OF ARTS AND SCIENCES. 243 



water. The reason for adopting the method of titrition is, that the 

 ignition of amnionic meta-vanadate, even when conducted with great 

 care, usually yields a mixture of vanadic pentoxide and dioxide, which 

 it is difficult to oxidize completely. The precipitation of ammonic 

 meta-vaiiadate by ammonic chloride is, as is well known, not abso- 

 lutely complete, but the results of the analyses are on the whole fairly 

 good. 



It is sometimes more advisable to employ a combination of the two 

 methods described. After precipitating the two oxides by means of 

 mercurous nitrate and mercuric oxide, the greater part of the molyb- 

 dic oxide may be expelled by heat, after which the mixture of oxides, 

 which now consists chiefly of vanadic pentoxide, may be dissolved in 

 ammonia and treated with ammonic chloride in the manner above 

 described. The quantity of molybdic oxide is most easily estimated 

 by difference, the sum of the weights of the two oxides being deter- 

 mined by the method given above. It is not possible to reduce the 

 vanadic pentoxide in the vanadio-molybdates to vanadic dioxide by 

 means of sulphydric acid gas, by sulphurous oxide, or by any other 

 reagent which I have tried, without at the same time reducing a por- 

 tion of the molybdic teroxide to the blue oxide M02O5. This difficulty 

 may however be overcome by first adding phosphoric acid to the 

 solution in quantity about equal to the weight of the salt analyzed. 

 After the addition of a little sulphuric acid the reduction of the vana- 

 dic pentoxide may be effected either by boiling with a solution of sul- 

 phurous oxide, or by passing sulphydric acid gas for some time through 

 the hot solution of the salt. 



In separating vanadic from molybdic oxide by means of ammonic 

 chloride it is necessary to keep the solution always slightly alkaline by 

 the occasional addition of a little ammonia, since, as soon as the liquid 

 becomes acid by the dissociation of the ammonium salts present, a 

 vanadio-molybdate is again formed in greater or less quantity. To 

 resolve a vanadio-molybdate into a mixture of vanadate and molyb- 

 date, it is necessary to employ an excess of free ammonia, and also to 

 heat the solution. Actual boiling is most advantageous. 



Another method of estimating vanadic oxide in the presence of 

 molybdic teroxide consists in boiling the compound with strong chlor- 

 hydric acid, passing the chlorine set free into a solution of potassic 

 iodide, and then determining the free iodine volumetrically by sodic 

 hyposulphite and starch. I have not actually employed this method, 

 however, though it gives excellent results in the analysis of the vana- 

 dio-tungstates. 



