ALTERATIONS IN THE DIELECTRIC 143 



of a condenser like a thermometer with a long cylindrical bulb 

 showed that 



Maxwell's electric stresses do not explain the effect. 

 The result just given is important as showing at once that no such 

 system of stresses as that considered in the last chapter will 

 account for the effect. That system would give extension parallel 

 to the lines of force, and compression perpendicular to them. But 

 it fails to account for the effect in that it would give on the whole 

 a contraction in volume. If, as in the last chapter (Fig. 93), we take a 

 cube of the dielectric, with its vertical edges parallel to the lines 

 of force, it is acted on by its surroundings with the set of forces 

 represented in the figure. 



The vertical tensions and one pair of side forces will give a 

 shear with no change of volume, while the other pair of pressures 

 will produce a decrease in unit volume (1 2<r)P/Y, where <r is 

 Poisson's ratio and Y is Young's modulus. We can perhaps estimate 

 this effect as compared with the increase measured by Quincke. 

 In one case of flint glass with a thickness 014 cm. the change in 

 volume was 1 in 10 5 when the potential difference would spark over 

 1 2 mm. Ix'tween knobs i> cm. in diameter. We may perhaps take this 

 as a potential difference about equal to 25 E.S. units, and it gives 



O ^v 



intensity within the glass E = ^- X 1C 3 . For flint glass K is 



K ]';- 

 probably about 7, .so that P = - = 10 6 nearly. 



\\ C may take 1 - 2<r = ^ and Y == 5 X 10", so that the 



10 6 

 decrease in unit volume should be of the order - = 10~ 6 or 1 in 



1,000,000, as against the observed increase of 1 in 100,000. The 

 Maxwell .system of stresses then would only lessen the electric 

 expansion slightly, even if we could suppose them to produce 

 elastic- >train. 



Electric expansion of liquids. Quincke also experimented 

 with liquids. A conden>er was formed in the liquid by a pair of 

 vertical platinum plates. The containing vessel was entirely filled 

 with the liquid and closed except that a capillary tube rose from it 

 to show any change in ve>lume of the liquid. On charging the plates 

 as a concfenser the level in the tube was observed to alter. 

 The results indicated that the change in volume is probably 

 proportional to the sejuare of the intensity. In general there 



xpansion, but with fatty oils, such as colza and almond 

 oils, there is a elccre.-a.se in volume. 



