(PDF) Phylogeny of sipunculan worms: A combined analysis of four gene regions and morphology

Molecular Phylogenetics and Evolution 42 (2007) 171–192www.elsevier.com/locate/ympev

Phylogeny of sipunculan worms: A combined analysis of four gene regions and morphology

Anja Schulze a,¤,1, Edward B. Cutler a, Gonzalo Giribet a,b

a Museum of Comparative Zoology, Harvard University, 26 Oxford Street, Cambridge, MA 02138, USAb Department of Organismic and Evolutionary Biology, Harvard University, 16 Divinity Avenue, Cambridge, MA 02138, USA

Received 23 November 2005; revised 11 June 2006; accepted 16 June 2006Available online 5 July 2006

Abstract

The intra-phyletic relationships of sipunculan worms were analyzed based on DNA sequence data from four gene regions and 58 mor-phological characters. Initially we analyzed the data under direct optimization using parsimony as optimality criterion. An implied align-ment resulting from the direct optimization analysis was subsequently utilized to perform a Bayesian analysis with mixed models for thediVerent data partitions. For this we applied a doublet model for the stem regions of the 18S rRNA. Both analyses support monophyly ofSipuncula and most of the same clades within the phylum. The analyses diVer with respect to the relationships among the major groups butwhereas the deep nodes in the direct optimization analysis generally show low jackknife support, they are supported by 100% posteriorprobability in the Bayesian analysis. Direct optimization has been useful for handling sequences of unequal length and generating conser-vative phylogenetic hypotheses whereas the Bayesian analysis under mixed models provided high resolution in the basal nodes of the tree.© 2006 Elsevier Inc. All rights reserved.

Keywords: Sipuncula; Direct optimization; Bayesian; Doublet model

1. Introduction

Sipuncula (peanut worms or star worms) contain roughly150 exclusively marine species and are currently recognizedas a phylum. Sipunculan systematics is rooted in a long andconvoluted history, as outlined in Cutler (1994); Maxmenet al. (2003); Staton (2003) and Schulze et al. (2005). Today,annelid aYnities are suggested by both mitochondrial genearrangement data and DNA sequence analysis.

Using mitochondrial gene order data, both Boore andStaton (2002) and Bleidorn et al. (2005) place the sipunculanPhascolopsis gouldii in a clade with annelids. The latter Wndhigh support for a close relationship with the orbiniid poly-chaete Orbinia latreillii. However, taxon sampling for mito-

* Corresponding author. Present address: Smithsonian Marine Station,701 Seaway Drive, Fort Pierce, FL 34949, USA.

E-mail address: [emailprotected] (A. Schulze).1 Current address: Department of Marine Biology, Texas A&M Univer-

sity at Galveston, 5007 Avenue U, Galveston, TX 77551.

chondrial genomes of lophotrochozoans is still limited.Apart from P. gouldii, only annelid, mollusc and brachiopodmitochondrial genomes are considered in the analyses. Inaddition, only about 50% of the mitochondrial genome ofP. gouldii genome has been sequenced. Jennings and Hala-nych (2005) conclude that the mitochondrial gene order isvery conserved and of limited use for analyzing relationshipswithin annelids. On the other hand, the correspondence ofthe partial mitochondrial gene sequence of P. gouldii withthe highly conserved annelid sequence is noteworthy.

Several recent publications focusing on systematics ofannelids based on DNA sequence data included sipuncu-lans and unanimously placed them in the annelid ingroup,although the number of outgroups was sometimes limited(Brown et al., 1999) and the branch support was generallylow for deep nodes in the tree (Bleidorn et al., 2003a,b; Hallet al., 2004). Better resolution and support for deep rela-tionships was achieved by Telford et al. (2005) who pre-sented a phylogeny of the Bilateria, based on 72 complete18S ribosomal RNA sequences. The sequences were

mailto: [emailprotected]

172 A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192

analyzed under mixed models for the stem and loop regionsof the ribosomal molecule, applying a doublet model thataccounts for correlation of substitutions in the complemen-tary stem regions. The included sipunculan sequencegrouped with the seven annelids, as the sister taxon to asabellid. Peterson and Eernisse (2001) and Staton (2003)also concluded that Sipuncula were more closely related toannelids than to molluscs contradicting the hypothesis ofScheltema (1996). On the other hand, total evidence analy-ses of ribosomal genes and morphology have suggested thatsipunculans are the sister group to a clade including anne-lids and molluscs (Giribet et al., 2000). Our recent analysesof approximately 2400 base pairs from three gene regionsand morphology strongly support the monophyly of Sipun-cula but provide no resolution at a higher level (Maxmenet al., 2003; Schulze et al., 2005).

The fossil record for sipunculans is generally sparse butHuang et al. (2004) reported three exquisitely preserved fos-sil sipunculan species from the Lower Cambrian Maotian-shan Shale in southwest China. This Wnding places theorigin of the Sipuncula to more than 520 Myr ago despitethe fact that the group presents little morphological varia-tion and low species diversity. The aYnity of the fossils tomodern sipunculans was supported by the presence of aretractable introvert, a mouth surrounded by tentacles, acaudal appendage (sometimes present in modern GolWngii-dae), a U-shaped gut with an anus on the anterior trunkand hooks, wrinkles and papillae on the body, but interest-ingly, the gut lacks the typical helical morphology of themodern members of the group.

As in our previous analyses (Maxmen et al., 2003; Schu-lze et al., 2005), the focus of the present paper is the internalphylogeny of the Sipuncula. Since submission of our latestanalyses (Schulze et al., 2005), we have increased the num-ber of sipunculan species from 29 to 52, representing morethan one third of all known sipunculan species and repre-sentatives of all but one of the 17 currently recognized gen-era. The only missing one is Siphonomecus, a monotypicgenus only known from the southeastern United States.Siphonomecus multicinctus is rare to emerge from its deepsediment burrows and we have not yet been able to obtainmaterial appropriate for DNA analysis.

The phylogenetic hypotheses that we are generating willenable us to test the current taxonomy of the Sipuncula asproposed by Cutler and Gibbs (1985) and Gibbs and Cutler(1987) who classiWed the group into seventeen genera, sixfamilies, four orders and two classes. The two classes,Sipunculidea and Phascolosomatidea, are morphologicallydistinguished by their tentacle arrangement: in Sipunculi-dea the tentacles form a circle around a central mouth,whereas in Phascolosomatidea they are arranged in ahorseshoe shape around the nuchal organ. For the diagno-ses of genera, families and orders, see Gibbs and Cutler(1987) and references to taxonomic revisions therein. How-ever, the monophyly of both classes has been questioned inrecent phylogenetic analyses (Maxmen et al., 2003; Staton,2003; Schulze et al., 2005), and the new hypothesis regards

certain characters of Sipunculidea as plesiomorphies for thephylum.

The phylogeny proposed in this study will also serve as aframework to detect and interpret evolutionary trendswithin the phylum. In addition, by including multiple repre-sentatives of widespread species from diVerent geographiclocations whenever possible, we are performing initial testsfor the phylogenetic cohesiveness of species and the puta-tive existence of hidden species diversity. This will allow usin the future to choose appropriate species for populationgenetic and phylogeographic studies.

2. Materials and methods

2.1. Collections and taxon sampling

Upon collection in the Weld, all sipunculan specimenswere frozen or Wxed in 70–100% ethanol or in one case inisopropanol (Xenosiphon branchiatus) prior to DNA extrac-tion. Outgroup taxa, chosen among molluscs, annelids,entoprocts and nemerteans, were treated in the same fash-ion. Collection data are listed in Appendix A. In total, wepresent data for 99 sipunculan individuals belonging to 52recognized morphospecies.

2.2. Sequence generation

Total DNA was extracted from a piece of tissue, prefera-bly from the retractor muscles, using the DNeasy tissue kit(Qiagen), following the instructions of the manufacturer.The desired gene regions were ampliWed from the DNAtemplates using polymerase chain reaction (PCR) (seeTable 1 for primers). The complete 18S rRNA was ampli-Wed in three fragments, using the primers 1F-3R, 1F-4R or1F-5R for the Wrst, 3F-18Sbi for the second and 18Sa2.0-9Rfor the last fragment. Occasionally, the second fragmentwas split up into 3F-5R and 4F-18Sbi, 4F-7R or 4F-8R.The total length of the 18S rRNA gene is approximately1800 bp, resulting in an implied alignment (see below) of2053 bp. Three other gene regions (lengths excluding ampli-Wcation primers) were ampliWed as single fragments: the D3fragment of the nuclear 28S ribosomal RNA (ca. 310 bp;implied alignment 409 bp), the nuclear coding gene for his-tone H3 (327 bp) and the mitochondrial coding gene forcytochrome c oxidase subunit I (650 bp). PCRs were per-formed in 25 �l volume according to standard protocolswith the annealing temperature varying between 40 °C forcoding genes and up to 50 °C for ribosomal genes. PCRproducts were visualized in 1–1.5% agarose gels andcleaned using the GENECLEAN II kit (Bio 101) or theQIAquick PCR puriWcation kit (Qiagen). Sequence reac-tions were performed in 10 �l volume, using 1 �l (if cleanedwith GENECLEAN) or 5 �l (if cleaned with QIAquick) oftemplate, 1 �M of primer, 2�l of ABI BigDye™ Terminatorv3.0 (Applied Biosystems) and 2 �l of halfTERM Dye Ter-minator reagent (Genpak). Sequence reactions were per-formed with the same thermal cycler as for PCR following

A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192 173

ABI standard protocols. After cleanup of the sequencereactions using Edge Biosystems gel Wltration cartridges,the sequences were analyzed with an ABI PRISM® 3100Genetic Analyzer.

2.3. Analyses

Electrochromatograms were visualized in Sequencher™4.0. Forward and reverse fragments were assembled and, inthe case of 18S rRNA, several fragments joined into a sin-gle sequence. From each sequence, external primers werecropped and discarded. Sequences were subsequently editedin MacGDE (Smith et al., 1994; Linton, 2005) and/or Bio-Edit Sequence Alignment Editor (Hall, 1994). All newsequences were deposited in GenBank under accessionnumbers DQ299950 through DQ300172 (Table 2).

Phylogeny reconstruction followed two diVerentapproaches: (1) A direct optimization approach (Wheeler,1996), as implemented in the computer program POY(Wheeler et al., 2004, 2006), using parsimony as the opti-mality criterion. (2) A Bayesian approach with mixed mod-els estimated for each independent partition, asimplemented in the program MrBayes 3.1.1 (Ronquist andHuelsenbeck, 2003). For both types of analyses, all parti-tions were analyzed separately and in combination. Theparsimony jackknife tree of the morphological data alonewas calculated in PAUP* (SwoVord, 2003). We performed1000 jackknife replicates, using the heuristic search option.For each heuristic search, 100 replicates of random taxonaddition were performed with tree bisection and reconnec-tion as the branch-swapping algorithm. Branches with lessthan 50% jackknife support were discarded. Both morphol-

ogy-based trees were rooted with Sipunculus nudus, becausewe could not score other outgroups for the characters spe-ciWc to sipunculans.

The direct optimization approach allows the analysis ofsequences of unequal length without prior alignment. Thealignment and tree generation are performed simulta-neously in a dynamic programming environment by takinginto account the same parameters (e.g., for transversion-to-transition ratios and gap penalties) throughout the entireanalysis; this is known as “one-step phylogenetics”. Inaddition we performed a sensitivity analysis in which wetested multiple parameter sets (Giribet, 2003; Wheeler,1995). For each data partition, 12 parameter sets were ana-lyzed with the transversion-to-transition ratios of 1, 2, 4and 8 and indel-to-transversion ratios of 1, 2, and 4.

Tree searches were conducted in parallel (using PVM—Parallel Virtual Machine) on a cluster of 30 dual-processornodes (between 1 and 2.4 GHz) assembled at Harvard Uni-versity (darwin.oeb.harvard.edu). Commands for load bal-ancing of spawned jobs were in eVect to optimizeparallelization procedures (-parallel, -dpm, -jobspernode 2,-multirandom). Initially trees were built through a randomaddition sequence procedure (20 replicates) followed by acombination of branch-swapping steps (SPR “subtreepruning and regrafting” and TBR “tree bisection andreconnection”), and continuing with tree fusing (GoloboV,1999, 2002) in order to further improve tree length. WhileSPR and TBR allow branch rearrangement within a giventree, tree fusing allows exchanging of branches of identicalcomposition among diVerent trees, as in other simulatedevolutionary algorithms (Moilanen, 1999, 2001). Discrep-ancies between heuristic and actual tree length calculations

Table 1Primers used for PCR ampliWcation and cycle sequencing

Primer Sequence Reference

COILCO1490 5� GGT CAA CAA ATC ATA AAG ATA TTG G 3� Folmer et al. (1994)HCO2198 5� TAA ACT TCA GGG TGA CCA AAA AAT CA 3� Folmer et al. (1994)COI-7 5� ACN AAY CAY AAR GAY ATY GGN AC 3� Saito et al. (2000)COI-D 5� TCN GGR TGN CCR AAN ARY CAR AA 3� Saito et al. (2000)

H3H3aF 5� ATG GCT CGT ACC AAG CAG AC(ACG) GC 3� Colgan et al. (1998)H3aR 5� ATA TCC TT(AG) GGC AT(AG) AT(AG) GTG AC 3� Colgan et al. (1998)

28S rRNA28Sa 5� GAC CCG TCT TGA AAC ACG GA 3� Whiting et al. (1997)28Sb 5� TCG GAA GGA ACC AGC TAC TA 3� Whiting et al. (1997)

18S rRNA1F 5� TAC CTG GTT GAT CCT GCC AGT AG 3� Giribet et al. (1996)3R 5� AGG CTC CCT CTC CGG AAT CGA AC 3� Giribet et al. (1996)3F 5� GTT CGA TTC CGG AGA GGG A 3� Giribet et al. (1996)4R 5� GAA TTA CCG CGG CTG CTG G 3� Giribet et al. (1996)4F 5� CCA GCA GCC GCG CTA ATT C 3� Giribet et al. (1996)5R 5� CTT GGC AAA TGC TTT CGC 3� Giribet et al. (1996)7R 5� GCA TCA CAG ACC TGT TAT TGC 3� Giribet et al. (1996)8R 5� ACG GGC GGT GTG TAC 3� Giribet et al. (1996)9R 5� GAT CCT TCC GCA GGT TCA CCT AC 3� Giribet et al. (1996)18Sa2.0 5� ATG GTT GCA AAG CTG AAA C 3� Giribet et al. (1999)18Sbi 5� GAG TCT CGT TCG TTA TCG GA 3� Giribet et al. (1999)

