Mixing of unfamiliar animals with a change of physical environment is a common practice in pig husbandry. The pigs have to cope with novelty in their life environment, whereas it is well known that pigs exhibit novel-induced anxiety when submitted to a novel area (Janczak et al., 2003). At the same time, pigs have to cope with social challenges, which may cause injuries and physiological reactions of acute stress. Fighting between newly mixed pigs over the first 24 h is part of the process necessary to establish a dominance order (Meese and Ewbank, 1972). As for other domestic mammals, once established, this social hierarchy regulates aggressive interactions, improves the predictability of social relationships between animals and thus reduces stressful social encounters in stable groups (Tennessen, 1989). However, the abrupt breaking of a stable social group leads to the re-establishment of a new hierarchy (Ewbank and Meese, 1971; Tan and Shackleton, 1990; Puppe, 1998) that may be detrimental for animal welfare. The aim of the present study was to evaluate the effects of repeated mixing and relocation in growing pigs using production, behavioural and physiological indices of stress. Thirty-two growing-finishing pigs were housed in pairs. Animals of half of the pairs were placed in a new pen with a new partner three times a week during 4 weeks (R pigs, n = 8 pairs). Other animals stayed in the same pen with the same partner (C pigs, n = 8 pairs). During the first 90 min following each regrouping, vigorous fights were frequent in R pigs (same pen same partner). Those pigs exhibited more agonistic behaviour than C pigs (new pen with new partner) throughout the experiment (P < 0.05). The number of offensive acts realized by R pigs decreased throughout the experiment (P < 0.05) whereas the number of defensive acts did not vary (P > 0.1). The time-budget determined during 24 h 2 days after the third, sixth, ninth and 12th regrouping was similar in both groups of pigs except the percentage of standing that was higher in R pigs after the last regrouping. R pigs tended to have a lower daily liveweight gain at the fifth regrouping (P = 0.07). Salivary cortisol concentrations were higher in R than in C pigs 4–5 h after the second and fifth regrouping (P < 0.05) but were similar in both groups after the ninth and 11th regrouping. Basal cortisol measured 2 days after the last regrouping in plasma and saliva was higher in R than C pigs. In response to a dexamethasone/ACTH challenge, cortisol in saliva and plasma was similar in both groups (P > 0.05). The behavioural and cortisol responses to regrouping and relocation in pigs decreased over time but were still obvious after 12 mixings. Repeated regroupings induce a chronic stress in pigs without altering severely growth rate.