U. S. Geological Survey, Great Lakes Science Center, Lake Superior Biological Station, 2800 Lake Shore Drive East, Ashland, Wisconsin 54806.
Present address: U.S. Fish and Wildlife Service, Fisheries Division, Bishop Henry Whipple Federal Building, 1 Federal Drive, Ft. Snelling, Minnesota 55111.

Summary

         Lake Superior benthic fish community structure was studied during the summers of 1972-1995 by analysis of samples collected in bottom trawls. Nine hundred twenty-eight trawl tows at 21 stations collected 395,077 fish from 26 taxa. Redundancy analysis showed that fish community structure was not well explained by depth and temperature gradients, because they explained only 10% of fish community structure. Instead, discriminant analysis correctly classified 80% of the trawl tows to three depth zones that were revealed using principal components analysis. Some taxa were rarely collected, and densities of 10 taxa explained most of the data variation. Multivariate analysis of variance of those 10 taxa densities across depths showed that three significantly different assemblages existed; one inhabited shallow water (5.0-39.9 m), one inhabited intermediate depths (40.0-79.9 m), and one inhabited the deepest depths sampled (80.0-141.0 m). Analysis of variance showed that rainbow smelt (Osmerus mordax) and trout-perch (Percopsis omiscomaycus) densities were highest in the shallowest zone, whereas lake whitefish (Coregonus clupeaformis) density was highest in the intermediate zone, ninespine stickleback (Pungitius pungitius) and slimy sculpin (Cottus cognatus) densities were highest in the shallow and intermediate zones, siscowet lake trout (Salvelinus namaycush siscowet), bloater (Coregonus hoyi), kiyi (Coregonus kiyi), and deepwater sculpin (Myoxocephalus thompsoni) densities were highest in the deepest zone, and spoonhead sculpin (Cottus ricei) densities were similar across zones. Management objectives can be developed for each assemblage and each of the 10 most structurally important taxa. Understanding the existence and structure of each assemblage will aid in the scientific management of them, the taxa that compose them, and their habitats. Information on assemblage existence and structure will be useful for assessing changes in those assemblages that result from management actions and changes in the lake and its habitats.