The Biological Problem of To-day: Preformation Or Epigenesis? / The Basis of a Theory of Organic Development

by Oscar Hertwig

PART I.

WEISMANN'S THEORY OF THE GERMPLASM AND DOCTRINE OF DETERMINANTS.

As may be seen in his essays, On Life and Death, On the Duration of Life, etc., Weismann believes himself to have established a fundamental distinction between unicellular and multicellular organisms. Unicellular organisms (he would have it) do not undergo natural death, but, since they are able to reproduce themselves continuously by a process of simple division, are immortal. Multicellular organisms, on the other hand, must perish after a definite duration of life, and so are mortal. He makes an exception of the sexual cells, which, like unicellular organisms, are able to multiply indefinitely, and so are immortal. Thus Weismann came to make a distinction between the mortal (somatic) cells and the immortal (germ) cells of multicellular organisms. The latter he regarded as arising directly from the egg-cell, and never from somatic cells.

Nussbaum has given utterance to similar views, holding that the dividing egg at a very early period cleaves into the cells from which the individual grows and the cells for the maintenance of the species. He has enunciated the proposition that, when the sexual cells have been separated from the cells of the young embryo, the material of the germ has been divided into shares for the individual and shares for the species; that the sexual cells take no part in the formation of the body, and that body-cells never give rise to ova or spermatozoa.

Weismann differs from Nussbaum in one important point. He lays no stress on the direct origin of the sexual cells, as cells, from the egg at the beginning of its development. He found, for instance, that, in the case of hydroids, the sexual cells did not arise in such a fashion. He considers, therefore, that the chain of events is as follows: The whole of the protoplasm of an egg-cell is not required to build up the new being, and the superfluous part remains unaltered to form the sexual cells of the new generation. Unlike Nussbaum, then, he asserts a continuity, not for the sexual cells, but for the germinal protoplasm which he believes to pass along definite cell-tracks until it forms the sexual cells. From this germinal protoplasm, which makes the germ-cells, he distinguishes the somatic protoplasm which makes the mortal, somatic cells.

The germplasm theory entered a new phase in the year 1885, after the independent appearance in 1884 of essays by Strasburger and by me, in which we gave reason for thinking that the cell nucleus was, as I expressed it, the bearer of the characters which were transmitted by parents to their offspring; that, in fact, the nucleus was the material basis of heredity.

Weismann laid hold of this idea, but transmuted it to fit in with his original theory of the germplasm. Shortly put, his view is as follows: The whole of the nuclear material is not hereditary material, but only a definite part is such, and this part, throughout the development of the individual, remains unaltered in composition, and finally becomes the starting-point for the generations to come. The remaining and greater part of the nuclear material does not remain in an unaltered condition. The layers of cells, first formed in the embryo, grow unlike each other, and give rise to different organs and tissues; Weismann draws the inference that the nuclear substance as well alters during the process of development, transforming itself in a regular, orderly fashion, until, finally, each different kind of cell in the whole body contains a specific nuclearplasm. This segregation and transformation begins with the process of cleavage itself, and thus 'the two daughter-cells that arise from the first cleavage of the egg-cell become different, so that the one contains all the hereditary characters for the ectoderm, the other for the endoderm. In further course the ectodermal nuclearplasm divides into that containing the primary germs of the nervous system, and that containing the similar constituents for the outer skin. By further cellular and nuclear divisions the inherited germs for the nervous system separate into those for the sense organs, those for the central nervous system, and so forth, until there are separated the germs for all the separate organs, and for the production of the minutest histological differentiation.'

Weismann calls the diverging nuclearplasms into which the primitive germplasm is gradually transformed histogenous, because they determine the specific characters of the tissues. He assumes that the primitive, original germplasm has a most complicated molecular structure, while the histogenous nuclearplasms for tissue-cells, like muscle-cells, nerve-cells, sense-cells, gland-cells, and so forth, have relatively simpler structures. As, during the growth of the embryo, the germplasm becomes transformed into the histogenous plasms, its molecular structure becomes simpler in proportion to the fewer different possibilities of development each separated portion of it comes to contain.

Following out this chain of ideas, Weismann attributes only to those cells which contain unaltered germplasm the power of giving rise to complete new individuals, while cells with histogenous nuclearplasm, whether these be embryonal cells or cells of the ectoderm or of the endoderm, he regards as having lost this capacity, because nuclearplasm of a simpler molecular structure cannot retransform itself into that with the more complicated structure. The further conclusion is necessary that a part of the nuclearplasm of the original nucleus of the fertilised egg-cell must remain unaltered throughout the various nuclear divisions, although it may be mingled with the nuclearplasms of certain series of cells. For these reasons, ova and spermatozoa can arise only when the germplasm which has been handed on from the original nucleus to certain cells is able to overcome the histogenous plasm of these cells. In this respect Weismann has amended his original proposition that the germ-cells were immortal, like unicellular organisms. In a strict and literal interpretation such a proposition would be incorrect, for the germ-cells are immortal only so far as they contain the germplasm, the immortal part of the organism.

In its further elaboration Weismann's conception was influenced considerably by publications of Naegeli, De Vries, and Wiesner. These dealt with the composition of the hereditary material, and they contained new hypotheses concerning the primary structure of the cell-body. Weismann avowedly accepted the suggestion of De Vries, who had rehabilitated and modernized Darwin's doctrine of pangenesis, according to which gemmules, small particles endowed with the power of division, were the material bearers of hereditary characters.

From these different sources Weismann has now worked out, in minutest detail, a theory to which he considers his former writings but as the preface; none the less, he has taken from his own writings the most essential and characteristic sequences of idea, in a fashion but slightly modified. Let me give the most important parts of his conception.

The substance which is the bearer of the hereditary character of a species (the idioplasm of Naegeli) lies not in the general protoplasm of the ovum and spermatozoon, but in their nuclear matter (hypothesis of Hertwig and Strasburger). Weismann calls this the germplasm, so altering the previous connotation of the word. The germplasm of every species has an extremely complicated, stable architecture, an architecture that has been elaborated gradually in the course of past time. In this he distinguishes simple and complex component parts, the biophores, determinants, ids, and idants.

The biophores are his smallest material units, and to them are due the fundamental qualities of life—assimilation, metabolism, and reproduction by division. Thus, they correspond to Herbert Spencer's physiological units, Darwin's gemmules, De Vries' pangenes, and Hertwig's idioblasts. They are the bearers of the various characters of cells, and there are present in the germplasm a very large multitude of different kinds of them, corresponding to the number of cells with different characters.

The determinants are units of the rank next higher; they have qualities of their own, but are composed of groups of several kinds of biophores. They, too, have the power of division which is associated with, and comes about by, multiplication of the coherent company of biophores which lies within them.

The histological character of every cell in a multicellular organism is determined by a single determinant (cell-determinants). Weismann has framed his conception of determinants so as to avoid the supposition that every single cell is represented in the germplasm by its own biophores. There are small parts in the body in which the cells are all alike, and for these parts a single determinant suffices, afterwards multiplying by division. On the other hand, each cell or cell-group in the body, that is independently variable, must have its special determinant in the germplasm. And so the germplasm of a species must possess as many determinants, or guiding particles, as there are in the organism cells or cell-groups that are independently variable in the germ or in later stages (hereditary pieces or determinates).

As every cell or group of cells which corresponds to determinants has a definite position in the body, Weismann infers that the determinants are definitely placed in the germplasm, and form an ordered, complicated community. He has given the name id to these communities, which are higher units with definite constitution and with complicated architecture. These ids are bodies containing all the determinants necessary to build up the individual of a species, and correspond to what Weismann previously called ancestral plasms. Every id must be able to grow and multiply, for it is by their multiplication that the germplasm for new individuals is formed.

A single id would suffice for the conduct of a single life-history; Weismann, however, in the pursuit of a chain of thought connected with the relation of sexual reproduction to heredity, and which I shall not discuss here, regards the germplasm as being still more complicated, and consisting of many, sometimes more than a hundred, ancestral plasms or ids, which have been derived from near or distant ancestors, the peculiarities of whose structures they retain, and may at some time actually produce (explanation of atavism).

But how does this fabric, endowed with an architecture so complicated, actually produce the development of the adult from the egg? The natural mechanism for this purpose is cell division and nuclear division.

According to Weismann's supposition—a supposition which forms, as we shall see, a chief corner-stone of his system—there are two kinds of nuclear division, the difference between which has not been observed, but is a corollary from the difference between their results. The one kind is denoted as integral, or doubling division; the other as differential, or differentiating division. The first method has only an incidental importance in Weismann's hypothesis: it consists of the doubling by growth of the rudiments, and of a perfectly fair division of them between the half-chromosomes; it occurs in tissues-cells, where parent-cells divide into daughter-cells exactly similar to each other and to their parents.

On the other hand, in differentiating division the rudiments become irregularly grouped during their growth; consequently, on division of the ids, which are composed of determinants, totally different combinations of the determinants are included in the daughter-ids. This method of division of the germplasm plays the chief part in the transformation of the egg into the adult. It has to take place so that the numberless determinants, or guiding particles, of the germplasm may be disentangled and brought forward at the time and place necessary for them to guide the formations of the determinates, or independently variable parts of the adult body.

To take an example: Weismann's hypothesis requires that when the egg first divides into two, the germplasm should divide into two halves, each containing only one half of the total assemblage of determinants. In each subsequent cell-division this process of segregation is continued, so that the ids, as the phases of embryonic growth occur, contain more and more few different kinds of determinants. Supposing the germplasm to be composed of a million determinants at one stage, in the next it would contain only half a million, and in the next, again, only a quarter-million. In this manner the architecture of the ids becomes simpler and simpler, reaching the simplest conceivable condition in the active cells of the adult body. In these the germplasm consists only of the kind of determinants peculiar to the cells in which they lie; and these determinants are broken up into biophores, or bearers of cell qualities.

'The disintegration of the germplasm,' says Weismann,[7] 'is a wonderfully complicated process; it is a true "development," in which the idic stages necessarily follow one another in a regular order, and thus the thousands and hundreds of thousands of hereditary parts are gradually formed, each in its right place, and each provided with the proper determinants. The construction of the whole body, as well as its differentiation into parts, its segmentation, and the formation of its organs, and even the size of these organs—determined by the number of cells composing them—depend upon this complicated disintegration of the determinants in the id of germplasm. The transmission of characters of the most general kind—that is to say, those which determine the structure of an animal as well as those characterising the class, order, family, and genus to which it belongs—are due exclusively to this process.'

This mechanism of differentiating division fails to explain the phenomena of reproduction and of regeneration. For these Weismann has the following ancillary suppositions:

The first is the already-described hypothesis of continuity of the germplasm. As the disintegration of the germplasm into determinants, occurring in the development of an egg into an organism, is a process which cannot be retraced, and, as the future reproductive cells of the organism must contain undisintegrated, perfect germplasm, it follows that the germplasm in the germ-cells of the child must have come directly from the original germplasm of the parent. During the development, as Weismann assumes, only a few of the ids, each of which contains all the necessary germs, break up by differentiating division into the determinants which control the course of the ontogeny, and decide the final characters of the cells. Another set of ids remains undisintegrated, with their determinants fast bound together, and, in the cell divisions, is not broken up into dissimilar groups. The first set of ids is the active, disintegrating germplasm; the second set is a passive, latent germplasm, which may be described as accessory germplasm (Nebenkeimplasma). The active ids are his explanation of the embryonic events, which they direct; the accessory germplasm is reserved to form the germ-cells, and, in fast-bound condition, is handed on through a short or long series of cell-divisions alongside the active germplasm. Handed on in this passive state, it finally reaches a group of cells which may be many or few generations distant from the original egg-cell, and impresses upon them the character of sexual cells. This transfer of germplasm from the egg to the sexual cell occurs in orderly fashion, along prescribed series of cells which Weismann has called the germ-tracks. Only these cells, which contain part of the perfect, undisintegrated germplasm, serve for the preservation of the species and are immortal; the other cells, since, from the disintegration due to differentiating division, they contain only fragments of the perfect plasm (groups of determinants or single determinants), are mortal, somatic cells.

