AT KEW OBSERVATORY, 1890 TO 1900. 
225 
Absolute Ranges. 
§ 18. The regular diurnal inequality is of special interest for theorists, particularly 
those whose ambition it is to discover an explanation of the phenomena of terrestrial 
magnetism. There are, however, other facts connected with the daily changes which 
merit a close study. If we take the case of a surveyor or explorer making frequent 
use of an accurate compass, it is, no doubt, well that he should know the general 
features of the regular diurnal inequality ; but what immediately concerns him is 
how far the needle is pointing from its mean position at a given hour of a given day. 
Now, no examination of past data can supply exact prophetic information as to the 
future, but the present investigation has provided data which will, I trust, enable the 
practical man to see more exactly how the matter stands, and will give him a more 
adequate idea of the risks he runs in accepting individual observational data as 
representative. The statistical data which seem most likely to serve this purpose 
consist of the mean values of the absolute daily range (absolute maximum less 
absolute minimum) for different months and years, the relative frequency of absolute 
ranges of specified size, and the frequency with which the absolute maximum and 
minimum fall at different hours of the day. 
§ 19. Table XIY. gives the average value of the absolute daily range for each month 
of the eleven years as derived from all days disturbed and undisturbed, with the 
means thence derived for each separate year, and for the twelve months of the 
average year. The two last rows contain corresponding monthly means derived 
respectively from all ordinary days and from the Astronomer Royal’s quiet days. The 
two last columns contain yearly means from ordinary days and from quiet days. It 
is interesting to compare the mean absolute ranges in the second last line of Table 
XIV. with the mean inequality ranges in the last line of Table VI., these tables both 
applying to individual months and depending on the observations of exactly the same 
days. Taking the ratio borne by the absolute to the inequality range, we obtain the 
following results :—- 
j January. 
ci 
^a3 
March. 
April. 
. 
c5 
i-—1 
June. 
August. 
September. 
October. 
<D 
s 
O 
> 
o 
December. 
2-01 
00 
r-H 
1 -55 
1-30 
1-29 
1-21 
1-26 
1-24 
1-43 
1-67 
1-90 
1-97 
The arithmetic mean of these twelve values of the ratio is 1'48. So large an excess 
in the absolute over the inequality range in ordinary days would hardly, I think, have 
been anticipated. Relatively considered, the excess of the absolute range is much 
greater in winter than in summer. The extreme values actually found for the ratio 
VOL. CCVIII.-A. 2 G 
