B.O. Basioccipital. A.S. Alisphenoid. Foramina for nerves. B.S. Basisphenoid. O.S. Orbitosphenoid. 1. Olfactory; 2. optic; 3. & P.S. Presphenoid. Pf. Prefrontal. 4. oculomotor and pathetic Eth. Ethmoid (lamina Sq. Squamosal. nerves; 5. third division of per-pendicularis). Ep. Epiotic. trigeminal; 7. portio dura and E.O. Exoccipital. S.O. Supraoccipital. mollis; 8. pneumogastric; M. Mastoid. Pa. Parietal. Epiph. Pineal gland, or P. or P.S. Petrosal. F. Frontal. epiphysis cerebri. P.M. Petromastoid.

The sella turcica lodges the pituitary body, and the synchondrosial union between the basisphenoid and presphenoid is situated so far forwards that the anterior wall of the fossa is almost wholly formed by the rostrum-like anterior prolongation of the basisphenoid. The [544] spinal cord passes out behind the posterior margin of the basi-occipital. The olfactory nerves leave the skull on each side of the ethmovomerine division of the craniofacial axis.

The walls of the cranial cavity are formed by a number of bones, which are divisible into two series, a superior and a lateral. Of the latter, four pairs of bones, separated by natural lines of demarcation, or sutures, are distinguishable, three of which abut directly upon the cranio-facial axis, while the fourth pair are only indirectly connected with it. Behind are the exoccipitals,³ united with the basioccipital and forming the lateral boundaries of the occipital foramen. In front of these are the petromastoids, complex bones which contain the auditory labyrinth, and are connected with the anterior part of the basioccipital and the posterior and superior part of the basisphenoid, only by cartilage.

Next come the alisphenoids, which are attached to the inferoposterior and the anterior portions of the basisphenoid. And, lastly, the orbitosphenoids articulate with the upper margins of the vertically elongated presphenoid.

In the superior series only four bones can be counted, of which two are single and two are pairs. The hindermost is the supraoccipital bone. It articulates with both the exoccipitals and the petromastoids. The next, in front, is the parietal, single in the adult sheep, but composed of two symmetrical halves in the lamb. It articulates with the petromastoids and with the alisphenoids. The frontals, or anterior paired bones, lastly, unite with the orbitosphenoids, and, in front of them, with the ethmoid.

Most important relations exist between the contents of the cranium and these constituent elements of its walls. The par vagum makes its exit between the exoccipital and the petromastoid; the portio dura and portio mollis enter the petromastoid; the third division of the trigeminal passes through the large "foramen ovale," which, in the sheep, has the exceptional peculiarity of being situated nearly in the middle of the alisphenoid; the optic nerve passes through a foramen included between the orbito- and pre-sphenoids, while, as has been mentioned above, the olfactory nerve passes out beside the ethmoid and in front of the orbitosphenoid. The relation [545] of the pituitary body, or hypophysis cerebri, to the upper surface of the basisphenoid, has already been alluded to; it, of course, gives more or less nearly the position of the third ventricle and crura cerebri. A style passed horizontally through the gemina, or mesencephalon, would strike against, or close to, the anterior margin of the petromastoid bone.

On turning to the exterior of the skull, certain bones come into view which were before invisible, as they take no share in forming the lateral walls of the cranial cavity, but are as it were, stuck on to the outer surface of these walls. The principal of these is the great squamosal bone, applied to the outer surfaces of the petro-mastoid, parietal and alisphenoid bones, sending off its zygomatic process to unite with the jugal, and furnishing the articular surface for the condyle of the lower jaw.

Partly articulated with the squamosal and partly with the petro-mastoid, is the irregular capsule of the tympanic bone, to which the tympanic membrane is attached, on whose removal the ossicule auditûs come into view, consisting of the malleus, incus, and stapes. The processus gracilis of the first of these bones lies between the tympanic and the squamosal. The short process of the incus abuts against the inner wall of the tympanum, just below the squamosal and close to the line of junction of the petrous and mastoid. These are the leading points in the structure of the. sheep's cranium to which I wish to direct attention at present. Bearing them in mind, let us now proceed to the consideration of the skull of a bird.

Composition of the Skull of a Bird (fig 2).

In most adult birds, as is well known, the bones of the cranium have coalesced so completely as to be undistinguishable. But in the chick, and to a greater or less extent, in the adult struthious bird, the boundaries of the various bones are obvious enough; and I will therefore select for comparison with the mammalian skull that of an ostrich, and that of a young chicken.

The craniofacial axis of the bird has the same general figure as that of the sheep, consisting of a thick, solid, median portion, lodging the sella turcica; of a posterior, horizontally, and of an anterior, vertically, expanded division; but it is comparatively shorter and thicker in correspondence with the greater shortness, in proportion to its depth, of the cranial cavity. The sella turcica is very deep, and its front wall is very thick. The lower and anterior half of this wall is produced into a long tapering process, which extends forwards [546] far beyond the anterior limit of the bony lamina perpendicularis of the ethmoid, to end in a point.

Overlying this process, and articulated with more than the posterior half of its upper surface, there is, in the ostrich, a strong thick, vertical, bony plate, narrower in front and behind than in the middle, and below than above. A curved vertical ridge on each lateral surface marks the line of its greatest transverse diameter and seems to indicate a primitive division of the mass into two parts, an anterior and a posterior. The latter is connected above with the bony plates representing the orbitosphenoids. The former exhibits on each side, posteriorly and superiorly, a groove, in which the olfactory nerve rests and, above this, expands into an arched process, which supports the anterior extremity of the frontal bone. Anteriorly, the superior end of the bone widens into a rhomboidal plate, which appears externally between the nasal bones. These anterior and posterior processes of the superior edge of the bone are connected by a delicate ridge, which passes from one to the other above, but leaves an irregular oval gap below.

Fig. 2.–Longitudinal section of the Skull of a young Ostrich.

The anterior edge of the bony plate in question is continued into the unossified septum narium, which below supports the delicate bony representative of the vomer.

In the chick, the whole of the parts just described are unossified, but the composition and structure of the rest of the axis is essentially the same as in the ostrich.

It is not difficult to identify in the craniofacial axis of the birds parts corresponding with those which have been shown to exist in the mammal. In the chick, the basioccipital can be readily separated [547] from the basisphenoid. The latter has the same relation to the sella turcica in the bird as in the mammal; and only differs from it in that singular beaklike process, into which its inferior portion is prolonged anteriorly, and which is produced, according to Kölliker,⁴ by the coalescence with the basisphenoid of a distinct ossification, which is developed in the presphenoidal cartilage and partially represents the presphenoid of the mammal. The rest of the presphenoidal cartilage is more or less completely ossified, and appears to be represented in the ostrich by that part of the "vertical bony plate" which lies behind the curved ridge referred to above; while that part of the plate which is situated in front of the ridge, answers to the lamina perpendicularis of the ethmoid.

Nothing can be more variable, in fact, than the mode in which the ossification of the presphenoidal and ethmoidal portions of the craniofacial axis takes place in birds; while nothing is more constant than the general form preserved by these regions, and their relation to other parts, irrespectively of the manner in which ossification takes place in them. And in these respects birds do but typify the rest of the oviparous Vertebrate

If we compare the inferolateral walls of the ostrich's cranium with those of the sheep, we find the most singular correspondences. Posteriorly are the exoccipitals, which contribute to form the single condyloid head for articulation with the atlas, but otherwise present no important differences. In front of the exoccipital lies a considerable bony mass, which unites, internally and inferiorly, with the basioccipital and basisphenoid bones, and posteriorly is confluent with the exoccipitals. Its anterior margin is distinguishable into two portions, a superior and an inferior, which meet at an obtuse angle. The anterior inferior portion articulates with the alisphenoid; the anterior superior portion with the parietal. The anterior, posterior and inferior, relations of this bone are therefore the same as those of the petromastoid of the sheep.

Superiorly and posteriorly, a well-marked groove (which, however, is not a suture) appears to indicate the line of demarcation between the supraoccipital and this bone, whose pointed upper extremity appears consequently to be wedged in between the supraoccipital and the parietal.

The par vagum passes out between the bony mass under description and the exoccipital; the third division of the trigeminal leaves the skull between it and the alisphenoid. The portio dura and the portio mollis enter it by foramina very similarly disposed to those in [548] the sheep. Superiorly there is a fossa on the inner face of the bone which corresponds with a more shallow depression in the sheep, and, like it, supports a lobe of the cerebellum. Finally, the anterior inferior edge of the bone traverses the middle of the fossa which receives the mesencephalon. In every relation of importance, therefore, this bony mass corresponds exactly with the petromastoid of the sheep, while it differs from it only in its union with the exoccipitals and the supraoccipital posteriorly, and its contact with the craniofacial axis below.

If from the ostrich we turn to the young chick (fig. 3), the condition of this part of the walls of the skull will be found to be still more instructive. The general connexions of the corresponding bony mass, Pt. M. Ep., are as in the ostrich; but while it is even more evident that the groove appearing to separate its upper end from the supraoccipital is no longer a real suture (whatever it may have been), a most distinct and clear suture, of which no trace is visible in the ostrich's skull, traverses the bone at a much lower point, dividing it into an inferior larger piece, united with the exoccipital, and a superior portion, anchylosed with the supraoccipital. The latter contains the upper portions of the superior and external semicircular canals.

Moreover, on endeavouring to separate the inferior bone from the exoccipital, it readily parts along a plane which traverses the fenestra ovalis externally, and the anterior boundary of the foramen of exit of the par vagum internally. The posterior smaller portion remains firmly adherent to the exoccipital, while the other larger portion comes away as a distinct bone.

