and on the Elemcntanj Colours of the Solar Spectrum. '21'i 



prism, and at another through a limited space of the prism 

 narrower than the pupil of the eye, it may happen that the 

 greater extei^t of the tints transmitted by the blue glass in the 

 first observation may arise from an indistinct vision, and not 

 from a real overlapping of the colours from the superior and 

 inferior parts of the prism. This conjecture seems the more 

 plausible, since all the rays refracted by the prismatic elements 

 are hot perceived by the observer, but only those that pass 

 through the aperture of the pupil. 



To determine whether this was in reality the cause of the 

 phaenomenon, I placed four small bands of tin around the 

 circular hole in the shutter, and arranged them so that they 

 formed by crossing one another a perfectly square aperture, 

 the sides of which were horizontal and vertical. Then placing 

 before the prism a slip of glass of a deeper blue than the pre- 

 ceding, I saw on looking successively through the central 

 part of the prism and the part from which the paint was par- 

 tially removed, that the two spectra modified by the interpo- 

 sition of the coloured medium were composed of a red rect- 

 angle, almost square, followed by a broad dark zone, and then 

 by a very brilliant yellow rectangle, of which the longer sides 

 were directed vertically and parallel to the length of the spec- 

 trum. There came then a deep indistinct colour, then the 

 blue, to the modifications of which it is unnecessary here to 

 attend, but only to the changes of colour and darkness of the 

 space already mentioned which precedes the yellow. 



On observing attentively the rectangular form of the space 

 occupied by the yellow rays in each of these spectra, it will be 

 distinctly perceived to be less elongated in the elementary 

 spectrum tiian in the compound spect7-um. Now distinct vision 

 may diminish the magnitude of the image formed upon the 

 retina, and render the contours more decided and sharp, but 

 it cannot vary the relations of its dimensions. The shorter 

 length of the yellow rectangle in the elementary spectrum must 

 then arise from an effect different from that which is produced 

 on vision by the smallness of the aperture through which the 

 prismatic image is observed. We cannot therefore suppose, 

 the vertical sides of the rectangle being a little sharper and 

 more vivid than the horizontal ones, that this difference of 

 illumination can give origin to the phaenomenon in question ; 

 for any alteration arising in that way would be in the opposite 

 direction to that actually observed. For, in reality, the most 

 luminous images being those which undergo the greater re- 

 duction of size in passing from confused to distinct vision, the 

 yellow rectangle of the spectrum observed in the elementary 

 prism ought to be shorter in the horizontal than in the vertical 

 yV///. Ma^. S. 3. Vol. 32. No, 215. April 1848. T 



