1. Introduction
Thirty years ago zoologists were mainly occupied with that comparative description of development which had aroused intense interest owing to the overwhelming stimulus of the theory of evolution. The analogy of the developing egg and the developing species from a single undifferentiated cell to a differentiated mass of cells was regarded as an even stronger support to that theory than the direct proof of palæontology itself. At that time not many realised how far zoologists were becoming committed to far-reaching speculations whose truth it was difficult, if not impossible, to test by actual observation. Owing largely to the genius of a few distinguished men a new science has arisen whose scope it is to analyse by experiment some at least of the actual factors which control the development of the organism, and to build up thereby a new conception of the organism itself. Two outstanding pioneers may be mentioned—Wilhelm Roux and Jacques Loeb, both of whom died in 1924.
Among the first problems to be attacked by the new science of developmental mechanics were those concerned with the process of, differentiation. Is differentiation caused and controlled by some inherent factor in the organism or is it impressed on an animal by various factors in the external environment? Is the organism a preformed mosaic or are the parts a function of their position in the whole ? Amongst the methods adopted to answer these questions those of removing parts of eggs or animals and of transplanting them into other positions have become well known. The discovery of post-generation in the frog’s egg by Roux, and of polar heteromorphosis in hydroids by Loeb had given a strong impetus to the new experiments on regeneration, whilst T. H. Morgan and the author of this summary had justified the belief that regeneration is to be regarded as a primary function of living matter just as is normal development.*
Since then, transplantation has been more extensively used for studying the fate of parts removed at any of the main periods of an animal’s life-history: (1) in the egg before or during segmentation, (2) in the embryo at early or late stages, (3) in the hatched larva, (4) in the metamorphosed or adult animal I will now try to summarise briefly those experiments which have been carried on from 1915 to 1924 by my own and other schools, and to draw the conclusions which I think are thereby justified.
2. Uni- and pluri-potent systems
The statement of Barfurth and Przibram that the regenerative faculties diminish in both ontogeny and phylogeny with differentiation and with age has now been widely accepted by biologists. The theoretical explanation of the empirical facts may be found in the progressive expansion of those regions which do not receive the full potentialities of the central area from which growth ensues, but only of those parts that are still to be developed from the secondary regions themselves. Thus the body of the developing animal is gradually divided up into smaller systems which may be equipotential in themselves (to use Driesch’s appropriate terminology) but which differ from each other in their potentialities. Heidenhain (24, 25) has called these particular systems “Teil-systeme,” and has given several examples without fully discussing the significance of symmetry. If we accept the theoretical conception I have just expounded, we may cut away step by step, as it were, the proximal toti- and pluri-potent regions of an animal’s body and arrive at last at unipotent systems, “simple Teil-systeme,” which are only able to produce one sort of differentiation. Let us cut out a small ring just behind the tentacles of the fresh-water hydroid Pelmaiohydra oligactis or a ring of similar size from its foot region and leave it without further treatment in water. It will soon perish without regenerating, whereas pieces from the middle of the animal will live on and develop a new hypostome with tentacles at the head end, and a new tail with adhesive glands at the aboral end. We thus get the impression that all pieces of this Hydra, which are capable of survival, are equi- and toti-potent. This, however, is not actually the case. D. R. R. Burt (5) succeeded in keeping these small rings of tissue from the head or foot region alive by transplanting them to longer pieces of Pebnatohydra by an ingenious modification of the threading technique. It was thus demonstrated that the head ring would always regenerate a head region (including hypostome, tentacles, etc.), but not a foot, and that the foot ring would always regenerate a foot but not a head region whatever the orientation of graft to stock may have been, The head region and the foot region are thus unipotent with regard to the regeneration of head or foot. It has long been known that the tentacles themselves are not able to produce a new hypostome or neck, but that in some species they have been observed to form a new tentacle at their basal cut end. Thus the tentacles are unipotent with regard to tentacle formation, while the head ring, which at first sight appears to be unipotent with regard to head formation, has in reality still got the potentiality of forming different organs, namely, hypostome and tentacles. It is also questionable whether the head and foot regions of this Hydra would prove unipotent with regard to head or foot formation if the experiments could be extended in time and modified so as to allow the middle part or the fixed end of the graft to regenerate once more. It has been demonstrated in planarians (Morgan, etc.) that the region in front of the eyes will form, at the cut end, a head directed posteriorly ; but, after some time, a tail will develop between the two heads, grow out at right angles to both of them. Finally two complete worms are formed (Lang (58), Keil (30)). This proves that the potentiality of body and tail formation is not really altogether lost in the normal head region of planarians, it is only much weaker than in the middle or tail region. In our interpretation we could say that the extrusion of peripheral parts has not been absolute, but that these still retain the potentialities of other regions.
