ROSA, VINAL AND McDANIELI THE SILVER VOLTAMETER 41 



deviation of individual values from the mean of 2 or 3 cups is 

 only 0.001 per cent. 



Comparison of the large porous cup voltameter with the smaller 

 size sho\yed consistently a heavier deposit in the former by about 

 5 parts in 100,000 and the siphon form which required a very 

 large volume of electrolyte gave a still greater excess. It was 

 at length discovered that this excess of deposit in the larger sizes 

 was roughly proportional to the volume of the electrolj^te and the 

 authors have called this the ^'volume effect." It seemed prob- 

 able that this was due to traces of impurity w^hich had not been 

 eliininated from the electrolyte and led to an extended investi- 

 gation of the purifying and testing of silver nitrate. As a result 

 of this work, electrolyte was prepared which satisfied all the 

 criteria for purity and brought the large sizes of voltameters into 

 agreement with the small porous cup voltameter, thus justifying 

 its previous use as a standard form. The effect of slight con- 

 tamination of the electrolyte in this small size must have been 

 almost negligible as the results obtained with it have been amply 

 justified by the subsequent work. As the experiments progressed 

 the ''volume effect" was found to be a very useful indication of 

 the purity of the electrolyte, revealing the presence of impurities 

 too small in amount to visibly affect the crystalline structure of 

 the silver deposit. It is not claimed that this phenomenon of 

 heavier deposits in large voltameters is a new discovery, for it 

 was observed in the work of Lord Rayleigh nearly thirty years 

 ago and has been ascribed to various causes, but the authors do 

 believe that its significance has not been appreciated before, nor 

 the correct explanation for it given. 



Using the mean of 44 of the most reliable determinations made 

 in the small porous cup form of voltameter during the period 

 December, 1909, to April, 1910, the authors have obtained as the 

 value for the Weston Normal Cell at 20°C. on the present basis: 



1.018275 volts. 

 This is very close to the value recorded in the first series of 

 measurements and is probably more reliable. 



Several determinations were made in which two voltameters 

 were maintained at 50°C. by external heating coils during the 



