We studied the performance and adaptability of 40 nurses (median age 35 years), 20 on permanent day shift and 20 on permanent night shift with fast rotation of work and days off, matched for age, gender, and socio-familial responsibilities. For 15 days prior to the study, subjects maintained sleep logs and trained for performance tests. Questionnaires were administered to evaluate adaptability to shift work. During the experimental phase, sleep/wake patterns were monitored using sleep logs and activity/inactivity with wrist actigraphy. Performance levels were measured with the four choice reaction time and memory test for seven letters, eight times/day during the wake period, days on and off. On the last day of work and first day off, 6-sulfatoxy-melatonin levels were assayed from urine samples collected every 2 hours. Estimated total sleep time during the 15-day experimental period was not significantly different in the dayshift and nightshift nurses. Night nurses shifted regularly to daytime activities on days off and, as a group, were significantly sleep deprived on work days with napping on the job in 9 of the 20 night shift nurses (mean of 114+/-45 minutes per shift) and a significant performance decrement during the work period. Further analysis revealed two subgroups of night nurses: The majority (14 nurses) had a mean peak of 6-sulfatoxy-melatonin at 0718 hours on days off and no peak during night work while the other 6 night shift nurses presented a fast melatonin shift with two clear peaks on both work and days off. Comparison of performance scores revealed that all nurses performed similarly on days off. Daytime nurses and fast-shifting night nurses had similar scores on work days, while nonshifting night nurses had significantly lower scores at work. Despite similar gender, age, social conditions, and light exposure levels, a minority of the nurses studied possessed the physiological ability to adapt to a fast-shifting sleep-wake schedule of more than 8 hours and were able to perform appropriately in both conditions. This shift was associated with a change in the acrophase of 6-sulfatoxy-melatonin.