Weismannism up to date (1893).

Hitherto we have been considering Professor Weismann’s system as it stood prior to the publication of his most recent works on Amphimixis and The Germ-plasm, in 1891 and 1893 respectively. These later and highly elaborate essays present considerable modifications of the system, as it stood when the foregoing criticism was written. But, for reasons already stated in the Preface, it appears to me desirable to leave that criticism as it was originally constructed, and to supply this further chapter for the purpose of dealing with the large alterations of, and important additions to, the theory of germ-plasm, which the maturer thought of its gifted author has led him to announce.

A few general remarks may be most conveniently made at the outset.

In the first place, these recent publications present the advantage over their predecessors of being systematic treatises, instead of more or less independent papers. On this account they present a logical sequence of thought, which renders the task of examination much less difficult than it was in the case of the first volume of the Essays.

In the second place, as a result of his more matured reflection, Professor Weismann has himself perceived a considerable number of the difficulties and objections which I have set forth in the preceding chapters. And not only has he thus anticipated many of my criticisms; but, as a result of doing so, he has changed not a few of the most important parts of his previous system, with the result of greatly improving it.

But, in the third place, notwithstanding that his remarkable power of speculative thinking is everywhere united with adequate knowledge in the sundry branches of biological science with which it deals, I confess to a serious doubt whether it has not been permitted to enjoy an undue amount of liberty. If only they can be laced together by a thread of logical connexion, hypotheses are added to hypotheses in such profusion as we are acquainted with in the works of metaphysicians, but which has rarely been approached in those of naturalists. The whole mechanism of heredity has been now planned out in such minuteness of detail and assurance of accuracy, that in reading the account one is reminded of that which is given by Dante of the topography of Inferno. For not only is the “sphere” of germ-plasm now composed of nine circles (molecules, biophores, determinants, ids, idants, idio-plasm, somatic-idioplasm, morpho-plasm, apical-plasm), but in most of these regions our guide is able to show us such strange and interesting phenomena, that we return to the fields of science with a sense of having been indeed in some other world. Or, to change the metaphor, if it be the case that “a true scientific judgement consists in giving a free rein to speculation with one hand, while holding ready the break of verification with the other,” I think it must be admitted that, in as far as he has erred, Professor Weismann has done so by driving a chariot which is unprovided with any break at all.

Hence, fourthly, it is needless to follow, even in epitome, the innumerable windings of these never-ending speculations. For, on the one hand, it would be impossible to do so without adding an unduly extended chapter to our already tediously prolonged consideration of Weismann’s views; while, on the other hand, we should have to deal merely with matters of comparative detail. The additions which have been made to his theory by his most recent publications are chiefly concerned with the matter just alluded to—viz., a minute elaboration of the hypothetical mechanism of heredity, in accordance with the general theory of germ-plasm. Without question this elaboration is everywhere thoughtful, and often highly ingenious; but until the general theory in question shall have been satisfactorily grounded, it seems premature to supply so immense a design of purely deductive construction. Beautiful though it may be in its imposing elevation, this drawing of “the architecture of germ-plasm” must be regarded as a work of artistic imagination rather than as one of scientific generalization. From the latter point of view it is at most a temple in posse, and even if it is ever to be realized in esse, we cannot allow the actual building to begin until we are much more sure than anybody is at present entitled to be touching the foundations on which it is proposed to rear so great an edifice.

Again, and fifthly, even if Weismann should ever be able to satisfy us upon this matter, or fully to demonstrate his basal proposition touching the perpetual continuity of germ-plasm, there would still be a far cry between accepting this sufficiently simple proposition and supposing that there is any adequate reason for entertaining so complex a scheme of the structure of germ-plasm. No doubt Weismann himself would be quite ready to admit, that from his basal proposition of the continuity of germ-plasm it is logically possible to construct many other designs of the architecture of germ-plasm, besides the one which he has so beautifully drawn. And although most of such alternative designs would doubtless embody some one or other of the features which are presented by his own, no one could say which features common to any two of the designs represent the facts. For in the case of all alike there would be a necessary absence of verification: the architects would all and equally have to acknowledge that their imposing pictures of “the palace of truth” were but imaginary. Such, in my opinion, has been the case with all theories of the ultimate mechanism of heredity hitherto published; but the difference between them and Weismann’s theory in this respect is, that while most of the others have not gone into speculative details further than was necessary as a means of substantiating their basal postulates, Weismann’s, as now developed in The Germ-plasm, is mainly concerned with such speculative details as an end, or object, per se.

But, it may be replied, by thus constructing an ideal mechanism of heredity Weismann is greatly strengthening his fundamental postulate of the continuity of germ-plasm, because he shows how all the main facts of heredity, and allied phenomena, admit of being explained if once the postulate be accepted. If this were urged, however, I should have two remarks to offer. The first is that Weismann, in constructing his ideal mechanism, has gone very much further in the way of elaboration than can possibly be required for this purpose. So much further, indeed, that his purpose has evidently been the constructing of his ideal mechanism, as I have just said, for its own sake, and not for the sake of substantiating its basal proposition by showing how well the latter can be made to work in explaining the phenomena of heredity, &c. Moreover—and this is my second remark—however well the basal proposition may be made to work in this respect, we must not be deceived into supposing that such a fact is equivalent to a substantiation of the proposition. This proposition—the continuity of germ-plasm—is the inverse of that which constitutes the basis of the theory of pangenesis. For while the latter assumes that in the last resort it is always somatic tissues which produce the substance of heredity, the former simply inverts the terms of this assumption, and holds that it is always the substance of heredity which produces the somatic tissues. Now, in all cases where one theory consists in thus simply inverting the terms of another, it will be found that the facts which they both seek to explain lend themselves equally to explanation by either, up to some certain and usually distant point, where a crucial test becomes possible. Take, as an example, the geocentric and heliocentric theories of the solar system. Here the question was whether the earth moved round the sun, or vice versa; and so many of the facts of observation lent themselves equally well to either interpretation, that it was very many centuries before the crucial tests were forthcoming. So, in the present instance, the question is as to whether the carriers of heredity move from body-cells to germ-cells, or vice versa; and it is because the theory which sustains the latter view has merely to invert the terms of the one which takes the former, that so many of the facts of observation lend themselves equally well to both—as we have seen in chapter III (pp. 56-59).

Lastly, yet another reason for not considering in any detail Professor Weismann’s intricate speculations on the ultimate mechanism of heredity is, that by so doing I should have found it impossible to avoid obscuring the main issues. For even Professor Weismann himself, by the extreme care which he has taken in fully presenting his scheme of this ultimate mechanism, has not found it practicable to keep distinctly before our view the relative insignificance of such details, as compared with the fundamental importance of his original postulates. Hence, I have deemed it best in the present chapter to restrict our attention to the changes which he has recently made in these the foundations of his entire system.

For these reasons, then, I will mention only those main features in the “architecture of germ-plasm” which it is necessary to understand for the purposes of the following criticism touching the general theory of germ-plasm in the most recent phase of its evolution.

To begin with, Weismann has now seen the desirability of ceasing to designate the ultimate “carriers of heredity” by the term “molecules.” Indeed, in these later volumes he has fully anticipated my remarks touching the use of this term in his previous “Essays[33].” The result of his more mature reflection may be presented in epitome thus.

A number of “molecules,” in the proper or chemical sense of the word, go to form a “biophore,” which is the ultimate unit of living substance.

A number of “ biophores” go to form a “determinant,” which is a special element in the germ-plasm, capable of directing the ontogeny of such and such a group of cells as is independently variable from the germ onwards.

A number of “determinants” go to form an “id,” which is the same hypothetical body as Weismann has hitherto designated by the term “ancestral germ-plasm.” That is to say, it is a group of determinants indissolubly united in phylogeny, and therefore transmitted by heredity as one complex whole. Ids are, perhaps, microscopically visible; and, if so, they probably correspond to the small granules (microsomata), which are familiar to the histologist in the structure of chromosomes.

