Changes in functional residual capacity and their relationship to alveolararterial oxygen difference were observed during the course of respiratory failure in twelve patients. The influence of recumbency and oxygen breathing on frc in normal subjects was also examined. Low resting lung volumes and large alveolar-arterial oxygen differences were often observed at the time of admission in respiratory failure. The reduction in frc cannot be explained on the basis that these measurements were made in recumbent patients breathing a high concentration of oxygen in the inspired gas. An increase in frc was often associated with an improvement in oxygenation, and it is inferred that in these patients correction of frc is associated with the opening up of previously non-ventilated alveoli, and that therapy designed with this objective is therefore justified. However, when extremely low resting lung volume is due to causes not reversible mechanically, such as diffuse pneumonia, large tidal volume ventilation determined on a body weight basis may not be indicated. Measurement of frc is not recommended for routine clinical management.