The apparent diet of predators and biases due to different handling times of their prey.

In order to estimate the true diet of predators, the prey of a number of predators is recorded at one time. Such sampling underestimates the true diet during a period of time. Where handling times of different types of prey are very different, these estimates will be biased, because prey that take a relatively long time to be eaten will be overestimated. We examined a rocky intertidal predatorprey system and demonstrated the existence of such bias. A number of hypothetical correlates of the bias were also investigated. As anticipated, variations in handling times were a major factor, but neither taxonomic affinity nor absolute size of the prey could predict the degree of underestimation in the true diet for any given type of prey. A previously described correction for this type of bias was tested, but found to be unsatisfactory. We suggest that it was too insensitive to variability in handling times.A simple computer model incorporating differences among prey in their handling times was also unable to predict the bias, but did indicate that non-random selection of prey was occurring.We concluded that where such biases are likely to occur, information on the handling times of different prey and/or accurate estimates of the true diets of the predators are essential for the predatory interaction to be interpreted properly. These results were discussed in relation to published accounts of diets of predators in rocky intertidal habitats. Many studies have not presented data on handling times of prey in the field, and the magnitude and importance of potential biases in these studies are therefore difficult to assess.