Abstract
Non-target pesticide exposure causes a myriad of detrimental sub-lethal effects in a broad range of taxa, with pesticide-induced changes in behaviour, physiology and reproduction documented widely. Most studies have focused on vertebrate models, while insects are among the least-studied taxa in this field. Within insect-related literature, focal pesticides are often specifically insecticidal; however, several insect species have been observed feeding on pesticide baits used for mammalian predator control in the wild, introducing the need to investigate the effects of these pesticides on insects. New Zealand is one of the largest consumers of vertebrate pesticides globally, largely due to an overarching conservation goal of eradicating mammalian predators from the country by 2050. Unfortunately, non-target consumption of pesticide baits has been frequently observed in the iconic endemic taxa, weta. Limited studies suggest that these baits are non-lethal to weta, but research never extends beyond mortality-related studies. In this thesis, 34 Wellington tree weta (Hemideina crassidens) were collected from Wellington; half were exposed to an anticoagulant rodenticide bait, brodifacoum, for 40 days, while the other half served as a control group. A range of behavioural and other sub-lethal measurements were taken, none of which have been investigated in weta before, to determine whether exposure to brodifacoum has any sub-lethal effects.
Behavioural effects of brodifacoum exposure were measured in chapter 2 using video recordings, analysed with EthoVision XT behavioural quantification software, and observational data. Four biologically important behaviours, emergence (presence outside of the refuge), activity (total distance moved), refuge-seeking (i.e. boldness; latency and tendency to enter a refuge) and aggression (response to a facial probe), were compared between the treatment and control groups to detect pesticide-induced behavioural differences. Repeatability of the behaviours was also tested to determine whether the baits had an effect on behavioural consistency. Brodifacoum exposure was shown to alter expression of all four behaviours to some extent, with treated individuals showing increased emergence but decreased activity, aggression and boldness compared to the controls by the end of the treatment period. All four behaviours were repeatable overall, but most varied between sexes and/or treatment groups, suggesting that H. crassidens show both natural behavioural differences and sensitivity to pesticide exposure. Results from this study indicate that pesticide baits can alter behaviours directly related to foraging, mating, competitive interactions and anti-predator responses in a non-target insect, which could have additional detrimental effects on fitness.
Brodifacoum-induced physiological and reproductive changes were measured in chapter 3. All 34 weta were weighed regularly over the 40-day treatment period, and the 5 heaviest females from each group were dissected at the end of the trial to record fat composition (as a measure of body condition), parasitic infection intensity, number of eggs and developmental stage (early or late) of those eggs. Brodifacoum treatment did not affect weight or parasite intensity, but brodifacoum-exposed females had significantly less fat and more eggs than the controls, with a greater proportion of those eggs in the early stage of development. While these results are preliminary due to small sample size, the findings suggest that even short-term exposure can alter reproductive output and body condition, which could result in reduced fecundity and longevity of H. crassidens populations.
The results from this study clearly show that brodifacoum exposure has the potential to induce several sub-lethal responses in a non-target insect, H. crassidens. This thesis fills several gaps in the literature, including behavioural responses of an insect to a non-insecticidal pesticide, multi-faceted effects of a single pesticide on a non-target organism, and how New Zealand’s predator-free 2050 goal may actually be damaging to some endemic species in ways not previously acknowledged. Comparable studies should be conducted to investigate sub-lethal responses of weta to other widely-used pesticide baits, such as 1080 and diphacinone, especially as pesticide use in New Zealand continues to increase. These results should then be used to determine which pesticides have the least detrimental impacts on non-target wildlife, including weta, and regulations on pesticide use should be altered accordingly. This thesis highlights the fundamental need to consider sub-lethal effects of pesticide exposure in non-target consumers, as changes in behaviour, physiology and reproduction undoubtedly affect other fitness-related factors such as inter- and intra-specific interactions, foraging efficacy and ultimately, survival. Importantly, the goal of this research is not to devalue the use of pesticides as a conservation tool, but rather to make suggestions to reduce their non-target impacts.