32 
only to results of uncertainty. At the same time it is but 
proper to remark that appearances have been observed at 
eclipses of the sun suggestive of this phenomenon, and 
which have been interpreted by Flamsteed as indicating the 
existence of a lunar atmosphere. The particular circum- 
stances in which, as I take it, chromatic dispersion may 
afford weighty evidence of the existence of a lunar atmo- 
sphere occur when the body occulted by the moon is one of 
considerable angular magnitude and great intrinsic splen- 
dour. Then it is manifest that the chromatic dispersion 
effected by such an atmosphere would cause the projections 
of prismatic bands upon the earth, forming as it were an 
iris on the borders of the shadow, and bathing the land- 
scape and the clouds in all the rainbow hues. These 
circumstances it will be seen exist during the totality of a 
solar eclipse, and the rainbow hues bathing alike the land- 
scape and the sky, which I have indicated as the inevitable 
consequences of chromatic dispersion by a lunar atmosphere, 
would seem to be almost constant accompaniments of such 
eclipses. ‘'As early as the year 840 it was remarked that 
during the total eclipse of the sun which happened in that 
year the colours of objects on the earth were changed.’' 
“ Kepler mentions that during the eclipse which occurred in 
the Autumn of 1590 the reapers in Styria noticed that 
everything had a yellow tinge,” whilst during that which 
took place in 1706 objects were observed to change their 
colour, now appearing of an orange yellow, and now of a 
reddish tinge. 
The illustrious Edmund Halley remarked that the face 
and colour of the sky were changed during the eclipse ob- 
served by him in 1715. “The serene azure of the sky,” he 
says, “turned to a more dusky livid colour, intermingled 
with a tinge of purple, and grew darker and darker until 
the total immersion of the sun.” Sir John Clarke, in his 
account of the eclipse of 1737, states that “the ground v/as 
