This study examines the cross-sectional association of hysterectomy and oophorectomy status, chronological age, and years since menopause with plasma levels of total and bioavailable testosterone and estradiol, androstenedione, estrone, and sex hormone-binding globulin (SHBG) in community-dwelling postmenopausal women who were not using estrogen replacement therapy. Six hundred and eighty-four women, aged 50-89 yr, were surveyed for hysterectomy and oophorectomy status and had plasma obtained between 1984-1987. Of these, 438 (67%) had not undergone hysterectomy or oophorectomy (intact), 123 (18%) reported hysterectomy with bilateral oophorectomy, and 123 (18%) reported hysterectomy with conservation of 1 or both ovaries. After adjustment for age and body mass index, both total and bioavailable testosterone levels were reduced by more than 40% (P < 0.001) in hysterectomized women with bilateral oophorectomy compared to those in intact women, with intermediate levels observed in hysterectomized women with ovarian conservation. Androstenedione levels were about 10% lower in hysterectomized women with or without ovarian conservation compared to those in intact women (P = 0.039). Total estradiol levels tended to be lower (P = 0.095) in bilaterally oophorectomized women. Levels of bioavailable estradiol, estrone, and SHBG did not differ by hysterectomy and oophorectomy status. Among intact women, total, but not bioavailable, testosterone levels increased with age (P = 0.015), reaching premenopausal levels for the 70-79 decade with relatively stable levels thereafter. Among oophorectomized women, total and bioavailable testosterone levels did not vary with age and were 40-50% lower than those in intact women throughout the 50-89 yr age range. Androstenedione levels decreased 27% and SHBG levels increased 30% (P < 0.001) with age in intact, but not oophorectomized, women. Levels of other hormones did not vary with age. Stratification by years since menopause or surgery yielded similar results. These results demonstrate that the postmenopausal ovary remains a critical source of androgen throughout the lifespan of older women. The clinical consequences of lower testosterone levels years after oophorectomy are unknown. Reconsideration of prophylactic oophorectomy and clinical trials to evaluate the effects of androgen replacement after oophorectomy are needed.