ON GERM-PLASM.

As already stated in the text (p. 71), Weismann’s general reasoning in support of his own theory of germ-plasm, as against Darwin’s theory of gemmules in any form, admits of being reduced to arguments in favour of three propositions—viz., first, that there is no evidence of the transmission of somatogenetic characters; secondly, that the theory of pangenesis, which seeks to explain their supposed transmission, is “inconceivable”; and, thirdly, that its logical antithesis—the theory of germ-plasm—is so much less beset with difficulties, that by comparison it is simple, self-coherent, and offers a real, as distinguished from a “formal,” explanation of the facts of heredity.

The first of these propositions will be discussed at considerable length in my next volume. The second and third propositions, however, may be dealt with here.

The following paragraph, which I shall quote sentence by sentence, sets forth the grounds on which Weismann bases the second proposition, namely, that any theory belonging to the order of pangenesis—i. e., which supposes the carriers of heredity ever to travel centripetally—is, from its very nature, inconceivable.

At first sight this hypothesis seems to be quite reasonable. It is not only conceivable that particles might proceed from the somatic to the reproductive cells, but the very nutrition of the latter at the expense of the former is a demonstration that such a passage actually takes place. But a closer examination reveals immense difficulties. In the first place, the molecules of the body devoured are never simply added to those of the feeding individual without undergoing any change, but, as far as we know, they are really assimilated, that is, converted into the molecules of the latter. We cannot therefore gain much by assuming that a number of molecules can pass from the growing somatic cells into the growing reproductive cells, and can be deposited unchanged in the latter, so that, at their next division, the molecules are separated to become the somatic cells of the following generation[69].

The obvious answer to this is, that no one has ever supposed “gemmules” to be merely “molecules,” in the chemical sense of this word; nor has any one ever imagined that they are “devoured” by the germ-cells into which they pass. Of course, if this were the case—i.e., if gemmules serve merely as food to the germ-cells—they would become disintegrated down even to their chemically molecular structure, and there would be an end of them as organized “carriers of heredity.”

In the second place, it is asked:—

How can such a process [i.e. the passage of gemmules into growing germ-cells] be conceivable, when the colony becomes more complex, when the number of somatic cells becomes so large that they surround the reproductive cells with many layers, and when at the same time, by an increasing division of labour, a great number of different tissues and cells are produced, all of which must originate de novo from a single reproductive cell?

Here, again, the obvious answer is, that no one has ever propounded such a statement. Far from supposing that “all the different cells and tissues of a complex organism must originate de novo from a single reproductive cell,” the theory of pangenesis supposes the very contrary—viz., that somatic changes in the past history of the phyla have not thus originated in any reproductive cell. The idea of somatic changes originating in reproductive cells belongs to the theory of germ-plasm; but even this theory does not suppose all the great number of different cells and tissues which compose a complex organism to have ever originated de novo from a single reproductive cell.

The difficulty touching germ-cells becoming isolated, or buried, by the phylogenetic increase of somatic cells, is enforced in the immediately succeeding sentences, thus:—

Each of these various elements [somatic cells] must, ex hypothesi, give up certain molecules to the reproductive cells; hence those which are in immediate contact with the latter would obviously possess an advantage over those which are more remote. If, then, any somatic cell must send the same number of molecules to each reproductive cell[70], we are compelled to suspend all known physical and physiological conceptions, and must make the entirely gratuitous assumption of an affinity on the part of the molecules for the reproductive cells. Even if we admit the existence of this affinity, its origin and means of control remain perfectly unintelligible if we suppose that it has arisen from differentiation of the complete colony. An unknown controlling force must be added to this mysterious arrangement, in order to marshal the molecules which enter the reproductive cell in such a manner that their arrangement corresponds with the order in which they must emerge as cells at a later period.

