336 M. MELLONI ON THE POLARIZATION OF HEAT. 
We were already aware that the rays immediately transmitted through 
bodies differing in their nature pass in very different proportions through 
a given plate of a diathermanous substance*: we were also aware that 
those rays are differently absorbed by the surfaces of certain opake 
bodies+. To these distinctive characteristics we are now enabled to add, 
that they undergo in the same system of tourmalines an apparent vari- 
able polarization. 
We see, in fact, that of every hundred rays of heat transmitted by 
the tourmalines when they are placed with their axes parallel, about 22 
are made to disappear by merely crossing the axes. This proportion suf- 
fers no very decided change in the rays transmitted by the common, the 
red, the orange, the yellow, the blue, the indigo, and the violet glass ; 
but is reduced to +45 or +30 when we employ green or opake-black 
glass ; and when we employ sulphate of lime, yellow amber, water either 
pure or saturated with salts, and alum, the quantity of heat polarized 
amounts successively to 77%, 2080 a5o'0> 1°0'0> and =3,8;- 
It is a fact worthy of remark, that the character derived from the in- 
dex of polarization leads to the same consequences that we have deduced 
from the experiments of transmission. Indeed, the latter analytical 
process had authorized us to admit that the colouring matter introduced 
into the composition of coloured glass merely extinguishes a part of the 
calorific stream transmitted by the colourless glass, without sensibly af- 
fecting the proportions which the groups of rays composing that stream 
bear to each other in respect to quantity; so that the effect produced 
by that matter relatively to radiant heat is analogous to that which 
would be produced relatively to light by brown or blackish substances 
diluted in a liquid having no chemical influence on those substances f. 
Now, since the proportion of heat polarized by the tourmalines varies 
with the quality of the calorific rays transmitted by the different screens, 
the constancy of this proportion between the rays which issue from the 
coloured and those which issue from the uncoloured glass clearly shows, 
as in the experiments of transmission, that the colouring matters do not 
affect the composition of the calorific stream transmitted by the glass. 
True, the green and the opake-black glasses furnish a very marked ex- 
ception; but the experiments of transmission furnish an exception 
completely analogous §. 
* Ann. de Chim. et de Phys., tom. ly. p. 384. 
+ Id., p. 388. 
¢ Id., p. 381. 
§ The same consequences are derived from the experimentsof refraction. 
With this view, one of the faces of the refracting angle of a rock-salt prism is 
covered with a plate of coloured glass, and the distribution of temperature in the 
bands of the spectrum produced by exposing this system to the light of the sun 
is then observed. If we change the colour of the glass, we not only find the 
