Essays Upon Heredity and Kindred Biological Problems / Authorised Translation

It is at any rate a delusion to believe that we have explained natural death, by deriving it from the starvation of the soma of the Orthonectides, by the aid of the unproved assumption of the transmission of acquired variations. We must first explain why these organisms produce only a limited number of reproductive cells which are all extruded at once, so that the soma is rendered helpless. Why should not the reproductive cells ripen in succession as they do indirectly among the Monoplastides, that is to say in a succession of generations, and as they do directly in great numbers among the Metazoa? There would then be no necessity for the soma to die, for a few reproductive cells would always be present, and render the persistence of the individual possible. In fact, the whole arrangement—the formation of reproductive cells at one time only, and their sudden extrusion,—presupposes the mortality of the somatic cells, and is an adaptation to it, just as this mortality itself must be regarded as an adaptation to the simultaneous ripening and sudden extrusion of the generative cells. In short, there is no alternative to the supposition stated above, viz. that the mortality of the somatic cells arose with the differentiation of the originally homogeneous cells of the Polyplastids into the dissimilar cells of the Heteroplastids. And this is the first beginning of natural death.

Probably at first the somatic cells were not more numerous than the reproductive cells, and while this was the case the phenomenon of death was inconspicuous, for that which died was very small. But as the somatic cells relatively increased, the body became of more importance as compared with the reproductive cells, until death seems to affect the whole individual, as in the higher animals, from which our ideas upon the subject are derived. In reality, however, only one part succumbs to natural death, but it is a part which in size far surpasses that which remains and is immortal,—the reproductive cells.

Götte combats the statement that the idea of death necessarily implies the existence of a corpse. Hence he maintains that the cellular sac which is left after the extrusion of the reproductive cells among the Orthonectides, and which ultimately dies, is not a corpse; ‘for it does not represent the whole organism, any more than the isolated ectoderm of any other Heteroplastid’ (l. c., p. 48). But it is only a popular notion that a corpse must represent the entire organism. In cases of violent death this idea is correct, because then the reproductive cells are also killed. But as soon as we recognise that the reproductive cells on the one side, and the somatic cells on the other, form respectively the immortal and mortal parts of the Metazoan organism, then we must acknowledge that only the latter,—that is, the soma without the reproductive cells,—suffers natural death. The fact that all the reproductive cells have not left the body (as sometimes happens) before natural death takes place, does not affect this conception. Among insects, for instance, it may happen that natural death occurs before all the reproductive cells have matured, and these latter then die with the soma. But this does not make any difference to their potential immortality, any more than it modifies the scientific conception of a corpse. The idea of natural death involves that of a corpse, which consists of the soma, and when the latter happens to contain reproductive cells, these do not succumb to a natural death, which can never apply to them, but to an accidental death. They are killed by the death of the soma just as they might be killed by any other accidental cause of death.

The scientific conception of a corpse is not affected, whether the dead soma remains whole for some time, or falls to pieces at once. I cannot therefore agree with Götte when he denies that an Orthonectid possesses ‘the possibility of becoming a corpse’ (in his sense of the word) because ‘its death consists in the dissolution of the structure of the organism.’ When the young of the Rhabdites form of Ascaris nigrovenosa bore through the body-walls of their parent, cause it to disintegrate and finally devour it, the whole organism disappears, and it would be difficult to say whether a corpse exists in the popular sense of the word. But, scientifically speaking, there is certainly a corpse; the real soma of the animal dies, and this, however subdivided, must be considered as a corpse. The fact that natural death is so difficult to define without any accurate conception of what is meant by a corpse, proves the necessity for arriving at a scientific idea as to the meaning of the latter. There is no death without a corpse—whether the latter be small or large, whole or in pieces.

