Most published studies have failed to show a greater food intake in obese subjects than in nonobese. However, the sample sizes in most of these studies are small and the methodologies open to question. HANES I is based on a probability sample of 20,749 people, representative of the civilian noninstitutional population of the U.S. We analyzed a subsample of 6219 nonpregnant adults whose diet was not influenced by illness or drugs and who stated that their intake, estimated by dietary interview, represented their usual pattern. Neither the caloric intake nor the caloric intake adjusted for physical activity level and age was higher in the obese subjects. This suggests that, unless estimates of food intake differ in accuracy between obese and nonobese subjects, factors other than overeating should be given increased consideration in the etiology of obesity.