To describe bronchial responsiveness in populations, the dose-response slope (DRS), which is the linear slope of the dose-response curve, may be a more useful measure of severity than the provoking dose that causes a 20% fall in FEV1 (PD20 FEV1). To examine the distribution of DRS measurements and their relation to respiratory symptoms in children, we have analyzed data collected during a population study designed to measure the prevalence of bronchial hyper-responsiveness and respiratory symptoms. In this study, respiratory symptom history was measured by a self-administered questionnaire to parents and bronchial responsiveness was measured using the rapid inhalation method. Of 1,217 children studied, DRS values were obtained for 1,203 children who had technically satisfactory bronchial challenge data. In asymptomatic children, DRS values were distributed symmetrically on a logarithmic scale. The geometric mean DRS was reliable measure of the central position of the curve, and 1.96 standard deviations only slightly underestimated the 95% interval. In children with recent wheeze, the distribution was skewed toward larger DRS values. Mean DRS values were significantly different between groups determined according to symptom frequency. It appears that DRS values are more useful than PD20 FEV1 values in epidemiologic studies of respiratory illness in children because a value that relates well to symptom history can be calculated for the entire sample. The major advantages are that the measurement more clearly discriminates between symptom severity groups and that a value that represents abnormal responsiveness can be calculated.