Chapter XII: , which, according to Kowalewsky and others, give origin to the (1)

free leucocytes.

We see, then, that the conception of a syncytial neuro-epithelial host holding in its meshes a number of free cells leads directly to the questions: What is the coelom? To which category does its lining membrane belong? and further, also, What is the origin of these free cells?

The Metazoa have been divided into two great groups--those which possess a coelom (the Coelomata; Lankester's Coelomocoela) and those which do not (Coelenterata; Lankester's Enterocoela). As an example of the latter we may take Hydra, because it is a very {473}primitive form, and because its development has been carefully worked out recently by Brauer.

In Hydra we find a dermal layer of cells and an inner layer of cells separated by a gelatinous mass known as mesogloea; in this mass between the dermal and inner layers scattered cells are found, the interstitial cells. Now, according to Brauer the position of the germ in Hydra is the interstitial cell-layer. One cell of the ovarium becomes the egg-cell, the others have their substance changed into yolk-grains, forming the so-called pseudo-cells, and as such afford pabulum to the growing egg-cell. Thus we see that in between the dermal and gastral layer of cells a third layer of cells is found, composed of free living germ-cells, some of which, by the formation of yolk-granules, become degraded into pabulum for their more favoured kinsfolk. These interstitial cells are said to arise from the dermal layer, or ectoderm, but clearly, as in other cases, germ-cells constitute a class by themselves and cannot be spoken of as originating from ectoderm-cells or from hypoderm-cells.

So also in Porifera, Minchin states: "In addition to the collared cells of the gastral layer, and the various cell-elements of the dermal layer, the body-wall contains numerous wandering cells or amoebocytes, which occur everywhere among the cells and tissues. Though lodged principally in the dermal layer, they are not to be regarded as belonging to it, but as constituting a distinct class of cells by themselves. They are concerned probably with the functions of nutrition and excretion, and from them arise the genital products." Further (p. 31): "At certain seasons some of these cells become germ-cells; hence the wandering cells and the reproductive cells may be included together under the general term archæocytes." Also (p. 51): "The mesogloea is the first portion to appear as a structureless layer between the dermal and gastral epithelia, and is probably a secretion of the former."

He also points out that in these, the very lowest of the Metazoa, the separate origin of these archæocytes can be traced back to a very early period of embryonic life. Thus in _Clathrina blanca_ the ovum undergoes a regular and total cleavage, resulting in the formation of a hollow ciliated blastula of oval form. At one point, the future posterior pole of the larva, are a pair of very large granular cells with vesicular nuclei, which represent undifferentiated blastomeres and are destined to give rise to the archæocytes, and, therefore, also to the {474}sexual cells of the adult. Thus, as he says, from the very earliest period a distinction is made between the "tissue-forming" cells (my syncytial host) and the archæocytes.

We see, then, that the origin of all these free-living cells can be traced back to the very earliest of the Metazoa. Here between the dermal and gastral layers a gelatinous material, the mesogloea is secreted by these layers. This material is non-living, non-cellular. In it live free cells which may either be germ-cells, amoebocytes, or 'collencytes' (connective tissue cells). If this mesogloea were a fluid secretion, then we should have a tissue of the nature of blood or lymph; if it were solid, then we should have the foundation of connective tissue, cartilage, and bone.

From this primitive tissue it is easy to see how the special elements of the vascular, lymphatic, and skeletal tissues gradually arose, the matrix being provided by the cells of the syncytial host and the cellular elements by the archæocytes. In fact, we have no right to speak of these lowest members of the Metazoa as not being triploblastic, as possessing nothing corresponding to mesoblast, for in these free cells in the mesogloea we have the origin of the mesenchyme of the higher groups. Thus Lankester, talking of mesenchyme, says: "I think we are bound to bring into consideration here the existence in many Coelentera of a tissue resembling the mesenchyme of Coelomocoela. In Scyphomedusæ, in Ctenophora, and in Anthozoa, branched fixed and wandering cells are found in the mesogloea which seem to be the same thing as a good deal of what is distinguished as mesenchyme in Coelomocoela. These appear to be derived from both the primitive layers; some produce spicules, others fibrous substance, others again seem to be amoebocytes with various functions. It appears to be probable that, though it may be necessary to distinguish other elements in it, the mesenchyme of Coelomocoela is largely constituted by cells, which are the mother-cells of the skeletotrophic group of tissues, and are destined to form connective tissues, blood-vessels, and blood."

Thus we see that the earliest Metazoa were composed of a dermal and gastral epithelium, with a sub-epithelial nervous system connecting the parts together, which formed, as it were, a host, carrying around free living cells of varying function, all of which may be looked on as derived from archæocytes, _i.e._ germ-cells. From these the coelomatous animals arose, and here also we find, according to {475}present-day opinion, that the coelom arose in the first place in the very closest connection with the germ-cells or gonads. Thus Lankester, in his review of the history of the coelom, states:--

"The numerous embryological and anatomical researches of the past twenty years seem to me to definitely establish the conclusion that the coelom is primarily the cavity, from the walls of which the gonad cells (ova or spermata) develop, or which forms around those cells. We may suppose the first coelom to have originated by a closing or shutting off of that portion of the general archenteron of Enterocoela (Coelentera), in which the gonads developed as in Aurelia or as in Ctenophora. Or we may suppose that groups of gonad mother cells, having proliferated from the endoderm, took up a position between it and the ectoderm, and there acquired a vesicular arrangement, the cells surrounding the cavity in which liquid accumulated.

