Target-enriched DNA sequencing clarifies the identity of name-bearing types of the Gephyromantis plicifer complex and reveals a new species of mantellid frog from Madagascar

Mantellid frogs of the Madagascar-endemic Gephyromantis plicifer complex consist of three nominal species (G. luteus, G. plicifer and G. sculpturatus) as well as several genetically divergent lineages (candidate species), but uncertainties surround the identity of the name-bearing types of all three established nomina. We applied laboratory techniques to isolate archival DNA from the three old, liquidpreserved, name-bearing type specimens and conducted target-enriched DNA sequencing of a fragment of the mitochondrial 16S rRNA gene as basis for the revision of their taxonomy. Along with 16S sequences of 161 fresh samples, sequences of a fragment of the nuclear-encoded Rag-1 gene for 97 of these samples, a multi-gene data set of selected samples, and a comprehensive analysis of available advertisement call recordings, the new data suggest a novel taxonomic hypothesis: (1) The name G. plicifer applies to a highland clade of populations from the southern Central East, previously referred to as Gephyromantis sp. Ca21, with G. sculpturatus as a junior synonym. (2) The name G. luteus applies to a widespread clade distributed along most of Madagascar’s eastern rainforest band, and more specifically, is assigned to a subclade occurring mostly in coastal regions of the northern Central East and North East. (3) Large-sized specimens from Ranomafana (southern Central East) and Andohahela (South East) previously considered to belong to G. plicifer in fact represent a hitherto undescribed species, which we here formally name as G. pedronoi sp. nov. The three species are distinguished by very high genetic distances (over 8 % in the 16S gene), an absence of haplotype sharing in Rag-1, as well as morphological and bioacoustic characteristics. Moreover, several infraspecific clades with 16S divergences > 4 % may represent distinct species, especially within the widespread G. luteus, but cannot be unambiguously delimited by available data. We suggest dedicated sampling at contact zones and more extensive bioacoustic study of genotyped call vouchers to clarify their taxonomic status.