Butter Record oe the Buckhoed Herd. 315 



days in a month), give the individual butter record of each cow 

 for the year as obtained by this method (see column XIV.) ; and 

 the total at the foot of that column gives the yearly butter record 

 of the herd as obtained by this method, i.e., 14,838 lb. 



Reasons for deciding which of the two Records, A or B, is the 



more reliable. 



The record B does not always agree quite accurately with 

 the record A. One reason is that the former record entirely 

 depends on the chemical test (it is, however, checked each 

 month by the churn), whereas the record A depends principally 

 on the churn ; and churning, however carefully it may be carried 

 out, is liable to slight errors. Thus, on one day a little more than 

 12 per cent, of water, and on other days a little less, may be in- 

 cluded in the butter. For this reason the chemical test would 

 seem to be the more exact, as butter-fat is a very nearly constant 

 substance, while butter, even in the hands of tl e very best dairy- 

 men, varies slightly from day to day. But tht/'j is another reason 

 why the two accounts do not completely agree, namely, that the 

 chemical test is taken only on one day in the month, while the 

 churning takes place here on all week days ; and it may be that 

 the day of the chemical test is a little unfavourable, as would 

 happen if it were taken on a very hot day in summer or on a 

 very cold day in winter, or opposite conditions might prevail, 

 and the test be too favourable. It will further be noticed that 

 as the chemical test is reckoned for only 360 days (30 x 12) the 

 result is about 1^ per cent, too low, 



The question then arises, as these two records check one 

 another and yet do not quitej coincide, which of them should be 

 considered to be the more reliable ? The record A has been taken 

 as the standard as being probably more exact than the other for 

 the reasons given above, and also because in general there is 

 greater confidence in results obtained by the churn than in those 

 obtained by a chemical test. It has been our custom to add to 

 the record B or to subtract from it until it is nearly equal to 

 the record A, and this is the reason for the addition of 3 per 

 cent, to the individual records in column XV. on pp. 318, 319. It 

 may be mentioned that no greater difference than 3 per cent, 

 between the two records has arisen in the last six years. The 



