Ridges and rivers: a test of competing hypotheses of Amazonian diversification using a dart-poison frog (Epipedobates femoralis).

Mitochondrial DNA cytochrome b sequence data from a dart-poison frog, Epipedobates femoralis, were used to test two hypotheses of Amazonian diversification: the riverine barrier and the ridge hypotheses. Samples were derived from sites located on both banks of the Rio Juruá and on both sides of the Iquitos Arch in western Amazonia. The phylogeographic structure was inconsistent with predictions of the riverine barrier hypothesis. Haplotypes from opposite river banks did not form monophyletic clades in any of our phylogenetic analyses, nor was the topology within major clades consistent with the riverine hypothesis. Further, the greatest differentiation between paired sites on opposite banks was not at the river mouth where the strongest barrier to gene flow was predicted to occur. The results instead were consistent with the hypothesis that ancient ridges (arches), no longer evident on the landscape, have shaped the phylogeographic relationships of Amazonian taxa. Two robustly supported clades map onto opposite sides of the Iquitos Arch. The mean haplotypic divergence between the two clades, in excess of 12%, suggests that this cladogenic event dates to between five and 15 million years ago. These estimates span a period of major orogenesis in western South America and presumably the formation of these ancient ridges.