Second Series. — Polarization by Reflection. 365 



Rays out of 100 polarized. 

 Argand-lamp without reflector, . . .55 



Dark heat, from brass about 700°, . . .61 



Incandescent platinum, . .... 65 



32. This result would be explicable on the supposition that 

 the angle of incidence was that which corresponds to the pola- 

 rizing angle for heat from incandescent platinum, whilst it was 

 too small for the more refrangible heat from an Argand burner, 

 and too great for the less refrangible heat wholly unaccompanied by 

 light ; nor is this view devoid of plausibility, as will immediately 

 be seen, though the troublesome nature of the experiments, and 

 the smallness of the numerical results, prevented me from prose- 

 cuting them farther in this way, than merely to verify generally 

 the preceding results. 



33. These experiments were made, as usual, by disclosing 

 the source of heat only for the time that the needle was making 

 its first swing. Thus the effect of absorbed heat was to a great 

 extent avoided. It was worth inquiring, however, whether the 

 acquired heat of the first plate AB, Fig. 6, could by possibility 

 produce an effect similar to that of polarization in rectangular 

 positions. For this purpose I used the tin water-canister al- 

 ready described (art. 19), which was subsituted for the plate AB. 

 Whilst, however, it remained vertically below CD, as in Fig. 6, 

 such a stream of hot air was carried upwards to the pile as to 

 spoil the experiment. But when the plane of reflection at 

 CD was made horizontal, as shewn in Fig. 8, the effect could 

 be accurately observed, and the result was the same as in the 

 case of transmission, namely, that even on this enormously exag- 

 gerated scale of error, the quantities of heat reflected to the pile 

 in parallel and rectangular positions of the surface were almost 

 precisely the same. (Dec. 21. 1835.) 



34. A mica-plate placed between the two reflecting surfaces 

 in Fig. 6, perpendicularly to the reflected ray, is capable of de- 

 polarizing the heat, as in the case of heat polarized by transmis- 



