We treated 17 narcolepsy patients in a placebo-controlled, double-blind, crossover trial with 10-, 20-, 30-, and 40-mg daily doses of selegiline, a monoamine oxidase inhibitor widely used in Parkinson's disease. There was a dose-dependent as well as a statistically and clinically significant improvement in narcoleptic symptoms and polygraphic measures. At 40 mg, there was a 36% reduction in the number of daytime sleep episodes and a 34% reduction in their duration (compared with placebo, mean values). The number of excessive sleepiness episodes decreased by 43%, and the duration decreased by 47%. The number of cataplectic attacks was reduced by 89%. On the multiple sleep latency test, the REM sleep latency increased from 5.0 to 13.3 minutes, and the number of sleep-onset REM periods decreased from 3.1 to 0.6. Sleep (S1) latency was not changed. No intolerable adverse events occurred. The effective dose range was 20 to 40 mg, requiring a low-tyramine diet, which was easy to maintain. In conclusion, selegiline alleviates both main symptoms of narcolepsy--the abnormal sleep tendency and cataplexy. Thus, treatment with selegiline makes it possible to avoid polypharmacy and to use a potent stimulant without known addiction risk.