Speciation in little: the role of range and body size in the diversification of Malagasy mantellid frogs

Background: The rate and mode of lineage diversification might be shaped by clade-specific traits. In Madagascar, many groups of organisms are characterized by tiny distribution ranges and small body sizes, and this high degree of microendemism and miniaturization parallels a high species diversity in some of these groups. We here investigate the geographic patterns characterizing the radiation of the frog family Mantellidae that is virtually endemic to Madagascar. We integrate a newly reconstructed near-complete species-level timetree of the Mantellidae with georeferenced distribution records and maximum male body size data to infer the influence of these life-history traits on each other and on mantellid diversification. Results: We reconstructed a molecular phylogeny based on nuclear and mitochondrial DNA for 257 species and candidate species of the mantellid frog radiation. Based on this phylogeny we identified 53 well-supported pairs of sister species that we used for phylogenetic comparative analyses, along with whole tree-based phylogenetic comparative methods. Sister species within the Mantellidae diverged at 0.2-14.4 million years ago and more recently diverged sister species had geographical range centroids more proximate to each other, independently of their current sympatric or allopatric occurrence. The largest number of sister species pairs had non-overlapping ranges, but several examples of young microendemic sister species occurring in full sympatry suggest the possibility of non-allopatric speciation. Range sizes of species included in the sister species comparisons increased with evolutionary age, as did range size differences between sister species, which rejects peripatric speciation. For the majority of mantellid sister species and the whole mantellid radiation, range and body sizes were associated with each other and small body sizes were linked to higher mitochondrial nucleotide substitution rates and higher clade diversity. In contrast, small range sizes were unexpectedly associated with a slow-down of mitochondrial substitution rates. Conclusions: Based on these results we define a testable hypothesis under which small body sizes result in limited dispersal capabilities and low physiological tolerances, causing smaller and more strongly fragmented ranges. This can be thought to facilitate reproductive isolation and thus favor speciation. Contrary to the expectation of the faster speciation of such microendemic phenotype species, we only found small body sizes of mantellid frogs to be linked to higher diversification and substitution rates, but not small range sizes. A joint analysis of various species-rich regional anuran radiations might provide enough species with all combinations of range and body sizes for a more conclusive test of this hypothesis.