82 Testing Milk and Its Products. 



It is very convenient to weigh out 18 grams of cream 

 (or a fraction thereof) so that the readings may be 

 taken directly from the neck of the bottle. The smaller 

 the quantity of cream taken for a sample, the greater 

 is the error introduced by inaccurate weighings or read- 

 ings. The result is rendered more accurate and certain 

 if two or three tests of a sample are made, and the 

 readings averaged. 



92. Measuring cream for testing. Where a special 

 cream scale or a small balance is not available, fairly 

 satisfactory results may be obtained with cream of low 

 or average quality by measuring out the sample with 

 a 17.6 pipette and correcting the results as indicated 

 below. One cf the cream test bottles or a common milk 

 test bottle may be used for this purpose. The table 

 on p. 76 shews that a 17.6 cc. pipette, in the case of 

 cream fresh from the separator, containing less than 

 25 per cent, of fat, will deliver only 17.2 grams of 

 cream, that is, the results will bei**i 5 =l.l per cent, too 

 low. In the same way in case of 40 per cent, cream, 

 only 16.3 grams of cream would be delivered, and the 

 results therefore 3.8 per cent, too low. When the cream 

 has been ripened or is thick, less cream would be deliv- 

 ered than the amounts given, and the error introduced 

 by measuring out the samples correspondingly increased. 

 A table of correction for testing such cream by meas- 

 uring the samples has been prepared by Prof. Eckles, 

 formerly of Iowa experiment station. 1 



1 Press bull, dated August, 1901. Bonn 1 croMimrit>s boat tin- sam- 

 ples of en-am in a w:il-r bath to about 110 K. brforo tin- test -;mipl-s 



