klasifikasi zoin

Syst Biol 49(3)539ndash562 2000

Triploblastic Relationships with Emphasis on the Acoelomates and thePosition of Gnathostomulida Cycliophora Plathelminthes

and Chaetognatha A Combined Approach of 18S rDNASequences and Morphology

GONZALO G IRIBET15 DANIEL L DISTEL2 MARTIN POLZ3 WOLFGANG STERRER4

AND WARD C WHEELER1

1Division of Invertebrate Zoology American Museum of Natural History New York New York 10024-5192 USA2Department of Biochemistry Microbiology and Molecular Biology University of Maine Orono Maine 04469 USA

3Department of Civil and Environmental Engineering Massachusetts Institute of Technology 48-421 CambridgeMassachusetts 02139 USA

4Bermuda Natural History Museum Flatts FLBX Bermuda

AbstractmdashTriploblastic relationships were examined in the light of molecular and morphologicalevidence Representatives for all triploblastic ldquophylardquo (except Loricifera) were represented by bothsources of phylogenetic data The 18S ribosomal (rDNA) sequence data for 145 terminal taxa and276 morphological characters coded for 36 supraspecic taxa were combined in a total evidenceregime to determine the most consistent picture of triploblastic relationships for these data Onlytriploblastic taxa are used to avoid rooting with distant outgroups which seems to happen be-cause of the extreme distance that separates diploblastic from triploblastic taxa according to the18S rDNA data Multiple phylogenetic analyses performed with variable analysis parameters yieldlargely inconsistent results for certain groups such as Chaetognatha Acoela and NemertodermatidaA normalized incongruence length metric is used to assay the relative merit of the multiple analy-ses The combined analysis having the least character incongruence yields the following scheme ofrelationships of four main clades (1) Deuterostomia [((Echinodermata + Enteropneusta) (Cephalo-chordata (Urochordata + Vertebrata)))] (2) Ecdysozoa [(((Priapulida + Kinorhyncha) (Nematoda +Nematomorpha)) ((Onychophora + Tardigrada) Arthropoda))] (3) Trochozoa [((Phoronida + Bra-chiopoda) (Entoprocta (Nemertea (Sipuncula (Mollusca (Pogonophora (Echiura + Annelida)))))))]and (4) Platyzoa [((Gnathostomulida (Cycliophora + Syndermata)) (Gastrotricha + Plathelminthes))]Chaetognatha Nemertodermatida and Bryozoa cannot be assigned to any one of these four groupsFor the rst time adata analysis recognizes a clade of acoelomates the Platyzoa (sensu Cavalier-SmithBiol Rev 73203ndash266 1998) Other relationships that corroborate some morphological analyses arethe existence of a clade that groups Gnathostomulida + Syndermata (= Gnathifera) which is ex-panded to include the enigmatic phylum Cycliophora as sister group to Syndermata [EcdysozoaMetazoa morphology phylogeny Platyzoa 18S rRNA Triploblastica]

Metazoan and in particular triploblasticrelationships are becoming clearer as phylo-genetic techniques are applied to morpho-logical and molecular characters althoughthe rst high-level analysis combining bothmorphology and molecules (total evidence)was not published until 1998 (Zrzavy et al1998) That study constituted the most com-prehensive analysis published to date forthree reasons (1) all metazoan phyla andproblematic groups were coded for morphol-ogy (2) the most complete molecular sam-pling (in terms of number of phyla) was in-cluded and (3) this was the rst time thatmetazoan relationships were examined on

5Address correspondence to this author at Depart-ment of Organismic and Evolutionary Biology HarvardUniversity 26 Oxford St Cambridge Massachusetts02138 USA E-mail ggiribetoebharvardedu

the basis of the total evidence As a resultof this analysis Zrzavy et al (1998) elevateda considerable number of groups to phylumstatus and merged others

Almost simultaneously with Zrzavy et alCavalier-Smith (1998) proposed a ldquonew sys-tem of liferdquo based on six kingdoms withmajor changes in metazoan classicationHis new system tried to synthesize manyof the new relationships that molecular sys-tematists (in particular ldquo18S systematistsrdquo)had proposed in recent years although itwas not based on a character-based phy-logenetic analysis One of these groupsEcdysozoa might seem already to be a well-recognized metazoan group although it wasjust proposed in 1997 (Aguinaldo et al 1997)Another new group was Platyzoa which in-cludes a series of acoelomate and pseudo-coelomate (nonecdysozoan) taxa Also many

539

540 SYSTEMATIC BIOLOGY VOL 49

new phyla as well as other supra and inf-raphylum categories were erected and newnames introduced

The ultimate responsibility for the en-larged number of phylogenetic hypotheses ofrelationships among animal phyla lies withthe exponential growth of molecular dataand the renement of morphological andanatomical information through electron mi-croscopy The goal of this paper is to evalu-ate these hypotheses of relationship by usingnew molecular data in combination with themorphological data set published by Zrzavyet al (1998)

BACKGROUND

Cavalier-Smith (1998) in his six-kingdomsystem of life proposed dividing triploblas-tic animals into four infrakingdoms of pro-tostome animals (Lophozoa ChaetognathiEcdysozoa and Platyzoa) and into two in-frakingdoms of deuterostomes (Coelomo-pora and Chordonia) Lophozoa (previouslyreferred to as Eutrochozoa [Ghiselin 1988Eernisse et al 1992] or Lophotrochozoa[Halanych et al 1995]) included the classicalprotostome coelomates (Annelida EchiuraSipuncula Pogonophora Mollusca Ne-mertea) Entoprocta and Cycliophora plusthe ldquolophophoratesrdquo (Bryozoa Phoronidaand Brachiopoda) Zrzavy et al (1998) usedthe name Trochozoa for all Lophozoa ex-cept Brachiopoda and Phoronida Ecdysozoa(Aguinaldo et al 1997) included the molt-ing animals (Nematoda NematomorphaKinorhyncha Priapulida Loricifera Ony-chophora Tardigrada and Arthropoda) (seealso Giribet and Ribera 1998 Zrzavy et al1998) Chaetognathi include the enigmaticphylum Chaetognatha Perhaps the mostsurprising taxon in this new system of classi-cation is the so-called Platyzoa The compo-nents of this group were dened as ciliatednonsegmented acoelomates or pseudocoelo-mates that lack a vascular system and have astraight gut (when present) with or withoutanus Platyzoa thus include Rotifera Acan-thocephala Gastrotricha Gnathostomulidaand Plathelminthes (Cavalier-Smith 1998)

This system relocates many taxa far fromtheir previous positions in the classicationsFor example Platyzoa were divided into twophyla Acanthognatha (= Rotifera Acantho-cephala Gastrotricha and Gnathostomul-ida) and a monophyletic Plathelminthes

However Gastrotricha and Gnathostomul-ida are considered independent phyla bymost authors whereas Plathelminthes havebeen found to be nonmonophyletic in severalmorphological and molecular analyses

What Is the Evidence for the Monophylyof the Platyzoa

Winnepenninckx et al (1995) presentedan ldquoaschelminthrdquo phylogeny based on 18SrDNA sequences including platyzoan se-quences of Gastrotricha Rotifera Acantho-cephala and Rhabditophora which turnedout to constitute a monophyletic cladeCarranza et al (1997) also used 18S rDNAsequences of several platyzoan groups (Gas-trotricha Acanthocephala Acoela Nemerto-dermatida Catenulida and Rhabditophora)to address the question of Plathelminthesmonophyly The data suggested mono-phyly of Rhabditophora (including the Ne-mertodermatida) and nonmonophyly ofPlathelminthes (to the exclusion of Catenul-ida and Acoela)

In general molecular phylogenies haveproposed monophyly of Plathelminthes +Syndermata (eg Eernisse 1998 Littlewoodet al 1998) Plathelminthes + Gastrotricha +Syndermata (eg Winnepenninckx et al1995 Carranza et al 1997) Gnathostomul-ida + Gastrotricha (Zrzavy et al 1998)and Syndermata + Cycliophora (Winnepen-ninckx et al 1998) Together all these groupsexcept Cycliophora constitute the Platyzoasensu Cavalier-Smith (1998)

Different relationships among the platy-zoan taxa have been proposed by severalauthors on the basis of morphological data(eg Lorenzen 1985 Wallace et al 19951996 Neuhaus et al 1996 Ahlrichs 19951997 Nielsen 1995 Nielsen et al 1996Haszprunar 1996) a summary of these hy-potheses is shown in Figure 1 Meglitsch andSchram (1991) (see also Schram 1991) madethe rst attempt to analyze metazoan phy-logeny using parsimony algorithms Schramand Ellis (1994) reanalyzed an amendeddata set by using more-standard parsimo-ny procedures Their consensus cladogramcontained a basal clade of triploblastic ani-mals consisting of Gastrotricha RotiferaAcanthocephala Chaetognatha NematodaNematomorpha Kinorhyncha Priapulidaand Loricifera (= ldquoAschelminthesrdquo) Thisclade was followed by a clade containing

2000 GIRIBET ET ALmdashTRIPLOBLASTIC AND ACOELOMATE PHYLOGENY 541

FIGURE 1 Phylogenetic hypotheses for differentplatyzoan taxa (a) Lorenzen (1985) (b) Wallace et al(1995 1996) (c) Neuhaus et al (1996) (d) Ahlrichs(1995 1997) (e) Nielsen (1995) Nielsen et al (1996)(f ) Haszprunar (1996)

Gnathostomulida + Plathelminthes (= Pla-thelminthomorpha) as sister group tocoelomates Eernisse et al (1992) recodedSchramrsquos 1991 data matrix and createda new morphological data matrix for 26selected taxa Plathelminthomorpha weremonophyletic but Gastrotricha RotiferaAcanthocephala and some minor phylawere not included in the analyses Backel-jau et al (1993) performed another analy-sis recoding some of Schramrsquos 1991 charac-ters Their results showed little resolutionat the base of the triploblastic animals Theposition of Gastrotricha Chaetognatha andSyndermata among others was unresolvedin the consensus tree Plathelminthomorpha

were monophyletic within a clade of spi-ralian worms plus panarthropods

Wallace et al (1995 1996) analyzed re-lationships among pseudocoelomate taxaand found the following pattern (Syn-dermata (Gnathostomulida (Gastrotricha +Introverta))) (Fig 1b) In this case onlyPlathelminthes and Polychaeta were usedto test ldquopseudocoelomaterdquo monophyly Theparsimony analysis of Nielsen et al (1996)recognized three independent groups that in-volve platyzoan taxa Cycloneuralia (includ-ing the Gastrotricha) Parenchymia (includ-ing the Plathelminthes) and SyndermataWith respect to Gnathostomulida Nielsen(1995) considered them to be a modied an-nelid and thus did not include this taxon inhis analysis More recently the morphologi-cal parsimony analysis of Zrzavy et al (1998)supported monophyly of Gastrotricha +Syndermata as well as monophyly ofPlathelminthes (Catenulida + Acoela +Nemertodermatida + Rhabditophora) andPlathelminthomorpha (Gnathostomulida +Plathelminthes) Plathelminthomorpha wasthe sister group to Trochozoa (including Cy-cliophora and excluding the ldquolophophor-atesrdquo) and Gastrotricha + Syndermata weresister group to Plathelminthomorpha + Tro-chozoa Cycliophora appeared as the sistergroup to Entoprocta

GNATHOSTOMULIDA ENIGMATICAS EVER

The marine worm group Gnathostomul-ida was originally described by Ax (1956)from two species Gnathostomula paradoxafrom the North and Baltic Seas and Gnathos-tomaria lutheri from the Mediterranean al-though he considered them an order ofthe phylum Plathelminthes Later Ax (1960)separated them from the Turbellaria andcreated a new class Riedl (1969) elevatedthe Gnathostomulida to the rank of phy-lum and Sterrer (1972) divided the Gnathos-tomulida into the orders Filospermoidea(two families including Haplognathia) andBursovaginoidea (nine families includingGnathostomula) More recently Ax (eg 19841985 1996) considered Gnathostomulidaas the sister group to Plathelminthes thetwo of them constituting the supraphyleticcategory Plathelminthomorpha (Ax 1984)Sterrer et al (1985) questioned the pu-tative sister group relationship between


Gnathostomulida and Plathelminthes andpostulated alternative hypotheses of re-lationships for Gnathostomulida One ofthese hypotheses supported a relationship ofGnathostomulida + Gastrotricha based onthe monociliary epithelium and the struc-ture of protonephridia (see also Rieger andMainitz 1977) The other hypothesis postu-lated a grouping of Gnathostomulida + Ro-tifera based on tubular reinforcement of theinner jaw lamella (see also Rieger and Rieger1977 1980 Rieger and Tyler 1995)