If we compare the bodies of the higher Metazoa with those of the lower, we see at once that not only has the structure of the body increased in size and complexity as far as the soma is concerned, but we also see that another factor has been introduced, which exercises a most important influence in lengthening the duration of life. This is the replacement of cells by multiplication. Somatic cells have acquired (at any rate in most tissues) the power of multiplying, after the body is completely developed from the reproductive cells. The cells which have undergone histological differentiation can increase by fission, and thus supply the place of those which are being continually destroyed in the course of metabolism. The difference between the higher and lower Metazoa in this respect lies in the fact that there is only one generation of somatic cells in the latter, and these are used up in the process of metabolism at almost the same time that the reproductive cells are extruded, while among the former there are successive generations of somatic cells. I have elsewhere endeavoured to render the duration of life in the animal kingdom intelligible by the application of this principle, and have attempted to show that its varying duration is determined in different species by the varying number of somatic cell-generations^[86]. Of course, the varying duration of each cell-generation materially influences the total length of life, and experience teaches us that the duration of cell-generations varies, not only in the lowest Metazoa as compared with the highest, but even in the various kinds of cells in one and the same species of animal.

We must, for the present, leave unanswered the question—upon what changes in the physical constitution of protoplasm does the variation in the capacity for cell-duration depend; and what are the causes which determine the greater or smaller number of cell-generations. I mention this obvious difficulty because it is the custom to meet every attempt to search deeper into the common phenomena of life with the reproach that so much is still left unexplained. If we must wait for the explanation of these processes until we have ascertained the molecular structure of cells, together with the changes that occur in this structure and the consequences of the changes, we shall probably never understand either the one or the other. The complex processes of life can only be followed by degrees, and we can only hope to solve the great problem by attacking it from all sides.

Therefore it is, in my opinion, an advance if we may assume that length of life is dependent upon the number of generations of somatic cells which can succeed one another in the course of a single life; and, furthermore, that this number, as well as the duration of each single cell-generation, is predestined in the germ itself. This view seems to me to derive support from the obvious fact that the duration of each cell-generation, and also the number of generations, undergo considerable increase as we pass from the lowest to the highest Metazoa.

In an earlier work^[87] I have attempted to show how exactly the duration of life is adapted to the conditions by which it is surrounded; how it is lengthened or shortened during the formation of species, according to the conditions of life in each of them; in short, how it is throughout an adaptation to these conditions. A few points however were not touched upon in the work referred to, and these require discussion; their consideration will also throw some light upon the origin of natural death and the forms of life affected by it.

I have above explained the limited duration of the life of somatic cells in the lower Metazoa—Orthonectides—as a phenomenon of adaptation, and have ascribed it to the operation of natural selection, at the same time pointing out that the existence of immortal Metazoan organisms is conceivable. If the Monoplastides are able to multiply by fission, through all time, then their descendants, in which division of labour has induced the antithesis of reproductive and somatic cells, might have done the same. If the Homoplastid cells reproduced their kind uninterruptedly, equal powers of duration must have been possible for the two kinds of Heteroplastid cells; they too might have been immortal so far as immortality only depends upon the capacity for unlimited reproduction.

But the capacity for existence possessed by any species is not only dependent upon the power within it; it is also influenced by the conditions of the external world, and this renders necessary the process which we call adaptation. Thus it is just as inconceivable that either a homogeneous or a heterogeneous cell-colony possessing the physiological value of a multicellular individual should continue to grow to an unlimited extent by continued cell-division, as it is inconceivable that a unicellular being should increase in size to an unlimited extent. In the latter case the process of cell-division imposes a limit upon the size attained by growth. In the former, the requirements of nutrition, respiration, and movement must prescribe a limit to the growth of the cell-colony which constitutes the individual of the higher species, just as in the case of the unicellular Monoplastides, and it does not affect the argument if we consider this limitation to be governed by the process of natural selection. It would only be possible to regulate the relations of the single cells of the colony to each other by fixing the number of cells within narrow limits. During the development of Magosphaera—one of the Homoplastides—the cells arrange themselves in the form of a hollow sphere, lying in a gelatinous envelope. But the fact that reproduction does not follow the simple unvarying rhythm of unicellular organisms is of more importance; for a rhythm of a higher order appears, in which each cell of the colony separates from its neighbours, when it has reached a certain size, and proceeds by very rapid successive divisions to give rise to a certain number of parts which arrange themselves as a new colony. The number of divisions is controlled by the number of cells to which the colony is limited, and at first this number may have been very small. With the introduction of this secondary higher rhythm during reproduction, the first germ of the Polyplastides became evident; for then each process of fission was not, as in unicellular organisms, equivalent to all the others; for in a colony of ten cells the first fission differs from the second, third, or tenth, both in the size of the products of division and also in remoteness from the end of the process. This secondary fission is what we know as segmentation.

