80 
mr: airy on the observed deviations of the compass 
3. For the same ship, the mass of unmagnetized iron, if adjusted at one port, will 
produce its due effect at all parts of the world, without ever requiring change or 
adjustment. The quadrantal deviation may thus be accurately corrected without 
difficulty, leaving only the polar-magnet-deviation uncorrected. 
4. The elements of polar-magnet- deviation are liable to changes, but of very 
different amounts in different ships. In some (as the Trident), even in the voyage 
of an iron steamer from the Thames to Rio Janeiro, the ship’s subpermanent mag- 
netism is so little altered, that, if the compass were rigorously corrected in the 
Thames, it would (as to sense) be rigorously correct at Rio Janeiro ; in others there 
is such a change, in going to the Cape of Good Hope, that the compass might be in 
error 5 ° or 6°. This is nearly the greatest error that appears in the observations 
discussed above. 
5. It is therefore imperatively the duty of every captain of a ship, particularly of 
an iron-built ship, to examine the state of the compasses at every opportunity. For 
the correctness of the compasses may be vitiated, not only by the changes in the 
polar-magnetism of the ship, but also by changes in the intensity of the magnets used 
for the correction. But as the correction of the quadrantal deviation is not liable to 
any doubt whatever, it is sufficient, for ascertaining the existence and recording the 
anjount of error of the polar-magnet-deviation, to observe the error when the ship’s 
head is N. or S., and when it is E. or W. 
6. From whatever cause the changes in the elements of polar-magnet-deviation 
arise (whether from a real change in the subpermanent magnetism of the ship, or 
from the variation of that part of induced magnetism which is similar to polar- 
magnetism but which changes in different magnetic latitudes), they may be pre- 
cisely corrected by readjusting the position of the magnets, leaving the unmagnet- 
ized iron undisturbed. And the change (if there is any) in the intensity of the 
correcting magnets will also be corrected, as to its effect on the compass, by the same 
readjustment of position. 
7. It is therefore highly desirable that the magnets should be mounted in such a 
manner that their distance from the compass can be delicately changed. And, as 
the easiest way of preserving a register of the ship’s magnetism, it is desirable that 
there should be means of registering the positions of the magnets. 
8. In a ship’s first voyage, there are no means of correcting the errors of the 
compass at different parts of the earth, except by such adjustment of the distances 
of the magnets. But if, on the ship’s return to England, her subpermanent mag- 
netism is found to be unaltered (which affords presumption that it has been unaltered 
during the voyage), and if the elements of magnetism have been registered either by 
record of the positions of the correcting magnets or by such discussions as those 
which occupy this Memoir, then it will be possible to correct by magnets that part 
of the polar-magnet-deviation which is due to subpermanent magnetism only (and 
which, alone, would be sensible at the magnetic equator), and to correct the remain- 
