Musculature in Sipunculan Worms: Ontogeny and Ancestral States

Home , Golfingia vulgaris, Phascolionidae, Phascolosomatidae, Phascolosoma granulatum, Themiste (worm)

EVOLUTION & DEVELOPMENT 11:1, 97–108 (2009) DOI: 10.1111/j.1525-142X.2008.00306.x

Musculature in sipunculan worms: ontogeny and ancestral states

Anja SchulzeÃ and Mary E. Rice Smithsonian Marine Station, 701 Seaway Drive, Fort Pierce, FL 34949, USA ÃAuthor for correspondence (email: [email protected]). Present address: Department of Marine Biology, Texas A & M University at Galveston, 5007 Avenue U, Galveston, TX 77551, USA.

SUMMARY Molecular phylogenetics suggests that the introvert retractor muscles as adults, go through devel- Sipuncula fall into the Annelida, although they are mor- opmental stages with four retractor muscles that are phologically very distinct and lack segmentation. To under- eventually reduced to a lower number in the adult. The stand the evolutionary transformations from the annelid to the circular and sometimes the longitudinal body wall musculature sipunculan body plan, it is important to reconstruct the are split into bands that later transform into a smooth sheath. ancestral states within the respective clades at all life history Our ancestral state reconstructions suggest with nearly 100% stages. Here we reconstruct the ancestral states for the head/ probability that the ancestral sipunculan had four introvert introvert retractor muscles and the body wall musculature in retractor muscles, longitudinal body wall musculature in bands the Sipuncula using Bayesian statistics. In addition, we and circular body wall musculature arranged as a smooth describe the ontogenetic transformations of the two muscle sheath. Species with crawling larvae have more strongly systems in four sipunculan species with different de- developed body wall musculature than those with swimming velopmental modes, using F-actin staining with fluo- larvae. To interpret our findings in the context of annelid rescent-labeled phalloidin in conjunction with confocal laser evolution, a more solid phylogenetic framework is needed for scanning microscopy. All four species, which have smooth the entire group and more data on ontogenetic trans- body wall musculature and less than the full set of four formations of annelid musculature are desirable.

INTRODUCTION roots of the respective clades. Whenever possible this should encompass all life history stages, because some stages may Sipuncula have long been regarded as a distinct protostome retain ancestral states more readily than others. This is ex- phylum (Hyman 1959; Stephen and Edmonds 1972; Cutler emplified in the Echiura, in which only the larvae retain a 1994). Most analyses have placed them in the lophotrochozoan segmented nervous system, indicative of their annelid origins clade (Zrzavyét al. 1998; Giribet et al. 2000; Peterson and (Hessling and Westheide 2002). Ontogenetic transformations Ernisse 2001; Passamaneck and Halanych 2006), but until re- can also be informative characters themselves. cently relationships within the lophotrochozoans were largely In the Sipuncula, the body wall musculature and the in- unresolved. Today there is a growing consensus that sipuncul- trovert retractor muscles are important taxonomic characters. ans fall into the annelids (Boore and Staton 2002; Bleidorn The longitudinal and circular body wall muscles are either et al. 2005; Struck et al. 2007; Dunn et al. 2008), sparking a arranged in continuous sheaths or are broken up into nu- renewed interest in their morphology and development. merous bands. The number of introvert retractor muscles in As adults, their body is divided into a trunk and a re- the adults varies from one to four. They originate in the body tractable introvert with a crown of tentacles at its anterior wall of the trunk and insert in the head region. end, showing little resemblance to any known annelid group. Here we reconstruct the ancestral states of the two muscle More similarities to annelids are apparent in the larval stages, systems using Bayesian statistics. For this purpose, we rean- such as the trochophore larva with prototrochal and meta- alyze a simplified dataset of sipunculan sequence and mor- trochal ciliary bands, the retention of the egg envelope to phological data previously generated by the first author and form the larval cuticle and the paired ventral nerve cord in the collaborators (Schulze et al. 2007). In addition, we analyze pelagosphera larva of several species (Rice 1985). On the the ontogenetic transformations of the musculature using other hand, the absence of morphological segmentation in phalloidin-staining of F-actin and confocal laser scanning any life history stage is striking (Wanninger et al. 2005). microscopy. As these transformations may depend on the In general, to understand the morphological transforma- developmental mode of the species, we studied them in tions from one body plan to another, it is important to de- four sipunculan species displaying different developmental termine the ancestral states, that is character states at the patterns.

& 2009 The Author(s) 97 Journal compilation & 2009 Wiley Periodicals, Inc. 98 EVOLUTION & DEVELOPMENT Vol. 11, No. 1, January^February 2009

I The head retractor muscles eventually transform into the introvert retractor muscles of the juvenile. The four species we chose for the present study all have less II than the full set of four introvert retractor muscles and continuous layers of circular and longitudinal body wall musculature. Phascolion cryptum (Fig. 2A) displays developmental III mode I (Rice 1975a). In the adult stage, a single introvert retractor muscle is present with two separate roots at the

e a

Fig. 1. The four developmental modes in the Sipuncula. I. Direct np development. II. Indirect development with a single pelagic stage, ir the lecithotrophic trochophore. III. Indirect development with a A lecithotrophic trochophore and a lecithotrophic pelagosphera. IV. Indirect development with a lecithotrophic trochophore and a planktotrophic pelagosphera. (Modiﬁed from Rice 1975a, b.) ee B Rice (1967, 1975a, b, 1976) describes four distinct develop- i mental patterns in the Sipuncula (Fig. 1): (1) direct development; (2) one pelagic larval stage: lecithotrophic trochophore; j (3) two pelagic larval stages: lecithotrophic trochophore and lecithotrophic pelagosphera; (4) two pelagic larval stages: lecithotrophic trochophore and planktotrophic pelagosphera. The majority of sipunculan species of which development C has been studied, develop according to mode IV. The lecithotrophic trochophore stage is generally short lived, but the pelagosphera may remain pelagic for 4–8 months (Scheltema and Hall 1975). Pelagosphera larvae swim by means of a dr strongly developed metatrochal ciliary band. When disturbed, vr their head region, including the metatroch, completely retracts into the trunk by a contraction of the head retractor muscles. D E

