I9 o S .] HALL AND SMITH— COLUMBIUM. 205 



columbium hydrate and evaporating the filtrate to dryness and 

 weighing the potassium sulphate after ignition. 



The oxygen and water were determined in another sample by 

 weighing a portion of the substance in a tube sealed at one end, 

 covering it with a plug of ignited asbestos, connecting with a gas 

 burette and igniting. The oxygen was collected and measured, 

 the tube was reweighed, the loss being water and oxygen. The 

 water was obtained by difference. 



Analysis ; 



0.4004 gram of the salt gave o. 1672 gram of oxide and 

 0.2196 gram of K 2 S0 4 



0.4432 gram of the salt lost 0.0470 gram, which contained 17.9 

 c.c. of oxygen at 24 and under 742 mm. pressure, or 0.0229 

 gram, the difference — 0.0241 gram — being water. 



Calculated. Found. 



K 2 S0 4 54. S9 54.84 



Oxide 42.28 41-84 



O (active) 5.05 5.16 



H 2 5.68 5.44 



This salt was obtained and the above composition ascribed to it 

 by Piccini {Zeit. anorg. Chan., 2, 21). He regarded it as a deriv- 

 ative of percolumbic acid and not an addition product of potassium 

 columbium fluoride and hydrogen peroxide, because the water was 

 lost on heating at ioo°, while the oxygen did not escape until the 

 temperature reached 150 . 



On crystallizing this salt from concentrated hydrofluoric acid 

 and hydrogen peroxide in the hope of getting a perfluoride 

 large plates were obtained, which were quite yellow in color with 

 a green tint when dry. They did not seem to differ if little or 

 much hydrofluoric acid was used. The crystals taken for analysis 

 were obtained from a solution consisting of one-half hydrofluoric 

 acid, 48 per cent., and one-half hydrogen peroxide, 3 per cent. 

 They were dried between filter paper and promptly weighed out 

 for analysis. 



Analysis : 



0.7260 gram of salt gave 0.3274 gram of oxide and 



0.3982 gram of potassium sulphate. 



