parallel to it another plate (_t'_ _t'_): the green of the crystal is
a little deepened, nothing more; this agrees with our conclusion. By
means of an endless screw, I now turn one of the crystals gradually
round, and you observe that as long as the two plates are oblique to
each other, a certain portion of light gets through; but that when
they are at right angles to each other, the space common to both is a
space of darkness (fig. 28). Our conclusion, arrived at prior to
experiment, is thus verified.

Let us now return to a single plate; and here let me say that it is on
the green light transmitted by the tourmaline that you are to fix your
attention. We have to illustrate the two-sidedness of that green
light, in contrast to the all-sidedness of ordinary light. The white
light surrounding the green image, being ordinary light, is reflected
by a plane glass mirror in all directions; the green light, on the
contrary, is not so reflected. The image of the tourmaline is now
horizontal; reflected upwards, it is still green; reflected sideways,
the image is reduced to blackness, because of the incompetency of the
green light to be reflected in this direction. Making the plate of
tourmaline vertical, and reflecting it as before, it is the light of
the upper image that is quenched; the side image now shows the green.
This is a result of the greatest significance. If the vibrations of
light were longitudinal, like those of sound, you could have no action
of this kind; and this very action compels us to assume that the
vibrations are transversal. Picture the thing clearly. In the one case
the mirror receives, as it were, the impact of the _edges_ of the
waves, the green light being then quenched. In the other case the
_sides_ of the waves strike the mirror, and the green light is
reflected. To render the extinction complete, the light must be
received upon the mirror at a special angle. What this angle is we