A mid-Pleistocene rainforest corridor enabled synchronous invasions of the Atlantic Forest by Amazonian anole lizards.

Shifts in the geographic distribution of habitats over time can promote dispersal and vicariance, thereby influencing large-scale biogeographic patterns and ecological processes. An example is that of transient corridors of suitable habitat across disjunct but ecologically similar regions, which have been associated with climate change over time. Such connections likely played a role in the assembly of tropical communities, especially within the highly diverse Amazonian and Atlantic rainforests of South America. Although these forests are presently separated by open and dry ecosystems, paleoclimatic and phylogenetic evidence suggest that they have been transiently connected in the past. However, little is known about the timing, magnitude and the distribution of former forest connections. We employ sequence data at multiple loci from three codistributed arboreal lizards (Anolis punctatus, Anolis ortonii and Polychrus marmoratus) to infer the phylogenetic relationships among Amazonian and Atlantic Forest populations and to test alternative historical demographic scenarios of colonization and vicariance using coalescent simulations and approximate Bayesian computation (ABC). Data from the better-sampled Anolis species support colonization of the Atlantic Forest from eastern Amazonia. Hierarchical ABC indicates that the three species colonized the Atlantic Forest synchronously during the mid-Pleistocene. We find support of population bottlenecks associated with founder events in the two Anolis, but not in P. marmoratus, consistently with their distinct ecological tolerances. Our findings support that climatic fluctuations provided key opportunities for dispersal and forest colonization in eastern South America through the cessation of environmental barriers. Evidence of species-specific histories strengthens assertions that biological attributes play a role in responses to shared environmental change.