New Analysis of Solar Light. 415 



however, the solar spectrum be enfeebled by reflexion from un- 

 coated glass plates, or by a pair of Nicbol's prisms crossed nearly 

 at right angles, the yellow recedes and appears as in the daylight 

 spectrum. If, besides this, the colours belonging to various groups 

 of Fraunhofer's lines be isolated and determined in a spectrum 

 of moderately strong sunlight and in one of daylight, they will 

 be found quite alike. 



Of the facts which Brewster adduces in support of his theory 

 one remains over, with regard to which I do not know whether 

 I can say that I have succeeded in repeating it, and some others 

 which I was unable to repeat, not having the proper absorbing 

 media in my possession. The first is obtained with Peruvian 

 balsam, sulphur-- balsam, pitch and mica. The red of the spec- 

 trum seen through these media is said to appear orange. With 

 a moderate intensity of light I could observe nothing of the kind, 

 no matter what might be the degree of thickness of the pitch or 

 the balsam ; the red retained its colour quite unchanged. Only 

 ■nith light of a greater intensity, and when a brown luminosity 

 which surrounded the place gave evidence that a considerable 

 quantity of light was dispersed, did I see the red somewhat 

 orange. This, however, in the case under consideration, is to 

 be referred to the admixture of the dispersed brown of the light 

 compounded of red, yellow, and a little green, and to the ten- 

 dency of red to appear yellow when the light is intense. Perhaps 

 Brewster also made use of such a bright spectrum. When the 

 red is isolated according to my method, its colour remains totally 

 unchanged. 



Various experiments have been made by Brewster with trans- 

 parent wafers formed of gelatine. I could not obtain such here ; 

 and as their colours alone, and not the colouring matters, are 

 mentioned, I was unable to make them myself. They seem, 

 however, to me to be not free from objection, at least when placed 

 between the piism and the eye, inasmuch as the best glutinous 

 plates when formed of the purest isinglass between jjlates of 

 glass, do not belong to the class of transparent bodies. Even 

 when we are able to see pretty clearly through a single one, 

 several placed one over the other make the image cloudy — a proof 

 that they disperse a considerable quantity of light. This would 

 indeed explain the action said to be exhibited by orange, yellow, 

 and green wafers — the turning of the red of the spectrum 

 orange. The dispersion of the predominant coloured light over 

 the red is sufficient for this. How a green wafer generates a 

 white band in the blue I am unable to say, inasmuch as I can- 

 not repeat the experiment. 



A pale red glass which absorbs the green between b and F 

 (probably coloured with purple of Cassius) and a pale yellow 



