538 M. MELLONI ON THE POLARI'ZATIOT^ OF HEAT. 



and the feeble absorption wliich they suffer from white surfaces* ; and 

 this analogy is completed here by the almost total polarization of the 

 same rays under the influence of the tourmalines. 



It will now be easy to account for the differences between the indices 

 of polarization produced by the different tourmalines. All the calorific 

 rays emitted by the flame of the lamp do not indiscriminately pass 

 through the tourmalines, each of which, according to its nature, is per- 

 meable to particular quantities and qualities of heat. This fact, which 

 is observable in diathermanous substances in general, is so true in the 

 particular instance under consideration, that each of the plates composing 

 the polarizing system indicated in the first table by the numbers 1, 2, 3, i, 

 when combined with a plate of alum, fails to afford an appreciable trans- 

 mission ; an evident proof that the heat which alum is capable of trans- 

 mitting is not to be found in the calorific stream emerging from these 

 four systems. Now we have just seen that the different species of heat 

 contained in the radiation of flame give very different indices of polari- 

 zation. The calorific stream admitted by each polarizing pair will 

 therefore necessarily have a mean index of polarization varying with the 

 quality of the tourmalines. 



If we place outside the polarizing system a screen indued with the 

 same diathermancy as the plates composing this system, namely, a screen 

 permeable to the same species and the same portions of calorific rays, 

 then the effect of absolute transmission will, no doubt, be more or less 

 diminished in proportion as this screen is more or less diathermanous, 

 but the tourmalines will present no change in their index of polarization: 

 such is the case with respect to the white, the red, the orange, the yel- 

 low, the blue, the indigo, and the violet glass. Water, oil, amber, alum, 

 green, or opake-black glass affect this index in a greater or less degree, 

 because their diathermancy differs from that of the tourmalines em- 

 ployed. 



But let the polarizing system be changed. It is clear that, if the new 

 system has not the same diathermancy, the order and the direction of 

 the variations produced by the different screens in the value of the in- 

 dex of polarization will be no longer the same. The following is in fact 

 a series of observations made on the pair of green tourmalines marked 

 No. 5 in the first table : 



* yinn. de Chim. et de Phys., p. 390. 