Fig. 2. Developmental stages of Phascolion cryptum; (A) Introvert of adult emerging from gastropod shell, macrophotography; (B) Drawing of interior anatomy of adult; note single introvert retractor muscle with two roots (arrows); (C) scanning electron micrograph of juvenile in the process of elongation; 36 h post fertilization (p.f.); egg envelope still clearly visible in anterior part; (D–G) juvenile stages, confocal laser scanning projections after fluorescent staining for F-actin with phalloidin Alexa 488. (D) Juvenile, 36 h p.f., dorsal slice. (E) Same specimen as (D), ventral slice. (F) Crawling stage, dorsal view, 3 days p.f.; note distinct bands of cm longitudinal and circular body wall musculature. (G) Same spec- vr imen, interior slice; note four head retractor muscles. a, anus; cm, circular body wall musculature; dr, dorsal head retractor muscles; dr e, esophagus; ee, egg envelope; ir, introvert retractor muscle; j, lm juvenile; lm, longitudinal body wall musculature; nc, nerve cord; F G np, nephridium; vr, ventral head retractor muscles. Scale bars: 20 mm. Schulze and Rice Myogenesis and phylogeny in sipuncula 99 posterior end of the trunk (Fig. 2B). Themiste lageniformis MATERIAL AND METHODS (Fig. 3A) develops according to mode III (Pilger 1987). Its pelagosphera larva is capable of swimming short distances but Ancestral state reconstruction Ancestral state reconstruction was performed by reanalyzing a spends most of its time crawling on the bottom. The adult has simplified version of the dataset first presented by Schulze et al. two introvert retractor muscles originating in the posterior (2007), with a constrained topology, using Bayesian statistics. The third of the trunk (Fig. 3B). Phascolion psammophilum (Fig. analysis included the same four gene regions (18S ribosomal RNA, 4A) also displays developmental mode III (Rice 1993a), but 28S ribosomal RNA, Histone H3, and cytochrome c oxidase sub- its pelagosphera larva is purely pelagic. As an adult, the spe- unit I) (Table 1) and 58 morphological characters and utilized the cies has a single large dorsal retractor muscle and a weaker same mixed models for the different data partitions. One difference ventral muscle with two separate roots (Fig. 4B). The fourth to the original analysis was that due to computing constraints only species, Nephasoma pellucidum (Fig. 5A) develops according 1,000,000 generations of Monte Carlo Markov chains were per- to mode IV (M. E. Rice, personal observation). It has two formed instead of 1,500,000. Of the 1,000,000 generations the initial introvert retractor muscles as an adult (Fig. 5B). 500,000 were discarded as burn-in. Another change from the pre- vious analysis was that for most species only single representatives were included in the analysis, except for two species that clearly t appeared to be polyphyletic in the initial analysis. The two species were Aspidosiphon parvulus and Phascolosoma granulatum (see ey comments on their status in Schulze et al. 2007). a e Ancestral states were reconstructed for the root node of all ir dr Sipuncula. To accomplish ancestral state reconstruction, the topology was constrained to render the Sipuncula monophyletic. The np np high support for sipunculan monophyly in all recent analyses (Maxmen et al. 2003; Staton 2003; Schulze et al. 2005, 2007) jus- C tified constraining the topology. Ancestral states were only reconstructed for the morphological characters. AB

i vr Collecting adults Adults of the four species were collected in March and April 2003. P. cryptum inhabits abandoned gastropod shells, which were sieved from intertidal sand in the vicinity of the Harbor Branch Ocean- pt D ographic Institution, Fort Pierce, FL. Specimens of T. lageniformis mt m were retrieved from the crevices of oyster shells around Jack Island, Indian River Lagoon, near the Fort Pierce Inlet. P. psammophilum, an interstitial species, was collected 6 miles offshore from Fort Pierce, using sediment dredges on the R/V SUNBURST and cm hr tr Fig. 3. Developmental stages of Themiste lageniformis. (A) Adult, macrophotography; (B) Drawing of interior anatomy; note two introvert retractor muscles; (C, D, F, H, I) Confocal laser scanning to projections after fluorescent staining with phalloidin Alexa 488, (E, G) Scanning electron micrographs. (C) Early pelagosphera, 37 h EF post fertilization (p.f.), dorsal slice. (D) Same specimen as (C), ventral slice. (E) Pelagosphera larva, 5 days p.f., lateral view. (F) Pelagosphera larva, 5 days p.f.; note strongly developed circular body wall musculature, lateral view. (G) Juvenile, 15 days p.f., head retracted, anterior is up; (H) Juvenile, 20 days p.f., head retracted. vr Two dorsal head retractors are strongly developed. At least one ventral retractor is still present but less strongly developed than dorsal retractors. (I) Juvenile, 20 days p.f. No sign of the ventral head retractors is detectable in this specimen. Arrows indicate dr dorsal retractors; circular body wall musculature less clearly organized in bands. a, anus; cm, circular body wall musculature; dr, dorsal head retractor muscles; e, esophagus; ey, eye; hr, head retractor muscles; ir, introvert retractor muscles; m, mouth; mt, metatroch; np, nephridium; pt, prototroch; t, tentacles; to, terminal GHI organ; tr, terminal organ retractor muscle; vr, ventral head retractor muscles. Scale bars: 50 mm. 100 EVOLUTION & DEVELOPMENT Vol. 11, No. 1, January^February 2009 subsequent sieving of the sediment. Specimens of N. pellucidum tory, generally after changing the water. Whenever eggs were ob- were collected from rubble (including mollusc shells, sand dollar served in the culture dishes, they were pipetted into a clean dish and tests, and rocks), 4 miles offshore from Fort Pierce, FL, using an observed for development. Larval cultures were kept for up to six echinoderm dredge on the R/V SUNBURST. weeks. Water was changed at least every 2 days. The larvae were periodically fed with unicellular algae or diatoms (Isochrysis, Dunaliella,orNanochloropsis). Spawning and rearing larvae Multiple adults of each species were kept in glass dishes in ap- Sample preparation for confocal laser scanning proximately 200 ml seawater at room temperature. None of the species shows sexual dimorphism. T. lageniformis reproduces pri- microscopy marily parthenogenetically (Pilger 1987), whereas the other species Live specimens were relaxed by floating menthol dissolved in 95% 1 display bisexual reproduction. Spawning occurred in the labora- ethanol on the surface of a shallow dish with seawater at 4 Cfor 20 min. They were fixed for up to 24 h in a 4% solution of pa- raformaldehyde in 0.1 M So¨rensen’s phosphate buffer (PBS) (Clark et al. 1981) at 41C. Samples were then washed in PBS and per- a t meabilized for 1 h in a solution of 0.2% Triton X-100 in PBS (PBT). Fluorescent staining was accomplished in 1.5 ml microcen- trifuge tubes wrapped in aluminum foil using a 1:20 solution of Phalloidin Alexa 488 in PBT. Staining times were either 1 h at h room temperature or 12–24 h at 41C. Following staining, specimens np were washed in PBS and transferred to 30% isopropanol. They i were then pipetted onto poly-L-lysine-coated microscope slides to dr which they attached. The slides with the attached specimens were taken through a dehydration series of isopropanol (30%, 50%, 70%, 95%, and 100%). Specimens were cleared and mounted AB nonpermanently with Murray’s clear (2:1 benyl benzoate:benzyl acohol) and viewed with a Radiance 2100 confocal laser scanning vr microscope. Z-series of optical slices were produced for each specimen with slices every 0.1–0.5 mm, depending on the specimen. Im- age projections were generated in Volocity version 3.6.1 for MacIntosh (improvision) or Confocal Assistant version 4.02 (Brelje 1994–1996).

