Patterns of species richness, endemism and environmental gradients of African reptiles

Aim: To map and assess the richness patterns of reptiles (and included groups: amphisbaenians, crocodiles, lizards, snakes and turtles) in Africa, quantify the overlap in species richness of reptiles (and included groups) with the other terrestrial vertebrate classes, investigate the environmental correlates underlying these patterns, and evaluate the role of range size on richness patterns. Location: Africa. Methods: We assembled a data set of distributions of all African reptile species. We tested the spatial congruence of reptile richness with that of amphibians, birds and mammals. We further tested the relative importance of temperature, precipitation, elevation range and net primary productivity for species richness over two spatial scales (ecoregions and 1° grids). We arranged reptile and vertebrate groups into range-size quartiles in order to evaluate the role of range size in producing richness patterns. Results: Reptile, amphibian, bird and mammal richness are largely congruent (r = 0.79–0.86) and respond similarly to environmental variables (mainly productivity and precipitation). Ecoregion size accounts for more variation in the richness of reptiles than in that of other groups. Lizard distributions are distinct with several areas of high species richness where other vertebrate groups (including snakes) are species-poor, especially in arid ecoregions. Habitat heterogeneity is the best predictor of narrow-ranging species, but remains relatively important in explaining lizard richness even for species with large range sizes. Main conclusions: Reptile richness varies with similar environmental variables as the other vertebrates in Africa, reflecting the disproportionate influence of snakes on reptile richness, a result of their large ranges. Richness gradients of narrow-ranged vertebrates differ from those of widespread taxa, which may demonstrate different centres of endemism for reptile subclades in Africa. Lizard richness varies mostly with habitat heterogeneity independent of range size, which suggests that the difference in response of lizards is due to their ecological characteristics. These results, over two spatial scales and multiple range-size quartiles, allow us to reliably interpret the influence of environmental variables on patterns of reptile richness and congruency.