i 4 4 SCIENCE PROGRESS 



thus be grafted on to the original smooth density curve a 

 " hump" precisely similar to that represented by the correcting 

 factor of Riicker's equation. Pickerings conception, if followed to 

 its logical sequel, leads therefore to the conclusion that the formation 

 of the pentahydrate of sulphuric anhydride may be responsible, not 

 only for the break at 59 per cent. H 2 S0 4 , corresponding with the 

 composition of the hydrate, but also for the adjacent breaks at 

 50 per cent, and 72 per cent. H 2 S0 4 . If the breaks at 78 per cent, 

 and at 92 per cent. H 2 S0 4 are attributed in a similar way to the 

 hydrate SO3.2H0O, which produces the break at 84 per cent. 

 H 2 S0 4 , it is possible to account for the whole of the six breaks 

 between 50 per cent, and 92 per cent. H 2 S0 4 in terms of the hydrates 

 H 2 S0 4 . H 2 0, H0SO4 . 2H0O, and H 2 S0 4 . 4HX), which have actually 

 been crystallised out from the solutions. 



If the breaks at 19 per cent., 30 per cent., and 38 per cent. 

 H 2 S0 4 are attributed in a similar manner to the formation of 

 an unknown hydrate containing a larger proportion of water, it 

 is possible to account for all the breaks from 10 per cent, to 

 90 per cent. H 2 S0 4 by assuming the existence of three known 

 and one unknown hydrates, in place of the three known and six 

 unknown hydrates postulated by Pickering. 



The breaks at 3*99 and 9*69 per cent. H 2 S0 4 demand separate 

 consideration. They cannot be accounted for by the formation 

 of hydrates, since compounds such as S0 3 . 52H0O or S0 3 . 132I-LO 

 could scarcely be formed in quantities sufficient to disturb the 

 course of the density curve. Nevertheless the most casual 

 inspection of the curve of the first differentials reveals the fact 

 that the most violent twist occurs at a concentration of less than 

 5 per cent. It is therefore natural to attribute it to alterations 

 in the properties of the solvent, the complex character of which 

 is now well recognised. The presence in liquid water of two 

 kinds of molecules — a dense "water "-molecule and a lighter 

 though more complex " ice "-molecule — was postulated by 

 Rontgen {Wied. Ann. 1891, 45, 91) in order to account for the con- 

 traction which takes place when water is heated from o° to 4°, and 

 Sutherland, who has elaborated this view, has suggested {Phil. 

 Mag. 1900, 50, 460) that the ice-molecules may consist of trihydrol, 

 H 6 3 , and the water-molecules of dihydrol, H 4 2 . The presence 

 of a third kind of molecule of simpler structure and lower density 

 than the typical water-molecule has also been suggested 

 (Bousfield and Lowry, Phil. Trans. 1905, 204, 283) as affording 



