BY R. GREIG-SMITII. 681 



These authors agree with me in considering that the agricere 

 is derived from the remains of the organic matter of plants. 



With the removal or segregation of the agricere, the sokible 

 matter will more readily diffuse out from the soil-particles, 

 and as this may be divided into toxic and nutritive sub- 

 stances, the benefit derived from the removal will depend, in 

 the first place, upon the relative preponderance of the one 

 or the other of these. In experiments with soil-extracts, the 

 toxin* generally predominates while, in soils, the nutrients 

 preponderate. 



It is difficult to explain the difference between the behaviour 

 of the moist soil and of the extract. A plausible explanation 

 may be found in the relative quantities of water used in the 

 two methods, and in the faint acidityt of the soil. In the 

 soil-tests, the water varied from one-tenth to one-quarter of 

 the weight of the soil, just enough to make it damp ; while, in 

 the extracts, an equal weight of water was employed. With 

 the smaller proportion of water, the acidity of the soil would 

 not be so much weakened, and, taken in conjunction with the 

 longer exposure, more of the presumably less soluble nutri- 

 ents would be dissolved. This reasoning, however, does not 

 appear to hold. 



Twenty gr. portions of garden-soil were placed in deep 

 Petri dishes, and to each, 2 cc. of a suspension of Bar. jirodi- 



*I have used the word.s "toxin" and "toxins" indiscriminately. 

 Doubtless lliere are as many varieties of toxin as there are groups of bac- 

 teria in the soil. It is known that the toxins of one group may not be 

 toxic to another group, but, so far as these experiments go, Baa. prodigiosus 

 behaves to the soil-toxins and nutrients like the ordinary soil-bacteria, and, 

 since its growth is rapid and the colonies easily detected, it acts after the 

 manner of an indicator. 



t From the literature upon tho subject, one is led to believe that all 

 normal soils are faintly alkaline. Mr. R. S. Symnionds, of the Department 

 of Agriculture, drew my attention to the fact that the majorit\' of soils 

 are acid to litmus, and this holds for the soils I have tested in New South 

 Wales. The method of testing consists in stirring the soil with a small 

 quantity of water, thus forming a paste; and, after from 5 to 10 minutes 

 pressing a piece of litmus paper upon the mass, 