The cladistic analyses of Meglitsch andSchram (1991) Schram (1991) Eernisse et al(1992) Backeljau et al (1993) and Schramand Ellis (1994) suggested monophyly ofPlathelminthomorpha However a study onthe epidermis of several members of thePlatyzoa (Ahlrichs 1995 1997) indicateda putative clade Gnathifera (= Gnathos-tomulida + Syndermata) as opposed toPlathelminthomorpha (Fig 1d) This grouphas also been proposed by Kristensen (1995)based on the jaw apparatus of a new freshwa-ter taxon from Greenland (New group A) byHaszprunar (1996) based on four morpho-logical characters (Fig 1f ) and by Herlynand Ehlers (1997) based on an anatomicalstudy of the pharynx of one gnathostomulid

Analyses of molecular data from 18SrDNA sequences published to date are am-biguous with respect to the position ofgnathostomulids Littlewood et al (1998)used a sequence of Gnathostomula paradoxathat came out as a member of EcdysozoaHowever using the same gnathostomulidse-quence Zrzavy et al (1998) found Gnathosto-mula to be a sister taxon to the two gastrotrichsequences included in their analyses

In summary morphology has suggestedthat gnathostomulids are either a memberof the Plathelminthes the Annelida or aresister group to Plathelminthes GastrotrichaRotifera or Syndermata The most recentmorphological analyses seem to favor a sis-ter group relationship between Gnathos-tomulida and Syndermata Additionally arelationship with Chaetognatha and withGastrotricha has been proposed by molecu-lar data

Cycliophora

The phylum Cycliophora was describedby Funch and Kristensen (1995) on the basisof a single species Symbion pandora living

on the mouthparts of the Norway lobsterNephrops norvegicus Originally a morpho-logic resemblance of cycliophorans with ro-tifers had been postulated by Wingstrand(unpublished observations cited by Funchand Kristensen 1997) but later Symbion pan-dora was thought to be related to Entoproctaand Ectoprocta(Funch and Kristensen 1995)or just to Entoprocta (Funch and Kristensen1997) The rst cladistic morphological anal-ysis including the phylum Cycliophora wasthat of Zrzavy et al (1998) which placedSymbion as sister group to Entoprocta Thisputative relationship of (Symbion + Ento-procta) was also postulated in the clas-sication of the metazoan kingdom byCavalier-Smith (1998) who considered themto be members of the phylum Kampto-zoa The rst molecular analysis of Sym-bion pandora used 18S rDNA sequences toplace it as the sister group to Syndermata(Winnepenninckx et al 1998) as predictedby Wingstrand but other taxamdashpostulatedto be related to Syndermata (Gnathosto-mulida Plathelminthes and Gastrotricha)mdashwere not included in the molecular analysis

Plathelminthes

Modern classications of Plathelminthesrecognize three groups Acoelomorpha(= Acoela + Nemertodermatida) Catenul-ida and a third group Rhabditophora thatcontains the remaining ldquoturbellarianrdquo ordersplus the parasitic classes (eg Karling 1974Ehlers 1985 1986 1995 Ax 1984 1996) Theposition of the parasitic groups (= Neoder-mata) within the Rhabditophora has beendisputed (eg Ehlers 1985 Ax 1984)

Molecular analyses of Plathelminthes arenumerous in the literature but only a fewhave investigated higher-level relationshipswithin the group (Katayama et al 1993 19951996 Rohde et al 1993 1995 Katayamaand Yamamoto 1994 Carranza et al 1997Giribet and Ribera 1998 Zrzavy et al1998 Littlewood et al 1999 Ruiz-Trilloet al 1999) The common conclusions ofthese studies are (1) all the major parasiticgroups constitute a single lineage (2) Rhab-ditophora is monophyletic including theparasitic forms (3) Plathelminthes may bepolyphyletic with acoels forming a sepa-rate clade (4) nemertodermatids are ei-ther included within Rhabditophora or forman independent clade and (5) catenulids


either form a separate clade or are sister toRhabditophora

Catenulida has been considered to beone of the earliest diverged groups ofPlathelminthes (Karling 1974 Ax 19631996 Ehlers 1985 1986) because of the pres-ence of several morphological traits consid-ered plesiomorphic for the Plathelminthestwo cilia in the terminal cell of the pro-tonephridium and epidermal cells weaklymulticiliated usually with two cilia per cellOther authors consider the putativesynapomorphies uniting Catenulida andRhabditophoramdashstructure of protonephri-dia structure of cilia epidermal movement(Ehlers 1985 1986)mdashto be convergent(Sterrer and Rieger 1974 Smith et al 1986)Furthermore the absence of synapomor-phies for Catenulida + Acoelomorpha hasled some authors to consider that catenulidsare not plathelminths (eg Haszprunar1996) The morphological parsimony anal-ysis of Zrzavy et al (1998) suggested apolytomy among Catenulida Acoelomor-pha and Rhabditophora whereas the mor-phological analysis of Littlewood et al(1999) showed the Catenulida as sistergroup of the Rhabditophora Phylogeneticstudies using 18S rDNA data consideredCatenulida either as sister group to Rhabdi-tophora (Rohde et al 1993 Carranza 1997Giribet and Ribera 1998) included withinRhabditophora (Katayama et al 1996) ororiginating independently of Rhabditophora(Carranza et al 1997 Zrzavy et al 1998)The combined (morphological + 18S rDNA)parsimony analysis of Zrzavy et al (1998)concluded that the origin of Catenulida wasindependent of that of other Plathelminthesand elevated Catenulida to the phylumrank the combined analysis of Littlewoodet al (1999) however placed Catenulidaand Rhabditophora as sister taxa

Acoelomorpha (Acoela + Nemertoder-matida) have been considered monophyletic(Karling 1974 Tyler and Rieger 1977Ehlers 1985 1986 Ax 1984 1996) Threesynapomorphies were proposed by Ax(1996) network formed by interconnectingrootlets of epidermal cilia shaft regionin epidermal cilia and absence of pro-tonephridia Other authors have consideredthe rootlet system (Lundin 1997 1998) andthe presence of degenerating epidermal(pulsatile) bodies (Lundin and Hendelberg1996) as the synapomorphies for the groupEmbryology of Acoela is autopomorphic

(Boyer 1971) and that of Nemertodermatidais unknown Reuter et al (1998) have alsosuggested that the brain-like structure ofAcoela is not homologous with the brains ofother Plathelminthes which may indicate anindependent origin of Acoela Nemertoder-matida constitute an enigmatic group withonly eight known species (Sterrer 1998)dened by the presence of a statocyst withtwo statoliths and several parietal cells (Ax1996) Steinbock (1930) described the rstspecies of the group Nemertoderma bathycolawhich was dredged from a muddy bottom300ndash400 m deep off Greenland and at thetime was considered to be the most primitivebilaterian

Whether the Acoelomorpha or theCatenulida represents the rst branchingevent of Plathelminthes has been disputedKarling (1974) and Zrzavy et al (1998morphological tree) did not resolve theposition of Catenulida and AcoelomorphaThe morphological analysis of Zrzavy et al(1998) and Littlewood et al (1999) supportedmonophyly of Acoelomorpha but themolecular and the total evidence analysesdid not Acoela and Nemertodermatidawereeach elevated to the phylum rank by Zrzavyet al (1998) Other molecular analyses of 18SrDNA sequences have suggested that Acoelahave a different origin from the remainingPlathelminthes (Katayama et al 1995 1996Carranza et al 1997 Zrzavy et al 1998Littlewood et al 1999 Ruiz-Trillo et al1999) whereas Nemertodermatida havebeen found to belong within Rhabditophora(Carranza et al 1997 Zrzavy et al 1998Littlewood et al 1999)

Chaetognatha

Chaetognaths still constitute one of themostenigmatic animal phyla in terms of theirphylogenetic relationships Darwin (1844)mentioned the obscurity of their afnitiesand later Ghirardelli (1968) argued thatchaetognaths were not closely related to anyextant metazoan phylum These ideas reectthe lack of unambiguous synapomorphies tounite chaetognaths with other phyla Oftenplaced within the Deuterostomia because oftheir method of coelom formation (Hyman1959) Meglitsch and Schram (1991) ques-tioned the enterocoelic nature of their bodycavities Both development and morphologyof the adult nervous system are typicallyprotostomian the early determination of the


germ cells is typical of many ldquoaschelminthsrdquoand the toothed oral membrane with its veryhigh content of chitin resembles the mastaxof rotifers (Nielsen 1995)

The rst cladistic analysis of the meta-zoan phyla (Meglitsch and Schram 1991)placed chaetognaths within a clade ofldquoaschelminthsrdquo Nielsen (1995) also placedChaetognaths in a clade of ldquoaschelminthsrdquothat included Gastrotricha NematodaNematomorpha Priapulida KinorhynchaLoricifera Rotifera and Acanthocephalain his cladistic analysis however Chaetog-natha constituted the sister group to Synder-mata but did not group with the remainingaschelminths The morphological analysis ofZrzavy et al (1998) placed Chaetognatha asthe sister group of a clade containing all otherprotostome phyla except lophophorates

Chaetognath afnities have been proposedon the basis of molecular analyses using 18SrDNA sequence data (Telford and Holland1993 Wada and Satoh 1994 Halanych1996) Telford and Holland (1993) and Wadaand Satoh (1994) concluded that chaetog-naths were not deuterostomes although thetaxonomic sampling used did not allow theirphylogenetic position to be established moreaccurately Halanych (1996) concluded thatchaetognaths were sister group to nema-todes postulating an evolutionary scenariofor the origin of chaetognaths from a vermi-form benthic organism The clade contain-ing (Chaetognatha + Nematoda) was sis-ter group to Plathelminthes Other analyses(Eernisse 1998) placed chaetognaths eitheras sister group to Nematomorpha withinthe Ecdysozoa with the Nematoda at thebase of the tree or in a clade containing(Nematomorpha + Nematoda + Chaetog-natha) within Ecdysozoa Littlewood et al(1998) placed Chaetognatha as sister group toGnathostomulida with both as sister groupto Nematoda within Ecdysozoa Chaetog-naths have extremely divergent 18S rDNAsequences in comparison with other meta-zoans and in all the analyses published sofar have tended to group with other diver-gent sequences such as those of nematodesor gnathostomulids

Metazoan PhylogenymdashIs There a RootingProblem for the Bilateria

Recent analyses of metazoan taxa basedon large 18S rDNA data sets (Eernisse1998 Giribet and Ribera 1998 Littlewoodet al 1998 Zrzavy et al 1998 Giribet

and Wheeler 1999) basically agreed inthe monophyly of the triploblastic animals(= Bilateria) and in the presence of four maingroups of triploblastic animals Deuteros-tomia Ecdysozoa Platyzoa and TrochozoaHowever the relationships among these fourgroups some of which appear to be para-phyletic have been problematic It also hasnot been possible to resolve the internal rela-tionships among these four main clades con-sistently especially with regard to whetherthe rst dichotomy within the Bilateria isDeuterostomia versus Protostomia or Platy-zoa versus coelomates (Zrzavy et al 1998)For this reason it has been suggested thatthere might be a rooting problem for the Bi-lateria because the branch separating themfrom the diploblastic animals is too longand thus may have accumulated too manychanges (Giribet and Wheeler 1999Fig 2)As Wheeler (1990) pointed out distant out-groups may lead to spurious relationshipsbased on random similarity a phenomenonthat may apply also to the phylogenetic re-construction of the Bilateria on the basis of18S rDNA sequences

THE NEW ANALYSIS

In an attempt to resolve the myriad of hy-potheses proposed for the interrelationshipsamong triploblastic phyla especially for theinterrelationships among the ldquoaschelminthrdquoand platyzoan phyla we have analyzed anenlarged 18S rDNA data set combined witha morphological data matrix

1 Taxonomic sampling was improvedwithin each phylum with 23 unpublished18S rDNA sequences including twonew Gnathostomulida (Gnathostomulasp and Haplognathia sp) one ldquoarchian-nelidrdquo (Dinophilus gyrociliatus) one newNemertodermatida (Meara stichopi) andnew sequences of other several phylaMollusca Sipuncula Echiura NemerteaBrachiopoda Phoronida Bryozoa Pria-pulida Onychophora Arthropoda andEnteropneusta (see Appendix 1)

2 Only triploblastic taxa were analyzedto avoid rooting with distant outgroupsThis strategy may be useful to test severalof the hypotheses formulated here andto avoid problems with putative sequenceheterogeneity

3 Data were analyzed by using the direct op-timization method (Wheeler 1996) whichavoids intermediate alignment steps


4 A sensitivity analysis using seven param-eter sets for three data sets (18S com-plete 18S without ve variable regions18S without variable regions + morphol-ogy) were conducted to avoid formulatingnonrobust hypotheses

