Thin Metal Partition in a Voltameter. 291 



through the gold-leaf partition, This method, in detail, was 

 to set up two similar voltameters at the same time, leaving 

 both open, and making quantitative analyses at stated in- 

 tervals ; then, closing the current on both in series, to again 

 make quantitative analyses of the solution on the kathode 

 side and weigh the Cu on the kathodes : finally to break the 

 circuit and again repeat the analyses. 



Curves plotted from these results show no break nor change of 

 slope for the intervals during which the current was passing. 

 In the foregoing experiments there was no deposit of Cu on the 

 gold-leaf partition provided the current was less than about 

 0*3 of an ampere. The time-curves for the two specimens of 

 gold-leaf are not identical, showing a different rate of diffusion, 

 but are quite similar in shape. 



Using a Cu anode there was no visible gas on the anode, 

 but the anode was dissolved by the current in the usual 

 manner. Using a platinum kathode and a platinum anode, 

 there was, of course, an escape of gas from the anode pro- 

 portional to the current-strength. 



Gas always escaped from the Pt kathode, proportionally but 

 less as the amount of Cu deposited increased. In the case 

 where the partition was removed and the solutions thoroughly 

 mixed there was no gas, but the full amount of Cu deposited 

 on the kathode. 



Thus it seems that the passage of the current in such a 

 voltameter is not accomplished by the passage of the Cu 

 through the gold-leaf partition. Some of the CuS0 4 passes 

 the partition by difusion, and then does its proportional part 

 of the conduction of the mixed solution of CuS0 4 aod H 2 S0 4 , 

 which accounts for the increased percentage . of Cu deposited 

 on the kathode as the interval from the time the voltameter 

 was filled increases. 



In the time-curves which may be constructed from Table III. 

 (PI. IY. fig. 2) it will be noticed that the amount of Cu at the 

 end of the first interval seems too small, causing a similar bend 

 in both curves. This is doubtless due to the fact that the 

 current was not continued long enough to deposit all the Cu 

 in the 10 cub. cent, sample. The amount of CuS0 4 present 

 being very small, it would be deposited very slowly ; and the 

 circuit should have remained closed several hours instead of 

 less than an hour, as was the case. A similar bend occurs in 

 both curves, because the specimens were analysed in series 

 and the same defect applies to both alike. If the current 

 was increased beyond about 0*3 ampere, with CuS0 4 and 

 H 2 S0 4 as the solutions, separated by a gold-leaf partition, 

 there was development of gas and Cu on the partition. This 



