f the rate
at which nitrification may go on under different circumstances, the
results furnished by actual analyses of soils and their drainage-waters
are of still more practical value; and the Rothamsted experiments
fortunately furnish us with a number of these valuable results.

_Quantity of Nitrates formed in the soils of Fallow Fields._

These researches had to be carried out on soil taken from fields lying
in bare fallow; for no true estimate of the amount of nitrates formed
could have been obtained from _cropped_ fields. In the first 27 inches
of soil of six separate fields, nitrate-nitrogen was found to vary from
36.3 lb. to 59.9 lb. per acre. In four of these fields the largest
proportion was found in the first 9 inches of soil; in the remaining
two, in the second 9 inches; while the third 9 inches in two fields
showed almost as large a proportion as the first 9 inches.[121]

_Position of Nitrates depends on Season._

The position of nitrates in the soil depends largely on the season; for,
as has been already pointed out, their production is almost entirely
limited to the surface-soil, and it is only by being washed down in rain
that they find their way to the lower layers. A wet season, therefore,
has the effect of increasing their percentage in the lower soil-layers.

_Nitrates in Drainage-waters._

As there is a certain proportion of nitrates that finds its way even
below the first 27 inches of soil, the above results do not show their
total production. To accurately estimate this amount we must ascertain
the quantity escaping in drainage-water. Here, again, the Rothamsted
experiments furnish us with valuable data. The amount found in
drainage-waters of course naturally varies very much, and depends
largely on the rainfall; but taking an average of twelve years, this has
been found to amount to between 30 and 40 lb. per acre--an amount not so
very far short of that found in the first 27 inches of the soil itself.
This was from comparatively poor soil, it must be remembered, and a much
larger quantity would undoubtedly be produced in the case of richer
soils. Adding then the results together, we find that in soils like
those at Rothamsted, when in bare fallow, between 80 and 90 lb. of
nitrogen are converted into nitrates in some fourteen months' time--an
amount equal to about 5 cwt. of nitrate of soda. It is a fact of no
little practical significance that nearly one-half of this large
quantity is f