234 ANNXJAL REPORT SMITHSONIAN INSTITUTION, 191f{. 



molecule of magnetite the results were more complex and must be 

 divided by 3 to compare them with the atom of iron. These also 

 led to whole multiples of the same number, the factor of propor- 

 tionality changing abruptly at certain temperatures as if the atom 

 of iron underwent corresponding alterations. The number 1,235, 

 of which all the atomic magnetic moments are multiples, will be 

 called the magneton-gram, and its quotient by the Avogadro num- 

 ber (the number of atoms per gram-atom) is the magneton, 

 16.4X10~^'^- The properties of a ferromagnetic body are then well 

 explained by supposing that the magnetic moments of their atoms 

 are simple multiples of a magneton. Magnetism will then have a 

 granular structure like electricity. 



Interesting confirmations have been made of this theory through 

 measures of various experimenters upon paramagnetic salts or, in- 

 deed, upon other bodies. The numbers of Pascal ^ and those of Mile. 

 Feytis^ are in qualitative and quantitative accord with the hypothe- 

 sis of the magneton. As these numbers were calculated with refer- 

 ence to water as a standard, an exact knowledge of the diamagnetic 

 constant of water became necessary. Its measure is difficult and 

 has led to discrepant results. It has been remeasured separately by 

 Seve^ and by P. Weiss and Piccard,* who have reached concordant 

 results close to 0.72X10^° at 20° C. The theory of the magneton 

 thus has had the merit of fixing definitely this important constant. 



We are obliged to admit, however, that for ferromagnetic bodies 

 the atom does not possess a unique magnetic moment, but has a cer- 

 tain number of different values according to the temperature and 

 the chemical compound into which it enters. All these values, how- 

 ever, have integral ratios. The actual existence of the magneton 

 has been demonstrated in the atoms of iron, nickel, cobalt, manganese, 

 vanadium, calcium, mercury, and uranium. We therefore seem to 

 have here a real, very general constituent element of matter. We 

 may therefore think of adding the mag-neton to the other known 

 fundamental elementary bodies. The attempt made hj Langevin^ 

 to deduce the magneton from the quantum of Planck will doubtless 

 serve as a stimulus in this direction. 



V. THE PRODL'CTION AND NATURE OF GASEOUS IONS. 



We will not discuss here the simple, ordinary ions, such as origi- 

 nate from the X rays, radium, the Hertz effect, etc. For several 

 years the accepted theory (Langevin, J. J. Thomson. Townsend. and 

 others) was this: the negative electron, torn from a molecule by the 



1 Pascal, .\nn. CIi. Pliys., vol. 16, p. 531, 1909; vol. 19, p. 5, 1910. 



2 Mile. Feytis, Comptes Rendus, vol. 152, p. 708, 1911. 

 sSeve, Ann. Ch. Phys., vol. 27. p. 189, 1912. 



* Weiss and Piccard, Comptes Rendus, vol. 155, p. 12.34, 1912. 

 6 Langevin, Rapport & la Conference de Bnixelles, 1911. 



