It is an error to suppose that we gain any further insight into the formation of germ-cells by referring to these cells of so-called ‘embryonic character,’ which are contained in the body of the mature organism. It is of course well known that many cells are characterized by very sharply defined histological differentiation, while others are but slightly differentiated; but it is as difficult to imagine that germ-cells can arise from the latter as from the former. Both classes of cells contain idioplasm with a structure different from that which is contained in the germ-cell, and we have no right to assume that any of them can form germ-cells until it is proved that somatic idioplasm is capable of undergoing re-transformation into germ-idioplasm.

The same argument applies to the cells of the embryo itself, and it therefore follows that those instances of early separation of sexual from somatic cells, upon which I have often insisted as indicating the continuity of the germ-plasm, do not now appear to be of such conclusive importance as at the time when we were not sure about the localization of the idioplasm in the nuclei. In the great majority of cases the germ-cells are not separated at the beginning of embryonic development, but only in some one of the later stages. A single exception is found in the pole-cells (‘Polzellen’) of Diptera, as was shown many years ago by Robin^[126] and myself^[127]. These are the first cells formed in the egg, and according to the later observations of Metschnikoff^[128] and Balbiani^[129], they become the sexual glands of the embryo. Here therefore the germ-plasm maintains a true unbroken continuity. The nucleus of the egg-cell directly gives rise to the nuclei of the pole-cells, and there is every reason to believe that the latter receive unchanged a portion of the idioplasm of the former, and with it the tendencies of heredity. But in all other cases the germ-cells arise by division from some of the later embryonic cells, and as these belong to a more advanced ontogenetic stage in the development of the idioplasm, we can only conclude that continuity is maintained, by assuming (as I do) that a small part of the germ-plasm persists unchanged during the division of the segmentation nucleus and remains mixed with the idioplasm of a certain series of cells, and that the formation of true germ-cells is brought about at a certain point in the series by the appearance of cells in which the germ-plasm becomes predominant. But if we accept this hypothesis it does not make any difference, theoretically, whether the germ-plasm becomes predominant in the third, tenth, hundredth, or millionth generation of cells. It therefore follows that cases of early separation of the germ-cells afford no proof of a direct persistence of the parent germ-cells in those of the offspring; for a cell the offspring of which become partly somatic and partly germ-cells cannot itself have the characters of a germ-cell; but it may nevertheless contain germ-idioplasm, and may thus transfer the substance which forms the basis of heredity from the germ of the parent to that of the offspring.

If we are unwilling to accept this hypothesis, nothing remains but to credit the idioplasm of each successive ontogenetic stage with a capability of re-transformation into the first stage. Strasburger accepts this view; and he believes that the idioplasm of the nuclei changes during the course of ontogeny, but returns to the condition of the first stage of the germ, at its close. But the rule of probability is against such a suggestion. Suppose, for instance, that the idioplasm of the germ-cell is characterized by ten different qualities, each of which may be arranged relatively to the others in two different ways, then the probability in favour of any given combination would be represented by the fraction (1/2)¹⁰ = 1/1024: that is to say, the re-transformation of somatic idioplasm into germ-plasm will occur once in 1024 times, and it is therefore impossible for such re-transformation to become the rule. It is also obvious that the complex structure of the germ-plasm which potentially contains, with the likeness of a faithful portrait, the whole individuality of the parent, cannot be represented by only ten characters, but that there must be an immensely greater number; it is also obvious that the possibilities of the arrangement of single characters must be assumed to be much larger than two; so that we get the formula (1/p), where p represents the possibilities, and n the characters. Thus if n and p are but slightly larger than we assumed above, the probabilities become so slight as to altogether exclude the hypothesis of a re-transformation of somatic idioplasm into germ-plasm.

It may be objected that such re-transformation is much more probable in the case of those germ-cells which separate early from the somatic cells. Nothing can in fact be urged against the possibility that the idioplasm of (e. g.) the third generation of cells may pass back into the condition of the idioplasm of the germ-cell; although of course the mere possibility does not prove the fact. But there are not many cases in which the sexual cells are separated so early as the third generation: and it is very rare for them to separate at any time during the true segmentation of the egg. In Daphnidae (Moina) separation occurs in the fifth stage of segmentation^[130], and although this is unusually early it does not happen until the idioplasm has changed its molecular structure six times. In Sagitta^[131] the separation does not take place until the archenteron is being formed, and this is after several hundred embryonic cells have been produced, and thus after the germ-plasm has changed its molecular structure ten or more times. But in most cases, separation takes place at a much later stage; thus in Hydroids it does not happen until after hundreds or thousands of cell-generations have been passed through; and the same fact holds in the higher plants, where the production of germ-cells frequently occurs at the end of ontogeny. In such cases the probability of a re-transformation of somatic idioplasm into germ-plasm becomes infinitely small.

It is true that these considerations only refer to a rapid and sudden re-transformation of the idioplasm. If it could be proved that development is not merely in appearance but in reality a cyclical process, then nothing could be urged against the occurrence of re-transformation. It has been recently maintained by Minot^[132] that all development is cyclical, but this is obviously incorrect, for Nägeli has already shown that direct non-cyclical courses of development exist, or at all events courses in which the earliest condition is not repeated at the close of development. The phyletic development of the whole organic world clearly illustrates a development of the latter kind; for although we may assume that organic development is not nearly concluded, it is nevertheless safe to predict that it will never revert to its original starting-point, by backward development over the same course as that which it has already traversed. No one can believe that existing Phanerogams will ever, in the future history of the world, retrace all the stages of phyletic development in precise inverse order, and thus return to the form of unicellular Algae or Monera; or that existing placental mammals will develope into Marsupialia, Monotremata, mammal-like reptiles, and the lower vertebrate forms, into worms and finally into Monera. But how can a course of development, which seems to be impossible in phylogeny, occur as the regular method of ontogeny? And quite apart from the question of possibility, we have to ask for proofs of the actual occurrence of cyclical development. Such a proof would be afforded if it could be shown that the nucleoplasm of those somatic cells which (e.g. in Hydroids) are transformed into germ-cells passes backwards through many stages of development into the nucleoplasm of the germ-cell. It is true that we can only recognise differences in the structure of the idioplasm by its effects upon the cell-body, but no effects are produced which indicate that such backward development takes place. Since the course of onward development is compelled to pass through the numerous stages which are implied in segmentation and the subsequent building-up of the embryo, etc., it is quite impossible to assume that backward development would take place suddenly. It would be at least necessary to suppose that the cells of embryonic character, which are said to be transformed into primitive germ-cells, must pass back through at any rate the main phases of their ontogeny. A sudden transformation of the nucleoplasm of a somatic cell into that of a germ-cell would be almost as incredible as the transformation of a mammal into an amoeba; and yet we are compelled to admit that the transformation must be sudden, for no trace of such retrogressive stages of development can be seen. If the appearance of the whole cell gives us any knowledge as to the structure of its nuclear idioplasm, we may be sure that the development of a primitive germ-cell proceeds without a break, from the moment of its first recognizable formation, to the ultimate production of distinct male or female sexual cells.