Now I do not see much force in the suggestion that those somatic cells which happen to be in immediate contact with germ-cells, “must obviously possess an advantage over those which are more remote.” On the contrary, I do not see that mere proximity of one species of cell to another species within the same organism need have anything to do with the matter—still less that “we must suspend all physical and physiological conceptions,” if we demur to the statement that it “obviously must.” As for “physical conceptions,” how many thousands of cases might not be pointed to among chemical and mechanical processes where contact or proximity are conditions of comparatively little importance? And as for “physiological conceptions,” do we find that any part of the organism is affected by its distance, say, from the liver and kidneys, for getting rid of its effete products? Is it not rather the case that every gland in the body is wholly unaffected by its distance from any part of the body, in regard to its function of draining off the particular substances with which it is concerned? Why then should the reproductive gland constitute a conspicuous exception? Or how do we suspend all physiological conceptions, if we suppose that this gland resembles every other gland in being specialized to secrete a particular kind of “molecule,” which, because thus specially selected, may be said to have for that gland a special “affinity”? If there are such things as gemmules, I do not see any violation of physiological analogies—still less an “entirely gratuitous assumption”—in supposing that they can be filtered out from all parts of the body by the sexual glands, and there aggregated as a special product to be discharged in the form of sexual elements[71].

But, it is further represented, “even if we admit the existence of this affinity, an unknown controlling force must be added to this mysterious arrangement, in order to marshal the molecules which enter the [growing] reproductive cell in such a manner that their arrangement corresponds with the order in which they emerge as cells at a later period.” Surely, however, for Weismann of all naturalists it ought not to be difficult to find this “unknown controlling force.” For of all naturalists he is perhaps the most ready to invoke the agency of natural selection as sufficient to explain every case—actual or imaginable—of adaptation. Now, here is a case where natural selection, one would think, is positively bound to act—supposing that there be such things as gemmules. For, if “the carriers of heredity” are gemmules, it is evident that their mutual “affinities” must be adaptively “marshalled” at each step of phylogenetic evolution, before any further advance of such evolution can be possible. And I do not see anything more “inconceivable” in supposing the establishment of such mutual affinities step by step through natural selection, than in supposing any other course of adaptive development by similar means. For, as Darwin has well shown, while anticipating this particular objection to his theory,—“The assumed elective affinity of each gemmule for that particular cell which precedes it in due order of development is supported by many analogies.” The analogies which he then gives are so numerous that I must here refer to his own discussion of the subject[72]—a discussion which is entirely ignored by Weismann.

Lastly, the principal ground, as far as I can see, which Weismann has for regarding Darwin’s theory in any shape “inconceivable,” is his own supposition that there is as complete an anatomical separation between the soma and its germ-cells as there is, for example, between the mammalian soma and these same cells when afterwards detached from the ovary and developing as foetuses in utero. In other words, the only connexion is supposed to be that of deriving nourishment by way of imbibition. But, as regards the germ-cell while still forming in the ovary or testicle, there is for this supposition no basis in fact. There is nothing in the histology of spermatogenesis that lends countenance to the supposition, while in the case of the ovum such histological evidence as we possess makes altogether against it. As Professor Vines has remarked:—

It cannot be seriously maintained that the whole body of the embryo is developed solely from the germ-plasm of the ovum. On the contrary, since the embryo is developed from the whole of the nucleus and more or less of the cytoplasm of the ovum, it must be admitted that the non-germ-plasm of the ovum provides a large part of the material in embryogeny. It is an obvious inference that, under these circumstances, hereditary characters may be transmitted from the parent to the offspring, not only by the germ-plasm, but also by the somato-plasm, of the ovum[73].

Again, and apart from this consideration, it is now known that a very intimate network of protoplasmic fibres connects the cell-contents of cellular tissues, both in plants and animals. So that here we have another very possible means of communication between the germ-cells and the somatic-cells which together constitute a multicellular organism.

Therefore, in so far as histology can be trusted to constitute a basis for generalizations of this kind at all, it does not sustain the supposition that there can be no medium of communication between the general cellular tissues of an organism and its specially reproductive elements. On the contrary, the microscope is able to demonstrate possible roads of connexion—and this even upon Weismann’s own view as to a specialized germinal substance which is restricted to the nucleus of an ovum. In short, the supposition as to an absolute anatomical separation between germ-plasm and somato-plasm is a deduction from Weismann’s theory itself: it is not supported—it is discredited—by histological observation. Hence, it cannot be accepted as valid evidence in favour of the theory from which alone it is derived, or as a valid objection to the rival theory of pangenesis.

