The law of heredity, I have said above, may be regarded as that of persistence exemplified in a series of organic generations. Variation results—it is clear that it must result—from some kind of differentiating influence. Such statements as these, however, though they are true enough, do not help us much in understanding either heredity or variation.

Let us first notice that normal cases of reproduction exemplify both phenomena—heredity with variation; hereditary similarity to the parents in all essential respects, individual variations in minor points. This is seen in man. Brothers and sisters may present family resemblances among each other and to their parents, but each has individual traits of feature and of character. Only in particular cases of so-called "identical twins" are the variations so slight as not to be readily perceptible by even a casual observer.

Now, when we seek an explanation of these well-known facts, we may be tempted to find it in the supposition that the character of the parents does not remain constant, that the character influences the offspring, and that therefore the children born at successive periods will differ from each other, while twins born in the same hour will naturally resemble each other. As Darwin himself says,[BD] "The greater dissimilarity of the successive children of the same family in comparison with twins, which often resemble each other in external appearance, mental disposition, and constitution, in so extraordinary a manner, apparently proves that the state of the parents at the exact period of conception, or the nature of the subsequent embryonic development, has a direct and powerful influence on the character of the offspring." But a little consideration will show that, though this might, in the absence of a better explanation, account for variation in character, it could not account for variation in form and feature, unless we regard these as in some way determined by the character. Moreover, as we shall see presently, it is open to question whether acquired modifications of structure or character in the parent can in any way influence the offspring. Again, in the litter of puppies born of the same bitch by the same dog there are individual variations, often as well marked as those in successive births.

The facts, then, to be accounted for are—first, the close hereditary resemblance in all essential points of offspring to parent; and, secondly, the individual differences in minor points among the offspring produced simultaneously or successively by the same parents. These are the facts as they occur in the higher animals. It will be well to lead up to our consideration of them by a preliminary survey of the facts as they are exemplified by some of the lower organisms.

In the simpler protozoa, where fission occurs, and where the organism is composed of a single cell, where also there is a single nucleus which apparently undergoes division into two equal and similar parts, it is easy to understand that the two organisms thus resulting from the halving of a single organism partake completely of its nature. If the fission of an amœba is such as to divide it into two similar parts, there is no reason why these two similar parts should not be in all respects alike, and should not, by the assimilation of new material, acquire the size and all the characteristics of the parent form. In the higher and more differentiated protozoa, the case is not quite so simple; for the two halves are not each like the whole parent, but have to be remodelled into a similar organism. But if we suppose, as we seem to have every right to suppose, that it is the nucleus that controls the formative processes in the cell, there is not much difficulty in understanding how, when the nucleus divides into two similar portions, each directs, so to speak, the similar refashioning of its own separated protoplasmic territory.

From the protozoa we may pass to such a comparatively simple metazoon as the hydra. Here the organism is composed, not of a single cell, but of a number of cells. These cells are, moreover, not all alike, but have undergone differentiation with physiological division of labour. There is an inner layer of large nutritive cells, and an outer layer of protective cells, some of which are conical with fine processes proceeding from the point of the cone; others are smaller, and fill in the interstices between the apices of the cones, while others have developed into thread-cells, each with a fine stinging filament. Between the two layers there is a thin supporting lamella. The essential point we have here to notice is that there are two distinct layers with cells of different form and function.

Now, it has again and again been experimentally proved that if a hydra be divided into a number of fragments, each will grow up into a complete and perfect hydra. All that is essential is that, in the separated fragment, there shall be samples of the cells of both layers. Under these conditions, the separated cells of the outer layer regenerate a complete external wall, and the separated cells of the inner layer similarly regenerate a complete internal lining. From these facts, it would appear that such a small adequately sampled fragment has the power, when isolated, of assimilating nutriment and growing by the multiplication of the constituent cells, and that the growth takes such lines that the original form of the hydra is reproduced.

Here we may note, by way of analogy, what takes place in the case of inorganic crystals. If a fragment of an alum crystal be suspended in a strong solution of alum, the crystal will be recompleted by the growth of new parts along the broken edges. We say that this is effected under the influence of molecular polarity. Similarly, we may say that the fragment of the hydra rebuilds the complete form under the influence of an hereditary morphological tendency residing in the nuclei of the several cells. The case, though still comparatively simple, is more complex than that of the higher protozoa. There the divided nucleus in two separated cells directs each of these in hereditary lines of morphological growth. Here not only do the cells and their nuclei divide, but they are animated by a common morphological principle, and in their multiplication combine to form an organism possessing the ancestral symmetry. If, however, we call this an hereditary morphological tendency or a principle of symmetry; or, with the older physiologists, a nisus formativus; or, with Darwin, "the co-ordinating power of the organization" (all of these expressions being somewhat unsatisfactory);—we must remember that these terms merely imply a play of molecular forces analogous to that which causes the broken crystal of alum to become recompleted in suitable solution. The inherent molecular processes in the nuclei[BE] in the one case enable the cells to regenerate the hydra; the inherent molecular stresses in the crystalline fragment in the other case lead to the reproduction of the complete crystal. In either case there is no true explanation, but merely a restatement of the facts under a convenient name or phrase.