I am well aware that Strasburger has stated that, in the ultimate maturation of the sexual cells, the substance of the nuclei returns to a condition similar to that which existed at the beginning of ontogenetic development; still such a statement is no proof, but only an assumption made to support a theory. I am also aware that Nussbaum and others believe that, in the formation of spermatozoa in higher animals, a backward development sets in at a certain stage; but even if this interpretation be correct, such backward development would only lead as far as the primitive germ-cell, and would afford no explanation of the further transformation of the idioplasm of this cell into germ-plasm. But this latter transformation is just the point which most needs proof upon any theory except the one which assumes that the primitive germ-cell still contains unchanged germ-plasm. Every attempt to render probable such a re-transformation of somatic nucleoplasm into germ-plasm breaks down before the facts known of the Hydroids, in which only certain cells in the body, out of the numerous so-called embryonic cells, are capable of becoming primitive germ-cells, while the rest do not possess this power.

I must therefore consider as erroneous the hypothesis which assumes that the somatic nucleoplasm may be transformed into germ-plasm. Such a view may be called ‘the hypothesis of the cyclical development of the germ-plasm.’

Nägeli has tried to support such an hypothesis on phyletic grounds. He believes that phyletic development follows from an extremely slow but steady change in the idioplasm, in the direction of greater complexity, and that such changes only become visible periodically. He believes that the passage from one phyletic stage to another is chiefly due to the fact that ‘in any ontogeny, the very last structural change upon which the separation of germs depends, takes place in a higher stage, one or more cell-generations later’ than it occurred in a lower stage. ‘The last structural change itself remains the same, while the series of structural changes immediately preceding it is increased.’ I believe that Nägeli, being a botanist, has been too greatly influenced by the phenomena of plant-life. It is certainly true that in plants, and especially in the higher forms, the germ-cells only make their appearance, as it were, at the end of ontogeny; but facts such as these do not hold in the animal kingdom: at any rate they are not true in the great majority of cases. In animals, as I have already mentioned several times, the germ-cells are separated from the somatic cells during embryonic development, sometimes even at its very commencement; and it is obvious that this latter is the original, phyletically oldest, mode of formation. The facts at our disposal indicate that the germ-cells only appear, for the first time, after embryological development, in those cases where the formation of asexually produced colonies takes place, either with or without alternation of generations; or in cases where alternation of generations occurs without the formation of such colonies. In a colony of polypes, the germ-cells are produced by the later generations, and not by the founder of the colony which was developed from an egg. This is also true of the colonies of Siphonophora, and the germ-cells appear to arise very late in certain instances of protracted metamorphosis (Echinodermata), but on the other hand, they arise during the embryonic development of other forms (Insecta) which also undergo metamorphosis. It is obvious that the phyletic development of colonies or stocks must have succeeded that of single individuals, and that the formation of germ-cells in the latter must therefore represent the original method. Thus the germ-cells originally arose at the beginning of ontogeny and not at its close, when the somatic cells are formed.

This statement is especially supported by the history of certain lower plants, or at any rate chlorophyll-containing organisms, and I think that these forms supply an admirable illustration of my theory as to the phyletic origin of germ-cells, as explained in my earlier papers upon the same subject.

The phyletic origin of germ-cells obviously coincides with the differentiation of the first multicellular organisms by division of labour^[133]. If we desire to investigate the relation between germ-cells and somatic cells, we must not only consider the highly developed and strongly differentiated multicellular organisms, but we must also turn our attention to those simpler forms in which phyletic transitions are represented. In addition to solitary unicellular organisms, we know of others living in colonies of which the constituent units or cells (each of them equivalent to a unicellular organism) are morphologically and physiologically identical. Each unit feeds, moves, and under certain circumstances is capable of reproducing itself, and of thus forming a new colony by repeated division. The genus Pandorina (Fig. I), belonging to the natural order Volvocineae, represents such ‘homoplastid’ (Götte) organisms. It forms a spherical colony composed of ciliated cells, all of which are exactly alike: they are embedded in a colourless gelatinous mass. Each cell contains chlorophyll, and possesses a red eye-spot, and a pulsating vacuole. These colonies are propagated by the sexual and asexual (Fig. II) methods alternately, although in the former case the conjugating swarm-cells cannot be distinguished with certainty as male or female. In both kinds of reproduction, each cell in the colony acts as a reproductive cell; in fact, it behaves exactly like a unicellular organism.

