To determine whether the incidence of asthma had increased in Rochester, Minnesota, from January 1, 1964 through December 31, 1983, we used a population-based computer-linked medical diagnosis system to identify individual medical records with diagnosis of asthma or other conditions mimicking asthma. All records were reviewed using explicit predetermined diagnostic criteria; we identified 3,622 incident cases of asthma, including definite asthma (n = 1,547), probable asthma (n = 952), and single wheezing episodes (n = 1,123). The annual age- and sex-adjusted incidence of definite + probable asthma rose from 183 per 100,000 in 1964 to 284 per 100,000 in 1983. This rise was entirely accounted for by increased incidence rates in children and adolescents (age range, 1 to 14 yr); incidence rates for infants younger than 1 yr of age and for adults remained constant. For definite + probable asthma cases, the incidence rates for males were higher than for females from infancy through 9 yr of age and for persons older than 50; incidence rates for females were higher than for males from 15 through 49 yr of age. The median age at onset was 3 yr for males and 8 yr for females. We conclude that asthma begins in early childhood, with a higher incidence and earlier onset in males, and that the increase in incidence rates seen from 1964 to 1983 occurred only in children and in adolescents.