DECLINATION. 
allowing a proper variation upon them, and put me mendes 
the neceffity of endeavouring to find out fome method 
recting or allowing for thefe This tenes 
then relates cons inefieet 
e degrees, and was at lait in 
duced to draw the area ne inferences. 
1, #6 uppo ole,” he fays, ‘ the attractive power of the 
different bodies in the fhip, which are capable of affecting 
the compafs, to be collected into oe likea foes) se 
or centre : & ravity, and t this point is nearly in 
centre of the thip where the fhot are - epo ofited, for here i 
ogether 
greate’ quantity of iron is colleéted 
‘¢ T fuppofe this point to be endued with “the fame kind 
ebatec cn as the pole of the hemifphere where the fhip is ; 
confequently, in New Holland the fouth ends of the needle 
would attracted by it, and the north end repelled.” 
3." at the attractive power of this point is fufficiently 
fro in a fhip of war to interfere with the action of the mag- 
netic poles up mpafs placed upon or in the binnacle.’’ 
With inftruments upon land. a much grea 
this accuracy that the daily or hourly bela has been dif- 
covered ; forthe ene of it from bour to hour is exceed. 
This acc 
ion and the indications of the 
have an influence 
needle hefe particulars Hel ore are; rit. that the 
agnetic oe oles me _ nee dle, be exa al in 
the axis of its fizure. But ast s feldom the cafe, t 
only expedient by which the error » ocifng therefrom ma be 
afcertained ‘and allowed for, is to turn the needle with the 
rue, oint 
magnetic ax needle from nee of its igus ba 
when once ae ee may afterwards be allow 
out any farther inverfions of the needle... 2dly. 
ridian line muft be had, in order to compare the direétion of 
the needle with it. See Meripian Line. jdly. Care muft 
be taken that the needle be not within the influence hd bel 
mafs of-iron or other ferruginous matter that may 
tained in the ae in ea Sei or about the perion of 
o better m 
1, The hea te in itis houfe, at the time of obfervation, 
was placed in the ae window, on the fouth fide of the 
fociety’s meetin up mahogany board. 
Againft the esi: building the dial-plate of a watch is 
fixed, making an a with the true meridian 
8 
pafs was adjufted. th 
ith the true meridian, I fixed a tranfit-initrument on the 
mahogany board sen air ete precifely in the fame 
lace where the compafs had laced, and having ad- 
jufted its telefcope to the faid ae the trantits of the fon. 
and flars over a vertical circle asa through oe bw ~ 
ie mark, were obferved ; and the angle conta aan 
the faid mark and the true aedan. Was ae — 
tation to be 31° 8/.8 as above. 
2. © For “the pure rpofe of afcertaining what error there 
might be, from a want of parallelifm between the line j joining 
the indices and the magnetifm of the needle, and thereby to 
determine whether, in the ufual metho obferving, the 
indices fhew the true angle which the direétion of magnetifm 
pa with the ie divifion or zero, a great many obferva.- 
tions were made on both ends of the needle, and with both 
fides of the qceaie uppermat (the cap of the needle being 
made to fit on readily on either face for this purpofe ); vis. 
north end and fouth end in its upright pofition, and north 
cen 
end and fouth end with the needle inverted, and the mean 
of the four giving the angle greater by 2’, than that fhewn 
the north end in the upright pofiti f the needle, 
y n 
aie was the end nih ted | in . hofe able tvations,) two 
minutes eee been added to all the obfervations read from 
the initrument, as the eareaci for this erro angles on - 
the eaft fide of zero, and fubtracted from isc on a weit 
fide, to obtain the true angle ; which error te augles on the 
welt fide, however, only occurred, when the inftrument was 
taken out of doors to-determine the effect of the iron-work 
of the building. 
variation compafs being placed in the ae for 
oblervation pak not be fuppoled to be entirely of the 
influe a3 was, therefore, defirous to pe rtain . 
how ae oe ee might extend; for the determination . 
of which, the following method was adopted. 
« Having caufed to be funk into the earth to fome depth 
a ftrong poft, in the wood-yard of Somerfet- houfe, ata. 
confiderable diftance from the influence of any iron, 
there at thofe tim 
tionary ; viz. morning and afternoon e compafs 
was carried out of doors, obfervations were made in the 
; then it wags taken out of doors to the above-men- 
paed ftation for dieceen there; and the Geeeations 
r 
occurred between the cba made in he houfe before, 
ane after, thofe taken e yard, 
«© The obfervaticns, ae efore, made in the yard, 
pared with thofe taken in-the ont. both before and ae 
thofe taken out of it, formed the comparifon for obtaining 
the error, or the effe&t of the iron-werk of the room on the 
needle in the houfe ; and there is reaton to believe that con- 
fiderable accuracy has been obtained. They are as fal. 
lows : 
“ By a mean of 20 — or 200 obfervations taken with 
the compafs in the yard, c red with twice that number 
taken in the houfe, before and ae thofe oo in the aie 
in e was eater 
the variation obferved 
n 
II. The obfervations of the quantity of magnetical decli-< . 
nation which have been made and regiftered in journals and - 
other 
