Amphioxus amongst vertebrates, and in many invertebrates with holoblastic (alecithal) ova, the bilaminar blastoderm is produced, not by delamination, but by the invagination of one pole of an originally simple hollow spherical blastodermic vesicle, the invaginated portion becoming the primitive entoderm and the remaining part of the wall of the vesicle forming the primitive ectoderm (fig. 24). This condition, which was discovered by Kowalewsky, is known as the gastrula stage, and it is regarded by most embryologists, following Haeckel, as typical of the mode of formation of the bilaminar blastoderm throughout the animal kingdom. The aperture of invagination by which the cavity of the entoderm communicates for a time with the exterior has been termed the blast opore (Lankester).

It is not possible in this account of the embryology of the mammal (which must necessarily be very short) to examine at any length the evidence upon which the opinion rests that a gastrula stage can be shown to exist at an early stage in the development of the meroblastic ova of the lower vertebrata. It will be sufficient for the present purpose to state that in fishes, reptiles, and birds, the ova of all of which are of a markedly meroblastic type, that part of the ovum in which alone segmentation has occurred, and in which active development subsequently proceeds, produces a bilaminar blastoderm as in the mammal by the separation off as a distinct layer of a lower or inner stratum of cells to form the primitive entoderm, whilst the remaining cells arrange themselves into an upper or outer stratum, the primitive ectoderm. At one part of the circular blastoderm which has thus been formed there now occurs a crescentic thickening of the ectoderm, on the surface of which a pit or depression becomes formed by an invagination of the ectoderm. This pit extends inwards until it abuts against a subjacent entodermal thickening, and it may even penetrate the entoderm and communicate with the cavity below the blastoderm (which afterwards becomes in part converted into the posterior end of the alimentary canal). The invagination in question has been regarded as a rudimentary blastopore, its time of formation having become shifted to a later period, and the entoderm having already been formed by delamination altogether independently of, in place of resulting from, the invagination, as in the typical mode of gastrula formation. In the mammal a similar invagination of the ectoderm also occurs at the posterior extremity of the embryonic area, and this invagination has been described by Heape in the mole as communicating for a time with the cavity of the blastodermic vesicle (fig. 22, blp), which sub

Fig. 25.-SURFACE VIEW OF AN EMBRYONIC AREA OF THE MOLE

IN WHICH THE MEDULLARY GROOVE HAS BEGUN TO APPEAR
IN FRONT OF THE PRIMITIVE STREAK. AT THE JUNCTION OF
THE TWO A SMALL APERTURE IS SEEN: THIS IS THE DORSAL

OPENING OF THE OBLIQUE NEURENTERIC CANAL. (Heape.) sequently becomes converted in part into the alimentary canal. In birds and reptiles as well as mammals the invagination in question soon becomes extended forward along the middle line of the blastoderm as a linear groove (primitive groove), which indents an ectodermal thickening (primitive streak), and if the posterior invagination represents a blast opore, this groove must be looked upon as an extension of such blastopore, a view which derives support from the fact that there appears to be a tendency for the primitive groove, at least its anterior end, to penetrate to the entoderm, and thus to form here also a canal of communication between the cavity below the entoderm and the exterior. Such a canal is designated "neurenteric," because the anterior end of the primitive streak and groove becomes eventually enclosed by the neural tube, and the caual then effects a (temporary) communication between the neural tube and the enteric canal.

Another important point of resemblance between this invagination and the blastopore of the typical gastrula is the fact that the middle layer of the trilaminar blastoderm begins to develope from the margins of the invagination. But in this respect again there is a difference, for whereas in the simplest and most typical forms, such as Sagitta amongst invertebrates, and Amphioxus amongst vertebrates, the middle layer (mesoblast) originates as a pair of hollow protrusions of the primitive entoderm (cœlom-invaginations of Hertwig, figs. 28, 29); in mammals and birds it makes its first appearance in the form of solid outgrowths from the primitive streak.3

Other views concerning the gastrulation of vertebrates.-Kupffer regards the part

Also, according to Hofmann, to some extent in elasmobranch fishes.

2 No layer corresponding with Rauber's layer of the mammal is known to exist in lower vertebrates, unless that layer is to be regarded as the homologue of the external (corneous) stratum of the epiblast, which is found at a later stage in fishes and amphibia.

3 Rückert has described an imperfect form of colom-invagination in elasmobranch fishes, and Hertwig in amphibia.

which has been above alluded to as invaginated ectoderm (primitive groove) as the homologue of part of the entoderm of more typical forms. If this view be correct, many of the difficulties in the way of regarding the aperture of the invagination as the blastopore, and in explaining the differences in the mode of origin of the mesoblast are removed; but, on the other hand, other difficulties are introduced, and the subject is left by no means clear.