"The coelom is thus essentially and primarily (as first clearly formulated by Hatschek) the perigonadial cavity or gonocoel, and the lining cells of gonadial chambers are coelomic epithelium. In some few groups of Coelomocoela the coeloms have remained small and limited to the character of gonocoels. This seems to be the case in the Nemertina, the Planarians, and other Platyhelmia. In some Planarians they are limited in number, and of individually large size; in others they are numerous."

When Lankester says that "the lining cells of gonadial chambers are coelomic epithelium," that is equivalent to saying that the lining cells of the coelom form an epithelium which was originally gonadial, provided that, as seems to me most probable, his second suggestion, of the coelom being formed from gonadial mother-cells which have taken up an intermediate position between endoderm and ectoderm and there acquired a vesicular arrangement, is the true one. It does not seem to me possible to conceive of the gonads arising from cells of the epiblast or of the hypoblast, in the sense that such cells are differentiated cells belonging to a layer with a definite meaning. When we consider that the gonad gives origin to the whole of a new individual, that in the protozoan ancestors of the Metazoa their ultimate aim and object was the formation of gonads, it seems a wrong conception to speak of the gonads as formed from cells belonging either to the gut-wall or to the external epithelium. The gonads must stand in a category by themselves; they represent a whole, {476}while the other cells represent only a part; they cannot therefore be derived from the latter. They may, and indeed do, give rise to cells of a subordinate character, but they cannot rightly be spoken of as derived from such cells. The very fact mentioned by Lankester, that in the lowest coelomatous Metazoa, the Platyhelminthes, the coeloms are limited to the character of simple gonocoels, strongly points to the conclusion that all the coelomic cells were originally of the nature of gonadial cells, and therefore free-living and independent of the rest of the cells of the body. Whether the germ-cells appear, as in Hydra, to be derived from the ectoblast, or, as is usually stated, from the endoblast, in neither case ought they to be classed with the internal or external epithelium; they are germ-cells, and the epithelium which they form is neither epiblastic nor hypoblastic, but germinal, forming originally a simple gonocoele, afterwards, in the higher forms, the coelom with its cells of various function. Thus, to quote again from Lankester, "The coelomic fluid and the coelomic epithelium, as well as the floating corpuscles derived from that epithelium, acquire special properties and importance over and above the original functions subservient to the maturation of the gonadial cells ... the most important developments of the coelom are in connection with the establishment of an exit for the generative products through the body-wall to the outer world, and further in the specialization of parts of its lining epithelium for renal excretory functions."

Such exits led very early to the formation of coelomoducts, which are true outgrowths of the coelom itself (p. 14): "The coelomoducts and the gonocoels of which they are a part, frequently acquire a renal excretory function, and may retain both the function of genital conduits and of renal organs, or may, where several pairs are present (metamerized or segmented animals), subserve the one function in some segments of the body, and the other function in other segments."

The origin of the coelom and its derivatives from a germinal membrane, as suggested by Lankester, appears to me most probable, and, if true, it carries with it conclusions of far-reaching importance, for it necessitates that all the cells which line true coelomic cavities, and their derivatives, belong to the category of free-living cells, and are not connected with the nervous system. The cells in question are essentially those which line serous cavities and those which form excretory glands such as the kidneys. In the latter organ we ought especially to be able to obtain a clear answer to this question, for is {477}it not a gland which secretes into a duct and might therefore be expected to be innervated in the same way as other secretory glands? Although there is a strong _primâ facie_ presumption in favour of the existence of renal secretory nerves, yet according to the universal opinion of physiologists no evidence in favour of such nerves has hitherto been found; all the phenomena of excretion of urine consequent on nerve stimulation are explicable by the action of nerves on the renal vessels, not on the renal cells. Not only is the physiological evidence negative up to the present time, but also, I think, the histological. On the one hand, Retzius has failed to find nerve-connections with kidney-cells; on the other, Berkley has obtained such evidence with the Golgi method, but failed entirely with methylene blue. I do not myself think that the evidence of the Golgi method alone is sufficient without corroboration by other methods, and, in any case, Berkley's evidence does not show the nerve-fibres terminating in the kidney-cells, in the same way as can be shown by modern methods to exist in the case of epithelial cells of the surface, etc. Quite recently another paper on this subject has appeared by Smirnow, who appears to have obtained better results than those given by Berkley.

Apart from these physiological and histological considerations, this question is also dependent upon the nature of the development of the excretory organs, for, according to Lankester, all excretory organs may be divided into the two classes of nephridial organs and coelomostomes, of which the former are largely derived from epiblast. We should, therefore, expect to find secretory nerves to nephridial organs, though possibly not to coelomostomes. The kidneys of the Mammalia are supposed to be true coelomostomes, although, according to Goodrich's researches, the excretory organs in Amphioxus are solenocytes, _i.e._ true nephridia.

As to the lining epithelium of the peritoneal, pleural, and pericardial cavities--_i.e._ the mesothelium--there is no definite evidence that these cells are provided with nerves. Such surfaces are remarkably insensitive in the healthy condition, and the pain in such cavities is essentially a pressure phenomenon and referable to special sense-organs, such as Pacinian bodies, etc., rather than to the mesothelium itself.

