162 Testing Milk and Its Products. 



each lot of milk. This can easily be done by means of 

 special sampling devices (see 182 et seq.). As the quan- 

 tities of the milk delivered from day to day by each 

 patron vary but little, perhaps not exceeding 10 per 

 cent, of the milk delivered, the error introduced by 

 taking a uniform sample, e. g., an ounce of milk, each 

 time is, however, small and it may not be necessary to 

 take cognizance of it in factory work. This method of 

 composite sampling described is quite generally adopted 

 in separator creameries and cheese factories, where the 

 payment for the milk is based on its quality. 



In order to obtain reliable results by composite sam- 

 pling it is essential that each lot of milk sampled shall 

 be sweet and in good condition, containing no lumps of 

 curdled milk or butter granules. The milk should of 

 course always be evenly mixed before the sample is 

 taken. 



i82. b. Drip sample. Composite samples are some- 

 times taken at creameries and cheese factories by col- 

 lecting in a small dish the milk that drips through a 

 fine hole in the bottom of the conductor spout through| ^^ 

 which the milk runs from the weighing can to the •'re- 

 ceiving vat or tank. A small portion of the drip col- 

 lected each day is placed in the composite sample jar, 

 or the quantity of drip is regulated so that all of it 

 may be taken. In the latter case the quantity of milk 

 delivered will enter into the composite sampling as well 

 as its quality and the sample from, say 200 lbs. of milk, 

 will be twice as large as the sample from 100 lbs. of milk. 



Where it is desired to vary the size of samples accord- 

 ing to the quantity of milk delivered from day to day. 



