Composite Samples of MUk. 161 



patron of the factory and is labeled with his name or 

 number. A small quantity of preservative (bi-chromate 

 of potash, corrosive sublimate, etc., see 190) is added to 

 each jar; these are placed on shelves or somewhere 

 within easy reach of the operator who inspects and 

 weighs the milk as it is received at the factory. When 

 all the milk delivered by a patron is poured into the 

 weighing can and weighed, a small portion thereof, 

 usually about an ounce, is put into the jar labeled with 

 the name or number of the patron. The samples are 

 conveniently taken by means of a small tin dipper hold- 

 ing about an ounce. This sampling is continued for a 

 week, ten days, or sometimes two weeks, a portion of 

 each patron's milk being added to his particular jar 

 every time he delivers milk. A test of these composite 

 samples takes the place of separate daily tests and givp.s 

 information regarding the average quality of the milk 

 delivered by each patron during the period of sampling. 

 The weight of butter fat which each patron brought 

 to the factory in his milk during this time, is obtained 

 by multiplying the total weight of milk delivered dur- 

 ing the sampling period by the test of the composite 

 sample, dividing the product by 100. 



i8i. This method of taking composite samples has 

 been proved to be practically correct. It is absolutely 

 correct only when the same weight of milk is delivered 

 daily by the patron. If this is not the case, the size of 

 the various small samples should bear a definite relation 

 to the milk delivered; one sixteen-hundredth, or one 

 two-thousandth of the amount of milk furnished should, 

 for instance, he taken for the composite sample from 



