Energy expenditure (EE) was determined, using an open-flow indirect calorimetry system in a group of 20 clinically normal, apparently resting, client-owned dogs. Five evaluations were performed over an 8-hour period to determine reliability of the method. The intraclass correlation coefficient was calculated as the ratio of within- and between-subject variances, using repeated-measures ANOVA. When only the middle 3 evaluations were included, the intraclass correlation coefficient was 0.87, indicating good reliability. The first evaluation was higher than the subsequent 4 evaluations for rate of O2 consumption (Vo2/kg and Vo2/kg0.75; (P < or = 0.01), and EE/kg and EE/kg0.75 (P < or = 0.005). The respiratory quotients at the first (P = 0.004) and second (P = 0.013) evaluations were different from the respiratory quotient at the fourth evaluation. Therefore, the first evaluation may not be representative of the actual EE. The mean value of at least 3 subsequent evaluations after an adequate adaptation period (5 to 10 minutes) to the equipment will be useful for predicting energy requirements of apparently resting, clinically normal dogs.