BY R. GREIG-SMITH. 617 



that a fixation of nitrogen had occurred. Such proved to be the 

 case, for duplicate determinations showed a gain of 0000392 grm. 

 of nitrogen in 20 c.c. of medium, which is equal to 0-01 % of 

 asparagin. The test to which - 01 % of asparagin had been added 

 showed a total over the blank tests of - 000742 grins., which is 

 equivalent to 002 °/ o of asparagin, or a gain of - 01 %. A gain 

 of nitrogen by the mixed culture is therefore established. 



Meanwhile the other components of the mixed culture had been 

 separated. They were identified as Bad. radiobacter and Bac. 

 levaniformans. Beijerinck and v. Delden* examined a similar 

 combination of bacteria, but instead of using Bac. levaniformans 

 they employed other spore-bearing bacteria, viz., varieties of 

 Granulobacter and Bac. mesentericus vulgatus. The various races 

 of Bac. levaniformans differ so greatly in their growth-characters 

 that, were it not for the characteristic product, gum levan, being 

 so easily detected, one would conclude that species of bacteria 

 were being dealt with instead of races of one bacterium. It is 

 extremely probable that Beijerinck and v. Delden's Granulobacter 

 is really Bac. levaniformans, the more so as Granulobacter, like 

 Bac. levaniformans, does not form slime upon glucose-agar by 

 itself, but produces a luxuriant slime in combination with the 

 other bacteria. These authors ascribe the fixation of the nitrogen 

 to Azotobacter, and not so much, if at all, to the others, and from 

 their experiments this appears to be the case. 



The quantity of slime formed by the pure culture of Azotobacter 

 did not appear to confirm this view, and an experiment! was' 



* Oentrbl. f. Bakt. 2te Abt. ix.(1902), 1. 

 t The medium contained dextrose 2%, sodium phosphate 0*2%, 

 and agar 2%. 20 e.e. of this medium were poured into Petri-dishes, 

 and, after setting, the medium was smeared with cultures of the bacteria. 

 The slimes were scraped off one set at the end of seven days and the 

 plates were reinfected. All the slimes were scraped off the plates at the end 

 of a month after the primary infection, when nitrogen determinations were 

 made by the Kjeldahl-Gunning process. Due allowance was made for the 

 nitrogen in the original media and for the errors of experiment. 



48 