5 Character congruence (ILD of Mickevichand Farris 1981) was used as an ex-ternal criterion to choose the parameterset that minimizes incongruence amongmolecules and morphology

In total we used complete 18S rDNA se-quences of 145 species of Bilateria ( laquo 1800ndash2300 bp) and 276 morphological charactersfor 36 morphological terminals All majortriploblastic groups or putative phyla (ex-cept Pterobranchia and Loricifera) were rep-resented The morphological data set was ex-tracted from Zrzavy et al (1998) and adjustedfor the taxa represented by use of moleculardata We tested ve main hypotheses

1 Are the members of the Platyzoa (sensuCavalier-Smith 1998) directly related toeach other or are certain platyzoans moreclosely related to the introvertan pseudo-coelomates

2 Is the phylum Gnathostomulida relatedto other acoelomate taxa (Platyzoa) orto the introvertan pseudocoelomates in-cluded within Ecdysozoa

3 Are the Plathelminthes monophyletic4 Is Symbion pandora (Cycliophora) more

closely related to Entoprocta or to Synder-mata

5 Is the phylum Chaetognatha a member ofthe Ecdysozoa

MATERIALS AND METHODS

Taxon Sampling

Multiple sampling within each triploblas-tic phylum was attempted and additionaltaxa were added to phyla for which onlyone sequence was available This was notpossible for Kinorhyncha Cephalochordataor Cycliophora Sequence data from thephylum Loricifera are not existent and forthe hemichordate class Pterobranchia onlya small fragment of the 18S rDNA geneis available in GenBank and it was notincluded in our analyses Complete 18SrDNA sequences of the following groups(and no of species sampled) were in-cluded in this investigation Polychaeta (9)

Clitellata (4) Mollusca (10) Sipuncula (2)Echiura (2) Pogonophora (2) Nemertea (3)Brachiopoda (4) Phoronida (2) Bryozoa (4)Entoprocta (2) Cycliophora (1) Rotifera (2)Acanthocephala (4) Gastrotricha (2) Gnath-ostomulida (3) Acoela (2) Nemertoder-matida (2) Catenulida (2) Rhabditophora(15) Priapulida (2) Kinorhyncha (1) Ne-matomorpha (2) Nematoda (8) Onychop-hora (2) Tardigrada (2) Arthropoda (22)Enteropneusta (2) Echinodermata (10) Uro-chordata (3) Cephalochordata (1) Crani-ata (11) and Chaetognatha (2) Among the145 18S rDNA sequences used in this investi-gation 37 have been obtained by the presentauthors (25) Higher taxonomic ranks usedin the paper are summarized in Tables 1and 2 Taxonomy of the terminal taxa is givenin Appendix 1

DNA Sequences

Genomic DNA samples were obtainedfrom fresh frozen or ethanol-preserved tis-sues in a solution of guanidinium thio-cyanate homogenization buffer followinga modied protocol for RNA extraction(Giribet et al 1999) The 18S rDNA loci wereamplied by polymerase chain reaction intwo or three overlapping fragments of laquo 950900 and 850 bp each with primer pairs 1Fndash5R 3Fndash18Sbi and 5Fndash9R respectively andwere sequenced by using standard cycle-sequencing protocols (see primers and de-tailed protocols in Giribet et al 1999) All thenew sequences have been deposited in Gen-Bank (see accession codes in Appendix 1)

Morphological Data

The morphological data set was extractedfrom the data matrix presented by Zrzavyet al (1998) Taxa representing diploblas-tic animals as well as certain triploblastictaxa for which no molecular data were avail-able (Xenoturbellida Pterobranchia Lori-cifera Buddenbrockia Lobatocerebromorphaand Myzostomida) were excluded from themorphological data set In total 36 termi-nal higher-taxa and 276 characters were usedfor the morphologicalanalysis All characterswere treated as unordered and no differen-tial weighting was applied The coding strat-egy combines an exemplar approach of themolecular taxa (real taxa) with a groundplanapproach for the morphological terminalsThis approach has been adopted because of


Cavallier-Smith (1998) Brusca and Brusca (1990)

Bryozoa

Kamptozoa

Mollusca

Brachiozoa

Annelida

SipunculaNemertina

ChaetognathaArthropoda

Lobopoda

Nemathelminthes

Acanthognatha

Platyhelminthes

HemichordataEchinodermata

UrochordataChordata

Zrzavy et al (1998)

Bryozoa

EntoproctaCycliophora

Mollusca

Phoronozoa

AnnelidaPogonophora

Echiura

SipunculaNemertea


OnychophoraTardigrada

NematodaNematomorphaCephalorhyncha

Syndermata

GastrotrichaGnathostomulida

AcoelaNemertodermatida

RhabditophoraCatenulida


Chordata

Xenoturbellida

Chordata

Ectoprocta

Entoprocta

Mollusca

PhoronidaBrachiopoda

AnnelidaPogonophoraVestimentifera

Echiura

SipunculaNemertea


Pentastomida


NematodaNematomorpha

PriapulaKinorhyncha

Loricifera

RotiferaAcanthocephala


Platyhelminthes


UrochordataCephalochordata

TABLE 1 Triploblastic phyla recognized by three different authors

the difculties of coding all morphologicaltraits for the exact same species as are used inthemolecular approach further codingof themorphologicaldata for exemplar species willbe presented in the future because it seemsto us to be a more defensible strategy

Phylogenetic Analyses

Homology concept in sequence datamdashAl-though most molecular analyses use strictbase-to-base correspondences (a xed align-ment) as their primary homology statementthis introduces ambiguity and does not al-low the accommodation of sequences of sub-stantially unequal length (see Wheeler 1996)In contrast our rst hypothesis of homologycorresponds to secondary structure features

(see below) followed by a dynamic base-to-base correspondence as described by the ldquodi-rect optimizationrdquo method (Wheeler 1996)To do this the sequences are divided into thesmallest possible unambiguously recogniz-able homologous regions For the rst splitwe used primer regions and then we iden-tied secondary structure features In to-tal the 18S rDNA molecule was divided into47 regions (excluding the external primers 1Fand 9R) for each terminal taxon The 47 inputles contained the unaligned sequences of allterminal taxa All 47 sequence les param-eter les and batch les are available fromthe anonymous ftp site ftpscienceamnhorgpubmoleculardatagnathostomulida

Sequence data analysis direct optimizationmdashSequence data were analyzed by using the


TABLE 2 Some proposed nomenclature involving acoelomate groups Names in bold type are accepted in thepresent work

Acanthognatha (Cavalier-Smith 1998)Rotifera + Acanthocephala + Gastrotricha + Gnathostomulida

Cephalorhyncha (Malakhov 1980 Nielsen 1995) (= Scalidophora [Lemburg 1995])Priapulida + Kinorhyncha + Loricifera

Cycloneuralia (Nielsen 1995) (= Nemathelminthes ss [Neuhaus 1994 Schmidt-Rhaesa 1997])Gastrotricha + Nematoda + Nematomorpha + Priapulida + Kinorhyncha + Loricifera

Ecdysozoa (Aguinaldo et al 1997)Nematoda + Nematomorpha + Priapulida + Kinorhyncha + Loricifera + Onychophora + Tardigrada +

ArthropodaGnathifera (Ahlrichs 1995 1997)

Gnathostomulida + Rotifera + AcanthocephalaIntroverta (Nielsen 1995) (= Cycloneuralia [sensu Ahlrichs 1995])

Nematoda + Nematomorpha + Priapulida + Kinorhyncha + LoriciferaMonokonta (Cavalier-Smith 1998) (= Neotrichozoa [Zrzavy et al 1998])

Gastrotricha + GnathostomulidaNematoida (Schmidt-Rhaesa 1996)

Nematoda + NematomorphaNematozoa (Zrzavy et al 1998)

Nematoda + Nematomorpha + Onychophora + Tardigrada + ArthropodaParenchymia (Nielsen 1995)

Nemertini + PlathelminthesPlathelminthomorpha (Ax 1984)

Gnathostomulida + PlathelminthesPlatyzoa (Cavalier-Smith 1998)

Rotifera + Acanthocephala + Gastrotricha + Gnathostomulida + PlathelminthesSyndermata (Ahlrichs 1995 1997 Zrzavy et al 1998) (= Trochata [Cavalier-Smith 1998])

Rotifera + Seison + Acanthocephala

direct optimization method described byWheeler (1996) and implemented in the com-puter program POY (Gladstein and Wheeler1997) The method assesses directly the num-ber of DNA sequence transformations (evo-lutionary events) required by a phylogenetictopology without the use of multiple se-quence alignment This is accomplishedthrough ageneralization of existingcharacteroptimization procedures to include insertionand deletion events (indels) in addition tobase substitutions The crux of the proce-dure is the treatment of indels as processesrather than as patterns implied by multi-ple sequence alignment That is the neces-sary indel events that occurred during recon-structing the ancestor of two given sequencesare counted but are not reected in a ldquoxedalignmentrdquo The result of this procedure is di-rectly compatible with parsimony-based treelengths and appears to generate more ef-cient (simpler) explanations of sequence vari-ation than doesmultiple alignment (Wheeler1996) Moreover the method is much lessdemanding than parsimony-based multiplesequence alignment algorithms and yieldsmore congruent results than multiple se-quence alignments when congruence amongpartitions is used as a criterion (Wheeler andHayashi 1998)

Sensitivity analysismdashCharacter transfor-mations were weighted differentially to seehow they affect phylogenetic conclusions(sensitivity analysis sensu Wheeler 1995)A parameter space of two analytical vari-ables was examined insertionndashdeletion ra-tio and transversionndashtransition ratio (tvts)(as in Wheeler 1995) When the tvts wasset as 6= 1 the insertionndashdeletion cost wasset according to the cost of transversionsthat is gapchange = 2 tvts = 2 gaps costtwiceas much as transversions and transver-sions cost twice as much as transitions In to-tal seven combinations of parameters wereused in the analysis gaptvts = 111 121141 211 221 241 and 411 We consideredthis an effective way to explore the data andto discern between well-supported relation-ships (those supported throughout a widerange of parameters) and poorly supportedrelationships (those that appear only withvery particular parameter sets) The molec-ular analyses were performed for the com-plete sequences (all 47 fragments) and fora reduced data set that excluded 5 of the47 regions (E10-1 E10-2 E21-1-2 41 and 47)showing large variations in sequence lengthamong the sampled taxa

Morphological data analysismdashA parsimonyanalysis of the morphological data set was


TABLE 3 Tree length for the individual partitions (18S = 18S rDNA data set Morph = morphological data setweighted as the highest molecular cost [molecC]) and for the combined data set (Total) The congruence amongpartitions is measured by using the ILD metrics (Mickevich and Farris 1981) IndelC = insertionndashdeletion cost ratioTvTi = transversion-transition ratio molecC = highest molecular cost (is calculated multiplying IndelC x TvTi)

IndelC TvTi molecC 18S Morph Total ILD

1 1 1 11336 2185 13628 0007851 2 2 16892 4370 21563 0013961 4 4 27848 8740 36952 0009852 1 2 12799 4370 17329 0009232 2 4 19691 8740 28769 0011752 4 8 33013 17480 51312 0015964 1 4 15213 8740 24373 001723

performed with the computer programNONAv 19 (Goloboff 1998) The tree searchstrategy adopted involved a heuristics algo-rithm with random addition-sequence andTBR branch-swapping No further more-sophisticated strategies were required be-cause all of the 100 replications performedyielded the same result Branch support(Bremer 1988) up to ve extra steps was cal-culated by using a heuristic procedure andholding a maximum of 10000 trees withNONA (Goloboff 1998)

Combined analysismdashMorphological andmolecular data were combined directly andanalyzed using the ldquodirect optimizationrdquomethod (Wheeler 1996) For the combinedanalyses the reduced molecular data set (theone excluding theve hypervariable regions)was used The morphological transforma-tions were weighted as equal to the greatestof the molecular costs (= indels) Branch sup-port (Bremer 1988) was calculated by usinga heuristic procedure implemented in POY

Character congruencemdashCongruence am-ong partitions (morphological and molec-ular) was measured by the ILD metrics(Mickevich and Farris 1981) (see Table 3)This value is calculated by dividing thedifference between overall tree length andthe sum of its data components ILD =(lengthcombined iexcl R lengthindividual sets)lengthcombined Character congruence is thusused as the criterion to choose our best (mostcongruent) tree the tree that minimizescharacter conict among all the data Thisis understood as an extension of parsimony(or any other minimizing criteria) in thesame sense that parsimony tries to minimizethe number of overall steps in a tree theldquocharacter congruence analysisrdquo tries to ndthe model that minimizes incongruence forall the data sources

