Ecological release in lizard assemblages of neotropical savannas.

We compare lizard assemblages of Cerrado and Amazonian savannas to test the ecological release hypothesis, which predicts that niche dimensions and abundance should be greater in species inhabiting isolated habitat patches with low species richness (Amazonian savannas and isolated Cerrado patches) when compared with nonisolated areas in central Cerrado with greater species richness. We calculated microhabitat and diet niche breadths with data from 14 isolated Cerrado patches and Amazon savanna areas and six central Cerrado populations. Morphological data were compared using average Euclidean distances, and lizard abundance was estimated using the number of lizards captured in pitfall traps over an extended time period. We found no evidence of ecological release with respect to microhabitat use, suggesting that historical factors are better microhabitat predictors than ecological factors. However, data from individual stomachs indicate that ecological release occurs in these areas for one species (Tropidurus) but not others (Ameiva ameiva, Anolis, Cnemidophorus, and Micrablepharus), suggesting that evolutionary lineages respond differently to environmental pressures, with tropidurids being more affected by ecological factors than polychrotids, teiids, and gymnophthalmids. We found no evidence that ecological release occurs in these areas using morphological data. Based on abundance data, our results indicate that the ecological release (density compensation) hypothesis is not supported: lizard species are not more abundant in isolated areas than in nonisolated areas. The ecology of species is highly conservative, varying little from assemblage to assemblage. Nevertheless, increases in niche breadth for some species indicate that ecological release occurs as well.