and Polarization of Heat. 145 



27. M. Mellon i has proved that the more light that accom- 

 panies heat, the greater power it has to traverse most media, such 

 as clear glass or alum. I made several experiments on the quality 

 of the heat which passed through the tourmalines in their dark- 

 est and in their brightest positions, and I always found that the 

 presence of the light materially increased the power of the heat 

 to permeate such screens, though we have seen how little it add- 

 ed to the quantity. 



28. This fact, namely, that by sifting, as it were, heat sepa- 

 rate from light, we give to it the characters of non-luminous 

 heat, or heat of low temperature, and small refrangibility, such as 

 exists beyond the red extremity of the spectrum, seems so far con- 

 genial with analogy. But according to Melloni's experiments, 

 this does not hold with other degrees of sifting of heat. Thus 

 the absorption of all rays of light, except the blue, the yellow, or 

 the red, by coloured glasses, does not give the peculiar character 

 to the heat which it possesses, when it accompanies light in the 

 process of refraction, namely, that of permeating screens (in ge- 

 neral) more readily as the refrangibility is greater. Hence I con- 

 ceive we must conclude, that heat in the spectrum accompanies 

 the light, and has corresponding properties, but that in general 

 these properties are independent of the nature of the accompany- 

 ing light. 



29. The only fact which appeared to militate against this 

 view, so far as coloured media were concerned, was the case of 

 green light. It appeared probable that this arose from some pe- 

 culiarity in the absorptive nature of the material, not from its co- 

 lour. To investigate this point, I tried the relative transparency 

 (or diathermancy, to borrow a word from M.Melloni) of screens 

 for the heat of various coloured flames. I did not find that 

 marked peculiarity in the green, which M. Mellon i observed in 

 the absorptive action of green glass. The following results are 

 not pretended to be numerically accurate, but they are probably 

 nearly comparable. The flames were obtained from alcohol, com- 

 bined with the following substances : — for the red, nitrate of 



VOL. XIII. PART I. T 



