Maxwell's Electro-maynetic Theory of Light. 387 



atom may have its own vibration periods, a much more complicated 

 dispersion formula might be anticipated. In this connection the 

 fact that the absorption spectra of the permanganates of potassium, 

 sodium, lithium &c., are identical* appears to be very suggestive. 



For infinitely quick vibrations /t becomes equal to unity. If the 

 radiations discovered by Rontgen are ultimately proved to be due to 

 transverse periodic disturbances of the ether, of very short wave 

 length, this would explain why no refraction of these radiations is 

 produced by material media. 



10. The application of (16) to explain the optical properties of crys- 

 tals and metals is obvious. In a crystal the co-efficients of the terms 

 involving ^ and T 2 will depend on the direction of vibration, which 

 will result in 11 varying with the direction of propagation of light. 

 Further, in accordance with a well-known mechanical principle 7 

 must be taken into account, when T is nearly equal to TI or T 3 . In 

 this case (16) may be reduced to the form 



yti 2 = R 2 (cos 2a + 1 sin 2a) . 



Sin 2a is essentially positive, whilst cos 2a may be either positive or 

 negative. This is generally considered sufficient to explain the 

 optical properties of metals, and of the quasi-metallic aniline dyes 

 when in a solid state. 



Kundt has pointed out that the velocity of red light in a metal is 

 proportional to the electrical conductivity of that metal. A sug- 

 gestive relation in this connection may be derived from (16). 



Let 2a = TT 2x . Experiments with metallic films and prisms 

 alike show that 2a' is small for the majority of metals. We may 

 write 



f.i = /R(cosa' /sin a'). 



The real part of the refractive index is therefore equal to 

 R sin a. 



(16) may also be written in the form 



Then, since siii2a' = tan 2 a' = j-~ , 



R sin a' = jLAli , approximately. 

 2\/( /(T)) 



Mr. Tomlinsonf has shown that in a number of cases the' molecular X 

 viscosities of metals are in the same order of magnitude as their 



* ' Ostwald, 'Zeits. Phys. Chem.,' TO!. 9, p. 579. 

 f II. Tomlinson, 'Phil. Trans.,' 1883, p. 168. 



