INVESTIGATIONS ON THE PRODUCTION OF PLANT FOOD. 195 



This is equally true of the minor events. Changes in conditions 

 do not cease to be effective as soon as the old conditions are restored ; 

 they leave their mark, which may persist for a long time and lead to 

 very unexpected results. Experiments at Rothamsted and elsewhere 

 have brought out the apparent paradox that conditions harmful 

 to life lead to greater bacterial activity as soon as they have passed, while 

 conditions favourable to life finally cause decreased bacterial activity. 



Population Bacteria in Bacteria in I 01 



of Great I 02. of ol PartiaJly 



Britain Sick Soil Sterilised Soil 



Fig. 65. — Comparison of the Number of Bacteria in Sick and 

 in Partially Sterilised Soils. 



Thus, if a soil is dried for a time and then re-moistened, it becomes 

 a better medium for the growth of plants and of bacteria, the produc- 

 tion of nitrates is increased, and the supply of phosphate becomes 

 more available. The bacterial numbers do not, however, undergo 

 any visible change. 



Productiveness of Soils stored dry. 

 (Oats : Gedroix, 1908.) 



No. of years of 

 Storage. 



No 

 Manure. 



Complete 

 Manure. 



Complete without 

 Nitrogen. 



Manure without 

 Phosphate. 



O 



10-3 



83'5 



I3'5 





I 



iy8 



83-9 



32'3 



19 



3 



24*6 



90-9 



236 



35'4 



5 



25-0 



102-8 



32-2 



[ 



42 



When soil is exposed to severe cold there is an increase in nitrate 

 production, and, in this case, in bacterial numbers also. Exposure to 



