The chemical process by which nitrogen is fixed is quite unknown, although a number of speculative suggestions have been made. The appearance of considerable amounts of amino acids in young cultures of _Azotobacter_ suggests that these may be a step in the process, but at present the data are too inconclusive to form a basis for theorising. _Azotobacter_ is very rich in phosphorus, an analysis of the surface growth in _Azotobacter_ cultures, made by Stoklasa, giving about 60 per cent. of phosphoric acid in the ash. In cultures it has been found that a considerable amount of phosphate is needed to produce full development. As would be expected, therefore, nitrogen fixation in soil is often greatly stimulated by the addition of phosphates. Christensen has, indeed, found soils where lack of phosphate was the limiting factor for _Azotobacter_ growth. _Azotobacter_ is very intolerant of an acid medium, and is very dependent on the presence of an available base. In cultures this is usually provided in the form of calcium or magnesium carbonate. Gainey[21] found that _Azotobacter_ occurred in soils having an acidity not greater than P_{H} 6·0, and Christensen,[7],[9] in Denmark, has found a close association between the occurrence of _Azotobacter_ in soils and the presence of an adequate supply of calcium carbonate. So close was this association that he devised a technique based on this fact for detecting a deficiency of lime in a soil sample. In addition to the groups already discussed, there is a remarkable and important group of nitrogen-fixing bacteria that inhabit and can carry on their functions within the root tissues of higher plants. It has been known at least from classical times that certain leguminous plants would, under suitable conditions, render the soil more productive. On the roots of leguminosæ small tubercles are commonly found. These were noted and figured by Malpighi in the seventeenth century, and for a long time were regarded as root-galls. As was described in Chapter I., the true nature of these tubercles was finally elucidated by Hellriegel and Wilfarth in 1886. As the result of a series of pot experiments, they made the very brilliant deduction that the ability to fix nitrogen, possessed by the legumes, was due to bacteria associated with them in the tubercles.