RICHARDS, COLLINS, AND HEIMROD. — COPPER AND SILVER. 131 



the gain on an area of twice the size. This method of correction would 

 certainly give a correction which is too large, for during the electrolysis 

 the solution is steadily approaching saturation with cuprous ions, which 

 are deposited but not accounted for by the correction. Nevertheless it 

 was decided to carry out a series of observations with plates of two 

 sizes, calculating the results according to this method of correction, and 

 comparing both corrected and uncorrected values with the results of an 

 included silver voltameter. In this way an upper and a lower limit for 

 the electrochemical equivalent of copper might be fixed ; and by using a 

 low temperature, it seemed probable that the two limits might be brought 

 near together. Agitation would of course increase the amount of corro- 

 sion * and diminish the error due to hydrogen ; but we preferred to use 

 low current densities and rely upon the natural convection of the elec- 

 trolyte, which is then sufficient if the error is kept low by cold. 



Experimental Details. 

 The cupric sulphate used for the following experiments, after having 

 been purified in usual ways, was allowed to stand for some time, first with 

 floccnlent cupric hydroxide and then with pure copper, in order to free 

 it from possible trace9 of bismuth, silver, etc. Subsequently it was 

 recrystallized from the purest water. Metallic copper for the anodes 

 was prepared by electrolysis from this material, after strongly acidifying 

 its solution with pure sulphuric and nitric acids. 



The silver used for the anodes was prepared according to the well 

 known chloride-milk-sugar method of Stas, and was amply pure enough 

 for its purpose. The argentic nit rate used as an electrolyte was made 

 from this silver by the action of pure nitric acid. This salt was recrys- 

 tallized. dried, and fused in order to insure its neutrality. For particulars 

 concerning these and other precautions, the reader is referred to recent 

 papers upon the atomic weight of copper and other atomic weights.f 



In any given experiment, the galvanic current was sent through two 

 silver and at least two copper cells, all being arranged in series in such 

 a way as to avoid liability to leakage. Every point of contact with the 

 desk was amply insulated by glass plates, and the wires were arranged 

 as much as possible in air-lines. 



The copper cells were contained in wide-mouthed bottles with a ca- 

 pacity of 500 cubic centimeters each, provided with well cleaned rubber 



* Compare Foerster's results. 



t These Proceedings, 22, 342 (1887), 23, 177 (1887), 25, 95 (1890), 26, 240 

 (1891), etc. 