A number of “ids” go to form an “idant,” which is a chromosome, or chromatin fibre[34].

In my opinion the most important advance which Weismann has made in his theory by means of this scheme has reference to the third of these divisions—the determinant. It is a matter of observation that every cell of a multicellular organism does not vary independently: it appears to be always the case that in the phenomena of variation a smaller or a larger group of cells is concerned. Now there must be something that determines the similar and simultaneous variation of such a whole group of cells; and, in all cases where such a variation is congenital, it is certain that this something must be contained in the substance of heredity. So far, I think, we must all agree, whether or not we regard this substance as “germ-plasm.” In other words, whether we regard the carriers of heredity as proceeding centrifugally (germ-plasm) or centripetally (gemmules), it seems to me that we ought to accept Weismann’s doctrine of determinants. Indeed, pathologists have already furnished a foreshadowing of such a doctrine in regard to the phenomena presented by certain diseases, such as cancer; but it is an important step to have extended the idea from pathology to biology in general—and, at the same time, to have given it a more definite shape than it has hitherto presented. In Weismann’s hands it serves to render more conceivable—if not also more intelligible—that process of marshalling cell-formations, which, be our theories what they may, is assuredly the most distinctive and remarkable fact of ontogenetic organization.

Again, as regards the id, I do not see how any one can attentively read Professor Weismann’s discussion without acknowledging that, if we once accept his doctrine of determinants, his sequent doctrine of ids becomes a logical necessity.

On the other hand, however, I do not see that such is the case with respect to idants; and still less do I see any reason for identifying the latter with chromosomes—even assuming that chromosomes are the visible repositories of the carriers of heredity[35].

Referring the reader to Weismann’s own exposition for a full account of these and many other additions to his general theory of germ-plasm, I will at once proceed to consider the alterations or emendations of that theory which have been published in his last two volumes, and which, as we shall find, have in large measure anticipated some of the most important points in the foregoing criticism. Therefore in the following criticism I will consider seriatim what he has now said touching all these points, and conclude by offering some general remarks on the resulting position of his general system of theories up to the present date.

Pursuing the same method of criticism as that adopted in the preceding chapters, we will first consider the further modifications of Weismann’s theory of heredity, and next those of his theory of organic evolution.

Weismann’s theory of Heredity (1893).

First of all, Weismann has now profoundly modified his theory of polar bodies. For, owing to certain more recent researches of Professor O. Hertwig, he very candidly allows:—“My previous interpretation of the first polar body as the removal of ovogenetic nucleo-plasm from the egg must fall to the ground: about this there is no possible doubt[36].”

He now regards both polar bodies as concerned in the same function of removing superfluous germ-plasm. Therefore one-half of his previous theory is abandoned: “the ovogenetic idio-plasm” is now supposed to be simply absorbed in the course of ontogeny, as I had suggested in one of the preceding chapters (pp. 42-46). The consequence is that he has now nothing to oppose to the view which is likewise there suggested (pp. 43-44)—viz., that his whole theory of polar bodies is rendered needless and improbable by the fact that the very mode in which ova are produced renders ample provision for the removal of any amount of superfluous germ-plasm which the theory of germ-plasm may require.

It is needless to say, after what has already been said in the pages just referred to, that in my opinion Professor Weismann has improved his main system of theories by dropping this part of his subordinate and, for the most part, separate theory of polar bodies. I only wish he could have seen his way to dropping the whole.

Again, he has now fully considered the phenomena of repair, regeneration, reproduction from somatic tissues, budding, and graft-hybridization.

Touching the four former he takes the view which I have supposed that he would (p. 53). As regards the latter, he fully accepts the fact of an occasional transmission of characters from one species or variety of plant to another by mere grafting[37]. But, although the explanation which he gives of this fact may pass muster so far as the only case which he deals with in detail is concerned, I do not see how it can do so to many others. For the case which he considers is that of Cystisus adami, where a bud of one species of Laburnum having been inserted in the wood of another produced a shoot which presented intermediate characters; and these have ever since been propagated by cuttings. Weismann’s interpretation of the facts here is, “that they were due to an abnormal kind of amphimixis, so that the idants of both species were combined in the apical cell of the first shoot[38].” Now, although this explanation may well apply to a case of graft-hybridization by means of buds, it obviously cannot do so to any case where hybridization is produced by the grafting of woody tissues. For here there is no “apical cell” in the question; and therefore the difficulties which I have adduced on page 82 remain. Possibly Weismann may dispute the fact of hybridization in any of these cases; but, as he has not expressly done so, I will not go into the question of evidence[39].

One important addition to this side of Weismann’s system has been made in order to meet the class of difficulties which are presented by the apparent inheritance of certain climatic variations, as already mentioned on pp. 67-8. For example, his own butterflies seemed to render definite proof of somatogenetic variations caused by changed conditions of life being transmitted to progeny. Therefore, it will be remembered, Weismann candidly admitted, “even now I cannot explain the facts otherwise than by supposing a passive acquisition of characters produced by the direct influence of climate”—i.e., an exactly representative copying in progeny of characters acquired by parents. I have already quoted these words in order to show their logical inadmissibility as used by Weismann. He cannot be allowed thus to entertain the Lamarckian factors and at the same time to maintain his theory of germ-plasm, which excludes them as physiologically impossible. Doubtless he was himself aware of this, for he immediately added that “new experiments will be necessary to afford the true explanation[40].”

The explanation, however, which he now gives is not based on any new experiments, but on a new suggestion to the effect that all such seemingly conclusive instances of the inheritance of acquired characters are, in truth, illusory. This suggestion is that “Many climatic variations may be due wholly or in part to the simultaneous variation of corresponding determinants in some parts of the soma, and in the germ-plasm of the reproductive cells.[41]” For example, if, as Weismann now supposes, determinants of the same kinds occur in the somatic tissues as well as in the germ-cells, when a particular spot occurs on a butterfly’s wing, it has been due to a particular kind of determinant which in the course of ontogeny was transmitted from the germ-cell for the express purpose of controlling the size and colour of the spot. But a residue of precisely similar determinants was reserved in the germ-cell (germ-plasm), for the purpose of determining a precisely similar spot in the next generation. Hence, if a rise of temperature, or any other external change, is capable of so acting on the determinant in the soma as to cause it to impart an abnormal colour to the spot when formed, a similar change is likely to be simultaneously effected in the corresponding determinants which are lying dormant in the germ-plasm. Therefore, when the latter become active in the ontogeny of the next generation, they will produce spots presenting the same variations as those of the preceding generation. Obviously, however, there would not be here any transmission of acquired characters. The change would be “specialized,” but not “representative.”

No doubt we have here a sufficiently ingenious method of circumventing an awkward class of facts. But I should like to make two observations with regard to it.

In the first place, the suggestion is highly speculative, and has been advanced solely for the sake of saving the theory of germ-plasm. There are no facts adduced in its favour, and it could scarcely be entertained as in the least degree probable by any one who has not already accepted the theory in question. Hence, unless we are to embark on a course of circular reasoning, we must refuse to accept the explanation of hereditary climatic variation now offered, until it shall have been fully corroborated by the experimental enquiry to which Weismann says he is now submitting it.

