You are here


1. The Foundations of the Physiology of Development. “EVOLUTIOANDEPIGENESIS


Of all the purely hypothetic theories on morphogenesis, that of August Weismann1 can claim to have had the greatest influence, and to be at the same time the most logical and the most elaborated. The “germ-plasma” theory of the German author is generally considered as being a theory of heredity, and that is true inasmuch as problems of inheritance proper have been the starting-point of all his hypothetic speculations, and also form in some respects the most valuable part of them. But, rightly understood, Weismann’s theory consists of two independent parts, which relate to morphogenesis and to heredity separately, and it is only the first which we shall have to take into consideration at present; what is generally known as the doctrine of the “continuity of the germ-plasm” will be discussed in a later chapter.

Weismann assumes that a very complicated organised structure, below the limits of visibility even with the highest optical powers, is the foundation of all morphogenetic processes, in such a way that, whilst part of this structure is handed over from generation to generation as the basis of heredity, another part of it is disintegrated during the individual development, and directs development by being disintegrated. The expression, “part” of the structure, first calls for some explanation. Weismann supposes several examples, several copies, as it were, of his structure to be present in the germ cells, and it is to these copies that the word “part” has been applied by us: at least one copy has to be disintegrated during ontogeny.

The morphogenetic structure is assumed to be present in the nucleus of the germ cells, and Weismann supposes the disintegration of his hypothetic structure to be accomplished by nuclear division. By the cleavage of the egg, the most fundamental parts of it are separated one from the other. The word “fundamental” must be understood as applying not to organs or tissues, but to the chief relations of symmetry; the first cleavage, for instance, may separate the right and the left part of the primordial structure, the second one its upper and lower parts, and after the third or equatorial cleavage all the principal eighths of our minute organisation are divided off: for the minute organisation, it must now be added, had been supposed to be built up differently in the three directions of space, just as the adult organism is.

At the end of organogenesis the structure is assumed to have been broken up into its elements, and these elements, each of them localised in a particular cell, determine the fate of the single cells of the adult organism, the whole process having been accompanied by growth.

Here let us pause for a moment. There cannot be any doubt that Weismann’s theory resembles to a very high degree some of the old “evolutio” doctrines of the eighteenth century; those doctrines, though not pretending to “see” the chick in the hen’s egg, yet postulated its existence in a practically invisible form but qua chick with all its parts. Weismann’s theory, however, is a little less crude. The chick quâ chick is not supposed to be present in the hen’s egg before development, and ontogeny is not regarded as a mere growth of that chick in miniature; but what really is supposed to be present in the egg is nevertheless a something that in all its parts corresponds to all the parts of the chick, only under a somewhat different aspect, while all the relations of the parts of the one correspond to the relations of the parts of the other. Indeed, only on such an hypothesis of a fairly fixed and rigid relation between the parts of the morphogenetic structure could it be possible for the disintegration of the structure to go on, not by parts of organisation, but by parts of symmetry; which, indeed, is a very strange, but not an illogical, feature of Weismann’s doctrine.

Weismann is absolutely convinced that there must be a theory of “evolutio” to account for the ontogenetic facts; that “epigenesis” has its place only in descriptive embryology, where, indeed, as we know, manifoldness in the visible sense is produced, but that epigenesis can never form the foundation of a real morphogenetic theory: theoretically one pre-existing material manifoldness is transformed into the other. An epigenetic theory would lead right beyond natural science, Weismann thinks, as in fact all such theories, if fully worked out, have carried their authors to vitalistic views. But vitalism is regarded by him as dethroned for ever.

Under these circumstances we have a good right, it seems to me, to speak of a dogmatic basis of Weismann’s theory of development.

But to complete the outlines of the theory itself: Weismann was well aware that there were some grave difficulties attaching to his statements: all the facts of so-called adventitious morphogenesis in plants, of regeneration in animals, proved that the morphogenetic organisation could not be fully disintegrated during ontogeny. But these difficulties were not absolute: they could be overcome: indeed, Weismann assumes that in certain specific cases—and he regarded all cases of restoration of a destroyed organisation as due to specific properties of the subjects, originated by roundabout variations and natural selection—that in specific cases, specific arrangements of minute parts were formed during the process of disintegration, and were surrendered to specific cells during development, from which regeneration or adventitious budding could originate if required. “Plasma of reserve” was the name bestowed on these hypothetic arrangements.

Almost independently, another German author, Wilhelm Roux,2 has advocated a theoretical view of morphogenesis which very closely resembles the hypothesis of Weismann.

But in spite of this similarity of the outset, we enter an altogether different field of biological investigation on mentioning Roux’s name: we are leaving hypothetic construction, at least in its absoluteness, and are entering the realms of scientific experiment in morphology.


We have already said that an hypothesis about the foundation of individual development was Roux’s starting-point. Like Weismann, he supposed that there exists a very complicated structure in the germ, and that nuclear division leads to the disintegration of that structure. He next tried to bring forward what might be called a number of indicia supporting his view.

A close relation had been found to exist in many cases between the direction of the first cleavage furrows of the germ and the direction of the chief planes of symmetry in the adult: the first cleavage, for instance, very often corresponds to the median plane, or stands at right-angles to it. And in other instances, such as have been worked out into the doctrine of so-called “cell-lineages”, typical cleavage cells were found to correspond to typical organs. Was not that a strong support for a theory which regarded cellular division as the principal means of differentiation? It is true, the close relations between cleavage and symmetry did not exist in every case, but then there had always happened some specific experimental disturbances, e.g. influences of an abnormal direction of gravity on account of a turning over of the egg, and it was easy to reconcile such cases with the generally accepted theory on the assumption of what was called “anachronism” of cleavage.

But Roux was not satisfied with mere indicia; he wanted a proof, and with this intention he carried out an experiment which has become very celebrated.3 With a hot needle he killed one of the first two blastomeres of the frog’s egg after the full accomplishment of its first cleavage, and then watched the development of the surviving cell. A typical half-embryo was seen to emerge—an organism, indeed, which was as much a half as if a fully-formed embryo of a certain stage had been cut in two by a razor. It was especially in the anterior part of the embryo that its “halfness” could most clearly be demonstrated.

That seemed to be a proof of Weismann’s and Roux’s theory of development, a proof of the hypothesis that there is a very complicated structure which promotes ontogeny by its disintegration, carried out during the cell divisions of embryology by the aid of the process of nuclear division, the so-called “karyokinesis”. But things were far from being decided in a definitive manner.


Roux’s results were published for the first time in 1888; three years later I tried to repeat his fundamental experiment on another subject and by a somewhat different method. It was known from the cytological researches of the brothers Hertwig and Boveri that the eggs of the common sea-urchin (Echinus microtuberculatus) are able to stand well all sorts of rough treatment, and that, in particular, when broken into pieces by shaking, their fragments will survive and continue to segment. I took advantage of these facts for my purposes. I shook the germs rather violently during their two-cell stage, and in several instances I succeeded in killing one of the blastomeres, while the other one was not damaged, or in separating the two blastomeres from one another.4

Let us now follow the development of the isolated surviving cell. It went through cleavage just as it would have done in contact with its sister-cell, and there occurred cleavage stages which were just half of the normal ones. The stage, for instance, which corresponded to the normal sixteen-cell stage, and which, of course, in my subjects was built up of eight elements only, showed two micromeres, two macromeres, and four cells of medium size, exactly as if a normal sixteen-cell stage had been cut in two; and the form of the whole was that of a hemisphere. So far there was no divergence from Roux’s results.

The development of our Echinus proceeds rather rapidly, the cleavage being accomplished in about fifteen hours. I quickly noticed, on the evening of the first day of the experiment, when the half-germ was composed of about two hundred elements, that the margin of the hemispherical germ bent together a little, as if it were about to form a whole sphere of smaller size, and, indeed, the next morning a whole diminutive blastula was swimming about. I was so much convinced that

Fig. 5.—Illustration of Experiments on Echinus.
a1 and b1. Normal gastrula and normal pluteus.
a2 and b2. “Half”—gastrula and “half”—pluteus, that ought to result from one of the two first blastomeres, when isolated, according to the theory of “evolutio”.
a3 and b3. The small but whole gastrula and pluteus, that actually do result.

I should get Roux’s morphogenetical result in all its features that, even in spite of this whole blastula, I now expected that the next morning would reveal to me the half-organisation of my subject once more; the intestine, I supposed, might come out quite on one side of it, as a half-tube, and the mesenchyme ring might be a half one also.

But things turned out as they were bound to do and not as I had expected; there was a typically whole gastrula on my dish the next morning, differing only by its small size from a normal one and this small but whole gastrula was followed by a whole and typical small pluteus-larva (Fig. 5).

That was just the opposite of Roux’s result: one of the first two blastomeres had undergone a half-cleavage as in his case, but then it had become a whole organism by a simple process of rearrangement of its material, without anything that resembled regeneration, in the sense of a completion by budding from a wound.

If one blastomere of the two-cell stage was thus capable of performing the morphogenetical process in its totality, it became, of course, impossible to allow that nuclear division had separated any sort of “germ-plasm” into two different halves, and not even the protoplasm of the egg could be said to have been divided by the first cleavage furrow into unequal parts, as the postulate of the strict theory of so-called “evolutio” had been. This was a very important result, sufficient alone to overthrow at once the theory of ontogenetical “evolutio”, the “Mosaiktheorie” as it had been called, in its exclusiveness.

After first widening the circle of my observations by showing that one of the first four blastomeres is also capable of performing a whole organogenesis, and that three of the first four blastomeres together result in an absolutely perfect organism, I went on to follow up separately one of the two fundamental problems which had been suggested by my first experiment: Was there anything more to find out about the importance or unimportance of the single nuclear divisions in morphogenesis?5

By raising the temperature of the medium or by diluting the sea-water to a certain degree it proved at first to be possible to alter in a rather fundamental way the type of the cleavage-stages without any damage to the resulting organism. There may be no micromeres at the sixteen-cell stage, or they may appear as early as in the stage of eight cells; no matter, the larva is bound to be typical. So it certainly is not necessary for all the cleavages to occur just in their normal order.

But of greater importance for our purposes was what followed. I succeeded in pressing the eggs of Echinus between two glass plates, rather tightly, but without killing them; the eggs became deformed to comparatively flat plates of a large diameter. Now, in these eggs all nuclear division occurred at right-angles to the direction of pressure, that is to say, in the direction of the plates, as long as the pressure lasted; but the divisions began to occur at right-angles to their former direction, as soon as the pressure ceased. By letting the pressure be at work for different times I therefore, of course, had it quite in my power to obtain cleavage types just as I wanted to get them. If, for instance, I kept the eggs under pressure until the eight-cell stage was complete, I got a plate of eight cells, one beside the other, instead of two rings, of four cells each, one above the

Fig. 6—Pressure-experiments on Echinus.
a1 and b1. Two normal cleavage stages, consisting of eight and sixteen cells.
a2 and b2. Corresponding stages modified by exerting pressure until the eight-cell stage was finished. See text.

other, as in the normal case; but the next cell division occurred at right-angles to the former ones, and a sixteen-cell stage, of two plates of eight cells each, one above the other, was the result. If the pressure continued until the sixteen-cell stage was reached, sixteen cells lay together in one plate, and two plates of sixteen cells each, one above the other, were the result of the next cleavage.

We are not, however, studying these things for cytological but for morphogenetical purposes, and for these the cleavage phenomenon itself is less important than the organogenetic result of it: all our subjects resulted in absolutely normal organisms. Now, it is clear that the spatial relations of the different nuclear divisions to each other are anything but normal in the eggs subjected to the pressure experiments; that, so to say, every nucleus has got quite different neighbours if compared with the “normal” case. If that makes no difference, then there cannot exist any close relation between the single nuclear divisions and organogenesis at all, and the conclusion we have drawn moire provisionally from the whole development of isolated blastomeres has been extended and proved in the most perfect manner. There ought to result a morphogenetic chaos according to the theory of real “evolutio” carried out by nuclear division, if the positions of the single nuclei were fundamentally changed with regard to one another (Fig. 6). But now there resulted not chaos, but the normal organisation: therefore it was disproved, in the strictest way, that nuclear divisions have any bearing on the origin of organisation—at least, as far as the divisions during cleavage come into account.

On the egg of the frog (O. Hertwig), the annelids (E. B. Wilson), and ciona (T. H. Morgan) my pressure experiments have been carried out with the same result.6

An important experiment carried out by Spemann with the egg of the newt confirms the result of my pressure experiments in a different way: He corded the egg after fertilisation, by the aid of a hair, in such a way that one of the parts, which still were in communication by a bridge of protoplasm, was without the nucleus. Later on, at the stage of 8 or 16 cells, one of the nuclei migrated into that part. But both parts resulted in the formation of complete organisms, whilst, according to Weismann, that part which had received a nucleus of the 8–or 16–cell stage “should” have given but one-eighth or one-sixteenth of the organisation.

Finally, I have succeeded in raising one giant organism from two sea-urchin eggs, fused in the blastula stage; and the same result has been obtained with the egg of the newt by Mangold, in a slightly different way.

This, of course, is the real counterpart of my first experiments: in these, I got many (two or four) complete organisms from a material that normally should have given one; now, there is one instead of many. The second result is perhaps even more striking than the first. Both of them, in any case, contradict Weismann’s hypothesis.


Nuclear division, then, cannot be the basis of organogenesis, and all we know about the whole development of isolated blastomeres seems to show that there exists nothing responsible for differentiation in the protoplasm either.

But would that be possible? It cannot appear possible on a more profound consideration of the nature of morphogenesis, it seems to me: as the untypical agents of the medium cannot be responsible in any way for the origin of a form combination which is most typical and specific, there must be somewhere in the egg itself a certain factor which is responsible at least for the general orientation and symmetry of it. Considerations of this kind led me, as early as 1893,7 to urge the hypothesis that there existed, that there must exist, a sort of intimate structure in the egg, including polarity and bilaterality as the chief features of its symmetry, a structure which belongs to every smallest element of the egg, and which might be imagined by analogy under the form of elementary magnets.8 This hypothetic structure could have its seat in the protoplasm only. In the egg of echinoderms it would be capable of such a quick rearrangement after being disturbed, that it could not be observed but only inferred logically; there might, however, be cases in which its real discovery would be possible. Indeed Roux’s frog-experiment seems to be a case where it is found to be at work: at least it seems very probable to assume that Roux obtained half of a frog’s embryo because the protoplasm of the isolated blastomere had preserved the “halfness” of its intimate structure, and had not been able to form a small whole out of it.

Of course, it was my principal object to verify this hypothesis, and such verification became possible in a set of experiments which my friend T. H. Morgan and myself carried out together,9 in 1895, on the eggs of ctenophores, a sort of pelagic animals, somewhat resembling the jelly-fish, but of a rather different inner organisation. The zoologist Chun had found, even before Roux’s analytical studies, that isolated blastomeres of the ctenophore egg behave like parts of the whole and result in a half-organisation like the frog’s germ does; Chun had not laid much stress on his discovery, which now, of course, from the new points of view, became a very important one. We first repeated Chun’s experiment and obtained his results, with the sole exception that there was a tendency of the endoderm of the half-larva of Beroë to become more than “half”. But that was not what we chiefly wanted to study. We succeeded in cutting away a certain mass of the protoplasm of the ctenophore egg just before it began to cleave, without damaging its nuclear material in any way: in all cases, where the cut was performed at the side, there resulted a certain type of larvae from our experiments which showed exactly the same sort of defects as were present in larvae developed from one of the first two blastomeres alone.

