310 R. Sissingh on Kerr's 



The numbers marked with an asterisk are values chosen at 

 random, and serve to show the influence of a small alteration 

 in the observed rotations upon the calculated amplitudes and 

 phases of the magnetic component of the light. A glance at 

 the Table will show that the mean error in /a is 0*03 x 10~ 3 , 

 and in in about 1°. For angles of incidence smaller than 5 1°, the 

 relative error in the phases m obtained from the values of <j> m 

 is great; while for i = 24:° or smaller, a small alteration in the 

 differences of </>™ and c/>™, or 0™ and $f a alters the amplitude 

 fju very considerably. 



The results obtained show that, within the limits of experi- 

 mental error, <#p° ro = K° ro , and #£ oro = <fe oro . It follows 

 from these relations that fi t =fi t ; m i =-m l . All previous 

 observers have found the same result. We must, however, 

 have regard to the assumptions with respect to signs made in 

 order to completely specify ^ and m*. 



In the neighbourhood of the angle of incidence i = 60°, the 

 zero-rotations <£? and <$>° ia are equal to zero. The magneti- 

 zation then produces only a rotation of the plane of polarization. 

 For an angle of incidence of about 80° the minimum-rotations 

 are zero, i. e. <j>ip= : 4 > Z := Q- ^ ne reversal of sign of these 

 rotations can be seen in the Table, since at the principal angle 

 of incidence J =76° 30' # 5, the minimum- and zero-rotations 

 are equal to each other. For this angle of incidence, there is a 

 difference of 90° in the retardations of phase of the magneto- 

 optic component and that component of the metallic reflexion 

 which is polarized perpendicularly to the plane of incidence. 



If a pencil of light polarized parallel to the plane of inci- 

 dence falls on a magnetized iron mirror, the reflected light is 

 in general elliptically polarized, but in the neighbourhood of 

 i = Q0° it is plane-polarized. If the incident ray is polarized 

 perpendicularly to the plane of incidence, we have at 2 = 80° 

 the peculiarity that the axes of the ellipse which represent 

 the paths of the aether molecules, are parallel and perpendi- 

 cular respectively to the plane of incidence f. 



* Righi found, for example, that *;=-«; C=~4V By an 

 alteration of the direction which we have called the positive direction of 

 rotation of one of the Nicol prisms, we can make the signs agree with 

 our results. 



t Righi, Ann. de Chim. et dePhijs. (6) x. pp. 212,219 (1887) assumes 

 that the influence of magnetization consists in producing a rotation of the 

 plane of polarization together with ellipticity of the reflected light. For 

 an angle of incidence ^=60°, the ellipticity is zero, and for ?=80° the 

 rotation is zero. Eighi calls the first angle "incidence singuliere," and 

 the second, the principal angle of incidence of magnetized iron. From 

 the observations given above, it is obvious that this can only apply to a 

 particular plane of polarization of the incident light. 



