Under wild conditions, the ancestors of the domestic pig cohabit in small, genetically related matriarchal groups. Outside of the mating season, aggression between individuals is infrequent and rarely injurious; a feature common to domestic pigs housed in an extensive enclosure (Stolba and Wood-Gush, 1984, 1989; Mendl, 1995). Group social stability is facilitated by the infrequent and gradual integration of new members to the group (Mauget, 1981), close kinship, the preservation of individual space and the use of threats and non-aggressive behaviour to maintain dominance relationships (Mendl, 1995). In contrast, many commercial pigs are repeatedly exposed to sudden mixing episodes with unrelated animals in an environment that limits effective dispersal and the display of appropriate submissive behaviour. Under these conditions, post-mixing aggression is intense during the first 24 h after mixing (Meese and Ewbank, 1973), but varies considerably between members of the group (Mount and Seabrook, 1993; Erhard et al., 1997). The number of skin lesions (lesion score, LS) has frequently been used as a proxy indicator of the extent of post-mixing aggressive behaviour. This approach offers a rapid means of assessing aggressiveness of a large number of pigs and has been used commonly when investigating the development of aggression over several days or weeks or when using large social group sizes (e.g. Francis et al., 1996; Erhard et al., 1997; Spoolder et al., 1999; Turner et al., 1999, 2000, 2002). The objective of this work was to investigate the factors that contribute to LS in order to assess the validity of LS as a measure of individual aggressiveness. After mixing into new groups of 12, the durations of reciprocal fighting and bullying, the proportions of fights initiated, won and lost and physical attributes of the pigs were investigated for their impact on LS using multiple regression on a sample of 342 growing pigs. Pig live weight was the single greatest determinant of LS. The duration spent in reciprocal fighting and being bullied were also significant determinants of the LS and contributed to the total LS in proportions of 0.17 and 0.14, respectively. The proportion of fights initiated, won and lost did not influence the LS for pigs of a given weight. On an individual pig basis, it was impossible to establish the relative importance of reciprocal fighting and the receipt of bullying to the accumulation of lesions by reference to the total LS alone. Engagement in reciprocal fighting was found to result in lesions to the anterior third of the body, whilst the receipt of bullying resulted in lesions accruing on the caudal third of the body. Reference to the location of lesions, in addition to their number, is a refinement of the methodology that potentially allows discrimination of pigs that accumulate lesions as a result of reciprocal fighting or receipt of bulling. A partial replication of the experiment at a second unit (n = 84 pigs) with different genotypic, environmental and husbandry conditions confirmed that the duration spent in reciprocal fighting and being bullied were significant determinants of the LS and that the LS approach is applicable across units. These results suggest that the LS methodology provides a rapid means of estimating aggressive behavioural phenotypes when reference is made to both the number and location of lesions and pig weight and pen identity effects are accounted for.