The hypothesis of the morphogenetic importance of protoplasm had thus been proved. In our experiments there was all of the nuclear material, but there were defects on one side of the protoplasm of the egg; and the defects in the adult were found to correspond to these defects in the protoplasm.

And now O. Schultze and Morgan succeeded in performing some experiments which directly proved the hypothesis of the part played by protoplasm in the subject employed by Roux, viz., the frog’s egg. The first of these investigators managed to rear two whole frog embryos of small size, if he slightly pressed the two-cell stage of that form between two plates of glass and turned it over; and Morgan,10 after having killed one of the first two blastomeres, as was done in the original experiment of Roux, was able to bring the surviving one to a half or to a whole development according as it was undisturbed or turned. There cannot be any doubt that in both of these cases it is the possibility of a rearrangement of protoplasm, offered by the turning over, which allows the isolated blastomere to develop as a whole. The regulation of the frog’s egg, with regard to its becoming whole, may be called facultative, whilst the same regulation of the egg of Echinus is obligatory. It is not without interest to note that the first two blastomeres of the common newt, i.e., of a form which belongs to the other class of Amphibia, after a separation of any kind, always develop as wholes, their faculty of regulation being obligatory, like that of Echinus.

Whole or partial development may thus be dependent on the power of regulation contained in the intimate polar-bilateral structure of the protoplasm. Where this is so, the regulation and the differences in development are both connected with the chief relations of symmetry. The development becomes a half or a quarter of the normal because there is only one-half or one-quarter of a certain structure present, one-half or one-quarter with regard to the very wholeness of this structure; the development is whole, in spite of disturbances, if the intimate structure became whole first. We may describe the “wholeness”, “halfness”, or “quarterness” of our hypothetic structure in a mathematical way, by using three axes, at right-angles to one another, as the base of orientation. To each of these, x, y, and z, a certain specific state with regard to the symmetrical relations corresponds; thence it follows that, if there are wanting all those parts of the intimate structure which are determined, say, by a negative value of y, by minus y, then there is wanting half of the intimate structure; and this halfness of the intimate structure is followed by the halfness of organogenesis, the dependence of the latter on the intimate structure being established. But if regulation has restored, on a smaller scale, the whole of the arrangement according to all values of x, y and z, development also can take place completely (Fig. 7).

I am quite aware that such a discussion is rather empty and purely formal, nevertheless it is by no means without

Fig. 7.—Diagram illustrating the intimate Regulation of Protoplasm fromHalftoWhole”.
The large circle represents the original structure of the egg. In all cases where cleavage-cells of the two-cell stage are isolated this original structure is only present as “half” in the beginning, say only on the right (+Y) side. Development then becomes “half”, if the intimate structure remains half; but it becomes “whole” (on a smaller scale) if a new whole-structure (small circle!) is formed by regulatory processes.

value, for it shows most clearly the differences between what we have called the intimate structure of germs, responsible only for the general symmetry of themselves and of their isolated parts, and another sort of possible structure of the egg-protoplasm which we now shall have to consider, and which, at the first glance, seems to form a serious difficulty to our statements, as far at least as they claim to be of general importance. The study of this other sort of germinal structure at the same time will lead us a step further in our historical sketch of the first years of experimental embryology, and will bring this sketch to its end.


It was known already about 1890, from the careful study of what has been called “cell-lineage”, that in the eggs of several families of the animal kingdom the origin of certain organs may be traced back to individual cells of cleavage, having a typical

Fig. 8.—The Mollusc Dentalium (after E. B. Wilson).
a. The egg, consisting of three different kinds of protoplasmatic material.
b. First cleavage-stage. There are two cells and one “pseudo-cell”, the yolk-sac, which contains no nucleus. This was removed in Crampton’s experiment.

histological character of their own. In America especially such researches have been carried out with the utmost minuteness, E. B. Wilson’s study of the cell-lineage of the Annelid Nereis being the first of them. If it were true that nuclear division is of no determining influence upon the ontogenetic fate of the blastomeres, only peculiarities of the different parts of the protoplasm could account for such relations of special cleavage cells to special organs. I advocated this view as early as in 1894, and it was proved two years later by Crampton, a pupil of Wilson’s, in some very fine experiments performed on the germ of a certain mollusc.11 The egg of this form contains a special sort of protoplasm near its vegetative pole, and this part of it is separated at each of the first two segmentations by a sort of pseudo-cleavage, leading to stages of three and five separated masses instead of two and four, the supernumerary mass being the so-called “yolk-sac” and possessing no nuclear elements (Fig. 8). Crampton removed this yolk-sac at the two-cell stage, and he found that the cleavage of the germs thus operated upon was normal except with regard to the size and histological appearance of one cell, and that the larvae originating from these germs were complete in every respect except in their mesenchyme, which was wanting. A special part of the protoplasm of the egg had thus been brought into relation with quite a special part of organisation, and that special part of the protoplasm contained no nucleus.


This experiment of Crampton’s, afterwards confirmed by Wilson himself, may be said to have closed the first period of the new science of physiology of form, a period devoted almost exclusively to the problem whether the theory of nuclear division or, in a wider sense, whether the theory of a strict “evolutio” as the basis of organogenesis was true or not.

It was shown, as we have seen, that the theory of the “qualitatively unequal nuclear division” (“qualitativ-ungleiche Kernteilung” in German) certainly was not true, and that there also was no strict “evolutio” in protoplasm. Hence Weismann’s theory was clearly disproved. There certainly is a good deal of real “epigenesis” in ontogeny, a good deal of “production of manifoldness”, not only with regard to visibility but in a more profound meaning. But some sort of pre-formation had also been proved to exist, and this pre-formation, or, if you like, this restricted evolutio, was found to be of two different kinds. First an intimate organisation of the protoplasm, spoken of as its polarity and bilaterality, was discovered, and this had to be postulated for every kind of germs, even when it was overshadowed by immediate obligatory regulation after disturbances. Besides that, there were cases in which a real specificity of special parts of the germ existed, a relation of these special parts to special organs; but this sort of specification also was shown to belong to the protoplasm.

It follows from all we have mentioned about the organisation of protoplasm and its bearings on morphogenesis, that the eggs of different animals may behave rather differently in this respect, and that the eggs indeed may be classified according to the degree of their organisation. Though we must leave a detailed discussion of these topics to zoology proper, we yet shall try shortly to summarise what has been ascertained about them in the different classes of the animal kingdom. A full regulation of the intimate structure of isolated blastomeres to a new whole has been proved to exist in the highest degree in the eggs of all echinoderms, medusae, nemertines, Amphioxus, fishes, and in one class of the Amphibia (the Urodela); it is facultative only among the other class of Amphibia, the Anura, and seems to be only partly developed or to be wanting altogether among ctenophora, ascidia, annelids, and mollusca. Peculiarities in the organisation of specific parts of protoplasm have been proved to occur in more cases than at first had been assumed; they exist even in the echinoderm egg, as experiments of later years have shown; even here a sort of specification exists at the vegetative pole of the egg, though it is liable to a certain land of regulation; the same is true in medusae, nemertines, etc.; but among molluscs, ascidians, and annelids no regulation about the specific organisation of the germ in cleavage has been found in any case.

The differences in the degree of regulability of the intimate germinal structure may easily be reduced to simple differences in the physical consistency of their protoplasm.12 But all differences in specific organisation must remain as they are for the present; it will be one of the aims of the future theory of development to trace these differences also to a common source.

That such an endeavour will probably be not without success, is clear, I should think, from the mere fact that differences with regard to germinal specific pre-formation do not agree in any way with the systematic position of the animals exhibiting them; for, strange as it would be if there were two utterly different kinds of morphogenesis, it would be still more strange if there were differences in morphogenesis which were totally unconnected with systematic relationship: the ctenophores behaving differently from the medusae, and Amphioxus differently from ascidians.


We now might close this chapter, which has chiefly dealt with the disproof of a certain sort of ontogenetic theories, and therefore has been almost negative in its character, did it not seem desirable to add at least a few words about the later discoveries relating to morphogenetic restorations of the adult. We have learnt that Weismann created his concept of “reserve plasma” to account for what little he knew about “restitutions”—that is, about the restoration of lost parts; he only knew regeneration proper in animals and the formation of adventitious buds in plants. It is common to both of these phenomena that they take their origin from typically localised points of the body in every case; each time they occur a certain well-defined part of the body is charged with the restoration of the lost parts. To explain such cases Weismann’s hypothesis was quite adequate, at least in a logical sense. But at present, as we shall discuss more fully in another chapter, we know of some very widespread forms of restitution, in which what is to be done for a replacement of the lost is not entrusted to one typical part of the body in every case, but in which the whole of the morphogenetic action to be performed is transferred in its single parts to the single parts of the body which is accomplishing restoration: each of its parts has to take an individual share in the process of restoration, effecting what is properly called a certain kind of “re-differentiation” (“Umdifferenzierung”), and this share varies according to the relative position of the part in each case. Later on, these statements will appear in more correct form than at present, and then it will become clear that we are fully entitled to emphasise, at the end of our criticism of Weismann’s theory, that his hypothesis relating to restorations can be no more true than his theory of development proper was found to be.

And now we shall pass on to our positive work.

We shall try to sketch the outlines of what might properly be called an analytical theory of morphogenesis; that is, to explain the sum of our knowledge about organic form-production, gained by experiment and by logical analysis, in the form of a real system, in which each part will be, or at least will try to be, in its proper place and in relation with every other part. Our analytical work will give us ample opportunity of mentioning many important topics of so-called general physiology also, irrespective of morphogenesis as such. But morphogenesis is always to be the centre and starting-point of our analysis. As I myself approach the subject as a zoologist, animal morphogenesis, as before, will be the principal subject of what is to follow.

2. Analytical Theory of Morphogenesis

We must now study morphogenesis in a systematic way, i.e. by dividing it into various characteristics and problems, as they offer themselves to the unbiassed observer. Six different problems will thus enter the scene—and a certain feature of one of them will force us to discuss the whole morphogenetic problem once more.


Prospective Value and Prospective Potency

Wilhelm Roux did not fail to see that the questions of the locality and the time of all morphogenetic differentiations had to be solved first, before any problem of causality proper could be attacked. From this point of view he carried out his fundamental experiments.

It is only in terminology that we differ from his views, if we prefer to call our introductory chapter an analysis of the distribution of morphogenetic potencies. The result will be, of course, rather different from what Roux expected it would be.

Let us begin by laying down two fundamental concepts. Suppose we have here a definite embryo in a definite state of development, say a blastula, or a gastrula, or some sort of larva: then we are entitled to study any special element of any special elementary organ of this germ with respect to what is actually to develop out of this very element in the future actual course of this embryology; it is the actual, the real fate of our element, that we take into account. I have proposed to call this real fate of each embryonic part in this very definite line of morphogenesis its prospective value (“prospective Bedeutung” in German). The fundamental question of the first chapter of our analytical theory of development may now be stated as follows: Is the prospective value of each part of any state of the morphogenetic line constant, i.e. is it unchangeable, can it be nothing but one; or is it variable, may it change according to different circumstances?

We first introduce a second concept: the term prospective potency (“prospective Potenz” in German) of each embryonic element. The term “prospective morphogenetic potency” is to signify the possible fate of each of those elements. With the aid of our two concepts, we are now able to formulate our introductory question thus: Is the prospective potency of each embryonic part fully given by its prospective value in a certain definite case; is it, so to say, identical with it, or does the prospective potency contain more than the prospective value of an element in a certain case reveals?

We know already from our historical sketch that the latter is true: that the actual fate of a part need not be identical with its possible fate, at least in many cases; that the potency of the first four blastomeres of the egg of the sea-urchin, for instance, has a far wider range than is shown by what each of them actually performs in even this ontogeny. There are more morphogenetic possibilities contained in each embryonic part than are actually realised in a special morphogenetic case.

As the most important special morphogenetic case is, of course, the so-called “normal” one, we can also express our formula in terms of special reference to it: there are more morphogenetic possibilities in each part than the observation of the normal development can reveal. Thus we have at once justified the application of analytical experiment to morphogenesis, and have stated its most important results.

As the introductory experiments have shown already that the prospective potency of embryonic parts, at least in certain cases, can exceed their prospective value, the concept of prospective potency at the very beginning of our studies puts itself in the centre of analytical interest, leaving to the concept of prospective value the second place only. For that each embryonic part actually has a certain prospective value, a specified actual fate in every single case of ontogeny, is clear from itself and does not affirm more than the reality of morphogenetic cases in general; but that the prospective value of the elements may change, that there is a morphogenetic power in them which contains more than actuality—in other words, that the term “prospective potency” has not only a logical but a factual interest: all these points amount to a statement not only of the most fundamental introductory results but also of the actual problems of the physiology of form.

If at each point of the germ something else can be formed than actually is formed, why then does there happen in each case just what happens and nothing else? In these words indeed we may state the chief problem of our science, at least after the fundamental relation of the superiority of prospective potency to prospective value has been generally shown.

We consequently may shortly formulate our first problem as the question of the distribution of the prospective morphogenetic potencies in the germ. Now, this general question involves a number of particular ones. Up to what stage, if at all, is there an absolutely equal distribution of the potencies over all the elements of the germ? When such an equal distribution has ceased to exist at a certain stage, what are then the relations between the parts of different potency? How, on the other hand, does a newly arisen, more specialised sort of potency behave with regard to the original general potency, and what about the distribution of the more restricted potency?

The Potencies of the Blastomeres

At first we turn back to our experiments on the egg of the sea-urchin as a type of the germ in the very earliest stages. We know already that each of the first two, or each of the first four, or three of the first four blastomeres together may produce a whole organism. We may add that the swimming blastula, consisting of about one thousand cells, when cut in two quite at random, in a plane coincident with, or at least passing near, its polar axis, may form two fully developed organisms out of its halves.13 We may formulate this result in the words: the prospective potency of the single cells of a blastula of Echinus is the same for all of them, at least around the axis; their prospective value is as far as possible from being invariable.

But we may say even a little more: what actually will happen in each of the blastula cells in any special case of development experimentally determined depends on the position of that cell in the whole, if the “whole” is put into relation with any fixed system of co-ordinates; or more shortly, “the prospective value of any blastula cell is a function of its position in the whole”.

I know from former experience that this statement wants a few words of explanation. The word “function” is employed here in the most general, mathematical sense, simply to express that the prospective value, the actual fate of a cell, will change whenever its position in the whole is different.14 The “whole” may be related to any three axes drawn through the normal undisturbed egg, on the hypothesis that there exists a primary polarity and bilaterality of the germ; the axes which determine this sort of symmetry may, of course, conveniently be taken as co-ordinates; but that is not necessary.

The Potencies of Elementary Organs in General

Before dealing with other very young germs, I think it advisable to describe first an experiment which is carried out at a later stage of our well-known form. This experiment will easily lead to a few new concepts, which we shall want later on, and will serve, on the other hand, as a basis of explanation for some results, obtained from the youngest germs of some other animal species, which otherwise would seem to be rather irreconcilable with what our Echinus teaches us.

You know what a gastrula of our sea-urchin is. If you bisect this gastrula, when it is completely formed, or, still better, if you bisect the gastrula of the starfish, either along the axis or at right-angles to it, you get complete little organisms developed from the parts: the ectoderm is formed in the typical manner in the parts, and so is the endoderm; everything is proportionate and only smaller than in the normal case. So we have at once the important results, that, as in the blastula, so in the ectoderm and in the endoderm of our Echinus or of the starfish, the prospective potencies are the same for every single element: both in the ectoderm and in the endoderm the prospective value of each cell is a “function of its position” (Fig. 9).

