(Plate 8, fig. 2), or that of a Chelonian, it will be observed that the resemblance between the former and the Sauropsidan pelvis is, in most respects, closer than that which it bears to the higher Mammalian pelvis. In the reptiles both the pubes and the ischia unite in a ventral symphysis; the pubis has a strong pectineal process, which acquires very large dimensions in the Chelonia; the metischial processes are also often very strong. Nevertheless, there is an important difference, for, in all these animals, the iliac axis is either nearly perpendicular to the sacral axis, or slopes from above downwards and forwards; the obturator axis also inclines downwards and forwards. Hence, in most Lacertilia and Chelonia, the pubes slope forwards very obliquely, while the ischia come more and more forwards. In other words, such modifications of the pelvis as occur in the Lacertilia and the Chelonia are of an opposite kind to those which take place in Mammalia.

The same thing is true of the Crocodilia (Plate 8, fig. 3). Here the ilium is much broader than in the lizards and the Chelonia. This broadening is effected by the expansion of the ilium, both in front of and behind the iliac axis, which retains about the same inclination to the sacral axis that it has in lizards. The ischia have but small metischial processes, and their long axes lie further forwards than in most lizards. The obturator axis inclines forwards, and the iliopectineal axis is parallel with the sacral axis, as in lizards. As in Echidna, a space of the inner wall of the acetabulum is fibrous. The lower boundary of this space is constituted by a prolongation of the anterior end of the cotyloid extremity of the ischium. The interval between this and the anterior end of the ilium answers to the cotyloid end of the pubis in a lizard, but it does not ossify. The pubis corresponds exactly in direction with that of a lizard, but its form is very different. At first narrow and rounded, it gradually flattens from above downwards and, at the same time, widens into a broad trowel-shaped plate of cartilage enclosed in a dense fibrous perichondrium, which lies close beside the middle line in the ventral wall of the abdomen (Plate 8, fig. 12). Each of these flat cartilages is distinct from its fellow throughout the greater part of its extent; but, posteriorly, the two approach, and are united by a broad and strong ligamentous band (Sy. p.l.). The bony portion of the pubis commences just outside the acetabulum and extends to this band, terminating by a curved edge directed inwards and forwards. It is the osseous portions of the pubes which are commonly described as the entire pubes of the Crocodilia, and much speculative ingenuity has been expended upon the interpretation of these apparently anomalous elements of the pelvis, which

* For the latest of these interpretations see Hoffman's excellent memoir, "Beiträge zur Kenntniss des Beckens der Amphibien und Reptilien." "Nied. Archiv für Zoologie," 1876. I cannot but think that had Professor Hoffman studied the crocodile's pelvis in fresh or spirit specimens, he would not have put forward the hypothesis that the pubes of the crocodiles are epipubes." Rathke's account of

are readily moveable upon their fibro-cartilaginous connexions with the acetabulum. But in no essential respect do they differ from ordinary pubes. Throughout their whole length they give attachment to a muscle, which answers to the pectineus and short adductors of the thigh, while the aponeurosis which lies between them and the ischia gives origin as usual to the obturator externus; and the obturator nerves pass out close to the cotyloid ends of the pubes (Plate 8, fig. 12, Ob. n.). For the trowel-shaped forward continuations of the pubes on each side of the symphysis, I will adopt the name of epipubes, proposed by Hoffman for other structures which I believe to be homologous with them. They are firmly connected with the aponeurosis of the external oblique muscle, in which, just in front of their outer edges, lie, on each side, the first set of abdominal false ribs (Plate 8, fig. 12, r). In short, in all their most important relations, these appear to me to be structures homologous with the marsupial bones of the Monotremes and Marsupials, which in Thylacinus are represented by mere cartilages. But although homologous, they are very different in detail; and, in all other respects, the Crocodilian pelvis departs even further than that of the Chelonia and Lacertilia from the Mammalian type.

The Pterosauria seem to have possessed epipubes; and in the Dicynodontia there is an approximation to the backward elongation of the subsacral part of the ilium which is characteristic of Mammals; but, in both these groups there appears to have been no obturator fontanelle.

In the ornithoscelidan reptiles and in birds (Plate 8, figs. 7, 8, and 9), the pelvis, starting from the lacertilian and crocodilian type, undergoes a series of modifications of a new character, the ultimate result of which is a pelvis as much specialized as that of the higher Mammals, but totally different from it in principle.

