238 
DR. C. CHREE: MAGNETIC DECLINATION 
Table XXIII.—Algebraic Excess of Sunspot Areas over the Mean for the Year. 
Days of largest range. 
Days of least range. 
Year. 
n. 
n- 1. 
n - 2. 
n- 3. 
n. 
71-1. 
71-2. 
71 - 3. 
1890 
- 26 
- 23 
- 17 
- 5 
+ 8 
0 
- 6 
- 12 
1891 
- 13 
- 24 
- 19 
- 12 
+ 31 
+ 82 
+ 53 
+ 22 
1892 
+ 38 
+ 108 
+ 155 
+ 226 
-125 
-141 
-203 
-201 
1893 
+ 4 
+ 33 
+ 72 
+ 100 
+ 6 
- 20 
- 62 
- 122 
1894 
+ 170 
+ 147 
+ 127 
+ 108 
- 177 
-187 
- 179 
-165 
1895 
+ 14 
+ 2 
- 47 
- 92 
+ 6 
+ 20 
+ 40 
+ 85 
1896 
+ 28 
+ 37 
+ 35 
+ 48 
- 72 
- 71 
- 51 
- 6 
1897 
+ 23 
+ 7 
+ 24 
+ 48 
+ 37 
+ 15 
- 10 
- 36 
1898 
+ 118 
+ 106 
+ 93 
+ 66 
-143 
- 130 
- 114 
- 85 
1899 
+ 5 
+ 8 
+ 13 
+ 19 
+ 31 
+ 27 
+ 14 
+ 6 
1900 
+ 24 
+ 27 
+ 20 
+ 26 
- 7 
- 8 
- 15 
- 23 
Means. . . 
+ 35 
+ 39 
+ 41 
+ 48 
- 37 
- 38 
- 48 
- 49 
Percentages . 
+ 4-4 
+ 5-1 
+ 4-9 
+ 6-8 
-2-5 
-2-0 
-5-1 
-6-4 
some individual years, e.g. 1892, 1893, 1894, 1898, and 1900, the phenomena are so 
far distinctly favourable to views such as Arrhenius’, if we suppose the velocity 
of propagation from the sun highly variable. But even in these years the balance in 
excess of the S’s, and of deficit in the S"s, appears very small when we consider the 
great disproportion which exists between the means of the ten largest and the ten 
least absolute ranges. 
A remarkable feature in Table XXIII. is the comparative regularity in the values 
of the successive S’s and S's in any one year. But whether the trend of the sequence 
is to a rise or a fall seems largely fortuitous. Thus we have a regular rise from S to 
S_ 3 in 1892, but a regular fall in 1894. The phenomenon is probably due in part to 
the tendency for days of largest and of least sunspot area to occur in groups. 
Whilst the mean results in Table XXIII. suggest a closer connection of the range 
with the sunspot area two or three days previously than with that of the day itself, 
still we have a difference of 6‘9 between the final percentage values in the first and 
fifth columns of Table XXIII. as against a corresponding difference of 13'2 between 
the fourth and eighth columns. Any theory which requires as much as one day for 
the minimum time of transfer of disturbance from the sun to the earth has thus the 
evidence in its favour very considerably weakened. In every year, it should be 
mentioned, individual months gave results diametrically opposed to the final mean. 
§ 28. The excess of S over S', and the pre-eminence of S_ 3 in the final means, 
pointed to the desirability of further enquiry. The scheme of operations leading 