But a further experiment has been made on our gastrula. If at the moment, where the material of the future intestine is most distinctly marked in the blastoderm, but not yet grown into a tube—if at this moment the upper half of the larva is separated from the lower by an equatorial section, you will get a complete larva only from that part which bears the “Anlage” of the endoderm, while the other half will proceed in morphogenesis very well but will form only ectodermal organs. By another sort of experiment, which we cannot fully explain here, it has been shown that the endoderm if isolated is also only able to form such organs as are normally derived from it.

And so we may summarise both our last results by saying: though ectoderm and endoderm have their potencies equally distributed amongst their respective cells, they possess different potencies compared one with the other. We may also say that they are equipolential in themselves, but of different potencies compared with each other.

So much for the conclusions that could be drawn from my own experiments. In recent years Spemann and his followers have shown in the embryo of the newt, v. Ubisch whilst working with the sea-urchin, that the equipotentiality of the early embryonic cells goes much further: the cells of the so-called germ

Fig. 9.—The Starfish, Asterias.
a1. Normal gastrula; may be bisected along the main axis or at right-angles to it (see dotted lines).
a1. Normal larva, “Bipinnaria”.
b1. Small but whole gastrula that results by a process of regulation from the parts of a bisected gastrula.
b1. Small but wholeBipinnaria”, developed out of b1.

layers may replace one another under certain conditions, in any case so long as they have not yet acquired their definite histological structure, even though they are already in their typical position.

We shall come back to Spemann’s experiments in another paragraph and therefore only note in this place that, in spite of them, certain potential restrictions of embryonic cells exist beyond any doubt, at least in later stages. But these potential restrictions may be due to mere secondary circumstances.

But those of you who are familiar with morphogenetic facts will object to me, that what we have stated about potential restrictions in ontogeny is not true in any case, and you will censure me for having overlooked regeneration, adventitious budding, and so on. To some extent the criticism would be right, but I am not going to recant; I shall only introduce another new concept. We are dealing only with primary potencies in our present considerations, i.e. with potencies which lie at the root of true embryology, not with those serving to regulate disturbances of the organisation. It is true, we have in some way disturbed the development of our sea-urchin’s egg in order to study it; more than that, it would have been impossible to study it at all without some sort of disturbance, without some sort of operation. But, nevertheless, no potencies of what may properly be called the secondary or restitutive type have been aroused by our operations; nothing happened except on the usual lines of organogenesis. It is true, some sort of regulation occurred, but that is included among the factors of ontogeny proper.

We shall afterwards study more fully and from a more general point of view this very important feature of “primary regulation” in its contrast to “secondary regulation” phenomena. At present it must be enough to say that in speaking of the restriction of the potencies in form-building we refer only to potencies of the primary type, which contain within themselves some properties of a (primary) regulative character.

The Morphogenetic Function of Maturation in the Light of Recent Discoveries

Turning again to more concrete matters, we shall first try, with the knowledge acquired of the potencies of the blastoderm and the so-called germ layers of Echinus, to understand certain rather complicated results which the experimental morphogenetic study of other animal forms has taught us. We know from our historical sketch that there are some very important aberrations from the type to which the Echinus germ belongs,15 i.e. the type with an equal distribution of the potencies over all the blastomeres. We know not only that in cases where a regulation of the intimate structure of the protoplasm fails to occur a partial development of isolated cells will take place, but that there may even be a typical disposition of typical cells for the formation of typical organs only, without any regulability.

Let us first consider the last case, of which the egg of mollusca is a good type: here there is no equal distribution of potencies whatever, the cleavage-cells of this germ are a sort of real “mosaic” with regard to their morphogenetic potentialities. Is this difference between the germ of the echinoderms and the molluscs to remain where it is, and not to be elucidated any further? Then there would be rather important differences among the germs of different animals, at least with regard to the degree of the specification of their cleavage cells; or if we ascribe differences among the blastomeres to the organisation of the fertilised egg ready for cleavage, there would be differences in the morphogenetic organisation of the egg-protoplasm: some eggs would be more typically specialised at the very beginning of morphogenesis than others.

In the first years of experimental embryology I pointed out that it must never be forgotten that the egg itself is the result of organogenesis. If, therefore, there are real mosaic-like specifications in some eggs at the beginning of cleavage, or during it, there may perhaps have been an earlier stage in the individual history of the egg which did not show such specifications of the morphogenetic structure. Two American authors share the merit of having proved this hypothesis. Conklin showed that certain intracellular migrations and rearrangements of material do happen in the first stages of ovogenesis in certain cases, but it is to E. B. Wilson16 that science owes a proper and definitive elucidation of the whole subject. Wilson’s researches, pursued not only by descriptive methods,17 but also by means of analytical experiment, led him to the highly important discovery that the eggs of several forms (nemertines, molluscs), which after maturation show the mosaic type of specification in their protoplasm to a more or less high degree, fail to show any kind of specification in the distribution of their potencies before maturation has occurred. In the mollusc egg a certain degree of specification is shown already before maturation, but nothing to be compared with what happens afterwards; in the egg of nemertines there is no specification at all in the unripe egg.

Maturation thus becomes a part of ontogeny itself; it is not with fertilisation that morphogenesis begins—there is a sort of ontogeny anterior to fertilisation.

These words constitute a summary of Wilson’s researches. Taken together with the general results obtained about the potencies of the blastula and the gastrula of Echinus, they reduce what appeared to be differences of degree or even of kind in the specification of the egg-protoplasm to mere differences in the time of the beginning of real morphogenesis. What occurs in some eggs, as in those of Echinus, at the time of the definite formation of the germ layers, leading to a specification and restriction of their prospective potencies, may happen very much earlier in other eggs. But there exists in every sort of egg an earliest stage, in which all parts of its protoplasm are equal as to their prospectivity, and in which there are no potential diversities or restrictions of any kind.

So much for differences in the real material organisation of the germ and their bearing on inequipotentialities of the cleavage cells.

The Intimate Structure of Protoplasm: Further Remarks

Where a typical half-or quarter-development from isolated blastomeres happens to occur, we know already that the impossibility of a regulation of the intimate polar-bilateral structure may account for it. As this impossibility of regulation probably rests on rather simple physical conditions,18 it may properly be stated that equal distribution of potencies is not wanting but is only overshadowed here. In this respect there exists a logical difference of fundamental importance between those cases of so-called “partial” or (better) “fragmental” development of isolated blastomeres in which a certain embryonic organ is wanting, on account of its specific morphogenetic material being absent, and those cases in which the “fragmental” embryo lacks complete “halves” or “quarters” with regard to general symmetry, on account of the symmetry of its intimate structure being irregularly disturbed. This logical difference has not always received the attention which it undoubtedly deserves. Our hypothetical intimate structure in itself is, of course, also a result of factors concerned in ovogenesis. Only in one case do we actually know anything about its origin: Roux has shown that in the frog it is the accidental path of the fertilising spermatozoon in the egg which, together with the polar axis, normally determines the plane of bilateral symmetry; but this symmetry may be overcome and replaced by another, if gravity is forced to act in an abnormal manner upon the protoplasm—the latter showing parts of different specific gravity in the eggs of all Amphibia.

The Neutrality of the Concept of “Potency”

Now we may close our rather long chapter on the distribution of potencies in the germ; it has been made long, because it will prove to be very important for further analytical discussion; and its importance, in great measure, is due to its freedom from prepossessions. Indeed, the concept of prospective potency does not prejudice anything. We have said, it is true, that limitations of potencies may be due to the presence of specific parts of organisation in some cases; that, at least, they may be connected therewith. But we have not determined at all what a prospective potency really is, what the term really is to signify. It may seem that such a state of things gives an air of emptiness to our discussions, that it leaves uncertain what is the most important. But, I think, our way of argument, which tries to reach the problems of greatest importance by degrees, though it may be slow, could hardly be called wrong and misleading.


We now proceed to an analysis of what may properly be called the means of morphogenesis, the word “means” being preferable to the more usual one “conditions” in this connection, as the latter would not cover the whole field. It is in quite an unpretentious and merely descriptive sense that the expression “means” should be understood at present; what is usually called “conditions” is part of the morphogenetic means in our sense. We may say that morphogenetic means are circumstances which are necessary for a complete morphogenesis.

The Internal Elementary Means of Morphogenesis

We know that all morphogenesis, typical or atypical, primary or secondary, goes on by one morphogenetic elementary process following the other. Now, the very foundation of these elementary processes themselves lies in the elementary functions of the organism as far as they result in the formation of stable visible products. Therefore the elementary functions of the organism may properly be called the internal means of morphogenesis.

Secretion and migration are among such functions; and the same is true of cell division and growth and, last though not least, of the aggregative state of the organic substance in general, including so-called surface tension and osmotic pressure. But let us observe that the elementary means of morphogenesis are far from being morphogenesis themselves. The word “means” itself implies as much. It would be possible to understand each of these single acts in morphogenesis as well as anything, and yet to be as far from understanding the whole as ever. All means of morphogenesis are only to be considered as the most general frame of events within which morphogenesis occurs.

We must be cautious in admitting that any organic feature has been explained, even in the most general way, by the action of physical forces. What at first seems to be the result of mechanical pressure may afterwards be found to be an active process of growth, and what at first seems to be a full effect of capillarity among homogeneous elements may afterwards be shown to depend on specialised metabolic conditions of the surfaces as its principal cause.19 All these processes are only means of the organism, and can never do more than furnish the general type of events. They do not constitute life—they are used by life; let it remain an open question, for the present, how the phenomenon of “life” is to be regarded in general.

We do not by any means intend to discredit a thorough and detailed study of, say, osmosis or colloid chemistry in their relation to morphogenesis. On the contrary, we highly appreciate the results of such investigations, as they serve to formulate very sharply the central biological problem. But they never give us a solution of this problem.

Let us close this chapter with a few words on cell division.

The investigations of the last few years have made it quite clear that even in organisms with a high power of morphogenetic regulation it is always the form of the whole, but not the individual cell, which is subjected to the regulation processes. Starting from certain results obtained by T. H. Morgan, I was able to show that in all the small but whole larvae, reared from isolated blastomeres, the size of the cells remains normal, only their number being reduced; and Boveri has shown most clearly that it is always the size of the nucleus—more correctly, the mass of the chromatin—which determines how large a cell of a certain histological kind is to be. In this view, the cell appears even more as a sort of material used by the organism as supplied, just as workmen can build the most different buildings with stones of a given size.

The External Means of Morphogenesis

We now know what internal means of morphogenesis are, and so we may glance at some of the most important “outer means” or “conditions” of organisation.

Like the adult, the germ also requires a certain amount of heat, oxygen, and, when it grows up in the sea, salinity in the medium. For the germ, as for the adult, there exists not only a minimum but also a maximum limit of all the necessary factors of the medium; the same factor which at a certain intensity promotes development, disturbs it from a certain other intensity upwards.

Within the limits of this minimum and this maximum of every outside agent there generally is an increase in the rate of development corresponding to the increase of intensity of the agent. The acceleration of development by heat has been shown to follow the law of the acceleration of chemical processes by a rise of temperature; that seems to prove that certain chemical processes go on during the course of morphogenesis.

Almost all that has been investigated of the part played by the external conditions of development has little bearing on specific morphogenesis proper, and therefore may be left out of account here: we must, however, lay great stress on the general fact that there is a very close dependence of morphogenesis on the outside factors, lest we should be accused afterwards of having overlooked it.

Of course, all “external” means or conditions of morphogenesis can actually relate to morphogenetic processes only by becoming in some way “internal”, but we unfortunately have no knowledge whatever how this happens. We at present are only able to ascertain what must necessarily be accomplished in the medium, in order that normal morphogenesis may go on, and we can only suppose that there exist certain specific internal general states, indispensable for organogenesis but inaccessible to present modes of investigation.

There are but few points in the doctrine of the external means or conditions of organogenesis which have a more special bearing on the specification of proper form, and which therefore require to be described here a little more fully. All these researches, which have been carried out almost exclusively by Herbst,20 relate to the effect of the chemical components of sea-water upon the development of the sea-urchin. If we select the most important of Herbst’s results, we must in the first place say a few words on the part taken by lime or calcium, not only in establishing specific features of form, in particular of the skeleton, but in rendering individual morphogenesis possible at all. Herbst has found that in sea-water which is deprived of calcium the cleavage cells and many tissue cells also completely lose contact with each other: cleavage goes on quite well, but after each single division the elements are separated; at the end of the process you find the 808 cells of the germ together at the bottom of the dish, all swimming about like infusoria. There seems to be some influence of the calcium salts upon the physical state of the surfaces of the blastomeres.

It is not without interest to note that this discovery has an important bearing on the technical side of all experiments dealing with the isolation of blastomeres. Since the separation of the single cleavage elements ceases as soon as the germs are brought back from the mixture without lime into normal sea-water, it of course is possible to separate them up to any stage which it is desired to study, and to keep them together afterwards. Thus, if, for instance, you want to study the development of isolated cells of the eight-cell stage, you will leave the egg in the artificial mixture containing no calcium until the third cleavage, which leads from the four-to the eight-cell stage, is finished. The single eight cells brought back to normal sea-water at this point will give you the eight embryos you want. All researches upon the development of isolated blastomeres since the time of Herbst’s discovery have been carried out by this method, and it would have been quite impossible by the old method of shaking to pursue the study into such minute detail as actually has been done.

Among all the other very numerous studies of Herbst, we need only mention that potassium is necessary for the typical growth of the intestine, just as this element has been found necessary for normal growth in plants, and that there must be the ion SO4, or, in other terms, sulphur salts, present in the water, in order that the germs may acquire their pigments and their bilateral symmetry. This is indeed a very important result, though it cannot be said to be properly understood. It is a fact that in water without sulphates the larvae of Echinus retain the radial symmetry they have had in the very earliest stages, and may even preserve that symmetry on being brought back to normal sea-water if they have spent about twenty-four hours in the artificial mixture.

We may now leave the subject of Herbst’s attempts to discover the morphogenetic function of the single constituents of normal sea-water, and may devote a few words to the other branch of his investigations, those dealing with the morphogenetic effects of substances which are not present in the water of the sea, but have been added to it artificially. Here, among many other achievements, Herbst has made the most important discovery that all salts of lithium effect radical changes in development.21 I cannot describe fully here how the so-called “lithium larva” originates; let me only mention that its endoderm is formed outside instead of inside, that it is far too large, that there is a spherical mass between the ectodermal and the endodermal part of the germ, that a radial symmetry is established in place of the normal bilateralism, that no skeleton exists, and that the mesenchyme cells are placed in a quite abnormal position. All these features, though abnormal, are typical of the development in lithium. The larvae present no really pathological appearance at all, and, therefore, it may indeed be said that lithium salts are able to change fundamentally the whole course of morphogenesis. It detracts nothing from the importance of these discoveries that, at present, they stand quite isolated: only with lithium salts has Herbst obtained such strange results; and only upon the eggs of echinids, not even upon those of asterids, do lithium salts act in this specific way.


The Definition of Cause

We cannot begin the study of the “causes” of the differentiation of form without a few words of explanation about the terminology which we shall apply. Causality is a very disputed concept; many modern scientists, particularly in physics, try to avoid the concept of cause altogether, and to replace it by mere functional dependence, in the mathematical meaning of the term. They claim to express completely by an equation all that is discoverable about any sort of phenomena constantly connected.

I cannot convince myself that such a very restricted view is the right one: it is very cautious, no doubt, but it is incomplete, for the concept of the acting “cause” is a legitimate concept and we are forced to search for applications of it in Nature. On the other hand, it does not at all escape me that there are many difficulties, or rather ambiguities, in applying it.

