OBJECTIVE: This study characterized objectively the hypersomnia frequently seen in the depressed phase of bipolar affective disorder. On the basis of previous work in sleep and affective disorders, it has been hypothesized that the hypersomnia is related to greater REM sleep. This hypothesis was tested by using a multiple sleep latency test to compare bipolar affective disorder with narcolepsy, a well-defined primary sleep disorder associated with known REM sleep dysfunction. METHOD: Twenty-five bipolar depressed patients were selected on the basis of complaints of hypersomnia. They underwent 2 nights of polysomnography followed by a multiple sleep latency test. Data on their nocturnal sleep and daytime naps were compared with similar data on 23 nondepressed narcoleptic patients referred for sleep evaluation. RESULTS: Despite their complaints of hypersomnia, no abnormalities were noted for the bipolar group in the results from the multiple sleep latency test. Contrary to the working hypothesis, REM sleep was notably absent during daytime naps in the depressed patients, in marked contrast to the findings for the narcoleptic group. CONCLUSIONS: The complaint of sleepiness in the hypersomnic bipolar depressed patient appears to be related to the lack of interest, withdrawal, decreased energy, or psychomotor retardation inherent in the anergic depressed condition, rather than an increase in true sleep propensity or REM sleep propensity.