Once more, even if it were true that histology proves an absolute anatomical isolation on the part of germ-cells, it would still have remained unquestionable that there is no absolute physiological isolation. For, at least, the germ-plasm derives its nourishment from the soma in which it resides; and who shall say that the process of mere imbibition is not amply sufficient to admit of the passage of “gemmules”? Call them what we choose, the “carriers of heredity” must be so unimaginably small, that in relation to histological cells they must be as gnats to camels. Yet we know that even camels in the form of “migrating cells” of various kinds are able to pass through living membranes; and we also know that the microbes of syphilis can penetrate both ova and spermatozoa. Why then should it be deemed inconceivable that, where all such things can pass, gemmules can do so likewise?

Lastly, I have recently spoken of the detached condition of a ripe ovum in utero. Now it seems to me more “inconceivable” that such an ovum should be capable of announcing, as it were, to the walls of the uterus whether or not it is in a fertilized condition, than it is that, before quitting the ovary, it should have had some kind of physiological converse with its environing soma. Yet it is certain that, without any visible medium of communication, the impregnated ovum is able to inform the uterus that it is impregnated; and thereupon the uterus behaves towards that ovum in an altogether astonishing manner, such as it never displays towards an unimpregnated ovum. Of course various hypotheses may now be formed to account for this fact, seeing that no one can question it as a fact. But supposing that the fact could be questioned, with how much greater effect might it be argued that any communication between the ovum and its soma is even more antecedently incredible when the ovum is entirely free than when it is still contained within its ovary.

Now these, as far as I can find, are the only grounds for Weismann’s repeated assertion that the theory of pangenesis in any form is “inconceivable.” I have therefore endeavoured to show that this is too strong a statement. All the facts and considerations whereby he seeks to support it were present to the mind of Darwin; and, quite apart from any question of relative authority, I cannot avoid agreeing with Darwin that, whether or not the theory is true, at all events the “difficulties” attaching to it on these merely a priori grounds are not insuperable, or such as to render his “pet child” an unconceived monstrosity in logic, or a proved absurdity in science.

Be it understood, however, that I am not here defending the theory of pangenesis. I am investigating the theory of germ-plasm; and it is because Weismann seeks to sustain the latter by excluding the former as preposterous, that I have been obliged thus to consider the validity of his criticism. For the point to which I am leading is, that Weismann gains nothing in the way of support to his own theory by this disparagement of Darwin’s, unless he can show that the former supplies some more “conceivable” explanation touching the mechanism of heredity. Now I am unable to see that he has shown this. What I do see is that his a priori argument from “inconceivability” cuts both ways, and that it makes at least as much against germ-plasm as it does against gemmules. Therefore, having now considered what Weismann has said against the conceivability of gemmules on grounds of general reasoning, I shall proceed to show that quite as much—or even more—may be said in the way of a tu quoque. In other words, we have now finished with the second of the three propositions which we are examining (see p. 71), and proceed to our consideration of the third.

First of all, I do not see any greater difficulty in supposing that the “carriers of heredity” proceed centripetally from somatic-cells to germ-cells, than in supposing that they proceed centrifugally from the germ-cells to the somatic-cells which they are engaged in constructing. Nor do I see any more difficulty in imagining these “carriers of heredity” to be capable of constructing a new organism if they have first proceeded centripetally, and are thus severally representative of all parts of the parent organism after its construction has been completed, than I do if they have proceeded centrifugally, and are thus similarly representative of all parts of that organism before its construction has been commenced[74].

Similarly, it seems to me, whatever cogency there may be in Weismann’s objection to Darwin’s theory on the score that it must assume “an unknown controlling force in order to marshal the molecules,” is equally great as regards his own. True, Weismann has a lot to say about the control which nucleo-plasm can exercise on cell-formation, and germ-plasm on marshalling successive stages of ontogeny; but all that this amounts to is a re-statement of the facts. Such a controlling force must be equally assumed by both theories; but in each alike there is an absence of any ghost of an explanation.

