On a recent phylogenetic reanalysis of Sphaenorhynchini (Anura: Hylidae: Hylinae): Does it all come down to the method?

Abstract

Sphaenorhynchini comprises 15 species of small, greenish treefrogs from cis-Andean South America. Araujo-Vieira et al. (2019, A total evidence analysis of the phylogeny of hatch-faced treefrogs [Anura: Hylidae: Sphaenorhynchus], Cladistics 35, 469–486) conducted a total evidence parsimony analysis of DNA sequences and phenotypic data, corroborating the monophyly of Sphaenorhynchus, identifying three species groups, and placing the unusual species S. pauloalvini as sister to all other species. On this basis and numerous phenotypic differences, Araujo-Vieira et al. (2020, A new genus of lime treefrogs [Anura: Hylidae: Sphaenorhynchini], Zool. Anz. 286, 81–89) erected the genus Gabohyla for S. pauloalvini. Subsequently, Pereira et al. (2022, The dispersal between Amazonia and Atlantic Forest during the Early Neogene revealed by the biogeography of the treefrog tribe Sphaenorhynchini [Anura, Hylidae], Ecol. Evol. 12, e8754) performed a Bayesian analysis of a subset of Araujo-Vieira et al.’s 2019) molecular data and found G. pauloalvini to be nested within Sphaenorhynchus, which they attributed exclusively to choice of analytical method. To test this claim, we performed parsimony and Bayesian analyses of the total evidence dataset and the complete and partial molecular datasets with either the entire outgroup sample or a single terminal. The topology from the Bayesian analysis of the complete dataset is almost identical to that of Araujo-Vieira et al. (2019), with G. pauloalvini sister to Sphaenorhynchus, thereby refuting Pereira et al.’s claim. Moreover, the monophyly of Sphaenorhynchus sensu stricto was remarkably robust, being recovered in all analyses except the Bayesian analysis of the partial molecular dataset with a single outgroup terminal (i.e., the analysis performed by Pereira et al.). In addition to supporting the continued recognition of Gabohyla, our results underscore the importance of considering not only the choice of analytical method, but also character and taxon sampling—including outgroup sampling—before rejecting prior findings.