18 SCIENCE. 
Even if we allow them their full value, however, 
they do not in all cases cover the entire loss of nitro- 
‘gen. On the basis of the results just described, the 
authors have calculated the average annual amount 
of nitrogen in the manure, crop, and drainage of 
fourteen experimental wheat-plots for a period of 
thirty years. In seven cases the nitrogen found in 
crop and drainage is from 14 to 40 pounds per acre 
less than the amount applied as manure. Analyses 
of the soil of one plot in 1865, and again in 1881, 
showed that about a third of this amount was still 
present in the soil, the latter having gained nitrogen. 
The authors believe the remainder either to have 
escaped the drain-tiles, and been carried into the 
lower strata of the soil, or to have been set free in 
the gaseous state in the soil. 
Dehérain’s experiments were intended to determine 
the total loss of nitrogen by the soil. Three series, 
of four plots each, were laid out, each plot having an 
area of one are (equal to about four square rods). The 
first series bore fodder-maize; the second, potatoes; 
the third, beets, fodder-maize, and esparcette succes- 
sively. During the first three years, one plot in each 
series was unmanured; one received, per hectare, 
80,000 kilograms of stable-manure; one, 1,200 kilo- 
grams of nitrate of soda; and one, 1,200 kilograms of 
sulphate of ammonia. During the following four 
years none of the plots received anymanure. At the 
beginning of the experiments, and at the end of three 
and seven years respectively, the percentage of nitro- 
gen in the soil was determined. With the aid of these 
determinations, a balance was struck for each plot be- 
tween the nitrogen originally present and that added 
in the manure, on the one hand, and that removed in 
the crops and remaining in the soil at the close of the 
experiments, on the other hand. In every case except 
that of the esparcette, a very great loss of nitrogen 
was found to have occurred. The following table 
contains the annual loss of nitrogen from a portion 
of the plots, reduced to pounds per acre to compare 
with Lawes, Gilbert, and Warington’s results :— 
Maize plots. 
Manuring. First period. | Second period. 
Stable-manure . — 257 — 118 
Nitrate of soda. | = shi: ==) hee 
Sulphate of ammonia = 820 — 132 
Nothing 5 — 338 ==" 5 
Esparcette plots. 
First period | 1 4 
Manuring. (roots and psec 
maize). (eer . 
Stable-manure . — 808 cel 
Nitrate of soda . = — 663 + 134 
Sulphate of ammonia — 732 + 135 
Nothing 3 ont + 149 
Compared with the losses observed in Rothamsted, 
some of these figures are enormous, being over nine 
Oe oot Sw Aad 
[Von. IIL, No. 
times as great as the highest obtained there. When 
we consider that the soil was calculated to contain, — 
to a depth of 14 inches, only about 7,000 pounds of 
nitrogen per acre, they seem to show that but a com- 
paratively short time would be required to reduce the 
supply of nitrogen to the point at which culture ceases 
to be profitable. 
In order to be able to judge of the force of these 
results, it is desirable, in the first place, to consider 
somewhat more in detail the method by which they 
were obtained. At the beginning of the experiments 
the soil was found to contain .204% of nitrogen; and 
this was made the basis of the calculation for all the 
plots. Itis highly improbable that this assumption 
of uniformity among all the plots, as regards nitrogen, 
is correct; and, when we consider that a difference of 
.001% corresponds to a difference of about 34 pounds 
of nitrogen per acre, we are led to question, not only 
the accuracy of Dehérain’s results, but the possibility 
of discovering small losses of nitrogen by means of 
soil-analysis. The absolute quantity of nitrogen was 
calculated on the further assumption that one hectare 
of soil to a depth of 35 centimetres weighed 38,850,000 
kilograms, which, again, involves a possibility of 
error. Finally, the determination of the quantity 
of nitrogen removed in the weighed crops rests on 
assumptions as to the percentage of nitrogen they con- 
tained, while it is a well-established fact that consid- 
erable variations in this respect occur. Especially 
does heavy manuring with nitrogenous fertilizers, 
such as some of these plots received, tend to increase 
the percentage of nitrogen in the crop. 
To recapitulate: Dehérain’s conclusions involve 
three improbable assumptions; viz., pertect uniform- 
ity of soil as regards nitrogen, equal weight of soil 
over equal areas, and a uniform and average percent- 
age of nitrogen in the crops. 
On the other hand, all but a few of the plots show 
a loss of nitrogen ; and while, for the reasons just 
stated, the accuracy of the reults is very questionable, 
it would appear that we must admit some loss of ni- 
trogen to be probable in most of the experiments. | 
Aside from these considerations, however, there are 
others which should be borne in mind. During the 
first three years the manured plots were very heavily 
manured. Sulphate of ammonia and nitrate of soda. 
were applied at the rate of 1,070 pounds per acre, and 
stable-manure at the rate of over 35% tons per acre, — 
quantities much greater than would be used under 
any ordinary conditions. Moreover, from the fact 
that the unmanured plots yielded nearly as large crops 
at the close of the experiments as at their beginning, 
we may conclude that the soil was naturally of good 
quality. ; 
-Dehérain’s calculations show that but a very small 
proportion of the nitrogen of the manures was utilized 
by the crops; and, though the exactitude of his fig- 
ures may be questioned, the general result is what we 
should expect. A large excess of available nitrogen 
was evidently present in the soil. The latter was oc- 
cupied by crops for only four or five months of the year, 
at most (except the esparcette plots); and, during the 
remaining two-thirds of the year, leaching, and other 
