Stream inflow and predation risk affect littoral habitat selection by benthic fish

1. We examined small, fishless headwater streams to determine whether transport of macroinvertebrates into the littoral zone of an oligotrophic lake augmented food availability for Cottus asper, an abundant predatory fish in our study system. We sampled fish and macroinvertebrates during the recruitment and growth season of 2 years, either monthly (2004) or bi‐monthly (2005), to observe whether stream inputs increased prey availability and whether variation in total macroinvertebrate biomass was tracked by fish. 2. Observations from eight headwater streams indicated that streams did not increase the total macroinvertebrate biomass in the shallow littoral zone at stream inflows, relative to adjacent plots without stream inputs (controls). The taxonomic composition of stream macroinvertebrates drifting toward the lake differed from that in the littoral lake benthos itself, although there was no evidence of any species change in the composition of the littoral benthos brought about by stream inputs. 3. Although streams made no measurable contribution to the biomass or taxonomic composition of the littoral macroinvertebrate benthos, there was substantial temporal variation in biomass among the eight sites for each of the (n = 7) sample periods during which observations were made. Variation in total biomass was primarily a function of bottom slope and benthic substrata in the lake habitats. Dominant taxonomic groups were Baetidae, Ephemerellidae (two genera), Leptophlebiidae, Chironomidae (three subfamilies) and Perlodidae, although we did not determine the specific substratum affinities of each taxon. 4. Mixed effects linear models identified a significant interaction between macroinvertebrate biomass and plot type (stream inflow vs. control) associated with fish abundance. Across the observed range of macroinvertebrate biomass, fish showed a significant preference for stream inflows, but more closely tracked food availability in the controls. For young‐of‐the‐year (YOY), a negative effect of temperature was also included in the model, and we observed lower temperatures at stream inflows. However, abundance of predatory adults affected habitat selection for YOY. Lake‐bottom slope also accounted for variation in abundance in both fish models. 5. Our results suggest that the effect of fishless headwater streams on downstream fish may not always be through direct delivery of food. In this study system, fish preferred stream inflow plots, but this preference interacted with macroinvertebrate biomass in a manner that was difficult to explain. For YOY, predation risk was related to the preference for stream inflows, although the specific factor that mitigates predation risk remains poorly understood.