348 On the Rain and Drainage- Waters at Rothamsted. 
drainage is accompanied by an increase in the amount of 
nitrates removed from the soil, but the latter does not increase 
at the same rate as the former. The increase in the amount of 
chlorides removed from the soil is, on the other hand, at nearly 
the same rate as the increase in the drainage. Expressing in 
figures the average results of the three gauges, the increase 
of drainage in these two extreme seasons is from 100 to 268, 
the increase of chlorides from 100 to 252, and the increase 
of nitrates from 100 to 182. As the rain which produces the 
increased drainage also supplies the chlorides, it is easy to 
understand why both drainage and chlorides should increase at 
a similar rate. The rain, on the other hand, supplies but an in- 
significant amount of nitrates, and only up to a certain point 
increases the rate of nitrification in the soil ; the larger amount 
of nitrates removed by heavy rain is thus in great measure 
simply due to the more thorough washing of the soil. 
As the quantity of the nitrates removed from the soil depends 
so greatly on the amount of drainage, and as this has been very 
variable during the last four years, it is impossible to say 
whether, on the whole, the amount of nitrates yielded by the 
-drain-gauges is, or is not, diminishing. It is, of course, how- 
ever, probable that the production of nitrates is slowly becoming 
less, and that in the earlier years of the experiment the amount 
-contained in the drainage-waters was even larger than at 
.present. 
The large quantity of nitrogen as nitrates removed per acre by 
the drainage-water is a fact of great interest. The figures at the 
foot of Table XXXIV. (pp. 344-5) show that the annual amount 
of nitrogen as nitrates removed in the drainage-water was, on 
an average of four years, 45*51 lbs., 36'32 lbs., and 43'59 lbs. 
respectively from the three drain-gauges, the mean of all being 
41"81 lbs., equivalent to 268 lbs. of ordinary nitrate of 
sodium. If we suppose that the drainage-water contained at 
the same time 0 5 part of nitrogen per million in the form of 
organic matter and ammonia, we shall have a total of 43"77 lbs. 
as the quantity of nitrogen removed in one year from an acre of 
uncropped soil for 17"281 inches of drainage. Such a quantity 
of nitrogen is equal to that contained in an average crop of 
wheat or barley ; its loss to the soil in the drainage-water is thus 
a matter of grave importance. Are we to suppose that a similar 
soil kept as bare fallow for an entire year in ordinary agricul- 
tural practice would have suffered a similar loss? The question 
really resolves itself into two: 1. Would the production of 
nitrates he similar in the two soils? 2. Would the loss by 
drainage be equal ? 
The experimental soils being supported on perforated iron 
