502 CONTRIBUTIONS TO OUR KNOWLEDGE OF SOIL-FERTILITY, ii., 



observed. Finally, the nitrogen-fixing power was determined in 

 order to make certain that RJiizobium le<jumi')iosarum Jiad been 

 dealt with. 



Two plates were taken, at times wliicli chanced to be con- 

 venient, and from these, certain colonies were taken. The only 

 recommendations for the selection were, that the colonies should 

 be Khizohia, and that they should be free from moulds. Freedom 

 from moulds was not a general character of the colonies hy the 

 time that they were finally identified as Rhizobia, for the moulds 

 run over the surface of tlie plate, and contaminate the bacterial 

 colonies during the time that the giant colonies ai'e being grown 

 for identification. However, ten colonies were taken from a 

 plate of Expt. viii., containing thirteen colonies of Rhizohia, and 

 five colonies from a plate of garden-soil in Expt. ix., which had 

 ten colonies. 



The races were grown on saccharine media for several weeks, 

 and finally were smeared over plates of dextrose-agar.* 



The Fixation of Nitrogen by Rhizobia. 



With regard to the number of bacteria in soil, there will 

 naturally be great variations; and although we may obtain 

 certain results from their enumeration, it must be remembered 



'For technique, see these Proceedings, 1906, p. 608. 