My second observation is, that the suggestion is not new; but appears to have been derived from Professor Weismann’s recent study of Mr. Galton’s Theory of Heredity. At all events, the suggestion is there presented with sufficient lucidity, thus:—

It is said that the structure of an animal changes when he is placed under changed conditions; that his offspring inherit some of his change; and that they vary still further on their own account, in the same direction, and so on through successive generations, until a notable change in the congenital characteristics of the race has been effected. Hence, it is concluded that a change in the personal structure has reacted on the sexual elements. For my part, I object to so general a conclusion, for the following reasons. It is universally admitted that the primary agents in the processes of growth, nutrition, and reproduction are the same, and that a true theory of heredity must so regard them. In other words, they are all due to the development of some germinal matter, variously located. Consequently, when similar germinal matter is everywhere affected by the same conditions, we should expect that it would be everywhere affected in the same way. The particular kind of germ whence the hair sprang, that was induced to throw out a new variety in the cells nearest to the surface of the body under certain changed conditions of climate and food, might be expected to throw out a similar variety in the sexual elements at the same time. The changes in the germs would everywhere be collateral, although the moments when any of the changed germs happened to receive their development might be different[42].

This allusion to Mr. Galton’s Theory of Heredity leads me to consider what Professor Weismann has said with regard to it in this latest publication, where, for the first time, he has dealt with it. In my opinion he has done but scant justice to the views of his predecessor, and therefore I will occupy some considerable space in seeking to justify this opinion.

As already stated, from the time that Mr. Galton published his theory I have felt that in its main contention it presents a probably true solution of the main problem of heredity—viz., to account for the contrast between congenital and acquired characters in respect of transmissibility. And this solution, as likewise already stated, was substantially identical with that which Professor Weismann published in the next decade. Indeed, the only important difference between these two theories of heredity is, that while Weismann’s excludes on deductive grounds the physiological possibility of the inheritance of acquired characters, Galton’s more judiciously leaves to be determined, by subsequent enquiry of the inductive kind, the question whether acquired characters are ever transmitted in faint degrees, or whether they are never transmitted at all. In addition to this important difference, however, there are certain others which seem to me of very little consequence, inasmuch as they have reference to speculations on the ultimate mechanism of heredity, or the intimate morphology and physiology of the carriers of heredity—speculations which it would be absurd to suppose can be other than purely conjectural. Therefore in my previous criticism I did not allude to these subordinate points of difference, but stated merely, in general terms, that Galton’s view of the ultimate mechanism in question was such as to leave room for the possibility of the occasional transmission of acquired characters. And in this respect, it still seems to me, his theory has an advantage over that of Weismann. No doubt the latter is a much more elaborate and highly finished piece of work; but beauty of ideal construction is no guarantee of scientific truth—as we shall presently find exemplified in a striking manner with regard to Weismann’s theory of evolution. And if his theory of heredity, in its final shape, is a much more precise, detailed, and logically coherent structure than any which has ever been framed in this department of biological thought, there is all the more reason to scan critically the fundamental postulate on which it rests. Hence I cannot help feeling that it will be time enough to consider minor differences between the two theories when the physiological possibility of the occasional transmission of acquired characters, as entertained by Galton’s theory, shall have been ruled out as demonstrably opposed to fact.

Seeing, however, that Professor Weismann thinks otherwise, and appears to attach as much importance to differences concerning deductive minutiae as he does to those concerning fundamental principles, I will here contrast the two theories somewhat more in detail than heretofore, and with special reference to what he has now himself said touching their relationship.

It will be remembered that the primary or fundamental difference just alluded to is, that while the theory of germ-plasm postulates an absolute continuity, the theory of stirp postulates but a partial continuity, of the substance of heredity. Hence, according to Weismann’s view, we must go back to the unicellular organisms for the origin of this substance in the multicellular; and we must regard use-inheritance as physiologically impossible. On the other hand, according to Galton’s view, there is no necessity for us to do either of these things. The origin of stirp is to be found in the somatic tissues of the multicellular organisms themselves. Nevertheless, this theory differs greatly from pangenesis, in that the former supposes the origin of hereditary substance to be mainly given in the phylogeny of any group of multicellular organisms, while the latter supposes it to be given mainly in each ontogeny, Galton’s theory is, that in each ontogeny only a small part of the stirp derived from parents is consumed in making the new organism—the larger part being handed over in trust for passing on to the next generation, in the same way as Weismann supposes to be the case with germ-plasm. Darwin’s theory, on the other hand, does not entertain any such notion of “continuity” in the substance of heredity from germ-cell to germ-cell of parent and offspring; it supposes that in each successive generation the germ-cells are wholly supplied with their germinal material from somatic-cells of each individual organism. Or, adopting our previous terminology, the three theories may be ranked thus.

The particulate elements of heredity all proceed centripetally from somatic-cells to germ-cells (gemmules): the inheritance of acquired characters is therefore habitual.

These particulate elements proceed for the most part, though not exclusively, from germ-cells to somatic-cells (stirp): the inheritance of acquired characters is therefore but occasional.

The elements in question proceed exclusively in the centrifugal direction last mentioned (germ-plasm): the inheritance of acquired characters is therefore impossible[43].

Such being the fundamental points of difference between these three theories of heredity, we have now to consider more particularly those which obtain between Galton’s and Weismann’s.

The general doctrine of gemmules (i. e. somatic-cell-germs) is accepted by Galton; but instead of supposing, as Darwin supposed, that these minute bodies freely circulate through all the body tissues, so that some of them are absorbed from all the somatic-cells by the germ-cells, and there constitute the entire mass of hereditary material out of which the offspring will afterwards be formed, Galton supposes that gemmules circulate with comparative difficulty, and that only comparatively few of them gain access to the germ-cells in each generation. Hence, characters acquired in the individual lifetime are much less heritable than those which are called congenital. For congenital characters are due to the “continuity” of stirp through numberless generations in the phylogeny of the organism; hence such characters are represented by a vastly greater number of equivalent hereditary elements. Weismann, on the other hand, rejects the doctrine of gemmules in toto.

Again, according to Galton’s view, “individual [congenital] variation depends upon two factors; the one is the variability of the germ[44] and of its progeny; the other is that of all kinds of external circumstances, in determining which out of many competing germs, of nearly equal suitability, shall be the one that becomes developed. The variability of germs under changed conditions, and that of their progeny, may be small, but it is indubitable; absolute uniformity being scarcely conceivable in the condition and growth, and, therefore, in the reproduction of any organism. The law of heredity goes no further than to say, that like tends to produce like; the tendency may be very strong, but it cannot be absolute[45].”

Here, of course, there is a wide difference between stirp and germ-plasm. For while Galton does not entertain amphimixis among the “factors” of congenital variation, Weismann, as we are now well aware, has hitherto regarded it as the sole cause of such variation. Nevertheless, as we shall presently find, Weismann has now greatly modified his views upon this point, and does entertain, in The Germ-plasm, both the “factors” mentioned by Galton. Hence, the difference between the two theories in question with regard to this matter is not nearly so wide as it was prior to the publication of Weismann’s last work.

The next most important point of difference between the theories of stirp and germ-plasm has reference to the mechanism of ontogeny. According to Galton, this is simply a struggle between all the carriers of heredity composing the stirp of a fertilized ovum. It is not, however, a struggle for existence, but what may be called a struggle for development. In the fertilized ovum all the carriers of heredity are, to begin with, in a “latent” condition; but of this enormous multitude of “germs” or “gemmules,” only a very small proportional number are destined to become “patent”—i. e., developed into the tissue-cells composing the new organism. The vast majority of the gemmules, or those which fail to be thus developed, go to constitute the stirp of the new organism when this has been formed by the development of the comparatively few successful gemmules. Thus much understood, the following quotation will be fully intelligible.

My argument is this: Of the two groups of germs, the one consisting of those that succeed in becoming developed and in forming the bodily structure, and the other consisting of those that remain continually latent, the latent vastly preponderates in number. We should expect the latent germs to exercise a corresponding predominance in matters of heredity, unless it can be shown that, on the whole, the germ that is developed into a cell becomes thereby more fertile than if it had remained latent. But the evidence points the other way. It appears both that the period of fertility is shorter, and the fecundity even during that period is less in the germ that becomes developed into a cell, than they are in the germ that remains latent. Much less then would the entire bodily structure, which consists of a relatively small number of these comparatively sterile units, successfully compete in matters of heredity with the total effect of the much more numerous and more prolific units which are in a latent form[46].

