ON TEERESTKIAL GALVANIC CURRENTS. 
223 
The next step of treatment was, to resolve the series of terms for each month into a 
diurnal, a semidiurnal, and a tertio-diurnal series. With such numbers as we have here 
before us, I have found the following process very convenient. I take the numbers for 
the Westerly Inequality in January as an example. 
Arrange the numbers in two columns, 0 h to ll h and 12 h to 23 h , side by side, take the 
difference of corresponding numbers for double diurnal, and the sum (when corrected by 
subtracting its mean) for double semidiurnal. For convenience, the numbers may be 
left in the double form. Half the mean above-mentioned will be the true mean of the 
twenty-four hours. 
Hours. 
Westerly Tendencies. 
Differences for double 
diurnal. 
Sums for double 
semidiurnal. 
0 
12 
35 
20 
+ 15 
— 15 
Uncorrected. 
55 
Corrected. 
— 11 
1 
13 
39 
26 
+ 13 
-13 
65 
— 1 
2 
14 
40 
33 
+ 7 
- 7 
73 
+ 7 
3 
15 
39 
39 
0 
0 
78 
+ 12 
4 
16 
39 
43 
- 4 
+ 4 
82 
+ 16 
5 
17 
37 
44 
- 7 
+ 7 
81 
+ 15 
6 
18 
35 
42 
- 7 
+ 7 
77 
+ 11 
7 
19 
32 
38 
- 6 
+ 6 
70 
+ 4 
8 
20 
28 
33 
— 5 
+ 5 
61 
— 5 
9 
21 
25 
29 
— 4 
+ 4 
54 
— 12 
10 
22 
21 
29 
— 8 
+ 8 
50 
-16 
11 
23 
19 
31 
— 12 
+ 12 
50 
-16 
Mean 
Half, or mean of the 24 hours ... 
... 66 
... 33 
The tertio-diurnal term will contribute no part to the semidiurnal, but it will con- 
tribute a part to the diurnal. To find its value, arrange the diurnal terms (which I shall 
now call ‘ uncollected’) in three columns of eight hours each; take the means of the 
corresponding numbers, which will be the tertio-diurnal terms, and apply them nega- 
tively to correct the diurnal terms. 
Honrs. 
Uncorrected double 
diurnal. 
Sums. 
Means or double 
tertio-diurnal. 
Corrected double 
diurnal. 
0 
8 
16 
+ 15 
— 5 
+ 
4 
+ 14 
+ 5 
+ 10 
-10 
— i 
1 
9 
17 
+ 13 
- 4 
+ 
7 
+ 16 
+ 3 
+ 
8 
- 9 
+ 2 
2 
10 
18 
+ 7 
- 8 
+ 
7 
+ 6 
+ 2 
+ 
5 
-10 
+ 5 
3 
11 
19 
0 
-12 
+ 
6 
- 6 
-2 
+ 
2 
-10 
+ 8 
4 
12 
20 
- 4 
- 15 
+ 
5 
-14 
-5 
+ 
1 
-10 
+ 10 
5 
13 
21 
- 7 
- 13 
+ 
4 
-16 
-5 
— 
2 
- 8 
+ 9 
6 
14 
22 
- 7 
- 7 
+ 
8 
- 6 
-2 
— 
5 
- 5 
+ 10 
7 
15 
23 
- 6 
0 
+ 1 
2 
+ 6 
+ 2 
— 
8 
2 
+ 10 
A quarto-diurnal series would readily be formed from the semidiurnal series by the 
same process by which we have formed the semidiurnal from the original numbers ; and 
the semidiurnal would be corrected as in the operation for correcting the diurnal. But 
the numbers would be small, and would scarcely repay the trifling trouble. 
The numbers thus found ought to be multiplied by 0 - 5 ; and then the diurnal numbers 
ought to be multiplied by (secant 7 D 30') 8 , the semidiurnal numbers by (secant 15°) 6 , the 
tertio-diurnal numbers by (secant 22° 30') 6 . 
The following Tables exhibit all the results. 
2 o 2 
