287 
CONSTITUTION AND TEMPEKATUKE ON MAONETIC SUSCEIUTBILITY. 
peculiarities of the molecular configuration. It has been sliown that the energy 
change which occurs on crystallization is compatible with a volume change of the 
same order of magnitude as that accompanying crystallization, and we may therefore 
interpret the change of volume on crystallization as a magneto-striction effect of the 
local molecular forcive. The magneto-striction effect depends on molecular orientation 
which is proportional to the scpiare of the magnetic forcive (pp. 259-263). In Part HI. 
it was shown that the double refraction of crystalline media can be interpreted as 
due to the magnetic double refraction effect of the local molecular forcive which 
orientates the molecules into a crystalline space lattice. This effect is also propor¬ 
tional to the square of the magnetic forcive and the two effects mutually support 
one another. 
(V.) The above results are interesting in connexion with Tammann’s theory of the 
closed region of stability of the crystalline state, as represented on the pressure 
temperature diagram. Tammann’s experimental work gives an alternative method of 
determining intrinsic pressures, but the results are notably higher than those found in 
other ways. Possibly this is due to extrapolation over a wide pressure range. The 
pressure temperature diagrams showing the fusion curve are instructive in dealing 
with problems relating to thermal evolution (or absorption), and volume changes and 
possible interpretations of these peculiarities in the ferro-magnetic elements, iron, nickel 
and cobalt, have been given (pp. 263-266). 
(VI.) A physical interpretation has been given of the large values of the 
coefficients, N and a'^, of the molecular fields in ferro-magnetic and diamagnetic 
crystalline media respectively. These coefficients are the reciprocals of the limiting 
local susceptibilities of the media under field strengths equal to the respective 
molecidar fields. The local susceptibility of a diamagnetic molecule is comparable 
with that of a ferro-magnetic molecule and the two vary in the same way with field 
strength and temperature. In diamagnetic media, however, magnetic hysteresis will 
be inappreciable, since the molecule as a whole possesses a zero magnetic moment. 
Nevertheless, mechanical hysteresis in diamagnetic media will lie of the same order as 
in ferro-magnetic media (pp. 267-270 and p. 257). 
(VII.) From Tyndall’s experiments on the deportment of paramagnetic and 
diamagnetic crystals in a magnetic field, the positions of the planes of cleavage can be 
traced. These results show that the forces responsible for crystalline symmetry are 
very probably of a magnetic nature. If the forces are of an electrostatic nature, then, 
since an electric field must disclose the same planes of cleavage, the electric and 
magnetic symmetries must coincide. This is not "the case however. The magnetic 
forces are partly due to the valency or boundary electrons whose orbits are 
controlled by the atomic nuclei. The nuclei determine the crystalline sjunmetry, 
indirectly, through the medium of the magnetic forces of the electrons. This 
conclusion is not at variance with the results of X-ray diffraction experiments ; 
the latter determine only the positions of the diffuse diffracting cores and give no 
VOL. CCXX.-A. 2 R 
