"No matter! I wish we were there now! Wouldn't it be jolly, dear boys, to have old Mother Earth for our Moon, to see her always on our sky, never rising, never setting, never undergoing any change except from New Earth to Last Quarter! Would not it be fun to trace the shape of our great Oceans and Continents, and to say: 'there is the Mediterranean! there is China! there is the gulf of Mexico! there is the white line of the Rocky Mountains where old Marston is watching for us with his big telescope!' Then we should see every line, and brightness, and shadow fade away by degrees, as she came nearer and nearer to the Sun, until at last she sat completely lost in his dazzling rays! But--by the way--Barbican, are there any eclipses in the Moon ?" "O yes; solar eclipses" replied Barbican, "must always occur whenever the centres of the three heavenly bodies are in the same line, the Earth occupying the middle place.

However, such eclipses must always be annular, as the Earth, projected like a screen on the solar disc, allows more than half of the Sun to be still visible." "How is that ?" asked M'Nicholl, "no total eclipses in the Moon?
Surely the cone of the Earth's shadow must extend far enough to envelop her surface ?" "It does reach her, in one sense," replied Barbican, "but it does not in another.

Remember the great refraction of the solar rays that must be produced by the Earth's atmosphere.

It is easy to show that this refraction prevents the Sun from ever being totally invisible.

See here!" he continued, pulling out his tablets, "Let _a_ represent the horizontal parallax, and _b_ the half of the Sun's apparent diameter--" "Ouch!" cried the Frenchman, making a wry face, "here comes Mr._x_ square riding to the mischief on a pair of double zeros again! Talk English, or Yankee, or Dutch, or Greek, and I'm your man! Even a little Arabic I can digest! But hang me, if I can endure your Algebra!" "Well then, talking Yankee," replied Barbican with a smile, "the mean distance of the Moon from the Earth being sixty terrestrial radii, the length of the conic shadow, in consequence of atmospheric refraction, is reduced to less than forty-two radii.