It is very interesting to find in another genus belonging to the same natural order, that the transition from the homoplastid to the heteroplastid condition, and the separation into somatic and reproductive cells, have taken place. In Volvox (Fig. III) the spherical colony consists of two kinds of cells, viz. of very numerous small ciliated cells, and of a much smaller number of large germ-cells without cilia. The latter alone possess the power of producing a new colony, and this takes place by the asexual and sexual methods alternately: in the latter a typical fertilization of large egg-cells by small spermatozoa occurs. The sexual differentiation of the germ-cells is not material to the question we are now considering; the important point is to ascertain whether here, at the very origin of heteroplastid organisms, the germ-cells, sexually differentiated or not, arise from the somatic cells at the end of ontogeny, or whether the substance of the parent germ-cell, during embryonic development, is from the first separated into somatic and germ-cells. The former interpretation would support Nägeli’s view, the latter would support my own. But Kirchner^[134] distinctly states that the germ-cells of Volvox are differentiated during embryonic development, that is, before the escape of the young heteroplastid organism from the egg-capsule. We cannot therefore imagine that the phyletic development of the first heteroplastid organism took place in a manner different from that which I have previously advocated on theoretical grounds, before this striking instance occurred to me. The germ-plasm (nucleoplasm) of some homoplastid organism (similar to Pandorina) must have become modified in molecular structure during the course of phylogeny, so that the colony of cells produced by its division was no longer made up of identical units, but of two different kinds. After this separation, the germ-cells alone retained the power of reproduction possessed by all the parent cells, while the rest only retained the power of producing similar cells by division. Thus Volvox seems to afford distinct evidence that in the phyletic origin of the heteroplastid groups, somatic cells were not, as Nägeli supposes, intercalated between the mother germ-cell and the daughter germ-cells in each ontogeny, but that the somatic cells arose directly from the former, with which they were previously identical, as they are even now in the case of Pandorina. Thus the continuity of the germ-plasm is established at least for the beginning of the phyletic series of development.

The fact, already often mentioned, that in most higher organisms the separation of germ-cells takes place later, and often very late, at the end of the whole ontogeny, proves that the time at which this separation of the two kinds of cells took place, must have been gradually changed. In this respect the well-established instances of early separation are of great value, because they serve to connect the extreme cases. It is quite impossible to maintain that the germ-cells of Hydroids or of the higher plants, exist from the time of embryonic development, as indifferent cells, which cannot be distinguished from others, and which are only differentiated at a later period. Such a view is contradicted by the simplest mathematical consideration; for it is obvious that none of the relatively few cells of the embryo can be excluded from the enormous increase by division, which must take place in order to produce the large number of daughter individuals which form a colony of polypes. It is therefore clear that all the cells of the embryo must for a long time act as somatic cells, and none of them can be reserved as germ-cells and nothing else: this conclusion is moreover confirmed by direct observation. The sexual bud of a Coryne arises at a part of the Polype which does not in any way differ from surrounding areas, the body wall being uniformly made up of two single layers of cells, the one forming the ectoderm and the other the endoderm. Rapid growth then takes place at a single spot, and some of the young cells thus produced are transformed into germ-cells, which did not previously exist as separate cells.

Strictly speaking I have therefore fallen into an inaccuracy in maintaining (in former works) that the germ-cells are themselves immortal; they only contain the undying part of the organism—the germ-plasm; and although this substance is, as far as we know, invariably surrounded by a cell-body, it does not always control the latter, and thus confer upon it the character of a germ-cell. But this admission does not materially change our view of the whole subject. We may still contrast the germ-cells, as the undying part of the Metazoan body, with the perishable somatic cells. If the nature and the character of a cell is determined by the substance of the nucleus and not by the cell-body, then the immortality of the germ-cells is preserved, although only the nuclear substance passes uninterruptedly from one generation to another.

G. Jäger^[135] was the first to state that the body in the higher organisms is made up of two kinds of cells, viz., ontogenetic and phyletic cells, and that the latter, the reproductive cells, are not a product of the former (the body-cells), but that they arise directly from the parent germ-cell. He assumed that the formation of germ-cells takes place at the earliest stage of embryonic life, and he thus believed the connexion between the germ-plasm of the parent and of the offspring had received a satisfactory explanation. As I have previously mentioned in the introduction, Nussbaum also brought forward this hypothesis at a later period, and also based it upon a continuity of the germ-cells. He assumed that the fertilized egg is divided into the cells of the individual and into the cells which effect the preservation of the species, and he supported this view by referring to the few known cases of early separation of the sexual cells. He even maintained this hypothesis when I had proved in my investigations on Hydromedusae that the sexual cells are not always separated from the somatic cells during embryonic development, but often at a far later period. Not only is the hypothesis of a direct connexion between the germ-cells of the offspring and parent broken down by the facts known in the Hydroids, and in the Phanerogams^[136] which resemble them in this respect, but even the instances of early separated germ-cells quoted by Jäger and Nussbaum do not as a matter of fact support their hypothesis. Among existing organisms it is extremely rare for the germ-cells to arise directly from the parent egg-cell (as in Diptera). If, however, the germ-cells are separated only a few cell-generations later, the postulated continuity breaks down; for an embryonic cell, of which the offspring are partly germ-cells and partly somatic cells, cannot itself possess the nature of a germ-cell, and its idioplasm cannot be identical with that of the parent germ-cell. In order to prove this, it is only necessary to refer to the arguments as to the ontogenetic stages of the idioplasm. In the above-mentioned instances, the continuity from the germ-substance of the parent to that of the offspring can only be explained by the supposition that the somatic nucleoplasm still contains some unchanged germ-plasm. I believe that the fundamental idea of Jäger and Nussbaum is quite correct: it is the same idea which has led me to the hypothesis of the continuity of the germ-plasm, viz., the conviction that heredity can only be understood by means of such an hypothesis. But both these writers have worked out the idea in the form of an hypothesis which does not correspond with the facts. That this is the case is also shown by the following words of Nussbaum—‘the cell-material of the individual (somatic cells) can never produce a single sexual cell.’ Such production undoubtedly takes place, not only in Hydroids and Phanerogams, but in many other instances. The germ-cells cannot indeed be produced by any indifferent cell of embryonic character, but by certain cells, and under circumstances which allow us to positively conclude that they have been predestined for this purpose from the beginning. In other words, the cells in question contain germ-plasm, and this alone enables them to become germ-cells.

