We have previously shown that in some asthmatic patients, plethysmographic total lung capacity, measured from mouth pressure versus plethysmographic volume (Vbox) changes (TLCm), is overestimated, whereas TLCes, derived from esophageal pressure versus Vbox changes, is an accurate estimate of lung volume (VL). We studied 24 patients with chronic airflow obstruction (mean forced expiratory volume in one second, 1.23 L(SD:0.64 L) and mean specific airway conductance, SGaw, 0.052 +/- 0.016 cm H2O-1 X S-1) to reassess the relative merits of TLCm and the 7-min closed-circuit He dilution technique (TLCHe) in this condition. The TLCHe (6.54 +/- 1.09 L) was significantly less (p less than 0.001) than both TLCm (7.53 +/- 1.20 L) and TLCes (7.16 +/- 1.12 L), whereas TLCm was greater than TLCes (p less than 0.001). Differences between TLCes and TLCHe (ranging from -0.11 to 2.17 L) and between TLCm and TLCes (from -0.10 to 1.48 L) were correlated to the degree of airway obstruction as assessed by SGaw (p less than 0.001 and p less than 0.01, respectively). We conclude that VL is overestimated by TLCm and underestimated by TLCHe in moderate to severe chronic air-flow obstruction. The "trapped gas" volume (TLCm to TLCHe difference) is actually made up by these two additive errors.