i?. JV. Maxson — Small Amounts of Gold. 469 



It may be seen from these results that the decomposition of 

 the aurous iodide under the conditions is comparatively slow, 

 although potassium iodide was present in large excess. In 

 these experiments the coloration ran through various grada- 

 tions from pink to blue, as the starch used, in common with 

 most specimens, had suffered hydrolytic change, the pink color 

 indicating, of course,* the incipient excess of iodine, and not 

 under-titration as Rupp has stated. 



As I have previously shown, a complete determination by 

 the method of Gooch and Morley can be easily accomplished 

 in ten minutes ; the maximum, time actually necessary for the 

 completion of the titration after the formation of the aurous 

 iodide need not exceed four minutes. As can be seen from the 

 experiments tabulated above, the minimum time required for 

 the incipient decomposition of the aurous iodide was seventeen 

 minutes. In no case did decomposition take place immediately, 

 although potassium iodide was present in large excess. It is 

 to be especially noted that these experiments begin where the 

 analytical process ends, so that the probability of any iodine 

 being freed by a too rapid decomposition is rendered yet more 

 remote. The facts, then, afford absolutely no foundation for 

 the statement that the decomposition of aurous iodide is an 

 inherent source of error in the method of Gooch and Morley. 



Concerning the accuracy of the method proposed by Hupp 

 it is unnecessary to speak further, since Rupp now disputes 

 the desirability of estimating less than the milligram ; but to 

 the reduction of gold chloride by arsenious oxide, the reaction 

 used by Rupp, it is necessary to recur. Rupp states that the 

 process of reduction goes on in acid solution. As I said in a 

 former paper, I have been unable to effect the precipitation of 

 the gold when free acid is present, but after the addition of 

 acid potassium carbonate, reduction takes place. Rupp added 

 to his solution of gold chloride containing hydrochloric acid, 

 10 cm3 of nearly ~N/2 solution of arsenious oxide. It is to be 

 supposed that this solution of arsenious oxide was made, as 

 usual, by dissolving the oxide in acid potassium carbonate. 

 The question arises as to whether Rupp did not have enough 

 acid carbonate in his arsenic solution to neutralize the free 

 acid. The fact that the gold was precipitated, and within ten 

 or fifteen minutes, points strongly, according to my experience, 

 to an alkaline condition of the solution. So, if to heat a solu- 

 tion of acid potassium carbonate upon the water bath, under 

 the experimental conditions, brings about the formation of an 

 iodine-binding carbonate, and is indeed an error as* Rupp says, 

 it would seem that Rupp makes that error. But was it an 

 error? It has seemed worth while to investigate this point. 



*See Hale, this Journal (4), xiii, 390, 1902. 



