612 THE FIXATION OF NITROGEN BY RHIZOBIUM LEGUMINOSARUM, 



The experiment was repeated, with the difference that no 

 sodium carbonate was added to the medium and the time 

 extended over 31 days. 



The last two experiments show that saccharose is as capable 

 as dextrose of assisting in the fixation of nitrogen. Further- 

 more, as the slimes and media were alkaline, the fixation can 

 occur in alkaline as well as in acid media. 



The fixation of nitrogen by certain races of Rhizobium has been 

 already proved by Maze,* but many investigators, f myself 

 included, could not confirm his results. Then Moore,! and 

 finally Lohnis,§ showed that a gain had taken place. The 

 quantities of nitrogen demonstrated were about the same as I 

 have found. Maze obtained up to 4 milligrams per cent, of 

 medium in 14 days; Moore of from 2 to 3 mgrms. in three weeks, 

 and Lohnis of from to 1*5 mgrms. in the same time. 



From my own experiments, it would appear that rhe failure of 

 investigators to detect a fixation has been caused either by using 

 races which were incapable of forming slime or by testing the 

 races under conditions which precluded slime-formation. The 

 fixation appears to be proportional to the amount of slime formed, 

 and although in my experiments the individual tests do not bring 

 this out clearly, yet if we take averages of all the tests with 

 Rhizobium alone and in combination with Bac. levaniformans, we 

 obtain numbers which indicate a certain relationship. 



The intimate relation between the fixation of nitrogen and the 

 formation of slime is best brought out in the last table because 

 the number of tests which are included in the averages are greater 

 and one irregularity is probably counterbalanced by another. 

 These irregularities may have been caused by (1) experimental 

 errors greatly influencing the small gains of nitrogen; (2) varia- 



* Maz6, Ann. de l'lnst. Past. xii.(1898). 

 tBeijerinck and v. Delden found no fixation of nitrogen with cultures of 

 Rhizobium + Granvlobacler (B. hvaniformans?). 



J Moore, Bulletin Bureau of Plant Industry, No.71(1905). 

 § Lohnis, Centrbl. f. Bakt. 2te, Abt. xiv.(190S), 582. 



