Predicting tail-biting from behaviour of pigs prior to outbreaks

Predicting that an outbreak of tail-biting is going to occur in a group of pigs would be a useful tool for farmers. In a prospective longitudinal study, 24 groups of c.30 undocked pigs were followed from birth to slaughter weight. Four groups had to be excluded from the analysis, the remaining groups were classified as having No Outbreak (n = 6), Underlying Outbreak (n = 8) or Severe Outbreak (n = 6) of tail-biting. The hypotheses examined were that pigs would be more active, perform more tail-orientated behaviours, or have their tails tucked under their body more in groups that went on to have outbreaks than in those that did not. Direct observations were made at 7, 11, 15 and 19 weeks of age and video recordings were examined for the 4 days prior to an outbreak. All outbreaks occurred after the behavioural observation at 11 weeks of age. Activity levels were significantly higher in Severe Outbreak groups in the 4 days prior to an outbreak, with more pigs Standing (p < 0.05) and significantly fewer pigs Sitting (p < 0.05) or Lying Inactive (p < 0.05) than in matched control groups. Comparisons of the Severe, Underlying and No Outbreak groups at 7 and 11 weeks of age, prior to any outbreaks, showed no difference in activity levels. However, at these ages, levels of Tail Interest were higher in No Outbreak groups than in those with Severe Outbreaks (p < 0.05), while the opposite was true for damaging Tail-Biting (p < 0.05). Thus high levels of damaging Tail-Biting may be a good predictor of impending outbreaks, but high levels of Tail Interest are not. Tail position also differed between groups at this age, with fewer Tails Tucked Under in No Outbreak groups (p < 0.01). Some outbreaks were predicted by multiple variables, while others had no clear predictors. In five outbreaks a small runty pig was tail bitten in the absence of a full outbreak. Whenever this happened, a full tail-biting outbreak always went on to take place sometime afterwards. The occurrence of single tail-biting events may thus be reliable indicators of future outbreaks. In summary, measurement of pig activity has potential for predicting tail-biting outbreaks on commercial farms as do levels of tails tucked under and damaging tail contact. Further work is needed to understand the relationship between different forms of tail contact and tail-biting. We highlight the difficulty in predicting all outbreaks from a single measure and conclude that tail-biting outbreaks vary considerably. The presence of an 'indicator pig' might be a useful sign that an outbreak will occur.