Quantitative autoradiographic analysis was used to identify regions in the brain of the male primate where androgen binding sites may be involved in the actions of testosterone. Three days after castration, adult male rhesus monkeys received a subcutaneous injection of either dihydrotestosterone propionate (DHTP, 20 mg, n = 6), testosterone propionate (TP, 100 mg, n = 2), or oil vehicle (control males, n = 4). Three hours later, 5 mCi [3H]testosterone was administered as an i.v. bolus. At 60 min, brains were rapidly removed and the left halves were used for autoradiography. In control males, highest percentages of labeled neurons (20-84% using a rigorous Poisson criterion) were observed in the ventromedial, arcuate and premammillary nuclei (n.) of the hypothalamus, medial preoptic n., bed n. of stria terminalis, intercalated mammillary n., lateral septal n. and the medial, cortical and accessory basal n. of the amygdala. Pretreatment with DHTP eliminated labeling in androgen target tissues of the genital tract, and reduced the percentages of labeled neurons to 4-22% of control values in the arcuate, lateral septal, premammillary and intercalated mammillary n., indicating that in these regions testosterone acted predominantly at androgen binding sites. However, in the medial preoptic n., the ventromedial hypothalamic n. and the accessory basal amygdaloid n., DHTP pretreatment resulted in much less blocking which, together with other data, suggested that in these sites, testosterone's actions involved aromatization and interaction with estrogen-binding sites.