166 Prof. J. D. Forbes on the Classification of Colours. 



whole, then, in this experiment, half the light is still absorbed 

 and half of the remainder is yellow, the rest blue, constituting 

 a green colour. 



Similar reasoning will hold true of any number of separate 

 colours combined into one; and as perhaps no pigment reflects 

 only one pure colour of the spectrum, the mixtures will always 

 be more compound than they are assumed to be, and give 

 hues of always increasing impurity. 



The process of mixture cannot in any case be expected to 

 improve the power of reflecting the pure colours residing in 

 the constituent pigments. It is much more likely to deterio- 

 rate it, which will tend to give a tone of always less absolute 

 brightness to more complex colours. 



The process of mixing pigments may often affect their rela- 

 tive strength, especially if moisture be used. There may be 

 a mutual action, which will cause an undue preponderance of 

 one or other constituent. These difficulties have constantly 

 been felt by those who have endeavoured to compound colours 

 from their elements. 



We have assumed above that the reflective power of pig- 

 ments of different colours is the same. But this does not 

 appear to be the case. The ingenious experiments of Lam- 

 bert*, which though probably but rough approximations are 

 yet valuable as such, inform us that the whitest surface re- 

 flects Y%ths of the white incident rays, and king's yellow almost 

 as much of rays of its own colour : but the brightest red 

 (cinnabar) reflects but |rd of its own coloured light, and a 

 blue surface (mountain Berlin blue) but yth of the blue i*ays. 



On these principles we may expect that if the circumference 

 of a wheel be painted with stripes of red, yellow and blue, alter- 

 nating with one another, so tliat the extent occupied by these 

 colours shall be in certain determinate proportions, the mix- 

 ture shall appear white, or rather tieufral gj-ay, the wheel being 

 put in rapid rotation. We can estimate the illumination of the 

 surface compared to that of white paper in the following 

 manner. 



The proportional extent of the surfaces may be found by 

 direct experiment, or otherwise thus. By Newton's rule of 

 compounding colours (see p. 163), we may deduce that a white 

 compounded of red, yellow and blue, must consist of 38*6, 

 19*6, and 41 '8 rays out of 100 of these colours respectively; 

 for the centres of gravity of the red, yellow and blue sectors 

 make angles of 91° 54>'-5 between the red and yellow, 1 15° 26' 

 between the yellow and blue, and 152° 39'*5 between the blue 

 and red. In order that the centre of gravity of the whole shall 

 * Photometria, § 747; and Farbenpyr amide, § 5. 



