[20] Cf. Balfour’s Manual of Comparative Embryology, vol. ii; Theodore Morgan’s The Development of the Frog’s Egg.

The remarkable processes of gastrulation, ovum-segmentation, and formation of germinal layers present a most conspicuous variety. There is to-day only the lowest of the vertebrates, the amphioxus, that exhibits the original form of those processes, or the palingenetic gastrulation which we have considered in the preceding chapter, and which culminates in the formation of the archigastrula (Fig. 38). In all other extant vertebrates these fundamental processes have been more or less modified by adaptation to the conditions of embryonic development (especially by changes in the food-yelk); they exhibit various cenogenetic types of the formation of germinal layers. However, the different classes vary considerably from each other. In order to grasp the unity that underlies the manifold differences in these phenomena and their historical connection, it is necessary to bear in mind always the unity of the vertebrate-stem. This “phylogenetic unity,” which I developed in my General Morphology in 1866, is now generally admitted. All impartial zoologists agree to-day that all the vertebrates, from the amphioxus and the fishes to the ape and man, descend from a common ancestor, “the primitive vertebrate.” Hence the embryonic processes, by which each individual vertebrate is developed, must also be capable of being reduced to one common type of embryonic development; and this primitive type is most certainly exhibited to-day by the amphioxus.

It must, therefore, be our next task to make a comparative study of the various forms of vertebrate gastrulation, and trace them backwards to that of the lancelet. Broadly speaking, they fall first into two groups: the older cyclostoma, the earliest fishes, most of the amphibia, and the viviparous mammals, have holoblastic ova—that is to say, ova with total, unequal segmentation; while the younger cyclostoma, most of the fishes, the cephalopods, reptiles, birds, and monotremes, have meroblastic ova, or ova with partial discoid segmentation. A closer study of them shows, however, that these two groups do not present a natural unity, and that the historical relations between their several divisions are very complicated. In order to understand them properly, we must first consider the various modifications of gastrulation in these classes. We may begin with that of the amphibia.

The most suitable and most available objects of study in this class are the eggs of our indigenous amphibia, the tailless frogs and toads, and the tailed salamander. In spring they are to be found in clusters in every pond, and careful examination of the ova with a lens is sufficient to show at least the external features of the segmentation. In order to understand the whole process rightly and follow the formation of the germinal layers and the gastrula, the ova of the frog and salamander must be carefully hardened; then the thinnest possible sections must be made of the hardened ova with the microtome, and the tinted sections must be very closely compared under a powerful microscope.

The ova of the frog or toad are globular in shape, about the twelfth of an inch in diameter, and are clustered in jelly-like masses, which are lumped together in the case of the frog, but form long strings in the case of the toad. When we examine the opaque, grey, brown, or blackish ova closely, we find that the upper half is darker than the lower. The middle of the upper half is in many species black, while the middle of the lower half is white.^[21] In this way we get a definite axis of the ovum with two poles. To give a clear idea of the segmentation of this ovum, it is best to compare it with a globe, on the surface of which are marked the various parallels of longitude and latitude. The superficial dividing lines between the different cells, which come from the repeated segmentation of the ovum, look like deep furrows on the surface, and hence the whole process has been given the name of furcation. In reality, however, this “furcation,” which was formerly regarded as a very mysterious process, is nothing but the familiar, repeated cell-segmentation. Hence also the segmentation-cells which result from it are real cells.

[21] The colouring of the eggs of the amphibia is caused by the accumulation of dark-colouring matter at the animal pole of the ovum. In consequence of this, the animal cells of the ectoderm are darker than the vegetal cells of the entoderm. We find the reverse of this in the case of most animals, the protoplasm of the entoderm cells being usually darker and coarser-grained.

The unequal segmentation which we observe in the ovum of the amphibia has the special feature of beginning at the upper and darker pole (the north pole of the terrestrial globe in our illustration), and slowly advancing towards the lower and brighter pole (the south pole). Also the upper and darker hemisphere remains in this position throughout the course of the segmentation, and its cells multiply much more briskly. Hence the cells of the lower hemisphere are found to be larger and less numerous. The cleavage of the stem-cell (Fig. 40 A) begins with the formation of a complete furrow, which starts from the north pole and reaches to the south (B). An hour later a second furrow arises in the same way, and this cuts the first at a right angle (Fig. 40 C). The ovum is thus divided into four equal parts. Each of these four “segmentation cells” has an upper and darker and a lower, brighter half. A few hours later a third furrow appears, vertically to the first two (Fig. 40 D). The globular germ now consists of eight cells, four smaller ones above (northern) and four larger ones below (southern). Next, each of the four upper ones divides into two halves by a cleavage beginning from the north pole, so that we now have eight above and four below (Fig. 40 E). Later, the four new longitudinal divisions extend gradually to the lower cells, and the number rises from twelve to sixteen (F). Then a second circular furrow appears, parallel to the first, and nearer to the north pole, so that we may compare it to the north polar circle. In this way we get twenty-four segmentation-cells—sixteen upper, smaller, and darker ones, and eight smaller and brighter ones below (G). Soon, however, the latter also sub-divide into sixteen, a third or “meridian of latitude” appearing, this time in the southern hemisphere: this makes thirty-two cells altogether (H). Then eight new longitudinal lines are formed at the north pole, and these proceed to divide, first the darker cells above and afterwards the lighter southern cells, and finally reach the south pole. In this way we get in succession forty, forty-eight, fifty-six, and at last sixty-four cells (I, K). In the meantime, the two hemispheres differ more and more from each other. Whereas the sluggish lower hemisphere long remains at thirty-two cells, the lively northern hemisphere briskly sub-divides twice, producing first sixty-four and then 128 cells (L, M). Thus we reach a stage in which we count on the surface of the ovum 128 small cells in the upper half and thirty-two large ones in the lower half, or 160 altogether. The dissimilarity of the two halves increases: while the northern breaks up into a great number of small cells, the southern consists of a much smaller number of larger cells. Finally, the dark cells of the upper half grow almost over the surface of the ovum, leaving only a small circular spot

