Testing Milk on the Farm. 151 



held; a sample for testing purposes is then taken at 

 once. A correct sample of a cow's milk cannot be ob- 

 tained by milking directly into a small bottle from one 

 teat, or by filling the bottle with a little milk from each 

 teat, or by taking some of the first, middle and last milk 

 drawn from the udder. Such samples cannot possibly 

 represent the average quality of the milk of one entire 

 milking, since there is as much difference between the 

 first and the last portions of a milking, as between milk 

 and cream. ^ Lack of care in taking a fair sample is 

 the cause of many surprising results obtained in testing 

 the milk of single cows. 



169. Composite samples. When a cow is to be tested 

 for milk production she should be milked dry the last 

 milking previous to the day when the test is to be made. 

 The entire quantity of milk obtained at each milking is 

 mixed and sampled separately. On account of the vari- 

 ations in the composition of the milk, a number of tests 

 of successive milkings must be made. As this involves 

 considerable labor, the plan of taking composite samples 

 is preferable ; the method of composite sampling and test- 

 ing is explained in detail under the second subdivision of 

 Chapter X (180) ; suffice it here to say that the method 

 followed in the case of single cows' or herd milk is to 

 take about an ounce of the thoroughly mixed milk of 

 each milking; this is placed in a pint or quart glass jar 

 containing a small quantity of some preservative, prefer- 

 ably about half a gram (8 grains) of powdered 

 potassium bi-chromate. If a number of composite sam- 

 ples of the milk of single cows are taken, each jar should 



' Agricultural Science, 6. pp. 540-42. 



