THE HYDRATE THEORY OF SOLUTIONS 125 



solution with that of the salt at very great dilutions. On the 

 other hand, it had been found by van t'Hoff and others that, 

 in contrast with non-electrolytes, salt solutions exhibited 

 abnormally high osmotic pressures, just as would be the case 

 if the constituent ions were capable of acting separately in 

 contributing to the osmotic pressure of the solution : by 

 following up this clue it was possible to deduce, either from 

 direct measurements of osmotic pressure, or indirectly from 

 observations of boiling-point and freezing-point, a second 

 series of values for the co-efficient of ionisation of a salt. 

 Although much collateral evidence was adduced in favour of 

 the view that the salt-radicles existed independently in solution, 

 the approximate agreement between the values for the co- 

 efficient of ionisation deduced by the two methods was an 

 experimental result of the utmost importance, and undoubtedly 

 afforded one of the strongest arguments in favour of the 

 dissociative theory. 



In the case of the hydrate theory, on the other hand, the 

 existence of a reverse action did not form an essential feature, 

 but was rather a disturbing factor, the existence of which was 

 somewhat grudgingly admitted. Like many other theories this 

 conception of the nature of solutions had existed in a nebulous 

 form for many years prior to the publication of the paper in 

 which it first took definite shape. The idea that an aqueous 

 solution might contain compounds of solvent and solute, 

 similar to those which frequently separate when the solution is 

 allowed to crystallise, was widely disseminated and had formed 

 the subject of frequent and prolonged discussion. Mendeleef s 

 paper was, however, remarkable in that it introduced a method 

 by which the nature and composition of these hydrates could 

 be determined from a study of the solution itself. He thus 

 gave to the theory a definite quantitative character, and brought 

 it, practically for the first time, within the region of exact 

 experimental investigation. In measuring the densities of 

 aqueous solutions of sulphuric acid and of alcohol, he had 

 found that the curves showing the relationship between con- 

 centration and density were not continuous, but exhibited 

 abrupt changes of direction at certain points. In the case of 

 sulphuric acid, these changes occurred at the compositions 

 3*5, 47-6, 73*9, 84/5, and 100 per cent. H 2 S0 4 , corresponding 

 with the formulae H 2 S0 4 . 150HA H 2 S0 4 . 6H 2 0, H 2 S0 4 .H 2 0, 



