342 M. MELLONI ON THE POLARIZATION OF HEAT. 
Let us first direct our attention to the two last systems of tourma- 
lines. Their indices undergo a gradual increase in passing from the 
copper to the Argand lamp. This shows that the radiation of each of 
the four sources contains a greater quantity of heat capable of being 
polarized by the tourmalines in proportion as the temperature of the 
source is more elevated. 
Yet the indices of polarization of the systems 1 and 5 undergo, from 
the action of the first three sources, changes completely opposed to 
those which we have been considering. In order to account for such an 
anomaly, we must keep in view the very imperfect diathermancy of this 
sort of tourmalines. It is true that the calorific streams of the Argand 
and the Locatelli lamps contain rays more capable of being polarized 
by the tourmalines than any of those contained in the calorific streams 
from the inferior sources; but in the present case such rays scarcely 
contribute to increase the index of polarization; for we have seen that 
they cannot penetrate the plates which form the polarizing systems. The 
green tourmalines, however, are permeable to several species of heat ; 
and, as in the radiation of each source there are several of these species, 
it is easy to see that if a certain group of rays, possessing an index of 
polarization inferior to those of the excluded but superior to those of 
the transmitted rays, is more abundant in the calorific stream of the in- 
candescent platina than in that of the Locatelli, the index of polariza- 
tion in the two systems of green tourmalines will, in this case, suffer 
a decrease in passing from the first to the second source. The same 
reasoning will apply to the Locatelli as compared with the Argand 
lamp; so that, notwithstanding the fact that the rays are more suscep- 
tible of polarization as we proceed from the incandescent platina to the 
higher sources, the two systems of green tourmalines will give lower 
indices of polarization. 
Without a knowledge of the laws of calorific transmission and the 
analytical resources they afford, we should perhaps find it impossible 
to extricate ourselves from the perplexing difficulty presented by these 
singularly anomalous phenomena of polarization. We are now able to 
offer the following brief recapitulation of our observations on them. 
“The different calorific rays coexisting in the radiation of the same 
Argand lamp, gives but a difference of 0°-1 in the two positions of the axes, 
the arcs of impulsion being from 26° to 27°; a difference which could not be 
rendered perceptible otherwise than by taking the mean of 10 observations; and 
I am not yet quite sure that it would not vanish altogether if the experiments 
were more numerous, for I frequently obtained a stronger transmission with 
the axes perpendicular. The faci is that in operating upon ares of from 15° to 
20° the transmission of this system of tourmalines seemed to undergo no vatia- 
tion whatsoever in consequence of the axes being crossed ; and indeed, after the 
preceding experiments, there is nothing more surprising in the existence of tour- 
malines that give no sign of calorific polarization than there would be in the 
discovery of tourmalines capable of completely polarizing heat. — 
a 
