Even when propagated by budding, a multicellular organism has been ultimately derived from a germ-cell.
Now, if the theory of evolution is true, what should we expect to happen when these germ-cells are fertilized, and so enter upon their severally distinct processes of development? Assuredly we should expect to find that the higher organisms pass through the same phases of development as the lower organisms, up to the time when their higher characters begin to become apparent. If in the life-history of species these higher characters were gained by gradual improvement upon lower characters, and if the development of the higher individual is now a general recapitulation of that of its ancestral species, in studying this recapitulation we should expect to find the higher organism successively unfolding its higher characters from the lower ones through which its ancestral species had previously passed. And this is just what we do find. Take, for example, the case of the highest organism, Man. Like that of all other organisms, unicellular or multicellular, his development starts from the nucleus of a single cell. Again, like that of all the Metazoa and Metaphyta, his development starts from the specially elaborated nucleus of an egg-cell, or a nucleus which has been formed by the fusion of a male with a female element. When his animality becomes established, he exhibits the fundamental anatomical qualities which characterize such lowly animals as polyps and jelly-fish. And even when he is marked off as a Vertebrate, it cannot be said whether he is to be a fish, a reptile, a bird, or a beast. Later on it becomes evident that he is to be a Mammal; but not till later still can it be said to which order of mammals he belongs.
 It has already been stated that both parthenogenesis and gemmation are ultimately derived from sexual reproduction. It may now be added, on the other hand, that the earlier stages of parthenogenesis have been observed to occur sporadically in all sub-kingdoms of the Metaxoa, including the Vertebrata, and even the highest class, Mammalia. These earlier stages consist in spontaneous segmentations of the ovum; so that even if a virgin has ever conceived and borne a son, and even if such a fact in the human species has been unique, still it would not betoken any breach of physiological continuity. Indeed, according to Weismann's not improbable hypothesis touching the physiological meaning of polar bodies, such a fact need betoken nothing more than a slight disturbance of the complex machinery of ovulation, on account of which the ovum failed to eliminate from its substance an almost inconceivably minute portion of its nucleus.
Here, however, we must guard against an error which is frequently met with in popular expositions of this subject. It is not true that the embryonic phases in the development of a higher form always resemble so many adult stages of lower forms. This may or may not be the case; but what always is the case is, that the embryonic phases of the higher form resemble the corresponding phases of the lower forms. Thus, for example, it would be wrong to suppose that at any stage of his development a man resembles a jelly-fish. What he does resemble at an early stage of his development is the essential or groundplan of the jelly-fish, which that animal presents in its embryonic condition, or before it begins to assume its more specialized characters fitting it for its own particular sphere of life. The similarities, therefore, which it is the function of comparative embryology to reveal are the similarities of type or morphological plan: not similarities of specific detail. Specific details may have been added to this, that, and the other species for their own special requirements, after they had severally branched off from the common ancestral stem; and so could not be expected to recur in the life-history of an independent specific branch. The comparison therefore must be a comparison of embryo with embryo; not of embryos with adult forms.
* * * * *
In order to give a general idea of the results thus far yielded by a study of comparative embryology in the present connexion, I will devote the rest of this chapter to giving an outline sketch of the most important and best established of these results.
Histologically the ovum, or egg-cell, is nearly identical in all animals, whether vertebrate or invertebrate. Considered as a cell it is of large size, but actually it is not more than 1/100, and may be less than 1/200 of an inch in diameter. In man, as in most mammals, it is about 1/120. It is a more or less spherical body, presenting a thin transparent envelope, called the zona pellucida, which contains—first, the protoplasmic cell-substance or "yolk," within which lies, second, the nucleus or germinal vesicle, within which again lies, third, the nucleolus or germinal spot. This description is true of the egg-cells of all animals, if we add that in the case of the lowest animals—such as sponges, &c.—there is no enveloping membrane: the egg-cell is here a naked cell, and its constituent protoplasm, being thus unconfined, is free to perform protoplasmic movements, which it does after the manner, and with all the activity, of an amoeba. But even with respect to this matter of an enveloping membrane, there is no essential difference between an ovum of the lowest and an ovum of the highest animals. For in their early stages of development within the ovary the ova of the highest animals are likewise in the condition of naked cells, exhibiting amoebiform movements; the enveloping membrane of an ovum being the product of a later development. Moreover this membrane, when present, is usually provided with one or more minute apertures, through which the spermatozooen passes when fertilizing the ovum. It is remarkable that the spermatozoa know, so to speak, of the existence of these gate-ways,—their snake-like movements being directed towards them, presumably by a stimulus due to some emanation therefrom. In the mammalian ovum, however, these apertures are exceedingly minute, and distributed all round the circumference of the pellucid envelope, as represented in this illustration (Fig. 32).
 The spermatozooids of certain plants can be strongly attracted towards a pipette which is filled with malic acid—crowding around and into it with avidity.
[Illustration: FIG. 31.—Amoeboid movements of young egg-cells, a, Amoeboid ovum of Hydra (from Balfour, after Kleitnenberg); b, early ovum of Toxopneustes variegatus, with pseudopodia-like processes (from Balfour, after Selenka); c, ovum of Toxopneustes lividus, more nearly ripe (from Balfour, Hertwig). A1 to A4, the primitive egg-cell of a Chalk-Sponge (Leuculmis echinus), in four successive conditions of motion. B1 to B8, ditto of a Hermit-Crab (Chondracanthus cornutus), in eight successive stages (after E. von Beneden). C1 to C5, ditto of a Cat, in five successive stages (after Pflueger). D, ditto of Trout; E, of a Hen; F, of Man. The first series is taken from the Encycl. Brit.; the second from Haeckel's Evolution of Man.]
In thus saying that the ova of all animals are, so far as microscopes can reveal, substantially similar, I am of course speaking of the egg-cell proper, and not of what is popularly known as the egg. The egg of a bird, for example, is the egg-cell, plus an enormous aggregation of nutritive material, an egg-shell, and sundry other structures suited to the subsequent development of the egg-cell when separated from the parent's body. But all these accessories are, from our present point of view, accidental or adventitious. What we have now to understand by the ovum, the egg, or the egg-cell, is the microscopical germ which I have just described. So far then as this germ is concerned, we find that all multicellular organisms begin their existence in the same kind of structure, and that this structure is anatomically indistinguishable from that of the permanent form presented by the lowest, or unicellular organisms. But although anatomically indistinguishable, physiologically they present the sundry peculiarities already mentioned.
Now I have endeavoured to show that none of these peculiarities are such as to exclude—or even so much as to invalidate—the supposition of developmental continuity between the lowest egg-cells and the highest protozoal cells. It remains to show in this place, and on the other hand, that there is no breach of continuity between the lowest and the highest egg-cells; but, on the contrary, that the remarkable uniformity of the complex processes whereby their peculiar characters are exhibited to the histologist, is such as of itself to sustain the doctrine of continuity in a singularly forcible manner. On this account, therefore, and also because the facts will again have to be considered in another connexion when we come to deal with Weismann's theory of heredity, I will here briefly describe the processes in question.
We have already seen that the young egg-cell multiplies itself by simple binary division, after the manner of unicellular organisms in general—thereby indicating, as also by its amoebiform movements, its fundamental identity with such organisms in kind. But, as we have likewise seen, when the ovum ceases to resemble these organisms, by taking on its higher degree of functional capacity, it is no longer able to multiply itself in this manner. On the contrary, its cell-divisions are now of an endogenous character, and result in the formation of many different kinds of cells, in the order required for constructing the multicellular organism to which the whole series of processes eventually give rise. We have now to consider these processes seriatim.
First of all the nucleus discharges its polar bodies, as previously mentioned, and in the manner here depicted on the previous page. (Fig. 33.) It will be observed that the nucleus of the ovum, or the germinal vesicle as it is called, gets rid first of one and afterwards of the other polar body by an "indirect," or karyokinetic, process of division. (Fig. 33.) Extrusion of these bodies from the ovum (or it may be only from the nucleus) having been accomplished, what remains of the nucleus retires from the circumference of the ovum, and is called the female pronucleus. (Fig. 33. f. pn.) The ovum is now ready for fertilization. A similar emission of nuclear substance is said by some good observers to take place also from the male germ-cell, or spermatozooen, at or about the close of its development. The theories to which these facts have given rise will be considered in future chapters on Heredity.
Turning now to the mechanism of fertilization, the diagrams (Figs. 34, 35) represent what happens in the case of star-fish.
The sperm-cell, or spermatozooen, is seen in the act of penetrating the ovum. In the first figure it has already pierced the mucilaginous coat of the ovum, the limit of which is represented by a line through which the tail of the spermatozooen is passing: the head of the spermatozooen is just entering the ovum proper. It may be noted that, in the case of many animals, the general protoplasm of the ovum becomes aware, so to speak, of the approach of a spermatozooen, and sends up a process to meet it. (Fig. 35, A, B, C.) Several—or even many—spermatozoa may thus enter the coat of the ovum; but normally only one proceeds further, or right into the substance of the ovum, for the purpose of effecting fertilization. This spermatozooen, as soon as it enters the periphery of the yolk, or cell-substance proper, sets up a series of remarkable phenomena. First, its own head rapidly increases in size, and takes on the appearance of a cell-nucleus: this is called the male pronucleus. At the same time its tail begins to disappear, and the enlarged head proceeds to make its way directly towards the nucleus of the ovum which, as before stated, is now called the female pronucleus. The latter in its turn moves towards the former, and when the two meet they fuse into one mass, forming a new nucleus. Before the two actually meet, the spermatozooen has lost its tail altogether; and it is noteworthy that during its passage through the protoplasmic cell-contents of the ovum, it appears to exercise upon this protoplasm an attractive influence; for the granules of the latter in its vicinity dispose themselves around it in radiating lines. All these various phenomena are depicted in the above wood-cuts. (Figs. 34, 35.)