The latter answers exactly to the mammalian petrosal, while the small posterior segment corresponds with the mammalian mastoid. Like that of the mammal, it is eventually anchylosed with the petrosal; but unlike that of the mammal, it is also, and indeed at an earlier period, confluent with the exoccipital.⁵

Thus, to return to the ostrich's skull, the bony mass interposed between the exoccipital, supraoccipital and parietal bones, and the craniofacial axis, is in reality composed of three bones, an anterior, petrosal, a posterior, mastoid, and a third, which is distinct from the petrosal and mastoid in the chick, but is anchylosed with them in the ostrich, and which has as yet received no name. I shall term it, from its position with respect to the organ of hearing, the epiotic bone, "os epioticum."⁶

[549] The homology of the bone here called petrosal, with that of the mammal, is admitted by all anatomists. The bone which lies immediately in front of the petrosal is, with a no less fortunate unanimity, admitted to be the homologue of the mammalian alisphenoid. But is worthy of particular remark, in reference to the shifting of the relative positions of the lateral elements of the cranial wall, which has been imagined to take place in the ovipara, in consequence of the supposed invariable disappearance of the squamosal from the interior of their skulls; that although precisely the same bones are visible on the inner surface of the cranial cavity in the ostrich as in the sheep, the squamosal being absent in both, yet in the ostrich the third division of the trigeminal does not pass through the middle of the alisphenoid, but between it and the petrosal.⁷

The orbitosphenoids appear like mere processes of the presphenoid, and their relation to the optic nerves is altered in the same way (when compared with the corresponding bones in the sheep) as that of the alisphenoids to the trigeminal, that is to say the nerves pass behind, and not through them.

The superior series of bones in the cranial wall is exactly the same as in the sheep, and the parietals are distinct in the young ostrich, as in the lamb.

Attached to the exterior of the skull of the ostrich are, as in the sheep, several bones; but the appearance of some of these is widely different from that of the parts which correspond with them in the mammal. This is at least the case with the largest and uppermost of these bones, which lies upon the parietal above, the alisphenoid in front, and the exoccipital behind; while internally it is in relation with the petromastoid.

This bone lies immediately above an articular surface, which is furnished to the os quadratum by the petrosal, and more remotely it helps to roof in the tympanic cavity but takes no share in the formation of the fenestra ovalis. It sends a free pointed process downwards and forwards, which does not articulate with the jugal. Except in this particular, however, the bone in question resembles in every essential relation the squamosal of the sheep, while to the same extent it differs from the mastoid of that animal.

I have stated that in the ostrich this bone does not appear upon the inner surface of the wall of the skull, and in this respect, while it resembles the squamosal of the sheep and Ruminants generally, it differs from that of most other Mammalia, in which the squamosal makes its appearance in the interior of the skull, between the parietal, [550] frontal, alisphenoid and petrosal bones, and so contributes more or less largely to the completion of the cranial wall.

But it has been most strangely forgotten that the relations of the bone in question in birds, are by no means always those which obtain in the ostrich. In the young of the commonest and most accessible of domestic birds, in the chicken, the squamosal may be readily seen to enter largely into the cranial wall; a rhomboidal portion of its anterior and internal surface being interposed in front of the petrosal, between this bone, the parietal, the frontal, and the alisphenoid (Sq. fig. 3).

Fig. 3.–Longitudinal section of the Skull of a young Chicken.

There is therefore not a single relation (save the connexion of the jugal) in which this bone does not resemble the squamosal of the Mammalia–there is not one in which it does not differ from their mastoid.

The second bone applied externally to the cranium in the bird, is that large and important structure, the os quadratum, which intervenes between the petrosal and squamosal bones above, and the articular portion of the lower jaw below; which articulates with the pterygoid internally, and with the quadratojugal externally, which gives attachment to a part of the tympanic membrane, posteriorly, and which is very generally termed the tympanic bone, from its supposed homology with the bone so named in the Mammalia. The resemblance to the tympanic bone, however, hardly extends beyond its relation to the tympanic membrane; for in no other of the particulars mentioned above do the connexions of the two bones correspond. The tympanic of the mammal does not articulate with the lower jaw, nor with the pterygoid,⁸ nor with the jugal or quadratojugal. On the other hand, if the connexions of the tympanic membrane were sufficient to determine the point, not only the quad[551]ratum, but the articular element of the lower jaw, and even some cranial bones, must be regarded as tympanic.⁹

Again, if we trace the modifications which the tympanic bone undergoes in the mammalian series, we find that in those mammals, such as Echidna and Ornithorhynchus, which approach nearest to the Ovipara, and which should therefore furnish us with some hint of the modifications to which the tympanic bone is destined in that group, the bone, so far from increasing in size and importance, and taking on some of the connexions which it exhibits in the oviparous Vertebrata, absolutely diminishes and becomes rudimentary, so that the vast bony capsule of the placental mammal is reduced, in the monotreme, to a mere bony ring.

But it is no less worthy of remark, that in these very same animals the malleus and incus have attained dimensions out of all proportion to those which they exhibit in other mammals, and that they even contribute to the support of the tympanic membrane.

So far, therefore, from being prepared by the study of those Mammalia which most nearly approach the Ovipara, to find, in the most highly organised of the latter, an immense os tympanicum, with a vanishing malleus and incus, we are, on the contrary, led to anticipate the disappearance of the tympanicum, and the further enlargement of the ossicula auditûs. Thus far the cautious application of the method of gradations leads us, and leads us rightly–though the demonstration of the justice of its adumbrations can only be obtained by the application of the criterion of development.

It is twenty-one years since this criterion was applied by Reichert. Since his results were published, they have been, in their main features, verified and adopted by Rathke, the first embryologist of his age; and yet they are ignored, and the quadratum of the bird is assumed to be the tympanic of the mammal, in some of the most recent, if not the newest discussions of the subject. Reichert and Rathke have proved, that in the course of the development of either a mammal or a bird, a slender cartilaginous rod makes its appearance in the first visceral arch, and eventually unites with its fellow, at a point corresponding with the future symphysis of the lower jaw. Superiorly, this rod is connected with the outer surface of the cartilage, in which the petrosal bone subsequently makes its appearance. Near its proximal end, the rod-like "mandibular cartilage" sends off another slender cartilaginous process, which extends forwards parallel with the base of the skull. With the progress of development, ossification takes place in the last-named cartilage, and [552] converts it, anteriorly, into the palatine, and posteriorly, into the pterygoid bone. The mandibular cartilage itself becomes divided into two portions, a short, proximal, and a long, distal, by an articulation which makes its appearance just below the junction of the pterygopalatine cartilage. The long distal division is termed, from the name of its original discoverer, Meckel's cartilage. It lengthens, and an ossific deposit takes place around, but, at first, not in it. The proximal division in the mammal ossifies, but usually loses its connexion with the pterygoid, remains very small and becomes the incus.

In the bird the corresponding part enlarges, ossifies, and becomes the os quadratum, retaining its primitive connexion with the pterygoid. In the mammal, the proximal end of Meckel's cartilage ossifies and becomes the malleus, while the rest ultimately disappears. The ossific mass which is formed around Meckel's cartilage remains quite distinct from the proximal end of that cartilage, or the malleus gradually acquires the form of the ramus of the lower jaw, and eventually developes a condyle which comes into contact and articulates with, the squamosal. In the bird, on the contrary, the ramus [553] of the jaw unites with the ossified proximal end of Meckel's cartilage which becomes anchylosed with the ramus, but retaining its moveable connexion with the quadratum (or representative of the incus), receives the name of the articular piece of the jaw. The rest of Meckel's cartilage disappears.

Thus the primitive composition of the mandibular cartilaginous arch is the same in the bird as in the mammal; in each, the arch becomes subdivided into an incudal and a Meckelian portion; in each, the incudal and the adjacent extremity of the Meckelian cartilage, ossify, while the rest of the cartilaginous arch disappears and is replaced by a bony ramus deposited round it. But from this point the mammal and the bird diverge. In the former, the incudal and Meckelian elements are so completely applied to the purposes of the organ of hearing, that they are no longer capable of supporting the ramus, which eventually comes into contact with the squamosal bone. In the latter, they only subserve audition so far as they help to support the tympanic membrane, their predominant function being the support of the jaw.

The tympanic bone of every mammal is, at first, a flat, thin, curved plate of osseous matter, which appears on the outer side of the proximal end of Meckel's cartilage, but is as completely independent of it as is the ramus of the jaw of the rest of that cartilage. In most birds it has no bony representative.¹⁰

It is clear, then, as Professor Goodsir¹¹ has particularly stated, that the os quadratum of the bird is the homologue of the incus of the mammal, and has nothing to do with the tympanic bone; while the apparently missing malleus of the mammal is to be found in the os articulare of the lower jaw of the bird.

It would lead me too far were I to pursue the comparison of the bird's skull with that of the mammal further. But sufficient has been said, I trust, to prove that, so far as the cranium proper is concerned there is the most wonderful harmony in the structure of the two, not a part existing in the one which is not readily discoverable in the same position, and performing the same essential functions, in the other. I have the more willingly occupied a considerable time in the demonstration of this great fact, because it must be universally admitted that the bones which I have termed petrous, squamosal, mastoid, quadratum, articulare in the bird, are the homologues of particular bones in other oviparous Vertebrata, and consequently, if [554] these determinations are correct in the bird, their extension to the other Ovipara is a logical necessity. But the determination of these bones throughout the vertebrate series is the keystone of every theory of the skull–it is the point upon which all further reasoning must turn; and therefore it is to them, in considering the skulls of the other Ovipara, that I shall more particularly confine myself.

Composition of the Skull of the Turtle

It has been seen that in birds the presphenoid, ethmoid, and orbitosphenoid regions are subject to singular irregularities in the mode and extent of their ossification. In the turtle, not only are the parts of the cranium which correspond with these bones unossified, but its walls remain cartilaginous for a still greater extent. In fact, if a vertical section be made through the longitudinal axis of a turtle's skull, it will be observed that a comparatively small extent of the cranial wall, visible from within, is formed by bone, and that the large anterior moiety is entirely cartilaginous and unossified. The anterior part of the posterior, bony, moiety of the cranial wall is formed by a bone (Pt.), whose long, vertical, anterior-inferior margin forms the posterior boundary of the foramen by which the third division of the trigeminal nerve makes its exit from the skull. The anterior and superior margin of the bone is very short, and articulates with the parietal bone. The superior margin is inclined backwards and articulates with the supraoccipital The posterior margin is straight, and abuts against a cartilaginous plate interposed between [555] this bone and that which succeeds it. The inner face of the bone is, as it were, cut short and replaced by this cartilage, whence the inferior edge is also short and is connected only with the basisphenoid, and not with the basioccipital. The anterior margin of the bone corresponds with the middle of the mesencephalon, while its inner face presents apertures for the portio dura and portio mollis. The posterior margin of its outer face forms half the circumference of the fenestra ovalis, and it contains the anterior and inferior portions of the labyrinth. Thus, with the exception of the absence of an inferior connexion with the basioccipital,–a circumstance fully explained by the persistence in a cartilaginous state of part of the bone, it corresponds in the closest manner with the petrosal of the bird. I confess I cannot comprehend how those who admit the homology of the bone called petrosal in the bird with that called petrosal in the mammal (as all anatomists do), can deny that the bone in question is also the petrosal, and affirm it to be an alisphenoid. The general adoption of such a view would, I do not hesitate to say, throw the Theory of the Skull into a state of hopeless confusion, and render a consistent terminology impossible. Where then is the alisphenoid? I reply, that it is unossified. The posterior portion of the cartilaginous side-wall of the skull, in fact, unites with the parietal, the petrosal, and the basisphenoid, just in the same way as the bony alisphenoid of the bird unites with those bones. Furthermore, as in the bird, it bounds the foramen for the third division of the trigeminal nerve anteriorly, and is specially perforated by the second division of the fifth, while the optic and the other divisions of the fifth pass out in front of or through its anterior margin.

