COMPASS IN IRON ARMOUR-PLATED, IRON, AND COMPOSITE SHIPS. 619 
coefficients 33 and (S, arising from vertical induction in soft iron, may be computed 
whenever the magnetic inclination or dip is known, or can be taken from charts of 
that element. 
Before further investigating the effects of a change of magnetic latitude on the 
coefficients, the question of how far time affects the constants P and Q requires 
consideration. 
On looking over the values of P and Q, it will be remarked that in some ships a 
change takes place immediately after leaving England, which appears to be neither 
due to time nor change of latitude, but to another cause which will hereafter be 
referred to. Taking the values obtained subsequently at different times in the 
same geographical position it will be found— 
Bellerophon.—Quebec.| 
Iron Duke.—Hong-Kong.| 
Northampton.—Halifax, N.S.| 
Active.—Simon’s Bay, C. G. Hope . . | 
Raleigh.—Spithead. 
Inconstant—Spithead.^ 
f Plymouth...... | 
Himalaya.—< i 
f Simon’s Bay, 0. G. Hope < 
Albatross.—Sheerness and Plymouth . ^ 
Boxer.—Esquimault, Y. I.| 
Encounter.—Simon’s Bay, C. G. Hope . | 
P. 
Change. 
Q. 
1874 
-•0021 
[‘018 
-•003 
1876 
— '020 j 
-•008 
1880 
-•208' 
[>•015 
-•002 
1881 
— T93 _ 
-•037 
vi. 
1880 
+ •024' 
[>•008 
-•037 
X. 
1880 
+ "032 j 
-•057 
X. 
1874 
— T201 
[ -024 
-•019 
X. 
1876 
— •096 j 
-■006 
ix. 
1874 
+ ■0431 
[ -014 
+ •026 
vi. 
1876 
+ •057 j 
+ •001 
viii. 
1869 
-•0231 
[•021 
+ •016 
xi. 
1871 
— •044 j 
+ •017 
vii. 
1872 
-•0301 
[•009 
-•061 
Vll. 
1874 
-•021 j 
-•058 
viii. 
1872 
-■0301 
[ -009 
-•042 
i. 
1875 
-•021 j 
-•081 
iii. 
1874 
-•0661 
[•002 
+ •015 
X. 
1874 
— ‘064 j 
+ -022 
X. 
1869 
-•0081 
> +04 
-■049 
X. 
1874 
-•004 J 
-■030 
iv. 
1874 
+ •002 1 
>■ -050 
+ •022 
i. 
1876 
— 048 J 
-•001 
Change. 
j-005 
|-035 
j-020 
| -013 
| 025 
j-001 
| -003 
J -039 
j-007 
j-019 
\ -023 
Thus, in the worst case amongst the armour-plated, iron, and composite vessels, P 
takes a year to alter '015, and generally two or three years for about half that amount. 
A change of '015 in P would make about a degree change in the deviation. This 
evidently slow change of P bv time is important, as should P alter during the ship s 
sailing over a long range of magnetic latitude, the values of c, as found by the above 
formulae, are correspondingly untrustworthy. 
With regard to the constant Q, although more subject to change by time than P, it 
is comparatively of less importance, as, with few exceptions, the value of the constant 
/, which depends on the constancy of Q, is, if not zero, so small as to be neglected. 
Having accepted as the result of fifteen years’ trial that the correcting bar magnets 
