Be it noted, however, that I am not now expressing my own opinion. There are weighty reasons against thus identifying climatic variations with good species--reasons which will be dealt with in the next chapter. Kerner does not seem to appreciate the weight of these reasons, and therefore I do not call him as a witness to the subject as a whole; but only to that part of it which has to do with the great and general importance of climatic variability in relation to diagnostic work. And thus far his testimony is fully corroborated by every other botanist who has ever attended to the subject. Therefore it does not seem worth while to quote further authorities in substantiation of this point, such as Gärtner, De Candolle, Nägeli, Peter, Jordan, &c. For nowadays no one will dispute the high generality and the frequently great extent of climatic variation where the vegetable kingdom is concerned. Indeed, it may fairly be doubted whether there is any one species of plant, whose distribution exposes it to any considerable differences in its external conditions of life, which does not present more or less considerable differences as to its characters in different parts of its range. The principal causes of such climatic variation appear to be the chemical, and, still more, the mechanical nature of soil; temperature; intensity and diurnal duration of light in spring and summer; moisture; presence of certain salts in the air and soil of marine plants, or of plants growing near mineral springs; and sundry other circumstances of a more or less unknown character. Before closing these remarks on climatic variation in the vegetable kingdom, prominent attention must be directed to a fact of broad generality and, in relation to our present subject, of considerable importance. This is that the same external causes very frequently produce the same effects in the way of specific change throughout large numbers of _unrelated_ species--i.e. species belonging to different genera, families, and orders. Moreover, throughout all these unrelated species, we can frequently trace a uniform correlation between the degrees of change and the degrees to which they have been subjected to the causes in question.