154 SMITH— COLUMBIUM AND TANTALUM. [April 13, 



tating potassium columbium fluoride incompletely with ammonia, 

 to get a columbium oxide, which apparently gave no response to the 

 hydrogen peroxide test upon its application, or to the reagent called 

 chromotropic acid. Hence he inferred that he had eliminated the 

 titanium from the columbium. More recent work with large quan- 

 tities of material has demonstrated that in the latter cases it was 

 impossible to entirely remove the metallic acid which gave the 

 color tests. It was further found that by the action of sulphur 

 monochloride upon the oxides of columbium and titanium, cor- 

 responding chlorides were produced ; but again, it proved im- 

 possible to wholly expel the titanium from the columbium by 

 this process, notwithstanding titanium chloride is an exceed- 

 ingly volatile liquid and columbium chloride a solid, crystalline 

 body. 



In the present year, we have, in this laboratory, prepared large 

 quantities of the double fluoride of tantalum and potassium, and 

 found no difficulty whatsoever in eliminating from it every trace of 

 what was supposed to be the titanium double fluoride. 



Having thus, at our disposal, such generous amounts of pure 

 tantalic oxide, free from columbic oxide, in short, really pure tan- 

 talic oxide, it was determined to make a new study of the double 

 fluorides of tantalum with the alkali metals and organic bases. 

 This was undertaken in order to discover, if possible, why Dr. 

 Pennington, when working in this laboratory in 1895, obtained 

 double fluorides of tantalum and columbium with caesium, which 

 showed these unusual formulas: 15CSF. TaF5 and yCsF. CbF5, 

 which varied so widely from those generally followed by the 

 double fluorides of tantalum and columbium, and were not in ac- 

 cord with the law proposed for double halides (Amer. Chem. Jour., 

 v. 291). 



At the outstart it was thought that this re-investigation of the 

 double fluorides would prove to be an easy and simple matter. 

 But it was not long until it was seen that the discordant results of 

 Dr. Pennington were probably due to the fact that there was more 

 than one csesium tantalum fluoride. Indeed, the latest work done 

 in this laboratory, by Mr. C. W. Balke, proves that there are two 

 coasium tantalum fluorides, two rubidium tantalum fluorides, two 

 sodium tantalum fluorides, two ammonium tantalum fluorides, and 

 so forth, of these ratios : 



