Examination of Weismann’s Theory of Evolution (1891).

Having now considered germ-plasm as perpetually continuous, we have next to regard it as unalterably stable.

First, let it be noted that these two fundamental and distinctive postulates of the whole Weismannian system are so intimately connected as to be in large measure mutually dependent. For, on the one hand, if germ-plasm has not been perpetually continuous since the first origin of life, it cannot have been absolutely stable “since the first origin of sexual propagation”: every time that its hereditary characters are modified by its containing soma (whether or not representatively so), its stability has been so far upset. On the other hand, if germ-plasm has not been absolutely stable, it cannot have been perpetually continuous “since the first origin of life.” As often as its stability has been upset, its “molecular structure” has been modified by causes ab extra, as distinguished from mixtures of germ-plasms in sexual unions. Therefore, it can no longer have been continuous in the sense of having borne an ineffaceable record of all congenital variations, due to sexual unions, throughout the entire phylogeny of the Metaphyta and Metazoa. At most it can have been continuous only in the attenuated sense, that however much and however often its hereditary characters may have been modified by somatic changes on the one hand or by changes in the external conditions of life on the other, they can never have been thus modified representatively, as supposed by the theory of pangenesis.

From which it follows that, while examining in our last chapter Weismann’s doctrine of the perpetual continuity of germ-plasm, we have been indirectly examining also his companion doctrine of the unalterable stability of germ-plasm. Nevertheless, for the sake of doing justice to both these doctrines, I have thought it desirable to examine each on its own merits, without prejudice arising from our criticism of the other. To such a separate and independent examination of the doctrine of unalterable stability we will, therefore, now proceed.

As we have already and repeatedly seen, this doctrine of the unalterable or absolute stability of germ-plasm “since the first origin of sexual propagation” is a logically essential part of Weismann’s theory of evolution, or of his system of hypotheses considered as a whole. It is so because upon this doctrine depends his reference of individual variations in the Metazoa to an ultimate origin in the Protozoa, the significance of sexual reproduction in the theory of natural selection, &c., &c. Therefore this doctrine of the absolute stability of germ-plasm is enunciated by Weismann, not merely for the purpose of meeting any one class of facts, such as those of atavism, persistence of rudimentary organs, &c. The doctrine is enunciated for the purpose of constituting one of the foundation-stones of his general theory of evolution. We have now to consider how far the quality of this stone renders it trustworthy as a basis to build upon.

In the first place, we can scarcely fail to perceive that this doctrine of the absolute stability of germ-plasm is not only gratuitous, but intrinsically improbable. That the most complex material in nature should likewise be the most stable is opposed to all the analogies of nature, and therefore to all the probabilities of the case.

Again, the germ-plasm, as it originally occurred (and still exists) in unicellular organisms, is supposed to be exactly the same kind of material as now occurs in the germ-cells of multicellular organisms. Yet the very same theory which supposes so absolute a stability on the part of germ-plasm when located in germ-cells (or diffused through somatic-cells), likewise supposes so high a degree of variability on the part of germ-plasm when not thus located, as to represent that all individual variations which have ever taken place in the unicellular organisms—and all the innumerable species of such organisms which have arisen therefrom—have been due to the direct action of external conditions of life; or, in other words, to the instability of germ-plasm. The very same substance which at one time and in one place is supposed to be so absolutely unchangeable, at another time and in another place is supposed to be highly susceptible of change.

Lastly—and this is, perhaps, the most curious part of the whole matter—the place where germ-plasm is supposed to be unchangeable is not the place where it is most likely to be so, but the place where it is least likely. For germ-plasm as it occurs in the germ-cells of multicellular organisms must have a constitution greatly more complex even than that which it has in unicellular organisms—seeing that in the former case, and by hypothesis, it bears a living record of the whole phylogeny of the Metaphyta and Metazoa in all their innumerable branchings. And not only so, but when germ-plasm occurs in germ-cells it becomes exposed to much greater vicissitudes: its environment has become vastly more complex, as well as greatly more liable to change with the changing conditions of life of the many mutable species in which it resides, and on the individual somas of which it now depends for its nourishment. So that, altogether, we have here on merely a priori grounds about as strong a case against this doctrine of absolute stability as it is well conceivable that on merely a priori grounds a case can be.

Turning next to arguments a posteriori, let us begin by considering those which Weismann has adduced in support of the doctrine.

