Vapour in Experiments on the Absorption of Heat. 415 



to admit or lead away air. The lateral apertures nearest the 

 thermopile were connected by caoutchouc tubes with two tubes, 

 one of which contained pieces of pumice soaked in sulphuric 

 acid, while the other contained pumice moistened with water. 

 These tubes were connected, by means of a forked tube, with a 

 small compression-pump provided with a windbag, such as is 

 used for the blowpipe-table. In this manner one of the tubes 

 received dry, and the other moist air. In order to cause these 

 to spread in the tube and prevent them from escaping at the 

 nearest open ends, both the other lateral apertures were con- 

 nected with an air-pump. Professor Wild subsequently pre- 

 ferred to connect the apertures most distant from the pile with 

 the compression-pump, and those nearest it with the exhausting- 

 pump. 



Instead of a caoutchouc force-pump I used a somewhat larger 

 double bellows, while instead of pumice I used pieces of glass 

 moistened with water; in other respects the apparatus was the 

 same. The air in my experiments, when not otherwise stated, 

 was blown in by the more distant lateral aperture. I obtained 

 the same results as Professor Wild — that is, heating of the pile 

 by blowing in dry air, cooling by moist ; the deflections of the 

 galvanometer, however, were somewhat greater than he states, 

 probably because the instrument I used was more delicate. 

 Hence it was also not necessary to blow dry air into one tube 

 and at the same time moist into the other, and then to change ; 

 it was sufficient to allow dry, and afterwards moist air to enter 

 the same tube. I soon found also that the exhausting-pump 

 was superfluous; for the deflections were the same, whether it 

 was or was not used. This is not surprising, considering how 

 small is the quantity of air which can be removed by such a 

 pump, and how little must be the action it can exert on a tube 

 open at both ends. 



I soon found, however, that the discrepancy between the pre- 

 sent results and those I had formerly observed depended on a 

 circumstance which I had previously neglected. 



If no air is blown into the two brass tubes, and the needle of 

 the galvanometer is at rest, both sides of the thermopile receive 

 the same amount of heat. If, then, into both brass tubes either 

 moist air or dry air be simultaneously blown, the galvanometer 

 ought to remain at rest, provided both tubes are simultaneously 

 filled with the air. When, however, the experiment was made 

 it was seen that the galvanometer, which was at rest when dry 

 air was forced in, was no longer so on forcing in moist air — that, 

 therefore, both sides of the pile do not receive the same amount 

 of heat. By interchanging the tubes, altering the connexions, 

 &c, I was convinced that the cause of this difference did not 



