376 O. U. VONWILLER. 



in the first case, and at a time when the change had been 

 completely produced in the first case it would be still 

 going on in the second. If in the case of fig. 7 the illumin- 

 ation were increased so that A M' is the effect at the 

 surface, the initial rate of rise and the total effect would 

 both be increased, but the final value would not be 

 attained any sooner than before, since corresponding to 

 any layer E with a change EF with the weaker light 

 there is a layer E,' with a change E' F' equal to this, 

 with the stronger illumination, both these changes taking 

 the same length of time to occur, and unless the intensity 

 jS so great that an appreciable effect is produced at the 

 lower surface, the time occupied in getting to a state 

 where no measurable change occurs will be no shorter with 

 strong illumination than with weak. 



Now in the observations given above on the rates of rise 

 of conductivity upon exposure to red and green light, it is 

 seen that under the action of red light the selenium behaves 

 as though the effect varied with the depth below the 

 surface as in fig. 6, and under the action of green light 

 as though it varied as in fig. 7. The results obtained 

 when the two surfaces were illuminated separately and 

 together confirm the view that the effect with green light 

 falls off more rapidly as the depth penetrated increases, 

 for in a case such as that of fig. 7, the increase in con- 

 ductivity obtained when both sides are illuminated would 

 be almost equal to the sum of the increases obtained 

 when the sides are separately illuminated, as there is 

 practically no overlapping of the regions affected, whereas 

 in the case of fig. 6, where considerable overlapping occurs, 

 the illumination of both sides produces an effect consider- 

 ably less than the sum of the changes produced when the 

 sides are separately illuminated. 



That the rapidity with which the effect is produced 

 should increase with the magnitude of the effect might be 



