Theory and Measurement of Residual Charges. 439 

 electrons, it will be seen that the change in the value of I is 

 probably very small. We have, therefore, 



Tq^Ti 



zij being the number of free electrons per unit volume at the 

 anode, and 7? the corresponding number at the cathode. 

 According to this view, the residual charge is due to changes 

 of the surface resistances consequent on a displacement of 

 the free electrons under the action of the Held of force, and 



£ _ Vl — 'h m When the field is removed by connecting the 

 ^1 ^o . /'•••-.■: 



two metal plates the residual charge disappears gradually, 

 and, after a time, nj becomes equal to n . 



It is not necessary, then, for a residual charge, that the 

 dielectric should be heterogeneous, unless the rearrangement 

 of the free electrons caused by electric force be regarded as 

 constituting heterogeneity. In this connexion it may be 

 worth while quoting a remark of Maxwell (' Electricity and 

 Magnetism/ vol. i. 3rd edition, p, 457). 6i It by no means 

 follows," he says, "that every substance which exhibits the 

 phenomenon is so composed, for it may indicate a new kind 

 of electric polarization of which a homogeneous substance 

 maybe capable, and this in some cases may perhaps resemble 

 electrochemical polarization much more than dielectric 

 polarization." 



Tig. 1, 



EAHfht 



To €tt#fffiim ma 



Earthing Hey. 



TO ElECTfflMETffi- 



f/lftTH. 



The method of measuring the residual discharge will be 

 understood from fig. 1. The dielectric was a plate D of 

 hard brittle amorphous sulphur of a uniform thickness of 

 •153 cm. placed between two smooth circular plates of zinc 

 A and B, each having a diameter of 10 cm. Both the sur- 

 faces of D were like smooth glass. The plate B was insulated 

 by two blocks of paraffin, and the whole condenser was inside 

 an earthed cylindrical box E. B was connected by a wire 

 passing through a plug of paraffin wax, to a key, by means 

 of which it could be charged to any available potential, or 

 earthed, and A was connected by a wire passing through 

 another plug to the electrometer. Another zinc plate (J, 



2Gf 2 



