The contribution of nonspecific bronchial reactivity to the day-to-day clinical expression of asthma is uncertain. We have examined this relationship in a longitudinal study of eight children and 12 adults. Measurements of reactivity to methacholine were made every 2 to 3 wk over a period of 12 to 18 months, deriving the dose that caused a 20% fall in FEV1 (PD20). Throughout the study, all patients kept a daily record of symptoms and treatment and twice daily measurements of peak expiratory flow (PEF). A significant relationship was found between subjects' overall reactivity (median PD20) and both their average day-to-day variation in morning PEF (Spearman's rho = -0.53, p = 0.016) and diurnal variation in PEF (Spearman's rho = -0.60, p = 0.004). However, examining the temporal relationship between reactivity and asthma within subjects, individual PD20 measurements were not consistently related to concurrent asthma severity: in only six subjects did changes in PD20 generally reflect simultaneous trends in symptoms or PEF. In several patients, exacerbations of asthma occurred in the absence of bronchial hyperreactivity (PD20 greater than 12.8 mumol). We conclude that nonspecific bronchial reactivity is only one mechanism underlying airflow obstruction in asthma, and that its relationship to the clinical state of asthma is not sufficiently close to be of practical clinical use.