Macroevolution and climate changes: a global multi-family test supports the resource-use hypothesis in terrestrial mammals

Elisabeth S. Vrba's resource-use hypothesis suggests that speciation in biomes subjected to successive expansion-contraction-fragmentation during periods of climatic change generates high frequency of species restricted to a single biome (stenobiomic species). We compiled biome occupation for all terrestrial mammals and, using Monte Carlo simulations, demonstrated that patterns of biome occupation are congruent with those predicted by the resource-use hypothesis. Biome specialists are much more speciose than expected by chance, while there are fewer moderate biome generalists than expected. Despite their scarcity, extreme eurybiomic lineages show significant overrepresentation, which suggests they are seldom affected by climate-related extinction processes. Additionally, stenobiomic species are concentrated in biomes placed at the extremes of the climatic gradient, such as equatorial rainforest, subtropical desert, steppe, and tundra. Although this pattern is fairly maintained across different mammalian families, highlighting its universality, our analysis also found great variability. Exceptions to the predictions of the resource-use hypothesis seem to be associated to biome climatic or geographical heterogeneity, which favours vicariance in some biomes not placed in extremes of the climatic gradient (tropical deciduous woodland, sclerophyllous woodland-shrubland), as well as life-history differences across taxa, which generates a stronger trend to specialisation in small body size lineages than in larger mammals.