«^ 2 }, (7) 



Inductive Capacity of Electrolytes. 203 



of the condenser, the medium having a specific conducti- 

 vity K. 



^ K 1 4^/cd . n 



XV 



where C 2 is the capacity of the condenser with air as dielec- 

 tric. Neglecting polarization effects, which in these experi- 

 ments were usually negligible, equation (3) becomes 



or F = AKY 2 + a 1 (47r«C 2 V) 2 (6) 



For convenience we may take the case of a guard-ring 



2*77-0 2 



condenser, in which case the value of A is — w-*-, S being the 

 area of the disk. Then equation (6) becomes 



F = | r C 2 2 Y 2 [l 



where b = SiT^>, a constant numerical factor. <Zj is our unde- 

 termined coefficient ; we only know of it that it is indepen- 

 dent of V. We should expect it, however, to be a constant, 

 at least for a given liquid. By assumption, u-Jjk 2 is large 

 compared with K, so that any considerable change in F would 

 be due chiefly to a change in a-Jj/c 2 . 



Experiment shows that when C 2 is altered by varying the 



F 



distance between the plates, -^ is constant. Hence a A is 



independent of C 2 . 



When a very small quantity of acid or salt is added to the 

 liquid, k is increased — in a particular case about ten times. 

 F was found to be but slightly altered"*. Therefore, since k 2 

 w T as increased 100 times, a 2 must have decreased 100 times, 

 u x k 2 remaining constant. We have thus found that a x is 

 independent of V and C 2 , and is inversely proportional to k 2 . 



When alcohol is heated from 3° to 26°, the force falls off 

 10 per cent (see fig. 6). But, according to Foussereau f, 

 k in this case increases about 30 per cent., or k 2 69 per cent. 

 We know that the change in F due to change in K is 

 small, when K is small compared with u^k 2 . Hence ol x is no 



* This assumes that the observed diminution was chiefly due to 

 polarization, which is very probable, 

 t Journ. de Physique, 1885. 



