390 Note 15: on the theory of dielectrics 



Probably for the sake of being able to apply his mathematical theorems, 

 he takes the case in which the conducting parts of the glass are in the form of 

 strata parallel to the surfaces of the glass. He is perfectly aware that this is 

 not a true physical theory, for if such conducting strata existed in a plate of 

 glass, they would make it a good conductor for an electric current parallel to 

 its surfaces. As this is not the case, Cavendish is obliged to stipulate, as in 

 this proposition, that the conducting strata conduct freely perpendicularly to 

 their surfaces, but do not conduct in directions parallel to their surfaces. 



The idea of some peculiar structure in plates of glass was not peculiar to 

 Cavendish. Franklin had shown that the surface of glass plates could be charged 

 with a large quantity of electricity, and therefore supposed that the electric 

 fluid was able to penetrate to a certain depth into the glass, though it was not 

 able to get through to the other side, or to effect a junction with the negative 

 charge on the other side of the plate. 



The most obvious explanation of this was by supposing that there was a 

 stratum of a certain thickness on each side of the plate into which electricity 

 can penetrate, but that in the middle of the plate there was a stratum im- 

 pervious to electricity. Franklin endeavoured to test this hypothesis by grinding 

 away five-sixths of the thickness of the glass from the side of one of his vials, 

 but he found that the remaining sixth was just as impervious to electricity as 

 the rest of the glass*. 



It was probably for reasons of this kind, as well as to ensure that his thin 

 plates were of the same material as his thick ones, that Cavendish prepared 

 his thin plate of crown glass by grinding equal portions off both sides of a 

 thicker plate. [Art. 378.] 



It appears, however, from the experiments, that the proportion of the 

 thickness of the conducting to the non-conducting strata is the same for the 

 thin plates as the thick ones, so that the operation of grinding must have re- 

 moved non-conducting portions as well as conducting ones, and we cannot 

 suppose the plate to consist of one non-conducting stratum with a conducting 

 stratum on each side, but must suppose that the conducting portions of the 

 glass are very small, but so numerous that they form a considerable part of 

 the whole volume of the glass. If we suppose the conducting portions to be of 

 small dimensions in every direction, and to be completely separated from each 

 other by non-conducting matter, we can explain the phenomena without 

 introducing the possibility of conduction through finite portions of glass. 



It was probably because Cavendish had made out the mathematical theory 

 of stratified condensers, but did not see his way to a complete mathematical 

 theory of insulating media, in which small conducting portions are disseminated, 

 that he here expounds the theory of strata which conduct electricity perpen- 

 dicularly to their surfaces but not parallel to them. 



* Franklin's Works, 2nd Edition, vol. I, p. 301, Letter to Dr Lining, March 18, 



