereas genera asso-ciated with Molongum (Ambelanieae sensuLeeuwenberg 1994a) have syncarpous fruitsand lackanarillus HoweverMacoubea formsa link between the two having a syncar-pous fruit with arillate seeds (Zarucchi et al1995)
DenitionmdashTribe Tabernaemontaneae isthe most inclusive clade that includes the pri-mary types of Tabernaemontana and Schizozy-gia but not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae PlumerieaeVinceae Willughbeieae or Apocynoidina
Tribus Aspidospermeae Miers
Aspidosperma and Vallesia (Aspidosper-meae) form a strongly supported clade (uw-boot and swboot D 100) rst suggestedin Sennblad (1997 see also Sennblad and
402 SYSTEMATIC BIOLOGY VOL 51
Bremer 2000) Subsequently GeissospermumAllemao Haplophyton ADC MicroplumeriaBaill and Strempeliopsis Benth were indi-cated to belong to this clade (Potgieter andAlbert 2001) The Aspidospermeae havesimple style heads lacking both stigmatichollow and upper hair wreath and haveapocarpous fruits with variation in eshi-ness and sclerication Alstonia has tradi-tionally been placed in the Plumerieae Be-cause its seeds have a hairy margin it hasbeen suggested to form a link to subfam-ily Apocynoideae This was contradicted bySennblad and Bremerrsquos study (1996) how-ever where Alstonia was placed in an iso-lated position as the sister group to therest of the Apocynaceae sl Here it groupswith Aspidospermeae Sensu Endress andBruyns (2000) However this association isvery weakly supported (uwboot and swbootlt 50) and in other recent studies As-pidospermeae and Alstonia do not form aclade (Potgieter and Albert 2001) Becauseof this we do not use it as a reference taxonin the denition Thus although the Aspi-dospermeae include Alstonia on the presenttree some other position of Alstonia canbe accommodated without amending thedenition
DenitionmdashTribe Aspidospermeae is themost inclusive clade that includes theprimary types of Aspidosperma and Vallesiabut not Alyxieae Carisseae HunterieaeMelodineae Plumerieae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Melodineae G DonThe well-supported association between
Melodinus and Craspidospermum (uwboot D70 swboot D 75) was rst reportedby Sennblad (1997 see also Sennblad andBremer 2000) and has subsequently beenreected in the tribe Melodineae sensuEndress and Bruyns (2000 their inclusion ofDiplorhynchus however is not supported)Craspidospermum although its fruit is dryand dehiscent as opposed to the indehis-cent fruits of Melodinus was noted by Pichon(1948b) to present several characters of theCarisseae such as stamens inserted near thebase of the corolla tube and a syncarpousovary In particular Pichon found the verydense suprastaminal indumentum type rem-iniscent of Melodinus From his descriptions
of the two genera (Pichon 1948ab) the fol-lowing similarities also emerge presence ofa stipular line pollen in tetrads and a punc-tiform hilum
DenitionmdashTribe Melodineae is the mostinclusive clade that includes the pri-mary types of Melodinus and Craspidosper-mum but not Alyxieae AspidospermeaeCarisseae Hunterieae Plumerieae Taber-naemontaneae Vinceae Willughbeieae orApocynoidina
Tribus Hunterieae K Schum
The Picralima and Pleiocarpa clade isstrongly supported (uwboot and swboot D100) and corresponds to the subtribeHunterieae sensu Endress andBruyns (2000)The study of Potgieter and Albert (2001)also supports the inclusion of Hunteria RoxbThese taxa are characterized by apocar-pous sometimes pluricarpous ovaries andeshy brous fruit walls (Omino 1996)Diplorhynchus (Melodineae) is here weaklyassociated (uwboot and swboot lt 50) withthe Pleiocarpeae It shares a few similari-ties such as stipular lines and a style headwithout a stigmatic hollow and hair wreathsbut also has differences such as a dry de-hiscent fruit With the present denitionDiplorhynchus will tentatively be included inthe Hunterieae
DenitionmdashTribe Hunterieae is the mostinclusive clade that includes the primarytypes of Hunteria and Picralima but not Alyx-ieae Aspidospermeae Carisseae Melod-ineae Plumerieae TabernaemontaneaeVinceae Willughbeieae or Apocynoidina
Tribus Plumerieae EndlThe strongly supported relationship
(uwboot and swboot D 100) between Alla-manda and Plumeria corresponds to earlierstudies (Endress et al 1996 Sennblad andBremer 1996 Civeyrel et al 1998 Potgieterand Albert 2001) In all earlier treatmentsAllamanda has had uncertain relationshipsThe association with Plumeria nds supportin pollen morphology such as perforatemesocolpial depressions and similar innerexine pattern The two genera also containsecoiridoids rather than the indole alka-loids and cardenolides that are commonin the traditional Plumerioideae (Endresset al 1996) Cameraria Cerbera and Thevetiaform a strongly supported association
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 403
(uwboot D 83 swboot D 91) with Ane-chites which has been suggested by Fallen(1983) This association was based mainlyon the latrorsesublatrorse anthers andthe broad style head with large apicalappendages and stigmatic hollow The studyby Potgieter and Albert (2001) suggesteda further inclusion of Cerberiopsis Viell ExPancher amp Sebert and Skytanthus Meyen inthis tribe These taxa form a well-supportedclade with Plumeria and Allamanda whichcorresponds to the Plumerieae sensuEndress and Bruyns (2000) Characterssupporting this relationship are the presenceof infrastaminal as well as suprastaminalappendages (not present in Anechites andPlumeria) and winged seeds (not present inAnechites) (Endress et al 1996)
DenitionmdashTribe Plumerieae is themost inclusive clade that includes theprimary types of Plumeria and Allamandabut not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Carisseae Dumort
Acokanthera and Carissa (both Carisseae)are traditionally joined and are sometimeseven treated as one genus (eg Pichon1948a) In the present study they form astrongly supported clade (uwboot D 96swboot D 99) Among the characters sup-porting this tribe are syncarpous fruits with-out axile placentation and endocracks on theinner exine of the pollen (Endress et al 1996)
DenitionmdashTribe Carisseae is the most in-clusive clade that includes the primary typesof Carissa and Acokanthera but not AlyxieaeAspidospermeae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
Tribus Alyxieae G Don
Chilocarpus has been a genus with un-certain afnities In Endress et al (1996) itwas the sister taxon to the ApocynoideaeHere it is associated with Alyxia and Lep-inia in a clade corresponding to the Alyxieaesensu Endress and Bruyns (2000) Thisclade is strongly supported with succes-sive weighted bootstrap (swboot D 91) butnot with unit-weighted bootstrap support(uwboot D 58) such clades will henceforthbe termed moderatelywell supported clades
A further inclusion of Condylocarpon DesfLepiniopsis Valeton and Plectaneia Thouarswas indicated by Potgieter and Albert (2001)The constituent taxa are the only Apocy-naceae taxa outside the Apocynoidina thathave porate pollen
DenitionmdashTribe Alyxieae is the most in-clusive clade that includes the primary typesof Alyxia and Lepinia but not Aspidosper-meae Carisseae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
APOCYNOIDINA
On the present tree the traditionalsubfam-ily Apocynoideae includes all subfamilies(Periplocoideae Secamonoideae and Ascle-piadoideae) of the Asclepiadaceae sensuLiede and Albers (1994) All of these taxaare established names and selecting one asa subfamily at the expense of the othersmight be considered unfortunate from dif-ferent points of view We have here chosennot to recognize subfamilies instead we willdene the traditional Apocynoideae Ascle-piadaceae and Asclepiadoideae as nonstan-dard taxa The correspondence to the tradi-tional taxa will be indicated in the names byreplacing the sufx -eae with a neutral suf-x -ina (Kron 1997) We will also for practi-cal reasonsrecognize as a nonstandardtaxonthe informal group euapocynoids suggestedby Sennblad et al (1998)
The taxa of the traditional Apocynoideaeand Asclepiadaceae (uwboot D 86 swboot D99) clade will in many cases be more dif-cult to safely delimit to tribes becausemany of the groups in this clade have weakor no support (although preliminary resultsfrom an unpublished collaborative study bySennblad et al indicate additional supportfor the tribes discussed below) The tribaland subtribal classication of the traditionalApocynoideae has been shown to be prob-lematic (Leeuwenberg 1994a Endress andBruyns 2000) The circumscription of tribesfor these taxa might therefore in some casesbe preliminary
In all following taxa the basal part of theanther connective called the retinacle is ad-nate to the style head
DenitionmdashApocynoidina are the mostinclusive clade that includes the pri-mary types of Apocynum and Wrightia butnot Alyxieae Aspidospermeae Carisseae
404 SYSTEMATIC BIOLOGY VOL 51
Hunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Willughbeieae
Tribus Wrightieae G DonThe Stephanostema and Wrightia (both
Wrightieae) clade is strongly supported(uwboot and swboot D 100) and correspondsto the Wrightieae sensu Sennblad et al (1998)who also suggested an inclusion of Pleio-ceras Baill The tribe could be characterizedby a combination of synapomorphic and ple-siomorphic characters for example presenceofa chalazalandabsence of micropylar comaleft contorted aestivation and absence of airspaces in the anthers (1998)
DenitionmdashTribe Wrightieae is the mostinclusive clade in the Apocynoidina that in-cludes the primary types of Wrightia andStephanostema but not Malouetieae Nerieaeor Euapocynoidina
Tribus Nerieae (Benth) M Pichon
Adenium Nerium and Strophanthus (allWrightieae) constituted the Nerieae inSennblad et al (1998) however the cladereceived weak support Here Adenium andNerium form a clade that is absent in halfof the most-parsimonious trees in the com-plete rbcL analysis but is strongly supportedin the combined ndhF and rbcL analysis(uwboot D 98 swboot D 99) Strophan-thus groups with the Mascarenhasia Pachy-podium Kibatalia Funtumia and Holarrhenaclade which corresponds to the Malouetieaesensu Sennblad et al (1998) This position ofStrophanthus is however weakly supported(uwboot and swboot lt 50) and is not presentin the combined ndhF and rbcL tree Fur-thermore certain morphological characters(unfused slits in the corolla tube and anapical anther appendage) indicate a posi-tion closer to the Nerieae (Sennblad et al1998)
DenitionmdashTribe Nerieae is the most in-clusive clade in the Apocynoidina that in-cludes the primary types of Nerium and Ade-nium but not Malouetieae Wrightieae orEuapocynoidina
Tribus Malouetieae Mull-Arg
The association of Mascarenhasia Pachy-podium Kibatalia Funtumia and Holar-rhena (all Wrightieae) is moderately wellsupported in the complete rbcL analysis
(uwboot D 57 swboot D 74) whereasthe corresponding clade is well supportedin the combined rbcL and ndhF analysis(uwboot D 66 swboot D 100) On the ba-sis of identied synapomorphies (presenceof calcium oxalate crystals in the stomium ofthe anthers absence of interpetal vein ses-sile laments and absence of air spaces inthe anthers) Sennblad et al (1998) predictedan inclusion of Alaa Thouars Kibatalia andMalouetia ADC in the tribe The inclusion ofKibatalia is strongly supported in this studyFollowing these suggestions we will here as-sume the inclusion of the reference genusMalouetia in this tribe (The denition belowwill include Strophanthus on the present treebut see comments under Nerieae)
DenitionmdashTribe Malouetieae is the mostinclusive clade in the Apocynoidina that in-cluded the primary types of Malouetia andFuntumia but not Nerieae Wrightieae orEuapocynoidina
EUAPOCYNOIDINA
In all the remaining representatives ofthe Apocynoideae sensu Endress andBruyns (2000) that is Mandevilla MesechitesAganosma Apocynum TrachelospermumRhabdadenia Beaumontia Parsonsia andPrestonia the anthers are adnate to the stylehead both by the retinacle and by the thecaeThese taxa together with the taxa of the tra-ditional Asclepiadaceae form a moderatelysupported clade (uwboot lt 50 swboot D86) that corresponds to the informal taxoneuapocynoids sensu Sennblad et al (1998)
DenitionmdashEuapocynoidina are the mostinclusive clade that includes the primarytypes of Apocynum and Echites but notMalouetieae Wrightieae or Nerieae
Tribus Mesechiteae Miers
Mandevilla and Mesechites (both Mese-chiteae) form a moderately supported clade(uwboot lt 50 swboot D 87) equivalentto the Mesechiteae This tribe could be char-acterized by anthers with obtuse truncatedtails and a retinacle lacking hairs and also astrongly pentagonal style head with a stig-matic hollow
DenitionmdashTribe Mesechiteae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Mesechitesand Mandevilla but not Apocyneae EchiteaePeriploceae or Asclepiadacina
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 405
Tribus Apocyneae Bercht et J Presl
Aganosma Apocynum Beaumontia andTrachelospermum (Apocyneae) form a weaklysupported clade with Rhabdadenia (Echiteae)(uwboot and swboot lt 50) in the completerbcL analysis whereas the correspondingclade in the combined ndhF and rbcL analy-sis gains a little better support (uwboot lt 50swboot D 91) Likewise this clade (with theinclusion of Chonemorpha G Don) is weaklysupported in Potgieter and Albert (2001) Po-tential characterizing traits could be a stylehead without stigmatic hollow in combina-tion with thecae that are adnate to the stylehead (Sennblad et al 1998)
DenitionmdashTribe Apocyneae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Apocynumand Trachelospermum but not Echiteae Mese-chiteae Periploceae or Asclepiadacina
Tribus Echiteae Bartl
Parsonsia Peltastes and Prestonia (allEchiteae) form a moderately supported orunsupported clade with the Periplocoideaerepresentatives of the study (uwboot andswboot lt 50 and uwboot lt 50 swboot D89 in the complete rbcL and the combinedrbcL and ndhF analysis respectively) In thestudy of Sennblad et al (1998) Prestonia andParsonsia formed a weakly supported cladeand in the study of Potgieter and Albert(2001) inclusion of both Parsonsia and EchitesP Browne in the Echiteae is supportedParsonsia and Echiteae also share certaincharacters such as a horseshoe-shaped reti-nacle and a style head with stigmatic hollow(see Sennblad et al 1998)
DenitionmdashTribe Echiteae is the most in-clusive clade in the Euapocynoidina thatincludes the primary types of Echites andPrestonia but not Apocyneae MesechiteaePeriploceae or Asclepiadacina
Tribus Periploceae BartlThe representatives of the Periplo-
coideae sensu Endress and Bruyns (2000)mdashPentopetia Parquetina Petopentia PeriplocaTacazzea and Mondiamdashgroup together ina well-supported clade (uwboot D 67swboot D 93) The monophyly of the tra-ditional periplocoid taxa is also supportedin earlier studies that included a wider ordifferent sampling (Civeyrel et al 1998
Potgieter and Albert 2001) The taxa ofthe traditional Asclepiadaceae have ldquotrans-latorsrdquo structures related to pollinationspecialization In the Periploceae the pollenis deposited as tetrads in a spoon-liketranslator with an adhesive disc
DenitionmdashTribe Periploceae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Periplocaand Pentopetia but not Apocyneae EchiteaeMesechiteae or Asclepiadacina
ASCLEPIADACINA
The taxa of the Asclepiadoideae sensuLiede and Albers (1994) form a clade togetherwith the two Secamone species and Baissea(Apocyneae uwboot lt 50 swboot D 73)The exclusion of Periploceae (traditionallyassociated with this clade see Civeyrel et al1998) from this taxon is weakly supportedTherefore Periploceae is not included as areference taxon in the stem-based part ofthe denition of the Asclepiadacina thusleaving open the possibility for a later in-clusion of this taxon In the Asclepiadacina(except Baissea) the pollen is agglutinatedinto pollinia that are connected to a claspingtranslator
DenitionmdashAsclepiadacina are the mostinclusive monophyletic clade in the Euapoc-ynoidina that includes the primary types ofAsclepias and Secamoneae but not ApocyneaeEchiteae or Mesechiteae
Tribus Secamoneae G DonThe tribe Secamoneae sensu Bruyns
(uwboot D 89 swboot D 96) is mono-phyletic in the present analysisHowever be-cause representatives from only one of theconstituent genera are included monophylyis not well tested but the study of Civeyrelet al (1998) which included a larger numberof taxa including Pervillea Decne showedthat monophyly of the tribe was supportedThe traditional character for this tribe is fourpollinia per translator
DenitionmdashTribe Secamoneae is the mostinclusive clade in the Asclepiadacina that in-cludes the primary types of Secamone andPervillea but not Asclepioidina or Baissea
BaisseaThe position of Baissea corresponds with
that in the study of Sennblad et al (1998)
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
390 SYSTEMATIC BIOLOGY VOL 51
ourselves with familytribal-level plant clas-sication Likewise our use of the term theLinnaean system mainly refers to the In-ternational Code of Botanical Nomenclature(ICBN Greuter et al 1994) We will notdiscuss species-level nomenclature nor thatof extinct taxa (see eg de Queiroz andGauthier 1992 Eriksson et al 1998 Mishler1999)
The Apocynaceae sl belong to the or-der Gentianales and have a mainly pantrop-icalsubtropical distribution with a fewtemperate representatives The family com-prises many well-known ornamentals suchas Nerium (oleander) and Hoya (wax ower)The plants are typically laticiferous and pro-duce various alkaloids and cardenolidessome of which have medical properties Themost well-known example is Catharanthus(rose periwinkle) which contains vinblas-tine and vincristine compounds now usedworldwide to treat childhood leukemia
The family Asclepiadaceaewas segregatedfrom the Apocynaceae sl by Brown (1810)Although some authors have questioned thisseparation (eg Demeter 1922 Safwat 1962Judd et al 1994 Struwe et al 1994) mostclassications subsequent to that of Brownhave followed his ldquotwo-familyrdquo treatmentMolecular studies of the Apocynaceae sl(Sennblad and Bremer 1996 Civeyrel et al1998 Potgieter and Albert 2001) have indi-cated that the Asclepiadaceae form a sub-clade of the Apocynaceae sstr thus render-ing the Apocynaceae sstr nonmonophyleticThese studies also indicated major problemswith the infrafamilial classicationin the twofamilies A detailed discussion of differentclassications of the Apocynaceae sstr andthe Asclepiadaceae is given in Sennblad andBremer (1996) and only a short review willbe given here concentrating on the most re-cent classications and results from molecu-lar studies
Since the classication of Schumann(1895) the Apocynaceae have tradition-ally been divided into two subfamilies thePlumerioideae and the Apocynoideae andapproximately into 12 tribes and 27 sub-tribes (Pichon 1948abc1950Leeuwenberg1994a) The studies by Sennblad and Bremer(1996) and Endress et al (1996) showed thatboth the Plumerioideae and Apocynoideaeare nonmonophyletic Four of the tribesof the Plumerioideae have been shown tobe nonmonophyletic (Endress et al 1996
Sennblad and Bremer 1996 Civeyrel et al1998 Potgieter and Albert 2001) threetribes were monogeneric and the mono-phyly of one tribe has not been evalu-ated The study of Sennblad et al (1998)was aimed mainly at the tribe Wrightieae(Apocynoideae) sensu Leeuwenberg (1994a)but also included representatives of the othertribes of the Apocynoideae as well as of theAsclepiadaceae Sennblad et al showed thatnone of the described tribes within the Apoc-ynoideae is monophyletic and suggesteda reclassication of the Wrightieae sensuLeeuwenberg (1994a)into the Wrightieae theNerieae and the Malouetieae
The infrafamilial classication of Ascle-piadaceae proposed by Brown (1810) whendescribing the family has been followedby most subsequent authors (Liede andAlbers 1994) and comprises three subfam-ilies of which the Periplocoideae and Se-camonoideae are monotribal and the As-clepiadoideae usually comprise ve tribesAn exception from this traditional view isthe classication of Swarupanandan et al(1996) which reduced the number of tribesof the Asclepiadoideae to two and includedthe Secamonoideae as a third tribe in theAsclepiadoideae Furthermore a revision ofthe Periplocoideae by Venter and Verhoeven(1997) recognizes three tribes in this subfam-ily The monophyly of the Asclepiadaceaeis uncertain in the study of Sennblad andBremer (1996) however the subfamilyPeriplocoideae itself monophyletic did notform a monophyletic group with the othertwo subfamilies (sensu Liede and Albers1994) whereas in the study of Civeyrelet al (1998) the three subfamilies did form amonophyletic group The study of Civeyrelet al (1998) also indicated that the tradi-tional Secamonoideae is monophyletic witha position as sister to Asclepiadoideae (seealso Sennblad and Bremer 1996 Potgieterand Albert 2001) The traditional Asclepi-adoideae receive good support (Sennbladand Bremer 1996 Civeyrel et al 1998Potgieter and Albert 2001) Of the tribesof the Asclepiadoideae in the classicationof Liede and Albers (1994) Asclepiadeaeand Marsdenieae have been shown to benonmonophyletic
The ndings of the molecular studies indi-cated above and also from a previous versionof this paper (included in the PhD thesisof BS 1997 see also Sennblad and Bremer
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 391
2000) were summarized and developed fur-ther by use of morphology in the most re-cent classication of the Apocynaceae sl byEndress and Bruyns (2000) The ve tradi-tional subfamilies were kept but 19 newlyrecircumscribed tribes were also recognized(see Table 1)
However the results of the present studyindicate that although the classication ofEndress and Bruyns (2000) are a major ad-vance in making the classication congru-ent with current best estimates of phylogenyproblematic cases still need revision Fur-thermore for many of the recognized tribestaxon sampling for molecular data is weakand the morphological homologies in thesegroups are difcult to interpret (Endress andBruyns 2000) Further rearrangements oftribal circumscription will almost certainlybe needed in the future
The Apocynaceae sl may thus provide asuitable test for the classication system pro-posed in this study We therefore propose anew classication of the Apocynaceae basedmainly on an rbcL analysis that includesextended taxon sampling as well as addi-tional data from length variation of rbcL in its30-end and downstream Certain relation-ships of special interest have been tested witha smaller data set for the subfamily Apoc-ynoideae and the traditionalAsclepiadaceaeincluding additionalnucleotide sequences ofthe chloroplast gene ndhF This gene oftenhas a greater substitution rate than rbcL (seeOlmstead and Sweere 1994 Kim and Jansen1995) and may therefore form a complementto rbcL when the latter proves too conservedWe have used only the 3rsquo region of the genebecause that region appears to show the mostvariability (see Kim and Jansen 1995) Fur-thermore results from other recent molecu-lar studies have also been taken into accountin the classication
MATERIALS AND METHODS
We sampled 77 representatives of theApocynaceae sl (Table 1) All subfamiliesand tribes of the classication of Endress andBruyns (2000) are represented in the analy-sis Gelsemium Juss and Mostuea Didr of theGelsemiaceae were chosen as outgroup taxabecause they were indicated to be the closest
TABLE 1 Classication of the Apocynaceae sl andsampling of taxa Classication are according to Endressand Bruyns (2000)
RAUVOLFIOIDEAEAlstonieae Melodineae
Alstonia RBr Craspidospermum ADCAspidosperma Mart Diplorhynchus Ficalho
et Zucc et HiernVallesia Ruiz Melodinus JR Forst
et Pav et G ForstVinceae Hunterieae
Catharanthus G Don Picralima PierreKopsia Blume Pleiocarpa BenthOchrosia Juss PlumerieaeRauvola L Allamanda LVinca L Anechites Griseb
Willughbeieae Cameraria LAncylobotrys Pierre Cerbera LDictyophleba Pierre Plumeria LVahadenia Stapf Thevetia Adans
Tabernaemontaneae CarisseaeCarvalhoa K Schum Acokanthera G DonMolongum Pichon Carissa LSchizozygia Baill AlyxieaeTabernaemontana L Alyxia RBrTabernanthe Baill Chilocarpus Blume
Lepinia DecneAPOCYNOIDEAE
Wrightieae ApocyneaeAdenium Roem Aganosma (Blume)
et Schult G DonNerium L Apocynum LStephanostema K Schum Baissea ADCStrophanthus DC Beaumontia WallWrightia RBr Trachelospermum Lem
Malouetieae MesechiteaeFuntumia Stapf Mandevilla LindlHolarrhena RBr Mesechites Mull-ArgKibatalia GDon EchiteaeMascarenhasia ADC Parsonsia RBrPachypodium Lindl Peltastes Woodson
Prestonia RBrRhabdadenia Mull-Arg
PERIPLOCOIDEAEMondia Skeels Periploca LParquetina Baill Petopentia BullockPentopetia Decne Tacazzea Decne
SECAMONOIDEAESecamone RBr
ASCLEPIADOIDEAEMarsdenieae Asclepiadeae
Fockea Endl Araujia BrotHoya RBr Asclepias LMicholitzia NEBr Calotropis RBrStephanotis Thouars Orthosia Decne
Ceropegieae Fischeria DCCeropegia L Matelea AublStapelia L Schizostephanus Benth
et Hook fTweedia Hook et ArnTylophora RBrVincetoxicum Wolf
392 SYSTEMATIC BIOLOGY VOL 51
sister group to the Apocynaceae sl in therecent analysis of Gentianales by Backlundet al (2000)
Forty-one new sequences of rbcL are pub-lished in this study rbcL sequences forthe outgroup taxa and for 35 ingroup taxawere published previously (Table 2) ForndhF the 30 region of the gene was se-quenced for 18 taxa representing the tra-ditional Apocynoideae and AsclepiadaceaeTotal DNA was extracted from fresh leavesor herbarium material by using the meth-ods of Saghai-Maroof et al (1984) and Doyleand Doyle (1987) The extractions were pu-ried by ultracentrifugation in CsCl gra-dients or ethanol precipitation Additionalpurication with the Qiaquick PCR puri-cation kit (Qiagen Inc) was performed incases with problematic polymerase chain re-action (PCR) amplication Double-strandedDNA was amplied by PCR by using Taq-polymerase kit (Promega Corp) SyntheticPCR primer sequences for rbcL were takenfrom Olmstead et al (1992) ndhF primerswere taken from Olmstead and Sweere (1994)and Oxelman et al (1999) In cases wherePCR amplication proved difcult a PCRreaction using Taq extender PCR additive(Stratagene Inc) was performed followingthe protocol provided by the supplier A sec-ond round of PCR with only one of theprimers was performed to obtain single-stranded DNA (Kaltenboeck et al 1992)Single-stranded DNA was sequenced man-ually (Sanger et al 1977) by using internalprimers designed for rbcL by G Zurawski(DNAX Research Institute) and for ndhF byOlmstead and Sweere (1994) and Oxelmanet al (1999)
The length of the sequences of rbcL ex-cluding the 26 rst nucleotides but includ-ing positions just downstream from the geneobtained in this study varied between 1411and 1509 nt For all taxa except Parsonsia thesequence includes the stop codon In posi-tions 27ndash1425 no structural mutations occurin rbcL Alignment of this part of the geneis thus very simple However after position1425 that is just before the typical positionof the stop codon gaps of different sizes oc-cur complicating alignment Similarly gapsoccur in the coding region of the ndhF se-quences Because of these alignment prob-lems the sequences were truncated 23 basesbefore the typical position of the stop codonin Apocynaceae sl The reading frame was
inferred through comparisonswith completeApocynaceae ndhF-sequences from the studyof Backlund et al (2000)
Multiple sequence alignment was made byhand to reduce the number of gaps while in-creasing the percentage similarity We usedthe criteria of Golenberg et al (1993) with mi-nor modications
Two cladistic analyses were performedThe data matrix for the rst analysis coveredall included taxa and comprised characterscorresponding to the nucleotide positions27ndash1425 of the rbcL gene and characterscorresponding to gaps and informative nu-cleotide positions downstream of position1425 EMBL database accession numbersfor rbcL sequences are given in Table 2Only parsimony informative characters wereanalyzed
The second analysis was of a subset ofthe taxa included in the rst analysis andincludes 18 genera representing the Apoc-ynoideae sensu Leeuwenberg (1994a) andthe traditional Asclepiadaceae The data ma-trix consisted of two submatrices One sub-matrix comprised positions 27ndash1425 of therbcL gene The second submatrix comprised733 aligned nucleotide positions from the3rsquo region of ndhF EMBL database accessionnumbers for the ndhF sequences are given inTable 2 Only parsimony informative charac-ters were analyzed
The cladistic analyses were performed us-ing PAUP 40b2 (Swofford 1998) An initialheuristic search (PAUP settings HSEARCH[ADDSEQ D RANDOM NREPS D 200SWAP D TBR] other options withdefault set-tings) with all characters given a unit weightwas followed by a successiveweighting anal-ysis (Farris 1969 1989) also using heuris-tic searches (PAUP settings HSEARCH[ADDSEQ D RANDOM NREPS D 10SWAP D TBR] other options withdefault set-tings) and characters reweighted accordingto their rescaled consistency index
Bootstrap frequencies (Felsenstein 1985)were calculated with 10000 replicates on theunit-weighted (uwboot) and successiveweighted (swboot) data matrices (PAUPsettings BOOTSTRAP [NREPS D 10000METHOD D HEURISTIC CONLEVEL D50] [ADDSEQ D RANDOM NREPS D 1SWAP D SPR NOMULPARS] other optionswith default settings)
To simplify the discussion of the resultswe used a rough scale of the relative support
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 393
TABLE 2 Vouchers andEMBLaccession numbers for sequences published in this study Forpreviously publishedsequences and for sequences from previously published voucher specimens the reference is given
Taxon Vouchersourceb EMBL number rbcLndhF
Acokanthera oblongifolia (Hochst) Codd Sennblad and Bremer 1996 X91758-Adenium obesum (Forssk) Roem et Schult Sennblad et al 1998 AJ002880AJ420130Aganosma marginata (Roxb) G Don Ryding 490 UPS AJ419730-Allamanda cathartica L Sennblad and Bremer 1996 X91759-Alstonia scholaris RBr Sennblad and Bremer 1996 X91760-Alyxia ruscifolia RBr Sennblad 236 UPS AJ419731-Ancylobotrys petersiana Pierre Sennblad 201 UPS AJ419732-Anechites nerium Urb Bremer et al 3386 UPS AJ419733-Apocynum cannabinum L Sennblad and Bremer 1996 X91761AJ420113Araujia hortorum E Fourn Bremer 3006 UPS AJ419734-Asclepias curassavica L Sennblad and Bremer 1996 X91774-Aspidosperma triternatum Rojas Acosta Bremer 3029 UPS AJ419735-Baissea leonensis Benth Sennblad et al 1998 Yallah 112 UPSa AJ002881AJ420114Beaumontia grandiora Wall Sennblad et al 1998 AJ002882AJ420115Calotropis procera (Aiton) WT Aiton Sennblad 222 UPS AJ419736-Cameraria latifolia L Houghton 1085 FTG AJ419737-Carissa bispinosa (L) Merxm Sennblad 235 UPS AJ419738-Carvalhoa campanulata K Schum Sennblad 217 UPS AJ419739-Catharanthus roseus G Don Sennblad and Bremer 1996 X91757-Cerbera venenifera Steud No voucher AJ419740-Ceropegia woodii Schltr Sennblad and Bremer 1996 X91775-Chilocarpus suaveolens Blume Endress et al 1996 X92445-Fockea multiora K Schum Specks 248 cult MSTRc AJ419741-Craspidospermum verticillatum Bojer Pettersson and Nilsson 742 AJ419742-Dictyophleba lucida Pierre Sennblad and Bremer 1996 X91762-Diplorhynchus condylocarpon (Mull-Arg) Pichon Sennblad 203 UPS AJ419743-Fischeria stellata (Vell) E Fourn Clark and Watt 793 UPS AJ419744-Funtumia elastica Stapf Sennblad et al 1998 AJ002884-Gelsemium sempervirens Aiton Olmstead et al 1993 L14397-Holarrhena pubescens G Don Sennblad et al 1998 AJ002884-Hoya bella Hook Sennblad and Bremer 1996 X91776-Kibatalia gitingense (Elmer) Woodson Liede 3268 Z AJ419745-Kopsia fruticosa ADC Sennblad and Bremer 1996 X91763-Lepinia taitensis Decne Perlman 15071 NTBG AJ419746-Mandevilla sanderi (Hemsl) Woodson Sennblad and Bremer 1996 X91764AJ420116Mascarenhasia arborescens ADC Sennblad et al 1998 AJ002885AJ420117Matelea hirsuta (Vahl) Woodson Sennblad 263 UPS AJ419747-Melodinus monogynus Roxb Sennblad 261 UPS AJ419748-Mesechites trida Mull-Arg Bremer et al 3351 UPS AJ419749-Micholitzia obcordata NEBr Bremer 3010 UPS AJ419750-Molongum laxum (Benth) Pichon Sennblad and Bremer 1996 X91765-Mondia ecornuta (NEBr) Bullock Sennblad 215 UPS AJ419751-Mostuea brunonis Didr Olmstead et al 1993 L14404-Nerium oleander L Sennblad 265 UPS AJ002886AJ420118Ochrosia coccinea Miq v d Laan 389 WAG AJ419752-Orthosia serpyllifolia Decne Bremer et al 3372 UPS AJ419753-Pachypodium lamerei Drake Sennblad et al 1998 AJ002887AJ420119Parquetina nigrescens (Afzel) Bullock Sennblad and Bremer 1996 X91777-Parsonsia heterophylla A Cunn Sennblad et al 1998 AJ002888-Peltastes peltatum (Vell) Woodson Sennblad 262 UPS AJ419754-Pentopetia sp No voucher AJ419755-Periploca graeca L Sennblad et al 1998 AJ002889AJ420120Petopentia natalensis (Schltr) Bullock Sennblad 237a and sn cult MSTR AJ419756AJ420121Picralima nitida T et H Dur Sennblad and Bremer 1996 X91766-Pleiocarpa mutica Benth Bremer 3017 UPS AJ419757-Plumeria inodora Jacq Sennblad and Bremer 1996 X91767-Prestonia quinquangularis Spreng Sennblad and Bremer 1996 X91768AJ420122Rauvola mannii Stapf Sennblad and Bremer 1996 X91769-Rhabdadenia biora Mull-Arg Zona 616 FTG AJ419759AJ420123Schizostephanus alatus K Schum Cult MSTR AJ419758-Schizozygia coffaeoides Baill Sennblad 207 UPS AJ419760-Secamone afzelii (Schult) K Schum Sennblad and Bremer 1996 X91779AJ420124
394 SYSTEMATIC BIOLOGY VOL 51
TABLE 2 Continued
Taxon Vouchersource EMBL number rbcLndhF
Secamone geayi Constatin et Gallaud Civeyrel 1200 LC AJ419761-Stapelia leendertziae NEBr Sennblad and Bremer 1996 X91778-Stephanotis oribunda Brongn Sennblad 256 UPS AJ419762AJ420125Stephanostema stenocarpum K Schum Sennblad and Bremer 1996 X91770AJ420126Strophanthus eminii Pax Sennblad and Bremer 1996 X91771AJ420127Tabernaemontana divaricata Roem et Schult Sennblad and Bremer 1996 X91772-Tabernanthe iboga Baill Leeuwenberg 12544 WAG AJ419763-Tacazzea apiculata Oliver Venter 9188 cult MSTR AJ419764-Thevetia peruviana (Pers) K Schum Sennblad and Bremer 1996 X91773-Trachelospermum jasminoides (Lindl) Lem Sennblad et al 1998 AJ002890AJ420128Tweedia coerulea Sweet Sennblad 254 UPS AJ419765-Tylophora sylvatica Decne Sennblad and Bremer 1996 X91789-Vahadenia caillei (AChev) Hutch et Dalziel Leeuwenberg 12275 WAG AJ419766-Vallesia antillana Woodson Meagher 966 FTG AJ419767-Vinca minor L Sennblad 230 UPS AJ419768-Vincetoxicum hirundinaria Medik Sennblad 257 UPS AJ419769-Wrightia arborea (Dennst) Mabb Sennblad et al 1998 AJ002891AJ420129
aVoucher for rbcL sequence onlybHerbarium abbreviations are according to Index Herbariorum except LC D Private herbarium of Laure CiveyrelcThis specimen was originally determined as Cibirhiza albersiana in Kunze et al (1994) but has recently been redetermined as
Fockea multiora K Schum (Verhoeven et al 2002)
of the clades in the successive weightinganalysis Clades with a successive weightedbootstrap gt63 were considered well sup-ported those with a successive weightedbootstrap gt85 were considered stronglysupported In a theoretical context a cladesupported by one uncontradicted charactercorresponds to a bootstrap support of 63anda cladesupported by two uncontradictedcharacters to 85 (Harshman 1994 Farriset al 1996) Here however this scale merelyrelates to the support distribution within theanalysis
RESULTS
The unit weight analysis of the completerbcL data set (237 characters total 214 ofthe cells were scored as missing data) re-sulted in 19003 trees 828 steps long witha consistency index (ci) D 0378 and a re-tention index (ri) D 0678 The consensustree from the successive weighting analysisof 252 most-parsimonious trees (each 1773steps long ci D 0661 ri D 0880) is presentedin Figure 1 The length of these trees withunit weight characters is 830 steps thus thesetrees are not identical to any of the most-parsimonious trees from the unit weightanalysis Branches not present in the strictconsensus tree from the unit weight analysisare indicated on the combinable consensustree from the successive weighting analysis
(Fig 1) The unit weight analysis of the com-bined set of the rbcL and ndhF data (153 char-acters total 043 of the cells were scored asmissing data) resulted in four trees 345 stepslong (ci D 0539 ri D 0482) The successiveweighting analysis gave three trees 978 stepslong (ci D 0840 ri D 0849) The strict consen-sus tree of the successive weighting analysisis presented in Figure 2 These trees are onestep longer with unit weights than the treesfrom the unit weight analysis Unless other-wise stated the following discussion refersto the successive weighting analyses of thecomplete rbcL data set
Investigating the justication of a prioriweighting (Sennblad and Bremer 2000) wemade some additional analyses of our rbcLdata set including tests for base compositionbias and rate heterogeneity This did not pro-duce signicantly different results and allwell-supported groups were congruent withthose of the present study The general struc-ture and the well-supported groups of theresulting trees (Figs 1 and 2) are to a largedegree congruent with the molecular stud-ies of Sennblad and Bremer (1996) whichare based on rbcL data of Civeyrel et al(1998) based mainly on matK sequence dataand of Potgieter and Albert (2001) basedmainly on trnLndashF data Although there aredifferences in the weakly supported rela-tionships among the well-supported groupsin particular the exact position of the
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 395
FIGURE 1 Combinable component consensus tree of the 252 most-parsimonious trees from the successiveweighting analysis of the complete rbcL data set Subfamilial and tribal classication is according to Endress andBruyns (2000) For subfamilies a three-letter code is used RAU APO PER SEC and ASC denote RauvoloideaeApocynoideae Periplocoideae Secamonoideae and Asclepiadoideae respectively OUT denotes the outgroupNumbers below branches are unit-weighted bootstrap values successive weighted bootstrap values are indicatedabove branches Branches not present in the strict consensus tree from the unit weight analysis are indicated witha cross (dagger) For branches marked A or B the frequencies in the combinable component consensus are 95 or 50respectively all other branches have 100 frequency
396 SYSTEMATIC BIOLOGY VOL 51
FIGURE 2 Strict consensus tree from the successive weighting analysis of the combined ndhF and rbcL dataset Tribal classication sensu Endress and Bruyns (2000) is indicated Numbers below branches are unit-weightedbootstrap values whereas successive weighted bootstrap values are indicated above branches Branches not presentin the strict consensus tree from the unit weight analysis are indicated with a cross (dagger)
Periplocoideae and of the root of the in-group all well-supported groups in our anal-ysis are congruent with those studies Thecombined analysis using data from rbcL-and matK-sequences and oral and pollenmorphology in the study of the Apocy-naceae sstr by Endress et al (1996) andthe analysis of Sennblad et al (1998) ofthe tribe Wrightieae sensu Leeuwenberg(1994a) and related taxa using rbcL andmorphological data are largely congruentwith the result from the present study Fur-thermore preliminary results from collab-orative studies by Sennblad et al (aimedat the traditional Apocynoideae and As-clepiadaceae pers comm) and Endresset al (aimed at the traditional Alyxieaepers comm) based on combined analy-sis of several molecular and morphologi-cal datasets are also congruent with ourresults
Support is good for the split betweenthe ingroup and the two outgroup taxaGelsemium and Mostuea (uwboot D 96swboot D 100) Thus the initial assump-tion of monophyly of the Apocynaceaesl is not violated Conforming to ear-lier studies (eg Sennblad and Bremer1996 Civeyrel et al 1998 Sennblad et al1998 Potgieter and Albert 2001) the tra-ditional Asclepiadaceae are nested withinthe traditional subfamily Apocynoideae(eg sensu Endress and Bruyns 2000)
The ApocynoideaendashAsclepiadaceae cladewhich is strongly supported (uwboot D 86swboot D 99) is in turn nested within thesubfamily Rauvoloideae sensu Endress andBruyns (2000)
Comparison of our results to the mostrecent classications of the ApocynaceaeEndress and Bruyns (2000 see Table 1 andFig 1) shows that although their circum-scription of the subfamilies PeriplocoideaeSecamonoideae and Asclepiadoideae aremonophyletic in our analysis their sub-families Rauvoloideae and Apocynoideaeare nonmonophyletic Similarly the tribesof the Rauvoloideae except Melodineaeare monophyletic but three of the tribesof the Apocynoideaemdashthe Apocyneae theEchiteae and the Wrightieaemdashare nonmono-phyletic Of the tribes of the Asclepiadoideaethe Ceropegieae and Asclepiadeae are mono-phyletic The position of Fockea as sisterto the rest of the Asclepiadoideae ren-ders the Marsdenieae nonmonophyletic Allthe included Periplocoideae taxa belongto the tribe Periploceae sensu Venter andVerhoeven (1997) except Pentopetia whichthey placed in the tribe CryptolepideaeGiven this unresolved position of Pentopetiathe monophyly of these tribes cannot be eval-uated however the reduction of Parquetinato synonymy with Periploca proposed byVenter and Verhoeven (1997) is not sup-ported by the present study
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 397
DISCUSSION
One of the most important tasks for anomenclature system is to provide a sta-ble means to communicate classications(which we here will understand as represent-ing phylogenetic relationships) Althoughboth the Linnaean and the Phylocode sys-tems are able to communicate phylogeniesthey are sensitive to changes in phylogenetichypotheses which in both systems will de-mand taxonomic changes In addition theLinnaean system is subject to a potentialinstability relating to subjective changes incircumscription that is lumpers and split-ters may refer to different groups of organ-isms by the same name (see de Queiroz andGauthier 1992) For many taxa a consensuscircumscription has eventually been estab-lished but for other groups this is still a prob-lem A current example of this from Apoc-ynaceae is the genus Tabernaemontana sensuLeeuwenberg (1991 1994b) which in theclassication of Allorge (1985) is split intoeight genera The Phylocode system was de-signed to avoid this problem by use of strictdenitions of taxa subjective changes incircumscription are avoided
However this rigidity in denitions leadsto a potential instability and high turnoverof names associated with a group of nestedtaxa Relatively small changes in phyloge-netic views eg when a dening taxonldquospecierrdquo receives a less nested positionmaycause some(or all) of these nested namesto become synonymous (eg de Queirozand Gauthier 1994 Bryant 1996 Liden andOxelman 1996Sereno 1999)The junior syn-onyms should according to the Phylocodebe rejected and new names may then beneeded for the new system of nested clades(de Queiroz and Gauthier 1994) The namesassociated with this group of taxa are thusreplaced by new names This can be espe-cially problematic when careless denitionshave been made as have been noted bySchander and Thollesson (1995) and Cantinoet al (1997) among others Consider a hy-pothetical worst-case example The Ascle-piadaceae were traditionally considered asseparate from Apocynaceae A stem-baseddenition of Apocynaceae reecting thisview could be the largest clade that in-cludes Apocynum but not Asclepias On thepresent tree such a denition would limitthe Apocynaceae to be synonymous with the
tribe Apocyneae and a new name wouldbe needed for the taxon corresponding tothe current view of Apocynaceae (eg sensuEndress and Bruyns 2000) A node-baseddenition would handle this particular prob-lem but may have other problems as illus-trated here with another hypothetical worst-case scenario Early classications of theApocynaceae included Plocosperma BenthA node-based denition representing thisview could have included as reference taxasay Apocynum and Plocosperma The recentlyproposed position of Plocosperma close tothe Boraginaceae (Backlund et al 2000)would with this denition have made Apoc-ynaceae a very large taxon (containing atleast both Gentianales and Boraginales) pos-sibly synonymous with (a hypothetical def-inition of) the informal taxon EuasteridsII (Bremer et al 1998) and again a newname for the Apocynaceae sensu Endressand Bruyns (2000 ie excluding Plocosperma)would have been needed Because theLinnaean system allows recircumscription ofsynonymous names the potential turnoverin the set of names in use is reduced withthis system (albeit at the expense of allowingsubjectivity) Additionally the ranks of theLinnaean system provide a set of names thatwork as relatively stable universal standardnames in communication note that equalranks does not guarantee comparable evo-lutionary units however Continuity in us-age of names is important for example injournals literature databases oras teach-ing and inventory work and the need forsuch continuity has been recognized by pro-ponents of both nomenclature systems (deQueiroz and Gauthier 1992 Greuter et al1994 Reveal 1996)
We propose a compromise approach thatuses solutions relating to these aspects fromboth systems concentrating mainly onbotanical classication of extant plant taxaat the tribal and familial levels and onlyconsidering monophyletic taxa We recom-mend a system that uses ldquostandard namesrdquo(ie universally used communication unitsat convenient hierarchical levels) such asthose provided by the principal ranks of theLinnaean system (Greuter et al 1994) but inwhich sensitivity toward subjective changesin circumscription is reduced One way to dothis is to adopt the denitions of the phy-logenetic system but use constraints similar
398 SYSTEMATIC BIOLOGY VOL 51
to those that apply between ranks in theLinnaean system To reduce sensitivity to-ward changes in phylogenetic hypotheses(Schander and Thollesson 1995 Bryant1996 1997 Cantino et al 1997 Moore 1998)we propose using combined node C stem-based denitions that will provoke explicitincompatibilities between taxa in case ofunfortunate changes in circumscription (deQueiroz 1996 has proposed a similar type ofdenition for designating nonmonophyletictaxa somewhat similar variants are alsodiscussed in the Phylocode) A node-basedpart of the denition provides a minimalcircumscription whereas a stem-based partprovides boundaries to competing taxa(maximal circumscription) If on a particu-lar phylogeny the minimal (monophyletic)circumscriptions of competing taxa are over-lapping then the two taxa are incompatibleTo make this meaningful at least two taxashould be included in the node-based partof the denition this will also avoid unnec-essary monotypic taxa This incompatibilitywill provide an objective point when nomen-claturalreconsideration isneeded Returningto the rst of our examples above node Cstem-based denitions of the Apocynaceaeand Asclepiadaceae could be as follows
The family Apocynaceae is the most inclusive clade inthe order Gentianales including the type specimensof Apocynum and Dictyophleba but not the taxa Ascle-piadaceae [ Gentianaceae Rubiaceae ]
and
The family Asclepiadaceae is the most inclusive cladein the order Gentianales including the type speci-mens of Asclepias and Periploca but not the taxa Apoc-ynaceae [ Gentianaceae Rubiaceae ]
On the tree in Figure 1 these two denitionsare mutually exclusive and a decision onwhich name to keep must be made Mini-mizing the number of classicatory changesshould take precedence in this decision Thereason for this is to avoid the large incon-venient ldquostandard namerdquo changes that couldresult from relatively small rearrangementsRejected names are then ignored when oc-curring in denitions of other taxa This isbecause the reference taxa in the stem-basedpart of the denitions are ldquodenedrdquo taxarather than physical types (thus reliance ondened reference taxa which has been crit-icized for stem-based denitions by Sereno1999 is what we aim for on the other hand
the explicit reference of type specimens in theminimum circumscription provided by thenode-based part of the denitions circum-vents the tautology problem raised byBryant 1996) In the current example thiswould mean that regardless of which nameof Apocynaceae and Asclepiadaceae is re-jected the other will be dened as corre-sponding to the Apocynaceae sensu Endressand Bruyns (2000) The choice will there-fore simply be which name to keep In thiscase because Apocynaceae has priority un-der the Linnaean system we would chooseto keep Apocynaceae to promote consistencywith the current system Nevertheless out-side these recommendations choices of whatnames to choose will probably to a certaindegree be arbitrary We also propose thatthe rejected name should be reinstated withan emended denition if appropriate to po-tentially allow for more stability in the setof names in use Thus in our second ex-ample above a node C stem-based deni-tion reecting the inclusion of Plocospermain Apocynaceae (eg by exchanging Dic-tyophleba for Plocosperma in the denitionabove) would presumably be incongruentwith the denition of Gentianales as well asthe denitions of Rubiaceae Gentianaceaeand so forth and therefore would be rejectedIn such acasewhere the only real change is inthe position of a single taxon a reinstatementof the Apocynaceae with an emended def-inition is motivated However a drawbackis that this may also introduce elements ofsubjectivity in the circumscriptions of taxa
Not all named clades need to be providedwith standard names In fact for the purposeof a universal set used in journals databasesand so forth it may sufce to use standardnames corresponding to the principal ranksof the Linnaean system (Greuter et al 1994)such as species genus family and orderNevertheless to provide compatibility withthe present system we suggest that addi-tional names corresponding to commonlyused secondary ranks (eg tribes) may berecognized The hierarchical level of a stan-dard name needs to be indicated This doesnot imply that taxa of the same hierarchi-cal level are comparable evolutionary unitsbut simply indicates the hierarchical levelof a taxon relative to nested taxa Becausesuch misinterpretations will in any case bedifcult to prohibit one may as well pro-vide compatibility with the present Linnaean
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 399
system by using the names of the Linnaeanranks (eg species genus tribe familyorder) to indicate hierarchical level The prin-ciple of exhaustive subsidiary taxa is notaccepted so redundant taxa need not be rec-ognized Other ldquononstandardrdquo taxa couldbe dened by using Phylocode denitionsmdashthat is without the restraints discussed forstandard taxa above
One advantage of the described systemis that utilizing the correspondence betweenthe standard names and Linnaean ranks al-lows the system to be largely compatiblewiththe establishedLinnaean system A change ofsystem could therefore be gradual with thetwo systems coexisting during a conversiontime (see eg de Quieroz 1997b)
Even though the node C stem-baseddenitions reduce the sensitivity towardschanges in circumscriptions they may still besensitive towardldquounfortunaterdquo or ldquobadrdquo def-initions (see eg Cantino et al 1997Cantinoand de Queiroz 2000) Many of the recom-mendations for such things as choices of ref-erence taxa (speciers) in phylogenetic def-initions are applicable also to the node Cstem-based denitions (eg Schander andThollesson 1995 Bryant 1996 Cantino et al1997 Sereno 1999) Most likely the compro-mise system described above will have sev-eral further problems and we hope that thispaper will invite further discussion on thesubject
A New Classication of the Apocynaceae slOur results indicate that some problem-
atic taxa remain in the classication ofEndress and Bruyns (2000) Monophyly fortwo of the subfamilies and ve of the tribesis called into question If other recent molec-ular studies (eg Potgieter and Albert 2001)are taken into consideration a further fourtribes (Alstonieae Alyxieae Plumerieae andVinceae) may be nonmonophyletic This canin some cases be simply analysis artifactsbut as Endress and Bruyns (2000) themselvespoint out their classication is to be consid-ered preliminary Insufcient taxonsamplingin molecular studies and difcult homologydecisions relating to morphological charac-ters make the circumscription of many ofthe tribes for example in the Apocynoideaesensu Endress and Bruyns (2000) uncertainThere is thus risk for future rearrangementsin many of the tribes A classication using
node C stem-based denitions may be ableto subsume such rearrangements withoutnumerous explicit recircumscriptions Wewill here use the result fromthe present studyas a basis to propose such a classication ofthe Apocynaceae sl
Our aim is to base taxa on clades that arewell supported A second aim is to makeour classication compatible with the cur-rent system We will therefore primarily rec-ognize nonoverlapping tribes that are validunder the Linnaean system We will furtheradopt the principle of nested referencing (assuggested by Lee 1999 and Sereno 1999)from the Linnaean system Thus the rst ref-erence taxon in the node-based part of thedenition is the primary type which corre-sponds to the Linnaean type used for thename of the taxon Instead of the (optional)enumeration of subsumed taxa of a lowerrank (eg genera or subtribes) we will in-clude node C stem-based denitions as dis-cussed above Citations to primary types ofreference taxa in node-based part of de-nitions are also nested and thus refer toldquothe type specimen of the type species ofthe type genus of the dened taxonrdquo Aproblem is that the relationships between thetribes areweakly supported This may lead tovery cumbersome denitions enumeratingall competing tribes in the stem-based partof the denition (eg Moore 1998) One wayto reduce this problem is to dene interme-diate well-supported taxa and use them ascompeting taxaWe therefore will further rec-ognize four nonstandard taxa (note that anassignment of ldquononstandardrdquo taxa does notimply less reliability standard names relateto communication purposes only) The taxadiscussed are indicated in Figure 3 We willdiscuss the tree from the top of Figure 3 start-ing with the taxa of the traditional Plumeri-oideae Tribes and subtribes given in paren-theses in the text refer to the classicationsof Endress and Bruyns (2000) in other casesour classication is implied For some tribeswe have chosen a name with priority underLinnaean system that implies inclusion of agenus not sampled in the present study Forall but one (Willughbeieae) of these casesadditional molecular or morphological phy-logenetic analyses support the inclusion ofthe genus Even if too uncertain denitionsshould be avoided this is less critical withnode C stem-based denitions In the caseof Willughbeieae we have decided from
400 SYSTEMATIC BIOLOGY VOL 51
FIGURE 3 The tribal reclassication discussed in the text indicated on the combinable component consensustree from the successive weighting analysis of the complete rbcL data set Four nonstandard taxamdashApocynoidinaEuapocynoidina Asclepiadacina and Asclepiadoidinamdashare also indicated
discussions with a morphological expert(M Endress pers comm) that morphologi-cal support exists for including WillughbeiaRoxb in Willughbeieae If this is correctthe denition given below will be valid if
it is not our system allows redenition ofthe taxon We have also tried to providecomments on potential morphological char-acters taken from literature for the differ-ent taxa recognized However we have not
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 401
performed any morphological analysis butrely on information external to this studyfor these comments In most cases we cantherefore not identify synapomorphies forthe taxa
TAXONOMIC TREATMENT
Tribus Willughbeieae ADCThe Ancylobotrys Dictyophleba and Va-
hadenia clade is strongly supported (uwbootand swboot D 100) and corresponds to theWillughbeieae of Endress and Bruyns (2000)The present sampling contains closely re-lated genera that have been traditionallykept together Landolphia P Beauv sl in-cludes all these taxa (eg Schumann 1895)In addition to the taxa sampled here thestudy of Potgieter and Albert (2001) sup-ports the inclusion of Couma Aubl LacmelleaH Karst Pacouria Aubl and Saba (Pichon)Pichon in this clade Plants in this cladeall have a gynoecium that is congenitallysyncarpous and indehiscent and that con-tains a eshy pulp including numerous seedswith copious horny endosperm (Fallen 1986Persoon et al 1992) This probably con-stitutes synapomorphies for this clade(M Endress pers comm) although not un-ambiguously for example syncarpy is par-alleled in Carisseae We will therefore as-sume that the reference genus Willughbeia isincluded in this tribe
DenitionmdashTribe Willughbeieae is themost inclusive clade that includes the prim-ary types of Willughbeia and Dictyophleba butnot Alyxieae Aspidospermeae CarisseaeHunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Apocynoidina
Tribus Vinceae Bartl
Ochrosia Rauvola Vinca and Catharanthus(all Vinceae) form a strongly supported clade(uwboot D 89 swboot D 99) The associa-tion between Rauvola and Catharanthus wasindicated by Sennblad and Bremer (1996)The study of Potgieter and Albert (2001) sup-ports inclusion of the genera LaxoplumeriaMarkgr Neisosperma Raf and Tonduzia Pit-tier previously classied with AspidospermaThe Vinceae contain both herbaceous andfruticose taxa The fruits are apocarpousbut whereas the fruits of Catharanthus Lax-oplumeria Neisosperma Tonduzia and Vincahave dry fruit walls the fruits of Ochrosia andRauvola are drupes Plants in Vinceae share
a similar type of style head (ie the apicalenlarged part of the style) with a stigmatichollow and upper and lower hair wreaths(except in some species of Ochrosia) the pres-ence of a nectar disk and a linear hilum onthe seed (Pichon 1948b) On the present treeVinceae also include Kopsia however this po-sition of Kopsia is very weakly supported (seealso Sennblad and Bremer 1996)
DenitionmdashTribe Vinceae is the mostinclusive clade that includes the primarytypes of Vinca and Rauvola but not AlyxieaeAspidospermeae Carisseae HunterieaeMelodineae Plumerieae Tabernaemon-taneae Willughbeieae or Apocynoidina
Tribus Tabernaemontaneae G DonThe strongly supported association be-
tween Tabernaemontana Tabernanthe Carval-hoa Schizozygia and Molongum (Tabernae-montaneae uwboot D 84 swboot D 98)has been suggested previously (Boiteau et al1978 Fallen 1986 Endress et al 1996Sennblad and Bremer 1996) Inclusion of taxatraditionally associated with Tabernaemon-tana (eg Voacanga Thouars and CallichiliaStapf see also Tabernaemontaneae sensuLeeuwenberg 1994a) and Macoubea Aublis supported by the analysis of Potgieterand Albert (2001) The taxa share sclerenchy-matic anthers free from the style head anda characteristic band of heavily cutinizedcells at the insertion on the staminal rib(Endress et al 1996) The taxa tradition-ally associated with Tabernaemontana arecharacterized by their apocarpous fruitswith arillate seeds whereas genera asso-ciated with Molongum (Ambelanieae sensuLeeuwenberg 1994a) have syncarpous fruitsand lackanarillus HoweverMacoubea formsa link between the two having a syncar-pous fruit with arillate seeds (Zarucchi et al1995)
DenitionmdashTribe Tabernaemontaneae isthe most inclusive clade that includes the pri-mary types of Tabernaemontana and Schizozy-gia but not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae PlumerieaeVinceae Willughbeieae or Apocynoidina
Tribus Aspidospermeae Miers
Aspidosperma and Vallesia (Aspidosper-meae) form a strongly supported clade (uw-boot and swboot D 100) rst suggestedin Sennblad (1997 see also Sennblad and
402 SYSTEMATIC BIOLOGY VOL 51
Bremer 2000) Subsequently GeissospermumAllemao Haplophyton ADC MicroplumeriaBaill and Strempeliopsis Benth were indi-cated to belong to this clade (Potgieter andAlbert 2001) The Aspidospermeae havesimple style heads lacking both stigmatichollow and upper hair wreath and haveapocarpous fruits with variation in eshi-ness and sclerication Alstonia has tradi-tionally been placed in the Plumerieae Be-cause its seeds have a hairy margin it hasbeen suggested to form a link to subfam-ily Apocynoideae This was contradicted bySennblad and Bremerrsquos study (1996) how-ever where Alstonia was placed in an iso-lated position as the sister group to therest of the Apocynaceae sl Here it groupswith Aspidospermeae Sensu Endress andBruyns (2000) However this association isvery weakly supported (uwboot and swbootlt 50) and in other recent studies As-pidospermeae and Alstonia do not form aclade (Potgieter and Albert 2001) Becauseof this we do not use it as a reference taxonin the denition Thus although the Aspi-dospermeae include Alstonia on the presenttree some other position of Alstonia canbe accommodated without amending thedenition
DenitionmdashTribe Aspidospermeae is themost inclusive clade that includes theprimary types of Aspidosperma and Vallesiabut not Alyxieae Carisseae HunterieaeMelodineae Plumerieae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Melodineae G DonThe well-supported association between
Melodinus and Craspidospermum (uwboot D70 swboot D 75) was rst reportedby Sennblad (1997 see also Sennblad andBremer 2000) and has subsequently beenreected in the tribe Melodineae sensuEndress and Bruyns (2000 their inclusion ofDiplorhynchus however is not supported)Craspidospermum although its fruit is dryand dehiscent as opposed to the indehis-cent fruits of Melodinus was noted by Pichon(1948b) to present several characters of theCarisseae such as stamens inserted near thebase of the corolla tube and a syncarpousovary In particular Pichon found the verydense suprastaminal indumentum type rem-iniscent of Melodinus From his descriptions
of the two genera (Pichon 1948ab) the fol-lowing similarities also emerge presence ofa stipular line pollen in tetrads and a punc-tiform hilum
DenitionmdashTribe Melodineae is the mostinclusive clade that includes the pri-mary types of Melodinus and Craspidosper-mum but not Alyxieae AspidospermeaeCarisseae Hunterieae Plumerieae Taber-naemontaneae Vinceae Willughbeieae orApocynoidina
Tribus Hunterieae K Schum
The Picralima and Pleiocarpa clade isstrongly supported (uwboot and swboot D100) and corresponds to the subtribeHunterieae sensu Endress andBruyns (2000)The study of Potgieter and Albert (2001)also supports the inclusion of Hunteria RoxbThese taxa are characterized by apocar-pous sometimes pluricarpous ovaries andeshy brous fruit walls (Omino 1996)Diplorhynchus (Melodineae) is here weaklyassociated (uwboot and swboot lt 50) withthe Pleiocarpeae It shares a few similari-ties such as stipular lines and a style headwithout a stigmatic hollow and hair wreathsbut also has differences such as a dry de-hiscent fruit With the present denitionDiplorhynchus will tentatively be included inthe Hunterieae
DenitionmdashTribe Hunterieae is the mostinclusive clade that includes the primarytypes of Hunteria and Picralima but not Alyx-ieae Aspidospermeae Carisseae Melod-ineae Plumerieae TabernaemontaneaeVinceae Willughbeieae or Apocynoidina
Tribus Plumerieae EndlThe strongly supported relationship
(uwboot and swboot D 100) between Alla-manda and Plumeria corresponds to earlierstudies (Endress et al 1996 Sennblad andBremer 1996 Civeyrel et al 1998 Potgieterand Albert 2001) In all earlier treatmentsAllamanda has had uncertain relationshipsThe association with Plumeria nds supportin pollen morphology such as perforatemesocolpial depressions and similar innerexine pattern The two genera also containsecoiridoids rather than the indole alka-loids and cardenolides that are commonin the traditional Plumerioideae (Endresset al 1996) Cameraria Cerbera and Thevetiaform a strongly supported association
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 403
(uwboot D 83 swboot D 91) with Ane-chites which has been suggested by Fallen(1983) This association was based mainlyon the latrorsesublatrorse anthers andthe broad style head with large apicalappendages and stigmatic hollow The studyby Potgieter and Albert (2001) suggesteda further inclusion of Cerberiopsis Viell ExPancher amp Sebert and Skytanthus Meyen inthis tribe These taxa form a well-supportedclade with Plumeria and Allamanda whichcorresponds to the Plumerieae sensuEndress and Bruyns (2000) Characterssupporting this relationship are the presenceof infrastaminal as well as suprastaminalappendages (not present in Anechites andPlumeria) and winged seeds (not present inAnechites) (Endress et al 1996)
DenitionmdashTribe Plumerieae is themost inclusive clade that includes theprimary types of Plumeria and Allamandabut not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Carisseae Dumort
Acokanthera and Carissa (both Carisseae)are traditionally joined and are sometimeseven treated as one genus (eg Pichon1948a) In the present study they form astrongly supported clade (uwboot D 96swboot D 99) Among the characters sup-porting this tribe are syncarpous fruits with-out axile placentation and endocracks on theinner exine of the pollen (Endress et al 1996)
DenitionmdashTribe Carisseae is the most in-clusive clade that includes the primary typesof Carissa and Acokanthera but not AlyxieaeAspidospermeae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
Tribus Alyxieae G Don
Chilocarpus has been a genus with un-certain afnities In Endress et al (1996) itwas the sister taxon to the ApocynoideaeHere it is associated with Alyxia and Lep-inia in a clade corresponding to the Alyxieaesensu Endress and Bruyns (2000) Thisclade is strongly supported with succes-sive weighted bootstrap (swboot D 91) butnot with unit-weighted bootstrap support(uwboot D 58) such clades will henceforthbe termed moderatelywell supported clades
A further inclusion of Condylocarpon DesfLepiniopsis Valeton and Plectaneia Thouarswas indicated by Potgieter and Albert (2001)The constituent taxa are the only Apocy-naceae taxa outside the Apocynoidina thathave porate pollen
DenitionmdashTribe Alyxieae is the most in-clusive clade that includes the primary typesof Alyxia and Lepinia but not Aspidosper-meae Carisseae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
APOCYNOIDINA
On the present tree the traditionalsubfam-ily Apocynoideae includes all subfamilies(Periplocoideae Secamonoideae and Ascle-piadoideae) of the Asclepiadaceae sensuLiede and Albers (1994) All of these taxaare established names and selecting one asa subfamily at the expense of the othersmight be considered unfortunate from dif-ferent points of view We have here chosennot to recognize subfamilies instead we willdene the traditional Apocynoideae Ascle-piadaceae and Asclepiadoideae as nonstan-dard taxa The correspondence to the tradi-tional taxa will be indicated in the names byreplacing the sufx -eae with a neutral suf-x -ina (Kron 1997) We will also for practi-cal reasonsrecognize as a nonstandardtaxonthe informal group euapocynoids suggestedby Sennblad et al (1998)