First, he alleges that there is a total absence of variability on the part of all organisms which have been produced parthenogenetically, or from unfertilized ova. We may look in vain, he says, for any individual differences on the part of any multicellular organisms, which have been brought into existence independently of the blending of germ-plasms in a previous act of sexual union. Now, unquestionably, if this statement could be corroborated by sufficiently extensive observation, the fact would become one of immense significance—so much so, indeed, that of itself it would go far to neutralize all antecedent objections, and to verify his theory as to sexual propagation being the sole cause of congenital variation. But seeing that the alleged fact stands so entirely out of analogy with the phenomena of bud-variation (which will be alluded to later on), it is highly improbable, even on antecedent grounds; while Professor Vines has refuted the statement on grounds of actual fact. Thus, speaking of the Basidiomycetes, he says—

These Fungi are not only entirely a-sexual, but it would appear that they have been evolved in a purely a-sexual manner from a-sexual ascomycetous or Æcidiomycetous ancestors. The Basidiomycetes, in fact, afford an example of a vast family of plants, of the most varied form and habit, including hundreds of genera and species, in which, so far as minute and long-continued investigation has shown, there is not, and probably never has been, any trace of a sexual process[25].

Here, then, we have actual proof of “hereditary individual variations” among a-sexually propagating organisms, sufficient in amount to have given origin, not merely to “individual differences,” but to innumerable species, and even genera. Consequently Weismann allows that the criticism abolishes this line of evidence in favour of the absolute stability of germ-plasm[26]. Consquently, also, we must now add, in whatever measure the alleged fact would have corroborated the theory had it been proved to be a fact, in that measure is the theory discredited by proof that it is not a fact. For, if the theory were sound, this particular fact would certainly have admitted of demonstration: therefore the proof that it is not a fact—but the reverse of a fact—amounts at the same time to a disproof of the theory[27].

The only other line of evidence to be adduced in favour of the absolute stability of germ-plasm is that which is furnished by the high antiquity of some specific types, by the facts of atavism, and by the persistency of vestigial organs. But this line of evidence is as futile as the other. Nobody has ever questioned that hereditary characters are persistently stable as long as they are persistently maintained by natural selection; and this, according to Weismann himself, must have been the case with all long-enduring species: these, therefore, fail to furnish any evidence of the inherent stability of germ-plasm, which is the only point in question.

Again, as regards the facts of atavism, nobody is disputing these facts. What we are disputing is whether the degree of inherent stability which they unquestionably prove can be rationally regarded as such that it may endure, not merely for such a comparatively small number of generations as these facts imply, but actually for any number of generations, or through the practically infinite series of generations that now intervene between the higher metazoa and their primeval parentage in the protozoa. Clearly, the ratio between these two things is such that no argument derived from the facts of atavism can be of any avail for the purposes of this Weismannian doctrine.

Lastly, as regards vestigial organs, the consideration that, surprisingly persistent as they unquestionably are, nevertheless they do eventually disappear, seems to prove that the power of heredity does in time become exhausted, even in cases most favourable to its continuance. That it should thus become finally exhausted is no more than Darwin’s theory of perishable gemmules, or Galton’s theory of a not absolutely stable stirp, would expect. But the fact is irreconcilable with Weismann’s theory of an absolutely stable germ-plasm.

Hence, we can only conclude that there is no evidence in favour of the hypothesis that germ-plasm has been unalterably stable “since the first origin of sexual propagation”; while the suggestion that it may have been so is on antecedent grounds improbable, and on inductive grounds untenable. It only remains to add that the degree of stability has been proved in not a few cases to be less than even the theory of gemmules might anticipate. Many facts in proof of this statement might be given, but it will here suffice to quote one, which I select because it has been dealt with by Professor Weismann himself.

Professor Hoffmann has published an abstract of a research, which consisted in subjecting plants with normal flowers to changed conditions of life through a series of generations. In course of time, certain well-marked variations appeared. Now, in some cases such directly-produced variations were transmitted by seed from the affected plants; and therefore Weismann acknowledges,—“I have no doubt that the results are, at any rate in part, due to the operation of heredity.” Hence, whether these results be due to the transmission of somatogenetic characters (“representative changes”), or to the direct action of changed conditions of life on the germ-plasm itself (“specialized changes”), it is equally certain that the hereditary characters of the plants were congenitally modified to a large extent, within (at most) a few generations. In other words, it is certain that, if there be such a material as germ-plasm, it has been proved in this case to have been highly unstable. Therefore, in dealing with these and other similar facts, Weismann himself can only save his postulate of continuity by surrendering for the time being his postulate of stability[28].

