The prevalence and associations of bronchial hyperresponsiveness were investigated in a general practice population. The sample was obtained by using every 12th patient on the practice age-sex register, replacing non-responders with corresponding age and sex matched individuals from up to two further 1 in 12 samples. The response rate was 43%; 366 patients were studied. Doubling concentrations of methacholine were given to a maximum of 32 mg/ml or until a 20% fall in forced expiratory volume in one second (FEV1) occurred (provocation concentration, PC20FEV1). Bronchial hyperresponsiveness was defined arbitrarily as a PC20FEV1 of 2 mg/ml or less (or 11 mumol cumulative dose, PD20FEV1). The prevalence of bronchial hyperresponsiveness was 23%. Bronchial hyperresponsiveness was not associated with age but was more prevalent in women than men (31%:13%). It was also more common in those who had ever wheezed (39%) and in those who had had an attack of rhinitis in the preceding month (45%, p less than 0.1), in atopic individuals (30%), and in smokers (32%), but it was not associated with cough or dyspnoea. There was a positive correlation between PC20FEV1 and resting FEV1 (r = 0.288) and a negative correlation between PC20FEV1 and mean daily peak flow variability (r = -0.356). Stepwise binary logistic regression analysis showed significant independent effects on PC20FEV1 for mean daily peak flow variability, gender, number of positive skin test responses, resting FEV1, and mean histamine skin weal area, but no relation with smoking or mean allergen weal area. The prevalence of bronchial hyperresponsiveness was much higher than the prevalence of diagnosed asthma in the practice in 1984 (4.9%). Analysis of case notes of 169 individuals showed that those with bronchial hyperresponsiveness had not attended the practice more frequently for respiratory complaints during the previous five years.