The taxa of the traditional Apocynoideaeand Asclepiadaceae (uwboot D 86 swboot D99) clade will in many cases be more dif-cult to safely delimit to tribes becausemany of the groups in this clade have weakor no support (although preliminary resultsfrom an unpublished collaborative study bySennblad et al indicate additional supportfor the tribes discussed below) The tribaland subtribal classication of the traditionalApocynoideae has been shown to be prob-lematic (Leeuwenberg 1994a Endress andBruyns 2000) The circumscription of tribesfor these taxa might therefore in some casesbe preliminary
In all following taxa the basal part of theanther connective called the retinacle is ad-nate to the style head
DenitionmdashApocynoidina are the mostinclusive clade that includes the pri-mary types of Apocynum and Wrightia butnot Alyxieae Aspidospermeae Carisseae
404 SYSTEMATIC BIOLOGY VOL 51
Hunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Willughbeieae
Tribus Wrightieae G DonThe Stephanostema and Wrightia (both
Wrightieae) clade is strongly supported(uwboot and swboot D 100) and correspondsto the Wrightieae sensu Sennblad et al (1998)who also suggested an inclusion of Pleio-ceras Baill The tribe could be characterizedby a combination of synapomorphic and ple-siomorphic characters for example presenceofa chalazalandabsence of micropylar comaleft contorted aestivation and absence of airspaces in the anthers (1998)
DenitionmdashTribe Wrightieae is the mostinclusive clade in the Apocynoidina that in-cludes the primary types of Wrightia andStephanostema but not Malouetieae Nerieaeor Euapocynoidina
Tribus Nerieae (Benth) M Pichon
Adenium Nerium and Strophanthus (allWrightieae) constituted the Nerieae inSennblad et al (1998) however the cladereceived weak support Here Adenium andNerium form a clade that is absent in halfof the most-parsimonious trees in the com-plete rbcL analysis but is strongly supportedin the combined ndhF and rbcL analysis(uwboot D 98 swboot D 99) Strophan-thus groups with the Mascarenhasia Pachy-podium Kibatalia Funtumia and Holarrhenaclade which corresponds to the Malouetieaesensu Sennblad et al (1998) This position ofStrophanthus is however weakly supported(uwboot and swboot lt 50) and is not presentin the combined ndhF and rbcL tree Fur-thermore certain morphological characters(unfused slits in the corolla tube and anapical anther appendage) indicate a posi-tion closer to the Nerieae (Sennblad et al1998)
DenitionmdashTribe Nerieae is the most in-clusive clade in the Apocynoidina that in-cludes the primary types of Nerium and Ade-nium but not Malouetieae Wrightieae orEuapocynoidina
Tribus Malouetieae Mull-Arg
The association of Mascarenhasia Pachy-podium Kibatalia Funtumia and Holar-rhena (all Wrightieae) is moderately wellsupported in the complete rbcL analysis
(uwboot D 57 swboot D 74) whereasthe corresponding clade is well supportedin the combined rbcL and ndhF analysis(uwboot D 66 swboot D 100) On the ba-sis of identied synapomorphies (presenceof calcium oxalate crystals in the stomium ofthe anthers absence of interpetal vein ses-sile laments and absence of air spaces inthe anthers) Sennblad et al (1998) predictedan inclusion of Alaa Thouars Kibatalia andMalouetia ADC in the tribe The inclusion ofKibatalia is strongly supported in this studyFollowing these suggestions we will here as-sume the inclusion of the reference genusMalouetia in this tribe (The denition belowwill include Strophanthus on the present treebut see comments under Nerieae)
DenitionmdashTribe Malouetieae is the mostinclusive clade in the Apocynoidina that in-cluded the primary types of Malouetia andFuntumia but not Nerieae Wrightieae orEuapocynoidina
EUAPOCYNOIDINA
In all the remaining representatives ofthe Apocynoideae sensu Endress andBruyns (2000) that is Mandevilla MesechitesAganosma Apocynum TrachelospermumRhabdadenia Beaumontia Parsonsia andPrestonia the anthers are adnate to the stylehead both by the retinacle and by the thecaeThese taxa together with the taxa of the tra-ditional Asclepiadaceae form a moderatelysupported clade (uwboot lt 50 swboot D86) that corresponds to the informal taxoneuapocynoids sensu Sennblad et al (1998)
DenitionmdashEuapocynoidina are the mostinclusive clade that includes the primarytypes of Apocynum and Echites but notMalouetieae Wrightieae or Nerieae
Tribus Mesechiteae Miers
Mandevilla and Mesechites (both Mese-chiteae) form a moderately supported clade(uwboot lt 50 swboot D 87) equivalentto the Mesechiteae This tribe could be char-acterized by anthers with obtuse truncatedtails and a retinacle lacking hairs and also astrongly pentagonal style head with a stig-matic hollow
DenitionmdashTribe Mesechiteae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Mesechitesand Mandevilla but not Apocyneae EchiteaePeriploceae or Asclepiadacina
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 405
Tribus Apocyneae Bercht et J Presl
Aganosma Apocynum Beaumontia andTrachelospermum (Apocyneae) form a weaklysupported clade with Rhabdadenia (Echiteae)(uwboot and swboot lt 50) in the completerbcL analysis whereas the correspondingclade in the combined ndhF and rbcL analy-sis gains a little better support (uwboot lt 50swboot D 91) Likewise this clade (with theinclusion of Chonemorpha G Don) is weaklysupported in Potgieter and Albert (2001) Po-tential characterizing traits could be a stylehead without stigmatic hollow in combina-tion with thecae that are adnate to the stylehead (Sennblad et al 1998)
DenitionmdashTribe Apocyneae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Apocynumand Trachelospermum but not Echiteae Mese-chiteae Periploceae or Asclepiadacina
Tribus Echiteae Bartl
Parsonsia Peltastes and Prestonia (allEchiteae) form a moderately supported orunsupported clade with the Periplocoideaerepresentatives of the study (uwboot andswboot lt 50 and uwboot lt 50 swboot D89 in the complete rbcL and the combinedrbcL and ndhF analysis respectively) In thestudy of Sennblad et al (1998) Prestonia andParsonsia formed a weakly supported cladeand in the study of Potgieter and Albert(2001) inclusion of both Parsonsia and EchitesP Browne in the Echiteae is supportedParsonsia and Echiteae also share certaincharacters such as a horseshoe-shaped reti-nacle and a style head with stigmatic hollow(see Sennblad et al 1998)
DenitionmdashTribe Echiteae is the most in-clusive clade in the Euapocynoidina thatincludes the primary types of Echites andPrestonia but not Apocyneae MesechiteaePeriploceae or Asclepiadacina
Tribus Periploceae BartlThe representatives of the Periplo-
coideae sensu Endress and Bruyns (2000)mdashPentopetia Parquetina Petopentia PeriplocaTacazzea and Mondiamdashgroup together ina well-supported clade (uwboot D 67swboot D 93) The monophyly of the tra-ditional periplocoid taxa is also supportedin earlier studies that included a wider ordifferent sampling (Civeyrel et al 1998
Potgieter and Albert 2001) The taxa ofthe traditional Asclepiadaceae have ldquotrans-latorsrdquo structures related to pollinationspecialization In the Periploceae the pollenis deposited as tetrads in a spoon-liketranslator with an adhesive disc
DenitionmdashTribe Periploceae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Periplocaand Pentopetia but not Apocyneae EchiteaeMesechiteae or Asclepiadacina
ASCLEPIADACINA
The taxa of the Asclepiadoideae sensuLiede and Albers (1994) form a clade togetherwith the two Secamone species and Baissea(Apocyneae uwboot lt 50 swboot D 73)The exclusion of Periploceae (traditionallyassociated with this clade see Civeyrel et al1998) from this taxon is weakly supportedTherefore Periploceae is not included as areference taxon in the stem-based part ofthe denition of the Asclepiadacina thusleaving open the possibility for a later in-clusion of this taxon In the Asclepiadacina(except Baissea) the pollen is agglutinatedinto pollinia that are connected to a claspingtranslator
DenitionmdashAsclepiadacina are the mostinclusive monophyletic clade in the Euapoc-ynoidina that includes the primary types ofAsclepias and Secamoneae but not ApocyneaeEchiteae or Mesechiteae
Tribus Secamoneae G DonThe tribe Secamoneae sensu Bruyns
(uwboot D 89 swboot D 96) is mono-phyletic in the present analysisHowever be-cause representatives from only one of theconstituent genera are included monophylyis not well tested but the study of Civeyrelet al (1998) which included a larger numberof taxa including Pervillea Decne showedthat monophyly of the tribe was supportedThe traditional character for this tribe is fourpollinia per translator
DenitionmdashTribe Secamoneae is the mostinclusive clade in the Asclepiadacina that in-cludes the primary types of Secamone andPervillea but not Asclepioidina or Baissea
BaisseaThe position of Baissea corresponds with
that in the study of Sennblad et al (1998)
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 391
2000) were summarized and developed fur-ther by use of morphology in the most re-cent classication of the Apocynaceae sl byEndress and Bruyns (2000) The ve tradi-tional subfamilies were kept but 19 newlyrecircumscribed tribes were also recognized(see Table 1)
However the results of the present studyindicate that although the classication ofEndress and Bruyns (2000) are a major ad-vance in making the classication congru-ent with current best estimates of phylogenyproblematic cases still need revision Fur-thermore for many of the recognized tribestaxon sampling for molecular data is weakand the morphological homologies in thesegroups are difcult to interpret (Endress andBruyns 2000) Further rearrangements oftribal circumscription will almost certainlybe needed in the future
The Apocynaceae sl may thus provide asuitable test for the classication system pro-posed in this study We therefore propose anew classication of the Apocynaceae basedmainly on an rbcL analysis that includesextended taxon sampling as well as addi-tional data from length variation of rbcL in its30-end and downstream Certain relation-ships of special interest have been tested witha smaller data set for the subfamily Apoc-ynoideae and the traditionalAsclepiadaceaeincluding additionalnucleotide sequences ofthe chloroplast gene ndhF This gene oftenhas a greater substitution rate than rbcL (seeOlmstead and Sweere 1994 Kim and Jansen1995) and may therefore form a complementto rbcL when the latter proves too conservedWe have used only the 3rsquo region of the genebecause that region appears to show the mostvariability (see Kim and Jansen 1995) Fur-thermore results from other recent molecu-lar studies have also been taken into accountin the classication
MATERIALS AND METHODS
We sampled 77 representatives of theApocynaceae sl (Table 1) All subfamiliesand tribes of the classication of Endress andBruyns (2000) are represented in the analy-sis Gelsemium Juss and Mostuea Didr of theGelsemiaceae were chosen as outgroup taxabecause they were indicated to be the closest
TABLE 1 Classication of the Apocynaceae sl andsampling of taxa Classication are according to Endressand Bruyns (2000)
RAUVOLFIOIDEAEAlstonieae Melodineae
Alstonia RBr Craspidospermum ADCAspidosperma Mart Diplorhynchus Ficalho
et Zucc et HiernVallesia Ruiz Melodinus JR Forst
et Pav et G ForstVinceae Hunterieae
Catharanthus G Don Picralima PierreKopsia Blume Pleiocarpa BenthOchrosia Juss PlumerieaeRauvola L Allamanda LVinca L Anechites Griseb
Willughbeieae Cameraria LAncylobotrys Pierre Cerbera LDictyophleba Pierre Plumeria LVahadenia Stapf Thevetia Adans
Tabernaemontaneae CarisseaeCarvalhoa K Schum Acokanthera G DonMolongum Pichon Carissa LSchizozygia Baill AlyxieaeTabernaemontana L Alyxia RBrTabernanthe Baill Chilocarpus Blume
Lepinia DecneAPOCYNOIDEAE
Wrightieae ApocyneaeAdenium Roem Aganosma (Blume)
et Schult G DonNerium L Apocynum LStephanostema K Schum Baissea ADCStrophanthus DC Beaumontia WallWrightia RBr Trachelospermum Lem
Malouetieae MesechiteaeFuntumia Stapf Mandevilla LindlHolarrhena RBr Mesechites Mull-ArgKibatalia GDon EchiteaeMascarenhasia ADC Parsonsia RBrPachypodium Lindl Peltastes Woodson
Prestonia RBrRhabdadenia Mull-Arg
PERIPLOCOIDEAEMondia Skeels Periploca LParquetina Baill Petopentia BullockPentopetia Decne Tacazzea Decne
SECAMONOIDEAESecamone RBr
ASCLEPIADOIDEAEMarsdenieae Asclepiadeae
Fockea Endl Araujia BrotHoya RBr Asclepias LMicholitzia NEBr Calotropis RBrStephanotis Thouars Orthosia Decne
Ceropegieae Fischeria DCCeropegia L Matelea AublStapelia L Schizostephanus Benth
et Hook fTweedia Hook et ArnTylophora RBrVincetoxicum Wolf
392 SYSTEMATIC BIOLOGY VOL 51
sister group to the Apocynaceae sl in therecent analysis of Gentianales by Backlundet al (2000)
Forty-one new sequences of rbcL are pub-lished in this study rbcL sequences forthe outgroup taxa and for 35 ingroup taxawere published previously (Table 2) ForndhF the 30 region of the gene was se-quenced for 18 taxa representing the tra-ditional Apocynoideae and AsclepiadaceaeTotal DNA was extracted from fresh leavesor herbarium material by using the meth-ods of Saghai-Maroof et al (1984) and Doyleand Doyle (1987) The extractions were pu-ried by ultracentrifugation in CsCl gra-dients or ethanol precipitation Additionalpurication with the Qiaquick PCR puri-cation kit (Qiagen Inc) was performed incases with problematic polymerase chain re-action (PCR) amplication Double-strandedDNA was amplied by PCR by using Taq-polymerase kit (Promega Corp) SyntheticPCR primer sequences for rbcL were takenfrom Olmstead et al (1992) ndhF primerswere taken from Olmstead and Sweere (1994)and Oxelman et al (1999) In cases wherePCR amplication proved difcult a PCRreaction using Taq extender PCR additive(Stratagene Inc) was performed followingthe protocol provided by the supplier A sec-ond round of PCR with only one of theprimers was performed to obtain single-stranded DNA (Kaltenboeck et al 1992)Single-stranded DNA was sequenced man-ually (Sanger et al 1977) by using internalprimers designed for rbcL by G Zurawski(DNAX Research Institute) and for ndhF byOlmstead and Sweere (1994) and Oxelmanet al (1999)
The length of the sequences of rbcL ex-cluding the 26 rst nucleotides but includ-ing positions just downstream from the geneobtained in this study varied between 1411and 1509 nt For all taxa except Parsonsia thesequence includes the stop codon In posi-tions 27ndash1425 no structural mutations occurin rbcL Alignment of this part of the geneis thus very simple However after position1425 that is just before the typical positionof the stop codon gaps of different sizes oc-cur complicating alignment Similarly gapsoccur in the coding region of the ndhF se-quences Because of these alignment prob-lems the sequences were truncated 23 basesbefore the typical position of the stop codonin Apocynaceae sl The reading frame was
inferred through comparisonswith completeApocynaceae ndhF-sequences from the studyof Backlund et al (2000)
Multiple sequence alignment was made byhand to reduce the number of gaps while in-creasing the percentage similarity We usedthe criteria of Golenberg et al (1993) with mi-nor modications
Two cladistic analyses were performedThe data matrix for the rst analysis coveredall included taxa and comprised characterscorresponding to the nucleotide positions27ndash1425 of the rbcL gene and characterscorresponding to gaps and informative nu-cleotide positions downstream of position1425 EMBL database accession numbersfor rbcL sequences are given in Table 2Only parsimony informative characters wereanalyzed
The second analysis was of a subset ofthe taxa included in the rst analysis andincludes 18 genera representing the Apoc-ynoideae sensu Leeuwenberg (1994a) andthe traditional Asclepiadaceae The data ma-trix consisted of two submatrices One sub-matrix comprised positions 27ndash1425 of therbcL gene The second submatrix comprised733 aligned nucleotide positions from the3rsquo region of ndhF EMBL database accessionnumbers for the ndhF sequences are given inTable 2 Only parsimony informative charac-ters were analyzed
The cladistic analyses were performed us-ing PAUP 40b2 (Swofford 1998) An initialheuristic search (PAUP settings HSEARCH[ADDSEQ D RANDOM NREPS D 200SWAP D TBR] other options withdefault set-tings) with all characters given a unit weightwas followed by a successiveweighting anal-ysis (Farris 1969 1989) also using heuris-tic searches (PAUP settings HSEARCH[ADDSEQ D RANDOM NREPS D 10SWAP D TBR] other options withdefault set-tings) and characters reweighted accordingto their rescaled consistency index
Bootstrap frequencies (Felsenstein 1985)were calculated with 10000 replicates on theunit-weighted (uwboot) and successiveweighted (swboot) data matrices (PAUPsettings BOOTSTRAP [NREPS D 10000METHOD D HEURISTIC CONLEVEL D50] [ADDSEQ D RANDOM NREPS D 1SWAP D SPR NOMULPARS] other optionswith default settings)
To simplify the discussion of the resultswe used a rough scale of the relative support
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 393
TABLE 2 Vouchers andEMBLaccession numbers for sequences published in this study Forpreviously publishedsequences and for sequences from previously published voucher specimens the reference is given
Taxon Vouchersourceb EMBL number rbcLndhF
Acokanthera oblongifolia (Hochst) Codd Sennblad and Bremer 1996 X91758-Adenium obesum (Forssk) Roem et Schult Sennblad et al 1998 AJ002880AJ420130Aganosma marginata (Roxb) G Don Ryding 490 UPS AJ419730-Allamanda cathartica L Sennblad and Bremer 1996 X91759-Alstonia scholaris RBr Sennblad and Bremer 1996 X91760-Alyxia ruscifolia RBr Sennblad 236 UPS AJ419731-Ancylobotrys petersiana Pierre Sennblad 201 UPS AJ419732-Anechites nerium Urb Bremer et al 3386 UPS AJ419733-Apocynum cannabinum L Sennblad and Bremer 1996 X91761AJ420113Araujia hortorum E Fourn Bremer 3006 UPS AJ419734-Asclepias curassavica L Sennblad and Bremer 1996 X91774-Aspidosperma triternatum Rojas Acosta Bremer 3029 UPS AJ419735-Baissea leonensis Benth Sennblad et al 1998 Yallah 112 UPSa AJ002881AJ420114Beaumontia grandiora Wall Sennblad et al 1998 AJ002882AJ420115Calotropis procera (Aiton) WT Aiton Sennblad 222 UPS AJ419736-Cameraria latifolia L Houghton 1085 FTG AJ419737-Carissa bispinosa (L) Merxm Sennblad 235 UPS AJ419738-Carvalhoa campanulata K Schum Sennblad 217 UPS AJ419739-Catharanthus roseus G Don Sennblad and Bremer 1996 X91757-Cerbera venenifera Steud No voucher AJ419740-Ceropegia woodii Schltr Sennblad and Bremer 1996 X91775-Chilocarpus suaveolens Blume Endress et al 1996 X92445-Fockea multiora K Schum Specks 248 cult MSTRc AJ419741-Craspidospermum verticillatum Bojer Pettersson and Nilsson 742 AJ419742-Dictyophleba lucida Pierre Sennblad and Bremer 1996 X91762-Diplorhynchus condylocarpon (Mull-Arg) Pichon Sennblad 203 UPS AJ419743-Fischeria stellata (Vell) E Fourn Clark and Watt 793 UPS AJ419744-Funtumia elastica Stapf Sennblad et al 1998 AJ002884-Gelsemium sempervirens Aiton Olmstead et al 1993 L14397-Holarrhena pubescens G Don Sennblad et al 1998 AJ002884-Hoya bella Hook Sennblad and Bremer 1996 X91776-Kibatalia gitingense (Elmer) Woodson Liede 3268 Z AJ419745-Kopsia fruticosa ADC Sennblad and Bremer 1996 X91763-Lepinia taitensis Decne Perlman 15071 NTBG AJ419746-Mandevilla sanderi (Hemsl) Woodson Sennblad and Bremer 1996 X91764AJ420116Mascarenhasia arborescens ADC Sennblad et al 1998 AJ002885AJ420117Matelea hirsuta (Vahl) Woodson Sennblad 263 UPS AJ419747-Melodinus monogynus Roxb Sennblad 261 UPS AJ419748-Mesechites trida Mull-Arg Bremer et al 3351 UPS AJ419749-Micholitzia obcordata NEBr Bremer 3010 UPS AJ419750-Molongum laxum (Benth) Pichon Sennblad and Bremer 1996 X91765-Mondia ecornuta (NEBr) Bullock Sennblad 215 UPS AJ419751-Mostuea brunonis Didr Olmstead et al 1993 L14404-Nerium oleander L Sennblad 265 UPS AJ002886AJ420118Ochrosia coccinea Miq v d Laan 389 WAG AJ419752-Orthosia serpyllifolia Decne Bremer et al 3372 UPS AJ419753-Pachypodium lamerei Drake Sennblad et al 1998 AJ002887AJ420119Parquetina nigrescens (Afzel) Bullock Sennblad and Bremer 1996 X91777-Parsonsia heterophylla A Cunn Sennblad et al 1998 AJ002888-Peltastes peltatum (Vell) Woodson Sennblad 262 UPS AJ419754-Pentopetia sp No voucher AJ419755-Periploca graeca L Sennblad et al 1998 AJ002889AJ420120Petopentia natalensis (Schltr) Bullock Sennblad 237a and sn cult MSTR AJ419756AJ420121Picralima nitida T et H Dur Sennblad and Bremer 1996 X91766-Pleiocarpa mutica Benth Bremer 3017 UPS AJ419757-Plumeria inodora Jacq Sennblad and Bremer 1996 X91767-Prestonia quinquangularis Spreng Sennblad and Bremer 1996 X91768AJ420122Rauvola mannii Stapf Sennblad and Bremer 1996 X91769-Rhabdadenia biora Mull-Arg Zona 616 FTG AJ419759AJ420123Schizostephanus alatus K Schum Cult MSTR AJ419758-Schizozygia coffaeoides Baill Sennblad 207 UPS AJ419760-Secamone afzelii (Schult) K Schum Sennblad and Bremer 1996 X91779AJ420124
394 SYSTEMATIC BIOLOGY VOL 51
TABLE 2 Continued
Taxon Vouchersource EMBL number rbcLndhF
Secamone geayi Constatin et Gallaud Civeyrel 1200 LC AJ419761-Stapelia leendertziae NEBr Sennblad and Bremer 1996 X91778-Stephanotis oribunda Brongn Sennblad 256 UPS AJ419762AJ420125Stephanostema stenocarpum K Schum Sennblad and Bremer 1996 X91770AJ420126Strophanthus eminii Pax Sennblad and Bremer 1996 X91771AJ420127Tabernaemontana divaricata Roem et Schult Sennblad and Bremer 1996 X91772-Tabernanthe iboga Baill Leeuwenberg 12544 WAG AJ419763-Tacazzea apiculata Oliver Venter 9188 cult MSTR AJ419764-Thevetia peruviana (Pers) K Schum Sennblad and Bremer 1996 X91773-Trachelospermum jasminoides (Lindl) Lem Sennblad et al 1998 AJ002890AJ420128Tweedia coerulea Sweet Sennblad 254 UPS AJ419765-Tylophora sylvatica Decne Sennblad and Bremer 1996 X91789-Vahadenia caillei (AChev) Hutch et Dalziel Leeuwenberg 12275 WAG AJ419766-Vallesia antillana Woodson Meagher 966 FTG AJ419767-Vinca minor L Sennblad 230 UPS AJ419768-Vincetoxicum hirundinaria Medik Sennblad 257 UPS AJ419769-Wrightia arborea (Dennst) Mabb Sennblad et al 1998 AJ002891AJ420129
aVoucher for rbcL sequence onlybHerbarium abbreviations are according to Index Herbariorum except LC D Private herbarium of Laure CiveyrelcThis specimen was originally determined as Cibirhiza albersiana in Kunze et al (1994) but has recently been redetermined as
Fockea multiora K Schum (Verhoeven et al 2002)
of the clades in the successive weightinganalysis Clades with a successive weightedbootstrap gt63 were considered well sup-ported those with a successive weightedbootstrap gt85 were considered stronglysupported In a theoretical context a cladesupported by one uncontradicted charactercorresponds to a bootstrap support of 63anda cladesupported by two uncontradictedcharacters to 85 (Harshman 1994 Farriset al 1996) Here however this scale merelyrelates to the support distribution within theanalysis
RESULTS
The unit weight analysis of the completerbcL data set (237 characters total 214 ofthe cells were scored as missing data) re-sulted in 19003 trees 828 steps long witha consistency index (ci) D 0378 and a re-tention index (ri) D 0678 The consensustree from the successive weighting analysisof 252 most-parsimonious trees (each 1773steps long ci D 0661 ri D 0880) is presentedin Figure 1 The length of these trees withunit weight characters is 830 steps thus thesetrees are not identical to any of the most-parsimonious trees from the unit weightanalysis Branches not present in the strictconsensus tree from the unit weight analysisare indicated on the combinable consensustree from the successive weighting analysis
(Fig 1) The unit weight analysis of the com-bined set of the rbcL and ndhF data (153 char-acters total 043 of the cells were scored asmissing data) resulted in four trees 345 stepslong (ci D 0539 ri D 0482) The successiveweighting analysis gave three trees 978 stepslong (ci D 0840 ri D 0849) The strict consen-sus tree of the successive weighting analysisis presented in Figure 2 These trees are onestep longer with unit weights than the treesfrom the unit weight analysis Unless other-wise stated the following discussion refersto the successive weighting analyses of thecomplete rbcL data set
Investigating the justication of a prioriweighting (Sennblad and Bremer 2000) wemade some additional analyses of our rbcLdata set including tests for base compositionbias and rate heterogeneity This did not pro-duce signicantly different results and allwell-supported groups were congruent withthose of the present study The general struc-ture and the well-supported groups of theresulting trees (Figs 1 and 2) are to a largedegree congruent with the molecular stud-ies of Sennblad and Bremer (1996) whichare based on rbcL data of Civeyrel et al(1998) based mainly on matK sequence dataand of Potgieter and Albert (2001) basedmainly on trnLndashF data Although there aredifferences in the weakly supported rela-tionships among the well-supported groupsin particular the exact position of the
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 395
FIGURE 1 Combinable component consensus tree of the 252 most-parsimonious trees from the successiveweighting analysis of the complete rbcL data set Subfamilial and tribal classication is according to Endress andBruyns (2000) For subfamilies a three-letter code is used RAU APO PER SEC and ASC denote RauvoloideaeApocynoideae Periplocoideae Secamonoideae and Asclepiadoideae respectively OUT denotes the outgroupNumbers below branches are unit-weighted bootstrap values successive weighted bootstrap values are indicatedabove branches Branches not present in the strict consensus tree from the unit weight analysis are indicated witha cross (dagger) For branches marked A or B the frequencies in the combinable component consensus are 95 or 50respectively all other branches have 100 frequency
396 SYSTEMATIC BIOLOGY VOL 51
FIGURE 2 Strict consensus tree from the successive weighting analysis of the combined ndhF and rbcL dataset Tribal classication sensu Endress and Bruyns (2000) is indicated Numbers below branches are unit-weightedbootstrap values whereas successive weighted bootstrap values are indicated above branches Branches not presentin the strict consensus tree from the unit weight analysis are indicated with a cross (dagger)
Periplocoideae and of the root of the in-group all well-supported groups in our anal-ysis are congruent with those studies Thecombined analysis using data from rbcL-and matK-sequences and oral and pollenmorphology in the study of the Apocy-naceae sstr by Endress et al (1996) andthe analysis of Sennblad et al (1998) ofthe tribe Wrightieae sensu Leeuwenberg(1994a) and related taxa using rbcL andmorphological data are largely congruentwith the result from the present study Fur-thermore preliminary results from collab-orative studies by Sennblad et al (aimedat the traditional Apocynoideae and As-clepiadaceae pers comm) and Endresset al (aimed at the traditional Alyxieaepers comm) based on combined analy-sis of several molecular and morphologi-cal datasets are also congruent with ourresults
Support is good for the split betweenthe ingroup and the two outgroup taxaGelsemium and Mostuea (uwboot D 96swboot D 100) Thus the initial assump-tion of monophyly of the Apocynaceaesl is not violated Conforming to ear-lier studies (eg Sennblad and Bremer1996 Civeyrel et al 1998 Sennblad et al1998 Potgieter and Albert 2001) the tra-ditional Asclepiadaceae are nested withinthe traditional subfamily Apocynoideae(eg sensu Endress and Bruyns 2000)
The ApocynoideaendashAsclepiadaceae cladewhich is strongly supported (uwboot D 86swboot D 99) is in turn nested within thesubfamily Rauvoloideae sensu Endress andBruyns (2000)
Comparison of our results to the mostrecent classications of the ApocynaceaeEndress and Bruyns (2000 see Table 1 andFig 1) shows that although their circum-scription of the subfamilies PeriplocoideaeSecamonoideae and Asclepiadoideae aremonophyletic in our analysis their sub-families Rauvoloideae and Apocynoideaeare nonmonophyletic Similarly the tribesof the Rauvoloideae except Melodineaeare monophyletic but three of the tribesof the Apocynoideaemdashthe Apocyneae theEchiteae and the Wrightieaemdashare nonmono-phyletic Of the tribes of the Asclepiadoideaethe Ceropegieae and Asclepiadeae are mono-phyletic The position of Fockea as sisterto the rest of the Asclepiadoideae ren-ders the Marsdenieae nonmonophyletic Allthe included Periplocoideae taxa belongto the tribe Periploceae sensu Venter andVerhoeven (1997) except Pentopetia whichthey placed in the tribe CryptolepideaeGiven this unresolved position of Pentopetiathe monophyly of these tribes cannot be eval-uated however the reduction of Parquetinato synonymy with Periploca proposed byVenter and Verhoeven (1997) is not sup-ported by the present study
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 397
DISCUSSION
One of the most important tasks for anomenclature system is to provide a sta-ble means to communicate classications(which we here will understand as represent-ing phylogenetic relationships) Althoughboth the Linnaean and the Phylocode sys-tems are able to communicate phylogeniesthey are sensitive to changes in phylogenetichypotheses which in both systems will de-mand taxonomic changes In addition theLinnaean system is subject to a potentialinstability relating to subjective changes incircumscription that is lumpers and split-ters may refer to different groups of organ-isms by the same name (see de Queiroz andGauthier 1992) For many taxa a consensuscircumscription has eventually been estab-lished but for other groups this is still a prob-lem A current example of this from Apoc-ynaceae is the genus Tabernaemontana sensuLeeuwenberg (1991 1994b) which in theclassication of Allorge (1985) is split intoeight genera The Phylocode system was de-signed to avoid this problem by use of strictdenitions of taxa subjective changes incircumscription are avoided
However this rigidity in denitions leadsto a potential instability and high turnoverof names associated with a group of nestedtaxa Relatively small changes in phyloge-netic views eg when a dening taxonldquospecierrdquo receives a less nested positionmaycause some(or all) of these nested namesto become synonymous (eg de Queirozand Gauthier 1994 Bryant 1996 Liden andOxelman 1996Sereno 1999)The junior syn-onyms should according to the Phylocodebe rejected and new names may then beneeded for the new system of nested clades(de Queiroz and Gauthier 1994) The namesassociated with this group of taxa are thusreplaced by new names This can be espe-cially problematic when careless denitionshave been made as have been noted bySchander and Thollesson (1995) and Cantinoet al (1997) among others Consider a hy-pothetical worst-case example The Ascle-piadaceae were traditionally considered asseparate from Apocynaceae A stem-baseddenition of Apocynaceae reecting thisview could be the largest clade that in-cludes Apocynum but not Asclepias On thepresent tree such a denition would limitthe Apocynaceae to be synonymous with the
tribe Apocyneae and a new name wouldbe needed for the taxon corresponding tothe current view of Apocynaceae (eg sensuEndress and Bruyns 2000) A node-baseddenition would handle this particular prob-lem but may have other problems as illus-trated here with another hypothetical worst-case scenario Early classications of theApocynaceae included Plocosperma BenthA node-based denition representing thisview could have included as reference taxasay Apocynum and Plocosperma The recentlyproposed position of Plocosperma close tothe Boraginaceae (Backlund et al 2000)would with this denition have made Apoc-ynaceae a very large taxon (containing atleast both Gentianales and Boraginales) pos-sibly synonymous with (a hypothetical def-inition of) the informal taxon EuasteridsII (Bremer et al 1998) and again a newname for the Apocynaceae sensu Endressand Bruyns (2000 ie excluding Plocosperma)would have been needed Because theLinnaean system allows recircumscription ofsynonymous names the potential turnoverin the set of names in use is reduced withthis system (albeit at the expense of allowingsubjectivity) Additionally the ranks of theLinnaean system provide a set of names thatwork as relatively stable universal standardnames in communication note that equalranks does not guarantee comparable evo-lutionary units however Continuity in us-age of names is important for example injournals literature databases oras teach-ing and inventory work and the need forsuch continuity has been recognized by pro-ponents of both nomenclature systems (deQueiroz and Gauthier 1992 Greuter et al1994 Reveal 1996)
We propose a compromise approach thatuses solutions relating to these aspects fromboth systems concentrating mainly onbotanical classication of extant plant taxaat the tribal and familial levels and onlyconsidering monophyletic taxa We recom-mend a system that uses ldquostandard namesrdquo(ie universally used communication unitsat convenient hierarchical levels) such asthose provided by the principal ranks of theLinnaean system (Greuter et al 1994) but inwhich sensitivity toward subjective changesin circumscription is reduced One way to dothis is to adopt the denitions of the phy-logenetic system but use constraints similar
398 SYSTEMATIC BIOLOGY VOL 51
to those that apply between ranks in theLinnaean system To reduce sensitivity to-ward changes in phylogenetic hypotheses(Schander and Thollesson 1995 Bryant1996 1997 Cantino et al 1997 Moore 1998)we propose using combined node C stem-based denitions that will provoke explicitincompatibilities between taxa in case ofunfortunate changes in circumscription (deQueiroz 1996 has proposed a similar type ofdenition for designating nonmonophyletictaxa somewhat similar variants are alsodiscussed in the Phylocode) A node-basedpart of the denition provides a minimalcircumscription whereas a stem-based partprovides boundaries to competing taxa(maximal circumscription) If on a particu-lar phylogeny the minimal (monophyletic)circumscriptions of competing taxa are over-lapping then the two taxa are incompatibleTo make this meaningful at least two taxashould be included in the node-based partof the denition this will also avoid unnec-essary monotypic taxa This incompatibilitywill provide an objective point when nomen-claturalreconsideration isneeded Returningto the rst of our examples above node Cstem-based denitions of the Apocynaceaeand Asclepiadaceae could be as follows
The family Apocynaceae is the most inclusive clade inthe order Gentianales including the type specimensof Apocynum and Dictyophleba but not the taxa Ascle-piadaceae [ Gentianaceae Rubiaceae ]
and
The family Asclepiadaceae is the most inclusive cladein the order Gentianales including the type speci-mens of Asclepias and Periploca but not the taxa Apoc-ynaceae [ Gentianaceae Rubiaceae ]
On the tree in Figure 1 these two denitionsare mutually exclusive and a decision onwhich name to keep must be made Mini-mizing the number of classicatory changesshould take precedence in this decision Thereason for this is to avoid the large incon-venient ldquostandard namerdquo changes that couldresult from relatively small rearrangementsRejected names are then ignored when oc-curring in denitions of other taxa This isbecause the reference taxa in the stem-basedpart of the denitions are ldquodenedrdquo taxarather than physical types (thus reliance ondened reference taxa which has been crit-icized for stem-based denitions by Sereno1999 is what we aim for on the other hand
the explicit reference of type specimens in theminimum circumscription provided by thenode-based part of the denitions circum-vents the tautology problem raised byBryant 1996) In the current example thiswould mean that regardless of which nameof Apocynaceae and Asclepiadaceae is re-jected the other will be dened as corre-sponding to the Apocynaceae sensu Endressand Bruyns (2000) The choice will there-fore simply be which name to keep In thiscase because Apocynaceae has priority un-der the Linnaean system we would chooseto keep Apocynaceae to promote consistencywith the current system Nevertheless out-side these recommendations choices of whatnames to choose will probably to a certaindegree be arbitrary We also propose thatthe rejected name should be reinstated withan emended denition if appropriate to po-tentially allow for more stability in the setof names in use Thus in our second ex-ample above a node C stem-based deni-tion reecting the inclusion of Plocospermain Apocynaceae (eg by exchanging Dic-tyophleba for Plocosperma in the denitionabove) would presumably be incongruentwith the denition of Gentianales as well asthe denitions of Rubiaceae Gentianaceaeand so forth and therefore would be rejectedIn such acasewhere the only real change is inthe position of a single taxon a reinstatementof the Apocynaceae with an emended def-inition is motivated However a drawbackis that this may also introduce elements ofsubjectivity in the circumscriptions of taxa
Not all named clades need to be providedwith standard names In fact for the purposeof a universal set used in journals databasesand so forth it may sufce to use standardnames corresponding to the principal ranksof the Linnaean system (Greuter et al 1994)such as species genus family and orderNevertheless to provide compatibility withthe present system we suggest that addi-tional names corresponding to commonlyused secondary ranks (eg tribes) may berecognized The hierarchical level of a stan-dard name needs to be indicated This doesnot imply that taxa of the same hierarchi-cal level are comparable evolutionary unitsbut simply indicates the hierarchical levelof a taxon relative to nested taxa Becausesuch misinterpretations will in any case bedifcult to prohibit one may as well pro-vide compatibility with the present Linnaean
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 399
system by using the names of the Linnaeanranks (eg species genus tribe familyorder) to indicate hierarchical level The prin-ciple of exhaustive subsidiary taxa is notaccepted so redundant taxa need not be rec-ognized Other ldquononstandardrdquo taxa couldbe dened by using Phylocode denitionsmdashthat is without the restraints discussed forstandard taxa above
One advantage of the described systemis that utilizing the correspondence betweenthe standard names and Linnaean ranks al-lows the system to be largely compatiblewiththe establishedLinnaean system A change ofsystem could therefore be gradual with thetwo systems coexisting during a conversiontime (see eg de Quieroz 1997b)
Even though the node C stem-baseddenitions reduce the sensitivity towardschanges in circumscriptions they may still besensitive towardldquounfortunaterdquo or ldquobadrdquo def-initions (see eg Cantino et al 1997Cantinoand de Queiroz 2000) Many of the recom-mendations for such things as choices of ref-erence taxa (speciers) in phylogenetic def-initions are applicable also to the node Cstem-based denitions (eg Schander andThollesson 1995 Bryant 1996 Cantino et al1997 Sereno 1999) Most likely the compro-mise system described above will have sev-eral further problems and we hope that thispaper will invite further discussion on thesubject
A New Classication of the Apocynaceae slOur results indicate that some problem-
atic taxa remain in the classication ofEndress and Bruyns (2000) Monophyly fortwo of the subfamilies and ve of the tribesis called into question If other recent molec-ular studies (eg Potgieter and Albert 2001)are taken into consideration a further fourtribes (Alstonieae Alyxieae Plumerieae andVinceae) may be nonmonophyletic This canin some cases be simply analysis artifactsbut as Endress and Bruyns (2000) themselvespoint out their classication is to be consid-ered preliminary Insufcient taxonsamplingin molecular studies and difcult homologydecisions relating to morphological charac-ters make the circumscription of many ofthe tribes for example in the Apocynoideaesensu Endress and Bruyns (2000) uncertainThere is thus risk for future rearrangementsin many of the tribes A classication using
node C stem-based denitions may be ableto subsume such rearrangements withoutnumerous explicit recircumscriptions Wewill here use the result fromthe present studyas a basis to propose such a classication ofthe Apocynaceae sl
Our aim is to base taxa on clades that arewell supported A second aim is to makeour classication compatible with the cur-rent system We will therefore primarily rec-ognize nonoverlapping tribes that are validunder the Linnaean system We will furtheradopt the principle of nested referencing (assuggested by Lee 1999 and Sereno 1999)from the Linnaean system Thus the rst ref-erence taxon in the node-based part of thedenition is the primary type which corre-sponds to the Linnaean type used for thename of the taxon Instead of the (optional)enumeration of subsumed taxa of a lowerrank (eg genera or subtribes) we will in-clude node C stem-based denitions as dis-cussed above Citations to primary types ofreference taxa in node-based part of de-nitions are also nested and thus refer toldquothe type specimen of the type species ofthe type genus of the dened taxonrdquo Aproblem is that the relationships between thetribes areweakly supported This may lead tovery cumbersome denitions enumeratingall competing tribes in the stem-based partof the denition (eg Moore 1998) One wayto reduce this problem is to dene interme-diate well-supported taxa and use them ascompeting taxaWe therefore will further rec-ognize four nonstandard taxa (note that anassignment of ldquononstandardrdquo taxa does notimply less reliability standard names relateto communication purposes only) The taxadiscussed are indicated in Figure 3 We willdiscuss the tree from the top of Figure 3 start-ing with the taxa of the traditional Plumeri-oideae Tribes and subtribes given in paren-theses in the text refer to the classicationsof Endress and Bruyns (2000) in other casesour classication is implied For some tribeswe have chosen a name with priority underLinnaean system that implies inclusion of agenus not sampled in the present study Forall but one (Willughbeieae) of these casesadditional molecular or morphological phy-logenetic analyses support the inclusion ofthe genus Even if too uncertain denitionsshould be avoided this is less critical withnode C stem-based denitions In the caseof Willughbeieae we have decided from
400 SYSTEMATIC BIOLOGY VOL 51
FIGURE 3 The tribal reclassication discussed in the text indicated on the combinable component consensustree from the successive weighting analysis of the complete rbcL data set Four nonstandard taxamdashApocynoidinaEuapocynoidina Asclepiadacina and Asclepiadoidinamdashare also indicated
discussions with a morphological expert(M Endress pers comm) that morphologi-cal support exists for including WillughbeiaRoxb in Willughbeieae If this is correctthe denition given below will be valid if
it is not our system allows redenition ofthe taxon We have also tried to providecomments on potential morphological char-acters taken from literature for the differ-ent taxa recognized However we have not
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 401
performed any morphological analysis butrely on information external to this studyfor these comments In most cases we cantherefore not identify synapomorphies forthe taxa
TAXONOMIC TREATMENT
Tribus Willughbeieae ADCThe Ancylobotrys Dictyophleba and Va-
hadenia clade is strongly supported (uwbootand swboot D 100) and corresponds to theWillughbeieae of Endress and Bruyns (2000)The present sampling contains closely re-lated genera that have been traditionallykept together Landolphia P Beauv sl in-cludes all these taxa (eg Schumann 1895)In addition to the taxa sampled here thestudy of Potgieter and Albert (2001) sup-ports the inclusion of Couma Aubl LacmelleaH Karst Pacouria Aubl and Saba (Pichon)Pichon in this clade Plants in this cladeall have a gynoecium that is congenitallysyncarpous and indehiscent and that con-tains a eshy pulp including numerous seedswith copious horny endosperm (Fallen 1986Persoon et al 1992) This probably con-stitutes synapomorphies for this clade(M Endress pers comm) although not un-ambiguously for example syncarpy is par-alleled in Carisseae We will therefore as-sume that the reference genus Willughbeia isincluded in this tribe
DenitionmdashTribe Willughbeieae is themost inclusive clade that includes the prim-ary types of Willughbeia and Dictyophleba butnot Alyxieae Aspidospermeae CarisseaeHunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Apocynoidina
Tribus Vinceae Bartl
Ochrosia Rauvola Vinca and Catharanthus(all Vinceae) form a strongly supported clade(uwboot D 89 swboot D 99) The associa-tion between Rauvola and Catharanthus wasindicated by Sennblad and Bremer (1996)The study of Potgieter and Albert (2001) sup-ports inclusion of the genera LaxoplumeriaMarkgr Neisosperma Raf and Tonduzia Pit-tier previously classied with AspidospermaThe Vinceae contain both herbaceous andfruticose taxa The fruits are apocarpousbut whereas the fruits of Catharanthus Lax-oplumeria Neisosperma Tonduzia and Vincahave dry fruit walls the fruits of Ochrosia andRauvola are drupes Plants in Vinceae share
a similar type of style head (ie the apicalenlarged part of the style) with a stigmatichollow and upper and lower hair wreaths(except in some species of Ochrosia) the pres-ence of a nectar disk and a linear hilum onthe seed (Pichon 1948b) On the present treeVinceae also include Kopsia however this po-sition of Kopsia is very weakly supported (seealso Sennblad and Bremer 1996)
DenitionmdashTribe Vinceae is the mostinclusive clade that includes the primarytypes of Vinca and Rauvola but not AlyxieaeAspidospermeae Carisseae HunterieaeMelodineae Plumerieae Tabernaemon-taneae Willughbeieae or Apocynoidina
Tribus Tabernaemontaneae G DonThe strongly supported association be-
tween Tabernaemontana Tabernanthe Carval-hoa Schizozygia and Molongum (Tabernae-montaneae uwboot D 84 swboot D 98)has been suggested previously (Boiteau et al1978 Fallen 1986 Endress et al 1996Sennblad and Bremer 1996) Inclusion of taxatraditionally associated with Tabernaemon-tana (eg Voacanga Thouars and CallichiliaStapf see also Tabernaemontaneae sensuLeeuwenberg 1994a) and Macoubea Aublis supported by the analysis of Potgieterand Albert (2001) The taxa share sclerenchy-matic anthers free from the style head anda characteristic band of heavily cutinizedcells at the insertion on the staminal rib(Endress et al 1996) The taxa tradition-ally associated with Tabernaemontana arecharacterized by their apocarpous fruitswith arillate seeds whereas genera asso-ciated with Molongum (Ambelanieae sensuLeeuwenberg 1994a) have syncarpous fruitsand lackanarillus HoweverMacoubea formsa link between the two having a syncar-pous fruit with arillate seeds (Zarucchi et al1995)
DenitionmdashTribe Tabernaemontaneae isthe most inclusive clade that includes the pri-mary types of Tabernaemontana and Schizozy-gia but not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae PlumerieaeVinceae Willughbeieae or Apocynoidina
Tribus Aspidospermeae Miers
Aspidosperma and Vallesia (Aspidosper-meae) form a strongly supported clade (uw-boot and swboot D 100) rst suggestedin Sennblad (1997 see also Sennblad and
402 SYSTEMATIC BIOLOGY VOL 51
Bremer 2000) Subsequently GeissospermumAllemao Haplophyton ADC MicroplumeriaBaill and Strempeliopsis Benth were indi-cated to belong to this clade (Potgieter andAlbert 2001) The Aspidospermeae havesimple style heads lacking both stigmatichollow and upper hair wreath and haveapocarpous fruits with variation in eshi-ness and sclerication Alstonia has tradi-tionally been placed in the Plumerieae Be-cause its seeds have a hairy margin it hasbeen suggested to form a link to subfam-ily Apocynoideae This was contradicted bySennblad and Bremerrsquos study (1996) how-ever where Alstonia was placed in an iso-lated position as the sister group to therest of the Apocynaceae sl Here it groupswith Aspidospermeae Sensu Endress andBruyns (2000) However this association isvery weakly supported (uwboot and swbootlt 50) and in other recent studies As-pidospermeae and Alstonia do not form aclade (Potgieter and Albert 2001) Becauseof this we do not use it as a reference taxonin the denition Thus although the Aspi-dospermeae include Alstonia on the presenttree some other position of Alstonia canbe accommodated without amending thedenition
DenitionmdashTribe Aspidospermeae is themost inclusive clade that includes theprimary types of Aspidosperma and Vallesiabut not Alyxieae Carisseae HunterieaeMelodineae Plumerieae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Melodineae G DonThe well-supported association between
Melodinus and Craspidospermum (uwboot D70 swboot D 75) was rst reportedby Sennblad (1997 see also Sennblad andBremer 2000) and has subsequently beenreected in the tribe Melodineae sensuEndress and Bruyns (2000 their inclusion ofDiplorhynchus however is not supported)Craspidospermum although its fruit is dryand dehiscent as opposed to the indehis-cent fruits of Melodinus was noted by Pichon(1948b) to present several characters of theCarisseae such as stamens inserted near thebase of the corolla tube and a syncarpousovary In particular Pichon found the verydense suprastaminal indumentum type rem-iniscent of Melodinus From his descriptions
of the two genera (Pichon 1948ab) the fol-lowing similarities also emerge presence ofa stipular line pollen in tetrads and a punc-tiform hilum
DenitionmdashTribe Melodineae is the mostinclusive clade that includes the pri-mary types of Melodinus and Craspidosper-mum but not Alyxieae AspidospermeaeCarisseae Hunterieae Plumerieae Taber-naemontaneae Vinceae Willughbeieae orApocynoidina
Tribus Hunterieae K Schum
The Picralima and Pleiocarpa clade isstrongly supported (uwboot and swboot D100) and corresponds to the subtribeHunterieae sensu Endress andBruyns (2000)The study of Potgieter and Albert (2001)also supports the inclusion of Hunteria RoxbThese taxa are characterized by apocar-pous sometimes pluricarpous ovaries andeshy brous fruit walls (Omino 1996)Diplorhynchus (Melodineae) is here weaklyassociated (uwboot and swboot lt 50) withthe Pleiocarpeae It shares a few similari-ties such as stipular lines and a style headwithout a stigmatic hollow and hair wreathsbut also has differences such as a dry de-hiscent fruit With the present denitionDiplorhynchus will tentatively be included inthe Hunterieae
DenitionmdashTribe Hunterieae is the mostinclusive clade that includes the primarytypes of Hunteria and Picralima but not Alyx-ieae Aspidospermeae Carisseae Melod-ineae Plumerieae TabernaemontaneaeVinceae Willughbeieae or Apocynoidina
Tribus Plumerieae EndlThe strongly supported relationship
(uwboot and swboot D 100) between Alla-manda and Plumeria corresponds to earlierstudies (Endress et al 1996 Sennblad andBremer 1996 Civeyrel et al 1998 Potgieterand Albert 2001) In all earlier treatmentsAllamanda has had uncertain relationshipsThe association with Plumeria nds supportin pollen morphology such as perforatemesocolpial depressions and similar innerexine pattern The two genera also containsecoiridoids rather than the indole alka-loids and cardenolides that are commonin the traditional Plumerioideae (Endresset al 1996) Cameraria Cerbera and Thevetiaform a strongly supported association
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 403
(uwboot D 83 swboot D 91) with Ane-chites which has been suggested by Fallen(1983) This association was based mainlyon the latrorsesublatrorse anthers andthe broad style head with large apicalappendages and stigmatic hollow The studyby Potgieter and Albert (2001) suggesteda further inclusion of Cerberiopsis Viell ExPancher amp Sebert and Skytanthus Meyen inthis tribe These taxa form a well-supportedclade with Plumeria and Allamanda whichcorresponds to the Plumerieae sensuEndress and Bruyns (2000) Characterssupporting this relationship are the presenceof infrastaminal as well as suprastaminalappendages (not present in Anechites andPlumeria) and winged seeds (not present inAnechites) (Endress et al 1996)
DenitionmdashTribe Plumerieae is themost inclusive clade that includes theprimary types of Plumeria and Allamandabut not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Carisseae Dumort
Acokanthera and Carissa (both Carisseae)are traditionally joined and are sometimeseven treated as one genus (eg Pichon1948a) In the present study they form astrongly supported clade (uwboot D 96swboot D 99) Among the characters sup-porting this tribe are syncarpous fruits with-out axile placentation and endocracks on theinner exine of the pollen (Endress et al 1996)
DenitionmdashTribe Carisseae is the most in-clusive clade that includes the primary typesof Carissa and Acokanthera but not AlyxieaeAspidospermeae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
Tribus Alyxieae G Don
Chilocarpus has been a genus with un-certain afnities In Endress et al (1996) itwas the sister taxon to the ApocynoideaeHere it is associated with Alyxia and Lep-inia in a clade corresponding to the Alyxieaesensu Endress and Bruyns (2000) Thisclade is strongly supported with succes-sive weighted bootstrap (swboot D 91) butnot with unit-weighted bootstrap support(uwboot D 58) such clades will henceforthbe termed moderatelywell supported clades
A further inclusion of Condylocarpon DesfLepiniopsis Valeton and Plectaneia Thouarswas indicated by Potgieter and Albert (2001)The constituent taxa are the only Apocy-naceae taxa outside the Apocynoidina thathave porate pollen
DenitionmdashTribe Alyxieae is the most in-clusive clade that includes the primary typesof Alyxia and Lepinia but not Aspidosper-meae Carisseae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
APOCYNOIDINA
On the present tree the traditionalsubfam-ily Apocynoideae includes all subfamilies(Periplocoideae Secamonoideae and Ascle-piadoideae) of the Asclepiadaceae sensuLiede and Albers (1994) All of these taxaare established names and selecting one asa subfamily at the expense of the othersmight be considered unfortunate from dif-ferent points of view We have here chosennot to recognize subfamilies instead we willdene the traditional Apocynoideae Ascle-piadaceae and Asclepiadoideae as nonstan-dard taxa The correspondence to the tradi-tional taxa will be indicated in the names byreplacing the sufx -eae with a neutral suf-x -ina (Kron 1997) We will also for practi-cal reasonsrecognize as a nonstandardtaxonthe informal group euapocynoids suggestedby Sennblad et al (1998)
The taxa of the traditional Apocynoideaeand Asclepiadaceae (uwboot D 86 swboot D99) clade will in many cases be more dif-cult to safely delimit to tribes becausemany of the groups in this clade have weakor no support (although preliminary resultsfrom an unpublished collaborative study bySennblad et al indicate additional supportfor the tribes discussed below) The tribaland subtribal classication of the traditionalApocynoideae has been shown to be prob-lematic (Leeuwenberg 1994a Endress andBruyns 2000) The circumscription of tribesfor these taxa might therefore in some casesbe preliminary
In all following taxa the basal part of theanther connective called the retinacle is ad-nate to the style head
DenitionmdashApocynoidina are the mostinclusive clade that includes the pri-mary types of Apocynum and Wrightia butnot Alyxieae Aspidospermeae Carisseae
404 SYSTEMATIC BIOLOGY VOL 51
Hunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Willughbeieae
Tribus Wrightieae G DonThe Stephanostema and Wrightia (both
Wrightieae) clade is strongly supported(uwboot and swboot D 100) and correspondsto the Wrightieae sensu Sennblad et al (1998)who also suggested an inclusion of Pleio-ceras Baill The tribe could be characterizedby a combination of synapomorphic and ple-siomorphic characters for example presenceofa chalazalandabsence of micropylar comaleft contorted aestivation and absence of airspaces in the anthers (1998)
DenitionmdashTribe Wrightieae is the mostinclusive clade in the Apocynoidina that in-cludes the primary types of Wrightia andStephanostema but not Malouetieae Nerieaeor Euapocynoidina
Tribus Nerieae (Benth) M Pichon
Adenium Nerium and Strophanthus (allWrightieae) constituted the Nerieae inSennblad et al (1998) however the cladereceived weak support Here Adenium andNerium form a clade that is absent in halfof the most-parsimonious trees in the com-plete rbcL analysis but is strongly supportedin the combined ndhF and rbcL analysis(uwboot D 98 swboot D 99) Strophan-thus groups with the Mascarenhasia Pachy-podium Kibatalia Funtumia and Holarrhenaclade which corresponds to the Malouetieaesensu Sennblad et al (1998) This position ofStrophanthus is however weakly supported(uwboot and swboot lt 50) and is not presentin the combined ndhF and rbcL tree Fur-thermore certain morphological characters(unfused slits in the corolla tube and anapical anther appendage) indicate a posi-tion closer to the Nerieae (Sennblad et al1998)
DenitionmdashTribe Nerieae is the most in-clusive clade in the Apocynoidina that in-cludes the primary types of Nerium and Ade-nium but not Malouetieae Wrightieae orEuapocynoidina
Tribus Malouetieae Mull-Arg
The association of Mascarenhasia Pachy-podium Kibatalia Funtumia and Holar-rhena (all Wrightieae) is moderately wellsupported in the complete rbcL analysis
(uwboot D 57 swboot D 74) whereasthe corresponding clade is well supportedin the combined rbcL and ndhF analysis(uwboot D 66 swboot D 100) On the ba-sis of identied synapomorphies (presenceof calcium oxalate crystals in the stomium ofthe anthers absence of interpetal vein ses-sile laments and absence of air spaces inthe anthers) Sennblad et al (1998) predictedan inclusion of Alaa Thouars Kibatalia andMalouetia ADC in the tribe The inclusion ofKibatalia is strongly supported in this studyFollowing these suggestions we will here as-sume the inclusion of the reference genusMalouetia in this tribe (The denition belowwill include Strophanthus on the present treebut see comments under Nerieae)
DenitionmdashTribe Malouetieae is the mostinclusive clade in the Apocynoidina that in-cluded the primary types of Malouetia andFuntumia but not Nerieae Wrightieae orEuapocynoidina
EUAPOCYNOIDINA
In all the remaining representatives ofthe Apocynoideae sensu Endress andBruyns (2000) that is Mandevilla MesechitesAganosma Apocynum TrachelospermumRhabdadenia Beaumontia Parsonsia andPrestonia the anthers are adnate to the stylehead both by the retinacle and by the thecaeThese taxa together with the taxa of the tra-ditional Asclepiadaceae form a moderatelysupported clade (uwboot lt 50 swboot D86) that corresponds to the informal taxoneuapocynoids sensu Sennblad et al (1998)
DenitionmdashEuapocynoidina are the mostinclusive clade that includes the primarytypes of Apocynum and Echites but notMalouetieae Wrightieae or Nerieae
Tribus Mesechiteae Miers
Mandevilla and Mesechites (both Mese-chiteae) form a moderately supported clade(uwboot lt 50 swboot D 87) equivalentto the Mesechiteae This tribe could be char-acterized by anthers with obtuse truncatedtails and a retinacle lacking hairs and also astrongly pentagonal style head with a stig-matic hollow
DenitionmdashTribe Mesechiteae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Mesechitesand Mandevilla but not Apocyneae EchiteaePeriploceae or Asclepiadacina
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 405
Tribus Apocyneae Bercht et J Presl
Aganosma Apocynum Beaumontia andTrachelospermum (Apocyneae) form a weaklysupported clade with Rhabdadenia (Echiteae)(uwboot and swboot lt 50) in the completerbcL analysis whereas the correspondingclade in the combined ndhF and rbcL analy-sis gains a little better support (uwboot lt 50swboot D 91) Likewise this clade (with theinclusion of Chonemorpha G Don) is weaklysupported in Potgieter and Albert (2001) Po-tential characterizing traits could be a stylehead without stigmatic hollow in combina-tion with thecae that are adnate to the stylehead (Sennblad et al 1998)
DenitionmdashTribe Apocyneae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Apocynumand Trachelospermum but not Echiteae Mese-chiteae Periploceae or Asclepiadacina
Tribus Echiteae Bartl
Parsonsia Peltastes and Prestonia (allEchiteae) form a moderately supported orunsupported clade with the Periplocoideaerepresentatives of the study (uwboot andswboot lt 50 and uwboot lt 50 swboot D89 in the complete rbcL and the combinedrbcL and ndhF analysis respectively) In thestudy of Sennblad et al (1998) Prestonia andParsonsia formed a weakly supported cladeand in the study of Potgieter and Albert(2001) inclusion of both Parsonsia and EchitesP Browne in the Echiteae is supportedParsonsia and Echiteae also share certaincharacters such as a horseshoe-shaped reti-nacle and a style head with stigmatic hollow(see Sennblad et al 1998)
DenitionmdashTribe Echiteae is the most in-clusive clade in the Euapocynoidina thatincludes the primary types of Echites andPrestonia but not Apocyneae MesechiteaePeriploceae or Asclepiadacina
Tribus Periploceae BartlThe representatives of the Periplo-
coideae sensu Endress and Bruyns (2000)mdashPentopetia Parquetina Petopentia PeriplocaTacazzea and Mondiamdashgroup together ina well-supported clade (uwboot D 67swboot D 93) The monophyly of the tra-ditional periplocoid taxa is also supportedin earlier studies that included a wider ordifferent sampling (Civeyrel et al 1998
Potgieter and Albert 2001) The taxa ofthe traditional Asclepiadaceae have ldquotrans-latorsrdquo structures related to pollinationspecialization In the Periploceae the pollenis deposited as tetrads in a spoon-liketranslator with an adhesive disc
DenitionmdashTribe Periploceae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Periplocaand Pentopetia but not Apocyneae EchiteaeMesechiteae or Asclepiadacina
ASCLEPIADACINA
The taxa of the Asclepiadoideae sensuLiede and Albers (1994) form a clade togetherwith the two Secamone species and Baissea(Apocyneae uwboot lt 50 swboot D 73)The exclusion of Periploceae (traditionallyassociated with this clade see Civeyrel et al1998) from this taxon is weakly supportedTherefore Periploceae is not included as areference taxon in the stem-based part ofthe denition of the Asclepiadacina thusleaving open the possibility for a later in-clusion of this taxon In the Asclepiadacina(except Baissea) the pollen is agglutinatedinto pollinia that are connected to a claspingtranslator
DenitionmdashAsclepiadacina are the mostinclusive monophyletic clade in the Euapoc-ynoidina that includes the primary types ofAsclepias and Secamoneae but not ApocyneaeEchiteae or Mesechiteae
Tribus Secamoneae G DonThe tribe Secamoneae sensu Bruyns
(uwboot D 89 swboot D 96) is mono-phyletic in the present analysisHowever be-cause representatives from only one of theconstituent genera are included monophylyis not well tested but the study of Civeyrelet al (1998) which included a larger numberof taxa including Pervillea Decne showedthat monophyly of the tribe was supportedThe traditional character for this tribe is fourpollinia per translator
DenitionmdashTribe Secamoneae is the mostinclusive clade in the Asclepiadacina that in-cludes the primary types of Secamone andPervillea but not Asclepioidina or Baissea
BaisseaThe position of Baissea corresponds with
that in the study of Sennblad et al (1998)
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
392 SYSTEMATIC BIOLOGY VOL 51
sister group to the Apocynaceae sl in therecent analysis of Gentianales by Backlundet al (2000)
Forty-one new sequences of rbcL are pub-lished in this study rbcL sequences forthe outgroup taxa and for 35 ingroup taxawere published previously (Table 2) ForndhF the 30 region of the gene was se-quenced for 18 taxa representing the tra-ditional Apocynoideae and AsclepiadaceaeTotal DNA was extracted from fresh leavesor herbarium material by using the meth-ods of Saghai-Maroof et al (1984) and Doyleand Doyle (1987) The extractions were pu-ried by ultracentrifugation in CsCl gra-dients or ethanol precipitation Additionalpurication with the Qiaquick PCR puri-cation kit (Qiagen Inc) was performed incases with problematic polymerase chain re-action (PCR) amplication Double-strandedDNA was amplied by PCR by using Taq-polymerase kit (Promega Corp) SyntheticPCR primer sequences for rbcL were takenfrom Olmstead et al (1992) ndhF primerswere taken from Olmstead and Sweere (1994)and Oxelman et al (1999) In cases wherePCR amplication proved difcult a PCRreaction using Taq extender PCR additive(Stratagene Inc) was performed followingthe protocol provided by the supplier A sec-ond round of PCR with only one of theprimers was performed to obtain single-stranded DNA (Kaltenboeck et al 1992)Single-stranded DNA was sequenced man-ually (Sanger et al 1977) by using internalprimers designed for rbcL by G Zurawski(DNAX Research Institute) and for ndhF byOlmstead and Sweere (1994) and Oxelmanet al (1999)
The length of the sequences of rbcL ex-cluding the 26 rst nucleotides but includ-ing positions just downstream from the geneobtained in this study varied between 1411and 1509 nt For all taxa except Parsonsia thesequence includes the stop codon In posi-tions 27ndash1425 no structural mutations occurin rbcL Alignment of this part of the geneis thus very simple However after position1425 that is just before the typical positionof the stop codon gaps of different sizes oc-cur complicating alignment Similarly gapsoccur in the coding region of the ndhF se-quences Because of these alignment prob-lems the sequences were truncated 23 basesbefore the typical position of the stop codonin Apocynaceae sl The reading frame was
inferred through comparisonswith completeApocynaceae ndhF-sequences from the studyof Backlund et al (2000)
Multiple sequence alignment was made byhand to reduce the number of gaps while in-creasing the percentage similarity We usedthe criteria of Golenberg et al (1993) with mi-nor modications
Two cladistic analyses were performedThe data matrix for the rst analysis coveredall included taxa and comprised characterscorresponding to the nucleotide positions27ndash1425 of the rbcL gene and characterscorresponding to gaps and informative nu-cleotide positions downstream of position1425 EMBL database accession numbersfor rbcL sequences are given in Table 2Only parsimony informative characters wereanalyzed
The second analysis was of a subset ofthe taxa included in the rst analysis andincludes 18 genera representing the Apoc-ynoideae sensu Leeuwenberg (1994a) andthe traditional Asclepiadaceae The data ma-trix consisted of two submatrices One sub-matrix comprised positions 27ndash1425 of therbcL gene The second submatrix comprised733 aligned nucleotide positions from the3rsquo region of ndhF EMBL database accessionnumbers for the ndhF sequences are given inTable 2 Only parsimony informative charac-ters were analyzed