at the south pole, where the large and clear cells of the lower half are visible. This white region at the south pole corresponds, as we shall see afterwards, to the primitive mouth of the gastrula. The whole mass of the inner and larger and clearer cells (including the white polar region) belongs to the entoderm or ventral layer. The outer envelope of dark smaller cells forms the ectoderm or skin-layer.

In the meantime, a large cavity, full of fluid, has been formed within the globular body—the segmentation-cavity or embryonic cavity (blastocoel, Figs. 41–44 F). It extends considerably as the cleavage proceeds, and afterwards assumes an almost semi-circular form (Fig. 41 F). The frog-embryo now represents a modified embryonic vesicle or blastula, with hollow animal half and solid vegetal half.

Now a second, narrower but longer, cavity arises by a process of folding at the lower pole, and by the falling away from each other of the white entoderm-cells (Figs. 41–44 N). This is the primitive gut-cavity or the gastric cavity of the gastrula, progaster or archenteron. It was first observed in the ovum of the amphibia by Rusconi, and so called the Rusconian cavity. The reason of its peculiar narrowness here is that it is, for the most part, full of yelk-cells of the entoderm. These also stop up the whole of the wide opening of the primitive mouth, and form what is known as the “yelk-stopper,” which is seen freely at the white round spot at the south pole (P). Around it the ectoderm is much thicker, and forms the border of the primitive mouth, the most important part of the embryo (Fig. 44 k, k). Soon the primitive gut-cavity stretches further and further at the expense of the segmentation-cavity (F), until at last the latter disappears altogether. The two cavities are only separated by a thin partition (Fig. 43 s). With the formation of the primitive gut our frog-embryo has reached the gastrula stage, though it is clear that this cenogenetic amphibian gastrula is very different from the real palingenetic gastrula we have considered (Figs. 30–36).

In the growth of this hooded gastrula we cannot sharply mark off the various stages which we distinguish successively in the bell-gastrula as morula and gastrula. Nevertheless, it is not difficult to reduce the whole cenogenetic or disturbed development of this amphigastrula to the true palingenetic formation of the archigastrula of the amphioxus.

This reduction becomes easier if, after considering the gastrulation of the tailless amphibia (frogs and toads), we glance for a moment at that of the tailed amphibia, the salamanders. In some of the latter, that have only recently been carefully studied, and that are phylogenetically older, the process is much simpler and clearer than is the case with the former and longer known. Our common water-salamander (Triton taeniatus) is a particularly good subject for observation. Its nutritive yelk is much smaller and its formative yelk less obscured with black pigment-cells than in the case of the frog; and its gastrulation has better retained the original palingenetic character. It was first described by Scott and Osborn (1879), and Oscar Hertwig especially made a careful study of it (1881), and rightly pointed out its great importance in helping us to understand the vertebrate development. Its globular blastula (Fig. 45) consists of loosely-aggregated, yelk-filled entodermic cells or yelk-cells (dz) in the lower vegetal half; the upper, animal half encloses the hemispherical segmentation-cavity (fh), the curved roof of which is formed of two or three strata of small ectodermic cells. At the point where the latter pass into the former (at the equator of the globular vesicle) we have the border zone (rz). The folding which leads to the formation of the gastrula takes place at a spot in this border zone, the primitive mouth (Fig. 46 u).

Unequal segmentation takes place in some of the cyclostoma and in the oldest fishes in just the same way as in most of the amphibia. Among the cyclostoma (“round-mouthed”) the familiar lampreys are particularly interesting. In respect of organisation and development they are half-way between the acrania (lancelet) and the lowest real fishes (Selachii); hence I divided the group of the cyclostoma in 1886 from the real fishes with which they were formerly associated, and formed of them a special class of vertebrates. The ovum-segmentation in our common river-lamprey (Petromyzon fluviatilis) was described by Max Schultze in 1856, and afterwards by Scott (1882) and Goette (1890).

Unequal total segmentation follows the same lines in the oldest fishes, the selachii and ganoids, which are directly descended from the cyclostoma. The primitive fishes (Selachii), which we must regard as the ancestral group of the true fishes, were generally considered, until a short time ago, to be discoblastic. It was not until the beginning of the twentieth century that Bashford Dean made the important discovery in Japan that one of the oldest living fishes of the shark type (Cestracion japonicus) has the same total unequal segmentation as the amphiblastic plated fishes (ganoides).^[22] This is particularly interesting in connection with our subject, because the few remaining survivors of this division, which was so numerous in paleozoic times, exhibit three different types of gastrulation.

[22] Bashford Dean, Holoblastic Cleavage in the Egg of a Shark, Cestracion japonicus Macleay. Annotationes zoologicae japonenses, vol. iv, Tokio, 1901.

