Hubrecht,1 and of the more recent references of Bateson in his British Association address. That saltation is a constant phenomenon in nature, a vera causa of evolution, no one can longer deny. Bateson shows that it harmonizes with Mendel's conceptions of heredity, and it may be regarded as par excellence the contribution of the experimental method. Similarly, I regard mutation as a quite distinct phenomenon, and as par excellence the contribution of the paleontological method; it is the gradual rise of new adaptive characters neither by the selection of accidental variations nor by saltation, but by origin in an obscure and almost invisible form, followed by direct increase and development in successive generations until a stage of actual usefulness is reached, where perhaps selection may begin to operate. While clearly setting forth the difficulties, I at one time attributed definite variation or mutation to Lamarck's principle of the inherited effects of habit as the only assignable cause; subsequently I realized that it was not explainable by the Lamarckian hypothesis. I then attributed it to an unknown law of evolution, and there I believe it rests to-day, namely, as a process of which we do not know the cause. Still more recently, however, comes the discovery that original kinship is, partly at least, a control-principle. For example, in the descent of independent stocks of hornless animals arising from a common stock, rudimentary horn-cores are found to appear independently in exactly the same region of the skull, indicating a kind of predetermination in the stock, or potential of similar evolutions. The facts on which this law of mutation, properly called, rests have been misunderstood, totally denied, or explained away by selectionists as survivals of favorable out of indiscriminate variations. Even my colleague, Scott, has identified these phenomena with the saltations of De Vries. Nevertheless, I regard the genesis of new adaptive characters from almost imperceptible beginnings as a vera causa, and as one of the greatest problems we have to solve. That a natural solution will be found goes without saying, although this principle, as stated, is undoubtedly of a teleological nature. Its philosophical bearings are of far-reaching importance. Just as we demand a continent to transfer land animals from Australia to South America, so we demand a natural law to explain these facts. The creative factors of fitness coöperating with selection, which, in my judgment, are now well demonstrated, reside either primarily A. A. W. Hubrecht, Hugo De Vries's Theory of Mutations, The Popular Science Monthly, vol. LXV, no. 3, July, 1904, pp. 205-228. 2 W. Bateson. See Report of the British Association for the Advancement of Science, Cambridge meeting, 1904; (with Miss E. R. Saunders) Reports to the Evolution Committee of the Royal Society, 1902. [Gives summary of Mendel's life and work.] 3 H. F. Osborn, Recent Zoopaleontology, Science, 1905, N. S. vol. xxi, no. 530 pp. 315-316. in the environment, in the bodies of animals, or in the germinal cells they all ultimately find their way into the germinal cells. They may be summarized as follows: (1) Segregation. Besides the familiar geographical segregation of animals, which reaches its highest expression in insular forms, such as the pygmy fossil elephants of Malta and those recently discovered in Cyprus (Bate), there is the no less effective segregation of habit among animals existing in the same geographical regions and under the same climatic conditions, but seeking different varieties of food on different kinds of soil. These give rise to what I have called local adaptive radiations, a principle which explains the occurrence in the same country, and almost side by side, of very conservative as well as very progressive forms. (2) Adaptive Modification. This is a plastic principle which tends in the course of life to an increasing fitness of the bodies of individuals to their special environments and habits, well illustrated among men in the influence of various trades and occupations and operating both in active and in passive structures. Consistent with the adaptive modification principle is the fact that every individual requires habit and environment to model it into its parental form; and in every change of environment or habit every individual is carried an infinitesimal degree beyond the parental form; the wonderful phenomena of correlated development which puzzled Spencer so much are chiefly attributable to this principle. These adaptive modifications are not directly inherited, as Lamarck supposed, but acting through long periods of time there results the organic selection (Morgan, Baldwin,ʻ Osborn') of those individuals in which hereditary predisposition happens most closely to coincide with adaptive modification, and there thus finally comes about an apparent, but not real, inheritance of acquired characters, as Lamarck, Spencer, and Cope sup posed. (3) Variations of Degree. We should by no means exclude as true causes of evolution associated with both the above factors the selection of those variations of degree or around a mean which conform to Quetelet's curve, the subject of the chief investigations of the Galton 1 L. Adams, On the Dentition and Osteology of the Maltese Fossil Elephants, Transactions of the Zoological Society, vol. IX, pt. 1, 1874. 