28 
INVESTIGATIONS ON ROTHAMSTED SOILS. 
Table 10. — Broadbalk wheat soils, sampled in October, 1S93 — Percentages of nitro- 
gen in fine dry soil, as determined, respectively, by the soda-lime method and by 
the Kjeldahl method: also the percentages of " nitric" nitrogen {nitrogen existing 
in the form of nitrates). 
[Summary of preceding results in Table 9— averages of all plats sampled at each depth.] 
Depth. 
Num- 
ber of 
plats. 
Total nitrogen. 
By soda- By Kjel- 
lime. dahl. 
Kjeldahl 
(+)or (-) 
soda-lime. 
** Nitric " 
nitrogen. 
Per cent. 
Per cent. 
Per cent. 
Per cent. 
First 9 inches (excluding 2a and 2b) 
19 
0. 1098 
0.1149 
+0. 0051 
0.000501 
First 9 inches! all plats) 
21 
.1167 
.1222 
+ .0055 
.000520 
Second 9 inches 
21 
.0728 
.0784 
+ .0056 
.000931 
Third 9 inches 
21 
.0613 
.0666 
+ .0053 
.000356 
Fourth 9 inches 
12 
.0474 
.0511 
+ .0037 
.000223 
Fifth 9 inches 
12 
.0437 
.0472 
+ .0035 
. 000174 
12 
.0392 
.0430 
+ .0038 
.000144 
Seventh 9 inches 
12 
.0386 
.0420 
+ .0034 
.000143 
Eighth 9 inches 
12 
.0375 
.0396 
+ .0021 
.000144 
Ninth 9 inches. 
4 
.0358 
.0391 
+ .0033 
.000115 
Tenth 9 inches 
4 
.0358 
.0375 
+ .0017 
.000106 
It will be seen that in almost every individual instance, and always 
in the case of the average results for each depth of soil, the Kjeldahl 
method gives a higher percentage of nitrogen than the soda-lime 
method. The actual difference is greater in the case of the higher 
than of the lower depths. Reckoned, however, in percentage on the 
total nitrogen obtained, the deficiencj 7 by the soda-lime method is on 
the average considerabl}' less in the case of the samples of the first 
depth (which, as will be seen hereafter, contain not only more nitro- 
gen, but also more carbon, and a higher ratio of carbon to nitrogen) 
than in the lower or subsoil depths. But, with some marked excep- 
tions, the relative deficiency of the soda-lime method, though higher 
in t he subsoils than in the surface soils, is on the average fairly uni- 
form from depth to depth, and averages between 7 and 8 per cent of 
the total nitrogen. A comparison of the column of the summary 
table (Table 10), showing the greater amount of nitrogen found by 
the Kjeldahl method, with the column showing the nitrogen existing 
as nitrates, will show that the higher results cannot be accounted 
lor by more complete determination of the nitric nitrogen by the 
Kjeldahl method; nor do the figures showing the relation of carbon 
to 1 1 1 1 rogen (to be hereafter given) indicate any deficiency of carbon 
for the decomposition of the organic nitrogenous compounds by the 
soda-lime. The conclusion is rather that the action of the liquid acid 
in breaking up the organic nitrogenous compounds, in order lo con- 
vert the nitrogen into ammonia, in the Kjeldahl process, is more com- 
plete 1 ban is t he act ion of dry heat on the mixture of soda-lime and soil. 
On tin; whole it appears that the Kjeldahl results should be adopted 
;i- mosl correctly indicating the actual nitrogen in tin; 1893 samples — 
t hough t he soda-lime results will be more properly taken in any compari- 
son which it is desired to draw be1 ween these and the 1865 or 1881 
samples. For 1 he present, therefore, we will direct our attention to 
the nitrogen figures as obtained by the Kjeldahl method. 
