Revisão taxonômica das rãzinhas-de-barriga-colorida Paratelmatobius cardosoi e P. segallai (Leptodactylidae, Paratelmatobiinae)

Abstract

The use of methods based on the multispecies coalescent model represents a significant advancement in the field of species delimitation. However, the model may detect population splits that do not necessarily correspond to interspecific divergences, especially in early divergent lineages with high levels of geographic structure. One possible solution is the use of different lines of evidence and delimitation methods in an integrative taxonomic approach. The colorful-bellied frogs Paratelmatobius cardosoi and P. segallai are sister species distributed throughout the Brazilian Atlantic Forest in the states of São Paulo and Paraná. In a recent study, four divergent and geographically structured lineages were identified in P. cardosoi, and four others in P. segallai. However, only molecular data were analyzed to date. The aim of this study was to assess the taxonomic status of these lineages within P. cardosoi and P. segallai by using an integrative taxonomy protocol, which involves the application of a set of discovery and validation methods based on sequences of ten molecular markers and morphometric data, followed by a verification of congruence with morphological and acoustic characters. For the discovery step, we used two methods based on the concatenated mitochondrial markers (Assemble Species by Automatic Partitioning, ASAP, and multi-rate Poisson Tree Processes, mPTP) and one method using the nuclear markers (multigenic distance network). The lineages delimited in the strict consensus of the three discovery methods were subjected to a two-step validation analysis. The first step was performed by applying methods based on the multispecies coalescent model and for this we used BPP (Bayesian Phylogenetics and Phylogeography) with molecular data and iBPP (integrative Bayesian Phylogenetics and Phylogeography) integrating morphometric data with molecular data. Finally, the second validation step was performed by evaluating congruence with morphological and acoustic characters. The different methods resulted in partially discordant delimitations. In the discovery step, analyses based on mitochondrial markers (ASAP and mPTP) delimited eight and nine lineages, respectively, whereas nuclear networks delimited six lineages. All the six lineages delimited in the strict consensus of the discovery methods were validated by BPP and iBPP analyses. However, the second validation step did not corroborate the lineages of P. cardosoi. Therefore, the hypothesis that the structure observed in the species corresponds to intraspecific divergences was better supported. On the other hand, one lineage of P. segallai was validated based on congruence with bioacoustic data, supporting the hypothesis that structure in this lineage corresponds to a speciation process. However, a broader sampling of vocalizations for this lineage needs to be conducted before proposing any taxonomic act. Finally, the use of genomic-scale data sets as well as other character systems such as larval and cytogenetic traits may contribute to clarify the uncertainty in the phylogenetic placement of some lineages and increase the accuracy of species delimitation.