Social network analysis of tail-biting interactions among growing-finishing pigs with intact tails housed in two group sizes

Tail biting compromises the welfare and performance of pigs. Understanding the dynamics of this behavior is essential for developing management strategies to minimize outbreaks. We used social network analysis (SNA) to evaluate the effect of group size on tail-biting interactions among pigs and to examine differences in tail-biting network position between barrows and gilts and biters and non-biters. Pigs (n = 315, initial weight = 22.1 ± 3.76 kg) with intact tails of mixed sexes were housed in small (9 pigs/pen) or large (18 pigs/pen) pens on slatted floors for 14 wk until market weight (120.2 ± 11.47 kg). Floor and feeder space allowances were identical for pigs in both group sizes. Growth performance, tail injury, and behavior were recorded. Videos recorded the day before the first TBO were manually reviewed to document tail-biting events and identify pig roles. The top 25% of pigs responsible for the most tail-biting incidents were classified as tail-biters and the remaining pigs as non-biters. Tail-biting network metrics were calculated using RStudio. Data were analyzed using SAS. Pigs in small pens exhibited higher density, reciprocity, out- and in-degree centrality, betweenness centrality, and closeness centrality (all P < 0.05), suggesting greater direct and indirect involvement in tail-biting interactions than pigs in large pens. Gilts had higher out-degree centrality (P < 0.01), indicating they initiated more tail biting than barrows. Both sexes had similar in-degree centrality, suggesting they were bitten equally. In gilts, biters tended to have lighter final weight (P = 0.08) and lower ADG (P = 0.10) and were injured less frequently (P = 0.02) by tail biting than non-biters. In both sexes, biters and non-biters had similar in-degree centrality, suggesting that they were bitten equally. These results indicate that pigs housed in different group sizes exhibit different social structures and positions in tail-biting networks, which potentially affects tail damage. Further research should evaluate network metrics across a wider range of group sizes and during multiple TBO.