Understanding the host finding behavior of a biological weed control candidate specialist as a contribution to pre‐release risk assessments.

Pre‐release host range assessments of weed biocontrol agent (BCA) candidates typically rely on no‐choice and choice feeding, oviposition, and development tests. However, these tests may exclude potentially environmentally safe BCA candidates from consideration if they develop on nontarget plant species in no‐choice tests that they would not colonize in the field because of behavioral barriers during host selection. Here, we examined the behavioral responses of Ceutorhynchus rusticus Gyllenhal (Coleoptera: Curculionidae) to olfactory and visual plant cues of nine native North American and two Eurasian confamilial plant species of invasive Eurasian mustard dyer's woad, Isatis tinctoria L. (Brassicaceae). In behavioral bioassays with olfactory, visual, and combined plant cues of one plant species vs. purified air and/or empty arms (control treatments), weevils responded with attraction to all offered I. tinctoria plant cues but showed no preference for most nontarget plant species relative to control treatments. The weevil also responded faster in these bioassays to I. tinctoria cues when compared to most nontarget species. In bioassays with olfactory, visual, or combined cues of nontarget species vs. those of I. tinctoria, C. rusticus preferred I. tinctoria over most nontarget plant species (olfactory cues: eight of 11, visual cues: two of 11, and combined cues: 11 out of 11 nontarget plant species). The results also suggest that C. rusticus mainly uses olfactory or combined olfactory and visual plant cues rather than visual cues alone in the early stages of host finding. Additionally, the findings suggest that during host finding, C. rusticus prefers cues of its Eurasian field host over those of nontarget species, reducing risks of post‐release nontarget attack in the field. Assessing behavioral responses to multimodal cues can advance our understanding of BCA candidates' host discrimination and facilitate the accurate interpretation of conventional no‐choice and choice feeding and development data during pre‐release environmental safety assessments. [ABSTRACT FROM AUTHOR]