We may call the “cause” of any event, the sum total of all the constellations of facts which must be completed in order that the event may occur; it is in this meaning, for instance, that the first principle of energetics applies the term in the words causa aequat effectum. But, by using the word only in this very general sense, we deprive ourselves of many conveniences in the further and more particular study of Nature. Would it be better to say that the “cause” of any event is the very last change which, after all the constellations necessary for its start are accomplished, must still take place in order that the event may actually occur? Let us see what would follow from such a use of the word causality. We here have an animal germ in a certain stage, say a larva of Echinus, which is just about to form the intestine; all the internal conditions are fulfilled, and there is also a certain temperature, a certain salinity, and so on, but there is no oxygen in the water: the intestine, of course, will not grow in such a state of things, but it soon will when oxygen is allowed to enter the dish. Is, therefore, oxygen the “cause” of the formation of the intestine of Echinus? Nobody, I think, would care to say so. By such reasoning, indeed, the temperature, or sodium, might be called the “cause” of any special process of morphogenesis. It, therefore, seems to be of little use to give the name of cause to that factor of any necessary constellation of events which accidentally happens to be the last that is realised. But what is to be done then?

Might we not say that the cause of any morphogenetic process is that typical property, or quality, or change on which its specific character depends—on which depends, for example, the fact that now it is the intestine which appears, while at another time it is the lens of the eye? We might very well, but we already have our term for this sort of cause, which is nothing else than our prospective potency applied to that elementary organ from which the new process takes its origin. The prospective potency indeed is the truly immanent cause of every specification affecting single organogenetic processes. But we want something more than this.

We may find what we want by considering that each single elementary process or development not only has its specification, but also has its specific and typical place in the whole—its locality. Therefore we shall call the “cause” of a single morphogenetic process, that occurrence on which depends its localisation, whether its specific character also partly depends on this “cause” or not.22

This definition of “cause” in morphology may be artificial; in any case it is clear. And at the same time the concepts of the prospective potency and of the “means” of organogenesis now acquire a clear and definite meaning: potency is the real basis of the specific character of every act in morphogenesis, and “means”, including conditions, are the sum of all external and internal general circumstances which must be present in order that morphogenetic processes may go on, without being responsible for their specificity or localisation.

It is implied in these definitions of cause and potency, that the former almost always will be of that general type which usually is called a stimulus. There is no quantitative correspondence between our “cause” and the morphogenetic effect.

Some Instances of Formative and Directive Stimuli

Again it is to Herbst that we owe not only a very thorough logical analysis of what he calls formative and directive stimuli23 but also some important discoveries on this subject. We cannot do more here than barely mention some of the most characteristic facts.

Amongst plants it has long been known that the direction of light or of gravity may determine where roots or branches or other morphogenetic formations are to arise; in hydroids also we know that these factors of the medium may be at work24 as morphogenetic causes, though most of the typical architecture of hydroid colonies certainly is due to internal causes, as is also much of the organisation in plants.

Light and gravity are external formative causes; beside that they are merely “localisers”. But there also are some external formative stimuli, on which depends not only the place of the effect, but also part of its specification. The galls of plants are the most typical organogenetic results of such stimuli. The potencies of the plant and the specific kind of the stimulus equally contribute to their specification; for several kinds of galls may originate on one sort of leaves.

No exterior formative stimuli are responsible for the intimate details of animal organisation; and one would hardly be wrong in saying that this morphogenetic independence in animals is due to their comparatively far-reaching functional independence of those external agents which have any sort of direction. But many organogenetic relations are known to exist between the single parts of animal germs, each of these parts being in some respect external to every other; and, indeed, it might have been expected already a priori that such formative relations between the parts of an animal embryo must exist, after all we have learned about the chief lines of early embryology. If differentiation does not go on after the scheme of Weismann, that is, if it is not carried out by true “evolutio” from within, how could it be effected except from without? Indeed, every embryonic part may, in some respect, be a possible cause for morphogenetic events which are to occur on every other part: it is here that the very roots of epigenesis are to be found.

Heliotropism and geotropism are among the well-known physiological functions of plants: the roots are seen to bend away from the light and towards the ground; the branches behave just in the opposite way. It now has been supposed by Herbst that such “directive stimuli” may also be at work among the growing or wandering parts of the embryo, that their growth or their migration may be determined by the typical character of other parts, and that real morphogenetic characters can be the result of some such relation; a sort of “chemotropism” or “chemotaxis” may be at work here. Herbst himself has discussed theoretically several cases of organogenesis in which the action of directive stimuli is very probable.

What has become actually known by experiment is not very much at present: the mesenchyme cells of Echinus are directed in their migration by specified places in the ectoderm (Driesch), the pigment cells of the yolk-sac of the fish Fundulus are attracted by its blood-vessels (J. Loeb), and the nerves of amphibian embryos, growing out from the ganglion cells of the central nervous system (His, Harrison), are attracted in some way on their path by the organs which need innervation. Transplantation experiments (Detwiler, Weiss) have proved, however, that innervation may occur in very abnormal ways, an implanted leg, for example, being innervated by the trigeminous or facialis nerve, according to circumstances. The attractive relation among peripheral organs and nerves, therefore, is by no means specific.

The first case of an “internal formative stimulus” in the proper sense, that is, of one embryonic part causing another to appear, was discovered by Herbst himself. The arms of the so-called pluteus of the sea-urchin are in formative dependence on the skeleton—no skeleton, no arms; so many skeleton primordia,25 in abnormal cases, so many arms; abnormal position of the skeleton, abnormal position of the arms: these three experimental observations form the proof of this morphogenetic relation.

It may be simple mechanical contact, or it may be some chemical influence that really constitutes the “stimulus” in this case; certainly, there exists a close and very specific relation of the localisation of one part of the embryo to another. Things are much the same in another case, which, after having been hypothetically stated by Herbst on the basis of pathological data, was proved experimentally by Spemann. The lens of the eye of certain Amphibia is formed of their skin in response to a formative stimulus proceeding from the so-called primary optical vesicle. If this vesicle fails to touch the skin, no lens appears; and, on the other hand, the lens may appear in quite abnormal parts of the skin if they come into contact with the optic vesicle after transplantation.

But formative dependence of parts may also be of different types.

We owe to Herbst the important discovery that the eyes of crayfishes, after being cut off, will be regenerated in the proper way if the optic ganglion is present, but that an antenna will arise in their place if this ganglion has also been removed. There must in this case be some unknown influence of the formative kind on which depends, if not regeneration itself, at least its special character.

In other cases there seems to be an influence of the central nervous system on the regenerative power in general. Amphibia, for instance, are said to regenerate neither their legs (Wolff) nor their tail (Godlewski), if the nervous communications have been disturbed. But this influence is not of a specific nature (Weiss), and in other animals there is no such influence at all; in yet others, as for instance, in Planarians, it must seem doubtful at present whether the morphogenetic influence of the nervous system upon processes of restoration is more than indirect; the movements of the animal, which become very much reduced by the extirpation of the ganglia, being one of the main conditions of a good regeneration.

Of course, all we have said about the importance of special materials in the ripe germ, as bearing on specifically localised organisations, might be discussed again in our present chapter, and our intimate polar-bilateral structure of germs may also be regarded as embracing formative stimuli, at any rate as far as the actual poles of this structure are concerned. This again would bring us to the problem of so-called “polarity” in general, and to the “inversion” of polarity, that is, to a phenomenon well known in plants and in many hydroids and worms, viz., that morphogenetic processes, especially of the type of restitutions, occur differently, according as their point of origin represents, so to speak, the positive or the negative, the terminal or the basal end of an axis, but that under certain conditions the reverse may also be the case. But a fuller discussion of these important facts would lead us deeper and deeper into the science of morphogenesis proper, without being of much use for our future considerations.26

A few words may be devoted to the problem of the determination of sex, which, according to the latest researches, seems to depend on cytological events occurring in the very earliest embryonic stages, say even before ontogeny, and not on formative stimuli proper. It seems, indeed, as if the sexual products themselves would account for the sex of the individual produced by them, particularly if there were differences in their chromatin. But the problem is by no means solved. In the worm Bonellia external conditions do certainly play a rôle in sex determination (Baltzer, Herbst).

The influence of the sexual glands on the so-called secondary sexual characters of Vertebrates is beyond any doubt, as the transplantation experiments of Steinach have shown. But this statement must not be generalised and is certainly not true for Arthropods (Meisenheimer), where the secondary characters develop quite independently of the sex-glands. It is important to notice that, in Vertebrates, not the sex—cells themselves but certain intermediate cells are the source of the formative influence in question.

Very important additions to the theory of formative stimuli have recently been made by Spemann and his school (Mangold, Goerttler, etc.), by the aid of transplantation of embryonic parts, either in one and the same individual or from one individual to the other. These experiments, all of them carried out in embryos of the newt (Triton), have revealed quite an enormous variety of formative dependences. The authors speak of homœ;ogeneous induction, if an implanted part makes other cells equal to itself (morphological assimilation), whilst they call heterogeneous induction the action of formative stimuli proper. The roof of the primordial intestine, e.g., induces heterogeneously the formation of the axial nervous system, whilst cells of this system, when abnormally transplanted, increase in number by morphogenetic assimilation.

The intimate nature of the morphogenetic influence is unknown. In any case, no nervous influence is in question. The concept of hormon (Starling), of so much use in physiology proper, may be applied here.

The ultimate source of all these formative influences is called organiser by Spemann,27 and, in Triton, certain cells near the mouth of the gastrula constitute the organiser kat’ exochen. The original organiser may induce secondary ones, etc. Later on we shall raise the question whether or not the sequence of organisers may in fact be dissolved into a play of single formative acts.

We have spoken of the morphogenetic potencies in a former paragraph, and may repeat in this place that the discoveries of the Spemann school have shown the embryo of Triton to be equipotential in quite an unexpected degree. Not only is every elementary organ equipotential in itself, as we already knew, but the so-called germ layers—in their “presumptive” state at least, i.e. as long as they have not yet reached their typical histological structure—are equipotential with regard to one another. “Presumptive” cells of the ectoderm may become mesoderm, etc.

It should well be kept in mind that this equipotentiality refers to what we have called primary potencies (page 58), and not to the secondary ones, which serve for regeneration.


The duration of individual life is specific for each living species, and the same is true for the duration of the embryo-logical process, the “speed” of embryology, as we may call it. Both statements require the addition of the words ceteris paribus, for cold-blooded animals live longer in a low than in a high temperature (J. Loeb), and the duration of embryology depends on temperature, amount of oxygen, etc. But under equal conditions even two different species of sea-urchins differ in the speed of their embryology.

Steinach has shown that in vertebrates the duration of life may be increased by certain manipulations with the sexual glands; and Gudernatsch has discovered that the metamorphosis of tadpoles occurs much earlier than normally if they are fed with materials taken from the thyreoidea; whilst, according to Adler, no metamorphosis occurs, if the hypophysis or the thyreodea is extirpated, or if the tadpoles are fed with thymus material. All these results have been confirmed and elaborated in detail. But not very much is really understood in these cases; something of a chemical nature seems to play its rôle.

What seems to me to be most important with regard to the problem of embryology and time is the fact that every single embryological process occupies a particular and specific temporal position within a well-ordered sequence of events, i.e. a particular relative moment in this sequence at which it is due. It neither takes place earlier nor later—both words to be understood in a relative sense—and if, for some reason, it does not take place when due, it will never occur.

Herbst, e.g., has oppressed the growth of the intestine in the gastrula of echinoderms by eliminating the element kalium from the sea-water. Everything else, then, goes on quite normally as far as the ectodermal organs are concerned. But there never will be any endoderm, even if the larvae are put back into normal sea-water.

Even the single steps in cleavage are due at specific relative moments. According to Boveri a whole normal egg may occasionally segment, as if it were one of the two or four first cleavage cells. In this case there have been nuclear divisions without any protoplasmic division in the egg. Let us assume that the two first divisions have affected the nucleus exclusively, whilst the third division has led to two equal cells. Then the next division will not lead to a stage of four equal cells, but will give us two mesomeres, one macromere and one micromere. For at the fourth nuclear division micromeres are “due” (cp. p. 21).

This phenomenon of “being due” requires further investigation. And so does the question whether an embryonic process n happens to be realised because the process n − 1 has occurred or because the result of this process does exist, which, of course, is not the same. To this problem we shall come back later.

Certain observations in the course of Spemann’s and Mangold’s experiments (cp. p. 73 f.) tend in the same direction: implanted cells do not show their formative influence upon the cells of the host if the process in question is not “due”, i.e. if the host is “too old” already—though it may be a little “too young”. And this agrees with what Uhlenhuth and Kornfeld have called “synchronic metamorphosis”: if you implant embryonic eyes or gills of a salamander upon a host of greater age, they will change their structure according to the age of the host. The host gives the temporal rule.

To sum up: it seems as if there were a continuous stream of becoming at the basis of the embryological process, of which the visible embryonic phases are nothing but indexes. And this ultimate continuous foundation seems to give its temporal position to every single embryological performance.

All this relates to embryology proper exclusively. It does not relate to “secondary regulations” nor to the morphogenesis of “open forms” (p. 31) as, e.g., the plants. The periodicity in the origin and the fall of the leaves of our trees, according to Klebs, has no internal but only external reasons, and the same is true for the alternation of the various generations of fungi and in many other botanical cases.


Let us now turn to considerations of a more abstract kind: we have become acquainted with some morphogenetic interactions among the parts of a developing embryo; and, indeed, we can be sure that there exist far more of such interactions than we know at present.

But it is far from being true that the development of each embryonic part depends on the existence or development of every other one.

On the contrary, it is a very important and fundamental feature of organogenesis that it occurs in separate lines, that is to say, in lines of processes which may start from a common root, but which are absolutely independent of one another in their manner of differentiation. Roux has coined the term self-differentiation to denote this phenomenon, and we admit that this term may be conveniently used for the purpose, if only it can be kept in mind that its sense is always relative,

Fig. 10.—PluteusLarva of Sphaerechinus.
The Intestine (i) is developed outside instead of inside (by means of raising the temperature); but the mouth (r) is formed in its normal place. S = Skeleton.

and that it is also negative. Suppose a part, A, shows the phenomenon of self-differentiation: this means that the further development of A is not dependent on certain other parts, B, C, and D; it does not mean at all that A has not been formatively dependent on some other parts, E or F, at the time of its first appearance, nor does it imply that there might not be many formative actions among the constituents of A itself.

We indeed are entitled to say that the ectoderm of Echinus shows “self-differentiation” with regard to the endoderm; it acquires its mouth, for instance, as has been shown by experiment, even in cases where no intestine is present at all (Fig. 10); but ectoderm and endoderm both are formatively dependent on the intimate and the material organisation of the blastoderm. In the same way, in the amphibian embryo, the formation of the “Anlage” of the legs and of the central nervous system are mutually independent.

The phenomenon of self-differentiation, properly understood, now may help to the discovery of one most general character of all development. If the phenomenon of self-differentiation really occurs in ontogeny in its most different aspects, and if, on the other hand, in spite of this relative morphogenetic independence of embryonic parts, the resulting organism is one whole in organisation and in function, some sort of harmony of constellation, as it may properly be styled, must be said to be one of the most fundamental characters of all production of individual form. In establishing this harmony, we do nothing more than describe exactly what happens: the harmony is shown by the fact that there is a whole organism at the end, in spite of the relative independence of the single events leading to it.

