=282. Freeing Fats Of Moisture.=—Any excess of water in glycerids will accumulate at the bottom of the liquid sample and can be removed by decanting the fat or separating it from the oil by any other convenient method. The warm oil may be almost entirely freed of any residual moisture by passing it through a dry filter paper in a jacket funnel kept at a high temperature. A section showing the construction of such a funnel with a folded filter paper in place, is shown in Fig. 81. The final drying, when great exactness is required, is accomplished in a vacuum, or in an atmosphere of inert gas, or in the cold in an exsiccator over sulfuric acid. In drying, it is well to expose the hot oil as little as possible to the action of the air. Wherever convenient, it should be protected from oxidation by some inert gas or a vacuum. =283. Sampling for Analysis.=—It is a matter of some difficulty to secure a representative sample of a fat or oil for analytical purposes. The moisture in a fat is apt to be unevenly distributed, and the sampling is to be accomplished in a manner to secure the greatest possible uniformity. When the quantity of material is of considerable quantity a trier may be used which will remove a cylindrical or partly cylindrical mass from the whole length or depth. By securing several subsamples of this kind, and well mixing them, an average sample of the whole mass may be secured. Where the fat is found in different casks or packages samples should be drawn from each as described above. The subsamples are mixed together in weights corresponding to the different casks from which they are taken and the mass obtained by this mixture divided into three equal portions. Two of these parts are melted in a dish at a temperature not exceeding 60°, with constant stirring, and when fully liquid the third part is added. As a rule, the liquid fat retains enough heat to melt the added quantity. As soon as the mixed fats begin to grow pasty the mass is vigorously stirred to secure an intimate mixture of the water and other foreign bodies.[232] [Illustration: FIG. 81.—SECTION SHOWING CONSTRUCTION OF A FUNNEL FOR HOT FILTRATION.] In the case of butter fat the official chemists recommend that subsamples be drawn from all parts of the package until about 500 grams are secured. The portions thus drawn are to be perfectly melted in a closed vessel at as low a temperature as possible, and when melted the whole is to be shaken violently for some minutes till the mass is homogeneous, and sufficiently solidified to prevent the separation of the water and fat. A portion is then poured into the vessel from which it is to be weighed for analysis, and this should nearly or quite fill it. This sample should be kept in a cold place till analyzed.[233]