The diffusing capacity of carbon monoxide (D(L,CO)) and its value normalized to alveolar volume (D(L,CO)/V(A)) are usually estimated with the single breath method at total lung capacity (TLC). Severely ill patients and small children are not able to deliver a satisfactory vital capacity (VC) or hold their breath for 10 s at TLC. The aim of this study was to develop a rebreathing procedure in which diffusing capacity can be determined during spontaneous tidal breathing. The conventional rebreathing method during hyperventilation was modified so that rebreathing volume and gas concentrations were kept constant by CO2 absorption and O2 supplementation. In adult healthy volunteers and in patients who were able to perform both tests, the diffusion indices obtained with this rebreathing method during resting ventilation were compared with those obtained by the single breath method. Predicted (reference) values for rebreathing D(L,CO) and D(L,CO)/V(A) for children were determined. D(L,CO)/V(A) decreased with alveolar volume (V(A)) and increased with alveolar ventilation (V'(A)). In adults at V'(A) above 35 L x min(-1) rebreathing D(L,CO)/V(A) was similar to single breath D(L,CO)/V(A) at similar alveolar volumes. The D(L,CO)/V(A) relative to their corresponding reference values were the same for both methods in adult patients, irrespective of ventilation distribution disturbances. In children the newly described diffusing capacity obtained by the rebreathing method during resting ventilation can serve as a valuable index to assess and follow a diffusion disorder.