Optimal Foraging and Habitat Shift in Perch (Perca Fluviatilis) in a Resource Gradient

To better understand how and why foraging shifts between habitats and among prey occur, and to evaluate the ability of an optimal foraging model to predict diet breadth quantitatively, we studied the food and habitat use of perch (Perca fluviatilis) in an experimentally induced resource gradient. The purpose of the study was also to see whether prey—type—dependent search rates were important for observed resource utilizations. The resource level in one (pelagic) of two (pelagic and benthic) habitats was selectively reduced in a field experiment by varying the densities of roach (Rutilus rutilus), and interspecific competitor of perch (Perca fluviatilis) that feeds preferentially on pelagic resources. We tested whether a simple optimal foraging model could predict (1) the diet breadth of perch and (2) when perch should shift from feeding in the pelagic habitat to the benthic habitat. The experiment was carried out in a pond divided into 10 enclosures. Perch density was held constant between enclosures, whereas the density of the planktivorous competitor was varied among enclosures. The smallest size classes of pelagic prey used by perch correspondence with what was predicted by the basic prey model. However, the mean prey size in the diet of perch was larger than that predicted by the model. For benthic prey, diet breadth was poorly predicted, reflecting the problem of extrapolating laboratory—estimated encounter rates and handling times to more complex habitats in the field. During the first half of the experiment, but not during the second half, there was a positive correlation between the growth rate of perch and the predicted net return rate from feeding on pelagic prey, based on the foraging model. Conversely, no significant relationship between perch growth performance and predicted net energy intake from the benthic habitat was present during the first half of the experiment, while a positive relationship was present during the second half. There was a time lag between predicted and observed shifts in habitat use from pelagic to benthic feeding. In addition, the habitat shift was never complete since perch continued to feed on some pelagic prey after having shifted to feeding mainly on benthic prey. Different optimal search rates for prey in different habitats may have given rise to time lags in habitat shifts as well as to partial habitat use. Habitat—dependent search rates may thus form a complementary explanation to learning and sampling behavior, for observed time lags and partial habitat use. Flexible predator search rates reinforce the importance of considering the whole behavioral repertoire of predator and prey, not only the predator's choice once prey has been encountered.