We assessed the cross-sectional relationship of age, menopausal years, body mass, previous estrogen use, and ethnic background to bone mineral status in a sample of 875 healthy postmenopausal women at the time they were recruited from the community to participate in a multicenter clinical trial. The women were 1-10 years postmenopause, 45-64 years of age, and had not received estrogen replacement therapy within 3 months of enrollment. Of the participants, 89% were white, 69% had a spontaneous menopause, and 53% had a history of previous estrogen replacement therapy. Bone mineral density (BMD) of the lumbar spine (L2-4) and proximal femur was measured by dual-energy x-ray absorptiometry. Results were consistent with a significant negative linear regression of BMD on age or years from menopause. Body mass index (BMI) correlated significantly with BMD at all sites (L 2-4 r = 0.28; femoral neck r = 0.34, p < 0.0001). BMD adjusted for age and BMI were higher at both sites in women who had taken estrogen versus those who had not (L2-4 0.976 +/- 0.009 versus 0.932 +/- 0.01; femoral neck 0.740 +/- 0.006 versus 0.708 +/- 0.008, p < 0.05). Adjusted BMD also increased with duration of ERT. Parity was negatively associated with L2-4 BMD (p = 0.03) but did not correlate significantly with BMD at the femoral neck. Black women had the highest L2-4 BMD, and Hispanic women had the highest femoral neck BMD, even when results were adjusted for age and BMI. When data were corrected for differences in bone size, these interethnic differences were no longer significant. We conclude that increased body mass is positively correlated with BMD, and this may confer a degree of skeletal protection to heavier postmenopausal women. Exposure for 5 years to exogenous estrogen is associated with significantly increased age- and BMI-adjusted BMD.