1910.] Conservation of the Fertility of the Soil. 115 



of soil nitrogen when the growth of the plant is returned to 

 the soil. 



(2) Various bacteria are capable of bringing atmospheric 

 nitrogen into combination, and so increasing the stock of soil 

 nitrogen. They may either live in symbiosis with higher 

 plants (Pseiidomonas), or exist free in the soil (Azotobactet, 

 Clostridium), 



(3) Another group of bacteria in the process of breaking- 

 down organic matter liberate the nitrogen in the free state, 

 and so reduce the stock of soil nitrogen. 



(4) Natural drainage waters contain nitrates which have 

 been derived from the soil nitrogen by bacterial oxidation. 



(5) The rain annually contributes a certain amount of com- 

 bined nitrogen to the soil. The amount is greater in the 

 proximity of towns; the average amount at Rothamsted is 

 3*84 lb. per acre per annum, and other results would show that 

 this is a very representative figure for ordinary country air. 



In practice most of these factors giving rise to gain or loss 

 are at work together; which of them will predominate will 

 depend upon the style of farming and cultivation the land 

 receives. Some of the Rothamsted plots with their long- 

 recorded history afford an opportunity of estimating the inter- 

 play of the various factors. 



A. The simplest case to take is that of land under arable 

 cultivation when nothing is restored to the soil. The un- 

 manured plot on the Broadbalk wheatfield affords a useful 

 example, and we have figures which show the change in its 

 fertility between 1865 an d I ^93, a period of 28 years. 



Broadbalk, Plot 3. Nitrogen, lb. per acre. 



In soil, In soil, Loss in Added Removed Unaccounted 



1865. 1 %93- ' 2 8 years. by rain. by crop. for. 



2722 2437 285 107 428 +36 



In the first place it will be seen that the nitrogen in the soil 

 declines when the crop is wholly removed and no manure 

 is added, though the decline becomes slow after the first 

 large loss of condition has taken place. It will be remem- 

 bered that the yield of the unmanured wheat plot at Rotham- 

 sted fell off pretty quickly for a few years, but for the last 

 fifty years has remained almost constant at an average of 

 about 1 2 '5 bushels per acre. 



If we make out a balance-sheet and set off the nitrogen 

 removed in the crop against that which the soil has lost, as 



