132 Hillebrand and Ransome — Carnotite and Associated 



purified were redissolved, precipitated by ammonia, weighed, 

 redissolved by acid potassium sulphate, the ferric iron was 

 reduced by hydrogen sulphide and titrated with permanganate 

 after boiling out the reducing agent. The titanium was then 

 found colorimetrically by hydrogen peroxide. 



The combined alkaline filtrates were heated with excess of 

 ammonium nitrate to separate alumina from most of the 

 vanadium. The precipitate was redissolved in nitric acid, 

 reprecipitated by ammonia, then ignited and weighed and redis- 

 solved by acid potassium sulphate. The sulphuric solution was 

 reduced by hydrogen sulphide, boiled and filtered from plati- 

 num sulphide, again boiled in a current of carbon dioxide and 

 titrated hot with permanganate. Sulphur dioxide gas was 

 then introduced, boiled out in a current of carbon dioxide, and 

 the titration repeated, this second result being taken as repre- 

 senting the V 2 4 equivalent to the V 2 6 that still contaminated 

 the A1 2 3 . Deducting the V 2 6 thus found from the combined 

 weight of A1 2 3 and Y 2 5 gave the A1 2 3 of the soluble con- 

 stituent of the sandstone. 



The combined filtrates from the alumina were evaporated, 

 ignited to remove ammoniacal salts, the residue converted to 

 sulphates, and united with the earlier solution of vanadium sul- 

 phate. The vanadium in it was then determined precisely in 

 the manner already described, and when added to that found 

 with the alumina and calculated to V 2 3 , gave the total in the 

 sandstone. 



A check was made by dissolving the sandstone in hydro- 

 fluoric and sulphuric acids in a current of carbon dioxide and 

 titrating the V 2 3 with permanganate. The figure thus 

 obtained was a trifle below that found as above detailed, and 

 this is to be attributed to the presence of a very little vanadium 

 as Y 2 5 in carnotite, and probably as a lead vanadate. All the 

 uranium and most of the trifling amount of lead could be 

 extracted by cold dilute nitric acid, thus showing them to be 

 foreign to the green substance. 



It is preferable to regard iron as in the ferric state rather 

 than as ferrous, otherwise an equivalent amount .of vanadium 

 would have to be considered as V 2 6 , for which there is no 

 occasion. It is quite possible that the iron is in part, if not 

 wholly, foreign to the green matter, for it would be surprising 

 to find an exposed sandstone free from ferric oxide. 



If the iron is disregarded as probably extraneous, the ratios 

 given below afford the following empirical formula : 



H 396 R 'l 9 4 R " 3 . R '"eiO S5 763 2 77 6 - 



or H 396 R' IM R" 39 R'" 610 (Si0 4 ) 555 (Si 3 6 ) 69 . 

 Simplified this is H 40 R' 20 R" 4 R'" 61 X 62 . 4 . 



