Scaphiophryne, Boulenger, 1882

Grosjean, Stéphane, Glos, Julian, Teschke, Meike, Glaw, Frank & Vences, Miguel, 2007, Comparative larval morphology of Madagascan toadlets of the genus Scaphiophryne: phylogenetic and taxonomic inferences, Zoological Journal of the Linnean Society 151 (3), pp. 555-576 : 572-573

OF SCAPHIOPHRYNE

Microhylids, in the past, have sometimes been considered to be related to basal frog lineages, one argument for this placement being their larval morphology which is reminiscent of that of the archaic pipid and rhinophrynid frogs (e.g. Orton, 1953; Hecht, 1963; Inger, 1967). Current data are unambiguous with regard to placing the family, instead, firmly among modern (neobatrachian) frogs, and more precisely, as one of three major lineages in the superfamily Ranoidea (e.g. Van der Meijden et al., 2004), which is equivalent to the phylocode taxon Ranoides of Frost et al. (2006). Although most microhylids are characterized by filter-feeding tadpoles which are derived compared with the basic ranoid type, many microhylids have further derived larval morphologies, such as the surface-feeding tadpoles of some Microhyla ( M. heymonsi, Wassersug, 1980 ; Chou & Lin, 1997; M. achatina, Smith, 1916 ), the non-feeding tadpoles of the subfamily Cophylinae ( Blommers-Schlösser, 1975) , the burrowing tadpoles of the South American Otophryne with keratinized ‘teeth’ in place of the beak and a sinistral spiracle ( Wassersug & Pyburn, 1987), or a fully reduced larval stage such as in the directly developing Papuan taxa (e.g. Tyler, 1963; Zweifel, 1972; Menzies, 1976; McDiarmid & Altig, 1999).

The tadpoles of the genus Scaphiophryne , as emphasized by Blommers-Schlösser (1975) and Wassersug (1984), bear some characters typical of ranid tadpoles, others typical of microhylid tadpoles and others intermediate between the two families. Among these intermediate characters are the structure and shape of branchial food traps, the position of the glottis, the shape of the ventral velum and, at the level of the oral disc, the presence of marginal papillae and of a beak but the absence of keratodonts, and an intermediate position of the spiracle. Further intermediate characters of Scaphiophryne , and their impact on morphological reconstruction of anuran phylogeny, are discussed by Haas (2003). It is tempting to interpret this morphology as an evolutionary transitory state, and scaphiophrynines as the most basal microhylid group. In fact, Scaphiophryne was placed basal to other microhylids in the molecular study of Van der Meijden et al. (2004), albeit with negligible statistical support.

This interpretation, however, is strongly questioned if the second scaphiophrynine genus, Paradoxophyla , is taken into account. This genus contained a single species, P. palmata , until recently, which was strongly divergent from Scaphiophryne based on its general habitus, its partially aquatic habits and fully webbed feet; the recent discovery of a second species, P. tiarano , with rudimentary foot webbing ( Andreone et al., 2006) decreases the importance of these differences. Blommers-Schlösser & Blanc (1991) hypothesized relationships between Scaphiophryne and Paradoxophyla based on osteological similarities, especially on the presence of an undivided vomer and sphenethmoid. Molecular data by Van der Meijden et al. (2007) in fact support the monophyletic group constituted by these two genera. However, the tadpoles of Paradoxophyla palmata exhibit typical morphology and buccopharyngeal features of microhylids ( Parker, 1934; Wassersug, 1980; Chou & Lin, 1997; Shimizu & Ota, 2003; S. Grosjean, pers. observ.), as obvious from their similarity to the tadpoles of Dyscophus insularis described herein (see Table 3). Tadpoles assigned to Paradoxophyla tiarano , decribed by Andreone et al. (2006), are of the same morphology, although buccopharyngeal features have not been examined. In the morphological phylogeny of Haas (2003), based on primarily larval characters, Paradoxophyla was grouped sister to the African microhylid Phrynomantis while Scaphiophryne was placed basal to all other microhylids, which emphasizes the strong morphological differences between these larvae and the morphological similarities of the Paradoxophyla tadpole to those of other microhylids.

Paradoxophyla appears to be the sister group of Scaphiophryne and shares with other microhylids their derived tadpole morphology, including detailed similarity in many specialized character states, indicating that the Scaphiophryne tadpoles have evolved their unique morphology from an ancestor of typical microhylid larval morphology. Any other hypothesis would need the assumption that the filter-feeding microhylid tadpole has evolved twice independently, once in Paradoxophyla and once in the lineage leading to the remaining microhylid taxa. Given the close similarity between larvae of Paradoxophyla and other microhylids, which extends to many synapomorphic states not found in Scaphiophryne , and not found in pipid and rhinophrynoid filter-feeding tadpoles (such as the U-shaped lower labium, the unique medioventral spiracle lying near the ventral tube, the position of the glottis far anterior on the buccal floor and its anterodorsal orientation; Sokol, 1962, 1977; Gradwell, 1975), we consider this alternative as highly unlikely. Hence, we favour an evolutionary scenario in which the larvae of Scaphiophryne obtained their unique features by reversing some of the filter-feeding adaptations of their ancestors and adapting to their psammonektonic lifestyle which is not known from any other anuran species thus far, thereby constituting an astonishing exception from Dollo’s rule. Together with probable examples of re-aquisition of larval stages from endotrophically developing ancestors ( Duellman & Hillis, 1987; Vences & Glaw, 2001; Chippindale et al., 2004), this provides remarkable evidence for the evolutionary plasticity of developmental modes and larval morphology in amphibians.