a strong attraction on the
magnetic needle. Another complication is introduced by the fact that
the iron of the ship will always become more or less magnetic, and when
the ship is built of steel, as modern ones are, this magnetism will be
more or less permanent.

We have already said that a magnet has the property of making steel or
iron in its neighborhood into another magnet, with its poles pointing
in the opposite direction. The consequence is that the magnetism of the
earth itself will make iron or steel more or less magnetic. As a ship
is built she thus becomes a great repository of magnetism, the
direction of the force of which will depend upon the position in which
she lay while building. If erected on the bank of an east and west
stream, the north end of the ship will become the north pole of a
magnet and the south end the south pole. Accordingly, when she is
launched and proceeds to sea, the compass points not exactly according
to the magnetism of the earth, but partly according to that of the ship
also.

The methods of obviating this difficulty have exercised the ingenuity
of the ablest physicists from the beginning of iron ship building. One
method is to place in the neighborhood of the compass, but not too near
it, a steel bar magnetized in the opposite direction from that of the
ship, so that the action of the latter shall be neutralized. But a
perfect neutralization cannot be thus effected. It is all the more
difficult to effect it because the magnetism of a ship is liable to
change.

The practical method therefore adopted is called "swinging the ship,"
an operation which passengers on ocean liners may have frequently
noticed when approaching land. The ship is swung around so that her bow
shall point in various directions. At each pointing the direction of
the ship is noticed by sighting on the sun, and also the direction of
the compass itself. In this way the error of the pointing of the
compass as the ship swings around is found for every direction in which
she may be sailing. A table can then be made showing what the pointing,
according to the compass, should be in order that the ship may sail in
any given direction.

This, however, does not wholly avoid the danger. The tables thus made
are good when the ship is on a level keel. If, from any cause whatever,
she heels over to one side, the action will be different. Thus there is
a "heeling error" which must be allowed for. It is supposed to have
be