656 CONTRIBUTIONS TO OUE KNOWLEDGE OF SOIL-FERTILITY, vi. 



But I have taken care to use soil that had not been overheated, and 

 to have controls of un filtered soil-extracts to compare with the fil- 

 tered, presumably protozoa-free, extracts. Taken as a whole, my 

 experiments show that Russell's contention cannot be sustained; 

 the protozoa have little or no action in limiting the number of 

 soil-bacteria. This is in agreement with Fred,* who, in one experi- 

 ment upon the nitrification of compost-soil, found a slight gain in 

 the test which had been treated with ether; he ascribed this to the 

 stimulating effect of the disinfectant, rather than to the destruc- 

 tion of phagocytes. 



The action of the soil-baeteriotoxins has been little investigated, 

 althougli their effect has been known for a considerable time. Mar- 

 tin, for example, found that typhoid bacteria disappeared from 

 raw soil in two days, but persisted for over a year in sterilised 

 soil. Althougli ascribed to the competition of other bacteria, there 

 can be little doul)t that the typhoid bacteria were destroyed by the 

 bacteriotoxins in the raw soil. When some kinds of actively-grow- 

 ing bacteria are added to soil, they rapidly increase, and then die 

 down. For example, Bac. prodigiosus, is at the height of its 

 growth betwen the second and third day at 28°; after that, the 

 numbers rapidly fall away. In experiments with soil bacteria, the 

 height of the rise is generally greatest at a later period, on account 

 of the smaller number of bacteria at the start, and also on account 

 of the slower growth of the natural microbes. In plate-cultures, 

 the toxic influence of bacteria, such as Bact. putidum, is readily 

 seen, for when many are present upon a plate, few colonies of 

 other bacteria develop. Thus it comes about that the weaker dilu- 

 tions show a greater number of bacteria than the stronger, for. in 

 the presence of relatively fewer colonies of Bact. putidum, the 

 other bacteria are not inhibited. The actual diminution of bacteria, 

 by the products of Bac. prodipiosus, was numerically shown in a 

 previous paper. t 



The spore-producing soil-bacteria, such as Bac. rulgatus, Bac. 

 suhtilis, and Bac. mycoides, which resist the action of the volatile 



*r'entrall). f. Bakt., 26 Abt.. xxxi., 23^. 

 t These I'roceedings, 1911, p.GSG. 



