550 



Dr. F. W. Edridge-Green. 



[Dec. 4, 



after-image formed on the white ground is rose colour for the green portion 

 and dark grey for the white portion. In this case, as with white paper, 

 not a trace of green is seen in the after-image of the grey. 



A simple experiment described by Waller* illustrates this relativity of 

 perception very well. If a strip of grey paper be placed upon a sheet of 

 white paper, and then a piece of green paper be placed on either side of the 

 middle third of the grey paper, and the whole covered with a piece of tissue 

 paper, no contrast colour, or very little, will be visible. If, now, the middle 

 third of the strip of grey paper be isolated by means of two pins placed 

 transversely, the middle third becomes strongly tinged with the contrast 

 colour, rose. On repeating this experiment, I find that, when the grey strip 

 is seen as grey, the after-image is also grey, but when it is seen as rose, the 

 after-image is green, and the rose after-image of the green appears less 

 saturated. Also, when the contrast colour is developed, the objective green 

 appears less saturated than when the grey strip appears grey. 



A definite amount of saturation is necessary before a colour can be 

 recognised. This colour becomes much more marked on contrast. A tinted 

 paper which appears pure white without comparison may, when laid on a 

 pure white surface, appear very definitely coloured. If a square of cream- 

 coloured paper be placed on a white ground, it will appear of a decided pale 

 yellow ; the colour of the white ground will, however, not be altered. If the 

 after-image of this paper on a white surface be examined, it will appear as a 

 pale blue square on a white ground, but no adjacent yellow is to be seen. 

 The estimation of colour is always relative ; for instance, if a pale yellow 

 diamond be given to a man, who has to classify diamonds, as a standard 

 white, he will classify the pure white diamonds as blue, and not sufficiently 

 estimate the amount of yellow in those diamonds which are yellow. 



Our power of discrimination of colours is much more limited than is 

 usually supposed. I have shownf that most persons can only differentiate 

 about 18 separate regions in a pure spectrum, and that if one of these 

 regions be examined with a double-image prism, so that the red side of one 

 image be adjacent to the violet side of the other, no difference will be 

 detected. Here we should expect that, if any colour induction were 

 produced, an immediate difference would be observable between lights which 

 are objectively so different. 



In examining the two images, the greatest care should be taken to have 

 them both of the same apparent luminosity. It will be noticed that, when 

 the images are of different luminosity, the hue is also different. Both 



* ' Journ. Phys.,' 1891, p. 44. 



t ' Koy. Soe. Proc.,' B, 1911, p. 116. 



