■ BAXTER. — A REVISION OF THE ATOMIC WEIGHT OF IODINE. 433 



after the usual treatment were re-fused in air with a loss in weight of 

 only 0.00002 gram. The same chloride, when fused and cooled in chlorine, 

 gained five milligrams. 



It was soon discovered that a porcelain crucible, when used for the 

 conversion of silver iodide into silver chloride, gradually gained in weiglit. 

 In the first experiment this gain was over three milligrams, and in several 

 subsequent experiments with the same crucible, amounted to slightly less 

 than a milligram in eacli case. This gain did not take place to the 

 slightest extent when the crucible was heated alone in chlorine, nor after 

 the reaction was complete, for the weight of the crucible and silver chlor- 

 ide very soon became constant. Probably it was due partially to solution 

 of the silver salt in the glaze, as was shown by slight discoloration on the 

 bottom of the crucible. Possibly, however, it was caused by the attacking 

 of the glaze by the " nascent" chlorine and iodine set free in the process, 

 perhaps with evolution of oxygen. At any rate too great an uncertainty 

 existed as to the weight of the crucible at the end of the experiment. 

 Accordingly quartz crucibles were next employed for the same purpose. 

 These crucibles behaved in an ideal fashion, for they remained practically 

 constant in weight through the process. It was only necessary to rotate 

 the crucibles during the solidification of the fused salts in order that the 

 salt might solidify in a thin layer on the sides of the crucible. Neglect 

 to do this almost invariably resulted in the cracking of the crucible. 



After the first weighing of the crucible, it was again heated in chlorine 

 for an hour, and again cooled and weighed. In no case did a loss in 

 weight of more than a tenth of a milligram take place, showing both that 

 the iodine was completely replaced, and that no silver chloride had vola- 

 tilized. 



Four determinations (Analyses 20 to 23) were completed with quartz 

 crucibles in the manner described. Two more determinations (Analyses 

 24 and 25) were made by heating silver iodide first in a current of carbon 

 dioxide and bromine, and then in chlorine. The weight of the silver 

 bromide was obtained, but although the results for the atomic weight of 

 iodine calculated from the weight of the bromide agreed closely with those 

 obtained by the other methods, on account of some uncertainty as to the 

 purity of the bromine, these results are withheld until the experiments can 

 be repeated. The only possible impurity in the bromine, chlorine, would 

 not affect the final weight of the silver chloride. 



At the time these experiments were performed, the atomic weight of 

 chlorine was thought to be 35.456. The atomic weight of iodine calcu- 

 lated from the ratio between silver iodide and silver chlorine, using this 



VOL. XL. — 28 



