Recent studies of deep-sea faunas considered the influence of mid-domain models in the distribution of species diversity and richness with depth. In this paper, I show that separating local diversity from regional species richness in benthic isopods clarifies mid-domain effects in the distribution of isopods in the Gulf of Mexico. Deviations from the randomised implied species ranges can be informative to understanding general patterns within the Gulf of Mexico. The isopods from the GoMB study contained 135 species, with a total of 156 species including those from an earlier study. More than 60 species may be new to science. Most families of deep-sea isopods (suborder Asellota) were present, although some were extremely rare. The isopod family Desosomatidae dominated the samples, and one species of Macrostylis (Macrostylidae) was found in many samples. Species richness for samples pooled within sites ranged from 1 to 52 species. Because species in pooled samples were highly correlated with individuals, species diversity was compared across sites using the expected species estimator (n = 15 individuals, ES15). Six depth transects had idiosyncratic patterns of ES15, and transects with the greatest short-range variation in topography, such as basins and canyons, had the greatest short-range disparity. Basins on the deep slope did not have a consistent influence (i.e., relatively higher or lower than surrounding areas) on the comparative species diversity. ES15 of all transects together showed a weak mid-domain effect, peaking around 1200-1500m, with low values at the shallowest and deepest samples (Sigsbee Abyssal Plain): no longitudinal (east-west) pattern was found. The regional species pool was analyzed by summing the implied ranges of all species. The species ranges in aggregate did not have significant patterns across longitudes, and many species had broad depth ranges, suggesting that the isopod fauna of the Gulf of Mexico is well dispersed. The summed ranges, as expected, had strong mid-domain patterns, contrasting with the local species richness estimates. The longitudinal ranges closely matched a randomized pattern (species ranges placed randomly, 1000 iterations), with significant deviations in the east attributable to lower sampling effort. The depth pattern, however, deviated from the mid-domain model, with a bimodal peak displaced nearly 500 in shallower than the mode of the randomized distribution. The deviations from random expectation were significantly positive above 1600 rn and negative below 2000 m, with the result that regional species richness peaked between 800 and 1200 m, and decreased rapidly at deeper depths. The highest species richness intervals corresponded to the number of individuals collected, Residuals from a regression of the deviations on individual numbers, however, still deviated from the randomized pattern. In this declining depth-diversity pattern, the Gulf of Mexico resembles other partially enclosed basins, such as the Norwegian Sea, known to have suffered geologically recent extinction events. This displaced diversity pattern and broad depth ranges implicate ongoing re-colonization of the deeper parts of the Gulf of Mexico. The Sigsbee Abyssal Plain sites could be depauperate for historical reasons (e.g., one or more extinction events) rather than ongoing ecological reasons (e.g., low food supply). Crown Copyright (C) 2008 Published by Elsevier Ltd. All rights reserved.
