FROM ORDINARY DAYS OF THE ELEVEN YEARS 1890 TO 1900. 
191 
sunspot frequency, 1893, though showing very large regular diurnal variations, was 
decidedly quieter than the adjacent years. 
Table I.— Number of Days not included in “ Ordinary.” 
Element. 
1890. 
1891. 
1892. 
1893. 
1894. 
1895. 
1896. 
1897. 
1898. 
1899. 
1900.' 11 
- years. 
D 
7 
26 
31 
17 
21 
21 
41 
16 
19 
21 
.8 I 228 
H 
6 
25 
36 
14 
26 
29 
45 
19 
21 
22 
9 252 
V 
14 
48 
53 
36 
36 
39 
54 
26 
28 
21 
11 1 366 
While 1892 and 1894 contained most of the outstanding magnetic storms, 1896 
was remarkable for the persistence of disturbed conditions. A point to be 
remembered is that it is easier to recognise the general trend of the regular diurnal 
variation when the range is large than when it is small. Thus a disturbance 
sufficient to mask the regular diurnal variation when least— i.e., at midwinter, in 
sunspot minimum—might prove no serious obstacle to smoothing curves at mid¬ 
summer near sunspot maximum. Making all due allowance for the increased 
amplitude of the regular diurnal variation in summer, Table II. shows clearly that 
the annual variation of disturbance has a well marked double period, with minima 
at midwinter and midsummer, and maxima in March (or late February) and October. 
The maximum in Spring is the more prominent of the two. 
Table II.—Number of Days not included in “ Ordinary.” 
Element. 
January. 
•Tamuqaq; 
March. 
April. 
>> 
krH 
J une. 
r-q 
August. 
1 
September. 
October. 
November. 
CJ ! 
CD 
O 
<D 
rH 
D 
22 
26 
32 
17 
18 
7 
12 
11 
19 
30 
20 
14 
H 
26 
26 
34 
19 
18 
8 
13 
18 
21 
29 
24 
16 
V 
40 
42 
45 
31 
28 
19 
23 
25 
30 
37 
26 
20 
The basis on which the curves were treated was largely determined by the fact 
that the D, H and V curves had to be considered at different times. If a careful 
consideration of the curves from all three elements had preceded all measurements, 
a common selection of ordinary days would probably have been made. As it is, 
a little more disturbance enters into the ordinary day H inequality than into that 
for V, and a little more into the D inequality than into that for H. This is most to be 
regretted perhaps in relation to the derived inequalities. The north component (N) 
and west component (W) inequalities depend on both D and IT, while the total force 
(T) and inclination (I) inequalities depend on H and . Lhese inequalities, not 
