ayid on the Elementary Colours of the Solar Spectrum. 271 



upon the blue of the same image. A prism of water, prepared 

 in like manner, of which the refracting angle was 79°, gave 

 analogous results. In one, as in the other case, the red ex- 

 tremity of the left spectrum was on a level with the green of 

 the central image, when the observations were made at the 

 distance of a metre from the prism. 



"Let us imagine the central unpainted part of our prism 

 to be divided into a series of longitudinal elements, the width 

 of each being equal to that of the lateral bands. It is clear 

 that each of these elements will produce a refracted image 

 similar to the two pale spectra arising from the lateral bands ; 

 and that the two last images of the series will be as it were 

 continuations of those spectra. Therefore, the red and the 

 violet which we see by the side of the yellow and blue of the 

 central coloured spectrum, equally exist in that central spec- 

 trum, and enter into the composition of those tints. This 

 argument is unanswerable ; it establishes that, instead of being 

 absolutely pure, or even nearly so, the gradations of the 

 image formed by the central compartment contain different 

 colours.*" 



I had already proved, in 1843, that the colours obtained by 

 ordinary prisms at a distance less than two metres, are com- 

 posed of a mixture of colours belonging to the spectra of the 

 different elementary bands of the prism ; and that the red, 

 the violet, and consequently all the other prismatic colours of 

 the two extreme elements, are so much the nearer to the 

 centre as the observation is made nearer to the prism. Now, 

 in the experiment of Sir D. Brewster the prism is very close 

 to the eye, — the spectrum obtained in this experiment must 

 necessarily be formed of very impure tints ; and the colour 

 which ap))ears in a given zone which has lost by absorption 

 the dominant tint, does not belong to a ray of the same re- 

 frangibilily as the tint absorbed, but rather to colours of the 

 elementary spectra of the superior or inferior parts of the 

 prism. 



To prove this directly, I have repeated the fundamental 

 experiment of Sir D. Brewster. It consists, as is well known, 

 in interposing between the eye and the spectrum, produced 

 by the refracted image of a luminous object seen through the 

 prism, a slij) of glass deeply coloured blue by the oxide of 

 cobalt. The spectrum was formed fiom the light of a circular 

 aperture, ten iiiillinietres in diameter, made in a metallic plate 

 j)laced in the shutter of a dark chamber. T!ie prism was of 

 flint glass, ecjuiiateral, twenty-five millimetres wide, and suf- 



• Mtuco di Scicnzc, Lcltcre edArii, vol. i. fasc. 1, Napoli cli 1843. Bib- 

 Vwthcquc UniverscUe, 1844, vol. xlix, p. 141. 



