24 THE FERTILITY OF THE SOIL [ch. 



e.g. wherever vegetation is allowed to die back. It 

 has been found by actual measurement at Rothamsted 

 that nitrogen does accumulate in soil left to run wild 

 and to cover itself with the varied assortment of 

 plants cropping up in these conditions. How much 

 of this is due to Azotobacter is not certain, because 

 leguminous plants occur among the herbage and fix 

 an unknown quantity of nitrogen. 



The organisms that we have been considering 

 represent the constructive agencies in the nitrogen 

 cycle in the soil, bringing in new supplies from the 

 air and so making good the losses already discussed. 

 It is necessary to remember — this point will con- 

 stantly recur in future chapters — that these con- 

 structive processes only manifest themselves in soils 

 covered ivith permane^it vegetation such as grass- 

 land, woodland, etc. One exception only is known, 

 viz. where a leguminous crop is growing, when the 

 amount of nitrogen fixed may be considerable. With 

 this exception one general rule holds : losses of 

 nitrogen preponderate on soil that is cultivated, and 

 gains of nitrogen preponderate on soils covered with 

 permanent vegetation. In either case the action 

 does not go on indefinitely: the losses become less 

 and less as the soil becomes poorer, till finally they 

 are so small that it is difficult to detect them, and 

 the gains also become less and less as the soil be- 

 comes richer, till finally they also cease or are 



