The three higher classes of vertebrates which we call the amniotes—the mammals, birds, and reptiles—are notably distinguished by a number of peculiarities of their development from the five lower classes of the stem—the animals without an amnion (the anamnia). All the amniotes have a distinctive embryonic membrane known as the amnion (or “water-membrane”), and a special embryonic appendage—the allantois. They have, further, a large yelk-sac, which is filled with food-yelk in the reptiles and birds, and with a corresponding clear fluid in the mammals. In consequence of these later-acquired structures, the original features of the development of the amniotes are so much altered that it is very difficult to reduce them to the palingenetic embryonic processes of the lower amnion-less vertebrates. The gastræa theory shows us how to do this, by representing the embryology of the lowest vertebrate, the skull-less amphioxus, as the original form, and deducing from it, through a series of gradual modifications, the gastrulation and cœlomation of the craniota.

It was somewhat fatal to the true conception of the chief embryonic processes of the vertebrate that all the older embryologists, from Malpighi (1687) and Wolff (1750) to Baer (1828) and Remak (1850), always started from the investigation of the hen’s egg, and transferred to man and the other vertebrates the impressions they gathered from this. This classical object of embryological research is, as we have seen, a source of dangerous errors. The large round food-yelk of the bird’s egg causes, in the first place, a flat discoid expansion of the small gastrula, and then so distinctive a development of this thin round embryonic disk that the controversy as to its significance occupies a large part of embryological literature.

One of the most unfortunate errors that this led to was the idea of an original

[ 116 ]

antithesis of germ and yelk. The latter was regarded as a foreign body, extrinsic to the real germ, whereas it is properly a part of it, an embryonic organ of nutrition. Many authors said there was no trace of the embryo until a later stage, and outside the yelk; sometimes the two-layered embryonic disk itself, at other times only the central portion of it (as distinguished from the germinative area, which we will describe presently), was taken to be the first outline of the embryo. In the light of the gastræa theory it is hardly necessary to dwell on the defects of this earlier view and the erroneous conclusions drawn from it. In reality, the first segmentation-cell, and even the stem-cell itself and all that issues therefrom, belong to the embryo. As the large original yelk-mass in the undivided egg of the bird only represents an inclosure in the greatly enlarged ovum, so the later contents of its embryonic yelk-sac (whether yet segmented or not) are only a part of the entoderm which forms the primitive gut. This is clearly shown by the ova of the amphibia and cyclostoma, which explain the transition from the yelk-less ova of the amphioxus to the large yelk-filled ova of the reptiles and birds.

Fig. 105—Severance of the discoid mammal embryo from the yelk-sac, in transverse section (diagrammatic). A The germinal disk (h, hf) lies flat on one side of the branchial-gut vesicle (kb). B In the middle of the germinal disk we find the medullary groove (mr), and underneath it the chorda (ch). C The gut-fibre-layer (df) has been enclosed by the gut-gland-layer (dd). D The skin-fibre-layer (hf) and gut-fibre-layer (df) divide at the periphery; the gut (d) begins to separate from the yelk-sac or umbilical vesicle (nb). E The medullary tube (mr) is closed; the body-cavity (c) begins to form. F The provertebræ (w) begin to grow round the medullary tube (mr) and the chorda (ch): the gut (d) is cut off from the umbilical vesicle (nb). H The vertebræ (w) have grown round the medullary tube (mr) and chorda; the body-cavity is closed, and the umbilical vesicle has disappeared. The amnion and serous membrane are omitted. The letters have the same meaning throughout: h horn-plate, mr medullary tube, hf skin-fibre-layer, w provertebræ, ch chorda, c body-cavity or cœloma, df gut-fibre-layer, dd gut-gland-layer, d gut-cavity, nb umbilical vesicle.

It is precisely in the study of these difficult features that we see the incalculable value of phylogenetic considerations in explaining complex ontogenetic facts, and the need of separating cenogenetic phenomena from palingenetic. This is particularly clear as regards the comparative embryology of the vertebrates, because here the phylogenetic unity of the stem has been already established by the well-known facts of paleontology and comparative anatomy. If this unity of the stem, on the basis of the amphioxus, were always borne in mind, we should not have these errors constantly recurring.

