Background/Question/Methods Plant defenses can both deter herbivores and affect trophic interactions. The consequences of plant defenses for herbivores may influence higher trophic levels by affecting the quality of herbivores as prey or altering their susceptibility to enemies. With regards to herbivore suppression, plant defenses work synergistically with predators in some situations and antagonistically in others. In order to investigate interactive effects of defense and predation on herbivores, and to ascertain whether these effects are density-dependent, we conducted a fully factorial experiment in which predation and plant defense were manipulated across a range of naturally observed herbivore densities. To manipulate these factors, we assembled artificial communities in screened cages in a greenhouse. Cages contained four tomato plants (Lycopersicon esculentum), 5, 10, 20 or 40 beet armyworm herbivores (Spodoptera exigua) and either a pair of predatory stinkbugs (Podisus maculiventris) or no predators. Tomato plants either had constitutively low defenses (JAI-1 mutants that are unable to produce inducible defenses) or constitutively high defenses (variety Castlemart treated with jasmonic acid). After 48 hours the abundance and developmental stage of surviving herbivores were recorded. Predation rates and dynamic interaction index values were determined by comparing herbivore abundance in cages with and without predators.

Results/Conclusions High levels of plant defense resulted in increased predation rates and stronger predator-prey interactions. Herbivore density also significantly affected predation. Higher densities of herbivores were associated with weaker interaction strengths and a greater number of total prey killed by predators. The statistical significance of plant defense effects on interaction strength depended on the level of prey density. Dynamic interaction indices were not significantly different between plant defense treatments at densities of 10 or 40 herbivores per cage. However, interactions were significantly stronger when plants were highly defended at densities of 5 and 20 herbivores per cage (116% and 168% changes in average index magnitude, respectively). The developmental stage of caterpillars was not affected by different levels of plant defenses, implying that effects of defense on predation were not due to slowed herbivore growth, but instead mediated by behavioral responses. Herbivores tend to abandon plants that are highly defended, so defenses could have increased the likelihood of predator-prey encounters by increasing prey movement rate. Our results suggest that plant defense effects on herbivore behavior can act synergistically with predators to suppress herbivore populations, and that such synergism is density-dependent.