174 A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192

Table 2Taxon sampling and GenBank accession numbers for each sequenced locus

Species MCZ Catalogue # 18S rRNA 28S rRNA Histone H3 COI

SipunculidaePhascolopsis gouldii DNA100199 AF123306 AF519272 AF519297 DQ300134Siphonosoma cumanense DNA100235 AF519241 AF519271 AF519296Siphonosoma cumanense DNA100464 DQ300001 DQ300088 DQ300155Siphonosoma cumanense DNA100622 AY326201 AY445139 AY326296 DQ300156Siphonosoma cumanense DNA100991 DQ300002 DQ300047 DQ300089 DQ300157Siphonosoma vastum DNA100625 DQ300003 AY445137 AY326297 DQ300158Sipunculus (S.) norvegicus DNA101069 DQ300004 DQ300090 DQ300159Sipunculus (S.) nudus DNA100234 DQ300005 DQ300160Sipunculus (S.) nudus DNA100245 AF519239 AF519269Sipunculus (S.) nudus DNA100246 AF519240 AF519270 AF519295 DQ300161Sipunculus (S.) nudus DNA100468 DQ300006 DQ300048 DQ300091 DQ300162Sipunculus (S.) nudus DNA100629 DQ300007 DQ300092 DQ300163Sipunculus (S.) nudus DNA100993 DQ300008 DQ300049 DQ300093 DQ300164Sipunculus (S.) phalloides DNA101337 DQ300009 DQ300094 DQ300165Sipunculus (S.) polymyotus DNA101121 DQ300010 DQ300095 DQ300166Xenosiphon branchiatus DNA101086 DQ300016 DQ300050 DQ300101 DQ300172

GolWngiidaeGolWngia elongata DNA100465 DQ299969 DQ300031 DQ300065 DQ300121GolWngia elongata DNA100466 AF519242 DQ300066 DQ300122GolWngia elongata DNA101003 DQ299970 DQ300123GolWngia elongata DNA101066 DQ299971 DQ300067 DQ300124GolWngia elongata DNA101081 DQ299972 DQ300068 DQ300125GolWngia margaritacea DNA100738 DQ299973 DQ300032 DQ300069 DQ300126GolWngia vulgaris DNA100207 AF519244 AF519273 DQ300127Nephasoma diaphanes DNA101072 DQ299975 DQ300071 DQ300128Nephasoma Xagriferum DNA100439 AF519243 AF519299Nephasoma Xagriferum DNA100440 DQ299976 DQ300033 DQ300072 DQ300129Nephasoma Xagriferum DNA101071 DQ299977 DQ300073 DQ300130Nephasoma pellucidum DNA101009 DQ299978 DQ300131Thysanocardia catherinae DNA101068 DQ300015 DQ300099Thysanocardia nigra DNA100606 AF519247 AF519274 DQ300100

ThemistidaeThemiste (T.) dyscrita DNA101095 DQ300011 DQ300167Themiste (T.) hennahi DNA100627 DQ300012 DQ300096 DQ300168Themiste (L.) lageniformis DNA100229 AF519249 AF519276 AF519302 DQ300169Themiste (L.) minor DNA100210 AF519250 F519277 AF519303Themiste (L.) minor DNA101083 DQ300013 DQ300097 DQ300170Themiste (T.) pyroides DNA101084 DQ300014 DQ300098 DQ300171

PhascolionidaeOnchnesoma steenstrupii DNA101080 DQ299979 DQ300034 DQ300074Phascolion (L.) cryptum DNA101007 DQ299980 DQ300035 DQ300075 DQ300132Phascolion (I.) gerardi DNA101002 DQ299981 DQ300076Phascolion (P.) psammophilum DNA101006 DQ299982 DQ300036 DQ300133Phascolion (P.) strombus DNA100101 AF519248 AF519275 AF519301Phascolion (P.) strombus DNA100739 DQ299983Phascolion (P.) strombus DNA101077 DQ299984 DQ300077

PhascolosomatidaeAntillesoma antillarum DNA100390 AF519259 AF519286 AF519311Antillesoma antillarum DNA100759 DQ299950Antillesoma antillarum DNA101008 DQ299951 DQ300051 DQ300102Apionsoma (A.) misakianum DNA100231 AF519260 AY445142Apionsoma (A.) misakianum DNA100737 DQ299952 DQ300017 DQ300052 DQ300103Apionsoma (A.) murinae DNA100446 DQ299953 DQ300018Apionsoma (E.) pectinatum DNA100624 AY326293 AY445142 AY326300 DQ300104Phascolosoma (P.) agassizii DNA101096 DQ299985 DQ300037 DQ300078 DQ300135Phascolosoma (P.) albolineatum DNA100396 AF519251 AF519278 DQ300136Phascolosoma (F.) capitatum DNA101070 DQ299986 DQ300079 DQ300137Phascolosoma (P.) granulatum DNA100201 AF519252 AF519279 AF519304 DQ300138Phascolosoma (P.) granulatum X79874Phascolosoma (P.) nigrescens DNA100621 AY326292 AY445140 AY326299 DQ300139

A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192 175

Table 2 (continued)

Species MCZ Catalogue # 18S rRNA 28S rRNA Histone H3 COI

Phascolosoma (P.) nigrescens DNA100736 DQ299987 DQ300038 DQ300080 DQ300140Phascolosoma (P.) nigrescens DNA100822 DQ299988 DQ300039 DQ300081 DQ300141Phascolosoma (P.) nigrescens DNA101010 DQ299989 DQ300040 DQ300142Phascolosoma (P.) nigrescens DNA101082 DQ299990 DQ300041 DQ300143Phascolosoma (P.) noduliferum DNA100208 AF519253 AF519280 AF519305 DQ300144Phascolosoma (P.) perlucens DNA100228 AF519254 AF519281 AF519306 DQ300145Phascolosoma (P.) perlucens DNA100233 DQ299991Phascolosoma (P.) perlucens DNA100395 DQ299992 DQ300082 DQ300146Phascolosoma (P.) perlucens DNA100748 DQ299993 DQ300042 DQ300147Phascolosoma (P.) perlucens DNA100819 DQ299994 DQ300148Phascolosoma (P.) perlucens DNA100829 DQ299995 DQ300043 DQ300083 DQ300149Phascolosoma (P.) scolops DNA100373 AF519255 AF519282 AF519309Phascolosoma (P.) scolops DNA100394 DQ299996 DQ300150Phascolosoma (P.) scolops DNA100735 DQ299997 DQ300084 DQ300151Phascolosoma (P.) scolops DNA100813 DQ299998 DQ300044 DQ300085 DQ300152Phascolosoma (P.) stephensoni DNA100203 DQ299999 DQ300045 DQ300086Phascolosoma (P.) stephensoni DNA100209 AF519257 AF519284 AF519307Phascolosoma (P.) stephensoni DNA100469 AF519256 AF519283 AF519310 DQ300153Phascolosoma (P.) stephensoni DNA100485 AF519258 AF519285 AF519308Phascolosoma (P.) turnerae DNA100230 DQ300000 DQ300046 DQ300087 DQ300154

AspidosiphonidaeAspidosiphon (A.) albus DNA101017 DQ299954 DQ300053 DQ300105Aspidosiphon (A.) albus DNA101336 DQ299955 DQ300054Aspidosiphon (A.) elegans DNA100977 DQ299956 DQ300019 DQ300055Aspidosiphon (A.) elegans DNA101016 DQ299957 DQ300020 DQ300056 DQ300106Aspidosiphon (P.) Wscheri DNA100620 AY326294 AY326301 DQ300107Aspidosiphon (P.) Wscheri DNA100981 DQ299958 DQ300021 DQ300108Aspidosiphon (A.) gosnoldi DNA101014 DQ299959 DQ300022 DQ300057 DQ300109Aspidosiphon (A.) gracilis schnehageni DNA101087 DQ299960 DQ300023 DQ300058 DQ300110Aspidosiphon (P.) laevis DNA100467 AF519261 DQ300024 DQ300059 DQ300111Aspidosiphon (P.) laevis DNA100992 DQ299961 DQ300112Aspidosiphon (A.) misakiensis DNA100205 AF119090 AF519288 AF519312Aspidosiphon (A.) muelleri DNA100206 DQ299962 DQ300025 DQ300060 DQ300113Aspidosiphon (P.) parvulus DNA100202 AF119075 DQ300026 DQ300061Aspidosiphon (P.) parvulus DNA100375 DQ299963 DQ300062 DQ300114Aspidosiphon (P.) parvulus DNA100982 DQ299964 DQ300027 DQ300063 DQ300115Aspidosiphon (P.) steenstrupii DNA100232 AF519262 AF519291 AF519315 DQ300116Aspidosiphon (P.) steenstrupii DNA100372 DQ299965 DQ300028 DQ300117Aspidosiphon (P.) steenstrupii DNA100391 DQ299966 DQ300029 DQ300118Aspidosiphon (P.) steenstrupii DNA100630 DQ299967 DQ300064 DQ300119Cloeosiphon aspergillus DNA100393 AF519263 AF519292 AF519316Cloeosiphon aspergillus DNA100825 DQ299968 DQ300030 DQ300120Lithacrosiphon cristatus DNA100623 AY326295 AY445142 AY326302Lithacrosiphon cristatus DNA100986 DQ299974 DQ300070

NemerteaAmphiporus sp. AF119077 AF519265 AF519293 AJ436899Argonemertes australiensis AF519235 AF519264 AF519293 AY428840

MolluscaLepidopleurus cajetanus AF120502 AF120565 AY070142 AF120626Rhabdus rectius AF120523 AF120580 AY070144 AY260826Yoldia limatula AF120528 AF120585 AY070149 AF120642

AnnelidaLumbrineris latreilli AF519238 AF519267 AF185253 AY364855Lamellibrachia spp. AF168742 AF185235 U74055Owenia fusiformis AF448160 AY428824 AY428832 AY428839Urechis caupo AF119076 AF519268 X58895 U74077Lumbricus terrestris AJ272183 — AF185262 NC001673

EntoproctaLoxosomella murmanica AY218100 AY218129 AY218150 AY218083

176 A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192

were addressed by adjusting slop values (-checkslop 10).While doing tree reWnements using TBR, -checkslop naccepts all trees that are within n tenths of a percent of thecurrent minimum value. For example, -checkslop 10accepts all trees up to 1% above the current minimumlength while doing TBR.

POY facilitates eYcient sensitivity analysis (Wheeler,1995; Giribet, 2003). All data sets (individual genes anddiVerent combinations of genes) were analyzed under 12parameter sets, for a range of indel-to-transversion ratiosand transversion-to-transition ratios (see Table 3). Impliedalignments—a topological-unique “alignment” or synapo-morphy scheme (Wheeler, 2003; Giribet, 2005)—can be eas-ily generated for each tree. Only when multiple partitionsare combined in a single analysis will reciprocally informeach other about the hom*ology statement—and this occursonly for those partitions that are not analyzed as prealigned(i.e., 18S rRNA and 28S rRNA).

To identify the optimal parameter set we employed acharacter-congruence technique which is a modiWcation ofthe ILD (Incongruence Length DiVerence) metric devel-oped by Mickevich and Farris (1981; see also Farris et al.1995), as proposed by Wheeler (1995) (Table 3). The valueis calculated for each parameter set by subtracting the sumof the scores of all partitions from the score of the com-bined analysis of all partitions, and normalizing it for thescore of the combined length. Although the reliability of theILD measure employed here has been questioned because itmay show a trivial minimum in circ*mstances in which par-titions are given disproportionate weights (Aagesen et al.,2005), this is not the case here. The ILD has been inter-

Table 3Tree lengths for the diVerent partitions analyzed (18S, 18S rRNA; 28S,28S rRNA; H3, histone H3; COI, cytochrome c oxidase subunit I; MOR,morphological data; MOL, 4 loci combined; TOT, morphology + 4 locicombined) and congruence value (ILD) for the combined analysis ofmorphology + 4 molecular loci combined at diVerent parameter sets (leftcolumn)

The Wrst numeral used in the parameter set column corresponds to theratio between indel-to-transversion and the following two numbers corre-spond to the ratio between transversion-to-transition; e.g., 111 is equalweights; 121 corresponds to a indel-to-transversion ratio of 1 and a trans-version-to-transition ratio of 2:1—so indels have a cost of 2, transversionshave a cost of 2 and transitions have a cost of 1 (for a list of the speciWcstep matrices that this involves see Giribet et al., 2002: App. 4). OptimalILD value is indicated in bold.

18S 28S H3 COI MOR MOL TOT ILD

111 3256 721 1191 5539 154 11662 11909 0.08800121 4878 1101 1712 8134 308 17256 17731 0.09012141 8033 1798 2712 13056 616 28052 29030 0.09697181 14229 3185 4688 22848 1232 49487 51480 0.10291211 3677 846 1191 5582 308 12333 12800 0.09344221 5627 1323 1712 8192 616 18428 19363 0.09776241 9482 2258 2712 13184 1232 30338 32227 0.10423281 17122 4099 4688 23127 2464 54117 57865 0.11000411 4311 1037 1191 5595 616 13297 14239 0.10457421 6868 1674 1712 8217 1232 20380 22207 0.11276441 11946 2917 2712 13227 2464 34202 38004 0.12467481 22061 5368 4688 23190 4928 61769 69254 0.13023

preted as a meta-optimality criterion for choosing theparameter set that best explains all partitions in combina-tion, the one that maximizes overall congruence and mini-mizes character conXict among all the data (Giribet, 2003).But this congruence maximization comes from the individ-ual partitions which are most congruent with the combinedone. This parameter set was given special consideration inthe analysis of data from each individual gene and isreferred to throughout this paper as the “optimal parame-ter set”. Additionally, we discuss results from the strict con-sensus of all parameter sets explored, which has beeninterpreted as a measure of stability to parameter choice, asapplied in statistical sensitivity analyses (Wheeler, 1995;Giribet, 2003). Nodal support for all topologies was mea-sured by parsimony jackkniWng (Farris, 1997; Farris et al.,1996).

In order to evaluate the potential eVect of treating eachgap as an independent character (e.g. Giribet and Wheeler,1999; Simmons and Ochoterena, 2000), we ran the com-bined analysis of all data using a non-linear (aYne) gapfunction, where the gap opening value was higher (2) thanthat of the gap extension (1). This was used for generatingthe implied alignment to be used in the subsequent Bayes-ian analysis.

