Learning performance of gestating sows called to the feeder

A call feeding station in which sows learn to be allowed to enter a feeding station only after being called by an individual acoustic signal has been shown to reduce agonistic interactions in front of the feeding station. Here, we tested important prerequisites for integration of a call feeding station in common husbandry practice of group-housed sows. In particular we tested whether the learning task was accepted by the animals in a reasonable time, whether sows which already knew a feeding station would perform different from sows naive towards any feeding station, whether sows would remember their individual acoustic call after returning from farrowing and service period and whether feeding frequency in the gestation pen would affect learning performance. A total of 62 sows involved in this experiment. 22 of those sows did not have any practice with a feeding station before the experiments (NLS = new learning sows) and 40 sows were already accustomed to a conventional feeding station and had to adapt to the new task of the call feeding station (RLS = relearning sows). Subgroups of sows were trained in a learning pen for 19 days and then integrated into a larger group of sows in the gestation pen where they were fed either once or twice a day by a call feeding station. During training the proportion of correct responses in both groups increased (RLS: rs = 0.6247; p = 0.0169; NLS: rs = 0.6169; p = 0.0188). NLS sows showed a higher learning performance compared to RLS sows (median: 62.75% vs. 47.5% correct responses). In the gestation pen learning success of sows which had learned the call feeding task for the first time still increased throughout the gestating period of 80 days (rs = 0.5426; p < 0.0001) but showed a steady course of correct responses of 100% in sows which returned into the gestation pen for a second time or even more often. After six weeks of farrowing and service period in which sows did not hear their individual call both NLS and RLS sows remembered it to 100% in median. Across the whole stay in the gestation pen, calling the animals to feed once a day resulted in a slightly better learning performance than twice a day (N1 = 80, N2 = 80; Z = 3.6297, p < 0.0003). Our results suggest that with respect to the learning ability of sows the prerequisites for implementation a call feeding station in the normal production process are given.