I90 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 



traced to the circumstance that the soil population is complex and is 

 not formed of bacteria only. The figures do not give the effect of 

 temperature or moisture supply on the whole soil population, but only 

 on part of it, and they indicate a competition or a destructive effect. 

 When the soil population is simplified by killing the less resistant 

 forms one obtains much more consistent results. Thus soil which has 

 been treated with mild poisons such as toluene shows the expected 

 increase in bacterial numbers with rise in temperature or moisture. 



These facts, and others which need not now be dealt with, indicate 

 that the soil bacteria are subject to the operation of some limiting 

 factor quite distinct from temperature, moisture content, or food 

 supply, and I have in other papers argued that this limiting factor is 

 to be found in the action of some of the less resistant and larger forms, 

 such as the protozoa, which keep down the numbers of the bacteria. 

 This hypothesis explains all the facts that have yet been ascertained, 

 but so many kinds of protozoa have now been discovered in the soil 

 that it is difficult to pick out the exact offenders and render the hypo- 

 thesis more precise from the purely zoological standpoint. We shall 

 find, however, that the simplest way of interpreting the phenomena 

 is to recognize the complex nature of the soil population and to admit 

 the possibility of competition among them, just as one has to admit 

 it in the case of any other population. We shall run into great diffi- 

 culties if we make the common mistake of supposing that all soil 

 organisms are- there for the express benefit of our plants and our 

 crops. 



The amount of nitrate production does increase with the tempera- 

 ture, and in this respect it differs from the numbers of bacteria. This 

 is in accordance with expectation ; up to a certain point an organism 

 may be expected to do more work as the temperature rises, but the 

 increase is not as great as one would get if the numbers increased as 

 well (fig. 62). 



When now we turn to the field conditions and try to follow the 

 production of nitrate in the soil, matters are complicated by the fact 

 that the nitrates produced do not all remain in the soil, but are liable 

 to be washed out or taken up by plants. Analytical determinations, 

 therefore, only give the difference between the amount formed and the 

 amount lost ; they do not show how much is actually produced, 

 nor give the rate of production that we desire to obtain. For some 

 time we could see no way of getting over the difficulty, but a simple 

 solution was ultimately found. It is evident that if the curves showing 

 the amount of some other substance produced in the same way as the 

 nitrate, but lost in a different way, are of the same general shape as the 

 nitrate curves, then the shape is due mainly to the production factors ; 

 if, on the other hand, the two sets of curves are different in shape, then 

 the loss factors control the situation. The carbon dioxide in the soil 

 air satisfies these requirements ; it is produced, like nitrates, by 

 bacterial action, but it is lost largely by gaseous diffusion, and only 

 in very wet weather by leaching. Carbon dioxide was therefore deter- 



