236 Mr. Shelford Bidwell on the 



apparent change of its resistance under the influence of light. 

 According to my hypothesis, the conductivity of Se depended 

 mainly upon the selenides which it contained. I supposed 

 the conduction to be of a truly electrolytic character, the 

 current decomposing the selenides, and depositing upon the 

 anode amorphous Se and upon the kathode the various 

 metals. Amorphous Se is a non-conductor : unless, therefore, 

 the deposited Se combined more or less freely with the metal 

 of the anode, forming upon it a layer of conducting selenide, 

 the resistance of a Se cell would, under the influence of a 

 current, soon rise to infinity. It has been shown that Se 

 does slowly unite with at all events some metals when merely 

 brought into contact with them, and I ventured to suggest 

 that light played its part in reducing the resistance by 

 facilitating this union. The action would occur not only at 

 the anode, but throughout the body of the Se, the recombina- 

 tion of the separated molecules of Se and the metal (forming 

 the ci Grotthuss chain ") being similarly assisted. 



This hypothesis would, as I showed, explain many observed 

 facts which could not otherwise be readily accounted for, and 

 I adduced in support of it several arguments which need not 

 here be repeated. But the hypothesis postulated two im- 

 portant assumptions which I was not able at the time (nor 

 when I returned to the subject five years later) to justify 

 by experimental evidence. It assumed that solid metallic 

 selenides conduct electrolytically, and that the combination of 

 selenium with a metal is accelerated by the action of light. 



My recent experiments, some of which are discussed in the 

 present paper, were undertaken primarily with the object of 

 testing these two assumptions. In the course of the investi- 

 gation, however, many other interesting points presented 

 themselves for consideration, and some of the results arrived 

 at are also here recorded. The experiments are described in 

 the order which appeared to be most convenient ; it is of 

 course not that in which they were actually performed. 



§ 4. The Effect of Annealing upon the Specific Resistance of 



Crystalline Selenium. 



The resistance of Se which has been crystallized and 

 annealed out of contact with any metal appears to be always 

 much greater than when metallic electrodes have been fused 

 into it in the usual fashion. It also depends to a large extent, 

 and in a manner which is not yet fully explained, upon the 

 treatment which the Se has received. 



Exp. 1. — A number of disks were prepared by casting Se 



