Retrospective and behavioural studies to improve host specificity testing of parasitoids introduced as biological control agents

Abstract

Classical biological control is an important component of pest management, and has proven to be a highly cost effective long term method and often considered the only way to establish and maintain a self-sustaining reduction of insect pest populations. However, concern exists in regards to the risks that new biological control agents pose to non-target species in the country of introduction, and increasing attention is being paid to the estimation of such risk. Most countries have developed regulations in order to improve decision making in relation to the introduction of biological control agents, and the scientific community continue to discuss what laboratory host testing methods and/or guidelines should be followed to best predict the ecological host ranges of proposed biological control agents. In this study, I used the lepidopteran pest Uraba lugensWalker (Lepidoptera: Nolidae) and the parasitoid Cotesia urabae Austin & Allen (Hymenoptera: Braconidae) as a host-parasitoid model system to conduct retrospective post-release assessments of the host range of C. urabae and verify predictions made from host specificity testing prior to release in New Zealand. Firstly, while investigating methods to improve mating success in the laboratory, C. urabae males were proven to be attracted to odour volatiles of virgin females but not mated females in a Y-tube olfactometer, suggesting the importance of sex pheromones. Male mating success was highest in the presence of a male competitor, and larger males had greater mating success than smaller males when paired with a single conspecific female. Secondly, investigations on C. urabae odour attraction revealed that females respond to chemical cues specific to the host plant and target host larvae but not to target larval frass, and were significantly more attracted to odours of the entire target plant-host complex. In addition, the attraction of C. urabae to the odours of three non-target species (Pseudocoremia suavis, Nyctemera annulata, and Tyria jacobaeae) and their host plants was tested. I found that C. urabae females are attracted to the odours of the non-target larvae in the closely-related Erebidae family when tested on their own and when feeding on their host plants, but not to their non-target host plants alone. In a multiple comparison bioassay where these non-targets feeding on their corresponding host plants were tested simultaneously with the target host, C. urabae was significantly more attracted to odours of its target plant-host complex. Additionally, when Cotesia urabae females were given prior exposure and the opportunity to oviposit within either non-target Erebidae species, no increase in the attraction towards the non-target plant-host complexes was detected, suggesting that associative learning is unlikely to increase non-target attack. The use of Y-tube and 4-way olfactometers to assess odour attraction towards non-targets and their plant-host complexes is a novel way to increase knowledge on potential ecological host range during host specificity testing of potential biological control agents. Retrospective laboratory host specificity testing of C. urabae was conducted to evaluate the potential risk posed to the rare endemic Nolid moth Celama parvitis, the sole representative of the same family as the target in New Zealand. A high level of developmental failure of C. urabae within C. parvitis was observed, which confirms this non-target species as an unsuitable physiological host. However, behavioural observations suggest that host deprivation and prior non-target host experience may play an important role in increasing C. urabae responsiveness to non-target species, in this case C. parvitis. Additional retrospective post-release assessments, involving laboratory bioassays and multiple choice testing under semi-field conditions using field cages, was conducted to assess potential non-target impacts of C. urabae on P. suavis, N. annulata, and T. jacobaeae. Laboratory testing and rearing proved N. annulata to be a suitable physiological host of C. urabae although poorer compared to the target host U. lugens. The field-cage experiments suggest that the probability of non-target impacts of C. urabae on N. annulata in the field are likely to be low, but this has the potential to be higher in December which is the time of year between U. lugens larval generations and when N. annulata larvae are present in New Zealand. The research presented in this thesis reveals the importance of conducting behavioural studies and the value of continued retrospective host-range assessment to improving risk assessment of parasitoids introduced as biological control agents. Furthermore, this thesis highlights the benefits that can be obtained by gaining knowledge of odour attraction in parasitoids to assist in more accurately predicting the host ranges of candidate biological control agents.