250 PROCEEDINGS OF THE AMERICAN ACADEMY 



potassic or soclic nitrate may be added to the mixture in the crucible. In 

 this manner we may avoid the error which would result from the pres- 

 ence of vanadic dioxide in the small quantity which often exists in con- 

 sequence of the reducing agency of dust or traces of organic matter. 



After a great number of trials, the methods given above are the only 

 ones by which I have found it possible to effect even an approximate 

 separation of vanadic and tuugstic oxides in combination. 



In place of the actual separation of the oxides, it is much more con- 

 venient, and I believe also in general much more accurate, to employ 

 the following process, which permits us to estimate vanadium quanti- 

 tatively in presence of tungstic, as well as of various other oxides. 

 This process depends upon the fact that vanadic pentoxide, either free 

 or in combination, when boiled with strong chlorhydric acid, evolves 

 chlorine, which may be collected in a solution of potassic iodide. The 

 iodine set free may then be estimated with great accuracy by means of 

 sodic hyposulphite after adding a solution of starch. 



The fact that vanadic pentoxide evolves chlorine on heating with 

 chlorhydric acid was, so far as I have been able to find, first remarked 

 by Mohr, who in his well-known work on volumetric analysis sug- 

 gested as possible the method of determining vanadium which I have 

 employed, without, however, testing it in any single case. 



It is possible, in some cases at least, to determine vanadic pentoxide 

 in the presence of tungstic oxide by titrition with hypermanganate. 

 To the solution of the substance containing both metals, phosphoric and 

 sulphuric acids are to be added, and the vanadic pentoxide is then to 

 be reduced to dioxide by means of sulphydric acid gas passed into 

 the boiling solution until comjjlete reduction is obtained. No appre- 

 ciable quantity of tungstic oxide is reduced under these circumstances. 

 The titrition with hypermanganate may then usually, though not 

 always, be executed, and gives a fairly good result. 



Finally, the method of titrition by ferrous salts may often be used 

 with success. The details have already been given, and do not need 

 to be repeated. 



5 : 1 Vanadlo-tungstate of Ammonium. — When 12:5 ammonic 

 tunijstate is boiled with a solution of ammonic meta-vanadate an orange 

 solution is obtained, which, on evaporation deposits orange-yellow crusts 

 of indistinct crystals. The color of this salt is somewhat dull ; it is 

 not very soluble in cold, but soluble in boiling water, and may be re- 

 crystallized without sensible decomposition. The salt becomes green- 

 ish upon the surface when dried upon woollen paper, from a slight 

 superficial reduction of the vanadic pentoxide. Of this salt, — 



