explained that the vibrations of the individual ether-particles are
executed _across_ the line of propagation. In the case of ordinary
light we are to figure the ether-particles as vibrating in all
directions, or azimuths, as it is sometimes expressed, across this
line.

Now, in the case of a plate of tourmaline cut parallel to the axis of
the crystal, a beam of light incident upon the plate is divided into
two, the one vibrating parallel to the axis of the crystal, the other
at right angles to the axis. The grouping of the molecules, and of
the ether associated with the molecules, reduces all the vibrations
incident upon the crystal to these two directions. One of these beams,
namely, that whose vibrations are perpendicular to the axis, is
quenched with exceeding rapidity by the tourmaline. To such vibrations
many specimens of the crystal are highly opaque; so that, after having
passed through a very small thickness of the tourmaline, the light
emerges with all its vibrations reduced to a single plane. In this
condition it is what we call _plane polarized light_.

[Illustration: Fig. 27.]

[Illustration: Fig. 28.]

A moment's reflection will show that, if what is here stated be
correct, on placing a second plate of tourmaline with its axis
parallel to the first, the light will pass through both; but that, if
the axes be crossed, the light that passes through the one plate will
be quenched by the other, a total interception of the light being the
consequence. Let us test this conclusion by experiment. The image of a
plate of tourmaline (_t_ _t_, fig. 27) is now before you. I place