This study compared three indirect calorimetry determinations, as kcals/minute, over three consecutive days on 28 healthy, sedentary women of varying body composition. No significant within-individual variation for VO2, CO2, respiratory quotient (RQ), or Kcals/minute was found among the three days. A low coefficient of variation (3.4 +/- 3%) and a relatively small standard deviation in mean Kcals/day (1383 +/- 214) suggests possibly one or only a few measures are necessary for predicting resting metabolic rate (RMR) within a range applicable for clinical use. There were strong correlations of body weight and body composition variables (fat and LBM) with RMR. Knowing both LBM and fat mass increased the ability to predict RMR significantly over the prediction with either variable alone. The Harris Benedict equation over-predicted RMR by 11.1% compared to RMR measured by indirect calorimetry. When equations are based on body weight, rather than LBM, metabolic rate may be over-predicted in obese populations.