Second Series. — Polarization by Reflection. 361 



lity is less completely polarized by a bundle of plates placed at a 

 given angle, than the more refrangible rays. 



25. Mica is pre-eminently adapted for the purpose of polariz- 

 ing by its considerable diathermancy, and by the extreme thin- 

 ness of its laminae. I have, however, succeeded in polarizing 

 heat by transmission through a bundle of rock-salt plates with 

 parallel surfaces. * Two bundles consisting of three plates, or six 

 surfaces each, polarized about one-seventh of the heat which 

 passed in the parallel position, the angle of inclination to the 

 incident heat being about 55° ; but when all the six plates were 

 combined into one bundle, and the mica plate I used along with 

 it, not far from a half of the transmitted heat was polarized. f 



26. A very convenient mode of mounting the mica plates for 

 polarizing is shewn in Fig. 5. A cylindrical wooden tube is cut 

 across at an obliquity of 34° to the axis. The plate of split mica 

 is interposed and the parts reunited. The plane of polarization 

 or analyzation may thus be made to shift through any angle by 

 turning the tube containing the mica round its axis, and a small 

 support A is provided to preserve it in any position ; whilst a 

 graduation may easily be applied to the exterior of the tube, so 

 as to mark the angular revolution. The convenience of this 

 construction will afterwards appear. 



§ 4. On the Laws of Polarization of Heat by Reflection. 



27. The general fact of the polarization of heat by reflection 

 was ascertained by me in December 1834, and 1 stated the re- 



* For a supply of this valuable substance I have been greatly indebted both to 

 Sir Philip Grey Egerton, Baronet, of Oulton Park, Cheshire, and to Dr Traill. 



-f- Such plates being equally permeable to every kind of heat, as M. Melloni's 

 admirable experiment shews, would probably enable us to polarize cold, or to shew 

 the negative effects due to a reduction of temperature. This experiment I have not 

 tried. 



3 n2 



