Abstract

Pauly (1980. J. Cons. Int. Explor. Mer, 39: 175-192) showed that natural mortality rates in fish could be predicted from body growth parameters and environmental temperature but found no evidence for ecological or taxonomic influences. Using an updated database and techniques that avoid some of the earlier analytical problems, we confirm Pauly's conclusion that mortality is correlated with growth and temperature. A path model supports the role of ecological effects on mortality. A phylogenetic effect is also apparent: perciform fishes occupy warmer environments than other species, but in the predator-rich reef habitat they suffer much lower mortality rates. Species that are cryptic or hide in burrows or have morphological defences against predators show the lowest mortality rates.