he
breaking-down of the sensitive element of the tissue, or both of these.
In such a case we should expect that this fatigue, with its consequent
negative image, would gradually and finally disappear on the restoration
of the retina to its normal condition.

We find, however, that this is not the case, for the negative image
recurs with alternate positive. The accepted theory of fatigue is
incapable of explaining this phenomenon.

In the sensitive silver cell, we found that the molecular strain
produced by light gave rise to a current of response, and that on the
cessation of light an oscillatory after-effect was produced. The
alternating after-effect in the retina points to an exactly similar
process.

#Binocular alternation of vision.#--It was while experimenting on the
phenomena of recurrent vision that I discovered the curious fact that in
normal eyes the two do not see equally well at a given instant, but that
the visual effect in each eye undergoes fluctuation from moment to
moment, in such a way that the sensation in the one is complementary to
that in the other, the sum of the two sensations remaining approximately
constant. Thus they take up the work of seeing, and then, relatively
speaking, resting, alternately. This division of labour, in binocular
vision, is of obvious advantage.

As regards maximum sensation in the two retinae there is then a relative
retardation of half a period. This may be seen by means of a
stereoscope, carrying, instead of stereo-photographs, incised plates
through which we look at light. The design consists of two slanting
cuts at a suitable distance from each other. One cut, R, slants to the
right, and the other, L, to the left (see fig. 111). When the design is
looked at through the stereoscope, the right eye will see, say R, and
the left L, the two images will appear superimposed, and we see an
inclined cross. When the stereoscope is turned towards the sky, and the
cross looked at steadily for some time, it will be found, owing to the
alternation already referred to, that while one arm of the cross begins
to be dim, the other becomes bright, and _vice versa_. The alternate
fluctuations become far more conspicuous when the eyes are closed; the
pure oscillatory after-effects are then obtained in a most vivid manner.
After looking through the stereoscope for ten seconds or more, the eyes
are closed. The first effect observed is one of darkness, due to the
rebound. Then _one_ lumi