100 DIFFUSION OF GASES 267 



expect that the variation of D with the ratio of mixture, 

 determined by the values of JYj and N 2 , would be quite 

 small. The formula takes the form 



B.N.IN, A + B 



when for shortness we put 



A = wo-VW + IV), B l = irV*W2, 



and it then shows that the first of the two terms becomes 

 smaller when the ratio of N-^ to N 2 increases, but the second 

 becomes larger; the value of D will therefore be able to 

 change with N^ and jV 2 much less than either of these 

 numbers itself changes. 1 



The alteration of D will be very small when the values 

 of the magnitudes B are nearly equal ; and that will occur 

 for the most part. For one of the two gases possesses a 

 greater speed fl, and a smaller molecular section TTS* along 

 with a less molecular weight than the other. 



After these considerations and experiments we cannot 

 assert that the theoretical formula for the coefficient of dif- 

 fusion here obtained has been proved by experiment to be 

 perfectly correct. But it is highly probable from the 

 observations that have hitherto been made that the formula 

 corresponds to the true circumstances. In any case there 

 is no fact known which contradicts it ; on the contrary 

 the other formula developed in 99, which gives a constant 

 value of the coefficient, is certainly inaccurate. 



To finally decide as to the validity of our formula we 

 must, asHausmaniger has already noted, experiment upon 

 gaseous mixtures 2 which contain only small masses of one 

 of the constituents. For when N = the formula gives 

 the limiting value 



D = V/aMrV(*V + n 2 2 ), 



and on the contrary, when A T 2 = 0, the limiting value 



+ j, 



1 Guglielmo, Atti d. Accad. di Torino, 1882, xvii. p. 106; Exner's 

 Repert. d. Phys. 1883, xix. p. 580. 



2 Perhaps by the method recommended by M. Topler, Wied. Ann. 1896, 

 Iviii. p. 599. 



