THE BALANCE OF NUTRITION 243 



edge as possible of the functions of each separate machine 

 and of the changes undergone by the materials submitted to 

 its action. 



It may fairly be presumed, however, that these critics, even 

 with the fullest knowledge of the technical details of such a 

 factory, would hardly undertake to conduct it as a business 

 enterprise without keeping account of the stock purchased and 

 the output realized, i.e., exactly the sort of bookkeeping which 

 the balance experiment attempts for the animal body. The 

 truth is that both types of investigation are equally necessary 

 and each aids in the interpretation of the other. The balance 

 experiment has been especially prominent in the past, while at 

 present attention is being directed to a greater extent to in- 

 vestigations of the intermediary metabolism, but neither can 

 say to the other " I have no need of thee." 



331. The balance experiment in agricultural investigations. 

 As already indicated, the methods of the balance experiment 

 have been quite largely applied in agricultural investigations. 

 Such investigations have been made, in the majority of cases, 

 not with single chemical compounds, but with feeding stuffs 

 or rations as a whole, and the effect observed in such an ex- 

 periment is obviously a summation of the effects of all the 

 ingredients contained in the feed consumed. The result, there- 

 fore, while entirely adequate to determine the total nutritive 

 effect of the particular material experimented with is less capable 

 of generalization than one obtained with a single chemical 

 compound like starch or fat, and from this point of view may 

 even be regarded as being in a sense empirical. A compre- 

 hensive knowledge of the nutritive value of a feeding stuff im- 

 plies, first, a determination of the kinds and amounts of chemical 

 compounds contained in it and, second, a determination of the 

 exact physiological functions of each. Obviously, however, 

 such a complete determination of the nutritive value of any 

 considerable number of feeding stuffs is a work requiring a vast 

 expenditure of time and labor. One justification, therefore, for 

 the " short-cut " method of determining summarily by a. bal- 

 ance experiment the effect of a feeding stuff or ration is that it 

 appears possible to secure in this way within a reasonable time 

 data which can be put to practical use in the comparison of 

 feeding stuffs and rations. Moreover, it is to be anticipated 