In the present case seven parameter setswere analyzed (a total of 21 analyses) thattook laquo 4 months of computing time on a clus-ter of 10 PC processors (200 MHz) workingin parallel Other phylogenetic studies haveexplored wider ranges of parameter sets aswell as different relative weights betweenmorphology and molecules (ie Wheelerand Hayashi 1998) but the currently avail-able computer technology did not allow us toundertake the same approach in a reasonableamount of time

RESULTS

The trees presented here have been rootedarbitrarily in the branch that separatesdeuterostomes from protostomes as hasbeen found in several morphological analy-ses (eg Nielsen et al 1996 Zrzavy et al1998) However other possibilities havebeen also proposed Thus the term mono-phyly does not strictly apply to the unrootedtopologies presented

Molecular Analysis

Figure 2 shows summary trees (Deuteros-tomia collapsed) of the results for parameterset 111 (gap cost = tv = ts = 1 the parame-ter set that minimizes incongruence [see be-low]) for the complete data set (all 47 frag-ments included Fig 2a) and for the reduceddata set (42 fragments included Fig 2b)When the complete data set is used theplaty-zoan taxa (in red) appear in four independentclades (1) grouping the acoels (Amphiscolopsand Convoluta) plus the two gnathostomulidspecies of the genus Gnathostomula withinthe Ecdysozoa (in green) (2) a clade con-taining the cycliophoran (Symbion) the twogastrotrics (Chaetonotus and Lepidomermella)the other gnathostomulid(Haplognathia) and


FIGURE 2 18S rDNA trees for parameter set 111 (gap = change tv = ts) when the complete gene sequenceis used (a) or when the most heterogeneous regions are removed (b) The deuterostomes have been collapsedColors represent major protostome groups Ecdysozoa (green) Platyzoa (red) and Trochozoa (blue) Branches ofthe platyzoan taxa are represented in red


one nemertodermatid (Meara) (3) a cladecontaining the Syndermata and (4) a cladecontaining the remaining plathelminths (in-cluding the nemertodermatid Nemertinoides)(Fig 2a) The last three clades appear scat-tered among the trochozoan taxa (in blue)

The same parameter set for the data setthat excludes the ve hypervariable regions(Fig 2b) yields platyzoan monophyly (inred) except for Meara (one of the two ne-mertodermatid sequences) which groupswith the chaetognaths Platyzoa is the sis-ter group to a clade of trochozoans (exclud-ing Capitella Caberea and Membraniporamdasha polychaete and two gymnolaematan bry-ozoans respectively) Platyzoa + Trochozoaconstitute the Spiralia In this tree four maingroups (deuterostomes ecdysozoans tro-chozoans and platyzoans) are recognizedexcept for the taxa previously mentionedplus onychophorans and chaetognaths Thisstructure is unstable to parameter variation

Morphological Analysis

The parsimony analysis of the morpholog-ical data matrix resulted in 16 trees of 455steps (consistency index = 0466 retentionindex = 0701) The strict consensus of these16 trees is represented in Figure 3 In theconsensus tree (arbitrarily rooted betweendeuterostomes and protostomes) Deuteros-tomia Trochozoa Ecdysozoa and Platy-zoa are recognized The ldquolophophoratesrdquoappear between deuterostomes and proto-stomes (their relationship depends on thepoint where the tree is rooted) Other rele-vant results (that conict with the molecu-lar analyses) are that Syndermata branchesoutside the Platyzoa Chaetognatha appearsas sister group to Ecdysozoa and the cyclio-phoran is placed within the Trochozoa

Branch support = 1 for almost all supra-phyletic nodes except for the following allthe chordate phyla (Bremer support [bs] =3) the Trochozoa (excluding Entoprocta)(bs = 3) and some other groupings withinthis clade Syndermata (bs = 5) Nematoida(bs = 3) (Priapulida (Kinorhyncha (Ony-chophora Tardigrada Arthropoda))) (bs =2 2 gt5 for each respective node) andPlathelminthomorpha (bs = 2) All the cladeswith branch support gt2 are also found in thetotal evidence tree except for Acoelomorpha(Acoela + Nemertodermatida) which has amorphological branch support of 4 but is notfound in the combined analysis tree

FIGURE 3 Strict consensus of 16 trees of 455 steps(consistency index = 0466 retention index = 0701)based on the morphological data of Zrzavy et al (1998)Bremer support values gt1 are indicated

Combined Analysis

The total evidence analyses achieved aminimum of incongruence at equal weights(indel = tv = ts = morphology parameterset 111) (Table 3) Thus this tree (Fig 4) rep-resents our best hypothesis of relationshipsbeing the one that minimizes incongruenceamong the different sources of data A sum-mary tree with the taxonomic categories ascoded for the morphological analysis is rep-resented in Figure 5


Our best tree presented four main convex(nonpolyphyletic) groupings the Deuteros-tomia and three main groups of proto-stome animalsmdashEcdysozoa Platyzoa andTrochozoa Three major taxa (ChaetognathaNemertodermatida and Bryozoa) could notbe assigned to any of the four main groups

Deuterostomia (bs = 17) contained twomain clades one of nonchordate animals(Echinodermata + Enteropneusta) and oneof chordate animals (Cephalochordata (Uro-chordata + Craniata)) (Coelomopora andChordonia respectively sensu Cavalier-Smith 1998) Ecdysozoa (bs = 11) couldalso be divided into two main clades oneincluding the introvertan pseudocoelomatephyla ((Priapulida + Kinorhyncha) (Ne-matoda + Nematomorpha)) (= Introvertasensu Nielsen 1995) and another includ-ing arthropods and related phyla ((Ony-chophora + Tardigrada) Arthropoda) Tro-chozoa (bs = 3) includes the followinggroups ((Phoronida + Brachiopoda) (En-toprocta (Nemertea (Sipuncula (Mollusca(Pogonophora (Echiura + Annelida)))))))being the grouping with the lowest branchsupport

Platyzoa (bs = 5) constitute a clade thatincludes all the classical acoelomate phyla(except Nemertodermatida) Two mainsubgroups were recognized one containing(Gnathostomulida (Cycliophora (Mono-gononta (Bdelloidea + Acanthocephala))))and another containing (Gastrotricha +Plathelminthes) The internal structure ofPlathelminthes was as follows (Catenulida(Macrostomum ((Microstomum + Polycladida)(ldquoNeoophorardquo + Acoela)))) Catenulids weresister group to the remaining PlathelminthesMacrostomidswere nonmonophyletic beingMicrostomum sister to Polyclads Neoophoracontained the remaining orders includingthe Acoela as sister group to Tricladida

DISCUSSION

Platyzoa

Platyzoan taxa to the exclusion of Nemer-todermatida were convex in our favored tree(monophyletic if the rooting was correct orparaphyletic if the root was placed withinthe Platyzoa) This result is consistent withsome molecular analyses (eg Winnepen-ninckx et al 1995) However putative ldquolong-branchrdquo problems with certain groups ofPlathelminthes and other acoelomates mayhave been inuential in the failure by other

authors to obtain this clade by molecular dataanalyses (see below)

Most morphological analyses have consid-ered platyzoans to be polyphyletic or para-phyletic although many times this hypothe-sis has been assumed without testing ThusSyndermata and Gastrotricha (and some-times Gnathostomulida) have been often re-lated to the introvertan ecdysozoans (ieLorenzen 1985 Wallace et al 1995 1996Neuhaus et al 1996 [shown in Figs 1andashc this paper]) Other authors consideredSyndermata as an independent clade (ieFig 1e) Haszprunar (1996 shown in Fig 1fhere) proposed a paraphyletic grade of platy-zoans that led to the ldquohigher spiraliansrdquo

The absence of coelom (as dened histo-logically) is the only morphological synapo-morphy that might dene Platyzoa althoughits optimization is ambiguous (this characterstate is also present in Entoprocta Nemer-todermatida Kinorhyncha Nematoda andNematomorpha it is coded as unknown forSyndermata) This clade is thus mainly de-ned by molecular characters

Gnathifera Gnathostomulida Cycliophoraand Syndermata

The two classes of the phylum Gnathos-tomulida (Bursovaginoidea and Filisper-moidea) appeared to be monophyletic undermost parameter sets in the molecular anal-yses although not in those with high gapcosts The monophyly of the phylum wasnot tested in the morphological analysisfor which the three species were codedidentically Our combined tree also sup-ported gnathostomulid monophyly not sur-prisingly given that the three species havebeen coded identically for the morphology

Gnathostomulids appeared as sister taxato a clade containing (Symbion + Synder-mata) The close relationship of Symbion (Cy-cliophora) with Syndermata agrees with themolecular analysis of Winnepenninckx et al(1998) but contradicts the original hypoth-esis of a close relationship to Entoprocta(Funch and Kristensen 1995 1997) That theGnathostomulida constitute the sister taxonof (Syndermata + Cycliophora) could be con-sidered an extension of the Gnathifera hy-pothesis (Ahlrichs 1995 1997 Haszprunar1996 Herlyn and Ehlers 1997 Kristensen1995 Rieger and Rieger 1977 1980 Riegerand Tyler 1995) (see Figs 1d 1f ) whichdid not include the phylum Cycliophora


FIGURE 4 Total evidence tree (molecular data with the ve most heterogeneous regions removed) for parameterset 111 Numbers on branches indicate Bremer support values


FIGURE 4 Continued


FIGURE 5 Summary tree of Figure 4 with the termi-nal taxa as coded for the morphology The dashed linein Macrostomida indicates nonmonophyly Numberson branches indicate Bremer support values

In consequence the Plathelminthomorphahypothesis (Ax 1984 Meglitsch and Schram1991 Schram 1991 Eernisse et al 1992Backeljau et al 1993 Schram and Ellis1994 Zrzavy et al 1998) which postulateda sister group relationship of Gnathosto-mulida with Plathelminthes was rejectedOther hypotheses such as the relation-ship of Gnathostomulida with Ecdysozoa(Littlewood et al 1998) were rejected aswell by the current analyses Under one para-

meter set (221) however the reduced molec-ular data set placed gnathostomulids withinecdysozoans

The monophyly of Gnathifera is supportedby one unambiguous morphological synapo-morphy (presence of protonephridia withchannel cell completely surrounding lumen)although this state was coded as unknownfor Cycliophora The inclusion of Cyclio-phora within the Gnathifera was entirely be-cause of the molecular data

Syndermata was proposed by Zrzavyet al (1998) as a new phylum combiningRotifera and Acanthocephala based on theparaphyletic status of Rotifera Paraphylyof Rotifera with respect to Acanthocephalahad already been proposed by Lorenzen(1985) and subsequent authors (Ahlrichs1995 1997 Garey et al 1996 Giribet andRibera 1998 Littlewood et al 1998 1999Winnepenninckx et al 1998) Our resultsagree with this hypothesis and also suggesta sister group relationship between Cyclio-phora and Syndermata

Plathelminthes Nemertodermatida Acoelaand Putative ldquoLong Branchesrdquo

Nemertodermatids were represented inour analyses by two sequences of the genusMeara and Nemertinoides and morphologywas coded identically for both taxa Themolecular dataanalyses placed Nemertinoidestogether with Urastoma in eight of nine pa-rameter sets related either to other Platyzoaor within Deuterostomia Only in one casedid Nemertinoides come out as sister taxonto Meara within a clade of platyzoan taxa(parameter set 121) as was expected How-ever the position of Meara was extremelyparameter-dependent

The 18S rDNA of Nemertinoides elongatuswas sequenced by S Carranza although itdid notappear related toUrastoma in his anal-yses (Carranza et al 1997) The same specieswas used by Zrzavy et al (1998) but in theiranalysis it appeared as sister to the rhab-docoelan Bothromesostoma sp This Nemerti-noides sequence was also used by Littlewoodet al (1999) but in that case it grouped withsome proseriate sequences In our analysesBothromesostoma appeared as sister taxon toMesocastrada (both Rhabdocoela) throughoutall the parameters explored Meara stichopi isa new sequence that has not been includedin any previous analyses It seems that to


resolve their denitive position more sam-pling might be needed within the Nemerto-dermatida even though only eight speciesare known The combined analyses ldquoforcedrdquothe two nemertodermatids to group togetherbecause they had been coded identically forthe morphology In the best tree they ap-peared to be sister group to chaetognathsbut in the second best tree (parameter set 121tree not shown) they appear to be the sistergroup to all the remaining Plathelminthesthus making the phylum monophyletic

According to our analyses the only mor-phological ldquosynapomorphyrdquo for Nemerto-dermatida + Chaetognatha would be thepresence of hermaphroditism which ishighly homoplastic and is found in somePriapulida Kinorhyncha GnathostomulidaGastrotricha Plathelminthes Nemertoder-matida Clitellata and Mollusca

