480 Roi/al Society of Edinburgh. 



pears to me, from those observations, that the quantity of heat re- 

 flected from different transparent bodies is independent, or nearly 

 so, of the nature of the source of heat, and the diathermancy of the 

 reflecting body, and that at 55° of incidence the intensity of reflected 

 heat is nearly that which Fresnel's theory gives. The substances, 

 however, were not all prepared so as wholly to exclude the action 

 of second surfaces. 



•' I have this winter resumed the subject. I have had an appa- 

 ratus constructed for securing sufficient accuracy in determining the 

 angle of incidence, and I have used reflecting surfaces, both trans- 

 parent and metallic : the former are wedges of plate-glass, by means 

 of which reflection from the first surface only may be observed, and 

 the latter are plane specula of steel and silver. The prosecution, 

 however, of these apparently simple experiments has been attended 

 with unforeseen difficulties ; and although the relative proportions 

 of heat at diff^ereht angles of incidence are now pretty well deter- 

 mined for glass in several cases, I am not prepared to say whether 

 the absolute amount is exactly the same as Fresnel's formula would 

 give, assigning to heat its proper refractive index. It is satisfac- 

 tory, however, to know, that the approximation to it is much greater 

 than direct photometrical measures have yet given, with the single 

 exception of two experiments of M. Arago already referred to ; and 

 that I have reason to believe that the experimental law which 

 Mr. Potter has given from direct observation in the case of light, 

 represents my results much less accurately than the theory of 

 Fresnel. 



" With respect to reflection at the metals, I believe I may 

 assert that I have verified the remark of Mr. Potter, that metallic 

 reflection is less intense at the higher angles of incidence. I have 

 attempted to ascertain whether it reaches a minimum, and then in- 

 creases up to 90*^ of incidence, as Mr. MaccuUagh supposes, but 

 I have not obtained decisive results. The quantity of heat reflected 

 by the metals is so much greater than Mr. Potter's estimate for 

 light, as to lead me to suspect that his jihotometric ratios are all too 

 small, which would nearly account for their deviation from Fresnel's 

 law. 



" The most complete verification of Fresnel's law would be found 

 in observations made on heat polarized in opposite planes. These 

 I have attempted, and, so far as they go, they seem to confirm the 

 analogies of heat and light. But the intensity is so much reduced 

 in the process of polarizing, that I fear we must wait for yet more 

 delicate instruments to measure it. In the mean time I may state 

 the method which I propose to employ. 



" When heat is polarized by transmission through inclined mica- 

 plates, the polarization is incomplete ; but if the plates be inclined 

 at the polarizing angle, the transmitted heat is undoubtedly com- 

 posed of a portion p polarized (in a plane which we will call + ), 

 and a portion 1 —p unpolarized. This latter part philosophers are 



\—p 

 content to consider as compounded of a part —^- polarized -f , and 



