t varies according to whether the
acceleration is parallel with or perpendicular to the direction of
this velocity. In other words, there seems to be a longitudinal; and a
transversal mass which need not be the same.

All these results would persist even if the material mass were very
small relatively to the electromagnetic mass; and the electron
possesses some inertia even if its ordinary mass becomes slighter and
slighter. The apparent mass, it can be easily shown, increases
indefinitely when the velocity with which the electrified particle is
animated tends towards the velocity of light, and thus the work
necessary to communicate such a velocity to an electron would be
infinite. It is in consequence impossible that the speed of an
electron, in relation to the ether, can ever exceed, or even
permanently attain to, 300,000 kilometres per second.

All the facts thus predicted by the theory are confirmed by
experiment. There is no known process which permits the direct
measurement of the mass of an electron, but it is possible, as we have
seen, to measure simultaneously its velocity and the relation of the
electric charge to its mass. In the case of the cathode rays emitted
by radium, these measurements are particularly interesting, for the
reason that the rays which compose a pencil of cathode rays are
animated by very different speeds, as is shown by the size of the
stain produced on a photographic plate by a pencil of them at first
very constricted and subsequently dispersed by the action of an
electric or magnetic field. Professor Kaufmann has effected some very
careful experiments by a method he terms the method of crossed
spectra, which consists in superposing the deviations produced by a
magnetic and an electric field respectively acting in directions at
right angles one to another. He has thus been enabled by working _in
vacuo_ to register the very different velocities which, starting in
the case of certain rays from about seven-tenths of the velocity of
light, attain in other cases to ninety-five hundredths of it.

It is thus noted that the ratio of charge to mass--which for ordinary
speeds is constant and equal to that already found by so many
experiments--diminishes slowly at first, and then very rapidly when
the velocity of the ray increases and approaches that of light. If we
represent this variation by a curve, the shape of this curve inclines
us to think that the ratio tends toward zero when the velocit