of Dry and Moist Air. 27 



gas to be examined is brought into immediate contact with the 

 source of heat. The latter, according to Prof. Tyndall, is thereby- 

 cooled; and this cooling is said to take place at different rates 

 with the various gases and with a vacuum. In support of this 

 assertion, Prof. Tyndall quotes an experiment* in which, his 

 " front chamber " being full of air, it produced a much smaller 

 effect on the thermo-pile than when exhausted of air, the state 

 in which it was commonly employed. 



This front chamber, however, consists of a horizontal cylinder 

 which is heated at one end, while the entire cylinder is surrounded 

 with water in order to keep it cold. Under such circumstances 

 it is plain that no one could doubt that a circulation of air must 

 arise, and that this would produce a cooling effect. In my ap- 

 paratus the heat is applied from above — an arrangement which 

 precludes the formation of internal air-currents, unless indeed a 

 lateral cooling or heating should also take place. The essential 

 thing about the apparatus is the heating from above, and accord- 

 ingly this ought not to be lost sight of in estimating its merits. 



I have modified somewhat the arrangement of the apparatus 

 which I formerly employed, and which I described in Poggen- 

 dorff's Annalen, vol. cxii. p. 516. By reference to fig. 2, Plate I. 

 it will be easily understood. I am thus enabled to construct the 

 tube R R, which serves to contain the gas under examination, 

 out of the several glass tubes R A, B C, D R, and to divide it 

 into sections of various lengths by interposing plates of rock-salt, 

 or of any other diathermanous material. Each section can be 

 pumped out separately and filled with whatever gas may be 

 desired. Moreover, such gases as would injure the air-pump can 

 be passed through any particular section so as to displace the air 

 contained in it. The lower part, E F, of the tube is somewhat 

 wider than the rest, and contains the thermo-pile a c provided 

 with its conical reflector ah. The conducting wires of the 

 thermo-pile pass insulated through the brass plate F F, which 

 closes the bottom of the tube. The portions which project from 

 this plate are surrounded with caoutchouc, as is also the whole 

 length of the wires which pass through the water M N whereby 

 the lower part of the apparatus is surrounded, and which is kept 

 constantly at 15° C. 



In order to be able to unite easily the separate tubes, they are 

 provided at each end with a brass mounting carrying a projecting 

 rim. These rims being ground perfectly true and smeared with 

 a very little grease, the tubes are placed one on the top of another ; 

 a strong piece of brass shaped like a horse-shoe is then pushed 

 over each mounting so as to press against the projecting rim, 

 and by now screwing together each pair of these horse-shoe pieces 

 * Phil. Trans. 1862, p. 93. [Phil. Mag. vol. xxiv. p. 430.] 



