The reliability and concurrent validity of a simple questionnaire to assess leisure time physical activity has been investigated on 306 self-selected healthy adults of both sexes (163 M; 143 F). Values of body fat (BF) and maximum oxygen intake (VO2 max) expressed as percentiles of appropriate age and sex categories were used as criteria of validity for the questionnaire. BF and VO2 max were predicted from the Durnin and Womersley skinfold equations, and the laboratory version of the Canadian Home Fitness Test respectively. The strongest correlation was between VO2 max (percentile) and reported strenuous exercise (r = 0.35). The optimum discriminant function for VO2 max was based on a combination of reported strenuous and light activity. This yielded a correct 2-way classification of 69% of the subjects. A combination of sweat-inducing and moderate exercise yielded a correct 2-way classification of BF for 66% of subjects. The reliability coefficients for the optimum discriminant functions classifying VO2 max and BF were 0.83 and 0.85 respectively. We conclude that this simple instrument has potential value for the assessment of leisure time exercise behavior, offering the possibility of examining changes in behavior following the implementation of health and physical fitness promotion programmes in the community.