74 PROCEEDINGS OP THE AMERICAN ACADEMY. 



ing the atomic weight of bromine to be 79.955, the value found by 

 Stas and repeatedly verified by experimenters in this laboratory,* these 

 two determinations give the values 126.987 and 126.985. The results 

 obtained by reference both to chlorine and to bromine thus became 

 almost identical. However, in order to conlirm these two determinations 

 with bromine, further experiments were carried out in the same manner 

 as before, as follows : 



Pure silver iodide was prepared as described in the previous paper, 

 by precipitating pure silver nitrate with a considerable excess of the 

 purest amnionic iodide,f washing the precipitate with one per cent nitric 

 acid and pure water, and drying the iodide, in an air bath, and then by 

 fusion in a tared quartz crucible, first in an atmosphere containing free 

 iodine, and finally in air to expel dissolved iodine. Fusion in an atmos- 

 phere containing iodine must have completely converted occluded silver 

 nitrate into silver iodide. The iodide was then heated in a current of air 

 and bromine. The air was purified by being conducted successively over 

 beads moistened with silver nitrate solution, sodium carbonate, and finally 

 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. Four different samples of bromine 

 were employed, each one of which had been thrice distilled from a solution 

 of a bromide, in each distillation the bromide having been made from a 

 portion of the product of the previous distillation. Chlorine must have 

 been completely separated in this way. The presence of iodine in the 

 bromine was of less importance. Nevertheless each sample, while in 

 the form of hydrobromic acid, was freed from iodine by boiling with 

 several small portions of potassic permanganate. The apparatus for 

 purifying the air and saturating the air with bromine was so constructed 

 that the gases came in contact only with glass. The quartz crucibles 

 were always contained in large porcelain crucibles during the heating in 

 bromine, as well as in the initial fusion of the silver iodide with iodine. 

 Volatilization of the silver halides was prevented by heating very gently 

 at first, at a temperature insufficient to fuse the mixture, then, when the 

 greater part of the iodine had been replaced, until the silver bromide 

 barely fused. Furthermore, the crucible was very deep and the current 

 of gases very slow, so that any volatilized silver halides had opportunity 

 to condense upon the cool walls of the crucible. That no loss from vola- 



* Richards, Proc. Amer. Phil. Soc, 43, 119. 



t The method of purifying the iodine is described briefly on page 77 of this 

 paper. 