These sense-organs are identical in structure with those in the skin, and, as Anderson has shown, the nerves of these organs {478}medullate at the same time as those in the skin, and both obtain their medullary sheaths earlier than any other nerves, whether afferent or efferent. However difficult it may be to explain this fact, only one conclusion seems to me possible--these Pacinian bodies, like the skin Pacinians, originate from a nest of surface epithelial cells, a conclusion which is extremely probable on my theory of the origin of vertebrates, but not, as far as I can see, on any other.

At the present moment the weight of evidence is, to my mind, in favour of the lining endothelium of the coelomic cavities being composed of free cells, unconnected with the nervous system rather than the reverse, but I must confess that the question is undecided. If it be true that the coelomic lining is partly enterocoelic and partly gonocoelic, as Lankester teaches, then it would be natural that its cells should be in connection with the nervous system, to some extent at all events. This view is, however, based on very slender foundations. If the mesothelium is composed of cells capable of becoming free, it cannot give rise to the skeletal muscles, and it cannot therefore be right to speak of the skeletal muscles as derived from the lining cells of a part of the primary coelom. The phylogenetic history of the musculature of the different animals points strongly to its intimate connection with and derivation from surface epithelial cells rather than from coelomic mesothelial cells. Thus in the coelenterates, as seen in Hydra, the muscular layer arises directly from a modification of the surface epithelial cells; and right up to the annelids, even to the highest form in the Polychæta, we still see it stated that the musculature, both circular and longitudinal, arises from the ectoderm. In the Oligochæta and Hirudinea, according to Bergh, there are five rows of teloblasts on each side, of which four are ectodermic and give rise to the nerve-ganglia and the circular muscles, while one is mesoblastic and forms the nephridial organs and the longitudinal muscles. (The latter statement is, according to Bergh, well known, and is not particularly shown by him. These longitudinal muscle-bands always lie close against the nervous system at their first formation, and may well have been derived in connection with it.)

It is apparently only in the Vertebrata that the lining cells of the coelomic cavity are definitely stated to give origin to the body-musculature, and taking into account on the one hand the evidence of Graham Kerr as to the intimate connection between nerve-cell and {479}muscle-cell from the very beginning, and on the other the manner in which all the skeletal muscles of the adult are lined with a lymphatic endothelium, I am strongly inclined to believe that at the closing up of the myocoele, when the myomere separates from the mesomere, the lining cells remain scattered in among the forming muscle-cells and form the ultimate lymphatic tissue of the muscles. If this is really so, then the evidence in favour of the mesothelium being composed of free cells not connected with the nervous system would be much strengthened, for, on the one hand, an intimate relation exists between the connective tissue cells and the endothelium of the roots of the lymphatic vessels, a relation which, according to Virchow, has rendered it impossible to draw any sharp line of distinction between the two; and, on the other, the lymphatic endothelium merges into the lining cells of the great serous cavities of the body.

It is impossible to conceive of an animal possessing a nervous system which is not in connection with sensory and muscular tissues; an isolated nerve-cell is a meaningless possession; but it is equally natural to conceive of a germ-cell being isolated, capable of living an independent existence. Such a difference between the two kinds of tissues must have existed from the very commencement of the Metazoa, so that we must, it seems to me, imagine that in the formation of the Metazoa from the Protozoa the whole of the body of the latter did not break up into a mass of separate gonads, each capable of becoming a free-living protozoan similar to its parent, but that a portion proliferated into a multinucleated syncytium while the remainder formed the free-living gonads. This multinucleated syncytium, or host, as it might be called, would still continue to exist for the purpose of carrying further afield the immortal gonads, which need no longer be all shed at one time.

In such an animal as _Volvox globator_ we have an indication of the very kind of animal postulated as connecting the single-celled Protozoa and the multi-cellular Metazoa, for it consists of a many-celled case which forms a hollow sphere, each of the cells being provided with flagella for the purpose of locomotion of the sphere, except a certain number which are not flagellated; the latter leave the case to swim freely in the fluid contained within the sphere, and forming spermaries and ovaries, conjugate, maturate, and then are set free by the rupture of the encircling locomotor host.

{480}This conception of the predecessors of the Metazoa being composed of a mortal host, holding within itself the immortal sexual products, leads naturally to the idea of the separate development of the host from that of the germ-cells _ab initio_, so that the study of the development of the Metazoa means the study of two separate constituents of the metazoan individual--on the one hand, the elaboration of the elements forming the syncytial host, on the other, of those derived from the free-living independent germ-cells. The elaboration of the host means the differentiation of the protoplasm into epithelial, muscular, and nervous elements, by means of which the gonads were carried further afield and their nourishment as well as that of the host ensured.