The power of regeneration of lost parts, which is thus seen in the hydra, is also seen, in a less degree so far as amount is concerned, but in a higher degree so far as complexity goes, in animals far above the hydra in the scale of life. The lobster that has lost a claw, the snail whose tentacle has been removed, the newt which has been docked of a portion of its tail or a limb, are able more or less completely to regenerate these lost parts. And the regeneration may involve complex structures. With the tentacle of the snail the eye may be removed, and this, not once only, but a dozen times. After such mutilation, no part of the eye remains, though the stump of its nerve is, of course, left; still the perfect organ is reconstructed again and again, as often as the tentacle is removed. The cells at the cut end of the nerve-stump divide and multiply, as do also those of the surrounding tissues, and the growing nerve terminates in an optic cup, as it did previously under the influences of normal development before the mutilation. Here we have phenomena analogous to, and in some respects more complex than, those which are seen in the regenerative process in hydra. It is well known, however, that, in the case of higher animals, in birds and mammals, this power of regenerating lost parts does not exist. When a bone is broken, osseous union of the broken pieces may indeed take place; and in flesh-wounds, the gash is filled in and heals over, not without permanent signs of its existence, as may often be seen in the faces of German students. But beyond this there is normally no regeneration. The soldier who has lost an arm in battle cannot return home and in quiet seclusion reproduce a new limb. That which seems to be among lower animals a well-established law of organic growth does not here obtain. This is probably due to the fact that the higher histological differentiation of the tissues in the more highly developed forms of life is a bar to regeneration. In their devotion to special and minute details of physiological work, the cells have, so to speak, forgotten their more generalized reproductive faculties. In any case, however the fact is to be explained, the higher organisms have in many cases almost completely lost the power of regenerating lost parts. But this loss of the regenerative power in the more highly differentiated animals does not alter or invalidate the law of organic growth we are considering. The law may be thus stated: Whenever, after mutilation, free growth of the mutilated surface occurs, that growth is directed in such lines as to reproduce the lost part and restore the symmetrical integrity of the organism. This is a matter of heredity. And we may regard the hereditary reconstructive power as residing either (1) in those cells at or adjoining the mutilated surface which are concerned in the regrowth of the lost part; or (2) in the general mass of cells of the mutilated organism.

There are difficulties in either view. Professor Sollas, supporting the former, says,[BF] "This power [in the snail] of growing afresh so complex and specialized an organ as an eye is certainly, at first sight, not a little astonishing, but it appears to be capable of a very simple explanation. The cells terminating the cut stump of the tentacle are the ancestors of those which are removed; a fresh series of descendants are derived from them, similarly related to the ancestral cells as their predecessors which they replace; the first generation of descendants become in turn ancestors to a second generation, similarly related to them as were the second tier of extirpated cells; and this process of descent being repeated, the completed organ will at length be rebuilt." This explanation is, however, misleading in its simplicity. The cells terminating the cut stump are not the direct ancestors of those which are removed, except in the same sense as gorillas are ancestors of men. They are rather collateral descendants of common ancestors. I think that Professor Sollas would probably agree that, though the lens and "retina" are of epiblastic (outer layer) origin, their relationship with the epiblastic cells at the cut stump is a somewhat distant one. In the reproduction of the lens the cell-heredity is not direct, but markedly indirect. And it is somewhat difficult to understand by what means the ordinary epiblastic cells of the cut stump, which have had no part in the special and peculiar work of lens-production, should be enabled to produce cell-offspring, some of which, and those in a special relation to other deeper-lying cells, possess this peculiar power.

On the other hand, if we turn to the view that the reproduction is effected, not by the cells of the cut surface alone, but by the general mass of cells in the mutilated organism, we have to face the difficulty of understanding how the influence of cells other than those partaking in the regrowth can be brought to bear on these so as to direct their lines of development. If we say that the organism is pervaded by a principle of symmetry such that both growth and regrowth, whenever they take place, are constrained to follow the lines of ancestral symmetry, we are really doing little more than restating the facts without affording any real organic explanation. That which we want to know is in what organic way this symmetrical growth is effected—how the hereditary tendency is transmitted through the nuclear network which is concerned in cell-division. I do not think that we are at present in a position to give a satisfactory answer to this question.

Let us now return to the hydra, the artificial fission of which has suggested these considerations. Multiplication in this way is probably abnormal. Under suitable conditions, however, if well fed, the hydra normally multiplies by budding. At some spot, generally not far from the "foot," or base of attachment, a little swelling occurs, and the growth of the cells in this region takes such lines that a new hydra is formed. This is at first in direct connection with the parent stem, the two having a common internal cavity; but eventually it separates and lives a free existence as a distinct organism (see Fig. 9, p. 45).

Now, here we may notice, as an implication from these facts, that the size of the organism is limited. When the normal limits of size are reached, any further assimilation of nutriment ministers, not to the further growth of the organism, but to the formation of a new outgrowth, or bud. What determines that the outgrowth, or bud, should originate in this or that group of cells, we do not know. But, like the isolated fragment in the hydra subdivided by fission, the little group in which budding commences contains a fair sample of the various kinds of cells which constitute the hydra. And here, too, we see that their growth and development follow definite lines of hereditary symmetry.

But there is a third method of multiplication in hydra: this is the sexual mode of reproduction, and occurs generally in the autumn. On the body-wall of certain individuals, near the tentacles, conical swellings appear. Within these swellings are great numbers of minute sperms, with small oval heads and active, thread-like tails. They appear to originate from the interstitial cells of the outer layer (see p. 124). Nearer the foot, or base of attachment, and generally, but not quite always, in separate individuals, there are other larger swellings, different in appearance, of which there is generally only one in the same individual at the same time. Each contains a single ovum, or egg-cell, surrounded by a capsule. It, too, and the cells which surround it would appear to be developed from the interstitial cells. It grows rapidly at the expense of the surrounding tissue, but when mature, it bursts through the enveloping capsule, and is freely exposed. A sperm-cell, which seems, in some cases at least, to be produced by the same individual, now unites with it; the egg-cell then begins to undergo division, becomes detached, falls to the bottom, and develops into a young hydra.

