320 Prof. A. Kundt on the Electromagnetic Rotation 



quenched by a nicol, then upon exciting the electromagnet 

 the field should become bright. This brightness should not 

 disappear upon any rotation of the analyzing nicol, but should 

 increase, whether it be rotated to the one side or to the other. 



To this my experiments are directly in opposition ; inas- 

 much as, when the electromagnet was excited, the analyzing 

 nicol had always to be rotated through a definite angle in 

 order to obtain good extinction. Whether the extinction 

 itself is absolute with homogeneous light, and consequently 

 whether the reflected light remains strictly plane-polarized or 

 whether a certain small amount of ellipticity results, cannot 

 of course be decided. In any case Fitzgerald's theory is 

 refuted by the experiments at normal incidence, and it is 

 therefore not necessary to discuss the admissibility of the 

 assumptions upon which it is based. The question how far 

 Voigt's theory on the reflection of light at circularly pola- 

 rizing media is applicable to our case may remain undecided 

 for the present. 



VI. 



Rotation of the Plane of Polarization upon Oblique Reflection of 



Light from the pole-surfaces and side-faces of a Magnet. 



For these experiments a large electromagnet with erect 

 legs of iron was employed. These had a height of 47 centim. 

 and a diameter of 11 centim. Each leg was provided with 

 three coils of wire 3 millim. thick : each coil had a resistance 

 of about 0*65 Siemens unit. Two large iron parallelepipeds 

 served as poles. The magnetizing current was furnished by 

 the Gramme machine, and had usually an intensity of 16 am- 

 peres. Since the experiments were made in winter, when the 

 sun was little to be relied upon, a Drurnmond lime-light was 

 employed as source of light. The readings of the analyzing 

 nicol were taken upon a divided circle graduated into quarter- 

 degrees, reading with the vernier to 4 minutes, although 

 the reading could be estimated to 2 minutes with certainty. 



The angle of incidence of the rays of light upon the reflect- 

 ing iron surface was carefully read off upon a divided circle. 

 The plane of polarization of the polarizing nicol was adjusted 

 as exactly as possible in the plane of incidence or at right 

 angles to it, as determined by the good extinction of the light 

 bv the analyzer. When the magnet was excited, it was gene- 

 rally not possible to get rid of the light which now passed by 

 rotation of the analyzer ; since after reflection the light was 

 no longer plane-polarized, but elliptically polarized. It was 

 always adjusted to the minimum of illumination, so that the 

 plane of oscillation of the analyzing nicol was brought at 

 right angles to the long axis of the rotated elliptical rays. 