The formation of buds is explained in much the same way as the origin of germ-cells. There is handed along from the egg, through prescribed series of cells, a quantity of accessory, or bud, idioplasm.

The phenomena of alternation of generations require the supposition that in those animals and plants in which it occurs 'two kinds of germplasm exist, both of which always are present in the egg or in the bud, but of which one only is active at any time and rules the ontogeny, while the other remains inactive.' The alternating activity of these two produces the alternation of generations. So also dimorphism, which is exhibited most frequently as differences between the sexes, is explained by the assumption that 'double determinants' are present in the germplasm for all the cells, cell-groups, or entire organisms which have different characters in the male and female. One set of these double determinants remains latent, the other becomes active.

Finally, to explain the phenomena of regeneration, it is assumed that in the complicated cases where large parts of the body, like the head, the tail, or a bone, can be replaced after accidental loss, the cells with this power of regeneration contain, in addition to the determinants proper to them, supplementary determinants, which contain the germs needed for regeneration of the lost parts. These were handed on, during the ontogeny, through definite series of cells, in a passive condition, to become active when the conditions for their growth are supplied by the loss of the parts they can replace.

Criticism of the Germplasm Theory.[8]

At first sight, much of Weismann's fabric of hypotheses gives the impression of being a closed system, thought out as a whole, and it has been treated as such in most of the notices and criticisms which I have seen. As a matter of fact, Weismann has spared no pains in the elaboration of his system, and has attempted to bring under his theory the many different phenomena of heredity and development, as well as alternation of generations, regeneration, atavism, and so forth. But, on the other hand, he has been careless in testing the stability and security of the foundations upon which he has built. It is on solid foundations that lie deep in the earth, and that avoid all reproach of being scamped or superficial work, that the durability of a structure depends. In this criticism the details of the superstructure will be disregarded, but the foundation will be tested thoroughly.

Cells and cell-properties are essential parts of Weismann's theory; while Naegeli has attempted to make his theory of the idioplasm independent of the whole conception of cells. In this matter I agree with Weismann, as, indeed, with De Vries and others, and I consider that the course taken by Naegeli has made his position untenable.

Naegeli would make his theory of the idioplasm quite independent of the theory of cells, because, while cells are important units in morphological structure, independently of this they cannot be regarded as important units. 'By a unit,' he insists, 'we must understand, in a physical sense, a system of material particles. In the organic world there are very many kinds of higher and lower units; vegetable and animal individuals, organs, tissues, groups of cells (in the vegetable kingdom, for instance, vessels and sieve-tubes), cells, parts of cells (plant cell-membranes, plasma, granules, and crystalloids, starch—grains, fat-globules, and so forth), micellæ, molecules, atoms. In morphology and physiology, sometimes one kind of unit, sometimes another, comes characteristically and notably into evidence. That being so, there is no reason why a special kind of unit should be exalted in a general theory.'

Although, with Naegeli, we must recognise and keep in view the presence of a large number of higher and lower units in the organic world, a fact upon which I shall lay considerable emphasis later, we must none the less recognise that, among all elementary units, cells are most the conspicuous, morphologically and physiologically, in the whole organic realm. In actual research this is avowed very practically, as a glance at the biological literature of the last thirty years will show. Especially in the study of heredity, the cell is a unit that cannot be neglected, for it has been established that spores, ova, and spermatozoa, the units by which species are preserved in reproduction, both in the animal and in the vegetable kingdom, have the morphological value of cells.

In this point I am in opposition to Naegeli, although otherwise I agree with much in his conceptions.

A theory of heredity must be reconciled with the cell theory. In investigating Darwin's pangenesis, Galton's doctrine of the stirp, Naegeli's idioplasm, Weismann's germplasm, the intracellular pangenesis of De Vries, His' doctrinal of germinal foci for the formation of organs, or Roux's mosaic theory, I believe that one must face the question: How far do these doctrines agree with what we know about the structure and function of the cell? Moreover, in deciding between the alternatives—preformation and epigenesis—I believe that it will profit us to start our critical investigation with the cell itself. With this object, I shall now sum up in a few sentences as much of our present knowledge of the life of cells as, I believe, must be reckoned with in any theory of propagation.

The cell, which consists of protoplasm and a nucleus, is an elementary organism, that, by itself, or in combination with other cells, forms the basis of all animal and vegetable organisation. In minute structure it is so extraordinarily complicated that its essential constitution (its micellar or molecular structure) eludes our observation. It is a medley, composed of numerous, chemically distinct particles that may be divided into two groups, organised and unorganised. The latter are free, or in solution; they are such as albuminates, fats, carbo-hydrates, water, salts, and they serve as material for the nutrition and growth of the cell. The former make up the living cell body (in the narrow sense). They are able to multiply by growth and division, and they are therefore the elementary parts, units of life of lower rank, of which the cell, a unit of higher rank, is composed. They are the gemmules of Darwin, the physiological units of Spencer, the bioblasts of Altmann, the pangenes of De Vries, the plasomes of Wiesner, the idioblasts of Hertwig, and the biophores of Weismann.

The cells of every organic species possess a proper, specific organisation, more or less complicated, and, in correspondence with this, they are composed of more or less numerous and varied organised particles.

The nucleus is a special organ of cells, which is always present. It displays a collection of numerous, peculiar, elementary living units, the idioblasts. These show chemical, morphological, and functional differences from the plasomes, the living units of the protoplasm; but perhaps the idioblasts, by absorption of different material, may transform themselves into the plasomes, just as these last, by a similar process, may produce the plasma-products. In my view, the nucleus is the bearer of the idioplasm or hereditary material, that is to say, of a substance that is more stable than protoplasm, and, because it is less subject to influences of the outer world, it stamps its specific character upon the organism.[9]

A mass of protoplasm with several nuclei (like the myxomycetes, cœloblasts, etc.) has the morphological value of a number of cells (synergides), corresponding to the number of the nuclei.

The means by which the continuity of life is maintained is the capacity of the cell to manifold itself by division, so forming two or more separate pieces. The process, which in most cases is associated with complicated changes of the nuclear contents, appears essentially to consist of the following: The elementary units of the cell (centrosomes, chromatin bodies in the case of nuclear division), being endowed with special energy resulting from the processes of growth, divide, and the elementary products of division separate into two groups, which move from the middle line; upon this there follows a division of the general body of the cell, i.e., of the protoplasm and its contents.

From the point of view of cells, I believe myself compelled to raise several objections to most important bases of Weismann's germplasm theory. For convenience of exposition these may be divided into two groups: Objections to the hypothesis of differentiating division; objections to Weismann's doctrine of determinants.

I. Objections to the Hypothesis of Differentiating Division.

A corner-stone of Weismann's theory is his assumption of nuclear divisions which are differentiating. Proof of this fundamental assumption may be sought in vain in Weismann's writings. Instead of that, a series of abstract arguments are brought forward in favour of it. Thus on p. 31 (of the English translation) Weismann treats the chromatin in the nucleus of the fertilised egg as the substance which accomplishes inheritance, and he denotes all the nuclei of the organism arising from the nucleus of the egg by divisions as the chromatin-tree, and then goes on to ask whether or no the pieces of hereditary material that make up the chromatin-tree of an organism are like each other or different. 'It can easily be shown,' the answer runs, 'that the latter must be the case.' For 'the chromatin is in a condition to impress the specific character on the cell in the nucleus of which it is contained. As the thousands of cells which constitute an organism possess very different properties, the chromatin which controls them cannot be uniform; it must be different in each kind of cell.'

Moreover, on p. 45 (of the English edition), 'The fact itself' (the capacity on the part of the idioplasm for regular and spontaneous change) 'is beyond doubt. When once it is established that the morphoplasm of each cell is controlled, and its character decided, by the idioplasm of the nucleus, the regular changes occurring in the egg-cell, and the products of its division in each embryogeny, must then be referred to the corresponding changes of the idioplasm.'

Finally, on p. 205 (of the English edition), 'The cells of the segmenting ovum are completely dissimilar as regards their hereditary value, although they are all young and embryonic, and are not infrequently quite similar in appearance. It therefore seems to me to follow from this, as a logical necessity, that the hereditary substance of the egg-cell, which contains all the hereditary tendencies of the species, does not transmit them in toto to the segmentation cells, but separates them into various combinations, and transmits them in groups to the cells. I have taken account of these facts in considering the regular distribution of the determinants of the germplasm, and the conversion of the latter into the idioplasm of the cells in the different stages of ontogeny.'

In the different propositions I have quoted, we have to deal with what is merely a fallacy in rhetorical disguise. For, from the premiss that the chromatin has the power of impressing specific character upon the protoplasm of the cell, it by no means follows that two cells, distinguishable by the nature of their plasma-products, must therefore contain different kinds of protoplasm. There are other possibilities to be reckoned with. Weismann himself knows that there is no logical necessity for the conclusion, for he himself suggests another possibility in the following: 'If we wished to assume that the whole of the determinants of the germplasm are supplied to all the cells of the entogeny, we should have to suppose that differentiation of the body is due to all the determinants except one particular one remaining dormant in a regular order, and that, apart from special adaptations, only one determinant reaches the cell, viz., that which has to control it. If, however, we do make the assumption,' etc. (p. 63, English edition).

Here, then, Weismann himself points out that what in other places he has attempted to represent as a necessary conclusion is but one of two alternatives.

Not only does he grant the possibility of the alternative, but uses it himself in explanation of the phenomena of reproduction and development. He attributes to certain series of cells, in addition to the active rudiments controlling the normal characters of their protoplasm, the possession of numerous latent rudiments which become active when opportunity presents itself.

This non sequitur in his argument Weismann excuses with the remark that the presence of latent rudiments in special cases 'depends, as I believe, upon special adaptations, and is not primitive, at any rate not in higher animals and plants. Why should Nature, who always manages with economy, indulge in the luxury of always providing all the cells of the body with the whole of the determinants of the germplasm, if a single kind of them is sufficient? Such an arrangement will presumably have occurred only in cases where it serves definite purposes' (p. 63, English edition). Here, again, is a rhetorical flourish instead of a proof.

But the dilemma which we are examining is not yet at an end. Supposing for the moment that we accept the assumption that different character in cells implies different character in their nuclear matter, we have at once a new and important decision to make. Does the nuclear matter in the different cells, that has arisen by division from the nuclear matter of the egg-cell, become unlike by the process of division itself? or is it only after the division that it becomes different, and in consequence of the action of outer forces upon the nuclei?

Weismann decides boldly—but again without bringing forward proof—in favour of the former interpretation. 'For the chromatin,' he remarks,[10] 'cannot become different in the cells of the fully formed organism; the differences in the chromatin controlling the cells must begin with the development of the egg-cell and must increase as development proceeds; for otherwise the different products of the division of the egg-cell could not give rise to entirely different hereditary tendencies. This is, however, the case.' Weismann represents to himself that[11] 'the changes of the idioplasm depend on purely internal causes, which lie in the physical nature of the idioplasm. In obedience to these, a division of the nucleus accompanies each qualitative change in the idioplasm, in which process the different qualities are distributed between the two resulting halves of the chromatin rods.'

I shall proceed to show that this conception involves material difficulties and contradictions. It will be found that characters totally contradictory are ascribed to Weismann's idioplasm. On the one hand, it is credited with being a stable substance, possessing a coherent, complicated architecture; in the form of ancestral plasms it is supposed to be handed on, from one individual to another, unchanged through many generations; on the other hand there is ascribed to it a labile architecture, that allows a free and perpetual casting loose of rudiments, of such a kind that at each division there is caused a complete rearrangement and unequal division of these rudiments. In the one case, the inner forces produce a reciprocal, coherent bond between the numerous rudiments; in the other case, permit change of their position and relations to one another, and this not only once but in orderly, definite fashion, different in each of many successive divisions, so that the id comes to possess a completely altered architecture. 'Each id in every stage' (p. 77 of the English edition), has its definitely inherited architecture; its structure is a complex, but a perfectly definite one, which, originating in the id of germplasm, is transferred by regular changes to the subsequent idic stages. The structure exhibited in all these stages exists potentially in the architecture of the id of germplasm: to this architecture is due, not only the regular distribution of the determinants—that is to say the entire construction of the body from its primary form.'