Sample preparation for scanning electron microscopy Specimens were relaxed as described above. They were fixed in 2.5% glutaraldehyde in Millonig’s phosphate buffer (Millonig C 1964) for at least 12 h at 41C. Fixation was followed by three washes in a 1:1 mixture of Millonig’s phosphate buffer and 0.6 M cm sodium chloride and postfixation in 1% osmium tetroxide (1:1:2 mix of 4% OsO4:Millonig’s buffer:0.75 M NaCl). Samples were then dehydrated in an ethanol series up to 100% and critical point

Fig. 4. Developmental stages of Phascolion psammophilum,(A) D Scanning electron micrograph of adult introvert and tentacles; (B) Drawing of interior anatomy of adult; note single large dorsal introvert retractor and smaller ventral introvert retractor muscle; vr (C, D, E) Confocal laser scanning projections after ﬂuorescent staining with phalloidin Alexa 488. (C) Late trochophore, dorsal view, 24 h post fertilization (p.f.). Note four head retractor muscles dr (arrows). (D) Pelagosphera larva, 2 days, dorsal view. Four head retractor muscles of similar size are obvious in anterior body region (arrows). Circular body wall musculature in bands. (E) Pelagosph- era larva 5 days (contracted). Dorsal pair of head retractor muscles E much more strongly developed than ventral (only one ventral retractor observed at this stage). a, anus; cm, circular body wall musculature; dr, dorsal head retractor muscles; h, hooks; i, intestine; np, nephridium; t, tentacles; vr, ventral head retractor muscles. Scale bars: (A) 200 mm, (B-F) 50 mm. Schulze and Rice Myogenesis and phylogeny in sipuncula 101

dried. Specimens were mounted on SEM stubs using double-sticky t tape and viewed in a JEOL 6400 Visions scanning electron micro- np np scope (Peabody, MA, USA). All images were stored digitally. a no m RESULTS ir i hr Ancestral state reconstruction nc a tr Ancestral states could be reconstructed with nearly 100% A B C probability for the musculature-related characters. The average probability over 1000 sampled trees (500,000 generations of Monte Carlo Markov chains in a Bayesian analysis with every 500th tree sampled) for the longitudinal body wall musculature to be arranged in distinct bands at the root node of the sipunculan tree is 97.5%. With an average probability m bo of 98.1% the circular body wall musculature was smooth. cm With 98.1% probability, the ancestral sipunculan had two pairs of introvert retractor muscles. vr dr The phylogenetic analysis with the constrained topology a resulted in a very similar tree (Fig. 6) to the one presented in Schulze et al. (2007). The same ﬁve clades were recovered, DEFtr except that Phascolosoma capitatum falls into clade VI instead of V as in the original analysis.

m P. cryptum mt The ontogenetic transformations of the musculature of bo P. cryptum are illustrated in Fig. 2. P. cryptum omits both the trochophore and the pelagosphera stage. A crawling, cm vr highly contractile juvenile emerges from the egg envelope (Fig. dr dr 2C) approximately 36 h post fertilization (p.f.). Although partially obscured by yolk, a dorsal and a ventral pair of introvert retractor muscles can be distinguished(Fig.2,DandE).At 48 h p.f. the juvenile crawls on the bottom of the dish by a contractions of its body wall musculature. At this stage, both to GHtr the circular and the longitudinal body wall musculature are strongly developed and organized in bands (Fig. 2F). Four separate retractor muscles are apparent at this stage (Fig. 2G). Fig. 5. Developmental stages of Nephasoma pellucidum; (A, D, G) scanning electron micrographs; (B) Drawing of interior anatomy of T. lageniformis adult; note two introvert retractor muscles; (C, E, F, H) confocal laser scanning projections after ﬂuorescent staining for F-actin with Figure 3 shows the development of T. lageniformis,which phalloidin Alexa 488. (A) Anterior introvert of adult, showing lacks the trochophore stage. Although the early pelagosphera arrangement of the tentacles. (C) Late trochophore stage, 40 h post at 37 h p.f. has four head retractors (Fig. 3, C and D), this is fertilization (p.f.) (see supporting information for rotational view no longer the case at 5 days p.f. The pelagosphera at 5 days which clearly shows four separate head retractor muscles). (D) p.f. (Fig. 3, E and F) shows weakly deleloped prototroch and Early pelagosphera, 40 h p.f., ventral view. (E) Early pelagosphera, 40 h p.f., ventral slice of larva. (F) Same specimen as E, dorsal slice metatroch but very strongly developed body wall muscula- of larva. (G) Pelagosphera, 5 days p.f., ventral view. (H) Pelagosph- ture, especially the circular body wall musculature which era, 5 days p.f., ventrolateral view. a, anus; bo, buccal organ; cm, shows an organization in broad bands (Fig. 3F). Dorsal and circular body wall musculature; dr, dorsal head retractor muscles; ventral head retractor muscles cannot clearly be distinguished hr, head retractor muscles; i, intestine; ir, introvert retractor mus- at this stage. The retractor musculature consists of loosely cles; m, mouth; mt, metatroch; nc, nerve cord; no, nuchal organ; np, nephridium; t, tentacles; to, terminal organ; tr, terminal organ arranged individual ﬁbers that reach from the head region to retractor muscles; vr, ventral head retractor muscles; scale bars: about 3/4 toward the posterior end where they insert into the 20 mm. body wall. At about 15 days the larva loses its ability to swim and starts to undergo metamorphosis. Most notably, the head 102 EVOLUTION&DEVELOPMENT Vol. 11, No. 1, January^February 2009

Table 1. GenBank accession numbers for all sipunculan and outgroup sequences included in the phylogenetic analysis and reconstruction of ancestral states

