BAXTER. — REVISION OF THE ATOMIC WEIGHT OF IRON. 249 



owing to the fact that ferrous salts precipitate silver from its solutions 

 under certain conditions. Obviously it was necessary to oxidize the 

 ferrous bromide solution before addition of the silver nitrate. Hydrogen 

 peroxide proved too slow an oxidizing agent at ordinary temperatures, 

 potassium permanganate is too vigorous in its action, and sets free bromine 

 from the salt. Even potassium dichromate, at concentrations such as are 

 used in ordinary analytical work, sets free bromine rapidly. But if 

 a very dilute solution of dichromate is used, and if the sulphuric acid 

 is added gradually with the dichromate and only in quantities theoreti- 

 cally necessary, so as to keep the velocity of reaction as low as possible, 

 and finally if the solution of dichromate and sulphuric acid is introduced 

 beneath the surface of the solution of ferrous bromide, so that any 

 bromine set free may be taken up by the ferrous salt before it can escape 

 from the solution, oxidation can be accomplished without the loss of 

 enough bromine to be detected with starch iodide paper. Furthermore 

 complete oxidation is unnecessary, for although dilute solutions of ferrous 

 and silver salts deposit silver upon standing, if considerable ferric salt is 

 present no precipitation of silver takes place owing to the retarding 

 influence of the ferric salt. 



Purification of Materials. 



Pure ferric oxide was prepared in an exactly similar fashion to that 

 used in the previous investigation upon the oxide.* In order to convert 

 this oxide into the metal, one portion was heated in a current of ammonia. 

 But since metal reduced by ordinary ammonia, when dissolved in sul- 

 phuric acid, leaves a residue of carbon owing to the presence of amines 

 in the ammonia, the material used in all analyses except Analysis 1 was 

 reduced in pure electrolytic hydrogen.t 



Silver was purified in the manner usually employed in 'bis laboratory, 

 except that the electrolytic crystals were finally fused in a current of 

 pure hydrogen instead of in a vacuum. Even spongy silver occludes no 

 appreciable amount of hydrogen when heated in this gas, t hence no 

 error could have resulted from this source. 



Bromine was partially freed from chlorine by solution in concentrated 

 aqueous calcic bromide made from a portion of the same sample of 

 bromine. It was then converted into hydrobromic acid by means of 



* Richards and Baxter, These Proceedings, 35, 256. 



t These Proceedings, 34, 357. 



t Baxter, Amer. Chem. Jour., 22, 362. 



