810 
MR. J. LARMOR ON A DYNAMICAL THEORY OF 
On the other hand the sharpness of the spectral lines shows that the waves in the 
a3ther are absolutely simple harmonic, and this would point to atomic rather than 
molecular vibrations, were it not that the molecule is so small compared with a wave¬ 
length and also the periods far too great for such an origin.* 
116. A difficulty has been felt as to how the centre of rotational strain which 
represents an electron is possible without a discrete structure of the medium ; the 
following explanations may therefore be 2 :»ertinent. In the first place, it is essential 
to any simple elastic theory of the aether that the charge of an ion shall be repre¬ 
sented by some permanent state of strain of the aether, which is associated with the 
ion and carried along by it. Such a strain-configuration (in the light of what follows) 
can hardl}^ be otherwise than symmetrical all round the ion ; even if the nucleus be 
not itself symmetrical, this symmetry will be attained at a sufficient distance away 
from it. Now in an isotropic medium a steady configuration of strain of this kind 
must consist of a radial displacement such as we could imagine to be produced by an 
intrinsic pressure in the nucleus, or of a radial twist as above described, or it may 
combine the two. But for a great variety of reasons, electric and optical phenomena 
have no relation to any compression of the cether; therefore the notion of an intrinsic 
radial twist is the only representation that is available. An ideal process for the 
creation of such a twist-centre has already been described in § 51 for the case of the 
rotational rether. A filament of the aether ending at the nucleus is supposed to be 
removed, and the proper amount of circulatory motion is to be imparted to the walls 
of the channel so formed, at each point of its length, so as to produce throughout the 
medium the radial rotational strain that is to be associated with the electron; uffien 
this has been accomplished the channel is to be filled up again with aether which is 
to be made continuous ^Yith its walls. On now removino; the constraint from the 
walls of the channel, the circulation imposed on them will tend to undo itself, until 
the reaction against rotation of the aether with which the channel has been tilled up 
balances that tendency, and an equilibrium state thus supervenes -with intrinsic 
rotational strain symmetrically surrounding the nucleus. If on the other hand the 
aether had the properties of an elastic solid, and resisted shear but not rotation, the 
equations oihodily elasticity would remain just the same (§ 19); but the surfaces of 
sheair of such a nucleus would be conical, with the channel by -which the shear is 
introduced as their common axis, and when the constraint is removed the rotaitioii 
imjjosed on che surface of this channel will undo itself and the shear thus aill come 
out again, because the medium with which the channel is now filled up opposes no 
resistance to being rotated. Thus an elastic solid aether does not admit of any con¬ 
figuration of intrinsic strain such as would be required to represent an electric charge; 
and this forms an additional ground for limitation of that medium to a rotatioiially 
elastic structure. For an isotropic medium must be either elastic like a solid or fluid, 
* See Gr. Johnstone Stonei', “ On the Cause of Double Lines and of Equidistant Satellites in the 
Spectra of Cases,” ‘ Trans. Roy. Dublin Soc.,’ 1891. 
