24 Mr. Christie on the effects of temperature on 
the preceding results I first took those values of A derived 
from observations where the mean of the temperatures was 
near to 65®, and from these obtained a mean value of A . 
when the mean temperature of the magnets was nearly 65® : 
in like manner I obtained a mean value of A . ^ when the 
mean temperature of the magnets was nearly 60°, and also 
when nearly 54° : taking a mean of these three means, I oh- 
tained a value of A . which could not be sensibly different 
from its true value in any of the observations which were to 
be corrected. These are collected in the following table. 
Table of Results for obtaining the Mean value of Is.. — . 
0 
Mean Temperature 65 nearly. 
0 
Mean Temperature 60 nearly. 
0 
Mean Temparature 54 nearly. 
Date. 
Temperatures. 
Mean. 
Value of 
F 
^ M 
Date. 
Temperatures. 
Mean. 
Value of 
A.X. 
M 
Date. 
Temperatures. 
Mean. 
Value of 
F 
A . —. 
M 
June 
6 
13 
*7 
17 
18 
18 
18 
19 
*9 
0 
62.05 
61 .08 
60.25 
61.75 
56.90 
55.40 
55.40 
55-45 
55-55 
0 
67.15 
71.05 
68.25 
73-80 
73.60 
73-80 
73-80 
73.90 
73-90 
0 
64.600 
66.065 
64.250 
67-775 
65.250 
64.600 
64 . 600 
64-675 
64-725 
.1233 
-1430 
-1391 
.1260 
•1314 
.1237 
.1251 
.1360 
.1311 
June 
6 
17 
18 
19 
19 
19 
59.05 
49-30 
55-58 
55-45 
55-55 
55-15 
62.05 
68.25 
66.00 
64.30 
64. 28 
66.75 
0 
60.550 
58- 775 
60.790 
59- 875 
59-915 
60.950 
. 1268 
.1263 
-1315 
-1299 
.1220 
.1224 
June 
17 
20 
20 
20 
22 
22 
0 
49-30 
50.65 
51-35 
51-35 
51-35 
51-35 
0 
60.25 
55.80 
55.80 
57-20 
54.08 
55-94 
54-775 
53-225 
53 - 575 
54- 275 
52- 715 
53- 645 
.1179 
- 1063 
-1059 
.1125 
-1203 
. 1064 
Mean 
60. 143 
.12648 
Mean 
53-701 
.11155 
F 
Mean value of A . to be applied in correcting the observations i 
^ Jo. 12300 
on the Diurnal changes in the positions of the Points of Equilibrium ) 
Mean 
65-171 
-13097 
F 
To apply the value of A . thus determined, to the cor- 
rection of the observed directions of the needle, for the 
changes which took place in the temperature of the magnets, 
let A (p be the increment of the azimuth of the needle corres- 
