NO. II STRUCTURE OF THE ATOM — PARSON 45 



sense in which Ag is monovalent, is of doubtful existence, for the 

 mercurous ion has been shown to be double : Hg 2 ++ .) 



The case of Ag is remarkable. We would certainly expect AgCl 2 

 or AgCl 3 to exist, even if they were not as stable as AgCl, but Ag is 

 monovalent in its salts almost without exception. However, this 

 atom's power to form complex ions, which, as I hope to show later, 

 is a characteristic property of these unsaturated atoms and condi- 

 tioned by their unsaturation, is good evidence for the existence of 

 the 3y + 9 phase. 



The whole of this explanation, apart from its simplicity and con- 

 sistency, is strongly supported by the nature of the physical properties 

 of these elements ; for there is marked parallelism between their high 

 melting points, electrical conductivities, and magnetic susceptibilities, 

 and the large numbers of free magnetons in their atoms. Their small 

 atomic volumes are also in accordance with the results of this paper, 

 as may be seen in the next section (§14). 



PART IV. VOLUME 

 §14. The Volume of the Positive Sphere 



The atomic volume of an element in the liquid or solid state is, as is 

 well known, far from being, even approximately, a simple function 

 of its atomic weight, or of the number of magnetons that are in the 

 atom according to the present theory. The elements at the maxima 

 of the well-known atomic volumes curve have atomic volumes that 

 are about seven times as great as those of the elements at the minima. 

 The periodic nature of these fluctuations, however, and their obvious 

 relation to fluctuations in other properties of the elements, such as 

 their valences or melting points (the relation being of an inverse 

 character in these two cases), have made it fairly clear that they are 

 to be ascribed to differences in the forces acting between atoms rather 

 than to corresponding fluctuations in the volumes that the atoms 

 might have if they could be isolated. 



This is the " Hypothesis of Compressible Atoms " for which T. W. 

 Richards has brought forward many kinds of evidence (Faraday 

 Lecture, 1911; Journ. Amer. Chem. Soc, 36, 617-634, 1914; etc.) ; 

 and the final justification for bringing this idea very concretely into 

 the present theory (which has already been done in §7) is that the 

 elements which lie along the minima of the atomic volumes curve 

 are just those to which we have ascribed the maximum numbers of 

 magnetons not bound in groups of eight or tending to form them 

 (the constitutions assigned to (Fe Co Ni), Cu, Zn, Ga, and their ana- 



