JEther and Electrons. 707 



relation between electric displacement and electric force 

 ceases to hold good ; that, in fact, when a certain displace- 

 ment is exceeded, increase of displacement may imply de- 

 crease of electric force, the electric force being even reversed 

 for very high values of the displacement*. Now although 

 it is usual, in the absence of definite knowledge, to assume 

 linearity of the electromagnetic equations, even for the in- 

 tense field in the immediate neighbourhood of an electron, 

 this assumption is recognized as involving an extrapolation 

 of no ordinary kind. Since it leads to something like de- 

 finite dimensions for a negative electron whose charge and 

 mass have been determined, we know at once what order of 

 electric force is implied as existing in immediate proximity 

 to the electron. With 10 -13 cm. for the radius of the electron, 

 and 5 x 10 -10 for its charge in electrostatic measure, the 

 field-intensity just adjacent to the electron would be 

 £/a 2 = 5xl0 16 electrostatic units, or about 1/5 x 10 19 volts 

 per centimetre. Apparently, then, the assumption that 

 electric force is proportional to electric displacement through- 

 out the region surrounding an electron, implies that the 

 proportionality holds good for field-intensities of this order 

 of magnitude, which is many millions of millions of times 

 greater than any intensity for which such proportionality 

 has been experimentally established. It seems hardly pro- 

 bable that electric force would remain proportional to electric 

 displacement if the value of either quantity could be increased 

 without limit, and in supposing that a sensible deviation 

 from linear relations occurs long before such an intensity as 

 1*5 x 10 19 volts per centimetre is reached, no serious a priori 

 difficulty appears to be encountered. It is even suggested 

 that a free strain-figure theory, by postulating excessive 

 strains and stresses as an essential of electronic constitution, 

 gains rather than loses in plausibility. None the less it may 

 be assumed that, for the relatively feeble values of electric 

 force involved in radiation phenomena and in realizable 

 laboratory experiments, linearity of relation holds good to an 

 extremely high order of accuracy. 



21. Moreover, it is to be observed, as a corollary to 

 §§11, 12, that, to a first order, the elastic properties of a 

 medium are not modified by the presence of assemblages of 

 locked strain-figures ; no property corresponding to dielectric 



* When the strain has reached such a point that the electromagnetic 

 relations are no longer linear, some further definition may be needed as 

 to what precisely is meant by electric displacement and electric force. 

 But even in the absence of such definition, the uncertainty in the meaning 

 of the terms as here used is hardlv a serious inconvenience. 



3 B 2 



