NATURAL PHILOSOPHY. 183 



a yellow with green, &c. The compound tints obtained by a combi- 

 nation of two prismatic colors, have for the most part their analogues 

 . in the solar spectrum, except the white tint produced by yellow and 

 blue and the purple tints produced by red and violet, but there is 

 almost always an appreciable difference from the analogous tints. 

 Thus the red and yellow give an orange not to be distinguished from 

 the orange of the spectrum, and the violet and blue reproduce almost 

 completely, the pure indigo ; but the yellowish green and bluish green 

 of the spectrum, in combination, give a green much more white and 

 much more dull than the green of the spectrum ; the appearance of 

 this latter color it is almost impossible to imitate by a mixture of two 

 simple colors. The yellow is very well reproduced by means of orange 

 and yellowish green, and the blue by bluish green and indigo, but the 

 violet can only be very imperfectly imitated by means of a combina- 

 tion of indigo with a small quantity of red ; the red itself cannot be re- 

 produced in any manner, the combination of orange and violet giving 

 a carmine tint very different from the red of the spectrum. 



It results that to reproduce simply the tints of the spectrum, in a 

 manner at all satisfactory, requires at least five simple colors, the red, 

 the yellow, the green, the blue, and the violet. If we confine our- 

 selves to three simple colors, but very imperfect imitations will be pro- 

 duced ; it is necessary always to use the red, green and violet; a choice 

 of any other elementary colors does not permit us to imitate but par- 

 tially, the tints of the spectrum, and the combination of red, yellow 

 and blue, which is indicated by most authors, is the worst of all. 



M. Helmholtz likewise made certain experiments upon the results 

 of the mixture of three simple colors. To effect this, the slit in form 

 of a V was replaced by another having nearly the form of N inclined, 

 and studied as in the preceding manner by the superposition of the 

 three spectrums. He investigated, chiefly, the combination of three 

 colors which would give the white, and found a great number. There 

 results from these experiments an important fact, the effect of the 

 combination of a simple color with two others, produces a different 

 effect from the combination of the first simple color with the color of 

 the spectrum analogous to the tint resulting from the mixture of the 

 two colors ; for example, the red and the bluish green of the spectrum 

 give yellow ; the red, and the bluish green produced by the combina- 

 tions of green and indigo, give white. 



The whole of the phenomena conduce evidently to the rejection of 

 the hypothesis cited at the commencement of this memoir ; since it is 

 impossible to reproduce the different tints of the spectrum with any 

 three others, nor does it affect less the three fundamental colors if the 

 same are considered as three distinct physiological impressions. 



At the last meeting of the British Association, at Hull, M. Helm- 

 holtz stated that he had repeated the above experiments, following an- 

 other method, similar to that lately described by M. Foucault^ for 

 obtaining larger fields equally dyed with the mixture of two homo- 

 geneous colors, and has found that there are more pairs of comple- 

 mentary colors in the spectrum. These colors are situated at both 



17 



