To take a specific illustration. The octopus, or devil-fish, belongs to a widely different class of animals from a true fish; and yet its eye, in general appearance, looks wonderfully like the eye of a true fish. Now, Mr. Mivart pointed to this fact as a great difficulty in the way of the theory of evolution by natural selection, because it must clearly be a most improbable thing that so complicated a structure as the eye of a fish should happen to be arrived at through each of two totally different lines of descent. And this difficulty would, indeed, be a formidable one to the theory of evolution, if the similarity were not only analogical but homological. Unfortunately for the objection, however, Darwin clearly showed in his reply that in no one anatomical or homologous feature do the two structures resemble one another; so that, in point of fact, the two organs do not resemble one another in any particular further than it is necessary that they should, if both are to be analogous, or to serve the same function as organs of sight. But now, suppose that this had not been the case, and that the two structures, besides presenting the necessary superficial or analogical resemblance, had also presented an anatomical or homologous resemblance, with what force might it have then been urged,--Your hypothesis of hereditary descent with progressive modification being here excluded by the fact that the animals compared belong to two widely different branches of the tree of life, how are we to explain the identity of type manifested by these two complicated organs of vision? The only hypothesis open to us is intelligent adherence to an ideal plan or mechanism. But as this cannot now be urged in any comparable case throughout the whole organic world, we may on the other hand present it as a most significant fact, that while within the limits of the same large branch of the tree of life we constantly find the same typical structures modified so as to perform very different functions, we never find any of these particular types of structure in other large branches of the tree. That is to say, we never find typical structures appearing except in cases where their presence may be explained by the hypothesis of hereditary descent; while in thousands of such cases we find these structures undergoing every conceivable variety of adaptive modification.