Again, whatever difficulty there may be in conceiving the transition of somatic substance, mutatis mutandis there must be an equal difficulty in conceiving the transition of germinal substance into somatic substance. Indeed, as far as I can see, the difficulty is even greater in the latter case than it is in the former. For the very essence of Weismann’s view is that germ-plasm differs from all or any other “plasm” in origin or kind: germ-plasm, and germ-plasm alone, has been immortal, perpetually continuous, capable of indefinite self-multiplication, and so of differentiating itself into an endless number and variety of somatic tissues. But, according to Darwin’s view, there is not, and never has been, any such fundamental difference between the essential nature of somatic elements, and the essential nature of sexual elements. On the contrary, it is supposed that both formative and formed material are one in kind—that all the cellular tissues of a multicellular organism, like the single cell of a unicellular organism, are per se endowed with the vital property of self-multiplication; and that whether this property finds its expression in normal growth, in abnormal increments of growth (such as tumours), in processes of repair, in the various forms of a-sexual reproduction, or in the more specialized form of sexual fertilization, there is everywhere an exhibition of one and the same capacity. Now, without going further than this contrast between the fundamental principles of the two theories, does it not become evident that the difficulty of conceiving a transition of A into A´ is at any rate no greater than that of conceiving a transition of A into B, where A is in both cases the formative substance, A´ this same substance in another stage of evolution (i.e., elaborated for the performance of some special function, but never so as to lose its original function A), while B is a substance which differs from A almost as much as a woven texture differs from the hands that weave it?

Once more, in all his arguments which are directed to prove the continuity of germ-plasm, Weismann nowhere seems to perceive the necessity of arguing the correlative hypothesis—viz., that of the discontinuity of somato-plasm. Yet, as Professor Vines has remarked, it is as incumbent on him to disprove any possible continuity on the part of somato-plasm, as it is to prove a perpetual continuity on the part of germ-plasm. And here I am disposed to go further than Professor Vines has gone; for it appears to me even more incumbent on Weismann to argue a discontinuity on the part of somato-plasm, than it is on him to argue a continuity on the part of germ-plasm.

This must be immediately apparent if we remember that, unless the discontinuity of somato-plasm be assumed, the theory of the continuity of germ-plasm in telluric time (as distinguished from eternity) becomes identical in form with all those theories of heredity to the family of which pangenesis belongs. All these theories go upon the assumption that living material has been continuous in telluric time—i.e., always derived from pre-existing material of the same kind; but they embody the further assumption that all living material is material of the same kind—i.e., everywhere presents the same fundamental properties. Weismann’s theory on the other hand, while adopting the first assumption, rejects the second; and assumes in its stead that living material exists in “two kinds,” only one of which has been continuous, while the other is discontinuous—being, in fact, formed anew at each ontogeny. Therefore, to my mind, it seems more needful to argue the point wherein his theory differs from these other theories of heredity, than it is to argue the point wherein it agrees with them. We look to him for a proof of the discontinuity of somato-plasm much more than we do for a proof of the continuity of germ-plasm. Now the only proof that he has to give of the discontinuity of somato-plasm—or, in other words, that the self-multiplication of somatic cells cannot take place unless the nucleus of each contains a self-multiplying idio-plasm derived from the nucleus of a germ-cell—is the non-transmissibility of somatogenetic characters. Here, however, there is an obvious equivoque. For his only test of characters as somatogenetic and blastogenetic consists in observing whether or not they are inherited: if they are inherited, he says they are blastogenetic: if they are not inherited, he says they are somatogenetic. But this is manifestly circular reasoning, so long as the question in debate is as to the truth of his theory. What we require in proof of the distinguishing feature of that theory—i.e., the discontinuity of the hypothetical somato-plasm—is not merely the obvious fact that some characters are inherited while others are not, but independent proof that inherited and non-inherited characters correspond to a continuity of germ-plasm on the one hand, and a discontinuity of somato-plasm on the other. He shows us, indeed, what was well known before, that characters developed during the lifetime of the individual are seldom (if ever) inherited, while characters developed during the lifetime of the species are always inherited. Obviously, however, this fact is no proof of the assumed correlation just mentioned, because, as Darwin has clearly pointed out, it may very well be due to the much shorter time which has been allowed for what may be termed the impress of heredity. Therefore, supposing (with Darwin and others) that living material is all of one kind, and continuous, the fact on which Weismann relies admits of being explained without resorting to his more complex supposition of living material in two kinds, the one perpetually continuous, and the other interrupted at each ontogeny.