Unfortunately, Weismann's hypothesis tells us nothing at all about these internal causes, that depend upon the physical nature of the idioplasm; that is to say, nothing at all about the causes which, working in a fashion so contradictory and astonishing, really produce the whole development.

In such a state of affairs it is better to turn to Nature herself; and to see whether or no the occurrence of differentiating division of the nucleus in the organic world is at all supported by the actual observations and investigations of those who study cells.

We shall examine (1) Unicellular organisms; (2) Lower multicellular organisms; (3) The phenomena of generation and regeneration; (4) alteration of structural growth due to external interferences (heteromorphosis); (5) A number of physiological indications that cells and tissues, in addition to their patent characters, contain latent characters which have reached them by doubling division, and which are representative of the species.

FIRST GROUP OF FACTS.—UNICELLULAR ORGANISMS.

Doubling division alone exists, or could exist, among unicellular organisms. The maintenance of the species depends upon this. Our belief that a species produces only its own species, that like begets only like, a belief that finds continual confirmation all through the study of systematic and embryological natural history, would disappear, were it possible that in the division of unicellular organisms the hereditary mass should be split into two unequal components and be bestowed unequally upon the daughter-cells. All research shows that unicellular fungi, algæ, infusoria, and so forth, in dividing, transmit specific characters so strongly and in detail so minute that their descendants, a million generations off, resemble them in every respect. No one has doubted the fact, and Weismann himself recognises that division, among unicellular organisms, is always doubling. The process of division, as such, appears never to be the means by which new species are called into existence among unicellular organisms. This is a fundamental proposition of cell-life, not to be doubted, and to be taken into account in the presentation of theories of heredity.

From the proposition that like begets only like the corollary by no means follows that mother- and daughter-cells must appear identical from the beginning. For the identity under consideration belongs only to the substance that is the bearer of specific characters, to the hereditary mass; besides that, a unicellular organism contains other substances, substances that change from time to time during its life. Many unicellular organisms pass through a regular series of developmental stages; the stages themselves being inherited, and following each other as infallibly as in the case of embryonic stages of higher animals.

The following will serve as examples of this. Podophrya gemmipara, an Acinetan, in the adult condition is attached by a long stalk, while the free end, at which is the mouth, is provided with suctorial tentacles. It reproduces by giving rise to many little buds, ciliated on the upper surface like free-swimming, hypotrichous infusoria. These, in appearance, are quite unlike the parent organism, and, after a vagrant existence in the water for some time, they attach themselves to a surface and produce a stalk, tentacles with suctorial pseudopodia, and so for the first time attain the maternal form.

Some Gregarines are large, jointed cells, divided into two pieces, a protomerite and a deutomerite; they are clad with a cuticle, under which lies a layer of muscular fibrils. After conjugation they encyst, the nucleus divides, and they break up into numerous peculiarly-shaped boat-like structures, (pseudonavicellæ), which afterwards are set free as small, sickle-shaped embryos. These exceedingly small germ-cells afterwards develop into the very different, adult gregarine-cells.

If the characters of a species be associated with a hereditary mass, an actual substance that is handed on from the parent-cell to the offspring, it is clear that the infusoria-like vagrant young of the Acinetan, and the sickle-shaped embryos of the Gregarine possess it, although for some time they are quite unlike the parent organism. For at last they become an Acinetan or a Gregarine, exactly like the parent-cell from which they arose as embryos.

These circumstances, among unicellular organisms, are a weighty indication of the error of concluding, with Weismann, in the case of multicellular forms, that because cells are unlike in outward appearance, the hereditary mass, or, as I call it, the nuclear matter, within them is also unlike. Such an assumption would involve us in the greatest contradictions. For the supposition that the nucleus is the hereditary mass transmitting the characters of the species necessitates the conclusion, in the case of unicellular forms, that the hereditary mass remains in possession of all the rudiments of the cell while it passes through the various phases of its cycle of development. Otherwise, these phases would have to be acquired anew in each case. We must, therefore, represent the possibilities of exchange between the nucleus, in its capacity of bearer of the hereditary mass, and the protoplasm as being such that all the rudiments are not simultaneously in activity, but that some of them can remain latent for a time.

SECOND GROUP OF FACTS.—THE LOWER MULTI-CELLULAR ORGANISMS.

Although in the development of unicellular organisms the way by which like begets like is plain and intelligible enough, at least in the cases dealt with, it is different with multicellular organisms, which have reached a higher grade of development. Among them we have to do with a continuous process of development, in which the highly-differentiated, multicellular organism arises from an egg, and in turn gives rise to an egg, and so on in unending sequence. But the succeeding stages of the sequence are so exceedingly dissimilar in appearance that the question how one step of the series turns into the next, and, above all, the question how the similarity of organisms, separated by the egg-stage, can be transmitted through the egg-stage, form the deepest riddle offered to biological investigation. Here, in a completeness so wonderful that our intelligence can hardly apprehend it, are presented to us the qualities of the organic material of which cells are made. Here lies that dark secret into which the various theories of generation try to direct a beam of light, and seek to find out the direction in which explanation may be found.

An intermediate stage which may serve towards the explanation of these circumstances is presented by the lower multicellular organisms, such as threadlike algæ, fungi, and other simple creatures. In them cells arise by division from the egg or from the spore, and become united into an individual of a higher rank; these cells resemble one another so completely in appearance and in qualities that there can be as little doubt as in the case of unicellular organisms that they arose by doubling division.

It is certain, then, that there exist multicellular bodies, often consisting of many thousand cells, in which each part retains the qualities of the egg from which it arose by doubling division, and which, as that method implies, possess the rudiments of the whole of which each is a part.

In this category there naturally fall the multinucleated masses of protoplasm, sometimes highly organised, in which every nucleus, surrounded by a shell of protoplasm, is capable of reproducing the whole. I am thinking of the slime-fungi (Myxomycetes), with their peculiar formation of reproductive bodies; of the 'acellular plants,' which in some cases closely resemble multicellular species in their formation of leaf and root, and in their mode of growth, as, for instance, Caulerpa, the multinucleated Foraminifera and Radiolarians. For, according to our definition of the cell, a multinucleated organism potentially is a multicellular organism.

In this matter Weismann has assumed a position which leads to peculiar consequences. In his opinion, somatic cells and germ-cells were sharply distinct at their first appearance in evolution, and have remained so ever since. Transitional forms between them are nowhere to be found. It would be inconsistent with his theory of the germplasm had somatic cells contained germplasm as their idioplasm, even when the soma first came into existence. The phyletic origin of the somatic cells depended directly upon an unequal separation of the determinants contained in the germplasm. It would totally contradict his presentation if the somatic cells, even at their first origin in phylogeny, contained, in addition to their patent special qualities, the qualities common to the whole species in a latent condition.

Weismann's conception, therefore, implies that many of the lower multicellular organisms, having no somatic-cells, have no body. Take the closely-allied creatures Pandorina morum and Volvox globator, which Weismann himself brings forward as instances for his view; the latter has a body, the former has no body, as all its cells are able to serve for reproduction!

It is enough to have pointed out how contradictory are the interpretations in this matter. Enlarging upon them may be postponed at present, for we are concerned here not with the interpretation of individual cases, but with the principles involved in the question, and, therefore, we must pass on to show further reason for considering the existence of differentiating division highly improbable in the whole organic world.

THIRD GROUP OF FACTS.—THE PHENOMENA OF REPRODUCTION AND REGENERATION IN PLANTS AND ANIMALS.

The numerous phenomena of reproduction and regeneration appear to support the principle of doubling division—that is, of division in which the germinal substance is handed on to every part of the organism. Our review may be short, as the phenomena are matters of common knowledge.

In nearly all plants there exist, widely spread through the body, cells and cell-groups, which may be induced, by inner or outer influences, to give rise to a bud; the bud grows out into a shoot, ultimately producing flowers and genital products. Such happens both in parts of the plant above the ground and below it; in the latter case shoots arise from roots, and reproduce the species in the ordinary sexual fashion by bearing sexual products.

Thus, in the case of Funaria hygrometrica, a little moss, one may chop up the plant into tiny fragments, scatter these on damp earth, and see numerous moss-plants reproduced from the little groups of cells. By cutting little pieces from a willow, an experimenter may cause the production from slips of thousands of willow-trees, each with all the characters of the species, so that there must have been contained in each of the little pieces of tissue hereditary masses with the characters of the whole plant. Separate pieces of the leaves of many plants, as of the begonia, produce buds from which the whole plant may grow out.

An aptitude for reproduction like that in plants exists in many cœlenterates, worms, and tunicates. The polyps of hydroids and of bryozoa, the stolons of an ascidian (Clavellina lepadiformis), may give rise to buds in many places, and these grow up into the perfect hydroid, bryozoon, or ascidian. There must, then, be contained in the cells of the bud the germinal rudiments of the whole animal; this conclusion is more necessary as the individuals, produced from the buds, in due course bear sexual products.

Although in many higher animals and plants one sees that cells with the capacity for reproduction are limited to special areas, still, the capacity for regeneration often is very great. In a wonderful fashion animals will reproduce lost parts, sometimes of most complicated structure; just as a crystal, from which a corner has been chipped, will perfect itself again when brought into a solution of its own salt. A Hydra, from which the oral disc and tentacles have been cut off, a Nais deprived of its head or of its tail, a snail of which a tentacle with its terminal eye has been amputated, will reproduce the lost parts, sometimes in a very short time. The cells lying at the wounded spot begin to bud, producing a layer or lump, the cells of which resemble embryonic cells. From this embryonic mass of cells the lost organs and tissues arise—in Hydra, the oral disc with its tentacles; in Nais, the anterior end with its sense-organs and special groups of muscles; in the snail, the tentacle with its compound eye built up of elements so different as retinal-rods, pigment-cells, nerve-cells, lens, and so forth.

Even among vertebrates, in which the capacity for regeneration is the least, as in the restoration of the wounded parts small defects occur, lizards can reproduce a lost tail, tritons an amputated limb. From a bud of embryonic tissue there are elaborated in the one case whole vertebræ, with their muscles and tendons, and part of the spinal cord with its ganglia and nerves, in the other case, the numerous, differently-shaped, skeletal pieces of the hand or foot, with their appropriate muscles and nerves. The regeneration, moreover, is in strict conformity with the characters of the species concerned. Thus, from the facts of regeneration also, we must infer that cells in the vicinity of these casual wounds possess not only the special qualities which they possess as definite parts of a definite whole, but also the characters of the whole, and thus have the power of becoming buds, from which a complicated part of the body may be reproduced with the appropriate characters of the species.

FOURTH GROUP OF FACTS.—THE PHENOMENA OF HETEROMORPHOSIS.[12]

Of all the facts brought forward here, the phenomena of heteromorphosis perhaps bear most strongly in favour of my conception, and offer difficulties most irreconcilable with Weismann's theory.

Loeb uses the word 'heteromorphosis' to denote the ability possessed by organisms, under the stimulus of external forces, to produce organs on parts of the organism where such do not occur normally, or the power to replace lost parts by parts unsimilar to them in form and function. Regeneration is the reproduction of parts like those lost; heteromorphosis is the reproduction of parts unlike those lost.

Heteromorphoses are well known in plant physiology. When one cuts a slip from a willow, one may make the cut at the bottom of the slip and the cut at the top in any part of the willow-twig, yet still the lower end of the slip always produces rootlets, which are organs not normal to that part of the twig, while shoots will rise from the upper end. Moreover, either end of the slip may be made the root portion, and it is clear, therefore, that in every small area there are cell-groups present able to bear roots or shoots according to the determining conditions; and therefore that, in addition to the characters active at any time, there are present the germinal rudiments for shoots and roots, and, indeed, for the whole organism, since the shoots ultimately may bear genital products.

