Association patterns across multiple gestation cycles within a dynamic sow pen

Abstract

The transition of housing gestating sows from individual stalls to group pens allows for a greater freedom of movement and social interactions, but it comes with many challenges to ensure group cohesion and safety of all animals. Previous research on the behavior and welfare of group housed sows has predominately focused on aggression at mixing, feeding systems, and flooring type, with little information on association patterns within the pen. In order to examine the complex direct and indirect relationships among group housed sows, this study recorded dyadic associations (defined as resting in proximity, <1.5 m) among a subset of multiparous sows (n = 129) housed in the same dynamic group pen. Stochastic simulations were used to compare the association patterns observed to random permutations of association patterns given the pen logistics. For this study pen, 13.4 % of the dyads had significantly higher associations than we would expect under the null model of completely random lying patterns, while 43 % of dyads were never recorded to rest in proximity. Distinct subpopulations demonstrating clear preferences for different subareas of the gestation pen were also found. Younger sows demonstrated association preferences above and beyond what would be expected if there were only a mutual preference for a general area of the pen. The older sows, on the other hand, seldom formed strong associations, and their distribution within the pen appears to be driven by preferences for specific lying areas rather than for resting partners. Beyond this age dynamic, our results also indicated that sows that rested most often on the slatted floors, tended to be younger, smaller, had more severe lesions and had higher feed rank scores (i.e., ate last in the day) – all indicators of lower social rank – relative to the sows which were consistently recorded in the sow cubicles (i.e., concrete lying areas). While the results of this observational study have shed some light on the factors which influence positive-to-neutral social interactions amongst sows in group housed pens, we hope that they will ultimately serve as a jumping-off point for future experimental work capable of producing more comprehensive and conclusive insights into these important social dynamics.