154 
SIE DAVID BEEWSTEE AXD DE. J. H. GLADSTONE OX 
become rather less luminous as the sun descends ; but the greatest effect is observable in 
the narrow band a 8, the nearest to B, the appearance of which at that time is repre- 
sented in fig. 9, where the scale is double that of fig. 2. C and most of the lines between 
it and C6 are deepened; and the last-mentioned line is particularly noticeable, as it begins 
to intensify when the sun is at a considerable altitude ; indeed in Great Britain it is 
very evident during winter at any hour of the day. When the sun is about to set, or 
when it is just rising, it is one of the most strongly marked lines in the whole spectmm. 
C 15 increases to a black band, C 16 becomes more evident than before, and the double 
line D is very strongly developed ; D 1 deepens in shade ; and at about D 2 commences 
a band, marked by the Greek letter S in the diagram (fig. 7), which is one of the most 
characteristic features of the prismatic image of light that has passed through a long 
space of air. It is discernible in the diffuse light of a dull day at any hour ; it is that 
which Professor W. A. Miller observed manifesting itself on the occasion of a thimder 
shower*; and it becomes evident in the dmect solar rays when the lu min ary is several 
degrees above the horizon : from its occurrence in a most luminous part of the ordinary 
spectrum, and from its great breadth, it intercepts a large amount of light, even at that 
time ; and when the sun is just setting, it becomes a broad space of almost total darkness. 
It appears to cover a larger amount of the image in the direction of E, as it deepens in 
shade. It is succeeded by other bands, z and 2, and by a line n. The latter becomes 
very prominent, and from its vicinity to the comparatively faint line E, wliich is not per- 
ceptibly deepened by the atmosphere, has sometimes been mistaken for E itself. Beyond 
1) there are several remarkable bands, especially those designated in the map by the leite!> 
I and n. F itself seems to become nebulous ; and between it and G appear seven bauds. 
A, pj, v, I, 0, TT, ^ : the representation of the six last given in the map is on the authority 
of one single observation and drawing by Sir David Brewster. It is but rai'ely that 
this part of the spectrum can be seen at sunset ; and the portion beyond G is so seldom 
visible, and when seen is so faint, that no atmospheric lines or bands have been descried 
in it. 
Allusion has already been made to the fact that the band ^ is easily recognizable hi 
diffused light ; and this is more or less true, as might be expected, of the other bands 
and lines. Indeed the western sky after sunset affords a favourable opportunity for 
studying those in the more luminous parts of the spectrum. If the sky be red. C, I’ 6. 
D, and ^ generally appear as four very black bands ; if it be yellow, they are not so well 
defined. The various colours of the clouds also afford a great variety of appearance, but 
not so much in the character of the dark spaces, as of the luminous image which they 
intersect. There seems to be a difference in the visibility of these bands at diftercnt 
times, which is not readily accounted for; thus on October 29, 1837, at Allerly. near 
Melrose, at the instant of sunset the luminous sky gave a spectrum in which C 6. though 
distinctly seen, was not black, nor was D, nor while the line B was I'evy broad and 
deep. For thirty-seven minutes after sunset this black B was discernible; but even 
then, and indeed until the twilight had gone, the forenientioned bands, usually so 
* Pliilosophieal Magazine, August, 1815, p. 85. 
