Confinement of breeding sows to stalls is a controversial welfare issue, and there is a worldwide move to house gestating sows in groups. We examined the effects of day of mixing following insemination on aggression, injuries and stress in sows. A total of 800 sows were used in this experiment and we examined the effects of mixing sows in groups within 1 to 7 days post-insemination (Group0) or at 35 days (36-42 days) post-insemination after housing in stalls from insemination (Group35). Groups of 85 sows were housed on concrete floors covered in rice hulls with a floor space allowance of 2.3m2 per sow. Mixing sows at day 35 post-insemination, instead of early post-insemination, resulted in a reduction (P<0.05) in the frequency of aggressive behaviour (1.4-0.8 bouts per sow) and cortisol concentrations at mixing (5.2-2.5nM) and the number of old injuries (13.2-4.6 injuries per sow) at 7 days after mixing (7 and 42 days post-insemination, respectively). By 7 days after mixing, the frequency of aggressive behaviours and cortisol concentrations were substantially lower than at mixing and there were no differences between the treatments. There were no treatment effects on the ratio of neutrophils to lymphocytes at day 51 post-insemination, cortisol concentrations at day 91 post-insemination, or reproductive performance (farrowing rate and litter size). The results show that the stage of gestation when sows are mixed affects aggression, injuries and stress in sows and that challenges associated with aggression, injuries and stress at mixing are greater early after insemination than later. With the increasing interest in further reducing the period during which sows are housed in stalls, research examining ways of reducing risks to sow welfare and productivity associated with grouping needs to focus on mixing sows both after their piglets are weaned and early after insemination.