the vessels which, with their surrounding membranes, constitute the umbilical cord-the medium of communication between the foetus and the placenta, when that organ is fully developed.

The egg membranes of the Monotremes present many points of agreement with those of the ovum of the Marsupials,1 and differ from those of the Placental types. Thus Monotremes and Marsupials agree in having a vitelline membrane, which appears between the young ovum and the follicular epithelium, persisting in the one case until the time of hatching, and in the other till a late uterine stage. There are also several other common features fully described in Mr. Caldwell's memoir, but which cannot be detailed in this work.

3

In the Marsupialia the observations made many years ago by Sir R. Owen upon the development of the Kangaroo have been confirmed by those of Dr. H. C. Chapman,2 while Dr. Selenka, and Professor H. F. Osborn have contributed important evidence as to the structure and relations of the foetal membranes of the Opossums and others. It thus appears that up to the period of the very premature birth of these animals the outer covering of the ovum, or false chorion, is free from persistent villi, and not adherent to the epithelium of the uterine walls; for, although fitting into the folds of the latter, it is perfectly and readily separable in its entire extent from them. The umbilical vesicle or yolk-sac is large, vascular, and adherent to a considerable portion of the false chorion or subzonal membrane, while the allantois is relatively small, and although the usual blood-vessels can be traced into it, it does not appear to contract any connection with the false chorion, and, therefore, much less with the walls of the uterus, of such a nature as to constitute a placenta. In some forms, however, such as the Opossums, the umbilical vesicle or yolk-sac develops temporary villi, which unite with the subzonal membrane, or false chorion, to form a disc-like area closely attached to the cells covering the utricular glands of the uterine epithelium, and thus forming a so-called yolk-sac placenta. The function of this organ is considered to be the transmission of the secretions of the utricular glands to the embryo by means of the umbilical vesicle; the function of the allantois being either respiratory or the absorption of the fluid secreted in the uterine cavity by the utricular glands.

While in the uterus the nourishment of the fœtus seems, therefore, to be derived from the umbilical vesicle, as in reptiles and

1 See B. H. Caldwell—“The Embryology of Monotremata and Marsupialia,” Phil. Trans. for 1887, p. 463.

2 Proc. Acad. Nat. Sci. Philadelphia, 1881, p. 468.

3 "Studien uëber Entwickelungeschichte der Thiere," pt. 4, Wiesbaden, 1886. 4 Journal of Morphology, vol. i. p. 373 (1887).

birds, rather than from the uterine walls by means of the allantoic vessels, as in the higher mammals. The latter vessels, in fact, play even a much less important part in the development of these animals, not only than in the placental mammals, but even than in the Sauropsida, for they can scarcely have the respiratory function assigned to them in that group: pulmonary respiration and the lacteal secretion of the mother very early superseding all other methods of providing the due supply both of oxygen and of food required for the development and growth of the young animal. In this sense the Marsupials may be looked upon as the most typically "mammalian" of the whole class. In no other group do the milk-secreting glands play such an important part in providing for the continuity of the race.

In the third primary division of the Mammalia, the so-called Placentalia, the umbilical vesicle generally does not quite unite with the chorion, and disappears as development proceeds, so that no trace of it can be seen in the membranes of an advanced embryo; but it may persist until the end of the intra-uterine life. as a distinct sac in the umbilical cord, or lying between the allantois and amnion. The disappearance or persistence of the umbilical vesicle does not, according to our present knowledge, appear to be correlated with a higher or lower general grade of development, as might be presupposed. It is stated to have been found in Man even up to the end of intra-uterine life, and also in the Carnivora, while in the Ungulata and Cetacea it disappears at an earlier age. In many, if not all, of the Rodentia, Insectivora, and Chiroptera, it plays a more important part, becoming adherent to a considerable part of the inner surface of the chorion, to which it conveys blood-vessels, although villi do not appear to be developed from the surface of this part, as they are on the portion of the chorion supplied by the allantoic vessels. These orders thus present to a certain extent a transitional condition from the Marsupials, although essentially different, in possessing the structures next to be described.

The special characteristic of the whole of the placental mammals constituting the majority of the class, is that the allantois and its vessels become intimately blended with a smaller or greater part of the parietes of the ovum, forming a structure on the outer surface of which villi are developed, and which, penetrating into corresponding cavities of the "decidua," or soft, vascular, hypertrophied lining membrane of the uterus, constitutes the placenta. This organ may be regarded, as Sir William Turner says, both in its function and in the relative arrangement of its constituent textures, as a specially modified secreting gland, the ducts of which are represented by the extremities of the blood-vessels of the foetal system. The passage of material from the maternal to the foetal system of vessels is not

a simple percolation or diffusion through their walls, but is occasioned by the action of a layer of cells derived from the maternal or uterine structures, and interposed between the blood-vessels of the maternal part of the placenta and those of the villi covering the chorion, in which the embryonic vessels ramify.

The numerous modifications in the details of the structure of this organ relate to augmenting the absorbing capacity of the vessels of the chorion, and are brought about either by increasing the complexity of the fœtal villi and maternal crypts over a limited area, or by increasing the area of the part of the chorion covered by the placental villi, or by various combinations of the two methods.

