346 On the Rain and Drainage - Waters at Rothamsted. 
show a high proportion of nitrates at the usual period of maxi- 
mum ; while in the dry autumn and winter which followed the 
proportion of nitrates is for the season unusually high. 
We have evidence again of the greater relative richness of the 
drainage from the 20-inch gauge-during the season of maximum 
proportion of nitrates, and its relative poverty, as compared 
with the water from the deepest gauge, as the season of minimum 
nitrates approaches. This gradual change of relation between 
the waters of the 20-inch and GO-inch gauges is most con- 
spicuously seen in the autumn and winter of 1877-78, and of 
1880-81. It follows from what has just been stated that the 
range in composition is greatest in the water from the 20-inch 
gauge, less in that from the 40-inch gauge, and least in that 
from the 60-inch gauge. At a still greater depth the drainage- 
water would probably have a uniform composition all the year 
round. 
The quantity of nitrates removed in the autumn drainage- 
waters is generally greater than at any other period of the year, 
the drainage-waters being most concentrated at this season, and 
the drainage also usually abundant. This excess during autumn 
is most marked in the drainage from the shallowest soil. Thus 
Avith the 20-inch gauge the nitrates removed during the last six 
months of the year have been on an average 65'3 per cent, of the 
•annual quantity ; with the 40-inch gauge, 62'6 per cent. ; and 
with the 60-inch gauge, 59'5 per cent. The amount of drainage 
for the same period being respectively 56"4, 55'2, and 54*5 
per cent. 
The effect of a heavy and continuous rain in removing 
exceptionally large quantities of nitrates from the soil is strik- 
ingly shown by the results obtained in September 1880. In 
this month 5*1 10 inches of rain fell at Rothamsted in five days, 
and the quantity of nitrogen as nitrates in the drainage-water 
from the 20-inch gauge amounted to nearly 16 lbs. per acre, an 
amount far larger than that obtained in any other month during 
the four years. This is quite in accordance with what has been 
already said (page 331). A heavy rain, falling in a short time, 
should be especially effective in discharging the soluble salts 
from a soil, as the smallest opportunity is then afforded for the 
retention of the salts through upward diffusion within the wet 
soil. 
The proportion of chlorides in the drainage-waters shows far 
less range of variation than that of nitrates. The chlorides 
present in these soils being derived from the rain are not, like 
the nitrates, produced chiefly at one season of the year, but are 
supplied whenever a shower falls. Small rainfalls, we have 
already seen, are richer in chlorides than large rainfalls, and 
