130 



PROCEEDINGS OF THE AMERICAN ACADEMY. 



1 .000 



0.999 ' 



o 

 Q 



Area per ampere 



0.998 



Figure 3. — Diagram showing Decrease in Copper with small Cathodes 



(due to Hydrogen). 



The table above is here jjlotted by taking the higliest weight of copper in each 

 determination as unity. Great exactness could not be expected from this method, 

 but it is sufBuiently definite to show a marked deflection of the line when the 100 

 sq. cm. line is approached, and a sudden downward tendency when the current 

 density approaclies 0.02 ampere per square centimeter. 



of the action of the ocduded hydrogen, and the deficiency assumes 

 gigantic proportions. Indeed it has been well known for a long time 

 that this current density is the limit if one makes any pretence to 

 accuracy ; our results merely show that the action of the hydrogen 

 begins to affect the deposit loug before it becomes manifest to crude 

 observation. Thus it seems probable that the left-handed tendency of 

 Gray's curves, instead of being due to the approach to the true value and 

 a lessening of the corrosion, was really due to the beginning of a new 

 source of inaccuracy. From a theoretical standpoint such a phenome- 

 non appears only natural ; with more dilute solutions the deficiency would 

 of course appear at lower current densities. Probably the fact that in 

 analytical work the amount of the deposited metal never corresponds 

 to the quantity of electricity may often be due to the same cause, in 

 addition to the oxidizing action pointed out by Danneel.* 



Returning now to weaker currents in fairly strong solutions (Exi^eri- 

 ments a and b), it is evident both from Gray's results and from our own 

 that the weight of the deposit in this case is approximately a linear 

 function of the area of the cathode ; in other words, the dissolving of 

 copper in cupric sulphate is approximately proportional to the surface 

 of copper exposed, even while the current is running. Since this is the 

 case, the obvious method of obtaining the true weight of copper would 

 be to extrapolate for a surface of zero area, — or to add to the weight 

 deposited upon a medium-sized area the difference between its gain and 



* Zeitschr. Electrochem., 4, 153 (1898). 



