34 
LIEUT.-COLONEL SABINE ON TERRESTRIAL MAGNETISM. 
Declination . — I have collected in the subjoined Table all the recorded observations 
of the declination at Matavai Bay with which I am acquainted, from the earliest dis- 
covery of the island to the present time. 
Table XV. — Declination observed at Matavai Bay, Otaheite. 
Year. 
Month. 
Observer. 
Declination. 
(East.) 
Year. 
Month. 
Observer. 
Declination. 
(East.) 
1765 
1767 
1769 
1773 
1774 
1777 
June 
July 
Byron 
Wallis 
5 00 
5 36 
4 46 
5 40 
5 46 
5 34 
1792 
1823 
1824 
1826 
1835 
1840 
January .... 
May 
Vancouver 
Duperrey 
6 12 
6 40 
6 50 
7 33 
7 34 
6 30 
June 
August 
April and May 
December . . 
Cook 
Wales and Bayley 
Wales and Bayley 
Cook 
March 
April 
November . . 
May 
Kotzebue 
Beechey 
FitzRoy 
Belcher 
A first glance at these observations shows that the easterly declination has been in- 
creasing at Otaheite from the time of the first discovery of that island. It is scarcely 
probable that the progression has been strictly uniform throughout the whole period, 
but the deficiency of determinations in the years that form the middle portion of 
the interval, renders the data that we possess unsuitable for deducing the variation in 
the rate of the secular change ; and we must be content with that approximate repre- 
sentation which may be given by an uniform rate. Assuming, therefore, the change 
of declination to be proportional to the time, I have computed by the method of least 
squares from the data contained in the Table, the following formula for the declina- 
tion h at Otaheite : 
§ = 6° IT-85 + T-65 Qt, 
t being the interval of time elapsed since January 1 , 1800, expressed in terms of a 
year. 
The declinations computed according to this formula, and the differences from the 
observed declinations, are as follows : — 
Computed. 
Differences. 
Computed. 
Differences. 
1765 . 
/ O 
. . 5 15 
+ 15 
1792 . 
o / 
. . 5 59 
- 13 
1767 . 
. . 5 18 
- 18 
1823 . 
. . 6 51 
+ 11 
1769 . 
. . 5 21 
+ 35 
1824 . 
. . 6 52 
+ 2 
1773 . 
. . 5 28 
+ 12 
1826 . 
. . 6 55 
- 38 
1774 . 
. . 5 29 
- 17 
1835 . 
. . 7 12 
— 22 
1 777 • 
. . 5 35 
+ 1 
1840 . 
. . 7 19 
+ 49 
It will be seen that the discordances with 
each other of 
the observations of recent 
date are as great, and even greater, than those of the earlier observers ; which ought 
to be an indication that the larger discrepances are occasioned rather by local 
disturbing influences than by errors of observation. The probable error of a single 
determination, as resulting from the tabulated differences, would be about fifteen 
minutes. 