It seems to me of little importance whether the first process of segmentation takes place in the water or within a cyst, although it is quite possible that the necessity for some protective structure appeared at a very early period, in order to shield the segmenting cell from danger.

It is impossible to accept Götte’s conception of the germ (Keim), and at this point the question arises as to its true meaning. I should propose to include under this term every cell, cytode, or group of cells which, while not possessing the structure of the mature individual of the species, possesses the power of developing into it under certain circumstances. The emphasis is now laid upon the expression development, which is something opposed to simple growth, without change of form. A cell which becomes a complete individual by growth alone is not a germ but an individual, although a very small one. For example, the small encapsuled Heliozoon, which arises as the product of multiple fission, is not a germ in our sense of the word. It is an individual, provided with all the characteristic marks of its species, and it has only to protrude the retracted processes (pseudopodia) and to take in the expelled water (formation of vacuoles) in order to become capable of living in a free state. In this sense of the word, germs are not confined to the Polyplastides, but are found in many Monoplastides. There is nevertheless, in my opinion, a profound and significant difference between the germs of these two groups. And this lies not so much in the morphological as in the developmental significance of these structures. As far as I have been able to compare the facts, I may state that the germs of the Monoplastides are entirely of secondary origin, and have never formed the phyletic origin of the species in which they are found. For instance, the spore-formation of the Gregarines resulted from a gradually increasing process of division, which was concentrated into the period of encystment; and it was induced by a necessity for rapid multiplication due to the parasitic life and unfavourable surroundings of these animals. If Gregarines were free-living animals, they would not need this method of reproduction. The encysted animal would probably divide into eight, four, or two parts, or perhaps, like many Infusoria^[88], it would not divide at all, so that the whole reproduction would depend on simple fission alone during the free state.

The original mode of reproduction among the Monoplastides was undoubtedly simple fission. This became connected with encystment, which originally took place without multiplication; and only when the divisions in the cyst became excessively numerous did such minute plastids appear that a genuine process of development had to be undergone in order to produce complete individuals. Here we have the general conception of the germ as I defined it. Its limitations are naturally not very sharply defined, for it is impossible to draw an absolute distinction between simple growth and true development accompanied by changes in form and structure. For instance, Häckel’s Protomyxa aurantiaca divides within its cyst into numerous plastids, which might be spoken of as germs. But the changes of form which they undergo before they become young Protomyxae are very small, and for the most part depend upon the expansion of the body, which existed in the capsule as a contracted pear-shaped mass. It is therefore more correct to speak only of the simple growth of the products of the fission of the parent organism, and to look upon these products as young Protomyxae rather than germs. On the other hand, the young animals which creep out of the germs (the ‘spores’) of Gregarina gigantea, described by E. van Beneden, differ essentially from the adult, and pass through a series of developmental stages before they assume the characteristic form of a Gregarine.

This is true development^[89]. But such a method of germ-formation and development are found most frequently, although not exclusively, among the parasitic Monoplastides, and this fact alone serves to indicate their secondary origin. It is a form of ontogenetic development differing from that of the Polyplastides in that it does not revert to a phyletically primitive condition of the species, but, on the contrary, exhibits stages which first appear in the phyletic development of the specific form. The Psorosperms were only formed after the Gregarines had become established as a group. The amoeboid organisms which creep out of them are in no way to be regarded as the primitive forms of the Gregarines, even if the latter may have resembled them, but they are coenogenetic forms produced by the necessity for a production of numerous and very minute germs. The necessity for a process of genuine development perhaps depends upon the small amount of material contained in one of these germs, and on other conditions, such as change of host, change of medium, etc. It therefore results that the fundamental law of biogenesis does not apply to the Monoplastides; for these forms are either entirely without a genuine ontogeny and only possess the possibility of growth, or else they are only endowed with a coenogenetic ontogeny^[90].

