The mechanisms underlying age-related changes in the signal from the biological clock have yet to be determined. The authors sought to determine if the phase advance of circadian melatonin rhythm during the middle years of life is related to different patterns of habitual light exposure. Forty-one healthy subjects between the ages of 22 and 58 y were studied. Habitual light exposure was measured by a wrist monitor for 7 days. Participants underwent a 25-h constant routine. They provided saliva samples every 30 min, and melatonin concentration was determined by radioimmunoassay to assess salivary dim light melatonin onset (S-DLMO(1.3)). Aging was associated with earlier S-DLMO(1.3). Increasing age was not related to the time spent at different light intensities. However, it was associated with lower percentage of light exposure during the night (between 0200-0400, 0600-0700, and 2300-2400 h) and with higher percentage of light exposure in the morning (between 0800-1100 h). Earlier S-DLMO(1.3) was associated with lower percentage of light exposure early on in the night (between 2200-0000, 0000-0100, and 0200-0300 h) as well as in the afternoon (between 1500-1600 h) and with higher percentage of light exposure in the morning (between 0800-1100 h). When the effects of age were controlled, there was no significant relationship between S-DLMO(1.3) and percentages of light exposure. Yet increasing age was associated with earlier S-DLMO(1.3) regardless of light exposure patterns. Earlier habitual wake time explained the earlier light exposure patterns of older subjects. Both habitual wake time and age contributed to the prediction of S-DLMO(1.3). The results suggest a phase advance of circadian rhythms in the middle years of life. Whereas a clear change in habitual light exposure patterns was associated with aging and with shifts in S-DLMO(1.3), it did not explain entirely the age-related advance of melatonin circadian phase.