The reproducibility of total respiratory resistance (Rrs) measured with a simplified forced oscillatory method (Siemens Siregnost FD 5) was measured and compared with that of slow inspiratory vital capacity (IVC) and forced expiratory volume in one second (FEV1). The former technique has the advantage that assessment of bronchial obstruction can be made without a forced maneuver, which may be difficult in patients with chronic obstructive pulmonary disease (COPD). We used the criteria proposed by the American Thoracic Society for the diagnosis of COPD. Pulmonary function tests (IVC, FEV1 and Rrs) were measured in seven healthy subjects and in two groups of patients with COPD. First one technician performed six measurements of IVC, FEV1 and Rrs in all subjects during a period of 90 minutes on the same day. Second to evaluate intraindividual variability the measurements were performed on ten subsequent days. The median interval (range) between the first and last measurements in days was 38 (20-186). The mean +/- SD coefficient of variation (CV) Rrs in patients was 15.7% +/- 5.0% and in normals, 10.8% +/- 3.2%. There was less variation in the FEV1 value of 11.0% +/- 6.2% and normals, 2.2% +/- 1.0%; and IVC, 6.9% +/- 5.0% and normals, 2.4% +/- 0.7%. There is no correlation between age and CV. It is concluded from the study that oscillatory Rrs has a larger coefficient of variation within one patient than FEV1 or IVC. If Rrs is used for longitudinal follow-up in COPD patients, we suggest that variations less than 26% (mean +/- 2 SD) can be considered the result of "spontaneous" variation in lung function.