of Dry and Moist Air. 25 



tainty of the experiment, I have repeated the blowing in of dry 

 and moist air alternately many hundred times ; but in no single 

 case was the deflection such as to indicate a greater absorption 

 by moist air. 



It would be out of place to relate the numerous experiments 

 which were undertaken, partly in order to make myself master 

 of the phenomena, and partly in order to explain the surprising 

 contradiction between my results and the conclusions which Prof. 

 Tyndall has drawn from his experiments. I found, in the first 

 place, that the deflection took place only when the air was driven 

 in with a certain amount of force. It was found, further, that 

 when the air was pressed in continuously, the deflection of the 

 galvanometer was not maintained constant, but that the instru- 

 ment gradually returned to its position of equilibrium. Hence 

 it resulted that the air did not cause the deflection by absorp- 

 tion. I suspected that possibly moisture might be condensed 

 on the internal surface of the tube, and that a heating effect 

 might be thus produced ; but this supposition was likewise 

 found to be erroneous. It appears, on the other hand, that the 

 phenomenon is occasioned by an absorption which takes place at 

 the surface of the pile itself. 



In fact, when air is blown in at the side of the tube, the greater 

 part of it escapes at the end nearest the hole through which it 

 enters, but the air also escapes from the tube at the opposite 

 end. This can be seen very distinctly by bringing a small flame 

 in front of the open ends of the tube. If the air-pump is con- 

 nected with the side of the tube near the other end, it causes the 

 air to escape with less force, or not quite continuously ; never- 

 theless, if the pressure under which the air enters the tube is suf- 

 ficient to cause any of it to escape at the other end, the air-pump 

 is incapable of altogether counteracting the motion, its action 

 being neither sufficiently continuous nor sufficiently powerful. 

 The air which issues from the tube continues its motion in the 

 same direction, and thus reaches, first, the conical reflector, and 

 then the pile itself, even when the latter is at a considerable 

 distance. 



Now, if the air is saturated with moisture, it appears that water 

 is condensed on the surface of the pile; a heating effect is thus 

 produced, and the galvanometer is deflected. If, however, the air 

 is dry, it takes up from the pile the moisture previously condensed 

 upon it, and thus an evaporation and consequent cooling is oc- 

 casioned. 



This explanation at once shows why the galvanometer gra- 

 dually returns to its position of equilibrium when an unbroken 

 current of saturated air is kept up. For when so much aqueous 

 vapour has been condensed that no further condensation can 



