low, cnidophorous or sting-cell-bearing tentacles found on the inner surface of the ovarian membrane of Medusa aurita and Lucernaria auricula; they are simply, as are the grapelike bodies, prolongations of the endoderm and gelatinous layer of the ovarian membrane.

Although the testicles of Chrysaora are apparently not homologous with those of other zoophytes, yet in reality they differ but little from those of Actinia and Lucernaria. I have given, in Plate IX. fig. 2, a section of the testicle of Chrysaora, and in fig. 3, of one of the same bodies in Actinia mesembryanthemum. In Chrysaora, the thin endoderm (a) forms the distant sperm-sacs which project from the surface. In Actinia, the thick endoderm (a) also forms the more closely aggregated sperm-sacs, and fills up the interstices. between them. The testicle of Lucernaria, again, resembles in shape and structure fig. 3; but the sperm-sacs are so closely moulded together, that they form hexagonal prisms divided from each other by exceedingly delicate walls of endoderm.

The sperm-sac of Chrysaora (fig. 4), as well as of other Steganophthalmatous Medusa, Lucernarias, and Actinias, is thus always formed of the endoderm or lining membrane of the digestive system, while the sperm-sac of Hydra (fig. 5), the Hydroid Polyps, and the Gymnopthalmatous Medusæ is formed of the ectoderm. In the first class of animals the spermatic cells (fig. 4) become first matured into spermatozoa in the centre (c), or at the base of the sperm-sac, the part most distant from the endoderm (a). In the second class they ripen at the periphery, or at the summit of the sperm-sac (fig. 5), the part also most distant from the endoderm (a).

My friend Mr Hincks, in his valuable paper on "Clavatella,"* appears to consider that the ova of that creature may be developed from the ectoderm. But an examination of the embryology of a very large number of zoophytes forbids me to entertain this idea. The endoderm of the generative capsule in these creatures consists of two layers intimately connected with each other. The external layer, or that in * Annals and Magazine of Natural History for February 1861.

tact with the generative elements, is transparent and structureless. The internal layer, communicating with the cavity. of the digestive system, is loaded with brown granules. In Coryne glandulosa, the ova are at an early period observed attached to the transparent layer of the endoderm, and separated from the ectoderm by a wide space of fluid. In Hydractinia, the reproductive polyps of which possess a muscular coat, that coat intervenes between the ova and the ectoderm.

In the subject of this paper, the ectoderm does not enter at all into the constitution of the sperm-sac. We may therefore conclude that the ova and spermatic plasma are detached or secreted from the external surface of the endoderm, which continues to convey nutriment to the former until they are fully developed.

EXPLANATION OF PLATE IX.

Fig. 1. Male organs of Chrysaora hyoscella: a, grape-like bodies dotted with sperm-sacs and attached to the ovarian membrane, b; cc, tentacular processes bearing tubercles and sperm-sacs.

Fig. 2. Section of tubercle bearing sperm-sacs, from the extremity of long oral tentacles: a, endoderm; b, "jelly;" c, ectoderm.

Fig. 3. Section of similar tubercle from Actinia mesembryanthemum, showing sperm-sacs formed by and imbedded in endoderm, a; b, interstitial tissue.

Fig. 4. Single sperm-sac of C. hyoscella: a, endoderm; b, unripe spermatic cells; c, spermatozoa; d, "jelly.”

Fig. 5. Sperm-sac of Hydra viridis: a, endoderm; d, ectoderm; b, unripe spermatic cells; c, spermatozoa bearing the same relations to the constituents of the sperm-sac as in fig. 4.

III. On the Serial Homologies of the Articular Surfaces of the Mammalian Axis, Atlas, and Occipital Bone. By JOHN CLELAND, M.D., Demonstrator of Anatomy in the University of Edinburgh.

In works on human anatomy it has been customary to compare the articular surfaces of the atlas, and the superior articular surfaces of the axis, with those of the oblique processes of other vertebræ, as if they were homologous, notwithstanding the apparently anomalous manner in which, according to that view, the first and second spinal nerves must be considered as emerging from the spinal canal. The

circumstances which have led to this comparison being made, are merely the rapid diminution in size of the intervertebral discs from the thoracic region up to the axis, and a general similarity of appearance between the articular surfaces of the atlas and axis and those of succeeding vertebræ; and though the impropriety of this comparison has been exposed in very explicit terms by Professor Henle,* there is still room for a few remarks as to the precise parts of other vertebræ to which the surfaces in question correspond.

