The energy expenditure of free-living animals has been studied extensively by the doubly-labeled water (DLW) technique. This method provides a reasonably accurate estimate of daily energy needs. However, there is considerable interest in the energy demands of animals over much shorter timescales, for which the DLW technique is less useful. We examined the possibility of measuring the expenditure of small animals over these shorter timescales from the washout kinetics of a bolus dose of 13C labeled bicarbonate. The study involved 19 laboratory mice which were injected either i.p. or s.c. with 0.2 ml of 13C labeled bicarbonate in water. Mice were placed in a standard respirometry system, maintained at different temperatures to precipitate a 3 fold variation in metabolism. Samples of breath were collected from the chamber into vacutainers at one minute intervals for approximately 40 minutes to an hour. Samples were analyzed by admission to a mass spectrometer (VG Optima) via a GC interface which identified and admitted the CO2 peak. The log converted isotope elimination was linear (r2 > 98% in all cases) indicating a single pool was involved. We evaluated the pool size from a dilution series of the injectate in equilibrium with CO2 gas. Conventional compartmental analysis produced an estimate which on average across the 19 individuals provided a reasonable estimate of the CO2 production. Individual estimates were however imprecise and the overall correlation between isotope and calorimeter estimates had an r2 of only 15%. Reasons for this discrepancy are unclear. Nevertheless an empirical model, using the elimination gradient, pool size and route of isotope administration as predictors explained 86% of the variation in CO2 production. Elimination of a bolus dose of 13C labeled bicarbonate provides a useful tool for estimating the energy metabolism of mice over intervals between 15 and 40 minutes.