[Vol. S 



4 ANNALS OF THE MISSOURI BOTANICAL GARDEN 



iron, aluminium, silicates, and phosphates proved valueless in 

 their hands, but a faintly alkaline mannite solution containing 

 iron and aluminium phosphates appeared promising. Follow- 

 ing this out through a series of experiments they finally suc- 

 ceeded in preparing a nutrient medium which contained, in 

 addition to the regular constituents of Beijerinck's nutrient, 

 a so-called "ferric hydroxide solution," with cane sugar as 

 a protection against precipitation ("Fallingsschutz") by al- 

 kalis. This iron-containing solution was prepared by dissolv- 

 ing in 1 liter of water 1 gm. of FeCU, 10 gms. cane sugar, and 

 0.80 gm. iron-free NaOH. Two drops of H 2 S0 4 were then 

 added. This final solution contained 0.50 gm. Fe 2 0:? and 0.26 

 gm. free NaOH per liter. Numerous distinctly beneficial re- 

 sults are reported from the addition of Fe 2 03 to Beijerinck's 

 nutrient solution in the above form. For instance, when 15 

 mgs. Fe 2 03 were added to 100 cc. of Beijerinck's solution 7.88 

 mgs. of nitrogen were fixed by Azotobacter per gm. of mannite 

 in 2 weeks, whereas ordinarily only 1 to 2 mgs. were assim- 

 ilated per gram of mannite. 



By increasing the amount of Fe 2 (>3 Remy and Rosing state 

 that better results are obtained, although they unfortunately 

 omitted the nitrogen determinations for these larger amounts 

 of iron. The other substances, NaCl, cane sugar, and NaOH, 

 present in the so-called ferric hydroxide solutions, were with- 

 out effect, hence these workers felt that the following conclu- 

 sions were justified: 



(1) A relation exists between the iron content of the nutri- 

 ent solution, Azotobacter development, and nitrogen fixation. 

 In the case of ferric silicate the optimum lies above 10 mgs. 

 Fe 2 03 per 100 cc. of Beijerinck's mannite solution. 



(2) The value of the iron varies according to its form. 

 The most favorable is an alkaline solution in which iron hy- 

 droxide is dissolved by means of cane sugar. Then follows 

 ferric silicate, while all other iron compounds stand far below. 

 Thomas phosphate acts strongly, which is to be attributed in 

 part to the content in silicic acid and basic lime. 



In discussing the results of their work Remy and Rosing 

 point out that the action of the iron cannot be that of a nutri- 



