FOOD-COLORING SUBSTANCES. II 
well. Many dyes exist in more or less associated condition in ordinary 
solutions. However, it is found in practice that in most cases the 
distribution ratios with given acidity do not vary greatly, but that 
fair results can be obtained on the assumption that they will remain 
constant. 1 
Both basic and acid triphenylmethane colors tend to undergo slow 
intermolecular changes when treated, with acids and alkalies (adjust- 
ment of equilibrium between carbinol, imid, and ammonium forms) 
and their complete separation by means of solvents is less simple than 
that of most other classes. 
It will be noticed from the solubilhy table, pages 22-33, that amyl 
alcohol, amyl alcohol gasoline mixtures, and ether, although differing 
greatly in their power as solvents, show a sort of general correspon- 
dence in properties. They are especially suited for fractionations of 
such dyes as the sulphonated phenolic compounds, the distribution 
ratios of these changing greatly with varying hydrogen ion con- 
centrations. Dichlorhy drin, because of its solubility and non- 
volatility, is not very convenient as a solvent; nevertheless, it is 
almost indispensable for the separation of many colors. Anilin is 
an excellent solvent, but usually must be completely removed from 
a color solution before tests are made, and will be employed only 
for a few separations. Both anilin and dichlorhy drin are con- 
veniently removed from water solutions by shaking out with 
carbon tetrachloride 
It may be remarked that in working with the coal-tar dyes the plan 
of acidifying strongly, extracting, and then washing the solvent with 
a more dilute acid, is in nearly all cases preferable to the practice of 
gradually increasing the acidity and using a number of portions of 
solvent. Much more solvent is required in this latter way, with a 
corresponding increase in the proportion of the accompanying 
impurities and in the difficulties caused by emulsification. 
The writer usually prefers to begin the treatment with immiscible 
solvents by shaking out with amyl alcohol from the neutral solution 
after addition of some sodium chlorid. The outline which follows 
will indicate approximately the order in which the solvents will be 
chosen for a complex mixture. 
PROCEDURE OF SEPARATION. 
The solution of the coloring matter, as free as possible from suspended matter, is 
treated carefully with sodium carbonate if it contains free mineral acid or with acetic 
acid if it is alkaline. It should finally be neutral or very faintly acid. Tt should not 
contain the coloring matter in too great concentration, although when working with 
extracts of food products this latter condition is seldom encountered. Concentrations 
of about 0.01 per cent may be taken as most suitable in general, and only in excep- 
tional cases would stronger solutions (0.1 per cent, for instance) be chosen by pref- 
i See W. Reinders and C. Sely, Zeit. fur Chem. und Ind. d. Koll. 13 (1913), 96. 
