To each of these he added a small quantity of mineral food--the required phosphorus, potassium, iron, sulphur, and so on. To the first he supplied no nitrogen at all in any form; the second he fertilised with saltpetre, which is largely composed of nitrogen in a form in which plants may readily absorb it through their roots; the third of the jars he inoculated with some of his bacteria culture. Then he planted beans in all three jars, and awaited the results, as may be imagined, somewhat anxiously. Perfectly pure sterilised water was supplied to each jar in equal amounts and the seeds sprouted, and for a week the young shoots in the three jars were almost identical in appearance. But soon after that there was a gradual but striking change. The beans in the first jar, having no nitrogen and no inoculation, turned pale and refused to grow, finally dying down completely, starved for want of nitrogenous food, exactly as a man would starve for the lack of the same kind of nourishment. The beans in the second jar, with the fertilised soil, grew about as they would in the garden, all of the nourishment having been artificially supplied. But the third jar, which had been jealously watched, showed really a miracle of growth. It must be remembered that the soil in this jar was as absolutely free of nitrogen as the soil in the first jar, and yet the beans flourished greatly, and when some of the plants were analysed they were found to be rich in nitrogen. Nodules had formed on the roots of the beans in the third or inoculated jar only, thereby proving beyond the hope of the experimenter that soil inoculation was a possibility, at least in the laboratory. With this favourable beginning Professor Nobbe went forward with his experiments with renewed vigour. He tried inoculating the soil for peas, clover, lupin, vetch, acacia, robinia, and so on, and in every case the roots formed nodules, and although there was absolutely no nitrogen in the soil, the plants invariably flourished. Then Professor Nobbe tried great numbers of difficult test experiments, such as inoculating the soil with clover bacteria and then planting it with beans or peas, or vice versa, to see whether the bacteria from the nodules of any one leguminous plant could be used for all or any of the others. He also tried successive cultures; that is, bean bacteria for beans for several years, to see if better results could be obtained by continued use. Even an outline description of all the experiments which Professor Nobbe made in the course of these investigations would fill a small volume, and it will be best to set down here only his general conclusions. [Illustration: Trees Growing in Water at Professor Nobbe's Laboratory.]