The preoptic area (POA) of the male ring dove is a target for specific behavioral effects of estrogen that are separable from those of androgen. Activity of the aromatase system in the POA, which converts testosterone to 17β-estradiol (E2 1.0 mm from POA nuclei were ineffective in increasing preoptic aromatase activity irrespective of whether the implants were unilateral, or bilateral with twice the potential output of testosterone. Within the 1.0-mm range, distance of implant from the POA nuclei was negatively correlated with induced aromatase activity in POA samples, indicating a direct effect of testosterone or its metabolites on enzyme activity. Induction of aromatase activity was higher in the right side of the POA than the left, suggesting asymmetry in inducible aromatase. Inactive 5β-reduced androstanes, 5α-dihydrotestosterone and 17β-estradiol were formed from intracranial 3H-testosterone in POA. Since estradiol benzoate implants did not induce aromatase activity, this metabolite does not appear to act directly on the POA, although it is effective if administered systemically. Implants of TP in the region of the POA caused vocal behavior (perch calling) to be shown by some males. There was no correlation between the behavioral effectiveness of implants and induced POA aromatase activity. Since increase in E2 formation occurred in the absence of vocal behavior, activation of androgen-dependent behavior is not an absolute requirement for the induction effect. We conclude that testosterone can influence aromatase activity required for local production of E2 in the brain by direct action on POA ce