204 



COAL-TAR COLORS USED IN FOOD PRODUCTS. 



Percentages based on theoretical composition Continued. 



INDIGO DISULPHO ACID. 



ARSENIC DETERMINATIONS ON 86 BATCHES. 



During the course of this work it became necessary to determine 

 the exact amount of arsenic contained in 86 of the various batches 

 of certified colors ; for this purpose the arsenic method of Seeker and 

 Smith (see p. 212) was devised. These results are expressed in the 

 number of parts of color containing one part of metallic arsenic (As). 

 The numbers in parentheses indicate the number of specimens of the 

 quality indicated. 



Naphthol Yellow S (17 specimens). 770,000, 833,000, 1,250,000 (2), 1,428,000 (2), 

 1,666,000 (3), 2,000,000 (6), 3,3^3,000 (2). 



Ponceau (15 specimens). 588,000, 625,000, 667,000, 714,000 (2), 769,000 (2), 

 1,000,000 (2), 1,111,000, 1,250,000, 1,666,000 (2), 2,000,000, 2,500,000. 



Orange (7 specimens). 200,000, 250,000, 588,000, 1,000,000, 1,111,000, 1,429,000, 

 5,000,000. 



Amaranth (27 specimens). 909,000, 1,000,000, 1,111,000, 1,250,000 (2), 1,438,000 (3), 

 2,000,000(4), 2,500,000 (4), 3,333,000 (9), 5,000,000 (2). 



Green (8 specimens). 166,000, 200,000, 370,000 (2), 500,000, 666,000, 770,000, 833,000. 



Erythrosin (7 specimens). 2,000,000(2), 5,000,000 (2), 10,000,000 (2), 20,000,000 (1). 



Blue (5 specimens). 285,000, 666,000, 1,428,000, 3,333,000 (2). 



The United States Pharmacopoeia test for arsenic is sensitive to 

 0.005 mg of arsenious oxid (As 2 O 3 ) which on a sample containing 2 

 grams of actual or real color would amount to one part of arsenious 

 oxid in 400,000 of color; calculated to metallic arsenic, the basis 

 employed in the foregoing, this would mean 528,000 parts of color for 

 each one part of metallic arsenic. 



There are, therefore, in the foregoing 86 lots of certified colors 

 8 which did not comply with that requirement, namely: 



Orange (2 specimens). 200,000 and 250,000. 



Green (5 specimens). 166,000; 200,000; 370,000 (2); 500,000. 



Blue (1 specimen). 285,000. 



The reason for this discrepancy is that in the preparation of the 

 samples for analysis by the United States Pharmacopoeia test a loss 

 of arsenic ensued, which is avoided in the Seeker-Smith method 

 now employed in making these determinations. Had the existence 



