loo SOIL CONDITIONS AND PLANT GROWTH 



the action of this organism and that of azotobacter is that it gets its 

 carbohydrates from the plant, and is, therefore, independent of soil 

 organic matter. Thus, it operates perfectly well in the poorest soils 

 provided potassium salts, phosphates and calcium carbonate are present 

 in sufficient quantity for the host plant, while azotobacter (except where 

 it is associated with algae, a case that requires further investigation) 

 requires a supply of organic matter in the soil, and therefore only works 

 in fairly rich soils where its effects are more difficult to measure. 



Few improvements in agriculture have produced more marked effects 

 than the extension of leguminous cropping. Where a new leguminous 

 crop is being grown for the first time it may be necessary to introduce 

 the appropriate organism, as has been successfully done in Canada by 

 Harrison and Barlow (126), and on the North German moors by Hilt- 

 ner (135 ; see also 217 and 218). But, in general, these inoculations 

 have not proved useful, and they have never come into farming prac- 

 tice : the high hopes sometimes entertained that the whole problem of 

 nitrogenous manuring the most costly item in the farmer's fertiliser 

 bill might reduce itself solely to bacterial inoculation have never been 

 realised. The problem is extraordinarily fascinating and never fails to 

 arouse immense popular interest, but it is elusive and treacherous, and 

 has raised more false hopes and led to more disappointments than any 

 other in our subject. 



Denitrification. 



If the air supply of the soil is cut off by water logging, or in the 

 laboratory by means of an air pump, the nitrates rapidly disappear, 

 whilst nitrites, ammonia, or gaseous nitrogen are formed. The con- 

 ditions can be so arranged that the decomposition of nitrate-bouillon 

 by soil shall give rise to notable quantities of gaseous nitrogen, nitrous 

 oxide (18 and 279), or nitric oxide (168). 



This decomposition of nitrates has long been known. The reduction 

 to nitrite was shown by Meusel in 1875 ( ! 98) to be bacterial, since it 

 could be stopped by antiseptics. The property appears to be generally 

 possessed by bacteria and was shown by no fewer than 85 out of 109 

 kinds investigated by Maassen (191). 



The formation of gaseous products is effected by a smaller but still 

 considerable number of organisms ; these were first investigated by 

 Gayon and Dupetit (101-102), and by D6herain and Maquenne (81). 



The physiological significance of the reduction appears to be that 

 nitrates can supply oxygen to the organisms when free oxygen is no 

 longer obtainable. It is not simply a reaction between the organism 

 and the nitrate : easily oxidisable organic matter must be present at 



