Abstract: The Minnesota Zoological Garden established a bachelor (i.e. all-male) group of 8 Przewalski horses (Equus przewalskii) in 1984. Possible methods for mixing males emerged through preliminary introductions and review of the relevant literature. These included: (1) grouping colts (<=3 years old) after removing them from their natal bands; (2) introducing adult stallions to colts in spacious, controlled environments to reduce the risk of combat-related injury or death associated with largely unpredictable adult temperament; (3) introducing more than 1 animal at a time to established groups. Based on the frequency and orientation of agonistic interactions observed among the bachelors, a linear dominance hierarchy was discerned among the 4 adult stallions; all 4 colts were determined to be of equally low or seemingly non-ranking status. Social rank among the adults was most influenced by individual temperament, and contention for rank was limited almost exclusively to the 3 lower-ranking stallions; the alpha male participated in 73% of all agonistic interactions observed, but was identified as the recipient in less than 3% of these interactions. Intra-band associations were detected in the bachelor group. The most prevalent of these was among the 4 colts, found together in over 70% of the observations. Other affinities commonly observed were between the colts (as a sub-group) and different stallions. The dominant stallion, the colts' most frequent companion, actively discouraged the 3 lower-ranking stallions from associating with the colts. Further, the colts were subjected to significantly fewer agonistic behaviors than the non-alpha stallions (per colt=1.05+/-0.24 h-1; per non-alpha stallion=2.01+/-0.04 h-1). These findings support the hypothesis developed by other workers for ungulates that young bachelors are sheltered from aggression by association with an alpha male. With one exception, adult associations were rarely noted. Although combat-related injuries (superficial wounds and temporary lameness) were common among the bachelors, few required medical attention. No permanent disabilities or deaths have occurred since the group's inception. The existence of social organization and intra-band associations in the bachelor group support the suggestion that such groups were once a natural part of this species' social milieu. Bachelor groups are therefore recommended as an ethologically sound, space- and cost-effective means of managing “surplus” (i.e. non-harem) male Przewalski horses.