THE HYDRATE THEORY OF SOLUTIONS 137 



process of distilling off the water. But, while he thus paid 

 formal homage to the idea of mass-action, the effects which it 

 produced were treated as of very slight significance. The 

 points obtained by differentiating the experimental numbers 

 were represented by sharply intersecting curves, and their 

 essential continuity was treated as a matter of remote and 

 theoretical, rather than of immediate and practical, importance. 

 It is, moreover, noteworthy that the breaks which he recognised 

 were not limited to a small number, such as might be produced 

 by the formation of a few hydrates of exceptional stability, but 

 were recorded at sixteen different concentrations between 0*104 

 and 99-493 per cent. H 2 S0 4 . The admission that dissociation 

 might occur was thus accompanied, not by a restriction of 

 Mendeleefs method, but by its application to cases even more 

 extreme than those which Mendeleef had himself considered. 

 It will therefore be of interest to consider, both from the 

 theoretical and from the experimental standpoint, the extent to 

 which— in spite of its theoretical limitations — the method may 

 still be utilised for the detection of hydrates in solution. 



It is obvious from everyday experience of phenomena such 

 as the efflorescence of hydrated crystals of sodium sulphate or of 

 sodium carbonate, that the stability of a hydrate decreases as it 

 becomes more and more complex. This general view is con- 

 firmed in detail by measurements of vapour-pressure, such as 

 those recorded above for the hydrates of copper sulphate, and 

 also by the thermo-chemical data. It is evident, for example, 

 that although the hydrate S0 3 . H 2 0, in the formation of which 

 19,400 calories are liberated, may be sufficiently stable to melt 

 without dissociation, this is less likely to be true of the hydrate 

 S0 3 . 2H 2 0, in the formation of which an additional 6,400 calories 

 only are liberated. The hydrates S0 3 .3HX> and S0 3 .5H,0, 

 which have also been isolated in a crystalline state, are probably 

 still less stable compounds since the additional heat liberated in 

 their formation amounts to only 3,040 and 2,700 calories. Whilst, 

 therefore, the formation of the monohydrate and dihydrate may 

 give rise to more or less abrupt changes in the physical 

 properties of the solution, the effects produced by the formation 

 of the more complex hydrates are likely to be spread over such 

 a range of concentration as to be recognised only with difficulty. 

 In the extreme case of the supposed hydrate with 5 i82H 2 0, the 

 heat of formation amounts only to three or four calories per 



