The changes indicated in the amount of nitrogen in the soil 

 are not large, being, indeed, very close to the experimental error 

 but, reviewing the numbers altogether, it may safely be concluded 

 that the first plot, IJ / X 4, from which the roots are carted and where 

 no clover is grown, is declining in fertility. The two plots 9 / IO 

 and IJ /,g are practically stationary; if anything, 9 / IO without clover 

 may be losing ground, whereas IJ / l6 with clover, but from which 

 the roots are carted, may be gaining ground; while the plot Sl / I2t 

 on which both the roots are returned and clover is grown, is still 

 more probably gaining a little fertility. This last plot has yielded 

 on the average over 52 years 35 bushels of wheat, 34 bushels of 

 barley, over 9 tons of swedes, and nearly 2 T / 2 tons of clover hay 

 per acre, which is nearly equal to, if anything higher than, the 

 average production of the whole of Great Britain during that period. 

 Such a yield, which, though equal to the average, may be taken 

 as lower than a good fanner would expect from that class of land, 

 has been obtained without bringing in any external source of ni- 

 trogen, without even returning to the land all that would come 

 back under ordinary conditions of farming. In practice it would 

 not always be possible to feed off the root crop on the land, and 

 even then not so much carbonaceous material would be returned 

 as is the case in the experiment, where it has been necessary to 

 cut up the roots and plough them in, but, on the other hand, the 

 manure made from the straw of both the wheat and barley crops 

 and from the clover hay would also come back to the land. 



The evidence provided by this plot is strengthened by the 

 results obtained on the three other plots, on which the recuperative 

 operations of clover growing, and returning the roots to the land, 

 are either singly or together omitted; under such conditions the 

 gross production is distinctly less, and the fertility of the land is 

 stationary or declining very slowly, so that an equilibrium at a 

 lower level of production has been or will shortly be attained. 



We may then conclude from these Agdell Field results that a 

 conservative system of farming on the four-course system, in which 

 clover is grown at least once in every two rotations, in which the 

 roots are consumed on the land, and the dung made by the straw 

 and hay comes back to the land, will maintain the fertility of the 

 soil and support for an indefinite period a gross production at about 

 a 4 qr. of wheat per acre level without any necessity for importing 

 nitrogen. The natural agencies of nitrogen fixation due to the 

 growth of the clover crop and the bacteria depending on the supply 



