284 THE FEEDING OF ANIMALS 



feeds, but Hills showed again that if Woll had left out 

 rye bran these prices would be greatly changed. It 

 appears that varying individual judgments as to the list 

 of feeds which shall determine values may cause absurd 

 differences in the calculated market cost of the nutrients, 

 and introducing into the list or withdrawing from it a 

 comparatively unimportant feeding-stuff may lower or 

 raise the price of one nutrient even one-half. 



A still more serious difficulty arises from the fact 

 that often when an apparently typical and proper list 

 of feeds is used from which to calculate prices, the 

 use of the method of least squares results in giving a 

 negative value to one of the nutrients. In several cases 

 of this kind the fat was shown to be worth less than 

 nothing, a most absurd conclusion. This mathematical 

 method is, therefore, not available for the valuation 

 of feeding-stuffs, and so far no mathematician has offered 

 one that is. 



380. Physiological values. — ^We are left now to inquire 

 whether we may not use physiological values, in other 

 words the work which a nutrient wiU perform in the ani- 

 mal body, as a starting-point from which to calculate 

 relative values. If, for instance, it could be demon- 

 strated that protein has a fixed physiological value twice, 

 and fats three times, that of carbohydrates, it would 

 then be a very simple matter to ascertain what propor- 

 tion of the cost of a ton of cottonseed meal should be 

 applied to each class of nutrients. To illustrate, a ton of 

 high-grade cottonseed meal contains about 590 pounds 

 of carbohydrates, 860 pounds of protein, and 260 poimds 

 of fat. If these ingredients are assumed to have a ratio 

 of value of 1, 2, and 3, then the whole would be equiva- 

 lent to 3,090 units of carbohydrates, the cost of one unit 