Plathelminthes (to the exclusion of Ne-mertodermatida) appeared as a clade in ourbest tree with a branch support of eight (pa-rameter set 121 also included Nemertoder-matida with the remaining Plathelminthes)Catenulida appeared as the sister taxon tothe remaining groups and Acoela was sis-ter group to Tricladida The phylogenetic po-sition of catenulids has been problematicas noted above (see for example Carranzaet al 1997 Giribet and Ribera 1998) Zrzavyet al (1998) elevated the group to the phy-lum category however our total evidenceresults disagree (see also Rhode et al 1993Carranza 1997 Giribet and Ribera 1998Littlewood et al 1999 Ruiz-Trillo et al1999)

The Acoela were also elevated to the phy-lum category by Zrzavy et al (1998) The phy-logenetic position of acoel plathelminths cer-tainly has been problematic Morphologicalsynapomorphies with other plathelminthsare scarce (densely multiciliated epidermalcells frontal organ or frontal glandularcomplex [Ax 1996]) Furthermore becausemost of the molecular analyses publishedso far did not place acoels with the remain-ing plathelminths many authors have con-cluded that plathelminths are polyphyletic(Katayama et al 1995 Carranza et al 1997Zrzavy et al 1998 Littlewood et al 1999Ruiz-Trillo et al 1999) although a molecularanalysis (Noren and Jondelius 1997) showedmonophyly of Plathelminthes including theacoel Praesagittifera Some of these authorshave also proposed that acoels might repre-

sent a basal triploblastic animal (Zrzavy et al1998 Ruiz-Trillo et al 1999)

The available 18S rDNA sequences ofacoels have large phenetic dissimilaritieswith respect to other metazoan taxa whichresult in branches considerably longer thanthe average branches of the tree (seeCarranza et al 1997Fig 3) When such a phe-nomenon occurs positioning of these taxabased solely on sequence data is difcultThis phenomenon often referred to as ldquolongbranch attractionrdquo (among other names) hasbeen used perhaps excessively to explainmany anomalies in phylogenetic trees andalthough ldquolong branchrdquo problems may existthey are difcult to demonstrate empirically

A putative ldquolong branchrdquo example in theliterature is the position of nematodes inmetazoan trees based on 18S rDNA sequencedata Nematodes are recognized as mem-bers of the Introverta (sensu Nielsen 1995)in many morphological analyses but theyhave also repeatedly been found at basalpositions within triploblastic animals (as isthe case of Acoela) (Aguinaldo et al [1997]summarize this problem) Two subsequentstrategies were followed that seemed to solvethis putative ldquolong branchrdquo problem Therst adopted by Aguinaldo et al (1997) wasto sequence the 18S rDNA loci of severalnematodes and use only the most slowlyevolving taxa This strategy led them toldquorepositionrdquo nematodes with their morpho-logically recognized relatives the other intro-vertan pseudocoelomates and to dene theclade Ecdysozoa The second strategy fol-lowing the recommendation of Hillis (1996)for ldquobreaking up long branchesrdquo by using alarger taxonomic sampling was followed byGiribet and Ribera (1998) leading to the sameconclusions reached by Aguinaldo et al

Acoels (as well as in nematodes and otherputative ldquolong branchrdquo metazoans) tend toldquomigraterdquo to the base of the triploblastictaxa in 18S rDNA phylogenetic trees in-stead of grouping with the bulk of thePlathelminthes This could be their actualposition in the metazoan phylogenetic tree(eg Ruiz-Trillo et al 1999) However itcould also be a problem because of thebranch lengths present in the phylogenetictrees separating diploblastic from triploblas-tic animals In consequence the removal ofthis branch could be a third way of avoid-ing certain ldquolong branch attractionrdquo prob-lems If acoels were basal triploblastics and


the remaining plathelminths were derivedspiralians acoels should not group withthe remaining Plathelminthes On the otherhand if their basal position was an artifactwhen eliminating the diploblastic taxa fromthe analyses the acoel sequences shouldbe repositioned together with the otherplathelminths And this seems to be the casein the current analysis although we recog-nize that the position of this taxon is stronglyparameter dependent

In the analyses of Zrzavy et al (1998)molecular data were used to suggest that ne-mertodermatids (Nemertinoides) were relatedto other Plathelminthes whereas acoels werenot Our molecular tree (without variable re-gions) using parameter set 111 (Fig 2b) sug-gests that Nemertinoides and the acoels arederived Rhabditophora whereas Meara (theother nemertodermatid) is not This resultwas extremely unstable to parameter varia-tion For example parameter set 121 sug-gested that both nemertodermatids (Meara +Nemertinoides) were sister to catenulids andthat acoels were related to rhabditophoransparameter 141 suggested that Nemertinoideswas related to Urastoma and Meara was re-lated to the Acoela both clades unrelated tothe remaining platyzoans and so on Thuswe can be condent neither of the posi-tion of Acoela nor of Nemertodermatida Inthe absence of additional evidence we can-not consider Acoela Nemertodermatida orCatenulida as independent animal phyla assuggested by Zrzavy et al (1998) and we rec-ommend caution against erecting new taxo-nomic ranks that are based on poorly sup-ported analyses

Other Protostome Groups

Phoronozoa is another new phylum pro-posed by Zrzavy et al (1998) (= Brachio-zoa of Cavalier-Smith 1998) This was basedmainly on the paraphyletic status of Bra-chiopoda with respect toPhoronida in molec-ular analyses and stands in contrast to themorphological analyses of Carlson (1995)Zrzavy et al (1998) used sequences ofone phoronid (Phoronis vancouverensis) twoinarticulate brachiopods (Glottidia pyrami-data and Lingula lingua) and one articulatebrachiopod (Terebratalia transversa) In thepresent analysis we used two phoronid se-quences (P architecta and P australis) butexcluded the P vancouverensis sequence as

potentially being of doubtful origin For bra-chiopods we used the same sequences as inZrzavy et al (1998) plus a second articulatebrachiopod (Argyrotheca cordata) Our sam-pling suggested monophyly of Brachiopodamonophyly of Phoronida and monophylyof (Phoronida + Brachiopoda) This resultprompts us to consider Brachiopoda andPhoronida as monophyletic groups andwe thus recommend preserving the phy-lum status for both groups Phoronida andBrachiopoda (but see Cohen et al 1998Cohen 2000)

Synapomorphies supporting Trochozoaare (1) the presence of a haemal system (alsopresent in Deuterostomes Panarthropodsreduced in Nemertea) (2) the presence of res-piratory pigments (also present in Enterop-neusta Chordonia Priapulida and Nema-toda) and (3) the presence of a primary larva(also present in Enteropneusta Echinoder-mata Bryozoa Cycliophora and Polycla-dida reduced in Clitellata)

Synapomorphies for Ecdysozoa are (1) anabsence of epidermal ciliation (also absentin Chaetognatha and Acanthocephala) (2)the presence of two-layered cuticle (also inGastrotricha Entoprocta Sipunculida Mol-lusca Pogonophora and Annelida) (3)the presence of cuticular molting (also inClitellata) (4) the presence of ecdysone(known only in Nematoda Onychophoraand Arthropoda) and (5) nonciliated intesti-nal cells (also in Gnathostomulida)

Chaetognatha

The position of the phylum Chaetognathacontinues to be one of the most enigmatic is-sues in metazoan phylogeny The groupingof Chaetognatha with Nemertodermatida(bs = 19) in our favored tree (Figs 4 5) isdifcult to justify on the basis of morpholog-icalanatomical characters (see above) De-spite the high branch support this result isnot corroborated by any other parameter setThe placement of chaetognaths with nema-todes gnathostomulids and nematomorphs(Halanych 1996 Eernisse 1998 Littlewoodet al 1998) is also unstable Disregardingtheir ldquorelationshiprdquo with nemertodermatidschaetognaths appear between the proto-stomes and the deuterostomes a conictingphylogenetic position that has sustained con-siderable debate for decades (eg Hyman[1959] vs Meglitsch and Schram [1991])


CONCLUSIONS

Gnathostomulida appear to be a mem-ber of Gnathifera a group that mightalso include Cycliophora In our analysesGnathifera are more closely related to theother platyzoan acoelomates (Gastrotrichaand Plathelminthes) than to the introver-tan pseudocoelomates that form part of theEcdysozoa Platyzoa thus may be mono-phyletic (if the root is placed betweendeuterostome and protostome animals) andsister group to Trochozoa Platyzoa + Tro-chozoa would constitute the group namedSpiralia Other possibilities are that Platyzoais a sister group to the remaining triploblas-tic animals (acoelomates vs coelomates) orthat Platyzoa constitutes a grade of basaltriploblastic animals

Higher groupings found in our favoredtree were Cephalorhyncha (Nielsen 1995)(= Scalidophora [Lemburg 1995]) Ecdyso-zoa (Aguinaldo et al 1997) Gnathifera(Ahlrichs 1995 1997) Introverta (Nielsen1995) (= Cycloneuralia [Ahlrichs 1995])Nematoida (Schmidt-Rhaesa 1996) Platy-zoa (Cavalier-Smith 1998) and Syndermata(Ahlrichs 1995 1997 Zrzavy et al 1998)(= Trochata [Cavalier-Smith 1998]) Highergroups not supported by our best tree areAcanthognatha (Cavalier-Smith 1998) Cy-cloneuralia (Nielsen 1995) (= Nemathelmin-thes ss [Neuhaus 1994 Schmidt-Rhaesa1997]) Monokonta (Cavalier-Smith 1998)(= Neotrichozoa [Zrzavy et al 1998]) Ne-matozoa (Zrzavy et al 1998) Parenchymia(Nielsen 1995) and Plathelminthomorpha(Ax 1984)

Our analysis also disagrees with the phy-lum status of Catenulida (Zrzavy et al 1998)Acoela (Zrzavy et al 1998) and Phoronozoa(Zrzavy et al 1998) (= Brachiozoa [Cavalier-Smith 1998]) Relationships of Bryozoa Ne-mertodermatida and Chaetognatha to thefour main triploblastic groups (Deuterosto-mia Ecdysozoa Platyzoa and Trochozoa)were unresolved and to some extent they de-pend on the root position

Commonly Rotifera AcanthocephalaGastrotricha and Introverta are includedin the ldquoNemathelminthesrdquo (see Lorenzen1985 Malakhov 1980) Neuhaus (1994) ex-cluded Rotifera and Acanthocephala andnamed Nemathelminthes s str to the pu-tative clade containing the remaining taxaNone of those concepts (Nemathelminthesor Nemathelminthes s str) is supported byour data

The results shown here are certainly sub-ject to revision because sampling withinsome of the most diverse phyla and in par-ticular for certain aberrant taxa must be im-proved We also recognize that the additionof the diploblastic animals to the tree coulddetermine the position of certain groups oftriploblastics although the difculty in do-ing so based on 18S rDNA sequence dataaccording to previously published stud-ies made us investigate the new approachadopted here Despite the recognition ofthese problems our analyses seem novelin discerning among well-corroborated ver-sus unstable hypotheses of relationships Theplacement of certain taxa are stable andprobably denitely established through thecombination of molecular and morphologi-cal data such as the position of the phylumCycliophora Others such as ChaetognathaAcoela and Nemertodermatida (among oth-ers) are highly parameter-sensitive and in-ferences based on the currently available dataare at least poorly supported

NOTE ADDED IN PROOF

The sequence of Nemertinoides elongatuspublished in GenBank and used in this studyhas been demonstrated to be a sequence ar-tifact (Jaume Baguna pers comm)

ACKNOWLEDGMENTS

We thank Salvador Carranza Richard OlmsteadMark W Westneat and two anonymous reviewers fordiscussion and comments on earlier versions of thisarticle and Kimberlee Wollter for stylistic advice Wealso thank Steven Haddock Jerry Harasewych DanielMartotilden Cruz Palacotilden Lorenzo Prendini Jose IgnacioSaiz-Salinas and Xavier Turon for assistance with speci-mensand Kathy Coatesand HankTrapido-Rosenthal foreld work assistance GG was supported by a LernerGray Research Fellowship from the American Museumof Natural History and by a Grant-in-Aid Fellowshipaward from the BermudaBiological Station for ResearchThis article is BBSR contribution number 1555

REFERENCES

AGUINALDO A M A J M TURBEVILLE L SLINDFORD M C RIVERA J R GAREY R A RAFFAND J A LAKE 1997 Evidence for a clade of nema-todes arthropods and other moulting animals Nature387489ndash493

AHLRICHS W H 1995 Seison annulatus and SeisonnebaliaemdashUltrastruktur und Phylogenie Verh DtschZool Ges 88155

AHLRICHS W H 1997 Epidermal ultrastructure ofSeison nebaliae and Seison annulatus and a compar-ison of epidermal structures within the GnathiferaZoomorphology 117 41ndash48


