ACTION OF LIGHT ON SELENIUM. 
341 
Deflections. 
(1) + Current in at tlie marked end 55 to the left. 
ditto ditto light on marked end 47 „ „ 
(2) + Current in at the unmarked end 56 to the right. 
ditto ditto light on marked end 66 „ „ 
(3) -f- Current in at the marked end 62 to the right. 
ditto ditto light on unmarked end 72 „ „ 
(4) + Current in at the unmarked end 56 to the left. 
ditto ditto light on unmarked end 48 „ „ 
Hence the experiments with all three pieces point to the same conclusions, viz. : — 
(1) That when light falls on the end of the selenium at which the positive current 
from the thermoelectric pile is entering, it opposes the passage of the current. 
(2) That when light falls on the end of the selenium at which the positive current 
from the thermoelectric pile is leaving the selenium, it assists the passage of the current. 
(3) That when no battery-current is passing, the action of light is to cause a flow of 
electricity from the selenium to the platinum at the illuminated junction. 
We have been led to infer, from the experiments with those pieces of selenium through 
which no battery-current had ever been passed, that the currents thus produced in the 
selenium by the action of light are not thermoelectric currents due to the heating 
of the junctions between the platinum electrodes and the selenium, from the following 
considerations : — 
(1) Assuming that platinum stands above selenium in the thermoelectric scale, the 
positive direction of the current, due to heating a junction of platinum and selenium, 
would be across the heated junction from the platinum to the selenium. Hence in all 
the above experiments the current, supposing it to have been a thermoelectric one, 
Avould always have been away from the illuminated end — that is to say, from the platinum 
to the selenium. In the above experiments, however, we find that, in general, the 
current is from the selenium to the platinum at the illuminated end, and, in those 
cases where this is not found to be the case, we have the current almost always in one 
direction with respect to the selenium, whichever end is illuminated. 
(2) A thermoelectric current does not usually attain its maximum strength imme- 
diately, nor does it cease altogether immediately after the withdrawal of the source of 
heat. In these experiments we have invariably found that, on interposing a screen in 
the path of the beam of light, the current immediately ceased, and on withdrawing the 
screen, the current was at once renewed in its original strength. The phenomena 
noticed are exactly similar to those observed when a tangent-galvanometer and a 
galvanic cell are joined up by a key in simple circuit. On closing the circuit, the 
needle swings at once up to, and then oscillates about its ultimate position of rest, and 
stays there while the current is passing. On breaking the circuit, the needle at once 
returns, oscillates on either side of, and then stops at zero. 
During these experiments we were frequently struck by the analogy of the two cases, 
and impressed with the idea that cutting off the light was, in point of fact, removing 
the electromotor of the current. 
