Identifying food web linkages between biocontrol agents of invasive plants and native species is crucial for predicting indirect non-target effects. Biocontrol insects can integrate into food webs within recipient habitats and influence native insects through apparent competition (altering shared natural enemies) or density-mediated exploitation competition (changing density of native plants). However, whether and how trait-mediated exploitation competition (modifying native plant chemical defenses and volatiles profiles) can produce indirect non-target effects remains largely overlooked, despite plant phenotypic responses to insect herbivory being common and widely documented.

The beetle Agasicles hygrophila was introduced into China for management of weedy Alternanthera philoxeroides, but it also attacks the native congener A. sessilis, which may have indirect non-target effects on the native beetle Cassida piperata that also feeds on A. sessilis. Here, we examined the relationships among the abundance of A. hygrophila and C. piperata, and A. sessilis coverage in the field. Then, we investigated the impact of A. hygrophila herbivory on C. piperata development and oviposition, as well as A. sessilis primary metabolites and leaf volatiles.

Cassida piperata abundance was not related to A. sessilis coverage, but C. piperata was less abundant on plants with more A. hygrophila in the field survey and on plants with more prior A. hygrophila damage in the field cage experiment. Cassida piperata development was inhibited by prior A. hygrophila herbivory in bioassays in enemy-free conditions and they preferred to oviposit on A. sessilis plants that had experienced little or no A. hygrophila damage. Agasicles hygrophila herbivory decreased A. sessilis foliar glucose and protein, and substantially changed its leaf volatile blend.

Synthesis. Our results show that A. hygrophila has its major indirect non-target effects on C. piperata through trait-mediated exploitation competition, rather than apparent competition or density-mediated exploitation competition. Our results demonstrate a new mechanism for indirect non-target effects of biocontrol agents. Furthermore, our results indicate that minor and temporary negative impacts on non-target plants might propagate to higher trophic levels and such negative impacts can strengthen with increasing intensity of the direct non-target effect.

Data were collected form field survey, field cage experiment, bioassays and chemical analyses.

Data were analyses using SAS.