438 INVESTIGATION BY CULTURE EXPERIMENTS 



limestone soil, will furnish the very favorable conditions, although 

 the data thus far reported from Agdell field (Table 75) show a 

 greater average loss of nitrogen (245 pounds from the surface 9 

 inches only) from the phosphorus plots than from the untreated 

 plots (105 pounds) in the legume rotation; while, as an average of 

 the four plots, the fallow rotation lost less nitrogen than the legume 

 system. 



It should be kept in mind, also, that the organic matter of the 

 soil contains nitrogen as well as carbon, and that the amount of 

 combined nitrogen liberated from this organic matter may be 

 nearly or quite sufficient to meet the needs of the bacteria that can 

 be supported by the carbonaceous food. In laboratory cultures 

 the fixation of nitrogen amounts to about 10 milligrams for each 

 gram of sugar (mannite) consumed by the "free-living" bacteria. 1 

 Thus the amount of nitrogen fixed is equal to about i per cent of 

 the carbonaceous food consumed; whereas the organic matter of 

 the soil contains, as a rule, more than 2 per cent of nitrogen. 



On Broadbalk field the nitrogen 2 content of the surface 9 inches 

 decreased during 28 years (1865 to 1893) by 285 pounds (from 2722 

 to 2437) on plot 3 (unfertilized), by 265 pounds (from 2782 to 2517) 

 on plot 5 (minerals), and by 63 pounds (from 3034 to 2971) on 

 plot 7 (minerals and 86 pounds of nitrogen) ; while the only in- 

 creases shown are 633 pounds (from 4343 to 4976) on plot 2 (farm 

 manure), and 131 pounds (from 2991 in 1865 to 3015 in 1881 and 

 to 3122 in 1893) on plot. 14, which receives ammonium salts (86 

 Ib. N), acid phosphate, and magnesium sulfate. (The possibility 

 of erosion or deposit from surface washing should not be over- 

 looked. Compare the nitrogen content of plots u, 12, 13, and 14 

 with respect to each stratum, as shown in Table 73.) 



1 In this connection attention is called to the point that if increased growth of 

 plants is caused by the use of pyrogallol, as reported from the unverified experi- 

 ments of Whitney and Cameron, it may be due to the fixation of free' nitrogen by 

 the nonsymbiotic bacteria that find in pyrogallol a suitable carbonaceous food 

 supply. It is known that the addition of sugar to ordinary soil deficient in nitrogen 

 will increase the growth of nonleguminous plants because of the increased nitrogen 

 fixation by the "free-living" bacteria. 



J All of these determinations were made by the older soda-lime method and 

 are considered trustworthy for comparison, but the 1893 analyses reported in Table 

 73 were made by the newer Kjeldahl method, which gives somewhat higher and 

 more nearly correct results. 



