In 1923 Gamble, Ross, and Tisdall (1) presented data correlating the fixed base and water balance of two fasting epileptic children. In these observations, the water balance was represented as the difference between total body weight loss and the sum of the protein and fat loss, the latter calculated from the urinary nitrogen and ketones. In 1929 Newburgh (2) and his coworkers described a method for the calculation of the water exchange of normal adults which depended on a knowledge of the total caloric expenditure and the composition of the metabolic mixture. The limitations of this so-called metabolic method have been discussed in detail for adults by Lavietes (3), Peters (4), and Newburgh et al. (5) and for infants by Levine et al. (6). In 1935, Lavietes, D'Esopo, and Harrison (7) proposed a method for estimating water balance from base balance and changes in concentration of base in serum. This method was suggested by the earlier observations of Gamble and his coworkers (1) that water and base are lost from the body in approximately the proportions in which they appear in body fluids. In recent years the " electrolyte " method for measuring water exchange has been applied, particularly by Darrow, Yannet, and Harrison (8) and Hastings and his coworkers (9), to quantitative studies of the relative distribution of intracellular and extracellular water.Both the "metabolic" and the " electrolyte" methods of measuring total water exchange are open to criticism on theoretical grounds since the former assumes that the respiratory exchange is a reliable index of the composition of the metabolic mixture and the latter, that certain electrolytes are retained solely as extracellular or intra-1 Respiratory Metabolism in Infancy and Childhood,