When the prothallus of a fern has developed normally, it is a flattened leaf-like structure which bears rootlets and male and female genital organs on the lower surface, i.e., on that turned from the light. But the experimenter may reverse this order, by artificially shading the upper surface, and strongly illuminating the lower surface.

Among the most interesting heteromorphoses are the galls, produced upon young plants when certain insects lay eggs on them, or when plant-lice irritate their tissues. From these abnormal stimuli there result active masses of cells which grow into organs of definite form and of complex structure. The galls, moreover, differ widely, in correspondence with the specific stimulus which was their initial cause, and with the specific substance, the stimulation of which resulted in the formation of a gall. By the action of different insects upon the same plant different galls are produced, and the galls of different plants may be distinguished systematically.

Blumenbach has already brought forward the existence of galls as an argument against preformation, holding them to be structures produced epigenetically, and, therefore, unrepresented by rudiments in the germ. I, also, consider them witnesses against Weismann's germplasm. They teach us that the cells of the plant-body may serve purposes quite different from those arranged for in the course of development; that cells modify their form in correspondence with novel conditions, and that they are forced into forming special structures, not by special determinants in the germ, but by external stimulants.

Galls exhibit yet another instructive kind of heteromorphosis.

Even the tissue of a leaf, turned into a gall by pathological conditions, retains the power of producing roots. Beyerinck has shown that galls of Salix purpurea, planted in moist earth, bear rootlets identical with those of the normal plant. As the roots of all woody plants are able to bear adventitious buds, De Vries thinks it probable that one could rear a whole willow-tree from a gall. That would imply that all the inheritable characters of the willow were contained even in the gall.

Loeb has produced heteromorphoses experimentally upon many lower animals, among which were Tubularia, Cerianthus, and Cione intestinalis.

In Tubularia mesembryanthemum, a hydroid polyp, there are stalk, root, and polyp-head. If one cut off the head, a new head will be formed in a few days, this being a case of regeneration. On the other hand, a heteromorphosis may be produced by modifying the experiment as follows: Both root and head must be cut from the stem; if the lopped piece of the stem be stuck in the sand of the aquarium by the end that bore the head, then the original aboral pole in a few days produces a head; if the lopped piece of stem be supported horizontally in the water, then each end of it produces a head.

In a Cerianthus membranaceus (Fig. 1), the body was opened by a cut some distance below the mouth, whereupon buds appeared on the lower edge of the slit, where the experimenter had prevented coalescent growth. These buds gave rise to inner and outer tentacles, and an oral disc was produced. Thus, artificially, an animal with two mouth-openings or two heads was produced; and, similarly, animals with a row of three or more heads may be produced.

Fig. 1.—Cerianthus membranaceus, in which a second oral aperture, surrounded by tentacles, has appeared as the result of an artificial slit. (After Loeb.)

Fig. 2.—Cione intestinalis, in which eye-specks resembling those surrounding the mouth have appeared in the neighbourhood of an artificial opening (a).

The third animal in which heteromorphosis was produced artificially was Cione intestinalis, a solitary ascidian, an animal more highly organized. In Cione (Fig. 2) the edges of the mouth-opening and of the cloaca are provided with numerous, simple eye-spots. Loeb, in a series of experiments, made incisions either into the inhalent or the exhalent tube; after a time eye-spots appeared round the edges of the cut; then the margin of the artificial oral opening grew out into a tube, even longer than the normal oral tube. 'If several incisions be made simultaneously at different places on the same animal, then several new tubes arise simultaneously.'

In the three cases, the cut surfaces, from which in Tubularia, a head, in Cerianthus, tentacles, and in Cione, eye-spots, took their origin, were made in different parts of the bodies and in different directions. Thus, again, we have an indication that there are present in most regions of the body cell-groups, which may give rise to complex organs in unnatural positions, and yet bearing the specific stamp.

These examples might easily be multiplied, and they serve to show that heteromorphosis in plants and animals implies the presence of numerous latent characters in cells and tissues, in addition to the characters proper to their normal position in the organism. These latent characters, under the impulse of stimulation from without, manifest themselves in abnormal formation of organs in abnormal situations. Save that they are in abnormal situation, the induced organs conform to the specific type in all respects, and indicate that all the cells of an organism contain, as the result of doubling division, the characters of germinal rudiments of the whole organism. On the other hand, heteromorphoses bear heavily against the doctrine of determinants. For it is impossible that, in the architecture of the germplasm, there can be provision, in the form of special determinants, for events so foreign to the natural course of development as these arbitrary, outer stimulants.

Heteromorphosis may be extended to include more than Loeb intended by reckoning under it artificially-produced modification of the early stages in the cleavage of the egg. I have in mind those experiments by Driesch, Wilson, and myself, in which the first cells of the embryonic history were induced to form parts of the embryo, to which in the normal course they would not have given rise. In these cases heteromorphosis begins from the first cleavage of the egg.

In an ingenious way Driesch compressed fertilised echinoderm eggs between glass plates, and so secured that the first sixteen cells were separated, not by alternate vertical and horizontal planes, as in the normal development, but only by vertical planes. In the resulting one-layered plate of cells the nuclei had relative positions quite different from the normal. As, notwithstanding this, the distorted eggs developed into normal plutei larvæ, Driesch inferred that the cell material composing the earliest cells of echinoids is equivalent in all the cells, and that the cells may be pushed over one another like a heap of balls without disturbing in the slightest their capacity to develop. Such a permutation could be without injury to the developmental product only if one nucleus had the same qualities as another; that is to say, only if all the nuclei had arisen from the nucleus of the fertilized egg by doubling division.

Driesch is right to regard these experiments as incompatible with Weismann's theory. 'Only consider,' he remarks, 'how great a number of "supplemental hypotheses," how many "accessory determinants," would be required to make specification of the early stages of a development in which any nucleus may take the place of any other nucleus in the whole embryo.'

I myself have carried out similar experiments upon frogs' eggs—experiments with a double interest. The frog's egg has the poles different, and so has a definite orientation. Weismann and Roux themselves have used these objects to support their view that, at the first cleavage, nuclei with different qualities are formed.

On p. 64 of the English edition Weismann remarks: 'The fact that the right and left halves of the body can vary independently in bilaterally symmetrical animals points to the conclusion that all the determinants are present in pairs in the germplasm. As, moreover, in many of these animals—e.g., in the frog—the division of the ovum into the two first embryonic cells indicates a separation of the body into right and left halves, it follows that the id of germplasm itself possesses a bilateral structure, and that it also divides so as to give rise to the determinants of the right and left halves of the body. This illustration may be taken as a further proof of our view of the constant architecture of the germplasm.'

Roux[13] has based his mosaic theory upon experiments upon frogs' eggs. According to the theory, the first two segmentation spheres contain not only all the formative material for the right and left halves of the embryo respectively, but also the differentiating and elaborating forces for these, so that on the destruction of one cell, the other can give rise only to one lateral half of the embryo (hemiembryo lateralis). Roux, therefore, considers that by the first cleavage the nuclear material is broken up into unlike halves, by which the development of the corresponding cells is directed diversely, i.e., is determined in a specific fashion.

Fig. 3.—Diagrams of the Eggs of Frogs, which show how alteration of the cleavage process changes the mode in which the nuclear material is distributed. The nuclei indicated by the same numbers have the same descent in all the diagrams. All the eggs are viewed from the animal pole. A. Normally developing eggs. B. Eggs developing under compression by horizontal plates. C. Eggs developing under compression by vertical plates.

The error in these representations of Weismann and of Roux has been shown by varied experiments of my own. The eggs of frogs on the point of cleaving were flattened to a disc between vertically or horizontally placed glass-plates. In the first case they were flattened in the dorsoventral direction, i.e., the axis passing through the animal and vegetative pole was shortened; in the second case an axis at right angles to this was shortened. In both cases the course of cleavage, and the resulting distribution of the nuclei in the yolk, was artificially modified.

The diagrams A, B, C (Fig. 3) will make the results plain to the reader. A, represents the distribution of the nuclei after normal cleavage; B, the same, when the egg was pressed between horizontally-arranged parallel glass-plates; C, the same, where the flattening was produced by vertically-placed parallel glass-plates.[14]

The diagrams show the positions of the segmentation spheres and of the contained nuclei as seen from the animal pole. In stages where two layers of cells as a result of division lay one above the other, the cells of the lower layer are distinguished in the figure by shading. In the three diagrams the nuclei are numbered so that the reader may know how far they are removed from the nuclei of the first two segmentation spheres. The numbers are further exhibited in the following two genealogical trees:

In the three diagrams the nuclei with the same numbers have the same rank in descent, and therefore, according to the theory of Roux and Weismann, have the same qualities, while the nuclei with unlike numbers differ in qualities.

Let us now notice how the nuclei in the three processes of division, of which two are abnormal, are placed in the mass of the egg.

After the first division, the nuclei are alike in all three cases; after the second difference appears. In A1 and B1 nuclei 3 and 5 lie to the left; 4 and 6 to the right of the second cleavage-plane, which, according to Roux's hypothesis, corresponds to the median-plane of the future embryo; while in C they are forced into two layers, one above the other, nuclei 4 and 6 being dorsal, 3 and 5 ventral.

In the third cycle of division there is no agreement between the three cases.

In the diagrams A2 and B2 the nuclei still lie similarly to the right and left of the middle line; but in A2 they are arranged in two layers, in B2 in a single layer. The nuclei 8, 10, 12, and 14, which compose the upper layer in A2, form the middle of the disc in B2; and 7 and 9, 11 and 13, the ventral nuclei of A2, occupy the ends of the single-layered disc of B2, being closely pressed against each other.

In the diagram C2 there is actually no median-plane after the third cycle of division. The nuclei 9, 10, 14, 13, which in A and B form the right side of the mass, here form a dorsal layer with nuclei 7, 8, 12, 11, forming a ventral layer. In the fourth cycle of division the nuclear matter is still more variously distributed through the mass, as may be seen from comparison of diagrams A3, B3, C3.

Although, under normal conditions, the multiplication and division of the nuclear material occurs in an almost invariable and definite fashion, the mere altering of the spherical form to a cylinder or to a disc produces a method of division completely different, so far as the nuclei are related to each other in a genealogical tree. In the one and the other method of division the nuclei are brought into relation with different regions of the protoplasmic mass, and are united with these regions to form cellular individuals.

I had quite enough reason for what I said in my essay: 'If the doctrine of Roux and Weismann be true, and the successive divisions by which nuclei arise really place different qualities in the nuclei—qualities according to which the masses of protoplasm surrounding them become different and definite parts of the embryo—what a pretty set of malformations must result from eggs in which the nuclear matter has been shuffled about so wantonly! As such malformations do not occur, it is plain that the doctrine is untenable.'

We reach the same conclusion from consideration of the interesting experiments made by Driesch and Wilson upon the early stages of segmentation of the egg. In the cases of an echinoid and of amphioxus (Fig. 4) they succeeded in shaking apart the first two and the first four cells that arose in division of the egg; and they traced the subsequent development of these separated segmentation spheres.

Fig. 4.—Normal and Fractional Gastrulæ Amphioxus.
(After Wilson.)
A Gastrula from a whole egg; B, C and D, gastrulæ from single cells artificially separated, (B) from the two-celled stage, (C) from the four-celled, and (D) from the eight-celled stages of normal development.

From one of the first two segmentation spheres of an echinoid egg, Driesch was able to rear successive embryonic stages (Gastrula and Pluteus), which were normal in shape, but one-half the usual size. Wilson's results, obtained by shaking apart the segmentation spheres, were even more interesting, as they were performed upon amphioxus, a more highly-organized animal. He reared gastrulæ and older embryos with notochord and nerve-tube, which were perfect and normal, except in size. They were one-half, one-quarter, or one-eighth of the usual size, according as they were reared from cells isolated from the two, four, or eight-celled stage of the segmenting egg.

