A repeated testing paradigm was used to assess the efficacy for the management of daytime sleepiness in narcolepsy-cataplexy of single long, multiple short and no-nap sleep/wake schedule conditions, with total sleep per 24 hours held constant. Eight narcoleptic subjects participated and followed each experimental schedule for two consecutive days, the second of which served as a test day during which simultaneous electroencephalogram (EEG) polygraphic recordings were made. Performance tests reported here include a grammatical transformation test and a four-choice reaction time test. A single long nap placed 180 degrees out-of-phase with the nocturnal midsleep time improved sustained performance over the no-nap condition. Reaction time performance was significantly improved in the long nap condition over the no-nap condition. Time-of-day analyses found that the greatest improvement was in the afternoon and evening. By contrast, the grammatical transformation test results suffered under the napping compared to no-nap schedules, suggesting that continuity of wakefulness and/or a long nocturnal sleep period may be important for this test. In addition, unscheduled sleep episodes tended to occur earlier in the day than the period of maximum afternoon sleep tendency seen in normal subjects. Two napping strategies are suggested for further study.