the middle of the boxes kept the roots of the weeds separate from those of the corn. When the partition was present, each sort of plant was confined to half the soil in the box, but, where lacking, one sort of roots appeared to exercise a noxious influence upon the other.
In studying the general question of soil fertility, the laboratories of the Bureau of Soils in the United States Department of Agriculture have obtained some instructive evidence bearing upon this problem of plant excretions. The data obtained from different lines of experimentation all go to prove the truth of the general assumption that plants do excrete substances which may have a toxic action on the species of plants producing them.
The Bureau of Soils has recently published the outcome of some experiments in which several sets of wheat plants were grown in rapid succession on the same soil. The soil was kept in pots and each set of wheat plants was smaller than the one immediately preceding, as might have been predicted, since it is well known that the continuous culture of a given crop on a soil "exhausts" it more rapidly than the culture of diversified crops. In the experiments cited, the growth of the fourth set of plants was in some cases only 30 per cent, of the growth of the first set.
Such cases of apparent exhaustion have generally been assumed to be due to the removal of plant nutrients from the soil by previous crops. Accordingly, experiments were made in which mineral plant nutrients in the form of pure chemicals were added to the soil at the time of planting the successive crops in amounts equal or greater than those removed by the preceding crop. The addition of these supplies of plant food constituents failed, however, to produce a growth of plants equal to the first crop, although they helped the growth of the plants. This possibility of soil exhaustion is still further rendered improbable by the fact that the wheat plants grew only three weeks and during that time they derived much of their food supply from the reserve nutrients stored in the seed; hence the amounts taken from the soil were very small.
To emphasize the action of root excretions upon the soil, experiments were made in which the first set of wheat plants was allowed to grow for only five days, or until the plumules were just beginning to appear above the surface of the soil in the pots, a length of time obviously insufficient to "exhaust" the soil. All the plants were carefully removed and the soil again planted with wheat, simultaneously with the same number of pots of fresh soil. The plants following the five-day crop made only 80 per cent, of the growth of the plants in fresh soil.
The outcome of these experiments is quite strong proof that the cause of these decreased yields is not the depletion of mineral nutri-