Results which Chabry and I gained by destroying, by puncture, one of the first two segmentation spheres, assist the present argument. Although one-half of the mass had been destroyed, Chabry obtained, in the case of an ascidian, and I obtained, in the common frog, embryos with notochord and nerve-plate. These developed directly and normally, although, in the case of the frog, there was a slight defect at the ventral posterior part of the body, where the arrested protoplasmic mass came to lie.

All these experiments show that the first two (and in some cases the first four) results of division can assume a quite different bearing as regards their function in the mechanical building of the embryo, according to whether they remain bound with each other into a whole or are separated and develop by themselves. In the former case, each forms only one-half (in some cases only a fourth) of the whole. In the latter case, each by itself produces the whole. The half and the whole, then, of the first cleavage-cells are identical in real nature, and, according to the circumstances, can develop, now in this way, now in that.

Even if Weismann were to admit the correctness of these experiments, perhaps he would not consider that they contradicted his theory of the germplasm and the segregation of the hereditary mass, but would make a supplemental hypothesis, which, from the spirit of his theory, could be none other than this: each of the first cleavage-cells, in addition to its specific part of the hereditary mass, the part that controls its normal course of development, possesses an accessory idioplasm, an undivided fragment of the germplasm, left behind to be ready for unforeseen emergencies; this part takes command when, in consequence of violence, a separated part develops into the whole.

But such an assumption does not go far enough, if it be confined to the first cleavage-cells. By compression of the frog's egg, I have shown that the pole passing through the blastopore, which coincides with the chief axis of the future embryo, may assume different relations to the first segmentation-plane, sometimes coinciding with that, sometimes making a right or an acute angle with it. It is clear that in each of these cases the embryonal-cells take a different share in the formation of the regions of the body, and that they must be fore-endowed with the capacity of playing different parts.

The developmental history of double monsters enforces the same doctrine; such are common among the embryos of fish, and rather less common among chicks. From causes of which we are ignorant two, instead of one, gastrula stages may arise at separate regions of the germinal layer of the egg. According to the position of these two invaginations, which may be regarded as crystallisation-points for the formation of the future embryo, the cells of the germinal disc will be drawn into the process of development, and, falling into groups, will build up organs. In relation to this double gastrulation, there may arise, for instance, four instead of two primitive ears, eyes, and nasal organs; and these arise from cell-groups, the choice of which is determined by their relation to the position of the gastrula-invagination.

From various other experiments, conducted so as to distort the normal course of development, I have obtained parallel results.

Taking frogs' eggs immediately after fertilisation, I compressed them strongly between parallel, horizontally placed glass plates. I then inverted them, so that the vegetative pole came to lie uppermost. In spite of their unnatural relation to gravity, they developed further, and became abnormal, quite unsymmetrical embryos.

In another experiment, taking a triton's eggs after they had divided into two spheres, I surrounded them with a silk thread in the plane of the first cleavage, and tightened the thread until the embryo assumed the form of a sand-glass. The deformity of the resulting larvæ was very different, and perhaps depended on the tightness of the constriction. Some became greatly elongated, and had developed so that the thread surrounded the dorsal nerve-cord. In other cases the dorsally-placed organs arose only from one-half of the sand-glass-shaped embryo, while the other half gave rise to the ventral part of the body. In this case the dorsal organs (nerve-tube and notochord) were doubled over like a snare, the head and tail ends, the mouth and the region of the anus, being bent in at the position of the constricting thread.

The important point is that in both the experiments, in the case of the frog and of the triton, the cell-material, separated at the first cleavage, was turned to a use quite different to its use in the formation of a normal embryo.

We may conclude with a very convincing proof. In the above-mentioned abnormal development of the frog's egg it happened that one edge of the blastopore, on account of its weight, was very much bent outwards. In consequence of this the cleft of the blastopore lay between the normal blastopore-lip and the everted border of the other lip. When the notochord and the nerve-plate appeared, as a result of this abnormal condition, they grew from a cell-material that was quite different to that which gives them origin in normal cases.[15]

In these cases Weismann cannot apply his accessory conception, the existence of supplementary idioplasm, only to the nuclei arising from the first division; he must extend it to the thousands of embryonic cells that arise by division up to the time for the appearance of the nerve-tube and notochord. The behaviour of these cells under fortuitously changed conditions shows them all to be endowed with the capacity of development in different directions.

FIFTH GROUP OF FACTS.—PHENOMENA OF VEGETATIVE AFFINITY.[16]

Many considerations, taken from the region of general physiology, support the view that all the cells of an individual, of any species, are alike, and are to be distinguished from one another only by the special development of one character.

Formerly, indeed, many biologists, relying upon the optical appearances presented in microscopical investigation, have been inclined to the view that the visible qualities of a tissue, as revealed by the microscope, were the only, or the chief, distinctive characters. For instance, by microscopical investigation one cannot distinguish the tendons, nerves, bones, and cartilages of a dog from the corresponding tissues in a horse. So far as their special use in the organism goes, one might interchange the corresponding parts in these two mammals. A tendon from the dog, if large enough, might be attached to the muscle of a horse, and would transmit the pull of the muscle on the bone just as well, and would completely satisfy the mechanical duties of the horse's tendon. The same might happen in the case of a bone, of a cartilage, or of a nerve-fibre.

As a matter of fact, the idea that parts of the tissues of different animals may serve to replace one another has been employed repeatedly in science, especially in the science of medicine. But I believe that our ideas are not yet clear upon the matter. The erroneous impression to which I have alluded has arisen because we do not bear in mind that each tissue, each part of an organ, each cell, possesses, in addition to its obvious characters, very many characters that are invisible to us. Such characters are inherent in the tissue-cells because these are parts of a definite organism. In consequence of their specific tissue characters, which are visible to us, we assign cells their place in histological classification; in contrast, we may denote the other characters as constitutional, or species, characters.

No doubt tissue cells are in the same case as genital cells. So far as microscopical characters go, egg cells and spermatozoa are wonderfully alike in all the mammalia; in many cases we could not distinguish between those of different animals. But, because they bear the specific characters, we cannot doubt but that they are as distinct as are the species, although invisibly to us.

The products of the sexual cells show us clearly enough that out of each kind of egg only its own species of organism can be developed. Certainly it is not so plain that, besides their visible microscopical characters, the tissues and organic parts are in possession of more general characters, identical in all the differently-specialised tissues of a single organism; but we may infer the existence of such latent characters, at least partly, from the results obtained, in the case of plants, by grafting, in the case of animals, by transplantation and transfusion.

In the case of plants one may graft a twig cut from one tree upon the stem or lower part of another tree of the same kind, and so bring about a firm and lasting union between the two. In a short time the corresponding tissues of the parts brought into connection quietly unite. Thus from two different individuals a single living organism may be produced artificially.

One would expect, therefore, that a twig and stem, chosen from two closely allied species, such as, for instance, the pear and the apple, would unite when the suitable tissues were put together. But this does not happen. Successful grafting depends far less on the conjunction of obviously appropriate parts than upon characters unrecognisable by us, such as deep-seated kinship between the parts, and the specific characters of their cells; while in the case of individuals of the same species two pieces will unite even if they are not brought together in appropriate conjunction, or when they belong to different parts of the organism, as, for instance, to the root and the leaf; yet in the absence of deep-seated kinship union will not take place.

Generally this kinship, which has been called vegetative affinity, depends, like sexual affinity, upon the degree of systematic relationship. It appears that the same condition of things occurs as when, in ordinary fertilisation, sexual cells from different varieties, or species, are united. In both cases it happens, on the average, that union is the more to be expected the more closely the plants concerned are akin, in a natural system of classification.

But in grafting, as in cross-fertilisation, unexpected exceptions to this rule occur. Relying upon these, Naegeli thought that the external distinguishing tokens do not always indicate correctly the intrinsic constitutional differences. Frequently union will not take place between plants most near akin in classification, most alike in external characters; while it will occur between plants most different in outward aspect and belonging to different genera or even families. In other words, external characters give no certain index to the degree of vegetative affinity or of sexual affinity between two kinds of plants.

As an example of this, Vöchting, in his treatise upon transplantation of plant-tissues, takes the tribes of pear-trees. Grafting between these and apple-trees takes place only with difficulty, although the apple is a close kinsman and belongs to the same genus. On the other hand, most of them graft easily upon the quince, although that belongs to a different genus. In this case, also, there is no sexual affinity between the pollen and the ova. Hybrids are not formed between the pear and the apple.

It seems probable to me, although as yet I cannot get complete proof of it, that sexual and vegetative affinity, that is to say, the relationship between the egg-cell and the pollen of two species, and the relation between twig and stem, depend upon the same intrinsic qualities of that elementary organism the cell.

Vöchting distinguishes as harmonic or disharmonic the modes of union between twig and stem, according to whether or no they reach the formation of functional unity. Among cases of disharmony there are several interesting gradations. Generally speaking, in the case of plants not adapted to each other, no attempt at union occurs, and the grafted twig speedily perishes; sometimes even the stem dies, as if it had been poisoned by the graft. In other cases the disharmony is not shown so strongly. The twig and the stem begin to unite, but, sooner or later, disturbances occur, and complete destruction results. According to Vöchting, in the case of some Cruciferæ the disturbances are as follows: the twig begins to form roots at its lower end, and these grow into the stem of the host. Through them the twig uses as food the juices and salts of the stem, refusing to unite with the stem so as to form a single individual. As Vöchting says, this formation of roots simply is an attempt on the part of the twig to complete its own individuality. Instead of growing into corporate union with the stem, the twig attempts to become a parasite upon it. A further consequence often is, that the stem, too, begins to respond to the unadaptive stranger's influence. Thus, when Vöchting grafted a Rhipsalis paradoxa on an Opuntia labouretiana, he found that round the roots of the graft the tissues of the host threw out a protective sheath of cork, or turned in places to a gelatinous mass.

In some cases experimenters have overcome disharmony between two species, A and B, by making use of a third species, C, with a vegetative affinity for both A and B. Thus, an intermediary between the two disharmonic forms is made, and by such an arrangement a single functional individual is produced from pieces of three different species. Thus, upon A, as stock, a shoot of C is grafted, while upon this shoot of C, as stock, a shoot of B in turn is grafted.

In the matter of these different grades of disharmony, a comparison may be made between sexual and vegetative affinities. In many cases the spermatozoa of one species will not impregnate the eggs of another species. In other cases, the alien spermatozoon may penetrate the egg and unite with its nucleus, making, however, an unsatisfactory combination in various degrees of infertility. Sometimes the fertilised egg divides only a few times and then dies; sometimes development proceeds to the stage of the blastula, the gastrula, or even further; but it then comes to an end, through intrinsic causes beyond our ken, and, finally, complete destruction follows.

Our acquaintance with what happens in transplantation of animal tissues is smaller than in the sphere of botany.

Long ago, Trembley attempted to cause, by grafting, the union of two pieces of hydroid polyps into a single individual. He divided, across their middles, two specimens of Hydra fusca, and then, in a watch-glass, applied the upper end of one to the lower end of the other. In one case he was rewarded by the occurrence of complete union; for, after a few days, on feeding the upper end with a worm, it was passed on into the lower end. Later on buds arose, both above and below the point of union. Trembley, however, was unable to graft on each other parts of different species, parts of the green hydra, Hydra viridis, upon the common hydra.

Transplantations of single tissues or organs have been made more often, and by several investigators. I shall mention only the older results of Ollier and M. Bert, and those made in 1893 by A. Schmitt and Beresowsky.

Ollier exposed the bone of an animal, and, carefully removing a part of the periosteum, planted it in the connective tissue under the skin in another part of the body. The consequences differed according as the transplanted tissue was imbedded in another animal of the same species, or of another species. In the first case the piece of periosteum grew, obtaining a supply of blood from vessels which grew out into it from the surrounding connective tissue in which it was embedded. In a short time lamellæ of bone were formed by the layer of osteoblasts, so that a small plate of bone was formed under the skin. This, however, proved always but a temporary structure, for, being formed in an inappropriate spot, and, therefore, being functionless, it was soon reabsorbed. In the second case, however, in which the piece of periosteum was removed from the bone of a dog and planted in a cat, rabbit, goat, camel, or fowl (or vice versâ), formation of bone did not occur; either the piece of periosteum was absorbed, or set up suppuration around it, or became enclosed in a cyst.

