For instance, butter fat contains a large proportion of certain volatile compounds, which are either absent or are present in much smaller quantity in the fats used to adulterate butter; and thus an estimation of these volatile compounds affords a means of judging of the purity of the butter. Thus, if only half the normal quantity of volatile compounds is present, the conclusion is drawn that the butter is adulterated with an equal quantity of foreign fat, and so on. The task would not be difficult if butter fat were always constant in composition; but, unfortunately, there are often wide variations in the proportion of ingredients, and it frequently happens that the public analyst has to give his judgment upon a sample, which might either be a butter very rich in the characteristic volatile substances and adulterated with 10 per cent. of foreign fat; or it might be a genuine butter that was very deficient in these volatile compounds. This, then, is the dilemma. If the analyst condemn such a sample on the strength of this and other tests, he may be confronted by the evidence of other analysts who will give their opinion that the butter is genuine; and if, then, the matter be referred to the Government analysts, their report may or may not corroborate his, and in the latter alternative the authority instituting the prosecution may have to pay heavy costs. It is well known that butters are scientifically blended with foreign fats so as to fall just on the border line between abnormal and adulterated samples, and the analyst is frequently compelled to pass such a butter as genuine, lest he should unwittingly do an injustice. A large proportion of Dutch butter is abnormal in this respect, and has been so frequently condemned as adulterated by English chemists, that protests have been made by the leading Dutch analysts. The reason for the abnormal character of Dutch butter appears to be that the cows are left too long in the fields, for as soon as they are stalled for the winter, the character of the butter gradually becomes normal again. These details have been given at some length, for they are typical of the problem which the public analyst has to solve in the case of many natural products, _i.e._, to decide whether a food is adulterated or only naturally of poor quality. There is no special difficulty in the analyses; it is a question of interpretation of the results. The chief culprit in the matter of the adulteration of butter is the small dealer, who buys margarine from the margarine manufacturer and skilfully blends it with butter in a proportion that is small in a single instance, but is sufficient to bring him in a handsome profit in the course of a year. Owing to the difficulty of detecting such small additions of margarine to butter (which, as was explained above, is due to the variations in the natural product) a most ingenious device has been adopted in some countries.