Contributions to Zoology, 73 (3) 207-252 (2004) SPB Academic Publishing bv, The Hague Morphological data, extant Myriapoda, and the myriapod stem-group Gregory+D. Edgecombe Australian Museum, 6 College Street, Sydney, NSW 2010, Australia, e-mail: [email protected] Keywords: Myriapoda, phylogeny, stem-group, fossils Abstract Tagmosis; long-bodied fossils 222 Fossil candidates for the stem-group? 222 Conclusions 225 The status ofMyriapoda (whether mono-, para- or polyphyletic) Acknowledgments 225 and controversial, position of myriapods in the Arthropoda are References 225 .. fossils that an impediment to evaluating may be members of Appendix 1. Characters used in phylogenetic analysis 233 the myriapod stem-group. Parsimony analysis of319 characters Appendix 2. Characters optimised on cladogram in for extant arthropods provides a basis for defending myriapod Fig. 2 251 monophyly and identifying those morphological characters that are to taxon to The necessary assign a fossil the Myriapoda. the most of the allianceofhexapods and crustaceans need notrelegate myriapods “Perhaps perplexing arthropod taxa 1998: to the arthropod stem-group; the Mandibulatahypothesis accom- are the myriapods” (Budd, 136). modates Myriapoda and Tetraconata as sister taxa. No known pre-Silurianfossils have characters that convincingly place them in the Myriapoda or the myriapod stem-group. Because the Introduction strongest apomorphies ofMyriapoda are details ofthe mandible and tentorial endoskeleton,exceptional fossil preservation seems confound For necessary to recognise a stem-group myriapod. Myriapods palaeontologists. all that Cambrian Lagerstdtten like the Burgess Shale and Chengjiang have contributed to knowledge of basal Contents arthropod inter-relationships, they are notably si- lent on the matter of myriapod origins and affini- Introduction 207 ties. Few comparisons have been made between Arthropod phylogeny: the Recent tree 208 Cambrian marine organisms and members of the Taxonomic and character sampling 209 the myriapod crown-group, i.e., Chilopoda (centi- Cladistic methods 210 and and Results 210 pedes) Progoneata (symphylans, pauropods The dearth of well-founded Reconstructing the myriapod ground pattern 213 millipedes). compari- 213 Autapomorphies of Mandibulata ' sons is a reflection ofreal patterns in the fossil record Mandible 213 (a lack of appropriately-aged soil and litter faunas) First maxilla 216 the terrestrial habits of and cryptic, crown-group Ommatidium with crystalline cone [Mandibulata] myriapods, but it is also influenced by imprecise and multilayered rhabdom [Myriapoda] 217 Ollier characters 218 or flawed concepts of myriapod morphology and Autapomorphies of Myriapoda 218 fossils that be relationships. Identifying may mem- Tentorial bars and tentorial mobility 218 bers of the myriapod stem-group requires a well- Separated, independently musculated gnathal lobe 220 founded hypothesis of myriapod phytogeny in the Other characters 221 broader context of Arthropoda, based on Mandibular comb lamellae 221 precisely defined characters that ‘Tracheate” characters in Myriapoda 221 apomorphic can potentially Downloaded from Brill.com10/10/2021 12:05:33AM via free access - 208 G.D. Edgecombe Myriapod stem-group be examined in fossil material. Budd al. As et (2001: coding genes (Regier and Shultz, 1997, 2001a, b; noted when the well from 38) evaluating “myriapod-like” Shultz and Regier, 2001) as as several Cambrian fossil Xanthomyria, “The ground-plan kinds of non-sequence data. The latter includes brain features for remain many groups [of arthropods] ultrastructure(Strausfeld, 1998; Locsel et al., 2002), uncertain”. That uncertainty impedes understand- neurogenesis (Duman-Scheel and Patel, 1999; Simp- the or contributions of ultrastructure ing significance possible son, 2001), eye (Melzer et al., 1997), fossils. and mitochondrial order gene (Boore et al., 1998). The palaeontological context of Myriapoda has These studies [reviewed by Dohle (2001) and Richter been reconsidered based on developments in ar- (2002a)] have contributed to a perception that myria- thropod phylogcny that have come from exclusively pods are more “basal” than crustaceans and hexa- neontological data, such as gene order, neurogenesis, pods. and molecular Budd for sequences. (1999: 286), In short, Myriapoda is variably seen, even in the example, considered the possibility of a crustacean- latest literature, as either monophyletic (Edgecombe hexapod alliance to invite a radical repositioning and Giribet, 2002), paraphyletic (Kraus, 2001) or “It of the Myriapoda: is hard to see how the myri- polyphyletic (Loesel et al., 2002). Myriapods are be apods may be considered to the sister group to variably allied to either hexapods, chelicerates, or of any the arthropod stem-group taxa discussed a crustacean-hexapod clade, or are left unassigned here...; if the Tracheata [=Myriapoda + Hexapoda] in the euarthropod stem-group. The present review be are to abandoned, the possibility of myriapods aims to establish constraints on the systematic posi- representing an independent line of arthropodisation tion of myriapods by synthesising character evi- a remains dence available for their in the (from lobopodous ancestor) open...”