_Unhumified Organic Matter does not Nitrify._—There can be little doubt that the formation of ammonia from the amido-compounds in humus is also the work of bacteria; but this, really the initial phase of the nitrogen-nutrition of plants, has not yet been fully elucidated. That, however, it is essentially only the ready-formed humus and not the unhumified debris of the soil which participate in nitrification was shown by the experiments of the writer, see chapter 19. [Illustration: FIG. 17.—Bacillus denitrificans I. (Burri.)] _Denitrifying Bacteria._—Among the sources of _loss of nitrates_ in the soil is the action of denitrifying bacteria; some of which cause merely the reduction of nitrates to nitrites and progressively to ammonia, while others cause gaseous nitrogen to be given off from nitrites and nitrates, resulting in their complete loss to the soil. While there are probably several kinds of the latter class, the most rapidly effective is an organism contained abundantly in fresh horse dung, and also on the surface of old straw. This can readily be shown by subjecting a very dilute solution (1-3 per cent.) of Chile saltpeter to the action of fresh horse dung in a close flask, when nitrogen and carbonic dioxid gases are evolved, and in a few days the nitrate has totally disappeared. In the course of time this power of horse-manure disappears; so that “rotted manure” is practically free from it and under proper conditions serves nitrification so effectively, that in the past it has served extensively for the production of saltpeter in the “niter-plantations” for the industrial purposes; the material of which was loose earth, marl and manure, kept moist and frequently forked over for better aeration. Saltpeter is similarly produced in stables, corroding the mortar of brick foundations. Nevertheless, it is necessary to avoid the use, either together or at short intervals apart, of Chile saltpeter and fresh manure; the manure if used first should be allowed to remain at least two months in the soil before saltpeter is applied. The reduction of nitrates to nitrites and ammonia is brought about by quite a number of bacteria, mostly anaerobic, and such as consume combined oxygen in their development. Thus the butyric ferment, which in the absence of readily reducible compounds evolves free hydrogen, will in presence of nitrates reduce the latter to nitrites, or form ammonia by addition of hydrogen to nitrogen just set free by reduction. Such reductive processes of course occur chiefly in soils rich in organic matter, or ill-aerated. The ammonia so formed, while at first simply combining with any humus acids present, may in the course of time be itself reduced to the amidic condition, being thereby rendered relatively inert, until again brought into action by ammonia-forming bacteria.