In many cases the cenogenetic relation of the embryo to the food-yelk has until now given rise to a quite wrong idea of

[ 117 ]

the first and most important embryonic processes in the higher vertebrates, and has occasioned a number of false theories in connection with them. Until thirty years ago the embryology of the higher vertebrates always started from the position that the first structure of the embryo is a flat, leaf-shaped disk; it was for this reason that the cell-layers that compose this germinal disk (also called germinative area) are called “germinal layers.” This flat germinal disk, which is round at first and then oval, and which is often described as the tread or cicatricula in the laid hen’s egg, is found at a certain part of the surface of the large globular food-yelk. I am convinced that it is nothing else than the discoid, flattened gastrula of the birds. At the beginning of germination the flat embryonic disk curves outwards, and separates on the inner side from the underlying large yelk-ball. In this way the flat layers are converted into tubes, their edges folding and joining together (Fig. 105). As the embryo grows at the expense of the food-yelk, the latter becomes smaller and smaller; it is completely surrounded by the germinal layers. Later still, the remainder of the food-yelk only forms a small round sac, the yelk-sac or umbilical vesicle (Fig. 105 nb). This is enclosed by the visceral layer, is connected by a thin stalk, the yelk-duct, with the central part of the gut-tube, and is finally, in most of the vertebrates, entirely absorbed by this (H). The point at which this takes place, and where the gut finally closes, is the visceral navel. In the mammals, in which the remainder of the yelk-sac remains without and atrophies, the yelk-duct at length penetrates the outer ventral wall. At birth the umbilical cord proceeds from here, and the point of closure remains throughout life in the skin as the navel.

As the older embryology of the higher vertebrates was mainly based on the chick, and regarded the antithesis of embryo (or formative-yelk) and food-yelk (or yelk-sac) as original, it had also to look upon the flat leaf-shaped structure of the germinal disk as the primitive embryonic form, and emphasise the fact that hollow grooves were formed of these flat layers by folding, and closed tubes by the joining together of their edges.

This idea, which dominated the whole treatment of the embryology of the higher vertebrates until thirty years ago, was totally false. The gastræa theory, which has its chief application here, teaches us that it is the very reverse of the truth. The cup-shaped gastrula, in the body-wall of which the two primary germinal layers appear from the first as closed tubes, is the original embryonic form of all the vertebrates, and all the multicellular invertebrates; and the flat germinal disk with its superficially expanded germinal layers is a later, secondary form, due to the cenogenetic formation of the large food-yelk and the gradual spread of the germ-layers over its surface. Hence the actual folding of the germinal layers and their conversion into tubes is not an original and primary, but a much later and tertiary, evolutionary process. In the phylogeny of the vertebrate embryonic process we may distinguish the following three stages:—

As this theory, a logical conclusion from the gastræa theory, has been fully substantiated by the comparative study of gastrulation in the last few decades, we must exactly reverse the hitherto prevalent mode of treatment. The yelk-sac is not to be treated, as was done formerly, as if it were originally antithetic to the embryo, but as an essential part of it, a part of its visceral tube. The primitive gut of the gastrula has, on this view, been divided into two parts in the higher animals as a result of the cenogenetic formation of the food-yelk—the permanent gut (metagaster), or permanent alimentary canal, and the yelk-sac (lecithoma), or umbilical vesicle. This is very clearly shown by the comparative ontogeny of the fishes and amphibia. In these cases the whole yelk undergoes cleavage at first, and forms a yelk-gland, composed of yelk-cells, in the ventral wall

[ 118 ]

of the primitive gut. But it afterwards becomes so large that a part of the yelk does not divide, and is used up in the yelk-sac that is cut off outside.

When we make a comparative study of the embryology of the amphioxus, the frog, the chick, and the rabbit, there cannot, in my opinion, be any further doubt as to the truth of this position, which I have held for thirty years. Hence in the light of the gastræa theory we must regard the features of the amphioxus as the only and real primitive structure among all the vertebrates, departing very little from the palingenetic embryonic form. In the cyclostoma and the frog these features are, on the whole, not much altered cenogenetically, but they are very much so in the chick, and most of all in the rabbit. In the bell-gastrula of the amphioxus and in the hooded gastrula of the lamprey and the frog the germinal layers are found to be closed tubes or vesicles from the first. On the other hand, the chick-embryo (in the new laid, but not yet hatched, egg) is a flat circular disk, and it was not easy to recognise this as a real gastrula. Rauber and Goette have, however, achieved this. As the discoid gastrula grows round the large globular yelk, and the permanent gut then separates from the outlying yelk-sac, we find all the processes which we have shown (diagrammatically) in Figure 1.108—processes that were hitherto regarded as principal acts, whereas they are merely secondary.

