A REVISION OF THE ATOMIC WEIGHT OF BROMINE. 53 



Considerable uncertainty exists as to the purity of the materials employed 

 in much of the foregoing work. Richards and Wells ^ have already exhaustively 

 investigated the various methods of preparing pure silver, and have found that 

 while it is a comparatively simple matter to free this substance from metallic 

 impurities, the absence of gaseous impurities is by no means so easy to secure. 

 Oxygen may be eliminated best by fusion in an atmosphere of pure hydrogen 

 gas,2 or by prolonged fusion in a vacuum, while a lime boat was found to be the 

 most suitable support for the silver during fusion. 



In most of the experiments cited on page 52, one of the final steps in the 

 purification of the silver was fusion of electrolytic crystals on lime, in many 

 cases in a vacuum, but without especial care to prolong the fusion. Silver pre- 

 pared in this way was found by Richards and Wells to contain traces of oxy- 

 gen, derived from silver nitrate occluded by the electrolytic crystals. In cases 

 8, 9, 10, 17, and 18, however, the silver was fused in hydrogen. Richards and 

 Wells showed also that Stas's silver contained at least o.oi per cent of impurity, 

 since it yielded o.oi per cent less silver chloride than their purest silver.^ Scott's 

 silver in three cases was merely heated, not fused, in hydrogen, and in two of 

 the others was fused before a blowpipe on calcic phosphate. In one experiment 

 only the metal was fused on lime. No details are given as to the purification 

 of the silver used by Marignac. 



Bromine also may be freed from impurities only with some difficulty. Ex- 

 perience in this laboratory has shown that chlorine may be eliminated most 

 conveniently by distilling or precipitating the bromine from solution in a bro- 

 mide. One such distillation is sufficient to remove chlorine completely only 

 when the substance is initially comparatively pure. If, however, the process 

 is repeated by converting a portion of the partially purified product into a bro- 

 mide, and dissolving the remainder of the bromine in this comparatively pure 

 bromide, the chlorine is eliminated so completely that further repetition of 

 this process has no apparent effect.^ The removal of iodine may be easily 

 effected by converting the bromine into hydrobromic acid or a soluble bromide, 

 and boiling the solution with a small quantity of free bromine. Here again it is 

 well to repeat the process several times, since the reaction between free bro- 

 mine and the iodine ion, like that between free chlorine and the bromine ion, 

 is undoubtedly incomplete. 



The greater part of the results given on page 52 were obtained with bro- 

 mine which had been purified with due observance of these precautions. Of 

 the other investigators, Stas seems to have been the only one to use sufficient 



^ Pub. Car. Inst. No. 28, 16; Jour. Amer. Chem. Soc, 27, 472; Zeii. anorg. Ghent., 47, 70. 

 2 Baxter: Proc. Amer. Acad., 39, 249 (1903); Zeii. anorg. Chem., 38, 232 (1904). 

 ^ Loc. cit., page 62. 



* Attention has already been called to these points by Richards and Wells: Proc. Amer. 

 Acad., 41, 440 (1906); Zcit. physikal. Chem., 56, 354. 