Not only is the alisphenoid cartilaginous, but the orbitosphenoid is in the same condition, and a great vertical plate of cartilage represents the whole anterior part of the craniofacial axis, or the presphenoid and ethmovomerine bones.¹² It has been imagined, indeed, that the rostrum-like termination of the basisphenoid represents the presphenoid, but I think this comes of studying dry skulls. Those who compare a section of the fresh skull of a turtle with the like section of the skull of a lamb, will hardly fail to admit that the rostrum of the basisphenoid in the turtle is exactly represented by that part of the sheep's basisphenoid, which forms the anterior and inferior boundary of the sella turcica, and that the suture between the basisphenoid and the presphenoid in the sheep corresponds [556] precisely with the line of junction between the rostrum of the basisphenoid and the presphenoidal cartilage in the turtle.

Connected with the posterior edge of the petrosal by the cartilaginous plate, which has been referred to above, and between this and the exoccipital, there appears, on the inner aspect of the longitudinal section of the turtle's skull, a narrow plate of bone connected above, with the supraoccipital, behind, with the exoccipital, below, with the basioccipital, and leaving between its posterior margin and the exoccipital an aperture whereby the par vagum leaves the skull. In fact, except in being separated from the petrosal by cartilage, this bone presents all the characters of the mastoid of the bird, which it further resembles in forming one-half of the circumference of the fenestra ovalis. In other respects it is more like the mastoid of the sheep, for it is not anchylosed with the exoccipital; it is produced externally into a great bony apophysis, which gives attachment to the representative of the digastric muscle; and it is largely visible external to the exoccipital, when the skull is viewed from behind. Indeed, the resemblance to the mastoid of the mammal is more striking than that to the corresponding bone in the bird. And I think it is hardly possible for any unprejudiced person to rise from the comparison of the chelonian skull with that of the mammal, with any doubt on his mind as to the homology of the two bones.

When the sheep's skull is viewed from behind, the posterior half of the squamosal is seen entering into its outer boundary above the mastoid. On regarding the turtle's skull in the same way, there is seen, occupying the same position, the bone which Cuvier, as I venture to think, most unfortunately, named "mastoid." But if the arguments brought forward above be, as I believe with Hallmann, they are, irrefragable, this bone cannot be the mastoid; and I can discover no valid reason why it should not be regarded as what its position and relations naturally suggest it to be–the squamosal. Its connections with the mastoid, petrosal, and quadratum are essentially the same as those of the squamosal in the bird and the mammal. The quadratum and articulare of the turtle are on all hands admitted to be the homologues of the similarly-named bones in the bird, and therefore all the reasonings which applied to the one apply to the other. When the petrosal, mastoid, and squamosa are determined in the turtle, they are determined in all the Reptilia. But the Crocodilia, Lacertilia, and Ophidia differ from the turtle and Chelonia generally, in that their mastoid is, as in the bird, anchylosed with the exoccipital. The squamosal, again, which in the Crocodilia essentially resembles that of the turtle, becomes a slender and elon[557]gated bone in the Lacertilia, and still more in the Ophidia, in which the quadratum is carried at its extremity.¹³

In the Amphibia the petrous and mastoid have the same relations as in the Reptilia; but it is interesting to remark, that in some Amphibia the anterior margins of the petrosal encroach upon the lateral walls of the skull so as completely to enclose the exit of the trigeminal, just as the posterior margin of the alisphenoid encroached so as to inclose it, in the sheep. It can be hardly necessary to remark, however, that this result has nothing to do with the disappearance of any element in the postero-lateral cranial walls, which have the same composition in the frog as in the crocodile or lizard.

The determination of the homologues of the squamosal, incudal, Meckelian, and tympanic elements in the amphibian skull is by no means an easy matter, but one requiring a much more careful investigation than it has yet received.

In Mammalia, a second arch, the hyoid, is connected with the outer surface of the skull, immediately behind the mandibular, and more particularly with that of the mastoid bone or its rudiment. The proximal end of this arch (which is, at first, like the mandibular arcade, a simple cartilaginous rod), in fact, usually becomes continuously ossified with the mastoid, forming part of the walls of the styloid canal; while below this, and external to the tympanum, it is converted into that slender bone, which is known as the styloid process.

In adult birds and most reptiles, the upper end of the hyoid arch is free, but in some Reptilia ¹⁴ it is attached by a styloid process to the representative of the mastoid. Whether attached to the cranium or not, in all abranchiate Vertebrata the proximal end of the hyoidean arch is quite distinct from that of the mandibular arch.

In the Amphibia, however, I find a condition of the proximal ends of these two arches, which seems to foreshadow that intimate connexion between them which obtains in fishes. On the outer side of the petrosal, and of that part of the exoccipital which represents the mastoid, there lies a cartilaginous mass, which is continued downwards into a pedicle, with whose lower end the mandible is articulated. From the anterior edge of the proximal half of this pedicle, the narrow cartilaginous basis of the pterygoid passes forwards and [558] upwards, to become directly continuous with the palatine bone in the frog, but to stop short of that point in the newt. Posteriorly, close to its proximal end, the pedicle becomes connected by a slender, fibrous or fibro-cartilaginous ligament with the upper extremity of the cornu of the hyoid. The hyoid and the mandibular arches are thus suspended to the skull by a common peduncle, which, to avoid all theoretical suggestion, I will simply term the "suspensorium."

The extent of the ossification which takes place, in and about this primitively cartilaginous suspensorium, varies greatly in different genera of Amphibia. Sometimes its distal end remains wholly unossified; sometimes, as in the common frog, a small outer portion of its lower extremity is ossified and sends a process forwards, becoming what is termed the quadratojugal bone; sometimes, as in the Triton, the distal half of the cartilage becomes more or less completely enclosed in a bony mass.

Another ossific deposit usually takes place in the outer half of the proximal end of the suspensorium, extending for a greater or less distance down towards the distal end, which it may even completely reach. It may be a simple triangular plate, as in Triton, or a T-shaped bone, as in Rana. In either case its lower end is the narrower, and fits into a kind of groove in the posterior and outer margin of the distal ossification.

This bone was considered by Cuvier to be the equivalent of the tympanic and the temporal (=squamosal); by Dugès it was called "temporomastoid."

The last constituent of this region of the skull in the Amphibia is one which is frequently overlooked altogether. In the frog, the membrana tympani is supported by a well-defined cartilaginous and partially ossified hoop, which is originally quite distinct from any of the elements of the suspensorium which have just been described, and which clearly deprives any of them of the right of being considered the homologue of the "tympanicum" of Mammalia.

I must defer the attempt to decide what the parts of the suspensorium really are, until the Piscine skull has been under consideration.

Composition of the Skull of the Carp.

The skulls of fishes present difficulties which necessitate, even for my present limited purpose, the entering into greater detail regarding them, than respecting those of the Reptilia or Amphibia. I select the cranium of the carp for description, as it departs far less widely [559] from the common plan, and therefore forms a better type for comparison with the skulls of other Vertebrata than that of any acanthopterygian or ganoid fish.

The craniofacial axis presents only four distinguishable bones. Behind, is the short basioccipital, with its cup for articulation with the first vertebra of the spinal column. In front of this is a greatly elongated bone, which, as in the bird, sends a process as far as the vomer, and forms the greater part of the axis of the skull; and which, I believe, represents, as in the bird, the basisphenoid and more or less of the presphenoid. The short vomer terminates the craniofacial axis anteriorly, and bears upon its upper surface a vertical septum, which, as in the bird, expands into a broad plate above, and is the ethmoid.

Fig. 6.–Longitudinal section of the Skull of a Carp (Cyprinus carpio).

The orbitosphenoids, united below, spring from the upper and anterior part of the presphenoid. Behind them the lateral walls of the skull are formed by the alisphenoid. These bones have the same essential relations as in the bird, for the olfactory nerves pass out of the skull over, and in front of, the orbito-sphenoids; the optic nerves make their exit behind and beneath these and the alisphenoid, while the trigeminal makes its exit behind the posterior edge of the alisphenoid. When viewed from within, the foramen ovale is seen to be as in the bird, a mere conjugational foramen between the alisphenoid and the bone which follows it; and on an external view, the third division of the trigeminal is seen to pass entirely in front of the last-named bone.