The cladistic analyses were performed us-ing PAUP 40b2 (Swofford 1998) An initialheuristic search (PAUP settings HSEARCH[ADDSEQ D RANDOM NREPS D 200SWAP D TBR] other options withdefault set-tings) with all characters given a unit weightwas followed by a successiveweighting anal-ysis (Farris 1969 1989) also using heuris-tic searches (PAUP settings HSEARCH[ADDSEQ D RANDOM NREPS D 10SWAP D TBR] other options withdefault set-tings) and characters reweighted accordingto their rescaled consistency index
Bootstrap frequencies (Felsenstein 1985)were calculated with 10000 replicates on theunit-weighted (uwboot) and successiveweighted (swboot) data matrices (PAUPsettings BOOTSTRAP [NREPS D 10000METHOD D HEURISTIC CONLEVEL D50] [ADDSEQ D RANDOM NREPS D 1SWAP D SPR NOMULPARS] other optionswith default settings)
To simplify the discussion of the resultswe used a rough scale of the relative support
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 393
TABLE 2 Vouchers andEMBLaccession numbers for sequences published in this study Forpreviously publishedsequences and for sequences from previously published voucher specimens the reference is given
Taxon Vouchersourceb EMBL number rbcLndhF
Acokanthera oblongifolia (Hochst) Codd Sennblad and Bremer 1996 X91758-Adenium obesum (Forssk) Roem et Schult Sennblad et al 1998 AJ002880AJ420130Aganosma marginata (Roxb) G Don Ryding 490 UPS AJ419730-Allamanda cathartica L Sennblad and Bremer 1996 X91759-Alstonia scholaris RBr Sennblad and Bremer 1996 X91760-Alyxia ruscifolia RBr Sennblad 236 UPS AJ419731-Ancylobotrys petersiana Pierre Sennblad 201 UPS AJ419732-Anechites nerium Urb Bremer et al 3386 UPS AJ419733-Apocynum cannabinum L Sennblad and Bremer 1996 X91761AJ420113Araujia hortorum E Fourn Bremer 3006 UPS AJ419734-Asclepias curassavica L Sennblad and Bremer 1996 X91774-Aspidosperma triternatum Rojas Acosta Bremer 3029 UPS AJ419735-Baissea leonensis Benth Sennblad et al 1998 Yallah 112 UPSa AJ002881AJ420114Beaumontia grandiora Wall Sennblad et al 1998 AJ002882AJ420115Calotropis procera (Aiton) WT Aiton Sennblad 222 UPS AJ419736-Cameraria latifolia L Houghton 1085 FTG AJ419737-Carissa bispinosa (L) Merxm Sennblad 235 UPS AJ419738-Carvalhoa campanulata K Schum Sennblad 217 UPS AJ419739-Catharanthus roseus G Don Sennblad and Bremer 1996 X91757-Cerbera venenifera Steud No voucher AJ419740-Ceropegia woodii Schltr Sennblad and Bremer 1996 X91775-Chilocarpus suaveolens Blume Endress et al 1996 X92445-Fockea multiora K Schum Specks 248 cult MSTRc AJ419741-Craspidospermum verticillatum Bojer Pettersson and Nilsson 742 AJ419742-Dictyophleba lucida Pierre Sennblad and Bremer 1996 X91762-Diplorhynchus condylocarpon (Mull-Arg) Pichon Sennblad 203 UPS AJ419743-Fischeria stellata (Vell) E Fourn Clark and Watt 793 UPS AJ419744-Funtumia elastica Stapf Sennblad et al 1998 AJ002884-Gelsemium sempervirens Aiton Olmstead et al 1993 L14397-Holarrhena pubescens G Don Sennblad et al 1998 AJ002884-Hoya bella Hook Sennblad and Bremer 1996 X91776-Kibatalia gitingense (Elmer) Woodson Liede 3268 Z AJ419745-Kopsia fruticosa ADC Sennblad and Bremer 1996 X91763-Lepinia taitensis Decne Perlman 15071 NTBG AJ419746-Mandevilla sanderi (Hemsl) Woodson Sennblad and Bremer 1996 X91764AJ420116Mascarenhasia arborescens ADC Sennblad et al 1998 AJ002885AJ420117Matelea hirsuta (Vahl) Woodson Sennblad 263 UPS AJ419747-Melodinus monogynus Roxb Sennblad 261 UPS AJ419748-Mesechites trida Mull-Arg Bremer et al 3351 UPS AJ419749-Micholitzia obcordata NEBr Bremer 3010 UPS AJ419750-Molongum laxum (Benth) Pichon Sennblad and Bremer 1996 X91765-Mondia ecornuta (NEBr) Bullock Sennblad 215 UPS AJ419751-Mostuea brunonis Didr Olmstead et al 1993 L14404-Nerium oleander L Sennblad 265 UPS AJ002886AJ420118Ochrosia coccinea Miq v d Laan 389 WAG AJ419752-Orthosia serpyllifolia Decne Bremer et al 3372 UPS AJ419753-Pachypodium lamerei Drake Sennblad et al 1998 AJ002887AJ420119Parquetina nigrescens (Afzel) Bullock Sennblad and Bremer 1996 X91777-Parsonsia heterophylla A Cunn Sennblad et al 1998 AJ002888-Peltastes peltatum (Vell) Woodson Sennblad 262 UPS AJ419754-Pentopetia sp No voucher AJ419755-Periploca graeca L Sennblad et al 1998 AJ002889AJ420120Petopentia natalensis (Schltr) Bullock Sennblad 237a and sn cult MSTR AJ419756AJ420121Picralima nitida T et H Dur Sennblad and Bremer 1996 X91766-Pleiocarpa mutica Benth Bremer 3017 UPS AJ419757-Plumeria inodora Jacq Sennblad and Bremer 1996 X91767-Prestonia quinquangularis Spreng Sennblad and Bremer 1996 X91768AJ420122Rauvola mannii Stapf Sennblad and Bremer 1996 X91769-Rhabdadenia biora Mull-Arg Zona 616 FTG AJ419759AJ420123Schizostephanus alatus K Schum Cult MSTR AJ419758-Schizozygia coffaeoides Baill Sennblad 207 UPS AJ419760-Secamone afzelii (Schult) K Schum Sennblad and Bremer 1996 X91779AJ420124
394 SYSTEMATIC BIOLOGY VOL 51
TABLE 2 Continued
Taxon Vouchersource EMBL number rbcLndhF
Secamone geayi Constatin et Gallaud Civeyrel 1200 LC AJ419761-Stapelia leendertziae NEBr Sennblad and Bremer 1996 X91778-Stephanotis oribunda Brongn Sennblad 256 UPS AJ419762AJ420125Stephanostema stenocarpum K Schum Sennblad and Bremer 1996 X91770AJ420126Strophanthus eminii Pax Sennblad and Bremer 1996 X91771AJ420127Tabernaemontana divaricata Roem et Schult Sennblad and Bremer 1996 X91772-Tabernanthe iboga Baill Leeuwenberg 12544 WAG AJ419763-Tacazzea apiculata Oliver Venter 9188 cult MSTR AJ419764-Thevetia peruviana (Pers) K Schum Sennblad and Bremer 1996 X91773-Trachelospermum jasminoides (Lindl) Lem Sennblad et al 1998 AJ002890AJ420128Tweedia coerulea Sweet Sennblad 254 UPS AJ419765-Tylophora sylvatica Decne Sennblad and Bremer 1996 X91789-Vahadenia caillei (AChev) Hutch et Dalziel Leeuwenberg 12275 WAG AJ419766-Vallesia antillana Woodson Meagher 966 FTG AJ419767-Vinca minor L Sennblad 230 UPS AJ419768-Vincetoxicum hirundinaria Medik Sennblad 257 UPS AJ419769-Wrightia arborea (Dennst) Mabb Sennblad et al 1998 AJ002891AJ420129
aVoucher for rbcL sequence onlybHerbarium abbreviations are according to Index Herbariorum except LC D Private herbarium of Laure CiveyrelcThis specimen was originally determined as Cibirhiza albersiana in Kunze et al (1994) but has recently been redetermined as
Fockea multiora K Schum (Verhoeven et al 2002)
of the clades in the successive weightinganalysis Clades with a successive weightedbootstrap gt63 were considered well sup-ported those with a successive weightedbootstrap gt85 were considered stronglysupported In a theoretical context a cladesupported by one uncontradicted charactercorresponds to a bootstrap support of 63anda cladesupported by two uncontradictedcharacters to 85 (Harshman 1994 Farriset al 1996) Here however this scale merelyrelates to the support distribution within theanalysis
RESULTS
The unit weight analysis of the completerbcL data set (237 characters total 214 ofthe cells were scored as missing data) re-sulted in 19003 trees 828 steps long witha consistency index (ci) D 0378 and a re-tention index (ri) D 0678 The consensustree from the successive weighting analysisof 252 most-parsimonious trees (each 1773steps long ci D 0661 ri D 0880) is presentedin Figure 1 The length of these trees withunit weight characters is 830 steps thus thesetrees are not identical to any of the most-parsimonious trees from the unit weightanalysis Branches not present in the strictconsensus tree from the unit weight analysisare indicated on the combinable consensustree from the successive weighting analysis
(Fig 1) The unit weight analysis of the com-bined set of the rbcL and ndhF data (153 char-acters total 043 of the cells were scored asmissing data) resulted in four trees 345 stepslong (ci D 0539 ri D 0482) The successiveweighting analysis gave three trees 978 stepslong (ci D 0840 ri D 0849) The strict consen-sus tree of the successive weighting analysisis presented in Figure 2 These trees are onestep longer with unit weights than the treesfrom the unit weight analysis Unless other-wise stated the following discussion refersto the successive weighting analyses of thecomplete rbcL data set
Investigating the justication of a prioriweighting (Sennblad and Bremer 2000) wemade some additional analyses of our rbcLdata set including tests for base compositionbias and rate heterogeneity This did not pro-duce signicantly different results and allwell-supported groups were congruent withthose of the present study The general struc-ture and the well-supported groups of theresulting trees (Figs 1 and 2) are to a largedegree congruent with the molecular stud-ies of Sennblad and Bremer (1996) whichare based on rbcL data of Civeyrel et al(1998) based mainly on matK sequence dataand of Potgieter and Albert (2001) basedmainly on trnLndashF data Although there aredifferences in the weakly supported rela-tionships among the well-supported groupsin particular the exact position of the
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 395
FIGURE 1 Combinable component consensus tree of the 252 most-parsimonious trees from the successiveweighting analysis of the complete rbcL data set Subfamilial and tribal classication is according to Endress andBruyns (2000) For subfamilies a three-letter code is used RAU APO PER SEC and ASC denote RauvoloideaeApocynoideae Periplocoideae Secamonoideae and Asclepiadoideae respectively OUT denotes the outgroupNumbers below branches are unit-weighted bootstrap values successive weighted bootstrap values are indicatedabove branches Branches not present in the strict consensus tree from the unit weight analysis are indicated witha cross (dagger) For branches marked A or B the frequencies in the combinable component consensus are 95 or 50respectively all other branches have 100 frequency
396 SYSTEMATIC BIOLOGY VOL 51
FIGURE 2 Strict consensus tree from the successive weighting analysis of the combined ndhF and rbcL dataset Tribal classication sensu Endress and Bruyns (2000) is indicated Numbers below branches are unit-weightedbootstrap values whereas successive weighted bootstrap values are indicated above branches Branches not presentin the strict consensus tree from the unit weight analysis are indicated with a cross (dagger)
Periplocoideae and of the root of the in-group all well-supported groups in our anal-ysis are congruent with those studies Thecombined analysis using data from rbcL-and matK-sequences and oral and pollenmorphology in the study of the Apocy-naceae sstr by Endress et al (1996) andthe analysis of Sennblad et al (1998) ofthe tribe Wrightieae sensu Leeuwenberg(1994a) and related taxa using rbcL andmorphological data are largely congruentwith the result from the present study Fur-thermore preliminary results from collab-orative studies by Sennblad et al (aimedat the traditional Apocynoideae and As-clepiadaceae pers comm) and Endresset al (aimed at the traditional Alyxieaepers comm) based on combined analy-sis of several molecular and morphologi-cal datasets are also congruent with ourresults
Support is good for the split betweenthe ingroup and the two outgroup taxaGelsemium and Mostuea (uwboot D 96swboot D 100) Thus the initial assump-tion of monophyly of the Apocynaceaesl is not violated Conforming to ear-lier studies (eg Sennblad and Bremer1996 Civeyrel et al 1998 Sennblad et al1998 Potgieter and Albert 2001) the tra-ditional Asclepiadaceae are nested withinthe traditional subfamily Apocynoideae(eg sensu Endress and Bruyns 2000)
The ApocynoideaendashAsclepiadaceae cladewhich is strongly supported (uwboot D 86swboot D 99) is in turn nested within thesubfamily Rauvoloideae sensu Endress andBruyns (2000)
Comparison of our results to the mostrecent classications of the ApocynaceaeEndress and Bruyns (2000 see Table 1 andFig 1) shows that although their circum-scription of the subfamilies PeriplocoideaeSecamonoideae and Asclepiadoideae aremonophyletic in our analysis their sub-families Rauvoloideae and Apocynoideaeare nonmonophyletic Similarly the tribesof the Rauvoloideae except Melodineaeare monophyletic but three of the tribesof the Apocynoideaemdashthe Apocyneae theEchiteae and the Wrightieaemdashare nonmono-phyletic Of the tribes of the Asclepiadoideaethe Ceropegieae and Asclepiadeae are mono-phyletic The position of Fockea as sisterto the rest of the Asclepiadoideae ren-ders the Marsdenieae nonmonophyletic Allthe included Periplocoideae taxa belongto the tribe Periploceae sensu Venter andVerhoeven (1997) except Pentopetia whichthey placed in the tribe CryptolepideaeGiven this unresolved position of Pentopetiathe monophyly of these tribes cannot be eval-uated however the reduction of Parquetinato synonymy with Periploca proposed byVenter and Verhoeven (1997) is not sup-ported by the present study
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 397
DISCUSSION
One of the most important tasks for anomenclature system is to provide a sta-ble means to communicate classications(which we here will understand as represent-ing phylogenetic relationships) Althoughboth the Linnaean and the Phylocode sys-tems are able to communicate phylogeniesthey are sensitive to changes in phylogenetichypotheses which in both systems will de-mand taxonomic changes In addition theLinnaean system is subject to a potentialinstability relating to subjective changes incircumscription that is lumpers and split-ters may refer to different groups of organ-isms by the same name (see de Queiroz andGauthier 1992) For many taxa a consensuscircumscription has eventually been estab-lished but for other groups this is still a prob-lem A current example of this from Apoc-ynaceae is the genus Tabernaemontana sensuLeeuwenberg (1991 1994b) which in theclassication of Allorge (1985) is split intoeight genera The Phylocode system was de-signed to avoid this problem by use of strictdenitions of taxa subjective changes incircumscription are avoided
However this rigidity in denitions leadsto a potential instability and high turnoverof names associated with a group of nestedtaxa Relatively small changes in phyloge-netic views eg when a dening taxonldquospecierrdquo receives a less nested positionmaycause some(or all) of these nested namesto become synonymous (eg de Queirozand Gauthier 1994 Bryant 1996 Liden andOxelman 1996Sereno 1999)The junior syn-onyms should according to the Phylocodebe rejected and new names may then beneeded for the new system of nested clades(de Queiroz and Gauthier 1994) The namesassociated with this group of taxa are thusreplaced by new names This can be espe-cially problematic when careless denitionshave been made as have been noted bySchander and Thollesson (1995) and Cantinoet al (1997) among others Consider a hy-pothetical worst-case example The Ascle-piadaceae were traditionally considered asseparate from Apocynaceae A stem-baseddenition of Apocynaceae reecting thisview could be the largest clade that in-cludes Apocynum but not Asclepias On thepresent tree such a denition would limitthe Apocynaceae to be synonymous with the
tribe Apocyneae and a new name wouldbe needed for the taxon corresponding tothe current view of Apocynaceae (eg sensuEndress and Bruyns 2000) A node-baseddenition would handle this particular prob-lem but may have other problems as illus-trated here with another hypothetical worst-case scenario Early classications of theApocynaceae included Plocosperma BenthA node-based denition representing thisview could have included as reference taxasay Apocynum and Plocosperma The recentlyproposed position of Plocosperma close tothe Boraginaceae (Backlund et al 2000)would with this denition have made Apoc-ynaceae a very large taxon (containing atleast both Gentianales and Boraginales) pos-sibly synonymous with (a hypothetical def-inition of) the informal taxon EuasteridsII (Bremer et al 1998) and again a newname for the Apocynaceae sensu Endressand Bruyns (2000 ie excluding Plocosperma)would have been needed Because theLinnaean system allows recircumscription ofsynonymous names the potential turnoverin the set of names in use is reduced withthis system (albeit at the expense of allowingsubjectivity) Additionally the ranks of theLinnaean system provide a set of names thatwork as relatively stable universal standardnames in communication note that equalranks does not guarantee comparable evo-lutionary units however Continuity in us-age of names is important for example injournals literature databases oras teach-ing and inventory work and the need forsuch continuity has been recognized by pro-ponents of both nomenclature systems (deQueiroz and Gauthier 1992 Greuter et al1994 Reveal 1996)
We propose a compromise approach thatuses solutions relating to these aspects fromboth systems concentrating mainly onbotanical classication of extant plant taxaat the tribal and familial levels and onlyconsidering monophyletic taxa We recom-mend a system that uses ldquostandard namesrdquo(ie universally used communication unitsat convenient hierarchical levels) such asthose provided by the principal ranks of theLinnaean system (Greuter et al 1994) but inwhich sensitivity toward subjective changesin circumscription is reduced One way to dothis is to adopt the denitions of the phy-logenetic system but use constraints similar
398 SYSTEMATIC BIOLOGY VOL 51
to those that apply between ranks in theLinnaean system To reduce sensitivity to-ward changes in phylogenetic hypotheses(Schander and Thollesson 1995 Bryant1996 1997 Cantino et al 1997 Moore 1998)we propose using combined node C stem-based denitions that will provoke explicitincompatibilities between taxa in case ofunfortunate changes in circumscription (deQueiroz 1996 has proposed a similar type ofdenition for designating nonmonophyletictaxa somewhat similar variants are alsodiscussed in the Phylocode) A node-basedpart of the denition provides a minimalcircumscription whereas a stem-based partprovides boundaries to competing taxa(maximal circumscription) If on a particu-lar phylogeny the minimal (monophyletic)circumscriptions of competing taxa are over-lapping then the two taxa are incompatibleTo make this meaningful at least two taxashould be included in the node-based partof the denition this will also avoid unnec-essary monotypic taxa This incompatibilitywill provide an objective point when nomen-claturalreconsideration isneeded Returningto the rst of our examples above node Cstem-based denitions of the Apocynaceaeand Asclepiadaceae could be as follows
The family Apocynaceae is the most inclusive clade inthe order Gentianales including the type specimensof Apocynum and Dictyophleba but not the taxa Ascle-piadaceae [ Gentianaceae Rubiaceae ]
and
The family Asclepiadaceae is the most inclusive cladein the order Gentianales including the type speci-mens of Asclepias and Periploca but not the taxa Apoc-ynaceae [ Gentianaceae Rubiaceae ]
On the tree in Figure 1 these two denitionsare mutually exclusive and a decision onwhich name to keep must be made Mini-mizing the number of classicatory changesshould take precedence in this decision Thereason for this is to avoid the large incon-venient ldquostandard namerdquo changes that couldresult from relatively small rearrangementsRejected names are then ignored when oc-curring in denitions of other taxa This isbecause the reference taxa in the stem-basedpart of the denitions are ldquodenedrdquo taxarather than physical types (thus reliance ondened reference taxa which has been crit-icized for stem-based denitions by Sereno1999 is what we aim for on the other hand
the explicit reference of type specimens in theminimum circumscription provided by thenode-based part of the denitions circum-vents the tautology problem raised byBryant 1996) In the current example thiswould mean that regardless of which nameof Apocynaceae and Asclepiadaceae is re-jected the other will be dened as corre-sponding to the Apocynaceae sensu Endressand Bruyns (2000) The choice will there-fore simply be which name to keep In thiscase because Apocynaceae has priority un-der the Linnaean system we would chooseto keep Apocynaceae to promote consistencywith the current system Nevertheless out-side these recommendations choices of whatnames to choose will probably to a certaindegree be arbitrary We also propose thatthe rejected name should be reinstated withan emended denition if appropriate to po-tentially allow for more stability in the setof names in use Thus in our second ex-ample above a node C stem-based deni-tion reecting the inclusion of Plocospermain Apocynaceae (eg by exchanging Dic-tyophleba for Plocosperma in the denitionabove) would presumably be incongruentwith the denition of Gentianales as well asthe denitions of Rubiaceae Gentianaceaeand so forth and therefore would be rejectedIn such acasewhere the only real change is inthe position of a single taxon a reinstatementof the Apocynaceae with an emended def-inition is motivated However a drawbackis that this may also introduce elements ofsubjectivity in the circumscriptions of taxa
Not all named clades need to be providedwith standard names In fact for the purposeof a universal set used in journals databasesand so forth it may sufce to use standardnames corresponding to the principal ranksof the Linnaean system (Greuter et al 1994)such as species genus family and orderNevertheless to provide compatibility withthe present system we suggest that addi-tional names corresponding to commonlyused secondary ranks (eg tribes) may berecognized The hierarchical level of a stan-dard name needs to be indicated This doesnot imply that taxa of the same hierarchi-cal level are comparable evolutionary unitsbut simply indicates the hierarchical levelof a taxon relative to nested taxa Becausesuch misinterpretations will in any case bedifcult to prohibit one may as well pro-vide compatibility with the present Linnaean
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 399
system by using the names of the Linnaeanranks (eg species genus tribe familyorder) to indicate hierarchical level The prin-ciple of exhaustive subsidiary taxa is notaccepted so redundant taxa need not be rec-ognized Other ldquononstandardrdquo taxa couldbe dened by using Phylocode denitionsmdashthat is without the restraints discussed forstandard taxa above
One advantage of the described systemis that utilizing the correspondence betweenthe standard names and Linnaean ranks al-lows the system to be largely compatiblewiththe establishedLinnaean system A change ofsystem could therefore be gradual with thetwo systems coexisting during a conversiontime (see eg de Quieroz 1997b)
Even though the node C stem-baseddenitions reduce the sensitivity towardschanges in circumscriptions they may still besensitive towardldquounfortunaterdquo or ldquobadrdquo def-initions (see eg Cantino et al 1997Cantinoand de Queiroz 2000) Many of the recom-mendations for such things as choices of ref-erence taxa (speciers) in phylogenetic def-initions are applicable also to the node Cstem-based denitions (eg Schander andThollesson 1995 Bryant 1996 Cantino et al1997 Sereno 1999) Most likely the compro-mise system described above will have sev-eral further problems and we hope that thispaper will invite further discussion on thesubject
A New Classication of the Apocynaceae slOur results indicate that some problem-
atic taxa remain in the classication ofEndress and Bruyns (2000) Monophyly fortwo of the subfamilies and ve of the tribesis called into question If other recent molec-ular studies (eg Potgieter and Albert 2001)are taken into consideration a further fourtribes (Alstonieae Alyxieae Plumerieae andVinceae) may be nonmonophyletic This canin some cases be simply analysis artifactsbut as Endress and Bruyns (2000) themselvespoint out their classication is to be consid-ered preliminary Insufcient taxonsamplingin molecular studies and difcult homologydecisions relating to morphological charac-ters make the circumscription of many ofthe tribes for example in the Apocynoideaesensu Endress and Bruyns (2000) uncertainThere is thus risk for future rearrangementsin many of the tribes A classication using
node C stem-based denitions may be ableto subsume such rearrangements withoutnumerous explicit recircumscriptions Wewill here use the result fromthe present studyas a basis to propose such a classication ofthe Apocynaceae sl
Our aim is to base taxa on clades that arewell supported A second aim is to makeour classication compatible with the cur-rent system We will therefore primarily rec-ognize nonoverlapping tribes that are validunder the Linnaean system We will furtheradopt the principle of nested referencing (assuggested by Lee 1999 and Sereno 1999)from the Linnaean system Thus the rst ref-erence taxon in the node-based part of thedenition is the primary type which corre-sponds to the Linnaean type used for thename of the taxon Instead of the (optional)enumeration of subsumed taxa of a lowerrank (eg genera or subtribes) we will in-clude node C stem-based denitions as dis-cussed above Citations to primary types ofreference taxa in node-based part of de-nitions are also nested and thus refer toldquothe type specimen of the type species ofthe type genus of the dened taxonrdquo Aproblem is that the relationships between thetribes areweakly supported This may lead tovery cumbersome denitions enumeratingall competing tribes in the stem-based partof the denition (eg Moore 1998) One wayto reduce this problem is to dene interme-diate well-supported taxa and use them ascompeting taxaWe therefore will further rec-ognize four nonstandard taxa (note that anassignment of ldquononstandardrdquo taxa does notimply less reliability standard names relateto communication purposes only) The taxadiscussed are indicated in Figure 3 We willdiscuss the tree from the top of Figure 3 start-ing with the taxa of the traditional Plumeri-oideae Tribes and subtribes given in paren-theses in the text refer to the classicationsof Endress and Bruyns (2000) in other casesour classication is implied For some tribeswe have chosen a name with priority underLinnaean system that implies inclusion of agenus not sampled in the present study Forall but one (Willughbeieae) of these casesadditional molecular or morphological phy-logenetic analyses support the inclusion ofthe genus Even if too uncertain denitionsshould be avoided this is less critical withnode C stem-based denitions In the caseof Willughbeieae we have decided from
400 SYSTEMATIC BIOLOGY VOL 51
FIGURE 3 The tribal reclassication discussed in the text indicated on the combinable component consensustree from the successive weighting analysis of the complete rbcL data set Four nonstandard taxamdashApocynoidinaEuapocynoidina Asclepiadacina and Asclepiadoidinamdashare also indicated
discussions with a morphological expert(M Endress pers comm) that morphologi-cal support exists for including WillughbeiaRoxb in Willughbeieae If this is correctthe denition given below will be valid if
it is not our system allows redenition ofthe taxon We have also tried to providecomments on potential morphological char-acters taken from literature for the differ-ent taxa recognized However we have not
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 401
performed any morphological analysis butrely on information external to this studyfor these comments In most cases we cantherefore not identify synapomorphies forthe taxa
TAXONOMIC TREATMENT
Tribus Willughbeieae ADCThe Ancylobotrys Dictyophleba and Va-
hadenia clade is strongly supported (uwbootand swboot D 100) and corresponds to theWillughbeieae of Endress and Bruyns (2000)The present sampling contains closely re-lated genera that have been traditionallykept together Landolphia P Beauv sl in-cludes all these taxa (eg Schumann 1895)In addition to the taxa sampled here thestudy of Potgieter and Albert (2001) sup-ports the inclusion of Couma Aubl LacmelleaH Karst Pacouria Aubl and Saba (Pichon)Pichon in this clade Plants in this cladeall have a gynoecium that is congenitallysyncarpous and indehiscent and that con-tains a eshy pulp including numerous seedswith copious horny endosperm (Fallen 1986Persoon et al 1992) This probably con-stitutes synapomorphies for this clade(M Endress pers comm) although not un-ambiguously for example syncarpy is par-alleled in Carisseae We will therefore as-sume that the reference genus Willughbeia isincluded in this tribe
DenitionmdashTribe Willughbeieae is themost inclusive clade that includes the prim-ary types of Willughbeia and Dictyophleba butnot Alyxieae Aspidospermeae CarisseaeHunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Apocynoidina
Tribus Vinceae Bartl
Ochrosia Rauvola Vinca and Catharanthus(all Vinceae) form a strongly supported clade(uwboot D 89 swboot D 99) The associa-tion between Rauvola and Catharanthus wasindicated by Sennblad and Bremer (1996)The study of Potgieter and Albert (2001) sup-ports inclusion of the genera LaxoplumeriaMarkgr Neisosperma Raf and Tonduzia Pit-tier previously classied with AspidospermaThe Vinceae contain both herbaceous andfruticose taxa The fruits are apocarpousbut whereas the fruits of Catharanthus Lax-oplumeria Neisosperma Tonduzia and Vincahave dry fruit walls the fruits of Ochrosia andRauvola are drupes Plants in Vinceae share
a similar type of style head (ie the apicalenlarged part of the style) with a stigmatichollow and upper and lower hair wreaths(except in some species of Ochrosia) the pres-ence of a nectar disk and a linear hilum onthe seed (Pichon 1948b) On the present treeVinceae also include Kopsia however this po-sition of Kopsia is very weakly supported (seealso Sennblad and Bremer 1996)
DenitionmdashTribe Vinceae is the mostinclusive clade that includes the primarytypes of Vinca and Rauvola but not AlyxieaeAspidospermeae Carisseae HunterieaeMelodineae Plumerieae Tabernaemon-taneae Willughbeieae or Apocynoidina
Tribus Tabernaemontaneae G DonThe strongly supported association be-
tween Tabernaemontana Tabernanthe Carval-hoa Schizozygia and Molongum (Tabernae-montaneae uwboot D 84 swboot D 98)has been suggested previously (Boiteau et al1978 Fallen 1986 Endress et al 1996Sennblad and Bremer 1996) Inclusion of taxatraditionally associated with Tabernaemon-tana (eg Voacanga Thouars and CallichiliaStapf see also Tabernaemontaneae sensuLeeuwenberg 1994a) and Macoubea Aublis supported by the analysis of Potgieterand Albert (2001) The taxa share sclerenchy-matic anthers free from the style head anda characteristic band of heavily cutinizedcells at the insertion on the staminal rib(Endress et al 1996) The taxa tradition-ally associated with Tabernaemontana arecharacterized by their apocarpous fruitswith arillate seeds whereas genera asso-ciated with Molongum (Ambelanieae sensuLeeuwenberg 1994a) have syncarpous fruitsand lackanarillus HoweverMacoubea formsa link between the two having a syncar-pous fruit with arillate seeds (Zarucchi et al1995)
DenitionmdashTribe Tabernaemontaneae isthe most inclusive clade that includes the pri-mary types of Tabernaemontana and Schizozy-gia but not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae PlumerieaeVinceae Willughbeieae or Apocynoidina
Tribus Aspidospermeae Miers
Aspidosperma and Vallesia (Aspidosper-meae) form a strongly supported clade (uw-boot and swboot D 100) rst suggestedin Sennblad (1997 see also Sennblad and
402 SYSTEMATIC BIOLOGY VOL 51
Bremer 2000) Subsequently GeissospermumAllemao Haplophyton ADC MicroplumeriaBaill and Strempeliopsis Benth were indi-cated to belong to this clade (Potgieter andAlbert 2001) The Aspidospermeae havesimple style heads lacking both stigmatichollow and upper hair wreath and haveapocarpous fruits with variation in eshi-ness and sclerication Alstonia has tradi-tionally been placed in the Plumerieae Be-cause its seeds have a hairy margin it hasbeen suggested to form a link to subfam-ily Apocynoideae This was contradicted bySennblad and Bremerrsquos study (1996) how-ever where Alstonia was placed in an iso-lated position as the sister group to therest of the Apocynaceae sl Here it groupswith Aspidospermeae Sensu Endress andBruyns (2000) However this association isvery weakly supported (uwboot and swbootlt 50) and in other recent studies As-pidospermeae and Alstonia do not form aclade (Potgieter and Albert 2001) Becauseof this we do not use it as a reference taxonin the denition Thus although the Aspi-dospermeae include Alstonia on the presenttree some other position of Alstonia canbe accommodated without amending thedenition
DenitionmdashTribe Aspidospermeae is themost inclusive clade that includes theprimary types of Aspidosperma and Vallesiabut not Alyxieae Carisseae HunterieaeMelodineae Plumerieae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Melodineae G DonThe well-supported association between
Melodinus and Craspidospermum (uwboot D70 swboot D 75) was rst reportedby Sennblad (1997 see also Sennblad andBremer 2000) and has subsequently beenreected in the tribe Melodineae sensuEndress and Bruyns (2000 their inclusion ofDiplorhynchus however is not supported)Craspidospermum although its fruit is dryand dehiscent as opposed to the indehis-cent fruits of Melodinus was noted by Pichon(1948b) to present several characters of theCarisseae such as stamens inserted near thebase of the corolla tube and a syncarpousovary In particular Pichon found the verydense suprastaminal indumentum type rem-iniscent of Melodinus From his descriptions
of the two genera (Pichon 1948ab) the fol-lowing similarities also emerge presence ofa stipular line pollen in tetrads and a punc-tiform hilum
DenitionmdashTribe Melodineae is the mostinclusive clade that includes the pri-mary types of Melodinus and Craspidosper-mum but not Alyxieae AspidospermeaeCarisseae Hunterieae Plumerieae Taber-naemontaneae Vinceae Willughbeieae orApocynoidina
Tribus Hunterieae K Schum
The Picralima and Pleiocarpa clade isstrongly supported (uwboot and swboot D100) and corresponds to the subtribeHunterieae sensu Endress andBruyns (2000)The study of Potgieter and Albert (2001)also supports the inclusion of Hunteria RoxbThese taxa are characterized by apocar-pous sometimes pluricarpous ovaries andeshy brous fruit walls (Omino 1996)Diplorhynchus (Melodineae) is here weaklyassociated (uwboot and swboot lt 50) withthe Pleiocarpeae It shares a few similari-ties such as stipular lines and a style headwithout a stigmatic hollow and hair wreathsbut also has differences such as a dry de-hiscent fruit With the present denitionDiplorhynchus will tentatively be included inthe Hunterieae
DenitionmdashTribe Hunterieae is the mostinclusive clade that includes the primarytypes of Hunteria and Picralima but not Alyx-ieae Aspidospermeae Carisseae Melod-ineae Plumerieae TabernaemontaneaeVinceae Willughbeieae or Apocynoidina
Tribus Plumerieae EndlThe strongly supported relationship
(uwboot and swboot D 100) between Alla-manda and Plumeria corresponds to earlierstudies (Endress et al 1996 Sennblad andBremer 1996 Civeyrel et al 1998 Potgieterand Albert 2001) In all earlier treatmentsAllamanda has had uncertain relationshipsThe association with Plumeria nds supportin pollen morphology such as perforatemesocolpial depressions and similar innerexine pattern The two genera also containsecoiridoids rather than the indole alka-loids and cardenolides that are commonin the traditional Plumerioideae (Endresset al 1996) Cameraria Cerbera and Thevetiaform a strongly supported association
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 403
(uwboot D 83 swboot D 91) with Ane-chites which has been suggested by Fallen(1983) This association was based mainlyon the latrorsesublatrorse anthers andthe broad style head with large apicalappendages and stigmatic hollow The studyby Potgieter and Albert (2001) suggesteda further inclusion of Cerberiopsis Viell ExPancher amp Sebert and Skytanthus Meyen inthis tribe These taxa form a well-supportedclade with Plumeria and Allamanda whichcorresponds to the Plumerieae sensuEndress and Bruyns (2000) Characterssupporting this relationship are the presenceof infrastaminal as well as suprastaminalappendages (not present in Anechites andPlumeria) and winged seeds (not present inAnechites) (Endress et al 1996)
DenitionmdashTribe Plumerieae is themost inclusive clade that includes theprimary types of Plumeria and Allamandabut not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Carisseae Dumort
Acokanthera and Carissa (both Carisseae)are traditionally joined and are sometimeseven treated as one genus (eg Pichon1948a) In the present study they form astrongly supported clade (uwboot D 96swboot D 99) Among the characters sup-porting this tribe are syncarpous fruits with-out axile placentation and endocracks on theinner exine of the pollen (Endress et al 1996)
DenitionmdashTribe Carisseae is the most in-clusive clade that includes the primary typesof Carissa and Acokanthera but not AlyxieaeAspidospermeae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
Tribus Alyxieae G Don
Chilocarpus has been a genus with un-certain afnities In Endress et al (1996) itwas the sister taxon to the ApocynoideaeHere it is associated with Alyxia and Lep-inia in a clade corresponding to the Alyxieaesensu Endress and Bruyns (2000) Thisclade is strongly supported with succes-sive weighted bootstrap (swboot D 91) butnot with unit-weighted bootstrap support(uwboot D 58) such clades will henceforthbe termed moderatelywell supported clades
A further inclusion of Condylocarpon DesfLepiniopsis Valeton and Plectaneia Thouarswas indicated by Potgieter and Albert (2001)The constituent taxa are the only Apocy-naceae taxa outside the Apocynoidina thathave porate pollen
DenitionmdashTribe Alyxieae is the most in-clusive clade that includes the primary typesof Alyxia and Lepinia but not Aspidosper-meae Carisseae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
APOCYNOIDINA
On the present tree the traditionalsubfam-ily Apocynoideae includes all subfamilies(Periplocoideae Secamonoideae and Ascle-piadoideae) of the Asclepiadaceae sensuLiede and Albers (1994) All of these taxaare established names and selecting one asa subfamily at the expense of the othersmight be considered unfortunate from dif-ferent points of view We have here chosennot to recognize subfamilies instead we willdene the traditional Apocynoideae Ascle-piadaceae and Asclepiadoideae as nonstan-dard taxa The correspondence to the tradi-tional taxa will be indicated in the names byreplacing the sufx -eae with a neutral suf-x -ina (Kron 1997) We will also for practi-cal reasonsrecognize as a nonstandardtaxonthe informal group euapocynoids suggestedby Sennblad et al (1998)
The taxa of the traditional Apocynoideaeand Asclepiadaceae (uwboot D 86 swboot D99) clade will in many cases be more dif-cult to safely delimit to tribes becausemany of the groups in this clade have weakor no support (although preliminary resultsfrom an unpublished collaborative study bySennblad et al indicate additional supportfor the tribes discussed below) The tribaland subtribal classication of the traditionalApocynoideae has been shown to be prob-lematic (Leeuwenberg 1994a Endress andBruyns 2000) The circumscription of tribesfor these taxa might therefore in some casesbe preliminary
In all following taxa the basal part of theanther connective called the retinacle is ad-nate to the style head
DenitionmdashApocynoidina are the mostinclusive clade that includes the pri-mary types of Apocynum and Wrightia butnot Alyxieae Aspidospermeae Carisseae
404 SYSTEMATIC BIOLOGY VOL 51
Hunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Willughbeieae
Tribus Wrightieae G DonThe Stephanostema and Wrightia (both
Wrightieae) clade is strongly supported(uwboot and swboot D 100) and correspondsto the Wrightieae sensu Sennblad et al (1998)who also suggested an inclusion of Pleio-ceras Baill The tribe could be characterizedby a combination of synapomorphic and ple-siomorphic characters for example presenceofa chalazalandabsence of micropylar comaleft contorted aestivation and absence of airspaces in the anthers (1998)
DenitionmdashTribe Wrightieae is the mostinclusive clade in the Apocynoidina that in-cludes the primary types of Wrightia andStephanostema but not Malouetieae Nerieaeor Euapocynoidina
Tribus Nerieae (Benth) M Pichon
Adenium Nerium and Strophanthus (allWrightieae) constituted the Nerieae inSennblad et al (1998) however the cladereceived weak support Here Adenium andNerium form a clade that is absent in halfof the most-parsimonious trees in the com-plete rbcL analysis but is strongly supportedin the combined ndhF and rbcL analysis(uwboot D 98 swboot D 99) Strophan-thus groups with the Mascarenhasia Pachy-podium Kibatalia Funtumia and Holarrhenaclade which corresponds to the Malouetieaesensu Sennblad et al (1998) This position ofStrophanthus is however weakly supported(uwboot and swboot lt 50) and is not presentin the combined ndhF and rbcL tree Fur-thermore certain morphological characters(unfused slits in the corolla tube and anapical anther appendage) indicate a posi-tion closer to the Nerieae (Sennblad et al1998)
DenitionmdashTribe Nerieae is the most in-clusive clade in the Apocynoidina that in-cludes the primary types of Nerium and Ade-nium but not Malouetieae Wrightieae orEuapocynoidina
Tribus Malouetieae Mull-Arg
The association of Mascarenhasia Pachy-podium Kibatalia Funtumia and Holar-rhena (all Wrightieae) is moderately wellsupported in the complete rbcL analysis
(uwboot D 57 swboot D 74) whereasthe corresponding clade is well supportedin the combined rbcL and ndhF analysis(uwboot D 66 swboot D 100) On the ba-sis of identied synapomorphies (presenceof calcium oxalate crystals in the stomium ofthe anthers absence of interpetal vein ses-sile laments and absence of air spaces inthe anthers) Sennblad et al (1998) predictedan inclusion of Alaa Thouars Kibatalia andMalouetia ADC in the tribe The inclusion ofKibatalia is strongly supported in this studyFollowing these suggestions we will here as-sume the inclusion of the reference genusMalouetia in this tribe (The denition belowwill include Strophanthus on the present treebut see comments under Nerieae)
DenitionmdashTribe Malouetieae is the mostinclusive clade in the Apocynoidina that in-cluded the primary types of Malouetia andFuntumia but not Nerieae Wrightieae orEuapocynoidina
EUAPOCYNOIDINA
In all the remaining representatives ofthe Apocynoideae sensu Endress andBruyns (2000) that is Mandevilla MesechitesAganosma Apocynum TrachelospermumRhabdadenia Beaumontia Parsonsia andPrestonia the anthers are adnate to the stylehead both by the retinacle and by the thecaeThese taxa together with the taxa of the tra-ditional Asclepiadaceae form a moderatelysupported clade (uwboot lt 50 swboot D86) that corresponds to the informal taxoneuapocynoids sensu Sennblad et al (1998)
DenitionmdashEuapocynoidina are the mostinclusive clade that includes the primarytypes of Apocynum and Echites but notMalouetieae Wrightieae or Nerieae
Tribus Mesechiteae Miers
Mandevilla and Mesechites (both Mese-chiteae) form a moderately supported clade(uwboot lt 50 swboot D 87) equivalentto the Mesechiteae This tribe could be char-acterized by anthers with obtuse truncatedtails and a retinacle lacking hairs and also astrongly pentagonal style head with a stig-matic hollow
DenitionmdashTribe Mesechiteae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Mesechitesand Mandevilla but not Apocyneae EchiteaePeriploceae or Asclepiadacina
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 405
Tribus Apocyneae Bercht et J Presl
Aganosma Apocynum Beaumontia andTrachelospermum (Apocyneae) form a weaklysupported clade with Rhabdadenia (Echiteae)(uwboot and swboot lt 50) in the completerbcL analysis whereas the correspondingclade in the combined ndhF and rbcL analy-sis gains a little better support (uwboot lt 50swboot D 91) Likewise this clade (with theinclusion of Chonemorpha G Don) is weaklysupported in Potgieter and Albert (2001) Po-tential characterizing traits could be a stylehead without stigmatic hollow in combina-tion with thecae that are adnate to the stylehead (Sennblad et al 1998)
DenitionmdashTribe Apocyneae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Apocynumand Trachelospermum but not Echiteae Mese-chiteae Periploceae or Asclepiadacina
Tribus Echiteae Bartl
Parsonsia Peltastes and Prestonia (allEchiteae) form a moderately supported orunsupported clade with the Periplocoideaerepresentatives of the study (uwboot andswboot lt 50 and uwboot lt 50 swboot D89 in the complete rbcL and the combinedrbcL and ndhF analysis respectively) In thestudy of Sennblad et al (1998) Prestonia andParsonsia formed a weakly supported cladeand in the study of Potgieter and Albert(2001) inclusion of both Parsonsia and EchitesP Browne in the Echiteae is supportedParsonsia and Echiteae also share certaincharacters such as a horseshoe-shaped reti-nacle and a style head with stigmatic hollow(see Sennblad et al 1998)
DenitionmdashTribe Echiteae is the most in-clusive clade in the Euapocynoidina thatincludes the primary types of Echites andPrestonia but not Apocyneae MesechiteaePeriploceae or Asclepiadacina
Tribus Periploceae BartlThe representatives of the Periplo-
coideae sensu Endress and Bruyns (2000)mdashPentopetia Parquetina Petopentia PeriplocaTacazzea and Mondiamdashgroup together ina well-supported clade (uwboot D 67swboot D 93) The monophyly of the tra-ditional periplocoid taxa is also supportedin earlier studies that included a wider ordifferent sampling (Civeyrel et al 1998
Potgieter and Albert 2001) The taxa ofthe traditional Asclepiadaceae have ldquotrans-latorsrdquo structures related to pollinationspecialization In the Periploceae the pollenis deposited as tetrads in a spoon-liketranslator with an adhesive disc
DenitionmdashTribe Periploceae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Periplocaand Pentopetia but not Apocyneae EchiteaeMesechiteae or Asclepiadacina
ASCLEPIADACINA
The taxa of the Asclepiadoideae sensuLiede and Albers (1994) form a clade togetherwith the two Secamone species and Baissea(Apocyneae uwboot lt 50 swboot D 73)The exclusion of Periploceae (traditionallyassociated with this clade see Civeyrel et al1998) from this taxon is weakly supportedTherefore Periploceae is not included as areference taxon in the stem-based part ofthe denition of the Asclepiadacina thusleaving open the possibility for a later in-clusion of this taxon In the Asclepiadacina(except Baissea) the pollen is agglutinatedinto pollinia that are connected to a claspingtranslator
DenitionmdashAsclepiadacina are the mostinclusive monophyletic clade in the Euapoc-ynoidina that includes the primary types ofAsclepias and Secamoneae but not ApocyneaeEchiteae or Mesechiteae
Tribus Secamoneae G DonThe tribe Secamoneae sensu Bruyns
(uwboot D 89 swboot D 96) is mono-phyletic in the present analysisHowever be-cause representatives from only one of theconstituent genera are included monophylyis not well tested but the study of Civeyrelet al (1998) which included a larger numberof taxa including Pervillea Decne showedthat monophyly of the tribe was supportedThe traditional character for this tribe is fourpollinia per translator
DenitionmdashTribe Secamoneae is the mostinclusive clade in the Asclepiadacina that in-cludes the primary types of Secamone andPervillea but not Asclepioidina or Baissea
BaisseaThe position of Baissea corresponds with
that in the study of Sennblad et al (1998)
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 393
TABLE 2 Vouchers andEMBLaccession numbers for sequences published in this study Forpreviously publishedsequences and for sequences from previously published voucher specimens the reference is given
Taxon Vouchersourceb EMBL number rbcLndhF
Acokanthera oblongifolia (Hochst) Codd Sennblad and Bremer 1996 X91758-Adenium obesum (Forssk) Roem et Schult Sennblad et al 1998 AJ002880AJ420130Aganosma marginata (Roxb) G Don Ryding 490 UPS AJ419730-Allamanda cathartica L Sennblad and Bremer 1996 X91759-Alstonia scholaris RBr Sennblad and Bremer 1996 X91760-Alyxia ruscifolia RBr Sennblad 236 UPS AJ419731-Ancylobotrys petersiana Pierre Sennblad 201 UPS AJ419732-Anechites nerium Urb Bremer et al 3386 UPS AJ419733-Apocynum cannabinum L Sennblad and Bremer 1996 X91761AJ420113Araujia hortorum E Fourn Bremer 3006 UPS AJ419734-Asclepias curassavica L Sennblad and Bremer 1996 X91774-Aspidosperma triternatum Rojas Acosta Bremer 3029 UPS AJ419735-Baissea leonensis Benth Sennblad et al 1998 Yallah 112 UPSa AJ002881AJ420114Beaumontia grandiora Wall Sennblad et al 1998 AJ002882AJ420115Calotropis procera (Aiton) WT Aiton Sennblad 222 UPS AJ419736-Cameraria latifolia L Houghton 1085 FTG AJ419737-Carissa bispinosa (L) Merxm Sennblad 235 UPS AJ419738-Carvalhoa campanulata K Schum Sennblad 217 UPS AJ419739-Catharanthus roseus G Don Sennblad and Bremer 1996 X91757-Cerbera venenifera Steud No voucher AJ419740-Ceropegia woodii Schltr Sennblad and Bremer 1996 X91775-Chilocarpus suaveolens Blume Endress et al 1996 X92445-Fockea multiora K Schum Specks 248 cult MSTRc AJ419741-Craspidospermum verticillatum Bojer Pettersson and Nilsson 742 AJ419742-Dictyophleba lucida Pierre Sennblad and Bremer 1996 X91762-Diplorhynchus condylocarpon (Mull-Arg) Pichon Sennblad 203 UPS AJ419743-Fischeria stellata (Vell) E Fourn Clark and Watt 793 UPS AJ419744-Funtumia elastica Stapf Sennblad et al 1998 AJ002884-Gelsemium sempervirens Aiton Olmstead et al 1993 L14397-Holarrhena pubescens G Don Sennblad et al 1998 AJ002884-Hoya bella Hook Sennblad and Bremer 1996 X91776-Kibatalia gitingense (Elmer) Woodson Liede 3268 Z AJ419745-Kopsia fruticosa ADC Sennblad and Bremer 1996 X91763-Lepinia taitensis Decne Perlman 15071 NTBG AJ419746-Mandevilla sanderi (Hemsl) Woodson Sennblad and Bremer 1996 X91764AJ420116Mascarenhasia arborescens ADC Sennblad et al 1998 AJ002885AJ420117Matelea hirsuta (Vahl) Woodson Sennblad 263 UPS AJ419747-Melodinus monogynus Roxb Sennblad 261 UPS AJ419748-Mesechites trida Mull-Arg Bremer et al 3351 UPS AJ419749-Micholitzia obcordata NEBr Bremer 3010 UPS AJ419750-Molongum laxum (Benth) Pichon Sennblad and Bremer 1996 X91765-Mondia ecornuta (NEBr) Bullock Sennblad 215 UPS AJ419751-Mostuea brunonis Didr Olmstead et al 1993 L14404-Nerium oleander L Sennblad 265 UPS AJ002886AJ420118Ochrosia coccinea Miq v d Laan 389 WAG AJ419752-Orthosia serpyllifolia Decne Bremer et al 3372 UPS AJ419753-Pachypodium lamerei Drake Sennblad et al 1998 AJ002887AJ420119Parquetina nigrescens (Afzel) Bullock Sennblad and Bremer 1996 X91777-Parsonsia heterophylla A Cunn Sennblad et al 1998 AJ002888-Peltastes peltatum (Vell) Woodson Sennblad 262 UPS AJ419754-Pentopetia sp No voucher AJ419755-Periploca graeca L Sennblad et al 1998 AJ002889AJ420120Petopentia natalensis (Schltr) Bullock Sennblad 237a and sn cult MSTR AJ419756AJ420121Picralima nitida T et H Dur Sennblad and Bremer 1996 X91766-Pleiocarpa mutica Benth Bremer 3017 UPS AJ419757-Plumeria inodora Jacq Sennblad and Bremer 1996 X91767-Prestonia quinquangularis Spreng Sennblad and Bremer 1996 X91768AJ420122Rauvola mannii Stapf Sennblad and Bremer 1996 X91769-Rhabdadenia biora Mull-Arg Zona 616 FTG AJ419759AJ420123Schizostephanus alatus K Schum Cult MSTR AJ419758-Schizozygia coffaeoides Baill Sennblad 207 UPS AJ419760-Secamone afzelii (Schult) K Schum Sennblad and Bremer 1996 X91779AJ420124
394 SYSTEMATIC BIOLOGY VOL 51
TABLE 2 Continued
Taxon Vouchersource EMBL number rbcLndhF
Secamone geayi Constatin et Gallaud Civeyrel 1200 LC AJ419761-Stapelia leendertziae NEBr Sennblad and Bremer 1996 X91778-Stephanotis oribunda Brongn Sennblad 256 UPS AJ419762AJ420125Stephanostema stenocarpum K Schum Sennblad and Bremer 1996 X91770AJ420126Strophanthus eminii Pax Sennblad and Bremer 1996 X91771AJ420127Tabernaemontana divaricata Roem et Schult Sennblad and Bremer 1996 X91772-Tabernanthe iboga Baill Leeuwenberg 12544 WAG AJ419763-Tacazzea apiculata Oliver Venter 9188 cult MSTR AJ419764-Thevetia peruviana (Pers) K Schum Sennblad and Bremer 1996 X91773-Trachelospermum jasminoides (Lindl) Lem Sennblad et al 1998 AJ002890AJ420128Tweedia coerulea Sweet Sennblad 254 UPS AJ419765-Tylophora sylvatica Decne Sennblad and Bremer 1996 X91789-Vahadenia caillei (AChev) Hutch et Dalziel Leeuwenberg 12275 WAG AJ419766-Vallesia antillana Woodson Meagher 966 FTG AJ419767-Vinca minor L Sennblad 230 UPS AJ419768-Vincetoxicum hirundinaria Medik Sennblad 257 UPS AJ419769-Wrightia arborea (Dennst) Mabb Sennblad et al 1998 AJ002891AJ420129
aVoucher for rbcL sequence onlybHerbarium abbreviations are according to Index Herbariorum except LC D Private herbarium of Laure CiveyrelcThis specimen was originally determined as Cibirhiza albersiana in Kunze et al (1994) but has recently been redetermined as
Fockea multiora K Schum (Verhoeven et al 2002)
of the clades in the successive weightinganalysis Clades with a successive weightedbootstrap gt63 were considered well sup-ported those with a successive weightedbootstrap gt85 were considered stronglysupported In a theoretical context a cladesupported by one uncontradicted charactercorresponds to a bootstrap support of 63anda cladesupported by two uncontradictedcharacters to 85 (Harshman 1994 Farriset al 1996) Here however this scale merelyrelates to the support distribution within theanalysis
RESULTS
The unit weight analysis of the completerbcL data set (237 characters total 214 ofthe cells were scored as missing data) re-sulted in 19003 trees 828 steps long witha consistency index (ci) D 0378 and a re-tention index (ri) D 0678 The consensustree from the successive weighting analysisof 252 most-parsimonious trees (each 1773steps long ci D 0661 ri D 0880) is presentedin Figure 1 The length of these trees withunit weight characters is 830 steps thus thesetrees are not identical to any of the most-parsimonious trees from the unit weightanalysis Branches not present in the strictconsensus tree from the unit weight analysisare indicated on the combinable consensustree from the successive weighting analysis
(Fig 1) The unit weight analysis of the com-bined set of the rbcL and ndhF data (153 char-acters total 043 of the cells were scored asmissing data) resulted in four trees 345 stepslong (ci D 0539 ri D 0482) The successiveweighting analysis gave three trees 978 stepslong (ci D 0840 ri D 0849) The strict consen-sus tree of the successive weighting analysisis presented in Figure 2 These trees are onestep longer with unit weights than the treesfrom the unit weight analysis Unless other-wise stated the following discussion refersto the successive weighting analyses of thecomplete rbcL data set
Investigating the justication of a prioriweighting (Sennblad and Bremer 2000) wemade some additional analyses of our rbcLdata set including tests for base compositionbias and rate heterogeneity This did not pro-duce signicantly different results and allwell-supported groups were congruent withthose of the present study The general struc-ture and the well-supported groups of theresulting trees (Figs 1 and 2) are to a largedegree congruent with the molecular stud-ies of Sennblad and Bremer (1996) whichare based on rbcL data of Civeyrel et al(1998) based mainly on matK sequence dataand of Potgieter and Albert (2001) basedmainly on trnLndashF data Although there aredifferences in the weakly supported rela-tionships among the well-supported groupsin particular the exact position of the
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 395
FIGURE 1 Combinable component consensus tree of the 252 most-parsimonious trees from the successiveweighting analysis of the complete rbcL data set Subfamilial and tribal classication is according to Endress andBruyns (2000) For subfamilies a three-letter code is used RAU APO PER SEC and ASC denote RauvoloideaeApocynoideae Periplocoideae Secamonoideae and Asclepiadoideae respectively OUT denotes the outgroupNumbers below branches are unit-weighted bootstrap values successive weighted bootstrap values are indicatedabove branches Branches not present in the strict consensus tree from the unit weight analysis are indicated witha cross (dagger) For branches marked A or B the frequencies in the combinable component consensus are 95 or 50respectively all other branches have 100 frequency
396 SYSTEMATIC BIOLOGY VOL 51
FIGURE 2 Strict consensus tree from the successive weighting analysis of the combined ndhF and rbcL dataset Tribal classication sensu Endress and Bruyns (2000) is indicated Numbers below branches are unit-weightedbootstrap values whereas successive weighted bootstrap values are indicated above branches Branches not presentin the strict consensus tree from the unit weight analysis are indicated with a cross (dagger)
Periplocoideae and of the root of the in-group all well-supported groups in our anal-ysis are congruent with those studies Thecombined analysis using data from rbcL-and matK-sequences and oral and pollenmorphology in the study of the Apocy-naceae sstr by Endress et al (1996) andthe analysis of Sennblad et al (1998) ofthe tribe Wrightieae sensu Leeuwenberg(1994a) and related taxa using rbcL andmorphological data are largely congruentwith the result from the present study Fur-thermore preliminary results from collab-orative studies by Sennblad et al (aimedat the traditional Apocynoideae and As-clepiadaceae pers comm) and Endresset al (aimed at the traditional Alyxieaepers comm) based on combined analy-sis of several molecular and morphologi-cal datasets are also congruent with ourresults
Support is good for the split betweenthe ingroup and the two outgroup taxaGelsemium and Mostuea (uwboot D 96swboot D 100) Thus the initial assump-tion of monophyly of the Apocynaceaesl is not violated Conforming to ear-lier studies (eg Sennblad and Bremer1996 Civeyrel et al 1998 Sennblad et al1998 Potgieter and Albert 2001) the tra-ditional Asclepiadaceae are nested withinthe traditional subfamily Apocynoideae(eg sensu Endress and Bruyns 2000)
The ApocynoideaendashAsclepiadaceae cladewhich is strongly supported (uwboot D 86swboot D 99) is in turn nested within thesubfamily Rauvoloideae sensu Endress andBruyns (2000)
Comparison of our results to the mostrecent classications of the ApocynaceaeEndress and Bruyns (2000 see Table 1 andFig 1) shows that although their circum-scription of the subfamilies PeriplocoideaeSecamonoideae and Asclepiadoideae aremonophyletic in our analysis their sub-families Rauvoloideae and Apocynoideaeare nonmonophyletic Similarly the tribesof the Rauvoloideae except Melodineaeare monophyletic but three of the tribesof the Apocynoideaemdashthe Apocyneae theEchiteae and the Wrightieaemdashare nonmono-phyletic Of the tribes of the Asclepiadoideaethe Ceropegieae and Asclepiadeae are mono-phyletic The position of Fockea as sisterto the rest of the Asclepiadoideae ren-ders the Marsdenieae nonmonophyletic Allthe included Periplocoideae taxa belongto the tribe Periploceae sensu Venter andVerhoeven (1997) except Pentopetia whichthey placed in the tribe CryptolepideaeGiven this unresolved position of Pentopetiathe monophyly of these tribes cannot be eval-uated however the reduction of Parquetinato synonymy with Periploca proposed byVenter and Verhoeven (1997) is not sup-ported by the present study
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 397
DISCUSSION
One of the most important tasks for anomenclature system is to provide a sta-ble means to communicate classications(which we here will understand as represent-ing phylogenetic relationships) Althoughboth the Linnaean and the Phylocode sys-tems are able to communicate phylogeniesthey are sensitive to changes in phylogenetichypotheses which in both systems will de-mand taxonomic changes In addition theLinnaean system is subject to a potentialinstability relating to subjective changes incircumscription that is lumpers and split-ters may refer to different groups of organ-isms by the same name (see de Queiroz andGauthier 1992) For many taxa a consensuscircumscription has eventually been estab-lished but for other groups this is still a prob-lem A current example of this from Apoc-ynaceae is the genus Tabernaemontana sensuLeeuwenberg (1991 1994b) which in theclassication of Allorge (1985) is split intoeight genera The Phylocode system was de-signed to avoid this problem by use of strictdenitions of taxa subjective changes incircumscription are avoided
However this rigidity in denitions leadsto a potential instability and high turnoverof names associated with a group of nestedtaxa Relatively small changes in phyloge-netic views eg when a dening taxonldquospecierrdquo receives a less nested positionmaycause some(or all) of these nested namesto become synonymous (eg de Queirozand Gauthier 1994 Bryant 1996 Liden andOxelman 1996Sereno 1999)The junior syn-onyms should according to the Phylocodebe rejected and new names may then beneeded for the new system of nested clades(de Queiroz and Gauthier 1994) The namesassociated with this group of taxa are thusreplaced by new names This can be espe-cially problematic when careless denitionshave been made as have been noted bySchander and Thollesson (1995) and Cantinoet al (1997) among others Consider a hy-pothetical worst-case example The Ascle-piadaceae were traditionally considered asseparate from Apocynaceae A stem-baseddenition of Apocynaceae reecting thisview could be the largest clade that in-cludes Apocynum but not Asclepias On thepresent tree such a denition would limitthe Apocynaceae to be synonymous with the
tribe Apocyneae and a new name wouldbe needed for the taxon corresponding tothe current view of Apocynaceae (eg sensuEndress and Bruyns 2000) A node-baseddenition would handle this particular prob-lem but may have other problems as illus-trated here with another hypothetical worst-case scenario Early classications of theApocynaceae included Plocosperma BenthA node-based denition representing thisview could have included as reference taxasay Apocynum and Plocosperma The recentlyproposed position of Plocosperma close tothe Boraginaceae (Backlund et al 2000)would with this denition have made Apoc-ynaceae a very large taxon (containing atleast both Gentianales and Boraginales) pos-sibly synonymous with (a hypothetical def-inition of) the informal taxon EuasteridsII (Bremer et al 1998) and again a newname for the Apocynaceae sensu Endressand Bruyns (2000 ie excluding Plocosperma)would have been needed Because theLinnaean system allows recircumscription ofsynonymous names the potential turnoverin the set of names in use is reduced withthis system (albeit at the expense of allowingsubjectivity) Additionally the ranks of theLinnaean system provide a set of names thatwork as relatively stable universal standardnames in communication note that equalranks does not guarantee comparable evo-lutionary units however Continuity in us-age of names is important for example injournals literature databases oras teach-ing and inventory work and the need forsuch continuity has been recognized by pro-ponents of both nomenclature systems (deQueiroz and Gauthier 1992 Greuter et al1994 Reveal 1996)
We propose a compromise approach thatuses solutions relating to these aspects fromboth systems concentrating mainly onbotanical classication of extant plant taxaat the tribal and familial levels and onlyconsidering monophyletic taxa We recom-mend a system that uses ldquostandard namesrdquo(ie universally used communication unitsat convenient hierarchical levels) such asthose provided by the principal ranks of theLinnaean system (Greuter et al 1994) but inwhich sensitivity toward subjective changesin circumscription is reduced One way to dothis is to adopt the denitions of the phy-logenetic system but use constraints similar
398 SYSTEMATIC BIOLOGY VOL 51
to those that apply between ranks in theLinnaean system To reduce sensitivity to-ward changes in phylogenetic hypotheses(Schander and Thollesson 1995 Bryant1996 1997 Cantino et al 1997 Moore 1998)we propose using combined node C stem-based denitions that will provoke explicitincompatibilities between taxa in case ofunfortunate changes in circumscription (deQueiroz 1996 has proposed a similar type ofdenition for designating nonmonophyletictaxa somewhat similar variants are alsodiscussed in the Phylocode) A node-basedpart of the denition provides a minimalcircumscription whereas a stem-based partprovides boundaries to competing taxa(maximal circumscription) If on a particu-lar phylogeny the minimal (monophyletic)circumscriptions of competing taxa are over-lapping then the two taxa are incompatibleTo make this meaningful at least two taxashould be included in the node-based partof the denition this will also avoid unnec-essary monotypic taxa This incompatibilitywill provide an objective point when nomen-claturalreconsideration isneeded Returningto the rst of our examples above node Cstem-based denitions of the Apocynaceaeand Asclepiadaceae could be as follows
The family Apocynaceae is the most inclusive clade inthe order Gentianales including the type specimensof Apocynum and Dictyophleba but not the taxa Ascle-piadaceae [ Gentianaceae Rubiaceae ]
and
The family Asclepiadaceae is the most inclusive cladein the order Gentianales including the type speci-mens of Asclepias and Periploca but not the taxa Apoc-ynaceae [ Gentianaceae Rubiaceae ]
On the tree in Figure 1 these two denitionsare mutually exclusive and a decision onwhich name to keep must be made Mini-mizing the number of classicatory changesshould take precedence in this decision Thereason for this is to avoid the large incon-venient ldquostandard namerdquo changes that couldresult from relatively small rearrangementsRejected names are then ignored when oc-curring in denitions of other taxa This isbecause the reference taxa in the stem-basedpart of the denitions are ldquodenedrdquo taxarather than physical types (thus reliance ondened reference taxa which has been crit-icized for stem-based denitions by Sereno1999 is what we aim for on the other hand
the explicit reference of type specimens in theminimum circumscription provided by thenode-based part of the denitions circum-vents the tautology problem raised byBryant 1996) In the current example thiswould mean that regardless of which nameof Apocynaceae and Asclepiadaceae is re-jected the other will be dened as corre-sponding to the Apocynaceae sensu Endressand Bruyns (2000) The choice will there-fore simply be which name to keep In thiscase because Apocynaceae has priority un-der the Linnaean system we would chooseto keep Apocynaceae to promote consistencywith the current system Nevertheless out-side these recommendations choices of whatnames to choose will probably to a certaindegree be arbitrary We also propose thatthe rejected name should be reinstated withan emended denition if appropriate to po-tentially allow for more stability in the setof names in use Thus in our second ex-ample above a node C stem-based deni-tion reecting the inclusion of Plocospermain Apocynaceae (eg by exchanging Dic-tyophleba for Plocosperma in the denitionabove) would presumably be incongruentwith the denition of Gentianales as well asthe denitions of Rubiaceae Gentianaceaeand so forth and therefore would be rejectedIn such acasewhere the only real change is inthe position of a single taxon a reinstatementof the Apocynaceae with an emended def-inition is motivated However a drawbackis that this may also introduce elements ofsubjectivity in the circumscriptions of taxa
