A study of agonistic strategies after mixing in group housed sows

Group-housing systems offer sows more physical space, and the opportunity for exploration and social interaction. However, sows in groups may have to adopt different agonistic strategies to deal with the social environment, and research has shown distinct differences between these agonistic strategies in associated injuries and stress in sows (Mendl et al. 1992; Verdon et al. 2013). The aim of the present study was to examine some behavioural strategies that sows may adopt to cope with aggression at mixing. It was hypothesised that submissive sows will have less access to feed, will perform less aggressive and more avoidance behaviours, and will have less fresh skin injuries compared with other sows.

This study was conducted at a commercial piggery in Victoria, Australia. Over three replicates (one replicate per week), 155 recently-weaned sows (parity1 to 8) were allocated to one of two mixing pens. Pen 1 housed an average of 27 sows per week (range 24–30 sows) at an average space allowance of 2.7 m²/sow (range 2.4–3.0 m²/sow), whereas Pen 2 housed an average of 24 sows per week (range 21–27 sows) at an average space allowance of 2.7 m²/sow (range 2.4–3.1 m²/sow).

During each replicate, behavioural observations were made on 10 focal animals randomly selected from each pen. All behavioural observations were made by a single observer using video records. Focal sows were observed for a total of 45 min in the first, third and fifth hours after mixing on d 1 and for 45 min after each feed drop at d 2 (0730 and 1300 h). Agonistic behaviours (threat, parallel pressing, head and body knocking, bites, fights, submission and displacements) were continuously observed for a total of 45 min for each observation period. All aggressive interactions delivered and received by the focal sows were recorded to calculate an ‘aggression index’ using the formula: aggression delivered/(aggression delivered + aggression received) (after Verdon et al. 2013). Sows were then classified into three aggression categories according to the calculated ratio [dominant (D), subdominant (SD) and submissive (S)]. In addition, 1-min point sampling was used to record time spent feeding (TF) and resting (TR) and the area of the pen where the sows were located. Areas of the pen that provided food and bedding were classified as preferred resting areas, and less preferred areas were defined as those with no food and bedding materials. The TF and TR were analysed as a proportion over the total observation time (225 min), and those sows culled for lameness after the experiment were not included in the statistical analysis for TR. Skin injuries were measured in focal sows using the method described by Karlen et al. (2007), at 1500 h on d 2. Data were appropriately transformed when the assumption of normality was not fulfilled. One-way ANOVA was used to compare differences in agonistic behaviour, space utilisation and skin injuries of D, SD and S sows. Multiple comparisons between means were performed using the least significant difference test (SAS^®; USA).

Significant differences existed between the aggressive interactions delivered (P < 0.001), the submissions performed (P = 0.015) and the aggression received (P = 0.016) between the three aggression categories. The D and SD individuals delivered significantly more aggression than S sows, the D sows performed less submission, and the SD sows received more aggression than the other two categories. The SD sows also presented more skin injuries (old and fresh) (P = 0.009) compared with the D sows. Significant differences also existed in relation to space utilisation: S animals spent less time feeding (P = 0.031) and more time resting in less preferred areas of the pen (P = 0.021) than D sows.

In conclusion, each aggression category appeared to have costs and benefits with regard to aggression received and delivered, injury, and resource access. The D and SD sows delivered similar levels of aggression, but SD were more persistent in displaying aggressive behaviour regardless of defeat, and thus had higher numbers of skin injuries (old and fresh) compared with D sows. In addition, S sows may have experienced difficulties in gaining access to resources such as feed and preferred lying areas of the pen. Further research on features of a mixing pen, such as provision of a barrier and increased floor space, is required to examine opportunities to minimise aggression and safeguard the vulnerable individuals.