Determinations of pulmonary diffusing capacity for CO (DLCO) by physiological and morphometric techniques have resulted in substantially different values for both DLCO and its major components. To evaluate the differences in these methods of measurement of DLCO, measurements were made under controlled conditions on isolated perfused dog lungs. Multiple gas-rebreathing techniques were used to measure DLCO, the membrane component of the diffusing capacity for CO (DmCO), and pulmonary capillary blood volume (Vc) in both anesthetized dogs and after isolation and perfusion of their lungs. The isolated perfused lungs were than perfusion fixed for morphometric analysis of the components of DLCO. The values obtained morphometrically for Vc were similar to those measured by physiological techniques. Perfusion fixation did not substantially alter the morphometric estimate of DmCO when compared with previous values obtained on inflation fixed lungs. However, the morphometric estimate of DmCO was over 10 times higher than that estimated physiologically. Analysis of the potential errors in the techniques suggests that the correct value for DmCO is substantially higher than that commonly estimated by use of physiological techniques and that the explanation for the difference is due to a number of factors that can influence the binding of CO to hemoglobin under in vivo conditions. The net effect of these factors can be represented by an unknown in each component of the Roughton-Forster relationship so that 1/DL = 1/(U1.Dm) + 1/(U2.theta Vc), where theta is the binding rate for CO to hemoglobin. Because the magnitudes of the unknown terms (U1 and U2) in the Roughton-Forster relationship are likely to be large, this relationship cannot be reliably used to determine Dm and Vc.