328 Profs. H. du Bois and H. Rubens on Polarization of 



the axis R. The deviation between the brass plate and the 

 axis o£ revolution o£ the tube can be read off by means o£ 

 the circular scale IT. V serves to make the plate fast in any 

 required position. The investigated grating is fixed to the 

 brass plate by means of clampscrews, in such a way that 

 the wires of the grating remain exactly parallel to the axis of 

 rotation R. The pencil of rays is bounded by a right-angled 

 diaphragm, fixed to the back of! the plate Q. Thus when the 

 grating is moved the effective cross-section of the ray-pencil 

 remains unchanged. This is necessary if the direct action 

 of the rays is to be compared with that after transmission 

 through the grating. A circular diaphragm of 25 mm. inner 

 diameter is attached to the straight end of the tube. 



The concave mirror B is so adjusted that an image of 

 the source A is formed near the centre of the tube C. 

 The rays afterwards pass into the box K, which contains 

 the necessary reflecting surfaces for the production of the 

 Reststrahlen *, and after reflexion at these surfaces and at 

 the concave mirror G reach the microradiometer H con- 

 tained in an air-tight bell f. The polarization of the rays is 

 very complete. This is to be expected from their small 

 divergence, which in these experiments scarcely exceeded +3°.. 



Fis-. 2. 



MlA 



§ 7. The experimental arrangement for the long waves is 

 illustrated in fig. 2, and resembles in many important points 



* For the flttorite Reststrahlen three and for the rocksalt four reflecting 

 surfaces were applied. By using the rocksalt screen the rays appeared 

 to be very pure in both cases; at any rate the impurity was less than 

 1 per cent. The high purity of the rocksalt Reststrahlen is partly ex- 

 plained by the effect of the polarizing quartz plate, which, on account of 

 its high refractive index for long waves, reflects these rays much better 

 than the short- waved heat rays which can be transmitted through the 

 quartz window. 



f With respect to the construction and arrangement of the radio- 

 micrometer reference may be made to the following : — H. Schmidt, Ann. 

 d. Phys. xxix. p. 1003 (1909) ; H. Rubens and H. Hollnagel, Phil. Mag. 

 [6] xix. p. 764 (1910). ' 



