must follow, for
example, Professor Lorentz or Dr Larmor, and look with them for a mode
of representation which appears, besides, to be a natural consequence
of the fundamental ideas forming the basis of Hertz's experiments.

The moment we look upon a wave in the ether as an electromagnetic
wave, a molecule which emits light ought to be considered as a kind of
excitant. We are thus led to suppose that in each radiating molecule
there are one or several electrified particles, animated with a
to-and-fro movement round their positions of equilibrium, and these
particles are certainly identical with those electrons the existence
of which we have already admitted for so many other reasons.

In the simplest theory, we will imagine an electron which may be
displaced from its position of equilibrium in all directions, and is,
in this displacement, submitted to attractions which communicate to it
a vibration like a pendulum. These movements are equivalent to tiny
currents, and the mobile electron, when animated with a considerable
velocity, must be sensitive to the action of the magnet which modifies
the form of the trajectory and the value of the period. This almost
direct consequence was perceived by Lorentz, and it led him to the new
idea that radiations emitted by a body ought to be modified by the
action of a strong electromagnet.

An experiment enabled this prevision to be verified. It was made, as
is well known, as early as 1896 by Zeeman; and the discovery produced
a legitimate sensation. When a flame is subjected to the action of a
magnetic field, a brilliant line is decomposed in conditions more or
less complex which an attentive study, however, allows us to define.
According to whether the observation is made in a plane normal to the
magnetic field or in the same direction, the line transforms itself
into a triplet or doublet, and the new lines are polarized
rectilinearly or circularly.

These are the precise phenomena which the calculation foretells: the
analysis of the modifications undergone by the light supplies,
moreover, valuable information on the electron itself. From the
direction of the circular vibrations of the greatest frequency we can
determine the sign of the electric charge in motion and we find it to
be negative. But, further than this, from the variation of the period
we can calculate the relation of the force acting on the electron to
its material mass, and, in addition, the relation of the cha