Thus, Galton’s theory of the mechanism of geny is a theory of struggle; and this constitutes a point of difference on which Weismann lays much stress in his latest work. For, as we know, Weismann regards the mechanism of ontogeny as characterized by a peaceful succession of “stages,” which are “pre-determined from the germ onwards”; and in his latest work this idea of orderly sequence has been further elaborated in his doctrine of “determinants.” In short, to adopt their own metaphors, while Galton tells us that the mechanism of ontogeny is like that of a political election, where rival candidates compete to “represent” the nation (stirp) in Parliament (individual organism); Weismann likens it to the mechanism of a well-drilled army, where ultimate carriers of heredity (privates) are banded together in companies, regiments, battalions, &c., under the command of corresponding officers (determinants).

Lastly, there is yet one further point of difference between stirp and germ-plasm, which is thus stated by Weismann:—

Galton’s idea is only conceivable on the presupposition of the occurrence of sexual reproduction, while the theory of the continuity of the germ-plasm is entirely independent of any assumption as to whether each primary constituent is present in the germ singly or in numbers. According to my idea, the active and the reserve germ-plasm contain precisely similar primary constituents, gemmules, or determinants; and on this the resemblance of a child to its parent depends. The theory of the continuity of the germ-plasm, as I understand it, is not based on the fact that each “gemmule” necessary for the construction of the soma is present many times only, so that a residue remains from which the germ-cells of the next generation may be formed: it is founded on the view of the existence of a special adaptation, which is inevitable in the case of multicellular organisms, and which consists in the germ-plasm of the fertilized egg-cell becoming doubled primarily, one of the resulting portions being reserved for the formation of germ-cells[47].

These being the main points of difference between the theories of stirp and of germ-plasm to which Professor Weismann has alluded, I will now proceed to consider them separately, in reverse order to that in which they have been here stated.

The point of difference last mentioned need not detain us long, because it seems to me one of very little importance. “Whether each primary constituent is present in the germ singly or in numbers” cannot greatly signify, so long as both theories agree that, sooner or later, they must be present plurally. Galton supposes them to be thus present from the first (i. e. in the unfertilized ovum), while Weismann supposes them to be so only as a result of their self-multiplication at a somewhat later stage (i. e. in the segmenting ovum, and onwards throughout the procreative life of the individual). Doubtless Weismann does not suppose that they ever become so numerous as Galton imagines; but the whole question is so highly speculative that I do not see how any useful purpose can be served by debating it. Nor do I see why Weismann should conclude that “Galton’s idea is only conceivable on the presupposition of the occurrence of sexual reproduction.” It is true that Galton has discussed exclusively the case of sexual reproduction; but I cannot perceive that any of his ideas are inapplicable to a-sexual.

Touching the question whether the phenomena of ontogeny had best be ascribed to a competition among a vast number of “germs,” or to a strictly ordered evolution of a comparatively small number of “determinants,” a considerable array of arguments might be adduced in support of either view. Thus, Galton might well maintain that his interpretation of the observable facts is most in accordance with the general analogies supplied by organic nature as a whole. The ancient aphorism of Heraclitus, “Struggle is the father, king, and lord of all things,” has been in large measure justified by Darwin and his followers, at any rate within the range of biology. Not only have we the “struggle for existence” where “the origin of species” is concerned; but Roux has well argued, in his remarkable work on Der Kampf der Theile im Organismus, that the principle of “struggle” is concerned to an equally important extent as between all the constituent parts of the same individual. But if this is so—if every tissue-cell of the organism owes its maintenance to success in a general contest for nutriment, &c.,—do we not find at least a probability that it owes its origin as a visible cell to a similar success in a similarly general contest among the invisible elements from which tissue-cells are developed? Nay, does it not seem well nigh incredible that when this selection-principle is seen to be the governing cause of evolution everywhere else, it should cease to play any part at all just at the place where we are unable to see what is going on? As we are agreed that this “father of all things” is of prime importance in phylogeny—to say nothing of physiology, psychology, and sociology,—must we not deem it absurd to suppose that it is supplanted in ontogeny by the opposite principle of absolute peace?

On the other hand, Weismann adduces many forcible considerations per contra; so that, in the result, I deem it best to dispose of the question with two general remarks. The first is, that the rival views are not necessarily incompatible. Each may present one aspect of the truth. Weismann’s doctrine of determinants may be—and, to the best of my judgement, must be—sound; but this does not hinder that Galton’s doctrine of struggling “germs” may be so likewise. For, as we have already seen, these germs present the same compound character which belong to determinants; in fact I do not suppose that Galton would object to identifying them with determinants. On the other hand, I do not see why Weismann should object to supposing that similar determinants compete among themselves for ontogenetic development. Indeed, he has already argued, in his suggestive theory of “germ-tracts,” that it is usually only one among a number of similar determinants which does succeed in achieving such development—or, as he expresses it, which “becomes active.” But what is it that causes this activity? Surely it must be some superiority on the part of the active determinant over its passive companions. And, if so, it is the selection-principle that is here at work. In fact, he has himself laid no small stress on what he calls “the struggle of the determinants of the two parents in ontogeny,” and has even supplied a long section on “the Struggle of the Ids in Ontogeny.” Therefore I do not see why he should so emphatically dissent from Galton’s view upon this matter as he does in his work on The Germ-plasm[48].

My second remark is a brief one—viz., that the whole question is of so very speculative a character, that I cannot see the smallest use in debating it.

The only remaining point of difference between strip and germ-plasm is the one referring to stability. Needless to say, Galton is at one with Weismann in recognizing a high degree of stability on the part of the substance of heredity; but the agreement extends only so far as is necessitated by the facts of atavism, &c. Indeed, he does not even mention—although he perhaps implies—what Weismann has called amphimixis as among the factors of individual congenital variation. Weismann, on the other hand, has hitherto regarded amphimixis as the sole cause of all such variations. But, as we shall presently find, in his recent work on The Germ-plasm he has now greatly modified his views upon this subject, and, in fully recognizing the “factors” of variability to which Galton alludes, has correspondingly lessened the difference between germ-plasm and stirp. But this is a point which can be better dealt with when we come to consider the important modifications which in this respect the theory of germ-plasm has undergone.

The only other matter which has to be mentioned in connexion with Weismann’s theory of heredity is, that in The Germ-plasm he has for the first time given us his views upon the influence of a previous sire on the progeny of a subsequent one by the same dam. The phenomena in question, which I have already detailed in pp. 77-9, 110, he designates by the term “telegony.” The analogous phenomena in plants he calls, following Focke, “xenia.”

With regard to telegony, he adopts, almost precisely, the position which I surmised that he would. That is to say, he first disputes the alleged facts, and then argues that, even if they be facts, they admit of being explained on the theory of germ-plasm by supposing that some of the germ-plasm from the first sire penetrates the unripe ova which are afterwards fertilized by the second[49]. The only difference between his views and my own upon this matter is, therefore, as follows.

Supposing that the phenomena alleged ever occur in fact, I have said that the only way of explaining them would seem to be, “that the life of ‘germ-plasm’ is not conterminous with that of the spermatozoa which convey it, and hence that, if the carriers of heredity, after the disintegration of their containing spermatozoa, should ever penetrate an unripe ovum, the germ-plasm thus introduced might remain dormant in the ovum until the latter becomes mature, and is then fertilized by another sire. In this way it is conceivable that the hitherto dormant germ-plasm of the previous sire might exercise some influence on the ontogeny of the embryo[50].”

