Conventional gas-exchange instruments are confined to the measurement of O(2) consumption (VO(2)) and CO(2) production (VCO(2)) and are subject to a variety of errors. This handicaps the performance of these devices at inspired O(2) fraction (FI(O(2))) > 0.40 and limits their applicability to indirect calorimetry only. We describe a device based on the automation of the Douglas bag technique that is capable of making continuous gas-exchange measurements of multiple species over a broad range of experimental conditions. This system is validated by using a quantitative methanol-burning lung model modified to provide reproducible (13)CO(2) production. The average error for VO(2) and VCO(2) over the FI(O(2)) range of 0.21-0.8. is 2.4 and 0.8%, respectively. The instrument is capable of determining the differential atom% volume of known references of (13)CO(2) to within 3.4%. This device reduces the sources of error that thwart other instruments at FI(O(2)) > 0. 40 and demonstrates the capacity to explore other expressions of metabolic activity in exhaled gases related to the excretion of (13)CO(2).