The incidence of anorexia nervosa during a 45-year period (1935 through 1979) was determined for the population of Rochester, Minnesota, from the epidemiology archives at the Mayo Clinic. The medical records of local residents with the diagnosis of anorexia nervosa were the primary source for case ascertainment, although records of those whose diagnoses might have been shielding anorexia nervosa were also screened. Standard criteria for diagnosis were applied. We identified 140 (128 female and 12 male) residents of the community who fulfilled the diagnostic criteria for anorexia nervosa. No significant long-term trend in rates was ascertained. The overall age- and sex-adjusted incidence rate was 7.3 per 100,000 person-years. The highest age-specific incidence rate (56.7 per 100,000 person-years) occurred in female residents 15 to 19 years old. The prevalence rate on Jan. 1, 1980, for Rochester residents with a history of the disease, age- and sex-adjusted to the 1970 US white population, was 113.1 per 100,000 (203.9 for female and 16.9 for male residents).