Mr. Darwin therefore is in the following dilemma. If he does not treat natural selection as a cause of variation, the 'Origin of Species' will turn out to have no _raison d'etre_. It will have professed to have explained to us the manner in which species has originated, but it will have left us in the dark concerning the origin of those variations which, when added together, amount to specific and generic differences. Thus, as I said in 'Life and Habit,' Mr. Darwin will have made us think we know the whole road, in spite of his having almost ostentatiously blindfolded us at every step in the journey. The 'Origin of Species' would thus prove to be no less a piece of intellectual sleight-of-hand than Paley's 'Natural Theology.' If, on the other hand, Mr. Darwin maintains natural selection to be a cause of variation, this comes to saying that when an animal has varied in an advantageous direction, the fact of its subsequently surviving in the struggle for existence is the cause of its having varied in the advantageous direction--or more simply still--that the fact of its having varied is the cause of its having varied. And this is what we have already seen Mr. Darwin actually to say, in a passage quoted near the beginning of this present book. When writing of the eye he says, "Variation will cause the slight alterations;"[349] but the "slight alterations" _are_ the variations; so that Mr. Darwin's words come to this--that "variation will cause the variations." There does not seem any better way out of this dilemma than that which Mr. Darwin has adopted--namely, to hold out natural selection as "a means" of modification, and thenceforward to treat it as an efficient cause; but at the same time to protest again and again that it is not a cause. Accordingly he writes that "Natural Selection _acts only by the preservation and accumulation_ of small inherited modifications,"[350]--that is to say, it has had no share in inducing or causing these modifications. Again, "What applies to one animal will apply throughout all time to all animals--_that is, if they vary, for otherwise natural selection can effect nothing_"[351]; and again, "for natural selection only _takes advantage of such variations as arise_"[352]--the variations themselves arising, as we have just seen, from variation. Nothing, then, can be clearer from these passages than that natural selection is not a cause of modification; while, on the other hand, nothing can be clearer, from a large number of such passages, as, for instance, "natural selection may be _effective_ in _giving_ and _keeping_ colour,"[353] than that natural selection is an efficient cause; and in spite of its being expressly declared to be only a "means" of modification, it will be accepted as cause by the great majority of readers.