TEKKESTKIAL MAC4NETISM AT KEW OBSERVATOKY. 
loy 
the column headed n. This relates to the 5 days of largest H range in each month. 
These days total 180 in each of the 3-year periods, and 660 in the eleven years. Of 
the 180 selected days in the first 3-year group only 54 attained character “ 2.” The 
corresponding numbers in the second and third 3-year groups were respectively 132 
and 108, and of the 660 days selected from the 11 years 365 reached character “ 2.” 
The three last lines in Table XIV. all relate to the whole 11 years. Still considering 
column we learn that 266 of the 660 days had character “ 1.” Combining these 
with the 365 days of character “2,” we have in all 631 disturbed days, leaving 29 
days of character “0.” Nearly all these 29 quiet days came from the last months of 
1900, when days of character “ 2 ” were non-existent, and days of character “ 1 ” were 
rare. 
The data in columns n—b to n—1, to n-\-l to n4-35, have an exactly similar 
significance. They refer, as in Table XI., to the 5 days before and the 35 days after the 
selected days of largest H range. The last column in the table shows what numbers 
we should have got from average days of the years concerned. Eleven-year data in 
columns n —5 to n + S5 which exceed the corresponding figures in the last column 
are in heavy type. 
The results relating to the character “2” figures from the 11 years are shown 
graphically in the broken line curve of fig. 2, p. 100. 
The number of days of character “ 1 ” in the 11 years is very notably in excess of 
the mean in columns n + 27 to 7i + ^2, but it is the great development of character “ 2 ” 
figures that is mainly responsible for the prominence of the 27-28-day period in 
Table XI. Taking the 11-year data, the number of days of character “ 2 ” is above the 
mean only in columns n—2 to n4-2 and n + 2b to n-t30. The rise to the crest and the 
subsequent subsidence in the primary pulse of character “ 2 ” figures are exceedingly 
rapid, and the same is true to only a slightly less extent of the secondary pulse. 
The numbers of days “ 2 ” in columns n-t-27 and n-l-28 exceed those in columns n — 2 
and n-\-2, and simply tower over the numbers in columns n + 2> to n-f 24. The relative 
prominence of the character “ 2 ” figures in columns n + 27 and ?i + 28 is even greater 
for the two first 3-year periods in Table XIV. than for the 11-year period. In the 
third or sunspot maximum group of yeai's this prominence is less. 
The number of days of character “ 1 ” is much less variable in Table XIV. than the 
number of days of character “ 2.” When, however, there are a large number of the 
latter days there are so many the fewer available for other character figures. Thus 
the number of days of character “ i ” is perhaps less instructive than the number of 
disturbed days as a whole. This latter number subsides much less rapidly than it 
rises in both the primary and secondary pulses. This presumably is a direct con¬ 
sequence of the known fact that while a highly disturbed day often immediately 
follows a quiet day, the converse is rare. 
The fewness of quiet days in columns n + 26 to w + 30 presents the 27-28-day period 
in perhaps as striking a light as any. 
