The purpose of placing sand at the bottom of each lysimeter is to offer a porous stratum in which free water may collect and rise to the level of the perforated copper tube, which would prevent any further rise by conveying the surplus above into the vault as drainage water. The soil above the tube will therefore be constantly drained and the sand below constantly saturated, unless the water be drawn up by the capillary action of the soil as the result of evaporation from the surface.

By means of a proper arrangement within the vault, of a kind of Mariotte’s bottle, the water may be caused to flow back through the drainage tube into the lysimeter to take the place of that lost by evaporation, and thus maintain the level of free water just below the drainage tube. The water flowing back to the lysimeter, and the amount of drainage water, are carefully measured by a system of graduated tubes.

The lysimeters thus constructed represent tile-drained land; in one case the tile being three feet below the surface and in the other six feet below. The drainage waters collected in the receiving bottles can be measured and analyzed from time to time, as occasion may require, to determine the amount of plant food which is removed.

179. Improved Method of Deherain.^[120]—Deherain’s earlier experiments were made in pots containing about sixty kilos of soil. These vases serve very well for some kinds of plants, but there are other kinds which do not grow at all normally when their roots are imprisoned. For instance, in pots, even of the largest size, wheat is always poor, beets irregular, maize never acquires its full development, and the conclusions which can be drawn from the experiments can not be predicated of the action of the plant under conditions entirely normal.

It is necessary therefore to carry on the work in an entirely different way, and to construct boxes so large as to make the conditions of growth entirely normal. The arrangement of these boxes is shown in Fig. 24.

They are placed in a large trench, two meters wide, one meter deep, and forty meters long. There are twenty boxes in this trench, the upper surface of each containing four square meters area. The boxes are one meter deep, and therefore can contain four cubic meters of soil. The sides and bottoms of the boxes are made of iron lattice work, covered with a cement which renders them impervious to water.

The bottom inclines from the sides towards the middle, and from the back to the front, thus forming a gutter which permits of the easy collection of the drainage. The drainage water is conveyed, by means of a pipe and a funnel, into a demijohn placed in the ditch in front of the apparatus, as shown in the figure. These receptacles stand in niches under the front of the cases, and are separated by the brick foundations. Access to them is gained by means of the inclined plane shown in the figure, and this plane permits the demijohns in which the drainage water is collected, to be removed with a wheelbarrow for the purpose of weighing. This apparatus is especially suitable for a study of the distribution of the nitrogen to the crop, the soil and the drainage waters. The loss in drainage waters of potash and phosphoric acid is insignificant in comparison with the loss in nitrogen.

The cases having been placed in position they are filled with the natural soil, which is taken to the depth of one meter, in such a way that the relative positions of the soil and subsoil are not changed.

While the soil is transferring to the cases it is carefully sampled in order to have a portion representing accurately the composition of both the soil and subsoil. These samples are subjected to analysis and the quantities of nitrogen, phosphoric acid, and potash contained therein carefully noted.

One or two cases should be left without crop or fertilizer to determine the relations of the soil and subsoil to the rain-fall. Three or four cases should be kept free of vegetation and receive treatment with different fertilizer, in order to determine the influences of these on the deportment of the soil to rain-fall. The rest of the cases should be seeded with plants representing the predominant field culture of the locality, and some of them should be fertilized with the usual manures used in farm culture.