The oldest and most conservative forms of the modern ganoids are the scaly sturgeons (Sturiones), plated fishes of great evolutionary importance, the eggs of which are eaten as caviar; their cleavage is not essentially different from that of the lampreys and the amphibia. On the other hand, the most modern of the plated fishes, the beautifully scaled bony pike of the North American rivers (Lepidosteus), approaches the osseous fishes, and is discoblastic like them. A third genus (Amia) is midway between the sturgeons and the latter.

The group of the lung-fishes (Dipneusta or Dipnoi) is closely connected with the older ganoids. In respect of their whole organisation they are midway between the gill-breathing fishes and the lung-breathing amphibia; they share with the former the shape of the body and limbs, and with the latter the form of the heart and lungs. Of the older dipnoi (Paladipneusta) we have now only one specimen, the remarkable Ceratodus of East Australia; its amphiblastic gastrulation has been recently explained by Richard Semon (cf. Chapter XXI). That of the two modern dipneusta, of which Protopterus is found in Africa and Lepidosiren in America, is not materially different. (Cf. Fig. 51.)

All these amphiblastic vertebrates, Petromyzon and Cestracion, Accipenser and Ceratodus, and also the salamanders and batrachia, belong to the old, conservative groups of our stem. Their unequal ovum-segmentation and gastrulation have many peculiarities in detail, but can always be reduced with comparative ease to the original cleavage and gastrulation of the lowest vertebrate, the amphioxus; and this is little removed, as we have seen, from the very simple archigastrula of the Sagitta and Monoxenia (see Fig. 29–36). All these and many other classes of animals generally agree in the circumstance that in segmentation their ovum divides into a large number of cells by repeated cleavage. All such ova have been called, after Remak, “whole-cleaving” (holoblasta), because their division into cells is complete or total.

In a great many other classes of animals this is not the case, as we find (in the vertebrate stem) among the birds, reptiles, and most of the fishes; among the insects and most of the spiders and crabs (of the articulates); and the cephalopods (of the molluscs). In all these animals the mature ovum, and the stem-cell that arises from it in fertilisation, consist of two different and separate parts, which we have called formative yelk and nutritive yelk. The formative yelk alone consists of living protoplasm, and is the active, evolutionary, and nucleated part of the ovum; this alone divides in segmentation, and produces the numerous cells which make up the embryo. On the other hand, the nutritive yelk is merely a passive part of the contents of the ovum, a subordinate element which contains nutritive material (albumin, fat, etc.), and so represents in a sense the provision-store of the developing embryo. The latter takes a quantity of food out of this store, and finally consumes it all. Hence the nutritive yelk is of great indirect importance in embryonic development, though it has no direct share in it. It either does not divide at all, or only later on, and does not generally consist of cells. It is sometimes large and sometimes small, but generally many times larger than the formative yelk; and hence it is that it was formerly thought the more important of the two. As the respective significance of these two parts of the ovum is often wrongly described, it must be borne in mind that the nutritive yelk is only a secondary addition to the primary cell, it is an inner enclosure, not an external appendage. All ova that have this independent nutritive yelk are called, after Remak, “partially-cleaving” (meroblasta). Their segmentation is incomplete or partial.

There are many difficulties in the way of understanding this partial segmentation and the gastrula that arises from it. We have only recently succeeded, by means of comparative research, in overcoming these difficulties, and reducing this cenogenetic form of gastrulation to the original palingenetic type. This is comparatively easy in the small meroblastic ova which contain little nutritive yelk—for instance, in the marine ova of a bony fish, the development of which I observed in 1875 at Ajaccio in Corsica. I found them joined together in lumps of jelly, floating on the surface of the sea; and, as the little ovula were completely transparent, I could easily follow the development of the germ step by step. These ovula are glossy and colourless globules of little more than the 50th of an inch. Inside a structureless, thin, but firm membrane (ovolemma, Fig. 52 c) we find a large, quite clear, and transparent globule of albumin (d). At both poles of its axis this globule has a pit-like depression. In the pit at the upper, animal pole (which is turned downwards in the floating ovum) there is a bi-convex lens composed of protoplasm, and this encloses the nucleus (k); this is the formative yelk of the stem-cell, or the germinal disk (b). The small fat-globule (f) and the large albumin-globule (d) together form the nutritive yelk. Only the formative yelk undergoes cleavage, the nutritive yelk not dividing at all at first.

The segmentation of the lens-shaped formative yelk (b) proceeds quite independently of the nutritive yelk, and in perfect geometrical order.

When the mulberry-like cluster of cells has been formed, the border-cells of the lens separate from the rest and travel into the yelk and the border-layer. From this the blastula is developed; the regular bi-convex lens being converted into a disk, like a watch-glass, with thick borders. This lies on the upper and less curved polar surface of the nutritive yelk like the watch glass on the yelk. Fluid gathers between the outer layer and the border, and the segmentation-cavity is formed. The gastrula is then formed by invagination, or a kind of turning-up of the edge of the blastoderm. In this process the segmentation-cavity disappears.

The space underneath the entoderm corresponds to the primitive gut-cavity, and is filled with the decreasing food-yelk (n). Thus the formation of the gastrula of our fish is complete. In contrast to the two chief forms of gastrula we considered previously, we give the name of discoid gastrula (discogastrula, Fig. 54) to this third principal type.

Very similar to the discoid gastrulation of the bony fishes is that of the hags or myxinoida, the remarkable cyclostomes that live parasitically in the body-cavity of fishes, and are distinguished by several notable peculiarities from their nearest relatives, the lampreys. While the amphiblastic ova of the latter are small and develop like those of the amphibia, the cucumber-shaped ova of the hag are about an inch long, and form a discoid gastrula. Up to the present it has only been observed in one species (Bdellostoma Stouti), by Dean and Doflein (1898).

