98 
LIEUT.-COLONEL SABINE ON TERRESTRIAL MAGNETISM. 
the observations at N.E., N.W., S.E. and S.W. give Z>=+’984; those at E.N.E., 
W.N.W., E.S.E. and W.S.W., b —' 984 ; and those at E. and W. b=- 982. 
In the case of c, we have from equation (3.), 
-^-sin0'= ccos0cos^+c?sin0 ; 
from the observations at N. to N.E. and N.W. inclusive, and from S. to S.E. and 
S.W. inclusive, eliminating d, we have 
c= +’009. 
The constant d is perhaps the most difficult of the constants to ascertain satisfac- 
torily, as its value derivable from the observations depends on a knowledge of the true 
geographical dip at the place of observation, free from what is now known as station 
error. Experience has fully shown the general fact, that inclinations observed on land 
cannot safely be assumed as free from local disturbance. The discrepancies of gravi- 
tation at the Falkland Islands are well known from the experiments with the pendu- 
lum ; and from the geological character of these islands, we might be prepared to expect 
the existence of magnetic discrepancies also. By the needles in both ships, the incli- 
nation was found a third of a degree higher at the magnetic observatory on shore 
than when observed on board in the harbour; if the observatory dip were to be 
assumed as an undisturbed one, we should obtain d in both ships considerably less 
than unity, whereas from the comparison of the observations in both ships in the 
preceding December and January, with the inclination observed at the same time 
on the ice over a deep sea, where no local attraction can be imagined to exist, we 
have d (as far as the small differences of geographical position will permit us to judge) 
differing scarcely, if at all, from unity in either ship. The preference is certainly 
due to the deduction from the results obtained on the ice. Taking therefore d— 1 , 
c=+’01, b - :’984 and a=+’026, we have the corrections, and the corrected incli- 
nation, of the observations in the Terror between the 6th and 16th of January as 
follows : 
Ship’s head. 
No. of 
observations. 
Inclination 
observed. 
Correction. 
Corrected Inclination. 
N. 
4 
—81 19’5 
° ‘ n 
+ 1 26 
— 79 53-51 
N.W. 
N.E. 
4 
-80 58-5 
+ 1 09 
-79 49-5 
w. i S. 
E. \ S. 
w. by s. 
E. by s. 
9 
-79 58-5 
-79 49-0 
+ 0 12 
— 0 01 
-79 46-5 
-79 50-0 
>-79 52-4 
w.s.w. 
E.S.E. 
' 
4 
-79 29’0 
lO 
O 
1 
-79 56-5 
S.W. 
S.E. 
6 
-79 04-5 
— 0 51-5 
-79 56-0 
s. 
6 
— 78 30-0 
— 1 24-5 
-79 54-5 J 
Slight differences in the corrected results must be looked for, as the observations 
were not all taken precisely at the same geographical spot : those which appear in 
the table are, however, very slight ; the accord produced by the corrections seems as 
