A circadian rhythm disturbance in narcolepsy has been postulated. To investigate this issue, the 24-h pattern of rectal temperature variation was monitored under entrained conditions in narcoleptics selected on the basis of documented sleep-onset REM episode(s). The 24-h temperature mesor of narcoleptics was elevated in comparison with control subjects. This is attributable to higher nocturnal temperatures, perhaps related to increased sleep disturbance. The actual recorded temperature minimum in narcoleptics appeared 1 h after sleep onset, independent of the occurrence of sleep-onset REM, compared with 4-5 h after sleep onset in control subjects. In contrast, the results of cosinor analyses indicated that the estimated temperature nadir is phaseadvanced in those narcoleptics who exhibited sleep-onset REM during the 24- h temperature study compared with those without sleep-onset REM. An early nocturnal temperature minimum in narcolepsy is consistent with recent reports that the latency to the appearance of REM sleep in normals varies as a function of the phase of the circadian temperature rhythm when sleep onset occurs.