A REVISION OF THE ATOMIC WEIGHT OF MANGANESE. 35 



num flask. The resulting manganous carbonate, after thorough washing with 

 water containing a small amount of ammonia to prevent colloidal solution of 

 the carbonate, was readily converted into bromide by solution in hydrobromic 

 acid. Since it was probable that the carbonate contained occluded nitrate, and 

 since a portion of the material had been oxidized to the manganic state during 

 the washing, it was obvious that bromine would be set free during the solution 

 in hydrobromic acid. The use of a platinum vessel for this purpose was there- 

 fore precluded. In order to avoid the introduction of silica, fused quartz 

 dishes were employed, instead of glass vessels. The former have been shown 

 to be practically insoluble in acid solutions.^ 



The free bromine was expelled from the solution of manganous bromide by 

 prolonged heating on a steam-bath in a quartz dish. Finally it was crystallized 

 six times, thrice in quartz, and, after filtration with a platinum funnel, thrice in 

 platinum with centrifugal drainage after each crystallization. The crystals 

 were dried as far as possible over stick potash in a vacuum desiccator. From 

 the mother-liquors, by means of six similar crystallizations, Sample A2 was 

 obtained. 



In the conversion of Sample B from permanganate to bromide minor changes 

 were introduced. The ammonium carbonate was prepared in a pure state by 

 distilling a solution of commercial ammonium carbonate in a platinum still. 

 Instead of expelling free bromine from the solution of manganous bromide by 

 prolonged heating upon the steam-bath, the solution was evaporated as far as 

 possible upon the steam-bath and the residue was heated to 200° in an electric 

 oven. The bromide was dissolved in water, and after filtration of the solution, 

 was crystallized three times in a platinum dish. The third crop of crystals is 

 designated as Sample B. 



Sample C was prepared from a commercial specimen of pyrolusite. This was 

 first dissolved in hydrochloric acid and the solution was boiled to expel chlorine. 

 Hydrogen sulphide was passed into the diluted solution of manganous chloride 

 to saturation, and the precipitate of sulphur and sulphides was removed by fil- 

 tration. After the excess of hydrogen sulphide had been expelled by boiling, the 

 solution was fractionally precipitated with sodium hydroxide until the precipi- 

 tate was free from iron. Finally, the manganese was precipitated with ammo- 

 nium carbonate and the precipitate was washed and dissolved in nitric acid. 

 The nitrate was recrystallized and converted into bromide exactly as in the case 

 of Sample A. 



The source of Sample D was Merck's "chemically pure" manganous sul- 

 phate. A solution of 500 gm. of this salt was first saturated with hydrogen sul- 

 phide, and the precipitate, which consisted chiefly of manganous sulphide, was 

 removed by filtration. After the addition of a small amount of ammonia, hy- 

 drogen sulphide was again passed into the solution to saturation, and the pre- 



1 Mylius and Meusser: Zeit. anorg. Chem., 44, 221 (1905). (See also page 36 of this paper.) 



