The objective of this investigation was to clarify the epidemiology of idiopathic Parkinson's disease (PD) in the county of Rogaland, Norway. Total case ascertainment and a high diagnostic accuracy were attempted through a detailed community study and the use of a new clinical diagnostic classification. The study population comprised 220,858 inhabitants, and a total of nearly 400 patients was interviewed and examined by a neurologist. On prevalence day, January 1, 1993, 245 patients were included in the study. The diagnostic classification revealed 135 patients with clinically definite, 74 with probable, and 36 with possible PD. The crude prevalence rate was shown to be 110.9 per 100,000 inhabitants. The total age-adjusted prevalence was calculated to be 102.4 per 100,000 and to 120.9 per 100,000 men and 89.8 per 100,000 women. Among the 245 patients, 28 patients had a tremor-dominant disease, 50 patients an akinetic-dominant disease, and 167 patients a mixed clinical pattern of PD. Age-adjusted prevalence figures were slightly higher for rural compared to urban areas. About 50% of the PD patients were in need of public help, 15% had complaints about pain related to their parkinsonism, and after approximately 6 years of levodopa treatment, 20% were suffering from clinical fluctuations. The study showed that 40% of the patients had some degree of thought disorder. The prevalence figures for PD in this study are slightly lower than those reported from most previous prevalence studies with a comparable study design for case finding. This may be due to a careful diagnostic evaluation with the use of specified diagnostic criteria, excluding patients with other parkinsonian syndromes.