Roughton and Forster (RF) proposed to split the lung diffusing capacity into two contributions describing first, diffusion to red blood cells (RBC), and second, capture by diffusion from the RBC surface and reaction with haemoglobin. Solving the diffusion-reaction equations for simplified capillary-RBC structures, we investigate the RF interpretation. This reveals first that the conventional extrapolation to zero pressure of 1/DLCO on PO2 is not a correct measure of the diffusive component. Consequently the capillary volumes deduced from this extrapolation are erroneous. Secondly, capture mechanisms are different for CO and NO: while DLCO characterizes "volume absorption" in the RBC and is correlated with hematocrit, DLNO quantifies "surface absorption" and provide information about the morphology of the space between the alveolar surface and the RBC surfaces. In conclusion, the RF approach may lead to erroneous physiological interpretations of DLCO; nevertheless, the measurement of DLCO and DLNO bring different types of information that give the potential for a better understanding of respiratory diseases.