2 Dorothy M. A. Bate, Further Note on the Remains of Elephas Cypriotes, Philosophic Transactions of the Royal Society, London, ser. B, vol. 197, 1904, pp. 347-360. C. Lloyd Morgan [Organic Selection], Science, Nov. 27, 1896. and July, 1896; Development and Evolution, 8vo, New York, 1902. 'J. Mark Baldwin, A New Factor in Evolution, American Naturalist, June York Academy of Science, March and April, 1896, pp. 141-148; The Limits of Inheritance of Acquired Characters (Organic Selection), Transactions of the New H. F. Osborn, A Mode of Evolution requiring neither Natural Selection nor the Organic Selection, American Naturalist, Nov. 1897, pp. 944-951; Modification and Variation and the Limits of Organic Selection, Proceedings, American Association for the Advancement of Science, 46th meeting, 1898, pp. 239-242. school, of Pearson 1 and of Weldon, and which form the strongest remaining ground for Darwin's theory of selection in connection with fortuitous variation. For example, I regard the appearance of longnecked giraffes, of slender-limbed ruminants and horses, of longsnouted aquatic vertebrates, as instances of the selection of variations around a mean rather than of the selection of saltations. The selection of such variations, where they happen to be adaptive, has been an incessant cause of evolution. (4) Saltation. Although Geoffroy Saint-Hilaire argued for paleontological evolution by saltation, I do not think we have much evidence in paleontology for the saltation theory. In the nature of the case, we cannot expect to recognize such evidence even where it may exist, because, wherever a new form appears or a new character arises, as it were, suddenly, we must suspect that this appearance is due to absence of the connecting transitional links to an older form. The whole tendency of paleontological discovery is to resolve what are apparently saltations or discontinuities into processes of continuous change. This, however, by no means precludes saltation from being a vera causa in past time, as rising from "unknown" causes in the germ-cells and as forming the materials from which nature may select the saltations which are adaptive from those which are inadaptive: The paleontologist has every reason to believe that he finds saltations in the sudden variations in the number of vertebræ of the neck, of the back, of the sacral region, for example. In the many familiar cases of the abbreviation or elongation of the vertebral column in adaptation to certain habits, a vertebra in the middle of a series cannot dwindle out of existence; it must suddenly drop out or suddenly appear. (5) Mutation. These new characters are also germinal in origin, because they appear in the teeth, which are structures fully formed beneath the surface before they pierce the gum, and therefore not subsequently modeled by adaptive modification, as the bones, muscles, and all the other tissues of the body are. Mutations are found arising according to partly known influences of kinship. They do not, so far as we observe, possess adaptive value when they first appear, but then frequently, if not always, develop into a stage of usefulness. Fitness is, therefore, the central thought of modern paleontology in its most comprehensive sense, as embracing fitness in the very remote past, in its evolution toward the present, and in its tendencies for the future. Just as the uniformitarian method of Lyell transformed 1 K. Pearson. See articles in Proceedings of the Royal Society of London, and in the Grammar of Science, London, 1900. Geoffroy Saint-Hilaire, Recherches sur des grands Sauriens trouvés à l'état fossile, Mémoires de l'Académie des Sciences, Paris, 1831; Influence du monde ambiant pour modifier les formes animales, Mémoires de l'Académie des Sciences, XII, 1833, p. 63. geology, so the uniformitarian method is penetrating paleontology and making observations of animal and plant life as it is to-day the basis of the understanding of animal and plant life as it was from the beginning. Here again paleontology is not merely an auxiliary to zoölogy; it is chief of a division and enjoys certain unique advantages. We pass in review, with the pedigrees and the prodigies of fitness, the entirely unreasonable, irrational, paradoxical extremes of structure such, for example, as the pterosaurs, which far surpass in boldness and ingenuity of design any of the creations of the modern yacht-builder which are mistakenly regarded by some as having reached an absurd extreme. Problem of Historical Study The paleontologist must also be an historian; he has to deal with lineage, with ancestors, he comes directly upon the problem of kinship or relationship, and he has to determine the various means of distinguishing the true from the apparent relationships. It happens that fitness, while fascinating in itself, has led even the most faithful and skillful into the most devious paths away from the truth. The explanation of this apparent contradiction is in this wise. The ingenuity of nature in adapting animals is astounding, but it is not infinite; the same devices are resorted to repeatedly to accomplish the same purposes. In the evolution of long-snouted rapacious swimming forms, for example, we have already discovered that nature has repeated herself twenty-four times in employing the same processes to accomplish the same ends in entirely different families of animals. 