Not all named clades need to be providedwith standard names In fact for the purposeof a universal set used in journals databasesand so forth it may sufce to use standardnames corresponding to the principal ranksof the Linnaean system (Greuter et al 1994)such as species genus family and orderNevertheless to provide compatibility withthe present system we suggest that addi-tional names corresponding to commonlyused secondary ranks (eg tribes) may berecognized The hierarchical level of a stan-dard name needs to be indicated This doesnot imply that taxa of the same hierarchi-cal level are comparable evolutionary unitsbut simply indicates the hierarchical levelof a taxon relative to nested taxa Becausesuch misinterpretations will in any case bedifcult to prohibit one may as well pro-vide compatibility with the present Linnaean
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 399
system by using the names of the Linnaeanranks (eg species genus tribe familyorder) to indicate hierarchical level The prin-ciple of exhaustive subsidiary taxa is notaccepted so redundant taxa need not be rec-ognized Other ldquononstandardrdquo taxa couldbe dened by using Phylocode denitionsmdashthat is without the restraints discussed forstandard taxa above
One advantage of the described systemis that utilizing the correspondence betweenthe standard names and Linnaean ranks al-lows the system to be largely compatiblewiththe establishedLinnaean system A change ofsystem could therefore be gradual with thetwo systems coexisting during a conversiontime (see eg de Quieroz 1997b)
Even though the node C stem-baseddenitions reduce the sensitivity towardschanges in circumscriptions they may still besensitive towardldquounfortunaterdquo or ldquobadrdquo def-initions (see eg Cantino et al 1997Cantinoand de Queiroz 2000) Many of the recom-mendations for such things as choices of ref-erence taxa (speciers) in phylogenetic def-initions are applicable also to the node Cstem-based denitions (eg Schander andThollesson 1995 Bryant 1996 Cantino et al1997 Sereno 1999) Most likely the compro-mise system described above will have sev-eral further problems and we hope that thispaper will invite further discussion on thesubject
A New Classication of the Apocynaceae slOur results indicate that some problem-
atic taxa remain in the classication ofEndress and Bruyns (2000) Monophyly fortwo of the subfamilies and ve of the tribesis called into question If other recent molec-ular studies (eg Potgieter and Albert 2001)are taken into consideration a further fourtribes (Alstonieae Alyxieae Plumerieae andVinceae) may be nonmonophyletic This canin some cases be simply analysis artifactsbut as Endress and Bruyns (2000) themselvespoint out their classication is to be consid-ered preliminary Insufcient taxonsamplingin molecular studies and difcult homologydecisions relating to morphological charac-ters make the circumscription of many ofthe tribes for example in the Apocynoideaesensu Endress and Bruyns (2000) uncertainThere is thus risk for future rearrangementsin many of the tribes A classication using
node C stem-based denitions may be ableto subsume such rearrangements withoutnumerous explicit recircumscriptions Wewill here use the result fromthe present studyas a basis to propose such a classication ofthe Apocynaceae sl
Our aim is to base taxa on clades that arewell supported A second aim is to makeour classication compatible with the cur-rent system We will therefore primarily rec-ognize nonoverlapping tribes that are validunder the Linnaean system We will furtheradopt the principle of nested referencing (assuggested by Lee 1999 and Sereno 1999)from the Linnaean system Thus the rst ref-erence taxon in the node-based part of thedenition is the primary type which corre-sponds to the Linnaean type used for thename of the taxon Instead of the (optional)enumeration of subsumed taxa of a lowerrank (eg genera or subtribes) we will in-clude node C stem-based denitions as dis-cussed above Citations to primary types ofreference taxa in node-based part of de-nitions are also nested and thus refer toldquothe type specimen of the type species ofthe type genus of the dened taxonrdquo Aproblem is that the relationships between thetribes areweakly supported This may lead tovery cumbersome denitions enumeratingall competing tribes in the stem-based partof the denition (eg Moore 1998) One wayto reduce this problem is to dene interme-diate well-supported taxa and use them ascompeting taxaWe therefore will further rec-ognize four nonstandard taxa (note that anassignment of ldquononstandardrdquo taxa does notimply less reliability standard names relateto communication purposes only) The taxadiscussed are indicated in Figure 3 We willdiscuss the tree from the top of Figure 3 start-ing with the taxa of the traditional Plumeri-oideae Tribes and subtribes given in paren-theses in the text refer to the classicationsof Endress and Bruyns (2000) in other casesour classication is implied For some tribeswe have chosen a name with priority underLinnaean system that implies inclusion of agenus not sampled in the present study Forall but one (Willughbeieae) of these casesadditional molecular or morphological phy-logenetic analyses support the inclusion ofthe genus Even if too uncertain denitionsshould be avoided this is less critical withnode C stem-based denitions In the caseof Willughbeieae we have decided from
400 SYSTEMATIC BIOLOGY VOL 51
FIGURE 3 The tribal reclassication discussed in the text indicated on the combinable component consensustree from the successive weighting analysis of the complete rbcL data set Four nonstandard taxamdashApocynoidinaEuapocynoidina Asclepiadacina and Asclepiadoidinamdashare also indicated
discussions with a morphological expert(M Endress pers comm) that morphologi-cal support exists for including WillughbeiaRoxb in Willughbeieae If this is correctthe denition given below will be valid if
it is not our system allows redenition ofthe taxon We have also tried to providecomments on potential morphological char-acters taken from literature for the differ-ent taxa recognized However we have not
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 401
performed any morphological analysis butrely on information external to this studyfor these comments In most cases we cantherefore not identify synapomorphies forthe taxa
TAXONOMIC TREATMENT
Tribus Willughbeieae ADCThe Ancylobotrys Dictyophleba and Va-
hadenia clade is strongly supported (uwbootand swboot D 100) and corresponds to theWillughbeieae of Endress and Bruyns (2000)The present sampling contains closely re-lated genera that have been traditionallykept together Landolphia P Beauv sl in-cludes all these taxa (eg Schumann 1895)In addition to the taxa sampled here thestudy of Potgieter and Albert (2001) sup-ports the inclusion of Couma Aubl LacmelleaH Karst Pacouria Aubl and Saba (Pichon)Pichon in this clade Plants in this cladeall have a gynoecium that is congenitallysyncarpous and indehiscent and that con-tains a eshy pulp including numerous seedswith copious horny endosperm (Fallen 1986Persoon et al 1992) This probably con-stitutes synapomorphies for this clade(M Endress pers comm) although not un-ambiguously for example syncarpy is par-alleled in Carisseae We will therefore as-sume that the reference genus Willughbeia isincluded in this tribe
DenitionmdashTribe Willughbeieae is themost inclusive clade that includes the prim-ary types of Willughbeia and Dictyophleba butnot Alyxieae Aspidospermeae CarisseaeHunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Apocynoidina
Tribus Vinceae Bartl
Ochrosia Rauvola Vinca and Catharanthus(all Vinceae) form a strongly supported clade(uwboot D 89 swboot D 99) The associa-tion between Rauvola and Catharanthus wasindicated by Sennblad and Bremer (1996)The study of Potgieter and Albert (2001) sup-ports inclusion of the genera LaxoplumeriaMarkgr Neisosperma Raf and Tonduzia Pit-tier previously classied with AspidospermaThe Vinceae contain both herbaceous andfruticose taxa The fruits are apocarpousbut whereas the fruits of Catharanthus Lax-oplumeria Neisosperma Tonduzia and Vincahave dry fruit walls the fruits of Ochrosia andRauvola are drupes Plants in Vinceae share
a similar type of style head (ie the apicalenlarged part of the style) with a stigmatichollow and upper and lower hair wreaths(except in some species of Ochrosia) the pres-ence of a nectar disk and a linear hilum onthe seed (Pichon 1948b) On the present treeVinceae also include Kopsia however this po-sition of Kopsia is very weakly supported (seealso Sennblad and Bremer 1996)
DenitionmdashTribe Vinceae is the mostinclusive clade that includes the primarytypes of Vinca and Rauvola but not AlyxieaeAspidospermeae Carisseae HunterieaeMelodineae Plumerieae Tabernaemon-taneae Willughbeieae or Apocynoidina
Tribus Tabernaemontaneae G DonThe strongly supported association be-
tween Tabernaemontana Tabernanthe Carval-hoa Schizozygia and Molongum (Tabernae-montaneae uwboot D 84 swboot D 98)has been suggested previously (Boiteau et al1978 Fallen 1986 Endress et al 1996Sennblad and Bremer 1996) Inclusion of taxatraditionally associated with Tabernaemon-tana (eg Voacanga Thouars and CallichiliaStapf see also Tabernaemontaneae sensuLeeuwenberg 1994a) and Macoubea Aublis supported by the analysis of Potgieterand Albert (2001) The taxa share sclerenchy-matic anthers free from the style head anda characteristic band of heavily cutinizedcells at the insertion on the staminal rib(Endress et al 1996) The taxa tradition-ally associated with Tabernaemontana arecharacterized by their apocarpous fruitswith arillate seeds whereas genera asso-ciated with Molongum (Ambelanieae sensuLeeuwenberg 1994a) have syncarpous fruitsand lackanarillus HoweverMacoubea formsa link between the two having a syncar-pous fruit with arillate seeds (Zarucchi et al1995)
DenitionmdashTribe Tabernaemontaneae isthe most inclusive clade that includes the pri-mary types of Tabernaemontana and Schizozy-gia but not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae PlumerieaeVinceae Willughbeieae or Apocynoidina
Tribus Aspidospermeae Miers
Aspidosperma and Vallesia (Aspidosper-meae) form a strongly supported clade (uw-boot and swboot D 100) rst suggestedin Sennblad (1997 see also Sennblad and
402 SYSTEMATIC BIOLOGY VOL 51
Bremer 2000) Subsequently GeissospermumAllemao Haplophyton ADC MicroplumeriaBaill and Strempeliopsis Benth were indi-cated to belong to this clade (Potgieter andAlbert 2001) The Aspidospermeae havesimple style heads lacking both stigmatichollow and upper hair wreath and haveapocarpous fruits with variation in eshi-ness and sclerication Alstonia has tradi-tionally been placed in the Plumerieae Be-cause its seeds have a hairy margin it hasbeen suggested to form a link to subfam-ily Apocynoideae This was contradicted bySennblad and Bremerrsquos study (1996) how-ever where Alstonia was placed in an iso-lated position as the sister group to therest of the Apocynaceae sl Here it groupswith Aspidospermeae Sensu Endress andBruyns (2000) However this association isvery weakly supported (uwboot and swbootlt 50) and in other recent studies As-pidospermeae and Alstonia do not form aclade (Potgieter and Albert 2001) Becauseof this we do not use it as a reference taxonin the denition Thus although the Aspi-dospermeae include Alstonia on the presenttree some other position of Alstonia canbe accommodated without amending thedenition
DenitionmdashTribe Aspidospermeae is themost inclusive clade that includes theprimary types of Aspidosperma and Vallesiabut not Alyxieae Carisseae HunterieaeMelodineae Plumerieae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Melodineae G DonThe well-supported association between
Melodinus and Craspidospermum (uwboot D70 swboot D 75) was rst reportedby Sennblad (1997 see also Sennblad andBremer 2000) and has subsequently beenreected in the tribe Melodineae sensuEndress and Bruyns (2000 their inclusion ofDiplorhynchus however is not supported)Craspidospermum although its fruit is dryand dehiscent as opposed to the indehis-cent fruits of Melodinus was noted by Pichon(1948b) to present several characters of theCarisseae such as stamens inserted near thebase of the corolla tube and a syncarpousovary In particular Pichon found the verydense suprastaminal indumentum type rem-iniscent of Melodinus From his descriptions
of the two genera (Pichon 1948ab) the fol-lowing similarities also emerge presence ofa stipular line pollen in tetrads and a punc-tiform hilum
DenitionmdashTribe Melodineae is the mostinclusive clade that includes the pri-mary types of Melodinus and Craspidosper-mum but not Alyxieae AspidospermeaeCarisseae Hunterieae Plumerieae Taber-naemontaneae Vinceae Willughbeieae orApocynoidina
Tribus Hunterieae K Schum
The Picralima and Pleiocarpa clade isstrongly supported (uwboot and swboot D100) and corresponds to the subtribeHunterieae sensu Endress andBruyns (2000)The study of Potgieter and Albert (2001)also supports the inclusion of Hunteria RoxbThese taxa are characterized by apocar-pous sometimes pluricarpous ovaries andeshy brous fruit walls (Omino 1996)Diplorhynchus (Melodineae) is here weaklyassociated (uwboot and swboot lt 50) withthe Pleiocarpeae It shares a few similari-ties such as stipular lines and a style headwithout a stigmatic hollow and hair wreathsbut also has differences such as a dry de-hiscent fruit With the present denitionDiplorhynchus will tentatively be included inthe Hunterieae
DenitionmdashTribe Hunterieae is the mostinclusive clade that includes the primarytypes of Hunteria and Picralima but not Alyx-ieae Aspidospermeae Carisseae Melod-ineae Plumerieae TabernaemontaneaeVinceae Willughbeieae or Apocynoidina
Tribus Plumerieae EndlThe strongly supported relationship
(uwboot and swboot D 100) between Alla-manda and Plumeria corresponds to earlierstudies (Endress et al 1996 Sennblad andBremer 1996 Civeyrel et al 1998 Potgieterand Albert 2001) In all earlier treatmentsAllamanda has had uncertain relationshipsThe association with Plumeria nds supportin pollen morphology such as perforatemesocolpial depressions and similar innerexine pattern The two genera also containsecoiridoids rather than the indole alka-loids and cardenolides that are commonin the traditional Plumerioideae (Endresset al 1996) Cameraria Cerbera and Thevetiaform a strongly supported association
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 403
(uwboot D 83 swboot D 91) with Ane-chites which has been suggested by Fallen(1983) This association was based mainlyon the latrorsesublatrorse anthers andthe broad style head with large apicalappendages and stigmatic hollow The studyby Potgieter and Albert (2001) suggesteda further inclusion of Cerberiopsis Viell ExPancher amp Sebert and Skytanthus Meyen inthis tribe These taxa form a well-supportedclade with Plumeria and Allamanda whichcorresponds to the Plumerieae sensuEndress and Bruyns (2000) Characterssupporting this relationship are the presenceof infrastaminal as well as suprastaminalappendages (not present in Anechites andPlumeria) and winged seeds (not present inAnechites) (Endress et al 1996)
DenitionmdashTribe Plumerieae is themost inclusive clade that includes theprimary types of Plumeria and Allamandabut not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Carisseae Dumort
Acokanthera and Carissa (both Carisseae)are traditionally joined and are sometimeseven treated as one genus (eg Pichon1948a) In the present study they form astrongly supported clade (uwboot D 96swboot D 99) Among the characters sup-porting this tribe are syncarpous fruits with-out axile placentation and endocracks on theinner exine of the pollen (Endress et al 1996)
DenitionmdashTribe Carisseae is the most in-clusive clade that includes the primary typesof Carissa and Acokanthera but not AlyxieaeAspidospermeae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
Tribus Alyxieae G Don
Chilocarpus has been a genus with un-certain afnities In Endress et al (1996) itwas the sister taxon to the ApocynoideaeHere it is associated with Alyxia and Lep-inia in a clade corresponding to the Alyxieaesensu Endress and Bruyns (2000) Thisclade is strongly supported with succes-sive weighted bootstrap (swboot D 91) butnot with unit-weighted bootstrap support(uwboot D 58) such clades will henceforthbe termed moderatelywell supported clades
A further inclusion of Condylocarpon DesfLepiniopsis Valeton and Plectaneia Thouarswas indicated by Potgieter and Albert (2001)The constituent taxa are the only Apocy-naceae taxa outside the Apocynoidina thathave porate pollen
DenitionmdashTribe Alyxieae is the most in-clusive clade that includes the primary typesof Alyxia and Lepinia but not Aspidosper-meae Carisseae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
APOCYNOIDINA
On the present tree the traditionalsubfam-ily Apocynoideae includes all subfamilies(Periplocoideae Secamonoideae and Ascle-piadoideae) of the Asclepiadaceae sensuLiede and Albers (1994) All of these taxaare established names and selecting one asa subfamily at the expense of the othersmight be considered unfortunate from dif-ferent points of view We have here chosennot to recognize subfamilies instead we willdene the traditional Apocynoideae Ascle-piadaceae and Asclepiadoideae as nonstan-dard taxa The correspondence to the tradi-tional taxa will be indicated in the names byreplacing the sufx -eae with a neutral suf-x -ina (Kron 1997) We will also for practi-cal reasonsrecognize as a nonstandardtaxonthe informal group euapocynoids suggestedby Sennblad et al (1998)
The taxa of the traditional Apocynoideaeand Asclepiadaceae (uwboot D 86 swboot D99) clade will in many cases be more dif-cult to safely delimit to tribes becausemany of the groups in this clade have weakor no support (although preliminary resultsfrom an unpublished collaborative study bySennblad et al indicate additional supportfor the tribes discussed below) The tribaland subtribal classication of the traditionalApocynoideae has been shown to be prob-lematic (Leeuwenberg 1994a Endress andBruyns 2000) The circumscription of tribesfor these taxa might therefore in some casesbe preliminary
In all following taxa the basal part of theanther connective called the retinacle is ad-nate to the style head
DenitionmdashApocynoidina are the mostinclusive clade that includes the pri-mary types of Apocynum and Wrightia butnot Alyxieae Aspidospermeae Carisseae
404 SYSTEMATIC BIOLOGY VOL 51
Hunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Willughbeieae
Tribus Wrightieae G DonThe Stephanostema and Wrightia (both
Wrightieae) clade is strongly supported(uwboot and swboot D 100) and correspondsto the Wrightieae sensu Sennblad et al (1998)who also suggested an inclusion of Pleio-ceras Baill The tribe could be characterizedby a combination of synapomorphic and ple-siomorphic characters for example presenceofa chalazalandabsence of micropylar comaleft contorted aestivation and absence of airspaces in the anthers (1998)
DenitionmdashTribe Wrightieae is the mostinclusive clade in the Apocynoidina that in-cludes the primary types of Wrightia andStephanostema but not Malouetieae Nerieaeor Euapocynoidina
Tribus Nerieae (Benth) M Pichon
Adenium Nerium and Strophanthus (allWrightieae) constituted the Nerieae inSennblad et al (1998) however the cladereceived weak support Here Adenium andNerium form a clade that is absent in halfof the most-parsimonious trees in the com-plete rbcL analysis but is strongly supportedin the combined ndhF and rbcL analysis(uwboot D 98 swboot D 99) Strophan-thus groups with the Mascarenhasia Pachy-podium Kibatalia Funtumia and Holarrhenaclade which corresponds to the Malouetieaesensu Sennblad et al (1998) This position ofStrophanthus is however weakly supported(uwboot and swboot lt 50) and is not presentin the combined ndhF and rbcL tree Fur-thermore certain morphological characters(unfused slits in the corolla tube and anapical anther appendage) indicate a posi-tion closer to the Nerieae (Sennblad et al1998)
DenitionmdashTribe Nerieae is the most in-clusive clade in the Apocynoidina that in-cludes the primary types of Nerium and Ade-nium but not Malouetieae Wrightieae orEuapocynoidina
Tribus Malouetieae Mull-Arg
The association of Mascarenhasia Pachy-podium Kibatalia Funtumia and Holar-rhena (all Wrightieae) is moderately wellsupported in the complete rbcL analysis
(uwboot D 57 swboot D 74) whereasthe corresponding clade is well supportedin the combined rbcL and ndhF analysis(uwboot D 66 swboot D 100) On the ba-sis of identied synapomorphies (presenceof calcium oxalate crystals in the stomium ofthe anthers absence of interpetal vein ses-sile laments and absence of air spaces inthe anthers) Sennblad et al (1998) predictedan inclusion of Alaa Thouars Kibatalia andMalouetia ADC in the tribe The inclusion ofKibatalia is strongly supported in this studyFollowing these suggestions we will here as-sume the inclusion of the reference genusMalouetia in this tribe (The denition belowwill include Strophanthus on the present treebut see comments under Nerieae)
DenitionmdashTribe Malouetieae is the mostinclusive clade in the Apocynoidina that in-cluded the primary types of Malouetia andFuntumia but not Nerieae Wrightieae orEuapocynoidina
EUAPOCYNOIDINA
In all the remaining representatives ofthe Apocynoideae sensu Endress andBruyns (2000) that is Mandevilla MesechitesAganosma Apocynum TrachelospermumRhabdadenia Beaumontia Parsonsia andPrestonia the anthers are adnate to the stylehead both by the retinacle and by the thecaeThese taxa together with the taxa of the tra-ditional Asclepiadaceae form a moderatelysupported clade (uwboot lt 50 swboot D86) that corresponds to the informal taxoneuapocynoids sensu Sennblad et al (1998)
DenitionmdashEuapocynoidina are the mostinclusive clade that includes the primarytypes of Apocynum and Echites but notMalouetieae Wrightieae or Nerieae
Tribus Mesechiteae Miers
Mandevilla and Mesechites (both Mese-chiteae) form a moderately supported clade(uwboot lt 50 swboot D 87) equivalentto the Mesechiteae This tribe could be char-acterized by anthers with obtuse truncatedtails and a retinacle lacking hairs and also astrongly pentagonal style head with a stig-matic hollow
DenitionmdashTribe Mesechiteae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Mesechitesand Mandevilla but not Apocyneae EchiteaePeriploceae or Asclepiadacina
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 405
Tribus Apocyneae Bercht et J Presl
Aganosma Apocynum Beaumontia andTrachelospermum (Apocyneae) form a weaklysupported clade with Rhabdadenia (Echiteae)(uwboot and swboot lt 50) in the completerbcL analysis whereas the correspondingclade in the combined ndhF and rbcL analy-sis gains a little better support (uwboot lt 50swboot D 91) Likewise this clade (with theinclusion of Chonemorpha G Don) is weaklysupported in Potgieter and Albert (2001) Po-tential characterizing traits could be a stylehead without stigmatic hollow in combina-tion with thecae that are adnate to the stylehead (Sennblad et al 1998)
DenitionmdashTribe Apocyneae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Apocynumand Trachelospermum but not Echiteae Mese-chiteae Periploceae or Asclepiadacina
Tribus Echiteae Bartl
Parsonsia Peltastes and Prestonia (allEchiteae) form a moderately supported orunsupported clade with the Periplocoideaerepresentatives of the study (uwboot andswboot lt 50 and uwboot lt 50 swboot D89 in the complete rbcL and the combinedrbcL and ndhF analysis respectively) In thestudy of Sennblad et al (1998) Prestonia andParsonsia formed a weakly supported cladeand in the study of Potgieter and Albert(2001) inclusion of both Parsonsia and EchitesP Browne in the Echiteae is supportedParsonsia and Echiteae also share certaincharacters such as a horseshoe-shaped reti-nacle and a style head with stigmatic hollow(see Sennblad et al 1998)
DenitionmdashTribe Echiteae is the most in-clusive clade in the Euapocynoidina thatincludes the primary types of Echites andPrestonia but not Apocyneae MesechiteaePeriploceae or Asclepiadacina
Tribus Periploceae BartlThe representatives of the Periplo-
coideae sensu Endress and Bruyns (2000)mdashPentopetia Parquetina Petopentia PeriplocaTacazzea and Mondiamdashgroup together ina well-supported clade (uwboot D 67swboot D 93) The monophyly of the tra-ditional periplocoid taxa is also supportedin earlier studies that included a wider ordifferent sampling (Civeyrel et al 1998
Potgieter and Albert 2001) The taxa ofthe traditional Asclepiadaceae have ldquotrans-latorsrdquo structures related to pollinationspecialization In the Periploceae the pollenis deposited as tetrads in a spoon-liketranslator with an adhesive disc
DenitionmdashTribe Periploceae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Periplocaand Pentopetia but not Apocyneae EchiteaeMesechiteae or Asclepiadacina
ASCLEPIADACINA
The taxa of the Asclepiadoideae sensuLiede and Albers (1994) form a clade togetherwith the two Secamone species and Baissea(Apocyneae uwboot lt 50 swboot D 73)The exclusion of Periploceae (traditionallyassociated with this clade see Civeyrel et al1998) from this taxon is weakly supportedTherefore Periploceae is not included as areference taxon in the stem-based part ofthe denition of the Asclepiadacina thusleaving open the possibility for a later in-clusion of this taxon In the Asclepiadacina(except Baissea) the pollen is agglutinatedinto pollinia that are connected to a claspingtranslator
DenitionmdashAsclepiadacina are the mostinclusive monophyletic clade in the Euapoc-ynoidina that includes the primary types ofAsclepias and Secamoneae but not ApocyneaeEchiteae or Mesechiteae
Tribus Secamoneae G DonThe tribe Secamoneae sensu Bruyns
(uwboot D 89 swboot D 96) is mono-phyletic in the present analysisHowever be-cause representatives from only one of theconstituent genera are included monophylyis not well tested but the study of Civeyrelet al (1998) which included a larger numberof taxa including Pervillea Decne showedthat monophyly of the tribe was supportedThe traditional character for this tribe is fourpollinia per translator
DenitionmdashTribe Secamoneae is the mostinclusive clade in the Asclepiadacina that in-cludes the primary types of Secamone andPervillea but not Asclepioidina or Baissea
BaisseaThe position of Baissea corresponds with
that in the study of Sennblad et al (1998)
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
394 SYSTEMATIC BIOLOGY VOL 51
TABLE 2 Continued
Taxon Vouchersource EMBL number rbcLndhF
Secamone geayi Constatin et Gallaud Civeyrel 1200 LC AJ419761-Stapelia leendertziae NEBr Sennblad and Bremer 1996 X91778-Stephanotis oribunda Brongn Sennblad 256 UPS AJ419762AJ420125Stephanostema stenocarpum K Schum Sennblad and Bremer 1996 X91770AJ420126Strophanthus eminii Pax Sennblad and Bremer 1996 X91771AJ420127Tabernaemontana divaricata Roem et Schult Sennblad and Bremer 1996 X91772-Tabernanthe iboga Baill Leeuwenberg 12544 WAG AJ419763-Tacazzea apiculata Oliver Venter 9188 cult MSTR AJ419764-Thevetia peruviana (Pers) K Schum Sennblad and Bremer 1996 X91773-Trachelospermum jasminoides (Lindl) Lem Sennblad et al 1998 AJ002890AJ420128Tweedia coerulea Sweet Sennblad 254 UPS AJ419765-Tylophora sylvatica Decne Sennblad and Bremer 1996 X91789-Vahadenia caillei (AChev) Hutch et Dalziel Leeuwenberg 12275 WAG AJ419766-Vallesia antillana Woodson Meagher 966 FTG AJ419767-Vinca minor L Sennblad 230 UPS AJ419768-Vincetoxicum hirundinaria Medik Sennblad 257 UPS AJ419769-Wrightia arborea (Dennst) Mabb Sennblad et al 1998 AJ002891AJ420129
aVoucher for rbcL sequence onlybHerbarium abbreviations are according to Index Herbariorum except LC D Private herbarium of Laure CiveyrelcThis specimen was originally determined as Cibirhiza albersiana in Kunze et al (1994) but has recently been redetermined as
Fockea multiora K Schum (Verhoeven et al 2002)
of the clades in the successive weightinganalysis Clades with a successive weightedbootstrap gt63 were considered well sup-ported those with a successive weightedbootstrap gt85 were considered stronglysupported In a theoretical context a cladesupported by one uncontradicted charactercorresponds to a bootstrap support of 63anda cladesupported by two uncontradictedcharacters to 85 (Harshman 1994 Farriset al 1996) Here however this scale merelyrelates to the support distribution within theanalysis
RESULTS
The unit weight analysis of the completerbcL data set (237 characters total 214 ofthe cells were scored as missing data) re-sulted in 19003 trees 828 steps long witha consistency index (ci) D 0378 and a re-tention index (ri) D 0678 The consensustree from the successive weighting analysisof 252 most-parsimonious trees (each 1773steps long ci D 0661 ri D 0880) is presentedin Figure 1 The length of these trees withunit weight characters is 830 steps thus thesetrees are not identical to any of the most-parsimonious trees from the unit weightanalysis Branches not present in the strictconsensus tree from the unit weight analysisare indicated on the combinable consensustree from the successive weighting analysis
(Fig 1) The unit weight analysis of the com-bined set of the rbcL and ndhF data (153 char-acters total 043 of the cells were scored asmissing data) resulted in four trees 345 stepslong (ci D 0539 ri D 0482) The successiveweighting analysis gave three trees 978 stepslong (ci D 0840 ri D 0849) The strict consen-sus tree of the successive weighting analysisis presented in Figure 2 These trees are onestep longer with unit weights than the treesfrom the unit weight analysis Unless other-wise stated the following discussion refersto the successive weighting analyses of thecomplete rbcL data set
Investigating the justication of a prioriweighting (Sennblad and Bremer 2000) wemade some additional analyses of our rbcLdata set including tests for base compositionbias and rate heterogeneity This did not pro-duce signicantly different results and allwell-supported groups were congruent withthose of the present study The general struc-ture and the well-supported groups of theresulting trees (Figs 1 and 2) are to a largedegree congruent with the molecular stud-ies of Sennblad and Bremer (1996) whichare based on rbcL data of Civeyrel et al(1998) based mainly on matK sequence dataand of Potgieter and Albert (2001) basedmainly on trnLndashF data Although there aredifferences in the weakly supported rela-tionships among the well-supported groupsin particular the exact position of the
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 395
FIGURE 1 Combinable component consensus tree of the 252 most-parsimonious trees from the successiveweighting analysis of the complete rbcL data set Subfamilial and tribal classication is according to Endress andBruyns (2000) For subfamilies a three-letter code is used RAU APO PER SEC and ASC denote RauvoloideaeApocynoideae Periplocoideae Secamonoideae and Asclepiadoideae respectively OUT denotes the outgroupNumbers below branches are unit-weighted bootstrap values successive weighted bootstrap values are indicatedabove branches Branches not present in the strict consensus tree from the unit weight analysis are indicated witha cross (dagger) For branches marked A or B the frequencies in the combinable component consensus are 95 or 50respectively all other branches have 100 frequency
396 SYSTEMATIC BIOLOGY VOL 51
FIGURE 2 Strict consensus tree from the successive weighting analysis of the combined ndhF and rbcL dataset Tribal classication sensu Endress and Bruyns (2000) is indicated Numbers below branches are unit-weightedbootstrap values whereas successive weighted bootstrap values are indicated above branches Branches not presentin the strict consensus tree from the unit weight analysis are indicated with a cross (dagger)
Periplocoideae and of the root of the in-group all well-supported groups in our anal-ysis are congruent with those studies Thecombined analysis using data from rbcL-and matK-sequences and oral and pollenmorphology in the study of the Apocy-naceae sstr by Endress et al (1996) andthe analysis of Sennblad et al (1998) ofthe tribe Wrightieae sensu Leeuwenberg(1994a) and related taxa using rbcL andmorphological data are largely congruentwith the result from the present study Fur-thermore preliminary results from collab-orative studies by Sennblad et al (aimedat the traditional Apocynoideae and As-clepiadaceae pers comm) and Endresset al (aimed at the traditional Alyxieaepers comm) based on combined analy-sis of several molecular and morphologi-cal datasets are also congruent with ourresults
Support is good for the split betweenthe ingroup and the two outgroup taxaGelsemium and Mostuea (uwboot D 96swboot D 100) Thus the initial assump-tion of monophyly of the Apocynaceaesl is not violated Conforming to ear-lier studies (eg Sennblad and Bremer1996 Civeyrel et al 1998 Sennblad et al1998 Potgieter and Albert 2001) the tra-ditional Asclepiadaceae are nested withinthe traditional subfamily Apocynoideae(eg sensu Endress and Bruyns 2000)
The ApocynoideaendashAsclepiadaceae cladewhich is strongly supported (uwboot D 86swboot D 99) is in turn nested within thesubfamily Rauvoloideae sensu Endress andBruyns (2000)
Comparison of our results to the mostrecent classications of the ApocynaceaeEndress and Bruyns (2000 see Table 1 andFig 1) shows that although their circum-scription of the subfamilies PeriplocoideaeSecamonoideae and Asclepiadoideae aremonophyletic in our analysis their sub-families Rauvoloideae and Apocynoideaeare nonmonophyletic Similarly the tribesof the Rauvoloideae except Melodineaeare monophyletic but three of the tribesof the Apocynoideaemdashthe Apocyneae theEchiteae and the Wrightieaemdashare nonmono-phyletic Of the tribes of the Asclepiadoideaethe Ceropegieae and Asclepiadeae are mono-phyletic The position of Fockea as sisterto the rest of the Asclepiadoideae ren-ders the Marsdenieae nonmonophyletic Allthe included Periplocoideae taxa belongto the tribe Periploceae sensu Venter andVerhoeven (1997) except Pentopetia whichthey placed in the tribe CryptolepideaeGiven this unresolved position of Pentopetiathe monophyly of these tribes cannot be eval-uated however the reduction of Parquetinato synonymy with Periploca proposed byVenter and Verhoeven (1997) is not sup-ported by the present study
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 397
DISCUSSION
One of the most important tasks for anomenclature system is to provide a sta-ble means to communicate classications(which we here will understand as represent-ing phylogenetic relationships) Althoughboth the Linnaean and the Phylocode sys-tems are able to communicate phylogeniesthey are sensitive to changes in phylogenetichypotheses which in both systems will de-mand taxonomic changes In addition theLinnaean system is subject to a potentialinstability relating to subjective changes incircumscription that is lumpers and split-ters may refer to different groups of organ-isms by the same name (see de Queiroz andGauthier 1992) For many taxa a consensuscircumscription has eventually been estab-lished but for other groups this is still a prob-lem A current example of this from Apoc-ynaceae is the genus Tabernaemontana sensuLeeuwenberg (1991 1994b) which in theclassication of Allorge (1985) is split intoeight genera The Phylocode system was de-signed to avoid this problem by use of strictdenitions of taxa subjective changes incircumscription are avoided
However this rigidity in denitions leadsto a potential instability and high turnoverof names associated with a group of nestedtaxa Relatively small changes in phyloge-netic views eg when a dening taxonldquospecierrdquo receives a less nested positionmaycause some(or all) of these nested namesto become synonymous (eg de Queirozand Gauthier 1994 Bryant 1996 Liden andOxelman 1996Sereno 1999)The junior syn-onyms should according to the Phylocodebe rejected and new names may then beneeded for the new system of nested clades(de Queiroz and Gauthier 1994) The namesassociated with this group of taxa are thusreplaced by new names This can be espe-cially problematic when careless denitionshave been made as have been noted bySchander and Thollesson (1995) and Cantinoet al (1997) among others Consider a hy-pothetical worst-case example The Ascle-piadaceae were traditionally considered asseparate from Apocynaceae A stem-baseddenition of Apocynaceae reecting thisview could be the largest clade that in-cludes Apocynum but not Asclepias On thepresent tree such a denition would limitthe Apocynaceae to be synonymous with the
tribe Apocyneae and a new name wouldbe needed for the taxon corresponding tothe current view of Apocynaceae (eg sensuEndress and Bruyns 2000) A node-baseddenition would handle this particular prob-lem but may have other problems as illus-trated here with another hypothetical worst-case scenario Early classications of theApocynaceae included Plocosperma BenthA node-based denition representing thisview could have included as reference taxasay Apocynum and Plocosperma The recentlyproposed position of Plocosperma close tothe Boraginaceae (Backlund et al 2000)would with this denition have made Apoc-ynaceae a very large taxon (containing atleast both Gentianales and Boraginales) pos-sibly synonymous with (a hypothetical def-inition of) the informal taxon EuasteridsII (Bremer et al 1998) and again a newname for the Apocynaceae sensu Endressand Bruyns (2000 ie excluding Plocosperma)would have been needed Because theLinnaean system allows recircumscription ofsynonymous names the potential turnoverin the set of names in use is reduced withthis system (albeit at the expense of allowingsubjectivity) Additionally the ranks of theLinnaean system provide a set of names thatwork as relatively stable universal standardnames in communication note that equalranks does not guarantee comparable evo-lutionary units however Continuity in us-age of names is important for example injournals literature databases oras teach-ing and inventory work and the need forsuch continuity has been recognized by pro-ponents of both nomenclature systems (deQueiroz and Gauthier 1992 Greuter et al1994 Reveal 1996)
We propose a compromise approach thatuses solutions relating to these aspects fromboth systems concentrating mainly onbotanical classication of extant plant taxaat the tribal and familial levels and onlyconsidering monophyletic taxa We recom-mend a system that uses ldquostandard namesrdquo(ie universally used communication unitsat convenient hierarchical levels) such asthose provided by the principal ranks of theLinnaean system (Greuter et al 1994) but inwhich sensitivity toward subjective changesin circumscription is reduced One way to dothis is to adopt the denitions of the phy-logenetic system but use constraints similar
398 SYSTEMATIC BIOLOGY VOL 51
to those that apply between ranks in theLinnaean system To reduce sensitivity to-ward changes in phylogenetic hypotheses(Schander and Thollesson 1995 Bryant1996 1997 Cantino et al 1997 Moore 1998)we propose using combined node C stem-based denitions that will provoke explicitincompatibilities between taxa in case ofunfortunate changes in circumscription (deQueiroz 1996 has proposed a similar type ofdenition for designating nonmonophyletictaxa somewhat similar variants are alsodiscussed in the Phylocode) A node-basedpart of the denition provides a minimalcircumscription whereas a stem-based partprovides boundaries to competing taxa(maximal circumscription) If on a particu-lar phylogeny the minimal (monophyletic)circumscriptions of competing taxa are over-lapping then the two taxa are incompatibleTo make this meaningful at least two taxashould be included in the node-based partof the denition this will also avoid unnec-essary monotypic taxa This incompatibilitywill provide an objective point when nomen-claturalreconsideration isneeded Returningto the rst of our examples above node Cstem-based denitions of the Apocynaceaeand Asclepiadaceae could be as follows
The family Apocynaceae is the most inclusive clade inthe order Gentianales including the type specimensof Apocynum and Dictyophleba but not the taxa Ascle-piadaceae [ Gentianaceae Rubiaceae ]
and
The family Asclepiadaceae is the most inclusive cladein the order Gentianales including the type speci-mens of Asclepias and Periploca but not the taxa Apoc-ynaceae [ Gentianaceae Rubiaceae ]
On the tree in Figure 1 these two denitionsare mutually exclusive and a decision onwhich name to keep must be made Mini-mizing the number of classicatory changesshould take precedence in this decision Thereason for this is to avoid the large incon-venient ldquostandard namerdquo changes that couldresult from relatively small rearrangementsRejected names are then ignored when oc-curring in denitions of other taxa This isbecause the reference taxa in the stem-basedpart of the denitions are ldquodenedrdquo taxarather than physical types (thus reliance ondened reference taxa which has been crit-icized for stem-based denitions by Sereno1999 is what we aim for on the other hand
the explicit reference of type specimens in theminimum circumscription provided by thenode-based part of the denitions circum-vents the tautology problem raised byBryant 1996) In the current example thiswould mean that regardless of which nameof Apocynaceae and Asclepiadaceae is re-jected the other will be dened as corre-sponding to the Apocynaceae sensu Endressand Bruyns (2000) The choice will there-fore simply be which name to keep In thiscase because Apocynaceae has priority un-der the Linnaean system we would chooseto keep Apocynaceae to promote consistencywith the current system Nevertheless out-side these recommendations choices of whatnames to choose will probably to a certaindegree be arbitrary We also propose thatthe rejected name should be reinstated withan emended denition if appropriate to po-tentially allow for more stability in the setof names in use Thus in our second ex-ample above a node C stem-based deni-tion reecting the inclusion of Plocospermain Apocynaceae (eg by exchanging Dic-tyophleba for Plocosperma in the denitionabove) would presumably be incongruentwith the denition of Gentianales as well asthe denitions of Rubiaceae Gentianaceaeand so forth and therefore would be rejectedIn such acasewhere the only real change is inthe position of a single taxon a reinstatementof the Apocynaceae with an emended def-inition is motivated However a drawbackis that this may also introduce elements ofsubjectivity in the circumscriptions of taxa
Not all named clades need to be providedwith standard names In fact for the purposeof a universal set used in journals databasesand so forth it may sufce to use standardnames corresponding to the principal ranksof the Linnaean system (Greuter et al 1994)such as species genus family and orderNevertheless to provide compatibility withthe present system we suggest that addi-tional names corresponding to commonlyused secondary ranks (eg tribes) may berecognized The hierarchical level of a stan-dard name needs to be indicated This doesnot imply that taxa of the same hierarchi-cal level are comparable evolutionary unitsbut simply indicates the hierarchical levelof a taxon relative to nested taxa Becausesuch misinterpretations will in any case bedifcult to prohibit one may as well pro-vide compatibility with the present Linnaean
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 399
system by using the names of the Linnaeanranks (eg species genus tribe familyorder) to indicate hierarchical level The prin-ciple of exhaustive subsidiary taxa is notaccepted so redundant taxa need not be rec-ognized Other ldquononstandardrdquo taxa couldbe dened by using Phylocode denitionsmdashthat is without the restraints discussed forstandard taxa above
One advantage of the described systemis that utilizing the correspondence betweenthe standard names and Linnaean ranks al-lows the system to be largely compatiblewiththe establishedLinnaean system A change ofsystem could therefore be gradual with thetwo systems coexisting during a conversiontime (see eg de Quieroz 1997b)
Even though the node C stem-baseddenitions reduce the sensitivity towardschanges in circumscriptions they may still besensitive towardldquounfortunaterdquo or ldquobadrdquo def-initions (see eg Cantino et al 1997Cantinoand de Queiroz 2000) Many of the recom-mendations for such things as choices of ref-erence taxa (speciers) in phylogenetic def-initions are applicable also to the node Cstem-based denitions (eg Schander andThollesson 1995 Bryant 1996 Cantino et al1997 Sereno 1999) Most likely the compro-mise system described above will have sev-eral further problems and we hope that thispaper will invite further discussion on thesubject
A New Classication of the Apocynaceae slOur results indicate that some problem-
atic taxa remain in the classication ofEndress and Bruyns (2000) Monophyly fortwo of the subfamilies and ve of the tribesis called into question If other recent molec-ular studies (eg Potgieter and Albert 2001)are taken into consideration a further fourtribes (Alstonieae Alyxieae Plumerieae andVinceae) may be nonmonophyletic This canin some cases be simply analysis artifactsbut as Endress and Bruyns (2000) themselvespoint out their classication is to be consid-ered preliminary Insufcient taxonsamplingin molecular studies and difcult homologydecisions relating to morphological charac-ters make the circumscription of many ofthe tribes for example in the Apocynoideaesensu Endress and Bruyns (2000) uncertainThere is thus risk for future rearrangementsin many of the tribes A classication using
node C stem-based denitions may be ableto subsume such rearrangements withoutnumerous explicit recircumscriptions Wewill here use the result fromthe present studyas a basis to propose such a classication ofthe Apocynaceae sl
Our aim is to base taxa on clades that arewell supported A second aim is to makeour classication compatible with the cur-rent system We will therefore primarily rec-ognize nonoverlapping tribes that are validunder the Linnaean system We will furtheradopt the principle of nested referencing (assuggested by Lee 1999 and Sereno 1999)from the Linnaean system Thus the rst ref-erence taxon in the node-based part of thedenition is the primary type which corre-sponds to the Linnaean type used for thename of the taxon Instead of the (optional)enumeration of subsumed taxa of a lowerrank (eg genera or subtribes) we will in-clude node C stem-based denitions as dis-cussed above Citations to primary types ofreference taxa in node-based part of de-nitions are also nested and thus refer toldquothe type specimen of the type species ofthe type genus of the dened taxonrdquo Aproblem is that the relationships between thetribes areweakly supported This may lead tovery cumbersome denitions enumeratingall competing tribes in the stem-based partof the denition (eg Moore 1998) One wayto reduce this problem is to dene interme-diate well-supported taxa and use them ascompeting taxaWe therefore will further rec-ognize four nonstandard taxa (note that anassignment of ldquononstandardrdquo taxa does notimply less reliability standard names relateto communication purposes only) The taxadiscussed are indicated in Figure 3 We willdiscuss the tree from the top of Figure 3 start-ing with the taxa of the traditional Plumeri-oideae Tribes and subtribes given in paren-theses in the text refer to the classicationsof Endress and Bruyns (2000) in other casesour classication is implied For some tribeswe have chosen a name with priority underLinnaean system that implies inclusion of agenus not sampled in the present study Forall but one (Willughbeieae) of these casesadditional molecular or morphological phy-logenetic analyses support the inclusion ofthe genus Even if too uncertain denitionsshould be avoided this is less critical withnode C stem-based denitions In the caseof Willughbeieae we have decided from
400 SYSTEMATIC BIOLOGY VOL 51
FIGURE 3 The tribal reclassication discussed in the text indicated on the combinable component consensustree from the successive weighting analysis of the complete rbcL data set Four nonstandard taxamdashApocynoidinaEuapocynoidina Asclepiadacina and Asclepiadoidinamdashare also indicated
discussions with a morphological expert(M Endress pers comm) that morphologi-cal support exists for including WillughbeiaRoxb in Willughbeieae If this is correctthe denition given below will be valid if
it is not our system allows redenition ofthe taxon We have also tried to providecomments on potential morphological char-acters taken from literature for the differ-ent taxa recognized However we have not
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 401
performed any morphological analysis butrely on information external to this studyfor these comments In most cases we cantherefore not identify synapomorphies forthe taxa
TAXONOMIC TREATMENT
Tribus Willughbeieae ADCThe Ancylobotrys Dictyophleba and Va-
hadenia clade is strongly supported (uwbootand swboot D 100) and corresponds to theWillughbeieae of Endress and Bruyns (2000)The present sampling contains closely re-lated genera that have been traditionallykept together Landolphia P Beauv sl in-cludes all these taxa (eg Schumann 1895)In addition to the taxa sampled here thestudy of Potgieter and Albert (2001) sup-ports the inclusion of Couma Aubl LacmelleaH Karst Pacouria Aubl and Saba (Pichon)Pichon in this clade Plants in this cladeall have a gynoecium that is congenitallysyncarpous and indehiscent and that con-tains a eshy pulp including numerous seedswith copious horny endosperm (Fallen 1986Persoon et al 1992) This probably con-stitutes synapomorphies for this clade(M Endress pers comm) although not un-ambiguously for example syncarpy is par-alleled in Carisseae We will therefore as-sume that the reference genus Willughbeia isincluded in this tribe
DenitionmdashTribe Willughbeieae is themost inclusive clade that includes the prim-ary types of Willughbeia and Dictyophleba butnot Alyxieae Aspidospermeae CarisseaeHunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Apocynoidina
Tribus Vinceae Bartl
Ochrosia Rauvola Vinca and Catharanthus(all Vinceae) form a strongly supported clade(uwboot D 89 swboot D 99) The associa-tion between Rauvola and Catharanthus wasindicated by Sennblad and Bremer (1996)The study of Potgieter and Albert (2001) sup-ports inclusion of the genera LaxoplumeriaMarkgr Neisosperma Raf and Tonduzia Pit-tier previously classied with AspidospermaThe Vinceae contain both herbaceous andfruticose taxa The fruits are apocarpousbut whereas the fruits of Catharanthus Lax-oplumeria Neisosperma Tonduzia and Vincahave dry fruit walls the fruits of Ochrosia andRauvola are drupes Plants in Vinceae share
a similar type of style head (ie the apicalenlarged part of the style) with a stigmatichollow and upper and lower hair wreaths(except in some species of Ochrosia) the pres-ence of a nectar disk and a linear hilum onthe seed (Pichon 1948b) On the present treeVinceae also include Kopsia however this po-sition of Kopsia is very weakly supported (seealso Sennblad and Bremer 1996)
DenitionmdashTribe Vinceae is the mostinclusive clade that includes the primarytypes of Vinca and Rauvola but not AlyxieaeAspidospermeae Carisseae HunterieaeMelodineae Plumerieae Tabernaemon-taneae Willughbeieae or Apocynoidina
Tribus Tabernaemontaneae G DonThe strongly supported association be-
tween Tabernaemontana Tabernanthe Carval-hoa Schizozygia and Molongum (Tabernae-montaneae uwboot D 84 swboot D 98)has been suggested previously (Boiteau et al1978 Fallen 1986 Endress et al 1996Sennblad and Bremer 1996) Inclusion of taxatraditionally associated with Tabernaemon-tana (eg Voacanga Thouars and CallichiliaStapf see also Tabernaemontaneae sensuLeeuwenberg 1994a) and Macoubea Aublis supported by the analysis of Potgieterand Albert (2001) The taxa share sclerenchy-matic anthers free from the style head anda characteristic band of heavily cutinizedcells at the insertion on the staminal rib(Endress et al 1996) The taxa tradition-ally associated with Tabernaemontana arecharacterized by their apocarpous fruitswith arillate seeds whereas genera asso-ciated with Molongum (Ambelanieae sensuLeeuwenberg 1994a) have syncarpous fruitsand lackanarillus HoweverMacoubea formsa link between the two having a syncar-pous fruit with arillate seeds (Zarucchi et al1995)
DenitionmdashTribe Tabernaemontaneae isthe most inclusive clade that includes the pri-mary types of Tabernaemontana and Schizozy-gia but not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae PlumerieaeVinceae Willughbeieae or Apocynoidina
Tribus Aspidospermeae Miers
Aspidosperma and Vallesia (Aspidosper-meae) form a strongly supported clade (uw-boot and swboot D 100) rst suggestedin Sennblad (1997 see also Sennblad and
402 SYSTEMATIC BIOLOGY VOL 51
Bremer 2000) Subsequently GeissospermumAllemao Haplophyton ADC MicroplumeriaBaill and Strempeliopsis Benth were indi-cated to belong to this clade (Potgieter andAlbert 2001) The Aspidospermeae havesimple style heads lacking both stigmatichollow and upper hair wreath and haveapocarpous fruits with variation in eshi-ness and sclerication Alstonia has tradi-tionally been placed in the Plumerieae Be-cause its seeds have a hairy margin it hasbeen suggested to form a link to subfam-ily Apocynoideae This was contradicted bySennblad and Bremerrsquos study (1996) how-ever where Alstonia was placed in an iso-lated position as the sister group to therest of the Apocynaceae sl Here it groupswith Aspidospermeae Sensu Endress andBruyns (2000) However this association isvery weakly supported (uwboot and swbootlt 50) and in other recent studies As-pidospermeae and Alstonia do not form aclade (Potgieter and Albert 2001) Becauseof this we do not use it as a reference taxonin the denition Thus although the Aspi-dospermeae include Alstonia on the presenttree some other position of Alstonia canbe accommodated without amending thedenition
DenitionmdashTribe Aspidospermeae is themost inclusive clade that includes theprimary types of Aspidosperma and Vallesiabut not Alyxieae Carisseae HunterieaeMelodineae Plumerieae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Melodineae G DonThe well-supported association between
Melodinus and Craspidospermum (uwboot D70 swboot D 75) was rst reportedby Sennblad (1997 see also Sennblad andBremer 2000) and has subsequently beenreected in the tribe Melodineae sensuEndress and Bruyns (2000 their inclusion ofDiplorhynchus however is not supported)Craspidospermum although its fruit is dryand dehiscent as opposed to the indehis-cent fruits of Melodinus was noted by Pichon(1948b) to present several characters of theCarisseae such as stamens inserted near thebase of the corolla tube and a syncarpousovary In particular Pichon found the verydense suprastaminal indumentum type rem-iniscent of Melodinus From his descriptions
of the two genera (Pichon 1948ab) the fol-lowing similarities also emerge presence ofa stipular line pollen in tetrads and a punc-tiform hilum
DenitionmdashTribe Melodineae is the mostinclusive clade that includes the pri-mary types of Melodinus and Craspidosper-mum but not Alyxieae AspidospermeaeCarisseae Hunterieae Plumerieae Taber-naemontaneae Vinceae Willughbeieae orApocynoidina
Tribus Hunterieae K Schum
The Picralima and Pleiocarpa clade isstrongly supported (uwboot and swboot D100) and corresponds to the subtribeHunterieae sensu Endress andBruyns (2000)The study of Potgieter and Albert (2001)also supports the inclusion of Hunteria RoxbThese taxa are characterized by apocar-pous sometimes pluricarpous ovaries andeshy brous fruit walls (Omino 1996)Diplorhynchus (Melodineae) is here weaklyassociated (uwboot and swboot lt 50) withthe Pleiocarpeae It shares a few similari-ties such as stipular lines and a style headwithout a stigmatic hollow and hair wreathsbut also has differences such as a dry de-hiscent fruit With the present denitionDiplorhynchus will tentatively be included inthe Hunterieae
DenitionmdashTribe Hunterieae is the mostinclusive clade that includes the primarytypes of Hunteria and Picralima but not Alyx-ieae Aspidospermeae Carisseae Melod-ineae Plumerieae TabernaemontaneaeVinceae Willughbeieae or Apocynoidina
Tribus Plumerieae EndlThe strongly supported relationship
(uwboot and swboot D 100) between Alla-manda and Plumeria corresponds to earlierstudies (Endress et al 1996 Sennblad andBremer 1996 Civeyrel et al 1998 Potgieterand Albert 2001) In all earlier treatmentsAllamanda has had uncertain relationshipsThe association with Plumeria nds supportin pollen morphology such as perforatemesocolpial depressions and similar innerexine pattern The two genera also containsecoiridoids rather than the indole alka-loids and cardenolides that are commonin the traditional Plumerioideae (Endresset al 1996) Cameraria Cerbera and Thevetiaform a strongly supported association
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 403
(uwboot D 83 swboot D 91) with Ane-chites which has been suggested by Fallen(1983) This association was based mainlyon the latrorsesublatrorse anthers andthe broad style head with large apicalappendages and stigmatic hollow The studyby Potgieter and Albert (2001) suggesteda further inclusion of Cerberiopsis Viell ExPancher amp Sebert and Skytanthus Meyen inthis tribe These taxa form a well-supportedclade with Plumeria and Allamanda whichcorresponds to the Plumerieae sensuEndress and Bruyns (2000) Characterssupporting this relationship are the presenceof infrastaminal as well as suprastaminalappendages (not present in Anechites andPlumeria) and winged seeds (not present inAnechites) (Endress et al 1996)
DenitionmdashTribe Plumerieae is themost inclusive clade that includes theprimary types of Plumeria and Allamandabut not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Carisseae Dumort
Acokanthera and Carissa (both Carisseae)are traditionally joined and are sometimeseven treated as one genus (eg Pichon1948a) In the present study they form astrongly supported clade (uwboot D 96swboot D 99) Among the characters sup-porting this tribe are syncarpous fruits with-out axile placentation and endocracks on theinner exine of the pollen (Endress et al 1996)
DenitionmdashTribe Carisseae is the most in-clusive clade that includes the primary typesof Carissa and Acokanthera but not AlyxieaeAspidospermeae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
Tribus Alyxieae G Don
Chilocarpus has been a genus with un-certain afnities In Endress et al (1996) itwas the sister taxon to the ApocynoideaeHere it is associated with Alyxia and Lep-inia in a clade corresponding to the Alyxieaesensu Endress and Bruyns (2000) Thisclade is strongly supported with succes-sive weighted bootstrap (swboot D 91) butnot with unit-weighted bootstrap support(uwboot D 58) such clades will henceforthbe termed moderatelywell supported clades
A further inclusion of Condylocarpon DesfLepiniopsis Valeton and Plectaneia Thouarswas indicated by Potgieter and Albert (2001)The constituent taxa are the only Apocy-naceae taxa outside the Apocynoidina thathave porate pollen
DenitionmdashTribe Alyxieae is the most in-clusive clade that includes the primary typesof Alyxia and Lepinia but not Aspidosper-meae Carisseae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
APOCYNOIDINA
On the present tree the traditionalsubfam-ily Apocynoideae includes all subfamilies(Periplocoideae Secamonoideae and Ascle-piadoideae) of the Asclepiadaceae sensuLiede and Albers (1994) All of these taxaare established names and selecting one asa subfamily at the expense of the othersmight be considered unfortunate from dif-ferent points of view We have here chosennot to recognize subfamilies instead we willdene the traditional Apocynoideae Ascle-piadaceae and Asclepiadoideae as nonstan-dard taxa The correspondence to the tradi-tional taxa will be indicated in the names byreplacing the sufx -eae with a neutral suf-x -ina (Kron 1997) We will also for practi-cal reasonsrecognize as a nonstandardtaxonthe informal group euapocynoids suggestedby Sennblad et al (1998)
The taxa of the traditional Apocynoideaeand Asclepiadaceae (uwboot D 86 swboot D99) clade will in many cases be more dif-cult to safely delimit to tribes becausemany of the groups in this clade have weakor no support (although preliminary resultsfrom an unpublished collaborative study bySennblad et al indicate additional supportfor the tribes discussed below) The tribaland subtribal classication of the traditionalApocynoideae has been shown to be prob-lematic (Leeuwenberg 1994a Endress andBruyns 2000) The circumscription of tribesfor these taxa might therefore in some casesbe preliminary
In all following taxa the basal part of theanther connective called the retinacle is ad-nate to the style head
DenitionmdashApocynoidina are the mostinclusive clade that includes the pri-mary types of Apocynum and Wrightia butnot Alyxieae Aspidospermeae Carisseae
404 SYSTEMATIC BIOLOGY VOL 51
Hunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Willughbeieae
Tribus Wrightieae G DonThe Stephanostema and Wrightia (both
Wrightieae) clade is strongly supported(uwboot and swboot D 100) and correspondsto the Wrightieae sensu Sennblad et al (1998)who also suggested an inclusion of Pleio-ceras Baill The tribe could be characterizedby a combination of synapomorphic and ple-siomorphic characters for example presenceofa chalazalandabsence of micropylar comaleft contorted aestivation and absence of airspaces in the anthers (1998)
DenitionmdashTribe Wrightieae is the mostinclusive clade in the Apocynoidina that in-cludes the primary types of Wrightia andStephanostema but not Malouetieae Nerieaeor Euapocynoidina
Tribus Nerieae (Benth) M Pichon
Adenium Nerium and Strophanthus (allWrightieae) constituted the Nerieae inSennblad et al (1998) however the cladereceived weak support Here Adenium andNerium form a clade that is absent in halfof the most-parsimonious trees in the com-plete rbcL analysis but is strongly supportedin the combined ndhF and rbcL analysis(uwboot D 98 swboot D 99) Strophan-thus groups with the Mascarenhasia Pachy-podium Kibatalia Funtumia and Holarrhenaclade which corresponds to the Malouetieaesensu Sennblad et al (1998) This position ofStrophanthus is however weakly supported(uwboot and swboot lt 50) and is not presentin the combined ndhF and rbcL tree Fur-thermore certain morphological characters(unfused slits in the corolla tube and anapical anther appendage) indicate a posi-tion closer to the Nerieae (Sennblad et al1998)
DenitionmdashTribe Nerieae is the most in-clusive clade in the Apocynoidina that in-cludes the primary types of Nerium and Ade-nium but not Malouetieae Wrightieae orEuapocynoidina
Tribus Malouetieae Mull-Arg
The association of Mascarenhasia Pachy-podium Kibatalia Funtumia and Holar-rhena (all Wrightieae) is moderately wellsupported in the complete rbcL analysis
(uwboot D 57 swboot D 74) whereasthe corresponding clade is well supportedin the combined rbcL and ndhF analysis(uwboot D 66 swboot D 100) On the ba-sis of identied synapomorphies (presenceof calcium oxalate crystals in the stomium ofthe anthers absence of interpetal vein ses-sile laments and absence of air spaces inthe anthers) Sennblad et al (1998) predictedan inclusion of Alaa Thouars Kibatalia andMalouetia ADC in the tribe The inclusion ofKibatalia is strongly supported in this studyFollowing these suggestions we will here as-sume the inclusion of the reference genusMalouetia in this tribe (The denition belowwill include Strophanthus on the present treebut see comments under Nerieae)
DenitionmdashTribe Malouetieae is the mostinclusive clade in the Apocynoidina that in-cluded the primary types of Malouetia andFuntumia but not Nerieae Wrightieae orEuapocynoidina
EUAPOCYNOIDINA
In all the remaining representatives ofthe Apocynoideae sensu Endress andBruyns (2000) that is Mandevilla MesechitesAganosma Apocynum TrachelospermumRhabdadenia Beaumontia Parsonsia andPrestonia the anthers are adnate to the stylehead both by the retinacle and by the thecaeThese taxa together with the taxa of the tra-ditional Asclepiadaceae form a moderatelysupported clade (uwboot lt 50 swboot D86) that corresponds to the informal taxoneuapocynoids sensu Sennblad et al (1998)
DenitionmdashEuapocynoidina are the mostinclusive clade that includes the primarytypes of Apocynum and Echites but notMalouetieae Wrightieae or Nerieae
Tribus Mesechiteae Miers
Mandevilla and Mesechites (both Mese-chiteae) form a moderately supported clade(uwboot lt 50 swboot D 87) equivalentto the Mesechiteae This tribe could be char-acterized by anthers with obtuse truncatedtails and a retinacle lacking hairs and also astrongly pentagonal style head with a stig-matic hollow
DenitionmdashTribe Mesechiteae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Mesechitesand Mandevilla but not Apocyneae EchiteaePeriploceae or Asclepiadacina
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 405
Tribus Apocyneae Bercht et J Presl
Aganosma Apocynum Beaumontia andTrachelospermum (Apocyneae) form a weaklysupported clade with Rhabdadenia (Echiteae)(uwboot and swboot lt 50) in the completerbcL analysis whereas the correspondingclade in the combined ndhF and rbcL analy-sis gains a little better support (uwboot lt 50swboot D 91) Likewise this clade (with theinclusion of Chonemorpha G Don) is weaklysupported in Potgieter and Albert (2001) Po-tential characterizing traits could be a stylehead without stigmatic hollow in combina-tion with thecae that are adnate to the stylehead (Sennblad et al 1998)
DenitionmdashTribe Apocyneae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Apocynumand Trachelospermum but not Echiteae Mese-chiteae Periploceae or Asclepiadacina
Tribus Echiteae Bartl
Parsonsia Peltastes and Prestonia (allEchiteae) form a moderately supported orunsupported clade with the Periplocoideaerepresentatives of the study (uwboot andswboot lt 50 and uwboot lt 50 swboot D89 in the complete rbcL and the combinedrbcL and ndhF analysis respectively) In thestudy of Sennblad et al (1998) Prestonia andParsonsia formed a weakly supported cladeand in the study of Potgieter and Albert(2001) inclusion of both Parsonsia and EchitesP Browne in the Echiteae is supportedParsonsia and Echiteae also share certaincharacters such as a horseshoe-shaped reti-nacle and a style head with stigmatic hollow(see Sennblad et al 1998)
