244 On the Irjlucncc oj successive Impulses of Light. 



produced by the action of light falling on the retina at a di- 

 stance from the axis of vision. When we look, for example, 

 indirectly, or rather obliquely, at a candle for some time, the 

 image of the candle itself becomes bluish purple surrounded 

 with a nebulosity ofycllo~o light, the accidental colour being 

 the invariable companion of the primitive one. 



In order to explain why no colours appear during a very 

 rapid rotation of the disc, and why the primitive and the ac- 

 cidental colour succeed each other in the pattern, let us call 

 T the time in which the disc revolves, 

 n the number of apertures in its margin, 

 D the duration of the impression of direct light, and 

 d the duration of its complementary colour. 



T 



It is obvious that — will be the time which elapses between 



11 ' 



each consecutive impulse of light on the retina, or the time 

 during which the eye has the opake part of the disc opposite 



T 



the pupil. When — is very small, or the velocity very 



T 



great, or when — is very much less than D, (D is = eight 



thirds, or nearly one eighth of a second,) the ground will be 

 uniformly luminous, because the direct impression of the one 

 aperture has not begun to fade away before the succeeding 



T 



aperture makes a new impression. When — , however, is 



nearly equal to D, the impression of the direct light is nearly 

 gone, and hence arises the flickering or wavering appearance 

 of the luminous ground, which becomes a maximum when 



T 



— = D; for when this takes place the direct impression of 

 n 



the one aperture is just gone before the other aperture renews 



T 



it. When — is greater than D, the accidental colour of the 



u ° 



direct impression begins to show itself; and when — = D + - , 



the accidental colour will be about its brightest, and will be 

 seen to succeed the direct impression, the latter being now 



T 



bluish purple, and the former Icmon-yelloio. When — = D 



+ d, the opake space between the apertures will begin to be 

 visible, and the phenomena will disappear. 



As the reticulated pattern is marked out by different co- 

 lours, and even by the same colour in different states of in- 