3. Reversal of symmetry
By the method of transplantation it has also been shown that the tails of tadpoles are partipotent systems, regenerating tails at their proximal cut ends (Giardina, Roux’s Archiv., 1907). The articulate appendages of Crustacea and of insects (as well as the legs of vertebrates with their internal chains of bones) furnish excellent material for testing the progressive loss of potentiality with a peripheral extension. It has never been possible to obtain regeneration of limb regions shaped like those which normally lie proximally to a cut surface ; further, a monstrosity in which the same joint appears twice in the series has never been described. Only such joints as normally stand peripherally or distally to the cut surface are regenerated or are even found to protrude in a proximal direction from a broken limb. I was able to show that all the triple monstrosities, that have been described and compiled by W. Bateson (in his Materials for the Study of Variation, 1894), and by others, originate from such breaks, and I have termed them “Bruchdreifachbildung.” It is the proximally directed component of these triplications which proves the loss of regenerative faculties for forming more proximal regions ; otherwise the inherent polarity of formative substances (which I have termed “Schichtungspolarität”) ought to assert itself by producing proximal parts. It has also been possible in amphibians to prove the validity of this view by artificial breakage (Della Valle (58), Puppe (63)), or by transplantation of reversed extremities (Kurz (42), Græper(17, 18), Milojevié (48)). By means of these results the rules of symmetry emphasized by Bateson have been transformed into laws which systematize such empirical statements as that “the dorsal and ventral surfaces are never changed in development or regeneration after they have once been established” (Przibram, Roux’s Arckiv., 1910), and that “the anterior and posterior borders of an appendage keep on growing with unchanged qualities after transverse amputation.” But we have not yet been able to determine precisely the conditions under which a partly resected or split leg will produce mirror images of itself. It is certain that reversal by interchange of the anterior and posterior borders in one of the components may occur. This shows once more the greater stability of the dorsoventral as opposed to the anteroposterior differentiation. Once the two axes of the growing limb are determined the third axis will be fixed also, so that in the developing embryo the laterality is induced as soon as dorsoventral and anteroposterior axes are differentiated (Przibram (61), B. G. Smith (74)). The symmetry of right or left limbs will depend on the direction of growth (“Richtungspolarität”), and from a proximal cut surface an appendage of reversed laterality may thus originate. I think we can apply the same considerations to the painstaking experiments which Harrison (21-23) and his school (Detwiler (10-14), Swett (79), Nicholas (49-51)) have undertaken in transplanting the “Anlagen” of limbs in the Ambystoma embryo in the tail-bud stage. As I have tried to demonstrate elsewhere by models (62), there is no necessity to invoke an influence of the whole body on the polarity of the graft. The observed reversal of laterality in the limbs occurs in just those combinations and with just that direction of outgrowth which can be accounted for by regeneration of reversed distal parts from a proximal cut end. In accordance with our explanation Harrison obtained a mirrored limb when he simply cut away the half of the limb bud facing the head of the embryo ; this he calls the anterior limb half, but which is, according to Swett’s results, really the proximal half of the limb-anlage. Brandt (3) failed to get reversal in Triton at the tail-bud stage, but succeeded in earlier stages. He concluded that in this newt the corresponding stage of limb-anlage is reached sooner than in Ambystoma in accordance with the presumed higher phylogenetic development of the former. While this argument would seem plausible, we might just as well assume that the homology of the early stage in Triton with the later in Amblystoma extends to the circumstances of operation as well, and that the latter are responsible for the possibility of outgrowth from a proximal cut surface or a development of the graft in a distal direction. In the latter case of course no reversal would ensue. While I cannot agree with Harrison and his followers in the assumption of a reversal of the dorsoventral axis in the transplanted limb-anlagen (since ulnar and radial limb-borders may be interchangeable), I must concede (Goette(i6)) that my former conceptions of regeneration as a simple multiplication and re-arrangement of the cells in the tissues reaching to the wound surface cannot any longer be deemed satisfactory in view of the many new facts brought out by observations on “Bruchdreifachbildung” in lower organisms and by experiments on regeneration and transplantation in the limbs of amphibians. Huber-Pestalozzi (29) describes triplicities in protozoa conforming to the scheme of “Bruchdreifachbildung.” It is clear that in the case of unicellular animals an arrangement of cells cannot be made responsible for the phenomenon. Secondly, Hadzi(19) could not corroborate in his histological sections the turning round of cells at the proximal cut end of a wound in the coenosarc of Tubularia which I had inferred to exist from his preparations. Altogether the view is more and more gaining in strength that the importance of the cell as a unit has been greatly over-estimated, and that the cells play a greater rôle in subdividing living matter in the way necessary for restoration of nucleo-plasmatic equilibrium, than in morphogenesis. While retaining the conception of “Bruchdreifachbildung” outlined above, I must admit that my special application (1906, 1909) of Zur Strassen’s cell and nuclear movements is not a wholly adequate explanation of the facts.
