[Palaeontology, 2013, pp. 1–15] ONTOGENY, MORPHOLOGY AND TAXONOMY OF THE SOFT-BODIED CAMBRIAN ‘MOLLUSC’ WIWAXIA by MARTIN R. SMITH1,2,3 1Department of Ecology and Evolutionary Biology, University of Toronto, Toronto, Ontario, M5S 3G5, Canada 2Palaeobiology Section, Department of Natural History, Royal Ontario Museum, Toronto, Ontario, M5S 2C6, Canada 3Current address: Department of Earth Sciences, University of Cambridge, Cambridge, CB2 3EQ, UK; e-mail: [email protected] Typescript received 26 September 2012; accepted in revised form 25 May 2013 Abstract: The soft-bodied Cambrian organism Wiwaxia sclerites. I recognize a digestive tract and creeping foot in poses a taxonomic conundrum. Its imbricated dorsal scleri- Wiwaxia, solidifying its relationship with the contemporary tome suggests a relationship with the polychaete annelid Odontogriphus. Similarities between the scleritomes of worms, whereas its mouthparts and naked ventral surface Wiwaxia, halkieriids, Polyplacophora and Aplacophora hint invite comparison with the molluscan radula and foot. 476 that the taxa are related. A molluscan affinity is robustly new and existing specimens from the 505-Myr-old Burgess established, and Wiwaxia provides a good fossil proxy for Shale cast fresh light on Wiwaxia’s sclerites and scleritome. the ancestral aculiferan – and perhaps molluscan – body My observations illuminate the diversity within the genus plan. and demonstrate that Wiwaxia did not undergo discrete moult stages; rather, its scleritome developed gradually, with Key words: halwaxiids, scleritomorphs, Aculifera, Mollusca, piecewise addition and replacement of individually secreted evolution, Cambrian explosion. T HE slug-like Cambrian organism Wiwaxia perennially Nevertheless, Butterfield (2006, 2008) identified char- resists classification, in part due to its unusual scleritome acters that could place these taxa deeper within of originally chitinous scales and spines. These sclerites, Lophotrochozoa, and the affinity of Wiwaxia remains preserved in fossils as carbonaceous films, seem to have ambiguous. been secreted by microvilli – placing Wiwaxia among the Wiwaxia has also been connected to Halkieria, another Lophotrochozoa (Butterfield 1990, 2008). The sclerites Cambrian scleritome bearer; organisms of this ilk may were first identified with the elytra (modified bristles belong to a ‘halwaxiid’ clade or grade (Conway Morris resembling fleshy scales) of annelid worms, although and Caron 2007; Sigwart and Sutton 2007; Vinther et al. annelid paleae (flattened chaetae) more closely match 2008). However, unlike Wiwaxia – whose sclerites are the sclerites’ shape, microstructure and arrangement in a solid and nonmineralized (Butterfield 1990) – Halkieria dorsal, nonmineralized, imbricated scleritome (Butterfield bore hollow sclerites composed of fibrous aragonite. It 1990). However, paleae are invariably paired with neu- has thus been proposed that Halkieria is unrelated to rosetae (chaetae borne on the ventral portion of a bira- Wiwaxia and instead lies with the sponge-like chancellori- mous unit), which are not evident in Wiwaxia. ids and other coeloscleritophorans (Porter 2008). Phylogenetic analyses reflect the limited morphological The species-level taxonomy of Wiwaxia represents a resemblance between Wiwaxia and annelids; parsimony- further outstanding question. Besides the canonical based reconstructions oppose a position in the annelid Wiwaxia corrugata Matthew 1899, the Burgess Shale con- crown group (Eibye-Jacobsen 2004) and instead suggest a tains smaller articulated specimens that lack spines. These molluscan affinity (Sigwart and Sutton 2007; Vinther could represent either juvenile W. corrugata (Conway et al. 2008). The latter affiliation is consistent with Morris 1985) or a separate species. Furthermore, isolated Wiwaxia’s aplacophoran-like body form and radula-like sclerites have been recovered from the Burgess Shale and mouthparts (Conway Morris 1985; Caron et al. 2006, several other early-to-middle Cambrian deposits (Butter- 2007; Smith 2012). Wiwaxia’s close relative Odontogriphus field 1990, 1994; Fatka et al. 2011; Harvey and Butterfield omalus Conway Morris 1976 also has molluscan synapo- 2011; Butterfield and Harvey 2012; Harvey et al. 2012). morphies – notably a radula and a ventral creeping These tend to be an order of magnitude smaller than foot surrounded by gills (Caron et al. 2006; Smith 2012). those typical of body fossils and show subtle differences © The Palaeontological Association doi: 10.1111/pala.12063 1 2 PALAEONTOLOGY in morphology; they may represent multiple species or SCLERITOME ontogenetic stages. Articulated specimens from Kaili, China may belong to W. corrugata (Sun et al. 2013) but Previous research have been proposed as a separate