a well known fact that the rays of light are very much refracted when passing out of a rare into a denser medium; and about the poles of the earth it is believed, (and this belief is confirmed by navigators) that refraction increases very considerably, owing to the great density of the atmosphere. We have good reason then to believe that refraction throws the rays of the sun several degrees further within the sphere. But let us take the known refraction of the horizontal ray, at or near the equator (say one half of a degree) it would throw the rays of light so much further into the concave, and not leave quite thirty-seven degrees in the centre of the sphere deprived of the sun's rays. The motion of the earth causes the apparent motion of the sun to be about fifteen degrees in an hour, as the diurnal revolution of the earth causes the sun to move apparently through three hundred and sixty degrees in twenty-four hours. Now it is a well known fact to all that the sun gives us light sufficient to be called day-light, for about an hour after he descends below the horizon; consequently he must afford us light when he is fifteen degrees obscured from our view. Accordingly, the sun, though he might not be visible, would illuminate the concave part of the sphere fifteen degrees further than his direct rays fall, which reduces the space in the interior of the sphere to the breadth of not quite seven degrees which would still remain unlighted. But this is making calculations on the most unfavorable premises possible. Considering the form of the earth, and the power of refraction, I have no doubt but the direct rays of the sun would fall on every part of the inner sphere. However, I have proceeded on such premises as, I conclude, the most sceptical must admit. Light, we know, is reflected from solid bodies on which it falls, and also from the atmosphere: the rays of the sun, then, which would pass the lower part of the verge and fall on the opposite concave surface, would be reflected back in all directions, and most probably light the whole of the interior of the sphere sufficient for the ordinary purposes of life. By way of further illustration, suppose a perpendicular wall were raised on a plain, one mile high, does any person believe that there would be no light on the side of the wall opposite to the sun; although his rays would have to form an angle of one hundred and forty, or one hundred and fifty degrees, to reach the earth on that side of the wall? No axiom is more evident than that the rays of light are communicated to other places than those on which the rays of the sun fall directly; for example, we all know that a close room, however large, with a north window, will be sufficiently lighted by refraction and reflection from the atmosphere, provided there is no obstruction opposite the window, although the rays of the sun would have to form an angle of one hundred and fifty degrees to enter it, and why might not