Five male and four female normal weight research volunteers, participating in 13-day residential studies, received oral fenfluramine (20, 40 mg) or placebo at 09:30 and 17:00. Food intake, performance, and subjective ratings were measured throughout the day. Carbohydrate intake was manipulated by providing lunch meals high (males: 120 g; females: 80 g) or low (males: 25 g; females: 16 g) in carbohydrate on 8 days; on the remaining days subjects self-selected lunch. Total caloric intake (approximately 2800 Kcal) did not differ among the low- and high-carbohydrate, and self-selected lunch conditions when subjects received placebo, indicating caloric compensation. Total carbohydrate intake was significantly less, however, when subjects consumed the low-carbohydrate lunch compared to the other lunch conditions. Fenfluramine significantly decreased total caloric intake (approximately 500 kcal) by decreasing meal size, not number, only when subjects consumed the low-carbohydrate lunch. Fenfluramine was only an effective anorectic drug when subjects consumed a lunch with fewer calories and a lower carbohydrate:protein ratio than self-selected baseline. Also, fenfluramine improved performance on a range of computer tasks and increased ratings of "Alert," "Friendly," and "Talkative," while decreasing ratings of "Tired" and "Irritable."