In three foxhounds, diffusing capacity for carbon monoxide (DL(CO)) was reduced by 25-30% after left pneumonectomy. Based on previous morphometric data in animals and physiologic data in humans, this reduction should not result in any impairment in gas exchange. However, experimental evidence indicates that diffusion limitation develops during exercise after pneumonectomy. Our objective is to determine whether this diffusion limitation to gas exchange can be predicted from physiologic measurements of DL(CO). DL(CO) measured by the rebreathing technique was translated into diffusing capacity for O2 (DL(O2)) using an average conversion factor for canids obtained morphometrically (Weibel et al., Respir. Physiol. 54: 173-188, 1983). Arterial O2 saturation (Sa(O2)) at various intensities of steady state exercise was calculated from DL(O2) and measured values of O2 consumption, alveolar P(O2), hemoglobin and arterial pH, and compared to observed Sa(O2). After pneumonectomy, Sa(O2) declined progressively with increasing exercise load. In all dogs, the observed pattern of arterial O2 saturation could be predicted from DL(CO) measured at similar work loads. The relationship between predicted (Pr) and observed (Ob) Sa(O2) is: Sa(O2)(Pr) = 22.73 + 0.77Sa(O2)(Ob), r = 0.92. The slope is significantly less than 1.0 (P < 0.005), indicating that other factors must also contribute to arterial desaturation. We conclude that physiologic measurement of DL(CO) is a meaningful indicator of diffusion limitation to gas exchange. In the foxhound, a modest reduction in DL(CO) significantly impairs O2 transport during exercise; but other gas exchange abnormalities, e.g. ventilation perfusion inhomogeneity, must also develop.