SOURCES OF KNOWLEDGE 207 



the ultra-practical feeding experiment in the direction 

 of an accurate history of what occurs wheii the animal 

 is eating a particular ration is the measurement of the 

 digested nutrients and the determination of the gain or 

 loss of nitrogen. This is accomplished, as heretofore 

 stated, by ascertaining the quantity of various com- 

 pounds eaten and the amount of the same in the feces, 

 the difference being the digested portion. The vaine is 

 also collected, and if the nitrogen in it is less or more than 

 that in the digested protein, then the animal is either 

 gaining or losing nitrogenous body substance, unless 

 the measurement is with a milch cow, when the nitrogen 

 in the mUk must be taken into account. By an experi- 

 ment conducted in this way, with careful and continued 

 weighings of the experimental animal, it is possible to 

 secure a probable relation between a unit of digested dry 

 matter and a unit of production. Such a method has been 

 used to determine what is a maintenance ration for ani- 

 mals of several classes, and in those cases where the 

 experiments have been continued for a sufficient length 

 of time and have shown on repetition a reasonable agree- 

 ment, we are justified in accepting the results as a close 

 approximation to fact. When a ration keeps an animal in 

 nitrogen equilibrium for one or more months and no 

 material gain or loss of weight occurs, we may safely 

 regard it as approximately a maintenance ration under 

 the conditions involved. Experiments of the same kind 

 are equally useful in testing the productive power of 

 various food combinations, and whenever by such con- 

 tinued tests one ration shows no superiority over another, 

 it is safe to assume that no differences exist which would 

 be especially important to the farmer's pocketbook. This 

 may be accepted as a business fact. 