Another view, which was formerly extensively held, regards the blastopore of the meroblastic vertebrate ovum as bounded by the thickened edge of the bilaminar blastoderm (Haeckel). According to this view, the cavity of the gastrula is entirely filled up by a mass of unsegmented or but partially segmented yolk, which also projects for a considerable distance through the blastopore, forming in fact the great mass of the ovum. The primitive groove is regarded as a linear prolongation of this thickened edge of the blastoderm towards the centre of the blastoderm (Balfour), so that the embryo, which developes in front of the primitive streak, thus comes to have a pseudo-central position in the blastoderm instead of developing altogether from its margin as in the lower vertebrata and in invertebrates. In conformity with this idea, it may be noted that at the thickened rim of the blastoderm of these meroblastic ova, the two primary layers are continuous with one another as in the primitive streak. In elasmobranchs an intermediate condition is observed, viz., a short groove, the margins of which are freely continuous with the margin of the blastoderm. If, as His and others have described (vide infra), the mesoblast is in part (vascular and connective tissue part) derived from the thickened rim of the blastoderm, this would furnish another point of resemblance between the primitive streak and that margin.

Ed. v. Beneden has promulgated an entirely different opinion as to the mammalian blastopore from those above described. He regards the condition of the ovum, after the completion of segmentation and before the formation of a blastodermic vesicle, as representing the gastrula stage, and looks upon the point where the granular inner mass of cells comes to the surface between the clear cells which form the outer investment as the blastopore (fig. 14, a). In conformity with this view he considers the layer of clear cells to represent the whole of the primitive ectoderm, and the granular inner mass the primitive entcderm. But since all the more recent observations upon early mammalian ova agree in affirming the formation of the three blastodermic layers from the granular inner mass, and that Rauber's layer either takes no part at all, or only a subordinate part in the formation of the ectoderm of the embryonic area, v. Beneden's view, in spite of the superficial resemblance of the ovum at this stage to certain gastrula forms, has not met with general acceptance from embryologists. Inversion of the blastodermic layers in some mammals. - In the guinea-pig (Bischoff), rat and mouse (Fraser, Selenka), and in some other rodents, an inversion of the usual position of the blastodermic layers is found to occur, the epiblast being innermost, the hypoblast outermost. The foundation of this inversion is laid early by a process of invagination and formation of a central cavity in the mass of entomeres, so that when the blastoderm is differentiated, the innermost cells which are next the (secondary) cavity thus formed become the epiblast, and the outermost the hypoblast, the mesoblast subsequently forming between the two by proliferation of epiblast at the primitive groove, as in other mammals. (For details as to this process of invagination, the student is referred to the papers by Selenka.)

Historical. The existence of several laminæ in the germinal substance of the egg was first suggested by C. F. Wolff in his celebrated work Theoria Generationis, published in 1759, and in his later Memoir On the Development of the Intestine, first published in Nov. Comment. Acad. Petropol. in 1767 and republished in German by J. F. Meckel in 1812. It is, however, to the researches of Pander, conducted under the direction of Döllinger of Würzburg, and published in 1817, and those of v. Baer (1826-1837), that we owe the first consistent attempt to connect the development of the several organs and systems of the embryo with the different constituent parts or layers of the blastoderm. Pander recognised a trilaminar structure of the blastoderm and distinguished the three layers composing it, in their order from above downwards, or from without inwards in the egg, as the serous, vascular, and mucous layers.

In 1850-54 a further important advance was made in the knowledge of the constitution of the blastodermic layers, by the discovery by Remak that the greater part of the middle layer soon after its formation comes to be divided into two laminæ, separated by a space which corresponds to the perivisceral cavity (calom)—a fact which had been partially stated by von Baer. So marked a division of the middle layer and distinction of the parts which are afterwards developed from its two laminæ, has seemed sufficient to some authors to warrant the recognition of four distinct layers in the blastoderm; but it will be found on the whole more convenient to consider the fundamental layers as only three, to which, following the nomenclature of Foster and Balfour, the designations of epiblast, mesoblast, and hypoblast are applied, terms which are synonymous with those of ectoderm, mesoderm, and entoderm, employed by many authors.

