Morbidity and mortality from asthma is increased during the grass-pollen season and during the autumn months in the United Kingdom. It is not apparent why this seasonal variation occurs nor whether the variation in morbidity and mortality is associated with variation in bronchial reactivity. We have measured bronchial reactivity on four occasions during 12 months in 60 subjects selected from a community population in the south of England. All subjects had had a histamine challenge test and skin tests to common antigens as part of a survey of asthma prevalence in March 1984. Further measurements of the provocative dose causing a 20% fall in FEV1 (PD20) were made at the peak of the grass-pollen season in June, at the end of September, and in the following March, and current symptoms of respiratory tract infection (RTI) were assessed on all four occasions. Geometric mean PD20 demonstrated significant seasonal variation between 1.38, 0.82, 0.92, and 1.20 mumol in March, June, September, and March, respectively (p less than 0.02). Relative to March 1984, PD20 was significantly decreased in June and September (p less than 0.005 and p less than 0.02, respectively) but not in March 1985 (p = 0.39). Within subjects atopy was significantly related to decrease in PD20 in September (p less than 0.05) and in March 1985 (p less than 0.025) but not in June (p = 0.40). Change in PD20 between occasions was unrelated to RTI symptoms, age, or smoking status, but it was related to change in baseline FEV1/FVC (p less than 0.001).(ABSTRACT TRUNCATED AT 250 WORDS)