But still another sort of harmony is revealed in morphogenesis, by an analysis of the general conditions of the formative actions themselves. In order that these actions may go on properly, the possibility must be guaranteed that the formative causes may always find something upon which to act, and that those parts which contain the potencies for the next ontogenetic stage may properly receive the stimuli awaking these potencies: otherwise there would be no typical production of form at all. This, the second species of harmonious relations to be described in ontogeny, may be called causal harmony; the term simply expresses the unfailing relative condition of formative causes and cause-recipients.

Finally, in functional harmony we have an expression descriptive of the unity of organic function, and so we may state, as the latest result of our analytical theory of development up to this point, that individual morphogenesis is marked by a threefold, harmony among its parts.

The three sides or parts of this threefold harmony are united in a very close way. They penetrate one another, as it were. For the embryological process lays the foundations of all the future functions of the adult; and, on the other hand, that process itself depends on a harmony of functions, every embryonic stage being also a functioning totality. We must never forget that the organism is in each of its stages a unity of form and of function, these two features being dependent on one another.


At this stage we leave for a while our analytical studies of ontogeny proper. We must not forget that typical ontogenesis is not the only form in which morphogenesis can occur: the organic form is able to restore disturbances of its organisation, and it certainty is to be regarded as one of the chief problems of analytical morphogenesis to discover the specific and real stimulus which calls forth the restoring processes. For simply to say that the disturbance is the cause of the restoration would be to evade the problem instead of attacking it. But there are still some other problems peculiar to the doctrine of restitutions.

A few Remarks on Secondary Potencies and on Secondary Morphogenetic Regulations in General

We have only briefly mentioned in a previous chapter (page 58) that there exist many kinds of potencies of what we call the secondary or truly restitutive type, and that their distribution may be most various and quite independent of all the potencies for the primary processes of ontogeny proper. Let us first add a few words about the concept of “secondary restitution” and about the distribution of secondary potencies in general.

Primary ontogenetic processes founded upon primary potencies may imply regulation, or, more correctly, restitution in many cases: so it is when fragments of the blastula form the whole organism or when the mesenchyme cells of Echinus reach their normal final position by an attraction on the part of specific localities of the ectoderm in spite of a very abnormal original position enforced upon them by experiment. In these cases we speak of primary regulations or restitutions; disturbances are neutralised by the very nature of the process in question. We speak of secondary restitution whenever a disturbance of organisation is rectified by processes foreign to the realm of normality; and these abnormal lines of events are revealed to us in the first place by the activity of potencies which remain latent in ontogeny proper.

We know already that a certain kind of secondary restitution has been discovered, very contradictory to the theoretical views of Weismann; the process of restoration being carried out not by a definite part of the disturbed organisation, but by all the single elements of it. The problem of the distribution of secondary potencies in these cases of so-called “re-differentiation” is to form our special study in the next chapter.

In all other cases restoration processes start from specific localities; if they occur on the site of the wound which caused the disturbance, we speak of regeneration; if they occur at some distance from the wound, we call them adventitious processes.

Besides these three types of processes of restitution, there may be mentioned a fourth one, consisting in what is generally called compensatory hypertrophy; the most simple case of such a compensatory process is when one of a pair of organs, say a kidney, becomes larger after the other has been removed. But real compensatory differentiation occurs in the cases of so-called “hypertypy” as first discovered by Przibram and afterwards studied by Zeleny: here the two organs of a pair show a different degree of differentiation. Whenever the more specialised organ is removed the less developed one assumes its form. Similar cases, which might simply be called “compensatory heterotypy”, are known in plants, though only relating to the actual fate of undifferentiated “Anlagen” in these organisms. A leaf may be formed out of the “Anlage” of a scale, if all the leaves are cut off, and so on.

Finally, at least in plants, a change of the directive irritability, of so-called “geotropism” for instance, in certain parts may serve to restore other more important parts.

In two of these general types of restitution, in regeneration proper and in the production of adventitious organs, the potencies which underlie these processes may be said to be “complex”. It is a complicated series of events, a proper morphogenesis in itself, for which the potency has to account, if, for instance, a worm newly forms its head by regeneration, or if a plant restores a whole branch in the form of an adventitious bud.

Such generalisations as are possible about the distribution of complex potencies are also reserved for a special part of our future discussion.

Secondary restitution is always, like ontogeny, a process of morphogenesis, and therefore all the questions about single formative stimuli, and about internal and external conditions or means, occur again. But of course we cannot enter into these problems a second time, and may only say that, especially in regeneration proper, the specific type of the regenerative formation of any part may differ very much from the ontogenetic type of its origin: the end of both is the same, but the way can be even fundamentally different in every respect.

The Stimuli of Restitutions29

But now we turn to the important question: what is the precise stimulus30 that calls forth processes of restitution; or, in other words, what must have happened in order that restitution may occur?

That the operation in itself, by its removing of mechanical obstacles, cannot be the true stimulus of any restitutions, is simply shown by all those restitutions that do not happen at the place of the wound. If we took a narrower point of view, and if we only considered regeneration proper from the wound itself, we might probably at first be inclined to advocate the doctrine that the removing of some obstacles might in fact be the stimulus to the process of restoration; but, even then, why is it that just what is wanted grows out? Why is there not only growth, but specific growth, growth followed by specification? The removing of an obstacle could hardly account for that. But, of course, taking account of all the adventitious restitutions—that is, all restorations not beginning at the wound itself—the theory that the removing of obstacles is the stimulus to restoration becomes, as we have said, quite impossible.

But where then is the stimulus to be found? There is another rather simple theory of the “Auslösung” of restitutions, which starts from the phenomena of compensatory hypertrophy and some occurrences among plants. The removal of some parts of the organism, it is said, will bring its other parts into better conditions of nutrition, and therefore these parts, particularly if they are of the same kind, will become larger. Granted for the moment that such a view may hold in cases when one of a pair of glands becomes larger after the other has been removed, or when pruning of almost all the leaves of a tree leads to the rest becoming larger, it certainly must fail to explain the fact that in other cases true new formations may arise in order to restore a damaged part, or that the latter may be regenerated in its proper way. For merely quantitative differences in the mixture of the blood or of the nourishing sap in plants can never be a sufficient reason for the highly typical and qualitative structure of newly formed restitutions. And even in the most simple cases of a mere increase in the size of some parts, that is in the simplest cases of so-called compensatory hypertrophy, it is at least doubtful, if not very improbable, that the compensation is accomplished in such a purely passive way, because we know that in other cases it is usually the growth of the young parts that actively attracts the nourishment: there is, first, differentiation and growth, and afterwards there is a change in the direction of the nourishing fluids.

The process of true regeneration, beginning at the locality of the wound itself, has been shown by Morgan, even as regards its rate, to occur quite irrespectively of the animal being fed or not. There could hardly be a better demonstration of the fundamental fact that food assists restitution, but does not “cause” it in any way.

But in spite of all we have said, there seems to be some truth in regarding the nutritive juices of animals and plants as somehow connected with the stimulus of restitutions: only in this very cautious form, however, may we make the hypothesis. It has been shown, for both animals and plants, that morphogenesis of the restitutive type may be called forth even if the parts now to be “regenerated” have not been actually removed; e.g. in the so-called super-regeneration of legs and tails in Amphibia, of the head in Planarians, of the root-tip in plants and in some other cases. Here it has always been a disturbance of the normal connection of some parts with the rest of the organism which proved to be the reason of the new formation. This shows that something to do with the communication among parts is at least connected with restitution, and this communication may go on either by the unknown action of specific tissues or by the aid of the blood or sap. But in what this change or break of specific communication consists, is absolutely unknown. One might suppose that each part of the organisation constantly adds some sort of ferment to the body fluids outside or inside the cells, that the removing of any part will change the composition of these fluids in this particular respect, and that this change acts as a sort of communication to summon the restituting parts of the whole to do their duty.31

But I see quite well that such a theory is not very satisfactory; for what has to be done in restitution in each case is not a simple homogeneous act, for which one special material might account, but is a very complicated work in itself. It was the defect of the theory of “organ-forming substances” as advocated by Sachs, that it overlooked this point.

So all we know about the proper stimuli of restitutions is far from resting on any valid grounds at all. No doubt, there will be something discovered some day, and the idea of the “whole” in organisation will probably play some part in it. For two facts must be well kept in mind: first, that in all restitutions the cells in action always perform that which is necessary in this particular case in order to restore normal organisation; and secondly, that the same cells might perform something else if required. Does it not seem, then, as if the cells had some “knowledge” about the specificity of the disturbance? If so, there would be an analogy to—specific sensation.

But nothing more about this topic at present.

This is the first time that, hypothetically at least, the idea of the whole has entered into our discussion. The same idea may be said to have entered it already in a more implicit form in the statement of the threefold harmony in ontogeny.

Let us now see whether we can find the same problem of the “whole” elsewhere, and perhaps in more explicit and less hypothetical form. Let us see whether our analytical theory of development is in fact as complete as it seemed to be, whether there are no gaps left in it which will have to be filled up.

3. The Problem of Morphogenetic Localisation


We have come to the central point of the first part of this book; we shall try in this chapter to decide a question which is to give life its place in Nature, and biology its place in the system of sciences. One of the foundation-stones is to be laid, upon which our future philosophy of the organism will rest.

The General Problem

Our analytical theory of morphogenesis has been founded upon three elementary concepts—the prospective potency, the means, and the formative stimulus; and its principal object has been to show that all morphogenesis may be resolved into the three phenomena expressed by them. Have we indeed succeeded in attaining this object? Is it really possible to explain every morphogenetic event, at least in the most general way, by the aid of the terms potency, means, and stimulus?

All of these questions are apt to lead us to further considerations. Perhaps these considerations will give us a very clear and simple result by convincing us that it is indeed possible to analyse morphogenesis in our schematic way.

But if the answer were a negative one? What would that suggest?

The full analysis of morphogenesis into a series of single formative occurrences, brought about by the use of given means and on the basis of given potencies, might assure us, perhaps, that, though not yet, still at some future time, a further sort of analysis will be possible: the analysis into the elemental facts studied by the sciences of inorganic nature. The organism might prove to be a machine, not only in its functions but also in its very origin.

But what are we to say if even the preliminary analysis, which possibly might lead to such an ultimate result, fails?

Let us, then, set to work. Let us try to consider most carefully the topic in which our concept of the formative cause or stimulus may be said to be centred, the localisation of all morphogenetic effects. Is it always possible, in fact, to account for the typical localisation of every morphogenetic effect by the discovery of a single specific formative stimulus? You will answer me, that such an analysis certainly is not possible at present. But I ask you again, are there any criteria that it is possible, at least in principle; or are there any criteria which will render such an aim of science impossible for all future time?

The Morphogenetic “System”

We know from our experimental work that many, if not all, of the elementary organs in ontogeny show one and the same prospective potency distributed equally over their elements. If we now borrow a very convenient term from mechanics, and call any part of the organism which is considered as a unit from any morphogenetic point of view, a morphogenetic “system”, we may sum up what we have learnt by saying that both the blastoderm of the echinoderms, at least around its polar axis, and also the germ-layers of these animals, are “systems” possessing an equal potentiality in all of their elements, or, in short, that they are equipotential systems.

But such a term would not altogether indicate the real character of these systems.

Later on, we shall analyse, more carefully than before, the distribution of potencies which are the foundation both of regeneration proper and of adventitious growth, and then we shall see that, in higher plants for instance, there is a certain “system” which may be called the organ proper of restitutions, and which also in each of its elements possesses the same restoring potency; I refer to the well-known cambium. This cambium, therefore, also deserves the name of an “equipotential system”. But we know already that its potencies are of the complex type, that they consist in the faculty of producing the whole of such a complicated organisation as a branch or a root, that the term “equipotential system” is here only to signify that such a complicated unit may arise out of each of the cells of the cambium.

The potencies we have been studying in the blastula or gastrula of echinoderms are not of the complex type: our systems are equipotential to the extent that each of their elements may play every single part in the totality of what will occur in the whole system; it is to this single part that the term “function of the position” relates. We therefore might call our systems equipotential systems with single potencies; or, more shortly, singular-equipotential systems.

But even this terminology would fail to touch precisely the very centre of facts: it is not only the simplicity or singularity of their potencies which characterises the rôle of our systems in morphogenesis, but far more important with respect to the production of form are two other leading results of the experimental researches. The proper act to be performed by every element in each actual case is, in fact, a single one, but the potency of any element as such consists in the possibility of many single acts: that, then, might justify us in speaking of our systems as “indefinite equipotential”, were it not that another reason makes another title seem still more preferable. For the name of indefinite equipotential systems might also be applied to elementary organs, the single potencies of which are awaked to organogenesis by specific formative stimuli, as in the experiments carried out by Spemann and his school, where the reaction of one and the same cell varies according to the stimulus in question. But this is not the case in the system of which we are now speaking. There are, indeed, indefinite singular potencies at work in our systems during ontogeny; but what happens to arise in every case out of the totality of the single acts performed by all of the single equipotential cells is not a sum induced from without, but a unit guaranteed from within. That is to say, there exists a sort of inner harmony in every case among the real products of our systems, these products being due to the inner forces of the systems exclusively. The term harmonious-equipotential system, therefore, seems to be the right one to denote them.

We now shall try, first, to analyse to its very extremes the meaning of the statement that a morphogenetic system is harmonious-equipotential.

The “Harmonious-Equipotential System”

We have an ectoderm of the gastrula of a starfish here before us; we know that we may cut off any part of it in any direction, and that nevertheless the differentiation of the ectoderm may go on perfectly well and result in a typical little embryo, which is only smaller in its size than it would normally be. It is by studying the formation of the highly complicated ciliary band that these phenomena can be most clearly understood.

Now let us imagine our ectoderm to be a cylinder instead of being approximately a sphere, and let us imagine the surface of this cylinder unrolled. It will give us a plane of two definite dimensions, a and b. And now we have all the means necessary for the analytical study of the differentiation of an harmonious-equipotential system.

Our plane of the dimensions a and b is the basis of the normal, undisturbed development; taking the sides of the plane as fixed localities for orientation, we can say that the actual fate, the “prospective value” of every element of the plane stands in a fixed and definite correlation to the length of two lines, drawn at right-angles to the bordering lines of the plane; or, to speak analytically, there is a definite actual fate corresponding to each possible value of x and of y. Now, we have been able to state by our experimental work, that the prospective value of the elements of our embryonic organ is not identical with their “prospective potency”, or their possible fate, this potency being very much richer in content than is shown by a single case of ontogeny. What will be the analytical expression of such a relation?

Let us put the question in the following way: on what factors does the fate of any element of our system depend in all possible cases of development obtainable by means of operations? We may express our results in the form of an equation—

p.v. (X) = f (…),

i.e. “the prospective value of the element X is a function of…”—of what?

We know that we may take off any part of the whole, as to quantity, and that a proportionate embryo will result, unless the part removed is of a very large size. This means that the prospective value of any element certainly depends on, certainly is a function of, the absolute size of the actually existing part of our system in the particular case. Let s be the absolute size of the system in any actual experimental case of morphogenesis: then we may write p.v. (X) = f (s…). But we shall have to add still some other letter to this s.

The operation of section was without any restriction either as to the amount of the material removed from the germ or as to the direction of the cut. Of course, in almost every actual case there will be both a definite size of the actual system and a definite direction of the cut going hand-in-hand. But in order to study independently the importance of the variable direction alone, let us imagine that we have isolated at one time that part of our system which is bounded by the lines a1 b1, and at another time an equal amount of it which has the lines a2 b2 as its boundaries. Now, since in both cases a typical small organism may result on development, we see that, in spite of their equal size, the prospective value of every element of the two pieces cut out of the germ may vary even in relation to the direction of the cut itself. Our element, X, may belong to both of these pieces of the same size: its actual fate nevertheless will be different. Analytically, it may be said to change in correspondence to the actual position of the actual boundary lines of the piece itself with regard to the fundamental lines of orientation, a and b; let this actual position be expressed by the letter l, l marking the distance of one32 of the actual boundary lines of our piece from a or b: then we are entitled to improve our formula by writing p.v. (X) = f (s, l…) (Fig. II).