We performed the Bayesian analysis under mixed mod-els with the complete dataset of the morphological data andall four gene regions, based on the implied alignment fromthe POY analyses. For this purpose, the 18S rRNAsequence was divided into two partitions corresponding tostem and loop regions. To identify secondary structure fea-tures, the annotated sequence for Phascolosoma granulatum(GenBank Accession No. X79874) was downloaded fromthe European Ribosomal Database (Van de Peer et al.,2000) and used as an annotation reference for the remain-ing taxa. A list of 388 nucleotide pairings was assembledmanually for the stem regions. The stem regions were ana-lyzed under a doublet model with a single rate parameterand sixteen states (Schöninger and von Haeseler, 1994),representing all possible nucleotide pairings. Currently, thedoublet model cannot be tested for goodness of Wt againstother models with common software programs, such asModeltest (Posada and Crandall, 1998) or MrModeltest(Nylander, 2004), but recent studies have shown that the16-state doublet model signiWcantly outperformed simpler,4-state, single-nucleotide models for stem regions of ribo-somal sequences (Savill et al., 2001; Telford et al., 2005). Nosecondary structure information was used for the fragmentof 28S rRNA because (1) no appropriate sipunculan refer-ence sequence was available, and (2) only ca. 10% of thetotal 28S rRNA gene was sequenced for this study, result-ing in few nucleotide pairings in this fragment.

All four gene regions were separately tested for theappropriateness of 24 diVerent models of nucleotide substi-tution in MrModeltest 2.2 (Nylander, 2004). The same testwas performed on the 18S rRNA loop regions only. Thedata sets were submitted to four hierarchical likelihoodratio tests (hLRT). In addition, the Akaike Information

A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192 177

Criterion (AIC) was calculated for each of the 24 models.For the 18S loop regions, one hLRT favored a general timereversible model but the remaining three hLRTs as well asthe AIC favored a symmetrical model, which assumes equalbase frequencies (Zarkihk, 1994). The symmetrical modelwas implemented with corrections for a discrete gammadistribution of substitution rates (four categories) (G) and aproportion of invariable sites (I). For the complete 18SrRNA and the other three gene regions, all hLRTs and theAIC favored the General Time Reversible (GTR) model(Tavaré, 1986) which was also implemented with correc-tions for a gamma distribution and a proportion of invari-able sites. The 58 morphological characters (Appendix B,C) were analyzed under the discrete likelihood model pro-posed by Lewis (2001), as implemented in MrBayes 3.1.This model is similar to a Jukes–Cantor model (Jukes andCantor, 1969), except that the number of states can varyfrom 2 to 10.

Phylogenetic analysis was performed using Bayesian sta-tistics with a Metropolis coupled Markov Chain MonteCarlo algorithm as implemented in MrBayes 3.1.1 (Ron-quist and Huelsenbeck, 2003). Two runs with four chainseach were performed simultaneously for 1,500,000 genera-tions, sampling trees every 500 generations. The tempera-ture parameter was set to 0.08. We chose a lowertemperature parameter than the default setting of 0.2,because the cold chain barely swapped among the fourchains of a run under the default setting. The value of 0.08was chosen after several initial trials under diVerent temper-ature parameters. It increased the number of cold chainswaps and stationarity was achieved sooner than withhigher parameter values. The initial 500,000 generationswere discarded as burn-in. To test for the eVect of the dou-blet model, the 18S rRNA partition was analyzed bothunder a single model (GTR+I+G) as well as under mixedmodels for stem and loop regions (doublet and symmetricalmodel, respectively).

3. Results

For the POY analysis, parameter set “111” which spec-iWes a ratio of 1:1 for both transversion-to-transition andindel-to-transversion ratios was determined to be the opti-mal parameter set, because it minimized overall incongru-ence as measured by the ILD test (Table 3). The combinedanalysis of all four gene regions and the morphologicaldataset resulted in 30 most parsimonious trees of length11,909. The strict consensus of all most parsimonioustrees for the optimal parameter set with jackknife supportvalues is shown in Fig. 1A. Fig. 1B illustrates the strictconsensus of all parameter sets for the combined analysisof all data. All parameter sets identify monophyly ofSipuncula, and a major split uniting the two sipunculi-dean genera Sipunculus and Xenosiphon (97% jackknifesupport; JF hereafter) as the sister group to all othersipunculans (90% JF). This clade shows paraphyly ofSiphonosoma and identiWes several clades, some well sup-

ported and stable to parameter variation, such as theclade including all Aspidosiphon + Lithacrosiphon or aclade including most species of Phascolosoma, except forP. capitatum and P. turnerae. Monophyly of the genusThemiste is also well supported. Finally, Phascolopsisgouldii and GolWngia vulgaris form a clade throughout theentire parameter space and have a JF of 97%, one of thehighest values in the analysis.

The Bayesian analysis of 18S rRNA under mixed modelsresulted in signiWcantly higher log likelihoods-lnL thanthose of the analysis under a single model. Among the 4000sampled trees (2000 from each of the two runs), the bestlikelihood-lnL score was ¡16988.8 (mean: ¡17037.7§16.5)under mixed models and ¡17761.00 (mean:¡17820.00§ 27.9) under a single model.

The 50% majority rule consensus tree resulting from thecombined Bayesian analysis is shown in Fig. 2. The treetopology is identical to the tree excluding morphologicaldata. The analysis resulted in Wve major clades of whichfour are supported by 100% posterior probability (pp here-after). The Wve clades are referred to as clade I through Vfor comparative purpose and are also indicated in Fig. 1A.Clade I is the sister group to all other sipunculans and con-tains all representatives of the genera Xenosiphon andSipunculus, with Xenosiphon branchiatus nested withinSipunculus. Clade II includes the two Siphonosoma species,with S. vastum as the sister group to the four representa-tives of S. cumanense. Clade III includes representatives ofnine diVerent genera. Of those, only Themiste and Thysano-cardia are monophyletic. Phascolion, Nephasoma and Gol-Wngia are clearly polyphyletic. The monophyly ofOnchnesoma could not be tested because only one species(O. steenstrupii) was included in the analysis. Phascolopsisgouldii represents a monotypic genus. Phascolosoma turne-rae and Apionsoma murinae are part of larger genera thatfall into other clades in the tree. Clade IV includes all shal-low-water Phascolosoma species. P. turnerae and P. capita-tum, both deep-sea inhabitants, are included in clades IIIand V, respectively. Clade V is only supported by 56% pp,and appears polyphyletic in the direct optimization analy-sis. It contains representatives of six genera. Phascolosomacapitatum and the two Apionsoma species form the basalbranches in the clade. The two representatives of Apion-soma misakianum form a clade with A. pectinatum,although with low pp (50%). The rest of the clade is wellresolved: with 100% pp, Cloeosiphon, Antillesoma, Lithacro-siphon and Aspidosiphon form a clade. The two representa-tives of Cloeosiphon aspergillus, three representatives ofAntillesoma antillarum and two representatives of Lithacro-siphon cristatus all form clades. Lithacrosiphon and Aspido-siphon form a highly supported clade (100% pp).Lithacrosiphon is nested within Aspidosiphon. Sipunculidea,as deWned in Gibbs and Cutler (1987), does not appear asmonophyletic. Clades IV + V basically correspond to thePhascolosomatidea, but Apionsoma murinae and Phascolo-soma turnerae, two species that morphologically fall intoPhascolosomatidea, appear in clade III.

178 A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192

A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192 179

Of the Wve clades as deWned above, clade I is well sup-ported in the direct optimization analysis under the optimalparameter set (97% JF) and stable across all parameters. Itstopology diVers from the topology of the Bayesian tree onlywith respect to Sipunculus nudus 100629 and Sipunculusnorvegicus 101069. Clade IV has 86% jackknife support andit is also found under all parameter sets, with the exceptionof one parameter set that resolves the clade under some ofthe most parsimonious trees, but not all. Its topology is sim-ilar to the topology in the Bayesian analysis: Phascolosomanoduliferum is the sister taxon the other Phascolosoma spe-cies which group into two clades. Clade III appears in theconsensus tree of the analysis under the optimal parameterset but has jackknife support below 50%. Furthermore, it isunstable to parameter set variation. However, there are sev-eral well-supported groups within the clade found in bothanalyses. Clades II and V are not supported in the directoptimization analysis, although there is high support for anAspidosiphon + Lithacrosiphon clade which is also stableunder diVerent parameter sets.

Of the 21 species of which more than one individualwas sequenced, eleven appeared as monophyletic in bothanalyses; one (Aspidosiphon steenstrupii) appeared asmonophyletic only in the direct optimization analysis(Table 4).

Fig. 3 shows a tree based on a combination of the par-simony jackknife tree and the 50% majority rule consen-sus tree resulting from the Bayesian analysis of themorphological dataset. Both trees diVer only in theirdegree of resolution. For the most part, the tree resultingfrom the Bayesian analysis shows higher resolution thanthe parsimony jackknife tree. However, the clade thatincludes all sipunculans except Sipunculus and Xenosiphonis weakly supported in the parsimony analysis only andremains unresolved in the Bayesian analysis. Of the Wvemajor clades identiWed in the Bayesian combined analysis(Fig. 2), only clade II also Wnds support in the morpholog-ical data with both types of analyses. The Bayesian analy-sis of the morphological data supports a monophyleticPhascolosomatidea.

4. Discussion

4.1. Phylogeny of the Sipuncula

Our analyses show that the large, sediment-burrowingspecies of the genera Sipunculus, Xenosiphon and Siphono-soma represent early branches in the sipunculan phyloge-netic tree (clades I and II). S. nudus has long been used as amodel sipunculan for physiological and biochemical stud-ies, however, Cutler (1994) pointed out that the species is in

Fig. 1. (A) strict consensus tree of 30 most parsimonious trees (length 11,909)optimal parameter set 111. Shaded boxes refer to clades deWned in the Bayesiwere monophyletic in the Bayesian analysis. Roman numerals correspond to cages (only >50%) at the nodes. (B) Strict consensus of most parsimonious treecies names indicate polyphyletic species.

many ways not a “typical” sipunculan because it diVersfrom the majority of species in embryology, structure of theventral nerve cord, coelomic urn cells, regeneration capabil-ities, osmoregulation and chromosome number. Because ofthese peculiarities, Cutler suspected that S. nudus wouldhave a highly derived position in the sipunculan tree. OurBayesian analysis shows that the Sipunculus/Xenosiphonclade is separated from the other sipunculans by longbranch lengths, reXecting its morphological distinctness,but instead of being highly derived in the tree, the clade isthe sister group to the remaining sipunculan species.

While the basal position of Sipunculus has been pro-posed previously (Maxmen et al., 2003; Schulze et al., 2005),our present study is the Wrst to suggest that Siphonosoma isthe next to branch oV in the sipunculan tree. This is notonly supported by the molecular data but also by the mor-phological data when analyzed under parsimony undrooted with Sipunculus nudus. In both previous studies,Siphonosoma appeared at the base of the Phascolosomati-dea clade. Siphonosoma shows many morphological simi-larities with Sipunculus and Xenosiphon, such as the largeelongated body with a short introvert, prominent papillaealong the introvert and circular and longitudinal muscula-ture split into bands. Among others, these similarities werethe basis for the deWnition of the family SipunculidaeRaWnesque, 1814, but they appear to be plesiomorphic forthe phylum. The family Sipunculidae is not supported inour analyses.

Clade III shows a high degree of morphological diver-sity. It contains representatives of nine diVerent genera,only two of which—Thysanocardia and Themiste—aremonophyletic according to our analyses. Both Thysanocar-dia and Themiste are morphologically clearly deWned bytheir tentacular crown. In Thysanocardia, the tentacles arevery numerous and extend some way along the introvert asfestoons (Cutler, 1994). Themiste is the only sipunculangenus with branched tentacles, presumably an adaptationfor Wlter feeding, whereas all other sipunculans have simpletentacles and seem to be primarily deposit feeders. We havenot been able to detect any morphological trends amongthe other members of clade III. The number of introvertretractor muscles varies between one and four; introverthooks may be absent or present; if present, they may bescattered or form rings.

Clade III basically corresponds to GolWngiiformes, exceptthat Phascolopsis gouldii, previously thought to belong toSipunculidae, and two species that morphologically fall intoPhascolosomatiformes are included. Of the latter, Apionsomamurinae, is a small deep-water species of which we only had asingle specimen. The specimen was identiWed by a sipunculanexpert (J.I. Saiz Salinas) but it is a small, thread-like species

resulting from the direct optimization analysis of all data in POY under thean analysis (Fig. 2). Dashed boxes refer to non-monophyletic groupings thatlade designations in Fig. 2. Branch support is indicated by jackknife percent-s found across all parameter sets analyzed. Squares and circles following spe-

180 A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192

A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192 181

for which identiWcation is diYcult. After Wxation in ethanoland use of most of the tissue for DNA extraction, the identi-Wcation can no longer be veriWed. This clearly demonstratesthe need to keep adequately preserved voucher material incollections whenever possible. Furthermore, only approxi-mately 1300bp of 18S and 28S rRNA were sequenced forthis species, potentially causing uncertainties in its placement.The second species that unexpectedly does not group withPhascolosomatidea is Phascolosoma turnerae. In both thedirect optimization and the Bayesian analysis, it splits oVearly in clade III. P. turnerae is a deep-water wood-dwellingspecies (Rice, 1985). There is no doubt about its identiWcationand the only gene region missing is COI. Although P. turne-rae is unusual with respect to its depth range and habitat—most other Phascolosoma species either inhabit crevices ofrocks or burrow into soft rocks in shallow water—it mor-phologically Wts the description of the genus. Currently, ourbest explanation for its unexpected placement in our trees isthat the morphological similarities between P. turnerae andother Phascolosoma species may be convergences, and thisdeserves further anatomical work.

Our placement of P. gouldii in clade III, and speciWcally asthe sister species to GolWngia vulgaris, conWrms the Wndings ofMaxmen et al. (2003) and Schulze et al. (2005). P. gouldii is aspecies with a confusing taxonomic history. Prior to Cutler

Fig. 2. Fifty percent majority rule consensus tree resulting from the Bayesiannumerals indicate major clades. Branch support is indicated as percent posterbility. Squares and circles following species names indicate polyphyletic specie

and Gibbs (1985) who placed it within Sipunculidae, it hadbeen associated with species that are now considered GolWn-giidae. Our Wndings therefore conWrm original ideas aboutthe evolutionary relationships of the species.

The division of the Sipuncula into the two majorgroups, Sipunculidea and Phascolosomatidea, is morpho-logically plausible. However, none of our analyses showedsupport for a monophyletic Sipunculidea, not even theanalyses based on morphological data alone. The Bayes-ian analyses show strong support for a Phascolosomati-dea clade, although the combined analysis excludesApionsoma murinae and Phascolosoma turnerae from thisclade, as discussed above. Other species with uncertainaYnities in both analyses are Phascolosoma (Fisherana)capitatum, Apionsoma (Edmondsius) pectinatum and Api-onsoma (Apionsoma) misakianum. In the Bayesian analy-sis, they form basal branches in clade V, but the branchsupport for such position is not signiWcant. In the directoptimization analysis, Phascolosoma capitatum is the sis-ter group to clade III. P. capitatum is morphologically dis-tinguished from other Phascolosoma species by having thebody wall musculature organized as a smooth sheet andnot in bands. The aYliation of this species has changedrepeatedly in the past: Stephen and Edmonds (1972)treated Fisherana as a distinct genus, Cutler (1979) associ-

analysis of four gene regions and morphology. Shaded boxes with romanior probability. Asterisks indicate branch support of 100% posterior proba-s.

