ON THE ABSORPTION AND RADIATION OP HEAT. 353 
transmission of radiant heat through gases, would be to pass 
heat from a constant source through an exhausted vessel, and 
to determine the amount transmitted ; to fill then the vessel, 
with the gas or vapour to be examined, and determine, whilst 
the vessel is still exposed to the same source of heat, the 
transmitted heat again. The difference between both results 
would express the heat absorbed by the gas. This method is 
not sufficiently sensitive in the case of air and some other 
gases. Mr. Tyndall therefore had recourse to a most ingenious 
process, which will be best understood from the following 
description. 
The instrument used for measuring the heat was the well- 
known thermo-pile of Seebeck and Melloni. It consists of a 
bundle of bars of bismuth and antimony, the single bars 
separated from each other by gypsum or some other non-con- 
ductor of electricity, but soldered together alternately at their 
ends. The two end-pieces of this so-called pile are connected 
with the copper wire of a galvanometer. Rays of heat impinging 
on one of the faces of the thermo-pile, develop in the latter 
an electric current, which during its passage through the 
copper wire deflects the magnetic needle of the galvanometer. 
An increase of the heat would augment the deflection of the 
needle. If we suppose our source of heat and its distance 
from the pile so chosen, that the radiant heat after passing 
through an exhausted tube would cause a deflection of 1 0° of 
the needle, then admission of an absorbing gas would cut 
off some of the rays of heat, and the needle would recede to a 
proportionate extent. W e will assume the needle to point to 
9 x 9 o° after admission of air. A difference of T V° could scarcely 
be appreciated, and would therefore be a very doubtful result. 
The absolute amount of heat to cause the needle to move 
through an arc of 10° is very small. A moro powerful flux of 
heat would increase the deflection and also the absorption. If 
we suppose the needle to have been deflected 81° bv radiant 
heat passing through an exhausted tube, an absorption of 
1 per cent, of this amount could scarcely be detected by a 
retrograde movement of the needle, for it requires much more 
heat to move the needle from 80° to 81°, than from 9° to 10°; 
the galvanometer is more sensitive when the needle is near 
zero. Hence, if the instrument could be kept in this sensitive 
condition, whilst the substances to be examined are exposed 
to a very powerful flux of heat, the attending’ absorption might 
be detected. 
If the needle of the galvanometer is in its normal position, 
heat, which falls on one side of the thermo-pile, will always cause 
a deflection in the same direction ; but if the heat is allowed to 
meet the opposite face of the pile, a deflection opposite in 
