Abstract

The rate of mitochondrial DNA (mtDNA) evolution has been carefully calibrated only in primates. Similarity between the primate calibration and rates estimated for other vertebrates has led to widespread assumption of a constant molecular clock in vertebrates even though this has never been rigorously tested. We report here the examination of mtDNA sequence variation for 13 species of sharks from two orders that are well represented in the fossil record to test the constancy hypothesis. Nucleotide substitution rates in the cytochrome b and cytochrome oxidase I genes in sharks are seven- to eightfold slower than in primates or ungulates. This difference in substitution rate cannot be explained by nucleotide composition bias, codon-usage bias, selection, or choice of genes sequenced, and was confirmed by comparing species recently separated by the rise of the Isthmus of Panama. Such differences in mtDNA substitution rates among taxa indicate that it is inappropriate to use a calibration for one group to estimate divergence times or demographic parameters for another group. High-resolution studies of molecular evolutionary rates require taxon-specific calibrations.