It is clear that the important features which distinguish the discoid gastrula from the other chief forms we have considered are determined by the large food-yelk. This takes no direct part in the building of the germinal layers, and completely fills the primitive gut-cavity of the gastrula, even protruding at the mouth-opening. If we imagine the original bell-gastrula (Figs. 30–36) trying to swallow a ball of food which is much bigger than itself, it would spread out round it in discoid shape in the attempt, just as we find to be the case here (Fig. 54). Hence we may derive the discoid gastrula from the original bell-gastrula, through the intermediate stage of the hooded gastrula. It has arisen through the accumulation of a store of food-stuff at the vegetal pole, a “nutritive yelk” being thus formed in contrast to the “formative yelk.” Nevertheless, the gastrula is formed here, as in the previous cases, by the folding or invagination of the blastula. We can, therefore, reduce this cenogenetic form of the discoid segmentation to the palingenetic form of the primitive cleavage.

This reduction is tolerably easy and confident in the case of the small ovum of our deep-sea bony fish, but it becomes difficult and uncertain in the case of the large ova that we find in the majority of the other fishes and in all the reptiles and birds. In these cases the food-yelk is, in the first place, comparatively colossal, the formative yelk being almost invisible beside it; and, in the second place, the food-yelk contains a quantity of different elements, which are known as “yelk-granules, yelk-globules, yelk-plates, yelk-flakes, yelk-vesicles,” and so on. Frequently these definite elements in the yelk have been described as real cells, and it has been wrongly stated that a portion of the embryonic body is built up from these cells. This is by no means the case. In every case, however large it is—and even when cell-nuclei travel into it during the cleavage of the border—the nutritive yelk remains a dead accumulation of food, which is taken into the gut during embryonic development and consumed by the embryo. The latter develops solely from the living formative yelk of the stem-cell. This is equally true of the ova of our small bony fishes and of the colossal ova of the primitive fishes, reptiles, and birds.

The gastrulation of the primitive fishes or selachii (sharks and rays) has been carefully studied of late years by Ruckert, Rabl, and H.E. Ziegler in particular, and is very important in the sense that this group is the oldest among living fishes, and their gastrulation can be derived directly from that of the cyclostoma by the accumulation of a large quantity of food-yelk. The oldest sharks (Cestracion) still have the unequal segmentation inherited from the cyclostoma. But while in this case, as in the case of the amphibia, the small ovum completely divides into cells in segmentation, this is no longer so in the great majority of the selachii (or Elasmobranchii). In these the contractility of the active protoplasm no longer suffices to break up the huge mass of the passive deutoplasm completely into cells; this is only possible in the upper or dorsal part, but not in the lower or ventral section. Hence we find in the primitive fishes a blastula with a small eccentric segmentation-cavity (Fig. 55 b), the wall of which varies greatly in composition. The circular border of the germinal disk which connects the roof and floor of the segmentation-cavity corresponds to the border-zone at the equator of the amphibian ovum. In the middle of its hinder border we have the beginning of the invagination of the primitive gut (Fig. 56 ud); it extends gradually from this spot (which corresponds to the Rusconian anus of the amphibia) forward and around, so that the primitive mouth becomes first crescent-shaped and then circular, and, as it opens wider, surrounds the ball of the larger food-yelk.

Essentially different from the wide-mouthed discoid gastrula of most of the selachii is the narrow-mouthed discoid gastrula (or epigastrula) of the amniotes, the reptiles, birds, and monotremes; between the two—as an intermediate stage—we have the amphigastrula of the amphibia. The latter has developed from the amphigastrula of the ganoids and dipneusts, whereas the discoid amniote gastrula has been evolved from the amphibian gastrula by the addition of food-yelk. This change of gastrulation is still found in the remarkable ophidia (Gymnophiona, Coecilia, or Peromela), serpent-like amphibia that live in moist soil in the tropics, and in many respects represent the transition from the gill-breathing amphibia to the lung-breathing reptiles. Their embryonic development has been explained by the fine studies of the brothers Sarasin of Ichthyophis glutinosa at Ceylon (1887), and those of August Brauer of the Hypogeophis rostrata in the Seychelles (1897). It is only by the historical and comparative study of these that we can understand the difficult and obscure gastrulation of the amniotes.

The bird’s egg is particularly important for our purpose, because most of the chief studies of the development of the vertebrates are based on observations of the hen’s egg during hatching. The mammal ovum is much more difficult to obtain and study, and for this practical and obvious reason very rarely thoroughly investigated. But we can get hens’ eggs in any quantity at any time, and, by means of artificial incubation, follow the development of the embryo step by step. The bird’s egg differs considerably from the tiny mammal ovum in size, a large quantity of food-yelk accumulating within the original yelk or the protoplasm of the ovum. This is the yellow ball which we commonly call the yolk of the egg. In order to understand the bird’s egg aright—for it is very often quite wrongly explained—we must examine it in its original condition, and follow it from the very beginning of its development in the bird’s ovary. We then see that the original ovum is a quite small, naked, and simple cell with a nucleus, not differing in either size or shape from the original ovum of the mammals and other animals (cf. Fig. 13 E). As in the case of all the craniota (animals with a skull), the original or primitive ovum (protovum) is covered with a continuous layer of small cells. This membrane is the follicle, from which the ovum afterwards issues. Immediately underneath it the structureless yelk-membrane is secreted from the yelk.