If to this it be replied that Hoffmann’s facts are exceptional—that GÄrtner, NÄgeli, De Candolle, Peter, Jordan, and others, did not find individual variations produced in plants by changed conditions of life to be inherited,—the reply would be irrelevant. It does not require to be proved that all variations produced by changed conditions of life are inherited. If only some—even though it be but an extremely small percentage—of such variations are proved to be inherited, the many millions of years that separate the germ-plasm of to-day from its supposed origin in the protozoa, must have furnished opportunities enough for the occurrence of such variations to have obliterated, and re-obliterated numberless times, any aboriginal differences in the germ-plasms of incipiently sexual organisms. Moreover, it is probable that when further experiments shall have been made in this direction, Hoffmann’s results will be found not so exceptional as they at present appear. Mr. Mivart, for example, has mentioned several instances[29]; while there are not a few facts of general knowledge—such as the modifications undergone by certain Crustacea as a direct result of increased salinity of the water in which they live—that will probably soon be proved to be facts of the same order. But here attention must be directed to another large body of facts, which are of high importance in the present connexion.

The phenomena of what is called bud-variation in plants are phenomena of not infrequent occurrence, and they consist in the sudden appearance of a peculiarity on the part of a shoot which develops from a single bud. When such a peculiarity arises, it admits of being propagated, not only by cuttings and by other buds from that shoot, but sometimes also by seeds which the flowers of the shoot subsequently produce—in which case all the laws of inheritance that apply to congenital variations are found to apply also to bud-variation. Or, as Darwin puts it, “there is not any particular in which new characters arising by bud-variation can be distinguished from those due to seminal variation”; and, therefore, any theory which deals with the latter is bound also to take cognizance of the former. Now, as far as I can find, there is only one paragraph in which Weismann alludes to bud-variation, and what he there says I do not find very easy to understand. Therefore I will quote the whole paragraph verbatim.

I have not hitherto considered budding in relation to my theories, but it is obvious that it is to be explained, from my point of view, by supposing that the germ-plasm which passes on into a budding individual consists not only of the unchanged germ-plasm of the first ontogenetic stage, but of this substance altered, so far as to correspond with the altered structure of the individual which arises from it—viz., the rootless shoot which springs from the stem or branches. The alteration must be very slight, and perhaps quite insignificant, for it is possible that the differences between the secondary shoots and the primary plant may depend chiefly on the changed conditions of development, which takes place beneath the earth in the latter case, and in the tissues of the plant in the former. Thus we may imagine that the idio-plasm [? of that particular bud], when it develops into a flowering shoot, produces at the same time the germ-cells which are found in the latter. We thus approach an understanding of Fritz MÜller’s observation; for if the whole shoot which produces the flower arises from the same idio-plasm which also forms its germ-cells, we can readily understand why the latter should contain the same hereditary tendencies which were previously expressed in the flower which produced them. The fact that variations may occur in a single shoot depends on the changes explained above, which occur in the idio-plasm during the course of its growth, as a result of the varying proportions in which the ancestral idio-plasms may be contained in it.[30]

The meaning here appears to be twofold. For there are only two ways of explaining the phenomena of bud-variation. Either they are due to the influence of external conditions acting on the particular bud in question, or else they are due to so-called “spontaneous” changes taking place within the bud itself. Possibly it may be both, but at least it must be either. Well, in the above passage, Weismann appears to assume that it is both. For at the beginning of the passage he speaks of the “germ-plasm of the first ontogenetic stage” becoming “altered so far as to correspond with the altered structure of the individual which arises therefrom,” and he goes on to say that the alteration “may depend chiefly on the changed conditions of development”—that is, as I understand, the influence of external conditions. But at the end of the paragraph he says that “the changes which occur in the idio-plasm during the course of its growth” in the sporting bud, are due to “the varying proportions in which the ancestral idio-plasms may be contained in it.” Thus, I take it, Weismann here entertains both explanations of the phenomena in question: he appears to regard these phenomena as partly due to peculiar admixtures of ancestral idio-plasms in the bud itself (or “spontaneous” variation), but partly also to an alteration of the germ-plasm by its changed condition of development (or variation caused by external conditions).

However, it is but of little consequence whether or not this is the meaning which Weismann intends to convey. For the point we are coming to is, that, whatever he intends to convey, “from the point of view” of the theory of germ-plasm, there is only one interpretation possible. It is not open to Weismann (as it was to Darwin, or even to Galton,) to entertain both the explanations, whether separately or in conjunction. For germ-plasm (unlike gemmules, or even stirp) must be held always and everywhere unalterably stable: else the whole superstructure of Weismann’s theory of evolution falls to the ground. We cannot consent to his retaining this theory on the one hand, and, on the other, explaining bud-variation by “germ-plasm of the first ontogenetic stage” becoming altered “chiefly by changed conditions of development.” Even if it were true that “the alteration must be very slight, if not quite insignificant,” there would here be a rift in the lute, which must finally stop any further harping on the subject of Evolution.

