244 PROCEEDINGS OF THE AMERICAN ACADEMY 



It will be convenient to speak in this place of the best method of 

 determining vanadium in alkaline vanadates without special reference 

 to its separation from other elements. I find that this estimation may 

 be effected with great accuracy by first oxidizing any lower oxide 

 which may be present by means of nitric acid to vanadic pentoxide, 

 and then i>recipitating the faintly acid solution by mercurous nitrate 

 and an excess of mercui'ic oxide at a boiling heat. Precipitation by 

 mercurous nitrate alone was long since suggested by Berzelius, but the 

 addition of mercuric oxide to render and to keep the solution neutral 

 makes precisely the difference between a very accurate and an only 

 tolerably good result. The precipitate may be collected upon a paper 

 filter, and after washing with hot water ignited with free access of air. 

 On cooling, a fine orange-brown crystalline mass of vanadic pentoxide 

 remains. The filtrate from the precipitated oxides contains no appre- 

 ciable trace of vanadium. The great facility with which a solution 

 of vanadic pentoxide containing free chlorhydric acid is reduced by fer- 

 rous salts, suggested the possibility of determining vanadium by titri- 

 tion with ferrous sulphate, a solution of potas.sic ferricyanide being 

 employed to determine the point of complete reduction. 



A special study of this j^rocess was made in my laboratory by my 

 assistant, Mr. Charles A. French, to whom the following details are 

 exclusively due. The solution containing vanadic pentoxide is to be 

 concentrated, and chlorhydric acid solution containing about one tenth 

 of its weight of pure acid added in large excess. After nearly saturat- 

 ing with a solution of ferrous or ammonio-ferrous sulphate of known 

 titre, the solution is to be heated nearly to boiling, and the titrition 

 then completed with the ferrous salt, the end reaction being deter- 

 mined by trial drops upon a porcelain plate with a very dilute solu- 

 tion of potassic ferricyanide. Of this solution two drops of the 

 ordinary laboratory reagent in a test-tube full of water give an ap- 

 propriate strength. The process may be varied by adding the ferrous 

 solution in excess, then titrating back by means of potassic dichromate 

 of known titre, and finally adding the ferrous salt until the end reac- 

 tion with ferricyanide is obtained. The ferrous solution may contain 

 with advantage 0.005 gr. of iron in one cubic centimeter. A sample 

 of potassic di vanadate containing no other impurity than a little potas- 

 sic nitrate gave, by the gravimetric process above described, 65.12^ 

 V2O5. In three successive titritions with ammonio-ferrous suljihate 

 Mr. French found G5.16, 65.27, and 64.95%. 



As ammonic meta-vanadate is now largely used in the manufacture 

 of aniline black, and has become an article of commerce, the method 



