REVERSIBLE AND IRREVERSIBLE SYSTEMS UNDER INFLUENCE OF LIGHT. 
341 
of the system in the dark being 0‘00079 volt per 1° C., we ought to have got 
8‘56 centims. deflection, if the system were also heated by light 0°‘1208 C. 
Now the same light passing a liquid column of a bath has evidently a very much 
smaller heating effect than when it passes air only; the rise of the temperature of 
the same thermopile would have been very much smaller if it could be immersed into 
water instead of being in the air; it would have been smaller if, instead of very thin 
plates, thicker plates were used as in our experiments, &c. The first problem of all 
thus became to determine in each case quantitatively and by correct methods whether 
the observed E.M.F.’s were not entirely a thermo E.M.F.; or, if not, which part in 
the total phenomena was due to one effect and which part to the other. 
We shall now illustrate the method used for this investigation : Let the system of 
two silver plates in a 0T normal AgN0 3 solution be investigated. First, the thermo 
E.M.F. of this combination (Ag wires in 0T normal AgN0 3 ) was determined (as will 
be described in a future publication in a section dealing with thermo-galvanic cells) 
and found to be, per 1° C., about O’OOOll volt. Next the rise of temperature in the 
quartz vessel for the same system (Ag pl 0'1 normal AgN0 3 solution) under the action 
of light, under the given experimental conditions, was determined. For this a 
thermopile with four junctions of very thin iron-nickel wires was made and covered 
with shellac; 1° C. difference of temperature was found to give 65 - 2 millims. 
deflection on Nalder Nl. The same thermopile was then placed in the quartz vessel 
containing the same two silver plates immersed in a 0'1 normal AgN0 3 solution (which 
at one time were connected • with 850 ohms to a circuit, and the other time left in 
open circuit); two junctions of the thermopile were placed in front of the plate 
exposed to light, the other two near the plate kept in the dark. The deflection of 
Nalder Nl, connected with the thermopile, on the same scale (when the arc of 
intensity 12‘7 centims. was removed 27 centims. from quartz vessel), was in open and 
closed circuit between 2 and 3 millims. The heating effect upon the exposed plate 
under these conditions of experiments was therefore between 0°‘03 and 0°‘045 C., and 
the thermo E.M.F. for Ag in AgN0 3 solution, under the action of the light, thus was 
between 4‘8x 10 -6 and 3'2 x 10 -u volt. 
The total deflection produced by the arc upon the silver plates, immersed in the 
0T normal AgN0 3 solution was next measured under the same experimental conditions 
(distance of the arc from quartz vessel, 26 centims., the same intensity of light, 
sensitiveness of Nalder Nl, &c.) (see Table I., June 6, 1903). The total deflection 
obtained was 55 millims. Since the resistance of the solution in the quartz vessel was 
13‘4 ohms, of the galvanometer 834‘4 ohms, the deflection caused by the thermo E.M.F. 
of 3‘2 x 10 -6 volt was 3'2 millims., i.e., about 5’8 per cent, of the total deflection 
observed (the total E.M.F. at 37 centims. distance from arc, this being 55 x 10 -6 volt). 
We thus got direct proof that under the action of light upon metallic plates two 
E.M.F.’s are produced : one is due to the heating of one of the plates by light and is 
a thermo E.M.F., the chemical potential of one of the two plates becoming, owing to 
