448 PROCEEDINGS OF THE AMERICAN ACADEMY. 



Summary. 



The results of this paper may be summed up in the following 

 sentences : 



1. Pure sodic bromide is not to be obtained by recrystallizing the 

 ordinary commercial samples, but must be made from pure bromine and 

 pure sodic carbonate. 



2. Prepared in this way, our salt upon analysis was found to corres- 

 pond very closely with the new value of the atomic weight of sodium, 

 23.008, if silver be taken as 107.93 and bromine 79.955. Therefore, it 

 was presumably pure. 



3. The less pure material on successive recrystallization gave in 

 every case a slightly rising transition temperature as the crystallization 

 proceeded. Only the purest material melted at a perfectly constant 

 point, therefore constancy of melting point is an indicator of purity ; 

 but it is safer to analyze the salt as well. 



4. When all precautions are taken it is possible to duplicate the 

 results for the transition temperature with samples of salt prepared in 

 different ways and at different times without great difficulty ; a value 

 within .01° of the truth may be easily obtained, and further precautions 

 enable one to approach much closer. Therefore, the point is one suit- 

 able to use in the calibration of thermometers, although its determina- 

 tion requires more chemical skill than that involving sodic sulphate. 



5. The actual value of the transition temperature on the international 

 hydrogen scale is 50.674°. 