Now, this is substantially the position which Weismann takes up: only instead of supposing that it is the “carriers of heredity” of the first sire which gain access to the unripe ovum “after the disintegration of their containing spermatozoa,” he supposes that it is one of the spermatozoa which does so before its disintegration has commenced. Of course there is here no difference in principle, but only a question touching the mode in which the access is presumably effected. But, as regards this question, I retain my original opinion. For, while I can see no theoretical difficulty in supposing that “the carriers of heredity,” when set free by the disintegration of their containing spermatozoa, may reach the unripe ova while still embedded in the depths of the ovary, I do see a difficulty, amounting almost to a physiological impossibility, in supposing that a whole spermatozoon can perform such a feat. From all that we know about the powers and functions of spermatozoa in the vertebrata, it appears simply absurd to imagine that these bodies are able to penetrate the dense coating of an ovary, and then delve their way through the stroma. There is, indeed, a remarkable investigation which was published a year or two ago by Mr. Whitman[51] which appears to prove that in certain leeches the male injects his seminal fluid into any part of the body of the female, and that the spermatozoa then reach the ova by wandering about her general tissues until some of them happen to hit upon her ovary. But in this case the spermatozoa are specially adapted to perform such acts of penetration—being spear-like bodies provided with a sharp point. Hence, if Weismann should quote this instance, it would not tend to support his view, seeing that the spermatozoa of mammals do not exhibit any such specializations of structure; and therefore, before any one of them can effect fertilization, must wait for the ovum to mature, reach the surface of the ovary, and rupture its follicle.

But, as already observed, it does not signify, so far as we are here concerned with the matter, in what precise manner the telegonous influence may be supposed to be exercised—provided that it may be so directly, and not necessarily through first having to influence the whole material organism. Therefore I quite agree with Weismann that the facts—supposing them to be facts—are quite as explicable by the theory of germ-plasm as by that of pangenesis[52].

Again, with respect to xenia, Weismann writes:—

As such eminent botanists as Focke, and more recently De Vries, have expressed much doubt with regard to these observations—or rather interpretations—we must wait until these cases have been critically re-investigated before attempting to account for them theoretically. The chief difficulty we should meet with in any such explanation would be due to the fact that we are here concerned with the influence of the germ-plasm of the sperm-cell on a tissue of another plant which only constitutes a part of this plant. It would thus be necessary to assume that all the determinants of this germ-plasm are not active, and that only those take effect which determine the nature of the fruit.

Now, it does not appear that De Vries has looked into the matter on his own account, as he merely refers to what Focke has said. And this amounts merely to showing the dubious character of some half-dozen cases which Focke gives as those which alone have fallen within his cognizance. Why he does not mention any of the numerous cases which are quoted by Darwin, I do not understand. Nor can I understand why he does not consider what seem to be the particularly conclusive facts given on p. 80,—i. e., where xenia appears to constitute “a needful preliminary to fertilization.” But the whole matter is one for botanists to deal with, and if any doubt attaches to it, at least the grounds of such doubt should be fully stated. Still more, in my opinion, should the matter be freed from any such doubt. The question—if there be a question—is one of great interest from a merely physiological point of view, while in relation to the fundamental problems of heredity its importance is immense. Surely, then, any competent botanist who disputes the facts ought to test them by way of experiment.

But, be this as it may, I must call prominent attention to the following very remarkable words wherewith Weismann concludes the passage above quoted. For he there says, that even supposing there were no doubt as to the facts or their interpretation, “the chief difficulty” which they would oppose to the theory of germ-plasm would be, “that we are here concerned with the influence of the germ-plasm of the sperm-cell on a tissue of another plant which only constitutes a part of this plant.” In other words, Weismann now freely entertains the possibility of a direct action of germ-plasm on the somatic tissues, even though these belong to another individual! Thus he now concedes the only point for the establishment of which I adduced the phenomena of xenia, in Chapter III: the whole of one side of that “reciprocal action between the sphere of germinal-substance and the sphere of body-substance,” which I contended for on pp. 76-85, is now conceded; and although it is the less important side, its surrender goes far to weaken the doctrine of a perpetual isolation of germinal-substance to a “sphere” of its own. If we suppose that the germinal substance of one organism may thus directly act upon the somatic tissues of another, and that changed conditions of life are able to produce simultaneously an acquired character in the soma and a precisely identical character as congenital in the germ (pp. 129-30), we are plainly inviting ourselves to abandon the complex explanation of living material in “two kinds,” where one is capable in all sorts of ways of communicating with the other, while the possibility of any reciprocal action is excluded. For the simpler hypothesis of living material as all of one kind encounters no such antinomies. So long as one kind of this material was supposed to be as distinct from the other as a parasite is distinct from its host, there was not so much to choose between the theory of germ-plasm and that of gemmules in this respect of simplicity. But the more that the former theory has had to be adjusted to facts, the greater has its complexity become, until now its own author is obliged to make so many additional assumptions for the purpose of mantaining it, that we begin to wonder how long it can continue to support the weight of its accumulating difficulties.

So much for the main modifications which have this year been made in Weismann’s postulate of the perpetual continuity of germ-plasm. We must next consider the changes which he has effected in his companion postulate of the absolute stability of germ-plasm.

Weismann’s Theory of Evolution (1893).

Of far more importance than any of the alterations which Professor Weismann has recently made in his theory of heredity, are those whereby he has modified his sequent theory of evolution. For while, as we have just seen, his work on The Germ-plasm leaves the former theory substantially unaltered,—although largely added to in matters of detail,—it so profoundly modifies the latter that careful readers will find no small difficulty in ascertaining how much of it has been allowed to remain. I will consider only the main modifications, and these I will take separately.

It will be remembered that one distinctive feature in Weismann’s theory of evolution has hitherto been, that the unicellular organisms differ from the multicellular in the following important particulars.

1. There being no division in unicellular organisms between germ-cells and somatic-cells, there is no possibility in them of the occurrence of amphimixis.

2. Consequently, there is no possibility in them of congenital variations, in the sense that these occur in multicellular organisms.

3. Hence the only causes of individual variation and of the origin of species in the unicellular organisms are the Lamarckian factors, just as in the multicellular the only cause of these things is natural selection.

4. Hence, also, the unicellular organisms are potentially immortal, while the multicellular have acquired mortality for certain adaptive reasons.

But now, with the exception of No. 4, all these positions have been abandoned. For, chiefly on account of the beautiful researches of Maupas, Weismann has come to perceive that no real distinction can be drawn between an act of sexual union in the multicellular organisms, and an act of conjugation in the unicellular. Amphimixis, therefore, is now held by him to occur equally in both these divisions of organic nature, with the consequence that the Protozoa and Protophyta owe their individual variations, and therefore the origin of their innumerable species, as exclusively to the action of natural selection as is the case with the Metazoa and Metaphyta. In fact, the term “amphimixis” has been coined in express relation to these very points.

It will be seen, however, that this important change of view merely postpones the question as to the origin of amphimixis, if the object of this process be that which Weismann supposes—viz., the providing of material in the way of congenital variations on which natural selection can act. Therefore he is obliged to assume that there now are, or once have been, organisms of a less organized character than even the lowest of the unicellular forms—organisms, that is to say, which possess no nucleus, but are wholly composed of undifferentiated bioplasm. These most primitive organisms it must have been that were not subject to any process of natural selection, but, in virtue of an exclusive action of the Lamarckian factors upon their protoplasmic substance, gave rise to individual variations which subsequently gave rise to a unicellular progeny—when the process of natural selection was immediately inaugurated, and thereafter entirely superseded the Lamarckian factors. Or, to state the matter in Weismann’s own words:—

My earlier views on unicellular organisms as the source of individual differences, in the sense that each change called forth in them by external influences, or by use and disuse, was supposed to be hereditary, must therefore be dismissed to some stage less distant from the origin of life. I now believe that such reactions under external influences can only obtain in the lowest organisms which are without any distinction between nucleus and cell-body. All variations which have arisen in them, by the operation of any causes whatever, must be inherited, and their hereditary individual variability is due to the direct influence of the external world.... If I am correct in my view of the meaning of conjugation as a method of amphimixis, we must believe that all unicellular organisms possess it, and that it will be found in numerous low organisms, in which it has not yet been observed[53].

