RICHARDS AND WELLS. — TRANSITION TEMPERATURE. 439 



20°, to which the reaction heat corresponds. In the next place, the 

 transition is not a simple melting, although, indeed, it is true that the 

 equation of vaii't Iloff applies as a close approximation. It is not, how- 

 ever, worth while to attempt a closer calculation of these quantities for 

 the reason that further elaborate experimentation would be required in 

 order to demonstrate the disposition of the sodic chloride in this mixture. 

 It is not known whether or not the salts are isomorphously mixed in 

 both solid phases, — a circumstance presenting a series of complications 

 quite untraceable on the basis of data known at present. However, the 

 above calculation probably shows the maximum effect to be produced by 

 a substance with even as low a molecular weight as 58.5. Of course, 

 substances with higher molecular weights would produce proportionately 

 smaller effects for a given percentage composition, and therefore be even 

 less objectionable for the present purpose. Practically we found that 

 sodic bromide containing about .02 per cent of sodic chloride actually 

 "melted" at a temperature only about 0.013 or 0.014 too low. 



It is worth while now to compare the effect of impurity in sodic 

 bromide with the effect of impurity in water. The lowering caused by 

 a gram molecule of non-electrolyte in a litre of water is 1.86°, therefore 

 an impurity of .01 per cent of the electrolyte common salt would cause a 

 change of 0.006°. This is a somewhat smaller effect of impurity than 

 in the case of the sodic bromide ; the comparison emphasizes strongly 

 the importance of purity in the materials in the latter case. 



Preparation of Sodic Bromide. 



The preparation of pure sodic bromide proved to be unexpectedly 

 difficult, but the difficulties have been surmounted, and the result of the 

 present work must be considered as essentially final. The problem is 

 harder than that involving sodic sulphate, and we cannot recommend 

 sodic bromide with quite the unqualified approval which the other salt 

 received; nevertheless, although not so easy to repeat, the preparation 

 is by no means impossible, and the results will certainly serve a useful 

 practical purpose. 



At first it was attempted to purify the two different specimens of 

 commercial bromide by recrystallization as dihydrate with the hope that 

 the chloride would be thus easily eliminated, because of the very low in- 

 version temperature of the corresponding sodium salt. However, the 

 result in each case was very disappointing. The bromide, although 

 increasing in purity with each successive crystallization as shown by 



