274 Alterations in the Light of the Sun. 



at noon, fig. 2, 185; at four o'clock P. M., fig. 3, 110; ana 

 ut ten minutes past four, fig. 4, 100. 



When the spectrum was longest, the yellow interposed 

 between the green and the orange was pure; when the 

 violet disappeared, the yellow assumed a deeper tinge : it 

 was coloured red, and partook of the orange colour. 



I repealed these observations several times, at various 

 times of the year, and always with the same success; the 

 spectrum increased or diminished in length according as 

 the colour of the disk was whiter or yellower. 



Finally, on the 15th of January 1605, I remarked, on 

 observing the decomposition of the light of the setting sun, 

 that when the disk of that orb was of a fine red, the length 

 of the spectrum was diminished more than one half: it 

 was no more than 70 millimetres long; whereas at mid- 

 day it was 1S5 ; and in the series of colours separated by the 

 prism, we could only distinguish the red, orange, and green. 



M. Gerard, draftsman to the polytechnic school, having 

 been present on the above occasion, drew and coloured the 

 solar spectrum in question. Figure 5 is the copy of the 

 drawing, and fig. 6 presents the image of the sun when it 

 was received on a white card in the dark room : this tint is 

 deep orange, approaching the blush of dawn. 



The subtraction of one or several coloured rays in the 

 fasciculus which the sun sends us, when its disk is yellow, 

 orange, or red, may be easily remarked in the irises observed 

 at different hours of the day, either in the series of colours 

 which they present, or in the breadth of the coloured arcs. 

 I have several times verified this fact since my experiment 

 in 1802; an experiment which, I must confess, then 

 seemed to be very important. I have even remarked in the 

 sky, when the disk of the sun was red, irises which con- 

 tained only red, orange, and green, like the figure of the 

 spectrum which I have presented to the class. 



From these facts we may conclude, that among the causes 

 which may produce the alterations observed in the colour oi 

 the sun's disk, one of the most important is the subtrac- 

 tion of the coloured rays intercepted by the medium which 

 they pass through ; and the coloured' molecules separated 

 from the fasciculus of white light are, the purples and a 

 part of the violets, when the disk appears yellow: the pur- 

 ples, the violets, the yellows, and a part of the indigo blues, 

 when the disk appears orange : the purples, the violets, the 

 indigo blues, the blues, the yellows, and a Utile of the 

 orange, when the disk appears red : finally, that there may 

 be a arm, a colouring of the disk in red, undir the polar 

 circle, at which all the coloured molecules, at least the red, 



are 



