Food restriction as a cause of stereotypic behaviour in tethered gilts

The behaviour of Large White × Landrace gilts tethered in stalls was studied by scan sampling from 07.00 to 17.00 h on 3 days consecutively every 5 weeks. Results are presented on 30 gilts sampled during their first 5 weeks in the house, period 1, and on 36 gilts during the following 5-week period, period 2. Twenty individuals were common to both samples. Gilts were fed at 08.00 h with a pelleted complete diet, either on low food levels (1·25 to 2·00 kg) or on high food levels (2·2 to 4·0 kg). In both periods behaviour was strongly related to food allowance in individual gilts. Gilts on low food levels spent a greater proportion of their time in the standing position than those on high food levels (period 1, medians 0·31 and 0·12 respectively; period 2, medians 0·27 and 0·11 respectively), spent more time in repetitive behaviour (period 1, medians 0·22 and 0·04 respectively; period 2, medians 0·16 and 0·07 respectively) and spent a greater proportion of their standing time in repetitive behaviour (period 1, medians 0·46 and 0 respectively; period 2, medians 0·30 and 0 respectively). Similarly, all three measures of behaviour correlated negatively with food allowance overall, but there was some indication that food level had a threshold effect, with more activity and repetitive behaviour occurring on food levels below about 2 kg. These results suggest that if food allowances are low tether stalls are not a suitable system of housing for pregnant sows.