ally speaking, were made zinc-like, and the rest copper-like.
From such unfamiliar conceptions, we shall now pass easily to others to
which we are more accustomed. Instead of two opposed surfaces, we may
obtain a similar response by unequally lighting different portions of
the same surface. Taking a sheet of metal, we may expose one half, say
A, to light, the other half, B, being screened. Electrolytic contacts
are made by plunging the two limbs in two vessels which are in
connection with the two non-polarisable electrodes E and E' (fig. 98,
_a_). On illumination of A and B alternately, we shall now obtain
currents flowing alternately in opposite directions.

[Illustration: FIG. 98.--MODIFICATION OF THE SENSITIVE CELL]

Just as in the strain cells the galvanometer contact was transferred
from the electrolytic part to the metallic part of the circuit, so we
may next, in an exactly similar manner, cut this plate into two, and
connect these directly to the galvanometer, electrolytic connection
being made by partially plunging them into a cell containing water. The
posterior surfaces of the two half-plates may be covered with a
non-conducting coating. And we arrive at a typical photo-electric cell
(fig. 98, _b_). These considerations will show that the eye is
practically a photo-electric cell.

[Illustration: FIG. 99.--RESPONSES TO LIGHT IN FROG'S RETINA
Illumination L for one minute, recovery in dark for two minutes during
obscurity D. (Waller.)]

We shall now give detailed experimental results obtained with the
sensitive silver-bromide cell, and compare its response-curve with those
of the retina. A series of uniform light stimuli gives rise to uniform
responses, which show very little sign of fatigue. How similar these
response-curves are to those of the retina will be seen from a pair of
records given below, where fig. 99 shows responses of frog's retina, and
fig. 100 gives the responses obtained with the sensitive silver cell
(fig. 100).

It was said that the responses of the retina are uniform. This is only
approximately true. In addition to numerous cases of uniform responses,
Waller finds instances of 'staircase' increase, and its opposite, slight
fatigue. In the record here given of the silver cell, the staircase
effect is seen at the beginning, and followed by slight fatigue. I have
other records where for a very long time the responses are perfectly
uniform, there being no sign of fatigue.

[Illustrati