It is well known that the total energy metabolism of animals falls during fasting (e.g. Tigerstedt, 1910;Lusk, 1931). On the other hand, the available data demonstrate much disagreement on the extent of the change in either the total or the basal metabolic rate (M.R. and B.M.R.) during fasting.If an animal is starved for a period it loses weight, and its metabolism can be expected to fall by an amount corresponding to (but not necessarily proportional to) the decrement of active body tissue. Where the total metabolism is measured, the fact that the animal is deprived of food reduces its heat production also by reducing its activity and the heat increment due to obtaining, eating, digesting and metabolizing food. While, however, the energy cost of eating is largely eliminated in complete deprivation of food (although refection and rumination may continue in some animals), the other components of the heat increment of feeding are not completely eliminated and may, in certain circumstances, increase. The question arises, therefore, whether the fall in total energy expenditure during fasting can be entirely accountedfor by the changes in metabolizing mass and foodintake, orwhether there is also a component accountable to fasting qua fasting; i.e. whether there is a true change in metabolic rate when changes in bodyweight and in' supra-basal 'energy costs are included in the reckoning.The analysis of this type of problem is always complicated by difficulty in taking body size into account. Body size is most conveniently measured as body weight, and that has been used in this work. Many other measures of body size have been suggested as reference bases for metabolic rate, but they have never been shown to be applicable to total metabolism. When body weight, or any function of body weight, is used as a measure of body size, then, if the decrease of each of the chemical components of body weight is in proportion to the corresponding concentration of each component in the normally fed animal, it would be reasonable to expect the decrements of body weight and of metabolizing potential to be proportional. If, however, the components alter in different proportions, then it is not strictly legitimate to equate the fall in metabolizing mass with fall in body weight. It is apparent that any resultant change in the manifest 15 PHYSIO. CLIV