Subjects followed a 26-h sleep-wake schedule and were exposed to various light patterns while living at home exposed to the conflicting 24-h zeitgebers. In one protocol, a 26-h light pattern containing evening bright light was compared with a natural-light-only pattern. In another protocol, the evening-light pattern was compared with a morning-light pattern. Rectal temperature was continuously measured. Sleep times were estimated from daily sleep logs. The sleep times of most subjects conformed fairly well to the 26-h sleep-wake schedule, although the evening-light condition produced slightly better results. A larger proportion of subjects had their temperature rhythm entrained to the 26-h schedule during the evening-light condition than during the morning-light or natural-light conditions. Entrainment to the 26-h schedule was achieved in 74% (14/19) of the subjects tested in the evening-light condition. This study shows that non-24-h bright light and sleep-wake schedules can be used to phase shift and entrain human circadian rhythms, despite the presence of the conflicting 24-h zeitgebers.