366 O. U. VONWILLER. 



used with cell 12, and of 100 volts with cell 22. In obser- 

 vations 1 to 4 with cell 22 the lamp used was an 8 candle 

 power incandescent carbon lamp, in observation 5 a small 

 incandescent carbon lamp of about 1/20 candle power, and 

 in observations 6 to 9 a 23 candle power Tantalum lamp 

 was used. In the case of observation 6 the light was 

 intensified by means of a large concave mirror held at the 

 back of the lamp. 



While the absolute conductivity, in the dark, of cell 22 

 is only about 1/500 of that of cell 12, its sensitiveness is 

 very much greater, an 8 c.p. lamp at a distance of 100 

 centimetres with the red screen producing an increase of 

 conductivity of more than 1000°/° when both sides are 

 illuminated, while with cell 12 the increase under similar 

 conditions is only about 40°/°. 



With both cells the ratio , n \ * is much less than 



unity when red light is used. In the case of cell 22 the 

 ratio does not differ much for the first four intensities used, 

 while with cell 12 the ratio shows an appreciable rise as 

 the intensity falls, a similar result being obtained with 

 cell 22 with green light. The effect then seems not to 

 be limited to an extremely thin surface layer but to 

 penetrate more than half-way through the selenium so that 

 the regions affected overlap when both sides are illuminated. 



In fig. 2 are drawn curves for Cell 22, showing the rela- 

 tion between the intensities of illumination (abscissas) and 

 the increases in conductivity, (Di), (Da), and (D 1 + D 2 ), 

 respectively (ordinates), when using red light, the unit 

 intensity being that due to a source of one candle power 

 at a distance of 1 metre. It will be seen that the effect 

 produced when both sides are illuminated with any intensity 

 I exceeds that produced when one side only is illuminated 

 with an intensity 2 1. If there were no decrease in the 



