240 
DR. C. CHREE: MAGNETIC DECLINATION 
to the effect occurring 24 hours in advance of the cause. An explanation may, to 
some extent, be forthcoming from, the fact that there is a tendency for days of large 
range and of small range to occur in groups. This tendency is, however, markedly 
less than the corresponding tendency in the case of sunspot areas already alluded to 
(a difference of significance in itself), and it is unusual to have more than two or three 
successive days on which the ranges are all markedly greater or markedly less than 
the mean for the month. 
Table XXV.—Algebraic Excess of Sunspot Areas over the Mean for the Year. 
Aear. 
Days of largest range. 
Days of least range. 
71+1. 
n. 
71-1. 
71-2. 
77-3. 
n - 4. 
77+1. 
77. 
77-1. 
77-2. 
77-3. 
77-4. 
1890 
- 46 
- 35 
- 41 
- 41 
- 28 
- 11 
+ 28 
+ 31 
+ 18 
+ 1 
- 13 
- 28 
1891 
■ 44 
■ 59 
- 84 
■ 56 
- 32 
+ 1 
- 24 
- 26 
- 46 
- 64 
- 69 
- 98 
1892 
- 105 
+ 32 
+ 161 
+ 223 
+ 321 
+ 408 
- 53 
- 186 
-205 
-230 
- 98 
- 142 
1893 
- 49 
+ 90 
+ 196 
+ 240 
+ 245 
+ 207 
+ 140 
+ 163 
+ 186 
+ 162 
+ 160 
+ 83 
1894 
+ 106 
+ 184 
+ 201 
+ 182 
+ 154 
+ 212 
-320 
-366 
- 364 
-334 
-296 
-251 
1895 
- 130 
- 152 
- 161 
- 194 
-229 
-216 
■ 53 
- 27 
- 29 
+ 8 
+ 130 
+ 220 
1896 
+ 57 
+ 72 
+ 69 
+ 83 
+ 67 
+ 30 
- 128 
- 123 
- 112 
- 127 
- 108 
-114 
1897 
- 20 
+ 28 
+ 15 
+ 33 
+ 76 
+ 135 
+ 78 
+ 78 
+ 80 
+ 46 
+ 30 
• 42 
1898 
+ 207 
+ 171 
+ 180 
+ 158 
+ 110 
+ 70 
-246 
-258 
-208 
- 166 
- 130 
- 64 
1899 
+ 46 
+ 50 
+ 53 
+ 51 
+ 56 
+ 41 
+ 11 
+ 24 
+ 47 
+ 48 
+ 42 
+ 34 
1900 
+ 27 
+ 54 
+ 57 
+ 55 
+ 57 
+ 66 
- 8 
- 9 
- 13 
- 4 
- 3 
+ 15 
Means . 
+ 4 
+ 40 
+ 59 
+ 67 
+ 72 
+ 86 
- 52 
- 64 
- 59 
- 60 
- 32 
- 35 
§ 29. A minute comparison of Tables XXY. and XXIII. brings out some very 
curious features, which seem explicable only in the view that accident plays a large 
part in the mean results for any single year. In 1890, 1891, and 1895, the entries 
under n, n— 1, n— 2, and n —3 for the days of largest range are not merely negative 
in Table XXV., but are conspicuously more so than the corresponding entries in 
Table XXIII. In 1893 the entries under n— 1, n — 2, and n — 3 for the days of least 
range in Table XXIII. were negative, but in Table XXV. the corresponding entries 
are positive and large. Thus in this year the phenomena for the five days of largest 
and the five of least range a month differ notably from the phenomena for the average 
day, but they differ in the same direction. With the exception of 1894, 1898 is the 
year in which the sunspot phenomena from the two groups of days differ most; but 
this difference is largest for the columns headed n + 1, i.e. for the days subsequent to 
those of largest range. 
§ 30. Considering the contradictions in Table XXV., it appeared desirable to obtain 
corresponding results for the different months of the year from the 11 years combined. 
They appear in Table XXVI. The headings have the same meaning as before. As, 
