Standards for high-quality lung function testing have not yet been formally adopted in the RSA, despite the increase in the performance of spirometry. A study was undertaken to determine the standard of spirometry in clinical practice in this country. Forty-five spirometer users agreed to participate. Responses to a questionnaire, administered by personal interview, were graded according to the proportion of correct or appropriate answers: completely unsatisfactory less than 33.3%; poor 33.3-66.6%; and satisfactory greater than 66.6%. Only 6 spirometers were regularly calibrated. Knowledge of international standards for spirometry, the basic working mechanism of the spirometer being used and calibration ranged from poor to completely unsatisfactory in most of the practices. Instructions to the patients for performing the manoeuvres were satisfactory, but elevation of the result for acceptability, reproducibility, selection of the best test and diagnosis of a positive bronchodilator response were generally completely unsatisfactory. Only 18 practitioners knew the source of the reference values they used and 5 had proved their validity with a sample. Fourteen of the 45 practitioners were unable to make the correct diagnoses from two typical test results. Of the 45 practitioners, 40 were interested in a course in practical, clinical spirometry. In the light of the predominantly unsatisfactory results obtained in this study, we consider clinical spirometry to be an urgent priority for all levels of medical education.