He found that the magnetic needle behaved just as if the iron sphere were the earth itself, pointing to the poles and showing different directions at different distances from them. The thought flashed into his mind that the earth itself might be a great magnet. This thought was entirely new. The ancients had imagined that the attraction of the magnetic needle was due to some magnetic mountains in the north, or some star in the tail of the Great Bear, a group of seven stars which appear to move round the north pole of the sky. [Illustration: Fig. 20.--A magnetic Compass.] This supposition could not be maintained as the facts became better known. In the first place, since the stars appear to be carried round the north celestial pole, one would naturally expect that the direction of the magnetic compass would change in the course of every twenty-four hours, which it did not. [Illustration: Fig. 21.--Old form of Magnetic Dip. Circle.] Then, as to the magnetic mountains, if they were situated at the north pole, the compass would point in a horizontal direction only in the region near those mountains. In southern lands and seas the needle would have to point below the horizon, since the pole itself is at a considerable angle below the horizon. The angle between the needle and the horizon would have to become less and less as the ship sailed towards the pole. This is, however, just the reverse of what actually takes place. When a needle is balanced carefully before it is magnetised, then on touching it with the magnet, and so magnetising it, it is found that it balances no longer. The pole which points to the north points down, and the pole which points to the south is raised. This might be taken to indicate the existence of magnetic mountains at the pole of the earth. But on proceeding to the equator, this so-called “dip” of the needle becomes less, and on sailing towards the pole it becomes greater and greater, until a point is reached at which the needle points vertically downwards. Dr. Gilbert found that a small needle brought near a magnetic sphere behaves in a similar manner. At a distance midway between the two poles the needle points in a direction parallel to the axis of the sphere. At the two poles the needle points vertically downwards towards the centre of the sphere. If this variation of the “dip” were quite regular, sailors might discover their latitude on the sea by simply observing the “dip,” and Dr. Gilbert thought that it was quite possible. In this he was mistaken, because not only does the “dip” vary in an irregular manner, but it also varies from century to century, being greater or less at the same place from time to time.