Fertilization having been thus effected by fusion of the male and female pronuclei into a single (or new) nucleus, this latter body proceeds to exhibit complicated processes of karyokinesis, which, as before shown, are preliminary to nuclear division in the case of egg-cells. Indeed the karyokinetic process may begin in both the pronuclei before their junction is effected; and, even when their junction is effected, it does not appear that complete fusion of the so-called chromatin elements of the two pronuclei takes place. For the purpose of explaining what this means, and still more for the purpose of giving a general idea of the karyokinetic processes as a whole, I will quote the following description of them, because, for terseness combined with lucidity, it is unsurpassable.
[Illustration: FIG. 36.—Karyokinesis of a typical tissue-cell (epithelium of Salamander). (After Flemming and Klein.) The series from A to I represents the successive stages in the movement of the chromatin fibres during division, excepting G, which represents the "nucleus-spindle" of an egg-cell. A, resting nucleus; D, wreath-form; E, single star, the loops of the wreath being broken; F, separation of the star into two groups of U-shaped fibres; H, diaster or double star; I, completion of the cell-division and formation of two resting nuclei. In G the chromatin fibres are marked a, and correspond to the "equatorial plate"; b, achromatin fibres forming the nucleus-spindle; c, granules of the cell-protoplasm forming a "polar star." Such a polar star is seen at each end of the nucleus-spindle, and is not to be confused with the diaster H, the two ends of which are composed of chromatin.]
Researches, chiefly due to Flemming, have shown that the nucleus in very many tissues of higher plants and animals consists of a capsule containing a plasma of "achromatin," not deeply stained by re-agents, ramifying in which is a reticulum of "chromatin" consisting of fibres which readily take a deep stain. (Fig. 36, A). Further it is demonstrated that, when the cell is about to divide into two, definite and very remarkable movements take place in the nucleus, resulting in the disappearance of the capsule and in the arrangement of its fibres first in the form of a wreath (D), and subsequently (by the breaking of the loops formed by the fibres) in the form of a star (E). A further movement within the nucleus leads to an arrangement of the broken loops in two groups (F), the position of the open ends of the broken loops being reversed as compared with what previously obtained. Now the two groups diverge, and in many cases a striated appearance of the achromatin substance between the two groups of chromatin loops is observable (H). In some cases (especially egg-cells) this striated arrangement of the achromatin is then termed a "nucleus-spindle," and the group of chromatin loops (G, a) is known as "the equatorial plate." At each end of the nucleus-spindle in these cases there is often seen a star consisting of granules belonging to the general protoplasm of the cell (G, c). These are known as "polar stars." After the separation of the two sets of loops (H) the protoplasm of the general substance of the cell becomes constricted, and division occurs, so as to include a group of chromatin loops in each of the two fission products. Each of these then rearranges itself together with the associated chromatin into a nucleus such as was present in the mother cell to commence with (I).
 Ray Lankester, Encyclop. Brit., 9th ed., Vol. XIX, pp. 832-3.
Since the above was published, however, further progress has been made. In particular it has been found that the chromatin fibres pass from phase D to phase F by a process of longitudinal splitting (Fig. 37 g, h; Fig. 38, VI, VII)—which is a point of great importance for Weismann's theory of heredity,—and that the protoplasm outside the nucleus seems to take as important a part in the karyokinetic process as does the nuclear substance. For the so-called "attraction-spheres" (Fig. 38 II a, III, III a, VIII to XII), which were at first supposed to be of subordinate importance in the process as a whole, are now known to take an exceedingly active part in it (see especially IX to XI). Lastly, it may be added that there is a growing consensus of authoritative opinion, that the chromatin fibres are the seats of the material of heredity, or, in other words, that they contain those essential elements of the cell which endow the daughter-cells with their distinctive characters. Therefore, where the parent-cell is an ovum, it follows from this view that all hereditary qualities of the future organism are potentially present in the ultra-microscopical structure of the chromatin fibres.
[Illustration: FIG. 37.—Study of successive changes taking place in the nucleus of an epithelium cell, preparatory to division of the cell. (From Quain's Anatomy, after Flemming.) a, resting cell, showing the nuclear network; b, first stage of division, the chromatoplasm transformed into a skein of closely contorted filaments; c to f, further stages in the growth and looping arrangement of the filaments; g, stellate phase, or aster; h, completion of the splitting of the filaments, already begun in f and g; i, j, k, successive stages in separation of the filaments into two groups; l, the final result of this (diaster); m to q, stages in the division of the whole cell into two, showing increasing contortion of the filaments, until they reach the resting stage at q].
[Illustration: FIG. 38.—Formation and conjugation of the pronuclei in Ascaris megalocephala. (From Quain's Anatomy, after E. von Beneden.) f, female pronucleus; m, male pronucleus; p, one of the polar bodies.
I. The second polar body has just been extruded; both male and female pronuclei contain two chromatin particles; those of the male pronucleus are becoming transformed into a skein. II. The chromatin in both pronuclei now forms into a skein.
II a. The skeins are more distinct. Two attraction (or protoplasmic) spheres, each with a central particle united with a small spindle of achromatic fibres, have made their appearance in the general substance of the egg close to the mutually approaching pronuclei. The male pronucleus has the remains of the body of the spermatozooen adhering to it.
III. Only the female pronucleus is shown in this figure. The skein is contracted and thickened. The attraction-spheres are near one side of the ovum, and are connected with its periphery by a cone of fibres forming a polar circle, p.c.; e.c., equatorial circle.
III a. The pronuclei have come into contact, and the spindle-system is now arranged across their common axis.
IV. Contraction of the skein, and formation of two U-or V-shaped chromatin fibres in each pronucleus.
V. The V-shaped chromatin filaments are now quite distinct: the male and female pronuclei are in close contact.]
[Illustration: (38 continued)
VI., VII. The V-shaped filaments are splitting longitudinally; their structure of fine granules of chromatin is apparent in VII., which is more highly magnified. The conjugation of the pronuclei is apparently complete in VII. The attraction-spheres and achromatic spindle, although present, are not depicted in IV., V., VI., and VII.
VIII. Equatorial arrangement of the four chromatin loops in the middle of the now segmenting ovum: the achromatic substance forming a spindle-shaped system of granules with fibres radiating from the poles of the spindle (attraction-spheres); the chromatin forms an equatorial plate. (Compare Fig. 36 G.)
IX. Shows diagrammatically the commencing separation of the chromatin fibres of the conjugated nuclei, and the system of fibres radiating from the attraction-spheres. (Compare again Fig. 36 G.) p.c., polar circle; e.c., equatorial circle; c.c., central particle.
X. Further separation of the chromatin filaments. Each of the central particles of the attraction-spheres has divided into two.
XI. The chromatin fibres are becoming developed into the skeins of the two daughter-nuclei. These are still united by fibres of achromatin. The general protoplasm of the ovum is becoming divided.
XII. The two daughter-nuclei exhibit a chromatin network. Each of the attraction-spheres has divided into two, which are joined by fibres of achromatin, and connected with the periphery of the cell in the same way as in the original or parent sphere, III.]
As I shall have more to say about these processes in the next volume, when we shall see the important part which they bear in Weismann's theory of heredity, it is with a double purpose that I here introduce these yet further illustrations of them upon a somewhat larger scale. The present purpose is merely that of showing, more clearly than hitherto, the great complexity of these processes on the one hand, and, on the other, the general similarity which they display in egg-cells and in tissue-cells. But as in relation to this purpose the illustrations speak for themselves, I may now pass on at once to the history of embryonic development, which follows fertilization of the ovum.
* * * * *
We have seen that when the new nucleus of the fertilized ovum (which is formed by a coalescence of the male pronucleus with the female) has completed its karyokinetic processes, it is divided into two equal parts; that these are disposed at opposite poles of the ovum; and that the whole contents of the ovum are thereupon likewise divided into two equal parts, with the result that there are now two nucleated cells within the spherical wall of the ovum where before there had only been one. Moreover, we have also seen that a precisely similar series of events repeat themselves in each of these two cells, thus giving rise to four cells (see Fig. 29). It must now be added that such duplication is continued time after time, as shown in the accompanying illustrations (Figs. 39, 40).
All this, it will be noticed, is a case of cell-multiplication, which differs from that which takes place in the unicellular organisms only in its being invariably preceded (as far as we know) by karyokinesis, and in the resulting cells being all confined within a common envelope, and so in not being free to separate. Nevertheless, from what has already been said, it will also be noticed that this feature makes all the difference between a Metazooen and a Protozooen; so that already the ovum presents the distinguishing character of a Metazooen.
I have dealt thus at considerable length upon the processes whereby the originally unicellular ovum and spermatozooen become converted into the multicellular germ, because I do not know of any other exposition of the argument from Embryology where this, the first stage of the argument, has been adequately treated. Yet it is evident that the fact of all the processes above described being so similar in the case of sexual (or metazoal) reproduction among the innumerable organisms where it occurs, constitutes in itself a strong argument in favour of evolution. For the mechanism of fertilization, and all the processes which even thus far we have seen to follow therefrom, are hereby shown to be not only highly complex, but likewise highly specialized. Therefore, the remarkable similarity which they present throughout the whole animal kingdom—not to speak of the vegetable—is expressive of organic continuity, rather than of absolute discontinuity in every case, as the theory of special creation must necessarily suppose. And it is evident that this argument is strong in proportion to the uniformity, the specialization, and the complexity of the processes in question.
