36 Testing milk and its Products. 



gravimetric analyses have shown that the readings ob- 

 tained in this manner gi^ correct results. While the 

 lower line of the fat column is nearly straight, the upper 

 one is curved, and errors in the reading are therefore 

 easily made, unless the preceding rule is observed. 



The fat obtained should form a clear yellowish liquid 

 distinctly separated from the acid solution beneath it. 

 There should be no black or white sediment in or below 

 the column of fat, and no bubbles or foam on its sur- 

 face. The bottles must be kept warm until the read- 

 ings are made, so that the column of fat will have a 

 sharply defined upper and lower meniscus. When the 

 testing is done in a cold room, it is a good plan to place 

 the bottles in a pail with water of 140 F. be- 

 fore readings are made. The readings should always 

 be made when the fat has a temperature of about 140 

 F. ; too low results will be obtained if the fat is allowed 

 to cool below 120 F., and too high if readings are 

 taken above 150. The fat separated in the Babcock 

 test solidifies at about 100 F. If the fat is partly sol- 

 idified, it is impossible to make an accurate reading. 1 



42. Readings of tests of milk made in steam turbine 

 testers with tightly closed covers which prevent the free 

 escape of exhaust steam (71) , will come .2 to .3 per cent. 



x The effect of differences in the temperature of the fat on the read- 

 ings obtained will be seen from the following: If 110 and 150 F. be 

 taken as the extreme temperatures at which readings can be made, 

 this difference of 40 F. (22.8 O.) would HIM UP a difference in the volume 

 of the fat column obtained in case of 10 per cent, milk, of .00064 x 2 x 22.3 

 = .028544 cc., or .14 per cent., .00064 being the expansion coefficient of pure 

 butter fat per degree Centigrade between 50 and 100 O. fZune, Anal y si- 

 des Beurre*, I, 87), and 2, the volume of the fat in cc. contained In 17.6 cc. 

 of 10 per cent. milk. On 5 per cent milk this -xtnme difference would 

 therefore be about .07, or nearly one-tenth, of one per cent. 



