M. H. Helmholtz on the Theory of compound Colours. 523 



tersect each other at right angles. Every coloured band of the 

 one intersects in the common field of the spectra each band of 

 the other, and thus we at once obtain the total combinations 

 capable of being formed out of every two simple colours. 



As it is necessaiy to illuminate the slit uniformly through its 

 entire extent, direct sunlight cannot be well applied, and we 

 must content ourselves with the light of the firmament, or of a 

 white sm-face shone upon by the sun. These lights are in 

 general completely adequate to the purpose. 



The flint-glass prism which I made use of, permitted, when 

 du-ect sunlight and a narrow slit were applied, a great number 

 of the finer Fraunhofer's lines to be seen. In the spectrum of the 

 angular, and somewhat wider slit above described, the stronger 

 Imes, at least, were distinctly visible, particularly those which 

 Fraunhofer has distinguished by the letters A, B, D, E, b, F, G, H. 

 The presence of these hues assures us, in the first place, that in 

 the spectrum of each distinct limb of the slit the difi'erently 

 coloured rays could not overlap each other, and hence that we 

 had to deal with pure coloured rays ; and, secondly, they increase 

 greatly the facility of examining the mixed field, through which 

 they are distinctly seen to run. My telescope possessed a pair of 

 cross wires which cut each other at right angles, and these I set 

 parallel to the dark lines of the coincident spectra. The wires 

 thus mark, at the upper and lower rim of the illuminated field, 

 the two simple colours which are mixed together at their point 

 of intersection. 



It is necessary to be able to alter the relative intensity of the 

 mixed colours. This I accomplished by bringing the prism from 

 its vertical position into a more or less inclined one. It was so 

 attached to the forward cylindrical end of the telescope as to 

 permit of its being turned round the axis of the latter, and thus 

 might be brought into any required position relative to the hori- 

 zon. In order to explain how, by this means, the intensity of 

 the light of the spectrum is changed, let us fix our attention 

 upon a single slit. The intensity of the spectrum depends upon 

 the quantity of light wliich falls from the slit upon the prism 

 and telescope, and on the apparent magnitude of the spectrum 

 to the illumination of which this light is apjjlied. The quantity 

 of light received does not change when the prism is turned round 

 the axis of the telescope, but the illuminated surface of the spec- 

 tral image changes. The latter, as already remarked, possesses 

 the form of a parallelogram. Two of its sides are parallel to the 

 slit, and always of the same length as the slit ajipears in the 

 telescope; the two other sides stand per])endicular to the re- 

 fracting edge of the i»rism, and their length depends solely upon 

 the dispersive power of the latter. The spectrum therefore forms 



