12. Evidence of a CARNIVOROUS REPTILE (CYNODRACO MAJOR, OW.) about the size of a Lion, with Remarks thereon. By Prof. OWEN, C.B., F.R.S., F.G.S., &c. (Read February 2, 1876.)

[PLATE XI.]

SEARCHING Over the residuary, more or less shapeless blocks of matrix from the Karoo lacustrine deposits of South Africa, transmitted by their discoverer the late Andrew Geddes Bain, Esq., F.G.S., I came upon a lump about the size of one's fist, in which the sole indication of organic remains was a pair of mutilated canine-shaped teeth. Part of the outer enamelled surface was exposed to view, and suggested that the block might contain a fragment of the upper jaw of a Dicynodon. On clearing away the matrix from the teeth, however, they presented a difference of shape from the tusks in that genus: the crown was narrow instead of round; the transverse section was a long oval (Pl. XI. fig. 3) with the small end pointed: when the whole of the outer side of the crown became exposed, the shape of the best-preserved canine (ib. fig. 2) resembled that in Machairodus, and most so that of Machairodus latidens*. On carefully relieving the hinder trenchant margin of the tooth from the matrix, I was much interested in finding that it carried the resemblance to the canine of the extinct feline mammal to correspondence in the minute serration of that margin (ib. ib. c').

A small part only of the non-enamelled base was preserved, the canines having been broken off a little above their exit from the socket; but the breadth of the broken base and the reduction of the pulp-cavity to a linear trace were indicative of a long and deeply implanted fang. Nevertheless it was evident that, in proportion to the antero-posterior diameter of the base of the enamelled crown, that part of the tooth was relatively longer than in Machairodus latidens: it differed also in the absence of the serrate border at the fore part of the crown (ib. ib. c). The only part of the skull preserved posterior to that supporting the upper canines was a portion of the lower jaw, of which the alveolar border of the right ramus extended one inch behind the right better-preserved upper canine. trace of tooth could be detected in this border.

Νο

The symphysial part of the lower jaw (fig. 5) extended forward about 2 inches in advance of the upper canines (ib. c. c). The animal had become fossilized with its mouth shut; and the upper canines descended along a laterally compressed part of the lower jaw, also as in Machairodus, one on each side, with their points projecting beyond the lower border of the jaw (fig. 1, c, c). The next thought was as to possible evidences of the teeth at the fore end of the lower jaw; but not until the extremely hard matrix had been ground down to the alveolar border were any such traces visible. The crowns of these teeth seemed to have been broken away prior to interment; but the bases of eight incisors and of two lower canines were exposed, in transverse sections at the level stated. * British Fossil Mammals, 8vo, 1846, p. 180, fig. 69. Q. J. G. S. No. 126.

H

The lower canines (ib. fig. 5, c', c') had risen, as in Machairodus, immediately in front of the upper ones, and presented the same inferiority of size; but they are divided, in Cynodraco, by a toothless interval, or " diastema," from the lower incisors (ib. i, 1, 2, 3, 4). In this character, as in the number of incisors, the SouthAfrican Karoo fossil resembles the Marsupial genus Didelphis. The lower incisors are subequal, subcompressed, and elliptical in transverse section, at least at the base of the crown, with the long axis of the section directed from the fore to the hind part of the alveolar border; and they are close-set, as in carnivorous mammals. As in these, also, the dentine, in both canines and incisors, is of the hard unvascular kind, and the enamel as distinct in tissue and as thick. The decrease in size is from the 1st to the 4th; but the degree shown in the fossil and fig. 5 may be due to section at different heights from the base.

Associated with this fossil, or from near the same locality, was a larger oblong block of the same matrix, with the ends of a long bone partially visible. Out of this block an entire humerus was developed (Pl. XI. figs. 6-9, half nat. size). It is of a left fore limb, in length 10 inches 6 lines, with some loss by abrasion of both articular extremities, the shaft showing well-marked developments for muscular attachments and other characters unusual or unknown in the Reptilian class.

The breadth of the distal end-the extension of strong ridges from both the outer (e, e') and inner (f, f') sides, just above the elbowjoint, indicative of strong supinators, flexors, and extensors of the forearm and paw-the modification of the articular surfaces of that end, better preserved than those above, for the combination of due attachment of two bones of the forearm with freedom of motion, not only in the bending and extending, but in rotating on each other, so that the paw could be turned "prone" and "supine," whereby its application as an instrument for seizing and lacerating is advantaged, add to this the structure, hitherto known only in the Mammalian class and preeminently in the feline family, of a defence of the main artery and nerve of the forearm from compression during the action of the above-named muscles by a strong bridge of bone (h) spanning across them, furthermore the extensive and powerful ridge (b, b') at the proximal half of the humerus for the attachment of arm-muscles, especially the deltoid,-the combination of these characteristics, which Cuvier dwells upon in contrasting the humerus of the feline and bovine mammals *, are here exemplified in a fossil homologue, from a formation of the Triassic or Permian division of geological time.

