We compared bronchial responsiveness to isocapnic hyperventilation of cold dry air at -18 degrees C and 0% humidity with bronchial responsiveness to inhaled methacholine in 24 subjects with current or previous asthma and 2 nonasthmatics. Two inhalation tests with each agent were carried out in random order on 4 consecutive days. The response to cold air was expressed as the respiratory heat exchange required to reduce the FEV1 by 10% (PD10 RHE) and the response to methacholine as the provocation concentration required to reduce the FEV1 by 20% (PC20 methacholine). There was a close positive linear correlation between PD10 RHE and PC20 (r = 0.86, p less than 0.001). The responsiveness to each agent was highly reproducible. The PD10 RHE could be measured in all 21 subjects with current symptoms of asthma and it could be obtained by extrapolation in 2 normal subjects, but it could not be measured in 3 subjects with a past history of asthma. The PC20 in the current asthmatics was 6.3 mg/ml or less, in the 2 nonasthmatic subjects, it was 14 and 16 mg/ml, and in the previous asthmatics it was between 26 and 54 mg/ml. The results indicate that nonspecific bronchial responsiveness is an important factor influencing the bronchial response to cold air, that either cold air or methacholine are suitable stimuli to measure nonspecific bronchial responsiveness, and that the differences in bronchial responsiveness observed between asthmatics and nonasthmatic subjects are in keeping with a quantitative rather than a qualitative difference in responsiveness.