species, Wiwaxia taiji- angensis (Zhao et al. 1994). The dorsal and lateral surfaces of W. corrugata are cov- This study revisits features in Wiwaxia and Odontogri- ered with imbricating leaf-like sclerites. Conway Morris phus that have previously been contentious, and provides (1985) found the sclerites to be consistently arranged in a detailed account of the Wiwaxia scleritome. This abun- seven to nine transverse rows and five distinct regions. dant new material fills a gap in the Wiwaxia size distribu- From the midline out, he recognized zones of (1) tion, resolving the ontogeny of the scleritome and asymmetrical dorsal sclerites; (2) rounded, symmetrical confirming that smaller specimens are more likely to rep- upper-lateral sclerites; (3) more oval-shaped symmetri- resent juveniles than separate taxa. Novel observations cal lower-lateral sclerites; and (4) sickle-shaped ventrolat- substantiate Wiwaxia’s molluscan affinity, but its position eral sclerites. A distinct anterior row forms the fifth within Mollusca remains unsettled. sclerite zone; sclerites in this row resemble the upper laterals. Eibye-Jacobsen (2004), however, saw a different number of transverse rows in each scleritome zone, MATERIALS AND METHODS reporting 11–14 rows of upper-lateral sclerites in contrast to seven to eight rows of lower-lateral sclerites. I examined 476 Wiwaxia and 170 Odontogriphus speci- In addition to these ‘body sclerites’, a variable number mens (Table 1 and Table S1), most collected by the Royal of elongate spines (7–11) emerge from the dorsal/lateral Ontario Museum. Backscatter electron micrographs of sclerite zones on each side of larger specimens. Spine uncoated specimens were obtained under environmental length is variable, with spines usually shorter near the pressure SEM (Orr et al. 2002), complementing tradi- front and rear; maximum spine length grows nonlinearly tional light microscopy and digital interference of images with respect to sagittal length (Conway Morris 1985). obtained under plane- and cross-polarized light (Bengtson Each sclerite comprises a hollow tubular root that 2000). Scaling relationships were recovered from log- opens out to form a flattened, one-sided blade (Butter- transformed dimension measurements (obtained from field 1990). The blade is ornamented by a number of lon- digital images) using linear models; sclerite zones were gitudinal ribs, of which more are present in larger treated separately where this improved the Akaike infor- specimens (Conway Morris 1985). These external ribs are mation criterion. restricted to the upper surface of sclerites and are sometimes accompanied by irregularly scattered pustules (Butterfield 1990). In some sclerites from Mount Cap and Kaili, ribs have two thicknesses: prominent ribs alternate TABLE 1. Details of examined material. with fainter ones (Butterfield 1994; Harvey et al. 2012). Taxon Wiwaxia Odontogriphus The full width of each sclerite is striated by finely spaced longitudinal lineations. Parker (1998) argued that Total specimens studied 476 170 these were superficial – although they are not visible on Provenance surfaces imaged under SEM and do not exhibit interfer- Walcott Quarry 447 167 ence under transmitted light, so might be better inter- Talus below Walcott Quarry 23 1 preted as internal channels indicating microvillar Talus below Raymond Quarry 1 1 secretion (Butterfield 1990). Tulip Beds (locality S7 on 51 Butterfield (1990) suggested that sclerites were moulted Mount Stephen) Status individually, as in Halkieria (Vinther and Nielsen 2005). Articulated specimen 431 160 Conversely, whole-body ecdysis is suggested by a puta- ‘Juvenile’ specimen 52 n/a tively moulting specimen and the absence of specimens in (<25 mm in length) the 15- to 20-mm size range (Conway Morris 1985). Institution Royal Ontario Museum, 443 169 Toronto, Canada (ROM) Sclerite morphology Smithsonian Institution National 31 1 Museum of Natural History, In the examined material, sclerites are uniformly orna- Washington, DC, USA (NMNH) mented with a single order of ribs that are most prominent – Geological Survey of Canada, 2 in the smallest specimens. The ribs’ number generally Ottawa, Canada increases with the fossils’ sagittal length, reaching 9–15 in SMITH: ONTOGENY, MORPHOLOGY AND TAXONOMY OF WIWAXIA 3 adult body sclerites (intercept = 0.88 Æ 0.26, expo- results in a partly open tube (Fig. 1C). The spines are À nent = 0.27 Æ 0.03, R2 = 0.38; p < 10 7) and 5–12 in gently curved in cross section (Fig. 1D). spines (intercept = 0.58 Æ 0.50, exponent = 0.20 Æ 0.06, In six smaller specimens (sclerite length c. 1 mm), scle- R2 = 0.16, p = 0.0037); ventrolateral sclerites are the rites’ fine longitudinal striations (the ‘chambers’ of But- exception, with no statistically significant increase in terfield