(7) Samples which, on pulverizing, change their content of water, must have the water content estimated in both the coarse and powdered condition and the results of the analysis must be calculated to the water content of the original coarse substance. =15. Special Cases.=—Many cases arise of such a nature as to make it impossible to lay down any rule which can be followed with success. As in almost every other process in agricultural chemistry the analyst in such cases must be guided by his judgment and experience. Keeping in view the main object, _viz._, to secure in a few grams of material a fair representation of large masses he will generally be able to reach the required result by following the broad principles already outlined. In many cases the details of the work and the adaptations necessary to success must be left to his own determination. =16. Drying Samples of Fertilizers.=—The determination of the uncombined moisture in a sample of fertilizer is not an easy task. In some cases, as in powdered minerals, drying to constant weight at the temperature of boiling water is sufficient. In organic matters containing volatile nitrogenous compounds, these must first be fixed by oxalic or sulfuric acid, before the desiccation begins. If any excess of sulfuric acid be added, however, drying at 100° becomes almost impossible. Particular precautions must be observed in drying superphosphates. In drying samples preparatory to grinding for analysis, it is best to stop the process as soon as the materials can be pulverized. In general, samples should be dried only to determine water, and the analytical processes should be performed on the undried portions. It is not necessary, as a rule, to dry samples of fertilizers in an inert atmosphere, such as hydrogen or carbon dioxid. Drying in vacuo may be practiced when it is desired to secure a speedy desiccation or one at a low temperature. =17. Official Methods.=—The Official Agricultural Chemists direct, in the case of potash salts, sodium nitrate, and ammonium sulfate, to heat from one to five grams in a flat platinum or aluminum dish at 130° until the weight is constant.[9] The loss in weight is taken to represent the water. In all other cases heat two grams, or five grams if the sample be very coarse, for five hours in a steam-bath. In the German stations in the case of untreated phosphates and bone-black the moisture is estimated at from 105° to 110°. Samples which lose ammonia should have the weight of ammonia given off at that temperature, determined separately.