The effect of sleep at regular times on the stability of rhythms was studied in young male and female students. In the first design, subjects slept for 8 hours at irregular times. In the second design, they slept in two 4 hour long periods per day. In another design, one 4 hour sleep period was taken at a constant time on each day. This period designated anchor sleep (AS). The second 4 hour sleep was taken at a different time. Meals were eaten at the same times as during the control phase. Circadian rhythms were obtained by fitting cosine curves to the data. Mean rhythm periods for group- A, male subjects sleeping 8 hours per day at different times, and group-B, female students sleeping in two 4 hour periods, were 24.516 and 24.683 hours, respectively. Acrophases of cosine curves for group-C, female students with AS at 0000 to 0400 hours, and for group-D, male students with AS as group-C, fit to experimental phase. Results from subjects of group-E, who took AS sleep at the same time as C and D groups, but had reverse ordering of irregularly taken sleep, were very similar. In groups-F, G, and H, who took AS at 0400 to 0800, 0800 to 1200, and 1200 to 1600 hours, respectively, acrophase was sometimes a few hours later in the experimental phase compared to the control phase. Mean period of rhythms was close to 24 hours. A statistically significant increase in period of rhythms was observed in the initial phase in groups-G and H. Analyses suggested that when AS was either 0000 to 0400 or 0400 to 0800, stable 24 hour rhythms were rapidly obtained and maintained thereafter. The authors conclude that the phase of the stable 24 hour rhythms in AS experiments depends on time of AS, with phase shift approximately equal to time by which midsleep is shifted.
The Twenty-Four Hour Workday: Proceedings of a Symposium on Variations in Work-Sleep Schedules, Division of Biomedical and Behavioral Science, NIOSH, U.S. Department of Health and Human Services, Cincinnati, Ohio, DHHS (NIOSH) Publication No. 81-127