Objective: This study evaluated the diagnostic process and prevalence rates of Asperger syndrome (AS) according to the DSM-IV, ICD-10, and criteria developed by Gillberg and Gillberg and by Szatmari and colleagues and clarified confusion about AS. Method: An epidemiological study of 5,484 eight-year-old children in Finland, 4,422 (80.6%) of whom rated on the high-functioning Autism Spectrum Screening Questionnaire by parents and/or teacher, 125 of them screened and 110 examined by using structured interview, semistructured observation, IQ measurement, school day observation, and patient records. Diagnoses were performed by following the DSM-IV, ICD-10, and criteria developed by Gillberg and Gillberg and by Szatmari and colleagues in detail. Results: The prevalence rates per 1,000 were 2.5 according to the DSM-IV, 2.9 to ICD-10, 2.7 to Gillberg and Gillberg's criteria, and 1.6 to the criteria of Szatmari et al. Conclusions: The results emphasize the need to reconsider the diagnostic criteria of AS. The importance of multi-informant sources came up, and the need of several informants was highlighted, especially when diagnosing the broader pervasive developmental disorders. (Contains 3 tables and 1 figure.)