Memoirs presented to the Cambridge philosophical society on the occasion of the jubilee of Sir George Gabriel Stokes, bart., Hon. LL. D., Hon. SC. D., Lucasian professor.

AND THE INFLUENCE OF A MAGNETIC FIELD ON OPTICAL PHENOMENA. 395
The polarization of a dielectric medium by an imposed electric field is effected in a
cognate manner. The electric force slightly modifies the orbital system by exerting
opposite forces on the positive and negative ions. In this case these forces are independent of the velocities or masses of the ions. The fact that the polarization is
proportional to the inducing field shows that the influence produced by the field on the
orbital system is always a slight one. Yet the numerical value of the coefficient of
electric polarization is always considerable, in contrast with the very small value of
the magnetic coefficient; which arises from the very great intrinsic electric polarity
of the molecule, due to the magnitude of the electric charge e of an ion. Taking the
effective molecular diameter as of the order 10-8 cm., there will be 1024 molecules per
unit volume in a solid or liquid, and the aggregate of their intrinsic electric polarities may
be as high as 1024. 10- ee electrostatic units, where ec is 3. 10-11. Now the moment
of polarization per unit volume for an inducing field F is (K - 1) F/87r; thus even for very
strong fields this involves very slight change in the orbital configuration. A similar
remark applies to the polarization induced by mechanical pressure in dielectric crystals. It
would be unreasonable to expect any aggregate rotational effect around an axis, such
as constitutes magnetization, from the polarizing action of an electric field; in fact if it
were present, reversal of the direction of the field could not affect its total amount
considered as arising from molecules orientated in all directions.
The possibilities as regards the aggregate intrinsic magnetic polarities of all the
molecules are of the same high order, viz. eAn/T, where A is the area and r the period
of a molecular orbit, which is elnv or 10-5v per cubic centimetre, where v is the velocity
in a molecular orbit whose linear dimension I is 10-8. Thus the superior limit of the
magnetization if the molecules were all completely orientated would be of the order
10-5v, which is large enough to include even the case of iron if v were as much as one
per cent. of the velocity of radiation.
In the case of iron a marked discrepancy exists between the enormous Faraday
optical effect of a very thin sheet in a magnetic field on the one hand, and the slight
Zeeman effect of the radiating molecule, as also the absence of peculiarity in optical
reflexion from iron, and the absence of special influence on Hertzian waves, on the other:
which must be in relation with the circumstance that at a moderately high temperature
the iron loses its intense magnetic quality and comes into line with other kinds of matter.
This suggests the explanation that the magnetization of iron at ordinary temperatures
depends essentially on retentiveness, owing to facility possessed by groups of molecules for
hanging together when once they are put into a new configuration. This is the well-known
explanation of the phenomena of hysteresis, which can be effectively dirninished by
mechanical disturbance of the mass. In soft iron the magnetic cohesion would be less
strong and more plastic, and thus readily shaken down by slight disturbance in the
presence of a demagnetizing field, so that retentiveness would not be prominent. It is
conceivable that the primary effect of an inducing field is to slightly magnetize the
different molecules: that then the molecules thus altered change their condition of
aggregation, and so are retained mutually in new positions independently of the field,
50-