Search

We used morphological, mitochondrial DNA sequence, paleontological, and biogeographical information to examine the evolutionary history of crabs of the genus Cancer. Phylogenies inferred from adult morphology and DNA sequence of the cytochrome oxidase I (COI) gene were each well resolved and well supported, but differed substantially in topology. Four lines of evidence suggested that the COI data set accurately reflected Cancer phylogeny: (1) in the phylogeny inferred from morphological data, each Atlantic species was sister taxon to an ecologically similar Pacific species, suggesting convergence in morphology; (2) a single trans-Arctic dispersal event, as indicated by the phylogeny inferred from COI, is more parsimonious than two such dispersal events, as inferred from morphology; (3) test and application of a maximum likelihood molecular clock to the COI data yielded estimates of origin and speciation times that fit well with the fossil record; and (4) the tree inferred from the combined COI and morphology data was closely similar to the trees inferred from COI, although notably less well supported by the bootstrap. The phylogeny inferred from maximum likelihood analysis of COI suggested that Cancer originated in the North Pacific in the early Miocene, that the Atlantic species arose from a North Pacific ancestor, and that Cancer crabs invaded the Atlantic from the North Pacific 6-12 mya. This inferred invasion time is notably prior to most estimates of the date of submergence of the Bering Strait and the trans-Arctic interchange, but it agrees with fossil evidence placing at least one Cancer species in the Atlantic about 8 mya., (Copyright 1999 Academic Press.)