Evidence from human free-running studies has suggested a close relationship between the timing of the circadian rhythm of core body temperature and the rhythm of sleep propensity. However, this relationship may be questioned by variations of sleep and wakeful activity which could have masked the endogenous temperature rhythm. A constant routine was used here to 'unmask' the endogenous temperature rhythm in addition to frequent sleep trials across a 24-h period to confirm the relationship between temperature and sleep propensity rhythms. Of the 14 healthy, good sleeping subjects 13 had significant 24-h cosine rhythms of sleep propensity. Eight of these also had a significant 12-h cosine rhythm. The eight subjects with both 24-h and 12-h rhythms showed a minor peak of sleep propensity in the early afternoon followed by a though in the early evening (20.00 hours). Sleep propensity then rose rapidly at about midnight to a major peak in the early morning. This was followed by a second trough of sleep propensity in the late morning. The average times of the sleep propensity phases relative to the circadian temperature rhythm were very similar to the earlier free-running studies. Furthermore, the times of the sleep propensity phases were highly correlated with the body temperature minimum. These results suggested the possibility that a common oscillator determines the timing of both the body temperature rhythm and the phases of the sleep propensity rhythm.