The small primitive ovum of the bird begins very early to take up into itself a quantity of food-stuff through the yelk-membrane, and work it up into the “yellow yelk.” In this way the ovum enters on its second stage (the metovum), which is many times larger than the first, but still only a single enlarged cell. Through the accumulation of the store of yellow yelk within the ball of protoplasm the nucleus it contains (the germinal vesicle) is forced to the surface of the ball. Here it is surrounded by a small quantity of protoplasm, and with this forms the lens-shaped formative yelk (Fig. 15 b). This is seen on the yellow yelk-ball, at a certain point of the surface, as a small round white spot—the “tread” (cicatricula). From this point a thread-like column of white nutritive yelk (d), which contains no yellow yelk-granules, and is softer than the yellow food-yelk, proceeds to the middle of the yellow yelk-ball, and forms there a small central globule of white yelk (Fig. 15 d). The whole of this white yelk is not sharply separated from the yellow yelk, which shows a slight trace of concentric layers in the hard-boiled egg (Fig. 15 c). We also find in the hen’s egg, when we break the shell and take out the yelk, a round small white disk at its surface which corresponds to the tread. But this small white “germinal disk” is now further developed, and is really the gastrula of the chick. The body of the chick is formed from it alone. The whole white and yellow yelk-mass is without any significance for the formation of the embryo, it being merely used as food by the developing chick. The clear, glarous mass of albumin that surrounds the yellow yelk of the bird’s egg, and also the hard chalky shell, are only formed within the oviduct round the impregnated ovum.

When the fertilisation of the bird’s ovum has taken place within the mother’s body, we find in the lens-shaped stem-cell the progress of flat, discoid segmentation (Fig. 57). First two equal segmentation-cells (A) are formed from the ovum. These divide into four (B), then into eight, sixteen (C), thirty-two, sixty-four, and so on. The cleavage of the cells is always preceded by a division of their nuclei. The cleavage surfaces between the segmentation-cells appear at the free surface of the tread as clefts. The first two divisions are vertical to each other, in the form of a cross (B). Then there are two more divisions, which cut the former at an angle of forty-five degrees. The tread, which thus becomes the germinal disk, now has the appearance of an eight-rayed star. A circular cleavage next taking place round the middle, the eight triangular cells divide into sixteen, of which eight are in the middle and eight distributed around (C). Afterwards circular clefts and radial clefts, directed towards the centre, alternate more or less irregularly (D, E). In most of the amniotes the formation of concentric and radial clefts is irregular from the very first; and so also in the hen’s egg. But the final outcome of the cleavage-process is once more the formation of a large number of small cells of a similar nature. As in the case of the fish-ovum, these segmentation-cells form a round, lens-shaped disk, which corresponds to the morula, and is embedded in a small depression of the white yelk. Between the lens-shaped disk of the morula-cells and the underlying white yelk a small cavity is now formed by the accumulation of fluid, as in the fishes. Thus we get the peculiar and not easily recognisable blastula of the bird (Fig. 58). The small segmentation-cavity (fh) is very flat and much compressed. The upper or dorsal wall (dw) is formed of a single layer of clear, distinctly separated cells; this corresponds to the upper or animal hemisphere of the triton-blastula (Fig. 45). The lower or ventral wall of the flat dividing space (vw) is made up of larger and darker segmentation-cells; it corresponds to the lower or vegetal hemisphere of the blastula of the water-salamander (Fig. 45 dz). The nuclei of the yelk-cells, which are in this case especially numerous at the edge of the lens-shaped blastula, travel into the white yelk, increase by cleavage, and contribute even to the further growth of the germinal disk by furnishing it with food-stuff.

The invagination or the folding inwards of the bird-blastula takes place in this case also at the hinder pole of the subsequent chief axis, in the middle of the hind border of the round germinal disk (Fig. 59 s). At this spot we have the most brisk cleavage of the cells; hence the cells are more numerous and smaller here than in the fore-half of the germinal disk. The border-swelling or thick edge of the disk is less clear but whiter behind, and is more sharply separated from contiguous parts. In the middle of its hind border there is a white, crescent-shaped groove—Koller’s sickle-groove (Fig. 59 s); a small projecting process in the centre of it is called the sickle-knob (sk). This important cleft is the primitive mouth, which was described for a long time as the “primitive groove.” If we make a vertical section through this part, we see that a flat and broad cleft stretches under the germinal disk forwards from the primitive mouth; this is the primitive gut (Fig. 60 ud). Its roof or dorsal wall is formed by the folded upper part of the blastula, and its floor or ventral wall by the white yelk (wd), in which a number of yelk-nuclei (dk) are distributed. There is a brisk multiplication of these at the edge of the germinal disk, especially in the neighbourhood of the sickle-shaped primitive mouth.

We learn from sections through later stages of this discoid bird-gastrula that the primitive gut-cavity, extending forward from the primitive mouth as a flat pouch, undermines the whole region of the round flat lens-shaped blastula (Fig. 61 ud). At the same time, the segmentation-cavity gradually disappears altogether, the folded inner germinal layer (ik) placing itself from underneath on the overlying outer germinal layer (ak). The typical process of invagination, though greatly disguised, can thus be clearly seen in this case, as Goette and Rauber, and more recently Duval (Fig. 61), have shown.

