BUTTER 63 



mixed by violently agitating the bottom of the tube. When now 

 allowed to stand and separate, the lower layer of mixed acids 

 will be strongly colored wine-red if the azo-color be present. 

 Pure butter-fat imparts no color to the acids, or, at most, only 

 a faint brownish tinge. 



For turmeric and annatto mixtures, Martin's test will 

 usually be satisfactory: 2 c.c. carbon disulfid are mixed with 

 15 c.c. of alcohol, by adding small portions of the disulfid 

 to the alcohol and shaking gently; 5 grams of the butter- 

 fat are added to this mixture in a test-tube and shaken. The 

 disulfid falls to the bottom of the tube, carrying with it the 

 fatty matter, while any artificial coloring-matter remains in 

 the alcohol. The separation takes place in from one to three 

 minutes. If the amount of the coloring- matter is small, more 

 of the fat may be used. If the alcoholic solution be evaporated 

 to dryness and the residue treated with concentrated sulfuric 

 acid, annatto will be indicated by the production of a greenish- 

 blue color. With many samples of oleomargarin, a pink tint 

 will be produced, which indicates an azo-color. 



Palm oil is sometimes used as a coloring agent in butter- 

 substitutes. Crampton & Simons have found that two tests 

 devised for detection of rosin-oil can be satisfactorily adapted 

 to detection of palm oil. Success depends on several points. 

 The sample must be kept in a cool dark place until used, filtered 

 at a temperature not above 70°, the heating as brief as possible, 

 and promptly tested. The reagents must be pure and colorless. 



Cochran finds that annatto will simulate palm oil in these 

 tests, and hence the absence of the former must be assured (see 

 above) before inferring the presence of the latter. 



Halphen method. 100 c.c. of the filtered fat are dissolved in 

 300 c.c. petroleum spirit and shaken out with 50 c.c. of potas- 

 sium hydroxid solution (0.5 per cent, of hydroxid). The water 

 is drawn off, made distinctly acid with hydrochloric acid, and 

 shaken out with 10 c.c. of carbon tetrachlorid. This solution 



