The criteria of the American Thoracic Society (ATS) for determining the end-point of the forced vital capacity (FVC) maneuver were compared with an algorithm employed by a computer system (Hewlett-Packard), which yields a later end-point. The 150 subjects tested were divided into five groups of 30 on the basis of the following spirometric diagnoses: normal; small airways' obstruction (SAO); mild airways' obstruction (MAO); severe chronic obstructive pulmonary disease (COPD); or restrictive disease. The subjects performed a minimum of three maneuvers according to ATS standards, and the flow-volume curves with the greatest sum of FVC and forced expiratory volume in one second (FEV) were chosen for analysis by the two algorithms. Hewlett-Packard (HP) values for FVC and FVC time were always higher than the corresponding ATS values, and the HP values for flows were always lower than the corresponding ATS values. The higher differences were observed in the SAO group. In the group with severe COPD, high FVC differences and low flow differences were observed; on the contrary, in the restrictive group, low FVC differences and high flow differences were present. These results, in addition to the different location of the FVC end point, may be explained by the different morphology of the flow-volume curves. In conclusion, the ATS algorithm caused a systematic underevaluation of FVC and a systematic overevaluation of flows, which cause practical consequences only in the SAO group. In fact, 28 percent (17/60) of the subjects with SAO characteristics were considered "normal" using ATS criteria for the end-point.