274 On the Rain and Drainage - Waters at Rothamsted. 
heavy one on the surface of the 60 -inch gauge, the figures for 
the month have been corrected ; the record of the 20-inch gauge, 
where there was no drift, being adopted for the other gauges for 
those days of the month during which the thawing of the snow 
occurred. The records thus altered appear in the Table as 2*349, 
2"524, and 2'428 inches respectively. In all cases in which the 
correction has been sufficiently large to become important, the 
figure in the Table will be found enclosed in a bracket. 
In a few cases, as in November 1870, February 1879, and 
January 1880, the monthly drainage from some of the gauges 
has exceeded the monthly rainfall. Generally this has been 
more or less due to rain, or especially snow, falling at the end 
of one month and appearing as drainage in the next. It 
occasionally happens also in severe winters that a considerable 
amount of frozen water is retained in the upper layer of the soil 
for some time, and appears as drainage only when a complete 
thaw takes place. There is, however, another possible ex- 
planation of excessive drainage in relation to rainfall, namely, 
the condensation of water by the soil directly from the atmo- 
sphere. That such condensation must take place whenever the 
temperature of the soil is below the dew-point of the atmosphere 
is quite plain.* During a clear frosty night both rain-gauge and 
soil will condense water from the atmosphere, and the soil, 
perhaps, somewhat the more. It seems, however, very probable 
that after a long-continued frost, followed by mild weather, the 
soil may continue for some time to condense from the air very 
appreciable quantities of water of which the rain-gauge will give 
no account. There is some evidence, which will be mentioned 
by-and-by, that such a condensation of water took place in the 
soils of the drain-gauges during the severe winters of 1878-9 and 
1879-80. The total amount of water obtained by the soil from 
the atmosphere without the records of the rain-gauge being 
affected, is probably, however, save in exceptional seasons, not 
considerable. We shall have some evidence further on that the 
condensation by soil during the winter months is at all events no 
greater than the condensation by a water-surface. 
Before discussing the results obtained in the drainage ex- 
periments, it will be well to consider briefly what takes place 
when rain falls upon a soil. It would be a mistake to regard 
an ordinary soil as a uniform porous mass, which simply be- 
comes saturated with water, and then parts with its surplus 
by drainage ; soil is, in fact, penetrated by innumerable small 
channels, and through these more or less of the drainage always 
* A soil baked by a summer sun may re-absorb a certain amount of water from 
the air during tlio night without its temperature falling below that of the air ; 
but the water thus absorbed is hygroscopic and will not appear as drainage. 
