of dark Heat-rays by Gases and Vapours. 9 



When vapour forms in the empty chamber M from the 

 liquid contained in the space below it, a portion condenses on 

 the walls and runs down unseen. When the chamber is shut 

 off from the liquid, therefore, it does not contain only vapour, 

 but also the liquid which has condensed on the walls. Hence 

 when communication is made with the large tube, not only 

 does the vapour diffuse itself through the tube, but the liquid 

 condensed in M also evaporates. There can be no doubt that 

 condensation takes place in the large tube also ; but since the 

 evaporation is limited, equilibrium is attained. Hence it 

 seems to us we have an explanation of the results of the above 

 calculations; and at the same time we see that vapour-adhesion 

 is so important a source of error in TyndalFs experiments that 

 all his results must be considered doubtful. Still further, in 

 his later experiments Tyndall changed the method of introdu- 

 cing the vapour into the tube, allowing it to enter directly from 

 a small flask in which the liquid was contained free from air, 

 the tension being measured by a mercurial manometer. The 

 enormous condensation which we have previously found on 

 the walls of the small chamber M, now takes place in the large 

 tube. 



Another inaccuracy must be pointed out in Tyndall's method 

 of dealing with percentages. It is the result of the method 

 of compensation which he employs. The sensitiveness which 

 is attained by giving the absorption in units of the galvano- 

 meter, the first degree =1, is more than compensated by the 

 inaccuracy in determining the deflection produced by the total 

 heat in these units, since these deflections amount to 70° or 

 80°, and one degree in this position is equal to about 20 units. 



Possibly this is the reason why Tyndall scarcely ever gives 

 the deflection produced by the total heat. His tables hardly 

 ever permit of our obtaining in any case of interest any thing 

 else than the ratio of absorptions in such a case ; so that it is 

 impossible to express the absorptions in percentages. 



We are acquainted with only one other research on the ab- 

 sorption of radiant heat by vapour, besides the numerous 

 researches of Tyndall in this direction, that, namely, of Hoor- 

 weg*. This experimentalist endeavoured to avoid vapour- 

 adhesion by allowing a stream of vapour to rise freely in the 

 room between the source of heat and the thermopile. It is 

 obvious that as a quantitative determination this method must 

 yield less accurate results than that which we shall immediately 

 describe. Moreover, Hoorweg's chief object was to observe the 

 behaviour of aqueous vapour, and his results are much less in 

 contradiction with ours than are those of Tyndall. We shall 

 * Pogg. Ann. civ. 