Table 4Jackknife support or posterior probability (both in %) for species with multiple representatives in the two analyses of combined datasets presented in Figs.1 and 2

Numbers of individuals sequenced given in parentheses. NM, non-monophyletic.

Species Number of individuals

Direct optimization analysis Bayesian analysis

Branch support for monophyly (jackknife %)

Maximum branch support contradicting monophyly (jackknife %)

Branch support for monophyly (% posterior probability)

Maximum branch support contradicting monophyly (% posterior probability)

Antillesoma antillarum 3 58 100Apionsoma misakianum 2 52 100Aspidosiphon albus 2 51 99Aspidosiphon elegans 2 Nm <50 Nm 100Aspidosiphon Wscheri 2 81 100Aspidosiphon laevis 2 72 100Aspidosiphon parvulus 3 Nm 88 Nm 100Aspidosiphon steenstrupii 4 <50 Nm 54Cloeosiphon aspergillus 2 62 100GolWngia elongata 5 90 100Lithacrosiphon cristatus 2 98 100Nephasoma Xagriferum 3 78 100Phascolion strombus 3 <50 100Phascolosoma granulatum 2 Nm 54 Nm 99Phascolosoma nigrescens 5 Nm <50 Nm 99Phascolosoma perlucens 6 Nm <50 Nm <50Phascolosoma scolops 4 Nm 72 Nm 100Phascolosoma stephensoni 4 Nm <50 Nm 97Siphonosoma cumanense 4 100 100Sipunculus nudus 6 Nm <50 Nm 75Themiste minor 2 Nm <50 Nm 53

182 A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192

ated P. capitatum with GolWngia (in the subgenus Apion-soma) and with Apionsoma (Cutler and Cutler, 1987).Later, assuming that the state of the body wall muscula-ture was hom*oplastic, it was moved back into Phascolo-soma, as originally proposed. Apionsoma pectinatum andA. misakianum belong to two diVerent subgenera,Edmondsius and Apionsoma, respectively. While they forma weakly supported clade in the Bayesian analysis, there isno support for this clade in the direct optimization analy-sis. Similar to Phascolosoma capitatum, A. pectinatum haschanged its generic aYliations in the past; it was origi-nally described as a Phascolosoma by Keferstein (1867).

The two species Antillesoma antillarum (with three repre-sentatives) and Cloeosiphon aspergillus (with two representa-tives) clearly fall into clade V in the Bayesian analysis, andform a clade in the direct optimization analysis. The presenceof an anal shield has in the past justiWed the inclusion of Clo-eosiphon in the within Aspidosiphonidae (Cutler and Gibbs,1985), although the anal shield is constructed very diVerentlyin this genus than in Aspidosiphon and Lithacrosiphon. It isnot restricted to the dorsal side but extends all around thepre-anal trunk and is constructed of relatively large rhom-boid plates. Our Bayesian analysis conWrms that, despitethese morphological diVerences, Cloeosiphon is indeed closely

Fig. 3. Unrooted phylogenetic tree based on results of the analyses of morphological data alone. The parsimony and Bayesian analyses resulted in similartrees that only diVered in degree of resolution. Branch support values are indicated as Bayesian posterior probabilities and parsimony jackknife percent-ages (the latter marked with asterisks). Roman numerals correspond to clade designations in Fig. 2. The shaded box corresponds to clade II in the Bayes-ian analysis. Dashed boxes refer to non-monophyletic groupings that were monophyletic in the Bayesian analysis.

A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192 183

related to the Aspidosiphon/Lithacrosiphon clade, however,Antillesoma antillarum also falls into this clade. Antillesoma, amonotypic genus, seems to be morphologically more similarto Phascolosoma than to the Aspidosiphonidae. It is lackingan anal shield, the main deWning characteristic of the Aspid-osiphonidae. The placement of Antillesoma in the Aspidosi-phonidae has previously been suggested (Maxmen et al.,2003) and is gaining support with the addition of more data.The anal shield in Lithacrosiphon is distinctly cone-shapedbut is restricted to the dorsal side as in Aspidosiphon. Thereare no other clear distinctions between Lithacrosiphon andAspidosiphon and all our analyses agree that Lithacrosiphonis nested within Aspidosiphon. Within Aspidosiphon, there isno support for the monophyly of the three diVerent subgen-era Akrikos, Aspidosiphon and Paraspidosiphon.

4.2. Species

In our study, only slightly more than half of the specieswith multiple representatives appeared monophyletic (Table4). It can be argued that biological species are not necessarilymonophyletic. In some cases, species paraphyly may be dueto a lack of resolution (e.g. Themiste minor, Aspidosiphonsteenstrupii, and Phascolosoma perlucens). Species may alsoappear paraphyletic due to incomplete lineage sorting. This isto be expected when conserved genes are used to resolve rela-tionships among very closely related species. However, of thenon-monophyletic species, at least three (Aspidosiphon ele-gans, A. parvulus, and P. scolops) are clearly polyphyletic (seeFigs. 1 and 2). None of the partitions or the combined analy-sis supports their monophyly. In a fourth polyphyletic spe-cies, P. granulatum, we cannot exclude a faulty identiWcationin one of the two specimens, since the 18S rRNA sequenceX79874 was downloaded from GenBank and the identiWca-tion cannot be veriWed. All represented specimens from theother four species, however, are from our own collections.

Cryptic species are common in marine organisms (forreviews see Knowlton, 1993, 2000) and there is evidencefor cryptic speciation in the sipunculan Apionsoma misa-kianum (Staton and Rice, 1999). However, it is unusual toWnd two morphologically cryptic species that are not sis-ter species, as is the case in several of the sipunculan sam-ples examined. All three polyphyletic species aregeographically widespread and inhabit shallow-waterhard substrates, mostly coral rubble. In Aspidosiphonparvulus, the two specimens from the Gulf of Mexico fallinto a separate clade from the specimen from Belize. It ispossible that the specimen from Belize represents Paraspi-dosiphon spinososcutatus, a species known from severalCaribbean locations that Cutler (1973) synonymized withAspidosiphon parvulus, due to the species’ very similarmorphology and ecological habits. Paraspidosiphon is nota valid genus any longer; it has been reduced to a subge-nus of Aspidosiphon and includes all Aspidosphon speciesin which the longitudinal body wall musculature splitsinto bands. A. parvulus was originally described as havingsmooth body wall musculature but at closer inspection,

Cutler (1973) detected longitudinal muscle bands in thetype specimens and therefore placed the species in Parasp-idosiphon. Other characters that originally distinguishedthe two species are introvert length, hook morphologyand the pattern of furrows in the anal shield. According toCutler (1973) all of these characters show too much varia-tion within populations or even individuals—e.g. both sin-gle-pointed and double-pointed hooks occur in the samespecimens—to warrant the status as separate species.However, in light of the molecular evidence, it would beworthwhile to re-evaluate the validity of Aspidosiphon(Paraspidosiphon) spinososcutatus.

Aspidosiphon elegans is another species with many juniorsynonyms. Cutler (1994) described a remarkable range inintraspeciWc morphological variation, but we cannotexclude the possibility that some of this variation actuallyrepresents interspeciWc diVerences. One of its junior syn-onyms is Aspidosiphon brocki, the only example of an asex-ually reproducing sipunculan.

The morphology of the anal shield and introvert hooks arethe most important taxonomic characters in Aspidosiphon.The anal shield seems to function primarily as an operculumto seal the burrow after the introvert has been retracted intothe trunk. The hooks are probably used to scrape epifauna oVthe rock surface when foraging. As the functions of hooksand anal shields seem to be the same across the genus, mor-phological convergence may be more common among Aspi-dosiphon species than originally thought. Another possibleexplanation for morphological convergence may be hybrid-ization. There is currently no evidence for hybridization insipunculans, but the possibility cannot be ruled out.

Similar to the polyphyletic Aspidosiphon species, Phascolo-soma scolops has a number of junior synonyms. In our analy-ses, three of the four P. scolops specimens, from the Red Seaand the Indian Ocean, group relatively closely together in apoorly resolved clade (unresolved in the direct optimizationanalysis) with P. perlucens and P. albolineatum, but thefourth specimen from Hawaii clearly falls into a separateclade, together with P. stephensoni and P. albolineatum.

Given that most of the non-monophyletic species and allof the polyphyletic species fall into Aspidosiphon and Phas-colosoma, we are currently focusing on the taxonomy ofthese two genera, using detailed morphological as well asmolecular techniques.

4.3. Methodology

Generally, alignment precedes tree-building as a separatestep in the analysis of sequence data. Direct optimization, asimplemented in POY can avoid some of the problems associ-ated with the analysis of sequences of unequal length bycombining the alignment and tree building steps in a single,dynamic process. Alignment and tree building are performedunder the same analytical parameters, and both are opti-mized according to the same optimality criterion. In ouranalyses, the optimality criterion was parsimony, althoughoptimization in a maximum likelihood or Bayesian frame-

184 A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192

work is also possible (Fleissner et al., 2005; Redelings andSuchard, 2005; Wheeler, 2006). POY also allowed us to per-form a sensitivity analysis to test the stability of clades underdiVerent parameters. On the other hand, when data are com-bined, POY does not allow for implementation of mixedmodels for the diVerent partitions, except between molecularand morphological data. As its goal is to minimize overalltree length (under parsimony as the optimality criterion), alldata are simultaneously submitted to an analysis under theoverall optimal parameter set. However, it does not accountfor the possibility that diVerent gene regions might evolveaccording to diVerent models. This was therefore addressedin the Bayesian analysis under mixed models.

A doublet model (Schöninger and von Haeseler, 1994)accounts for correlated substitutions in the complementarystrands of the stem regions in ribosomal molecules. Telfordet al. (2005) used a novel permutation approach on bilaterian18S rRNA and found that the likelihood of their treesimproved signiWcantly when the nucleotides in the stemregions were correctly paired and allowed to evolve under adoublet model as compared to the permutated data sets withunpaired nucleotides and a single model for the completemolecule. Their analysis under mixed models was clearlysuperior to simpler analyses that did not take into accountthe secondary structure. Similarly, the likelihoods for our 18SrRNA trees also signiWcantly improved when mixed modelsfor stem and loop regions were implemented. We were onlyable to apply the doublet model to the 18S rRNA stemregions. Even better resolution might be achieved if the full28S rRNA sequences (and possibly mitochondrial ribosomalsequences) were sequenced and models of secondary struc-ture could be applied for those. Bayesian analysis can overes-timate branch support (e.g., Simmons et al., 2004); however,almost all of our basal nodes in the Bayesian analysis havemaximum branch support and long branches separate mostof our major clades.

4.4. Conclusions

Our analysis covers almost the entire diversity within thephylum and is by far the most comprehensive analysis ofsipunculan phylogeny published to date. Yet some questionsremain open, such as the phylogenetic aYnities of some“stragglers”, i.e. species that appear in diVerent parts of thetree depending on the method or parameters of analysis.

Both our analyses strongly support the monophyly ofSipuncula and most of the same clades within Sipuncula. Inparticular, both agree that the Sipunculus/Xenosiphon cladeis the sister group to all other sipunculans. This had beensuggested in our previous analyses (Maxmen et al., 2003;Schulze et al., 2005) and has now gained further corrobora-tion. Maxmen et al. (2003) also showed that the rootingbetween Sipunculus and the remaining sipunculans is notdependent on outgroup choice.

Direct optimization with POY has been a useful approachto handle the ribosomal sequences of unequal length andgenerating conservative estimates of phylogeny. The Bayes-

ian analysis under mixed models provides high resolutionwith maximum branch support in the deep nodes of the tree.

Our study complements previous taxonomic work onsipunculans, such as Cutler’s (1994) monograph and hisprevious generic revisions. Morphological data alone hadreached their limitations due to the simple morphologies ofthese ancient worms. Molecular analyses have added a newperspective to sipunculan phylogeny. However, our studyalso pinpoints some taxonomic problems, especially in thegenera Phascolosoma and Aspidosiphon. We are currentlyaddressing these issues using detailed morphological andmolecular methods.

Acknowledgments

We thank all individuals who contributed to the collectionof specimens, as listed in Appendix A, without whom thisstudy would have been impossible. We are especially gratefulto Harlan Dean and Iñaki Saiz Salinas who have helped us inmany ways through the years. This project was fundedthrough a MarCraig Grant at Harvard University to G.G.and E.B.C., which included postdoctoral support to A.S. Itwas continued under a postdoctoral fellowship to A.S. at theSmithsonian Marine Station at Fort Pierce, FL (Contribu-tion No. 658). Additional funding was provided by theCaribbean Coral Reef Foundation (Contribution No. 758)and by internal funds from Harvard University and theMuseum of Comparative Zoology to G.G. We owe thanks tothe captain and the crew of the R/V Oceanus andchief scientist K. Halanych (NSF EAR-0120646). We grate-fully acknowledge Mary E. Rice, Cheryl Hayashi, ChristophBleidorn and an anonymous reviewer for comments and sug-gestions that greatly improved this manuscript.

Appendix A. Collection data for specimens used in this study, in the following format: MCZ DNA voucher number—collecting location, collection date (collector)

Antillesoma antillarum (Grübe & Oersted, 1858):DNA100390—phu*ket, Thailand, Jan. 31, 2001 (J. Hylle-berg); DNA100759—Six Men’s Bay, Barbados, June 27,2002 (A. Schulze, J. I. Saiz-Salinas, E. B. Cutler);DNA101008—Bessie Cove, South Hutchinson Island, FL,USA, March 20, 2003 (A. Schulze).

Apionsoma (Apionsoma) misakianum (Ikeda, 1904):DNA100231—Pickles Reef, Key Largo, USA, Nov. 27,1993 (S. Taylor); DNA100737—Eilat, Israel, Sept, 30, 2000(N. Ben-Eliahu).

Apionsoma (Apionsoma) murinae(E. Cutler, 1969):DNA100446—Meteor station Me48/1#AT339, Antarc-tica, Nov. 30, 1999 (J.I. Saiz Salinas).

Apionsoma (Edmondsius) pectinatum (Keferstein, 1867):DNA100624—Six Mens Bay, Barbados, June 27, 2002 (A.Schulze, J.I. Saiz Salinas, E. B. Cutler).

Aspidosiphon (Akrikos) albus Murina, 1967:DNA101017—R/V Sunburst, cruise 521, Capron Shoals,FL, USA, March 18, 2003 (A. Schulze, W. Lee, H. Reic-

A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192 185

hardt); DNA101336—Capron Shoals, FL, USA, dateunknown (A. Schulze, W. Lee, H., H. Reichardt).