Fig. 106—The visceral embryonic vesicle (blastocystis or gastrocystis) of a rabbit (the “blastula” or vesicula blastodermica of other writers), a outer envelope (ovolemma), b skin-layer or ectoderm, forming the entire wall of the yelk-vesicle, c groups of dark cells, representing the visceral layer or entoderm. Fig. 107—The same in section. Letters as above. d cavity of the vesicle. (From Bischoff.)

The oldest, oviparous mammals, the monotremes, behave in the same way as the reptiles and birds. But the corresponding embryonic processes in the viviparous mammals, the marsupials and placentals, are very elaborate and distinctive. They were formerly quite misinterpreted; it was not until the publication of the studies of Edward van Beneden (1875) and the later research of Selenka, Kuppfer, Rabl, and others, that light was thrown on them, and we were in a position to bring them into line with the principles of the gastræa theory and trace them to the embryonic forms of the lower vertebrates. Although there is no independent food-yelk, apart from the formative yelk, in the mammal ovum, and although its segmentation is total on that account, nevertheless a large yelk-sac is formed in their embryos, and the “embryo proper” spreads leaf-wise over its surface, as in the reptiles and birds, which have a large food-yelk and partial segmentation. In the mammals, as well as in the latter, the flat, leaf-shaped germinal disk separates from the yelk-sac, and its edges join together and form tubes.

How can we explain this curious anomaly? Only as a result of very characteristic and peculiar cenogenetic modifications of the embryonic process, the real causes of which must be sought in the change in the rearing of the young on the part of the viviparous mammals. These are clearly connected with the fact that the ancestors of the viviparous mammals were oviparous amniotes like the present monotremes, and only gradually became viviparous. This can no longer be questioned now that it has been shown (1884) that the monotremes, the lowest and oldest of the mammals, still lay eggs, and that these develop like the ova of the reptiles and birds. Their nearest descendants, the marsupials, formed the habit of retaining the eggs, and developing them in the

[ 119 ]

oviduct; the latter was thus converted into a womb (uterus). A nutritive fluid that was secreted from its wall, and passed through the wall of the blastula, now served to feed the embryo, and took the place of the food-yelk. In this way the original food-yelk of the monotremes gradually atrophied, and at last disappeared so completely that the partial ovum-segmentation of their descendants, the rest of the mammals, once more became total. From the discogastrula of the former was evolved the distinctive epigastrula of the latter.

It is only by this phylogenetic explanation that we can understand the formation and development of the peculiar, and hitherto totally misunderstood, blastula of the mammal. The vesicular condition of the mammal embryo was discovered 200 years ago (1677) by Regner de Graaf. He found in the uterus of a rabbit four days after impregnation small, round, loose, transparent vesicles, with a double envelope. However, Graaf’s discovery passed without recognition. It was not until 1827 that these vesicles were rediscovered by Baer, and then more closely studied in 1842 by Bischoff in the rabbit (Figs. 106, 107). They are found in the womb of the rabbit, the dog, and other small mammals, a few days after copulation. The mature ova of the mammal, when they have left the ovary, are fertilised either here or in the oviduct immediately afterwards by the invading sperm-cells.¹ (As to the womb and oviduct see Chapter XXIX) The cleavage and formation of the gastrula take place in the oviduct. Either here in the oviduct or after the mammal gastrula has passed into the uterus it is converted into the globular vesicle which is shown externally in Fig. 106, and in section in Fig. 107. The thick, outer, structureless envelope that encloses it is the original ovolemma or zona pellucida, modified, and clothed with a layer of albumin that has been deposited on the outside. From this stage the envelope is called the external membrane, the primary chorion or prochorion (a). The real wall of the vesicle enclosed by it consists of a simple layer of ectodermic cells (b), which are flattened by mutual pressure, and generally hexagonal; a light nucleus shines through their fine-grained protoplasm (Fig. 108). At one part (c) inside this hollow ball we find a circular disc, formed of darker, softer, and rounder cells, the dark-grained entodermic cells (Fig. 109).

Fig. 108—Four entodermic cells from the embryonic vesicle of the rabbit. Fig. 109—Two entodermic cells from the embryonic vesicle of the rabbit.