Paul Bert's experiments were the following. He removed pieces two or three centimetres long from the tails of white rats a few days old, skinned each piece, and planted it in the connective tissue under the skin of the same animal. In a few days circulation of blood was established in the pieces of the tails, by union with vessels from the connective tissue in which they were embedded. Muscles and nerves degenerated, but the other tissues, bones, cartilages, and connective tissue, grew vigorously, so that, in animals killed and examined a month after the operation, the pieces of tail, implanted when they were two or three centimetres long, had grown five to nine centimetres long.

The result was totally different when the transplantation was made from one species to another. When the tip of the tail of a Mus decumanus or a Mus rattus was transplanted to a squirrel, guinea-pig, rabbit, cat, dog (or vice versâ), either extensive suppuration took place, and the piece was extruded, while sometimes the subject of the experiment died; or, after a less turbulent course, the alien piece was absorbed. The continuance of life and growth in the piece only took place when the two animals concerned were allied very closely. Thus success followed transplantation from Mus rattus to Mus decumanus (or vice versâ), but not when it was from Mus sylvaticus to Mus rattus.

The recent experiments of A. Schmitt and Beresowsky lead to the same conclusion. The former succeeded in making pieces of living bone 'take' only when the transplantation was from one individual to another of the same species, or to another part of the same individual. Beresowsky transplanted pieces of frog's skin to the dog and the guinea-pig, and pieces of dog's skin to the guinea-pig, and always found that they died, or were thrust out as foreign bodies.

Precisely the same results follow transfusion of blood between animals of different species. There is complete agreement among investigators. When the blood is made to flow directly from the vessels of one animal to the vessels of an animal of a different species, as from the dog to rabbit, or from dog to sheep (or vice versâ); or when it has been first freed from fibrin and then injected, the result is always the same. 'We have always found,' says Ponfick, summing up the results of the investigation, 'not only that blood of another species acts in strong doses as a poison, and in weaker or smaller doses is harmful, but that (and this seems to me my most important result) in every case the blood-corpuscles are destroyed almost completely, probably quite completely.' In a very few minutes, in the case of disharmonic kinds of blood, the red corpuscles degenerate, and the hæmoglobin, becoming dissolved in the blood-plasma, soon appears in the urine. In the case of transfusion of similar blood between individuals of the same or of very closely related species, the hæmoglobin does not appear in the urine except after very large doses; and Ponfick infers that the red blood-corpuscles, either all of them or most of them, remain unchanged in the new animal.

Landois has carried out transfusion between the remotest species, between different families of mammals, and between mammals, birds, and amphibia; from these he drew 'the inference, important for classification of animals, that those animals anatomically most nearly allied have their blood most closely alike.' In fact, 'the destruction of the foreign blood happens the more slowly the more nearly the animals are allied.' 'Thus, in doubtful cases, experiments on transfusion might settle degrees of relationship. Between individuals of the same species transfusion is a complete success; when the species are closely allied, the transfused blood disappears only very gradually, and large quantities may be transfused without harm. The further apart the animals may be, in a system of classification, the more violently the destruction of the foreign blood takes place, and the smaller is the quantity that can be endured in the vessels. Thus, in the extent to which blood transfusion may occur, I see a step towards the foundation of a Darwinian theory applied to cells.'

As yet, transplantations and transfusions between animals of different species have been considered with a view to their importance in surgery and in medicine, rather than from their purely physiological side. From the results given above, in which I believe, although there might be drawn from literature contradictory results—in which, however, I cannot feel confident—I am prepared to extend a conclusion to the animal kingdom that is better supported in botany: the conclusion that the cells and tissues possess, in addition to their definite microscopical characters, more general, intrinsic, specific characters, and, that one may speak of the vegetative affinities between tissues exactly as one speaks of the sexual affinities between reproductive cells.

SUMMARY OF THE CONCLUSIONS IN THE FIRST SECTION.

Summing up what has been said in the preceding pages, we find a large series of facts supporting our contention that cells multiply only by doubling division. First comes the fundamental circumstance that single-celled organisms exhibit only doubling division, as by that alone the permanence of species, which experience shows us to exist, is possible.

Secondly, some facts of reproduction were considered. The formation of germinal tissues, and, in the case of lower plants and animals, the occurrence of budding in almost any part of the body, are easily intelligible if every cell, like the egg-cell, has been formed by doubling division, and so contains the rudiments of all parts of the organism; and if thus, on the call of special conditions, every cell may become a germ-cell again.

Thirdly, great stress is to be laid on those experiments in which the process of development was interfered with at different stages, as these showed that the separate cells which arose by division were not predestined unalterably for a particular rôle, according to a predetermined plan (facts of regeneration and heteromorphosis).

Fourthly, the results of grafting, transplantation, and transfusion indicate that the cells and tissues of an organism possess, in addition to their patent microscopical characters, latent characters, which show themselves to be peculiar to the species.

How does Weismann attempt to reconcile his hypothesis of differentiating division with these facts? By the provision of different complementary hypotheses, which, as we have seen, amount to this, that he allows the set of rudiments which he had turned out by differentiating division of the cell to creep in again by a back-door. He accomplishes this by his idea that the germplasm may undergo, simultaneously, doubling and differentiating division. In these cases cell-division has a double aspect. According to Weismann, this is possible, because the egg contains many, sometimes as many as a hundred, ids, each of which is a combination representing the species. Weismann believes that in an egg, while it is preparing for its first division, the ids are arranged in two groups—an active army and a reserve army. By differentiating division the active army is broken up into the divisions, brigades, and regiments of determinants appropriate to the separate groups of cells, and so the course of the development is conducted according to a preconceived plan. On the other hand, the passive, reserve army multiplies by doubling division, and is sent along with definite parts of the active army as baggage in a fixed or inactive condition, so that it has no influence upon the normal course of development nor upon the characters of the cells (fixed germplasm, inactive, accessory idioplasm, bud-idioplasm).

In spite of this purely arbitrary, complementary hypothesis, the facts seem to me to show that Weismann assumed an untenable position when he attributed a reserve army of 'stable plasma' only to the sets of cells in which it was necessary to suppose its existence. The experiments of Driesch, Wilson, and myself show that a complete embryo may spring from a half or quarter of the egg, and that the set of nuclei first to arise may be shifted about in the egg like a heap of billiard-balls. In the face of such facts there seems nothing left for the theory of Weismann but to endow every cell with accessory germplasm to prepare it for unforeseen events. This, however, would sterilize the other part of the theory, the doctrine of determinants, and the mechanism of development dependent on a rigid architecture of the germplasm. Consider the confusion that would arise when the deploying of the active army was disarranged by external influences, now in one fashion, now in another, if the reserve army, with its store of latent rudiments, had to come to the help of the broken pieces. What would compel the rudiments disposed to activity according to the prearranged plan to become latent where they were no longer wanted? And what would stir into activity in the necessary places the originally quiescent rudiments of the reserve army? In fact, if the rôles of activity and quiescence are even once to be exchanged by the rudiments in the cell, what object is there in drawing a distinction so sharp between the two armies—the active army which carries out the process of development according to a plan prearranged in its minutest details, and a passive reserve army ordered into quiescence and carried as baggage?

But here we come upon the scarlet thread that continuously has traversed the theory of germplasm in all its changes. Weismann attaches the greatest importance to the distinction. The twofold nature of the process of development is a cardinal point in his theory, linked to his doctrine of immortality for unicellular organisms and germ-cells and mortality for somatic cells.

Between somatic cells and reproductive cells Weismann places a gulf that cannot be bridged. Only the reproductive cells contain real germplasm, and only these contain the conditions for maintaining the species, as they alone serve for the starting of new generations of development. The somatic cells, on the other hand, are endowed only with fragments of germplasm, and hence they are incapable of preserving the species, and are doomed to death. The reproductive cells, like unicellular organisms, are regarded as immortal, the somatic cells as mortal. According to Weismann, cells cannot pass from the one category to the other.

As I see Nature, this contrast has been artificially reasoned into her. From several reasons, I do not think that it exists. In the first place, I consider that the facts I have given show the hypothesis of a differentiating division of cells and germplasm to be not proven and arbitrary. Next, the reproductive-cells must be considered as much a part of the organism as any other tissue. Sometimes they form the greater part of the body, as in many parasites, and, like the other tissues, they are subject to death, unless the conditions necessary to their further development have occurred in time. But under such conditions other cell-complexes may have death averted from them, as, for instance, when a slip cut from a willow-tree is planted. Thirdly, the reproductive cells are derived from the egg-cell just in the same way as other tissue cells are derived from it. Like tissue cells in multicellular organisms, they arise by the specialisation of material separated from the egg-cell, and, like every other organ, attain the position assigned them in the plan of development in the course of the general metamorphosis of position that all the cells pass through. Often the sexual cells, like those of other tissues, appear at a distance of several cell-generations from the egg. The intervening generations are specially numerous in those animals and plants in which several sexless generations come between the sexual generations (e.g., many plants, cœlenterates, worms, tunicates).

I cannot agree to the existence (in Weismann's sense) of special germ-tracks. Naturally, I do not deny that the sexual cells arise from the egg after definite sequences of cell-divisions; but this happens in the case of all specialised cells, such as muscle, liver, kidney, and bone cells. The conception of special germ-tracks has no more significance than there would be in the conception of muscle, liver, kidney, and bone tracks. Though Weismann associates with germ-tracks the idea that germplasm travels along them, proof of this has yet to be brought forward.

Finally, a word about the meaning of 'immortal.' In a scientific work the word must be used in a philosophical sense. In calling a being immortal one implies both individuality and indivisibility. This, at least, was the view of the old philosophers, who have defined the idea of immortality. Thus says Leibnitz in his Theodice: 'I hold that the souls which one day become the souls of men existed already in the seed, that they have existed always in organised form in the ancestors, back to Adam—that is to say, to the beginning of things.'

In his doctrine of immortality, Weismann has not concerned himself with the two implications—individuality and indivisibility. He calls a unicellular organism immortal, simply because its life is preserved in the organisms arising from it by division. The immortality of the unicellular forms depends upon their divisibility, upon a property which, according to the philosophical use of the word, is incompatible with immortality. According to Weismann, one immortal organism gives rise to several immortal organisms, but, as these are subject to destruction by external agents, the separate individuals are mortal. The unicellular organism is not immortal in itself, but only in as much as it may give rise to other organisms. In this way Weismann comes in conflict with the idea of individuality, and is compelled to transform his conception. For he says 'that among unicellular organisms there are not individuals separated from each other in the sense of time, but that each living being is separated into parts so far as space is considered, but is continuous with its predecessors and successors, and is, in reality, a single individual from the point of view of time.' Consequently Weismann must take the same view of the germ-cells, which, according to his theory, are immortal in the same way as unicellular organisms, and, in the same sense, he must make a single individual of all the germ cells arising from a single germ cell, and, with them, of all the organisms developed out of them. Adam is immortal quite as much as unicellular organisms, for he survives in his successors.

In brief, Weismann assigns immortality not to the unicellular individual, but to the sum of all the individuals arising from it, all the individuals of the same species, living contemporaneously and successively—in fact, to the conception of a species.

In my view, what Weismann has tried to express by the word 'immortality' is no more than the continuity of the process of development. So he himself says in the course of a defence in which, however, he did not intend to give up the standpoint he had taken; he wishes to imply, by the immortality of unicellular organisms, only 'the deathless transformation of organic material,' or 'a transformation of organic material that always comes back to its original form again.'

Thus, Weismann himself really has implied that his distinction between immortal unicellar organisms, immortal germplasm, and mortal somatic cells, is a misconception. For the continuity of the process of development, or the mode of transformation of organic material, depends upon the continual formation and eventual destruction of newly-formed material, but in no way implies the continuous existence of the organised material in a state of organisation. From this point of view, the immortality of unicellular organisms and of the germplasm breaks down, and, above all, the artificial distinction between somatic cells and reproductive cells. For, in the latter, the organic process of development, with its transformation of organic material, also occurs.

