Since it was impossible to establish the complete measure of the income and outgo of each of these men in a respiration calorimeter of sufficient size to enable them to carry on physical activity comparable to their ordinary collegiate activities, it became necessary for us to rely in the last analysis upon the changes in body-weight as an index of the nutritive condition. For short periods nothing could be more erroneous than to follow this rule. The literature is full of illustrations of undue emphasis being placed on either considerable changes in body-weight or on slight changes incidental to alterations in diet. As is well known, since the body contains so large a proportion of water, a change of 1 or 2 kg. in body-weight may be in large part due to changes in water content and not to a change in the amount of organized body-tissue. Still in this research body-weights determined over a long period of time could fairly be taken as indices of the condition of nutrition. Since they were so important a factor in noting the effect of the reduced diet, it was increasingly necessary to take unusual precautions as to accuracy, regularity of record, and suitability of technique.

The body-weights were all taken in the early morning, with the nude subject in the post-absorptive condition and after the bladder had been emptied. This did not make allowance for undischarged fecal matter, but with the average individual these variations would not be very great, especially over a long period. One of our subjects, however, Pec, defecated regularly but once every 5 days; it is important to note that his body-weights may be affected by the presence or absence of a large amount of fecal matter in the colon. The subjects were weighed either upon a platform balance of unusual sensitivity (the so-called "silk scale" manufactured by the Howe Scale Company, with a capacity of 150 kg. and a sensitivity of 10 grams) or, as was done in the latter part of the test, upon an accurate and carefully calibrated spring platform balance. The latter balance, which is manufactured by John Chatillon and Sons, is designed especially for recording body-weights, and has a large dial which can easily be read to 0.25 kg.; a large number of body-weights were obtained with this balance. The records of the body-weight were checked by some member of the Laboratory staff. In no instance was an individual record made by the subject or an unchecked figure recorded.