BY R. GREIG-SMITH. 



609 



The main point brought out by the experiment is that con- 

 currently with the formation of slime there is a fixation of atmos- 

 pheric nitrogen by Rhizobium. We already know that Azoto- 

 bacter can gain nitrogen, and indeed Stoklasa says that it is the 

 most powerful agent in this respect that the soil possesses. This 

 experiment shows that the combination of Azotobacter and Rhizo- 

 bium appears to be still more powerful. Bac. levaniformavs 

 cannot fix nitrogen, and although it does not form slime from 

 dextrose, yet it undoubtedly assists other bacteria to do so. Bad. 

 radiobacter is another bacterium of much the same nature. It is 



The Formation of Slime and Fixation of Nitrogen by Races of Rhizobium 

 alone and in combination with bac. levaniformans. 





Rhizobium alone. 



Rhizobium and 

 Bac. levaniformans. 



Race. 



Slime, 

 grms. 



Nitrogen 

 gained, 

 mgrms. 



Slime, 



grms. 



Nitrogen 

 gained, 

 mgrms. 



Macrozamia 



Robinia 



French Bean 



3 

 6 

 2 

 3 



o 



2 





 

 

 5 

 4 

 11 





 

 1 

 

 

 

 

 

 

 

 

 2 



5 



11 



1 



5 

 5 



3 



(1) 

 (1) 

 (1) 

 8 

 12 

 22 



2 

 2 

 1 



Tare 







Pea (March) 



1 



,, (October) 



Lupin(Kr;il) ... 



Pea ,, 



1 





 







Lupin "a" 



2 



)} "c" 



] 



,, "d" 



4 



most difficult to maintain it in pure culture; it soon dies out, 

 although in mixed culture it keeps well. Both of these bacteria 

 have assisted the Lupin race to form slime, and probably also to 



process, using tenth-normal alkali and methyl-orange for the final titration. 

 Due allowance was made for the nitrogen in the original medium + infecting 

 material by means of check tests which also controlled the error of reagents 

 and apparatus. The milligram of nitrogen per 100 c.c. of medium has been 

 taken as the unit. This is equivalent to - 15 c.c. of tenth-normal alkali in 

 the 20 c.c. test. Quantities less than 0i5 c.c. are expressed as 0, 0-15 to - 25 

 c.c. as 1, Q'25 to - 4 c.c as 2 and so on. The quantities are small, but the 

 method is sufficiently delicate, the check tests agreeing in three cases out of 

 four, the fourth differing by 0"05 c.c. 



47 