4. Metaplasia and homœosis
A further and most urgent cause for not regarding old cells at the cut surface as responsible for regeneration of the distal parts is found in the experiments of Paul Weiss (89) who succeeded in inducing the regeneration of the lower arm and hand of Urodeles after the extraction of the humerus bone and transverse section of the limb at the elbow. As no humerus bone regenerated inside the upper arm in accordance with former experience of other authors (Philippeaux, Wendelstadt), it is clear that the bone or cartilaginous tissue itself is not necessary for the production of distal bony parts, since the ulna, radius, and hand bones were present in the regenerated lower arm and hand. It is not the tissue which is “actu” regenerating, but regeneration must arise from the potentiality immanent in the limb stump. These experiments have been confirmed independently by Bischler (2). It is probable that some other tissues also do not require the presence of their own kind for regeneration. As with the cells so also the significance of the tissue layers has been overrated. Nusbaum and Oxner (Roux’s Archiv., 1912) found that in nemertines the regenerating heads or tails do not arise by proliferation from all the tissues, but that entodermic and ectodermic cells alone may be involved. Let us remember that in the eyes of amphibians the regenerating lens may be derived from the iris, as demonstrated by Wolff and studied later by many other authors (for a review, see Ubisch 8). In connection with these studies the question has been debated whether in embryogeny the lens arises by contact of the eye-cup with the skin or by self-differentiation of the latter. Different stages, species, and technique have given different results. Most biologists are inclined to believe in a gradual disappearance of a specific eye - cup influence with individual age and with higher phylogenetic development, and to attach more importance to a gradual capacity for self-differentiation by the skin over the eye-bud. From our viewpoint this would correspond to the extrusion of a partipotent system with total extinction of its potentialities in the proximal parts, that is eye-cup, and all the other distal parts, as is shown by the fact that skin from another body region grafted over the eye-bud fails in certain species to be converted into a lens.
The disappearance of potentialities with the removal of partipotent systems which have been extruded without loss of potentialities for proximal parts is clearly demonstrated in another class of phenomenon appropriately designated by Bateson as “Homceosis,” or the replacement of one appendage by another normally belonging to a different segment of the same animal. I have proved by comparison and experiment that these homceotic appendages originate by regeneration not only in Crustacea but also in insects (Przibram (53, 55-57)). One of the best examples is the antenna of the stick-insect Carausias (= Dixippus). morosus, as was first observed by Schmit-Jensen (67). We have analysed the conditions and found that an antenna is replaced by a front leg when the former is cut off near its base, but by another antenna when only the flagellum has been removed (L. Brecher (4)).* Other cases are similar. The alleged necessity of the nervous centre in the eye of the stalk-eyed crayfish (Herbst, Morgan, Steele) does not seem to be borne out by experiments in other forms. For example, the regeneration of the eye in snails is independent of the ganglionic mass (Nonne). In all cases, however, the stage of development is a factor not to be neglected. Chantran, who was the first to obtain antennulæ instead of eyes in the crayfish, mentioned that this was only the case in older specimens since young animals regenerated eyes. The same is true with homœosis in insects. Increasing age at the time of operation makes the homœosis more complete and normal regeneration becomes less obvious. In intermediate stages the loss of the potentiality for forming the normal appendage may not be permanent, as is shown in experiments on the antenna of Mantids (Przibram (53)). When in the early larval stages the antenna is cut off near the base a foot-like termination first appears, but later on this is replaced by a normal antenna. Only in the last larval stages do leg-like appendages remain permanently, as is, of course, always the case as soon as the imaginai instar is reached, as, for example, in operations on the nymph of Mantids or the last larval stage of Cimbex (Przibram (57)).