As a result of my investigations on Hydroids^[137], I concluded that the germ-plasm is present in a very finely divided and therefore invisible state in certain somatic cells, from the very beginning of embryonic development, and that it is then transmitted through innumerable cell-generations, to those remote individuals of the colony in which sexual products are formed. This conclusion is based upon the fact that germ-cells only occur in certain localized areas (‘Keimstätten’) in which neither germ-cells nor primitive germ-cells (the cells which are transformed into germ-cells at a later period) were previously present. The primitive germ-cells are also only formed in localized areas, arising from somatic cells of the ectoderm. The place at which germ-cells arise is the same in all individuals of the same species; but differs in different species. It can be shown that such differences correspond to different phyletic stages of a process of displacement, which tends to remove the localized area from its original position (the manubrium of the Medusa) in a centripetal direction. For the purposes of the present enquiry it is unnecessary to discuss the reasons for this change of position. The phyletic displacements of the localized areas are brought about during ontogeny by an actual migration of primitive germ-cells from the place where they arose to the position at which they undergo differentiation into germ-cells. But we cannot believe that primitive germ-cells would migrate if the germ-cells could be formed from any of the other young cells of indifferent character which are so numerous in Hydroids. Even when the localized area undergoes very slight displacement, e.g. when it is removed from the exterior to the interior of the mesogloea^[138], the change is always effected by active migration of primitive germ-cells through the substance of the mesogloea. Although the localized area has been largely displaced in the course of phylogeny, the changes in position have always taken place by very gradual stages, and never suddenly, and all these stages are repeated in the ontogeny of all existing species, by the migration of the primitive germ-cells from the ancestral area to the place where the germ-cells now arise. Hartlaub^[139] has recently added a further instance (that of Obelia) to the numerous minute descriptions of these phyletic displacements of the localized area, and ontogenetic migrations of the primitive germ-cells, which are given in my work already referred to. The instance of Obelia is of especial interest as the direction of displacement is here reversed, taking place centrifugally instead of in a centripetal direction.

But if displacements of the localized areas can only take place by the frequently roundabout method of the migration of primitive germ-cells, we are obliged to conclude that such is the only manner in which the change can be effected, and that other cells are unable to play the role of the primitive germ-cells. And if other cells are unable to take this part, it must be because nucleoplasm of a certain character has to be present in order to form germ-cells, or according to the terms of my theory, the presence of germ-plasm is indispensable for this purpose. I do not see how we can escape the conclusion that there is continuity of the germ-plasm; for if it were supposed that somatic idioplasm undergoes transformation into germ-plasm, such an assumption would not explain why the displacement occurs by small stages, and with extreme and constant care for the preservation of a connexion with cells of the ancestral area. This fact can only be explained by the hypothesis that cell-generations other than those which end in the production of the cells of the ancestral area, are totally incapable of transformation into germ-cells.

Strasburger has objected that the transmission of germ-plasm along certain lines, viz. through a certain succession of somatic cells, is impossible, because the idioplasm is situated in the nucleus and not in the cell-body, and because a nucleus can only divide into two exactly equal halves by the indirect method of division, which takes place, as we must believe, in these cases. ‘It might indeed be supposed,’ says Strasburger, ‘that during nuclear division certain molecular groups remain unchanged in the nuclear substance which is in other respects transformed, and that these groups are uniformly distributed through the whole organism; but we cannot imagine that their transmission could only be effected along certain lines.’

I do not think that Strasburger’s objections can be maintained. I base this opinion on my previous criticism upon the assumed equality of the two daughter-nuclei formed by indirect division. I do not see any reason why the two halves must always possess the same structure, although they may be of equal size and weight. I am surprised that Strasburger should admit the possibility that the germ-plasm, which, as I think, is mixed with the idioplasm of the somatic cells, may remain unchanged in its passage through the body; for if this writer be correct in maintaining that the changes of nuclear substance in ontogeny are effected by the nutritive influence of the cell-body (cytoplasm), it follows that the whole nuclear substance of a cell must be changed at every division, and that no unchanged part can remain. We can only imagine that one part of a nucleus may undergo change while the other part remains unchanged, if we hold that the necessary transformations of nuclear substance are effected, by purely internal causes, viz. that they follow from the constitution of the nucleoplasm. But that one part may remain unchanged, and that such persistence does, as a matter of fact, occur is shown by the cases above described, in which the germ-cells separate very early from the developing egg-cell. Thus in the egg of Diptera, the two nuclei which are first separated by division from the segmentation nucleus, form the sexual cells, and this proves that they receive the germ-plasm of the segmentation nucleus unchanged. But during or before the separation of these two nuclei, the remaining part of the segmentation nucleus must have become changed in nature, or else it would continue to form ‘pole-cells’ at a later period instead of forming somatic cells. Although in many cases the cell-bodies of such early embryonic cells fail to exhibit any visible differences, the idioplasm of their nuclei must undoubtedly differ, or else they could not develope in different directions. It seems to me not only possible, but in every way probable, that the bodies of such early embryonic cells are equal in reality as well as in appearance; for, although the idioplasm of the nucleus determines the character of the cell-body, and although every differentiation of the latter depends upon a certain structure of its nucleoplasm, it does not necessarily follow that the converse proposition is true, viz. that each change in the structure of the nucleoplasm must effect a change in the cell-body. Just as rain is impossible without clouds, but every cloud does not necessarily produce rain, so growth is impossible without chemical change, but chemical processes of every kind and degree need not produce growth. In the same manner every kind of change in the molecular structure of the nucleoplasm need not exercise a transforming influence on the cytoplasm, and we can easily imagine that a long series of changes in the nucleoplasm may appear only in the kind and energy of the nuclear divisions which take place, the cell-substance remaining unchanged, as far as its molecular and chemical structure is concerned. This suggestion is in accordance with the fact that during the first period of embryonic development in animals, the cell-bodies do not exhibit any visible differences, or only such as are very slight; although exceptional instances occur, especially among the lower animals. But even these latter (e.g. the difference in appearance of the cells of the ectoderm and endoderm in sponges and Coelenterata) perhaps depend more largely upon a different admixture of nutritive substances than upon any marked difference in the cytoplasm itself. It is obvious that, in the construction of the embryo, the amount of cell-material must be first of all increased, and that it is only at a later period that the material must be differentiated so as to possess various qualities, according to the principle of division of labour. Facts of this kind are also opposed to Strasburger’s view, that the cause of changes in the nucleoplasm does not lie within this substance itself but within the cell-body.

