The relationship between symptoms and pulmonary function in asthma is important if the latter is to be held relevant to management guidelines and their audit. Associations between reported symptoms, pulmonary function and therapy were studied in 824 asthmatics (mean FEV1 75.4% predicted; best FEV1 84.6% predicted; and actual/best peak flow (PEF) 87.5%). Bronchodilator usage (reflecting symptomatic wheeze) was evenly distributed up to eight times daily; 22.5% of subjects had nocturnal disturbance and 46.3% persistent daytime symptoms. The univariate relationships between symptoms and function were generally closer with best rather than actual/best. They were further explored using quintiles of function. Symptoms were consistently less as best function increased, but were highly significantly greater in the fifth than in the third and fourth quintiles of actual/best FEV1. There was a trend to a similar U-shaped relationship of actual/best PEF with nocturnal disturbance and daytime symptoms. Best function is a good determinant of expected symptom load in an asthmatic population. Below 85% actual/best function reflects the prevalence of symptoms. In asymptomatic patients a level of at least 85-90% is a useful check of physiological control but will not exclude some symptomatic patients, irrespective of best function.