the whole interior of the sphere be lighted in the same manner, even supposing the rays of the sun should never enter directly. The north polar opening being about four thousand one hundred and fifty miles in diameter, and the southern six thousand three hundred and fifty, with the whole force of the direct rays of the sun falling on and passing through the atmosphere at either polar opening, it would not require refraction, or reflection, to make an angle of ninety degrees to light the whole of the interior concave; and certainly the polar openings are sufficiently large for the purpose, when we compare a common window with the dimensions of an ordinary sized room. It is believed, by the adherents of the new theory, that the atmosphere, within the concave, and about the polar openings, is much denser than our atmosphere; which appears inevitably to be the case, as the centrifugal force on the convex has the tendency to throw the atmosphere _from_ the surface, and on the concave to force it from the centre of motion, and nearer _to_ the surface. This admitted, the rays of the sun passing out of a rare medium into a denser, would be refracted much further into the sphere; and the sun-shine on the surface of one sphere would be reflected obliquely, according to the angle of incidence, to the next sphere, and in this manner might be extended even beyond the centre of the concave. It is also believed, that near the verges of the polar openings, and perhaps in many other parts of the unfathomable ocean, the spheres are water quite through, (at least all except the _mid-plane-spaces_, or cavities) which being the case, light would probably be transmitted between the spheres. The apparent elevation of celestial bodies above their true altitude, is greatest when the body is on the horizon, which is ascertained to be a little more than half a degree; hence, in our climate, the sun appears three minutes sooner, and sets three minutes later than is really the case, which increases the length of our day six minutes, by refraction. This gradually increases in proceeding from the equator to the frigid zones; and at the poles, were the earth entire, the day should become thirty-six hours longer, by refraction alone, than it would otherwise be.[38] It was doubtless owing to some peculiar refractive power in the northern regions, that caused the Dutch, who wintered on Nova-Zembla, (which is in latitude between seventy and seventy-eight degrees,) on the approach of summer, to see the sun about two weeks sooner than he should have appeared in that latitude, according to astronomical calculation.[39] This tends to show that there is more refraction in the northern regions than is observable in the south.[40]