Twelve obese and four lean subjects increased their energy expenditure by 100 kcal/day by cycling on a bicycle ergometer. The order of testing was alternated so that half of the subjects did exercise on the first three days and the remainder on the last three days. Ad lib energy intake was measured over six days using an automated food dispensing machine. The obese subjects ate 1196 +/- 517 kcal/day and the lean subjects ate 1162 +/- 301 kcal/day and showed no overall significant difference in energy intake or eating patterns. However, these obese subjects ate 18 kcal/day less and the lean subjects ate 155 kcal/day more during the exercise period. The difference in response confirms the hypothesis that lean subjects tend to regulate energy intake more accurately than obese subjects. Exercise increased the frequency of eating and drinking and decreased hunger and appetite for all subjects. There was a significantly greater decrease in hunger and appetite with exercise for the obese subjects. However, as there was no accompanying decrease in energy intake and some subjects had difficulty in completing the exercise, the value of small increases in voluntary exercise in the treatment of obesity is questioned.