Some authorities may be inclined to limit the above proposition, and to maintain that we must admit the possibility that we are likely to occasionally meet with an ontogeny of which the stages largely correspond with the most important stages in the phyletic development of the species, and that the ontogenetic repetition of the phylogeny, although not the rule, may still occur as a rare exception in the Protozoa.

A careful consideration of the subject indicates, however, that the occurrence of such an exception is very improbable. Such an ontogeny would, for instance, occur if one of the lowest Monoplastides, such as a Moneron, were to develope into a higher form, such as one of the Flagellata, with mouth, eye-spot, and cortical layer, under such external conditions that it would be advantageous for the existence of its species that it should no longer reproduce itself by simple fission, but that the periodical formation of a cyst (which was perhaps previously part of the life-history) should be associated with the occurrence of numerous divisions within the cyst itself, and with the formation of germs. We must suppose either that these germs were so minute that the young animals could not become Flagellata directly, or that it was advantageous for them to move and feed as Monera at an early period, and to assume the more complex structure of the parent by gradual stages. In other words, the phyletic development would proceed hand in hand with the ontogeny corresponding to it, although not from any internal cause, but as an adaptation to the existing conditions of life. But the supposed transformation of the species also depended upon these same conditions of life, which must therefore have been of such a nature as to bring about simultaneously, by an intercalation of germs and by a genuine development, the evolution of the form in question in the last stage of its ontogeny, and the maintenance of its original condition during the initial stage. Such a combination of circumstances can have scarcely ever happened. Against the occurrence of such a transformation as we have supposed, it might be argued, indeed, that the assumed production of very numerous germs does not occur among free-living Monoplastides. Those which have acquired parasitic habits must be younger phyletic forms, for their first host—whether a lowly or a highly organized Metazoon—must have appeared before they could gain access to it and adapt themselves to the conditions of a parasitic life, and by this time the Flagellate Infusoria were already established. It is by far less probable that the persistence or rather the intercalation of the ancestral form would occur in an ontogenetic cycle, consisting of a series of stages, and not of two only, as in our example. For as soon as reproduction can be effected by the simple fission of the adult, not only is there no reason why the earlier phyletic stages should be again and again repeated, but such recapitulation is simply impossible. We cannot, therefore, conclude that the anomalous early stages of a Monoplastid such as Acineta correspond with an early form of phyletic development.

Supposing, for instance, that the Acinetaria were derived from the Ciliata, then this transformation must have taken place in the course of the continued division of the ciliate ancestor—partially connected with encystment, but for the most part independently of it. Of the myriads of generations which such a process of development may have occupied, perhaps the first set moved with suctorial processes, while the second gradually adopted sedentary habits, and throughout the whole of the long series, each succeeding generation must have been almost exactly like its predecessor, and must always have consisted of individuals which possessed the characters of the species.

This does not exclude the possibility that in spite of an assumed sedentary mode of life, the need for locomotion and for obtaining food in fresh places may have arisen at some period of life. But whenever formation of swarm-spores takes place instead of simple fission, this does not depend upon the persistence of an ancestral form in the ontogenetic cycle, but is due to the intercalation of an entirely new ontogenetic stage, which happens to resemble an ancestral form, in the possession of cilia, etc.

I imagine that I have now sufficiently explained the above proposition, that the repetition of the phylogeny in the ontogeny does not and cannot occur among unicellular organisms.

With the Polyplastides the opposite is the case. There is no species, as far as we know, which does not—either in each individual, or after long cycles which comprise many individuals (alternation of generations)—invariably revert to the Monoplastid state. This applies from the lowest forms, such as Magosphaera and the Orthonectides, up to the very highest. In the latter a great number of intermediate phyletic stages always occur, although some have been omitted as the result of concentration in the ontogeny, while others have sometimes been intercalated.

