FEEDING ORDER OF SOWS AT AN INDIVIDUAL ELECTRONIC FEED STATION IN A DYNAMIC GROUP-HOUSING SYSTEM

Patterns, in the use of a feed station, by sows, in a dynamic group-housing situation, were studied because deviations from such patterns might indicate disease, oestrus, reproductive or other problems. Groups of ten sows of various parities were transported to our experimental farm on the day their piglets were weaned. They had no prior experience with group-housing. After 1 day elsewhere for settling dominance relationships, the sows were group-housed in a ser-vice house. After 10, 17 or 31 days in the service house, the groups were introduced as subgroups into a gestation house and stayed there until a few weeks before the expected date of farrowing. In both the service house and the gestation house the animals were fed by an individual electronic feed station. The feeding order was defined as the order in which the sows obtained the first food from the feed station each day after the start of the feed cycle. The consistency of feeding order within and between subgroups, and the difference in rank number for feeding within and between subgroups on 2 consecutive days were studied during the period the sows stayed in the gestation house. Furthermore, the relation between the number of sows with uneaten food in the gestation house and the time the sows spent in the service house was studied. Although the feeding order was not random within subgroups, it did not remain stable over a period of several weeks. In general, sows from the subgroup most recently introduced into the gestation house fed later than sows introduced earlier. The difference in rank number for feeding within subgroups of ten sows on 2 consecutive days was three or more in 19.1% of the studied cases. Overall (28-30 sows), it was six or more in 17.3% of the studied cases. Allowing longer periods for learning to operate the feed station reduced the time the sows needed to become accustomed to the feed station in a new environment. No sufficiently stable patterns could be identified to make it feasible to monitor the sows using deviations from such patterns.