Obese subjects have increased bone density relative to non-obese subjects, yet this relationship is not fully understood. We examined whether alterations in glucose or insulin concentrations might explain the effect of obesity on bone density in 317 women from the San Antonio Heart Study, a population-based study of diabetes. We measured fasting and 2-hour plasma glucose and fasting and 2-hour serum insulin levels. Bone density was assessed by a Hologic dual-photon absorptiometer. Lumbar spine and femoral neck density were positively correlated with body mass index (BMI). Femoral neck density also was positively correlated with fasting insulin level in younger women after adjustment for age (r = .214, P < .01). After further adjustment for BMI, femoral neck density was not significantly correlated with fasting insulin level (P = .08). The magnitude of the estimated femoral neck density difference for a 32.2-microU/mL decrease in fasting insulin level (the difference in insulin concentrations between nondiabetics and diabetics) was 0.13 g/cm2. Adjustment for glucose and insulin concentrations does not explain the association between bone density and obesity.