Comparative mtDNA phylogeography of neotropical freshwater fishes: testing shared history to infer the evolutionary landscape of lower Central America

Historical biogeography seeks to explain contemporary distributions of taxa in the context of intrinsic biological and extrinsic geological and climatic factors. To decipher the relative importance of biological characteristics vs. environmental conditions, it is necessary to ask whether groups of taxa with similar distributions share the same history of diversification. Because all of the taxa will have shared the same climatic and geological history, evidence of shared history across multiple species provides an estimate of the role of extrinsic factors in shaping contemporary biogeographic patterns. Similarly, differences in the records of evolutionary history across species will probably be signatures of biological differences. In this study, we focus on inferring the evolutionary history for geographical populations and closely related species representing three genera of primary freshwater fishes that are widely distributed in lower Central America (LCA) and northwestern Colombia. Analysis of mitochondrial gene trees provides the opportunity for robust tests of shared history across taxa. Moreover, because mtDNA permits inference of the temporal scale of diversification we can test hypotheses regarding the chronological development of the Isthmian corridor linking North and South America. We have focused attention on two issues. First, we show that many of the distinct populations of LCA fishes diverged in a relatively brief period of time thus limiting the phylogenetic signal available for tests of shared history. Second, our results provide reduced evidence of shared history when all drainages are included in the analysis because of inferred dispersion events that obscure the evolutionary history among drainage basins. When we restrict the analysis to areas that harbour endemic mitochondrial lineages, there is evidence of shared history across taxa. We hypothesize that there were two to three distinct waves of invasion into LCA from putative source populations in northwestern Colombia. The first probably happened in the late Miocene, prior to the final emergence of the Isthmus in the mid-Pliocene; the second was probably coincident with the rise of the Isthmus in the mid-Pliocene, and the third event occurred more recently, perhaps in the Pleistocene. In each case the geographical scale of the dispersion of lineages was progressively more limited, a pattern we attribute to the continuing development of the landscape due to orogeny and the consequent increase in the insularization of drainage basins. Thus, the fisheye view of LCA suggests a complex biogeographic history of overlaid cycles of colonization, diversification, sorting and extinction of lineages.