io8 2 VISION. 



one of green would be represented in Fig. 391 by o when og = 2 or; if B 

 = quantity of red and G = quantity of green, Bxro = Gx og. 



If this resultant colour is mixed with a third (B), the resultant of the 

 mixture of the three will be on the line 6b, at a point determined by 

 the quantities of each. Maxwell found that by taking these components 

 in different proportions, one may determine the colour corresponding to 

 every point of the triangle of which the three colours form the angles. 

 If three suitable spectral colours are chosen, it will be possible to repre- 

 sent all other colours by means of such a triangle. Complementary 

 colours in such a diagram will be placed at a distance from one another, 

 and the resultant of their mixture to form white will be represented by 

 a definite position within the triangle ; the lines joining any two comple- 

 mentaries must pass through this definite point ; the lines joining the 

 point for white with any point of one side of the triangle will represent 

 all degrees of saturation of the colour of that point from spectral purity 

 to white. In Fig. 391, if and B are complementaries and mixed together 

 in approximately equal proportions to match white, w will represent the 

 point for white, and points between o and w will represent orange in its 

 various grades of saturation. The fact that it is possible to find three 

 colours, and by means of their geometrical representation to deduce 

 the results of their mixture in any proportions, is the basis of one of the 

 chief theories of colour vision, and will be returned to in that connection 

 and also in connection with colour blindness. 



Laws of colour mixture.— The validity of this procedure depends on 

 certain assumptions which were formulated by Grassmann, 1 as follows : — 



1. Every impression of light may be imitated by mixing a homogene- 

 ous colour of a certain intensity with colourless light of a certain intensity. 



2. If one of the two mingling lights be continuously altered (whilst 

 the other remains unchanged), the impression of the mixed light also 

 continuously changes. 



3. Two colours, each of which has a constant tint and a constant 

 intensity of the intermixed white, also give constant mixed colours, no 

 matter of what homogeneous colours they may be composed. 



A fourth assumption which Grassmann regarded as less well 

 established was — 



4. The total intensity of a mixture is the sum of the intensities of 

 the lights mixed. 



Grassmann's propositions have been investigated experimentally. 

 The third and fourth are those about which there is most doubt. The 

 third proposition implies that a mixture good at one intensity should 

 hold good for all intensities, i.e. multiplication of both sides of an 

 equation by the same factor should not affect the validity of the 

 equation. According to several observers, this proposition is not true, 

 and the alteration of an equation with altered intensity is usually 

 described as a departure from Newton's law or rule. 



The most marked examples of this deviation occur in the colour- 

 blind, and will be considered later (p. 1090). According to Konig and 

 his followers, it also occurs in normal vision. Thus Tonn 2 found that the 

 complementary to a red of 670-8 X varied with the intensity, shifting 

 from 511-8 X to 547*3 X when the intensity was reduced to a fortieth. 

 The complementary to 586-5 X, on the other hand, showed no change 



1 London, Edinb. and Dublin Phil. Mag., London, 1854, Ser. 4, vol. vii. p. 254. 



2 Ztschr.f. Psychol, u. Physiol, d. Sinnesorg., Hamburg u. Leipzig, 1894, Bd. vii. S. 2/9. 



