Nine lean and nine obese subjects participated in two laboratory studies comparing the effects of a traditional high-fat dish (2.42 MJ and 49% as fat) and its convert fat-reduced version (1.53 MJ and 23% as fat) on subsequent food intake. Each version was consumed at the beginning of a free-choice lunch (experiment 1) or alone at lunch time (experiment 2). All subjects experienced both experiments. Three and 7 h after lunch, subjects were free to consume any of the items available in a vending machine. Food intake was measured directly. In experiment 1, by 8 h after lunch, mean cumulative energy intake (test dish included) was not different between the test dish conditions in both lean and obese subjects, and ad libitum energy intake (i.e. without test dish) was significantly higher under the low-fat than under the high-fat test dish condition. Thus, lean compensated 86% and obese 70% of the initial energy reduction. The cumulative fat intake (test dish included) was lower under the low-fat than under the high-fat dish condition for both weight groups. Nevertheless, the percentage of ad libitum energy intake derived from fat was significantly higher after the low-fat test dish in obese but not in lean subjects. In experiment 2, both weight groups maintained the energy and fat reduction. Thus, obese subjects did not respond differently to lean subjects to the fat and energy manipulation of the test dish, except for the percentage of fat in their diet. These results suggest that different ways of including fat-reduced foods into the habitual diet might have different effects on subsequent energy compensation.