Effects of two different light intensities during daytime were examined on human circadian rhythms in plasma melatonin, core body temperature, and wrist activity under a fixed sleep schedule. Sleep qualities as indicated by polysomnography and subjective sleepiness were also measured. In the first week, under dim light conditions ( approximately 10 lx), the onset and peak of nocturnal melatonin rise were significantly delayed, whereas the end of melatonin rise was not changed. The peak level of melatonin rise was not affected. As a result, the width of nocturnal melatonin rise was significantly shortened. In the second week, under bright light conditions ( approximately 5,000 lx), the phases of nocturnal melatonin rise were not changed further, but the peak level was significantly increased. Core body temperature at the initial sleep phase was progressively elevated during the course of dim light exposure and reached the maximum level at the first night of bright light conditions. Subjective sleepiness gradually declined in the course of dim light exposure and reached the minimum level at the first day of bright light. These findings indicate that repeated exposures to daytime bright light are effective in controlling the circadian phase and increasing the peak level of nocturnal melatonin rise in plasma and suggest a close correlation between phase-delay shifts of the onset of nocturnal melatonin rise or body temperature rhythm and daytime sleepiness.