Species MCZ catalogue # 18S rRNA 28S rRNA Histone H3 COI Sipunculidae Phascolopsis gouldii DNA100199 AF123306 AF519272 AF519297 DQ300134 Siphonosoma cumanense DNA100991 DQ300002 DQ300047 DQ300089 DQ300157 Siphonosoma vastum DNA100625 DQ300003 AY445137 AY326297 DQ300158 Sipunculus (S.) norvegicus DNA101069 DQ300004 DQ300090 DQ300159 Sipunculus (S.) nudus DNA100246 AF519240 AF519270 AF519295 DQ300161 Sipunculus (S.) phalloides DNA101337 DQ300009 DQ300094 DQ300165 Sipunculus (S.) polymyotus DNA101121 DQ300010 DQ300095 DQ300166 Xenosiphon branchiatus DNA101086 DQ300016 DQ300050 DQ300101 DQ300172 Golfingiidae Golfingia elongata DNA100465 DQ299969 DQ300031 DQ300065 DQ300121 Golfingia margaritacea DNA100738 DQ299973 DQ300032 DQ300069 DQ300126 Golfingia vulgaris DNA100207 AF519244 AF519273 DQ300127 Nephasoma diaphanes DNA101072 DQ299975 DQ300071 DQ300128 Nephasoma flagriferum DNA100440 DQ299976 DQ300033 DQ300072 DQ300129 Nephasoma pellucidum DNA101009 DQ299978 DQ300131 Thysanocardia catherinae DNA101068 DQ300015 DQ300099 Thysanocardia nigra DNA100606 AF519247 AF519274 DQ300100 Themistidae Themiste (T.) dyscrita DNA101095 DQ300011 DQ300167 Themiste (T.) hennahi DNA100627 DQ300012 DQ300096 DQ300168 Themiste (L.) lageniformis DNA100229 AF519249 AF519276 AF519302 DQ300169 Themiste (L.) minor DNA101083 DQ300013 DQ300097 DQ300170 Themiste (T.) pyroides DNA101084 DQ300014 DQ300098 DQ300171 Phascolionidae Onchnesoma steenstrupii DNA101080 DQ299979 DQ300034 DQ300074 Phascolion (L.) cryptum DNA101007 DQ299980 DQ300035 DQ300075 DQ300132 Phascolion (I.) gerardi DNA101002 DQ299981 DQ300076 Phascolion (P.) psammophilum DNA101006 DQ299982 DQ300036 DQ300133 Phascolion (P.) strombus DNA100101 AF519248 AF519275 AF519301 Phascolosomatidae Antillesoma antillarum DNA101008 DQ299951 DQ300051 DQ300102 Apionsoma (A.) misakianum DNA100737 DQ299952 DQ300017 DQ300052 DQ300103 Apionsoma (A.) murinae DNA100446 DQ299953 DQ300018 Apionsoma (E.) pectinatum DNA100624 AY326293 AY445142 AY326300 DQ300104 Phascolosoma (P.) agassizii DNA101096 DQ299985 DQ300037 DQ300078 DQ300135 Phascolosoma (P.) albolineatum DNA100396 AF519251 AF519278 DQ300136 Phascolosoma (F.) capitatum DNA101070 DQ299986 DQ300079 DQ300137 Phascolosoma (P.) granulatum DNA100201 AF519252 AF519279 AF519304 DQ300138 Phascolosoma (P.) granulatum X79874 Phascolosoma (P.) nigrescens DNA100822 DQ299988 DQ300039 DQ300081 DQ300141 Phascolosoma (P.) noduliferum DNA100208 AF519253 AF519280 AF519305 DQ300144 Phascolosoma (P.) perlucens DNA100228 AF519254 AF519281 AF519306 DQ300145 Phascolosoma (P.) scolops DNA100813 DQ299998 DQ300044 DQ300085 DQ300152 Phascolosoma (P.) stephensoni DNA100469 AF519256 AF519283 AF519310 DQ300153 Phascolosoma (P.) turnerae DNA100230 DQ300000 DQ300046 DQ300087 DQ300154 Aspidosiphonidae Aspidosiphon (A.) albus DNA101017 DQ299954 DQ300053 DQ300105 Aspidosiphon (A.) elegans DNA101016 DQ299957 DQ300020 DQ300056 DQ300106 Aspidosiphon (P.) fischeri DNA100620 AY326294 AY326301 DQ300107 Aspidosiphon (A.) gosnoldi DNA101014 DQ299959 DQ300022 DQ300057 DQ300109 Aspidosiphon (A.) gracilis schnehageni DNA101087 DQ299960 DQ300023 DQ300058 DQ300110 Aspidosiphon (P.) laevis DNA100467 AF519261 DQ300024 DQ300059 DQ300111 Aspidosiphon (A.) misakiensis DNA100205 AF119090 AF519288 AF519312 Aspidosiphon (A.) muelleri DNA100206 DQ299962 DQ300025 DQ300060 DQ300113 Schulze and Rice Myogenesis and phylogeny in sipuncula 103

Table 1. (Contd.)