Having occupied so much space with supplying what appear to me the deficiencies in previous expositions of the argument from Embryology, I can now afford to take only a very general view of the more important features of this argument as they are successively furnished by all the later stages of individual development. But this is of little consequence, seeing that from the point at which we have now arrived previous expositions of the argument are both good and numerous. The following then is to be regarded as a mere sketch of the evidences of phyletic (or ancestral) evolution, which are so abundantly furnished by all the subsequent phases of ontogenetic (or individual) evolution.
The multicellular body which is formed by the series of segmentations above described is at first a sphere of cells (Fig. 40). Soon, however, a watery fluid gathers in the centre, and progressively pushes the cells towards the circumference, until they there constitute a single layer. The ovum, therefore, is now in the form of a hollow sphere containing fluid, confined within a continuous wall of cells (Fig. 41 A). The next thing that happens is a pitting in of one portion of the sphere (B). The pit becomes deeper and deeper, until there is a complete invagination of this part of the sphere—the cells which constitute it being progressively pushed inwards until they come into contact with those at the opposite pole of the ovum. Consequently, instead of a hollow sphere of cells, the ovum now becomes an open sac, the walls of which are composed of a double layer of cells (C). The ovum is now what has been called a gastrula; and it is of importance to observe that probably all the Metazoa pass through this stage. At any rate it has been found to occur in all the main divisions of the animal kingdom, as a glance at the accompanying figures will serve to show (Fig. 42). Moreover many of the lower kinds of Metazoa never pass beyond it; but are all their lives nothing else than gastrulae, wherein the orifice becomes the mouth of the animal, the internal or invaginated layer of cells the stomach, and the outer layer the skin. So that if we take a child's india-rubber ball, of the hollow kind with a hole in it, and push in one side with our fingers till internal contact is established all round, by then holding the indented side downwards we should get a very fair anatomical model of a gastraea form, such as is presented by the adult condition of many of the most primitive Metazoa—especially the lower Coelenterata. The preceding figures represent two other such forms in nature, the first locomotive and transitory, the second fixed and permanent (Figs. 43, 44).
 In most vertebrated animals this process of gastrulation has been more or less superseded by another, which is called delamination; but it scarcely seems necessary for our present purposes to describe the latter. For not only does it eventually lead to the same result as gastrulation—i. e. the converting of the ovum into a double-walled sac,—but there is good evidence among the lower Vertebrata of its being preceded by gastrulation; so that, even as to the higher Vertebrata, embryologists are pretty well agreed that delamination has been but a later development of, or possibly improvement upon, gastrulation.
Here, then, we leave the lower forms of Metazoa in their condition of permanent gastrulae. They differ from the transitory stage of other Metazoa only in being enormously larger (owing to greatly further growth, without any further development as to matters of fundamental importance), and in having sundry tentacles and other organs added later on to meet their special requirements. The point to remember is, that in all cases a gastrula is an open sac composed of two layers of cells—the outer layer being called the ectoderm, and the inner the endoderm. They have also been called the animal layer and the vegetative layer, because it is the outer layer (ectoderm) that gives rise to all the organs of sensation and movement—viz. the skin, the nervous system, and the muscular system; while it is the inner layer (endoderm) that gives rise to all the organs of nutrition and reproduction. It is desirable only further to explain that gastrulation does not take place in all the Metazoa after exactly the same plan. In different lines of descent various and often considerable modifications of the original and most simple plan have been introduced; but I will not burden the present exposition by describing these modifications. It is enough for us that they always end in the formation of the two primary layers of ectoderm and endoderm.
 The most extreme of them is that which is mentioned in the last foot-note.
The next stage of differentiation is common to all the Metazoa, except those lowest forms which, as we have just seen, remain permanently as large gastrulae, with sundry specialized additions in the way of tentacles, &c. This stage of differentiation consists in the formation of either a pouch or an additional layer between the ectoderm and the endoderm, which is called the mesoderm. It is probably in most cases derived from the endoderm, but the exact mode of its derivation is still somewhat obscure. Sometimes it has the appearance of itself constituting two layers; but it is needless to go into these details; for in any case the ultimate result is the same—viz. that of converting the Metazooen into the form of a tube, the walls of which are composed of concentric layers of cells. The outermost layer afterwards gives rise to the epidermis with its various appendages, and also to the central nervous system with its organs of special sense. The median layer gives rise to the voluntary muscles, bones, cartilages, &c., the nutritive systems of the blood, the chyle, the lymph, and the muscular tube of the intestine. Lastly, the innermost layer developes into the epithelium lining of the intestine, with its various appendages of liver, lungs, intestinal glands, &c.
I have just said that this three or four layered stage is shared by all the Metazoa, except those very lowest forms—such as sponges and jelly-fish—which do not pass on to it. But from this point the developmental histories of all the main branches of the Metazoa diverge—the Vermes, the Echinodermata, the Mollusca, the Articulata, and the Vertebrata, each taking a different road in their subsequent evolution. I will therefore confine attention to only one of these several roads or methods, namely, that which is followed by the Vertebrata—observing merely that, if space permitted, the same principles of progressive though diverging histories of evolution would equally well admit of being traced in all the other sub-kingdoms which have just been named.
In order to trace these principles in the case of the Vertebrata, it is desirable first of all to obtain an idea of the anatomical features which most essentially distinguish the sub-kingdom as a whole. The following, then, is what may be termed the ideal plan of vertebrate organization, as given by Prof. Haeckel. First, occupying the major axis of body we perceive the primitive vertebral column. The parts lying above this axis are those which have been developed from the ectoderm and mesoderm—viz. voluntary muscles, central nervous system, and organs of special sense. The parts lying below this axis are for the most part those which have been developed from the endoderm—namely, the digestive tract with its glandular appendages, the circulating system and the respiratory system. In transverse section, therefore, the ideal vertebrate consists of a solid axis, with a small tube occupied by the nervous system above, and a large tube, or body-cavity, below. This body-cavity contains the viscera, breathing organs, and heart, with its prolongations into the main blood-vessels of the organism. Lastly, on either side of the central axis are to be found large masses of muscle—two on the dorsal and two on the ventral. As yet, however, there are no limbs, nor even any bony skeleton, for the primitive vertebral column is hitherto unossified cartilage. This ideal animal, therefore, is to all appearance as much like a worm as a fish, and swims by means of a lateral undulation of its whole body, assisted, perhaps, by a dorsal fin formed out of skin.
Now I should not have presented this ideal representation of a primitive vertebrate—for I have very little faith in the "scientific use of the imagination" where it aspires to discharge the functions of a Creator in the manufacture of archetypal forms—I say I should not have presented this ideal representative of a primitive vertebrate, were it not that the ideal is actually realized in a still existing animal. For there still survives what must be an immensely archaic form of vertebrate, whose anatomy is almost identical with that of the imaginary type which has just been described. I allude, of course, to Amphioxus, which is by far the most primitive or generalized type of vertebrated animal hitherto discovered. Indeed, we may say that this remarkable creature is almost as nearly allied to a worm as it is to a fish. For it has no specialized head, and therefore no skull, brain, or jaws: it is destitute alike of limbs, of a centralized heart, of developed liver, kidneys, and, in short, of most of the organs which belong to the other Vertebrata. It presents, however, a rudimentary backbone, in the form of what is called a notochord. Now a primitive dorsal axis of this kind occurs at a very early period of embryonic life in all vertebrated animals; but, with the exception of Amphioxus, in all other existing Vertebrata this structure is not itself destined to become the permanent or bony vertebral column. On the contrary, it gives way to, or is replaced by, this permanent bony structure at a later stage of development. Consequently, it is very suggestive that so distinctively embryonic a structure as this temporary cartilaginous axis of all the other known Vertebrata should be found actually persisting to the present day as the permanent axis of Amphioxus. In many other respects, likewise, the early embryonic history of other Vertebrata refers us to the permanent condition of Amphioxus. In particular, we must notice that the wall of the neck is always perforated by what in Amphioxus are the gill-openings, and that the blood-vessels as they proceed from the heart are always distributed in the form of what are called gill-arches, adapted to convey the blood round or through the gills for the purpose of aeration. In all existing fish and other gill-breathing Vertebrata, this arrangement is permanent. It is likewise met with in a peculiar kind of worm, called Balanoglossus—a creature so peculiar, indeed, that it has been constituted by Gegenbaur a class all by itself. We can see by the wood-cuts that it presents a series of gill-slits, like the homologous parts of the fishes with which it is compared—i. e. fishes of a comparatively low type of organization, which dates from a time before the development of external gills. (Figs. 48, 49, 50.) Now, as I have already said, these gill-slits are supported internally by the gill-arches, or the blood-vessels which convey the blood to be oxygenized in the branchial apparatus (see below, Figs. 51, 52, 53); and the whole arrangement is developed from the anterior part of the intestine—as is likewise the respiratory mechanism of all the gill-breathing Vertebrata. That so close a parallel to this peculiar mechanism should be met with in a worm, is a strong additional piece of evidence pointing to the derivation of the Vertebrata from the Vermes.