Sexual reproduction is the obvious cause of this very important arrangement. Even if this is an hypothesis rather than a fact we must nevertheless accept it unconditionally, because it is a method of reproduction found everywhere. It is the rule in every group of the animal kingdom, and is only absent in a few species in which it is replaced by parthenogenesis. In these latter instances sexual reproduction may be local, and entirely absent in certain districts only (Apus), or it may be only apparently wanting; in some cases where it is undoubtedly absent, it is equally certain that it was present at an earlier period (Limnadia Hermanni). We cannot as yet determine whether its loss will not involve the degeneration and ultimate extinction of the species in question.

If the essential nature of sexual reproduction depends upon the conjugation of two equivalent but dissimilar morphological elements, then we can understand that a multicellular being can only attain sexual reproduction when a unicellular stage is present in its development; for the coalescence of entire multicellular organisms in such a manner that fusion would only take place between equivalent cells, would seem to be impracticable. In the necessity for sexual reproduction, there is therefore also implied the necessity for reverting to the original condition of the Polyplastides—that of a single cell—and upon this alone depends the fundamental law of biogenesis. This law is therefore confined to the Polyplastides, and does not apply to the Monoplastides; and Götte’s suggestion that the latter fall back into the primitive condition of the organism during their encystment (rejuvenescence), finds no support in this aspect of the question.

I have on a previous occasion^[91] referred the utility of death to the ultimate fact that the unending life of the Metazoan body would be a useless luxury, and to the fact that the individuals would necessarily become injured in the course of time, and would be therefore ‘not only valueless to the species, but ... even harmful, for they take the place of those which are sound’ (l. c., p. 24). I might also have said that such damaged individuals would sooner or later fall victims to some accidental death, so that there would be no possibility of real immortality. I now propose to examine this statement a little more closely, and to return to a question which has already been alluded to before.

It is obvious that the advantages above set forth did not form the motive which impelled natural selection to convert the immortal life of the Monoplastides into the life of limited duration possessed by the Heteroplastides, or more correctly, which led to the restriction of potential immortality to the reproductive cells of the latter. It is at any rate theoretically conceivable that a struggle might arise between the mortal and immortal individuals of a certain Metazoan species, and that natural selection might secure the success of the former, because the longer the immortal individuals lived, the more defective they became, and as a result gave rise to weaker offspring in diminished numbers. Probably no one would be bold enough to suggest such a crude example of natural selection. And yet I venture to think that the principle of natural selection is here also to be taken into account, and even plays, although in a negative rather than a positive way, a very essential part in determining the duration of life in the Metazoa.

When the somatic cells of the first Heteroplastides ceased to be immortal, such a loss would not in any way have precluded them from regaining this condition. Just as, with the differentiation of the first somatic cells of the lowest Heteroplastides, their duration was limited to that of a single cell-generation,—so it must have been possible for them, at a later period and if the necessity arose, to lengthen their duration over two, three, or more generations. And if my theory of the duration of life in the Metazoa is well founded, these cells have as a matter of fact increased their duration, to an extent about equal to that of the organism to which they belong. There is no ground whatever for the assumption that it is impossible to fix the number of cell-generations at infinity,—as actually happens in the case of the reproductive cells,—but on the other hand it has already been shown to be obvious that such an extension is opposed to the principle of utility. It could never be to the advantage of a species to produce crippled individuals, and therefore the infinite duration of individuals has never reappeared among the Metazoa. So far the limited duration of Metazoan life may be attributed to the worthlessness or even the injurious nature of individuals, which although immortal, were nevertheless liable to wear and tear. This fact explains why immortality has never reappeared, it explains the predominance of death, but it was not the single primary cause of this phenomenon. The perishable and vulnerable nature of the soma was the reason why nature made no effort to endow this part of the individual with a life of unlimited length.

Götte considers that death is inherent in reproduction, and in a certain sense this is true, but not in the general way supposed by him.

I have endeavoured to show above that it is most advantageous for the preservation of the species among the lowest Metazoa, that the body should consist of a relatively small number of cells, and that the reproductive cells should ripen simultaneously and all escape together. If this conclusion be accepted, the uselessness of a prolonged life to the somatic cells is obvious, and the occurrence of death at the time of the extrusion of the reproductive cells is explained. In this manner death (of the soma) and reproduction are here made to coincide.

