ing more vigorously at the expense of the other," and that free nitrogen is appropriated by the microbes.
In 1883 Hellriegel began experiments with leguminous plants in pots of washed quartz sand, to which no nitrogen was added. Marked differences were observed in the growth of the plants under these conditions, but tubercles were found on the roots of the plants that made the best growth, while they were ab sent in other cases. He was then led to attempt the production of the root-tubercles by seeding or inoculating sterilized sand with a water-extract of a soil in which leguminous plants were growing. To some of the pots, in which peas and vetches were planted, from twenty-five c. c. to fifty c. c. of a water-extract of a fertile soil were added. When this soil-extract was not sterilized, there was a luxuriant growth of the plants in the pots to which it was applied, with abundant formation of root-nodules; but when the soil-extract was sterilized, this result was not obtained.
This soil-extract, however, was without effect on lupines and some other plants; but when the lupine pots were inoculated with an extract of a soil in which lupines were growing, the plants made a luxuriant growth, and root-tubercles were abundantly developed. In all cases the nitrogen supply of the plants was coincident with the development of root-tubercles, that were produced by inoculation with the extract of a fertile soil.
In 1888 a preliminary series of experiments, on the same lines, were begun at Rothamsted by Sir John B. Lawes and Prof. J. H. Gilbert; and in 1889 they were continued, on a more extended scale, with modified conditions suggested by the results of the preceding year. Their first experiments were made with peas, blue lupines, and yellow lupines, in pots seven inches high and about six inches in diameter. For our present purpose we need only call attention to the experiments in 1888 with peas.
Pots 1, 2, and 3 were filled with a washed yellow sand, to which was added 0*5 per cent of the ash of pea plants to furnish the required mineral constituents. Pot 4 was filled with a rich garden soil. Distilled water was used for watering the plants, and no other application was made to pot 1. Care was taken to determine the nitrogen of the soils, and of the seeds planted, which we need not describe in detail.
An extract of a rich garden soil was prepared by shaking in a stoppered bottle one part of soil with five parts of distilled water, and, after the coarser particles had subsided, twenty-five c. c. of the liquid was applied to each of pots 2 and 3. A chemical analysis of this soil-extract showed that the amount of plant food contained in it was so small that it could be safely neglected as an element of plant growth, and that its effect must be attributed solely to the soil microbes it contained.
