NO. II STRUCTURE OF THE ATOM PARSON 7 1 



In the second half of the long periods we find that the elements 

 are -all diamagnetic. For the first three of them, 

 Cu Zn Ga 



3Y + 9 3r + io [37+11], 



f f ^ 



[4y+l] 4y + 2 4y + 3 



this is somewhat surprising, although their 4y phases, certainly, 

 would not be expected to be particularly magnetic. We may note, 

 however, that Cu and Zn are much less diamagnetic than elements 

 which are expected to be so ; e. g., A, Bi, CI, Br, etc. As, Se, Br 

 and Sb, Te, I are diamagnetic for the same reasons as P, S, and 

 CI are. 



As regards the cases just noted where diamagnetism is found 

 instead of the paramagnetism at first expected, we may say that, as 

 already indicated, the interference of free magnetons with one 

 another's orientation, both in the same atom and in neighboring 

 atoms, holds out a possibility of a sufficient explanation, although 

 the rules governing this interference must be very complex and are 

 at present obscure. More definite corroboration of the principles of 

 the magneton theory is got from a consideration of certain com- 

 pounds, which follows now. 



§21. The Magnetic Properties of Compounds 



The typical saturated compound has no free magnetons ; and of 

 most saturated inorganic molecules we can say more : all the valence 

 magnetons of the constituent atoms have gone to make up groups 

 of eight. Thus it is that, while 2 is paramagnetic and the metals 

 Li, Na, K, Be, Mg, Ca, Sr ( ?) , Ba are all paramagnetic, the oxides 



and chlorides of these metals, such as Mg=FO), K-hCU, are 

 without exception diamagnetic. 



The ordinary salts of Fe, Ni, Co, which according to the present 

 views must contain free magnetons, are strongly paramagnetic, while 

 the complex Fe salts and the " cobaltammines," in which, presumably, 

 these free magnetons are bound, are very slightly paramagnetic, or, 

 more often, diamagnetic. 



A table of data illustrating these points and the arguments which 

 follow is given herewith. It was found necessary to supplement the 

 available data by a few determinations for the complex compounds of 

 Cu, Ag, Au, Zn, Hg: the rough method used gave results which 

 agreed with known values (in the case of other substances) to 10 per 

 cent or so. 



