w cup, sensitising the inside
by exposing it for a short time to vapour of bromine. The cup may now be
filled with water, and connection made with a galvanometer by
non-polarisable electrodes. There will now be a current due to
difference between the inner surface and the rod. This may be balanced,
however, by a compensating E.M.F.

[Illustration: FIG. 96.--RECORD OF RESPONSES TO LIGHT GIVEN BY THE
SENSITIVE CELL
Thick lines represent the effect during illumination, dotted lines the
recovery in darkness. Note the preliminary negative twitch, which is
sometimes also observed in responses of frog's retina.]

We have thus an arrangement somewhat resembling the eye, with a
sensitive layer corresponding to the retina, and the less sensitive rod
corresponding to the conducting nerve-stump (fig. 96, _a_).

The apparatus is next placed inside a black box, with an aperture at the
top. By means of an inclined mirror, light may be thrown down upon the
sensitive surface through the opening.

On exposing the sensitive surface to light, the balance is at once
disturbed, and a responsive current of positive character produced. The
current, that is to say, is from the less to the more stimulated
sensitive layer. On the cessation of light, there is fairly quick
recovery (fig. 96, _b_).

The character and the intensity of E.M. variation of the sensitive cell
depend to some extent on the process of preparation. The particular cell
with which most of the following experiments were carried out usually
gave rise to a positive variation of about .008 volt when acted on for
one minute by the light of an incandescent gas-burner which was placed
at a distance of 50 cm.

#Typical experiment on the electrical effect induced by light.#--This
subject of the production of an electrical current by the stimulus of
light would appear at first sight very complex. But we shall be able to
advance naturally to a clear understanding of its most complicated
phenomena if we go through a preliminary consideration of an ideally
simple case. We have seen, in our experiments on the mechanical
stimulation of, for example, tin, that a difference of electric
potential was induced between the more stimulated and less stimulated
parts of the same rod, and that an action current could thus be
obtained, on making suitable electrolytic connections. Whether the more
excited was zincoid or cuproid depended on the substance and its
molecular condition.