Seventeen healthy adults exercised for 30 min (70% VO2max), at different times of the day and night throughout a 24 h period. The effect on the phase of core temperature rhythm was assessed by comparing the rhythms on the day immediately before and immediately after the day of exercise with the participants living normally on these 2 days. Assessments were made upon data that had been corrected for the thermogenic effects of sleep and activity. When exercise was performed in the period between 4 h before and 1 h after the temperature minimum, there was a phase delay of 1.03 +/- 0.78 h (mean +/- s; n = 6); when performed between 3 and 8 h after the temperature minimum, there was a phase advance of 1.07 +/- 1.23 h (n = 9). Both means were significantly different from zero (Student's one-sample t-tests: P < 0.05). Performed at other times, exercise had no significant effect on the phase of the temperature rhythm. The phase changes were not significantly different from those previously observed by us in sedentary individuals exposed to domestic lighting. We conclude that the amount of physical activity has very little effect on the body clock.