314 R. Sissingh on Kerr's Magneto-optic Phenomenon. 



Perhaps this difference of nearly 90° points to a difference in 

 the behaviour of ordinary electric currents and the periodic 

 motions of the electric theory of light. As we have already 

 seen, this can be shown by other methods *. 



18. Kerr, who discovered the influence of magnetization on 

 reflexion at the surface of iron, gives only the sign, and not 

 the magnitude, of minimum-rotations. His results agree with 

 my observations. In his case, too, there was a change of 

 sign at an angle slightly greater than the principal angle of 

 incidence. Kerr found a maximum of the minimum-rotation in 

 the neighbourhood of 60°. In my observations the maximum 

 amplitude of the magnetic component of the light was reached 

 between 60° and 70°, jut the maxima of rotation and ampli- 

 tude do not coincide. It is worthy of note that Kerr has also 

 used the motion of the dark band in the field of the analyser 

 as the most accurate means of adjustment. With respect to 

 the phase of the magneto-optic component, Kerr mentions 

 that it approximates more nearly to zero than to 90° in the 

 case of angles of incidence less than the principal angle. 

 According to my observations, this is the case for every angle 

 of incidence for which the phase can be accurately determined. 



The signs and magnitudes of the minimum-rotations of the 

 analyser given by Kundt f agree with mine, with the excep- 

 tion of some slight differences of magnitude. 



Kighi's investigation J appeared in the year 1887, after I 

 had commenced these measurements. I have calculated the 

 amplitudes and phases of the magnetic component of the 

 light from the numerous angles of incidence given by Eighi 

 (Table opposite). The principal angle of incidence and 

 azimuth were assumed the same as that for my mirror No. II. 



Where two values of fi and m are given for the same angle 

 of incidence, the first values are calculated from the minimum- 

 and the second from the zero-rotations. We cannot conclude 

 from these observations that the observed and calculated 

 amplitudes are proportional to each other, much less that the 

 difference between the calculated and observed phases is 

 constant. The rotations must be measured with greater 

 accuracy before we can draw the conclusion. To reach this 

 degree of accuracy it is absolutely necessary to test the 

 method of observation. The influence of errors of observa- 

 tion masks these relations. Since Righi did not calculate ary 



* Although the theory of magnetic reflexion which Fitzgerald has 

 given really applies only to transparent media, yet the amplitudes reckoned 

 from it are also proportional to the observed values, 

 t Kundt, Wied. Ann. xxiii. p. 244 (1884). 

 * X Eighi, Ann. de Chim. et de Phys. (6) x. p. 200 (1887). 