From the point of view of this theory, then, there is only one interpretation open,—viz., that a bud-variation is ultimately due to a peculiar admixture of germ-plasms in the seed from which the bud was ultimately derived. But the objections to entertaining this as even a logically possible explanation of the phenomena in all cases, is insuperable.

In the first place, such a variation, when it does arise, is usually a variation of an extremely pronounced character; therefore it is very far from supporting Weismann’s view, that the “alteration” of germ-plasm which is needed to produce it “must be very slight, and perhaps quite insignificant.” In most cases where it occurs bud-variation presents so extreme a departure from the normal type, that no other kind of variation can be fitly compared with it in this respect. In particular, the degree of variation is usually very much greater than that which customarily obtains in congenital variations of the ordinary kind; and, therefore, if these be supposed due to particular admixtures of germ-plasm in sexual propagation, much more must those admixtures which give rise to sporting buds be characterized by peculiarities of no “insignificant” order. And much more, therefore, ought they to assert themselves in sister-buds developed from the same individual seed (ovule), than we find to be the case with any sister-organisms which are developed from different individual seeds. Yet, in the second place, so far is this from being the case, that the most remarkable feature connected with bud-variation—next to the suddenness and extreme amount of the variation itself—is the usually isolated nature of its occurrence. There may be thousands of other buds on the same plant, and yet it is one bud alone that deviates so suddenly and so widely from its ancestral characters. Nay, more, a single bud-variation may—and usually does—occur in plants which are habitually propagated by cuttings and graftings; so that there may not only be thousands, but millions of buds all derived from one original seed, and all for many years remaining perfectly true to their parent type, with the single exception of the sporting bud, which, while it departs so widely from that type, is usually capable of transmitting its extraordinary characters indefinitely by a-sexual, and not infrequently also by sexual, methods. So that, altogether, it seems impossible to suppose that in millions and millions of sister-buds, which through years and years exhibit no variation, a highly peculiar admixture of germ-plasm (which was originally present in the parent-seed) should have been latent; that it should then suddenly become so patent in a single bud, after which it never occurs in any other bud, save in the progeny of the sporting one.

On the whole, then, while it thus seems impossible to attribute all cases of bud-variation to mixtures of germ-plasms in sexual propagation, the theory of germ-plasm is unable to entertain any other explanation, on pain of surrendering its postulate touching the unalterable stability of germ-plasm, on which the Weismannian theory of evolution is founded.

So much for Weismann’s evidence touching the extreme, or virtually everlasting, stability of germ-plasm. We have seen that this evidence is not merely of a very poor character per se, or on antecedent grounds; but that it is directly negatived as evidence by the a-sexual origin of species in the plants alluded to by Professor Vines; by certain facts which prove so high a degree of instability on the part of this hypothetical substance, that in some cases it admits of being very considerably modified in the course of only two or three generations by exposure to changed conditions of life; while in other cases it may “sport,” so as to produce “hereditary individual variations,” which are much more pronounced than any of those that ordinarily result from a blending of hereditary qualities in an act of sexual union.

It will be well to conclude our examination of Weismann’s system by stating exactly the effect produced on his theory of evolution by the foregoing disproof of its fundamental postulate—the absolute stability of germ-plasm.

Clearly, in the first place, if germ-plasm has not been absolutely stable “since the first origin of sexual propagation,” the hereditary characters of germ-plasm may have been modified any number of times, and in always accumulating degrees. It matters not whether the modifications have been due mainly to external or to internal causes. It is enough to have shown that modifications occur. For, it will be remembered, the doctrine of the absolute stability of germ-plasm is, that inasmuch as the “molecular” structure of germ-plasm cannot be affected either from without or from within, the only source of “hereditary individual variations” is to be found in admixtures of germ-plasms taking place in sexual fertilization. Slight “molecular” differences having been originally impressed upon different masses of germ-plasm when these were severally derived from their unicellular sources, so unalterable has been the stability of germ-plasm ever since, that these slight “molecular” differences have never been in any degree effaced; and although in sexual unions they have for untold ages been obliged to mix in ever-varying proportions, they still continue—and ever must continue—to assert themselves in each ontogeny. Therefore, as Weismann himself formulates this astonishing doctrine,—“The origin of hereditary individual variations cannot indeed be found in the higher organisms, the Metazoa and Metaphyta; but is to be sought for in the lowest—the unicellular organisms.” Or again,—“The formation of new species, which among the lower Protozoa could be achieved without amphigony, could only be attained by means of this process in the Metazoa and Metaphyta. It was only in this way that hereditary individual differences could arise and persist[31].”