The _rôle_ of the nervous system as the middleman between internal and external muscular and epithelial surfaces was, I imagine, initiated from the very earliest time. The further evolution of the host consisted in a greater and greater differentiation and elaboration of this neuro-epithelial syncytium, with the result of a steadily increasing concentration and departmental centralization of the main factor of the syncytium; in other words, it led to the origin and elaboration of a central nervous system. In the interstices of this syncytium the gonads were placed, and at first, doubtless, the life of the host ended when all the germ-cells had been set free. 'Reproduce and die' was, I imagine, the law of the Metazoa at its earliest origin, and throughout the ages, during all the changes of evolution, the reminiscence of such law still manifests itself even up to the highest forms as yet reached. With the differentiation of the syncytial host there came also differentiation of the free-living gonads, so that only some of them attained to the perfection of independent existence, capable of continuing the species; while others became subordinate to the first and provided them with pabulum, manufacturing within themselves yolk-spherules, and thus in the shape of yolk-cells ministered to the developing egg-cell. Thus arose a germinal epithelium of which only a few of the elements passed out of the host as perfect individuals, the remainder being utilized for the nutrition of these few. Such yolk-cells of the germinal epithelium would still, however, retain their character as free cells totally independent of the syncytial host, and, situated as they were between the internal and external epithelium, capable of amoeboid movement, would naturally have their phagocytic action {481}utilized either as yolk-cells for the providing of pabulum to the egg-cell, or as excretory cells for the removal and rendering harmless of deleterious products of all kinds. Thus the free cells of the body would become differentiated into the three classes of germ-cells, yolk-cells, and excretory cells.

Further, the mass of gonads, which originally occupied so large a space within the interior of the host, necessarily, as the tissues of the host differentiated more and more, took up less and less space in proportion to the whole bulk of the host and formed a germinal mass of cells between the outer and inner epithelial layers. This germinal mass formed an epithelium, some of the members of which acted as scavengers for the inner and outer layers of the host, with the result that fluid accumulated between the two parts of the germinal epithelium in connection respectively with the external and internal epithelial surfaces of the host, and thus led to the formation of a gonocoele, which, by obtaining an external opening, a coelomostome, gave origin to the coelom.

Again, with the longer life of the host, the setting free of the gonads no longer necessitating the destruction of the host, and also the gonads themselves requiring a longer and longer time to be fed up to maturity, the bulk and complexity of the whole organism increased and special supporting structures became a necessity. The host itself could and did provide these to a certain extent by secretions from its epithelial elements, but the intermediate supports were provided by the system of phagocytic cells utilizing the fluids of the body, at first in the shape of plasma-cells able to move from place to place, then settling down to form a connective tissue framework, and, later on, cartilage and bone.

So also were gradually evolved the whole of the endothelial structures; the lymph-cells, blood-cells, etc., all having their origin from the free cells of the body, which themselves originated in the extension of a germinal epithelium. Just as in a bee-hive the egg-cells may form the fully developed sexual animal, whether drone or queen bee, or the asexual host of workers, so in the body of the Metazoa the free cells may form either male or female germ-cells spermatozoa, or ova, or a host of workers, scavengers, repairers, food-providers, all useful to the community, all showing their common origin by their absolute independence of the nervous system.

Two points of great importance follow from this method of looking {482}at the problem. First, the evolution of the animal kingdom means essentially the evolution of the host, for that is what forms the individual; secondly, as the host is composed of a syncytium, the common factor of whose elements is the neural moiety, it follows that the tissue of central importance for the evolution of the host must be, as indeed it is, the nervous system. Further, seeing that the growth of the individual means the orderly spreading out of the epithelial moiety away from the neural moiety, it follows that the germ-band or germ-area from which growth starts must be in the position of the nervous system. If then, the nervous system in the animal is a concentrated one, then the growth will emanate from the position of such nervous system. If, on the other hand, the nervous system is diffused, then the growth will also be diffused.

In this book I have throughout argued that the ancestors of vertebrates belonged to a great group of animals which gave origin also to Limulus and scorpion-like animals; it is therefore instructive to see what is the nature of the development of such animals. For this purpose I will take the development of the scorpion, as given by Brauer, for he has worked out its development with great thoroughness and care. His papers show that the segmentation is discoidal, and results in an oval blastodermic area lying on a large mass of yolk. Very early there separates out in this area genital cells and yolk-cells, which latter move freely into the yolk and prepare it into a fluid pabulum for the nutrition of the cells of the embryonic shield or germ-band. These free yolk-cells do not take part in the formation of the germinal layers, nor does the endoderm when formed give origin to free yolk-cells.

The cells of the germ-band form a small compact area, in which by continual mitosis the cells become more than one-layered, and soon it is found that those cells which lie close against the fluid pabulum form a continuous layer and absorb the nutritious material for themselves and the rest of the embryo. While this area is thus increasing in thickness by continuous development, the group of genital cells remains always apart, increasing in number, but being always in a state of isolation from the cells of the rest of the growing area. Thus from the very first Brauer's observations on the development of the scorpion point to the formation of a syncytial host containing separate genital cells. The continuous layer of cells against the fluid pabulum, which is already functioning as a gut, and may {483}therefore be called hypoblast, spreads continuously over the yolk, as also does the surface epithelial layer, or epiblast. Such spreading is always a continuous one for both surfaces, so that the yolk is gradually enclosed by a continuous orderly growth from the germ-band, and not by the settling down of free cells in the yolk here and there to form the gut-lining. This steady orderly development proceeds owing to the nourishment afforded by the activity of the free cells or vitellophags and the absorbing power of the hypoblast, a steady growth round the yolk which results in the formation of the gut-tube, the outer covering and all the muscular and excretory organs. Where, then, is this starting-point, this germ-band from which the whole embryo grows? It forms the mid ventral area of the adult animal, it corresponds exactly to the position of the central nervous system. The whole phenomenon of embryonic growth in the scorpion is exactly what must take place on the argument deduced from the study of the adult that the animal arises as a neuro-epithelial syncytium, and we see that that layer of cells which is situated next to the food-material forms the alimentary tube. It is not a question whether such layer is ventral or dorsal to the neural cells, but whether it is contiguous to or removed from the food-material.

