This is an excerpt from the report (1992) of the Epidemiology Data Committee of the Japan Diabetes Society. Portions regarding the community-based prevalence data on adult diabetes and the data of official sources are shown. Data on the prevalence and incidence of diabetes were collected from Japanese medical literature from 1959-1991 by various means. In early studies, people were first screened by urine glucose test, and subjects positive for glucosuria were further evaluated by blood glucose test. Recent studies mostly used 75 g GTT without prior urine screening. Before 1984, the criteria of the JDS (1970) were chiefly used, while after 1985 the WHO criteria are used to define diabetes. Prevalence of diabetes in people older than 40 years was 1.3-4.7% in earlier studies, but it increased to 4-11% in recent studies despite the fact that the WHO criteria are more stringent in defining diabetes than the JDS criteria of 1970. The prevalence of diabetes was higher in men than in women in most reports. Some publications from the Ministry of Health and Welfare included the data on diabetes. They were based on the information of known diabetic patients seen by doctors, and gave much lower estimates of prevalence than community-based survey. The estimated prevalence of diabetes was 1.7% from the 1987 Patient Survey and 1.1% by the Comprehensive Survey of Living Conditions. There was a sharp 30-fold increase during these 30 years. Efforts are now being made to collect more accurate data on the prevalence of diabetes in Japan by the Epidemiology Study Group sponsored by the Ministry of Health and Welfare.