. crown-group context insects as derived ofother Data come from Likewise, “Regarding crustaceans extant Arthropoda. a range makes in the difficult... of certainly fitting myriapods non-sequence sources, including external mor- But other if they do not fit in here, then there is no phology, internal anatomy, ultrastructure, embry- obvious for them to be tied into and order. Cladistic place arthropod ology, gene expression, gene phylogeny” (Budd, 2001: 71). analysis of this evidence permits the ground pat- The enigmatisation of myriapods has not been tern of the myriapod crown-group to be clarified. exclusively palaeontological. Rejection of a myri- This in turn allows fossils, including potential stem- alliance Tracheata be evaluated apod-hexapod (the or Atelocerata) group Myriapoda, to more precisely. after decades ofalmost universal acceptance stem- med first from analyses of molecular sequence data, such as studies based on small subunit rRNA (Frie- Arthropod phylogeny: the Recent tree drich and Tautz, 1995, 2001; Giribet et al., 1996). These analyses offered an alternative resolution of It hardly need be said that extant taxa have some sister of fossils. classes of character myriapods as group Chelicerata, though advantages over Some this result was or at best rendered am- such as mitochondrial order and rejected, data, gene sperm biguous, for the same genes with denser taxonomic ultrastructure, are confined to extant taxa, as is true sampling (Giribet and Ribera, 1998, 2000). A di- for most other genetic and ultrastructural informa- vision of into + tion and soft anatomical Euarthropoda (Chelicerata many characters. The es- Myriapoda) and (Crustacea + Hexapoda) has, how- calation of missing data for fossils is not confined ever, been endorsed in some other molecular analy- to molecular data; a sobering proportion of the non- Hox ses, including those based on gene sequences sequence characters analysed in this study (see (Cook et al., 2001), sequences for much of the Appendix) are unknown for all fossils, e.g., whole mitochondrial genome(Hwang et al., 2001; Delsuc blocks of characters for embryology, brain and eye et and and muscles. al., 2003), hemocyanin sequences (Kusche structure, For extant taxa, conjectures and Kusche et Inde- of refer Burmester, 2001; al., 2002). primary homology can to gene expression pendent support for the exclusion of Myriapoda from patterns (see the example ofDistal-less and dachs- a crustacean-hexapod clade has come from nuclear hund expression in the mandible below), embryo- Downloaded from Brill.com10/10/2021 12:05:33AM via free access Contributions to Zoology , 73 (3) - 2004 209 logical development, and details of ultrastructure tively, for summaries from opposing perspectives) that will be unknown for fossils. These details permit have agreed on the monophyly of Onychophora + a rigour in formulating hypotheses of homology Tardigrada + Euarthropoda. that be with and The taxonomic is may not possible fossils, larger sample designed to span the suites of character data more kinds of within the extant Taxo- (including major groups Arthropoda. character I nomic based terminal data) are available. shall not, however, sampling is on taxa for which endorse the conclusion drawn by Patterson and at least four of eight widely-sampled nuclear ribo- Rosen (1977) that these epistemological matters somal, nuclear protein-coding, and mitochondrial mean that fossils are subordinate in available Giribet to extant taxa genes are (see et al., 2001). The It has al. phylogenetic analysis. been amply demon- terminal taxa used here are as in Giribet et (2001) strated that a Recent-only tree may be overturned with the addition of the myriapods Cryptops (Chilo- on the basis of including extinct taxa in the sample, poda: Scolopendromorpha) and Spirostreptoidea a phenomenon that has been defended both theo- (Diplopoda: Spirostreptida) for which most of the retically and empirically (Gauthier et al., 1988; genes used in the molecular character set are now available. In marker is maxi- O’Leary, 1999). many cases, diversity The mised present analysis is confined to extant taxa by combining sequence data from different because terminal code the taxa are selected to include species to for a supraspecific terminal taxon, groups for which