Although the body-weight over considerable periods of time may properly be taken as an index of nutritional level, or at least of varia-tions in nutritional level, it of itself gives no idea as to the exact caloric requirements. It seemed imperative, therefore, to measure quantitatively the caloric output of these men at the different nutritional or different weight-levels. The ideal apparatus for measuring the caloric output is the respiration calorimeter, but its use was precluded, first, on account of the impossibility of transporting this complicated apparatus to Springfield, and second, because of the length of time required for the several experiments. Consequently the determination of the gaseous exchange as frequently as possible throughout the entire period of dietetic control seemed to be the best method for obtaining information regarding the caloric requirements.

As stated in the chapter on technique (see p. 82), the gaseous metabolism of nine of the men was determined nearly every morning in Springfield, these measurements including observations of the oxygen consumption and in many cases of the carbon-dioxide production. In addition, the carbon-dioxide output of the squad as a group was measured biweekly during sleep in a large respiration chamber in Boston. Since in the Springfield measurements, the subjects were always in the post-absorptive condition - that is, without food for at least 12 hours - we have clear records of the basal metabolism. The experiments made in Boston during the night were not strictly basal in that they did not conform to the fullest demand for the post-absorptive condition, since the experiment began less than 12 hours after the preceding meal. On the other hand, the measurements were made for the most part during deep sleep and after a very light supper and in all probability they represent the basal demand.

It was extremely important to determine the basal metabolism prior to the dietetic restriction and thus establish the normality of these men and find whether the variations in metabolism were within the so-called normal limits. As has been frequently pointed out in this monograph, the period for the measurement of basal values during normal diet was too short with all factors studied in this research. Since there was a possibility that it might be necessary to conclude the observations at Christmas, we were obliged to choose between sacrificing the total length of the experiment by prolonging the period of normal diet, or shortening this period so as to obtain extended measurements with restricted diet. The second alternative seemed to us the more desirable; consequently, the values during normal diet for these men were determined on a relatively few days in the latter part of September and during the first few days of October.