hanical disturbance of the card caused by the vibration of the
hull in ships driven by powerful engines.

The effects of the iron and steel used in the construction of ships upon
the compass occupied the attention of the ablest physicists of the 19th
century, with results which enable navigators to conduct their ships
with perfect safety. The hull of an iron or steel ship is a magnet, and
the distribution of its magnetism depends upon the direction of the
ship's head when building, this result being produced by induction from
the earth's magnetism, developed and impressed by the hammering of the
plates and frames during the process of building. The disturbance of the
compass by the magnetism of the hull is generally modified, sometimes
favourably, more often unfavourably, by the magnetized fittings of the
ship, such as masts, conning towers, deck houses, engines and boilers.
Thus in every ship the compass needle is more or less subject to
deviation differing in amount and direction for every azimuth of the
ship's head. This was first demonstrated by Commander Matthew Flinders
by experiments made in H.M.S. "Investigator" in 1800-1803, and in 1810
led that officer to introduce the practice of placing the ship's head on
each point of the compass, and noting the amount of deviation whether to
the east or west of the magnetic north, a process which is in full
exercise at the present day, and is called "swinging ship." When
speaking of the magnetic properties of iron it is usual to adopt the
terms "soft" and "hard." Soft iron is iron which becomes instantly
magnetized by induction when exposed to any magnetic force, but has no
power of retaining its magnetism. Hard iron is less susceptible of being
magnetized, but when once magnetized it retains its magnetism
permanently. The term "iron" used in these pages includes the "steel"
now commonly employed in shipbuilding. If an iron ship be swung when
upright for deviation, and the mean horizontal and vertical magnetic
forces at the compass positions be also observed in different parts of
the world, mathematical analysis shows that the deviations are caused
partly by the permanent magnetism of hard iron, partly by the transient
induced magnetism of soft iron both horizontal and vertical, and in a
lesser degree by iron which is neither magnetically hard nor soft, but
which becomes magnetized in the same manner as hard iron, though it
gradually loses its magnetism on change of conditi