Now this doctrine is the most distinctive, as it is the most original feature in Weismann’s system of theories. That it is of interest as an example of boldly carrying the premises of a theory to their logical termination, no one will deny. But as little can it be denied that the very stringency of this logical process brings the theory itself into collision with such facts as those which have now been stated, and which, as far as I can see, are destructive of the theory—or, at any rate, of all that side of the theory which depends on the doctrine of absolute stability.

Take, for instance, the sequent doctrine that natural selection is inoperative among the unicellular organisms. Here, indeed, we have another of those doctrines which are so improbable on merely antecedent grounds, that their presence might well be deemed a source of irremediable weakness to the whole theory of evolution of which they form integral, or logically essential, parts. For seeing that the rate of increase in most of the unicellular organisms is quite as high as—and in most cases very much higher than—the rate that obtains in any of the multicellular, it becomes on merely antecedent grounds incredible that the struggle for existence should here not lead to any survival of the fittest. When, for instance, we learn from Maupas that a single Stylonichia is potentially capable of yielding a billion descendants within a week, we should need some extraordinarily good evidence to make us believe that as regards this organism natural selection is inoperative. But the point at present is that, quite apart from all general and a priori considerations of this kind, Weismann’s doctrine that unicellular organisms cannot be influenced by natural selection must be abandoned. For this doctrine followed deductively from the premiss that in the multicellular organisms congenital variations can only be due to admixtures of germ-plasms in acts of sexual fertilization; so that, in the absence of such admixtures, there could be no material for natural selection to work upon. But now we have found that this premiss must be given up; and, therefore, the deduction with respect to the unicellular organisms falls to the ground. Although it is true that the unicellular organisms propagate by fission, and although we grant, for the sake of argument, that they never propagate by way of sexual unions—even so this can no longer be taken to argue that none of their innumerable species owe their origin to natural selection. And, although it is probably true that the sexual methods of propagation constitute one source of hereditary individual variation among the multicellular organisms, there is no vestige of any independent reason for supposing that this is the only source of such variation; while the sundry facts which have now been given amount to nothing short of a demonstration to the contrary[32].

Lastly, and as regards the multicellular organisms, it is evident that Weismann’s essay On the Significance of Sexual Reproduction in the Theory of Natural Selection must be cancelled. For, apart from the contradictory manner in which this matter has been stated (pp. 70, 93, notes), and apart also from the consideration that other and quite as probable reasons have been suggested for the origin of sexual reproduction, there is the fact that Weismann’s theory is no longer tenable after the above destruction of its logical postulate in the absolute stability of germ-plasm. For, in the absence of this postulate, there is no basis for the theory that admixtures of germ-plasms in sexual reproduction furnish the sole means whereby heritable variations can be supplied for the working of natural selection.

Summary.

The theory of germ-plasm is not only a theory of heredity: it is also, and more distinctively, a theory of evolution. As a theory of heredity it is grounded on its author’s fundamental postulate—the continuity of germ-plasm; and, further, on a fact well recognized by all other theories of heredity, which he expresses by the term stability of germ-plasm. But as a theory of evolution it requires two additional postulates for its support—viz., that germ-plasm has been perpetually continuous “since the first origin of life,” and absolutely stable “since the first origin of sexual reproduction.” It is clear that these two additional postulates are not needed for his theory of heredity, but only for his additional theory of evolution. There have been other theories of heredity, prior to this one, which, like it, have been founded on the postulate of “continuity” (in Weismann’s sense) of the substance of heredity; but it has not been needful for any of these theories to postulate further that this substance has been always thus isolated, or even that it is now invariably so. For even though the isolation be frequently invaded by influences of body-changes on the congenital characters of this substance, it does not follow that the body-changes must be transmitted to offspring exactly as they occurred in parents. They may produce in offspring what we have agreed to call “specialized” hereditary changes, even if they never produce “representative” hereditary changes,—i.e., the transmission of acquired characters. But it is essential to Weismann’s theory of evolution that body-changes should not exercise a modifying influence of any kind on the ancestral endowments of this substance; hence, for the purposes of this further theory he has to assume that germ-plasm presents, not only continuity, but continuity unbroken since the first origin of life.

Similarly as regards his postulate of the stability of germ-plasm as absolute. It is enough for all the requirements of his theory of heredity, that the substance in question should present the high degree of stability which the facts of atavism, persistence of vestigial organs, &c., prove it to possess. But for his further theory of evolution it is necessary to make this further postulate of the stability of germ-plasm as undisturbed since the first origin of sexual propagation: otherwise there would be no logical foundation for any of the distinctive doctrines which go to constitute that theory.