The characteristic embryonic form that the developing mammal now exhibits has up to the present usually been called the “blastula” (Bischoff), “sac-shaped embryo” (Baer), “vesicular embryo” (vesicula blastodermica, or, briefly, blastosphæra). The wall of the hollow vesicle, which consists of a single layer of cells, was called the “blastoderm,” and was supposed to be equivalent to the cell-layer of the same name that forms the wall of the real blastula of the amphioxus and many of the invertebrates (such as Monoxenia, Fig. 29 F, G). Formerly this real blastula was generally believed to be equivalent to the embryonic vesicle of the mammal. However, this is by no means the case. What is called the “blastula” of the mammal and the real blastula of the amphioxus and many of the invertebrates are totally different embryonic structures. The latter (blastula) is palingenetic, and precedes the formation of the gastrula. The former (blastodermic vesicle) is cenogenetic, and follows gastrulation. The globular wall of the blastula is a real blastoderm, and consists of homogeneous (blastodermic) cells; it is not yet differentiated into the two primary germinal layers. But the globular wall of the mammal vesicle is the differentiated ectoderm, and at one point in it we find a circular disk of quite different cells—the entoderm. The round

1. In man and the other mammals the fertilisation of the ova probably takes place, as a rule, in the oviduct; here the ova, which issue from the female ovary in the shape of the Graafian follicle, and enter the inner aperture of the oviduct, encounter the mobile sperm-cells of the male seed, which pass into the uterus at copulation, and from this into the external aperture of the oviduct. Impregnation rarely takes place in the ovary or in the womb.

[ 120 ]

cavity, filled with fluid, inside the real blastula is the segmentation-cavity. But the similar cavity within the mammal vesicle is the yelk-sac cavity, which is connected with the incipient gut-cavity. This primitive gut-cavity passes directly into the segmentation-cavity in the mammals, in consequence of the peculiar cenogenetic changes in their gastrulation, which we have considered previously (Chapter IX). For these reasons it is very necessary to recognise the secondary embryonic vesicle in the mammal (gastrocystis or blastocystis) as a characteristic structure peculiar to this class, and distinguish it carefully from the primary blastula of the amphioxus and the invertebrates.

Fig. 110—Ovum of a rabbit from the uterus, one sixth of an inch in diameter. The embryonic vesicle (b) has withdrawn a little from the smooth ovolemma (a). In the middle of the ovolemma we see the round germinal disk (blastodiscus, c), at the edge of which (at d) the inner layer of the embryonic vesicle is already beginning to expand. (Figs. 110–114 from Bischoff. Fig. 111—The same ovum, seen in profile. Letters as in Fig. 110. Fig. 112—Ovum of a rabbit from the uterus, one-fourth of an inch in diameter. The blastoderm is already for the most part two-layered (b). The ovolemma, or outer envelope, is tufted (a). Fig. 113—The same ovum, seen in profile. Letters as in Fig. 112. Fig. 114—Ovum of a rabbit from the uterus, one-third of an inch in diameter. The embryonic vesicle is now nearly everywhere two-layered (k) only remaining one-layered below (at d).

[ 121 ]

The small, circular, whitish, and opaque spot which the gastric disk (Fig. 106) forms at a certain part of the surface of the clear and transparent embryonic vesicle has long been known to science, and compared to the germinal disk of the birds and reptiles. Sometimes it has been called the germinal disk, sometimes the germinal spot, and usually the germinative area. From the area the further development of the embryo proceeds. However, the larger part of the embryonic vesicle of the mammal is not directly used for building up the later body, but for the construction of the temporary umbilical vesicle. The embryo separates from this in proportion as it grows at its expense; the two are only connected by the yelk-duct (the stalk of the yelk-sac), and this maintains the direct communication between the cavity of the umbilical vesicle and the forming visceral cavity (Fig. 105).

Fig. 115—Round germinative area of the rabbit, divided into the central light area (area pellucida) and the peripheral dark area (area opaca). The light area seems darker on account of the dark ground appearing through it. Fig. 116—Oval area, with the opaque whitish border of the dark area without.

The germinative area or gastric disk of the animal consists at first (like the germinal disk of birds and reptiles) merely of the two primary germinal layers, the ectoderm and entoderm. But soon there appears in the middle of the circular disk between the two a third stratum of cells, the rudiment of the middle layer or fibrous layer (mesoderm). This middle germinal layer consists from the first, as we have seen in Chapter X, of two separate epithelial plates, the two layers of the cœlom-pouches (parietal and visceral). However, in all the amniotes (on account of the large formation of yelk) these thin middle plates are so firmly pressed together that they seem to represent a single layer. It is thus peculiar to the amniotes that the middle of the germinative area is composed of four germinal layers, the two limiting (or primary) layers and the middle layers between them (Figs. 96, 97). These four secondary germinal layers can be clearly distinguished as soon as what is called the sickle-groove (or “embryonic sickle”) is seen at the hind border of the germinative area. At the borders, however, the germinative area of the mammal only consists of two layers. The rest of the wall of the embryonic vesicle consists at first (but only for a short time in most of the mammals) of a single layer, the outer germinal layer.