AX P 1956 Die Gnathostomulida eine ratselhafteWurmgruppe aus dem Meeressand Akad Wiss LitMainz Abh Math-Naturwiss Kl Jg 81ndash32

AX P 1960 Die Entdeckung neuer Organisation-stypen im Tierreich Die neue BrehmBucherei Band258 Ziemsen Berlag Wittemberg

AX P 1963 Relationships and phylogeny of theTurbellaria Pages 191ndash224 in The lower Metazoa(E C Dougherty ed) Univ of California PressBerkeley

AX P 1984 Das Phylogenetische System Systema-tisierung der lebenden Natur aufgrund ihrer Phylo-genese G Fischer Stuttgart

AX P 1985 The position of Gnathostomulida andPlatyhelminthes in the phylogenetic system of the Bi-lateria Pages 168ndash180 in The origins and relation-ships of lower invertebrates (S Conway Morris J DGeorge R Gibson and H M Platt eds) Oxford UnivPress Oxford

AX P 1996 Multicellular animals A new approachto the phylogenetic order in nature Springer-VerlagBerlin

BACKELJAU T B WINNEPENNINCKX AND L DE BRUYN1993 Cladistic analysis of metazoan relationships Areappraisal Cladistics 9167ndash181

BOYER B C 1971 Regulative development in a spi-ralian embryo as shown by cell deletion experimentson the acoel Childia J Exp Zool 17697ndash105

BREMER K 1988 The limits of amino acid sequencedata in angiosperm phylogenetic reconstruction Evo-lution 42795ndash803

BRUSCA R C AND G J BRUSCA 1990 InvertebratesSinauer Sunderland Massachusetts

CARLSON S J 1995 Phylogenetic relationshipsamong extant brachiopods Cladistics 11131ndash197

CARRANZA S 1997 Taxonomia molecular mitjancantla sequenciacio del DNA ribosomic 18S Aplicacio alrsquoorigen i logenia dels Platihelmints PhD ThesisUniv Barcelona Barcelona

CARRANZA S J BAGUNA AND M RIUTORT 1997Are the Platyhelminthes a monophyletic primitivegroup An assessment using 18S rDNA sequencesMol Biol Evol 14485ndash497

CAVALIER-SMITH T 1998 A revised six-kingdomsystem of life Biol Rev 73203ndash266

COHEN B L 2000 Monophyly of brachiopods andphoronids Reconciliation of molecular evidencewith Linnaean classication (the subphylum Phoroni-formea nov) Proceedings of the Royal SocietyLondon Series B 267225ndash231

COHEN B L GAWTHROP A B AND CAVALIER-SMITHT 1998 Molecular phylogeny of brachiopods andphoronids based on nuclear-encoded small subunitribosomal RNA gene sequences Philosophical Trans-actions of the Royal Society B 3532039ndash2061

DARWIN C R 1844 Observations on the structureand propagation of the genus Sagitta Ann Mag NatHist Ser 1 131ndash6

EERNISS E D J 1998 Arthropod and annelid relation-ships re-examined Pages 43ndash56 in Arthropod rela-tionships (R A Fortey and R H Thomas eds) Chap-man amp Hall London

EERNISS E D J J S ALBERT AND F E ANDERSON1992 Annelida and Arthropoda are not sister taxaA phylogenetic analysis of spiralian metazoan mor-phology Syst Biol 41305ndash330

EHLERS U 1985 Das Phylogenetische System derPlathelminthes Gustav Fischer Stuttgart

EHLERS U 1986 Commentson a phylogenetic systemof the Plathelminthes Hydrobiologia 1321ndash12

EHLERS U 1995 The basic organization of the Platy-helminthes Hydrobiologia 30521ndash26

FUNCH P AND R M KRISTENSEN 1995 Cycliophorais a new phylum with afnities to Entoprocta and Ec-toprocta Nature 378711ndash714

FUNCH P AND R MKRISTENSEN 1997 CycliophoraPages 409ndash474 in Lophophorates Entoprocta andCycliophora (F W Harrison and R M Woollacotteds) Wiley-Liss New York

GAREY J R T J NEAR M R NONNEMACHER AND SA NADLER 1996 Molecular evidence for Acantho-cephala as a subtaxon of Rotifera J Mol Evol 43287ndash292

GHIRARDELLI E 1968 Some aspects of the biology ofthe chaetognaths Adv Mar Biol 6271ndash375

GHIS ELIN M T 1988 The origin of molluscs in thelightof molecular evidence Oxf Surv Evol Biol 566ndash95

G IRIBET G S CARRANZA M RIUTORT J BAGUNAAND C RIBERA 1999 Internal phylogeny of theChilopoda (Arthropoda lsquoMyriapodarsquo) A combinedapproach using complete 18S rDNA and partial 28SrDNA sequences Philos Trans R Soc London B354215ndash222

G IRIBET G AND C RIBERA 1998 The position ofarthropods in the animal kingdom A search for areliable outgroup for internal arthropod phylogenyMol Phylogenet Evol 9481ndash488

G IRIBET G AND W C WHEELER 1999 The positionof arthropods in the animal kingdom Ecdysozoa is-lands trees and the ldquoparsimony ratchetrdquo Mol Phy-logenet Evol 13619ndash623

GLADSTEIN D S AND W C WHEELER 1997 POYThe optimization of alignment characters AmericanMuseum of Natural History New York

GOLOBOFF P A 1998 NONA version 19 AmericanMuseum of Natural History New York

HALANYCH K M 1996 Testing hypotheses ofchaetognath origins Long branches revealed by 18Sribosomal DNA Syst Biol 45223ndash246

HALANYCH K M J D BACHELLER A M AGUINALDOS M LIVA D M HILLIS AND J A LAKE 1995 Ev-idence from 18S ribosomal DNA that the lophophor-ates are protostome Science 2671641ndash1643

HASZPRUNAR G 1996 Plathelminthes and Plathel-minthomorphamdashparaphyletic taxa J Zool Syst EvolRes 3441ndash48

HERLYN H AND U EHLERS 1997 Ultrastructure andfunction of the pharynx of Gnathostomula paradoxa(Gnathostomulida) Zoomorphology 117135ndash145

HILLIS D M 1996 Inferring complex phylogeniesNature 383130ndash131

HYMAN L H 1959 The invertebrates Smaller coelo-mate groups McGraw-Hill New York

KARLING T G 1974 On the anatomy and afnitiesof the turbellarian orders Pages 1ndash16 in Biology ofthe Turbellaria (N W Riser and M P Morse eds)McGraw-Hill New York

KATAYAMA T M NISHIOKA AND M YAMAMOTO1996 Phylogenetic relationships among turbellarianorders inferred from 18S rDNA sequences Zool Sci13747ndash756

KATAYAMA T H WADA H FURUYA N SATOHAND M YAMAMOTO 1995 Phylogenetic position ofthe dicyemid Mesozoa inferred from 18S rDNA se-quences Biol Bull 18981ndash90

KATAYAMA T AND M YAMAMOTO 1994 Phylo-genetic position of the catenulid atworm (Platy-helminthes Catenulida) inferred from the 18S ribo-somal DNA sequence Zool Sci 1136


KATAYAMA T M YAMAMOTO H WADA AND NSATOH 1993 Phylogenetic position of Acoel turbel-larians inferred from partial 18S rDNA sequencesZool Sci 10529ndash536

KRISTENSEN R M 1995 Are Aschelminthes pseu-docoelomate or acoelomate Pages 41ndash43 in Bodycavities Function and phylogeny (G LanzavecchiaR Valvassori and M D Candia-Carnevali eds)Mucchi Editore Modena Italy

LEMBURG C 1995 Ultrastructure of the introvert andassociated structures of the larvae of Halicryptus spin-ulosus (Priapulida) Zoomorphology 11511ndash29

LITTLEWOOD D T J K ROHDE AND K A CLOUGH1999 The interrelationships of all major groups ofPlatyhelminthes Phylogenetic evidence from mor-phology and molecules Biol J Linn Soc 6675ndash114

LITTLEWOOD D T M J TELFORD K A CLOUGHAND K ROHDE 1998 Gnathostomulidamdashan enig-matic metazoan phylum from both morphologicaland molecular perspectives Mol Phylogenet Evol972ndash79

LORENZEN S 1985 Phylogenetic aspects of pseudo-coelomate evolution Pages 210ndash223 in The originsand relationships of lower invertebrates (S ConwayMorris J D George R Gibson and H M Platt eds)Oxford Univ Press Oxford England

LUNDIN K 1997 Comparative ultrastructure of epi-dermal ciliary rootlets and associated structuresin species of the Nemertodermatida and Acoela(Plathelminthes) Zoomorphology 11781ndash92

LUNDIN K 1998 The epidermal ciliary rootlets ofXenoturbella bocki (Xenoturbellida) revisited Newsupport for a possible kinship with the Acoelomor-pha (Platyhelminthes) Zool Scr 27263ndash270

LUNDIN K AND J HENDELBERG 1996 Degenerat-ing epidermal bodies (lsquopulsatile bodiesrsquo) in Meara sti-chopi (Plathelminthes Nemertodermatida) Zoomor-phology 1161ndash5

MALAKHOV V V 1980 Cephalorhyncha a new typeof animal kingdom uniting Priapulida KinorhynchaGordiacea and a system of Aschelminthes wormsZool Zh 59485ndash499

MEGLITSCH P A AND F R SCHRAM 1991 Inverte-brate zoology Oxford Univ Press Oxford England

MICKEVICH M F AND J S FARRIS 1981 The impli-cations of congruence in Menidia Syst Zool 27143ndash158

NEUHAUS B 1994 Ultrastructure of alimentary canaland body cavity ground pattern and phylogenetic re-lationships of the Kinorhyncha Microfauna Mar 961ndash156

NEUHAUS B N P KRISTENSEN AND C LEMBURG1996 Ultrastructure of the cuticle of the Nemath-elminthes and electron microscopical localization ofchitin Verh Dtsch Zool Ges 89221

NIELSEN C 1995 Animal evolution interrelation-ships of the living phyla Oxford Univ Press OxfordEngland

NIELSEN C N SCHARFF AND D EIBYE-JACOBSEN1996 Cladistic analyses of the animal kingdom BiolJ Linn Soc 57385ndash410

NOREN M AND U JONDELIUS 1997 Xenoturbellarsquosmolluscan relatives [ and molluscan embryoge-nesis] Nature 39031ndash32

REUTER M O I RAIKOVA AND M K GUSTAFSSON1998 An endocrine brain The pattern of FMRF-amide immunoreactivity in Acoela (Plathelminthes)Tissue Cell 3057ndash63

RIEDL R J 1969 Gnathostomulida from AmericaScience 163445ndash452

RIEGER G E AND R M RIEGER 1977 Comparativene structure study of the gastrotrich cuticle and as-pects of cuticle evolution within the AschelminthesZ Zool Syst Evolutionsforsch 1581ndash124

RIEGER G E AND R M RIEGER 1980 Fine structureand formation of the gastrotrich eggshell Zoomor-phology 96215ndash229

RIEGER R M AND M MAINITZ 1977 Comparativene structural study of the body wall in Gnathos-tomulids and their phylogenetic position betweenPlatyhelminthes and Aschelminthes Z Zool SystEvolutionsforsch 159ndash35

RIEGER R M AND S TYLER 1995 Sister-grouprelationship of Gnathostomulida and Rotifera-Acanthocephala Invert Biol 114186ndash188

ROHDE K C HEFFORD J T ELLIS P R BAVERSTOCKA M JOHNSON N A WATSON AND S DITTMANN1993 Contributions to the phylogeny of Platy-helminthes based on partial sequencing of 18S riboso-mal DNA Int J Parasitol 23705ndash724

ROHDE K A M JOHNSON P R BAVERSTOCK ANDN A WATSON 1995 Aspects of the phylogenyof Platyhelminthes based on 18S ribosomal DNAand protonephridial ultrastructure Hydrobiologia30527ndash35

RUIZ-TRILLO I M RIUTORT D T LITTLEWOOD E AHERNIOU AND J BAGUNA 1999 Acoel atwormsEarliest extant bilaterian Metazoans not members ofPlatyhelminthes Science 2831919ndash1923

SCHMIDT-RHAESA A 1996 Ultrastructure of the an-terior end in three ontogenetic stages of Nectonemamunidae (Nematomorpha) Acta Zool 77267ndash278

SCHMIDT-RHAESA A 1997 Phylogenetic relation-ships of the Nematomorphamdasha discussion of currenthypotheses Zool Anz 236203ndash216

SCHRAM F R 1991 Cladistic analysis of metazoanphyla and the placement of fossil problematicaPages 35ndash46 in The early evolution of Metazoa andthe signicance of problematic taxa (A M Simonettaand S Conway Morris eds) Cambridge Univ PressCambridge England

