294 
MR. .1. LAR^rOR ON A DYNAMICAL THEORY OF 
is straight, so that if F denote the electric force in air and F/K that in the liquid, the 
traction on the surface due to these two causes is (K — 1 )® F 7 SttK^ + (K — 1 ) F'/SttK-, 
that is (K — 1 ) F^SttK upwards, which is the formula employed, by Qctixcke ; while 
F is determined by the relation aF + bF/K =y, provided the bubble is so broad 
that the middle tube of induction is practically cylindrical, a and b being the lengths 
of this tube that are in air and the liquid, and V the difference of potential between 
the plates. Alter an error in the direct determination of K, due to an experimental 
oversight whose existence was suggested by Hopkinson, had been corrected, all the 
results showed substantial agreement, for thirteen liquids that were examined, with 
these theoretical formulse. But the agreement was not quite complete; a subsequent 
examination* still showed that the attraction between the plates always came out 
less and the change of pressure in the liquids greater than the formulm would give, 
though these discrepancies were within the limits of experimental error, except for 
the case of rape oil in wdiich they amounted to as much as ten per cent. The neglect 
of the capillary correction above mentioned would account for a discrepancy in the 
same direction as the first; and the irregularity of the electric distribution near the 
edge of the plates would account for one in the same direction as the second. 
In a paper on the bearing of the phenomena of electric stress on electrodynamic 
theory,! I had previously been led to inferences militating against the possibility of 
dielectric polarization being of molecular type, from a comparison of these experimental 
results with an electric traction formula including both the molar and the molecular 
forcives. According to the present argument {§ 44), the latter forcive being separately 
compensated, the difficulty there encountered does not exist. The remaining con¬ 
siderations in that paper retain their validity; they show for instance that the 
formulm for the experimental reductions can be derived from a .knowledge of the 
distribution of the organized energy alone. But in the light of the present views, 
we are no longer restricted or even allowed to consider the induction in a dielectric 
as all of one kind ; the total circuital induction is in fact made up of a material 
polarity combined with an mthereal elastic displacement, giving an apparent but 
natural complexity which it had previously been an aim to evade. 
79. Experiments on Electric Expcmsion in Solids and Eluids. —The results 
obtained in § 70 may be applied to the discussion of a very thorough series of 
experiments on electric expansion, made by Quincke,| which appear hitherto 
not to have been correct!}^ interpreted. Following the early experiments of 
Fontana, and more recent ones by Govi and Duter, a condenser of the form of a 
glass thermometer-bulb was used, and the expansion of volume arising from electric 
excitation was read directly on its tube, By employing a long cylindrical bulb, the 
* G. Quincke ; ‘ Wied. Ann.,’ 32, 1887, p. 537. 
t ‘ Roy. Soc. Proc.,’ 52, 1892, pp. 65-6. 
t Abstracted in ‘ Sitz. Akad. Berlin,’ February, 1880, and ‘ Pbil. Mag.,’ July, 1880, pp. 30-39 : in 
full in ‘Wied. Ann.,’ 10, 1860, pp. 161-202, and 513-553. 
