REVERSIBLE AND IRREVERSIBLE SYSTEMS UNDER INFLUENCE OF LIGHT. 383 
alcohol are inconstant cells, showing polarisation, i.e. , the E.M.F. does not give the 
maximum work performed under the action of light. 
(2) If only strips of the front plate are illuminated (Becquerel), local action 
takes place between the illuminated and noil-illuminated parts of the front plate. 
(3) In their publications no data are given to show how far the E.M.F.’s in the 
dark were brought in their experiments not only to constant, but also to sufficiently 
small values; the thermo E.M.F.’s in the cell, &c., do not seem to have been 
sufficiently guarded against. They did not possess a constant source of light, nor 
measure and check the intensity of their source of light, or take into account the 
induction and deduction periods. We have no evidence as to how far they reached 
the constant maximum deflections in each case—this is not always easily obtainable— 
nor how far the E.M.F. in the dark was influenced during their experiments by 
previous exposure of the system to light, &c., &c. 
§ 8. Experimental Proof of tlie Pliysico-Mathematical Theory of “ Constant Rever¬ 
sible Cells Created by Light ” (given by the Author in ‘ Roy. Soc. Proc.,’ 
November, 1904). 
In the above paper the same was dealt with by the author in detail. The general 
equation found was 
2E = E 1 + E 8 -E 3 = 0-860 T flog, ^ - — log A • 10" 4 volt 
\ id u + v pdj 
for systems reversible in respect of the cation, and 
SE = E.+E.-E, = 0-860 T(-log,.U + AT l 0 g, A). Hr* volt 
\ Pd u + v pdj 
for systems reversible in respect of the anion. 
The present research confirms the conclusions drawn there from the above 
equations :— 
(1) Table IV., Plates 30 of July 31, 1903, August 1, 1903, prove for the cells 
reversible in respect of the anion; experiments of June 19, 1903, §6, Plates 7, 8, 9, 
Table I., prove the cells reversible in respect of the cation, that E.M.F. is independent 
of concentration. 
(2) Experiments, § 7, XI., Table X., Plates N22 and N22', of the 25th to the 27th 
of July, 1904, and of February 27, 1905, as well as § 7, XI., Table X., show that 
both the value 
0'860T ^ v log c 1— • 10~ 4 volt, or 0 - 860T - log e — • 10~ 4 volt, 
u + v & pd u + v & pd 
and the electrical potential between the illuminated and non-illuminated parts of the 
solution can be neglected, so that instead of (I) we can use 
P/ p,7 
SE ~ 0‘860T log ? ' 10 4 volt, and SE = 0‘860T log e • 10 4 volt. 
