Energy and the tempo of evolution in amphibians

The evolutionary speed hypothesis (ESH) attempts to explain global patterns of species richness on the basis that rates of molecular evolution and speciation in warmer climates have led to a greater accumulation of taxa at lower latitudes. A substantial alternative hypothesis to the ESH is the tropical conservatism hypothesis (TCH). However, recent tests of the TCH, using amphibians as the model taxon, have relied on the assumption that rates of molecular evolution are stable across latitudes and elevations. Here, we test for the first time for systematic variation in rates of molecular evolution across latitude and elevation among amphibians. The dataset is geographically diverse with samples from all continents except Antarctica and also from many of the earth's major tropical-warm temperate archipelagos. We tested for substitution rate heterogeneity across climatically varying habitats with the mitochondrial RNA genes 12S and 16S. Thus, we report here on our findings for amphibians - a taxon whose phylogenetic and trophic contexts are remote from those previously tested - using genes that have also not been examined before. The study utilized paired contrasts of sister species (188 species across 18 families, including both caudates and anurans) that are spatially separated in either latitudinal or elevational dimensions. We found substantially faster substitution rates for species living in warmer habitats (P= 0.001-0.002) at both lower latitudes (P < 0.02) and lower elevations (P < 0.01). The consistency of these results with the previous studies that used quite different organisms - and in this instance also using different genes - suggests that this is a ubiquitous pattern in nature consistent with the predictions of the ESH. Recent tests of the TCH that, in estimating diversification rates, have relied on the assumption that DNA evolution occurs at a constant rate across latitudes and elevations, require reconsideration in light of the findings presented here. Our results indicate that greater caution is required when estimating dates of divergence using DNA sequence data.