slowly, so that settlers, knowing its utility, have carried it in boxes and bottles to their new inland stations."

It must be long before facts enough to theorize upon can be collected. Meanwhile, the inquiry appears to be perhaps the most interesting and important in all Biology, and as such it is most earnestly to be desired that all who are favourably circumstanced to pursue it, will do so both systematically and very carefully.

XXI.-ON THE DEVELOPMENT OF THE CRANIUM IN THE VERTEBRATA. By Prof. H. Rathke.*

[Concluded from Vol. III. p. 251.]

In my Development of Blennius viviparus' (Abh. z. Bildungs-und Entwickelungs-geschichte § 22) I have stated that the sheath of the cephalic part of the notochord (or, to speak more correctly, the investing mass thereof) divides itself into three successive portions. This error probably arose from the circumstance that so minute an object, if it be not kept sufficiently moist under the microscope, dries rapidly, and that its central part breaks up into two or three portions, so as to give rise to the appearance of articulation.

Perhaps in no fish does the bony plate referred to above, which I regard as the representative of the presphenoid of higher animals, attain so large a relative size as in the Sturgeons. In these it extends far back, so as to cover, not merely the chondrified investing mass of the notochord, but also the bodies of the four or five anterior vertebræ. If we consider this bone (which arises, not by the modification and higher development of a portion of the chondrified investing mass of the cephalic part of the notochord, but independently of it) in different animals we observe that it is but little governed by the investing mass and the bones which arise therefrom, for it extends sometimes for a greater, sometimes for a less distance backwards, and assumes, in comparison to the investing mass, a very various breadth and form. It may be added that, in the Sturgeon, the bodies of the most anterior vertebræ are not, as Joh. Müller states, divided below, and their lateral halves separated, by the bony plate in question; but it is the ribs of the anterior vertebræ, together with a corresponding number of small pieces of cartilage to which they are attached,

Translated from the original Memoir published at Königsberg, in 1839. See Nat. Hist. Rev. 1863, p. 234.

(and which perhaps may be regarded as transverse processes), which are separated by that piece of bone from their centres, so as to be connected with the latter only by fibrous tissue.

§ 16. The simplest skull at present known among the Vertebrata is that of Ammocætes branchialis; and its form and composition agree in the most surprising manner with the first rudiments of the skull of other vertebrates; so that, in this fish, the skull, like the rest of the skeleton, exhibits, in a persistent form, a structure which belongs to the earliest period of development of other members of the same sub-kingdom.

The anterior pointed end of the notochord extends almost to the anterior edge of the auditory capsules. On each side of it, a short distance behind these capsules, begin two thin bands, consisting of a solid mass of cartilage, which, closely applied to the cephalic part of the notochord, approach, anteriorly, so as to be nearest opposite its point; and then, as they pass forward, diverge considerably and finally arch round towards one another, so as to meet and coalesce at the anterior end of the cranial cavity. From the arch, a very short, though tolerably broad and moderately thick, median process passes forward in the middle line of the head and is rounded off anteriorly. Johannes Müller has already described and accurately figured the thin cartilaginous bands in his work upon the Myxinoids (Part I., p. 117 and 118, and Pl. iv. fig. 7 and 8.).

That part of them which extends anteriorly, beyond the notochord, plainly answers to those structures in other Vertebrata which I have termed the trabeculæ cranii. Hereafter, I shall endeavour to explain the nature of the part which lies at the sides of the notochord. Furthermore, the relatively very large space, which is enclosed by the cartilaginous bands in front of the notochord, I found to be filled by a plate, the entire edge of which is united with the bands; it is thinner behind than in front, but even in front it appears to be somewhat thinner than the bands. Under the microscope it exhibits no trace of cartilage corpuscles, but is quite membranous, or rather, consists of condensed connective tissue. Partly on this ground, partly by reason of its position and connexions, it corresponds with the substance which fills up the space between the trabecula in the embryos of the higher Vertebrata. Still thinner than the plates just described, are, with the exception of the very dense cartilaginous auditory capsules, the lateral walls and the roof of the cerebral capsule, for they are constituted only by a very delicate membrane, united

with the 'bands' and with the auditory capsules. Ammocætes presents not the slightest trace of facial elements.

The nature of the posterior parts of the cartilaginous bands must, I conceive, be interpreted as follows. In the embryos of the higher vertebrates, the investing mass of the larger portion of the notochord, or of that part which lies behind the head, is first deposited upon each side of the notochord and represents, chiefly, the first rudiments of the bodies of the vertebræ. This mass is the first to chondrify; and it would seem that the investing mass of the cephalic part of the notochord conducts itself in a similar manner. This mass is at first deposited on each side of the notochord, but then grows upwards and downwards, around the cephalic part, until the halves coalesce.

My investigations into the development of the cranium of different Vertebrata have not enabled me to speak decidedly upon this point, but the circumstance that the investing mass is very early accumulated at the two sides of the cephalic part of the notochord, is in favour of this view. But if this conception of the mode of origin of the investing mass of the cephalic part of notochord is correct, the above-mentioned posterior halves of the two cartilaginous bands; that is, those parts of them between which, in Ammocætes, the cephalic notochord lies, correspond with the first rudiments of that part of the investing mass of the cephalic portion of the notochord of other vertebrates, in which the basi-occipital bone is developed.

