352 
ME. EVA^’-S ON THE DE'^’IATION OF THE COMPASS 
This error, as also a smaller one considered to exist after the ship’s amval at Holy- 
head, was, as I understood, eliminated prior to my Hsit at that port in the latter part of 
October ; but on then determining the deviations of the Admiralty Standard compass, 
and referring the results to the ship’s compensated standard, a residual en'or in the latter 
of nearly 5° still existed on some of its points. 
[Although the question of compensation is extraneous to the general tenor of this Ee- 
port, I deem it of interest to record the above circumstances, to coiTect a misapprehension 
as to the employment of long needles, carried, in the compasses of the Great Eastern, 
to the extent of 11 inches for standard and steering, and 16 inches for the mast-compass. 
A series of experiments, still in progress, which I have instituted in elucidation of the 
comparative errors due to the use of long and short needles, both where the distm'bing 
cause is comparatively close to the compass, as in the case of the rudder-head near the 
steering compass, or the top sides and iron beams when heeling, and in compensated 
compasses from the necessary proximity of bar magnets, lead to the conclusion that no 
compass-needle should exceed in length, for seiuice in an ii'on ship, 6 or 7 inches ; the 
card of course may be enlarged as practically convenient. 
The errors of long needles, under the conditions just named, of proximity to perma- 
nent magnetic bodies, arise from this notable fact, that their- deviations assume the foi-m, 
not of a curve which is simply “ semicircular,” as the cui-ve is found to be which results 
from the ship’s general magnetic force when the length of the needle may be relatively 
considered as infinitesimally small, but of a “ sextantal ” superposed on a “ semicircular ” 
curve, the various maxima of the “ sextantal ” curve increasing with the square of the 
length of the needle. Assuming the zero of both curves (the semicircular and sextantal) 
to be at the magnetic North and South, the maxima of the sextantal deviations (by 
ship’s compass) are respectively at about N. 30° E., East, S. 30° E., S. 30° W., West, 
N 30° W., the points of coincidence with the semicircular curve being at North, N. 60° E., 
S. 60° E., South, S. 60° W., and N. 60° W.] 
The tendency of the direction of the ship’s force in the Great Eastern is to assume 
a fore-and-aft line, supporting the view, that time, with the vibrations and concussions 
due to sea-service, leads to a settled distribution of the magnetic lines ; the respective 
sections of the hull which hav^e north and south polarity being separated by lines ap- 
proximating more nearly to a horizontal equatorial plane through the body of the ship, 
instead of the inclined equatorial plane of separation due to the magnetic dip of the 
locality, and divergence from the magnetic meridian of the hull while building. 
A similar tendency of the direction of the ship’s force to approach a fore-and-aft line 
will be observed in several examples of Table I. 
The example of the Great Eastern offers this practical information ; that prior to a 
newly built iron ship being sent to sea, her head (while being equipped) should be 
secured in an opposite direction to that in which she was built ; and that the magnetic 
lines should be assisted to be “ shaken down ” by the vibrations of the machinery in a 
short preparatory trip prior to the determination of her compass errors, or their com- 
