PS 16-229 - Phloem-feeding herbivores completely suppress parasitism of caterpillars on oak trees

Abstract: Tri-trophic interactions mediated by herbivore-induced plant volatile organic compounds (VOCs) are well established from laboratory experiments in model systems. However, relatively few studies have been done in field settings, especially in natural ecosystems. This research gap has spawned persistent questions about the importance of volatile-mediated tri-trophic interactions in complex ecological networks and variable environmental conditions. The current study embraces these complexities by investigating the indirect effect of phloem-feeding insects on caterpillar-parasitoid interactions on a shared host plant. We studied these indirect interactions on white oak trees (Quercus alba) in forests of Connecticut, U.S.A. Research from agricultural and model systems motivated our leading hypothesis that phloem-feeders would indirectly reduce parasitism of caterpillars via plant-mediated phytohormonal crosstalk. Such research shows that phloem-feeders typically induce the salicylic acid (SA) pathway, which often suppresses the jasmonic acid (JA) pathway induced by caterpillars. Because the JA pathway controls plant emission of VOCs known to attract parasitoids of caterpillars, we predicted that phloem-feeders would indirectly reduce parasitism of co-occurring caterpillars. Our alternative hypothesis was that phloem-feeders might indirectly reduce parasitism of caterpillars via an ant-mediated mechanism. Namely, ants attracted to the honeydew of phloem-feeders were predicted to deter parasitoids that might otherwise attack nearby caterpillars. We tested these hypotheses with a field experiment in which we manipulated phloem-feeders (removal/removal and replacement) on tree branches, stocked each branch with randomly assigned wild caterpillars, and measured the parasitism rates of caterpillars recovered after 5-7 days. This experiment was conducted at five field sites over four years. In two years, the phloem-feeder manipulation was factorially crossed with an ant manipulation (exclusion/access) to test our ant-mediated hypothesis. We found that caterpillar mortality from parasitoids was completely suppressed when phloem-feeders were replaced (0% parasitism) versus removed (17% parasitism). Ant exclusion had no effect on parasitism of caterpillars. The effect of phloem-feeders on parasitism was observed in multiple caterpillar species. The lack of ant-mediated and caterpillar-specific effects strongly suggests that the effect of phloem-feeders on caterpillar-parasitoid interactions is plant-mediated, providing ecological validation for the large body of mechanistic research conducted in isolated environments. The complete suppression of caterpillar-parasitoid interactions by phloem-feeders on oak trees is a surprisingly dramatic indirect interaction with important implications for community and evolutionary ecology, biological control, and forest management.