Aspidosiphon (Aspidosiphon) elegans (Chamisso &Eysenhardt, 1821): DNA100977—Carrie Bow Cay, Belize,April 17, 2003 (A. Schulze, M. E. Rice); DNA101016—R/VSunburst cruise 520, Capron Shoals, FL, USA, March 11,2003 (A. Schulze).

Aspidosiphon (Aspidosiphon) gosnoldi (E. Cutler, 1981):DNA101014—R/V Sunburst cruise 521, Capron Shoals,FL, USA, March 18, 2003 (A. Schulze).

Aspidosiphon (Aspidosiphon) gracilis schnehageni (W.Fischer, 1913): DNA101087—Punta Moralia, Costa Rica,Aug. 27, 2003 (H. K. Dean, J. A. Vargas).

Aspidosiphon (Aspidosiphon) misakiensis Ikeda, 1904:DNA100205—Cova Blava, Cabrera, Balearic Islands,Spain, May 31, 1997 (X. Turon).

Aspidosiphon (Aspidosiphon) muelleri Diesing, 1851:DNA100206—Banyuls-sur-Mer, France, July 19, 2000 (G.Giribet).

Aspidosiphon (Paraspidosiphon) Wscheri ten Broeke,1925: DNA 100620—Martin’s Bay, Barbados, June 21,2002 (A. Schulze, J. I. Saiz Salinas, E. B. Cutler, G. Y.Kawauchi); DNA100981—Twin Cays, Belize, April 24,2003 (M. E. Rice, A. Schulze).

Aspidosiphon (Paraspidosiphon) laevis de Quatrefa*ges,1865: DNA100467—Hungary Bay, Bermuda, Aug. 9, 2001(T. Nishikawa); DNA100992—Twin Cays, Belize, April 20,2003 (M. E. Rice, A. Schulze).

Aspidosiphon (Paraspidosiphon) parvulus Gerould, 1913:DNA100202—unspeciWed locality, purchased from GulfSpecimens Co. [this specimen corresponds to the sequenceerroneously published as Themiste alutacea by Giribet et al.(2000)]; DNA100375—unspeciWed locality, purchased fromGulf Specimens Co.; DNA100982—Twin Cays, Belize,April 20, 2003 (M. E. Rice, A. Schulze).

Aspidosiphon (Paraspidosiphon) steenstrupii Diesing,1859: DNA100232—Pickles Reef, Key Largo, USA, Nov. 27,1993 (S. Taylor); DNA100372—Kewalo Reef, Honolulu,USA, Jan. 25, 2001 (J. Brock); DNA100391—phu*ket, Thai-land, Jan. 31, 2001 (J. Hylleberg); DNA100630—Bank Reef,Barbados, June 26, 2002 (J. I. Saiz Salinas, A. Schulze).

Cloeosiphon aspergillus (de Quatrefa*ges, 1865):DNA100393—phu*ket, Thailand, Jan. 21, 2001 (J. Hylle-berg); DNA100825—Perrier’s Rock, South Africa, Oct. 18,2002 (R. Biseswar).

GolWngia elongata (Keferstein, 1862): DNA100465—SW ofTrunk Is., Harrington Sound, Bermuda, Aug. 8, 2001 (T. Nishik-awa); DNA100466—South coast of Stock’s Harbor, St. DavisIsland, Bermuda, Aug. 7, 2001 (T. Nishikawa); DNA101003—Twin Cays, Belize, April 20, 2003 (M. E. Rice, A. Schulze);DNA101066—R/V Oceanus, Southern New England, USA,40°27.299�N, 69°54.601�W, June 11, 2003 (A. Schulze);DNA101081—R/V Oceanus, Southern New England, USA,40°27.299�N, 69°54.601�W, June 11, 2003 (A. Schulze).

GolWngia margaritacea (Sars, 1851): DNA100738—Kongsfjord Svalbard, Norway, June 23, 2002 (D.Hughes).

GolWngia vulgaris (de Blainville, 1827): DNA100207—Banyuls sur Mer, France, July 19, 2000 (G. Giribet).

Lithacrosiphon cristatus (Sluiter, 1902): DNA100623—Bank Reef, Barbados, June 26, 2002 (A. Schulze, J.I. SaizSalinas); DNA100986—Carrie Bow Cay, Belize, April 17,2003 (M. E. Rice, A. Schulze).

Nephasoma diaphanes (Gerould, 1913): DNA101072—R/V Oceanus, Southern New England 40°20.410�N,70°46.765�W, June 11, 2003 (A. Schulze).

Nephasoma Xagriferum (Selenka, 1885): DNA100439—Meteor Station Me48/1#345/7, Antarctica, Nov. 11, 1999(J.I. Saiz Salinas); DNA100440—Meteor Station Me48/1#349, Antarctica, Nov. 11, 1999 (J.I. Saiz Salinas);DNA101071—R/V Oceanus, Southern New England, USA,39°47.230�N, 70°46.295�W, June 14, 2003 (A. Schulze).

Nephasoma pellucidum (Keferstein, 1865):DNA101009—R/V Sunburst, cruise 526, 4 miles oV Ft.Pierce, FL, USA, March 28, 2003 (A. Schulze).

Onchnesoma steenstrupii Koren & Danielssen, 1875:DNA101080—R/V Oceanus, Southern New England, USA,39°56.172�N, 69°34.563�W, June 15, 2003 (A. Schulze).

Phascolion (Isomya) gerardi Rice, 1993: DNA101002—Pinnacles between Sand bores, south of Carrie Bow Cayand Curlew Bank, Belize, April 21, 2003 (M. E. Rice, A.Schulze).

Phascolion (Lesenka) cryptum Hendrix, 1975:DNA101007—Harbor Branch Oceanographic Institution,Indian River Lagoon, Ft. Pierce, FL, USA, March 9, 2003(A. Schulze).

Phascolion (Phascolion) psammophilum Rice, 1993:DNA101006—R/V Sunburst cruise 523, Capron Shoals,FL, March 18, 2003 (A. Schulze).

Phascolion (Phascolion) strombus (Montagu, 1804):DNA100101—Banyuls sur Mer, France, July 20, 2000(G. Giribet); DNA100739—Kristineberg Marine BiologicalStation, Fiskebäckskil, Sweden, Dec. 31, 1997 (A. Okusu);DNA101077—R/V Oceanus, Southern New England, USA,39°47.230�N, 70°46.295�W, June 14, 2003 (A. Schulze).

Phascolopsis gouldii (Portalés, 1851): DNA100199—Woods Hole, USA, Sept. 30, 1997 (Marine BiologicalLaboratory).

Phascolosoma (Fisherana) capitatum (Gerould, 1913):DNA101070—R/V Oceanus, Southern New England, USA,39°47.230�N, 70°46.295�W, June 14, 2003 (A. Schulze).

Phascolosoma (Phascolosoma) agassizii Keferstein,1866: DNA101096—Cape Arago (North Cove), OR, USA,Aug. 28, 2003 (M. E. Rice, S. Rumrill, C. Young).

Phascolosoma (Phascolosoma) albolineatum (Baird,1868): DNA100396—phu*ket, Thailand, Jan. 31, 2001 (J.Hylleberg).

Phascolosoma (Phascolosoma) granulatum Leuckart,1828: DNA100201—Blanes, Girona, Catalonia, Spain,Aug. 12, 1997 (G. Giribet, C. Palacín).

Phascolosoma (Phascolosoma) nigrescens (Keferstein,1865): DNA100621—Six Mens Bay, Barbados, June 27,2002 (A. Schulze, J.I. Saiz Salinas); DNA100736—Eilat,Israel, Sept. 30, 2002 (N. Ben-Eliahu); DNA100822—Per-

186 A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192

rier’s Rock, South Africa, Oct. 18, 2002 (R. Biseswar);DNA101010—Bessie Cove, FL, USA, March 20, 2003 (A.Schulze); DNA101082—Broome, Australia, dateunknown (G. Rouse).

Phascolosoma (Phascolosoma) noduliferum Stimpson,1855: DNA100208—Nielsen Park Shore, Port Jackson,Sydney Harbor, Australia, April 12, 2000 (G. Giribet, P.Hutchings).

Phascolosoma (Phascolosoma) perlucens Baird, 1868:DNA100228—García House, Puerto Peñasco, Sonora,Mexico, October 12, 2000 (M. K. Nishiguchi);DNA100233—Missouri Key, FL, USA Oct. 8, 1993 (J.Wise); DNA100395—phu*ket, Thailand, Jan. 31, 2001 (J.Hylleberg); DNA100748—Bank Reef, Barbados, June 26,2002 (A. Schulze, J. I. Saiz Salinas, E. B. Cutler, G.Y.Kawauchi); DNA100819—Perrier’s Rock, South Africa,Oct. 18, 2002 (R. Biseswar); DNA100829—Farfan, Pan-ama, June 19, 2002 (T. Nishikawa).

Phascolosoma (Phascolosoma) scolops (Selenka & deMan, 1883): DNA100373—Kewalo Reef, Honolulu, HI,USA, Jan. 25, 2001 (J. Brock); DNA100394—phu*ket, Thai-land, July 31, 2001 (J. Hylleberg); DNA100735—Eilat,Israel, Sept. 30, 2002 (N. Ben-Eliahu); DNA100813—Per-rier’s Rock, South Africa, Oct. 18, 2002 (R. Biseswar).

Phascolosoma (Phascolosoma) stephensoni (Stephen, 1942):DNA100203—Cova Blava, Cabrera, Balearic Islands, Spain,May 31, 1997 (X. Turon); DNA100209—Nielsen Park Shore,Port Jackson, Sydney Harbor, NSW, Australia, April 12, 2000(G. Giribet, P. Hutchings); DNA100469—Baileys Bay, Hamil-ton Island, Bermuda, Aug. 7, 2001 (E.B. Cutler); DNA100485—Terceira, Azores, Portugal, Oct. 31, 2001 (P. Wirtz).

Phascolosoma (Phascolosoma) turnerae Rice, 1985:DNA100230—Southwest Reef, Bahamas, Jan. 31, 2000 (S.Brooke, T. GriYn).

Siphonosoma cumanense (Keferstein, 1867):DNA100235—unspeciWed locality, Puerto Rico June 3,1993 (J. Staton, H. Reichardt); DNA100464—Baileys Bay,Hamilton, and south coast of St. Davus, Bermuda, Aug. 7,2001 (T. Nishikawa); DNA100622—Bath, Barbados, June24, 2002 (A. Schulze, J.I. Saiz Salinas); DNA100991—TwinCays, Belize, April 24, 2004 (M. E. Rice, A. Schulze).

Siphonosoma vastum (Selenka & von Bülow, 1883):DNA100625—Bath, Barbados, June 24, 2002 (A. Schulze,J.I. Saiz Salinas, E.B. Cutler).

Sipunculus (Sipunculus) norvegicus Danielssen, 1869:DNA101069—R/V Oceanus, Southern New England, USA,39°47.230�N, 70°46.295�W, June 14, 2003 (A. Schulze).

Sipunculus (Sipunculus) nudus Linnaeus, 1766:DNA100234—Station 53F, No Name Cay, July 28, 1993(J. Staton, H. Reichardt); DNA100245—near Arcachon(Wshermen’s locality unspeciWed), Oct. 30, 2000;DNA100246—unspeciWed locality, Vietnam, Oct. 30,2000; DNA100468—South coast of Stock’s Harbor, StDavid’s Island, Aug. 7, 2001 (E. Cutler); DNA100629—Isla Taboguilla, oV Panama City, Panama, June 20, 2002(T. Nishikawa); DNA100993—Twin Cays, Belize, April24, 2003 (M. E. Rice, A. Schulze).

Sipunculus (Sipunculus) phalloides (Pallas, 1774):DNA101337—Ponta do Araça, Sao Sebastiao, Brazil,23°49�02�S, 45 °24�19�W (G.Y. Kawauchi).

Sipunculus (Sipunculus) polymyotus Fisher, 1947:DNA101121—Pelican Beach, Belize, Oct. 24, 2002 (D. Fel-der, R. Robles).

Themiste (Lagenopsis) lageniformis (Baird, 1868):DNA100229—Jack Island oyster beds, Fort Pierce, FL,USA, Oct. 18, 2000 (S. Reed).

Themiste (Lagenopsis) minor (Ikeda, 1904):DNA100210—Nielsen Park Shore, Port Jackson, SydneyHarbor, April 12, 2000 (G. Giribet, P. Hutchings);DNA101083—unspeciWed locality, South Africa, Sept. 28,2002 (G. Rouse).

Themiste (Themiste) dyscrita (Fisher, 1952):DNA101095—Cape Arago (North Cove), OR, USA, Aug.29, 2002 (M. E. Rice, S. Rumrill, C. Young).

Themiste (Themiste) hennahi Gray, 1828:DNA100627—BahÂ´a de Concepción, Lirquen Playa sectorLa Cata, Chile, April 26, 2001 (E. Tarifeño).

Themiste (Themiste) pyroides (Chamberlin, 1920):DNA101084—WhiYn Spit, Vancouver Island, B.C., Can-ada, Sept. 9, 2003 (A. Schulze, M. E. Rice).

Thysanocardia catherinae (Grübe, 1868): DNA101068—R/V Oceanus, Southern New England, USA, 39°47.230�N,70°48.295�W, June 14, 2003 (A. Schulze).

Thysanocardia nigra (Ikeda, 1904): DNA100606—LopezIsland, WA, USA, May 17, 2002 (D. McHugh).

Xenosiphon branchiatus (Fischer, 1895): DNA101086—Tamarindo Beach, Costa Rica, Feb. 8, 2003 (R. Quiros).

Appendix B. Description of morphological characters

The main sources for the following morphological char-acters were either direct observations or the following pub-lications: Cutler (1994); Rice (1993) and Stephen andEdmonds (1972).

Characters 1-4 (tentacles): A crown of tentacles is usu-ally present at the anterior end of the introvert. In a por-tion of the species, the tentacles encircle the mouthperipherally and the nuchal organ lies outside of that cir-cle (1). In other species, the tentacles form an arc aroundthe nuchal organ (2). Representatives of Thysanocardiaare the only species in which both nuchal and peripheraltentacles are present. In Themiste species, 4-6 stem-likeoutgrowths give rise to numerous branches (3). In Phasco-lion cryptum and representatives of the genus Thysanocar-dia tentacles are not restricted to the very tip of theintrovert but extend for a short distance along it (4).Character coding - 1. Peripheral tentacles: 0D absent,1D present; 2. Nuchal tentacles: 0D absent, 1D present; 3.Branched tentacles: 0 D absent, 1D present; 4. Tentaclesalong introvert: 0D absent, 1D present.

Charcters 5-8 (Nephridia): Nephridia are either pairedor single (5). They usually form simple sacs but are dis-tinctly bilobed in some species (6). They also vary withregard to their position relative to the anus (7) and may be

A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192 187

attached to the body wall (8). Character coding—5. Nephri-dia: 0Dpaired, 1D single; 6. Nephridial attachment:0Dmostly unattached, 1D at least 50% attached; 7.Nephridial shape: 0Dunilobed, 1Dbilobed. 8. Position ofnephridiopores: 0D anterior to anus, 1D same level as anus,2Dposterior to anus.