The terms ectoderm and entoderm were first applied to the two fundamental layers, shown by Huxley in 1849 to constitute the whole body of colenterates, and which were correctly regarded by him as homologous with the two layers of the bilaminar blastoderm, to which we

have applied the terms primitive ectoderm and primitive entoderm. Since the middle layer is developed from one or both of these primitive layers their permanent representatives are morphologically different, having lost the elements which go to form the middle layer, and it is therefore convenient to accentuate this distinction by the adoption of different terms to represent the permanent layers.

The generalisation that the formation of a bilaminar blastoderm is typically produced by the invagination of a hollow spherical unilaminar blastodermic vesicle is due to Haeckel, and was based largely upon the important researches of Kowalevsky, especially those on Sagitta and Amphioxus. The process of delamination which in some animals produces the two primary layers was originally regarded by Ray Lankester as the typical mode of formation, but is now generally admitted to be a secondary modification. Finally, it has been shown (Balfour, Lankester, R. and O. Hertwig), as is set forth below, that the coelom or body cavity is typically developed. not by a process of splitting of the mesoblast (although in some animals this may occur as a secondary modification), but as hollow protrusions from the primitive alimentary cavity, the cells which bound these protrusions forming the mesoblast. Thus from an originally single blastodermic layer by successive processes of invagination or folding, the three permanent laminæ are ultimately produced.

Such folds may be regarded as formed mechanically by local hypertrophic multiplication of the cells of the laminæ, an increased surface being thus found for the increased number of cells. In analogous manner the folds which accompany the formation and separation of the body and the development of the several organs, e.g., the nervous system, alimentary canal, amnion, may also be regarded as resulting mechanically from cell-multiplication. This mechanical theory of development was first enunciated by Pander, and has of late years been applied extensively by several embryologists, notably by His (Entwickl. d. Huhnchens, 1868, and Unsere Korperform, 1874), and Rauber.

Characters of the blastodermic layers.-The three layers of the blastoderm show from the first distinctive characters (fig. 26). The outer layer, or epiblast, is epithelial in nature and consists of somewhat irregularly columnar cells closely set side

Fig. 26.-TRANSVERSE SECTION THROUGH THE FRONT END OF THE PRIMITIVE STREAK AND
BLASTODERM OF THE CHICK. (From Balfour.)

pr, primitive groove; m, mesoblast; ep, epiblast; hy, hypoblast.

by side, forming a single stratum for the most part, except near the middle line, and becoming thinner and flatter towards the margins of the embryonic area.

The inner layer or hypoblast is also epithelial, but the cells are at first all flattened, and appear therefore quite thin and linear in sections of the blastoderm. At a later stage, the hypoblast cells become markedly columnar and enlarged, so that they considerably exceed the epiblast cells in size.

The middle layer, or mesoblast, which differs, as we have seen, in its mode of origin, being formed secondarily from one or both of the primary layers, also differs from them entirely in its appearance and structure. Instead of consisting of cells closely joined together into a continuous membrane after the manner of an epithelium, the mesoblast is at first composed of cells which are not thus closely arranged, but have, on the contrary, a considerable amount of intercellular fluid between them. They are most irregular in shape, and are often branched and united with one another, so that much of the mesoblast early resembles an embryonic connective tissue.

Before proceeding to describe the commencing development of the embryo it will be instructive to enumerate the parts which are formed respectively from the three blastodermic layers. The following is the relation given in tabular form :

The whole of the nervous system, including not only the central organs (brain and spinal cord), but also the peripheral nerves and sympathetic.

The epithelial structures of the organs of special sense.

The epidermis and its appendages, including the hair and nails.

The epithelium of all the glands opening upon the surface of the skin, including the mammary glands, the sweat glands, and the sebaceous glands.

The muscular fibres of the sweat glands.

The epithelium of the mouth (except that covering the tongue and the adjacent posterior part of the floor of the mouth, which is derived from hypoblast), and that of the glands opening into it. The enamel of the teeth.

The epithelium of the nasal passages, of the adjacent upper part of the pharynx, and of all the cavities and glands opening into the nasal passages.

The urinary and generative organs (except the epithelium of the urinary bladder and urethra).

All the voluntary and involuntary muscles of the body (except the muscular fibres of the sweat glands).

The whole of the vascular and lymphatic system, including the serous membranes and spleen.

The skeleton and all the connective tissue structures of the body.

The epithelium of the alimentary canal from the back of the mouth to the anus, and that of all the glands which open into this part of the alimentary tube.

The epithelium of the Eustachian tube and tympanum.

The epithelium of the bronchial tubes and air sacs of the lungs.

The epithelium lining the vesicles of the thyroid body.

The epithelial nests of the thymus.

