Abstract
We describe an Oligocene newt specimen from western Germany that has gone practically unnoticed in the literature despite having been housed in the Museum für Naturkunde (Berlin) for a century. It is referable to the coeval Chelotriton, but is unusually peramorphic; for many characters it is more peramorphic than all other caudates or even all other lissamphibians. Most noticeable are the position of the jaw joints far caudal to the occiput, the honeycombed sculpture on the maxilla, and the possible presence of a septomaxilla (which would be unique among salamandrids). Referral to a species would require a revision of the genus, but the specimen likely does not belong to the type species. A phylogenetic analysis of nonmolecular characters of Salamandridae, far larger than all predecessors, confirms the referral to Chelotriton. It further loosely associates the Oligocene Archaeotriton and the Miocene Carpathotriton with the extant Lissotriton, though the former may alternatively lie outside Pleurodelinae altogether. The Miocene? I. randeckensis may not belong to the extant Ichthyosaura. The Miocene "Triturus" roehrsi is found neither with the extant Ommatotriton nor with Lissotriton, but inside an Asian/aquatic clade or, when geographic distribution is included as a character, as the sister-group to all other European molgins. The main cause for discrepancies between the results and the molecular consensus is not heterochrony, but adaptations to a life in mountain streams; this is the most likely reason why the Paleocene Koalliella from western Europe forms the sister-group to some or all of the most aquatic extant newts in different analyses. We would like to urge neontologists working on salamandrids to pay renewed attention to the skeleton, not limited to the skull, as a source of diagnostic and phylogenetically informative characters.