34 
accounted for. It is not unlikely that further and special 
observations will be required to say of either of these 
theories that it may or may not be maintained. 
In the meantime, considering that the non-existence of a 
lunar atmosphere is undemonstrated and undemonstrable, 
that it is in opposition to analogy, and that even simple 
refraction has given evidence of such an inconsiderable 
atmospheric envelope as we might at most expect a body of 
the moon’s small mass to have, it certainly seems to me 
that the balance of probability lies in favour of the theory 
that the rainbow hues observed at total eclipses of the sun 
are really the results of chromatic dispersion effected by a 
lunar atmosphere. 
