THE FOOD AS A SOURCE OF ENERGY. 279 
similar to that of the proteids of the grains). Correcting these 
values for the error involved in the usual computation of protein 
from nitrogen, he obtains as the average metabolizable energy of 
the protein (N X 6.25) of a mixed diet 4.1 Cals. per gram. 
For the fat and carbohydrates it is assumed that all their poten- 
tial energy is metabolizable, but an allowance is made in the latter 
case for the error due to the ordinary computation of the carbo- 
hydrates by difference and for some minor sources of uncertainty. 
Rubner’s final averages are— 
Protein (N X6.25)......... 4.1 Cals. per gram. 
Pati cacciiicatedetauatedd Q.3: fe oe 
Carbohydrates... .......... 7 a 
The value for protein, by the method of computation, includes 
an allowance for the metabolic products contained in, the feces, but 
neither it nor the values for the other nutrients include any estimate 
for the loss through imperfect digestion. In other words, they 
refer to the digested nutrients. 
These figures were designed expressly for computing the metab- 
olizable energy of human dietaries, and even for that purpose are 
confessedly only approximations. In the absence of more exac 
figures, however, they have been somewhat extensively used for 
computing the metabolizable energy of the digested portion of the 
food of domestic animals. For purposes of approximate estimates 
such a use of them was perhaps justifiable, but in too many cases 
their origin seems to have been forgotten and a degree of accuracy 
ascribed to them which they do not possess. As will be shown 
presently, later investigations have yielded materially different 
results for the metabolizable energy of the several classes of nutri- 
ents in the fuod of herbivorous animals. 
Later Experiments.—Quite recently Rubner * has published the 
results of some experimental investigations into the validity of the 
averages or “standard figures” given above. In these experiments 
the weights and heats of combustion of food, feces, and urine were 
determined calorimetrically and the metabolizable energy as ob- 
tained from these data was compared with that computed by the use 
of the above factors. In making the latter calculation an allowance 
* Zeit. f. Biol., 42, 261. 
