While, therefore, the earth is nearly a round ball, it must not be supposed that it is exactly spherical in form. We may disregard the small irregularities of its surface, for even the greatest mountains are insignificant in height when compared with the entire diameter of the earth itself. But even leaving these out of account, the earth is not perfectly spherical. We can describe it best as a flattened sphere. It is as though one were to press a round rubber ball between two smooth boards. It would be flattened at the top and bottom and bulged out in the middle. This is the shape of the earth. It is flattened at the poles and bulges out near the equator. The shortest straight line that can be drawn through the earth's centre and terminated by the flattened parts of its surface may be called the earth's axis of figure; and the two points where this axis meets the surface are called the poles of figure. But the earth has another axis, called the axis of rotation. This is the one about which the planet turns once in a day, giving rise to the well-known phenomena called the rising and setting of sun, moon, and stars. For these motions of the heavenly bodies are really only apparent ones, caused by an actual motion of the observer on the earth. The observer turns with the earth on its axis, and is thus carried past the sun and stars. This daily turning of the earth, then, takes place about the axis of rotation. Now, it so happens that all kinds of astronomical observations for the determination of latitude lead to values based on the rotation axis of the earth, and not on its axis of figure. We have seen how the earth's equator, from which we count our latitudes, is everywhere 90 degrees distant from the pole. But this pole is the pole of rotation, or the point at which the rotation axis pierces the earth's surface. It is not the pole of figure. It is clear that the latitude of any observatory will remain constant only if the pole of figure and the rotation pole maintain absolutely the same positions relatively one to the other. These two poles are actually very near together; indeed, it was supposed for a very long time that they were absolutely coincident, so that there could not be any variations of latitude. But it now appears that they are separated slightly.