One of the fundamentally important features of this research was to find the caloric intake requirement at the lower weight-level, with a view to determining whether, when the body-weight is reduced, the caloric needs are reduced proportionately, or less or more? If the body-weight could be held at the lower level for a considerable period of time, such as several weeks or months, an accurate measurement of the caloric intake would give a true measure of the caloric output or daily need. With practically all of the men in Squad A at certain times during the diet restriction the weight was maintained at a lower level with approximate constancy, but in no instance were these periods of weight maintenance of sufficient length to establish absolutely the true caloric needs. Although we were not able to hold the men at the lower weight-level for a sufficiently long time to determine definitely the caloric needs, further evidence may be obtained from the results of the gaseous metabolism experiments, which will be considered in a succeeding chapter. The period of diet restriction with Squad B was too short to establish a lower weight level for these men.

Reference has already been made, particularly in the discussion of the body-weight charts for Squad A, to the caloric intake at the different periods of the experiment. These caloric intakes have been given in actual figures in the energy tables for the individual subjects. Very considerable variations in the calories for each day were observed, which were due not only to irregularity in the character and amounts of food served from day to day in the dining hall but to the unrestricted days, i. e., the uncontrolled Sundays. These latter were entirely beyond our control, other than as previously outlined at several points in this discussion. In preparing the estimate of the average caloric intake, however, we have invariably included the estimates based upon the reports of the men of the quantities of food eaten on the unrestricted Sundays, as given in table 34. (See p. 270.) It was necessary in our final tables to assume that on the first unrestricted Sunday, October 14, the quantities taken were exactly those on the second unrestricted Sunday, the estimates of which were furnished by the men. No information is at hand in regard to the energy intakes during the Thanksgiving and Christmas recesses. Hence a break must occur in the caloric intake shown on the body-weight charts at these points.

An analysis of the relationship between the caloric intake and the body-weight, and particularly the body-weight at maintenance level, may be made either from the tables showing the balance between the income and outgo, or from the body-weight charts. It is perhaps somewhat more satisfactory to study the subject graphically from the body-weight curves by noting the heights of the various blocks which correspond to the caloric intake. (See figs. 57 to 68, pp. 210 to 222.) In practically every instance the caloric intake was increased somewhat in the second week in December; in general the height of the block at this time and the height of the block for the last few weeks of the experiment are not far from the same. In other words, we have two periods which represent a fairly close approximation to the caloric intake for weight maintenance, i. e., when the body-weight was either constant or not materially increasing or decreasing. At neither of these times had we the perfect control desired. Nevertheless, these two independent periods, some six weeks apart, give fairly good evidence of the probable maintenance requirement of these men at the lower weight-level.

If we examine the body-weight chart for Bro, (fig. 57, p. 210), we find that the initial requirement in the early part of this test on the uncontrolled days was somewhat over 3,000 calories. The calories here, as well as on the other charts, refer only to the net calories, i. e., calories of food less those of feces and urine. On October 4 a diet restriction took place with a fall in energy intake to about 2,200 calories. Further reductions were made but in the early part of December the energy was increased. This increase was determined, not by calculating the number of calories beforehand, but simply by a gradual increase of the diet during this period until the body-weight had become constant. Exactly the same procedure was carried out in the latter part of January, but the assistant in charge of the apportionment at no time determined the exact caloric intake of the food. Thus, both of these levels were adjusted without a previous knowledge of the caloric requirement. This holds true for all of the subjects. On the return of the men to college in January, all of the subjects received a low diet for a short time, to compensate for the increase in body-weight during their absence. With Bro we find that the average of the period of maintenance diet in December and January is not far from 2,000 calories. Reference to the individual balance-tables (tables 46 to 58) confirms this. To make the details still clearer, an abstract is given in table 35 of the principal data in the several balance-tables, grouped with regard to the several periods of diet ingestion. Reference to the actual energy available to the body with Bro during the period from December 3 to 20 shows that he had 2,091 calories, and from January 16 to February 3, 1,931 calories, making an average for the two periods of 2,011 calories per day. Hence, we may argue that Bro at the lower weight-level required 2,000 net calories.