Here, then, we have that sexual mode of reproduction which occurs in all the higher animals. It is, however, in some respects peculiar in hydra. In the first place, the ovum is nearly always in other animals (but occasionally not in hydra) fertilized by the sperm from a separate and distinct individual. In the second place, the germinal cells are generally produced, not from the outer layer, but from the middle layer, which appears between the two primitive layers. In some allies of hydra, however, they take their origin in the inner layer; and it has been suggested that, even in hydra, the true germinal cells may migrate from the inner to the outer layer. But of this there does not seem to be at present sufficient evidence. In any case, however, the essential fact to bear in mind is that a new individual is produced by the union of a single cell produced by one organism and of another cell produced in most cases (but not always in the hydra) from a different individual. In the higher forms of animal life, the organisms are either female (egg-producing) or male (sperm-producing). But there are many hermaphrodite forms which produce both eggs and sperms, as in the common snail and earthworm. Even in these cases, however, there are generally special arrangements by which it is ensured that the sperm from one individual should fertilize the ovum produced by another individual.

What, we must next inquire, is the relation in the higher forms of life—for we may now leave the special consideration of hydra—of the ovum or sperm to the organism which produces it? This is but one mode of putting a very old question—Does the hen produce the egg, or does the egg produce the hen? Of course, in a sense, both are true; for the hen produces an egg which, if duly fertilized, will develop into a new hen. But the question has of late been asked in a new sense; and many eminent naturalists reply, without hesitation—The egg produces the hen, but under no circumstances does the hen produce the egg. What, then, it may be asked, does produce the egg? To this it is replied—The egg was produced by a previous egg. At first sight, this may seem a mere quibble; for it may be said that, of course, if an egg produces a hen which contains other eggs, these eggs may be said to be produced by the first. But it is really more than a quibble, and we must do our best clearly to grasp what is meant.

We have seen that, in development, the fertilized egg-cell undergoes division into two cells, each of which again divides into two, and so on, again and again, until from one there arises a multitude of cells. Nor is this all. The multitude are organized into a whole. The constituent cells have different forms and structures, and perform diverse functions. Some are skeletal, such as bone and connective tissue; some are protective, such as those which give rise to feathers or scales; some form nerves or nerve-centres; some, muscles; some give rise to glandular tissue; and lastly, some form the essential elements in reproduction. If, now, we express the development of tissues and the sequence of organisms in the following scheme, the continuity of the reproductive cells will be apparent:—

It is clear that there is a continuity of reproductive cells, which does not obtain with regard to nerve, gland, or skeleton. If, then, we class together as body-cells those tissue-elements which constitute what we ordinarily call the body, i.e. the head, trunk, limbs—all, in fact, except the reproductive cells, our scheme becomes—

From this, again, it is clear that the body does not produce the egg, or reproductive cell, but that the reproductive cell does produce the body. Of course, it should be noted that we are here using the term "body" as distinguished from, and not as including, the reproductive cells. But this is convenient, in that it emphasizes the fact that the muscular, nervous, skeletal, and glandular cells take (on this view) no part whatever in producing those reproductive cells which are concerned in the continuance of the species.

Such, in brief, is the view that the egg produces the hen. We will return to it presently when we have glanced at the alternative view that the hen produces the egg.

On this view, the reproductive elements are not merely cells, the result of normal cell-division, which have been set aside for the continuance of the species. They are, so to speak, the concentrated extract of the body, and contain minute or infinitesimal elements derived from all the different tissues of the organism which produces them. Darwin[BG] suggested that all the cells of the various tissues produce minute particles called gemmules, which circulate freely throughout the body, but eventually find a home in the reproductive cells. Just as the organism produces an ovum from which an organism like itself develops, so do the cells of the organism produce gemmules, which find their way to the ovum and become the germs of similar cells in the developing embryo. "The child, strictly speaking," says Darwin, "does not grow into a man, but includes germs which slowly and successively become developed and form the man." "Each animal may be compared with a bed of soil full of seeds, some of which soon germinate, some lie dormant for a period, whilst others perish." Or, to vary the analogy, "an organic being is a microcosm—a little universe formed of a host of self-propagating organisms." This is Darwin's provisional hypothesis of pangenesis, which has recently been accepted in a modified form by Professor W. K. Brooks in America, to some extent by De Vries on the Continent, by Professor Herdman of Liverpool, and by other biologists. The ovum on this view is to be regarded as a composite germ containing the germs of the cellular constituents of the future organism. The scheme representing this view will stand thus—

