Discussion of the Observed Deviations of the Compass in several Ships , YWood-built and Iron-built : with a General Table for facilitating the examination of CompassDeviations .
By G. B. AIRY , Esq. , Astronomer Royal .
Received September 14 , -Read November 22 , 1855 .
IN the year 1839 I communicated to the Royal Society a paper ( printed in the Philosophical Transactions of that year ) containing the results of examination of the compass in two iron-built ships , and a general theory of the effect of the transient induced magnetism of iron in disturbing the direction of the compass-needle .
The result of the theory of induced magnetism may be stated as follows .
First , I must premise , in explanation of the ternlm " polar-magnet-deviation " which I shall frequently have occasion to use , the following theorem on the disturbance of the cotpass by a magnetized steel bar , or one which possesses independent polar magnetism , in no way referred to the influence of the existing terrestrial magnetism .
Let the line CA , fig. 1 , represent in magnitude and in direction the Fig , 1 .
terrestrial directive force .
[ In the applications of the theory to E ships , the terrestrial directive force is diminished in a constant ratio differing little from unity ; and then it must be understood that CA represents the terrestrial force so diminished .
] A is understood to / be the magnetic-north end of the line .
And let AB represent , in / // proportional magnitude and in direction , the directive force of the magnetized steel bar or " polar-magnet , " B corresponding to that end of the polar-magnet which possesses boreal magnetism .
Then the directive force which really acts on the compass-needle will be c represented in proportional magnitude and direction by CB ; and the angle ACB will be the angle of deviation of the compass .
And if the polar-magnet be turned round in azimuth , so that the point B occupies successively different points in the circumference of the circle , the angle of deviation will have successively the different magnitudes and the different directions ( right or left of the line CA ) given by this construction for these different circumstances .
This theorem is very simply founded on the ordinary " composition of forces , " and is abundantly proved by experiment .
The deviation ACB is what I shall call " polar-magnet-deviation .
" In some cases it is convenient to refer the azimuth of the polar-magnet to the true magnetic meridian or CA , and then the polar-magnet-deviation is given by this formula : AB .
sin BAD tan ACB-=CA+AB.cos BAD ' where BAD is the true azimuth of the polar-magnet .
In other cases it is convenient to refer the azimuth of the polar-magnet to the disturbed direction of the compassneedle or BE , and then the polar-magnet-deviation is given by this formula : AB sin ACB-= sin BF , CA where EBF is the apparent azimuth of the polar-magnet .
In either case , the law which connects the polar-magnet-deviation with the azimuth ( true or apparent ) of the polar-magnet is what I shall call " the law of polar-magnet-deviation .
" Secondly : the disturbing effect of the polar-magnet , whose power is represented by AB , may be completely neutralized by attaching to the same fr-ame ( whether it be a ship , or an experimental wood fiame , &amp;c. ) which carries that polar-magnet , another polar-magnet in the opposite position , its power and direction being represented by the line Ab .
Thirdly : if , instead of the polar magnetism of a steel bar , the distulbing forlce upon the compass be that of the transient induced magnetistm in a nearly spherical mass of soft iron possessing no permanent mragnetism , plane as the compass ; and if NOS , fig. 2 , represent the position taken by the needle under the action of terrestrial magnetism only ; then if the mass of soft iron be in either of the positions M , , M , , M , , M it will not disturb the needle NOS : if the mtass of soft iron be placed either in the quadrant between M , and M2 ( as at M , ) or in the quadrant between M3 and M4 ( as at M6 ) , it will make the point N deviate towards n , ; and if the soft iron be placed in either of the remaining quadrants ( as at M7 or M8 ) it will make the point N deviate towards n7 .
The amount of deviation is proplaced in the same horizontal Fig. 2 .
M.7^ , --"'. .
/ ' ?
M I_tI\\ / \ -%.6 / *N 6\\ s SI M.S 8 x\ 6 v-s x , , portional to the sine of double the angle of azimuth of the disturbing mass , that is to the sine of double the angle M1OM , , or M , OM8 , &amp;c. If the disturbing mass be carried round the circle in the direction MM2M , M4M , I , the deviation of the needle ( estimated positive when the point N is moved towards the right , and negative when towards the left ) will in the four quadrants have the signs +-+- .
The deviation following this law I shall call " quadrantal deviation .
" Fourthly : the deviation produced by the mass of soft iron at M , will not be corrected by placing a similar mass at M , ( which , instead of cotrrecting the deviation , will double it ) , but it will be corrected by placing a similar mass either at m , or at m6 , the angles M5Om , and M,50m being supposed to be right angles .
Similarly , the deviation produced by the mass at M6 will be neutralized by a similar tmass either at mS or at m6 , if the angles MoOms , M6Om6 are right angles .
Thus the " quadrantal deviation " may be corrected by attaching to the same frame ( whether it be a ship , or a wooden experimental fiamre , &amp;c. ) , which carries the mass that produces the " quadrantal deviation , " another mass , at the same level as the compass but in an 54 azimuth differing 90 ?
from that of the disturbing mass .
And , if it be found that when a ship 's head is in the quadrant between N. and E. , or between S. and W. , the needle deviates to the right , and the opposite way for the remaining quadrants ; or that , in respect to the quadrants of azimuth of ship 's head , the quadlantal deviation follows the law +-which I shall call " positive quadrantal deviation ; " the inference is that the deviation is of the same kind as would be produced by a mass of iron at the same level as the compass , either headward or sternward of the conmpass : and it may be neutralized by placing a mass of iron at the same level as the compass , either on the starboard or on the port side .
But if the deviation follow the law ++ in respect of the four quadrants of azimuth , which I shall call " negative quadrantal deviation , " it may be neutralized by placing a mass of iron at the same level as the compass , either headward or steirnward of the compass .
All these laws I have abundantly confirmed by experiment .
These being premised , the laws of the deviation of the compass produced by the transient induced magnetism of a ship , as shown by swinging the ship round in a given locality , will ( according to the theory to which 1 have referred ) be as follows : ( 1 ) .
There will be a force , sinilar to the force of a polar-magnet , and producing a polar-magnet-deviation .
In northern magnetic latitudes , the nature of the effect will usually be the same as if the boreal magnetismn were towards the ship 's head : in southern magnetic latitudes , it will be usually the same as if the austral magnetism were towards the ship 's head .
