and Polarization of Heat. 141 



the refrangibility of the rays diminished with their temperature. 

 I also obtained a slight refraction of non-luminous heat through 

 a glass prism. 



17. But if heat be capable of refraction by the ordinary 

 agents, an important question arises, Is the phenomenon of dou- 

 ble refraction common to heat and light ? Rock-salt, the only 

 substance yet discovered, which transmits dark heat in large 

 quantity, does not possess this power. To attempt it with Ice- 

 land spar would certainly be fruitless, from the very small trans- 

 mitting power which it possesses, besides some other practical 

 difficulties which suggest themselves. It must be by more re- 

 fined processes that we can detect this property. Such will be 

 stated in the sequel. 



§ 2. On the Polarization of Heat by Tourmaline. 



1 8. It is well known that two slices of tourmaline cut parallel 

 to the axis of the crystal, as they are looked through with their 

 axes parallel or perpendicular to one another, transmit a great 

 portion of the incident light in the one case, and almost wholly 

 intercept it in the other. 



19. It occurred to me as a curious question, at an early pe- 

 riod of my researches, whether non-luminous heat would undergo 

 any similar change in similar circumstances. I made a prelimi- 

 nary experiment with heat from an oil-lamp (not an argand), and 

 though, when the axes were crossed, the whole light was stopped, 

 the heat transmitted appeared to be as intense as before. The 

 tourmalines which I employed were mounted on glass, and were 

 kindly lent to me by the Reverend Mr Craig. Struck with the 

 singularity of the result, I repeated the experiment with addi- 

 tional precautions, and I found that some circumstance prevented 

 this statement from being true in all its generality. The quan- 

 tity of heat transmitted being very small, the lamp, the tourma- 

 lines, and the pile were very near to one another ; and, as the 

 tourmaline absorbs heat with great rapidity, I found that a mi- 

 nute difference might exist if the experiment was made first with 



