under the usual headings H , EL, H . There is nothing 



Refractivity and Atomic Interaction. 529 



The reader will notice that the interatomic distances R in 

 the actual molecules of the gases H 2 , 2 , N 2 are considerably 

 larger than the corresponding critical values (Re), thus 

 satisfying the condition of optical stability. For the sake 

 of future reference in working out numerical examples, the 

 values of the atomic refractivities, calculated by means of the 

 formula 



N = — ' u = \-«- .... (12) 



Uq — U 



for the first three members of the Balmer 



y 



surprising in the fact that our atomic refractivities differ 



considerably from those hitherto used by chemists (cfr. L, 



p- 94 )- 



Comparison with Results of Kinetic Theory of Gases. — It 

 may be interesting to compare the above interatomic dis- 

 tances, viz., in H 2 , 2 , N 2 respectively, 



22.= . 1-2-22 1-448 1-764.10" 8 , . . (20) 



with the semi diameters \<t of the corresponding molecules, 

 obtained by various methods based on the kinetic theory of 

 gases. Let us take for this purpose Jeans' latest free-path 

 semidiameters * } 



\<j = 1-34 1-81 L90.18" 8 , . . (J.) 



which seem thoroughly well-founded, the more so as they 

 represent, at the same time, the viscosity-, the heat conduc- 

 tion- and the diffusion-values, i. e. all values following from 

 the free-path phenomena. They differ only, as they should, 

 from the Boyle's-Law values of the semidiameters of the 

 corresponding molecules. To avoid confusion, these latter 

 will be denoted by j^a^, and the former, free-path ones, 

 simply by a. 



The above kinetic semidiameters (J.) are considerably 

 greater than our interatomic distances R. The two sets, 

 however, show a marked parallelism, and not only a mere 

 agreement in order of magnitude. A geometric interpreta- 

 tion of this parallelism may be of some interest. 



* J. II. Jeans, Dyn. Theory of Gases, 2nd edition (1916;, p. 341, 

 "where 7r = 2 , 75.10 ln , agreeing with the value I have adopted throughout, 

 is taken for Loschmidt's number. In the 1st edition of his work (1904) 

 Jeans has adopted the considerably higher value 4.10 19 . This I have 

 noticed only through the huge contrast afforded by the figures of Jeans' 

 2nd edition (which is only in a small extent due to the replacing of the 

 original coefficient 044 by, Chapman's 0"499). The number 2*75 or 

 2"76.10 11 ' has been in fairly general use since 1901. 