The absolute magnitude of the polar-magnet-force will be a multiple of the vertical tetrestrial magnetism ; the proportion which it bears to the terrestrial directive force , which is the proportion of AB to AC in fig. 1 ( supposing no other polar-magnet-force to act ) , will be a multiple of the tangent of dip , without regard to the absolute force .
( 2 ) .
There will be a quadrantal deviation ; and this deviation will be the same in all magtletic latitudes , and whatever be the magnitude of the earth 's magnetic force .
It will usually be a positive quadrantal deviation .
These are the disturbances that are produced by transient induced magnetism only .
But if the iron that enters into the composition of a ship possess independent polar magnetism similar to that of a magnetized steel bar ( i. e. not depending on the terrestrial magnetism at the present moment for its existence ; and not changing its amount or quality or direction in regard to the ship 's keel , while the ship is swung round in different positions ) , which from the slowness of its changes , though probably more variable than that of a steel bar , I propose to call " subpermanent magnetism ; " it will be necessary for us to consider how the expression for the effects of this subpermanent magnetism can be most easily combined with those for the induced transient magnetism .
It is readily seen that the polar-magnet-force of subpermanent magnetism must be combined with the polar-magnet-force of induced transient magnetism ; and that , at a given locality , they cannot be separated .
In 55 fig. 3 , let SH represent the magnitude and direction of the polar-magnet-force of induced magnetism , directed from the ship 's stern to her head ( this diagram having no relation to the direction of terrestrial magnetism ) , and let HP F 3 .
represent the magnitude and direction of the subpermanent mag7 Q_ netism , which , inasmuch as its direction is invariable with respect / to the ship , is inclined at a constant angle to SH : then the force resulting from the composition of these two will be represented in magnitude and direction by SP , which is invariable in magnitude , and inclined at a constant angle to SH .
And this force will appear , in the phenomena of compass-disturbance at any one locality , as a whole , and cannot immediately be separated into the two parts SH and HP .
All that can be done is this .
At a given locality we can find the direction of SP with regard to the ship 's keel , if , by methods to be explained below , we can find the " neutral position of the ship in reference to the polar-magnet-force , " or ( which is the same thing ) the azimuth of the ship 's head , or of the line SH , when the polarmagnet-deviation vanishes , or when the line SP coincides with the magnetic meridian .
And we can find the magnitude of SP by methods to be explained below .
Therefore we can find SQ and QP .
And if we assume HQ to be constant , and SH proportional to the earth 's vertical magnetic force at the given locality , we shall be able , by comparison of the results at different localities where the vertical force has different magnitudes , to discover the value of SH at each place and the value of HQ .
The value of QP however requires no combination of results found at different places , and is not liable to any uncertainty .
But without at present insisting on this separation of the subpermanent magnetism from the polar-magnet-force of induced magnetism , we can lay down the following rule:(3 ) .
The whole disturbance of the compass , whether the ship be wood-built or ironbuilt , will be represented by the sum of the effects of two forces , which separately would produce these two disturbances : one , a polar-magnet-deviation whose neutral point may be in any direction ; the other , a quadrantal deviation , which may be expected to be a positive quadrantal deviation , following the law of signs +-+-as depending on the quadrants of azimuth of the ship 's head .
And the whole disturbance will be very nearly ( but not exactly ) the algebraic sum of these two disturbances : the slight departure from that law will be the subject of examination below .
The practical problem , then , of analysing a given series of compass-deviations , is reduced to the dividing of them into two parts , of which one follows the law of polarmagnet-deviation , and the other follows the law of quadrantal deviation , subject to the trifling correction to which I have just alluded .
In the experiments with the 'Rainbow ' and 'Ironsides ' ( which are treated in my 56 paper in the Philosophical Transactions , 1839 ) , I was guided by experiments on hotizontal intensity as determined by vibrations ; but even then I found the computation of polar-magnet-deviation so troublesome , that I executed the calculations by graphical construction .
But in other cases , where there is no determination of horizontal intensity , the computation of polar-magnet-deviation would be very much more troublesome .
This consideration , together with the paucity of instances in which a comparison of the ship 's magnetism in different localities was possible , prevented me from entering further into the numerical calculations of ships ' magnetism .
But having lately received from Captain WASHINGTON , R.N. , Hydrographer to the Admiralty , the recolds of observations in several ships , which I desired to treat numerically ; I remarked that the trouble of calculation might be much diminished , and the process might be made perfectly direct and definitive , by the previous preparation of a Table of Polar-Magnet-Deviation ; and I proceeded therefore at once to compute the Table which is appended to this paper .
Of this Table I will now give a short description .
The table is a Table : of double-entry .
One of the arguments is the " Modulus , " which is the same as the proportion of AB to AC in fig. 1 .
It is given to every ?
01 from '00 to '80 .
The other argument is the " Apparent Azimuth of the Ship 's Head from the Neutral Position , " which is the same as the " apparent azimuth of the polar magnet " or " azimuth of the polar magnet as measured from the disturbed position of the compass-needle , " or the angle EBF in fig. 1 .
This is used as the algument of the Table , because , in the examination of the disturbance of ships ' compasses , it is usually most convenient to fix the ship in position by means of its own compass ; and in fact all the observations supplied to me have been made in positions of the ship so determined .
As the observations of deviation of ships ' compasses are usually made from " point " to " point " of azimuth , the division of the circle here employed is that by points and decimals of a point .
The Table is carried to 8 points only , as the polar-magnet-deviations from 8 to 16 points are the same in revelrsed order ; and those from 16 to 32 points are the same as those from 0 to 16 points with change of sign .
At the bottom of each column is the " Mean of all the Polar-MagnetDeviations for each value of the Modulus , " which is necessary for enabling us to determine the value of the modulus in any given case .
In ascertaining , from a given series of observed compass-deviations , the neutral position and modulus to be used in the application of this Table , it will be necessary to recognize the existence of a deviation following very nearly the law of quadrantal deviation , and the given numbers must therefore be so combined that quadrantal deviation will be ipsofacto eliminated .
This will be done by so arranging the process that the numbers for a whole semicircle of apparent azimuth will be added together algebraically .
This being understood , we may now proceed with advantage to investigate the nature of the terms produced by combining the effects of the polar-magnet-force and the quadrantal force .
