ANIMAL NUTRITION DISCUSSION AT DUNDEE 427 



show some definite relation between these units and increase. 

 Some of Ingle's graphs do show a certain measure of correlation 

 but viewing the whole display in light of the law of error the 

 "Starch Equivalent" and "Digestible Protein" are correlated 

 with such widely different feeding effects that they must stand 

 for different things in the different rations. 



Further, the conclusions drawn from such correlation as 

 exists in these graphs do not confirm Kellner's standard 

 rations. They indicate a lower protein requirement and show 

 no correlation whatever between " Albuminoid Ratio " and 

 increase in cattle or sheep. 



The heterogeneous nature of all the analytical units on which 

 scientific feeding values are based seems a sufficient reason for 

 these failures. 



An ordinary analysis gives : 



(1) "Protein" or "Albuminoids." Every animal ration, to 

 be efficient, must contain a certain minimum of certain proteins. 

 But a minimum of specific proteins is a very different thing 

 from a minimum of insoluble and precipitated nitrogen multiplied 

 by six and a quarter. Even in a mixed ration, an animal may 

 have to consume a great superfluity of other proteins before 

 it obtains the necessary amount of specific proteins. The 

 various amounts of protein recommended by different authorities 

 and the great divergences in Ingle's charts indicate that this 

 difference in effect of a unit of mixed proteins frequently occurs 

 in practice. 



(2) "Amides, etc.," account for a large proportion of the 

 nitrogen in home-grown foods. Theoretically different quantities 

 and kinds of concentrated food would have to be added to 

 turnips and straw according to the method adopted in evalu- 

 ating this group of varying composition and unknown function. 



(3) " Ether Extract " is also a varying mixture. In young 

 grass only about 35 per cent, is fat ' and all oil is not 

 linseed oil. 



(4) " Carbohydrates " and " Fibre " are equally heterogeneous. 

 The "Fibre" of undecorticated cotton cake and of swedes and 

 the " Carbohydrates " of maize and wheat straw are, so far as 

 our analyses go, the same things. We are therefore entirely 

 dependent on average " digestibility " and " value " factors and in 

 the case of the foods most largely used here the " probable error " 



1 Highland and Agric. Society Trans. 1889, p. 44. 



