January 13, 1923] 
stances, which show generally a larger and positive 
induced moment, varying inversely as the absolute 
temperature, and (3) ferro-magnetic substances, which 
show still larger positive magnetic moments, which 
vary with the temperature and external field in a 
complex way. In each case the total induced moment 
per gram of the substance, divided by the applied 
field, is called the specific susceptibility. On Lan- 
gevin’s theory, a molecule consists of a congeries of 
revolving positive and negative charges ; if the total 
initial magnetic moment of these is zero, the substance 
is diamagnetic, if it is not zero, the substance is either 
agnetic or ferro-magnetic. The diamagnetic 
effect must exist in all matter, but is masked by the 
larger para- or ferro-magnetic effects in the latter 
case. Langevin’s theory indicated that when there is 
no interaction between the molecules, the diamagnetic 
effect is independent of temperature, while the para- 
magnetic effect varies inversely as the absolute tem- 
perature in accordance with the Curie rules mentioned 
above. Langevin did not consider ferro-magnetism ; 
this was done by Weiss. 
These remarks hold only in so far as there is no 
appreciable mutual action between the molecules. 
In ferro-magnetic substances, such action is pro- 
nounced, and Weiss (1907) extended Langevin’s theory 
by introducing an intrinsic molecular field to represent 
this mutual molecular interference. According to 
Weiss the molecular field has not necessarily a mag- 
netic origin ; it corresponds to the forces determining 
crystallisation, but for magnetic purposes it may be 
ed as a magnetic field proportional to the in- 
tensity of magnetisation, and its value is then of the 
order 107 gauss. Weiss further showed that the 
energy of this field is a measure of the thermal change 
when, at the critical temperature, the substance 
passes from the ferro-magnetic to the paramagnetic 
state. The results obtained with magnetite above 
the critical temperature showed that Curie’s rule of 
etism held but that the constant of pro- 
portionality had a series of different values over 
certain temperature ranges. These values were inter- 
preted by Weiss as due to sudden changes of the molec- 
ular magnetic moment by a unit, the value of which 
was found to be 16:4 x 10~™ c.g.s.e.m.u. This is the 
Weiss magneton ; its value has later been corrected 
to 18-54 x 10° *. Weiss and others claim that this 
unit exists in many ferro-magnetic substances and in 
paramagnetic salts, though in the latter substances 
the evidence is not quite so conclusive. Further 
aya and theoretical extensions of the work have 
made by Weiss, Kunz, Honda, and Frivold, but 
_ lack of space prevents an extended account of these 
_ parallel to the field. 
here. 
Honda (1910) made an extensive examination of the 
variation of susceptibility of many elements with 
temperature, and concluded that, in general, the Curie 
rules did not hold. In 1914 he submitted that the 
magnetic moments of molecules were not constant but 
depended on the temperature, and that they exert 
forces on one another which hinder their lining up 
In solids which are paramag- 
netic, the magnetic unit is a spherical group of mole- 
cules. This sphere becomes elongated in the ferro- 
magnetic state. A second theory due to Honda (1914) 
NO. 2776, VOL. 111] 
NATURE 
55 
advocates a gyroscopic motion of the molecule to 
account for diamagnetism and paramagnetism. This 
is very similar to the theories of Gans (1910-1916). 
There appears to be no doubt that certain gyroscopic 
motions are involved, but more recent evidence (see 
below) indicates that these do not arise from molecular 
rotations but from a gyroscopic property of the electron 
itself, ice., the electron is a magneton. 
Certain departures from the Curie rules for para- 
magnetic crystals at low temperatures have been 
examined by Onnes, Oosterhuis, and others, and inter- 
preted in terms of a molecular field in a manner similar 
to that of Weiss. 
The variation of diamagnetism accompanying the 
transition from the liquid to the crystalline state has 
been investigated by the writer (1911-22), who found 
that organic compounds changed their specific suscep- 
tibility by a few per cent. The theoretical explanation 
of the results was obtained by including in the Langevin 
theory of diamagnetism a term depending on the local 
polarisation which determines a local molecular field. 
Weiss regarded his molecular field as uniform, but in 
the present case it must be of an alternating character 
as we pass from molecule to molecule of the crystal. 
It exists whether the substance is subjected to an 
external field or not, and distorts the electron orbits, 
producing a few per cent. change of specific suscepti- 
bility on crystallisation. It can be shown that (1) 
these local fields are of the same order of intensity as 
Weiss’s field, namely, 107 gauss, (2) the energy of this 
field is a measure of the latent heat of fusion, (3) the 
existence of such a field would induce a magnetic 
double refraction which is comparable with the natural 
double refraction of crystals, (4) the change of volume 
on crystallisation is the magneto-striction effect of 
this molecular field, and (5) the energy of the local 
molecular field per unit volume represents the tensile 
strength of the crystal. 
Thus it appears that in all crystalline media there 
are intense local fields, the linking up of which from 
molecule to molecule determines the rigidity of the 
crystal. We are not certain what is the true nature of 
the field ; it is probably partly electrostatic and partly 
magnetic. That the magnetic forces are important in 
determining the distribution of the planes of cleavage 
in crystals has been emphasised by the writer (1920), 
a uniform magnetic field being capable of isolating the 
cleavages, i.e., of distinguishing between an open or 
close packing of the molecules in certain directions. 
The present position of magnetic theories is fascinat- 
ing. There appears to be evidence that the ultimate 
magnetic particle is neither the molecule nor the atom 
but the electron itself ; in other words, the electron is 
not merely an electrostatic charge but also a magnetic 
doublet or magneton. Such a structure no doubt 
accounts for the spiral tracks of the #-particles as 
observed by C. T. R. Wilson. The problem of the 
interaction between such doublets in crystalline media 
is far from being solved. It appears that a useful 
picture of the mechanism is obtained on the Lewis- 
Langmuir theory (elaborated by Langmuir in 1919) of 
the cubical atom. In non-ionised media the coupling 
force between atoms is formed by units, each consisting 
of a pair of electrons, and each pair corresponds to a 
single valency bond of chemistry, The influence of 