It is not very clear, at first sight, how Professor Weismann, after having thus abandoned the propositions 1, 2, and 3, as above stated, manages to retain his former view as given in No. 4. Nevertheless he does so, by representing that a unicellular organism, even though it present such a considerable degree of organization as we meet with in the higher Protozoa, still resembles a germ-cell of a multicellular organism, in that it consists of all the essential constituents of a germ-cell, including germ-plasm in its nucleus. And inasmuch as a germ-cell is potentially immortal, so it must be with a unicellular organism; in the one case, as in the other, the design of the structure is that its contained germ-plasm shall fuse with the germ-plasm contained in the nucleus of another individual cell, when the life of both will be preserved. For my own part, however, I cannot see that in either case the cell, as distinguished from its contained germ-plasm, is thus shown to be potentially immortal. On the contrary, it appears to me a mere accident of the case that in a unicellular organism the immortal substance (germ-plasm) is contained in a single cell, which is at the same time a free cell, and, as such, is denominated an “organism.” We might just as well call a germ-cell an “organism,” whether as an ovum it happens to be embedded in a mass of somatic-cells, or as a locomotive spermatozoon it happens to be free. In fact Weismann himself appears to recognize this. But, if so, it is surely a distinction without a difference to say that unicellular organisms are immortal, while multicellular are mortal. For in neither case is the organism immortal, while in both cases it is the germ-plasm (i.e., the substance of heredity) that is so. Where the cell containing the germ-plasm happens to be a free cell, it is called an “organism”; but whether it be a germ-cell or a protozoan, it alike ceases to be a cell when it has given origin to a multitude of other cells, whether these happen to be other germ-cells (plus somatic-cells) or other protozoan cells. In short, qu cell, all cells are mortal: it is only the substance of heredity which some cells contain that can be said, in any sense of the term, to be immortal. For the immortality in question does not belong to unicellular organisms as such, but to the germ-plasm which they contain. And from this it follows that, as the immortality of germ-plasm is one and the same thing as the continuity of germ-plasm, by alleging an immortality as belonging to the unicellular organisms, Weismann is merely restating his fundamental postulate. Hence, also, he is but denying, in a somewhat round-about way, the occurrence of spontaneous generation.

I conclude, therefore, that his sole remaining distinction between the unicellular and the multicellular organisms is but illusory, or unreal. And, with regard to the great change which he has thus effected in his system by expressly abolishing all the other distinctions, I have only to say that in my opinion he has thereby greatly improved his system. For he has thus relieved it of all the formidable difficulties which he had needlessly created for himself, and which I have already enumerated in the foregoing pages (88-89). In his ever-shifting drama of evolution the unicellular organisms have left the stage en masse, and, so far as they are concerned, we are all as we were before the curtain rose.

But of even more importance than this fundamental change of view with regard to the unicellular organisms, is a further and no less fundamental change with regard to the multicellular. That such is the case will immediately become apparent by a simple statement of the fact, that Weismann has now expressly surrendered his postulate of the absolute stability of germ-plasm!

We have already seen that, even in the first volume of his Essays, there were some passages which gave an uncertain sound with regard to this matter. But as they seemed attributable to mere carelessness on the part of their author, after quoting a sample of them, I showed it was necessary to ignore such inconsistent utterances—necessary, that is, for the purpose of examining the theory of germ-plasm as even so much as a logically coherent system of ideas[54]. For we have seen that if any doubt were to be entertained touching the absolute stability of germ-plasm “since the first origin of sexual propagation,” a corresponding measure of doubt would be cast on Weismann’s theory of congenital variation as solely due to amphimixis, with the result that his whole theory of evolution would be similarly rendered dubious. Since then, however, he has gone very much further in this direction. First, in reply to Professor Vines he says (1890):—

I am at present inclined to believe that Professor Vines is correct in questioning whether sexual reproduction is the only factor which maintains Metazoa and Metaphyta in a state of variability. I could have pointed out in the English edition of my “Essays” that my views on this point had altered since their publication; my friend Professor de Bary, too early lost to science, had already called my attention to those parthenogenetic Fungi which Professor Vines justly cites against my views; but I desired, on grounds already mentioned, to undertake no alteration in the essays[55].

Next, in his essay on Amphimixis (1892), there are several passages to somewhat the same effect; while, lastly, in his Germ-plasm (1893), the fundamental postulate in question is, as I have said, expressly surrendered. For example, we have in the following words the final conclusions of his recent arguments. Speaking of amphimixis, he says:—

It is not the primary cause of hereditary variation. By its means those specific variations which already exist in a species may continually be blended in a fresh manner, but it is incapable of giving rise to new variations, even though it often appears to do so.... The cause of hereditary variation must lie deeper than this. It must be due to the direct effects of external influences on the biophores and determinants.[56]

These quotations are enough to show that Weismann has now abandoned his original theory of congenital variations being exclusively due to amphimixis, and adopts in its stead the precisely opposite view—viz., that the origin of all such variations must be ascribed to the direct influence of causes acting on germ-plasm from without. Up to the present year the very essence of the whole Weismannian theory of evolution has been that, owing to the stability of germ-plasm since the first origin of sexual propagation, “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,” because “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[57].”

But about the beginning of the present year we have this fundamental doctrine directly contradicted in such words as:—

The origin of a variation is equally independent of selection and amphimixis, and is due to the constant occurrence of slight inequalities of nutrition in the germ-plasm[58].

This complete reversal of his previous doctrine brings Weismann into line with Darwin, who long ago gave very good reasons for the following conclusion:—

Those authors who, like Pallas, attribute all variability to the crossing either of distinct races, or to distinct individuals belonging to the same race but somewhat different from each other, are in error; as are those authors who attribute all variability to the mere act of sexual union [amphimixis][59].

And again:—

These several considerations alone render it probable that variability of every kind is directly or indirectly caused by changed conditions of life. Or, to put it under another point of view, if it were possible to expose all the individuals of a species during many generations to absolutely uniform conditions of life, there would be no variability[60].

Hence, Darwin was disposed to find the main, if not the only, causes of congenital variations in circumstances depending for their efficacy on the instability of what Weismann calls germ-plasm. And the noteworthy fact is, that Weismann has now adopted this view, to the destruction of his originally fundamental postulate touching the stability of germ-plasm since the first origin of sexual propagation.

By such a right-about-face manoeuvre, Weismann has placed his critics in a somewhat difficult position. For, in the first place, it is only towards the close of The Germ-plasm that the manoeuvre is executed, and then only in a few sentences such as I have just quoted—italicized, it is true, but otherwise so slightly emphasized that, as Professor Hartog has observed, no one of his reviewers has noticed it[61]. In the second place, he nowhere expressly recognizes the effects upon his theory of evolution, which necessarily follow from the change. And, lastly, the manner in which he endeavours to underpin that theory after having thus removed its logical foundation in his former postulate of the absolute stability of germ-plasm, is so peculiar that it is hard to epitomize his reasoning with due regard to brevity.

Speaking for myself, I can only say that my first impulse, after reading the sentences above quoted, was to cancel the whole of Chapter IV, as well as all those parts of Chapters I and III where the Weismannian theory of evolution is alluded to; and then to start anew with a bare statement that this theory had now been wholly discarded by its author. But after due consideration it seemed desirable to leave the criticism as it was originally written, not only on account of the reasons already stated in the Preface, but still more because I found it would be impracticable to start a new criticism of the greatly modified theory of evolution without introducing many and lengthy parts of the old one, for the purpose of showing how the most recent theory had been arrived at. Hence, seeing that my previous criticism was far from having been rendered obsolete by the large changes which had taken place in Weismann’s system of theories, I concluded that it was best to retain what I had written, and add the present paragraphs for the purpose of dealing exclusively with the changes in question.

