BANCROFT. — TERNARY MIXTURES. 325 



whether liquid or solid in the pure state, behaves like a gas at that 

 temperature. If applied to any dissolved substance, the statement 

 just quoted is too inaccurate to need any comment. The precipitation 

 of salts by alcohol is a well known instance where it does not apply, 

 and, in general, adding to a solution a substance in which the solute is 

 practically insoluble diminishes the solubility of the latter. This is 

 recognized by Nernst, for he has based a method for determining re- 

 acting weights upon it.* Even if limited to solids, the proposition can- 

 not be admitted. We have the precipitation of lactones by potassium 

 carbonate as an intermediate step, and the precipitation of salts by 

 phenol as a definite case of diminished solubility without the presence 

 of a common ion. Other cases could be cited, if necessary, and there 

 are also examples where an increase of solubility takes place when a 

 solid substance is added to a solution containing another solid as solute. 

 The explanation usually offered under these circumstances is, that 

 " double molecules " are formed, a mode of getting round the facts 

 which is not always entirely satisfactory. 



Since in the application of the gas laws to solutions there has been 

 observed no difference between a solid and a liquid when dissolved, 

 I am inclined to think that the general statement should be, that in 

 all cases where a third substance, B, is added to a solution of ^ in S, the 

 solubility of A undergoes a change. This variation may be large or 

 small, positive or negative, depending on the nature of the three sub- 

 stances A, B, and S. When both A and B are liquids, or even when 

 only one of them is, the effect is so marked as to be familiar to all ; 

 when both are solids, the effect is not yet recognized by so competent 

 an authority as Nernst. 



The work of the last few years on solutions has been devoted to 

 bringing out the analogy between the dissolved substance and gases. 

 In the cases of changed solubility, no common ion being present, the 

 analogy is no longer with gases, but with liquids. The added sub- 

 stance acts as a liquid, precipitating the solute more or less in propor- 

 tion as the dissolved substance happens to be more or less soluble in it. 

 The laws governing these displacements are entirely unknown, with 

 the exception of Nernst's Distribution Law,t which is only a first ap- 

 proximation, in that it takes no account of the changing mutual 

 solubilities of the hypothetically non-raiscible liquids. Under these 

 circumstances it seemed to me desirable to investigate the laws gov- 

 erning systems composed of three substances, and the experiments 



* Zeitschr. f. ph. Chem., VI. IG. 1890. t Teilingssatz. 



