Objective: To evaluate the relation between body fatness (%Fat) and body mass index (BMI) and to evaluate the validity of the BMI standards for obesity established by the NIH in older black and white postmenopausal women. Research methods: Height, weight, BMI, and %Fat, assessed by DXA, were determined for 296 healthy, independently living women ranging in age from 50 to 80 years (M7s.d.; 64.477.8 years). Results: Per NIH guidelines, 32% were classified as obese (X30 kg/m 2 , mean BMI ¼ 28.175.5 kg/m 2 ). In contrast, using the %Fat criterion of 38% advocated by Lohman to define obesity, 47% of our sample was obese (mean %Fat ¼ 37.376.2%). A moderately high curvilinear relation existed between BMI and %Fat (R ¼ 0.82, SEE ¼ 3.57 %Fat, Po0.05). Race added meaningfully to the prediction of %Fat (Po0.05) such that for the same BMI, black women will have 1% lower body fatness than white women. Based on a %Fat X38 as the criterion for obesity, receiver operating characteristic (ROC) analysis, performed separately by race, indicated that the currently accepted BMI cutpoint for obesity produced low sensitivity (69% and 61% for black and white women, respectively). Alternatively, BMI valuesX28.4 kg/m 2 for black women and X26.9 kg/m 2 for white women to define obesity maximized classification accuracy. Conclusion: We conclude that current BMI categories may not be appropriate for identifying obesity among postmenopausal women. Furthermore, the relation between BMI and %Fat is different in black compared to white women but remains constant from the sixth through the eighth decade of life.