Here I may give the conclusion of this division of my argument. Cells multiply only by doubling division. Between somatic cells and reproductive cells there is no strong contrast, no gulf that cannot be bridged. The continuity of the process of development depends upon the power of the cells to grow and to divide, and has already been set forth in the sayings—Omnis cellula e cellula, omnis nucleus e nucleo. Whatever novelty the doctrine of the continuity of the germplasm brings into this saying depends upon error, and is in contradiction to known natural facts.

II. ARGUMENTS AGAINST THE DOCTRINE OF DETERMINANTS.

Weismann has united his doctrine of determinants with his assumption of a differentiating division. He conceives that every little group of cells in the adult body possessed of definite character and of definite position in the body—in fact, every group of cells that is independently variable—is represented in the egg and in the spermatozoon by a number of little particles—the biophores—and that these, joined in a system, form the determinants. The innumerable determinants, he thinks, are, so arranged in the germplasm, and are endowed with such powers, that, during the process of development, they reach, at the right time, the right place for their expansion into cells. For instance, in the case of a mammal with parti-coloured fur, as many architecturally arranged determinants would be present as there were different spots and stripes in the fur, due to colour and length of the hairs.

This chain of ideas, made sharp and definite by Weismann, has recurred again and again in theoretical biological literature in a vague way. In my view, it rests upon a false use of the conception of causality, and upon a false implication given to the relation between the rudiment and the product of the rudiment, each mistake involving the other.

Because, if its development be not interfered with, a definite egg necessarily gives rise to a definite kind of animal, a complete identity between the rudiment and the product, between cause and consequence, has been assumed more or less consciously. The conception of the sequence has been as if an organism caused its own development in a closed system of forces, in a kind of organic perpetual motion. It has been overlooked that, in the course of the development, many other conditions must be fulfilled, as without them the product never would come from the rudiment.

That the same adults may come from the eggs depends upon the egg-cells, in the ordinary course of events, being in similar conditions of anabolism and katabolism, being affected by gravity, light, temperature, and so forth, in the same way. Thus, when we are attempting to grasp the fundamental nature of the course of organic development, we must not omit the part played by these factors.

We may dwell for a moment upon this weighty point, as its significance is commonly misunderstood.

The course of each organic development depends in the first place, upon the absorption and metamorphosis of matter. Inorganic matter perpetually is being turned into organic material to serve for the growth and development of the rudiments. Thus, what in one stage of the development is mere inorganic material, and an external condition of the development of the rudiment, in the next stage is become a part of the rudiment. The food-yolk of an egg, for instance, like the oxygen of the atmosphere, appears, in its relation to the material of the rudiments, to be something supplied from outside, an external condition of the development; yet it is continually passing into the rudiments and altering them, even though the alteration may be purely quantitative. From this follows the very simple inference that during the course of an organic development external matter is always being changed into internal matter, or that the rudiments are continually growing and changing at the expense of the surroundings.

Now, let one reflect that the egg and the adult are two terminal states of organised material, and that they are separated from each other by an almost inconceivably long series of connecting, intermediate states; consider that each stage of the development is the rudiment and the producer of the succeeding stage, of the stage that follows, as the consequence of it; consider that what was external in each antecedent stage has entered the rudiment and become part of it in the succeeding stage. Then it will be understood that it is a logical error to assume that all the characters present in the last link of the chain of development have their determining causes in the first link of the chain. The mistake lies in this: in the failure to distinguish between the causes contained in the egg at the beginning of the development, and the causes entering it during the course of development from the accession of external material in the various stages. As there can be no absolute identity between rudiment and product, it is erroneous to transmute the visible complexity of the final stage of the development into an invisible complexity of the first stage, as the old evolutionists did, and as the new evolutionists are attempting to do.

But there is another error in the doctrine of determinants. This is in intimate union with the error just discussed, and, to put it shortly, consists in attributing to a cell—and the egg and spermatozoon are cells—the possession of characters not peculiar to cells, but resulting from the co-operation of many cells.

The characters of an adult active organism, like a plant or an animal, are exceedingly numerous, most varied in their nature, and essentially different. Some characters depend upon the healthy co-operation of nearly all the parts of the body, or of a group of organs; others are peculiar to an organ, and may be referred to its shape, structure, position, function, and so forth. Others, again, depend upon individual cells, or even upon separate parts of cells. Is it really possible that all these characters, so many and so heterogeneous, have special, material bearers in the germ, and that these bearers are either simple biophores or determinants—that is to say, groups of biophores?

I can conceive a cell as endowed only with the material bearers of such characters as really belong to a cell itself. Thus, a reproductive cell might have material particles as the rudiments for producing horn, chitin, chondrin, ossein, pigment, or chlorophyll, or for nerve-fibrils, muscle-fibrils; but not for producing a hair, or a separate ganglion of the spinal cord or the biceps muscle. The rudiments for hairs, nerve-ganglia, muscles, and so forth, must be groups of cells, for only groups of cells, and not specially arranged groups of particles within a cell, are able to grow into hairs, spinal ganglia, or muscles.

In a short statement, made in 1892, I said: 'The mistake into which speculations upon the nature of organic development has led so many investigators is this: they reflect the characters of the adult upon the undivided egg, and so people that sphere of yolk with a system of tiny particles, corresponding to the parts of the adult, qualitatively and in spacial relations. But in this method of thinking, it is left out of count that the egg is an organism which multiplies by division into numerous organisms like itself, and that, in each stage of the development, it is only by the mutual action of all these numerous elementary organisms that the development of the whole organism slowly proceeds.'

Weismann himself, in a discussion of the pangenes of De Vries, has partly shown that one cannot assume the existence in the cell of material particles that are the bearers of qualities foreign to the nature of a cell and transcending it. In reference to the attempt to explain zebra-striping by pangenes, he says (Germplasm, English edition, p. 16): 'There can be no "zebra-pangenes," because the striping of a zebra is not a cell character. There may perhaps be black and white pangenes, whose presence causes the black or white colour of a cell; but the striping of a zebra does not depend on the development of these colours within a cell, but is due to the regular alternation of thousands of black and white cells arranged in stripes.' Again (p. 17), he says: 'The serrated margin of a leaf, for instance, cannot depend on the presence of "serration-pangenes," but is due to the peculiar arrangement of the cells. The same argument would apply to almost all the obvious "characters" of the species, genus, family, and so on. For instance, the size, structure, veining, and shape of leaves, the characteristic and often absolutely constant patches of colour on the petals of flowers, such as orchids, may be referred to similar causes. These qualities can only arise by the regular co-operation of many cells.'

Notwithstanding so correct a declaration, Weismann himself, in his doctrine of determinants, has fallen into the error he himself has exposed. To represent characters of the adult due to groups of cells and organisms, he imagines in the egg-cell, not simple particles like pangenes, but architecturally arranged groups of particles, determinants.

No real change has been made. Conditions are reflected upon the cell that in their real nature surpass its possibilities. With right and reason one may adduce, against his own determinants, what Weismann has said about pangenes, for exactly the same reasons: 'There cannot be zebra-determinants or serration-determinants, because zebra-striping, like the serrated edge of a leaf, is no cell character.'

The error in Weismann's doctrine of determinants may be made clearer by an analogy.

The human state may be conceived as a high and compound organism that, by the union of many individuals, and by their division into classes with different functions, has developed into a form always becoming more complicated. To carry out our comparison better, let us assume that all the individuals united in the human state arose from a single pair. The single pair would be the rudiment of the whole state, and would bear the same significance in the development of the state, as the fertilised egg bears to the development of the adult. The characters of the state, its different organisations for protection, for tilling the soil, for trade, for government, and for education, must be explained causally from the characters of the first pair, which we take as the human rudiment, and from the outer conditions under which that pair and the generations that arose from it had to live.

As the state develops, urban and district communities, unions for husbandry and manufactures, colleges of physicians, parliaments, ministries, armies, and so forth, appear. All this visible complexity depends upon individuals associated for definite purposes and specialised in different directions. It would certainly not occur to anyone to explain the growth of this complexity in the developing state by the assumption that this secondary complexity was preformed as definite material particles present in the first pair, although the first pair is the rudiment of the whole. Much comment is unnecessary; everyone must feel that this attempt to explain the causal relations is on the wrong track, that it is perverse to try to explain the complex characters of the human state by a system of architecturally arranged particles stored within the first pair. The organisations arising from the co-operation of many men are something new, and cannot be regarded as present in the organizations of one man. No doubt they depend, in the last resort, upon human nature, but by no means in this crude, mechanical fashion.

But what applies to the causal relations between the state-organism and men applies also, ceteris paribus, to the explanation of the causal relations between the rudiments in the egg and the organism to which the egg gives rise. For these an explanation cannot be expected on the lines of Weismann's doctrine of determinants, as that implies a fundamentally erroneous assumption. It refers organizations that depend upon cell-communities to organizations of material particles within a cell.

'To understand inheritance,' says Naegeli, with truth, 'we require not an independent, special symbol for every difference resulting from time, space, and quality, but a substance that, by the linking of the limited number of elements in it, can exhibit every possible combination of differences, and that by permutation can pass into another combination of differences.'

This standpoint is clearer when interpreted in terms of cells. The hereditary masses contained in the egg and spermatozoon can be composed only of such particles as are the bearers of cell-characters. Every compound organism can inherit characters only in the form of cell-characters. The innumerable, and endlessly variable, characters of plants and animals are of composite nature. They find their expression in differences of shape, structure, and function in the organs and tissues, and in the special methods in which these are interrelated. They depend upon the co-operation of many cells, and, for this reason, cannot be carried into the hereditary mass of any cell by material bearers. They are secondary formations, that can arise only after the multiplication of cells, and from the varied combination of cell-characters that accompanies the multiplication of cells.

In the foregoing pages I have attempted to prove the untenability of the doctrine of determinants from general considerations. I shall now attempt the same by analysis of a concrete case. The frog's egg may serve for this. It is a familiar object, frequently studied. Consider its mode of division, and the formation of the blastula, gastrula, and germinal layers.

In cleavage the nucleus plays the chief part, and thus has been accepted as the bearer of the hereditary mass. But no single, special determinant gives the impulse for cleavage; rather, the co-operation of all the particles that are essential to the nature of the nucleus. The chromosomes, which we may regard as independently growing and dividing units, must have doubled by assimilation of food material from the yolk; perhaps, also, the centrosome may have doubled in the same way before the nucleus is in a condition to divide. This condition itself appears the necessary result of many different processes of nutrition and growth, as the result of complicated chemical processes that run their course within the separate, elementary, vital units of the nucleus.

The multiplication of the nucleus into two, four, and eight daughter-nuclei, and so forth, gives the impulse for the breaking up of the yolk into a corresponding number of cells. In that process the direction of the cleavage-planes, the relative positions and the different sizes of the cells exhibit, under normal conditions, the most marked regularity. But it may be shown directly that this regularity is not the result of special determinants lying within the nucleus. For all these phenomena, which are characteristic in the cleavage of the frog's egg, as well as in the cleavage of all other eggs, are determined directly by the qualities of the yolk surrounding the nucleus.

In several publications I have shown clearly that the external form of an egg and the arrangement of its contents, according to the different specific gravities of the component particles, determine the position of the nucleus and of the successive planes of division. Similarly, the different sizes of the cells first formed and the unequal rate of division shown at the two poles of the egg depend upon the constitution of the yolk, upon the cleavage of the yolk into a portion richer in protoplasm and a portion poorer in protoplasm, and upon the differences in the bulk of protoplasm that in this way reaches each of the first-formed cells.

In many cases it has been shown that there is a constant relation between the first three cleavage-planes of the egg and the long axis of the animal that arises from the egg. Weismann and Roux make this a proof that, in nuclear division, the nuclei that arise have different qualities; that the protoplasmic masses lying to the right and left of the median plane are set apart to build up the right and left halves of the embryo; that, similarly, the first transverse and horizontal cleavage-planes divide the protoplasm of the egg into pieces predetermined for the formation of the anterior and posterior, dorsal and ventral, parts of the embryo.