Use the notation of the paper of 1839 , as far as it goes , and use also the following notation : In fig. 1 , let CA represent the terrestrial horizontal force multiplied by ( 1-M ) , M being a constant peculiar to each ship .
p the modulus , or the proportion of AB to CA .
It will be remarked that SP in fig. 3 is the representative of AB in fig. 1 .
a the angle QSP , fig. 3 , or the starboard angle made by the compound polarmagnet-force with the ship 's keel .
A the true eastern azimuth of the ship 's head .
A ' the eastern azimuth of the ship 's head as referred to the needle disturbed by polar-magnet-force only .
A " the eastern azimuth of the ship 's head as referred to the needle disturbed by polar-magnet-force and quadrantal force .
tA + the corresponding azimuths of the compound polar-magnet-force .
A+oS is Ai ^ Jthe same as the angle BAD in fig. 1 .
A ' the compass-deviation to the east produced by polar-magnet-force only =A-A ' = ( A+ ( A ' ) .
A " the additional deviation produced by the quadrantal force =A'-A"=(A'+a ) And the following double equation is accurate:sin A ' sin A+ a. .
sl.sn A'+ ?
/{1+ 2 cos A+ + } ' Then , neglecting MP only , the formulae in the paper of 1839 giveWhole force to north =I.cos.(1--M).(1 +f.cos A+a+P.cos2A ) Whole force to east =I.cos.(1 -M).(p^.sin A+os+P.sin 2A ) .
P..sinA ?
a+P.sin2A Therefore tanA'+A"= siA P 2A .
1 +-.csA+ P.cos+2A But tanA'sinA 1 +c.cosA+ a Therefore , retaining the complete multiplier of the first power of P , but no higher powers of P , tan A " , P. sin2A+ pP .
sinA-a 1+ 2p .
cosA +tc2 This quantity , however , is not that which we shall have occasion to use , for the following reason .
The polar-magnet-deviation which we shall take out from the Table is taken for an argument which is referred to the position of the compassneedle as disturbed by all causes ; it is therefore taken out , not for argument A'+os ( which would give us A exactly ) , but for argument A"+c- .
Let the quantity 58 thus taken from the Table be called A , , and the correction required be called A , , .
Then A , +A , , A'+A " ' ; and A , ,= ( A'A , ) +A"=(sin A'+c-sinAA"+- ) +Az " nearly =/ p(sin A"-+ a+A"-sin A"+a ) +A"=Al"(I +p .
cos A"+ , ) nearly .
If in the computation of this small quantity we reject powers of p following the first , , , =(1 --.cos A"+ ).P .
( sin 2A+p.sin A-oc ) .
But sin 2A=sin 2A"+2p .
cos 2A".sin A"+nearly .
After all reductions , A , ,=P .
sin 2A"/ +pP.cos 2A".sin A"+oa .
The first term of this expression is in the very convenient form of a quadrantal term referred to the apparent azimuth of the ship 's head .
The general influence of the second term is , that it produces no effect on the maxima of the quadrantal terms , that it slightly increases the polar-magnet-deviation when A"=0 or 180 ?
, and slightly diminishes it when A"=90 or 270 ?
; and this will be practically a sufficient description of its characteristic effects .
But as , in the aggregate of numbers , small terms become sensible which are scarcely sensible in the individual numbers , it will be desirable to ascertain the effect of this on a semicircular group .
Combine the term pP.cos 2A".sin A"+a with the approximate polar-magnet-deviation.sin A"+-o , and integrate from A"+a=0 to A " !
+S=s'r : the result is 2p ( 1-cos 2ca .
Without the small term we should have obtained 2p .
Hence it appears that the result for modulus found from semicircular groups , which .
* Expanding the last term , this quantity becomes , after all reductions , P.sin 2A"+PiP sin 3A"+a-sin A+3a ) It will scarcely be necessary to tabulate the small terms ; an estimate of their general effect can very well be formed in the mind .
The entire process will therefore be the following : 1 .
For the nautical terms N. , N.b.E. , N.N.E. , &amp;c. , use the numeral reckoning of points 0 , 1 , 2 , &amp;c. , as far as 31 , which will correspond to N.b.W. And for deviation E and deviation W , use the algebraical signs deviation + and deviation - .
It will always be convenient to place the + deviations and thedeviations in separate columns .
2 .
Clear the deviations of constant error by adding together all the + deviations , adding together all the deviations , combining them algebraically , taking the mean of the sum , and applying this mean with sign changed to every deviation .
The deviations thus corrected will be the base of all the following operations .
3 .
In writing down , in columns , the corrected deviations , repeat those from 0 to 15 points , in sequence to those from 0 to 31 points ; so that the Table contains forty-eight lines .
4 .
A conjecture will easily be formed as to the approximate value of the azimuth for the ' neutral position : " and then two or three neighbouring half-points are to be adopted for trial .
Thus , if the azimuth for neutral position appears to be near 3P or 4P , the positions to be tried may be 2P-5 , 3P'5 , 4P'5 .
5 .
The trial of these azimuths will be effected by dividing the series of observed deviations , not at these azimuths , but at azimuths distant from them 8 points on each side .
Thus , to make trial of the assumption 2P-5 , the observed deviations are to be divided at 26P'5 and 10P5 .
And the criterion will be given by adding algebraically all the deviations from 27P to 10P , both included ; a little accuracy will be gained if we also add in a separate sum all the deviations from 1lp to 26P , both included , and subtract this sum from the former .
It will be remarked that the quadrantal deviation is here eliminated .
6 .
If our assumption 2P'5 for the neutral position were strictly correct , the sum or difference of sums found in the manner just stated would =0 .
As this usually will not prove to be true , we must try the next assumption 3P-5 in like manner .
The comparison of the surms or differences of sums will give the correction to be applied to 2P'5 with very great accuracy .
The azimuth thus determined is strictly an " approximate neutral position , " and its supplement to 32P is the " approximate starboard angle .
" 7 .
The approximate neutral position being thus determined , the observed deviations are to be divided into two groups , one division being at the interval in which the neutral position falls , the other at the interval distant from it by 16P .
The algebraic 60 sums of the deviations in the two groups are to be taken : one is to be subtracted from the other , and the remainder is to be divided by 32 .
The quotient is the mean deviation .
