Let us start with Mercury in perihelion at the point _A_. The little cross on the planet stands exactly under the sun and in the center of the illuminated hemisphere. The large arrows show the direction in which the planet travels in its revolution about the sun, and the small curved arrows the direction in which it rotates on its axis. Now, in moving along its orbit from _A_ to _B_ the planet, partly because of its swifter motion when near the sun, and partly because of the elliptical nature of the orbit, traverses a greater angular interval with reference to the sun than the cross, moving with the uniform rotation of the planet on its axis, is able to traverse in the same time. As drawn in the diagram, the cross has moved through exactly ninety degrees, or one right angle, while the planet in its orbit has moved through considerably more than a right angle. In consequence of this gain of the angle of revolution upon the angle of rotation, the cross at _B_ is no longer exactly under the sun, nor in the center of the illuminated hemisphere. It appears to have shifted its position toward the west, while the hemispherical cap of sunshine has slipped eastward over the globe of the planet. In the next following section of the orbit the planet rotates through another right angle, but, owing to increased distance from the sun, the motion in the orbit now becomes slower until, when the planet arrives at aphelion, _C_, the angular difference disappears and the cross is once more just under the sun. On returning from aphelion to perihelion the same phenomena recur in reverse order and the line between day and night on the planet first shifts westward, attaining its limit in that respect at _D_, and then, at perihelion, returns to its original position. Now, if we could stand on the sunward hemisphere of Mercury what, to our eyes, would be the effect of this shifting of the sun's position with regard to a fixed point on the planet's surface? Manifestly it would cause the sun to describe a great arc in the sky, swinging to and fro, in an east and west line, like a pendulum bob, the angular extent of the swing being a little more than forty-seven degrees, and the time required for the sun to pass from its extreme eastern to its extreme western position and back again being eighty-eight days. But, owing to the eccentricity of the orbit, the sun swings much faster toward the east than toward the west, the eastward motion occupying about thirty-seven days and the westward motion about fifty-one days. [Illustration: THE REGIONS OF PERPETUAL DAY, PERPETUAL NIGHT, AND ALTERNATE DAY AND NIGHT ON MERCURY. IN THE LEFT-HAND VIEW THE OBSERVER LOOKS AT THE PLANET IN THE PLANE OF ITS EQUATOR; IN THE RIGHT-HAND VIEW HE LOOKS DOWN ON ITS NORTH POLE.]