I believe I have shown that theoretically hardly any objections can be raised against the view that the nuclear substance of somatic cells may contain unchanged germ-plasm, or that this germ-plasm may be transmitted along certain lines. It is true that we might imagine a priori that all somatic nuclei contain a small amount of unchanged germ-plasm. In Hydroids such an assumption cannot be made, because only certain cells in a certain succession possess the power of developing into germ-cells; but it might well be imagined that in some organisms it would be a great advantage if every part possessed the power of growing up into the whole organism and of producing sexual cells under appropriate circumstances. Such cases might exist if it were possible for all somatic nuclei to contain a minute fraction of unchanged germ-plasm. For this reason, Strasburger’s other objection against my theory also fails to hold; viz. that certain plants can be propagated by pieces of rhizomes, roots, or even by means of leaves, and that plants produced in this manner may finally give rise to flowers, fruit and seeds, from which new plants arise. ‘It is easy to grow new plants from the leaves of Begonia which have been cut off and merely laid upon moist sand, and yet in the normal course of ontogeny the molecules of germ-plasm would not have been compelled to pass through the leaf; and they ought therefore to be absent from its tissue. Since it is possible to raise from the leaf a plant which produces flower and fruit, it is perfectly certain that special cells containing the germ substance cannot exist in the plant.’ But I think that this fact only proves, that in Begonia and similar plants, all the cells of the leaves or perhaps only certain cells contain a small amount of germ-plasm, and that consequently these plants are specially adapted for propagation by leaves. How is it then that all plants cannot be reproduced in this way? No one has ever grown a tree from the leaf of the lime or oak, or a flowering plant from the leaf of the tulip or convolvulus. It is insufficient to reply that, in the last-mentioned cases, the leaves are more strongly specialized, and have thus become unable to produce germ-substance; for the leaf-cells in these different plants have hardly undergone histological differentiation in different degrees. If, notwithstanding, the one can produce a flowering plant, while the others have not this power, it is of course clear that reasons other than the degree of histological differentiation must exist; and, according to my opinion, such a reason is to be found in the admixture of a minute quantity of unchanged germ-plasm with some of their nuclei.

In Sachs’ excellent lectures on the physiology of plants, we read on page 723^[140]—‘In the true mosses almost any cell of the roots, leaves and shoot-axes, and even of the immature sporogonium, may grow out under favourable conditions, become rooted, form new shoots, and give rise to an independent living plant.’ Since such plants produce germ-cells at a later period, we have here a case which requires the assumption that all or nearly all cells must contain germ-plasm.

The theory of the continuity of the germ-plasm seems to me to be still less disproved or even rendered improbable by the facts of the alternation of generations. If the germ-plasm may pass on from the egg into certain somatic cells of an individual, and if it can be further transmitted along certain lines, there is no difficulty in supposing that it may be transmitted through a second, third, or through any number of individuals produced from the former by budding. In fact, in the Hydroids, on which my theory of the continuity of the germ-plasm has been chiefly based, alternation of generations is the most important means of propagation.

II. The Significance of the Polar Bodies.

We have already seen that the specific nature of a cell depends upon the molecular structure of its nucleus; and it follows from this conclusion that my theory is further, and as I believe strongly, supported, by the phenomenon of the expulsion of polar bodies, which has remained inexplicable for so long a time.

For if the specific molecular structure of a cell-body is caused and determined by the structure of the nucleoplasm, every kind of cell which is histologically differentiated must have a specific nucleoplasm. But the egg-cell of most animals, at any rate during the period of growth, is by no means an indifferent cell of the most primitive type. At such a period its cell-body has to perform quite peculiar and specific functions; it has to secrete nutritive substances of a certain chemical nature and physical constitution, and to store up this food-material in such a manner that it may be at the disposal of the embryo during its development. In most cases the egg-cell also forms membranes which are often characteristic of particular species of animals. The growing egg-cell is therefore histologically differentiated: and in this respect resembles a somatic cell. It may perhaps be compared to a gland-cell, which does not expel its secretion, but deposits it within its own substance^[141]. To perform such specific functions it requires a specific cell-body, and the latter depends upon a specific nucleus. It therefore follows that the growing egg-cell must possess nucleoplasm of specific molecular structure, which directs the above-mentioned secretory functions of the cell. The nucleoplasm of histologically differentiated cells may be called histogenetic nucleoplasm, and the growing egg-cell must contain such a substance, and even a certain specific modification of it. This nucleoplasm cannot possibly be the same as that which, at a later period, causes embryonic development. Such development can only be produced by true germ-plasm of immensely complex constitution, such as I have previously attempted to describe. It therefore follows that the nucleus of the egg-cell contains two kinds of nucleoplasm:—germ-plasm and a peculiar modification of histogenetic nucleoplasm, which may be called ovogenetic nucleoplasm. This substance must greatly preponderate in the young egg-cell, for, as we have already seen, it controls the growth of the latter. The germ-plasm, on the other hand, can only be present in minute quantity at first, but it must undergo considerable increase during the growth of the cell. But in order that the germ-plasm may control the cell-body, or, in other words, in order that embryonic development may begin, the still preponderating ovogenetic nucleoplasm must be removed from the cell. This removal takes place in the same manner as that in which differing nuclear substances are separated during the ontogeny of the embryo: viz. by nuclear division, leading to cell-division. The expulsion of the polar bodies is nothing more than the removal of ovogenetic nucleoplasm from the egg-cell. That the ovogenetic nucleoplasm continues to greatly preponderate in the nucleus up to the very last, may be concluded from the fact that two successive divisions of the latter and the expulsion of two polar bodies appear to be the rule. If in this way a small part of the cell-body is expelled from the egg, the extrusion must in all probability be considered as an inevitable loss, without which the removal of the ovogenetic nucleoplasm cannot be effected.

This is my theory of the significance of polar bodies, and I do not intend to contrast it, in extenso, with the theories propounded by others; for such theories are well known and differ essentially from my own. All writers agree in supposing that something which would be an obstacle to embryonic development is removed from the egg; but opinions differ as to the nature of this substance and the precise reasons for its removal^[142]. Some observers (e. g. Minot^[143], van Beneden, and Balfour) regard the nucleus as hermaphrodite, and assume that in the polar bodies the male element is expelled in order to render the egg capable of fertilization. Others speak of a rejuvenescence of the nucleus, others again believe that the quantity of nuclear substance must be reduced in order to become equal to that of the generally minute sperm-nucleus, and that the proportions for nuclear conjugation are in this way adjusted.

