82 
J. HOPKINSOX-HALF-A-CENTURF’s 
Table II gives the mean rainfall “for each decade in half a 
century for the number of stations shown in Table I, and also the 
mean for the whole period. Owing to the varying number of 
stations, this may require a slight correction. In the first decade 
we have one rather wet station, Nash Mills; in the second we have 
that coupled with a decidedly dry one, Hitchin. The excess in 
one decade would therefore seem to be approximately compensated 
by the defect in the other, and afterwards the number of stations 
is sufficient to give a fairly reliable mean for the county. It may 
be well, however, to examine this question rather more carefully. 
The close agreement between the mean rainfall in 1880-89 as 
deduced from the returns for our 18 selected stations and from the 
returns for the whole of our stations (26 to 30), as shown below, 
proves that the continuous records of the smaller number of stations 
give as accurate a mean for the whole county for this period as we 
can expect to obtain. Accepting, then, these 18 stations as our 
standard to represent the county, it is an easy matter to correlate 
the returns from a smaller number of stations in the previous 
decades with the returns for these. The means thus corrected are 
as follows:—1840-49, 24-83 ins.; 1850-59, 25*81 ins.; 1860-69, 
25*88 ins.; 1870-79, 27-53 ins.; to which must be added the mean 
for 1880-89, 26*74 ins. The mean for half a century for the 
whole county thus derived is 26*16 ins., being about one per cent, 
less than that directly derived from the actual means for each 
decade as given in this table, and this difference is almost entirely 
due to the comparatively small rainfall at the one station, Nash 
Mills, in the decade 1840-49. But this decade had generally over 
the country the average rainfall of the 50 years 1830-79 (see 
Symons’ ‘British Rainfall’ for 1881, p. 32), and if for 1830-39 
we substitute 1880-89, it will have nearly the average for the 
50 years 1840-89. I therefore think that it will be best to apply 
no theoretical correction to the mean values shown in Table II, 
accepting them as giving a sufficiently near approximation to the 
actual mean rainfall in Hertfordshire for the half-century 1840-89. 
In Table III* is given the mean monthly and annual rainfall 
for the decade 1880-89 at the eighteen stations for which complete 
returns for the decade have been received, and also the mean at 
27 stations in the wettest year of the period (1880) and at 28 in 
the driest year (1887). In order to ascertain whether the eighteen 
stations fairly represent the county, I computed the mean for all 
the stations (26 to 30) for the decade, with the following result:— 
Jan. 1-73 April 1-88 July 2 70 Oct. 3-18 
Feb. 1*94 May 2-06 Aug. 193 Nov. 2-83 
Mar. 1-61 June 2*02 Sept. 2-58 Dec. 2-32 
giving for the year 26*78 ins. This differs by only 0*04 in., or 
0*15 per cent., from the mean for the 18 stations, a remarkably 
close agreement, and the greatest difference in any one month 
is only 0*06 in. 
* It is scarcely necessary to state that in this and other tables the amount of 
rain is expressed in inches. 
