PROTEIN REQUIREMENTS OF CATTLE: MITCHELL 53 



duction is nioasured by the protein content of the milk produced. This 

 fact was first recooiiized by Haecker, who says in the bulletin describ- 

 ing his fuiulaniental investigations in milk production (^^) : 



During the time when the feeding experiments in milk-production in review 

 were in progress, it occun-ed to the writer that in order to determine the actual 

 net nutrients required to produce a given animal product, the composition of 

 the product should be known, as well as the composition and the available nutrients 

 in the food which is to be fed for its production, so that the nutrients in the 

 ration might be provided in the proportions needed by the animal. Before a 

 builder bids on a contract, he determines the quantity needed of each of the 

 materials that are to appear in the structure. Without such specifications he 

 would not know how much of each of the different materials would have to be 

 provided. 



The fat-to-protein ratio in mill-. — It is well known that cows' milk is 

 subject to wide variations in composition, depending upon the breed and 

 individuality of the animal, the stage of lactation, the plane of nutrition, 

 the character of the feed consumed, and possibly other factors. The fat 

 content of milk appears to be the most readily affected by these factors, 

 the sugar and ash content the least. The commercial importance of the 

 fatty constituents of milk and the relative ease of their quantitative 

 determination in toto are responsible for the fact that different milk 

 samples are distinguished and graded in accordance with their content 

 of fat. Therefore, the most profitable and practically significant method 

 of studying the protein content of milk is to consider the fat-to-protein 

 ratio, with the hope of finding some formula from which the protein 

 content of milk may bo satisfactorily predicted if the fat content is 

 known. 



The protein of milk appears to be more closely correlated with the fat 

 content than any other constituent. Gaines (^') lias determined the 

 correlation existing between the fat content and the content of protein, 

 sugar and ash from Haecker's(^®) published analyses of mixed milk. 

 The correlation coefficient for fat and protein was +0.812 ±0.010, for 

 fat and sugar, +0.263 ±0.027, and for fat and ash, +0.232 ±0.027. 

 Overman and Sanmann(^®), working with a smaller set of data, report 

 a correlation coefficient of +0.729 ±0.017 for fat and protein, and one of 

 + 0.184 ±0.045 for fat and lactose. 



The prospect would seem favorable, therefore, for devising a predic- 

 tion formula for protein from fat. In 1899, Timpe(^^) derived the rela- 

 tion p = 2 + 0.35f between the protein and fat contents of milk from 

 analyses of milk from 21 cows of various breeds. Van Slyke(^®) in 1908 

 proposed the relation p=1.6 + 0.4f from a large, but unspecified, num- 

 ber of analyses of herd milk and milk from individual cows of dif- 



