The samples should be taken as described in paragraph =65=. The locality and date are written on a label attached to the tube. The tube contains about sixty or seventy grams of soil, and the moisture determination is made on this in the laboratory in the usual way. It would be desirable to have this sample represent a depth of from six to nine inches, thus rejecting the surface three inches which are more liable to sudden and accidental changes. These tubes are very inexpensive, and a sufficient number should be purchased to keep each station supplied. The sample represents a definite depth, and it does not have to be subsampled or even transferred in the field. This record of the moisture of the soil will show the amount of moisture which the different soils can maintain at the disposal of the plants, which, together with the temperature of the soil, is believed to be a most important factor in crop distribution and development. =295. Method of Berthelot and André.=—The estimation of the water according to Berthelot and André[189] should be made under three forms; _viz._, 1. Water eliminated spontaneously at ordinary temperatures. 2. Water eliminated by drying to constant weight at 110°. 3. Water eliminated at a red heat. The water may be determined directly on a sample weighed at the time of taking and afterwards dried in the open air, and finally, if necessary, in a desiccator. For a general idea the desiccation should be made on a sample of 100 grams, for exact work on ten grams. The dish in which the drying takes place should be shallow, and during the time the sample should be frequently stirred and thoroughly pulverized with a spatula which is weighed with the dish. The drying in the air should continue several days. The data obtained are not fixed since they depend on the temperature and the degree of saturation of the air with aqueous vapor. The variations due to these causes, however, are not very wide. The process may be regarded as practically finished when successive weights sensibly constant are obtained. In this state the soils contain very little water eliminable at 110°. =296. Estimation of Water Remaining after Air-Drying.=—The sifted sample is placed in quantities of five or ten grams in a flat-bottomed dish and dried at 110° to constant weight. This treatment not only removes the moisture, but all matters volatile at that temperature. Petermann,[190] in the Agricultural Station, at Gembloux, practices drying the sample to constant weight at 150°. It is further recommended by Petermann to determine total volatile and combustible matters by igniting to incipient redness, allowing to cool, moistening with distilled water, and drying at 150°.