CHAP, iv.] ELECTRICITY AND MAGNETISM. 227 



inductive power sufficiently disproves the idea that 

 induction is merely an " action at a distance," for it is 

 evident that the dielectric medium is itself concerned in 

 the propagation of induction, and that some media allow 

 induction to take place across them better than others. 

 The existence of a residual charge (Art. 53) can be 

 explained either on the supposition that the dielectric is 

 composed of heterogenous particles which have unequal 

 conducting powers, as Maxwell has suggested, or on the 

 hypothesis that the molecules are actually subjected to 

 a strain from which, especially if the stress be long-con- 

 tinued, they do not recover all at once. Kohlrausch and 

 others have pointed out the analogy between this pheno- 

 menon and that of the "elastic recovery" of solid bodies 

 after being subjected to a bending or a twisting strain. 

 A fibre of glass, for example, twisted by a certain force, 

 flies back when released to almost its original position, 

 a slight sub -permanent set remains, from which, how- 

 ever, it slowly recovers itself, the rate of its recovery 

 depending upon the amount and duration of the original 

 twisting strain. Hopkinson has shown that it is possible 

 to superpose several residual charges, even charges of 

 opposite signs, which apparently " soak out " as the 

 strained material gradually recovers itself. Perry and 

 Ayrton have also investigated the question, and have 

 shown that the polarisation charges in voltameters exhibit 

 a similar recovery. 1 Air condensers exhibit no residual 

 charges. 



When a condenser is discharged a sound is often heard. 

 This was noticed by Sir W. Thomson in the case of air 

 condensers ; and Varley even constructed a telephone in 

 which the rapid charge and discharge of a condenser 

 gave rise to distinct tones. 



1 It would appear, therefore, probable that Maxwell's suggestion of hetero- 

 geneity of structure, as leading to residual electrification at the bounding 

 surface of the particles whose electric conductivities differ, is the true 

 explanation of the "residual" charge. The phenomenon of elastic recovery' 

 may itself be due to heterogeneity of structure. 



