The single breath method underestimates total lung capacity by comparison with the multiple breath method (TLCmb) because of inhomogeneity of ventilation distribution. This study proposes a simple correction for the single breath TLC (TLCsb), using inert gas phase III slope to account for the effects of uneven ventilation distribution. A model of a non-uniform lung ventilation was designed, composed of a serial dead space and two alveolar compartments arranged in parallel, whose relative ventilations were determined from the phase III plateau. Before correction TLCsb was 104-44% of TLCmb in 64 subjects (17 with diffuse interstitial disease, 42 with chronic obstructive pulmonary disease, and five healthy subjects). The limit of acceptability for the correction (TLCcorr) was determined from the 95% confidence interval of TLCsb/TLCmb in the healthy subjects. The correction resulted in a significant increase in TLCsb (p less than 0.004). TLCcorr remained under the limit of acceptability for only 12 patients with emphysema, and all 12 showed a large improvement in the TLC estimate. The presence of poorly ventilated zones during a single breath in these patients may explain this partial correction.