Three experiments were used to tests the applicability of the aromatization hypothesis of androgen action to aggressive behaviour among Norway rats. In Expt 1, administration of testosterone propionate was highly effective in restoring aggressive behaviour to castrated rats while 17 beta-hydroxy-5 alpha-androstan-3-one was of intermediate effectiveness. Of the steriods tested in Expt 2, androstenedione and testosterone were highly effective, 17 beta, 19-dihydroxyandrost-4-en-3-one was of intermediate effectiveness and cholesterol was ineffective. The results of Expt 3 indicated that treatment with testosterone or oestradiol both resulted in increased aggression while treatment with (5 alpha,17 beta)-17,19-bis(acetyloxy)-andostan-3-one diacetate (5 alpha-19-hydroxytestosterone) was without effect. Androgens which were aromatizable and could be 5 alpha reduced, i.e. testosterone, testosterone propionate and androstenedione, were highly effective in restoring aggressive behaviour; however, two other steroids, 5 alpha,19-hydroxytestosterone which is 5 alpha reduced, and 19-hydroxytestosterone, which can be aromatized, were respectively of low or medium effectiveness on behaviour. However, oestradiol, which did not maintain sexual development of accessory glands, was highly effective in the restoration of aggressive behaviour. Since the behaviourally active steroids in the present experiments were not only those predicted by the aromatization hypothesis, it is proposed that several steroids are capable of activating aggressive behaviour and that the aromatization hypothesis does not adequately explain the hormonal basis of aggressive behaviour among Norway rats.