Locomotory responses to olfactory cues during host‐finding can inform environmental safety assessments of biological weed control agents

Abstract

In weed biological control programs, the safety assessment of an agent currently relies on no‐choice and choice feeding, development, and oviposition cage tests. However, confined conditions restrict insect behavior, potentially producing false positive host‐use results that can prematurely disqualify agents, reducing opportunities for successful weed biocontrol. To minimize risks of nontarget attack and maximize successful biocontrol outcomes, it is useful to investigate an agent’s pre‐contact responses to olfactory cues from target and nontarget plant species during host‐selection. Here, we tested how the locomotory behavior of the biocontrol agent Mogulones crucifer Pallas (Coleoptera: Curculionidae) differed when exposed to olfactory cues from its host, the invasive herbaceous Cynoglossum officinale L. (Boraginaceae), versus cues from closely related nontarget species. We used olfactometer behavioral assays, comparing M. crucifer responses to volatiles from C. officinale and nontarget plant species. We observed and measured the duration of resting and walking behaviors, as well as frequencies of arrestment or persisted resting, by M. crucifer in response to these odor sources. In total, 22 confamilial plant species were tested, including five species federally listed as threatened or endangered (T&E) in the USA. Mogulones crucifer consistently responded to C. officinale volatiles with greater duration of resting and frequencies of arrestment behavior compared to other plant species. In contrast, M. crucifer behavior was unaffected by odors of confamilials of C. officinale or responded with repellence to these odors in the olfactometer. These data suggest that in the invaded North American range, M. crucifer is unlikely to use nontarget olfactory cues to find or come into contact with tested nontarget plants, including the tested T&E species. Results from this study improve predictions of M. crucifer’s host‐use in the introduced range and illustrate the potential of this novel approach as a component of pre‐release environmental safety assessments of weed biocontrol agents.

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

“…Because our data suggest strong host discrimination by the weevils, previously documented nontarget attacks from traditional host specificity experiments could be an experimental artifact. The data presented here, however, should be treated with caution as herbivores' responsiveness to plant cues can be affected by factors like learning or physiological status, environmental cues (Park et al 2018), and studying additional behavioral components of weevils is merited, such as walking, resting, and arrestment (e.g., Fung et al, 2021). Studies such as ours, Park et al ( 2018), and Fung al.…”

“…However, few such studies exist, and of these most focus on only one plant cue modality, i.e., olfactory cues (Andreas et al, 2008;Fung et al, 2021) or visual cues (Reddy et al, 2009;Müller & Nentwig, 2011). Studies investigating two plant cue modalities simultaneously are scarce in weed biological control research (but see Park et al, 2018Park et al, , 2019 and present an important gap as host selection behavior by insect herbivores often involves multiple modalities acting simultaneously (Harris et al, 1982(Harris et al, , 1990.…”

“…Studies of the host selection of BCA candidates have been recommended to complement traditional laboratory and field host range testing methods to improve predictions of post‐release host use (Wheeler & Schaffner, 2013). However, few such studies exist, and of these most focus on only one plant cue modality, i.e., olfactory cues (Andreas et al, 2008; Fung et al, 2021) or visual cues (Reddy et al, 2009; Müller & Nentwig, 2011). Studies investigating two plant cue modalities simultaneously are scarce in weed biological control research (but see Park et al, 2018, 2019) and present an important gap as host selection behavior by insect herbivores often involves multiple modalities acting simultaneously (Harris et al, 1982, 1990).…”

Biological control of weeds – Context dependency matters

“…These typically rely on no‐choice and choice feeding, oviposition, and development rates of individual agents in the laboratory, and the resulting predictions for attacks on non‐target plant species are >99% accurate (Hinz et al., 2020). However, two papers in this issue warn that these traditional studies may exclude potentially safe agents that can to some extent feed, oviposit, or develop on non‐targets in such conservative setups, but that are not expected to attack them in the field (see also Paynter et al., 2020; Fung et al., 2021). Therefore, they suggest additional methods to better assess risks to non‐targets in the field for pre‐release environmental safety assessments.…”

Using Chemical Ecology to Enhance Weed Biological Control

“…For instance, the discovery that Galerucella males fail to emit aggregation pheromone when feeding on non-target suboptimal hosts [19] provides additional evidence of host specificity and safety of the loosestrife beetles. There have been several recent studies addressing differences in VOC profiles between biocontrol target weeds and native nontarget species [61][62][63][64][65]. These studies focused on the behavioral responses to the VOCs and the implications for host selection and safety.…”