The importance of an audio cue warning in training sheep to a virtual fence and differences in learning when tested individually or in small groups

Abstract

In previous studies, livestock have been effectively trained to virtual fencing systems by learning an association between an audio cue and an electrical stimulus (collectively, ‘the virtual fence’), however, there was a large variation shown in ability to learn. The aim of this study was to determine the influence of temperament on learning this association, and to determine the importance of the audio warning cue in learning and behavioural responses to the virtual fence. Forty-six ewes underwent temperament testing, and 40 were assigned to one of two groups (n = 20/group), Group 1 (AS) with both an audio cue warning (Audio cue 2 s) and electrical stimulus (<1 s, after audio cue), and Group 2 without an audio warning (S - Electrical stimulus <1 s). Both groups were fitted with manually controlled collars and were trained to a virtual fence over two weeks. The number of approaches, number of audio cues and electrical stimuli and behavioural responses to the virtual fence were recorded. Throughout training the AS group approached the fence more often than the S group (average number of approaches were 11.2 ± 0.83 vs 4.5 ± 0.46, P < 0.05), possibly because S group did not have an audio warning of the fence and therefore could not avoid receiving an electric stimulus if they were in the vicinity of the fence, however there was no difference in the number of electrical stimuli received (P = 0.3). Sheep approached the virtual fence significantly less over each of the training days (P < 0.05). Behaviour displayed in response to the electrical stimulus showed a tendency to decrease in severity across sessions for the AS group (P = 0.06) but did not change for the S group. Temperament had no effect on the number of interactions with the fence (P = 0.10), or on the responses to the electrical stimulus, (P = 0.2). Next, the sheep in the AS group were ranked according to the ratio of electrical stimuli to audio cues and split into the top and bottom six depending on ratios of S:A (low and high). They were tested for their ability to respond to an audio cue in a group paddock test. Sheep in all treatment groups were able to learn to respond to the audio cue with no influence of temperament on learning. Further studies to understand the how individual differences influence learning in groups are recommended.