In now proceeding to do this, I think it is needless to occupy space by giving the reasons which have caused Weismann thus to abandon his doctrine of the universal stability of germ-plasm since the first origin of sexual propagation, and to substitute the precisely opposite doctrine of its universal instability. It is enough to say that these reasons all arise by way of logical necessity from the further working out in The Germ-plasm of his theory of heredity—or, more correctly, from the additions which he has there made to his previous views on the mechanism of heredity. Thus he has reversed his former doctrine touching the absolute stability of germ-plasm, not so much on account of any of the general considerations or particular facts which I have adduced against it in Chapter IV, as because it would not tally with the recent additions which he has made to other parts of his system. Any one who cares to follow this matter will find the reasons in question fully and lucidly stated in Chapter XIV of The Germ-plasm[62].

It is of more importance to consider the means whereby Weismann seeks to save his theory of evolution after he has thus removed its foundation in his former postulate of the absolute stability of germ-plasm. As far as I can understand, he seeks to do so as follows.

In the first place, it must be noted that after his changes of view with regard to polar bodies, unicellular organisms, and the significance of sexual reproduction, nothing remains of his original theory of evolution save what he can manage to retain of his original theory of variation as due to amphimixis. But, as we have just seen, he has surrendered this latter theory also. Therefore, at first sight it appears that no part of the former can possibly remain. Beginning at the apex, he has removed, stone by stone, his doctrine of descent, and, on arriving at its fundamental postulate—the absolute stability of germ-plasm—simply turns it upside down. Surely, therefore, it may be thought, there is here as complete a destruction as well could be of all this side of Weismann’s system. Such, however, he endeavours to show is not the case. He regards it as still possible to retain so much of his theory of descent as is presented by what he can save of his theory of variation, thus:—

Although he now represents that the instability of germ-plasm is such that in no case can amphimixis have had anything to do with the origin of congenital variations, he continues to regard the stability of germ-plasm sufficiently great to necessitate, in all cases, the occurrence of amphimixis in order to promote the development of congenital variations. In other words, notwithstanding that he now thinks all congenital variations must be begun by external conditions acting directly on an unstable germ-plasm, he also thinks that the amount of variation thus produced is likely to be exceedingly minute, and must therefore be increased by subsequent amphimixis in order to fall within the range of natural selection. So that, although powerless to initiate congenital variation, amphimixis must still play an indispensable part in the process of evolution, as in all cases a necessary condition to the occurrence of natural selection. External conditions first cause slight changes in the determinants of a species; but these are so slight that they have to be augmented by amphimixis before they constitute material on which natural selection can act, and hence before they can become of any significance either in ontogeny or phylogeny.

Such, I take it, is what Professor Weismann would now have us to understand; for otherwise I should have expected from him as frank a surrender of his theory of evolution (or the remnant thereof in his theory of variation) as he has made of its fundamental postulate. But, if such is his meaning, I may mention the reasons which appear to me to render it nugatory.

In the first place, it is evident that in thus minimizing the possible range of congenital variation due to the action of external conditions on a non-absolutely stable substance of heredity, Weismann is making a wholly gratuitous assumption, for the sole purpose of saving what remains of his theory of evolution—i. e., the doctrine of the immense importance of amphimixis.

We have already seen in the foregoing chapter that his original assumption of the absolute stability of germ-plasm was a gratuitous one, made for the purpose of supplying a foundation for constructing his theory of evolution. But still more gratuitous is the assumption which he has now substituted, for the purpose of saving as much of this theory as is left—the assumption, namely, that germ-plasm, although universally unstable, nevertheless everywhere presents only a certain low degree of instability, which serves to accommodate his modified theory of heredity on the one hand, and all that is possible of his previous theory of evolution on the other. His original assumption, untenable though it was, furnished at least a logical basis for the necessary conclusion that amphimixis was the only possible cause of congenital variations. But there is not so much as any logical sequence in the now substituted assumption, that (A) all congenital variations are ultimately due to the universal instability of germ-plasm, and (B) that nevertheless they are all more proximately due to such a high degree of stability of germ-plasm as necessitates amphimixis as the only means whereby variations can be made “perceptible.” These statements are as independent of one another as any two statements can well be; and, therefore, if the second of them is to be substantiated, it can only be so by some totally distinct line of reasoning. The first statement does not even tend to suggest the second; in fact it tends to suggest the precise contrary. For, obviously, there is nothing in the logic of the matter to show why, if all congenital variations depend for their origin on the instability of germ-plasm, such instability must nevertheless be always so slight that the variations due to it must afterwards depend on amphimixis for their development to the point where they become “perceptible.” As above indicated, it is surely little short of absurd thus to assume that a universally unstable germ-plasm universally presents only that particular degree of instability which will serve to accommodate Professor Weismann’s newer theory of heredity, and at the same time to save thus much of his previous theory of evolution.

But now, in the second place, not only is this assumption wholly gratuitous, but there are many considerations which render it in the highest degree improbable, while there are not wanting facts which appear to demonstrate that it is false. For, unquestionably, most of the considerations which have already been advanced in the preceding chapter against the assumption of an absolute stability of germ-plasm, are here equally available against the assumption of an imperceptibly small amount of instability[63]. Similarly, all the facts there given with regard to the a-sexual origin of species—and even genera—of parthenogenetic organisms, bud-variation[64], &c., amply demonstrate that congenital variations due to the instability of germ-plasm alone, or apart from amphimixis, are sometimes enormous. Hence, we cannot accept the gratuitous suggestion that in all other cases they are too insignificant to count for anything till they have been augmented by amphimixis, even although we may be prepared to agree that amphimixis is probably one important factor in the production of congenital variations. What degree of importance it presents in this connexion, however, we have not at present any means of determining; all we can conclude with certainty is, that in some cases it is demonstrably very much less than Weismann supposes, while it is extremely improbable that it is ever in any case the sole and necessary antecedent to the operation of natural selection.

This extreme improbability is shown, not only by what I have already said in the previous chapter, and need not here repeat; but likewise by the “several considerations” which Darwin has adduced with regard to this very point, and which, as he says, “alone render it probable that variability of every kind is directly or indirectly caused by changed conditions of life,” with the consequence that “those authors who attribute all variability to the mere act of sexual union are in error.” I have already quoted these words further back in the present chapter, in order to show that by now attributing the origin of all congenital variations to the direct action of external conditions, Weismann has brought himself into line with Darwin so far as this fundamental point of doctrine is concerned. But I here re-quote the words in order to show that by further attributing the development of congenital variations “to the mere act of sexual union,” Weismann is again falling out of line with Darwin. So to speak, he first performs a right-about-face movement as regards his original position towards the “stability of germ-plasm,” and immediately afterwards makes a half-turn back again. Now, it is this half-turn to which I object as unwarranted in logic and opposed to fact.

In a previous chapter (pp. 66-7) I presented to him the dilemma, that germ-plasm must be either absolutely stable or else but highly stable, and that in the former case his theory of amphimixis as the sole cause of congenital variations would be valid, while in the latter case the theory would collapse. But it did not then occur to me that Weismann might seek a narrow seat between the horns of this dilemma, by representing that germ-plasm is universally unstable up to a certain very low degree of instability—viz., exactly that degree which is required for starting a congenital variation by means of external causes, without its being possible for the variation to become perceptible unless afterwards increased by means of amphimixis. And now that this extremely sophistical position has been adopted, I cannot see any imaginable reason for adopting it other than a last endeavour to save as much as possible of his former theory of evolution. There can be nothing in the nature of things thus to limit, within the narrowest possible range, the instability of a universally unstable germ-plasm—distributed, as this most complex of known substances is, throughout all species of plants and animals, and exposed to inconceivably varied conditions of life in all quarters of the globe. And these considerations are surely of themselves enough to dispose of the assumption as absurd, without again rehearsing the facts of congenital variation which definitely prove it to be false.

