302 PRINCIPLES OF ANIMAL NUTRITION. 
the averages of the tables are of less significance than. the results 
given in the following pages, where the digestible nutrients are 
taade the basis of the computation. 
ENERGY OF DIGESTIBLE NUTRIENTS. 
The foregoing paragraphs have dealt with the apparent 
metabolizable energy of feeding-stuffs, and the results have 
been expressed in terms of total or of digestible organic matter, 
or as percentages of gross energy. We now turn to a con- 
sideration of such data as are available regarding the several con- 
ventional groups of nutrients into which the food of herbivorous 
animals is ordinarily divided and inquire whether it is possible to 
compute average factors for- their metabolizable energy which 
shall be useful in themselves and be of value particularly for pur- 
poses of comparison with earlier experiments. This was the special 
purpose of Kellner’s investigations, and his experiments supply 
valuable data on-these points as regards cattle and ‘presumably 
other ruminants, which may be supplemented to a certain extent 
from experiments by other investigators upon horses and swine. 
In considering the experiments from this standpoint, Kellner’s 
discussion and methods of computation have been closely followed, 
the attempt being made to compute as accurately as possible the 
real metabolizable energy of the several nutrients. 
Gross Energy. 
If it were possible to add pure nutrients to a basal ration and 
be sure that they would have no effect upon the utilization of the 
latter, it would be a comparatively simple matter to determine their 
real metabolizable energy. As a matter of fact, however, as has 
been seen, this is not possible. Not only is it impracticable to secure 
large quantities of pure nutrients, but each such addition to the basal 
ration is liable to affect especially the digestibility of the latter. 
Consequently the difference in metabolizable energy between the 
two rations fails to represent correctly the real metabolizable energy 
of the nutrient added. In order to compute the latter we must 
have a basis for correcting the results for the small variations in the 
amounts of other nutrients digested, and for this purpose we need 
to know the total or gross energy of the digested matters. 