DenitionmdashTribe Echiteae is the most in-clusive clade in the Euapocynoidina thatincludes the primary types of Echites andPrestonia but not Apocyneae MesechiteaePeriploceae or Asclepiadacina
Tribus Periploceae BartlThe representatives of the Periplo-
coideae sensu Endress and Bruyns (2000)mdashPentopetia Parquetina Petopentia PeriplocaTacazzea and Mondiamdashgroup together ina well-supported clade (uwboot D 67swboot D 93) The monophyly of the tra-ditional periplocoid taxa is also supportedin earlier studies that included a wider ordifferent sampling (Civeyrel et al 1998
Potgieter and Albert 2001) The taxa ofthe traditional Asclepiadaceae have ldquotrans-latorsrdquo structures related to pollinationspecialization In the Periploceae the pollenis deposited as tetrads in a spoon-liketranslator with an adhesive disc
DenitionmdashTribe Periploceae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Periplocaand Pentopetia but not Apocyneae EchiteaeMesechiteae or Asclepiadacina
ASCLEPIADACINA
The taxa of the Asclepiadoideae sensuLiede and Albers (1994) form a clade togetherwith the two Secamone species and Baissea(Apocyneae uwboot lt 50 swboot D 73)The exclusion of Periploceae (traditionallyassociated with this clade see Civeyrel et al1998) from this taxon is weakly supportedTherefore Periploceae is not included as areference taxon in the stem-based part ofthe denition of the Asclepiadacina thusleaving open the possibility for a later in-clusion of this taxon In the Asclepiadacina(except Baissea) the pollen is agglutinatedinto pollinia that are connected to a claspingtranslator
DenitionmdashAsclepiadacina are the mostinclusive monophyletic clade in the Euapoc-ynoidina that includes the primary types ofAsclepias and Secamoneae but not ApocyneaeEchiteae or Mesechiteae
Tribus Secamoneae G DonThe tribe Secamoneae sensu Bruyns
(uwboot D 89 swboot D 96) is mono-phyletic in the present analysisHowever be-cause representatives from only one of theconstituent genera are included monophylyis not well tested but the study of Civeyrelet al (1998) which included a larger numberof taxa including Pervillea Decne showedthat monophyly of the tribe was supportedThe traditional character for this tribe is fourpollinia per translator
DenitionmdashTribe Secamoneae is the mostinclusive clade in the Asclepiadacina that in-cludes the primary types of Secamone andPervillea but not Asclepioidina or Baissea
BaisseaThe position of Baissea corresponds with
that in the study of Sennblad et al (1998)
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 395
FIGURE 1 Combinable component consensus tree of the 252 most-parsimonious trees from the successiveweighting analysis of the complete rbcL data set Subfamilial and tribal classication is according to Endress andBruyns (2000) For subfamilies a three-letter code is used RAU APO PER SEC and ASC denote RauvoloideaeApocynoideae Periplocoideae Secamonoideae and Asclepiadoideae respectively OUT denotes the outgroupNumbers below branches are unit-weighted bootstrap values successive weighted bootstrap values are indicatedabove branches Branches not present in the strict consensus tree from the unit weight analysis are indicated witha cross (dagger) For branches marked A or B the frequencies in the combinable component consensus are 95 or 50respectively all other branches have 100 frequency
396 SYSTEMATIC BIOLOGY VOL 51
FIGURE 2 Strict consensus tree from the successive weighting analysis of the combined ndhF and rbcL dataset Tribal classication sensu Endress and Bruyns (2000) is indicated Numbers below branches are unit-weightedbootstrap values whereas successive weighted bootstrap values are indicated above branches Branches not presentin the strict consensus tree from the unit weight analysis are indicated with a cross (dagger)
Periplocoideae and of the root of the in-group all well-supported groups in our anal-ysis are congruent with those studies Thecombined analysis using data from rbcL-and matK-sequences and oral and pollenmorphology in the study of the Apocy-naceae sstr by Endress et al (1996) andthe analysis of Sennblad et al (1998) ofthe tribe Wrightieae sensu Leeuwenberg(1994a) and related taxa using rbcL andmorphological data are largely congruentwith the result from the present study Fur-thermore preliminary results from collab-orative studies by Sennblad et al (aimedat the traditional Apocynoideae and As-clepiadaceae pers comm) and Endresset al (aimed at the traditional Alyxieaepers comm) based on combined analy-sis of several molecular and morphologi-cal datasets are also congruent with ourresults
Support is good for the split betweenthe ingroup and the two outgroup taxaGelsemium and Mostuea (uwboot D 96swboot D 100) Thus the initial assump-tion of monophyly of the Apocynaceaesl is not violated Conforming to ear-lier studies (eg Sennblad and Bremer1996 Civeyrel et al 1998 Sennblad et al1998 Potgieter and Albert 2001) the tra-ditional Asclepiadaceae are nested withinthe traditional subfamily Apocynoideae(eg sensu Endress and Bruyns 2000)
The ApocynoideaendashAsclepiadaceae cladewhich is strongly supported (uwboot D 86swboot D 99) is in turn nested within thesubfamily Rauvoloideae sensu Endress andBruyns (2000)
Comparison of our results to the mostrecent classications of the ApocynaceaeEndress and Bruyns (2000 see Table 1 andFig 1) shows that although their circum-scription of the subfamilies PeriplocoideaeSecamonoideae and Asclepiadoideae aremonophyletic in our analysis their sub-families Rauvoloideae and Apocynoideaeare nonmonophyletic Similarly the tribesof the Rauvoloideae except Melodineaeare monophyletic but three of the tribesof the Apocynoideaemdashthe Apocyneae theEchiteae and the Wrightieaemdashare nonmono-phyletic Of the tribes of the Asclepiadoideaethe Ceropegieae and Asclepiadeae are mono-phyletic The position of Fockea as sisterto the rest of the Asclepiadoideae ren-ders the Marsdenieae nonmonophyletic Allthe included Periplocoideae taxa belongto the tribe Periploceae sensu Venter andVerhoeven (1997) except Pentopetia whichthey placed in the tribe CryptolepideaeGiven this unresolved position of Pentopetiathe monophyly of these tribes cannot be eval-uated however the reduction of Parquetinato synonymy with Periploca proposed byVenter and Verhoeven (1997) is not sup-ported by the present study
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 397
DISCUSSION
One of the most important tasks for anomenclature system is to provide a sta-ble means to communicate classications(which we here will understand as represent-ing phylogenetic relationships) Althoughboth the Linnaean and the Phylocode sys-tems are able to communicate phylogeniesthey are sensitive to changes in phylogenetichypotheses which in both systems will de-mand taxonomic changes In addition theLinnaean system is subject to a potentialinstability relating to subjective changes incircumscription that is lumpers and split-ters may refer to different groups of organ-isms by the same name (see de Queiroz andGauthier 1992) For many taxa a consensuscircumscription has eventually been estab-lished but for other groups this is still a prob-lem A current example of this from Apoc-ynaceae is the genus Tabernaemontana sensuLeeuwenberg (1991 1994b) which in theclassication of Allorge (1985) is split intoeight genera The Phylocode system was de-signed to avoid this problem by use of strictdenitions of taxa subjective changes incircumscription are avoided
However this rigidity in denitions leadsto a potential instability and high turnoverof names associated with a group of nestedtaxa Relatively small changes in phyloge-netic views eg when a dening taxonldquospecierrdquo receives a less nested positionmaycause some(or all) of these nested namesto become synonymous (eg de Queirozand Gauthier 1994 Bryant 1996 Liden andOxelman 1996Sereno 1999)The junior syn-onyms should according to the Phylocodebe rejected and new names may then beneeded for the new system of nested clades(de Queiroz and Gauthier 1994) The namesassociated with this group of taxa are thusreplaced by new names This can be espe-cially problematic when careless denitionshave been made as have been noted bySchander and Thollesson (1995) and Cantinoet al (1997) among others Consider a hy-pothetical worst-case example The Ascle-piadaceae were traditionally considered asseparate from Apocynaceae A stem-baseddenition of Apocynaceae reecting thisview could be the largest clade that in-cludes Apocynum but not Asclepias On thepresent tree such a denition would limitthe Apocynaceae to be synonymous with the
tribe Apocyneae and a new name wouldbe needed for the taxon corresponding tothe current view of Apocynaceae (eg sensuEndress and Bruyns 2000) A node-baseddenition would handle this particular prob-lem but may have other problems as illus-trated here with another hypothetical worst-case scenario Early classications of theApocynaceae included Plocosperma BenthA node-based denition representing thisview could have included as reference taxasay Apocynum and Plocosperma The recentlyproposed position of Plocosperma close tothe Boraginaceae (Backlund et al 2000)would with this denition have made Apoc-ynaceae a very large taxon (containing atleast both Gentianales and Boraginales) pos-sibly synonymous with (a hypothetical def-inition of) the informal taxon EuasteridsII (Bremer et al 1998) and again a newname for the Apocynaceae sensu Endressand Bruyns (2000 ie excluding Plocosperma)would have been needed Because theLinnaean system allows recircumscription ofsynonymous names the potential turnoverin the set of names in use is reduced withthis system (albeit at the expense of allowingsubjectivity) Additionally the ranks of theLinnaean system provide a set of names thatwork as relatively stable universal standardnames in communication note that equalranks does not guarantee comparable evo-lutionary units however Continuity in us-age of names is important for example injournals literature databases oras teach-ing and inventory work and the need forsuch continuity has been recognized by pro-ponents of both nomenclature systems (deQueiroz and Gauthier 1992 Greuter et al1994 Reveal 1996)
We propose a compromise approach thatuses solutions relating to these aspects fromboth systems concentrating mainly onbotanical classication of extant plant taxaat the tribal and familial levels and onlyconsidering monophyletic taxa We recom-mend a system that uses ldquostandard namesrdquo(ie universally used communication unitsat convenient hierarchical levels) such asthose provided by the principal ranks of theLinnaean system (Greuter et al 1994) but inwhich sensitivity toward subjective changesin circumscription is reduced One way to dothis is to adopt the denitions of the phy-logenetic system but use constraints similar
398 SYSTEMATIC BIOLOGY VOL 51
to those that apply between ranks in theLinnaean system To reduce sensitivity to-ward changes in phylogenetic hypotheses(Schander and Thollesson 1995 Bryant1996 1997 Cantino et al 1997 Moore 1998)we propose using combined node C stem-based denitions that will provoke explicitincompatibilities between taxa in case ofunfortunate changes in circumscription (deQueiroz 1996 has proposed a similar type ofdenition for designating nonmonophyletictaxa somewhat similar variants are alsodiscussed in the Phylocode) A node-basedpart of the denition provides a minimalcircumscription whereas a stem-based partprovides boundaries to competing taxa(maximal circumscription) If on a particu-lar phylogeny the minimal (monophyletic)circumscriptions of competing taxa are over-lapping then the two taxa are incompatibleTo make this meaningful at least two taxashould be included in the node-based partof the denition this will also avoid unnec-essary monotypic taxa This incompatibilitywill provide an objective point when nomen-claturalreconsideration isneeded Returningto the rst of our examples above node Cstem-based denitions of the Apocynaceaeand Asclepiadaceae could be as follows
The family Apocynaceae is the most inclusive clade inthe order Gentianales including the type specimensof Apocynum and Dictyophleba but not the taxa Ascle-piadaceae [ Gentianaceae Rubiaceae ]
and
The family Asclepiadaceae is the most inclusive cladein the order Gentianales including the type speci-mens of Asclepias and Periploca but not the taxa Apoc-ynaceae [ Gentianaceae Rubiaceae ]
On the tree in Figure 1 these two denitionsare mutually exclusive and a decision onwhich name to keep must be made Mini-mizing the number of classicatory changesshould take precedence in this decision Thereason for this is to avoid the large incon-venient ldquostandard namerdquo changes that couldresult from relatively small rearrangementsRejected names are then ignored when oc-curring in denitions of other taxa This isbecause the reference taxa in the stem-basedpart of the denitions are ldquodenedrdquo taxarather than physical types (thus reliance ondened reference taxa which has been crit-icized for stem-based denitions by Sereno1999 is what we aim for on the other hand
the explicit reference of type specimens in theminimum circumscription provided by thenode-based part of the denitions circum-vents the tautology problem raised byBryant 1996) In the current example thiswould mean that regardless of which nameof Apocynaceae and Asclepiadaceae is re-jected the other will be dened as corre-sponding to the Apocynaceae sensu Endressand Bruyns (2000) The choice will there-fore simply be which name to keep In thiscase because Apocynaceae has priority un-der the Linnaean system we would chooseto keep Apocynaceae to promote consistencywith the current system Nevertheless out-side these recommendations choices of whatnames to choose will probably to a certaindegree be arbitrary We also propose thatthe rejected name should be reinstated withan emended denition if appropriate to po-tentially allow for more stability in the setof names in use Thus in our second ex-ample above a node C stem-based deni-tion reecting the inclusion of Plocospermain Apocynaceae (eg by exchanging Dic-tyophleba for Plocosperma in the denitionabove) would presumably be incongruentwith the denition of Gentianales as well asthe denitions of Rubiaceae Gentianaceaeand so forth and therefore would be rejectedIn such acasewhere the only real change is inthe position of a single taxon a reinstatementof the Apocynaceae with an emended def-inition is motivated However a drawbackis that this may also introduce elements ofsubjectivity in the circumscriptions of taxa
Not all named clades need to be providedwith standard names In fact for the purposeof a universal set used in journals databasesand so forth it may sufce to use standardnames corresponding to the principal ranksof the Linnaean system (Greuter et al 1994)such as species genus family and orderNevertheless to provide compatibility withthe present system we suggest that addi-tional names corresponding to commonlyused secondary ranks (eg tribes) may berecognized The hierarchical level of a stan-dard name needs to be indicated This doesnot imply that taxa of the same hierarchi-cal level are comparable evolutionary unitsbut simply indicates the hierarchical levelof a taxon relative to nested taxa Becausesuch misinterpretations will in any case bedifcult to prohibit one may as well pro-vide compatibility with the present Linnaean
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 399
system by using the names of the Linnaeanranks (eg species genus tribe familyorder) to indicate hierarchical level The prin-ciple of exhaustive subsidiary taxa is notaccepted so redundant taxa need not be rec-ognized Other ldquononstandardrdquo taxa couldbe dened by using Phylocode denitionsmdashthat is without the restraints discussed forstandard taxa above
One advantage of the described systemis that utilizing the correspondence betweenthe standard names and Linnaean ranks al-lows the system to be largely compatiblewiththe establishedLinnaean system A change ofsystem could therefore be gradual with thetwo systems coexisting during a conversiontime (see eg de Quieroz 1997b)
Even though the node C stem-baseddenitions reduce the sensitivity towardschanges in circumscriptions they may still besensitive towardldquounfortunaterdquo or ldquobadrdquo def-initions (see eg Cantino et al 1997Cantinoand de Queiroz 2000) Many of the recom-mendations for such things as choices of ref-erence taxa (speciers) in phylogenetic def-initions are applicable also to the node Cstem-based denitions (eg Schander andThollesson 1995 Bryant 1996 Cantino et al1997 Sereno 1999) Most likely the compro-mise system described above will have sev-eral further problems and we hope that thispaper will invite further discussion on thesubject
A New Classication of the Apocynaceae slOur results indicate that some problem-
atic taxa remain in the classication ofEndress and Bruyns (2000) Monophyly fortwo of the subfamilies and ve of the tribesis called into question If other recent molec-ular studies (eg Potgieter and Albert 2001)are taken into consideration a further fourtribes (Alstonieae Alyxieae Plumerieae andVinceae) may be nonmonophyletic This canin some cases be simply analysis artifactsbut as Endress and Bruyns (2000) themselvespoint out their classication is to be consid-ered preliminary Insufcient taxonsamplingin molecular studies and difcult homologydecisions relating to morphological charac-ters make the circumscription of many ofthe tribes for example in the Apocynoideaesensu Endress and Bruyns (2000) uncertainThere is thus risk for future rearrangementsin many of the tribes A classication using
node C stem-based denitions may be ableto subsume such rearrangements withoutnumerous explicit recircumscriptions Wewill here use the result fromthe present studyas a basis to propose such a classication ofthe Apocynaceae sl
Our aim is to base taxa on clades that arewell supported A second aim is to makeour classication compatible with the cur-rent system We will therefore primarily rec-ognize nonoverlapping tribes that are validunder the Linnaean system We will furtheradopt the principle of nested referencing (assuggested by Lee 1999 and Sereno 1999)from the Linnaean system Thus the rst ref-erence taxon in the node-based part of thedenition is the primary type which corre-sponds to the Linnaean type used for thename of the taxon Instead of the (optional)enumeration of subsumed taxa of a lowerrank (eg genera or subtribes) we will in-clude node C stem-based denitions as dis-cussed above Citations to primary types ofreference taxa in node-based part of de-nitions are also nested and thus refer toldquothe type specimen of the type species ofthe type genus of the dened taxonrdquo Aproblem is that the relationships between thetribes areweakly supported This may lead tovery cumbersome denitions enumeratingall competing tribes in the stem-based partof the denition (eg Moore 1998) One wayto reduce this problem is to dene interme-diate well-supported taxa and use them ascompeting taxaWe therefore will further rec-ognize four nonstandard taxa (note that anassignment of ldquononstandardrdquo taxa does notimply less reliability standard names relateto communication purposes only) The taxadiscussed are indicated in Figure 3 We willdiscuss the tree from the top of Figure 3 start-ing with the taxa of the traditional Plumeri-oideae Tribes and subtribes given in paren-theses in the text refer to the classicationsof Endress and Bruyns (2000) in other casesour classication is implied For some tribeswe have chosen a name with priority underLinnaean system that implies inclusion of agenus not sampled in the present study Forall but one (Willughbeieae) of these casesadditional molecular or morphological phy-logenetic analyses support the inclusion ofthe genus Even if too uncertain denitionsshould be avoided this is less critical withnode C stem-based denitions In the caseof Willughbeieae we have decided from
400 SYSTEMATIC BIOLOGY VOL 51
FIGURE 3 The tribal reclassication discussed in the text indicated on the combinable component consensustree from the successive weighting analysis of the complete rbcL data set Four nonstandard taxamdashApocynoidinaEuapocynoidina Asclepiadacina and Asclepiadoidinamdashare also indicated
discussions with a morphological expert(M Endress pers comm) that morphologi-cal support exists for including WillughbeiaRoxb in Willughbeieae If this is correctthe denition given below will be valid if
it is not our system allows redenition ofthe taxon We have also tried to providecomments on potential morphological char-acters taken from literature for the differ-ent taxa recognized However we have not
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 401
performed any morphological analysis butrely on information external to this studyfor these comments In most cases we cantherefore not identify synapomorphies forthe taxa
TAXONOMIC TREATMENT
Tribus Willughbeieae ADCThe Ancylobotrys Dictyophleba and Va-
hadenia clade is strongly supported (uwbootand swboot D 100) and corresponds to theWillughbeieae of Endress and Bruyns (2000)The present sampling contains closely re-lated genera that have been traditionallykept together Landolphia P Beauv sl in-cludes all these taxa (eg Schumann 1895)In addition to the taxa sampled here thestudy of Potgieter and Albert (2001) sup-ports the inclusion of Couma Aubl LacmelleaH Karst Pacouria Aubl and Saba (Pichon)Pichon in this clade Plants in this cladeall have a gynoecium that is congenitallysyncarpous and indehiscent and that con-tains a eshy pulp including numerous seedswith copious horny endosperm (Fallen 1986Persoon et al 1992) This probably con-stitutes synapomorphies for this clade(M Endress pers comm) although not un-ambiguously for example syncarpy is par-alleled in Carisseae We will therefore as-sume that the reference genus Willughbeia isincluded in this tribe
DenitionmdashTribe Willughbeieae is themost inclusive clade that includes the prim-ary types of Willughbeia and Dictyophleba butnot Alyxieae Aspidospermeae CarisseaeHunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Apocynoidina
Tribus Vinceae Bartl
Ochrosia Rauvola Vinca and Catharanthus(all Vinceae) form a strongly supported clade(uwboot D 89 swboot D 99) The associa-tion between Rauvola and Catharanthus wasindicated by Sennblad and Bremer (1996)The study of Potgieter and Albert (2001) sup-ports inclusion of the genera LaxoplumeriaMarkgr Neisosperma Raf and Tonduzia Pit-tier previously classied with AspidospermaThe Vinceae contain both herbaceous andfruticose taxa The fruits are apocarpousbut whereas the fruits of Catharanthus Lax-oplumeria Neisosperma Tonduzia and Vincahave dry fruit walls the fruits of Ochrosia andRauvola are drupes Plants in Vinceae share
a similar type of style head (ie the apicalenlarged part of the style) with a stigmatichollow and upper and lower hair wreaths(except in some species of Ochrosia) the pres-ence of a nectar disk and a linear hilum onthe seed (Pichon 1948b) On the present treeVinceae also include Kopsia however this po-sition of Kopsia is very weakly supported (seealso Sennblad and Bremer 1996)
DenitionmdashTribe Vinceae is the mostinclusive clade that includes the primarytypes of Vinca and Rauvola but not AlyxieaeAspidospermeae Carisseae HunterieaeMelodineae Plumerieae Tabernaemon-taneae Willughbeieae or Apocynoidina
Tribus Tabernaemontaneae G DonThe strongly supported association be-
tween Tabernaemontana Tabernanthe Carval-hoa Schizozygia and Molongum (Tabernae-montaneae uwboot D 84 swboot D 98)has been suggested previously (Boiteau et al1978 Fallen 1986 Endress et al 1996Sennblad and Bremer 1996) Inclusion of taxatraditionally associated with Tabernaemon-tana (eg Voacanga Thouars and CallichiliaStapf see also Tabernaemontaneae sensuLeeuwenberg 1994a) and Macoubea Aublis supported by the analysis of Potgieterand Albert (2001) The taxa share sclerenchy-matic anthers free from the style head anda characteristic band of heavily cutinizedcells at the insertion on the staminal rib(Endress et al 1996) The taxa tradition-ally associated with Tabernaemontana arecharacterized by their apocarpous fruitswith arillate seeds whereas genera asso-ciated with Molongum (Ambelanieae sensuLeeuwenberg 1994a) have syncarpous fruitsand lackanarillus HoweverMacoubea formsa link between the two having a syncar-pous fruit with arillate seeds (Zarucchi et al1995)
DenitionmdashTribe Tabernaemontaneae isthe most inclusive clade that includes the pri-mary types of Tabernaemontana and Schizozy-gia but not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae PlumerieaeVinceae Willughbeieae or Apocynoidina
Tribus Aspidospermeae Miers
Aspidosperma and Vallesia (Aspidosper-meae) form a strongly supported clade (uw-boot and swboot D 100) rst suggestedin Sennblad (1997 see also Sennblad and
402 SYSTEMATIC BIOLOGY VOL 51
Bremer 2000) Subsequently GeissospermumAllemao Haplophyton ADC MicroplumeriaBaill and Strempeliopsis Benth were indi-cated to belong to this clade (Potgieter andAlbert 2001) The Aspidospermeae havesimple style heads lacking both stigmatichollow and upper hair wreath and haveapocarpous fruits with variation in eshi-ness and sclerication Alstonia has tradi-tionally been placed in the Plumerieae Be-cause its seeds have a hairy margin it hasbeen suggested to form a link to subfam-ily Apocynoideae This was contradicted bySennblad and Bremerrsquos study (1996) how-ever where Alstonia was placed in an iso-lated position as the sister group to therest of the Apocynaceae sl Here it groupswith Aspidospermeae Sensu Endress andBruyns (2000) However this association isvery weakly supported (uwboot and swbootlt 50) and in other recent studies As-pidospermeae and Alstonia do not form aclade (Potgieter and Albert 2001) Becauseof this we do not use it as a reference taxonin the denition Thus although the Aspi-dospermeae include Alstonia on the presenttree some other position of Alstonia canbe accommodated without amending thedenition
DenitionmdashTribe Aspidospermeae is themost inclusive clade that includes theprimary types of Aspidosperma and Vallesiabut not Alyxieae Carisseae HunterieaeMelodineae Plumerieae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Melodineae G DonThe well-supported association between
Melodinus and Craspidospermum (uwboot D70 swboot D 75) was rst reportedby Sennblad (1997 see also Sennblad andBremer 2000) and has subsequently beenreected in the tribe Melodineae sensuEndress and Bruyns (2000 their inclusion ofDiplorhynchus however is not supported)Craspidospermum although its fruit is dryand dehiscent as opposed to the indehis-cent fruits of Melodinus was noted by Pichon(1948b) to present several characters of theCarisseae such as stamens inserted near thebase of the corolla tube and a syncarpousovary In particular Pichon found the verydense suprastaminal indumentum type rem-iniscent of Melodinus From his descriptions
of the two genera (Pichon 1948ab) the fol-lowing similarities also emerge presence ofa stipular line pollen in tetrads and a punc-tiform hilum
DenitionmdashTribe Melodineae is the mostinclusive clade that includes the pri-mary types of Melodinus and Craspidosper-mum but not Alyxieae AspidospermeaeCarisseae Hunterieae Plumerieae Taber-naemontaneae Vinceae Willughbeieae orApocynoidina
Tribus Hunterieae K Schum
The Picralima and Pleiocarpa clade isstrongly supported (uwboot and swboot D100) and corresponds to the subtribeHunterieae sensu Endress andBruyns (2000)The study of Potgieter and Albert (2001)also supports the inclusion of Hunteria RoxbThese taxa are characterized by apocar-pous sometimes pluricarpous ovaries andeshy brous fruit walls (Omino 1996)Diplorhynchus (Melodineae) is here weaklyassociated (uwboot and swboot lt 50) withthe Pleiocarpeae It shares a few similari-ties such as stipular lines and a style headwithout a stigmatic hollow and hair wreathsbut also has differences such as a dry de-hiscent fruit With the present denitionDiplorhynchus will tentatively be included inthe Hunterieae
DenitionmdashTribe Hunterieae is the mostinclusive clade that includes the primarytypes of Hunteria and Picralima but not Alyx-ieae Aspidospermeae Carisseae Melod-ineae Plumerieae TabernaemontaneaeVinceae Willughbeieae or Apocynoidina
Tribus Plumerieae EndlThe strongly supported relationship
(uwboot and swboot D 100) between Alla-manda and Plumeria corresponds to earlierstudies (Endress et al 1996 Sennblad andBremer 1996 Civeyrel et al 1998 Potgieterand Albert 2001) In all earlier treatmentsAllamanda has had uncertain relationshipsThe association with Plumeria nds supportin pollen morphology such as perforatemesocolpial depressions and similar innerexine pattern The two genera also containsecoiridoids rather than the indole alka-loids and cardenolides that are commonin the traditional Plumerioideae (Endresset al 1996) Cameraria Cerbera and Thevetiaform a strongly supported association
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 403
(uwboot D 83 swboot D 91) with Ane-chites which has been suggested by Fallen(1983) This association was based mainlyon the latrorsesublatrorse anthers andthe broad style head with large apicalappendages and stigmatic hollow The studyby Potgieter and Albert (2001) suggesteda further inclusion of Cerberiopsis Viell ExPancher amp Sebert and Skytanthus Meyen inthis tribe These taxa form a well-supportedclade with Plumeria and Allamanda whichcorresponds to the Plumerieae sensuEndress and Bruyns (2000) Characterssupporting this relationship are the presenceof infrastaminal as well as suprastaminalappendages (not present in Anechites andPlumeria) and winged seeds (not present inAnechites) (Endress et al 1996)
DenitionmdashTribe Plumerieae is themost inclusive clade that includes theprimary types of Plumeria and Allamandabut not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Carisseae Dumort
Acokanthera and Carissa (both Carisseae)are traditionally joined and are sometimeseven treated as one genus (eg Pichon1948a) In the present study they form astrongly supported clade (uwboot D 96swboot D 99) Among the characters sup-porting this tribe are syncarpous fruits with-out axile placentation and endocracks on theinner exine of the pollen (Endress et al 1996)
DenitionmdashTribe Carisseae is the most in-clusive clade that includes the primary typesof Carissa and Acokanthera but not AlyxieaeAspidospermeae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
Tribus Alyxieae G Don
Chilocarpus has been a genus with un-certain afnities In Endress et al (1996) itwas the sister taxon to the ApocynoideaeHere it is associated with Alyxia and Lep-inia in a clade corresponding to the Alyxieaesensu Endress and Bruyns (2000) Thisclade is strongly supported with succes-sive weighted bootstrap (swboot D 91) butnot with unit-weighted bootstrap support(uwboot D 58) such clades will henceforthbe termed moderatelywell supported clades
A further inclusion of Condylocarpon DesfLepiniopsis Valeton and Plectaneia Thouarswas indicated by Potgieter and Albert (2001)The constituent taxa are the only Apocy-naceae taxa outside the Apocynoidina thathave porate pollen
DenitionmdashTribe Alyxieae is the most in-clusive clade that includes the primary typesof Alyxia and Lepinia but not Aspidosper-meae Carisseae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
APOCYNOIDINA
On the present tree the traditionalsubfam-ily Apocynoideae includes all subfamilies(Periplocoideae Secamonoideae and Ascle-piadoideae) of the Asclepiadaceae sensuLiede and Albers (1994) All of these taxaare established names and selecting one asa subfamily at the expense of the othersmight be considered unfortunate from dif-ferent points of view We have here chosennot to recognize subfamilies instead we willdene the traditional Apocynoideae Ascle-piadaceae and Asclepiadoideae as nonstan-dard taxa The correspondence to the tradi-tional taxa will be indicated in the names byreplacing the sufx -eae with a neutral suf-x -ina (Kron 1997) We will also for practi-cal reasonsrecognize as a nonstandardtaxonthe informal group euapocynoids suggestedby Sennblad et al (1998)
The taxa of the traditional Apocynoideaeand Asclepiadaceae (uwboot D 86 swboot D99) clade will in many cases be more dif-cult to safely delimit to tribes becausemany of the groups in this clade have weakor no support (although preliminary resultsfrom an unpublished collaborative study bySennblad et al indicate additional supportfor the tribes discussed below) The tribaland subtribal classication of the traditionalApocynoideae has been shown to be prob-lematic (Leeuwenberg 1994a Endress andBruyns 2000) The circumscription of tribesfor these taxa might therefore in some casesbe preliminary
In all following taxa the basal part of theanther connective called the retinacle is ad-nate to the style head
DenitionmdashApocynoidina are the mostinclusive clade that includes the pri-mary types of Apocynum and Wrightia butnot Alyxieae Aspidospermeae Carisseae
404 SYSTEMATIC BIOLOGY VOL 51
Hunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Willughbeieae
Tribus Wrightieae G DonThe Stephanostema and Wrightia (both
Wrightieae) clade is strongly supported(uwboot and swboot D 100) and correspondsto the Wrightieae sensu Sennblad et al (1998)who also suggested an inclusion of Pleio-ceras Baill The tribe could be characterizedby a combination of synapomorphic and ple-siomorphic characters for example presenceofa chalazalandabsence of micropylar comaleft contorted aestivation and absence of airspaces in the anthers (1998)
DenitionmdashTribe Wrightieae is the mostinclusive clade in the Apocynoidina that in-cludes the primary types of Wrightia andStephanostema but not Malouetieae Nerieaeor Euapocynoidina
Tribus Nerieae (Benth) M Pichon
Adenium Nerium and Strophanthus (allWrightieae) constituted the Nerieae inSennblad et al (1998) however the cladereceived weak support Here Adenium andNerium form a clade that is absent in halfof the most-parsimonious trees in the com-plete rbcL analysis but is strongly supportedin the combined ndhF and rbcL analysis(uwboot D 98 swboot D 99) Strophan-thus groups with the Mascarenhasia Pachy-podium Kibatalia Funtumia and Holarrhenaclade which corresponds to the Malouetieaesensu Sennblad et al (1998) This position ofStrophanthus is however weakly supported(uwboot and swboot lt 50) and is not presentin the combined ndhF and rbcL tree Fur-thermore certain morphological characters(unfused slits in the corolla tube and anapical anther appendage) indicate a posi-tion closer to the Nerieae (Sennblad et al1998)
DenitionmdashTribe Nerieae is the most in-clusive clade in the Apocynoidina that in-cludes the primary types of Nerium and Ade-nium but not Malouetieae Wrightieae orEuapocynoidina
Tribus Malouetieae Mull-Arg
The association of Mascarenhasia Pachy-podium Kibatalia Funtumia and Holar-rhena (all Wrightieae) is moderately wellsupported in the complete rbcL analysis
(uwboot D 57 swboot D 74) whereasthe corresponding clade is well supportedin the combined rbcL and ndhF analysis(uwboot D 66 swboot D 100) On the ba-sis of identied synapomorphies (presenceof calcium oxalate crystals in the stomium ofthe anthers absence of interpetal vein ses-sile laments and absence of air spaces inthe anthers) Sennblad et al (1998) predictedan inclusion of Alaa Thouars Kibatalia andMalouetia ADC in the tribe The inclusion ofKibatalia is strongly supported in this studyFollowing these suggestions we will here as-sume the inclusion of the reference genusMalouetia in this tribe (The denition belowwill include Strophanthus on the present treebut see comments under Nerieae)
DenitionmdashTribe Malouetieae is the mostinclusive clade in the Apocynoidina that in-cluded the primary types of Malouetia andFuntumia but not Nerieae Wrightieae orEuapocynoidina
EUAPOCYNOIDINA
In all the remaining representatives ofthe Apocynoideae sensu Endress andBruyns (2000) that is Mandevilla MesechitesAganosma Apocynum TrachelospermumRhabdadenia Beaumontia Parsonsia andPrestonia the anthers are adnate to the stylehead both by the retinacle and by the thecaeThese taxa together with the taxa of the tra-ditional Asclepiadaceae form a moderatelysupported clade (uwboot lt 50 swboot D86) that corresponds to the informal taxoneuapocynoids sensu Sennblad et al (1998)
DenitionmdashEuapocynoidina are the mostinclusive clade that includes the primarytypes of Apocynum and Echites but notMalouetieae Wrightieae or Nerieae
Tribus Mesechiteae Miers
Mandevilla and Mesechites (both Mese-chiteae) form a moderately supported clade(uwboot lt 50 swboot D 87) equivalentto the Mesechiteae This tribe could be char-acterized by anthers with obtuse truncatedtails and a retinacle lacking hairs and also astrongly pentagonal style head with a stig-matic hollow
DenitionmdashTribe Mesechiteae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Mesechitesand Mandevilla but not Apocyneae EchiteaePeriploceae or Asclepiadacina
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 405
Tribus Apocyneae Bercht et J Presl
Aganosma Apocynum Beaumontia andTrachelospermum (Apocyneae) form a weaklysupported clade with Rhabdadenia (Echiteae)(uwboot and swboot lt 50) in the completerbcL analysis whereas the correspondingclade in the combined ndhF and rbcL analy-sis gains a little better support (uwboot lt 50swboot D 91) Likewise this clade (with theinclusion of Chonemorpha G Don) is weaklysupported in Potgieter and Albert (2001) Po-tential characterizing traits could be a stylehead without stigmatic hollow in combina-tion with thecae that are adnate to the stylehead (Sennblad et al 1998)
DenitionmdashTribe Apocyneae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Apocynumand Trachelospermum but not Echiteae Mese-chiteae Periploceae or Asclepiadacina
Tribus Echiteae Bartl
Parsonsia Peltastes and Prestonia (allEchiteae) form a moderately supported orunsupported clade with the Periplocoideaerepresentatives of the study (uwboot andswboot lt 50 and uwboot lt 50 swboot D89 in the complete rbcL and the combinedrbcL and ndhF analysis respectively) In thestudy of Sennblad et al (1998) Prestonia andParsonsia formed a weakly supported cladeand in the study of Potgieter and Albert(2001) inclusion of both Parsonsia and EchitesP Browne in the Echiteae is supportedParsonsia and Echiteae also share certaincharacters such as a horseshoe-shaped reti-nacle and a style head with stigmatic hollow(see Sennblad et al 1998)
DenitionmdashTribe Echiteae is the most in-clusive clade in the Euapocynoidina thatincludes the primary types of Echites andPrestonia but not Apocyneae MesechiteaePeriploceae or Asclepiadacina
Tribus Periploceae BartlThe representatives of the Periplo-
coideae sensu Endress and Bruyns (2000)mdashPentopetia Parquetina Petopentia PeriplocaTacazzea and Mondiamdashgroup together ina well-supported clade (uwboot D 67swboot D 93) The monophyly of the tra-ditional periplocoid taxa is also supportedin earlier studies that included a wider ordifferent sampling (Civeyrel et al 1998
Potgieter and Albert 2001) The taxa ofthe traditional Asclepiadaceae have ldquotrans-latorsrdquo structures related to pollinationspecialization In the Periploceae the pollenis deposited as tetrads in a spoon-liketranslator with an adhesive disc
DenitionmdashTribe Periploceae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Periplocaand Pentopetia but not Apocyneae EchiteaeMesechiteae or Asclepiadacina
ASCLEPIADACINA
The taxa of the Asclepiadoideae sensuLiede and Albers (1994) form a clade togetherwith the two Secamone species and Baissea(Apocyneae uwboot lt 50 swboot D 73)The exclusion of Periploceae (traditionallyassociated with this clade see Civeyrel et al1998) from this taxon is weakly supportedTherefore Periploceae is not included as areference taxon in the stem-based part ofthe denition of the Asclepiadacina thusleaving open the possibility for a later in-clusion of this taxon In the Asclepiadacina(except Baissea) the pollen is agglutinatedinto pollinia that are connected to a claspingtranslator
DenitionmdashAsclepiadacina are the mostinclusive monophyletic clade in the Euapoc-ynoidina that includes the primary types ofAsclepias and Secamoneae but not ApocyneaeEchiteae or Mesechiteae
Tribus Secamoneae G DonThe tribe Secamoneae sensu Bruyns
(uwboot D 89 swboot D 96) is mono-phyletic in the present analysisHowever be-cause representatives from only one of theconstituent genera are included monophylyis not well tested but the study of Civeyrelet al (1998) which included a larger numberof taxa including Pervillea Decne showedthat monophyly of the tribe was supportedThe traditional character for this tribe is fourpollinia per translator
DenitionmdashTribe Secamoneae is the mostinclusive clade in the Asclepiadacina that in-cludes the primary types of Secamone andPervillea but not Asclepioidina or Baissea
BaisseaThe position of Baissea corresponds with
that in the study of Sennblad et al (1998)
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
396 SYSTEMATIC BIOLOGY VOL 51
FIGURE 2 Strict consensus tree from the successive weighting analysis of the combined ndhF and rbcL dataset Tribal classication sensu Endress and Bruyns (2000) is indicated Numbers below branches are unit-weightedbootstrap values whereas successive weighted bootstrap values are indicated above branches Branches not presentin the strict consensus tree from the unit weight analysis are indicated with a cross (dagger)
Periplocoideae and of the root of the in-group all well-supported groups in our anal-ysis are congruent with those studies Thecombined analysis using data from rbcL-and matK-sequences and oral and pollenmorphology in the study of the Apocy-naceae sstr by Endress et al (1996) andthe analysis of Sennblad et al (1998) ofthe tribe Wrightieae sensu Leeuwenberg(1994a) and related taxa using rbcL andmorphological data are largely congruentwith the result from the present study Fur-thermore preliminary results from collab-orative studies by Sennblad et al (aimedat the traditional Apocynoideae and As-clepiadaceae pers comm) and Endresset al (aimed at the traditional Alyxieaepers comm) based on combined analy-sis of several molecular and morphologi-cal datasets are also congruent with ourresults
Support is good for the split betweenthe ingroup and the two outgroup taxaGelsemium and Mostuea (uwboot D 96swboot D 100) Thus the initial assump-tion of monophyly of the Apocynaceaesl is not violated Conforming to ear-lier studies (eg Sennblad and Bremer1996 Civeyrel et al 1998 Sennblad et al1998 Potgieter and Albert 2001) the tra-ditional Asclepiadaceae are nested withinthe traditional subfamily Apocynoideae(eg sensu Endress and Bruyns 2000)
The ApocynoideaendashAsclepiadaceae cladewhich is strongly supported (uwboot D 86swboot D 99) is in turn nested within thesubfamily Rauvoloideae sensu Endress andBruyns (2000)
Comparison of our results to the mostrecent classications of the ApocynaceaeEndress and Bruyns (2000 see Table 1 andFig 1) shows that although their circum-scription of the subfamilies PeriplocoideaeSecamonoideae and Asclepiadoideae aremonophyletic in our analysis their sub-families Rauvoloideae and Apocynoideaeare nonmonophyletic Similarly the tribesof the Rauvoloideae except Melodineaeare monophyletic but three of the tribesof the Apocynoideaemdashthe Apocyneae theEchiteae and the Wrightieaemdashare nonmono-phyletic Of the tribes of the Asclepiadoideaethe Ceropegieae and Asclepiadeae are mono-phyletic The position of Fockea as sisterto the rest of the Asclepiadoideae ren-ders the Marsdenieae nonmonophyletic Allthe included Periplocoideae taxa belongto the tribe Periploceae sensu Venter andVerhoeven (1997) except Pentopetia whichthey placed in the tribe CryptolepideaeGiven this unresolved position of Pentopetiathe monophyly of these tribes cannot be eval-uated however the reduction of Parquetinato synonymy with Periploca proposed byVenter and Verhoeven (1997) is not sup-ported by the present study
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 397
DISCUSSION
One of the most important tasks for anomenclature system is to provide a sta-ble means to communicate classications(which we here will understand as represent-ing phylogenetic relationships) Althoughboth the Linnaean and the Phylocode sys-tems are able to communicate phylogeniesthey are sensitive to changes in phylogenetichypotheses which in both systems will de-mand taxonomic changes In addition theLinnaean system is subject to a potentialinstability relating to subjective changes incircumscription that is lumpers and split-ters may refer to different groups of organ-isms by the same name (see de Queiroz andGauthier 1992) For many taxa a consensuscircumscription has eventually been estab-lished but for other groups this is still a prob-lem A current example of this from Apoc-ynaceae is the genus Tabernaemontana sensuLeeuwenberg (1991 1994b) which in theclassication of Allorge (1985) is split intoeight genera The Phylocode system was de-signed to avoid this problem by use of strictdenitions of taxa subjective changes incircumscription are avoided
However this rigidity in denitions leadsto a potential instability and high turnoverof names associated with a group of nestedtaxa Relatively small changes in phyloge-netic views eg when a dening taxonldquospecierrdquo receives a less nested positionmaycause some(or all) of these nested namesto become synonymous (eg de Queirozand Gauthier 1994 Bryant 1996 Liden andOxelman 1996Sereno 1999)The junior syn-onyms should according to the Phylocodebe rejected and new names may then beneeded for the new system of nested clades(de Queiroz and Gauthier 1994) The namesassociated with this group of taxa are thusreplaced by new names This can be espe-cially problematic when careless denitionshave been made as have been noted bySchander and Thollesson (1995) and Cantinoet al (1997) among others Consider a hy-pothetical worst-case example The Ascle-piadaceae were traditionally considered asseparate from Apocynaceae A stem-baseddenition of Apocynaceae reecting thisview could be the largest clade that in-cludes Apocynum but not Asclepias On thepresent tree such a denition would limitthe Apocynaceae to be synonymous with the
tribe Apocyneae and a new name wouldbe needed for the taxon corresponding tothe current view of Apocynaceae (eg sensuEndress and Bruyns 2000) A node-baseddenition would handle this particular prob-lem but may have other problems as illus-trated here with another hypothetical worst-case scenario Early classications of theApocynaceae included Plocosperma BenthA node-based denition representing thisview could have included as reference taxasay Apocynum and Plocosperma The recentlyproposed position of Plocosperma close tothe Boraginaceae (Backlund et al 2000)would with this denition have made Apoc-ynaceae a very large taxon (containing atleast both Gentianales and Boraginales) pos-sibly synonymous with (a hypothetical def-inition of) the informal taxon EuasteridsII (Bremer et al 1998) and again a newname for the Apocynaceae sensu Endressand Bruyns (2000 ie excluding Plocosperma)would have been needed Because theLinnaean system allows recircumscription ofsynonymous names the potential turnoverin the set of names in use is reduced withthis system (albeit at the expense of allowingsubjectivity) Additionally the ranks of theLinnaean system provide a set of names thatwork as relatively stable universal standardnames in communication note that equalranks does not guarantee comparable evo-lutionary units however Continuity in us-age of names is important for example injournals literature databases oras teach-ing and inventory work and the need forsuch continuity has been recognized by pro-ponents of both nomenclature systems (deQueiroz and Gauthier 1992 Greuter et al1994 Reveal 1996)
We propose a compromise approach thatuses solutions relating to these aspects fromboth systems concentrating mainly onbotanical classication of extant plant taxaat the tribal and familial levels and onlyconsidering monophyletic taxa We recom-mend a system that uses ldquostandard namesrdquo(ie universally used communication unitsat convenient hierarchical levels) such asthose provided by the principal ranks of theLinnaean system (Greuter et al 1994) but inwhich sensitivity toward subjective changesin circumscription is reduced One way to dothis is to adopt the denitions of the phy-logenetic system but use constraints similar
398 SYSTEMATIC BIOLOGY VOL 51
to those that apply between ranks in theLinnaean system To reduce sensitivity to-ward changes in phylogenetic hypotheses(Schander and Thollesson 1995 Bryant1996 1997 Cantino et al 1997 Moore 1998)we propose using combined node C stem-based denitions that will provoke explicitincompatibilities between taxa in case ofunfortunate changes in circumscription (deQueiroz 1996 has proposed a similar type ofdenition for designating nonmonophyletictaxa somewhat similar variants are alsodiscussed in the Phylocode) A node-basedpart of the denition provides a minimalcircumscription whereas a stem-based partprovides boundaries to competing taxa(maximal circumscription) If on a particu-lar phylogeny the minimal (monophyletic)circumscriptions of competing taxa are over-lapping then the two taxa are incompatibleTo make this meaningful at least two taxashould be included in the node-based partof the denition this will also avoid unnec-essary monotypic taxa This incompatibilitywill provide an objective point when nomen-claturalreconsideration isneeded Returningto the rst of our examples above node Cstem-based denitions of the Apocynaceaeand Asclepiadaceae could be as follows
The family Apocynaceae is the most inclusive clade inthe order Gentianales including the type specimensof Apocynum and Dictyophleba but not the taxa Ascle-piadaceae [ Gentianaceae Rubiaceae ]
and
The family Asclepiadaceae is the most inclusive cladein the order Gentianales including the type speci-mens of Asclepias and Periploca but not the taxa Apoc-ynaceae [ Gentianaceae Rubiaceae ]
On the tree in Figure 1 these two denitionsare mutually exclusive and a decision onwhich name to keep must be made Mini-mizing the number of classicatory changesshould take precedence in this decision Thereason for this is to avoid the large incon-venient ldquostandard namerdquo changes that couldresult from relatively small rearrangementsRejected names are then ignored when oc-curring in denitions of other taxa This isbecause the reference taxa in the stem-basedpart of the denitions are ldquodenedrdquo taxarather than physical types (thus reliance ondened reference taxa which has been crit-icized for stem-based denitions by Sereno1999 is what we aim for on the other hand
the explicit reference of type specimens in theminimum circumscription provided by thenode-based part of the denitions circum-vents the tautology problem raised byBryant 1996) In the current example thiswould mean that regardless of which nameof Apocynaceae and Asclepiadaceae is re-jected the other will be dened as corre-sponding to the Apocynaceae sensu Endressand Bruyns (2000) The choice will there-fore simply be which name to keep In thiscase because Apocynaceae has priority un-der the Linnaean system we would chooseto keep Apocynaceae to promote consistencywith the current system Nevertheless out-side these recommendations choices of whatnames to choose will probably to a certaindegree be arbitrary We also propose thatthe rejected name should be reinstated withan emended denition if appropriate to po-tentially allow for more stability in the setof names in use Thus in our second ex-ample above a node C stem-based deni-tion reecting the inclusion of Plocospermain Apocynaceae (eg by exchanging Dic-tyophleba for Plocosperma in the denitionabove) would presumably be incongruentwith the denition of Gentianales as well asthe denitions of Rubiaceae Gentianaceaeand so forth and therefore would be rejectedIn such acasewhere the only real change is inthe position of a single taxon a reinstatementof the Apocynaceae with an emended def-inition is motivated However a drawbackis that this may also introduce elements ofsubjectivity in the circumscriptions of taxa
Not all named clades need to be providedwith standard names In fact for the purposeof a universal set used in journals databasesand so forth it may sufce to use standardnames corresponding to the principal ranksof the Linnaean system (Greuter et al 1994)such as species genus family and orderNevertheless to provide compatibility withthe present system we suggest that addi-tional names corresponding to commonlyused secondary ranks (eg tribes) may berecognized The hierarchical level of a stan-dard name needs to be indicated This doesnot imply that taxa of the same hierarchi-cal level are comparable evolutionary unitsbut simply indicates the hierarchical levelof a taxon relative to nested taxa Becausesuch misinterpretations will in any case bedifcult to prohibit one may as well pro-vide compatibility with the present Linnaean
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 399
system by using the names of the Linnaeanranks (eg species genus tribe familyorder) to indicate hierarchical level The prin-ciple of exhaustive subsidiary taxa is notaccepted so redundant taxa need not be rec-ognized Other ldquononstandardrdquo taxa couldbe dened by using Phylocode denitionsmdashthat is without the restraints discussed forstandard taxa above
One advantage of the described systemis that utilizing the correspondence betweenthe standard names and Linnaean ranks al-lows the system to be largely compatiblewiththe establishedLinnaean system A change ofsystem could therefore be gradual with thetwo systems coexisting during a conversiontime (see eg de Quieroz 1997b)
Even though the node C stem-baseddenitions reduce the sensitivity towardschanges in circumscriptions they may still besensitive towardldquounfortunaterdquo or ldquobadrdquo def-initions (see eg Cantino et al 1997Cantinoand de Queiroz 2000) Many of the recom-mendations for such things as choices of ref-erence taxa (speciers) in phylogenetic def-initions are applicable also to the node Cstem-based denitions (eg Schander andThollesson 1995 Bryant 1996 Cantino et al1997 Sereno 1999) Most likely the compro-mise system described above will have sev-eral further problems and we hope that thispaper will invite further discussion on thesubject
A New Classication of the Apocynaceae slOur results indicate that some problem-
atic taxa remain in the classication ofEndress and Bruyns (2000) Monophyly fortwo of the subfamilies and ve of the tribesis called into question If other recent molec-ular studies (eg Potgieter and Albert 2001)are taken into consideration a further fourtribes (Alstonieae Alyxieae Plumerieae andVinceae) may be nonmonophyletic This canin some cases be simply analysis artifactsbut as Endress and Bruyns (2000) themselvespoint out their classication is to be consid-ered preliminary Insufcient taxonsamplingin molecular studies and difcult homologydecisions relating to morphological charac-ters make the circumscription of many ofthe tribes for example in the Apocynoideaesensu Endress and Bruyns (2000) uncertainThere is thus risk for future rearrangementsin many of the tribes A classication using
node C stem-based denitions may be ableto subsume such rearrangements withoutnumerous explicit recircumscriptions Wewill here use the result fromthe present studyas a basis to propose such a classication ofthe Apocynaceae sl
Our aim is to base taxa on clades that arewell supported A second aim is to makeour classication compatible with the cur-rent system We will therefore primarily rec-ognize nonoverlapping tribes that are validunder the Linnaean system We will furtheradopt the principle of nested referencing (assuggested by Lee 1999 and Sereno 1999)from the Linnaean system Thus the rst ref-erence taxon in the node-based part of thedenition is the primary type which corre-sponds to the Linnaean type used for thename of the taxon Instead of the (optional)enumeration of subsumed taxa of a lowerrank (eg genera or subtribes) we will in-clude node C stem-based denitions as dis-cussed above Citations to primary types ofreference taxa in node-based part of de-nitions are also nested and thus refer toldquothe type specimen of the type species ofthe type genus of the dened taxonrdquo Aproblem is that the relationships between thetribes areweakly supported This may lead tovery cumbersome denitions enumeratingall competing tribes in the stem-based partof the denition (eg Moore 1998) One wayto reduce this problem is to dene interme-diate well-supported taxa and use them ascompeting taxaWe therefore will further rec-ognize four nonstandard taxa (note that anassignment of ldquononstandardrdquo taxa does notimply less reliability standard names relateto communication purposes only) The taxadiscussed are indicated in Figure 3 We willdiscuss the tree from the top of Figure 3 start-ing with the taxa of the traditional Plumeri-oideae Tribes and subtribes given in paren-theses in the text refer to the classicationsof Endress and Bruyns (2000) in other casesour classication is implied For some tribeswe have chosen a name with priority underLinnaean system that implies inclusion of agenus not sampled in the present study Forall but one (Willughbeieae) of these casesadditional molecular or morphological phy-logenetic analyses support the inclusion ofthe genus Even if too uncertain denitionsshould be avoided this is less critical withnode C stem-based denitions In the caseof Willughbeieae we have decided from
400 SYSTEMATIC BIOLOGY VOL 51
FIGURE 3 The tribal reclassication discussed in the text indicated on the combinable component consensustree from the successive weighting analysis of the complete rbcL data set Four nonstandard taxamdashApocynoidinaEuapocynoidina Asclepiadacina and Asclepiadoidinamdashare also indicated
discussions with a morphological expert(M Endress pers comm) that morphologi-cal support exists for including WillughbeiaRoxb in Willughbeieae If this is correctthe denition given below will be valid if
it is not our system allows redenition ofthe taxon We have also tried to providecomments on potential morphological char-acters taken from literature for the differ-ent taxa recognized However we have not
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 401
performed any morphological analysis butrely on information external to this studyfor these comments In most cases we cantherefore not identify synapomorphies forthe taxa
TAXONOMIC TREATMENT
Tribus Willughbeieae ADCThe Ancylobotrys Dictyophleba and Va-
hadenia clade is strongly supported (uwbootand swboot D 100) and corresponds to theWillughbeieae of Endress and Bruyns (2000)The present sampling contains closely re-lated genera that have been traditionallykept together Landolphia P Beauv sl in-cludes all these taxa (eg Schumann 1895)In addition to the taxa sampled here thestudy of Potgieter and Albert (2001) sup-ports the inclusion of Couma Aubl LacmelleaH Karst Pacouria Aubl and Saba (Pichon)Pichon in this clade Plants in this cladeall have a gynoecium that is congenitallysyncarpous and indehiscent and that con-tains a eshy pulp including numerous seedswith copious horny endosperm (Fallen 1986Persoon et al 1992) This probably con-stitutes synapomorphies for this clade(M Endress pers comm) although not un-ambiguously for example syncarpy is par-alleled in Carisseae We will therefore as-sume that the reference genus Willughbeia isincluded in this tribe
DenitionmdashTribe Willughbeieae is themost inclusive clade that includes the prim-ary types of Willughbeia and Dictyophleba butnot Alyxieae Aspidospermeae CarisseaeHunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Apocynoidina
Tribus Vinceae Bartl
Ochrosia Rauvola Vinca and Catharanthus(all Vinceae) form a strongly supported clade(uwboot D 89 swboot D 99) The associa-tion between Rauvola and Catharanthus wasindicated by Sennblad and Bremer (1996)The study of Potgieter and Albert (2001) sup-ports inclusion of the genera LaxoplumeriaMarkgr Neisosperma Raf and Tonduzia Pit-tier previously classied with AspidospermaThe Vinceae contain both herbaceous andfruticose taxa The fruits are apocarpousbut whereas the fruits of Catharanthus Lax-oplumeria Neisosperma Tonduzia and Vincahave dry fruit walls the fruits of Ochrosia andRauvola are drupes Plants in Vinceae share
a similar type of style head (ie the apicalenlarged part of the style) with a stigmatichollow and upper and lower hair wreaths(except in some species of Ochrosia) the pres-ence of a nectar disk and a linear hilum onthe seed (Pichon 1948b) On the present treeVinceae also include Kopsia however this po-sition of Kopsia is very weakly supported (seealso Sennblad and Bremer 1996)
DenitionmdashTribe Vinceae is the mostinclusive clade that includes the primarytypes of Vinca and Rauvola but not AlyxieaeAspidospermeae Carisseae HunterieaeMelodineae Plumerieae Tabernaemon-taneae Willughbeieae or Apocynoidina
Tribus Tabernaemontaneae G DonThe strongly supported association be-
tween Tabernaemontana Tabernanthe Carval-hoa Schizozygia and Molongum (Tabernae-montaneae uwboot D 84 swboot D 98)has been suggested previously (Boiteau et al1978 Fallen 1986 Endress et al 1996Sennblad and Bremer 1996) Inclusion of taxatraditionally associated with Tabernaemon-tana (eg Voacanga Thouars and CallichiliaStapf see also Tabernaemontaneae sensuLeeuwenberg 1994a) and Macoubea Aublis supported by the analysis of Potgieterand Albert (2001) The taxa share sclerenchy-matic anthers free from the style head anda characteristic band of heavily cutinizedcells at the insertion on the staminal rib(Endress et al 1996) The taxa tradition-ally associated with Tabernaemontana arecharacterized by their apocarpous fruitswith arillate seeds whereas genera asso-ciated with Molongum (Ambelanieae sensuLeeuwenberg 1994a) have syncarpous fruitsand lackanarillus HoweverMacoubea formsa link between the two having a syncar-pous fruit with arillate seeds (Zarucchi et al1995)
DenitionmdashTribe Tabernaemontaneae isthe most inclusive clade that includes the pri-mary types of Tabernaemontana and Schizozy-gia but not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae PlumerieaeVinceae Willughbeieae or Apocynoidina
Tribus Aspidospermeae Miers
Aspidosperma and Vallesia (Aspidosper-meae) form a strongly supported clade (uw-boot and swboot D 100) rst suggestedin Sennblad (1997 see also Sennblad and
402 SYSTEMATIC BIOLOGY VOL 51
Bremer 2000) Subsequently GeissospermumAllemao Haplophyton ADC MicroplumeriaBaill and Strempeliopsis Benth were indi-cated to belong to this clade (Potgieter andAlbert 2001) The Aspidospermeae havesimple style heads lacking both stigmatichollow and upper hair wreath and haveapocarpous fruits with variation in eshi-ness and sclerication Alstonia has tradi-tionally been placed in the Plumerieae Be-cause its seeds have a hairy margin it hasbeen suggested to form a link to subfam-ily Apocynoideae This was contradicted bySennblad and Bremerrsquos study (1996) how-ever where Alstonia was placed in an iso-lated position as the sister group to therest of the Apocynaceae sl Here it groupswith Aspidospermeae Sensu Endress andBruyns (2000) However this association isvery weakly supported (uwboot and swbootlt 50) and in other recent studies As-pidospermeae and Alstonia do not form aclade (Potgieter and Albert 2001) Becauseof this we do not use it as a reference taxonin the denition Thus although the Aspi-dospermeae include Alstonia on the presenttree some other position of Alstonia canbe accommodated without amending thedenition
DenitionmdashTribe Aspidospermeae is themost inclusive clade that includes theprimary types of Aspidosperma and Vallesiabut not Alyxieae Carisseae HunterieaeMelodineae Plumerieae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Melodineae G DonThe well-supported association between
Melodinus and Craspidospermum (uwboot D70 swboot D 75) was rst reportedby Sennblad (1997 see also Sennblad andBremer 2000) and has subsequently beenreected in the tribe Melodineae sensuEndress and Bruyns (2000 their inclusion ofDiplorhynchus however is not supported)Craspidospermum although its fruit is dryand dehiscent as opposed to the indehis-cent fruits of Melodinus was noted by Pichon(1948b) to present several characters of theCarisseae such as stamens inserted near thebase of the corolla tube and a syncarpousovary In particular Pichon found the verydense suprastaminal indumentum type rem-iniscent of Melodinus From his descriptions
of the two genera (Pichon 1948ab) the fol-lowing similarities also emerge presence ofa stipular line pollen in tetrads and a punc-tiform hilum
DenitionmdashTribe Melodineae is the mostinclusive clade that includes the pri-mary types of Melodinus and Craspidosper-mum but not Alyxieae AspidospermeaeCarisseae Hunterieae Plumerieae Taber-naemontaneae Vinceae Willughbeieae orApocynoidina
Tribus Hunterieae K Schum
The Picralima and Pleiocarpa clade isstrongly supported (uwboot and swboot D100) and corresponds to the subtribeHunterieae sensu Endress andBruyns (2000)The study of Potgieter and Albert (2001)also supports the inclusion of Hunteria RoxbThese taxa are characterized by apocar-pous sometimes pluricarpous ovaries andeshy brous fruit walls (Omino 1996)Diplorhynchus (Melodineae) is here weaklyassociated (uwboot and swboot lt 50) withthe Pleiocarpeae It shares a few similari-ties such as stipular lines and a style headwithout a stigmatic hollow and hair wreathsbut also has differences such as a dry de-hiscent fruit With the present denitionDiplorhynchus will tentatively be included inthe Hunterieae
DenitionmdashTribe Hunterieae is the mostinclusive clade that includes the primarytypes of Hunteria and Picralima but not Alyx-ieae Aspidospermeae Carisseae Melod-ineae Plumerieae TabernaemontaneaeVinceae Willughbeieae or Apocynoidina
Tribus Plumerieae EndlThe strongly supported relationship
(uwboot and swboot D 100) between Alla-manda and Plumeria corresponds to earlierstudies (Endress et al 1996 Sennblad andBremer 1996 Civeyrel et al 1998 Potgieterand Albert 2001) In all earlier treatmentsAllamanda has had uncertain relationshipsThe association with Plumeria nds supportin pollen morphology such as perforatemesocolpial depressions and similar innerexine pattern The two genera also containsecoiridoids rather than the indole alka-loids and cardenolides that are commonin the traditional Plumerioideae (Endresset al 1996) Cameraria Cerbera and Thevetiaform a strongly supported association
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 403
