Differential Inductometer and in the Electrophorus. 135 



divide, therefore, the resistances of the two branches in 

 like ratio. Then no current passes round a galvanometer 

 introduced between these points. In the differential inducto- 

 meter the beginning of the branching is the middle inductric 

 plate, whose electric level is V, and the end of the branching 

 is on the two inducteous plates, where the potential is zero. 

 Between the inductric plate and each of the inducteous plates 

 let us now introduce a metal intermediate plate of equal po- 

 tential, so that there are four divisions whose inductive capa- 

 cities are proportional to one another in pairs. If the induc- 

 tive capacity of one of these divisions is altered by introducing 

 another dielectric than air, then the capacity of the other divi- 

 sion must also be altered, in order to restore the original propor- 

 tion. The capacity of this second division is changed by alter- 

 ing the distance between its limiting plates until the original 

 equality of the electrical potential of the two middle plates is 

 again reached. This is recognized by finding that a quadrant 

 electrometer connecting the two shows no deviation. Thus 

 the measurement of inductive capacity is reduced to measur- 

 ing a length. This method is essentially the same as that 

 recommended by Maxwell and Sir W. Thomson to Mr. Gordon 

 for his measurements of inductive capacity, published a year 

 ago (Phil. Trans.). 



Theory of the Electrophones. 



A further instance of the branching of induction is found in 

 the electrophorus. Being about to deduce its theory and to 

 discuss the number of electrically effective layers, I shall start 

 from Faraday's differential inductometer. Such a theory of 

 the electrophorus as takes Faraday's views into account 

 has not yet been given. Faraday has not developed it; 

 and the question as to how many layers are electrically 

 effective is not treated by Maxwell, who views the electro- 

 phorus not as a symmetrical construction consisting of a plate 

 of ebonite with two movable metallic coverings, but as an un- 

 symmetrical arrangement consisting of a single ebonite plate 

 covered at the back with metal, and having only one movable 

 metal plate. 



In this special manner we shall obtain a theory of the elec- 

 trophorus, and get a fixed point from which to view the dif- 

 ferent theories based on electrical action at a distance and 

 expressed in the language of the influence theory. We shall 

 approach nearest to that one of these theories which has lately 

 been supported by Herr v. Bezold (Pogg. Ann. cxliii.). 



The electrophorus, as I view it, is symmetrical. It is a 

 non-conducting plate with two movable coverings, and can 



