The History of a Field neivly laid down to Permanent Ch-ass. 11 
soil have been derived from vegetable residue, partly of the 
manure applied, but in great part also of the crops grown ; 
and it has undergone decomposition, and thereby lost most of 
its carbon, whilst what remains exists in more stable com- 
bination. 
Snch accumulations are retained in the soil with great 
tenacity, and they contribute to its fertility. From the un- 
manured plot of the Ilothamsted permanent grass experiments, 
taking first and second cuttings together, 43 ci'ops of hay have 
been removed during the last 33 years ; but the greater part of 
the long previously accumulated fertility still remains. When, 
however, grass-land is broken up and converted into arable, the 
accumulated fertility rapidly diminishes. 
. When it is borne in mind that all our arable soils were 
originally covered with natural herbage or timber, and that 
during centuries of arable culture much of the previously accu- 
mulated fertility has thereby been removed, it will readily be 
understood that, in the re-conversion of such arable land into 
permanent grass, much of the used-up accumulated fertility 
must be restored, before we can get it back into a satisfactory 
condition. 
Assuming these to be the general conditions essential for 
the conversion of arable into permanent grass-land, we may 
now endeavour to ascertain what changes have actually taken 
place in the soil, during the formation of the meadow, the 
history of which we have under consideration. Any estimate 
of the accumulations which have taken place can, however, 
only be approximately correct, since we have not any exact 
data on the important point of the amount of nitrogen which 
the soil contained when it was first taken in hand in 1856 as 
arable land. 
But we have made some thousands of determinations of 
nitrogen in the surface-soils and subsoils of the Rothamsted 
experimental fields, which are very near to this newly formed 
meadow ; and we are in a position, therefore, to make a fair 
estimate on the subject. 
Careful samples were, however, taken in January 1879 — that 
is, when no crop had been removed since the previous July — and 
again in September 1888, after a second crop of hay had been 
taken. It will be instructive to go into some detail as to the 
methods of experimenting adopted, and the results obtained, as 
the discussion will illustrate some of the difficulties which are 
encountered in endeavouring properly to interpret analyses of 
soils made on samples taken at different periods, and in different 
conditions of the land. 
