Although commercial farrowing sheds keep individual litters separated, previous studies have suggested that housing systems that allow socialisation of piglets pre-weaning can reduce 15 aggression after weaning. This study tested whether pigs socialised with non-litter mates preweaning would show less aggression during mixing at weaning (when piglets are taken from their sows and mixed in group housing), and whether socialisation influenced the time budgets or behavioural expression of piglets at weaning. In total, 353 piglets were followed from birth through to one week after weaning. Piglets from 24 sows were allowed to socialise in groups of four litters 20 ('socialised' treatment group) from 10 d of age; litters from nine sows were followed as controls.Socialised piglets were monitored to determine the prevalence of cross-suckling. Body weight was recorded at birth, prior to weaning and one week after weaning. Continuous video footage was collected for 1.5 days after weaning for behavioural analyses. There was no difference in the body weight of socialised pigs compared to control pigs at weaning or one week after weaning. 25Quantitative scoring of behaviour revealed no significant difference in aggression displayed between treatment groups or between the sexes; however, compared with overall averages, a greater proportion of socialised males spent time lying (57% of time compared with an average of 43% for the other sex-treatment groups, P < 0.001; but less eating/drinking 4% cf. average 8%, P < 0.001), and a greater proportion of socialised females were investigating (17% cf. average 12%, P < 30 0.001 with less lying 40% cf. 48%, P < 0.001). Qualitative Behavioural Assessment (QBA) was used to assess the body language of pigs during an active period (the middle of the day after weaning).Observers reached consensus in regard to their assessments of pig behavioural expression (P < 0.001). Two main dimensions of behavioural expression were identified, which accounted for 41% and 19% of the correlation between pigs. There were significant socialisation treatment effect (P = 35 0.002 and P = 0.007) on both dimensions, with socialised pigs more likely to be described as 'sleepy'/'tired' or 'content'/'relaxed' than control pigs (described as more 'active'/'curious' or 'aggressive'/'dominant'). Because socialising piglets had no effect on body weight pre-weaning, and there was a low occurrence of cross-suckling (2.9 ± 6.5 % of piglets recorded suckling), socialisation was not disadvantageous. On the contrary, the behavioural difference at weaning suggests 40 socialising piglets may be beneficial from a welfare perspective.