This study compared circadian rhythms in physiological, subjective, and performance measures between groups exhibiting different levels of habitual physical activity. Fourteen male subjects, aged 19–29 years, were assigned to a physically active (group I, n=7) or a physically inactive (group II, n= 7) group on the basis of leisure–time physical activity. Rectal temperature, oral temperature, resting pulse rate, subjective arousal and sleepiness were measured at 02:00,06:00, 10:00, 14:00, 18:00 and 22:00 in a counter–balanced sequence for each subject. Whole–body flexibility, back and leg strength, grip strength (right and left), flight time in a vertical jump, PWC150 and self–chosen work–rate were also recorded at each time point. At least 8h separated each test session. Subjects avoided exercise 48h prior to, and during the experiment. Data were subjected to the group cosinor method. Group I evidenced 1·5–2·5 times greater rhythm amplitudes than Group II for oral temperature, subjective arousal, sleepiness, flexibility, left and right grip strength, submaximal heart rate, and self–chosen work–rate (p<0.05). Oral temperature and arousal for Group I were lower than Group II only at 06:00. Early morning troughs in most of the performance measures were significantly greater for Group I (p<0.05). The groups did not differ with respect to phasing of the rhythms (p<0.05). These results confirm with physical performance measures that rhythm amplitudes are higher for physically fit subjects. This could be attributed to greater early–morning troughs in the measures for active individuals. Since the subjects were sedentary immediately prior to testing, it is plausible that these findings are training effects of physical activity.