It is clear that, on this hypothesis, we may frame an apparently simple and, on first sight, satisfactory theory of heredity. Since all the body-cells produce gemmules, which collect in or give rise to the reproductive cells, and since each gemmule is the germ of a similar cell, what can be more natural than that the ovum, thus composed of representative cell-germs, should develop into an organism resembling the parent organism? Modifications of structure acquired during the life of the organism would thus be transmitted from parent to offspring; for the modified cells of the parent would give rise to modified gemmules, which would thus hand on the modification. The inheritance of ancestral traits from grandparent or great-grandparent might be accounted for by supposing that some of the gemmules remained latent to develop in the second or third generation. The regeneration of lost parts receives also a ready explanation. If a part be removed by amputation, regrowth is possible because there are disseminated throughout the body gemmules derived from each part and from every organ. A stock of nascent cells or of partially developed gemmules may even be retained for this special purpose, either locally or throughout the body, ready to combine with the gemmules derived from the cells which come next in due succession. Similarly, in budding, the buds may contain nascent cells or gemmules in a somewhat advanced stage of development, thus obviating the necessity of going through all the early stages in the genesis of tissues. The gemmules derived from each part being, moreover, thoroughly dispersed through the system, a little fragment of such an organism as hydra may contain sufficient to rebuild the complete organism; or, if it contains an insufficient number, we may assume that the gemmules, in their undeveloped state, are capable of multiplying indefinitely by self-division. Finally, variations might arise from the superabundance of certain gemmules and the deficiency of others, and from the varying potency of the gemmules contained in the sperm and ovum. Where the maternal and paternal gemmules are of equal potency, the cell resulting from their union will be a true mean between them; where one or other is prepotent, the resulting cell will tend in a corresponding direction. And since the parental cells are subject to modification, transmitted through the gemmules to the reproductive elements, it is clear that there is abundant room and opportunity for varietal combinations.

It is claimed, as one of the chief advantages of some form of pangenetic hypothesis, that it, and it alone, enables us to explain the inheritance of characters or modifications of structure acquired by use (or lost by disuse) during the life of the organism, or imprinted on the tissues by environmental stresses. The evidence for the transmission of such acquired characters we shall have to consider hereafter. We may here notice, however, that at first sight the hypothesis seems to prove too little or too much. For while modifications of tissues, the effects of use and disuse, are said to be inherited, the total removal of tissues by amputation, even if repeated generation after generation, as in the docking of the tails of dogs and horses, formerly so common, does not have the effect of producing offspring similarly modified. Professor Weismann has recently amputated the tails of white mice so soon as they were born, for a number of generations, but there is no curtailment of this organ in the mice born of parents who had not only themselves suffered in this way, but whose parents, grandparents, and great-grandparents were all rendered tailless. The pangenetic answer to this objection is that gemmules multiply and are transmitted during long series of generations. We have only to suppose that the gemmules of distantly ancestral tails have been passing through the mutilated mice in a dormant condition, awaiting an opportunity to develop, and the constant reappearance of tails is seen to be no real anomaly. In this case the gemmules of the parental and grandparental tail are simply absent. But if the muscles of the parental tail were modified through unwonted use, the modified cells would give rise to modified gemmules, which would unite in generation with ancestral gemmules, and to a greater or less degree modify them. The difference is between the mere absence of gemmules and the presence of modified gemmules. And the fact that it takes some generations for the effects of use or disuse to become marked is (pangenetically) due to the fact that it takes some time for the modified gemmules to accumulate and be transmitted in sufficient numbers to affect seriously the numerous ancestral gemmules.

The direction in which Professor W. K. Brooks has recently sought to modify Darwin's pangenetic hypothesis may here be briefly indicated. He holds that it is under unwonted and abnormal conditions that the cells are stimulated to produce gemmules, and that the sperm is the special centre of their accumulation. Hence it is the paternal influence which makes for variation, the maternal tendency being conservative. The reproductive cell is not merely or chiefly a microcosm of gemmules. It is a cell produced by ordinary cell-division from other reproductive cells. The ovum remains comparatively unaffected by changes in the body; but it receives from the sperm, with which it unites, gemmules from such tissues in the male as were undergoing special modification. The hen does not produce the egg; but the cock does produce the sperm; and the union of the two hits the happy mean between the conservatism of the one view and the progressive possibilities of the other.

Mr. Francis Galton, in 1876,[BH] suggested a modification of Darwin's hypothesis, which included, as does that of Professor Brooks, the idea of germinal continuity which had been suggested by Professor (now Sir Richard) Owen, in 1849. He calls the collection of gemmules in the fertilized ovum the "stirp." Of the gemmules which constitute the stirp only a certain number, and they the most dominant, develop into the body-cells of the embryo. The rest are retained unaltered to form the germinal cells and keep up a continuous tradition. Mr. Galton's place in the history of theories of heredity can scarcely be placed too high. Only one further modification of pangenesis can here be mentioned, namely, that proposed in 1883 by Professor Herdman, of Liverpool. He suggested "that the body of the individual is formed, not by the development of gemmules alone and independently into cells, but by the gemmules in the cells causing, by their affinities and repulsions, these cells so to divide as to give rise to new cells, tissues, and organs."

