In boars, social relationships could influence pubertal development and feeding behaviour. The objectives of the present study were to determine the relationships between behaviour (agonistic, mounting and feeding behaviours), plasma sex steroids (oestradiol, testosterone) and fat androstenone. A total of 270 Pietrain x Large White boars, derived from four distinct genotypes were used. They were raised in groups of 11-12 pigs/pen. Animals were observed for about two times 10 h/day at the beginning or about 6 h at the end of fattening. Agonistic (fighting, hitting, biting, threatening, chasing) and sexual (mounting) behaviours were counted in early fattening shortly after a social mixing (for 177 of the 270 animals) and in late fattening after several weeks of social stability. Skin lesions were counted in the same periods and on carcasses. Data obtained from electronic feeders were used to determine the number and duration of meals, and the feed intake over 96-h periods in the middle and at the end of fattening. At the end of fattening, blood was sampled to measure oestradiol and testosterone. At slaughter, fat was collected to measure androstenone. Using the agonistic behaviours, a dominance index was calculated (DRrank). As expected, numbers of skin lesions and agonistic acts were higher in early fattening (P < 0.0001) whereas that of mounting acts was lower in early than in late fattening (P < 0.05). The feeding characteristics (r = 0.25 to 0.39) were significantly correlated between ages, whereas numbers of mountings (r = 0.18), of skin lesions (r = 0.05) and agonistic acts (r =-0.05) were not. The number of agonistic acts was significantly correlated with that of mountings only-in early fattening (r = 0.38). At both ages, no difference between dominance groups was observed for feeding characteristics. This was probably related to low competition between boars for access to feed. Dominant boars had a higher concentration of androstenone in fat than subordinates (P < 0.05) but no difference was observed for plasma steroids. An active role of androstenone in controlling social behaviour but not reproductive function may explain this phenomenon.