The usefulness of lung-thorax compliance (or elastance) as an index of pulmonary compliance (or elastance) was examined in 15 patients being ventilated for acute respiratory failure. Mean lung-thorax elastance (ELT) was 27.9 +/- 2.6 cm H2O/L, and the chest wall accounted for 34 +/- 2 percent of the mean total value. Changes in ELT caused by increments of positive end-expiratory pressure correlated only with changes in pulmonary elastance (r = 0.96; P less than 0.001) and not with chest wall elastance, although individual patients varied as to the contribution of the chest wall component. Lung-thorax elastance increased in direct proportion (1:1) to increases in pulmonary elastance, whereas the changes in lung-thorax compliance were only half those in pulmonary compliance. We conclude that elastance is a more useful clinical index than compliance.