320 On the Rain and Drainagr - Waters at Rolhamsted. 
however, to the height of summer, the soil must generally tend 
to become drier, and the calculated evaporation lor the months 
lying in this period will consequently be on the whole too low. 
The hottest season passed, a reverse action will take place, the 
soil gradually becoming more saturated with water ; the figures 
given as representing evaporation during this portion of the 
year will therefore be too high. A comparison of the amounts 
of evaporation calculated for the Rothamsted soil with the 
average monthly evaporation determined by Mr. Greaves for 
a water-surface (see Table XXV., p. 325) is most instructive, and 
shows very clearly the existence of the two errors in opposite 
directions which affect our calculations for spring and autumn. 
During December, January, February and March, the amounts 
of evaporation Irom a water-surface are seen to be all but identical 
with the evaporation calculated for the soil of the 20-inch gauge 
at Rothamsted, plainly showing that the soil is during these 
months sufficiently saturated with water to yield at all times 
a maximum rate of evaporation. During the warmer months of 
the year we should expect to find the evaporation from the water 
uniformly greater than from the soil, as the soil is not at this 
time of the year permanently saturated. VVe find, however, that 
while the evaporation from a water-surface considerably exceeds 
that from the soil from April to August, the evaporation from 
the soil is distinctly greater during September, October and 
November, the difference in favour of the soil during these 
months amounting on the whole to 1 inch. Under spring and 
autumn seasons of sim.ilar temperature we find, therefore, that in 
the spring the evaporation from water exceeds that calculated for 
the soil, while in autumn the converse holds true. This is quite 
in accordance with our previous reasoning. There can be little 
doubt that the evaporation credited to the soil for September and 
October is somewhat too high, while that reckoned for April and 
May is rather too low. 
On comparing the records of the 20, 40, and 60-inch gauges 
given in Table XXII. (p. 318), it is evident that the relation 
between them is different at different times of the year. In the 
spring months drainage will continue to take place from the 
lower layers of a deep soil when it has altogether ceased in a 
shallower soil. In summer and autumn, on the other hand, the 
drainage will be less from the deeper soil, as there is in this 
case a larger mass of dry soil to be saturated with water before 
drainage can commence. In March and April the drainage 
from the soil 40 and 60 inches in depth, is on an average 
rather greater than that from similar soil 20 inches deep ; but 
in August, September, and October, the relation is reversed, the 
shallowest soil yielding the largest drainage. These facts 
appear in the records of the gauges for the last six years, as 
