Studies of phenotypic plasticity frequently ask how organisms respond to a change in their environment, but most organisms do not experience single environmental changes. Therefore, we need to move to the next step and understand how organisms respond to combinations of environmental changes. Recent studies of predator-induced plasticity have addressed how prey respond to different combinations of predators. I briefly review 22 studies of combined predator effects on prey phenotypes and identify four factors that make it difficult to interpret the results of these studies: (1) uncontrolled prey consumption, (2) a low number of prey traits, (3) a low number of predator combinations, and (4) confounded predator composition and total predator density.
I address these challenges in an experiment that examined how wood frog tadpoles (Rana sylvatica) altered 12 behavioral, morphological, and life historical traits in response to four different caged predators (Erythemis, Belostoma, Dytiscus, and Anax). The predators were present alone at low density, alone at high density (23), or combined into six pairwise combinations. When each predator was alone (at either low or high density), tadpoles
discriminated among different predators and produced predator-specific phenotypes. The doubling of predator density rarely induced more extreme prey phenotypes. When predators were combined, the tadpoles generally developed phenotypes that were similar to those induced by the more risky predator alone (90% of all traits examined, at either low or high density). These results suggest that tadpoles perceive the risk of combined predators as being similar to the risk of the most dangerous predator in the pair, and not as a summed or averaged predation risk. The actual risk from these predator combinations remains to be tested. This appears to be the first study to take a comprehensive approach that controls prey consumption, examines a large number of prey traits, uses a large number of predator combinations, and separates the effects of predator composition and predator density. There is a clear need for more such studies to determine whether these results can be generalized to other taxa.