The minutest scrutiny of the relations of this bone only strengthens the conviction suggested by the first view of it, that it is the homologue of the petrosal of birds, and therefore of mammals and reptiles. As in the bird, the anterior margin of the fish's petrosal is divided into a superior and an inferior portion, which meets at an angle, [560] the superior portion articulating with the parietal (and squamosal), the inferior with the alisphenoid. Inferiorly, the petrous articulates with the basisphenoid, and, to a small extent, with the basioccipital. Posteriorly it articulates with a bone through which the pneumogastric passes, and which, guided by the analogy of most Reptilia, of Amphibia, and of birds, I believe to represent the coalesced or connate mastoid and exoccipital. The bone lodges the anterior part of the auditory labyrinth; its middle region corresponds with the middle of the mesencephalon. But as it does not separate the auditory organ from the cavity of the skull, it naturally presents no foramina corresponding with those through which the portio dura and portio mollis pass in abranchiate Vertebrata and Amphibia. There is one relation of the petrosal in the fish, however, in which it seems to differ from that of any of the oviparous Vertebrata hitherto described. Superiorly and posteriorly, in fact, it does not unite with the supraoccipital, which is small, comparatively insignificant, and occupies the middle of the posterior and superior region of the skull; but with a large and distinct bone which forms the internal of the two posterolateral angles of the skull, unites internally with the supraoccipital, anteriorly with the parietal and petrosal, inferiorly with the conjoined mastoid and exoccipital. It is the bone which was called "occipital externe" by Cuvier; and he and others have supposed it to be the homologue of that bone in the turtle which, following Hallmann, I have endeavoured to prove to be the mastoid. As I have already shown, the true mastoid of the fish must be sought elsewhere, and consequently the Cuvierian determination is inadmissible. And I must confess, that if our comparisons be confined to adult Vertebrata, the only conclusion which can be arrived at seems to be, that this bone is peculiar to fishes.

But a remarkable and interesting observation of Rathke, combined with the peculiar structure of the skull of the chick described above, leads me to believe that when their development is fully worked out, we shall find a distinct representative of this bone in many, if not all, vertebrate crania.

In his account of the development of Coluber natrix (see Note IV.), Rathke states that three centres of ossification make their appearance in that part of the cartilaginous wall of the cranium which immediately surrounds the auditory labyrinth. One of these is anterior, and becomes the petrosal; one is posterior, and eventually unites with the exoccipital; the third is superior, and in the end coalesces with the supraoccipital. The posterior ossification clearly [561] represents the mastoid, and it is most interesting to find it, in this early condition, as distinct as in the Chelonian.

The superior ossification has only to increase in size and remain distinct in the same way as the mastoid of the turtle remains distinct, to occupy the precise position of the "occipital externe" of the fish. But, further, it is most important to remark, that when this primarily distinct bone has coalesced with the supraoccipital, it stands in just the same relation to that bone, to the petrosal, to the mastoid and to the semicircular canals, in the snake, as that lateral element, early confluent or connate with the supraoccipital in the chick, which I have termed the "os epioticum." I believe, then, that this "os epioticum," distinct in the young snake, but afterwards confluent with the supraoccipital, and becoming what may be termed the epiotic ala of that bone in the adult, is the homologue of the corresponding bone, or confluent ala of the supraoccipital, in birds and reptiles, while in the fish it remains distinct, and constitutes the "occipital externe."

For the rest, the superior part of the cranial arch in the carp resembles that of the bird. There are a supraoccipital, two parietals, and two frontals; the squamosal occupies the same position as in the chick, and as in the latter, is, in the dry skull, visible from within, in front of the petrosal.

As in the Amphibia, both the mandibular and the hyoidean arches are suspended by a pedicle or suspensorium, which is, to a certain extent, common to both, and presents a complexity of structure which can only be elucidated by the most careful study of development.

In ordinary fishes, such as the carp, stickleback, &c., the proximal end of the suspensorium is constituted by a single bone, Cuvier's "temporal," whose cranial end abuts against the squamosal, petrosal, and post-frontal bones.

This temporal¹⁵ gives off posteriorly a process to which the cornu of the hyoid arch is attached; anteriorly and distally it ends in an expanded plate, with which two bones are connected, in front the tympanal, behind the symplectique. The distal end of the suspensor is constituted by the triangular jugal, whose distal and narrower extremity furnishes the condyle with which the mandible is articulated.

The elongated styliform symplectique is received into a groove on the posterior part of the inner surface of the jugal, and extends nearly to the condyle. In front, the jugal articulates with the transverse, [562] and more or less with the pterygoidien, which again are anteriorly connected with the palatine. The flat tympanal is fitted in between the pterygoidien, jugal, and temporal.

Besides these numerous bones, there are four others which enter less directly into the composition of the suspensorium. These are the pre-opercule, a sort of splint-like bone which lies on the outer and posterior faces of the temporal and jugal, and binds the two together; the operate, which articulates with a special condyle developed for it from the posterior edge of the temporal, above the attachment of the hyoid; the sousopercule, which lies in the opercular membrane beneath this; and lastly, the interopercule, the lowest of all, and commonly more or less closely connected with the angle of the lower jaw.

On examining the region in which these bones are eventually found, in an embryonic fish, I discovered, in their place, a delicate inverted cartilaginous arch, attached anteriorly, by a very slender pedicle, to the angles of the ethmoidal cartilage, and posteriorly connected by a much thicker crus with the anterior portion of that part of the cranial wall which encloses the auditory organ (fig. 8).

The crown of the inverted arch exhibits an articular condyle for the cartilaginous rudiment of the mandible. The posterior crus is not, as it appears at first, a single continuous mass, but is composed of two perfectly distinct pieces of cartilage applied together by their edges. The anterior of these juxtaposed pieces is continuous below [563] with the condyle-bearing crown of the arch, and with its anterior crus or pedicle (P.Q.). It is inclined backwards and upwards, and terminates close to the base of the skull in a free pointed extremity.

The posterior piece (S.Y.H.M.), on the other hand, has its broad and narrow ends turned in the opposite direction. Distally, or below, it is a slender cylindrical rod terminating in a rounded free extremity behind, but close to, the condyle for the mandible; above, it gradually widens and becomes connected with the cranial walls. On its posterior edge there is a convexity which articulates with the rudimentary operculum, and below this it gives off a short styloid process, to which the cartilaginous cornu of the hyoid is articulated. Thus the cartilaginous arch, which stretches from the auditory capsule to the ethmo-presphenoidal cartilage, consists, in reality, of two perfectly distinct and separate portions–the anterior division V-shaped, having its anterior crus fixed and its posterior crus free above; the posterior, styliform, parallel with the posterior leg of the V and free below. The anterior division supports the mandibular cartilage, the posterior the hyoidean cornu.

As ossification takes place, that part of the anterior crus of the V-shaped cartilage which is attached to the ethmo-presphenoidal cartilage becomes the palantine; its angle becomes the jugal ; between these two the transverse and pterygoidien (represented by only one bone in Gasterosteus) are developed in and around the anterior crus: the tympanal arises in the same way around the free end of the posterior crus. Thus these bones constitute an assemblage which is at first quite distinct from the other elements of the suspensorium, and immediately supports the mandibular cartilage.

The proximal end (H.M.) of the posterior styliform division gradually becomes articulated with the cranial walls, and, ossifying, is converted into the temporal . The distal cylindrical end (S.Y.) becomes surrounded by an osseous sheath, which at first leaves its distal end unenclosed. The bone thus formed is the symplectique, which is at first free, but eventually becomes enclosed within a sheath furnished to it by the jugal, and so strengthens the union of the two divisions of the arch already established by the junction of the tympanal with the temporal . The symplectique and temporal do not meet, but leave between them a cartilaginous space, whence the supporting pedicle of the hyoid, which ossifies and becomes the osselet styloide, arises.

The operculum, suboperculum, interoperculum, and preoperculum are not developed from the primitive cartilaginous arch, but make their appearance as osseous deposits in the branchiostegal membrane, behind, and on the outer side of, the posterior crus.

If we turn to the higher Vertebrata, we find, as I have stated above, that, at an early period of their embryonic existence, they also present a cartilaginous arch, stretching from the ethmo-presphenoidal cartilage to the auditory capsule, and supporting the mandibular or Meckelian cartilage on the condyle furnished by its inverted crown. The anterior part of the anterior crus of this arch becomes the palatine bone, which is therefore truly the homologue of the fishes' palatine. The posterior part of it becomes the pterygoid, which therefore is the homologue of the pterygoidien (and transverse ?) of the fish.

The produced crown of the arch in the higher Vertebrata becomes either the incus, or its equivalent, the quadratum. I therefore entertain no doubt that the jugal is really the homologue of the quadratum of other oviparous Vertebrata. That the tympanal has no relation whatsoever with the bone of the same name in the [565] higher Vertebrata is indubitable; and I am unable to discover among them any representative of it. It seems to me to be an essentially piscine bone, to be regarded either as a dismemberment of the quadratum or of the pterygoid. It may be termed the "metapterygoid."

Still less do I find among the higher Vertebrata in their adult state, any representative of the posterior division of the suspensor, constituted by the temporal and symplectique. It is quite clear, that the temporal is not, as Cuvier's name would indicate, the homologue of the squamosal. The whole course of its development would negative such an idea, even if we had not a squamosal already; and I shall therefore henceforward term it, from its function of affording support to both the hyoid and mandibular arches, the hyomandibular bone, "os hyomandibulare," while the other bone of this division may well retain the name of symplectic.

It is commonly supposed that the hyomandibular, symplectic, metapterygoid, and quadrate are all to be regarded as mere subdivisions of the quadratum of higher Vertebrata. Such a view, however, completely ignores and fails to explain, the connexion of the hyoidean arch with the hyomandibular bone. In no one of the higher Vertebrata does such a connexion ever obtain between any part of the quadratum and the hyoid, which are quite distinct, and attached separately to the walls of the cranium, in even young embryos of the abranchiateVertebrata.

Nevertheless, in their very earliest conditions, these embryos are said to present a structure, which, if I mistake not, shadows forth the organization of the fish. The visceral arches, in which the mandibular and hyoid cartilages are developed, are at first separated to the very base of the cranium by a deep cleft, the anterior visceral cleft, so that the semi-cartilaginous rudiments of the mandibular and hyoid are completely separate. Subsequently they are said to coalesce above, as the visceral cleft diminishes, so as to have a common root of attachment to the cranium; and this, I apprehend, answers to the hyomandibular bone, and its prolongation to the symplectic. With advancing development, however, this part does not advance, but remains stationary, and becomes confounded with the wall of the cranium; so that the two arches subsequently appear to be attached to the latter quite independently, and there is nothing left to represent this division of the suspensorium in fishes.

I am strengthened in this view by the structure and development of the palatosuspensorial apparatus in the Amphibia, whose consideration I deferred when speaking of the skull in that class. [566] On examining a young tadpole (fig. 9), a cartilaginous process is seen to arise from the walls of the cranium, opposite the anterior part of the auditory capsule, and, passing obliquely downwards and forwards, to end in a rounded condyloid head, which articulates with the representative of Meckel's cartilage. At the anterior boundary of the orbit the process gives off a broad, nearly vertical apophysis (O), Which ends superiorly in a free, rounded, and incurved edge.

