SKELETON.] REPTILES 449 In addition to the ordinary articular processes (or zygapo- physes), which are here broad and flattened, often with an accessory process, there are two noteworthy structures. The first of these is called the zygosphene, and consists of a wedge-shaped process with two articular surfaces, which projects forward from the anterior surface of each neural arch. The other is called the zyg- ns ns antrum, and is a corresponding excavation with two articular sur- faces on the hinder side of the neural arch, and receives the zygo- sphene of the vertebra next behind it. It was structures such as these that were referred to in the last paragraph as existing so exception- trunk vertebra o ally amongst Lacertilians in Iguana, nature), a. articularprocesses; mi , i 1-1 na, neural arches; ns, neural I he atlas and axis have a form spines; i.transverseprocesses; similar to that already described zs > zygosphene. as existing in the Lacertilia. Long azygous hypapophyses nearly always depend from the centre of the anterior trunk vertebrae, or even from the whole of them, being especially large and numerous in some of the most poisonous Serpents, e.g., Crolalus and Naja. In Dasypeltis scaber some of these processes in the ante- rior dorsal region are made to minister to alimentation, their tips being COated with tooth Sub- Fl - 7 Posterior aspect of a , . . , . , trunk vertebra of Python (from Stance, and penetrating Within the nature), a, articular processes; nliinpntirv pannl n will VIP no-nin 6 ' bal1 on tlle sul 'f ace of the alimentary canal, as win >e again centrum . , Wi neural arch . nSj mentioned further On. neural spine t, transverse pro- mi [ j-i f L' c cess; zg zygantrum. Ihe ribs assume the function of locomotion, and are therefore very movably articulated to short transverse processes. Each has two proximal articu- lar surfaces, but they are close to- gether, there being no diverging articular and tubercular processes. Each rib terminates in a short car- tilage. The caudal vertebrae may, in rare instances, be very few, and they may be very numerous from five to two hundred. They do not FIG. 8. Anterior aspect of s, but bifold depending hypapophyses, and they possess Chevron bones, but bifold trunk vertebra of PythonQram nature), a, articularprocesses; cup on the surface of the 835? WSSESfJZ cess : ". zygosphene, have transverse processes which also are generally bent downwards. Sometimes, at the limit between the body and "the tail, the transverse process, or the rib, as the case may be, bifurcates or develops an ascending process, as, e.g., in Boa, Naja, and Echidna. No Ophidian has any sternum. Ckelonia. The Tortoises and Turtles present a spinal structure remarkably divergent from that of all other Reptiles, and especially divergent from that of Ophidians. Nevertheless the Chelonia agree with the Ophidia, in having no sternum, and in having at least a portion of the vertebral column formed for extreme mobility, in spite of the excessive rigidity and immobility of the trunk. Only in Sphargis do we find a structure generally resembling that of other Reptiles and diverging from that common to other Chelonians. With the exception of Sphargis, the structure of the vertebral column of which will be sub- sequently noticed, the Chelonians present the following characters. The endoskeleton of the dorsal region is intimately united with those ossified exoskeletal plates which have been already described (supra, p. 447) as investing both dorsally and ventrally the entire trunk. The median series of plates are united with the neural spines and arches, and the lateral plates are similarly united with the ribs, so that the carapace is formed both of endoskeletal and exoskeletal elements intimately united. The constancy of the number of vertebrae is very excep- tional, as is a'lso the very small number of those of the body. There are always 8 cervical, 12 dorsal, and 2 sacral vertebrae. The number of caudal vertebrae varies, but is never great. The cervical vertebrae have very small processes, whereby they are better fitted for great mobility. They vary greatly as to the form of their centra, some being opisthocoelous, others procoelous, others biconvex, while one at least is flattened both in front and behind the arrangement differing even in different species of the same genus. The atlas consists of three pieces, and the os odontoideum is separate. The neuro-central suture per- sists. The centra of the first ten dorsal vertebra are amphicoelous, and but loosely connected with twin neural arches, each neural arch being superimposed over the posterior half of one vertebra and the anterior half of the vertebra next behind. The sacral vertebrae have either stout ribs suturally attached or transverse processes anchy- losed to the vertebrae. The caudal vertebrae are proccelous, and generally have a pair of separate descending processes, which may (as in Chdydra) unite distally and form chevron bones. In Sphargis the endoskeleton is quite distinct from the exoskeleton. Its neural arches are very loosely united to the centra, so that they can be separated with the greatest ease. The trans- verse processes of the caudal vertebrae also remain unan- chylosed to their centra. The dorsal vertebrae pass so gradually into the caudal that the boundary can only be determined by p JG _ g_ Three vertebra! of Sphargis coriacea (from the Somewhat in- nature), c, vertebral centra; n, neural arches; r, ribs. creased thickness of the transverse process of the two sacral vertebrae. The summits of the neural arches are somewhat dilated (as if from superincumbent pressure) or obtusely keeled. The ribs are not much expanded, their heads fitting into a hollow formed by the centra of two vertebrae together with the neural arch superimposed upon the two. They become shorter and narrower behind, that of the tenth dorsal vertebra being even smaller than the sacral ribs. The Skull. The anterior portion of the axial skeleton, The or cranium, differs markedly from that of every other skull; class of Vertebrates, while it presents a much greater diversity of structure than does the cranium of Mammals or of Birds. It differs from the cranium of the lower Vertebrate classes by its more complete ossification and by a greater prolongation backwards of the^ nasal cavity, which causes a greater expanse and density of the palatal structure. The skull, however, is not so much ossified as in Mammals, and the bones do not become anchylosed together so quickly and certainly as in Birds. The occipital region always consists of four occipital elements, and there is an ossified basisphenoid in front of a well- developed basioccipital. The skull almost always arti- culates with the atlas by a single occipital condyle. The lower jaw is always suspended to the skull by the inter- vention of a quadrate bone with which the os articulare of the mandible articulates. Besides the last-mentioned bone the mandible consists of as many bones as in Fishes, with the addition of a supra-angular and a complementary bone. The bones of the cranium may form a very solid and XX. - 57
