232 Maine Agricultural Experiment Station. 1918. 



in the ratios of solids-not-fat to butter-fat as compared with 

 other breeds. 



The correlations and linearity of regression between the 

 variables butter-fat per cent, amount of milk and age at com- 

 mencement of test lead to the following conclusions. 



As the amount of milk given by the cows in this test in- 

 creases, the percentage composition of the butter-fat decreases. 

 The amount of this decrease is highly significant measured sta- 

 tistically. Considered practically this fall in butter-fat content 

 could not be easily detected in small samples. No such effect 

 is noted for the solids-not-fat or put in another way, the quan- 

 tity of milk produced for one year is independent of the con- 

 centration of the solids-not-fat. This, from the view of inheri- 

 tance, means that the hereditary units for high or low milk 

 production are separate and distinct from those causing a high 

 or low percentage of solids-not-fat. 



The correlations of the age when the yearly record com- 

 mences with butter-fat and with solids-not-fat brings out the 

 following points. Age of the cow does not effect the percentage 

 of butter-fat in the milk significantly. As the age of a cow 

 advances each lactation brings with it a decrease in the per- 

 centage of the solids-not-fat found in the milk. 



This differential action of amount of milk produced and 

 age gives us the criterion to prove that butter-fat and solids- 

 not-fat cannot have a common mother substance from which • 

 they are derived by splitting. 



The correlations between the variables, pounds of milk, 

 butter-fat and solids-not-fat lead to the following conclusion. 

 (a). Some of the factors responsible for high concentration 

 of butter-fat are also responsible for high concentration of some 

 of the solids-not-fat in cow's milk. (b). Practically consid- 

 ered this means that if it is desired to improve either the but- 

 ter-fat or solids-not-fat concentration of the milk of a given 

 1 erd, the determination of the concentration of either solid and 

 selection of the animals accordingly will result in a correspond- 

 ing increase for the other solid. 



Taken in conjunction with the above results the diurnal 

 variations of cow's milk furnish the facts necessary to test the 

 hypothesis to account for the mode of secretion of the milk 

 solids. The data show that evening milk is between 0.678 and 



