Habitual sleep duration varies greatly among individuals. The biological meaning of this variation is unknown. The present project proposes that differences in habitual sleep duration are associated with differences in the output of the endogenous circadian pacemaker, which programs the 'internal biological night'. The protocol hypothesizes that the duration of the internal biological night as defined by the nocturnal intervals of detectable plasma melatonin levels, low body temperature, low EEG activity in the high-frequency alpha band, increasing sleepiness and increasing plasma cortisol levels is longer in habitual long sleepers (sleep duration greater than 9 hours) than in short sleepers (less than 6 hours). Since recent results in healthy subjects suggest that melatonin has some soporific or hypnotic properties, we furthermore hypothesize that the level of nocturnal melatonin secretion is higher in long sleepers than in short sleepers. The circadian rhythms of plasma melatonin, body temperature, waking EEG activity, subjective sleepiness and plasma cortisol will be assessed in a ~40-hour constant routine protocol. In addition, a 36-hour extended bed rest protocol will be carried out in order to study spontaneous sleep duration and sleep structure under ad lib conditions while time cues and social cues are absent. It is hypothesized that as a consequence of differences in the circadian output, long sleepers will sleep more, have a different sleep structure, and spontaneously wake up on a lower level of homeostatic sleep pressure than short sleepers. Finally, we intend to determine whether differences in habitual sleep duration between long and short sleepers are associated with dysthymic and hypomanic characteristics of mood, respectively. We are asking to study a total of 15 long sleepers and 15 short sleepers. Given that we have already studied 9 long sleepers and 9 short sleepers, we anticipate that 72 patient-days per year for one year would be required to complete the study.