Food choices and diet composition have been studied less often than energy intake in subjects with varying levels of physical activity. The reported effects of exercise on food choices are not fully consistent, especially on the short term. Type of exercise, intensity, duration can affect the results as well as subjects' characteristics (gender, age, previous training and fitness). A crucial role could also be played by psychological (chronic dieting, attitudes toward health and food, long-established food habits and preferences) and social (traditions, food availability, appropriate times and places) factors. In short-term intervention studies, where a meal is ingested a few minutes following a bout of exercise of varying duration and intensity, an increase in CHO intake is most often reported, while increased protein intake is an occasional observation. In long-term (several weeks) training interventions, intake is assessed from dietary records. Again CHO intake is augmented in exercised subjects as compared to controls, while that of saturated fats and cholesterol may also be affected. Epidemiological studies (without dietary or exercise intervention) often report that habitually active persons eat more and ingest more fruits and vegetables than less active peers. It is not known to what extent such food choices are driven by biological needs (e.g. replacement of glycogen) or elicited by social and psychological factors.