Hilgard recommends the following procedure: The clay water from the previous separation is placed in a cylindrical vessel of such a diameter as to allow the column of water to be 200 millimeters high where it is allowed to settle for twenty-four hours. When the clay is very abundant a longer time may be allowed; _viz._, from forty to sixty hours. The line of separation between the dark silt below and the translucent clay above is sharply defined. Finally the clay water is decanted and the remaining liquid poured off leaving the sediment as sharply defined as possible. The sediment is rubbed with a rubber pestle and a few drops of ammonia water added. Distilled water is added, the beaker well shaken or stirred to break up the floccules that may have formed and subsidence permitted as before. This operation is repeated from six to nine times until the water remains quite clear after subsidence or the decanted turbid water fails to be precipitated by brine showing the suspended matter to be fine silt and not clay. The diameter of the particles of silt thus obtained is from 0.001 to 0.02 millimeter, and it is impossible to obtain it quite free from any admixture with clay. =227. Estimation of the Colloidal Clay.=—The importance of the colloidal constituent of the clay is such as to make its direct determination desirable. The volume of the clay waters at this stage of the analysis may amount to twenty liters. One method of determination consists in evaporating an aliquot portion and this method will yield good results if the sample be free from soluble salts and the quantity taken be not too small. At least 500 cubic centimeters should be used for this purpose. A better method consists in precipitating the clay by means of a saline solution. A saturated solution of salt is recommended for this purpose of which fifty cubic centimeters are sufficient to precipitate the clay from one liter of the clay water. The precipitation is hastened by heating. Each portion of the clay water should be precipitated as soon as obtained, the total volume of the precipitate at the end of twenty-four hours is thus reduced to a minimum. The clay water from the succeeding separations of the same analysis can be mixed with the precipitate which diffuses therein, thus promoting the precipitation of the rest of the clay inasmuch as the separation takes place more readily where more clay is present. When all the clay is thus collected it can be gathered on a tared filter and washed with weak brine. Pure water may not be used because of the diffusibility of clay therein. After drying at 100° and weighing it is washed with a weak solution of ammonium chlorid until all sodium is removed. The filtrate is evaporated to dryness, ignited at low redness, and weighed. The weight of the sodium chlorid thus obtained plus the weight of the filter deducted from the total weight gives the weight of the clay precipitate. Whenever the clay