134 
MR. A. E. OXLEY ON THE INFLUENCE OF MOLECULAR 
connected with the change of volume on crystallization ; for in the case of water # the 
effect is small, and is normal as far as sign is concerned, while a large expansion takes 
place on crystallization. In the case of benzene there is a fall and a rise of x on 
crystallization, while the change of state involves a contraction simply. 
With regard to the crystalline forms of the elements, Barlow and PoPEf point 
out that those elements which show allotropic modifications crystallize in forms of low 
symmetry, while other elements not so easily obtainable in allotropic forms crystallize 
in forms of high symmetry. It seems necessary, therefore, to distinguish between the 
forces holding the atoms in position in the molecule and the forces holding the 
molecules together in the crystal structure. The atoms of an element are assumed to 
be identical, and the molecular aggregate to which they give rise and the accompanying 
distortion depend upon the kind of packing which these equal atomic spheres of 
influence assume. The only element which has been investigated in this research is 
mercury. The large change of x on freezing indicates a large change of molecular 
configuration. 
The allotropic forms of an element, which possess characteristic crystalline structures, 
will have different molecular distortions and consequently different magnetic proper¬ 
ties. Moreover, as these forms are of different degrees of stability, the disintegration 
of the structure with change of temperature will in general be different for each form. 
The x — ^ curves may therefore be expected to show irregularities. Preference has 
already been made to the work of DU Bois and Honda| on the magnetic properties 
of the allotropic modifications of phosphorus, antimony, carbon, and tin. But these 
authors used the irregularities which their observations disclosed as an argument 
against the truth of the foundations of L angevin’s electron theory of diamagnetism 
instead of interpreting them as due to the variation of molecular complexity. The 
experiments described above show that the peculiarities of the x — ^ curves can in all 
cases examined be traced to the variation of molecular complexity. In the liquid 
state most of the compounds that have been investigated are unassociated, and the 
absence of any variation of x with B is conclusive proof of the truth of the Curie law 
of diamagnetism when the molecules are not appreciably distorted. This law is true 
for distorted molecules, if the distortion is not a function of the temperature, as the 
parallelism of the curves for the various crystals and for the empty phial shows. 
It is clear from the experimental results that diamagnetism is not wholly an atomic 
property, for the near approach of the molecules, which takes place on crystallization, 
* Piccard, ‘ Comptes Rendus,’ vol. 155, p. 1497, December 23, 1912, and Weiss and Piccard, ‘ Comptes 
Rendus,’ vol. 155, p. 1234, December 9, 1912. These authors find that the variation of the diamagnetic 
property of water with temperature can be interpreted as due to the co-existence of two types of 
complexes, one type growing at the expense of the other as the temperature is varied. This explanation 
is in accordance with equation (1) which I had published earlier. 
f ‘ Trans. Chem. Soc.,’ vol. LXXXIX., p. 1741, 1906. 
J 'Versl. Kon. Ak. v. Wetensch.,’ Amsterdam, XII., p. 601, 1910. 