Such are Darwin's provisional hypothesis of pangenesis, and some more recent modifications thereof. Bold and ingenious as was Darwin's speculation, supported as it at first sight seems to be by organic analogies, it finds to-day but few adherents. With all our increased modern microscopical appliances, no one has ever seen the production of gemmules. Although it appears sufficiently logical to say that, just as a large organism produces a small ovum, so does each small cell produce an exceedingly minute gemmule; when closely investigated, the analogy is not altogether satisfactory. It is denied, as we have seen, by many biologists that the organism does produce the ovum. Multiplication is normally by definite, visible cell-division. Nuclear fission can be followed in all its phases. But where is the nuclear fission in the formation of gemmules? It is true that the conjugation of monads is followed by the pouring forth of a fluid which must be crowded with germs from which new monads arise, and that these germs are so minute as to be invisible, even under high powers of the microscope. It might be suggested, then, that in every tissue some typical cell or cells might thus break up into a multitude of invisible gemmules. But there is at present no evidence that they do so. And even if this were the case, it would not bear out Darwin's view, that every cell is constantly throwing off numerous gemmules. It is known, however, or at least generally believed, that there is a constant replacement of tissues during the life of the organism. It is said, for example, that in the course of seven years the whole cellular substance of the human body is entirely renewed. The fact is, I think, open to question. Granting it, however, it might be suggested that the effete cells, ere they vanish, give rise to minute gemmules, which find their way to the ova. But it must be remembered that the new tissue-cells in the supposed successional renewal of the organs are the descendants of the old tissue-cells; that these are, therefore, already reproducing their kind directly; and that the formation of gemmules would thus be a special superadded provision for a future generation. Still, there is no reason why cells should not have this double mode of reproduction, if any definite evidence of its existence could be brought forward. Without such definite evidence, we may well hesitate before we accept it even provisionally.

The existence of gemmules, then, is unproven, and their supposed mode of origin not in altogether satisfactory accordance with organic analogies. Furthermore, the whole machinery of the scheme of heredity is complicated and hyper-hypothetical. It is difficult to read Darwin's account of reversion, the inheritance of functionally acquired characters, and the non-inheritance of mutilation, or to follow his skilful manipulation of the invisible army of gemmules, without being tempted to exclaim—What cannot be explained, if this be explanation? and to ask whether an honest confession of ignorance, of which we are all so terribly afraid, be not, after all, a more satisfactory position.

That the hen produces the egg, that "gemmules are collected from all parts of the system to constitute the sexual elements, and that their development in the next generation forms a new being," is further rendered improbable by direct observation upon the mode of origin of the germinal cells, ova, or sperms.

It will be remembered that the view that the egg produces the hen, while the hen does not produce the egg, suggested the question—What, then, does produce the egg? to which the answer was—The egg is the product of a previous egg. On this view, then, the germinal cells, ova, or sperms are the direct and unmodified descendants of an ovum and sperm which have entered into fertile union. Now, in certain cases, notably in the fly Chironomus, studied by Professor Balbiani, but also in a less degree in some other invertebrate forms, it is possible to trace the continuity of the germinal cells with the fertilized ovum from which they are derived. In Chironomus, for example, "at a very early stage in the embryo, the future reproductive cells are distinguishable and separable from the body-forming cells. The latter develop in manifold variety, into skin and nerve, muscle and blood, gut and gland; they differentiate, and lose almost all protoplasmic likeness to the mother ovum. But the reproductive cells are set apart; they take no share in the differentiation, but remain virtually unchanged, and continue unaltered the protoplasmic tradition of the original ovum."[BI] In such a case, then, observation flatly negatives the view that the germinal cells are "constituted" by gemmules collected from the body-cells, though, of course (on a modified pangenetic hypothesis), they might be the recipients of such gemmules.

It is only in a minority of cases, however, that the direct continuity of germinal cells as such is actually demonstrable. In the higher vertebrates, for instance, the future reproductive cells can first be recognized only after differentiation of some of the body-cells and the tissues they constitute is relatively advanced. While in cases of alternation of generations, "an entire asexual generation, or more than one, may intervene between one ovum and another." In all such cases the continuity of the chain of recognizably germinal cells cannot be actually demonstrated.

The impracticability of actually demonstrating a continuity of germinal cells in the majority of cases has induced Professor Weismann to abandon the view that there is a continuity of germinal cells, and to substitute for it the view that there is a continuity of germ-plasm (keimplasma). "A continuity of germ-cells," he says,[BJ] "does not now take place, except in very rare instances; but this fact does not prevent us from adopting a theory of the continuity of the germ-plasm, in favour of which much weighty evidence can be brought forward." It might, however, be suggested that, although a continuity of germ-cells cannot be demonstrated, such continuity may, nevertheless, obtain, the future germinal cells remaining undifferentiated, while the cells around them are undergoing differentiation. The comparatively slight differentiation of the body-cells in hydroids renders such a view by no means improbable. But Professor Weismann does not regard such an idea as admissible, at all events, in certain cases. "It is quite impossible," he says,[BK] "to maintain that the germ-cells of hydroids, or of the higher plants, exist from the time of embryonic development, as undifferentiated cells, which cannot be distinguished from others, and which are only differentiated at a later period." The number of daughter-cells in a colony of hydroid zoophytes is so great that "all the cells of the embryo must for a long time act as body-cells, and nothing else." Moreover, actual observation (e.g. in Coryne) convinces Dr. Weismann that ordinary body-cells are converted into reproductive cells. After describing the parts of the body-wall in which a sexual bud arises as in no way different from surrounding areas, he says, "Rapid growth, then, takes place at a single spot, and some of the young cells thus produced are transformed into germ-cells which did not previously exist as separate cells."[BL]

This transformation of body-cells or their daughter-cells into germ-cells seems therefore, if it be admitted, to negative the continuity of germ-cells as such. But this fact, says Weismann, does not prevent us from adopting a theory of the continuity of germ-plasm. "As a result of my investigations on hydroids," he says,[BM] "I concluded that the germ-plasm is present in a very finely divided and therefore invisible state in certain body-cells, from the very beginning of embryonic development, and that it is then transmitted, through innumerable cell-generations, to those remote individuals of the colony in which the sexual products are formed."