5. Induction of form
By experiments on transplanting “buds” or later “anlagen” of embryonic organs or regenerating blastemas most important conclusions have been reached as to the time at which the self-differentiating stage of the different parts or organs occurs. When extremities of amphibians are cut off at stages which are still capable of regeneration and the formation of conical regeneration buds is then waited for, it is found that these may be grafted to another place, eg. the regeneration bud grown on a foreleg can be grafted to a fresh-cut stump of the hindleg. In the case of young buds the development which occurs in the new position is different from what it would have been had the buds been left in their place of origin : the appendage will assume the character of the part to which it has been transferred (Schaxel (65, 66), Milojević (48)). If, on the contrary, one waits longer before taking the bud so that differentiation has begun to set in (the bud developing to an anlage) and this anlage is then grafted to another place, it is found to develop on its own lines and does not conform as before to its new situation. Its form is “herkunftsgemäss,” not “standortsgemäss.” We must therefore suppose that the material formed at the cut surface is at first undifferentiated, and that the stump later on impresses its polarities on to this mass. As soon as this impression has been received, however, the anlage proceeds by self-differentiation. I do not see that the facts brought forward so far make it probable that the dorsoventral and anteroposterior axes are impressed at different stages as has been suggested (Gräper (18)). In complete agreement with the results of regenerative differentiation are those from early embryonic development obtained by Spemann’s (75-77) technique of grafting (H. Mangold, O. Mangold (46,47), Pröscholdt, Geinitz, Ruud, Sternberg (78, etc.)). Not only was it possible to produce secondary medullary folds by implanting a bit of the upper lip of the blastopore into the side of Triton embryos, but also to induce the entoderm to produce such structures as would never have been produced had no such “organiser” of a later developmental stage been embedded in it. Furthermore, the structure induced by an organiser is capable of again inducing surrounding areas in a germ not further advanced than itself. The organiser may be taken from another race or species or even class, without the effect of its impression being changed. On the specific characters of the receiving germ the organiser has no influence except such as may eventually be explained by the diffusion of a ferment. This, again, is in accordance with the facts of transplantation in larvæ and in adults where the graft tends to retain its specific characters as also does the stock. Only in a few instances do apparent exceptions occur, thus melanin-producing ferments change the colour of the graft, as in the eyes of albinotic axolotls grafted to coloured specimens (Koppanyi (40)). There are similar cases in limbs (Schaxel (66)). The stock is mainly a nurse to the graft, and for the survival of the latter only the restoration of blood circulation is absolutely necessary. The restoration of functional nerves, though usually possible if the autophoric method of transplantation (Przibram (60)) is closely followed, is not necessary for the preservation or restoration of form. Regeneration in cut limbs is also independent of the connection of the nerve with its own centre in amphibia, as in arthropods, though possibly some sort of nerve, perhaps belonging to the sympathetic system, has to be present before regenerative morphogenesis proceeds (Weiss (86-88), Schotté (68-72)). Even for a limb to express its functions it is not necessary that all the nerves should be present. This is shown by experiments on transplanting limbs of the larvæ of Salamandra maculosa to the neighbourhood of another limb (Weiss (86, 87)). Just as in embryogeny the organs all develop before the nervous connection functions, and loss of nerve centres does not prevent the morphogeny of the peripheral organs, there being no difference in this respect between first development and regeneration. Eyes severed from their nerve centre and grafted to the body even at an unusual place do not degenerate permanently in the amphibia as was first proved by Uhlenhuth (1912) and confirmed in 1917(81). If eyes are replanted into the socket, the optic nerve may regenerate. This has been repeatedly observed in fish (Koppányi (37, 40), Ask(1)); urodeles (Pardo 1906, Koppányi (37, 39, 40), Kolmer(31)) ; frogs (Koppányi (37)), Trapesonzewa in Koltzoffs laboratory, etc.; rats (Koppányi (37-39), Carlson (6), Koppányi and Baker (41), and in a rabbit (Kolmer 31)).