426 Prof. J. H. Poynting on the Discharge 



forming one positive tube, as 5 (fig. 3), which will move off 



Fisr. 3. 



WIRE 



to the wire, the inner parts uniting to form a negative tube 6 

 in B. 



When the difference of potential between the end plates is 

 zero, suppose the wire to be removed. The induction still 

 remaining decays. If it decayed in the same proportion 

 throughout, the difference of potential would always remain 

 zero. But it decays in greater proportion in the negative 

 layers, since in these the dissipation is by hypothesis most 

 rapid. Hence in the line-integral of the induction from plate 

 to plate the negative terms decrease more rapidly than the 

 positive, and so the total value becomes positive. Then on 

 again connecting with a wire, another positive discharge 

 occurs. The process may evidently be repeated, the discharge 

 always being positive, until finally it becomes insensible. 



The analogy between the residual discharge and the phe- 

 nomenon of elastic recovery in strained solids, pointed out bv 

 Kohlrausch, suggests a simple illustration. 



Suppose that we build up a cube with successive layers of 

 substances with the same instantaneous rigidity but with dif- 

 ferent viscosities. Let this be placed between two plates, 

 A, B (fig. 4), the lower plate being fixed. Let rigid trans- 







k 



s\ 4. 



















1 

































M^M BH 



verse partitions, r, r, be passed through the layers and attached 



