386 VISION. LECT. XX. 



These images become weaker after having undergone this 

 series of alternations. 



Accidental colours form, with each other, combinationslike 

 real ones. The following fact puts this curious observation 

 beyond a doubt: place upon a black ground two squares of 

 paper, one violet, for example, the other orange, both having 

 black points in their centres. Carry the eyes alternately 

 from one point to the other, and after each passage shut 

 the eyes, and you will fancy you see three squares : one 

 yellow, which is complementary to violet; one blue, com- 

 plementary to the orange; and a third, between the two 

 others, of a green colour, which is the precise shade formed 

 by the yellow and blue. In this experiment the primitive 

 impressions upon the retina, are only the superposition of 

 the two partial impressions which they would manifest if 

 we had observed only one black point ; but as the direc- 

 tion of the optic axes differs according as we look at one or 

 the other of these points, it follows that the points of the 

 retina which receive these two partial impressions are not 

 symmetrical. From the juxtaposition of the two squares it 

 results, that the accidental image of the orange colour for 

 the first partial impression, is superposed on that of the 

 violet for the second. The intermediate square, which we 

 perceive when closing the eyes, is owing to this superposi- 

 tion; consequently, we must conclude that the accidental 

 yellow and blue form the green, as real blue and yellow 

 would do. We should arrive at the same conclusion ? 

 whatever were the colours of the two squares. Yet we 

 observe a difference when they are tinted with complemen- 

 tary colours ; in this case the intermediate square, arising 

 from the superposition of the accidental images, is black, 

 and not white, as takes place in the mixture of two real 

 colours. 



Accidental colours combine with real ones exactly as the 

 latter would do with each other. To be convinced of this 



