IM 
of considerable attenuation. Arago, for instance, has ob- 
served on more than one occasion the apparent adhesion of 
a star for three or four seconds to the dark limb of the 
moon after it had been perceived to be in contact with it, 
and he has also observed a very sensible diminution of 
brightness previous to immersion. 
From the distortion of the visible segment of the solar 
disc observed by Euler during the eclipse of 1748, Du 
Sejour, after making corrections for the effects of irradia- 
tion, arrived at the conclusion that the moon possesses an 
atmosphere having a horizontal refraction of 1.5", and 
which is therefore 1,400 times more rare than common 
atmospheric air upon the surface of the earth. 
A phenomenon similar in kind to the twilight of the 
earth has been recognised by Shroeter, in the form of a 
faint crepuscular light extending from each of the lunar 
cusps along the circumference of the unenlightened portion 
of her disc, from which he has been enabled to deduce the 
existence of an atmospheric envelope about our satellite 
capable at an elevation of 5,000 feet above her surface of 
causing a sensible inflexion of the light proceeding from a 
celestial body. But as the moon would describe an arc 
representing this amount in less than two seconds of time, 
“ the circumstance has been adduced as affording: a sufflcient 
explanation of the difficulty of detecting a lunar atmosphere 
in the phenomena of occultations.” 
Chromatic dispersion is the test which in certain circum- 
stances at any rate seems to offer better opportunities for 
ascertaining positively the existence of a lunar atmosphere 
than the test of simple refraction. It would cause the 
colour of an occulted object to change, making it green, and 
finally blue at the instant of disappearance. 
The direct telescopic observation, by which alone this 
appearance could be noticed, would, from the operation of 
circumstances upon which I need not dwell, probably lead 
