A REVISION OF THE ATOMIC WEIGHT OF MANGANESE. 45 



discussed elsewhere (page 14), and in the same place it has been shown that 

 hydrochloric-acid gas attacks concentrated sulphuric acid to so slight an ex- 

 tent that the latter may safely be used to dry the former. 



THE METHOD OF ANALYSIS. 



The principal point of difference between the analyses of manganous bro- 

 mide and manganous chloride was occasioned by the greater solubility of silver 

 chloride. This introduced no difficulty in the case of the titrations, although 

 the opalescence in both nephelometer tubes was of course more marked at the 

 end-point. In order to make the precipitation of the silver chloride more 

 complete, however, a much larger excess of silver nitrate was necessary than in 

 the case of silver bromide, about 0.15 gm. being added in each analysis. Fur- 

 thermore, the precipitate was washed at first with a solution of silver nitrate 

 containing 0.04 gm. in a liter, although the final six washings were performed 

 with water. The silver chloride dissolved in both filtrate and washings was de- 

 termined by comparison with standard solutions in a nephelometer as previously 

 described. Here even the filtrate showed a trace of dissolved silver chloride. 



In the case of manganous chloride also the extent of the adsorption of man- 

 ganese compounds by the filter-paper was investigated by filtering a solution 

 of the salt through a filter-paper, and, after thoroughly washing the paper, de- 

 termining the residue after ignition. The correction for adsorption thus found, 

 0.00006 gm., is in some cases larger than the weight of the residue minus the loss 

 in weight of the boat, owing possibly to the fact that not all the platinum lost 

 by the boat is collected upon the filter-paper. At any rate the uncertainty in- 

 troduced by applying the correction wherever possible is very small. 



The filter upon which the asbestos shreds were collected was washed with 

 warm dilute hydrochloric acid to eliminate adsorbed manganese compounds, so 

 that no correction for adsorption is here necessary. This acid was then tested 

 for silver chloride with the nephelometer, and if any was found it was added to 

 the main weight of this substance. 



In analyses i, 2, 3, 4, and 7 the silver nitrate was added to the manganous 

 chloride, while in analyses 5 and 6 precipitation was performed in the reverse 

 fashion. 



The possibihty of the existence of manganic compounds in the salt was con- 

 sidered. In order to determine whether or not this was the case, one specimen 

 of manganous chloride, after fusion in hydrochloric acid, was tested for man- 

 ganic compounds by adding a solution of starch paste and potassium iodide. 

 No coloration resulted even after adding hydrochloric acid, although a mere 

 trace of permanganate produced immediate coloration in the same solution. 



The fused bromide and chloride were both pink, hence Forchhammer's state- 

 ment that the pink color of manganous salt is due to the presence of manganic 

 compounds is incorrect.' 



' Ann. Phil, N. S., i, 50 (1821). 