This is to be compared with the Means of Polar-Magnet-Deviations at the foot of the columns of the Table .
Two adjacent Tabular Means being found , one greater and one less than the mean deviation just obtained , and the values of modulus .
corresponding to those two tabular means being noted , there is no difficulty in finding by interpolation the value of modulus corresponding to the mean deviation just- , obtained .
This is the " Approximate Modulus .
" 8 .
By use of the approximate neutral position , the angle of apparent azimuth from the neutral position will be formed for every observation .
Using this as the argument of Azimnuth in the Table .
the Polar-Magnet-Deviation is to be taken out for every observation with two tabular values of Modulus , one greater and one less than the approximate modulus just found .
Between these , the Polar-Magnet-Deviation will be interpolated for the approximate modulus ; and thus the Tabular PolarMagnet-Deviation corresponding to the Approxilnate Modulus will be obtained for every observation .
9 .
Subtracting this Tabular Polar-Magnet-Deviation algebraically from the Observed Deviation , the residual quantity will consist of Qtadrantal Deviation , of the small correction A , , and of errors of observation .
Neglecting the two last mentioned , a pretty accurate estimate of the coefficient of quadrantal deviation may be got by omitting the values for OP , 8P , 16P , 24P , and dividing the sum of each group of seven numbers by 5 ; the quotient will be the coefficient , or the Quadrantal Deviation for 4P , 12P , 20P , 28P .
The conversion of this coefficient into abstract number ( radius =1 ) gives the numerical coefficient P. 10 .
The angle 5X coefficient of quadrantal deviation X sine of twice the approximate starboard angle is to be subtracted from the approximate starboard angle to give the " True Starboard Angle .
" And the approximate modulus is to be multiplied by P 1+3X cosine of twice the approximate starboard angle to give the " True Modulus .
" 11 .
The Headward Modulus = True Modulus X cosine True Starboard Angle ; and the Startboard Modulus = True Modulus X sineTrue Starboalrd Angle .
As the modulus is the proportion of the disturbing force to the terrestrial horizontal force ( slightly diminished everywhere in the same proportion ) , we must , for the exhibition of the absolute values of the disturbing forces , multiply these quantities by the terrestrial horizontal force .
Then ( referring to the statements at the commencement of this paper for the results of theory ) we shall have , Headward Modulus X Terrestrial Horizontal Folce=H+N Terrestl ialVertical Force , Starboard Modulus X Terrlestrial Horizontal Force -S , where H and S are the forces of the ship 's subpermanent magnetism in the headward and starboard directions ; and N is a constant peculiar to the ship , depending 61 on the arrangement of the mass of iron , and having relation only to the ship 's capacity for induced magnetism , but in no way related to terrestrial magnetism .
If we can reconcile the observations made in the same ship at various localities by making 1I and S constant , then the subpermanent magnetism is truly permanent .
In any case , N must be constant for the same ship .
Perhaps the process of obtaining the various elements of a ship 's magnetism will be rendered a little more intelligible by exhibitingthe work in a single instance .
Iron-Steamer " Trident , " examined at Greenhithe , 1852 , September .
1 .
Deviations as registered , substituting only the 0 , 1,2 , 3 , &amp;c. points for N. , N.b.E. , N.N.E. , N.E.b.N. , &amp;c. , and the signs + and for E. and W. Apparent Deviation .
Apparent Deviation .
azimuth azimuth of ship 's ' , of ship 's i head .
+ head .
P00r_002 55 16 2 30 13 40 17 0029 15 18 2 30 3 12 32 19 5 50 4 15 50 20 9 10 5 17 40 21 11 52 6 19 30 22 14 35 7 19 5 23 16 47 8 18 40 24 19 09 17 15 20 22 10 15 50 26 21 45 11 14 10 27 22 10 12 12 30 28 20 0 13 11 30 29 16 30 14 9 20 30 13 '0 15 5 55 31 7 32 The sum of the + deviations is +205 ? ?
12 ' ; the sum of the deviations is -203 ? ?
58 ' ; the algebraical sum of all is +1 ? ?
14 ' ; which implies a mean error of +0 ? ?
2 ' .
Applying the correction -0 ? ?
2 ' to every deviation , the next table is formed .
62 &amp; 3 .
Deviations as corrected for constant error , Apparent Deviation .
Apparent Deviation .
azimuth azimuth of ship 's of ship 's head .
head .
+ ; pP o/ 0/ P/ 002 57 24 19 213 38 25 20 24 29 13 26 21 47 3 12 30 27 22 12 4 15 48 28 20 25 17 38 29 16 32 6 19 28 30 13 27 19 3 31 7 34 8 18 38 02 57 9 17 13 13 38 10 15 48 29 13 11 14 83 12 3)0 12 12 28 4 15 48 1 13 11 128 5 17 38 14 9 18 6 19 28 1-5 5 53 7 19 3 16 28 8 18 38 17 0297 13 18 2 32 10 15 48 19 i5 52 11 14 13 20 i 92 12 12 8 21 11 54 13 1 2B8 22 14 i 14 37 14 9 23 16 49 156 9 553 ' '~~ ~ ~~~~ ~ ~~ * } .
'* 63 4 .
The neutral position appears to be somewhere near IP , and therefore trial be made with the two assumptions OP*5 and 1P'5 .
5 .
For trial 0P'5 , the sums must be taken from 9P to 24P , and from 25P to 8p .
p.o 0o J01 Sum from 9 to 24=+ 88 4480 0=+ 8 44 Sum from 25 to 8=+-115 56-124 30-8 34 Excess of the first ... +17 18 may For trial 1P'5 , the sums must be taken from lOP to 25P , and from 26P to 9p .
pp0i00 Sum from 10 to 25 71 31-100 24= -28 53 Sum from 26 to 9=+-133 9-104 6-=+29 3 Excess of the first ... .
-57 56 6 .
A change of 1P in the assumption has changed the " excess " from + 17 ? ?
18 ' to -57 ? ?
56 ' , or has changed it by 75 ? ?
14 ' .
Hence , the assumption which would make the " Excess " --0 is 0r'5-]lp X 17 ? ?
18 ' the " Excess " =O is OP5+lPX750 14-05+ lpX0O23=OP'5+0p23=0P73 .
