104 Mr. J. Larmor on Wiener s Localization of the 



atoms are themselves intrinsic mobile configurations of stress 

 or motion, or both together, in the ultimate medium. 



[It is not without interest to consider how far the conception 

 mentioned above of an isotropic solid medium of very small 

 density, with very massive minute nodules imbedded in it but 

 exerting no direct forcives on each other, will carry us in 

 forming a representation of optical phenomena. The theory 

 is of the Young-Sellmeier type, because each nodule has one 

 or more free periods conditioned by its form and by the 

 surrounding elasticity. 



On eliminating (f b r) x , £\) from the equations expressed 

 above, we obtain the vibrational equations of the aether, 

 supposed thus loaded. Its elastic properties are found to be 

 conserved intact, but the effective density as regards vibra- 

 tions of period t is increased by apM a Pi )• When the 



coefficient a is of aeolotropic type, by reason either of the 

 form or the distribution of the nodules, we have effectively 

 an isotropically elastic medium with aeolotropic inertia ; this 

 leads to Fresnel's wave-surface, provided the elasticity is 

 labile in Lord Kelvin's sense. The theory also leads to a 

 formula for ordinary dispersion, of the usually admitted type 

 (Ketteler's) for isotropic media ; but, on the other hand, it is 

 in default by assigning a dispersional origin to double refrac- 

 tion. Jf we wish to include the minute effect known as the 

 dispersion of the optic axes in crystals, it will be necessary to 

 assume for the elastic stress between aether and matter a 

 somewhat more general form, involving (after von Helmholtz) 

 absolute as well as relative displacement, but always of course 

 remaining- linear. 



The assumption of elasticity of labile type also allows an 

 escape from the usual difficulties of a solid aether in the 

 matter of reflexion. In that problem the elasticity would 

 naturally be taken continuous across the interface, the volume 

 occupied by the molecules being on this hypothesis extremely 

 small compared with that occupied by the aether. 



We may further amend the theory by getting rid of the 

 difficulties associated with lability, at the same time avoiding 

 the difficulty as to how a body can move through a perfect 

 solid medium, if we take the aether to be a rotationally elastic 

 fluid, and retain the material load as before. 



But an essential and fundamental difficulty will still remain. 

 It is the extremely small volume-density of the energy involved 

 in radiation which permits a very small inertia, and conse- 

 quently a small elasticity, to be assigned to the aether, and so 

 prevents it from acting as an appreciable drag or exerting an 