Characters 9-16 (body wall): In several sipunculan gen-era the coelom extends into the body wall (9). In Siphono-soma and Siphonomecus (the latter not included in thisstudy) the extensions are sac-like whereas they formcanals in Sipunculus and Xenosiphon (Ruppert and Rice,1995) (10). The canals either run longitudinally betweenthe longitudinal muscle bands (Sipunculus), or diagonallyas short, subcutaneous canals (Xenosiphon) (11). The bodywall consists of an outer layer of circular and an innerlayer of longitudinal musculature. Both can either formcontinuous sheets or a variable number of bands (12, 13,15). Muscle bands may be distinct for most of their lengthor anastomosing (14, 16). Character coding—9. Coelomicextensions in body wall: 0D absent, 1D present; 10. Typeof coelomic extensions: 0D sacs, 1D canals; 11. Orienta-tion of coelomic canals: 0D longitudinal, 1D in bands; 12.Longitudinal musculature: 0D distinct, 1D anastomosing;13. Number of longitudinal muscle bands: 0D commonly< 25; 1D 25-40, 2 D > 40; 14. Longitudinal muscle bands:0D distinct, 1D anastomosing. 15: Circular musculature:0D continuous sheet, 1D in bands; 16. Circular musclebands: 0D distinct, 1D anastomosing.

Characters 17-20 (anal shield): All currently recognizedAspidosiphonidae are characterized by a calcareous orhorny protein shield shield at the anterior end of thetrunk. However, its chemical composition, extend andmorphology are variable among the species (17). The analshield is cone-shaped in Lithacrosiphon and more or lessXat in Aspidosiphon (18) and may be relatively smooth orbear distinct grooves (19). In both these genera it isrestricted to the dorsal side. In Cloeosiphon it is composedof numerous calcareous plates that surround the anteriorintrovert (“pineapple shield”) (20). Character coding—17.Dorsal shield: 0D absent, 1D present; 18. Shape of dorsalshield: 0D Xat, 1D cone-shaped; 19. Grooves in dorsalshield: 0D absent, 1D present. 20. Pineapple shield:0D absent, 1D present.

Characters 21-25 (spindle muscle): The spindle muscleis a slender, thread-like muscle that runs through theintestinal coil. Anteriorly, it is either attached to the bodywall or on the rectum or wing muscle (22, 23). In mostcases there is only one anterior point of insertion, exceptin Siphonosoma and two of the Themiste species includedhere where the muscle sends branches anteriorly with sev-eral points of insertion (24). Posteriorly, it attaches to thebody wall or ends in the gut coil (25). Character coding—21. Spindle muscle: 0D absent, 1D present; 22. Anteriorattachment of spindle muscle: 0D body wall, 1D on rec-tum. 23. Level of attachment on body wall: 0D anterior toanus, 1D same level as anus, 2 D posterior to anus. 24.Anterior roots of spindle muscle: 0 D one, 1D two or

more. 25. Posterior attachment of spindle muscle:0D posterior body wall, 1Dwithin gut coil.

Characters 26-40 (hooks): Other than in polychaetes, thehooks in sipunculans are non-chitinous epidermal struc-tures (Andreae, 1882; Andrews, 1890; Voss-Foucart et al.,1977; Voss-Foucart et al., 1978). In representatives of theSipunculidea, hooks, if present, are generally scattered, sim-ple and only slightly curved. In the Phascolosomatideahooks are usually arranged in rings, are sharply curved pos-teriorly and show a distinct internal anatomy when viewedwith light microscopy. The number of rings of hooks variesgreatly and can exceed 100. Bidentate hooks (30) are oftenfound in the Aspidosiphonidae and are distinct from hooksin Phascolosoma (31). In both cases, there is a primarytooth and a secondary tooth, but whereas the secondarytooth in Aspidosiphon hooks is usually pointed, close to theprimary tooth, similar in shape and only slightly smaller, inPhascolosoma it is closer to the base of the hook and alwaysmore blunt (somewhat pointed in P. stephensoni but stillwith broad base). The internal anatomy of Phascolosomahooks may include an anterior clear triangle (33), a clearstreak (34) and a posterior, crescent-shaped space (35).Phascolosoma hooks are also characterized by posteriorbasal structures (36) that can take the shape of toes, root-lets or warts (37). Warts are only found in P. glabrum, notincluded in this study. The angle of the hook relative to thebody axis is usually greater than 90 °, except in two Phasco-lion species, Cloeosiphon aspergillus and Apionsoma misa-kianum where it forms a smaller angle (38).

In Aspidosiphonidae, several zones of hooks can bedistinguished along the introvert. In the distal zone, thehooks are laterally compressed and usually arranged inrings. In many cases, the anterior rings consist of biden-tate hooks, followed by a zone of rings with unidentatehooks. Behind the hooks arranged in rings, there is often azone of scattered hooks. These may include hooks that aresimilar in shape to the hooks in the rings, or pyramidal(39) and conical (40) hooks. Pyramidal hooks can be dis-tinguished from laterally compressed hooks by their trian-gular base but intermediate forms between the two maybe present. Conical hooks have a nearly circular cross sec-tion and are only found in the dorsal region of the intro-vert in A. elegans (Cutler and Cutler, 1989).

Character coding—26. Hooks on introvert in adults:0D absent, 1D present; 27. Hooks in rings: 0D absent,1D present; 28. Number of rings of hooks: 0D< 50, 1D>50. 29. Scattered hooks: 0D absent, 1D present. 30. Biden-tate hooks: 0D absent, 1D present. 31. Secondary toothon hook: 0D absent, 1D present; 32. Shape of secondarytooth: 0D blunt, 1D pointed; 33. Anterior clear triangle:0D absent, 1D present; 34. Clear streak: 0D absent,1D present; 35. Crescent-shaped space: 0D absent,1D present; 36. Posterior basal structures: 0D absent,1D present; 37. Type of posterior basal structures:0Dwarts, 1D rootlets; 38. Angle of hook: 0D< 90%,1D> 90%. 39. Pyramidal hooks: 0D absent, 1D present.40. Conical hooks: 0D absent, 1D present.

188 A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192

Character 41 (anus location): The anus is usually locateddorsally at the anterior trunk. However, in two of thespecies included here, Phascolion gerardi and Onchnesomasteenstrupii it is shifted anteriorly onto the introvert(Rice, 1993; Shipley, 1892). Character coding—41. Loca-tion of anus: 0Don anterior trunk, 1Don introvert.

Character 42 (pigmented introvert bands): In Phascolo-soma species, the dorsal side of the introvert is usuallydarker than the ventral side and the pigment is often dis-tributed in distinct broad stripes. The presence or absenceof these pigment bands is constant within species (Cutlerand Cutler, 1990). Character coding—42. Pigmentedintrovert bands: 0D absent, 1D present.

Characters 43, 44 (contractile vessel villi): The contrac-tile vessel is part of the tentacular coelomic system. It runsdorsally along the esophagus and has a coelomic lining. Itcontains hemocytes and is considered an analogue to ablood vascular system (Pilger, 1982). In some species, inparticular of the genus Themiste, villi are present alongthe length of the vessel (Rice, 1993) (43). These vary innumber and length: whereas they are relatively short inThysanocardia species, Antillesoma antillarum, Siphono-soma cumanense and Themiste (Lageniformis) the villi areshort and digitiform, they are long thread-like and fewerin number in Themiste (Themiste) (Edmonds, 1980). Char-acter coding—43. Contractile vessel villi: 0D absent, 1D 1 present; 44. Type of contractile vessel villi:0D digitiform, 1D elongate tubules.

Characters 45-51 (introvert retractor muscles): This setof strong muscles insert anteriorly near the brain and areposteriorly attached to the body wall. The number ofintrovert retractors varies from a single column to four(45, 46) although all pelagosphera larvae or early juvenilesseem to have four retractor muscles (A.S. pers. obs.). Iffour retractors are present in the adult, they are arrangedin a dorsal and a ventral pair. The pairs are fused to vari-ous degrees (47) but separate origins on the body wall areusually discernable. If only one pair of retractors is pres-ent the muscles originate on the ventral side and are hereregarded as ventral retractors. However, this needs furtherinvestigation as they might actually be the dorsal retrac-tors that have shifted ventrally or a product of fusionbetween dorsal and ventral retractors. The point of originof the muscles along the body wall also varies among spe-cies (48, 49). In some cases (Antillesoma antillarum andPhascolosoma scolops), there is evidence of fusion betweenthe dorsal and the ventral retractor muscle on each sidebut this varies intraspeciWcally and is not included in thisdata set. The points of origin of the retractor musclesalong the body wall vary greatly and can be in the ante-rior third of the trunk (e.g. Sipunculus), in the middle third(e.g. most Phascolosoma species) or near the posterior end(e.g. Onchnesoma). Character coding – 45. Introvertretractor muscles in adult: 0D two pairs, 1D less than twopairs; 46. Fusion of dorsal retractors: 0D not fused (<10%), 1D partially fused, 2D completely fused; 47. Fusionin ventral retractors: 0D not fused (< 10%), 1D partially

fused, 2D completely fused; 48. Fusion of dorsal and ven-tral retractors: 1D absent, 2D present; 49: Retractor col-umn: 0D absent, 1D present; 50. Origin of dorsalretractor muscles: 0D anterior 1/3 of body, 1Dmiddle 1/3of body, 2D posterior 1/3 of body; 51: Origin of ventralretractor muscles: 0D anterior 1/3 of body, 2Dmiddle 1/3of body; 3D posterior 1/3 of body.

Character 52 (protractor muscle): In Xenosiphon, a pairof short muscles connects the introvert near the brain withthe anterior body wall at the level of the anus. Charactercoding—52. Protractor muscle: 0D absent, 1D present.

Character 53 (ratio of introvert/trunk length): Theratio between the fully extended introvert and the trunklength is here roughly divided into three categories.Although some intraspeciWc variation is often observed,most species clearly fall into one of the categories. Char-acter coding—53. Ratio of introvert/trunk length: 0D<0.75, 1D 0.75-2, 2D> 2.

Character 54 (holdfast papillae with hardened borders):These are specialized papillae found in Phascolion speciesthat inhabit abandoned shells of gastropods, scaphopods orforaminiferans. The papillae are usually located in the pos-terior or mid-trunk region and have sclerotinized borders.The borders may surround the anterior margin of a roundpapilla (e.g. P. hedraeum), be U-shaped or V-shaped (e.g.Phascolion cryptum), or form pointed projections (P.caupo). Character coding—54. Holdfast papillae with hard-ened borders: 0Dabsent, 1Dpresent.

Character 55 (caudal appendage): This tail-likeprojection at the posterior end of the trunk occurs inNephasoma Xagriferum and two GolWngia species. Thelatter are not included in this dataset, so that the charac-ter is here an autapomorphy for Nephasoma Xagriferum.Character coding—55. Caudal appendage: 0 D absent,1 D present.

Characters 56-58 (nuchal organ): Nuchal organs areprobably chemosensory organs known also in annelids. Insipunculans, nuchal organs are located close to the tenta-cles at the anterior introvert in sipunculans (Åkesson,1958; Rice, 1993). The nuchal organs of only one specieshave been examined with ultrastructural methods (Purs-chke, 1997). Åkesson (1958) originally concluded thatnuchal organs in annelids and sipunculans are hom*olo-gous. Later authors came to the opposite conclusion(Purschke, 1997, 2002) but suggested that more studies onsipunculans be undertaken to answer the question withmore certainty. Three main designs are here distinguishedin sipunculans (56): simple pits as in Sipunculus species,ciliated bands as in Phascolion cryptum and P. gerardi orpatches (all other species where known). The number ofpatches varies between one and two (57). If only a singlepatch is present, it might be triangular to heart-shaped,distinctly bilobed or multilobed (58). Character coding—56. Nuchal organ: 0D pit, 1D ciliated band, 2D patch orpatches; 57. Number of nuchal patches: 0 D one, 1D two;58. Shape of nuchal patch: 0D triangular, 1D bilobed,2Dmulti-lobed.

A. S

chulze et al. / Molecular P

hylogenetics and Evolution 42 (2007) 171–192

189Appendix C. Morphological data matrix for the 58 morphological characters described in Appendix B. “x” indicates inapplicable characters while “?” indicates

5 46 47 48 49 50 51 52 53 54 55 56 57 58

1 1 1 0 1 1 0 0 0 0 2 0 1

0 0 0 0 1 0 0 2 0 0 ? ? ?

0 0 0 0 1,2 1,2 0 2 0 0 ? ? ?

0 0 0 0 0 0 0 2 0 0 2 0 1

x 2 x 0 x 2 0 2 0 0 2 0 ?

x 1 x 0 x 2 0 2 0 0 2 0 0

x 1 x 0 x 2 0 1,2 0 0 2 0 ?

x 1 x 0 x 2 0 1 0 0 2 0 ?

x 1 x 0 x 2 0 1 0 0 2 0 0

x 2 x 0 x 2 0 1 0 0 2 0 0

x 2 x 0 x 2 0 1 0 0 2 0 ?

x 2 x 0 x 2 0 1 0 0 2 0 0

x 1 x 0 x 2 0 1 0 0 2 0 ?

