An attempt was made to determine: (i) the prevalence of exercise-induced bronchoconstriction among white and coloured schoolchildren in Cape Town; and (ii) the validity of an exercise test for the diagnosis of asthma in the general population. Children (698 white and 494 coloured) were randomly drawn from schools in the northern suburbs of Cape Town. Each child participated in a standard 6-minute exercise test and spirometric measurements were taken before and 10 minutes after exercise with a portable spirometer. The diagnosis of asthma was based on a questionnaire and personal questioning and included those children who, in the past or at present, suffered episodic or continuous airflow obstruction, which was responsive to a bronchodilator. The criterion for the diagnosis of exercise-induced bronchoconstriction was a 10% decline in forced expiratory volume in 1 second after exercise. The prevalence of exercise-induced bronchoconstriction was significantly higher among white children (5.87%) than coloured children (4.05%). The sensitivity of the exercise test was 0.31 and the specificity 0.97. In contrast to the results of hospital-based studies, the negative predictive value of an exercise test (95%) was found to be greater than the positive predictive value (46%). It is therefore concluded that exercise testing is not a useful screening test for epidemiological use; it is probably useful as a challenge test for detecting asthma in the individual patient when the likelihood of the disease is high.