The broadening of the ilium seen in the crocodile increases, so that the antero-posterior length of the bone eventually becomes very great, chiefly in consequence of the elongation of the præaxial region of the ilium. But, with all this, the direction of the iliac axis does not sensibly change, and it remains more or less inclined downwards and forwards (Plate 8, fig. 9). The inner wall of the acetabulum is largely membranous. The iliopectineal axis becomes slightly inclined to the sacral axis, but never so much even as in Echidna. The main change in the pelvis is, in fact, effected by the extraordinary elongation of the pubes and the ischia, and their rotation backwards and upwards; while, at the same time, the symphysial union of the bones of opposite sides altogether disappears. In Rhea, the ischia unite with some of the post-sacral vertebræ as they do in many Mammalia. The pubis becomes very slender and, as it lies parallel with the ischium, the obturator space is the development of the crocodile's pelvis affords conclusive evidence respecting the homologies of the pubes in these animals.

reduced to a mere slit, often bridged over by a process of the ischium.* The pectineal process is immensely elongated in some Ornithoscelida (as Hulke has shown in Iguanodon, and Marsh in Laosaurus (Plate 8, fig. 8)); but, in birds, it is usually short (fig. 9), and may be absent, and no epipubes have been discovered, either in the Ornithoscelida or in birds.†

Thus, it appears to be useless to attempt to seek among any known Sauropsida for the kind of pelvis which analogy leads us to expect among those vertebrated animals which immediately preceded the lowest known Mammalia. For, if we prolong the series of observed modifications of the pelvis in this group backwards, the "Promammalia ” antecedent to the Monotremes may be expected to have the iliac and obturator axes perpendicular to the sacral axis, and the iliopectineal axis parallel with it; something, in short, between the pelvis of an Ornithorhynchus and that of a land tortoise; and provided, like the former, with large epipubes intermediate in character between those of the lower mammals and those of crocodiles. In fact, we are led to the construction of a common type of pelvis, whence all the modifications known to occur in the Sauropsida and in the Mammalia may have diverged.

It is a well-known peculiarity of the urodele Amphibia, that each os innominatum consists of a continuous cartilage, the ventral half of which is perforated by a foramen for the obturator nerve, but has no large fibrous fontanelle, or obturator foramen, in the ordinary sense of the word. At the junction of the dorsal with the ventral moiety, the acetabulum marks off the iliac portion of the pelvic arch above, from the pubic and ischial regions below; and these are further distinguishable, even apart from their ossifications, by the position of the foramen for the obturator nerve and the origins of the muscles. In full-grown specimens of Salamandra maculosa, the pelvis presents the following characters (Plate 8, figs. 1, 10, and 11):-The iliac axis is slightly inclined forwards, while the iliopectineal axis is practically parallel with the sacral axis. The iliac ossification extends into the acetabulum, and forms a triangular segment of its roof with the apex downwards, exactly as in lizards. The posterior and inferior side of the triangle is separated by a thin band of the primitive cartilage from the upper edge of the similarly triangular cotyloid end of the ischial ossification, the anterior edge of which is vertical, again as in lizards. Between this edge and the anterior and inferior edge of the iliac ossification there is a cartilagi

*I was, at one time, inclined to think that this represented the union of the pubes and ischia of the same side in ordinary Sauropsida, and that the rest of the ischium represented an unusually elongated metischial process; but the study of the development of the pelvis in the chick has convinced me that this is not the case. + See, however, the observations of Mr. Garrod on a "marsupial bone" in ostriches. "Proceedings of the Zoological Society," 1872.

nous interspace, as in crocodiles, which represents the cotyloid end of the pubis (Plate 8, fig. 1). This cartilaginous part of the pubis gives rise to a pectineal process (p. p.), which has the same position as in birds and in Ornithorhynchus. In the floor of the acetabulum, the pubic ossification (fig. 1, Pb.) makes its appearance as a very thin lamina, which extends, underneath the pectineal process, inwards; and gradually surrounds the whole of the thickened transverse ridge of cartilage which corresponds with the pubis. The pubis is thus represented by an axis of cartilage surrounded by bone, and the thick inner extremities of the two pubes are largely united by fibrous tissue (fig. 10, Sy. p.). The ischia are relatively large, and are united, partly by cartilage and partly by ligament, in a long symphysis (fig. 10, Sy. I.). Their posterior and external angles are produced into short metischial processes. In one specimen, I observed a distinct sutural line (Plate 8, fig. 11, s), between the anterior curved edge of the right ischium and the corresponding pubis, while no such suture could be traced upon the other side.

