Species interaction strengths in food webs are variable over time and space, with most webs having a few strong interactions and many weak interactions. Such variation has important consequences for population- and community dynamics, yet empirical studies that have quantified interaction strengths across replicated food webs or explored factors underlying observed variation are rare. In the present study, we used an observational approach to quantify taxon-specific feeding rates of reticulate sculpin (Cottus perplexus) on their diverse assemblage of aquatic invertebrate prey at nine stream sites in Western Oregon. Our approach combined prey-specific gut clearance rates obtained in the laboratory with sculpin diet surveys and estimates of predator and prey abundances from field data. We assessed the relative importance of four factors in driving variation in feeding rates: 1) predator density, 2) prey density, 3) predator-prey body mass ratios, and 4) a suite of five abiotic environmental variables (summarized using a principal components analysis). We used AIC to compare the relative performance of models that variously included these factors. Based on food web theory, we predicted that feeding rates would correlate positively with prey abundances and predator/prey body size ratios, and negatively with predator abundances.
Results/Conclusions
Sculpin consumed over 70 prey taxa, with the majority being aquatic insects in the orders Ephemeroptera, Diptera, Trichoptera, and Plecoptera. Taxon-specific feeding rates varied over three orders of magnitude and showed a highly skewed distribution at the site-level, with a few strong and many weak interactions. In contrast, feeding rates on each given prey taxon varied little between the nine stream sites. Taxon-specific feeding rates correlated positively with prey density, such that mean feeding rates increased 40-fold from the lowest to the highest density prey taxa. Feeding rates also correlated negatively with the density of sculpin themselves, consistent with predator interference. A model including only prey density and a model including both prey- and predator density performed equally well at explaining variation in feeding rates. Incorporating abiotic conditions (i.e., water temperature, stream flow, light levels, substrate size, and stream width) or predator-prey body mass ratios did not improve model performance. The highest feeding rates were observed at intermediate predator/prey body mass ratios, which varied by over 500-fold in the dataset. Our results emphasize how community structure and the densities of interacting species contribute to driving variation in interaction strengths within complex food webs.