QUESTION OP THE SOURCES OP THE NITROGEN OP VEGETATION. 
17 
the increase was in most cases much greater on the various legnminons plots. Thus, 
comparing the final means in the foregoing Table with those in Table V. (p. 14), tlie per¬ 
centage in the wheat-fallow soil is 0'1028 against 0'09G9 in 1881 ; in the Trifolium 
repens soil it is 0T179, in the Vida sativa soil 0T191, in the Melilotus leucantJia soil 
0T130, and in the Medicago sativa boW 0T2i9, against a general mean of 0T058 in 1881. 
It should be stated that before taking the samples all above-ground growth is 
carefully cut off by scissors and removed; and that in the preparation of the soil- 
samples for analysis, all roots, indeed all visible vegetable debris, is carefully picked 
out; so that the results only include the nitrogen of that part of the crop-residue 
which has become thoroughly disintegrated, and may be considered as a proper 
constituent of the surface mould. It maybe further stated that the separated residue 
from the leguminous crop soils contained more nitrogen than that from the wheat- 
fallow soil. 
Going a little more into detail, it is seen that the Trifolium rip)ens soil shows 
a mean of 0T140 per cent, nitrogen in 1882, of 0T128 per cent, in 1883, and of 0T269 
per cent, in 1885. That is to say, the lowest percentage is in 1883 when there had 
been no growth, when there had been a whole season for the disintegration and nitri¬ 
fication of the residue of the previous year, and wdien 146 lbs. of nitrogen as nitric 
acid were found to the depth of 108 inches, and more than four-fifths of it below the 
surface soil. On the other hand, the percentage is the highest in 1885, when nearly 
100 lbs. of nitrogen had recently been removed in the crop, and the crop-residue 
would be comparatively large. 
In the Melilotus leucantha soil somewdrat more nitrogen was found in 1885, wdiich 
was the eighth year of continuous crop, than in 1882, which was only the fifth season, 
but which yielded more than twice the amount of nitrogen in the crop, and left 
considerably more visible and separated residue in the surface soil. 
Lastly, of the Medicago sativa soils, we have samples only in the sixth year of the 
growth, which had rapidly increased to an enormous amount in the fifth year, 1884, 
and yielded very large, though somewhat less, produce in 1885. Under these 
circumstances the mean percentage of nitrogen in the surface soil is 0T219, or higher 
than in the case of any other plant or year, excepting in the Trifolium repens soil of 
the same year. This is the case notwithstanding that more visible crop-residue had 
been separated from the Medicago sativa soil samples than from any ol‘ the others ; 
and in fact about three times as much as from the Trifolium reiJens soils of the same 
year. Indeed, it w’as estimated that the separated residue from the Medicago sativa 
soil-samples, the nitrogen in which was determined, represented a removal of about 
100 lbs. of nitrogen per acre. 
Without relying on the exact figures as representing exact gains or losses of nitrogen 
by the surface soils, we think it will be granted that the results are too consistent to 
leave any doubt that by the growth of the leguminous crops the surface soils had 
gained nitrogen, and that this gain bore some relation to the amount of growth and 
removal in the crops. 
MDCCCLXXXIX.—B. D 
