M. W. Feddersen on Thermo diffusion of Gases. 61 



After this, the bent end of the tube on the cold side was im- 

 mersed in sulphuric hydrate. The sulphuric acid rose quickly 

 to 53 milliras. height, while the drop received a displacement 

 corresponding to the dimensions of the closed spaces. The po- 

 sition remained unchanged during the first hour ; but after the 

 lapse of the second hour the acid had risen to 6.2 millims.,and after 

 another 1^ hour to about 80 millims., where the bend of the 

 tube commenced; and so any further diminution of pressure in 

 the tube could not be observed. Sustained blowing with the 

 mouth into the open end on the warm side very soon reduced 

 (as was to be expected) the column to zero. 



The numbers observed with magnesia, again, point to an ab- 

 sorption of the gas by the substance in the form of powder. 



From the preceding experiments with the most heterogeneous 

 substances it appears to follow that it is a universal property of 

 porous bodies, when in the form of diaphragms, to draw gases 

 through them in the direction from the cold to the hot side. 

 We have thus a phenomenon of diffusion which, contrary to 

 ordinary diffusion, occurs even when the same gas under the 

 same pressure is found on both sides of the diaphragm. This 

 is a singular, hitherto unknown phenomenon ; and we are there- 

 fore justified in giving it the name of 



Thermo diffusion . 



Whether this kind of diffusion depends also on the nature of 

 the gas, so that with mixed gases a selective diffusion might take 

 place (that is, a mechanical separation of the gases), analogous 

 to that which occurs in the diffusion of liquids, I have not yet 

 been able to discover; it appears at all events probable when 

 Dufour's experiments are taken into consideration. Further, it 

 may be an argument in favour of this view, that I often observed 

 a progressive diminution in the velocity of the motion in the en- 

 closed space, starting from the attachment of the tubes, such as 

 must necessarily occur if, on the cooler side, one of the consti- 

 tuents of atmospheric air diffuses through to the warmer end 

 more quickly than the other ; for the closed space on the former 

 side would be sooner exhausted of the more easily diffusing gas 

 than of the other. 



Dufour, /. c, states that when gases are caused to diffuse, there 

 is a rise of temperature on the side where the more quickly dif- 

 fusing gas enters the porous diaphragm, and a fall of temperature 

 on the opposite side ; so that in his experiments diffusion pro- 

 duces a change of temperature ; in mine a change of tempera- 

 ture causes diffusion, and the latter in a direction such that the 



