of Hydrogen and some other Gases. 347 



and hence we cannot attach too much weight to their value. 

 Since Pascal's value was obtained by calculation from the 

 measurement of the susceptibility of organic compounds, 

 we cannot expect that his value will coincide with that 

 of the susceptibility of the gaseous hydrogen directly 

 determined. 



Lastly, we shall consider the above results by experiments 

 in the light of the electron theory. According to Langevin's 

 theory, Bohr's model for hydrogen molecules gives a strong 

 paramagnetism by magnetization, as shown by J. Kunz *, 

 while the observed polarization is diamagnetic. If, however, 

 we suppose the Bohr molecules to be revolving about an 

 axis through the middle point of the line joining two 

 positive nuclei and perpendicular to it, then, as shown by 

 Professors K. Honda and J. Okubo f, the magnetization 

 produces a diamagnetic effect. That is, the velocity of this 

 rotation, though there is no thermal agitation, is supposed 

 to have a definite value characteristic of hydrogen molecules, 

 which is, however, far smaller than the high velocit}^ of 

 revolving electrons ; then the radius and the velocity of 

 t electrons in Bohr's model are very little affected by the 

 characteristic rotation of the positive nuclei. Hence 

 denoting by K the moment of inertia of the molecule about 

 the axis of characteristic rotation, and 12 its angular velocity, 

 then the energy of rotation will be, n being Loschmidt's 

 number, 



The diamagnetic susceptibility per one gram molecule 

 is then 



_ 1 Q"o 2 ctq 2 



X g »EO.« 4bK "°" 



where cr is the magnetic moment of saturation per gram 

 molecule, that is, 



cr =ne(or 2 , 



where e is the charge of an electron, co the angular velocity, 

 and r the radius of the electron orbit. Thus the suscept- 

 ibility of hydrogen gas is diamagnetic, as actually observed, 



* J. Kunz, Phys. Bey. xii. p. 59 (1918). 

 t K. Honda and J. Okubo, loc. cit, 