Well, I have just said that in all the gill-breathing Vertebrata, this mechanism of gill-slits and vascular gill-arches in the front part of the intestinal tract is permanent. But in the air-breathing Vertebrata such an arrangement would obviously be of no use. Consequently, the gill-slits in the sides of the neck (see Figs. 16 and 57, 58), and the gill-arches of the large blood-vessels (Figs. 54, 55, 56), are here exhibited only as transitory phases of development. But as such they occur in all air-breathing Vertebrata. And, as if to make the homologies as striking as possible, at the time when the gill-slits and the gill-arches are developed in the embryonic young of air-breathing Vertebrata, the heart is constructed upon the fish-like type. That is to say, it is placed far forwards, and, from having been a simple tube as in Worms, is now divided into two chambers, as in Fish. Later on it becomes progressively pushed further back between the developing lungs, while it progressively acquires the three cavities distinctive of Amphibia, and finally the four cavities belonging only to the complete double circulation of Birds and Mammals. Moreover, it has now been satisfactorily shown that the lungs of air-breathing Vertebrata, which are thus destined to supersede the function of gills, are themselves the modified swim-bladder or float, which belongs to Fish. Consequently, all these progressive modifications in the important organs of circulation and respiration in the air-breathing Vertebrata, together make up as complete a history of their aquatic pedigree as it would be possible for the most exacting critic to require.
If space permitted, it would be easy to present abundance of additional evidence to the same effect from the development of the skeleton, the skull, the brain, the sense-organs, and, in short, of every constituent part of the vertebrate organization. Even without any anatomical dissection, the similarity of all vertebrated embryos at comparable stages of development admits of being strikingly shown, if we merely place the embryos one beside the other. Here, for instance, are the embryos of a fish, a salamander, a tortoise, a bird, and four different mammals. In each case three comparable stages of development are represented. Now, if we read the series horizontally, we can see that there is very little difference between the eight animals at the earliest of the three stages represented—all having fish-like tails, gill-slits, and so on. In the next stage further differentiation has taken place, but it will be observed that the limbs are still so rudimentary that even in the case of Man they are considerably shorter than the tail. But in the third stage the distinctive characters are well marked.
* * * * *
So much then for an outline sketch of the main features in the embryonic history of the Vertebrata. But it must be remembered that the science of comparative embryology extends to each of the other three great branches of the tree of life, where these take their origin, through the worms, from the still lower, or gastraea, forms. And in each of these three great branches—namely, the Echinodermata, the Mollusca, and the Arthropoda—we have a repetition of just the same kind of evidence in favour of continuous descent, with adaptive modification in sundry lines, as that which I have thus briefly sketched in the case of the Vertebrata. The roads are different, but the method of travelling is the same. Moreover, when the embryology of the Worms is closely studied, the origin of these different roads admits of being clearly traced. So that when all this mass of evidence is taken together, we cannot wonder that evolutionists should now regard the science of comparative embryology as the principal witness to their theory.
The present Chapter will be devoted to a consideration of the evidence of organic evolution which has been furnished by the researches of geologists. On account of its direct or historical nature, this branch of evidence is popularly regarded as the most important—so much so, indeed, that in the opinion of most educated persons the whole doctrine of organic evolution must stand or fall according to the so-called "testimony of the rocks." Now, without at all denying the peculiar importance of this line of evidence, I must begin by remarking that it does not present the denominating importance which popular judgment assigns to it. For although popular judgment is right in regarding the testimony of the rocks as of the nature of a history, this judgment, as a rule, is very inadequately acquainted with the great imperfections of that history. Knowing in a general way what magnificent advances the science of geology has made during the present century, the public mind is more or less imbued with the notion, that because we now possess a tolerably complete record of the chronological succession of geological formations, we must therefore possess a correspondingly complete record of the chronological succession of the forms of life which from time to time have peopled the globe. Now in one sense this notion is partly true, but in another sense it is profoundly false. It is partly true if we have regard only to those larger divisions of the vegetable or animal kingdoms which naturalists designate by the terms classes and orders. But the notion becomes progressively more untrue when it is applied to families and genera, while it is most of all untrue when applied to species. That this must be so may be rendered apparent by two considerations.
In the first place, it does not follow that because we have a tolerably complete record of the succession of geological formations, we have therefore any correspondingly complete record of their fossiliferous contents. The work of determining the relative ages of the rocks does not require that every cubic mile of the earth's surface should be separately examined, in order to find all the different fossils which it may contain. Were this the case, we should hitherto have made but very small progress in our reading of the testimony of the rocks. The relative ages of the rocks are determined by broad comparative surveys over extensive areas; and although the identification of widely separated deposits is often greatly assisted by a study of their fossiliferous contents, the mere pricking of a continent here and there is all that is required for this purpose. Hence, the accuracy of our information touching the relative ages of geological strata does not depend upon—and, therefore, does not betoken—any equivalent accuracy of knowledge touching the fossiliferous material which these strata may at the present time actually contain. And, as we well know, the opportunities which the geologist has of discovering fossils are extremely limited, if we consider these opportunities in relation to the area of geological formations. The larger portion of the earth's surface is buried beneath the sea; and much the larger portion of the fossiliferous deposits on shore are no less hopelessly buried beneath the land. Therefore it is only upon the fractional portion of the earth's surface which at the present time happens to be actually exposed to his view that the geologist is able to prosecute his search for fossils. But even here how miserably inadequate this search has hitherto been! With the exception of a scratch or two in the continents of Asia and America, together with a somewhat larger number of similar scratches over the continent of Europe, even that comparatively small portion of the earth's surface which is available for the purpose has been hitherto quite unexplored by the palaeontologist. How enormously rich a store of material remains to be unearthed by the future scratchings of this surface, we may dimly surmise from the astonishing world of bygone life which is now being revealed in the newly discovered fossiliferous deposits on the continent of America.
But, besides all this, we must remember, in the second place, that all the fossiliferous deposits in the world, even if they could be thoroughly explored, would still prove highly imperfect, considered as a history of extinct forms of life. In order that many of these forms should have been preserved as fossils, it is necessary that they should have died upon a surface neither too hard nor too soft to admit of their leaving an impression; that this surface should afterwards have hardened sufficiently to retain the impression; that it should then have been protected from the erosion of water, as well as from the disintegrating influence of the air; and yet that it should not have sunk far enough beneath the surface to have come within the no less disintegrating influence of subterranean heat. Remembering thus, as a general rule, how many conditions require to have met before a fossil can have been both formed and preserved, we must conclude that the geological record is probably as imperfect in itself as are our opportunities of reading even the little that has been recorded. If we speak of it as a history of the succession of life upon the planet, we must allow, on the one hand, that it is a history which merits the name of a "chapter of accidents"; and, on the other hand, that during the whole course of its compilation pages were being destroyed as fast as others were being formed, while even of those that remain it is only a word, a line, or at most a short paragraph here and there, that we are permitted to see. With so fragmentary a record as this to study, I do not think it is too much to say that no conclusions can be fairly based upon it, merely from the absence of testimony. Only if the testimony were positively opposed to the theory of descent, could any argument be fairly raised against that theory on the grounds of this testimony. In other words, if any of the fossils hitherto discovered prove the order of succession to have been incompatible with the theory of genetic descent, then the record may fairly be adduced in argument, because we should then be in possession of definite information of a positive kind, instead of a mere absence of information of any kind. But if the adverse argument reaches only to the extent of maintaining that the geological record does not furnish us with so complete a series of "connecting links" as we might have expected, then, I think, the argument is futile. Even in the case of human histories, written with the intentional purpose of conveying information, it is an unsafe thing to infer the non-occurrence of an event from a mere silence of the historian—and this especially in matters of comparatively small detail, such as would correspond (in the present analogy) to the occurrence of species and genera as connecting links. And, of course, if the history had only come down to us in fragments, no one would attach any importance at all to what might have been only the apparent silence of the historian.
In view, then, of the unfortunate imperfection of the geological record per se, as well as of the no less unfortunate limitation of our means of reading even so much of the record as has come down to us, I conclude that this record can only be fairly used in two ways. It may fairly be examined for positive testimony against the theory of descent, or for proof of the presence of organic remains of a high order of development in a low level of strata. And it may be fairly examined for negative testimony, or for the absence of connecting links, if the search be confined to the larger taxonomic divisions of the fauna and flora of the world. The more minute these divisions, the more restricted must have been the areas of their origin, and hence the less likelihood of their having been preserved in the fossil state, or of our finding them even if they have been. Therefore, if the theory of evolution is true, we ought not to expect from the geological record a full history of specific changes in any but at most a comparatively small number of instances, where local circumstances happen to have been favourable for the writing and preservation of such a history. But we might reasonably expect to find a general concurrence of geological testimony to the larger fact—namely, of there having been throughout all geological time a uniform progression as regards the larger taxonomic divisions. And, as I will next proceed to show, this is, in a general way, what we do find, although not altogether without some important exceptions, with which I shall deal in an Appendix.
There is no positive proof against the theory of descent to be drawn from a study of palaeontology, or proof of the presence of any kind of fossils in strata where the fact of their presence is incompatible with the theory of evolution. On the other hand, there is an enormous body of uniform evidence to prove two general facts of the highest importance in the present connexion. The first of these general facts is, that an increase in the diversity of types both of plants and animals has been constant and progressive from the earliest to the latest times, as we should anticipate that it must have been on the theory of descent in ever-ramifying lines of pedigree. And the second general fact is, that through all these branching lines of ever-multiplying types, from the first appearance of each of them to their latest known conditions, there is overwhelming evidence of one great law of organic nature—the law of gradual advance from the general to the special, from the low to the high, from the simple to the complex.
