We measured diffusing capacities for carbon monoxide (DLCO) and nitric oxide, lung volume, and cardiac output by a rebreathing technique at two alveolar O2 tensions (PAO2) at rest and exercise. Membrane diffusing capacity for CO (DMCO) and VC were estimated from DLCO by the Roughton-Forster (RF) method and also from simultaneous lung diffusing capacity for NO and DLCO measured at one O2 tension (modified RF method). Estimates by these methods agreed closely in normal subjects (Tamhane et al., Chest 2001;120:1850-1856). Using these methods, we studied patients with stages II-III pulmonary sarcoidosis to determine (1) whether the modified RF method accurately estimates DMCO and VC in parenchymal disease and (2) whether sarcoidosis alters recruitment of diffusing capacity with respect to cardiac output. In patients, DMCO and VC estimated by the two methods agreed closely. DMCO was disproportionately reduced relative to VC at any given cardiac output, and the slope of the relationship between DLCO and cardiac output was moderately, though significantly, below normal. We conclude that in sarcoidosis (1) the modified RF method provides comparable estimates of DMCO and VC as the standard RF method and (2) the limitation to diffusive gas transport resides primarily in the membrane barrier, although recruitment of microvascular reserves is also modestly impaired.