344 



Size of Molecules 



[CH. XIX 



differences, measured algebraically, between the two sets of figures, we find 

 that the order is roughly that of the boiling points of the substances. 



This circumstance suggests that as regards the gases which occur near 

 the bottom of the table, a considerable error may have been introduced into 

 the determination of ^cr, by the neglect of molecular aggregation. 



This consideration is of course by no means independent of that of the 

 last section. The two considerations together seem to suggest that the 

 determination of \a for some of the vapours tabulated, and for mercury in 

 particular, is almost worthless except as regards order of magnitude. 



420. Finally mention must be made of an interesting determination of 

 the size of molecules which does not depend upon the Kinetic Theory at all. 

 Regarding molecules as spheres which are perfect conductors of electricity *, 

 the dielectric capacity K of a medium containing N molecules per unit 

 volume is found to be given by 



K-l 4 



For a gas K will be very nearly equal to unity, so that the relation becomes 



It is therefore possible to calculate a- from values of K. On Maxwell's 

 Theory K is equal to /z 2 where fi is the refractive index for light of infinite 

 wave-length, so that the foregoing equation becomes 



p, - 1 = frNv 3 . 



It is difficult to calculate either //, or K directly, but values of //, for 

 different substances can be compared with great accuracy by an interference- 

 method. In the following table, many of the values are calculated in this 



* Mossotti's Hypothesis: see Maxwell, Elect, and Mag. p. 70 (Third Edn.). 