Species MCZ catalogue # 18S rRNA 28S rRNA Histone H3 COI Aspidosiphon (P.) parvulus DNA100202 AF119075 DQ300026 DQ300061 Aspidosiphon (P.) parvulus DNA100982 DQ299964 DQ300027 DQ300063 DQ300115 Aspidosiphon (P.) steenstrupii DNA100232 AF519262 AF519291 AF519315 DQ300116 Cloeosiphon aspergillus DNA100825 DQ299968 DQ300030 DQ300120 Lithacrosiphon cristatus DNA100623 AY326295 AY445142 AY326302 Nemertea Amphiporus sp. AF119077 AF519265 AF519293 AJ436899 Argonemertes australiensis AF519235 AF519264 AF519293 AY428840 Mollusca Lepidopleurus cajetanus AF120502 AF120565 AY070142 AF120626 Rhabdus rectius AF120523 AF120580 AY070144 AY260826 Yoldia limatula AF120528 AF120585 AY070149 AF120642 Annelida Lumbrineris latreilli AF519238 AF519267 AF185253 AY364855 Lamellibrachia spp. AF168742 AF185235 U74055 Owenia fusiformis AF448160 AY428824 AY428832 AY428839 Urechis caupo AF119076 AF519268 X58895 U74077 Lumbricus terrestris AJ272183 F AF185262 NC_001673 Entoprocta Loxosomella murmanica AY218100 AY218129 AY218150 AY218083 is mostly retracted and the metatroch is lost (Fig. 3G). At 20 planktotrophic pelagosphera (Fig. 5). The introvert retractor days p.f., two strongly developed dorsal head retractors are muscles and the terminal organ retractor muscles start form- clearly visible (Fig. 3, H and I). In one case, a less strongly ing in the late trochophore, ca. 40 h p.f. In the early developed ventral retractor muscle was present (Fig. 3H). The pelagosphera, a ventral and a dorsal pair of retractor mus- density of fibers in the body wall musculature has increased cles can be distinguished (Fig. 5C, also see animation in sup- and an arrangement into bands is much less obvious than in porting information). The body wall musculature is indistinct. earlier stages. At 40 h p.f. (Fig. 5, D–F), the two pairs of retractor muscles are still discernible, although individual fibers from the dorsal P. psammophilum and ventral muscle on each side have almost approached each other. Similar to P. psammophilum, the circular body wall The development of this species, with a trochophore and a musculature forms rings and is most strongly developed at the lecithotrophic pelagosphera isshowninFig.4.Inthelate constriction between head and trunk region (Fig. 5, E and F). trochophore, approximately 24 h p.f., a dorsal and a ventral In the later stage pelagosphera of N. pellucidum (Fig.5,Gand pair of retractor muscles are visible (Fig. 4C). Some circular H) the terminal organ and the associated musculature are body wall musculature is discernible in the central part of the very strongly developed (Fig. 5H). larva but is still rudimentary and not clearly in bands. At 2 days p.f. (Fig. 4D) the circular body wall musculature is most strongly developed in the constriction between head and trunk and in the posterior area of the trunk. Some longitu- DISCUSSION dinal muscle fibers are present in the body wall. The four Ancestral state reconstruction retractor muscles are still clearly distinguishable and of ap- Traditionally, parsimony has been the only method available proximately equal size. At five days p.f. (Fig. 4E) the number to reconstruct ancestral states but the results can depend on of individual muscle fibers has increased greatly in the dorsal the preference settings in the reconstruction algorithm and pair of retractor muscles. We have not observed how the leave no room for uncertainty. In case of ambiguities, the number of dorsal retractor muscles is reduced to one. The ‘‘accelerated transformation’’ algorithm (ACCTRAN in circular body wall musculature is still discernible in rings. PAUPÃ; Swofford 2003) places character state changes close to the root of the tree whereas ‘‘delayed transformation’’ N. pellucidum algorithm (DELTRAN in PAUPÃ) places them closer to the N. pellucidum shows the most common developmental mode leaves. The probabilities calculated with Bayesian statistics, as in sipunculans, with a lecithotrophic trochophore and a applied in this study, allow a better judgement of the degree of 104 EVOLUTION & DEVELOPMENT Vol. 11, No. 1, January^February 2009

Loxosomella Owenia Lepidopleurus 78 Yoldia 81 Rhabdus Argonemertes * * Amphiporus Lumbricus Urechis Lumbrineris * Lamellibrachia 95 66 Sipunculus norvegicus 101069 Clade I: Sipunculus nudus 100246 IRM - 4 87 Xenosiphon branchiatus 101086 * LMB - y 66 Sipunculus phalloides 101337 93 Sipunculus polymyotus 101121 CMB - y * Siphonosoma cumanense 100991 Clade II: IRM - 4, LMB - y; CMB - y * Siphonosoma vastum 100625 Phascolosoma turnerae 100230 * Phascolion gerardi 101002 72 Nephasoma pellucidum 101009 Phascolion psammophilum 101006 72 Phascolion cryptum101007 Onchnesoma steenstrupii 101080 59 66 Phascolion strombus 100101 * Nephasoma diaphanes 101072 Nephasoma flagriferum 100440 Golfingia elongata 100465 72 Golfingia vulgaris 100207 Clade III: 79 * Phascolopsis gouldii 100199 IRM - 1-4 52 Apionsoma murinae 100446 Golfingia margaritacea 100738 LMB - n* Thysanocardia catherinae 101068 CMB - n 65 * * Thysanocardia nigra 100606 Themiste lageniformis 100229 89 *except P. gouldii * Themiste minor 101083 Themiste hennahi 100627 * Themiste dyscrita 101095 * 70 Themiste pyroides 101084 Phascolosoma capitatum 101070 Phascolosoma granulatum 100201 Phascolosoma agassizii 101096 Clade IV: Phascolosoma nigrescens 100822 IRM - 4 65 * * Phascolosoma granulatum x79874 * Phascolosoma stephensoni 100469 LMB - y* 99 * Phascolosoma noduliferum 100208 CMB - n 72 Phascolosoma scolops 100813 * * Phascolosoma albolineatum 100396 *except P. capitatum 63 Phascolosoma perlucens 100228 Apionsoma misakianum 100737 Apionsoma pectinatum 100624 79 Cloeosiphon aspergillus 100825 Antillesoma antillarum 101008 * Lithacrosiphon cristatus 100623 Clade V: 68 Aspidosiphon muelleri 100206 IRM - 2-4 Aspidosiphon albus 101017 LMB - y/n * Aspidosiphon misakiensis 100205 * CMB - n 63 98 Aspidosiphon steenstrupii 100232 98 Aspidosiphon parvulus 100202 Aspidosiphon laevis 100467 * Aspidosiphon parvulus 100982 Aspidosiphon gosnoldi 101014 * Aspidosiphon gracilis 101087 99Aspidosiphon elegans 101016 * Aspidosiphon fischeri 100620 0.1

Fig. 6. 50% Majority rule consensus tree resulting from the Bayesian analysis of four gene regions and morphology. Shaded boxes outline the five major clades. Branch support is given as Bayesian posterior probability. Asterisks: 100% posterior probability. Black box indicates the root node of the Sipuncula for which the ancestral states were reconstructed. IRM, number of introvert retractor muscles; CMB, cirucular muscle bands (yes/no); LMB, longitudinal muscle bands (yes/no). confidence in the reconstructions. Here the ancestral states of the only unambiguously identified fossils of sipunculans the musculature at the sipunculan root node could be recon- are uninformative with regard to musculature, although structed with nearly 100% posterior probability. parts of the internal anatomy are well preserved (Huang Direct evidence from fossils would be the most reliable et al. 2004). No fossils of sipunculan larvae have ever been source of information about ancestral states. However, discovered. Schulze and Rice Myogenesis and phylogeny in sipuncula 105

