114 ILLINOIS STATE DAIRYMEN'S ASSOCIATION. 



obliged to allow a fraction of a per cent for the individuality 

 of the reader himself. All of these factors conspire to make us 

 liable to certain reasonable error which is always present and 

 must be reckoned with. However, these errors should not al- 

 ways fall in the same direction nor need they change results 

 materially, since they tend to divide themselves about equally 

 on either side of a medium line, thus tending to equalize each 

 other at the end of any given testing period. 



To further illustrate, a table has been prepared, showing the 

 tendency for variation of two experienced testers, determining 

 fat in the same cream. The samples tested represent the com- 

 posite tests of patrons' cream for the two halves of the months 

 named. Reading the columns in either direction, shows seldom 

 a variation of more than half of one per cent, and that there is 

 no regular variation in either direction. From our own creamery 

 records we might tabulate data from ten months, on one hundred 

 patrons, illustrating the same point. Evidence enough to con- 

 vince us that in testing cream samples the reading is likely to be 

 one-half per cent too high, or one-half per cent too low. 



Errors greater than this may usually be accounted for in 

 one of the following ways: burned tests, poor scales, careless 

 weighing, incorrect sampling, and misreading. Burned or cloudy 

 tests should be the least of our troubles, since they are easily over- 

 come by employing proper temperatures of acid and milk and 

 regulating the strength of acid. As an aid in securing more uni- 

 form readings and less discolored fat columns, we have found 

 it very beneficial to change the order of manipulation for cream 

 and add water to the neck previous to the first whirling before 

 the acid has had a chance to char the fat. Then handle in the 

 usual manner, whirling twice for periods of 3 and 5 minutes, 

 respectively. This method, intelligently followed, seldom gives 

 occasion for cloudy tests. 



It is not an uncommon thing to find in many places where 

 cream is weighed for testing, scales that are not sensitive enough 

 to be used for that purpose, or dirt and dust has accumulated in 

 the bearings, causing them to stick. Scales in this condition are 



