Abstract
The reliability of fecal analysis for dietary assessment was evaluted in four species of insectivorous mammals. In separate experiments, invertebrates were fed to captive animals by one author, and feces were analyzed, without prior knowledge of diet composition, by the other author. There was overall agreement between ranked frequencies of different taxa of invertebrates consumed and recovered in feces. However, underestimation of numbers of invertebrates in feces was more evident (1) as increasing numbers of the same type of invertebrate were ingested, (2) for soft-bodied than for hard-bodied invertebrates, (3) for small than for large invertebrates, and (4) for well-fed than for hungry insectivores. We conclude that fecal analysis is a relatively reliable method for determining the diet of generalist insectivores that eat hard-bodied prey, especially if the results are expressed as the number of animals in which a prey item is found rather than as the minimum number of prey eaten per animal. However, we caution that there are more potential sources of bias in fecal analysis than realized previously.