Different protocols for pulmonary function testing were evaluated in 611 children, 6-9 years old. Attention was paid to the necessary minimum number of acceptable maneuvers, the necessary maximum number of attempts, and the best of seven different methods for summarizing variables from maximum expiratory flow-volume (MEFV) curves. The results of the study indicate that it is not useful to continue pulmonary function testing after eight attempts. The differences between the results from three or five acceptable maneuvers are not substantial for any method; in children, obtaining three acceptable curves seems sufficient. When flow variables are selected from one acceptable curve, as in the American Thoracic Society (ATS) guidelines, the reproducibility is somewhat less good than in methods in which results of various curves are averaged or selection is made from different curves. The method that has been recommended by the European Community for Coal and Steel for the selection of MEFV variables in teenagers and adults in Europe, was also demonstrated to be suitable for younger children.