278 MAJOR A. E. OXLEY ON THE INFLUENCE OF MOLECULAR 



negative constant for a certain type of peculiarity of molecular configuration. In 

 normally saturated compounds A = 0. The elements carbon, hydrogen, chlorine, 

 bromine and iodine have constant atomic susceptibilities in a large variety of simple 

 and complex organic compounds. This suggests that the origin of the valencies of 

 these elements is also the origin of a definite amount of diamagnetism, under the 

 influence of a definite magnetic field. In other words, a hydrogen atom, in whatever 

 organic compound it is found, has a constant atomic susceptibility equal to 30'5 x 10~ 7 , 

 while the carbon atom has a constant atomic susceptibility equal to 62'5x 10~ 7 , and 

 so on. 



This result of PASCAL'S, in conjunction with 



(1) The enormous magnitude of the local molecular field in diamagnetic media, and 



(2) The conception of diamagnetism as due to an induction effect in oppositely 

 spinning electrons (as developed in Parts L, II. and III.), led me to suspect that the 

 magneton may be a constituent of the diamagnetic hydrogen molecule. The calculation 

 showed* that if there is one electron in each hydrogen atom whose period is equal to 



* 'Eoy. Soc. Proc.,' A, vol. 95, p. 58, 1918. 



At the time this paper was written, I was out of touch with the latest available data concerning the 

 values of AVAGADKO'S constant (N) and the ratio (/m. The calculation was to determine M from the 



relation 



M = 





N . n . . T 



Taking x = molecular susceptibility of hydrogen, 



-61 -Ox 10~ 7 (PASCAL), 

 N - 6-06xl0 23 (MiLLiKAN), 

 (fin - 1 -77 x 10 T e.m.u. (BUCHERER), 

 n = 2, the number of electrons per molecule, 



T = 2 19 x 10~ lr ' sec., the period of revolution for the line Ha, 

 we find on calculation 



M = 16 -3 x 10~ 22 for the moment of the magneton. 



.0 



This gives for ?, as calculated from M = -^-, the value 0'85 x 10~ 8 cm. 



T 



In this connexion it should be pointed out that, on PLANCK'S theory of quanta of energy, the constant 

 h is consistent with the existence of a unit of magnetism. Assuming, as NICHOLSON and BOHR have 



done, that the angular momentum of the electron is an integral multiple of , CHALMERS showed that 



2iTT 



the magnetic moment of the electron orbit is 



M-I.A.. 



m 4?r 



This gives M = 92 4 x 10~ 22 e.m.u., which is 5 times the experimental value of the moment of the 

 magneton. If we leave aside PLANCK'S theory of energy quanta and adopt instead SOMMERFELD'S theory 

 of quanta of action, LANGEVIN showed that a remarkable relation between h and M exists. He found 



M= L. 



h 



m 247T 1 



when the law of attraction between the nucleus and the electron is the inverse square. This gives for the 

 magnetic moment of the electron orbit M = 15- 4 x 10~ 22 e.m.u., a value nearly equal to the most recent 

 experimental value of the moment of the magneton, viz., 18 '5 x 10~ 22 e.m.u. 