The development of the skull of Petromyzon has advanced far beyond that of Ammocates. The cephalic portion of the notochord, which, in the Lampreys, also extends to between the auditory capsules, is already surrounded by a cartilaginous mass, as by a sheath, and this sends upwards the processes which lie behind the auditory capsules and correspond with the exoccipitals of other animals. Anteriorly, however, it is prolonged, as a flattened horizontal plate, as far forwards as the infundibulum. This plate, again, is continued forwards by two moderately distinct flat, narrow and short cartilaginous processes, which answer to the trabecule of other Vertebrata, and pass, anteriorly, into a great shield-shaped plate of cartilage which is strongly arched and has its convex side directed upwards. Obviously, this plate is a greater development of the very short and moderately broad process which occurs in Ammocætes, formed by the coalescing of the two trabeculæ anteriorly, and without doubt it corresponds with the lateral part of the ethmoid of other vertebrates.

It is a very remarkable circumstance, however, that the olfactory

N. H. R.-1864.

K

organ of the Cyclostomes is not double but single and median, and that it is arranged, not at the two sides of the part analogous to the ethmoid of other animals, but, upon it, in the middle plane of the body. Consequently, the olfactory organ, becoming as it does greatly elongated backwards, not only in Petromyzon but in Ammocætes and the other Cyclostomes, can find no other road (quite contrary to the rule which holds good for other vertebrates) than between the ethmoid and the brain, through the space which lies between the trabecula and which, in the Lampreys as in Ammocætes, is filled by a dense almost fibrous, connective tissue. The cartilages of the face, excepting the ethmoid, in the Cyclostomes in general, differ so widely in form, position and connexions, from the elements of the facial skeletons of other vertebrates that they seem to be constructed on a totally different principle from these.

§ 17. It is probably true for all Vertebrata, that at a certain period of fœtal life, a simple vesicle makes its appearance upon each side of the head, which communicates by a wide aperture with the third cerebral vesicle, and which may be regarded as an outgrowth of this vesicle. This aperture, however, soon narrows and the trunk of the auditory nerve becomes gradually evolved between the vesicle and the brain.

In order to receive this vesicle, which is the rudiment of the membraneous labyrinth of the internal ear, the gelatinous wall of the capsule, which incloses the brain, and which chiefly becomes developed into the brain case, bulges out and so forms a more or less extensive projection, a part of which becomes chondrified, so as to form a cartilaginous capsule widely open towards the cranial cavity. In many Fishes this auditory capsule, which, as it ossifies, becomes the petrous bone, undergoes but little modification of form and so appears, throughout life, only as a dilatation of the brain case.

In most Vertebrata, however, it contracts so strongly where it passes into the rest of the brain case that only one, or several, small apertures remain through which the trunk of the auditory nerve, or its branches, penetrate into the interior of the capsule. The thickness of the auditory capsule gradually increases more or less, and its substance grows between the different parts of the membranous labyrinth as these develop, to a greater or less extent; so that the intervals between the parts of the labyrinth, in many animals, (as Birds and Mammalia) become almost entirely filled up, whilst in others, as in many osseous Fishes, there is hardly a trace of any such process.

The ossification of the auditory capsule proceeds in many animals from several points, from which a corresponding number of bones arise. I have convinced myself of this in the Snake, the Lizard and the Chick, in which three such parts are developed in each auditory capsule. In the Snake, the separate ossifications do not coalesce, but two unite with the adjacent bones, viz. one with the supra-occipital, the other with the exoccipital. The same thing occurs, at first, in the Chick, but, at a subsequent period, the three pieces of the petrosal coalesce with one another. Resting partly upon these observations, partly upon the well known fact, that in bony Fishes, Chelonia and Sauria, single constituents of the membranous labyrinth lie inclosed in yet other cranial bones than in those which are usually alone held to be elements of the petrosal, I am led to believe that in the animals just mentioned, the petrosal is far more complex than is usually believed, and that it consists in them of many pieces and not of one. In osseous Fishes, especially, all those are to be reckoned as constituents of the petrosal which inclose the membraneous vestibule and the semicircular canals.

§ 18. As I have already observed (§ 3.) the investing mass of the notochord grows, in vertebrate animals in general, into the two dorsal laminæ, so that, after a time, a part of it forms, in each dorsal lamina a band, which extends throughout the entire length of the future spinal marrow. By partial thickening of this band the rudiments of the crura of the vertebral arches are originated; and when, subsequently, the investing mass of the notochord chondrifies, the process (if we leave the Cyclostomes out of consideration) extends upwards, without interruption, into the thickened parts of the band, so that the parts which are developed out of these portions-the crura of the vertebral arches - appear, even when they are chondrified to be immediate outgrowths of the vertebral centra.

Strictly speaking, the two bands described are continued still further forwards, so that they may be said to end only in the head. How far they extend in different animals remains yet to be seen. Assuredly, however, the exoccipitals arise from them in the same way as the arches of the vertebræ do, so that when they are chondrified they appear as outgrowths of the basi-occipital cartilage. But nothing can be said, with certainty, as to whether the substance out of which the auditory capsules and the alisphenoids are developed are parts of those bands (and in fact their anterior terminations) or not. All appearances are in favour of supposing that the auditory capsules