138 
J. HOPKINSON-WATER AND WATER-SUPPLY 
them 10 V 0 of an acre in area. The gauge should be filled with the 
ordinary soil of the district, and should have grass growing on its 
surface ; its rim should project about two inches above the surface 
of the ground; and a pipe from the bottom should lead to a col¬ 
lecting vessel, the top of which should of course be below the 
bottom of the gauge. Eor experimental purposes percolation- 
gauges may be filled with any material; in a chalk district, for 
instance, it would be well to have one filled with chalk, in a sand¬ 
stone district with sandstone, etc., grass being grown on the surface, 
but such gauges in every case should be supplemental to the gauge 
which is filled with the soil which is most prevalent in the district, 
and by which in most cases it will probably be surrounded. At 
Bothamsted the soil was left undisturbed, the gauges being built 
up around them, and for the special purpose of the investigation 
vegetation is not allowed to grow on the surface. 
The three gauges—for rainfall, evaporation, and percolation— 
should be read daily at 9 a.m., and their records should be entered 
to the previous day, as in all rainfall investigations. If, in 
addition, a river flowing by were gauged, and the height of water 
in a well at some distance from the river were measured daily, or 
even monthly, very valuable information which we do not as yet 
possess would in course of time be obtained. Such an investigation 
might with great advantage be undertaken by the Hertfordshire 
County Council. 
Bainfall observations have been carried on in Hertfordshire since 
the year 1833, and I have elsewhere shown that the mean rainfall 
for the whole county for the 50 years 1840-79, as deduced from 
observations taken at a varying number of stations, has been between 
26 and 26^ inches.* The mean rainfall in Hertfordshire for each 
year during this period is shown in the accompanying table (p. 139) 
which is compiled from continuous observations taken at the 
following stations:—Bash Mills, Hemel Hempstead, f 1840-90; 
Hitchin and Boyston, 1850-90; Bothamsted, Harpenden, 1853— 
90; Berkhamsted, 1855-90; Gorhambury, St. Albans, and Bav- 
fordbury, Hertford, 1860-90; and Much Hadham, 1865-90; each 
year being from 1st October to 30th September, and the years being 
divided into summer and winter halves. 
If we assume the mean rainfall in Hertfordshire for the half- 
century to be 26^ inches, we shall probably be as near as possible 
to the truth. 
But the rainfall in this part of England has been considerably 
greater during the last 50 years of the present century than it was 
during the first 40 years, and therefore this is too high a mean for 
the county for the century, the true mean being a little under 25 
inches. This determination has been arrived at by combining with 
* ‘ Trans. Herts Bat. Hist. Soc.,’ Yol. YI, p. 82. 
f This being the only rainfall station in Hertfordshire during the decade 
1840-50, and having a rainfall 5 per cent, in excess of the mean fall in the 
county, the values from it for this decade have been reduced by this amount in 
order to give the true mean as nearly as possible. 