This relation of reproduction to death still exists in a great number of the higher animals. But such an association, together with the simultaneous ripening of the reproductive cells, has not been maintained continuously in the past. As the soma becomes larger and more highly organized, it is able to withstand more injuries, and its average duration of life will extend: pari passu with these changes it will become increasingly advantageous not only for the number of reproductive cells to be multiplied, but also for the time during which they are produced to be prolonged. In this manner a lengthening of the reproductive period arises, at first continuously and then periodically. It is beyond my present purpose to consider in detail the conditions upon which this lengthening depends, but I would emphasize the fact that a lengthening of life is connected with the increase in the duration of reproduction, while on the other hand there is no reason to expect life to be prolonged beyond the reproductive period; so that the end of this period is usually more or less coincident with death.

A further prolongation of life could only take place when the parent begins to undertake the duty of rearing the young. The most primitive form of this is found among those animals, which do not expel their reproductive cells as soon as they are ripe but retain them within their bodies, so that the early stages of development take place under the shelter of the parent organism. Associated with such a process there is frequently a necessity for the germs to reach a certain spot, where alone their further development can take place. Thus a segment of a tapeworm lives until it has brought the embryos into a position which affords the possibility of their passive transference to the stomach of their special host. But the duration of life is first materially lengthened when the offspring begin to be really tended, and as a general rule the increase in length is exactly proportional to the time which is demanded by the care of the young. Accurately conducted observations are wanting upon this precise point, but the general tendency of the facts, as a whole, cannot be doubted. Those insects of which the care for their offspring terminates with the deposition of eggs at the appropriate time, place, etc., do not survive this act; and the duration of life in such imagos is shorter or longer according as the eggs are laid simultaneously or ripen gradually. On the other hand, insects—such as bees and ants—which tend their young, have a life which is prolonged for years.

But the lengthening of the reproductive period alone may result in a marked increase in the length of life, as is proved by the queen-bee. In all these cases it is easy to imagine the operation of natural selection in producing such alterations in the duration of life, and indeed we might accurately calculate the amount of increase which would be produced in any given case if the necessary data were available, viz. the physiological strength of the body, and its relations to the external world, such as, for instance, the power of obtaining food at various periods of life, the expenditure of energy necessary for this end, and the statistics of destruction, that is, the probabilities in favour of the accidental death of a single individual at any given time. These statistics must be known both for the imagos, larvae, and eggs; for the lower they are for the imagos, and the higher for the larvae and eggs, the more advantageous will it be, ceteris paribus, for the number of eggs produced by the imago to be increased, and the more probable it would therefore be that a long reproductive period, involving a lengthening of the life of the imago, would be introduced. But we are still far from being able to apply mathematics to the phenomena of life; the factors are too numerous, and no attempt has been made as yet to determine them with accuracy.

But we must at least admit the principle that both the lengthening and shortening of life are possible by means of natural selection, and that this process is alone able to render intelligible the exact adaptation of the length of life to the conditions of existence.

A shortening of the normal duration of life is also possible; this is shown in every case of sudden death, after the deposition of the whole of the eggs at a single time. This occurs among certain insects, while nearly allied forms of which the oviposition lasts over many days therefore possess a correspondingly long imago-life. The Ephemeridae and Lepidoptera afford many examples of this, and in an earlier work I have collected some of them^[92]. The humming-bird hawk-moth flies about for weeks laying an egg here and there, and, like the allied poplar hawk-moth and lime hawk-moth, probably dies when it has deposited all the eggs which can be matured with the amount of nutriment at its disposal. Many other Lepidoptera, such as the majority of butterflies, fly about for weeks depositing their eggs, but others, such as the emperor-moths and lappet-moths, lay their eggs one after another and then die. The eggs of the parthenogenetic Psychidae are laid directly after the imago has left the cocoon, and death ensues immediately, so that the whole life of the imago only lasts for a few hours. No one could look upon this brief life as a primitive arrangement among Lepidoptera, any more than we do upon the absence of wings in the female Psychidae; shortening of life here is therefore clearly explicable.

