CARBON AND NITROGEN CYCLES IN THE SOIL 95 



to Koch's experiments ceases to work at 7 C. Thiele read tempera- 

 tures daily for three years of arable and grass soils at different depths 

 at Breslau (282), and concluded that only rarely were they favourable 

 for azotobacter. But it is impossible to argue from a culture solution 

 to the soil, and indeed Lohnis has shown that the mixed cultures of the 

 soil are almost as effective at 10 as at 20 : 1 



IO-I2 C. 20-22 C. 30-32 C. 



3-15 mg. 4-55 mg. 4-27 mg. nitrogen fixed. 



It seems legitimate to conclude that azotobacter fixes nitrogen in 

 well-aerated soils sufficiently provided with calcium carbonate, potassium 

 salts and phosphates, carbonaceous material of the right kind and 

 moisture, so long as the temperature is high enough. Where the air 

 supply is diminished owing to the close texture of the soil there is still 

 the possibility of fixation by clostridium. Ashby (7) found that the 

 relative distribution of azotobacter and clostridium at Rothamsted de- 

 pended on the amount of calcium carbonate in the soil ; wherever any 

 notable quantity was present, azotobacter invariably occurred : other- 

 wise clostridium alone was found. It is not certain whether this result 

 is due to some specific action of calcium carbonate or to the shortage 

 of air supply consequent on the bad mechanical state always induced 

 at Rothamsted when calcium carbonate is absent. 



Nitrogen Fixation by Bacteria in Symbiosis with Leguminosae. 



After Hellriegel and Wilfarth's great discovery of the relationship 

 between bacteria and leguminosse (p. 16) many unsuccessful attempts 

 were made to isolate and study the organisms by the methods then in 

 vogue. In 1888 Beijerinck (n) broke away from the ordinary meat- 

 bouillon-gelatin plate and substituted a slightly acid medium made up 

 of infusion of pea leaves, gelatin (7 per cent), asparagine ('25 per 

 cent.) and sucrose (-5 per cent.). Growth readily took place and the 

 colonies yielded rods I p wide and 4 to 5 p long, some of which showed 

 signs of bacteroid formation, and "swarmers" 0*9 //, long and O'i8 //, 

 wide, these being among the smallest organisms known. 2 



None of these organisms, however, could be found in the soil, nor 

 indeed has any one yet succeeded in finding them there although their 

 existence cannot be doubted. Their mode of entry into the pea was 

 studied by Prdzmowski (228), and later by Nobbe and Hiltner (216). 

 The root hair is attacked, presumably by the " swarmer," and a filament, 



1 Mitt. Landw. Inst., Leipsic, 1905, vii., 94. 



2 Golding has shown that they will even pass through a porcelain filter and has pre- 

 pared pure cultures in this way. 



