The choice of host plant species for polyphagous insect herbivores can have strong repercussions for insect fitness. While traditionally examined from a bi-trophic perspective, host plant identity can also influence patterns of natural enemy-mediated mortality through its effects on predator density, the strength of apparent competition, and via diet-induced changes in herbivore development time and antipredator/antiparasitoid behaviors and physiological responses. Using larvae of a common skipper butterfly, Epargyreus clarus, we measured predation by comparing the survival of exposed vs. protected larvae on each of 6 commonly used leguminous host plants in a common garden over 8 generations (3 years). Diet-induced variation in development time was related to predation to test the slow-growth-high-mortality hypothesis. In the same garden, we also compared the densities of naturally colonizing arthropod prey and predators among host plants. In addition, we reared naturally occurring E. clarus caterpillars from a subset of these same host plants growing naturally at local field sites and estimated parasitism rates over multiple years. Finally, in the laboratory, we tested the effects of host plant diet on antipredator behaviors and resulting mortality from captive paper wasps.
Results/Conclusions
Predation pressure, as estimated from the log response ratio of survival in exposed vs. protected treatments, varied significantly among host plants in most generations. For exposed larvae only, the effects of host plant were stronger in the late season, which is also when natural enemy densities peaked. Arthropod predator community structure differed among host plants and densities tended to covary with total prey availability, as estimated by insect herbivore abundance. Host plant differences in predation, however, were not related to differences in caterpillar development time in any of the generations, providing no support for the slow-growth-high-mortality hypothesis. Parasitism of naturally occurring E. clarus larvae collected from four of the six host plants over a four year period (2013-2016) exhibited strong patterns of host plant dependence, with the gregarious braconid Apanteles argynnidis differentially attacking larvae reared from black locust and the solitary braconid Bassus spiraculis primarily attacking hosts reared from kudzu. Finally, levels of aggression and tendency to regurgitate in response to simulated predation of E. clarus larvae differed among rearing hosts, as did survival rates in predator arena trials using captive paper wasps.