In order to arrive at a just conclusion upon this subject, we shall find it advantageous to examine the atlas in the bird. In it we find on the posterior aspect a pair of true oblique processes passing backwards, to articulate above the intervertebral foramina with a corresponding pair of processes of the axis, similar to those of succeeding vertebræ ; while inferiorly there is a cartilaginous surface which forms, with the body of the axis and its odontoid process, a joint similar to those between the succeeding bodies of vertebræ. On the anterior aspect of the atlas there are no articular processes like the posterior pair; and there is presented for articulation with the condyle of the occipital bone, a single surface, exactly corresponding in extent with that which articulates with the body of the axis. As regards the occipital condyle, its constitution will be best understood by looking at the quite similar condyle of the occipital of the turtle. In it the middle and lower portions are formed by the basi-occipital, in precisely the same manner as the body of a vertebra is formed principally by the centrum, but has its superior angles derived from the arch. Thus there can be no doubt that the atlo-occipital articulation in birds, as well as the inferior atlo-axoid articulation, belongs to the same series as those between the bodies of the succeeding vertebræ.

It remains for us to show that they also correspond to the atlo-occipital and atlo-axoid articulations in mammals; and that they do so will readily appear, on making a more careful examination of the anterior articular surface of the atlas of the bird in the recent condition. It presents the form of

* Henle, Handbuch der Syst. Anat. des Menschen, i. p. 42.

*

a cup perforated by a small foramen, through which a ligament passes from the tip of the odontoid process to the occipital condyle, and the part of the cup which lies above the foramen is formed by a transverse ligament. This transverse ligament corresponds to those which pass from side to side of the bodies of other vertebræ, and are attached to the superior angles of their anterior aspects-those angles which are derived from the arches. Now, in mammalia, not only is the function of the transverse ligament of the atlas the same as in birds; but in many of them the heads of the ribs of opposite sides are united above the intervertebral discs by transverse ligaments (ligamenta conjugalia costarum), which very obviously correspond to the ligaments just mentioned on the vertebræ of the bird; for, though they do not, like them, pass from angle to angle of the bodies of the vertebræ, they are attached to structures interpolated between these angles. It appears, therefore, that the transverse ligaments of the atlas and other vertebræ in birds, and the ligamentum conjugale costarum, and transverse ligament of the atlas in mammals, are all homologous structures; and, in that case, the only difference between the atlo-occipital articulation in the mammal and in the bird is, that while in the latter it is single, in the former it is divided into two lateral parts. But this is not an important distinction; for in the atlo-axoid articulation, we find the arrangement in many mammals, as in the human subject, similar to that of the atlo-occipital; while in others, as in the sheep, a single joint extends across the middle line exactly as in the bird.

The serial correspondences of the vertebral articulations are very well illustrated in the human foetus. The articular surfaces of the oblique processes are situated immediately behind the transverse processes, and in the cervical region. the arches are bulged outwards at the points where they are placed (fig. 5). The axis is shaped altogether like one of

I have described and figured the ligament here referred to in a paper "On the Structure, Actions, and Morphological Relations of the Ligamentum Conjugale Costarum," in the Edinburgh New Philosophical Journal, April 1859.

the succeeding vertebræ, except only that the odontoid process is superadded to the centrum; and the bulging of the arch on each side behind the transverse process is well marked, and bears the inferior articular surface on its under side. On the other hand, the superior articular surface is placed partly on the odontoid process, but principally on the most anterior part of the arch, viz., that part which, in all the succeeding vertebræ, forms the posterior angle of the body (fig. 3). So also on the anterior extremity of the arch are placed the articular surfaces (both superior and inferior) of the atlas (fig. 4); and also, in the dorsal region, the surfaces for the heads of the ribs. The occipital condyles are placed upon the most anterior parts of the arch of the occipital bone, and to a small extent upon the centrum.

The foregoing examination of vertebral articulations leads us to observe, that, when surfaces for a synovial joint are present upon the body of a vertebra, however little of the body they may cover, they are never absent from those angles which are formed by the arches.

The synovial articulations between the bodies of vertebræ in mammals are arranged in the following manner: In the dorsal region are the synovial capsules for the heads of the ribs, which always occupy the angles of the bodies, but are also, in many animals, united across the middle line between the intervertebral disc and the conjugal ligament; while in some cases, as in the horse and the sheep, a small line of cartilage is stretched along the superior margin of the posterior of the two vertebræ concerned in each joint. In the cervical region in the human subject, the minute joints described by Luschka,* are situated between those parts of the bodies which are formed by the arches. Lastly, in the atloaxoid and atlo-occipital articulations, the principal parts of the articular surfaces are placed upon those parts of the arches which correspond to the angles of the bodies of succeeding vertebræ, while the intervertebral discs have disappeared.

I may here remark that, if the odontoid process be re*Luschka, Die Halbgelenke des Menschlichen Körpers, 1858, p. 71, and Tab. I., fig. 2.