Group-housed castrated (but not intact) males show high incidences of attack on lactating (but not non-lactating) female intruders. In Experiment 1, responses by isolated and group (in 3's)-housed intact and castrated males to lactating intruders were studied. The latter males were given sham, testosterone (T) or estradiol (E2) implants before testing. Castration augmented attack by animals in both housing conditions, confirming that intra-group fighting could not account for androgen-related changes in fighting behaviour in group-housed animals. Both T and E2 implants suppressed attack in animals in both housing conditions. Consequently, Experiment 2 evaluated the influence of three dose levels of T, E2 and DHT (dihydrotestosterone), in attack by castrated group-housed male residents on lactating intruders. Treatment of castrated males with T or E2 (but not DHT) inhibited attack in these animals. These findings provide support for the idea that aromatization is also a prerequisite for the inhibitory action of T on male aggression towards lactating females.