NATURAL PHILOSOPHY. 181 



spectrurns to that which they presented when the prism was vertical, 

 it is clear the intensity could be determined, being in exactly an in- 

 verse ratio. Nevertheless, this arrangement was not applicable when 

 it was. desired to diminish very greatly the intensity of the light of one 

 spectrum, for upon inclining the prism sufficiently, the colors ceased 

 to be pure and distinct. In this case, leaving the prism in its vertical 

 position, one of the branches of the slit was covered with screens more 

 or less transparent, such as paper oiled and not oiled. 



The results of the experiments upon the combination of two colors 

 may be summed up as follows : 



1st. The red gives with orange an orange-red ; with yellow, an 

 orange, the two compound tints not being distinguished sensibly from 

 the orange-red and the orange of the spectrum ; with the green the 

 red gives a yellow more pale than the yellow of the spectrum, which 

 passes into an orange, or red, or greenish yellow or a green, according 

 to the increase of the proportion of one or the other of the composing 

 colors ; with the bluish green there is produced a flesh-colored tint ; 

 with azure blue, a rose-red tint passing into a violet or crimson-red, 

 according to the predominance of blue or red. 



2d. The orange with yellow gives yellowish orange ; with the green, 

 a yellow very pale ; with the blue, a flesh-color ; and with the indigo 

 and violet, a carmine red. 



3d. The yellow gives a greenish yellow with the green, and a green- 

 ish white with azure blue ; with indigo blue the yellow gives pure white ; 

 with the violet, the white assumes a flesh-tint, passing to whitish violet 

 or yellowish white, according as we increase the proportion of violet 

 or yellow. 



4th. The green gives with blue a bluish green ; with indigo a lighter 

 blue, more dull and mixed with white than the blue of the spectrum ; 

 with the violet, the green gives likewise a light blue. 



5th. The blue mixed with indigo gives intermediate shades, mixed 

 with violet it gives a dark blue, but not so deep as the indigo of the 

 spectrum. 



6th. The indigo with the violet gives intermediate shades. 



The most remarkable fact contained in the above observations is 

 evidently the production of white by the combination of yellow and 

 blue. We are so accustomed to consider green as the only color re- 

 sulting from blue and yellow, that M. Helniholtz was not willing to 

 trust to his own personal judgment, and he exhibited the phenomenon 

 to a great number of persons. To obtain pure white it is necessary to 

 take a yellow which has no orange or greenish tint, that is, the yellow 

 between Fraunholer's lines D and E, about three times more distant 

 from E than from D, the color being very near that of chromate of 

 lead ; as for the blue, we may take indifferently the whole of that part 

 of the spectrum ordinarily termed indigo, extending from the middle 

 of the interval between the rays F and G, as far as the ray G, the tint 

 of this color is rather that of sombre ultramarine than of natural 

 iiidigo. 



Again, this observation appears to controvert the daily experience 



