372 Gooch and Stookey — Reduction of Vanadic Acid 



were charged with water containing 3 grm. of potassium 

 iodide ; the vanadate was introduced into the dry flask and the 

 apparatus was adjusted. Hydrochloric acid was put in the 

 stoppered separating funnel and, when all was ready, allowed 

 to run into the flask. Connection was made between the 

 funnel and a carbon dioxide generator, so that by passing car- 

 bon dioxide through the apparatus all danger of regurgitation 

 might be avoided. Sometimes in the course of our work, too, 

 a branched connection with the funnel tube was so arranged 

 that either hydrochloric acid gas or carbon dioxide or both 

 might, at pleasure, be sent into the apparatus. We have used 

 in the experiments to be detailed ammonium vanadate either 

 purchased in pure condition or made by precipitation with 

 alcohol, or prepared from the oxychloride obtained by acting 

 with chloroform upon ignited vanadium pentoxide. In each 

 case the content in vanadium pentoxide was determined by 

 ignition and confirmed by Holverscheit's"^ bromide process. 



In Table I A are gathered results obtained by the use of 

 the ordinary pure hydrochloric acid of the laboratory of sp. gr. 

 1*17 ; while in B of the same table are recorded results got by 

 using acid of sp. gr. 1*20, specially prepared by saturation of 

 the former acid kept cold by ice, with hydrochloric acid gas. 



In these experiments, in which the precautions described to 

 catch all the chlorine evolved were taken, the actual registra- 

 tion of chlorine amounts in the average to 92*62 per cent of 

 the theory when the ordinary pure acid of the laboratory of 

 sp. gr. 1'I7 was used, and to 9446 per cent when the specially 

 concentrated acid of sp. gr. 1-20 was employed. While the 

 error of these determinations would amount to a considerable 

 figure when large amounts of vanadic acid are to be deter- 

 mined, it is obvious that a correction of, let us say, 6 per cent 

 or 8 per cent upon the small proportions of vanadium pen- 

 toxide present in many of the complex salts to which Gibbs 

 applied this method, would not change the significance of the 

 analysis. Indeed, reference to the original shows that the 

 formulae deduced by Gibbs would in many cases apply as well 

 to the corrected results as to the original figures. Thus it 

 appears that, so far as concerns the special cases in which Gibbs 

 used this method of analysis, the conclusion of Milch that the 

 process affords no ground for the calculation of the amounts 

 of pentoxide present, and Holverscheit's difficulty in suppos- 

 ing that results were even approximately true, prove to be 

 unfounded, while Rosenheim's supposititious oxide disappears. 



In the light of these results it is of interest to note the effect 

 of a repetition of this treatment with strong acid of the residue 



^Inang. Dissert., p. 48. 



