BY R. GHEIG-SMITH. 503 



that all artificial media are more or less selective. The numbers 

 ol)taine(l per grni. of soil will, therefore, generall}' be low. Lipman 

 says that ricli loam soils may contain the enormous total of from 

 15 to 20 millions per grm. Conn found that the bacteria rangerl 

 up to 22 millions in sunnner, and 3.3 millions in winter. Tf we 

 assume, however, that there were 25 millions in the Hawkesbury 

 soil No.l, we realise that the nitrogen-gathering Rhizobium is 

 pre.sent to the extent of 10 { of the total bacteria. 



In order to get some iletiriite information, however, two soils 

 were obtained fi'c^m the Hawkesbury Agricultural College in 

 May: one was designated "good," the other "medium." The 

 total bacteria were determined by sowing dilute suspensions upon 

 ordinary nutrient agar, and the Bhizobia upon levulose-citrate- 

 agar. The "good" soil contained 4,160,000 bacteria, and 145,600 

 Rhizobia per grm. of dry soil; that is, the Rhizobia were present 

 to the extent of 3-5% of the so-called total bacteria. In the 

 "medium" soil, there were 4,858,000 total bacteria, and 20,200 

 Rhizobia per grm., or 0-4%. 



In a garden-soil. Mi-. C. W. H. Powell, a student in this labor- 

 atory, found 12,000,000 bacteria per grm., and 830,000 Rhizobia, 

 or 6-75%. 



Sum^yiart/. 



1 . The number of nitrogen-fixing Rhizobia in agricultui^al soils 

 varied up to three or four millions per gi'am. 



2. The numbers present aftbrd an indication of the comparative 

 fertility of the soil. 