(uwboot D 83 swboot D 91) with Ane-chites which has been suggested by Fallen(1983) This association was based mainlyon the latrorsesublatrorse anthers andthe broad style head with large apicalappendages and stigmatic hollow The studyby Potgieter and Albert (2001) suggesteda further inclusion of Cerberiopsis Viell ExPancher amp Sebert and Skytanthus Meyen inthis tribe These taxa form a well-supportedclade with Plumeria and Allamanda whichcorresponds to the Plumerieae sensuEndress and Bruyns (2000) Characterssupporting this relationship are the presenceof infrastaminal as well as suprastaminalappendages (not present in Anechites andPlumeria) and winged seeds (not present inAnechites) (Endress et al 1996)
DenitionmdashTribe Plumerieae is themost inclusive clade that includes theprimary types of Plumeria and Allamandabut not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Carisseae Dumort
Acokanthera and Carissa (both Carisseae)are traditionally joined and are sometimeseven treated as one genus (eg Pichon1948a) In the present study they form astrongly supported clade (uwboot D 96swboot D 99) Among the characters sup-porting this tribe are syncarpous fruits with-out axile placentation and endocracks on theinner exine of the pollen (Endress et al 1996)
DenitionmdashTribe Carisseae is the most in-clusive clade that includes the primary typesof Carissa and Acokanthera but not AlyxieaeAspidospermeae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
Tribus Alyxieae G Don
Chilocarpus has been a genus with un-certain afnities In Endress et al (1996) itwas the sister taxon to the ApocynoideaeHere it is associated with Alyxia and Lep-inia in a clade corresponding to the Alyxieaesensu Endress and Bruyns (2000) Thisclade is strongly supported with succes-sive weighted bootstrap (swboot D 91) butnot with unit-weighted bootstrap support(uwboot D 58) such clades will henceforthbe termed moderatelywell supported clades
A further inclusion of Condylocarpon DesfLepiniopsis Valeton and Plectaneia Thouarswas indicated by Potgieter and Albert (2001)The constituent taxa are the only Apocy-naceae taxa outside the Apocynoidina thathave porate pollen
DenitionmdashTribe Alyxieae is the most in-clusive clade that includes the primary typesof Alyxia and Lepinia but not Aspidosper-meae Carisseae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
APOCYNOIDINA
On the present tree the traditionalsubfam-ily Apocynoideae includes all subfamilies(Periplocoideae Secamonoideae and Ascle-piadoideae) of the Asclepiadaceae sensuLiede and Albers (1994) All of these taxaare established names and selecting one asa subfamily at the expense of the othersmight be considered unfortunate from dif-ferent points of view We have here chosennot to recognize subfamilies instead we willdene the traditional Apocynoideae Ascle-piadaceae and Asclepiadoideae as nonstan-dard taxa The correspondence to the tradi-tional taxa will be indicated in the names byreplacing the sufx -eae with a neutral suf-x -ina (Kron 1997) We will also for practi-cal reasonsrecognize as a nonstandardtaxonthe informal group euapocynoids suggestedby Sennblad et al (1998)
The taxa of the traditional Apocynoideaeand Asclepiadaceae (uwboot D 86 swboot D99) clade will in many cases be more dif-cult to safely delimit to tribes becausemany of the groups in this clade have weakor no support (although preliminary resultsfrom an unpublished collaborative study bySennblad et al indicate additional supportfor the tribes discussed below) The tribaland subtribal classication of the traditionalApocynoideae has been shown to be prob-lematic (Leeuwenberg 1994a Endress andBruyns 2000) The circumscription of tribesfor these taxa might therefore in some casesbe preliminary
In all following taxa the basal part of theanther connective called the retinacle is ad-nate to the style head
DenitionmdashApocynoidina are the mostinclusive clade that includes the pri-mary types of Apocynum and Wrightia butnot Alyxieae Aspidospermeae Carisseae
404 SYSTEMATIC BIOLOGY VOL 51
Hunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Willughbeieae
Tribus Wrightieae G DonThe Stephanostema and Wrightia (both
Wrightieae) clade is strongly supported(uwboot and swboot D 100) and correspondsto the Wrightieae sensu Sennblad et al (1998)who also suggested an inclusion of Pleio-ceras Baill The tribe could be characterizedby a combination of synapomorphic and ple-siomorphic characters for example presenceofa chalazalandabsence of micropylar comaleft contorted aestivation and absence of airspaces in the anthers (1998)
DenitionmdashTribe Wrightieae is the mostinclusive clade in the Apocynoidina that in-cludes the primary types of Wrightia andStephanostema but not Malouetieae Nerieaeor Euapocynoidina
Tribus Nerieae (Benth) M Pichon
Adenium Nerium and Strophanthus (allWrightieae) constituted the Nerieae inSennblad et al (1998) however the cladereceived weak support Here Adenium andNerium form a clade that is absent in halfof the most-parsimonious trees in the com-plete rbcL analysis but is strongly supportedin the combined ndhF and rbcL analysis(uwboot D 98 swboot D 99) Strophan-thus groups with the Mascarenhasia Pachy-podium Kibatalia Funtumia and Holarrhenaclade which corresponds to the Malouetieaesensu Sennblad et al (1998) This position ofStrophanthus is however weakly supported(uwboot and swboot lt 50) and is not presentin the combined ndhF and rbcL tree Fur-thermore certain morphological characters(unfused slits in the corolla tube and anapical anther appendage) indicate a posi-tion closer to the Nerieae (Sennblad et al1998)
DenitionmdashTribe Nerieae is the most in-clusive clade in the Apocynoidina that in-cludes the primary types of Nerium and Ade-nium but not Malouetieae Wrightieae orEuapocynoidina
Tribus Malouetieae Mull-Arg
The association of Mascarenhasia Pachy-podium Kibatalia Funtumia and Holar-rhena (all Wrightieae) is moderately wellsupported in the complete rbcL analysis
(uwboot D 57 swboot D 74) whereasthe corresponding clade is well supportedin the combined rbcL and ndhF analysis(uwboot D 66 swboot D 100) On the ba-sis of identied synapomorphies (presenceof calcium oxalate crystals in the stomium ofthe anthers absence of interpetal vein ses-sile laments and absence of air spaces inthe anthers) Sennblad et al (1998) predictedan inclusion of Alaa Thouars Kibatalia andMalouetia ADC in the tribe The inclusion ofKibatalia is strongly supported in this studyFollowing these suggestions we will here as-sume the inclusion of the reference genusMalouetia in this tribe (The denition belowwill include Strophanthus on the present treebut see comments under Nerieae)
DenitionmdashTribe Malouetieae is the mostinclusive clade in the Apocynoidina that in-cluded the primary types of Malouetia andFuntumia but not Nerieae Wrightieae orEuapocynoidina
EUAPOCYNOIDINA
In all the remaining representatives ofthe Apocynoideae sensu Endress andBruyns (2000) that is Mandevilla MesechitesAganosma Apocynum TrachelospermumRhabdadenia Beaumontia Parsonsia andPrestonia the anthers are adnate to the stylehead both by the retinacle and by the thecaeThese taxa together with the taxa of the tra-ditional Asclepiadaceae form a moderatelysupported clade (uwboot lt 50 swboot D86) that corresponds to the informal taxoneuapocynoids sensu Sennblad et al (1998)
DenitionmdashEuapocynoidina are the mostinclusive clade that includes the primarytypes of Apocynum and Echites but notMalouetieae Wrightieae or Nerieae
Tribus Mesechiteae Miers
Mandevilla and Mesechites (both Mese-chiteae) form a moderately supported clade(uwboot lt 50 swboot D 87) equivalentto the Mesechiteae This tribe could be char-acterized by anthers with obtuse truncatedtails and a retinacle lacking hairs and also astrongly pentagonal style head with a stig-matic hollow
DenitionmdashTribe Mesechiteae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Mesechitesand Mandevilla but not Apocyneae EchiteaePeriploceae or Asclepiadacina
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 405
Tribus Apocyneae Bercht et J Presl
Aganosma Apocynum Beaumontia andTrachelospermum (Apocyneae) form a weaklysupported clade with Rhabdadenia (Echiteae)(uwboot and swboot lt 50) in the completerbcL analysis whereas the correspondingclade in the combined ndhF and rbcL analy-sis gains a little better support (uwboot lt 50swboot D 91) Likewise this clade (with theinclusion of Chonemorpha G Don) is weaklysupported in Potgieter and Albert (2001) Po-tential characterizing traits could be a stylehead without stigmatic hollow in combina-tion with thecae that are adnate to the stylehead (Sennblad et al 1998)
DenitionmdashTribe Apocyneae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Apocynumand Trachelospermum but not Echiteae Mese-chiteae Periploceae or Asclepiadacina
Tribus Echiteae Bartl
Parsonsia Peltastes and Prestonia (allEchiteae) form a moderately supported orunsupported clade with the Periplocoideaerepresentatives of the study (uwboot andswboot lt 50 and uwboot lt 50 swboot D89 in the complete rbcL and the combinedrbcL and ndhF analysis respectively) In thestudy of Sennblad et al (1998) Prestonia andParsonsia formed a weakly supported cladeand in the study of Potgieter and Albert(2001) inclusion of both Parsonsia and EchitesP Browne in the Echiteae is supportedParsonsia and Echiteae also share certaincharacters such as a horseshoe-shaped reti-nacle and a style head with stigmatic hollow(see Sennblad et al 1998)
DenitionmdashTribe Echiteae is the most in-clusive clade in the Euapocynoidina thatincludes the primary types of Echites andPrestonia but not Apocyneae MesechiteaePeriploceae or Asclepiadacina
Tribus Periploceae BartlThe representatives of the Periplo-
coideae sensu Endress and Bruyns (2000)mdashPentopetia Parquetina Petopentia PeriplocaTacazzea and Mondiamdashgroup together ina well-supported clade (uwboot D 67swboot D 93) The monophyly of the tra-ditional periplocoid taxa is also supportedin earlier studies that included a wider ordifferent sampling (Civeyrel et al 1998
Potgieter and Albert 2001) The taxa ofthe traditional Asclepiadaceae have ldquotrans-latorsrdquo structures related to pollinationspecialization In the Periploceae the pollenis deposited as tetrads in a spoon-liketranslator with an adhesive disc
DenitionmdashTribe Periploceae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Periplocaand Pentopetia but not Apocyneae EchiteaeMesechiteae or Asclepiadacina
ASCLEPIADACINA
The taxa of the Asclepiadoideae sensuLiede and Albers (1994) form a clade togetherwith the two Secamone species and Baissea(Apocyneae uwboot lt 50 swboot D 73)The exclusion of Periploceae (traditionallyassociated with this clade see Civeyrel et al1998) from this taxon is weakly supportedTherefore Periploceae is not included as areference taxon in the stem-based part ofthe denition of the Asclepiadacina thusleaving open the possibility for a later in-clusion of this taxon In the Asclepiadacina(except Baissea) the pollen is agglutinatedinto pollinia that are connected to a claspingtranslator
DenitionmdashAsclepiadacina are the mostinclusive monophyletic clade in the Euapoc-ynoidina that includes the primary types ofAsclepias and Secamoneae but not ApocyneaeEchiteae or Mesechiteae
Tribus Secamoneae G DonThe tribe Secamoneae sensu Bruyns
(uwboot D 89 swboot D 96) is mono-phyletic in the present analysisHowever be-cause representatives from only one of theconstituent genera are included monophylyis not well tested but the study of Civeyrelet al (1998) which included a larger numberof taxa including Pervillea Decne showedthat monophyly of the tribe was supportedThe traditional character for this tribe is fourpollinia per translator
DenitionmdashTribe Secamoneae is the mostinclusive clade in the Asclepiadacina that in-cludes the primary types of Secamone andPervillea but not Asclepioidina or Baissea
BaisseaThe position of Baissea corresponds with
that in the study of Sennblad et al (1998)
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 397
DISCUSSION
One of the most important tasks for anomenclature system is to provide a sta-ble means to communicate classications(which we here will understand as represent-ing phylogenetic relationships) Althoughboth the Linnaean and the Phylocode sys-tems are able to communicate phylogeniesthey are sensitive to changes in phylogenetichypotheses which in both systems will de-mand taxonomic changes In addition theLinnaean system is subject to a potentialinstability relating to subjective changes incircumscription that is lumpers and split-ters may refer to different groups of organ-isms by the same name (see de Queiroz andGauthier 1992) For many taxa a consensuscircumscription has eventually been estab-lished but for other groups this is still a prob-lem A current example of this from Apoc-ynaceae is the genus Tabernaemontana sensuLeeuwenberg (1991 1994b) which in theclassication of Allorge (1985) is split intoeight genera The Phylocode system was de-signed to avoid this problem by use of strictdenitions of taxa subjective changes incircumscription are avoided
However this rigidity in denitions leadsto a potential instability and high turnoverof names associated with a group of nestedtaxa Relatively small changes in phyloge-netic views eg when a dening taxonldquospecierrdquo receives a less nested positionmaycause some(or all) of these nested namesto become synonymous (eg de Queirozand Gauthier 1994 Bryant 1996 Liden andOxelman 1996Sereno 1999)The junior syn-onyms should according to the Phylocodebe rejected and new names may then beneeded for the new system of nested clades(de Queiroz and Gauthier 1994) The namesassociated with this group of taxa are thusreplaced by new names This can be espe-cially problematic when careless denitionshave been made as have been noted bySchander and Thollesson (1995) and Cantinoet al (1997) among others Consider a hy-pothetical worst-case example The Ascle-piadaceae were traditionally considered asseparate from Apocynaceae A stem-baseddenition of Apocynaceae reecting thisview could be the largest clade that in-cludes Apocynum but not Asclepias On thepresent tree such a denition would limitthe Apocynaceae to be synonymous with the
tribe Apocyneae and a new name wouldbe needed for the taxon corresponding tothe current view of Apocynaceae (eg sensuEndress and Bruyns 2000) A node-baseddenition would handle this particular prob-lem but may have other problems as illus-trated here with another hypothetical worst-case scenario Early classications of theApocynaceae included Plocosperma BenthA node-based denition representing thisview could have included as reference taxasay Apocynum and Plocosperma The recentlyproposed position of Plocosperma close tothe Boraginaceae (Backlund et al 2000)would with this denition have made Apoc-ynaceae a very large taxon (containing atleast both Gentianales and Boraginales) pos-sibly synonymous with (a hypothetical def-inition of) the informal taxon EuasteridsII (Bremer et al 1998) and again a newname for the Apocynaceae sensu Endressand Bruyns (2000 ie excluding Plocosperma)would have been needed Because theLinnaean system allows recircumscription ofsynonymous names the potential turnoverin the set of names in use is reduced withthis system (albeit at the expense of allowingsubjectivity) Additionally the ranks of theLinnaean system provide a set of names thatwork as relatively stable universal standardnames in communication note that equalranks does not guarantee comparable evo-lutionary units however Continuity in us-age of names is important for example injournals literature databases oras teach-ing and inventory work and the need forsuch continuity has been recognized by pro-ponents of both nomenclature systems (deQueiroz and Gauthier 1992 Greuter et al1994 Reveal 1996)
We propose a compromise approach thatuses solutions relating to these aspects fromboth systems concentrating mainly onbotanical classication of extant plant taxaat the tribal and familial levels and onlyconsidering monophyletic taxa We recom-mend a system that uses ldquostandard namesrdquo(ie universally used communication unitsat convenient hierarchical levels) such asthose provided by the principal ranks of theLinnaean system (Greuter et al 1994) but inwhich sensitivity toward subjective changesin circumscription is reduced One way to dothis is to adopt the denitions of the phy-logenetic system but use constraints similar
398 SYSTEMATIC BIOLOGY VOL 51
to those that apply between ranks in theLinnaean system To reduce sensitivity to-ward changes in phylogenetic hypotheses(Schander and Thollesson 1995 Bryant1996 1997 Cantino et al 1997 Moore 1998)we propose using combined node C stem-based denitions that will provoke explicitincompatibilities between taxa in case ofunfortunate changes in circumscription (deQueiroz 1996 has proposed a similar type ofdenition for designating nonmonophyletictaxa somewhat similar variants are alsodiscussed in the Phylocode) A node-basedpart of the denition provides a minimalcircumscription whereas a stem-based partprovides boundaries to competing taxa(maximal circumscription) If on a particu-lar phylogeny the minimal (monophyletic)circumscriptions of competing taxa are over-lapping then the two taxa are incompatibleTo make this meaningful at least two taxashould be included in the node-based partof the denition this will also avoid unnec-essary monotypic taxa This incompatibilitywill provide an objective point when nomen-claturalreconsideration isneeded Returningto the rst of our examples above node Cstem-based denitions of the Apocynaceaeand Asclepiadaceae could be as follows
The family Apocynaceae is the most inclusive clade inthe order Gentianales including the type specimensof Apocynum and Dictyophleba but not the taxa Ascle-piadaceae [ Gentianaceae Rubiaceae ]
and
The family Asclepiadaceae is the most inclusive cladein the order Gentianales including the type speci-mens of Asclepias and Periploca but not the taxa Apoc-ynaceae [ Gentianaceae Rubiaceae ]
On the tree in Figure 1 these two denitionsare mutually exclusive and a decision onwhich name to keep must be made Mini-mizing the number of classicatory changesshould take precedence in this decision Thereason for this is to avoid the large incon-venient ldquostandard namerdquo changes that couldresult from relatively small rearrangementsRejected names are then ignored when oc-curring in denitions of other taxa This isbecause the reference taxa in the stem-basedpart of the denitions are ldquodenedrdquo taxarather than physical types (thus reliance ondened reference taxa which has been crit-icized for stem-based denitions by Sereno1999 is what we aim for on the other hand
the explicit reference of type specimens in theminimum circumscription provided by thenode-based part of the denitions circum-vents the tautology problem raised byBryant 1996) In the current example thiswould mean that regardless of which nameof Apocynaceae and Asclepiadaceae is re-jected the other will be dened as corre-sponding to the Apocynaceae sensu Endressand Bruyns (2000) The choice will there-fore simply be which name to keep In thiscase because Apocynaceae has priority un-der the Linnaean system we would chooseto keep Apocynaceae to promote consistencywith the current system Nevertheless out-side these recommendations choices of whatnames to choose will probably to a certaindegree be arbitrary We also propose thatthe rejected name should be reinstated withan emended denition if appropriate to po-tentially allow for more stability in the setof names in use Thus in our second ex-ample above a node C stem-based deni-tion reecting the inclusion of Plocospermain Apocynaceae (eg by exchanging Dic-tyophleba for Plocosperma in the denitionabove) would presumably be incongruentwith the denition of Gentianales as well asthe denitions of Rubiaceae Gentianaceaeand so forth and therefore would be rejectedIn such acasewhere the only real change is inthe position of a single taxon a reinstatementof the Apocynaceae with an emended def-inition is motivated However a drawbackis that this may also introduce elements ofsubjectivity in the circumscriptions of taxa
Not all named clades need to be providedwith standard names In fact for the purposeof a universal set used in journals databasesand so forth it may sufce to use standardnames corresponding to the principal ranksof the Linnaean system (Greuter et al 1994)such as species genus family and orderNevertheless to provide compatibility withthe present system we suggest that addi-tional names corresponding to commonlyused secondary ranks (eg tribes) may berecognized The hierarchical level of a stan-dard name needs to be indicated This doesnot imply that taxa of the same hierarchi-cal level are comparable evolutionary unitsbut simply indicates the hierarchical levelof a taxon relative to nested taxa Becausesuch misinterpretations will in any case bedifcult to prohibit one may as well pro-vide compatibility with the present Linnaean
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 399
system by using the names of the Linnaeanranks (eg species genus tribe familyorder) to indicate hierarchical level The prin-ciple of exhaustive subsidiary taxa is notaccepted so redundant taxa need not be rec-ognized Other ldquononstandardrdquo taxa couldbe dened by using Phylocode denitionsmdashthat is without the restraints discussed forstandard taxa above
One advantage of the described systemis that utilizing the correspondence betweenthe standard names and Linnaean ranks al-lows the system to be largely compatiblewiththe establishedLinnaean system A change ofsystem could therefore be gradual with thetwo systems coexisting during a conversiontime (see eg de Quieroz 1997b)
Even though the node C stem-baseddenitions reduce the sensitivity towardschanges in circumscriptions they may still besensitive towardldquounfortunaterdquo or ldquobadrdquo def-initions (see eg Cantino et al 1997Cantinoand de Queiroz 2000) Many of the recom-mendations for such things as choices of ref-erence taxa (speciers) in phylogenetic def-initions are applicable also to the node Cstem-based denitions (eg Schander andThollesson 1995 Bryant 1996 Cantino et al1997 Sereno 1999) Most likely the compro-mise system described above will have sev-eral further problems and we hope that thispaper will invite further discussion on thesubject
A New Classication of the Apocynaceae slOur results indicate that some problem-
atic taxa remain in the classication ofEndress and Bruyns (2000) Monophyly fortwo of the subfamilies and ve of the tribesis called into question If other recent molec-ular studies (eg Potgieter and Albert 2001)are taken into consideration a further fourtribes (Alstonieae Alyxieae Plumerieae andVinceae) may be nonmonophyletic This canin some cases be simply analysis artifactsbut as Endress and Bruyns (2000) themselvespoint out their classication is to be consid-ered preliminary Insufcient taxonsamplingin molecular studies and difcult homologydecisions relating to morphological charac-ters make the circumscription of many ofthe tribes for example in the Apocynoideaesensu Endress and Bruyns (2000) uncertainThere is thus risk for future rearrangementsin many of the tribes A classication using
node C stem-based denitions may be ableto subsume such rearrangements withoutnumerous explicit recircumscriptions Wewill here use the result fromthe present studyas a basis to propose such a classication ofthe Apocynaceae sl
Our aim is to base taxa on clades that arewell supported A second aim is to makeour classication compatible with the cur-rent system We will therefore primarily rec-ognize nonoverlapping tribes that are validunder the Linnaean system We will furtheradopt the principle of nested referencing (assuggested by Lee 1999 and Sereno 1999)from the Linnaean system Thus the rst ref-erence taxon in the node-based part of thedenition is the primary type which corre-sponds to the Linnaean type used for thename of the taxon Instead of the (optional)enumeration of subsumed taxa of a lowerrank (eg genera or subtribes) we will in-clude node C stem-based denitions as dis-cussed above Citations to primary types ofreference taxa in node-based part of de-nitions are also nested and thus refer toldquothe type specimen of the type species ofthe type genus of the dened taxonrdquo Aproblem is that the relationships between thetribes areweakly supported This may lead tovery cumbersome denitions enumeratingall competing tribes in the stem-based partof the denition (eg Moore 1998) One wayto reduce this problem is to dene interme-diate well-supported taxa and use them ascompeting taxaWe therefore will further rec-ognize four nonstandard taxa (note that anassignment of ldquononstandardrdquo taxa does notimply less reliability standard names relateto communication purposes only) The taxadiscussed are indicated in Figure 3 We willdiscuss the tree from the top of Figure 3 start-ing with the taxa of the traditional Plumeri-oideae Tribes and subtribes given in paren-theses in the text refer to the classicationsof Endress and Bruyns (2000) in other casesour classication is implied For some tribeswe have chosen a name with priority underLinnaean system that implies inclusion of agenus not sampled in the present study Forall but one (Willughbeieae) of these casesadditional molecular or morphological phy-logenetic analyses support the inclusion ofthe genus Even if too uncertain denitionsshould be avoided this is less critical withnode C stem-based denitions In the caseof Willughbeieae we have decided from
400 SYSTEMATIC BIOLOGY VOL 51
FIGURE 3 The tribal reclassication discussed in the text indicated on the combinable component consensustree from the successive weighting analysis of the complete rbcL data set Four nonstandard taxamdashApocynoidinaEuapocynoidina Asclepiadacina and Asclepiadoidinamdashare also indicated
discussions with a morphological expert(M Endress pers comm) that morphologi-cal support exists for including WillughbeiaRoxb in Willughbeieae If this is correctthe denition given below will be valid if
it is not our system allows redenition ofthe taxon We have also tried to providecomments on potential morphological char-acters taken from literature for the differ-ent taxa recognized However we have not
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 401
performed any morphological analysis butrely on information external to this studyfor these comments In most cases we cantherefore not identify synapomorphies forthe taxa
TAXONOMIC TREATMENT
Tribus Willughbeieae ADCThe Ancylobotrys Dictyophleba and Va-
hadenia clade is strongly supported (uwbootand swboot D 100) and corresponds to theWillughbeieae of Endress and Bruyns (2000)The present sampling contains closely re-lated genera that have been traditionallykept together Landolphia P Beauv sl in-cludes all these taxa (eg Schumann 1895)In addition to the taxa sampled here thestudy of Potgieter and Albert (2001) sup-ports the inclusion of Couma Aubl LacmelleaH Karst Pacouria Aubl and Saba (Pichon)Pichon in this clade Plants in this cladeall have a gynoecium that is congenitallysyncarpous and indehiscent and that con-tains a eshy pulp including numerous seedswith copious horny endosperm (Fallen 1986Persoon et al 1992) This probably con-stitutes synapomorphies for this clade(M Endress pers comm) although not un-ambiguously for example syncarpy is par-alleled in Carisseae We will therefore as-sume that the reference genus Willughbeia isincluded in this tribe
DenitionmdashTribe Willughbeieae is themost inclusive clade that includes the prim-ary types of Willughbeia and Dictyophleba butnot Alyxieae Aspidospermeae CarisseaeHunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Apocynoidina
Tribus Vinceae Bartl
Ochrosia Rauvola Vinca and Catharanthus(all Vinceae) form a strongly supported clade(uwboot D 89 swboot D 99) The associa-tion between Rauvola and Catharanthus wasindicated by Sennblad and Bremer (1996)The study of Potgieter and Albert (2001) sup-ports inclusion of the genera LaxoplumeriaMarkgr Neisosperma Raf and Tonduzia Pit-tier previously classied with AspidospermaThe Vinceae contain both herbaceous andfruticose taxa The fruits are apocarpousbut whereas the fruits of Catharanthus Lax-oplumeria Neisosperma Tonduzia and Vincahave dry fruit walls the fruits of Ochrosia andRauvola are drupes Plants in Vinceae share
a similar type of style head (ie the apicalenlarged part of the style) with a stigmatichollow and upper and lower hair wreaths(except in some species of Ochrosia) the pres-ence of a nectar disk and a linear hilum onthe seed (Pichon 1948b) On the present treeVinceae also include Kopsia however this po-sition of Kopsia is very weakly supported (seealso Sennblad and Bremer 1996)
DenitionmdashTribe Vinceae is the mostinclusive clade that includes the primarytypes of Vinca and Rauvola but not AlyxieaeAspidospermeae Carisseae HunterieaeMelodineae Plumerieae Tabernaemon-taneae Willughbeieae or Apocynoidina
Tribus Tabernaemontaneae G DonThe strongly supported association be-
tween Tabernaemontana Tabernanthe Carval-hoa Schizozygia and Molongum (Tabernae-montaneae uwboot D 84 swboot D 98)has been suggested previously (Boiteau et al1978 Fallen 1986 Endress et al 1996Sennblad and Bremer 1996) Inclusion of taxatraditionally associated with Tabernaemon-tana (eg Voacanga Thouars and CallichiliaStapf see also Tabernaemontaneae sensuLeeuwenberg 1994a) and Macoubea Aublis supported by the analysis of Potgieterand Albert (2001) The taxa share sclerenchy-matic anthers free from the style head anda characteristic band of heavily cutinizedcells at the insertion on the staminal rib(Endress et al 1996) The taxa tradition-ally associated with Tabernaemontana arecharacterized by their apocarpous fruitswith arillate seeds whereas genera asso-ciated with Molongum (Ambelanieae sensuLeeuwenberg 1994a) have syncarpous fruitsand lackanarillus HoweverMacoubea formsa link between the two having a syncar-pous fruit with arillate seeds (Zarucchi et al1995)
DenitionmdashTribe Tabernaemontaneae isthe most inclusive clade that includes the pri-mary types of Tabernaemontana and Schizozy-gia but not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae PlumerieaeVinceae Willughbeieae or Apocynoidina
Tribus Aspidospermeae Miers
Aspidosperma and Vallesia (Aspidosper-meae) form a strongly supported clade (uw-boot and swboot D 100) rst suggestedin Sennblad (1997 see also Sennblad and
402 SYSTEMATIC BIOLOGY VOL 51
Bremer 2000) Subsequently GeissospermumAllemao Haplophyton ADC MicroplumeriaBaill and Strempeliopsis Benth were indi-cated to belong to this clade (Potgieter andAlbert 2001) The Aspidospermeae havesimple style heads lacking both stigmatichollow and upper hair wreath and haveapocarpous fruits with variation in eshi-ness and sclerication Alstonia has tradi-tionally been placed in the Plumerieae Be-cause its seeds have a hairy margin it hasbeen suggested to form a link to subfam-ily Apocynoideae This was contradicted bySennblad and Bremerrsquos study (1996) how-ever where Alstonia was placed in an iso-lated position as the sister group to therest of the Apocynaceae sl Here it groupswith Aspidospermeae Sensu Endress andBruyns (2000) However this association isvery weakly supported (uwboot and swbootlt 50) and in other recent studies As-pidospermeae and Alstonia do not form aclade (Potgieter and Albert 2001) Becauseof this we do not use it as a reference taxonin the denition Thus although the Aspi-dospermeae include Alstonia on the presenttree some other position of Alstonia canbe accommodated without amending thedenition
DenitionmdashTribe Aspidospermeae is themost inclusive clade that includes theprimary types of Aspidosperma and Vallesiabut not Alyxieae Carisseae HunterieaeMelodineae Plumerieae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Melodineae G DonThe well-supported association between
Melodinus and Craspidospermum (uwboot D70 swboot D 75) was rst reportedby Sennblad (1997 see also Sennblad andBremer 2000) and has subsequently beenreected in the tribe Melodineae sensuEndress and Bruyns (2000 their inclusion ofDiplorhynchus however is not supported)Craspidospermum although its fruit is dryand dehiscent as opposed to the indehis-cent fruits of Melodinus was noted by Pichon(1948b) to present several characters of theCarisseae such as stamens inserted near thebase of the corolla tube and a syncarpousovary In particular Pichon found the verydense suprastaminal indumentum type rem-iniscent of Melodinus From his descriptions
of the two genera (Pichon 1948ab) the fol-lowing similarities also emerge presence ofa stipular line pollen in tetrads and a punc-tiform hilum
DenitionmdashTribe Melodineae is the mostinclusive clade that includes the pri-mary types of Melodinus and Craspidosper-mum but not Alyxieae AspidospermeaeCarisseae Hunterieae Plumerieae Taber-naemontaneae Vinceae Willughbeieae orApocynoidina
Tribus Hunterieae K Schum
The Picralima and Pleiocarpa clade isstrongly supported (uwboot and swboot D100) and corresponds to the subtribeHunterieae sensu Endress andBruyns (2000)The study of Potgieter and Albert (2001)also supports the inclusion of Hunteria RoxbThese taxa are characterized by apocar-pous sometimes pluricarpous ovaries andeshy brous fruit walls (Omino 1996)Diplorhynchus (Melodineae) is here weaklyassociated (uwboot and swboot lt 50) withthe Pleiocarpeae It shares a few similari-ties such as stipular lines and a style headwithout a stigmatic hollow and hair wreathsbut also has differences such as a dry de-hiscent fruit With the present denitionDiplorhynchus will tentatively be included inthe Hunterieae
DenitionmdashTribe Hunterieae is the mostinclusive clade that includes the primarytypes of Hunteria and Picralima but not Alyx-ieae Aspidospermeae Carisseae Melod-ineae Plumerieae TabernaemontaneaeVinceae Willughbeieae or Apocynoidina
Tribus Plumerieae EndlThe strongly supported relationship
(uwboot and swboot D 100) between Alla-manda and Plumeria corresponds to earlierstudies (Endress et al 1996 Sennblad andBremer 1996 Civeyrel et al 1998 Potgieterand Albert 2001) In all earlier treatmentsAllamanda has had uncertain relationshipsThe association with Plumeria nds supportin pollen morphology such as perforatemesocolpial depressions and similar innerexine pattern The two genera also containsecoiridoids rather than the indole alka-loids and cardenolides that are commonin the traditional Plumerioideae (Endresset al 1996) Cameraria Cerbera and Thevetiaform a strongly supported association
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 403
(uwboot D 83 swboot D 91) with Ane-chites which has been suggested by Fallen(1983) This association was based mainlyon the latrorsesublatrorse anthers andthe broad style head with large apicalappendages and stigmatic hollow The studyby Potgieter and Albert (2001) suggesteda further inclusion of Cerberiopsis Viell ExPancher amp Sebert and Skytanthus Meyen inthis tribe These taxa form a well-supportedclade with Plumeria and Allamanda whichcorresponds to the Plumerieae sensuEndress and Bruyns (2000) Characterssupporting this relationship are the presenceof infrastaminal as well as suprastaminalappendages (not present in Anechites andPlumeria) and winged seeds (not present inAnechites) (Endress et al 1996)
DenitionmdashTribe Plumerieae is themost inclusive clade that includes theprimary types of Plumeria and Allamandabut not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Carisseae Dumort
Acokanthera and Carissa (both Carisseae)are traditionally joined and are sometimeseven treated as one genus (eg Pichon1948a) In the present study they form astrongly supported clade (uwboot D 96swboot D 99) Among the characters sup-porting this tribe are syncarpous fruits with-out axile placentation and endocracks on theinner exine of the pollen (Endress et al 1996)
DenitionmdashTribe Carisseae is the most in-clusive clade that includes the primary typesof Carissa and Acokanthera but not AlyxieaeAspidospermeae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
Tribus Alyxieae G Don
Chilocarpus has been a genus with un-certain afnities In Endress et al (1996) itwas the sister taxon to the ApocynoideaeHere it is associated with Alyxia and Lep-inia in a clade corresponding to the Alyxieaesensu Endress and Bruyns (2000) Thisclade is strongly supported with succes-sive weighted bootstrap (swboot D 91) butnot with unit-weighted bootstrap support(uwboot D 58) such clades will henceforthbe termed moderatelywell supported clades
A further inclusion of Condylocarpon DesfLepiniopsis Valeton and Plectaneia Thouarswas indicated by Potgieter and Albert (2001)The constituent taxa are the only Apocy-naceae taxa outside the Apocynoidina thathave porate pollen
DenitionmdashTribe Alyxieae is the most in-clusive clade that includes the primary typesof Alyxia and Lepinia but not Aspidosper-meae Carisseae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
APOCYNOIDINA
On the present tree the traditionalsubfam-ily Apocynoideae includes all subfamilies(Periplocoideae Secamonoideae and Ascle-piadoideae) of the Asclepiadaceae sensuLiede and Albers (1994) All of these taxaare established names and selecting one asa subfamily at the expense of the othersmight be considered unfortunate from dif-ferent points of view We have here chosennot to recognize subfamilies instead we willdene the traditional Apocynoideae Ascle-piadaceae and Asclepiadoideae as nonstan-dard taxa The correspondence to the tradi-tional taxa will be indicated in the names byreplacing the sufx -eae with a neutral suf-x -ina (Kron 1997) We will also for practi-cal reasonsrecognize as a nonstandardtaxonthe informal group euapocynoids suggestedby Sennblad et al (1998)
The taxa of the traditional Apocynoideaeand Asclepiadaceae (uwboot D 86 swboot D99) clade will in many cases be more dif-cult to safely delimit to tribes becausemany of the groups in this clade have weakor no support (although preliminary resultsfrom an unpublished collaborative study bySennblad et al indicate additional supportfor the tribes discussed below) The tribaland subtribal classication of the traditionalApocynoideae has been shown to be prob-lematic (Leeuwenberg 1994a Endress andBruyns 2000) The circumscription of tribesfor these taxa might therefore in some casesbe preliminary
In all following taxa the basal part of theanther connective called the retinacle is ad-nate to the style head
DenitionmdashApocynoidina are the mostinclusive clade that includes the pri-mary types of Apocynum and Wrightia butnot Alyxieae Aspidospermeae Carisseae
404 SYSTEMATIC BIOLOGY VOL 51
Hunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Willughbeieae
Tribus Wrightieae G DonThe Stephanostema and Wrightia (both
Wrightieae) clade is strongly supported(uwboot and swboot D 100) and correspondsto the Wrightieae sensu Sennblad et al (1998)who also suggested an inclusion of Pleio-ceras Baill The tribe could be characterizedby a combination of synapomorphic and ple-siomorphic characters for example presenceofa chalazalandabsence of micropylar comaleft contorted aestivation and absence of airspaces in the anthers (1998)
DenitionmdashTribe Wrightieae is the mostinclusive clade in the Apocynoidina that in-cludes the primary types of Wrightia andStephanostema but not Malouetieae Nerieaeor Euapocynoidina
Tribus Nerieae (Benth) M Pichon
Adenium Nerium and Strophanthus (allWrightieae) constituted the Nerieae inSennblad et al (1998) however the cladereceived weak support Here Adenium andNerium form a clade that is absent in halfof the most-parsimonious trees in the com-plete rbcL analysis but is strongly supportedin the combined ndhF and rbcL analysis(uwboot D 98 swboot D 99) Strophan-thus groups with the Mascarenhasia Pachy-podium Kibatalia Funtumia and Holarrhenaclade which corresponds to the Malouetieaesensu Sennblad et al (1998) This position ofStrophanthus is however weakly supported(uwboot and swboot lt 50) and is not presentin the combined ndhF and rbcL tree Fur-thermore certain morphological characters(unfused slits in the corolla tube and anapical anther appendage) indicate a posi-tion closer to the Nerieae (Sennblad et al1998)
DenitionmdashTribe Nerieae is the most in-clusive clade in the Apocynoidina that in-cludes the primary types of Nerium and Ade-nium but not Malouetieae Wrightieae orEuapocynoidina
Tribus Malouetieae Mull-Arg
The association of Mascarenhasia Pachy-podium Kibatalia Funtumia and Holar-rhena (all Wrightieae) is moderately wellsupported in the complete rbcL analysis
(uwboot D 57 swboot D 74) whereasthe corresponding clade is well supportedin the combined rbcL and ndhF analysis(uwboot D 66 swboot D 100) On the ba-sis of identied synapomorphies (presenceof calcium oxalate crystals in the stomium ofthe anthers absence of interpetal vein ses-sile laments and absence of air spaces inthe anthers) Sennblad et al (1998) predictedan inclusion of Alaa Thouars Kibatalia andMalouetia ADC in the tribe The inclusion ofKibatalia is strongly supported in this studyFollowing these suggestions we will here as-sume the inclusion of the reference genusMalouetia in this tribe (The denition belowwill include Strophanthus on the present treebut see comments under Nerieae)
DenitionmdashTribe Malouetieae is the mostinclusive clade in the Apocynoidina that in-cluded the primary types of Malouetia andFuntumia but not Nerieae Wrightieae orEuapocynoidina
EUAPOCYNOIDINA
In all the remaining representatives ofthe Apocynoideae sensu Endress andBruyns (2000) that is Mandevilla MesechitesAganosma Apocynum TrachelospermumRhabdadenia Beaumontia Parsonsia andPrestonia the anthers are adnate to the stylehead both by the retinacle and by the thecaeThese taxa together with the taxa of the tra-ditional Asclepiadaceae form a moderatelysupported clade (uwboot lt 50 swboot D86) that corresponds to the informal taxoneuapocynoids sensu Sennblad et al (1998)
DenitionmdashEuapocynoidina are the mostinclusive clade that includes the primarytypes of Apocynum and Echites but notMalouetieae Wrightieae or Nerieae
Tribus Mesechiteae Miers
Mandevilla and Mesechites (both Mese-chiteae) form a moderately supported clade(uwboot lt 50 swboot D 87) equivalentto the Mesechiteae This tribe could be char-acterized by anthers with obtuse truncatedtails and a retinacle lacking hairs and also astrongly pentagonal style head with a stig-matic hollow
DenitionmdashTribe Mesechiteae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Mesechitesand Mandevilla but not Apocyneae EchiteaePeriploceae or Asclepiadacina
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 405
Tribus Apocyneae Bercht et J Presl
Aganosma Apocynum Beaumontia andTrachelospermum (Apocyneae) form a weaklysupported clade with Rhabdadenia (Echiteae)(uwboot and swboot lt 50) in the completerbcL analysis whereas the correspondingclade in the combined ndhF and rbcL analy-sis gains a little better support (uwboot lt 50swboot D 91) Likewise this clade (with theinclusion of Chonemorpha G Don) is weaklysupported in Potgieter and Albert (2001) Po-tential characterizing traits could be a stylehead without stigmatic hollow in combina-tion with thecae that are adnate to the stylehead (Sennblad et al 1998)
DenitionmdashTribe Apocyneae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Apocynumand Trachelospermum but not Echiteae Mese-chiteae Periploceae or Asclepiadacina
Tribus Echiteae Bartl
Parsonsia Peltastes and Prestonia (allEchiteae) form a moderately supported orunsupported clade with the Periplocoideaerepresentatives of the study (uwboot andswboot lt 50 and uwboot lt 50 swboot D89 in the complete rbcL and the combinedrbcL and ndhF analysis respectively) In thestudy of Sennblad et al (1998) Prestonia andParsonsia formed a weakly supported cladeand in the study of Potgieter and Albert(2001) inclusion of both Parsonsia and EchitesP Browne in the Echiteae is supportedParsonsia and Echiteae also share certaincharacters such as a horseshoe-shaped reti-nacle and a style head with stigmatic hollow(see Sennblad et al 1998)
DenitionmdashTribe Echiteae is the most in-clusive clade in the Euapocynoidina thatincludes the primary types of Echites andPrestonia but not Apocyneae MesechiteaePeriploceae or Asclepiadacina
Tribus Periploceae BartlThe representatives of the Periplo-
coideae sensu Endress and Bruyns (2000)mdashPentopetia Parquetina Petopentia PeriplocaTacazzea and Mondiamdashgroup together ina well-supported clade (uwboot D 67swboot D 93) The monophyly of the tra-ditional periplocoid taxa is also supportedin earlier studies that included a wider ordifferent sampling (Civeyrel et al 1998
Potgieter and Albert 2001) The taxa ofthe traditional Asclepiadaceae have ldquotrans-latorsrdquo structures related to pollinationspecialization In the Periploceae the pollenis deposited as tetrads in a spoon-liketranslator with an adhesive disc
DenitionmdashTribe Periploceae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Periplocaand Pentopetia but not Apocyneae EchiteaeMesechiteae or Asclepiadacina
ASCLEPIADACINA
The taxa of the Asclepiadoideae sensuLiede and Albers (1994) form a clade togetherwith the two Secamone species and Baissea(Apocyneae uwboot lt 50 swboot D 73)The exclusion of Periploceae (traditionallyassociated with this clade see Civeyrel et al1998) from this taxon is weakly supportedTherefore Periploceae is not included as areference taxon in the stem-based part ofthe denition of the Asclepiadacina thusleaving open the possibility for a later in-clusion of this taxon In the Asclepiadacina(except Baissea) the pollen is agglutinatedinto pollinia that are connected to a claspingtranslator
DenitionmdashAsclepiadacina are the mostinclusive monophyletic clade in the Euapoc-ynoidina that includes the primary types ofAsclepias and Secamoneae but not ApocyneaeEchiteae or Mesechiteae
Tribus Secamoneae G DonThe tribe Secamoneae sensu Bruyns
(uwboot D 89 swboot D 96) is mono-phyletic in the present analysisHowever be-cause representatives from only one of theconstituent genera are included monophylyis not well tested but the study of Civeyrelet al (1998) which included a larger numberof taxa including Pervillea Decne showedthat monophyly of the tribe was supportedThe traditional character for this tribe is fourpollinia per translator
DenitionmdashTribe Secamoneae is the mostinclusive clade in the Asclepiadacina that in-cludes the primary types of Secamone andPervillea but not Asclepioidina or Baissea
BaisseaThe position of Baissea corresponds with
that in the study of Sennblad et al (1998)
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
398 SYSTEMATIC BIOLOGY VOL 51
to those that apply between ranks in theLinnaean system To reduce sensitivity to-ward changes in phylogenetic hypotheses(Schander and Thollesson 1995 Bryant1996 1997 Cantino et al 1997 Moore 1998)we propose using combined node C stem-based denitions that will provoke explicitincompatibilities between taxa in case ofunfortunate changes in circumscription (deQueiroz 1996 has proposed a similar type ofdenition for designating nonmonophyletictaxa somewhat similar variants are alsodiscussed in the Phylocode) A node-basedpart of the denition provides a minimalcircumscription whereas a stem-based partprovides boundaries to competing taxa(maximal circumscription) If on a particu-lar phylogeny the minimal (monophyletic)circumscriptions of competing taxa are over-lapping then the two taxa are incompatibleTo make this meaningful at least two taxashould be included in the node-based partof the denition this will also avoid unnec-essary monotypic taxa This incompatibilitywill provide an objective point when nomen-claturalreconsideration isneeded Returningto the rst of our examples above node Cstem-based denitions of the Apocynaceaeand Asclepiadaceae could be as follows
The family Apocynaceae is the most inclusive clade inthe order Gentianales including the type specimensof Apocynum and Dictyophleba but not the taxa Ascle-piadaceae [ Gentianaceae Rubiaceae ]
and
The family Asclepiadaceae is the most inclusive cladein the order Gentianales including the type speci-mens of Asclepias and Periploca but not the taxa Apoc-ynaceae [ Gentianaceae Rubiaceae ]
On the tree in Figure 1 these two denitionsare mutually exclusive and a decision onwhich name to keep must be made Mini-mizing the number of classicatory changesshould take precedence in this decision Thereason for this is to avoid the large incon-venient ldquostandard namerdquo changes that couldresult from relatively small rearrangementsRejected names are then ignored when oc-curring in denitions of other taxa This isbecause the reference taxa in the stem-basedpart of the denitions are ldquodenedrdquo taxarather than physical types (thus reliance ondened reference taxa which has been crit-icized for stem-based denitions by Sereno1999 is what we aim for on the other hand
the explicit reference of type specimens in theminimum circumscription provided by thenode-based part of the denitions circum-vents the tautology problem raised byBryant 1996) In the current example thiswould mean that regardless of which nameof Apocynaceae and Asclepiadaceae is re-jected the other will be dened as corre-sponding to the Apocynaceae sensu Endressand Bruyns (2000) The choice will there-fore simply be which name to keep In thiscase because Apocynaceae has priority un-der the Linnaean system we would chooseto keep Apocynaceae to promote consistencywith the current system Nevertheless out-side these recommendations choices of whatnames to choose will probably to a certaindegree be arbitrary We also propose thatthe rejected name should be reinstated withan emended denition if appropriate to po-tentially allow for more stability in the setof names in use Thus in our second ex-ample above a node C stem-based deni-tion reecting the inclusion of Plocospermain Apocynaceae (eg by exchanging Dic-tyophleba for Plocosperma in the denitionabove) would presumably be incongruentwith the denition of Gentianales as well asthe denitions of Rubiaceae Gentianaceaeand so forth and therefore would be rejectedIn such acasewhere the only real change is inthe position of a single taxon a reinstatementof the Apocynaceae with an emended def-inition is motivated However a drawbackis that this may also introduce elements ofsubjectivity in the circumscriptions of taxa
Not all named clades need to be providedwith standard names In fact for the purposeof a universal set used in journals databasesand so forth it may sufce to use standardnames corresponding to the principal ranksof the Linnaean system (Greuter et al 1994)such as species genus family and orderNevertheless to provide compatibility withthe present system we suggest that addi-tional names corresponding to commonlyused secondary ranks (eg tribes) may berecognized The hierarchical level of a stan-dard name needs to be indicated This doesnot imply that taxa of the same hierarchi-cal level are comparable evolutionary unitsbut simply indicates the hierarchical levelof a taxon relative to nested taxa Becausesuch misinterpretations will in any case bedifcult to prohibit one may as well pro-vide compatibility with the present Linnaean
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 399
system by using the names of the Linnaeanranks (eg species genus tribe familyorder) to indicate hierarchical level The prin-ciple of exhaustive subsidiary taxa is notaccepted so redundant taxa need not be rec-ognized Other ldquononstandardrdquo taxa couldbe dened by using Phylocode denitionsmdashthat is without the restraints discussed forstandard taxa above
One advantage of the described systemis that utilizing the correspondence betweenthe standard names and Linnaean ranks al-lows the system to be largely compatiblewiththe establishedLinnaean system A change ofsystem could therefore be gradual with thetwo systems coexisting during a conversiontime (see eg de Quieroz 1997b)
Even though the node C stem-baseddenitions reduce the sensitivity towardschanges in circumscriptions they may still besensitive towardldquounfortunaterdquo or ldquobadrdquo def-initions (see eg Cantino et al 1997Cantinoand de Queiroz 2000) Many of the recom-mendations for such things as choices of ref-erence taxa (speciers) in phylogenetic def-initions are applicable also to the node Cstem-based denitions (eg Schander andThollesson 1995 Bryant 1996 Cantino et al1997 Sereno 1999) Most likely the compro-mise system described above will have sev-eral further problems and we hope that thispaper will invite further discussion on thesubject
A New Classication of the Apocynaceae slOur results indicate that some problem-
atic taxa remain in the classication ofEndress and Bruyns (2000) Monophyly fortwo of the subfamilies and ve of the tribesis called into question If other recent molec-ular studies (eg Potgieter and Albert 2001)are taken into consideration a further fourtribes (Alstonieae Alyxieae Plumerieae andVinceae) may be nonmonophyletic This canin some cases be simply analysis artifactsbut as Endress and Bruyns (2000) themselvespoint out their classication is to be consid-ered preliminary Insufcient taxonsamplingin molecular studies and difcult homologydecisions relating to morphological charac-ters make the circumscription of many ofthe tribes for example in the Apocynoideaesensu Endress and Bruyns (2000) uncertainThere is thus risk for future rearrangementsin many of the tribes A classication using
node C stem-based denitions may be ableto subsume such rearrangements withoutnumerous explicit recircumscriptions Wewill here use the result fromthe present studyas a basis to propose such a classication ofthe Apocynaceae sl
Our aim is to base taxa on clades that arewell supported A second aim is to makeour classication compatible with the cur-rent system We will therefore primarily rec-ognize nonoverlapping tribes that are validunder the Linnaean system We will furtheradopt the principle of nested referencing (assuggested by Lee 1999 and Sereno 1999)from the Linnaean system Thus the rst ref-erence taxon in the node-based part of thedenition is the primary type which corre-sponds to the Linnaean type used for thename of the taxon Instead of the (optional)enumeration of subsumed taxa of a lowerrank (eg genera or subtribes) we will in-clude node C stem-based denitions as dis-cussed above Citations to primary types ofreference taxa in node-based part of de-nitions are also nested and thus refer toldquothe type specimen of the type species ofthe type genus of the dened taxonrdquo Aproblem is that the relationships between thetribes areweakly supported This may lead tovery cumbersome denitions enumeratingall competing tribes in the stem-based partof the denition (eg Moore 1998) One wayto reduce this problem is to dene interme-diate well-supported taxa and use them ascompeting taxaWe therefore will further rec-ognize four nonstandard taxa (note that anassignment of ldquononstandardrdquo taxa does notimply less reliability standard names relateto communication purposes only) The taxadiscussed are indicated in Figure 3 We willdiscuss the tree from the top of Figure 3 start-ing with the taxa of the traditional Plumeri-oideae Tribes and subtribes given in paren-theses in the text refer to the classicationsof Endress and Bruyns (2000) in other casesour classication is implied For some tribeswe have chosen a name with priority underLinnaean system that implies inclusion of agenus not sampled in the present study Forall but one (Willughbeieae) of these casesadditional molecular or morphological phy-logenetic analyses support the inclusion ofthe genus Even if too uncertain denitionsshould be avoided this is less critical withnode C stem-based denitions In the caseof Willughbeieae we have decided from
400 SYSTEMATIC BIOLOGY VOL 51
FIGURE 3 The tribal reclassication discussed in the text indicated on the combinable component consensustree from the successive weighting analysis of the complete rbcL data set Four nonstandard taxamdashApocynoidinaEuapocynoidina Asclepiadacina and Asclepiadoidinamdashare also indicated
discussions with a morphological expert(M Endress pers comm) that morphologi-cal support exists for including WillughbeiaRoxb in Willughbeieae If this is correctthe denition given below will be valid if
it is not our system allows redenition ofthe taxon We have also tried to providecomments on potential morphological char-acters taken from literature for the differ-ent taxa recognized However we have not
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 401
performed any morphological analysis butrely on information external to this studyfor these comments In most cases we cantherefore not identify synapomorphies forthe taxa
TAXONOMIC TREATMENT
Tribus Willughbeieae ADCThe Ancylobotrys Dictyophleba and Va-
hadenia clade is strongly supported (uwbootand swboot D 100) and corresponds to theWillughbeieae of Endress and Bruyns (2000)The present sampling contains closely re-lated genera that have been traditionallykept together Landolphia P Beauv sl in-cludes all these taxa (eg Schumann 1895)In addition to the taxa sampled here thestudy of Potgieter and Albert (2001) sup-ports the inclusion of Couma Aubl LacmelleaH Karst Pacouria Aubl and Saba (Pichon)Pichon in this clade Plants in this cladeall have a gynoecium that is congenitallysyncarpous and indehiscent and that con-tains a eshy pulp including numerous seedswith copious horny endosperm (Fallen 1986Persoon et al 1992) This probably con-stitutes synapomorphies for this clade(M Endress pers comm) although not un-ambiguously for example syncarpy is par-alleled in Carisseae We will therefore as-sume that the reference genus Willughbeia isincluded in this tribe
DenitionmdashTribe Willughbeieae is themost inclusive clade that includes the prim-ary types of Willughbeia and Dictyophleba butnot Alyxieae Aspidospermeae CarisseaeHunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Apocynoidina
Tribus Vinceae Bartl
Ochrosia Rauvola Vinca and Catharanthus(all Vinceae) form a strongly supported clade(uwboot D 89 swboot D 99) The associa-tion between Rauvola and Catharanthus wasindicated by Sennblad and Bremer (1996)The study of Potgieter and Albert (2001) sup-ports inclusion of the genera LaxoplumeriaMarkgr Neisosperma Raf and Tonduzia Pit-tier previously classied with AspidospermaThe Vinceae contain both herbaceous andfruticose taxa The fruits are apocarpousbut whereas the fruits of Catharanthus Lax-oplumeria Neisosperma Tonduzia and Vincahave dry fruit walls the fruits of Ochrosia andRauvola are drupes Plants in Vinceae share
a similar type of style head (ie the apicalenlarged part of the style) with a stigmatichollow and upper and lower hair wreaths(except in some species of Ochrosia) the pres-ence of a nectar disk and a linear hilum onthe seed (Pichon 1948b) On the present treeVinceae also include Kopsia however this po-sition of Kopsia is very weakly supported (seealso Sennblad and Bremer 1996)
DenitionmdashTribe Vinceae is the mostinclusive clade that includes the primarytypes of Vinca and Rauvola but not AlyxieaeAspidospermeae Carisseae HunterieaeMelodineae Plumerieae Tabernaemon-taneae Willughbeieae or Apocynoidina
Tribus Tabernaemontaneae G DonThe strongly supported association be-
tween Tabernaemontana Tabernanthe Carval-hoa Schizozygia and Molongum (Tabernae-montaneae uwboot D 84 swboot D 98)has been suggested previously (Boiteau et al1978 Fallen 1986 Endress et al 1996Sennblad and Bremer 1996) Inclusion of taxatraditionally associated with Tabernaemon-tana (eg Voacanga Thouars and CallichiliaStapf see also Tabernaemontaneae sensuLeeuwenberg 1994a) and Macoubea Aublis supported by the analysis of Potgieterand Albert (2001) The taxa share sclerenchy-matic anthers free from the style head anda characteristic band of heavily cutinizedcells at the insertion on the staminal rib(Endress et al 1996) The taxa tradition-ally associated with Tabernaemontana arecharacterized by their apocarpous fruitswith arillate seeds whereas genera asso-ciated with Molongum (Ambelanieae sensuLeeuwenberg 1994a) have syncarpous fruitsand lackanarillus HoweverMacoubea formsa link between the two having a syncar-pous fruit with arillate seeds (Zarucchi et al1995)
DenitionmdashTribe Tabernaemontaneae isthe most inclusive clade that includes the pri-mary types of Tabernaemontana and Schizozy-gia but not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae PlumerieaeVinceae Willughbeieae or Apocynoidina
Tribus Aspidospermeae Miers
Aspidosperma and Vallesia (Aspidosper-meae) form a strongly supported clade (uw-boot and swboot D 100) rst suggestedin Sennblad (1997 see also Sennblad and
402 SYSTEMATIC BIOLOGY VOL 51
Bremer 2000) Subsequently GeissospermumAllemao Haplophyton ADC MicroplumeriaBaill and Strempeliopsis Benth were indi-cated to belong to this clade (Potgieter andAlbert 2001) The Aspidospermeae havesimple style heads lacking both stigmatichollow and upper hair wreath and haveapocarpous fruits with variation in eshi-ness and sclerication Alstonia has tradi-tionally been placed in the Plumerieae Be-cause its seeds have a hairy margin it hasbeen suggested to form a link to subfam-ily Apocynoideae This was contradicted bySennblad and Bremerrsquos study (1996) how-ever where Alstonia was placed in an iso-lated position as the sister group to therest of the Apocynaceae sl Here it groupswith Aspidospermeae Sensu Endress andBruyns (2000) However this association isvery weakly supported (uwboot and swbootlt 50) and in other recent studies As-pidospermeae and Alstonia do not form aclade (Potgieter and Albert 2001) Becauseof this we do not use it as a reference taxonin the denition Thus although the Aspi-dospermeae include Alstonia on the presenttree some other position of Alstonia canbe accommodated without amending thedenition
DenitionmdashTribe Aspidospermeae is themost inclusive clade that includes theprimary types of Aspidosperma and Vallesiabut not Alyxieae Carisseae HunterieaeMelodineae Plumerieae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Melodineae G DonThe well-supported association between
Melodinus and Craspidospermum (uwboot D70 swboot D 75) was rst reportedby Sennblad (1997 see also Sennblad andBremer 2000) and has subsequently beenreected in the tribe Melodineae sensuEndress and Bruyns (2000 their inclusion ofDiplorhynchus however is not supported)Craspidospermum although its fruit is dryand dehiscent as opposed to the indehis-cent fruits of Melodinus was noted by Pichon(1948b) to present several characters of theCarisseae such as stamens inserted near thebase of the corolla tube and a syncarpousovary In particular Pichon found the verydense suprastaminal indumentum type rem-iniscent of Melodinus From his descriptions
of the two genera (Pichon 1948ab) the fol-lowing similarities also emerge presence ofa stipular line pollen in tetrads and a punc-tiform hilum
DenitionmdashTribe Melodineae is the mostinclusive clade that includes the pri-mary types of Melodinus and Craspidosper-mum but not Alyxieae AspidospermeaeCarisseae Hunterieae Plumerieae Taber-naemontaneae Vinceae Willughbeieae orApocynoidina
Tribus Hunterieae K Schum
The Picralima and Pleiocarpa clade isstrongly supported (uwboot and swboot D100) and corresponds to the subtribeHunterieae sensu Endress andBruyns (2000)The study of Potgieter and Albert (2001)also supports the inclusion of Hunteria RoxbThese taxa are characterized by apocar-pous sometimes pluricarpous ovaries andeshy brous fruit walls (Omino 1996)Diplorhynchus (Melodineae) is here weaklyassociated (uwboot and swboot lt 50) withthe Pleiocarpeae It shares a few similari-ties such as stipular lines and a style headwithout a stigmatic hollow and hair wreathsbut also has differences such as a dry de-hiscent fruit With the present denitionDiplorhynchus will tentatively be included inthe Hunterieae
DenitionmdashTribe Hunterieae is the mostinclusive clade that includes the primarytypes of Hunteria and Picralima but not Alyx-ieae Aspidospermeae Carisseae Melod-ineae Plumerieae TabernaemontaneaeVinceae Willughbeieae or Apocynoidina
Tribus Plumerieae EndlThe strongly supported relationship
(uwboot and swboot D 100) between Alla-manda and Plumeria corresponds to earlierstudies (Endress et al 1996 Sennblad andBremer 1996 Civeyrel et al 1998 Potgieterand Albert 2001) In all earlier treatmentsAllamanda has had uncertain relationshipsThe association with Plumeria nds supportin pollen morphology such as perforatemesocolpial depressions and similar innerexine pattern The two genera also containsecoiridoids rather than the indole alka-loids and cardenolides that are commonin the traditional Plumerioideae (Endresset al 1996) Cameraria Cerbera and Thevetiaform a strongly supported association
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 403
(uwboot D 83 swboot D 91) with Ane-chites which has been suggested by Fallen(1983) This association was based mainlyon the latrorsesublatrorse anthers andthe broad style head with large apicalappendages and stigmatic hollow The studyby Potgieter and Albert (2001) suggesteda further inclusion of Cerberiopsis Viell ExPancher amp Sebert and Skytanthus Meyen inthis tribe These taxa form a well-supportedclade with Plumeria and Allamanda whichcorresponds to the Plumerieae sensuEndress and Bruyns (2000) Characterssupporting this relationship are the presenceof infrastaminal as well as suprastaminalappendages (not present in Anechites andPlumeria) and winged seeds (not present inAnechites) (Endress et al 1996)
DenitionmdashTribe Plumerieae is themost inclusive clade that includes theprimary types of Plumeria and Allamandabut not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Carisseae Dumort
Acokanthera and Carissa (both Carisseae)are traditionally joined and are sometimeseven treated as one genus (eg Pichon1948a) In the present study they form astrongly supported clade (uwboot D 96swboot D 99) Among the characters sup-porting this tribe are syncarpous fruits with-out axile placentation and endocracks on theinner exine of the pollen (Endress et al 1996)
DenitionmdashTribe Carisseae is the most in-clusive clade that includes the primary typesof Carissa and Acokanthera but not AlyxieaeAspidospermeae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
Tribus Alyxieae G Don
Chilocarpus has been a genus with un-certain afnities In Endress et al (1996) itwas the sister taxon to the ApocynoideaeHere it is associated with Alyxia and Lep-inia in a clade corresponding to the Alyxieaesensu Endress and Bruyns (2000) Thisclade is strongly supported with succes-sive weighted bootstrap (swboot D 91) butnot with unit-weighted bootstrap support(uwboot D 58) such clades will henceforthbe termed moderatelywell supported clades
A further inclusion of Condylocarpon DesfLepiniopsis Valeton and Plectaneia Thouarswas indicated by Potgieter and Albert (2001)The constituent taxa are the only Apocy-naceae taxa outside the Apocynoidina thathave porate pollen
