344 M. MELLONI ON THE POLARIZATION OF HEAT. 
take a sufficiently thick plate of tourmaline cut parallel to the axis, and 
attenuate it obliquely at one side, so that the planes of its two faces 
may intersect each other exactly on one of the edges, and thus form a 
certain angle. A very narrow slip of white paper, or any other object 
equally minute, if viewed across this angle and in the direction of the 
edge, will give two images neither lying one upon the other nor con- 
founded together, as they are when the faces are parallel, but separated 
by the double refraction of the tourmaline. If we achromatize the re- 
fringent angle by means of a glass prism in order to have a clearer view, 
we find that these two images, viewed through the thinnest part of the 
tourmaline, are nearly of the same intensity ; but by passing the thicker 
parts successively before the eye, we perceive the image formed by the 
ordinary refraction becoming gradually weaker and weaker until it is 
finally extinguished. 
Thus we see that it isin consequence of the unequal absorption of 
the two pencils formed by double refraction that the polarization in a 
plate of tourmaline becomes perceptible. If the absorbency of the mat- 
ter of which the plate is composed acted with the same intensity on each 
of them, the two pencils would emerge intermixed, and exhibit all the 
properties of ordinary light ; so that a second plate of tourmaline would 
no longer, by having its axis placed transversely to that of the first, pro- 
duce any diminution of intensity in the light transmitted. 
Let us now apply these notions to calorific polarization. Let us sup- 
pose that all the rays of heat, like those of light, undergo a complete 
polarization as soon as they enter a plate of tourmaline, and that each 
of them is consequently divided into two pencils or bundles possessing 
equal intensities and polarized at right angles. Let us admit, besides, 
that the inequality of absorption effected by the matter of the tourma- 
line in the two pencils varies with the different calorific rays; that it is 
very great with respect to some rays, and little or none with respect to 
others : it is evident that the former will issue from the tourmaline en- 
tirely polarized in one plane, while the latter will be more or less po- 
larized in the two planes standing at right angles to one another, and 
will therefore present little or no appearance of polarization. 
All the facts which we have stated may then be explained on the hy- 
pothesis of a complete polarization of the calorific rays; and we shall 
see, indeed, that this hypothesis is rendered more and more probable, nay, 
certain, by the following experiments. But before we conclude our 
observations on this part of the subject, it will not, perhaps, be useless 
to set in contrast, by means of an example easily to be comprehended, 
the two different effects which the tourmalines have on the rays of light 
and the rays of heat. 
Let us imagine a series of alcohol flames coloured by different salts, 
and several common flames masked by glasses of different colours. If 
