214 RECENT PROGRESS OF THE THEORY OF VISION. 



and which must therefore be placed at the greatest distance from 

 the central white, will not arrange themselves in the form of a 

 circle. The circumference of the diagram presents three pro- 

 jections corresponding to the red, the green, and the violet, so 

 that the colour circle is more properly a triangle, with the 

 corners rounded off, as seen in Fig. 34. The continuous line 

 represents the curve of the colours of the spectrum, and the 

 small circle in the middle the white. At the corners are the 

 three colours I have mentioned, 1 and the sides of the triangle 

 show the transitions from red through yellow into green, from 

 green through bluish-green and ultramarine to violet, and from 

 violet through purple to scarlet. 



Newton used the diagram of the colours of the spectrum (in 

 a somewhat different form from that just given) only as a con- 

 venient way of representing the facts to the eye.; but recently 

 Maxwell has succeeded in demonstrating the strict and even 

 quantitative accuracy of the principles involved in the construc- 

 tion of this diagram. His method is to produce combinations of 

 colours on swiftly rotating discs, painted of various tints in 

 sectors. When such a disc is turned rapidly roiand, so that the 

 eye can no longer follow the separate hues, they melt into a uni- 

 form combination-colour, and the quantity of light which belongs 

 to each can be directly measured by the breadth of the sector of 

 the circle it occupies. Now the combination-colours which are 

 produced in this manner are exactly those which would result if 

 the same qualities of coloured light ilhiminated the same surface 

 continuously, as can be experimentally proved. Thus have the 

 relations of size and number been introduced into the apparently 

 inaccessible region of colours, and their differences in quality 

 have been reduced to relations of quantity. 



All differences between colours may be reduced to three, which 

 may be described as difference of tone, difference of fulness, or, as 

 it is technically called, ' saturation,' and difference of brightness. 

 The differences of tone are those which exist between the several 



1 The author has restored violet as a primitive colour in accordance with 

 the experiments of J. J. Miiller, having in the first edition adopted the opinion 

 of Maxwell that it is blue. 



