In a within-subjects (n = 18), counterbalanced design, the circadian phase-shifting effects of 3 h of 1) bright light (3,000 lx) alone 2) and bright light combined with vigorous exercise were compared. For each treatment, volunteers spent 3 nights and 2 days in the laboratory, typically receiving the treatment from approximately 2300 to 0200 on night 2. Bedtimes and waketimes were fixed to the volunteers' habits. Illumination was 50 lx during other wake hours and 0 lx during sleep. Bright Light Alone elicited a significant phase delay in rectal temperature minimum (70 min), but not in urinary 6-sulphatoxymelatonin (6-SMT) acrophase (20 min). Bright Light + Exercise elicited a significant phase delay in 6-SMT (68 min), but did not result in a significant difference in shift compared with Bright Light Alone. The study had adequate statistical power (80%) to detect phase-shift differences between treatments of approximately 2-2.5 h. Thus any antagonism of light shifts with exercise could not have been revealed. Within the limited exercise and light parameters of this study, the results suggest that exercise does not reliably modulate phase-shifting effects of late night bright light in humans.