A new test for assessing clinical dyspnoea with greater precision has been devised. This involves walking on a treadmill while ventilation is measured and dyspnoea is assessed serially with visual analogue scales. The test has been applied successfully in patients with chronic bronchitis and emphysema and in asthmatic patients. The reproducibility of the test was determined by comparing identical periods of exercise separated by an interval of five days in patients with chronic bronchitis and emphysema. The relationship between dyspnoea and ventilation showed acceptable within-patient reproducibility. The sensitivity of the test was investigated in asthmatic patients by comparing exercise after placebo and after salbutamol which had increased PEFR by a mean of 33%. After salbutamol, lower breathlessness scores usually occurred for a given level of ventilation and this indicates that the test was sensitive enough to detect an effective treatment. Thus more precise assessment of dyspnoea in patients is feasible and generally reliable. However, a proportion of patients respond unpredictably and it is necessary to apply this test critically and with an appreciation of its limitations.