116 Mr. Spencer U. Pickering on the 



remark that although there must of course be a change of 

 curvature at the point where one crystallization gives place 

 to another, yet these changes are totally different from those 

 previously discussed, and. do not necessarily imply any special 

 changes in the solutions at the points where they occur ; 

 they are changes in the actual phenomena occurring, whereas 

 the others are changes in the rate at which one particular 

 phenomenon or property is influenced by changes of strength 

 of the solution. 



The strongest solution used contained nearly 69 per cent, 

 of acid, and was made by passing the gas into some of its 

 solution cooled in a freezing-mixture. (The solubility of the 

 gas, I may mention by way of parenthesis, is but very slightly 

 affected by temperature.) The whole liquid soon solidified 

 into a mass of transparent gritty crj^stals. These could not 

 be melted (in order to take their melting-point) without 

 giving off a small amount of gas, so that the melting-point 

 determined is slightly below that of the pure crystals. Adding 

 more and more water, the series of freezing-points shown by 

 the curve marked 2H 2 was obtained. The maximum of the 

 curve, as determined by drawing a smoothed curve through 

 the experimental points with a bent lath, is at about 69 per 

 cent., and the hydrate is, therefore, the dihydrate, which 

 contains 69*23 per cent. This hydrate was obtained by 

 Berthelot {Ann. Chim. Phys. [5] xiv. p. 369), and examined 

 by Roozeboom [Rec. Chim. des Pays-Bas, iv. p. 331), who 

 gives its melting-point as — 11 0, 3. The present determina- 

 tions give it as — 11°*2. 



At about 60 per cent, the appearance of the crystallization 

 suddenly changes, the crystals being now much finer and less 

 gritty than those of the dihydrate ; the freezing-points at 

 the same time cease to fall rapidly, and form a new curve 

 marked 3H 2 in the figure. This curve has not been followed 

 on the right-hand side beyond its maximum, owing to the 

 crystallization of the dihydrate interfering with doing so, 

 but the maximum can be located with a fair amount of 

 certainty by a bent-lath curve drawn through the experi \ 

 mental points, and is thus found to be at 59'6 per cent., corre- 

 sponding to HBr, 3*05H 2 O. This new hydrate is, therefore, 

 the trihydrate (containing 60*00 per cent.), the existence of 

 which was indicated by the break in the densities. Its melting- 

 point is —48°. 



At 65 per cent, the crystallization again changes in a very 

 marked manner, the crystals now being larger and more 

 transparent: the freezing-points rise at first and then fall. 

 The position of the maximum in this case is at 53 per cent., 



