"Let us now turn to the second of these two assumptions, viz., that some at least among migratory birds must possess, by inheritance alone, a very precise knowledge of the particular direction to be pursued. It is without question an astonishing fact that a young cuckoo should be prompted to leave its foster parents at a particular season of the year, and without any guide to show the course previously taken by its own parents, but this is a fact which must be met by any theory of instinct which aims at being complete. Now upon our own theory it can only be met by taking it to be due to inherited memory." {234d} Mr. Romanes says in a note that this theory was first advanced by Canon Kingsley in _Nature_, January 18, 1867, a piece of information which I learn for the first time; otherwise, as I need hardly say, I should have called attention to it in my own books on evolution. _Nature_ did not begin to appear till the end of 1869, and I can find no communication from Canon Kingsley bearing upon hereditary memory in any number of _Nature_ prior to the date of Canon Kingsley's death; but no doubt Mr. Romanes has only made a slip in his reference. Mr. Romanes also says that the theory connecting instinct with inherited memory "has since been independently 'suggested' by many writers." A little lower Mr. Romanes says: "Of what kind, then, is the inherited memory on which the young cuckoo (if not also other migratory birds) depends? We can only answer, of the same kind, whatever this may be, as that upon which the old bird depends." {235} I have given above most of the more marked passages which I have been able to find in Mr. Romanes' book which attribute instinct to memory, and which admit that there is no fundamental difference between the kind of memory with which we are all familiar and hereditary memory as transmitted from one generation to another. But throughout his work there are passages which suggest, though less obviously, the same inference. The passages I have quoted show that Mr. Romanes is upholding the same opinions as Professor Hering's and my own, but their effect and tendency is more plain here than in Mr. Romanes' own book, where they are overlaid by nearly 400 long pages of matter which is not always easy of comprehension. The late Mr. Darwin himself, indeed--whose mantle seems to have fallen more especially and particularly on Mr. Romanes--could not contradict himself more hopelessly than Mr. Romanes often does. Indeed in one of the very passages I have quoted in order to show that Mr. Romanes accepts the phenomena of heredity as phenomena of memory, he speaks of "heredity as playing an important part _in forming memory_ of ancestral experiences;" so that whereas I want him to say that the phenomena of heredity are due to memory, he will have it that the memory is due to the heredity, {236a} which seems to me absurd.