Species delimitation in frogs from South American temperate forests: The case of Eupsophus, a taxonomically complex genus with high phenotypic variation.

One of the most characteristic and abundant amphibian taxa of South American temperate forests is Eupsophus. The ten currently recognized species of the genus have been divided in two species groups, roseus and vertebralis, but most of them, eight, belong to the roseus group. Recent phylogeographic and phylogenetic studies have suggested that species diversity of the roseus group could be underestimated. An examination of the literature shows that species of the roseus group exhibit high levels of variation in their external characteristics, particularly those used as diagnostic characters, which compromises their taxonomy and hinders their field recognition. High levels of variation were also observed in several new populations of the roseus group discovered in southern Chile (36°-40°S), which could not be identified to the species level by their external characteristics. On the other hand, the literature reveals a scarse karyotype differentiation and a high bioacoustic uniformity among the species of the roseus group. We performed a Bayesian phylogenetic analysis using mitochondrial and nuclear genes to reevaluate the species diversity of the roseus group, including all the nominal species of Eupsophus and new populations. This analysis was complemented with three species delimitation approaches, General Mixed Yule Coalescent, multi-rate Poisson Tree Process and Automatic Barcode Gap Discovery. We favored a conservative delimitation of only four species for the roseus group, a result more consistent with the distribution of pairwise genetic distances, and the available chromosome and bioacoustic evidence. The four recognized lineages, which have nearly completely allopatric distributions, are named after the earliest nominal species that they include, but because high levels of phenotypic variation, they are not diagnosable by consistent differences in external morphology. We discuss the implications of this new proposal for the taxonomy and conservation of the genus, and the possible causes of the difficulty to estimate its species diversity.