* Grafted eyes of urodele larvæ will metamorphosize simultaneously with their host without having obtained connection with their brain centre (Uhlenhuth), showing that metamorphosis also is independent of the nerve connection. It is the same in anura (Vrtelówna (85)). Similar cases have recently been recorded for transplanted intestines in these animals (Sembrat (73)). Kopć (35) has grafted eyes to the abdomens of caterpillars, and these eyes metamorphosed, although no nervous connection was made with their host. Malpighian tubes and wing rudiments also developed after they had been transplanted to a much younger caterpillar, again without any nervous connection (Kopć 34)). It must be added, however, that Kopć (33) believes an impulse from the larval brain to be necessary for the onset of metamorphosis. The character of this influence is not clear, but it may be rather that of the production of an endocrine substance than of nervous origin. Once initiated, the process of metamorphosis proceeds even in decapitated caterpillars (Przibram (59)). Whereas the peripheral organs do not seem to depend to any large extent on their nerve centres for morphogenesis, the nerve centres themselves are not independent of the presence or absence of their peripheral organs. Kopć (35) has observed that in the caterpillars deprived of their eyes the optic ganglia degenerate, while Dürken(15) reports progressive degeneration of the central nervous system after extirpation of sensory or motor organs in amphibian embryos, a degeneration which even begins to affect peripheral parts in a general way and which weakens their morphogenetic faculties (Hamburger (20)). A loading of centres with new peripheral organs has, on the contrary, strengthened these centres in Detwiler’s (12-14) experiments on transplantation in amblystoma members on the tail-bud stage. I think we can no longer uphold a specific influence of nerve centres on developing form ; nor maintain, as Roux suggested, that there is a period of self-differentiation without influence of nerve or function on the form followed by a period of functional activity where self-differentiation is reduced and the form largely dependent on nervous function. Indeed some early processes of development seem to be more dependent on self-differentiation than the later ones, and a nervous influence on developing organs does not seem to enter into morphogenesis proper at all, neither into early development nor into regeneration.
6. Facts against preformation
Yet another belief has been badly shattered by the experiments of the last ten years, namely, that of the immutability of the germ layers. We have already seen that ectodermic structures may arise from entodermic tissue, that the regenerative blastema does not seem to arise from these different layers at all, and many more instances could be cited pointing to the conclusion that the formation of different layers may not be the essential thing in development. Vogt’s (83, 84) experiments with intra vitam staining have lately shown that in the amphibian embryo (Triton) there is nothing which corresponds to the cœlomic foldings which we used to believe necessary for the “Coelomata.” Recently J. Runnström (64) has described cœlom-rudiments in Echinoids which arise from ectoderm after injury to the primitive entoderm. We may also remember that Selenka some time ago described an inversion of the primitive layers of rodents which is hardly compatible with an immutable fixation of the germ layers. It seems to me we have not to look to the unfolding of primitive organs or the unfolding of visible anlagen for the explanation of the facts of ontogeny and regeneration, but to a system of potentialities unfolding progressively with age. These are all included in the totipotential regions as long as such exist, and are always all present in the germ cells. I do not wish to enter here into discussion of physical probability,* but I would like to emphasise that the germ cells are confined to those regions of the body we would deem the most proximal of any of the parts of the body, and that they mostly arise from that undifferentiated plasma which connects ectoderm and endoderm, namely, from the mesoderm, mesenchyme or parenchyme. In those low types in which there are large totipotent somatic regions the germ cells may be regenerated after removal of the germinal segments, e.g. hydroids, planarians (Vandel (32)), some annelids, echinoderms, and tunicates ; in forms which lack such regions the germ-parenchyme alone may regenerate germ cells, e.g. arthropods (Inachus), vertebrates, even mammals after removal of all formed ovarian tissue (Davenport (9)). It is perhaps noteworthy that those types in which an early separation of the germ blastomere has been observed