Reproductive Biology and Phylogeny of Gymnophiona (Caecilians). Jean-Marie Exbrayat, editor.

Abstract

Fifteen years ago, we published a critical review of caecilian phylogeny and classification (Nussbaum and Wilkinson 1989). We hoped to establish some stability in caecilian classification in the face of some highly divergent phylogenetic hypotheses and alternative taxonomic treatments (Wake and Campbell 1983; Duellman and Trueb 1986; Lescure et al. 1986; Laurent 1986). We concluded that caecilian phylogeny was too poorly known to provide the basis for a working phylogenetic classification that recognized only well-founded monophyletic groups. Instead, we provided an interim, conservative classification in which 154 nominate species were partitioned into six families with no sub-or suprafamilial ranks other than 34 genera. Up until 1968, all caecilians were placed into a single family, the Caeciliidae, and we recognized that the subsequent removal of distinctive subsets of species in the establishment of additional families had left the Caeciliidae a most likely paraphyletic assemblage of caecilians that did not fit into one of the better circumscribed families. This is reflected in the relative numbers of taxa: well over half of all recognized caecilian species and genera are caeciliids. Some important milestones have appeared over the last 15 years in caecilian phylogenetics. Since our previous review, the first phylogenetic study of the interrelationships of caecilians based on DNA sequence data was published. Hedges et al. (1993) analyzed partial 16S and 12S mt rDNA sequences for 13 caecilian species in 9 genera, including members of four of the family-level taxa recognized in our 1989 classification. Recently, the taxonomic coverage for these molecular markers has begun to expand, so that comparative sequence data are now available for 23 species, 16 genera 40 Reproductive Biology and Phylogeny of Gymnophiona and for representatives of all six families (Wilkinson et al. 2002; 2003b). Gower et al. (2002), in a study focusing on ichthyophiid caecilians, demonstrated the potential for sequence data from ribosomal and protein coding (cytochrome B) mt DNA to help resolve low level taxonomic problems when the taxonomic sampling is sufficiently dense, and Gower et al. (2005) tentatively identified an undescribed cryptic species of Sri Lankan Ichthyophis using molecular data. Recently, San Mauro et al. (2004) addressed relationships among single representatives of each of the six families with a combination of complete mitochondrial genomes and RAG-1 nuclear gene sequences. Morphological data sets have also been expanded in terms of taxa and through the discovery of additional characters, and previously assembled data have been critically reviewed and revised to reduce errors (Naylor and Nussbaum 1980; Nussbaum and Naylor 1982; Scheltinga et al. 2003, see also Chapter 7 of this volume; Wilkinson and Nussbaum 1996; Wilkinson 1996, 1997). There have also been a few phylogenetic studies of monophyletic subgroups at the genus-or species-level using morphological data (Nussbaum and Hinkel 1994; Wilkinson and Nussbaum 1999; Wilkinson et al. 2004). In our 1989 classification, we provided diagnoses of caeciliid genera based on a core set of characters. Although each genus was understood to have a unique combination of characters, uniquely derived characters supporting the monophyly of most caeciliid genera were simply unknown, and knowledge of the diversity within the more speciose nominate genera was limited to one or a few species. Since 1989, a single genus and 16 species (one of which we consider invalid) have been newly described, and 5 species have been removed from synonymy. In the same period, two genera and 5 species have been lost to synonymy (in addition to those we excluded from our treatment and subsequently synonymized). Little else has changed, and the limited low-level taxonomic activity belies the fact that taxonomy at the species-and genus-level remains in need of careful study and stabilisation. The new genus, Atretochoana was established to receive a single species of typhlonectid caecilian with a radically divergent morphology discovered in the course of routine taxonomic work (Nussbaum and Wilkinson 1995). Atretochoana is the largest lungless tetrapod and the only known lungless caecilian, and it possesses many unique features associated with a novel cranial architecture (Wilkinson and Nussbaum 1997). Its discovery represents a substantial increase in the perceived diversity of caecilians, and of tetrapods (Donoghue and Alverson 2000), and it serves to emphasise the limited knowledge of caecilian biodiversity. Overall, taxonomic coverage has remained patchy in both morphological and molecular phylogenetic studies. Consequently, even where inferred relationships for the subset of sampled taxa are well-supported, they are not readily translated into a phylogenetic classification of the entire Order. This is exacerbated by the low-level taxonomic