624 
LIEnT.-GENERAL R. STRACHET ON HARMONIC ANALYSIS OF 
Table showing frequency of the Value of /x falling in the several Quadrants in 
different Months of the Year. 
jMobtIi. 
1st 
component. 
2ncl 
component. 
3rd component. 
4th component. 
Quadrant. 
Quadra nt. 
Quadrant. 
Quadrant. ! 
3 
4 
1 
4 1 2 3 
4 
1 
1 1 
2 13 4 : 
Jainiary .... 
100 
100 
. 100 
15 
60 
25 ’ i 
Febimai'y .... 
100 
100 
. 100 
25 
o 
.. ' 70 1 
March. 
100 
100 
20 75 , .. 
, , 
95 5 1 
April. 
100 
5 
95 
.. 100 .. 
. , 
20 70 10 ' 
May. 
100 
35 
65 
100 .. 
5 
45 
50 * 
Jn]ie. 
100 
45 
55 
.. 100 : .. 
100 
.Tuly. 
100 
10 
90 
.. 100 .. 
5 
35 
60 
Angu.st. 
100 
100 
100 .. 
, , 
55 45 
September 
100 
100 
84 16 
.. 
5 95 
October .... 
100 
100 
.. ' 40 60 
, , 
10 85 5 
November. 
100 
, , 
100 
. 100 
25 
40 
25 10 
December .... 
lOO 
100 
. 100 
15 
75 
10 
Note.— Where the 4th qiiadrant is shown as preceding the 1st quadrant, it is assumed that the phase 
■ has fallen back, or occurred earlier. 
These figures show that in their main features the variations in the value of /x for 
the second, third, and fourth components have a truly periodical character, and that 
as the year passes from winter to summer the maximum phase of the second and third 
travels backwards, that is, gradually occurs earlier in the day, while it retimns in 
tlie opposite direction, l)ecoming gradually later, in the change from summer to winter. 
In the fourtli component corresponding changes take place between the equinoxes and 
tlie solstices. 
A reference to Table B from the Greenwich series will show in more detail how 
well the consistency of the I’esidts is maintained in all the components. 
The first component is always in the first quadrant, i.e., between noon and 
6 B.M. The variation of the five-years mean from the 20-yea,rs mean is in no month 
more than 2^°, or 10 minutes of time, and the average for all the months is less than 
lialf that amount. 
In the second component the retrogression of the phase of maximum in the summer 
months leads to (which from August onwards to March is always in the first 
quadrant, or between midnight and 3 a.m.) falling back into the fourth quadrant, 
tliat is occurring before midnight, and the greater frequency of this in May and 
.lune is thei'efoi’e quite consistent with the regular periodical change which brings the 
mean value of p^ in June very nearly to zero, or midnight. The variation of the five- 
years mean fi’om the 20-years mean is in no month greater than (5° or 24 minutes of 
time, and for the whole year the average variation is only 2°‘3, or 9 minutes of time. 
