In the past 15 years, 411 sporadic narcolepsy patients have been diagnosed in the Hephata Klinik, Schwalmstadt, Germany. They were explored for presence or absence of excessive daytime sleepiness and narcolepsy in their relatives. A subset of 39 patients were explored for presence or absence of parasomnias. Six patients had more than one relative affected by narcolepsy-cataplexy. Forty-seven family members were investigated with the Stanford Center for Narcolepsy Sleep Inventory and a standardized parasomnia questionnaire. Twenty-four relatives had nocturnal polysomnographies and Multiple Sleep Latency Tests. HLA class I typing was performed in all sporadic and familial cases, class II and microsatellite typing was performed in all members of multicase families. Based on the Finnish prevalence study by Hublin et al., 1994, the relative risk for first degree relatives to develop narcolepsy-cataplexy was in our sample 16.5, 34.2 for excessive daytime sleepiness and 426.9 for parasomnias. Cataplexy, excessive daytime sleepiness and single narcoleptic symptoms in the multicase families segregate with the DRBI*1501, CARII:200, CARI: 103, DQBI*0602 haplotype. In two families, members with narcolepsy and isolated symptoms have inherited the DRBI*1501/DQBI*0602 haplotype from the nonaffected parent. The observed segregations in these two families may support the view that narcoleptic symptoms are expressed by DRBI*1501/DQBI*0602 carriers, independent of haplotype origin. Parasomnias do not segregate with a specific haplotype. The frequency of parasomnias in narcolepsy is much higher than in the general population. The empirical risk for first degree family members of narcolepsy patients to develop cataplexy seems to be low, whereas it is higher for EDS and highest for parasomnias.