It is a practice much to be desired, that not only should a proper choice be made of seeding examples, but that there be a change of situation, and, if possible, a time of storage before being planted for seed. These are all cultivative processes, and to the care with which they are carried out must we look for permanence in our derivative root-crops. It cannot be too strongly urged, that, as an efficient sort of root has only been arrived at as the result of great care—that is, by successful breeding,—so every care must be taken for its maintenance. Defect in seed results in defect in the produce of that seed; and downward tendencies of this kind are common results of even most careful cultivation. With carelessness in this respect we must not be surprised at rapid degeneracy. CHAPTER VI. ON THE ADULTERATION OF SEEDS, MORE PARTICULARLY OF TURNIPS. In order to make the experiments which illustrate this chapter tell their tale to the fullest extent, we would set out with the two following postulates:— 1st. All well-grown, well-preserved new seeds should be _capable_ of germinating to the extent of at least 90 per cent. 2nd. Seeds in general, and more especially turnip seeds, as usually delivered to the farmer, are generally _incapable_ of germinating to the extent of from 25 to 30 per cent., and very frequently even more. We shall hereafter see, that this want of germinating power is too often the result of mixing charlock, Indian rape, and the like, by way of adulteration, which latter are killed to prevent “their telling tales.” But to our experiments:— A number of tin cases were made of the following proportions: Length, 15 inches; width, 10 inches; depth, 4 inches. These, which were well perforated at the bottom, were divided across into ten equal parts, each of which was filled to within an inch of the rim, with a mixture of fine mould and silver sand. In these, seeds of different sorts of turnips were sown, and the whole was put into a bed of sand in our forcing-house. We could, however, see no difference in the results, nor could we trace any in the germinal or cotyledon leaves of swedes, turnips, or charlock. But, of course, samples of turnip-seed could not be tested as to freedom from charlock by this experiment, because charlock is killed before being mixed with the turnip. Now, seeing that we could get no trustworthy results by this kind of experiment, it struck us that our germination-pans might be used to test the germinating power, not only of the samples we had obtained for a different purpose, but of others also. We first, then, counted a hundred of each of the following sorts of seeds, and carefully dibbled them in a fresh mixture of soil, in September, 1860; the results, which were as carefully noted from day to day, are shortly given in the following table:— TABLE 1.—_Germination of Ten Sorts of Turnips._