68 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL, 65 



number of factors in the problem it is hard to say which configuration 

 to predict from the original assumptions. The two factors of mag- 

 netic attraction between the magnetons and electric attraction 

 between each magneton and its positive sphere would make for the 

 first configuration, but the electric repulsion between the magnetons 

 for the second. 



Nor do there seem to be any reliable data for the susceptibility of 

 this gas, the low density of which makes a determination very diffi- 

 cult. Quincke (loc. cit.) obtained for the susceptibility per cc. the 

 value +.008 at I atm., but ±.000 at 40 atm. The only other deter- 

 mination available seems to be that by Bernstein (loc. cit.), who gets 

 the value — .005 for the specific susceptibility at 1 atm., which corre- 

 sponds to an atomic susceptibility of —50 (xio -6 ). But his only 

 other result (according to Landolt's tables) — the value —78 for the 

 atomic susceptibility of gaseous Chlorine at 1 atm. — is nearly four 

 times as great as Pascal's more reliable value (—20.9) for Chlorine 

 in the liquid state, where the element certainly could not be less 

 diamagnetic (see §22). It seems, therefore, that the H 2 molecule, 

 if diamagnetic, is less so, and probably much less so, than two 

 He atoms. 



But the solid alkali metals, whose atoms, like H atoms, contain only 

 one valence magneton, are slightly paramagnetic. In their case, 

 however, one of the factors that made for the diamagnetic configura- 

 tion in the H 2 molecule — the attraction of the positive spheres for 

 the magnetons — is obviously modified by the presence of the groups 

 of eight in such a way as to make the diamagnetic configuration of 

 the valence magnetons less stable than in the case of H 2 : also, the 

 vibration of the neighboring atoms would tend to prevent the forma- 

 tion of a stable positive bond such as the H 2 molecule possesses, thus 

 leaving the magnetons freer than they could be in the latter, which, 

 though isolated, has its two atoms firmly united together. 



With regard to periodicity in magnetic properties, the first short 

 period of the Periodic Scheme is exceptional in some respects, but 

 as we pass along the second short period we find that the presence of 

 1, 2, and 3 valence magnetons makes the atom more and more para- 

 magnetic, though not in proportion to their number. The absence of 

 proportionality may be attributed partly to the superimposed dia- 

 magnetic effect of the two groups of eight, which is probably similar 

 to that of the Neon atom (2y) in all three; and partly to the inter- 

 ference of the valence magnetons, in the same and in contiguous 

 atoms, with one another's freedom. The intra-atomic part of this 

 interference culminates" towards the end of the period in the tendency 





