[ 55 ] 



V. On Thermo diffusion of Gases. 

 ByW. Feddersen, of Leipzig*. 



7 THEORETICAL considerations have opportunely been ap- 

 plied by Carl Neumann f to the proposition that if a 

 limited portion of a gas enclosed in a tube of unlimited length 

 (or one that returns into itself) be different in density from the 

 rest of the gas, an artificially produced difference of tempera- 

 ture at the two ends of this portion must occasion a continuous 

 motion throughout the endless cylinder of gas in a determined 

 direction — and, indeed, from the cooler to the warmer end 

 through the limited portion in question, if the gas in this be in 

 the state of condensation. 



By experiment I have answered, in a simple manner, to the 

 requirements of the theory. Already in the beginning of this 

 year (1872) I found that experiment confirmed that theory, in 

 that without exception the expected direction of the motion 

 could be recognized. My experiments, however, did not per- 

 mit the deduction of definite laws — partly because they were in 

 reality only qualitative, in order to prove the fact generally, and 

 partly on account, as it appears, of manifold disturbing influ- 

 ences which were not further ascertained. I therefore thought 

 I must defer their publication until I should have an opportu- 

 nity of furnishing more exact data. Meanwhile the publication 

 of the results of an investigation by M. Louis DufourJ has in- 

 duced me already, although I have only a few, properly speak- 

 ing, qualitative results, to publish what I have hitherto disco- 

 vered, because my observations are the converse of Dufour's and 

 in harmony with them. 



My experiments were performed in this manner : — A substance 

 in the form of powder was stuffed tight into a glass tube so as 

 to form therein an immovable plug. This tube was fixed in a 

 horizontal position ; and the two ends which projected on each 

 side of the plug were each connected air-tight, by means of 

 caoutchouc, with another horizontal glass tube, which was 

 stopped by a drop of liquid at any place in its interior. In this 

 manner, every displacement of the air cylinder contained in the 

 middle portion of the tubes must displace the two drops of 

 liquid at its extremities in the same direction. One end of the 

 plug was now exposed to a constant source of heat, the other 

 being left cold or artificially cooled. Then, without exception, 



* Translated ^froni a separate impression, communicated by the Author, 

 from Poggendorff's Annalen, vol. cxlviii. pp. 302-311. 



t Berichte der Konigl. Sachsischen Gesellschaft der Wissenschaften, 

 Feb. 15, 1872. 



X Archives de Geneve, Sept. 1872, pp. 10, 11. . 