Take, again, a meroblastic vertebrate egg as of the bird. Again we find free cells passing into the yolk to act as vitellophags, the so-called periblast cells; again we see that the embryo starts from a germ-band or embryonic shield, and spreads from there continuously and steadily; again we see that the layer of cells which lies against the yolk absorbs the fluid pabulum for the growing cells; again we see that the area from which the whole process of growth starts is that of the central nervous system, and again we see that those cells which are contiguous to the food form the commencing gut, and are therefore called hypoblast, though in this case they are ventral not dorsal to the neural layer.

The comparison of these two processes shows that there is one common factor, one thing comparable in the two, one thing that is homologous and is the essential in the formation of that part of the animal which I have called the host, and that is the central nervous system. Whether the epithelial layer which lies ventrally to it or the one that is dorsal forms the gut depends upon the position of the food-mass. Where the food is, there will be the absorbing layer. {484}Where the food is not, there will be no gut formation, whatever may have been the previous history of that layer. If, then, we suppose, as I do, that the vertebrate arose from a scorpion-like animal without any reversal of dorsal and ventral surfaces, and that the central nervous system remained the same in the two animals, then the comparison of the development of the two embryos shows that the one would be derived from the other if the yolk-mass shifted from the dorsal to the ventral side of the nervous system. This would leave the dorsal epithelial layer of the original syncytium free from pabulum; it would no longer form the definite gut, _but it would still tend to form itself in the same manner as before, would still grow from a ventrally situated germ-band dorsalwards to form a tube, would recapitulate its past history, and show how the alimentary canal of the arthropod became the neural canal of the vertebrate_. Although this alimentary canal is formed in the same way as before, it is no longer recognized as homologous with the scorpion's alimentary canal, but because it no longer absorbs pabulum, and does not therefore form the definite gut, it is called an epiblastic tube, and, in the words of Ray Lankester, has no developmental importance.

All the arthropods are built up on the same type, and in all the development may in its broad outlines be referred to the type just mentioned. So also with the vertebrate group; in both cases the position of the central nervous system determines the starting area of embryonic growth. In both cases the absorbing layer shows the position of the definite gut. A concentrated nervous system of this type is common to all the segmented animals from the annelids to the vertebrates, and in all cases the germ-band which indicates the first formation of the embryo is in the position of this nervous system.

As far as the embryo is concerned, there is no great difficulty in the conception that the yolk-mass may have shifted from one side to the other in passing from the arthropod to the vertebrate, for in the arthropod the embryo at first is surrounded by yolk and then passes to the periphery of the egg. If it is permissible to speak of a dorsal and ventral surface to an egg, and we may imagine the egg held with such dorsal surface uppermost, then the yolk would be situated ventrally to the embryo, as in the vertebrate, if the protoplasmic cells of the embryo rose from their central position to the surface through the yolk, while if they sank through the yolk, the yolk would be situated dorsally to the embryo, as in the arthropod.

{485}In cases where there is no yolk, or very little, as in Lucifer and Amphioxus respectively, the embryo is compelled to feed itself at a very early age; such embryos form a free-swimming pelagic ciliated blastula, the invagination of which, for the purpose of collecting food material out of the open sea, is the simplest method of obtaining nutriment. Here, as in other cases, it is the physiological necessity which determines the method of formation of the gut, and such similarity of appearance as exists between the gastrula of Lucifer and that of Amphioxus, by no means implies that the gut of the adult Lucifer is homologous with the gut of Amphioxus.

I have compared two meroblastic eggs of the two classes respectively, because the scorpion's egg is meroblastic. I imagine that no real difficulty arises with respect to holoblastic eggs, for the experiments of O. Hertwig and Samassa show that by centrifugalizing, stimulating, and breaking down of large spheres the holoblastic amphibian egg may be converted into a meroblastic one, and then development will proceed regularly, _i.e._ in this case also the growth proceeds from the animal pole; the large cells of the vegetal pole, like the yolk-cells of the meroblastic egg, manufacture pabulum for the growing syncytial host.

SUMMARY.

Any attempt to discover how vertebrates arose from invertebrates must be
based upon the study of Comparative Anatomy, of Palæontology, and of
Embryology. The arguments and evidence put forward in the preceding
chapters show most clearly how the theory of the origin of vertebrates
from palæostracans is supported by the geological evidence, by the
anatomical evidence, and by the embryological evidence. Of the three the
latter is the strongest and most conclusive, if it be taken to include
the evidence given by the larval stage of the lamprey.

The stronghold of embryology for questions of this sort is the Law of
Recapitulation, which asserts that the history of the race is
recapitulated to a greater or less extent in the development of the
individual. In the previous chapters such recapitulation has been shown
for all the organs of the vertebrate body. In this respect, then,
embryology has proved of the greatest value in confirming the evidence of
relationship between the palæostracan and the vertebrate, given by
anatomical and geological study.

There is, however, another side to embryology, which claims that the
tissues of all the Metazoa are built up on the same plan; that in all
cases in the very early stage of the embryo three layers are formed, the
epiblast, mesoblast, and hypoblast; that in all animals above the
Protozoa these three layers are {486}homologous, the epiblast in all
cases forming the external or skin-layer, the hypoblast the internal or
gut-layer.

