On the Rain and Drainage - Waters at Rothamsted. 
339 
spring. 2. That in early autumn the drainage from the 20-inch 
gauge is richest in nitrates, but that in late winter and spring 
the drainage from the GO-inch gauge becomes generally the 
richest. These two facts will be found further illustrated by the 
more recent analyses contained in Table XXXIV. (p. 346) ; they 
admit of ready explanation. The summer, as already mentioned,- 
is the season when nitrates are most abundantly produced in the 
surface soil ; but little drainage occurs in summer time, owing 
to the high rate of evaporation ; the nitrates therefore accumu- 
late in the soil. In the autumn drainage becomes active, and the 
washing-out of the nitrates commences ; the first drainage is not, 
however, always the strongest, as the nitrates are most abundant* 
at the surface, and must be displaced by rain, and allowed time 
for diffusion before they can appear in quantity in the drainage- 
water. The drainage from the shallowest soil is the first to 
show a maximum contents of nitrates, because the amount of 
displacement and diffusion required to bring the nitrates withiri 
the area of discharge is here the smallest ; for the very same 
reason the shallowest soil is also the most quickly washed out, 
while the deepest soil, having a larger mass available for the 
diffusion of the nitrates, parts with them more equably. 
Before considering the more complete series of monthly 
analyses, extending from May, 1877, to the present time, it will 
be convenient to give the results obtained in March and April^ 
1879 (Table XXXIII.), when a detailed examination was made 
of the runnings from the 60-inch gauge, as these results will 
help to interpret the remainder. 
The winter of 1878-9 had been extremely wet, and the drain- 
age collected up to the middle of February far exceeded the 
normal quantity. March, however, was dry, so that the drain- 
gauges almost ceased running, and it was necessary in the 
present case to allow the drainage-water to accumulate for two or 
three days in order to obtain sufficient for analysis. The rain 
credited to March 24-26 was really snow, which, having thawed, 
was measured on the last of these days. The snow on the soil of 
the drain-gauge would melt later. In April the amounts ofraiiv 
and drainage were much more considerable. 
The analyses of the drainage-waters during March display a 
considerable amount of uniformity. The more considerable 
rains, namely, those of the 10th, 14th, 26th, and 30th, do not^ 
except in the last instance, appreciably increase the amount of 
drainage, but they have all a more or less distinct effect in 
temporarily diminishing the proportion of nitrates in the 
drainage-water. 
The results obtained in April, with greater rainfalls, are much 
more striking. Here, as before, small rainfalls, as those of the 
