80 Mr Basset, A Provisional Theory of Kerrs [May 1, 



shews that for many metals a> 45°. Exactly the same objection 

 applies to the hypothesis that metallic reflection may be ac- 

 counted for on the so-called elastic solid theory by the introduc- 

 tion of a viscous term. In this theory the difficulty may to a 

 certain extent be got rid of by an extension of von Helmholtz' 

 theory of anomalous dispersion, as shewn in § 386 of my book ; 

 and I entertain very little doubt that if we were better acquainted 

 with the molecular motions which take place when matter is 

 disturbed by the action of electromagnetic waves, this difficulty 

 would be surmounted, and a satisfactory electromagnetic theory 

 of metallic reflection would be obtained. " When," as I stated 

 in § 486 of my book, " electromagnetic waves travel through a 

 medium which is susceptible to magnetic influence, the molecules 

 of matter will be thrown into vibration ; and the direction in which 

 we ought to look for a theory, which will take cognizance of 

 these hitherto unexplained phenomena, is one in which account 

 is taken of the mutual reaction of ether and matter, and which 

 will enable us to introduce the free periods of the vibrations 

 of the matter into our equations." 



[6. Since this paper was read, Prof J. J. Thomson's Recent 

 Researches in Electricity and Magnetism has been published, 

 which contains a theory of Kerr's experiments, see pp. 482 — 509. 

 The principles upon which this theory is based appear to be much 

 the same as those of the present paper, but the details of the 

 analysis are sufficiently different to render comparison difficult ; 

 moreover he has not entered into numerical calculations to the 

 same extent that I have done. He finds that a result consistent 

 with the experiment discussed in Case IV. cannot be obtained 

 unless the transverse electromotive intensity is proportional to the 

 polarization current instead of to the total current ; whereas in 

 the present paper it has been tacitly assumed that this quantity 

 depends upon the total current. 



The reversal of the direction of rotation, which takes place 

 when the incident light is polarized perpendicularly to the plane 

 of incidence, appears to arise from the circumstance that e/ 

 (where e^\/27r is the change of phase) increases from about 22° 

 at normal incidence to 180° at grazing incidence, and consequently 

 between these incidences cose/ vanishes and changes sign. But 

 when the light is polarized in the plane of incidence e^ (where 

 e^\/27r is the change of phase) diminishes from about 22° at 

 normal incidence to zero at grazing incidence, and therefore cos e^ 

 can never change sign, but always remains positive. Sep. 1893.] 



7. It may not be out of place to call attention to another 

 series of experiments made by Kerr *, in which he shewed that 



* Phil. Mag. Nov. 1875, p. 339 ; Dec. 1875, p. 446; March 1880, p. 157. 



