298 R. Sissingh on Kerr's 



section through this ring by a plane normal to the reflecting 

 surface. 



The coils do not project Fig"- 2. 



over the surface of the mirror. Mirror. 



The magnetizing current was 

 furnished by 15 Bunsen-cells, 



in 5 sets each containing 3 fW M^f^^%>, 7 Coil. 

 cells in parallel, the sets being 

 joined up in series ; the mag- 

 netization was measured by 

 Rowland's method *. 



In the centre of the mirror, 

 for a magnetizing current of King. 



15 amperes its value was 1400 C.G.S. units per unit volume. 

 The mirror was nearly saturated ; the magnetism increased 

 8 per cent, for a rise of current from 7*5 to 22 amperes, and 

 only 2 per cent, between 15 and 22 amperes. These rings 

 had the advantage that small masses of iron could be strongly 

 magnetized, and a spectrometer used for the investigation. 

 Two mirrors were experimented upon. The diameter of the 

 rings was 10 centim., the thickness 0'6 centim., the length of 

 the mirror 3 centim., and its breadth 2 # 8 centim. 



The optical constants were determined by a Babinet's 

 compensator f . 



Tvr- Polarizing Angle Principal Azimuth 



Wlllor - (J).° * (II). 



I. . . . 77° 23-5 ; t W° 34' 



II.. . . 76° 30-5' J 26° 44' 



7. In the observations a parallel plane-polarized bundle of 

 rays is reflected from, the mirror, and the intensity of the 

 reflected light is brought to zero or a minimum by rotating 

 the analyser. In order to make measurements with very 

 small or with normal incidence, the collimator, which gives 

 the parallel bundle of rays, and the polarizer are separated 

 from the spectrometer. These three parts — collimator, 

 polarizer, and spectrometer — are so arranged that they can be 

 moved round the edge of a circular platform. 



* Rowland, Phil. Mag. (4) xlvi. p. 140, 1873. 



t For the adjustment of the compensator and the method of using it, 

 cf. Sissingh, Dissertation, Leyden, 1885, and Arch, Neer. xx. p. 171. 



| There is a considerable difference between the polarizing- angles of 

 the two mirrors ground and polished in the same manner. If the two 

 mirrors were not ground and polished in exactly the same way, I often 

 observed differences in J and H. I hope to investigate this more fully in 

 the future. 





