26 



Dr. H. E. Ives on 



the behaviour of the flicker photometer shows. This appa- 

 rent mutual action does not, however, appear to occur 

 between the pho topic and the after-images. The latter 

 behave as though due to an entirely separate mechanism. 



(e) After-images. 



The most characteristic feature of the positive after-image, 

 if it actually is not a recurrent impression, but the slowly 

 transmitted impression through the rods, as the present 

 theory would indicate, would be that the interval between 

 it and the photopic image would be variable with the 

 intensity. With decreasing intensity the after-image should 

 move in towards the blue image, and at very low intensities 

 should merge into it. 



Another characteristic of this after-image would be that, 

 in accordance with the w r ell-esta,blished wave-length sensi- 

 bility of the eye for scotopic vision, the red end of the 

 spectrum is practically inoperative — that is, there should be 

 no after-image of a red stimulus. 



These phenomena are all exhibited by providing the slots 

 with red, green, and blue glasses, in the manner indicated 

 in fig. 6 a. On rotation the three colours are shifted out of 

 line, in accordance with the subject-matter of the previous 

 sections, and at some distance behind the green and blue 



Appearance of red, green, and blue apertures lying in straight line : 

 a, stationary; b, in motion. 



images — but not behind the red — is the colourless after-image, 

 the after-images of the two colours appearing in the same 

 straight line, as they should from the theory. The appear- 

 ance at this stage is as shown in fig. 6 b. On decreasing the 



