SOURCES OF KNOWLEDGE 205 



animals has gained more or less pounds of weight than 

 another set, but the why and the how are not explained. 



Practically the same considerations pertain to feed- 

 ing tests for milk production. When the milk flow from 

 one ration is larger than from another, we can easily 

 satisfy ourselves as to the comparative yield of milk 

 solids, which is the real test of such production; but we 

 are not able to decide whether the cow either may not 

 have contributed to the milk secretion from the substance 

 of her own body, or may not have gained in body sub- 

 stance, the extent of such loss or gain being greater, 

 perhaps, with one ration than with another. 



Even if these uncertainties did not exist, we have 

 the still greater disadvantage of not learning by this 

 means why a particular combination of feeds has superior 

 qualities for causing growth or sustaining milk secretion. 

 The mere data showing that an animal ate so many 

 pounds of food and produced so many pounds of beef or 

 milk are important business facts, but they reveal noth- 

 ing concerning the uses of the several classes of nutrients 

 and of themselves furnish slight basis for developing a 

 rational system of feeding. We must somehow learn the 

 function of protein, carbohydrates, and fats in main- 

 taining the various classes of animals and the real effect of 

 varying the source, quantity, and relative proportions of 

 these nutrients before we can draw safe general conclusions. 



292. Chemical and physiological studies. As pre- 

 liminary to more comprehensive and convincing methods 

 of investigating feeding problems, there has been going 

 on during many years a necessary study of the compounds 

 which are found in plants and animals. Much has been 

 learned about the ultimate composition and the consti- 

 tution of the proteins, carbohydrates, and fats, their 