The first class of variations has given rise to a distinction into two principal kinds of placenta : (1) simple or non-deciduate, and (2) deciduate. In the former the fœtal villi are received into corresponding depressions of the maternal surface, from which at the period of parturition they are simply withdrawn. In the second, or more complex form, the relation is more intimate, a layer of greater or less thickness of the lining membrane of the uterus, called "decidua," becoming so intimately blended with the chorion as to form part of the placenta proper, or that structure which is cast off as a solid body at parturition. In other words, in the one case the line of separation between the placenta and uterus at birth takes place at the junction of the fœtal and maternal structures, in the other through the latter, so that a portion of them, often of considerable thickness, and containing highly organised structures, is cast off with the former. It was once thought that the distinction between these two forms of placentation is so important as to constitute a sufficiently valid basis for a primary division of the placental mammals into two groups. It has, however, been shown that the distinction is one rather of degree than of kind, as intermediate conditions may exist, and it is probable that in different primary groups the simpler, non-deciduate form may have become developed independently into one or other of the more complex kinds.

Apart from its intimate structure, the placenta may be met with of very varied general form. It may consist of villi scattered more or less regularly over the greater part of the surface of the chorion, the two extremities or poles being usually more or less bare. This form is called the "diffused placenta." It is probably a primitive condition, from which most of the others are derived, although its existence must presuppose the absence of the umbilical vesicle as a constituent of the chorionic wall. It is found at present in the Manis among Edentates, the Cetacea, the Perissodactyle Ungulates, and the Camels, Pigs, and Chevrotains among the Artiodactyles. Such placenta are always non-deciduate. Recent observations by Sir W. Turner on the placentation of the Dugong show that the

Sirenia present the peculiarity of having a zonary placenta, which is either entirely or in great part non-deciduate, and is, therefore, transitional between the diffused and the true zonary type.

In the true Ruminants or Pecora, among the Artiodactyle Ungulates, the villi are aggregated in masses called cotyledons, with bare spaces between. Such a placentation is called "polycotyledonary." In another modification the villi are collected in a more or less broad band encircling the chorion, leaving a very large portion of the two poles bare, constituting the "zonary placenta," characteristic of the Carnivora, and also occurring in the Elephant, Hyrax, and Orycteropus. The fact of the form of the placenta of these three last-named animals agreeing together, and with that of the Carnivora, does not, however, necessitate the ascription of zoological affinities, as the same ultimate form may have been attained by different processes of development.

In another form one pole only of the chorion is non-vascular, the placenta assuming a dome or bell shape, as in the Lemurs and the Sloths. The transition from this, by the gradual restriction of the vascular area, is easy to the oval or discoidal form of placenta of the Anteaters, Armadillos, and higher Primates. The discoidal placenta of the Rodents, Insectivores, and Chiroptera, though showing so much superficial resemblance to that of the last-named order as to have led to the inclusion of all these forms in one primary group, is now known to be developed in another manner, not by the concentration of villi from a diffused to a limited area, but by retaining the area to which it was originally restricted in consequence of the large surface of the chorion occupied, as before mentioned, by the umbilical vesicle. To compensate for the smallness of area, the complex or deciduate structure has been developed. Among some Rodents there is evidence to show that the discoidal placenta has been derived from a zonary one, of which distinct vestiges have been detected in the Mouse. We may conclude that, although the characters and arrangement of the fœtal structures may not have that extreme importance which has been attributed to them by some zoologists, they will form, especially when more completely understood, valuable aids in the study of the natural affinities and evolution of the Mammalia.1

1 For a full exposition of the present state of knowledge on this subject, see the various memoirs of Sir William Turner, also F. M. Balfour's Treatise on Comparative Embryology, vol. ii. (1881), and J. A. Ryder in American Naturalist, vol. xxi. p. 780 (1887).

CHAPTER III

1

ORIGIN AND CLASSIFICATION OF THE MAMMALIA

Origin. Although, as stated in the first chapter, the mammalian class, as at present known either by existing or extinct forms, is completely isolated from all other groups of the animal kingdom, yet it is impossible to refrain from speculating as to its origin and nearest affinities. In arranging the classes of vertebrates in a linear series it is customary to place them in the following order-Pisces, Amphibia, Reptilia, Aves, Mammalia, an order which probably indicates the relative degree of elevation to which the mos highly developed members of each class has attained. Such an arrangement appears to express the true relationship of the first four classes to one another, but it is quite clear that the Mammalia have no sort of affinity with the Aves. Writing in 1879, Professor Huxley 1 came to the conclusion that, in looking among vertebrates for the progenitors of the Mammalia, we must pass over all known forms of birds and reptiles, and go straight down to the Amphibia. In addition to the characters derived from the conformation of the pelvis upon which the argument was primarily based, the following reasons were given for this conclusion: "The Amphibia are the only air-breathing Vertebrata which, like mammals, have a dicondylian skull. It is only in them that the articular element of the mandibular arch remains cartilaginous, while the quadrate ossification is small, and the squamosal extends down over it to the osseous elements of the mandible, thus affording an easy transition to the mammalian condition of those parts. The pectoral arch [girdle] of the Monotremes is as much amphibian as it is sauropsidian; the carpus and the tarsus of all Sauropsida, except the Chelonia, are modified away from the Urodele type, while those of the mammal are directly reducible to it. Finally, the fact that in all Sauropsida it is a right aortic arch which is the main conduit of arterial blood leaving the heart, while in mammals it is a left aortic arch which

1 Proceedings of the Royal Society of London, vol. xxviii. p. 395 (1879).