Fish assemblages and environmental variables were evaluated for 30 first- through seventh-order streams in the upper Snake River Basin, Idaho and western Wyoming. Data were collected as part of the National Water-Quality Assessment Program to characterize aquatic biota and associated habitats in surface water. Sampling sites represented major stream types in the basin large river, agricultural, and least-disturbed reference streams and springs in forested and (or) rangeland watersheds.

Twenty-four environmental variables representing various spatial scales, from watershed characteristics to instream habitat and physicochemical measures, were used to examine relations with fish assemblages. Twenty-six fish species in the families Catostomidae, Centrarchidae, Cottidae, Cyprinidae, Ictaluridae, Percidae, and Salmonidae were collected. Detrended correspondence analysis and canonical correspondence analysis differentiated fish assemblages on the basis of site type and showed that fish assemblages were most strongly correlated with percent agricultural and forest land uses, stream width, watershed size, and elevation. Fish assemblages did not correspond to the four major ecoregions in the basin. Comparisons between multiple-year and multiple-reach collections using Jaccard's coefficient of community similarity index generally indicated little difference in fish assemblages. Percent substrate fines, percent embeddedness, and specific conductance typically were higher for streams influenced by agricultural land use than for reference streams in forested and (or) rangeland watersheds. The number of native species, percent introduced species, percent omnivores, percent common carp, percent salmonids, and percent coldwater-adapted species varied according to site type. Percent omnivores and percent common carp were higher for large river and agricultural sites than for reference stream and spring sites. The introduction of intolerant salmonid species throughout the basin confounds the use of introduced species as a measure of environmental disturbance.

Analysis offish metrics identified some large river and agricultural sites in the lower part of the basin that did not support viable coldwater fish assemblages. These sites characteristically were dominated by tolerant, warmwater-adapted species. The findings of this study support the waterquality- limited designation for the middle reach of the Snake River between Milner Dam and King Hill and provide a framework for developing indices of biotic integrity by using fish assemblages to evaluate water quality of streams in the upper Snake River Basin.