The older embryologists (Pander, Baer, Remak), and, in recent times especially, His, KÖlliker, and others, said that the two primary germinal layers of the hen’s ovum—the oldest and most frequent subject of observation!—arose by horizontal cleavage of a simple germinal disk. In opposition to this accepted view, I affirmed in my GastrÆa Theory (1873) that the discoid bird-gastrula, like that of all other vertebrates, is formed by folding (or invagination), and that this typical process is merely altered in a peculiar way and disguised by the immense accumulation of food-yelk and the flat spreading of the discoid blastula at one part of its surface. I endeavoured to establish this view by the derivation of the vertebrates from one source, and especially by proving that the birds descend from the reptiles, and these from the amphibia. If this is correct, the discoid gastrula of the amniotes must have been formed by the folding-in of a hollow blastula, as has been shown by Remak and Rusconi of the discoid gastrula of the amphibia, their direct ancestors. The accurate and extremely careful observations of the authors I have mentioned (Goette, Rauber, and Duval) have decisively proved this recently for the birds; and the same has been done for the reptiles by the fine studies of Kupffer, Beneke, Wenkebach, and others. In the shield-shaped germinal disk of the lizard (Fig. 62), the crocodile, the tortoise, and other reptiles, we find in the middle of the hind border (at the same spot as the sickle groove in the bird) a transverse furrow (u), which leads into a flat, pouch-like, blind sac, the primitive gut. The fore (dorsal) and hind (ventral) lips of the transverse furrow correspond exactly to the lips of the primitive mouth (or sickle-groove) in the birds.

The gastrulation of the mammals must be derived from this special embryonic development of the reptiles and birds. This latest and most advanced class of the vertebrates has, as we shall see afterwards, evolved at a comparatively recent date from an older group of reptiles; and all these amniotes must have come originally from a common stem-form. Hence the distinctive embryonic process of the mammal must have arisen by cenogenetic modifications from the older form of gastrulation of the reptiles and birds. Until we admit this thesis we cannot understand the formation of the germinal layers in the mammal, and therefore in man.

I first advanced this fundamental principle in my essay On the Gastrulation of Mammals (1877), and sought to show in this way that I assumed a gradual degeneration of the food-yelk and the yelk-sac on the way from the proreptiles to the mammals. “The cenogenetic process of adaptation,” I said, “which has occasioned the atrophy of the rudimentary yelk-sac of the mammal, is perfectly clear. It is due to the fact that the young of the mammal, whose ancestors were certainly oviparous, now remain a long time in the womb. As the great store of food-yelk, which the oviparous ancestors gave to the egg, became superfluous in their descendants owing to the long carrying in the womb, and the maternal blood in the wall of the uterus made itself the chief source of nourishment, the now useless yelk-sac was bound to atrophy by embryonic adaptation.”

My opinion met with little approval at the time; it was vehemently attacked by KÖlliker, Hensen, and His in particular. However, it has been gradually accepted, and has recently been firmly established by a large number of excellent studies of mammal gastrulation, especially by Edward Van Beneden’s studies of the rabbit and bat, Selenka’s on the marsupials and rodents, Heape’s and LieberkÜhn’s on the mole, Kupffer and Keibel’s on the rodents, Bonnet’s on the ruminants, etc. From the general comparative point of view, Carl Rabl in his theory of the mesoderm, Oscar Hertwig in the latest edition of his Manual (1902), and Hubrecht in his Studies in Mammalian Embryology (1891), have supported the opinion, and sought to derive the peculiarly modified gastrulation of the mammal from that of the reptile.

In the meantime (1884) the studies of Wilhelm Haacke and Caldwell provided a proof of the long-suspected and very interesting fact, that the lowest mammals, the monotremes, lay eggs, like the birds and reptiles, and are not viviparous like the other mammals. Although the gastrulation of the monotremes was not really known until studied by Richard Semon in 1894, there could be little doubt, in view of the great size of their food-yelk, that their ovum-segmentation was discoid, and led to the formation of a sickle-mouthed discogastrula, as in the case of the reptiles and birds. Hence I had, in 1875 (in my essay on The Gastrula and Ovum-segmentation of Animals), counted the monotremes among the discoblastic vertebrates. This hypothesis was established as a fact nineteen years afterwards by the careful observations of Semon; he gave in the second volume of his great work, Zoological Journeys in Australia (1894), the first description and correct explanation of the discoid gastrulation of the monotremes. The fertilised ova of the two living monotremes (Echidna and Ornithorhynchus) are balls of one-fifth of an inch in diameter, enclosed in a stiff shell; but they grow considerably during development, so that when laid the egg is three times as large. The structure of the plentiful yelk, and especially the relation of the yellow and the white yelk, are just the same as in the reptiles and birds. As with these, partial cleavage takes place at a spot on the surface at which the small formative yelk and the nucleus it encloses are found. First is formed a lens-shaped circular germinal disk. This is made up of several strata of cells, but it spreads over the yelk-ball, and thus becomes a one-layered blastula.

If we then imagine the yelk it contains to be dissolved and replaced by a clear liquid, we have the characteristic blastula of the higher mammals. In these the gastrulation proceeds in two phases, as Semon rightly observes: firstly, formation of the entoderm by cleavage at the centre and further growth at the edge; secondly, invagination. In the monotremes more primitive conditions have been retained better than in the reptiles and birds. In the latter, before the commencement of the gastrula-folding, we have, at least at the periphery, a two-layered embryo forming from the cleavage. But in the monotremes the formation of the cenogenetic entoderm does not precede the invagination; hence in this case the construction of the germinal layers is less modified than in the other amniota.

