Coelacanths were believed to have gone extinct more than 80 million years ago - until the sensational rediscovery of one surviving member of this leneage, Latimeria chalumanae, in 1938. Since then, plaeontologists and comparative morphologists have argues whether coelacanths or lungfish (two groups of lobe-finned fish) are the living sistergroup of the third extant lineage, the tetrapods. Recent molecular phylogenetic data on this debate tend to favor the hypothesis that lungfish are the closest relatives of land vertebrates. Somewhat surprisingly, the strongest molecular support for this hypothesis stems from mitochondrial rather than nuclear DNA sequences, despite the expectation that the more-slowly evolving nuclear genes should be more appropriate in addressing a phylogenetic issue involving taxonomic groups that diverged around 400 million years ago. This molecular estimate might serve as a framework to test palepntological and physiological innovations and preadaptations that allowed Devanian lobe-finned fish to colonize land.