COMPASS IN IRON ARMOUR-PLATED, IRON, AND COMPOSITE SHIPS. 
617 
these coefficients have been computed, as they, together with the methods of making 
the original observations, were so fully described in the paper on “ The Magnetic 
Character of the Armour-plated ships of the Royal Navy,” &c., 1865, to which 
allusion has already been made. 
On comparing the results, however, in that paper with those now brought forward 
for discussion, a remarkable difference will be observed. In the paper for 1865, the 
coefficients of the semicircular deviation are those of compasses to which no 
mechanical correction by permanent magnets had been applied. The corresponding 
coefficients in the accompanying Table III. have, with the exception of the original 
values in England, been computed from the deviation of compasses for which a 
permanent bar magnet, or magnets, has been employed to annul or correct the 
semicircular deviation. 
This correcting magnet has in every case been permanently fixed horizontally in the 
compass pillar, in the resultant of the magnetic forces producing semicircular 
deviation, and at a distance found tentatively below the card, after the several 
horizontal and vertical forces affecting the compass had been ascertained. 
It may be asked, whether this application of correcting bar magnets of possible 
variable magnetic moment does not in itself introduce an element of change in the 
deviation in addition to those of the ship ? It may be answered, with the reasonable 
confidence induced by fifteen years’ trial, that the permanency of the magnetic moment 
of the magnets employed is considered to be assured. 
Thus, in the ships named in the tables, the correction by magnets may be considered 
as the introduction into them of a permanent magnetic force acting independently on 
their compasses, and in opposition to the permanent magnetic forces of the ships. 
It is now proposed to pass on to the chief object of this paper, which is to show 
the amount and direction of the changes which take place in the deviations of the 
standard compasses in six different classes of modern vessels in the Royal Navy, on 
change of magnetic latitude. 
Taking the exact coefficients in Table III. in the order in which they stand we have 
] first:— 
The constant deviation. 
21 . 
Some rather large values of this coefficient are found in the tables, but it has been 
proved that for standard compasses placed in the central fore and aft line of the vessel 
where the iron is symmetrically placed with respect to that position, little or no real 
value from magnetic causes has been observed. An error in the bearing of the distant 
object used for swinging the ship, swinging her too fast, or prism error in the azimuth 
circle, gives fictitious values of this coefficient, which those in the table are considered 
to be. 
