1905.] 



Butter Tests. 



485 



It will be seen that the ratios obtained vary very widely, 

 20 lb. of Shorthorn milk sufficing at one time to yield 1 lb. of 

 butter, while at another no less than 41 lb. were required. In 

 the same way something under 16 lbs of Jersey milk yielded 

 1 lb. of butter in November, 1902, while in August of the same 

 year 25 lb. of milk were needed. Although some very bad 

 ratios were obtained during the hot weather, when, ice or cold 

 spring water not being available, a certain amount of the butter 

 was not recovered, yet, generally speaking, the worst ratios were 

 obtained during the winter months, when the cattle were indoors 

 and living on artificial food. All recent work on the analysis of 

 the milk yielded by the same cows goes to show that the daily 

 variation in composition is very marked, and that it is quite im- 

 possible to explain such variations by reference to feeding or 

 external conditions. It would therefore be expected that the 

 amount of butter obtained from a given quantity of milk would 

 vary in like manner, and the experiments at Bickenhall bear 

 out in a striking way the wide variations in composition found 

 by other experimenters to exist. 



Butter Tests with Mixed Milk. — Although the milk of each 

 breed was separately tested, the bulk of the butter was of 

 course made from the mixed milk uf both breeds. Where two 

 milks of varying quality are mixed in this way, one would 

 expect that the quantity of butter obtained would hold some 

 intermediate position between that which would be obtained 

 from the milks churned separately. An interesting point arose 

 as to whether the quantity of butter obtained would vary 

 exactly, according to the quantity of each milk used, or whether 

 it would obey some independent law of its own ; in other words, 

 would the amount of butter obtained from two given quantities 

 of milk mixed before churning, be equal to the sum of the 

 butter obtained from the same milk if churned separately, or 

 would the addition of the richer milk to the poorer influence the 

 mixture in such a way as to give more butter than if they were 

 churned separately ? Much time and attention has been given to 

 the elucidation of this problem. The first experiment designed 

 to throw light upon it was started in November, 1901, and con- 

 tinued till March, 1903. It consisted of a series of monthly 

 tests, in which Shorthorn and Jersey milk were mixed in the 



