and Electrical Conductivity in Selenium. 487 



the kind of the result obtained. The point ft represents the 

 " dark " current, and, although the curve is still straight over 

 the range of illuminations used, it does not pass through the 

 point ft. This may be contrasted with curve I, which 

 represents the normal case for the same cell — i. <?., when the 

 voltage is only applied momentarily to determine the con- 

 ductivity. Here the straight line does pass through the 

 point a, which represents the observed "dark" current. 

 Curve II indicates that for considerable intensities the 

 " extra " current is still proportional to 1% if one estimates 

 from a fictitious dark current determined by finding the 

 intercept of the straight part of the curve upon the current- 

 axis. It is clear, however, that the dotted part of the curve 

 requires investigation by means of a more sensitive arrange- 

 ment than the present one. 



Significance of the Results. 



The mechanism of the action of light on selenium is not 

 yet known with certainty. It is probably very complicated, 

 and has to account, for example, for the recent observations 

 by White* that the effect is more marked near the electrodes 

 of the selenium call than elsewhere. The theory propounded 

 by Fournier d' Albe t that the change of conductivity is due 

 to ionization in the selenium is plausible, and may, indeed, 

 ultimately form the basis of a complete theory. It is worth 

 while to consider in what respects the present results suggest 

 modifications of Fournier d'Albe's assumptions. He assumes 

 that the rate of production of ions is proportional to the 

 incident flux of energy — i. e. to I, — and that the rate of recom- 

 bination is proportional to the square of their number. The 

 steady state is reached when the rates of production and 

 recombination are equal. In this state, therefore, 



I^(C-Co) 2 



for the "extra" C — C will be proportional to the final 

 number of ions. It is from this proportionality that 

 Fournier d' Albe deduces the theoretical law C — C cc 1^ men- 

 tioned in the beginning of this paper. 



From this point of view the significance of the results 

 here recorded would be that (C — C ) 2 is proportional to 

 Is, and not to I itself. In other words, the rate of pro- 

 duction of ions is proportional, not to the incident energy, but 

 to its square root. Now, for monochromatic light the square 



* Phil. Mag. ser. 6, vol. xxvii. p. 370. 

 t Loc. cit. 



