imes a sudden
increase in the retinal current, before the usual recovery takes place.
This is very well shown in the series of records taken by Waller
(fig. 106). It will be noticed that on illumination the response-curve
rises, that continued illumination produces a decline, and that on the
cessation of light there is a transient rise of current. I give here a
series of records which will show the remarkable similarity between the
responses of the cell and retina, in respect even of abnormalities so
marked as those described (fig. 107). I may mention here that some of
these curious effects, that is to say, the preliminary negative twitch
and sudden augmentation of the current on the cessation of light, have
also been noticed by Minchin in photo-electric cells.

[Illustration: FIG. 107.--RESPONSES IN SILVER CELL
The thick line represents response during light (half a minute's
exposure), and dotted line the recovery during darkness. Note the
terminal positive twitch.]

#4. Decline and reversal.#--We have seen that under the continuous action
of light, response begins to decline. Sometimes this process is very
rapid, and in any case, under continued light, the deflection falls.

(1) The decline may nearly reach zero. If now the light be cut off there
is a rebound towards recovery _downwards_, which carries it below zero,
followed by an after-oscillation (fig. 108, _a_).

[Illustration: FIG. 108--DECLINE UNDER THE CONTINUED ACTION OF LIGHT
(_a_) Decline short of zero; on stoppage of light, rebound downwards to
zero; after-oscillation.
(_b_) Decline below zero; on stoppage of light, rebound towards zero,
with preliminary negative twitch.
(_c_) The same, decline further down; negative twitch almost
disappearing.]

(2) If the light be continued for a longer time, the decline goes on
even below zero; that is to say, the response now becomes apparently
negative. If, now, the light be stopped, there is a rebound upwards to
recovery, with, generally speaking, a slight preliminary twitch
downwards (fig. 108, _b_, _c_). This rebound carries it back, not only
to the zero position, but sometimes beyond that position. We have here a
parallel to the following observation of Dewar and McKendrick: 'When
diffuse light is allowed to impinge on the eye of the frog, after it has
arrived at a tolerably stable condition, the natural E.M.F. is in the
first place increased, then diminished; during the continuance