Pathways for plant-mediated negative feedback to insect herbivores: accounting for non-linear effects of larval density on plant quality and quantity

Changes in plant traits induced by herbivore damage can produce a negative feedback to increasing herbivore densities. Several aspects of these plant-mediated feedbacks are predicted to influence herbivore population dynamics, but the degree to which feedbacks are non-linear, whether their strength varies among plant genotypes, and whether they occur via changes in the quality of plant tissue or the amount of tissue available have rarely been examined. In this study, we damaged five genotypes of the perennial weed Carolina horsenettle, Solanum carolinense L. (Solanaceae), with eight densities of the false potato beetle, Leptinotarsa juncta (Germar) (Coleoptera: Chysomelidae, Chrysomelini). To account for plant quantity available to beetles, we measured the leaf area of each plant after imposing larval density treatments. We then measured the oviposition preference of L. juncta adults for damaged vs. undamaged plants and the relative growth rate (RGR) of L. juncta larvae on each plant, as well as plant trypsin proteinase inhibitor (tryPI) expression. We found that L. juncta females strongly preferred to oviposit on undamaged S. carolinense, and that preference for a plant decreased as the density of damaging larvae increased. In addition, we found decreased larval RGR and increased production of tryPIs with greater initial larval density. Effects of larval density on insect preference, performance, and tryPI expression were linear and did not vary among plant genotypes. Larval density effects were not solely due to reduced plant quantity, as plant leaf area had no effect on oviposition preference or larval performance. This suggests that effects were most likely due to changes in the quality of plant tissue. Thus, negative feedback on increasing larval density can be mediated by the effects of induced resistance on both oviposition preference and larval performance. Published models suggest that the linear, quality-mediated feedbacks observed in this experiment are predicted to stabilize herbivore population dynamics.