Thus much understood, we proceeded to examine the theory of germ-plasm in each of its departments separately—i.e., first as a theory of heredity, and next as a theory of evolution. And we begun by comparing it as a theory of heredity with the preceding theories of Darwin and Galton. In the result we found that germ-plasm resembles gemmules in all the following respects. It is particulate; constitutes the material basis of heredity; is mainly lodged in highly specialized cells; is nevertheless also distributed throughout the general cellular tissues, where it is concerned in all processes of regeneration, repair, and a-sexual reproduction; presents an enormously complex structure, in that every constituent part of a potentially future organism is represented in a fertilized ovum by corresponding particles; is everywhere capable of virtually unlimited multiplication, without ever losing its hereditary endowments; is often capable of carrying these endowments in a dormant state through a long series of generations, until at last they reappear again in what we recognize as reversions. Such being the points of resemblance, the only points of difference may be summed up in the two words—continuity, and stability. For, as regards continuity, while Darwin’s theory supposes the substance of heredity to be more or less formed anew in each generation by the body-tissues of that generation, Weismann’s theory regards this substance as owing nothing to the body-tissues, further than lodgement and nutrition. Therefore, while the theory of gemmules can freely entertain the doctrines of Lamarck, the theory of germ-plasm excludes them as physiologically impossible, in all cases where sexual reproduction is concerned. Again, as regards stability, while Darwin’s theory simply accepts the fact of such a degree of stability appertaining to the substance of heredity as the phenomena of atavism, &c. prove, Weismann’s theory postulates the stability of this substance as absolute. But, as we have now so often seen, he does so in order to provide a hypothetical basis for his further theory of evolution. In as far as his theory of heredity is concerned, there is no reason why it should differ from Darwin’s in this respect.

Again, comparing Weismann’s theory of heredity with that of Galton, we found that germ-plasm resembles stirp in all the points wherein we have just seen that it resembles germ-plasm. Or, otherwise stated, all three theories are thus far coincident. But germ-plasm resembles stirp much more closely than it does gemmules, seeing that the theory of stirp is founded on the postulate of “continuity” in exactly the same manner as is the theory of germ-plasm. In point of fact, the only difference between these two theories consists in the two further postulates presented by the latter—viz., that the “continuity” in question has been unbroken since the origin of life, while the “stability” in question has been uninterrupted since the origin of sexual propagation. But seeing that both these additional postulates have reference to Weismann’s theory of evolution, we may say that his theory of heredity is, as regards all essential points, indistinguishable from that of Galton.

The truly scientific attitude of mind with regard to the problem of heredity is to say, as Galton says, “that we might almost reserve our belief that the structural [i.e., somatic] cells can react on the sexual elements at all, and we may be confident that at most they do so in a very faint degree; in other words, that acquired modifications are barely, if at all, inherited, in the correct sense of that word.” But for Weismann’s further theory of evolution, it is necessary to postulate the two additional doctrines in question; and it makes a literally immeasurable difference to the theory of evolution whether or not we entertain these two additional postulates. For no matter how faintly or how fitfully the substance of heredity may be modified by somatic tissues, by external conditions of life, or even by so-called spontaneous changes on the part of this substance itself, numberless causes of congenital variation are thus admitted, while even the Lamarckian principles are hypothetically allowed some degree of play. And although this is a lower degree than Darwin supposed, their influence in determining the course of organic evolution may still have been enormous; seeing that their action in any degree must always have been directive on the one hand, and cumulative on the other.

Having thus pointed out the great distinction between the theories of stirp and of germ-plasm, it became needful to note that Weismann himself is not consistent in observing it. On the contrary, in some passages he apparently expresses himself as willing to resign both his distinctive postulates—continuity as perpetual, and stability as absolute. But it is evident that such passages must be ignored by his critics, because, although as far as his theory of heredity is concerned they betoken an approach to the less speculative views of Galton, any such approach is proportionally destructive of his theory of evolution. It must not be supposed that I am taking an ungenerous advantage of these occasionally fundamental concessions. On the contrary, one cannot but admire the candour which they display. But, as I have said, it is necessary for us to ignore them, if only in order to examine the Weismannian theory of germ-plasm as a distinctive theory at all. And more than this. Seeing that his theory of heredity differs from Galton’s chiefly in being further an elaborate theory of evolution (founded on the two additional postulates in question), my main object has been to show the enfeeblement of the former which Weismann has caused by his addition of the latter. If he were to express his willingness to abandon his theory of evolution for the sake of strengthening his theory of heredity by identifying its main features with those of Galton’s, personally I should have no criticism to pass. Indeed, I was myself one of the first evolutionists who called in question the Lamarckian factors; and ever since the publication of Galton’s theory of heredity at about the same time, I have felt that in regard to its main principles—or those in which it agrees with Weismann’s—it is probably the true one. But I can nowhere find that Weismann is thus prepared to surrender his theory of evolution. Occasionally he plays fast and loose with the two additional postulates on which this theory is founded; but he does so without appearing to perceive the speculative impossibility of any longer sustaining his temple of evolution if he were to remove its pillars of germ-plasm.

