If, on the contrary, the plane of polarisation is parallel to the direction 

 of vibration, the velocity of propagation is a function of the position 

 of the plane which includes the direction of vibration, and the direc- 

 tion of transmission. 



If the velocity of polarised light in a crystalline substance depends 

 on the elasticity of the luminiferous medium alone, the latter view 

 must be adopted, and Fresnel's supposition rejected ; for a wave of 

 light is a wave of distortion ; and the rigidity, or elasticity which 

 resists distortion, is, in all conceivable media, a function of the posi- 

 tion of the plane of distortion, being the same for all directions of 

 distortion in a given plane. 



But the experiments of Mr Stokes on diffracted light (Cambr. 

 Trans., Vol. ix.. Part 1) prove that Fi-esnel's conjecture is correct, 

 the plane of polarisation being normal to the direction of vibration : 

 therefore the propagation of light in crystalline media does not de- 

 pend on elasticity alone. 



The author of this paper supposes, according to the hypothesis of 

 molecular vortices (Trans. Roy. Soc. Edin., Vol. xx., Part 1), that 

 the medium which transmits light and radiant heat consists of the 

 nuclei of the atoms of matter, of very small mass, but exerting in- 

 tense mutual forces, vibrating almost independently of the atmo- 

 spheres which surround them. Each nucleus, however, carries along 

 with it in its oscillations a small portion of atmosphere, which acts 

 as a load, retarding the velocity of propagation. In the celestial 

 space, this load is insensible, and it is, generally speaking, greater, 

 the more dense the substance. In crystalline media, the atmosphere 

 of each nucleus is distributed round it symmetrically with respect to 

 three axes, but not equally in all directions ; so that the load upon 

 the nucleus, and consequently the velocity of propagation, is a func- 

 tion of the direction of vibration, as conjectured by Fresnel. 



The author further shews, that according to this hypothesis, if the 

 range of variation of the velocity of propagation of luminiferous trans- 

 verse vibrations is small (as it is in all known media), that velocity 

 must vary sensibly as the reciprocal of the diameter of an ellipsoid, 

 drawn parallel to the direction of vibration. It is well known that 

 this law is the foundation of the whole of Fresnel's theory of double 

 refraction. 



