218 Prof. Minchin's Experiments in Photoelectricity. 



and immersed in a cell with an aqueous solution of pyrogallic 

 acid. On exposure to light in this cell, the plate was about 

 half as sensitive, and of the same sign (positive) ns in the 

 methyl alcohol ; and on replacing it in the alcohol cell, its 

 indication was less than it had been originally. Thus, there 

 was merely a falling off, but not an absolute destruction, of 

 the E.M.F. caused by the pyrogallic acid, the sensitive surface 

 having assumed a yellowish appearance. 



The conjecture of M. Becquerel, that the sign of the 

 E.M.F. developed by light depends on the thickness of the 

 sensitive layer on the plate, has been already mentioned ; but 

 it may be allowable to put forward another conjecture — that 

 when there are two, three, or any number of salts of a metal, 

 any one of which may be on the plate, the result depends on 

 the particular salt, and that the series may he alternately 

 positive and negative on exposure to light according to the 

 amount of oxygen (suppose) that they contain. A reducing 

 agent may convert one of these into an adjacent one, and so 

 alter the sign of the E.M.F. The question is not one with 

 which I am competent to deal ; it is for the chemist. And 

 it might also be worth while to investigate whether there is 

 any connexion between the sign and magnitude of the 

 E.M.F. generated by the action of light on a metallic salt 

 and the effect produced on the salt by electrolysis. 



Production of the Sensitive Surface. — After trying a very 

 great number of processes for producing a sensitive tin 

 surface which is positive on ex- 

 posure to light, the following is 

 that which I have found to be 

 most efficacious. 



Fis. 2. 



Take a strip of pure tin foil, say 

 1 inch long and |- inch broad, and 

 about as thick as a stiff sheet of 

 note-paper (the very thin foil is 

 not suitable ; but the exact thick- 

 ness is not of much consequence) : 

 clean it first in sodic hydrate, and 

 then in dilute hydrochloric acid, 

 or in hydrofluoric acid, until the 

 surface presents a bright crystal- 

 line appearance. Having thoroughly washed it in distilled 

 water until every trace of acid disappears from the water, 

 place it horizontally, at p, on a bent plate of glass of the 

 shape gg in fig. 2, this plate of glass being held by a 

 clamp C supported by an ordinary retort-stand, which is not 

 represented in the figure ; under the tin plate p and the 