Now, the importance of these large and general facts in the present connexion must be at once apparent; but it may perhaps be rendered more so if we try to imagine how the case would have stood supposing geological investigation to have yielded in this matter an opposite result, or even so much as an equivocal result. If it had yielded an opposite result, if the lower geological formations were found to contain as many, as diverse, and as highly organized types as the later geological formations, clearly there would have been no room at all for any theory of progressive evolution. And, by parity of reasoning, in whatever degree such a state of matters were found to prevail, in that degree would the theory in question have been discredited. But seeing that these opposite principles do not prevail in any (relatively speaking) considerable degree, we have so far positive testimony of the largest and most massive character in favour of this theory. For while all these large and general facts are very much what they ought to be according to this theory, they cannot be held to lend any support at all to the rival theory. In other words, it is clearly no essential part of the theory of special creation that species should everywhere exhibit this gradual multiplication as to number, coupled with a gradual diversification and general elevation of types, in all the growing branches of the tree of life. No one could adopt seriously the jocular lines of Burns, to the effect that the Creator required to practise his prentice hand on lower types before advancing to the formation of higher. Yet, without some such assumption, it would be impossible to explain, on the theory of independent creations, why there should have been this gradual advance from the few to the many, from the general to the special, from the low to the high.
 For objections which may be brought against this and similar statements, see the Appendix.
- - Epochs and Formations. Faunal Characters. C a - i POST-PLIOCENE. Man. Mammalia principally of living n Glacial Period. species. Mollusca exclusively recent. o + - z PLIOCENE, 3,000 feet. Mammalia principally of recent genera o living species rare. Mollusca very i modern. c + - MIOCENE, 4,000 ft. Mammalia principally of living o families; extinct genera numerous; r species all extinct. Mollusca largely OLIGOCENE, 8,000 ft. of recent species. T + - e EOCENE, 10,000 ft. Mammalia with numerous extinct families r and orders; all the species and i most of the genera extinct. Modern a type Shell-Fish. r y + - - LARAMIE, 4,000 ft. Passage beds. - M CRETACEOUS, 12,000 ft. Dinosaurian (bird-like) Reptiles; e Chalk. Pterodactyls (flying Reptiles); s toothed Birds; earliest Snake; bony o Fishes; Crocodiles; Turtles; z Ammonites. o + - i JURASSIC, 6,000 ft. Earliest Birds; giant Reptiles c Oolite. (Ichthyosaurs, Dinosaurs, Lias. Pterodactyls); Ammonites; Clam- and o Snail-Shells very abundant; decline r of Brachiopods; Butterfly. + - S TRIAS, 5,000 ft. First Mammalian (Marsupial); 2-gilled e New Red Sandstone. Cephalopods (Cuttle-Fishes, c Belemnites); reptilian Foot-Prints. o n d a r y - + - P PERMIAN, 5,000 ft. Earliest true Reptiles. a + - l CARBONIFEROUS, 26,000 ft. Earliest Amphibian (Labyrinthodont); e extinction of Trilobites; first o Coal. Cray-fish; Beetles; Cockroaches; z Centipedes; Spiders. o + - i DEVONIAN, 18,000 ft. Cartilaginous and Ganoid Fishes; c Old Red Sandstone. earliest and (snail) and freshwater Shells; Shell-Fish abundant; decline o of Trilobites; May-flies; Crab. r + - SILURIAN, 33,000 ft. Earliest Fish; the first Air-Breathers P (Insect, Scorpion); Brachiopods and r 4-gilled Cephalopods very abundant; i Trilobites; Corals; Graptolites. m - a CAMBRIAN, 24,000 ft. Trilobites; Brachiopod Mollusks. r y - - A ARCHAEEAN, 30,000 ft. z Huronian. Eozooen, (probably not a fossil). o Laurentian. i + - c PRIMEVAL. Non-sedimentary. +
I submit, then, that so far as the largest and most general principles in the matter of palaeontology are concerned, we have about as strong and massive a body of evidence as we could reasonably expect this branch of science to yield; for it is at once enormous in amount and positive in character. Therefore, if I do not further enlarge upon the evidence which we here have, as it were en masse, it is only because I do not feel that any words could add to its obvious significance. It may best be allowed to speak for itself in the millions of facts which are condensed in this tabular statement of the order of succession of all the known forms of animal life, as presented by the eminent palaeontologist, Professor Cope.
 For difficulties and objections, see Appendix.
Or, taking a still more general survey, this tabular statement may be still further condensed, and presented in a diagrammatic form, as it has been by another eminent American palaeontologist, Prof. Le Conte, in his excellent little treatise on Evolution and its Relations to Religious Thought. The following is his diagrammatic representation, with his remarks thereon.
When each ruling class declined in importance, it did not perish, but continued in a subordinate position. Thus, the whole organic kingdom became not only higher and higher in its highest forms, but also more and more complex in its structure and in the interaction of its correlated parts. The whole process and its result is roughly represented in the accompanying diagram, in which A B represents the course of geological time, and the curve, the rise, culmination, and decline of successive dominate classes.
I will here leave the evidence which is thus yielded by the most general principles that have been established by the science of palaeontology; and I will devote the rest of this chapter to a detailed consideration of a few highly special lines of evidence. By thus suddenly passing from one extreme to the other, I hope to convey the best idea that can be conveyed within a brief compass of the minuteness, as well as the extent, of the testimony which is furnished by the rocks.
* * * * *
When Darwin first published his Origin of Species, adverse critics fastened upon the "missing-link" argument as the strongest that they could bring against the theory of descent. Although Darwin had himself strongly insisted on the imperfection of the geological record, and the consequent precariousness of any negative conclusions raised upon it, these critics maintained that he was making too great a demand upon the argument from ignorance—that, even allowing for the imperfection of the record, they would certainly have expected at least a few cases of testimony to specific transmutation. For, they urged in effect, looking to the enormous profusion of the extinct species on the one hand, and to the immense number of known fossils on the other, it was incredible that no satisfactory instances of specific transmutation should ever have been brought to light, if such transmutation had ever occurred in the universal manner which the theory was bound to suppose. But since Darwin first published his great work palaeontologists have been very active in discovering and exploring fossiliferous beds in sundry parts of the world; and the result of their labours has been to supply so many of the previously missing links that the voice of competent criticism in this matter has now been well-nigh silenced. Indeed, the material thus furnished to an advocate of evolution at the present time is so abundant that his principal difficulty is to select his samples. I think, however, that the most satisfactory result will be gained if I restrict my exposition to a minute account of some few series of connecting links, rather than if I were to take a more general survey of a larger number. I will, therefore, confine the survey to the animal kingdom, and there mention only some of the cases which have yielded well-detailed proof of continuous differentiation.
It is obvious that the parts of animals most likely to have been preserved in such a continuous series of fossils as the present line of evidence requires, would have been the hard parts. These are horns, bones, teeth, and shells. Therefore I will consider each of these four classes of structures separately.
* * * * *
Horns wherever they occur, are found to be of high importance for purposes of classification. They are restricted to the Ruminants, and appear under three different forms or types—namely solid, as in antelopes; hollow, as in sheep; and deciduous, as in deer. Now, in each of these divisions we have a tolerably complete palaeontological history of the evolution of horns. The early ruminants were altogether hornless (Fig. 60). Then, in the middle Miocene, the first antelopes appeared with tiny horns, which progressively increased in size among the ever-multiplying species of antelopes until the present day. But it is in the deer tribe that we meet with even better evidence touching the progressive evolution of horns; because here not only size, but shape, is concerned. For deer's horns, or antlers, are arborescent; and hence in their case we have an opportunity of reading the history, not only of a progressive growth in size, but also of an increasing development of form. Among the older members of the tribe, in the lower Miocene, there are no horns at all. In the mid-Miocene we meet with two-pronged horns (Cervus dicrocerus, Figs. 61, 62, 1/5 nat. size). Next, in the upper Miocene (C. matheronis, Fig. 63, 1/8 nat. size), and extending into the Pliocene (C. pardinensis, Fig. 64, 1/18 nat. size), we meet with three-pronged horns. Then, in the Pliocene we find also four-pronged horns (C. issiodorensis, Fig. 65, 1/16 nat. size), leading us to five-pronged (C. tetraceros). Lastly, in the Forest-bed of Norfolk we meet with arborescent horns (C. Sedgwickii, Fig. 66, 1/35 nat. size). The life-history of existing stags furnishes a parallel development (Fig. 67), beginning with a single horn (which has not yet been found palaeontologically), going on to two prongs, three prongs, four prongs, and afterwards branching.
* * * * *
Coming now to bones, we have a singularly complete record of transition from one type or pattern of structure to another in the phylogenetic history of tails. This has been so clearly and so tersely conveyed by Prof. Le Conte, that I cannot do better than quote his statement.
It has long been noticed that there are among fishes two styles of tail-fins. These are the even-lobed, or homocercal (Fig. 68), and the uneven-lobed, or heterocercal (Fig. 69). The one is characteristic of ordinary fishes (teleosts), the other of sharks and some other orders. In structure the difference is even more fundamental than in form. In the former style the backbone stops abruptly in a series of short, enlarged joints, and thence sends off rays to form the tail-fin (Fig. 68); in the latter the backbone runs through the fin to its very point, growing slenderer by degrees, and giving off rays above and below from each joint, but the rays on the lower side are much longer (Fig. 69). This type of fin is, therefore, vertebrated, the other non-vertebrated. Figs. 68 and 69 show these two types in form and structure. But there is still another type found only in the lowest and most generalized forms of fishes. In these the tail-fin is vertebrated and yet symmetrical. This type is shown in Fig. 70.