The first view seems to me to be disproved by the fact that male as well as female qualities are transmitted by the egg-cell, while the sperm-cell also transmits female qualities. The germ-plasm of the nucleus of the egg cannot therefore be considered as female, and that of the sperm-nucleus cannot be considered as male: both are sexually indifferent. The last view has been recently formulated by Strasburger, who holds that the quantity of the idioplasm contained in the germ-nucleus must be reduced by one half, and that a whole nucleus is again produced by conjugation with the sperm-nucleus. Although I believe that the fundamental idea underlying this hypothesis is perfectly correct, viz. that the influence of each nucleus is as largely dependent upon its quantity as upon its quality, I must raise the objection that the decrease in quantity is not the explanation of the expulsion of polar bodies. The quantity of idioplasm contained in the germ-nucleus is, as a matter of fact, not reduced by one-half but by three-fourths, for two divisions take place one after the other. Thus by conjugation with the sperm-nucleus, which we may assume to be of the same size as the germ-nucleus, a nucleus is produced which can only contain half as much idioplasm as was present in the original germ-nucleus, before division. Strasburger’s view leaves unexplained the question why the size of the germ-nucleus, before the expulsion of polar bodies, was thus twice as large; and even if we neglect the theory of ovogenetic nucleoplasm and assume that this larger nucleus was entirely made up of germ-plasm, it must be asked why the egg did not commence segmentation earlier. The theory which explains the sperm-cell as the vitalizing principle which starts embryonic development, like the spark which kindles the gunpowder, would indeed answer this question in a very simple manner. But Strasburger does not accept this theory, and holds, as I do, a very different view, which will be explained later on.

If, on the other hand, we assume that the germ-nucleus contains two different kinds of nucleoplasm, the question is answered quite satisfactorily. In treating of parthenogenesis, further on, I shall mention a fact which seems to me to furnish a real proof of the validity of this explanation; and, if we accept this fact for the present, it will be clear that the simple explanation now offered of phenomena which are otherwise so difficult to understand, would also largely support the theory of the continuity of the germ-plasm. Such an explanation would, above all, very clearly demonstrate the co-existence of two nucleoplasms with different qualities in one and the same nucleus. My theory must stand or fall with this explanation, for if the latter were disproved, the continuity of the germ-plasm could not be assumed in any instance, not even in the simplest cases, where, as in Diptera, the germ-cells are the first-formed products of embryonic development. For even in these insects the egg possesses a specific histological character which must depend upon a specifically differentiated nucleus. If then two kinds of nucleoplasm are not present, we must assume that in such cases the germ-plasm of the newly formed germ-cells, which has passed on unchanged from the segmentation nucleus, is at once transformed entirely into ovogenetic nucleoplasm, and must be re-transformed into germ-plasm at a later period when the egg is fully mature. We could not in any way understand why such a re-transformation requires the expulsion of part of the nuclear substance.

At all events, my explanation is simpler than all others, in that it only assumes a single transformation of part of the germ-plasm, and not the later re-transformation of ovogenetic nucleoplasm into germ-plasm, which is so improbable. The ovogenetic nucleoplasm must possess entirely different qualities from the germ-plasm; and, above all, it does not readily lead to division, and thus we can better understand the fact, in itself so remarkable, that egg-cells do not increase in number by division, when they have assumed their specific structure, and are controlled by the ovogenetic nucleoplasm. The tendency to nuclear division, and consequently to cell-division, is not produced until changes have to a certain extent taken place in the mutual relation between the two kinds of nucleoplasm contained in the germ-nucleus. This change is coincident with the attainment of maximum size by the body of the egg-cell. Strasburger, supported by his observations on Spirogyra, concludes that the stimulus towards cell-division emanates from the cell-body; but the so-called centres of attraction at the poles of the nuclear spindle obviously arise under the influence of the nucleus itself, even if we admit that they are entirely made up of cytoplasm. But this point has not been decided upon, and we may presume that the so-called ‘Polkörperchen’ of the spindle (Fol) are derived from the nucleus, although they are placed outside the nuclear membrane^[144]. Many points connected with this subject are still in a state of uncertainty, and we must abstain from general conclusions until it has been possible to demonstrate clearly the precise nature of certain phenomena attending indirect nuclear division, which still remain obscure in spite of the efforts of so many excellent observers. We cannot even form a decided opinion as to whether the chromatin or the achromatin of the nuclear thread is the real idioplasm. But although these points are not yet thoroughly understood, we are justified in maintaining that the cell enters upon division under the influence of certain conditions of the nucleus, although the latter are invisible until cell-division has already commenced.

I now pass on to examine my hypothesis as to the significance of the formation of polar bodies, in the light of those ascertained facts which bear upon it.

If the expulsion of the polar bodies means the removal of the ovogenetic nucleoplasm after the histological differentiation of the egg-cell is complete, we must expect to find polar bodies in all species except those in which the egg-cell has remained in a primitive undifferentiated condition, if indeed such species exist. Wherever the egg-cell assumes the character of a specialized cell, e.g. in the attainment of a particular size or constitution, in the admixture of food-yolk, or the formation of membranes, it must also contain ovogenetic nucleoplasm, which must ultimately be removed if the germ-plasm is to gain control over the egg-cell. It does not signify at all, in this respect, whether the egg is or is not destined for fertilization.

If we examine the Metazoa in regard to this question, we find that polar bodies have not yet been discovered in sponges^[145], but this negative evidence is no proof that they are really absent. In all probability, no one has ever seriously endeavoured to find them, and there are perhaps difficulties in the way of the proofs of their existence, because the egg-cell lies free for a long time and even moves actively in the tissue of the mesogloea. We might expect that the formation of polar bodies takes place here, as in all other instances, when the egg becomes mature, that is, at a time when the eggs are already closely enveloped in the sponge tissue. At all events the eggs of sponges, as far as they are known, attain a specific nature, in the possession of a peculiar cell-body, frequently containing food-yolk, and of the nucleus which is characteristic of all animal eggs during the process of growth. Hence we cannot doubt the presence of a specific ovogenetic nucleoplasm, and must therefore also believe that it is ultimately removed in the polar bodies.