DenitionmdashTribe Alyxieae is the most in-clusive clade that includes the primary typesof Alyxia and Lepinia but not Aspidosper-meae Carisseae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
APOCYNOIDINA
On the present tree the traditionalsubfam-ily Apocynoideae includes all subfamilies(Periplocoideae Secamonoideae and Ascle-piadoideae) of the Asclepiadaceae sensuLiede and Albers (1994) All of these taxaare established names and selecting one asa subfamily at the expense of the othersmight be considered unfortunate from dif-ferent points of view We have here chosennot to recognize subfamilies instead we willdene the traditional Apocynoideae Ascle-piadaceae and Asclepiadoideae as nonstan-dard taxa The correspondence to the tradi-tional taxa will be indicated in the names byreplacing the sufx -eae with a neutral suf-x -ina (Kron 1997) We will also for practi-cal reasonsrecognize as a nonstandardtaxonthe informal group euapocynoids suggestedby Sennblad et al (1998)
The taxa of the traditional Apocynoideaeand Asclepiadaceae (uwboot D 86 swboot D99) clade will in many cases be more dif-cult to safely delimit to tribes becausemany of the groups in this clade have weakor no support (although preliminary resultsfrom an unpublished collaborative study bySennblad et al indicate additional supportfor the tribes discussed below) The tribaland subtribal classication of the traditionalApocynoideae has been shown to be prob-lematic (Leeuwenberg 1994a Endress andBruyns 2000) The circumscription of tribesfor these taxa might therefore in some casesbe preliminary
In all following taxa the basal part of theanther connective called the retinacle is ad-nate to the style head
DenitionmdashApocynoidina are the mostinclusive clade that includes the pri-mary types of Apocynum and Wrightia butnot Alyxieae Aspidospermeae Carisseae
404 SYSTEMATIC BIOLOGY VOL 51
Hunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Willughbeieae
Tribus Wrightieae G DonThe Stephanostema and Wrightia (both
Wrightieae) clade is strongly supported(uwboot and swboot D 100) and correspondsto the Wrightieae sensu Sennblad et al (1998)who also suggested an inclusion of Pleio-ceras Baill The tribe could be characterizedby a combination of synapomorphic and ple-siomorphic characters for example presenceofa chalazalandabsence of micropylar comaleft contorted aestivation and absence of airspaces in the anthers (1998)
DenitionmdashTribe Wrightieae is the mostinclusive clade in the Apocynoidina that in-cludes the primary types of Wrightia andStephanostema but not Malouetieae Nerieaeor Euapocynoidina
Tribus Nerieae (Benth) M Pichon
Adenium Nerium and Strophanthus (allWrightieae) constituted the Nerieae inSennblad et al (1998) however the cladereceived weak support Here Adenium andNerium form a clade that is absent in halfof the most-parsimonious trees in the com-plete rbcL analysis but is strongly supportedin the combined ndhF and rbcL analysis(uwboot D 98 swboot D 99) Strophan-thus groups with the Mascarenhasia Pachy-podium Kibatalia Funtumia and Holarrhenaclade which corresponds to the Malouetieaesensu Sennblad et al (1998) This position ofStrophanthus is however weakly supported(uwboot and swboot lt 50) and is not presentin the combined ndhF and rbcL tree Fur-thermore certain morphological characters(unfused slits in the corolla tube and anapical anther appendage) indicate a posi-tion closer to the Nerieae (Sennblad et al1998)
DenitionmdashTribe Nerieae is the most in-clusive clade in the Apocynoidina that in-cludes the primary types of Nerium and Ade-nium but not Malouetieae Wrightieae orEuapocynoidina
Tribus Malouetieae Mull-Arg
The association of Mascarenhasia Pachy-podium Kibatalia Funtumia and Holar-rhena (all Wrightieae) is moderately wellsupported in the complete rbcL analysis
(uwboot D 57 swboot D 74) whereasthe corresponding clade is well supportedin the combined rbcL and ndhF analysis(uwboot D 66 swboot D 100) On the ba-sis of identied synapomorphies (presenceof calcium oxalate crystals in the stomium ofthe anthers absence of interpetal vein ses-sile laments and absence of air spaces inthe anthers) Sennblad et al (1998) predictedan inclusion of Alaa Thouars Kibatalia andMalouetia ADC in the tribe The inclusion ofKibatalia is strongly supported in this studyFollowing these suggestions we will here as-sume the inclusion of the reference genusMalouetia in this tribe (The denition belowwill include Strophanthus on the present treebut see comments under Nerieae)
DenitionmdashTribe Malouetieae is the mostinclusive clade in the Apocynoidina that in-cluded the primary types of Malouetia andFuntumia but not Nerieae Wrightieae orEuapocynoidina
EUAPOCYNOIDINA
In all the remaining representatives ofthe Apocynoideae sensu Endress andBruyns (2000) that is Mandevilla MesechitesAganosma Apocynum TrachelospermumRhabdadenia Beaumontia Parsonsia andPrestonia the anthers are adnate to the stylehead both by the retinacle and by the thecaeThese taxa together with the taxa of the tra-ditional Asclepiadaceae form a moderatelysupported clade (uwboot lt 50 swboot D86) that corresponds to the informal taxoneuapocynoids sensu Sennblad et al (1998)
DenitionmdashEuapocynoidina are the mostinclusive clade that includes the primarytypes of Apocynum and Echites but notMalouetieae Wrightieae or Nerieae
Tribus Mesechiteae Miers
Mandevilla and Mesechites (both Mese-chiteae) form a moderately supported clade(uwboot lt 50 swboot D 87) equivalentto the Mesechiteae This tribe could be char-acterized by anthers with obtuse truncatedtails and a retinacle lacking hairs and also astrongly pentagonal style head with a stig-matic hollow
DenitionmdashTribe Mesechiteae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Mesechitesand Mandevilla but not Apocyneae EchiteaePeriploceae or Asclepiadacina
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 405
Tribus Apocyneae Bercht et J Presl
Aganosma Apocynum Beaumontia andTrachelospermum (Apocyneae) form a weaklysupported clade with Rhabdadenia (Echiteae)(uwboot and swboot lt 50) in the completerbcL analysis whereas the correspondingclade in the combined ndhF and rbcL analy-sis gains a little better support (uwboot lt 50swboot D 91) Likewise this clade (with theinclusion of Chonemorpha G Don) is weaklysupported in Potgieter and Albert (2001) Po-tential characterizing traits could be a stylehead without stigmatic hollow in combina-tion with thecae that are adnate to the stylehead (Sennblad et al 1998)
DenitionmdashTribe Apocyneae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Apocynumand Trachelospermum but not Echiteae Mese-chiteae Periploceae or Asclepiadacina
Tribus Echiteae Bartl
Parsonsia Peltastes and Prestonia (allEchiteae) form a moderately supported orunsupported clade with the Periplocoideaerepresentatives of the study (uwboot andswboot lt 50 and uwboot lt 50 swboot D89 in the complete rbcL and the combinedrbcL and ndhF analysis respectively) In thestudy of Sennblad et al (1998) Prestonia andParsonsia formed a weakly supported cladeand in the study of Potgieter and Albert(2001) inclusion of both Parsonsia and EchitesP Browne in the Echiteae is supportedParsonsia and Echiteae also share certaincharacters such as a horseshoe-shaped reti-nacle and a style head with stigmatic hollow(see Sennblad et al 1998)
DenitionmdashTribe Echiteae is the most in-clusive clade in the Euapocynoidina thatincludes the primary types of Echites andPrestonia but not Apocyneae MesechiteaePeriploceae or Asclepiadacina
Tribus Periploceae BartlThe representatives of the Periplo-
coideae sensu Endress and Bruyns (2000)mdashPentopetia Parquetina Petopentia PeriplocaTacazzea and Mondiamdashgroup together ina well-supported clade (uwboot D 67swboot D 93) The monophyly of the tra-ditional periplocoid taxa is also supportedin earlier studies that included a wider ordifferent sampling (Civeyrel et al 1998
Potgieter and Albert 2001) The taxa ofthe traditional Asclepiadaceae have ldquotrans-latorsrdquo structures related to pollinationspecialization In the Periploceae the pollenis deposited as tetrads in a spoon-liketranslator with an adhesive disc
DenitionmdashTribe Periploceae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Periplocaand Pentopetia but not Apocyneae EchiteaeMesechiteae or Asclepiadacina
ASCLEPIADACINA
The taxa of the Asclepiadoideae sensuLiede and Albers (1994) form a clade togetherwith the two Secamone species and Baissea(Apocyneae uwboot lt 50 swboot D 73)The exclusion of Periploceae (traditionallyassociated with this clade see Civeyrel et al1998) from this taxon is weakly supportedTherefore Periploceae is not included as areference taxon in the stem-based part ofthe denition of the Asclepiadacina thusleaving open the possibility for a later in-clusion of this taxon In the Asclepiadacina(except Baissea) the pollen is agglutinatedinto pollinia that are connected to a claspingtranslator
DenitionmdashAsclepiadacina are the mostinclusive monophyletic clade in the Euapoc-ynoidina that includes the primary types ofAsclepias and Secamoneae but not ApocyneaeEchiteae or Mesechiteae
Tribus Secamoneae G DonThe tribe Secamoneae sensu Bruyns
(uwboot D 89 swboot D 96) is mono-phyletic in the present analysisHowever be-cause representatives from only one of theconstituent genera are included monophylyis not well tested but the study of Civeyrelet al (1998) which included a larger numberof taxa including Pervillea Decne showedthat monophyly of the tribe was supportedThe traditional character for this tribe is fourpollinia per translator
DenitionmdashTribe Secamoneae is the mostinclusive clade in the Asclepiadacina that in-cludes the primary types of Secamone andPervillea but not Asclepioidina or Baissea
BaisseaThe position of Baissea corresponds with
that in the study of Sennblad et al (1998)
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 399
system by using the names of the Linnaeanranks (eg species genus tribe familyorder) to indicate hierarchical level The prin-ciple of exhaustive subsidiary taxa is notaccepted so redundant taxa need not be rec-ognized Other ldquononstandardrdquo taxa couldbe dened by using Phylocode denitionsmdashthat is without the restraints discussed forstandard taxa above
One advantage of the described systemis that utilizing the correspondence betweenthe standard names and Linnaean ranks al-lows the system to be largely compatiblewiththe establishedLinnaean system A change ofsystem could therefore be gradual with thetwo systems coexisting during a conversiontime (see eg de Quieroz 1997b)
Even though the node C stem-baseddenitions reduce the sensitivity towardschanges in circumscriptions they may still besensitive towardldquounfortunaterdquo or ldquobadrdquo def-initions (see eg Cantino et al 1997Cantinoand de Queiroz 2000) Many of the recom-mendations for such things as choices of ref-erence taxa (speciers) in phylogenetic def-initions are applicable also to the node Cstem-based denitions (eg Schander andThollesson 1995 Bryant 1996 Cantino et al1997 Sereno 1999) Most likely the compro-mise system described above will have sev-eral further problems and we hope that thispaper will invite further discussion on thesubject
A New Classication of the Apocynaceae slOur results indicate that some problem-
atic taxa remain in the classication ofEndress and Bruyns (2000) Monophyly fortwo of the subfamilies and ve of the tribesis called into question If other recent molec-ular studies (eg Potgieter and Albert 2001)are taken into consideration a further fourtribes (Alstonieae Alyxieae Plumerieae andVinceae) may be nonmonophyletic This canin some cases be simply analysis artifactsbut as Endress and Bruyns (2000) themselvespoint out their classication is to be consid-ered preliminary Insufcient taxonsamplingin molecular studies and difcult homologydecisions relating to morphological charac-ters make the circumscription of many ofthe tribes for example in the Apocynoideaesensu Endress and Bruyns (2000) uncertainThere is thus risk for future rearrangementsin many of the tribes A classication using
node C stem-based denitions may be ableto subsume such rearrangements withoutnumerous explicit recircumscriptions Wewill here use the result fromthe present studyas a basis to propose such a classication ofthe Apocynaceae sl
Our aim is to base taxa on clades that arewell supported A second aim is to makeour classication compatible with the cur-rent system We will therefore primarily rec-ognize nonoverlapping tribes that are validunder the Linnaean system We will furtheradopt the principle of nested referencing (assuggested by Lee 1999 and Sereno 1999)from the Linnaean system Thus the rst ref-erence taxon in the node-based part of thedenition is the primary type which corre-sponds to the Linnaean type used for thename of the taxon Instead of the (optional)enumeration of subsumed taxa of a lowerrank (eg genera or subtribes) we will in-clude node C stem-based denitions as dis-cussed above Citations to primary types ofreference taxa in node-based part of de-nitions are also nested and thus refer toldquothe type specimen of the type species ofthe type genus of the dened taxonrdquo Aproblem is that the relationships between thetribes areweakly supported This may lead tovery cumbersome denitions enumeratingall competing tribes in the stem-based partof the denition (eg Moore 1998) One wayto reduce this problem is to dene interme-diate well-supported taxa and use them ascompeting taxaWe therefore will further rec-ognize four nonstandard taxa (note that anassignment of ldquononstandardrdquo taxa does notimply less reliability standard names relateto communication purposes only) The taxadiscussed are indicated in Figure 3 We willdiscuss the tree from the top of Figure 3 start-ing with the taxa of the traditional Plumeri-oideae Tribes and subtribes given in paren-theses in the text refer to the classicationsof Endress and Bruyns (2000) in other casesour classication is implied For some tribeswe have chosen a name with priority underLinnaean system that implies inclusion of agenus not sampled in the present study Forall but one (Willughbeieae) of these casesadditional molecular or morphological phy-logenetic analyses support the inclusion ofthe genus Even if too uncertain denitionsshould be avoided this is less critical withnode C stem-based denitions In the caseof Willughbeieae we have decided from
400 SYSTEMATIC BIOLOGY VOL 51
FIGURE 3 The tribal reclassication discussed in the text indicated on the combinable component consensustree from the successive weighting analysis of the complete rbcL data set Four nonstandard taxamdashApocynoidinaEuapocynoidina Asclepiadacina and Asclepiadoidinamdashare also indicated
discussions with a morphological expert(M Endress pers comm) that morphologi-cal support exists for including WillughbeiaRoxb in Willughbeieae If this is correctthe denition given below will be valid if
it is not our system allows redenition ofthe taxon We have also tried to providecomments on potential morphological char-acters taken from literature for the differ-ent taxa recognized However we have not
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 401
performed any morphological analysis butrely on information external to this studyfor these comments In most cases we cantherefore not identify synapomorphies forthe taxa
TAXONOMIC TREATMENT
Tribus Willughbeieae ADCThe Ancylobotrys Dictyophleba and Va-
hadenia clade is strongly supported (uwbootand swboot D 100) and corresponds to theWillughbeieae of Endress and Bruyns (2000)The present sampling contains closely re-lated genera that have been traditionallykept together Landolphia P Beauv sl in-cludes all these taxa (eg Schumann 1895)In addition to the taxa sampled here thestudy of Potgieter and Albert (2001) sup-ports the inclusion of Couma Aubl LacmelleaH Karst Pacouria Aubl and Saba (Pichon)Pichon in this clade Plants in this cladeall have a gynoecium that is congenitallysyncarpous and indehiscent and that con-tains a eshy pulp including numerous seedswith copious horny endosperm (Fallen 1986Persoon et al 1992) This probably con-stitutes synapomorphies for this clade(M Endress pers comm) although not un-ambiguously for example syncarpy is par-alleled in Carisseae We will therefore as-sume that the reference genus Willughbeia isincluded in this tribe
DenitionmdashTribe Willughbeieae is themost inclusive clade that includes the prim-ary types of Willughbeia and Dictyophleba butnot Alyxieae Aspidospermeae CarisseaeHunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Apocynoidina
Tribus Vinceae Bartl
Ochrosia Rauvola Vinca and Catharanthus(all Vinceae) form a strongly supported clade(uwboot D 89 swboot D 99) The associa-tion between Rauvola and Catharanthus wasindicated by Sennblad and Bremer (1996)The study of Potgieter and Albert (2001) sup-ports inclusion of the genera LaxoplumeriaMarkgr Neisosperma Raf and Tonduzia Pit-tier previously classied with AspidospermaThe Vinceae contain both herbaceous andfruticose taxa The fruits are apocarpousbut whereas the fruits of Catharanthus Lax-oplumeria Neisosperma Tonduzia and Vincahave dry fruit walls the fruits of Ochrosia andRauvola are drupes Plants in Vinceae share
a similar type of style head (ie the apicalenlarged part of the style) with a stigmatichollow and upper and lower hair wreaths(except in some species of Ochrosia) the pres-ence of a nectar disk and a linear hilum onthe seed (Pichon 1948b) On the present treeVinceae also include Kopsia however this po-sition of Kopsia is very weakly supported (seealso Sennblad and Bremer 1996)
DenitionmdashTribe Vinceae is the mostinclusive clade that includes the primarytypes of Vinca and Rauvola but not AlyxieaeAspidospermeae Carisseae HunterieaeMelodineae Plumerieae Tabernaemon-taneae Willughbeieae or Apocynoidina
Tribus Tabernaemontaneae G DonThe strongly supported association be-
tween Tabernaemontana Tabernanthe Carval-hoa Schizozygia and Molongum (Tabernae-montaneae uwboot D 84 swboot D 98)has been suggested previously (Boiteau et al1978 Fallen 1986 Endress et al 1996Sennblad and Bremer 1996) Inclusion of taxatraditionally associated with Tabernaemon-tana (eg Voacanga Thouars and CallichiliaStapf see also Tabernaemontaneae sensuLeeuwenberg 1994a) and Macoubea Aublis supported by the analysis of Potgieterand Albert (2001) The taxa share sclerenchy-matic anthers free from the style head anda characteristic band of heavily cutinizedcells at the insertion on the staminal rib(Endress et al 1996) The taxa tradition-ally associated with Tabernaemontana arecharacterized by their apocarpous fruitswith arillate seeds whereas genera asso-ciated with Molongum (Ambelanieae sensuLeeuwenberg 1994a) have syncarpous fruitsand lackanarillus HoweverMacoubea formsa link between the two having a syncar-pous fruit with arillate seeds (Zarucchi et al1995)
DenitionmdashTribe Tabernaemontaneae isthe most inclusive clade that includes the pri-mary types of Tabernaemontana and Schizozy-gia but not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae PlumerieaeVinceae Willughbeieae or Apocynoidina
Tribus Aspidospermeae Miers
Aspidosperma and Vallesia (Aspidosper-meae) form a strongly supported clade (uw-boot and swboot D 100) rst suggestedin Sennblad (1997 see also Sennblad and
402 SYSTEMATIC BIOLOGY VOL 51
Bremer 2000) Subsequently GeissospermumAllemao Haplophyton ADC MicroplumeriaBaill and Strempeliopsis Benth were indi-cated to belong to this clade (Potgieter andAlbert 2001) The Aspidospermeae havesimple style heads lacking both stigmatichollow and upper hair wreath and haveapocarpous fruits with variation in eshi-ness and sclerication Alstonia has tradi-tionally been placed in the Plumerieae Be-cause its seeds have a hairy margin it hasbeen suggested to form a link to subfam-ily Apocynoideae This was contradicted bySennblad and Bremerrsquos study (1996) how-ever where Alstonia was placed in an iso-lated position as the sister group to therest of the Apocynaceae sl Here it groupswith Aspidospermeae Sensu Endress andBruyns (2000) However this association isvery weakly supported (uwboot and swbootlt 50) and in other recent studies As-pidospermeae and Alstonia do not form aclade (Potgieter and Albert 2001) Becauseof this we do not use it as a reference taxonin the denition Thus although the Aspi-dospermeae include Alstonia on the presenttree some other position of Alstonia canbe accommodated without amending thedenition
DenitionmdashTribe Aspidospermeae is themost inclusive clade that includes theprimary types of Aspidosperma and Vallesiabut not Alyxieae Carisseae HunterieaeMelodineae Plumerieae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Melodineae G DonThe well-supported association between
Melodinus and Craspidospermum (uwboot D70 swboot D 75) was rst reportedby Sennblad (1997 see also Sennblad andBremer 2000) and has subsequently beenreected in the tribe Melodineae sensuEndress and Bruyns (2000 their inclusion ofDiplorhynchus however is not supported)Craspidospermum although its fruit is dryand dehiscent as opposed to the indehis-cent fruits of Melodinus was noted by Pichon(1948b) to present several characters of theCarisseae such as stamens inserted near thebase of the corolla tube and a syncarpousovary In particular Pichon found the verydense suprastaminal indumentum type rem-iniscent of Melodinus From his descriptions
of the two genera (Pichon 1948ab) the fol-lowing similarities also emerge presence ofa stipular line pollen in tetrads and a punc-tiform hilum
DenitionmdashTribe Melodineae is the mostinclusive clade that includes the pri-mary types of Melodinus and Craspidosper-mum but not Alyxieae AspidospermeaeCarisseae Hunterieae Plumerieae Taber-naemontaneae Vinceae Willughbeieae orApocynoidina
Tribus Hunterieae K Schum
The Picralima and Pleiocarpa clade isstrongly supported (uwboot and swboot D100) and corresponds to the subtribeHunterieae sensu Endress andBruyns (2000)The study of Potgieter and Albert (2001)also supports the inclusion of Hunteria RoxbThese taxa are characterized by apocar-pous sometimes pluricarpous ovaries andeshy brous fruit walls (Omino 1996)Diplorhynchus (Melodineae) is here weaklyassociated (uwboot and swboot lt 50) withthe Pleiocarpeae It shares a few similari-ties such as stipular lines and a style headwithout a stigmatic hollow and hair wreathsbut also has differences such as a dry de-hiscent fruit With the present denitionDiplorhynchus will tentatively be included inthe Hunterieae
DenitionmdashTribe Hunterieae is the mostinclusive clade that includes the primarytypes of Hunteria and Picralima but not Alyx-ieae Aspidospermeae Carisseae Melod-ineae Plumerieae TabernaemontaneaeVinceae Willughbeieae or Apocynoidina
Tribus Plumerieae EndlThe strongly supported relationship
(uwboot and swboot D 100) between Alla-manda and Plumeria corresponds to earlierstudies (Endress et al 1996 Sennblad andBremer 1996 Civeyrel et al 1998 Potgieterand Albert 2001) In all earlier treatmentsAllamanda has had uncertain relationshipsThe association with Plumeria nds supportin pollen morphology such as perforatemesocolpial depressions and similar innerexine pattern The two genera also containsecoiridoids rather than the indole alka-loids and cardenolides that are commonin the traditional Plumerioideae (Endresset al 1996) Cameraria Cerbera and Thevetiaform a strongly supported association
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 403
(uwboot D 83 swboot D 91) with Ane-chites which has been suggested by Fallen(1983) This association was based mainlyon the latrorsesublatrorse anthers andthe broad style head with large apicalappendages and stigmatic hollow The studyby Potgieter and Albert (2001) suggesteda further inclusion of Cerberiopsis Viell ExPancher amp Sebert and Skytanthus Meyen inthis tribe These taxa form a well-supportedclade with Plumeria and Allamanda whichcorresponds to the Plumerieae sensuEndress and Bruyns (2000) Characterssupporting this relationship are the presenceof infrastaminal as well as suprastaminalappendages (not present in Anechites andPlumeria) and winged seeds (not present inAnechites) (Endress et al 1996)
DenitionmdashTribe Plumerieae is themost inclusive clade that includes theprimary types of Plumeria and Allamandabut not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Carisseae Dumort
Acokanthera and Carissa (both Carisseae)are traditionally joined and are sometimeseven treated as one genus (eg Pichon1948a) In the present study they form astrongly supported clade (uwboot D 96swboot D 99) Among the characters sup-porting this tribe are syncarpous fruits with-out axile placentation and endocracks on theinner exine of the pollen (Endress et al 1996)
DenitionmdashTribe Carisseae is the most in-clusive clade that includes the primary typesof Carissa and Acokanthera but not AlyxieaeAspidospermeae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
Tribus Alyxieae G Don
Chilocarpus has been a genus with un-certain afnities In Endress et al (1996) itwas the sister taxon to the ApocynoideaeHere it is associated with Alyxia and Lep-inia in a clade corresponding to the Alyxieaesensu Endress and Bruyns (2000) Thisclade is strongly supported with succes-sive weighted bootstrap (swboot D 91) butnot with unit-weighted bootstrap support(uwboot D 58) such clades will henceforthbe termed moderatelywell supported clades
A further inclusion of Condylocarpon DesfLepiniopsis Valeton and Plectaneia Thouarswas indicated by Potgieter and Albert (2001)The constituent taxa are the only Apocy-naceae taxa outside the Apocynoidina thathave porate pollen
DenitionmdashTribe Alyxieae is the most in-clusive clade that includes the primary typesof Alyxia and Lepinia but not Aspidosper-meae Carisseae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
APOCYNOIDINA
On the present tree the traditionalsubfam-ily Apocynoideae includes all subfamilies(Periplocoideae Secamonoideae and Ascle-piadoideae) of the Asclepiadaceae sensuLiede and Albers (1994) All of these taxaare established names and selecting one asa subfamily at the expense of the othersmight be considered unfortunate from dif-ferent points of view We have here chosennot to recognize subfamilies instead we willdene the traditional Apocynoideae Ascle-piadaceae and Asclepiadoideae as nonstan-dard taxa The correspondence to the tradi-tional taxa will be indicated in the names byreplacing the sufx -eae with a neutral suf-x -ina (Kron 1997) We will also for practi-cal reasonsrecognize as a nonstandardtaxonthe informal group euapocynoids suggestedby Sennblad et al (1998)
The taxa of the traditional Apocynoideaeand Asclepiadaceae (uwboot D 86 swboot D99) clade will in many cases be more dif-cult to safely delimit to tribes becausemany of the groups in this clade have weakor no support (although preliminary resultsfrom an unpublished collaborative study bySennblad et al indicate additional supportfor the tribes discussed below) The tribaland subtribal classication of the traditionalApocynoideae has been shown to be prob-lematic (Leeuwenberg 1994a Endress andBruyns 2000) The circumscription of tribesfor these taxa might therefore in some casesbe preliminary
In all following taxa the basal part of theanther connective called the retinacle is ad-nate to the style head
DenitionmdashApocynoidina are the mostinclusive clade that includes the pri-mary types of Apocynum and Wrightia butnot Alyxieae Aspidospermeae Carisseae
404 SYSTEMATIC BIOLOGY VOL 51
Hunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Willughbeieae
Tribus Wrightieae G DonThe Stephanostema and Wrightia (both
Wrightieae) clade is strongly supported(uwboot and swboot D 100) and correspondsto the Wrightieae sensu Sennblad et al (1998)who also suggested an inclusion of Pleio-ceras Baill The tribe could be characterizedby a combination of synapomorphic and ple-siomorphic characters for example presenceofa chalazalandabsence of micropylar comaleft contorted aestivation and absence of airspaces in the anthers (1998)
DenitionmdashTribe Wrightieae is the mostinclusive clade in the Apocynoidina that in-cludes the primary types of Wrightia andStephanostema but not Malouetieae Nerieaeor Euapocynoidina
Tribus Nerieae (Benth) M Pichon
Adenium Nerium and Strophanthus (allWrightieae) constituted the Nerieae inSennblad et al (1998) however the cladereceived weak support Here Adenium andNerium form a clade that is absent in halfof the most-parsimonious trees in the com-plete rbcL analysis but is strongly supportedin the combined ndhF and rbcL analysis(uwboot D 98 swboot D 99) Strophan-thus groups with the Mascarenhasia Pachy-podium Kibatalia Funtumia and Holarrhenaclade which corresponds to the Malouetieaesensu Sennblad et al (1998) This position ofStrophanthus is however weakly supported(uwboot and swboot lt 50) and is not presentin the combined ndhF and rbcL tree Fur-thermore certain morphological characters(unfused slits in the corolla tube and anapical anther appendage) indicate a posi-tion closer to the Nerieae (Sennblad et al1998)
DenitionmdashTribe Nerieae is the most in-clusive clade in the Apocynoidina that in-cludes the primary types of Nerium and Ade-nium but not Malouetieae Wrightieae orEuapocynoidina
Tribus Malouetieae Mull-Arg
The association of Mascarenhasia Pachy-podium Kibatalia Funtumia and Holar-rhena (all Wrightieae) is moderately wellsupported in the complete rbcL analysis
(uwboot D 57 swboot D 74) whereasthe corresponding clade is well supportedin the combined rbcL and ndhF analysis(uwboot D 66 swboot D 100) On the ba-sis of identied synapomorphies (presenceof calcium oxalate crystals in the stomium ofthe anthers absence of interpetal vein ses-sile laments and absence of air spaces inthe anthers) Sennblad et al (1998) predictedan inclusion of Alaa Thouars Kibatalia andMalouetia ADC in the tribe The inclusion ofKibatalia is strongly supported in this studyFollowing these suggestions we will here as-sume the inclusion of the reference genusMalouetia in this tribe (The denition belowwill include Strophanthus on the present treebut see comments under Nerieae)
DenitionmdashTribe Malouetieae is the mostinclusive clade in the Apocynoidina that in-cluded the primary types of Malouetia andFuntumia but not Nerieae Wrightieae orEuapocynoidina
EUAPOCYNOIDINA
In all the remaining representatives ofthe Apocynoideae sensu Endress andBruyns (2000) that is Mandevilla MesechitesAganosma Apocynum TrachelospermumRhabdadenia Beaumontia Parsonsia andPrestonia the anthers are adnate to the stylehead both by the retinacle and by the thecaeThese taxa together with the taxa of the tra-ditional Asclepiadaceae form a moderatelysupported clade (uwboot lt 50 swboot D86) that corresponds to the informal taxoneuapocynoids sensu Sennblad et al (1998)
DenitionmdashEuapocynoidina are the mostinclusive clade that includes the primarytypes of Apocynum and Echites but notMalouetieae Wrightieae or Nerieae
Tribus Mesechiteae Miers
Mandevilla and Mesechites (both Mese-chiteae) form a moderately supported clade(uwboot lt 50 swboot D 87) equivalentto the Mesechiteae This tribe could be char-acterized by anthers with obtuse truncatedtails and a retinacle lacking hairs and also astrongly pentagonal style head with a stig-matic hollow
DenitionmdashTribe Mesechiteae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Mesechitesand Mandevilla but not Apocyneae EchiteaePeriploceae or Asclepiadacina
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 405
Tribus Apocyneae Bercht et J Presl
Aganosma Apocynum Beaumontia andTrachelospermum (Apocyneae) form a weaklysupported clade with Rhabdadenia (Echiteae)(uwboot and swboot lt 50) in the completerbcL analysis whereas the correspondingclade in the combined ndhF and rbcL analy-sis gains a little better support (uwboot lt 50swboot D 91) Likewise this clade (with theinclusion of Chonemorpha G Don) is weaklysupported in Potgieter and Albert (2001) Po-tential characterizing traits could be a stylehead without stigmatic hollow in combina-tion with thecae that are adnate to the stylehead (Sennblad et al 1998)
DenitionmdashTribe Apocyneae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Apocynumand Trachelospermum but not Echiteae Mese-chiteae Periploceae or Asclepiadacina
Tribus Echiteae Bartl
Parsonsia Peltastes and Prestonia (allEchiteae) form a moderately supported orunsupported clade with the Periplocoideaerepresentatives of the study (uwboot andswboot lt 50 and uwboot lt 50 swboot D89 in the complete rbcL and the combinedrbcL and ndhF analysis respectively) In thestudy of Sennblad et al (1998) Prestonia andParsonsia formed a weakly supported cladeand in the study of Potgieter and Albert(2001) inclusion of both Parsonsia and EchitesP Browne in the Echiteae is supportedParsonsia and Echiteae also share certaincharacters such as a horseshoe-shaped reti-nacle and a style head with stigmatic hollow(see Sennblad et al 1998)
DenitionmdashTribe Echiteae is the most in-clusive clade in the Euapocynoidina thatincludes the primary types of Echites andPrestonia but not Apocyneae MesechiteaePeriploceae or Asclepiadacina
Tribus Periploceae BartlThe representatives of the Periplo-
coideae sensu Endress and Bruyns (2000)mdashPentopetia Parquetina Petopentia PeriplocaTacazzea and Mondiamdashgroup together ina well-supported clade (uwboot D 67swboot D 93) The monophyly of the tra-ditional periplocoid taxa is also supportedin earlier studies that included a wider ordifferent sampling (Civeyrel et al 1998
Potgieter and Albert 2001) The taxa ofthe traditional Asclepiadaceae have ldquotrans-latorsrdquo structures related to pollinationspecialization In the Periploceae the pollenis deposited as tetrads in a spoon-liketranslator with an adhesive disc
DenitionmdashTribe Periploceae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Periplocaand Pentopetia but not Apocyneae EchiteaeMesechiteae or Asclepiadacina
ASCLEPIADACINA
The taxa of the Asclepiadoideae sensuLiede and Albers (1994) form a clade togetherwith the two Secamone species and Baissea(Apocyneae uwboot lt 50 swboot D 73)The exclusion of Periploceae (traditionallyassociated with this clade see Civeyrel et al1998) from this taxon is weakly supportedTherefore Periploceae is not included as areference taxon in the stem-based part ofthe denition of the Asclepiadacina thusleaving open the possibility for a later in-clusion of this taxon In the Asclepiadacina(except Baissea) the pollen is agglutinatedinto pollinia that are connected to a claspingtranslator
DenitionmdashAsclepiadacina are the mostinclusive monophyletic clade in the Euapoc-ynoidina that includes the primary types ofAsclepias and Secamoneae but not ApocyneaeEchiteae or Mesechiteae
Tribus Secamoneae G DonThe tribe Secamoneae sensu Bruyns
(uwboot D 89 swboot D 96) is mono-phyletic in the present analysisHowever be-cause representatives from only one of theconstituent genera are included monophylyis not well tested but the study of Civeyrelet al (1998) which included a larger numberof taxa including Pervillea Decne showedthat monophyly of the tribe was supportedThe traditional character for this tribe is fourpollinia per translator
DenitionmdashTribe Secamoneae is the mostinclusive clade in the Asclepiadacina that in-cludes the primary types of Secamone andPervillea but not Asclepioidina or Baissea
BaisseaThe position of Baissea corresponds with
that in the study of Sennblad et al (1998)
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
400 SYSTEMATIC BIOLOGY VOL 51
FIGURE 3 The tribal reclassication discussed in the text indicated on the combinable component consensustree from the successive weighting analysis of the complete rbcL data set Four nonstandard taxamdashApocynoidinaEuapocynoidina Asclepiadacina and Asclepiadoidinamdashare also indicated
discussions with a morphological expert(M Endress pers comm) that morphologi-cal support exists for including WillughbeiaRoxb in Willughbeieae If this is correctthe denition given below will be valid if
it is not our system allows redenition ofthe taxon We have also tried to providecomments on potential morphological char-acters taken from literature for the differ-ent taxa recognized However we have not
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 401
performed any morphological analysis butrely on information external to this studyfor these comments In most cases we cantherefore not identify synapomorphies forthe taxa
TAXONOMIC TREATMENT
Tribus Willughbeieae ADCThe Ancylobotrys Dictyophleba and Va-
hadenia clade is strongly supported (uwbootand swboot D 100) and corresponds to theWillughbeieae of Endress and Bruyns (2000)The present sampling contains closely re-lated genera that have been traditionallykept together Landolphia P Beauv sl in-cludes all these taxa (eg Schumann 1895)In addition to the taxa sampled here thestudy of Potgieter and Albert (2001) sup-ports the inclusion of Couma Aubl LacmelleaH Karst Pacouria Aubl and Saba (Pichon)Pichon in this clade Plants in this cladeall have a gynoecium that is congenitallysyncarpous and indehiscent and that con-tains a eshy pulp including numerous seedswith copious horny endosperm (Fallen 1986Persoon et al 1992) This probably con-stitutes synapomorphies for this clade(M Endress pers comm) although not un-ambiguously for example syncarpy is par-alleled in Carisseae We will therefore as-sume that the reference genus Willughbeia isincluded in this tribe
DenitionmdashTribe Willughbeieae is themost inclusive clade that includes the prim-ary types of Willughbeia and Dictyophleba butnot Alyxieae Aspidospermeae CarisseaeHunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Apocynoidina
Tribus Vinceae Bartl
Ochrosia Rauvola Vinca and Catharanthus(all Vinceae) form a strongly supported clade(uwboot D 89 swboot D 99) The associa-tion between Rauvola and Catharanthus wasindicated by Sennblad and Bremer (1996)The study of Potgieter and Albert (2001) sup-ports inclusion of the genera LaxoplumeriaMarkgr Neisosperma Raf and Tonduzia Pit-tier previously classied with AspidospermaThe Vinceae contain both herbaceous andfruticose taxa The fruits are apocarpousbut whereas the fruits of Catharanthus Lax-oplumeria Neisosperma Tonduzia and Vincahave dry fruit walls the fruits of Ochrosia andRauvola are drupes Plants in Vinceae share
a similar type of style head (ie the apicalenlarged part of the style) with a stigmatichollow and upper and lower hair wreaths(except in some species of Ochrosia) the pres-ence of a nectar disk and a linear hilum onthe seed (Pichon 1948b) On the present treeVinceae also include Kopsia however this po-sition of Kopsia is very weakly supported (seealso Sennblad and Bremer 1996)
DenitionmdashTribe Vinceae is the mostinclusive clade that includes the primarytypes of Vinca and Rauvola but not AlyxieaeAspidospermeae Carisseae HunterieaeMelodineae Plumerieae Tabernaemon-taneae Willughbeieae or Apocynoidina
Tribus Tabernaemontaneae G DonThe strongly supported association be-
tween Tabernaemontana Tabernanthe Carval-hoa Schizozygia and Molongum (Tabernae-montaneae uwboot D 84 swboot D 98)has been suggested previously (Boiteau et al1978 Fallen 1986 Endress et al 1996Sennblad and Bremer 1996) Inclusion of taxatraditionally associated with Tabernaemon-tana (eg Voacanga Thouars and CallichiliaStapf see also Tabernaemontaneae sensuLeeuwenberg 1994a) and Macoubea Aublis supported by the analysis of Potgieterand Albert (2001) The taxa share sclerenchy-matic anthers free from the style head anda characteristic band of heavily cutinizedcells at the insertion on the staminal rib(Endress et al 1996) The taxa tradition-ally associated with Tabernaemontana arecharacterized by their apocarpous fruitswith arillate seeds whereas genera asso-ciated with Molongum (Ambelanieae sensuLeeuwenberg 1994a) have syncarpous fruitsand lackanarillus HoweverMacoubea formsa link between the two having a syncar-pous fruit with arillate seeds (Zarucchi et al1995)
DenitionmdashTribe Tabernaemontaneae isthe most inclusive clade that includes the pri-mary types of Tabernaemontana and Schizozy-gia but not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae PlumerieaeVinceae Willughbeieae or Apocynoidina
Tribus Aspidospermeae Miers
Aspidosperma and Vallesia (Aspidosper-meae) form a strongly supported clade (uw-boot and swboot D 100) rst suggestedin Sennblad (1997 see also Sennblad and
402 SYSTEMATIC BIOLOGY VOL 51
Bremer 2000) Subsequently GeissospermumAllemao Haplophyton ADC MicroplumeriaBaill and Strempeliopsis Benth were indi-cated to belong to this clade (Potgieter andAlbert 2001) The Aspidospermeae havesimple style heads lacking both stigmatichollow and upper hair wreath and haveapocarpous fruits with variation in eshi-ness and sclerication Alstonia has tradi-tionally been placed in the Plumerieae Be-cause its seeds have a hairy margin it hasbeen suggested to form a link to subfam-ily Apocynoideae This was contradicted bySennblad and Bremerrsquos study (1996) how-ever where Alstonia was placed in an iso-lated position as the sister group to therest of the Apocynaceae sl Here it groupswith Aspidospermeae Sensu Endress andBruyns (2000) However this association isvery weakly supported (uwboot and swbootlt 50) and in other recent studies As-pidospermeae and Alstonia do not form aclade (Potgieter and Albert 2001) Becauseof this we do not use it as a reference taxonin the denition Thus although the Aspi-dospermeae include Alstonia on the presenttree some other position of Alstonia canbe accommodated without amending thedenition
DenitionmdashTribe Aspidospermeae is themost inclusive clade that includes theprimary types of Aspidosperma and Vallesiabut not Alyxieae Carisseae HunterieaeMelodineae Plumerieae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Melodineae G DonThe well-supported association between
Melodinus and Craspidospermum (uwboot D70 swboot D 75) was rst reportedby Sennblad (1997 see also Sennblad andBremer 2000) and has subsequently beenreected in the tribe Melodineae sensuEndress and Bruyns (2000 their inclusion ofDiplorhynchus however is not supported)Craspidospermum although its fruit is dryand dehiscent as opposed to the indehis-cent fruits of Melodinus was noted by Pichon(1948b) to present several characters of theCarisseae such as stamens inserted near thebase of the corolla tube and a syncarpousovary In particular Pichon found the verydense suprastaminal indumentum type rem-iniscent of Melodinus From his descriptions
of the two genera (Pichon 1948ab) the fol-lowing similarities also emerge presence ofa stipular line pollen in tetrads and a punc-tiform hilum
DenitionmdashTribe Melodineae is the mostinclusive clade that includes the pri-mary types of Melodinus and Craspidosper-mum but not Alyxieae AspidospermeaeCarisseae Hunterieae Plumerieae Taber-naemontaneae Vinceae Willughbeieae orApocynoidina
Tribus Hunterieae K Schum
The Picralima and Pleiocarpa clade isstrongly supported (uwboot and swboot D100) and corresponds to the subtribeHunterieae sensu Endress andBruyns (2000)The study of Potgieter and Albert (2001)also supports the inclusion of Hunteria RoxbThese taxa are characterized by apocar-pous sometimes pluricarpous ovaries andeshy brous fruit walls (Omino 1996)Diplorhynchus (Melodineae) is here weaklyassociated (uwboot and swboot lt 50) withthe Pleiocarpeae It shares a few similari-ties such as stipular lines and a style headwithout a stigmatic hollow and hair wreathsbut also has differences such as a dry de-hiscent fruit With the present denitionDiplorhynchus will tentatively be included inthe Hunterieae
DenitionmdashTribe Hunterieae is the mostinclusive clade that includes the primarytypes of Hunteria and Picralima but not Alyx-ieae Aspidospermeae Carisseae Melod-ineae Plumerieae TabernaemontaneaeVinceae Willughbeieae or Apocynoidina
Tribus Plumerieae EndlThe strongly supported relationship
(uwboot and swboot D 100) between Alla-manda and Plumeria corresponds to earlierstudies (Endress et al 1996 Sennblad andBremer 1996 Civeyrel et al 1998 Potgieterand Albert 2001) In all earlier treatmentsAllamanda has had uncertain relationshipsThe association with Plumeria nds supportin pollen morphology such as perforatemesocolpial depressions and similar innerexine pattern The two genera also containsecoiridoids rather than the indole alka-loids and cardenolides that are commonin the traditional Plumerioideae (Endresset al 1996) Cameraria Cerbera and Thevetiaform a strongly supported association
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 403
(uwboot D 83 swboot D 91) with Ane-chites which has been suggested by Fallen(1983) This association was based mainlyon the latrorsesublatrorse anthers andthe broad style head with large apicalappendages and stigmatic hollow The studyby Potgieter and Albert (2001) suggesteda further inclusion of Cerberiopsis Viell ExPancher amp Sebert and Skytanthus Meyen inthis tribe These taxa form a well-supportedclade with Plumeria and Allamanda whichcorresponds to the Plumerieae sensuEndress and Bruyns (2000) Characterssupporting this relationship are the presenceof infrastaminal as well as suprastaminalappendages (not present in Anechites andPlumeria) and winged seeds (not present inAnechites) (Endress et al 1996)
DenitionmdashTribe Plumerieae is themost inclusive clade that includes theprimary types of Plumeria and Allamandabut not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Carisseae Dumort
Acokanthera and Carissa (both Carisseae)are traditionally joined and are sometimeseven treated as one genus (eg Pichon1948a) In the present study they form astrongly supported clade (uwboot D 96swboot D 99) Among the characters sup-porting this tribe are syncarpous fruits with-out axile placentation and endocracks on theinner exine of the pollen (Endress et al 1996)
DenitionmdashTribe Carisseae is the most in-clusive clade that includes the primary typesof Carissa and Acokanthera but not AlyxieaeAspidospermeae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
Tribus Alyxieae G Don
Chilocarpus has been a genus with un-certain afnities In Endress et al (1996) itwas the sister taxon to the ApocynoideaeHere it is associated with Alyxia and Lep-inia in a clade corresponding to the Alyxieaesensu Endress and Bruyns (2000) Thisclade is strongly supported with succes-sive weighted bootstrap (swboot D 91) butnot with unit-weighted bootstrap support(uwboot D 58) such clades will henceforthbe termed moderatelywell supported clades
A further inclusion of Condylocarpon DesfLepiniopsis Valeton and Plectaneia Thouarswas indicated by Potgieter and Albert (2001)The constituent taxa are the only Apocy-naceae taxa outside the Apocynoidina thathave porate pollen
DenitionmdashTribe Alyxieae is the most in-clusive clade that includes the primary typesof Alyxia and Lepinia but not Aspidosper-meae Carisseae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
APOCYNOIDINA
On the present tree the traditionalsubfam-ily Apocynoideae includes all subfamilies(Periplocoideae Secamonoideae and Ascle-piadoideae) of the Asclepiadaceae sensuLiede and Albers (1994) All of these taxaare established names and selecting one asa subfamily at the expense of the othersmight be considered unfortunate from dif-ferent points of view We have here chosennot to recognize subfamilies instead we willdene the traditional Apocynoideae Ascle-piadaceae and Asclepiadoideae as nonstan-dard taxa The correspondence to the tradi-tional taxa will be indicated in the names byreplacing the sufx -eae with a neutral suf-x -ina (Kron 1997) We will also for practi-cal reasonsrecognize as a nonstandardtaxonthe informal group euapocynoids suggestedby Sennblad et al (1998)
The taxa of the traditional Apocynoideaeand Asclepiadaceae (uwboot D 86 swboot D99) clade will in many cases be more dif-cult to safely delimit to tribes becausemany of the groups in this clade have weakor no support (although preliminary resultsfrom an unpublished collaborative study bySennblad et al indicate additional supportfor the tribes discussed below) The tribaland subtribal classication of the traditionalApocynoideae has been shown to be prob-lematic (Leeuwenberg 1994a Endress andBruyns 2000) The circumscription of tribesfor these taxa might therefore in some casesbe preliminary
In all following taxa the basal part of theanther connective called the retinacle is ad-nate to the style head
DenitionmdashApocynoidina are the mostinclusive clade that includes the pri-mary types of Apocynum and Wrightia butnot Alyxieae Aspidospermeae Carisseae
404 SYSTEMATIC BIOLOGY VOL 51
Hunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Willughbeieae
Tribus Wrightieae G DonThe Stephanostema and Wrightia (both
Wrightieae) clade is strongly supported(uwboot and swboot D 100) and correspondsto the Wrightieae sensu Sennblad et al (1998)who also suggested an inclusion of Pleio-ceras Baill The tribe could be characterizedby a combination of synapomorphic and ple-siomorphic characters for example presenceofa chalazalandabsence of micropylar comaleft contorted aestivation and absence of airspaces in the anthers (1998)
DenitionmdashTribe Wrightieae is the mostinclusive clade in the Apocynoidina that in-cludes the primary types of Wrightia andStephanostema but not Malouetieae Nerieaeor Euapocynoidina
Tribus Nerieae (Benth) M Pichon
Adenium Nerium and Strophanthus (allWrightieae) constituted the Nerieae inSennblad et al (1998) however the cladereceived weak support Here Adenium andNerium form a clade that is absent in halfof the most-parsimonious trees in the com-plete rbcL analysis but is strongly supportedin the combined ndhF and rbcL analysis(uwboot D 98 swboot D 99) Strophan-thus groups with the Mascarenhasia Pachy-podium Kibatalia Funtumia and Holarrhenaclade which corresponds to the Malouetieaesensu Sennblad et al (1998) This position ofStrophanthus is however weakly supported(uwboot and swboot lt 50) and is not presentin the combined ndhF and rbcL tree Fur-thermore certain morphological characters(unfused slits in the corolla tube and anapical anther appendage) indicate a posi-tion closer to the Nerieae (Sennblad et al1998)
DenitionmdashTribe Nerieae is the most in-clusive clade in the Apocynoidina that in-cludes the primary types of Nerium and Ade-nium but not Malouetieae Wrightieae orEuapocynoidina
Tribus Malouetieae Mull-Arg
The association of Mascarenhasia Pachy-podium Kibatalia Funtumia and Holar-rhena (all Wrightieae) is moderately wellsupported in the complete rbcL analysis
(uwboot D 57 swboot D 74) whereasthe corresponding clade is well supportedin the combined rbcL and ndhF analysis(uwboot D 66 swboot D 100) On the ba-sis of identied synapomorphies (presenceof calcium oxalate crystals in the stomium ofthe anthers absence of interpetal vein ses-sile laments and absence of air spaces inthe anthers) Sennblad et al (1998) predictedan inclusion of Alaa Thouars Kibatalia andMalouetia ADC in the tribe The inclusion ofKibatalia is strongly supported in this studyFollowing these suggestions we will here as-sume the inclusion of the reference genusMalouetia in this tribe (The denition belowwill include Strophanthus on the present treebut see comments under Nerieae)
DenitionmdashTribe Malouetieae is the mostinclusive clade in the Apocynoidina that in-cluded the primary types of Malouetia andFuntumia but not Nerieae Wrightieae orEuapocynoidina
EUAPOCYNOIDINA
In all the remaining representatives ofthe Apocynoideae sensu Endress andBruyns (2000) that is Mandevilla MesechitesAganosma Apocynum TrachelospermumRhabdadenia Beaumontia Parsonsia andPrestonia the anthers are adnate to the stylehead both by the retinacle and by the thecaeThese taxa together with the taxa of the tra-ditional Asclepiadaceae form a moderatelysupported clade (uwboot lt 50 swboot D86) that corresponds to the informal taxoneuapocynoids sensu Sennblad et al (1998)
DenitionmdashEuapocynoidina are the mostinclusive clade that includes the primarytypes of Apocynum and Echites but notMalouetieae Wrightieae or Nerieae
Tribus Mesechiteae Miers
Mandevilla and Mesechites (both Mese-chiteae) form a moderately supported clade(uwboot lt 50 swboot D 87) equivalentto the Mesechiteae This tribe could be char-acterized by anthers with obtuse truncatedtails and a retinacle lacking hairs and also astrongly pentagonal style head with a stig-matic hollow
DenitionmdashTribe Mesechiteae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Mesechitesand Mandevilla but not Apocyneae EchiteaePeriploceae or Asclepiadacina
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 405
Tribus Apocyneae Bercht et J Presl
Aganosma Apocynum Beaumontia andTrachelospermum (Apocyneae) form a weaklysupported clade with Rhabdadenia (Echiteae)(uwboot and swboot lt 50) in the completerbcL analysis whereas the correspondingclade in the combined ndhF and rbcL analy-sis gains a little better support (uwboot lt 50swboot D 91) Likewise this clade (with theinclusion of Chonemorpha G Don) is weaklysupported in Potgieter and Albert (2001) Po-tential characterizing traits could be a stylehead without stigmatic hollow in combina-tion with thecae that are adnate to the stylehead (Sennblad et al 1998)
DenitionmdashTribe Apocyneae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Apocynumand Trachelospermum but not Echiteae Mese-chiteae Periploceae or Asclepiadacina
Tribus Echiteae Bartl
Parsonsia Peltastes and Prestonia (allEchiteae) form a moderately supported orunsupported clade with the Periplocoideaerepresentatives of the study (uwboot andswboot lt 50 and uwboot lt 50 swboot D89 in the complete rbcL and the combinedrbcL and ndhF analysis respectively) In thestudy of Sennblad et al (1998) Prestonia andParsonsia formed a weakly supported cladeand in the study of Potgieter and Albert(2001) inclusion of both Parsonsia and EchitesP Browne in the Echiteae is supportedParsonsia and Echiteae also share certaincharacters such as a horseshoe-shaped reti-nacle and a style head with stigmatic hollow(see Sennblad et al 1998)
DenitionmdashTribe Echiteae is the most in-clusive clade in the Euapocynoidina thatincludes the primary types of Echites andPrestonia but not Apocyneae MesechiteaePeriploceae or Asclepiadacina
Tribus Periploceae BartlThe representatives of the Periplo-
coideae sensu Endress and Bruyns (2000)mdashPentopetia Parquetina Petopentia PeriplocaTacazzea and Mondiamdashgroup together ina well-supported clade (uwboot D 67swboot D 93) The monophyly of the tra-ditional periplocoid taxa is also supportedin earlier studies that included a wider ordifferent sampling (Civeyrel et al 1998
Potgieter and Albert 2001) The taxa ofthe traditional Asclepiadaceae have ldquotrans-latorsrdquo structures related to pollinationspecialization In the Periploceae the pollenis deposited as tetrads in a spoon-liketranslator with an adhesive disc
DenitionmdashTribe Periploceae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Periplocaand Pentopetia but not Apocyneae EchiteaeMesechiteae or Asclepiadacina
ASCLEPIADACINA
The taxa of the Asclepiadoideae sensuLiede and Albers (1994) form a clade togetherwith the two Secamone species and Baissea(Apocyneae uwboot lt 50 swboot D 73)The exclusion of Periploceae (traditionallyassociated with this clade see Civeyrel et al1998) from this taxon is weakly supportedTherefore Periploceae is not included as areference taxon in the stem-based part ofthe denition of the Asclepiadacina thusleaving open the possibility for a later in-clusion of this taxon In the Asclepiadacina(except Baissea) the pollen is agglutinatedinto pollinia that are connected to a claspingtranslator
DenitionmdashAsclepiadacina are the mostinclusive monophyletic clade in the Euapoc-ynoidina that includes the primary types ofAsclepias and Secamoneae but not ApocyneaeEchiteae or Mesechiteae
Tribus Secamoneae G DonThe tribe Secamoneae sensu Bruyns
(uwboot D 89 swboot D 96) is mono-phyletic in the present analysisHowever be-cause representatives from only one of theconstituent genera are included monophylyis not well tested but the study of Civeyrelet al (1998) which included a larger numberof taxa including Pervillea Decne showedthat monophyly of the tribe was supportedThe traditional character for this tribe is fourpollinia per translator
DenitionmdashTribe Secamoneae is the mostinclusive clade in the Asclepiadacina that in-cludes the primary types of Secamone andPervillea but not Asclepioidina or Baissea
BaisseaThe position of Baissea corresponds with
that in the study of Sennblad et al (1998)
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 401
performed any morphological analysis butrely on information external to this studyfor these comments In most cases we cantherefore not identify synapomorphies forthe taxa
TAXONOMIC TREATMENT
Tribus Willughbeieae ADCThe Ancylobotrys Dictyophleba and Va-
hadenia clade is strongly supported (uwbootand swboot D 100) and corresponds to theWillughbeieae of Endress and Bruyns (2000)The present sampling contains closely re-lated genera that have been traditionallykept together Landolphia P Beauv sl in-cludes all these taxa (eg Schumann 1895)In addition to the taxa sampled here thestudy of Potgieter and Albert (2001) sup-ports the inclusion of Couma Aubl LacmelleaH Karst Pacouria Aubl and Saba (Pichon)Pichon in this clade Plants in this cladeall have a gynoecium that is congenitallysyncarpous and indehiscent and that con-tains a eshy pulp including numerous seedswith copious horny endosperm (Fallen 1986Persoon et al 1992) This probably con-stitutes synapomorphies for this clade(M Endress pers comm) although not un-ambiguously for example syncarpy is par-alleled in Carisseae We will therefore as-sume that the reference genus Willughbeia isincluded in this tribe
DenitionmdashTribe Willughbeieae is themost inclusive clade that includes the prim-ary types of Willughbeia and Dictyophleba butnot Alyxieae Aspidospermeae CarisseaeHunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Apocynoidina
Tribus Vinceae Bartl
Ochrosia Rauvola Vinca and Catharanthus(all Vinceae) form a strongly supported clade(uwboot D 89 swboot D 99) The associa-tion between Rauvola and Catharanthus wasindicated by Sennblad and Bremer (1996)The study of Potgieter and Albert (2001) sup-ports inclusion of the genera LaxoplumeriaMarkgr Neisosperma Raf and Tonduzia Pit-tier previously classied with AspidospermaThe Vinceae contain both herbaceous andfruticose taxa The fruits are apocarpousbut whereas the fruits of Catharanthus Lax-oplumeria Neisosperma Tonduzia and Vincahave dry fruit walls the fruits of Ochrosia andRauvola are drupes Plants in Vinceae share
a similar type of style head (ie the apicalenlarged part of the style) with a stigmatichollow and upper and lower hair wreaths(except in some species of Ochrosia) the pres-ence of a nectar disk and a linear hilum onthe seed (Pichon 1948b) On the present treeVinceae also include Kopsia however this po-sition of Kopsia is very weakly supported (seealso Sennblad and Bremer 1996)
DenitionmdashTribe Vinceae is the mostinclusive clade that includes the primarytypes of Vinca and Rauvola but not AlyxieaeAspidospermeae Carisseae HunterieaeMelodineae Plumerieae Tabernaemon-taneae Willughbeieae or Apocynoidina
Tribus Tabernaemontaneae G DonThe strongly supported association be-
tween Tabernaemontana Tabernanthe Carval-hoa Schizozygia and Molongum (Tabernae-montaneae uwboot D 84 swboot D 98)has been suggested previously (Boiteau et al1978 Fallen 1986 Endress et al 1996Sennblad and Bremer 1996) Inclusion of taxatraditionally associated with Tabernaemon-tana (eg Voacanga Thouars and CallichiliaStapf see also Tabernaemontaneae sensuLeeuwenberg 1994a) and Macoubea Aublis supported by the analysis of Potgieterand Albert (2001) The taxa share sclerenchy-matic anthers free from the style head anda characteristic band of heavily cutinizedcells at the insertion on the staminal rib(Endress et al 1996) The taxa tradition-ally associated with Tabernaemontana arecharacterized by their apocarpous fruitswith arillate seeds whereas genera asso-ciated with Molongum (Ambelanieae sensuLeeuwenberg 1994a) have syncarpous fruitsand lackanarillus HoweverMacoubea formsa link between the two having a syncar-pous fruit with arillate seeds (Zarucchi et al1995)
DenitionmdashTribe Tabernaemontaneae isthe most inclusive clade that includes the pri-mary types of Tabernaemontana and Schizozy-gia but not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae PlumerieaeVinceae Willughbeieae or Apocynoidina
Tribus Aspidospermeae Miers
Aspidosperma and Vallesia (Aspidosper-meae) form a strongly supported clade (uw-boot and swboot D 100) rst suggestedin Sennblad (1997 see also Sennblad and
402 SYSTEMATIC BIOLOGY VOL 51
Bremer 2000) Subsequently GeissospermumAllemao Haplophyton ADC MicroplumeriaBaill and Strempeliopsis Benth were indi-cated to belong to this clade (Potgieter andAlbert 2001) The Aspidospermeae havesimple style heads lacking both stigmatichollow and upper hair wreath and haveapocarpous fruits with variation in eshi-ness and sclerication Alstonia has tradi-tionally been placed in the Plumerieae Be-cause its seeds have a hairy margin it hasbeen suggested to form a link to subfam-ily Apocynoideae This was contradicted bySennblad and Bremerrsquos study (1996) how-ever where Alstonia was placed in an iso-lated position as the sister group to therest of the Apocynaceae sl Here it groupswith Aspidospermeae Sensu Endress andBruyns (2000) However this association isvery weakly supported (uwboot and swbootlt 50) and in other recent studies As-pidospermeae and Alstonia do not form aclade (Potgieter and Albert 2001) Becauseof this we do not use it as a reference taxonin the denition Thus although the Aspi-dospermeae include Alstonia on the presenttree some other position of Alstonia canbe accommodated without amending thedenition
DenitionmdashTribe Aspidospermeae is themost inclusive clade that includes theprimary types of Aspidosperma and Vallesiabut not Alyxieae Carisseae HunterieaeMelodineae Plumerieae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Melodineae G DonThe well-supported association between
Melodinus and Craspidospermum (uwboot D70 swboot D 75) was rst reportedby Sennblad (1997 see also Sennblad andBremer 2000) and has subsequently beenreected in the tribe Melodineae sensuEndress and Bruyns (2000 their inclusion ofDiplorhynchus however is not supported)Craspidospermum although its fruit is dryand dehiscent as opposed to the indehis-cent fruits of Melodinus was noted by Pichon(1948b) to present several characters of theCarisseae such as stamens inserted near thebase of the corolla tube and a syncarpousovary In particular Pichon found the verydense suprastaminal indumentum type rem-iniscent of Melodinus From his descriptions
of the two genera (Pichon 1948ab) the fol-lowing similarities also emerge presence ofa stipular line pollen in tetrads and a punc-tiform hilum
DenitionmdashTribe Melodineae is the mostinclusive clade that includes the pri-mary types of Melodinus and Craspidosper-mum but not Alyxieae AspidospermeaeCarisseae Hunterieae Plumerieae Taber-naemontaneae Vinceae Willughbeieae orApocynoidina
Tribus Hunterieae K Schum
The Picralima and Pleiocarpa clade isstrongly supported (uwboot and swboot D100) and corresponds to the subtribeHunterieae sensu Endress andBruyns (2000)The study of Potgieter and Albert (2001)also supports the inclusion of Hunteria RoxbThese taxa are characterized by apocar-pous sometimes pluricarpous ovaries andeshy brous fruit walls (Omino 1996)Diplorhynchus (Melodineae) is here weaklyassociated (uwboot and swboot lt 50) withthe Pleiocarpeae It shares a few similari-ties such as stipular lines and a style headwithout a stigmatic hollow and hair wreathsbut also has differences such as a dry de-hiscent fruit With the present denitionDiplorhynchus will tentatively be included inthe Hunterieae
DenitionmdashTribe Hunterieae is the mostinclusive clade that includes the primarytypes of Hunteria and Picralima but not Alyx-ieae Aspidospermeae Carisseae Melod-ineae Plumerieae TabernaemontaneaeVinceae Willughbeieae or Apocynoidina
Tribus Plumerieae EndlThe strongly supported relationship
(uwboot and swboot D 100) between Alla-manda and Plumeria corresponds to earlierstudies (Endress et al 1996 Sennblad andBremer 1996 Civeyrel et al 1998 Potgieterand Albert 2001) In all earlier treatmentsAllamanda has had uncertain relationshipsThe association with Plumeria nds supportin pollen morphology such as perforatemesocolpial depressions and similar innerexine pattern The two genera also containsecoiridoids rather than the indole alka-loids and cardenolides that are commonin the traditional Plumerioideae (Endresset al 1996) Cameraria Cerbera and Thevetiaform a strongly supported association
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 403
(uwboot D 83 swboot D 91) with Ane-chites which has been suggested by Fallen(1983) This association was based mainlyon the latrorsesublatrorse anthers andthe broad style head with large apicalappendages and stigmatic hollow The studyby Potgieter and Albert (2001) suggesteda further inclusion of Cerberiopsis Viell ExPancher amp Sebert and Skytanthus Meyen inthis tribe These taxa form a well-supportedclade with Plumeria and Allamanda whichcorresponds to the Plumerieae sensuEndress and Bruyns (2000) Characterssupporting this relationship are the presenceof infrastaminal as well as suprastaminalappendages (not present in Anechites andPlumeria) and winged seeds (not present inAnechites) (Endress et al 1996)
DenitionmdashTribe Plumerieae is themost inclusive clade that includes theprimary types of Plumeria and Allamandabut not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Carisseae Dumort
Acokanthera and Carissa (both Carisseae)are traditionally joined and are sometimeseven treated as one genus (eg Pichon1948a) In the present study they form astrongly supported clade (uwboot D 96swboot D 99) Among the characters sup-porting this tribe are syncarpous fruits with-out axile placentation and endocracks on theinner exine of the pollen (Endress et al 1996)
DenitionmdashTribe Carisseae is the most in-clusive clade that includes the primary typesof Carissa and Acokanthera but not AlyxieaeAspidospermeae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
Tribus Alyxieae G Don
Chilocarpus has been a genus with un-certain afnities In Endress et al (1996) itwas the sister taxon to the ApocynoideaeHere it is associated with Alyxia and Lep-inia in a clade corresponding to the Alyxieaesensu Endress and Bruyns (2000) Thisclade is strongly supported with succes-sive weighted bootstrap (swboot D 91) butnot with unit-weighted bootstrap support(uwboot D 58) such clades will henceforthbe termed moderatelywell supported clades
A further inclusion of Condylocarpon DesfLepiniopsis Valeton and Plectaneia Thouarswas indicated by Potgieter and Albert (2001)The constituent taxa are the only Apocy-naceae taxa outside the Apocynoidina thathave porate pollen
