Prey-predator size-dependent functional response: derivation and rescaling to the real world

1. We derived a functional response function that accounted for both predator and prey size. Laboratory experiments were then performed to investigate the size relationships of the functional response of the benthic isopod Saduria entomon (L.) feeding on Monoporeia affinis (Lindstrom) amphipods. 2. The S. entomon predators, in general, showed sigmoid (type III) functional responses. The attack rate of S. entomon was a bent ridge-shaped function of predator and prey size. The maximum attack rates of small predators were found on small prey, and then attack rate declined rapidly as prey size increased. In large predators attack rate increased slowly up to a maximum for large prey. The shape of the attack rate was dependent on the size scaling of predator and prey body mass in the attack rate function. We suggest that the ridge-shaped attack rate function found in our study is likely to be general for predators that feed on prey with a large size range. The handling time of small S. entomon increased rapidly as prey size increased. In contrast, we found only small differences in handling time of large predators. The complete functional response may provide size-refuges for very small and large M. affinis depending on the size-structure of S. entomon. 3. To investigate the realism in our functional response estimates, we incorporated the functional response in a consumption model based on the densities and size-structures of the predator and the prey in the field (Bothnian Sea). We adjusted the attack rate using a correction factor to predict the observed survival of the different age-classes of M. affinis from field samples. The field data-based consumption model accounted for 1 year of size-dependent predation by the size-structured population of S. entomon on M. affinis. In addition, using a growth model for S. entomon we checked that the predicted consumption by the predator gave realistic growth curves. We found that the experimentally derived attack rates may have been over-estimated by a factor of 400.