ACCIDENTAL IMAGES. 175 



luminous rays, that is to say, from the positive condition of 

 impression to a negative one, and then forced by the re- 

 action, it passes the point of repose, and recedes in an 

 opposite direction. These oscillations continue for a variable 

 time, growing feebler and feebler. The reaction of the 

 retina, and the negative phases of impression, give rise to a 

 new sensation independent of any external agent, by pro- 

 ducing what are termed accidental or consecutive images. 



We know that two colours are complementary to each 

 other, which when mingled together produce white; but the 

 accidental images have the peculiarity of presenting them- 

 selves in the colour complementary to that of the luminous 

 rays which have excited the retina; thus, if we look steadily 

 for a certain length of time in a very clear light at a wall 

 painted red, the accidental image is green, and if the wall is 

 orange, the image will be blue, &c. 



If, on going into a dimly lighted gallery, we fix our eyes 

 for a minute or two on a window which receives the diffused 

 light, and then shut them suddenly and cover them so as to 

 place them in complete darkness, the primitive impression 

 of the window, with the panes lighted and the sashes dark, 

 remains for a time, but very soon the consecutive image 

 appears with the frame luminous and the glass obscure. 

 This last image will appear sooner if a little light be admitted 

 through the closed lids; but in all experiments of this kind, 

 the eyeballs must be kept perfectly still under the veil with 

 which they are covered, for the slightest change in the direc- 

 tion of the optic axes will cause the images, whether primi- 

 tive or accidental, immediately to disappear. 



One of the most important facts in this portion of the 

 history of the eye, we owe to the observation of M. Plateau. 

 It is that the duration of the uniform intensity of the retinal 

 impression, up to the moment when it begins to decrease, is 

 short in proportion to the intensity, that is in proportion as 

 the light which produced it was brilliant and white ; so the 

 impression is less and less durable in its first intensity ac- 

 cording as it is produced by looking at a blue, red, yellow, or 

 white disk; if, on the contrary, we measure the impression 

 not only in its period of uniform intensity, but from its 



