268 
DR. C. CHREE: DISCUSSION OF KEW MAGNETIC DATA 
exposed to undue uncertainty owing to the large temperature coefficient of the 
magnet and the absence of a temperature correction. The uncorrected temperature 
effect on R/ is of course not always in one direction, but amongst the least ranges of 
the month the chances are that some will be reduced by the temperature effect. Thus 
the natural consequence of the absence of a temperature correction will be to give too 
low a value for the monthly minimum range. The H ranges are not wholly free 
from this uncertainty, but the temperature coefficient of the H magnet is only about 
a seventh of that of the V magnet, while the average range in H is fully 80 per cent, 
in excess of that in V, thus the uncertainty is of quite a different order. 
Table XLVII.—Horizontal Force. Absolute Daily Ranges. 
Smallest Values (Unit ly). 
Year. 
January. 
February. 
March. 
April. 
>> 
June. 
13 
+3 
co 
3 
'op 
< 
September. 
October. 
November. 
December. 
Year. 
1890 
16 
18 
17 
31 
25 < 
26 
31 
24 
23 
25 
16 
13 
13 
1891 
12 
17 
18 
38 
28 
35 
39 
41 
40 
29 
29 
13 
12 
1892 
26 
36 
37 
30 
41 
44 
46 
45 
36 
38 
19 
17 
17 
1893 
20 
20 
31 
52 
40 
45 
42 
50 
38 
36 
30 
19 
19 
1894 
21 
30 
25 
36 
46 
60 
50 
49 
31 
36 
22 
19 
19 
1895 
16 
29 
34 
50 
45 
52 
44 
27 
33 
30 
21 
16 
16 
1896 
24 
29 
23 
43 
35 
35 
29 
33 
48 
23 
15 
12 
12' 
1897 
13 
15 
24 
34 
36 
33 
31 
28 
27 
21 
18 
12 
12 
1898 
10 
13 
29 
16 
23 
33 
28 
25 
33 
23 
18 
11 
10 
1899 
11 
16 
22 
30 
31 
27 
22 
32 
37 
16 
13 
9 
9 
1900 
12 
12 
21 
25 
26 
29 
23 
22 
22 
14 
11 
7 
7 
Whole 1 
period J 
10 
12 
17 
16 
23 
26 
22 
22 
22 
14 
11 
7 
i 
The D range is entirely free from this source of uncertainty, so far as is known ; it 
has the further advantage that no error can arise from an unrecognised variation in 
the scale value. There is, however, one slight drawback attending it, viz., that the 
force required to alter D by 1' alters if H alters. Strictly speaking, the force 
equivalent to a change of 1' in D is different at different times of a highly disturbed 
day. 
The slight increase in accuracy that would result from allowing for changes in H 
in the course of a single day would be no adequate return for the labour involved. 
Apart from any disturbance effect, the force equivalent to a change of U in D is 
influenced by secular change in H. The mean value of H in 1900 exceeded that in 
