Most authorities would agree that all absorption of free nitrogen, if by means of bacteria, must be through the roots. As a matter of fact, legumes, especially when young, use nitrogen, like all other plants, derived from the soil. It has been pointed out that, unless the soil is somewhat poor in nitrogen, there appears to be but little assimilation of free nitrogen and but a poor development of root nodules.[48] The free nitrogen made use of by the micro-organism is in the air contained in the interstices of the soil. For in all soils, but especially in well-drained and light soils, there is a large quantity of air. Although it is not known how the micro-organisms in legumes utilise free nitrogen and convert it into organic compounds in the tissues of the rootlet or plant, it is known that such nitrogen compounds migrate into the stem and leaves, and so make the roots really poorer in nitrogen than the foliage. But the ratio is a fluctuating one, depending chiefly on the stage of growth or maturity of the plant. If the nodules from the rootlets of _Leguminosæ_ be examined, the nitrogen-fixing bacteria can be readily seen. The writer has isolated these and grown them in pure culture as follows: The nodules are removed, if possible at an early stage in their growth, and placed for a few minutes in a steam steriliser. This is advisable in order to remove the various extraneous organisms attached to the outer covering of the nodule. They may then be washed in antiseptic solution, and their capsules softened by soaking. When opened with a sterile knife, thick creamy matter exudes. On microscopic examination this is found to be densely crowded with small round-ended bacilli or oval bodies, known as _bacteroids_. By a simple process of hardening and using the microtome, excellent sections of the nodules can be obtained which show these bacteria _in situ_. In the central parts of the section may be seen densely crowded colonies of the bacteria, which in some cases invade the cellular capsule of the nodule derived from the rootlet. Aërobic and anaërobic pure cultures of these bacteria were made. In some cases these cultures very closely resembled the feathery growth of the bacillus of anthrax.