Such a theory, therefore, as is advocated in this book, which turns the
gut of the arthropod into the neural canal of the vertebrate, and makes a
new gut for the vertebrate from the external surface must be wrong, as it
flatly contradicts the fundamental germ-layer theory.

Of recent years grave doubts have been thrown upon the validity of this
theory, doubts which have increased in force year by year as more and
more facts have been discovered which are not in agreement with the
theory. So much is it now discredited that any criticism against my
theory, which is based upon it, weighs nothing in the balance against the
positive evidence of recapitulation already stated. If the germ-layer
theory is no longer credited, upon what fundamental laws is embryology
based?

In this chapter I have ventured to suggest a reply to this question,
based on the uniformity of the laws of growth throughout the existence of
the individual.

In the adult animal the body is composed of two kinds of tissues, those
which are connected with or at all events are under the control of the
nervous system, and those which are capable of leading a free life
independent of the nervous system. These two kinds of tissues can be
traced back from the adult to the embryo, and it is the task of
embryology to find out how these two kinds of tissue originate.

The following out of this line of thought leads to the conception that,
throughout the Metazoa, the body is composed of a host which consists of
the master-tissues of the body, and takes the form of a neuro-epithelial
syncytium, within the meshes of which free living independent organisms
or cells live, so to speak, a symbiotic existence.

The evidence points to the origin of all these free cells from
germ-cells, and thus leads to the conception that the blastula stage of
every embryo represents two kinds of cells, the one which will form the
mortal host being the locomotor neuro-epithelial cell, the other the
independent immortal symbiotic germ-cell. Such conception leads directly
to the conclusion that the blastula stage of every member of the Metazoa
is the embryonic representation of a Protozoan ancestor of the Metazoa;
an ancestor, whose nature may be illustrated by such a living form as
_Volvox globator_, which, like a blastula, is composed of a layer of
cells forming a hollow sphere. These cells partly bear cilia, and so form
a locomotor host, partly are of a different character, and form male and
female germ-cells. The latter leave the surface of the sphere, pass as
free individuals into its fluid contents, form spermaries and ovaries,
and then by the rupture of the mortal locomotor host pass out into the
external medium, as free swimming young Volvox.

It is of interest to note that such members of the Protozoa are among the
most highly developed of the members of this great group.

From such a beginning arose in orderly evolution, on the one hand, all
the neuro-muscular and neuro-epithelial structures of the body--the
so-called master-tissues; on the other, the germ-cells, the
blood-corpuscles, lymph-corpuscles plasma and excretory cells, connective
tissue cells, cartilage and bone-cells, etc., all of them independent of
the central nervous system, all traceable to a modification of the
original germ-cells.

{487}Such a view of the processes of embryology brings embryology into
harmony with comparative anatomy and phylogeny, for it makes the central
nervous system and not the alimentary canal the most important factor in
the development of the host.

The growth of the individual, whether arthropod or vertebrate, spreads
from the position of the central nervous system, regardless of whether
that position is a ventral or dorsal one with respect to the yolk-mass.
Where the pabulum is, there is the definite gut, the lining walls of
which are called in the embryo, hypoblast; but when the pabulum is no
longer there, although a tube is formed in the same manner as the
alimentary canal of the arthropod, it is now called an epiblastic tube,
and is known as the neural tube of the vertebrate.

This is the great fallacy of the germ-layer theory, a fallacy which
consists of an argument in a vicious circle: thus the alimentary canal is
homologous in all of the Metazoa, because it is formed of hypoblast, but
there is no definition of hypoblast, except that it is always that layer
which forms the definitive alimentary canal.

When, after the process of segmentation has been completed, a free
swimming blastula results, unprovided with any store of pabulum in the
shape of yolk, then the same physiological necessity causes such a form
to obtain its nutriment from the surrounding medium. The simplest way to
do this is by a process of invagination, in consequence of which food
particles are swept into the invaginated part and then absorbed. For this
reason in such cases true gastrulæ are formed, as in the case of
Amphioxus among the vertebrates, and Lucifer among the crustaceans; such
a formation does not in the least imply that the gut of the arthropod is
homologous with that of the vertebrate. The resemblance between the two
is not a morphological one, but due to the same physiological necessity.
They are analogous formations, not homologous.

The muscular tissues are found to be formed in close connection with the
nervous tissues, and in very many cases are described as formed from
epiblast, so that there are strong reasons for placing them in a special
category of the so-called mesoblastic tissues. If they be separated out,
then it seems to me, the rest of the mesoblast would consist of the
free-living cells of the body, which are not connected with the central
nervous system. In watching, then, the formation of mesoblast, defined in
this way, we are watching the separation out from the master-tissues of
the body of the independent skeletal and excretory cells.

{488}CHAPTER XV

_FINAL REMARKS_

Problems requiring investigation--

Giant nerve-cells and giant-fibres; their comparison in fishes and in
arthropods; blood- and lymph-corpuscles; nature of the skin; origin of
system of unstriped muscles; origin of the sympathetic nervous system;
biological test of relationship.

Criticism of Balanoglossus theory.--Theory of parallel
development.--Importance of the theory advocated in this book for all
problems of Evolution.

