328 
Proceedings of the Koyal Society of Edinburgh. [Sess. 
by the 24-equidistant ordinate method. The values of the first four ampli- 
tudes and phases are given in Table III. In the calculation of these series 
the time was measured from midnight mean local time, the other times 
used in the paper being Greenwich mean time. The mean local time at 
Edinburgh is thirteen minutes behind Greenwich mean time. 
Table III. 
a 0 . 
<h- 
a 2 . 
CL§. 
u 4 . 
“i- 
«2- 
03" 
«4- 
Ui/n 2 . 
a x /aJ 
a 2 /a 0 . 
January . 
255 
71-3 
11-9 
10-6 
6-1 
219-7 
102-6 
140-8 
324-6 
5-99 
•280 
•047 
February . 
183 
52-6 
24-2 
9-2 
12-6 
221-9 
110-4 
356-7 
119-6 
2-17 
•287 
•132 
March 
124 
20 6 
19-6 
71 
2°8 
275-8 
184-9 
33-6 
78-2 
1-05 
•166 
•158 
April 
182 
12-0 
32-9 
15-1 
4-6 
270-1 
187-5 
91-0 
288-0 
0-36 
•066 
•181 
May . 
182 
15-8 
246 
8*6 
3-7 
277-1 
212-8 
96-0 
10-2 
0-64! 
•087 
•135 
June 
163 
171 
21-5 
3-8 
1-8 
270-4 
205-3 
172-8 
354-9 
0-80 
•105 
•132 
July. 
150 
281 
19-5 
11-3 
5-2 
238-2 
182-0 
100-0 
226-4 
1-44 
•187 
•130 
August 
144 
15-0 
31-5 
12-3 
17-7 
328-9 
188-4 
86-4 
296-1 
0-48 
•104 
•219 
September 
144 
8-4 
17-7 
3-5 
2-2 
260-6 
184-3 
80-2 
248-2 
0-47 
•058 
•123 
October . 
126 
20-3 
6-8 
1-5 
0-8 
213-3 
178-6 
343-6 
198-5 
2-99 
•161 
•054 
November 
194 
16-3 
12-6 
4-5 
7-3 
202-1 
237-1 
42-4 
208-4 
1-29 
•084 
•065 
December 
158 
331 
190 
4*4 
10-6 
217-8 
180-2 
25-6 
357-0 
1-74 
•209 
•120 
Year. 
167 
23-7 
17-0 
5-5 
2-0 
233-1 
183-0 
80-2 
301-9 
1-39 
142 
•102 
Winter 
174 
37-3 
11-3 
3-4 
1 13 
221-7 
160-7 
42-9 59-8 
3-30 
•214 
•065 
Summer . 
161 
13-8 
24-1 
8-5 
4-5 
267-8 
' 193-4 
94-9 292-4 
0-57 
•086 
| -150 
I 
It is interesting to compare the relative amplitudes of the 24-hour and 
12-hour waves, and for this purpose the ratio clJcl 2 has been calculated, and 
is also given in the table. If aja 2 is greater than unity, the 24-hour term 
is of a greater importance than the 12-hour term, and a study of the values 
of oq/<x 2 during the year shows that for the six months October to March 
the ratio exceeds unity, while for the six months April to September, with 
the single exception of July, the ratio is less than unity. This implies that 
at Edinburgh during the winter the 24-hour wave is of most importance, 
while during the summer the 12-hour wave is predominant. It was this 
consideration which led to the division of the year into the two portions, 
October to March and April to September. 
A comparison of the values of a x and a 2 for the different months shows 
that a ± is a maximum in January and decreases to a minimum in September, 
its range of values being extremely wide — from 8*4 in September to 71*3 in 
January. On the other hand, a 2 reaches a maximum in April, and a 
minimum in October. These variations are shown graphically in fig. 6. 
