274 MANUAL OF MILK PRODUCTS 



pose of heating and mixing is to distribute the moisture evenly 

 throughout the sample. If the butter is melted too far, it will 

 have to be cooled before being weighed out for the test. The 

 sample must be kept constantly stirred while cooling in order 

 to keep the moisture evenly distributed ; if this stirring is not 

 done, the moisture will tend to collect in the center and an accu- 

 rate test cannot be made. Three or four grams of the prepared 

 sample is weighed out into a cream bottle; if more than this 

 quantity is used, the fat column will be so large that it cannot 

 be read in the bottle. The sample is then made up to approxi- 

 mately 18 grams by adding water. The purpose of adding water 

 is the same as in testing cream : it retards the action of the acid. 

 About 12 c.c. of acid is then added ; this quantity is sufficient be- 

 cause the solids not fat in butter are present in comparatively 

 small proportions. The tests should be made in duplicate. 

 From this point the test is conducted in the same manner as is 

 the test for fat in cream (see p. 119). The fat column should 

 be kept at the proper temperature, and glymol should be used 

 in reading the completed test. 



BUTTER-MOISTURE TESTS 



A number of methods have been devised for testing the per- 

 centage moisture in butter which give satisfactory results in 

 commercial work. The following is representative. 



Cornell butter-moisture test (Ross) 



When butter is heated, a covering of casein will collect over 

 the surface of the sample. When the sample becomes quite hot, 

 this covering is of snow-white color. After heating the sample 

 for a time, the foam begins to subside and loses its snow-white 

 color, changing to a dirty brown. By comparison with the 

 chemical method, it was found that when the foam had lost its 



