ten and gradually disappear from view
without the body of the ship to which they belong becoming visible at
all. Since one knows from experience that there is no _edge_ at the
horizon over which a vessel can drop down, the appearance which we have
been describing can only be explained by supposing that the surface of
the earth is always curving gradually in every direction.

The distance at which what is known as the _horizon_ lies away from us
depends entirely upon the height above the earth's surface where we
happen at the moment to be. A ship which has appeared to sink below the
horizon for a person standing on the beach, will be found to come back
again into view if he at once ascends a high hill. Experiment shows that
the horizon line lies at about three miles away for a person standing at
the water's edge. The curving of the earth's surface is found, indeed,
to be at the rate of eight inches in every mile. Now it can be
ascertained, by calculation, that a body curving at this rate in every
direction must be a globe about 8000 miles in diameter.

Again, the fact that, if not stopped by such insuperable obstacles as
the polar ice and snow, those who travel continually in any one
direction upon the earth's surface always find themselves back again at
the regions from which they originally set out, is additional ground for
concluding that the earth is a globe.

We can find still further evidence. For instance, in an eclipse of the
moon the earth's shadow, when seen creeping across the moon's face, is
noted to be _always_ circular in shape. One cannot imagine how such a
thing could take place unless the earth were a sphere.

Also, it is found from observation that the sun, the planets, and the
satellites are, all of them, round. This roundness cannot be the
roundness of a flat plate, for instance, for then the objects in
question would sometimes present their thin sides to our view. It
happens, also, that upon the discs which these bodies show, we see
certain markings shifting along continually in one direction, to
disappear at one side and to reappear again at the other. Such bodies
must, indeed, be spheres in rotation.

The crescent and other phases, shown by the moon and the inferior
planets, should further impress the truth of the matter upon us, as such
appearances can only be caused by the sunlight falling from various
directions upon the surfaces of spherical bodies.

Another proof, perhaps indeed the weig