396 NATURE 
the temperature of 100°C. The inside we shall suppose 
to be a vacuum. Let us in the first place hang up in this 
chamber two thermometers, one covered on the outside 
of its bulb with lamp-black, the other with polished silver. 
The former of these will absorb all the rays that fall upon 
it from the walls of the chamber, the latter, on the other 
hand, will absorb very few of these rays. Ultimately, 
however, both thermometers will attain the temperature of 
the walls. Since, therefore, according to the theory of 
exchanges the equilibrium of temperature is kept up by 
an equality of absorption and radiation, it is manifest 
that the radiation from the lamp-black thermometer must 
be great, because the absorption is great, and the radia- 
tion from the silvered thermometer small, because the 
absorption is small. 
It will be noticed that this connexion between the two 
qualities, absorption and radiation, is deduced from a 
hypothetical case where everything is at a constant tem- 
perature. To prove it experimentally we may without 
any breach of scientific propriety take out the two ther- 
mometers from the enclosure, exposing them to a lower 
temperature, and noticing their velocity of cooling, when 
it will be found that the blackened thermometer cools 
more rapidly than the silvered one. 
Or we may allow their radiation to fall upon a thermo- 
pile, and to be registered by a galvanometer, when it will 
be found that the indication of the galvanometer will be 
much greater for the blackened than for the silvered 
thermometer. 
Let us next hang up in our enclosure a plate of glass 
and one of polished rock-salt. 
The plate of glass will absorb all or nearly all the rays 
of dark heat that fall upon it from the sides of the en- 
closure. The plate of rock-salt will, on the other hand, 
absorb only a few of these rays. A similar argument to 
that already given will enable us to see that if the theory 
of exchanges be true, the radiation from a plate of rock- 
salt must be decidedly less than from one of glass, and 
this is found to be the case. 
Next, let us hang up two plates of rock-salt,a thick one 
and a thin one. The thick one will absorb more rays 
than the thin one, and we shall therefore expect it to 
radiate more. This, too, will be found to hold experi- 
mentally, thus proving the fact of internal radiation. On 
the other hand, we shall observe no sensible difference if 
we hang up two plates of glass, one thick and one thin, 
the reason being that the thin plate of glass already 
absorbs all the heat which falls upon it, so that no in- 
crement of absorption, and hence of radiation, can take 
place by increasing the thickness. We thus see that it is 
only in the case of diathermanous bodies that the radiation 
increases with the thickness, while for athermanous bodies 
there is no such increase. 
We are now ina better position for realising what takes 
place in our hypothetical enclosure. 
There is a stream of heat from the walls: which falls 
upon any substance which we may introduce into our 
chamber. Now this stream is not altered in intensity by 
altering eithez the shape or substance of the walls. Sup- 
pose, for instance, that they are of polished metal instead 
of being covered with lamp-black, then, while the heat 
radiated from them will be less, the reflexion of this heat 
will be so bandied backwards and forwards between these 
walls as to swell up the total amount to an equality with 
the lamp-black radiation, the only difference being that 
in the lamp-black radiation there is little or no reflexion, 
while in the other there is much reflexion and compara- 
tively little radiation. Nor will the stream from the walls 
be altered by hanging up a plate of any substance between 
them and the body we introduce. For the plate will 
radiate on its own account just as much heat as it absorbs 
from the walls, so that the joint radiation of the two will 
be the same as if the plate were taken away. 
Our remarks have hitherto applied only to the total 
intensity of this stream of radiant heat, and not to its 
quality—that is to say we have left out of consideration 
the specific mixture of various kinds of rays differing 
either in wave-length or in polarisation which go to make 
up the whole heterogeneous radiation. Now a little re- 
flection will convince us that this specific mixture—this 
quality of the radiation-stream—must, as well as its 
guantity remain the same under any change made in the 
shape or substance of the enclosure. For suppose that 
we introduce a thermometer coated with some substance 
which exercises a selective absorption for certain rays of 
the stream, and not for others, then a change of quality 
would mean for this thermometer a change of absorption 
as truly as if there were a change of quantity. But by the 
theory of exchanges the absorption must remain the 
same, being equal to the radiation, and hence this can 
only be brought about by the quantity and the quality of 
the radiation-stream remaining each unaltered whatever 
change be made in the walls of the enclosure, or whatever 
substance be introduced between these walls and the 
thermometer. Carrying out this train of thought, we see 
why, as was proved by Provostaye and Desains, the sum 
of the radiated and reflected heat from any portion of the 
walls must be unpolarised, the reason being that the 
radiated heat from Jamp-black is unpolarised and the one 
radiation must be equal to the other not only in quantity 
but in quality also. Again, the radiation of any surface 
or of any plate must be equal to its absorption, both as 
regards quantity and quality, so that the stream of heat 
may emerge from the surface or from the plate unaltered 
both in quality and in quality. 
Thus the putting up of a plate between the walls and 
our coated thermometer will produce no effect, inasmuch 
as the stream of radiant heat which falls upon the coating 
will be unaltered both in quantity and quality by the 
interposition of the plate. We thus see why the radiation 
from a thin plate of rock salt should be of a quality which 
renders it much absorbed by a cold plate of the same 
material, the reason being that a body radiates that kind 
of heat which it absorbs. 
We see, too, why heat from a thin plate of rock salt 
should be more absorbed by a cold screen of this material 
than that from a thick plate, inasmuch as the former con- 
sists of that kind of heat which is strongly absorbed, even 
by a thin plate, while the latter contains likewise a 
number of other rays which are not so strongly absorbed. 
The conclusion to be derived from these remarks is 
that we have in reality a separate equilibrium for every 
description of heat, an equilibrium which is independent 
of the shape of the enclosure and of the substances of 
which it is composed. Furthermore, the stream of radiant 
heat may be supposed to circulate in the interior of a 
substance such as glass, water, or even metal, the radia- 
tion of each particle which it meets being exactly equal 
to its absorption, so that the stream proceeds through the 
interior, being virtually the same at one part of its path 
as at another. Again, it can be shown that it is essential 
to equilibrium that in the interior of a substance this 
stream of heat should be proportional to the sgvare of the 
refractive index. That is to say, in an enclosure con- 
taining glass whose refractive index is 1°5 the stream of 
radiant heat in the heart of the glass will be 2'25 times 
greater than that proceeding through a vacuum ; we can- 
not, however, tell what takes place in the heart of a 
crystal. It also appears that, for an enclosure of given 
temperature, the stream of a given kind of heat has a 
definite value, the amount of this increasing as the tem- 
perature increases. 
We are, however, ignorant of the exact function of the 
temperature which expresses the value of this stream, but 
we know that this value increases more rapidly for the 
more refrangible rays of the spectrum than for those of 
greater wave-length and smaller refrangibility. 
We now come to consider the luminous rays, and here 
[August 27, 1885, 
