160 R. N. Maxson — lodometric Determination of Gold. 



standard arsenic is insufficient to neutralize the free acid the 

 gold may be precipitated upon the further addition of a suit- 

 able amount of acid potassium carbonate. The flask contain- 

 ing the gold, the arsenite and the acid carbonate was heated 

 thirty minutes upon a steam bath, then cooled and filled to the 

 100"='"^ mark. Measured portions of the solution, SS*^""' each, 

 were titrated with the standard iodine in presence of acid 

 potassium carbonate and potassium iodide. The details of 

 these experiments are shown in the accompanying table. 



A comparison of the figures obtained in the individual titra- 

 tions of portions of a single reduction show that the difference 

 due to the errors of analysis are reasonably small, rarely exceed- 

 ing a few units in the fifth decimal place — or a few hundredths 

 of a milligram. The errors of the process, however, computed 

 by comparing the average of these single determinations with 

 amount of gold, actually taken, range from —0*6°'^ to -f O'GS'^'s, 

 are wholly irregular, and are approximately ten times as large, 

 reckoned absolutely or- in percentages, as the errors obtained 

 by Gooch and Morley when handling similar amounts of gold. 

 So far as the evidence of these experiments goes, it is plain 

 that the process of Rupp is wholly untrustworthy for the deter- 

 mination of the very small amounts of gold concerned. 



I take this opportunity to thank Professor F. A. Gooch for 

 much kindly aid and advice in the preparation of this paper. 