The re-analysis of the simplified dataset from Schulze et al. dorsal retractors might actually be a pair of protractor mus- (2007) resulted in a very similar tree as the analysis with the cles. Scheltema and Rice (1990) describe a very similar, pos- full dataset. The placement of P. capitatum (clade V in sibly the same, larval type, named ‘‘leura’’ type ‘‘i’’ and only Schulze et al. 2007, clade IV in the present analysis) is the only mention one pair of dorsal and ventral retractors each. difference with regard to the composition of the major clades, P. cryptum belongs to the subgenus Lesenka which is but it has weak support in both of the analyses (56% and characterized by a single retractor muscle with two origins in 65% posterior probability, respectively). The clades that the posterior body wall. In the adult of P. cryptum both or- branch off early in the Sipuncula (clades I and II) are com- igins are to the left of the ventral nerve cord (Fig. 2B). Four posed of species with four introvert retractor muscles and retractors are present in the juvenile until at least 3 days p.f. circular and longitudinal muscle bands as adults. All of the (Fig.2,D,E,andG).Wehavenotobservedwhenfusionor more derived clades have a continuous sheath of circular body loss of muscles occurs. wall musculature. Within Clade III there is variation in the In T. lageniformis, it appears that the ventral pair of re- number of introvert retractors, but the body wall musculature tractors is reduced (Fig. 3, H and I) and the dorsal pair is generally organized in smooth layers, with the exception of transforms into the adult introvert retractors. Phascolopsis gouldii. In this species, the longitudinal body wall P. psammophilum belongs to the subgenus Phascolion, musclulature is split into anastomosing bands, that is bands characterized by a single strong dorsal retractor muscle and a that may fuse to neighboring bands at intervals throughout more weakly developed ventral retractor with two separate the length of the trunk. Members of clade IV consistently origins in the body wall (Fig. 4B). We have shown that in the have four introvert retractor muscles. All except P. capitatum larvae of this species the head retractor muscles are of equal have longitudinal muscle bands. Members of clade V vary in size until the early pelagosphera stage (Fig. 4E). In later both the number of retractor muscles as well as in the ar- pelagosphera larvae, the dorsal retractors are already more rangement of the longitudinal body wall musculature. strongly developed than the ventral retractors (Fig. 4E). We Within the five clades there are some minor discrepancies have not been able to observe the process by which the num- between the analysis of the reduced and the full dataset (for ber of dorsal retractors is reduced to one. details consult Schulze et al. 2007). The support for the A˚ kesson (1958a, b) describes the larva of Phascolion monophyly of clade III is lower in this study than in the (Phascolion) strombus, which has four introvert retractor original analysis (72% pp vs. 100% pp) whereas that of clade muscles and a pair of protractor muscles. He reports that the Vishigher(79%ppvs.56%pp). reduction in the number of retractor muscles does not occur until the larvae are about 1 month old. In the 5-day-old pelagosphera of N. pellucidum all four Head/introvert retractor muscles retractor muscles are of similar size (Fig. 5H). We have not Our analysis indicates with high probability (98.1%) that the observed which muscles are reduced to yield the adult con- presence four separate introvert retractor muscles is the an- dition with a single pair of retractor muscles. cestral state for adult Sipuncula. This condition is present in all representatives of clades I and II, as well as in the earliest branches in clades III and V, represented by Phascolosoma Body wall musculature turnerae and two Apionsoma species, respectively. It is Our reconstructions suggest that at the root of the Sipuncula also present in most representatives of clade IV, except in the circular body wall musculature was a continuous sheath, Phascolosoma captitatum for which the position is still uncer- whereas the longitudinal musculature was split into bands. In tain (see above). Our findings agree with Cutler and Gibbs contrast, Cutler and Gibbs (1985) and Cutler (1994) con- (1985) and Cutler (1994) who proposed a hypothetical ances- cluded that the ancestral sipunculan had both longitudinal tral sipunculan with four introvert retractor muscles based on and circular body wall musculature organized as continuous morphological analyses. sheaths. This was based on their understanding of sipunculan We were not able to reconstruct the ancestral states for the phylogeny, which differed in several respects from our more larvae in the same way as for the adults because larval anat- recent analyses. In particular, Cutler (1994) assumed that the omy has only been studied for a limited number of sipunculan Sipunculidae (with longitudinal and circular muscles in bands) species. However, in almost all known cases, two pairs of were a derived clade and that Apionsoma (with circular mus- head retractor muscles are present in the pelagosphera larvae cles always and longitudinal muscles mostly in continuous (A˚ kesson 1958a, b; Hall and Scheltema 1975) (Table 2). The sheaths) was morphologically closest to the sipunculan root. exceptions are ‘‘type J’’ with a single pair and ‘‘type S’’ with Consequently, he assigned continuous body wall musculature three pairs (Hall and Scheltema 1975). With regard to ‘‘type to his ‘‘revised hypothetical ancestral sipunculan.’’ J,’’ Hall and Scheltema (1975) mention that additional mus- The longitudinal body wall musculature generally develops cles are sometimes present. For ‘‘type S,’’ the second pair of later than the circular body wall musculature. None of the 106 EVOLUTION & DEVELOPMENT Vol. 11, No. 1, January^February 2009

Table 2. Summary of published data on musculature in sipunculan larvae and respective adults

Longitudinal body wall Circular body wall Head/introvert musculature musculature retractors Larval type Species Larva Adult Larva Adult Larva Adult Reference Phascolion Continuous Continuous Bands Continuous 2 pairs Single dorsal 1,2 strombus (11pairof and single protractors) ventral retractor Type A Aspidosiphon sp. Continuous ? Continuous Continuous 2 pairs 1 pair 3 (‘‘Baccardia citronella’’) Type B Xenosiphon Bands Bands Continuous Bands 2 pairs 2pairs 3,4 (‘‘Smooth’’) branchiatus (11pairof (11pairof protractors) protractor muscles) Type C Apionsoma 16 bands Continuous Continuous Continuous 2 pairs 2 pairs 3 (‘‘Baccardia misakianum oliva’’) Type E Siphonosoma 18 bands ? Continuous ? 2 pairs 2 pairs 3,5 cumanense Type F ? 24 bands ? Continuous ? ? ? 3 Type J ? 22 bands ? Continuous ? 1 pair ?3 (1metatrochal retractors) Type L ? Poorly developed ? Poorly ?2pairs?3 developed (11pairof protractors) Type O ? 28 bands ? Continuous ? 2 pairs ? 3 Type P ? Bands ? Continuous ? 2 pairs ? 3 Type S Sipunculus 50 bands Bands Poorly Bands 3 pairs 2pairs 3 polymyotus developed (2 dorsal11 ventral pair)

