or South, to be divested of the error which the attraction of the iron in the ship may produce. In making them on land, it should be done on the open shore, so as that no attraction, purely local, may interfere; and if the direction of that shore be North and South, the experiment would be more satisfactory.
In an investigation of the cause why the attraction of the iron in a ship, and in some cases of the land, should decrease with the dip of the needle, and cease at the magnetic equator, the position of the dipping needle must again be consulted. At the equator it is horizontal; and therefore the line connecting the north and south polarities in each piece of iron in the ship, if it still possess magnetism, will also be horizontal, and the two attracting parts equally near to the level of the binnacle; and it should follow, that the attractions on the north and south points of the compass would be equal, and counteract each other. But it seems not improbable that stanchions and other upright pieces of iron, and perhaps the whole lose most, if not all their magnetism at the equator; from the rotatory motion of the ship not allowing any piece to have one end directed to the North, and the other end to the South, a sufficiently long time to acquire or retain magnetism. This was not the case where the dipping needle approached the perpendicular; for there, however the ship were turned, the upper part or end of each fixed piece of iron still remained the upper part; and the more nearly the needle stood to the perpendicular, the more strongly would the magnetism of the iron be concentrated at the upper and lower extremities, and consequently the more strong would be the attractive power on the compass. This I take to be the true cause of the errors increasing and decreasing in close connexion with the dip of the needle.
With respect to land near the magnetic equator, the analogy should not hold, because the magnetic vein or mass is not, like the iron in a ship, subject to a rotatory motion. Suppose that in the upper part of an island near the equator, there be a mass of iron ore, or other stone possessing magnetism; the north end of this mass will have a power of attracting the south point of the compass, and the south end, the north point; and it should follow, that when the centre of this island hears S.W. or N.E., at a little distance, the west variation should be less than when it bears S.E. or N.W. At the small island Trinidad, where the south end of the needle