[333] Vol. ii. chap. i. [334] Vol. ii. chap, xxxiv., ed. 1872. [335] 'Philosophie Zoologique,' ed. M. Martins, Paris, Lyons, 1873, tom. i. p. 24. [336] 'Philosophie Zoologique,' tom. i. p. 72. [337] 'Origin of Species,' p. 3. [338] Ibid. p. 5. [339] 'Origin of Species,' p. 8. [340] Ibid. p. 9. [341] Ibid. p. 158. [342] Ibid. p. 159. [343] 'Origin of Species,' p. 4. [344] 'Genesis of Species,' p. 74, 1871. [345] See _ante_, p. 330, line 1 after heading. [346] 'Genesis of Species,' p. 76, ed. 1871. CHAPTER XX. NATURAL SELECTION CONSIDERED AS A MEANS OF MODIFICATION. THE CONFUSION WHICH THIS EXPRESSION OCCASIONS. When Mr. Darwin says that natural selection is the most important "means" of modification, I am not sure that I understand what he wishes to imply by the word "means." I do not see how the fact that those animals which are best fitted for the conditions of their existence commonly survive in the struggle for life, can be called in any special sense a "means" of modification. "Means" is a dangerous word; it slips too easily into "cause." We have seen Mr. Darwin himself say that Buffon did not enter on "the _causes or means_"[347] of modification, as though these two words were synonymous, or nearly so. Nevertheless, the use of the word "means" here enables Mr. Darwin to speak of Natural Selection as if it were an active cause (which he constantly does), and yet to avoid expressly maintaining that it is a cause of modification. This, indeed, he has not done in express terms, but he does it by implication when he writes, "Natural Selection _might be most effective in giving_ the proper colour to each kind of grouse, and in _keeping_ that colour when once acquired." Such language, says the late Mr. G. H. Lewes, "is misleading;" it makes "selection an agent."[348] It is plain that natural selection cannot be considered a cause of variation; and if not of variation, which is as the rain drop, then not of specific and generic modification, which are as the river; for the variations must make their appearance before they can be selected. Suppose that it is an advantage to a horse to have an especially hard and broad hoof, then a horse born with such a hoof will indeed probably survive in the struggle for existence, but he was not born with the larger and harder hoof _because of his subsequently surviving_. He survived because he was born fit--not, he was born fit because he survived. The variation must arise first and be preserved afterwards.