The crotaphite muscle passes to its insertion on the inner side of this, the so-called "orbitar process." From the condyle the cartilaginous process sweeps upwards and inwards, and ends by passing into the ethmo-presphenoidal cartilage. It consequently forms an inverted arch, whose keystone is the condyle for Meckel's cartilage, and is, in its connexions and form, strictly comparable with the cartilaginous arch which I have described in the embryo fish. The posterior crus [567] of the arch, it is true, is not divided into two parts, but nevertheless it represents the whole suspensorium of the fish, and not merely the quadratum of the abranchiate vertebrate, because immediately behind the orbitar process it presents an excavated surface, which articulates with the proximal end of the cornu of the hyoid. That part of the cartilaginous arch, therefore, which lies above and behind this point, corresponds with the proximal division of the suspensorium in the fish, or with the hyomandibular bone; while that portion which lies below and in front of it, corresponds with the distal division of the suspensorium and the anterior crus of the arch in the fish, or in other words, with the symplectic, quadratum, metapterygoid, pterygoid, transverse, and palatine bones.

In the course of development, in fact, the palatine bone appears, as in the fish, in that part of the arch which is immediately connected with the ethmo-presphenoidal cartilage, and a single pterygoid in that part of its anterior crus which lies between the palatine and the articular portion, which obviously represents the quadratum. But this pterygoid is, in the adult frog, a large bone, which, on the one hand, stretches down on the inner side of the quadrate cartilage, and, on the other, sends a process inwards and upwards, which nearly reaches the base of the skull. If the pterygoid, transverse, and metapterygoid of the fish were anchylosed into one bone, or if the corresponding region of the primitive cartilage were continuously ossified, the result would be a bone perfectly similar to the pterygoid of the frog; and I entertain no doubt that the amphibian pterygoid does really represent these bones.

The inferior ossification in the batrachian suspensorium certainly answers to the quadratum, in Triton–whether it should be regarded partly or wholly as a quadrato-jugale in the frog seems to be a question of no great moment–inasmuch as we may be quite sure that the lower end of the frog's suspensorium represents the quadrate or incudal element in other vertebrate.

It is well known that, in the course of the development of the frog, the end of the suspensorium, as it were, travels backwards, so that its axis, instead of forming an acute angle, open forwards, with that of the cranium, as in the tadpole (fig. 9), forms a very obtuse angle, open downwards, in the adult frog. This change is accompanied by a relative and absolute lengthening of that part of the suspensorium which lies between the articulation of the hyoid and that of Meckel's cartilage (containing its proper quadrate portion), and by a relative shortening of that part which lies between the articulation of the hyoid and the skull (or the hyomandibular portion). The consequence of this is, that the articular surface for the hyoid appears constantly to approach the cranial wall, until at length, in the adults it seems to be almost in contact with it. If a knife were passed obliquely between the pterygoid and the suspensorium, and then carried through the suspensorium to its posterior margin a little above the condyle for the mandible, it would divide the suspensor into a proximal and a distal portion, precisely resembling those which naturally exist in the embryonic fish. If the proximal division ossified, it would clearly represent the hyomandibular and symplectic bones. Now in the Amphibia, although the suspensor is not thus divided, it ossifies very nearly as if it were, and the superior or proximal ossification is the so-called "temporo-tympanic," "temporo-mastoid," or "squamosal" bone.¹⁶

That this bone is really the homologue of the hyomandibular and symplectic in the fish, becomes, I think, still more clear when we compare it with such an aberrant form of piscine suspensorium as is presented by some of the eel-tribe (Muræna, e.g.). In these fishes the suspensorium is formed by only two bones, a small distal quadratum, which, as usual, articulates with the lower jaw, and a large wide proximal bone, which articulates above with the post-frontal and squamosal, gives attachment to the operculum and to the cornu of the hyoid, and sends down a process towards the articular head of the quadratum. The single bone, which represents the three pterygoids of other fishes, is articulated for the most part with the quadratum, but partly with this proximal bone. The latter, therefore, clearly represents both the hyomandibular and the symplectic bones of ordinary fishes.

But if the suspensorium of Triton be compared with that of Muræna, e.g., it will, I think, be hardly doubted, that while the distal ossification in the former corresponds with the quadrature the proximal answers (at any rate, chiefly) to the hyomandibular bone of the Muræna. Indeed it differs from the latter principally in being an ossific deposit in the outer portion only of the primitive cartilages.¹⁷

Thus it would seem, that in the manner in which the lower jaw is [569] connected with the cranium, Pisces and Amphibia, as in so many other particulars, agree with one another, and differ from Reptilia and Aves on the one hand, as much as they do from Mammalia on the other. And the difference consists mainly, as might be anticipated, in the large development in the branchiate Vertebrata of a structure which aborts in the abranchiate classes. A most interesting series of modifications, all tending to approximate the ramus of the mandible more closely to the skull,¹⁸ is observable as we pass from the fish to the mammal. In the first, the two are separated by the hyomandibular, the quadrate, and the articular elements, the first of which becomes shortened in the Amphibia. In the oviparous abranchiate Vertebrata the cranium and the ramus are separated only by the quadratum and the articulare, the hyomandibulare having disappeared. Finally, in the mammal, the quadratum and the articulare are applied to new functions, and the ramus comes into direct contact with the cranium.

The operculum, suboperculum, and interoperculum appear to me to be specially piscine structures, having no unquestionable representatives in the higherVertebrata. Much might be said in favour of the identification of the preoperculum with the tympanic bone; but there are many arguments on the other side, and at present I do not see my way to the formation of a definite conclusion on this subject.

In the preceding discussion of the structure of the osseous. vertebrate skull, I have desired to direct your attention, more particularly, to the consideration of those fundamental bones, the determination of whose homologues throughout the vertebrate series is of the greatest importance for my present object. The presphenoid, ethmoid, mastoid, and petrosal are the Malakhoff and the Redan of the theory of the skull; and if anatomists were once agreed about their homologues, there would be comparatively little left to dispute about.

But besides the axial, inferolateral, and superior series of bones there are other, less constant, elements of the cranial wall, forming a discontinuous superolateral series. These are the epiotic, the squamosal, the postfrontal, the prefrontal, and lacrymal bones. Of the two first-named of these bones I have already spoken sufficiently. The postfrontal exists only in Reptiles and Fishes, and is always situated between the frontal, alisphenoid, petrosal, and squamosal–[570] the extent to which it is absolutely in contact with any one of these bones varying.

The prefrontal and lacrymal bones are always developed in or upon that lateral process of the ethmosphenoidal plate, which gives attachment externally to the palatopterygoid arch; consequently they lie at the anterolateral ends of the frontal, and have more or less close relations with it, the ethmoid and the palatine bones.

Finally, the nasal bones (or bone) never enter into the composition of the walls of the skull, but have the same relation to the anterior and upper expanded edge of the prolonged lamina perpendicularis or body of the ethmoid, as the vomer or vomers have to its lower edge.

If the conclusions which I have laid before you are correct, the following propositions are true of all the bony skulls of Vertebrata.

1. Their axis contains at most five distinct bones, which are, from before backwards, the basioccipital, the basisphenoid, the presphenoid, the ethmoid, and the vomer; but any of these bones; except the basisphenoid, may be represented by cartilage, and they may anchylose to an indefinite extent; so that the number distinguishable as separate bones in any skull cannot be predicated. The craniofacial axis invariably presents the same regions, but the histological character of these regions may vary.

2. Their roof contains at most, leaving Wormian bones out of consideration, five bones (supraoccipital, parietals and frontals), or seven, if we include the epiotic bones in the roof. The number falls below this in particular cases, for the same reason as that given for the apparent variations in composition of the axis.

3. Their inferolateral wall contains at most six pair of bones (exoccipitals, mastoids, petrosals, alisphenoids, orbitosphenoids, prefrontals), whose apparent number, however, is affected by the same causes.

4. The axial bones have definite relations to the brain and nerves. The basioccipital lies behind the pituitary body, the basisphenoid beneath it, the presphenoid in front of it. In fact the pituitary body may be regarded as marking the organic centre, as it were, of the skull–its relations to the axial cranial bones being the same, as far as I am aware, in all Vertebrata.

The olfactory nerves pass on either side of the ethmoid, which bounds the cranial cavity in front, the greater part of its substance and that of the vomer being outside the cranial cavity.

5. The lateral bones have definite relations to the brain, nerves, and organs of sense. The exoccipital lies behind the exit of the par [571] vagum; the mastoid lies in front of it; the petrosal lies behind the exit of the third division of the trigeminal; the alisphenoid lies in front of it; though either bone may, to a certain slight extent, encroach on the province of the other. The optic nerve passes out more or less in front of the alisphenoid, and behind, or through, the orbitosphenoid.

The organ of hearing is always bounded in front by the petrosal bone, which limits the anterior moiety of the fenestra ovalis.

The organs of smell always lie on each side of the ethmovomerine part of the axis.

The greater part, or the whole, of the petrosal lies behind the centre of the mesencephalon.

6. The attachment of the mandibular arch to the skull is never situated further forward than the posterior boundary of the exit of the trigeminal; consequently it cannot belong to any segment of the skull in front of the petrosal.

But if propositions of this generality can be enunciated with regard to all bony vertebrate skulls, it is needless to seek for further evidence of their unity of plan. These propositions are the expression of that plan, and might, if one so pleased, be thrown into a diagrammatic form. There is no harm in calling such a convenient diagram the 'Archetype' of the skull, but I prefer to avoid a word whose connotation is so fundamentally opposed to the spirit of modern science.

Admitting, however, that a general unity of plan pervades the organization of the ossified skull, the important fact remains, that many vertebrated animals–all those fishes, in fact, which are known as Elasmobranchii, Marsipobranchii, Pharyngobranchii, and Ditnoi–have no bony skull at all, at least in the sense in which the words have hitherto been used. In these Vertebrata the skull is either membranous or cartilaginous; or if ossified, the ossific matter presents no regular grouping around a few distinct centres.

Thus the cranium of the Amphioxus is nothing but a membranous capsule, whose walls are continuous with those of the canal for the spinal cord, and in whose floor lies a continuation of the notochord which underlies the spinal canal.

