270 Prof. W. McF. Orr on 



equilibrium of the fether* lead to the conditions that the 

 vector representing the rotational strain should at the surfaces 

 of the nuclei be normal, and outside them be the gradient of 

 a function of the coordinates. This strain-vector can further 

 be chosen so as to satisfy the conditions that outside the 

 inelastic regions its concentration shall be zero, and that the 

 surface-integral of the normal rotation over each such region 

 shall have any assigned value. A state in which the rotation 

 of the aether is identical with what the equilibrium electric 

 displacement of Maxwell would be in case the inelastic regions 

 were replaced by charged conductors, or if of exceedingly 

 minute dimensions by point charges, is thus a possible 

 (though not by any means the only possible) f state of 

 equilibrium of the aether ; and the explanation of electro- 

 static phenomena is so far satisfactory. 



3. But in the development of the theory of moving electric 

 charges, what is called the motion of an electron is the shifting 

 of the centre of the strain from one point in the aether to 

 another, the strain being altered, by continuous motion of 

 the aether, except for portions lying in the track of the 

 nucleus, from what it would be with the nucleus in one 

 position to that which it would be with the nucleus in the 

 other. What is contemplated is not a motion similar to that of 

 a solid body, or a bubble of air, or a vacuous region, through 

 a material fluid, or similar to that of a strain-nucleus through 

 a material solid ; not a displacement of the inelastic portion 

 of the aether, but a transference of the inelastic property 

 from one portion of the aether to another, or the annihilation 

 of the electron in one position and its creation in another, 

 and necessarily involves a constructive process of some kind. 

 An analogy is, however, drawn between such an electron in 

 the aether and a strah"unucleus in a material medium : in the 

 original development of the theory this was pushed too far ; 

 it was stated that each is freely mobile $ ; in the latest state- 

 ment it is, however, pointed out that a strain-nucleus in 

 ordinary matter does not possess mobility of the kind postu- 

 lated for an electron §. 



* See Phil. Trans. A. 1894, p. 747, where, however, it is really only 

 the varying part of the rotational strain-vector (/, g, h) that is expressed 

 as the curl of another vector, for such a representation of the total strain 

 implies that the surface-integral over any closed surface of the total 

 normal rotation is zero ; or § 9, below. 



t The concentration of the strain- vector might have arbitrarily assigned 

 values throughout the elastic regions, corresponding to a volume-dis- 

 tribution of electricity throughout the free aether. It might indeed be 

 urged against the theory that it does not explain why such a distribution, 

 never occurs. 



% Phil. Trans. 1897, A. p. 212; 'JEther and Matter/ -p. 97. 



§ ' iEther and Matter,' p. 336,. footnote.. 



