THE FOOD AS A SOURCE OF ENERGY. 333 
nutrients (as ordinarily determined) even for a single species of 
animal. 
The results tabulated above, however, are amply sufficient to 
justify the statement on p. 279 that Rubner’s averages are not appli- 
cable to herbivorous animals, and that the metabolizable energy 
as computed with their aid is likely to vary widely from the truth. 
Indeed, since Rubner’s factor for fat (9.3 Cals. per gram) is 2.27 
times that for carbohydrates and protein (4.1 Cals. per gram) a 
computation of the metabolizable energy of feeding-stuffs or rations 
as it has not uncommonly been made simply gives a series of figures 
about 4.1 times as great as that obtained for total digestible matter 
when the digestible fat is reduced to its starch equivalent by multi- 
plication by 24. So far, then, as a comparison of one feeding-stuff 
or ration with another is concerned, this process adds no whit to our 
knowledege. It does, it is true, give some idea, albeit an inade- 
quate one, of the total amount of metabolizable energy present. As 
yet, however, our accurate knowledge of the energy requirements of 
domestic animals for various purposes is comparatively meager. 
If we base our computations on the feeging standards now current, 
we simply repeat with them the useless multiplication performed 
on the feeding-stuffs. On the other hand, if we take the results of 
such exact experiments on the metabolism of energy as are available, 
then, as the above results show, we shall be computing the energy 
requirements upon one basis and the energy supply upon a mate- 
rially different one. 
Significance of the Results.—A much more fundamental prob- 
lem than that raised in the foregoing paragraph confronts us when 
we come to reflect upon the general method by which it has been 
attempted to compute the metabolizable energy of nutrients, and 
to consider the real significance of the results. In so doing we may 
properly confine ourselves to the results upon cattle, those for horses 
and for swine being more or less fragmentary and uncertain. By 
far the larger proportion of the results above tabulated, as well 
as the most important of them, are based on experiments in 
which additions were made to a basal ration, the computation being 
by difference. As was pointed out in discussing the apparent 
metabolizable energy of the organic matter on previous pages, 
and as was specifically illustrated in the case of one experiment on 