In the Marsipobranchii there is a marked increase in the capacity of the cranium as compared with that of the spinal canal, in correspondence with the decided differentiation of the cerebral masses; and, at the same time, the cranial walls have undergone a more or less extensive chondrification. The notochord terminates in the midst of the firm and solid cartilaginous plate which forms the posterior part of the basis cranii, and which sends forward two processes, including a membranous interspace. The auditory capsules are enclosed within prolongations of the sides of the basilar plate; and just in front of and below them, the root of each process of the basal plate gives off a solid prolongation, which passes at first outwards and downwards, and then bends upwards and forwards, to rejoin the anterior part of the process of the basilar plate of its side. An inverted arch is thus formed, and the space included between its crura and the sides of the cranium, constitutes the floor of the orbit.

The posterior crus of the arch is divided into two, more or less distinct, pillars, the posterior of which supports the hyoidean arc; the mandibular arc appears to be absent.

The apertures whereby the cranial nerves make their exit are situated in the side-walls of the capsule, that for the vagus lying immediately behind the auditory capsule, while that for the trigeminal is immediately in front of the same organ. The olfactory nerves perforate the anterior walls of the cranial capsule; the optic, its lateral walls between them and the trigeminal.

The skulls of the Elasmobranchii, again, appear at first to be something quite different from either of these. The cranium is here a cartilaginous box, more or less incomplete and membranous above, and presenting on each side posteriorly a transverse enlargement, in which the auditory organ is contained; while anteriorly it expands into a broad plate, which on each side overhangs the olfactory sacs. The notochord and the membranous space have disappeared, or their traces only are visible in the base of the cranium, whose walls are, as it were, crusted with a multitude of minute plates of bone.

In the Chimæræ the inferolateral walls of the cranium pass into a cartilaginous arch-like plate which form the floor of the orbit, and: whose posterior part, as in the Marsipobranchii, gives attachment to the hyoidean arch; besides which, a mandibular cartilage is connected with the condyloid surface developed from the crown of the arched plate.

In the Plagiostomes there is also an inverted suborbitar arch with a mandibular cartilage and a hyoidean apparatus, but the structure of the arch is different from what obtains in Chimera.

The outer wall of that portion of the cranium which lodges the auditory organ, in fact, furnishes an articular surface for a strong moveable peduncle, to which the hyoid arc is usually attached. At its lower end, however, this peduncle does not articulate with the mandibular cartilage, but is directly connected with a strong cartilaginous plate which forms the upper boundary of the gape, and is articulated anteriorly with the sides of the skull in front of the orbit. [573] This plate bears the upper series of teeth, and bites more or less directly against the mandible, which is moveably articulated with a condyle furnished by its posterior extremity.

The upper plate is commonly, though, as I think, erroneously, regarded as the homologue of the maxilla and premaxilla in other fishes; the peduncle as the homologue of their whole suspensorium.¹⁹

The par vagum leaves the skull behind the auditory organ; the trigeminal passes out in front of it; and then its third division traverses the space enclosed between the peduncle, the upper plate, and the skull. The optic nerve passes through the lateral walls of the skull in front of the trigeminal, and the olfactory perforates its anterior boundary.

So brief and simple a statement of the characters of the skulls of these three orders of fishes, while it brings their diversities into prominence, also exhibits an amount of uniformity among them which is not a little remarkable. The exits of the great nerves have fixed relations to the auditory capsules, to the anterior boundary of the skull, and to the pituitary body. The inferior arc of the hyoid is constant (except in the Pharyngobrachii), and has always, speaking broadly, the same relative position with respect to the auditory capsule and the posterior crus of the suborbitar arch. The suborbitar arch itself is always present (except in Pharyngobranchii); its posterior crus is always attached to the cranium behind the third division of the trigeminal nerve, while the anterior is invariably fixed to that part of the skull which lies behind, or beside, the base of the olfactory capsule.

Thus the employment of the method of gradation alone exhibits a surprising uniformity in the organization of these lower forms of skull; and on comparing them with the higher forms, it seems obvious that, so far as it goes, their plan is identical with that of the latter; for the relations of the auditory organ to the par vagum and trigeminal are the same in each; the posterior crus of the suborbitar arch answers to the suspensorium of Teleostei, its anterior crus to their palatopterygoid apparatus. But with all this, there are discrepancies in the structure of the skull itself, which would forbid too close an approximation between the bony and the unossified crania, if their adult forms alone were examined. The study of the development of the ossified vertebrate skull, however, eliminates this difficulty, and satisfactorily proves that the adult crania of the lower Vertebrata are but special developments of conditions through which the embryonic crania of the highest members of the subkingdom pass.

[574] It is to Rathke's luminous researches that we are indebted for the first, and indeed, even now, almost the only, demonstrative evidence of this great fact. Twenty years ago that great and laborious embryologist worked out the early stages of the development of the skull in each class of the Vertebrata. Confirmed and adopted by Vogt and Bischoff, his conclusions have been feebly controverted, but never confuted; and my own observations lead me to believe that they are destined to take a permanent place among the data of biological science. Nothing is easier than to verify Rathke's views in an embryonic fish or amphibian; and as it matters not which of the higher Vertebrata is selected for the study of cranial developments I will state at some length what I have observed in the embryonic frog.²⁰

Before the dorsal laminae have united so as to enclose the primitive craniospinal cavity, the anterior portion of the floor of that cavity is bent downwards. The angle which the deflexed portion forms with the rest becomes less and less obtuse, until, when the dorsal lamina have united and the visceral clefts have begun to appear, it constitutes a right angle.

On examining the floor of the craniospinal cavity at this period, it is seen that the notochord, at present formed by the aggregation of a number of yelk segments or embryo-cells, small in themselves, but larger than those of which the rest of the body is composed, ends in a point immediately behind the angular flexure.

The notochord has no sheath as yet, and is not in any sense prolonged into the deflexed portion of the floor of the craniospinal cavity.

When the visceral clefts first appear, they are best seen from the inner or pharyngeal aspect of the visceral wall. Five, of which the two anterior are the longest and about equal, while the others gradually diminish in length from before backwards, can be distinctly observed. They mark out the boundaries of a corresponding number of "visceral arches," and there is sometimes an appearance as of a sixth visceral arch behind the last cleft. A horizontal section shows that these arches differ in nothing but their relative size–in no other respect can one of them be distinguished from the other.

The anterior visceral cleft lies in a transverse plane, immediately behind the angular bend of the floor of the craniospinal cavity, or, as I shall henceforward term it, mesocephalic flexure. Consequently the posterior part of the first visceral arch passes into the future basis cranii close to the flexure.

The parts of the cerebrum are now distinguishable. It is bent in [575] correspondence with the mesocephalic flexure, and its most projecting portion, or the angle of the bend, is the rudiment of the mesencephalon. The large rudiment of the pituitary body lies immediately in front of the flexure, and is therefore altogether anterior to the end of the notochord and to the posterior part of the first visceral arch. The rudiment of the eye lies at first altogether in front of the flexure, and therefore anterior to the root of the first visceral arch.

The auditory vesicles make their appearance on each side of a line which would cut the chords a little behind its anterior termination. They are at first quite free and perfectly distinct from the walls of the cranium, which is in accordance with Remak's statement, that they are originally formed by the involution of the epidermic layer of the embryo. They long remain separate and easily detachable from the cranial walls.

Ten days after impregnation, larvæ with rudimentary external gills. and colourless blood, still exhibited some traces of the mesocephalic flexure, but the angle formed by the anterior and posterior portions of the cranium was very obtuse; the base of the cranium had, in fact, undergone a gradual straightening. The rudiments of the cranial skeleton had made their appearance, and consisted, behind the mesocephalic angle, of a broad semicartilaginous plate enclosing the anterior end of the notochord, but not covering it above or below. It is not as yet adherent to the auditory sacs.

That part of the middle of the basis cranii which underlies the pituitary body is not converted into cartilage, but remains membranous, and may be called the "subpituitary membrane." The delicacy of this membrane is so great that it is easily torn, when the pituitary body seems, as Rathke originally supposed, to unite with the palatine mucous membrane. But that this is not really the case, is readily demonstrable in an embryo whose tissues have been sufficiently hardened with alcohol or nitric acid.

The cartilaginous basal plate gives off a prolongation on either side of the subpituitary membrane. This, the "cranial trabecula" (Schädelbalke of Rathke), passes forwards with a slight convexity outwards, and then turning inwards comes into contact with its fellow (from which, however, it is at first distinct), and spreads out into a broad, flat, elongated process, which I shall term the ethmovomerine cartilage.

Behind the eye and just in front of the auditory capsule (in the posterior part of the first visceral arch, therefore), a cartilaginous process lies, which is connected proximally with the root of the [576] trabecula close to the basal plate, while at its distal end it sends a prolongation upwards to unite with the posterior end of the ethmovomerine cartilage. It then forms an arch, between which and the basis cranii is an interspace corresponding with, and lodging, the under surface of the large eyeball. The rudiments of the hyoid, mandibular and maxillary apparatus in larvæ at this stage are somewhat indistinct; and indeed not only in this, but in other respects, more instruction is to be derived from tadpoles which have advanced further.

In larvæ, with completely internal branchiae and very short tubercles in the place of hind limbs, the notochord suddenly narrows between the auditory capsules to hardly more than half its preceding dimensions, and then gradually tapers off, to what appears to be a rounded end, a short distance from the anterior boundary of the basal plate. On very careful examination, however, a delicate process (which may by possibility be nothing but a cavity in the cartilage) can be traced from it very nearly to the margin of the basal plate. But there is no continuation whatsoever, either of the notochord itself or of its sheath, into the subpituitary membrane, which is now composed of delicate connective tissue, and from its extreme thinness and transparency would exhibit the least trace of such a prolongation. And I speak the more confidently on this point, because the delicate process of the notochord or cavity in the cartilage, to which I have referred, contains opaque unchanged vitelline granules, and is therefore particularly conspicuous. The basal cartilage is still divided by the notochord into two lateral moieties, which are only united by a short band of cartilage in front of the end of the notochord. It sends off from its outer side a cartilaginous process, which envelopes the auditory capsule externally, but leaves on its inner side a wide aperture for the entrance of the auditory nerve. The oval auditory capsules thus formed have their long axis directed outwards and forwards.

