Penfield and Ford — Stibiotantalite. 73 



ignited in a current of carbon dioxide and weighed as Bi 2 S 8 . 

 The antimony separated from the bismuth was reprecipi- 

 tated as sulphide, collected on asbestos, and, after heating 

 in carbon dioxide, weighed as Sb 2 S 8 . Preliminary experi- 

 ments with known amounts of bismuth and antimony proved 

 that the methods as given above yielded correct results. The 

 filtrate from the original sulphide precipitate, containing the 

 niobium and tantalum, was evaporated to dryness and the resi- 

 due decomposed by fusion with acid potassium sulphate. 

 The fusion was then digested with water and the insoluble 

 niobium and tantalum oxides were filtered, ignited and 

 weighed. Qualitative tests indicated that there were not 

 determinable quantities of other substances present. The 

 results of the analysis of the original stibiotantalite from 

 Australia by G-oyder, together with those of two varieties of 

 the mineral from Mesa Grande, are as follows : — 



Australia. 



Mesa Grande. 



Mesa Grande. 





Specific gravity 7 '37 





6-72 



5-98 









I 



II 



Average I 



II 



Average 



(Ta,Nb) 2 5 58-69 



5522 



55-44 



55-33 50-57 



50-03 



50-30 



Sb Q 3 40-23 



44*3 1 



44-21 



44-26 49-07 



49-49 



49-28 



Bi 2 3 '82 



•31 



•34 



•33 -53 



•54 



•53 



NiO -08 



99*92 



100-11 



H 2 08 













99-90 

 The results of the analyses indicate that with decrease in 

 specific gravity there is an increase in antimony oxide, and a 

 proportional decrease in the combined acid oxides, due to the 

 isomorphous replacement of heavier Ta 2 6 by lighter Nb 2 5 . 

 As there is no satisfactory method for separating tantalum from 

 niobium, the attempt has been made to determine the propor- 

 tions of the two oxides by taking the specific gravity of the 

 mixed oxides as obtained from the analyses, and comparing the 

 results with those of pure Ta 2 5 and Sb 2 5 , which are quite 

 widely separated. This method has been tested sufficiently to 

 indicate that it gives fairly satisfactory results, and from the 

 values given below and a consideration of the ratios, which in 

 both analyses are almost exactly 1 : 1, it may be seen that the 

 percentage of the Ta 2 5 and Nb 2 5 must be reasonably exact : — 



Sp. gr. = 5-98. Eatio. 



■'«" ™ !$£ %%}■"" -° 



1543 1-01 49 ?J 'lj}\ i-1722 1-01 



'o3 -001 



99-92 100-11 



Sp. gr. 



= 6-72. 



Eatio. 



Ta o 0, 



36-35 



•0815 / 



Hb.O, 



18-98 



•0708 j 



Sb„0, 



44-26 



•1536 I 



Bi 2 3 



•33 



•0007 f 



