10 BULLETIN 448, U. S. DEPARTMENT OF AGRICULTURE. 
while 94.1 per cent of the other color, under the same conditions, 
remains in the aqueous layer. If such a mixture in water solution is 
brought to the given acidity and is shaken out successively with three 
portions of amyl alcohol, each equal to one-fourth its volume, calcu- 
lation shows that if the distribution ratios remain constant there will 
be present in the arid solution after the third shaking twenty-seven 
sixty-fourths or 42 per cent of the first dye. one sixty-fourth or 1.6 
per cent of the second. Conversely the first amyl alcohol portion 
after two washings with portions of acid of concentration similar to 
that of the original solution, will contain 42 per cent and 1.6 per cent 
of the second and first dyes, respectively. Obviously, for a practi- 
cable quantitative separation, somewhat greater differences in solu- 
bility must exist ; but it is usually sufficient to separate, in fairly pure 
condition, a portion of each of the colors that seem to be present, in 
order to characterize completely the components of the mixture. 1 
Emulsions occasionally cause trouble when the analyst is working 
wiih impure mixtures. These are most effectively broken by a good 
centrifuge. Should a solid stratum form between the two layers, it 
should be broken with a glass rod. the tube replaced in the cen- 
trifuge and whirled again. Heating tends to promote rapid separation. 
but the relative solubilities vary somewhat in hot mixtures. Strong 
- show much less tendency to emulsify than neutral or 
alkaline ones. 
Tiie final separation of mixtures of dyes of rather similar solu- 
bility will usually be made by selecting some pair of solvents in which 
they show a decided difference. Mixtures of dyes of practically 
identical solubility can. in most cases, be separated satisfactorily by 
chemical means or by precipitation reactions. Since the fractiona- 
tion will have removed all except a few dyes belonging to a known 
group, suitable chemical methods may usually be chosen without 
difficulty. 
While the scheme described is not intended to be applied to rela- 
tively concentrated solutions, hi practice in the examination of col- 
ored food products, such are seldom or never obtained. The chief 
concern of the analyst here will be to avoid, as far as possible, the 
dilution of the color by the use of unnecessarily large portions of 
organic solvents and washing liquids. Only in working with prod- 
ucts sold for use as coloring matters are solutions likely to be made 
too concentrated to be adapted to the scheme of separation. 
As the common food dyes are, for the most part, salts of polybasic 
acids, the equilibrium conditions are obviously quite complex and 
concern not only the relative solubilities and dissociation constants of 
the free color acid, but of the various acid salts and the sodium salt as 
» For procedure and calculation as to quanilTaiive fractionations, no coloring matter being rejected, 
see J. Ind. Eng. Chem. 5 (1913).. 26. 