The discussion in the last chapter on the "Principles of Embryology" completes the evidence which I am able to offer up to the present time in favour of my theory of the "Origin of Vertebrates." There are various questions which I have left untouched, but still are well worth discussion, and may be mentioned here. The first of these is the significance of the giant nerve-cells and giant nerve-fibres so characteristic of the brain-region of the lower vertebrates. In most fishes two very large cells are most conspicuous objects in any transverse section of the _medulla oblongata_ at the level of entrance of the auditory nerves. Each of these cells gives off a number of processes, some of which pass in the direction of the auditory nerves and one very large axis-cylinder process which forms a giant-fibre, known by the name of a Mauthnerian fibre. Each Mauthnerian fibre crosses the middle line soon after its origin from the giant-cell, and passes down the spinal cord on the opposite side right to the tail. Here, near the end of the spinal cord, it breaks up into smaller fibres, which are believed by Fritsch and others to pass out directly into the ventral roots to supply the muscles of the tail. Thus Bela Haller says: "The Mauthnerian fibres are known to give origin to certain fibres which supply the ventral roots of the last three spinal nerves, so that their terminal branches serve, in all probability, for the innervation of the muscles of the tail-fin." They do not occur in the eel, according to Haller, or in Silurus, according to Kölliker. {489}Their absence in those fishes, in which a well-developed tail-fin is also absent, increases the probability of the truth of Fritsch's original conclusion that these giant-fibres are associated axis-cylinders for certain definite co-ordinated movements of the fish, especially for the lateral movement of the tail.

In Ammocoetes, instead of two Mauthnerian fibres, a number of giant-fibres are found. They are called Müllerian fibres, and arise from giant-cells which are divisible into two groups. The first group consists of three pairs situated headwards of the level of exit of the trigeminal nerves. Two of these lie in front of the level of exit of the oculomotor nerves, and one pair is situated at the same level as the origin of the oculomotor nerves. The second group consists of a number of cells on each side at the level of the entrance of the fibres of the auditory nerves.

The Müllerian fibres largely decussate, as described by Ahlborn, and then become the most anterior portion of the white matter of the spinal cord, forming a group of about eight fibres on each side (Fig. 73). A few fibres are also found laterally, and slightly dorsally, to the grey matter. These giant-fibres pass down the spinal cord right to the anal region; their ultimate destination is unknown. Mayer considers that in the first part of their course they correspond to those tracts of fibres known as the "posterior longitudinal bundles" in other vertebrates. I imagine, therefore, that the spinal part of their course represents the two antero-lateral descending tracts. The second group of giant-cells, which appears to have some connection with the auditory nerves, may represent "Deiter's nucleus." The whole system is probably the central nervous part of a co-ordination mechanism, which arises entirely in the pro-otic or prosomatic region of the brain--the great co-ordinating and equilibrating region _par excellence_.

If we turn now to the arthropod it is a striking coincidence that in the crayfish and in the lobster the work of Retzius, of Celesia, of Allen, and of many others demonstrates the existence of an equilibration-mechanism for the swimming movements of the tail-muscles, which is carried out by means of giant-fibres. These giant-fibres are the axis-cylinder processes of giant-cells, situated exclusively in the brain-region, and they run through the whole ventral ganglionic chain in order to supply the muscles of the tail. In the ventral nerve-cord of the crayfish, according to Retzius, two specially large {490}giant-fibres exist, each of which breaks up, at the last abdominal ganglion, into smaller fibres, which pass directly out with the nerves to the tail-fin. Allen has shown that, in addition to these two specially large giant-fibres, there are a number of others, some of which, similarly to the Müllerian fibres of Ammocoetes, cross the middle line, while some do not. Each of these arises from a large nerve-cell and passes to one or other of the last pair of abdominal ganglia. The latter fibres, he says, send off collaterals, while the two specially large giant-fibres do not. The cells which give origin to all these large, long fibres are situated in or in front of the prosomatic region of the brain, similarly to the giant-cells, which give rise to the corresponding Müllerian fibres of Ammocoetes. I do not know how far this system is represented in Limulus or Scorpio.

It is, to my mind, improbable that the Mauthnerian fibres pass out directly as motor fibres to the muscles of the tail-fin; it is more likely that they are conducting paths between the equilibration-mechanism in connection with the VIIIth nerve and the spinal centres for the movements of the tail. Similarly, with respect to the arthropod, it is difficult to believe that the motor fibres for the tail-muscles arise in the brain-region. In either case, the striking coincidence remains that the movements of the tail-end of the body are regulated by means of giant-fibres which arise from giant-cells in the head-region of the body in both the Arthropoda and the lowest members of the Vertebrata.

The meaning of this system of giant-cells and giant-fibres in both classes of animals is well worthy of further investigation.

Another important piece of comparative work which ought to help in the elucidation of this problem is the comparison of the blood- and lymph-corpuscles of the vertebrate with those of the invertebrate groups. As yet, I have not myself made any observations in this direction, and feel that it is inadvisable to discuss the results of others until I know more about the facts from personal observation.

The large and important question of the manner of formation of the vertebrate skin has only been considered to a slight extent. A much more thorough investigation requires to be made into the nature of the skin of the oldest fishes in comparison with the skin of Ammocoetes on the one side, and of Limulus and the Palæostraca on the other.

The muscular system requires further investigation, not so much {491}the different systems of the striated voluntary musculature--for these have been for the most part compared in the two groups of animals in previous chapters--as the involuntary unstriped musculature, about which no word has been said. The origin of the different systems of unstriped muscles in the vertebrate is bound up with the origin of the sympathetic system and its relation to the cranial and sacral visceral systems. The reason why I have not included in this book the consideration of the sympathetic nervous system is on account of the difficulty in finding any such system in Ammocoetes. Also, so far as I know, the distribution of unstriped muscle in Ammocoetes has not been worked out.

