422 Mr. W. Sutherland on the 



his oscillating-disk apparatus, and were therefore specially 

 suitable for controlling Graham's results. But in 1879 Puluj 

 went into the subject more thoroughly by carrying out with 

 an oscillating-disk apparatus a large number of measurements 

 of the viscosity of mixtures of carbonic dioxide and hydrogen, 

 which had proved to be the most exceptional of all with Gra- 

 ham : his results were completely confirmatory of Graham's. 



But to my knowledge the theory of this interesting phe- 

 nomenon has not hitherto been undertaken, though apparently 

 so suited to serve as a test-case for the kinetic theory. Doubt- 

 less the general principles of the kinetic theory are now so 

 well established as to hardly require further confirmation ; 

 but still, apart from the interest in the problem for its own 

 sake, it will be advantageous to consider the theory of it, 

 because in pushing on with investigations of the properties 

 of molecules it is very desirable at the present juncture to 

 ascertain how far the simplest hypotheses will carry us and 

 where they cease to be sufficient. Moreover an inquiry into 

 the theory of the viscosity of mixed gases offers an additional 

 advantage in throwing light on the theory of the viscosity of 

 a single gas. 



In the theory of the viscosity of a gas we suppose the gas 

 to be confined between two parallel solid planes, one of which 

 is at rest and the other moving in a fixed direction in its own 

 plane with velocity to. It has been proved experimentally 

 that the slipping which takes place at either solid plane can 

 be neglected, unless the distance between the planes is only a 

 small multiple of the mean free path of the molecules of the 

 gas. Neglecting the slipping then, and denoting the distance 

 between the planes by D, the velocity of the gas in a layer at 

 distance x from the fixed plane is wxjD, and on the average 

 each molecule in the layer has this velocity in addition to its 

 velocity of agitation. The viscous action between the two 

 parts of the gas on opposite sides of the parallel plane at 

 distance x from the fixed one, is due to the velocities of agita- 

 tion carrying velocities greater than wx/D from distances 

 greater than x to distances less than x, and vice versa. To 

 follow the conversion of the molar energy due to the velocities 

 ivx/D into molecular energy, that is, to follow the conversion 

 of the molar energy into heat by the action of the gaseous 

 friction, we have to consider the collisions taking place in the 

 layer at x between molecules coming from its opposite sides, 

 and these may be said to come on the average from distances 

 x + l and x — l: consequently the velocity of the centre of 

 gravity of each colliding pair both before and after collision 

 is wx/D, which is the permanent velocity characteristic of the 





