DAIRYING. 45 



118. Any owner of cows can well afford to invest $10 in an 

 8-bottle tester, $3 in a milk-weighing scale with an adjustable loose 

 hand on the dial for balancing the weights of the pail, and 50 cents 

 in milk record sheets. The information obtained by the use of these 

 in getting an exact knowledge of the milk production of each cow 

 will be worth many times the price paid for it. 



G Methods of Testing the Milk Production of Cows. 

 1. Farm Tests by the Owner. 



119. Many good dairymen consider the information gained by 

 weighing each cow's milk at each milking, well worth the time 

 required to do this work. It has been shown, however, that weigh- 

 ing the milk of each cow two consecutive milkings one each week or 

 even one in two weeks, and multiplying the figures thus obtained by 

 the number of days elapsed since the preceding weighing, will give 

 a fairly close agreement with the results obtained by weighing the 

 milk of each milking. 



120. The number of tests necessary for showing the total pro- 

 duction of a cow depends largely on the uniformity of her milk in 

 quality from day to day. The milk flow of all cows gradually de-' 

 creases with the progress of the period of lactation, but the rich- 

 ness of some cows' milk varies more than others from day to day, 

 hence the number of tests necessary to show her average produc- 

 tion will vary with the peculiarity of the cow in this respect. Some 

 information on this point has been obtained by comparing the results 

 obtained by adding together the weights recorded at each milking 

 with those found by calculating the total production of milk and 

 butter fat from weights and tests made once each week, two weeks, 

 and four weeks. Such a comparison made with the records of each 

 one of six COW T S showed- that weighing and sampling the milk of two 

 consecutive milkings once a week gave 98 per cent, once in two 

 w r eeks 97 per cent, and once in four weeks, 96 per cent, of the 

 amount of milk and of butter fat obtained by daily weights. This 

 shows that there is a probable error of about 2 per cent, in the 

 calculation of a cow's annual production of milk and of butter fat 

 when such calculations are based on weights and tests made for 

 one day either once a week, or once in two weeks, and that a probable 

 error of about 4 per cent, exists in records based on weights and 



