A hole should be dug, and the soil and subsoil on one side removed to the depth at which the observation is to be made. A column of the soil or subsoil, two inches or more square, and four or five inches deep, is then to be carved out with a broad bladed knife, or a small saw can be conveniently used for cutting this out. A glass or metal frame, a little larger than the sample and three or four inches deep, is slipped over the column of soil, and melted paraffin is run in slowly to fill up the space between the soil and the frame. The soil is then struck off even with the top and bottom of the frame, preferably with a saw, or at any rate taking care not to smooth it over with a knife, which would disturb the surface and affect the rate of flow. The frame is then placed upon some coarse sand or gravel, contained in a funnel, to prevent the soil from falling out and to provide good drainage for the water to pass through. Another similar frame can then be placed on top and secured by a wide rubber band. A little coarse sand, which has been thoroughly washed and dried, is then placed on the soil, and water carefully poured on until it is level with the top of the frame. When the water begins to drop from the funnel more water must be added to the top, so as to have the initial depth of water over the soil the same in all the experiments. A graduated glass is then pushed under the funnel, and the time noted which is required for a quantity of water to pass through the soil. The quantity usually taken for measurement is equivalent to one inch in depth over the soil surface. In taking the sample, root and worm-holes are to be avoided, and these are particularly troublesome in clay lands. =178. Measurement of Percolation through the Soil in Situ.=—If lateral translocation could be prevented, the measurement of the quantity of water descending in the soil through a given area would be a matter of simplicity. But to secure accurate results all lateral communication of a given body of soil with adjacent portions must be cut off. Various devices have been adopted to secure this result. An elaborate system of lysimetric measurements is illustrated by the apparatus erected by the Agricultural Experiment Station, of Indiana. The plan and section of the apparatus are shown in Fig. 23.