For these reasons it appears to me that, so far as the argument from “inconceivability” is concerned, it makes at least as much against the theory of germ-plasm as it does against the theory of pangenesis; and, therefore, that no argumentative advantage is gained from its use by Weismann. The truth probably is that, whatever the mechanism of heredity may actually be, it is at once so minute and so complex that its action is “inconceivable,” or, more correctly, unimaginable. Be it again understood, therefore, that I am not arguing in favour of pangenesis. I am merely criticising what appears to me an unsound argument in favour of germ-plasm. All this general or merely a priori reasoning with regard to inconceivability is, as I have attempted to show, as available on the one side as on the other, and so fails to yield any observable advantage to either.

In conclusion it must be noticed, that Weismann now appears to have himself perceived the grave difficulties which lie against his antithesis between a hypothetical “germ-plasm” and a hypothetical “somato-plasm,” notwithstanding that the former becomes converted into the latter at each ontogeny. At any rate, he allows that Vines’ criticism upon this head is sound. But he is strongly of the opinion that, by means of a later emendation of his theory as originally published, he has succeeded in obviating these difficulties in toto. For my own part, as already several times observed in the text, I cannot in the least perceive that such is the case; and therefore I will quote in extenso what he has said in answer to Professor Vines. It will be seen that his newer emendation of the theory consists in substituting for his original “somato-plasm” two substances, which are called respectively “somatic idio-plasm” and “cytoplasm.” And it is by means of this substitution that he thinks he has, in some way or another, overcome the contradiction involved in the doctrine (and, as it still seems to me, the essential doctrine of his whole theory of heredity) that “germ-plasm” becomes converted into “somato-plasm” during the course of every ontogeny. The following, at any rate, is his latest utterance upon the subject:—

It may be dullness, but I confess that this does not appear to me an “answer” to Professor Vines’ criticism. Even though “idio-plasm of the first ontogenetic grade” has to become “idio-plasm of the second, third, fourth, hundredth, &c. grade,” before in each of the grades concerned it can give origin to the somatic-cells which are distinctive of that grade, I cannot see that it makes any difference (in relation to Vines’ criticism) whether we speak of those cells as containing “somato-plasm,” or as containing “somatic idio-plasm” of such and such a grade, plus “cytoplasm.” For whether we thus follow Weismann’s earlier terminology or his later, we are so far speaking about exactly the same thing, namely, the transformation of “germ-plasm” into all the constituent cells of the “soma.” The difficulty is, in Vines’ words above cited, “to conceive that the germ-plasm of the ovum can impress upon the somato-plasm of the developing embryo the hereditary characters of which it (the germ-plasm) is the bearer”; and Weismann says that this difficulty, which he acknowledges, can now be answered by substituting for his original statement that “germ-plasm” becomes changed into “somato-plasm,” the statement that it is “idio-plasm” derived from “germ-plasm” which thus “impresses its character on the cell containing it.” But, “as a matter of logical accuracy,” there is surely here a distinction without a difference. For what is the difference between saying that germ-plasm “impresses” its character on the contents of all somatic cells considered collectively under the term “somato-plasm,” and saying that every “ontogenetic grade” of germ-plasm “impresses” its character on each successive group of somatic cells considered severally under the term “idio-plasm” of such and such a grade? At best this newer terminology has reference merely to a superadded hypothesis touching the mode—or rather the history—of the transition in question: it does not affect the original and essential doctrine of the transition itself.