62 
THE YORK RAINFALL. 
This suspicion of possible error fortunately does not affect 
the very interesting question of relative monthly rainfall. 
Here we find, for the 70 years, that March comes out 
decidedly the month of least rain. There is also an accentua¬ 
tion of August as wetter than July and October, and of 
September as much less wet than these three. The 84 years 
accentuates more strongly the characters of March and 
September. 
On the other hand the addition of the 1811-24 observations 
actually deposes July and August in favour of October as the 
wettest month, and equalise these two. I wish to insist 
strongly upon the significance of this. A hundred years will 
perhaps give the annual rainfall within 1%, but probably 200 
years or more would be required to he sure of the monthly 
rainfall within, say 3%. For, while the 50 year annual fall is 
within about 1% of that for 84 years, the September fall 
differs by nearly 10% . This, it is true, is largely the result 
of the abnormal excess in the seventies decade (33J inches 
against an average of 21J in the five other decades), but this 
simply illustrates the possibilities of erratic monthly falls over 
a period of many years. The difference in March, which 
cimes second in divergence is 3%. Only in April is it 
inappreciable. 
How entirely inadequate a ten-year period proves to be is 
further illustrated by the following table* :— 
Decade 
Driest 
10 yrs. 
Wettest 
10 yrs. 
etuling. 
month. 
fall. 
month. 
fall. 
• 
O 
H- 
co 
H 
M ay 
L4'68 
November 
... 25*91 
1850. 
March 
15*30 
October 
29*11 
i860. 
February ... 
11*64 
August 
... 30*82 
1870. 
April 
14*11 
October 
... 27*42 
l880. 
March 
V5*t 
September 
••• 33 * 55 § 
H 
OO 
0 
« 
April 
16-15 
July ... 
••• 3 i *34 
1900. 
March 
13*20 
October 
30*28 
(1811- 
■24)^ March 
15*29 
October 
... 29*14 
Four months—Feb. to May). 
(Five months— 
-July to Nov.) 
* These are the o.hsolute falls, but the month, with one exception, would in 
every case be the same if reduced to a 30-day mean. § Only 10 - 49, or less than 
half, in 1891—1900. f January was only 0T5 more. J Taken reduced to 10 
years and 3'3% added as correction for height of gauge above ground. 