This germ-plasm resides in the nucleus of the cell; and it would seem that by a little skilful manipulation it can be made to account for anything that has ever been observed or is ever likely to be observed. It is one of those convenient invisibles that will do anything you desire. The regrowth of a limb shows that the cells contained some of the original germ-plasm. A little sampled fragment of hydra has it in abundance. It lurks in the body-wall of the budding polyp. It is ever ready at call. It conveniently accounts for atavism, or reversion; for[BN] "the germ-plasm of very remote ancestors can occasionally make itself felt. Even a very minute trace of a specific germ-plasm possesses the definite tendency to build up a certain organism, and will develop this tendency as soon as the nutrition is, for some reason, favoured above that of the other kinds of germ-plasm present in the nucleus."

In place, then, of the direct continuity of germ-cells as distinct from body-cells, we have here the direct continuity of germ-plasm as opposed to body-plasm. The germ-plasm can give rise to body-plasm to any extent; the body-plasm can never give rise to germ-plasm. If it seems to do so, this is only because the nuclei of the body-cells contain some germ-plasm in an invisible form. The body-plasm dies; but the life of the germ-plasm is, under appropriate conditions, indefinitely continuous.

So far as heredity is concerned, it matters not whether there be a continuity of germ-cells or of germ-plasma. In either case, the essential feature is that body-cells as such cannot give rise to the germ—that the hen cannot produce the egg. On either view, characters acquired by the body cannot be transmitted to the offspring through the ova or sperms. The annexed diagram illustrates how, on the view that the hen produces the egg, dints hammered into the body by the environment will be handed on; while, on the view that the hen does not produce the egg, the dints of the environment are not transmitted to the offspring. On the hypothesis of continuity, heredity is due to the fact that two similar things under similar conditions will give similar products. The ovum from which the mother is developed, and the ovum from which the daughter is developed, are simply two fragments separated at different times from the same continuous germ-plasm.[BO] Both develop under similar circumstances, and their products cannot, therefore, fail to be similar. How variation is possible under these conditions we shall have to consider presently.

Now, although I value highly Professor Weismann's luminous researches, and read with interest his ingenious speculations, I cannot but regard his doctrine of the continuity of germ-plasm as a distinctly retrograde step. His germ-plasm is an unknowable, invisible, hypothetical entity. Material though it be, it is of no more practical value than a mysterious and mythical germinal principle. By a little skilful manipulation, it may be made to account for anything and everything. The fundamental assumption that whereas germ-plasm can give rise to body-plasm to any extent, body-plasm can under no circumstances give rise to germ-plasm, introduces an unnecessary mystery. Biological science should set its face against such mysteries. The fiction of two protoplasms, distinct and yet commingled, is, in my opinion, little calculated to advance our knowledge and comprehension of organic processes. For myself, I prefer to take my stand on protoplasmic unity and cellular continuity.

The hypothesis of cellular continuity is one that the researches of embryologists tend more and more to justify. The fertilized ovum divides and subdivides, and, by a continuance of such processes of subdivision, gives rise to all the cells of which the adult organism is composed. It is true that in some cases, as in that of peripatus, as interpreted by Mr. Adam Sedgwick, the cells of the embryo run together or remain continuous as a diffused protoplasmic mass with several or many nuclei. But this seemingly occurs only in early stages as a step towards the separation of distinct cells. And even if the process should be proved of far wider occurrence, it would not disprove the essential doctrine of cellular continuity. The nucleus is the essence of the cell. And the doctrine of cellular continuity emphasizes the fact that the nuclei of all the cells of the body are derived by a process of divisional growth from the first segmentation-nucleus which results from the union of the nuclei of the ovum and the sperm. In this sense, then, however late the germinal cells appear as such, they are in direct continuity with the germinal cell from which they, in common with all the cells of the organism, derive their origin. In this sense there is a true continuity of germ-cells.

Now, it has again and again been pointed out that the simple cell of which an amœba is composed is able to perform, in simple fashion, the various protoplasmic functions. It absorbs and assimilates food; it is contractile and responds to stimulation; it respires and exhibits metabolic processes; it undergoes fission and is reproductive. The metazoa are cell-aggregates; and in them the cells exemplify a physiological division of labour. They differentiate, and give rise to muscle and nerve, gut and gland, blood and connective or skeletal tissue, ova and sperms. Are these germinal cells mysteriously different from all the other cells which have undergone differentiation? No. They are the cells which have been differentiated and set apart for the special work of reproduction, as others have been differentiated and set apart for other protoplasmic functions.

Cell-reproduction is, however, in the metazoa of two kinds. There is the direct reproduction of differentiated cells, by which muscle-cells, nerve-cells, or others reproduce their kind in the growth of tissues or organs; and there is the developmental reproduction, by which the germinal cells under appropriate conditions reproduce an organism similar to the parent. The former is in the direct line of descent from the simple reproduction of amœba. The latter is something peculiarly metazoan, and is, if one may be allowed the expression, specialized in its generality.

