Serum sex hormones may be related to the risk of several diseases in postmenopausal women. In the current report, the authors examined the epidemiology of serum sex hormones in 176 healthy, white postmenopausal women (mean age 58 years) recruited from the metropolitan Pittsburgh, Pennsylvania, area. The data were collected during 1982-1983; none of the women were on estrogen replacement therapy. Serum concentrations of estrone, estradiol, testosterone, and androstenedione were measured by a combination of extraction, column chromatography, and radioimmunoassay. Neither age nor time since menopause was a significant predictor of sex hormones. The degree of obesity was a major determinant of estrone and estradiol. The estrone levels of obese women were about 40% higher than the levels of nonobese women. There was a weak relation between obesity and the androgens. Cigarette smokers had significantly higher levels of androstenedione than nonsmokers, with little difference in serum estrogens between smokers and nonsmokers. Both estrone and estradiol levels tended to decline with increasing alcohol consumption. Physical activity was an independent predictor of serum estrone. More active women had lower levels of estrone. There was a positive relation of muscle strength with estrogen levels. The data suggest interesting relations between environmental and lifestyle factors and serum sex hormones. These environmental and lifestyle factors are potentially modifiable and, hence, if associations between sex hormones and disease exist, modification of these factors could affect disease risks.