We believe, however, that Grafe has made a fundamental error in assuming that the heat output of his dog outside the respiration chamber was the same as in the respiration chamber, since his own reports show that the respiration experiments were made at 22° C, while a very large part of the time the dog was in the cellar with a temperature of 15° C, and probably during the night with even a lower temperature. While exact quantitative data regarding the influence upon metabolism of temperature is still too scanty for most accurate computations, it has been the custom of physiologists for many years to use the figure obtained by Rubner, which implies an increase in metabolism of approximately 3 per cent per degree centigrade.1 We thus have a probable increase in metabolism of this animal on the days outside the respiration chamber of not far from 21 per cent or, roughly speaking, one-fifth. With a basal daily requirement of 1,200 calories, this would correspond to 240 calories per day. The four periods when excess food was given numbered 66 days, of which 12 were in the chamber. With an increase in the metabolism of 240 calories per day, there would be a total increase in the metabolism of 240 X 54 = 12,960 calories, or one-third of the excess computed by the above method. The excess calories, which amounted to 39,334 - 12,960 = 26,374, if stored as fat, would correspond to about 3 kg. of fat or 15 per cent of the body-weight of the dog. It should be stated, however, that in a recent investigation by Zuntz2 a series of experiments carried out with a dog at a temperature of 30° to 32° C. showed 44.4 calories per kilogram per 24 hours and another series at 15° to 16° C. showed 50.2 calories. This difference corresponds more nearly to 1 per cent per degree than to Rub-ner's figure of 3 per cent used in the above calculations.3

1 Benedict and Carpenter, Carnegie Inst. Wash. Pub. No. 261, 1918, p. 72.

2 Grafe had good precedent for using the fasting value for a base line as in similar comparisons.

Johansson, Landergren, Sonden, and Tigerstedt (Skand. Arch. f. Physiol., 1897, 7, p. 29) had previously used values obtained with a man during a fast of 5 days.

Since in our judgment the figures used in our recalculations for the basal metabolism and the estimates of probable energy requirements during repose are minimum rather than maximum, we believe that the experiment of Grafe does not positively prove the main point of his discussion, namely, that the dog, when given excess food, produced a sufficiently excess metabolism to counterbalance it.

A striking factor of Grafe's experiment must not be lost sight of in that throughout the entire period of realimentation there was a very great storage of nitrogen in the body. This storage, unaccompanied by a marked increase in metabolism, is, we believe, fully in conformity with the experience found by A. Muller4 with man, in which 210 grams of nitrogen were stored without appreciable increase in the heat production per kilogram of body-weight. Although personal conversation with Dr. Grafe gives no clue as to the probability of an increased activity on the part of the dog when outside the chamber, it would appear as if the estimates for the energy requirement on the days when the animal was not in the chamber are too low. If this is the case and if correction is made for the erroneous basal value obtained after a period of prolonged inanition, we believe that the figures would not positively prove Grafe's main contention. The addition of a large amount of protein and certainly of some fat can hardly be accounted for wholly by the character of the material added during the first 7 days of feeding. Although the body-weight regained its original level, a computation of the caloric output of the animal and the character of the food shows that the organized tissue could not have been replaced during this period.

1 Rubner, Die Geaetze des Energieverbrauchs bei der Ernabrung, Leipsic, 1902. Hari (Biochem. Zeitachr., 1914, 66, p. 2) has pointed out that in many of Rubner's experiments the percentage difference per degree is 3 to 5 per cent.

2 Zuntz, Biochem. Zeitschr., 1913, 55, p. 341.

3This difference is at least in part explained by the fact that 32° C. is considerably above the so-called critical temperature for most dogs.

4 A. Muller, Zentralbl. f. d. ges. Physiol, u. Pathol, d. Stoffw., 1911, N. F., 6, p. 617. See also p. 29 of this monograph.

The values given by Grafe of the metabolism on hunger days during excess feeding in the third period show an average value of 1,030 calories. This is not far from that found on the first hunger day, and does not indicate an increase in the basal metabolism during this period.

In consideration of the strong probability that during the first 7 days with food the dog had not regained the original pre-fasting nutritional state, that the basal value selected by Grafe should have been that at the beginning rather than at the end of the 21-day hunger period, and that the metabolism in the cellar at 15° C. would be somewhat greater than in the respiration chamber at 22° C, we believe that one of Grafe's most striking series of experiments fails to prove his point.

Dengler And Mayer, 1906

Two researches from von Noorden's clinic bear upon the relationship of the storage of nitrogen, increase in body-weight following excessive nitrogen ingestion, and the basal metabolism. Dengler and Mayer,1 using a Zuntz-Geppert apparatus and a subject unfortunately not normal, made a 3-months' experiment with a nitrogen-rich diet. They found that in spite of an increase in body-weight of 13 kg. and a nitrogen storage of 371 grams, the oxygen consumption rose very slowly from 222 to 242 c.c. per minute; per kilogram of body-weight there was, if anything, a slight fall. The authors interpret this as signifying that the addition of nitrogen, even in this very large amount, could not have been in the form of active protoplasmic tissue.