The mechanisms by which plants modify herbivore suppression by carnivores are important yet infrequently studied at a community-wide level. Interaction modifications can arise from two distinct mechanisms (trait-related and density-related) with different consequences for the evolutionary ecology of species interactions. Here we study how variation among tree species in food quality for caterpillars modifies herbivore suppression by birds in a temperate forest community via trait-related and density-related mechanisms. The trait-related mechanism, also known as the slow-growth/high-mortality hypothesis, entails heightened risk of predation on herbivores due to their relatively slow growth on tree species of relatively poor food quality. The density-related mechanism, which we are calling the high-performance/high-mortality hypothesis, gives an alternative scenario in which predation on herbivores peaks on tree species of relatively high food quality because such trees sustain relatively high herbivore densities, eliciting a relatively positive density-dependent predation response. We tested predictions of these mechanisms by quantifying effects of avian exclusion on caterpillar density across eight common tree species in a 2-year field experiment, and comparing growth performance of the most abundant generalist caterpillar species across these eight tree species in a 5-year study.
Results/Conclusions
We found evidence for both trait- and density-related mechanisms. Controlling for caterpillar density, the per capita effect of bird predation on caterpillars was highest on poor quality host-plant species. Without controlling for caterpillar density, the magnitude of bird predation on caterpillars was highest on high quality host-plant species, which also contained the highest caterpillar densities. These seemingly opposing results suggest that tri-trophic interactions in this forest community involve both trait- and density-related mechanisms of interaction modification. These observations represent the first community-wide test of the slow-growth/high-mortality hypothesis, and highlight some of the pitfalls in testing it. Our findings also suggest the ecological dominance of density-dependent interactions in this community. Finally, we hypothesize that the combination of trait- and density-related mechanisms determines evolutionary consequences for specialization of plant-herbivore interactions.