The relation of concentrations of endogenous estrogens and androgens to lipid and lipoprotein levels was examined in 176 white, postmenopausal women (mean age, 58 years) with an average of 9 years since the onset of menopause. All of the women were participants in a clinical trial of the effect of walking on postmenopausal bone loss. In that trial, women were randomized into either a walking group or a control group and were followed for 3 years. There were no differences in the serum hormones or lipids by randomized group, and hence, results from this study are presented for both groups combined. None of the women were on estrogen replacement therapy. Data were available from year 1 (1982-1983) of the trial for the estrogens, lipids, and lipoproteins. Information on androgens was available for 143 of these women. Hormone levels were determined by highly specific methods involving extraction, column chromatography, and radioimmunoassay. About 50% of the women had estradiol levels at or below the sensitivity level (2.5 pg/ml) of the assay; therefore, estradiol levels were viewed as dichotomous (measurable/not measurable), and the estradiol results should be interpreted with caution. There was little relation of the androgens to the lipid values. Univariate analyses suggested a direct relation between total cholesterol, low density lipoprotein cholesterol, and triglyceride levels with estradiol. An inverse relation was suggested between serum estrone and estradiol and total high density lipoprotein (HDL) cholesterol and HDL2 cholesterol, although none of these associations were statistically significant. Multiple regression analyses revealed that the primary determinant of the HDL cholesterol and triglyceride levels was the degree of obesity as estimated by the body mass index (weight (kg)/height (m)2). Addition of estrone or estradiol to the models did not contribute to the prediction of lipid levels. These results do not support the hypothesis of there being a relation between endogenous sex hormone levels and lipid levels in postmenopausal women. The results suggest that sex hormones cannot explain the sex difference in lipid levels and may not contribute to the rise in coronary heart disease that occurs in women around menopause.