Now, in the development of a teleost fish (Fig. 68), as has been shown by Alexander Agassiz, the tail-fin is first like Fig. 70; then becomes heterocercal, like Fig. 69; and, finally, becomes homocercal like Fig. 68. Why so? Not because there is any special advantage in this succession of forms; for the changes take place either in the egg or else in very early embryonic states. The answer is found in the fact that this is the order of change in the phylogenetic series. The earliest fish-tails were either like Fig. 69 or Fig. 70; never like Fig. 68. The earliest of all were almost certainly like Fig. 70; then they became like Fig. 69; and, finally, only much later in geological history (Jurassic or Cretaceous), they became like Fig. 68. This order of change is still retained in the embryonic development of the last introduced and most specialized order of existing fishes. The family history is repeated in the individual history.
Similar changes have taken place in the form and structure of birds' tails. The earliest bird known—the Jurassic Archaeopteryx—had a long reptilian tail of twenty-one joints, each joint bearing a feather on each side, right and left (Fig. 71): [see also Fig. 73]. In the typical modern bird, on the contrary, the tail-joints are diminished in number, shortened up, and enlarged, and give out long feathers, fan-like, to form the so-called tail (Fig. 72). The Archaeopteryx' tail is vertebrated, the typical bird's non-vertebrated. This shortening up of the tail did not take place at once, but gradually. The Cretaceous birds, intermediate in time, had tails intermediate in structure. The Hesperornis of Marsh had twelve joints. At first—in Jurassic strata—the tail is fully a half of the whole vertebral column. It then gradually shortens up until it becomes the aborted organ of typical modern birds. Now, in embryonic development, the tail of the modern typical bird passes through all these stages. At first the tail is nearly one half the whole vertebral column; then, as development goes on, while the rest of the body grows, the growth of the tail stops, and thus finally becomes the aborted organ we now find. The ontogeny still passes through the stages of the phylogeny. The same is true of all tailless animals.
The extinct Archaeopteryx above alluded to presents throughout its whole organization a most interesting assemblage of "generalized characters." For example, its teeth, and its still unreduced digits of the wings (which, like those of the feet, are covered with scales), refer us, with almost as much force as does the vertebrated tail, to the Sauropsidian type—or the trunk from which birds and reptiles have diverged.
We will next consider the palaeontological evidence which we now possess of the evolution of mammalian limbs, with special reference to the hoofed animals, where this line of evidence happens to be most complete.
I may best begin by describing the bones as these occur in the sundry branches of the mammalian type now living. As we shall presently see, the modifications which the limbs have undergone in these sundry branches chiefly consist in the suppression of some parts and the exaggerated development of others. But, by comparing all mammalian limbs together, it is easy to obtain a generalized type of mammalian limb, which in actual life is perhaps most nearly conformed to in the case of bears. I will therefore choose the bear for the purpose of briefly expounding the bones of mammalian limbs in general—merely asking it to be understood, that although in the case of many other mammalia some of these bones may be dwindled or altogether absent, while others may be greatly exaggerated as to relative size, in no case do any additional bones appear.
On looking, then, at the skeleton of a bear (Fig. 74), the first thing to observe is that there is a perfect serial homology between the bones of the hind legs and of the fore legs. The thigh-bone, or femur, corresponds to the shoulder-bone, or humerus; the two shank bones (tibia and fibula) correspond to the two arm-bones (radius and ulna); the many little ankle-bones (tarsals) correspond to the many little wrist-bones (carpals); the foot-bones (meta-tarsals) correspond to the hand-bones (meta-carpals); and, lastly, the bones of each of the toes correspond to those of each of the fingers.
The next thing to observe is, that the disposition of bones in the case of the bear is such that the animal walks in the way that has been called plantigrade. That is to say, all the bones of the fingers, as well as those of the toes, feet, and ankles, rest upon the ground, or help to constitute the "soles." Our own feet are constructed on a closely similar pattern. But in the majority of living mammalian forms this is not the case. For the majority of mammals are what has been called digitigrade. That is to say, the bones of the limb are so disposed that both the foot and hand bones, and therefore also the ankle and wrist, are removed from the ground altogether, so that the animal walks exclusively upon its toes and fingers—as in the case of this skeleton (Fig. 75), which is the skeleton of a lion. The next figures display a series of limbs, showing the progressive passage of a completely plantigrade into a highly digitigrade type—the curved lines of connexion serving to indicate the homologous bones (Figs. 76, 77).
I will now proceed to detail the history of mammalian limbs, as this has been recorded for us in fossil remains.
The most generalized or primitive types of limb hitherto discovered in any vertebrated animal above the class of fishes, are those which are met with in some of the extinct aquatic reptiles. Here, for instance, is a diagram of the left hind limb of Baptanodon discus (Fig. 78). It has six rows of little symmetrical bones springing from a leg-like origin. But the whole structure resembles the fin of a fish about as nearly as it does the leg of a mammal. For not only are there six rows of bones, instead of five, suggestive of the numerous rays which characterise the fin of a fish; but the structure as a whole, having been covered over with blubber and skin, was throughout flexible and unjointed—thus in function, even more than in structure, resembling a fin. In this respect, also, it must have resembled the paddle of a whale (see Fig. 79); but of course the great difference will be noted, that the paddle of a whale reveals the dwindled though still clearly typical bones of a true mammalian limb; so that although in outward form and function these two paddles are alike, their inward structure clearly shows that while the one testifies to the absence of evolution, the other testifies to the presence of degeneration. If the paddle of Baptanodon had occurred in a whale, or the paddle of a whale had occurred in Baptanodon, either fact would in itself have been well-nigh destructive of the whole theory of evolution.
Such, then, is the most generalized as it is the most ancient type of vertebrate limb above the class of fishes. Obviously it is a type suited only to aquatic life. Consequently, when aquatic Vertebrata began to become terrestrial, the type would have needed modification in order to serve for terrestrial locomotion. In particular, it would have needed to gain in consolidation and in firmness, which means that it would have needed also to become jointed. Accordingly, we find that this archaic type gave place in land-reptiles to the exigencies of these requirements. Here for example is a diagram, copied from Gegenbaur, of the right fore-foot of Chelydra serpentina (Fig. 78). As compared with the homologous limb of its purely aquatic predecessor, there is to be noticed the disappearance of one of the six rows of small bones, a confluence of some of the remainder in the other five rows, a duplication of the arm-bone into a radius and ulna, in order to admit of jointed rotation of the hand, and a general disposition of the small bones below these arm-bones, which clearly foreshadows the joint of the wrist. Indeed, in this fore-foot of Chelydra, a child could trace all the principal homologies of the mammalian counterpart, growing, like the next stage in a dissolving view, out of the primitive paddle of Baptanodon—namely, first the radius and ulna, next the carpals, then the meta-carpals, and, lastly, the three phalanges in each of the five digits.
Such a type of foot no doubt admirably meets the requirements of slow reptilian locomotion over swampy ground. But for anything like rapid locomotion over hard and uneven ground, greater modifications would be needed. Such modifications, however, need not be other in kind: it is enough that they should continue in the same line of advance, so as to reach a higher degree of firmness, combined with better joints. Accordingly we find that this took place, not indeed among reptiles, whose habits of cold-blooded life have not changed, but among their warm-blooded descendants, the mammals. Moreover, when we examine the whole mammalian series, we find that the required modifications must have taken place in slightly different ways in three lines of descent simultaneously. We have first the plantigrade and digitigrade modifications already mentioned (pp. 178, 179) Of these the plantigrade walking entailed least change, because most resembling the ancestral or lizard-like mode of progression. All that was here needed was a general improvement as to relative lengths of bones, with greater consolidation and greater flexibility of joints. Therefore I need not say anything more about the plantigrade division. But the digitigrade modification necessitated a change of structural plan, to the extent of raising the wrist and ankle joints off the ground, so as to make the quadruped walk on its fingers and toes. We meet with an interesting case of this transition in the existing hare, which while at rest supports itself on the whole hind foot after the manner of a plantigrade animal, but when running does so upon the ends of its toes, after the manner of a digitigrade animal.
It is of importance for us to note that this transition from the original plantigrade to the more recent digitigrade type, has been carried out on two slightly different plans in two different lines of mammalian descent. The hoofed mammals—which are all digitigrade—are sub-classified as artiodactyls and perissodactyls, i. e. even-toed and odd-toed. Now, whether an animal has an even or an odd number of toes may seem a curiously artificial distinction on which to found so important a classification of the mammalian group. But if we look at the matter from a less empirical and more intelligent point of view, we shall see that the alternative of having an even or an odd number of toes carries with it alternative consequences of a practically important kind to any animal of the digitigrade type. For suppose an aboriginal five-toed animal, walking on the ends of its five toes, to be called upon to resign some of his toes. If he is left with an even number, it must be two or four; and in either case the animal would gain the firmest support by so disposing his toes as to admit of the axis of his foot passing between an equal number of them—whether it be one or two toes on each side. On the other hand, if our early mammal were called upon to retain an odd number of toes, he would gain best support by adjusting matters so that the axis of his foot should be coincident with his middle toe, whether this were his only toe, or whether he had one on either side of it. This consideration shows that the classification into even-toed and odd-toed is not so artificial as it no doubt at first sight appears. Let us, then, consider the stages in the evolution of both these types of feet.
Going back to the reptile Chelydra, it will be observed that the axis of the foot passes down the middle toe, which is therefore supported by two toes on either side (Fig. 78). It may also be noticed that the wrist or ankle bones do not interlock, either with one another or with the bones of the hand or foot below them. This, of course, would give a weak foot, suited to slow progression over marshy ground—which, as we have seen, was no doubt the origin of the mammalian plantigrade foot. Here, for instance, to all intents and purposes, is a similar type of foot, which belonged to a very early mammal, antecedent to the elephant series, the horse series, the rhinoceros, the hog, and, in short, all the known hoofed mammalia (Fig. 80). It was presumably an inhabitant of swampy ground, slow in its movements, and low in its intelligence.