That the metazoa are derived from the protozoa is generally believed. How they were developed is to a large extent a matter of speculation. But, however originating, their evolution involved the production, from cells of one kind, of cells of two or more kinds, co-operating in the same organism. Whenever and however this occurred, the new phase of developmental reproduction must have had its origin. And if in cell-division there is any continuity of protoplasmic power, the faculty of producing diverse co-operating cells would be transmitted. On any view of the origin of the metazoa, this diverse or developmental reproduction is a new protoplasmic faculty; on any view, it must have been transmitted, for otherwise the metazoa would have ceased to exist. This new faculty of developmental reproduction, then, with the inception of the metazoa, takes its place among other protoplasmic faculties, and, with the progress of differentiation and the division of labour, will become the special business of certain cells. On this view, the specialization of the reproductive faculty and of germinal cells takes its place in line with other cell-differentiations with division of labour; and the difficulties of comprehending and following the process of differentiation in this matter are similar to those which attend physiological division of labour in general.

It is probable that, in the lower metazoa, in which differentiation has not become excessively stereotyped, the power of developmental reproduction is retained by a great number of cells, even while it is being specialized in certain cells. Hence the ability to produce lost parts and the reproduction of hydra by fission. But, on the other hand, the special differentiation of a tissue on particular lines has always a tendency to disqualify the cells from performing other protoplasmic faculties, and that of developmental reproduction among the number. I do not know of any definite, well-observed cases on record in the animal kingdom of ova or sperms being derived from cells which are highly differentiated in any other respect. In the vertebrata, the mesoblastic, or mid-layer, cells, from which the germinal epithelium arises, have certainly not been previously differentiated in any other line. And in the case of the hydroid zoophytes, quoted by Professor Weismann, the cells which give rise to the germinal products have never been so highly differentiated as to lose the protoplasmic faculty of developmental reproduction.

Some such view of developmental reproduction, based upon cellular continuity and the division of labour, seems to me more in accord with the general teachings of modern biology than a hypothetical and arbitrary distinction between a supposed germ-plasm and a supposed body-plasm.

To which category, then, does this hypothesis belong? Does it support the view that the hen produces the egg or that the egg produces the hen? Undoubtedly the latter. It is based on cellular continuity, and is summarized by the scheme on p. 131. It adequately accounts for hereditary continuity, for there is a continuity of the germinal cells, the bearers of heredity. But how, it may be asked, on this view, or on any continuity hypothesis, are the origin of variations and their transmission to be accounted for? To this question we have next to turn. But before doing so, it will be well to recapitulate and summarize the positions we have so far considered.

We saw at the outset that the facts we have to account for are those of heredity with variation. To lead up to the facts of sexual heredity, we considered fission, the regeneration of lost parts, and budding in the lower animals. We saw that, if a hydra be divided, each portion reproduces appropriately the absent parts. But we found it difficult to say whether this power resides, in such cases, in the cells along the plane of section or in the general mass of cells which constitute the regenerating portion.

Having led up to the sexual mode of reproduction, we inquired whether the egg produces the hen or the hen produces the egg. We saw that there is a marked difference between a direct continuity of reproductive cells, giving rise to body-cells as by-products, and an indirect continuity of reproductive cells, these cells giving rise to the hen, and then the hen to fresh reproductive cells, which, on this view, are to be regarded as concentrated essence of hen.

Darwin's hypothesis of pangenesis as exemplifying the latter view was considered at some length, and the modifications suggested by Professor Brooks, Mr. Galton, and Professor Herdman were indicated. The hypothesis, so far as it is regarded as a theory of the main facts of heredity, was rejected.

It was then pointed out that only in a few cases has a direct continuity of germinal cells as such been actually demonstrated. Whence Professor Weismann has been led to elaborate his doctrine of the continuity of germ-plasm. This germ-plasm can give rise to, but cannot originate from, body-plasm. It may lurk in body-cells, which may, by its subsequent development, be transformed into germ-cells. But any external influences which may affect these body-cells produce no change on the germ-plasm which they may contain. We regarded this hypothesis as a retrograde step, much as we admire the genius of its propounder, and considered that the fiction of two protoplasms, distinct and yet commingled, is little calculated to advance our comprehension of organic processes.

In the known and observed phenomena of cellular continuity and cell-differentiation, we found a sufficiently satisfactory hypothesis of heredity. The reproductive cells are the outcome of normal cell-division, and have been differentiated and set apart for the special work of developmental reproduction, as others have been differentiated and set apart for other protoplasmic functions. Such a view adequately accounts for hereditary continuity, for there is a continuity of the germinal cells, the bearers of heredity. But how, we repeat, on this view or any other hypothesis of direct continuity, are the origin of variations and their transmission to be accounted for?

Every individual organism reacts more or less markedly under the stress of environing conditions. The reaction may take the form of passive resistance, or it may be exemplified in the performance of specially directed motor-activities. The power to react in these ways is inborn; but the degree to which the power is exercised depends upon the conditions of existence, and during the life of the individual the power may be increased or diminished according to whether the conditions of life have led to its exercise or not. The effects of training and exercise on the performance of muscular feats and in the employment of mental faculties are too well known to need special exemplification. By manual labour the skin of the hand is thickened; and by long-continued handling of a rifle a bony growth caused by the weapon in drilling, the so-called exercierknochen of the Germans, is developed. Now, it is clear that if these acquired structures or faculties are transmitted from parent to offspring, we have here a most important source and origin of variations—a source from which spring variations just in the particular direction in which they are wanted. The question is—Are they transmitted? and if so, how?

