FROM SELECTED DAYS DURING THE SEVEN YEARS 1898 TO 1904. 
311 
that of c u the values for the winter months being conspicuously greater than those 
for the summer months. There is also a fairly systematic annual variation in the 
phase angle a 3 ; but this angle is largest at the season when is least, and con¬ 
versely. The 6-hour term shows an annual variation similar on the whole to that 
exhibited by the 12-hour term. In equinox' and summer the 6-hour wave seems 
fully larger than the 8-hour wave. 
Table Y.—Diurnal Inequality. Amplitudes and Phase Angles. 
C\. 
Oil. 
Co. 
Oi 
O- 
a 3 . 
Ci. 
a 4 . 
January . . 
22 
•33 
o 
206 
/ 
5 
21 
•50 
o 
169 
/ 
42 
6 
•82 
o 
10 
t 
36 
1 
•87 
o 
234 
/ 
39 
February. . 
22 
•32 
203 
37 
32 
•72 
170 
55 
10 
•31 
9 
6 
9 
mJ 
■64 
224 
36 
March. 
16 
61 
123 
8 
33 
•54 
185 
31 
4 
•68 
35 
53 
5 
•11 
307 
8 
April . . . 
12 
79 
71 
39 
30 
•57 
192 
33 
4 
•78 
96 
17 
6 
•28 
313 
55 
May . . . 
18 
21 
85 
57 
22 
•40 
188 
25 
3 
•14 
100 
22 
3 
•69 
313 
31 
June . . . 
8 
98 
78 
52 
22 
•57 
182 
45 
1 
•32 
124 
59 
1 
•40 
276 
49 
July . . . 
6 
12 
48 
13 
23 
77 
185 
8 
3 
•40 
124 
29 
4 
•44 
292 
39 
August . 
5 
96 
142 
14 
26 
30 
181 
30 
1 
50 
106 
39 
2 
■70 
312 
45 
September . 
9 
50 
153 
51 
22 
63 
198 
45 
2 
93 
16 
15 
3 
51 
329 
46 
October . . 
6 
58 
191 
52 
29 
60 
205 
46 
6 
35 
18 
21 
3 
25 
288 
23 
November . 
14 
14 
201 
40 
16 
98 
211 
53 
5 
43 
38 
17 
1 
56 
237 
34 
December 
29 
69 
208 
24 
21 
09 
175 
12 
7 
11 
35 
44 
2 
89 
249 
3 
Year . . . 
8- 
40 
165 
4 
24 
75 
186 
50 
3 
82 
39 
29 
2 
80 
292 
38 
Winter . . 
22- 
10 
205 
32 
22 
23 
178 
50 
7 
21 
21 
8 
9 
21 
236 
50 
32 
Equinox . . 
8- 
78 
124 
43 
28- 
83 
195 
5 
3 
98 
40 
26 
4 
42 
310 
Summer . 
8- 
82 
86 
7 
23- 
74 
184 
20 
2 
29 
113 
36 
2 
98 
301 
38 
Annual Vaviation. Fourier Series. 
§ 9. The annual variation may be represented by a series of the type 
M + Pi sin (t + di) +P 2 sin (2 1 + 0 2 ) + ..., 
where M is the arithmetic mean of the 12 monthly values, t is the time counted from 
the beginning of the year, 1 month being taken as equivalent to 30°, while P 4 , P 3 ,..., 
0 U 0 2 ,... are constants. 
Table YI. gives for a variety of elements the values of the amplitudes and phase 
angles of the yearly and half-yearly terms, and the ratios borne by these amplitudes 
to the mean value of the element and to one another. As before, in the amplitudes 
unity represents 1 volt per metre. The angles are given only to the nearest degree. 
The representation of the annual variation by two terms of the series is, as a 
matter of fact, not very close in any one of the elements. This may mean that other 
shorter period terms are by no means negligible, or merely that a larger number of 
year’s data are necessary for high precision. 
