Prof. Forbes on the Refraction and Polarization of Heat. 369 

 A, 2 a, &c. to the values — , — , — , &c. and again diminish 



A £t Zi 



in the same manner to the next limit. When the retardation 

 is — -, '— , —, &c., half the light exactly is depolarized ; it is 



T? Tr 'x 



then circularly polarized ; in other cases, it is plane or ellip- 

 tically polarized. 



73. Similar effects might be expected to occur in the case 

 of heat. But we must recollect that it is even more difficult 

 to obtain homogeneous heat^ than homogeneous lights and that 

 we shall have portions of heat differently depolarized by the 

 same plate, (in consequence of the different character of re- 

 frangibility, indicating a different length of undulation,) exactly 

 as when we operate upon white light. We know that heat of 

 various degrees of refrangibility constitutes the solar heat, and 

 probably all other kinds. Hence, no one plate can com- 

 pletely depolarize all these varieties. As far as my experi- 

 ments go, made similarly to that of (71), heat unaccompanied 

 by light is generally less depolarized by a plate of given thick- 

 ness than heat vividly luminous. In the case of contrasting 

 heat from an Argand lamp with that from incandescent pla- 

 tinum, and heat quite dark, this is strikingly marked, though 

 not so decisively in comparing the two last kinds. If the in- 

 accuracy be not in the experiments, it may very probably 

 arise from the want of homogeneity in the heat just alluded to. 

 The want of any apparent depolarizing power for dark heat 

 in the thin mica film mentioned in [5Q) is now easily explained. 

 Its thickness was such as to polarize (nearly) circularly, the 



mean luminous rays. Its retardation, or o — e was then = — - 



for these rays. But we know from Melloni's experiments, 

 that the heating rays are less refrangible than the luminous 

 rays (I mean in heat from terrestrial sources, as well as that 

 of the solar rays), and that generally in proportion to this ob- 

 scurity. Therefore, on the undulatory hypothesis, their waves 



are longer. Hence a retardation of — for light, would be a 



retardation of less than — , if A be the length of a wave of heat 



from an Argand lamp ; it would be still less for heat from in- 

 candescent platinum, and least of all for dark heat; hence, as 

 the retardation is a smaller fraction of A or approaches zero, 

 the depolarization or the value of E^ approaches zero. This 

 perfectly coincides with the experiment of (56). 



74<. Without attaching much weight to the numerical accu- 

 Third Sei'ies. Vol. 6. No. 35. Mai/ 1835. 3 B 