In other Coelenterata, in worms, echinoderms, and in molluscs polar bodies have been described, as well as in certain Crustacea, viz. in Balanus by Hoek and in Cetochilus septentrionale by Grobben. The latter instance appears to be quite trustworthy, but there is some doubt as to the former and also as regards Moina (a Daphnid), in which Grobben found a body, which he considered to be a polar body, on the upper pole of an egg which was just entering upon segmentation. In insects polar bodies have not been described up to the present time^[146], and only in a few cases in Vertebrata, as in Petromyzon by Kupffer and Benecke.

It must be left to the future to decide whether the expulsion of polar bodies occurs in those large groups of animals in which they have not been hitherto discovered. The fact, however, that they have not been so discovered cannot be urged as an objection to my theory, for we do not know a priori whether the removal of the ovogenetic nucleoplasm has not been effected in the course of phylogeny in some other and less conspicuous manner. The cell-body of the polar globules is so minute in many eggs that it was a long time before the cellular nature of these structures was recognized^[147]; and it is possible that their minute size may point to the fact that a phyletic process of reduction has taken place, to the end that the egg may be deprived of as little material as possible. It is at all events proved that in all Metazoan groups the nucleus undergoes changes during the maturation of the egg, which are entirely similar to those which lead to the formation of polar bodies in those eggs which possess them. In the former instances it is possible that nature has taken a shortened route to gain the same end.

It would be an important objection if it could be shown that no process corresponding to the expulsion of polar bodies takes place in the male germ-cells, for it is obvious that here also we should, according to my theory, expect such a process to occur. The great majority of sperm-cells differ so widely in character from the ordinary indifferent (i. e. undifferentiated) cells, that they are evidently histologically differentiated in a very high degree, and hence the sperm-cells, like the yolk-forming germ-cells, must possess a specific nuclear substance. The majority of sperm-cells therefore resemble the somatic cells in that they have a specific histological structure, but their characteristic form has nothing to do with their fertilizing power, viz. with their power of being the bearers of germ-plasm. Important as this structure is, in order to render it possible that the egg-cell may be approached and penetrated, it has nothing to do with the property of the sperm-cell to transmit the qualities of the species and of the individual to the following generation. The nuclear substance which causes such a cell to assume the appearance of a thread, or a stellate form (in Crustacea), or a boomerang form (present in certain Daphnids), or a conical bullet shape (Nematodes), cannot possibly be the same nuclear substance as that which, after conjugation with the egg-cell, contains in its molecular structure the tendency to build up a new Metazoon of the same kind as that by which it was produced. We must, therefore, conclude that the sperm-cell also contains two kinds of nucleoplasm, namely, germ-plasm and spermogenetic nucleoplasm.

It is true that we cannot say a priori whether the influence exercised on the sperm-cell by the spermogenetic nucleoplasm might not be eliminated by some means other than its removal from the cell. It is conceivable, for instance, that this substance may be expelled from the nucleus, but may remain in the cell-body, where it is in some way rendered powerless. We do not yet really know anything of the essential conditions of nuclear division, and it is quite impossible to bring forward any facts in support of my previous suggestion. The germ-plasm is supposed to be present in the nucleus of the growing egg-cell in smaller quantity than the ovogenetic nucleoplasm, and the germ-plasm gradually increases in quantity: thus when the egg has attained its maximum size, the opposition between the two different kinds of nucleoplasm becomes so marked, in consequence of the alteration in their quantitative relations, that their separation, viz. nuclear division, results. But although we are not able to distinguish, by any visible characteristics, the different kinds of nucleoplasm which may be united in one nuclear thread, the assumption that the influence of each kind bears a direct proportion to its quantity is the most obvious and natural one. The tendency of the germ-plasm contained in the nucleus cannot make itself felt so long as an excess of ovogenetic nucleoplasm is also present. We may imagine that the effects of the two different kinds of nucleoplasm are combined to produce a resultant effect; but when the two influences exerted upon the cell are nearly opposed, only the stronger can make itself felt, and in such a case the latter must exceed the former in quantity, because part of it is as it were neutralized by the other nucleoplasm working in an opposite direction. This metaphorical representation may give us a clue to explain the fact that the ovogenetic nucleoplasm comes to exceed the germ-plasm in quantity. For obviously these two kinds of nucleoplasm exert opposite tendencies in at least one respect. The germ-plasm tends to effect the division of the cell into the two first segmentation spheres; the ovogenetic nucleoplasm, on the other hand, possesses a tendency towards the growth of the cell-body without division. Hence the germ-plasm cannot make itself felt, and is not able to expel the ovogenetic nucleoplasm until it has reached such a relative size as enables it to successfully oppose the latter.

Applying these ideas to the sperm-cells we must see whether the expulsion of part of the nuclear substance, viz. of the spermogenetic nucleoplasm, corresponding to the ovogenetic nucleoplasm, takes place in them also.

As far as we can judge from thoroughly substantiated observations such phenomena are indeed found in many cases, although they appear to be different from those occurring in the egg-cell, and cannot receive quite so certain an interpretation.

The attempt to prove that a process similar to the expulsion of polar bodies takes place in the formation of sperm-cells has already been made by those observers who regard such expulsion as the removal of the male element from the egg, thus leading to sexual differentiation; for such a theory also requires the removal of part of the nuclear substance from the maturing sperm-cell. Thus, according to E. van Beneden and Ch. Julin, the cells which, in Ascaris, produce the spermatogonia (mother-cells of the sperm-cells), expel certain elements from their nuclear plate, a phenomenon which has not been hitherto observed in any other animal, and even in this instance has only been inferred and not directly observed. Moreover the sperm-cells have not attained their specific form (conical bullet-shaped) at the time when this expulsion takes place from the spermatogonia, and we should expect that the spermogenetic nucleoplasm would not be removed until it has completed its work, viz. not until the specific shape of the sperm-cell has been attained. We might rather suppose that phenomena explicable in this way are to be witnessed in those sperm-blastophores (mother-cells of sperm-cells) which, as has been known for a long time, are not employed in the formation of the nuclei of sperm-cells, but for the greater part remain at the base of the latter and perish after their maturation and separation. In this case an influence might be exerted by these nuclei upon the specific form of the sperm-cells, for the former arise and develope in the form of bundles of spermatozoa in the interior of the mother-cell.