1 Wanninger et al. (2005); 2 A˚ kesson (1958); 3 Hall and Scheltema (1975); 4 Ja¨gersten (1963); 5 Rice (1988).

early pelagic stages we examined had any longitudinal body has its evolutionary origins close to the root of the tree, does wall musculature but Table 2 shows that many later-stage not necessarily look like its ancestor; it evolved over the same pelagic pelagosphera larvae have longitudinal muscle bands, time period as other members of the clade and may have even if they have smooth longitudinal body wall musculature accumulated just as much morphological change. as adults, such as Apionsoma misakianum. We only observed longitudinal musculature in the crawling larval and juvenile stages of T. lageniformis and P. cryptum. An arrangement in bands was obvious in the 3-day-old juvenile of P. cryptum. CONCLUSIONS The circular musculature forms in bands in all examined species. Wanninger et al. (2005) observed the same in P. Even though Haeckel’s biogenetic law (‘‘ontogeny recapitu- strombus. They report that the number of circular muscle lates phylogeny’’) has been discredited, parallels between on- bands does not increase during initial growth of this species, togeny and phylogeny are not uncommon. In the case of the an indication that they are not segmental structures. Later Sipuncula, the four head/introvert retractor muscles reflect the larval stages (Table 2) and adult sipunculans rarely have cir- ancestral state of the group and are retained in nearly all of cular muscle bands. The condition is only present in adults of the pelagosphera larvae, even though adults may have a re- clades I and II. Even though these are early branches in the duced number. phylogeny, our ancestral state reconstruction indicates that The situation is more complex with regard to the body wall the condition of the circular body wall musculature is not musculature. The ancestral state for the longitudinal muscu- ancestral. This example shows that an extant group, which lature, an arrangement in separate bands, is not obvious in the Schulze and Rice Myogenesis and phylogeny in sipuncula 107 early larval stages, but is common in most of the later pelagic REFERENCES stages (Table 2) and in the crawling juvenile of P. cryptum. A˚ kesson, B. 1958a. A study of the nervous system of the Sipunculoideae, The circular body wall musculature forms early and always with some remarks on the development of the two species Phascolion in bands which later transform into a continuous sheath. Our strombi Montagu and Golfingia minuta Keferstein. Underso¨kningar o¨ver analyses indicate that the continuous sheath is the ancestral O¨resund 38: 1–249. A˚ kesson, B. 1958b. A Study of the Nervous System of the Sipunculoideae condition for adult Sipuncula, but we have no information with Some Remarks on the Development of the Two Species Phascolion about the ancestral condition in the larvae. It is possible that strombi Montagu and Golfingia minuta Keferstein.C.W.K.Gleerup, the transformation from separate bands in the larva to a Lund, Sweden. Bleidorn, C., Podsiadlowski, L., and Bartolomaeus, T. 2005. The complete continuous sheath in the adult is ancestral. mitochondrial genome of the orbiniid polychaete Orbinia latreillii The condition of the body wall musculature is also in part (Annelida, Orbiniidae) - a novel gene order for the Annelida and om- dependent on the developmental mode. Pelagic larval forms plications for annelid phylogeny. Gene 370: 96–103. Boore, J. L., and Staton, J. L. 2002. The mitochondrial genome of the generally have less strongly developed body wall musculature sipunculid Phascolopsis gouldii supports its association with Annelida than crawling larval stages. In our case, the pelagic pelagosph- rather than Mollusca. Mol. Biol. Evol. 19: 127–137. era stages of P. psammophilum and N. pellucidum have rel- Brelje, T. C. 1994-1996. Confocal Assistant. Software for 3D analysis of confocal images. Minneapolis, MN, USA. atively poorly developed body wall musculature. We find the Clark, G., Dougherty, W. J., and Kasten, F. H. 1981. General methods. In most prominently developed body wall musculature in the G. Clark (ed.). Staining Procedures. Williams & Wilkins, Baltimore, pp. benthic larva and juvenile of P. cryptum. T. lageniformis,with 1–103. Cutler,E.B.1994.The Sipuncula. Their Systematics, Biology and Evolution. a larva that both swims and crawls is intermediate. These Cornell University Press, Ithaca, NY. differences are easily explained from a functional morphology Cutler, E. B., and Gibbs, P. E. 1985. A phylogenetic analysis of higher taxa perspective: pelagic stages mostly move by ciliary action of in the phylum Sipuncula. Syst. Zool. 34: 162–173. Dunn, C. W., et al. 2008. Broad phylogenomic sampling improves reso- their metatroch, whereas benthic stages move by contractions lution of the animal tree of life. Nature 452: 745–749. of their body. Giribet,G.,Distel,D.L.,Polz,M.,Sterrer,W.,andWheeler,W.C.2000. To interpret our findings in the context of annelid phy- Triploblastic relationships with emphasis on the acoelomates and the position of Gnathostomulida, Cycliophora, Platyhelminthes, and Chate- logeny, the position of the Sipuncula within the Annelida ognatha: a combined approach of 18S rDNA sequences and morphol- needs to be determined with more certainty. There is some ogy. Syst. Biol. 49: 539–562. indication for a sister group relationship between Sipuncula Hall, J. R., and Scheltema, R. S. 1975. Comparative morphology of open- ocean pelagosphera. In M. E. Rice and M. Todorovic (eds.). Proceedings and the Terebelliformia but it is only weakly supported of the International Symposium on the Biology of the Sipuncula and (Struck et al. 2007). An increasing amount of information is Echiura. Naucno Delo Press, Kotor, pp. 183–197. available on the diversity of annelid musculature, but a solid Hessling, R., and Westheide, W. 2002. Are Echiura derived from a segmented ancestor? Immunohistochemical analysis of the nervous system phylogenetic framework necessary to determine plesiomor- in developmental stages of Bonellia viridis. Journal of Morphology 252: phic versus apomorphic states is still lacking. The longitudinal 100–113. body wall muscles are usually more developed that than the Huang,D.,Chen,J.,Vannier,J.,andSaizSalinas,J.I.2004.EarlyCam- brian sipunculan worms from southwest China. Proceedings of the Royal circular muscles, which are even absent in a number of taxa Society of London, Series B 271: 1671–1676. (Tzetlin and Filippova 2005; Purschke and Mu¨ller 2006). The Hyman, L. H. 1959. Phylum Sipunculida. In Smaller Coelomate Groups. longitudinal muscles in annelids are most commonly arranged vol. 5. McGraw-Hill, New York, pp. 610–690. Ja¨gersten, G. 1963. On morphology and behavior of pelagosphaera larvae in four to six distinct bands. The formation of a continuous (Sipunculoidea). Zoologiska Bidrag fran Uppsala 36: 27–36. sheath, as in the Clitellata, is regarded as derived (Purschke Maxmen,A.B.,King,B.F.,Cutler,E.B.,andGiribet,G.2003.Evolu- and Mu¨ller 2006). The longitudinal muscle bands in sipuncul- tionary relationships within the protostome phylum Sipuncula: a molecular analysis of ribosomal genes and histone H3 sequence data. Mol. ans are more numerous than in annelids, but more detailed Phylogenet. Evol. 27: 489–503. ontogenetic studies of the muscular system in different annelid Millonig, G. 1964. Study on the factors which influence preservation of fine groups might reveal more similarities to the Sipuncula than structure. In P. Buffa (ed.). SymposiumonElectronMicroscopy.Cons- iglio Nazionale delle Ricerce, Rome, pp. 347. obvious in the adults. Passamaneck, Y., and Halanych, K. M. 2006. Lophotrochozoan phylogeny assessed with LSU and SSU data: evidence of lophophorate polyphyly. Mol. Phylogenet. Evol. 40: 20–28. Acknowledgments Peterson, K. J., and Eernisse, D. J. 2001. Animal phylogeny and the an- This work was partially supported by a MarCraig grant at Harvard cestry of bilaterians: inferences from morphology and 18S rDNA gene University to G. Giribet and E. B. Cutler, which included a post- sequences. Evol. & Devo. 3: 170–205. doctoral fellowship to A. S. It was continued under a Smithsonian Pilger, J. F. 1987. Reproductive biology and development of Themiste lag- Marine Station fellowship to A. S. (contribution number XXX). eniformis, a parthenogenic sipunculan. Bull. Mar. Sci. 41: 59–67. We are grateful to the staff and postdoctoral fellows at the Smith- Purschke, G., and Mu¨ller,M.C.M.2006.Evolutionofbodywallmus- culature. Integrative Comparative Biol. 46: 497–507. sonian Marine Station for their technical and field support, notably Rice, M. E. 1967. A comparative study of the development of Phascolosoma J. Piraino, H. F. Reichardt and S. Santagata. Carolyn Gast is ac- agassizii, Golfingia pugettensis,andThemiste pyroides with a discussion of knowledged for Fig. 2B, 3B, and 5B and Charissa Baker for Fig. 4B. developmental patterns in the Sipuncula. Ophelia 4: 143–171. Figures2B,3B,4B,and5Bwerereproducedwithpermissionfrom Rice, M. E. 1975a. Observations on the development of six species of Rice (1993a, Fig. 2G) and Rice (1993b, Figs. 11–13). Caribbean Sipuncula with a review of development in the phylum. In 108 EVOLUTION & DEVELOPMENT Vol. 11, No. 1, January^February 2009

