Understanding the effects of exercise on food intake is complicated by limitations of methodology. One primary limitation with human subjects is food intake assessment. Highly accurate techniques may not represent usual eating patterns. Present techniques representative of usual eating habits lack accuracy. Several independent variables also affect the results of exercise-food intake studies, even with laboratory rats where food intake measurements can be quite accurate. These variables include exercise protocols, subject characteristics, types and amounts of food available for consumption, environmental conditions, and sociocultural factors. Most investigations have focused on the effects of exercise training protocols on energy intake. When forced to run on treadmills, energy intake of laboratory rats usually decreases in males and increases in females. In response to forced swimming protocols, energy intake in male rats is usually unchanged and increased by female rats. In contrast, voluntary running (in running wheels) usually results in an increase in energy intake in both male and female rats. However, energy intake may also be unaffected or reduced in male rats given access to running wheels. These gender differences may be related to much lower levels of voluntary wheel running observed in male rats. The gender difference observed with rats is not apparent in humans. Energy intake of humans is usually increased or unchanged in response to exercise training programmes. However, when energy intake is increased, it is usually below the increased expenditure, resulting in negative energy balance. Highly trained athletes and lean individuals usually increase energy intake in response to increased physical activity, whereas, obese untrained subjects commonly do not change energy intake when exposed to exercise training. Few studies have investigated the effects of exercise on changes in food selection in laboratory animals; the results of studies with humans also have been inconsistent. Research results range from increased carbohydrate, fat, or protein consumption to no change in diet composition. Rats respond to detraining (cessation of exercise) with increased or unchanged energy intake. Men who detrain after 9 months of training regain bodyweight and body fat lost during training despite a reduction of energy intake and the percentage of calories from dietary fat. The relationship between exercise and food intake is complex. These often inconsistent or conflicting results reflect this complexity. Further understanding awaits additional research to clarify confounding variables.