If now the aforesaid first sediment be suspended in pure water, and allowed to subside for the same time as before, the larger part of it will be again deposited, but some will remain in suspension, consisting of a considerable part of the finer matter of the first sediment. By pouring off these suspended particles with the water and agitating the sediment again with clear water as before, another portion of fine particles will be suspended and may be decanted from it. On continuing this process of repeated decantations it will soon be found that the soil has been separated into two grades. It is evident that in this way a separation can be made, but it is perhaps not so clear that such a separation would be sharp enough for the purposes of a mechanical soil analysis. If, for instance, the separation is to be made at 0.05 millimeter diameter, it is evident that by repeated decantations all below 0.01 millimeter can be washed out of that above 0.05 millimeter, but it may not appear so probable that all below 0.045 millimeter can be removed without removing some above 0.055 millimeter. Such a result may be easily attained, however, if the following principle be adhered to: Make the duration of the subsidence such that the liquid decanted the first few times shall contain nothing larger than the desired diameter. Then decant into another vessel, timing the subsidence so that the sediment shall contain nothing smaller than the chosen diameter. This can not be done without decanting much that is larger than the chosen diameter, but the greater part of the particles greater and less than the chosen diameter can be removed and an intermediate product obtained, the diameters of whose particles are not very far from that desired. If this intermediate portion be again subjected to the same process, two fractions may be separated from it, one containing particles larger than the chosen diameter and another containing particles smaller than this diameter, while a new intermediate product will remain which is less in amount than that resulting from the first operation. By frequent repetitions of this process this intermediate product can be reduced to a very small amount of substance the particles of which have diameters lying close to the chosen limit and may then be divided between the two fractions. The principles of the separation described by Osborne set forth with sufficient clearness the purposes to be achieved by the analysis. The chief methods of manipulation practiced will be found below. =199. Kühn’s Silt Cylinder.=—A simple form of apparatus for the determination of silt by the sedimentation process is the one described by Kühn.[134] The cylinder should be about twenty-eight centimeters high with a diameter of 8.5 centimeters. At the lower end of the cylinder five centimeters from the bottom it carries a tube 1.5 centimeter in diameter furnished with a pinch-cock and held in position by a rubber stopper.