But now, as we have seen, for more rapid progression on hard uneven ground, a stronger and better jointed foot would be needed. Therefore we find the bones of the wrist and ankle beginning to interlock, both among themselves and also with those of the foot and hand immediately below them. Such a stage of evolution is still apparent in the now existing elephant. (See Fig. 81.)
Next, however, a still stronger foot was made by the still further interlocking of the wrist and ankle bones, so that both the first and second rows of them were thus fitted into each other, as well as into the bones of the hand and foot beneath. This further modification is clearly traceable in some of the earlier perissodactyls, and occurs in the majority at the present time. Compare, for example, the greater interlocking and consolidation of these small bones in the Rhinoceros as contrasted with the Elephant (Fig. 81). Moreover, simultaneously with these consolidating improvements in the mechanism of the wrist and ankle joints, or possibly at a somewhat later period, a reduction in the number of digits began to take place. This was a continuation of the policy of consolidating the foot, analogous to the dropping out of the sixth row of small bones in the paddle of Baptanodon. (Fig. 78.) In the pentadactyl plantigrade foot of the early mammals, the first digit, being the shortest, was the first to leave the ground, to dwindle, and finally to disappear. More work being thus thrown on the remaining four, they were strengthened by interlocking with the wrist (or ankle) bones above them, as just mentioned; and also by being brought closer together.
The changes which followed I will render in the words of Professor Marsh.
Two kinds of reduction began. One leading to the existing perissodactyl foot, and the other, apparently later, resulting in the artiodactyl type. In the former the axis of the foot remained in the middle of the third digit, as in the pentadactyl foot. [See Fig. 81.] In the latter, it shifted to the outer side of this digit, or between the third and fourth toe. [See Fig. 82.]
In the further reduction of the perissodactyl foot, the fifth digit, being shorter than the remaining three, next left the ground, and gradually disappeared. [Fig. 81 B.] Of the three remaining toes, the middle or axial one was the longest, and retaining its supremacy as greater strength and speed were required, finally assumed the chief support of the foot [Fig. 81 C], while the outer digits left the ground, ceased to be of use, and were lost, except as splint-bones [Fig. 81 D]. The feet of the existing horse shows the best example of this reduction in the Perissodactyls, as it is the most specialized known in the Ungulates [Fig. 81 D].
In the artiodactyl foot, the reduction resulted in the gradual diminution of the two outer of the four remaining toes, the third and fourth doing all the work, and thus increasing in size and power. The fifth digit, for the same reasons as in the perissodactyl foot, first left the ground and became smaller. Next, the second soon followed, and these two gradually ceased to be functional, [and eventually disappeared altogether, as shown in the accompanying drawing of the feet of still existing animals, Fig. 82 B, C, D].
The limb of the modern race-horse is a nearly perfect piece of machinery, especially adapted to great speed on dry, level ground. The limb of an antelope, or deer, is likewise well fitted for rapid motion on a plain, but the foot itself is adapted to rough mountain work as well, and it is to this advantage, in part, that the Artiodactyls owe their present supremacy. The plantigrade pentadactyl foot of the primitive Ungulate—and even the perissodactyl foot that succeeded it—both belong to the past humid period of the world's history. As the surface of the earth slowly dried up, in the gradual desiccation still in progress, new types of feet became a necessity, and the horse, antelope, and camel were gradually developed, to meet the altered conditions.
The best instance of such progressive modifications in the case of perissodactyl feet is furnished by the fossil pedigree of the existing horse, because here, within the limits of the same continuous family line, we have presented the entire series of modifications.
There are now known over thirty species of horse-like creatures, beginning from the size of a fox, then progressively increasing in bulk, and all standing in linear series in structure as in time. Confining attention to the teeth and feet, it will be seen from the wood-cut on page 189 that the former grow progressively longer in their sockets, and also more complex in the patterns of their crowns. On the other hand, the latter exhibit a gradual diminution of their lateral toes, together with a gradual strengthening of the middle one. (See Fig. 83.) So that in the particular case of the horse-ancestry we have a practically complete chain of what only a few years ago were "missing links." And this now practically completed chain shows us the entire history of what happens to be the most peculiar, or highly specialized, limb in the whole mammalian class—namely, that of the existing horse. Of the other two wood-cuts, the former (Fig. 84) shows the skeleton of a very early and highly generalized ancestor, while the other is a partial restoration of a much more recent and specialized one. (Fig. 85.)
On the other hand, progressive modifications of the artiodactyl feet may be traced geologically up to the different stages presented by living ruminants, in some of which it has proceeded further than in others. For instance, if we compare the pig, the deer, and the camel (Fig. 82), we immediately perceive that the dwindling of the two rudimentary digits has proceeded much further in the case of the deer than in that of the pig, and yet not so far as in that of the camel, seeing that here they have wholly disappeared. Moreover, complementary differences are to be observed in the degree of consolidation presented by the two useful digits. For while in the pig the two foot-bones are still clearly distinguishable throughout their entire length, in the deer, and still more in the camel, their union is more complete, so that they go to constitute a single bone, whose double or compound character is indicated externally only by a slight bifurcation at the base. Nevertheless, if we examine the state of matters in the unborn young of these animals, we find that the two bones in question are still separated throughout their length, and thus precisely resemble what used to be their permanent condition in some of the now fossil species of hoofed mammalia.
Turning next from bones of the limb to other parts of the mammalian skeleton, let us briefly consider the evidence of evolution that is here likewise presented by the vertebral column, the skull, and the teeth.
As regards the vertebral column, if we examine this structure in any of the existing hoofed animals, we find that the bony processes called zygapophyses, which belong to each of the constituent vertebrae, are so arranged that the anterior pair belonging to each vertebra interlocks with the posterior pair belonging to the next vertebra. In this way the whole series of vertebrae are connected together in the form of a chain, which, while admitting of considerable movement laterally, is everywhere guarded against dislocation. But if we examine the skeletons of any ungulates from the lower Eocene deposits, we find that in no case is there any such arrangement to secure interlocking. In all the hoofed mammals of this period the zygapophyses are flat. Now, from this flat condition to the present condition of full interlocking we obtain a complete series of connecting links. In the middle Miocene period we find a group of hoofed animals in which the articulation begins by a slight rounding of the previously flat surfaces: later on this rounding progressively increases, until eventually we get the complete interlocking of the present time.
As regards teeth, and still confining attention to the hoofed mammals, we find that low down in the geological series the teeth present on their grinding surfaces only three simple tubercles. Later on a fourth tubercle is added, and later still there is developed that complicated system of ridges and furrows which is characteristic of these teeth at the present time, and which was produced by manifold and various involutions of the three or four simple tubercles of Eocene and lower Miocene times. In other words, the principle of gradual improvement in the construction of teeth, which has already been depicted as regards the particular case of the Horse-family (Fig. 83), is no less apparent in the pedigree of all the other mammalia, wherever the palaeontological history is sufficiently intact to serve as a record at all.
Lastly, as regards the skull, casts of the interior show that all the earlier mammals had small brains with comparatively smooth or unconvoluted surfaces; and that as time went on the mammalian brain gradually advanced in size and complexity. Indeed so small were the cerebral hemispheres of the primitive mammals that they did not overlap the cerebellum, while their smoothness must have been such as in this respect to have resembled the brain of a bird or reptile. This, of course, is just as it ought to be, if the brain, which the skull has to accommodate, has been gradually evolved into larger and larger proportions in respect of its cerebral hemispheres, or the upper masses of it which constitute the seat of intelligence. Thus, if we look at the above series of wood-cuts, which represents the comparative structure of the brain in the existing classes of the Vertebrata, we can immediately understand why the fossil skulls of Mammalia should present a gradual increase in size and furrowing, so as to accommodate the general increase of the brain in both these respects between the level marked "maml" and that marked "man," in the last of the diagrams. (Fig. 87.)
[Illustration: FIG. 86.—Comparative series of Brains. (After Le Conte.) The series reads from above downwards, and represents diagrammatically the brain of a Fish, a Reptile, a Bird, a Mammal, and a Man. In each case the letter A marks a side view, and the letter B a top view. The small italics throughout signify the following homologous parts: m, medulla; cb, cerebellum; op, optic lobes; cr, cerebrum and thalamus; ol, olfactory lobes. The series shows a progressive consolidation and enlargement of the brain in general, and of the cerebrum and cerebellum in particular, which likewise exhibit continually advancing structure in respect of convolution. In the case of Man, these two parts of the brain have grown to so great a size that they conceal all the other parts from the superficial points of view represented in the diagram.]
The tabular statement on the following diagram, which I borrow from Prof. Cope, will serve at a glance to reveal the combined significance of so many lines of evidence, united within the limits of the same group of animals.
To give only one special illustration of the principle of evolution as regards the skull, here is one of the most recent instances that has occurred of the discovery of a missing link, or connecting form (see Fig. 88). The fossil (B), which was found in New Jersey, stands in an intermediate position between the stag and the elk. In the stag (A) the skull is high, showing but little of that anterior attenuation which is such a distinctive feature of the skull of the elk (C). The nasal bones (N) of the former, again, are remarkably long when compared with the similar bones of the latter, and the premaxillaries (PMX), instead of being projected forward along the horizontal plane of the base of the skull, are deflected sharply downward. In all these points, it will be seen, the newly discovered form (Cervalces) holds an intermediate position (B). "The skull exhibits a partial attenuation anteriorly, the premaxillaries are directed about equally downward and forward, and the nasal bones are measurably contracted in size. The horns likewise furnish characters which further serve to establish this dual relationship."