From this stage, however, the whole wall of the embryonic vesicle becomes two-layered. The middle of the germinative area is much thickened by the growth of the cells of the middle layers, and the inner layer expands at the same time, and increases at the border of the disk all round. Lying close on the outer layer throughout, it grows over its inner surface at all points, covers first the upper and then the lower hemisphere, and at last closes in the middle of the inner layer (Figs. 110–114). The wall of the embryonic vesicle now consists throughout of two layers of cells, the ectoderm without and the entoderm within. It is only in the centre of the circular area, which becomes thicker and thicker through the growth of the middle layers, that it is made up of all four layers. At the same time, small structureless tufts or warts are deposited on the surface of the outer

[ 122 ]

ovolemma or prochorion, which has been raised above the embryonic vesicle (Figs. 112–114 a).

We may now disregard both the outer ovolemma and the greater part of the vesicle, and concentrate our attention on the germinative area and the four-layered embryonic disk. It is here alone that we find the important changes which lead to the differentiation of the first organs. It is immaterial whether we examine the germinative area of the mammal (the rabbit, for instance) or the germinal disk of a bird or a reptile (such as a lizard or tortoise). The embryonic processes we are now going to consider are essentially the same in all members of the three higher classes of vertebrates which we call the amniotes. Man is found to agree in this respect with the rabbit, dog, ox, etc.; and in all these animals the germinative area undergoes essentially the same changes as in the birds and reptiles. They are most frequently and accurately studied in the chick, because we can have incubated hens’ eggs in any quantity at any stage of development. Moreover, the round germinal disk of the chick passes immediately after the beginning of incubation (within a few hours) from the two-layered to the four-layered stage, the two-layered mesoderm developing from the median primitive groove between the ectoderm and entoderm (Figs. 82–95).

Fig. 117—Oval germinal disk of the rabbit, magnified. As the delicate, half-transparent disk lies on a black ground, the pellucid area looks like a dark ring, and the opaque area (lying outside it) like a white ring. The oval shield in the centre also looks whitish, and in its axis we see the dark medullary groove. (From Bischoff.) Fig. 118—Pear-shaped germinal shield of the rabbit (eight days old), magnified. rf medullary groove. pr primitive groove (primitive mouth). (From Kölliker.

The first change in the round germinal disk of the chick is that the cells at its edges multiply more briskly, and form darker nuclei in their protoplasm. This gives rise to a dark ring, more or less sharply set off from the lighter centre of the germinal disk (Fig. 115). From this point the latter takes the name of the “light area” (area pellucida), and the darker ring is called the “dark area” (area opaca). (In a strong light, as in Figs. 115–117, the light area seems dark, because the dark ground is seen through it; and the dark area seems whiter). The circular shape of the area now changes into elliptic, and then immediately into oval (Figs. 116, 117). One end seems to be broader and blunter, the other narrower and more pointed; the former corresponds to the anterior and the latter to the posterior section of the subsequent body. At the same time, we can already trace the characteristic bilateral form of the body, the antithesis of right and left, before and behind. This will be made clearer by the “primitive streak,” which appears at the posterior end.

At an early stage an opaque spot is seen in the middle of the clear germinative

[ 123 ]

area, and this also passes from a circular to an oval shape. At first this shield-shaped marking is very delicate and barely perceptible; but it soon becomes clearer, and now stands out as an oval shield, surrounded by two rings or areas (Fig. 117). The inner and brighter ring is the remainder of the pellucid area, and the dark outer ring the remainder of the opaque area; the opaque shield-like spot itself is the first rudiment of the dorsal part of the embryo. We give it briefly the name of embryonic shield or dorsal shield. In most works this embryonic shield is described as “the first rudiment or trace of the embryo,” or “primitive embryo.” But this is wrong, though it rests on the authority of Baer and Bischoff.

Fig. 119—Median longitudinal section of the gastrula of four vertebrates. (From Rabl.) A discogastrula of a shark (Pristiurus). B amphigastrula of a sturgeon (Accipenser). C amphigastrula of an amphibium (Triton). D epigastrula of an amniote (diagram). a ventral, b dorsal lip of the primitive mouth.