The epithelium of the urinary bladder and urethra.

PARABLAST THEORY OF HIS.

MESENCHYME THEORY OF HERTWIG.

The observations of His upon the development of the blood and connective tissues in the bird led him to regard these tissues as originating, not from the mesoblast which in the chick grows out from the sides of the primitive groove, but from cells which,

Fig. 27.--VERTICAL SECTION THROUGH THE BLASTODERM OF A HEN'S EGG TAKEN NEAR THE

PERIPHERY. (Stricker.)

E, epiblast; H, hypoblast, passing at the periphery into an undifferentiated mass of yolk, A, containing large cells filled with yolk granules; M (towards the centre of the blastoderm), mesoblast; M (nearer the periphery), granular cells, apparently derived from A, and lying between the epiblast and hypoblast.

originating either in the yolk or in the thickened rim of the spreading blastoderm, wander in centripetally between the primary layers and fill up all the interstices of the centrifugally. growing true mesoblast. These in-wandering cells being derived, not like the other cells of the embryonic area from the more active primarily differentiated central parts of the blastoderm, but from the peripheral non-embryonic portion, were collectively named by His parablast, and the tissues (blood and blood-vessels, and all the connective tissues) supposed to be formed from them were termed parablastic (all the other tissues of the embryo being termed, in contra-distinction, archiblastic).

His's theory was enunciated as long ago as 1868, although he afterwards introduced into it certain modifications. For a considerable time it met with little acceptance, but of

late years it has obtained, in its modified form as above given, the adherence of many embryologists, and especially of R. and O. Hertwig, Kupffer, Kollmann, and Waldeyer. R. and O. Hertwig have given the name of mesenchyme to His's parablast, while retaining the designation of middle germinal layer or mesoderm for the rest of the mesoblast, from which it differs (1) in its structure, consisting of loosely arranged wandering cells, as distinguished from the epithelium-like lamellæ, of which according to their description the rest of the mesoderm is composed; (2) in its derivation, arising as separate cells from the entoderm instead of in the form of a coherent layer; and (3) in its further development and destination, giving origin to the connective tissues and blood-vessels, and perhaps to the plain muscular tissue, whereas the mesoderm proper gives origin to the skeletal muscles and to the epithelium of the serous cavities, and of the genital and urinary organs. They describe the true mesoderm as consisting of two epithelial lamellæ, which form distinct layers of the blastoderm, so that according to this view the complete blastoderm would consist of four layers (epiblast, outer or somatic mesoblast, inner or splanchnic mesoblast and hypoblast) besides the mesenchyme; to which must be added a median strand of cells set aside for the formation of the notochord. It appears evident from the researches of Kowalevsky in Sagitta and Amphioxus that what is to be regarded as the typical origin of the mesoblast in Metazoa1 takes the form of a pair of diverticula from the primitive archenteric cavity (fig. 28, Ia; fig. 29, I), which hollow diverticula become pinched off from the remainder of that cavity, and. their cavities becoming compressed laterally, are converted into the cœlom or body-cavity (serous cavities of vertebrata), the two walls of this cavity on either side forming respectively the inner and outer mesodermic layers of R. and O. Hertwig. In Sagitta, the diverticula occur in the neighbourhood of the

Fig. 28.-FORMATION OF MESOBLASTIC SOMITES IN AMPHIOXUS, SHOWN IN LONGITUDINAL OPTICAL SECTION. (Hatschek.)

Ia., dorsal view of an embryo in which the mesoblast is beginning to form as two longitudinal folds of the hypoblast which are becoming subdivided from before back by constrictions into separate somites. I., the same viewed in profile, showing the anterior three somites of one side, with their cavities in free communication with the enteric cavity. The neural canal, n. c., is continued posteriorly by a neurenteric canal into the enteric cavity; ep, epiblast; hy, hypoblast. II., dorsal view of a more advanced embryo. The somites are more numerous and are completely separate. In all but the most anterior pair the communication with the enteric cavity is still seen. III., dorsal view at a still later stage. The somite cavities are now completely closed. The cellular rod, ch, shown running along the middle of the embryo is the notochord.

blastopore, which is also the typical seat of origin of the mesoblast, but in Amphioxus they are formed by longitudinal folds of the wall of the archenteric cavity, which grow from before backwards, and become separated up into segments in their progress. In most vertebrates above Amphioxus the mesoblastic outgrowths are from the first solid, not hollow (although a split may early, and does eventually in any case, occur in them, the cœlom being thus produced), nor do they originate so distinctly from the entoderm, but arise rather at the junction