Measurements of energy expenditures for domestic animals are used as a basis to determine the total energy requirements for maintenance and for production. Most studies on energy metabolism have been conducted under controlled, confinement conditions, and the results extrapolated to free - ranging animals. Such approaches do not always represent the real energy costs of free existence in the range or pasture environment. This study compared energy expenditures of sheep and goats under free-grazing conditions and assessed the accuracy of the carbon dioxide entry rate technique (CERT) as compared to the oxygen consumption method. In addition, 24-hour activity budgets were used to construct energy budgets for both animal species. These estimates were compared to CERT measurements. Dietary organic matter, crude protein, gross energy, digestible protein, and digestible energy intakes, as well as the apparent digestibility coefficients for crude protein and gross energy were compared for both species in one of the grazing trials. The validation of CERT yielded the linear regression equation: Y = 0.878 + 5.333 ER where Y energy expenditure in Kcal��min-1, and ER is the C02 entry rate in grams of C02 carbon��min-1. The coefficient of determination (r2) was 0.979, and the residual standard deviation of �� 0.12 Kcal��min-1. Daily energy expenditure of goats was higher (P High ambient temperatures during one of the grazing trials apparently caused heat stress to the grazing animals. Both species responded to this situation by using behavioral adaptations and avoidance mechanisms. The energy budget method over-estimated energy expenditures of sheep by 9 percent, and grossly underestimated the energy expenditures of goats by 39 percent. The nutritive value of the diets selected by both species was similar . However, goats had higher (P< .OS) organic matter, crude protein, gross energy, digestible protein and digestible energy intakes. The apparent crude protein digestibility coefficient for goats was 20 percent higher (P CERT can predict energy expenditures within �� 8 percent of the mean. It is a feasible technique for assessing energy expenditures of small ruminants under range conditions. The energy budget method was not reliable for estimating energy costs in small ruminants, especially goats.