EXPERIMENTS ON THE MIXED HERBAGE OF PERMANENT MEADOW. 379 
lation, and the annually renewed supply by manure, larger quantities of corn, or of 
straw, or of both, and also of hay, are every year obtained by the use (in conjunction 
with mineral manures) of much less than half as much nitrogen applied as ammonia- 
salts or as nitr ate of soda. The wheat plots so manured, and so yielding, at the same 
time show less than two-thirds as high a percentage of nitrogen in the first nine inches 
of depth. 
But, according to the estimates, besides the actually determined large, but compara¬ 
tively ineffective, residue within the soil, there was also in each case a very large 
amount unaccounted for, either in the increase of crop or in the soil, to the depths 
examined. The quantity to be estimated as lost would, of course, be less if the esti¬ 
mate of the amount supplied be too high ; again, by so much as may be retained as 
nitric acid, and not accounted for by the soda-lime method; and, again, by so much as 
may remain, either as nitric acid or in other forms, below the depth experimented on, 
but nevertheless not beyond the reach of root-collection, or of capillary action bringing 
up the stores from the lower to the upper layers. 
Direct experiments have shown that the soil in the experimental wheat field which 
is manured annually with farmyard manure retains near the surface, owing to its 
greatly increased porosity, very much more of the rainfall than the soil of the plots 
not so manured. It is, accordingly, found that the drain from the farmyard-manured 
plot runs much less frequently than do those from the unmanured or the artificially 
manured plots. There will, obviously, be less loss of water by drainage. Direct experi¬ 
ments further show, however, that a given volume of the drainage water which does 
run from the farmyard-manured plot contains from two to three or more times as 
much nitrogen as nitrates and nitrites as that from the unmanured plot, or from the 
plots with mineral wfithout artificial nitrogenous manure. We have here, therefore, 
a determined source of loss of the supplied nitrogen. But such calculations as the 
data admit of lead to the conclusion that the "whole of the estimated loss cannot be 
accounted for in this way. The probability is that there is a considerable additional 
loss by decomposition of the nitrogenous organic matter within the soil, and evolution 
as free nitrogen. 
There is, then, cumulative evidence to show T —that the nitrogen supplied as farmyard 
manure was recovered in very small proportion during the years of its application; 
that in after years it was recovered in constantly decreasing proportion ; that there 
nevertheless remained a considerable, but very slowly available, residue ; that there 
was a considerable loss of it by drainage; and, finally, that there is probably a further 
loss by decomposition, and evolution into the atmosphere. 
However significant these illustrations may be, it must be borne in mind that in 
ordinary agriculture much less farmyard manure would be applied than in these 
special experiments, and the losses by drainage would, from that cause alone, be propor¬ 
tionally less. Much, obviously, would also depend upon the character of the sod and 
the subsoil. Again, in an ordinary rotation of various crops, more of the supplied 
3 C 2 
