I9 o 5 .j HALL AND SMITH— COLUMBIUM. 211 



DIFFERENCE IN SOLUBILITY OF DOUBLE FLUORIDES. 



It is of interest to note that the solubility of potassium titanium 

 fluoride is increased upon the addition of hydrogen peroxide, while 

 that of the potassium columbium oxyfluoride is decreased. In 

 hydrofluoric acid this order is reversed, the columbium salt becom- 

 ing more soluble and the titanium salt less soluble. This suggests 

 alternating these solvents in the crystallization of columbium and 

 potassium double fluorides as one of the best means for removing 

 titanium. 



Recrystallization from hydrofluoric acid in the form of K 2 CbF. 

 will remove tin and probably also tungsten from impure potassium 

 columbium oxyfluoride. Two recrystallizations from that solvent 

 are sufficient to give an oxide, the ignition of which in a platinum 

 crucible gave no stain on the crucible. If partially dried oxide 

 wrapped in the filter paper be ignited directly in a platinum 

 crucible the presence of a stain on the crucible after removing the 

 oxide will be a very delicate test for tin. It is likely that when tin 

 is removed by crystallization tungsten is also, if they are present in 

 about equal amounts and in such cases where the total amount is 

 very small. The procedure would remove the necessity for the 

 tedious sodium carbonate and sulphur fusions used in this work. 



BEHAVIOR WITH PRECIPITANTS. 



Pennington {/our. Amer. Clicm. Soc, 18, 38) noted that 

 disodium hydrogen phosphate gave no precipitate in a solution of- 

 potassium columbium oxyfluoride, while it completely precipitated 

 titanium from a solution of its double fluoride. This was studied 

 briefly in order to determine if it might not serve as a quantitative 

 separation of columbium from titanium. It was found that when 

 the reagent was added to a solution containing a large excess of 

 columbium and only a little titanium no precipitate was produced 

 even on prolonged boiling. If the amount of titanium was increased 

 slightly both the titanium and the columbium were completely pre- 

 cipitated by the disodium hydrogen phosphate. This reagent, 

 therefore, does not separate the two elements. 



Geisow found that an alkaline formoxime solution precipitated 

 zirconium and titanium, but did not precipitate columbium. 



Formoxime, or its polymerization product, was prepared by 

 bringing together solutions of the calculated quantities of formal- 