One clue has arisen quite recently which is of great importance, and must be worked out in the future, viz. the extraordinary connection which exists between the action of the sympathetic nervous system and the action of adrenalin. This substance, which is obtained from the medullary part of the adrenal or suprarenal glands, when injected into an animal produces the same effects as stimulation of the nerves, which belong to the lumbo-thoracic outflow of visceral nerves, _i.e._ the system known as the sympathetic nervous system, which is distinct from both the cranial and sacral outflows of visceral nerves. The similarity of its action to stimulation of nerves is entirely confined to the nerves of this sympathetic system, and never resembles that of either the cranial or sacral visceral nerves.

Another most striking fact which confirms the great importance of this connection between the adrenals and the sympathetic nervous system from the point of view of the evolution of the latter system is that the extract of the adrenals always produces the same effect as that of stimulation of the nerves of the sympathetic system, whatever may be the animal from which the extract is obtained. Thus adrenalin obtained from the elasmobranch fishes will produce in the highest mammal all the effects known to occur upon stimulation of the nerves of its sympathetic system.

Further, the cells, which are always associated with the presence of this peculiar substance--adrenalin--stain in a characteristic manner in the presence of chromic salts. In Ammocoetes patches of cells which stain in this manner have been described in connection with blood-vessels in certain parts, so that, although I know of no definite evidence of the existence of cell-groups in Ammocoetes corresponding to the ganglia of the sympathetic system in other vertebrates, it is {492}possible that further investigation into the nature and connection of these "chromaffine" cells may afford a clue to the origin of the sympathetic nervous system. At present it is premature to discuss the question further.

Finally, another test as to the kinship of two animals of different species must be considered more fully than I have been able to do up to the present time. This test is of a totally different nature to any put forth in previous pages. It is known as the "biological test" of relationship, and is the outcome of pathological rather than of physiological or anatomical research. It is possible that this test may prove the most valuable of all. At present we do not know sufficiently its limitations and its sources of error, especially in the case of cold-blooded animals, to be able to look upon it as decisive in a problem of the kind considered in this book.

The nature of this test is as follows: It has been found that the serum of the blood of another animal, when injected in sufficient quantity into a rabbit, will cause such a change in the serum of that rabbit's blood that when it is added to the serum of the other animal a copious precipitate is formed, although the serum of normal rabbit's blood when mixed with that of another animal will cause no precipitate whatever. This extraordinary production of a precipitate in the one case and not in the other indicates the production of some new substance in the rabbit's serum in consequence of the introduction of the foreign serum into the rabbit, which brings about a precipitate when the rabbit's serum containing it is mixed with the serum originally injected. The barbarous name "antibody" has been used to express this supposed substance in accordance with the meaning of such a word as "antitoxin," which has been a long time in use in connection with preventive remedies against pathogenic bacteria. Now, it is found that the rabbit's serum containing a particular "antibody" will cause a precipitate only when added to the serum of the blood of the animal from which the "antibody" was produced or to the serum of the blood of a nearly related animal.

Further, if that animal is closely related a precipitate will be formed nearly as copious as with the original serum, if more distantly related a cloudiness will occur rather than a precipitate, and if the relationship is still more distant the mixture of the two sera will remain absolutely clear. Thus this test demonstrates the close relationship of man to the anthropoid apes and his more distant {493}relationship to monkeys in general. By this method very evident blood-relationships have been demonstrated, especially between members of the Mammalia.

I therefore started upon an investigation into the possibility of proving relationship in this way between Limulus and Ammocoetes, with the kind assistance of Mr. Graham Smith. I must confess I was not sanguine of success, as I thought the distance between Limulus and Ammocoetes was too great. Dr. Lee, of New York, kindly provided me with most excellent serum of Limulus, and the first experiments showed that the anti-serum of Limulus gave a most powerful precipitate with its own serum. Graham Smith then tried this anti-serum of Limulus with the serum of Ammocoetes, and to his surprise, and mine, he obtained a distinct cloudiness, indicative of a relationship between the two animals. This, however, is not considered sufficient, the reverse experiment must also succeed. I therefore, with Graham Smith, obtained a considerable amount of blood from the adult lampreys at Brandon, and produced an anti-serum of Petromyzon, which gave some precipitate with its own serum, but not a very powerful one. This anti-serum tried with Limulus gave no result whatever, but at the same time it gave no result with serum from Ammocoetes, so that the experiment not only showed that Petromyzon was not related to Limulus, but also was not related to its own larval form, which is absurd.

Considerable difficulties were encountered in preparing the Petromyzon anti-serum owing to the extreme toxic character of the lamprey's serum to the rabbit; in this respect it resembled that of the eel. It is possible that the failure of the lamprey's anti-serum was due to the necessity of heating the serum sufficiently to do away with its toxicity before injecting it into the rabbit. At this point the experiments have been at present left. It will require a long and careful investigation before it is possible to speak decisively one way or the other. At present the experiment is positive to a certain extent, and also negative; but the latter proves too much, for it proves that the larva is not related to the adult.

Some day I hope this "biological test" will be of use for determining the relationships of the Tunicata, the Enteropneusta, Amphioxus, etc., as well as of Limulus and Ammocoetes.