 Heilprin, Geological Evidences of Evolution, pp. 73-4 (1888).
Formation. No. of toes Feet Astragalus. Carpus and tarsus. Ulno-radius. Superior molars. Zygapophyses. Brain. Pliocene. 1-1, 2-2 Digitigrade. (Plantigrade.) Grooved. (Flat.) Interlocking. (Opposite.) Faceted. 4-tubercles, crested and cemented. Doubly involute. Singly involute. Hemispheres larger, convoluted. Upper Miocene. (Loup Fork.) 3-3, 4-4, (5-5) Middle. (John Day.) 2-2, 3-3, 4-4 Digitigrade. Grooved. Interlocking. Faceted. Smooth 4-tubercles, and crested Singly involute. Double involute. Hemispheres larger, convoluted. Lower (White River.) 3-3, 4-3 Digitigrade. Plantigrade. Grooved. Interlocking. Smooth. Faceted. 4-tubercles, and crested ? Singly involute. Hemispheres small, and largeer. Eocene. Upper (Bridger.) 3-3, 4-3, 4-5 (Digitigrade.) Plantigrade. Grooved. (Flat.) Opposite. Interlocking. Smooth. 4-tubercles. 3-tubercles, and crested. Singly involute. Plane Hemispheres small. Middle. (Wasatch.) 4-3, 4-5, 5-5 Plantigrade. (Digitigrade.) Flat. (Grooved.) Opposite. Interlocking. Smooth. 4-tubercles. 3-tubercles, a few crested. Plane. Singly involute. Hemispheres small; mesencephalon sometimes exposed Lower (Puerco.) 5-5 Plantigrade. Flat. Opposite. Smooth. 3-tubercles. (4-tubercles), none crested. Plane. Mesencephalon exposed; hemisphere small and smoother.
The evidence, then, which is furnished by all parts of the vertebral skeleton—whether we have regard to Fishes, Reptiles, Birds, or Mammals—is cumulative and consistent. Nowhere do we meet with any deviation or ambiguity, while everywhere we encounter similar proofs of continuous transformation—proofs which vary only with the varying amount of material which happens to be at our disposal, being most numerous and detailed in those cases where the greatest number of fossil forms has been preserved by the geological record. Here, therefore, we may leave the vertebral skeleton; and, having presented a sample of the evidence as yielded by horns and bones, I will conclude by glancing with similar brevity at the case of shells—which, as before remarked, constitute the only other sufficiently hard or permanent material to yield unbroken evidence touching the fossil ancestry of animals.
Of course it will be understood that I am everywhere giving merely samples of the now superabundant evidence which is yielded by palaeontology; and, as this chapter is already a long one, I must content myself with citing only the case of mollusk-shells, although shells of other classes might be made to yield highly important additions to the testimony. Moreover, even as regards the one division of mollusk-shells, I can afford to quote only a very few cases. These, however, are in my opinion the strongest single pieces of evidence in favour of transmutation which have thus far been brought to light.
Near the village of Steinheim, in Wuertemberg, there is an ancient lake-basin, dating from Tertiary times. The lake has long ago dried up; but its aqueous deposits are extraordinarily rich in fossil shells, especially of different species of the genus Planorbis. The following is an authoritative resume of the facts.
As the deposits seem to have been continuous for ages, and the fossil shells very abundant, this seemed to be an excellent opportunity to test the theory of derivation. With this end in view, they have been made the subject of exhaustive study by Hilgendorf in 1866, and by Hyatt in 1880. In passing from the lowest to the highest strata the species change greatly and many times, the extreme forms being so different that, were it not for the intermediate forms, they would be called not only different species, but different genera. And yet the gradations are so insensible that the whole series is nothing less than a demonstration, in this case at least, of origin of species by derivation with modifications. The accompanying plate of successive forms (Fig. 89), which we take from Prof. Hyatt's admirable memoir, will show this better than any mere verbal explanation. It will be observed that, commencing with four slight varieties—probably sexually isolated varieties—of one species, each series shows a gradual transformation as we go upward in the strata—i. e. onward in time. Series I branches into three sub-series, in two of which the change of form is extreme. Series IV is remarkable for great increase in size as well as change in form. In the plate we give only selected stages, but in the fuller plates of the memoir, and still more in the shells themselves, the subtilest gradations are found.
 Le Conte, loc. cit., pp. 236-7.
Here is another and more recently observed case of transmutation in the case of mollusks.
The recent species, Strombus accipitrinus, still inhabits the coasts of Florida. Its extinct prototype, S. Leidy, was discovered a few years ago by Prof. Heilprin in the Pliocene formations of the interior of Florida. The peculiar shape of the wing, and tuberculation of the whorl, are thus proved to have grown but of a previously more conical form of shell.
Lastly, attention may here again be directed to the very instructive series of shells which has already been shown in a previous chapter, and which serves to illustrate the successive geological forms of Paludina from the Tertiary beds of Slavonia, as depicted by Prof. Neumayr of Vienna. (Fig. 1, p. 19.)
The argument from geology is the argument from the distribution of species in time. I will next take the argument from the distribution of species in space—that is, the present geographical distribution of plants and animals.
Seeing that the theory of descent with adaptive modification implies slow and gradual change of one species into another, and progressively still more slow and gradual changes of one genus, family, or order into another genus, family, or order, we should expect on this theory that the organic types living on any given geographical area would be found to resemble or to differ from organic types living elsewhere, according as the area is connected with or disconnected from other geographical areas. For instance, the large continental islands of Australia and New Zealand are widely disconnected from all other lands of the world, and deep sea soundings show that they have probably been thus disconnected, either since the time of their origin, or, at the least, through immense geological epochs. The theory of evolution, therefore, would expect to find two general facts with regard to the inhabitants of these islands. First, that the inhabitants should form, as it were, little worlds of their own, more or less unlike the inhabitants of any other parts of the globe. And next, that some of these inhabitants should present us with independent information touching archaic forms of life. For it is manifestly most improbable that the course of evolutionary history should have run exactly parallel in the case of these isolated oceanic continents and in continents elsewhere. Australia and New Zealand, therefore, ought to present a very large number, not only of peculiar species and genera, but even of families, and possibly of orders. Now this is just what Australia and New Zealand do present. The case of the dog being doubtful, there is an absence of all mammalian life, except that of one of the oldest and least highly developed orders, the Marsupials. There even occurs a unique order, still lower in the scale of organization—so low, in fact, that it deserves to be regarded as but nascent mammalian: I mean, of course, the Monotremata. As regards Birds, we have the peculiar wingless forms alluded to in a previous chapter (viz. that on Morphology); and, without waiting to go into details, it is notorious that the faunas of Australia and New Zealand are not only highly peculiar, but also suggestively archaic. Therefore, in both the respects above mentioned, the anticipations of our theory are fully borne out. But as it would take too long to consider, even cursorily, the faunas and floras of these immense islands, I here allude to them only for the sake of illustration. In order to present the argument from geographical distribution within reasonable limits, I think it is best to restrict our examination to smaller areas; for these will better admit of brief and yet adequate consideration. But of course it will be understood that the less isolated the region, and the shorter the time that it has been isolated, the smaller amount of peculiarity should we expect to meet with on the part of its present inhabitants. Or, conversely stated, the longer and the greater the isolation, the more peculiarity of species would our theory expect to find. The object of the present chapter will be to show that these, and other cognate expectations, are fully realized by facts; but, before proceeding to do this, I must say a few words on the antecedent standing of the argument.
Where the question is, as at present, between the rival theories of special creation and gradual transmutation, it may at first sight well appear that no test can be at once so crucial and so easily applied as this of comparing the species of one geographical area with those of another, in order to see whether there is any constant correlation between differences of type and degrees of separation. But a little further thought is enough to show that the test is not quite so simple or so absolute—that it is a test to be applied in a large and general way over the surface of the whole earth, rather than one to be relied upon as exclusively rigid in every special case.
In the first place, there is the obvious consideration that lands or seas which are discontinuous now may not always have been so, or not for long enough to admit of the effects of separation having been exerted to any considerable extent upon their inhabitants. Next, there is the scarcely less important consideration, that although land areas may long have been separated from one another by extensive tracts of ocean, birds and insects may more or less easily have been able to fly from one to the other; while even non-flying animals and plants may often have been transported by floating ice or timber, wind or water currents, and sundry other means of dispersal. Again, there is the important influence of climate to be taken into account. We know from geological evidence that in the course of geological time the self-same continents have been submitted to enormous changes of temperature—varying in fact from polar cold to almost tropical heat; and as it is manifestly impossible that forms of life suited to one of these climates could have survived during the other, we can here perceive a further and most potent cause interfering with the test of geographical distribution as indiscriminately applied in all cases. When the elephant and hippopotamus were flourishing in England amid the luxuriant vegetation which these large animals require, it is evident that scarcely any one species of either the fauna or the flora of this country can have been the same as it was when its African climate gave place to that of Greenland. Therefore, as Mr. Wallace observes, "If glacial epochs in temperate lands and mild climates near the poles have, as now believed by men of eminence, occurred several times over in the past history of the earth, the effects of such great and repeated changes both on migration, modification, and extinction of species, must have been of overwhelming importance—of more importance perhaps than even the geological changes of sea and land."