4 6 



SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 65 



logues being, in that respect, the new features of the present treat- 

 ment, §§7, 13). Also the fact that the maxima of magnetic suscepti- 

 bility are almost coincident with the minima of atomic volumes (even 

 in the case of Cu salts) takes on a new significance when the forces 

 between atoms are attributed to specific attractions between mag- 

 netons, as they are here. 



Evidence about the volumes of isolated atoms is not entirely lack- 

 ing; and, as might be expected, it has been got from the behavior 

 of gases, especially those with monatomic molecules. If /x is the 

 refractive index of a monatomic gas for very long waves, and the 

 atom is assumed to be made up of electrons within a uniform positive 

 sphere, it can be shown that /x— 1 is proportional to the volume of this 

 sphere. Values for /* have been got for various gaseous elements 

 by Cuthbertson and Metcalfe (Phil. Trans. A., 207, 138, 1907), and 

 the corresponding values for /x-i are tabulated in Thomson's " Cor- 

 puscular Theory of Matter," chap. VII, p. 165. I have tabulated 

 them below together with the atomic volumes of the elements in the 

 solid or liquid state, their atomic weights, and their magneton 

 numbers (N). 



♦Hydrogen . 



Helium 



Neon 



Argon 



Krypton 



Xenon 



Zinc 



Cadmium. ... 

 Mercury. . . . 



*Oxygen. . . . 

 *Sulphur . . . 

 ♦Selenium.. . 

 *Tellurium . 



♦Nitrogen . . 

 ♦Phosphorus 

 ♦Arsenic. . . . 



0*— i)xio 6 



oc gaseous atomic 



volume 



oc Normal volume of 

 the positive sphere 



16 

 24 

 40 

 56 



34 

 50 

 82 



14 

 22 

 38 



54 



13 

 21 



37 



: Not monatomic. 



It may be seen from this table that the values of ju,— 1 for the 

 elements of any group show a much greater parallelism to the mag- 

 neton numbers than do the ordinary atomic volumes (except in the 

 single case of Hg) ; but if these values are plotted against the 