This is the " approximate neutral position " ; and its supplement or 31P'27 is the " approximate starboard angle , " or the approximate value of a. 7 .
The neutral position being OP-73 , the observed deviations which are to be grouped for ascertaining the approximate modulus are those from IP to 16P in one part , and those from 17P to OP in the other part .
The sums are respectively +204 ? ?
40 ' and -204 ? ?
30 ' ; the excess of the first is 409 ? ?
10 ' ; dividing by 32 , the quotient is 12 ? ?
47 ' .
On examining the " Mean of Polar-Magnet-Deviations " at the foot of the columns of the Table , it is found that the Mean corresponding to Modulus 0-34 is 12 ? ?
35 ' , and that corresponding to Modulus 0-35 is 12 ? ?
58 ' .
Hence the Modulus for mean 12 ? ?
47 ' is 0-3452 .
8 &amp; 9 .
The neutral position being 0OP73 , the " Apparent Inclination from Neutral Position " for azimuth of ship 's head OP will be 31P-27 ; and so for the others .
Then two series of numbers are interpolated from the Table of Polar-Magnet-Deviations for the fractional parts of the points ; one for Modulus 0'34 , and the other for Modulus 0'35 .
Between the pairs of corresponding numbers thus found , a third interpolation is made for the fiactional part of the Modulus , to 03452 ; and thus are obtained the Tabular Polar-Magnet-Deviations required .
Subtracting these from :1 .
I t , &gt; Apparent of Apparent iApparent Tabular polar-magnet-deviation .
Excess of azimuth of inclnaton corrected One-fifth of ship 's from observed quadrantal head .
neutral Modulus Modulus Modulus deviation .
group .
position .
0'34 .
0-35 .
0-3452 .
PPO0 3127 2 47 2 52 2 50 ( -0 7 ) 1 027 +1 2+ 1 4+1 3 +2 35 2 1-27 +4 49 +4 57 +4 53 +4 20 3 2-27 +8 26 +8 40 +8 33 +3 57 4 3'*7 +11 45 +12 6 +11 56 +3 52 +3 57 5 4-27 -+-14 38 +15 5 +14 52 -+246 6 5.27 +16 59 +17 31 +17 16 +2 12 7 6-27 +18 42 +19 16 +19 0 +0 38 7-27 +19 40 +20 16 +19 59 ( -1 21 ) 9 8127 +19 51 +20 27 +20 10 -2 57 10 927 +19 14 +19 50 +19 33 -3 45 11 10-27 +17 52 +18 24 +18 9 -4 1 12 11-27 +15 48 +16 16 +16 3 -3 35 -3 33 13 12-27 +13 9 +13 32 +13 21 -1 53 14 13-27 +10 0 +10 18 +10 9 -0 51 15 14-27 +6 31 +6 42 +6 37 -0 44 16 15-27 +2 47 +2 52 +2 50 ( -0 22 ) 17 16-27 121413 +1 1 18 17-27 4 49 4 57 4 53 +2 21 19 18-27 8 26 8 40 8 33 +2 41 20 19-27 -11 45 -12 6-11 56 +2 44 +3 19 21 20'27 -14 38 -15 5 -14 52 +2 58 22 21-27 -16 59 -17 31 -17 16 +2 39 23 22-27 -18 42 -19 16 -19 0 +2 11 24 23-27 -19 40 -20 16 -19 59 ( +0 57 ) x 25 24-27 -19 51 -20 27 -20 10 -0 14 26 25-27 -19 14 -19 50 -19 33 -2 14 27 26-27 -17 52 -18 24 -18 9 -4 3 28 27-27 -15 48 -16 16 -16 3 -3 59 -3 30 29 28-27 -13 9 -13 32 -13 21 -3 11 30 29-27 -10 0 -10 18 -10 9 -2 53 31 30-27 6 31 -6 42 6 37 -0 57 the corrected observed deviations , we have the excess , which clearly follows with considerable accuracy , a Positive Quadrantal law , perhaps slightly disturbed by the residual small telms in the manner explained above .
Its Mean Coefficient is +3 ? ?
35 ' or OP'3185 , whose equivalent in abstract number is +0-0626=P .
10 .
Frorn these numbers , 5 quadrantal coefficient X sine of twice approximate starboard angle =-OP'06 ; which subtracted from approximate starboard angle gives P True Starboard Angle =31P'33 .
And 3X cosine of twice approximate starboard angle =O0020 : adding this to unity and multiplying the Approximate Modulus , we have the True Modulus =0352 .
11 .
Expressing the magnitude of the terrestrial magnetic force in the manner introduced by GAUss for Absolute Measure , and adopting the English foot and English grain as the units of length and weight , the mteasure of terrestrial horizontal force at Greenhithe is 3'79 , and that of terrestrial vertical force is 9'66 .
Forming the quantities " True Modulus X cosine True Starboard Angle X Terrestrial Horizontal Force , " and " 'True Modulus X sine True Starboard Angle X Terrestrial Horizontal Force , " we have for the 'Trident ' at Greenhithe in 1852 , -1375=H+NX 9'663 -0'182=S ; and these results , with that just obtained for the Coefficient of Quadrantal Deviation , are the most advanced that can be obtained from the deviations of the compass in the 'Trident ' observed at Greenhithe only .
I shall now exhibit , in a tabular fortm , the results of the twenty-nine series of deviations which have reached me .
Nos. 1 to 13 are extracted from the work of the late Captain JOHNSON , R.N. , " On the Deviations of the Comtpass .
" The signs of the compass-deviations of the ' Erebus ' at St. Helena are changed , on the authority of Colonel SABINE , as conveyed to me by ARCHIBALD SMITH , Esq. Nos. 14 to 29 have been communicated to me by Captain WASHINGTON , R.N. , Hydrographer to the Admiralty .
It must be remarked that the first column , in every case , is the registered deviation as given by the observer ( a few numbers in brackets being supplied by interpolation ) , and not the deviation as cleared of mean error or index error .
In some cases this mean error is large ( thus with the 'Simoom ' at Simon 's Town it amounts to 1 ? ?
47 ' ) , and here it greatly modifies the true deviation , and even causes the original deviation to appear less on some points than the residual error .
