Patterns of diversity, depth range and body size among pelagic fishes along a gradient of depth

Studies of geographical patterns of diversity have focused largely on compiling and analysing data to evaluate alternative hypotheses for the near-universal decrease in species richness from the equator to the poles. Valuable insights into the mechanisms that promote diversity can come from studies of other patterns, such as variation in species distributions with elevation in terrestrial systems or with depth in marine systems. To obtain such insights, we analysed and interpreted data on species diversity, depth of occurrence and body size of pelagic fishes along an oceanic depth gradient. We used a database on pelagic marine fishes native to the north-east Pacific Ocean between 40degreesN and 50degreesN. We used data from the Pacific Rim Fisheries Program that were obtained from commercial, management and scientific surveys between 1999 and 2000. Depth of occurrence and maximum body length were used to assess the distributions of 409 species of pelagic fishes along a depth gradient from 0 to 8000 in. A presence-absence matrix was used to classify the depth range of each species into 100-m intervals. Atmar & Patterson's (1995) software was used to quantify the degree of nestedness of species distributions. Pelagic fish species diversity decreased steeply with increasing depth; diversity peaked at less than 200 m and more than half of the species had mean depths of occurrence between 0 and 300 m. The distribution of species showed a very strong nested subset pattern along the depth gradient. Whereas species with narrow ranges were generally restricted to shallow waters, wide-ranging species occurred from near the surface to great depths. The relationship between maximum body size and mean depth range differed between teleost and elasmobranch fishes: being positive for teleosts, but negative for elasmo-branches. Results support hypotheses that some combination of high productivity and warm temperature promote high species diversity, and reject those that would attribute the pattern of species richness to the mid-domain effect, habitat area, or environmental constancy. The data provided a clear example of Rapoport's rule, a negative correlation between average depth range and species diversity.