As a matter of fact, we already have the embryo in the stem-cell, the gastrula, and all the subsequent stages. The embryonic shield is simply the first rudiment of the dorsal part, which is the earliest to develop. As the older names of “embryonic rudiment” and “germinative area” are used in many different senses—and this has led to a fatal confusion in embryonic literature—we must explain very clearly the real significance of these important embryonic parts of the amniote. It will be useful to do so in a series of formal principles:—

1. The so-called ”first trace of the embryo” in the amniotes, or the embryonic shield, in the centre of the pellucid area, consists merely of an early differentiation and formation of the middle dorsal parts.

2. Hence the best name for it is ”the dorsal shield,” as I proposed long ago.

3. The germinative area, in which the first embryonic blood-vessels appear at an early stage, is not opposed as an external area to the ”embryo proper,” but is a part of it.

4. In the same way, the yelk-sac or the umbilical vesicle is not a foreign external

[ 124 ]

appendage of the embryo, but an outlying part of its primitive gut.

5. The dorsal shield gradually separates from the germinative area and the yelk-sac, its edges growing downwards and folding together to form ventral plates.

6. The yelk-sac and vessels of the germinative area, which soon spread over its whole surface, are, therefore, real embryonic organs, or temporary parts of the embryo, and have a transitory importance in connection with the nutrition of the growing later body; the latter may be called the ”permanent body” in contrast to them.

The relation of these cenogenetic features of the amniotes to the palingenetic structures of the older non-amniotic vertebrates may be expressed in the following theses: The original gastrula, which completely passes into the embryonic body in the acrania, cyclostoma, and amphibia, is early divided into two parts in the amniotes—the embryonic shield, which represents the dorsal outline of the permanent body; and the temporary embryonic organs of the germinative area and its blood-vessels, which soon grow over the whole of the yelk-sac. The differences which we find in the various classes of the vertebrate stem in these important particulars can only be fully understood when we bear in mind their phylogenetic relations on the one hand, and, on the other, the cenogenetic modifications of structure that have been brought about by changes in the rearing of the young and the variation in the mass of the food-yelk.

We have already described in Chapter IX the changes which this increase and decrease of the nutritive yelk causes in the form of the gastrula, and especially in the situation and shape of the primitive mouth. The primitive mouth or prostoma is originally a simple round aperture at the lower pole of the long axis; its dorsal lip is above and ventral lip below. In the amphioxus this primitive mouth is a little eccentric, or shifted to the dorsal side (Fig. 39). The aperture increases with the growth of the food-yelk in the cyclostoma and ganoids; in the sturgeon it lies almost on the equator of the round ovum, the ventral lip (a) in front and the dorsal lip (b) behind (Fig. 119 b). In the wide-mouthed, circular discoid gastrula of the selachii or primitive fishes, which spreads quite flat on the large food-yelk, the anterior semi-circle of the border of the disk is the ventral, and the posterior semicircle the dorsal lip (Fig. 119 A). The amphiblastic amphibia are directly connected with their earlier fish-ancestors, the dipneusts and ganoids, and further the oldest selachii (Cestracion); they have retained their total unequal segmentation, and their small primitive mouth (Fig. 119 C, ab), blocked up by the yelk-stopper, lies at the limit of the dorsal and ventral surface of the embryo (at the lower pole of its equatorial axis), and there again has an upper dorsal and a lower ventral lip (a, b). The formation of a large food-yelk followed again in the stem-forms of the amniotes, the protamniotes or proreptilia, descended from the amphibia (Fig. 119 D). But here the accumulation of the food-yelk took place only in the ventral wall of the primitive-gut, so that the narrow primitive mouth lying behind was forced upwards, and came to lie on the back of the discoid ”epigastrula” in the shape of the ”primitive groove”; thus (in contrast to the case of the selachii, Fig. 119 A) the dorsal lip (b) had to be in front, and the ventral lip (a) behind (Fig. 119 D). This feature was transmitted to all the amniotes, whether they retained the large food-yelk (reptiles, birds, and monotremes), or lost it by atrophy (the viviparous mammals).

This phylogenetic explanation of gastrulation and cœlomation, and the comparative study of them in the various vertebrates, throw a clear and full light on many ontogenetic phenomena, as to which the most obscure and confused opinions were prevalent thirty years ago. In this we see especially the high scientific value of the biogenetic law and the careful separation of palingenetic from cenogenetic processes. To the opponents of this law the real explanation of these remarkable phenomena is impossible. Here, and in every other part of embryology, the true key to the solution lies in phylogeny.

Title and Contents
Glossary
Chapter XI
Chapter XIII
Figs. 1–209
Figs. 210–408