BV K. GREIG-SMITH. 



501 



.111(1 ix., and are given in the following table, wliich also includes 

 the notes and fertility-numbers* that accompanied the soils. 



There is, to a certain extent, a parallel between the fertility 

 and the number of Rhizobia in the soil, and if we omit the virgin 

 soil and No. 4, which is exceptional in other ways, the numbers 

 expx'essing the ratios are virtually the same. Thus, so far as we 

 know, the number of Rhizobia in agricultural soils is proportional 

 to the fertility. 



The effect of cultivation in increasing the numbers of Rhizobia 

 is shown in Nos.l and 2; the si)il of No. 4 was subsequently found 

 to be very toxic, and this probably accounts for the absence of 

 Rhizobia. It improved greatly after heavy rains. 



The main object of the research, however, was the preparcition 

 of a medium selective enough to enable the Rhizobia to be deter- 

 mined, and this has been done. These Rhizobia are presumably 

 nitrogen-gatherers. On several occasions, the bacteria were 

 tested to see if they were nitrogen-destroyers, such as is Vibrio 

 denitrijicans, but no evolution of gas in nitrate bouillon was ever 



* To prevent any unconscious bias during the research, regarding the 

 numbers of lihizohia in tiiese soils, I opened the envelope containing the 

 description and fertility-numbers after the averages had been made. 



