42 EOSA, VINAL AND McDANIEL: THE SILVER VOLTAMETER 



passage of the current, for comparison with two similar voltame- 

 ters at room temperature (about 20°C.). The results indicate 

 that there is no temperature coefficient as Kohlrausch and Weber 

 also have shown in a different way. It is believed that the 

 reasons other observers have found increased deposits in voltame- 

 ters at higher temperatures is because the chemical activity of 

 impurities in their electrolyte would naturally increase with the 

 rise in temperature. 



Two gold dishes similar in design to the smaller platinum 

 dishes previously used were purchased for comparison. By tabu- 

 lating the differences in deposit between the gold and platinum 

 cathodes in each case where no other differences in the voltameter 

 existed it was found as a result of twenty-nine comparisons that 

 the deposits in the gold dishes were lighter by 1.4 parts in 100,000 

 which may be regarded as indicating the substantial agreement 

 of the two materials. 



Some further work was done with the Poggendorff form but 

 the results were not encouraging and hence attention was directed 

 chiefly to the forms that had proved more reliable and the purifi- 

 cation and testing of the electrolyte in anticipation of the cooper- 

 ative work with the English, French and German National Lab- 

 oratories which began in Washington April 1, 1910. 



The most difficult problems connected with the preparation of 

 silver nitrate for use in the voltameter are concerned with the 

 exclusion or removal of reducing impurities and colloidal silver 

 on the one hand and of uncombined acid and base on the other. 

 The tests that have been developed for determining the acidity 

 of the solution are for the purpose of determining the uncombined 

 acid or base rather than the absolute hydrogen ion concentration. 



The usual methods of testing the neutrality of salts do not 

 apply to silver nitrate. Thus, toward litmus a silver nitrate solu- 

 tion will react alkaline even after the addition of 1 part in 100,000 

 of nitric acid and a similar solution will react acid toward methyl 

 orange after the addition of alkali provided the silver is not 

 completely precipitated. Other indicators also were tried. If 

 the silver be precipitated from the silver nitrate solution by neu- 

 tral KCl solution and the AgCl precipitate filtered off on asbestos 



