74 Mr Basset, A Provisional Theory of Kerr's [May 1, 



By means of these values it can be seen by inspection that each 

 term of (11) is positive for all incidences. The same result follows 

 from Eisenlohr's corrected table. 



Having regard to the directions in which the quantities are 

 measured in my paper in the Phil. Trans., it follows that when 

 q is positive the amperean current which magnetizes the electro- 

 magnet circulates from the right hand towards the left hand of 

 an observer who is looking at the point of incidence through 

 the analyser * ; whence d is positive, and consequently in order 

 to extinguish the light produced by magnetic action the analyser 

 must be rotated towards the right hand of the observer, that is 

 in the opposite direction to that of the current. On the other 

 hand a rotation in the same direction as that of the current, or 

 a reversal of the direction of the current strengthens the light 

 restored from extinction. All these results agree with Kerr's 

 experiments. 



It ought to be noticed that the reflected light is really ellipti- 

 cally polarized, but as the magnetic term q is exceedingly small 

 the portion of the intensity which depends upon q^ is compara- 

 tively unimportant, and the reflected light is approximately plane 

 polarized. 



Case II. We must now consider what happens when the 

 incident light is polarized perpendicularly to the plane of in- 

 cidence. 



In this case A = 0, and we may take B =1; whence the re- 

 Fig. 3. 

 V 



fleeted vibrations are 



^ = -q{P-iQ)e"^ = -q(Pcos(f>+ Qsmcj)), 



^=^e'(*+^''^=i3 cos ((^ -{-<)• 



* [The value of q is C|^/4t, where C is Hall's constant, pj the magnetizing force 

 which acts along the axis of z and t the period. The value of C for iron is positive 

 (see J. J. Thomson, Electricity and Magnetism, p. 488). On p. 398 of mj' book, 

 lime 19, read " Hall's constant" for " Hall's effect." Sep. 1893.] 



