On the Rain and Drainarje - Waters at Rotliamsted. 341 
per million in the form of nitrates. When a small rainfall 
occurred the quantity of the discharge was scarcely increased, 
but it became considerably diluted, the drainage from the 
lowest layer of soil being now mixed with rain-water, which 
had come through open channels directly from the surface. 
With a heavier rain, pressure was brought to bear on the water 
column in the soil, and the discharge from the lowest layer 
was then much increased, but diluted as before with direct 
channel water. 
Another experiment of the same character may be quoted. 
The drainage from the 20-inch and 60-inch gauges was collected 
both on the morning and evening of January 14th, 1879 ; 
between the two collections a thaw of snow had taken place. 
The nitrogen existing as nitric acid per million of water was as 
follows : — 
20-Incli 
Gauge. 
GO-Inch 
Gauge. 
Morning Collection .. 
8-4 
1-2 -7 
Evening Collection . . 
5-7 
G-3 
We have, therefore, to bear in mind that the strongest 
drainage-waters are those obtained after rain has ceased ; and 
that the composition of the drainage-water may be considerably 
influenced by the varying amount and distribution of the rain ; 
the rain applied to the surface of a soil not simply displacing 
the water below, but in part proceeding directly to the area of 
■discharge through the open channels of the soil. 
The results of the analyses of the mixed monthly samples of 
drainage-water since May 1877 will be found in Table XXXIV. 
•(p. 344). The amounts of monthly drainage there given will 
be found in a few cases not to correspond with those found in 
Table XIX. (p. 272). Thus the amounts credited to the 40- 
and 60-inch gauges in April 1878, and to the 20-inch gauge in 
February 1879, are in excess of the numbers previously given. 
The quantities now stated are the actual amounts of drainage 
passing through the soil ; but, as already explained (p. 271), 
they are known to be excessive in these particular cases from 
accidental circumstances, and were therefore corrected in the 
earlier table. Such corrections are here, however, inadmissible, 
as our object is to ascertain the total quantity of nitric acid 
extracted from the soil. In the case of August 1879, the quan- 
tity of drain age- water sampled for analysis from all the gauges 
was less, and in the case of the 60-inch gauge much less, than 