The marsupials, a second sub-class, come next to the oviparous monotremes, the oldest of the mammals. But as in their case the food-yelk is already atrophied, and the little ovum develops within the mother’s body, the partial cleavage has been reconverted into total. One section of the marsupials still show points of agreement with the monotremes, while another section of them, according to the splendid investigations of Selenka, form a connecting-link between these and the placentals.

The fertilised ovum of the opossum (Didelphys) divides, according to Selenka, first into two, then four, then eight equal cells; hence the segmentation is at first equal or homogeneous. But in the course of the cleavage a larger cell, distinguished by its less clear plasm and its containing more yelk-granules (the mother cell of the entoderm, Fig. 64 en), separates from the others; the latter multiply more rapidly than the former. As, further, a quantity of fluid gathers in the morula, we get a round blastula, the wall of which is of varying thickness, like that of the amphioxus (Fig. 38 E) and the amphibia (Fig. 45). The upper or animal hemisphere is formed of a large number of small cells; the lower or vegetal hemisphere of a small number of large cells. One of the latter, distinguished by its size (Fig. 64 en), lies at the vegetal pole of the blastula-axis, at the point where the primitive mouth afterwards appears. This is the mother-cell of the entoderm; it now begins to multiply by cleavage, and the daughter-cells (Fig. 65 i) spread out from this spot over the inner surface of the blastula, though at first only over the vegetal hemisphere. The less clear entodermic cells (i) are distinguished at first by their rounder shape and darker nuclei from the higher, clearer, and longer entodermic cells (e), afterwards both are greatly flattened, the inner blastodermic cells more than the outer.

The unnucleated yelk-balls and curd (Fig. 65 d) that we find in the fluid of the blastula in these marsupials are very remarkable; they are the relics of the atrophied food-yelk, which was developed in their ancestors, the monotremes, and in the reptiles.

In the further course of the gastrulation of the opossum the oval shape of the gastrula (Fig. 66) gradually changes into globular, a larger quantity of fluid accumulating in the vesicle. At the same time, the entoderm spreads further and further over the inner surface of the ectoderm (e). A globular vesicle is formed, the wall of which consists of two thin simple strata of cells; the cells of the outer germinal layer are rounder, and those of the inner layer flatter. In the region of the primitive mouth (p) the cells are less flattened, and multiply briskly. From this point—from the hind (ventral) lip of the primitive mouth, which extends in a central cleft, the primitive groove—the construction of the mesoderm proceeds.

Gastrulation is still more modified and curtailed cenogenetically in the placentals than in the marsupials. It was first accurately known to us by the distinguished investigations of Edward Van Beneden in 1875, the first object of study being the ovum of the rabbit. But as man also belongs to this sub-class, and as his as yet unstudied gastrulation cannot be materially different from that of the other placentals, it merits the closest attention. We have, in the first place, the peculiar feature that the two first segmentation-cells that proceed from the cleavage of the fertilised ovum (Fig. 68) are of different sizes and natures; the difference is sometimes greater, sometimes less (Fig. 69). One of these first daughter-cells of the ovum is a little larger, clearer, and more transparent than the other. Further, the smaller cell takes a colour in carmine, osmium, etc., more strongly than the larger. By repeated cleavage of it a morula is formed, and from this a blastula, which changes in a very characteristic way into the greatly modified gastrula. When the number of the segmentation-cells in the mammal embryo has reached ninety-six (in the rabbit, about seventy hours after impregnation) the foetus assumes a form very like the archigastrula (Fig. 72). The spherical embryo consists of a central mass of thirty-two soft, round cells with dark nuclei, which are flattened into polygonal shape by mutual pressure, and colour dark-brown with osmic acid (Fig. 72 i). This dark central group of cells is surrounded by a lighter spherical membrane, consisting of sixty-four cube-shaped, small, and fine-grained cells which lie close together in a single stratum, and only colour slightly in osmic acid (Fig. 72 e). The authors who regard this embryonic form as the primary gastrula of the placental conceive the outer layer as the ectoderm and the inner as the entoderm. The entodermic membrane is only interrupted at one spot, one, two, or three of the ectodermic cells being loose there. These form the yelk-stopper, and fill up the mouth of the gastrula (a). The central primitive gut-cavity (d) is full of entodermic cells. The uni-axial type of the mammal gastrula is accentuated in this way. However, opinions still differ considerably as to the real nature of this “provisional gastrula” of the placental and its relation to the blastula into which it is converted.

As the gastrulation proceeds a large spherical blastula is formed from this peculiar solid amphigastrula of the placental, as we saw in the case of the marsupial. The accumulation of fluid in the solid gastrula (Fig. 73 A) leads to the formation of an eccentric cavity, the group of the darker entodermic cells (hy) remaining directly attached at one spot with the round enveloping stratum of the lighter ectodermic cells (ep). This spot corresponds to the original primitive mouth (prostoma or blastoporus). From this important spot the inner germinal layer spreads all round on the inner surface of the outer layer, the cell-stratum of which forms the wall of the hollow sphere; the extension proceeds from the vegetal towards the animal pole.

