22 MM. E. du Bois-Reymond and W. Beetz on the 



of error, the finiteness of the plate or the charge, exerts 

 most influence upon the result of the experiment, I had a cir- 

 cular disc of brass, a foot in diameter, thickly gilt and polished. 

 The rings which I obtained upon it extended so as to form a 

 radius of more than three inches, but then exhibited almost 

 the same irregularities in the proportions of the radii as those 

 on the smaller plates. The influence of the charge is thus far 

 greater than that of the finiteness^ 



As M. Becquerel in his experiments made use of German 

 silver as the receptacle of his coloured rings, it appeared to 

 me desirable to compare my admeasurements on plates of the 

 noble metals with some made under exactly the same circum- 

 stances as those of the French philosopher; for in these 

 plates of German silver we should not expect the same dis- 

 turbing influence of the polarization as in surfaces of gold 

 and platinum, which receive the charge to a far greater ex- 

 tent. The coloured rings which I procured upon plates of 

 German silver, appeared at first of a very brilliant colour. 

 This however, on account of the oxidizability of the metals, did 

 not continue long; the surfaces became spotted, and the rings, 

 on monochromatic illumination, were marked with curves, 

 differing from the circular form by numerous indentations. 

 The German silver plates, like the other ignoble metals, were 

 not well-calculated for the present object, so that I made no 

 further use of them ; however, the experiments made with 

 them convinced me that here also the law of the cubes as the 

 proportion of the radii is correct, and on account of the slight 

 influence exerted by the charge, it is attained in this case 

 with greater accuracy than with plates of the noble metals. 

 An example will serve as a proof of this : 



Table XIII. 

 Coloured rings upon German silver. 



Lastly, a third cause might urge the outermost rings nearer 

 to the centre. Thus, when the point forming the cathode is 

 not approximated so closely to the anode that this distance 



