THE HYDRATE THEORY OF SOLUTIONS 145 



an explanation of certain observations in connection with the 

 density of aqueous solutions of sodium hydroxide. The addition 

 of a solute has the effect of eliminating from the solvent water 

 the more complex molecules, so that even in comparativel) 

 dilute solutions the maximum density can no^onger be observed. 

 There can be little doubt that the abnormal course of the density- 

 concentration curves is due to the same causes, and that the 

 breaks recorded for dilute solutions of sulphuric acid are due, not to 

 the formation of hydrates, but to the recovery of the density curve 

 from the distortion produced by alterations in the character of the 

 complex water-molecules. 



Similar considerations may be applied in the case of the most 

 concentrated solutions, the changes at 97*4 and 93*9 per cent. 

 H 2 S0 4 being attributed to the elimination of complex acid 

 molecules of the type (H 2 S0 4 ) x and not to the production of 

 hydrates with one-sixth, one-third, etc., molecules of water. The 

 net result would then be that the whole of the thirteen breaks 

 between o to 100 per cent. H 2 SO can be accounted for by 

 assuming the existence of four hydrates, together with a certain 

 number of changes due to depolymerisation of the water and 

 acid. These assumptions are fully in accord with present-day 

 conceptions of the nature of solutions, and have the advantage 

 of eliminating most of the difficulties that have hitherto been felt 

 in recognising the actual existence of Pickering's breaks. 



It will be of interest, as confirming the accuracy of these 

 conclusions, to refer briefly to one of the few cases which have 

 been studied with something like the care which was bestowed 

 on the investigation of sulphuric acid. In the case of caustic 

 soda the hydrates which separate on cooling the solutions include 

 compounds with 1, 2, 3-5, 4, 5, and 7 H 2 (Pickering, Trans. 

 Chem. Soc. 1893, 63, 890), and breaks corresponding with most of 

 these were detected iin the somewhat complex density curve 

 (Phil. Mag. 1894, 37, 359). In addition to these, however, breaks 

 were detected at 20 per cent, 12*1 percent, and 3-4 per cent. NaOH, 

 corresponding with the formulae NaOH.9H 2 0, NaOH.i6H 2 0, and 

 NaOH.53H 2 0. In this case a later series of independent density 

 measurements is available (Bousfield and Lowry, Phil. Trans. 

 1905, 204, 265-73), from which it appears that the degree of 

 accuracy attained in the observations in which the breaks were 

 detected was considerably over-estimated ; the absolute error in 

 the concentration of the 50 per cent, stock solution, estimated at 



10 



