One objective of this study was to investigate whether the assumption in the single-breath method is correct, that gas concentrations measured in the second 750 mL are representative of the entire lung. This assumption seems acceptable for CH4 in healthy volunteers but not in COPD patients. TLC is best approximated with the multibreath helium dilution technique that lasts several minutes (TLCMB). Theoretically, TLCSB cannot be larger than TLCMB because of the shorter time available for wash-in of inert gas. TLCMB was determined by helium dilution, while TLCSB and TLCTB were determined by CH4 dilution. In some cases, we found that TLCMB was smaller than TLCSB or TLCTB. This might be explained by the difference in solubility of helium and methane. Both in healthy volunteers and in patients with obstruction, average TLCSB is significantly smaller than TLCMB. In healthy volunteers, this significant but small difference results from sequential filling and emptying and inhomogeneities caused by hydrostatic pressure differences due to gravitation, resulting in a small downward slope in the CH4 vs exhaled volume relationship.5,17,22,24–26 In all COPD patients, not only should these hydrostatic pressure differences be taken into account, but also uneven distribution of time constants resulting in asynchronous and inhomogeneous ventilation.4 As a result, the downward slope in the CH4 vs exhaled volume relationship is steeper in patients with obstruction than in healthy volunteers, causing a larger difference between TLCSB and TLCTB. Differences are larger in the patients with severe obstruction than in patients with moderate obstruction. The single-breath method increasingly underestimates RV with increasing inhomogeneity of ventilation.,16,27–29 The Bland-Altman plots in Figure 3, top, 3a, and bottom, 3b show that the differences between TLCMB and TLCSB or TLCTB are larger in COPD patients than in healthy volunteers. In the total-breath method, the differences are smaller than in the single-breath method because in Figure 3, bottom, 3b, the mean differences of patients (dashed lines) are closer to the 95% confidence interval of healthy volunteers (area between the solid lines). Therefore, we conclude that TLCTB is a better approximation of TLC than TLCSB.