20 BULLETIN 448, TJ. S. DEPARTMENT OF AGBICTJLTTJBE. 
The original mixture, from which the seven colors described were 
separated by adding acid and shaking out with amyl alcohol, may still 
contain Light green S F yellowish, which will be colorless or nearly 
so in the acid solution. To separate this dye the mixture is treated 
with strong ammonia or potassium hydroxid solution until slightly 
alkaline, then neutralized with acetic acid. Any green that is pres- 
ent will now be apparent by the color of the mixture and may be 
extracted by shaking with a few small portions of chcmorhydrku 
After washing the dichlorhydrin extract with a little water it should 
be diluted with several volumes of benzene or carbon tetrachloride 
The dye is then taken out with water. 
Besides the well-known reactions of acids and alkalies on the dyed 
fiber or with the solution, a few tests may be mentioned as best 
suited for the quick characterization of the different colors obtained 
in this separation. Incligo Carmine is extracted in rather small pro- 
portions from acid aqueous solutions by dichlorhydrin, differing in 
this respect from nearly all the ether common blue dyes. The bromin 
test (p. 47) and the reactions with acids and alkalies usually are suf- 
ficient for the identification of Tartrazin after its separation. New 
Coecin (Xo. 106) and Ponceau 6 R. (Xo, 108) are the only other com- 
mon red dyes of solubility similar to that of Amaranth. Both are yel- 
lowish in shade, the former markedly so. Ponceau 3 R, when treated 
with barium chlorid and enough sodium acetate to neutralize any free 
hydrochloric acid present, gives a bluish-red flocculent precipitate, 
the supernatant liquid being left practically colorless. Ponceau 2 R 
gives a carmine red precipitate: but most other red dyes of similar 
solubility do not form particularly insoluble barium salts. Xaphthol 
yellow S in solutions treated with an excess of ammonia or sodium 
carbonate becomes intensely rose colored on addition of sodium hydro- 
sulphite, the color gradually fading again as complete reduction takes 
place. Orange I is well characterized by its solubility and behavior 
with acids and alkalies. Eryihrosin is perhaps best further iden- 
tified by a test for iodin. Some of the color solution containing a 
slight excess of alkali is evaporated to dryness, the residue heated to 
redness, ancl the ash taken up with water and acidified with sulphuric 
acid. Iodin may be tested for m the usual ways, such as with carbon 
tetrachlorid and a small drop of sodium nitrite, or with starch paste 
and an oxidizing agent. It is useless to test for iodin with very 
small amounts of dye, but in most cases sufficient coloring matter 
will be available to give satisfactory results. 
TREATMENT OF SOLUTION RESERVED FOR TESTING FOR NATURAL COLORING 
SUBSTANCES. 
If coal-tar dyes have been found, the treatment for their separa- 
tion will have given much information as to natural colors that may 
be present. Many of the natural colors will have been separated in 
