of Electricity in an Imperfect Insulator. 427 



by hinges to the two plates, and then let the upper plate be 

 acted on by a force in a direction perpendicular to the par- 

 titions, so that a shearing-strain is given to the whole cube. 

 The partitions r, r are merely put so that the distortion from 

 the original position shall always be the same throughout. 

 When a given strain has been produced, let the upper plate be 

 also fixed. Now if the rate of dissipation of strain-energy 

 were the same throughout the layers the stress would also be 

 the same throughout, though gradually decreasing, and on 

 removing the constraint the upper plate would return by a 

 certain amount and then remain in its new position. But the 

 dissipation is not uniform, and after a time the stress in some 

 of the layers is greater than in others. Hence, on removing 

 the constraint from A and allowing it to return, when those 

 in which dissipation has been most rapid have become entirely 

 free from strain-energy, there is still some remaining in the 

 other layers. These latter will therefore strain the former, 

 and we shall have a reverse stress in some of the layers. Thus 

 A will come to a new position of equilibrium ; not so far, 

 however, as its first position. Suppose that it is now again 

 fixed. At first no force is necessary to keep it in position ; 

 but the stress exerted by the negative layers decays more 

 rapidly than that exerted by the positive, and soon, on releasing 

 A, it will return still further towards its original position. 

 The process may be repeated, the successive discharges of 

 momentum imparted to A being always in the same direction. 



[Added April 15, 1886. — The supposition of stratification 

 made by Maxwell is, no doubt, very artificial, and was made 

 for the sake of simplicity in the mathematical treatment. He 

 states that " an investigation of the cases in which materials 

 are arranged otherwise than in strata would lead to similar 

 results, though the calculations would be more complicated ; 

 so that we may conclude that the phenomena of electric ab- 

 sorption may be expected in the case of substances composed 

 of parts of different kinds, even though these individual parts 

 should be microscopically small. 



" It by no means follows that every substance which 



exhibits this phenomenon is so composed " (Electricity 



and Magnetism, 2nd ed. vol. i. p. 419.) 



Probably in the case of blown glass or any dielectric made 

 up of heterogeneous parts, which has been flattened by rolling, 

 there is more or less approach to the stratified condition; but 

 in other cases, such as shellac or paraffin, we might fairly 

 expect the dielectric to be similarly constituted in all direc- 

 tions. We can only therefore take Maxwell's investigation 