But the formula is not yet complete: s and l are what the mathematicians call variables: they may have any actual value and there will always be a definite value of p.v., i.e. of

Fig. xi.—Diagram to show the Characteristics of anHarmonious-Equipotential System.”
The element X forms part of the systems a b or a1 b1 or a2 b2; its prospective value is different in each case.

the actual fate which is being considered; to every value of s and l, which as we know are independent of each other, there corresponds a definite value of the actual prospectivity. Now, of course, there is also a certain factor at work in every actual case of experimental or normal development, which is not a variable, but which is the same in all cases. This factor is a something embraced in the prospective potency of our system, though not properly identical with it.

The prospective potency of our system, that is to say of each of its elements, is the sum total of what can be done equally well by all; but the fact that a typically proportionate development occurs in every possible case, proves that this sum comes into account, not merely as a sum, but as a sort of order: we may call this order the “relation of localities in the absolutely normal case”. If we keep in mind that the term “prospective potency” is always to contain this order, or, as we may also call it, this “relative proportionality”, which, indeed, was the reason for calling our systems “harmonious”, then we may apply it without further explanation in order to signify the non-variable factor on which the prospective value of any element of our systems depends; and if we denote the prospective potency, embracing order, by the letter E, we are now able to complete our formula by saying p.v. (X) = f (s, l, E).

So far the merely analytical study of the differentiation of harmonious-equipotential systems.33

Instances of “Harmonious-Equipotential Systems”

We must try at first to learn a few more positive facts about our systems, in order that we may know how important is the part which they play in the whole animal kingdom, and in order that our rather abstract analysis may become a little more familiar to us. We know already that many of the elementary morphogenetic organs have been really proved to be harmonious-equipotential systems, and that the same probably is true of many others; we also know that the immature egg of almost all animals belongs to this type, even if a fixed determination of its parts may be established just after maturation. Moreover, we said, when speaking about some new discoveries on form-restitution, that there are many cases in which the processes of restitution do not proceed from single localities, the seat of complex potencies in the organism, but in which each single part of the truncated organism left by the operation has to perform one single act of restoration, the full restitution being the result of the totality of all.

Let me mention in the first place a few more purely embryological cases.

The mesenchyme cells of the sea-urchin form a harmonious-equipotential system; for we may eliminate any number of these cells we like, or may change their relative position with respect to one another, and the skeleton is always normal. It is just as when 100 workmen have to construct a bridge and a certain number of them, n, fall ill and become unable to work. Then the rest, 100 − n,—(n being variable!)—must carry out the work, assuming that we cannot engage other workmen, and are able to do so!

Beautiful cases of equipotentiality have been discovered by Braus and Harrison: the Anlage of the limb of Amphibians, including the complicated skeleton, is equipotential to the very highest degree. You can do with this Anlage what you like: take off any part, unite two of them, turn it over, etc. The result is always normal.

The same is true with regard to regeneration: the small buds of cells, by which regeneration of the limbs of Amphibians begins, are equipotential. And it even seems, that the initial prolification of cells, which used to be the first step of regenerative processes, is harmonious-equipotential in every case.

But now let us give some details about some forms of restitution discovered during the last thirty years.

All of you have seen common sea-anemones or sea-roses, and many of you will also be familiar with the so-called hydroid polyps. Tubularia is one genus of them: it looks like a sea-anemone in miniature placed on the top of a stem like a flower. It was known already to Allman that Tubularia is able to restore its flower-like head when that is lost, but this process was taken to be an ordinary regeneration, until an American zoologist, Miss Bickford, succeeded in showing that there was no “regeneration” process at all, in the proper sense of the word no budding of the missing part from the wound, but that the new tubularian head was restored by the combined work of many parts of the stem. Further analysis then taught us that Tubularia indeed is to be regarded as the perfect type of an harmonious-equipotential system: you may cut the stem at whatever level you like: a certain length of the stem will always restore the new head by the co-operation of its parts. As the point of section is of course absolutely at our choice, it is clear, without any further discussion, that the prospective value of each part of the restoring stem is a

Fig. 12.—Tubularia.
a. Diagram of the “Hydranth”, with its short and long tentacles.
b. Restitution of a new hydranth inside the perisarc (p).
c. The same—later stage; the tentacles are complete; the whole hydranth will be driven out of the perisarc by a process of growth that occurs at the locality marked ↑.
d. A stem of Tubularia cut either at a1 b1 or at a2 b2, or at a1 c.
e. Position of tentacles in the piece cut at a1 b1.
f. ” ” ”a2 b2, which is equal in length to a1 b1.
g. ” ” ”a1 c, which is half as long as a1 b1.

“function of its position”, that it varies with its distance from the end of the stem; and so at once we discover one of the chief characteristics of our systems. But also the second point which enters into our formula can be demonstrated in Tubularia: the dependence of the fate of every element on the actual size of the system. You would not be able to demonstrate this on very long stems, but if you cut out of a Tubularia stem pieces which are less than ten millimetres in length, you will find the absolute size of the head restored to be in close relation to the length of the stem piece, and this dependence, of course, includes the second sort of dependence expressed in our formula.

The figures will serve to show you a little more concretely what has been described. The head of Tubularia consists of a sort of broad base with a thin proboscis upon it, both bearing a large number of tentacles; these tentacles are the first things to be seen as primordia (“Anlagen”) in the process of restitution. You notice two rings of longitudinal lines inside the stem; the lines will become walls and then will separate from the stem until they are only connected with it at their basal ends; the new tentacles are ready as soon as that has happened, and a process of growth at the end will serve to drive the new head out of the so-called perisarc, or horny skeleton, which surrounds the stem. By comparing the two figures, 12 e and g, you easily find out that the absolute lengths of the two tentacle rings are very different, and that both are in proportion34 to the actual size of the stem (Fig. 12).

So we find our formula p.v. (X) = f (s, l, E) very well illustrated in Tubularia. The formula indeed may help us to predict, in any case, where a certain part of the polyp’s organisation is to originate, at least if we know all that is included under our letter E, i.e. the normal proportion of our form.

Another very typical case of a morphogenetic system of the harmonious type is supplied by the phenomena of restoration in the ascidian Clavellina. I cannot fully describe the organisation of this form (Fig. 13 a), and it must suffice to say that it is very complicated, consisting of two very different chief parts, the branchial apparatus and the so-called intestinal sac; if these two parts of the body of Clavellina are separated one from the other, each may regenerate the other in the typical way, by budding processes from the wound. But, as to the branchial apparatus, there may happen something very different: it may lose almost all of its organisation and become a small white sphere, consisting only of epithelia corresponding to the germ-layers, and of mesenchyme between them; and then, after a certain period of rest, a new organisation will appear. Now, this new organisation is not that of a branchial apparatus but represents a very small but complete ascidian (Fig. 13). Such a fact certainly seems to be very important, not to say

Fig. 13.—Clavellina.
a. Diagram of the normal animal: E and J = openings; K = branchial apparatus; D = intestine; M = stomach; H = heart.
b. The isolated branchial apparatus.
c–e. Different stages of reduction of the branchial apparatus.
f. The new whole little ascidian.

very surprising. But still another phenomenon may be demonstrated on the animal, which seems to be even more important. You first isolate the branchial apparatus from the other part of the body, and then you cut it in two, in whatever direction you please. Provided they survive and do not die, as indeed many of them do, the pieces obtained by this operation will each lose their organisation, as did the whole branchial apparatus, and then will each acquire another one—and this new organisation is also that of a complete little Clavellina. So we see that not only is the branchial apparatus of our animal capable of being transformed into a whole animal by the co-operative work of all its parts, but even each part of it may be transformed into a small whole, and it is quite at our disposal how large this part shall be, and what sort of a fragment of the original branchial apparatus it shall represent.

We could hardly imagine a better instance of an harmonious-equipotential system.

I cannot give you a description of all the other types of our systems subservient to restitution, and I can only mention here that the common hydra and the flatworm Planaria are very fine examples of them. But to one special case of harmonious equipotentiality you must allow me to direct your further attention.

It has been known for many years that the Protozoa are also capable of a restoration of their form and organisation after disturbances, if at least they contain a certain amount of their nuclear substance. This process of restoration used to be regarded as belonging to the common type of regeneration proper, until T. H. Morgan succeeded in showing that in the genus Stentor it follows just the very lines which we know already from our study of embryonic organs or from Tubularia: that an harmonious-equipotential system is at the basis of what goes on. Now, you know that all Protozoa are but one highly organised cell: we have therefore here an instance where the so-called “elements” of our harmonious-morphogenetic system are not cells, but something inside of cells; and this feature must appear to be of very great moment, for it first shows, as we have already pointed out on another occasion, that morphogenesis is not necessarily dependent on cell-division, and it states at the same time that our concept of the harmonious-equipotential system may cover a very great area—that, in fact, it is a scheme of a very wide extent.

The Problem of the Factor E

We turn back again to considerations of a more abstract form. We left our analysis of the differentiation of the harmonious-equipotential systems, and particularly of the phenomena of localisation during this differentiation, at the point where we had succeeded in obtaining an equation as the expression of all those factors on which the prospective value, the actual fate, of any element of our systems depends. p.v. (X) = f (s, l, E) was the short expression of all the relations involved; s and l, the absolute size of the system and the relative position of the element with respect to some fixed points, were independent variables; E was a constant, namely, the prospective potency, with special regard to the proportions embraced by it.

We shall now study the significance of the factor E.

What does this E mean? Is it a short expression merely for an actual sum of elemental agents having a common resultant? And, if so, of what kind are these agents? Or what may E mean, if it can be shown not to be a short sign for a mere sum?

No Explanation offered by “Means” or “Formative Stimuli”

For practical purposes it seems better if we modify the statement of our question. Let us put it thus: E is one of the factors responsible, among variables, for the localisation of organic differentiation; what then do we actually know about the causal factors which play a localising part in organogenesis? We, of course, have to look back to our well-studied “formative stimuli”. These stimuli, be they “external” or “internal”, come from without with respect to the elementary organ in which any sort of differentiation, and therefore of localisation, occurs: but in our harmonious systems no localising stimulus comes from without, as was the case, for instance, in the formation of the lens of the eye in response to the optical vesicle touching the skin. We know absolutely that it is so, not to speak of the self-evident fact that the general “means” of organogenesis have no localising value at all.35

Some authors have objected to my arguments that the germ, say in the shape of sixteen cells, might be regarded as a typically ordered physico-chemical system, in which all sorts of diffusions and other kinds of transport of materials might go on in a well-regulated pre-established way. Very well—for the case of normal embryology. But there are the results of experiment! I take away one of the first four cleavage cells: the result is the normal one. And, if I may add another type of experiment not yet mentioned, the result is also the normal one if in the 16–cell stage I take, say, two micromeres, one macromere, and three mesomeres—that is, if I allow development to start from very “unharmoniously composed” conditions.

In face of such facts, the theory of the well-ordered pre-established system of surfaces, diffusions, etc., breaks down completely.

So we see there is nothing to be done, either with the means or with the formative stimuli; both are entirely unable to account for those kinds of localisation during differentiation which appear in our harmonious systems.

But is there no possibility of explaining the phenomena of organogenetic localisation by any other sort of interaction of parts? Two such possibilities may at the first glance seem to exist.

No Explanation offered by a Chemical Theory of Morphogenesis

Though never set forth in the form of a properly worked-out theory, the view has sometimes been advocated by biologists, that a chemical compound of a very high degree of complication might be the very basis of both development and inheritance, and that such a chemical compound by its disintegration might direct morphogenesis.

Let us first examine if such a view may hold for the most general features of organic morphogenesis. It seems to me that from the very beginning there exists one very serious objection to every purely chemical theory of form-building,36 in the mere fact of the possibility of the restoration of form starting from atypical localities. The mere fact, indeed, that there is such a thing as the regeneration of a leg of a newt—to say nothing about restitution of the harmonious type—simply contradicts,37 it seems to me, the hypothesis that chemical disintegration of one compound may govern the course of morphogenetic events: for whence comes the re-existence of the hypothetical compound, newly to be disintegrated, after disintegration has been completed once already? And we even know that regeneration may go on several times running from the same locality, and that a regenerated part may later on be itself the starting-point of regeneration.

But, if we intentionally disregard this difficulty, in spite of its fundamental character, how could the hypothesis of chemical disintegration give the reason for the differentiation of our harmonious-equipotential systems, with special regard to the localisation of it; how could it account, in other words, for the appearance of typically localised specifications in an organ for which no external localising causes can be predicated?

Let us remember that a few original intimate differences exist in our harmonious systems: the main directions of the intimate protoplasmic structure including polarity and bilaterality. There are therefore three times two specified poles in each of these systems, at least in bilateral organisms, but no other differences are present in them. A few very simple cases of harmonious differentiation might indeed be understood on the theory of a disintegrating chemical compound in connection with these few differences. Imagine that the original compound of the quantity a is disintegrated to the amount of a1; from a1 are formed the two more simple compounds, b and c, both of them in definite quantities; then we have the three chemical individuals, aa1, b and c, as the constituents of our harmonious system; and it now might be assumed, without any serious difficulty, though with the introduction of some new hypotheses, that the two poles of one of the fundamental axes of symmetry attract b and c respectively, aa1 remaining unattracted between them. We thus should have the three elementary constituents of the system separated into three parts, and as they all three are of a definite quantity, their separation would mean that the system had been divided into three parts, aa1, b and c, also with regard to its proper form. It is clear, that by taking away any part of the original system, by means of operations, there would be taken away a certain amount of the original compound; say that a/n is left; then, of course, the three constituents after the partial disintegration would be aa1/n, b/n and c/n, and so it follows that the proportionality of localisation would really be preserved in any case.

But these considerations, evident as they seem to be in the most simple case, fail to satisfy in a really general sense: for two different reasons. First, they could never account for the fact that the differentiated organism by no means consists of so many different compounds as it shows single parts of its differentiation, but that, on the contrary, it only consists, as we know, of a certain rather limited number of true different morphogenetic elements, these elements occurring again and again—as, for instance, nervous or muscular elements—but typical each time in locality, quantity, and form. And in the second place, the very form of elementary organs, their form as such, does not at all go hand-in-hand with chemical differences; this feature alone would absolutely overthrow any sort of a chemical morphogenetic theory to account for the problem of localisation. Take the typically arranged ring of the mesenchyme cells in our Echinus-gastrula, with its two spherical triangles, so typically localised; look at any sort of skeleton, in Radiolaria, or in starfishes, or in vertebrates: here you have form, real form, but form consisting of only one material. Not only is the arrangement of the elements of form typical here, e.g. the arrangement of the single parts of the skeleton of the hand or foot, but also the special form of each element is typical, e.g. the form of each single bone of the foot; and on a purely chemical theory of morphogenesis the sufficient reason for the production of typical form in such a sense would be wanting. For atoms or molecules by themselves can only account for form which is arranged, so to speak, according to spatial geometry—as in fact they do in crystallography; but they can never account for form such as the skeleton of the nose, or hand, or foot. You will answer me, perhaps, that there may be non-chemical agents in the germ,38 responsible for typical form-localisation, but by such reasoning you would be departing from a purely chemical theory. Our next paragraph will be devoted to this side of the question.

