CARBON AND NITROGEN CYCLES IN 'I HE SOIL 195 



decomposed by bacteria in dilute solutions saturated with 

 oxygen there is no loss of nitrogen. Russell and Richards 

 (241^) were also unable to find any loss of nitrogen in the 

 aerobic fermentation of urine or faeces, even though in the 

 case of urine considerable nitrification occurred. 



The present position of the problem, therefore, is that this 

 loss of nitrogen does not occur under anaerobic conditions, 

 nor under aerobic conditions ; it requires a combination of both, 

 such as obtains in manure heaps, in rich soils and in the system 

 of beds and filters Jn a sewage works. 



These facts are inconsistent with either a direct oxidation 

 or a direct reduction process ; both Wagner's and Immendorf s 

 hypotheses, therefore, are ruled out. They are consistent with 

 Miintz's alternate nitrification and denitrification hypothesis, 

 but they do not prove it. Russell and Richards have given a 

 more general explanation, which seems to be free from some 

 of the mechanical objections that can be urged against alter- 

 nate nitrification and denitrification. They suppose that 

 molecular groupings arise under anaerobic conditions which 

 become unstable as soon as aerobic conditions set in, and 

 decompose, splitting off nitrogen. The change is parallel to 

 the shortening of the propionic acid chain with formation of 

 an acetic acid derivative, which is known to take place in these 

 conditions. 1 



Although the investigations have been mainly on manure 

 heaps and sewage beds there is evidence that the same process 

 occurs in heavily manured soils and in virgin soils rich in 

 organic matter brought into arable cultivation, and that it 

 causes the losses already discussed on p. 1 8 1. 



The Fixation of Nitrogen. 



The first systematic search for a recuperative agency to 

 make good the losses of nitrogen from the soil was started 

 thirty years ago by Berthelot. He found that certain organic 



* See Barger, The Simple Natural Bases (Longmans' Biochemical Mono- 

 graphs), where various instances are given. 



13* 