Extending, however, the comparison of the present humerus beyond the salient features above defined, the head of the bone (fig. 8) differs from that of the feline humerus in being broader transversely, instead of from before backward; the articular part is oblong and narrow, not hemispheroid nor nearly so convex; there is no clevation of an outer transverse tuberosity. The representative of the

*Ossemens Fossiles, vol. i. (4to, 1821) p. xlvii.

deltoid ridge (ib. b, b′), of similar basal extent to that in Felis, is more produced and is much thinner, extending as a broad plate of bone outward (radiad) and forward (thenad); it seems to be rather a development in those directions of the entire shaft than to be a superadded process of the shaft. The more rounded, thickened character which prevails through nearly the whole extent of the humerus of the feline as of most mammals, is here confined to a constriction barely an inch in extent between the subsidence of the deltoidal or delto-pectoral crest and the beginning of the neuro-arterial bridge (h) on the inner side, and of the supinator crest (e) on the outer side of the lower two fifths of the bone. The perforation itself (k) is more highly placed, is further from the articular surface for the ulna, than in Felis. The ulnar articulation or "trochlea" is less defined; the inner boundary ridge, so prominent in Felis, is here undeveloped. The convexity (fig. 6, g) for the radius, on the other hand, is relatively larger, and its ball advances further upon the fore part of the shaft, in the present fossil. In this, likewise, the anconal pit (fig. 7, d), so deep and well-defined by outer and inner ridges in Felis, is a mere wide and shallow triangular depression.

The delto-pectoral crest (b, b') which seems to be a forward and inward production of the outer border of the proximal part of the humerus, is essentially a reptilian character of the bone. Its ordinary proportions in existing Saurians are shown in Uromastix spinipes (fig. 10, b, b') and in Monitor niloticus; in Crocodiles the shortening of its base gives it more the character of a distinct process; in Pterodactyles its development is chiefly in transverse extent; in Cynodraco the longitudinal extent prevails; in Omosaurus the development in both directions of the deltopectoral crest is such as to have suggested the generic name of this Dinosaur. But the perforation or canal (h, k, fig. 6) is not present in any existing Reptile. There is, indeed, what may be loosely termed a "supracondyloid foramen" in several Reptilia. It is noticed in the osteology of the genus Trionyx; but its position in the humerus is defined :-"The bone is perforated from before backwards at the outer angle of the distal extremity". The homologous supracondyloid foramen is shown in the humerus of the Monitor niloticus, in the under-cited monograph (Note *), in pl. xvii. fig. 6, at e', where it perforates a low supinator crest. Its position in Uromastix spinipes is shown at l' and m in figs. 10 and 11 (Plate XI.). Dr. Günther notes the homologous foramen in the humerus of Testudo elephantopus and its allies (Testudo ephippium e.g.), and describes it as "the canal on the radial edge of the bone, close to the elbow-joint, perforating the substance of the bone from the front to the hinder side "§.

*Monogr. on Omosaurus, pl. xvii. fig. 6, b, and fig. 6'.

+ Monogr. on Cretaceous Pterodactyles, 4to, 1860, pl. iii. fig. 9, b. Descriptive Catalogue of the Osteological Series contained in the Museum of the Royal College of Surgeons of England,' 4to, 1853, p. 184, Nos. 942, 943.

§ Philos. Trans. 1875, p. 266, plate xlii.; "b, radial canal for blood-vessels," ib. p. 283.