0 0,1 0 0 0 1 0 0 0 0 2 0 ?0 0 0 0 0 1 0 0 0 0 2 1 00 0 0 0 0 1 0 0,1 0 0 2 1 0x 1 x 0 x 2 0 1 0 0 2 0 ?x 1 x 0 x 1,2 0 0 0 0 ? ? ?x 1 x 0 x 1 0 1 0 1 2 ? ?x 1 x 0 x 1 0 0 0 0 2 1 0x 2 x 1 x 2 0 2 0 0 2 0 02 1 0 0 2 2 0 1 0 0 1 x xx 2 x 1 x 2 0 1 1 0 1 x x2 2 0 0 2 2 0 1 0 0 2 0 0

2 2 0 0 2 2 0 1 1 0 2 1 0

0 0 0 0 0 1 0 0 0 0 2 ? ?(continued on next page)

missing observations1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 4

Antillesoma antillarum

0 1 0 0 0 1 0 2 0 x x 1 0 1 0 x 0 x x 0 1 0 1 0 0 0 x x x x x x x x x x x x 0 0 0 0 1 0 0

Apionsoma (A.) misakianum

0 1 0 0 0 0 1 0 0 x x 0 x x 0 x 0 x x 0 1 0 0 0 0 1 1 0 0 0 0 x 0 0 0 0 x 1 0 0 0 0 0 x 0

Apionsoma (A.) murinae

0 1 0 0 0 0 0,1 1 0 x x 0 x x 0 x 0 x x 0 1 0 0 0 0 1 1 0 0 0 0 x 0 0 0 0 x 0 0 0 0 0 0 x 0

Apionsoma (E.) pectinatum

0 1 0 0 0 1 1 1 0 x x 1 0,1 1 0 x 0 x x 0 1 0 0 0 1 1 1 0 0 0 0 x 0 0 0 0 x 0 0 0 0 0 0 x 0

Aspidosiphon (A.) albus

0 0 0 0 0 1 0 1 0 x x 0 x x 0 x 1 0 0 0 1 0 ? ? 0 0 x x x x x x x x x x x x 0 0 0 0 0 x 1

Aspidosiphon (A.) elegans

0 1 0 0 0 1 0 2 0 x x 0 x x 0 x 1 0 0 0 1 0 2 0 0 1 1 1 1 1 0 x 0 0 0 0 x 0 0 1 0 0 0 x 1

Aspidosiphon (A.) gosnoldi

0 1 0 0 0 1 0 ? 0 x x 0 x x 0 x 1 0 0 0 1 0 0 0 0 1 1 1 1 1 0 x 0 0 0 0 x 0 1 0 0 0 0 x 1

Aspidosiphon (A.) gracilis

0 1 0 0 0 1 0 ? 0 x x 0 x x 0 x 1 0 0 0 1 0 ? ? 0 1 1 1 1 0 0 x 0 0 0 0 x 0 1 0 0 0 0 x 1

Aspidosiphon (A.) misakiensis

0 1 0 0 0 1 0 2 0 x x 0 x x 0 x 1 0 0 0 1 0 0 0 0 1 1 1 1 1 0 x 0 0 0 0 x 0 0 0 0 0 0 x 1

Aspidosiphon (A.) muelleri

0 1 0 0 0 1 0 1 0 x x 0 x x 0 x 1 0 1 0 1 0 0 0 0 1 1 1 1 1 0 x 0 0 0 0 x 0 1 0 0 0 0 x 1

Aspidosiphon (P.) Wscheri

0 1 0 0 0 1 0 0 0 x x 1 0 1 0 x 1 0 0 0 1 0 0 0 0 1 1 1 1 1 0 x 0 0 0 0 x 0 1 0 0 0 0 x 1

Aspidosiphon (P.) laevis

0 1 0 0 0 1 0 ? 0 x x 1 1,2 1 0 x 1 0 1 0 1 0,1 0,2 1 0 1 1 1 1 0 0 x 0 0 0 0 x 0 0 0 0 0 0 x 1

Aspidosiphon (P.) parvulus

0 1 0 0 0 1 0 ? 0 x x 1 0 1 0 x 1 0 0 0 1 0 0 0 0 1 1 1 1 1 0 x 0 0 0 0 x 0 1 0 0 0 0 x 1

Aspidosiphon (P.) steenstrupii

0 1 0 0 0 1 0 1 0 x x 1 0 1 0 x 1 0 0 0 1 0 0 0 0 1 1 1 1 1 0 x 0 0 0 0 x 0 1 0 0 0 0 x 1

Cloeosiphon aspergillus

0 1 0 0 0 1 0 1 0 x x 0 x x 0 x 0 x x 1 1 1 x x 0 1 1 1 0 1 0 x 0 0 0 0 x 1 0 0 0 0 0 x 1

GolWngia elongata 1 0 0 0 0 0 0 1 0 x x 0 x x 0 x 0 x x 0 1 1 x x 1 1 0 x 0 0 0 x 0 0 0 0 x 0 0 0 0 0 0 x 0GolWngia margaritacea 1 0 0 0 0 0 0 1 0 x x 0 x x 0 x 0 x x 0 1 1 x x 1 0 x x x x x x x x x x x x 0 0 0 0 0 x 0GolWngia vulgaris 1 0 0 0 0 0 0 0 0 x x 0 x x 0 x 0 x x 0 1 1 x x 1 1 0 x 0 0 0 x 0 0 0 0 x 0 0 0 0 0 0 x 0Lithacrosiphon cristatus 0 1 0 0 0 0 0 2 0 x x 1 0 1 0 x 1 1 1 0 1 0 0 0 0 1 1 1 0 1 0 x 0 0 0 0 x 0 0 0 0 0 0 x 1Nephasoma diaphanes 1 0 0 0 0 0 0 1 0 x x 0 x x 0 x 0 x x 0 1 ? ? ? 1 1 0 x 1 0 0 x 0 0 0 0 x 0 0 0 0 0 0 x 1Nephasoma Xagriferum 1 0 0 0 0 0 0 1 0 x x 0 x x 0 x 0 x x 0 1 0 0 0 1 0 x x x x x x x x x x x x 0 0 0 0 0 x 1Nephasoma pellucidum 1 0 0 0 0 0 0 1 0 x x 0 x x 0 x 0 x x 0 1 1 x x 1 1 0 x 1 0 0 x 0 0 0 0 x 0 0 0 0 0 0 x 1Onchnesoma steenstrupii 0 0 0 0 1 0 0 2 0 x x 0 x x 0 x 0 x x 0 0 x x x x 0 x x x x x x x x x x x x 0 0 1 0 0 x 1Phascolion (I.) gerardi 1 0 0 0 1 1 0 2 0 x x 0 x x 0 x 0 x x 0 0 x x x x 1 0 x 1 0 0 x 0 0 0 0 x 0 0 0 1 0 0 x 0Phascolion (L.) cryptum 1 0 0 1 1 0 0 2 0 x x 0 x x 0 x 0 x x 0 0 x x x x 0 x x x x x x x x x x x x 0 0 0 0 0 x 1Phascolion

(P.) psammophilum1 0 0 0 1 0 0 2 0 x x 0 x x 0 x 0 x x 0 0 x x x x 0 x x x x x x x x x 0 x 0 0 0 0 0 0 x 0

Phascolion (P.) strombus

1 0 0 0 1 1 0 2 0 x x 0 x x 0 x 0 x x 0 0 x x x x 1 0 x 1 0 0 x 0 0 0 0 x 0 0 0 0 0 0 x 0

Phascolopsis gouldi 1 0 0 0 0 0 0 2 0 x x 1 1 1 0 x 0 x x 0 1 0 0 0 1 0 x x x x x x x x x x x x 0 0 0 0 0 x 0

190A

. Schulze et al. / M

olecular Phylogenetics and E

volution 42 (2007) 171–192

39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58

Phascolos(F.) ca

0 0 0 0 0 x 0 0 1 1 0 1,2 2 0 1 0 0 2 0 1

Phascolos(P.) ag

0 0 0 1 0 x 0 0 0 0 0 1,2 1 0 1 0 0 2 0 1

Phascolos(P.) alb

0 0 0 1 0 x 0 0 0 0 0 1 1 0 1 0 0 2 0 1

Phascolos(P.) gr

0 0 0 0 0 x 0 0 0 1 0 1 1 0 1 0 0 2 0 1

Phascolos(P.) nig

0 0 0 1 0 x 0 0 0 0 0 1 1 0 1 0 0 2 ? ?

Phascolos(P.) no

0 0 0 1 0 x 0 0 0 0 0 1 1 0 1 0 0 2 ? ?

Phascolos(P.) pe

0 0 0 1 0 x 0 0 0 0 0 1 1 0 1 0 0 2 0 1

Phascolos(P.) sc

0 0 0 1 0 x 0 0 0 1 0 1 1 0 1 0 0 2 ? ?

Phascolos(P.) ste

0 0 0 1 0 x 0 0 0 0 0 1 1 0 1 0 0 2 0 ?

Phascolos(P.) tu

0 0 0 0 0 x 0 0 0 0 0 1 2 0 1 0 0 2 0 1

Siphonoso 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 ? ? ?Siphonoso 0 0 0 0 0 x 0 0 0 0 0 0 0 0 0 0 0 ? ? ?Sipunculu 0 0 0 0 0 x 0 0 0 0 0 0 0 0 0 0 0 0 x xSipunculu 0 0 0 0 0 x 0 0 0 0 0 0 0 0 0 0 0 0 x xSipunculu 0 0 0 0 0 x 0 0 0 0 0 0 0 0 0 0 0 0 x xSipunculu 0 0 0 0 0 x 0 0 0 0 0 0 0 0 0 0 0 0 x xThemiste

(L.) lag0 0 0 0 1 0 1 x 1 x 0 x 2 0 0 0 0 2 0 2

Themiste 0 0 0 0 1 0 1 x 0 x 0 x 1 0 0 0 0 2 0 2Themiste 0 0 0 0 1 1 1 x 0 x 0 x 2 0 0 0 0 2 0 ?Themiste 0 0 0 0 1 1 1 x 0 x 0 x 1 0 0 0 0 2 0 ?Themiste 0 0 0 0 1 1 1 x 0 x 0 x 2 0 0 0 0 2 0 ?Thysanoca

catherin0 0 0 0 1 0 1 x 1 x 0 x 2 0 1 0 0 2 0 0

Thysanoca 0 0 0 0 1 0 1 x 1 x 0 x 1,2 0 1 0 0 2 0 0Xenosipho 0 0 0 0 0 x 0 0 0 0 0 0 0 1 0 0 0 ? ? ?

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38

oma pitatum

0 1 0 0 0 1 0 1 0 x x 0 x x 0 x 0 x x 0 1 0 0 0 0 1 1 0 0 0 0 x 0 1 0 0 x 0

oma assizii

0 1 0 0 0 1 0 1 0 x x 1 0,1 1 0 x 0 x x 0 1 0 0 0 0 1 1 0 0 0 0,1 0 1 1 0 1 0 0

oma olineatum

0 1 0 0 0 1 0 1 0 x x 1 1 1 0 x 0 x x 0 1 0 0 0 0 1 1 0 0 0 1 0 1 1 0 1 0 1

oma anulatum

0 1 0 0 0 1 0 1 0 x x 1 0 1 0 x 0 x x 0 1 0 0 0 0 1 1 1 0 0 0,1 0 0 1 0 1 0 0

oma rescens

0 1 0 0 0 1 0 1 0 x x 1 1 1 0 x 0 x x 0 1 0 0 0 0 1 1 1 0 0 0,1 0 0 1 0 1 0 0

oma duliferum

0 1 0 0 0 1 0 1 0 x x 1 0 1 0 x 0 x x 0 1 0 0 0 0 1 1 1 0 0 0 x 0 1 0 1 0 0

oma rlucens

0 1 0 0 0 1 0 1 0 x x 1 0 1 0 x 0 x x 0 1 0 0 0 0 1 1 0 0 0 1 0 1 1 0 1 0 0

oma olops

0 1 0 0 0 1 0 1 0 x x 1 0 1 0 x 0 x x 0 1 0 0 0 0 1 1 0 0 0 0,1 0 1 1 0 1 0 0

oma phensoni

0 1 0 0 0 1 0 1 0 x x 1 1 1 0 x 0 x x 0 1 0 0 0 0 1 1 1 0 0 1 1 1 1 1 1 0 0

oma merae

0 1 0 0 0 1 0 1 0 x x 1 1 1 0 x 0 x x 0 1 0 0 0 0 1 1 1 0 0 0 x 0 0 0 1 1 0,1

ma cumanense 1 0 0 0 0 0 0 2 1 0 x 1 0 1 1 1 0 x x 0 1 0 0 1 0 0 x x x x x x x x x x x xma vastum 1 0 0 0 0 1 0 0 1 0 x 1 0 1 1 1 0 x x 0 1 0 0 1 0 1 1 0 0 0 0 x 0 0 0 0 x 0s norvegicus 1 0 0 0 0 0 0 0 1 1 0 1 0 0 1 0 0 x x 0 1 0 0 0 1 0 x x x x x x x x x x x xs nudus 1 0 0 0 0 0 0 0 1 1 0 1 1 0 1 0 0 x x 0 1 0 0 0 1 0 x x x x x x x x x x x xs phalloldes 1 0 0 0 0 0 0 0 1 1 0 1 1 0 1 0 0 x x 0 1 0 0 0 1 0 x x x x x x x x x x x xs polymyotus 1 0 0 0 0 0 0 1 1 1 0 1 2 0 1 0 0 x x 0 1 0 0 0 1 0 x x x x x x x x x x x x

enlformes1 0 1 0 0 0 0 1 0 x x 0 x x 0 x 0 x x 0 1 0 2 1 1 0 x x x x x x x x x x x x

(L.) minor 1 0 1 0 0 0 0 1 0 x x 0 x x 0 x 0 x x 0 1 0 1 0 1 1 0 x 1 0 0 x 0 0 0 0 x 0(T.) dyscrita 1 0 1 0 0 0 0 2 0 x x 0 x x 0 x 0 x x 0 1 0 1 0 1 0 x x x x x x x x x x x x(T.) hennahi 1 0 1 0 0 0 0 2 0 x x 0 x x 0 x 0 x x 0 1 0 2 1 1 0 x x x x x x x x x x x x(T.) pyroides 1 0 1 0 0 0 0 2 0 x x 0 x x 0 x 0 x x 0 1 0 2 0 1 1 0 x 1 0 0 x 0 0 0 0 x 0rdia ae

1 1 0 1 0 0 0 0 0 x x 0 x x 0 x 0 x x 0 1 0 1 0 1 0 x x x x x x x x x x x x

rdia nigra 1 1 0 1 0 0 0 2 0 x x 0 x x 0 x 0 x x 0 1 0 1 0 1 0 x x x x x x x x x x x xn branchiatus 1 0 0 0 0 0 0 2 1 1 1 1 1 0 1 0 0 x x 0 1 0 0 0 1 0 x x x x x x x x x x x x

A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192 191

References

Aagesen, L., Petersen, G., Seberg, O., 2005. Sequence length variation,indel costs and congruence in sensitivity analysis. Cladistics 21, 15–30.

Åkesson, B., 1958. A study of the nervous system of the sipunculoideae,with some remarks on the development of the two species Phascolionstrombi Montagu and GolWngia minuta Keferstein. Undersokningarover Oresund 38, 1–249.

Andreae, J., 1882. Beitrage zur Anatomie und Histologie des Sipunculusnudus L. Zeitschr. Wiss. Zool. 36, 201–255.

Andrews, E.A., 1890. Notes on the anatomy of Sipunculus gouldii Pour-tales. Stud. Biol. Lab. Johns Hopkins Univ. 4, 384–430.

Bleidorn, C., Vogt, L., Bartolomaeus, T., 2003a. A contribution to seden-tary polychaete phylogeny using 18S rRNA sequence data. J. Zool.Syst. Evol. Res. 41, 186–195.

Bleidorn, C., Vogt, L., Bartolomaeus, T., 2003b. New insights into poly-chaete phylogeny (Annelida) inferred from 18S rDNA sequences. Mol.Phylogenet. Evol. 29, 279–288.

