OBJECTIVE: Concern that disturbances of sleep and light exposures at night might increase cancer risks have been expressed, but little actual exposure data has been collected. Measurements from a representative population sample were examined to understand the magnitude of in-bed light exposure at night and possible correlates. METHODS: From 1990 to 1994, a home survey of sleep disorders among adults ages 40-64 was conducted in the City of San Diego California, using stratified representative sampling techniques. Along with questionnaires, sleep logs, and 3-night wrist activity and pulse oximetry measures, bedside illumination was measured with a computer recording system. Questionnaires included the CESD depression scale and a scale of symptoms typical of winter depression. RESULTS: Complete data were available from 286 men and women, whose mean in-bed intervals averaged 7 hours and 42 minutes. The mean room illumination during the first part of the night was mean 12.7 lux (median 3.2 lux) and during the last 2 hours in bed averaged 28.7 lux (median 18.9 lux). Nocturnal light exposure was positively correlated with age, male gender, summer season, time in bed, wake-up time, and depressive symptoms. CONCLUSION: Complex bi-directional interactions may take place between sleep disturbances, depression, time in bed, wake-up-time, and in-bed illumination. The most crucial light exposures appear to occur in the last 2 hours in bed, largely after dawn, so daylight exposure may be an important factor.