28 RESEARCHES UPON ATOMIC WEIGHTS. 



The most probable value for this ratio has been shown by Baxter to be 



57-4453-' 



In the preceding paper Bailey and Fowler's ^ statement that hydrochloric- 

 acid gas is contaminated with volatile compounds of phosphorus, if it is dried 

 with phosphorus pentoxide, is confirmed. Although Bailey and Fowler attrib- 

 ute to hydrobromic acid an effect similar to that of hydrochloric acid, for 

 several reasons it is certain that in the experiments upon cadmium bromide no 

 appreciable amount of phosphorus was introduced into the salt by the action 

 of the hydrobromic acid upon the phosphorus pentoxide. In the first place, the 

 experiment of passing into water hydrobromic-acid gas, formed as in our work 

 by passing nitrogen through bromine and then through an emulsion of red 

 phosphorus in concentrated hydrobromic-acid solution, and dried first by 

 fused calcium bromide and then by phosphorus pentoxide, was performed in 

 this laboratory some years ago in connection with the analysis of cobalt and 

 nickel bromides. In this experiment no phosphorus could be discovered in the 

 aqueous solution. In the second place, in two analyses of bromides which had 

 been heated in hydrobromic-acid gas, the filtrates from the silver-bromide pre- 

 cipitates were evaporated to small bulk and tested for phosphoric acid, with 

 negative results, while the slight residues obtained by filtering the aqueous solu- 

 tions of the original bromides also showed in one case the complete absence of 

 phosphorus, and in the other the presence of only a minute trace of this sub- 

 stance, although in the latter case the salt had been sublimed in a current of 

 hydrobromic acid and therefore contained maximum amounts of phosphorus.^ 

 This result was to be expected from a consideration of the fact that the hydro- 

 bromic-acid gas used in these experiments was diluted with at least twice its 

 volume of nitrogen. In the light of this evidence it seems safe to assume that 

 in the numerous analyses of bromides which have been carried out in this labo- 

 ratory in recent years, no error was introduced by the use of phosphorus pent- 

 oxide as drying agent for the hydrobromic-acid gas. Nevertheless, with more 

 concentrated hydrobromic acid, doubtless it would be unwise to use this drying 

 agent. 



It is interesting to compare the results of the analyses of the different frac- 

 tions of material in this research and in the preceding one. (See table on op- 

 posite page.) 



The close agreement of the results from fraction II by different methods and 

 of the results from all three fractions leaves no doubt of the identity of the dif- 

 ferent specimens of material. 



No matter how the results are averaged, the same conclusion is reached as in 

 the previous paper, i. e., that the atomic weight of cadmium lies very near the 



' Proc. Amer. Acad., 42, 210 {igo6); Jour. Amer. Chem. Soc, 28, 1332; Zeit. anorg. Chem., 

 50, 398. (Seepage 51.) 



2 Chem. News, 58, 22 (1888). 



' Baxter: Proc. Amer. Acad., 39, 248 (1903); Zeit. anorg. Chem., 38, 236. 



