Circular Polarization, 557- 



700°, I estimated the polarizing angle to be 57° nearly (16th 

 March 1836). By experiment I found that the polarizing 

 angle for the same mica surface and for homogeneous red 

 light was 59°. 



§ 5. On the Circular Polarization of Heat. 



In my last paper I showed (art. 75) that heat may be circu- 

 larly polarized, like light, by the doubly refracting action of 

 a plate of proper thickness. This circumstance is indicated 

 by an equal quantity of heat reaching the pile in all positions 

 of the analysing plate. 



Last summer it occurred to me that it was probable that 

 rock-salt, refracting heat almost as it does light, would cause 

 it to undergo total reflection at a proper incidence. Sup- 

 posing this to be the case (and I had afterwards reason to be- 

 lieve that such had been shown to be the fact by M. Melloni), 

 I then foresaw the possibility of trying an experiment of the 

 most conclusive character, as to the nature of heat, — its sus- 

 ceptibility of becoming circularly polarized by means of two 

 total internal reflections, as in the admirable experiment of 

 Fresnel in the case of light. 



Various circumstances prevented me from trying this expe- 

 riment until the end of January last [1 836], when I procured a 

 rock-salt rhomb, similar to that of glass used by Fresnel, but 

 having its angles calculated by FresneFs formula, for the 

 refractive index for light of rock-salt. I took the smaller of 

 the two angles which the double solution of the quadratic 

 equation gives, on account of the smaller dimensions required 

 for the rhomb. This angle is nearly 45°. On the 1st of 

 February I performed the experiment with complete success, 

 though with an apparatus less perfect than I afterwards pro- 

 cured. 



When the plane of reflection coincided with, or was per^ 

 pendicular to, the plane of primitive polarization, the heat 

 (whether wholly dark, or derived from incandescent platinum) 

 came out unchanged, that is, on placing the analysing plate 

 in azimuth 0° and 90° relatively to the polarizing plate, the 

 ratio of the effects was the same as if no reflection had taken 

 place. 



When the plane of first polarization was inclined -f-4:>° or 

 — 45° to the plane of reflection, and the analysing plate was 

 placed in the parallel and rectangular positions to the polar- 

 izing plate, the ratio of the effects was totally changed, and 

 was, in some instances, reduced nearly to unity. I'his took 

 place whether the rhomb or the polarizing plate was move- 

 able. 