Ignoring, then, these inconsistencies, we proceeded to examine separately, and on their own respective merits, the two distinctive postulates of the theory of germ-plasm—perpetual continuity since the first origin of life, and absolute stability since the first origin of sexual propagation.

It does not appear to me that very much has to be said, either for or against the former postulate, on merely antecedent grounds, or grounds of general reasoning. Therefore I relegated to an Appendix my examination of what Weismann has argued on these grounds, while in the text I considered only what he has advanced as evidence a posteriori. Here, as we saw, he has developed three distinct lines of verification—viz. (A) the migration of germ-cells in some of the Hydromedusae,(B) the early separation of germ-cells in the ontogeny of certain Invertebrata, and (C) the alleged invariability of organisms which are produced parthenogenetically. But we have seen, with respect to (A), that the specialized character of germinal cells is a fact which every theory of heredity must more or less recognize; and, therefore, that the migration of these cells, wherever it may be found to occur, does not lend any peculiar countenance to Weismann’s theory. There may be many reasons for such migration other than the one which this theory assigns; while the reason which it does assign is rendered improbable by the consideration that in the Hydromedusae the material of heredity is already and richly diffused throughout the general tissues. (B) and (C) are both contrary to fact; and, therefore, in whatever measure they would have corroborated the theory had they proved to be true, in that measure must they be held to discountenance the theory now that they have been shown to be false.

It appears, then, that there is no evidence in support of the postulate of the perpetual continuity of germ-plasm. There is nothing to show the necessary non-inheritance of acquired characters. The only evidence which one can recognize as good, is that which makes equally in favour of the theory of stirp—or rather, of the well-known fact that congenital characters are at any rate much more heritable than are acquired: which, it is needless to repeat, is a widely different thing from proving—or even rendering probable—the absolute restriction of germ-plasm to a separate “sphere” of its own “since the origin of life.”

But now, although there is no evidence in support of this postulate, there is no small amount of evidence against it. For this evidence goes to indicate that no small amount of reciprocal action habitually takes place between body-tissues and germinal elements: indeed it seems almost to prove that the orbits of germ-plasm and somato-plasm are not mutually exclusive, but touch and cut each other to a considerable extent. The evidence in question, it will be remembered, is derived from the effects of puberty, senility, castration, &c.; the occasional effect of pollenization on the somatic tissues of plants; the influence which a stock occasionally exercises upon a scion, or vice versa, which proves the possibility of a transmission of hereditary characters by a mere grafting together of somatic tissues; the direct evidence given by De Vries that in certain Algae constituents of cellular tissue pass immediately from the maternal ovum to the daughter organism; and the evidence, both direct and indirect, which remains to be given on a larger scale in my subsequent volume, where we shall have to challenge the validity of Weismann’s fundamental postulate touching the non-occurrence of Lamarckian factors in any of the multicellular organisms.

It must here again be noticed that in those passages where he concedes the possibly “occasional” transmission of acquired characters Weismann is annihilating his own theory, root and branch. Thus, for example, in allusion to De Vries’ observation just mentioned, he says that we cannot exclude the possibility of “changes being induced by external conditions in the organism as a whole, and then communicated to the germ-cells after the manner indicated in Darwin’s hypothesis of pangenesis.” But it is obvious that the theory of germ-plasm must “exclude the possibility of such a transmission occasionally occurring”; for the very essence of that theory consists in its postulating a difference between germ-plasm and the general body-substance in kind, such that there never can be any “communication” from the one to the other “after the manner indicated by Darwin’s hypothesis of pangenesis.” Any prevarication over this point amounts simply to abandoning the theory of germ-plasm altogether, and opening up a totally distinct issue—namely, the relative importance of natural selection and the Lamarckian factors in the process of organic evolution. It may be perfectly true—and I myself believe it is perfectly true—that Darwin attributed too large a measure of importance to the Lamarckian factors; but whether or not he did so is quite a different question from that which obtains between his theory of pangenesis and Weismann’s theory of germ-plasm. The former question is whether we are to “modify” the theory of pangenesis, so as to constitute it the theory of stirp; the latter question is whether we are to “abolish” the theory of pangenesis, in favour of its logical antithesis, the theory of germ-plasm. And this question remains to be dealt with in my next volume.