Conclusion.

For reasons stated at the commencement of this chapter, I have restricted its subject-matter almost exclusively to a consideration of the more fundamental changes which Professor Weismann has wrought in his general system of theories by the publication of his most recent works. In other words, I have purposely avoided considering those immensely elaborate additions to his theory of heredity which constitute by far the largest portion of his essays on Amphimixis and The Germ-plasm, and which have for their object an ideal construction of “the architecture of germ-plasm.”

The fundamental changes to which allusion has just been made are as follows.

Professor Weismann has to a large extent abandoned his theory of polar bodies, and in my opinion would have done well had he taken a further step and surrendered the theory in toto.

Similarly, he has withdrawn his previous distinctions between the unicellular and multicellular organisms. The Protozoa and Protophyta are now included by him in the same category as the Metazoa and Metaphyta, as regards all matters of individual variation, reproduction, subjection to the law of natural selection, and so forth. The only difference which he continues to allege is the somewhat metaphysical one touching mortality and immortality. But I have given what appears to me sufficiently good reasons for ignoring this distinction; and therefore, as it seems to me, every one of Weismann’s previous doctrines respecting unicellular organisms have vanished—very much to the benefit of his system as a whole.

By far the greatest change, however, which he has made in this general system is that which he has effected by surrendering the postulate of the absolute stability of germ-plasm. The rift in his lute which has been noticed with regard to this matter has now been widened to an extent which does prevent any further harping on the theme of evolution. It is true that Weismann endeavours to retain as far as possible the general character of his former postulate of the universal stability of germ-plasm, with the consequent “significance of sexual reproduction” as the sole cause of congenital variation. For although he now reverses both these doctrines by saying that germ-plasm is universally unstable, and that sexual reproduction is in no case the sole cause of congenital variation, he seeks at the same time to minimize the logical consequences of such reversal by making an ingenious assumption, the possibility of which I had not foreseen when writing the previous chapters. The assumption is, that although germ-plasm is universally unstable, the degree of its instability is everywhere restricted within the narrowest possible limits; so that sexual propagation is still necessary for the purpose of developing congenital variations to the point where they can fall within the range of natural selection, notwithstanding that they must all have been originated by external causes acting directly on a germ-plasm universally unstable within the narrow limits assumed. But clearly this assumption is arbitrary to the last degree, and, no less clearly, it is made by Weismann for the sole purpose of saving as much as he can of his previous theory of variation. His more recent speculations touching the mechanism of heredity are incompatible with his former view of amphimixis as the sole cause of congenital variations, and therefore he makes this arbitrary assumption for the purpose of representing that amphimixis may nevertheless still be regarded as a necessary con-cause. I need not here repeat what has so recently been said touching the sophistry of this assumption in theory, or the demonstrable falsity of it in fact. It is enough to remark, in conclusion, that the game is not worth the candle. It was originally well worth Weismann’s while to sustain his fundamental postulate of the absolute stability of germ-plasm, because he was able to rear upon it his whole theory of evolution. But the only part of this theory which he has now left standing, or which he can now save by his newer postulate of a germ-plasm both stable and unstable at the same time, is his doctrine of variation. So to speak, it is his desire to reserve as much as is speculatively possible from the general ruin of his theory of descent, that causes him to go so far to attempt so little. For I cannot suppose that he himself will expect any of his readers to entertain so arbitrary, fanciful, and demonstrably false an assumption as the one in question. Surely it would have been better to have surrendered in toto this “Weismannian theory of variation,” rather than to have attempted its rescue by means so plainly nugatory. It might still have been held that amphimixis plays a large and important part as one of the causes of variation, and therefore also as one of the factors of organic evolution. After having reversed his postulate of amphimixis being the sole cause of variability, and therefore having agreed with Darwin that “those writers are in error who attribute all variability to the mere act of sexual union,” he might well have questioned Darwin’s further statement as to its being “probable that variability of every kind is directly or indirectly caused by changed conditions of life.” But by now assuming that variations due to any causes other than amphimixis must be “imperceptible” until they have been augmented by amphimixis, Weismann is shutting out, with a futile hypothesis, the important question as to whether, or how far, amphimixis really is a cause of variation. Observe, the case is not as it might have been were there no reasons assignable for the occurrence of sexual propagation, other than that of assisting in the production of congenital variations. The theory of “rejuvenescence,” for example, is prima facie a more probable one than that which ascribes to sexual propagation the function of causing variability[65]; while Galton’s hypothesis, which supposes the object of this form of propagation to be that of conserving the “germs” (= “determinants”) of the phyla, has a good deal to say for itself[66]. Of course such alternative hypotheses touching “the significance of sexual reproduction” are not necessarily exclusive of one another: the process may subserve two or more adaptive purposes[67]. But he would be a bold man who, in the present state of our knowledge, could accept unreservedly the particular view of this process which Darwin so emphatically rejected; and I think he must be a biased man who could entertain for an instant the modification of this view which Weismann has now substituted.

Thus, the Weismannian theory of evolution has entirely fallen to pieces with the removal of its fundamental postulate—the absolute stability of germ-plasm. It only remains to mention once more the effects of this removal upon the other side of his system—viz., the companion postulate of the uninterrupted continuity of germ-plasm, with its superstructure in his theory of heredity.

Briefly, these effects are as follows:—

1. Germ-plasm ceases to be continuous in the sense of having borne a perpetual record of congenital variations from the first origin of sexual propagation.

2. On the contrary, as all such variations have been originated by the direct action of external conditions, the continuity of germ-plasm in this sense has been interrupted at the commencement of every inherited change during the phylogeny of all plants and animals, unicellular as well as multicellular.

3. But germ-plasm remains continuous in the restricted, though still highly important sense, of being the sole repository of hereditary characters of each successive generation, so that acquired characters can never have been transmitted to progeny “representatively,” even although they have frequently caused those “specialized” changes in the structure of germ-plasm which, as we have seen, must certainly have been of considerable importance in the history of organic evolution.

4. By surrendering his doctrine of the absolute stability of germ-plasm on the one hand, and of its perpetual[68] continuity on the other, Weismann has greatly improved his theory of heredity. For, whatever may be thought of his recent additions to this theory in the way of elaborate speculation touching the ultimate mechanism of heredity, it is a great gain to have freed his fundamental postulate of the continuity of germ-plasm from the two further postulates which have just been mentioned, and the sole purpose of which was to provide a basis for his untenable theory of evolution.

5. In my opinion it only remains for him to withdraw the last remnant of his theory of evolution by cancelling his modified and even less tenable views on amphimixis, in order to give us a theory of heredity which is at once logically intact and biologically probable.

6. The theory of germ-plasm would then resemble that of stirp in all points of fundamental importance, save that while the latter leaves the question open as to whether acquired characters are ever inherited in any degree, the former would dogmatically close it, chiefly on the grounds which I have considered in Appendix II. It seems to me that in the present state of our knowledge it is more prudent to follow Galton in suspending our judgement with regard to this question, until time shall have been allowed for answering it by the inductive methods of observation and experiment.

7. Hence, in conclusion, we have for the present only to repeat what Weismann himself has said in one of the wisest of his utterances,—“The question as to the inheritance of acquired characters remains, whether the theory of germ-plasm be accepted or rejected.”

It is now close upon twenty years that I accepted the substance of this theory under the name of stirp; and since that time the question as to the inheritance of acquired characters remains exactly where it was. No new facts, and no new considerations of much importance, have been forthcoming to assist us in answering it. Therefore, as already stated in the Preface, I intend to deal with this question hereafter as a question per se, or one which is not specially associated with the labours of Professor Weismann.