The trabeculæ are still better developed than before, but instead of remaining distinct anteriorly, they have become fused together into a single trapezoidal cartilage, which may be termed the ethmo-presphenoidal plate. This plate, as it were, divides anteriorly into two flat, elongated and somewhat divergent processes, which are concave downwards and end in truncate extremities. Fibrous tissue connects the ends of these ethmovomerine processes with a crescentic cartilaginous plate which supports the horny upper jaw of the tadpole.

The posterior crus of the palatosuspensorial, or suborbitar, arch is [577] not yet united with that portion of the cranial wall which encloses the auditory capsule; but for the rest the same description applies to it which has already been given of the palatosuspensorial arch and its appendages in more advanced tadpoles. In this state, the roof, and all the lateral walls of the cranium, but that part into which the auditory capsule enters, are membranous.

If the skull of the larval frog just described, be laid open and the exit of the nerves observed (fig. 9), it will be seen that the par vagum makes its way out by a foramen situated immediately behind the auditory capsule; that the third division of the trigeminal leaves the cranium in front of the auditory capsule, passing over the posterior crus of the palatosuspensorial arch; and that the optic traverses the membranous walls of the skull between this and the olfactory nerve, which perforates the anterolateral region to enter the olfactory capsules. The latter are situated wide apart, on each side and in front of, the broad ethmo-presphenoidal cartilage and the anterior crus of the palatosuspensorial arch, and are even a little overlapped by the edges of the ethmovomerine processes.

In the further course of development, the trabeculæ approximate and elongate, so as to obliterate the subpituitary membrane, and form with the enlarged basal cartilage, the ethmoid cartilage and the ethmovomerine cartilages, the continuous cartilaginous craniofacial axis. A histological metamorphosis into cartilage is undergone by the roof of the occipital region of the skull, but in front of this it remains membranous; so that in the adult frog (in which this cartilaginous framework persists), the skull, when deprived of its bony matter, presents an anterior fontanelle. The ethmovomerine cartilages diverge still more, and form the broad mass whose lateral cavities shelter the olfactory sacs in the adult frog. If, bearing in mind the changes which are undergone by the palatosuspensorial apparatus and which have already been described, we now compare the stages of development of the frog's skull with the persistent conditions of the skull in the Amphioxus, the Lamprey, and the Shark, we shall discover the model and type of the latter in the former. The skull of the Amphioxus presents a modification of that plan which is exhibited by the frog's skull, when its walls are still membranous and the notochord is not as yet embedded in cartilage. The skull of the lamprey is readily reducible to the same plan of structure as that which is exhibited by the tadpole, while its gills are still external and its blood colourless. And finally, the skull of the shark is at once intelligible when we have studied the cranium in further advanced larvae, or its cartilaginous basis in the adult frog.

[578] Thus, I conceive, the study of the mode in which the skulls of vertebrate animals are developed, demonstrates the great truth which is foreshadowed by a careful and comprehensive examination of the gradations of form which they present in their adult state; namely that they are all constructed upon one plan; that they differ, indeed, in the extent to which this plan is modified, but that all these modifications are foreshadowed in the series of conditions through which the skull of any one of the higher Vertebrata passes.

But if these conclusions be correct, the first problem which I proposed to you,–Are all vertebrate skulls constructed upon a common plan?–is solved affirmatively.

We have thus attained to a theory of general expression of the laws of structure of the skull. All vertebrate skulls are originally alike; in all (save Amphioxus ?) the base of the primitive cranium undergoes the mesocephalic flexure, behind which the notochord terminates, while immediately in front of it, the pituitary body is developed; in all, the cartilaginous cranium has primarily the same structure,–a basal plate enveloping the end of the notochord and sending forth three processes, of which one is short and median, while the other two, the lateral tuberculæ, pass on each side of the space, on which the pituitary body rests, and unite in front of it; in all, the mandibular arch is primarily attached behind the level of the pituitary space, and the auditory capsules are enveloped by a cartilaginous mass, continuous with the basal plate between them. The amount of further development to which the primary skull may attain varies, and no distinct ossifications at all may take place in it; but when such ossification does occur, the same bones are developed in similar relations to the primitive cartilaginous skull. But the theory of the skull thus enunciated is not a 'vertebral theory'; one may have a perfectly clear notion of the unity of organisation of all skulls without thinking of vertebræ.

So much for the first problem before us. I now proceed to the second question, which was, you will recollect, Given the existence of a common plan of organization of all vertebrate skulls; is this plan the same as that of a spinal column?

To deal properly with this question, we must know what is the plan of organization of a spinal column, and that can be learnt only by a careful study of its development, as well as of its adult modifications. Indeed, the latter are unintelligible without a knowledge of the former.

It is impossible to form a clear conception of the essential nature of the process of development of a spinal column, or to compare it [579] with that of the skull, unless we analyse very carefully, and distinguish from one another, the successive steps of that process.²¹

1. The primary changes of form exhibited by the blastoderm in the region of the spinal column, are, in all the Vertebrata whose development has yet been studied, precisely the same. Two ridges, the "laminæ dorsales," bounding a narrow elongated groove, rise up and eventually unite with one another so as to enclose a cavity–the neural canal. External to the junction of the laminæ dorsales with the blastoderm, the latter is converted more or less completely into the "laminæ ventrales," which become incurved, unite, and eventually enclose the visceral cavity.

A transverse section of the embryo in this state shows a very thin and narrow median plate, separating the neural canal above, from the hæmal or visceral canal below, and passing on each side into thickened masses of blastoderm, which give rise to the laminæ dorsales on the one hand, and to the laminæ ventrales on the other.

For convenience of description, I shall term the median plate the "diaphysial plate," and the lateral ridges the "paraphysial thickenings."

2. The primary histological differentiations, which take place in the rudimentary spinal column just described, are the same in all Vertebrata.

A long filament, composed of indifferent tissue, makes its appearance in the middle of the diaphysial plate, and constitutes the notochord, or chorda dorsalis.

Next, the substance of the paraphysial thickenings undergoes a certain change of tissue at regular intervals, so that they acquire a segmented appearance; solid, broad, darker masses of blastema lying opposite one another in each paraphysial thickening, and being separated by clear, narrow interspaces.

These segments are what the Germans term "Urwirbel," or "primitive vertebræ;" a somewhat misleading name, as they are in every way distinct from what are commonly understood under the name of "vertebræ," even if we use that word in its broadest signification. Professor Goodsir's terms of Somatomes for the segments and Metasomatomes for their interspaces, appear to me to be well worthy of adoption as the equivalents of these "Urwirbel."

3. The next step in the development of a vertebral column, is the histological differentiation of the somatomes. Leaving out of consideration the epithelial and other minor tissues, it may be said [580] that each somatome gives rise to (a) epiaxial muscles, (b) a nerve and its ganglion, (c) the blastema for a vertebral centrum and its neural and haemal arches, and (d) possibly hypaxial muscles; while the metasomatome becomes for the greater part of its extent an "intermuscular septum."

It is unnecessary for my present purpose to trace out particularly the development of any of these parts, except the centrum and its arches.

The blastema, which is specially intended for these parts, appears, in a distinct form, first, in the paraphysial thickenings, and then extends inwards above and below, so as gradually to enclose the notochord in a sheath, while, externally, it passes in the posterior half of each somatome, upwards into the neural arches, and downwards into the hæmal arches.

4. In some Vertebrata the spinal column never gets beyond this stage, nor even so far; but for the present it will be well to confine our attention to those which become completely ossified. In these chondrification is the next step. The blastema of the centra and its prolongations becomes converted into cartilage, but not continuously. On the contrary, at points corresponding with the intervals between every pair of metasomatomes, or with the middle of each somatome, the cartilage is replaced by more or less fibrous tissue. As a consequence, the cartilaginous sheath of the notochord is now divided into regular segments, which alternate with the somatomes, so that each metasomatome abuts upon the middle of one of these cartilaginous vertebral centra.

In every centrum it is necessary to distinguish three tracts or regions:–1. A diaphysial region immediately surrounding the notochord. 2. Two paraphysial regions lying in the paraphysial thickenings. The paraphysial regions give rise to the cartilaginous neural and hæmal semi-arcs, which are primitively continuous with them; so that all parts of the vertebra form one connected whole.

The neural semi-arcs eventually unite in the middle line, and ordinarily send a prolongation upwards from their junction. The hæmal semi-arcs also tend to unite below, but in a somewhat different manner.

5. The last step in the development of the vertebra is the differentiation of its various parts from one another, and their final metamorphosis into their adult form. The notochord, which primitively traversed the centra and the intercentra (intervertebral ligaments, synovial membranes, or the like, between the centra), becomes more or less completely obliterated.

[581] The distal, larger part of the hæmal semi-arc is commonly distinguished from its proximal smaller part, by the conversion of its cartilage into osseous or other tissue, and thus the semi-arc becomes separated into a rib and an articular surface or process, for the head of that rib, to which last the term Parapophysis may be conveniently restricted.

In the dorsal vertebræ of many Vertebrata, the neural semi-arc sends out a process, the Diapophysis, which is eventually met by a corresponding outgrowth of the rib, its so-called tubercle, and the two become firmly connected together.

When ossification occurs, it is a very general, if not invariable rule, that an annular deposit around the notochord takes place in the centrum. I term this the Diaphysis of the vertebra. In some fishes a distinct centre of ossification appears in each paraphysial region, and this may be termed the Paraphysis of the vertebra.

In mammals each end of the vertebra ossifies from a distinct point, and constitutes a central Epiphysis of the vertebra; and in many Vertebrata a part of the under surface of a centrum ossifies separately as a distinct Hypophysis. It is another very general, if not invariable rule, that a distinct centre of ossification appears in, or on, each neural semi-arc or Neurapophysis, and passes upwards, into the spine or Metaneurapophysis; downwards, to unite sooner or later with the diaphysis, or diaphysis and paraphysis; and outwards into the diapophysis.

It is doubtful whether the paraphysis appears as a distinct osseous element in any Vertebrata above the class of fishes, in very few of which even, is it distinguishable in the adult state. Consequently in the higher Vertebrata the paraphysial region is ossified, either from the diaphysis or from the neurapophysis, or from both; and a suture exists for a longer or shorter time at the point of junction of the neural and central ossifications. I will term this the Neurocentral suture. Its position is no certain or constant indication of the nature of the parts above or below it, for it may vary in the same vertebral column from the base of the neurapophysis, to the junction of the paraphysial with the diaphysial region of the centrum.

