An analysis of spirograms of 6,486 subjects from the general population, ages 8 to 90, was conducted to determine their ability to satisfy the American Thoracic Society's (ATS) acceptability and reproducibility criteria. The results indicate that both older and younger subjects had more difficulty satisfying the ATS acceptability and reproducibility criteria. The difficulty in satisfying the ATS reproducibility criterion, particularly in younger subjects, was in part associated with their smaller heights and lung volumes. A relatively uniform within-subject variability of FVC and FEV1 in terms of the mean differences between the largest and second largest FVC and FEV1, for all heights, was observed. In addition, unlike the ATS reproducibility criterion, when a constant 200-ml reproducibility criterion for FVC and FEV1 was used, there was no longer a significant difference between the number of reproducibility criterion failures for the 14 different height groups used. These results suggest that the ATS reproducibility criterion, based on a percentage of the FVC and FEV1, may inappropriately classify a higher percentage of subjects with smaller heights and lung volumes as having a nonreproducible test. In contrast, subjects with larger heights and lung volumes are much less likely to fail the ATS reproducibility requirement. These results emphasize the importance of following the ATS recommendation of using the reproducibility criterion only as a goal during data collection, not to classify a subject as having an invalid test.