Reply to Joseph.

TO THE EDITOR: Many thanks to Dr. Joseph (1) for requesting clarification of methods used to measure metabolic rate of the rodent in our recent study (3). Measurements of V̇O2 and V̇CO2 in the rat do indeed require a great deal of precision, given that the differences between inspired and expired fractions of O2 and CO2 are very small. Dr. Joseph’s letter highlights some very important potential sources of error in these measurements; however, we believe that several features of our flowthrough system minimize these threats to accuracy (4). First, a single set of O2 and CO2 analyzers sample the inspired and expired air in sequential fashion. Thus there can be no differences in linearity of the sensors. Each day, a three-point calibration of these analyzers is performed using gases that bracket the range of inspired and expired O2 and CO2 that we expect under our experimental conditions based on pilot testing. The calibration routine provides us with daily assessments of linearity and drift of the two sensors. In the set of calibrations represented in our study, drift of the O2 sensor averaged 0.002, 0.016, and 0.003%, respectively, for the three calibration gases. The corresponding drift of the CO2 sensor averaged 0.002, 0.014, and 0.000%. Finally, in our system, inspired air flows directly, via a blender, from tanks of air and nitrogen to the O2 and CO2 analyzers before humidification and entry into the plethysmograph. Expired gas is dried by a dessicant prior to analysis. Thus sampling of both inspired and expired air is performed on dry gas only, which we deemed sufficient for quantifying changes in V̇O2 and V̇CO2 under our varying experimental conditions. However, as detailed by Lighton and Halsey (2), for estimation of absolute values of V̇O2 and V̇CO2 it is likely more accurate to incorporate measures of water vapor in both inspired and expired gases.