Certain of the cells in the centre of the nodule become greatly enlarged, and in the fully grown nodule are seen to be filled with bacteria. Differences have been described in the morphology of the organisms in different parts of the nodule.[62] Whether the different stages of the organism are equally capable of fixing nitrogen, or what is the significance of these stages within the nodule, is not certainly known. It has been held that it is the irregular bacteroid forms that are chiefly concerned with nitrogen fixation. In older nodules the organisms become irregular and stain faintly, and the bacteroidal tissue breaks down, the nodule finally decaying. In the fixation of nitrogen that occurs in the nodules, the bacteria without doubt derive the necessary energy from the carbohydrates of the host-plant. There is evidence that the plant assists the process of fixation by removing soluble metabolic products from the neighbourhood of the bacteria. Golding[22] was able to obtain a greatly increased fixation of nitrogen in artificial cultures by arranging a filtering device so as to remove the products of metabolism. The great practical importance of leguminous crops in agriculture has led to numerous attempts being made to increase their growth, and the fixation of nitrogen in them, by inoculating the seed or the soil with suitable nodule-bacteria. This inoculation can be effected either with soil in which the host-plant has been successfully grown, and which should consequently contain the organism in fair numbers, or else pure cultures of the organisms isolated from nodules may be used. Very varying results have been obtained with inoculation trials. In farm practice a leguminous crop has often been introduced into a new area where it has never previously grown. In such soil it is very probable that varieties of the nodule organism capable of infecting the roots may not exist. In such cases inoculation with the right organism or with infected soil often produces good results. The more difficult case, however, is that in which the legume crop has been grown for a long time in the locality, and where the soil is already infected with right organisms. This, the more fundamental problem, applies especially to this country. Here it would seem that inoculation with a culture of the organism will benefit the plant only (1) if the naturally occurring organisms are present in very small numbers; or (2) if the organisms in the culture added are more virulent than those already in the soil. The problem of successful inoculation would therefore seem to be bound up with that of grading up the infective virulence of the organism to a higher level.