Cuvier thus defines the supracondyloid canal in the Lion :—“ Audessus du condyle interne, la ligne âpre est aussi percée d'un trou pour le passage de l'artère cubitale "*. But this is not the homologue of the commonly called "supracondyloid canal" in the humerus of certain Reptilia. The "internal condyle" of Cuvier and anthropotomy is the ulnar condyle;" the "external condyle" is that which is termed "radial" in vertebrate anatomy. Thus the zootomist has to take into account, in the application of his science to Palæontology, of an "ulnar supracondyloid foramen," and a "radial supracondyloid foramen." There is, also, a third perforation of the distal end of the humerus distinct from both, which may be termed an "inter-condyloid foramen." It is that which is present in the humerus of the wolf and some other mammals, but of which I have not found any example in recent or extinct Reptilia. While upon this comparison I am tempted further to remark that, in comparative anatomy, zootomy, or anatomy properly so called, the term "condyle" is restricted to portions of bone more or less convex, modified for the articulation of two bones, and usually covered with synovial cartilage. Anthropotomy, however, in reference to such portions of the distal articular surface of the human humerus, calls the radial condyle the "radial head," and the ulnar condyle the "ulnar trochlea," and restricts the term "condyle" to the tuberosities which project beyond and somewhat above the articular prominences, which, in the humerus, answer to the corresponding prominences rightly termed "condyles" in the femur. I have found it useful, in comparisons of the humerus akin to those in the present paper, to call the anthropotomical condyles "epicondyles," as being parts projecting somewhat above, or proximad of, the true condyles, and to distinguish them as "entepicondyle" and "ectepicondyle" respectively. Thus the canal which gives passage to an artery and commonly a nerve at the distal end of the humerus in Felines, and most Marsupials, is an entepicondylar canal or foramen; and this it is which characterizes the humerus of Cynodraco. The canal which gives passage to a blood-vessel in the humerus of certain Chelonia and Lacertilia is an ectepicondylar canal; and its presence in no way affects the resemblance (I will not say affinity) to the feline Mammalia which the extinct Cynodraco presents in its

humerus as in its dentition.

6

I am of opinion, though it is difficult to judge from the woodcuts of small and fragmentary humeri ascribed to Dicynodon by Prof. Huxley in the Memoirs of the Geological Survey of India,' that the "supracondyloid foramen" there indicated at a, figs A and B, p. 10, is not the homologue of the "supracondyloid foramen which occurs not unfrequently among Lacertian Reptiles," but that it differs not only in "form and proportions," but likewise in relative position, and that in Dicynodon it concurs with the pair of

*Leçons d'Anatomie comparée,' ed. 1835, vol. i. p. 384.

+ British Fossil Mammals,' 8vo, 1846, p. 129, fig. 47, a.

See for example, the instructive plate "77, left human humerus," p. 92, of the Anatomy Descriptive and Surgical,' Svo, 1858, by Henry Gray, F.R.S.

long upper tusks in exemplifying Mammalian characters, not Lacertian ones. The same concurrence, not only with well-defined canines, but likewise with equally well-defined incisors and molars, has been observed by me in several species of extinct Reptilia from the Karoo deposits of South Africa.

The earliest examples (Galesaurus, Cynochampsa) of such Reptilia were made known through this Society by its Quarterly Journal,’ vol. xvi. (for the year 1859-60). Since that date I have had evidences of the genera Lycosaurus, Tigrisuchus, Cynosuchus, Nythosaurus, Scaloposaurus, Procolophon, Gorgonops, as well as of the genus Cynodraco, of one species of which the dental and humeral characters have supplied the chief subject of the present communication.

Most of the genera above mentioned are represented by more than one species and both genera and species are classifiable in groups characterized by modifications of the structure of the external bony nostrils-as, for example, into the " tectinarial," " binarial," and "mononarial" families, the skull in the latter presenting, in the aspect and position of the single terminal nostril, a strikingly mammalian facies. For the name of these extinct carnivorous Saurians I find it convenient, and believe it will be generally acceptable, to form a distinct order of Reptilia under the denomination of Theriodontia, with the following characters :-" Dentition of the carnivorous type; incisors defined by position and divided from the molars by a large laniariform canine on each side of both upper and lower jaws, the lower canine crossing in front of the upper; no ectopterygoids; humerus with an entepicondylar foramen; digital formula of fore foot 2, 3, 3, 3, 3 phalanges."

When we contrast the grand development and notable modifications of the extinct species of Reptilia with the poor scanty evidences of the class still lingering on in life, we seem to be witnessing a course of degeneration, of retrogradation, rather than of organic progression in the course of time.

Although there be none now, there have been Saurians in which the articular surfaces of certain vertebræ in the same column were modified-as, e. g., of the anterior ones in great herbivorous Dinosaurs, for freer movements of the head, by means of an anterior ball playing in a posterior cup, such as we now find in the Rhinocercs and some other large mammalian herbivores. In another part of the same vertebral column of these extinct reptiles we find a sacrum not limited to two, but composed of five or six ankylosed vertebræ, also as in these and many other mammals.

Again, some of these vegetarian Dinosaurs had masticatory teeth of a complex structure + unknown in existing phytophagous Sauria, but resembling those teeth of certain phytophagous mammals, e. g. Megatherium, Mylodon §. With the modification of a portion of the

* These will be characterized and the characters illustrated in the Plates of the Catalogue of South-African Reptilia, on which I am now engaged by direction of the Trustees of the British Museum.

+ Odontography,' p. 250, pl. 71: 4to, 1840.

+ Ibid. pl. 84.

§ Ibid. pl. 79.