It has been already shown in many groups of animals that parts of the sperm-mother-cells^[148] perish, without developing into sperm-cells, as in Selachians, in the frog, in many worms and snails, and also in mammals (Blomfield). But the attention of observers has been directed to that part of the cell-body which is not used in the formation of sperm-cells, rather than to the nucleus; and the proof that part of the nucleus also perishes is still wanting in many of these cases. Fresh investigation must decide whether the nucleus of the sperm-mother-cell perishes as a general rule, and whether part of the nucleus is rendered powerless in some other way, where such mother-cells do not exist. Perhaps the paranucleus (Nebenkern) of the sperm-cell, first described by La Valette St. George, and afterwards found in many animals of very different groups, is the analogue of the polar body. It is true that this so-called paranucleus is now considered as a condensed part of the cell-body, but we must remember that it has been hitherto a question whether the head of the spermatozoon is formed from the nucleus of the cell or from the paranucleus; and that the observers who held the former view were in consequence obliged to regard the paranucleus as a product of the cell-body. But according to the most recent investigations of Fol^[149], Roule^[150], Balbiani^[151], and Will^[152], upon the formation of the follicular epithelium in the ovary of different groups, it is not improbable that parts of the nucleus may become detached without passing through the process of karyokinesis. Thus it is very possible that the paranucleus may be a product of the main nucleus and not a condensed part of the cell-body. This view is supported by its behaviour with staining reagents, while the other view, that it arises from the cell-substance, is not based upon direct observation. Consequently future investigation must decide whether the paranucleus is to be considered as the spermogenetic nucleoplasm expelled from the nucleus. But even if this question is answered in the affirmative, we should still have to explain why this nuclear substance, remaining in the cell-body, does not continue to exercise any control over the latter.

Strasburger has recently enumerated a large number of cases from different groups of plants, in which the maturation of both male and female germ-cells is accompanied by phenomena similar to the expulsion of polar bodies. In this respect the phenomena occurring in the pollen-grains of Phanerogams bear an astonishing resemblance to the maturation of the animal egg. For instance, in the larch, the sperm-mother-cell divides three times in succession, the products of division being very unequal on each occasion; and exactly as in the case of polar bodies, the three small so-called vegetative cells shrink rapidly after separation, and have no further physiological value. According to Strasburger, the so-called ‘ventral canal-cell,’ which, in mosses, ferns, and Conifers, separates from the female germ-cell, reminds us, in every way, of the polar bodies of animal eggs. Furthermore, the spermatozoids in the mosses and vascular cryptogams throw off a small vesicle before performing their functions^[153]. On the other hand the equivalents of ‘polar bodies’ (the ‘ventral canal-cells’) are said to be absent in the Cycads, although these are so nearly allied to Conifers. Furthermore, ‘no phenomenon occurs in the oospheres (ova) of Angiosperms which can be compared to the formation of polar bodies.’ Strasburger therefore concludes that the separation of certain parts from the germ-cells is not in all cases necessary for maturation, and that such phenomena are not fundamental, like those of fertilization, which must always take place along the same morphological lines. He further concludes that the former phenomena are only necessary in the case of the germ-cells of certain organisms, in order to bring the nuclei destined for the sexual act into the physiological condition necessary for its due performance.

I am unwilling to abandon the idea that the expulsion of the histogenetic parts of the nuclear substance, during the maturation of germ-cells, is also a general phenomenon in plants; for the process appears to be fundamental, while the argument that it has not been proved to occur universally is only of doubtful value. The embryo-sac of Angiosperms is such a complex structure that it seems to me to be possible (as it does to Strasburger) that ‘processes which precede the formation of the egg-cell have borne relation to the sexual differentiation of the nucleus of the egg.’ Besides, it is possible that the vegetable egg-cell may, in certain cases, possess so simple a structure and so small a degree of histological specialization, that it would not be necessary for it to contain any specific histogenetic nucleoplasm: thus it would consist entirely of germ-plasm from the first. In such cases, of course, its maturation would not be accompanied by the expulsion of somatic nucleoplasm.

I have hitherto abstained from discussing the question as to whether the process of the formation of polar bodies may require an interpretation which is entirely different from that which I have given it, whether it may receive a purely morphological interpretation.

In former times it could only be regarded as of purely phyletic significance: it could only be looked upon as the last remnant of a process which formerly possessed some meaning, but which is now devoid of any physiological importance. We are indeed compelled to admit that a process does occur in connexion with the true polar bodies of animal eggs, which we cannot explain on physiological grounds; I mean the division of the polar bodies after they have been expelled from the egg. In many animals the two polar bodies divide again after their expulsion, so as to form four bodies, which distinctly possess the structure of cells, as Trinchese observed in the case of gastropods. But, in the first place, this second division does not always take place, and, secondly, it is very improbable that a process which occurs during the first stage of ontogeny, or more properly speaking, before the commencement of ontogeny, and which is, therefore, a remnant of some excessively ancient phyletic stage, would have been retained up to the present day unless it possessed some very important physiological significance. We may safely maintain that it would have disappeared long ago if it had been without any physiological importance. Relying on our knowledge of the slow and gradual, although certain, disappearance, in the course of phylogeny, of organs which have lost their functions, and of processes which have become meaningless, we are compelled to regard the process of the formation of polar bodies as of high physiological importance. But this view does not exclude the possibility that the process possessed a morphological meaning also, and I believe that we are quite justified in attempting (as Bütschli^[154] has recently done) to discover what this morphological meaning may have been.

Should it be finally proved that the expulsion of polar bodies is nothing more than the removal of histogenetic nucleoplasm from the germ-cell, the opinion (which is so intimately connected with the theory of the continuity of the germ-plasm) that a re-transformation of specialised idioplasm into germ-plasm cannot occur, would be still further confirmed; for we do not find that any part of an organism is thrown away simply because it is useless: organs that have lost their functions are re-absorbed, and their material is thus employed to assist in building up the organism.