large a per cent of the total fat as possible into the butter, sec- 

 ond, of the portion not recovered get as little as possible in the 

 butter-milk. 



The milk used in these tests was sampled and analyzed. 

 Morning's milk had S()2% ^^ ^^t, night's milk 3.99% of fat. 



The following table shows the amount of fat in the milk* 

 the amount lost in the skim-milk and butter-milk and the amount 

 recovered in the butler : 



Moselev and r- i Open c 



Stodda.d. <^°°'^>'- p.ns. Separator. 



lbs. lbs. lbs. lbs. 



Total fat in milk, 5.9400 5.9400 5.6500 5.9400 



Lost in skim-milk. -3-46 •3C94 -3976 -4315 



Lost in butter-milk, .6020 .8718 .2140 .1089 



Recovered in butter, 48776 47734 51450 5.41 18 



] er cent. per cent. per cent. 



Per cent of Total fat lost in 



skim-milk, 5.47 5.21 7.03 7.26 



Per cent of Total fat lost in 



butter-milk, '0-i3 14-67 3.78 1.83 



Per cent of Total fat 



recovered in Butter, 82.11 80.36 91-07 91.10 



If perfect work could be done and if there was such a thing 

 as absolute accuracy, the sum of the fat in butter, skim milk 

 and butter-milk would equal the total fat in the milk used, but 

 there are small errors in weighing, and in chemical analysis and 

 sampling which are unavoidable. Every chemist recognizes 

 certain limits of error in his work ; every investigator realizes 

 that weighings made with standard scales are imperfect; as a 

 result we find just what would be expected, namely, that the sum 

 of the fat in butter, skim-milk and butter- milk in each system 

 does not correspond exactly with the fat in the milk, but if we 

 take the four lots it appears that we had, in the milk used 23.47 

 pounds of fat and there is accounted for 23.46 pounds, an error 

 of less than one-sixth of an ounce, hence we may reasonably 

 conclude that these errors, as a whole, balance. 



The figures are so plain that comment is unnecessary, and so 

 far as one test conducted by competent and careful men can be 

 relied upon we may place the Separator first, the Open pans 

 second, the Moseley and Stoddard third, and the Cooley last, in 

 point of efficiency. 



8 



