THE ILLUMINATION OF THE ECLIPSED MOON. 



367 



ivith a difference ; and it is this difference tliat I wiali par- 

 ticularly to insist upon. 



I have ca]<!ulatod for each position the height in the 

 atmosphoro through which rays from the ceritre of the sun 

 must have passed to reach the moon, and I have made a 

 sinrilar calculation for directly refracted rays from the limb 

 of the sun. 



In those calculations I have assumed that the mean 

 barometric pressure at sea-level is thirty inches, that a solar 

 ray passing through air of that density suffers a refraction 

 of sixty-six minutes, and that a ray traversing the atmos- 

 phere at any given height above the earth suffers a refrac- 

 tion proportional to the maximum pressure which it 

 encounters. In converting pressure into height I have 

 followed tables constructed from actual observations. The 

 following are the results : — 



Centre of sun 

 Iximb of sun 



Moon in Perigee. 

 1,752 feet 

 9,714 „ 



Moon in Apogee. 

 5,7(51 feet. 

 i;5,a07 „ 



I cannot but thiidt that wo have here the principal caxiso 

 of the differences observed in the visibility of the moon in 

 different eclipses. I suppose no one will doubt that the 

 transparency of the air increases with elevation, whether 

 this bo duo to diminishing vapoui- or to increasing rarity, 

 or to both combined. Moreover, it is certain that above the 

 limit of the region of cloud there will be less obstruction to 

 the passn,go of tlio sun's rays tlian bolovv that limit. Now 

 it so happens that the heights through which the moon in 

 its two extreme positions will receive the sun's rays are 

 just such that the difference of three or four thousand feet 

 would be most important with reference to the amount of 

 cloud. The inference, to my mind, is very clear, that the 



G G 



