A self-training device to teach conservation-working dogs to avoid poison baits

Context: Accidental poisoning of domestic dogs is a potential risk when using baits to control invasive animals. We developed and trialled an electrical device attached to a non-toxic bait to assess whether we could induce a learned aversion towards baits in conservation-working dogs. Aims: We tested the device on conservation-working dogs licenced to enter conservation estate as part of feral pig control, and consequently are potentially exposed to lethal baits for controlling red foxes. Methods: Over 1 year (up to seven separate training sessions per dog), 27 dogs were sequentially presented with electrified and non-electrified non-toxic baits and their behavioural responses were recorded. On-farm training (Days 0, 1, 7, Months 1, 12) comprised dogs being called by their owner standing nearby the electrified bait. If the dog touched the bait and demonstrated aversive behaviour (we assume that it received an electric shock or 'correction'), it was then presented with a non-electrified bait. If they ate either bait, they were shown another electrified bait (up to three electrified baits per session). Key results: Seventeen dogs (17/27) touched the bait and received a correction. Eleven dogs required only a single correction and did not touch another bait, three dogs needed two corrections, and two dogs needed three or four corrections. These 17 dogs showed increasing avoidance to the bait over successive training sessions (χ212 = 67.96, P < 0.001), including avoiding looking at the bait, refusing to come within 5-10 m of the bait and their owner, or leaving the training vicinity. All these dogs (17/17) avoided baits encountered in a working environment (1/17 touched but did not consume a bait) and bait-aversion was still detected up to 1-year post-initial training. Nine dogs (9/27) did not appear to receive a correction or show any change in bait-aversion behaviour. One dog (1/27) showed no aversion to the stimulus and continued to eat baits. Conclusions: Here we present a proof of concept for a deterrent device and associated experimental protocol to produce learned aversion behaviour in conservation-working dogs. Implications: We demonstrated that it is possible to induce a learned aversion to baits in conservation-working dogs, thereby reducing the risk of accidental poisoning