But I think I have shown beyond possibility of doubt that these events are due not to the existence of special, mysteriously working groups of determinants within the nucleus, but merely to the specific shape of the whole egg and to the segregation of the yolk. It is self-evident that, as the body of the embryo builds itself up from the actual material of the egg, the way in which the material of the egg is disposed must be of great influence upon the formation of the shape of the embryo. And so, in a recently published work, I stated that the growing embryo, especially in its early stages, must conform in many ways to the shape of the fertilised egg.

Thus, to bear out what I have been saying by actual examples, the distribution of the actual particles of the fertilised egg must correspond to the disposition of the bulk of material in the blastosphere; for, in the breaking up into cells, the spacial arrangement of the substances of different weights undergoes no change. Thus, amphibia, the eggs of which have the poles different in character, produce blastospheres the poles of which are unlike; while eggs, like those of the fowl, where the yolk does not divide, give rise to blastospheres with unsegmented yolk. In such cases the more or less complete segregation of the yolk and gravity, which causes a separation of the contents of the egg according to the weights of the particles, are agencies determining the particular kind of development. It is no case of special groups of determinants within the nucleus.

Thus, an oval and an elongate egg produce respectively an oval and an elongate blastosphere. The blastosphere determines the orientation of the gastrula, and so forth. In fact, the original distribution of mass in the material of the egg is carried directly on to the following stages of development (oval eggs of triton, insects, etc.).

So, finally, in many eggs, where, in addition to a polar differentiation, there is also a bilateral symmetry in the distribution of substances of different specific gravities and of different physiological value, the resulting blastospheres, from the reasons given above, assume a bilaterally symmetrical form.

Although, then, in eggs with polar differentiation, which have either one axis longer or are bilaterally symmetrical, under normal conditions the planes of the first two segmentations may correspond to the principal axes of the future embryo, the cause for this agreement lies in the structure of the egg, and is not to be looked for, as Roux and Weismann suppose, in differentiating processes of cleavage, undergone by the nuclei in their first divisions. It is in this way that there are to be explained the investigations made by Van Beneden and Jülin upon the eggs of ascidians, by Wilson upon the egg of Nereis, by Roux upon the egg of Rana esculenta, and by me on the egg of Triton.

As it fails with the process of cleavage, so Weismann's doctrine of determinants fails when we analyse the formation of the blastosphere, the gastrula, and the germinal layers.

The formation of the blastosphere seems to me to be due to the co-operation of the following processes:

(1) In the division of the egg-cell cavities arise between the four, eight, and sixteen pieces, and thus the whole contents of the egg become arranged more loosely. (2) The more the cells multiply by division and become smaller in circumference, the more closely they apply their lateral surfaces to each other, especially at the outer surface of the whole, so assuming the arrangement of cell-epithelia. (3) By the secretion of fluid, a constantly growing central cavity is formed pari passu with the approximation of the superficial cells, and this probably also brings with it an increase of the internal pressure, and a wider curvature of the wall of the sphere.

Now, is there any part of these processes that has to do with the breaking of the nuclear contents into groups of determinants with different qualities? By no means. The egg divides into many pieces, because such division is a general property of cells, and it is not associated with separate, special material bearers. The appearance of spaces between the cells, resulting from division, is due to forces some of which reside within the single cells, some of which come from without. In especial, the assumption of a spherical shape—an assumption occurring also to a greater or less degree when the results of division leave each other—is caused by the yolk actively arranging itself round the two nuclei as centres of attraction. The attempt to become spherical is opposed by other forces, in accordance with which the cells resulting from division press against each other. These forces that press the cells together seem to increase, as the size of the cells diminishes, so that the cells approximate their lateral faces continually more closely. The secretion of fluid into the interior of the sphere and the resulting increase of the outer surface results from the characters of the whole wall, and cannot be explained by single, specially determined cells.

Finally, to take the case of the special kinds of blastospheres (e.g., of amphioxus, amphibia, reptiles, birds, and so forth), it has been already shown that these are produced by the shape of the egg, by the bulk of the yolk, and by the segregation of the yolk-particles under the influence of gravity; that, in fact, the shapes are determined by the general gross conditions of the structure of the egg.

Plainly, the blastosphere cannot be pre-existing as a structure of particles in the fertilised nucleus; there cannot be blastosphere determinants. The conditions for the origin of the blastosphere come into existence only by the process of segmentation, and it is only by its capacity to divide that the egg contains the conditions for blastosphere formation. Here we have epigenesis—the appearance of a new formation, not the becoming visible of pre-existing complexity.

The conditions of gastrulation and of the formation of the germinal layers are similar. The invagination of the blastosphere comes about by the co-operation of all the cells of its wall, by local differences in the rates of growth in that wall, from dissimilarities in its curvature, from many causes which have not yet been sufficiently sought out and investigated. As cell division itself depends not upon special particles, but upon changes in the entire nuclear contents, it follows that the growth of the blastosphere-wall, which is merely the sum of the growth of all the cells in it, cannot be determined by special groups of determinants.

As an attempt to explain gastrulation, the origin of the germinal layers and many other events of development, the doctrine of determinants has reversed cause and effect. Certain cells do not become invaginated into the segmentation cavity because they possess groups of determinants that impel them to the assumption of inner layer characters. The reverse is the truth. Local conditions of growth cause the invagination of a set of the cells of the blastosphere-wall. This invaginated layer of cells, brought into a new position with regard to its environment, becomes the endoderm and receives the stimulus to assume the character appropriate to the new environment. It is unlogical to speak of endoderm in the fashion of many textbooks and treatises on embryology, while the so-called endoderm cells still form part of the outer surface of the blastosphere, or even while they are still in process of formation by cleavage. For 'inner germinal layer' implies a condition of position which is created by the invagination.

In fact, it is impossible, in thinking of the gastrula as in thinking of the blastosphere, to conceive that in the egg, which is a simple cell, there can be preformed by material particles in the nucleus a condition which implies the existence of two layers of cells.

Thus analysis of a special case leads to the same conclusion as is reached by the general reasoning of the earlier part of this section.

FOOTNOTES:

[7] The Germplasm, pp. 68, 69.

[8] The following treatises contain criticisms of Weismann's theories: W. Haacke, Gestaltung und Vererbung; Leipzig, 1893; Herbert Spencer, articles in Contemporary Review (1893-94); Romanes, An Examination of Weismannism; Longmans, 1893.

[9] Notwithstanding the objections raised by Bergh, Verworn, and Haacke, I abide by the supposition that the nucleus of reproductive cells contains the hereditary mass or germinal material. My reasons may be found in my text-book on The Cell (English edit., p. 274). Briefly they are: 1. The equivalence of the male and female hereditary masses. 2. The equal distribution of the growing nuclear mass of the primary egg-cell among the daughter-cells that, arising from it, build up the organism. 3. The preservation of a constancy of bulk of the hereditary mass when fertilization occurs. 4. The isotropism of protoplasm. Following Pflüger, I mean by isotropism that the protoplasm of the egg does not contain local areas for the formation of different organs; but that, according to the conditions, any part of the protoplasm may be employed in the formation of any organ. Isotropism is merely the negation of His' doctrine of the presence of local areas for definite organs, and without losing its meaning, is compatible with the fact that many eggs have their poles different, and that others have a bilateral symmetry which determines the plane of the first division. 5. The fact that the first stages of many embryonic developments consist in the multiplication of the nuclear material and its distribution in the yolk, following which the yolk-mass cleaves into cells.

[10] English edition, p. 32.

[11] English edition, p. 34.

[12] In this section upon heteromorphosis I rely upon the following treatises, which have appeared recently. Loeb, Untersuchungen zur physiologischen Morphologie der Thiere. Organbildung und Wachsthum. Heft, 1 and 2 (1891-1892). H. de Vries, Intracellulare Pangenesis (1889). H. Driesch, Entwicklungsmechamische Studien, i.-vi.; Zeitschrift f. wissenschaft, Zool., vol. liii.-lv. The same, Zur Theorie der thierischen Formbildung. Biol. Centralblatt, vol. xiii., 1893. Chabry, Contribution à l'embryologie normale et tératologique des Ascidies simples. Jour. de l'Anat. et de Physiol. (1887). Wilson, Amphioxus and the Mosaic Theory. Journal of Morph. (1893). See also Anatomischer Anzeiger (1892).

[13] Roux tried to give experimental evidence in favour of his mosaic theory in a treatise On the Artificial Productions of Half-Embryos by the Destruction of one of the first two Cleavage-Cells, and on the Reconstruction of the Lost Parts. Virchow's Archiv., vol. cxiv., 1888. Roux defends his mosaic theory against Driesch and myself in (1) Ueber das entwicklungsmechanische Vermögen jeder der beiden ersten Furchungszellen des Eies. Verhandl. der Anat. Gesellsch. der 6^ten Versamml. in Wien, 1892. (2) Ueber Mosaikarbeit und neuere Entwicklungshypothesen. Anatomische Hefte von Merkel und Bonnet (1893). Also in Biol. Centralblatt (1893); in the Anatom. Anzeiger (1893), and in the treatise Die Methoden zur Erzeugung halber Froschembryonen und zum Nachweis der Beziehung der ersten Furchungsebenen des Froscheies zur Medianebene des Embryo. Anatom. Anzeiger. (1894); Nos. 8 and 9.

If, as would appear from the last treatise, Roux would avoid being reckoned with evolutionists, he must abandon his mosaic theory, and this he has not done. I think in the present essay, on theoretical and experimental grounds I have shown the untenability of Roux's mosaic theory.

[14] The terms vertical and horizontal refer to the vertical axis of the egg, which passes through the animal and vegetative poles.—Translator's note.

[15] Further details concerning these experiments may be found in Hertwig, Ueber den Werth der ersten Furchungszellen für die Organbildung des Embryo. Experimentelle Studien am Froschund Tritonei. Archiv. für Mikrosk. Anatomie, vol. xlii., 1893, p. 710; Plate xli.; Figs. 1, 2, 27.

[16] For the facts in this section I rely in particular upon the writings of Vöchting, Bert, Ollier, Trembley, Landois, Ponfick, and others:

H. Vöchting: Ueber Transplantation auf Pflanzenkörper. Untersuchungen zur Physiologie und Pathologie; Tübingen, 1892.

Von Gärtner: Versuche und Beobachtungen ueber die Bastarderzeugung im Pflanzenreich, 1849.

Léopold Ollier: Recherches expérimentales sur la production artificielle des os au moyen de la transplantation du périoste, etc. Journal de la physiologie de l'homme et des animaux, tom. ii., 1859, pp. 1, 169, 468.

Léopold Ollier: Recherches expérimentales sur les greffes osseuses. The same, tom. iii., p. 88, 1860.

Paul Bert: Recherches expérimentales pour servir à l'histoire de la vitalité propre des tissus animaux. Annales des Sciences naturelles, Ser. V., Zoologie, tom. v., 1886.

Von Recklinghausen: Die Wiedererzeugung (Regeneration) und die Ueberpflanzung (Transplantation). Handbuch d. Allgem. Pathologie des Kreislaufs aus Deutsche Chirurgie, 1883.

Trembley: Mémoires pour servir à l'histoire d'un genre de Polypes d'eau douce, 1744.

Landois: Die Transfusion des Blutes; Leipzig, 1875.

Adolf Schmitt: Ueber Osteoplastik in klinischer und experimenteller Beziehung. Arbeiten aus der chirurgischenklinik der Königl. Universität, Berlin.

Ponfick: Experimentelle Beaträge zur Lehre von der Transfusion. Virchow's Archiv., vol. lxii.

Beresowsey: Ueber die histologischen Vorgänge bei der Transplantation von Hautstücken auf Thiere einer anderen Species. Ziegler's Beiträge zur pathologischen Anatomie und zur allgemeinen Pathologie; Jena, 1893.

---