379 



The average of these differences is less than a tenth of a pound of 

 butter fat in 4 days. Had the samples been taken for 7 days tlie differ- 

 ences woukl have been still smaller. The largest difference is two 

 thirds of a pound, and this is on a large lot of over a hundred pounds 

 of butter fat. 



In the light of these results it is evident that the taking of samples 

 for analysis that var^^ in size with the quantity of milk is a needless 

 precaution, and that in actual practice it is better to have a single 

 sampling cup for all the work. The author advocates the adoption of 

 the fat test as a basis for paying for milk at cheese factories, as well 

 as creameries. 



" In general, it is true that the more fat the milk contains tbe more 

 casein it also contains, so that the amount of cheese to be made out of 

 the milk can be very accurately gauged by the amount of fat it contains. 

 It is also true that the market value of the cheese comes almost entirely 

 from the fat. * * * a method to be just to all, however, must take 

 into account both the casein and the fat. This can be arrived at quite 

 closely by paying a certain amount for the milk D^ weight, without 

 regard to its quality, and a certain amount additional for each pound 

 of butter fat it contains." 



Notes foe the laboratory. — It is believed that in the hands of 

 an experienced manipulator the method above described is capable of 

 being brought to a high degree of accuracy. Suggestions for its use 

 in laboratories in testing cream, butter, and cheese, as well as milk, 

 are given. 