The cenogenetic gastrulation of the placental has been greatly modified by secondary adaptation in the various groups of this most advanced and youngest sub-class of the mammals. Thus, for instance, we find in many of the rodents (guinea-pigs, mice, etc.) apparently a temporary inversion of the two germinal layers. This is due to a folding of the blastodermic wall by what is called the “girder,” a plug-shaped growth of Rauber’s “roof-layer.” It is a thin layer of flat epithelial cells, that is freed from the surface of the blastoderm in some of the rodents; it has no more significance in connection with the general course of placental gastrulation than the conspicuous departure from the usual globular shape in the blastula of some of the ungulates. In some pigs and ruminants it grows into a thread-like, long and thin tube.

Thus the gastrulation of the placentals, which diverges most from that of the amphioxus, the primitive form, is reduced to the original type, the invagination of a modified blastula. Its chief peculiarity is that the folded part of the blastoderm does not form a completely closed (only open at the primitive mouth) blind sac, as is usual; but this blind sac has a wide opening at the ventral curve (opposite to the dorsal mouth); and through this opening the primitive gut communicates from the first with the embryonic cavity of the blastula. The folded crest-shaped entoderm grows with a free circular border on the inner surface of the entoderm towards the vegetal pole; when it has reached this, and the inner surface of the blastula is completely grown over, the primitive gut is closed. This remarkable direct transition of the primitive gut-cavity into the segmentation-cavity is explained simply by the assumption that in most of the mammals the yelk-mass, which is still possessed by the oldest forms of the class (the monotremes) and their ancestors (the reptiles), is atrophied. This proves the essential unity of gastrulation in all the vertebrates, in spite of the striking differences in the various classes.

In order to complete our consideration of the important processes of segmentation and gastrulation, we will, in conclusion, cast a brief glance at the fourth chief type—superficial segmentation. In the vertebrates this form is not found at all. But it plays the chief part in the large stem of the articulates—the insects, spiders, myriapods, and crabs. The distinctive form of gastrula that comes of it is the “vesicular gastrula” (Perigastrula).

In the ova which undergo this superficial cleavage the formative yelk is sharply divided from the nutritive yelk, as in the preceding cases of the ova of birds, reptiles, fishes, etc.; the formative yelk alone undergoes cleavage. But while in the ova with discoid gastrulation the formative yelk is not in the centre, but at one pole of the uni-axial ovum, and the food-yelk gathered at the other pole, in the ova with superficial cleavage we find the formative yelk spread over the whole surface of the ovum; it encloses spherically the food-yelk, which is accumulated in the middle of the ova. As the segmentation only affects the former and not the latter, it is bound to be entirely “superficial”; the store of food in the middle is quite untouched by it. As a rule, it proceeds in regular geometrical progression. In the end the whole of the formative yelk divides into a number of small and homogeneous cells, which lie close together in a single stratum on the entire surface of the ovum, and form a superficial blastoderm. This blastoderm is a simple, completely closed vesicle, the internal cavity of which is entirely full of food-yelk. This real blastula only differs from that of the primitive ova in its chemical composition. In the latter the content is water or a watery jelly; in the former it is a thick mixture, rich in food-yelk, of albuminous and fatty substances. As this quantity of food-yelk fills the centre of the ovum before cleavage begins, there is no difference in this respect between the morula and the blastula. The two stages rather agree in this.

When the blastula is fully formed, we have again in this case the important folding or invagination that determines gastrulation. The space between the skin-layer and the gut-layer (the remainder of the segmentation-cavity) remains full of food-yelk, which is gradually used up. This is the only material difference between our vesicular gastrula (perigastrula) and the original form of the bell-gastrula (archigastrula). Clearly the one has been developed from the other in the course of time, owing to the accumulation of food-yelk in the centre of the ovum.^[23]

[23] On the reduction of all forms of gastrulation to the original palingenetic form see especially the lucid treatment of the subject in Arnold Lang’s Manual of Comparative Anatomy (1888), Part I.

We must count it an important advance that we are thus in a position to reduce all the various embryonic phenomena in the different groups of animals to these four principal forms of segmentation and gastrulation. Of these four forms we must regard one only as the original palingenetic, and the other three as cenogenetic and derivative. The unequal, the discoid, and the superficial segmentation have all clearly arisen by secondary adaptation from the primary segmentation; and the chief cause of their development has been the gradual formation of the food-yelk, and the increasing antithesis between animal and vegetal halves of the ovum, or between ectoderm (skin-layer) and entoderm (gut-layer).

The numbers of careful studies of animal gastrulation that have been made in the last few decades have completely established the views I have expounded, and which I first advanced in the years 1872–76. For a time they were greatly disputed by many embryologists. Some said that the original embryonic form of the metazoa was not the gastrula, but the “planula”—a double-walled vesicle with closed cavity and without mouth-aperture; the latter was supposed to pierce through gradually. It was afterwards shown that this planula (found in several sponges, etc.) was a later evolution from the gastrula.

It was also shown that what is called delamination—the rise of the two primary germinal layers by the folding of the surface of the blastoderm (for instance, in the GeryonidÆ and other medusÆ)—was a secondary formation, due to cenogenetic variations from the original invagination of the blastula. The same may be said of what is called “immigration,” in which certain cells or groups of cells are detached from the simple layer of the blastoderm, and travel into the interior of the blastula; they attach themselves to the inner wall of the blastula, and form a second internal epithelial layer—that is to say, the entoderm. In these and many other controversies of modern embryology the first requisite for clear and natural explanation is a careful and discriminative distinction between palingenetic (hereditary) and cenogenetic (adaptive) processes. If this is properly attended to, we find evidence everywhere of the biogenetic law.