The residual error is formted by computing the polar-magnet-deviation from the approximate elements at the top of the Table , and subtracting it from the deviation corrected for constant mean error only ; it therefore contains the quadrantal deviation , the small terms produced by combination of polar-magnet-deviation with quadrantal deviation , and the accidental errors of observation .
T'he coefficients of quadrantal deviation below are formed by omitting the residual errors for OP , 8P , 16P , 24P , and taking one-fifth part of the sums of the residual eI rrors in the groups between them ; and the mean coefficient is formed by changing the signs of the second and fourth coefficients of quadrautal deviation , and taking one-fourth of the algebraical sum .
The true elements are formed by correcting the approximate elements in the manner just explained .
shall first proceed to consider the quadrantal deviations in these ships .
It will be remembered that , according to theory , the coefficient of quadrantal deviation in each ship ought to be sensibly the same in all localities ; but that the coefficient in one ship may differ in any degree from that in another ship .
Coefficients of Quadrantal Deviation .
1 .
Wood-built sailing-ships .
In the Erebus , at In the Pandora , at Gillingham ( No. ) ... +0 23 Porto Praya ( No. 2 ) ... +0 28 St. Helena ( No. 3 ) ... .
+0 34 Cape of Good Hope ( No. 4 ) , +0 35 Kerguelen 's Land ( No. 5). .
+0 38 Plymouth ( No. 14 ) ... .
+0 38 Auckland ( No. 15). .
+0 30 In the Maeander , at .
Sheerness ( No. 16 ) ... .
--0 12 Simon 's Bay ( No. 17 ) ... -0 11 In the Spy , at. .
Sheerness ( No. 22). .
+0 46 St. Paul 's Loando ( No. 23)..-0 14 The accordance of the numbers in the three first instances is very remarkable ; and the more so because the compound polar magnetism has changed considerably .
In the Spy there is a discordance , but not important in nautical experiments .
It will be seen , on referring to the columns of Residual Error for the Spy , that there is general irregularity .
2 .
Wood-built steamers .
In the Virago , at .
In the Plumper , at Plymouth ( No. 18 ) .
Valparaiso ( No. 19 ) .
Portsmouth ( No. 20 ) St. Catherine 's ( No. 21 ) 0 1. .
+1 9. .
+1 44. .
+1 30 ... +1 42 The agreement is sufficiently close .
3 .
Iron-built steamers .
In the Bloodhound , at In the Jackal , at. .
Plymouth ( No. 6 ) .
Constantinople ( No. 7 ) Pireus ( No. 8). .
Plymouth ( No. 9). .
Lisbon ( No. 10). .
Piraeus ( No. 11). .
0 ; . .
+3 48 ... +327 ... +3 59 .
+4 14. .
+4 11 +3 32 li smA eliI I &amp; r , .
SHIPS .
In the Trident , at In the Vulcan , at .
In the Simoom , at Greenhithe ( No. 12 ) .
Malta ( No. 13). .
Greenhithe ( No. 24 ) .
Rio de Janeiro ( No. 25 ) Portsmouth ( No. 26 ) Simon 's Bay ( No. 27 ) Portsmouth ( No. 28 ) .
Simon 's Town ( No. 29 ) 0 ... +4 *. .
+3. .
+3 The general accordance here is extremely good .
The petty discordances appear to be purely accidental .
It is evident , at least , that they are not dependent on the geographical locality .
Thus at ( 8 ) the Piraeus produces the largest and at ( 11 ) the smallest in the group ; at ( 12 ) Greenhithe produces the largest and at ( 24 ) the smallest ; at ( 27 ) Simon 's Bay produces the larger and at ( 29 ) the smaller .
Nor have I been able to connect these differences with any other law .
Regarding their accidental character as established , they give a measure of the range of accident in these observations , and they show that that in the Spy , though large , is not excessive .
I think it therefore certain that this part of the theory is entirely supported by the observations ; and therefore that this quadrantal deviation may be perfectly neutralized in all localities by a mass of soft iron placed in the manner described at the beginning of this paper , leaving only a deviation which follows accurately in every place , separately considered , the laws of polar-lnagnet-deviation , and which therefore in every place , separately considered , may be neutralized by the application of permanent steel magnets .
I shall now proceed to consider the deductions from the two elements of polarmagnet-deviations namely the starboard angle of the polar-magnet-force and the modulus .
For this purpose , I premise the following elements of terrestrial magnetism .
The forces are expressed in GAUss 's method , adopting as units the English foot and the English grain .
For some of the elements I am indebted to the kindness of Colonel SABINE : others were obtained from other sources .
None were furnished to me precisely in the form in which they are herei exhibited , and some calculation therefore has been required to adapt them to my wants .
It is possible that they may be affected by trifling inaccuracies .
1 .
Gillingham ... 2 .
Porto Praya. .
3 .
St. Helena. .
4 .
Cape of Good Hope .
5 .
Kerguelen 's Land 6 .
Plymouth ... Horizontal Force. .
378. .
6'26 ... 5-97. .
4'56. .
3'88. .
* 3'82 L2 Vertical Force .
+ .
9-94 .
+ 6'38 .
-1-97 -6'08.10 ?
68 + .
+ 9-67 I 11 44 35 54 .
+3 36..+4 9 .
-4 31. .
+4 14 75 Horizontal Force .
Constantinople Pi raus. .
Plymouth. .
Lisbon. .
Piraeus. .
Greenhithe .
Malta. .
Plymouth .
Auckland. .
Sheerness .
Simon 's Bay Plymouth. .
Valparaiso. .
Ports mouth St. Cathetine 's Sheerness St. Paul 's Loando Gr , eenhithe .
Rio de Janeiro Portsmouth .
Simon 's Bay .
Portsmouth .
Simon 's Town ... 563 5'85 ? .
3'82.462 .
5'85.3'79..574 3'82 .
632 .
378 .
446..382 .
717. .
3*83 6*48 .
3'78 .
5-57 .
3 79.649.3-83 .
446 .
383.4*46 Vertical Force ... ..-+ 734 ... ++ 6-85 .
+ .
9-'67 ... -+ 9-19 ... + 685 ... + 9-66 ... + 7-03 +9 .
4 9-67 - ... .
-11'32 * ... + 9-67 ... -6119 ' ... + 9'67 ... 5*23. .
+ 9*63. .
, .
, r 2'54 ... + 9i67 ?
^ , .
-309 ... .
-4-9-63 ... .
