ly of any value, as it cannot be
corrected. The deviation observed when the ship inclines to either side
is due--(1) to hard iron acting vertically upwards or downwards; (2) to
vertical soft iron immediately below the compass; (3) to vertical
induction in horizontal soft iron when inclined. To compensate (1)
vertical magnets are used; (3) is partly corrected by the soft iron
correctors of D; (2) and the remaining part of (3) cannot be
conveniently corrected for more than one geographical position at a
time. Although a compass may thus be made practically correct for a
given time and place, the magnetism of the ship is liable to changes on
changing her geographical position, and especially so when steaming at
right angles or nearly so to the magnetic meridian, for then
sub-permanent magnetism is developed in the hull. Some vessels are more
liable to become sub-permanently magnetized than others, and as no
corrector has been found for this source of deviation the navigator must
determine its amount by observation. Hence, however carefully a compass
may be placed and subsequently compensated, the mariner has no safety
without constantly observing the bearings of the sun, stars or distant
terrestrial objects, to ascertain its deviation. The results of these
observations are entered in a compass journal for future reference when
fog or darkness prevails.

Every compass and corrector supplied to the ships of the British navy is
previously examined in detail at the Compass Observatory established by
the admiralty at Deptford. A trained observer acting under the
superintendent of compasses is charged with this important work. The
superintendent, who is a naval officer, has to investigate the magnetic
character of the ships, to point out the most suitable positions for the
compasses when a ship is designed, and subsequently to keep himself
informed of their behaviour from the time of the ship's first trial. A
museum containing compasses of various types invented during the 19th
century is attached to the Compass Observatory at Deptford.

The mariner's compass during the early part of the 19th century was
still a very imperfect instrument, although numerous inventors had
tried to improve it. In 1837 the Admiralty Compass Committee was
appointed to make a scientific investigation of the subject, and
propose a form of compass suitable alike for azimuth and steering
purposes. The committee reported in July 1840, and aft