254 H. Wild on the Absorption of 



was apparent from the observations between the absorptions by 

 dry and by moist air. And in fact such a difference could not 

 have been detected, in consequence of the unsteadiness of the 

 magnetic needle, since its effect would have corresponded at 

 most to a few scale-divisions; for when I introduced coal-gas 

 in place of dry air, the diminution of the deflection which the 

 latter had produced did not amount to more than about forty 

 scale-divisions. It was the more surprising to me, therefore, 

 that even after waiting for a long time subsequent to the intro- 

 duction of air or to its exhaustion down to a pressure of 16 mil- 

 lims., the deflection caused by the radiation of the source of heat 

 t in the latter case was always from 100 to 200 scale-divisions or 

 millimetres greater than in the former one. Now since the differ- 

 ence of the absorptions by very rare and by dense air can certainly 

 not have caused this large difference between the deflections, we 

 must conclude that in these experiments also conduction and 

 convection had not been quite removed. It is true that, ordi- 

 narily, we assume that when heat is applied from above, as in the 

 present apparatus, no currents of heated air can occur ; to me, 

 however, this assumption does not appear to be well founded. 

 In the neighbourhood of the source of heat the surfaces of the 

 tube, or at least the interior one, must gradually acquire a higher 

 temperature by conduction, as well as by the effect of radiation, 

 even when the tube is constantly cooled by contact with water; 

 the heated sides of the tube must soon raise the temperature of 

 the adjacent air, which thus becomes specifically lighter, rises, and 

 gives place to other air ; and this process must continue until the 

 horizontal stratum of air under consideration has the same tem- 

 perature as the sides of the tube. As soon as this occurs, how- 

 ever, the temperature of the sides of the tube at this height may 

 again rise a little, or the same temperature may establish itself 

 at a lower place, and thus, under the influence of the sides of the 

 tube, the heat may gradually be propagated downwards from 

 stratum to stratum, more through the mechanical motions of the 

 heated air than through a proper conduction of the latter. 



The insufficient sensitiveness on the one hand, and the above- 

 described currents on the other ? at length induced me to abandon 

 the experiments by Magnus's method. 



Although, therefore, this method of investigating absorption 

 may, in the hands of so experienced and expert an experimenter 

 as Professor Magnus, be an appropriate one for determining ab- 

 solute values with great certainty, I feel bound, from my own 

 experiences, to give a decided preference to TyndalFs method, 

 not only on account of the greater facility with which it fur- 

 nishes qualitative results, but also in consequence of its greater 

 delicacy. It is principally in consequence of this greater delicacy 