--1'49 ... .
+9-63 ... .
, 6 19 ... .
-+ 963 6119 From these , with the starboard angle and the modulus , the resolved polar-magnetforce in the two directions of " headward " and " starboard " are formed by the rules given above .
It will be remembered that H is the ship 's subpermanent magnetism headward , S the subpermanent magnetism to the starboard side , and Na constant of capacity for induced magnetism peculiar to each ship .
As the total directive force 3'79 at Greenhithe is 3-79 , a change of disturbing force represented by 57- , or 066 nearly , would produce at Greenhithe a change of disturbance whose maximutm is 1 ?
nearly ; and so , mutatis mutandis , at other places .
1 .
2 .
3 .
4 .
5 .
1 .
WFood-built Sailing Ships .
The Erebus .
H+Nx 9'94=+0-261 S= +0'020 H+NX 6'38=+0'202 S=+0'037 H-NX 1-97=+0-043 S=+0-017 H--Nx 6-808=-0'090 S=+0-018 H -NX 10'68=--0'262 =S-+0-010 7 .
8 .
9 .
10 .
11 .
12 .
13 .
14 .
15 .
16 .
17 .
18 .
19 .
20 .
21 .
22 .
23 .
24 .
25 .
26 .
27 .
28 .
29 .
76 and S have small positive values , but so small that their combination would not produce , in the Thamnes , an error of half a degree .
N has a well-marked positive value .
In this ship the magnetism would be sensibly corrected , by placing ( by trial ) a mass of soft iron abaft the compass and at a lower level , in such a position as to correct the deviation with head east and head west , and then placing a mass of soft iron at the level of the compass , starboard or larboard , so as to correct the combination of the original quadrantal deviation and the new quadrantal deviation produced by the first corrector ; and this would be sensibly effective , without change , in all localities .
Or , permanent magnets might be applied to neutralize the errors , at the cardinal points , and soft iron at the level of the compass , starboard or larboard , to correct the quadrantal deviation ; the soft iron would then be effective in all latitudes , but one of the magnets would require alteration in different latitudes , and would require reversion in opposite hemispheres .
The Pandora .
14 .
H+NX 9-67=+0-166 S=+0'009 15 .
H--NX 1132= -0-272 S-+0-082 S has changed sufficiently to produce at Auckland ( No. 15 ) a disturbance whose maximum is 40 ' .
If this be neglected , the compass may be sensibly corrected in the same manner as for the Erebus .
H is small .
The Maeander .
16 .
H+Nx9'67=+0'070 S= +0-034 17 .
I--NX6'19=0-195 S= +0'008 H would appear to .
have a value of 0091 , which at Sheerness would produce a deviation of 1 ? ?
20 ' .
S is practically insensible .
This ship would require for perfect correction a weak magnet with its marked end towards the stern , in addition to the soft iron as in the Erebus and Pandora .
It will probably be better to use , for the polar-magnet-correction , a magnet alone , adjusting or reversing it as may be necessary .
The Spy .
22 .
H+Nx 967=-+0-097 S= -0'113 23 .
H--NX3'09=+0431 S=-0006 If the headward subpermanent magnetism has not changed , H is positive and N is negative .
These forces would require correctors in positions opposite to those of the Maeander .
S seems to have changed , to an amount which would produce at Sheerness a deviation of 1 ? ?
30 ' .
It will probably be best to correct the polar-magnetdeviation by acljustible magnets .
77 Wood-built Steamers .
The Virago .
18 .
H+N X 967=+0-493 S= +0-055 19 .
HNx 523= +0-029 S=+ 0079 S presents no sufficient evidence of change .
H and N are both positive : and the correction of polar-magnet-deviation would be made , either by a magnet with marked end towards the head , and soft iron as is described ' for the Erebus , or by an adjustible magnet .
The Plumper .
20 .
H+NX 9'63= +0'402 S= +0*069 21 .
H-N X 254=--+ 0301 S== +0'091 The remarks on the Virago apply also to the Plumper .
3 .
Iron-built Steamers .
The Bloodhound .
6 .
H+NX 967= +0'997 S=-0'153 7 .
H+NX 7'34= + 1043 S=---0'209 8 .
H+NX 6-85 =+0-931 S= -0220 The evidence for a distinct value of N is not very clear .
There seem to have been changes in the values of H and S ( subpermanent magnetism ) which on the whole would produce , at the stations Nos. 7 and 8 , a maximum deviation of nearly 1 ? .
The correction of polar-magnet-deviation should be madle by adjustible magnets .
The Jackal .
9 .
H+Nx9-67==+-1157 S==-0'123 10 .
H1+Nx9-19=1'005 S=0162 11 .
H+Nx 685-=+ 1-084 S=0'082 The remarks on the Bloodhound will nearly apply to the Jackal .
The changes in H and S would produce at the stations Nos. 10 and 11 a maximum deviation of a little more than 1 ? .
The Trident .
There seems to sensibly changed , 12 .
H+Nx9'66=+-1419 S=+0-080 13 .
H+NX 703 =+1-402 S--.-0.091 24 .
Hq+Nx9'66==+ 1323 S=-Q0-175 25 .
H-NX 1 49-=+1-418 S= -0198 be good reason for thinking that N is insensible , that H has not but that S has changed gradually , in the cor of sev eral years , 78 by a quantity which at station No. 25 would produce a maximum effect of 20 .
The change in the voyage between No. 24 and No. 25 is nearly insensible .
The Vulcan .
26 .
H+NX 963---0-603 S=-+0-046 27 .
H-Nx6-19=-=-1-236 S=+0-059 In this instance , as in some others , we feel greatly the want of observations after the ship 's return , to inform us whether the ship 's subpermanent magnetism has really undergone a change .
Assluming that it has not ( and it is certain that S has not sensibly changed ) , then H has a sensible negative value and Na sensible positive value : and the correction for all stations would be effected by magnets and masses of iron as has been described for the Maeander .
But it would probably be better to rely solely on adjustible magnets for the correction of the polar-magnet-disturbance .
The Simoom .
28 .
H+NX9-63-+1-364 S=-0'451 29 .
H-NX 619=+-1055 S=-0'184 The change in the value of S would produce at the station No. 29 a maximum error of about 3 ?
40 ' .
