of the Plane of Polarization of Light. 317 



IV. Rotation of the Plane of Polarization upon Normal 

 Reflection from Iron, Cobalt, and Nickel. 



For these experiments also Ruhmkorff's electromagnet was 

 employed. The perforated pole furthest from the observer 

 was replaced by a solid pole without perforation. By re- 

 flection at a glass plate inclined at an angle of 45° to the axis 

 of the pole, light was sent through the perforated limb of the 

 electromagnet, reflected at the metal plate placed between the 

 poles, and then passed through the glass plate, and so through 

 the analyzing nicol. The reflecting surfaces could be ad- 

 justed exactly at right angles to the direction of the incident 

 rays of light, and as nearly as possible at right angles to the 

 axis of the electromagnet, by observations for the purpose. 



The plane of polarization of the polarizing nicol was made 

 to coincide as exactly as possible with the plane of incidence, 

 in order to have as much light as possible. If upon reflection 

 from the metallic surface any rotation took place, so that the 

 plane of polarization of tKe reflected light no longer coincided 

 with the plane of incidence upon the glass plate inclined at 

 45°, then upon passing through this a further rotation took 

 place corresponding to Fresnel's formula for the refraction of 

 polarized light. The analyzing nicol is therefore used to 

 measure, not simply the electromagnetic rotation at the metallic 

 surface, but the sum of this and the rotation produced by the 

 glass plate. The latter, of course, was placed so far from the 

 magnet that no electromagnetic rotation could be produced 

 in it. 



If a denote the angle which the plane of polarization of a 

 ray of light falling upon a glass plate makes with the plane 

 of incidence, then the angle 7 between the two planes, after 

 passage through the plate, is given by the formula 



tana 

 cos" (i—r) 



where i and r denote the angles of incidence and refraction 

 respectively. 



Since in the observations i was always nearly 45° and the 

 refractive index of the glass about 1*48, calculation shows 

 that the values of the directly observed rotations must be 

 decreased by about 10 per cent, in order to obtain the actual 

 electromagnetic effect. Dr. Kerr as well as Mr. Gordon 

 seem to have overlooked the effect of the glass plate ; at least 

 they do not mention it. 



I have retained the same arrangement of the glass plate as 

 these observers, that it might be possible to have the light 



