A REVISION OF THE ATOMIC WEIGHT OF BROMINE. 59 



and the rinsings were added to the evaporated filtrate and wash-waters. The 

 whole was then tested in a nephelometer for silver and the quantity found was 

 estimated by comparison with standard silver solutions. In most cases the 

 correction thus obtained was less than o.i mg. 



The asbestos which formed the felt in the Gooch crucible, after having been 

 shredded, was digested for some hours with aqua regia and was then thor- 

 oughly washed with water. Before the empty crucible was weighed, the felt was 

 ignited with a Bunsen burner. Crucibles thus treated and then, after being 

 moistened with water, again heated to i8o°, did not change in weight. 



In the table on page 58 are cited all the analyses which were completed with- 

 out accident. Vacuum corrections of —0.000031 for every apparent gram of 

 silver and of -^o.oooo4I for every apparent gram of silver bromide are applied.^ 



The platinum-plated brass weights were standardized from time to time and 

 were found to retain their original values within a very few hundredths of a 

 milligram in all cases. 



CONVERSION OF SILVER BROMIDE INTO SILVER CHLORIDE. 



The ratio of silver bromide to silver chloride was determined much as de- 

 scribed in previous papers upon the atomic weight of iodine.^ Pure silver 

 bromide was prepared by precipitation of silver nitrate with an excess of 

 ammonium bromide. The silver employed was purified either by precipitation 

 as chloride and reduction with invert sugar, or by electrolysis, or by precipita- 

 tion with ammonium formate. The metal was then fused before a blowpipe 

 upon a crucible of the purest lime, and the buttons were thoroughly cleansed 

 with nitric acid. No further purification was considered necessary since the 

 weight of the metal was of no consequence. 



After the silver bromide had been washed by decantation with water, in some 

 cases it was collected in a Gooch crucible in which a disk of filter paper was 

 employed instead of asbestos, and after drying at 100° it was carefully sepa- 

 rated from the filter paper. In other cases the precipitate was transferred to a 

 platinum dish, and was drained with a platinum reverse filter ^ with a disk of 

 filter paper. In still others a platinum Gooch crucible with small holes was 

 found to belsufficiently effective as a filtering medium without the use of either 

 asbestos or filter paper. 



Before being weighed the silver bromide was fused in a current of air saturated 

 with bromine in a weighed quartz crucible. The air was purified by passing 

 successively over beads moistened with silver nitrate solution, over sodium 

 carbonate, and finally over concentrated sulphuric acid which had been heated 

 to its boiling-point with a small quantity of recrystallized potassium dichromate 

 to eliminate volatile and oxidizable impurities, The air was then passed 



' See page 41. 



"^ Baxter: Ptoc. Amer. Acad., 40, 432 (1904); 41, 75 (1905); Jour. Amer. Ohem. Soc, 26, 

 1590; 27,876; Zeit. anorg. Chem., 43, 27; 46,36. (See page 102.) 

 ' Cooke: Proc. Amer. Acad., 12, 121 (1876). 



