In an attempt to predict the severity of emphysema in patients with moderately severe and severe chronic air-flow obstruction, antemortem pulmonary function data, including spirometry, subdivisions of lung volumes, diffusing capacity (transfer factor) for carbon monoxide, and elastic recoil were assessed in 46 patients who were autopsied during the National Institutes of Health Intermittent Positive Pressure Breathing Clinical Trial and we compared these to the morphologic severity of emphysema. The severity of emphysema was graded by the panel grading method using whole lung, paper-mounted (Gough-Wentworth) sections and the mean linear intercept (Lm). The FEV1 and FEF25-75 of the FVC, the DLCO, the diffusing capacity for carbon monoxide, and subdivisions of lung volumes showed significant but low order correlations with the emphysema score and Lm. Total lung capacity, determined by plethysmography, was better related to emphysema in this study than in others in which TLC was measured by helium dilution. Volume-pressure data fitted to the exponential equation (V = A - Be-KP) yielded a low order, but a significant relationship between Lm and the exponential constant (K), but not between K and the panel emphysema score. We conclude that the recognition of the presence and severity of emphysema continues to require a multivariate approach including clinical history, assessment of air-flow obstruction via routine spirometry, radiologic assessment of the lung with emphasis on total lung capacity, and evaluation of the diffusing capacity for carbon monoxide.