1)E RANCE : UNDERGRUUXD WATER-SUPPLY AND RIVER ELOODS. 200 
until the top was level with the surface. He found 25 per cent, of 
the rain to be absorbed. These observations were probably taken with 
the soil resulting from the decomposition of vegetable matter mixed 
with the stiff Lancashire Boulder Clay. M. Maurier's observations 
gave the larger figxire of 39 per cent ; while the very careful observa- 
tions taken by Mr. Dickinson between 1836 and 1843 at Xasli Mills 
near Hemel Hampstead, Herts., gave 42 per cent. ; this is with a 
chalky porous soil. The Xash Mills observations have been con- 
tinued to the present time by Mr. Dickinson's partner, Dr. Evans F. R.S. 
The gauge, on two cast-iron cylinders, 18 inches in diameter, 
and 3 feet in length, turned to a knife edge on the top, are sunk to 
a depth of 3 feet below the level of the gi'ound in which they are 
placed, so that the edge just projects. One of the cylinders is filled 
with the ordinary surface soil of the neighbourhood, and the other 
with fragmentary chalk. Grass on and around the cylinders. The 
rainfall is observed in the immediate neighbourhood of the gauges, 
and the register is made up at 9 a.m. The year is divided into two, 
the winter half commencing 1st of October, and the summer half 
1st of April. 
The following figures give the average for 30 years : — 
Winter. 
Summer. 
Pxain. 1 Soil. 
Chalk. 
Rain. 
Soil. 
Chalk. 
13-02 1 5-20 
779 
13-35 
0-65 
171 
Annual average : — 
26-38 
5-85 
9-51 
The varying capacity of soils for the downward transmission of water 
is well-shown in the above table, but Dr. Evans points out it is 
doubtful whether a 3 foot tube is a fair test of the water that would 
really make its way down to feed the chalk water, as it is possible 
that atmospheric influence may reach even 10 feet, and maybe drawn 
up by capillarity within the limit of superficial evaporation. 
Mr. Green's observations at Lea Bridge from 1852 to 1873, with 
an average rainfall of 26 '8 inches, gives an average percolation of 
26-6 per cent. He finds the abundance of water, in a river, to be 
more closly dependent on percolation than on mere rainfall; for many 
