An observer at the North Pole of the earth, for instance, would see the stars moving round in circles parallel to the horizon. No star would either rise or set--one half of the heavens would be always visible above his horizon, and the other half invisible; whereas an observer at the South Pole would see that half of the stars invisible to the observer at the northern one, because it was the half below the N. horizon. If the observer be on the equator, the movements of the stars all appear as indicated in the above diagram--that is, all the stars will rise and set, and each star in turn will be half its time above the horizon, and half its time below it. But if we consider the position of an observer in middle latitude, say in London, we find that some stars will always be above the horizon, some always below--that is, they will neither rise nor set. All other stars will both rise and set, but some of them will be above the horizon for a long time and below for a short time, whereas others will be a very short time above the horizon and a long time below it. [Illustration: THE CELESTIAL SPHERE VIEWED FROM A MIDDLE LATITUDE. AN OBLIQUE SPHERE.] At _O_ we imagine an observer to be in latitude 45° (that is, half-way between the equator in latitude 0°, and the North Pole in latitude 90°), hence the North Celestial Pole will be half-way between the zenith and the horizon; and close to the pole he will see the stars describing circles, inclined, however, and not retaining the same distance from the horizon. As the eye leaves the pole, the stars rise and set obliquely, describe larger circles, gradually dipping more and more under the horizon, until, when the celestial equator is reached, half their journey is performed below it. Still going south, we find the stars rising less and less above the horizon, until, as there were northern stars that never dip below the horizon, so there are southern stars which never appear above it. _D D'_ shows the apparent path of a circumpolar star; _B B′ B″_ the path and rising and setting points of an equatorial star; _C C′C″_ and _A A′ A″_ those of stars of mid-declination, one north and the other south. [Illustration: A TERRESTRIAL GLOBE WITH WAFER ATTACHED TO SHOW THE VARYING CONDITIONS OF OBSERVATION IN A MIDDLE LATITUDE.] Wherever we are upon the earth we always imagine that we are on the top of it. The idea held by all the early peoples was that the earth was an extended plain: they imagined that the land that they knew and just the surrounding lands were really in the centre of the extended plain. Plato, for instance, as we have seen, was content to put the Mediterranean and Greece upon the top of his cube, and Anaximander placed the same region at the top of his cylinder.