The pelvic arch of Salamandra, therefore, contains all the elements which are found in the higher Vertebrata, but the obturator fontanelle is wanting.

Proceeding from the symphysis pubis, with which it is connected by ligament, is the "ypsiloid" cartilage or epipubis (figs. 1, 11, Ep. p.), which was called by the accurate and acute Dugès the "marsupial cartilage." This indication of the homology of the part has been adopted by Cuvier and others, but I do not know that it has been generally accepted. I believe, however, that the identification is perfectly just.

1

The ypsiloid cartilage proceeds forwards in the middle line, as a stem (Ep. p.) of variable length, and then divides into two branches (Ep. p., Ep. p.), which diverge at right angles to one another, and terminate by rounded extremities. The pedicle of the ypsiloid cartilage is broad and triangular in section, the apex of the triangle being dorsal. The manner in which the abdominal muscles are connected with this cartilage is very instructive (Plate 8, figs. 13, 15). The anterior pecti nations of the external oblique muscle (O. e.) are inserted into a thin but dense fascia, which is hardly separable from the superjacent dermis, and which extends across the middle line to the muscle of the other side. The most posterior bundles of the external oblique, however, are inserted by a rounded tendon into the pectineal process (Plate 8, fig. 15, p.p.), which therefore answers, to the spine or tuberosity of the pubis in Mammalia. Internal to this, the fascia arches over a small space (i.r.) (which corresponds with the inguinal ring), and is then inserted into the whole length of the pedicle and into the posterior half of each ramus of the ypsiloid cartilage. Beyond this point the rami are free externally, though they lie close against the fascia. On their inner sides, however,

they are connected together and with the fascia of the external oblique, in the middle line, by a delicate sheet of fibrous tissue (figs. 13 and 15, ƒ). On comparing this disposition of the tendon of the external oblique muscle with that which obtains in Ornithorhynchus (Plate 8, fig. 14) the correspondence is obvious. For, in the latter, the posterior fibres of that muscle are inserted directly into the spine of the pubis (t. p.). Between these and the outer edge of the marsupial bone there is fibrous interspace, corresponding with the inguinal ring (i. r.); while the more anterior fibres of the external oblique are inserted into the apex of the bone, which is, as it were, imbedded in the fascia. If the pedicle of the ypsiloid cartilage were reduced, the rami at the same time widening behind, until their outer angles reached the pectineal processes, and their free apices shortening, the ypsiloid cartilage of the salamander would be converted into two cartilages having exactly the same relation to the tendon of the external oblique that the marsupial bones have in the Ornithorhynchus.

In the Monotremes (Plate 8, fig. 14, Py.), there are two very large pyramidales muscles, which spring from the whole inner margins of the marsupial bones; their posterior and middle fibres run to the middle line, but the anterior ones constitute a longitudinal band, which extends forwards, and is inserted along with the rectus (R.), of which, indeed, it looks like a part. In the salamander, muscular fibres similarly take their origin from the inner edges of the rami of the epipubis; and the most anterior of these fibres pass forwards, and become more or less confounded with the inner edges of the recti (fig. 13, Py.). But the region which answers to the linea alba (l. a.) is very broad, whence the pyramidales are separated by a wide interval.

Thus far, the muscles which are connected with the epipubis are strictly comparable with those which are attached to the marsupial bones. But, in Salamandra, there is a muscle (Plate 8, fig. 13, A.), of which I have been able to find no representative in the Monotremes I have dissected. This is a thin band of longitudinal fibres, which spring partly from the pubis, close to the symphysis, and partly from the outer edge of the pedicle of the epipubis, and are inserted into the outer edge of the ramus (Plate 8, fig. 13, A.). External to this, a broad flat band of muscular fibres (fig. 13, B.) takes its origin from the pubis and its pectineal process, and runs forwards to be inserted into the modified branchial arches and the tongue. These are the hebosteoglossi of von Siebold ("Observationes quædam de Salamandris et Tritonibus"). Superficial to this muscle, is the proper rectus (R.) itself, marked by its tendinous intersections; while, on the dorsal or deep side of both, there is a curious fan-shaped muscle (C.), which takes its origin from the pectineal process, by a narrow tendon, and spreads out to be inserted into the outer face of the pedicle and the outer edge of the ramus of its side. The ventral face of the pedicle