CHARACTER OF THE ARMOUR-PLATED SHIPS OF THE ROYAL NAVY. 277 
much from the vertical soft iron and the hard iron, the first = ^2)+ 1^ ^ an ^ *°’ 
the second =■ tan H°; and these two parts are tabulated in the next columns. 
If we have not an opportunity of observing the vertical force on a sufficient number 
of points to obtain its mean value, the values observed will be affected by soft iron 
represented by the rod g , in the following figure : — 
Fig. 6. 
the value of [m on any azimuth £ being in fact increased by + cos where 6 is the 
dip. It is therefore convenient to know the values of q or and these are also 
^ ^ tan 0 
tabulated ; g does not change with a change of geographical position. 
In comparing the heeling error when the ship’s head is North or South, we must 
beware of falling into the error of confusing the two senses in which these words may 
be used. It may seem most natural to suppose the ship’s head to be North or South 
when upright, and that she is then heeled over without altering her direction. In that 
case we should have (nearly) 
Heeling error head North : heeling error head South : : 1 — 25 : 1 +93. 
In fact the heeling error is nearly inversely proportional to the directive force on the 
needle. 
But this is not the sense in which the term is generally used. In general we suppose 
the ship swung when heeled to starboard and again when heeled to port, and the devia- 
tions tabulated in the usual way, according to the ship’s azimuth by disturbed compass. 
In this case, which is the simplest mode of considering the error for the purpose of 
correction, the heeling error, head North, will only differ from the heeling error, head 
South, by reason of the quantity g , i. e. by reason of the difference of the vertical and 
not of the horizontal forces in the two positions. 
The importance of the heeling error, owing to its large amount in certain ships, will be 
seen in the discussion of the values given in the Tables ; and the importance of being 
able to determine it by observations easily made, and without the necessity of actually 
heeling over the ship, can hardly be overrated. 
We are now in a position to consider the numerical values of the coefficients given in 
the Tables. 
