THIRTY-FIFTH ANNUAL CONVENTION. 115 



likely to cause more error than the use of the pipette, a method 

 incorrect enough to make its use punishable by law. 



It is not uncommon to still find an error introduced by an 

 improper reading of the fat column, for instance, the older direc- 

 tions for reading cream state that we should read over all, that 

 is, between points indicated by upper and lower meniscus. Later 

 directions for reading cream state that we should read to the 

 bottom of the upper meniscus at a temperature of 140 degrees 

 F. This amounts to much the same as the method employed 

 by our own experiment station, but, since in practical work we 

 seldom find a tester who reads samples at 140 degrees, we have 

 compromised by dividing the meniscus nearly in half and read 

 at a temperature of 120 degrees F., according to the accompany- 

 ing chart. So far we have no authority in the way of a direct 

 comparison with the chemical analysis which would permit us 

 to use this form of reading, but comparisons of butter fat churned 

 with butter fat recovered on a large number of churnings corre- 

 spond so closely that we feel warranted in using this method 

 until it is sustained or disproved by the chemist. Another com- 

 mon mistake in reading is that of disregarding the half per cents. 

 When we consider that a per cent on 30 per cent cream will 

 change our overrun practically 2 per cent, we can readily under- 

 stand why in some cases we get such high or such low overruns. 

 To obtain anywhere near accurate results, the half must be taken 

 into account. 



Errors due to improper sampling of cream are not particu- 

 larly common where the samples can be taken by the creamery- 

 man himself, but we must have a great deal of sympathy for 

 that plant that must depend upon the accuracy of the cream 

 haulers' samples. Lack of knowledge of the importance of 

 sampling, combined with favoritism on the part of the hauler, 

 often work havoc with creamery records, and the only safe 

 method, where such samples are taken, is to run a daily composite 

 check on the hauler. Such a check need be no reflection on the 

 hauler, and often leads one directly to the source of creamery loss. 



The accompanying table shows what may happen when in- 

 correct samples are taken. 