Let us begin with the protozoa. Dr. Dallinger made some interesting experiments on monads. They extended over seven years, and were directed towards ascertaining whether these minute organisms could be gradually acclimatized to a temperature higher than that which is normal to them. Commencing at 60° Fahr., the first four months were occupied in raising the temperature 10° without altering the life-history. When the temperature of 73° was reached, an adverse influence appeared to be exerted on the vitality and productiveness of the organism. The temperature being left constant for two months, they regained their full vigour, and by gradual stages of increase 78° was reached in five months more. Again, a long pause was necessary, and during the period of adaptation a marked development of vacuoles, or internal watery spaces, was noticed, on the disappearance of which it was possible to raise the temperature higher. Thus by a series of advances, with periods of rest between, a temperature of 158° Fahr. was reached. It was estimated that the research extended over half a million generations. Here, then, these monads became gradually acclimatized to a temperature more than double that to which their ancestors had been accustomed to—a temperature which brought rapid death to their unmodified relatives.

Now, in such observations it is impossible to exclude elimination. It is probable that there were numbers of monads which were unable to accommodate themselves to the changed conditions, and were therefore eliminated. But in any case, the fact remains that the survivors had, in half a million generations, acquired a power of existing at a temperature to which no individual in its single life could become acclimatized. Here, then, we have the hereditary transmission of a faculty. But the organisms experimented on were protozoa. In them there is no distinction between germ-cell and body-cell. Multiplication is by fission. And if the cell which undergoes fission has been modified, the two separate cell-organisms which result from that fission will retain the special modification. In such cases the transmission of acquired characters is readily comprehensible. We have an hereditary summation of effects.

With the metazoa the case is different. In the higher forms the germinal cells are internal and sheltered from environing influences by the protecting body-wall. It is the body-cells that react to environmental stresses; it is muscle and nerve in which faculty is strengthened by use and exercise, or allowed to dwindle through neglect. The germ-cells are shielded from external influences. They lead a sheltered and protected life within the body-cavity. It is no part of their business to take part in either passive resistance or responsive activity. During the individual life, then, the body may be modified, may acquire new tissue, may by exercise develop enhanced faculties. But can the body so modified affect the germ-cells which it carries within it?

Biologists are divided on this question. Some say that the body cannot affect the germ; others believe that it can and does do so.

It might seem an easy matter to settle one way or another. But, in truth, it is by no means so easy. Suppose that a man by strenuous exercise brings certain muscles to a high degree of strength or co-ordination. His son takes early to athletics, and perhaps excels his parent. Is this a case of transmitted fibre and faculty? It may be. But how came it that the father took to athletics, and was enabled to develop so lithe and powerful a frame? It must have been "in him," as we say. In other words, it must have been a product of the germ-cells from which he was developed. And since his son was developed, in part at least, from a germ-cell continuous with these, what more natural than that he too should have an inherent athletic habit? Every faculty that is developed in any individual is potential in the germ-stuff from which he springs; the tendency to develop any particular faculty is there too; and both faculty and tendency to exercise it are handed on by the continuity of germ-protoplasm or germ-cells. Logically, there is no escape from the argument if put as follows: The body and all its faculties (I use the term "faculties" in the broadest possible sense) are the product of the germ; the acquisition of new characters or the strengthening of old faculties by the body is therefore a germinal product; there is continuity of the germs of parent and child; hence the acquisition by the child of characters acquired by the parent is the result of germinal or cellular continuity. It is not the acquired character which influences the germ, but the germ which develops what appears to be an acquired character. Finally, if an acquired character, so called, is better developed in the child than in the parent, what is this but an example of variation? And if, in a series of generations, the acquired character continuously increases in strength, this must be due to the continued selection of favourable variations. It is clear that the organism that best uses its organs has, other things equal, the best chance of survival. It will therefore hand on to its offspring germinal matter with an inherent tendency to make vigorous use of its faculties.

Those who argue thus deny that the body-cells can in any way affect the germ-cells. To account for any continuous increase in faculty, they invoke variation and the selection of favourable varieties. What, then, we may now ask, is, on their view, the mode of origin of variations?

In sexual reproduction, with the union of ovum and sperm, we seem to have a fertile source of variation. The parents are not precisely alike, and their individual differences are, ex hypothesi, germinal products. In the union of ovum and sperm, therefore, we see the union of somewhat dissimilar germs. And in sexual reproduction we have a constantly varying series of experiments in germinal combinations, some of which, we may fairly suppose, will be successful in giving rise to new or favourable variations. This view, however, would seem to involve an hypothesis which may be true, but which, in any case, should be indicated. For it is clear that if new or favourable variations arise in this way, the germinal union cannot be a mere mixture, but an organic combination.

An analogy will serve to indicate the distinction implied in these phrases. It is well known that if oxygen and hydrogen be mixed together, at a temperature over 100°C., there will result a gaseous substance with characters intermediate between those of the two several gases which are thus commingled. But if they are made to combine, there will result a gas, water-vapour, with quite new properties and characters. In like manner, if, in sexual union, there is a mere mixture, a mere commingling of hereditary characters, it is quite impossible that new characters should result, or any intensification of existing characters be produced beyond the mean of those of ovum and sperm. If, for example, it be true, as breeders believe, that when an organ is strongly developed in both parents it is likely to be even more strongly developed in the offspring, and that weakly parts tend to become still weaker, this cannot be the result of germinal mixture. Let us suppose, for the sake of illustration, that a pair of organisms have each an available store of forty units of growth-force, and that these are distributed among five sets of organs, a to e, as in the first two columns. Then the offspring will show the organs as arranged in the third column.[BP]