DenitionmdashTribe Alyxieae is the most in-clusive clade that includes the primary typesof Alyxia and Lepinia but not Aspidosper-meae Carisseae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
APOCYNOIDINA
On the present tree the traditionalsubfam-ily Apocynoideae includes all subfamilies(Periplocoideae Secamonoideae and Ascle-piadoideae) of the Asclepiadaceae sensuLiede and Albers (1994) All of these taxaare established names and selecting one asa subfamily at the expense of the othersmight be considered unfortunate from dif-ferent points of view We have here chosennot to recognize subfamilies instead we willdene the traditional Apocynoideae Ascle-piadaceae and Asclepiadoideae as nonstan-dard taxa The correspondence to the tradi-tional taxa will be indicated in the names byreplacing the sufx -eae with a neutral suf-x -ina (Kron 1997) We will also for practi-cal reasonsrecognize as a nonstandardtaxonthe informal group euapocynoids suggestedby Sennblad et al (1998)
The taxa of the traditional Apocynoideaeand Asclepiadaceae (uwboot D 86 swboot D99) clade will in many cases be more dif-cult to safely delimit to tribes becausemany of the groups in this clade have weakor no support (although preliminary resultsfrom an unpublished collaborative study bySennblad et al indicate additional supportfor the tribes discussed below) The tribaland subtribal classication of the traditionalApocynoideae has been shown to be prob-lematic (Leeuwenberg 1994a Endress andBruyns 2000) The circumscription of tribesfor these taxa might therefore in some casesbe preliminary
In all following taxa the basal part of theanther connective called the retinacle is ad-nate to the style head
DenitionmdashApocynoidina are the mostinclusive clade that includes the pri-mary types of Apocynum and Wrightia butnot Alyxieae Aspidospermeae Carisseae
404 SYSTEMATIC BIOLOGY VOL 51
Hunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Willughbeieae
Tribus Wrightieae G DonThe Stephanostema and Wrightia (both
Wrightieae) clade is strongly supported(uwboot and swboot D 100) and correspondsto the Wrightieae sensu Sennblad et al (1998)who also suggested an inclusion of Pleio-ceras Baill The tribe could be characterizedby a combination of synapomorphic and ple-siomorphic characters for example presenceofa chalazalandabsence of micropylar comaleft contorted aestivation and absence of airspaces in the anthers (1998)
DenitionmdashTribe Wrightieae is the mostinclusive clade in the Apocynoidina that in-cludes the primary types of Wrightia andStephanostema but not Malouetieae Nerieaeor Euapocynoidina
Tribus Nerieae (Benth) M Pichon
Adenium Nerium and Strophanthus (allWrightieae) constituted the Nerieae inSennblad et al (1998) however the cladereceived weak support Here Adenium andNerium form a clade that is absent in halfof the most-parsimonious trees in the com-plete rbcL analysis but is strongly supportedin the combined ndhF and rbcL analysis(uwboot D 98 swboot D 99) Strophan-thus groups with the Mascarenhasia Pachy-podium Kibatalia Funtumia and Holarrhenaclade which corresponds to the Malouetieaesensu Sennblad et al (1998) This position ofStrophanthus is however weakly supported(uwboot and swboot lt 50) and is not presentin the combined ndhF and rbcL tree Fur-thermore certain morphological characters(unfused slits in the corolla tube and anapical anther appendage) indicate a posi-tion closer to the Nerieae (Sennblad et al1998)
DenitionmdashTribe Nerieae is the most in-clusive clade in the Apocynoidina that in-cludes the primary types of Nerium and Ade-nium but not Malouetieae Wrightieae orEuapocynoidina
Tribus Malouetieae Mull-Arg
The association of Mascarenhasia Pachy-podium Kibatalia Funtumia and Holar-rhena (all Wrightieae) is moderately wellsupported in the complete rbcL analysis
(uwboot D 57 swboot D 74) whereasthe corresponding clade is well supportedin the combined rbcL and ndhF analysis(uwboot D 66 swboot D 100) On the ba-sis of identied synapomorphies (presenceof calcium oxalate crystals in the stomium ofthe anthers absence of interpetal vein ses-sile laments and absence of air spaces inthe anthers) Sennblad et al (1998) predictedan inclusion of Alaa Thouars Kibatalia andMalouetia ADC in the tribe The inclusion ofKibatalia is strongly supported in this studyFollowing these suggestions we will here as-sume the inclusion of the reference genusMalouetia in this tribe (The denition belowwill include Strophanthus on the present treebut see comments under Nerieae)
DenitionmdashTribe Malouetieae is the mostinclusive clade in the Apocynoidina that in-cluded the primary types of Malouetia andFuntumia but not Nerieae Wrightieae orEuapocynoidina
EUAPOCYNOIDINA
In all the remaining representatives ofthe Apocynoideae sensu Endress andBruyns (2000) that is Mandevilla MesechitesAganosma Apocynum TrachelospermumRhabdadenia Beaumontia Parsonsia andPrestonia the anthers are adnate to the stylehead both by the retinacle and by the thecaeThese taxa together with the taxa of the tra-ditional Asclepiadaceae form a moderatelysupported clade (uwboot lt 50 swboot D86) that corresponds to the informal taxoneuapocynoids sensu Sennblad et al (1998)
DenitionmdashEuapocynoidina are the mostinclusive clade that includes the primarytypes of Apocynum and Echites but notMalouetieae Wrightieae or Nerieae
Tribus Mesechiteae Miers
Mandevilla and Mesechites (both Mese-chiteae) form a moderately supported clade(uwboot lt 50 swboot D 87) equivalentto the Mesechiteae This tribe could be char-acterized by anthers with obtuse truncatedtails and a retinacle lacking hairs and also astrongly pentagonal style head with a stig-matic hollow
DenitionmdashTribe Mesechiteae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Mesechitesand Mandevilla but not Apocyneae EchiteaePeriploceae or Asclepiadacina
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 405
Tribus Apocyneae Bercht et J Presl
Aganosma Apocynum Beaumontia andTrachelospermum (Apocyneae) form a weaklysupported clade with Rhabdadenia (Echiteae)(uwboot and swboot lt 50) in the completerbcL analysis whereas the correspondingclade in the combined ndhF and rbcL analy-sis gains a little better support (uwboot lt 50swboot D 91) Likewise this clade (with theinclusion of Chonemorpha G Don) is weaklysupported in Potgieter and Albert (2001) Po-tential characterizing traits could be a stylehead without stigmatic hollow in combina-tion with thecae that are adnate to the stylehead (Sennblad et al 1998)
DenitionmdashTribe Apocyneae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Apocynumand Trachelospermum but not Echiteae Mese-chiteae Periploceae or Asclepiadacina
Tribus Echiteae Bartl
Parsonsia Peltastes and Prestonia (allEchiteae) form a moderately supported orunsupported clade with the Periplocoideaerepresentatives of the study (uwboot andswboot lt 50 and uwboot lt 50 swboot D89 in the complete rbcL and the combinedrbcL and ndhF analysis respectively) In thestudy of Sennblad et al (1998) Prestonia andParsonsia formed a weakly supported cladeand in the study of Potgieter and Albert(2001) inclusion of both Parsonsia and EchitesP Browne in the Echiteae is supportedParsonsia and Echiteae also share certaincharacters such as a horseshoe-shaped reti-nacle and a style head with stigmatic hollow(see Sennblad et al 1998)
DenitionmdashTribe Echiteae is the most in-clusive clade in the Euapocynoidina thatincludes the primary types of Echites andPrestonia but not Apocyneae MesechiteaePeriploceae or Asclepiadacina
Tribus Periploceae BartlThe representatives of the Periplo-
coideae sensu Endress and Bruyns (2000)mdashPentopetia Parquetina Petopentia PeriplocaTacazzea and Mondiamdashgroup together ina well-supported clade (uwboot D 67swboot D 93) The monophyly of the tra-ditional periplocoid taxa is also supportedin earlier studies that included a wider ordifferent sampling (Civeyrel et al 1998
Potgieter and Albert 2001) The taxa ofthe traditional Asclepiadaceae have ldquotrans-latorsrdquo structures related to pollinationspecialization In the Periploceae the pollenis deposited as tetrads in a spoon-liketranslator with an adhesive disc
DenitionmdashTribe Periploceae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Periplocaand Pentopetia but not Apocyneae EchiteaeMesechiteae or Asclepiadacina
ASCLEPIADACINA
The taxa of the Asclepiadoideae sensuLiede and Albers (1994) form a clade togetherwith the two Secamone species and Baissea(Apocyneae uwboot lt 50 swboot D 73)The exclusion of Periploceae (traditionallyassociated with this clade see Civeyrel et al1998) from this taxon is weakly supportedTherefore Periploceae is not included as areference taxon in the stem-based part ofthe denition of the Asclepiadacina thusleaving open the possibility for a later in-clusion of this taxon In the Asclepiadacina(except Baissea) the pollen is agglutinatedinto pollinia that are connected to a claspingtranslator
DenitionmdashAsclepiadacina are the mostinclusive monophyletic clade in the Euapoc-ynoidina that includes the primary types ofAsclepias and Secamoneae but not ApocyneaeEchiteae or Mesechiteae
Tribus Secamoneae G DonThe tribe Secamoneae sensu Bruyns
(uwboot D 89 swboot D 96) is mono-phyletic in the present analysisHowever be-cause representatives from only one of theconstituent genera are included monophylyis not well tested but the study of Civeyrelet al (1998) which included a larger numberof taxa including Pervillea Decne showedthat monophyly of the tribe was supportedThe traditional character for this tribe is fourpollinia per translator
DenitionmdashTribe Secamoneae is the mostinclusive clade in the Asclepiadacina that in-cludes the primary types of Secamone andPervillea but not Asclepioidina or Baissea
BaisseaThe position of Baissea corresponds with
that in the study of Sennblad et al (1998)
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
402 SYSTEMATIC BIOLOGY VOL 51
Bremer 2000) Subsequently GeissospermumAllemao Haplophyton ADC MicroplumeriaBaill and Strempeliopsis Benth were indi-cated to belong to this clade (Potgieter andAlbert 2001) The Aspidospermeae havesimple style heads lacking both stigmatichollow and upper hair wreath and haveapocarpous fruits with variation in eshi-ness and sclerication Alstonia has tradi-tionally been placed in the Plumerieae Be-cause its seeds have a hairy margin it hasbeen suggested to form a link to subfam-ily Apocynoideae This was contradicted bySennblad and Bremerrsquos study (1996) how-ever where Alstonia was placed in an iso-lated position as the sister group to therest of the Apocynaceae sl Here it groupswith Aspidospermeae Sensu Endress andBruyns (2000) However this association isvery weakly supported (uwboot and swbootlt 50) and in other recent studies As-pidospermeae and Alstonia do not form aclade (Potgieter and Albert 2001) Becauseof this we do not use it as a reference taxonin the denition Thus although the Aspi-dospermeae include Alstonia on the presenttree some other position of Alstonia canbe accommodated without amending thedenition
DenitionmdashTribe Aspidospermeae is themost inclusive clade that includes theprimary types of Aspidosperma and Vallesiabut not Alyxieae Carisseae HunterieaeMelodineae Plumerieae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Melodineae G DonThe well-supported association between
Melodinus and Craspidospermum (uwboot D70 swboot D 75) was rst reportedby Sennblad (1997 see also Sennblad andBremer 2000) and has subsequently beenreected in the tribe Melodineae sensuEndress and Bruyns (2000 their inclusion ofDiplorhynchus however is not supported)Craspidospermum although its fruit is dryand dehiscent as opposed to the indehis-cent fruits of Melodinus was noted by Pichon(1948b) to present several characters of theCarisseae such as stamens inserted near thebase of the corolla tube and a syncarpousovary In particular Pichon found the verydense suprastaminal indumentum type rem-iniscent of Melodinus From his descriptions
of the two genera (Pichon 1948ab) the fol-lowing similarities also emerge presence ofa stipular line pollen in tetrads and a punc-tiform hilum
DenitionmdashTribe Melodineae is the mostinclusive clade that includes the pri-mary types of Melodinus and Craspidosper-mum but not Alyxieae AspidospermeaeCarisseae Hunterieae Plumerieae Taber-naemontaneae Vinceae Willughbeieae orApocynoidina
Tribus Hunterieae K Schum
The Picralima and Pleiocarpa clade isstrongly supported (uwboot and swboot D100) and corresponds to the subtribeHunterieae sensu Endress andBruyns (2000)The study of Potgieter and Albert (2001)also supports the inclusion of Hunteria RoxbThese taxa are characterized by apocar-pous sometimes pluricarpous ovaries andeshy brous fruit walls (Omino 1996)Diplorhynchus (Melodineae) is here weaklyassociated (uwboot and swboot lt 50) withthe Pleiocarpeae It shares a few similari-ties such as stipular lines and a style headwithout a stigmatic hollow and hair wreathsbut also has differences such as a dry de-hiscent fruit With the present denitionDiplorhynchus will tentatively be included inthe Hunterieae
DenitionmdashTribe Hunterieae is the mostinclusive clade that includes the primarytypes of Hunteria and Picralima but not Alyx-ieae Aspidospermeae Carisseae Melod-ineae Plumerieae TabernaemontaneaeVinceae Willughbeieae or Apocynoidina
Tribus Plumerieae EndlThe strongly supported relationship
(uwboot and swboot D 100) between Alla-manda and Plumeria corresponds to earlierstudies (Endress et al 1996 Sennblad andBremer 1996 Civeyrel et al 1998 Potgieterand Albert 2001) In all earlier treatmentsAllamanda has had uncertain relationshipsThe association with Plumeria nds supportin pollen morphology such as perforatemesocolpial depressions and similar innerexine pattern The two genera also containsecoiridoids rather than the indole alka-loids and cardenolides that are commonin the traditional Plumerioideae (Endresset al 1996) Cameraria Cerbera and Thevetiaform a strongly supported association
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 403
(uwboot D 83 swboot D 91) with Ane-chites which has been suggested by Fallen(1983) This association was based mainlyon the latrorsesublatrorse anthers andthe broad style head with large apicalappendages and stigmatic hollow The studyby Potgieter and Albert (2001) suggesteda further inclusion of Cerberiopsis Viell ExPancher amp Sebert and Skytanthus Meyen inthis tribe These taxa form a well-supportedclade with Plumeria and Allamanda whichcorresponds to the Plumerieae sensuEndress and Bruyns (2000) Characterssupporting this relationship are the presenceof infrastaminal as well as suprastaminalappendages (not present in Anechites andPlumeria) and winged seeds (not present inAnechites) (Endress et al 1996)
DenitionmdashTribe Plumerieae is themost inclusive clade that includes theprimary types of Plumeria and Allamandabut not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Carisseae Dumort
Acokanthera and Carissa (both Carisseae)are traditionally joined and are sometimeseven treated as one genus (eg Pichon1948a) In the present study they form astrongly supported clade (uwboot D 96swboot D 99) Among the characters sup-porting this tribe are syncarpous fruits with-out axile placentation and endocracks on theinner exine of the pollen (Endress et al 1996)
DenitionmdashTribe Carisseae is the most in-clusive clade that includes the primary typesof Carissa and Acokanthera but not AlyxieaeAspidospermeae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
Tribus Alyxieae G Don
Chilocarpus has been a genus with un-certain afnities In Endress et al (1996) itwas the sister taxon to the ApocynoideaeHere it is associated with Alyxia and Lep-inia in a clade corresponding to the Alyxieaesensu Endress and Bruyns (2000) Thisclade is strongly supported with succes-sive weighted bootstrap (swboot D 91) butnot with unit-weighted bootstrap support(uwboot D 58) such clades will henceforthbe termed moderatelywell supported clades
A further inclusion of Condylocarpon DesfLepiniopsis Valeton and Plectaneia Thouarswas indicated by Potgieter and Albert (2001)The constituent taxa are the only Apocy-naceae taxa outside the Apocynoidina thathave porate pollen
DenitionmdashTribe Alyxieae is the most in-clusive clade that includes the primary typesof Alyxia and Lepinia but not Aspidosper-meae Carisseae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
APOCYNOIDINA
On the present tree the traditionalsubfam-ily Apocynoideae includes all subfamilies(Periplocoideae Secamonoideae and Ascle-piadoideae) of the Asclepiadaceae sensuLiede and Albers (1994) All of these taxaare established names and selecting one asa subfamily at the expense of the othersmight be considered unfortunate from dif-ferent points of view We have here chosennot to recognize subfamilies instead we willdene the traditional Apocynoideae Ascle-piadaceae and Asclepiadoideae as nonstan-dard taxa The correspondence to the tradi-tional taxa will be indicated in the names byreplacing the sufx -eae with a neutral suf-x -ina (Kron 1997) We will also for practi-cal reasonsrecognize as a nonstandardtaxonthe informal group euapocynoids suggestedby Sennblad et al (1998)
The taxa of the traditional Apocynoideaeand Asclepiadaceae (uwboot D 86 swboot D99) clade will in many cases be more dif-cult to safely delimit to tribes becausemany of the groups in this clade have weakor no support (although preliminary resultsfrom an unpublished collaborative study bySennblad et al indicate additional supportfor the tribes discussed below) The tribaland subtribal classication of the traditionalApocynoideae has been shown to be prob-lematic (Leeuwenberg 1994a Endress andBruyns 2000) The circumscription of tribesfor these taxa might therefore in some casesbe preliminary
In all following taxa the basal part of theanther connective called the retinacle is ad-nate to the style head
DenitionmdashApocynoidina are the mostinclusive clade that includes the pri-mary types of Apocynum and Wrightia butnot Alyxieae Aspidospermeae Carisseae
404 SYSTEMATIC BIOLOGY VOL 51
Hunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Willughbeieae
Tribus Wrightieae G DonThe Stephanostema and Wrightia (both
Wrightieae) clade is strongly supported(uwboot and swboot D 100) and correspondsto the Wrightieae sensu Sennblad et al (1998)who also suggested an inclusion of Pleio-ceras Baill The tribe could be characterizedby a combination of synapomorphic and ple-siomorphic characters for example presenceofa chalazalandabsence of micropylar comaleft contorted aestivation and absence of airspaces in the anthers (1998)
DenitionmdashTribe Wrightieae is the mostinclusive clade in the Apocynoidina that in-cludes the primary types of Wrightia andStephanostema but not Malouetieae Nerieaeor Euapocynoidina
Tribus Nerieae (Benth) M Pichon
Adenium Nerium and Strophanthus (allWrightieae) constituted the Nerieae inSennblad et al (1998) however the cladereceived weak support Here Adenium andNerium form a clade that is absent in halfof the most-parsimonious trees in the com-plete rbcL analysis but is strongly supportedin the combined ndhF and rbcL analysis(uwboot D 98 swboot D 99) Strophan-thus groups with the Mascarenhasia Pachy-podium Kibatalia Funtumia and Holarrhenaclade which corresponds to the Malouetieaesensu Sennblad et al (1998) This position ofStrophanthus is however weakly supported(uwboot and swboot lt 50) and is not presentin the combined ndhF and rbcL tree Fur-thermore certain morphological characters(unfused slits in the corolla tube and anapical anther appendage) indicate a posi-tion closer to the Nerieae (Sennblad et al1998)
DenitionmdashTribe Nerieae is the most in-clusive clade in the Apocynoidina that in-cludes the primary types of Nerium and Ade-nium but not Malouetieae Wrightieae orEuapocynoidina
Tribus Malouetieae Mull-Arg
The association of Mascarenhasia Pachy-podium Kibatalia Funtumia and Holar-rhena (all Wrightieae) is moderately wellsupported in the complete rbcL analysis
(uwboot D 57 swboot D 74) whereasthe corresponding clade is well supportedin the combined rbcL and ndhF analysis(uwboot D 66 swboot D 100) On the ba-sis of identied synapomorphies (presenceof calcium oxalate crystals in the stomium ofthe anthers absence of interpetal vein ses-sile laments and absence of air spaces inthe anthers) Sennblad et al (1998) predictedan inclusion of Alaa Thouars Kibatalia andMalouetia ADC in the tribe The inclusion ofKibatalia is strongly supported in this studyFollowing these suggestions we will here as-sume the inclusion of the reference genusMalouetia in this tribe (The denition belowwill include Strophanthus on the present treebut see comments under Nerieae)
DenitionmdashTribe Malouetieae is the mostinclusive clade in the Apocynoidina that in-cluded the primary types of Malouetia andFuntumia but not Nerieae Wrightieae orEuapocynoidina
EUAPOCYNOIDINA
In all the remaining representatives ofthe Apocynoideae sensu Endress andBruyns (2000) that is Mandevilla MesechitesAganosma Apocynum TrachelospermumRhabdadenia Beaumontia Parsonsia andPrestonia the anthers are adnate to the stylehead both by the retinacle and by the thecaeThese taxa together with the taxa of the tra-ditional Asclepiadaceae form a moderatelysupported clade (uwboot lt 50 swboot D86) that corresponds to the informal taxoneuapocynoids sensu Sennblad et al (1998)
DenitionmdashEuapocynoidina are the mostinclusive clade that includes the primarytypes of Apocynum and Echites but notMalouetieae Wrightieae or Nerieae
Tribus Mesechiteae Miers
Mandevilla and Mesechites (both Mese-chiteae) form a moderately supported clade(uwboot lt 50 swboot D 87) equivalentto the Mesechiteae This tribe could be char-acterized by anthers with obtuse truncatedtails and a retinacle lacking hairs and also astrongly pentagonal style head with a stig-matic hollow
DenitionmdashTribe Mesechiteae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Mesechitesand Mandevilla but not Apocyneae EchiteaePeriploceae or Asclepiadacina
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 405
Tribus Apocyneae Bercht et J Presl
Aganosma Apocynum Beaumontia andTrachelospermum (Apocyneae) form a weaklysupported clade with Rhabdadenia (Echiteae)(uwboot and swboot lt 50) in the completerbcL analysis whereas the correspondingclade in the combined ndhF and rbcL analy-sis gains a little better support (uwboot lt 50swboot D 91) Likewise this clade (with theinclusion of Chonemorpha G Don) is weaklysupported in Potgieter and Albert (2001) Po-tential characterizing traits could be a stylehead without stigmatic hollow in combina-tion with thecae that are adnate to the stylehead (Sennblad et al 1998)
DenitionmdashTribe Apocyneae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Apocynumand Trachelospermum but not Echiteae Mese-chiteae Periploceae or Asclepiadacina
Tribus Echiteae Bartl
Parsonsia Peltastes and Prestonia (allEchiteae) form a moderately supported orunsupported clade with the Periplocoideaerepresentatives of the study (uwboot andswboot lt 50 and uwboot lt 50 swboot D89 in the complete rbcL and the combinedrbcL and ndhF analysis respectively) In thestudy of Sennblad et al (1998) Prestonia andParsonsia formed a weakly supported cladeand in the study of Potgieter and Albert(2001) inclusion of both Parsonsia and EchitesP Browne in the Echiteae is supportedParsonsia and Echiteae also share certaincharacters such as a horseshoe-shaped reti-nacle and a style head with stigmatic hollow(see Sennblad et al 1998)
DenitionmdashTribe Echiteae is the most in-clusive clade in the Euapocynoidina thatincludes the primary types of Echites andPrestonia but not Apocyneae MesechiteaePeriploceae or Asclepiadacina
Tribus Periploceae BartlThe representatives of the Periplo-
coideae sensu Endress and Bruyns (2000)mdashPentopetia Parquetina Petopentia PeriplocaTacazzea and Mondiamdashgroup together ina well-supported clade (uwboot D 67swboot D 93) The monophyly of the tra-ditional periplocoid taxa is also supportedin earlier studies that included a wider ordifferent sampling (Civeyrel et al 1998
Potgieter and Albert 2001) The taxa ofthe traditional Asclepiadaceae have ldquotrans-latorsrdquo structures related to pollinationspecialization In the Periploceae the pollenis deposited as tetrads in a spoon-liketranslator with an adhesive disc
DenitionmdashTribe Periploceae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Periplocaand Pentopetia but not Apocyneae EchiteaeMesechiteae or Asclepiadacina
ASCLEPIADACINA
The taxa of the Asclepiadoideae sensuLiede and Albers (1994) form a clade togetherwith the two Secamone species and Baissea(Apocyneae uwboot lt 50 swboot D 73)The exclusion of Periploceae (traditionallyassociated with this clade see Civeyrel et al1998) from this taxon is weakly supportedTherefore Periploceae is not included as areference taxon in the stem-based part ofthe denition of the Asclepiadacina thusleaving open the possibility for a later in-clusion of this taxon In the Asclepiadacina(except Baissea) the pollen is agglutinatedinto pollinia that are connected to a claspingtranslator
DenitionmdashAsclepiadacina are the mostinclusive monophyletic clade in the Euapoc-ynoidina that includes the primary types ofAsclepias and Secamoneae but not ApocyneaeEchiteae or Mesechiteae
Tribus Secamoneae G DonThe tribe Secamoneae sensu Bruyns
(uwboot D 89 swboot D 96) is mono-phyletic in the present analysisHowever be-cause representatives from only one of theconstituent genera are included monophylyis not well tested but the study of Civeyrelet al (1998) which included a larger numberof taxa including Pervillea Decne showedthat monophyly of the tribe was supportedThe traditional character for this tribe is fourpollinia per translator
DenitionmdashTribe Secamoneae is the mostinclusive clade in the Asclepiadacina that in-cludes the primary types of Secamone andPervillea but not Asclepioidina or Baissea
BaisseaThe position of Baissea corresponds with
that in the study of Sennblad et al (1998)
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 403
(uwboot D 83 swboot D 91) with Ane-chites which has been suggested by Fallen(1983) This association was based mainlyon the latrorsesublatrorse anthers andthe broad style head with large apicalappendages and stigmatic hollow The studyby Potgieter and Albert (2001) suggesteda further inclusion of Cerberiopsis Viell ExPancher amp Sebert and Skytanthus Meyen inthis tribe These taxa form a well-supportedclade with Plumeria and Allamanda whichcorresponds to the Plumerieae sensuEndress and Bruyns (2000) Characterssupporting this relationship are the presenceof infrastaminal as well as suprastaminalappendages (not present in Anechites andPlumeria) and winged seeds (not present inAnechites) (Endress et al 1996)
DenitionmdashTribe Plumerieae is themost inclusive clade that includes theprimary types of Plumeria and Allamandabut not Alyxieae Aspidospermeae Caris-seae Hunterieae Melodineae Tabernae-montaneae Vinceae Willughbeieae orApocynoidina
Tribus Carisseae Dumort
Acokanthera and Carissa (both Carisseae)are traditionally joined and are sometimeseven treated as one genus (eg Pichon1948a) In the present study they form astrongly supported clade (uwboot D 96swboot D 99) Among the characters sup-porting this tribe are syncarpous fruits with-out axile placentation and endocracks on theinner exine of the pollen (Endress et al 1996)
DenitionmdashTribe Carisseae is the most in-clusive clade that includes the primary typesof Carissa and Acokanthera but not AlyxieaeAspidospermeae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
Tribus Alyxieae G Don
Chilocarpus has been a genus with un-certain afnities In Endress et al (1996) itwas the sister taxon to the ApocynoideaeHere it is associated with Alyxia and Lep-inia in a clade corresponding to the Alyxieaesensu Endress and Bruyns (2000) Thisclade is strongly supported with succes-sive weighted bootstrap (swboot D 91) butnot with unit-weighted bootstrap support(uwboot D 58) such clades will henceforthbe termed moderatelywell supported clades
A further inclusion of Condylocarpon DesfLepiniopsis Valeton and Plectaneia Thouarswas indicated by Potgieter and Albert (2001)The constituent taxa are the only Apocy-naceae taxa outside the Apocynoidina thathave porate pollen
DenitionmdashTribe Alyxieae is the most in-clusive clade that includes the primary typesof Alyxia and Lepinia but not Aspidosper-meae Carisseae Hunterieae MelodineaePlumerieae Tabernaemontaneae VinceaeWillughbeieae or Apocynoidina
APOCYNOIDINA
On the present tree the traditionalsubfam-ily Apocynoideae includes all subfamilies(Periplocoideae Secamonoideae and Ascle-piadoideae) of the Asclepiadaceae sensuLiede and Albers (1994) All of these taxaare established names and selecting one asa subfamily at the expense of the othersmight be considered unfortunate from dif-ferent points of view We have here chosennot to recognize subfamilies instead we willdene the traditional Apocynoideae Ascle-piadaceae and Asclepiadoideae as nonstan-dard taxa The correspondence to the tradi-tional taxa will be indicated in the names byreplacing the sufx -eae with a neutral suf-x -ina (Kron 1997) We will also for practi-cal reasonsrecognize as a nonstandardtaxonthe informal group euapocynoids suggestedby Sennblad et al (1998)
The taxa of the traditional Apocynoideaeand Asclepiadaceae (uwboot D 86 swboot D99) clade will in many cases be more dif-cult to safely delimit to tribes becausemany of the groups in this clade have weakor no support (although preliminary resultsfrom an unpublished collaborative study bySennblad et al indicate additional supportfor the tribes discussed below) The tribaland subtribal classication of the traditionalApocynoideae has been shown to be prob-lematic (Leeuwenberg 1994a Endress andBruyns 2000) The circumscription of tribesfor these taxa might therefore in some casesbe preliminary
In all following taxa the basal part of theanther connective called the retinacle is ad-nate to the style head
DenitionmdashApocynoidina are the mostinclusive clade that includes the pri-mary types of Apocynum and Wrightia butnot Alyxieae Aspidospermeae Carisseae
404 SYSTEMATIC BIOLOGY VOL 51
Hunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Willughbeieae
Tribus Wrightieae G DonThe Stephanostema and Wrightia (both
Wrightieae) clade is strongly supported(uwboot and swboot D 100) and correspondsto the Wrightieae sensu Sennblad et al (1998)who also suggested an inclusion of Pleio-ceras Baill The tribe could be characterizedby a combination of synapomorphic and ple-siomorphic characters for example presenceofa chalazalandabsence of micropylar comaleft contorted aestivation and absence of airspaces in the anthers (1998)
DenitionmdashTribe Wrightieae is the mostinclusive clade in the Apocynoidina that in-cludes the primary types of Wrightia andStephanostema but not Malouetieae Nerieaeor Euapocynoidina
Tribus Nerieae (Benth) M Pichon
Adenium Nerium and Strophanthus (allWrightieae) constituted the Nerieae inSennblad et al (1998) however the cladereceived weak support Here Adenium andNerium form a clade that is absent in halfof the most-parsimonious trees in the com-plete rbcL analysis but is strongly supportedin the combined ndhF and rbcL analysis(uwboot D 98 swboot D 99) Strophan-thus groups with the Mascarenhasia Pachy-podium Kibatalia Funtumia and Holarrhenaclade which corresponds to the Malouetieaesensu Sennblad et al (1998) This position ofStrophanthus is however weakly supported(uwboot and swboot lt 50) and is not presentin the combined ndhF and rbcL tree Fur-thermore certain morphological characters(unfused slits in the corolla tube and anapical anther appendage) indicate a posi-tion closer to the Nerieae (Sennblad et al1998)
DenitionmdashTribe Nerieae is the most in-clusive clade in the Apocynoidina that in-cludes the primary types of Nerium and Ade-nium but not Malouetieae Wrightieae orEuapocynoidina
Tribus Malouetieae Mull-Arg
The association of Mascarenhasia Pachy-podium Kibatalia Funtumia and Holar-rhena (all Wrightieae) is moderately wellsupported in the complete rbcL analysis
(uwboot D 57 swboot D 74) whereasthe corresponding clade is well supportedin the combined rbcL and ndhF analysis(uwboot D 66 swboot D 100) On the ba-sis of identied synapomorphies (presenceof calcium oxalate crystals in the stomium ofthe anthers absence of interpetal vein ses-sile laments and absence of air spaces inthe anthers) Sennblad et al (1998) predictedan inclusion of Alaa Thouars Kibatalia andMalouetia ADC in the tribe The inclusion ofKibatalia is strongly supported in this studyFollowing these suggestions we will here as-sume the inclusion of the reference genusMalouetia in this tribe (The denition belowwill include Strophanthus on the present treebut see comments under Nerieae)
DenitionmdashTribe Malouetieae is the mostinclusive clade in the Apocynoidina that in-cluded the primary types of Malouetia andFuntumia but not Nerieae Wrightieae orEuapocynoidina
EUAPOCYNOIDINA
In all the remaining representatives ofthe Apocynoideae sensu Endress andBruyns (2000) that is Mandevilla MesechitesAganosma Apocynum TrachelospermumRhabdadenia Beaumontia Parsonsia andPrestonia the anthers are adnate to the stylehead both by the retinacle and by the thecaeThese taxa together with the taxa of the tra-ditional Asclepiadaceae form a moderatelysupported clade (uwboot lt 50 swboot D86) that corresponds to the informal taxoneuapocynoids sensu Sennblad et al (1998)
DenitionmdashEuapocynoidina are the mostinclusive clade that includes the primarytypes of Apocynum and Echites but notMalouetieae Wrightieae or Nerieae
Tribus Mesechiteae Miers
Mandevilla and Mesechites (both Mese-chiteae) form a moderately supported clade(uwboot lt 50 swboot D 87) equivalentto the Mesechiteae This tribe could be char-acterized by anthers with obtuse truncatedtails and a retinacle lacking hairs and also astrongly pentagonal style head with a stig-matic hollow
DenitionmdashTribe Mesechiteae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Mesechitesand Mandevilla but not Apocyneae EchiteaePeriploceae or Asclepiadacina
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 405
Tribus Apocyneae Bercht et J Presl
Aganosma Apocynum Beaumontia andTrachelospermum (Apocyneae) form a weaklysupported clade with Rhabdadenia (Echiteae)(uwboot and swboot lt 50) in the completerbcL analysis whereas the correspondingclade in the combined ndhF and rbcL analy-sis gains a little better support (uwboot lt 50swboot D 91) Likewise this clade (with theinclusion of Chonemorpha G Don) is weaklysupported in Potgieter and Albert (2001) Po-tential characterizing traits could be a stylehead without stigmatic hollow in combina-tion with thecae that are adnate to the stylehead (Sennblad et al 1998)
DenitionmdashTribe Apocyneae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Apocynumand Trachelospermum but not Echiteae Mese-chiteae Periploceae or Asclepiadacina
Tribus Echiteae Bartl
Parsonsia Peltastes and Prestonia (allEchiteae) form a moderately supported orunsupported clade with the Periplocoideaerepresentatives of the study (uwboot andswboot lt 50 and uwboot lt 50 swboot D89 in the complete rbcL and the combinedrbcL and ndhF analysis respectively) In thestudy of Sennblad et al (1998) Prestonia andParsonsia formed a weakly supported cladeand in the study of Potgieter and Albert(2001) inclusion of both Parsonsia and EchitesP Browne in the Echiteae is supportedParsonsia and Echiteae also share certaincharacters such as a horseshoe-shaped reti-nacle and a style head with stigmatic hollow(see Sennblad et al 1998)
DenitionmdashTribe Echiteae is the most in-clusive clade in the Euapocynoidina thatincludes the primary types of Echites andPrestonia but not Apocyneae MesechiteaePeriploceae or Asclepiadacina
Tribus Periploceae BartlThe representatives of the Periplo-
coideae sensu Endress and Bruyns (2000)mdashPentopetia Parquetina Petopentia PeriplocaTacazzea and Mondiamdashgroup together ina well-supported clade (uwboot D 67swboot D 93) The monophyly of the tra-ditional periplocoid taxa is also supportedin earlier studies that included a wider ordifferent sampling (Civeyrel et al 1998
Potgieter and Albert 2001) The taxa ofthe traditional Asclepiadaceae have ldquotrans-latorsrdquo structures related to pollinationspecialization In the Periploceae the pollenis deposited as tetrads in a spoon-liketranslator with an adhesive disc
DenitionmdashTribe Periploceae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Periplocaand Pentopetia but not Apocyneae EchiteaeMesechiteae or Asclepiadacina
ASCLEPIADACINA
The taxa of the Asclepiadoideae sensuLiede and Albers (1994) form a clade togetherwith the two Secamone species and Baissea(Apocyneae uwboot lt 50 swboot D 73)The exclusion of Periploceae (traditionallyassociated with this clade see Civeyrel et al1998) from this taxon is weakly supportedTherefore Periploceae is not included as areference taxon in the stem-based part ofthe denition of the Asclepiadacina thusleaving open the possibility for a later in-clusion of this taxon In the Asclepiadacina(except Baissea) the pollen is agglutinatedinto pollinia that are connected to a claspingtranslator
DenitionmdashAsclepiadacina are the mostinclusive monophyletic clade in the Euapoc-ynoidina that includes the primary types ofAsclepias and Secamoneae but not ApocyneaeEchiteae or Mesechiteae
Tribus Secamoneae G DonThe tribe Secamoneae sensu Bruyns
(uwboot D 89 swboot D 96) is mono-phyletic in the present analysisHowever be-cause representatives from only one of theconstituent genera are included monophylyis not well tested but the study of Civeyrelet al (1998) which included a larger numberof taxa including Pervillea Decne showedthat monophyly of the tribe was supportedThe traditional character for this tribe is fourpollinia per translator
DenitionmdashTribe Secamoneae is the mostinclusive clade in the Asclepiadacina that in-cludes the primary types of Secamone andPervillea but not Asclepioidina or Baissea
BaisseaThe position of Baissea corresponds with
that in the study of Sennblad et al (1998)
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
404 SYSTEMATIC BIOLOGY VOL 51
Hunterieae Melodineae Plumerieae Taber-naemontaneae Vinceae or Willughbeieae
Tribus Wrightieae G DonThe Stephanostema and Wrightia (both
Wrightieae) clade is strongly supported(uwboot and swboot D 100) and correspondsto the Wrightieae sensu Sennblad et al (1998)who also suggested an inclusion of Pleio-ceras Baill The tribe could be characterizedby a combination of synapomorphic and ple-siomorphic characters for example presenceofa chalazalandabsence of micropylar comaleft contorted aestivation and absence of airspaces in the anthers (1998)
DenitionmdashTribe Wrightieae is the mostinclusive clade in the Apocynoidina that in-cludes the primary types of Wrightia andStephanostema but not Malouetieae Nerieaeor Euapocynoidina
Tribus Nerieae (Benth) M Pichon
Adenium Nerium and Strophanthus (allWrightieae) constituted the Nerieae inSennblad et al (1998) however the cladereceived weak support Here Adenium andNerium form a clade that is absent in halfof the most-parsimonious trees in the com-plete rbcL analysis but is strongly supportedin the combined ndhF and rbcL analysis(uwboot D 98 swboot D 99) Strophan-thus groups with the Mascarenhasia Pachy-podium Kibatalia Funtumia and Holarrhenaclade which corresponds to the Malouetieaesensu Sennblad et al (1998) This position ofStrophanthus is however weakly supported(uwboot and swboot lt 50) and is not presentin the combined ndhF and rbcL tree Fur-thermore certain morphological characters(unfused slits in the corolla tube and anapical anther appendage) indicate a posi-tion closer to the Nerieae (Sennblad et al1998)
DenitionmdashTribe Nerieae is the most in-clusive clade in the Apocynoidina that in-cludes the primary types of Nerium and Ade-nium but not Malouetieae Wrightieae orEuapocynoidina
Tribus Malouetieae Mull-Arg
The association of Mascarenhasia Pachy-podium Kibatalia Funtumia and Holar-rhena (all Wrightieae) is moderately wellsupported in the complete rbcL analysis
(uwboot D 57 swboot D 74) whereasthe corresponding clade is well supportedin the combined rbcL and ndhF analysis(uwboot D 66 swboot D 100) On the ba-sis of identied synapomorphies (presenceof calcium oxalate crystals in the stomium ofthe anthers absence of interpetal vein ses-sile laments and absence of air spaces inthe anthers) Sennblad et al (1998) predictedan inclusion of Alaa Thouars Kibatalia andMalouetia ADC in the tribe The inclusion ofKibatalia is strongly supported in this studyFollowing these suggestions we will here as-sume the inclusion of the reference genusMalouetia in this tribe (The denition belowwill include Strophanthus on the present treebut see comments under Nerieae)
DenitionmdashTribe Malouetieae is the mostinclusive clade in the Apocynoidina that in-cluded the primary types of Malouetia andFuntumia but not Nerieae Wrightieae orEuapocynoidina
EUAPOCYNOIDINA
In all the remaining representatives ofthe Apocynoideae sensu Endress andBruyns (2000) that is Mandevilla MesechitesAganosma Apocynum TrachelospermumRhabdadenia Beaumontia Parsonsia andPrestonia the anthers are adnate to the stylehead both by the retinacle and by the thecaeThese taxa together with the taxa of the tra-ditional Asclepiadaceae form a moderatelysupported clade (uwboot lt 50 swboot D86) that corresponds to the informal taxoneuapocynoids sensu Sennblad et al (1998)
DenitionmdashEuapocynoidina are the mostinclusive clade that includes the primarytypes of Apocynum and Echites but notMalouetieae Wrightieae or Nerieae
Tribus Mesechiteae Miers
Mandevilla and Mesechites (both Mese-chiteae) form a moderately supported clade(uwboot lt 50 swboot D 87) equivalentto the Mesechiteae This tribe could be char-acterized by anthers with obtuse truncatedtails and a retinacle lacking hairs and also astrongly pentagonal style head with a stig-matic hollow
DenitionmdashTribe Mesechiteae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Mesechitesand Mandevilla but not Apocyneae EchiteaePeriploceae or Asclepiadacina
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 405
Tribus Apocyneae Bercht et J Presl
Aganosma Apocynum Beaumontia andTrachelospermum (Apocyneae) form a weaklysupported clade with Rhabdadenia (Echiteae)(uwboot and swboot lt 50) in the completerbcL analysis whereas the correspondingclade in the combined ndhF and rbcL analy-sis gains a little better support (uwboot lt 50swboot D 91) Likewise this clade (with theinclusion of Chonemorpha G Don) is weaklysupported in Potgieter and Albert (2001) Po-tential characterizing traits could be a stylehead without stigmatic hollow in combina-tion with thecae that are adnate to the stylehead (Sennblad et al 1998)
DenitionmdashTribe Apocyneae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Apocynumand Trachelospermum but not Echiteae Mese-chiteae Periploceae or Asclepiadacina
Tribus Echiteae Bartl
Parsonsia Peltastes and Prestonia (allEchiteae) form a moderately supported orunsupported clade with the Periplocoideaerepresentatives of the study (uwboot andswboot lt 50 and uwboot lt 50 swboot D89 in the complete rbcL and the combinedrbcL and ndhF analysis respectively) In thestudy of Sennblad et al (1998) Prestonia andParsonsia formed a weakly supported cladeand in the study of Potgieter and Albert(2001) inclusion of both Parsonsia and EchitesP Browne in the Echiteae is supportedParsonsia and Echiteae also share certaincharacters such as a horseshoe-shaped reti-nacle and a style head with stigmatic hollow(see Sennblad et al 1998)
DenitionmdashTribe Echiteae is the most in-clusive clade in the Euapocynoidina thatincludes the primary types of Echites andPrestonia but not Apocyneae MesechiteaePeriploceae or Asclepiadacina
Tribus Periploceae BartlThe representatives of the Periplo-
coideae sensu Endress and Bruyns (2000)mdashPentopetia Parquetina Petopentia PeriplocaTacazzea and Mondiamdashgroup together ina well-supported clade (uwboot D 67swboot D 93) The monophyly of the tra-ditional periplocoid taxa is also supportedin earlier studies that included a wider ordifferent sampling (Civeyrel et al 1998
Potgieter and Albert 2001) The taxa ofthe traditional Asclepiadaceae have ldquotrans-latorsrdquo structures related to pollinationspecialization In the Periploceae the pollenis deposited as tetrads in a spoon-liketranslator with an adhesive disc
DenitionmdashTribe Periploceae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Periplocaand Pentopetia but not Apocyneae EchiteaeMesechiteae or Asclepiadacina
ASCLEPIADACINA
The taxa of the Asclepiadoideae sensuLiede and Albers (1994) form a clade togetherwith the two Secamone species and Baissea(Apocyneae uwboot lt 50 swboot D 73)The exclusion of Periploceae (traditionallyassociated with this clade see Civeyrel et al1998) from this taxon is weakly supportedTherefore Periploceae is not included as areference taxon in the stem-based part ofthe denition of the Asclepiadacina thusleaving open the possibility for a later in-clusion of this taxon In the Asclepiadacina(except Baissea) the pollen is agglutinatedinto pollinia that are connected to a claspingtranslator
DenitionmdashAsclepiadacina are the mostinclusive monophyletic clade in the Euapoc-ynoidina that includes the primary types ofAsclepias and Secamoneae but not ApocyneaeEchiteae or Mesechiteae
Tribus Secamoneae G DonThe tribe Secamoneae sensu Bruyns
(uwboot D 89 swboot D 96) is mono-phyletic in the present analysisHowever be-cause representatives from only one of theconstituent genera are included monophylyis not well tested but the study of Civeyrelet al (1998) which included a larger numberof taxa including Pervillea Decne showedthat monophyly of the tribe was supportedThe traditional character for this tribe is fourpollinia per translator
DenitionmdashTribe Secamoneae is the mostinclusive clade in the Asclepiadacina that in-cludes the primary types of Secamone andPervillea but not Asclepioidina or Baissea
BaisseaThe position of Baissea corresponds with
that in the study of Sennblad et al (1998)
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 405
Tribus Apocyneae Bercht et J Presl
Aganosma Apocynum Beaumontia andTrachelospermum (Apocyneae) form a weaklysupported clade with Rhabdadenia (Echiteae)(uwboot and swboot lt 50) in the completerbcL analysis whereas the correspondingclade in the combined ndhF and rbcL analy-sis gains a little better support (uwboot lt 50swboot D 91) Likewise this clade (with theinclusion of Chonemorpha G Don) is weaklysupported in Potgieter and Albert (2001) Po-tential characterizing traits could be a stylehead without stigmatic hollow in combina-tion with thecae that are adnate to the stylehead (Sennblad et al 1998)
DenitionmdashTribe Apocyneae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Apocynumand Trachelospermum but not Echiteae Mese-chiteae Periploceae or Asclepiadacina
Tribus Echiteae Bartl
Parsonsia Peltastes and Prestonia (allEchiteae) form a moderately supported orunsupported clade with the Periplocoideaerepresentatives of the study (uwboot andswboot lt 50 and uwboot lt 50 swboot D89 in the complete rbcL and the combinedrbcL and ndhF analysis respectively) In thestudy of Sennblad et al (1998) Prestonia andParsonsia formed a weakly supported cladeand in the study of Potgieter and Albert(2001) inclusion of both Parsonsia and EchitesP Browne in the Echiteae is supportedParsonsia and Echiteae also share certaincharacters such as a horseshoe-shaped reti-nacle and a style head with stigmatic hollow(see Sennblad et al 1998)
DenitionmdashTribe Echiteae is the most in-clusive clade in the Euapocynoidina thatincludes the primary types of Echites andPrestonia but not Apocyneae MesechiteaePeriploceae or Asclepiadacina
Tribus Periploceae BartlThe representatives of the Periplo-
coideae sensu Endress and Bruyns (2000)mdashPentopetia Parquetina Petopentia PeriplocaTacazzea and Mondiamdashgroup together ina well-supported clade (uwboot D 67swboot D 93) The monophyly of the tra-ditional periplocoid taxa is also supportedin earlier studies that included a wider ordifferent sampling (Civeyrel et al 1998
Potgieter and Albert 2001) The taxa ofthe traditional Asclepiadaceae have ldquotrans-latorsrdquo structures related to pollinationspecialization In the Periploceae the pollenis deposited as tetrads in a spoon-liketranslator with an adhesive disc
DenitionmdashTribe Periploceae is the mostinclusive clade in the Euapocynoidina thatincludes the primary types of Periplocaand Pentopetia but not Apocyneae EchiteaeMesechiteae or Asclepiadacina
ASCLEPIADACINA
The taxa of the Asclepiadoideae sensuLiede and Albers (1994) form a clade togetherwith the two Secamone species and Baissea(Apocyneae uwboot lt 50 swboot D 73)The exclusion of Periploceae (traditionallyassociated with this clade see Civeyrel et al1998) from this taxon is weakly supportedTherefore Periploceae is not included as areference taxon in the stem-based part ofthe denition of the Asclepiadacina thusleaving open the possibility for a later in-clusion of this taxon In the Asclepiadacina(except Baissea) the pollen is agglutinatedinto pollinia that are connected to a claspingtranslator
DenitionmdashAsclepiadacina are the mostinclusive monophyletic clade in the Euapoc-ynoidina that includes the primary types ofAsclepias and Secamoneae but not ApocyneaeEchiteae or Mesechiteae
Tribus Secamoneae G DonThe tribe Secamoneae sensu Bruyns
(uwboot D 89 swboot D 96) is mono-phyletic in the present analysisHowever be-cause representatives from only one of theconstituent genera are included monophylyis not well tested but the study of Civeyrelet al (1998) which included a larger numberof taxa including Pervillea Decne showedthat monophyly of the tribe was supportedThe traditional character for this tribe is fourpollinia per translator
DenitionmdashTribe Secamoneae is the mostinclusive clade in the Asclepiadacina that in-cludes the primary types of Secamone andPervillea but not Asclepioidina or Baissea
BaisseaThe position of Baissea corresponds with
that in the study of Sennblad et al (1998)
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
406 SYSTEMATIC BIOLOGY VOL 51
and is also supported by the combined ndhFand rbcL analysis (uwboot D 55 swboot D97) The unexpected position of the Baissea(rst suggested in Sennblad 1997 see alsoSennblad and Bremer 2000) nested withinthe traditional Asclepiadaceae prompted usto verify the rbcL sequence by resequencingthe results turned out to be identical Further-more in the study of Potgieter and Albert(2001) based on trnLndashF data an indepen-dent extraction (but sampled from the samespecimen) placed Baissea in the same posi-tion No obvious characters connect Baisseato the traditional Asclepiadoideae and Seca-monoideae but suggestive qualities may bethe tendencies of the stamens to have dor-sal staminal appendages of the style to haveelongated apices (compare certain Secamone)and of the bulbs or ridges to be below orat thelament insertions of certain species (sug-gesting a basal tube) Translators of a verysimple type are also found in Baissea Becauseof its uncertain position there are no obvioussister groups to Baissea Also to avoid mono-typic taxa we do not assign Baissea to a tribe(the principle of exhaustive subsidiary taxais not followed)
ASCLEPIADOIDINA
The genera of the traditional Asclepi-oideae form a moderately supported clade(uwboot D 58 swboot D 84) which willhere be treated as a nonstandard taxonThe traditional characters for the Asclepi-adoideae are clasping translators with twopollinia per translator
DenitionmdashAsclepiadoidina are the mostinclusive clade that includes the primarytypes of Asclepias and Fockea but not Seca-moneae
Tribus Fockeeae Kunze et al
Fockea (Marsdenieae) has been placed withCibirhiza Endl in the tribe Fockeeae In thisanalysis Fockea is the sister group to the restof the Asclepiadoideae (uwboot D 58 swbootD 84) This position is similar to the posi-tion of Fockea in the studies by Civeyrel et al(1998) and Potgieter and Albert (2001) Inclu-sion of the tribe Fockeeae (Kunze et al 1994)in the Marsdenieae (Endress and Bruyns2000) is thus not supported
DenitionmdashTribe Fockeeae is the most in-clusive clade in the Asclepiadoidina that in-
cludes the primary types of Fockea and Fockeabut not Ceropegieae Marsdenieae or Ascle-piadeae
Tribus Ceropegieae DecneThe tribe Ceropegieae sensu Endress and
Bruyns (2000) is represented by Stapelia andCeropegia and is very strongly supported(uwboot and swboot D 100) This traditionalgroup is strongly supported by morphol-ogy and the inclusion of further traditionalstapeliad taxa is indicated in Potgieter andAlbers (2001)
DenitionmdashTribe Ceropegieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Stapelia andCeropegia but not Asclepiadeae Fockeeae orMarsdenieae
Tribus Marsdenieae Benth
Stephanotis Micholitzia and Hoya repre-sent the tribe Marsdenieae sensu Endressand Bruyns (2000) and form a very wellsupported clade (uwboot D 91 swboot D95) The position of Fockea as sister groupto the rest of the Asclepiadoidina makesthe Marsdenieae sensu Endress and Bruyns(2000) nonmonophyletic Stephanotis is mostlikely congeneric with or sister to MarsdeniaRBr which we therefore will assume to beincluded in the tribe The study of Potgieterand Albers (2001) supports further inclusionof Dischidia RBr Dregea E Mey and TelosmaCoville
DenitionmdashTribe Marsdenieae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Marsdenia andHoya but not Asclepiadeae Fockeeae orCeropegieae
Tribus Asclepiadeae (R BR) Duby
Schizostephanus Asclepias CalotropisVince-toxicum Tylophora Orthosia Araujia TweediaFischeria and Matelea form a strongly sup-ported clade (uwboot D 89 swboot D 96) corre-sponding to Asclepiadeae sensu Endress andBruyns (2000) The inclusion of Tylophoraandthe traditional Gonolobeae (here representedby Fischeria and Matelea) in the Asclepiadeae(eg Liede 1996Sennblad and Bremer 19962000 Swarupanandan et al 1996 Civeyrelet al 1998 Endress and Bruyns 2000Potgieter and Albert 2001) is thus sup-ported The studies of Civeyrel et al (1998)
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 407
and Potgieter and Albert (2001) also indicatesupport for inclusion of other traditional As-clepiadeae taxa
DenitionmdashTribe Asclepiadeae is the mostinclusive clade in the Asclepiadoidina thatincludes the primary types of Asclepias andMatelea but not Fockeeae Marsdenieae orCeropegieae
ACKNOWLEDGMENTS
We thank K Bremer M Fishbein R Olmsteadand two anonymous reviewers for valuable commentsand discussion of the manuscript N Heidari is ac-knowledged for much help with the sequencing MBacklund M Chase L Civeyrel J Clark M Endress ALeeuwenberg U Meve B Oxelman B PetterssonE Robbrecht and S Zona kindly shared their mate-rial M Endress A Leeuwenberg and U Meve alsocontrolled the determination of many specimens KAndreasen A Backlund and J Kukka are thankedfor valuable help and company during eld tripsThe Uppsala Botanical Garden the Uppsala BotanicalMuseum the Fairchild Tropical Garden the MunsterBotanical Garden the Aarhus Botanical Garden andthe Stockholm University were most helpful in provid-ing material This study was supported by the SwedishResearch Council to B B and a travel grant from J AWahlbergrsquos memorial fund to B S
REFERENCES
ALLORGE L 1985 Monographie des ApocynaceesmdashTabernamontanoidees Americaines Mem Mus NatHist Nat Ser B Bot 30 216 pp
BACKLUND M B OXELMAN AND B BREMER 2000Phy-logenetic relationships within the Gentianales basedon ndhF and rbcL sequences with particular referenceto the Loganiaceae Am J Bot 871029ndash1043
BAUM D A W S ALVERSON AND R NYFFELER 1998Adurian by any other name Taxonomy and nomencla-ture of the core Malvales Harv Pap Bot 3313ndash330
BOITEAU P L ALLORGE AND C SASTRE 1978 Mor-phologie orale des Apocynaceae II Caracteresdistinctifs entre Ambelanieae (Plumerioideae) etMacoubeae (Tabernaemontanoideae) Adansonia ns18267ndash277
BREMER K M W CHASE P F STEVENS A AANDERBERG A BACKLUND B BREMER B G BRIGGS P K ENDRESS M F FAY P GOLDBLATT M H GGUSTAFSSON S B HOOT W S JUDD M KALLERSJOE A KELLOGG K A KRON D H LES C MMORTON D L NICKRENT R G OLMSTEAD R APRICE C J QUINN J E RODMAN P J RUDALLV SAVOLAINEN D E SOLTIS P S SOLTIS K JSYTSMA AND M THULIN 1998 An ordinal classi-cation for the families of owering plants Ann MBot Gard 85531ndash553
BROWN R 1810 On the Asclepiadeae Preprinted from1811 Mem Wernerian Nat Hist Soc 112ndash78
BRYANT H N 1996 Explicitness stability and uni-versality in the phylogenetic denition and usage oftaxon names A case study of the phylogenetic taxon-omy of the Carnivora (mammalia) Syst Biol 45174ndash189
BRYANT H N 1997 Cladistic information in phyloge-netic denitions and designated phylogenetic con-texts for the use of taxon names Biol J Linn Soc62495ndash503
CANTINO P D AND K DE QUEIROZ 2000 PhyloCodeA phylogenetic code of biological nomenclature Draftavailable on httpwwwohioueduphylocode
CANTINO P D R G OLMSTEAD AND S J WAGSTAFF1997 A comparison of phylogenetic nomenclaturewith the current system A botanical case study SystBiol 46313ndash331
CIVEYREL L A LE THOMAS K FERGUSON AND M WCHASE 1998 Critical reexamination of palynologicalcharacters used to delimit Asclepiadaceae in compari-son to the molecular phylogeny obtained from plastidmatK sequences Mol Phylogenet Evol 9 517ndash527
DE QUEIROZ K 1996 A phylogenetic approach to bi-ological nomenclature as an alternative to the Lin-nean system in current use In Biological nomencla-ture in the 21st century (J L Reveal) httpwwwinformumdeduPBIOnomcldequhtml Univ ofMaryland
DE QUEIROZ K 1997a Misunderstandings about thephylogenetic approach to biological nomenclature Areply to Liden and Oxelman Zool Scr 2667ndash70
DE QUEIROZ K 1997b The Linnaean hierarchy and theevolutionization of taxonomy with emphasis on theproblem of nomenclature Aliso 15115ndash144
DE QUEIROZ K AND J GAUTHIER 1992 Phylogenetictaxonomy Annu Rev Ecol Syst 23449ndash480
DE QUEIROZ K AND J GAUTHIER 1994 Toward a phy-logenetic system of biological nomenclature TrendsEcol Evol 927ndash31
DEMETER K 1922 Vergleichende AsclepiadeenstudienFlora (Jena) 115130ndash176
DOYLE J J AND J L DOYLE 1987 A rapid DNA isola-tion procedure for small quantities of fresh leaf tissuePhytochem bull 1911ndash15
ENDRESS M E AND P V BRUYNS 2000 A revised clas-sication of the Apocynaceae sl Bot Rev 661ndash56
ENDRESS M E B SENNBLAD S NILSSON L CIVEYRELM W CHASE S HUYSMANS E GRAFSTROM ANDB BREMER 1996 A phylogenetic analysis of Apoc-ynaceae sstr and some related taxa in Gentianales Amultidisciplinary approach Op Bot Belg 759ndash102
ERIKSSON T M J DONOGHUE AND M S HIBBS 1998 Phylogenetic analysis of Potentilla using DNAsequences of nuclear ribosomal internal transcribedspacers (ITS) and implications for the classication ofRosoideae (Rosaceae) Plant Syst Evol 211155ndash179
FALLEN M E 1983 A systematic revision of Anechites(Apocynaceae) Brittonia 35222ndash231
FALLEN M E 1986 Floral structure in the Apocy-naceae Morphological functional and evolutionaryaspects Bot Jahrb Syst Panzengesch Panzen-geogr 106245ndash286
FARRIS J S 1969 A successive approximations ap-proach to character weighting Syst Zool 18374ndash385
FARRIS J S 1989 The retention index and the rescaledconsistency index Cladistics 5417ndash419
FARRIS J S V A ALBERT M KALLERSJO D LIPS COMBAND A G KLUGE 1996 Parsimony jackkning out-performs neighbor-joining Cladistics 1299ndash124
FELSENSTEIN J 1985 Condence limits on phylogeniesAn approach using the bootstrap Evolution 39783ndash791
GOLENBERG E M M T CLEGG M L DURBINJ DOEBLEY AND D P MA 1993 Evolution of a
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
408 SYSTEMATIC BIOLOGY VOL 51
noncoding region of the chloroplast genome MolPhylogenet Evol 252ndash64
GREUTER W F R BARRIE H M BURDET W GCHALONER V DEMOULIN D L HAWKSWORTHP M JOslashRGENSEN D H NICHOLSON P C SILVAP TREHANE AND J MCNEILL 1994International Codeof Botanical Nomenclature (Tokyo Code) RegnumVegetabile 131
HARSHMAN J 1994 The effect of irrelevant characterson bootstrap values Syst Biol 43419ndash424
JUDD W S R W SANDERS AND M J DONOGHUE 1994Angiosperm family pairs Preliminary phylogeneticanalyses Harv Pap Bot 51ndash51
KALTENBOECK B J W SPATAFORA X ZHANG K GKOUSOULAS M BLACKWELL AND J STORZ 1992Efcient production of single-stranded DNA as longas 2 kb for sequencing of PCR-amplied DNABiotechniques 12164ndash171
KIM K-J AND R K JANSEN 1995 ndhF sequence evo-lution and the major clades in the sunower familyProc Nat Acad Sci USA 9210379ndash10383
KRON K A 1997 Exploring alternative systems of clas-sication Aliso 15105ndash112
KUNZE H U MEVE AND L LIEDE 1994 Cibirhiza alber-siana a new species of Asclepiadaceae and establish-ment of the tribe Fockeeae Taxon 43367ndash376
LEE M S Y 1999 Reference taxa and phylogeneticnomenclature Taxon 4831ndash34
LEEUWENBERG A J M 1991 Tabernaemontana 1 TheOld World species Royal Botanical Garden KewEngland
LEEUWENBERG A J M 1994a Taxa of the Apocynaceaeabove the genus level Series of revisions of Apocy-naceae XXXVIII Wageningen Agr Univ Pap 9445ndash60
LEEUWENBERG A J M 1994b A revision of Tabernae-montana 2 The New World species and StemmadeniaRoyal Botanic Gardens Kew England
LIDEN M AND B OXELMAN 1996 Point of viewmdashDowe need ldquophylogenetic taxonomyrdquo Zool Scr 25183ndash185
LIDEN M B OXELMAN A BACKLUND L ANDERSSONB BREMER R ERIKSSON R MOBERG I NORDALK PERSSON M THULIN AND B ZIMMER 1997Char-lie is our darling Taxon 46735ndash738
LIEDE S 1996 CynanchumndashRhodostegiellandashVincetoxicumndashTylophora (Asclepiadaceae) New considerations on anold problem Taxon 45193ndash211
LIEDE S AND F ALBERS 1994Tribal disposition of gen-era in the Asclepiadaceae Taxon 43201ndash231
MISHLER B D 1999 Getting rid of species Pages 307ndash315 In Species New interdisciplinary essays (R AWilson ed) MIT Press Cambridge Massachusetts
MOORE G 1998 A comparison of traditional and phy-logenetic nomenclature Taxon 47561ndash579
OLMSTEAD R G B BREMER K M SCOTT AND J DPALMER 1993 A parsimony analysis of the Asteridaesensu lato based on rbcL sequences Ann Mo BotGard 80700ndash722
OLMSTEAD R G C W DEPAMPHILIS A D WOLFEN D YOUNG W J ELISONS AND P A REEVES 2001Disintegration of the Scrophulariaceae Am J Bot88348ndash361
OLMSTEAD R G H J MICHAELS K M SCOTT AND JD PALMER 1992 Monophyly of the Asteridae andidentication of their major lineages inferred fromDNA sequences of rbcL Ann Mo Bot Gard 79249ndash265
OLMSTEAD R G AND J A SWEERE 1994 Combin-ing data in phylogenetic systematics An empiri-cal approach using three molecular data sets in theSolanaceae Syst Biol 43467ndash481
OMINO E 1996 A contribution to the leaf anatomy andtaxonomy of Apocynaceae in Africa The leaf anatomyof Apocynaceae in East Africa A monograph of Pleio-carpinae (Series of revisions of Apocynaceae XLI)Wageningen Agr Univ Pap 961ndash178
OXELMAN B M BACKLUND AND B BREMER 1999 Re-lationships of the Buddlejaceae s1 investigated usingparsimony jackknife and branch support analysis ofchloroplast ndhF and rbcL sequence data Syst Bot24164ndash182
PERSOON J G M F J H VAN DILST R P KUIJPERS A J M LEEUWENBERG and G J A VONK 1992 TheAfrican species of Landolphia PBeauv (Series of revi-sions of Apocynaceae XXXIV) WageningenAgr UnivPap 921ndash232
PICHON M 1948a Classication des Apocynacees ICarissees et Ambelaniees Mem Mus Nat Hist Nat24111ndash181
PICHON M 1948b Classication des ApocynaceesIX Rauvolees Alstoniees Allamandees etTabernaemontanoidees Mem Mus Nat HistNat 27153ndash252
PICHON M 1948c Classication des Apocynacees VCerberoidees Notul Syst 13212ndash229
PICHON M 1950 Classication des Apocynacees XXVEchito otilde dees Mem Mus Natl Hist Nat Ser B Bot11ndash143
POTGIETER K AND V A ALBERT 2001Phylogenetic re-lationships within Apocynaceae sl based on trnL in-tron and trnL-F spacer sequences and propagule char-acters Ann Mo Bot Gard 88523ndash549
REVEAL J L 1996 Solutions for biological nomencla-ture In Biological nomenclature in the 21st century(J L Reveal) httpwwwinformumdeduPBIOnomclrevehtml Univ of Maryland
SAFWAT F M 1962 The oral morphology of Secamoneand the evolution of the pollinating apparatus in As-clepiadaceae Ann Mo Bot Gard 4995ndash129
SAGHAI-MAROOF M A K M SOLIMAN R AJORGENSEN AND R W ALLARD 1984 Riboso-mal DNA spacer-length polymorphisms in barleyMendelian inheritance chromosomal location andpopulation dynamics Proc Nat Acad Sci USA818014ndash8018
SANGER F S NICKLEN AND A R COULSON 1977DNAsequencing with chain-terminating inhibitors ProcNat Acad Sci USA 745463ndash5467
SCHANDER C 1998 Types emendations and namesmdashAreply to Liden et al Taxon 47401ndash406
SCHANDER C AND M THOLLESSON 1995 Phyloge-netic taxonomymdashsome comments Zool Scr 24263ndash268
SCHUMANN K M 1895 Apocynaceae Pages 109ndash189 inDie naturlichen Panzenfamilien (H G A Engler andK A E Prantl eds) Wilhelm Engelmann LeipzigGermany
SENNBLAD B 1997 Phylogeny of the Apocynaceae slActa Univ Upsaliensis 29518pp
SENNBLAD B AND B BREMER 1996 The familial andsubfamilial relationships of Apocynaceae and Ascle-piadaceae evaluated with rbcL data Plant Syst Evol202153ndash175
SENNBLAD B AND B BREMER 2000 Is there a justi-cation for differential a priori weighting in coding
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead
2002 SENNBLAD AND BREMERmdashCLASSIFICATION OF APOCYNACEAE 409
sequences A case study from rbcL and ApocynaceaeSyst Biol 4943ndash55
SENNBLAD B M E ENDRESS AND B BREMER 1998Morphology and molecular data in phylogeneticfraternitymdashThe tribe Wrightieae (Apocynaceae) revis-ited Am J Bot 851143ndash1158
SERENO P C 1999 Denitions in phylogenetic taxon-omy Critique and rationale Syst Biol 48329ndash351
STRUWE L V A ALBERT AND B BREMER 1994 Cladis-tics and family level classication of the GentianalesCladistics 10175ndash206
SWARUPANANDAN K J K MANGALY T K SONNYK KISHOREKUMAR AND S CHAND BASHA 1996 Thesubfamilial and tribal classication of the family As-clepiadaceae Bot J Linn Soc 120327ndash369
SWOFFORD D L 1998PAUP Phylogenetic analysis us-ing parsimony (and other methods) version 40b3aSinauer Associates Sunderland Massachusetts
VENTER H J T AND R L VERHOEVEN 1997 A tribalclassication of the Periplocoideae (Apocynaceae)Taxon 46705ndash720
VERHOEVEN R L S LIEDE AND M ENDRESS 2002 Thetribal position of Fockea and Cibirhiza (ApocynaceaeAsclepiadoideae) Evidence from pollinium structureand cpDNA sequence data Grana (in press)
ZARUCCHI J L G N MORILLO M E ENDRESS B FHANSEN AND A J M LEEUWENBERG 1995 Apoc-ynaceae Pages 471ndash571 in Flora of the VenezuelanGuayana (J A Steyermark P A Berry and B KHolst eds)Missouri Botanical GardenTimber PressSt LouisPortland Oregon
First submitted 27 October 2000 revision submitted17 December 2001 nal accepted 31 December 2001
Associate Editor R Olmstead