Bleidorn, C., Podsiadlowski, L., Bartolomaeus, T., 2005. The completemitochondrial genome of the orbiniid polychaete Orbinia latreillii(Annelida, Orbiniidae)—a novel gene order for Annelida and implica-tions for annelid phylogeny. Gene 370, 96–103.

Boore, J.L., Staton, J.L., 2002. The mitochondrial genome of the sipunculidPhascolopsis gouldii supports its association with Annelida rather thanMollusca. Mol. Biol. Evol. 19, 127–137.

Brown, S., Rouse, G., Hutchings, P., Colgan, D., 1999. Assessing the useful-ness of histone H3, U2 snRNA and 28S rDNA in analyses of poly-chaete relationships. Austr. J. Zool. 47, 499–516.

Colgan, D.J., McLauchlan, A., Wilson, G.D.F., Livingston, S.P., Edge-combe, G.D., Macaranas, J., Cassis, G., Gray, M.R., 1998. Histone H3and U2 snRNA DNA sequences and arthropod molecular evolution.Austr. J. Zool. 46, 419–437.

Cutler, E.B., 1973. Sipuncula of the western North Atlantic. Bull. Amer.Mus. Nat. Hist. 152, 105–204.

Cutler, E.B., 1979. A reconsideration of the GolWngia subgenera FisheranaStephen, Mitosiphon Fisher, and Apionsoma Sluiter (Sipuncula). Zool.J. Linn. Soc. 65, 367–384.

Cutler, E.B., 1994. The Sipuncula. Their Systematics, Biology and Evolu-tion. Cornell University Press, Ithaca, NY.

Cutler, E.B., Cutler, N.J., 1987. Deep-water Sipuncula from the EasternAtlantic Ocean. Sarsia 72, 71–89.

Cutler, E.B., Cutler, N.J., 1989. A revision of the genus Aspidosiphon(Sipuncula, Aspidosiphonidae). Proc. Biol. Soc. Wash. 102, 826–865.

Cutler, N.J., Cutler, E.B., 1990. A revision of the subgenus Phascolosoma(Sipuncula, Phascolosoma). Proc. Biol. Soc. Wash. 3, 691–730.

Cutler, E.B., Gibbs, P.E., 1985. A phylogenetic analysis of higher taxa inthe phylum Sipuncula. Syst. Zool. 34, 162–173.

Edmonds, S.J., 1980. A revision of the systematics of Australian sipuncu-lans (Sipuncula). Rec. S. Austr. Mus. 18, 1–74.

Farris, J.S., 1997. The future of phylogeny reconstruction. Zool. Scr. 26,303–311.

Farris, J.S., Albert, V.A., Källersjö, M., Lipscomb, D., Kluge, A.G., 1996.Parsimony jackkniWng outperforms neighbor-joining. Cladistics 12,99–124.

Farris, J.S., Källersjö, M., Kluge, A.G., Bult, C., 1995. Testing signiWcanceof incongruence. Cladistics 10, 315–319.

Fleissner, R., Metzler, D., von Haeseler, A., 2005. Simultaneous statisticalmultiple alignment and phylogeny reconstruction. Syst. Biol. 54, 548–561.

Folmer, O., Black, M., Hoeh, W., Lutz, R., Vrijenhoek, R., 1994. DNAprimers for ampliWcation of mitochondrial cytochrome c oxidase sub-unit I from diverse metazoan invertebrates. Mol. Mar. Bio. Biotechnol.3, 294–295.

Gibbs, P.E., Cutler, E.B., 1987. A classiWcation of the phylum Sipuncula.Bull. Brit. Mus. Nat. Hist. (Zool.) 52, 43–58.

Giribet, G., 2003. Stability in phylogenetic formulations and its relation-ship to nodal support. Syst. Biol. 52, 554–564.

Giribet, G., 2005. Generating implied alignments under direct optimiza-tion using POY. Cladistics 21, 396–402.

Giribet, G., Carranza, S., Baguñà, J., Riutort, M., Ribera, C., 1996. Firstmolecular evidence for the existence of a Tardigrada + Arthropodaclade. Mol. Biol. Evol. 13, 76–84.

Giribet, G., Distel, D.L., Polz, M., Sterrer, W., Wheeler, W.C., 2000. Triplo-blastic relationships with emphasis on the acoelomates and the posi-tion of Gnathostomulida, Cycliophora, Plathelminthes, andChaetognatha: a combined approach of 18S rDNA sequences andmorphology. Syst. Biol. 49, 539–562.

Giribet, G., Edgecombe, G.D., Wheeler, W.C., Babbitt, C., 2002. Phylogenyand systematic position of Opiliones: a combined analysis of chelicer-ate relationships using morphological and molecular data. Cladistics18, 5–70.

Giribet, G., Rambla, M., Carranza, S., Baguñà, J., Riutort, M., Ribera, C.,1999. Phylogeny of the arachnid order Opiliones (Arthropoda) inferredfrom a combined approach of complete 18S and partial 28S ribosomalsequences and morphology. Mol. Phylogenet. Evol. 11, 296–307.

Giribet, G., Wheeler, W.C., 1999. On gaps. Mol. Phylogenet. Evol. 13, 132–143.

GoloboV, P.A., 1999. Analyzing large data sets in reasonable times: solu-tions for composite optima. Cladistics 15, 415–428.

GoloboV, P.A., 2002. Techniques for analyzing large data sets. In: DeSalle,R., Giribet, G., Wheeler, W.C. (Eds.), Techniques in Molecular System-atics and Evolution. Birkhäuser Verlag, Basel, pp. 70–79.

Hall, K.A., Hutchings, P., Colgan, D.J., 2004. Further phylogenetic studiesof the Polychaeta using 18S rDNA sequence data. J. Mar. Biol. Ass.UK 84, 949–960.

Hall, T. 1994. BioEdit Sequence Alignment Editor 6.0.7 (available at http://www.mbio.ncsu.edu/BioEdit/page2.html).

Huang, D., Chen, J., Vannier, J., Saiz Salinas, J.I., 2004. Early Cambriansipunculan worms from southwest China. Proc. Roy. Soc. Lond. B 271,1671–1676.

Jennings, R.M., Halanych, K.M., 2005. Mitochondrial genomes of Clyme-nella torquata (Maldanidae) and Riftia pachyptila (Siboglinidae): evi-dence for conserved gene order in Annelida. Mol. Biol. Evol. 22, 210–222.

Jukes, T., Cantor, C., 1969. Evolution of protein molecules. In: Munro, H.(Ed.), Mammalian Protein Metabolism. Academic Press, New York.

Keferstein, W., 1867. Untersuchungen uber einige amerikanische Sipuncul-iden. Zeitsch. Wiss. Zool. 17, 44–55.

Knowlton, N., 1993. Sibling species in the sea. Ann. Rev. Ecol. Syst. 24,189–216.

Knowlton, N., 2000. Molecular genetic analyses of species boundaries inthe sea. Hydrobiologia 420, 73–90.

Lewis, P.O., 2001. A likelihood approach to estimating phylogeny fromdiscrete morphological data. Syst. Biol. 50, 913–925.

Linton, E.W., 2005. MacGDE: Genetic Data Environment for MacOS X.Ver. 2.0. Software available at www.msu.edu/~lintone/macgde/.

Maxmen, A.B., King, B.F., Cutler, E.B., Giribet, G., 2003. Evolutionaryrelationships within the protostome phylum Sipuncula; a molecularanalysis of ribosomal genes and histone H3 sequence data. Mol. Biol.Evol. 27, 489–503.

Mickevich, M.F., Farris, J.S., 1981. The implications of congruence inMenidia. Syst. Zool. 27, 143–158.

Moilanen, A., 1999. Searching for most parsimonious trees with simulatedevolutionary optimization. Cladistics 15, 39–50.

Moilanen, A., 2001. Simulated evolutionary optimization and local search:introduction and application to tree search. Cladistics 17, S12–S25.

Nylander, J.A.A., 2004. MrModeltest v2. Program distributed by theauthor.

Pilger, J.F., 1982. Ultrastructure of the tentacles of Themiste lageniformis(Sipuncula). Zoomorph. 100, 143–156.

Peterson, K.J., Eernisse, D.J., 2001. Animal phylogeny and the ancestry ofbilaterians: inferences from morphology and 18S rDNA genesequences. Evol. Dev. 3, 170–205.

Posada, D., Crandall, K.A., 1998. Modeltest: testing the model of DNAsubstitution. Bioinformatics 14, 817–818.

Purschke, G., 1997. Ultrastructure of the nuchal organs in polychaetes(Annelida)—new results and review. Acta Zool. 78, 123–143.

http://www.mbio.ncsu.edu/BioEdit/page2.html

http://www.msu.edu/~lintone/macgde/

192 A. Schulze et al. / Molecular Phylogenetics and Evolution 42 (2007) 171–192

Purschke, G., 2002. On the ground pattern of Annelida. Org. Divers. Evol.2, 181–196.

Redelings, B.D., Suchard, M.A., 2005. Joint Bayesian estimation of align-ment and phylogeny. Syst. Biol. 54, 410–418.

Rice, M.E., 1985. Description of a wood dwelling sipunculan, Phascolo-soma turnerae, new species. Proc. Biol. Soc. Wash. 98, 54–60.

Rice, M.E., 1993. Sipuncula. In: Harrison, F.W., Rice, M.E. (Eds.), Micro-scopic Anatomy of Invertebrates: Onychophora, Chilopoda, andLesser Protostomata. Wiley-Liss, New York, pp. 237–325.

Ronquist, F., Huelsenbeck, J.P., 2003. MrBayes 3: Bayesian phylogeneticinference under mixed models. Bioinformatics 19, 1572–1574.

Ruppert, E.E., Rice, M.E., 1995. A functional organization of dermal coe-lomic canals in Sipunculus nudus (Sipuncula) with a discussion of respi-ratory designs in sipunculans. Invert. Zool. 114, 51–63.

Saito, M., Kojima, S., Endo, K., 2000. Mitochondrial CO1 sequences ofbrachiopods: genetic code shared with protostomes and limits ofutility for phylogenetic reconstruction. Mol. Phylogenet. Evol. 15,331–344.

Savill, N.J., Hoyle, D.C., Higgs, P.G., 2001. RNA sequence evolution withsecondary structure constraints: comparison of substitution rate mod-els using maximum likelihood methods. Genetics 157, 399–411.

Scheltema, A.H., 1996. Phylogenetic position of Sipuncula, Molluscaand the progenetic Aplacophora. In: Taylor, J. (Ed.), Origin andEvolutionary Radiation of the Mollusca. Oxford University Press,pp. 53–58.

Schöninger, M., von Haeseler, A., 1994. A stochastic model and the evo-lution of autocorrelated DNA sequences. Mol. Phylogenet. Evol. 3,2–240.

Schulze, A., Cutler, E.B., Giribet, G., 2005. Reconstructing the phylogenyof the Sipuncula. Hydrobiologia 535/536, 277–296.

Shipley, A.E., 1892. On a new species of Phymosoma. Proc. Camb. Philos.Soc. 7, 77–78.

Simmons, M.P., Ochoterena, H., 2000. Gaps as characters in sequence-based phylogenetic analysis. Syst. Biol. 49, 369–381.

Simmons, M.P., Pickett, K.M., Miya, M., 2004. How meaningful are Bayes-ian support values? Mol. Biol. Evol. 21, 188–199.

Smith, S.W., Overbeek, R., Woese, C.R., Gilbert, W., Gillevet, P.M., 1994.The Genetic Data Environment: an expandable GUI for multiplesequence analysis. Comput. Appl. Biosci. 10, 671–675.

Staton, J., Rice, M.E., 1999. Genetic DiVerentiation despite teleplanic lar-val dispersal: allozyme variation in sipunculans of the Apionsoma misa-kianum species complex. Bull. Mar. Sci. 65, 467–480.

Staton, J.L., 2003. Phylogenetic analysis of the mitochondrial cytochromec oxidase subunit 1 gene from 13 sipunculan genera: intra-and inter-phylum relationships. Invert. Biol. 122, 252–264.

Stephen, A.C., Edmonds, S.J., 1972. The phyla Sipuncula and Echiura.Trustees British Mus. (Nat. Hist.), London.

SwoVord, D.L., 2003. PAUP*: Phylogenetic Analysis Using Parsimony(and other methods), version 4.0b10 for Macintosh. Sinauer, Sunder-land, MA.

Tavaré, S., 1986. Some probabilistic and statistical problems in the analysisof DNA sequences. Lect. Math. Life Sci. 17, 57–86. 368–376.

Telford, M.J., Wise, M.J., Gowri-Shankar, V., 2005. Consideration ofRNA secondary structure signiWcantly improves likelihood-basedestimates of phylogeny: examples from the Bilateria. Mol. Biol. Evol.22, 1129–1136.

Van de Peer, Y., De Rijk, P., Wuyts, J., Winkelmans, T., De Wachter, R.,2000. The European small subunit ribosomal RNA database. NucleicAcids Res. 28, 175–176.

Voss-Foucart, M.F., Barzin, S., Jeuniaux, C., Bussers, J.C., 1977. Étudecomparée de la composition chimique des regions souples et durciesde la cuticule de quatre espèces de siponculiens. Cah. Biol. Mar. 18,135–145.

Voss-Foucart, M.F., Barzin, S., Toussaint, C., 1978. Étude comparée de lacomposition en acides amines d’extraits peptidiques de la cuticule de sipon-culiens et d’annelides polychetes. Arch. Zool. Exp. Gen. 118, 457–470.

Wheeler, W.C., 1995. Sequence alignment, parameter sensitivity, and thephylogenetic analysis of molecular data. Syst. Biol. 44, 321–331.

Wheeler, W.C., 1996. Optimization alignment: the end of multiplesequence alignment in phylogenetics? Cladistics 12, 1–9.

Wheeler, W.C., 2003. Implied alignment: a synapomorphy-based multiple-sequence alignment method and its use in cladogram search. Cladistics19, 261–268.

Wheeler, W.C., 2006. Dynamic hom*ology and the likelihood criterion. Cla-distics 22, 157–170.

Wheeler, W.C., Gladstein, D., DeLaet, J., 2004–2006. POY: the optimiza-tion of alignment characters, version 3.0. American Museum of Natu-ral History. Program and documentation available at ftp.amnh.org/pub/molecular.

Whiting, M.F., Carpenter, J.M., Wheeler, Q.D., 1997. The Strepsiptera prob-lem: phylogeny of the holometabolous insect orders inferred from 18Sand 28S ribosomal DNA sequences and morphology. Syst. Biol. 46, 1–68.

Zarkihk, A., 1994. Estimation of evolutionary distances between nucleo-tide sequences. J. Mol. Evol. 39, 315–329.

ftp://ftp.amnh.org/pub/molecular

(PDF) Phylogeny of sipunculan worms: A combined analysis of four gene regions and morphology - DOKUMEN.TIPS (2024)