Radiant Heat by Dry and by Moist Air. 243 



were, in fact, exactly opposite to those which Professor Tyndall 

 observed under like circumstances, so that they corresponded 

 to a diminished absorption due to the moisture of the air. 

 Professor Magnus therefore held to his previously expressed 

 view, that no difference exists between the absorption by dry 

 and by moist air ; and he sought to explain the opposite effect by 

 the fact, recently established by him with greater accuracy*, that 

 all substances are heated when air reaches them which is moister 

 than that which surrounds them, and that they are cooled when 

 impinged upon by air which contains less moisture than that 

 in which they are immersed. Since the principal contradiction, 

 therefore, lies in the experiments made according to Tyndall's 

 method, it was this method that I first employed. 



1. Experiments by Professor Tyndall' s Method. 

 The apparatus (without rock-salt plates) which I employed in 

 this part of the investigations differs but slightly from that of 

 Professor Tyndall. It consists of a thermopile of fifty bismuth- 

 antimony elements collected in a brass cylinder of 2 centims. in 

 diameter, which is provided at both ends with conical reflectors 12 

 centims. in length and having an external opening 6*5 centims. 

 in diameter. Placed on its feet, the thermopile can be moved 

 vertically and turned around a horizontal and a vertical axes. Its 

 soldered parts are, of course, covered over as evenly as possible 

 with lampblack. The poles of this thermopile are connected by 

 conducting wires, first with a gyrotrope, and afterwards with a 

 Meyerstein' s electro- galvanometer. This instrument differs from 

 the one described in PoggendorfFs Annalen, vol. cxiv. p. 132, 

 inasmuch as Herr Meyerstein, at my wish, has fixed the two 

 auxiliary magnets underneath the wooden plate which carries the 

 multiplier (for which purpose the feet of the plate were considerably 

 lengthened), and has suspended the magnet, with its mirror, 

 by a silk thread 60 centims. in length, which at its upper end is 

 fastened to a glass tube carried by a copper stirrup. Besides a 

 multiplier consisting of many windings of a thin wire, the instru- 

 ment is provided with a second one, which was employed in the 

 following experiments, and which consists of only twice 150 

 windings of a wire 1*5 millim. in thickness. Finally, in order 

 more easily to obtain sufficient inertia, Herr Meyerstein has 

 fixed a holder for a second magnet to the mirror- frame above 

 the multiplier. This was likewise used in these experiments, 

 by which means the greater of the auxiliary magnets could be 

 dispensed with. By the approach of the small auxiliary magnet, 

 the moment of inertia was increased so far as to obtain, without 

 damping, a time of oscillation of the system of magnets amount- 

 ing to 25 seconds in one set of experiments, and to 50 seconds in 



* PoggendorfFs Annalen, vol. cxxi. p. 174. [Phil. Mag. S. 4. vol. xxvii. 

 p. 241.] 



112 



