FOOD-COLOKING SUBSTANCES. 21 
fairly pure condition by the fractionation and the solutions obtained 
will be ready for identification tests. Obviously, no essential differ- 
ence exists between these and the so-called coal-tar colors; as a class, 
however, they show much less tendency to dye wool than do the 
common synthetic colors, and in addition are in many cases so sensi- 
tive to alkalies as to be completely destroyed in the double-dyeing 
test; i. e., by dyeing, stripping the fiber with dilute ammonia, acidi- 
fying, and dyeing again. The preliminary dyeing with wool described 
on page 8 serves fairly well in practice as a first indication of 
the course to be followed; but when for some reason the results 
obtained are not decisive, the general treatment for coal-tar colors 
with immiscible solvents should be carried out with consideration of 
solubilities of the coloring matters described in the footnotes in the 
tables of solubilities. In the dyeing test not only are certain syn- 
thetic dyes, especially Auramin, unstable under the conditions of 
treatment and likely to be destroyed ; but, on the other hand, Archil, and 
to a lesser extent other natural colors, may give well-marked dyeings. 
The crude products constituting the commercial natural coloring 
matters in most cases are mixtures containing several closely related 
chemical individuals. These may have different solubilities, but 
usually they contain the same chromophore groups, and are of closely 
similar shade. In practice, the analyst will scarcely attempt a full 
separation, but having identified the coloring matter in one fraction, 
can judge as to the likelihood of the other substances present being 
derived from the same source. 
The natural colors as a class do not contain strongly acidic groups, 
and their distribution ratios between immiscible solvents do not show 
wide variations with the acidity of the latter, at least not within 
convenient limits. The coloring matters of cochineal and turmeric 
give less trouble than the others, partly because they are less hete- 
rogeneous. 
When coal-tar dyes are absent, and it is desired to fractionate with 
immiscible solvents, it is best to begin extraction with neutral solu- 
tions; perhaps first using ether (petroleum ether is better when 
chlorophyll and the accompanying leaf colors are to be separated). 
The final extraction may be made with amyl alcohol from acid solu- 
tion, but it is of no advantage to have the acidity high, not, perhaps, 
above thirty-second normal. The anthocyans which constitute the 
coloring matters of the common red fruits and flowers are glucosids, 
and are extracted from acid solution only in very small amount by amyl 
alcohol and similar solvents. Their neutral solutions may be treated 
with excess of lead acetate solution (normal salt) when practically 
all of the glucosid will be precipitated. This mixture may be cen- 
trifuged and the precipitate washed in the centrifuge tubes with 
several portions of water until sugar and similar soluble substances 
