270 PROCEEDINGS OF THE AMERICAN ACADEMY 



oxidizing with bromine, and th^n determining the whole of the vanadium 

 as pentoxide by boiling with chlorhydric acid and titrating as before. 



rnOSPHO-VANADIO-VANADICO-TUNGSTATES. 



I have given this rather ponderous appellation to a class of com- 

 pounds in which vanadium exists partly as pentoxide and partly as 

 dioxide. They may be regarded simply as double salts, but in the 

 present state of our knowledge it will be found most convenient to 

 consider them as a special class. 



Salts of this type are formed whenever phospho-tungstates are mixed 

 with vanadates in presence of vanadic dioxide, when vanadio-vanadico- 

 tungstates are heated with solutions of alkaline phosphates or with 

 phosphoric acid, and when phospho-vanadates and alkaline tungstates are 

 brought together in presence of a reducing agent and of an acid. The 

 salts are sometimes green and sometimes orange-red. In the last case 

 they have frequently the peculiar aurora-red tint noticed among the 

 colors of the phospho-vanadio-tungstates. They pass readily by ox- 

 idation with bromine or nitric acid into salts of the last-mentioned 

 type, and they may be also derived from these by a partial reduction 

 of the vanadic pentoxide. 



Analytical Methods. — These are essentially the same as those which 

 have been described. All the non-basic oxides may be determined 

 together by means of mercurous nitrate and mercuric oxide, after com- 

 plete oxidation with nitric acid. Phosphoric oxide may usually be 

 determined directly, or after separation of a non-alkaline base, by double 

 precipitation with magnesia-mixture. The vanadic pentoxide can be 

 estimated by boiling with chlorhydric acid and titrition in the manner 

 already pointed out. To determine the vanadic dioxide the solution 

 must first be oxidized completely, so as to convert the dioxide into 

 pentoxide, any excess of the oxidizing agent employed being carefully 

 removed. I prefer to use bromine water for this purpose. The whole 

 of the vanadium may then be determined by titrition as above. The 

 difference between the amount of vanadic pentoxide found before and 

 after complete oxidation then gives, by a simple proportion, the corre- 

 sponding amount of vanadic dioxide. It must be remembered that the 

 sum of the non-basic oxides found by the mercury process requires to 

 be corrected by adding the amount of oxygen required to convert the 

 vanadic dioxide present into the equivalent of pentoxide. When the 

 whole quantity of vanadium present is small, this correction may be 

 neglected, as falling within the limits of the errors of analysis unavoid- 



