RADIANT HEAT. 351 



proportions of heat reflected, the quantity, even at the maximum 

 which would reach the thermoscope after two reflections, would be so 

 extremely small that no difference of effect in the two rectangular 

 positions could really have been perceptible in the form of the experi- 

 ment adopted by Berard. 



(4.) In the fourth section the author enters on the modifications 

 which polarized heat undergoes by the intervention of crystallized 

 plates between the polarizing and analysing parts of the apparatus — 

 an inquiry suggested by the obvious analogy in the case of light. In 

 the crossed position, when polarized heat is stopped, (if the analogy 

 hold good,) the intervention of a plate of double refracting crystal 

 would restore the effect. This apparently paradoxical result was fully 

 verified with plates of mica, and subsequently with selenite and other 

 substances, not only in the case of luminous sources, but even with 

 water below the boiling temperature. Of 157 experiments, with 

 three different mica plates, only one gave a neutral and one a nega- 

 tive result. Of these 157, 92 were made with heat below luminosity. 



The apparent paradox was increased by the circumstance that a 

 thin plate of mica which "depolarized" but feebly seemed to stop 

 more heat than a thick plate which depolarized more completely. 



The main fact was ascertained for the first time on December 16, 

 1834. The professor justly censures the use of the term ' ' depolarize," 

 and suggests " cfopolarize" as preferable. 



(5.) From the result thus unequivocally established, a train of highly 

 curious consequences follow. We have hence, as direct corollaries, 

 the double refraction of the rays of heat by the mica, and their inter- 

 ference according to the same laws as those of light. Hence also follow 

 the constancy of the sum of the intensities of the rays in the rectan- 

 gular positions, or their comjAementary character, agreeably to the 

 formulas of Fresnel for light. This again involves their retardation, 

 according to the well-known principles of the undulatory theory; and 

 hence, from Fresnel' s formulas, we are assured theoretically of the 

 existence of circular and elliptic polarization in the rays of heat under 

 the appropriate conditions. We have thus also the means of deducing 

 the length of a wave of heat. 



The whole of this most important series of investigations was 

 completed between November, 1834, and January, 1835, and their 

 originality and priority are thus placed beyond dispute. The main 

 practical improvement (which led to all the rest of the discoveries) 

 was the employment of the piles of mica for polarizing the heat. In 

 the summer of 1835, Professor Forbes was at Paris; and finding both 

 M. Biot and M. Melloni sceptical as to his results, he exhibited them 

 with mica piles, which he himself prepared on the occasion, and 

 which he left in M. Melloni' s hands. 



In these experiments the utmost care was taken to guard against 

 all the sources of fallacy from secondary radiation, &c. ; but, as Pro- 

 fessor Forbes observed, these always tended to disguise and not to 

 exaggerate the results. One consideration of this kind arising from 

 the mere mathematical question of the different amount of heat which 

 might be radiated from one pile to the other in the two rectangular 



