Because adult yellow perch were known to be cannibalistic in Oneida Lake, New York, this species was studied in 1965-71 to identify factors affecting the intensity of cannibalism and to evaluate the effect of cannibalism on the abundance of young perch. Adult perch changed their food habits in response to changes in the availability of different food items from June through September. Although young-of-the-year perch were most abundant in adult stomachs in August of most years, wide variations occurred in the annual intensity of cannibalism. Length and abundance of young perch, abundance of large chironomids and amphipods, and the size of the adult predator were identified as important factors for explaining variations in the frequency of occurrence of young perch in stomachs of adults. Correlations between the annual intensity of perch cannibalism and length and abundance of young perch imply that cannibalism operates as a compensatory or possibly an extrapensatory mortality process. The consumption of young perch by adults in 1971 was estimated from the incidence of young perch in stomachs, the rate of food evacuation, and the size of the adult perch population. About 25% of the perch fry that were available as forage in early June succumbed to cannibalism in June-September. This level of consumption of young perch by adults in a year of intense cannibalism, together with evidence that perch cannibalism likely does not operate in a manner consistent with the observed depensatory mortality of young perch, imply that cannibalism was not decisive in limiting the strength of year classes in 1965-71.