To assess the relationships among single-breath diffusing capacity for CO (DLCOsb) (13), respiratory symptoms, and cigarette smoking in a general population sample, the data of 718 men and 894 women 20 yr of age or older were analyzed, and comparisons were performed with flow-volume curve (MEFV) variables and the slope of the alveolar plateau (DN2%/L) as well. Percent predicted DLCOsb and its correction for alveolar volume (DL/VA) were significantly lower in smokers than in nonsmokers. The relationship of presence/absence of respiratory symptoms and cigarette smoking with DLCOsb and DL/VA was significant. DLCO indices were almost always selected as discriminant variables in multivariate analysis between asymptomatic and symptomatic subjects. Poor concordance among lung function tests was evident: in men, 30% with abnormal (i.e., lower than 97.5% percentile) and 21% with normal DLCO indices also had abnormal MEFV parameters and/or DN2%/L. In women, the corresponding figures were 24 and 10%, respectively. In men, when considering only DLCO indices, the percentage of symptomatic subjects with abnormal lung function tests ranged from 33% in those with at least one symptom to 45% in those complaining of dyspnea. When the proportion of symptomatic subjects with DN2%/L and MEFV abnormalities were added, it increased to 56 and 66%, respectively. However, in women the proportion of symptomatic subjects with abnormal lung function indices was very small. These results indicate the usefulness of including CO diffusing capacity in epidemiologic surveys in the detection of abnormalities.