While this state of things is primarily due to the intrinsic complexity and difficulty of the subject itself, it has unquestionably been materially aggravated by accidental, partly historic conditions. Foremost among these is the fact that until within recent times, soil studies have borne almost entirely on lands long cultivated and in most cases fertilized: thus changing them from their natural condition to a more or less artificial one, which obscures the natural relations of each soil to vegetation. The importance of these relations is obvious, both from the theoretical and from the practical standpoint. From the former, it is clear that the native vegetation represents, within the climatic limits of the regional flora, the result of a secular process of adaptation of plants to climates and soils, by natural selection and the survival of the fittest. The natural floras and sylvas are thus the expression of secular, or rather, millennial experience, which if rightly interpreted must convey to the cultivator of the soil the same information that otherwise he must acquire by long and costly personal experience. The general correctness of this axiom is almost self-evident; it is explicitly recognized in the universal practice of settlers in new regions, of selecting lands in accordance with the character of the forest growth thereon; it is even legally recognized by the valuation of lands upon the same basis, for purposes of assessment, as is practiced in a number of States. The accuracy with which experienced farmers judge of the quality of timbered lands by their forest growth, has justly excited the wonder and envy of agricultural investigators, whose researches, based upon incomplete theoretical assumptions, failed to convey to them any such practical insight. It was doubtless this state of the case that led a distinguished writer on agriculture to remark, nearly half a century ago, that he “would rather trust an old farmer for his judgment of land than the best chemist alive.”[1] [1] “The Soil Analyses of the Geological Surveys of Kentucky and Arkansas.” S. W. Johnson in _Am. Jour. Sci._, Sept. 1861. It is certainly true that mere physico-chemical analyses, unassisted by other data, will frequently lead to a wholly erroneous estimate of a soil’s agricultural value, when applied to cultivated lands. But the matter assumes a very different aspect when, with the natural vegetation and the corresponding cultural experience as guides, we seek for the factors upon which the observed natural selection of plants depends, by the physical and chemical examination of the respective soils. It is further obvious that, these factors being once known, we shall be justified in applying them to those cases in which the guiding mark of native vegetation is absent, as the result of causes that have not materially altered the natural condition of the soil.