On the Rain and Drainagc-JVaters at Rothamsted. 
63 
into the pipes than under the average conditions with spring 
sowing ; and so far as this is so, there will be less disappearance 
due to diffusion downwards and drainage other than through 
the drain-pipes. It would be expected, therefore, that the 
drainage collected after autumn sowing would, as a rule, more 
directly represent the supplies by manure than would that after 
spring sowing. After spring sowing, active vegetation com- 
mences, and the conditions of temperature and evaporation are 
less favourable for immediate and direct drainage through the 
pipes. Indeed, notwithstanding the conditions may be more 
favourable for return upwards by diffusion or capillary action, 
it is to be supposed that there will, upon the whole, be more 
gradual passage of nitrates downwards by diffusion, and bv 
drainage other than through the pipes. 
There is, in fact, no doubt that the estimates of loss of 
nitrogen bv drainage, founded on the composition of the col- 
lected drainage-waters, are frequently too low, and it may be 
that the whole of the otherwise unaccounted-for amounts are to 
be thus explained. 
On the other hand, it has long ago been shown bv Reiset, 
Boussingault, Ville, and at Rothamsted, that free nitrogen is 
frequently evolved in the decomposition of nitrogenous organic 
substances ; and in the case of the farmyard-manure plot, with 
its enormous accumulation of organic matter, and comparativelv 
little ascertained loss by drainage, there is a verv large amount 
of the estimated supplied nitrogen unaccounted for ; and there 
can be little doubt that here there has been considerable evolu- 
tion as free nitrogen. Of the artificially-manured plots it is 
those which receive the greatest excess of nitrogen in manure 
which retain the largest amount of vegetable debris; and this is 
a condition which, if sufficiently developed, is favourable for 
the evolution of free nitrogen : and so again, the greater un- 
accounted-for amount in the case of the highly-manured plots 
might in part be explained. But, independently of a possible 
loss by the evolution in the free state of the nitrogen of decom- 
posing organic matter within the soil, it has to be considered 
whether there may not be a loss of free nitrogen by the reduc- 
tion of nitric acid within the soil. Schlosing found such a 
destruction and evolution in experiments in which the access of 
gaseous oxygen was excluded, and the soil was saturated with 
moisture. He concluded that the combustion of organic matter 
in the soil had taken place at the cost of the oxvgen of the 
nitrate. In Part II., vol. xvii. p. 332, of this paper, similar results 
are recorded. In a water-logged soil, destitute of free oxvgen, 
nitrates were reduced by the organic matter of the soil, carbonic 
acid was evolved, and there was a considerable loss of nitric 
