Populations in nature face a number of factors that can alter their traits and subsequent performance. Predation is one factor that can have widespread effects on the mean trait value in a prey population because predators can impact prey traits through a number of processes. Predators can directly induce prey phenotypes through visual and chemical cues, predators can indirectly alter prey phenotypes by thinning the prey population (thereby reducing competition), and predators can cause selection on prey phenotypes through nonrandom killing. Ecologists are beginning to understand each of these three processes in isolation, but these processes act on prey phenotypes simultaneously, and we lack information on the relative importance of these processes in determining the final phenotype of prey.
I used a system of gray treefrog tadpoles (Hyla versicolor) and dragonfly predators (Anax longipes) to determine the separate and combined impacts of induction, thinning, and selection on the behavior, morphology, and growth of tadpoles reared in aquatic mesocosms. Using combinations of hand thinning, cues emitted from caged predators, and lethal predators, I demonstrated that the impact of lethal predators on growth was mediated primarily through thinning, the impact on morphology was primarily through induction, and the impact on behavior was affected similarly through thinning and induction. Surprisingly, while we know from numerous studies that the dragonflies kill tadpoles nonrandomly, selection did not have a significant impact on the final phenotypes of the tadpoles. This work appears to be the first study that identifies the relative magnitudes of the three predatory processes on a suite of prey phenotypes. The three processes can have opposing or supporting effects, the relative magnitudes of which likely differ among systems, illustrating why we often observe contradictory results when we examine the impact of lethal predators on prey phenotypes. Thus, it is important that we identify the mechanisms underlying the outcomes of predation, rather than simply observing the outcomes and then inferring which processes are responsible.