Synthesis of geological and comparative phylogeographic data point to climate, not mountain uplift, as driver of divergence across the Eastern Andean Cordillera

AimTo evaluate the potential role of the orogeny of the Eastern Cordillera (EC) of the Colombian Andes and the Merida Andes (MA) of Venezuela as drivers of vicariance between populations of 37 tetrapod lineages co-distributed on both flanks, through geological reconstruction and comparative phylogeographic analyses. LocationNorthwestern South America MethodsWe first reviewed and synthesized published geological data on the timing of uplift for the EC-MA. We then combined newly generated mitochondrial DNA sequence data with published datasets to create a comparative phylogeographic dataset for 37 independent tetrapod lineages. We reconstructed time-calibrated molecular phylogenies for each lineage under Bayesian inference to estimate divergence times between lineages located East and West of the Andes. We performed a comparative phylogeographic analysis of all lineages within each class of tetrapod using hierarchical approximate Bayesian computation (hABC) to test for synchronous vicariance across the EC-MA. To evaluate the potential role of life history in explaining variation in divergence times among lineages, we evaluated 13 general linear models (GLM) containing up to six variables each (maximum elevation, range size, body length, thermoregulation, type of dispersal, and taxonomic class). ResultsOur synthesis of geological evidence suggested that the EC-MA reached significant heights by 38-33 million years ago (Ma) along most of its length, and we reject the oft-cited date of 2-5 Ma. Based on mtDNA divergence from 37 lineages, however, the median estimated divergence time across the EC-MA was 3.26 Ma (SE = 2.84) in amphibians, 2.58 Ma (SE = 1.81) in birds, 2.99 Ma (SE = 4.68) in reptiles and 1.43 Ma (SE = 1.23) in mammals. Using Bayes Factors, the hypothesis for a single temporal divergence interval containing synchronous divergence events was supported for mammals and but not supported for amphibians, non-avian reptiles, or birds. Among the six life-history variables tested, only thermoregulation successfully explained variation in divergence times (minimum AICc, R2 0.10), with homeotherms showing more recent divergence relative to poikilotherms. Main conclusionsOur results reject the hypothesis of the rise Andean Cordillera as driver of vicariance of lowland population because divergence dates are too recent and too asynchronous. We discuss alternative explanations, including dispersal through mountain passes, and suggest that changes in the climatic conditions during the Pliocene and Pleistocene interacted with tetrapod physiology, promoting older divergences in amphibians and reptiles relative to mammals and birds on an already established orogen.