A novel method for the analysis of social structure allows in-depth analysis of sow rank in newly grouped sows

Chronic stress response in fearful animals can result in depression of growth and reproductive performance. It is therefore important to be aware of at risk animals in the herd. Thus far 'hierarchy' calculations have involved the use of fights won and lost on the day of mixing or successful displacements over a trial period, or a combination, but not analysis of the two separately and then combined. This experiment used 132, multiparous, Large White x Landrace sows. Following artificial insemination, sows were mixed into groups of 6, for 5 days. Salivary cortisol and behaviour were measured on the day before mixing (d-1) and after mixing d0, d1, d3 and d4. 'Hierarchy' was assessed using the number of successful displacements for each sow over all 4 days and the number of fights won and lost on the day of mixing. For both parameters, the sows were separated into three groups, 1D or 1F sows were involved in no fights or displacements, 2D and 2F sows lost more than they won and 3D and 3F sows won more than they lost. Sows ranked 1D1F received significantly more aggression [P < 0.05, eg. bites received; 1D1F = 0.7 ± 0.1 (5.8), average of other groups = 0.4 ± 0.1 (1.9)] suggesting that these sows have the lowest welfare out of the population of sows. 1D2F sows did not record similar findings suggesting that using this method to calculate rank is of use, as the subgroups were different. Sows ranked low by this calculation are at risk of facing more aggression and therefore stress.