52 
SIR J. B. LAWES AND PROFESSOR J. H. GILBERT ON THE 
We trust, therefore, that M. Deherain will accept our comments in all friendliness, 
when we say that our own dearly bought experience leads us to believe that the above 
facts are quite sufficient to render approximately accurate quantitative estimates at 
the different periods impossible. From the results, it seems probable that the samples 
taken at the commencement of the experiments in 1875 were less comparable with 
those of 1878 and 1879, than were those of these later dates with those of 1881 and 
1885. The losses indicated were, indeed, in most cases, much less over the second 
period of 4 years ; a result which is, however, doubtless partly due to the fact that no 
manure was applied during’ that period. Another reason for concluding that the 
samples were less truly representative at the commencement in 1875, than afterwards, 
is that the percentage of nitrogen found at that date (0’204), is high for the depth 
stated, of arable soil in ordinary agricultural condition. Though, if the soil is naturally 
very rich, or if it had been treated otherwise than in ordinary agricultural practice, 
such a percentage is by no means impossible. The percentages of 0T5 and upwards, 
as afterwards found, are however quite as high as is usual in good, but long worked 
arable soil, which is only manured and cropped in the ordinary way. 
Then as to the amounts of nitrogen estimated to be gained by the soil to the depth 
of 35 cm. (=13'8 inches) by the growth of sainfoin for 5 years, and of mixed grasses 
for 2 years. They were, both on the plot where farm-yard manure had previously 
been ajaplied, and on that which had been unmanured over the 7 years from the 
commencement, more than was taken off in the crops of that period. This is certainly 
more than our own experience would lead us to expect. 
From his determinations of the nitrogen in the subsoils at different periods, 
M. Deherain concluded that the gains were not from that source. The percentage 
in the subsoils of the different plots varied however considerably; and on this point 
it may be stated that in the subsoils at Rothamsted, the variations which are quite 
independent of manuring and cropping, are so great on the same plot, that we have 
found it quite impracticable to make calculations as to loss or gain in the total 
nitrogen of the subsoils. 
Upon the whole we conclude, that certainly the estimated losses of the surface 
soils, and probably also the estimated gains, are higher than can possibly happen in 
practice; and that the results are due to the method of taking the samples of soil not 
being such as to ensure strictly comparable estimates at the different periods. At 
the same time there can be no doubt that there would be losses beyond those due to 
the removal of the crops, under the conditions in which losses were found; that is, 
when the land was under arable culture. Nor can there be any doubt that there 
would be gains m the surface soil when the land was laid down in sainfoin and mixed 
grasses ; and M. Deherain points out the practical significance of such facts. 
M. Deherain concludes that the loss of nitrogen by arable soil, that is by soil that 
is mechanically worked, is due to the slow combustion of the nitrogenous organic 
matter of the soil; the nitrogen being either evolved as free nitrogen, or oxidated 