That is the principal reason for rejecting all sorts of purely chemical morphogenetic theories put forward to explain the problem of localisation; it is more explicit, and therefore, I suppose, still more convincing than the more general consideration that the very fact of restitutions in itself must contradict the hypothesis that a disintegration of compounds might be the directive agency in morphogenesis. To sum up: Specificity of organic form does not go hand-in-hand with specificity of chemical composition, and therefore cannot depend on it; and besides that, specific organic form is such that it can never be explained by atomic or molecular arrangement in the chemical sense; for, to state it in a short but expressive manner, the “form” of an atom or molecule can never be that of a lion or a monkey. To assume that, would be to go beyond the limits of chemistry in chemistry itself.

No Machine Possible Inside the Harmonious Systems

And now we turn to the last possibility which is left to us in our endeavour to “understand” the localisation of the differentiation in our harmonious-equipotential systems by the means of physics and chemistry. Outside causes have failed to account for it; chemical or any other kind of disintegration of a compound or mixture has failed too. But could there not exist some sort of complicated interactions amongst the parts of the harmonious systems themselves? Could there not exist some kind of a real machine in the system, which, if once set going by fertilisation or its equivalent, would result in the differentiations that are to take place? Then we might say that the “prospective potency” of the system is in fact that machine; we should know what the letter E of our equation stood for: viz. a resultant action of many complicated elemental interactions, and nothing more.

We shall understand the word “machine” in a most general sense. A machine is a typical configuration of physical and of chemical constituents, by the acting of which a typical effect is attained. We, in fact, lay much stress upon embracing in our definition of a machine the existence of chemical constituents also; we therefore understand by the word “machine” a configuration of a much higher degree of complication than for instance a steam-engine is. Of course, a machine whose acting is to be typical with regard to the three dimensions in space, has to be typically constructed with regard to these three dimensions itself; a machine that was an arrangement of elements in a strict plane could never have typical effects at right-angles to that plane. This is a point which must well be kept in mind in all hypothetical considerations about machines that claim to explain morphogenesis.

It must be granted that a machine, as we understand the word, might very well be the motive force of organogenesis in general, if only normal, that is to say, if only undisturbed development existed, and if a taking away of parts of our systems led to fragmental development.

But we know that, at least in our harmonious-equipotential systems, quite another process occurs after parts have been taken away: the development that occurs is not fragmental but whole, only on a smaller scale.

And we know, further, that this truly whole development sets in irrespective of the amount and direction of the separation. Let us first consider the second of these points. There may be a whole development out of each portion of the system—above certain limits—which is, say, of the volume V. Good! Then there ought to exist a machine, like that which exists in the whole undisturbed system, in this portion V also, only of smaller dimensions; but it also ought to exist in the portion V1 which is equal to V in amount, and also in V2, in V3, V4, and so on. Indeed, there do exist almost indefinitely many Vn, all of which can perform the whole morphogenesis, and all of which therefore ought to possess the complete machine. But these different portions Vn are only partly different from each other in spatial relation. Many parts of V2 are also parts of V1 and of V3 and of V4, and so on; that is to say, the different volumes Vn overlap each other successively and in such a manner that each following one exceeds the preceding one in the line by a very small amount only. But what then about our machines? Every volume which may perform morphogenesis completely must possess the machine in its totality. As now every element of one volume may play any possible elemental role in every other, it follows that each part of the whole harmonious system possesses any possible elemental part of the machine equally well, all parts of the system at the same time being constituents of different machines.

A very strange sort of machine indeed, which is the same in all its parts (Fig. 14)!

But we have forgotten, I see, that in our operation the absolute amount of substance taken away from the system was also left to our choice. From this feature it follows that not only all the different Vn, all of the same size, must possess the hypothetic machine in its completeness, but that all amounts of the values Vnn, n being variable, must possess the totality of the machine also: and all values Vnn, with their variable n, may again overlap each other.

Here we are led to real absurdities!

But what is the conclusion of our rather wild considerations?

It seems to me that there is only one conclusion possible. If we are going to explain what happens in our harmonious—

Fig. 14.—AnHarmonious-Equipotential SystemOF WHATEVER KIND.
According to the “machine-theory” of life this system ought to possess a certain unknown very complicated machine in its completeness:
(a) in its total length,
and (b) in each of the equal volumes v, v1, v2, v3, and so on,
and (c) in each of the unequal volumes w, x, y, and so on,
and (d) in every imaginable volume, no matter of what size.
Therefore the “machine-theory” of life is absurd.

equipotential systems by the aid of causality based upon the constellation of single physical or chemical factors and events, there must be some such thing as a machine. Now, just the assumption of the existence of a “machine” proves to be absolutely absurd in the light of the experimental facts. Therefore there can be neither any sort of a machine nor any sort of causality based upon constellation underlying the differentiation of harmonious-equipotential systems.

For a machine, typical with regard to the three chief dimensions of space, cannot remain itself if you remove parts of it or if you rearrange39 its parts at will.

Here we see that our long and careful study of morphogenesis has been worth while: it has afforded us a result of the very first importance.

The Autonomy of Morphogenesis proved

No kind of causality based upon the constellations of single physical and chemical acts can account for organic individual development; this development is not to be explained by any hypothesis about configuration of physical and chemical agents. Therefore there must be something else which is to be regarded as the sufficient reason of individual form-production. We now have got the answer to our question, what our constant E consists in. It is not the resulting action of a constellation. It is not only a short expression for a more complicated state of affairs, it expresses a true element of nature. Life, at least morphogenesis, is not a specialised arrangement of inorganic events; biology, therefore, is not applied physics and chemistry: life is something apart, and biology is an independent science.

All our results at present, indeed, are negative in their form; our evidence was throughout what is called per exclusionem, or indirect or apagogic. There were excluded from a certain number of possibilities all except one; a disjunctive proposition was stated in the form: E is either this, or that, or the other, and it was shown that it could not be any of all these except one, therefore it was proved to be that one. Indeed, I do not see how natural science could argue otherwise; no science dealing with inorganic phenomena does; something new and elemental must always be introduced whenever what is known of other elemental facts is proved to be unable to explain the facts in a new field of investigation.

We shall not hesitate to call by its proper name what we believe we have proved about morphogenetic phenomena. What we have proved to be true has always been called vitalism, and so it may be called in our days again. But if you think a new and less ambitious term to be better for it, let us style it the doctrine of the autonomy of life, as proved at least in the field of morphogenesis. I know very well that the word “autonomy” usually means the faculty of giving laws to oneself, and that in this sense it is applied with regard to a community of men, whilst in our phrase autonomy is to signify the being subjected to laws peculiar to the phenomena in question. But this meaning is etymologically defensible.

Vitalism then, or the autonomy of life, has been proved by us indirectly, and cannot be proved otherwise so long as we follow the lines of ordinary scientific reasoning.


But shall we not give a name to our vitalistic or autonomous factor E, concerned in morphogenesis? Indeed we will, and it was not without design that we chose the letter E to represent it provisionally. The great father of systematic philosophy, Aristotle, is also to be regarded as the founder of theoretical biology. Moreover, he is the first vitalist in history, for his theoretical biology is throughout vitalism; and a very conscious vitalism indeed, for it grew up in permanent opposition to the dogmatic mechanism maintained by the school of Democritus.

Let us then borrow our terminology from Aristotle, and let that factor in life phenomena which we have shown to be a factor of true autonomy be called Entelechy, though without identifying our doctrine with what Aristotle meant by the word evntele,ceia. We shall use this word only as a sign of our admiration for his great genius; his word is to be a mould which we have filled and shall fill with new contents. The etymology of the word evntele,ceia allows us such liberties, for indeed we have shown that there is at work a something in life phenomena “which bears the end in itself”, o] e;cei evn e`autw/| to. te,loj.

Our concept of entelechy marks the end of our analysis of individual morphogenesis. Morphogenesis, we have learned, is “epigenesis” not only in the descriptive but also in the theoretical sense: manifoldness in space is produced where no manifoldness was, real “evolutio” is limited to rather insignificant topics. But was there nothing “manifold” previous to morphogenesis? Nothing certainly of an extensive character, but there was something else: there was entelechy, and thus we may provisionally call entelechy an “intensive manifoldness”. That then is our result: not evolutio, but epigenesis—“epigenesis vitalistica”.

The Logic of our First Proof of Vitalism

Let us devote the end of our present chapter to an account of the logical means by which it has been possible to develop what we hope will be regarded as a true proof of life autonomy.

Firstly, we have looked upon the phenomena of morphogenesis without any prepossessions; we may say that we have fully surrendered ourselves to them; we have not attacked them with any sort of dogmatism except the inherent dogmatism of all reasoning. But this dogmatism, if it may be called so, does not postulate that the results of the inorganic doctrines must hold for the organic world, but only that both the inorganic and the organic must be subject to certain most general principles.

By studying fife as a given phenomenon, by fully devoting ourselves to our problem, we not only have analysed into its last elements what was given to us as our subject, but we also, more actively, have created new combinations out of those elements: and it was from the discussion of these positive constructions that our argument for vitalism was derived.

We have analysed morphogenesis into elementary processes, means, potency, formative stimulus, just as the physicist analyses mechanics into time, velocity, mass, and force; we have then rearranged our elements into “systems”—the equipotential systems, the harmonious-equipotential system in particular, just as the physicist composes his elements into the concepts of momentum or of kinetic energy or of work. And finally, we have discussed our compositions and have obtained our result, just as the physicist gets his ultimate results by discussing work and kinetic energy and momentum.

Of coarse the comparison is by no means intended to show that mechanics and biology are sciences of the same kind. In my opinion, they are not so at all; but nevertheless there do exist similarities of a logical kind between them.

And it is not the formal, logical character alone which allows us to compare biology with other natural sciences: there is still something more, there is one kind of assumption or postulate, or whatever you may choose to call it, without which all science whatever would be altogether impossible. I refer to the concept of universality. All concepts about nature which are gained by positive construction out of elements resulting from analysis, claim to be of universal validity; without that claim there could indeed be no science.

Of course this is no place for a discussion on methodology, and it therefore must suffice to make one remark with special regard to our purpose, which we should like to emphasise. Our concept of the harmonious-equipotential system—say rather, our concept of the prospective potency itself—presumes the understanding that indeed all blastomeres and all stems of Tubularia, including those upon which we have not carried out our experiments, will behave like those we have experimented with; and those concepts also presume that a certain germ of Echinus, A, the blastomeres of which were not separated, would have given two whole larvae, if separation had taken place, while another germ, B, which actually gave us two larvae after separation, would only have given one without it. Without this presumption the concept of “potency” is meaningless, and, indeed, every assumption of a “faculty” or a “possibility” would be meaningless in the whole area of science.

But this presumption can never be proved; it can only be postulated. It therefore is only with this postulate that our first proof of vitalism holds; but this restriction applies to every law of nature.

I cannot force you to agree with this postulate: but if you decline you are practically saying that there exists a sort of pre-established harmony between the scientific object and the scientist, the scientist always getting into his hands such objects only as have been “predestinated” from the very beginning to develop two larvae instead of one, and so on!

Of course, if that is so, no proof of natural laws is possible at all; but nature under such views would seem to be really daemonic.

And so, I hope, you will grant me the postulate of the universality of scientific concepts—the only “hypothesis” which we need for our argument.40

4. On Certain other Features of Morphogenesis advocating its autonomy

Our next studies on the physiology of form will be devoted in the first place to some additional remarks about our harmonious-equipotential systems themselves, and about some other kinds of morphogenetic “systems” which show a certain sort of relationship with them. For it is of the greatest importance that we should become as familiar as possible with all those facts in the physiology of form upon the analysis of which are to be based almost all of the future theories that we shall have to develop in biology proper and philosophical.


The type of the proper harmonious-equipotential system may go hand-in-hand with another type of “systems” which play a part in morphogenesis; a type which we have shortly mentioned already and which will be studied fully a few chapters later. We know that there are equipotential systems with complex potencies: that is to say, systems which may produce a whole organism equally well from any one of their elements; we know the cambium of Phanerogams to be such a system. Now it is easily understood that the germ of our Echinus, say in the stage of two or four or eight cleavage cells, is not only an harmonious-equipotential system, but a complex-equipotential system too. Not only may there arise a whole organism out of 2/4 or 3/4 or 3/8, 4/8, 5/8, 6/8, 7/8 of its elements, in which cases the harmonious role of the single element with regard to its single performance in a totality is variable, but there may also arise four whole single larvae out of the four cells of the four-cell stage, or eight single whole larvae out of the eight-cell stage.41 In these cases, of course, each of the four or eight elements has performed not a part of the totality, changing with its “position”, but the totality itself. With respect to these possible performances the “systems” present in the four-or eight-cell stages of cleavage must be called complex-equipotential ones.

We propose to give the name of mixed-equipotential systems to all those equipotential systems which, at the same time, may be regarded as belonging to the harmonious or to the complex type. It is not only among cleavage-stages that they are to be found; you may also find them very clearly exhibited in our ascidian Clavellina for instance. We know already that the branchial apparatus of this form is typically harmonious-equipotential, but it is complex-equipotential too, for it also may regenerate what is wanting in the proper way, by a budding from the wound; and the same is true of many other cases, the flatworm Planaria for instance.

Another type of systems, which might be said to be of a higher degree, is exhibited in some very strange phenomena of regeneration. It was first shown by Godlewski that a whole tail may be regenerated from a wound inflicted on the body of a newt, even if this wound involves section of only a portion of the body-diameter. Section of the whole of the body-diameter of course would cause the formation of the whole tail also; but it was found that even an incomplete cross-section of the body is capable of performing the whole on a smaller scale. The series of possible cross-sections which are all capable of regeneration would have to be called a system of the complex type in this case; but, now we learn that every single cross-section is of the harmonious type, we must speak of complex-harmonious systems. What we have described is not the only instance of our new type of morphogenetic systems. Weiss has observed the regeneration of a complete foot from a partial section through the leg of the newt, and in the flatworm Planaria a partial cross-section is also capable of forming a whole structure, say a head, and all cases of so-called “super-regeneration” after the infliction of a complicated wound probably belong here also.

You may say that our two additions to the theory of systems are merely formal, and indeed I am prepared to concede that we shall not learn anything altogether new from their discussion: their analysis would lead either to what was our “first proof” of the autonomy of life-phenomena or to what will be our “second” one. But the mere descriptions of the facts discovered here will interest you, I think, and will fill your minds with more vivid pictures of the various aspects of form-autonomy.

While dealing with our harmonious-equipotential systems as the starting-points of processes of restitution, e.g. in Tubularia, Clavellina, the flatworms, and other instances, we always have regarded cross-sections of the body as constituting the “elements” of equipotentiality. Now, cross-sections, of course, are by no means simple in themselves, but are made up of very different tissues, which are derivates of all three of the original germ layers—ectoderm, mesoderm, and endoderm. Owing to this composite character of the cross-sections, taken as elements of harmonious systems, a special phenomenon of morphogenesis is presented to us, which teaches somewhat more than the mere concept of harmonious-equipotentiality can express. If composite elements concerned in morphogenesis result in one whole organisation in spite of the development of the single tissues of these elements going on independently, then there must be a sort of correspondence or reciprocity of the harmonious development among these tissues themselves; otherwise a proportionate form could not be the final result. We may conveniently speak of a reciprocity of harmony as existing between the single tissues or germ layers which constitute many harmonious-equipotential systems, and there can be little doubt that we have here an important feature with regard to general morphogenesis.

Some very interesting complicated forms of equipotentiality have, finally, been discovered in recent years, in particular by Harrison, Schaxel, Weiss, etc. All of them imply our ordinary harmonious-equipotentiality, but exceed its realm in various ways.