Coming then, lastly, to the companion postulate of germ-plasm as absolutely stable since the first origin of sexual propagation, we had to observe that, unlike the one we have just been considering, there is an immensely strong presumption against it on merely antecedent grounds. That the most complex substance in nature should likewise be the most stable substance with regard to complexity of “molecular structure”; that the greater its complexity becomes the greater becomes its stability, so that while in the comparatively simple unicellular organisms it is eminently susceptible of modification by external conditions, it entirely ceases to be thus susceptible when it becomes evolved into the incomparably more complex and immensely more varied structures which form the bases of heredity in the multicellular organisms—where, also, it must come into ever more and more intricate as well as more and more diverse relations with the external world;—all this is, I repeat, well nigh incredible. At any rate, speaking for myself, I should require some enormous weight of evidence to balance so enormous an antecedent improbability, or before I could regard such a doctrine as meriting any serious attention.

What, then, is the evidence that has been adduced? We have found that this evidence is nil. On the other hand, we have found that the evidence against the doctrine is abundantly sufficient to annihilate the doctrine—and this quite apart from all the antecedent considerations just alluded to. For not only have we the sundry facts of bud-variation, a-sexual origin of species, &c., which contradict the doctrine; but we have also the results of direct experiment, which prove that the alleged stability of germ-plasm may be conspicuously upset by slight changes in the external conditions of life. So that both from within and from without the stability which is alleged in theory admits of being overturned by facts.

And here, in order to avoid all possible confusion, I must ask it once more to be noted that there is not, and never has been, any question touching the high degree of stability which is exhibited by whatever substance it is that constitutes the material basis of heredity. But this is a widely different thing from supposing the stability absolute, so that it can never have been affected in any degree since the first origin of multicellular organisms, or in any of the millions of species into which these organisms have ramified. And the fact that in some cases we are actually able to observe a change of congenital characters as resulting from some “spontaneous” change in the hereditary material itself (as in bud-variation), or from some change in the external conditions of life (as in Hoffmann’s experiments)—this fact is more than is required in order finally to overthrow the intrinsically untenable doctrine which is in question.

Now, with the collapse of this doctrine there collapses also the important chain of deductions therefrom, which together constitute Weismann’s new theory of evolution. In particular, that natural selection is the exclusive means of modification among all the Metazoa and Metaphyta, while it is as exclusively ruled out with respect to all the Protozoa and Protophyta; that individual variations among the former can only be determined by sexual unions, while among the latter they can only be determined by the direct action of the environment; that the origin of congenital variability in all the Metazoa and Metaphyta is to be sought, and can only be found, in variations which occurred millions of years ago in the Protozoa and Protophyta; that the “significance of sexual propagation” is to be found in the view, that by this means alone can congenital variations have been ever since produced; &c., &c.

Upon the whole then, it appears to me that both the fundamental postulates of the theory of germ-plasm are unsound. That the substance of heredity is largely continuous and highly stable I see many and cogent reasons for believing. But that this substance has been uninterruptedly continuous since the origin of life, and absolutely stable since the origin of sexual propagation, I see even more and better reasons for disbelieving. And inasmuch as these two latter, or distinctive, postulates are not needed for Weismann’s theory of heredity, while they are both essential to his theory of evolution, I cannot but regret that he should thus have crippled the former by burdening it with the latter. Hence my object throughout has been to display, as sharply as possible, the contrast that is presented between the brass and the clay in the colossal figure which Weismann has constructed. Hence, also, my emphatic dissent from his theory of evolution does not prevent me from sincerely appreciating the great value which attaches to his theory of heredity. And although I have not hesitated to say that this theory is, in my opinion, incomplete; that it presents not a few manifest inconsistencies, and even logical contradictions; that the facts on which it is founded have always been facts of general knowledge; that in all its main features it was present to the mind of Darwin, and distinctly formulated by Galton; that in so far as it has been constituted the basis of a more general theory of organic evolution, it has clearly proved a failure:—such considerations in no wise diminish my cordial recognition of the services which its distinguished author has rendered to science by his speculations upon these topics. For not only has he been successful in drawing renewed and much more general attention to the important questions touching the transmissibility of acquired characters, the causes of variation, and so on; but even those parts of his system which have proved untenable are not without such value as temporary scaffoldings present in relation to permanent buildings. Therefore, if I have appeared to play the rÔle of a hostile critic, this has only been an expression of my desire to separate what seems to me the grain of good science from the chaff of bad speculation. And the candour which Professor Weismann has always displayed towards criticism of this character enables me to hope with assurance, that I have said nothing which he himself will regard as inconsistent with high admiration of his work as a naturalist, or of his originality as a philosopher.