Hydrates of Hydrogen Bromide. 117 



which represents 3"99H 2 0, and the crystallizing substance is, 

 therefore, the tetrakydrate, the existence of which was indicated 

 by the other break in the densities. Its freezing-point is 

 -55*8. 



From 47° to 38° per cent, no freezing-points were obtained 

 at temperatures about —80°, nor, in the case of a 43 per cent, 

 solution, even at —95°. If it is permissible to venture a 

 conjecture, this gap might perhaps, if adequate cooling were 

 adopted, be filled up by the crystallization of a pentahydrate. 

 ~No definite break was observed, it is true, in the densities at 

 the point, and the determinations would have had to have 

 been a good deal more accurate than they were in order to 

 indicate one here as well as at 56 per cent. ; but there was a 

 general change in direction somewhere in the neighbourhood. 

 The chief ground, however, for imagining a pentahydrate to 

 exist is that a solution of that composition (47*4 per cent.) 

 possesses remarkable stability, its composition is affected to 

 but a very slight degree (1 per cent.) by boiling under con- 

 siderable ranges of pressure (750 to 1950 millim.), and also 

 that in the case of a solution of hydrochloric acid which 

 behaves in a similar manner, though to a less marked extent, 

 there is independent evidence to show that this behaviour is 

 due to the presence of a definite hydrate (Ber. xxvi. 

 pp. 279, 282). 



At 38 per cent, the crystallization of water begins, and the 

 freezing-points gradually rise to 0, the only feature of interest 

 in this curve being the existence of a break at 25 to 30 per. 

 cent, agreeing with that indicated by both Topsoe's and 

 my own density results. 



The two new isolated hydrates here described form the 

 fifth and sixth instances of hydrates isolated in cases where 

 the only grounds for regarding their existence as probable 

 were that they were indicated by changes of curvature in the 

 figures representing the properties of the. solutions. The 

 isolation of so many predicted hydrates would, even if no 

 other arguments existed, be sufficient to place the reality and 

 significance of these breaks beyond doubt, and the two special 

 instances here described are of particular importance in that 

 neither of the breaks were of at all a marked character, this 

 showing that minor breaks have the same significance as the 

 more marked ones * . 



* It must be noticed, however, that in comparison with results such 

 as those in the case of sulphuric-acid densities the experimental error here 

 is very large, and that these breaks might be very prominent if more 

 exact determinations were possible. They are, on the other hand, 

 decidedly minor breaks in comparison with those indicating the isolated 

 hydrates of nitric and of hydrochloric acid. 



