ON THE SOLUBILITY OF SALTS. 299 



solvent with equal force. It may be therefore that some salts do actually 

 retain an equivalent of their water of crystallization attached to them in 

 solution. Thomson's experiments lead to the conclusion that the third mole- 

 cule of water in terhydrated phosphoric acid is not held so firmly as the 

 others. On the other hand, Gerhardt has shown that when three equivalents 

 of the strong base soda are substituted for the three of water in terhydrated 

 phosphoric acid, the salt formed, 3NaO, P0 5 , has the power of retaining an 

 equivalent of HO at the temperature of 100° Cent., which can only be 

 removed at a temperature approaching redness, while the salt 2NaO, HO 

 F05 + 24HO loses the whole of its crystalline water at 100°. It is probable 

 that the equivalent of water thus retained, performs the same function as the 

 constitutional water of the magnesian sulphates, and mav be replaced by a 

 salt molecule. Such salts as SLiO, P0 5 + 2LiO, HO, PO. + 2HO and 3PbO 

 Pb0 3 + PO, N0 5 -f 2HO &c, may perhaps be referred to this type. 



If we possessed some method of measuring the force with which salts are 

 held in solution, we should be able to discover the law of its variation in the 

 same class of salts according as one base was substituted for another, and for 

 the same salt according as the temperature varied. The force with which 

 a salt may be held by the solvent does not seem to bear much relation 

 to its solubility. Charcoal removes salts of equal solubility from solutions 

 in different proportions. It may, however, be justly objected to this, that 

 the result is not due to a difference between the forces with which the salt 

 and water are held together, but rather to the greater power of adhesion which 

 the charcoal possesses for certain salts. I have begun a series of compara- 

 tive experiments on the relative amount of different salts which pure charcoal 

 and spongy platinum retain when solutions of various degrees of concentra- 

 tion are passed through them, the results of which will, I hope, throw some 

 light on this subject. 



The adoption of the hypothesis that crystalline salts are the result of 

 double decomposition, appears to involve another consequence of great im- 

 portance in the present inquiry, namely, that double salts do not exist as 

 such in solution. In the case of double salts formed by the substitution of 

 constitutional water, it is by no means necessary, as I have shown above, that 

 the constituent salts of the double salt should be considered as existing un- 

 combined in the solution, but simply in a different state of molecular arrange- 

 ment. Favre and Silbermann were of opinion, however, that they did not 

 exist in combination, and were only formed at the moment of their crystalli- 

 zation. Graham, from his experiments on diffusion, has likewise suggested 

 the possibility of the constituents of a double salt being dissolved together in 

 water without being chemically combined. The latter considered that diffu- 

 sion was capable of producing chemical decomposition ; for instance, that of bi- 

 sulphate of potash, alum, and sulphate of potash in lime-water. This opinion 

 harmonizes with the hypothesis of Favre and Silbermann ; and if understood 

 in the modified form above suggested, namely, that the double salt, such as 

 we know it in its crystallized state, does not exist in the solution, but that 

 the constituent salts are nevertheless not free in every case, would be per- 

 fectly accordant with it. Many isomorphous salts combine together in 

 almost every proportion, such for example as the chromates and sulphates. 

 H. Rose has analysed a compound of nitrate of silver and nitrate of soda to 

 which he assigns the formula AgO, NO 5 + 10NaO, N0 5 . If we adopt the 

 hypothesis that crystallized salts are compounds of several molecules of the 

 simple salt, we may consider the salt in question as n (NaO, N0 5 ), in which 

 one or more molecules of nitrate of silver replace a corresponding number of 

 nitrate of soda. In accordance with that hypothesis, such a salt could not 