If the whole change in the compound headward force depended upon H , that change would also producce an error of nearly the same magnitude ; and the combination of the two would produce , as the total result of the change of subpermanent magnetism , an error of about 5 ? ?
30 ' .
But it is probable that N has some positive value , and that the change of H is not so great .
I shall now state what appear to be the just practical inferences from the preceding investigations .
1 .
At any place , the deviation of the compass in any ship , whether wood-built or iron-built , may be accurately represented as the effect of the combination of two forces , of which one alone would produce a disturbance following the law of polarmagnet-deviation , and the other alone would produce a disturbance following the law of quadrantal deviation .
2 .
Consequently , at any place the deviation of the compass may be accurately corrected by well-known mechanical mettlods ; namely by a magnet in the athwartship direction , fixed at a distance determined by trial , for correcting the deviation when the ship 's head is N. or S. ; by a magnet in the head-and-stern direction , also at a distance determined by trial , for correcting the deviation when the ship 's head is E. or W. ; and by a mass of unmagnetized iron , at the same level as the compass , in the athwart-ship line or in the head-and-stern line according to ciicumstances ( usually in the former ) , also at a distance determined by trial , for correcting the deviation when the ship 's head is N.E. , S.E. , S.W. or N.W. 79 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 magnetism 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 eriro 5 ?
or 6 ? .
This is neally 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 sliip , to exatine 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 irecording the amount 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 magnetismr which is similar to polarmagnetism but which changes in different magnetic latitudes ) , they may be precisely corrected by readjusting the position of the magnets , leaving the unmagnetized iron undisturbed .
And the change ( if there is any ) in the intensity of the correcting magnets will also be co rrected , as to its effect on the compass , by the same readjustlnent 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 diffelrent 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 magnetismn is found to be unaltered ( which affords presumlption that it has been unaltered during the voyage ) , anti if the elements of magnetismr 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 b)y 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 remaining part by unmagnetized iron , as is described above for the Erebus ; and then the correction would be complete in all parts of the earth .
9 .
But , practically , it will perhaps always be easier and safer to readjust the positions of the magnets ( as in art .
6 .
) whenever the directions of one of the magnetic points N. and S. , and one of the magnetic points E. and W. , can be truly ascertained .
This can always be done in harbour , in a very short time .
Probably this can also be done at sea , in fine weather , by reference to a compass carried high up the ships ' masts .
It can also be done with the aid of astronomical observations and of a knowledge of the local " variation " or " declination .
" In all cases , the mere adjustment of the magnets is an extremely rapid process .
10 .
On reviewing the results of the preceding examinations , I think that I am justified in denouncing any system of navigating a ship by forming a table of compassdeviations at the starting port , and using that table until means of correction can be obtained from observations , as dangerous ; and I think that it ought to be at once discontinued .
It does not in the smallest degree provide against the effects of possible change in the ship 's subpermanent magnetism during the interval in which no observations are obtained ( which , with sometimes a minute change in the powers of the magnets , is the only risk to which the method of mechanical correction is liable ) ; and , as it does not recognize the effect of the variation in the magnitude of terrestrial horizontal magnetism at different places ( which alters the compassdeviation by changing the proportion of the ship 's subpermanent magnetism to the terrestrial horizontal magnetism , upon which proportion the compass-deviation depends ) , it gratuitously introduces a class of errors which are entirely avoided by correcting the compass by magnets and soft iron .
Thus , in the instance of the Trident ( 24 ) and ( 25 ) , sailing from Greenhithe to Rio Janeiro : suppose that there had been no good opportunity of making observations of azimuth on the voyage ; on the ship 's arrival at Rio Janeiro , the table of deviations formed at Greenhithe would have been found erroneous by 60 or 7 ?
in one direction with head eastward , and erroneous by 8 ?
or 9 ?
in the opposite direction with head westward .
But if the compass had been corrected by magnets and soft iron at Greenhithe , it would have been correct at Rio Janeiro without an error approaching to a single degree .
The change of compass-deviation , in fact , has been produced , not by the change of the ship 's subpermanent magnetism ( which has been sensibly constant ) , but by the change in the magnitude of the earth 's directive magnetism , which change has altered the proportion of the ship 's invariable magnetism to the earth 's variable magnetism ; and if this proportion had been reduced to zero by neutralization of the ship 's magnetism by means of magnets , the variation of the proportion as depending on the variation of the earth 's magnetism would also have been destroyed .
What has been said in regard to the errors arising during the whole voyage , applies , in a proportionate degree , to the errors arising during a part of the voyage : if there had been valid observations after making half the voyage , the errors perhaps would have been only half as great ; but these errors would have been equally gratuitous .
In other instances , such as that of the Simoom , in which the change of subpermanent magnetism is real and unusually great , the tabular method ( supposing , for illustration , that there had been no opporltunity of sufficiently investigating the errors during the whole voyage ) would have united the gratuitous errors with the errors produced by the real change , and would have produced at the Cape of Good Hope an error of 110 ; whereas , if the correction by magnets had been used , the error would have been under 6 ? .
At intermediate places , as the neighbourhood of St. Helena , where the earth 's directive force differs still more from that in England , the gratuitous error would have been much greater , and the error really depending on change in the ship would probably have been less , as occurring in an earlier part of the voyage .
The mere comparison of magnitudes of errors , however , in this way , does not sufficiently exhibit the disadvantage of the method of " Tables of Deviations .
" It is an important defect that no good new table can be formed , without observations for the error on numerous points of azirnuth ; whereas the operations for readjustment of magnets require observations on only two points of azimuth .
And , I apprehend , that the necessity of using a table at all ( that is , of steering by one nominal course when another course is intended ) is , especially in difficult channels , a very serious evil , from which the method of steering by a corrected compass is entirely free .
11 .
I have alluded above to the possible changes in the energy of the correcting magnets ; but I am bound to state that these changes ( when ordinary care is taken for the conservation of the magnets ) are , to the best of my knowledge , extremely minute .
It is known , as a matter of experience , that the diminution of the subpermanent magnetism of a new iron ship , though small , is usually greater than that of the magnets ; inasmuch as it usually becomes necessary to increase the distance of the magnets from the compass .
I subjoin the Table of Polar Magnet Deviations , which has been used in the preceding investigations , and which may perhaps be useful , in future , for similar investigations .
8 2