3 1 This introduces us to one of the two great ideas which we must employ in the interpretation of facts, namely, the idea of analogy. We see far more clearly than Huxley did the force of this idea. Owen,1 Cope, Scott, Fraas, and many others, under the terms "parallelism," convergence," 'homoplasy," have developed the force of the old Aristotelian notion that analogy is a similarity of habit, and that in the course of evolution a similarity of habit finally results in a close or exact similarity of structure; this similarity of structure is mistaken. as an evidence of kinship. Analogous evolution does not stop in its far-reaching consequences with analogies in organs; it moulds animals as a whole into similar form, as, for example, the ichthyosaurs, sharks, and dolphins; still more it moulds similar and larger groups of animals into similar lines or radii of specialization. Thus we reach the grand idea of analogy as operating in the divergencies or adaptive 1 R. Owen, On the Nature of Limbs, London, 1849; see also Homology in The Anatomy of Vertebrates, vols. I-III, 1866-68; see also Homology, Lectures on the Comparative Anatomy and Physiology of the Invertebrate Animals, London, 1843. E. D. Cope, Primary Factors of Organic Evolution, 8vo, New York, 1887. W. B. Scott, On some of the Factors in the Evolution of the Mammalia, . Journal of Morphology, vol. v, no. 3, 1891, pp. 379-402. radiations of groups, according to which great orders of animals tend in their families and suborders to mimic other orders, and the fauna or collective orders of continents to mimic the fauna of other continents. Amid this repetition on a grand scale of similar adaptations, which is altogether comparable to what we know as having occurred over and over again in human history, the paleontologist as an historian must keep constantly before him the second great idea of homogeny, of real ancestral kinship, of direct blood descent and hereditary relationship. The shark and the ichthyosaur superficially look alike, but their germ-cells are radically different, their external resemblances are a mere veneer of adaptation, so deceptive, however, that it may be a matter of half a century before we recognize the wolf beneath the clothing of the sheep, or the ass in the lion's skin. These two great ideas, of analogy, or similarity of habit, and homogeny, or similarity of descent, do not run on the same lines; they are the woof and the warp of animal history. Analogy corresponds to the woof, or horizontal strands, which tie animals together by their superficial resemblances in the present; homogeny to the warp, or the fundamental vertical strands which connect animals with their ancestors and their successors. The far-reaching extent of analogous evolution was only dimly perceived by Huxley, and this fact constituted his one great defect as a philosophical anatomist. Its power of transforming unlike and unrelated animals has accomplished miracles in the way of producing a likeness so exact that the inference of kinship is almost irresistible. The paleontologist who would succeed as historian must first, therefore, render himself immune to the misguiding influences of analogy by taking certain further precautions which will now be explained by watching his procedure as historian. Paleontology as the history of life takes its place in the background of recorded history and archeology, and simply from the standpoint of the human pedigree is of transcendent interest. Although it has progressed far beyond the dreams of Darwin and Huxley, the first general statement which must be made is that the actual points of contact between the grand divisions of the animal and plant kingdom, as well as between the lesser and even many of the minor divisions, have yet to be discovered. You recall that the older grand divisions of the Vertebrata, to which we must confine our attention, were suggested by the so-called Ages of Fishes, of Amphibians, of Reptiles, and of Mammals. Even within these grand divisions we observe a succession of more or less closely analogous groups. Each of these groups has its more or less central starting-point in a smaller and older group which contains a large number of primitive or generalized characters. The search for the primitive central form is always made by the |