We wanted to examine the predictive nature, over one year, of bronchial hyperresponsiveness (BHR) and recent wheeze (in the previous 12 months), which are the measurements frequently used to classify asthma for epidemiology. A prospective cohort study of 236 children, aged 8-11 yrs, was undertaken, with an initial baseline study, followed by four studies at three monthly intervals over one year. At each study, we measured bronchial responsiveness to histamine by the rapid method, respiratory symptoms by parent questionnaire, and atopy by skin prick tests to common allergens. Airflometer readings, which are closely related to FEV1, were self-recorded. Baseline data were used to classify children into the four categories of "current asthma" (BHR and wheeze), "wheeze only", "BHR only", or "normal". During the year following baseline study, the group initially classified as "wheeze only" had normal Airflometer variability, and 59% had wheeze, and 33% had BHR, which tended to be mild. In the group initially classified as "BHR only", 52% had wheeze with a peak in winter, and 62% had BHR during the following year. This group had more severe bronchial responsiveness and Airflometer variability than the normal and wheeze only groups. The group initially classified as current asthma had a more severe condition, with continued BHR (100%), and wheeze (93%), increased Airflometer variability and more atopy. Thus, the natural history of bronchial responsiveness, respiratory symptoms and allergic history in this group was different from the other three groups.(ABSTRACT TRUNCATED AT 250 WORDS)