SCHRAM F R AND W N ELLIS 1994 Metazoan rela-tionships A rebuttal Cladistics 10331ndash337

SMITH J P S S TYLER AND R M RIEGER 1986Is the Turbellaria polyphyletic Hydrobiologia 13213ndash21

STEINBOCK O 1930 Ergebnisse einer von EReisinger amp O Steinbock mit Hilfe des RaskO-rsted Fonds durchgefuhrten Reise in Gronland1926 2 Nemertoderma bathycola nov gen novspec eine eigenartige Turbellarie aus der Tiefe derDiskobay nebst einem Beitrag zur Kenntnis desNemertinenepithels Vidensk Medd Dan NaturhistForen 9047ndash84

STERRER W 1972 Systematics and evolution withinthe Gnathostomulida Syst Zool 21151ndash173

STERRER W 1998 New and known Nemertoder-matida (PlatyhelminthesndashAcoelomorpha)mdasha revi-sion Belg J Zool 12855ndash92

STERRER W M MAINITZ AND R M RIEGER 1985Gnathostomulida Enigmatic as ever Pages 181ndash199in The origins and relationships of lower invertebrates(S Conway Morris J D George R Gibson and H MPlatt eds) The Systematics Association and Claren-don Press Oxford England

STERRER W AND R M RIEGER 1974Retronectidaemdasha new cosmopolitan marine familyof Catenulida (Turbellaria) Pages 63ndash92 in Biologyof the Turbellaria (N W Riser and M P Morse eds)McGraw-Hill New York


TELFORD M J AND P W H HOLLAND 1993 Thephylogenetic afnities of the chaetognaths A molec-ular analysis Mol Biol Evol 10660ndash676

TYLER S AND R M RIEGER 1977 Ultrastructuralevidence for the systematic position of the Nemer-todermatida (Turbellaria) Acta Zool Fenn 154193ndash207

WADA H AND N SATOH 1994 Details of the evolu-tionary history from invertebrates to vertebrates asdeduced from the sequences of 18S rDNA Proc NatlAcad Sci USA 91 1801ndash1804

WALLACE R L C RICCI AND G MELONE 1995Through Alicersquos looking glass A cladistic analysisof pseudocoelomate anatomy Pages 61ndash67 in Bodycavities Function and phylogeny (G LanzavecchiaR Valvassori and M D Candia-Carnevali eds)Mucchi Editore Modena Italy

WALLACE R L C RICCI AND G MELONE 1996 Acladistic analysis of pseudocoelomate (aschelminth)morphology Invert Biol 115104ndash112

WHEELER W C 1990 Nucleic acid sequence phy-logeny and random outgroups Cladistics 6363ndash367

WHEELER W C 1995 Sequence alignment parameter

APPENDIX 1Taxon sampling used in the molecular analyses and GenBank accession codes Asterisks refer to the taxa

sequenced by the authors Taxonomy based in different sources (eg Brusca and Brusca 1990 Ax 1996)

Taxon sampled GenBank accession no

AnnelidandashPolychaeta (9 spp)Order Phyllodocida Nereis virens Z83754

Aphrodita aculeata Z83749Glycera americana U19519

Order Spionida Chaetopterus variopedatus U67324Order Capitellida Capitella capitata U67323Order Terebellida Lanice conchilega X79873Order Sabellida Sabella pavonina U67144

Protula sp U67142Order Dinophilida Dinophilus gyrociliatus AF119074 curren

AnnelidandashClitellata (4 spp)Order Prosothecata Enchytraeus sp Z83750Order Opisthopora Lumbricus rubellus Z83753Order Arhynchobdellida Hirudo medicinalis Z83752Order Rhynchobdellida Glossiphonia sp Z83751

Mollusca (10 spp)Class Caudofoveata Scutopus ventrolineatus X91977Class Polyplacophora Lepidopleurus cajetanus AF120502 curren

Acanthochitona crinita AF120503 curren

Class Gastropoda Diodora graeca AF120513 curren

Littorina obtusata X94274Siphonaria pectinata X94274

Class Scaphopoda Dentalium pilsbryi AF120522 curren

Rhabdus rectius AF120523 curren

Class Bivalvia Solemya velum AF120524 curren

Yoldia limatula AF120528 curren

Sipuncula (2 spp)Class Phascolosomida Aspidosiphon misakiensis AF119090 curren

Class Sipunculida Themiste alutacea AF119075 curren

Echiura (2 spp)Order Echiuroinea Ochetostoma erythrogrammon X79875Order Xenopneusta Urechis caupo AF119076 curren

Pogonophora (2 spp)Class Perviata Siboglinum ordicum X79876Class Obturata Ridgeia piscesae X79877

Nemertea (3 spp)Class Anopla Lineus sp X79878Class Enopla Prostoma eilhardi U29494 curren

Amphiporus sp AF119077 curren

sensitivity and the phylogenetic analysis of moleculardata Syst Biol 44321ndash331

WHEELER W C 1996 Optimization alignment Theend of multiple sequence alignment Cladistics 121ndash9

WHEELER W C AND C Y HAYASHI 1998 The phy-logeny of extant chelicerate orders Cladistics 14173ndash192

WINNEPENNINCKX B T BACKELJAU AND R MKRISTENSEN 1998 Relations of the new phylumCycliophora Nature 393636ndash638

WINNEPENNINCKX B T BACKELJAU L Y MACKEYJ M BROOKS R DE WACHTER S KUMAR ANDJ R GAREY 1995 18S rRNA data indicate that As-chelminthes are polyphyletic in origin and consist ofat least three distinct clades Mol Biol Evol 121132ndash1137

ZRZAVY J S MIHULKA P KEPKA A BEZDEK ANDD TIETZ 1998 Phylogeny of the Metazoa basedon morphological and 18S ribosomal DNA evidenceCladistics 14249ndash285

Received 19 March 1999 accepted 15 July 1999Associate Editor M Westneat


APPENDIX 1 Continued


Brachiopoda (4 spp)Class Inarticulata Glottidia pyramidata U12647

Lingula lingua X81631Class Articulata Terebratalia transversa U12650

Argyrotheca cordata AF119078 curren

Phoronida (2 spp) Phoronis architecta U36271Phoronis australis AF119079 curren

Bryozoa (4 spp)Class Stenolaemata Lichenopora sp AF119080 curren

Class Gymnolaemata Membranipora sp AF119081 curren

Caberea boryi AF119082 curren

Class Phylactolaemata Plumatella repens U12649Entoprocta (2 spp) Pedicellina cernua U36273

Barentsia hildegardae AJ001734Cycliophora (1 sp) Symbion pandora Y14811Rotifera (2 spp)

Class Bdelloidea Philodina acuticornis U41281Class Monogononta Brachionus plicatilis U29235

Acanthocephala (4 spp)Class Palaeoacanthocephala Plagiorhynchus cylindraceus AF001839

Echinorhynchus gadi U88335Class Archiacanthocephala Moniliformis moniliformis Z19562Class Eoacanthocephala Neoechinorhynchus pseudemydis U41400

Gastrotricha (2 spp)Order Chaetonotida Chaetonotus sp AJ001735

Lepidodermella squammata U29198Gnathostomulida (3 spp)

Order Filospermoidea Haplognathia sp AF119084 curren

Order Bursovaginoidea Gnathostomula paradoxa Z81325Gnathostomula sp AF119083 curren

PlathelminthesndashNemertodermatida (2 spp) Meara stichopi AF119085 curren

Nemertinoides elongatus U70084PlathelminthesndashCatenulida (2 spp) Stenostomum leucops U70085

Stenostomum sp U95947PlathelminthesndashAcoela (2 spp) Convoluta naikaiensis D83381

Amphiscolops sp D85099PlathelminthesndashRhabditophora (15 spp)

Order Macrostomida Macrostomum tuba U70081Microstomum lineare U70083

Order Polycladida Discocelis tigrina U70079Planocera multitentaculata D17562

Order Lecithoepitheliata Geocentrophora sp U70080Order Proseriata Archiloa rivularis U70077

Monocelis lineata U45961Order Rhabdocoela Bothromesostoma sp D85098

Mesocastrada sp U70082Order Prolecitophora Urastoma sp U70086Order Tricladida Dendrocoelum lacteum M58346

Ectoplana limuli D85088Cestoda Echinococcus granulosus U27015Trematoda Schistosoma mansoni M62652

Lobatostoma manteri L16911Priapulida (2 spp)

Priapulus caudatus D85088Tubiluchus corallicola AF119086 curren

Kinorhyncha (1 sp)Order Homalorhagida Pycnophyes kielensis U67997

Nematomorpha (2 spp)Class Gordioida Chordotes morgani AF036639

Gordius aquaticus X80233Nematoda (8 spp)

Order Araeolaimida Plectus aquatilis AF036602Order Desmodorida Desmodora ovigera Y16913Order Chromadorida Metachromadora sp AF036595Order Enoplida Enoplus brevis U88336




Order Trichocephalida Trichinella spiralis U60231Order Rhabditida Dolichorhabditis sp AF036591Order Tylenchida Globodera pallida AF036592Order Spirurida Dirolaria immitis AF036638

Onycophora (2 spp) Peripatopsis capensis AF119087 curren

Euperipatoides leukarti U49910Tardigrada (2 spp)

Class Eutardigrada Macrobiotus hufelandi X81442 curren

Milnesium tardigradum U49909Arthropoda (22 spp)

Class Pycnogonida Colossendeis sp AF005440 curren

Callipallene sp AF005439 curren

Class Chelicerata Limulus polyphemus U91490 curren

Belisarius xambeui U91491 curren

Liphistius bicoloripes AF007104 curren

Class Branchiopoda Branchinecta packardi L26512Lepidurus packardi L34048

Class Maxillopoda Argulus nobilis M27187Ulophysema oeresundense L26521Calanus pacicus L81939Porocephalus crotali M29931

Class Malacostraca Nebalia sp L81945Squilla empusa L81946

Class Myriapoda Thereuopoda clunifera AF119088 curren

Lithobius variegatus AF000773 curren

Polydesmus coriaceus AF005449 curren

Cylindroiulus punctatus AF005448 curren

Class Hexapoda Podura aquatica AF005452 curren

Dilta littoralis AF005457 curren

Lepisma sp AF005458 curren

Aeschna cyanea X89481Ephemera sp X89489

Enteropneusta (2 spp) Glossobalanus minutus AF119089 curren

Saccoglossus kowalevskii L28054Echinodermata (10 spp)

Class Crinoidea Antedon serrata D14357Endoxocrinus parrae Z80951

Class Holothuroidea Stichopus japonicus D14364Psychropotes longicauda Z80956

Class Echinoidea Echinus esculentus Z37125Brissopsis lyrifera Z37119

Class Ophiuroidea Amphipholis squamata X97156Astrobrachion constrictum Z80948

Class Asteroidea Asterias amurensis D14358Astropecten irregularis Z80949

Urochordata (3 spp)Class Apendicularia Oikopleura sp D14360Class Thaliacea Thalia democratica D14366Class Ascidiacea Styela plicata M97577

Cephalochordata (1 sp) Branchiostoma oridae M97571Craniata (11 spp)

Cephalaspidomorphi Lampetra aepyptera M97573Chondrichthyes Echinorhinus cookei M91181Actinopterygii Amia calva X98836

Polyodon spathula X98838Clupea harengus X98845

Coelacanthiformes Latimeria chalumnae L11288Amphibia Xenopus laevis X04025Testudines Trachemys scripta M59398Lepidosauria Heterodon platyrhinos M59392Archosauria Alligator mississippiensis M59383Eutheria Rattus norvegicus X01117

Chaetognatha (2 spp)Order Phragmophora Paraspadella gotoi D14362Order Aphragmophora Sagitta elegans Z19551


new phyla as well as other supra and inf-raphylum categories were erected and newnames introduced

The ultimate responsibility for the en-larged number of phylogenetic hypotheses ofrelationships among animal phyla lies withthe exponential growth of molecular dataand the renement of morphological andanatomical information through electron mi-croscopy The goal of this paper is to evalu-ate these hypotheses of relationship by usingnew molecular data in combination with themorphological data set published by Zrzavyet al (1998)