M. E. Rice and M. Todorovic (eds.). Proceedings of the International Staton, J. L. 2003. Phylogenetic analysis of the mitochondrial cytochrome c Symposium of Sipuncula and Echiura. Naucno Delo Press, Katar, pp. oxidase subunit 1 gene from 13 sipunculan genera: intra-and interphylum 141–160. relationships. Inv. Biol. 122: 252–264. Rice, M. E. 1975b. Sipuncula. In A. C. Giese and J. S. Pearse (eds.). Stephen, A. C., and Edmonds, S. J. 1972. The Phyla Sipuncula and Echiura. Reproduction of Marine Invertebrates. Academic Press Inc., New York, Trustees of the British Museum (Natural History), London. pp. 67–127. Struck, T. H., et al. 2007. Annelid phylogeny and the status of Sipuncula Rice, M. E. 1976. Larval development and metamorphosis in Sipuncula. and Echiura. BMC Evol. Biol. 7: 57. Ã Am. Zool. 16: 563–571. Swoﬀord, D. L. 2003. PAUP : Phylogenetic Analysis Using Parsimony (and Rice, M. E. 1985. Sipuncula: developmental evidence for phylogenetic in- Other Methods). Sinauer Associates, Sunderland. ference. In S. C. Morris, J. D. George, R. Gibson, and H. M. Platt (eds.). Tzetlin, A. B., and Filippova, A. V. 2005. Muscular system in polychaetes The Origins and Relationships of Lower Invertebrates.OxfordUniversity (Annelida). Hydrobiologia 535/536: 113–126. Press, Oxford, pp. 274–296. Wanninger, A., Koop, D., Bromham, L., Noonan, E., and Degnan, B. M. Rice, M. E. 1988. Observations on development and metamorphosis of 2005. Nervous and muscle system development in Phascolion strombus. Siphonosoma cumanense with comparative remarks on Sipunculus nudus Dev. Genes Evol 215: 508–518. (Sipuncula, Sipunculidae). Bull.Mar.Sci.42: 1–15. Zrzavy´,J.,Mihulka,S.,Kepka,P.,andBezdek, A. 1998. Phylogeny of the Rice, M. E. 1993a. Two new species of Phascolion (Sipuncula: Phascoli- Metazoa based on morphological and 18s ribosomal DNA evidence. onidae) from tropical and subtropical waters of the central western At- Cladistics 14: 249–285. lantic. Proc. Biol. Soc. Washington 106: 591–601. Rice,M.E.1993b.Sipuncula.InF.W.HarrisonandM.E.Rice(eds.). Microscocopic Anatomy of Invertebrates, Vol. 12: Onychophora, Chilo- poda and Lesser Protostomata. Wiley-Liss, New York, pp. 237–325. SUPPORTING INFORMATION Scheltema, R. S., and Hall, J. R. 1975. The dispersal of pelagosphera larvae by ocean currents and the geographical distribution of sipunculans. In Additional Supporting Information may be found in the on- M. E. Rice and M. Todorovic (eds.). Proceedings of the International Symposium on the Biology of the Sipuncula and Echiura.NaucnoDelo line version of this article: Press, Kotor, pp. 103–116. Scheltema, R. S., and Rice, M. E. 1990. Occurrence of teleplanic pelagosph- Movie 1. era larvae of sipunculans in tropical regions of the Paciﬁc and Indian oceans. Bull. Mar. Sci. 47: 159–181. Movie 2. Schulze, A., Cutler, E. B., and Giribet, G. 2005. Reconstructing the phy- Please note: Wiley-Blackwell are not responsible for the logeny of the Sipuncula. Hydrobiologia 535/536: 277–296. content or functionality of any supporting materials supplied Schulze, A., Cutler, E. B., and Giribet, G. 2007. Phylogeny of sipunculan worms: a combined analysis of four gene regions and morphology. Mol. by the authors. Any queries (other than missing material) Phylogenet. Evol. 42: 171–192. should be directed to the corresponding author for the article.