=7. Collecting Samples.=—When possible, the analyst should be his own collector. There is often as much danger from data obtained on non-representative samples as from imperfect manipulation. When personal supervision is not possible, the sample when received, should be accompanied by an intelligible description of the method of taking it, and of what it represents. In all cases the object of the examination must be kept steadily in view. Where comparisons are to be made the methods of collecting must be rigidly the same. The processes of analysis, as conducted with agricultural products, are tedious and difficult. The absolutely definite conditions that attend the analysis of mineral substances, are mostly lacking. The simple determinations of carbon, hydrogen, nitrogen and sulfur, which are required in the usual processes of organic analysis, are simplicity itself when compared with the operations which have to be performed on agricultural products to determine their character and their value as food and raiment. We have to do here with matters on which the sustenance, health and prosperity of the human race are more intimately concerned than with any other of the sciences. This fact also emphasizes the necessity for care in collecting the materials on which the work is to be performed. =8. Grinding Samples.=—In order to properly conduct the processes of agricultural analysis it is important to have the sample finely ground. This arises both from the fact that such a sample is apt to contain an average content of the various complex substances of which the material under examination is composed, and because the analytical processes can be conducted with greater success upon the finely divided matter. In mineral analysis it is customary to grind the sample to an impalpable powder in an agate mortar. With agricultural products, however, such a degree of fineness is difficult to attain, and moreover, is not necessary. There is a great difference of opinion among analysts respecting the degree of fineness desirable. In some cases we must be content with a sample which will pass a sieve with a millimeter mesh; in fact it may be found impossible, on account of the stickiness of the material, to sift it at all. In such cases a thorough trituration, so as to form a homogeneous mass will have to be accepted as sufficient. Where bodies can be reduced to a powder however, it is best to pass them through a sieve with circular perforations half a millimeter in diameter. A finer degree of subdivision than this is rarely necessary.