99 DIFFUSION OF GASES 263 



be a retardation of the velocities of flow. Hence the cause 

 of the remarkable slowness of the current of diffusion would 

 have to be sought not so much in the obstruction exerted on 

 each other in their flight by particles of the same kind as in 

 the resistance offered by the oppositely moving particles of 

 the other kind met with. It matters, therefore, the less 

 whether, and how often, particles of the same kind encounter 

 each other ; it is rather the collisions between molecules of 

 different kinds that have special weight in the determination 

 of the phenomena of diffusion. 



These views essentially agree with those which Maxwell l 

 and Stefan 2 have placed at the basis of their theories of 

 diffusion, as they assumed a resistance to diffusion which 

 arises from the encounters of the particles of the one gas 

 with the molecules of the other that meet them, while the 

 collisions between particles of the same kind are to remain 

 without influence on the phenomena. 



If we adopt this assumption, then, the magnitudes % 

 introduced into the formula for the coefficient of diffusion 

 would not really represent the actual free paths of the mole- 

 cules in the mixture of the two gases ; but fij would be the 

 mean value of the path which a particle of the first kind 

 would traverse if it moved solely among particles of the 

 second kind, and likewise 2 would be the free path of a 

 particle of the second kind in a medium composed of particles 

 of the first kind only. Hence in the expression for D the 

 values of ' x and 2/ 2 deduced in 96 must be substituted for 

 fcj and fc a . 



In this way we should obtain for the coefficient of 

 diffusion the value 



from which N { and N 2 have disappeared. It has therefore 

 become independent of the ratio of mixture of the two 

 gases, and is. therefore to be looked on as constant ; for it 

 has one and the same value at all points of the tube and at 

 all times during the progress of the experiment. It depends 



1 Phil Mag. 1868 [4] xxxv. p. 199 ; Scientific Papers, ii. p. 57. 



2 Wien. Sitzungsber. 1872, Ixv. Abth. 2, p. 323. 



