53° PRINCIPLES OF ANIMAL NUTRITION. 
pare p. 539) at 100,000 kgm., making the total additional work 
538,712 kgm. Kellner compares this difference with the increased 
amount of nitrogen-free extract digested, 613.79 grams, neglecting 
the small differences in the other nutrients. As corrected in a later 
publication,* the results are as follows: 
613.79 germs. starch = 2527 .601 Cals. = 1,071,698 kgm. 
538,712 + 1,071,698 = 50.27 per cent. 
If we base the calculation upon the difference in total organic 
matter digested, the percentage will of course be somewhat smaller. 
It was discovered later that the indications of the dynamometer 
used in these experiments and many subsequent ones were untrust- 
worthy, so that no value attaches to the percentage computed above, 
but it serves just as well to illustrate the method employed, and 
which was followed in the whole series of experiments. In brief, 
the attempt is to find in the indications of live weight and urinary 
nitrogen a partial substitute for the determination of the respira- 
tory products. As Kellner and Wolff do not fail to point out, the 
results are but approximations, and in any single experiment may 
vary considerably from the truth, but on the average of a large 
number of experiments it was hoped that satisfactory results might 
be reached. In later experiments rather more importance seems 
to be attached to the effects upon live weight than to those upon 
urinary nitrogen, but it should be noted that the live weight showed 
remarkably small variations from day to day, under the carefully 
regulated conditions of the experiments, and was quite sensitive 
to changes in the amount of work done. 
The experiments may be conveniently divided into three groups. 
The first of these | includes the years 1877 to 1886, inclusive, in which 
the work done was compared with the total digested food. The 
second { covers the experiments of 1886-1891, in which the digested 
crude fiber was omitted in computing the work-equivalent of the 
food, while the third group § includes the experiments of 1891-1894 
with a new and more accurate form of dynamometer. 
* Wolff, Grundlagen, etc., p. 89. 
+ Grundlagen fiir die rationelle Futterung des Pferdes, 1886, 66-155; Neue 
Beitriige, Landw. Jahrb., 16, Supp. III, 1-48. 
t Landw. Jahrb., 16, Supp. III, 49-131, and 24, 125-192. 
§ Ibid., 24, 193-271. 
