40 ROSA, VINAL AND McDANIEL: THE SILVER VOLTAMETER 



paper and connecting by straight lines each pair of points in 

 table 2 the composition and refractive index of a complete series 

 of fluids will be represented. Errors in n will not exceed ± 0.003 

 if fairly pure crystalline substances are used. Some of the fluids 

 having n between 1.61 and 1.683 may crystallize after standing, 

 but slight warming will restore their fluidity. Other materials 

 under investigation give permanent fluids over much of this range. 

 Where three constituents are specified in the table two of them 

 in fixed proportions are used to form one variable constituent. 



ELECTROCHEMISTRY.— r/ie silver voltameter. III. E. B. 

 Rosa, G. W. Vinal and A. S. McDaniel. To appear in 

 the Bulletin of the Bureau of Standards. 



The second series of quantitative experiments began in Decem- 

 ber, 1909, after several months spent in the qualitative work de- 

 scribed in Part II. As a result of the preceding work the authors 

 -had discarded the filter paper voltameter as an instrument of 

 precision and turned their attention to the problems related 

 to the porous cup form and the purification and testing of the 

 electrolyte. The apparatus and methods employed were similar 

 to those previously described. 



During this period of the work the small porous cup voltame- 

 ter was found to be the most convenient and reliable form to use 

 as a standard. With the purest salt available the deposits were 

 adherent and white, always non-striated, and crystalline as seen 

 under the microscope. As a test of the reproducibility of this 

 small porous cup form, fifty-four deposits made in pairs (except 

 two sets of three each) in which the electrolyte and other condi- 

 tions were as nearly identical as possible in the two cups of each 

 pair were tabulated. The average deviation of each value from 

 the mean of each group (of 2 or 3 cups) was found to be 1 part 

 in 100,000. That is, when the variations in the electrolyte and 

 the measurement of current and time are eliminated so that it 

 is simply a question of how nearly two similar voltameters agree 

 with one another and the variations are produced by loss of 

 silver in washing the deposit, fluctuation in moisture or impurity 

 in the deposit and the errors in weighing the cups the average 



