Radiant Heat by Dry and by Moist Air, 253 



ineffectually done that no (even approximately) constant position 

 of the galvanometer-needle could be secured. Besides this, strong 

 motions of the needle were caused whenever, by working the 

 pumps, dry or moist air or coal-gas was allowed to enter or leave 

 the tube through a lateral aperture. 



In order to avoid currents of heated air, I arranged the 

 apparatus for the following experiments more accurately, like 

 that of Professor Magnus. In the first place the tube was placed 

 vertically ; the vessel in which water was boiled was placed at its 

 upper end, and its lower portion was immersed in water contained 

 in a large vessel of zinc. Within the tube and near the bottom 

 the thermopile was again placed; and in so doing its lowest end 

 was left free, whilst on its upper end the conical funnel of 

 metal was replaced. This upper end was at a distance of 3 feet 

 from the vessel in which water was boiled, and the radiation of 

 the side walls was diminished by two interposed diaphragms. 

 The lateral aperture for introducing and expelling the several 

 gases lay between the boiling vessel and the uppermost diaphragm. 

 The water in the zinc vessel, from which of course the pole wires 

 of the thermopile were properly insulated, reached in the first 

 experiments to a little above the lower diaphragm, and in the 

 later experiments to the neighbourhood of the lateral aperture, 

 and consequently to a little above the upper diaphragm. 



The observations with this apparatus again showed, in the 

 first place, strong deflections of the magnet-mirror of our galva- 

 nometer when air was leaving or entering the tube; and in 

 fact in the former case these deflections indicated a cooling of the 

 upper side of the thermopile, and in the latter an opposite effect. 

 These deflections were much more intense than those produced by 

 the total radiation of the upper source of heat of 100° upon the 

 thermopile. These are the well-known thermal effects conse- 

 quent upon compression and dilatation, and accordingly they 

 manifested themselves when the boiling water was removed from 

 the upper vessel. The influence of a compression or a dilata- 

 tion of the air is always, in fact, first exerted on the upper side of 

 the thermopile, since the latter is nearer to the place of entrance 

 or exit*. To these thermal actions is probably due, to some 

 extent, the circumstance that I never, even when the water in 

 the upper vessel had been maintained in a boiling state for three 

 hours, could obtain a stationary condition of the temperature, 

 and consequently an approximately constant position of the mag- 

 net-mirror. As a consequence of this, no well-marked difference 



* The thermal effects of the compression and dilatation were in our case 

 so powerful and constant, that I found it most convenient, when showing 

 the same in my lectures, to employ an apparatus constructed on this 

 principle. 



