At rest, especially during starvation, protein (via hepatic gluconeogenesis and/or direct oxidation) contributes significantly to the body's total metabolic requirement. During exercise, carbohydrate (CHO) and free fatty acids (FFA) are generally regarded as the major energy sources. Although protein has the capability, it has seldom been considered as an important exercise energy source based largely on 24-h urinary nitrogen (N) excretion values. Contrary to classical belief, recent determinations of N in sweat suggest that total N excretion is in fact increased with exercise. This should not be surprising when one considers that prolonged exercise substrate shifts and hormonal changes are in many ways analogous to the situation that occurs during starvation. In addition, in vitro experiments have demonstrated that skeletal muscle, as well as liver, has the capacity to oxidize a number of amino acids. This information, together with the in vivo exercise observations of an increased alanine output from skeletal muscle and an increased 14CO2 evolution following [14C] leucine injection or ingestion, indicates that some amino acids contribute to exercise energy needs. Although a number of mechanisms have been suggested, the current evidence strongly suggests that the branched-chain amino acids (glucose-alanine cycle) are the most important. In conclusion, it appears that the earlier studies which indicated that protein/ amino acid utilization did not increase during exercise may have been premature. Although clearly not as important as either CHO or FFA, recent investigations employing both humans and laboratory animals suggest that protein/amino acids, under some conditions, may contribute significantly to total exercise calories.