In airway obstruction, thoracic gas volume derived from mouth pressure vs. plethysmographic volume changes (TGVm) is overestimated, whereas TGVes, derived from esophageal pressure vs. plethysmographic volume changes, is not (J. Appl. Physiol.: Respirat. Environ. Exercise Physiol. 52: 939-954, 1982). The reason appears to be that mouth pressure swings are less than alveolar pressure swings. We measured TGVm and TGVes in six normal subjects and in nine asthmatic patients before and during bronchospasm, while panting at the same lung volume at 0.8 Hz (low), 2-2.5 Hz (medium), and 4.5-5 Hz (high). No difference was observed between TGVm and TGVes (P greater than 0.05) at any frequency (f) in normal subjects or asthmatics before bronchospasm. During bronchospasm, TGVm and TGVes were similar at low f. However, TGVm increased from 5.66 +/- 1.16 (SD) liters at low f to 6.50 +/- 1.71 liters at medium f (P less than 0.01), resulting in a TGVm 1.16 +/- 0.95 liters higher than TGVes (P less than 0.01). In three asthmatics during bronchospasm, mean TGVm-TGVes difference was 0.01 liter at low f, 0.26 liter at medium f, and 0.73 liter at high f. Surprisingly TGVes was in average 5% higher at low f than at medium or high f, both in normal subjects and asthmatics. A similar pattern was observed for TGVm, except in asthmatics during bronchospasm. We conclude that in airway obstruction overestimation of TGVm is frequency dependent and can be avoided by panting at low f. However, at this f TGV is 5% larger than at higher f, difference which is not related to airway obstruction.