Attempts to compare bronchial responsiveness between populations have been hampered by between-study differences in the pharmacological agent of provocation, the method of administration and the summary statistic employed. The European Community Respiratory Health Survey used methacholine challenge delivered by Mefar dosimeter according to a standardized protocol used in 35 centres in 16 countries. Data were obtained from 13,161 men and women, aged 20-44 yrs at the start of the study. The dose of methacholine producing a 20% fall in forced expiratory volume in one second (FEV1) (PD20) and the regression coefficient of percentage decline in FEV1 with log dose, were calculated ("slope", after transformation), with and without calibration of nebulizers by weight and adjustment for nonresponse bias. Standardization for baseline lung function and variation in smoking prevalence was applied to slope. Results were robust to whichever summary measure was used, and to the various adjustments. Responsiveness was low in Iceland and Switzerland, and in most centres in Sweden, Italy and Spain, and high in New Zealand, Australia, the USA, Britain, France, Denmark and Germany. Bronchial responsiveness varies considerably in Europe, and high levels are not confined to the English-speaking world.