We investigated the prevalence of narcolepsy using a well-defined white population previously used for epidemiological investigations: the Finnish Twin Cohort. The Cohort consists of 13,888 monozygotic and dizygotic twin pairs born before 1958. There were 16,179 individuals who participated in the study, with a 77.3% response rate. The study methodolgy included a questionnaire covering sleep and alertness, the Ullanlinna Narcolepsy Scale (UNS), a scale specifically developed and tested for the study, telephone interviews, and finally, clinical evaluation, polygraphic recording, and HLA blood typing. Seventy-five subjects were selected for telephone interviews and laboratory evaluations based on data from both questionnaires. Five of them were strongly suspected of narcolepsy, but laboratory data identified only 3. All were dizygotic (fraternal) twins discordant for the disease with a negative family history and presence of DR2 DQw1 (i.e., DRw15 DQw6, new World Health Organization classification). The prevalence of narcolepsy in the Finnish population is 0.026% (95% confidence interval, 0.0–0.06). This prevalence is lower than that reported in studies performed without polygraphic recording and is close to that reported in 1945 in the black U.S. population. The tools developed to perform this study, the largest population study of its kind yet performed, can now be used for other population investigations.