Empirically derived estimates of prey resources using conventional sampling methods are generally poor indices of actual food available to visually feeding benthivorous fish. We preferentially sampled the most active fraction of the benthos (presumably that most detectable by fish) utilizing short-term (4 d) colonization of artificial substrates, and used colonist biomass to explain temporal variation in stomach contents of bluegill sunfish (Lepomis macrochirus) in Lake Opinicon, Ontario. Simple linear regressions showed that total biomass of colonists explained 75% of the variance in the biomass of stomach contents. Independent variables using a limited taxonomic range or size range of colonists explained less variance in stomach contents. The convenience and simplicity of this methodology, and its ability to explain within-lake variation in food exploitation by visual benthivores, suggest that artificial substrate colonization could be useful for surveys of prey availability to benthivorous fishes.