250 



THE ROYAL SOCIETY OF CANADA 



The effect of varying the position of the electrodes on face A 

 was investigated in a field of 4,000 volts per cm. It was found that 

 in some positions of the electrodes the dark current in one direction 

 was over 30 times that in the opposite direction, while in other posi- 

 tions there was no appreciable difference between the dark currents. 

 In the latter case, however, the current obtained on illuminating the 

 diamond was much larger in one direction than in the opposite, 

 showing that the unidirectional property still existed, although the 

 difference between the dark currents was not measurable. An illus- 

 tration of the results is given in Fig. 11, which shows two positions 

 of the electrodes on face A. In position "a," the dark current caused 

 a deflection of 1.7 cms. in both directions, but on illumination the 

 photo-electric current was 7.8 cms. in one direction and 2.5 cms. in 

 the opposite direction. When the electrodes were moved a small 

 distance to the position shown in "b," the dark current was 65 cms. 



(^} 



(t>) 



Fig. 11 



in one direction, and only 3.5 cms. in the opposite direction. Here 

 we have great variation in the dark current when the electrodes were 

 moved a little way in the same straight line. It is possible that 

 motion in this line may cause great alteration in the relative positions 

 of the electrodes and the crystallographic axes, which may account 

 for the variation in the magnitudes of the dark currents. In any 

 case, it is evident that the unidirectional property for various regions 

 of the crystal varied greatly in degree, but was always present. 



Polarisation. 



In view of the slow growth of the dark current, it was expected 

 that the crystal would become polarised during the passage of the 

 current, with consequent development of counter electromotive force 

 in the crystal. An attempt was made to show this polarisation by 

 subjecting the diamond to a high potential, suddenly cutting out the 



