opposite faces of the pile. Hence, if
after the anterior face has received the heat from our radiating
source, a second source, which we may call the compensating source, be
permitted to radiate against the posterior face, this latter radiation
will tend to neutralise the former. When the neutralisation is
perfect, the magnetic needle connected with the pile is no longer
deflected, but points to the zero of the graduated circle over which
it hangs.

And now let us suppose the glass tube, through which the waves from
the heated plate of copper are passing, to be exhausted by an
air-pump, the two sources of heat acting at the same time on the two
opposite faces of the pile. When by means of an adjusting screen,
perfectly equal quantities of heat are imparted to the two faces, the
needle points to zero. Let any gas be now permitted to enter the
exhausted tube; if its molecules possess any power of intercepting the
calorific waves, the equilibrium previously existing will be
destroyed, the compensating source will triumph, and a deflection of
the magnetic needle will be the immediate consequence. From the
deflections thus produced by different gases, we can readily deduce
the relative amounts of wave-motion which their molecules intercept.

In this way the substances mentioned in the following table were
examined, a small portion only of each being admitted into the glass
tube. The quantity admitted in each case was just sufficient to
depress a column of mercury associated with the tube one inch: in
other words, the gases were examined at a pressure of one-thirtieth of
an atmosphere. The numbers in the table express the relative amounts
of wave-motion absorbed by the respective gases, the quantity
intercepted by air being taken as unity.

Radiation through Gases.

Name of gas Relative absorption

Air 1

Oxygen 1

Nitrogen 1

Hydrogen 1

Carbonic oxide 750

Carbonic acid 972

Hydrochloric acid. 1,005

Nitric oxide 1,590

Nitrous oxide 1,860

Sulphide of hydrogen 2,100

Ammonia 5,460

Olefiant gas 6,030

Sulphurous acid 6,480

Every gas in this table is perfectly transparent to light, that is to
say, all waves within the limits of the visible spectrum pass through
it without obstruction; but for the waves of slower period, emanating