1. We examined indirect, community-level interactions in a stream that contained non-native brown trout (Salmo trutta Linnaeus), native brook trout (Salvelinus fontinalis Mitchill) and native slimy sculpin (Cottus cognatus Richardson). Our objectives were to examine benthic invertebrate composition and prey selection of fishes (measured by total invertebrate dry mass, dry mass of individual invertebrate taxa and relative proportion of invertebrate taxa in the benthos and diet) among treatments (no fish, juvenile brook trout alone, juvenile brown trout alone, sculpin with brook trout and sculpin with brown trout). 2. We assigned treatments to 1 m2 enclosures/exclosures placed in riffles in Valley Creek, Minnesota, and conducted six experimental trials. We used three designs of fish densities (addition of trout to a constant number of sculpin with unequal numbers of trout and sculpin; addition of trout to a constant number of sculpin with equal numbers of trout and sculpin; and replacement of half the sculpin with an equal number of trout) to investigate the relative strength of interspecific versus intraspecific interactions. 3. Presence of fish (all three species, alone or in combined-species treatments) was not associated with changes in total dry mass of benthic invertebrates or shifts in relative abundance of benthic invertebrate taxa, regardless of fish density design. 4. Brook trout and sculpin diets did not change when each species was alone compared with treatments of both species together. Likewise, we did not find evidence for shifts in brown trout or sculpin diets when each species was alone or together. 5. We suggest that native brook trout and non-native brown trout fill similar niches in Valley Creek. We did not find evidence that either species had an effect on stream communities, potentially due to high invertebrate productivity in Valley Creek. ?? 2007 Blackwell Publishing Ltd.