Publication

Trophic cascades are regarded as important signals for top-down control of food web dynamics. Although there is clear evidence supporting the existence of trophic cascades, the mechanisms driving this important dynamic are less clear. Trophic cascades could arise through direct population-level effects, in which predators prey on herbivores, thereby decreasing the abundance of herbivores that impact plant trophic levels. Trophic cascades could also arise through indirect behavioral-level effects, in which herbivore prey shift their foraging behavior in response to predation risk. Such behavioral shifts can result in reduced feeding time and increased starvation risk, again lowering the impact of herbivores on plants. We evaluated the relative importance of these two mechanisms, using field experiments in an old-field system composed of herbaceous plants, grasshopper herbivores, and spider predators. We created two treatments, Risk spiders that had their chelicerae glued, and Predation spiders that remained unmanipulated. We then systematically evaluated the impacts of these predator manipulations at behavioral, population, and food web scales in experimental mesocosms. At the behavioral level, grasshoppers did not distinguish between Risk spiders and Predation spiders. Grasshoppers exhibited significant shifts in feeding-time budget in the presence of spiders vs. when alone. At the grasshopper population level, Risk spider and Predation spider treatments caused the same level of grasshopper mortality, which was significantly higher than mortality in a control without spiders, indicating that the predation effects were compensatory to risk effects. At the food web level, Risk spider and Predation spider treatments decreased the impact grasshoppers had on grass biomass, supporting the existence of a trophic cascade. Moreover, Risk spider and Predation spider treatments produced statistically similar effects, again indicating that predation effects on trophic dynamics were compensatory to risk effects. We conclude that indirect effects resulting from antipredator behavior can produce trophic-level effects-that are similar in form and strength to those generated by direct predation events.