Previous epidemiological studies have demonstrated that obesity increases endometrial cancer risk two- to 10-fold. To test the hypothesis that abdominal adiposity further increases this relative risk, we conducted a nested case-control study of endometrial cancer incidence in a cohort of 41,873 women ages 55-69 years. Women were recruited by mail and asked to have a friend measure circumferences of several body parts using a tape measure and written instructions. Two-year follow-up for cancer incidence was conducted using a state-wide cancer registry. Compared to random controls (n = 1,274), cases (n = 63) had higher age-adjusted mean values of waist-to-hip circumference ratio (P = 0.10) and trunk-to-limb circumference ratio (waist plus hip circumferences divided by arm plus calf circumferences, P = 0.008). Other anthropometric variables, including current body mass index and current weight, were also greater (P less than 0.001) in cases than controls. After accounting for the association with body mass index, neither the waist-to-hip ratio nor the trunk-to-limb ratio remained associated with endometrial cancer incidence (P greater than 0.40). A 5 kg/m2 increase in body mass index was associated with an adjusted relative risk of endometrial cancer of 1.80 [95% CI = 1.46, 2.22] when other significant risk factors, namely age, education level, extended use of exogenous estrogens, and age at menopause, were taken into account. We conclude that endometrial cancer risk is increased in relation to the amount but not the distribution of adiposity. This is in contrast with several other diseases in which, in addition to overall body mass, the distribution of adiposity is also important.