In such cases have we any right to speak of the fatal effect of reproduction? We may certainly say that these insects die of exhaustion; their vital strength is used up in the last effort of laying eggs, and in the case of the males, in the act of copulation. Reproduction is here certainly the most apparent cause of death, but a more remote and deeper cause is to be found in the limitation of vital strength to the length and the necessary duties of the reproductive period. The fact that there are female Lepidoptera which, like the emperor-moths, do not feed in the imago-state, proves the truth of this statement. They still possess a mouth and a complete alimentary canal, but they have no spiral ‘tongue,’ and do not take food of any kind, not even a drop of water. They live in a torpid condition for days or weeks until fertilization is accomplished, and then they lay their eggs and die. The habit of extracting honey from flowers—common to most hawk-moths and butterflies—would not have thus fallen into disuse, if the store of nutriment, accumulated in the form of the fat-bodies, during the life of the caterpillar, had not been exactly sufficient to maintain life until the completion of oviposition. The fact that the habit of taking food has been thus abandoned is a proof that the duration of life beyond the reproductive period would not be to the advantage of the species.

The protraction of existence into old age among the higher Metazoa proves that death is not a necessary consequence of reproduction. It seems to me that Götte’s statement ‘that the appearances of senility must not be regarded as the general cause of death’ is not in opposition to my opinions but rather to those which receive general acceptance. I have myself pointed out that ‘death is not always preceded by senility or a period of old age^[93].’

The materials are wanting for a comprehensive investigation of the causes which first introduced this period among the higher Metazoa; in fact the most fundamental data are absent, for we do not even know the part of the animal kingdom in which it first appeared: we cannot even state the amount by which the duration of life exceeds that of the period of reproduction, or what is the value to the species of this last stage in the life of the individual.

It is in these general directions that we must seek for the significance of old age. It is obviously of use to man, for it enables the old to care for their children, and is also advantageous in enabling the older individuals to participate in human affairs and to exercise an influence upon the advancement of intellectual powers, and thus to influence indirectly the maintenance of the race. But as soon as we descend a step lower, if only as far as the apes, accurate facts are wanting, for we are, and shall probably long be, ignorant of the total duration of their life, and the point at which the period of reproduction ceases.

I must here break off in the midst of these considerations, rather than conclude them, for much still remains to be said. I hope, nevertheless, that I have thrown new light upon some important points, and I now propose to conclude with the following short abstract of the results of my enquiry.

I. Natural death occurs only among multicellular beings; it is not found among unicellular organisms. The process of encystment in the latter is in no way comparable with death.

II. Natural death first appears among the lowest Heteroplastid Metazoa, in the limitation of all the cells collectively to one generation, and of the somatic or body-cells proper to a restricted period: the somatic cells afterwards in the higher Metazoa came to last several and even many generations, and life was lengthened to a corresponding degree.

III. This limitation went hand in hand with a differentiation of the cells of the organism into reproductive and somatic cells, in accordance with the principle of division of labour. This differentiation took place by the operation of natural selection.

IV. The fundamental biogenetic law applies only to multicellular beings; it does not apply to unicellular forms of life. This depends on the one hand upon the mode of reproduction by fission which obtains among the Monoplastides (unicellular organisms), and on the other upon the necessity, induced by sexual reproduction, for the maintenance of a unicellular stage in the development of the Polyplastides (multicellular organisms).

V. Death itself, and the longer or shorter duration of life, both depend entirely on adaptation. Death is not an essential attribute of living matter; it is neither necessarily associated with reproduction, nor a necessary consequence of it.

In conclusion, I should wish to call attention to an idea which is rather implied than expressed in this essay:—it is, that reproduction did not first make its appearance coincidently with death. Reproduction is in truth an essential attribute of living matter, just as is the growth which gives rise to it. It is as impossible to imagine life enduring without reproduction as it would be to conceive life lasting without the capacity for absorption of food and without the power of metabolism. Life is continuous and not periodically interrupted: ever since its first appearance upon the earth, in the lowest organisms, it has continued without break; the forms in which it is manifested have alone undergone change. Every individual alive to-day—even the very highest—is to be derived in an unbroken line from the first and lowest forms.