BACKGROUND

Cavalier-Smith (1998) in his six-kingdomsystem of life proposed dividing triploblas-tic animals into four infrakingdoms of pro-tostome animals (Lophozoa ChaetognathiEcdysozoa and Platyzoa) and into two in-frakingdoms of deuterostomes (Coelomo-pora and Chordonia) Lophozoa (previouslyreferred to as Eutrochozoa [Ghiselin 1988Eernisse et al 1992] or Lophotrochozoa[Halanych et al 1995]) included the classicalprotostome coelomates (Annelida EchiuraSipuncula Pogonophora Mollusca Ne-mertea) Entoprocta and Cycliophora plusthe ldquolophophoratesrdquo (Bryozoa Phoronidaand Brachiopoda) Zrzavy et al (1998) usedthe name Trochozoa for all Lophozoa ex-cept Brachiopoda and Phoronida Ecdysozoa(Aguinaldo et al 1997) included the molt-ing animals (Nematoda NematomorphaKinorhyncha Priapulida Loricifera Ony-chophora Tardigrada and Arthropoda) (seealso Giribet and Ribera 1998 Zrzavy et al1998) Chaetognathi include the enigmaticphylum Chaetognatha Perhaps the mostsurprising taxon in this new system of classi-cation is the so-called Platyzoa The compo-nents of this group were dened as ciliatednonsegmented acoelomates or pseudocoelo-mates that lack a vascular system and have astraight gut (when present) with or withoutanus Platyzoa thus include Rotifera Acan-thocephala Gastrotricha Gnathostomulidaand Plathelminthes (Cavalier-Smith 1998)

This system relocates many taxa far fromtheir previous positions in the classicationsFor example Platyzoa were divided into twophyla Acanthognatha (= Rotifera Acantho-cephala Gastrotricha and Gnathostomul-ida) and a monophyletic Plathelminthes

However Gastrotricha and Gnathostomul-ida are considered independent phyla bymost authors whereas Plathelminthes havebeen found to be nonmonophyletic in severalmorphological and molecular analyses

What Is the Evidence for the Monophylyof the Platyzoa

Winnepenninckx et al (1995) presentedan ldquoaschelminthrdquo phylogeny based on 18SrDNA sequences including platyzoan se-quences of Gastrotricha Rotifera Acantho-cephala and Rhabditophora which turnedout to constitute a monophyletic cladeCarranza et al (1997) also used 18S rDNAsequences of several platyzoan groups (Gas-trotricha Acanthocephala Acoela Nemerto-dermatida Catenulida and Rhabditophora)to address the question of Plathelminthesmonophyly The data suggested mono-phyly of Rhabditophora (including the Ne-mertodermatida) and nonmonophyly ofPlathelminthes (to the exclusion of Catenul-ida and Acoela)

In general molecular phylogenies haveproposed monophyly of Plathelminthes +Syndermata (eg Eernisse 1998 Littlewoodet al 1998) Plathelminthes + Gastrotricha +Syndermata (eg Winnepenninckx et al1995 Carranza et al 1997) Gnathostomul-ida + Gastrotricha (Zrzavy et al 1998)and Syndermata + Cycliophora (Winnepen-ninckx et al 1998) Together all these groupsexcept Cycliophora constitute the Platyzoasensu Cavalier-Smith (1998)

Different relationships among the platy-zoan taxa have been proposed by severalauthors on the basis of morphological data(eg Lorenzen 1985 Wallace et al 19951996 Neuhaus et al 1996 Ahlrichs 19951997 Nielsen 1995 Nielsen et al 1996Haszprunar 1996) a summary of these hy-potheses is shown in Figure 1 Meglitsch andSchram (1991) (see also Schram 1991) madethe rst attempt to analyze metazoan phy-logeny using parsimony algorithms Schramand Ellis (1994) reanalyzed an amendeddata set by using more-standard parsimo-ny procedures Their consensus cladogramcontained a basal clade of triploblastic ani-mals consisting of Gastrotricha RotiferaAcanthocephala Chaetognatha NematodaNematomorpha Kinorhyncha Priapulidaand Loricifera (= ldquoAschelminthesrdquo) Thisclade was followed by a clade containing













Cycliophora



Plathelminthes









Chaetognatha









THE NEW ANALYSIS















Taxon Sampling



DNA Sequences


Morphological Data




Bryozoa

Kamptozoa

Mollusca

Brachiozoa

Annelida

SipunculaNemertina


Lobopoda

Nemathelminthes

Acanthognatha

Platyhelminthes


UrochordataChordata

Zrzavy et al (1998)

Bryozoa


Mollusca

Phoronozoa

AnnelidaPogonophora

Echiura

SipunculaNemertea




Syndermata





Chordata

Xenoturbellida

Chordata

Ectoprocta

Entoprocta

Mollusca



Echiura

SipunculaNemertea


Pentastomida



PriapulaKinorhyncha

Loricifera



Platyhelminthes































1 1 1 11336 2185 13628 0007851 2 2 16892 4370 21563 0013961 4 4 27848 8740 36952 0009852 1 2 12799 4370 17329 0009232 2 4 19691 8740 28769 0011752 4 8 33013 17480 51312 0015964 1 4 15213 8740 24373 001723





RESULTS


Molecular Analysis











Combined Analysis






DISCUSSION

Platyzoa











FIGURE 4 Continued



























Chaetognatha



CONCLUSIONS






NOTE ADDED IN PROOF


ACKNOWLEDGMENTS


REFERENCES













































































































































































































Cycliophora



Plathelminthes









Chaetognatha









THE NEW ANALYSIS















Taxon Sampling



DNA Sequences


Morphological Data




Bryozoa

Kamptozoa

Mollusca

Brachiozoa

Annelida

SipunculaNemertina


Lobopoda

Nemathelminthes

Acanthognatha

Platyhelminthes


UrochordataChordata

Zrzavy et al (1998)

Bryozoa


Mollusca

Phoronozoa

AnnelidaPogonophora

Echiura

SipunculaNemertea




Syndermata





Chordata

Xenoturbellida

Chordata

Ectoprocta

Entoprocta

Mollusca



Echiura

SipunculaNemertea


Pentastomida



PriapulaKinorhyncha

Loricifera



Platyhelminthes































1 1 1 11336 2185 13628 0007851 2 2 16892 4370 21563 0013961 4 4 27848 8740 36952 0009852 1 2 12799 4370 17329 0009232 2 4 19691 8740 28769 0011752 4 8 33013 17480 51312 0015964 1 4 15213 8740 24373 001723





RESULTS


Molecular Analysis











Combined Analysis






DISCUSSION

Platyzoa











FIGURE 4 Continued



























Chaetognatha



CONCLUSIONS






NOTE ADDED IN PROOF


ACKNOWLEDGMENTS


REFERENCES







































































































































































































Chaetognatha









THE NEW ANALYSIS















Taxon Sampling



DNA Sequences


Morphological Data




Bryozoa

Kamptozoa

Mollusca

Brachiozoa

Annelida

SipunculaNemertina


Lobopoda

Nemathelminthes

Acanthognatha

Platyhelminthes


UrochordataChordata

Zrzavy et al (1998)

Bryozoa


Mollusca

Phoronozoa

AnnelidaPogonophora

Echiura

SipunculaNemertea




Syndermata





Chordata

Xenoturbellida

Chordata

Ectoprocta

Entoprocta

Mollusca



Echiura

SipunculaNemertea


Pentastomida



PriapulaKinorhyncha

Loricifera



Platyhelminthes































1 1 1 11336 2185 13628 0007851 2 2 16892 4370 21563 0013961 4 4 27848 8740 36952 0009852 1 2 12799 4370 17329 0009232 2 4 19691 8740 28769 0011752 4 8 33013 17480 51312 0015964 1 4 15213 8740 24373 001723





RESULTS


Molecular Analysis











Combined Analysis






DISCUSSION

Platyzoa











FIGURE 4 Continued



























Chaetognatha



CONCLUSIONS






NOTE ADDED IN PROOF


ACKNOWLEDGMENTS


REFERENCES








































































































































































































THE NEW ANALYSIS















Taxon Sampling



DNA Sequences


Morphological Data




Bryozoa

Kamptozoa

Mollusca

Brachiozoa

Annelida

SipunculaNemertina


Lobopoda

Nemathelminthes

Acanthognatha

Platyhelminthes


UrochordataChordata

Zrzavy et al (1998)

Bryozoa


Mollusca

Phoronozoa

AnnelidaPogonophora

Echiura

SipunculaNemertea




Syndermata





Chordata

Xenoturbellida

Chordata

Ectoprocta

Entoprocta

Mollusca



Echiura

SipunculaNemertea


Pentastomida



PriapulaKinorhyncha

Loricifera



Platyhelminthes































1 1 1 11336 2185 13628 0007851 2 2 16892 4370 21563 0013961 4 4 27848 8740 36952 0009852 1 2 12799 4370 17329 0009232 2 4 19691 8740 28769 0011752 4 8 33013 17480 51312 0015964 1 4 15213 8740 24373 001723





RESULTS


Molecular Analysis











Combined Analysis






DISCUSSION

Platyzoa











FIGURE 4 Continued



























Chaetognatha



CONCLUSIONS






NOTE ADDED IN PROOF


ACKNOWLEDGMENTS


REFERENCES











































































































































































































Taxon Sampling



DNA Sequences


Morphological Data




Bryozoa

Kamptozoa

Mollusca

Brachiozoa

Annelida

SipunculaNemertina


Lobopoda

Nemathelminthes

Acanthognatha

Platyhelminthes


UrochordataChordata

Zrzavy et al (1998)

Bryozoa


Mollusca

Phoronozoa

AnnelidaPogonophora

Echiura

SipunculaNemertea




Syndermata





Chordata

Xenoturbellida

Chordata

Ectoprocta

Entoprocta

Mollusca



Echiura

SipunculaNemertea


Pentastomida



PriapulaKinorhyncha

Loricifera



Platyhelminthes































1 1 1 11336 2185 13628 0007851 2 2 16892 4370 21563 0013961 4 4 27848 8740 36952 0009852 1 2 12799 4370 17329 0009232 2 4 19691 8740 28769 0011752 4 8 33013 17480 51312 0015964 1 4 15213 8740 24373 001723





RESULTS


Molecular Analysis











Combined Analysis






DISCUSSION

Platyzoa











FIGURE 4 Continued



























Chaetognatha



CONCLUSIONS






NOTE ADDED IN PROOF


ACKNOWLEDGMENTS


REFERENCES



































































































































































































Bryozoa

Kamptozoa

Mollusca

Brachiozoa

Annelida

SipunculaNemertina


Lobopoda

Nemathelminthes

Acanthognatha

Platyhelminthes


UrochordataChordata

Zrzavy et al (1998)

Bryozoa


Mollusca

Phoronozoa

AnnelidaPogonophora

Echiura

SipunculaNemertea




Syndermata





Chordata

Xenoturbellida

Chordata

Ectoprocta

Entoprocta

Mollusca



Echiura

SipunculaNemertea


Pentastomida



PriapulaKinorhyncha

Loricifera



Platyhelminthes































1 1 1 11336 2185 13628 0007851 2 2 16892 4370 21563 0013961 4 4 27848 8740 36952 0009852 1 2 12799 4370 17329 0009232 2 4 19691 8740 28769 0011752 4 8 33013 17480 51312 0015964 1 4 15213 8740 24373 001723





RESULTS


Molecular Analysis











Combined Analysis






DISCUSSION

Platyzoa











FIGURE 4 Continued



























Chaetognatha



CONCLUSIONS






NOTE ADDED IN PROOF


ACKNOWLEDGMENTS


REFERENCES























































































































































































































1 1 1 11336 2185 13628 0007851 2 2 16892 4370 21563 0013961 4 4 27848 8740 36952 0009852 1 2 12799 4370 17329 0009232 2 4 19691 8740 28769 0011752 4 8 33013 17480 51312 0015964 1 4 15213 8740 24373 001723





RESULTS


Molecular Analysis











Combined Analysis






DISCUSSION

Platyzoa











FIGURE 4 Continued



























Chaetognatha



CONCLUSIONS






NOTE ADDED IN PROOF


ACKNOWLEDGMENTS


REFERENCES










































































































































































































Combined Analysis






DISCUSSION

Platyzoa











FIGURE 4 Continued



























Chaetognatha



CONCLUSIONS






NOTE ADDED IN PROOF


ACKNOWLEDGMENTS


REFERENCES





































































































































































































DISCUSSION

Platyzoa











FIGURE 4 Continued



























Chaetognatha



CONCLUSIONS






NOTE ADDED IN PROOF


ACKNOWLEDGMENTS


REFERENCES




































































































































































































FIGURE 4 Continued



























Chaetognatha



CONCLUSIONS






NOTE ADDED IN PROOF


ACKNOWLEDGMENTS


REFERENCES



























































































































































































































Chaetognatha



CONCLUSIONS






NOTE ADDED IN PROOF


ACKNOWLEDGMENTS


REFERENCES


































































































































































































CONCLUSIONS






NOTE ADDED IN PROOF


ACKNOWLEDGMENTS


REFERENCES

































































































































































































klasifikasi zoin

Documents

18s rdna loci

18s ribosomal

ve hypervariable

epidermal

18s rdna sequences

direct optimization

bremer support

morphological