The German experiment stations[3] estimate hygroscopic moisture for analytical calculations by drying to constant weight at 100°. In determining loss on ignition, however, the preliminary drying is made at 140°, with the exception of peaty samples where so high a temperature is not admissible. The Official Agricultural Chemists[191] place five grams of air-dried soil in a flat-bottomed and tared platinum dish; heat in an air-bath to 110° for eight hours; cool in a desiccator, and weigh; repeat the heating, cooling, and weighing, at intervals of an hour till constant weight is found, and estimate the hygroscopic moisture by the loss of weight. Weigh rapidly to avoid absorption of moisture from the air. In the German laboratories, according to König,[192] from ten to twenty grams of the fine earth, properly prepared by air-drying and sifting, for analysis, are heated at 100° to constant weight. For control, five grams are placed in a desiccator over sulfuric acid for two or three days. Wolff directs that a small portion of the well-mixed earth, for example, twenty grams, be spread out on a flat zinc plate, and its changes in weight observed through several days. These observations are continued until the variations are so slight that the means can be determined with sufficient exactness from the last weighings. The soil is then dried at 125° in a hot air-chamber. The loss in weight will give the mean hygroscopic moisture in the soil under the conditions in which the experiment is made. =297. Drying in a Desiccator.=—The sample dried as indicated previously by the method of Berthelot and André is placed in a desiccator over sulfuric acid. It is better to have the sample traversed by a current of perfectly dry air, and in this case it should be placed in a tube, which is closed while weighing, to prevent absorption of moisture. Much time is also required for this operation, and it does not possess the practical value of the method of drying in the free air. =298. Water Set Free at 110°.=—This is determined by Berthelot and André on a weight of five to ten grams of soil. The sample which has been employed for the preceding determination may be used. While this is going on in an air-bath heated to 110°, about ten times as much soil should be dried for the same time at the same temperature, and this should be preserved in a well-stoppered flask. All subsequent determinations are to be made with the soil dried at 110°. The loss of weight in a soil increases with the temperature to which it is exposed. The apparent quantity of water, therefore, determined at 140° or 180° is always greater than that obtained at 110°. But when the temperature exceeds 110° there is danger of decomposing organic bodies with the loss of a part of their constituent elements. Carbon dioxid and ammonia may also be lost, as well as acetic acid and other volatile bodies.