The number of the centres of ossification in each distal portion of the hæmal semi-arc may vary greatly; the uppermost is called a Pleurapophysis, the lower, Hæmapophyses and Met-hæmapophyses.

Besides these primary centres of ossification of a vertebra, there are others of less constancy. Thus the ends of the metaneurapophyses, diapophyses, and zygapophyses in many Mammalia are ossified from distinct centres; and in the caudal region of many of the higher [581] Vertebrata, outgrowths of the centra unite below to enclose the caudal vessels, and ossify as distinct apophyses.

If the development of the skull be now compared with that of the spinal column, it is found that (1) the very earliest changes undergone by the blastoderm in each are almost identical. The primitive groove extends to the extremity of the future cranial cavity; its lateral walls are continuous with the laminæ dorsales, and these pass into laminæ ventrales, also continuous with those of the spinal region. The laminæ dorsales of the head become the cranial walls and enclose the cerebrum–the continuation of the myelon; the laminæ ventrales give rise to the boundaries of the future buccal and pharyngeal cavities.

2. But at this point the identity of the skull with the spinal column ceases, and the very earliest steps in histological differentiation exhibit the fundamental differences between the two. For, in the first place, in no instance save the Amphioxus, has the notochord as yet been traced through the whole of the floor of the cranial cavity. In no other embryo has it been yet seen to extend beyond the middle vesicle of the cerebrum, or in other words, beyond the level of the rudiment of the infundibulum and pituitary body.

In the second place, the division into somatomes, in all known vertebrate embryos, stops short at the posterior boundary of the skull, and no trace of such segmentation has yet been observed in the head itself

3. Apparently as a consequence of these fundamental differences, the further course of the development of the skull is in many respects very different from that of a vertebral column. Chondrification takes place continuously on each side of the notochord, and beyond it, the two trabeculæ cranii, unlike anything in the spinal column, extend along the base of the cranium. No distinct cartilaginous centra, and consequently no intercentra, are ever developed. The occipital arch is developed in a manner remotely similar to that in which the neurapophysial processes are formed; but the walls of the auditory capsules, which lie in front of them, and which give rise to some of the parts, most confidently regarded as neurapophyses by the advocates of the current vertebral theories of the skull, are utterly unlike neurapophyses in their origin.

So, if we seek for hæmal semi-arcs, we find something very like them, arising from the substance of the basis cranii beneath the auditory cartilage; but there is none connected with the occipital cartilage, and none with the rudiment of the alisphenoid. The palatopterygoid cartilage might be regarded as the haemal semi-arc [583] of the presphenoidal region, though the grounds for so doing are not very strong; but the premaxillary cartilage is something quite without parallel in the spinal column.

4. The mode of ossification of the skull, and the ultimate arrangement of its distinct bony elements, are at once curiously like, and singularly unlike those presented by the spinal column. The basioccipital is ossified precisely after the manner of a vertebral centrum. Bony matter is deposited around the notochord, and gradually extends through the substance of the cartilaginous rudiment of the part.

The combined basi- and pre-sphenoid in Pisces and Amphibia is an ossific deposit, which takes place on the under surface of the basal cartilage in front of the basioccipital, and extends thence completely beneath the pituitary interspace as far as the ethmoid. It might be paralleled by the subchordal ossification in the coccyx of the frog, or by the cortical ossification of the atlas in many higher Vertebrata, if it really underlay a portion of the notochord; but at the very utmost the notochord only extends into its posterior extremity.

In some of the higher Vertebrata, as the snake, the osseous basisphenoid arises in the substance of its cartilaginous rudiment, while the osseous presphenoid underlies its cartilage. In others, both bones appear to arise directly in their cartilaginous forerunners. But nothing can be more irregular than the mode of ossification of the presphenoid, ethmoid and vomer in the vertebrate series, or less like the very constant and regular course of ossification of true vertebral centra.

With respect to the ossification of the lateral and superior constituents of the skull, the development of the exoccipital and supraoccipital does, without doubt, present a very close analogy to that of the separate pieces of the neural arch of some vertebrae in, e.g., a crocodile. The alisphenoids and orbitosphenoids follow in the train of the exoccipitals; but I know not where in the spinal column we are to find a parallel for the double parietals and frontals. But waiving this difficulty, and supposing, for the sake of argument, as was supposed by Oken, that the basisphenoid, alisphenoid and parietals, the presphenoid, orbitosphenoids, and frontals represent the elements of two vertebral centra and neural arches, what is to be made of the petrous and mastoid bones?

The difficulty has been eluded by terming the petrosal a "sense-capsule," the mastoid a "parapophysis." But I apprehend that neither of these explanations can be received for a moment by those who are acquainted with the development of the skull, or with the true homologues of the bones in question in the vertebrate series, or [584] who think that scientific terms should always possess a well-defined and single meaning.

What, in fact, is the origin of the petrous and mastoid bones? There is much reason for believing (according to Remak's late observations) that the membranous labyrinth is primarily an involution of the sensory or epidermic layer of the blastoderm; but however this may be, it is quite certain that the auditory organ is, primarily, altogether independent of the walls of the skull, and that it may be detached without causing any lesion of them, in young embryos.

It is also quite certain that this membranous labyrinth becomes invested by a coat of cartilage, continuous with the cranial wall; but I do not know that there is evidence, at present, to enable one to say positively, whether this cartilaginous auditory capsule is formed independently around the labyrinth, and then unites with the cranium; or whether it is an outgrowth from the cranial walls, which invests and encloses the labyrinth. If the latter be the case, a consistent vertebral theory of the skull must account for all the bones developed out of the auditory capsule; if the former, it must exclude them all, as parts of an extra-vertebral sensory skeleton.

Now the bones developed in the capsule are, in front, the petrosal; behind, the mastoid; above, the epiotic. The first-named bone is admitted, by the most zealous advocates of the vertebral theory, to be a neurapophysis, in all oviparous Vertebrata. Hence they are also bound to admit that, for three centra below and three neural spines bounding the cranial cavity above, there are four pairs of neural arches. More than this, I do not see how it is to be denied that the true mastoid is the morphological equivalent of the petrosal; and in that case there would be five neurapophyses to three central and three neural spines. Furthermore, it is precisely to these two superfluous elements that the only two clear and obvious hæmal arches, the mandibular and hyoid, are attached.

I confess I do not perceive how it is possible, fairly and consistently, to reconcile these facts with any existing theory of the vertebrate composition of the skull, except by drawing ad libitum upon the Deus ex machina of the speculator,–imaginary "confluences," "connations," "irrelative repetitions," and shiftings of position–by whose skilful application it would not be difficult to devise half a dozen very pretty vertebral theories, all equally true, in the course of a summer's day.

Those who, like myself, are unable to see the propriety and advantage of introducing into science any ideal conception, which is other than the simplest possible generalized expression of observed [585] facts, and who view with extreme aversion, any attempt to introduce the phraseology and mode of thought of an obsolete and scholastic realism into biology, will, I think, agree with me, not only in the negative conclusion, that the doctrine of the vertebral composition of the skull is not proven, but in the positive belief, that the relation of the skull to the spinal column is quite different from that of one part of the vertebral column to another.

The fallacy involved in the vertebral theory of the skull is like that which, before Von Bär, infested our notions of the relations between fishes and mammals. The mammal was imagined to be a modified fish, whereas, in truth, fish and mammal start from a common point, and each follows its own road thence. So I conceive what the facts teach us is this:–the spinal column and the skull start from the same primitive condition–a common central plate with its laming dorsales and ventrales–whence they immediately begin to diverge.

The spinal column in all cases becomes segmented into its somatomes; and, in the great majority of cases, distinct centra and intercentra are developed, enclosing the notochord more or less completely.

The cranium never becomes segmented into somatomes; distinct centra and intercentra, like those of the spinal column, are never developed in it. Much of the basis cranii lies beyond the notochord.

In the process of ossificatlon there is a certain analogy between the spinal column and the cranium, but that analogy becomes weaker and weaker as we proceed towards the anterior end of the skull.

Thus it may be right to say, that there is a primitive identity of structure between the spinal or vertebral column and the skull; but it is no more true that the adult skull is a modified vertebral column, than it would be, to affirm that the vertebral column is a modified skull.²²

While firmly entertaining this belief, however, I by no means wish to deny the interest and importance of inquiries into the analogies which obtain between the segments, which enter into the composition of the ossified cranium, and the vertebræ of an ossified spinal column. But all such inquiries must start with the recognition of the fundamental truths furnished by the study of development, which, as our knowledge at present stands, appear to me to be summed up in the following propositions:–

1. The notochord of the vertebrate embryo ends in that region of the basis cranii which ultimately lies behind the centre of the basisphenoid bone.

[586] 2. The basis cranii is never segmented.

3. The lamina perpendicularis of the ethmoid has the same morphological value as the presphenoid.

4. The petrosal has the same morphological value as the mastoid, if one is not an integral part of the skull, neither is the other.

5. The nasal bones are not neurapophyses.

6. The branchial arches have the same morphological value as the hyoid, and the latter as the mandibular arc.

7. The mandibular arc is primitively attached behind the point of exit from the skull, of the third division of the fifth nerve.

8. The premaxilla is originally totally distinct from the palatomaxillary arcade.

9. The pectoral arch is originally totally distinct from the skull.

Starting on this basis, it might not be difficult to show that the perfectly ossified skull is divisible into a series of segments, whose analogy with vertebræ is closer the nearer they lie to the occipital region; but the relation is an analogy and not an affinity, and these cephalic sclerotomes are not vertebræ.

¹ There is a wide difference, too, in the relative importance of either question to the student of comparative anatomy. Unless it can be shown that a general identity of construction pervades the multiform varieties of vertebrate skulls, a concise, uniform, anal consistent nomenclature becomes an impossibility, and the anatomist loses at one blow the most important of aids to memory, and the most influential of stimulants to research. The second question, on the other hand, though highly interesting, might be settled either way or the other without exerting any very important influence on the practice of comparative Anatomy.