Patterns of divergence in fish species separated by the Isthmus of Panama

Abstract

Background: The Pleistocene closure of Isthmus of Panama, separating the basins of the Eastern Pacific and the Caribbean Sea, created a unique natural experiment that reveals how marine faunas respond to environmental change. To explore how fishes have been affected by this tectonic event, I compare transisthmian patterns in phylogeny and morphology for geminate lineages in two families, Eleotridae (sleepers) and Apogonidae (cardinalfishes). Results: Time-calibrated phylogenies for these families show different diversification patterns. In Eleotridae, several independent shallow instances of transisthmian divergences occur, with one or a few species on either side of the Isthmus. Among Apogonidae, a single clade of Eastern Pacific species is nested within a broad Caribbean radiation that also includes a species known from the Mediterranean. Divergence time estimates for taxa isolated by closure of the Isthmus are broadly congruent. Hypotheses dated with deeper, fossil-based legacy calibrations put the divergences in the Miocene at 7.4-15.1 Ma, while those estimated with a shallow biogeographic calibration of final Isthmus closure range from 5.1 to 9.9 Ma, in the late Miocene/early Pliocene. Eleotridae are more euryhaline than Apogonidae, but do not exhibit shallower transisthmian divergences. In both families, descendent lineages on either side of the Isthmus of Panama exhibit significant shape differences, although that distinction disappears for Apogonidae when I apply a correction for phylogenetic relationships. To evaluate the tempo and mode of continuous character evolution, I fit several single and multiple rate evolutionary models to morphometric data reconstructed on the Apogonidae phylogeny. I find that the most highly favored model, as estimated on both legacy and isthmus calibrated hypotheses, is a multiple rate Ornstein-Uhlbeck model, with a mosaic of rate shifts postulated for shape changes among fishes in the Caribbean and Eastern Pacific. Conclusions: Although many transisthmian taxa have been compared and their phylogenies calibrated to estimate the dates associated with population sundering, few studies correlate these timing estimates with morphological change. I show that in transisthmian fish lineages, morphometric distinctions are detectable across the Isthmus, and that rates and patterns of shape change have also shifted, with variable manifestations across the body and between the Caribbean and Eastern Pacific.