The “Anlage” of a left leg of Amphibian embryos or regenerating Amphibians, if transplanted to the right side of the body, may give a right leg, and vice versa. A bud of the foreleg, when implanted upon the locality of a hindleg, gives a hindleg, and, if it is implanted upon the wound of the tail, it even gives a tail. We may speak of mirror equipotentiality and transgressing equipotentiality.

Here we find an influence of the locality of implantation upon the fate of the transplanted cells, a real formative influence. And the possibility of such an influence shows that the community of the transplanted cells is not only equipotential in the ordinary sense, i.e. that not only every cell of that community can perform, harmoniously, every single act within a given totality of such acts, but that the prospective area, so to speak, is enlarged.

Do not these facts tend to suggest that, in principle, in animal embryos also, every cell can do everything—as we know, e.g., from some lower forms of plants—and that restrictions of the prospective potency are nothing but a secondary phenomenon, due to a handicap, as it were?


In the hydroid polyp Tubularia, already familiar to us as being a most typical representative of the harmonious-equipotential systems, a very interesting phenomenon has been discovered,42 almost unparalleled at present but nevertheless of a general importance, a phenomenon that we may call a restitution of a restitution, or a restitution of the second order. You know that the first appearance of the new head of Tubularia, after an operation, consists in the formation of two rings of red lines, inside the stem, these rings being the primordia of the new tentacles. I removed the terminal ring by a second operation soon after it had arisen, disturbing in this way the process of restitution itself: and then the process of restitution itself became regulated. The organism indeed changed its course of morphogenesis, which was serving the purposes of a restitution, in order to attain its purpose in spite of the new disturbance which had occurred. For instance, it sometimes formed two rings out of the one that was left to it, or it behaved in a different way. As this difference of morphogenetic procedure is a problem by itself, to be discussed in the next paragraph, we shall postpone a fuller description of this case of a restitution of the second degree.

At present I do not see any way of proving independently the autonomy of life by a discussion of these phenomena; their analysis, I think, would again lead us to our problem of localisation and to nothing else; at least in such an exact form of reasoning as we demand.


I have told you already that Tubularia in the phenomena of the regulation of restitutions offers us a second problem of a great general importance, the problem of the Equifinality of Restitutions. There indeed may occur restitutions, starting from one and the same initial state and leading to one and the same end, but using very different means, following very different ways in the different individuals of one and the same species, taken from the same locality, or even colony.

Imagine that you have a piece of paper before you and wish to sketch a landscape. After drawing for some time you notice that you have miscalculated the scale with regard to the size of the paper, and that it will not be possible to bring upon the paper the whole of the landscape you want. What then can you do? You either may finish what you have begun to draw, and may afterwards carefully join a new piece of paper to the original one and use that for the rest of the drawing; or you may rub out all you have drawn and begin drawing to a new scale; or lastly, instead of continuing as you began, or erasing altogether, you may compromise as best you can by drawing here, and erasing there, and so you may complete the sketch by changing a little, according to your fancy, the proportions as they exist in nature.

This is precisely analogous to the behaviour of our Tubularia. Tubularia also may behave in three different ways, if, as I described to you, the terminal one of its two newly arisen rings of tentacle primordia is removed again. It may complete what is left, say the basal tentacle ring, then put forth from the horny skeleton (the “perisarc”) the new head as far as it is ready, and finally complete this head by a regular process of budding regeneration. But it also may behave differently. It may “erase” by a process of retro-differentiation all that has been left of what had already been formed, and then may form de novo the totality of the primordia of a new head. Or, lastly, it may remove a part of the middle of the one ring of tentacle rudiments which was left, and may use this one ring for the formation of two, which, of course, will not be quite in the normal relations of place with regard to each other and to the whole, but will be regulated afterwards by processes of growth. Thus, indeed, there is a sort of equifinality of restitution: one starting-point, one end, but three different means and ways.

It would, of course, contradict the principle of univocality, as we shall see more fully later on, to assume that there actually are different ways of regulation whilst all the conditions and stimuli are the same. We are obliged to assume, on the contrary, that this is not the case, that there are certain differences in the constellation, say of the general conditions of age or of metabolism, which are responsible for any given individual choosing one process of restitution instead of another; but even then the phenomenon of equifinality remains very striking.

It has long been known that restitution in general does not always follow the same lines of morphogenesis as are taken by ontogeny, and it was this feature that once led Roux to point out that the adult forms of organisms seem to be more constant than their modes of origin. But, comparing ontogeny with restitution in general, we see that only the ends are the same, not the points of starting; the latter are normal or typical in ontogeny, non-typical in restitution. In the new discoveries of an equifinality of restitutions we have the same starting-point, which is decidedly non-typical, i.e. dependent on our arbitrary choice, leading by different ways always to the same end.

There may be many who will regard the fact of equifinality as a proof of vitalism. I should not like to argue in this easy way; I indeed prefer to include part of the phenomena of equifinality in our first proof of autonomy, and part in the second one, which is to follow.

Another important phenomenon of the equifinality of regulation was discovered by Morgan. A species of the flatworm Planaria was found to restore its totality out of small pieces either by regeneration proper, if the pieces were fed, or by a sort of rearrangement of material, on the basis of its harmonious-equipotentiality, if they were kept fasting. It is important to note that here we see one of the conditions determining the choice of the way to restoration, as we also do in the well-known equifinal restitutions of the root in plants, where the behaviour of the organism depends on the distance of the operation-wound from the tip.44 In Tubularia the actual stage of restitution that has been already reached by the stem when the second operation takes place, may account for the specification of its future organogenesis, but this is not at all clearly ascertained at present.

Clavellina also shows equifinality in its restitution, as has already been shortly mentioned. The isolated branchial apparatus may restitute itself by retro-differentiation to an indifferent stage followed by renovation; or it may regenerate the intestine-sac in the proper way. Nothing is known here about the conditions, except perhaps that young individuals seem more apt to follow the first of these two ways, older ones the second; but there are exceptions to this rule.

The discussion of other instances of equifinality, though important in themselves, would not disclose anything fundamentally new, and so we may close the subject with the remark that nothing can show better than the fact of the equifinality of restitutions how absolutely inadequate all our scientific conceptions are when confronted with the actual phenomena of life itself. By analysis we have found differences of potencies, according as they are simple or complex; by analysis we have found differences of “systems”, differences of means, and indeed we were glad to be able to formulate these differences as strictly as possible: but now we see how, in defiance of our discriminations, one and the same species of animals behaves now like one sort of our “systems”, and now like the other; how it uses now one sort of “potencies”, now another.

But even if it is granted that, in the presence of such phenomena of life, our endeavour seems to be like a child’s play on the shores of the ocean, I do not see any other way for us to go, so long, at least, as our goal is human science—that is, a study of facts as demanded by our mental organisation.

Some Particular Features of Regeneration Proper

Regeneration may start from the same locality very many times (up to forty times have been observed). Can this be understood on mechanical lines?

And the regenerated half of, say, Clavellina or the little annelid Amphiglaena may regenerate the original other half, if this is cut off after the first regeneration. And, then, the half that was regenerated from the regenerated half may again regenerate the other half, etc., etc., etc.

Is there still the same organism? And what would be “the same” here?

But we are still moving on the grounds of science proper.

  • 1.

    Das Keimplasma, Jena, 1892.

  • 2.

    Die Bedeutung der Kernteilungsfiguren, Leipzig, 1883.

  • 3.

    Virchow’s Archiv. 114, 1888.

  • 4.

    Zeitschr. wiss. Zool. 53, 1891.

  • 5.

    Zeitschr. wiss. Zool. 55, 1892.

  • 6.

    In the pressure experiments I had altered the relative position of the nuclei in origine. In later years, I succeeded in disturbing the arrangement of the fully-formed cells of the eight-cell stage, and in getting normal larvae in spite of that in many cases. But as this series of experiments is not free from certain complications—which in part will be understood later on—it must suffice here to have mentioned them. (For further information, see my paper in Archiv. f. Entwickelungsmechanik, xiv., 1902, page 500.)

  • 7.

    Mitteil. Neapel. II, 1893.

  • 8.

    But the elementary magnets would have to be bilateral!

  • 9.

    Arch. Entw. Mech. 2, 1895.

  • 10.

    Anat. Anz. 10, 1895.

  • 11.

    Arch. Entw. Mech. 3, 1896.

  • 12.

    It deserves notice, in this connection, that in some cases the protoplasm of parts of a germ has been found to be more regulable in the earliest stages, when it is very fluid, than later, when it is more stiff.

  • 13.

    If the plane of section passes near the equator of the germ, two whole larvae may be formed also, but in the majority of cases the “animal” half does not go beyond the blastula. The specific features of the organisation of the protoplasm come into account here.

  • 14.

    A change of the position of the cell is, of course, effected by each variation of the direction of the cut, which is purely a matter of chance.

  • 15.

    The reader will remember (see page 42, note 1), that even the germ of Echinus is not quite equipotential along its main axis, but it is equipotential in the strictest sense around this axis. The germs of certain medusae seem to be equipotential in every respect, even in their cleavage stages.

  • 16.

    Journ. Exp. Zool. I, 1904.

  • 17.

    Great caution must be taken in attributing any specific morphogenetic part to differently coloured or constructed materials which may be observed in the egg-protoplasm in certain cases. They may play such a part, but in other cases they certainly do not (see Lyon, Arch. Entw. Mech. 23, 1907). The final decision always depends on experiment.

  • 18.

    It seems that these physical conditions also—besides the real specifications in the organisation of the egg—may be different before and after maturation or (in other cases) fertilisation. (See Driesch, Archiv. f. Entwickelungsmechanik, 7, p. 98, and Brachet, ibid. 22, p. 325.)

  • 19.

    According to Zur Strassen’s results, the early embryology of Ascaris proceeds almost exclusively by cellular surface-changes: the most typical morphogenetic processes are carried out by the aid of this “means”. As a whole, the embryology of Ascaris stands quite apart and presents a great number of unsolved problems.

  • 20.

    Arch. Entw. Mech. 17, 1904.

  • 21.

    Zeitschr. wiss. Zool. 55, 1892; and Mitt. Neapel. II, 1893.

  • 22.

    In certain cases, part of the specific feature of the process in question may also depend on the “cause” which is localising it, e.g. in the galls of plants.

  • 23.

    Herbst, “Über die Bedeutung der Reizphysiologie für die kausale Auffassung von Vorgängen in der tierischen Ontogenese” (Biol. Centralblatt, vols. xiv., 1894, and xv., 1895); Formative Reize in der tierischen Ontogenese, Leipzig, 1901. These important papers must be studied by every one who wishes to become familiar with the subject.

  • 24.

    Compare the important papers by J. Loeb, Untersuchungun zur physiologischen Morphologic der Tiere, Würzburg, 1891–2.

  • 25.

    I use the word “primordia” for the German “Anlage”; it is better than the word “rudiment”, as the latter may also serve to signify the very last stage of a certain formation that is disappearing (phylogenetically).

  • 26.

    A full analysis of the subject would not only have to deal with formative stimuli as inaugurating morphogenetic processes, but also with those stimuli which terminate or stop the single acts of morphogenesis. But little is actually known about this topic, and therefore the reader must refer to my other publications. I will only say here, that the end of each single morphogenetic act may either be determined at the very beginning or occur as an actual stopping of a process which otherwise would go on for ever and ever; in the first case some terminating factors are included in the very nature of the morphogenetic act itself.

  • 27.

    Croonian Lecture, Proc. Roy. Soc. vol. cii., 1927.

  • 28.

    Driesch, Die organischen Regulationen, Leipzig, 1901; Morgan, Regeneration, New York, 1901.

  • 29.

    For a fuller analysis, compare my opening address delivered before the section of “Experimental Zoology” at the Seventh International Zoological Congress, Boston, 1907: “The Stimuli of Restitutions” (see Proceeding of that Congress).

  • 30.

    The problem of the stimulus of a secondary restitution as a whole must not be confused with the very different question, what the single “formative stimuli” concerned in the performance of a certain restitutive act may be. With regard to restitution as a whole, these single “formative stimuli” might properly be said to belong to its “internal means”—in the widest sense of the word.

  • 31.

    The so-called “inner secretion” in physiology proper would offer a certain analogy to the facts assumed by such an hypothesis. Compare the excellent summary given by E. Starling at the seventy-eighth meeting of the German “Naturforscherversammlung”, Stuttgart, 1906.

  • 32.

    The distance of the other boundary line from a or b would be given by the value of s.

  • 33.

    A far more thorough analysis of this differentiation has been attempted in my paper, “Die Lokalisation morphogenetischer Vorgänge; Ein Beweis vitalistischen Geschehens,” Leipzig, 1899.

  • 34.

    This statement is not strictly correct for Tubularia. I found (Archiv. f. Entwickelungsmechanik, ix. 1899) that a reduction of the length of the stem is always followed by a reduction of the size of the hydranth-primordium, but there is no real proportionality between them. It is only for theoretical simplification that a strict proportionality is assumed here, both in the text and the diagram. But there is an almost strict proportionality in all cases of “closed forms”.

  • 35.

    One might object here that in a piece of a Tubularia stem, for instance, the tissues are in direct contact with the sea-water at the two points of the wounds only, and that at these very points a stimulus might be set up—say by a process of diffusion—which gradually decreases in intensity on its way inward. And a similar argument might apply to the small but whole blastula of Echinus, and to all other cases. But, in the first place, stimuli which only differ in intensity could hardly call forth the typical and typically localised single features realised in differentiation. On the other hand—and this will overthrow such an hypothesis completely—the dependence of the single localised effects in every case on the absolute size of the fragment or piece chosen for restoration renders quite impossible the assumption that all the singularities in the differentiation of the harmonious systems might be called forth by single stimuli originating in two fixed places in an independent way. These would never result in any “harmonious”, any proportionate structure, but a structure of the “normal” proportionality and size at its two ends and non-existent in the middle!

  • 36.

    Everything that is said in this paragraph against a “purely chemical” theory of morphogenesis is, of course, also valid against any theory which tries to explain morphogenesis exclusively by the dismingling of a given mixture.

  • 37.

    The question is rendered still more complicated by the fact that in the case of the regeneration, say, of a leg it is not the original “morphogenetic compound” which is again required for disintegration, after it has become disintegrated once already, but only a specific part of it: just that part of it which is necessary for producing the leg! On the other hand, it would be impossible to understand, on the basis of physical chemistry, how the isolated branchial apparatus of Clavellina could be transformed, by chemical processes exclusively, into a system of which only a certain part consists of that substance of which the starting-point had been composed in its completeness.

  • 38.

    Besides the specified poles determined by the polar-bilateral structure of the protoplasm.

  • 39.

    The pressure experiments and the dislocation experiments come into account here; for the sake of simplicity they have not been alluded to in the main line of our argument.

  • 40.

    In the Sitzungsberichte der Heidelberger Akademie der Wissenschaften (1918, No. 3, and 1919. No. 18) I have published a very careful and thorough, quasi-mathematical, analysis of harmonious equipotentiality, “which, so far, has not received much attention. Comp. also my little book Der Begriff der organischen Form, Berlin, 1919.

  • 41.

    The eight larvae would be incomplete in some respect, but not with regard to symmetry. They would be “whole” ones, only showing certain defects in their organisation.

  • 42.

    Driesch, Arch. Entw. Mech. 5, 1897.

  • 43.

    Driesch, Arch. Entw. Mech. 14, 1902.

  • 44.

    The root may be restored by regeneration proper, or by the production of adventitious roots, or by one of the side-roots changing its geotropism from horizontal to positive, according to the smaller or greater distance of the wound from the tip.