are those of exceptionally low regenerative power, e.g. Ascaris, Cyclops, Petromyzon, in comparison even with the higher developed types of worms, crayfish, and cold-blooded vertebrates respectively. It is possibly of importance to note that the regenerative blastemas seem to derive their material chiefly from mesodermal tissue. At any rate the contrast between the facts and the theory of preformative regenerative anlagen (Weissmann) is becoming stronger every year. There is little doubt left that such special anlagen cannot be found, and that the distribution of nuclear material during ontogeny has nothing to do with the building up of the organs themselves. This conclusion has also been reached by geneticists, for instance, Johannsen and T. H. Morgan, who emphasise that the genes are not to be considered as anlagen of the parts of an organism. In this connection it is of interest that in natural and artificial parthenogenesis the reduction of chromosome number may remain permanent or the missing set may be eventually regenerated (sea-urchin—Delage, frog—Loeb and Goldschmidt, Herlant, Hovasse(27, 28) without other differences in the resulting form than a variation in the size of nucleus and cell. If we regard sex as a race character as Mendelians are now wont to do, the problem arises how one is to account for the regeneration of a hetero-chromosome in those cases where both sexes may arise by parthenogenesis. At the last meeting of the German Anatomical Society held in Vienna in April 1925 Guenther Hertwig (26) reported that he had succeeded in transplanting limb anlagen of haploid Tirtons to diploid triton embryos. The haploid tritons had been obtained by Oscar and Paula Hertwig’s method of destroying the egg nucleus by radium and subsequent introduction of the sperm-nucleus. The haploid grafts were progressively converted into diploid limbs without disturbing their form, although they had not even been transplanted to their normal position. Experiments with heteroplastic grafts will perhaps show whether the diploid state has been restored in the formerly haploid cells or whether material derived solely from the diploid body has had the form of the limb anlagen impressed on it as G. Hertwig assumes.
7. Conclusion
In no case does the pure preformation theory afford a plausible explanation of the facts. The course of regeneration in animals is also not in accordance with an “Auslösungs-process” of preformed regenerative anlagen. It coincides rather with that of an energy flow from a point of higher to one of lower potential, as I have endeavoured to show by a comparison of all the quantitative data available (54). Loeb (43-45) by experiments on the relationship between the amount of regeneration produced by a unit mass of tissue arrived at the conclusion that in plants regeneration follows the rule of mass action. This is not necessarily in contradiction to Haberlandt’s idea that “wound hormones” are the initiators of the regenerative plasma flow. Remembering the facts of progressive localization and restriction of potentialities, we have, I think, to draw the conclusion that a purely epigenetic theory of regeneration or a vitalistic view of a special regulating entelechy (different from the forces inherent in the several “Schichten” with their characteristic potentialities) is no longer tenable. As I have suggested before, “regeneration” may be regarded as an acceleration of normal growth processes, restoring a morphogenetic equilibrium in consequence of some disturbance, and one which is made possible by the fact that the remaining parts still retain the potentialities to form those which are lost. Like other problems of morphogenesis regeneration is one not only of statics and dynamics but also of kinetics. Theories which attempt to solve such problems by the mere arrangement of “building stones” and those which postulate “entelechies” as hovering above physical matter will scarcely be of lasting satisfaction to natural science. Perhaps it may be useful to introduce a term for the theoretical view here advocated: “Apogenesis” (Lecture, Philos. Soc., Vienna, 6th March 1924) may designate development with gradual loss of potencies in its parts as revealed by the experiments on regeneration and transplantation.
References
For literature up to 1909 see Journ. Exp. Zool, 1, 1907 ; 2, 1909; 5, 1914; a comprehensive treatise on Regeneration and Transplantation will be found in Bethe’s Handbook der Physiologie (in press), Springer, Berlin.
Cuénot(7 and 8) interprets similar results in his own experiments as “mutations”
Extensive literature will be found in “Tierpfropfung,” Vieweg, Braunschweig, 1926.
A book on this subject, Die anorganischen Grenzgebiete der Biologie, is in press with Bomträger, Berlin.