This widespread power of producing ammonia makes it impossible in our present knowledge to regard any particular group of organisms as _par excellence_ promoters of fertility. Indeed, it is safest not to attempt to do so. The primary purpose of the activities of a soil organism is to obtain energy and cell material for itself; any benefit to the plant is purely incidental. For cell material it must have nitrogen and phosphorus; here it competes with the plant. If it produces more ammonia than it utilises--in other words, if it is driven to nitrogen compounds for its energy, then the plant benefits. If, on the other hand, it absorbs more ammonia than it produces, as happens when it derives its energy from non-nitrogenous substances, the plant suffers. Thus, addition of peptone to the soil or an increase in bacterial numbers effected without addition of external energy (e.g. by partial sterilisation) leads to increased ammonia supply, and, therefore, to increased fertility. But addition of sugar to the soil causes so great an increase of numbers of bacteria and other organisms that considerable absorption of ammonia and nitrate occurs, and fertility is for a time depressed. Both actions proceed in soils partially sterilised by organic substances, such as phenol, which are utilised by some of the soil organisms; there is first a great rise in numbers of these particular organisms with a depression of ammonia and nitrate, then a drop to the new level, higher than the old one, and an increased production of ammonia and nitrate resulting from the partial sterilisation effects. We must then regard the soil population as concerned entirely to maintain itself, and only incidently benefiting the plant, sometimes, indeed, injuring it; always essential, yet always taking its toll, and sometimes a heavy toll, of the plant nutrients it produces. This effect makes it difficult to deduce simple quantitative relationships between bacterial activity and soil fertility, and the difficulty is increased by the fact that bacteria and plants may both be injured or benefited by the same causes, so that high bacterial numbers in a fertile soil would not necessarily be the cause, but might be simply the result of fertility. The circumstance that certain soil organisms--bacteria, algæ, and fungi--themselves assimilate ammonia and nitrate may account for the remarkable slowness of nitrate accumulation, to which reference has already been made. The protein formed from the assimilated nitrogen remains in the bodies of the organisms, living or dead, till decomposition sets in. It is not difficult to picture a cycle of events in which much of the nitrate formed is at once reabsorbed by other organisms, and only little is actually thrown off into the soil. Such a process might continue almost interminably so long as any carbonaceous material remained. Finally, we come to the very interesting problem--is it possible to control the population of the soil?