114 
DE. C. CHREE; SOME PHENOMENA OF SUNSPOTS AND OF 
to that adopted earlier in the paper with daily magnetic values. Use was made of 
Wolfee’s frequencies instead of the Greenwich sunspot areas, because mean values 
of the former were available for all months of the 11 years. 
The three months of highest sunspot frequency in each year were entered in 
column w, the three months immediately preceding in column n—1, and the three 
months immediately subsequent in column n + 1. The months having been thus 
arranged in three columns, corresponding lists were made of Wolfee’s frequencies, 
and of the two sets of percentage figures in Table XVII. Exactly the same 
operations were then gone through, taking as basis the three months of least sunspot 
frequency in each year. 
The mean results thus found from the eleven years are given in Table XVIII. 
Table XVIII.—Data from Three Months of Largest and Three Months of Least 
Sunspot Frequency in each Year, 1890 to 1900. 
Sunspot 
frequencies. 
H ranges, 
percentages. 
Character figures, 
percentages. 
n- 1. 
n. 71+1. 
n-\. 11 . 
71+1. 
71-1. n. 77+1. 
Months of largest frequency 
Months of least frequency . 
43-4 
40-8 
54-7 43-6 
28-8 40-0 
103-0 104-5 
100-1 96-3 
102-4 
94-1 
110-9 104-1 102-6 
97-7 98-9 94-8 
Excess of first group . 
2-6 
25-9 3-6 
2-9 8-2 
8-3 
13-2 5-2 7-8 
The number of months was too small to eliminate accidental features. This is 
especially true of the character figures for reasons already stated. 
The greater uncertainty of the character figures is borne out by the fact that while 
the mean percentage value from the 3 months of largest sunspot frequency exceeded 
the corresponding mean from the 3 months of least frequency in every single year in 
tne case of H ranges, the same phenomenon occurred in only 7 of the 11 years in the 
case of character figures. 
The 4 days retardation, shown in Tables V. and VII., would lead to some association 
of sunspot frequency with magnetic phenomena in the following month, but even 
the H range figures in fable XVIII. suggest more connection than this would 
account for. 
§ 32. In the case of the range, R, of mean diurnal inequality for the year in H at 
Kew, the formula found by applying the method of least squares to the observations 
of the 11 years, 1890-1900, may be wifitten 
K = Ro(l + l'07x lO-^'S), 
(1) 
