CONSTITI'TEI) OF SPHERICALLY SYMMKTKICAL MiH.K< TI.KS. 



481 



number of experimental data entering into each of the above determinations of 

 molecular radii, the agreement shown is very good. It is notably better on the 

 hypothesis that the molecules attract one another than on the ordinary elastic sphere 

 theory (the large discrepancy in the case of CO,-N,O is probably due to the value of 

 C ia , which is larger than the value of (J,, for either gas, and this is somewhat improbable). 

 In the case of the latter theory the value of <r + <r' determined directly from diffusion 

 is in almost every case less than that obtained from separate observations of viscosity ; 

 there is a slight tendency for the same thing to show itself with the values deduced 

 from the other theory. This is probably to be attributed to the artificial nature of 

 our hypotheses ; the effective distance of reversal of motion between two different 

 molecules may not really be half the sum of the effective distances for two pairs of 

 like molecules, though the rigid sphere hypotheses assume this. The third hypothesis 

 we have considered does not do so ; but a comparison of diffusion and viscosity on 

 this hypothesis would only give us the relation between the force constants between 

 the molecules of one kind and different molecules, while the interes'ting question is the 

 one we have just tested, viz., how nearly alike will a molecule behave towards another 

 of the same kind, and towards one of a different kind. 



The following table* may be of interest, showing how the values of a- obtained 

 from VAN DEB WAALS' law agree with those obtained from viscosity (on the 

 attracting sphere theory) : 



22. The Variation of the Viscosity of Mixed (-rases mth Temperature. 



This question was deferred till after the similar questions regarding the viscosity 

 of simple gases, and diffusion, had been discussed. The formula we obtained for ^,, 



* The values of a- for the first four gases are calculated from values of b (in VAN DKR WAALS' equation) 

 given by JEANS (' Dynamical Theory of Gases,' p. 141) from the experiments of VAN DER WAALS and 

 ROSE INNES; but the value of v has been taken as 2-77 x 10 1 " instead of 4-0 x 10', as used by JEANS. 

 The values of b for the last four gases are taken from a paper by RuDORF, 'Phil. Mag.,' June, 1909, 

 p. 795. 



VOL. CCXI. A. 3 Q 



